NEUROSCIENCE 2016

THE JOURNAL OF EXPERIMENTAL MEDICINE

SELECTED ARTICLES NOVEMBER 2016 www.jem.org

World-Class Quality | Superior Customer Support | Outstanding Value

Toll-Free Tel: (US & Canada): 1.877.BIOLEGEND (246.5343)Tel: 858.768.5800

biolegend.com

08-0060-12

Learn more at: biolegend.com/neuroscience/jem

With the recent acquisition of the Covance legacy antibodies, BioLegend has expanded its portfolio to off er high quality and specifi city antibodies for neuroscience research. We off er a great selection of human- and rodent-specifi c antibodies provided in multiple formats for use in research applications such IHC, IF, WB, and more. BioLegend supports research within multiple neuroscience areas with a great focus on the fi eld of neurodegeneration.

Remember BioLegend as your Neuroscience Antibodies source

BioLegend is ISO 9001:2008 and ISO 13485:2003 Certifi ed

IHC staining of amyloid beta plaques with anti-β-Amyloid antibody (clone 6E10) conjugated to Alexa Fluor® 488 on FFPE human AD brain.

IHC staining of anti-GFAP antibody (clone MCA-5C10) on FFPE rat brain tissue. Nuclei were counterstained with Hoechst and are shown in blue.

β-Amyloid , 1-16 GFAP

A dural lymphatic vascular system that drains brain interstitial fluid and macromoleculesAleksanteri Aspelund, Salli Antila, Steven T. Proulx,Tine Veronica Karlsen, Sinem Karaman, Michael Detmar, Helge Wiig, and Kari Alitalo

Brain endothelial TAK1 and NEMO safeguard the neurovascular unitDirk A. Ridder, Jan Wenzel, Kristin Müller, Kathrin Töllner, Xin-Kang Tong, Julian C. Assmann, Stijn Stroobants, Tobias Weber, Cristina Niturad, Lisanne Fischer, Beate Lembrich, Hartwig Wolburg, Marilyn Grand’Maison, Panayiota Papadopoulos, Eva Korpos, Francois Truchetet, Dirk Rades, Lydia M. Sorokin, Marc Schmidt-Supprian, Barry J. Bedell, Manolis Pasparakis, Detlef Balschun, Rudi D’Hooge, Wolfgang Löscher, Edith Hamel, and Markus Schwaninger

Induced knockouts provide insights into human L1 syndromeJan Pielage

Conditional deletion of L1CAM in human neurons impairs both axonal and dendritic arborization and action potential generationChristopher Patzke, Claudio Acuna, Louise R. Giam, Marius Wernig, and Thomas C. Südhof

Changes in insulin and insulin signaling in Alzheimer’s disease: cause or consequence?Molly Stanley, Shannon L. Macauley, and David M. Holtzman

TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer’s disease mouse modelsTaylor R. Jay, Crystal M. Miller, Paul J. Cheng, Leah C. Graham, Shane Bemiller, Margaret L. Broihier, Guixiang Xu, Daniel Margevicius, J. Colleen Karlo, Gregory L. Sousa, Anne C. Cotleur, Oleg Butovsky, Lynn Bekris, Susan M. Staugaitis, James B. Leverenz, Sanjay W. Pimplikar, Gary E. Landreth, Gareth R. Howell, Richard M. Ransohoff, and Bruce T. Lamb

Cutting to the chase: How pathogenic mutations cause Alzheimer’sMichael S. Wolfe

Qualitative changes in human g-secretase underlie familial Alzheimer’s diseaseMaria Szaruga, Sarah Veugelen, Manasi Benurwar, Sam Lismont, Diego Sepulveda-Falla, Alberto Lleo, Natalie S. Ryan, Tammaryn Lashley, Nick C. Fox, Shigeo Murayama, Harrie Gijsen, Bart De Strooper, and Lucía Chávez-Gutiérrez

Neutrophil-related factors as biomarkers in EAE and MSJulie M. Rumble, Amanda K. Huber, Gurumoorthy Krishnamoorthy, Ashok Srinivasan, David A. Giles, Xu Zhang, Lu Wang, and Benjamin M. Segal

Therapeutic targeting of autophagy in neurodegenerative and infectious diseasesDavid C. Rubinsztein, Carla F. Bento, and Vojo Deretic

Selected Art ic les November 2016

®

Quantify with Confidence

908-231-0960 hamamatsucameras.com

Ever wonder about the quantitative nature of image data? We do. Introducing the new ORCA-Flash4.0 V3 and the W-VIEW GEMINI-2C dual channel, dual camera optics.

E N G I N E E R E D F O R S U P E R R E S O L U T I O N — R E A D Y F O R A N Y T H I N G .

THE JOURNAL OF EXPERIMENTAL MEDICINEExecutive Editor Teodoro Pulvirentiphone (212) 327-8575 email: jem@rockefeller.edu

Editors Jean-Laurent CasanovaDouglas T. FearonDavid HoltzmanLewis L. LanierWilliam A. MullerCarl NathanMichel NussenzweigAnne O’GarraAlexander RudenskyAlan SherSasha TarakhovskyAndreas TrumppDavid Tuveson

Senior Scientific Editors Thiago CarvalhoAlejo Chorny

Scientific Editor Zhe Wang

Editor Emeritus Alan N. Houghton

Managing Editor Sylvia F. Cuadradophone (212) 327-8575 email: jem@rockefeller.edu

Advisory EditorsShizuo AkiraKari AlitaloFrederick W. AltK. Frank AustenAlbert BendelacMichael J. BevanChristine A. BironChristian BogdanHal E. BroxmeyerMeinrad BusslingerArturo CasadevallAjay ChawlaYongwon ChoiRobert L. CoffmanDaniel J. CuaMyron I. CybulskyRiccardo Dalla FaveraGlenn DranoffMichael DustinVincent A. FischettiRichard A. FlavellAdolfo Garcia-SastrePatricia GearhartRonald N. GermainChristopher GoodnowSiamon GordonOr GozaniSergio GrinsteinPhilippe GrosKristian HelinChyi HsiehChristopher A. HunterKayo InabaGerard KarsentyJay KollsPaul Kubes

Vijay K. KuchrooRalf KuppersTomohiro KurosakiBart N. LambrechtKlaus F. LeyYong-Jun LiuClare LloydTak MakBernard MalissenJames S. MalterPhilippa MarrackDiane MathisIra MellmanMatthias MerkenschlagerSean J. MorrisonMuriel MoserChristian MünzCornelis MurreBenjamin G. NeelVictor NussenzweigJohn J. O’SheaPaul H. PattersonFiona PowrieLluis Quintana-MurciKlaus RajewskyGwendalyn J. RandolphJeffrey RavetchSergio RomagnaniNikolaus RomaniDavid L. SacksShimon SakaguchiMatthew D. ScharffOlaf SchneewindStephen P. SchoenbergerHans SchreiberGerold SchulerRobert A. Seder

Rafick-P. SékalyCharles N. SerhanNilabh ShastriEthan M. ShevachRoy L. SilversteinJonathan SprentJanet StavnezerAndreas StrasserStuart TangyeSteven L. TeitelbaumThomas J. TempletonKevin J. TraceyGiorgio TrinchieriShannon TurleyMarcel R.M. van den BrinkUlrich von AndrianHarald von BoehmerChristopher M. WalkerRaymond M. WelshE. John WherryLinda S. WickerIan WilsonThomas Wynn

Monitoring EditorsMarco ColonnaJason CysterStephen HedrickKristin A. HogquistAndrew McMichaelLuigi NotarangeloAnjana RaoFederica SallustoLouis M. StaudtToshio Suda

Print ISSN 0022-1007 Online ISSN 1540-9538

Copyright to articles published in this journal is held by the authors. Articles are published by The Rockefeller University Press under license from the authors. Conditions for reuse of the articles by third parties are listed at http://www.rupress.org/terms

Preflight Editor Rochelle Ritacco

Copy Editor Libby Goss

Asst. Production Editor Shauna O’Garro

Senior Production Editor Brianna Caszatt

Production Manager Camille Clowery

Production Designer Erinn A. Grady

950 N. Hague AveColumbus, OH 43204 U.S.A.

Phone: [614] 276.0861 Email: sales@colinst.com

www.colinst.comMetabolic Phenotyping Monitor: Oxymax CLAMS-HC

The Columbus Instruments Oxymax - CLAMS-HC (Comprehensive Lab Animal Monitoring System for Home Cages) is a versatile device for monitoring metabolic performance of mice and rats in their natural home cage environment.

Animal Activity MonitorThe Columbus Instruments Auto-Track Activity Meter presents the for measuring in home or special cages. Measures these parameters:

Distance Traveled

Animal TreadmillThe Exer 3/6 Treadmill provides 6 mouse lanes or 3 rat lanes for general purpose exercise. Adjustable speed, acceleration, and incline on all models. Available with or without electric stimulus or our

The Rotamex-5 measures coordination in up to four mice or rats by recording the latency to fall from a spinning rod. Key features include:

NEW

The loaded running wheel applys a precise resistance during voluntary running for the purpose of increasing the load and measuring Work.

0.00000038 Nm increments.

NEW

NEW

2 2

and LED light color control

For more information: clams@colinst.com

Neuroscience

1435

www.avantilipids.com Phone: 800-227-0651 Email: info@avantilipids.com

Check outour NEWWebsite!

AVANTI’S NEW LABELED NEUROTRANSMITTERS

AEA-d4 Avanti Number 857465

NBD AEA Avanti Number 810107

AEA-d11 Avanti Number 857466

More Endocannabinoids in stock at Avanti

NBD AA Avanti Number 810106

arachidonic acid-d11 Avanti Number 861810

2-AG Avanti Number 8704502-OG Avanti Number 870451 arachidonic acid-alkyne Avanti Number 900411C17:1 anandamide Avanti Number 870431 C18:1 anandamide Avanti Number 870432 C20:4 anandamide (AEA) Avanti Number 870430 C16:0 anandamide phosphate Avanti Number 870436C18:1 anandamide phosphate Avanti Number 870438C18:2 anandamide phosphate Avanti Number 870439

C20:4 anandamide phosphate Avanti Number 87044018:1 PE-N-20:4 Avanti Number 872100 16:0-20:4 PE-N-20:4 Avanti Number 872101 N-16:0 L-Serine Avanti Number 800740 N-16:0 L-Serine MeEster Avanti Number 800730 N-18:1 L-Serine Avanti Number 800741 N-18:1 L-Serine MeEster Avanti Number 800731 N-20:4 L-Serine (ARA-S) Avanti Number 800742 N-20:4 L-Serine MeEster Avanti Number 800732

-

The Original Fluoromount-G®

Fluoromount-G® is a registered trademark of Southern Biotechnology Associates, Inc.

www.SouthernBiotech.com info@SouthernBiotech.com

SouthernBiotech is ISO 9001:2008 Certified

The Mounting Media for Your Neuroscience Research

www.keystonesymposia.org/meetings | 1.800.253.0685 | 1.970.262.1230

a 501(c)(3) nonprofit educational organization

Neurogenesis during Development and in the Adult Brain joint with Transcriptional and Epigenetic Control in Stem Cells Scientific Organizers: Alysson R. Muotri, Kinichi Nakashima and Xinyu Zhao January 8–12, 2017 | Olympic Valley, California | USAwww.keystonesymposia.org/17J1 | www.keystonesymposia.org/17J2Registration is still open and abstracts are still being accepted for poster presentation.

Epigenetics and Human Disease: Progress from Mechanisms to Therapeutics Scientific Organizers: Johnathan R. Whetstine, Jessica K. Tyler and Rab K. Prinjha January 29–February 2, 2017 | Seattle, Washington | USAwww.keystonesymposia.org/17A9 Discounted Registration Deadline – Nov 30, 2016 (Abstracts are still being accepted for poster presentation.)

Autophagy Network Integration in Health and Disease Scientific Organizers: Ivan Dikic, Katja Simon and J. Wade Harper February 12–16, 2017 | Copper Mountain, Colorado | USAwww.keystonesymposia.org/17B2Abstract Deadline– Nov 15, 2016; Discounted Registration Deadline – Dec 13, 2017

Rare and Undiagnosed Diseases: Discovery and Models of Precision Therapy Scientific Organizers: William A. Gahl and Christoph Klein March 5–8, 2017 | Boston, Massachusetts | USAwww.keystonesymposia.org/17C2Abstract Deadline – Dec 6, 2016; Discounted Registration Deadline – Jan 10, 2017

Synapses and Circuits: Formation, Function and Dysfunction Scientific Organizers: Tony Koleske, Yimin Zou, Kristin Scott and A. Kimberley McAllister joint with Connectomics Scientific Organizers: Olaf Sporns, Danielle Bassett and Jeremy Freeman March 5–8, 2017 | Santa Fe, New Mexico | USAwww.keystonesymposia.org/17X1 | www.keystonesymposia.org/17X2Abstract Deadline – Dec 7, 2016; Discounted Registration Deadline – Jan 11, 2017

mRNA Processing and Human Disease Scientific Organizers: James L. Manley, Siddhartha Mukherjee and Gideon Dreyfuss March 5–8, 2017 | Taos, New Mexico | USAwww.keystonesymposia.org/17C3Abstract Deadline – Dec 6, 2016; Discounted Registration Deadline – Jan 10, 2017

Neuronal Control of Appetite, Metabolism and Weight Scientific Organizers: Lora K. Heisler and Scott M. Sternson joint with Gastrointestinal Control of Metabolism May 9–13, 2017 | Copenhagen | Denmarkwww.keystonesymposia.org/17Z5Abstract Deadline – Feb 9, 2017; Discounted Registration Deadline – Mar 9, 2017

Submit an abstract to participate fully in the conference via a poster presentation and possible selection for a short talk. Abstracts submitted by the abstract deadline will be considered for short talks.

For more details on these conferences, please visit www.keystonesymposia.org/neuro.

Don’t Miss Keystone Symposia’s2017 Neurobiology Conferences

The Rockefeller University Press $30.00J. Exp. Med. 2015 Vol. 212 No. 7 991–999www.jem.org/cgi/doi/10.1084/jem.20142290

991

Brief Definit ive Report

Lymphatic circulation extends throughout most of the body and contributes to tissue homeosta-sis and function by facilitating the clearance of excess fluid and macromolecules from the inter-stitium (Secker and Harvey, 2015). However, the central nervous system (CNS) is considered to lack lymphatic vasculature, which has raised long-standing questions about how cerebral interstitial fluid (ISF) is cleared of waste products (Iliff and Nedergaard, 2013). The exchange of compounds is limited by the blood–brain bar-rier, which functions as a diffusion barrier be-tween the brain and circulating blood. Therefore, the transvascular clearance of most compounds is dependent on specific active transporter mech-anisms (Zlokovic, 2011). In addition, the brain has adapted to use a unique paravascular route in which fluids may freely exchange between the brain ISF and the cerebrospinal fluid (CSF) along glial “lymphatic” (glymphatic) routes with-out crossing the tightly regulated endothelial cell (EC) layer (Iliff et al., 2012; Xie et al., 2013). Downstream of the glymphatic system, the ma-jority of the CSF is considered to drain into the venous circulation through arachnoid granulations.

Still, several studies have found that a substan-tial proportion of the CSF is also drained into extracranial lymphatic vessels and LNs (Koh et al., 2005). However, the mechanisms of CSF entry into the extracranial lymphatic compart-ment are unclear.

The visualization of lymphatic vessels has been markedly facilitated over the last decade by the identification of specific lymphatic EC markers, such as prospero homeobox protein 1 (PROX1) transcription factor, a master regulator in the program specifying the lymphatic EC fate (Hong et al., 2002), vascular endothelial growth factor receptor 3 (VEGFR3), a lymphan-giogenic tyrosine kinase receptor (Secker and Harvey, 2015), chemokine (C-C motif) ligand 21 (CCL21), a chemokine secreted by lym-phatic ECs which facilitates the migration of dendritic cells into LNs (Liao and von der Weid, 2015), lymphatic vessel endothelial hyaluronan

CORRESPONDENCE Kari Alitalo: kari.alitalo@helsinki.fi

Abbreviations used: A488-OVA, Alexa Fluor 488– conjugated OVA; CFS, cerebrospinal fluid; CNS, central nervous system; dcLN, deep cervical LN; DiI, 1,1-dioctadecyl-3,3,3,3-tetramethylindocarbocyanine; EC, endothelial cell; IFP, ISF pressure; ISF, interstitial fluid; PEG, poly(ethylene glycol); PFA, paraformaldehyde; RT, room temperature; scLN, superficial cervical LN; TG, transgenic.

A dural lymphatic vascular system that drains brain interstitial fluid and macromolecules

Aleksanteri Aspelund,1,2 Salli Antila,1,2 Steven T. Proulx,3 Tine Veronica Karlsen,4 Sinem Karaman,3 Michael Detmar,3 Helge Wiig,4 and Kari Alitalo1,2

1Wihuri Research Institute and 2Translational Cancer Biology Program, Biomedicum Helsinki, University of Helsinki, 00014 Helsinki, Finland3Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology (ETH Zurich), CH-8093 Zurich, Switzerland4Department of Biomedicine, University of Bergen, 5009 Bergen, Norway

The central nervous system (CNS) is considered an organ devoid of lymphatic vasculature. Yet, part of the cerebrospinal fluid (CSF) drains into the cervical lymph nodes (LNs). The mechanism of CSF entry into the LNs has been unclear. Here we report the surprising finding of a lymphatic vessel network in the dura mater of the mouse brain. We show that dural lymphatic vessels absorb CSF from the adjacent subarachnoid space and brain inter-stitial fluid (ISF) via the glymphatic system. Dural lymphatic vessels transport fluid into deep cervical LNs (dcLNs) via foramina at the base of the skull. In a transgenic mouse model expressing a VEGF-C/D trap and displaying complete aplasia of the dural lymphatic vessels, macromolecule clearance from the brain was attenuated and transport from the subarachnoid space into dcLNs was abrogated. Surprisingly, brain ISF pressure and water content were unaffected. Overall, these findings indicate that the mechanism of CSF flow into the dcLNs is directly via an adjacent dural lymphatic network, which may be important for the clearance of macromolecules from the brain. Importantly, these results call for a reexamination of the role of the lymphatic system in CNS physiology and disease.

© 2015 Aspelund et al. This article is distributed under the terms of an Attribution– Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http://creativecommons.org/licenses/ by-nc-sa/3.0/).

992 Dural lymphatic vessels drain cerebrospinal fluid | Aspelund et al.

Figure 1. Terminally differentiated lymphatic vessels in the dura mater of the brain. Visualization of CNS lymphatic vasculature using Prox1-GFP reporter mice with DiI counterstaining for blood vasculature, Vegfr3+/LacZ reporter mice and immunofluorescence for PECAM1, and the lymphatic markers PROX1, LYVE1, PDPN, CCL21, and VEGFR3, as indicated. White arrowheads denote lymphatic vessels, yellow arrowheads denote the skull exit sites, and asterisks denote valves. (A) A schematic image of the various areas analyzed. The letters in bold refer to the corresponding images below. MMA, middle meningeal artery; PPA, pterygopalatine artery; RGV, retroglenoid vein; RRV, rostral rhinal vein; SS, sigmoid sinus; SSS, superior sagittal sinus; TV, trans-verse vein. (B) Lymphatic vessels running down along the SS and exiting the skull. (C) Lymphatic vessels running down along the proximal MMA branches.

JEM Vol. 212, No. 7 993

Br ief Definit ive Repor t

accessory), exiting the skull along the nerve (Fig. 1, G and H). Lymphatic vessels could be observed also in the dural lining of the cribriform plate, where some vessels passed through the skull into the nasal mucosa (Fig. 1, G and J).

Generally, lymphatic vessels were relatively scarce in the superior portions of the skull, whereas the base of the skull contained a more extensive lymphatic vessel network (Fig. 1 A). Interestingly, only the lymphatic vessels at the base of the skull contained valves, but their distribution was relatively scarce. Valves were separated by long stretches of valveless vessel seg-ments (Fig. 1, B–D).

To determine the localization of these vessels in relation to the meninges, thick skull sections were analyzed. In these preparations, PROX1- and CCL21-positive lymphatic vessels were observed in the meninges underlying the bony parts of the skull, adjacent to the dural blood vasculature (Fig. 1 K).

Whole-mount immunofluorescence staining of the supe-rior sagittal lymphatic vessels showed that, like conventional lymphatic vessels, the dural lymphatic vessels express very low levels of PECAM1 (Fig. 1 L) but high levels of LYVE1, PDPN, VEGFR3, CCL21, and PROX1 (Fig. 1, M–P; Aspelund et al., 2014). Thus, the dural lymphatic vessels are lined by termi-nally differentiated lymphatic endothelium.

Overall, these data indicated that lymphatic vessels are pres-ent in the dura mater of the CNS and drain out of the skull via the foramina of the base of the skull alongside arteries, veins, and cranial nerves. We named these lymphatic vessels on the basis of their venous, arterial, or cranial nerve counterparts. The localization of the vessels suggested a possible role in CSF absorption through the arachnoid mater.

Dura mater lymphatic vessels drain brain ISF into deep cervical LNs (dcLNs)Tracers injected into the brain ISF have been shown to trans-locate into the CSF via the glymphatic system and further into dcLNs (Koh et al., 2005; Iliff et al., 2012; Plog et al., 2015). However, it is unclear how these tracers gain access into the LNs. We hypothesized that the dura mater lymphatic vessels absorb brain ISF and CSF. To test this, we injected an inert 20-kD poly(ethylene glycol) (PEG) conjugate of the bright near-infrared dye IRDye 680 (PEG-IRDye; Proulx et al., 2013) into the brain parenchyma of the Prox1-GFP mice. 2 h after injection, the tracer was observed to exit the brain via paravenous routes for entry into the CSF space (not depicted), as previously reported (Iliff et al., 2012). Lymphatic

receptor 1 (LYVE1), and podoplanin (PDPN; Oliver and Srinivasan, 2010). We have recently discovered that in the eye, another immune-privileged organ previously consid-ered to lack lymphatic circulation, the Schlemm’s canal is a lymphatic-like vessel (Aspelund et al., 2014). These intriguing inconsistencies and our recent discoveries led us to investi-gate the possibility of lymphatic circulation in the CNS in more detail.

RESULTS AND DISCUSSIONLymphatic vessels in the dura mater surrounding the brainThe brain is enveloped by meningeal linings consisting of three layers: the pia mater tightly attached to the surface of the brain, the avascular arachnoid mater overlying the subarach-noid space, and the vascularized dura mater fused to the cra-nial bones. To determine whether lymphatic vessels exist within the CNS and surrounding meninges, we analyzed the Prox1-GFP and Vegfr3+/LacZ reporter mice and whole-mount immuno-fluorescence preparations of the skull and brain of WT mice against LYVE1, PROX1, PDPN, CCL21, VEGFR3, and PECAM1. To visualize blood vessels, the Prox1-GFP mice were perfused with the fluorescent dye 1,1-dioctadecyl-3,3,3,3-tetramethylindocarbocyanine (DiI; Li et al., 2008).

After removing the brain from the skull, no lymphatic vessels were seen on the brain parenchyma or pia mater (not depicted). However, a surprisingly extensive network of lym-phatic vessels was observed in the meninges underlying the skull bones (Fig. 1, A–J; and Video 1). In sagittal planes of the inner skull, lymphatic vessels were observed to run down to-ward the base of the skull along the transverse sinus, the sig-moid sinus, the retroglenoid vein, the rostral rhinal vein, and the major branches of the middle and anterior meningeal arteries (Fig. 1, B and D; and Video 1). In preparations of the superior portions of the skull, the lymphatic vessels were visualized along the superior sagittal sinus, the transverse sinus, the rostral rhinal veins, and the middle meningeal artery (Fig. 1, E and F). A concentration of lymphatic vessels could be ob-served to exit the skull along the meningeal portions of the pterygopalatine artery, a branch of the internal carotid artery which dives in and out of the skull to give rise to the middle meningeal artery (Fig. 1 I). Lymphatic vessels along the sig-moid sinus and retroglenoid vein exited the skull along the veins (Fig. 1, B and D). In preparations of the base of the skull, lymphatic vessels could be seen in the distal portion of several cranial nerves (optic, trigeminal, glossopharyngeal, vagus, and

(D) Lymphatic vessels around the RGV with some vessels exiting the skull. (E–H) Whole-mount LYVE1 immunofluorescence of the skull top and base. (E) Lym-phatic vessels along the SSS and the distal parts of the anterior MMA branch extending toward the bregma. (F) Lymphatic vessels along the SSS, bifurcat-ing into the TVs at the confluence of sinuses. (G) Lymphatic vessels exiting the skull along the optic (II) and the trigeminal (V) nerves and through thecribriform plate (CP). CN, cranial nerve. (H) Lymphatic vessels associated with the glossopharyngeal (IX), vagus (X), and accessory (XI) nerves. XII, hypo-glossal nerve. (I and J) Stereomicrographs of tissues in a Vegfr3+/LacZ reporter mouse showing the skull exit of dural lymphatic vessels along the PPA(I) and through the CP into a nasal concha. OB, olfactory bulb area. (K) Immunofluorescence of thick skull section for PECAM1, PROX1, and CCL21. bv, bloodvessel; sas, subarachnoid space. (L–P) Whole-mount immunofluorescence staining of superior sagittal lymphatic vessels with antibodies against PECAM1(L), LYVE1 (M), PDPN (N), CCL21 (O), VEGFR3 (P), and PROX1 (M–P). LYVE1 and PECAM1 colocalization is indicated with the dashed lines. n = 2–3 perstaining. Data are from two to three independent experiments. Bars: (B–H and L–P) 100 µm; (K) 50 µm.

994 Dural lymphatic vessels drain cerebrospinal fluid | Aspelund et al.

data suggest that the dura mater lymphatic vessels absorb brain ISF/CSF from the subarachnoid space for transport into downstream dcLNs.

Absence of dura mater lymphatic vasculature in K14-VEGFR3-Ig miceVEGF-C/D signaling via VEGFR3 is a critical regulator of lymphangiogenesis (Secker and Harvey, 2015). To (a) study whether dura mater lymphatic vessels are regulated by VEGFC/D–VEGFR3 signaling and (b) characterize an animal model in which the functional consequences of dura mater lymphatic vessel aplasia can be examined, we investigated the K14-VEGFR3-Ig transgenic (TG) mouse, which has impaired

drainage of brain ISF was confirmed by visualization of an intense signal in the dcLN but not in the superficial cervi-cal LNs (scLNs; Fig. 2, A and B). Preferential drainage into the dcLN ipsilateral to the side of injection was observed (Fig. 2, A and C). When the tracer-filled afferent lymphatic vessels of the dcLNs were followed upstream, the vessels appeared to drain from the base of the skull (Fig. 2, C and D). Inside the skull, some PEG-IRDye filling of dura mater lymphatic ves-sels was observed only in the basal parts of the skull (Fig. 2, E and F), suggesting uptake by the lymphatic vessels but a quick washout. When the efferent lymphatic vessel of the dcLN was ligated (Fig. 2, G and H), enhanced filling of the dural lymphatic vessels was observed (Fig. 2, I–K). These

Figure 2. Dura mater lymphatic vessels drain brain ISF into dcLNs. (A–J) Analysis of lymphatic outflow routes of cerebral ISF by fluorescent stereo-microscopy in Prox1-GFP (green) mice 1 h after PEG-IRDye (red) injection into the brain parenchyma without (A–F) and with (G–J) ligation of the efferent lymphatic vessel of the dcLN. See K for schematic il-lustration of the experimental setup and summary of the results with and without ligation. (A and B) dcLNs and scLNs (both indicated with arrowheads) showing preferential filling of the ipsilateral dcLN but no fill-ing in the scLNs. (C) Drainage into the ipsilateral dcLN via the efferent carotid lymphatic vessels (arrow-heads). CCA, common carotid artery. (D) Internal carotid artery (ICA) and adjacent lymphatic vessels (white arrowheads) immediately below the osseous skull, showing drainage from the skull (yellow arrow-head). (E and F) Lymphatic vessels around the ptery-gopalatine artery (PPA), showing tracer uptake by the dura mater lymphatic vessels (arrowheads) only in the basal parts of the skull, nearby their exit site. MMA, middle meningeal artery. (G) Placement of a suture around the efferent lymphatic vessel (asterisk) of the dcLN. Arrowheads, afferent lymphatic vessels. (H) Afferent lymphatic vessel of the dcLN after liga-tion (asterisk), showing bulging of the afferent ves-sels (arrowheads). (I and J) Lymphatic vessels aroundthe posterior branch of the MMA, showing increasedfilling of lymphatic vessels after ligation, extendingabove the retroglenoid vein (RGV) level. n = 2–3/group. Data are representative of two independentexperiments. Bars: (A–E and G–J) 500 µm; (F) 100 µm.

JEM Vol. 212, No. 7 995

Br ief Definit ive Repor t

inhibited in the TG mice (Fig. 5). Overall, these data imply that the dura mater lymphatic vessels contribute to the clear-ance of macromolecules from the brain.

In this study, we report the surprising finding of a lym-phatic vessel network in the dura mater of the CNS and show that dura mater lymphatic vessels are lined by fully differ-entiated PROX1+/VEGFR3+/LYVE1+/PDPN+/CCL21+/PECAM1low lymphatic endothelium that is unique in its morphology and scarcity of valves. In the late eighteenth century the Italian anatomist Paolo Mascagni described what he called lymphatic vessels in the meninges and on the sur-face of the brain, but his finding could never be reproduced (Mascagni and Bellini, 1816; Lukić et al., 2003). The CNS proper has since been considered devoid of lymphatic vasculature.

VEGF-C/D–VEGFR3 signaling. These mice express a solu-ble VEGF-C/D trap protein consisting of the ligand-binding Ig homology domains 1–3 of VEGFR3 fused with the Fc domain of Ig (Mäkinen et al., 2001). Although the VEGF-C/D trap transgene is expressed in keratinocytes, the circulating protein inhibits lymphangiogenesis in most tissues, and the mice display LN hypoplasia (Mäkinen et al., 2001; Alitalo et al., 2013). Lymphatic vessels were absent from both supe-rior and basal parts of the skull in the TG mice compared with WT littermate mice (Fig. 3, A–F). Surprisingly, the mice displayed absence of only the scLNs but not dcLNs (Fig. 3, G–I; and Fig. 4 C). These data indicate that the dura mater lymphatic vessels are very sensitive to the inhibition of VEGF-C/D signaling and that the K14-VEGFR3-Ig TG mouse is a suitable model for studying the functional consequences of the absence of lymphatic drainage from the brain.

Lack of dural lymphatic vessels compromises CNS macromolecule clearanceFirst, we hypothesized that the absence of dura mater lym-phatic vessels would impair the clearance of ISF and solutes from the brain. Thus, brain water content and ISF pressures (IFPs) were measured in TG and WT mice. Surprisingly, the IFP (TG vs. WT: 2.50 ± 0.54 vs. 2.53 ± 0.53 mmHg, P = 0.92, n = 6/group) and water content (TG vs. WT: 3.68 ± 0.023 vs. 3.71 ± 0.043 g/g dry weight, P = 0.27, n = 4/group) were not significantly different between the two groups. These results suggest that in physiological conditions, the brain has alternative ways to manage fluid extravasated from the blood vessels.

Second, we hypothesized that the absence of dura mater lymphatic vessels may impair macromolecule clearance from the brain. To test this, we studied the cerebral clearance of Alexa Fluor 488–conjugated OVA (A488-OVA, 45 kD), a macromolecule which retains fluorescent signal during fixa-tion. We recorded cerebral, dcLN, and dura mater lymphatic vessel fluorescence from tissues 2 h after injection into the brain parenchyma of TG and WT littermate mice. Mice were perfusion fixed after sacrifice to prevent outflow of the tracer. Interestingly, the TG mice displayed a significant reduction in the amount of OVA cleared at the 2-h time point after injection (Fig. 4, A and B). Furthermore, a nearly complete abrogation of OVA accumulation was observed in the dcLNs of the TG mice (Fig. 4, C and D). Tracer-filled lymphatic vessels could be observed around the pterygopalatine artery and middle meningeal artery of WT mice, but this was absent in the TG mice (Fig. 4, E and F). To assess other possible causes for the drainage defect, we analyzed glymphatic function and the dcLN capacity for drainage. To this extent, the TG mice did not display qualitative defects in glymphatic function, as indicated by detectable paravascular outflow of the tracer in the subendothelial and perivascular space (Fig. 4, G and H), or a significant reduction in the amount of draining lym-phatic vessels in the dcLN (Fig. 4, I and J). We also studied PEG-IRDye transfer from the subarachnoid space into the dcLNs after cisterna magna injection, which was significantly

Figure 3. Absence of dural lymphatic vasculature in K14-VEGFR3-Ig TG mice. (A–F) Analysis of dura mater lymphatic vasculature in K14-VEGFR3-Ig TG and WT littermate control mice. (A–C) Immunofluorescence of the superior sagittal lymphatic vessels (arrowheads) for PECAM1, PROX1, and CCL21 (A and B) and quantification of PROX1+/CCL21+ lymphatic ECs (LECs)/grid (C). (D–F) Immunofluorescence of the pterygopalatine and mid-dle meningeal lymphatic vessels (arrowheads) for PECAM1 and PROX1 (D and E) and quantification of PROX1+ LECs/grid (F). (G–I) Stereomicroscopic photographs showing the absence of the scLNs (arrows) in the TG mice (G and H) and quantification of the (mean left/right) scLN and dcLN sur-face areas (I). Micrographs of the dcLNs are shown in Fig. 4 C. (A–F) n = 3 (TG) and 4 (WT). (G and H) n = 4/group. Data are representative of two independent experiments. Bars: (A, B, D, and E) 100 µm; (G and H) 2 mm. Error bars indicate SD. Statistical analysis: two-tailed Student’s t test (C and F) and two-way ANOVA followed by Šídák’s post-hoc test (I). ***, P < 0.001; ****, P < 0.0001.

996 Dural lymphatic vessels drain cerebrospinal fluid | Aspelund et al.

for transport into the cerebral venous sinuses (Pollay, 2010). However, recent discoveries have established the glymphatic system as a critical regulator of cerebral waste clearance, espe-cially during sleep (Iliff et al., 2012; Xie et al., 2013). In addi-tion to the CSF clearance via arachnoid granulations, several studies have established that a part of brain ISF and CSF is drained into cervical LNs, yet it has been unclear how CSF enters the LNs (Koh et al., 2005; Weller et al., 2009). The ob-servation of CSF tracers in the nasal lymphatic vessels under the cribriform plate has suggested clearance via olfactory

Incidentally, lymphatic vessels were mentioned in an electron microscopic study of the rat dura mater innervation. Fur-thermore, lymphatic vessels were detected in association with the murine cribriform plate and the human optic nerve (Andres et al., 1987; Gausas et al., 2007; Furukawa et al., 2008). How-ever, the extent of the dura mater lymphatic network, or its role in CSF clearance, has not been realized.

According to the classical textbook model, CSF is produced by the choroid plexus, flows through the ventricles and the subarachnoid space, and is absorbed by arachnoid granulations

Figure 4. Lack of dural lymphatic vessels compromises CNS macromolecule clearance. Analysis of A488-OVA distribution 2 h after intra-parenchymal injection in K14-VEGFR3-Ig TG mice and WT littermate controls. (A and B) Representative false color maps and quantification of the epifluorescence efficiency in the brain using IVIS imaging. (C and D) Representative images and quantification of the fluorescence in the dcLNs (indi-cated by arrows). (E and F) Representative fluorescent images of the A488-OVA tracer (indicated by arrowheads) accumulation in the LYVE1-stained lymphatic vessels around the PPA and MMA, with quantification of the A488-OVA–positive signal. Note the partial leakage of the tracer from the vessels caused by the perfusion fixation. (G) Fluorescent images of brain sections stained with DAPI and antibodies against endomucin (EMCN), show-ing the A488-OVA tracer distribution in the glymphatic system. (H) Plot profile analysis of the fluorescence along the indicated lines in G, showing A488-OVA signal in the subendothelial and perivascular spaces (arrows) in both TG and WT mice. (I) Immunofluorescent images of dcLNs stained with DAPI and antibodies against LYVE1. (J) Quantification of the LYVE1+ area in the dcLNs in TG mice and WT littermate controls. (A, B, and G–J) n = 4 (TG) and 3 (WT). (C–F) n = 3 (TG) and 4 (WT). Data are representative of two independent experiments. Bars: (C) 2 mm; (E) 100 µm; (G) 8 µm; (I) 1,000 µm. Error bars indicate SD. Statistical analysis: two-tailed Student’s t test. **, P < 0.01; ***, P < 0.001.

JEM Vol. 212, No. 7 997

Br ief Definit ive Repor t

Choi et al., 2011), and Vegfr3+/LacZ (FVB/N background; Dumont et al., 1998) mouse lines have been published previously. WT littermate mice were used as controls. For WT analysis, C57BL/6J mice were used. For tissue analysis, mice were given a lethal dose of ketamine and xylazine and perfusion fixed through the left ventricle with ice-cold 1% paraformaldehyde (PFA) after puncture of the right auricle. The tissues were immediately immersed in 4% ice-cold PFA and postfixed overnight at 4°C, washed in PBS, and processed for staining. Prox1-GFP mouse tissues were freshly imaged without fixation.

Immunostaining and X-gal staining. For whole-mount staining of the skull bones for laser-scanning confocal microscopy, the fixed skulls were dis-sected and underwent a mild decalcification with 0.5 M EDTA, pH 7.4, over-night at 4°C. For whole-mount staining of the basal skull for fluorescent stereomicroscopy, no decalcification was performed. After washes with PBS, the tissues were permeabilized in 0.3% Triton X-100 in PBS (PBS-TX) and blocked in 5% donkey serum/2% bovine serum albumin/0.3% PBS-TX. Pri-mary antibodies were added to the blocking buffer and incubated with the tis-sue overnight at room temperature (RT). After washes in PBS-TX, the tissues were incubated with fluorophore-conjugated secondary antibodies in PBS-TX overnight at RT, followed by washing in PBS-TX. After postfixation in 1% PFA, the superior portions of the skull were washed with PBS and mounted in Mowiol 4-88 mounting medium (Sigma-Aldrich) containing 1,4-diazabicy-clo[2.2.2]octane (DABCO; Sigma-Aldrich) and sealed with Cytoseal (Thermo Fisher Scientific). Clothespins were used to hold the coverslip and the micro-scopic slide together before the Cytoseal and Mowiol hardened. Tissues for fluorescent stereomicroscopy were stored in PBS and imaged immediately.

For cryosections of the skull, the fixed tissues underwent decalcification with 0.5 M EDTA, pH 7.4, for 72 h, immersion into 20% sucrose and 2% polyvinylpyrrolidone (PVP) for 24 h at 4°C, embedding in OCT compound (Tissue-Tek), and freezing for storage at 80°C. For other cryosections, the fixed tissues were immersed into 25% sucrose and embedded as above. Tissues were cut into 10–100-µm sections using a cryostat (Microm HM 550; Thermo Fisher Scientific). The sections were air-dried, encircled with a pap-pen, rehydrated in PBS, and blocked with 3% BSA in PBS-TX at RT. After primary antibody incubation at 4°C in 3% BSA in PBS overnight, the sections were washed with PBS and incubated for 2–3 h with the appropri-ate fluorophore-conjugated secondary antibody conjugates and 3% BSA in PBS. After washes with 0.1% PBS-TX, the sections were mounted with

nerve sheaths through the cribriform plate (Kida et al., 1993; Koh et al., 2005). Additionally, CSF clearance has been ob-served to occur along spinal and cranial nerve sheaths with subsequent entry into extracranial lymphatic vessels (Miura et al., 1998; Weller et al., 2009).

Our data indicated filling of the dura mater lymphatic ves-sels after intraparenchymal injection of the tracer and the lack thereof in the K14-VEGFR3-Ig TG mice. This suggests a model in which a part of the brain ISF, downstream of the glymphatic system, is cleared directly from the subarachnoid space as CSF into the dura mater lymphatic vasculature. Interestingly, we also observed lymphatic vessels draining out of the skull along the dura mater of cranial nerves. Furthermore, we observed lym-phatic vessels crossing the cribriform plate, which may explain some of the previous observations. Because of the lack of other known direct anatomical connections between the CSF space and extracranial lymphatic vessels, the dura mater lymphatic vessels are likely to represent the most important CSF source for the extracranial lymph compartment.

The importance of understanding the mechanisms of brain waste management are highlighted in patients suffering from Alzheimer’s disease and other neurodegenerative dis-eases characterized by the pathological accumulation of mis-folded proteins, such as amyloid , into the brain parenchyma (Deane et al., 2008; Huang and Mucke, 2012). In other tissues, lymphatic vessels are critical for the absorption of macromol-ecules (Tammela and Alitalo, 2010). In the brain under physi-ological conditions, a major part of the cerebral amyloid is removed by the transvascular route (Zlokovic, 2011; Zhao et al., 2015). However, recent evidence suggests that the glym-phatic system may also be key in amyloid clearance (Iliff et al., 2012). The present data show that the absence of dura mater lymphatic drainage results in inhibited clearance of OVA from the brain interstitium, suggesting that dura mater lym-phatic vessels are critical for the absorption of macromolecules from the brain ISF and CSF.

Importantly, these findings open new avenues for re-search. Several other potential roles of dura mater lymphatic vessels can be envisioned, such as in the trafficking of cerebral immune cells, in antigen presentation in the dcLNs, and in the clearance of brain edema. These data may also explain why primary brain tumors can rarely metastasize into cervical LNs (Mondin et al., 2010). Interestingly, surgical removal of the dcLN results in cognitive impairment in mice (Radjavi et al., 2014), and ligation of the deep cervical lymphatic vessels has been reported to aggravate cerebral ischemia after stroke by increasing brain water edema and infarction volume in rats (Si et al., 2006). Further studies should be conducted to de-fine the full contribution of dura mater lymphatic vasculature in CNS homeostasis and disease.

MATERIALS AND METHODSStudy approval. The study was approved by the Committee for Animal Experiments of the District of Southern Finland.

Mice and tissues. The K14-VEGFR31-3-Ig (FVB/N and C57BL/6J back-grounds; Mäkinen et al., 2001), Prox1-GFP (C57BL/6J albino background;

Figure 5. Lack of dural lymphatic vasculature inhibits CSF up-take into the dcLNs. (A) Schematic illustration of the experimental setup. (B) Representative fluorescent images of the dcLN in TG and WT mice 30 min after PEG-IRDye injection into the cisterna magna. AF, green channel autofluorescence. Bar, 1,000 µm. (C) Quantification of the dcLN fluorescence. n = 6 (TG) and 5 (WT). Data are representative of two inde-pendent experiments. Error bars indicate SD. Statistical analysis: two-tailed Student’s t test. *, P < 0.05.

998 Dural lymphatic vessels drain cerebrospinal fluid | Aspelund et al.

with a XFO-6 (Dolan-Jenner Fiber-Lite PL-900 Illuminator; quartz halogen lamp) for fluorescent imaging and an iXon+ 888 EMCCD camera (Andor Technology). Images were processed and region of interest efficiencies were calculated with the Living Image 3.2 software. Image brightness and contrast were adjusted using ImageJ (National Institutes of Health) or Photoshop (Adobe) software. Quantitative analysis of the micrographs was performed using the ImageJ software.

Measurement of brain water content. Mice were sacrificed with carbon dioxide. The brains were removed from the skull and placed on a preweighed piece of aluminum foil and immediately weighed to obtain the wet weight. The dry weight was recorded after dehydration for 5 d in an 80°C oven. Water content was calculated as (wet weight dry weight)/dry weight. The data shown in the text is representative of two independent experiments.

Measurement of brain IFP. Mice were anesthetized with a mixture of 120 mg/kg ketamine (Ketalar) and 0.24 mg/kg medetomidine (Domitor) in saline given s.c. and placed into a stereotactic device. Using a dental drill (2 mm OD), the skull bone was thinned at a site 2 mm caudal and lateral to the bregma. Brain ISF pressure was measured with micropipettes, tip diameter 2–4 µm, as described in detail previously (Wiig and Reed, 1983). Pipettes were inserted through an intact dura, and pressures were recorded 150–300 µm into brain tissue. The data shown in the text is representative of three inde-pendent experiments.

Statistical analysis. All values are expressed as mean ± SD. Quantitative data were compared between different groups by two-sample (unpaired Student’s) two-tailed t test assuming equal variance. Two-way ANOVA followed by Šídák’s post-hoc test was used for multiple comparisons. Differences were considered statistically significant at P < 0.05.

Online supplemental material. Video 1 shows dura mater lymphatic ves-sels in the lateral aspects of the interior portions of the skull in sagittal plane in the Prox1-GFP mouse. Online supplemental material is available at http://www.jem.org/cgi/content/full/jem.20142290/DC1.

We thank Drs. Paola Luciani and Jean-Christophe Leroux for the PEG-IRDye; Dr. Dontscho Kerjaschki for the antibodies against PDPN; Kirsi Lintula, Riitta Kauppinen, Jarmo Koponen, and Tapio Tainola (University of Helsinki, Helsinki, Finland) for technical assistance; the Biomedicum Imaging Unit for help with imaging; and the staff of the University of Helsinki Laboratory Animal Centre for technical assistance with the mouse work.

This study was supported by grants from the Academy of Finland, the Sigrid Juselius Foundation, the European Research Council (ERC-2010-AdG-268804), the Swiss National Science Foundation (310030B_147087), VESSEL-Marie Curie multipartner Initial Training Network for Vascular Biology (EU FP7-PEOPLE-2012-ITN), the Leducq Transatlantic Network of Excellence on Lymph Vessels in Obesity and Cardiovascular Disease (11CVD03; to K. Alitalo and M. Detmar), and the Research Council of Norway (222278/F20).

The authors declare no competing financial interests.

Submitted: 8 December 2014Accepted: 4 June 2015

REFERENCESAlitalo, A.K., S.T. Proulx, S. Karaman, D. Aebischer, S. Martino, M. Jost, N.

Schneider, M. Bry, and M. Detmar. 2013. VEGF-C and VEGF-D block-ade inhibits inflammatory skin carcinogenesis. Cancer Res. 73:4212–4221. http://dx.doi.org/10.1158/0008-5472.CAN-12-4539

Andres, K.H., M. von Düring, K. Muszynski, and R.F. Schmidt. 1987. Nerve fibres and their terminals of the dura mater encephali of the rat. Anat. Embryol. (Berl.). 175:289–301. http://dx.doi.org/10.1007/BF00309843

Aspelund, A., T. Tammela, S. Antila, H. Nurmi, V.-M. Leppänen, G. Zarkada, L. Stanczuk, M. Francois, T. Mäkinen, P. Saharinen, et al. 2014. The Schlemm’s canal is a VEGF-C/VEGFR-3-responsive lymphatic-like vessel. J. Clin. Invest. 124:3975–3986. http://dx.doi.org/10.1172/JCI75395

Vectashield mounting medium containing DAPI (Vector Laboratories). For X-gal staining of skulls from LacZ reporter mice, the tissues were fixed with0.2% glutaraldehyde and stained by the -galactosidase substrate X-Gal(Promega) with a published protocol (Karkkainen et al., 2004).

Blood vessel labeling with DiI. Blood vessels of Prox1-GFP mice were labeled by cardiac perfusion using a fluorescent lipophilic carbocyanine dye DiI as previously described (Li et al., 2008). DiI incorporated into the EC membranes was visualized with fluorescent stereomicroscopy.

Antibodies. The following primary antibodies were used for immunostain-ing of mouse tissues: rabbit anti–mouse PROX1 (1:200; Petrova et al., 2008), goat anti–human PROX1 (1:500; R&D Systems), polyclonal goat anti–mouse VEGFR3 (1:50; R&D Systems), rat anti–mouse PECAM-1 (1:500; clone MEC 13.3; BD), polyclonal rabbit anti–mouse LYVE1 (1:1,000; Karkkainen et al., 2004), goat anti-CCL21 (1:500; R&D Systems), IgG fraction of rabbit polyclonal anti–mouse PDPN (1:500; provided by D. Kerjaschki, University of Vienna, Vienna, Austria; Wick et al., 2007), and rat anti–mouse endomucin (EMCN; 1:500; clone V.7C7; Santa Cruz Biotechnology, Inc.). The primary antibodies were detected with the appropriate Alexa Fluor 488, 568/594, or 633/647 secondary antibody conjugates (1:500; Molecular Probes/Invitrogen).

Intraparenchymal PEG-IRDye and A488-OVA clearance. To evaluate the outflow pathways and clearance of tracers from the brain interstitium, mice were anesthetized with a mixture of 80 mg/kg ketamine (Ketalar; Pfizer) and 6 mg/kg xylazine (Rompun vet; KVP Pharma + Veterinär Produke GmbH) or 160 mg/kg ketamine (Narketan; Vétoquinol) and 0.4 mg/kg medetomi-dine (Domitor; Orion Pharma) and placed into a stereotactic device. A mid-line skin incision was made to reveal the skull bone, which was thinned with a dental drill 2 mm lateral and 2.5 mm caudal to the bregma. 4 mg/ml A488-OVA (O-34781; Molecular Probes) or 20 µM of 20-kD PEG-IRDye (provided by P. Luciani and J.-C. Leroux, ETH Zurich, Zurich, Switzerland; Proulx et al., 2013) was injected into a 2-mm depth from the bregma in 0.5 µl with a 34-G Hamilton needle at a 0.1 µl/min rate over 5 min with a syringe pump (Harvard Apparatus). After the indicated time, the mice were sacrificed with a lethal dose of anesthesia. For visualization of PEG-IRDye, tissues were immediately imaged ex vivo. For visualization of A488-OVA, mice were per-fusion fixed, and the entire head and neck with the cervical LNs were im-mersed in 4% ice-cold PFA and further postfixed overnight at 4°C with constant shaking. Fixed tissues were then washed in PBS and processed for imaging or staining as described above.

Cisterna magna PEG-IRDye clearance. To evaluate the PEG-IRDye clearance from the subarachnoid space, mice were anesthetized with a mixture of 160 mg/kg ketamine (Narketan) and 0.4 mg/kg medetomidine (Domitor) and placed into a stereotactic frame. The neck muscles were bluntly dissected through a small midline incision to reveal the dura mater overlying the cis-terna magna. 20 µM of 20-kD PEG-IRDye was injected into the subarach-noid space in 10 µl with a 34-G Hamilton needle at a rate of 2 µl/min over 5 min with a syringe pump.

Image acquisition, processing, and quantitative analysis. Laser-scanning confocal micrographs of the fluorescently labeled samples were acquired at high resolution with an LSM 780 microscope (Carl Zeiss; air objectives 10× Plan-Apochromat with NA 0.45 and 20× Plan-Apochromat with NA 0.80, oil objective 40× Plan-Neofluar with NA 1.3 and oil objective 63× Plan-Apochromat with NA 1.4) using multichannel scanning in frame. The ZEN 2010 software (Carl Zeiss) was used for image acquisition.

Fluorescent stereomicrographs of labeled samples were obtained with an Axio Zoom.V16 fluorescence stereo zoom microscope (Carl Zeiss) equipped with an ORCA-Flash 4.0 digital sCMOS camera (Hamamatsu Photonics) or an OptiMOS sCMOS camera (QImaging). The ZEN 2012 software (Carl Zeiss) was used for image acquisition. Brain epifluorescence images were ob-tained with a Caliper IVIS Kinetic imaging system (PerkinElmer) equipped

JEM Vol. 212, No. 7 999

Br ief Definit ive Repor t

Mascagni, P., and G.B. Bellini. 1816. Istoria Completa Dei Vasi Linfatici. Vol. II. Presso Eusebio Pacini e Figlio, Florence. 195 pp.

Miura, M., S. Kato, and M. von Lüdinghausen. 1998. Lymphatic drainage of the cerebrospinal fluid from monkey spinal meninges with special refer-ence to the distribution of the epidural lymphatics. Arch. Histol. Cytol. 61:277–286. http://dx.doi.org/10.1679/aohc.61.277

Mondin, V., A. Ferlito, K.O. Devaney, J.A. Woolgar, and A. Rinaldo. 2010. A survey of metastatic central nervous system tumors to cervical lymph nodes. Eur. Arch. Otorhinolaryngol. 267:1657–1666. http://dx.doi.org/ 10.1007/s00405-010-1357-1

Oliver, G., and R.S. Srinivasan. 2010. Endothelial cell plasticity: how to become and remain a lymphatic endothelial cell. Development. 137:363–372. http://dx.doi.org/10.1242/dev.035360

Petrova, T.V., A. Nykänen, C. Norrmén, K.I. Ivanov, L.C. Andersson, C. Haglund, P. Puolakkainen, F. Wempe, H. von Melchner, G. Gradwohl, et al. 2008. Transcription factor PROX1 induces colon cancer progression by promoting the transition from benign to highly dysplastic phenotype. Cancer Cell. 13:407–419. http://dx.doi.org/10.1016/j.ccr.2008.02.020

Plog, B.A., M.L. Dashnaw, E. Hitomi, W. Peng, Y. Liao, N. Lou, R. Deane, and M. Nedergaard. 2015. Biomarkers of traumatic injury are transportedfrom brain to blood via the glymphatic system. J. Neurosci. 35:518–526. http://dx.doi.org/10.1523/JNEUROSCI.3742-14.2015

Pollay, M. 2010. The function and structure of the cerebrospinal fluid outflow system. Cerebrospinal Fluid Res. 7:9. http://dx.doi.org/10.1186/1743- 8454-7-9

Proulx, S.T., P. Luciani, A. Christiansen, S. Karaman, K.S. Blum, M. Rinderknecht, J.-C. Leroux, and M. Detmar. 2013. Use of a PEG- conjugated bright near-infrared dye for functional imaging of rerouting of tumor lymphatic drainage after sentinel lymph node metastasis. Biomaterials. 34:5128–5137. http://dx.doi.org/10.1016/j.biomaterials.2013.03.034

Radjavi, A., I. Smirnov, N. Derecki, and J. Kipnis. 2014. Dynamics of the meningeal CD4+ T-cell repertoire are defined by the cervical lymph nodes and facilitate cognitive task performance in mice. Mol. Psychiatry. 19:531–532. http://dx.doi.org/10.1038/mp.2013.79

Secker, G.A., and N.L. Harvey. 2015. VEGFR signaling during lymphatic vascular development: From progenitor cells to functional vessels. Dev. Dyn. 244:323–331. http://dx.doi.org/10.1002/dvdy.24227

Si, J., L. Chen, and Z. Xia. 2006. Effects of cervical-lymphatic blockade on brain edema and infarction volume in cerebral ischemic rats. Chin. J. Physiol. 49:258–265.

Tammela, T., and K. Alitalo. 2010. Lymphangiogenesis: Molecular mecha-nisms and future promise. Cell. 140:460–476. http://dx.doi.org/10.1016/ j.cell.2010.01.045

Weller, R.O., E. Djuanda, H.-Y. Yow, and R.O. Carare. 2009. Lymphatic drain-age of the brain and the pathophysiology of neurological disease. Acta Neuropathol. 117:1–14. http://dx.doi.org/10.1007/s00401-008-0457-0

Wick, N., P. Saharinen, J. Saharinen, E. Gurnhofer, C.W. Steiner, I. Raab, D. Stokic, P. Giovanoli, S. Buchsbaum, A. Burchard, et al. 2007. Transcriptomal comparison of human dermal lymphatic endothelial cellsex vivo and in vitro. Physiol. Genomics. 28:179–192. http://dx.doi.org/10.1152/physiolgenomics.00037.2006

Wiig, H., and R.K. Reed. 1983. Rat brain interstitial fluid pressure measured with micropipettes. Am. J. Physiol. 244:H239–H246.

Xie, L., H. Kang, Q. Xu, M.J. Chen, Y. Liao, M. Thiyagarajan, J. O’Donnell, D.J. Christensen, C. Nicholson, J.J. Iliff, et al. 2013. Sleep drives metabolite clearance from the adult brain. Science. 342:373–377. http://dx.doi.org/ 10.1126/science.1241224

Zhao, Z., A.P. Sagare, Q. Ma, M.R. Halliday, P. Kong, K. Kisler, E.A. Winkler, A. Ramanathan, T. Kanekiyo, G. Bu, et al. 2015. Central role for PICALMin amyloid- blood-brain barrier transcytosis and clearance. Nat. Neurosci.http://dx.doi.org/10.1038/nn.4025

Zlokovic, B.V. 2011. Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders. Nat. Rev. Neurosci. 12:723–738. http://dx.doi.org/10.1038/mrm3114

Choi, I., H.K. Chung, S. Ramu, H.N. Lee, K.E. Kim, S. Lee, J. Yoo, D. Choi, Y.S. Lee, B. Aguilar, and Y.-K. Hong. 2011. Visualization of lymphatic vessels by Prox1-promoter directed GFP reporter in a bacterial artifi-cial chromosome-based transgenic mouse. Blood. 117:362–365. http://dx.doi.org/10.1182/blood-2010-07-298562

Deane, R., A. Sagare, and B.V. Zlokovic. 2008. The role of the cell surface LRP and soluble LRP in blood-brain barrier A clearance in Alzheimer’s disease. Curr. Pharm. Des. 14:1601–1605. http://dx.doi.org/10.2174/ 138161208784705487

Dumont, D.J., L. Jussila, J. Taipale, A. Lymboussaki, T. Mustonen, K. Pajusola, M. Breitman, and K. Alitalo. 1998. Cardiovascular failure in mouse embryos deficient in VEGF receptor-3. Science. 282:946–949. http://dx.doi.org/10.1126/science.282.5390.946

Furukawa, M., H. Shimoda, T. Kajiwara, S. Kato, and S. Yanagisawa. 2008. Topographic study on nerve-associated lymphatic vessels in the murine craniofacial region by immunohistochemistry and electron micros-copy. Biomed. Res. 29:289–296. http://dx.doi.org/10.2220/biomedres .29.289

Gausas, R.E., T. Daly, and F. Fogt. 2007. D2-40 expression demonstrates lym-phatic vessel characteristics in the dural portion of the optic nerve sheath. Ophthal. Plast. Reconstr. Surg. 23:32–36. http://dx.doi.org/10.1097/01 .iop.0000249431.04116.a8

Hong, Y.-K., N. Harvey, Y.-H. Noh, V. Schacht, S. Hirakawa, M. Detmar, and G. Oliver. 2002. Prox1 is a master control gene in the program specifyinglymphatic endothelial cell fate. Dev. Dyn. 225:351–357. http://dx.doi.org/10.1002/dvdy.10163

Huang, Y., and L. Mucke. 2012. Alzheimer mechanisms and therapeutic strate-gies. Cell. 148:1204–1222. http://dx.doi.org/10.1016/j.cell.2012.02.040

Iliff, J.J., and M. Nedergaard. 2013. The microcirculation—fantastic voyage: Is there a cerebral lymphatic system? Stroke. 44:S93–S95. http://dx.doi .org/10.1161/STROKEAHA.112.678698

Iliff, J.J., M. Wang, Y. Liao, B.A. Plogg, W. Peng, G.A. Gundersen, H. Benveniste, G.E. Vates, R. Deane, S.A. Goldman, et al. 2012. A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid . Sci. Transl. Med. 4:147ra111. http://dx.doi.org/10.1126/scitranslmed.3003748

Karkkainen, M.J., P. Haiko, K. Sainio, J. Partanen, J. Taipale, T.V. Petrova, M. Jeltsch, D.G. Jackson, M. Talikka, H. Rauvala, et al. 2004. Vascular endo-thelial growth factor C is required for sprouting of the first lymphatic vessels from embryonic veins. Nat. Immunol. 5:74–80. http://dx.doi.org/ 10.1038/ni1013

Kida, S., A. Pantazis, and R.O. Weller. 1993. CSF drains directly from the sub-arachnoid space into nasal lymphatics in the rat. Anatomy, histology and immunological significance. Neuropathol. Appl. Neurobiol. 19:480–488. http://dx.doi.org/10.1111/j.1365-2990.1993.tb00476.x

Koh, L., A. Zakharov, and M. Johnston. 2005. Integration of the subarach-noid space and lymphatics: is it time to embrace a new concept of cere-brospinal fluid absorption? Cerebrospinal Fluid Res. 2:6. http://dx.doi .org/10.1186/1743-8454-2-6

Li, Y., Y. Song, L. Zhao, G. Gaidosh, A.M. Laties, and R. Wen. 2008. Direct labeling and visualization of blood vessels with lipophilic carbocyanine dye DiI. Nat. Protoc. 3:1703–1708. http://dx.doi.org/10.1038/nprot .2008.172

Liao, S., and P.Y. von der Weid. 2015. Lymphatic system: An active pathway for immune protection. Semin. Cell Dev. Biol. 38:83–89. http://dx.doi.org/ 10.1016/j.semcdb.2014.11.012

Lukić, I.K., V. Gluncić, G. Ivkić, M. Hubenstorf, and A. Marusić. 2003. Virtual dissection: a lesson from the 18th century. Lancet. 362:2110–2113. http://dx.doi.org/10.1016/S0140-6736(03)15114-8

Mäkinen, T., L. Jussila, T. Veikkola, T. Karpanen, M.I. Kettunen, K.J. Pulkkanen, R. Kauppinen, D.G. Jackson, H. Kubo, S. Nishikawa, et al. 2001. Inhibi-tion of lymphangiogenesis with resulting lymphedema in transgenicmice expressing soluble VEGF receptor-3. Nat. Med. 7:199–205. http://dx.doi.org/10.1038/84651

Article

The Rockefeller University Press $30.00J. Exp. Med. 2015 Vol. 212 No. 10 1529–1549www.jem.org/cgi/doi/10.1084/jem.20150165

1529

Cerebral blood flow and the blood–brain bar-rier (BBB) are essential for brain homeostasis. Both rely on an intact brain endothelium. Under normal conditions, the BBB is tightly sealed, restricting the access of blood constitu-ents to the brain. However, during inflamma-tory states, the BBB may become leaky and tissue perfusion may be compromised. Indeed, inflammatory mediators, such as TNF and IL-1, and bacterial cell wall components, such as LPS, are able to open the BBB and impair microvas-cular perfusion in the brain (Tsao et al., 2001; Argaw et al., 2006; Taccone et al., 2010). The

capacity to open the BBB is essential for mount-ing an inflammatory response in the brain and may have developed during evolution to clear neurotropic viruses or other pathogens from the CNS (Roy and Hooper, 2007). Several known mechanisms increase the permeability of the BBB during inflammation involving peri-cytes, astrocytes, and endothelial cells (Zlokovic, 2008; Obermeier et al., 2013). However, the mechanisms that maintain and repair endothelial cell function in inflammation are still elusive. If

CORRESPONDENCE Markus Schwaninger: markus.schwaninger@ pharma.uni-luebeck.de

Abbreviations used: ACh, ace-tylcholine; BBB, blood-brain barrier; IKK, IB kinase com-plex; IP, incontinentia pigmenti; NEMO, NF-B essential mod-ulator; OZ, oxozeaenol; PBEC, primary brain endothelial cell; PCA, posterior cerebral artery; ROS, reactive oxygen species; TAK1, transforming growth factor -activated kinase-1.

*D.A. Ridder and J. Wenzel contributed equally tothis paper.

Brain endothelial TAK1 and NEMO safeguard the neurovascular unit

Dirk A. Ridder,1* Jan Wenzel,1,3* Kristin Müller,1 Kathrin Töllner,4,5 Xin-Kang Tong,6 Julian C. Assmann,1 Stijn Stroobants,7 Tobias Weber,1 Cristina Niturad,1 Lisanne Fischer,1 Beate Lembrich,1 Hartwig Wolburg,8 Marilyn Grand’Maison,6 Panayiota Papadopoulos,6 Eva Korpos,9 Francois Truchetet,10 Dirk Rades,2 Lydia M. Sorokin,9 Marc Schmidt-Supprian,11 Barry J. Bedell,6 Manolis Pasparakis,12 Detlef Balschun,7 Rudi D’Hooge,7 Wolfgang Löscher,4,5 Edith Hamel,6 and Markus Schwaninger1,3

1Institute of Experimental and Clinical Pharmacology and Toxicology and 2Department of Radiation Oncology, University of Lübeck, 23562 Lübeck, Germany3German Research Centre for Cardiovascular Research (DZHK), Partner Site Hamburg/Lübeck/Kiel, 23562 Lübeck, Germany4Department of Pharmacology, Toxicology, and Pharmacy, University of Veterinary Medicine Hannover, 30559 Hannover, Germany5Center for Systems Neuroscience, 30559 Hannover, Germany6Montreal Neurological Institute, McGill University, Montreal QC H3A 0G4, Canada7Laboratory of Biological Psychology, KU Leuven, 3000 Leuven, Belgium8Institute of Pathology and Neuropathology, University Hospital Tübingen, 72076 Tübingen, Germany9Institute of Physiological Chemistry and Pathobiochemistry, University of Münster, 48149 Münster, Germany10Service de Dermatologie, CHR Metz-Thionville, 57100 Thionville, France11Department of Hematology and Oncology, Klinikum rechts der Isar, Technische Universität München, 81675 Munich, Germany12Institute for Genetics, University of Cologne, 50674 Cologne, Germany

Inactivating mutations of the NF-B essential modulator (NEMO), a key component of NF-B signaling, cause the genetic disease incontinentia pigmenti (IP). This leads to severe neurological symptoms, but the mechanisms underlying brain involvement were unclear. Here, we show that selectively deleting Nemo or the upstream kinase Tak1 in brain endothelial cells resulted in death of endothelial cells, a rarefaction of brain microvessels, cerebral hypoperfu-sion, a disrupted blood–brain barrier (BBB), and epileptic seizures. TAK1 and NEMO protected the BBB by activating the transcription factor NF-B and stabilizing the tight junction protein occludin. They also prevented brain endothelial cell death in a NF-B–independent manner by reducing oxidative damage. Our data identify crucial functions of inflammatory TAK1–NEMO signaling in protecting the brain endothelium and maintaining normal brain function, thus explaining the neurological symptoms associated with IP.

© 2015 Ridder et al. This article is distributed under the terms of an Attribution– Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http://creativecommons.org/ licenses/by-nc-sa/3.0/).

1530 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

Although we did not detect any obvious pathology on hemo-toxylin and eosin (H&E)– and Nissl-stained sections (not de-picted), immunostainings of the endothelial cell marker CD31 and of collagen IV as an integral basement membrane com-ponent demonstrated numerous empty basement membrane strands, also known as string vessels (Brown, 2010), in the CNS of Nemo/+ mice (Fig. 1 A). Mice with a cell type–specific deletion of Nemo in neurons and glia (Nestin-Cre;NemoFl, NemongKO) had no increase of string vessels, suggesting that these cell populations are not responsible for the vascular pa-thology (Fig. 1 B).

To delete Nemo selectively in brain endothelial cells, we generated a tamoxifen-inducible CreERT2 driver line (Slco1c1- CreERT2) that affords selective and efficient recombination in brain endothelial cells but not in other organs (Ridder et al., 2011). Primary brain endothelial cells (PBECs) from tamoxifen-treated Slco1c1-CreERT2; NemoFl mice (NemobeKO) showed an 60% reduction in NEMO protein (Fig. 1 C) and lower NF-B activity under basal conditions and after treatment with TNF in comparison to cells from NemoFl control mice (Fig. 1, D and E). NemobeKO mice had numerous string vessels throughout the CNS (Fig. 1, B and F). Heterozygous deletion of Nemo in brain endothelial cells (NemobeKO/wt mice) was asso-ciated with an intermediate formation of string vessels (Fig. 1 G). We found frequent string vessels in the brain of a patient who suffered from IP, confirming a similar vascular pathology in the human disease (Fig. 1 H).

Endothelial cell death has been proposed to be the main cause of string vessels (Brown, 2010). Supporting this concept, we found many brain endothelial cells in NemobeKO mice stain-ing for active caspase 3, a marker of apoptotic cells (Fig. 2 A). After inducing recombination with tamoxifen, the number of apoptotic endothelial cells peaked at day 15 (Fig. 2 B), whereas the length of string vessels continuously increased even after day 15 (Fig. 2 C), demonstrating that apoptosis precedes string vessel formation and supporting the concept of a causal relationship. The angiography with sulfo-NHS-LC- biotin showed that string vessels were not perfused (Fig. 2 D). Overall, brain vessels of NemobeKO mice showed a normal cov-erage by astrocytic endfeet, pericytes, and smooth muscle cells, as well as a normal composition of the basement mem-brane (Fig. 2 E). However, some string vessels were appar-ently not covered by astrocytic endfeet and pericytes (Fig. 2, F and G, arrowheads).

Nemo deletion in brain endothelial cells disrupts the BBB and causes epileptic seizuresAfter inducing deletion of Nemo in brain endothelial cells, mice transiently lost weight in comparison to controls (Fig. 3 A) and 35% of those mice died (Fig. 3 B). A possible cause of the increased mortality is vasogenic brain edema, as water content in the brain was higher in NemobeKO mice than in NemoFl controls (Fig. 3 C). In accordance with this notion, MR imaging showed that the lateral ventricles were signifi-cantly smaller in NemobeKO mice (6.1 ± 0.5 mm3) than in NemoFl controls (8.3 ± 0.3 mm3; P < 0.05 Student’s t test; Fig. 3 D).

these mechanisms fail, an excessive opening of the BBB may lead to detrimental consequences, as illustrated by neurological disorders ranging from Alzheimer’s disease to zoster encephalitis (Erickson and Banks, 2013). When BBB permeability is in-creased, extravasation of blood components interferes with normal neural function and causes epileptic seizures (Zlokovic, 2011; Obermeier et al., 2013). Even under physiological con-ditions, inflammatory mediators, such as TNF, IL-1, and LPS, are present at low levels in the CNS and in the blood-stream, posing a constant challenge to the maintenance of the BBB (Boulanger, 2009; Gregor and Hotamisligil, 2011).

A central pathway in inflammation is mediated by NF-B. By using distinct adaptor proteins, such as TRAF6 in the case of IL-1 (Lomaga et al., 1999), inflammatory mediators acti-vate the protein kinase TAK1 (Map3k7; Sakurai, 2012). Sub-sequently, TAK1 stimulates the IB kinase complex (IKK), which consists of the NF-B essential modulator (NEMO; IKK) and the enzymatic subunits IKK1 (IKK) and IKK2 (IKK; Clark et al., 2013). In turn, IKK activates the tran-scription factor NF-B, which is formed by five subunits, one of them p65 (also known as RelA). In numerous diseases as-sociated with vascular dysfunction and opening of the BBB, NF-B signaling is activated in brain endothelial cells (Tripathi et al., 2009; Jacob et al., 2011; Kielland et al., 2012). However, at this stage, it is unclear whether endothelial NF-B causes the vascular dysfunction or represents a protective transcrip-tional program in brain endothelial cells.

The genetic disease incontinentia pigmenti (IP), which is caused by inactivating mutations of the Nemo gene (Ikbkg), hints at a protective role of NF-B signaling in the CNS. IP is named for its skin manifestations but the predominant problem is CNS involvement that often starts acutely and leads to residual deficits (Goldberg, 2004; Meuwissen and Mancini, 2012). The most frequent neurological symptoms are epileptic seizures (Meuwissen and Mancini, 2012). How Nemo mutations disrupt normal human brain function has been enigmatic.

To explore the mechanisms underlying the neurological symptoms of IP, we investigated mice with a germline dele-tion or with cell type–specific deletions of Nemo in the CNS. Deletion of Nemo in brain endothelial cells resulted in disrup-tion of the BBB and endothelial cell dysfunction and death. Our data dissect the pathways that disturb brain endothelial function and lead to the neurological manifestations of IP when NEMO is inactivated.

RESULTSNemo deletion induces the death of brain endothelial cellsFemale heterozygous mice with a germline deletion of the X-chromosomal Nemo gene (Nemo/+) developed skin erup-tions after birth mimicking IP as has been reported previously(Makris et al., 2000; Schmidt-Supprian et al., 2000). Locomotionof Nemo/+ mice was slowed and they died at postnatal day(P)7–P10. To search for brain manifestations of the disease, we performed a histological evaluation of brains at P6–8.

JEM Vol. 212, No. 10

Article

1531

brains on H&E stains (unpublished data). The inspection of brain sections of NemobeKO and control mice injected with 2,000 kD dextran demonstrated large parts of 2,000 kD dex-tran outside of CD31-positive endothelial cells in the perivas-cular space, suggesting that the majority of NEMO-deficient vessels were leaky (Fig. 3 K). In NEMO-deficient mice, ves-sels permeable for dextrans outnumbered the active caspase 3–positive vessels, and most leaky vessels did not show active caspase 3 staining (Fig. 3 L, arrows) implying that the death of endothelial cells is not the main cause of BBB disruption.

Vasogenic brain edema can be lethal by compressing the brain stem, a condition that leads to an altered reactivity of

A dramatic increase in albumin and Ig levels in whole-brain extracts of NemobeKO mice confirmed that the BBB was dam-aged (Fig. 3 E). Ig extravasation occurred in all brain areas (Fig. 3 F) and increased significantly starting at day 10 after tamoxifen induction (Fig. 3, G and H), preceding the increase in brain weight at day 15 (Fig. 3 I). Staining revealed Ig local-ization, mainly in the parenchyma of the tissue (Fig. 3 H). The loosening of the BBB of NemobeKO mice was not size-selective, as intravenously injected fluorescent tracers of various sizes, ranging from 4 to 2,000 kD, were detected in higher concentra-tions in brain extracts of NemobeKO mice than of control animals (Fig. 3 J). However, we found no hemorrhages in NemobeKO

Figure 1. Deletion of Nemo in brain endothelial cells causes string vessel formation. (A) Representative immunostainings demonstrating string vessels (arrows) in a Nemo/+ mouse but not in a control mouse at P8. String vessels were identified as capillaries that have lost CD31-positive endothe-lial cells and only consist of the basement membrane protein collagen IV. (right) Quantification of string vessels in cortex at P6–8 as percentage of total vessel lengths. (B) String vessel quantification in NemongKO mice that are deficient of Nemo in neurons and glial cells, NemobeKO mice that are deficient of Nemo in brain endothelial cells, and NemoFl controls. (right) Representative immunofluorescent staining showing string vessels (arrows) in a NemobeKO mouse. (inset) High magnification of a string vessel. (C) Quantification of NEMO by Western blotting in PBECs of NemobeKO and NemoFl control mice. (D) Quantification of IB by Western blotting of PBECs from NemobeKO mice upon treatment with murine TNF (10 ng/ml). (E) NF-B activity as deter-mined by luciferase reporter assays in PBECs of NemoFl and NemobeKO mice. PBECs were treated with murine TNF for 4 h before monitoring luciferaseactivity. (F) Distribution of string vessels in various brain areas of NemobeKO and NemoFl mice. (G) Quantification of string vessel lengths in mice with no,heterozygous, or homozygous deletion of Nemo in brain endothelial cells. (H) Representative immunofluorescent staining for collagen IV and CD34 re-vealing string vessels (arrows) in the cortex of a patient suffering from IP (left), but not in the cortex of an age-matched subject dying from an unrelateddisease (right). For further description of the IP patient, refer to (Bachevalier et al., 2003). Data are shown as means ± SEM (n = 5–8). *, P < 0.05; **, P <0.01; ***, P < 0.001, determined by Student´s t test (A-C), two-way ANOVA with Bonferroni´s post test (D-F), or one-way ANOVA with Tukey’s post test (G).Mice were investigated 15–18 d after start of tamoxifen treatment (B and F–G). Bars, 50 µm.

1532 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

Figure 2. Deletion of Nemo in brain endothelial cells causes brain endothelial cell death. (A) Representative immuno-staining for active caspase 3 (arrows) in the cortex of a NemobeKO mouse. (B and C) Quantification of active caspase 3–positive endothelial cells (B) and string vessels (C) at seven time points before and after the start of tamoxifen treatment. The difference in active caspase 3–positive cells and string vessel length between NemobeKO and NemoFl mice before tamoxifen induction (day 0) is due to a low recombination rate in the Slco1c1-CreERT2 driver line without tamoxi-fen treatment (not depicted). (D) Staining of brain sections for i.v.-injected sulfo-NHS-LC-biotin to detect perfused vessels. String vessels are indicated (arrows). (E) Quantifi-cation of vessel lengths immunostained for basement membrane components (pan-laminin, laminin 5, and collagen IV), as well as pericyte (CD13), smooth muscle cell (SMA), or astrocytic endfeet markers (aqua-porin 4). (F and G) Representative immuno-fluorescent stainings for aquaporin 4 (AQ4; F) and CD13 (G). String vessels covered by astrocytic endfeet and pericytes are indi-cated by arrows, while string vessels not covered are indicated by arrowheads. For the staining depicted in F and G, goat anti-CD31 antibodies had to be used (CD31*), leading to more unspecific staining. Data are shown as means ± SEM (n = 5–8). *, P < 0.05; **, P < 0.01; ***, P < 0.001, determined by two-way ANOVA with Bonferroni´s post test (B and C). Mice were investigated 15 d after start of tamoxifen treatment (A and D–G). Bars, 50 µm.

pupils in humans and mice (Matullo et al., 2010). However, pupillary reactivity was normal in NemobeKO mice (Fig. 3 M), arguing against brain stem compression as a cause of the in-creased mortality.

The extravasation of albumin has been reported to acti-vate astrocytes (Obermeier et al., 2013). We found increased GFAP staining, a sign of astrocytic activation, in NemobeKO mice (Fig. 3 N). Because BBB dysfunction and astrocytic

JEM Vol. 212, No. 10

Article

1533

Figure 3. Deletion of Nemo in brain endothelial cells disrupts the BBB and causes epileptic seizures. (A) Body weight of NemobeKO and NemoFl mice was measured before and after induction of recombination by tamoxifen (day 0). Values are presented as percentage of the body weight at day 0

1534 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

(n = 7–8). (B) Survival curves of NemobeKO and control mice (n = 41–46) after treatment with tamoxifen are shown. P < 0.01 (log-rank test). (C) Water content in brains of NemobeKO mice compared with control animals (n = 3–5). (D) Representative coronal anatomical MRI scans of a NemobeKO and a NemoFl control mouse. Arrows indicate the lateral ventricles. (E) Determination of IgG brain levels by Western blotting. (F) Quantification of Ig immunostaining in various brain areas of NemoFl and NemobeKO mice (n = 7–8). (G) Quantification of Ig immunostaining in the cortex at seven time points before and after the start of tamoxifen treatment (n = 7–8). (H) Representative images of immunofluorescent Ig detection co-stained with CD31 in the cortex at day 29 after tamoxifen treatment. (I) Brain weights at six time points before and after tamoxifen treatment. (J) Concentrations of fluorescently labeled dextrans of various molecular sizes in brain lysates after i.v. injections (n = 7–8). Fluorescence intensity in tissue extracts is expressed relative to NemoFl controls. (K) Imaging of fluorescent 2,000-kD dextran in brain slices of NemoFl and NemobeKO mice after staining of endothelial cells with anti-CD31. Dextranwas i.v. injected. (L) Representative image of fluorescent 2,000-kD dextran in a NemobeKO brain slice after staining with anti-CD31 and anti-active caspase3. Arrows, vessels negative for active caspase 3 staining. (M) Pupillary diameters were video-recorded while increasing the illumination rapidly from 1 lux(Max) to 500 lux (Min). Each dot represents the mean of two measurements for one mouse. (N) Quantification of GFAP staining in the cortex of NemobeKO

mice and control animals. (O) Representative EEG trace of a focal seizure recorded by cortex screws above the hippocampus of a NemobeKO mouse. Theseizure was accompanied by unspecific movement at the same place followed by a stretching of the body. (P) Representative EEG trace of a generalizedseizure recorded by fronto-parietal cortex screws in a NemobeKO mouse. The seizure was accompanied by forelimb clonus, rearing, and loss of balance withrunning and bouncing. Data are shown as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001, determined by Student´s t test (C and N), and two-wayANOVA with Bonferroni´s post test (F, G, I, and J). Mice were investigated 19–22 d (C) or 11–16 d (D, F, and J–P) after start of tamoxifen treatment. Bars,50 µm (H, L, K, top); 10 µm (K, bottom); and 5 s (O and P).

brains of NemobeKO mice (unpublished data). These data dem-onstrate that NEMO signaling in brain endothelial cells en-sures normal basal perfusion and neurovascular coupling.

Nemo deletion in brain endothelial cells has behavioral consequencesTo analyze whether disruption of the BBB and reduced blood flow translate into functional consequences, we performed behavioral tests. NemobeKO mice displayed intact contextual fear conditioning, reflecting normal learning and memory (Fig. 5 A). However, NemobeKO mice showed generally in-creased freezing during cued fear conditioning independent of the presence of the cue, although freezing behavior during the habituation phase, fear conditioning, and the context fear testing was normal, which may relate to increased state anxi-ety and/or overgeneralization of the fear response. Increased state anxiety is indeed consistent with two other observations. First, NemobeKO mice spent less time in the inner zone of the open field than NemoFl mice when the vascular pathology had developed (Fig. 5 B). Second, there was a trend that NemobeKO mice spent less time in the open arms of the elevated plus maze than NemoFl animals (Fig. 5 C), whereas their overall lo-comotion was unaffected (Fig. 5 D). In addition to elevated anxiety, NemobeKO mice showed less social behavior in the sociability test, as demonstrated by a significantly lower rela-tive number of approaches to and relative time in the proxim-ity of another mouse (Fig. 5 E).

TRAF6, TAK1, NEMO, and NF-B p65 differentially regulate brain endothelial function and survivalNEMO is an essential component of the canonical pathway activating NF-B. To explore the role of this pathway in brain endothelial cells, we deleted the upstream adaptor protein TRAF6 and the upstream kinase TAK1, as well as the p65 subunit of NF-B, by crossing the Slco1c1-CreERT2 strain with mice carrying loxP-flanked Traf6, Tak1, or p65 alleles and induced recombination by injecting tamoxifen (Traf6beKO, Tak1beKO, or p65beKO). Western blotting revealed that Traf6, p65

activation can trigger epileptic seizures, we monitored NemobeKO and control mice for seizures by combined EEG/video re-cording. 10 out of 12 NemobeKO animals, but none of the 13 control mice, displayed epileptic seizure activity (P < 0.0001, Fisher’s exact test). 3 out of 12 mice had periods of short focal seizure activity (<10 s duration; Fig. 3 O), and 7 out of 12 mice suffered from generalized epileptic seizures of >10 s dura-tion (Fig. 3 P). Two of the NemobeKO mice died during seizures, suggesting that epilepsy is a cause of the increased mortality.

Nemo deletion in brain endothelial cells compromises cerebrovascular reactivity, basal cerebral blood flow, and neurovascular couplingIn Slco1c1-CreERT2 mice, recombination occurs in endothelial cells of small and large diameter vessels of the brain (Fig. 4 A). Therefore, we investigated whether deletion of Nemo also af-fects cerebrovascular reactivity. Instead of dilating in response to acetylcholine (ACh; Zhang et al., 2013), posterior cerebral arteries (PCAs) isolated from NemobeKO mice weakly con-stricted (Fig. 4 B). In contrast, vessels responded normally to the NO donor sodium nitroprusside (SNP; Fig. 4 C), the po-tent vasoconstrictor endothelin-1 (ET-1; Fig. 4 D), and the NO synthase inhibitor L-NNA (Fig. 4 E; Tong et al., 2012; Zhang et al., 2013), demonstrating that smooth muscle func-tion and basal NO production were intact but endothelial-mediated relaxation was impaired.

The rarefaction of capillaries and impaired cerebrovascu-lar reactivity in NemobeKO mice may affect cerebral blood flow. Indeed, arterial spin labeling (ASL) perfusion MRI revealed a significantly reduced resting blood flow throughout the CNS of NemobeKO mice when compared with NemoFl mice (Fig. 4, F and G). Furthermore, the blood flow increase in response to whisker stimulation of NemobeKO mice was impaired (Fig. 4, H and I). The reduced hyperemia was not explained by struc-tural or functional changes of neurons as whisker pad stimula-tion evoked similar local field potentials in the barrel cortex of NemobeKO and NemoFl mice (Fig. 4, J–L). Moreover, we found no signs of neuronal cell death or loss of synapses in

JEM Vol. 212, No. 10

Article

1535

in brain endothelial cells TAK1 activates IKK and, subse-quently, the NF-B subunit p65 to maintain the BBB. The presence of other NF-B subunits that are able to compen-sate for the absence of p65 likely explains why p65 deletion had a less pronounced effect on Ig extravasation than Nemo deletion and had no effect on brain weight. The TAK1–IKK–NF-B pathway sealing the BBB appears to be independent of TRAF6, excluding upstream stimuli known to exert their effects via TRAF6, e.g., IL-1 (Lomaga et al., 1999).

By measuring the length of string vessels, we found a marked increase in Tak1beKO and NemobeKO mice, but not in Traf6beKO and p65beKO animals (Fig. 6 C). Conversely, the length of intact

(Fig. 5, F and G), and Tak1 (Ridder et al., 2011) had been ef-ficiently deleted in PBECs from mice of the different geno-types. Deletion of p65 in brain endothelial cells reduced basal and TNF-stimulated NF-B activity (Fig. 5 H) and the ex-pression of p65 in cortical vessels (Fig. 5 I).

Brain weight, as an indicator of brain edema, and brain Ig levels were not altered by deletion of Traf6 in brain endothe-lial cells (Fig. 6, A and B). However, both parameters were el-evated in Tak1beKO mice just as in NemobeKO animals (Fig. 6, A and B). In p65beKO mice, we found no change of brain weight and a smaller increase in Ig staining than in NemobeKO and Tak1beKO animals (Fig. 6, A and B). These findings suggest that

Figure 4. Deletion of Nemo in brain endothelial cells disturbs cerebrovascular reactivity and reduces basal cerebral blood flow and neuro-vascular coupling. (A) tdTomato expression after recombination in small and large CD31-positive vessels throughout the CNS of Slco1c1-CreERT2; Ai14 mice. A representative picture from the cortex is shown. Bar, 120 µm. (B–E) Responses of PCA segments of NemobeKO and NemoFl mice to Ach (B), the NO donor sodium nitroprusside (SNP; C) endothelin-1 (ET-1; D), and the NO synthase inhibitor N-nitro-L-arginine, L-NNA; E) are shown (n = 4–7). (F) Cere-bral perfusion as depicted by surface-projected, group-averaged ASL-MRI maps (n = 12–13). (G) Quantification of ASL-MRI tissue perfusion in three brain areas of NemobeKO and NemoFl mice (n = 12–13). (H and I) Cerebral blood flow (CBF) responses in the somatosensory cortex of NemobeKO and NemoFl mice evoked by whisker stimulation are shown (n = 6–7). The bar indicates stimulation. (J) Local field potentials were recorded in the somatosensory barrel cortex representing the contralateral whisker pad. (top) Basal neuronal activity. (bottom) Responses upon electrical stimulation of the whisker pad. Dots indicate stimulation. Note the different time scale for top and bottom parts. (K) Representative local field potentials of NemobeKO and NemoFl mice are shown. (L) Quantification of the amplitude of the local field potentials. Data are shown as means ± SEM. **, P < 0.01; ***, P < 0.001, determined by re-peated measures ANOVA with Bonferroni’s post test (B) and Student’s t test with Bonferroni correction (G and I). Mice were investigated 14–24 d after start of tamoxifen treatment.

1536 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

Figure 5. Deletion of Nemo in brain endothelial cells leads to more anxiety-related behavior and less sociability in mice. (A) Quantification of the relative freezing time in the different parts of the contextual fear conditioning test is shown (n = 13–18). (B) The open field test was performed at 7 time points before and after the start of tamoxifen treatment (n = 7–8). Injection of tamoxifen during the first 5 d led to a decreased time in the inner zone in both genotypes. (C and D) The time spent in the open arms of an elevated plus maze test is shown (C), as well as the overall activity (D) measured by total counts of entries (n = 13–18). (E) Sociability was tested by measuring the relative number of approaches and time spent in proximity of a conspe-cific in NemobeKO and NemoFl mice (n = 13–18). (F) TRAF6 was measured by Western blotting of PBECs from Traf6beKO and Traf6Fl control mice. (G) The amount of NF-B p65 was determined by Western blotting of PBECs from wild-type mice, p65beKO and p65Fl control mice. Note that endogenous p65 is replaced by a p65-eGFP fusion protein in p65Fl PBECs resulting in a higher molecular weight when compared with wild-type PBECs. (H) NF-B activity as determined by luciferase reporter assays in PBECs of p65Fl and p65beKO mice. PBECs were treated with murine TNF (10 ng/ml) or solvent for 4 h before monitoring luciferase activity (n = 5–6). (I) Immunostainings against GFP and CD31 of p65beKO and p65Fl mice are shown. To detect the GFP signal in

JEM Vol. 212, No. 10

Article

1537

BBB, we measured the water content and the BBB permea-bility by intravenously injecting the tracer sodium fluorescein. In Tak1beKO mice, the water content and the BBB permeabil-ity to fluorescein were elevated, confirming increased leaki-ness of the BBB (Fig. 7, A and B). TAK1 is upstream of the IKK complex. In the canonical NF-B pathway, IKK2 phos-phorylates and activates NF-B. To analyze whether IKK2 mediates the effects of TAK1 on BBB integrity, we expressed a constitutively active mutant of IKK2 (IKK2CA) specifi-cally in brain endothelial cells by crossing mice carrying the R26-StopFLIkk2ca (Ikk2caFl) allele (Sasaki et al., 2006) with Slco1c1-CreERT2 animals. IKK2CA activated NF-B, as shown by a lower steady-state level of the inhibitor IB in PBECs (Fig. 7 C) and elevated expression of the NF-B target gene VCAM-1 in brain endothelial cells of Tak1beKOIKK2beCA mice

CD31-stained vessels was shorter in Tak1beKO and NemobeKO mice but not in Traf6beKO and p65beKO animals (Fig. 6 D), confirming that the appearance of string vessels reflects a loss of capillary en-dothelium. In summary, deletion of Tak1 and Nemo led to a loss of brain vessels. Whereas the absence of p65 did not trigger endothelial cell death, it still caused increased BBB leakage, arguing for two independent mechanisms and confirming the observation that leakiness can occur in the absence of endothelial cell death (Fig. 3 L). The lack of a phenotype in Traf6beKO mice excludes an unspecific effect of the CreERT2 itself.

TAK1–NEMO–IKK2 signaling regulates BBB permeability, but not brain endothelial survivalTo verify that the increased brain weight and the Ig extravasa-tion in the brain of Tak1beKO mice reflect a disruption of the

endothelial cells we perfused animals with 4% paraformaldehyde without post-fixation before staining. Representative images from one out of three mice per genotype are shown. Bar, 100 µm. Data are shown as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001, determined by repeated measures ANOVA with Fisher LSD post test (A), Student’s t test (C), Mann-Whitney rank sum test (E), and two-way ANOVA with Bonferroni´s post test (B and H). Mice were investigated 19–24 d (E) or 28–32 d (A, C, and D) after start of tamoxifen treatment.

Figure 6. Brain endothelial deletion of the upstream kinase Tak1 but not of the adaptor protein Traf6 phenocopies Nemo knockout. Down-stream of NEMO, NF-B p65 regulates BBB permeability but not endothelial survival. (A) Brain weights of mice lacking different parts of the NF-B sig-naling cascade in brain endothelial cells are shown. (B) Quantifications of immunofluorescent stainings of Ig in the cortex of Traf6beKO, Tak1beKO, NemobeKO, p65beKO, and respective control mice. (C) String vessel lengths are quantified in the brains of the different mouse lines. (D) Quantifications of the total CD31-positive vessel lengths in the cortex. Data are shown as means ± SEM (n = 3–8 mice per genotype). *, P < 0.05; **, P < 0.01; ***, P < 0.001, deter-mined by Student´s t test. Mice were investigated 15–16 d after the start of tamoxifen treatment.

1538 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

junction proteins by immunostaining and Western blotting. Although protein levels of claudin-5 and ZO-1 were not changed (Fig. 8, A, B, F, and G), occludin expression was lower in the brains of NemobeKO (Fig. 8, C–E) and Tak1beKO mice (Fig. 8 H) than in controls. IKK2CA overexpression restored occludin levels in TAK1-deficient mice (Fig. 8 H). Thus, TAK1 signals via IKK to sustain occludin expression and to preserve the integrity of the BBB. In support of a role of NF-B signaling, occludin expression was also reduced in p65beKO mice (Fig. 8 I).

Dependence on NF-B p65 suggests that occludin ex-pression is regulated transcriptionally. Because occludin mRNA did not reflect the marked changes of its protein level in Tak1beKO mice (Fig. 8 J), we investigated whether an indirect mecha-nism regulates occludin protein. To evaluate the role of matrix metalloproteinases and the proteasome, which are able to de-grade occludin (Cummins, 2012), we used small-molecule in-hibitors in vitro. In PBECs of Tak1beKO mice or in wild-type PBECs treated with the TAK1 inhibitor 5Z-7-oxozeaenol (OZ), protein levels of occludin were reduced (Fig. 8, K, L and M). Down-regulation of occludin in OZ-treated cells

(Fig. 7 D). Overexpression of two copies of the Ikk2ca allele largely prevented the increase in brain weight and in brain Ig levels in Tak1beKO mice, confirming that TAK1 signals via IKK to maintain an intact BBB (Fig. 7, E and F). Additionally, over-expressing IKK2CA reduced the mortality of Tak1beKO mice (Fig. 7 G), supporting a causative relationship between BBB disruption and death of the mice. Although IKK2CA compen-sated for the TAK1 deficiency with respect to water content, Ig extravasation, and mortality, it did not prevent brain endothelial cell death induced by deleting Tak1, as the length of string vessels was not affected by IKK2CA overexpression (Fig. 7 H). This dissociation further supports the concept that different signaling pathways are effective downstream of TAK1 and NEMO in brain endothelial cells: an IKK2- and p65-dependent pathway stabilizing the BBB and an IKK2- and p65-independent pathway preventing endothelial cell death.

TAK1–NEMO–IKK2 signaling stabilizes the tight junction protein occludinAs the barrier properties of brain endothelial cells depend on their tight junctions, we investigated expression levels of tight

Figure 7. TAK1–NEMO signaling promotes BBB integrity via IKK2, whereas the effect on brain endothelial survival is independent of IKK2. (A) Brain water content in Tak1beKO and Tak1Fl mice is shown (n = 7–8). (B) BBB permeability in Tak1beKO and Tak1Fl mice was measured by fluorescein ex-travasation (n = 4). (C) Levels of the inhibitory IB in PBECs from Ikk2beCA mice were determined by Western blotting (n = 6). (D) Quantification of thevessel lengths expressing the endothelial NF-B target gene VCAM-1 in the cortex of control, Tak1beKO, and Tak1beKOIkk2beCA mice are shown (n = 5–9).In Tak1beKOIkk2beCA mice, Tak1 is deleted and two copies of the Ikk2ca allele are overexpressed in brain endothelial cells. (E and F) Brain weight (E) and Igstaining (F) was quantified in control, Tak1beKO, and Tak1beKOIkk2beCA mice (n = 9–14). (G) Survival curves of Tak1beKO and Tak1beKOIkk2beCA mice are shown(n = 21). P < 0.01 (log-rank test). (H) Quantification of string vessel lengths in the cortex 15 d after start of tamoxifen treatment (n = 10–14). Data areshown as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001, determined by two-way ANOVA with Bonferroni´s post test (B), one-way ANOVA withTukey´s post test (D–F and H), and Student´s t test (A and C). Mice were investigated 7 d (B) or 15 d (A, D–F, and H) after start of tamoxifen treatment.

JEM Vol. 212, No. 10

Article

1539

Figure 8. TAK1–NEMO inhibition leads to decrease of occludin and its deubiquitinase A20. (A–C) Quantifications of stainings of tight junction proteins ZO-1 (A), claudin 5 (B), and occludin (C). Sections were co-stained with anti-CD31 (n = 7–8). (D) Representative images of occludin staining are shown. (insets) High magnification of a capillary. (E) Western blotting of occludin in brain lysates of NemobeKO and control mice (n = 6). (F–I) Expression of claudin 5 (F), ZO-1 (G), and occludin (H and I) as determined by Western blotting in brain lysates of different mouse lines (n = 6–14). (J) Occludin mRNA levels were analyzed by real-time RT-PCR in brain lysates of Tak1beKO, Tak1beKOIkk2beCA, and control mice (n = 10–14). (K) Occludin protein levels in PBECs of Tak1Fl and Tak1beKO mice detected by Western blotting. (L–M) Effects of the proteasome inhibitors MG-132 (10 µM; L) or lactacystin (10 µM; M), as well as of the metalloproteinase inhibitor GM-6001 (10 µM; L) on occludin expression in PBECs was measured after treatment with the TAK1 inhibitor OZ (1 µM) for 24 h (n = 8–10) by Western blotting. (N–Q) Expression of A20, an occludin-targeting deubiquitinase, was measured in PBECs of Tak1beKOIkk2beCA, Tak1beKO (N and O) and p65beKO (P and Q) mice as determined by real-time RT-PCR (N and P) and Western blotting (O and Q; n = 5–6). Data are shown as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001, determined by one-way ANOVA with Tukey’s post test (H and N), Student´s t test (C, E, I, K, P, and Q), and two-way ANOVA with Bonferroni´s post test (L and M). Mice were investigated 15 d after start of tamoxifen treatment. Bar, 100 µm.

1540 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

lower in TAK1- and in p65-deficient PBECs, whereas IK-K2CA expression in PBECs of Tak1beKOIKK2beCA mice effi-ciently induced A20 mRNA (Fig. 8, N–Q) in accordance with regulation of A20 expression by the TAK1–IKK–p65 pathway (Catrysse et al., 2014). Collectively, these data show that TAK1 and NEMO prevent occludin degradation by the proteasome, possibly by inducing A20 through NF-B, and stabilize the BBB.

was reversed by the proteasome inhibitors MG-132 or lacta-cystin but not by the broad metalloproteinase blocker GM-6001 (Fig. 8, L and M), demonstrating that the proteasome degrades occludin when TAK1 is inhibited. Ubiquitination labels occludin for proteasomal degradation (Cummins, 2012). An endothelial enzyme that deubiquitinates occludin and protects it from proteasomal degradation is A20 (Tnfaip3; Kolodziej et al., 2011). A20 mRNA and protein levels were

Figure 9. TNF mediates death of brain endothelial cells upon disruption of TAK1 signaling. (A) Electron microscopy showing signs of endothelial cell death and a thickened basal lamina (bl) in brains of Tak1beKO mice. (left) Normal capillary of a Tak1Fl mouse. (middle) Endothelial cell (e) of a Tak1beKO mouse that is swollen; the lumen (asterisks) is obliterated; pericytes demonstrate signs of necrotic cytoplasm (arrows). tj, tight junction. (right) Endothe-lial cell (e) filled with electron-dark vacuoles; the basal lamina (bl) is thickened; lumen and tight junctions are no longer visible. (B) A representative immuno-staining for active caspase 3 (arrows) and CD31 of Tak1beKO cortex is shown. (C) Tnf mRNA was determined by real-time RT-PCR in brain lysates (n = 6). (D) Effects of TNF (50 ng/ml) and the TAK1 inhibitor OZ (1 µM) on cell death in brain endothelial bEnd.3 cells after 24 h. BHA, butylate hydroxyanisole(100 µM). #, P < 0.001 compared with groups treated with both TNF and OZ (n = 11). (E–G) The effect of neutralizing TNF antibodies (-TNF Ab) on thenumber of active caspase 3–positive endothelial cells (E), string vessel formation (F), and IgG extravasation determined by Western blotting (G) in Tak1beKO

and Tak1Fl mice is shown (n = 3–8). Data are shown as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001, determined by Student´s t test (C) and two-way ANOVA with by Bonferroni’s post test (D–G). Mice were investigated 10 d (A) or 14–15 d (B–G) after start of tamoxifen treatment. Bars: 1 µm (A);50 µm (B).

JEM Vol. 212, No. 10

Article

1541

TNF has been reported to induce cell death when TAK1–NEMO signaling is inhibited (Vanlangenakker et al., 2011; Xiao et al., 2011). Interestingly, in brains of Tak1beKO mice, Tnf expression was elevated (Fig. 9 C). In brain endothelial bEnd.3 cells, TNF strongly triggered cell death if TAK1 signaling was inhibited by OZ (Fig. 9 D). Therefore, we treated Tak1beKO mice with TNF-neutralizing antibodies. Anti-TNF partially protected against brain endothelial cell death as revealed by less endo-thelial cells positive for active caspase 3 and shorter lengths of string vessels (Fig. 9, E and F). However, TNF neutralization in the same animals did not significantly reduce BBB disruption as shown by a similar Ig brain content (Fig. 9 G).

Mechanisms of brain endothelial cell death upon disruption of TAK1–NEMO signalingIn electron microscopy, about half of the capillaries in Tak1beKO mice were pathologically altered, showing a substantial thick-ening of the basement membrane (160–250 nm compared with a normal thickness of 40–60 nm; Fig. 9 A). Vessel lu-mina were obliterated due to swelling of endothelial organ-elles and cytoplasm, consistent with a necrotic form of endo-thelial cell death (Fig. 9 A). Together with the observed signs of apoptosis, i.e., staining for active caspase 3 in Tak1beKO and NemobeKO mice (Figs. 2, A and B; and 9 B), this suggests that elements of necrosis and apoptosis occur in parallel.

Figure 10. ROS mediate death of brain endothelial cells upon disruption of TAK1 signaling. (A and B) A representative immunostaining for 4-HNE, a marker for lipid peroxidation, in a Tak1beKO mouse is shown (A). Quantification is shown in B (n = 5–7). (C and D) Effects of a diet containing theantioxidant BHA (0.7%) on string vessel formation (C) and occludin protein levels (D) in Tak1beKO mice (n = 7–11). (E) Response of PCA segments of NemoFl

control and NemobeKO mice to the TRPV4 agonist GSK1016790A (GSK; n = 4–7). (F) The effect of BHA on the response of vessels isolated from NemobeKO

mice were tested. Concentration-response curves for ACh are shown (n = 3–4). (G) Scheme summarizing the proposed mechanisms how TAK1 and NEMOprevent brain endothelial cell death and maintain the BBB. Ub, ubiquitination. Data are shown as means ± SEM. *, P < 0.05; ***, P < 0.001, determined bytwo-way ANOVA with Bonferroni’s post test (C), repeated measures ANOVA with Bonferroni’s post test (E and F), or Student´s t test (B). Mice were investi-gated 14–15 d (A–D), 14–18 d (E), or 19–22 d (F) after start of tamoxifen treatment. Bar, 50 µm.

1542 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

whereas deletion in brain endothelial cells (NemobeKO) induced a phenotype that resembles the clinical disease in several aspects. In accordance with a reduced motivational drive, NemobeKO mice showed signs of anorexia and less socia-bility but more anxiety-like behavior. Very much like IP pa-tients, NemobeKO mice had focal and generalized tonic-clonic seizures that seem to cause the increased mortality in the mouse model. Further support for the notion that NemobeKO mice are a model of the CNS involvement in IP patients comes from the detection of string vessels in the brain of an IP pa-tient, although we were only able to investigate a single patient at this stage.

Rarefaction of capillaries and string vessels in NemobeKO mice are explained by the death of endothelial cells. The signs of endothelial necrosis and apoptosis in the absence of TAK1 or NEMO were neither rescued by overexpression of IKK2CA nor mimicked by deletion of p65 NF-B. In con-trast to the general notion that NEMO exerts its effects via NF-B, these findings indicate that TAK1 and NEMO can prevent cell death independent of NF-B, at least of the p65 subunit and its main activator in the canonical pathway, IKK2. Previous work has already shown in vitro that TAK1 and NEMO can prevent TNF-induced cell death without the need for NF-B–mediated gene transcription (O’Donnell et al., 2012; Dondelinger et al., 2013). Our study now pro-vides the first in vivo evidence for this novel function of TAK1 and NEMO in brain endothelial cells. Interestingly, the effects of TAK1 or NEMO deficiency on survival differ between cell types. Neural cells, for instance, are resistant to TAK1 or NEMO deficiency (van Loo et al., 2006; Goldmann et al., 2013). In peripheral endothelial cells TAK1 deficiency leads to vessel regression and reduced cell migration during embryogenesis (Morioka et al., 2012). Based on the current data, we cannot exclude that NEMO has a similar role in pe-ripheral endothelial cells. However, the inducible deletion of Nemo in peripheral endothelial cells of adult mice prevented atherosclerotic changes (Gareus et al., 2008). Accordingly, IP patients do not suffer from peripheral vascular problems. In contrast, in the brain, deletion of Tak1 or Nemo compro-mised the survival of endothelial cells, underlining the dis-tinct features of these cells in the CNS.

In NemobeKO and Tak1beKO mice, death of endothelial cells led to a rarefaction of capillaries with a shorter total vessel length. Capillary pathology translates into hypoperfusion under basal conditions and, intriguingly, it may also explain why neurovascular coupling was impaired. According to a re-cent reanalysis of the flow–diffusion equation, an increased heterogeneity of capillary transit time caused by capillary pa-thology is associated with a deficit in neurovascular coupling (Ostergaard et al., 2013). Finally, Nemo deletion also affected large cerebral arteries. Endothelium-dependent stimuli were less effective in dilating the PCA of NemobeKO mice than of control animals. However, the antioxidant BHA rapidly nor-malized dilation in NemobeKO mice, suggesting that ROS impair endothelial function in the absence of endothelial NEMO. Thus, endothelial dysfunction of cerebral arteries and

Previous studies have demonstrated that TAK1–NEMO signaling protects against TNF-induced death of other cell types by down-regulating reactive oxygen species (ROS; Xiao et al., 2011; O’Donnell et al., 2012; Dondelinger et al., 2013). In Tak1beKO mice, we observed an increase in brain vessels positive for the lipid peroxidation product 4-hydroxy-nonenal (4-HNE), indicating ROS production (Fig. 10, A and B). The antioxidative agent butylated hyroxyanisole (BHA) inhibited TNF-induced cell death when TAK1 was blocked in vitro (Fig. 9 D) and reduced the occurrence of string vessels in Tak1beKO mice in vivo (Fig. 10 C); this gives evidence for the involvement of ROS in endothelial cell death. In contrast, BHA treatment did not prevent occludin down-regulation in Tak1beKO mice (Fig. 10 D). Thus, when TAK1–NEMO signal-ing is inhibited, endothelial cell death is partially dependent on TNF and ROS but not on IKK2 and p65.

Because ROS are a well-known cause of endothelial dys-function in arteries (Girouard and Iadecola, 2006), we asked whether ROS production could also underlie the impaired cerebrovascular reactivity in NemobeKO mice. The ROS-sensitive endothelial cation channel TRPV4 largely mediates endothelial dilation induced by acetylcholine (Sonkusare et al., 2012; Zhang et al., 2013). Dilations of PCAs in response to GSK1016790A (GSK), a selective TRPV4 agonist, were impaired if the vessels were isolated from NemobeKO mice, in accordance with the notion that ROS mediate endothelial dysfunction upon Nemo deletion (Fig. 10 E). Indeed, scavenging ROS by pretreating PCA segments of NemobeKO mice with BHA fully rescued the impaired dilatory response to acetylcholine (Fig. 10 F).

Together, NEMO and TAK1 use distinct mechanisms that differentially depend on IKK2 and p65 NF-B, as well as TNF and ROS to support the survival of brain endothelial cells, regu-late cerebrovascular reactivity, and protect the BBB (Fig. 10 G).

DISCUSSIONPatients with CNS involvement of IP usually present with the clinical picture of an acute encephalopathy, signs of which in-clude a reduced motivational drive and epileptic seizures (Hennel et al., 2003; Abe et al., 2011; Meuwissen and Mancini, 2012). Diffusion-weighted MR imaging in IP patients indi-cated cytotoxic brain edema that occurs in cerebral ischemia (Hennel et al., 2003; Abe et al., 2011). However, large cerebral arteries are not occluded in most IP patients (Hennel et al., 2003; Maingay-de Groof et al., 2008). This seeming contradic-tion is resolved by the finding of small vessel pathology. IP is caused by heterozygous mutations in the X-chromosomal Nemo gene and Nemo/+ mice mimic the skin manifestations of IP (Makris et al., 2000; Schmidt-Supprian et al., 2000). In the brain of Nemo/+ mice, we have detected so-called string ves-sels as a conspicuous sign of small vessel pathology that corre-sponds well to the clinical imaging data. String vessels are empty basement membrane strands without inner endothelial cells and without perfusion. A vascular cause of the neurolog-ical symptoms in IP is supported by the finding that cell type–specific deletion of Nemo in neurons and glial cells did not lead to any obvious deficits (Fig. 1 B; van Loo et al., 2006),

JEM Vol. 212, No. 10

Article

1543

findings for the role of NF-B in the permeability of periph-eral endothelial cells (Kisseleva et al., 2006; Ye et al., 2008).

Disruption of the BBB occurs in several pathological con-ditions, including traumatic brain injury, cerebral ischemia, and infections, all of which are associated with an increased risk of epilepsy. The concept that BBB breakdown induces epileptic seizures (Seiffert et al., 2004; Obermeier et al., 2013) is sup-ported by our observation that brain endothelial-specific de-letion of Nemo leads to epilepsy in NemobeKO mice. After leaking through the BBB, albumin triggers epileptic activity by trans-forming astrocytes (Ivens et al., 2007). Thus, the high amount of albumin in brain tissue of NemobeKO mice and the activation of astrocytes are likely causes of epileptic seizures and may serve as an explanation for the relationship between Nemo muta-tions and the development of epilepsy in IP patients.

A dense vascular network and the BBB are essential pre-requisites for brain function. Proinflammatory mediators open the BBB and disrupt microvascular perfusion, and therefore put brain homeostasis at risk. Our data now show that brain endothelial cells actually rely on an inflammatory pathway, TAK1–NEMO signaling, to maintain normal brain func-tion. Thus, brain endothelial-specific deletion of Nemo in mice mimics the neurological symptoms of IP. On the basis of these data, neutralizing TNF antibodies, antioxidative sub-stances such as BHA, or gene therapy with AAV vectors that selectively target brain endothelial cells (Chen et al., 2009) may provide rational treatment options for the neurological symptoms of IP.

MATERIALS AND METHODSMice. All mouse lines were established on a C57BL/6 background. We used littermate mice that were sex- and age-matched between experimental groups. In most experiments, mice were between 6 and 18 wk of age. All animal experiments were approved by the local animal ethics committee (Regierungspräsidium Karlsruhe; Ministerium für Landwirtschaft, Umwelt und ländliche Räume, Kiel, Germany). Heterozygous Nemo knockout mice (Nemo/+) were generated by crossing NemoFl mice (Schmidt-Supprian et al., 2000) with CMV-Cre mice (Schwenk et al., 1995). For neuronal and glial knockout of Nemo (NemongKO) we crossed NemoFl mice with the Nestin-Cre line (Tronche et al., 1999). Brain endothelial-specific knockout (beKO) ani-mals were generated by crossing the BAC-transgenic Slco1c1-CreERT2 strain (Ridder et al., 2011), which expresses the tamoxifen-inducible CreERT2 re-combinase under control of the mouse Slco1c1 regulatory sequences in brain endothelial cells and epithelial cells of the choroid plexus, with mice carry-ing different loxP-flanked alleles (NemoFl; Schmidt-Supprian et al., 2000), p65GFPFl (De Lorenzi et al., 2009), Tak1Fl (Sato et al., 2005), Traf6Fl (Polykratis et al., 2012), or the Cre reporter line B6.Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J (Ai14; Madisen et al., 2010). Mice with a brain endothelial deletion of Tak1 and overexpression of constitutively active IKK2 were created by crossing Tak1beKO mice with R26-StopFLIkk2ca (Ikk2caFL) mice (Sasaki et al., 2006). Mice were homo- or hemizygous for all loxP-flanked alleles with the exception of a group depicted in Fig. 1 G. All transgenic animals used in this study were injected with 1 mg tamoxifen dissolved in 90% miglyol 812/10% ethanol i.p. every 12 h for 5 consecutive days. Littermates lacking the Cre transgene were used as controls in all experiments. TNF function was blocked by injecting Tak1beKO and Tak1Fl mice i.p. weekly with TNF-neutralizing (0.2 mg, clone CNTO5048) or control antibodies (0.2 mg, clone CNTO1322) starting 1 d before tamoxifen treatment until the end of the experiment (15 d after the first tamoxifen injection). For antioxidative diet, animals were fed with BHA-enriched (0.7%; Ssniff) or control food starting 2 wk before

the loss of cerebral capillaries compromise cerebral blood flow in NemobeKO mice.

In addition to endothelial cell death, NemobeKO mice showed signs of a disturbed BBB and vasogenic brain edema, very much like IP patients with acute neurological manifesta-tions (Avrahami et al., 1985; Chatkupt et al., 1993). Conceiv-ably, gaps in the endothelial lining of cerebral blood vessels after endothelial cell loss might increase BBB permeability. However, our data demonstrate that disruption of the BBB and endothelial cell death are at least partially independent (Fig. 10 G), suggesting that capillaries with gaps in the endo-thelial lining are shut down before a major leak occurs. To safeguard the BBB, TAK1 and NEMO trigger an IKK2- and p65 NF-B–mediated pathway that ultimately up-regulates the tight junction protein occludin by limiting its proteasomal degradation. Occludin is possibly stabilized due to induction of A20, a well-established target gene of NF-B that is able to deubiquitinate occludin (Kolodziej et al., 2011).

Previous studies reported that NF-B activation in brain endothelial cells is associated with increased permeability of the BBB, suggesting that endothelial NF-B may promote disrup-tion of the barrier (Alvarez et al., 2013). However, when we deleted components of NF-B signaling exclusively in brain endothelial cells, we found that NF-B actually protects the BBB, at least under the basal conditions that we have investi-gated. At present, we cannot exclude that under inflammatory conditions the situation might be different and high levels of endothelial NF-B activity could damage the BBB, in the sense of a bell-shaped relationship between NF-B activity and BBB integrity. However, this assumption is not necessarily re-quired, because NF-B–independent mechanisms that are instead mediated by ARNO, ARF6, -catenin, and FoxO1 increase endothelial permeability in response to inflamma-tory factors (Zhu et al., 2012; Beard et al., 2014). In addition, activation of NF-B in neighboring astrocytes and pericytes leads to the release of inflammatory mediators that subsequently break down the BBB (Zhang et al., 2007; Bell et al., 2012). An attractive hypothesis is that the NF-B–mediated endothelial pathway identified by us serves to reverse the increased perme-ability of the BBB in the presence of inflammation.

Occludin is an integral component of tight junctions but is apparently not required for their normal morphology (Saitou et al., 2000). It is the claudin family of molecules that is mainly responsible for the barrier. Nevertheless, brain calci-fications in occludin knockout mice hint to a BBB distur-bance (Keller et al., 2013) and the acute interference with occludin function leads to an overt disruption of barriers (Everett et al., 2006), supporting the notion that occludin plays a modifying role in the BBB. As tight junctions and a high occludin expression are characteristic of brain endothelial cells (Hirase et al., 1997), it is expected that the TAK1–NEMO–NF-B signaling pathway-stabilizing occludin expression in brain endothelial cells may be less important for the tightness of peripheral vessels. However, our study did not address the contribution of NEMO to the permeability of the vasculature in other organs and previous studies have reported divergent

1544 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

and injected (100 µl per mouse) i.v. 2 h later, anesthetized mice were perfused and brains were homogenized in 50 mM Tris-HCl, pH 8.0 (500 µl per hemi-sphere). After centrifugation (16.1 g for 30 min), equal volumes of methanol were added to supernatants. After another centrifugation (16.1 g for 15 min), fluorescence in supernatants was detected at 528 nm with an excitation wavelength of 485 nm using a microplate reader (Fluostar Optima).

EEG monitoring. For the recording of EEG, cortical screw electrodes were stereotactically implanted under anesthesia with ketamine (70 mg/kg i.p.) and xylazine (14 mg/kg i.p.) above the dorsal hippocampus or the fronto-parietal cortex (coordinates from bregma: AP 1.7; L +/1.6 or AP +3.0; L +/1.5, respectively) in 6 NemobeKO and 2 NemoFl mice per electrode loca-tion. Furthermore, a control group of 10 wild-type C57BL/6 mice received fronto-parietal cortex screws. To prevent postoperative infection, mice were treated with marbofloxacin (4 mg/kg, s.c., once daily) for 7 d starting 2 d before electrode implantation.

2 wk after surgery, mice were treated with tamoxifen. For continuous (24 h/d) EEG-monitoring, mice were connected to the system consisting of one-channel amplifiers (Animal BioAmp ML136; ADInstruments) and analogue-digital converters (PowerLab 8/30 ML870; ADInstruments) via a flexible cable. The data were recorded and analyzed with the LabChart 6 soft-ware (ADInstruments; sampling rate 200 Hz; time constant 0.1 s; low pass filter 60 Hz). The EEG recording was directly linked to simultaneous digital video-recording using one high-resolution infrared camera for up to eight mice (NYCTO Vision; CaS Business Services). For monitoring during the dark phase, infrared LEDs were mounted above the cages. The monitoring period started at day 8 after beginning of the tamoxifen treatment and lasted 7–18 d depending on the detection of epileptiform activity. In addition, mice were observed by two experimenters for 1 h per day for 5 d to detect myoclonic twitches. Wild-type mice were recorded once for one week.

Antibodies and chemicals. For Western blotting and immunofluorescent staining, the following antibodies and dilutions were used: anti–4-hydroxynon-enal (4-HNE) 1:1,000 (EMD Millipore); anti-actin 1:1,000 (Santa Cruz Bio-technology, Inc.); anti–active-caspase-3 1:400 (Cell Signaling Technology); anti-CD31 1:500 (BD or Santa Cruz Biotechnology, Inc.); anti-CD34 1:50 (Leica); anti-aquaporin 4 1:100 (EMD Millipore); anti-CD13 1:400 (AbD Se-rotec); anti-GFP 1:2,000 (Abcam); anti-GFAP 1:500 (Dako); anti-SMA 1:200 (Acris); anti–pan-laminin 1:1,000 (Wu et al., 2009); anti-laminin 5 1:200 (Sorokin et al., 1997); anti–claudin-5 1:200 (Invitrogen); anti-collagen IV 1:1,000 (Abcam); anti-IB 1:500 (Santa Cruz Biotechnology, Inc.); anti-NEMO 1:500 (Santa Cruz Biotechnology, Inc.); anti–NF-B p65 1:500 (Santa Cruz Biotechnology, Inc.); anti-occludin 1:500 (Abcam or Proteintech; 1:1,000); anti-TAK1 1:500 (Cell Signaling Technology); anti-TRAF6 1:500 (Santa Cruz Biotechnology, Inc.); anti-ZO1 1:250 (Invitrogen); anti–VCAM-1 1:1,000 (BD); anti–mouse-IgM-Cy3 1:400 (Jackson ImmunoResearch Lab-oratories); anti–mouse-IgG-HRP 1:5,000 (Santa Cruz Biotechnology, Inc.). Butylated hydroxyanisole (BHA), MG-132, sodium perborate, sodium fluor-escein, and tamoxifen were obtained from Sigma-Aldrich. 5Z-7 OZ was pur-chased from Tebu-Bio. GM-6001 was obtained from Merck. Lactacystin was purchased from Santa Cruz Biotechnology, Inc. The diet containing 0.7% BHA and the respective control diet were manufactured by Ssniff. Anti-murine TNF (clone CNTO5048) and isotype-matched control antibodies (clone CNTO1322) were provided by D. Shealy (Janssen R&D, Raritan, NJ). Mouse TNF was obtained from PeproTech.

Cell culture. PBECs of mice were cultured as described previously (Ridder et al., 2011). bEnd.3 cells (ATCC) were grown in DMEM containing FCS (10%), glucose (4.5 g/liter), penicillin (100 IU/ml), streptomycin (100 µg/ml), and l-glutamine (2 mM). Cell death was measured using the Cytotoxicity Detection kit (LDH; Roche) according to the manufacturer´s instructions.

Western blotting. If not indicated otherwise, the cerebellum was homog-enized in cell lysis buffer (Cell Signaling Technology) supplemented with

tamoxifen treatment. We did not exclude mice from analysis unless they died before the endpoint was measured. Mice were randomly allocated to diet or treatment groups. Investigators were blinded for treatment or geno-type of mice or both in all experiments and analyses.

Histological quantifications. We perfused mice with PBS containing heparin (10 IU/ml) to minimize vascular Ig staining. To quantify endoge-nous Ig extravasation, cryosections were stained with Cy3-labeled donkey anti-mouse IgM antibodies upon methanol fixation. The intensity of two to four pictures per mouse was determined with ImageJ software (National Institutes of Health) and values were expressed relative to control animals. The same procedure was used to determine GFAP in cortical tissue stained with anti-GFAP antibodies. String vessels were defined as empty basement membrane tubes and quantified by measuring the length of collagen IV–positive and CD31-negative vessels with ImageJ software. Vessel length was mea-sured in anti-CD31–stained sections. Two to four images were taken in the cortex if not stated otherwise, and analyzed for all parameters. To detect string vessels in human paraffin-embedded brain samples, we performed antigen-retrieval procedures before staining and used anti–collagen IV and anti-CD34, an endothelial cell marker. The Ethics Committee of the University of Lübeck was informed about the study.

For quantifying 4-HNE- or active caspase 3–positive vessels, cryosec-tions were stained with anti–4-HNE or anti-active caspase 3 and CD31. Measurements were performed as described for string vessels.

Stainings for CD31, collagen IV, pan-laminin, laminin 5, and SMA were quantified in a partially automated manner using a custom macro im-plemented into the image analysis software Fiji. In brief, images were thres-holded using an auto-threshold method, despeckled, and smoothened to remove staining artifacts. After conversion into a binary image, staining length was quantified with the Skeleton 2D/3D plugin yielding the number and length of stained structures in each image. To ensure reliable quantifica-tion, the macro was initially compared with a manually quantified dataset showing a significant correlation (r2 = 0.9157; unpublished data).

For the quantification of tight junction proteins (ZO-1, claudin-5, and occludin) on histological stainings, double-staining against the protein of in-terest and CD31 was performed. A selection of the vessel area was created by thresholding the CD31-image with an auto-threshold method. The thresh-olded image was used to create a mask, which then was applied onto the image showing the staining for the tight junction protein. The mean inten-sity was determined for this mask.

The staining for GFP in p65Fl and p65beKO mice was performed on 100-µm vibratome slices after short perfusion of the animals with 4% paraformalde-hyde without post-fixation. Sections from one experiment were always stained in parallel; images were generated at the same magnification, fluor-escence excitation intensity, and gain.

Biotin angiography. Mice were i.v. injected with sulfo-NHS-LC-biotin (EZ-Link; 5 mg/100 µl per mouse; Thermo Fisher Scientific). 60 min later, mice were perfused with PBS and killed. The biotin was labeled on cryosec-tions with Alexa Fluor 488–coupled Streptavidin (1:1,000; Invitrogen).

Determination of brain water content and BBB permeability. After dissection, brains were weighed immediately and dried for 4 d at 85°C. The percentage of water content was calculated as (wet weight dry weight) × 100/wet weight.

To investigate the permeability to sodium fluorescein, mice were in-jected i.v. with sodium fluorescein (6 mg/ml dissolved in PBS, 200 µl per 25 g body weight; Schoch et al., 2002). 30 min later, anesthetized mice were transcardially perfused with 10 ml of ice-cold Ringer´s solution. Forebrain and cerebellum were homogenized in 0.5 M sodium perborate dissolved in dH2O and centrifuged (800 g) for 15 min at 4°C. Ethanol (1.2 ml) was added to the supernatant and the samples were again centrifuged (16,100 g) for 20 s before measurement of fluorescence in the supernatant.

Dextrans of various sizes labeled with FITC (BD) were suspended in PBS (4 and 70 kD, 12 mg/ml; 150 and 2,000 kD, 24 mg/ml; Carman et al., 2011)

JEM Vol. 212, No. 10

Article

1545

Vector Biolabs) in medium with reduced (2%) serum concentration for 30 h, subsequently incubated in serum-free medium for 18 h, and stimulated with murine TNF (10 ng/ml) for 4 h. Cells were then washed once with PBS and harvested in glycylglycine buffer (25 mM glycylglycine, 15 mM MgSO4, and 4 mM EDTA) containing 1% Triton X-100 and 1 mM DTT. Luciferin so-lution (50 µl; 1 µM d-luciferin [Sigma-Aldrich] in glycylglycine buffer con-taining 10 µM DTT) and 92 µl assay buffer (16.3 mM KH2PO4, 1.09 mM DTT, 2.17 mM ATP dissolved in glycylglycine buffer) were added to 25 µl of cell lysate. Firefly luciferase activity was determined with a microplate reader (Fluostar Optima; BMG Labtech) and normalized to protein content measured by Advanced Protein Assay Reagent (Cytoskeleton).

Real-time RT-PCR. RNA was isolated from cerebellum or PBECs by using the ABI Prism 6100 Nucleic Acid PrepStation (Applied Bioscience) according to the manufacturer’s instructions. RNA (400 ng) was transcribed with Moloney Murine Leukemia Virus Reverse transcription and random hexamer primers (Promega). The following primers were used for quantita-tive RT-PCR: Ocln forward, 5-TACTGGTCTCTACGTGGATCAAT-3, Ocln reverse, 5-TTCTTCGGGTTTTCACAGCAA-3, PCR product 137 bp; Ppia forward, 5-AGGTCCTGGCATCTTGTCCAT-3, Ppia reverse, 5-GAACCGTTTGTGTTTGGTCCA-3, PCR product 51 bp; Tnf for-ward, 5-TGTAGCCCACGTCGTAGCAAA-3, Tnf reverse, 5-GCTG-GCACCACTAGTTGGTTGT-3, PCR product 120 bp; Tnfaip3 forward, 5-ACAGTGGACCTGAACTTCGC-3, Tnfaip3 reverse, 5-TGCACA-GGGATCTCCATCAC-3, PCR product 151 bp. Quantitative RT-PCR was performed according to the following protocol: 2 min at 50°C, 2 min at 95°C, 15 s at 95°C, and 1 min at 60°C (40 cycles). Amplification was quanti-fied using Platinum SYBR Green qPCR SuperMix (Invitrogen). Quantified results were normalized to Ppia using the Ct method.

Cerebrovascular reactivity. Cerebrovascular dilatory and contractile func-tion was tested in vitro in isolated, cannulated, and pressurized (60 mm Hg) segments of the PCA isolated from NemobeKO or NemoFl mice. We used on-line video microscopy to monitor diameter changes (Tong et al., 2005, 2012). Concentration-dependent dilations to ACh (Sigma-Aldrich) were tested on vessels preconstricted with phenylephrine (200 nM; Sigma-Aldrich). The contractile response to endothelin-1 (American Peptide) was tested on vessels at resting tone as was the tonic production of nitric oxide (NO) mea-sured by superfusion with the NO synthase inhibitor N-nitro-L-arginine (L-NNA; 10 µM, 40 min; Sigma-Aldrich). Because TRPV4 channels are involved in the ACh-induced dilation of cerebral arteries (Zhang et al., 2013), their function was tested in preconstricted vessels with the TRPV4 selective agonist GSK1016790A. The dilatory capacity of the vessels was verified with the NO donor sodium nitroprusside. Reversibility of the vas-cular deficits in response to ACh was tested before and after preincubation (30–60 min) of the arterial segments with the antioxidant butylated hydoxy-anisole (BHA, 100 µM).

Electron microscopy. For electron microscopy, Tak1Fl and Tak1beKO mice were transcardially perfused with 2.5% glutaraldehyde (Paesel and Lorei) in 0.1 M cacodylate buffer (pH 7.4). Brains were isolated and postfixed in 2.5% glutaraldehyde for 2–4 h, and then stored in cacodylate buffer. Brain cortices were cut into small pieces and washed once with 0.1 M cacodylate buffer. The specimens were postfixed in 1% OsO4 in cacodylate buffer for 1 h and dehydrated in ascending series of ethanol and propylene oxide. For contrast enhancement, they were bloc stained in uranyl acetate in 70% ethanol for 4 h and flat-embedded in Araldite (Serva). Using an ultramicrotome (Ultracut R; Leica), semithin (1 µm) and ultrathin (50 nm) sections were cut. Ultrathin sections were stained with lead citrate, mounted on copper grids, and finally analyzed with an EM 10 electron microscope (Carl Zeiss).

Laser doppler flowmetry (LDF). LDF measurements of cerebral blood flow were performed as described previously (Tong et al., 2009). Mice were anesthetized with ketamine (85 mg/kg, i.m.) and xylazine (3 mg/kg, i.m.) and placed in a stereotaxic frame on a heating blanket to maintain a stable

0.5 M PMSF directly before use. PBECs were cultured on 6-well plates and lysed in hot 2× Laemmli buffer, incubated at 95°C for 5 min, and then loaded on SDS-PAGE gels. Proteins were transferred to nitrocellulose mem-branes, which were then incubated with the indicated primary antibodies overnight at 4°C and, subsequently, with HRP-conjugated secondary anti-bodies for 1–2 h at room temperature. For detection, we applied enhanced chemiluminescence (SuperSignal West Femto Substrate; Thermo Fisher Scien-tific) and a digital detection system (GelDoc 2000; Bio-Rad Laboratories).

In vivo MRI acquisition. Mice were anesthetized with an induction dose of 4–5% sevoflurane and secured in an MRI-compatible bed. All MRI stud-ies were performed under 2.5–3% sevoflurane in medical air and animals were allowed to breathe spontaneously. Respiration rate and body tempera-ture was continuously monitored using an MRI-compatible system (Small Animal Instruments Inc.), and the temperature was maintained at 37 ± 0.2°C throughout the study using a feedback-regulated warming system (Small Animal Instruments Inc.).

All MR images were acquired on a 7 T Bruker Pharmascan system (Bruker Biospin) using a 28-mm inner-diameter, quadrature volume resona-tor (RAPID MR International) and using previously established methods (Grand’Maison et al., 2013; Hébert et al., 2013). In brief, after the acquisition of scout images, ROI-based shimming (MAPSHIM; Bruker Biospin) was performed to increase the magnetic field homogeneity within the brain. An-atomical images were acquired using a 3D balanced Steady-State Free Preces-sion (b-SSFP) sequence with matrix size = 128 × 128 × 64, field-of-view = 1.8 × 1.8 × 0.9 cm, spatial resolution = 140 × 140 ×140 µm, number of phase-cycles = 4, and number of averages = 4. The phase-cycled images were combined using the sum-of-squares reconstruction method to mini-mize banding artifacts (Bangerter et al., 2004). Perfusion images were ac-quired with a customized, 3D pseudocontinuous arterial spin labeling (ASL) sequence with matrix size = 64 × 64 × 32, field-of-view = 1.8 × 1.8 × 0.9 cm, spatial resolution = 280 × 280 × 280 µm, and 48 averages. Perfusion la-beling was achieved by positioning a 2-mm-thick inversion slab in the neck region (7 mm inferior to the level of the brainstem) and inverting inflow-ing blood within this slab every 14 ms using spatially selective sinc radiofre-quency pulses. The entire scanning session lasted 2.5 h per animal.

MRI processing. An unbiased, symmetric, customized template was gener-ated from anatomical scans from the 25 mice using an iterative process (Lau et al., 2008; Fonov et al., 2011). Before template generation, each recon-structed image volume underwent image nonuniformity correction using the N3 algorithm (Sled et al., 1998), brain masking, and linear spatial normaliza-tion using a 12-parameter affine transformation (Collins et al., 1994) to map individual images from native coordinate space to reference space. In brief, the template-generation process involved an iterative (coarse-to-fine resolu-tion) estimation of the nonlinear transformation to match each MRI scan to the evolving average of the population. The final anatomical template (popu-lation average) was generated with an isotropic voxel resolution of 0.06 mm. This customized template was parcellated into an atlas consisting of the whole cortex, the hippocampus, and the lateral ventricles using the Montreal Neu-rological Institute (MNI) McConnell Brain Imaging Centre DISPLAY soft-ware package. The mask of the cortex was projected onto a standardized cortical surface template.

The control perfusion images were linearly registered to the anatomical images for each mouse to compensate for potential, slight movement during the scanning session. Parametric perfusion maps based on the fractional ASL signal, defined as the ratio between the difference [control – labeled] and control images (Hébert et al., 2013), were calculated on a voxel-by-voxel basis. The perfusion maps were spatially normalized to reference space using the transformations derived from the anatomical image registration. The individual, spatially normalized perfusion maps were averaged to produce group-mean parametric maps. ROI-based perfusion measures were derived using the MRI atlas in reference space.

NF-B luciferase assay. After 5 d in vitro, PBECs were infected with an adenovirus carrying a NF-B luciferase reporter gene (Ad-NFKb-Luc;

1546 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

(10 cm diameter and 11 cm high) in which stranger mice may be present (see below). Two cameras mounted above the setup transferred images to a PC with ANY-mazeVideo Tracking System software (Stoelting). Testing com-prised two trials: acclimation trial and sociability trial. During acclimation, the test mouse was placed in the central compartment for 5 min. Divider doors were closed so that the mouse did not have access to the left or right chamber. During the sociability trial, a stranger mouse was placed into the wire cup in either the left or right chamber. Divider doors were opened and the test mouse could freely explore all three chambers. After a 10-min ob-servation period, the test mouse was gently guided toward the central com-partment and the divider doors were closed again. Preferential exploration of the stranger mouse over the empty cup was recorded and analyzed. Approaching and spending time within 3 cm of the cup were used to mea-sure exploratory behavior. Stranger mice were female C57BL/6J mice that were group-housed and that had served as stranger mice in other socia-bility experiments.

Anxiety-related exploration was measured in an elevated plus maze, which mice could freely explore for 10 min. The arena consisted of a plus-shaped maze with two arms (5 cm wide) closed by side walls and two arms without walls. Four IR beams recording open and closed arm entries and one recording the percentage of time per min spent in the open arms were connected to a computerized activity logger. Overall activity was expressed as total counts of entries.

For open field recordings, mice were placed in an illuminated square arena (47 × 47 cm, 160 lux) and video-recorded for 10 min at the indicated time points after tamoxifen injection. Mice were measured during their ac-tivity phase (dark phase). Inner zone was centered and defined as 36% of the overall area. The software ANY-maze was used to analyze the percentage of time spent in the inner zone representing a parameter related to anxiety.

Statistical analysis. Based on data from previous projects or from prelimi-nary experiments, we calculated the sample size for key experiments to ensure adequate power to detect prespecified effect sizes. All values are ex-pressed as means ± SEM. Differences were considered to be significant at P < 0.05. If distribution was not normal (Kolmogorov-Smirnov test in Prism [GraphPad]), we used either the nonparametric Mann-Whitney or Kruskal-Wallis test. In other cases, we used the two-sided Student’s t test or ANOVA as indicated in the figure legends. Post-hoc tests were only performed if the ANOVA or Kruskal-Wallis test showed a significant difference. For statisti-cal analysis of MRI data, measurements from the left and right hemispheres were combined for the ROI analysis to maximize statistical power.

We thank Cornelia Magnussen and Gudrun Vierke (Lübeck, Germany); Nadine Gehrig (Heidelberg, Germany); Leen Van Aerschot and Nele De Ruyck (Leuven, Germany); Claudia Brandt, Edith Kaczmarek, and Friederike Twele (Hannover, Germany); as well as Gabriele Frommer-Kästle (Tübingen, Germany), for expert technical assistance; Dr. Jochen Ohnmacht (Lübeck, Germany) for help with the biotin angiography, Dr. Clotilde Leblond-Lecrux (Montréal, Canada) for help with the LFP recordings; Dr. Shizuo Akira (Osaka, Japan) for providing Tak1Fl mice; Prof. Wolfgang Bäumgärtner and co-workers (Hannover, Germany) for post mortem examinations of the mice; and Dr. David Shealy (Janssen R&D, Spring House, Pennsylvania) for providing anti-TNF antibodies.

The research leading to these results received funding from the Deutsche Forschungsgemeinschaft to M. Schwaninger (SCHW 413/5-2), and to W. Löscher and K. Töllner (Lo 274/11-2); the internal funding scheme of the University of Lübeck to D.A. Ridder and J. Wenzel; and a joint program of the Canadian Stroke Network and the European Stroke Network (European Union’s Seventh Framework Program FP7/2007–2013 under grant agreements 201024 and 202213) to M. Schwaninger, E. Hamel, and L.M. Sorokin. D. Balschun and R. D’Hooge were supported by interdisciplinary research grants from KU Leuven (IDO/06/004 and GOA 12/008).

The authors declare no competing financial interests.

Submitted: 28 January 2015Accepted: 7 August 2015

body temperature (37°C). The skull over the barrel cortex was thinned to translucency for the placing of the LDF probe. Contralateral cerebral blood flow was recorded and averaged (4–6 stimulations every 30–40 s) before, during, and after unilateral whisker stimulation (20 s at 8–10 Hz). Stimulus-evoked blood flow was expressed as percent change relative to baseline. All experiments were completed in less than 20 min, during which time mouse blood pressure, blood gases, and pH were physiologically stable (Tong et al., 2012).

Local field potential recordings. Mice were anaesthetized with ketamine (70 mg/kg, i.p.) and xylazine (14 mg/kg, i.p.) and placed on a heating blan-ket for maintaining body temperature at 37°C. Two holes were drilled to place the electrodes, one over the cerebellum for the reference electrode and a second one over the barrel cortex for the recording electrode (Tungsten microelectrodes; FHC). The whisker pad was stimulated electrically by in-troducing two electrodes into the whisker pad. A stimulator (multichannel system) was used to create a stimulus with the following parameters: 25 con-secutive bipolar pulses (0.75 mA, 1 ms duration, 5 Hz). These parameters led to a reproducible blood flow response in the contralateral barrel cortex. Local field potentials were recorded with an amplifier (2 channels Amplifier Model 1800; AM-Systems) and converted into a digital format by an analogue-digital converter (PowerLab 8/30 ML870; ADInstruments). The software LabChart 7 was used to record and to analyze the data (ADInstruments; sampling rate 20 kHz; notch filter 50 Hz). For each mouse, four to six different areas in the barrel cortex were recorded to find the region with the highest response upon whisker stimulation. Each region was stimulated three times with the stimulation protocol described above. The amplitudes of 30 pulses of the region with the highest amplitude were averaged for each mouse.

Pupillary reactivity. We measured pupillary reactivity in awake mice. Eyes were video recorded with a camera (MV830; Canon), which is sensi-tive in the dark, starting with an illumination of 1 lux. After taking the base-line pupillary diameter for some seconds, a lamp was switched on to increase the illumination to 500 lux. Again, pupils were video-recorded for some seconds to determine minimal pupillary diameter. Diameter was measured manually, each mouse was tested twice and values were averaged.

Behavioral tests. The contextual fear conditioning experiment was per-formed as described previously (Van der Jeugd et al., 2011). The test cham-ber (26 × 22 × 18 cm high) was made of clear Plexiglas, and the grid floor was used to deliver an electric shock using a constant current shocker (MED Associates). The test chamber was placed inside a sound attenuated chamber. The experimental period comprised 3 d. On the first day, animals were placed in the testing chamber and were allowed to acclimate for 5 min. On the second day, animals were again placed in the testing chamber and after 2 min of exploration (baseline score), a buzzer was sounded for 30 s. This auditory stimulus, the conditional stimulus, was followed by a 2-s foot shock (0.3 mA), the unconditional stimulus. After the shock, mice were allowed to explore the chamber once more for 1 min before they received a second conditional–unconditional stimulus pairing. Finally, they were allowed to explore for another min. Twenty-four h later, on the third and last day, the animals were placed in the same context for 5 min exploration (context score). After 90 min, the mouse was again placed in the test chamber. Envi-ronmental and contextual cues were changed: a white paper square insert was placed in the chamber to alter its color, and mint extract was used to alter the smell. After 3 min of free exploration (preconditional stimulus score), the auditory stimulus was delivered for 3 min (conditional stimulus score). Freezing behavior was recorded every 10 s during each trial block using the standard interval sampling procedure.

Sociability was assessed as described previously (Naert et al., 2013). The setup was made of transparent plexiglass and consisted of a central chamber (42 × 26 cm) that gave access to a left and right chamber (26 × 26 cm) via sliding doors. The left and right chambers contained cylindrical wire cups

JEM Vol. 212, No. 10

Article

1547

Erickson, M.A., and W.A. Banks. 2013. Blood-brain barrier dysfunction as a cause and consequence of Alzheimer’s disease. J. Cereb. Blood Flow Metab. 33:1500–1513. http://dx.doi.org/10.1038/jcbfm.2013.135

Everett, R.S., M.K. Vanhook, N. Barozzi, I. Toth, and L.G. Johnson. 2006. Specific modulation of airway epithelial tight junctions by apical ap-plication of an occludin peptide. Mol. Pharmacol. 69:492–500. http://dx.doi.org/10.1124/mol.105.017251

Fonov, V., A.C. Evans, K. Botteron, C.R. Almli, R.C. McKinstry, and D.L. Collins. Brain Development Cooperative Group. 2011. Unbiasedaverage age-appropriate atlases for pediatric studies. Neuroimage. 54:313–327. http://dx.doi.org/10.1016/j.neuroimage.2010.07.033

Gareus, R., E. Kotsaki, S. Xanthoulea, I. van der Made, M.J.J. Gijbels, R. Kardakaris, A. Polykratis, G. Kollias, M.P.J. de Winther, and M. Pasparakis. 2008. Endothelial cell-specific NF-kappaB inhibition protects mice from atherosclerosis. Cell Metab. 8:372–383. http://dx.doi.org/10.1016/ j.cmet.2008.08.016

Girouard, H., and C. Iadecola. 2006. Neurovascular coupling in the normal brain and in hypertension, stroke, and Alzheimer disease. J. Appl. Physiol. 100:328–335. http://dx.doi.org/10.1152/japplphysiol.00966.2005

Goldberg, M.F.M.D. 2004. The skin is not the predominant problem in incontinentia pigmenti. Arch. Dermatol. 140:748–750. http://dx.doi.org/ 10.1001/archderm.140.6.748

Goldmann, T., P. Wieghofer, P.F. Müller, Y. Wolf, D. Varol, S. Yona, S.M. Brendecke, K. Kierdorf, O. Staszewski, M. Datta, et al. 2013.A new type of microglia gene targeting shows TAK1 to be pivotal inCNS autoimmune inflammation. Nat. Neurosci. 16:1618–1626. http://dx.doi.org/10.1038/nn.3531

Grand’maison, M., S.P. Zehntner, M.K. Ho, F. Hébert, A. Wood, F. Carbonell, A.P. Zijdenbos, E. Hamel, and B.J. Bedell. 2013. Early cortical thickness changes predict -amyloid deposition in a mouse model of Alzheimer’s disease. Neurobiol. Dis. 54:59–67. http://dx.doi.org/10.1016/j.nbd.2013 .02.005

Gregor, M.F., and G.S. Hotamisligil. 2011. Inflammatory mechanisms in obesity. Annu. Rev. Immunol. 29:415–445. http://dx.doi.org/10.1146/annurev- immunol-031210-101322

Hébert, F., M. Grand’maison, M.K. Ho, J.P. Lerch, E. Hamel, and B.J. Bedell. 2013. Cortical atrophy and hypoperfusion in a transgenic mouse model of Alzheimer’s disease. Neurobiol. Aging. 34:1644–1652. http://dx.doi.org/10.1016/j.neurobiolaging.2012.11.022

Hennel, S.J., P.G. Ekert, J.J. Volpe, and T.E. Inder. 2003. Insights into the pathogenesis of cerebral lesions in incontinentia pigmenti. Pediatr. Neurol. 29:148–150. http://dx.doi.org/10.1016/S0887-8994(03)00150-4

Hirase, T., J.M. Staddon, M. Saitou, Y. Ando-Akatsuka, M. Itoh, M. Furuse, K. Fujimoto, S. Tsukita, and L.L. Rubin. 1997. Occludin as a possibledeterminant of tight junction permeability in endothelial cells. J. Cell Sci.110:1603–1613.

Ivens, S., D. Kaufer, L.P. Flores, I. Bechmann, D. Zumsteg, O. Tomkins, E. Seiffert, U. Heinemann, and A. Friedman. 2007. TGF-beta receptor- mediated albumin uptake into astrocytes is involved in neocortical ep-ileptogenesis. Brain. 130:535–547. http://dx.doi.org/10.1093/brain/ awl317

Jacob, A., B. Hack, P. Chen, R.J. Quigg, and J.J. Alexander. 2011. C5a/CD88 signaling alters blood-brain barrier integrity in lupus through nuclear factor-B. J. Neurochem. 119:1041–1051. http://dx.doi.org/10.1111/ j.1471-4159.2011.07490.x

Keller, A., A. Westenberger, M.J. Sobrido, M. García-Murias, A. Domingo, R.L. Sears, R.R. Lemos, A. Ordoñez-Ugalde, G. Nicolas, J.E. daCunha, et al. 2013. Mutations in the gene encoding PDGF-B causebrain calcifications in humans and mice. Nat. Genet. 45:1077–1082.http://dx.doi.org/10.1038/ng.2723

Kielland, A., L.M. Camassa, G. Døhlen, L.A. Munthe, R. Blomhoff, M. Amiry-Moghaddam, and H. Carlsen. 2012. NF-B activity in perina-tal brain during infectious and hypoxic-ischemic insults revealed by a reporter mouse. Brain Pathol. 22:499–510. http://dx.doi.org/10.1111/j.1750-3639.2011.00548.x

Kisseleva, T., L. Song, M. Vorontchikhina, N. Feirt, J. Kitajewski, and C. Schindler. 2006. NF-kappaB regulation of endothelial cell function dur-ing LPS-induced toxemia and cancer. J. Clin. Invest. 116:2955–2963. http://dx.doi.org/10.1172/JCI27392

REFERENCESAbe, S., A. Okumura, S. Hamano, M. Tanaka, T. Shiihara, K. Aizaki, T.

Tsuru, Y. Toribe, H. Arai, and T. Shimizu. 2011. Early infantile manifes-tations of incontinentia pigmenti mimicking acute encephalopathy. Brain Dev. 33:28–34. http://dx.doi.org/10.1016/j.braindev.2010.04.002

Alvarez, J.I., T. Katayama, and A. Prat. 2013. Glial influence on the blood brain barrier. Glia. 61:1939–1958. http://dx.doi.org/10.1002/glia .22575

Argaw, A.T., Y. Zhang, B.J. Snyder, M.L. Zhao, N. Kopp, S.C. Lee, C.S. Raine, C.F. Brosnan, and G.R. John. 2006. IL-1beta regulates blood-brain bar-rier permeability via reactivation of the hypoxia-angiogenesis program. J. Immunol. 177:5574–5584. http://dx.doi.org/10.4049/jimmunol.177.8.5574

Avrahami, E., S. Harel, U. Jurgenson, and D.F. Cohn. 1985. Computed to-mographic demonstration of brain changes in incontinentia pigmenti. Am. J. Dis. Child. 139:372–374.

Bachevalier, F., C. Marchal, M.P. Di Cesare, A. Antunes, and F. Truchetet. 2003. Atteinte neurologique létale au cours d’une incontinentia pigmenti. Ann. Dermatol. Venereol. 130:1139–1142.

Bangerter, N.K., B.A. Hargreaves, S.S. Vasanawala, J.M. Pauly, G.E. Gold, and D.G. Nishimura. 2004. Analysis of multiple-acquisition SSFP. Magn. Reson. Med. 51:1038–1047. http://dx.doi.org/10.1002/mrm .20052

Beard, R.S. Jr., R.J. Haines, K.Y. Wu, J.J. Reynolds, S.M. Davis, J.E. Elliott, N.L. Malinin, V. Chatterjee, B.J. Cha, M.H. Wu, and S.Y. Yuan. 2014. Non-muscle Mlck is required for -catenin- and FoxO1-dependent down-regulation of Cldn5 in IL-1-mediated barrier dysfunction in brain endothelial cells. J. Cell Sci. 127:1840–1853. http://dx.doi.org/10.1242/ jcs.144550

Bell, R.D., E.A. Winkler, I. Singh, A.P. Sagare, R. Deane, Z. Wu, D.M. Holtzman, C. Betsholtz, A. Armulik, J. Sallstrom, et al. 2012. Apolipoprotein E controlscerebrovascular integrity via cyclophilin A. Nature. 485:512–516. http://dx.doi.org/10.1038/nature11087

Boulanger, L.M. 2009. Immune proteins in brain development and synaptic plasticity. Neuron. 64:93–109. http://dx.doi.org/10.1016/j.neuron.2009 .09.001

Brown, W.R. 2010. A review of string vessels or collapsed, empty basement membrane tubes. J. Alzheimers Dis. 21:725–739. http://dx.doi.org/ 10.3233/JAD-2010-100219

Carman, A.J., J.H. Mills, A. Krenz, D.G. Kim, and M.S. Bynoe. 2011. Adenosine receptor signaling modulates permeability of the blood-brain barrier. J. Neurosci. 31:13272–13280. http://dx.doi.org/10.1523/ JNEUROSCI.3337-11.2011

Catrysse, L., L. Vereecke, R. Beyaert, and G. van Loo. 2014. A20 in inflam-mation and autoimmunity. Trends Immunol. 35:22–31. http://dx.doi.org/ 10.1016/j.it.2013.10.005

Chatkupt, S., A.O. Gozo, L.J. Wolansky, and S. Sun. 1993. Characteristic MR findings in a neonate with incontinentia pigmenti. AJR Am. J. Roentgenol. 160:372–374. http://dx.doi.org/10.2214/ajr.160.2.8424354

Chen, Y.H., M. Chang, and B.L. Davidson. 2009. Molecular signatures of dis-ease brain endothelia provide new sites for CNS-directed enzyme therapy. Nat. Med. 15:1215–1218. http://dx.doi.org/10.1038/nm.2025

Clark, K., S. Nanda, and P. Cohen. 2013. Molecular control of the NEMO family of ubiquitin-binding proteins. Nat. Rev. Mol. Cell Biol. 14:673–685. http://dx.doi.org/10.1038/nrm3644

Collins, D.L., P. Neelin, T.M. Peters, and A.C. Evans. 1994. Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space. J. Comput. Assist. Tomogr. 18:192–205. http://dx.doi.org/10.1097/ 00004728-199403000-00005

Cummins, P.M. 2012. Occludin: one protein, many forms. Mol. Cell. Biol. 32:242–250. http://dx.doi.org/10.1128/MCB.06029-11

De Lorenzi, R., R. Gareus, S. Fengler, and M. Pasparakis. 2009. GFP-p65 knock-in mice as a tool to study NF-kappaB dynamics in vivo. Genesis. 47:323–329. http://dx.doi.org/10.1002/dvg.20468

Dondelinger, Y., M.A. Aguileta, V. Goossens, C. Dubuisson, S. Grootjans, E. Dejardin, P. Vandenabeele, and M.J. Bertrand. 2013. RIPK3 contributes to TNFR1-mediated RIPK1 kinase-dependent apoptosis in conditions of cIAP1/2 depletion or TAK1 kinase inhibition. Cell Death Differ. 20:1381–1392. http://dx.doi.org/10.1038/cdd.2013.94

1548 TAK1 and NEMO protect the neurovascular unit | Ridder et al.

Roy, A., and D.C. Hooper. 2007. Lethal silver-haired bat rabies virus in-fection can be prevented by opening the blood-brain barrier. J. Virol. 81:7993–7998. http://dx.doi.org/10.1128/JVI.00710-07

Saitou, M., M. Furuse, H. Sasaki, J.D. Schulzke, M. Fromm, H. Takano, T. Noda, and S. Tsukita. 2000. Complex phenotype of mice lacking occludin, a component of tight junction strands. Mol. Biol. Cell. 11:4131–4142. http://dx.doi.org/10.1091/mbc.11.12.4131

Sakurai, H. 2012. Targeting of TAK1 in inflammatory disorders and can-cer. Trends Pharmacol. Sci. 33:522–530. http://dx.doi.org/10.1016/j .tips.2012.06.007

Sasaki, Y., E. Derudder, E. Hobeika, R. Pelanda, M. Reth, K. Rajewsky, and M. Schmidt-Supprian. 2006. Canonical NF-kappaB activity, dispensablefor B cell development, replaces BAFF-receptor signals and promotesB cell proliferation upon activation. Immunity. 24:729–739. http://dx.doi.org/10.1016/j.immuni.2006.04.005

Sato, S., H. Sanjo, K. Takeda, J. Ninomiya-Tsuji, M. Yamamoto, T. Kawai, K. Matsumoto, O. Takeuchi, and S. Akira. 2005. Essential functionfor the kinase TAK1 in innate and adaptive immune responses. Nat.Immunol. 6:1087–1095. http://dx.doi.org/10.1038/ni1255

Schmidt-Supprian, M., W. Bloch, G. Courtois, K. Addicks, A. Israël, K. Rajewsky, and M. Pasparakis. 2000. NEMO/IKK gamma-deficient mice model incontinentia pigmenti. Mol. Cell. 5:981–992. http://dx.doi.org/ 10.1016/S1097-2765(00)80263-4

Schoch, H.J., S. Fischer, and H.H. Marti. 2002. Hypoxia-induced vascular endothelial growth factor expression causes vascular leakage in the brain. Brain. 125:2549–2557. http://dx.doi.org/10.1093/brain/awf257

Schwenk, F., U. Baron, and K. Rajewsky. 1995. A cre-transgenic mouse strain for the ubiquitous deletion of loxP-flanked gene segments in-cluding deletion in germ cells. Nucleic Acids Res. 23:5080–5081. http://dx.doi.org/10.1093/nar/23.24.5080

Seiffert, E., J.P. Dreier, S. Ivens, I. Bechmann, O. Tomkins, U. Heinemann, and A. Friedman. 2004. Lasting blood-brain barrier disruption induces epileptic focus in the rat somatosensory cortex. J. Neurosci. 24:7829–7836. http://dx.doi.org/10.1523/JNEUROSCI.1751-04.2004

Sled, J.G., A.P. Zijdenbos, and A.C. Evans. 1998. A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans. Med. Imaging. 17:87–97. http://dx.doi.org/10.1109/42.668698

Sonkusare, S.K., A.D. Bonev, J. Ledoux, W. Liedtke, M.I. Kotlikoff, T.J. Heppner, D.C. Hill-Eubanks, and M.T. Nelson. 2012. Elementary Ca2+ signals through endothelial TRPV4 channels regulate vascular function. Science. 336:597–601. http://dx.doi.org/10.1126/science.1216283

Sorokin, L.M., F. Pausch, M. Frieser, S. Kröger, E. Ohage, and R. Deutzmann. 1997. Developmental regulation of the laminin alpha5 chain suggests a role in epithelial and endothelial cell maturation. Dev. Biol. 189:285–300. http://dx.doi.org/10.1006/dbio.1997.8668

Taccone, F.S., F. Su, C. Pierrakos, X. He, S. James, O. Dewitte, J.L. Vincent, and D. De Backer. 2010. Cerebral microcirculation is impaired during sepsis:an experimental study. Crit. Care. 14:R140. http://dx.doi.org/10.1186/cc9205

Tong, X.K., N. Nicolakakis, A. Kocharyan, and E. Hamel. 2005. Vascular remod-eling versus amyloid beta-induced oxidative stress in the cerebrovascular dysfunctions associated with Alzheimer’s disease. J. Neurosci. 25:11165–11174. http://dx.doi.org/10.1523/JNEUROSCI.4031-05.2005

Tong, X.K., N. Nicolakakis, P. Fernandes, B. Ongali, J. Brouillette, R. Quirion, and E. Hamel. 2009. Simvastatin improves cerebrovascular function and counters soluble amyloid-beta, inflammation and oxidative stress in aged APP mice. Neurobiol. Dis. 35:406–414. http://dx.doi.org/10 .1016/j.nbd.2009.06.003

Tong, X.K., C. Lecrux, P. Rosa-Neto, and E. Hamel. 2012. Age-dependent rescue by simvastatin of Alzheimer’s disease cerebrovascular and mem-ory deficits. J. Neurosci. 32:4705–4715. http://dx.doi.org/10.1523/ JNEUROSCI.0169-12.2012

Tripathi, A.K., W. Sha, V. Shulaev, M.F. Stins, and D.J. Sullivan Jr. 2009. Plasmodium falciparum-infected erythrocytes induce NF-kappaB regu-lated inflammatory pathways in human cerebral endothelium. Blood. 114:4243–4252. http://dx.doi.org/10.1182/blood-2009-06-226415

Tronche, F., C. Kellendonk, O. Kretz, P. Gass, K. Anlag, P.C. Orban, R. Bock, R. Klein, and G. Schütz. 1999. Disruption of the glucocorticoid

Kolodziej, L.E., J.P. Lodolce, J.E. Chang, J.R. Schneider, W.A. Grimm, S.J. Bartulis, X. Zhu, J.S. Messer, S.F. Murphy, N. Reddy, et al. 2011. TNFAIP3 maintains intestinal barrier function and supports epithelial cell tight junctions. PLoS ONE. 6:e26352. http://dx.doi.org/10.1371/journal.pone.0026352

Lau, J.C., J.P. Lerch, J.G. Sled, R.M. Henkelman, A.C. Evans, and B.J. Bedell. 2008. Longitudinal neuroanatomical changes determined by deformation-based morphometry in a mouse model of Alzheimer’s disease. Neuroimage. 42:19–27. http://dx.doi.org/10.1016/j.neuroimage.2008.04.252

Lomaga, M.A., W.-C. Yeh, I. Sarosi, G.S. Duncan, C. Furlonger, A. Ho, S. Morony, C. Capparelli, G. Van, S. Kaufman, et al. 1999. TRAF6deficiency results in osteopetrosis and defective interleukin-1, CD40,and LPS signaling. Genes Dev. 13:1015–1024. http://dx.doi.org/10.1101/gad.13.8.1015

Madisen, L., T.A. Zwingman, S.M. Sunkin, S.W. Oh, H.A. Zariwala, H. Gu, L.L. Ng, R.D. Palmiter, M.J. Hawrylycz, A.R. Jones, et al. 2010. A ro-bust and high-throughput Cre reporting and characterization system forthe whole mouse brain. Nat. Neurosci. 13:133–140. http://dx.doi.org/10.1038/nn.2467

Maingay-de Groof, F., M.H. Lequin, D.W. Roofthooft, A.P. Oranje, I.F. de Coo, L.A. Bok, P.J. van der Spek, G.M. Mancini, and P.P. Govaert. 2008. Extensive cerebral infarction in the newborn due to incontinen-tia pigmenti. Eur. J. Paediatr. Neurol. 12:284–289. http://dx.doi.org/10 .1016/j.ejpn.2007.09.001

Makris, C., V.L. Godfrey, G. Krähn-Senftleben, T. Takahashi, J.L. Roberts, T. Schwarz, L. Feng, R.S. Johnson, and M. Karin. 2000. Female miceheterozygous for IKK gamma/NEMO deficiencies develop a dermatopa-thy similar to the human X-linked disorder incontinentia pigmenti. Mol. Cell. 5:969–979. http://dx.doi.org/10.1016/S1097-2765(00)80262-2

Matullo, C.M., K.J. O’Regan, H. Hensley, M. Curtis, and G.F. Rall. 2010. Lymphocytic choriomeningitis virus-induced mortality in mice is trig-gered by edema and brain herniation. J. Virol. 84:312–320. http://dx .doi.org/10.1128/JVI.00727-09

Meuwissen, M.E., and G.M. Mancini. 2012. Neurological findings in in-continentia pigmenti; a review. Eur. J. Med. Genet. 55:323–331. http://dx.doi.org/10.1016/j.ejmg.2012.04.007

Morioka, S., M. Inagaki, Y. Komatsu, Y. Mishina, K. Matsumoto, and J. Ninomiya-Tsuji. 2012. TAK1 kinase signaling regulates embryonic an-giogenesis by modulating endothelial cell survival and migration. Blood. 120:3846–3857. http://dx.doi.org/10.1182/blood-2012-03-416198

Naert, A., I. Gantois, A. Laeremans, S. Vreysen, G. Van den Bergh, L. Arckens, Z. Callaerts-Vegh, and R. D’Hooge. 2013. Behavioural alterations relevant to developmental brain disorders in mice with neonatally induced ven-tral hippocampal lesions. Brain Res. Bull. 94:71–81. http://dx.doi.org/ 10.1016/j.brainresbull.2013.01.008

O’Donnell, M.A., H. Hase, D. Legarda, and A.T. Ting. 2012. NEMO inhibits programmed necrosis in an NFB-independent manner by restraining RIP1. PLoS ONE. 7:e41238. http://dx.doi.org/10.1371/journal.pone.0041238

Obermeier, B., R. Daneman, and R.M. Ransohoff. 2013. Development, maintenance and disruption of the blood-brain barrier. Nat. Med. 19: 1584–1596. http://dx.doi.org/10.1038/nm.3407

Østergaard, L., S.N. Jespersen, K. Mouridsen, I.K. Mikkelsen, K.Y. Jonsdottír, A. Tietze, J.U. Blicher, R. Aamand, N. Hjort, N.K. Iversen, et al. 2013. The role of the cerebral capillaries in acute ischemic stroke: the extended penumbra model. J. Cereb. Blood Flow Metab. 33:635–648. http://dx.doi .org/10.1038/jcbfm.2013.18

Polykratis, A., G. van Loo, S. Xanthoulea, M. Hellmich, and M. Pasparakis. 2012. Conditional targeting of tumor necrosis factor receptor-associated factor 6 reveals opposing functions of Toll-like receptor signaling in endothelial and myeloid cells in a mouse model of atherosclerosis. Circulation. 126:1739–1751. http://dx.doi.org/10.1161/CIRCULATIONAHA.112.100339

Ridder, D.A., M.F. Lang, S. Salinin, J.P. Röderer, M. Struss, C. Maser-Gluth, and M. Schwaninger. 2011. TAK1 in brain endothelial cells medi-ates fever and lethargy. J. Exp. Med. 208:2615–2623. http://dx.doi.org/ 10.1084/jem.20110398

JEM Vol. 212, No. 10

Article

1549

receptor gene in the nervous system results in reduced anxiety. Nat. Genet. 23:99–103. http://dx.doi.org/10.1038/12703

Tsao, N., H.P. Hsu, C.M. Wu, C.C. Liu, and H.Y. Lei. 2001. Tumour ne-crosis factor-alpha causes an increase in blood-brain barrier permeability during sepsis. J. Med. Microbiol. 50:812–821.

Van der Jeugd, A., T. Ahmed, S. Burnouf, K. Belarbi, M. Hamdame, M.-E. Grosjean, S. Humez, D. Balschun, D. Blum, L. Buée, and R. D’Hooge. 2011. Hippocampal tauopathy in tau transgenic mice coincides with impaired hippocampus-dependent learning and memory, and attenuated late-phase long-term depression of synaptic transmission. Neurobiol. Learn. Mem. 95:296–304. http://dx.doi.org/10.1016/j.nlm.2010.12.005

van Loo, G., R. De Lorenzi, H. Schmidt, M. Huth, A. Mildner, M. Schmidt-Supprian, H. Lassmann, M.R. Prinz, and M. Pasparakis. 2006. Inhibition of transcription factor NF-kappaB in the central nervous sys-tem ameliorates autoimmune encephalomyelitis in mice. Nat. Immunol. 7:954–961. http://dx.doi.org/10.1038/ni1372

Vanlangenakker, N., T. Vanden Berghe, P. Bogaert, B. Laukens, K. Zobel, K. Deshayes, D. Vucic, S. Fulda, P. Vandenabeele, and M.J. Bertrand.2011. cIAP1 and TAK1 protect cells from TNF-induced necrosis by pre-venting RIP1/RIP3-dependent reactive oxygen species production. CellDeath Differ. 18:656–665. http://dx.doi.org/10.1038/cdd.2010.138

Wu, C., F. Ivars, P. Anderson, R. Hallmann, D. Vestweber, P. Nilsson, H. Robenek, K. Tryggvason, J. Song, E. Korpos, et al. 2009. Endothelial basement mem-brane laminin alpha5 selectively inhibits T lymphocyte extravasation intothe brain. Nat. Med. 15:519–527. http://dx.doi.org/10.1038/nm.1957

Xiao, Y., H. Li, J. Zhang, A. Volk, S. Zhang, W. Wei, S. Zhang, P. Breslin, and J. Zhang. 2011. TNF-/Fas-RIP-1-induced cell death signaling

separates murine hematopoietic stem cells/progenitors into 2 distinct populations. Blood. 118:6057–6067. http://dx.doi.org/10.1182/blood- 2011-06-359448

Ye, X., J. Ding, X. Zhou, G. Chen, and S.F. Liu. 2008. Divergent roles of endothelial NF-kappaB in multiple organ injury and bacterial clearance in mouse models of sepsis. J. Exp. Med. 205:1303–1315. http://dx.doi .org/10.1084/jem.20071393

Zhang, X., J.A. Winkles, M.C. Gongora, R. Polavarapu, J.S. Michaelson, K. Hahm, L. Burkly, M. Friedman, X.J. Li, and M. Yepes. 2007.TWEAK-Fn14 pathway inhibition protects the integrity of the neu-rovascular unit during cerebral ischemia. J. Cereb. Blood Flow Metab.27:534–544. http://dx.doi.org/10.1038/sj.jcbfm.9600368

Zhang, L., P. Papadopoulos, and E. Hamel. 2013. Endothelial TRPV4 channels mediate dilation of cerebral arteries: impairment and recov-ery in cerebrovascular pathologies related to Alzheimer’s disease. Br. J. Pharmacol. 170:661–670. http://dx.doi.org/10.1111/bph.12315

Zhu, W., N.R. London, C.C. Gibson, C.T. Davis, Z. Tong, L.K. Sorensen, D.S. Shi, J. Guo, M.C. Smith, A.H. Grossmann, et al. 2012. Interleukinreceptor activates a MYD88-ARNO-ARF6 cascade to disrupt vas-cular stability. Nature. 492:252–255. http://dx.doi.org/10.1038/nature11603

Zlokovic, B.V. 2008. The blood-brain barrier in health and chronic neuro-degenerative disorders. Neuron. 57:178–201. http://dx.doi.org/10.1016/ j.neuron.2008.01.003

Zlokovic, B.V. 2011. Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders. Nat. Rev. Neurosci. 12:723–738. http://dx.doi.org/10.1038/nrn3114

INSIGHTS

466 INSIGHTS | The Journal of Experimental Medicine

<ID>JEM_2134insight1_Fig1.jpeg</ID>Unraveling the cellular changes elicited by pathogenic mutations is central to the development of novel targeted therapeutics. In this issue, <addart type=”rel” doi=”10.1084/jem.20150951”>Patzke et al.</addart> use advanced stem cell technologies to underscore the role of human L1CAM for neuronal development and function.

Mutations in the cell adhesion molecule L1CAM cause neurological alterations, including severe intel-lectual disabilities summarized as L1 syndrome. The requirements of L1CAM for neuronal development have been characterized extensively using in vitro, cell culture, and mouse model systems, which have im-plicated L1CAM in numerous signaling pathways with essential roles for neurite outgrowth, axon guidance, and synaptic transmission. However, distinct differences remain between the phenotypic strength of mouse models and clinical presentations in human patients.

<ID>JEM_2134insight1_Fig2.eps</ID>Patzke et al. address the requirements of human L1CAM using embryonic stem (ES) cell–derived neurons and demonstrate that L1CAM is cell-autonomously required for the control of neuronal morphology and excitability, features likely contributing to impaired neurological performances in human patients. The novel approach of generating induced con-ditional knockout neurons with perfectly matched controls enables precise qualitative and quantitative analyses of disease-associated alterations. Loss of L1CAM not only resulted in reduced axonal outgrowth but also uncovered a requirement for dendritic

development. In the absence of L1CAM, there is a reduction in the levels of ankyrinG and ankyrinB, two adaptor proteins that are essential organizers of axonal domains. Consequently, mutant neurons ex-hibited decreased excitability, resulting in impaired action potential generation. Pathogenic mutations in the ankyrin-binding site of L1CAM failed to re-store ankyrin distribution, suggesting that L1CAM and ankyrins form interdependent stabilizing com-plexes controlling the subcellular protein composition of axonal domains.

From a therapeutic point of view, this strategy of matched control and mutant human ES-derived neurons not only enables the identification of human-specific gene functions but also facilitates the testing of potential gene-based or pharmacological interven-tions to ameliorate disease-associated deficits. However, in the case of adhesion molecules like L1CAM, it is also important to consider that transcellular interac-tions may alter intrinsic signaling properties and con-tribute to the establishment of functional neuronal

circuits. In combination with in vivo animal models, the ES-based conditional knockout strategy provides an exciting and complementary route to human disease diagnostics and therapeutic development.

Patzke, C., et al. 2016. J. Exp. Med. http://dx.doi.org/10.1084/jem.20150951

<doi>10.1084/jem.2134insight1</doi><aid>jem.2134insight1</aid><au> Jan Pielage </au><AF1> Friedrich Miescher Institute for Biomedical Research </AF1><cor> jan.pielage@fmi.ch </cor><dochead>Insights</dochead><doctopic>News</doctopic>Induced knockouts provide insights into human L1 syndrome

Insight from

Jan Pielage

The cell adhesion molecule L1CAM is cell-autonomously required for the

development of neuronal compartments.

Jan Pielage, Friedrich Miescher Institute for Biomedical Research: jan.pielage@fmi.ch

Article

The Rockefeller University Press $30.00J. Exp. Med. 2016 Vol. 213 No. 4 499–515www.jem.org/cgi/doi/10.1084/jem.20150951

499

X-linked neurodevelopmental disorders that produce intel-lectual disability are relatively common diseases resulting frommutations in X-chromosomal genes, with ∼1/600–1/1,000males affected (Gécz et al., 2009). One particular gene associ-ated with X-linked intellectual disability is L1CAM, with an estimated incidence of 1/25,000–1/60,000 males (Halliday et al., 1986; Weller and Gärtner, 2001). L1CAM is a type 1 trans-membrane protein of the immunoglobulin superfamily that is conserved in vertebrates and invertebrates (Neuroglian/Sax-7) and that was initially described as a neuronal cell adhesionmolecule involved in axonal growth (Rathjen and Schachner, 1984; Chang et al., 1987). L1CAM contains six extracellularN-terminal Ig domains, followed by five fibronectin type 3(FN3) domains, a single transmembrane segment, and a shortC-terminal intracellular sequence. L1CAM is the foundingmember of the L1CAM family that includes neurofascin,NrCAM, and CHL1, which exhibit similar domain struc-tures and are composed of homologous sequences. L1CAMis highly expressed in the developing nervous system and hasalso been implicated in cancer progression (Schäfer and Al-tevogt, 2010; Kiefel et al., 2012; Schäfer and Frotscher, 2012).

Approximately 350 pathogenic mutations in the L1CAM gene have been described in patients with a broad

spectrum of neurological abnormalities and mental retarda-tion, summarized by the term L1 syndrome. This spectrum includes the MASA syndrome (mental retardation, aphasia, shuffling gait, and adducted thumbs), hydrocephalus due to stenosis of the aqueduct of Sylvius, agenesis of the corpus callosum, and SPG1 (X-linked hereditary spastic paraplegia type 1), which are referred to collectively as CRA SH syn-drome (Rosenthal et al., 1992; Stumpel and Vos, 1993; Jouet et al., 1994, 1995; Fransen et al., 1997; Weller and Gärtner, 2001; Vos et al., 2010). Besides a reported whole gene dele-tion (Chidsey et al., 2014), these mutations include frameshift, nonsense, and missense mutations, resulting in the production of truncated proteins or proteins with mutations in structur-ally defined key residues (Stumpel and Vos, 1993). Missense mutations most likely lead to alterations of intracellular traf-ficking and impaired function and mobility caused by addi-tional cysteines on the surface of the molecule or aberrant ligand binding (De Angelis et al., 1999, 2002; Kenwrick et al., 2000; Schäfer et al., 2010). Pathological mutations are known to affect binding of L1CAM to itself, Neuropilin-1, Tax-1/Axonin-1, ankyrins, and integrins, or to impair triggering of epidermal growth factor receptor and Erk1/2 signaling (De Angelis et al., 1999; Schäfer and Altevogt, 2010). Overall, most

Hundreds of L1CAM gene mutations have been shown to be associated with congenital hydrocephalus, severe intellectual disability, aphasia, and motor symptoms. How such mutations impair neuronal function, however, remains unclear. Here, we generated human embryonic stem (ES) cells carrying a conditional L1CAM loss-of-function mutation and produced precisely matching control and L1CAM-deficient neurons from these ES cells. In analyzing two independent conditionally mutant ES cell clones, we found that deletion of L1CAM dramatically impaired axonal elongation and, to a lesser extent, dendritic arboriza-tion. Unexpectedly, we also detected an ∼20–50% and ∼20–30% decrease, respectively, in the levels of ankyrinG and ankyrinB protein, and observed that the size and intensity of ankyrinG staining in the axon initial segment was significantly reduced. Overexpression of wild-type L1CAM, but not of the L1CAM point mutants R1166X and S1224L, rescued the decrease in ankyrin levels. Importantly, we found that the L1CAM mutation selectively decreased activity-dependent Na+-currents, al-tered neuronal excitability, and caused impairments in action potential (AP) generation. Thus, our results suggest that the clinical presentations of L1CAM mutations in human patients could be accounted for, at least in part, by cell-autonomous changes in the functional development of neurons, such that neurons are unable to develop normal axons and dendrites and to generate normal APs.

Conditional deletion of L1CAM in human neurons impairs both axonal and dendritic arborization and action potential generation

Christopher Patzke,1 Claudio Acuna,1 Louise R. Giam,1 Marius Wernig,3,4 and Thomas C. Südhof 1,2

1Department of Molecular and Cellular Physiology, 2Howard Hughes Medical Institute, 3Institute for Stem Cell Biology and Regenerative Medicine, and 4Department of Pathology, Stanford University School of Medicine, Stanford, CA 94305

© 2016 Patzke et al. This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http ://www .rupress .org /terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http ://creativecommons .org /licenses /by -nc -sa /3 .0 /).

Correspondence to Christopher Patzke: patzke@stanford.edu

Abbreviations used: AAV, adeno-associated virus; AIS, axon initial segment; AP, ac-tion potential; cKO, conditional KO; ES, embryonic stem; iN, induced neuron.

L1CAM mutation impairs human axon and dendrite development | Patzke et al.500

of the disease-causing mutations in L1CAM appear to be loss-of-function mutations.

Interestingly, an ethanol-binding site disrupting the in-terface between Ig-domains 1 and 4 of L1CAM has been identified. This site might explain the inhibitory effects of ethanol on L1CAM-mediated cell adhesion and neurite out-growth, and could contribute to neuropathological abnor-malities observed in fetal alcohol spectrum disorders, which exhibit features that are similar to those observed in L1 syn-drome patients (Ramanathan et al., 1996; Bearer et al., 1999; Arevalo et al., 2008). Electron microscopy studies on L1CAM and data from a crystal structure of the N-terminal Ig do-mains 1–4 of the L1CAM family member neurofascin, as well as a cryo-electron tomography report on liposomes supple-mented with L1CAM ectodomains, revealed a horseshoe-like structure of the Ig domains 1–4 (Schürmann et al., 2001; He et al., 2009; Liu et al., 2011). Based on the structure of the Ig domains 1–4 of the L1CAM homologue Axonin-1, it has been suggested that two horseshoes on opposing cells interact in a zipper-like manner, mediating homophilic cell adhesion (Freigang et al., 2000). Ethanol, and disease-causing missense mutations in the ethanol-binding pocket (e.g., Leu-120-Val and Gly-121-Ser), likely disrupt the horseshoe-shaped struc-ture and inhibit homophilic and heterophilic interactions of L1CAM (Bateman et al., 1996; De Angelis et al., 1999, 2002; Arevalo et al., 2008). However, in contrast to the notion that Ig domains 1–4 are essential for homophilic binding, neurons from a reported L1CAM mutant mouse line lacking only Ig domain 6, which contains the integrin-binding motif RGD, failed to attach to L1CAM in vitro (Itoh et al., 2004), sug-gesting a more complicated scenario for the homophilic ac-tivity of L1CAM on neurons.

Studies using constitutive L1CAM-deficient mice as a model system reported defects in axon guidance in the cor-ticospinal tract, impaired growth of pyramidal layer V neuron apical dendrites, reduced size of the corpus callosum, malfor-mations of the ventricular system and the cerebellar vermis, decreased association of axons with nonmyelinating Schwann cells, and reduced inhibitory synaptic transmission (Dahme et al., 1997; Cohen et al., 1998; Fransen et al., 1998; Demya-nenko et al., 1999; Saghatelyan et al., 2004). Puzzlingly, mutant mice expressing L1CAM with a truncated intracellular domain that lacks the ankyrin-binding region displayed no abnor-mal brain development, but exhibited a dramatic decrease in L1CAM expression and defects in motor functions in adult mice (Nakamura et al., 2010). However, conditional KO (cKO) mice where the L1CAM gene is inactivated in adult brain by crossing them to a calcium/calmodulin-dependent, kinase II promoter–driven Cre-line did not display these overt morpho-logical abnormalities, but instead exhibited an increase in basal excitatory synaptic transmission (Law et al., 2003). Thus, the precise functions of L1CAM and its mechanisms of action are not yet completely understood.

The clinical presentation of human L1CAM mutations, together with the phenotypes of the various mouse models,

suggests that L1CAM contributes to the migration of neurons, neurite outgrowth, axon guidance, and axon–glia interaction. Although the mouse mutants provided important information, the overall phenotypes of the mutant mice were surprisingly mild and the effect of L1CAM mutations on human neurons remains unknown. Here, we used cre/lox technology in human embryonic stem (ES) cell–derived neurons to study the effects of L1CAM mutations on human neuronal function at the cel-lular level in a controlled genetic background. Importantly, this manipulation enables the direct comparison of matching cell populations in all experiments. By creating a conditional hem-izygous KO (i.e., a KO of an X-chromosomal gene in a male cell), we could dissect the fundamental role of L1CAM in axo-nal and dendritic arborization in human neurons, and examine its function in action potential (AP) generation. Our data reveal a dramatic impairment in axon development that results in a decrease in axon length, a change in the protein composition of the axon initial segment (AIS), and a severe impairment in AP generation. In addition, we observed a less severe impairment in dendritic arborization. These results suggest that L1CAM mutations cell-autonomously impair neuronal function by im-peding normal axon function.

RES ULTSConditional hemizygous mutant human neurons as a model system for human L1CAM mutationsWe generated human neurons by forced expression of Ngn2 in human ES cells using our previously described induced neuron (iN) cell protocol (Zhang et al., 2013). We then exam-ined L1CAM expression in these WT neurons by immuno-cytochemistry (Fig. 1). We observed abundant expression of L1CAM in 5-wk-old mature neurons. L1CAM was specifi-cally localized to axons, but was absent from MAP2-positive dendrites, suggesting that iN cells could serve as a powerful tool to study L1 syndrome–related phenotypes at the cellu-lar level in human neurons (Fig. 1, A–C). When we assessed the distribution of L1CAM in developing iN cells, however, we detected L1CAM on all neurites at early stages (DIV 4), including MAP2-positive neurites, followed by selective L1CAM accumulation in MAP2-negative processes only at later stages (DIV 14–21 and mature neurons; Fig.  2). Thus, during neuronal development, L1CAM initially localizes to all neurites (which then differentiate into axons and den-drites), and only later becomes restricted to axons.

To examine the effect of L1CAM loss-of-function mu-tations on the properties of human neurons, we introduced a cKO mutation into the endogenous L1CAM gene of H1 ES cells (which are male, rendering a hemizygous mutation equivalent to a complete KO) by homologous recombina-tion. For this purpose, we flanked exon 3 (which encodes part of the first Ig domain) of the L1CAM gene with loxP sites (Fig. 3, A and B; and Materials and methods). Deletion of exon 3 creates an out-of-frame junction of exons 2 to 4, thus producing a KO. Two independently targeted ES cell clones were generated to exclude possible effects by unwanted clonal

501JEM Vol. 213, No. 4

variation or incidental genetic changes, and were converted into cKO cells by removal of the puromycin resistance cas-sette using transduced flp recombinase (Fig. 3 B).

We produced iN cells from both L1CAM cKO ES cell clones by expression of Ngn2 delivered by recombinant lenti-viruses. We also expressed WT active cre-recombinase (Cre) or mutant inactive cre-recombinase (ΔCre) in the iN cells duringinduction to generate precisely matching L1CAM-deficient mutant neurons and control neurons, respectively (Fig.  3, C and D; mutant neurons from the two clones are referred to as

−/y #1 and −/y #2, and WT neurons as +/y #1 and +/y #2). As analyzed by immunoblotting, we found that, consistent with previous studies (Wolff et al., 1988; Schäfer and Altevogt, 2010), control neurons produced multiple L1CAM protein variants with apparent molecular weights between 220 and 65 kD (Fig. 3 E). Upon cre recombination, neurons derived from the two cKO clones showed a complete lack of all L1CAM vari-ants, and exhibited no L1CAM expression as judged by immu-nocytochemistry (Fig. 3, E and F), demonstrating that the two cKO ES cell clones can be efficiently converted into control and L1CAM mutant human neurons with precisely matching backgrounds that only differ in the expression of L1CAM.

No change in survival, but decreased MAP2a/b and ankyrin expression in L1CAM mutant human neuronsL1CAM-deficient human neurons exhibited similar survival as control neurons, with no evidence of cell death or degen-eration as a result of the L1CAM deletion (Fig. 3, G and H). Furthermore, we observed no increase in phosphorylation of JNK/SAPK, which is activated in stressed cells (Fig. 4, C and D). Quantitative immunoblotting analyses revealed that the L1CAM deletion did not alter the majority of tested neuronal marker proteins; in particular, no changes in the levels of the majority of synaptic proteins were detected (Fig. 4, A and B). In contrast, we observed a decrease in the levels of the adaptor proteins ankyrinG and ankyrinB, which are axonal proteins that bind to the intracellular sequences of L1CAM (Davis and Bennett, 1994). We found a reduction of ∼50% and ∼20%of ankyrinG and ankyrinB, respectively (Fig.  4, A and B). In addition, we detected a ∼30% decrease in the dendriticmarker MAP2a/b and complexins. Because it was reported that L1CAM-activated signaling stimulates the mitogen- activated protein kinase (MAPK) pathway (Schaefer et al., 1999; Schmid et al., 2000; Maness and Schachner, 2007), we examined iN cells derived from both cKO clones for changes in MAPK activation, but did not detect significant changes in phosphorylation levels of Erk 1/2 or of related components in the mutant neurons (Fig. 4, C and D). Together, our data suggest that the L1CAM deletion in human neurons does not significantly impair neuronal survival or activate MAPK signaling, but decreases the levels of the axonal proteins ankyrinB and ankyrinG, the dendritic protein MAP2a/b, and the presynaptic protein complexin.

Axonal and dendritic arborizations are reduced in human L1CAM mutant neuronsTo test for changes in neuronal morphology, we next sparsely transfected (by the CaPO4 method) conditionally hemizygous mutant iN cells with a plasmid expressing TetO-EGFP and stained the neurons for MAP2a/b. Assuming that cytoplasmic EGFP visualizes the complete neuronal morphology and that MAP2a/b is selectively expressed in dendrites, this experiment allowed us to systematically distinguish between dendrites and axons (Fig. 5 A). By comparing mutant with WT cells, we ob-served in neurons derived from both independent cKO clones

Figure 1. Axonal localization of L1CAM in human WT neurons. (A–C) Representative confocal micrographs (from three independent exper-iments) of L1CAM on human neurons (iN cells) obtained by forced expres-sion of Ngn2 in ES cells and analyzed 5 wk after Ngn2 induction. Neurons were double labeled by immunofluorescence for L1CAM and MAP2 (A), TuJ1 (B), or Neurofilament-L (C). Each panel shows lower (top) and higher magnifications (bottom; boxed area in lower magnification images).

L1CAM mutation impairs human axon and dendrite development | Patzke et al.502

Figure 2. L1CAM partially colocalizes with MAP2 in the dendrites of human neurons at early developmental stages. (A) Representative con-focal micrographs of L1CAM on human neurons taken 4 and 21 d after Ngn2 induction. Neurons were triple-labeled by immunofluorescence for L1CAM, MAP2, and TuJ1. (B and D) Intensity profile of immunolabeling for the indicated proteins derived from line scans along the dashed line in the bottom left and second from the right images in A. (C and E) Summary graphs of Pearson’s correlation coefficients for co-localization of L1CAM versus MAP2 (C) and L1CAM versus TuJ1 (E) determined in multiple experiments at different times after Ngn2 induction as indicated (mature = 35 d after Ngn2 induction). Data are means ± SEM; statistical comparisons were performed by Student’s t test comparing various times after Ngn2 induction to mature neurons (*, P < 0.05; **, P < 0.01; ***, P < 0.001; numbers of images/independent experiments analyzed are shown in the bars).

503JEM Vol. 213, No. 4

Figure 3. Generation of human ES cells with a conditional L1CAM gene loss-of-function mutation andeffect of L1CAM mutation on neuronal survival. (A) L1CAM contains six Ig-domains and five FN-type 3 domains followed by a single transmembrane sequence, and a highly conserved short intracellular cytoplasmic domain. (B) Targeting strategy. The L1CAM gene was mutated by homologous recombination in male H1 ES cells using a recom-binant AAV, and correctly recombined mutant ES cells were verified by PCR. The introduced mutations flank anout-of-frame exon (exon 3) with loxP sites, and include a puromycin (Purom.) resistance cassette for selection; this cassette was removed using Frt-recombinase transduction to generate L1CAM cKO cells. For generation of precisely matched neurons containing or lacking L1CAM, mutant Cre-recombinase (ΔCre) or active Cre-recombinase (Cre) wereexpressed in iN cells during induction of neuronal differentiation as indicated. (C) Design of the lentiviruses used for conversion of ES cells into iN cells and for expression of inactive (ΔCre) or active Cre-recombinase (Cre). (D) Flow diagram of all experiments. cKO cells were infected at day –1 with lentivirusesshown in B, and differentiation was started on day 0 by activating Ngn2 expression with doxycycline. Analyses were conducted at least 21 d after induction of Ngn2. (E) Ponceau-stained blotting membranes (left) and immunoblots (right) of iN cells derived from two independent L1CAM cKO ES cell clones (#1 and #2). Immunoblots were stained with antibodies to human L1CAM (green) and GDI (red; loading control); signals were visualized with fluorescently labeled secondary antibodies (right). (F) Representative immunofluorescently labeled control and mutant L1CAM-deficient neurons. L1CAM, MAP2, and DAPI staining is shown. (G and H) Plot of the fraction of surviving neurons as a function of time in culture (G; dashed line indicates control neurons). As in all experiments, neurons from two independent L1CAM cKO ES cell clones (blue and orange) were analyzed. Data are means ± SEM (n = 3 independent biological experiments).

L1CAM mutation impairs human axon and dendrite development | Patzke et al.504

that deletion of L1CAM caused a massive decrease in the total length and number of branching points of axons, but produced no changes in cell soma size (Fig. 5, A, C, D, and G). More-over, deletion of L1CAM decreased the total length and branch numbers of dendrites, although the effect on dendrites was smaller than that on the length and branches of axons (Fig. 5, E and F), suggesting that the partial codistribution of L1CAM and MAP2 in very young neurons is important for dendritic tree development (Fig. 2). Interestingly, ∼15–25% of L1CAMmutant neurons developed two axons, compared with only 0–3% of WT neurons (Fig. 5, B and H). To test whether this morphological phenotype persists at later developmental stages, we analyzed iN cells at 180 d after induction. Because of the immense axonal growth constraining single-cell axon analyses (Fig. 6, A and B), we measured the dendritic arborization of these “old” human neurons and detected a defect in the mutant neurons (Fig. 6, C–F). Together with the quantitative immuno-

blotting results and the localization of L1CAM on developing and mature human neurons, these observations indicate that loss-of-function mutations of L1CAM impair dendritic and axonal arborization of human neurons.

L1CAM deletion selectively decreases ankyrinG levels in the AISThe marked decrease of ankyrinG levels in hemizygous mu-tant neurons (Fig. 4, A and B) suggests a potential defect in the AIS because ankyrinG is a key adaptor molecule that clusters Na+ and K+ channels at the AIS and partially co- localizes with L1CAM (Fig. 7 A; Zhou et al., 1998; Garrido et al., 2003; Kole and Stuart, 2012). To assess this aspect, we measured the area and intensity of ankyrinG staining in the AIS and found that both parameters decreased in L1CAM mutant human neurons (Fig.  7, C, E, and F). Investigating ankyrinB, which also co-localizes with L1CAM in axons

Figure 4. Protein levels in hemizygous mutant human neurons. (A and B) Representative immunoblots (A) and summary graph of protein levels (B) in iN cells produced from independent L1CAM cKO clones #1 and #2. Values in B represent mean protein levels observed in L1CAM KO neurons normalized to those of matching isogenic WT controls (dashed line), and corrected for blotting and loading variations using GDI as an internal standard. (C and D) Rep-resentative pictures and summary graph of phosphorylation levels of AKT, ERK1/2, CREB, and JNK in iN cells from clones #1 and #2 and matched controls (dashed line). Data are means ± SEM; statistical significance was assessed using Student’s t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

505JEM Vol. 213, No. 4

(Fig.  7  B), we additionally detected a ∼30% decrease inankyrinB staining intensity in mutant axons (Fig.  7, D and G). Together with the changes in axonal arborization (Fig. 5), these results suggest a fundamental role of L1CAM in the morphological development of axons in human neurons.

No decrease in synapse density in human mutant neuronsTo test the hypothesis that L1 syndrome symptoms and its associated severe intellectual disability are caused by mor-

phological changes at the synaptic level in human neurons, we quantified presynaptic synapsin-positive puncta (which partially overlap with L1CAM on axons from WT neurons; Fig.  8  A) on post-synaptic MAP2-positive dendrites. Both the size and the number of synapses per dendritic segment in WT and mutant neurons were compared (Fig. 8 B), and no significant differences could be detected (Fig. 8, C and D). No alterations of the intrinsic electrical properties were detected (Fig.  8, E–H), suggesting no alterations in the membrane

Figure 5. Deletion of L1CAM in human neurons dramatically impairs axonal growth and modestly decreases dendritic arborization. (A) Repre-sentative confocal images of control (top) and L1CAM-deficient iN cells (middle and bottom) that were sparsely transfected with EGFP at day 19 after Ngn2 induction to visualize the complete axonal and dendritic arbors of neurons, and that were additionally stained for MAP2 and DAPI as indicated at day 21 after Ngn2 induction. (B) Atypical iN cells with two independent axons. The multi-axon phenotype is significantly more frequently observed in L1CAM KO than in control neurons (see H). (C–G) Summary graphs of axonal and dendritic lengths and number of branches, as well as soma size of mutant and wild-type neurons from clones #1 and #2. Statistical significance was assessed using Student’s t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Three independent biological experiments were performed for each mutant clone and matching WT control. (H) Summary graph of the percentage of cells with one or two axons. Numbers in bars indicate number of cells/independent experiments analyzed.

L1CAM mutation impairs human axon and dendrite development | Patzke et al.506

composition. In summary, our results suggest that L1CAM is not required for synapse development as such, although we cannot exclude any functional synaptic changes at this point.

L1CAM deletion impairs AP generation and decreases neuronal excitability in mutant human neuronsTo investigate if L1CAM deficiency, and thus a decrease of ankyrinG at the AIS, leads to functional consequences in AP firing, we performed patch-clamp recordings and measured the whole-cell density of Na+ and K+ currents in 5-wk-old iN cells in voltage-clamp configuration (Fig. 9, A–D). Cells were held at −80 mV, and then the membrane potential was sequen-tially depolarized in 10-mV steps for 2 s (Fig. 9, A and B). We observed a significant decrease in the peak amplitude of Na+ currents (inward), whereas K+ currents (outward) measured in the same experiments remained unchanged (Fig. 9, C and D). Thus, the removal of L1CAM seems to selectively reduce the total density of functional Na+ channels in human neurons.

To test whether the reduction in Na+ channels impacts AP generation, we performed current-clamp recordings and sequentially depolarized the cells by injecting positive cur-rent. We found that L1CAM hemizygous mutant iN cells re-quired larger current injections to reach AP threshold than

precisely matching WT iN cells (Fig. 9, E–G, and M). We also assessed the firing threshold and the shape of the APs, but found no statistically significant differences between control and mutant cells (Fig. 9, H–L and N–R). These results indi-cate that deletion of L1CAM expression in human neurons leads to a partial loss of ankyrinG from the AIS, a decrease in the density of Na+ channels, and a change in a neuron’s response to somatic current injections.

Ankyrin-binding site of L1CAM is required for normal expression levels of ankyrinG and ankyrinB in human neuronsTo evaluate the mechanistic role of the binding of ankyrins to L1CAM in maintaining expression levels of ankyrins, we next performed rescue experiments by lentiviral-mediated expression (infection on day −1; Fig. 3 D) in iN cells of WTL1CAM, or mutant L1CAM. The two pathological point mutations, R1166X and S1224L, either completely lack the ankyrin-binding site or contain a substitution in a key residue of the ankyrin-binding site (Fig. 10 A; Vos and Hofstra, 2010). Overexpressed WT, as well as R1166X and S1224L mutant, L1CAM could be detected by immunoblots at similar levels, and were both expressed on the surface of human neurons

Figure 6. Reduced dendritic arboriza-tion of L1CAM mutant human neurons after 180 d of culture. (A) Representa-tive confocal micrographs showing WT cells 180 d after iN induction stained for L1CAM and counter-stained for MAP2a/b, TuJ1, or Neurofilament-L. (B) Summary graph of Pear-son’s correlation coefficients for co-localiza-tion of the indicated proteins analyzed by double immunofluorescence labeling. Values are means of three different pictures ± SEM from a single experiment. (C) Confocal micro-graph of dendritic tree visualized by MAP2a/b in mutant and matching control neurons 180 d after iN induction. (D–F) Summary graphs of dendritic lengths and number of branches, as well as soma size of mutant and control neurons. Statistical significance was assessed using Student’s t test (***, P < 0.001) of cells from three independent experiments.

507JEM Vol. 213, No. 4

(Fig. 10, B and C). Interestingly, in contrast to WT L1CAM, R1166X, and S1224L mutant L1CAM did not rescue the decrease of ankyrin levels in mutant human neurons (which in this experiment were suppressed to ∼20% for ankyrinGand ∼30% for ankyrinB), suggesting a requirement of theL1CAM ankyrin protein interaction for the stabilization of ankyrins (Fig. 10, D–F).

DIS CUS SIONIn this study, we used the cre/lox technology in human neu-rons to investigate the effect of hemizygous L1CAM loss-of-function mutations on neuronal function. This experimental design eliminates potentially confounding effects induced by genetic background changes or selection of cell clones. So far, hundreds of mutations in the L1CAM gene (which is X-chromosomal) were reported in patients with severe de-velopmental malformations, neurological abnormalities, andintellectual disability (Rosenthal et al., 1992; Stumpel andVos, 1993; Jouet et al., 1994, 1995; Fransen et al., 1997; Voset al., 2010; Adle-Biassette et al., 2013; Chidsey et al., 2014).Most of the disease-causing L1CAM mutations likely repre-sent loss-of-function mutations. We found that conditionallyL1CAM-deficient human neurons manifest three types ofdefects: first, a large reduction in the size of axons, a smallerdecrease in dendritic arborizations, and a significant reduc-tion in the size of the AIS; second, a decrease in the levelsof ankyrinG and ankyrinB; and third, impairment of APgeneration. It seems likely that these three types of defectsare commonly produced by the loss of the L1CAM com-plex with ankyrins upon deletion of L1CAM, and that thesedefects are responsible, at least in part, for the symptoms ofpatients with L1 syndrome.

In mutant mice and in in vitro systems, it has been well established that the intracellular segment of L1CAM binds to ankyrins (Davis and Bennett, 1994; Bennett and Chen, 2001). In general, it has been shown that ankyrinB binding promotes neuritogenesis by coupling L1CAM to the actin cytoskele-ton (Nishimura et al., 2003). Binding L1CAM to ankyrinG, conversely, is thought to recruit L1CAM to the AIS, a highly specialized compartment in the proximal axon where APs are generated (Malhotra et al., 1998; Yoshimura and Rasband, 2014). Both ankyrins are thought to recruit and compartmen-talize L1CAM and other L1CAM family members, such as neurofascin (Zhou et al., 1998; Boiko et al., 2007). For exam-ple, it has been reported that L1CAM and ankyrinB codis-tribute on axons in neonatal mouse brains before myelination. Mice deficient for ankyrinB exhibit reduced levels of L1CAM on these axons and, because of their hypoplasia of the cor-pus callosum and dilated ventricles, share features of L1 syn-drome patients (Scotland et al., 1998). Furthermore, it has been

Figure 7. Double labeling of L1CAM and ankyrins at the AIS and axons of WT and KO human neurons. (A and B) Confocal micrographs of L1CAM and ankyrinB-positive axons, as well as of the ankyrinG-positive AIS of axons in WT neurons. (C) Representative confocal micrographs of an L1CAM-deficient and control iN cell illustrating the AIS. Den-drites are labeled by anti-MAP2a/b. (D) Confocal micrographs of L1CAM hemizygous control and mutant iN cell labeling ankyrinB. The Pearson’s correlation for co-localization of L1CAM and ankyrinB is 0.77 ± 0.02 (de-termined from three independent experiments with three to four images each). (E and F) Summary graphs of the ankyrinG-labeled AIS area (E) and the ankyrinG staining intensity of the AIS (F). Data are means ± SEM; numbers in bars indicate the numbers of cells/independent experiments performed; statistical comparisons were done using Student’s t test (*, P < 0.05; **, P < 0.01). (G) Summary graph of mean ankyrinB intensi-ties (per ankyrinB-positive pixel), indicating a reduction of ∼30% in the

axons of L1CAM-deficient human neurons. Data are means ± SEM; sta-tistical comparisons were performed by Student’s t test (*, P < 0.05; num-bers of images/independent experiments analyzed are shown in the bars).

L1CAM mutation impairs human axon and dendrite development | Patzke et al.508

shown that the cytoplasmic SFI GQY motif of L1CAM (aa 1224–1229, which is present in all L1CAM family members and is highly conserved from invertebrates to vertebrates) is essential for ankyrin binding (Garver et al., 1997; Needham et al., 2001). By introducing two pathological missense mutations, namely S1224L and Y1229H, reduced recruitment of ankyrin to the plasma membrane in vitro was observed (Needham et al., 2001). Knock-in mice expressing the missense mutation Y1229H exhibit mistargeting of mutant ganglion cell axons from the ventral retina to lateral sites in the contralateral su-perior colliculus (Buhusi et al., 2008). Additionally, observa-tions of neonatal L1CAM-deficient mice suggest a decrease of ankyrinB expression in thalamocortical and corticothalamic axons, together with axonal hyperfasciculation and path- finding errors (Wiencken-Barger et al., 2004). However, mice with a truncated intracellular L1CAM sequence lacking this ankyrin-binding region exhibited morphologically normal brains (Nakamura et al., 2010), raising questions about the pre-cise role of the intracellular L1CAM sequence.

In the present study, we show that L1CAM codistributes with ankyrinG and ankyrinB and that the deletion of L1CAM

decreases ankyrins on the subcellular levels (ankyrinG at the AIS, ankyrinB on axons; Fig. 7), and on the global protein level (as revealed by quantitative immunoblot analyses; Fig. 4, A and B), suggesting that L1CAM may actually recruit and stabilize ankyrins to specific subcellular localizations. In addition to pre-vious studies, our data show that on human axons L1CAM is not only important for the expression of ankyrinB and axonal arborizations but it is also crucial to the levels of ankyrinG at the AIS and its proper function in Na+-mediated AP generation. It is tempting to speculate that L1CAM and ankyrins are in-terdependent subunits of a complex stabilizing each other, and thereby regulating axonal morphological and functional devel-opment. Evidence for this notion comes from our rescue ex-periments, suggesting that the integrity of the ankyrin-binding site is important for the stabilization of ankyrins (Fig. 10).

AnkyrinG is a key component of the AIS and has been shown to be responsible for clustering Na+ and K+ channels at the AIS (Zhou et al., 1998; Garrido et al., 2003; Kole and Stuart, 2012). In the absence of ankyrinG in mouse neurons, Na+ currents are highly reduced and the shape of the AP is altered (Zhou et al., 1998; Hedstrom et al., 2008; Barry

Figure 8. Hemizygous L1CAM KO does not alter synapse density. (A) Micrographs of double-labeling by immunofluorescence showing L1CAM and Synapsin in axons of WT human neurons. (B) Representative images of dendrites from control and hemizygous L1CAM mutant iN cells stained for MAP2a/b and Synapsin to visualize presynaptic terminals. Clones #1 and #2 are shown. (left) Zoom-out of L1CAM +/y #1. (C and D) Summary graphs of synaptic density and area of Synapsin-positive puncta on dendrites from mutant and matching WT. (E–H) Input resistance and capacitance of hemizygous L1CAM mutant iN cells. Data are means ± SEM; no statistically significant difference was detected using Student’s t test.

509JEM Vol. 213, No. 4

et al., 2014). This phenotype is similar to the Na+ channel phenotype we observed in L1CAM-deficient human neu-rons. AnkyrinG mutations in patients have been associated with severe cognitive deficits, autism, sleeping disorder, severe

attention deficit hyperactivity disorder, and bipolar disorder (Ferreira et al., 2008; Schulze et al., 2009; Iqbal et al., 2013).

To date, a change in neuronal excitability has not been observed in L1CAM mouse mutants and could represent a

Figure 9. Decreased Na+ currents and reduced excitability of mutant iN cells. (A and B) Protocol used for voltage-clamp experiments and representative traces of whole-cell voltage- clamp Na+ and K+ currents recorded in control and L1CAM mu-tant neurons. Cells were subjected to 10-mV step depolarizations (2 s each) from −80 mV to +10 mV. (C and D) Quantification ofI/V curves of Na+ and K+ current densities. Currents were aligned for the maximal Na+ current densities. Data are means ± SEM; statistical comparisons were performed by ANO VA (***, P < 0.001; number of cells from three independent experiments is indicated). (E and F) Representative traces and protocol of analyses of the AP firing properties of control and L1CAM mutant neurons. Neurons were maintained at near −80 mV in current-clamp mode and were depolarized with increasing current pulses (2–20 pA increments) for 2  s, until an AP was triggered. (G and M) Summary graph showing cell excitability normalized by the capacitance of the iN cells. (H–L and N–R) Summary graphs analyzing the properties of the first AP. Data in G–L and M–R are means ± SEM; statistical comparisons were performed Student’s t test (*, P < 0.05; num-bers of cells/independent experiments analyzed are shown in the bars of G and M). Access resistance was on average 9.5 ± 0.3 MΩand 8 ± 0.2 MΩ for the control and 9.0 ± 0.2 MΩ and 7.4 ± 0.2MΩ for the mutant cells of clones #1 and #2, respectively. Leakcurrent was 25.2 ± 3.5 pA and 36 ± 4.4 pA for the control cells and 27.4 ± 4.6 pA and 36.5 ± 5.2 pA for the mutant cell of clones #1 and #2, respectively.

L1CAM mutation impairs human axon and dendrite development | Patzke et al.510

feature that is more easily distinguished in human neurons (Dahme et al., 1997; Cohen et al., 1998; Fransen et al., 1998; Demyanenko et al., 1999; Law et al., 2003; Saghatelyan et al., 2004; Nakamura et al., 2010). However, a disintegration of the AIS, APs with modified waveform upon stimulation, abolition of spontaneous firing in Purkinje cells, and deficits in motor learning have been reported for mice deficient for

neurofascin in adult cerebellar Purkinje cells. But in develop-ing neurons, neurofascin was shown to be dispensable for AIS assembly, while ankyrinG staining intensities were unchanged (Zonta et al., 2011). Neurofascin is thought to be essential for the clustering of Na+ channels at the nodes of Ranvier, an-other excitable region of the axon, by its binding to ankyrinG (Sherman et al., 2005).

Figure 10. WT, but not mutant, L1CAM can rescue reduced ankyrinG and ankyrinB levels in L1CAM-deficient human neurons, and mutant L1CAM suppresses ankyrinG levels in WT neurons. (A) Schematic representation of rescue constructs. S1224L produces an amino acid substitution in the intracellular ankyrin-binding site of L1CAM, whereas R1166X introduces a stop codon N-terminal to the ankyrin-binding site. (B) Immunoblots showing expression of rescue L1CAM constructs in hemizygous mutant and matched control human neurons. Note that in contrast to the monoclonal mouse L1CAM antibody that binds to the extracellular part of L1CAM, the polyclonal rabbit L1CAM antibody that binds to intracellular sequences (aa 1153–1182) does not detect R1166X. (C) Confocal micrographs (from three independent experiments) of anti-L1CAM surface staining of hemizygous mutant neurons expressing the rescue constructs or WT L1CAM. (D) Immunoblots of iN cells as shown in B, detecting the levels of ankyrinG and ankyrinB. (E and F) Summary graphs of ankyrinG (top) and ankyrinB protein levels (bottom), corrected for blotting and loading variations using GDI as an internal standard. Data are means ± SEM; statistical significance was assessed using Student’s t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

511JEM Vol. 213, No. 4

We observed an ∼20–50% decrease in ankyrinG levels,possibly leading to a decrease in the concentration of voltage- gated Na+ channels at the AIS. This result might explain the rather modest reduction of Na+ currents and the lack of sig-nificant changes in the shape of APs. However, this points to an important aspect of L1 syndrome because the missing L1CAM–ankyrin interactions likely contribute to the clinical presentations of L1CAM and ankyrin mutations. Future work could use this model system to test ways of restoring neuronal excitability, for example, by reducing the activity of resting K+ channels at the AIS. It has been shown that pharmacological inhibition of Cyclin-dependent kinase 2 or 5 (Cdk2 or Cdk5) by roscovitine,enriched K+ channels at the AIS (Vacher et al., 2011). One possible way to reduce K+ channel activity would be to pharmacologically activate Cdk2 or Cdk5 function.

Overall, our study using cre/lox technology on human neurons presents a first step in understanding L1 syndrome at the neuronal level. Our results confirm a morphological phenotype previously observed in animal models (Dahme et al., 1997; Cohen et al., 1998; Demyanenko et al., 1999), but also presents an additional component in the human mutant neuron’s reduced ability to generate APs.

MAT ERI ALS AND MET HODSViral constructs. The following lentiviral constructs were used: FUW-TetO-Ng2-T2A-puromycin expressing TetO- Ng2-T2A-puromycin cassette (TetO promoter drives ex-pression of full-length mouse Ngn2 and of puromycin via the cleavage-peptide sequence T2A; Fig. 3 B); F-Ubiquitin- W (FUW)-rtTA containing rtTA; FUW-TetO-EGFP ex-pressing EGFP; FSW-NLS-mCherry expressing mCherry preceded by a nuclear localization sequence (under the con-trol of neuron-specific human synapsin promoter) to monitor cell death; FUW-Flp to express Flp-recombinase; FUW- GFP::Cre to express Cre-recombinase to delete L1CAM exon 3 and create via frameshift a null-allele or GFP::ΔCrefor the wild-type control; FUW-TetO-wild-type-L1CAM; FUW-TetO-R1166X-L1CAM; and FUW-TetO-S1224L-L1CAM. One adeno-associated virus (AAV) construct was used for gene targeting of X-chromosomal L1CAM allele of the male hESCs H1 (also shown in Fig.  3 B): the con-struct contains sequences from the region encoding exon 3 flanked by loxP sites and frt-sites flanked puromycin resis-tance cassette (Frt-PGK-Puro-SV40polyA-Frt) adjacent to the 5′ loxP site. For homologous recombination, the 5′ armof the construct includes 1.6 kb of sequences located up-stream of exon 3. The 3′ arm contains 1.4 kb of sequenceslocated downstream of exon 3.

Virus generation. Lentiviruses were produced as previously described (Zhang et al., 2013) in HEK293T cells (ATCC) by co-transfection with three helper plasmids (pRSV-REV, pMDLg/pRRE, and vesicular stomatitis virus G protein ex-pression vector) with 12 µg of lentiviral vector DNA and 6 µg of each of the helper plasmid DNA per 75 cm2 culture area)

using calcium phosphate. Lentiviruses were harvested in the medium 48  h after transfection, pelleted by centrifugation (49,000 g for 90 min), resuspended in MEM, aliquoted, and frozen at –80°C. Only virus preparations with >90% infec-tion efficiency as assessed by EGFP expression or puromycin resistance were used for experiments. AAV-DJ was used to deliver the targeting construct for generation of cKO cells. AAV-DJ was produced in HEK293T cells by co-transfection of pHelper, pDJ, and AAV vector (8.5 µg of DNA per 75 cm2 culture area) using calcium phosphate. Cells were harvested 72 h after transfection in PBS/1 mM EDTA and after one freezing/thawing cycle. AAVs were collected from cytoplasm using Benzonase nuclease at a final concentration of 50 U/ml at 37°C for 30 min. After clearing the suspension from cell debris by slow centrifugation (3,000 g for 30 min) AAVs were isolated after fast centrifugation (400,000 g for 120 min) in iodixanol (gradient from 15–60%) from the 40% layer and further concentrated using centricon concentrating tube (100,000 MWCO; EMD Millipore) according to the manu-facturer’s suggested protocol.

Cell culture. Experiments were performed as previously de-scribed (Zhang et al., 2013). H1 ES cells (WiCell Research Resources) were maintained as feeder-free cells in mTeSR1 medium (Stem Cell Technologies). Mouse glial cells were cul-tured from the forebrain of newborn WT CD1 mice. In brief, newborn mouse forebrain homogenates were digested with papain and EDTA for 30 min, and cells were dissociated by harsh trituration to avoid growing of neurons, and then plated onto T75 flasks in DMEM supplemented with 10% FBS. Upon reaching confluence, glial cells were trypsinized and replated at lower density a total of three times to remove po-tential trace amounts of mouse neurons before the glia cell cultures were used for co-culture experiment with iN cells.

Gene targeting in ES cells. H1 ES cells were transduced by AAVs to generate conditional hemizygous mutant ES cells by flanking exon 3 (ENSE00003680685, coding the aa MEP PVI TEQ SPR RLV VFP TDD ISL KCE ASG KPEV in the first Ig domain), which results in a frameshift and a premature stop codon upon cre recombination. Puromycin (1 µg/ml) was added and kept in mTeSR1 medium 2 d after transduction. The surviving ES cells were allowed to grow into colonies, and then individually picked. 7 correctly targeted colonies out of 76 colonies in total were confirmed by PCR screening using oligo sequences 5′-ACA GAT GAC ATC AGC CTC AAG TGT GAGG-3′ and 5′-GAC TGG GAG ATG GCG AGG ACT TG-3′, resulting in a 240-bp band for a correctly tar-geted clone or 185 bp for the untargeted WT. Correctly mu-tated alleles were confirmed after flp and cre recombination and iN differentiation.

Standard protocol for generation of iN cells from condi-tional mutant human ES cells and for activating the con-ditional mutations. iN cell generation has been previously

L1CAM mutation impairs human axon and dendrite development | Patzke et al.512

described (Zhang et al., 2013). In brief, targeted and flp-re-combined (after blasticidin selection) human ES cells were treated with Accutase (Innovative Cell Technologies) and plated as dissociated cells in 24-well plates (104 cells/well) on day −1 (Fig. 3 C). Cells were plated on Matrigel-coated(BD) coverslips in mTeSR1 containing 2  mM thiazovivin (Bio Vision). At the same time point, lentiviruses prepared as described in the paragraph "Virus generation"(0.3  µl/well of 24-well plate) were added. Two different types of lenti-viruses were co-infected: the lentiviruses used for iN cell induction as described, and lentiviruses expressing either Cre- recombinase (to create a null allele) under control of the ubiq-uitin promoter or an inactive mutated ΔCre-recombinase forthe WT control (Fig. 3 C). On day 0, the culture medium was re-placed with N2/DMEM/F12/NEAA (Invitrogen) containing human brain-derived neurotrophic factor (BDNF; 10 ng/ml; PeproTech), human NT-3 (10 ng/ml; PeproTech), and mouse Laminin-1 (0.2 µg/ml; Invitrogen). Doxycycline (2 µg/ml; Takara Bio Inc.) was added on day 0 to induce TetO gene expression and retained in the medium until the end of the experiment. On day 1, a 24-h puromycin selection (1 µg/ml) period was started. On day 2, mouse glia cells were added in neurobasal medium supplemented with B27/Glutamax (Invitrogen) containing BDNF, NT3, and Laminin-1; Ara-C (2 µM; Sigma-Aldrich) was added to the medium to inhibit astrocyte proliferation. After day 2, 50% of the medium in each well was exchanged every 2 d. FBS (2.5%) was added to the culture medium on day 10 to support astrocyte viability, and iN cells were assayed after at least 21 d or as indicated.

Quantification of neuronal morphology. iN cells were sparsely transfected by calcium phosphate on day 19 with a TetO-EGFP expression vector. On day 21, cells were fixed and stained for EGFP and MAP2a/b to discriminate be-tween dendrites and axons. Fixation at later time points would result in too abundant axonal growth. All MAP2- and EGFP-positive neurites were considered to be dendrites and all MAP2-negative but EGFP-positive neurites to be axons. Images were acquired using an A1RSi confocal mi-croscope system (Nikon), with a 10× or 20× objective at room temperature. Images of ∼10–30 neurons per condi-tion (n = 1 for axons; n = ∼100 in case dendrites) werereconstructed using the MetaMorph neurite application, scoring for total dendritic length, dendritic branch points, total axonal length, axonal branch points, and soma area. In the case of old human neurons, cells were fixed after DIV180 and immunolabeled.

Immunofluorescence and immunoblotting experiments. Im-munofluorescence experiments were performed essentially as previously described (Zhang et al., 2013). In brief, cultured iN cells were fixed in 4% paraformaldehyde and 4% sucrose in PBS for 20 min at room temperature, washed three times with PBS, and incubated in 0.2% Triton X-100 in PBS for 10 min at room temperature. Cells were blocked in PBS con-

taining 5% goat serum for 1 h at room temperature. Primary antibodies were applied overnight at 4°C, cells were washed in PBS three times, and fluorescently labeled secondary an-tibodies (Alexa Fluor; 1:1,000) were applied for 1 h at room temperature. The following antibodies were used in immu-nocytochemisty experiments: MAP2 a/b (Sigma-Aldrich; 1:1,000), MAP2 a/b (AB5622; EMD Millipore; 1:500), MAP2 (CPCA-MAP2; EnCor; 1:2,000), Synapsin (E028; 1:1,000), ankyrinG (106/36; Neuromab; 1:500), ankyrinB (105/13; Neuromab; 1:500), Neurofilament-L (171002; Syn-aptic Systems; 1:500), and GFP (Invitrogen; A11122; 1:2,000). To label L1CAM on human neurons, in some cases (Fig. 7) a post-fix staining using polyclonal Rb antibody (10140-R004-50; Sinobiological) was performed. In all other cases, a live staining was performed. The human-specific clone UJ127.11 (Sigma-Aldrich) was given into fresh culturing media for 20 min at 37°C (10 µg/ml). Next, cells were fixed and washed as described in this paragraph. Images were taken using an A1RSi confocal microscope system (Nikon), with a 10×, 20×, or 60× objective at room temperature. All quan-titative immunoblotting experiments were performed with fluorescently labeled secondary antibodies (LiCor; 1:5,000). Samples were separated by SDS-PAGE under reducing conditions and transferred onto nitrocellulose membranes. Blots were blocked in Tris-buffered saline containing 0.1% Tween 20 (Sigma-Aldrich) and 5% fat-free milk for 2 h at room temperature. The blocked membrane was incubated in blocking buffer containing the primary antibody overnight at 4°C, followed by three to five washes. The washed mem-brane was incubated in blocking buffer containing secondary antibody for 2  h at room temperature. Blots were scanned with the Odyssey-system (LiCor), followed by quantification with ImageStudio software (LiCor). For immunodetection, the following antibodies were used: NeuN (ABN78; EMD Millipore), TuJ1 (MMS-435P; Covance), Complexin 1/2 (L668), SNAP25 (P913), Synaptobrevin-2 (P939), Synapsin (E028), Syntaxin-1 (438B), β-Actin (A1978; Sigma-Aldrich),Synaptophysin (Synaptic Systems, 7.2), GDP-dissociation in-hibitor (Synaptic Systems; GDI; 81.2), calmodulin-associated serine/threonine kinase (BD; clone 7/CASK), L1CAM (UJ127.11; Sigma-Aldrich), L1CAM (ab123990; Abcam), SynCAM (T2412), MAP2 a/b (AB5622; EMD Millipore; 1:500), MAP2 a/b/c/d (Sigma-Aldrich; SAB4300658; 1:500), Neurofilament-L (171002,SySy; 1:500), Neurofilament-H (Abcam; RT97; 1:500), ankyrinG (Neuromab; 106/20, 1:500), ankyrinB (105/13; Neuromab; 1:500), Tau-5 (EMD Milli-pore; MAB361; 1:1,000), PSD95 (Abcam; ab76115; 1:500), GluR1 (Abcam; ab1504; 1:500), CREB (Cell Signaling Tech-nology; 86B10; 1:1,000), phospho-CREB (Cell Signaling Technology; 87G3; 1:1,000), Erk1/2 (Cell Signaling Tech-nology; 137F5; 1:1,000), phospho-Erk1/2 (Cell Signaling Technology; E10; 1:1,000), JNK (Cell Signaling Technology; 9252; 1:1,000), phospho-JNK (Cell Signaling Technology; G9; 1:1000), Akt (Cell Signaling Technology; 5G3; 1:1,000), phospho-Akt (Cell Signaling Technology; 9271; 1:1,000).

513JEM Vol. 213, No. 4

Quantification of synaptic density and survival. For syn-apsin puncta analyses, images were acquired using a A1RSi confocal microscope system (Nikon; 60× objective at room temperature), and the puncta density was determined using the software NIS-Elements (Nikon). Analyses of the sur-vival of iN cells was directly monitored using images of iN cell nuclei expressing mCherry at the same position of the culture dish taken every other day, using a DFC400 digital camera (Leica) attached to a DMIL LED inverted microscope with a 5× objective (Leica), driven by Appli-cation Suite image-acquisition software (Leica). Number of cells was determined using ImageJ software (National Institutes of Health). For each experimental condition, the mean of the pictures from 20–30 culture wells (of a 96-well culture plate) was considered as n = 1. In total, a mean of n = 3 was calculated.

Electrophysiology. All electrophysiological recordings were performed using whole-cell patch clamp. In brief, 5 wk after iN induction, on the day of recording, a coverslip containing relatively low-density induced human neurons was placed in a recording chamber mounted onto an Axioskop 2F upright microscope (Zeiss) equipped with DIC and fluorescence ca-pabilities. Cells were approached under DIC with ∼2 MΩpipettes pulled from borosilicate glass (Warner Instruments, Inc.) using a vertical PC-10 puller (Narishige), and impaled until a high-resistance seal (GΩ) was formed between therecording pipette and the cell membrane. Next, whole-cell configuration was established by gentle application of neg-ative pressure trough the recording pipette. The recording pipette contained (in mM): 125 K-gluconate, 20 KCl, 10 Hepes, 0.5 EGTA, 4 ATP-Magnesium, 0.3 GTP-Sodium, and 10 Na-Phosphocreatine; osmolarity, 312 mOsm, pH 7.2, adjusted with KOH. Electrical signals were recorded at 25 kHz with a two-channel Axoclamp 700B amplifier (Axon Instruments) and digitalized with a Digidata 1440 digitizer (Molecular Devices) that was, in turn, controlled by Clampex 10.1 (Molecular Devices). All recordings were performed at ∼24°C.

For whole-cell voltage-clamp recordings, neurons were maintained at −80 mV holding potentials. Series resistancevaried between 4–10 MΩ. All iNs in which series resistancewas higher than that were not included in the analysis. In all current-clamp experiments, the membrane potential was maintained approximately −80 mV by constant injectionof negative current through the recording pipette. Samples in the recording chamber were continuously perfused with oxygenated (95% O2/5%CO2) bath solution containing (in mM): 125 NaCl, 2.5 KCl, 1 MgCl2, 2 CaCl2, 25 glucose, 1.25 NaH2PO4, 0.4 ascorbic acid, 3 myo-inositol, 2 Na-pyruvate, and 25 NaHCO3, pH 7.4, and 315 mOsm.

Data presentation and statistics. All data shown are means ± SEMs. Number of measured cells and independent experiments is indicated inside each bar, or mentioned in the figure legend.

All statistical analyses were performed using either two-tailed Student’s t test or two-way ANO VA, comparing the test sample to the control sample examined in the same experiments.

Study approval. The present study was approved by Stem Cell Research Oversight (SCRO) at the Stanford Univer-sity Research Compliance Office (SCRO 518: Studying brain diseases affecting synaptic transmission by using human iNs). Experiments involving animals were approved by the Stanford University Institutional Animal Care and Use Com-mittee, Administrative Panel on Laboratory Animal Care Re-search Compliance Office.

ACK NOW LED GME NTSThis work was supported by grants from the National Institutes of Health (MH092931 to M. Wernig and AG010770 to S. Prusiner) and from the California Institute for Regenerative Medicine (RT2 02061 to M. Wernig), and by a postdoc-toral fellowship to C. Patzke (DFG PA 2110/1-1) and L.R. Giam (National Science Foundation Award 1202829 and the Helen Hay Whitney Foundation). M. Wernig is a New York Stem Cell Foundation-Robertson Investigator.

The authors declare no competing financial interests.

Submitted: 8 June 2015

Accepted: 12 February 2016

REfERENCESAdle-Biassette, H., P. Saugier-Veber, C. Fallet-Bianco, A.L. Delezoide, F.

Razavi, N. Drouot, A. Bazin, A.M. Beaufrère, B. Bessières, S. Blesson, et al. 2013. Neuropathological review of 138 cases genetically tested for X-linked hydrocephalus: evidence for closely related clinical entities ofunknown molecular bases. Acta Neuropathol. 126:427–442. http ://dx .doi .org /10 .1007 /s00401 -013 -1146 -1

Arevalo, E., S. Shanmugasundararaj, M.F. Wilkemeyer, X. Dou, S. Chen, M.E. Charness, and K.W. Miller. 2008. An alcohol binding site on the neural cell adhesion molecule L1. Proc. Natl. Acad. Sci. USA. 105:371–375. http ://dx .doi .org /10 .1073 /pnas .0707815105

Barry, J., Y. Gu, P. Jukkola, B. O’Neill, H. Gu, P.J. Mohler, K.T. Rajamani, and C. Gu. 2014. Ankyrin-G directly binds to kinesin-1 to transport voltage-gated Na+ channels into axons. Dev. Cell. 28:117–131. http ://dx .doi .org /10 .1016 /j .devcel .2013 .11 .023

Bateman, A., M. Jouet, J. MacFarlane, J.S. Du, S. Kenwrick, and C. Chothia. 1996. Outline structure of the human L1 cell adhesion molecule and the sites where mutations cause neurological disorders. EMBO J. 15:6050–6059.

Bearer, C.F., A.R. Swick, M.A. O’Riordan, and G. Cheng. 1999. Ethanol inhibits L1-mediated neurite outgrowth in postnatal rat cerebellar granule cells. J. Biol. Chem. 274:13264–13270. http ://dx .doi .org /10 .1074 /jbc .274 .19 .13264

Bennett, V., and L. Chen. 2001. Ankyrins and cellular targeting of diverse membrane proteins to physiological sites. Curr. Opin. Cell Biol. 13:61–67. http ://dx .doi .org /10 .1016 /S0955 -0674(00)00175 -7

Boiko, T., M. Vakulenko, H. Ewers, C.C. Yap, C. Norden, and B. Winckler. 2007. Ankyrin-dependent and -independent mechanisms orchestrate axonal compartmentalization of L1 family members neurofascin and L1/neuron-glia cell adhesion molecule. J. Neurosci. 27:590–603. http ://dx .doi .org /10 .1523 /JNE URO SCI .4302 -06 .2007

Buhusi, M., M.C. Schlatter, G.P. Demyanenko, R. Thresher, and P.F. Maness. 2008. L1 interaction with ankyrin regulates mediolateral topography in the retinocollicular projection. J. Neurosci. 28:177–188. http ://dx .doi .org /10 .1523 /JNE URO SCI .3573 -07 .2008

L1CAM mutation impairs human axon and dendrite development | Patzke et al.514

Chang, S., F.G. Rathjen, and J.A. Raper. 1987. Extension of neurites on axons is impaired by antibodies against specific neural cell surface glycoproteins. J. Cell Biol. 104:355–362. http ://dx .doi .org /10 .1083 /jcb .104 .2 .355

Chidsey, B.A., E.E. Baldwin, R. Toydemir, L. Ahles, H. Hanson, and D.A. Stevenson. 2014. L1CAM whole gene deletion in a child with L1 syndrome. Am. J. Med. Genet. A. 164A:1555–1558. http ://dx .doi .org /10 .1002 /ajmg .a .36474

Cohen, N.R., J.S. Taylor, L.B. Scott, R.W. Guillery, P. Soriano, and A.J. Furley. 1998. Errors in corticospinal axon guidance in mice lacking the neural cell adhesion molecule L1. Curr. Biol. 8:26–33. http ://dx .doi .org /10 .1016 /S0960 -9822(98)70017 -X

Dahme, M., U. Bartsch, R. Martini, B. Anliker, M. Schachner, and N. Mantei. 1997. Disruption of the mouse L1 gene leads to malformations of the nervous system. Nat. Genet. 17:346–349. http ://dx .doi .org /10 .1038 /ng1197 -346

Davis, J.Q., and V. Bennett. 1994. Ankyrin binding activity shared by the neu-rofascin/L1/NrCAM family of nervous system cell adhesion molecules. J. Biol. Chem. 269:27163–27166.

De Angelis, E., J. MacFarlane, J.S. Du, G. Yeo, R. Hicks, F.G. Rathjen, S. Kenwrick, and T. Brümmendorf. 1999. Pathological missense mutations of neural cell adhesion molecule L1 affect homophilic and heterophilic binding activities. EMBO J. 18:4744–4753. http ://dx .doi .org /10 .1093 /emboj /18 .17 .4744

De Angelis, E., A. Watkins, M. Schäfer, T. Brümmendorf, and S. Kenwrick. 2002. Disease-associated mutations in L1 CAM interfere with ligand interactions and cell-surface expression. Hum. Mol. Genet. 11:1–12. http ://dx .doi .org /10 .1093 /hmg /11 .1 .1

Demyanenko, G.P., A.Y. Tsai, and P.F. Maness. 1999. Abnormalities in neu-ronal process extension, hippocampal development, and the ventricular system of L1 knockout mice. J. Neurosci. 19:4907–4920.

Ferreira, M.A., M.C. O’Donovan, Y.A. Meng, I.R. Jones, D.M. Ruderfer, L. Jones, J. Fan, G. Kirov, R.H. Perlis, E.K. Green, et al. Wellcome Trust Case Control Consortium. 2008. Collaborative genome-wide association analysis supports a role for ANK3 and CAC NA1C in bipolar disorder. Nat. Genet. 40:1056–1058. http ://dx .doi .org /10 .1038 /ng .209

Fransen, E., G. Van Camp, L. Vits, and P.J. Willems. 1997. L1-associated diseases: clinical geneticists divide, molecular geneticists unite. Hum. Mol. Genet. 6:1625–1632. http ://dx .doi .org /10 .1093 /hmg /6 .10 .1625

Fransen, E., R. D’Hooge, G. Van Camp, M. Verhoye, J. Sijbers, E. Reyniers, P. Soriano, H. Kamiguchi, R. Willemsen, S.K. Koekkoek, et al. 1998. L1knockout mice show dilated ventricles, vermis hypoplasia and impairedexploration patterns. Hum. Mol. Genet. 7:999–1009. http ://dx .doi .org /10 .1093 /hmg /7 .6 .999

Freigang, J., K. Proba, L. Leder, K. Diederichs, P. Sonderegger, and W. Welte. 2000. The crystal structure of the ligand binding module of axonin-1/TAG-1 suggests a zipper mechanism for neural cell adhesion. Cell. 101:425–433. http ://dx .doi .org /10 .1016 /S0092 -8674(00)80852 -1

Garrido, J.J., P. Giraud, E. Carlier, F. Fernandes, A. Moussif, M.P. Fache, D. Debanne, and B. Dargent. 2003. A targeting motif involved in sodium channel clustering at the axonal initial segment. Science. 300:2091–2094. http ://dx .doi .org /10 .1126 /science .1085167

Garver, T.D., Q. Ren, S. Tuvia, and V. Bennett. 1997. Tyrosine phosphorylation at a site highly conserved in the L1 family of cell adhesion molecules abolishes ankyrin binding and increases lateral mobility of neurofascin. J. Cell Biol. 137:703–714. http ://dx .doi .org /10 .1083 /jcb .137 .3 .703

Gécz, J., C. Shoubridge, and M. Corbett. 2009. The genetic landscape of intellectual disability arising from chromosome X. Trends Genet. 25:308–316. http ://dx .doi .org /10 .1016 /j .tig .2009 .05 .002

Halliday, J., C.W. Chow, D. Wallace, and D.M. Danks. 1986. X linked hydrocephalus: a survey of a 20 year period in Victoria, Australia. J. Med. Genet. 23:23–31. http ://dx .doi .org /10 .1136 /jmg .23 .1 .23

He, Y., G.J. Jensen, and P.J. Bjorkman. 2009. Cryo-electron tomography of homophilic adhesion mediated by the neural cell adhesion molecule L1. Structure. 17:460–471. http ://dx .doi .org /10 .1016 /j .str .2009 .01 .009

Hedstrom, K.L., Y. Ogawa, and M.N. Rasband. 2008. AnkyrinG is required for maintenance of the axon initial segment and neuronal polarity. J. Cell Biol. 183:635–640. http ://dx .doi .org /10 .1083 /jcb .200806112

Iqbal, Z., G. Vandeweyer, M. van der Voet, A.M. Waryah, M.Y. Zahoor, J.A. Besseling, L.T. Roca, A.T. Vulto-van Silfhout, B. Nijhof, J.M. Kramer, et al. 2013. Homozygous and heterozygous disruptions of ANK3: at the crossroads of neurodevelopmental and psychiatric disorders. Hum. Mol. Genet. 22:1960–1970. http ://dx .doi .org /10 .1093 /hmg /ddt043

Itoh, K., L. Cheng, Y. Kamei, S. Fushiki, H. Kamiguchi, P. Gutwein, A. Stoeck, B. Arnold, P. Altevogt, and V. Lemmon. 2004. Brain development in micelacking L1-L1 homophilic adhesion. J. Cell Biol. 165:145–154. http ://dx.doi .org /10 .1083 /jcb .200312107

Jouet, M., A. Rosenthal, G. Armstrong, J. MacFarlane, R. Stevenson, J. Paterson, A. Metzenberg, V. Ionasescu, K. Temple, and S. Kenwrick. 1994. X-linked spastic paraplegia (SPG1), MASA syndrome and X-linkedhydrocephalus result from mutations in the L1 gene. Nat. Genet. 7:402–407. http ://dx .doi .org /10 .1038 /ng0794 -402

Jouet, M., A. Moncla, J. Paterson, C. McKeown, A. Fryer, N. Carpenter, E. Holmberg, C. Wadelius, and S. Kenwrick. 1995. New domains of neural cell-adhesion molecule L1 implicated in X-linked hydrocephalus and MASA syndrome. Am. J. Hum. Genet. 56:1304–1314.

Kenwrick, S., A. Watkins, and E. De Angelis. 2000. Neural cell recognition molecule L1: relating biological complexity to human disease mutations. Hum. Mol. Genet. 9:879–886. http ://dx .doi .org /10 .1093 /hmg /9 .6 .879

Kiefel, H., S. Bondong, J. Hazin, J. Ridinger, U. Schirmer, S. Riedle, and P. Altevogt. 2012. L1CAM: a major driver for tumor cell invasion andmotility. Cell Adhes. Migr. 6:374–384. http ://dx .doi .org /10 .4161 /cam .20832

Kole, M.H., and G.J. Stuart. 2012. Signal processing in the axon initial segment. Neuron. 73:235–247. http ://dx .doi .org /10 .1016 /j .neuron .2012 .01 .007

Law, J.W., A.Y. Lee, M. Sun, A.G. Nikonenko, S.K. Chung, A. Dityatev, M. Schachner, and F. Morellini. 2003. Decreased anxiety, altered place learn-ing, and increased CA1 basal excitatory synaptic transmission in mice with conditional ablation of the neural cell adhesion molecule L1. J. Neurosci. 23:10419–10432.

Liu, H., P.J. Focia, and X. He. 2011. Homophilic adhesion mechanism of neurofascin, a member of the L1 family of neural cell adhesion molecules. J. Biol. Chem. 286:797–805. http ://dx .doi .org /10 .1074 /jbc .M110 .180281

Malhotra, J.D., P. Tsiotra, D. Karagogeos, and M. Hortsch. 1998. Cis-activation of L1-mediated ankyrin recruitment by TAG-1 homophilic cell adhesion. J. Biol. Chem. 273:33354–33359. http ://dx .doi .org /10 .1074 /jbc .273 .50 .33354

Maness, P.F., and M. Schachner. 2007. Neural recognition molecules of the immunoglobulin superfamily: signaling transducers of axon guidance and neuronal migration. Nat. Neurosci. 10:19–26. http ://dx .doi .org /10 .1038 /nn1827

Nakamura, Y., S. Lee, C.L. Haddox, E.J. Weaver, and V.P. Lemmon. 2010. Role of the cytoplasmic domain of the L1 cell adhesion molecule in brain development. J. Comp. Neurol. 518:1113–1132. http ://dx .doi .org /10 .1002 /cne .22267

Needham, L.K., K. Thelen, and P.F. Maness. 2001. Cytoplasmic domain mu-tations of the L1 cell adhesion molecule reduce L1-ankyrin interactions. J. Neurosci. 21:1490–1500.

Nishimura, K., F. Yoshihara, T. Tojima, N. Ooashi, W. Yoon, K. Mikoshiba, V. Bennett, and H. Kamiguchi. 2003. L1-dependent neuritogenesis involves

515JEM Vol. 213, No. 4

ankyrinB that mediates L1-CAM coupling with retrograde actin flow. J. Cell Biol. 163:1077–1088. http ://dx .doi .org /10 .1083 /jcb .200303060

Ramanathan, R., M.F. Wilkemeyer, B. Mittal, G. Perides, and M.E. Charness. 1996. Alcohol inhibits cell-cell adhesion mediated by human L1. J. Cell Biol. 133:381–390. http ://dx .doi .org /10 .1083 /jcb .133 .2 .381

Rathjen, F.G., and M. Schachner. 1984. Immunocytological and biochemical characterization of a new neuronal cell surface component (L1 antigen) which is involved in cell adhesion. EMBO J. 3:1–10.

Rosenthal, A., M. Jouet, and S. Kenwrick. 1992. Aberrant splicing of neural cell adhesion molecule L1 mRNA in a family with X-linked hydrocephalus. Nat. Genet. 2:107–112. http ://dx .doi .org /10 .1038 /ng1092 -107

Saghatelyan, A.K., A.G. Nikonenko, M. Sun, B. Rolf, P. Putthoff, M. Kutsche, U. Bartsch, A. Dityatev, and M. Schachner. 2004. Reduced GABAergictransmission and number of hippocampal perisomatic inhibitorysynapses in juvenile mice deficient in the neural cell adhesion moleculeL1. Mol. Cell. Neurosci. 26:191–203. http ://dx .doi .org /10 .1016 /j .mcn .2004 .01 .008

Schaefer, A.W., H. Kamiguchi, E.V. Wong, C.M. Beach, G. Landreth, and V. Lemmon. 1999. Activation of the MAPK signal cascade by the neural cell adhesion molecule L1 requires L1 internalization. J. Biol. Chem. 274:37965–37973. http ://dx .doi .org /10 .1074 /jbc .274 .53 .37965

Schäfer, M.K., and P. Altevogt. 2010. L1CAM malfunction in the nervous system and human carcinomas. Cell. Mol. Life Sci. 67:2425–2437. http ://dx .doi .org /10 .1007 /s00018 -010 -0339 -1

Schäfer, M.K., and M. Frotscher. 2012. Role of L1CAM for axon sprouting and branching. Cell Tissue Res. 349:39–48. http ://dx .doi .org /10 .1007 /s00441 -012 -1345 -4

Schäfer, M.K., Y.C. Nam, A. Moumen, L. Keglowich, E. Bouché, M. Küffner, H.H. Bock, F.G. Rathjen, C. Raoul, and M. Frotscher. 2010. L1 syndrome mutations impair neuronal L1 function at different levels by divergent mechanisms. Neurobiol. Dis. 40:222–237. http ://dx .doi .org /10 .1016 /j .nbd .2010 .05 .029

Schmid, R.S., W.M. Pruitt, and P.F. Maness. 2000. A MAP kinase-signaling pathway mediates neurite outgrowth on L1 and requires Src-dependent endocytosis. J. Neurosci. 20:4177–4188.10818153

Schulze, T.G., S.D. Detera-Wadleigh, N. Akula, A. Gupta, L. Kassem, J. Steele, J. Pearl, J. Strohmaier, R. Breuer, M. Schwarz, et al. NIMH GeneticsInitiative Bipolar Disorder Consortium. 2009. Two variants in Ankyrin3 (ANK3) are independent genetic risk factors for bipolar disorder. Mol. Psychiatry. 14:487–491. http ://dx .doi .org /10 .1038 /mp .2008 .134

Schürmann, G., J. Haspel, M. Grumet, and H.P. Erickson. 2001. Cell adhesion molecule L1 in folded (horseshoe) and extended conformations. Mol. Biol. Cell. 12:1765–1773. http ://dx .doi .org /10 .1091 /mbc .12 .6 .1765

Scotland, P., D. Zhou, H. Benveniste, and V. Bennett. 1998. Nervous system defects of AnkyrinB (−/−) mice suggest functional overlap betweenthe cell adhesion molecule L1 and 440-kD AnkyrinB in premyelinated

axons. J. Cell Biol. 143:1305–1315. http ://dx .doi .org /10 .1083 /jcb .143 .5 .1305

Sherman, D.L., S. Tait, S. Melrose, R. Johnson, B. Zonta, F.A. Court, W.B. Macklin, S. Meek, A.J. Smith, D.F. Cottrell, and P.J. Brophy. 2005. Neurofascins arerequired to establish axonal domains for saltatory conduction. Neuron.48:737–742. http ://dx .doi .org /10 .1016 /j .neuron .2005 .10 .019

Stumpel, C., and Y.J. Vos. 1993. L1 Syndrome. In GeneReviews(R). R.A. Pagon, M.P. Adam, H.H. Ardinger, S.E. Wallace, A. Amemiya, L.J.H. Bean, T.D. Bird, C.R. Dolan, C.T. Fong, R.J.H. Smith, and K. Stephens, editors. University of Washington, Seattle, Seattle WA.

Vacher, H., J.W. Yang, O. Cerda, A. Autillo-Touati, B. Dargent, and J.S. Trimmer. 2011. Cdk-mediated phosphorylation of the Kvβ2 auxiliarysubunit regulates Kv1 channel axonal targeting. J. Cell Biol. 192:813–824. http ://dx .doi .org /10 .1083 /jcb .201007113

Vos, Y.J., and R.M. Hofstra. 2010. An updated and upgraded L1CAM mutation database. Hum. Mutat. 31:E1102–E1109. http ://dx .doi .org /10 .1002 /humu .21172

Vos, Y.J., H.E. de Walle, K.K. Bos, J.A. Stegeman, A.M. Ten Berge, M. Bruining, M.C. van Maarle, M.W. Elting, N.S. den Hollander, B. Hamel, et al. 2010. Genotype-phenotype correlations in L1 syndrome: a guide for geneticcounselling and mutation analysis. J. Med. Genet. 47:169–175. http ://dx.doi .org /10 .1136 /jmg .2009 .071688

Weller, S., and J. Gärtner. 2001. Genetic and clinical aspects of X-linked hydrocephalus (L1 disease): Mutations in the L1CAM gene. Hum. Mutat. 18:1–12. http ://dx .doi .org /10 .1002 /humu .1144

Wiencken-Barger, A.E., J. Mavity-Hudson, U. Bartsch, M. Schachner, and V.A. Casagrande. 2004. The role of L1 in axon pathfinding and fasciculation. Cereb. Cortex. 14:121–131. http ://dx .doi .org /10 .1093 /cercor /bhg110

Wolff, J.M., R. Frank, K. Mujoo, R.C. Spiro, R.A. Reisfeld, and F.G. Rathjen. 1988. A human brain glycoprotein related to the mouse cell adhesion molecule L1. J. Biol. Chem. 263:11943–11947.

Yoshimura, T., and M.N. Rasband. 2014. Axon initial segments: diverse and dynamic neuronal compartments. Curr. Opin. Neurobiol. 27:96–102. http ://dx .doi .org /10 .1016 /j .conb .2014 .03 .004

Zhang, Y., C. Pak, Y. Han, H. Ahlenius, Z. Zhang, S. Chanda, S. Marro, C. Patzke, C. Acuna, J. Covy, et al. 2013. Rapid single-step induction of functional neurons from human pluripotent stem cells. Neuron. 78:785–798. http ://dx .doi .org /10 .1016 /j .neuron .2013 .05 .029

Zhou, D., S. Lambert, P.L. Malen, S. Carpenter, L.M. Boland, and V. Bennett. 1998. AnkyrinG is required for clustering of voltage-gated Na channels at axon initial segments and for normal action potential firing. J. Cell Biol. 143:1295–1304. http ://dx .doi .org /10 .1083 /jcb .143 .5 .1295

Zonta, B., A. Desmazieres, A. Rinaldi, S. Tait, D.L. Sherman, M.F. Nolan, and P.J. Brophy. 2011. A critical role for Neurofascin in regulating action potential initiation through maintenance of the axon initial segment. Neuron. 69:945–956. http ://dx .doi .org /10 .1016 /j .neuron .2011 .02 .021

Review

The Rockefeller University Press $30.00J. Exp. Med. 2016 Vol. 213 No. 8 1375–1385www.jem.org/cgi/doi/10.1084/jem.20160493

1375

Alzheimer’s disease (AD) is a devastating neurodegenera-tive disorder affecting roughly 30 million people worldwide. Although some cases of AD (<1%) are caused by autoso-mal-dominant inherited mutations that typically lead to clin-ical disease onset before the age of 60, the majority of AD is late-onset AD (LOAD) where age, genetics, environment, and other diseases likely play a role (Holtzman et al., 2011; Musiek and Holtzman, 2015). AD is characterized by a cascade of pathological events, including the formation of amyloid plaques (made up of aggregated forms of Aβ), neurofibril-lary tangles (composed of aggregated, hyperphosphorylated tau), synapse loss, brain hypometabolism, neuroimflammation, and brain atrophy that is accompanied by severe and pro-gressive cognitive impairment. Amyloid plaques, consisting of aggregated forms of Aβ in the extracellular space, are gen-erated in a concentration-dependent manner. The buildup of hyperphosphorylated and aggregated tau protein leads to the development of intracellular neurofibrillary tangles. Ac-cumulation of Aβ occurs ∼15 yr before patients experiencecognitive decline, whereas tau begins to accumulate in the neocortex later but before the onset of dementia, adding to the complexity of this disease. Many risk factors for LOAD, both genetic and nongenetic, have been identified. Apart from aging, the strongest known risk factor for LOAD is genetic variation in the apolipoprotein E (APOE) gene. The APOE4 allele increases AD risk by 12-fold (two copies) or 3.7-fold (one copy) in part by influencing Aβ accumulation. However, APOE4 is only present in ∼50–60% of individuals with AD,suggesting that other factors are involved in AD pathogenesis (Holtzman et al., 2011).

One such risk factor for LOAD, which has received considerable attention is type 2 diabetes (T2D), which in-creases AD risk by at least twofold (Sims-Robinson et al., 2010). Also a disease of aging, T2D is characterized by hy-perglycemia, hyperinsulinemia, and insulin resistance (a lack of response in the insulin signaling [IS] pathway). Normally, insulin binds to the insulin receptor (IR) which phosphor-ylates IR substrate (IRS) on a tyrosine residue, leading to activation of the canonical signaling cascade (Fig. 1). In pe-ripheral tissues, such as muscle, fat, and liver, this signaling ultimately leads to the uptake and sequestration of glucose to satisfy cellular energy requirements and plays a key role in lipid metabolism (Dimitriadis et al., 2011). Contrary to the periphery, where glucose uptake is largely insulin depen-dent, the brain uses nearly 20% of all glucose in the body in a process that is largely insulin independent. However, brain IS is robust and has pleotropic effects due to the widespread distribution of IRs throughout the brain and the complexity of IR signaling. For example, hippocampal activation of IR signaling can modulate memory (McNay et al., 2010) and IR signaling in the hypothalamus can affect feeding behavior and peripheral metabolism (Brief and Davis, 1984). Similar to AD, pathological changes in insulin occur years before pa-tients receive a diagnosis of T2D, which typically occurs once pancreatic β cell dysfunction and insulin resistance producechronic hyperglycemia (Dankner et al., 2009). Interestingly, T2D alone has been associated with cognitive decline (Allen et al., 2004), brain hypometabolism (Roberts et al., 2014), and regional brain atrophy (Last et al., 2007). The cognitive defi-cits in T2D are proposed to be mediated by changes in brain IS (McNay and Recknagel, 2011), although there is little data from T2D patients measuring insulin/IS in the CNS to sup-port this assertion (Liu et al., 2011).

Individuals with type 2 diabetes have an increased risk for developing Alzheimer’s disease (AD), although the causal relation-ship remains poorly understood. Alterations in insulin signaling (IS) are reported in the AD brain. Moreover, oligomers/fibrils of amyloid-β (Aβ) can lead to neuronal insulin resistance and intranasal insulin is being explored as a potential therapy for AD. Conversely, elevated insulin levels (ins) are found in AD patients and high insulin has been reported to increase Aβ levels and tau phosphorylation, which could exacerbate AD pathology. Herein, we explore whether changes in ins and IS are a cause or consequence of AD.

Changes in insulin and insulin signaling in Alzheimer’s disease: cause or consequence?

Molly Stanley, Shannon L. Macauley, and David M. Holtzman

Department of Neurology, Hope Center for Neurological Disorders, Charles F. and Joanne Knight Alzheimer’s Disease Research Center, Washington University School of Medicine, St. Louis, MO

© 2016 Stanley et al. This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http ://www .rupress .org /terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http ://creativecommons .org /licenses /by -nc -sa /3 .0 /).

Correspondence to David M. Holtzman: holtzman@neuro.wustl.edu

Abbreviations used: Aβ, amyloid β; AD, Alzheimer’s disease; APOE, apolipoprotein E;APP, amyloid precursor protein; CSF, cerebrospinal fluid; GSK3, glycogen synthase kinase 3; ins, insulin level; IR, insulin receptor; IRS1, IR substrate 1; IS, insulin signaling; LOAD, late-onset AD; p-tau, phosphorylated tau; T2D, type 2 diabetes.

Insulin signaling in Alzheimer’s disease | Stanley et al.1376

There are two broad ways in which T2D could influ-ence the risk of AD: (1) T2D can lead to small vessel disease, which can cause or contribute to dementia, independent of or together with AD pathology, by disrupting proper function of the brain vasculature (Biessels and Reijmer, 2014), and (2) T2D can result in changes of brain function directly or in-teract with key proteins or pathways involved in AD pathol-ogy, such as Aβ or tau. This review will focus on mechanismsspecific to AD pathology, but acknowledges the significant impact that vascular alterations may have on the brain in AD and other dementias.

Over the past 15 yr, many studies have reported changes in insulin levels (ins) or IS (ins/IS) in LOAD patients (Table 1), suggesting that individuals with AD experience hyperinsulin-emia and brain insulin resistance. One interpretation is that the brain becomes insulin resistant as a consequence of AD pathology and hyperinsulinemia is compensatory, producing what has been termed type 3 diabetes (Steen et al., 2005). Insulin resistance in the AD brain may lead to cognitive im-pairment, similar to that observed in T2D patients, therefore treating individuals with AD with intranasal insulin to im-prove memory is currently under investigation in clinical trials (Reger et al., 2008; Craft et al., 2012; Wadman, 2012). Conversely, hyperinsulinemia and insulin resistance can mod-ulate Aβ and tau in ways that can put the brain at risk todevelop further AD pathology, so that changes in the ins/IS in AD patients may represent a contributor/cause of disease progression (Fig. 2). If so, increasing brain insulin could ex-acerbate AD pathology and cognitive decline over time. This proposes a unique problem as it relates to AD; whereas insulin

treatment may lead to modest cognitive improvement in AD patients, it could also worsen underlying pathology. In this review, we will analyze the changes in ins/IS that have been reported in AD and speculate if they are a cause or conse-quence of disease based on experimental evidence and the timeline of AD progression. Critical evaluation of this liter-ature as well as a determination of essential future experi-ments is crucial, as rates of both T2D and AD are on the rise and insulin therapy is actively being pursued worldwide in older adults and patients.

INS ULIN-REL ATED CHA NGES IN ADBrain insulin. Many groups have analyzed postmortem brain tissue from AD patients of varying severity and controls to look for alterations in Ins/IS through changes in mRNA, protein, or phosphorylation (Table 1). Insulin has been mea-sured in relatively low levels in brain tissue of humans and ro-dents (Banks et al., 1997; Frölich et al., 1998). Only one study reports insulin levels in the AD brain. They found that brain insulin was equally reduced in AD patients and age-matched controls, indicating that reductions in brain insulin are likely a result of age, not AD (Frölich et al., 1998).Two other groups report reductions in insulin mRNA in AD (Rivera et al., 2005; Steen et al., 2005), yet questions remain as to whether insulin is synthesized in the brain to an appreciable level be-cause there is evidence that a majority of brain insulin comes from the blood (Banks, 2004). Specifically, one study could not detect insulin mRNA in the cortex (Steen et al., 2005) but another did (Rivera et al., 2005), making this mRNA data difficult to interpret. Ultimately, a greater understand-

Figure 1. Canonical IR signaling cascade. Insulin binds to the insulin receptor (IR), a receptor tyrosine kinase, which autophosphorylates and activates a cascade of phosphorylation events. IRS1 is phosphorylated on a tyrosine residue to activate further signaling, which ultimately leads to the translocation of glucose transporter 4 (GLUT4) to the membrane and uptake of glucose for energy in peripheral tissues. Solid arrows represent activation upon insulin stim-ulation. Blocked arrows represent inhibition. Glycogen synthase kinase 3 (GSK3) is serine phosphorylated and inhibited in response to insulin stimulation. Dashed arrows represent downstream effectors that have been found to phosphorylate IRS1 on a serine residue (p(Ser)-IRS1), which is thought to lead to less activation of the signaling cascade through negative feedback (dashed blocked arrow). p(Ser)-IRS1 is a marker of insulin resistance.

1377JEM Vol. 213, No. 8

found no differences (Moloney et al., 2010; Liu et al., 2011; Ho et al., 2012; Talbot et al., 2012). The most convincing, consistent change in IS is a lower level of IR substrate 1 (IRS1) and higher p(Ser)-IRS1, a marker of insulin resistance, in AD brains (Steen et al., 2005; Moloney et al., 2010; Bom-fim et al., 2012; Talbot et al., 2012; Yarchoan et al., 2014). Higher p-JNK, which can lead to p(Ser)-IRS1, has also been found in AD brains (Bomfim et al., 2012; Talbot et al., 2012), suggesting some level of insulin resistance in AD. Questions still remain as to what are the physiological and pathological implications of increased markers such as p(Ser)-IRS1 and p-JNK and how that relates to AD pathology and brain function.

Parameter AD ↑ ↓ Study Details

Blood insulin ↑ Bucht et al., 1983; Fujisawa et al., 1991; Stolk et al., 1997; Craft et al., 1998; Ma et al., 2016

-Fasting or after glucose tolerance test -In women only (1 study) -Only in non-APOE4 and moderate/severe AD (1 study) -Meta-analysis of 11 studies: 5 report overall ↑, 1 ↑ in

women, 1 ↑ with advanced stage (Ma et al., 2016)CSF insulin ↑ Fujisawa et al., 1991 -Also found small increase with vascular dementia

↓ Craft et al., 1998; Gil-Bea et al., 2010

-Only in non-APOE4 and moderate/severe AD -No relationship to APOE or AD severity

No change Molina et al., 2002 -No relationship with AD severity or cognitionBrain insulin No change Frölich et al., 1998 -Comparing controls >65 y/o and AD patients

↓ Frölich et al., 1998; Rivera et al., 2005; Steen et al., 2005

-Comparing controls <65 y/o and AD patients -mRNA: in hippocampus and hypothalamus -mRNA: progressive reduction with Braak stage

Brain IR (total) ↓ Frölich et al., 1998; Rivera et al., 2005; Steen et al., 2005

-Comparing controls <65 y/o and AD patients -mRNA and protein -mRNA: progressive reduction with Braak stage

↑ Frölich et al., 1998 -Comparing controls >65 y/o and AD patientsNo change Moloney et al., 2010; Liu et al.,

2011; Ho et al., 2012; Talbot et al., 2012

-Potential changes in cellular distribution -Also no change in p-IR -Only reduced in patients with T2D and AD

Brain p-IR and activity ↓ Frölich et al., 1998; Rivera et al., 2005; Steen et al., 2005

-In hippocampus -Reduced insulin binding -TK activity reduced compared to all controls

Brain IRS1 (total) ↓ Steen et al., 2005; Moloney et al., 2010

-mRNA in 3 regions -Also reductions in IRS2

No change Liu et al., 2011; Talbot et al., 2012 -Also no change in IRS2 -Only reduced in patients with T2D and ADBrain p(Ser)-IRS1 ↑ Moloney et al., 2010 Talbot et

al., 2012 Bomfim et al., 2012 Yarchoan et al., 2014

-Regardless of APOE status and reduced ex vivo insulin stimulation -Highest in AD, but also elevated in some tauopathies

Brain AKT (total) ↓ Griffin et al., 2005; Liu et al., 2011 -Reduced in AD and in patients with T2D and ADNo change Steen et al., 2005; Talbot et al.,

2012Brain p-AKT ↑ Pei et al., 2003; Griffin et al.,

2005; Talbot et al., 2012; Yarchoan et al., 2014

-Associated with tangles

↓ Steen et al., 2005 -In hippocampusNo change Liu et al., 2011 -Only reduced in patients with T2D and AD

Brain GSK3 (total) ↓ Ho et al., 2012 -With advanced ADNo change Steen et al., 2005; Liu et al., 2011;

Talbot et al., 2012-Only reduced in patients with T2D or T2D and AD

Brain p(Ser)-GSK3 ↓ Steen et al., 2005 Griffin et al., 2005

-In hippocampus

No change Liu et al., 2011 -Only reduced in patients with T2D or T2D and ADBrain p-GSK3 ↑ Pei et al., 2003 -Associated with tanglesBrain p-JNK ↑ Bomfim et al., 2012; Talbot et

al., 2012Other IR signaling

molecules↓ Griffin et al., 2005; Liu et al., 2011;

Talbot et al., 2012-PDK1, p-PDK1 and p-PI3K -PIP3, PKC, p-mTOR, p-ERK2 -PTEN

Reported alterations in ins and brain IS in AD are categorized by the specific component measured, whether there have been reports of an increase, decrease (up and down arrows), or no change in individuals with AD, the studies that report this specific alteration, and important details. For blood and CSF insulin, AD diagnosis was based on clinical criteria. For postmortem analysis of brain insulin and IS components, AD was confirmed by clinical diagnosis and histological analyses. All reported changes were at the protein level unless mRNA is specified in the details. Overall, data from this table supports a higher level of blood insulin in individuals with AD and some degree of brain insulin resistance.

ing of insulin in the brain relative to the severity of AD and age-matched controls needs to be obtained in order to fully comprehend insulin’s function in healthy and diseased brains.

Brain IS. IRs are widely distributed throughout the brain, with relatively high concentrations in the olfactory bulb, hy-pothalamus, and hippocampus (Fernandez and Torres-Alemán, 2012). IRs are largely localized to neurons (Unger et al., 1989), although IR mRNA is present in glia and endothelial cells (Zhang et al., 2014). Although alterations in IR levels (Frölich et al., 1998; Steen et al., 2005) and its phosphoryla-tion (Steen et al., 2005) are reported in AD, other studies

Table 1. Insulin-related changes in AD

Insulin signaling in Alzheimer’s disease | Stanley et al.1378

Canonical IS posits that, upon activation of AKT by in-sulin, glycogen synthase kinase 3 (GSK3) is serine phosphor-ylated to reduce its activity. Active GSK3 phosphorylates tau among other substrates, suggesting that overly active GSK3 may exacerbate tau phosphorylation and, ultimately, its aggre-gation. In postmortem brain samples, there are reductions in p-AKT and p(Ser)-GSK3, suggesting increased GSK3 activity, which can lead to tau phosphorylation (Steen et al., 2005;Liu et al., 2011). In contrast, other groups report increases inp-AKT and p(Ser)-GSK3, even in the presence of elevatedphosphorylated tau (p-tau) and tangles (Pei et al., 2003; Grif-fin et al., 2005; Yarchoan et al., 2014), making this particularsignaling component difficult to interpret. Reductions in thelevel or phosphorylation of other IS molecules are reportedin AD brains (Liu et al., 2011; Talbot et al., 2012) and ex vivoactivity assays have shown that tyrosine kinase activity, insu-lin binding, and insulin stimulation are reduced in AD brains(Frölich et al., 1998; Rivera et al., 2005; Talbot et al., 2012).

Overall, there does appear to be some level of insulin re-sistance in the AD brain. However, this is not specific to insu-lin, as there are also reductions in both the levels and signaling of insulin-like growth factor (IGF) I and II (Steen et al., 2005; Moloney et al., 2010) and leptin signaling (Maioli et al., 2015). Although brain insulin and IGF resistance in the human AD

brain are seen as detrimental, mouse studies have paradoxi-cally shown that deleting IRs or IGF1 receptors in the brain is protective against amyloid plaque deposition (Freude et al., 2009; Stöhr et al., 2013) and improves survival (Freude et al., 2009), demonstrating that there is much left to elucidate about the physiological consequences of reduced insulin or IGF1 signaling in the brain.

Blood and cerebrospinal fluid (CSF) insulin. Studies reported alterations in blood insulin in AD as early as 1983 (Table 1). Fasted blood insulin or insulin in response to a glucose chal-lenge are higher in AD patients (Bucht et al., 1983; Fujisawa et al., 1991; Stolk et al., 1997; Craft et al., 1998; Ma et al., 2016). The transport of insulin from blood to brain and CSF is a receptor-mediated process. This transport is saturable within physiological levels and is affected by numerous vari-ables (Banks, 2004). The CSF/serum insulin ratio is subtly de-creased with age (Sartorius et al., 2015). Craft and colleagues also found the CSF/serum insulin ratio to be lower in AD, where higher blood and lower CSF insulin was more prom-inent with disease progression. Interestingly, they only found this change in AD patients without an APOE4 allele (Craft et al., 1998). Other groups who measured CSF insulin reported reductions but no difference with APOE4 (Gil-Bea et al.,

Figure 2. Connections between T2D and AD: cause or consequence? Big picture questions that need to be addressed to determine if insulin-related changes represent a cause or consequence of AD. In regards to the evidence that T2D increases the risk of AD, answering the questions in the top arrow will determine how and why T2D is a risk factor and the potentially causal role of insulin/IS. In regards to the idea that AD progression may lead to a dia-betic phenotype, answering the questions in the bottom arrow will determine if and how AD pathology may affect insulin homeostasis and the potential consequences of these changes on cognition.

1379JEM Vol. 213, No. 8

2010), no difference (Molina et al., 2002), or increases with higher CSF/plasma ratios (Fujisawa et al., 1991), making in-terpretations of this data difficult. While reports of CSF insulin are variable, current data suggests that AD patients are likely to experience higher blood insulin. When trying to determine if this hyperinsulinemia could be a cause or consequence of disease, it is important to consider when hyperinsulinemia occurs in relation to the development of AD. Table 1 demon-strates that blood insulin is higher in AD patients, and it may increase with disease progression (Craft et al., 1998), but other studies suggest that higher blood insulin, before diagnosis, may be present and influencing disease progression.

Changes in ins and IS: cause of AD?A longitudinal study found that fasting hyperinsulinemia, even without T2D, doubled the risk of developing AD (Luchsinger et al., 2004). A cross-sectional study found that in AD patients without an APOE4 allele, hyperinsulinemia was also associated with an increased risk of AD (Kuusisto et al., 1997) and higher insulin was associated with amyloid deposition, visualized by amyloid imaging on positron emis-sion tomography (PET) scans, before symptom onset (Wil-lette et al., 2015). Taken together, these studies suggest that high insulin could play a causative role in AD, although the AD population can be heterogeneous and it is possible that causal mechanisms differ across subgroups of patients. Since Aβ and tau deposition begin to occur ∼15 yr before symp-tom onset, these studies are difficult to interpret (Fig. 3). For example, high blood insulin before the onset of AD pathol-ogy could increase the risk of Aβ/tau deposition because

ins/IS is a primary instigator of disease. In contrast, amyloid accumulation which begins ∼15 yr before cognitive declinecould lead to brain insulin resistance, with hyperinsulinemia acting as a secondary indicator of underlying pathology in AD, and contribute to cognitive decline. To properly clarify whether insulin is a cause or consequence of disease, blood and CSF insulin should be tracked longitudinally, beginning before the onset of AD pathology, during AD pathology accumulation while individuals are still normal (preclinical AD), and then during the clinical stage of AD. To date, no such study has been reported. Recent studies have started to analyze the relationship between insulin resistance and AD biomarkers during the asymptomatic, preclinical stage in at-risk populations. In asymptomatic middle-age adults, insulin resistance was associated with higher CSF tau, p-tau (Starks et al., 2015) and Aβ42 (Hoscheidt et al., 2016). CSF insulinwas not measured, but baseline levels of blood glucose, insu-lin, and insulin resistance were no different between APOE4 carriers and noncarriers at this early stage of disease (Starks et al., 2015). Additional studies of this type, measuring CSF insulin in addition to traditional AD biomarkers, will help to determine the temporal relationship between insulin dys-regulation and AD progression.

Ins, IS, and Aβ. There are many studies suggesting that hyper-insulinemia may be directly influencing the risk of AD by modulating Aβ. In vitro studies demonstrate that high insulincan lead to higher extracellular Aβ by affecting clearancemechanisms and Aβ degrading enzymes. Both insulin and Aβare degraded by insulin degrading enzyme (IDE), and in the

Figure 3. Where and when do changes in ins and IS affect AD? Colored lines represent our current understanding of the pathological timeline in AD. It is currently unclear when and where changes in ins/IS occur in this timeline. If changes in ins/IS happen early, (1) they could initiate or potentiate amyloid accumulation to casu-ally influence AD. If ins/IS changes appear around the time of symptoms (4), this could be a consequence of years of patho-logical changes and may be directly related to cognitive decline. If changes in ins/IS occur in the presymptomatic period (2 and 3), they could be interacting with Aβ,tau, or metabolism to contribute to disease progression. Conversely, presymptomatic changes could be a result of tau or Aβaccumulation or metabolic perturbation. Additionally, it is possible that changes in ins/IS could simply push the symptomatic period to the left (earlier) without directly interacting with these pathologies.

Insulin signaling in Alzheimer’s disease | Stanley et al.1380

presence of high insulin, IDE will preferentially degrade insu-lin over Aβ (Qiu et al., 1998). In vivo experiments corrobo-rate in vitro findings where Aβ clearance is significantlyreduced in rats in the presence of high insulin (Shiiki et al., 2004). In addition to affecting Aβ clearance, in vitro workdemonstrates that high insulin increases extracellular Aβ con-centrations by increasing production through IS (Gasparini et al., 2001). Conversely, inhibition of PI3K leads to reduced Aβproduction (Stöhr et al., 2013). Crossing neuronal IR knock-out mice to an APP transgenic mouse abolishes IR signaling in the brain and leads to reduced Aβ levels and amyloid depo-sition, indicating that endogenous IS elevates Aβ in vivo(Stöhr et al., 2013). A recent study in mice found that injec-tion of supraphysiological levels of insulin increased brain IS and possibly Aβ (Sajan et al., 2016). In earlier work by Craftand colleagues, intravenous infusion of insulin in healthy older adults by hyperinsulinemic-euglycemic clamps im-proved performance on a declarative memory task, but in-creased CSF Aβ in ‘older’ participants (Watson et al., 2003).These clamps also increased plasma and CSF Aβ in anothercohort, and increased inflammatory markers in the CSF (Fishel et al., 2005). Collectively, data suggests that elevated ins can modulate Aβ, suggesting T2D could exacerbate ADpathogenesis over time via this mechanism.

A crucial, remaining question is whether physiolog-ical, peripheral hyperinsulinemia seen in AD or T2D raises brain insulin enough in vivo to actively compete with IDE, activate IS, or increase Aβ since insulin transport across theblood brain barrier (BBB) is saturable at normal physiologi-cal levels (Banks, 2004). A recent paper found that in several mouse models of T2D and in T2D monkeys, hyperinsulin-emia was associated with higher brain IS and higher Aβ levelsat baseline. Murine Aβ was reportedly measured by Westernblot in these experiments but the concentration is typically so low that this method may not be sufficient to detect small changes. Regardless, insulin injection did not further increase IS or Aβ, which could be due to transport or receptor satu-ration (Sajan et al., 2016). Overall, it still appears that hyper-insulinemia can positively regulate Aβ, though further studiesare needed to clarify this.

While high blood insulin may indicate higher brain ins/IS, it is also possible that lower brain ins/IS leads to higher blood insulin as an attempt to compensate for the reduction. A longitudinal study found that both the highest and lowest quartiles for fasting insulin were associated with the devel-opment of dementia (Peila et al., 2004). One mouse model of AD was found to have reduced brain insulin and some changes in IS that preceded the deposition of Aβ (Chua etal., 2012). These results suggest that low brain ins/IS may also affect Aβ. Unfortunately, they did not measure blood or CSFinsulin to know how these measures fluctuate overtime in relation to amyloid deposition and brain ins/IS.

GSK3 can increase Aβ levels in vitro and inhibition ofGSK3 by lithium can reduce APP processing and Aβ levels(Phiel et al., 2003). As shown in Fig.  1, insulin also inhib-

its GSK3 through canonical signaling which would suggest that adequate ins/IS may contain Aβ levels. So far, this hasnot been validated experimentally. Lowering blood insulin by destroying insulin-producing cells in a mouse model of AD reduces brain ins/IS and results in elevated Aβ levels (Wanget al., 2010). While the authors conclude that insulin defi-ciency raised Aβ, these mice also have extreme hyperglycemiawhich we and others have recently shown can independently increase extracellular Aβ (Macauley et al., 2015; Chao et al.,2016). Moreover, decreasing insulin production would lead to decreased IDE levels, which could affect Aβ since IDE isan Aβ degrading enzyme.

While hyperinisulinemia can initially increase brain insulin and positively regulate Aβ acutely, chronichyperinsulinemia can down-regulate transport leading to lower blood :brain insulin (Banks et al., 1997). This decrease in BBB transport may also increase Aβ, although there is nodirect experimental evidence supporting this claim. Since it is unclear whether chronic hyperinsulinemia (described in Table 1) leads to increased brain insulin or to decreased insulin transport, additional experimental confirmation is required to identify a causal relationship with Aβ and AD.

Ins, IS, and tau phosphorylation. GSK3 activity is often con-nected to the phosphorylation of tau. Neuronal IR knockout mice have higher p-tau, presumably due to more active GSK3 (Schubert et al., 2004). Interestingly, these IR knockout mice do not have any memory deficits despite evidence that IS is connected to cognition and p-tau is linked with memory deficits in mice (Schindowski et al., 2006). Peripheral injec-tion of supraphysiological insulin is capable of elevating IS in the mouse brain and increases p-tau even though GSK3 is serine phosphorylated by p-AKT to reduce activity and pre-sumably p-tau (Freude et al., 2005). Hyperinsulinemia also increases p-tau in aged, wild-type mice (Becker et al., 2012). In vivo studies from mice agree with the complex findings from AD brains (Table  1) which found GSK3 to be both more active and less active while p-tau was high. Overall, these data suggest, again, that both high ins/IS (from hyperin-sulinemia) and low IS (from insulin resistance) may put the brain at risk to exacerbate AD.

One study found that chronically raising blood insu-lin in mice via a high fat diet had no effect on p-tau in the brain, but they also reported no change in brain IS (Becker et al., 2012). When injected peripherally with supraphysio-logical insulin, p-tau was unchanged in mice with chronic hyperinsulinemia (Becker et al., 2012). Brain IS in response to the peripheral injection was not measured, but this likely demonstrates that insulin transport is reduced with chronic hyperinsulinemia. Additional evidence from a study in mice and monkeys demonstrates that chronic hyperinsulinemia re-sulted in higher brain IS and higher p-tau (Sajan et al., 2016). They also found no further increase in IS with insulin injec-tion, confirming that chronic hyperinsulinemia is most likely saturating BBB transport and/or IS. These studies highlight

1381

Aβ and brain insulin resistance. There is substantial evidence from in vitro experiments that Aβ may directly contribute toneuronal insulin resistance. Aβ can competitively inhibit thebinding of insulin to the IR (Xie et al., 2002). Aβ oligomers(AβO) are thought to be the more toxic species of Aβ, andsynthetic AβOs bind to and internalize IRs causing an in-crease in neuronal p(Ser)-IRS1 and p-JNK, markers of insulin resistance (Zhao et al., 2008; Bomfim et al., 2012). In vivo experiments found higher p(Ser)-IRS1 and p-JNK in human AD brains (Table 1) and aged APP transgenic mice. They also found that intracerebroventricular injection of AβO’s intomonkeys elevated p(Ser)-IRS1 and p-JNK, demonstrating that these AβO’s can cause resistance in vivo (Bomfim et al.,2012). AβO’s increase p-tau, which may result from activationof GSK3 as a result of reductions in IS (Ma et al., 2009). Taken together, these data suggest that Aβ has the ability to induceinsulin resistance in vitro and in vivo using several dif-ferent model organisms.

Insulin treatment for brain insulin resistance. Craft et al. (1996, 1999, 2003, 2012) continue to demonstrate that insulin treatment, either by intravenous infusion or intranasal deliv-ery, can modestly enhance performance on memory tasks in healthy adults and patients with AD or mild cognitive impair-ment (MCI) at specific doses (Watson et al., 2003; Reger et al., 2008; Claxton et al., 2015). Gender and APOE4 modulate the beneficial effects of insulin (Craft et al., 2000, 2003; Reger et al., 2008; Claxton et al., 2013). In mice, intranasal adminis-tration of insulin increases insulin in the cortex and hippo-campus to some extent (Salameh et al., 2015), but whether or not it increases IS and confers protection against neuronal damage is still unknown. In vitro studies show that synthetic AβOs co-cultured with neurons can bind to synapses and re-duce dendritic spines, an effect that is rescued when treated with high insulin and IS is activated (De Felice et al., 2009). Increasing hippocampal insulin and IS in rats also enhances memory. However, rats with T2D induced by a high fat diet did not show improvement with insulin treatment (McNay et al., 2010), suggesting that once brain insulin resistance has developed, insulin treatment may not be sufficient to over-come resistance at a cellular level. In a mouse model of AD, high fat diet led to insulin resistance and exacerbated amyloid pathology and memory impairment. A single, supraphysio-logical injection of insulin was found to improve IS in pe-ripheral tissues and reduce soluble Aβ in the brain, but brainIS was not measured (Vandal et al., 2014). If insulin resistance is severe and causing brain insulin to be too low, increasing insulin under these conditions may help mitigate Aβ levels;however, this hypothesis was not directly tested. If brain IS is within normal range, insulin injection can increase IS and possibly Aβ levels (Sajan et al., 2016). Additionally, raisingbrain insulin by intranasal delivery appears to elevate brain insulin high enough that it may increase Aβ or p-tau, al-though this has not been addressed in studies to date. Prelim-inary reports using intranasal insulin in mouse models suggests

the complexity between acute and chronic changes in blood insulin and insulin sensitivity in the brain that need to be considered when trying to understand how hyperinsulinemia and brain insulin resistance relate to each other and AD.

Changes in ins and IS: consequence of AD?One AD population study found that up to 80% of AD pa-tients had either T2D or insulin resistance, suggesting that AD may lead to a diabetic phenotype (Janson et al., 2004). Unfortunately, there are no longitudinal studies showing whether or not diabetic phenotypes occur after AD onset or precede AD diagnosis. Tissue analysis (Table 1) supports that brain insulin resistance worsens with advancing AD, but it is unclear how early in disease progression insulin resistance occurs (Fig. 3). If reduced brain IS is associated with cog-nitive decline, one would expect changes to occur around the time that symptoms start to appear, which may be after blood insulin is already high. In T2D patients, insulin resis-tance is also associated with brain hypometabolism (Baker et al., 2011). In AD, hypometabolism occurs before symptom onset and is likely related to synaptic dysfunction and neu-ronal loss, and worsens with disease progression (Sperling et al., 2011). It is unclear how hyperinsulinemia, insulin resis-tance, brain hypometabolism, and cognitive decline are tem-porally related in AD (Fig. 3). A recent study used plasma exosomes from neural sources to measure p(Ser)-IRS1, a marker of insulin resistance, longitudinally in individuals with AD, T2D, or healthy controls (Kapogiannis et al., 2015). In these exosomes, p(Ser)-IRS1 was found to be higher in AD and T2D compared with controls, and was elevated up to 1–10 yr before AD diagnosis. However, there was no association of exosome p(Ser)-IRS1 with AD severity or insulin resistance, which is contrary to results from postmor-tem brain tissue (Table 1). Determining which comes first, hyperinsulinemia or brain insulin resistance, would be the key to understanding which is a cause or consequence of AD and how they relate to AD pathogenesis.

In T2D, hyperinsulinemia is capable of leading to in-sulin resistance through negative feedback at the IR (Fig. 1). Thus, it is possible that the hyperinsulinemia seen in AD pa-tients may lead to brain insulin resistance or reduced insulin transport, which can modulate both Aβ and p-tau to con-tribute to AD in a causal way. Although there is currently no longitudinal data demonstrating that hyperinsulinemia precedes the onset of AD pathology, there is mechanistic ev-idence that it is possible for hyperinsulinemia to drive both insulin resistance and AD. Alternatively, insulin resistance in T2D can also be initiated by other processes, such as cellular stress and inflammation, which a ctivate s ignaling molecules like p-JNK to increase p(Ser)-IRS1, causing hyperinsulin-emia as a compensatory mechanism (Draznin, 2006). More-over, oxidative stress and neuroinflammation are key features of the AD brain, which may promote brain insulin resistance, but most experiments have yet to explore these mecha-nisms in the context of AD.

JEM Vol. 213, No. 8

Insulin signaling in Alzheimer’s disease | Stanley et al.1382

that this treatment may have some positive effects on memory and pathology (Chen et al., 2014; Zhang et al., 2016). Exten-din-4, an antidiabetic medication thought to enhance IS by activating similar pathways, was able to prevent the develop-ment of insulin resistance (p(Ser)-IRS1) in neuronal cultures and APP transgenic mice (Bomfim et al., 2012). Insulin sensi-tizers that are used to treat peripheral insulin resistance, rather than insulin itself, may also be beneficial in AD. However, it is unclear if enhancing IS through insulin sensitizers can also modulate Aβ or p-tau.

The current hypothesis is that intranasal insulin is di-rectly influencing cognition by acting on neuronal IRs to overcome resistance, but this has not been shown directly. Recent evidence has shown that intranasal insulin may be working through indirect pathways to influence cognition. For example, there have been reports of intranasal insulin in-creasing regional cerebral blood flow and cognition in T2D patients (Novak et al., 2014). Insulin is thought to act through IRs on endothelial cells in the brain, but a recent study could not detect significant IR peptide in this cell type, despite IR mRNA being abundant (Zuchero et al., 2016). Determining the mechanism of intranasal insulin’s cognitive enhancement, and whether or not neuronal IS is necessary, will ultimately allow a more targeted therapeutic approach that may not have the potential side effect of raising Aβ or p-tau.

Concluding remarksThere is substantial experimental evidence that hyperin-sulinemia and brain insulin resistance, seen in LOAD pa-tients, is capable of increasing Aβ and p-tau to initiate orexacerbate the pathological cascade associated with AD. If patients experience high blood insulin before early AD alterations, hyperinsulinemia could causally contribute to both AD pathology and insulin resistance. However, there is not significant data to confirm this hypothesis in humans. Alternatively, initial Aβ accumulation can lead to neuronalinsulin resistance and secondary hyperinsulinemia, which further exacerbates AD progression. Both longitudinal and biomarker studies need to be performed to properly under-stand how high insulin in the blood, brain, or CSF, com-bined with insulin resistance, relate to the progression of AD. Until that time, we should assume that the alterations in ins and IS represent both a cause and consequence of disease and need to be closely monitored as insulin therapy is investigated for AD.

ACkNoWLEDGMENTSThis work was supported by a National Science Foundation Graduate Research Fel-lowship (DGE-1143954; M. Stanley), National Institute on Ageing grant K01 AG050719 (S.L. Macauley), National Institute of Neurological Disorders and Stroke grants F32 NS080320 (S.L. Macauley) and P01 NS080675 (D.M. Holtzman), New Vi-sion Award through Donors Cure Foundation (S.L. Macauley), and the JPB Founda-tion (D.M. Holtzman).

D.M. Holtzman co-founded and is on the scientific advisory board of C2N Diag-nostics. D.M. Holtzman consults for Genentech, AbbVie, Eli Lilly, Neurophage, and

Denali. SML consults for Denali. The authors declare no additional competing fi-nancial interests.

Submitted: 5 April 2016

Accepted: 20 June 2016

REFERENCESAllen, K.V., B.M. Frier, and M.W. Strachan. 2004. The relationship between

type 2 diabetes and cognitive dysfunction: longitudinal studies and their methodological limitations. Eur. J. Pharmacol. 490:169–175. http ://dx .doi .org /10 .1016 /j .ejphar .2004 .02 .054

Baker, L.D., D.J. Cross, S. Minoshima, D. Belongia, G.S. Watson, and S. Craft. 2011. Insulin resistance and Alzheimer-like reductions in regional cerebral glucose metabolism for cognitively normal adults with prediabetes or early type 2 diabetes. Arch. Neurol. 68:51–57. http ://dx .doi .org /10 .1001 /archneurol .2010 .225

Banks, W.A. 2004. The source of cerebral insulin. Eur. J. Pharmacol. 490:5–12. http ://dx .doi .org /10 .1016 /j .ejphar .2004 .02 .040

Banks, W.A., J.B. Jaspan, and A.J. Kastin. 1997. Selective, physiological transport of insulin across the blood-brain barrier: novel demonstration by species-specific radioimmunoassays. Peptides. 18:1257–1262. http ://dx .doi .org /10 .1016 /S0196 -9781(97)00198 -8

Becker, K., S. Freude, J. Zemva, O. Stöhr, W. Krone, and M. Schubert. 2012. Chronic peripheral hyperinsulinemia has no substantial influence on tau phosphorylation in vivo. Neurosci. Lett. 516:306–310. http ://dx .doi .org /10 .1016 /j .neulet .2012 .04 .022

Biessels, G.J., and Y.D. Reijmer. 2014. Brain changes underlying cognitive dysfunction in diabetes: what can we learn from MRI? Diabetes. 63:2244–2252. http ://dx .doi .org /10 .2337 /db14 -0348

Bomfim, T.R., L. Forny-Germano, L.B. Sathler, J. Brito-Moreira, J.C. Houzel, H. Decker, M.A. Silverman, H. Kazi, H.M. Melo, P.L. McClean, et al.2012. An anti-diabetes agent protects the mouse brain from defectiveinsulin signaling caused by Alzheimer’s disease-associated Aβ oligomers. J. Clin. Invest. 122:1339–1353. http ://dx .doi .org /10 .1172 /JCI57256

Brief, D.J., and J.D. Davis. 1984. Reduction of food intake and body weight by chronic intraventricular insulin infusion. Brain Res. Bull. 12:571–575. http ://dx .doi .org /10 .1016 /0361 -9230(84)90174 -6

Bucht, G., R. Adolfsson, F. Lithner, and B. Winblad. 1983. Changes in blood glucose and insulin secretion in patients with senile dementia of Alzheimer type. Acta Med. Scand. 213:387–392. http ://dx .doi .org /10 .1111 /j .0954 -6820 .1983 .tb03756 .x

Chao, A.C., T.C. Lee, S.H. Juo, and D.I. Yang. 2016. Hyperglycemia Increases the Production of Amyloid β-peptide leading to decreased endothelialtight junction. CNS Neurosci. Ther. 22:291–297. http ://dx .doi .org /10 .1111 /cns .12503

Chen, Y., Y. Zhao, C.L. Dai, Z. Liang, X. Run, K. Iqbal, F. Liu, and C.X. Gong. 2014. Intranasal insulin restores insulin signaling, increases synaptic proteins, and reduces Aβ level and microglia activation in the brains of3xTg-AD mice. Exp. Neurol. 261:610–619. http ://dx .doi .org /10 .1016 /j .expneurol .2014 .06 .004

Chua, L.M., M.L. Lim, P.R. Chong, Z.P. Hu, N.S. Cheung, and B.S. Wong. 2012. Impaired neuronal insulin signaling precedes Aβ42 accumulationin female AβPPsw/PS1ΔE9 mice. J. Alzheimers Dis. 29:783–791.

Claxton, A., L.D. Baker, C.W. Wilkinson, E.H. Trittschuh, D. Chapman, G.S. Watson, B. Cholerton, S.R. Plymate, M. Arbuckle, and S. Craft. 2013. Sex and ApoE genotype differences in treatment response to two doses of intranasal insulin in adults with mild cognitive impairment or Alzheimer’s disease. J. Alzheimers Dis. 35:789–797. http ://dx .doi .org /10 .3233 /JAD -122308

Claxton, A., L.D. Baker, A. Hanson, E.H. Trittschuh, B. Cholerton, A. Morgan, M. Callaghan, M. Arbuckle, C. Behl, and S. Craft. 2015. Long actingintranasal insulin detemir improves cognition for adults with mild

1383JEM Vol. 213, No. 8

cognitive impairment or early-stage Alzheimer’s disease dementia. J. Alzheimers Dis. 45:1269–1270. http ://dx .doi .org /10 .3233 /JAD -15900

Craft, S., J. Newcomer, S. Kanne, S. Dagogo-Jack, P. Cryer, Y. Sheline, J. Luby, A. Dagogo-Jack, and A. Alderson. 1996. Memory improvement follow-ing induced hyperinsulinemia in Alzheimer’s disease. Neurobiol. Aging.17:123–130.

Craft, S., E. Peskind, M.W. Schwartz, G.D. Schellenberg, M. Raskind, and D. Porte Jr. 1998. Cerebrospinal fluid and plasma insulin levels in Alzheimer’s disease: relationship to severity of dementia and apolipoprotein E genotype. Neurology. 50:164–168. http ://dx .doi .org /10 .1212 /WNL .50 .1 .164

Craft, S., S. Asthana, J.W. Newcomer, C.W. Wilkinson, I.T. Matos, L.D. Baker, M. Cherrier, C. Lofgreen, S. Latendresse, A. Petrova, et al. 1999. Enhancement of memory in Alzheimer disease with insulin and somatostatin, but not glucose. Arch. Gen. Psychiatry. 56:1135–1140. http ://dx .doi .org /10 .1001 /archpsyc .56 .12 .1135

Craft, S., S. Asthana, G. Schellenberg, L. Baker, M. Cherrier, A.A. Boyt, R.N. Martins, M. Raskind, E. Peskind, and S. Plymate. 2000. Insulin effects on glucose metabolism, memory, and plasma amyloid precursor protein in Alzheimer’s disease differ according to apolipoprotein-E genotype. Ann. N. Y. Acad. Sci. 903:222–228. http ://dx .doi .org /10 .1111 /j .1749 -6632 .2000 .tb06371 .x

Craft, S., S. Asthana, D.G. Cook, L.D. Baker, M. Cherrier, K. Purganan, C. Wait, A. Petrova, S. Latendresse, G.S. Watson, et al. 2003. Insulin dose-response effects on memory and plasma amyloid precursor protein in Alzheimer’s disease: interactions with apolipoprotein E genotype. Psychoneuroendocrinology. 28:809–822. http ://dx .doi .org /10 .1016 /S0306 -4530(02)00087 -2

Craft, S., L.D. Baker, T.J. Montine, S. Minoshima, G.S. Watson, A. Claxton, M. Arbuckle, M. Callaghan, E. Tsai, S.R. Plymate, et al. 2012. Intranasalinsulin therapy for Alzheimer disease and amnestic mild cognitiveimpairment: a pilot clinical trial. Arch. Neurol. 69:29–38. http ://dx .doi .org /10 .1001 /archneurol .2011 .233

Dankner, R., A. Chetrit, M.H. Shanik, I. Raz, and J. Roth. 2009. Basal-state hyperinsulinemia in healthy normoglycemic adults is predictive of type 2 diabetes over a 24-year follow-up: a preliminary report. Diabetes Care. 32:1464–1466. http ://dx .doi .org /10 .2337 /dc09 -0153

De Felice, F.G., M.N. Vieira, T.R. Bomfim, H. Decker, P.T. Velasco, M.P. Lambert, K.L. Viola, W.Q. Zhao, S.T. Ferreira, and W.L. Klein. 2009. Protection of synapses against Alzheimer’s-linked toxins: insulin signaling prevents the pathogenic binding of Aβ oligomers. Proc. Natl. Acad. Sci.USA. 106:1971–1976. http ://dx .doi .org /10 .1073 /pnas .0809158106

Dimitriadis, G., P. Mitrou, V. Lambadiari, E. Maratou, and S.A. Raptis. 2011. Insulin effects in muscle and adipose tissue. Diabetes Res. Clin. Pract. 93:S52–S59. http ://dx .doi .org /10 .1016 /S0168 -8227(11)70014 -6

Draznin, B. 2006. Molecular mechanisms of insulin resistance: serine phosphorylation of insulin receptor substrate-1 and increased expression of p85alpha: the two sides of a coin. Diabetes. 55:2392–2397. http ://dx .doi .org /10 .2337 /db06 -0391

Fernandez, A.M., and I. Torres-Alemán. 2012. The many faces of insulin-like peptide signalling in the brain. Nat. Rev. Neurosci. 13:225–239. http ://dx .doi .org /10 .1038 /nrn3209

Fishel, M.A., G.S. Watson, T.J. Montine, Q. Wang, P.S. Green, J.J. Kulstad, D.G. Cook, E.R. Peskind, L.D. Baker, D. Goldgaber, et al. 2005.Hyperinsulinemia provokes synchronous increases in centralinflammation and beta-amyloid in normal adults. Arch. Neurol. 62:1539–1544. http ://dx .doi .org /10 .1001 /archneur .62 .10 .noc50112

Freude, S., L. Plum, J. Schnitker, U. Leeser, M. Udelhoven, W. Krone, J.C. Bruning, and M. Schubert. 2005. Peripheral hyperinsulinemia promotes tau phosphorylation in vivo. Diabetes. 54:3343–3348. http ://dx .doi .org /10 .2337 /diabetes .54 .12 .3343

Freude, S., M.M. Hettich, C. Schumann, O. Stöhr, L. Koch, C. Köhler, M. Udelhoven, U. Leeser, M. Müller, N. Kubota, et al. 2009. Neuronal IGF-1

resistance reduces Abeta accumulation and protects against premature death in a model of Alzheimer’s disease. FAS EB J. 23:3315–3324. http ://dx .doi .org /10 .1096 /fj .09 -132043

Frölich, L., D. Blum-Degen, H.G. Bernstein, S. Engelsberger, J. Humrich, S. Laufer, D. Muschner, A. Thalheimer, A. Türk, S. Hoyer, et al. 1998. Brain insulin and insulin receptors in aging and sporadic Alzheimer’s disease. J Neural Transm (Vienna). 105:423–438. http ://dx .doi .org /10 .1007 /s007020050068

Fujisawa, Y., K. Sasaki, and K. Akiyama. 1991. Increased insulin levels after OGTT load in peripheral blood and cerebrospinal fluid of patients with dementia of Alzheimer type. Biol. Psychiatry. 30:1219–1228. http ://dx .doi .org /10 .1016 /0006 -3223(91)90158 -I

Gasparini, L., G.K. Gouras, R. Wang, R.S. Gross, M.F. Beal, P. Greengard, and H. Xu. 2001. Stimulation of beta-amyloid precursor protein traffickingby insulin reduces intraneuronal β-amyloid and requires mitogen-acti-vated protein kinase signaling. J. Neurosci. 21:2561–2570.

Gil-Bea, F.J., M. Solas, A. Solomon, C. Mugueta, B. Winblad, M. Kivipelto, M.J. Ramirez, and A. Cedazo-Mínguez. 2010. Insulin levels are decreased in the cerebrospinal fluid of women with prodomal Alzheimer’s disease. J. Alzheimers Dis. 22:405–413.

Griffin, R.J., A. Moloney, M. Kelliher, J.A. Johnston, R. Ravid, P. Dockery, R. O’Connor, and C. O’Neill. 2005. Activation of Akt/PKB, increasedphosphorylation of Akt substrates and loss and altered distribution of Akt and PTEN are features of Alzheimer’s disease pathology. J. Neurochem.93:105–117. http ://dx .doi .org /10 .1111 /j .1471 -4159 .2004 .02949 .x

Ho, L., S. Yemul, L. Knable, P. Katsel, R. Zhao, V. Haroutunian, and G.M. Pasinetti. 2012. Insulin receptor expression and activity in the brains of nondiabetic sporadic Alzheimer’s disease cases. Int. J. Alzheimers Dis. 2012:321280. http ://dx .doi .org /1155 /2012 /321280

Holtzman, D.M., J.C. Morris, and A.M. Goate. 2011. Alzheimer’s disease: the challenge of the second century. Sci. Transl. Med. 3:77sr1. http ://dx .doi .org /10 .1126 /scitranslmed .3002369

Hoscheidt, S.M., E.J. Starks, J.M. Oh, H. Zetterberg, K. Blennow, R.A. Krause, C.E. Gleason, L. Puglielli, C.S. Atwood, C.M. Carlsson, et al. 2016.Insulin resistance is associated with increased levels of cerebrospinal fluid biomarkers of Alzheimer's disease and reduced memory function in at-risk healthy middle-aged adults. J. Alzheimers Dis. 12:1373–1383. http ://dx .doi .org /10 .3233 /JAD -160110

Janson, J., T. Laedtke, J.E. Parisi, P. O’Brien, R.C. Petersen, and P.C. Butler. 2004. Increased risk of type 2 diabetes in Alzheimer disease. Diabetes. 53:474–481. http ://dx .doi .org /10 .2337 /diabetes .53 .2 .474

Kapogiannis, D., A. Boxer, J.B. Schwartz, E.L. Abner, A. Biragyn, U. Masharani, L. Frassetto, R.C. Petersen, B.L. Miller, and E.J. Goetzl. 2015. Dysfunctionally phosphorylated type 1 insulin receptor substrate in neural-derived blood exosomes of preclinical Alzheimer’s disease. FAS EB J. 29:589–596. http ://dx .doi .org /10 .1096 /fj .14 -262048

Kuusisto, J., K. Koivisto, L. Mykkänen, E.L. Helkala, M. Vanhanen, T. Hänninen, K. Kervinen, Y.A. Kesäniemi, P.J. Riekkinen, and M. Laakso. 1997. Association between features of the insulin resistance syndrome and Alzheimer’s disease independently of apolipoprotein E4 phenotype: cross sectional population based study. BMJ. 315:1045–1049. http ://dx .doi .org /10 .1136 /bmj .315 .7115 .1045

Last, D., D.C. Alsop, A.M. Abduljalil, R.P. Marquis, C. de Bazelaire, K. Hu, J. Cavallerano, and V. Novak. 2007. Global and regional effects of type2 diabetes on brain tissue volumes and cerebral vasoreactivity. DiabetesCare. 30:1193–1199. http ://dx .doi .org /10 .2337 /dc06 -2052

Liu, Y., F. Liu, I. Grundke-Iqbal, K. Iqbal, and C.X. Gong. 2011. Deficient brain insulin signalling pathway in Alzheimer’s disease and diabetes. J. Pathol. 225:54–62. http ://dx .doi .org /10 .1002 /path .2912

Luchsinger, J.A., M.X. Tang, S. Shea, and R. Mayeux. 2004. Hyperinsulinemia and risk of Alzheimer disease. Neurology. 63:1187–1192. http ://dx .doi .org /10 .1212 /01 .WNL .0000140292 .04932 .87

Insulin signaling in Alzheimer’s disease | Stanley et al.1384

Ma, J., W. Zhang, H.F. Wang, Z.X. Wang, T. Jiang, M.S. Tan, J.T. Yu, and L. Tan. 2016. Peripheral blood adipokines and insulin levels inpatients with Alzheimer’s disease: a replication study and meta-analysis. Curr. Alzheimer Res. 13:223–233. http ://dx .doi .org /10 .2174 /156720501303160217111434

Ma, Q.L., F. Yang, E.R. Rosario, O.J. Ubeda, W. Beech, D.J. Gant, P.P. Chen, B. Hudspeth, C. Chen, Y. Zhao, et al. 2009. Beta-amyloid oligomers induce phosphorylation of tau and inactivation of insulin receptor substrate via c-Jun N-terminal kinase signaling: suppression by omega-3 fatty acidsand curcumin. J. Neurosci. 29:9078–9089. http ://dx .doi .org /10 .1523 /JNE URO SCI .1071 -09 .2009

Macauley, S.L., M. Stanley, E.E. Caesar, S.A. Yamada, M.E. Raichle, R. Perez, T.E. Mahan, C.L. Sutphen, and D.M. Holtzman. 2015. Hyperglycemia modulates extracellular amyloid-β concentrations and neuronal activityin vivo. J. Clin. Invest. 125:2463–2467. http ://dx .doi .org /10 .1172 /JCI79742

Maioli, S., M. Lodeiro, P. Merino-Serrais, F. Falahati, W. Khan, E. Puerta, A. Codita, R. Rimondini, M.J. Ramirez, A. Simmons, et al. Alzheimer’s Disease Neuroimaging Initiative. 2015. Alterations in brain leptin signalling in spite of unchanged CSF leptin levels in Alzheimer’s disease. Aging Cell. 14:122–129. http ://dx .doi .org /10 .1111 /acel .12281

McNay, E.C., and A.K. Recknagel. 2011. Brain insulin signaling: a key component of cognitive processes and a potential basis for cognitive impairment in type 2 diabetes. Neurobiol. Learn. Mem. 96:432–442. http ://dx .doi .org /10 .1016 /j .nlm .2011 .08 .005

McNay, E.C., C.T. Ong, R.J. McCrimmon, J. Cresswell, J.S. Bogan, and R.S. Sherwin. 2010. Hippocampal memory processes are modulated by insulin and high-fat-induced insulin resistance. Neurobiol. Learn. Mem. 93:546–553. http ://dx .doi .org /10 .1016 /j .nlm .2010 .02 .002

Molina, J.A., F.J. Jiménez-Jiménez, C. Vargas, P. Gómez, F. de Bustos, C. Gómez-Escalonilla, M. Zurdo, A. Tallón, A. Martínez-Salio, J. Porta-Etessam, et al. 2002. Cerebrospinal fluid levels of insulin in patients with Alzheimer’s disease. Acta Neurol. Scand. 106:347–350. http ://dx .doi .org /10 .1034 /j .1600 -0404 .2002 .01326 .x

Moloney, A.M., R.J. Griffin, S. Timmons, R. O’Connor, R. Ravid, and C. O’Neill. 2010. Defects in IGF-1 receptor, insulin receptor and IRS-1/2 in Alzheimer’s disease indicate possible resistance to IGF-1 and insulin signalling. Neurobiol. Aging. 31:224–243. http ://dx .doi .org /10 .1016 /j .neurobiolaging .2008 .04 .002

Musiek, E.S., and D.M. Holtzman. 2015. Three dimensions of the amyloid hypothesis: time, space and ‘wingmen’. Nat. Neurosci. 18:800–806. http ://dx .doi .org /10 .1038 /nn .4018

Novak, V., W. Milberg, Y. Hao, M. Munshi, P. Novak, A. Galica, B. Manor, P. Roberson, S. Craft, and A. Abduljalil. 2014. Enhancement ofvasoreactivity and cognition by intranasal insulin in type 2 diabetes.Diabetes Care. 37:751–759. http ://dx .doi .org /10 .2337 /dc13 -1672

Pei, J.J., S. Khatoon, W.L. An, M. Nordlinder, T. Tanaka, H. Braak, I. Tsujio, M. Takeda, I. Alafuzoff, B. Winblad, et al. 2003. Role of protein kinase Bin Alzheimer’s neurofibrillary pathology. Acta Neuropathol. 105:381–392.

Peila, R., B.L. Rodriguez, L.R. White, and L.J. Launer. 2004. Fasting insulin and incident dementia in an elderly population of Japanese-American men. Neurology. 63:228–233. http ://dx .doi .org /10 .1212 /01 .WNL .0000129989 .28404 .9B

Phiel, C.J., C.A. Wilson, V.M. Lee, and P.S. Klein. 2003. GSK-3alpha regulates production of Alzheimer’s disease amyloid-βpeptides. Nature. 423:435–439. http ://dx .doi .org /10 .1038 /nature01640

Qiu, W.Q., D.M. Walsh, Z. Ye, K. Vekrellis, J. Zhang, M.B. Podlisny, M.R. Rosner, A. Safavi, L.B. Hersh, and D.J. Selkoe. 1998. Insulin-degrading enzyme regulates extracellular levels of amyloid beta-protein by degradation. J. Biol. Chem. 273:32730–32738. http ://dx .doi .org /10 .1074 /jbc .273 .49 .32730

Reger, M.A., G.S. Watson, P.S. Green, C.W. Wilkinson, L.D. Baker, B. Cholerton, M.A. Fishel, S.R. Plymate, J.C. Breitner, W. DeGroodt, et al.

2008. Intranasal insulin improves cognition and modulates β-amyloid inearly AD. Neurology. 70:440–448. http ://dx .doi .org /10 .1212 /01 .WNL .0000265401 .62434 .36

Rivera, E.J., A. Goldin, N. Fulmer, R. Tavares, J.R. Wands, and S.M. de la Monte. 2005. Insulin and insulin-like growth factor expression and func-tion deteriorate with progression of Alzheimer’s disease: link to brain reductions in acetylcholine. J. Alzheimers Dis. 8:247–268.

Roberts, R.O., D.S. Knopman, R.H. Cha, M.M. Mielke, V.S. Pankratz, B.F. Boeve, K. Kantarci, Y.E. Geda, C.R. Jack Jr., R.C. Petersen, and V.J. Lowe. 2014. Diabetes and elevated hemoglobin A1c levels are associated with brain hypometabolism but not amyloid accumulation. J. Nucl. Med. 55:759–764. http ://dx .doi .org /10 .2967 /jnumed .113 .132647

Sajan, M., B. Hansen, R. Ivey III, J. Sajan, C. Ari, S. Song, U. Braun, M. Leitges, M. Farese-Higgs, and R.V. Farese. 2016. Brain insulin signaling is increased in insulin-resistant states and decreases in FOXOs and PGC-1α and increases in Aβ1-40/42 and phospho-tau may abetAlzheimer development. Diabetes. 65:1892–1903. http ://dx .doi .org /10 .2337 /db15 -1428

Salameh, T.S., K.M. Bullock, I.A. Hujoel, M.L. Niehoff, T. Wolden-Hanson, J. Kim, J.E. Morley, S.A. Farr, and W.A. Banks. 2015. Central nervoussystem delivery of intranasal insulin: mechanisms of uptake and effectson cognition. J. Alzheimers Dis. 47:715–728. http ://dx .doi .org /10 .3233 /JAD -150307

Sartorius, T., A. Peter, M. Heni, W. Maetzler, A. Fritsche, H.U. Häring, and A.M. Hennige. 2015. The brain response to peripheral insulin declineswith age: a contribution of the blood-brain barrier? PLoS One.10:e0126804. http ://dx .doi .org /10 .1371 /journal .pone .0126804

Schindowski, K., A. Bretteville, K. Leroy, S. Bégard, J.P. Brion, M. Hamdane, and L. Buée. 2006. Alzheimer’s disease-like tau neuropathology leads to memory deficits and loss of functional synapses in a novel mutated tau transgenic mouse without any motor deficits. Am. J. Pathol. 169:599–616. http ://dx .doi .org /10 .2353 /ajpath .2006 .060002

Schubert, M., D. Gautam, D. Surjo, K. Ueki, S. Baudler, D. Schubert, T. Kondo, J. Alber, N. Galldiks, E. Küstermann, et al. 2004. Role for neuronal insulin resistance in neurodegenerative diseases. Proc. Natl. Acad. Sci. USA.101:3100–3105. http ://dx .doi .org /10 .1073 /pnas .0308724101

Shiiki, T., S. Ohtsuki, A. Kurihara, H. Naganuma, K. Nishimura, M. Tachikawa, K. Hosoya, and T. Terasaki. 2004. Brain insulin impairs amyloid-beta(1-40) clearance from the brain. J. Neurosci. 24:9632–9637. http ://dx .doi .org /10 .1523 /JNE URO SCI .2236 -04 .2004

Sims-Robinson, C., B. Kim, A. Rosko, and E.L. Feldman. 2010. How does diabetes accelerate Alzheimer disease pathology? Nat. Rev. Neurol. 6:551–559. http ://dx .doi .org /10 .1038 /nrneurol .2010 .130

Sperling, R.A., P.S. Aisen, L.A. Beckett, D.A. Bennett, S. Craft, A.M. Fagan, T. Iwatsubo, C.R. Jack Jr., J. Kaye, T.J. Montine, et al. 2011. Toward defining the preclinical stages of Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. 7:280–292. http ://dx .doi .org /10 .1016 /j .jalz .2011 .03 .003

Starks, E.J., J. Patrick O’Grady, S.M. Hoscheidt, A.M. Racine, C.M. Carlsson, H. Zetterberg, K. Blennow, O.C. Okonkwo, L. Puglielli, S. Asthana, et al. 2015. Insulin resistance is associated with higher cerebrospinal fluid taulevels in asymptomatic APOEe4 carriers. J. Alzheimers Dis. 46:525–533. http ://dx .doi .org /10 .3233 /JAD -150072

Steen, E., B.M. Terry, E.J. Rivera, J.L. Cannon, T.R. Neely, R. Tavares, X.J. Xu, J.R. Wands, and S.M. de la Monte. 2005. Impaired insulin and insulin-like growth factor expression and signaling mechanisms in Alzheimer’s dis-ease--is this type 3 diabetes? J. Alzheimers Dis. 7:63–80.

Stöhr, O., K. Schilbach, L. Moll, M.M. Hettich, S. Freude, F.T. Wunderlich, M. Ernst, J. Zemva, J.C. Brüning, W. Krone, et al. 2013. Insulin receptorsignaling mediates APP processing and β-amyloid accumulation without altering survival in a transgenic mouse model of Alzheimer’s disease. Age (Dordr.). 35:83–101. http ://dx .doi .org /10 .1007 /s11357 -011 -9333 -2

1385JEM Vol. 213, No. 8

Stolk, R.P., M.M. Breteler, A. Ott, H.A. Pols, S.W. Lamberts, D.E. Grobbee, and A. Hofman. 1997. Insulin and cognitive function in an elderly population. The Rotterdam Study. Diabetes Care. 20:792–795. http ://dx .doi .org /10 .2337 /diacare .20 .5 .792

Talbot, K., H.Y. Wang, H. Kazi, L.Y. Han, K.P. Bakshi, A. Stucky, R.L. Fuino, K.R. Kawaguchi, A.J. Samoyedny, R.S. Wilson, et al. 2012. Demonstrated brain insulin resistance in Alzheimer’s disease patients is associated with IGF-1 resistance, IRS-1 dysregulation, and cognitive decline. J. Clin. Invest. 122:1316–1338. http ://dx .doi .org /10 .1172 /JCI59903

Unger, J., T.H. McNeill, R.T. Moxley III, M. White, A. Moss, and J.N. Livingston. 1989. Distribution of insulin receptor-like immunoreactivity in the rat forebrain. Neuroscience. 31:143–157. http ://dx .doi .org /10 .1016 /0306 -4522(89)90036 -5

Vandal, M., P.J. White, C. Tremblay, I. St-Amour, G. Chevrier, V. Emond, D. Lefrançois, J. Virgili, E. Planel, Y. Giguere, et al. 2014. Insulin reverses the high-fat diet-induced increase in brain Aβ and improves memory in ananimal model of Alzheimer disease. Diabetes. 63:4291–4301. http ://dx .doi .org /10 .2337 /db14 -0375

Wadman, M. 2012. US government sets out Alzheimer’s plan. Nature. 485:426–427. http ://dx .doi .org /10 .1038 /485426a

Wang, X., W. Zheng, J.W. Xie, T. Wang, S.L. Wang, W.P. Teng, and Z.Y. Wang. 2010. Insulin deficiency exacerbates cerebral amyloidosis and behavioral deficits in an Alzheimer transgenic mouse model. Mol. Neurodegener. 5:46. http ://dx .doi .org /10 .1186 /1750 -1326 -5 -46

Watson, G.S., E.R. Peskind, S. Asthana, K. Purganan, C. Wait, D. Chapman, M.W. Schwartz, S. Plymate, and S. Craft. 2003. Insulin increases CSFAβ42 levels in normal older adults. Neurology. 60:1899–1903. http ://dx.doi .org /10 .1212 /01 .WNL .0000065916 .25128 .25

Willette, A.A., S.C. Johnson, A.C. Birdsill, M.A. Sager, B. Christian, L.D. Baker, S. Craft, J. Oh, E. Statz, B.P. Hermann, et al. 2015. Insulin resistance predicts brain amyloid deposition in late middle-aged adults. Alzheimers Dement. 11:504–510.e1. http ://dx .doi .org /10 .1016 /j .jalz .2014 .03 .011

Xie, L., E. Helmerhorst, K. Taddei, B. Plewright, W. Van Bronswijk, and R. Martins. 2002. Alzheimer’s β-amyloid peptides compete for insulin bind-ing to the insulin receptor. J. Neurosci. 22:RC221.

Yarchoan, M., J.B. Toledo, E.B. Lee, Z. Arvanitakis, H. Kazi, L.Y. Han, N. Louneva, V.M. Lee, S.F. Kim, J.Q. Trojanowski, and S.E. Arnold. 2014. Abnormal serine phosphorylation of insulin receptor substrate 1 is associated with tau pathology in Alzheimer’s disease and tauopathies. Acta Neuropathol. 128:679–689. http ://dx .doi .org /10 .1007 /s00401 -014 -1328 -5

Zhang, Y., K. Chen, S.A. Sloan, M.L. Bennett, A.R. Scholze, S. O’Keeffe, H.P. Phatnani, P. Guarnieri, C. Caneda, N. Ruderisch, et al. 2014. An RNA-sequencing transcriptome and splicing database of glia, neurons, and vascular cells of the cerebral cortex. J. Neurosci. 34:11929–11947. http ://dx .doi .org /10 .1523 /JNE URO SCI .1860 -14 .2014

Zhang, Y., C.L. Dai, Y. Chen, K. Iqbal, F. Liu, and C.X. Gong. 2016. Intranasal insulin prevents anesthesia-induced spatial learning and memory deficit in mice. Sci. Rep. 6:21186. http ://dx .doi .org /10 .1038 /srep21186

Zhao, W.Q., F.G. De Felice, S. Fernandez, H. Chen, M.P. Lambert, M.J. Quon, G.A. Krafft, and W.L. Klein. 2008. Amyloid beta oligomers induce impairment of neuronal insulin receptors. FAS EB J. 22:246–260. http ://dx .doi .org /10 .1096 /fj .06 -7703com

Zuchero, Y.J., X. Chen, N. Bien-Ly, D. Bumbaca, R.K. Tong, X. Gao, S. Zhang, K. Hoyte, W. Luk, M.A. Huntley, et al. 2016. Discovery of novel blood-brain barrier targets to enhance brain uptake of therapeutic antibodies. Neuron. 89:70–82. http ://dx .doi .org /10 .1016 /j .neuron .2015 .11 .024

The Rockefeller University Press $30.00J. Exp. Med. 2015 Vol. 212 No. 3 287–295www.jem.org/cgi/doi/10.1084/jem.20142322

287

Brief Definit ive Report

It has historically been debated whether in-flammation plays an active role in Alzheimer’s disease (AD) pathogenesis or is ancillary to other AD pathologies (Wyss-Coray, 2006). How-ever, recent genome-wide association studies linked polymorphisms in inflammation-related genes to increased AD risk, supporting the con-clusion that inflammation can be a causative factor in disease pathology (Karch and Goate, 2015). Notably, variants of the immune cell–specific triggering receptor expressed on myeloid cells 2 (TREM2) confer dramatically elevated risk for developing AD (R. Guerreiro et al., 2013; Jonsson et al., 2013). TREM2 variants are also the genetic basis of Nasu-Hakola disease (Bird et al., 1983) and confer increased risk for fron-totemporal dementia (FTD; R.J. Guerreiro et al.,

2013), Parkinson’s disease (Rayaprolu et al., 2013), and amyotrophic lateral sclerosis (Cady et al., 2014). These data suggest that TREM2 may serve a common function that modifies risk for neurodegenerative disorders.

TREM2 is an important modulator of immune cell function. In the brain, the TREM2 receptor is expressed exclusively by myeloid cells (Colonna, 2003). TREM2, along with its obligate intracellular adaptor DAP12, was identified as a hub gene in systems biology analyses in AD (Forabosco et al., 2013; Zhang et al., 2013), highlighting the central impor-tance of these myeloid cell signaling elements

CORRESPONDENCE Bruce T. Lamb: lambb@ccf.org

Abbreviations used: A, amy-loid; AD, Alzheimer’s disease; DAB, 3,3-diaminobenzidine; FTD, frontotemporal dementia; IHC, immunohistochemistry; MAPT, microtubule-associated protein .

*T.R. Jay and C.M. Miller contributed equally to this paper.

TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer’s disease mouse models

Taylor R. Jay,1,3* Crystal M. Miller,1* Paul J. Cheng,1,3 Leah C. Graham,4 Shane Bemiller,1,6 Margaret L. Broihier,3 Guixiang Xu,1 Daniel Margevicius,1 J. Colleen Karlo,3 Gregory L. Sousa,4 Anne C. Cotleur,1 Oleg Butovsky,5

Lynn Bekris,1 Susan M. Staugaitis,1 James B. Leverenz,2 Sanjay W. Pimplikar,1,3

Gary E. Landreth,3 Gareth R. Howell,4 Richard M. Ransohoff,1

and Bruce T. Lamb1,2,3

1The Lerner Research Institute and 2Lou Ruvo Center for Brain Health, Cleveland Clinic, Cleveland, OH 441953Case Western Reserve University, Cleveland, OH 441064The Jackson Laboratory, Bar Harbor, ME 046095Brigham and Women’s Hospital, Boston, MA 021156Kent State University, Kent, OH 44340

Variants in triggering receptor expressed on myeloid cells 2 (TREM2) confer high risk for Alzheimer’s disease (AD) and other neurodegenerative diseases. However, the cell types and mechanisms underlying TREM2’s involvement in neurodegeneration remain to be estab-lished. Here, we report that TREM2 is up-regulated on myeloid cells surrounding amyloid deposits in AD mouse models and human AD tissue. TREM2 was detected on CD45hiLy6C+ myeloid cells, but not on P2RY12+ parenchymal microglia. In AD mice deficient for TREM2, the CD45hiLy6C+ macrophages are virtually eliminated, resulting in reduced inflammation and ameliorated amyloid and tau pathologies. These data suggest a functionally important role for TREM2+ macrophages in AD pathogenesis and an unexpected, detrimental role of TREM2 in AD pathology. These findings have direct implications for future development of TREM2-targeted therapeutics.

© 2015 Jay et al. This article is distributed under the terms of an Attribution– Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial– Share Alike 3.0 Unported license, as described at http://creativecommons.org/ licenses/by-nc-sa/3.0/).

288 TREM2 expression and function in AD | Jay et al.

RESULTS AND DISCUSSIONTREM2-expressing myeloid cells associate with amyloid (A) depositsWe defined the cellular localization and kinetics of TREM2 expression in tissues from two transgenic AD mouse models that exhibit age-related A plaque deposition. TREM2 immuno-histochemistry (IHC) revealed an age-dependent increase in TREM2-expressing cells around Congo red–positive plaques in both APPPS1 and 5XFAD mice (Fig. 1, f–m). We observed similar expression patterns in two neuropathologically con-firmed AD cases (Fig. 1, n–q). Despite reports of TREM2 ex-pression in microglia (Frank et al., 2008; Hickman et al., 2013), we did not detect TREM2 staining in nontransgenic mice (Fig. 1, b–e) or in nonplaque-associated cells (Fig. 1 o and see Fig. 4 l). Thus, TREM2 is expressed in parenchymal microglia

in neurodegeneration. In vitro, TREM2 promotes phagocytosis, suppresses toll-like receptor–induced inflammatory cytokine production and enhances antiinflammatory cytokine tran-scription (Neumann and Takahashi, 2007; Paradowska-Gorycka and Jurkowska, 2013). In contrast, Trem2/ mice have less inflammation and enhanced phagocytosis in models of stroke and lung infection (Sieber et al., 2013; Sharif et al., 2014). These findings suggest that the role of TREM2 in modulating inflammation may be more complex than previ-ously appreciated and may be dependent on the cell type in which it is expressed and the inflammatory context in which it is studied. In the AD brain, the role of TREM2 is poorly understood. In this study, we identify disease-relevant cell types that express TREM2 and examine the effects of loss of TREM2 function on AD pathologies.

Figure 1. TREM2 expression is increased around A plaques. (a) A novel Trem2 knockout with a LacZ reporter under the control of the Trem2 promoter was gener-ated. (b–m) TREM2 IHC was performed in WT mice at 2 (b), 4 (c), 6 (d), and 12 mo (e) of age, in APPPS1 mice at 2 (f), 4 (g), 6 (h), and 12 mo (i) of age, and in 5XFAD mice at 2 (j), 4 (k), 6 (l), and 8 mo (m) of age (n = 4–5). Insets show Congo red costaining with ar-rows indicating plaques. (n–q) TREM2 was also observed around plaques (n, p, and q) but not in regions lacking A deposition (o) in human AD cases (n = 2). Arrows pointto Congo red–positive plaques. (r) TREM2immunoreactivity was not observed inAPPPS1;Trem2/ mice (n = 7). Insets showCongo red costaining with arrows indicatingplaques. (s and t) TREM2 expression was alsoassessed by qRT-PCR (s; two-way ANOVA,age P = 0.083, genotype P = 0.0036, inter-action P = 0.185; Bonferroni-corrected Stu-dent’s t tests shown; n = 6–8 per group) andWestern blot (t; n = 2–3). Arrowhead indi-cates position of TREM2-specific band. Errorbars indicate SEM. *, P < 0.05. At least threeindependent experiments were performedfor all analyses. Bars, 100 µm (bars in b andn apply to b–p and r).

JEM Vol. 212, No. 3 289

Br ief Definit ive Repor t

the effect size previously observed in APPPS1;Trem2+/ mice (Ulrich et al., 2014). Cells expressing high levels of the peripheral monocyte marker Ly6C (Fig. 4 d) and CD45 (Fig. 4 g) are exclusively associated with Congo red–positive plaques in AD mouse models and in human AD tissue (Fig. 4 f), whereas TREM2-deficient animals revealed a near absence of plaque-associated CD45+Ly6C+ macrophages (Fig. 4, e and h). Consistent with these data, transcript levels of the myeloid cell markers CD11b, CD68, F4/80, and CD45 were all significantly decreased in brain lysates from APPPS1;Trem2/ animals compared with APPPS1;Trem2+/+ controls (Fig. 4 m). Interestingly, P2RY12, a purinergic receptor selectively expressed in microglia (Hickman et al.,

at levels below the limit of detection using IHC. TREM2 antibody specificity was confirmed in APPPS1;Trem2/ mice (Fig. 1 r). TREM2 RNA (Fig. 1 s) and protein (Fig. 1 t) analyses confirmed an age-dependent increase in TREM2 expression in the brains of AD mice.

We next wanted to assess the cellular localization of TREM2. We demonstrated that TREM2 transcripts colocalized with plaque-associated Iba1+ myeloid cells in APP/PS1 mice using in situ hybridization (Fig. 2 a). This was confirmed in mice in which exons 2–4 of the Trem2 gene were replaced with LacZ (Fig. 1 a) as a reporter for TREM2 gene expression (Fig. 1 a). X-gal staining in APPPS1;Trem2LacZ/+ mice visualized LacZ ex-pression in Iba1+ cells around A deposits (Fig. 2 b). Triple fluor-escent IHC confirmed that TREM2 protein colocalized with plaque-associated Iba1+ cells (Fig. 2 c), but not with markers of reactive astrocytes (GFAP, Fig. 2 d), neurons (MAP2, Fig. 2 e), oligodendrocytes (MBP, Fig. 2 f), or with parenchymal Iba1+ cells that were not associated with plaques (see Fig. 4 l). Collec-tively, these results demonstrate that TREM2 is selectively up-regulated by myeloid cells surrounding A deposits.

TREM2+ plaque–associated myeloid cells express markers characteristic of monocyte-derived macrophagesThe identity of TREM2-expressing cells in AD mouse mod-els was further examined using flow cytometry. The myeloid cell population was selected using CD11b (Fig. 3 a). Although there is no universally accepted marker to definitively distin-guish macrophages derived from microglia and those derived from infiltrating monocytes, differences in levels of CD45 ex-pression have been used extensively in flow cytometry to dis-tinguish these two cell populations (Sedgwick et al., 1991; Chiu et al., 2013; Hickman et al., 2013; Butovsky et al., 2014). In these studies, CD45hi was used to identify cells that origi-nate from bone marrow–derived monocytes, whereas CD45lo identified resident microglia. We found that CD11b+TREM2+ cells were exclusively CD45hi (Fig. 3 c). There was also a striking age-dependent increase in the percentage of brain TREM2+CD11b+CD45hi cells in two different AD mouse models (Fig. 3, d and e), whereas TREM2 expression by CD11b+CD45lo cells did not differ from that seen in wild-type mice at any time point. TREM2+ cells were also uniformly F4/80+, confirming their macrophage lineage (Fig. 3 f). TREM2 antibody specificity was validated in APPPS1;Trem2/ mice (Fig. 3, g and h). Collectively, these results show that the TREM2+ macrophages that surround the A deposits in the transgenic mouse models of AD are CD45hi, a canonical marker of peripherally derived macrophages.

TREM2-deficient mice have reduced A plaque–associated macrophagesTo determine the role of TREM2 in AD-like pathology, APPPS1 mice were crossed with Trem2/ mice (Fig. 1 a). Although APPPS1;Trem2+/+ mice exhibit robust accumu-lation of Iba1+ myeloid cells around plaques (Fig. 4 a), APPPS1;Trem2/ mice have a fivefold decrease in plaque-associated Iba1+ cells (Fig. 4, b and c), approximately twice

Figure 2. TREM2 is expressed in plaque-associated myeloid cells. (a) In situ hybridization with TREM2 probes colocalized with Iba1 (n = 2).(b) X-gal staining of brain tissue from 4-mo-old APPPS1;Trem2LacZ/+ micecolocalized with fluorescent IHC for Iba1 and 6E10 (n = 3). (c–f) Confocalmicroscopy was used to assess TREM2 colocalization with 6E10+ plaque-associated myeloid cells (c; Iba1), astrocytes (d; GFAP), neurons (e; MAP2),or oligodendrocytes (f; MBP; n = 8). At least two independent experimentswere performed for all analyses. Bars: (a) 20 µm; (b–f) 50 µm.

290 TREM2 expression and function in AD | Jay et al.

deficiency reduces transcripts for two cardinal inflammatory mediators and increases transcripts for selected antiinflamma-tory markers.

TREM2 deficiency reduces hippocampal A depositionGiven that TREM2 deficiency had robust effects on macro-phage accumulation and neuroinflammation, factors associated with A deposition, we assessed resultant changes in amyloid pathology in these mice. At 4 mo of age, APPPS1;Trem2/ mice displayed significantly reduced 6E10 area in the hippo-campus (Fig. 5, a and c), as well as other brain regions includ-ing the olfactory bulb and thalamus (not depicted), compared with APPPS1;Trem2+/+ controls. Interestingly, the effect of TREM2 deficiency on amyloid accumulation was not con-sistent across brain regions, as we observed only modest, non-significant effects in the cortex (Fig. 5, a and c) and brainstem (not depicted). Thioflavin S staining of dense core plaques revealed similar results (Fig. 5, b and d). Although the mechanism underlying this region specificity requires further investigation, similar effects have been observed in other con-texts (Riddell et al., 2007). The histological results were con-firmed by Western blot analysis (Fig. 5 e), which indicated no significant changes in APP (Fig. 5 f) but lower A levels in APPPS1;Trem2/ mice (Fig. 5 g). To assess which A spe-cies were affected by TREM2 deficiency, we performed ELISAs on brain lysates. These results demonstrated reduced levels of insoluble A42 (Fig. 5 h) and trends toward a reduction in soluble A42 (Fig. 5 h) and soluble and insolu-ble A40 (Fig. 5 i) in APPPS1;Trem2/ mice compared with APPPS1;Trem2+/+ mice. These data demonstrate that TREM2 deficiency ameliorates amyloid pathology in a region- specific manner.

TREM2 deficiency reduces astrocytosis and microtubule-associated protein (MAPT) pathologyIn addition to amyloid deposition, APPPS1 mice also exhibit astrocytosis and accumulation of hyperphosphorylated MAPT in dystrophic neurites around amyloid plaques. We examined the effects of TREM2 deficiency on these pathological hall-marks. TREM2-deficient APPPS1 mice had decreased num-bers of GFAP immunoreactive astrocytes around A deposits (Fig. 6, a–c) and a corresponding decrease in GFAP mRNA (Fig. 6 f). Additionally, there was a dramatic reduction in phosphorylated MAPT staining associated with A deposits in APPPS1;Trem2/ mice when compared with APPPS1;Trem2+/+ controls, as detected with AT8 (Fig. 6, d and e) and AT180 (Fig. 6, g–i) antibodies. Further investigation will be required to determine whether these changes in MAPT phosphoryla-tion are a result of reduced amyloid deposition or whether they are regulated independently in this context. However, TREM2 deficiency not only ameliorated amyloid pathology but also reduced reactive astrocyte accumulation and MAPT hyperphosphorylation.

In this study, we demonstrated several surprising results regarding the expression and function of TREM2 in AD. First, we found that TREM2+ cells in AD mouse models

2013; Butovsky et al., 2014), was expressed by parenchymal, nonplaque-associated cells in APPPS1;Trem2+/+ mice (Fig. 4 j) and in human AD tissue (Fig. 4 i), in contrast to CD45 and TREM2 (Fig. 4 l). Unlike the CD45+Ly6C+ cells, the P2RY12+ cells were not eliminated in TREM2-deficient mice (Fig. 4 k), and there was no change in P2RY12 RNA in APPPS1;Trem2/ mice compared with APPPS1;Trem2+/+ controls (Fig. 4 m). Collectively, these results demonstrate a dramatic reduction in A plaque–associated CD45+Ly6C+ macrophages, but not in parenchymal P2RY12+ microglia in TREM2-deficient mice.

TREM2 deficiency reduces neuroinflammationBecause plaque-associated myeloid cells have been shown to play an important role in modulating neuroinflammation in AD (Akiyama et al., 2000), we assessed inflammatory markers in APPPS1 mice lacking TREM2. Transcript levels of in-flammatory cytokines IL-1 and IL-6 were reduced (Fig. 4 n), whereas antiinflammatory markers chitinase-like 3/Ym1 and resistin B–like/Fizz1 were substantially increased in TREM2- deficient animals (Fig. 4 o). These data show that TREM2

Figure 3. TREM2 is specifically expressed on CD11b+CD45hiF4/80+ macrophages in AD mice. (a and b) Isolated brain myeloid cells were gated on CD11b (a) and divided into CD45lo and CD45hi populations (b). (c and f) TREM2+ cells were exclusively CD45hi (c) and F4/80+ (f; n = 41). (d and e) TREM2 expression was quantified on CD45lo and CD45hi popula-tions in APPPS1 mice (d; two-way ANOVA, age P < 0.0001; genotype/cell type P < 0.0001, interaction P < 0.0001; Bonferroni-corrected Student’s t tests shown; n = 2–8), 5XFAD mice (e; two-way ANOVA, age P = 0.025, genotype/cell type P < 0.0001, interaction P = 0.0003; Bonferroni-corrected Student’s t tests shown; n = 5–9), and WT controls (d and e; n = 4–14). Error bars indicate SEM. **, P < 0.01; ***, P < 0.001. (g and h) Flow cytom-etry on APPPS1;Trem2/ mice (g) revealed a lack of TREM2+ cells com-pared with APPPS1;Trem2+/+ mice (h; n = 7). At least two independent experiments were performed for all analyses.

JEM Vol. 212, No. 3 291

Br ief Definit ive Repor t

If TREM2+ cells are derived from the periphery, as their marker expression would suggest, then TREM2 deficiency could result in the striking reduction in macrophage accumu-lation observed in this study through several mechanisms: impaired transmigration of TREM2-negative cells across the blood–brain barrier or alterations directly to the brain vascu-lature that alter cell trafficking, reduced chemotaxis of TREM2- deficient cells to parenchymal A deposits, or shortened survival of these cells within the CNS. An alternative inter-pretation of the data in this study is that TREM2 is necessary for large-scale changes in microglial phenotype, including the up-regulation of CD45 and Ly6C and coincident down- regulation of P2RY12. In this case, TREM2 deficiency could

expressed high levels of CD45 and Ly6C, canonical markers of macrophages derived from peripheral monocytes. The potential role of peripherally derived macrophages in AD has been controversial (Prinz and Priller, 2014), but it is well accepted that CD45hiLy6C+CCR2+ monocytes enter the central nervous system (CNS) and modulate pathology in other disease contexts (Mildner et al., 2011). Although our experiments identify a marker signature on TREM2+ cells consistent with this peripherally derived population, further investigation will be required to conclusively dem-onstrate the functional relevance of peripherally derived cells in AD and the ontogeny of the TREM2+ cell population in the AD brain.

Figure 4. Plaque-associated myeloid cells are reduced in TREM2-deficient mice. (a–c) Confocal microscopy was used to examine Iba1 and 6E10 expression in 4-mo-old APPPS1;Trem2/ mice (b) and APPPS1;Trem2+/+ controls (a; quantified in c). (d and e) IHC and Congo red costaining (arrows) was performed for Ly6C in APPPS1;Trem2+/+ mice (d) and APPPS1;Trem2/ animals (e). (a, b, d, and e) Insetsshow higher-magnification images. (f–h) CD45 and Congored staining (arrows) was performed in human AD tissue(f; n = 2), APPPS1;Trem2+/+ mice (g), and APPPS1;Trem2/

mice (h). (i–k) P2RY12 and Congo red staining (arrows) wasperformed in human AD tissue (i; n = 2), APPPS1;Trem2+/+

mice (j), and APPPS1;Trem2/ animals (k). (l) TREM2 colo-calized with CD45 but not P2RY12. (m–o) qRT-PCR wasperformed on whole brain lysates to examine transcriptlevels of myeloid cell markers (m), proinflammatory cyto-kines (n), and antiinflammatory markers (o). All experimentsused n = 7–8 mice per group unless otherwise noted, andat least two independent experiments were performed forall analyses. Error bars indicate SEM. *, P < 0.05; **, P <0.01; ***, P < 0.001. Bars: (a and b) 50 µm; (d–l) 20 µm.

292 TREM2 expression and function in AD | Jay et al.

a change within the putative ligand-binding domain of TREM2, and alterations of its surface trafficking and function were less dramatic than the FTD variants (Kleinberger et al., 2014). Thus, it is possible that this variant could confer both loss- and gain-of-function phenotypes, which would be consistent with our findings. Going forward, it will be important to generate mice with knock-in TREM2 risk alleles to asses this possibility.

Although TREM2 has perhaps received the most attention for the high risk it confers for developing AD, TREM2 vari-ants also confer risk for developing other neurodegenerative diseases, including FTD, amyotrophic lateral sclerosis, and Parkinson’s disease (R.J. Guerreiro et al., 2013; Rayaprolu et al., 2013; Borroni et al., 2014; Cady et al., 2014). Thus, it is plau-sible that TREM2 plays a common role in modifying risk for developing these diverse CNS pathologies. It will be important to determine whether TREM2 is expressed on a common macrophage subset and promotes similar changes in neuroin-flammation in these other models. The results from the present study will help inform the future research agenda related to TREM2 biology in these other disease contexts.

reduce macrophage accumulation by eliminating microglial recognition of A as a relevant stimulus, impairing migration of microglia to plaques, or preventing these phenotypic changes. Additional studies will be required to determine which mechanisms are responsible for the changes in TREM2-deficient AD mice in the current study.

Regardless of mechanism, our results demonstrate that TREM2 deficiency is protective against disease pathogenesis in AD mouse models. These results are surprising given similar findings in AD mice in which TREM2 was overexpressed (Jiang et al., 2014), although in this study TREM2 expression was not restricted to the cell types in which TREM2 is ex-pressed physiologically. Our findings are also counterintuitive based on the human genetic data that demonstrate that pre-sumed loss-of-function variants in TREM2 promote AD pathogenesis. TREM2 variants such as the missense mutation Q33X (R. Guerreiro et al., 2013) and the FTD-related vari-ants T66M and Y38C which impair protein maturation (Kleinberger et al., 2014), almost certainly impair TREM2 function. However, the primary AD risk allele, R47H, harbors

Figure 5. TREM2 deficiency reduces A accumulation. (a) IHC with 6E10 was performed on brain slices from 4-mo-old APPPS1;Trem2+/+ and APPPS1;Trem2/ mice (quantified in c; n = 7–8). (b) Analysis of ThioS plaque number revealed similar results (quan-tified in d; n = 7–8). (a and b) Insets show a higher magnification of the hippocampus. Bar, 1 mm. (e) APP and A levels were as-sessed by Western blot using 6E10. (f and g) Quantified relative to GAPDH, there was no significant change in APP protein levels (f) but a significant reduction in A (g) inAPPPS1;Trem2/ mice (n = 3–4). (h and i)ELISAs on brain lysates also showed a signifi-cant reduction in insoluble A42 and a trendtoward a reduction in soluble A42 (h) and atrend toward a reduction in soluble andinsoluble A40 (i; n = 3–4). #, P < 0.10;*, P < 0.05; ***, P < 0.001. At least twoindependent experiments were performedfor all analyses.

on October 6, 2016

Dow

nloaded from

JEM Vol. 212, No. 3 293

Br ief Definit ive Repor t

AD. Postmortem interval was 6 h, and tissue was immediately frozen. Tissue labeled “AD Case 2” is from the hippocampus of a 78-yr-old male with an 8-yr-history of cognitive decline. The patient’s son and power-of-attorneyprovided written permission for release for research purposes of any autopsytissue not needed for diagnosis. Patient identifying information is housed inan Institutional Review Board–approved database. Postmortem interval was36 h. The tissue was stored at 4°C in 4% paraformaldehyde for 2 d, followedby cryoprotection solution (20% glycerol in 0.08 M phosphate buffer,pH 7.6) until use. Human experiments were approved by The ClevelandClinic and University of Washington Institutional Review Boards.

In situ hybridization. Mice were perfused with 4% PFA, and 25-µm-thick frozen brain sections were prepared. A digoxigenin (DIG)-labeled riboprobe for TREM2 was transcribed from a cDNA clone (MMM1013-202767203; Thermo Fisher Scientific). The plasmid was digested with Sal1, and in vitro transcription was performed with T7 polymerase. A TREM2 sense probe was digested with a Not1 restriction enzyme and transcribed by SP6 RNA polymerase. Colorimetric detection of hybridized mRNA was performed using an anti-DIG conjugated to alkaline phosphatase and was developed using NBT-BCIP substrate (Roche).

Quantitative RT-PCR (qRT-PCR). Mice were perfused with PBS, and their brains were removed, snap frozen, and kept at 80°C until use. Tissue was homogenized in 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, and 1:100 protease inhibitor cocktail in PBS. RNA was isolated using chloro-form extraction and was purified using Purelink RNA Mini kit (Life Tech-nologies) and treated with DNase Purelink (Life Technologies). cDNA was prepared from 1.5 µg RNA using a QuantiTect Reverse Transcription kit (QIAGEN), and real-time PCR was performed for 40 cycles with the StepOne Plus Real Time PCR system (Life Technologies). All primers and TaqMan probes were purchased from the Life Technologies database. Rela-tive gene expression was determined using the CT method. A two-way ANOVA was performed, and significance between CT values was deter-mined using a Bonferroni post-hoc test for the TREM2 qPCR and a Student’s t test used for qPCR assays comparing APPPS1;Trem2+/+ and APPPS1;Trem2/ animals.

Western blotting. Tissue was extracted and processed as described above for qRT-PCR. After sonication and centrifugation, protein concentration was determined using a BCA kit (Thermo Fisher Scientific). Proteins were denatured for 15 min at 95°C in 35% denaturing buffer containing LDS sample buffer (Life Technologies) and reducing agent (Life Technologies). 30–50 µg of protein per sample was loaded along with 5 µl Magic Mark XP protein ladder (Life Technologies) onto Novex 4–12% Bis-Tris gels (Life Technologies), run at 160 V for 30–45 min, and transferred onto PVDF membranes (EMD Millipore) in 1× TAE buffer at 100 mA overnight at room temperature. After transfer, membranes were blocked using Odyssey Blocking Buffer in PBS (LI-COR Biosciences) for 1 h at room temperature and incubated in the appropriate primary antibodies in blocking buffer with 0.1% Tween 20 overnight at 4°C, 6E10 (1:5,000; Signet), TREM2 (1:500; R&D Systems), CT15 (1:10,000; a gift from E.H. Koo, University of Cali-fornia, San Diego, La Jolla, CA), and GAPDH (1:10,000; Thermo Fisher Scientific). Membranes were washed in PBST (0.1% Tween 20), incubated in the appropriate IR dye–conjugated secondary antibody (Thermo Fisher Scientific) in blocking buffer/0.1% Tween 20, and imaged using an Odyssey IR Scanner (LI-COR Biosciences) system. ImageJ software (National Insti-tutes of Health) was used for densitometric analysis, and each experimental sample was normalized to GAPDH.

ELISA. A extraction was performed on microdissected brain tissue en-riched for cortex and hippocampus. Lysates were mixed with an equal vol-ume of 0.4% diethylamine in NaCl and centrifuged for 13,500 g for 1 h at 4°C. The supernatant was collected and neutralized with 0.5 M Tris, pH 6.8, and analyzed as the soluble A fraction. The pellet was sonicated with 70% formic acid and centrifuged at 105,000 g for 45 min at 4°C. The supernatant

MATERIALS AND METHODSMice. Two amyloid mouse models of AD were analyzed. APPPS1-21 (termed APPPS1) mice (provided by M. Jucker, German Center for Neuro-degenerative Diseases [DZNE], Tubingen, Germany) express human APP with the Swedish (K670M/N671L) and PSEN1 L166P mutations under control of the Thy1 promoter (Radde et al., 2006). This mouse was main-tained on the B6 background. 5XFAD mice (The Jackson Laboratory B6SJL-Tg(APPSwFlLon,PSEN1*M146*l286V)) express mutant human APP(695) with Swedish (K670N/M671L), Florida (I716V), and London (V717I) mutations and human PSEN1 with M146L and L286V mutations under control of the Thy1 promoter (Oakley et al., 2006). This mouse was maintained on a mixed B6/SJL background.

We also used a novel Trem2/ mouse model (Trem2tm1(KOMP)Vlcg), which has a LacZ reporter cassette knocked into the endogenous Trem2 locus in place of exons 2 and 3 and most of exon 4, resulting in a loss of TREM2 func-tion as well as expression of the LacZ reporter under the control of the TREM2 promoter (Fig. 1 a). This mouse was generated by the trans-NIH Knock-Out Mouse Project (KOMP). These mice were maintained on a B6 background. These mice were crossed with APPPS1 mice to yield APPPS1;Trem2+/LacZ and APPPS1;Trem2LacZ/LacZ genotypes (also termed throughout the paper APPPS1;Trem2+/ and APPPS1;Trem2/, respectively).

Mice were housed in the Cleveland Clinic Biological Resources Unit, Case Western Reserve University Animal Resource Center and The Jackson Laboratory, facilities fully accredited by the Association and Accreditation of Laboratory Animal Care. All experimental procedures were approved by the Institutional Animal Care and Use Committee at each respective institution.

Human tissue. Human AD tissue was obtained from two neuropathologi-cally confirmed AD patients. Tissue labeled “AD Case 1” is hippocampal tissue from an 88-yr-old female with a clinical diagnosis of Braak stage VI-C

Figure 6. TREM2 deficiency reduces astrocytosis and MAPT phosphorylation. (a and b) Astrocytosis was assessed in 4-mo-old APPPS1;Trem2+/+ (a) and APPPS1;Trem2/ mice (b) using IHC for GFAP and 6E10 (n = 7–8). (c and f) The number of GFAP+ cells surrounding plaques was quantified (c; n = 3–4), and results were confirmed by qRT-PCR (f; n = 7–8). (d, e, g, and h) Hyperphosphorylated MAPT was detected in APPPS1;Trem2/ (e and h) and APPPS1;Trem2+/+ mice (d and g) with AT8 (d and e) and AT180 antibodies (g and h; n = 7–8). Arrows indicate Congo red–positive plaques. (i) Quantification of the area of AT180 immunoreactivity revealed significant decreases in APPPS1;Trem2/ mice (n = 3–4). At least two independent experiments were performed for all analyses. Error bars indicate SEM. *, P < 0.05; **, P < 0.01. Bar, 50 µm.

294 TREM2 expression and function in AD | Jay et al.

subsets. TREM2+ events were assessed by overlaying FMO and ALL plots for each sample and gating on the population that was present only in the ALL sample. Significant differences between ages and genotypes were deter-mined using a two-way ANOVA and Bonferroni-corrected Student’s t tests between all groups.

Statistics. Statistical analyses were performed using Prism (GraphPad Software). Two-way ANOVAs and significance between individual groups were determined using a Bonferroni post-hoc test for analyses with multiple comparisons. Two-sided, unpaired Student’s t tests were used to determine statistical difference between samples in analyses that required only single comparisons. Although the data were not formally tested, based on previous results (Lee et al., 2014), we assumed they conform to a Gaussian distribution and that variance between groups was comparable. Biological replicates were used to define each n. Statistical outliers were excluded from all datasets. The mean of each group is graphed, and the error bars represent the SEM. De-gree of significance between groups is represented as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001. No tests were performed a priori to determine the sample size; however, the sample sizes used here are similar to those used in a previous study (Lee et al., 2014). 3–12 mice from at least two cohorts were included in each group.

We thank Rebecca Achey and Jiayang Li and the flow cytometry and imaging cores for their technical support.

This work was supported by an Alzheimer’s Association Multi-Center Program Grant, the Jane and Lee Seidman Fund, a generous donation from Chet and Jane Scholtz, National Institute on Aging grant AG023012, National Institute of Neurological Disorders and Stroke grants NS047804 and NS087298, Department of Defense grant W81XWH12-1-0629, BrightFocus Foundation grant A2013252F, and National Research Service Awards T32 NS067431 and T32 GM007250.

The authors declare no competing financial interests.

Submitted: 13 December 2014Accepted: 12 February 2015

REFERENCESAkiyama, H., S. Barger, S. Barnum, B. Bradt, J. Bauer, G.M. Cole, N.R.

Cooper, P. Eikelenboom, M. Emmerling, B.L. Fiebich, et al. 2000. Inflammation and Alzheimer’s disease. Neurobiol. Aging. 21:383–421. http://dx.doi.org/10.1016/S0197-4580(00)00124-X

Bird, T.D., R.M. Koerker, B.J. Leaird, B.W. Vlcek, and D.R. Thorning. 1983. Lipomembranous polycystic osteodysplasia (brain, bone, and fat disease): a genetic cause of presenile dementia. Neurology. 33:81–86. http://dx.doi.org/10.1212/WNL.33.1.81

Borroni, B., F. Ferrari, D. Galimberti, B. Nacmias, C. Barone, S. Bagnoli, C. Fenoglio, I. Piaceri, S. Archetti, C. Bonvicini, et al. 2014. Heterozygous TREM2 mutations in frontotemporal dementia. Neurobiol. Aging. 35:934.e7–934.e10. http://dx.doi.org/10.1016/j.neurobiolaging.2013.09.017

Butovsky, O., M.P. Jedrychowski, C.S. Moore, R. Cialic, A.J. Lanser, G. Gabriely, T. Koeglsperger, B. Dake, P.M. Wu, C.E. Doykan, et al. 2014. Identification of a unique TGF-–dependent molecular and functional signature in microglia. Nat. Neurosci. 17:131–143. http://dx .doi.org/10.1038/nn.3599

Cady, J., E.D. Koval, B.A. Benitez, C. Zaidman, J. Jockel-Balsarotti, P. Allred, R.H. Baloh, J. Ravits, E. Simpson, S.H. Appel, et al. 2014. TREM2 variant p.R47H as a risk factor for sporadic amyotrophic lat-eral sclerosis. JAMA Neurol. 71:449–453. http://dx.doi.org/10.1001/ jamaneurol.2013.6237

Chiu, I.M., E.T. Morimoto, H. Goodarzi, J.T. Liao, S. O’Keeffe, H.P. Phatnani, M. Muratet, M.C. Carroll, S. Levy, S. Tavazoie, et al. 2013. A neurodegeneration-specific gene-expression signature of acutely iso-lated microglia from an amyotrophic lateral sclerosis mouse model. Cell Reports. 4:385–401. http://dx.doi.org/10.1016/j.celrep.2013.06.018

Colonna, M. 2003. TREMs in the immune system and beyond. Nat. Rev. Immunol. 3:445–453. http://dx.doi.org/10.1038/nri1106

Cramer, P.E., J.R. Cirrito, D.W. Wesson, C.Y. Lee, J.C. Karlo, A.E. Zinn, B.T. Casali, J.L. Restivo, W.D. Goebel, M.J. James, et al. 2012.

was neutralized and analyzed as the insoluble A fraction. Samples were ana-lyzed by sandwich ELISA using 6E10 as the capture antibody and A1-42 as detection antibody (Covance) as previously described (Cramer et al., 2012).

IHC. Mice were deeply anesthetized with avertin and perfused with PBS. Brains were drop-fixed in 4% paraformaldehyde in PBS and cryoprotected in 30% sucrose. Brains were embedded in OCT, and free-floating 30-µm sagittal sections were collected and stored at 4°C in PBS. For 3,3- diaminobenzidine (DAB) staining, endogenous peroxidases were quenched by incubating sections in 1% H2O2 in PBS for 30 min. Sections were blocked in 5% NGS/0.3% Triton X-100 in 1× PBS for 1 h. The following primary antibodies were added overnight at 4°C: Iba1 (1:1,000; Wako Pure Chemical Industries), AT8 (1:500; Thermo Fisher Scientific), AT180 (1:500; Thermo Fisher Scientific), CD11b (1:500; EMD Millipore), CD45 (1:500; AbD Serotec), CD68 (1:500; AbD Serotec), TREM2 (1:100; R&D Systems), and P2RY12 (1:2,000; a gift from H. Weiner, Brigham and Women’s Hospital, Boston, MA). To block nonspecific staining with antibodies generated in mouse and rat, Mouse on Mouse Blocking Reagent (Vector Laboratories) was used at 1 µl/ml of block. Slices were incubated with appropriate bio-tinylated secondary antibodies (1:200; Vector Laboratories) and VECTA-STAIN Elite ABC kit (Vector Laboratories) and developed with DAB with nickel chloride. Indicated slices were counterstained with Congo red to vi-sualize dense core A plaques. Slices were mounted with Permount (Thermo Fisher Scientific).

TREM2 immunofluorescence was performed as described above for DAB staining except incubation with ABC was followed with incubation in the TSA Biotin System kit (PerkinElmer) and incubation with SA-488 (1:200; Life Technologies). Slices were then incubated with Iba1 (1:500; Wako Pure Chemical Industries), 6E10 (1:500; Covance), CT15 (1:250; a gift from E.H. Koo), GFAP (1:500; Sigma-Aldrich), MAP2 (1:500; EMD Millipore), or MBP (1:200; Abcam), followed by species-specific Alexa sec-ondary antibodies (1:1,000; Life Technologies) and mounted using Vecta-shield Hard Set mounting media (Vector Laboratories).

-Galactosidase activity was assessed in 30 µm free-floating sections with an X-gal staining solution (1 mg/ml X-gal, 5 mM potassium ferricya-nate, 5 mM potassium ferrocyante, and 25 µM sodium deoxycholate). Slices were incubated in the solution overnight at 37°C, washed in PBS, costained, and mounted with Permount.

Brightfield images were taken on a DMLS microscope (Leica), using QImaging camera (QImaging) using QCapture Software (QImaging). For quantification of plaque area, whole brain sections were imaged on the SCN400F slide scanner (Leica) with SCN Client Software (Leica) and ana-lyzed using Image Pro Plus Software (Media Cybernetics). Confocal images were taken on a LSM 510 META microscope (Carl Zeiss). 12–20 slices, 1 µm apart, were imaged and z-stacks were reconstructed in ImageJ.

Flow cytometry. Myeloid cells were isolated from APPPS1 and 5XFAD mice along with age-matched controls. Mice were perfused with HBSS. Brains were chopped and digested in papain (Roche) at 37°C with shaking for 30 min. Homogenates were passed through 18- and 22-gauge needles and transferred into 20% FBS. Isotonic Percoll was added to create a 30% solution, and 1 ml of 70% Percoll was underlaid. This was centrifuged at 800 g for 30 min, and 2–3 ml of the 70%/30% interface was collected and washed. Samples were resuspended in FACS buffer (PBS, 1% BSA, 0.1% NaN3, and 5 mM EDTA) and blocked with a CD16/CD32 antibody (1:200; BD) for 10 min. Pooled portions of samples were used for unstained and single-stained control, whereas individual samples were stained with a master mix of CD45/AX700 (1:500; BioLegend), CD11b/BV605 (1:500; BioLegend), F4/80/PEefluor610 (1:125; eBioscience), and Ly6C/PECy7 (1:50,000; BioLegend). Each sample was divided in two tubes, a fluorescence minus one “FMO” control and an “ALL” sample to which TREM2/APC (1:125; R&D Systems) was added. Events were acquired on a Fortessa SORP (BD) and analyzed using FlowJo. For analysis, events were gated on single cells and CD11b-positive events. Samples with >5,000 CD11b-positive events were used for analysis. This population was divided into CD45lo and CD45hi

JEM Vol. 212, No. 3 295

Br ief Definit ive Repor t

ApoE-directed therapeutics rapidly clear -amyloid and reverse defi-cits in AD mouse models. Science. 335:1503–1506. http://dx.doi.org/ 10.1126/science.1217697

Forabosco, P., A. Ramasamy, D. Trabzuni, R. Walker, C. Smith, J. Bras, A.P. Levine, J. Hardy, J.M. Pocock, R. Guerreiro, et al. 2013. Insights into TREM2 biology by network analysis of human brain gene expres-sion data. Neurobiol. Aging. 34:2699–2714. http://dx.doi.org/10.1016/ j.neurobiolaging.2013.05.001

Frank, S., G.J. Burbach, M. Bonin, M. Walter, W. Streit, I. Bechmann, and T. Deller. 2008. TREM2 is upregulated in amyloid plaque-associatedmicroglia in aged APP23 transgenic mice. Glia. 56:1438–1447. http://dx.doi.org/10.1002/glia.20710

Guerreiro, R., A. Wojtas, J. Bras, M. Carrasquillo, E. Rogaeva, E. Majounie, C. Cruchaga, C. Sassi, J.S. Kauwe, S. Younkin, et al. Alzheimer GeneticAnalysis Group. 2013. TREM2 variants in Alzheimer’s disease. N. Engl.J. Med. 368:117–127. http://dx.doi.org/10.1056/NEJMoa1211851

Guerreiro, R.J., E. Lohmann, J.M. Brás, J.R. Gibbs, J.D. Rohrer, N. Gurunlian, B. Dursun, B. Bilgic, H. Hanagasi, H. Gurvit, et al. 2013. Using exome sequencing to reveal mutations in TREM2 presenting as a fron-totemporal dementia-like syndrome without bone involvement. JAMA Neurol. 70:78–84. http://dx.doi.org/10.1001/jamaneurol.2013.579

Hickman, S.E., N.D. Kingery, T.K. Ohsumi, M.L. Borowsky, L.C. Wang, T.K. Means, and J. El Khoury. 2013. The microglial sensome revealed by direct RNA sequencing. Nat. Neurosci. 16:1896–1905. http://dx.doi .org/10.1038/nn.3554

Jiang, T., L. Tan, X.C. Zhu, Q.Q. Zhang, L. Cao, M.S. Tan, L.Z. Gu, H.F. Wang, Z.Z. Ding, Y.D. Zhang, and J.T. Yu. 2014. Upregulation of TREM2 ameliorates neuropathology and rescues spatial cognitive impairment in a transgenic mouse model of Alzheimer’s disease. Neuropsychopharmacology. 39:2949–2962. http://dx.doi.org/10.1038/npp.2014.164

Jonsson, T., H. Stefansson, S. Steinberg, I. Jonsdottir, P.V. Jonsson, J. Snaedal, S. Bjornsson, J. Huttenlocher, A.I. Levey, J.J. Lah, et al. 2013. Variantof TREM2 associated with the risk of Alzheimer’s disease. N. Engl. J.Med. 368:107–116. http://dx.doi.org/10.1056/NEJMoa1211103

Karch, C.M., and A.M. Goate. 2015. Alzheimer’s disease risk genes and mechanisms of disease pathogenesis. Biol. Psychiatry. 77:43–51. http://dx.doi.org/10.1016/j.biopsych.2014.05.006

Kleinberger, G., Y. Yamanishi, M. Suárez-Calvet, E. Czirr, E. Lohmann, E. Cuyvers, H. Struyfs, N. Pettkus, A. Wenninger-Weinzierl, F. Mazaheri, et al. 2014. TREM2 mutations implicated in neurodegeneration im-pair cell surface transport and phagocytosis. Sci. Transl. Med. 6:243ra86. http://dx.doi.org/10.1126/scitranslmed.3009093

Lee, S., G. Xu, T.R. Jay, S. Bhatta, K.W. Kim, S. Jung, G.E. Landreth, R.M. Ransohoff, and B.T. Lamb. 2014. Opposing effects of membrane- anchored CX3CL1 on amyloid and tau pathologies via the p38 MAPKpathway. J. Neurosci. 34:12538–12546. http://dx.doi.org/10.1523/JNEUROSCI.0853-14.2014

Mildner, A., B. Schlevogt, K. Kierdorf, C. Böttcher, D. Erny, M.P. Kummer, M. Quinn, W. Brück, I. Bechmann, M.T. Heneka, et al. 2011. Distinctand non-redundant roles of microglia and myeloid subsets in mousemodels of Alzheimer’s disease. J. Neurosci. 31:11159–11171. http://dx.doi.org/10.1523/JNEUROSCI.6209-10.2011

Neumann, H., and K. Takahashi. 2007. Essential role of the microglial triggering receptor expressed on myeloid cells-2 (TREM2) for cen-tral nervous tissue immune homeostasis. J. Neuroimmunol. 184:92–99. http://dx.doi.org/10.1016/j.jneuroim.2006.11.032

Oakley, H., S.L. Cole, S. Logan, E. Maus, P. Shao, J. Craft, A. Guillozet-Bongaarts, M. Ohno, J. Disterhoft, L. Van Eldik, et al. 2006. Intraneuronal -amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J. Neurosci. 26:10129–10140. http://dx.doi.org/10.1523/JNEUROSCI.1202-06.2006

Paradowska-Gorycka, A., and M. Jurkowska. 2013. Structure, expression pattern and biological activity of molecular complex TREM-2/DAP12. Hum. Immunol. 74:730–737. http://dx.doi.org/10.1016/j.humimm.2013 .02.003

Prinz, M., and J. Priller. 2014. Microglia and brain macrophages in the mo-lecular age: from origin to neuropsychiatric disease. Nat. Rev. Neurosci. 15:300–312. http://dx.doi.org/10.1038/nrn3722

Radde, R., T. Bolmont, S.A. Kaeser, J. Coomaraswamy, D. Lindau, L. Stoltze, M.E. Calhoun, F. Jäggi, H. Wolburg, S. Gengler, et al. 2006. A42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology. EMBO Rep. 7:940–946. http://dx.doi.org/10.1038/ sj.embor.7400784

Rayaprolu, S., B. Mullen, M. Baker, T. Lynch, E. Finger, W.W. Seeley, K.J. Hatanpaa, C. Lomen-Hoerth, A. Kertesz, E.H. Bigio, et al. 2013. TREM2 in neurodegeneration: evidence for association of the p.R47H variant with frontotemporal dementia and Parkinson’s disease. Mol. Neurodegener. 8:19. http://dx.doi.org/10.1186/1750-1326-8-19

Riddell, D.R., H. Zhou, T.A. Comery, E. Kouranova, C.F. Lo, H.K. Warwick, R.H. Ring, Y. Kirksey, S. Aschmies, J. Xu, et al. 2007. The LXR agonist TO901317 selectively lowers hippocampal A42 and improves memory in the Tg2576 mouse model of Alzheimer’s disease. Mol. Cell. Neurosci. 34:621–628. http://dx.doi.org/10.1016/ j.mcn.2007.01.011

Sedgwick, J.D., S. Schwender, H. Imrich, R. Dörries, G.W. Butcher, and V. ter Meulen. 1991. Isolation and direct characterization of resident mi-croglial cells from the normal and inflamed central nervous system. Proc. Natl. Acad. Sci. USA. 88:7438–7442. http://dx.doi.org/10.1073/pnas .88.16.7438

Sharif, O., R. Gawish, J.M. Warszawska, R. Martins, K. Lakovits, A. Hladik, B. Doninger, J. Brunner, A. Korosec, R.E. Schwarzenbacher, et al. 2014. The triggering receptor expressed on myeloid cells 2 inhibits complement component 1q effector mechanisms and exerts detrimental effects during pneumococcal pneumonia. PLoS Pathog. 10:e1004167. http://dx.doi.org/10.1371/journal.ppat.1004167

Sieber, M.W., N. Jaenisch, M. Brehm, M. Guenther, B. Linnartz-Gerlach, H. Neumann, O.W. Witte, and C. Frahm. 2013. Attenuated inflam-matory response in triggering receptor expressed on myeloid cells 2(TREM2) knock-out mice following stroke. PLoS ONE. 8:e52982.http://dx.doi.org/10.1371/journal.pone.0052982

Ulrich, J.D., M.B. Finn, Y. Wang, A. Shen, T.E. Mahan, H. Jiang, F.R. Stewart, L. Piccio, M. Colonna, and D.M. Holtzman. 2014. Altered microglialresponse to A plaques in APPPS1-21 mice heterozygous for TREM2. Mol. Neurodegener. 9:20. http://dx.doi.org/10.1186/1750-1326-9-20

Wyss-Coray, T. 2006. Inflammation in Alzheimer disease: driving force, bystander or beneficial response? Nat. Med. 12:1005–1015.

Zhang, B., C. Gaiteri, L.G. Bodea, Z. Wang, J. McElwee, A.A. Podtelezhnikov, C. Zhang, T. Xie, L. Tran, R. Dobrin, et al. 2013. Integrated sys-tems approach identifies genetic nodes and networks in late-onsetAlzheimer’s disease. Cell. 153:707–720. http://dx.doi.org/10.1016/j.cell.2013.03.030

JEM Vol. 212, No. 12 1991

How dominant mutations in presenilin (PSEN) cause early-onset familial Alzheimer’s disease (FAD) has been debated since the discovery of such mutations 20 years ago. A study by Szaruga et al. in this issue of JEM now appears to provide a definitive answer.

Presenilin is the catalytic subunit of γ-secretase, a protease that cuts the transmembrane domain ofthe amyloid precursor protein (APP) to produce the C terminus of the amyloid β-peptide (Aβ) thatnotoriously deposits in the Alzheimer brain. Some argue that reduction of presenilin’s proteolytic activ-ity (i.e., a loss-of-function effect) is responsible for the neurodegeneration caused by FAD mutations. Others have shown that some mutations do not reduce proteolytic activity, but all increase the propor-tion of aggregation-prone 42-residue Aβ (Aβ42) to 40-residue Aβ (Aβ40; i.e., a gain of toxic function).Further complicating matters, γ-secretase initially cuts APP substrate via an endopeptidase activity toproduce Aβ48 and Aβ49 and release the corresponding APP intracellular domain (AICD). These long

Aβ peptides are then sequentially trimmed via a carboxypeptidase function of γ-secretase along two primary pathways:Aβ49 → Aβ46 → Aβ43 → Aβ40 and Aβ48 → Aβ45 → Aβ42 → Aβ38.

To address the loss- versus gain-of-function question, Szaruga et al. examined γ-secretase proteolytic activity in samples frompost-mortem human brains from 24 FAD mutation carriers, covering nine different PSEN mutations. The samples contained endogenous human γ-secretase complexes and—importantly—both wild-type and PSEN mutant complexes. Under thesenatural conditions associated with the human disease state, the production of AICD—a measure of γ-secretase endoprotease activ-ity—was not significantly different from that seen in control non-AD brains. Thus, the presence of the wild-type PSEN allele apparently compensates for any loss of endoproteolytic activity from the mutant allele. In contrast, clear reduction of car-boxypeptidase activity—as measured by the ratio of Aβ38 from its precursorAβ42—was seen for every mutation.

These findings have implications for the mechanism of Alzheimer pathogen-esis and for drug discovery. In considering γ-secretase as a therapeutic target, oneshould first know what specific functional alterations in the enzyme lead to dis-ease, and that appears to be decreased carboxypeptidase activity. Therefore, a search for stimulators of this activity would make sense. Such compounds have al-ready been identified, although they appear to stimulate only the Aβ42 → Aβ38step, insufficient if other long Aβ pep-tides are augmented in Alzheimer’s and play pathogenic roles. As is so often the case, answering one key question leads to another.

Szaruga, M., et al. 2015. J. Exp. Med. http://dx.doi.org/10.1084/jem.20150892

Ab is derived from its precursor protein APP by sequential proteolysis, first by β-secretase (not depicted) and then by γ-secretase, the latter hydrolyzing within the transmembrane (TM) domain. Initial cleavage occurs at the so-called ε site (indicated by the scissors), releasing the APP intracellular domain or AICD (red intracellular piece) and leaving Aβ49 or Aβ48 fragments in the membrane. Aβ49 or Aβ48 fragments are successively cut by the carboxypeptidase-like activity of γ-secretase, increasing the probability of release from the plasma membrane to the extracellular medium. Both ε cleavage and carboxypeptidase TM trimming depend on PSEN, the catalytic subunit of γ-secretase. Pathogenic mutations in PSEN cause a qualitative shift in Aβ profile production, increasing the proportion of released longer Aβ peptides, which are prone to aggregate and form the plaques observed FAD.

Michael S. Wolfe, Brigham and Women’s Hospital and Harvard Medical School: mswolfe@partners.org

Insight from Michael Wolfe

Cutting to the chase: How pathogenic mutations cause Alzheimer’s

INSIGHTS

Br ief Definit ive Repor t

2003The Rockefeller University Press $30.00J. Exp. Med. 2015 Vol. 212 No. 12 2003–2013www.jem.org/cgi/doi/10.1084/jem.20150892

Early-onset familial Alzheimer’s disease (AD [FAD]), starting before age 65, is mainly caused by mutations in the Preseni-lin 1/2 (PSEN1/2) or the amyloid precursor protein (APP) genes and represents less than 0.1% of the total AD cases (Campion et al., 1999). Although rare, FAD offers a unique model to gain insights into the molecular mechanisms and etiology of sporadic AD (SAD).

PSEN is the catalytic subunit of the γ-secretase com-plex (De Strooper et al., 1998; Wolfe et al., 1999), an intram-embrane multimeric protease involved in the processing of many type 1 transmembrane proteins; among them, the Notch receptors and APP have received much attention be-

cause of their association with crucial cell signaling events or with AD pathogenesis, respectively (for a review see Ju-risch-Yaksi et al. [2013]). Nicastrin (Nct), PSEN enhancer 2 (Pen2), and anterior pharynx defective 1 (APH1) are, to-gether with PSEN, essential components of the protease complex (De Strooper, 2003).

More than 150 pathogenic mutations in PSEN1 have been reported so far (http://www.molgen.ua.ac.be/ADMutations); and notably, the vast majority are missense substitutions distributed throughout the primary structure of PSEN1. PSEN/γ-secretase hydrolyzes peptide bonds ina process called regulated intramembrane proteolysis, which allows translation of extracellular signals into the cell. Com-pelling evidence indicates that γ-secretase cuts membraneproteins sequentially: the first endopeptidase cleavage (ε)releases a soluble intracellular domain (ICD), which may

Presenilin (PSEN) pathogenic mutations cause familial Alzheimer’s disease (AD [FAD]) in an autosomal-dominant manner. The extent to which the healthy and diseased alleles influence each other to cause neurodegeneration remains unclear. In this study, we assessed γ-secretase activity in brain samples from 15 nondemented subjects, 22 FAD patients harboring nine dif-ferent mutations in PSEN1, and 11 sporadic AD (SAD) patients. FAD and control brain samples had similar overall γ-secretase activity levels, and therefore, loss of overall (endopeptidase) γ-secretase function cannot be an essential part of the patho-genic mechanism. In contrast, impaired carboxypeptidase-like activity (γ-secretase dysfunction) is a constant feature in all FAD brains. Significantly, we demonstrate that pharmacological activation of the carboxypeptidase-like γ-secretase activity with γ-secretase modulators alleviates the mutant PSEN pathogenic effects. Most SAD cases display normal endo- and car-boxypeptidase-like γ-secretase activities. However and interestingly, a few SAD patient samples display γ-secretase dysfunc-tion, suggesting that γ-secretase may play a role in some SAD cases. In conclusion, our study highlights qualitative shifts in amyloid-β (Aβ) profiles as the common denominator in FAD and supports a model in which the healthy allele contributes with normal Aβ products and the diseased allele generates longer aggregation-prone peptides that act as seeds inducing toxic am-yloid conformations.

Qualitative changes in human γ-secretase underlie familialAlzheimer’s disease

Maria Szaruga,1,2* Sarah Veugelen,1,2* Manasi Benurwar,1,2 Sam Lismont,1,2 Diego Sepulveda-Falla,3,4 Alberto Lleo,5,6 Natalie S. Ryan,7 Tammaryn Lashley,8 Nick C. Fox,7 Shigeo Murayama,10 Harrie Gijsen,11 Bart De Strooper,1,2,9 and Lucía Chávez-Gutiérrez1,2

1VIB Center for the Biology of Disease and 2Center for Human Genetics (CME) and Leuven Research Institute for Neuroscience and Disease (LIND), University of Leuven (KU Leuven), 3000 Leuven, Belgium

3Institut für Neuropathologie, Universitätsklinikum Hamburg-Eppendorf, 20246 Hamburg, Germany4Neuroscience Group of Antioquia, Faculty of Medicine, University of Antioquia, Medellín 1226, Colombia5Unidad de Memoria, Departamento de Neurología, Institut d’Investigacions Biomèdiques Sant Pau, Hospital de Sant Pau, 08025 Barcelona, Spain6Centro Investigación Biomédica en Red Enfermedades Neurodegenerativas (CIB ERNED), 28049 Madrid, Spain7Dementia Research Centre, 8Queen Square Brain Bank for Neurological Disorders, Department of Molecular Neuroscience, and 9Department of Molecular Neuroscience, Institute of Neurology, University College London, London WC1N 3AR, England, UK

10Tokyo Metropolitan Institute of Gerontology, Itabashi-ku, Tokyo 173-0015, Japan11Janssen Research and Development Division, Janssen Pharmaceutica NV, 2340 Beerse, Belgium

© 2015 Szaruga et al. This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http://creativecommons.org/licenses/by-nc-sa/3.0/).

*M. Szaruga and S. Veugelen contributed equally to this paper.

Correspondence to Lucía Chávez-Gutiérrez: Lucia.ChavezGutierrez@cme.vib-kuleuven.be; or Bart De Strooper: Bart.DeStrooper@cme.vib-kuleuven.be

Abbreviations used: Aβ, amyloid-β; AD, Alzheimer’s disease; APP, amyloid precursorprotein; DRM, detergent-resistant membrane; FAD, familial AD; GSM, γ-secretase modulator; ICD, intracellular domain; LOF, loss-of-function; Nct, Nicastrin; PSEN, Presenilin; SAD, sporadic AD.

Aβ production in familial Alzheimer’s brain | Szaruga et al.2004

translocate to the nucleus to regulate gene expression while the remaining N-terminal transmembrane domain (TMD) fragment is successively cut by the carboxypeptidase-like ac-tivity of γ-secretase (γ-cleavages). Endopeptidase products,either amyloid-β49 (Aβ49) or Aβ48, are then processed alongtwo major product lines: Aβ49 → Aβ46 → Aβ43 → Aβ40or Aβ48 → Aβ45 → Aβ42 → Aβ38 (Takami et al., 2009).Every γ-cleavage removes a short C-terminal peptide fromthe TMD, reducing its hydrophobicity and increasing the probability of release. Secretion of an N-terminal fragment into the extracellular/luminal space terminates this sequence (Qi-Takahara et al., 2005; Yagishita et al., 2008; Takami et al., 2009). Importantly, the efficiency of the endopeptidase cleavage determines ICD product levels, which acquires high physiological relevance in the case of the Notch substrate. The carboxypeptidase-like efficiency, the number of cuts per substrate, determines the length of the N-terminal products; the level of efficiency is pathologically very relevant in the case of the APP substrate, as lower efficiency results in the production of longer and more aggregation-prone Aβ pep-tides (Chávez-Gutiérrez et al., 2012).

How mutations in the PSENs cause FAD remains a hotly debated topic in the field. Because FAD is an autoso-mal-dominant disorder (patients carry both healthy and mu-tant alleles), a major unknown in the discussion remains the role of the healthy allele and, to a lesser extent, the role of the brain environment on the total (normal + mutant pro-teases) γ-secretase activity. To what degree do normal andmutant complexes contribute to total γ-secretase activity inthe patient brain? Does the healthy allele compensate for the disease allele effects? Despite their relevance, those questions have not been addressed.

Only one group, i.e., Potter et al. (2013), have estimated the Aβ production kinetics in the FAD brain by measuringisotope-labeled Aβ peptides in the cerebrospinal fluid of pa-tients (stable isotope-labeled kinetics [SILK]). Feeding the in vivo metabolic labeling patient data into a mathematical model, specifically generated for their approach, they de-scribed higher Aβ42 production rates in the central nervoussystem of PSEN mutation carriers (Potter et al., 2013). Ac-cordingly, Potter et al. (2013) suggest that increments in Aβ42play a decisive pathogenic role in AD.

In contrast, a “revised” loss-of-function (LOF) hypoth-esis has recently been proposed by Xia et al. (2015). In this view, loss of PSEN/γ-secretase physiological cell signalingfunction causes neurodegeneration, whereas changes in Aβpeptides are only secondary byproducts that arise from but do not trigger the disease (Xia et al., 2015). The idea is only tenable if FAD-linked PSEN mutations exert a LOF effect on PSEN/γ-secretase and, in addition, a dominant-negativeeffect on the healthy PSEN allele (normal γ-secretase) in pa-tients, a key part of this hypothesis (Heilig et al., 2013; Xia et al., 2015). However, it should be stressed that γ-secretase hap-loinsufficiency caused by nonsense, frameshift, and splice site mutations in genes coding for essential subunits of γ-secretase

(Nct, Pen2, and PSEN) is pathogenic in nature; such haplo-insufficiency causes a chronic inflammatory disease of hair follicles known as familial acne inversa. Most importantly, no clinical association between this disorder and AD has been reported (for a review see Pink et al. [2013]). Furthermore, if FAD-linked PSEN mutations were truly LOF mutations, resulting in “inactive” γ-secretase complexes, homozygousindividuals for the disease allele would not be viable because of disturbances in Notch signaling during embryonic devel-opment. However, six individuals with homozygous PSEN1 E280A gene mutation have been identified (Kosik et al., 2015).

An alternative view to both hypotheses is that pathogenic mutations in PSEN cause disease by qualitative shifts in Aβprofile production (γ-secretase dysfunction; Chávez-Gutiérrezet al., 2012). We have demonstrated that loss of endopeptidase activity is not necessarily observed in γ-secretase complexescontaining PSEN1/2 FAD-linked mutations, but reduced car-boxypeptidase-like efficiency (γ-secretase dysfunction) is theconstant denominator. Furthermore, FAD PSEN mutations may affect the carboxypeptidase-like γ-secretase activity atmultiple turnovers, resulting in increased Aβ43 and Aβ42 levelsas well as in other longer Aβ peptides, such as Aβ45 and Aβ46(Quintero-Monzon et al., 2011; Chávez-Gutiérrez et al., 2012; Fernandez et al., 2014). These data support a model in which relative, rather than absolute, changes in Aβ product profilesare at the basis of PSEN/γ-secretase–mediated pathogenicity. However, these findings were based on studies conducted in PSEN1/2-deficient MEFs, which does not fully recapitulate the in vivo heterozygous situation in the FAD patient’s brain.

In the current study, we investigated processing of APP by the γ-secretase complex in postmortem human brain sam-ples from FAD and SAD patients and healthy control sub-jects. Our investigation is the first to directly assess γ-secretaseactivity in brain material from FAD mutant carriers and to address how the FAD-linked mutant heterozygous situation in patients affects γ-secretase function in brain.

RES ULTS AND DIS CUSSI ONAβ production rates in FAD and SAD brainsWe aimed to evaluate the effects of pathogenic PSEN1 mu-tants on total γ-secretase activity (healthy and disease PSEN1 alleles) in human brain samples from FAD patients. As a first step, we sought to determine and contrast the production rates of Aβ peptides in (a) human control brains, i.e., con-taining two healthy PSEN1 alleles; (b) FAD brains carry-ing pathogenic mutations in PSEN1, heterozygous for the PSEN1 alleles; and (c) SAD brains, with two healthy PSEN1 alleles. Aβ peptides are generated from APP-C99 membranepeptide by consecutive γ-secretase proteolytic cleavages. Thefirst endopeptidase cut (ε) releases a soluble ICD (AICD)and generates a long membrane-associated Aβ peptide, eitherAβ49 or Aβ48, which are then processed along two majorproduct lines (Takami et al., 2009).

Active γ-secretase is associated with detergent-resistantmembranes (DRMs; Wahrle et al., 2002) and DRMs pre-

2005JEM Vol. 212, No. 12

pared from brain or cells are a bona-fide source of γ-secre-tase activity (Matsumura et al., 2014). Thus, we prepared DRMs from the prefrontal cortices of 15 control brain sam-ples, 22 FAD brain samples carrying nine different patho-logical PSEN1 mutations, V89L (1 case), intron 4 (2 cases), E120G (1 case), M139T (3 cases), I202F (1 case), P264L (2 cases), R278I (1 case), E280A (10 cases), and L286P (1 case; Table 1); and 11 SAD brain samples. DRMs were used as a source of the enzyme in in vitro activity assays. Importantly, DRMs contain γ-secretase in its native environment (themembrane) while maintaining the lipid composition of their origin (cells or brain).

The production rates of the Aβ38, Aβ40, and Aβ42 pep-tides (de novo) were determined by incubation of equivalent amounts of brain DRMs with the purified APPC99-3×FLAG γ-secretase substrate, under saturating conditions for 0 and4 h. Our results show a reduction in the Aβ38 production ratein the majority of FAD cases and a reduction in Aβ40 gener-ation in five of nine FAD cases (Fig. 1, A and B). In contrast, Aβ production in SAD samples did not show significant alter-ations (Fig. 1, A–C). However, the dispersion observed in the SAD group may be an indication of changes in γ-secretaseactivity in a fraction of late-onset AD cases (see next section for further discussion). Most interestingly, there were no dif-ferences in Aβ42 production rates in most of the FAD brainsamples, except for the intron 4 mutation cases, which display, on average, a 2.5-fold increment over controls (Fig. 1 C). A mild reduction in Aβ42 production rate was observed in theE120G-, P264L-, and R278I-PSEN1 brain samples, but the differences did not reach statistical significance. The γ-secre-

tase inhibitor X (a transition state analogue) abolished the production of Aβ peptides (not depicted), demonstrating thespecificity of the reaction. Unfortunately, the concentration of “de novo” generated Aβ37 and Aβ43 peptides in control andmost of the mutant samples were below the detection limits.

To correct for any potential differences in protein con-centrations in our assays, we also normalized Aβ productionto flotilin-1 levels, determined by immunoblot, which did not change our observations (not depicted). Collectively, our data support a model in which relative changes in Aβ pro-duction in FAD are more important for disease than abso-lute increments in Aβ42 levels (Tanzi and Bertram, 2005; DeStrooper, 2007; Kuperstein et al., 2010; Chávez-Gutiérrez et al., 2012). Thus, our findings contrast with those of Potter et al. (2013). The differences may arise from the fact that the SILK method measures Aβ released in the interstitial fluid,which provides an indirect and perhaps not so accurate assess-ment of γ-secretase activity in human brain. Although theirmathematical model should in principle correct for these po-tential issues, several not yet experimentally verified assump-tions were made with regard to production, secretion, and clearance mechanisms of Aβ peptides (for further discussionof the model see Edland and Galasko [2011]). For instance, it is assumed that different Aβs are generated independentlyfrom each other (Potter et al., 2013), which is not in line with the current knowledge showing that consecutive γ-secretasecleavages generate Aβ peptides (Takami et al., 2009). Further-more, we would like to draw attention to the heterogeneous behavior of the mutation carrier cohort, reported as a proof of concept in Potter et al. (2013): the Aβ42 production rateswere actually only elevated in three out of seven FAD-linked PSEN mutation carriers.

De novo production of AICD in FAD-PSEN brain samplesOur results indicate that Aβ38, Aβ40, and Aβ42 are the mainsecreted products, and the sum of “de novo” Aβ38, Aβ40, andAβ42 products reveals lower Aβ production in brain sam-ples carrying pathogenic mutations in PSEN1 in five out of nine cases, relative to control and SAD cases (Fig. 1 D). The observed effect on Aβ production could be caused by areduction in the endopeptidase activity or an impaired car-boxypeptidase-like efficiency in FAD brain samples. There-fore, we analyzed the production of AICD in our samples, which provides a relative indication for the efficiency of the γ-secretase endopeptidase activity in the tested human brainsamples. Decreased γ-secretase endopeptidase efficiency mayresult in alterations in cell signaling events involved in cel-lular communication (LOF). We found no significant differ-ences in the production rates of AICD (de novo AICD) in controls and FAD brain samples, although a slight reduction in AICD production rate was observed in the R278I brain sample (Fig. 1, E and F). These data clearly indicate that the overall γ-secretase endoprotease activity is unaffected in mostof the FAD brain samples, and therefore, this effect cannot be an essential part of the pathogenic mechanism.

Table 1. Clinical data of FAD patients whose brains were analyzed in this study

Mutation Sex Diagnosis Age of onset Age at death APOE

yr yrV89L M FAD 48      57 23Intron 4 F FAD 35      51.9 44Intron 4 F FAD 36      41.6 33E120G M FAD 34      44 33M139T M FAD 47      64 33M139T M FAD 48      57 33M139T M FAD 45      53 33I202F F FAD 48      59.3 44P264L F FAD 45      56 44P264L M FAD 53      60 34R278I F FAD 46      65.6 34L286P F FAD 35      56 33E280A F FAD 47      54 33E280A M FAD 44      52 33E280A M FAD 54      63 34E280A M FAD 47      56 33E280A F FAD 46      67 34E280A F FAD 48      64 33E280A F FAD 43      48 33E280A F FAD 50      60 33E280A F FAD 52      68 33E280A M FAD 47      58 33

Aβ production in familial Alzheimer’s brain | Szaruga et al.2006

Figure 1. APP processing in human brain samples from FAD, SAD, and control patients. (A–D) Aβ production rate in brains of FAD or SAD patientscompared with nondemented subjects. To determine de novo production of Aβ peptides, CHA PSO-resistant membranes prepared from brain tissue ofpatients were incubated with 1.5 µM C99-3×FLAG substrate and quantified using MSD ELI SA technology. Graphs show mean ± SD for groups with one case or mean of means ± SD for groups with number of cases greater than one. (E) SDS-PAGE/Western blot showing AICD product levels in reactions with human control and FAD brains. The molecular mass of AICD-3×FLAG is ∼10 kD. (F) De novo AICD product levels (endopeptidase activity levels) in humancontrol and FAD brain samples. Graph shows mean ± SE for groups with one case or mean of means ± SE for groups with number of cases greater than one. All experiments were repeated at least three times, and statistical significance was tested with one-way ANO VA and Dunnett's post test, taking the corresponding WT set as the control group (**, P < 0.01; *, P < 0.05).

2007JEM Vol. 212, No. 12

We have previously shown that pathogenic PSEN1 mu-tants display variable effects (including no effect) on the en-dopeptidase efficiency of mutant γ-secretase complexes butconsistently reduce the γ-secretase carboxypeptidase-like effi-ciency (γ-secretase dysfunction), all relative to the normal en-zyme (Chávez-Gutiérrez et al., 2012). To evaluate the effects of the pathogenic mutations analyzed here on γ-secretase ac-tivity, we stably expressed the different clinical mutant PSENs in a Psen1/2 knockout background (Fig. 3). With the excep-tion of the R278I PSEN1 mutant, which severely impairs ac-tivation of the γ-secretase complex, Psen1/2 knockout MEFstransduced with normal or mutant PSEN1s express com-parable levels of the mature γ-secretase complex (Fig. 3 A).Specific endopeptidase activities, defined as “de novo AICD” normalized against PSEN1-CTF subunit levels in the in vitro reactions, revealed variable effects (including no effect) of the PSEN1 mutations on the endopeptidase function of γ-secre-tase (Fig.  3 B). Interestingly, comparison of the endopepti-dase activities of samples containing either normal or mutant complexes (Fig.  3  B) with FAD brain samples (normal + mutant enzymes; Fig. 1 F) demonstrates that the healthy al-lele (normal enzyme) compensates for (if any) decrements in the endopeptidase cleavage rates caused by the disease allele (mutant γ-secretase complex). The above implies that normaland mutant proteases contribute to γ-secretase endopeptidaseactivity. Undoubtedly, our data do not support a mutant-me-diated “dominant-negative effect” on the healthy allele and, therefore, contrast with the hypothesis proposed recently by Heilig et al. (2013) and Xia et al. (2015).

Carboxypeptidase-like efficiency in FAD and SAD brain samplesInterestingly, brain samples carrying pathogenic PSEN1 mu-tations consistently display lower “Aβ38 + Aβ40 + Aβ42” production rates (Fig. 1 D) than for AICD (Fig. 1 F), sug-gesting a higher production of longer Aβ peptides (>Aβ42)in FAD versus control brain samples, which may be indic-ative of impaired carboxypeptidase-like efficiency (γ-secre-tase dysfunction). γ-Cleavage efficiency can be assessed bydetermining the Aβ38/Aβ42 ratio, which represents theproduct/substrate ratio for the fourth catalytic turnover of the γ-secretase. Significantly, changes in this ratio correlatedirectly with this particular cleavage efficiency. Thus, we cal-culated the Aβ38/Aβ42 ratios of total γ-secretase in FAD,SAD, and control brain samples. In the FAD brain samples, the observed decrement in the short Aβ38 and Aβ40 peptidestranslated into a significant reduction in the Aβ38/Aβ42 ratio(Fig. 2 A). Remarkably, regardless of the nature and position of the mutation in PSEN1, the Aβ38/Aβ42 ratios revealeda consistent reduction in total γ-secretase carboxypepti-dase-like efficiencies in FAD brain samples, relative to con-trols. Our data thus show no alterations in the total (normal + mutant complexes) γ-secretase endopeptidase activity andlower de novo production of Aβ38 and Aβ40 peptides. Thisis not accompanied by increased Aβ42 production levels in

most of the FAD patient brain samples. These results strongly suggest that longer Aβ peptides (such as Aβ43 and Aβ45) areproduced in patient brain samples. Unfortunately, technical limitations do not allow us currently to measure the produc-tion of these longer peptides.

The amplitude of the effects on the Aβ38/Aβ42 ratiosdid not correlate with the age at onset (Table  1), suggest-ing that additional genetic and/or environmental factors may play a role in the onset of FAD, and we speculate that altered processing of other substrates could contribute to this. We also looked at the Aβ42/Aβ40 ratio because in-crements in this (somewhat deliberately chosen) ratio have been used as hallmark of FAD mutations for decades. Our data show increased Aβ42/Aβ40 ratios in six (V89L, intron4, M139T, I202F, R278I, and E280A) of the nine PSEN1 mutant cohorts analyzed.

With regard to SAD, our data reveal similar γ-secretasecarboxypeptidase-like efficiencies relative to control brain samples. However, in accordance with the Aβ productionvelocities (Fig. 1, A–D), we observed a marked Aβ38/Aβ42ratio dispersion among the SAD brain samples (Fig. 2 A). This may indicate alterations in γ-secretase carboxypeptidase-likeefficiency in a subset of late-onset SAD cases but exclude the hypothesis that changes in γ-secretase efficiency play amajor role in the majority of SAD patients. Application of in vivo metabolic labeling in SAD patients found no changes in Aβ production in late AD (Mawuenyega et al., 2010). How-ever, analysis of other γ-secretase products in cerebrospinalfluid of late-onset patients revealed the apparent existence of subpopulations of SAD patients showing differential al-terations in γ-secretase activity, including enzyme dysfunc-tion (Hata et al., 2011, 2012). In addition, increments in the carboxypeptidase-like efficiency of γ-secretase in SAD brainsamples have also been reported (Kakuda et al., 2012). Thus, our data support the idea that the causes of late-onset AD are heterogeneous and raise the possibility that SAD patients showing γ-secretase dysfunction could benefit from γ-secre-tase activation (see next section). However, the analysis of a larger cohort is needed to accurately evaluate the relevance of γ-secretase dysfunction in the sporadic form of AD.

γ-Secretase modulators (GSMs) alleviate the FAD-associated effects of the majority of PSEN1 mutationsGSMs act as activators of the carboxypeptidase-like activity of the γ-secretase complex (Chávez-Gutiérrez et al., 2012;Takeo et al., 2014). Given the observed reduction in the γ-secretase carboxypeptidase-like efficiency in FAD patientbrain samples, we hypothesized that incubation with GSMs would enhance the carboxypeptidase-like activity and thereby correct for the pathogenic Aβ profiles associated with FAD.Thus, we decided to investigate the effects of two different GSM families on the efficiency of the γ-secretase carboxy-peptidase-like activity in FAD brain samples. Specifically, we tested an acid- and an imidazole-based modulator at 1 µM in our in vitro reactions (Fig. 2, B and C, respectively). Fig. 2

Aβ production in familial Alzheimer’s brain | Szaruga et al.2008

(B and C) shows that both GSMs restore the efficiency of the carboxypeptidase-like activity to control levels in eight out of nine FAD PSEN1 and seven out of nine FAD PSEN1 brain samples, respectively. Patient brain samples carrying the intron 4 mutation displayed limited responses to both GSMs. Similarly, the imidazole-based GSM did not restore the effi-ciency of the carboxypeptidase-like activity in the I202F case. The first extracellular loop of PSEN is part of an allosteric ligand-binding site within the N-terminal fragment of PSEN (Takeo et al., 2014); most likely, the intron 4 and I202F mu-tations disrupt the binding of GSMs to PSEN/γ-secretase. These data indicate that GSM treatment may be particularly useful in FAD, although the magnitude of the modulatory response may depend on the nature of the mutation and the

GSM chemistry. A very recent study reached similar conclu-sions on the effects of GSMs on FAD, using as a model neu-ronal cultures derived from FAD patient induced pluripotent stem cells (Moore et al., 2015). We would like to point out that full documentation of GSM-mediated effects on Aβ pro-files should be performed before any clinical applications. In particular, potential modulatory effects on the generation of long Aβ peptides (>42 amino acids) should be considered toprevent nondesired alterations at that level.

Collectively, these studies demonstrate no significant differences in the endopeptidase activity levels between FAD and control brain samples. The lack of effect on AICD pro-duction ascertains that the effects on the Aβ38/Aβ42 ratioare not simply caused by the extensive damage in the late-

Figure 2. AD-causing PSEN1 mutants impair γ-cleavage efficiency in FAD human brain samples, and GSMs correct for the pathogenic effect. (A) Carboxypeptidase-like efficiency seen as Aβ38/Aβ42 (product/substrate) ratio. (B and C) The re-sponses to GSMs observed in mutation carrier brain samples is contrasted with the carboxypeptidase-like efficiencies measured in brain samples from nondemented subjects (gray area, shown in panel A). To determine the response to GSMs, CHA PSO-resistant membranes prepared from brain tissue of patients were incu-bated with 1.5 µM C99-3×FLAG substrate in the presence of 1 µM GSM. Graphs show mean ± SD for groups with one case or mean of means ± SD for groups with number of cases greater than one. All experiments were repeated three to five times, and statistical significance was tested with one-way ANO VA and Dunnett's post test, taking the corresponding WT set as the control group (**, P < 0.01; *, P < 0.05).

2009JEM Vol. 212, No. 12

stage AD brains. Interestingly, when expressed in homozygous Psen1/2-deficient MEFs, some of the clinical mutants cause by themselves decrements in the overall γ-secretase endo-peptidase activity (seven mutations cause a reduction ranging from 70% to ∼30% of the normal γ-secretase activity, and twomutations did not affect activity; Fig. 3 B). The presence of the normal PSEN1 allele in heterozygous FAD patients prob-ably compensates for the decrease in γ-secretase endopepti-dase activity observed for some of the pathogenic PSEN1 mutants in homozygous Psen1/2-deficient MEFs (Fig. 3 B). Thus, our investigation does not support a dominant-nega-tive effect of the FAD allele over the healthy allele (normal PSEN1/γ-secretase), as proposed by others (Xia et al., 2015). However, we cannot discard the proposition that misprocess-ing of other γ-secretase substrates may contribute to diseasesymptoms (discussed in Chávez-Gutiérrez et al. [2012]) and help to explain the wide clinical spectrum observed in FAD patients (reviewed in Bergmans and De Strooper [2010]).

Significantly, this study is important in its demonstra-tion that γ-secretase dysfunction is the common denomina-tor in FAD patient brain samples (Fig. 2 A). These findings are entirely consistent with the effect of FAD-linked PSEN mutations on γ-secretase function shown in Fig. 3 C and inour previous work (Chávez-Gutiérrez et al., 2012). Our data

highlight that qualitative shifts in Aβ product profiles, towardlonger Aβ peptides, are the central feature in FAD patho-genesis, although elevated Aβ42 peptide could contribute topathogenesis in some FAD cases.

The qualitative shifts in the Aβ profiles observed withthe clinical mutations suggest that longer Aβ peptides (≥42)may promote neurotoxicity by providing the seeds for “toxic oligomers,” even at low concentrations. In this regard, a recent publication has put emphasis on the high amyloidogenicity and pathogenicity of Aβ43 (Saito et al., 2011). Althoughwe could not quantify this peptide in our tests on brain samples, because the amounts generated did not reach the threshold of detection, our previous work with membranes from Psen1/2-deficient fibroblasts expressing WT or mutant PSENs clearly indicates that FAD-linked mutants elevate the relative production of Aβ43 (Chávez-Gutiérrez et al., 2012).Furthermore, we show that the Aβ40 production rate is con-sistently low in FAD patient brain samples. Depletion of this particular peptide may also contribute to pathogenesis (Wang et al., 2006; Kim et al., 2007). Notably, minor changes in Aβprofiles have been reported to have drastic effects on neu-rotoxicity (Kuperstein et al., 2010). A phase II clinical trial using a humanized antibody against different Aβ42 assem-blies (crenezumab; Adolfsson et al., 2012) is currently being

Figure 3. FAD-PSEN1 mutations show variable effects on the endopeptidase cleavage, but all impair the fourth enzymatic turnover of γ-secre-tase. (A) Nct, PSEN1-NTF, PSEN1-CTF, and Pen2 protein levels in Psen1/2−/− MEFs transduced to express human WT or FAD-PSEN1. (B) Measurement of AICD production for WT and mutant PSEN1 γ-secretase complexes. To determine specific activities for WT and FAD complexes, AICD products were normalizedto PSEN1-CTF fragment levels quantified by Western blot. (C) FAD-PSEN1 mutations consistently impair the fourth catalytic cleavage seen as Aβ38/Aβ42(product/substrate) ratio, relative to WT activity. All experiments were repeated three to five times. Graphs show mean ± SE, and statistical significance was tested with one-way ANO VA and Dunnett's post test, taking the corresponding WT set as the control group (**, P < 0.01; *, P < 0.05).

Aβ production in familial Alzheimer’s brain | Szaruga et al.2010

conducted on PSEN1-E280A patients in Colombia (http://www.clinicaltrials.gov/ct2/show/NCT01998841). Accord-ing to the observed γ-secretase dysfunction in FAD, it wouldbe of relevance to test the affinity of crenezumab for other long, aggregation-prone Aβ peptides.

With regard to SAD, our data reveal a heterogeneous group in terms of γ-secretase activity, suggesting that sub-populations of late-onset patients may present alterations in Aβ production, which could be of relevance for future “per-sonalized” treatment strategies. However, the fact that most patients do not show altered γ-secretase activity supports theview that accumulation of Aβ peptides in the central ner-vous system of SAD patients is more frequently caused by impaired Aβ peptide clearance (Mawuenyega et al., 2010).The potential causes of alterations in γ-secretase function inthese few late-onset cases are intriguing, and the implications of such changes on disease onset, progression/duration, and therapy are currently unknown.

In conclusion, our investigation is the first to assess how the heterozygous situation in patients actually affects γ-secre-tase function in human brain. We find no evidence for a loss of overall γ-secretase endopeptidase function. Alternatively,we propose that qualitative changes in Aβ product profiles arethe basis of PSEN/γ-secretase–mediated pathogenicity. These findings imply that long Aβ peptides are potently pathogenic, and we speculate that a small alteration in the clearance of these long amyloidogenic peptides may contribute to late-on-set AD. Finally, our findings may have direct implications in therapy, as they indicate that activation of the carboxypepti-dase-like activity (while respecting the endopeptidase func-tion) could be a promising therapeutic concept in FAD.

MAT ERIALS AND MET HODSAntibodies and reagents. Antibodies were purchased as follows: MAB5232 against human PSEN1-CTF from EMD Millipore, 18189 rabbit polyclonal against human Pen2 from Abcam, 612290 and 610820 mouse monoclonal anti–human NCT and anti–human flotillin-1 from BD. ELI SA antibodies and GSMs were obtained through col-laboration with Janssen Pharmaceutica NV, Beerse, Bel-gium: JRF AB038 for Aβ1-38, JRF/cAb40/28 for Aβ1-40, JRF/cAb42/26 for Aβ1-42, and detection antibody JRF/AbN/25 against the N terminus of Aβ. Acid-based((2-[(1R,2S)-1-[4-methyl-1-[4-(trifluoromethyl)phenyl]pentyl]-2-[4-(trifluoromethyl)phenyl]-4-piperidyl] acetic acid) and imidazole-based (N-[2-fluoro-5-(trifluoro-methyl)phenyl]-5-[3-methoxy-4-(4-methylimidazol-1-yl)phenyl]-2-methyl-1,2,4-triazol-3-amine) GSMs were syn-thesized according to described procedures (Crump et al., 2011; Velter et al., 2014). γ-Secretase inhibitor refers to in-hibitor X purchased from EMD Millipore.

Expression and purification of C99-3×FLAG substrate. Sub-strate expression and purification was performed as previ-ously described (Chávez-Gutiérrez et al., 2008). Purity was

assessed by SDS-PAGE and Coomassie staining (gelcode re-agent; Thermo Fisher Scientific).

Subjects. Human cortical specimens for quantification of γ-secretase activity were obtained from Brain Bank at TokyoMetropolitan Institute of Gerontology, Queen Square Brain Bank for Neurological Disorders at University College Lon-don, throughout collaboration with the Neuroscience Group of Antioquia Brain Bank at University of Antioquia, Medellín, Colombia, and the Neurological Tissue Bank of the Bio-banc-Hospital Clinic-IDI BAPS. All of the samples came from brains that were removed and placed in −80°C within 65 hpostmortem (patients were moved to a cold room within 2 h after death). Samples were collected according to protocols approved by respective ethical boards, and written legal con-sents for the use of organs for medical research are available for each patient. A total of 48 brain samples were used for the reported project: 6 SAD cases and 10 controls from the Brain Bank at Tokyo (Brodmann areas 9–11); 4 different PSEN1 FAD mutations from the Queen Square Brain Bank at Uni-versity College London (1 patient per mutation, Brodmann areas 9–11); 10 E280A FAD cases, 5 SAD cases, and 4 controls from D. Sepulveda-Falla (Brodmann area 11); 5 different PSEN1 FAD mutations (8 patients, Brodmann areas 9–11; Pera et al., 2013); and 1 control from the Neurological Tissue Bank of the Biobanc-Hospital Clinic-IDI BAPS. All human protocols were approved by Medical Ethics Commit-tee UZ KU Leuven, Belgium.

Generation of MEFs. Psen1/Psen2−/− MEFs (Herreman et al., 2000) were cultured in Dulbecco’s modified Eagle’s medi-um/F-12 (Life Technologies) containing 10% fetal bovine serum. MEFs were transduced using pMSCV-puro, a replica-tion-defective recombinant retroviral expression system (Ta-kara Bio Inc.) harboring cDNA inserts coding for WT human PSEN1 or variants: V89L, intron 4 (p.L113_I114insT), E120G, M139T, I202F, I213T, P264L, R278I, and L286P. Stable cell lines were selected using 5 µg/ml puromycin (Sigma-Aldrich).

DRM preparation from human brains or MEFs. CHA PSO DRMs were prepared for human brain frontal cortices as pre-viously described (Kakuda et al., 2012) with minor modifica-tions or from MEFs expressing WT γ-secretase or mutantcomplexes containing PSEN1 mutations. In the first case, after careful removal of leptomeninges and blood vessels, <250 mg blocks of tissue were homogenized in ∼10 vol of10% sucrose in MBS buffer (25 mM MES, pH 6.5, 150 mM NaCl) containing 1% CHA PSO (Sigma-Aldrich) and prote-ase inhibitors (Complete; Roche). In the case of MEFs, total membranes were prepared from 12 big culture dishes (245 × 245 × 25), and membrane pellets were homogenized in ∼2.5 ml of 10% sucrose in MBS buffer (25 mM MES, pH 6.5, 150 mM NaCl) containing 1% CHA PSO and protease inhib-itors. Each homogenate was mixed with equal volume of 70%sucrose in MBS buffer, and 4 ml was placed at the bottom of

2011JEM Vol. 212, No. 12

an ultracentrifuge tube (344059; Beckman Coulter) and suc-cessively overlaid with 4 ml of 35% sucrose and 4 ml of 5% sucrose, both in MBS buffer. Samples were centrifuged at 39,000 rpm for 20 h at 4°C on an SW 41 Ti rotor (Beckman Coulter). After centrifugation, the DRM fraction (interface of 5%/35% sucrose) was carefully collected, rinsed in 20 mM PIP ES, pH 7.0, 250 mM sucrose, and 1 M EGTA, and recen-trifuged twice (100,000 g, 60 min, 4°C). The resultant pellet was resuspended with the aforementioned buffer using a 26G syringe and stored at −80°C until use. All DRM fractionsused in this study were set to 1 µg/µl with 20 mM PIP ES, pH 7.0, 250 mM sucrose, and 1 mM EGTA. Protein levels were tested by immunoblot using anti–flotillin-1 antibody.

Quantification of Aβ production rates by MSD ELI SA. To de-termine de novo production of Aβ peptides, 6 µg CHAPSO-resistant membranes were incubated for 0 or 4  h at 37°C with 1.5 µM C99-3×FLAG substrate. The activity as-says were performed in the presence of 2.5% DMSO (or 1 µM GSM in DMSO), 1 mM EGTA, 0.3% CHA PSO, and protease inhibitors (Complete; Roche). Aβ38, Aβ40, andAβ42 levels in reactions were quantified on Multi-Spot 96-well plates precoated with anti-Aβ38, -Aβ40, and -Aβ42 an-tibodies using multiplex MSD technology. MSD plates were blocked with 150 µl/well 0.1% casein buffer for 1.5 h at room temperature (600 rpm) and rinsed 5× with 200 µl/well wash-ing buffer (PBS + 0.05% Tween-20). 25  µl SUL FO-TAG JRF/AbN/25 detection antibody diluted in blocking buffer was mixed with 25 µl of standards (synthetic human Aβ1-38, Aβ1-40, and Aβ1-42 peptides) or reaction samples diluted inblocking buffer and loaded 50 µl per well.

After overnight incubation at 4°C, plates were rinsed with washing buffer and 150 µl/well of the 2× MSD Read Buffer T (Tris-based buffer containing tripropylamine, pur-chased from Meso Scale Discovery) was added. Plates were immediately read on a Sector Imager 6000 (Meso Scale Dis-covery). We determined the rates at which Aβ38, Aβ40, andAβ42 are produced in each sample by subtracting the 0-hvalue from the 4-h value obtained by ELI SA (Meso Scale Discovery) and normalizing Aβ amounts against time to ex-press rates in pM/h. To address the effects of modulators, we performed reactions in the presence of 1 µM GSMs. The ratio Aβ38/42 was taken as an estimation of the efficiency of thefourth turnover of the γ-secretase complex.

Statistical analysis. All statistical analysis was performed using Prism 6 software (GraphPad Software). An ANO VA test was used to test the significance of the changes between groups.

ACkNOwLEDGMENTSWe thank Dr. Amantha Thathiah for critical reading of the manuscript. We thank Dr. Francisco Lopera (University of Antioquia, Medellín, Colombia), Dr. Isidro Ferrer (Insti-tut Neuropatologia, Hospital Universitari Bellvitge, Barcelona, Spain), Dr. Ellen Gelpi (Neurological Tissue Bank of the Biobanc-Hospital Clinic-IDI BAPS, Barcelona, Spain), Dr. Hiroyuki Hatsuta (Tokyo Metropolitan Institute of Gerontology, Tokyo, Japan), and

Dr. Yasuo Ihara and Dr. Nobuto Kakuda (Doshisha University, Kyoto, Japan) for their invaluable assistance in the collection of postmortem human brain samples. We are grateful to all donors and their relatives for consent to autopsy and use of their tissues to advance scientific research. We thank Mark Mercken from Janssen Pharmaceutica for anti-Aβ monoclonal antibodies and Michel Vande Kerckhove for helpful discussions.

This work was funded by the Fund for Scientific Research, Flanders, the KU Leu-ven, a Methusalem grant from the KU Leuven and the Flemisch Government, IWT Agency for Innovation by Science and Technology, Janssen Pharmaceutica, Interuni-versity Attraction Poles Program of the Belgian Federal Science Policy Office, Instituto de Salud Carlos III (PI10/00018 to A. Lleo), and CIB ERNED. The London Neurodegen-erative Diseases Brain Bank received funding from the Medical Research Council and from the Brains for Dementia Research project (funded by the Alzheimer’s Society and Alzheimer’s Research UK). We acknowledge support from the National Institute for Health Research (NIHR) Queen Square Biomedical Unit in Dementia and the Leon-ard Wolfson Experimental Neurology Centre. N.C. Fox is an NIHR Senior Investigator. D. Sepulveda-Falla is supported by the Hamburg State Ministry of Science and Re-search, Landesforschungsförderung “Molekulare mechanismen der netzwerkmodifi-zierung.” N.S. Ryan is supported by a Brain Exit Fellowship. T. Lashley is supported by an Alzheimer’s Research UK fellowship. B. De Strooper is supported by the Arthur Bax and Anna Vanluffelen chair for Alzheimer’s disease.

H. Gijsen is an employee of Janssen Pharmaceutica NV and holds stock in John-son & Johnson. B. De Strooper is a consultant for Janssen Pharmaceutica, Envivo Pharmaceuticals, and Remynd NV. The authors declare no further competing fi-nancial interests.

Submitted: 28 May 2015

Accepted: 11 September 2015

REFERENCESAdolfsson, O., M. Pihlgren, N. Toni, Y. Varisco, A.L. Buccarello, K. Antoniello,

S. Lohmann, K. Piorkowska, V. Gafner, J.K. Atwal, et al.. 2012. An ef-fector-reduced anti-β-amyloid (Aβ) antibody with unique aβ bind-ing properties promotes neuroprotection and glial engulfment of Aβ. J. Neurosci. 32:9677–9689. http://dx.doi.org/10.1523/JNEUROSCI.4742-11.2012

Bergmans, B.A., and B. De Strooper. 2010. γ-secretases: from cell biol-ogy to therapeutic strategies. Lancet Neurol. 9:215–226. http://dx.doi.org/10.1016/S1474-4422(09)70332-1

Campion, D., C. Dumanchin, D. Hannequin, B. Dubois, S. Belliard, M. Puel, C. Thomas-Anterion, A. Michon, C. Martin, F. Charbonnier, et al.. 1999. Early-onset autosomal dominant Alzheimer disease: prevalence, geneticheterogeneity, and mutation spectrum. Am. J. Hum. Genet. 65:664–670. http://dx.doi.org/10.1086/302553

Chávez-Gutiérrez, L., A. Tolia, E. Maes, T. Li, P.C. Wong, and B. de Strooper. 2008. Glu332 in the Nicastrin ectodomain is essential for γ-secretasecomplex maturation but not for its activity. J. Biol. Chem. 283:20096–20105. http://dx.doi.org/10.1074/jbc.M803040200

Chávez-Gutiérrez, L., L. Bammens, I. Benilova, A. Vandersteen, M. Benurwar, M. Borgers, S. Lismont, L. Zhou, S. Van Cleynenbreugel, H. Esselmann, et al.. 2012. The mechanism of γ-secretase dysfunction in familialAlzheimer disease. EMBO J. 31:2261–2274. http://dx.doi.org/10.1038/emboj.2012.79

Crump, C.J., B.A. Fish, S.V. Castro, D.M. Chau, N. Gertsik, K. Ahn, C. Stiff, N. Pozdnyakov, K.R. Bales, D.S. Johnson, and Y.M. Li. 2011. Piperidine ace-tic acid based γ-secretase modulators directly bind to Presenilin-1. ACSChem. Neurosci. 2:705–710. http://dx.doi.org/10.1021/cn200098p

De Strooper, B. 2003. Aph-1, Pen-2, and Nicastrin with Presenilin gen-erate an active γ-secretase complex. Neuron. 38:9–12. http://dx.doi.org/10.1016/S0896-6273(03)00205-8

De Strooper, B. 2007. Loss-of-function presenilin mutations in Alzheimer disease. Talking Point on the role of presenilin mutations in Alzheimer disease. EMBO Rep. 8:141–146. http://dx.doi.org/10.1038/sj.embor.7400897

Aβ production in familial Alzheimer’s brain | Szaruga et al.2012

De Strooper, B., P. Saftig, K. Craessaerts, H. Vanderstichele, G. Guhde, W. Annaert, K. Von Figura, and F. Van Leuven. 1998. Deficiency of preseni-lin-1 inhibits the normal cleavage of amyloid precursor protein. Nature. 391:387–390. http://dx.doi.org/10.1038/34910

Edland, S.D., and D.R. Galasko. 2011. Fractional synthesis and clearance rates for amyloid β. Nat. Med. 17:1178–1179. http://dx.doi.org/10.1038/nm.2495

Fernandez, M.A., J.A. Klutkowski, T. Freret, and M.S. Wolfe. 2014. Alzheimer presenilin-1 mutations dramatically reduce trimming of long amyloid β-peptides (Aβ) by γ-secretase to increase 42-to-40-residue Aβ. J. Biol.Chem. 289:31043–31052. http://dx.doi.org/10.1074/jbc.M114.581165

Hata, S., S. Fujishige, Y. Araki, M. Taniguchi, K. Urakami, E. Peskind, H. Akatsu, M. Araseki, K. Yamamoto, R.N. Martins, et al. 2011. Alternative processing of γ-secretase substrates in common forms of mild cognitiveimpairment and Alzheimer’s disease: evidence for γ-secretase dysfunc-tion. Ann. Neurol. 69:1026–1031. http://dx.doi.org/10.1002/ana.22343

Hata, S., M. Taniguchi, Y. Piao, T. Ikeuchi, A.M. Fagan, D.M. Holtzman, R. Bateman, H.R. Sohrabi, R.N. Martins, S. Gandy, et al.; JapaneseAlzheimer’s Disease Neuroimaging Initiative. 2012. Multiple γ-secretaseproduct peptides are coordinately increased in concentration in the cere-brospinal fluid of a subpopulation of sporadic Alzheimer’s disease subjects. Mol. Neurodegener. 7:16. http://dx.doi.org/10.1186/1750-1326-7-16

Heilig, E.A., U. Gutti, T. Tai, J. Shen, and R.J. Kelleher III. 2013. Trans-dominant negative effects of pathogenic PSEN1 mutations on γ-secre-tase activity and Aβ production. J. Neurosci. 33:11606–11617. http://dx.doi.org/10.1523/JNE UROSCI.0954-13.2013

Herreman, A., L. Serneels, W. Annaert, D. Collen, L. Schoonjans, and B. De Strooper. 2000. Total inactivation of γ-secretase activity in presenilin-de-ficient embryonic stem cells. Nat. Cell Biol. 2:461–462. http://dx.doi.org/10.1038/35017105

Jurisch-Yaksi, N., R. Sannerud, and W. Annaert. 2013. A fast growing spec-trum of biological functions of γ-secretase in development and disease. Biochim. Biophys. Acta. 1828:2815–2827. http://dx.doi.org/10.1016/j.bbamem.2013.04.016

Kakuda, N., M. Shoji, H. Arai, K. Furukawa, T. Ikeuchi, K. Akazawa, M. Takami, H. Hatsuta, S. Murayama, Y. Hashimoto, et al.; Japanese Alzheimer’s Disease Neuroimaging Initiative. 2012. Altered γ-secretaseactivity in mild cognitive impairment and Alzheimer’s disease. EMBO Mol. Med. 4:344–352. http://dx.doi.org/10.1002/emmm.201200214

Kim, J., L. Onstead, S. Randle, R. Price, L. Smithson, C. Zwizinski, D.W. Dickson, T. Golde, and E. McGowan. 2007. Aβ40 inhibits amyloid depo-sition in vivo. J. Neurosci. 27:627–633. http://dx.doi.org/10.1523/JNE UROSCI.4849-06.2007

Kosik, K.S., C. Muñoz, L. Lopez, M.L. Arcila, G. García, L. Madrigal, S. Moreno, S. Ríos Romenets, H. Lopez, M. Gutierrez, et al.. 2015. Homozygosity of the autosomal dominant Alzheimer disease presenilin 1 E280A mutation. Neurology. 84:206–208. http://dx.doi.org/10.1212/WNL.0000000000001130

Kuperstein, I., K. Broersen, I. Benilova, J. Rozenski, W. Jonckheere, M. Debulpaep, A. Vandersteen, I. Segers-Nolten, K. Van Der Werf, V. Subramaniam, et al.. 2010. Neurotoxicity of Alzheimer’s disease Aβ pep-tides is induced by small changes in the Aβ42 to Aβ40 ratio. EMBO J. 29:3408–3420. http://dx.doi.org/10.1038/emboj.2010.211

Matsumura, N., M. Takami, M. Okochi, S. Wada-Kakuda, H. Fujiwara, S. Tagami, S. Funamoto, Y. Ihara, and M. Morishima-Kawashima. 2014. γ-Secretase associated with lipid rafts: multiple interactive pathways inthe stepwise processing of β-carboxyl-terminal fragment. J. Biol. Chem. 289:5109–5121. http://dx.doi.org/10.1074/jbc.M113.510131

Mawuenyega, K.G., W. Sigurdson, V. Ovod, L. Munsell, T. Kasten, J.C. Morris, K.E. Yarasheski, and R.J. Bateman. 2010. Decreased clearance of CNS β-amyloid in Alzheimer’s disease. Science. 330:1774. http://dx.doi.org/10.1126/science.1197623

Moore, S., L.D. Evans, T. Andersson, E. Portelius, J. Smith, T.B. Dias, N. Saurat, A. McGlade, P. Kirwan, K. Blennow, et al.. 2015. APP metabolism reg-ulates tau proteostasis in human cerebral cortex neurons. Cell Reports. 11:689–696. http://dx.doi.org/10.1016/j.celrep.2015.03.068

Pera, M., D. Alcolea, R. Sánchez-Valle, C. Guardia-Laguarta, M. Colom-Cadena, N. Badiola, M. Suárez-Calvet, A. Lladó, A.A. Barrera-Ocampo, D. Sepulveda-Falla, et al.. 2013. Distinct patterns of APP process-ing in the CNS in autosomal-dominant and sporadic Alzheimer dis-ease. Acta Neuropathol. 125:201–213. http://dx.doi.org/10.1007/s00401-012-1062-9

Pink, A.E., M.A. Simpson, N. Desai, R.C. Trembath, and J.N. Barker. 2013. γ-Secretase mutations in hidradenitis suppurativa: new insights intodisease pathogenesis. J. Invest. Dermatol. 133:601–607. http://dx.doi.org/10.1038/jid.2012.372

Potter, R., B.W. Patterson, D.L. Elbert, V. Ovod, T. Kasten, W. Sigurdson, K. Mawuenyega, T. Blazey, A. Goate, R. Chott, et al.. 2013. Increased in vivo amyloid-β42 production, exchange, and loss in presenilin mutationcarriers. Sci. Transl. Med. 5:189ra77.

Qi-Takahara, Y., M. Morishima-Kawashima, Y. Tanimura, G. Dolios, N. Hirotani, Y. Horikoshi, F. Kametani, M. Maeda, T.C. Saido, R. Wang, and Y. Ihara. 2005. Longer forms of amyloid β protein: implications forthe mechanism of intramembrane cleavage by γ-secretase. J. Neurosci.25:436–445. http://dx.doi.org/10.1523/JNE UROSCI.1575-04.2005

Quintero-Monzon, O., M.M. Martin, M.A. Fernandez, C.A. Cappello, A.J. Krzysiak, P. Osenkowski, and M.S. Wolfe. 2011. Dissociation between the processivity and total activity of γ-secretase: implications for the mech-anism of Alzheimer’s disease-causing presenilin mutations. Biochemistry. 50:9023–9035. http://dx.doi.org/10.1021/bi2007146

Saito, T., T. Suemoto, N. Brouwers, K. Sleegers, S. Funamoto, N. Mihira, Y. Matsuba, K. Yamada, P. Nilsson, J. Takano, et al.. 2011. Potent amyloi-dogenicity and pathogenicity of Aβ43. Nat. Neurosci. 14:1023–1032.http://dx.doi.org/10.1038/nn.2858

Takami, M., Y. Nagashima, Y. Sano, S. Ishihara, M. Morishima-Kawashima, S. Funamoto, and Y. Ihara. 2009. γ-Secretase: successive tripeptide and tet-rapeptide release from the transmembrane domain of β-carboxyl termi-nal fragment. J. Neurosci. 29:13042–13052. http://dx.doi.org/10.1523/JNE UROSCI.2362-09.2009

Takeo, K., S. Tanimura, T. Shinoda, S. Osawa, I.K. Zahariev, N. Takegami, Y. Ishizuka-Katsura, N. Shinya, S. Takagi-Niidome, A. Tominaga, et al..2014. Allosteric regulation of γ-secretase activity by a phenylimidaz-ole-type γ-secretase modulator. Proc. Natl. Acad. Sci. USA. 111:10544–10549. http://dx.doi.org/10.1073/pnas.1402171111

Tanzi, R.E., and L. Bertram. 2005. Twenty years of the Alzheimer’s disease amyloid hypothesis: a genetic perspective. Cell. 120:545–555. http://dx.doi.org/10.1016/j.cell.2005.02.008

Velter, A.I., F.P. Bischoff, D. Berthelot, M. De Cleyn, D. Oehlrich, L. Jaroskova, G. Macdonald, G. Minne, S. Pieters, F. Rombouts, et al.. 2014. Anilinotriazoles as potent gamma secretase modulators. Bioorg. Med. Chem. Lett. 24:5805–5813. http://dx.doi.org/10.1016/j.bmcl.2014.10.024

Wahrle, S., P. Das, A.C. Nyborg, C. McLendon, M. Shoji, T. Kawarabayashi, L.H. Younkin, S.G. Younkin, and T.E. Golde. 2002. Cholesterol-dependent γ-secretase activity in buoyant cholesterol-rich membranemicrodomains. Neurobiol. Dis. 9:11–23. http://dx.doi.org/10.1006/nbdi.2001.0470

Wang, R., B. Wang, W. He, and H. Zheng. 2006. Wild-type presenilin 1 protects against Alzheimer disease mutation-induced amyloid pathol-ogy. J. Biol. Chem. 281:15330–15336. http://dx.doi.org/10.1074/jbc.M512574200

Wolfe, M.S., W. Xia, B.L. Ostaszewski, T.S. Diehl, W.T. Kimberly, and D.J. Selkoe. 1999. Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and γ-secretase activity. Nature. 398:513–517. http://dx.doi.org/10.1038/19077

2013JEM Vol. 212, No. 12

Xia, D., H. Watanabe, B. Wu, S.H. Lee, Y. Li, E. Tsvetkov, V.Y. Bolshakov, J. Shen, and R.J. Kelleher III. 2015. Presenilin-1 knockin mice reveal loss-of-function mechanism for familial Alzheimer’s disease. Neuron. 85:967–981. http://dx.doi.org/10.1016/j.neuron.2015.02.010

Yagishita, S., M. Morishima-Kawashima, S. Ishiura, and Y. Ihara. 2008. Aβ46 isprocessed to Aβ40 and Aβ43, but not to Aβ42, in the low density mem-brane domains. J. Biol. Chem. 283:733–738. http://dx.doi.org/10.1074/jbc.M707103200

Article

The Rockefeller University Press $30.00J. Exp. Med. 2015 Vol. 212 No. 1 23–35www.jem.org/cgi/doi/10.1084/jem.20141015

23

It is widely believed that myelin-reactive CD4+ T cells initiate the formation of demyelinat-ing lesions in the central nervous system (CNS) during multiple sclerosis (MS). That premise is supported by extensive circumstantial evidence from animal models and genome-wide associa-tion studies (Steinman and Zamvil, 2006; Sawcer et al., 2011), and by the mechanism of action of disease-modifying agents (DMAs) that suppress clinical relapses by targeting lymphocytes (Stüve, 2008; Kowarik et al., 2011). Having crossed the blood-brain barrier (BBB), myelin-reactive CD4+ T cells induce chemokines and vasoactive mol-ecules, resulting in the local recruitment of a heterogeneous population of myeloid cells. Infil-trating myeloid cells secrete factors that escalate the inflammatory response and present anti-gen to reactivate encephalitogenic T cells within the CNS (Kawakami et al., 2004). Thus, MS dis-ease activity is dependent on an intricate interplay between the adaptive and innate immune sys-tems. Nevertheless, none of the FDA-approved

DMAs used to treat MS were designed to target innate immune cells.

Monocytes and macrophages can inflict dam-age in the CNS by phagocytosing the myelin sheath and by releasing factors that are toxic to oligodendrocytes and axons (Epstein et al., 1983; Lin et al., 1993; Toft-Hansen et al., 2004; Mantovani et al., 2011). Several studies have re-vealed dysregulation of peripheral monocytes and monocyte-derived dendritic cells in MS, mani-fested by increased expression of costimulatory molecules and polarizing cytokines (Balashov et al., 1997; Comabella et al., 1998; Karni et al., 2002, 2006; Vaknin-Dembinsky et al., 2006). Granulocytes have received less attention because they are relatively rare in mature MS and ex-perimental autoimmune encephalomyelitis (EAE)

CORRESPONDENCE Benjamin M. Segal: bmsegal@umich.edu

Abbreviations used: BBB, blood-brain barrier; CNS, central nervous system; CSF, cerebrospinal fluid; EAE, experimental autoimmune en-cephalomyelitis; EDSS, expanded disability status scale; G-CSF, granulocyte-colonystimulating factor; MS, multiplesclerosis; p.i., post immunization;PTx, Bordetella pertussis toxin.

Neutrophil-related factors as biomarkers in EAE and MS

Julie M. Rumble,1 Amanda K. Huber,1 Gurumoorthy Krishnamoorthy,6 Ashok Srinivasan,2 David A. Giles,1 Xu Zhang,3 Lu Wang,3 and Benjamin M. Segal1,4,5

1Holtom-Garrett Program in Neuroimmunology, Department of Neurology,2Department of Radiology, 3Department of Biostatistics, and 4Graduate Program in Immunology, University of Michigan, Ann Arbor, MI 48109

5Neurology Service, VA Ann Arbor Healthcare System, Ann Arbor, MI 481056Department of Neuroimmunology, Max Planck Institute for Neurobiology, 82152 Martinsried, Germany

A major function of T helper (Th) 17 cells is to induce the production of factors that activate and mobilize neutrophils. Although Th17 cells have been implicated in the patho-genesis of multiple sclerosis (MS) and the animal model experimental autoimmune enceph-alomyelitis (EAE), little attention has been focused on the role of granulocytes in those disorders. We show that neutrophils, as well as monocytes, expand in the bone marrow and accumulate in the circulation before the clinical onset of EAE, in response to systemic up-regulation of granulocyte colony-stimulating factor (G-CSF) and the ELR+ CXC chemokine CXCL1. Neutrophils comprised a relatively high percentage of leukocytes infiltrating the central nervous system (CNS) early in disease development. G-CSF receptor deficiency and CXCL1 blockade suppressed myeloid cell accumulation in the blood and ameliorated the clinical course of mice that were injected with myelin-reactive Th17 cells. In relapsing MS patients, plasma levels of CXCL5, another ELR+ CXC chemokine, were elevated during acute lesion formation. Systemic expression of CXCL1, CXCL5, and neutrophil elastase correlated with measures of MS lesion burden and clinical disability. Based on these results, we advo-cate that neutrophil-related molecules be further investigated as novel biomarkers and therapeutic targets in MS.

© 2015 Rumble et al. This article is distributed under the terms of an Attribution– Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http://creativecommons.org/licenses/ by-nc-sa/3.0/).

24 Neutrophil-related factors in EAE and MS | Rumble et al.

our additional finding that plasma levels of CXCL5, another CXCR2 ligand, are elevated in relapsing MS patients coinci-dent with acute lesion development. Furthermore, expression of CXCL1, CXCL5, and neutrophil elastase correlated with measures of MS lesion burden and clinical disability. The re-sults of our study endorse further evaluation of myeloid-related molecules as novel biomarkers and therapeutic targets in MS and other inflammatory demyelinating disorders.

RESULTSIntramedullary neutrophils and monocytes expand after active immunization with myelin antigensMyeloid cells are short-lived after tissue infiltration, raising the question of how they are replenished in the setting of re-lapsing or chronic autoimmune disease. We have previously shown that CD11b+CD115+Ly-6Chi blood monocytes are a precursor of CNS DCs and macrophages in EAE lesions (King et al., 2009). These cells are dynamically regulated during the

lesions. However, a major function of Th17 cells, identified as critical effector cells in EAE and MS, is to induce the expres-sion of neutrophil activating molecules such as granulocyte-colony stimulating factor (G-CSF) and ELR+ CXC chemokines (Kolls and Lindén, 2004; Khader et al., 2009; Onishi and Gaffen, 2010; Pelletier et al., 2010; Becher and Segal, 2011). Indeed, cerebrospinal fluid (CSF) samples obtained from newly diag-nosed MS patients at clinical relapse had elevated IL-17A lev-els which positively correlated with CSF neutrophil counts (Kostic et al., 2014). A pathogenic role of neutrophils in human autoimmune demyelinating disease is further suggested by the occurrence of severe exacerbations in some MS and NMO pa-tients when given recombinant G-CSF (Openshaw et al., 2000; Burt et al., 2001; Jacob et al., 2012). Transcripts encoding G-CSF are expressed in MS lesions but not normal appear-ing white matter (Lock et al., 2002), and the neutrophil- attracting chemokine CXCL8 has been detected in CSF ofMS patients (Ishizu et al., 2005; Campbell et al., 2010). It wasrecently reported that circulating neutrophils are more nu-merous, and exhibit a primed state, in individuals with MS(Naegele et al., 2012). These observations echo prior studiesthat documented enhanced neutrophil protease activity andintegrin receptor expression in patients with MS during relapsewhen compared with MS patients in remission, healthy con-trols, or individuals with other neurological diseases (Aoki et al., 1984; Guarnieri et al., 1985; Ziaber et al., 1998).

Despite the paucity of neutrophils in typical mature MS lesions, studies in the EAE model indicate that they comprise a higher frequency of infiltrating cells during the preclinical phase and could play a role in nascent lesion development by mediating BBB and blood-CSF barrier breakdown, and/or by stimulating the maturation of local APCs (Carlson et al., 2008; Christy et al., 2013; Steinbach et al., 2013). In the vast major-ity of MS tissue samples obtained by biopsy or autopsy, lesions are subacute or chronic. Hence, the impression that neutro-phils do not comprise a significant leukocyte population in the CNS during MS might reflect a sampling bias. Irrespective of their presence in the CNS, neutrophils could, conceivably, pro-mote disease activity from the periphery. Hence, activation of neutrophils within the bone marrow during autoimmune de-myelinating disease would drive the mobilization of mono-cytes, as well as neutrophils themselves, into the circulation, thereby increasing the pool of myeloid cells available for re-cruitment to the CNS (Singh et al., 2012).

In the current study, we serially analyzed the cellular com-position of bone marrow, blood, and CNS infiltrates after in-duction of EAE by active immunization or Th17 transfer. We found that neutrophils and monocytes expanded in the bone marrow and accumulated in the circulation during the preclin-ical phase of EAE in response to systemic up-regulation of G-CSF and the CXCR2 binding chemokine CXCL1. G-CSFreceptor (G-CSFR) deficiency and CXCR2 blockade sup-pressed the accumulation of circulating myeloid cells and ame-liorated the clinical course. In a model of spontaneous EAE,circulating neutrophils also expanded early in the clinical course.The translational relevance of these results is underscored by

Figure 1. Neutrophils and monocytes expand in the bone marrow after EAE induction. (A–C) WT mice were immunized with MOG35-55 in CFA and injected with PTx on days 0 and 2 p.i. Bone marrow cells were flushed from femurs and tibiae of representative mice at serial time points and analyzed by flow cytometry to enumerate neutrophils (CD31Ly6CintLy6G+, black bars/closed circles), monocytes (CD31+Ly6ChiLy6G, gray bars and circles), and lymphocytes (CD31+Ly6C, diagonal stripes/open triangles). (A) Percentage of leukocyte subsets within bone marrow cells. Data arerepresentative of six experiments (n ≥ 3 per time point). (B and C) Absolutenumbers of neutrophils (B) and monocytes and lymphocytes (C) recoveredper mouse. Data are representative of nine experiments (n ≥ 3 per timepoint). All graphs indicate means; error bars denote SEM. *, P < 0.05;**, P < 0.01 compared with unimmunized mice.

JEM Vol. 212, No. 1

Article

25

coadministration of Bordetella pertussis toxin (PTx), which is necessary for the clinical manifestation of disease in MOG/CFA immunized C57BL/6 mice, augmented the frequency of peripheral blood neutrophils (Fig. 2 A). Mice immunized with MOG peptide in IFA and injected with PTx on days 0 and 2 p.i. had comparable percentages of circulating myeloid cells to mice treated with PTx alone (Fig. 2 A). This suggests that my-eloid cells accumulate in the blood largely in response to my-cobacterial components in CFA. Immunization with a peptide of ovalbumin in CFA induced a significant expansion of circu-lating neutrophils and monocytes, demonstrating that this phe-nomenon is not antigen-specific (Fig. 2 A).

G-CSF and CXCL1 levels rise in the serumduring the preclinical stage of EAEWe next investigated the factors responsible for stimulating my-eloid cell expansion in the bone marrow and their mobilizationinto the circulation in our model. G-CSF levels rose dramaticallyin the serum of actively immunized mice, peaking at day 1 p.i. and remaining elevated through day 14 (Fig. 2 C). Reminiscentof its effects on the frequency of circulating neutrophils(Fig. 2 A), PTx enhanced the expression of serum G-CSF. Theneutrophil-attracting chemokine CXCL1, but not CXCL2, spiked in the serum on day 1 p.i. and fell to near baseline levels

course of EAE, accumulating in the blood and CNS imme-diately before clinical episodes. However the source of the inflammatory monocytes (i.e., from the bone marrow or ex-tramedullary sites), and the factors that modulate their fre-quency and migration patterns, have yet to be investigated in detail. Similarly, relatively little is known about the activation and distribution of neutrophils, or their interactions with other myeloid populations, in myelin-immunized mice.

We speculated that intramedullary myeloid cells expand in the bone marrow and then migrate into the circulation to support new CNS lesion formation. To assess our hypothesis, C57BL/6 mice were actively immunized with myelin oligo-dendrocyte glycoprotein peptide (MOG35-55) in CFA and bone marrow cells were analyzed at serial time points there-after. We found that the frequencies and absolute numbers of intramedullary neutrophils and monocytes rose during the preclinical phase (Fig. 1, A–C). The numbers of both myeloid subsets remained elevated above baseline at clinical onset (days 10–14 post immunization [p.i.]). In contrast, the num-ber and percentage of intramedullary lymphocytes fell over the same time frame (Fig. 1, A and C).

The expansion of monocytes and neutrophils in the bone marrow was mirrored in the blood and spleen, from preclinical time points though EAE onset (Fig. 2, A and B). Interestingly,

Figure 2. Serum G-CSF and CXCL1 are up-regulated and myeloid cells are mobi-lized into the circulation during EAE. (A–D) Peripheral blood cells and sera were collected from mice that had been primed with MOG35-55 or OVA323-339 in CFA or MOG35-55 in IFA, with or without administration of PTx. (A) Percentage of circulating neutrophils(CD11b+Ly6CintLy6G+, black bars) and mono-cytes (CD11b+Ly6ChiLy6G, white bars) onday 7 p.i. Shown is a representative of threeexperiments (n ≥ 6 mice per group). (B) Numbersof neutrophils (top panels, closed squares) andmonocytes (bottom panels, open circles) perml of blood or per spleen at serial time pointsafter active immunization with MOG35-55 in CFA. Data are pooled from 10 experiments (blood, n ≥ 6 per time point) or 5 experiments (spleen, n ≥ 3 per time point). (C and D) Serum levels of G-CSF (C) and CXCL1 (D) were mea-sured by ELISA. Data were pooled from 10 experiments (n ≥ 3 mice per time point). (E and F) G-CSF (E) and CXCL1 (F) were mea-sured in tissue homogenates and normalized to total protein. Shown is a representative of two experiments (n = 4 mice per time point). All graphs indicate means; error bars denote SEM. *, P < 0.05; **, P < 0.01 compared with unimmunized mice. #, P < 0.05; ##, P < 0.01 between groups. ND = not detectable.

26 Neutrophil-related factors in EAE and MS | Rumble et al.

MOG-immunized C57BL/6 WT and G-CSFR–deficient (Csf3r/) mice (Liu et al., 1996). Csf3r/ mice were highly resistant to EAE (Fig. 3 A). Whereas 80% of WT mice mani-fested neurological deficits beginning on day 10, only 14% of Csf3r/ mice succumbed to clinical EAE, with the earliest signs presenting on day 15. Histological analyses revealed mul-tifocal inflammatory disease in the spinal cords of WT mice at peak clinical EAE but no pathological changes in Csf3r/ mice euthanized on the same day p.i. (Fig. 3 B). Consistent with these findings, cells isolated from the CNS of WT mice were composed primarily of infiltrating hematopoietic cells, whereas microglia were the most prevalent cell type isolated from the CNS of Csf3r/ mice (Fig. 3 C). As previously

by day 7 (Fig. 2 D). Co-administration of PTx did not alter CXCL1 levels. To determine the source of G-CSF and CXCL1, we harvested an array of tissues on days 1 and 7 p.i. and per-formed ELISAs on homogenate supernatants. G-CSF and CXCL1 production were up-regulated in the spleen, lungs, liver, and spinal cord at both time points (Fig. 2, E and F).

The clinical manifestation of EAE is dependent on intact G-CSF signaling in hematopoietic cellsAs mentioned earlier, administration of G-CSF has been as-sociated with severe exacerbation of MS (Openshaw et al., 2000). To directly assess the role of endogenous G-CSF in the development of EAE, we compared the clinical courses of

Figure 3. EAE is dependent on G-CSF signaling in hematopoietic cells. (A–E) WT (closed triangles, black bars) and Csf3r/ (open triangles, white bars) mice were actively immunized with MOG35-55 in CFA. (A) Mean clinical scores (n = 25 WT, 21 Csf3r/ pooled from five independent experiments). (B) Representative paraffin sections of spinal cords stained with H&E. (C) Cell subsets recovered from spinal cords at peak of disease, shown as a percent-age of total CD45+ cells. Cell types were defined as follows: neutrophil (CD45hi, CD11b+, Ly6G+), monocyte (CD45hi, CD11b+, CD11c, Ly6G), DC (CD45hi,CD11b+, CD11c+), CD3+ (CD45hi, CD3+), and microglia (CD45mid CD11b+). Data are representative of two experiments (n ≥ 4 mice/group). (D and E) Circulat-ing and splenic neutrophils (D) and monocytes (E) were enumerated by flow cytometry. Data were pooled from two independent experiments (n ≥ 5 pergroup). (F) MOG-specific cytokine production by draining lymph node cells measured by ELIspot. Data are representative of three experiments (n = 3–5mice per group). In the experiment shown there were 2.4 × 105 total cells/well. (G) Mean clinical scores of WT to WT (closed triangles, n = 10) or Csf3r/

to WT (open triangles, n = 9) bone marrow chimeric mice after active immunization with MOG35-55 in CFA. Data are representative of three experiments.All graphs indicate means; error bars denote SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Bars, 100 µm.

JEM Vol. 212, No. 1

Article

27

myeloid cell populations that we observed in MOG-immunized mice. This prompted us to question whether peripheral my-eloid cells expand and mobilize in models of EAE that do not require the administration of exogenous antigen or adjuvant. Approximately 50% of C57BL/6 mice that coexpress MOG-specific T cell receptor and B cell receptor transgenes (OSE mice) spontaneously develop inflammatory demyelinating lesions and neurological deficits by 12 wk of age when maintained under specific pathogen-free conditions (Krishnamoorthy et al., 2006). We found that PBMC, splenocytes, bone marrow cells, and CNS mononuclear cells harvested from OSE mice with acute EAE (1–3 d after onset) had elevated percentages of neutrophils, but not monocytes, when compared with analo-gous cell preparations from healthy OSE mice (Fig. 4, A–D). The frequency of neutrophils fell to baseline during chronic EAE (>14 d after onset).

As an alternative to active immunization, EAE can be in-duced via the injection of myelin-primed, IL-23 modulated CD4+ Th17 cells or IL-12 modulated Th1 cells into naive WT C57BL/6 hosts (Kroenke et al., 2008; Kroenke and Segal, 2011). Similar to our findings in OSE mice with spontaneous EAE, the frequencies of circulating and splenic neutrophils, but not monocytes, rose above baseline in adoptive transfer recipients of either Th1- or Th17-polarized T cells at clinical onset (Fig. 4, E and F). Th17 effector cells mediated the most profound neutrophil expansion, consistent with the established role of IL-17A as an inducer of granulocyte mobilizing factors (Kolls and Lindén, 2004; Onishi and Gaffen, 2010; Pelletier et al., 2010). The frequency of bone marrow neutrophils did

described in infectious disease models (Basu et al., 2000), “emergency” mobilization of myeloid cells occurred in Csf3r/ mice after immunization with MOG in CFA in that pe-ripheral pools of neutrophils and monocytes expanded over baseline. However, at the time of clinical onset, the absolute numbers of circulating and splenic neutrophils, and of circu-lating monocytes, were significantly lower in Csf3r/ mice than in WT mice (Fig. 3, D and E). The low incidence and mild course of disease in Csf3r/ mice was not due to insuf-ficient CD4+ Th priming because MOG-immunized Csf3r/ and WT mice mounted comparable IL-17 and IFN- responses upon antigenic challenge ex vivo (Fig. 3 F).

Although primarily expressed on neutrophils, G-CSFR has been detected on nonhematopoietic cells, including glia and subsets of neurons (Kadota et al., 2012). Therefore, we con-structed bone marrow chimeric mice to establish whether G-CSFR deficiency in immune cells alone is sufficient to confer resistance to EAE. Lethally irradiated WT hosts were recon-stituted with either WT or Csf3r/ bone marrow cells. After active immunization with MOG in CFA, WT→WT bone marrow chimeras developed severe EAE at 100% incidence. In contrast, Csf3r/→WT bone marrow chimeras were highly resistant to disease induction (11% incidence), simulating the phenotype of germline Csf3r/ mice (Fig. 3 G).

Peripheral neutrophils expand at clinical onset in adjuvant-free models of EAEThe data in Fig. 2 A suggest that TLR ligands in CFA drive the systemic up-regulation of G-CSF and consequent shifts in

Figure 4. Neutrophils accumulate at onset of disease in adjuvant-free models of EAE. (A–D) OSE mice were sacrificed when healthy (n = 5; white bars), within 2 d of the onset of clinical EAE (n = 5; black bars) or during the chronic stages of EAE (n = 4; gray bars). Peripheral blood cells (A), splenocytes (B), BM cells (C) and spinal cord–infiltrating cells (D) were collected. Neutrophils and monocytes were enumerated by flow cytometry. (E and F) WT mice were injected with IL-12–polarized (Th1; black bars) or IL-23–polarized (Th17; gray bars) MOG-specific T cells. At day 7 after transfer, blood (E) and spleens (F) were collected and neutrophils and monocytes were enumerated by flow cytometry. Data are pooled from four (Th17 transfers) or two (Th1 transfers) experiments (n ≥ 10 mice per group). All graphs indicate means; error bars denote SEM. *, P < 0.05; **, P < 0.005; ***, P = 0.001; ****, P < 0.0001, by two-way ANOVA, correcting for multiple comparisons.

28 Neutrophil-related factors in EAE and MS | Rumble et al.

transfer (Fig. 6, B and C). At clinical onset, circulating neutro-phils had expanded approximately threefold over baseline in Csf3r/ hosts and approximately sevenfold over baseline in WT hosts (Fig. 6 D). The cytokine profile of CNS-infiltrating donor cells was comparable between the groups (Fig. 6 E).

Serum CXCL1 levels were elevated in Csf3r/ versus WT mice both at baseline and on day 7 after transfer (Fig. 6 F). Similarly, CXCL1 expression was significantly higher in the CNS of Csf3r/ compared with WT hosts at clinical onset (Fig. 6 G). This led us to question whether CXCL1 drives the residual neutrophil mobilization and CNS recruitment that occur in Csf3r/ hosts. In fact, treatment of Csf3r/ hosts with CXCR2 blocking antisera beginning on the day of trans-fer prevented the expansion of neutrophils in the circulation and development of neurological deficits (Fig. 6 H and not depicted). Mice succumbed to clinical disease 3 d after the final treatment, concurrent with rebound recovery of circu-lating neutrophils (data not shown). These data suggest that heightened expression of CXCL1 partially compensates for deficient G-CSF signaling in Csf3r/ hosts.

Neutrophil-related markers correlate with new lesion formation and measures of CNS injury in patients with relapsing MSAs shown in Figs. 2 and 6, CXCL1 levels are elevated in the blood at the onset of clinical EAE, whether induced by active immunization or Th17 transfer. To assess the association be-tween the expression of neutrophil-related factors and lesion formation during relapsing MS, we measured the plasma levels of a panel of chemokines in patients with active or inactive

not change significantly after the transfer of either Th1 or Th17 myelin-reactive T cells (unpublished data). Collectively, the above studies demonstrate for the first time that periph-eral neutrophils are modulated during the development of clinical EAE in the absence of adjuvant.

Plasma G-CSF levels rise and neutrophils expand in the circulation after the adoptive transfer of encephalitogenic Th17 cellsWe next interrogated the kinetics of neutrophil expansion after the adoptive transfer of encephalitogenic Th17 cells. We found that the number of circulating neutrophils consistently increased 1 wk after transfer around the time of clinical onset (Fig. 5 A). Concomitantly, G-CSF levels rose in the blood and the CNS (Fig. 5, B and C) and the numbers of infiltrat-ing myeloid cells peaked in the brain and spinal cord (Fig. 5, D and E). The number of splenic neutrophils also rose but not to a statistically significant extent (Fig. 5 F). In contrast to ac-tively immunized mice, adoptive transfer recipients did not up-regulate G-CSF in non-CNS tissues (Fig. 5 C).

G-CSFR deficiency confers resistance to Th17-mediatedEAE, which is partially rescued by overexpression of CXCL1We compared the clinical manifestation of EAE in Csf3r/ andWT adoptive transfer recipients of MOG-reactive Th17 cells.Csf3r/ hosts experienced a milder disease course (Fig. 6 A),which correlated with a relative diminution in the number ofneutrophils infiltrating the spinal cord at the time of peak EAEand accumulating in the circulation and spleen on day 7 after

Figure 5. Adoptive transfer of encephalitogenic Th17 cells induces the systemic up-regulation of G-CSF and neutrophil mobilization. (A–F) WT micewere injected with IL-23 polarized, MOG-specific CD4+

Th17 cells. (A) Circulating neutrophils (closed circles) andmonocytes (white squares) were enumerated by flowcytometry at serial time points. Data were pooled fromthree experiments (n ≥ 7 mice per group). (B) SerumG-CSF levels were measured by ELISA. Data were pooledfrom three experiments (n ≥ 10 mice per group).(C) G-CSF levels in tissue homogenates obtained fromnaive mice or from host mice on day 7 after transfer,measured by ELISA and normalized to total protein (n = 5mice per group). (D–F) Number of monocytes and neutro-phils in brain (D), spinal cord (E), and spleen (F), deter-mined by flow cytometry. Data were pooled from twoexperiments (n ≥ 6 per group). All graphs indicate means;error bars denote SEM. *, P < 0.05; **, P < 0.01 comparedwith naive or day 3 after transfer.

JEM Vol. 212, No. 1

Article

29

also correlated with overall brain lesion volume, as visualized on T2 weighted sequences. In contrast, brain parenchymal tissue volume correlated inversely with CXCL5, CXCL1, and neu-trophil elastase but was not significantly related to CXCL10 or CCL2.

DISCUSSIONThe formation of inflammatory demyelinating lesions is initi-ated by the trafficking of encephalitogenic T cells across the BBB and their reactivation within the CNS (Kawakami et al., 2004). These events are critical in the pathogenesis of relapsing-remitting MS, as suggested by the therapeutic efficacy of anti-4 integrin monoclonal antibodies that suppress exacer-bations, ostensibly by preventing effector CD4+ T cells from infiltrating the CNS (Stüve et al., 2006). However, adoptive transfer experiments with labeled myelin-reactive effector cells have shown that donor T cells comprise a small percentage of infiltrating leukocytes in established EAE lesions and tend to

disease, as determined by cerebral magnetic resonance imag-ing (MRI) and clinical course. We found that expression of CXCL5, a CXCR2 binding chemokine which activates and attracts neutrophils (Liu et al., 2011; Mei et al., 2012), was elevated in active patients concomitant with the presence of acute inflammatory lesions on MRI scan when compared with inactive patients with no inflammatory lesions (Fig. 7 A). In contrast, there was no significant association between new lesion formation and expression of CXCL10 or CCL2, chemokines that primarily target lymphocytes and monocytes, respectively (Fig. 7, B and C). The expanded disability status scale (EDSS) score, a measure of MS clinical disability (Kurtzke, 1983), correlated directly with plasma levels of CXCL1, CXCL5, and neutrophil elastase, as well as CCL2 (Table 1). CXCL1, CXCL5, and neutrophil elastase, but not CXCL10 or CCL2, correlated directly with cumulative MRI lesion volume on T1 weighted sequences, indicative of extensive tissue damage and axonal loss. CXCL5 and neutrophil elastase expression

Figure 6. ELR+ CXC chemokines partially compensate for loss of G-CSF signaling in Csf3r/ adoptive transfer recipients. (A–G) WT (closed triangles, black bars) and Csf3r/ (open triangles, white bars) mice were injected with MOG-specific Th17 cells. (A) Mean clinical scores, representative of seven independent experiments (n ≥ 7 mice per group). (B) Absolute number of neutrophils, monocytes, and microglia recovered from the spinal cord on day 7 after transfer, assessed by flow cytometry (n ≥ 7 per group, pooled from two experiments). (C) Number of neutrophils per milliliter of blood or per spleen at baseline and on day 7 d after transfer, assessed by flow cytometry (n ≥ 7 per group, pooled from two experiments). (D) Fold change in the num-ber of circulating and splenic neutrophils over baseline on day 7 after transfer. (E) Proportion of donor cells expressing IL-17 and IFN- immediately before adoptive transfer, and after isolation from the spinal cords of WT or Csf3r/ hosts on day 7 after transfer (n = 5 per group). (F) Levels of CXCL1 in sera from naive mice and adoptive transfer recipients, measured by ELISA (n ≥ 9, pooled from three experiments). (G) CXCL1 levels in spinal cord homog-enates were measured by ELISA and normalized to total protein (n ≥ 4). (H) Csf3r/ recipients of WT Th17 cells were treated with control serum (n = 7) or anti-CXCR2 (n = 6) every other day from days 0– 8 (arrows). Data are representative of two independent experiments. All graphs indicate means; error bars denote SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

30 Neutrophil-related factors in EAE and MS | Rumble et al.

as the spinal cord. The risk of MS relapse is higher in the set-ting of infection (Sibley et al., 1985; Rapp et al., 1995). Our results suggest that this association could be secondary, at least in part, to the release of Toll-like receptor ligands that modu-late myeloid-related factors. Campbell et al. (2010) detected increased hepatic expression of CXCL1, accompanied by neu-trophil recruitment to the liver, in Biozzi mice that were im-munized with spinal cord homogenate in CFA to induce EAE. Those authors also detected neutrophil infiltration in postmor-tem liver tissue from MS patients. Conversely, we found the CNS to be the primary site of G-CSF expression in the Th17 adoptive transfer model. Interestingly, Lock et al. (2002) found that G-CSF transcripts are present in MS lesions but not normal-appearing white matter. Because encephalitogenic Th17 cells secrete cytokines, such as IL-17A and TNF, that directly induce G-CSF and CXCL1 (Witowski et al., 2000; Kolls and Lindén, 2004; Iwakura et al., 2011), they could drive local production of neutrophil activating/chemoattractant factors upon being reactivated in the CNS. The cellular source of G-CSF and CXCL1 in the CNS may be meningeal epithelialcells (as previously reported in MOG-immunized mice; Soulikaet al., 2009), astrocytes, neurons, or cerebrovascular endothe-lial cells (Lenhoff and Olofsson, 1996; Jacob et al., 2012). Weare currently performing experiments to distinguish betweenthose possibilities.

In the current manuscript we demonstrate, for the first time that spikes in plasma levels of CXCL5, an ELR+ CXC chemokine, correspond with the development of new in-flammatory lesions in relapsing MS patients. Furthermore, ex-pression of CXCL1, CXCL5, and neutrophil elastase correlated with clinical and radiological measures of CNS injury in MS. These findings are consistent with recent reports of elevated

cluster in the perivascular or subpial space. Host phagocytes migrate to the CNS in a secondary wave, penetrate deep into the CNS parenchyma, and directly inflict damage to myelin, glial cells, and axons. The recruitment of innate immune cells from the periphery correlates with BBB breakdown and the clinical onset of autoimmune demyelinating disease (Ajami et al., 2011). In the current study, we show that neutrophils and monocytes expand in the bone marrow and accumulate in the circulation immediately before clinical exacerbations after either active immunization of C57BL/6 mice or injec-tion of Th1- or Th17-polarized encephalitogenic T cells. This substantiates our earlier observation that the frequency of cir-culating myeloid progenitor cells (measured as GM-CFU) in-creases in concert with the onset and relapse of EAE (King et al., 2009). These shifts in peripheral myeloid populations are driven by systemic up-regulation of G-CSF and ELR+ CXC chemokines. Reminiscent of their role in infectious diseases (Wengner et al., 2008), G-CSF and CXCL1 act synergistically to promote neutrophil mobilization during EAE. We found that Csf3r/ mice are relatively resistant to EAE induced by adoptive transfer, consistent with a role of neutrophils dur-ing the effector phase (McColl et al., 1998). Administration of pharmacological doses of recombinant G-CSF during EAE has yielded conflicting results (Lock et al., 2002; Zavala et al., 2002; Verda et al., 2006), reflecting the pleiotropic effects of the molecule. Compensatory pathways may be engaged in G-CSF–treated mice depending on the dosing regimen andtiming of administration.

In the active immunization model, mycobacterial com-ponents (most likely pathogen-associated molecular patterns) in CFA stimulate expression of G-CSF and CXCL1 in nu-merous tissues, including the spleen, liver, and lungs, as well

Figure 7. Plasma CXCL5, but not CCL2 or CXCL10, levels increase in association with new MS lesion formation. Patients with relapsing MS were classified as having “active” or “inactive” disease based on clinical course, neurological examination, and MRI scanning. Plasma levels of CXCL5, CCL2, and CXCL10 were measured by multiplex assay. Patients with active disease had enhancing MRI lesions and patients with inactive disease had no en-hancing lesions on the day of phlebotomy. Box plots show median, interquartile range, sample minimum, and maximum. Circles show outliers.

Table 1. Relationship between chemokine levels and clinical/radiological parameters

Parameter CXCL1 CXCL5 Neutrophil elastase CXCL10 CCL2

Correlation (R)/Significance (p-value)EDSS 0.58/<0.00001 0.43/<0.0001 0.26/0.029 NS 0.38/0.0006T2 lesion volume NS 0.47/0.0009 0.46/0.0043 NS NST1 lesion volume 0.47/0.0039 0.46/0.0005 0.42/0.0097 NS NSBPV 0.54/<0.0001 0.38/0.0019 0.45/<0.0001 NS NS

NS, not significant; BPV, brain parenchymal tissue volume.

JEM Vol. 212, No. 1

Article

31

BBB breakdown is prevented in myelin-immunized mice by treatment with neutrophil-depleting or blocking antibodies (Carlson et al., 2008).

Recent advances in MS therapeutics have revolved around strategies that target lymphocytes. Agents that block lymphocyte trafficking to the CNS (such as natalizumab and fingolimod) or deplete lymphocytes in the circulation (such as alemtu-zumab and ocrelizumab) are currently used in the clinic to de-crease the risk of MS relapse, or have yielded promising results in clinical trials. However, none of those agents fully suppress disease activity or are effective in all patients, underscoring the need for innovative medicinal approaches. The current study supports a growing body of evidence suggesting that myeloid cells and the factors critical for their survival, expan-sion, activation, and mobilization should be investigated as an alternative source of novel biomarkers and therapeutic tar-gets in MS.

MATERIALS AND METHODSMice. C57BL/6 and B6-Ly5.2/Cr mice were obtained from the National Cancer Institute. Csf3r/ mice were initially provided by D.C. Link (Wash-ington University School of Medicine, St. Louis, MO; Liu et al., 1996; Christopher et al., 2011) and bred in our vivarium. Double transgenic OSE mice (Krishnamoorthy et al., 2006) were bred in animal facilities at the Max Planck Institute of Neurobiology. All animals were housed under specific pathogen-free conditions.

Antibodies and reagents. The following monoclonal antibodies were used for flow cytometry: anti-Ly6C, anti-CD31, and anti-Ly6G (BD); and anti-CD4, anti-CD11b, anti-CD45.1, and anti-CD45.2 (eBioscience). The following monoclonal antibodies were used for ELIspot assays: anti–IL-17 (TC11-18H10), biotinylated anti–IL-17 (TC11-8H4), anti–IFN- (AN18), and biotinylated anti–IFN- (R4-6A2; e-Bioscience). Recombinant mouse IFN- and IL-12 were from R&D Systems.

Induction and assessment of EAE. Mice were immunized subcutaneously over the flanks with 100 µg MOG35-55 (MEVGWYRSP-FSRVVHLYRNGK; Biosynthesis) in CFA containing 250 µg Mycobacterium tuberculosis H37RA (BD). For induction of EAE by active immunization, animals were injected i.p. with heat inactivated PTx (List Biological Laboratories) on days 0 and 2.

For induction of EAE by adoptive transfer, B6-Ly5.2/Cr mice were immunized as described, but without administration of PTx. 10–14 d later, draining lymph nodes (inguinal, brachial, and axillary) were harvested and dissociated into single cell suspensions. Cells were cultured in standard me-dium with 50 µg/ml MOG35-55 and either Th17-polarizing factors (rmIL-23 at 8 ng/ml, rmIL-1 at 10 ng/ml, anti–mIFN- [clone XMG1.2] at 10 µg/ml, and anti-IL-4 [clone 11B11] at 10 µg/ml) or Th1-polarizing factors (rmIL-12 at 6 ng/ml, rmIFN- at 2 ng/ml, and anti–IL-4 [clone 11B11] at 10 µg/ml). After 4 d in culture, 2.5–5 × 106 CD4+ T cells were injected i.p. into naive congenic recipients. Clinical assessment of EAE was performedaccording to the following scale: 0, no disease; 1, limp tail; 2, hind-limbweakness; 3, partial hind-limb paralysis; 4, complete paralysis of hind-limbs;and 5, moribund state.

Anti-CXCR2 treatment. Adoptive transfer recipients were injected with rabbit anti-CXCR2 antisera (Biosynthesis) or control rabbit sera (Sigma-Aldrich) on alternate days beginning on the day of cell transfer. Antisera was generated against a mCXCR2 peptide (GCMGEFKVDKFNIEDFFSG; Mehrad et al., 1999; Hosking et al., 2009), and each bleed was tested for its efficacy in blocking neutrophil recruitment to the peritoneal cavity in re-sponse to thioglycollate administration. Only bleeds that showed >90% re-duction in neutrophil recruitment were used in studies.

numbers and enhanced priming of circulating neutrophils in patients with active MS (Naegele et al., 2012), as well as with the association between MS relapse and G-CSF administra-tion (Openshaw et al., 2000). Expression of CCL2, as well as CXCL1, CXCL5, and neutrophil elastase, all correlated with EDSS, possibly reflecting complementary roles of neutrophils and monocytes in MS pathogenesis.

Previous studies on the role of innate immunity in MS and EAE have focused on the monocyte/macrophage lineage because that subset is prominent in established white matter lesions. Although CNS-infiltrating neutrophils are prevalent in alternative forms of autoimmune demyelinating disease, such as neuromyelitis optica and acute disseminated encepha-lomyelitis, they are scarce in typical MS infiltrates. This might be interpreted as inconsistent with a role of neutrophil-related factors in MS (Godiska et al., 1995). However, EAE studies from our laboratory and others indicate that neutrophils par-ticipate in autoimmune demyelination during early lesion development, preceding the development of overt neurolog-ical deficits (Carlson et al., 2008; Soulika et al., 2009; Christy et al., 2013). 1–2 d before expected clinical onset, neutrophils comprise a significant percentage of CNS-infiltrating cells. By the time mice exhibit their first neurological signs, neutro-phils are greatly outnumbered by monocytes/macrophages. Hence, the paradox that neutrophils appear to be functionally important despite their scarcity in MS lesions may be a con-sequence of sampling bias, in that the vast majority of autop-sied CNS tissues are from patients with long-standing progressive disease, and biopsies are generally performed after lesions have matured over days, if not weeks. An alternative, but not mu-tually exclusive, explanation is that neutrophils primarily pro-mote neuroinflammation from the periphery. Activation of neutrophils in the bone marrow indirectly triggers the mobi-lization of monocytes and hematopoietic progenitor/stem cells (HPSCs) into the blood, thereby making them accessible for recruitment to the CNS (Singh et al., 2012). Some monocytic cells and HPSCs express G-CSFR themselves and, thus, can be directly stimulated by G-CSF to expand and migrate from the bone marrow (Demetri and Griffin, 1991; Christopher et al., 2011). It is also possible that neutrophils mediate increased permeability at the cerebrovascular interface without actually crossing the BBB. Neutrophils crawl on the luminal endo-thelial surface of CNS blood vessels before, and at the time of, clinical EAE onset (Richard et al., 2011). The adhesion of activated neutrophils to cerebrovascular endothelial cells alone may impair interendothelial cell-to-cell contacts via secretion of proteases and free oxygen radicals or by conferring confor-mational changes to adherens junctional proteins (Smedly et al., 1986; Tinsley et al., 1999; Scholz et al., 2007). Neutrophils have been shown to mediate BBB breakdown in animal models of brain trauma, cerebral hemorrhage, and viral encephalitis, and in response to intracerebral injection of recombinant IL-1 or CXCL2 (Bell et al., 1996; Anthony et al., 1997, 1998; Zhou et al., 2003; Scholz et al., 2007; Moxon-Emre and Schlichter, 2011). A comparable role of neutrophils in auto-immune demyelinating disease is suggested by the fact that

32 Neutrophil-related factors in EAE and MS | Rumble et al.

Multiplex assays. Plasma levels of cytokines and chemokines were mea-sured with customized multiplex magnetic bead based arrays (EMD Milli-pore) according to the manufacturer’s protocol. Data were collected using the Bio-Plex 200 system (Bio-Rad Laboratories). Standards were run in parallel to allow quantification of individual factors. The data shown indicates levels that fell within the linear portion of the corresponding standard curve.

MRI protocol and image analysis. All patients were evaluated with cra-nial MRI examinations on a 1.5 tesla strength magnet using axial T2-weighted, axial and sagittal T1-weighted sequences, and post-Gadolinium axial and coronal T1-weighted. Brain parenchymal tissue volume (defined as total brain volume subtracting CSF volume), and T1 and T2 lesion volume, were mea-sured using commercially available software developed by VirtualScopics. This involved coregistering each MRI to a presegmented anatomical atlas with manual refinement of automated brain boundaries by an expert analyst where necessary, as previously described (Ashton et al., 2003). Lesion bound-aries were identified in three dimensions using geometrically constrained re-gion growth (GEORG; Ashton et al., 1997, 2003). T1 and T2 lesion volumes were normalized to total brain parenchymal tissue volume.

Statistical analysis. Clinical courses of WT and Csf3r/ mice were compared by two-way ANOVA using Prism (GraphPad Software). Disease-free survival curves of anti-CXCR2 and control antisera treated mice were generated with Prism software and analyzed by Log-rank (Mantel-Cox) test. Immune parame-ters were compared between groups of mice by unpaired Student’s t tests.

Plasma chemokine levels were measured in patients with active disease on the day of their first MRI scan with a gadolinium enhancing lesion (in-dicative of acute lesion formation) and in inactive patients on the day of the initial blood draw, during which they had an MRI scan showing no enhanc-ing lesions. Levels between the active and inactive groups were compared using Box plots and Wilcoxon nonparametric tests.

Comparisons were made between immune parameters and EDSS, T1 lesion volume, T2 lesion volume, or brain parenchymal tissue volume using all available measurements. The analysis was done using within-cluster resam-pling methodology, considering the concern of the intraclass correlation. Specifically, we have randomly chosen one observation per patient and cal-culated Spearman’s correlations. This process was repeated 200×, and we merged all 200 estimated correlations to obtain the final results using estab-lished methods (Hoffman et al., 2001; Rieger and Weinberg, 2002).

Study approval. All animal experiments described herein were performed under protocols approved by the University of Michigan Committee on the Use and Care of Animals. The Institutional Review Boards of the University of Michigan and the University of Rochester approved our human study protocol. All subjects gave their written informed voluntary consent after the nature and possible consequences of the study were explained.

This research was supported by grants from the NINDS, NIH (R01 NS057670), and Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Rehabilitation and Development Service (B7545-R) to B.M. Segal. B.M. Segal is a Scholar of the A. Alfred Taubman Medical Research Institute. G. Krishnamoorthy is supported by the grants from Deutsche Forschungsgemeinschaft SFB TR 128, the German Competence Network on Multiple Sclerosis (KKNMS), Hertie foundation, and the Max Planck Society.

The authors declare no competing financial interests.

Submitted: 27 May 2014Accepted: 11 December 2014

REFERENCESAjami, B., J.L. Bennett, C. Krieger, K.M. McNagny, and F.M. Rossi. 2011.

Infiltrating monocytes trigger EAE progression, but do not contribute to the resident microglia pool. Nat. Neurosci. 14:1142–1149. http://dx.doi.org/10.1038/nn.2887

Anthony, D.C., S.J. Bolton, S. Fearn, and V.H. Perry. 1997. Age-related effects of interleukin-1 on polymorphonuclear neutrophil-dependent

Cell isolation. Blood was collected in EDTA-coated or heparinized tubes after cardiac puncture or retro-orbital bleed. Bone marrow cells were flushed from femurs and tibiae with cold media. Splenocytes were passed through a 70-µm nylon mesh filter to obtain a single-cell suspension. All cell prepara-tions were ACK lysed and washed before analysis. For collection of CNSmononuclear cells, spinal cords were harvested from PBS-perfused mice, me-chanically disrupted with an 18-gauge needle, and incubated with 2 µg/mlcollagenase A and 1 µg/ml DNase in HBSS for 30 min at 37°C. Mononuclearcells were then isolated on a 30–70% discontinuous Percoll gradient.

Flow cytometry. Cells were suspended in PBS with 2% FCS and Fc Block (50 ng/ml; hybridoma 2.4G2; American Type Culture Collection) before incu-bation with fluorochrome-conjugated antibodies (anti–mouse Ly-6C [AL21; BD], anti–mouse Ly-6G [1A8; BD], anti–mouse CD11b [M1/70; eBiosci-ence], anti–mouse CD4 [RM-45, eBioscience], anti–mouse CD45 [30-F11; eBioscience], and anti–mouse CD31 [MEC13.3; BD]). For intracellular cyto-kine staining, cells were stimulated with 50 ng/ml PMA and 2 µg/ml ionomy-cin in the presence of 10 µg/ml brefeldin A for 6 h, followed by fixation with 4% paraformaldehyde, permeabilization with 0.5% saponin, and staining with fluorochrome-conjugated anti–IFN- (XMG1.2; BD) and anti–IL-17 (17B7; eBioscience). Data were acquired on a FACSVerse or FACS Canto II flow cy-tometer (BD) and analyzed with FlowJo software (Tree Star).

ELISA. Sera was collected by cardiac puncture. Livers, spleens, spinal cords, and lungs were collected from PBS-perfused mice and homogenized in PBS with protease inhibitors (Roche). CXCL1 and G-CSF levels were measured by sandwich ELISA based on noncompeting pairs of antibodies using the manufacturer’s protocol (R&D Systems). Levels in homogenates were nor-malized to total protein as determined by Bradford assay.

ELISpot assay. Draining lymph nodes (inguinal, axillary, and brachial) were harvested from mice on day 11 p.i., pooled, and cultured in 96-well filter plates (EMD Millipore) at 2–5 × 105 cells/well, with or without 50 µg/ml MOG35-55, for 18 h. Streptavidin-alkaline phosphatase was from Southern-Biotech and the alkaline phosphatase substrate kit was from Vector Labora-tories. Spots were counted using the CTL ImmunoSpot Analyzer (Cellular Technology) with ImmunoSpot software.

Histology. After perfusion, spinal cords were fixed in 4% paraformalde-hyde, decalcified, and processed for paraffin embedding. 8-µm sections were cut with a microtome (Leica) and stained with Harris’ hematoxylin and eosin by standard procedure.

Bone marrow chimeras. Femurs and tibiae from C57BL/6 WT and Csf3r/ females were flushed with PBS using a 26-gauge needle and ACK lysed. Cells were suspended in cold PBS for intravenous administration to 6-wk-old B6-Ly5.2/Cr mice, which had been subjected to 13 Gy of irradia-tion (orthovoltage x-ray source), split into two doses, 3 h apart. Recipientswere given 6 × 106 total cells and allowed to reconstitute for 6 wk beforefurther use. Chimerism was verified by flow cytometric analysis of PBMC,which was >90% in all animals tested.

MS subjects and clinical assessments. Patients diagnosed with relapsing MS (n = 22), based on the revised McDonald Diagnostic Criteria, were re-cruited from the Multiple Sclerosis Clinic at the University of Michigan. All subjects underwent bimonthly gadolinium enhanced MRI scans and phle-botomy on the same day over a 1-yr observation period. Subjects who had at least one enhancing lesion on one or more MRI scan were classified as hav-ing active disease. Subjects who had no enhancing lesions detected through-out the course of study and who had no clinical relapses were classified as having inactive disease. Every MS patient underwent complete neurological examinations, including EDSS scoring by a board-certified neurologist who was blinded to the results of MRI scanning and immune assays. On the day of phlebotomy, all subjects were afebrile and had no signs and symptoms of infection based on history, physical examination, and responses to a survey.

JEM Vol. 212, No. 1

Article

33

Guarnieri, B., F. Lolli, and L. Amaducci. 1985. Polymorphonuclear neutral protease activity in multiple sclerosis and other diseases. Ann. Neurol. 18:620–622. http://dx.doi.org/10.1002/ana.410180519

Hoffman, E.B., P.K. Sen, and C.R. Weinberg. 2001. Within-cluster resa-mpling. Biometrika. 88:1121–1134. http://dx.doi.org/10.1093/biomet/ 88.4.1121

Hosking, M.P., L. Liu, R.M. Ransohoff, and T.E. Lane. 2009. A protective role for ELR+ chemokines during acute viral encephalomyelitis. PLoS Pathog. 5:e1000648. http://dx.doi.org/10.1371/journal.ppat.1000648

Ishizu, T., M. Osoegawa, F.J. Mei, H. Kikuchi, M. Tanaka, Y. Takakura, M. Minohara, H. Murai, F. Mihara, T. Taniwaki, and J. Kira. 2005.Intrathecal activation of the IL-17/IL-8 axis in opticospinal multiple scle-rosis. Brain. 128:988–1002. http://dx.doi.org/10.1093/brain/awh453

Iwakura, Y., H. Ishigame, S. Saijo, and S. Nakae. 2011. Functional spe-cialization of interleukin-17 family members. Immunity. 34:149–162. http://dx.doi.org/10.1016/j.immuni.2011.02.012

Jacob, A., S. Saadoun, J. Kitley, M. Leite, J. Palace, F. Schon, and M.C. Papadopoulos. 2012. Detrimental role of granulocyte-colony stimulating factor in neuromyelitis optica: clinical case and histological evidence. Mult. Scler. 18:1801–1803. http://dx.doi.org/10.1177/1352458512443994

Kadota, R., M. Koda, J. Kawabe, M. Hashimoto, Y. Nishio, C. Mannoji, T. Miyashita, T. Furuya, A. Okawa, K. Takahashi, and M. Yamazaki.2012. Granulocyte colony-stimulating factor (G-CSF) protects oligo-dendrocyte and promotes hindlimb functional recovery after spinal cordinjury in rats. PLoS ONE. 7:e50391. http://dx.doi.org/10.1371/journal.pone.0050391

Karni, A., D.N. Koldzic, P. Bharanidharan, S.J. Khoury, and H.L. Weiner. 2002. IL-18 is linked to raised IFN-gamma in multiple sclerosis and is induced by activated CD4+ T cells via CD40-CD40 ligand interactions. J. Neuroimmunol. 125:134–140. http://dx.doi.org/10.1016/S0165-5728(02)00018-8

Karni, A., M. Abraham, A. Monsonego, G. Cai, G.J. Freeman, D. Hafler, S.J. Khoury, and H.L. Weiner. 2006. Innate immunity in multiple scle-rosis: myeloid dendritic cells in secondary progressive multiple sclerosis are activated and drive a proinflammatory immune response. J. Immunol. 177:4196–4202. http://dx.doi.org/10.4049/jimmunol.177.6.4196

Kawakami, N., S. Lassmann, Z. Li, F. Odoardi, T. Ritter, T. Ziemssen, W.E. Klinkert, J.W. Ellwart, M. Bradl, K. Krivacic, et al. 2004. The ac-tivation status of neuroantigen-specific T cells in the target organ deter-mines the clinical outcome of autoimmune encephalomyelitis. J. Exp. Med. 199:185–197. http://dx.doi.org/10.1084/jem.20031064

Khader, S.A., S.L. Gaffen, and J.K. Kolls. 2009. Th17 cells at the crossroads of innate and adaptive immunity against infectious diseases at the mucosa. Mucosal Immunol. 2:403–411. http://dx.doi.org/10.1038/mi.2009.100

King, I.L., T.L. Dickendesher, and B.M. Segal. 2009. Circulating Ly-6C+ myeloid precursors migrate to the CNS and play a pathogenic role dur-ing autoimmune demyelinating disease. Blood. 113:3190–3197. http://dx.doi.org/10.1182/blood-2008-07-168575

Kolls, J.K., and A. Lindén. 2004. Interleukin-17 family members and inflam-mation. Immunity. 21:467–476. http://dx.doi.org/10.1016/j.immuni .2004.08.018

Kostic, M., T. Dzopalic, S. Zivanovic, N. Zivkovic, A. Cvetanovic, I. Stojanovic, S. Vojinovic, G. Marjanovic, V. Savic, and M. Colic. 2014. IL-17 andglutamate excitotoxicity in the pathogenesis of multiple sclerosis. Scand. J. Immunol. 79:181–186. http://dx.doi.org/10.1111/sji.12147

Kowarik, M.C., H.L. Pellkofer, S. Cepok, T. Korn, T. Kümpfel, D. Buck, R. Hohlfeld, A. Berthele, and B. Hemmer. 2011. Differential effectsof fingolimod (FTY720) on immune cells in the CSF and blood of pa-tients with MS. Neurology. 76:1214–1221. http://dx.doi.org/10.1212/WNL.0b013e3182143564

Krishnamoorthy, G., H. Lassmann, H. Wekerle, and A. Holz. 2006. Spontaneous opticospinal encephalomyelitis in a double-transgenic mouse model of autoimmune T cell/B cell cooperation. J. Clin. Invest. 116:2385–2392. http://dx.doi.org/10.1172/JCI28330

Kroenke, M.A., and B.M. Segal. 2011. IL-23 modulated myelin-specific T cells induce EAE via an IFN driven, IL-17 independent pathway. Brain Behav. Immun. 25:932–937. http://dx.doi.org/10.1016/j.bbi.2010 .10.001

increases in blood-brain barrier permeability in rats. Brain. 120:435–444. http://dx.doi.org/10.1093/brain/120.3.435

Anthony, D., R. Dempster, S. Fearn, J. Clements, G. Wells, V.H. Perry, and K. Walker. 1998. CXC chemokines generate age-related increases in neutrophil-mediated brain inflammation and blood-brain barrier breakdown. Curr. Biol. 8:923–927. http://dx.doi.org/10.1016/S0960- 9822(07)00373-9

Aoki, Y., T. Miyatake, N. Shimizu, and M. Yoshida. 1984. Medullasin ac-tivity in granulocytes of patients with multiple sclerosis. Ann. Neurol. 15:245–249. http://dx.doi.org/10.1002/ana.410150307

Ashton, E.A., K.J. Parker, M.J. Berg, and C.W. Chen. 1997. A novel volumet-ric feature extraction technique with applications to MR images. IEEE Trans. Med. Imaging. 16:365–371. http://dx.doi.org/10.1109/42.611343

Ashton, E.A., C. Takahashi, M.J. Berg, A. Goodman, S. Totterman, and S. Ekholm. 2003. Accuracy and reproducibility of manual and semiauto-mated quantification of MS lesions by MRI. J. Magn. Reson. Imaging. 17:300–308. http://dx.doi.org/10.1002/jmri.10258

Balashov, K.E., D.R. Smith, S.J. Khoury, D.A. Hafler, and H.L. Weiner. 1997. Increased interleukin 12 production in progressive multiple scle-rosis: induction by activated CD4+ T cells via CD40 ligand. Proc. Natl. Acad. Sci. USA. 94:599–603. http://dx.doi.org/10.1073/pnas.94.2.599

Basu, S., G. Hodgson, H.H. Zhang, M. Katz, C. Quilici, and A.R. Dunn. 2000. “Emergency” granulopoiesis in G-CSF-deficient mice in response to Candida albicans infection. Blood. 95:3725–3733.

Becher, B., and B.M. Segal. 2011. T(H)17 cytokines in autoimmune neuro-inflammation. Curr. Opin. Immunol. 23:707–712. http://dx.doi.org/10 .1016/j.coi.2011.08.005

Bell, M.D., D.D. Taub, and V.H. Perry. 1996. Overriding the brain’s intrinsic resistance to leukocyte recruitment with intraparenchymal injections of recombinant chemokines. Neuroscience. 74:283–292. http://dx.doi.org/ 10.1016/0306-4522(96)00083-8

Burt, R.K., A. Fassas, J. Snowden, J.M. van Laar, T. Kozak, N.M. Wulffraat, R.A. Nash, C.E. Dunbar, R. Arnold, G. Prentice, et al. 2001. Collection of hematopoietic stem cells from patients with autoimmune diseases. Bone Marrow Transplant. 28:1–12. http://dx.doi.org/10.1038/sj.bmt .1703081

Campbell, S.J., U. Meier, S. Mardiguian, Y. Jiang, E.T. Littleton, A. Bristow, J. Relton, T.J. Connor, and D.C. Anthony. 2010. Sickness behaviour is induced by a peripheral CXC-chemokine also expressed in multiple sclerosis and EAE. Brain Behav. Immun. 24:738–746. http://dx.doi.org/10.1016/j.bbi.2010.01.011

Carlson, T., M. Kroenke, P. Rao, T.E. Lane, and B. Segal. 2008. The Th17-ELR+ CXC chemokine pathway is essential for the development of central nervous system autoimmune disease. J. Exp. Med. 205:811–823. http://dx.doi.org/10.1084/jem.20072404

Christopher, M.J., M. Rao, F. Liu, J.R. Woloszynek, and D.C. Link. 2011. Expression of the G-CSF receptor in monocytic cells is sufficient to me-diate hematopoietic progenitor mobilization by G-CSF in mice. J. Exp. Med. 208:251–260. http://dx.doi.org/10.1084/jem.20101700

Christy, A.L., M.E. Walker, M.J. Hessner, and M.A. Brown. 2013. Mast cell activation and neutrophil recruitment promotes early and robust inflam-mation in the meninges in EAE. J. Autoimmun. 42:50–61. http://dx .doi.org/10.1016/j.jaut.2012.11.003

Comabella, M., K. Balashov, S. Issazadeh, D. Smith, H.L. Weiner, and S.J. Khoury. 1998. Elevated interleukin-12 in progressive multiple sclerosis correlates with disease activity and is normalized by pulse cyclophospha-mide therapy. J. Clin. Invest. 102:671–678. http://dx.doi.org/10.1172/ JCI3125

Demetri, G.D., and J.D. Griffin. 1991. Granulocyte colony-stimulating fac-tor and its receptor. Blood. 78:2791–2808.

Epstein, L.G., J.W. Prineas, and C.S. Raine. 1983. Attachment of myelin to coated pits on macrophages in experimental allergic encephalomyeli-tis. J. Neurol. Sci. 61:341–348. http://dx.doi.org/10.1016/0022-510X (83)90167-3

Godiska, R., D. Chantry, G.N. Dietsch, and P.W. Gray. 1995. Chemokine expression in murine experimental allergic encephalomyelitis. J. Neuroimmunol. 58:167–176. http://dx.doi.org/10.1016/0165-5728(95) 00008-P

34 Neutrophil-related factors in EAE and MS | Rumble et al.

pertussis toxin exposure through IL6, ICAM1 and ITGM. Brain Pathol. 21:661–671. http://dx.doi.org/10.1111/j.1750-3639.2011.00490.x

Rieger, R.H., and C.R. Weinberg. 2002. Analysis of clustered binary out-comes using within-cluster paired resampling. Biometrics. 58:332–341. http://dx.doi.org/10.1111/j.0006-341X.2002.00332.x

Sawcer, S., G. Hellenthal, M. Pirinen, C.C. Spencer, N.A. Patsopoulos, L. Moutsianas, A. Dilthey, Z. Su, C. Freeman, S.E. Hunt, et al. Wellcome Trust Case Control Consortium 2. 2011. Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature. 476:214–219. http://dx.doi.org/10.1038/nature10251

Scholz, M., J. Cinatl, M. Schädel-Höpfner, and J. Windolf. 2007. Neutrophils and the blood-brain barrier dysfunction after trauma. Med. Res. Rev. 27:401–416. http://dx.doi.org/10.1002/med.20064

Sibley, W.A., C.R. Bamford, and K. Clark. 1985. Clinical viral infections and multiple sclerosis. Lancet. 325:1313–1315. http://dx.doi.org/10 .1016/S0140-6736(85)92801-6

Singh, P., P. Hu, J. Hoggatt, A. Moh, and L.M. Pelus. 2012. Expansion of bone marrow neutrophils following G-CSF administration in mice results in osteolineage cell apoptosis and mobilization of hematopoi-etic stem and progenitor cells. Leukemia. 26:2375–2383. http://dx.doi .org/10.1038/leu.2012.117

Smedly, L.A., M.G. Tonnesen, R.A. Sandhaus, C. Haslett, L.A. Guthrie, R.B. Johnston Jr., P.M. Henson, and G.S. Worthen. 1986. Neutrophil-mediated injury to endothelial cells. Enhancement by endotoxin and es-sential role of neutrophil elastase. J. Clin. Invest. 77:1233–1243. http:// dx.doi.org/10.1172/JCI112426

Soulika, A.M., E. Lee, E. McCauley, L. Miers, P. Bannerman, and D. Pleasure. 2009. Initiation and progression of axonopathy in experi-mental autoimmune encephalomyelitis. J. Neurosci. 29:14965–14979. http://dx.doi.org/10.1523/JNEUROSCI.3794-09.2009

Steinbach, K., M. Piedavent, S. Bauer, J.T. Neumann, and M.A. Friese. 2013. Neutrophils amplify autoimmune central nervous system infil-trates by maturing local APCs. J. Immunol. 191:4531–4539. http://dx.doi.org/10.4049/jimmunol.1202613

Steinman, L., and S.S. Zamvil. 2006. How to successfully apply animal stud-ies in experimental allergic encephalomyelitis to research on multiple sclerosis. Ann. Neurol. 60:12–21. http://dx.doi.org/10.1002/ana.20913

Stüve, O. 2008. The effects of natalizumab on the innate and adaptive im-mune system in the central nervous system. J. Neurol. Sci. 274:39–41. http://dx.doi.org/10.1016/j.jns.2008.03.022

Stüve, O., C.M. Marra, A. Bar-Or, M. Niino, P.D. Cravens, S. Cepok, E.M. Frohman, J.T. Phillips, G. Arendt, K.R. Jerome, et al. 2006.Altered CD4+/CD8+ T-cell ratios in cerebrospinal fluid of natalizumab-treated patients with multiple sclerosis. Arch. Neurol. 63:1383–1387.http://dx.doi.org/10.1001/archneur.63.10.1383

Tinsley, J.H., M.H. Wu, W. Ma, A.C. Taulman, and S.Y. Yuan. 1999. Activated neutrophils induce hyperpermeability and phosphorylation of adherens junction proteins in coronary venular endothelial cells. J. Biol. Chem. 274:24930–24934. http://dx.doi.org/10.1074/jbc.274.35.24930

Toft-Hansen, H., R.K. Nuttall, D.R. Edwards, and T. Owens. 2004. Key metalloproteinases are expressed by specific cell types in experimental autoimmune encephalomyelitis. J. Immunol. 173:5209–5218. http://dx .doi.org/10.4049/jimmunol.173.8.5209

Vaknin-Dembinsky, A., K. Balashov, and H.L. Weiner. 2006. IL-23 is in-creased in dendritic cells in multiple sclerosis and down-regulation of IL-23 by antisense oligos increases dendritic cell IL-10 production. J. Immunol. 176:7768–7774. http://dx.doi.org/10.4049/jimmunol.176.12.7768

Verda, L., K. Luo, D.A. Kim, D. Bronesky, A.P. Kohm, S.D. Miller, L. Statkute, Y. Oyama, and R.K. Burt. 2006. Effect of hematopoietic growth factors on severity of experimental autoimmune encephalomy-elitis. Bone Marrow Transplant. 38:453–460. http://dx.doi.org/10.1038/ sj.bmt.1705473

Wengner, A.M., S.C. Pitchford, R.C. Furze, and S.M. Rankin. 2008. The coordinated action of G-CSF and ELR + CXC chemokines in neutro-phil mobilization during acute inflammation. Blood. 111:42–49. http://dx.doi.org/10.1182/blood-2007-07-099648

Witowski, J., K. Pawlaczyk, A. Breborowicz, A. Scheuren, M. Kuzlan-Pawlaczyk, J. Wisniewska, A. Polubinska, H. Friess, G.M. Gahl, U.

Kroenke, M.A., T.J. Carlson, A.V. Andjelkovic, and B.M. Segal. 2008. IL-12- and IL-23-modulated T cells induce distinct types of EAE based on histology, CNS chemokine profile, and response to cytokine inhibition. J. Exp. Med. 205:1535–1541. http://dx.doi.org/10.1084/jem.20080159

Kurtzke, J.F. 1983. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology. 33:1444–1452. http://dx.doi.org/10.1212/WNL.33.11.1444

Lenhoff, S., and T. Olofsson. 1996. Cytokine regulation of GM-CSF and G-CSF secretion by human umbilical cord vein endothelial cells (HUVEC). Cytokine. 8:702–709. http://dx.doi.org/10.1006/cyto.1996.0093

Lin, R.F., T.S. Lin, R.G. Tilton, and A.H. Cross. 1993. Nitric oxide local-ized to spinal cords of mice with experimental allergic encephalomyelitis: an electron paramagnetic resonance study. J. Exp. Med. 178:643–648. http://dx.doi.org/10.1084/jem.178.2.643

Liu, F., H.Y. Wu, R. Wesselschmidt, T. Kornaga, and D.C. Link. 1996. Impaired production and increased apoptosis of neutrophils in granu-locyte colony-stimulating factor receptor-deficient mice. Immunity. 5:491–501. http://dx.doi.org/10.1016/S1074-7613(00)80504-X

Liu, Y., J. Mei, L. Gonzales, G. Yang, N. Dai, P. Wang, P. Zhang, M. Favara, K.C. Malcolm, S. Guttentag, and G.S. Worthen. 2011. IL-17A and TNF- exert synergistic effects on expression of CXCL5 by al-veolar type II cells in vivo and in vitro. J. Immunol. 186:3197–3205. http://dx.doi.org/10.4049/jimmunol.1002016

Lock, C., G. Hermans, R. Pedotti, A. Brendolan, E. Schadt, H. Garren, A. Langer-Gould, S. Strober, B. Cannella, J. Allard, et al. 2002.Gene-microarray analysis of multiple sclerosis lesions yields new tar-gets validated in autoimmune encephalomyelitis. Nat. Med. 8:500–508.http://dx.doi.org/10.1038/nm0502-500

Mantovani, A., M.A. Cassatella, C. Costantini, and S. Jaillon. 2011. Neutrophils in the activation and regulation of innate and adaptive immunity. Nat. Rev. Immunol. 11:519–531. http://dx.doi.org/10.1038/nri3024

McColl, S.R., M.A. Staykova, A. Wozniak, S. Fordham, J. Bruce, and D.O. Willenborg. 1998. Treatment with anti-granulocyte antibodies inhib-its the effector phase of experimental autoimmune encephalomyelitis. J. Immunol. 161:6421–6426.

Mehrad, B., R.M. Strieter, T.A. Moore, W.C. Tsai, S.A. Lira, and T.J. Standiford. 1999. CXC chemokine receptor-2 ligands are necessary components of neutrophil-mediated host defense in invasive pulmonary aspergillosis. J. Immunol. 163:6086–6094.

Mei, J., Y. Liu, N. Dai, C. Hoffmann, K.M. Hudock, P. Zhang, S.H. Guttentag, J.K. Kolls, P.M. Oliver, F.D. Bushman, and G.S. Worthen. 2012. Cxcr2 and Cxcl5 regulate the IL-17/G-CSF axis and neutrophil homeostasis in mice. J. Clin. Invest. 122:974–986. http://dx.doi.org/10.1172/JCI60588

Moxon-Emre, I., and L.C. Schlichter. 2011. Neutrophil depletion reduces blood-brain barrier breakdown, axon injury, and inflammation after intracerebral hemorrhage. J. Neuropathol. Exp. Neurol. 70:218–235. http://dx.doi.org/10.1097/NEN.0b013e31820d94a5

Naegele, M., K. Tillack, S. Reinhardt, S. Schippling, R. Martin, and M. Sospedra. 2012. Neutrophils in multiple sclerosis are characterized by a primed phenotype. J. Neuroimmunol. 242:60–71. http://dx.doi.org/10 .1016/j.jneuroim.2011.11.009

Onishi, R.M., and S.L. Gaffen. 2010. Interleukin-17 and its target genes: mechanisms of interleukin-17 function in disease. Immunology. 129:311–321. http://dx.doi.org/10.1111/j.1365-2567.2009.03240.x

Openshaw, H., O. Stuve, J.P. Antel, R. Nash, B.T. Lund, L.P. Weiner, A. Kashyap, P. McSweeney, and S. Forman. 2000. Multiple sclero-sis flares associated with recombinant granulocyte colony-stimulatingfactor. Neurology. 54:2147–2150. http://dx.doi.org/10.1212/WNL.54.11.2147

Pelletier, M., L. Maggi, A. Micheletti, E. Lazzeri, N. Tamassia, C. Costantini, L. Cosmi, C. Lunardi, F. Annunziato, S. Romagnani, and M.A. Cassatella.2010. Evidence for a cross-talk between human neutrophils and Th17cells. Blood. 115:335–343. http://dx.doi.org/10.1182/blood-2009-04-216085

Rapp, N.S., J. Gilroy, and A.M. Lerner. 1995. Role of bacterial infection in exacerbation of multiple sclerosis. Am. J. Phys. Med. Rehabil. 74:415–418. http://dx.doi.org/10.1097/00002060-199511000-00004

Richard, J.F., M. Roy, J. Audoy-Rémus, P. Tremblay, and L. Vallières. 2011. Crawling phagocytes recruited in the brain vasculature after

JEM Vol. 212, No. 1

Article

35

Frei, and A. Jörres. 2000. IL-17 stimulates intraperitoneal neutrophil infiltration through the release of GRO chemokine from mesothelial cells. J. Immunol. 165:5814–5821. http://dx.doi.org/10.4049/jimmunol .165.10.5814

Zavala, F., S. Abad, S. Ezine, V. Taupin, A. Masson, and J.F. Bach. 2002. G-CSF therapy of ongoing experimental allergic encephalomyelitisvia chemokine- and cytokine-based immune deviation. J. Immunol.168:2011–2019. http://dx.doi.org/10.4049/jimmunol.168.4.2011

Zhou, J., S.A. Stohlman, D.R. Hinton, and N.W. Marten. 2003. Neutrophils promote mononuclear cell infiltration during viral-induced encephalitis. J. Immunol. 170:3331–3336. http://dx.doi.org/10.4049/ jimmunol.170.6.3331

Ziaber, J., J. Paśnik, Z. Baj, L. Pokoca, H. Chmielewski, and H. Tchórzewski.1998. The immunoregulatory abilities of polymorphonuclear neutro-phils in the course of multiple sclerosis. Mediators Inflamm. 7:335–338. http://dx.doi.org/10.1080/09629359890857

The Rockefeller University Press $30.00J. Exp. Med. 2015 Vol. 212 No. 7 979–990www.jem.org/cgi/doi/10.1084/jem.20150956

979

Review

Macroautophagy is one of the major routes for the degradation of intracytoplasmic contents, in-cluding proteins and organelles such as mitochon-dria. The earliest morphologically recognizable intermediates in this pathway are phagophores, which evolve into double-membraned, sac-shaped structures. After the edges of the phago-phores extend and fuse, engulfing a portion of cytoplasm, they become known as autophago-somes. These are then trafficked along micro-tubules in a direction that is biased toward the perinuclear microtubule-organizing center, where the lysosomes are clustered. This brings the autophagosomes close to lysosomes, enabling fusion of these different organelles, after which the lysosomal hydrolases degrade the autopha-gic contents (Fig. 1).

There are two additional forms of autoph-agy that will not be considered in detail in this review. Microautophagy involves the direct se-questration of portions of the cytoplasm by ly-sosomes, and has been mainly studied in yeast. Chaperone-mediated autophagy captures pro-teins that contain a pentapeptide motif related to KFERQ via Hsc70, which targets proteins to LAMP2A. LAMP2A then serves as a transloca-tion channel to enable import of such substrates into the lysosomes. This pathway is perturbed by proteins causing certain neurodegenerative

diseases and has been reviewed in detail else-where (Cuervo and Wong, 2014).

Much of the pioneering work in the mac-roautophagy (henceforth referred to as autoph-agy in this review) field was initiated in yeast, where autophagy protects against cellular star-vation. Although this role is conserved across evolution, more recent studies in mammalian systems have highlighted the importance of autophagy in diverse areas of physiology and disease. In this review, we will focus on the protective roles of autophagy in neurodegener-ative and infectious diseases (Fig. 2). We will start by outlining the basic models where autophago-somes engulf and degrade neurodegeneration-associated aggregate-prone proteins or infectious agents. We will then describe possible mecha-nisms for enhancing the capture of such sub-strates to extents greater than would occur with bulk autophagy, during which one assumes there is random sequestration of cytoplasmic contents. We will extend the discussion of the roles of autophagy in these diseases by considering more complex consequences, including con-trol of cell death, immunity, and inflammation. Although there are aspects that have been

CORRESPONDENCE David C. Rubinsztein: dcr1000@cam.ac.uk OR Vojo Deretic: vderetic@salud.unm.edu

Abbreviations used: AMPK, AMP-activated protein kinase; ATG, autophagy-related; DAMP, damage-associated molecular pattern; HD, Hun-tington’s disease; mTORC1, mammalian target of rapamycin complex 1; PAMP, pathogen-associated molecular pattern; SCA, spinocerebellar ataxia; TBK1, TANK-binding kinase 1; UBA, ubiquitin-associated.

Therapeutic targeting of autophagy in neurodegenerative and infectious diseases

David C. Rubinsztein,1 Carla F. Bento,1 and Vojo Deretic2,3

1Department of Medical Genetics, Cambridge Institute for Medical Research, University of Cambridge School of Clinical Medicine, Cambridge CB2 OSP, England, UK2Department of Molecular Genetics and Microbiology and 3Department of Neurology, University of New Mexico Health Sciences Center, Albuquerque, NM 87131

Autophagy is a conserved process that uses double-membrane vesicles to deliver cytoplasmic contents to lysosomes for degradation. Although autophagy may impact many facets of human biology and disease, in this review we focus on the ability of autophagy to protect against certain neurodegenerative and infectious diseases. Autophagy enhances the clear-ance of toxic, cytoplasmic, aggregate-prone proteins and infectious agents. The beneficial roles of autophagy can now be extended to supporting cell survival and regulating inflam-mation. Autophagic control of inflammation is one area where autophagy may have similar benefits for both infectious and neurodegenerative diseases beyond direct removal of the pathogenic agents. Preclinical data supporting the potential therapeutic utility of autoph-agy modulation in such conditions is accumulating.

© 2015 Rubinsztein et al. This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (Attribution– Noncommercial–Share Alike 3.0 Unported license, as described at http://creative-commons.org/licenses/by-nc-sa/3.0/).

980 Autophagy in neurodegeneration, inflammation, and infection | Rubinsztein et al.

calcium flow from the ER to mitochondria, and the lower intramitochondrial calcium levels inhibit oxidative phosphory-lation, thereby decreasing ATP levels, which activates AMPK (Cárdenas et al., 2010). Some signals activate autophagy by stimulating III phosphatidylinositol 3-kinase (called VPS34), which produces phosphatidylinositol 3-phosphate (PI3P); this, in turn, helps to recruit ATG16L1 to sites of autophago-some formation (Dooley et al., 2014). Some of these signals act via the ATG6 orthologue Beclin 1, which stimulates VPS34 activity (Furuya et al., 2005; Russell et al., 2013). However, PI3P-independent forms of autophagy have also been described, and some of these appear to be mediated via the use of PI5P as an alternative to PI3P (Vicinanza et al., 2015). Interestingly, many of the stimuli that induce autoph-agy are stress responses. For example, mTORC1 activity is in-hibited by amino acid starvation (Chen et al., 2014), the levels of PI5P are induced by glucose starvation (Vicinanza et al., 2015), and AMPK (a key sensor of ATP levels in the cells) is enhanced when ATP energy stores are reduced (Hardie et al., 2012). These pathways are also directly linked to antiinfective or general immune signaling players, such as IRGM (an antituberculosis and Crohn’s disease factor that interacts with ULK1 and Beclin 1, promoting their coassem-bly; Chauhan et al., 2015), TAK1 and NOD2/RIPK2 (which activate AMPK and ULK1, respectively), and NLRP (which interacts with Beclin 1). The pathways also receive input from TLRs, IL-1, and other immune system regulators (Deretic et al., 2013).

In the context of neurodegenerative diseases such as Huntington’s disease (HD), there appears to be a decrease in mTORC1 activity in neurons with large aggregates (Ravikumar et al., 2004). However, the ultimate consequences for au-tophagy may not be straightforward, as excitotoxicity will increase calcium levels, which in turn inhibits autophagosome biogenesis (Williams et al., 2008), whereas mutant huntingtin binds the autophagy inducer Rhes to impair autophagy (Mealer et al., 2014). Thus, the eventual consequences of a specific mutation or disease situation are frequently unpredictable, as multiple activating and inhibitory pathways may be affected. Furthermore, non–cell-autonomous effects may have an im-pact. For example, the increased nitric oxide released by glial cells in diseases such as Alzheimer’s disease impairs autophago-some biogenesis (Sarkar et al., 2011).

How autophagy clears aggregate-prone intracytoplasmic proteinsIntracellular protein misfolding and aggregation are features of many late-onset neurodegenerative diseases, which are re-ferred to as proteinopathies. These include Alzheimer’s disease, Parkinson’s disease, tauopathies, and polyglutamine expan-sion diseases (including HD and various spinocerebellar ataxias [SCAs]). Currently, there are no effective therapeutic strate-gies that slow or prevent the neurodegeneration resulting from these diseases in humans. The mutations that cause HD and many other proteinopathies (e.g., polyglutamine diseases and tauopathies) confer novel toxic functions on the specific protein,

explored more in neurodegenerative diseases than infectious diseases, and vice versa, we believe that the opportunity to consider both in parallel will enable consideration of new hypotheses and cross-fertilization. We propose that the two main areas of overlap between the roles of autophagy in neu-rodegeneration and infectious disease are: (a) similarities in the shared usage of autophagic receptors in defending against pathology-inducing agents in both classes of disease (Birgisdottir et al., 2013), and (b) the now well-documented antiinflam-matory action of autophagy (Deretic et al., 2013, 2015). This juxtaposition of autophagic roles in apparently distinct classes of diseases is a testament to the relevance of autophagy in cleansing the cellular interiors no matter what the disease con-text is, and is particularly timely in view of the explosion of data in the two fields. Finally, we will consider possible autophagy-related therapeutic strategies that may be of significance for these diseases, including the possibility of developing agents that may target both sets of conditions.

Autophagy biologyThe membranes that contribute to phagophore formation and elongation may derive from multiple sources, including the ER (including ER exit sites and ER–mitochondrial con-tact sites; Hayashi-Nishino et al., 2009; Hamasaki et al., 2013), the ER–Golgi intermediate compartment (Ge et al., 2013, 2014), recycling endosomes (Longatti et al., 2012; Puri et al., 2013), plasma membrane (Ravikumar et al., 2010; Moreau et al., 2011), the Golgi complex (Young et al., 2006; Ohashi and Munro, 2010), and, potentially, lipid droplets (Dupont et al., 2014; Shpilka et al., 2015). The coordination of the membrane rearrangements that enable autophagosome for-mation, and their subsequent delivery to the lysosomes, is regulated by multiple autophagy-related (ATG) proteins. Some of these participate in two ubiquitin-like conjugation reac-tions. The first involves ATG12 conjugation to ATG5. This ATG12–ATG5 conjugate binds noncovalently with ATG16L1 to form a complex essential for phagophore expansion (Rubinsztein et al., 2012a). These complexes are localized to the phagophore and dissociate after the autophagosome is formed. The completion of autophagosome formation is as-sisted by a second conjugation reaction involving ATG8/LC3. LC3 is first cleaved by ATG4 to form cytosolic LC3-I, which is conjugated to phosphatidylethanolamine on autophago-some precursors to form membrane-associated LC3-II.

Autophagy signalingA primordial signaling pathway regulating autophagy, which is conserved from yeast to humans, is mediated by the mam-malian target of rapamycin complex 1 (mTORC1), which inhibits autophagy by phosphorylating proteins such as ATG1 and ATG13 that act upstream in phagophore formation (Hosokawa et al., 2009; Jung et al., 2009). However, several mTORC1-independent pathways have been described, in-cluding low inositol triphosphate levels (Sarkar et al., 2005), which activate autophagy by activating AMP-activated pro-tein kinase (AMPK). Low inositol triphosphate levels reduce

JEM Vol. 212, No. 7 981

Review

inflammation (via recognition of damage-associated molec-ular patterns [DAMPs]; Deretic et al., 2013). In addition, autophagy may also protect by enhancing the removal of rele-vant toxins, such as Staphylococcus aureus -toxin (Maurer et al., 2015).

The direct elimination of microbes by autophagy (a pro-cess termed xenophagy) receives the most attention, although it is likely that the antiinflammatory role of autophagy inde-pendent of, or during, infection plays an equally important host protective role (Deretic et al., 2015). The former percep-tion is understandable, as intracellular microbes such as invad-ing bacteria or viruses are large intracytoplasmic objects that represent potential (and in many cases actual) substrates for autophagic removal. Prototypical examples of this are Myco-bacterium tuberculosis in infected macrophages (Gutierrez et al., 2004) and animal models (Castillo et al., 2012; Watson et al., 2012; Manzanillo et al., 2013) and the Group A Streptococcus that manages to invade host cells (Nakagawa et al., 2004), but many other bacteria (including Listeria, Salmonella, and Shigella) are at least partially susceptible to autophagic elimination when tested in cellular systems (Gomes and Dikic, 2014; Huang and Brumell, 2014). Similarly, viruses, including HIV (Kyei et al.,

and disease severity frequently correlates with expression lev-els. Thus, it is important to understand the factors regulating the expression levels of these aggregate-prone proteins. When these proteins are intracytoplasmic, they can be removed ei-ther via the ubiquitin-proteasome system or via autophagy. Whereas the former route is generally more rapid, it is re-stricted to species that can enter the narrow proteasome bar-rel, which precludes oligomers and higher order structures. These species can be cleared by autophagy. Consistent with the model above, the aggregate-prone forms of such proteins, including tau (Berger et al., 2006), -synuclein (Webb et al., 2003; Spencer et al., 2009), mutant huntingtin (Ravikumar et al., 2002), and mutant ataxin 3 (Berger et al., 2006) appear to have a higher dependency on autophagy for their clear-ance compared with the wild-type forms.

Autophagy in infectious and inflammatory diseasesIn the context of infectious and inflammatory diseases, autoph-agy plays at least three roles. Autophagy can clear intracellular microbes and moderate host innate immune responses to mi-crobial products (through recognition of pathogen-associated molecular patterns [PAMPs]) and endogenous sources of

Figure 1. Schematic of autophagy. Activation of AMPK and/or inhibition of mTORC1 by various stress signals induces activation of the ATG1–ULK1 complex, which positively regulates the activity of the VPS34 complex via phosphorylation-dependent mechanisms. Class III PI3K VPS34 provides PI3P to the phagophore, which seems to define the LC3-lipidation sites by assisting in the recruitment of the ATG12–ATG5–ATG16L1 complex to the membrane (asterisks). After the binding of ATG12–ATG5–ATG16L1 complex to the phagophore and LC3 conjugation to PE (LC3-II), the membrane elongates and en-gulfs portions of the cytoplasm, ultimately leading to the formation of the complete autophagosome. Proteins such as p62, NDP52, and NBR1 confer substrate selectivity to the pathway by establishing a bridge between LC3-II and specific ubiquitinated cargo (e.g., aggregates, microbes, mitochondria, and peroxisomes), through their LIR and UBA domains, respectively. In the final step of the process, autophagosomes fuse with lysosomes, resulting in the degradation of the vesicle contents. AMPK, AMP-activated protein kinase; mTORC1, mechanistic target of rapamycin complex 1; ULK, Unc-51-like kinase; VPS34, phosphatidylinositol 3-kinase VPS34; PI3P, phosphatidylinositol 3-phosphate; PE, phosphatidylethanolamine; LIR, LC3-interacting region; UBA, ubiquitin associated domain.

982 Autophagy in neurodegeneration, inflammation, and infection | Rubinsztein et al.

However, in several cases, evidence of microbial exploitation of autophagy (not just defense against it, but in some cases enhancing survival or promoting spread) suggests that this approach must be carefully tailored. Some examples of the latter include Brucella (Starr et al., 2008), Anaplasma (formerly Ehrlichia; Niu et al., 2012), and poliovirus (Bird et al., 2014).

Autophagy receptor proteinsWhereas autophagy was originally considered to be a nonse-lective bulk degradation process, accumulating data now sup-ports the concept of selective macroautophagy, where the cell uses receptor proteins to enhance the incorporation of spe-cific cargoes into autophagosomes. These receptor proteins include p62 (Bjørkøy et al., 2005; Pankiv et al., 2007), opti-neurin (Wild et al., 2011), NDP52 (Thurston et al., 2009), NBR1 (Kirkin et al., 2009), ALFY (Filimonenko et al., 2010), TRIM5 (Mandell et al., 2014), and Tollip (Lu et al., 2014). The canonical model for this process involves these receptors binding to cargoes, typically via interaction with ubiquitinated motifs, and the receptor binding to the autophagosome mem-brane protein LC3 via LC3-interacting domains (Birgisdottir et al., 2013; Stolz et al., 2014). However, some classical recep-tors, like p62 and NBR1, may not require LC3-binding to be

2009; Shoji-Kawata et al., 2013; Mandell et al., 2014; Campbell et al., 2015; Sagnier et al., 2015), as well as protozoans (Choi et al., 2014), can be targeted by conventional or modified forms of autophagy. In many cases, an evolutionary balance exists whereby the host’s ability to deploy autophagy against the microbe is countered by bacterial or viral adaptations, and in most instances a successful intracellular pathogen has very specific antiautophagy strategies (Huang and Brumell, 2014). Such adaptations are seen in a wide range of pathogens, in-cluding Shigella and Legionella (Huang and Brumell, 2014), Mycobacterium tuberculosis (Deretic et al., 2015), HSV-1 (Orvedahl et al., 2007; Lussignol et al., 2013), and HIV (Kyei et al., 2009; Borel et al., 2014). Interestingly, interactions between autophagy and viral products can lead to neurological manifes-tations; for example, HIV proteins have been associated with HIV-induced dementia and manifestations of neuroAIDS (Meulendyke et al., 2014; El-Hage et al., 2015; Fields et al., 2015). As with other host–pathogen interactions, a balance between a microbe and the host is established, leading to chronic disease or subclinical or latent infection, as in latent tuberculosis or persistent viral infections. This represents a therapeutic opportunity to tip the balance against the pathogen by enhancing autophagy using pharmacological intervention.

Figure 2. Protective roles of autophagy in neurodegenerative and infectious diseases. A major role for autophagy in neurodegenerative and in-fectious diseases involves the clearance of toxic aggregate-prone proteins and infectious agents, respectively. However, it also exerts ancillary beneficial roles by controlling cell death and exacerbated inflammatory responses associated with these pathologies.

JEM Vol. 212, No. 7 983

Review

(Ravikumar et al., 2006; Hou et al., 2010; Amir et al., 2013; Meunier et al., 2014). In a manner similar to what has been observed in yeast, autophagy inhibition sensitizes mammalian cells to nutrient deprivation, whereas autophagy compromise results in apoptosis (Boya et al., 2005). Consistent with this, autophagy activation protects against proapoptotic insults in culture and in vivo. This may be relevant in neurodegenera-tive diseases, where subapoptotic caspase activities may en-hance disease by processes including cleavage of proteins like mutant huntingtin (Wellington et al., 2002; Warby et al., 2008) or tau (Rohn et al., 2002) to increase their toxicities, or by trimming of dendritic spines (Pozueta et al., 2013; Ertürk et al., 2014).

Autophagy also regulates inflammation. As recently re-viewed (Deretic et al., 2015), the antiinflammatory functions of autophagy in principle involve: (a) prevention of spurious inflammasome activation and down-regulation of the response once inflammasome is activated (Saitoh et al., 2008; Nakahira et al., 2011; Zhou et al., 2011; Lupfer et al., 2013) and (b) in-hibition of type I IFN responses directly (Jounai et al., 2007; Saitoh et al., 2009; Konno et al., 2013; Liang et al., 2014) or indirectly (Tal et al., 2009; Liang et al., 2014). The underly-ing processes include autophagic elimination of endogenous DAMPs (e.g., depolarized mitochondria leaking ROS, mito-chondrial DNA, and oxidized mitochondrial DNA; Saitoh et al., 2008; Nakahira et al., 2011; Zhou et al., 2011; Lupfer et al., 2013), which lowers the threshold for inflammasome activation, or direct targeting and degradation of inflamma-some components and products such as NLRP3, ASC, and IL-1 (Harris et al., 2011; Shi et al., 2012; Chuang et al., 2013); this, in turn, tapers the intensity and duration of inflammasome activation. However, the engagement of autophagy with cel-lular outputs of IL-1, a prototypical unconventionally secreted protein, is more complex (Dupont et al., 2011; Ponpuak et al., 2015). Autophagy assists secretion of IL-1 (Dupont et al., 2011; Öhman et al., 2014; Wang et al., 2014), a cytosolic protein that lacks a signal peptide and is unable to enter the conventional secretory pathway via the ER and Golgi. Thus, autophagy also plays a positive role in delivering IL-1 and possibly other proinflammatory substrates, once they are prop-erly activated in the cytosol, to the extracellular space where they perform their signaling functions (Ponpuak et al., 2015).

The autophagic interference with type I IFN responses occurs either directly by targeting signaling molecules within the pathway, starting with RIG-I-like receptors or cGAMP synthase (sensors recognizing cytosolic nucleic acids) and con-verging upon stimulator of the interferon gene (STING) and interferon regulatory factors (Jounai et al., 2007; Saitoh et al., 2009; Konno et al., 2013; Liang et al., 2014), or indirectly by removing agonist sources that activate these pathways (Tal et al., 2009; Liang et al., 2014). The p62 receptor also appears to have a role in restraining TCR activation of NF-B sig-naling mediated by Bcl10. Although p62 enables the signal-ing to occur in the first place, it also serves as a receptor to degrade Bcl10, which becomes ubiquitinated as a response to TCR activation. Thus, this mechanism may serve to protect

incorporated into autophagosomes (Itakura and Mizushima, 2011). Although systematic studies have not yet been per-formed, many of these receptors, including p62 and optineurin, appear to be able to assist autophagic capture of both neuro-degenerative disease-causing proteins and infectious agents. In their antimicrobial role, these receptors are referred to as a new class of pattern recognition receptors termed sequesto-some 1/p62-like receptors (Birgisdottir et al., 2013; Deretic et al., 2013, 2015). The ability of receptor proteins to recruit substrates to autophagosomes can also be modulated by post-translational modifications. For example, the TANK-binding kinase 1 (TBK1) phosphorylates optineurin on Ser177, enhanc-ing LC3-binding affinity and autophagic clearance of substrates, such as expanded polyglutamines as seen with mutant hun-tingtin (Korac et al., 2013), and Salmonella (Wild et al., 2011). Likewise, TBK1- (Pilli et al., 2012) or casein kinase-mediated (Matsumoto et al., 2011) phosphorylation of p62 at residue S403 has additional benefits in enhancing recognition of ubiquitinated targets by the ubiquitin-associated (UBA) do-main of p62, as is observed in clearance of polyglutamine expansion targets (Matsumoto et al., 2011) or mycobacteria (Pilli et al., 2012). Enhancement of ubiquitin recognition by the p62 UBA is also under control of direct phosphorylation by ULK1, which phosphorylates Ser405 and Ser409 of mu-rine p62 (equivalent to human Ser403 and Ser407; Lim et al., 2015). ULK1-mediated phosphorylation of the former resi-due additionally destabilizes the UBA dimer interface, thus increasing binding affinity of p62 to ubiquitin in response to proteotoxic stress (Lim et al., 2015). In the case of p62, and possibly other molecules, the activity of receptors can them-selves be influenced by a disease protein. Huntingtin, the Huntington disease-causing protein, appears to act as a scaf-fold for selective macroautophagy but it is dispensable for bulk autophagy (Ochaba et al., 2014; Rui et al., 2015). Hun-tingtin interacts with p62 to enhance its interactions with LC3 and with ubiquitin-modified substrates (Rui et al., 2015). Interestingly, in some cases, such as with optineurin (Tumbarello et al., 2012) and TRIM5 (Mandell et al., 2014), the adaptor proteins themselves also can act as bulk autophagy regulators. It is interesting to note that several of these pro-teins, including p62, TBK1, and optineurin, are mutated in neurodegenerative diseases such as motor neuron disease and forms of frontotemporal dementia (Maruyama et al., 2010; Fecto et al., 2011; Freischmidt et al., 2015; Pottier et al., 2015). Of further note is the shared role of autophagy receptors in protection against neurodegeneration and infectious agents, a principle that may extend to new receptor categories (e.g., TRIMs or other classes), as their functions are further eluci-dated with future progress in selective autophagy.

Additional protective properties of autophagy in neurodegenerative and infectious diseasesA major consequence of autophagy in many of these diseases is promotion of the removal of toxic proteins or infectious agents, but there may be additional benefits. Autophagy is gen-erally an antiapoptotic process that reduces caspase activation

984 Autophagy in neurodegeneration, inflammation, and infection | Rubinsztein et al.

trehalose (Frake et al., 2015; this work also considers the points of action of many of these drugs). Conversely, autoph-agy inhibition enhances the toxicity of these proteins and, in parallel, leads to the accumulation of the relevant protein (Frake et al., 2015).

Similarly, autophagy up-regulation may enhance the clear-ance of a range of infectious agents, with some of the more developed aspects being shown with M. tuberculosis, including multidrug-resistant (MDR) strains. In some cases, the support for this type of strategy has been strengthened by mouse models and preclinical data. For example, drugs used for psychiatric and neurological disorders such as the antidepressants fluox-etine (Stanley et al., 2014) and nortriptyline (Sundaramurthy et al., 2013), and the antiepileptic carbamazepine (Rubinsztein et al., 2012b; Schiebler et al., 2015), have been shown to counter M. tuberculosis infection, possibly through autophagy. Notably, carbamazepine, an inducer of autophagy, has been shown to act on MDR M. tuberculosis in vivo (Schiebler et al., 2015). Furthermore, several tyrosine kinase inhibitors, which also act as inducers of autophagy, have been tested in vitro and in mouse models for their potential in host-directed ther-apy (HDT) in tuberculosis. This includes gefitinib, an inhibi-tor of the tyrosine kinase epidermal growth factor receptor (EGFR) shown to activate autophagy and suppress M. tuber-culosis in macrophages and, to some extent, in infected mice (Stanley et al., 2014). It also includes imatinib (Gleevec), a known inducer of autophagy (Ertmer et al., 2007) and in-hibitor of the tyrosine kinase Abl, whose depletion has been shown to suppress intracellular M. tuberculosis (Jayaswal et al., 2010), with imatinib reducing M. tuberculosis bacillary loads in infected macrophages (Bruns et al., 2012) and in a mouse model of tuberculosis (Napier et al., 2011). Other antituber-culosis HDT autophagy-inducing candidate drugs include antiparasitic pharmaceuticals such as nitozoxanide (Lam et al., 2012) and cholesterol-lowering drugs, i.e., statins (Parihar et al., 2014).

There may be a wide range of strategies that could be used in human conditions, including drugs (where several FDA-approved drugs show promise in preclinical models), peptides (Shoji-Kawata et al., 2013), and possibly topical agents for certain infectious agents. Furthermore, there may be oppor-tunities for modulating selective autophagy via adaptor pro-teins. Strategies could include regulating posttranslational modifications of proteins that could enhance their activities.

Neurodegenerative disease-causing proteins and various infectious agents can also impair autophagy. Although this issue has been dealt with in detail elsewhere (Menzies et al., 2015), one recent example includes the VPS35 D620N Parkinson’s disease mutation that impacts early stages of autophagosome biogenesis (Zavodszky et al., 2014). PICALM, an Alzheimer’s disease GWAS hit, impacts both autophagosome formation and autophagosome degradation, and altered PICALM activity in culture and in vivo leads to the accumulation and increased toxicity of tau, a protein which is an important driver of Alzheimer’s disease pathogenesis (Moreau et al., 2014). Likewise, infectious agents like Salmonella (Mesquita

cells from NF-B hyperactivation in response to TCR sig-naling (Paul et al., 2012). The antiinflammatory action of au-tophagy applies to both infectious and inflammatory diseases (either sterile or associated with microbial triggers), such as Crohn’s disease. These relationships may extend to neuroin-flammation in acute and chronic neurological disorders. Many neurodegenerative diseases are associated with inflam-matory responses in glia, which may contribute to pathology (Czirr and Wyss-Coray, 2012), and it is possible that autophagy in glial cells may play a role in keeping these processes in check, although this domain has not been carefully explored.

Autophagy also plays key roles in protecting cells against infectious agents that either remain within vacuoles or escape from phagosomes into the cytoplasm (Huang and Brumell, 2014). Examples of intracellular bacterial pathogens in most cases represent a mixed spectrum of retention within the para-sitophorous vacuole, partial permeabilization of such vacu-oles, or full escape of bacteria into the cytosol. Such mixed events are often skewed to one or the other end of the spec-trum, with Shigella (Ogawa et al., 2005; Dupont et al., 2009; Mostowy et al., 2011; Ogawa et al., 2011; Thurston et al., 2012) and Listeria (Py et al., 2007; Mostowy et al., 2011) pre-dominantly escaping into the cytosol, whereas Salmonella (Zheng et al., 2009; Wild et al., 2011; Huett et al., 2012; Thurston et al., 2012; Gomes and Dikic, 2014) and M. tuber-culosis (Gutierrez et al., 2004; Watson et al., 2012; Manzanillo et al., 2013; Deretic et al., 2015) primarily reside in undam-aged vacuoles although recent studies indicate that it pene-trates into the cytosol. Parallels may exist in neurodegenerative diseases, where autophagy may help glial cell clearance of extracellular -amyloid if the internalized peptide is found to gain access to the cytosol (Li et al., 2013). This principle may be also relevant to diseases like Parkinson’s disease and forms of frontotemporal dementia, where there is increasing evi-dence for extracellular spread of the relevant toxic proteins like -synuclein and tau via prion-like mechanisms (Desplats et al., 2009; Frost et al., 2009; Lee et al., 2010; Steiner et al., 2011). However, impaired clearance of autophagosomes due to defective lysosomal function may cause excess secretion of such proteins and exacerbate extracellular spread (Ejlerskov et al., 2013; Lee et al., 2013).

Therapeutic and clinical implicationsUp-regulation of autophagy via mTORC1-dependent and -independent routes has been shown to enhance the clearanceof neurodegenerative disease-causing proteins and reducetheir toxicity in a wide range of cells in Drosophila, zebrafish,and mouse models (Ravikumar et al., 2004; Furuya et al., 2005;Sarkar et al., 2007; Zhang et al., 2007; Pickford et al., 2008;Menzies et al., 2010; Spilman et al., 2010; Cortes et al.,2012; Schaeffer et al., 2012; Hebron et al., 2013; Frake et al.,2015). This strategy has shown promise in a range of diseasemodels, including tauopathies, -synucleinopathies, HD, spi-nocerebellar ataxia type 3, and familial prion disease. Thedrugs used in these diseases include a rapamycin analogue andmTOR-independent autophagy inducers like rilmenidine and

JEM Vol. 212, No. 7 985

Review

may be partially mitigated if suitable iPS stem cell–derived neu-ronal models are generated for sporadic cases. These difficul-ties may be less of an issue for monogenic diseases that can be more faithfully recapitulated in mice. However, even in these cases, the disease course is often much more rapid in the mod-els, which may have consequences for the way one interprets the preclinical data.

Future work will establish the potential for harnessing autophagy as a therapeutic option in various neurodegenera-tive and infectious diseases.

We are grateful to the Wellcome Trust (095317/Z/11/Z Principal Research Fellowship to D.C. Rubinsztein and strategic award 100140), the National Institute for Health Research Biomedical Research Unit in Dementia at Addenbrooke’s Hospital (D.C. Rubinsztein), and the National Institutes of Health (AI042999 and AI111935; V. Deretic) for funding our work.

D.C. Rubinsztein has received grant funding from MedImmune and is a scientific advisor for E3Bio and Bioblast. The authors declare no additional competing financial interests.

Note added in proof. While this manuscript was in production, further evidence of the extensive overlaps between inflammatory response systems and autophagy was documented in the context of cyclic GMP-AMP synthase (cGAS)-dependent type I IFN production and autophagic clearance of M. tuberculosis. (Collins, A.C., H. Cai, T. Li, L.H. Franco, X.D. Li, V.R. Nair, C.R. Scharn, C.E. Stamm, B. Levine, Z.J. Chen, and M.U. Shiloh. 2015. Cell host Microbe. 17:820-828; Watson, R.O., S.L. Bell, D.A. MacDuff, J.M. Kimmey, E.J. Diner, J. Olivas, R.E. Vance, C.L. Stallings, H.W. Virgin, and J.S. Cox. 2015. Cell host Microbe. 17:811-819).

REFERENCESAmir, M., E. Zhao, L. Fontana, H. Rosenberg, K. Tanaka, G. Gao, and M.J.

Czaja. 2013. Inhibition of hepatocyte autophagy increases tumor necro-sis factor-dependent liver injury by promoting caspase-8 activation. Cell Death Differ. 20:878–887. http://dx.doi.org/10.1038/cdd.2013.21

Berger, Z., B. Ravikumar, F.M. Menzies, L.G. Oroz, B.R. Underwood, M.N. Pangalos, I. Schmitt, U. Wullner, B.O. Evert, C.J. O’Kane, andD.C. Rubinsztein. 2006. Rapamycin alleviates toxicity of different ag-gregate-prone proteins. Hum. Mol. Genet. 15:433–442. http://dx.doi.org/10.1093/hmg/ddi458

Bird, S.W., N.D. Maynard, M.W. Covert, and K. Kirkegaard. 2014. Nonlytic viral spread enhanced by autophagy components. Proc. Natl. Acad. Sci. USA. 111:13081–13086. http://dx.doi.org/10.1073/pnas.1401437111

Birgisdottir, Å.B., T. Lamark, and T. Johansen. 2013. The LIR motif - cru-cial for selective autophagy. J. Cell Sci. 126:3237–3247. http://dx.doi .org/10.1242/jcs.126128

Birmingham, C.L., V. Canadien, N.A. Kaniuk, B.E. Steinberg, D.E. Higgins, and J.H. Brumell. 2008. Listeriolysin O allows Listeria mono-cytogenes replication in macrophage vacuoles. Nature. 451:350–354. http://dx.doi.org/10.1038/nature06479

Bjørkøy, G., T. Lamark, A. Brech, H. Outzen, M. Perander, A. Overvatn, H. Stenmark, and T. Johansen. 2005. p62/SQSTM1 forms protein aggre-gates degraded by autophagy and has a protective effect on huntingtin- induced cell death. J. Cell Biol. 171:603–614. http://dx.doi.org/10.1083/ jcb.200507002

Borel, S., V. Robert-Hebmann, J. Alfaisal, A. Jain, M. Faure, L. Espert, L. Chaloin, J.C. Paillart, T. Johansen, and M. Biard-Piechaczyk. 2014.HIV-1 viral infectivity factor interacts with microtubule-associated pro-tein light chain 3 and inhibits autophagy. AIDS. 29:275–286. http://dx.doi.org/10.1097/QAD.0000000000000554

Boya, P., R.A. González-Polo, N. Casares, J.L. Perfettini, P. Dessen, N. Larochette, D. Métivier, D. Meley, S. Souquere, T. Yoshimori, et al. 2005. Inhibition of macroautophagy triggers apoptosis. Mol. Cell. Biol. 25:1025–1040. http://dx.doi.org/10.1128/MCB.25.3.1025-1040.2005

Bruns, H., F. Stegelmann, M. Fabri, K. Döhner, G. van Zandbergen, M. Wagner, M. Skinner, R.L. Modlin, and S. Stenger. 2012. Abelson tyrosine kinase controls phagosomal acidification required for killing

et al., 2012; Owen et al., 2014), Legionella (Choy et al., 2012), Shigella (Ogawa et al., 2005), Listeria (Birmingham et al., 2008; Yoshikawa et al., 2009), and viruses (Orvedahl et al., 2007; Kyei et al., 2009; Lussignol et al., 2013; Borel et al., 2014) have multiple mechanisms that can at least partially counter or fully impair autophagy. In extreme cases, some infectious agents can convert autophagosomes into a replicative (Niu et al., 2012) or persistence (Birmingham et al., 2008) niche.

Understanding the biology of the relevant disease and the proposed treatment modality will enhance the probability of successful therapies. In diseases where there is impaired au-tophagosome degradation, including the lysosomal storage diseases, there may be concerns about the risks versus the benefits of increasing autophagosome biogenesis. However, this may depend on the extent of the block of autophago-some degradation, as stimulation of autophagosome biogene-sis appeared to enhance autophagic substrate clearance in cell culture models of Niemann-Pick Type C1 (Sarkar et al., 2013), a lysosomal storage disease associated with delayed autopha-gosome degradation.

Likewise, it is important to understand the actions and possible side effects of drugs used for these diseases. For ex-ample, azithromycin, a potent antibiotic, is used as a prophy-lactic against mycobacterial infections in cystic fibrosis patients. However, mycobacteria that develop resistance against azithro-mycin accumulate in culture and in vivo when treated with this agent, as azithromycin also impairs autophagosome deg-radation (Renna et al., 2011). Thus, the advantages of this drug as an antimicrobial for sensitive species may be, in part, counterbalanced by the risks of autophagy inhibition for re-sistant mycobacterial species. This possibility is suggested by preliminary clinical data which have reported increased risks of resistant nontuberculous mycobacterial infections in cystic fibrosis patients treated chronically with azithromycin.

Future directionsExtensive preclinical animal model data support the promise of the therapeutic use of autophagy up-regulation in various neurodegenerative and infectious diseases. This aim may be achievable with existing approved drugs using repurposing strategies. Here, a major challenge will be making the transi-tion between mice and humans, where one needs to contend with very different pharmacokinetics for drugs between the species. However, in these scenarios, the task is simplified by the existing human safety and pharmacokinetics data on the drugs. It is likely that most, if not all, of the approved drugs that influence autophagy have effects on other pathways, and although these may not be limiting or even disadvantageous, there would be major advantages both for experimental studies and possibly human treatments to identify more specific au-tophagy modulators. These may be more elusive than previously anticipated, given the increasing awareness of autophagy- independent roles of many ATG proteins.

A second major hurdle with such drug discovery efforts is disease modeling. It is currently impossible to model sporadic Alzheimer’s and Parkinson’s disease in rodents. These limitations

986 Autophagy in neurodegeneration, inflammation, and infection | Rubinsztein et al.

Dupont, N., S. Chauhan, J. Arko-Mensah, E.F. Castillo, A. Masedunskas, R. Weigert, H. Robenek, T. Proikas-Cezanne, and V. Deretic. 2014.Neutral lipid stores and lipase PNPLA5 contribute to autophago-some biogenesis. Curr. Biol. 24:609–620. http://dx.doi.org/10.1016/j.cub.2014.02.008

Ejlerskov, P., I. Rasmussen, T.T. Nielsen, A.L. Bergström, Y. Tohyama, P.H. Jensen, and F. Vilhardt. 2013. Tubulin polymerization-promoting protein (TPPP/p25) promotes unconventional secretion of -synuclein through exophagy by impairing autophagosome-lysosome fusion. J. Biol. Chem. 288:17313–17335. http://dx.doi.org/10.1074/jbc.M112.401174

El-Hage, N., M. Rodriguez, S.M. Dever, R.R. Masvekar, D.A. Gewirtz, and J.J. Shacka. 2015. HIV-1 and morphine regulation of autophagy in microglia: limited interactions in the context of HIV-1 infection and opioid abuse. J. Virol. 89:1024–1035. http://dx.doi.org/10.1128/ JVI.02022-14

Ertmer, A., V. Huber, S. Gilch, T. Yoshimori, V. Erfle, J. Duyster, H.P. Elsässer, and H.M. Schätzl. 2007. The anticancer drug imatinib induces cellular autophagy. Leukemia. 21:936–942.

Ertürk, A., Y. Wang, and M. Sheng. 2014. Local pruning of dendrites and spines by caspase-3-dependent and proteasome-limited mechanisms. J. Neurosci. 34:1672–1688. http://dx.doi.org/10.1523/JNEUROSCI.3121-13.2014

Fecto, F., J. Yan, S.P. Vemula, E. Liu, Y. Yang, W. Chen, J.G. Zheng, Y. Shi, N. Siddique, H. Arrat, et al. 2011. SQSTM1 mutations in familial and sporadic amyotrophic lateral sclerosis. Arch. Neurol. 68:1440–1446. http://dx.doi.org/10.1001/archneurol.2011.250

Fields, J., W. Dumaop, S. Elueteri, S. Campos, E. Serger, M. Trejo, K. Kosberg, A. Adame, B. Spencer, E. Rockenstein, et al. 2015. HIV-1 Tat alters neuronal autophagy by modulating autophagosome fu-sion to the lysosome: implications for HIV-associated neurocogni-tive disorders. J. Neurosci. 35:1921–1938. http://dx.doi.org/10.1523/ JNEUROSCI.3207-14.2015

Filimonenko, M., P. Isakson, K.D. Finley, M. Anderson, H. Jeong, T.J. Melia, B.J. Bartlett, K.M. Myers, H.C. Birkeland, T. Lamark, et al. 2010. The selective macroautophagic degradation of aggregated pro-teins requires the PI3P-binding protein Alfy. Mol. Cell. 38:265–279. http://dx.doi.org/10.1016/j.molcel.2010.04.007

Frake, R.A., T. Ricketts, F.M. Menzies, and D.C. Rubinsztein. 2015. Autophagy and neurodegeneration. J. Clin. Invest. 125:65–74. http://dx.doi.org/10.1172/JCI73944

Freischmidt, A., T. Wieland, B. Richter, W. Ruf, V. Schaeffer, K. Müller, N. Marroquin, F. Nordin, A. Hübers, P. Weydt, et al. 2015. Haploinsufficiency of TBK1 causes familial ALS and fronto-temporal dementia. Nat. Neurosci. 18:631–636. http://dx.doi.org/10.1038/ nn.4000

Frost, B., R.L. Jacks, and M.I. Diamond. 2009. Propagation of tau misfold-ing from the outside to the inside of a cell. J. Biol. Chem. 284:12845–12852. http://dx.doi.org/10.1074/jbc.M808759200

Furuya, N., J. Yu, M. Byfield, S. Pattingre, and B. Levine. 2005. The evo-lutionarily conserved domain of Beclin 1 is required for Vps34 binding, autophagy and tumor suppressor function. Autophagy. 1:46–52. http://dx.doi.org/10.4161/auto.1.1.1542

Ge, L., D. Melville, M. Zhang, and R. Schekman. 2013. The ER-Golgi intermediate compartment is a key membrane source for the LC3 lipi-dation step of autophagosome biogenesis. eLife. 2:e00947. http://dx.doi .org/10.7554/eLife.00947

Ge, L., M. Zhang, and R. Schekman. 2014. Phosphatidylinositol 3-kinase and COPII generate LC3 lipidation vesicles from the ER-Golgi in-termediate compartment. eLife. 3:e04135. http://dx.doi.org/10.7554/ eLife.04135

Gomes, L.C., and I. Dikic. 2014. Autophagy in antimicrobial immunity. Mol. Cell. 54:224–233. http://dx.doi.org/10.1016/j.molcel.2014.03.009

Gutierrez, M.G., S.S. Master, S.B. Singh, G.A. Taylor, M.I. Colombo, and V. Deretic. 2004. Autophagy is a defense mechanism inhibiting BCGand Mycobacterium tuberculosis survival in infected macrophages. Cell. 119:753–766. http://dx.doi.org/10.1016/j.cell.2004.11.038

Hamasaki, M., N. Furuta, A. Matsuda, A. Nezu, A. Yamamoto, N. Fujita, H. Oomori, T. Noda, T. Haraguchi, Y. Hiraoka, et al. 2013.

of Mycobacterium tuberculosis in human macrophages. J. Immunol. 189:4069–4078. http://dx.doi.org/10.4049/jimmunol.1201538

Campbell, G.R., R.S. Bruckman, Y.L. Chu, and S.A. Spector. 2015. Autophagy induction by histone deacetylase inhibitors inhibits HIV type 1. J. Biol. Chem. 290:5028–5040. http://dx.doi.org/10.1074/jbc .M114.605428

Cárdenas, C., R.A. Miller, I. Smith, T. Bui, J. Molgó, M. Müller, H. Vais, K.H. Cheung, J. Yang, I. Parker, et al. 2010. Essential regulation of cell bioenergetics by constitutive InsP3 receptor Ca2+ transfer to mitochon-dria. Cell. 142:270–283. http://dx.doi.org/10.1016/j.cell.2010.06.007

Castillo, E.F., A. Dekonenko, J. Arko-Mensah, M.A. Mandell, N. Dupont, S. Jiang, M. Delgado-Vargas, G.S. Timmins, D. Bhattacharya, H. Yang,et al. 2012. Autophagy protects against active tuberculosis by suppress-ing bacterial burden and inflammation. Proc. Natl. Acad. Sci. USA. 109:E3168–E3176. http://dx.doi.org/10.1073/pnas.1210500109

Chauhan, S., M.A. Mandell, and V. Deretic. 2015. IRGM Governs the Core Autophagy Machinery to Conduct Antimicrobial Defense. Mol. Cell. 58:507–521. http://dx.doi.org/10.1016/j.molcel.2015.03.020

Chen, R., Y. Zou, D. Mao, D. Sun, G. Gao, J. Shi, X. Liu, C. Zhu, M. Yang, W. Ye, et al. 2014. The general amino acid control pathway regulates mTOR and autophagy during serum/glutamine starvation. J. Cell Biol. 206:173–182. http://dx.doi.org/10.1083/jcb.201403009

Choi, J., S. Park, S.B. Biering, E. Selleck, C.Y. Liu, X. Zhang, N. Fujita, T. Saitoh, S. Akira, T. Yoshimori, et al. 2014. The parasitophorousvacuole membrane of Toxoplasma gondii is targeted for disruption by ubiquitin-like conjugation systems of autophagy. Immunity. 40:924–935. http://dx.doi.org/10.1016/j.immuni.2014.05.006

Choy, A., J. Dancourt, B. Mugo, T.J. O’Connor, R.R. Isberg, T.J. Melia, and C.R. Roy. 2012. The Legionella effector RavZ inhibits host au-tophagy through irreversible Atg8 deconjugation. Science. 338:1072–1076. http://dx.doi.org/10.1126/science.1227026

Chuang, S.Y., C.H. Yang, C.C. Chou, Y.P. Chiang, T.H. Chuang, and L.C. Hsu. 2013. TLR-induced PAI-2 expression suppresses IL-1 pro-cessing via increasing autophagy and NLRP3 degradation. Proc. Natl.Acad. Sci. USA. 110:16079–16084. http://dx.doi.org/10.1073/pnas.1306556110

Cortes, C.J., K. Qin, J. Cook, A. Solanki, and J.A. Mastrianni. 2012. Rapamycin delays disease onset and prevents PrP plaque deposition in a mouse model of Gerstmann-Sträussler-Scheinker disease. J. Neurosci. 32:12396–12405. http://dx.doi.org/10.1523/JNEUROSCI.6189-11.2012

Cuervo, A.M., and E. Wong. 2014. Chaperone-mediated autophagy: roles in disease and aging. Cell Res. 24:92–104. http://dx.doi.org/10.1038/ cr.2013.153

Czirr, E., and T. Wyss-Coray. 2012. The immunology of neurodegeneration. J. Clin. Invest. 122:1156–1163. http://dx.doi.org/10.1172/JCI58656

Deretic, V., T. Saitoh, and S. Akira. 2013. Autophagy in infection, inflam-mation and immunity. Nat. Rev. Immunol. 13:722–737. http://dx.doi .org/10.1038/nri3532

Deretic, V., T. Kimura, G. Timmins, P. Moseley, S. Chauhan, and M. Mandell. 2015. Immunologic manifestations of autophagy. J. Clin. Invest. 125:75–84. http://dx.doi.org/10.1172/JCI73945

Desplats, P., H.J. Lee, E.J. Bae, C. Patrick, E. Rockenstein, L. Crews, B. Spencer, E. Masliah, and S.J. Lee. 2009. Inclusion formation andneuronal cell death through neuron-to-neuron transmission of alpha-synuclein. Proc. Natl. Acad. Sci. USA. 106:13010–13015. http://dx.doi.org/10.1073/pnas.0903691106

Dooley, H.C., M. Razi, H.E. Polson, S.E. Girardin, M.I. Wilson, and S.A. Tooze. 2014. WIPI2 links LC3 conjugation with PI3P, autophagosome formation, and pathogen clearance by recruiting Atg12-5-16L1. Mol. Cell. 55:238–252. http://dx.doi.org/10.1016/molcel.2014.05.021

Dupont, N., S. Lacas-Gervais, J. Bertout, I. Paz, B. Freche, G.T. Van Nhieu, F.G. van der Goot, P.J. Sansonetti, and F. Lafont. 2009. Shigella phago-cytic vacuolar membrane remnants participate in the cellular response to pathogen invasion and are regulated by autophagy. Cell Host Microbe. 6:137–149. http://dx.doi.org/10.1016/j.chom.2009.07.005

Dupont, N., S. Jiang, M. Pilli, W. Ornatowski, D. Bhattacharya, and V. Deretic. 2011. Autophagy-based unconventional secretory pathway for extracellular delivery of IL-1. EMBO J. 30:4701–4711. http://dx.doi .org/10.1038/emboj.2011.398

JEM Vol. 212, No. 7 987

Review

pathway intersects with HIV-1 biosynthesis and regulates viral yields in macrophages. J. Cell Biol. 186:255–268. http://dx.doi.org/10.1083/ jcb.200903070

Lam, K.K., X. Zheng, R. Forestieri, A.D. Balgi, M. Nodwell, S. Vollett, H.J. Anderson, R.J. Andersen, Y. Av-Gay, and M. Roberge. 2012. Nitazoxanide stimulates autophagy and inhibits mTORC1 signaling and intracellular proliferation of Mycobacterium tuberculosis. PLoS Pathog. 8:e1002691. http://dx.doi.org/10.1371/journal.ppat.1002691

Lee, S.J., P. Desplats, C. Sigurdson, I. Tsigelny, and E. Masliah. 2010. Cell-to-cell transmission of non-prion protein aggregates. Nat Rev Neurol. 6:702–706. http://dx.doi.org/10.1038/nrneurol.2010.145

Lee, H.J., E.D. Cho, K.W. Lee, J.H. Kim, S.G. Cho, and S.J. Lee. 2013. Autophagic failure promotes the exocytosis and intercellular transfer of -synuclein. Exp. Mol. Med. 45:e22. http://dx.doi.org/10.1038/ emm.2013.45

Li, W., Y. Tang, Z. Fan, Y. Meng, G. Yang, J. Luo, and Z.J. Ke. 2013. Autophagy is involved in oligodendroglial precursor-mediated clear-ance of amyloid peptide. Mol. Neurodegener. 8:27. http://dx.doi.org/ 10.1186/1750-1326-8-27

Liang, Q., G.J. Seo, Y.J. Choi, M.J. Kwak, J. Ge, M.A. Rodgers, M. Shi, B.J. Leslie, K.P. Hopfner, T. Ha, et al. 2014. Crosstalk between the cGAS DNA sensor and Beclin-1 autophagy protein shapes innate an-timicrobial immune responses. Cell Host Microbe. 15:228–238. http://dx.doi.org/10.1016/j.chom.2014.01.009

Lim, J., M.L. Lachenmayer, S. Wu, W. Liu, M. Kundu, R. Wang, M. Komatsu, Y.J. Oh, Y. Zhao, and Z. Yue. 2015. Proteotoxic stress in-duces phosphorylation of p62/SQSTM1 by ULK1 to regulate selective autophagic clearance of protein aggregates. PLoS Genet. 11:e1004987. http://dx.doi.org/10.1371/journal.pgen.1004987

Longatti, A., C.A. Lamb, M. Razi, S. Yoshimura, F.A. Barr, and S.A. Tooze. 2012. TBC1D14 regulates autophagosome formation via Rab11- and ULK1-positive recycling endosomes. J. Cell Biol. 197:659–675. http://dx.doi.org/10.1083/jcb.201111079

Lu, K., I. Psakhye, and S. Jentsch. 2014. Autophagic clearance of polyQ proteins mediated by ubiquitin-Atg8 adaptors of the conserved CUET protein family. Cell. 158:549–563. http://dx.doi.org/10.1016/j.cell .2014.05.048

Lupfer, C., P.G. Thomas, P.K. Anand, P. Vogel, S. Milasta, J. Martinez, G. Huang, M. Green, M. Kundu, H. Chi, et al. 2013. Receptor interacting protein kinase 2-mediated mitophagy regulates inflammasome activa-tion during virus infection. Nat. Immunol. 14:480–488. http://dx.doi .org/10.1038/ni.2563

Lussignol, M., C. Queval, M.F. Bernet-Camard, J. Cotte-Laffitte, I. Beau, P. Codogno, and A. Esclatine. 2013. The herpes simplex virus 1 Us11 pro-tein inhibits autophagy through its interaction with the protein kinase PKR. J. Virol. 87:859–871. http://dx.doi.org/10.1128/JVI.01158-12

Mandell, M.A., A. Jain, J. Arko-Mensah, S. Chauhan, T. Kimura, C. Dinkins, G. Silvestri, J. Münch, F. Kirchhoff, A. Simonsen, et al. 2014. TRIM proteins regulate autophagy and can target autophagic substrates by direct recognition. Dev. Cell. 30:394–409. http://dx.doi .org/10.1016/j.devcel.2014.06.013

Manzanillo, P.S., J.S. Ayres, R.O. Watson, A.C. Collins, G. Souza, C.S. Rae, D.S. Schneider, K. Nakamura, M.U. Shiloh, and J.S. Cox. 2013. The ubiquitin ligase parkin mediates resistance to intracellular patho-gens. Nature. 501:512–516. http://dx.doi.org/10.1038/nature12566

Maruyama, H., H. Morino, H. Ito, Y. Izumi, H. Kato, Y. Watanabe, Y. Kinoshita, M. Kamada, H. Nodera, H. Suzuki, et al. 2010. Mutations of optineurin in amyotrophic lateral sclerosis. Nature. 465:223–226. http://dx.doi.org/10.1038/nature08971

Matsumoto, G., K. Wada, M. Okuno, M. Kurosawa, and N. Nukina. 2011. Serine 403 phosphorylation of p62/SQSTM1 regulates selective au-tophagic clearance of ubiquitinated proteins. Mol. Cell. 44:279–289. http://dx.doi.org/10.1016/j.molcel.2011.07.039

Maurer, K., T. Reyes-Robles, F. Alonzo III, J. Durbin, V.J. Torres, and K. Cadwell. 2015. Autophagy mediates tolerance to Staphylococcusaureus alpha-toxin. Cell Host Microbe. 17:429–440. http://dx.doi.org/10.1016/j.chom.2015.03.001

Mealer, R.G., A.J. Murray, N. Shahani, S. Subramaniam, and S.H. Snyder. 2014. Rhes, a striatal-selective protein implicated in Huntington disease,

Autophagosomes form at ER-mitochondria contact sites. Nature. 495:389–393. http://dx.doi.org/10.1038/nature11910

Hardie, D.G., F.A. Ross, and S.A. Hawley. 2012. AMPK: a nutrient and energy sensor that maintains energy homeostasis. Nat. Rev. Mol. Cell Biol. 13:251–262. http://dx.doi.org/10.1038/nrm3311

Harris, J., M. Hartman, C. Roche, S.G. Zeng, A. O’Shea, F.A. Sharp, E.M. Lambe, E.M. Creagh, D.T. Golenbock, J. Tschopp, et al. 2011.Autophagy controls IL-1beta secretion by targeting pro-IL-1beta fordegradation. J. Biol. Chem. 286:9587–9597. http://dx.doi.org/10.1074/jbc.M110.202911

Hayashi-Nishino, M., N. Fujita, T. Noda, A. Yamaguchi, T. Yoshimori, and A. Yamamoto. 2009. A subdomain of the endoplasmic reticulum forms a cradle for autophagosome formation. Nat. Cell Biol. 11:1433–1437. http://dx.doi.org/10.1038/ncb1991

Hebron, M.L., I. Lonskaya, and C.E. Moussa. 2013. Nilotinib reverses loss of dopamine neurons and improves motor behavior via autophagic deg-radation of -synuclein in Parkinson’s disease models. Hum. Mol. Genet. 22:3315–3328. http://dx.doi.org/10.1093/hmg/ddt192

Hosokawa, N., T. Hara, T. Kaizuka, C. Kishi, A. Takamura, Y. Miura, S. Iemura, T. Natsume, K. Takehana, N. Yamada, et al. 2009. Nutrient-dependent mTORC1 association with the ULK1-Atg13-FIP200 complex required for autophagy. Mol. Biol. Cell. 20:1981–1991. http://dx.doi.org/10.1091/mbc.E08-12-1248

Hou, W., J. Han, C. Lu, L.A. Goldstein, and H. Rabinowich. 2010. Autophagic degradation of active caspase-8: a crosstalk mechanism be-tween autophagy and apoptosis. Autophagy. 6:891–900. http://dx.doi .org/10.4161/auto.6.7.13038

Huang, J., and J.H. Brumell. 2014. Bacteria-autophagy interplay: a battle for survival. Nat. Rev. Microbiol. 12:101–114. http://dx.doi.org/10.1038/ nrmicro3160

Huett, A., R.J. Heath, J. Begun, S.O. Sassi, L.A. Baxt, J.M. Vyas, M.B. Goldberg, and R.J. Xavier. 2012. The LRR and RING domain protein LRSAM1 is an E3 ligase crucial for ubiquitin-dependent autophagy of intracellular Salmonella Typhimurium. Cell Host Microbe. 12:778–790. http://dx.doi.org/10.1016/j.chom.2012.10.019

Itakura, E., and N. Mizushima. 2011. p62 Targeting to the autophagosome formation site requires self-oligomerization but not LC3 binding. J. Cell Biol. 192:17–27. http://dx.doi.org/10.1083/jcb.201009067

Jayaswal, S., M.A. Kamal, R. Dua, S. Gupta, T. Majumdar, G. Das, D. Kumar, and K.V. Rao. 2010. Identification of host-dependent survival factors for intracellular Mycobacterium tuberculosis through an siRNA screen. PLoS Pathog. 6:e1000839. http://dx.doi.org/10.1371/journal .ppat.1000839

Jounai, N., F. Takeshita, K. Kobiyama, A. Sawano, A. Miyawaki, K.Q. Xin, K.J. Ishii, T. Kawai, S. Akira, K. Suzuki, and K. Okuda. 2007. The Atg5 Atg12 conjugate associates with innate antiviral immune re-sponses. Proc. Natl. Acad. Sci. USA. 104:14050–14055. http://dx.doi .org/10.1073/pnas.0704014104

Jung, C.H., C.B. Jun, S.H. Ro, Y.M. Kim, N.M. Otto, J. Cao, M. Kundu, and D.H. Kim. 2009. ULK-Atg13-FIP200 complexes mediate mTOR signaling to the autophagy machinery. Mol. Biol. Cell. 20:1992–2003. http://dx.doi.org/10.1091/mbc.E08-12-1249

Juris, L., M. Montino, P. Rube, P. Schlotterhose, M. Thumm, and R. Krick. 2015. PI3P binding by Atg21 organises Atg8 lipidation. EMBO J. 34:955–973. http://dx.doi.org/10.15252/embj.201488957

Kirkin, V., T. Lamark, Y.S. Sou, G. Bjørkøy, J.L. Nunn, J.A. Bruun, E. Shvets, D.G. McEwan, T.H. Clausen, P. Wild, et al. 2009. A role for NBR1 in autophagosomal degradation of ubiquitinated substrates. Mol. Cell. 33:505–516. http://dx.doi.org/10.1016/j.molcel.2009.01.020

Konno, H., K. Konno, and G.N. Barber. 2013. Cyclic dinucleotides trig-ger ULK1 (ATG1) phosphorylation of STING to prevent sustained innate immune signaling. Cell. 155:688–698. http://dx.doi.org/10 .1016/j.cell.2013.09.049

Korac, J., V. Schaeffer, I. Kovacevic, A.M. Clement, B. Jungblut, C. Behl, J. Terzic, and I. Dikic. 2013. Ubiquitin-independent function of op-tineurin in autophagic clearance of protein aggregates. J. Cell Sci. 126:580–592. http://dx.doi.org/10.1242/jcs.114926

Kyei, G.B., C. Dinkins, A.S. Davis, E. Roberts, S.B. Singh, C. Dong, L. Wu, E. Kominami, T. Ueno, A. Yamamoto, et al. 2009. Autophagy

988 Autophagy in neurodegeneration, inflammation, and infection | Rubinsztein et al.

2014. Dectin-1 pathway activates robust autophagy-dependent un-conventional protein secretion in human macrophages. J. Immunol. 192:5952–5962. http://dx.doi.org/10.4049/jimmunol.1303213

Orvedahl, A., D. Alexander, Z. Tallóczy, Q. Sun, Y. Wei, W. Zhang, D. Burns, D.A. Leib, and B. Levine. 2007. HSV-1 ICP34.5 confers neuro-virulence by targeting the Beclin 1 autophagy protein. Cell Host Microbe. 1:23–35. http://dx.doi.org/10.1016/j.chom.2006.12.001

Owen, K.A., C.B. Meyer, A.H. Bouton, and J.E. Casanova. 2014. Activation of focal adhesion kinase by Salmonella suppresses autophagy via an Akt/mTOR signaling pathway and promotes bacterial survival in macro-phages. PLoS Pathog. 10:e1004159. http://dx.doi.org/10.1371/journal .ppat.1004159

Pankiv, S., T.H. Clausen, T. Lamark, A. Brech, J.A. Bruun, H. Outzen, A. Øvervatn, G. Bjørkøy, and T. Johansen. 2007. p62/SQSTM1 binds di-rectly to Atg8/LC3 to facilitate degradation of ubiquitinated protein ag-gregates by autophagy. J. Biol. Chem. 282:24131–24145. http://dx.doi .org/10.1074/jbc.M702824200

Parihar, S.P., R. Guler, R. Khutlang, D.M. Lang, R. Hurdayal, M.M. Mhlanga, H. Suzuki, A.D. Marais, and F. Brombacher. 2014. Statin therapy reduces the mycobacterium tuberculosis burden in human macrophages and in mice by enhancing autophagy and phagosome maturation. J. Infect. Dis. 209:754–763. http://dx.doi.org/10.1093/infdis/jit550

Paul, S., A.K. Kashyap, W. Jia, Y.W. He, and B.C. Schaefer. 2012. Selective autophagy of the adaptor protein Bcl10 modulates T cell re-ceptor activation of NF-B. Immunity. 36:947–958. http://dx.doi.org/ 10.1016/j.immuni.2012.04.008

Pickford, F., E. Masliah, M. Britschgi, K. Lucin, R. Narasimhan, P.A. Jaeger, S. Small, B. Spencer, E. Rockenstein, B. Levine, and T. Wyss-Coray.2008. The autophagy-related protein beclin 1 shows reduced expression in early Alzheimer disease and regulates amyloid beta accumulation in mice. J. Clin. Invest. 118:2190–2199.

Pilli, M., J. Arko-Mensah, M. Ponpuak, E. Roberts, S. Master, M.A. Mandell, N. Dupont, W. Ornatowski, S. Jiang, S.B. Bradfute, et al. 2012. TBK-1 promotes autophagy-mediated antimicrobial defense by controlling autophagosome maturation. Immunity. 37:223–234. http://dx.doi.org/10.1016/j.immuni.2012.04.015

Ponpuak, M., M.A. Mandell, T. Kimura, S. Chauhan, C. Cleyrat, and V. Deretic. 2015. Secretory autophagy. Curr. Opin. Cell Biol. 35:106–116. http://dx.doi.org/10.1016/j.ceb.2015.04.016

Pottier, C., K.F. Bieniek, N. Finch, M. van de Vorst, M. Baker, R. Perkersen, P. Brown, T. Ravenscroft, M. van Blitterswijk, A.M. Nicholson, et al. 2015. Whole-genome sequencing reveals important role for TBK1 and OPTN mutations in frontotemporal lobar degenera-tion without motor neuron disease. Acta Neuropathol. http://dx.doi .org/10.1007/s00401-015-1436-x

Pozueta, J., R. Lefort, E.M. Ribe, C.M. Troy, O. Arancio, and M. Shelanski. 2013. Caspase-2 is required for dendritic spine and behavioural altera-tions in J20 APP transgenic mice. Nat. Commun. 4:1939. http://dx.doi .org/10.1038/ncomms2927

Puri, C., M. Renna, C.F. Bento, K. Moreau, and D.C. Rubinsztein. 2013. Diverse autophagosome membrane sources coalesce in recycling en-dosomes. Cell. 154:1285–1299. http://dx.doi.org/10.1016/j.cell.2013 .08.044

Py, B.F., M.M. Lipinski, and J. Yuan. 2007. Autophagy limits Listeria monocytogenes intracellular growth in the early phase of primary infec-tion. Autophagy. 3:117–125. http://dx.doi.org/10.4161/auto.3618

Ravikumar, B., R. Duden, and D.C. Rubinsztein. 2002. Aggregate-prone proteins with polyglutamine and polyalanine expansions are degraded by autophagy. Hum. Mol. Genet. 11:1107–1117. http://dx.doi .org/10.1093/hmg/11.9.1107

Ravikumar, B., Z. Berger, C. Vacher, C.J. O’Kane, and D.C. Rubinsztein. 2006. Rapamycin pre-treatment protects against apoptosis. Hum. Mol. Genet. 15:1209-1216. http://dx.doi.org/10.1093/hmg/ddl036

Ravikumar, B., C. Vacher, Z. Berger, J.E. Davies, S. Luo, L.G. Oroz, F. Scaravilli, D.F. Easton, R. Duden, C.J. O’Kane, and D.C. Rubinsztein. 2004. Inhibition of mTOR induces autophagy and reduces toxicity of polyglutamine expansions in fly and mouse models of Huntington dis-ease. Nat. Genet. 36:585–595. http://dx.doi.org/10.1038/ng1362

binds beclin-1 and activates autophagy. J. Biol. Chem. 289:3547–3554. http://dx.doi.org/10.1074/jbc.M113.536912

Menzies, F.M., J. Huebener, M. Renna, M. Bonin, O. Riess, and D.C. Rubinsztein. 2010. Autophagy induction reduces mutant ataxin-3 levels and toxicity in a mouse model of spinocerebellar ataxia type 3. Brain. 133:93–104. http://dx.doi.org/10.1093/brain/awp292

Menzies, F.M., A. Fleming, and D.C. Rubinsztein. 2015. Compromised autophagy and neurodegenerative diseases. Nat. Rev. Neurosci. 16:345–357. http://dx.doi.org/10.1038/nrn3961

Mesquita, F.S., M. Thomas, M. Sachse, A.J. Santos, R. Figueira, and D.W. Holden. 2012. The Salmonella deubiquitinase SseL inhibits selective autophagy of cytosolic aggregates. PLoS Pathog. 8:e1002743. http://dx.doi.org/10.1371/journal.ppat.1002743

Meulendyke, K.A., J.D. Croteau, and M.C. Zink. 2014. HIV life cycle, innate immunity and autophagy in the central nervous system. Curr Opin HIV AIDS. 9:565–571. http://dx.doi.org/10.1097/COH.0000000000000106

Meunier, E., M.S. Dick, R.F. Dreier, N. Schürmann, D. Kenzelmann Broz, S. Warming, M. Roose-Girma, D. Bumann, N. Kayagaki, K. Takeda,et al. 2014. Caspase-11 activation requires lysis of pathogen-containingvacuoles by IFN-induced GTPases. Nature. 509:366–370. http://dx.doi.org/10.1038/nature13157

Moreau, K., B. Ravikumar, M. Renna, C. Puri, and D.C. Rubinsztein. 2011. Autophagosome precursor maturation requires homotypic fusion. Cell. 146:303–317. http://dx.doi.org/10.1016/j.cell.2011.06.023

Moreau, K., A. Fleming, S. Imarisio, A. Lopez Ramirez, J.L. Mercer, M. Jimenez-Sanchez, C.F. Bento, C. Puri, E. Zavodszky, F. Siddiqi, et al. 2014. PICALM modulates autophagy activity and tau accumulation. Nat. Commun. 5:4998. http://dx.doi.org/10.1038/ncomms5998

Mostowy, S., V. Sancho-Shimizu, M.A. Hamon, R. Simeone, R. Brosch, T. Johansen, and P. Cossart. 2011. p62 and NDP52 proteins target in-tracytosolic Shigella and Listeria to different autophagy pathways. J. Biol.Chem. 286:26987–26995. http://dx.doi.org/10.1074/jbc.M111.223610

Nakagawa, I., A. Amano, N. Mizushima, A. Yamamoto, H. Yamaguchi, T. Kamimoto, A. Nara, J. Funao, M. Nakata, K. Tsuda, et al. 2004.Autophagy defends cells against invading group A Streptococcus. Science.306:1037–1040. http://dx.doi.org/10.1126/science.1103966

Nakahira, K., J.A. Haspel, V.A. Rathinam, S.J. Lee, T. Dolinay, H.C. Lam, J.A. Englert, M. Rabinovitch, M. Cernadas, H.P. Kim, et al. 2011. Autophagy proteins regulate innate immune responses by inhibiting the release of mitochondrial DNA mediated by the NALP3 inflammasome. Nat. Immunol. 12:222–230. http://dx.doi.org/10.1038/ni.1980

Napier, R.J., W. Rafi, M. Cheruvu, K.R. Powell, M.A. Zaunbrecher, W. Bornmann, P. Salgame, T.M. Shinnick, and D. Kalman. 2011.Imatinib-sensitive tyrosine kinases regulate mycobacterial pathogenesisand represent therapeutic targets against tuberculosis. Cell Host Microbe.10:475–485. http://dx.doi.org/10.1016/j.chom.2011.09.010

Niu, H., Q. Xiong, A. Yamamoto, M. Hayashi-Nishino, and Y. Rikihisa. 2012. Autophagosomes induced by a bacterial Beclin 1 binding protein facilitate obligatory intracellular infection. Proc. Natl. Acad. Sci. USA. 109:20800–20807. http://dx.doi.org/10.1073/pnas.1218674109

Ochaba, J., T. Lukacsovich, G. Csikos, S. Zheng, J. Margulis, L. Salazar, K. Mao, A.L. Lau, S.Y. Yeung, S. Humbert, et al. 2014. Potential function for the Huntingtin protein as a scaffold for selective autophagy. Proc. Natl. Acad. Sci. USA. 111:16889–16894. http://dx.doi.org/10.1073/pnas .1420103111

Ogawa, M., T. Yoshimori, T. Suzuki, H. Sagara, N. Mizushima, and C. Sasakawa. 2005. Escape of intracellular Shigella from autophagy. Science. 307:727–731. http://dx.doi.org/10.1126/science.1106036

Ogawa, M., Y. Yoshikawa, T. Kobayashi, H. Mimuro, M. Fukumatsu, K. Kiga, Z. Piao, H. Ashida, M. Yoshida, S. Kakuta, et al. 2011. A Tecpr1-dependent selective autophagy pathway targets bacterial pathogens. Cell Host Microbe. 9:376–389. http://dx.doi.org/10.1016/j.chom.2011 .04.010

Ohashi, Y., and S. Munro. 2010. Membrane delivery to the yeast autopha-gosome from the Golgi-endosomal system. Mol. Biol. Cell. 21:3998–4008. http://dx.doi.org/10.1091/mbc.E10-05-0457

Öhman, T., L. Teirilä, A.M. Lahesmaa-Korpinen, W. Cypryk, V. Veckman, S. Saijo, H. Wolff, S. Hautaniemi, T.A. Nyman, and S. Matikainen.

JEM Vol. 212, No. 7 989

Review

Shi, C.S., K. Shenderov, N.N. Huang, J. Kabat, M. Abu-Asab, K.A. Fitzgerald, A. Sher, and J.H. Kehrl. 2012. Activation of autophagy by inflammatory signals limits IL-1 production by targeting ubiquitinated inflammasomes for destruction. Nat. Immunol. 13:255–263. http://dx .doi.org/10.1038/ni.2215

Shoji-Kawata, S., R. Sumpter, M. Leveno, G.R. Campbell, Z. Zou, L. Kinch, A.D. Wilkins, Q. Sun, K. Pallauf, D. MacDuff, et al. 2013. Identification of a candidate therapeutic autophagy-inducing peptide. Nature. 494:201–206. http://dx.doi.org/10.1038/nature11866

Shpilka, T., E. Welter, N. Borovsky, N. Amar, M. Mari, F. Reggiori, and Z. Elazar. 2015. Lipid droplets and their component triglycerides andsteryl esters regulate autophagosome biogenesis. EMBO J. In press.

Spencer, B., R. Potkar, M. Trejo, E. Rockenstein, C. Patrick, R. Gindi, A. Adame, T. Wyss-Coray, and E. Masliah. 2009. Beclin 1 genetransfer activates autophagy and ameliorates the neurodegenerativepathology in alpha-synuclein models of Parkinson’s and Lewy bodydiseases. J. Neurosci. 29:13578–13588. http://dx.doi.org/10.1523/JNEUROSCI.4390-09.2009

Spilman, P., N. Podlutskaya, M.J. Hart, J. Debnath, O. Gorostiza, D. Bredesen, A. Richardson, R. Strong, and V. Galvan. 2010. Inhibition of mTOR by rapamycin abolishes cognitive deficits and reduces amy-loid-beta levels in a mouse model of Alzheimer’s disease. PLoS ONE. 5:e9979. http://dx.doi.org/10.1371/journal.pone.0009979

Stanley, S.A., A.K. Barczak, M.R. Silvis, S.S. Luo, K. Sogi, M. Vokes, M.-A. Bray, A.E. Carpenter, C.B. Moore, N. Siddiqi, et al. 2014. Identification of host-targeted small molecules that restrict intracel-lular Mycobacterium tuberculosis growth. PLoS Pathog. 10:e1003946. http://dx.doi.org/10.1371/journal.ppat.1003946

Starr, T., T.W. Ng, T.D. Wehrly, L.A. Knodler, and J. Celli. 2008. Brucella intracellular replication requires trafficking through the late endosomal/lysosomal compartment. Traffic. 9:678–694. http://dx.doi.org/10.1111/j.1600-0854.2008.00718.x

Steiner, J.A., E. Angot, and P. Brundin. 2011. A deadly spread: cellular mechanisms of -synuclein transfer. Cell Death Differ. 18:1425–1433. http://dx.doi.org/10.1038/cdd.2011.53

Stolz, A., A. Ernst, and I. Dikic. 2014. Cargo recognition and traffick-ing in selective autophagy. Nat. Cell Biol. 16:495–501. http://dx.doi .org/10.1038/ncb2979

Sundaramurthy, V., R. Barsacchi, N. Samusik, G. Marsico, J. Gilleron, I. Kalaidzidis, F. Meyenhofer, M. Bickle, Y. Kalaidzidis, and M. Zerial. 2013. Integration of chemical and RNAi multiparametric profiles iden-tifies triggers of intracellular mycobacterial killing. Cell Host Microbe. 13:129–142. http://dx.doi.org/10.1016/j.chom.2013.01.008

Tal, M.C., M. Sasai, H.K. Lee, B. Yordy, G.S. Shadel, and A. Iwasaki. 2009. Absence of autophagy results in reactive oxygen species-dependent am-plification of RLR signaling. Proc. Natl. Acad. Sci. USA. 106:2770–2775. http://dx.doi.org/10.1073/pnas.0807694106

Thurston, T.L., G. Ryzhakov, S. Bloor, N. von Muhlinen, and F. Randow. 2009. The TBK1 adaptor and autophagy receptor NDP52 restricts the proliferation of ubiquitin-coated bacteria. Nat. Immunol. 10:1215–1221. http://dx.doi.org/10.1038/ni.1800

Thurston, T.L., M.P. Wandel, N. von Muhlinen, A. Foeglein, and F. Randow. 2012. Galectin 8 targets damaged vesicles for autophagy to de-fend cells against bacterial invasion. Nature. 482:414–418. http://dx.doi .org/10.1038/nature10744

Tumbarello, D.A., B.J. Waxse, S.D. Arden, N.A. Bright, J. Kendrick-Jones, and F. Buss. 2012. Autophagy receptors link myosin VI to autopha-gosomes to mediate Tom1-dependent autophagosome maturation and fusion with the lysosome. Nat. Cell Biol. 14:1024–1035. http://dx.doi .org/10.1038/ncb2589

Vicinanza, M., V.I. Korolchuk, A. Ashkenazi, C. Puri, F.M. Menzies, J.H. Clarke, and D.C. Rubinsztein. 2015. PI(5)P regulates au-tophagosome biogenesis. Mol. Cell. 57:219–234. http://dx.doi.org/ 10.1016/j.molcel.2014.12.007

Wang, L.J., H.Y. Huang, M.P. Huang, W. Liou, Y.T. Chang, C.C. Wu, D.M. Ojcius, and Y.S. Chang. 2014. The microtubule-associated pro-tein EB1 links AIM2 inflammasomes with autophagy-dependent secre-tion. J. Biol. Chem. 289:29322–29333. http://dx.doi.org/10.1074/jbc.M114.559153

Ravikumar, B., K. Moreau, L. Jahreiss, C. Puri, and D.C. Rubinsztein. 2010. Plasma membrane contributes to the formation of pre-autophagosomal structures. Nat. Cell Biol. 12:747–757. http://dx.doi.org/10.1038/ ncb2078

Renna, M., C. Schaffner, K. Brown, S. Shang, M.H. Tamayo, K. Hegyi, N.J. Grimsey, D. Cusens, S. Coulter, J. Cooper, et al. 2011. Azithromycin blocks autophagy and may predispose cystic fibrosis patients to myco-bacterial infection. J. Clin. Invest. 121:3554–3563. http://dx.doi.org/ 10.1172/JCI46095

Rohn, T.T., R.A. Rissman, M.C. Davis, Y.E. Kim, C.W. Cotman, and E. Head. 2002. Caspase-9 activation and caspase cleavage of tau in theAlzheimer’s disease brain. Neurobiol. Dis. 11:341–354. http://dx.doi.org/10.1006/nbdi.2002.0549

Rubinsztein, D.C., T. Shpilka, and Z. Elazar. 2012a. Mechanisms of autophago-some biogenesis. Curr. Biol. 22:R29–R34. http://dx.doi.org/10.1016/ j.cub.2011.11.034

Rubinsztein, D.C., P. Codogno, and B. Levine. 2012b. Autophagy modula-tion as a potential therapeutic target for diverse diseases. Nat. Rev. Drug Discov. 11:709–730. http://dx.doi.org/10.1038/nrd3802

Rui, Y.N., Z. Xu, B. Patel, Z. Chen, D. Chen, A. Tito, G. David, Y. Sun, E.F. Stimming, H.J. Bellen, et al. 2015. Huntingtin functions as a scaf-fold for selective macroautophagy. Nat. Cell Biol. 17:262–275. http://dx.doi.org/10.1038/ncb3101

Russell, R.C., Y. Tian, H. Yuan, H.W. Park, Y.Y. Chang, J. Kim, H. Kim, T.P. Neufeld, A. Dillin, and K.L. Guan. 2013. ULK1 induces autoph-agy by phosphorylating Beclin-1 and activating VPS34 lipid kinase. Nat. Cell Biol. 15:741–750. http://dx.doi.org/10.1038/ncb2757

Sagnier, S., C.F. Daussy, S. Borel, V. Robert-Hebmann, M. Faure, F.P. Blanchet, B. Beaumelle, M. Biard-Piechaczyk, and L. Espert. 2015. Autophagy restricts HIV-1 infection by selectively degrading Tat in CD4+ T lymphocytes. J. Virol. 89:615–625. http://dx.doi.org/10.1128/ JVI.02174-14

Saitoh, T., N. Fujita, M.H. Jang, S. Uematsu, B.G. Yang, T. Satoh, H. Omori, T. Noda, N. Yamamoto, M. Komatsu, et al. 2008. Loss of the autophagy protein Atg16L1 enhances endotoxin-induced IL-1beta pro-duction. Nature. 456:264–268. http://dx.doi.org/10.1038/nature07383

Saitoh, T., N. Fujita, T. Hayashi, K. Takahara, T. Satoh, H. Lee, K. Matsunaga, S. Kageyama, H. Omori, T. Noda, et al. 2009. Atg9a controls dsDNA-driven dynamic translocation of STING and the innate immune re-sponse. Proc. Natl. Acad. Sci. USA. 106:20842–20846. http://dx.doi.org/10.1073/pnas.0911267106

Sarkar, S., R.A. Floto, Z. Berger, S. Imarisio, A. Cordenier, M. Pasco, L.J. Cook, and D.C. Rubinsztein. 2005. Lithium induces autophagy by in-hibiting inositol monophosphatase. J. Cell Biol. 170:1101–1111. http://dx.doi.org/10.1083/jcb.200504035

Sarkar, S., E.O. Perlstein, S. Imarisio, S. Pineau, A. Cordenier, R.L. Maglathlin, J.A. Webster, T.A. Lewis, C.J. O’Kane, S.L. Schreiber, and D.C. Rubinsztein. 2007. Small molecules enhance autophagy and re-duce toxicity in Huntington’s disease models. Nat. Chem. Biol. 3:331–338. http://dx.doi.org/10.1038/nchembio883

Sarkar, S., V.I. Korolchuk, M. Renna, S. Imarisio, A. Fleming, A. Williams, M. Garcia-Arencibia, C. Rose, S. Luo, B.R. Underwood, et al. 2011.Complex inhibitory effects of nitric oxide on autophagy. Mol. Cell.43:19–32. http://dx.doi.org/10.1016/j.molcel.2011.04.029

Sarkar, S., B. Carroll, Y. Buganim, D. Maetzel, A.H. Ng, J.P. Cassady, M.A. Cohen, S. Chakraborty, H. Wang, E. Spooner, et al. 2013. Impaired autophagy in the lipid-storage disorder Niemann-Pick type C1 dis-ease. Cell Reports. 5:1302–1315. http://dx.doi.org/10.1016/j.celrep .2013.10.042

Schaeffer, V., I. Lavenir, S. Ozcelik, M. Tolnay, D.T. Winkler, and M. Goedert. 2012. Stimulation of autophagy reduces neurodegeneration in a mouse model of human tauopathy. Brain. 135:2169–2177. http://dx.doi.org/10.1093/brain/aws143

Schiebler, M., K. Brown, K. Hegyi, S.M. Newton, M. Renna, L. Hepburn, C. Klapholz, S. Coulter, A. Obregón-Henao, M. Henao Tamayo, et al.2015. Functional drug screening reveals anticonvulsants as enhancers ofmTOR-independent autophagic killing of Mycobacterium tuberculosisthrough inositol depletion. EMBO Mol. Med. 7:127–139. http://dx.doi.org/10.15252/emmm.201404137

990 Autophagy in neurodegeneration, inflammation, and infection | Rubinsztein et al.

Warby, S.C., C.N. Doty, R.K. Graham, J.B. Carroll, Y.Z. Yang, R.R. Singaraja, C.M. Overall, and M.R. Hayden. 2008. Activated caspase-6 and caspase-6-cleaved fragments of huntingtin specifically colocalize in the nucleus. Hum. Mol. Genet. 17:2390–2404. http://dx.doi.org/10 .1093/hmg/ddn139

Watson, R.O., P.S. Manzanillo, and J.S. Cox. 2012. Extracellular M. tu-berculosis DNA targets bacteria for autophagy by activating the host DNA-sensing pathway. Cell. 150:803–815. http://dx.doi.org/10.1016/ j.cell.2012.06.040

Webb, J.L., B. Ravikumar, J. Atkins, J.N. Skepper, and D.C. Rubinsztein. 2003. Alpha-Synuclein is degraded by both autophagy and the protea-some. J. Biol. Chem. 278:25009–25013. http://dx.doi.org/10.1074/jbc .M300227200

Wellington, C.L., L.M. Ellerby, C.A. Gutekunst, D. Rogers, S. Warby, R.K. Graham, O. Loubser, J. van Raamsdonk, R. Singaraja, Y.Z. Yang, et al. 2002. Caspase cleavage of mutant huntingtin precedes neurode-generation in Huntington’s disease. J. Neurosci. 22:7862–7872.

Wild, P., H. Farhan, D.G. McEwan, S. Wagner, V.V. Rogov, N.R. Brady, B. Richter, J. Korac, O. Waidmann, C. Choudhary, et al. 2011. Phosphorylation of the autophagy receptor optineurin restricts Salmonella growth. Science. 333:228–233. http://dx.doi.org/10.1126/ science.1205405

Williams, A., S. Sarkar, P. Cuddon, E.K. Ttofi, S. Saiki, F.H. Siddiqi, L. Jahreiss, A. Fleming, D. Pask, P. Goldsmith, et al. 2008. Novel targets for Huntington’s disease in an mTOR-independent autophagy pathway. Nat. Chem. Biol. 4:295–305. http://dx.doi.org/10.1038/nchembio.79

Yoshikawa, Y., M. Ogawa, T. Hain, M. Yoshida, M. Fukumatsu, M. Kim, H. Mimuro, I. Nakagawa, T. Yanagawa, T. Ishii, et al. 2009. Listeriamonocytogenes ActA-mediated escape from autophagic recognition.Nat. Cell Biol. 11:1233–1240. http://dx.doi.org/10.1038/ncb1967

Young, A.R., E.Y. Chan, X.W. Hu, R. Köchl, S.G. Crawshaw, S. High, D.W. Hailey, J. Lippincott-Schwartz, and S.A. Tooze. 2006. Starvation and ULK1-dependent cycling of mammalian Atg9 between the TGN and endosomes. J. Cell Sci. 119:3888–3900. http://dx.doi .org/10.1242/jcs.03172

Zavodszky, E., M.N. Seaman, K. Moreau, M. Jimenez-Sanchez, S.Y. Breusegem, M.E. Harbour, and D.C. Rubinsztein. 2014. Mutation in VPS35 associated with Parkinson’s disease impairs WASH complex as-sociation and inhibits autophagy. Nat. Commun. 5:3828. http://dx.doi .org/10.1038/ncomms4828

Zhang, L., J. Yu, H. Pan, P. Hu, Y. Hao, W. Cai, H. Zhu, A.D. Yu, X. Xie, D. Ma, and J. Yuan. 2007. Small molecule regulators ofautophagy identified by an image-based high-throughput screen. Proc.Natl. Acad. Sci. USA. 104:19023–19028. http://dx.doi.org/10.1073/pnas.0709695104

Zheng, Y.T., S. Shahnazari, A. Brech, T. Lamark, T. Johansen, and J.H. Brumell. 2009. The adaptor protein p62/SQSTM1 targets invading bacteria to the autophagy pathway. J. Immunol. 183:5909–5916. http://dx.doi.org/10.4049/jimmunol.0900441

Zhou, R., A.S. Yazdi, P. Menu, and J. Tschopp. 2011. A role for mito-chondria in NLRP3 inflammasome activation. Nature. 469:221–225. http://dx.doi.org/10.1038/nature09663

Precision Instrumentation for the Sciences

MICROMANIPULATIONWorld-class micromanipulators, rock-solid mounting platforms, and precision X-Y translation are now offered in economically priced system bundles. Each system comes with 2 manipulators (MPC-385-2 or MPC-325-2) and either the classic moving scope design with translator, or large moving platform stage. The powerful Multi-Link Position Control ™ software allows you to link and coordinate movement of multiple devices connected to our MPC-200 controller.

AMPLIFIERSExplore our state-of-the-art patch clamp amplifi ers along with SutterPatch™, a comprehensive software package built on the foundation of Igor 7. The IPA™ (Integrated Patch Amplifi er) combines a low-noise patch clamp amplifi er, onboard D/A conversion and data acquisition via high-speed USB. Optimized for whole-cell patch clamp recordings, the IPA is backed by the service and support you’ve come to expect from Sutter Instrument at a low introductory price.

®

ONE D I G I T A L D R I V E , N O V A T O , CA. 94949 PHONE: 415.883.0128 | FAX: 415.883.0572 EMAIL: INFO@SUTTER.COM | WWW.SUTTER.COM

OPTICAL PRODUCTSThe Lambda VF-1™ Edge system allows selection of the band-width, as well as the center wavelength, upon command and uses a blocking fi lter to ensure the best blocking outside the bandpass. Other products include high-powered LED, Xenon, and plasma light sources, the MOM two-photon resonant scanning microscope, wavelength switchers, fi lter wheels, and the SmartShutter®.

MICROPIPETTE FABRICATIONSutter Instrument, the recognized leader in micropipettefabrication, offers leading edge technology in the P-2000 micropipette puller, allowing for fast and repeatable nanospray tip production with complete control of tip morphology. The P-2000 can be pre-programmed to produce tips to your specifi cations at no extra charge.

Visit us at Booth #1328