d activation by bexarotene promotes neuroprotection by ...for 3 days/week for 1 week, to 6-week-old...

SC I ENCE TRANS LAT IONAL MED I C I N E | R E S EARCH ART I C L E

HUNT INGTON ’S D I S EASE

1Department of Pediatrics, University of California, San Diego, La Jolla, CA 92093, USA.2Gene Expression Laboratory, Salk Institute for Biological Studies, San Diego, CA 92037,USA. 3Departments of Psychiatry, Neurology, and Pharmacology, Johns Hopkins Uni-versity School of Medicine, Baltimore, MD 21287, USA. 4Center for Drug Discovery andDepartment of Neurobiology, Duke University Medical Center, Durham, NC 27710,USA. 5Department of Neuroscience, Johns Hopkins University School of Medicine,Baltimore, MD 21287, USA. 6Department of Pathology, University of California, SanDiego, La Jolla, CA 92093, USA. 7Department of Neurosciences, University of California,San Diego, La Jolla, CA 92093, USA. 8HowardHughesMedical Institute, Salk Institute forBiological Studies, San Diego, CA 92037, USA. 9Department of Cellular and MolecularMedicine, University of California, San Diego, La Jolla, CA 92093, USA. 10Division ofBiological Sciences, University of California, San Diego, La Jolla, CA 92093, USA. 11Insti-tute for GenomicMedicine, University of California, San Diego, La Jolla, CA 92093, USA.12SanfordConsortium for RegenerativeMedicine, University of California, SanDiego, LaJolla, CA 92093, USA.*Corresponding author. Email: [email protected]

Dickey et al., Sci. Transl. Med. 9, eaal2332 (2017) 6 December 2017

Copyright © 2017

The Authors, some

rights reserved;

exclusive licensee

American Association

for the Advancement

of Science. No claim

to original U.S.

Government Works

http://stm.sciencem

Dow

nloaded from

PPARd activation by bexarotene promotesneuroprotection by restoring bioenergetic and qualitycontrol homeostasisAudrey S. Dickey,1 Dafne N. Sanchez,1 Martin Arreola,1 Kunal R. Sampat,1 Weiwei Fan,2

Nicolas Arbez,3 Sergey Akimov,3 Michael J. Van Kanegan,4 Kohta Ohnishi,1 Stephen K. Gilmore-Hall,1

April L. Flores,1 Janice M. Nguyen,1 Nicole Lomas,1 Cynthia L. Hsu,1 Donald C. Lo,4

Christopher A. Ross,3,5 Eliezer Masliah,6,7 Ronald M. Evans,2,8 Albert R. La Spada1,7,9,10,11,12*

Neuronsmustmaintain protein andmitochondrial quality control for optimal function, an energetically expensiveprocess. The peroxisome proliferator–activated receptors (PPARs) are ligand-activated transcription factors thatpromote mitochondrial biogenesis and oxidative metabolism. We recently determined that transcriptional dys-regulation of PPARd contributes to Huntington’s disease (HD), a progressive neurodegenerative disorder resultingfrom a CAG-polyglutamine repeat expansion in the huntingtin gene. We documented that the PPARd agonistKD3010 is an effective therapy for HD in a mouse model. PPARd forms a heterodimer with the retinoid X receptor(RXR), and RXR agonists are capable of promoting PPARd activation. One compound with potent RXR agonistactivity is the U.S. Food and Drug Administration–approved drug bexarotene. We tested the therapeutic potentialof bexarotene in HD and found that bexarotene was neuroprotective in cellular models of HD, including mediumspiny-like neurons generated from induced pluripotent stem cells (iPSCs) derived from patients with HD. To evaluatebexarotene as a treatment for HD, we treated the N171-82Qmousemodel with the drug and found that bexaroteneimproved motor function, reduced neurodegeneration, and increased survival. To determine the basis for PPARdneuroprotection,we evaluatedmetabolic function andnotedmarkedly impaired oxidativemetabolism inHDneurons,which was rescued by bexarotene or KD3010. We examined mitochondrial and protein quality control in cellularmodels of HD and observed that treatment with a PPARd agonist promoted cellular quality control. By boostingcellular activities that are dysfunctional in HD, PPARd activation may have therapeutic applications in HD andpotentially other neurodegenerative diseases.

ag.o
by guest on M
arch 4, 2021rg/
INTRODUCTION
Huntington’s disease (HD) is a progressive autosomal dominant neu-rodegenerative disorder in which patients develop motor and cognitiveimpairment (1). HD pathology is defined by degeneration and death ofmedium spiny neurons of the striatum, as well as cortical pyramidalneurons that project to the striatum (2, 3). In 1993, a CAG trinucleotiderepeat expansion mutation in the coding region of the huntingtin (htt)gene was identified as the cause of HD (4). As observed in other poly-glutamine (polyQ) repeat diseases, htt glutamine tracts that exceed acertain length threshold (~37 repeats inHD) adopt a new pathogenicconfirmation and are resistant to the normal processes of proteinturnover, leading to cellular toxicity and neurodegeneration (5). Thelength of the mutant htt polyQ expansion inversely correlates withthe age of disease onset and rate of disease progression in HD patients.

Neurons in the brain require continued production of high-energycompounds bymitochondria.We and others have linkedmitochondri-al dysfunction in HD to transcriptional dysregulation of peroxisomeproliferator–activated receptor (PPAR) gamma coactivator-1 alpha(PGC-1a), a coactivator that coordinates transcriptional programs thatculminate in mitochondrial biogenesis and enhanced oxidative metab-olism (6–8). The importance of PGC-1a for HD pathogenesis is under-scored by the observation that PGC-1a overexpression is sufficient torescue motor dysfunction, prevent accumulation of misfolded htt pro-tein, and stave off neurodegeneration inHDmice (9). To determine themechanistic basis for PGC-1a transcription interference inHD, we per-formed an unbiased screen that showed PPARs to be htt interactors anddocumented an interaction between PPARd and htt in non-neuronalcells, striatal-like neurons, and the cerebral cortex of HD mice (10).We noted that PPARd is highly expressed in neurons of the centralnervous system (CNS) and demonstrated that expression of dominant-negative PPARd in CNS is sufficient to produce motor phenotypes,neurodegeneration, mitochondrial defects, and transcriptional abnormal-ities that closely parallel HD disease phenotypes (10). We then evalu-ated a selective, potent PPARd agonist, KD3010, and after confirmingthat it crosses the blood-brain barrier to up-regulate expression of PPARdtarget genes in the cortex and striatum, we tested KD3010 in N171-82Qtransgenic mice, a rodent model of HD. This study established theefficacy of KD3010 PPARd agonist therapy as a potential therapeuticapproach for HD (10).

One facet of PPARd biology with relevance to therapy developmentis that PPARd forms a heterodimer with retinoid X receptor (RXR), andthe resulting “permissive” PPARd-RXR heterodimer is subject to dual

1 of 13

http://stm.sciencemag.org/


http://stm.sciencem

agD

ownloaded from

ligand regulation, meaning that RXR agonists can promote PPARdactivation (11). One drug compoundwith potent RXR agonist activityis bexarotene, a synthetic product structurally similar to retinoic acidcompounds, known endogenous RXR ligands. Bexarotene (Targretin)is U.S. Food and Drug Administration–approved for use in patientswith T cell cutaneous lymphoma. One provocative study reported thatbexarotene administration to a mouse model of Alzheimer’s disease(AD) yielded amarked rescue of cognitive, social, and olfactory deficits,accompanied by improved neural circuit function and enhanced clear-ance of soluble Ab oligomers (12). The mechanistic basis for this effectwas proposed to involve increased PPARg activation (13). BecausePPARd is highly expressed in CNS neurons, more so than PPARg (14),the mechanistic basis for the therapeutic action of bexarotene in ADdeserves reconsideration, in light of our discovery of a role for PPARd inmaintainingnormal nervous system function (10) and recentworkdem-onstrating the neuroprotective effect of PPARd agonist treatment inAD mice (15).

Here, we considered the neurotherapeutic potential of bexarotenein HD and found that bexarotene was neuroprotective in multiplecellularmodels ofHD, ranging frommouse striatal and cortical neuronsto medium spiny neurons generated from induced pluripotent stemcells (iPSCs) derived from patients with HD. We then treated theN171-82Q HD mouse model with bexarotene and observed improvedneuron survival and motor function. To determine the molecular basisfor PPARd agonist therapy, we evaluated metabolic dysfunction in HDand documented markedly impaired oxidative metabolism in HD neu-rons, whichwas rescued by bexarotene orKD3010.We examinedmito-chondrial and protein quality control in cellular models of HD andobserved that PPARd agonist therapy achieved neuroprotection bypromoting quality control function.

by guest on March 4, 2021

.org/

RESULTSBexarotene promotes PPARd activation resultingin neuroprotectionTo evaluate the effect of bexarotene on PPARd activation, we cotrans-fected PPARd and a 3x–PPAR response element luciferase reporterconstruct into primary cortical neurons derived from a bacterial artifi-cial chromosome transgenic HD (BAC-HD) mouse model containingthe full-length htt gene with 97 glutamine repeats or into primary cor-tical neurons derived from littermate control [wild-type (WT)] mice.We noted marked induction of PPARd activation upon bexarotenetreatment (Fig. 1A). Bexarotene promotion of PPARd activation wascomparable to PPARd activation using a PPARd-selective agonistGW501516 (Fig. 1A). We also observed a marked increase in PPARdactivation when we boosted RXR expression (fig. S1A), corroboratingthe permissive nature of PPARd activation upon heterodimerizationwith RXR (11). To determine whether bexarotene-induced PPARd ac-tivation countered mutant htt neurotoxicity, we assessed the effect ofbexarotene treatment on mitochondrial dysfunction and cell death inBAC-HD neurons. When we measured mitochondrial membranepotential using two different fluorescent probes, we noted significantlyreduced mitochondrial membrane potential in BAC-HD neurons,which was rescued upon bexarotene treatment (Fig. 1B and fig. S1B).Similarly, bexarotene treatment yielded a significant reduction in thefrequency of BAC-HD neurons that displayed caspase-3 activation,an indicator of impending cellular demise (Fig. 1C). Previous studiesof bexarotene neuroprotection in an AD mouse model have proposedthat bexarotene neuroprotection relies upon activation of PPARg (12).


To determine the basis for bexarotene neuroprotection in HD neurons,we repeated bexarotene treatment of BAC-HD primary neurons alongwith concurrent knockdown of either PPARa, PPARd, or PPARg,which was accomplished by coexpression of short hairpin RNA(shRNA) vectors directed against each of the different PPARs (fig.S1C). When we measured mitochondrial membrane potential inBAC-HDneurons treatedwith bexarotene and subjected to knockdownof either PPARa, PPARd, or PPARg, we found that bexarotene rescueof mitochondrial membrane potential occurred despite PPARa orPPARg knockdown but was prevented by concurrent knockdown ofPPARd (Fig. 1D). Bexarotene neuroprotection was not limited toBAC-HD primary cortical neurons because bexarotene treatment alsoimproved mitochondrial function and cell survival of WT neurons;enhanced WT neuron function upon exposure to bexarotene alsodepended upon PPARd (fig. S1, D and E). We then assayed bexarotenerescueofBAC-HDneuronal cell death andnoted thatPPARd knockdowneliminated any benefit from bexarotene treatment, whereas PPARaor PPARg knockdown did not significantly affect bexarotene neuro-protection (Fig. 1E). Finally, to determine whether bexarotene treatmentof BAC-HD primary neurons promoted increased PPARd activation,we measured the RNA expression of PPARd targets previously shownto respond to PPARd agonist treatment pharmacodynamically inmice (10). We documented bexarotene-induced expression of thesePPARd target genes (Fig. 1F). We noted that combined bexarotene +GW501516 treatment often yielded even greater increases in PPARdtarget gene expression (Fig. 1F).

To further evaluate bexarotene neuroprotection in HD, we pursuedexperiments in different models of mutant htt neurotoxicity thataim to recapitulate HD neuropathology. First, we transfected co-cultured mouse cortical and striatal neurons with N-terminal htt con-taining 90 amino acids with either an 8-glutamine repeat (Nt-90-8Q)or a 73-glutamine repeat expansion (Nt-90-73Q) and treated coculturedcortical and striatal neurons with increasing bexarotene concentrations.We observed a significant increase (P < 0.05 or P < 0.01) in neuronsurvival in a dose-dependent manner (Fig. 2, A and B). We then testedthe effect of bexarotene in primary cortical neurons transfected withN-terminal htt containing 586 amino acids with an 82-glutaminerepeat expansion (Nt-586-82Q) and noted a dose-dependent reduc-tion (P < 0.01 or P < 0.001) in neuronal cell death (Fig. 2C). We alsodifferentiated human HD iPSCs into striatal medium spiny-like neuronsand transferred them to brain-derived neurotrophic factor (BDNF)–freeneural induction medium because withdrawal of neurotrophic factorsupport promoted neuronal cell death. To evaluate rescue, we supple-mented the media with either bexarotene or BDNF, which robustlyprevented neuronal cell death in this system. We observed markedprotection of HD medium spiny-like neurons from cell death uponeither bexarotene or BDNF treatment (Fig. 2D). These findings indi-cate that bexarotene can ameliorate mutant htt neurotoxicity.

Bexarotene treatment improves motor function and rescuesneurodegeneration in HD miceBecause bexarotene displayed neuroprotection against mutant htttoxicity and is already approved for use in humans, we investigatedbexarotene treatment in N171-82Q mice, which recapitulate HD-likemotor phenotypes and neurodegeneration within a time frame of 5 to6months (16). Bexarotene is a lipophilic molecule that readily crossesthe blood-brain barrier in rodents (17, 18). To establish the dosage for apreclinical trial, we performed a pharmacodynamics study by deliveringbexarotene via intraperitoneal injection, at either 10 or 30 mg/kg

2 of 13




http://stm.sciencem

ag.org/D

ownloaded from

for 3 days/week for 1 week, to 6-week-old WT C57BL/6J mice andthen measuring PPARd target gene expression in the striatum. Weobserved comparable increases in PPARd target gene expression inthe brains of mice on the dosage regimen (10 mg/kg) and the dosageregimen (30 mg/kg) (fig. S2). However, we noted that the bexarotenedosage regimen (30 mg/kg) was not well tolerated, causing signifi-cant morbidity. The bexarotene dosage (10 mg/kg) regimen did notcause weight loss or visible side effects based on a neurological screen-ing exam, nor did we detect any evidence of organ toxicity at necropsy.

Next, we injected N171-82Q mice with either bexarotene or vehicle(10 mg/kg per day), three times per week, beginning at 6 weeks of age.We adhered to recommended preclinical trial guidelines, intended toavoid spurious results (19, 20). We tracked the progression of diseasephenotypes in vehicle- and bexarotene-treated HDmice by performing


a composite neurological examination (21) and rotarod analysis atmonthly intervals. Bexarotene treatment rescuedneurological dysfunctionand improvedmotor function inHDmice, as compared to vehicle-treatedHDmice (Fig. 3, A and B, and fig. S3, A toD). Bexarotene also extendedlife span inHDmice by an average of 9.8 days (Fig. 3C). Neuropathologyanalysis further indicated that bexarotene treatment yielded a reductionin htt protein aggregates andprevented neuronal cell loss in the striatumof HD mice (Fig. 3, D to F).

Bexarotene rescues PPARd target gene expression in bothCNS and skeletal muscle of HD miceTo confirm that bexarotene treatment elicited induction of PPARd ac-tivation in HD mice, we measured the expression of the PPARd targetgenes angiopoietin-like 4 (Angptl4) and uncoupling protein 2 (Ucp2)

0

1

2

3

4

A

Lu

cife

rase

act

ivit

y(a

rbit

rary

un

its)

**

**

Vehicle

GW501516

WT BAC-HD

**

Bexarotene

Mit

och

on

dri

al m

emb

ran

e p

ote

nti

al (

arb

itra

ry u

nit

s)

WTBAC-HD

0

0.5

1

1.5

2

2.5

3

Vehicle Bexarotene

*

*B

% N

euro

ns

wit

hac

tive

cas

pas

e-3

0 10 20 30 40 50 60 70

Vehicle Bexarotene

WTBAC-HD

* **

C

Mit

och

on

dri

al m

emb

ran

e p

ote

nti

al (

arb

itra

ry u

nit

s)

ControlPPARα

0

0.5

1

1.5

2

2.5

3

Vehicle Bexarotene

*

PPARδPPARγ

D shRNA:

% N

euro

ns

wit

hac

tive

cas

pas

e-3

Vehicle Bexarotene

E

*

ControlPPARαPPARδPPARγ

shRNA:F

0

0.2

0.4

0.6

0.8

1

1.2

PDK4 ANGPTL4 UCP2

UntreatedBexaroteneGW501516Bexarotene + GW501516

Rel

ativ

e g

ene

exp

ress

ion

******

**

****

****

80

0 10 20 30 40 50 60 70 80

eline as in (B). (E) We quantified active caspase-3 immunostaining of BAC-HD mouse primary cortical neurons, transfected

Fig. 1. Bexarotene promotes PPARd activa-tion to ameliorate the neurotoxicity ofmutanthuntingtin. (A) We measured 3x–PPAR re-sponse element luciferase reporter activity inprimary cortical neurons from WT control miceor bacterial artificial chromosome transgenicHuntington’s disease (BAC-HD) mice cotrans-fected with Renilla luciferase vector and treatedwith bexarotene (500 nM), GW501516 (100 nM), orvehicle. Peroxisome proliferator–activated recep-tor d (PPARd) activation in BAC-HDmouse neuronswas repressed at baseline compared to wild-type(WT) mouse neurons. **P < 0.01, Student’s t test.Bexarotene and GW501516 treatment promotedPPARd activation in BAC-HD mouse neurons.**P < 0.01, analysis of variance (ANOVA) withpost-hoc Tukey test. n = 3 biological replicates;n = 3 technical replicates. Results were normalizedto WT mouse neurons at baseline. (B) Mitochon-drial membrane potential of primary cortical neu-rons from WT and BAC-HD mice, treated withvehicle or bexarotene (500 nM), was determinedfrom the ratio of mitochondrial to cytosolic JC-1 flu-orescence. *P < 0.05, Student’s t test. n = 3 bio-logical replicates; n = 3 technical replicates. Resultswere normalized to WT mouse neurons at baseline.Similar results were obtained using tetramethyl-rhodamine methyl ester (TMRM) as the fluorescentprobe (fig. S1B). (C) We quantified active caspase-3immunostaining of primary cortical neurons fromWT and BAC-HD mice, treated with vehicle or bex-arotene (500 nM) for 24 hours and H2O2 (25 mM)for 4 hours. *P < 0.05, **P < 0.01, Student’s t test.n = 3 biological replicates; 30 to 50 cells werecounted per experiment. (D) Mitochondrial mem-brane potential was measured in BAC-HD mouseprimary cortical neurons, transfected with the in-dicated short hairpin RNA (shRNA) expressionvector (control = scrambled shRNA), and treatedwith vehicle or bexarotene (500 nM). Mitochondrialmembrane potential was determined from theratio ofmitochondrial to cytosolic JC-1 fluorescence.*P < 0.05, ANOVA with post-hoc Tukey test. n = 3biological replicates; n = 3 technical replicates.
Results were normalized to WT mouse neurons at baswith the indicated shRNA expression vectors, and treated with vehicle or bexarotene (500 nM) for 24 hours and H2O2 (25 mM) for 4 hours. *P < 0.05, ANOVA with post-hocTukey test. n = 3 biological replicates; 30 to 50 cells were counted per experiment. (F) We performed reverse transcription polymerase chain reaction (RT-PCR) analysis of RNAexpression of the PPARd target genes pyruvate dehydrogenase kinase isoform 4 (PDK4), angiopoietin-like 4 (Angptl4), and uncoupling protein 2 (UCP2) in BAC-HD mouseprimary cortical neurons, treated as indicated. **P < 0.01, ANOVA with post-hoc Tukey test. n = 6 independent experiments. Error bars represent SEM.
3 of 13





http://stm.sciencem

ag.org/D

ownloaded from

in the cortex and striatum of bexarotene-treated mice and documented markedincreases inAngptl4 andUcp2 expressionin comparison to vehicle-treatedHDmice(Fig. 4, A and B). In human patients,sampling of CNS tissues was not feasible.Because PPARd is highly expressed inskeletal muscle, a practical alternativestrategy would be to measure PPARdtarget gene expression in this highly ac-cessible tissue in the HD mice. We thusassayed the expression of five PPARd tar-get genes,Ucp2, adipose differentiation–related protein (Adfp), lipoprotein lipase(Lpl), pyruvate dehydrogenase kinase iso-form 4 (PDK4), and stearoyl–coenzymeAdesaturase 1 (Scd1), in quadricepsmusclesamples obtained fromWTcontrols andbexarotene- and vehicle-treated HDmice.For all tested PPARd target genes exceptLpl, we observed rescue of gene expres-sion; forPdk4 and Scd1, we notedmarkedup-regulation of expression at three- tofivefold that of WT HD mice (Fig. 4C).To assess the rapidity of productive PPARdtarget engagement in mouse skeletalmuscle, we assembled two additional co-horts of HD N171-82Q mice, and wetreated one HD cohort with vehicle andthe other HD cohort with bexaroteneat 10 mg/kg for 1 week. Induction ofPPARd target gene expression in quadri-ceps muscle was detected for three ofthe five targets in HDmice treated withbexarotene for just 1 week (fig. S4). Theseresults indicate that PPARd target geneexpression in mouse skeletal musclecould serve as a marker of response toPPARd agonist treatment.

PPARd agonist treatment restoresoxidative metabolic function inHD miceOur bexarotene preclinical trial in HDmice, together with a recent preclinicaltrial of the PPARd agonist KD3010 (10),suggests that PPARd agonists may be apotential treatment for HD. How doesPPARd agonist treatment achieve in vivoneuroprotection? PPARd has been shownto improve bioenergetics function by pro-moting mitochondrial adenosine tri-phosphate (ATP) generation in skeletalmuscle (22, 23). Mitochondrial dysfunc-tion is a key feature of HD pathogenesis(24). To determine whether PPARdagonist treatment affected mitochondrialfunction in HD, we performed a bio-energetics profile of HD mouse primary

Co

rtic

al n

euro

n s

urv

ival

(arb

itra

ry u

nit

s)

0

1.0

0.50

0.75

Nt-90-8Q Nt-90-73Q

A

**

0 μM0.1 μM0.3 μM

0.25

Bexarotene

Str

iata

l neu

ron

su

rviv

al(a

rbit

rary

un

its)

0

1.0

0.50

0.75

Nt-90-8Q Nt-90-73Q

B

**

0 μM0.1 μM0.3 μM

0.25

Bexarotene

0.1 0.5 1 Vehicle

Bexarotene (μM)

% N

euro

ns

wit

hco

nd

ense

d n

ucl

ei

0

10

20

30

40

50 Htt Nt-586-22QHtt Nt-586-82Q

C

*****

0

10

20

30

40

BDNF

% N

euro

ns

wit

hco

nd

ense

d n

ucl

ei

**

*D

**

Bexarotene

* *

Fig. 2. Bexarotene is neuroprotective in mouse and human HD neurons in vitro. (A) We measured cell survivalin mouse CD1 cortical neuron-striatal neuron cocultures transfected with the indicated Htt expression vector andtreated with bexarotene at the indicated concentrations. Total numbers counted for fluorescently transfected neu-rons for each condition were normalized to Htt Nt-90-8Q–transfected neurons at baseline, with neuron survivalarbitrarily set to 1. *P < 0.05, **P < 0.01, Student’s t test. n = 3 independent experiments. (B) We measured survivalof striatal neurons in cortical neuron-striatal neuron cocultures transfected with the indicated Htt expression vectorand treated with bexarotene at the indicated concentration. Total numbers counted for fluorescently transfectedneurons for each condition were normalized to Htt Nt-90-8Q–transfected neurons at baseline, with survival arbitrar-ily set to 1. *P < 0.05, **P < 0.01, Student’s t test. n = 3 independent experiments. (C) We quantified cell death inmouse primary cortical neurons, transfected with the indicated Htt expression vector, and treated with vehicle orbexarotene at the indicated concentration. **P < 0.01, ***P < 0.001, Student’s t test. n = 6 independent experiments.(D) We quantified cell death in medium spiny-like neurons differentiated from an induced pluripotent stem cell linederived from a patient with HD carrying a 60Q allele in the huntingtin gene and treated with bexarotene (1.0 mM) orbrain-derived neurotrophic factor (BDNF; 20 ng/ml). *P < 0.05, **P <0.01, Student’s t test. n = 3 independentexperiments. Error bars represent SEM.

4 of 13




http://stm.sciencem

ag.org/D

ownloaded from

cortical neurons in comparison toWTmouse primary cortical neuronsand evaluated the effects of bexarotene and KD3010. Extracellularflux analysis revealed that HD mouse neurons displayed a markedlyreduced oxygen consumption rate (OCR) at baseline compared toWT neurons and that the spare respiratory capacity of HD neurons,reflected by the maximal OCR, was also markedly reduced in compar-ison toWTneurons (Fig. 5A). Because PPARgwas shown to drive ATPgeneration throughmitochondrial oxidation of fatty acids (25), we sup-plied WT and HDmouse primary cortical neurons with palmitate, butthis substrate did not produce any further differences in oxidative me-tabolism betweenWT and HD neurons. Treatment with bexarotene orKD3010 yielded a marked improvement in both basal OCR and max-imal OCR inHD neurons (Fig. 5, A to C), restoring OCR to that ofWTneurons. In addition tomeasuring OCR, we also recorded rates of gly-colysis by assaying the extracellular acidification rate (ECAR) andconfirmed that HD neurons produce ATP primarily via glycolysisbut that bexarotene or KD3010 treatment reverts HD neurons to anoxidative mode of energy production (Fig. 5D). Given the potency ofPPARd agonist treatment for boosting oxidative metabolism in HDneurons, we measured the effect of bexarotene or KD3010 treatment


on WT neurons and noted significant increases in basal OCR (P <0.01) and maximal OCR (P < 0.05) (fig. S5A), indicating that PPARdactivation is also capable of boosting bioenergetics function inWTneu-rons. To determine the relevance of this bioenergetics profile of HDprimary neurons to the effects of bexarotene treatment in vivo, weperformed quantitative reverse transcription polymerase chain reaction(RT-PCR) analysis on striatal RNAs for subsets of PPARd target geneswhose protein products function in the oxidative phosphorylation,glycolysis, or gluconeogenesis pathways. We documented restorationof a gene expression pattern matching the WT oxidative profile forbexarotene-treatedmice (Fig. 5E).Whenwe extended this analysis toskeletal muscle from bexarotene- and KD3010-treated HD mice, weobserved a similar phenomenon (fig. S5B).

PPARd improves mitochondrial and protein quality controlto achieve neuroprotectionNeurons are placed under a constant demand for energy and are in acontinuous battle to maintain mitochondrial quality control and pro-tein quality control. Decompensation of proteostasis andmitochondrialquality control are defining features of neurodegenerative diseases,

0 1 2 3 4 5 6 7 8 9

HD-82Q + vehicle HD-82Q + bexarotene Non-Tg

Neu

rolo

gic

al d

ysfu

nct

ion

sco

re

Age (in months)

1.5 2.5 4.5 3.5 100

110

120

130

140

150

160

170

Lat

ency

to

fal

l (se

c)

Age (in months)

1.5 2.5 4.5 3.5

HD-82Q + vehicle HD-82Q + bexarotene

**

*

**

A B

EM48

NeuN

HD-82QNon-Tg Vehicle Bexarotene

D

0

1.0

1.5

2.0

Neu

ron

num

ber

(× 1

0 )6

0.5

Non-Tg

2.5

0

1

2

EM

48 p

unct

a/0.

1 m

33

4


Non-TgHD-82Q + vehicle HD-82Q + bexarotene

E F

% s

urv

ival

Age (in weeks)


C

180

Non-Tg

**

**

**

***

**

**

Fig. 3. Bexarotene is neuroprotective in the N171-82Q mouse model of HD. (A) We recorded neurological dysfunction scores (0 to 12, where 0 is normal and 12 isseverely affected) in cohorts (n = 10 to 19 mice per group) of nontransgenic control mice (Non-Tg), vehicle-treated HD mice, and bexarotene-treated HD mice atmonthly intervals, beginning at the initiation of the treatment. *P < 0.05, **P < 0.01, ANOVA with post-hoc Tukey test. (B) We measured times for the latency to fallon the rotarod test in cohorts (n = 12 to 24 mice per group) of nontransgenic control mice, vehicle-treated HD mice, and bexarotene-treated HD mice at monthlyintervals, beginning at the initiation of the treatment. *P < 0.05, **P < 0.01, ANOVA with post-hoc Tukey test. (C) We measured survival of vehicle-treated HD mice (n = 24)and bexarotene-treated HD mice (n = 26) over time. Bexarotene-treated HD mice lived longer than did vehicle-treated HD mice. P < 0.05, log-rank test. (D) Sections ofstriatum from 18-week-old nontransgenic control mice, vehicle-treated HD mice, and bexarotene-treated HD mice (n = 5 to 7 mice per group) were immunostained for theneuronal marker NeuN and with the anti-polyglutamine antibody EM48. Scale bar, 20 mm. (E) We quantified EM48 puncta from the data in (D). *P < 0.05, ANOVA with post-hoc Tukey test. n = 5 to 8 mice per group. (F) We quantified neuron numbers from data in (D). *P < 0.05, ANOVA with post-hoc Tukey test. n = 5 to 8 mice per group. Errorbars represent SEM.

5 of 13




http://stm.sciencem

ag.org/D

ownloaded from

including HD (24, 26). Because quality control processes are ex-tremely energy intensive, and PGC-1a genetic rescue prevents pro-tein aggregation and boosts mitochondrial function (9), we reasonedthat PPARd agonist treatment may ameliorate defects in mitochon-drial quality control and protein quality control through its rescue ofmetabolic function. To test this hypothesis, we examined the mito-


chondrial morphology of striatal-like neu-rons carrying homozygous CAG-111repeat expansion mutations in the mousehtt gene (ST-HdhQ111/Q111) (27). Com-parison of ST-Hdh Q111/Q111 cells withcontrol ST-Hdh Q7/Q7 cells revealed anincrease in mitochondrial fragmenta-tion in Q111/Q111 cells (Fig. 6, A and B).Bexarotene treatment of ST-Hdh Q111/Q111 striatal-like cells yielded a signifi-cant reduction (P < 0.05 or P < 0.01) inmitochondrial fragmentation in a dose-dependent fashion (Fig. 6B). BecausePPARd agonist treatment boosted meta-bolic function in normal neurons (fig.S4A), we tested the effect of PPARd ac-tivation on ST-HdhQ7/Q7 cells and ob-served an increase inmitochondrial length;we detected a reduction in mitochondriallength uponPPARd knockdown (Fig. 6C).Oxidative stress impairs mitochondrialquality control, promoting fragmentation(28). When we treated ST-Hdh Q7/Q7striatal-like cells with hydrogen peroxide,we observed a marked reduction in mito-chondrial length, and this mitochondriallength reduction could be amelioratedby combined PPARd overexpression andbexarotene treatment (Fig. 6D). Similarly,hydrogenperoxide yieldedmarkedly greatermitochondrial fragmentation in ST-HdhQ111/Q111 striatal-like cells, which wasrescued by combined PPARd overexpres-sion and bexarotene agonist treatment(Fig. 6E). To assess the in vivo relevanceof these findings, we quantified mito-chondrial genomic DNA (mitoDNA) andnuclear genomicDNA(nDNA) in the stri-atum of bexarotene-treated HD mice. Wedocumented an increase in themitoDNA/nDNA ratio, which is an index of theabundance of mitochondrial biomass, inHDmice treatedwith bexarotene (Fig. 6F).

To determine whether PPARd activa-tion status affects protein quality control,we used a Neuro2a cell culture model ofmutant htt aggregation, where transfec-tion of N-terminal htt protein with 104glutamine repeats yields marked aggregateformation in Neuro2a cells subjected tooxidative stress (9). Using this system,we found that treatment with the PPARdagonist GW501516 or bexarotene elicited

a marked reduction in htt-Q104 protein aggregation (Fig. 7A) and con-firmed that bexarotene-mediated turnover of mutant htt protein wasRXR-dependent (fig. S6). To establish which arm of the proteostasispathway was responsible for the bexarotene-mediated reduction inhtt aggregation, we performed htt-104Q transfection of Neuro2a cellsin the presence of bexarotene, in combination with either lactacystin or

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pre

ssio

n

Non-Tg HD + Veh HD + Bex

Angptl4

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pre

ssio

n


UCP2

*

*

A

0.0

0.5

1.0

1.5

2.0

Rel

ativ

e ex

pre

ssio

n


Angptl4

0.0

0.5

1.0

1.5

2.0R

elat

ive

exp

ress

ion


UCP2

**

B

C

0

1

2

3

4

5

UCP2 ADFP LPL PDK4 SCD1

Rel

ativ

e ex

pre

ssio

n


* *

**

**

Cortex

Striatum

Quadriceps muscle

Fig. 4. Bexarotene promotes PPARd activation of target genes in mouse brain and muscle. (A) We performedRT-PCR analysis to measure RNA expression of the PPARd target genes Angptl4 and UCP2 in the cortex of 18-week-old nontransgenic control mice, vehicle-treated HD mice, and bexarotene-treated HD mice. **P < 0.01, ANOVA withpost-hoc Tukey test. n = 9 to 12 mice per group. (B) We performed RT-PCR analysis of RNA expression of Angptl4and UCP2 in the striatum of 18-week-old nontransgenic control mice, vehicle-treated HD mice, and bexarotene-treated HD mice. **P < 0.01, ANOVA with post-hoc Tukey test. n = 9 to 12 mice per group. (C) We performed RT-PCRanalysis of RNA expression of five PPARd target genes in the quadriceps muscle of 18-week-old nontransgeniccontrol mice, vehicle-treated HD mice, and bexarotene-treated HD mice. ADFP, adipose differentiation–related pro-tein; LPL, lipoprotein lipase; SCD1, stearoyl–coenzyme A desaturase 1. *P < 0.05, **P < 0.01, ANOVA with post-hocTukey test. n = 9 mice per group. Error bars represent SEM.

6 of 13




http://stm.sciencem

ag.org/D

ownloaded from

spautin-1. We found that spautin-1 inhibition of autophagy blunted thereduction of htt aggregation achieved with bexarotene treatment, butlactacystin, an inhibitor of the ubiquitin-proteasome system, had no signif-icant effect (Fig. 7B). Knockdown of the autophagy-related 7 (Atg7) genesimilarly yielded a significant blunting of bexarotene-mediated htt aggre-gate rescue in this system (Fig. 7B). Todirectly evaluate the effect of PPARdactivation on autophagy, we assayed autophagy flux in Neuro2a cellssubjected to PPARd shRNA knockdown or PPARd agonist activationand noted a reduction in autophagy flux upon PPARd knockdown anda significant increase in autophagy flux with PPARd agonist activation(Fig. 7C). This increase in autophagy flux could also be achieved withbexarotene activation of RXR (Fig. 7D). Transcription factor EB (TFEB)is a master regulator of autophagy (29), and upon PGC-1a induction,TFEB expression increases, thereby promoting autophagy (9). To assessthe role of TFEB inPPARd activation of autophagy, wemeasured autoph-agy flux inHeLa cells deficient in TFEB (fig. S7, A and B).We detected an


increase in autophagy flux upon PPARd agonist treatment that was com-parable to the increased autophagy flux observed inWTHeLa cells treatedwith aPPARd agonist (fig. S7,C andD), thereby rulingout a role forTFEBinPPARd-mediatedautophagyactivation.Todeterminewhetherbexaroteneamelioration of mutant htt protein aggregation depended upon PPARdactivation, we quantified htt-Q104 aggregation in Neuro2a cells treatedwith bexarotene in the presence or absence of a specific PPARd inhib-itor (GSK3787) and noted that PPARd inhibition abrogated bexaroteneamelioration of htt protein aggregation (Fig. 7E). These findings indi-cate that PPARd activation achieves neuroprotection by improvingmitochondrial and protein quality control pathways.

DISCUSSIONHDand other neurodegenerative disorders, includingADandParkinson’sdisease (PD), share two key defining cellular pathologies: mitochondrial

OC

R (

pm

ol/m

in)

8 16 24 32 40 48 56 64 72 80

Time (min)

WT + vehicleHD + vehicleHD + BexHD + KD3010

A

Basal OCR

0

50

100

0

100

200Max OCR

OC

R (

pm

ol/m

in)

OC

R (

pm

ol/m

in)

B C

0

1

2

3

4

5

Veh Veh Bex KD3010

WT HD

OC

R/E

CA

R

Veh Veh Bex KD3010

WT HD

Veh Veh Bex KD3010

WT HD

D

****

*

0

1

2

3

A l3 A l1 HK4 HK2 Pck1 Pcx

OXPHOS Glycolysis Gluconeogenesis

Rel

ativ

e ex

pre

ssio

n

E

WT + vehicleHD + vehicleHD + bexarotene

**

**

*

* ****

Fig. 5. PPARd agonist treatment reverses impaired oxidative metabolism in mouse HD neurons. (A) Mitochondrial respiratory states were assessed by extracellular fluxanalysis in primary cortical neurons obtained from BAC-HDmice (HD) and littermate controls (WT) and treated with vehicle (Veh), bexarotene (Bex; 500 nM), or KD3010 (100 nM).(B) We quantified basal oxygen consumption rate (OCR) in neurons from (A). **P < 0.01, ANOVA with post-hoc Tukey test. n = 6 to 7 samples per condition. (C) We quantifiedmaximal OCR in neurons from (A). **P < 0.01, ANOVA with post-hoc Tukey test. n = 6 to 7 samples per condition. (D) We calculated OCR/extracellular acidification rate (ECAR) inneurons from (B). *P < 0.05, ANOVA with post-hoc Tukey test. n = 6 to 7 samples per condition. (E) We performed RT-PCR analysis for RNA expression of the PPARd target genesAscl3, Ascl1, Hk4, Hk2, Pck1, and Pcx (representative of different metabolic pathways) in the striatum of 18-week-old vehicle-treated nontransgenic control mice (WT), vehicle-treatedHDmice, and bexarotene-treated HDmice. PAL, palmitate; Olig, oligomycin; FCCP, carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone; A/R, Antimycin A/Rotenone B.*P < 0.05, **P < 0.01, ANOVA with post-hoc Tukey test. n = 3 mice per group. Error bars represent SEM.

7 of 13




http://stm.sciencem

ag.org/D

ownloaded from

dysfunction and impaired protein-organelle quality control. We andothers previously linkedmitochondrial dysfunction and transcriptionaldysregulation in HD to interference with the transcription coactivatorPGC-1a (6, 8, 30). Because mutant htt does not directly interact withPGC-1a to blunt its function, we pursued the mechanistic basis for thistranscription interference and identified the nuclear hormone receptorPPARd as a direct target of mutant htt neurotoxicity (10). BecausePPARd heterodimerizes with RXR to activate its target genes, and theresulting permissive PPARd-RXR heterodimer is subject to dual ligandregulation, RXR agonists are capable of promoting PPARd activation(11). Here, we examined bexarotene in various in vitro cellular models


of HD and observed robust neuroprotection, suggesting the neurother-apeutic potential of bexarotene in HD. We investigated bexarotenetreatment in HD N171-82Q mice using a study design that adheredto guidelines for rigor and reproducibility (19, 20). We documentedimprovements in motor function, htt protein aggregation, striatalneurodegeneration, and mouse survival.

Bexarotene (also called Targretin) is approved for use in humans forT cell cutaneous lymphoma but is also currently in clinical trials in ADpatients based on previous preclinical trial work in anADmousemodel(12). The mechanistic basis for the therapeutic efficacy of bexarotenewas proposed tobe increased activationof PPARg, supporting apresumed

A

0

1

2

3

4

5

Mit

och

on

dri

al le

ng

th (

m)

Empty vectorPPAR +GW501516PPAR shRNA

C B

0

10

20

30

40

50

% C

ells

wit

h f

rag

men

ted

m

ito

cho

nd

ria

0 10 1

Bexarotene (nM)

100

Q7/Q7Q111/Q111

****

***

0

1

2

3

4

5

Mit

och

on

dri

al le

ng

th (

m)

0

1

2

3

4

5

Mit

och

on

dri

al le

ng

th (

m)

Empty vector

+ PPAR /GW501516

Empty vector

Q7/Q7 Q111/Q111D E

****

** **

** **

0

1.0

1.5

2.0

0.5

Mit

och

on

dri

al D

NA

/nu

clea

r D

NA


F

**

Q7/Q7

Q111/Q111

H O2 2H O2 2 + PPAR /GW501516

H O2 2H O2 2

Fig. 6. Bexarotene activation of PPARd rescues altered mitochondrial morphology and protein quality control in HD mouse neurons. (A) Representativeimages of striatal-like cells from ST-Hdh Q7/Q7 (control) and ST-Hdh Q111/Q111 (HD) knock-in mice immunostained with Tom20 antibody are shown. Note the tubularappearance of mitochondrial network in the Q7/Q7 cells compared to fragmented appearance of mitochondrial network in the Q111/Q111 cells. Scale bars, 20 and 5 mm(inset). (B) We classified the mitochondrial network of ST-Hdh cells as tubular or fragmented as in (A) and then determined the percentage of cells with fragmented mitochon-dria. About 43% of Q111/Q111 cells contained a fragmented mitochondrial network, whereas only ~3% of Q7/Q7 cells contained a fragmented mitochondrial network. **P <0.01, Student’s t test. Treatment of Q111/Q111 cells with bexarotene reduced mitochondrial fragmentation in a dose-dependent manner. *P < 0.05, **P < 0.01, ***P < 0.001,ANOVA with post-hoc Tukey test. n = 51 to 87 cells per plate, six to nine plates per cell line. (C) We measured average mitochondrial length in ST-Hdh Q7/Q7 striatal-like cellsimmunostained with Tom20 antibody as in (A). We transfected these cells with a PPARd expression vector and then treated them with GW501516 or transfected them with aPPARd shRNA expression construct. **P < 0.01, ANOVA with post-hoc Tukey test. n = 42 to 56 cells per plate, three plates per condition. (D) We measured average mito-chondrial length in ST-Hdh Q7/Q7 striatal-like cells immunostained with Tom20 antibody as in (A). We treated these cells (with or without PPARd expression vector transfection)with hydrogen peroxide with or without GW501516 treatment. **P < 0.01, ANOVA with post-hoc Tukey test. n = 38 to 56 cells per plate, three plates per condition. (E) Wemeasured average mitochondrial length in ST-Hdh Q111/Q111 striatal-like cells immunostained with Tom20 antibody as in (A). We treated these cells (with or without PPARdexpression vector transfection) with hydrogen peroxide with or without GW501516 treatment. **P < 0.01, ANOVA with post-hoc Tukey test. n = 35 to 47 cells per replicate, threereplicates per condition. (F) We performed quantitative PCR analysis of a mitochondrial genomic amplicon and a nuclear genomic amplicon and determined the ratio of mito-chondrial DNA to nuclear DNA in the striatum of 18-week-old nontransgenic control mice, vehicle-treated HD mice, and bexarotene-treated HD mice. **P < 0.01, ANOVA withpost-hoc Tukey test. n = 9 to 12 mice per group. Error bars represent SEM.

8 of 13




http://stm.sciencem

ag.org/D

ownloaded from

0

10

20

30

40

50

% C

ells

wit

h a

gg

reg

ates

UntreatedGW501516Bexarotene

A

****

0

10

20

30

40

50

60

Spautin-1

Lactacystin

ATG7 shRNA

Bexarotene + + + + –+

+ + –

– ––––

––

–

–– –

B

% C

ells

wit

h a

gg

reg

ates **

**

**

-Actin

LC3-I LC3-II

0 1 2 3 4 5 6 7

Bafilomycin A + – – – + + Untreated PPAR

shRNAPPAR+ GW

LC

3-II/

acti

n

012345678

LC

3-II

flu

x

UntreatedPPAR shRNA

PPAR + GW501516

C

D

*

**

-Actin

LC3-I LC3-II

0 1 2 3 4 5 6 7

Bafilomycin A + – – + Untreated Bexarotene

LC

3-II/

acti

n

0

10

20

30

40

50

% C

ells

wit

h a

gg

reg

ates

UntreatedBexarotene

Bexarotene + PPAR inhibitor

** **

E

012345678

LC

3-II

flu

x

UntreatedBexarotene

**

Fig. 7. Bexarotene activation of PPARd promotes proteostasis by inducing the autophagy pathway in mouse neurons. (A) We quantified the percentage ofNeuro2a cells containing htt protein aggregates, when transfected with a htt-104Q expression vector, treated for 24 hours with GW501516 (500 nM) or bexarotene (1 mM), andexposed to H2O2 (25 mM) for 4 hours. **P < 0.01, ANOVA with post-hoc Tukey test. n = 30 to 50 cells per sample, nine samples per condition. (B) We quantified the percentageof Neuro2a cells containing htt protein aggregates, when transfected with a htt-104Q expression vector, treated for 24 hours with bexarotene (1 mM), spautin-1 (10 nM), orlactacystin (5 nM), and exposed to H2O2 (25 mM) for 4 hours. **P < 0.01, ANOVA with post-hoc Tukey test. n = 30 to 50 cells per sample, 9 to 12 samples per condition. ATG7,autophagy-related 7 gene. (C) We performed microtubule-associated protein 1A/1B-light chain 3 (LC3) immunoblot analysis of Neuro2a cells cultured in normal media in thepresence or absence of bafilomycin, transfected with a PPARd shRNA vector or a PPARd expression vector, and treated with GW501516 (100 nM). b-Actin served as a loadingcontrol. (D) We performed densitometry analysis of the LC3 immunoblotting results shown in (C) to determine autophagy flux. *P < 0.05, **P < 0.01, ANOVA with post-hocTukey test. n = 3 independent experiments. (E) We quantified the percentage of Neuro2a cells containing htt protein aggregates, when transfected with a htt-104Q expressionvector, treated with bexarotene (500 nM) alone, or bexarotene (500 nM) plus the PPARd inhibitor GSK3787 (200 nM), and exposed to H2O2 (25 mM) for 4 hours. **P < 0.01,ANOVA with post-hoc Tukey test. n = 30 to 50 cells per sample, nine samples per condition. Error bars represent SEM.

Dickey et al., Sci. Transl. Med. 9, eaal2332 (2017) 6 December 2017 9 of 13




http://stm.sciencem

ag.org/D

ownloaded from

role for enhanced Ab clearance by PPARg-expressingmicroglia (13). Be-cause PPARd is highly expressed in CNS neurons (14), and bexarotenecan also potently activate PPARd, we sought the basis for bexarotene-mediated neuroprotection in BAC-HD primary neurons by concur-rently knocking down the expression of PPARa, PPARd, or PPARg.For all tested readouts, we found that PPARd is required for amelio-ration of mutant htt neurotoxicity. Our findings thus suggested thatbexarotene neuroprotection involved PPARd activation and thatenhanced PPARd activation could be contributing to the beneficialeffects of bexarotene in AD (12, 31). Although bexarotene readilycrosses the blood-brain barrier in rodents (17, 18), it does not efficientlycross the blood-brain barrier in healthy human subjects (32), implyingthat bexarotene therapeutic responses observed in neurodegenerativedisease patients may stem from the peripheral benefits of RXR andPPARd activation or from alteration of blood-brain barrier functionin human patients (33).

An important question thatwe sought to answer in this investigationwas how PPARd achieved neuroprotection. Numerous studies haveexamined the role of PPARd in skeletal muscle, and these efforts ledto the realization that PPARd strongly favors oxidative metabolism,resulting in an increase in ATP (22, 23). Increased PPARd activationin skeletal muscle is sufficient to yield profound changes in musclephysiology and endurance exercise performance. These changes havebeen linked to altered gene expression that enhances the function ofthe tricarboxylic acid cycle and oxidative phosphorylation pathway(22, 23), preservation of glucose concentrations (34), and boosting ofPGC-1a (35). We recently surveyed the expression of the PPARs andfound that only PPARd is highly expressed in CNS neurons (10). Wefurther found that transgenic expression of a dominant-negativePPARd mutant in mice resulted in marked neurodegeneration in thecontext of diminished electron transport chain activity and greatly re-duced ATP (10). We therefore directly assayed metabolic function byperforming extracellular flux analysis on cortical neurons from BAC-HDmice and observed marked reductions in OCR. These metabolic de-fects were reversed when we treated HD neurons with bexarotene or thePPARd agonist KD3010, indicating that PPARd agonist therapy, whetherachieved with a PPARd agonist or RXR agonist, is capable of revertingHD neurons from glycolytic metabolism to oxidative metabolism.

Neurons are unique because they are postmitotic, have high energyrequirements, and are exquisitely vulnerable to misfolded protein stressand defects in organelle quality control. Misfolded proteins, or peptidefragments thereof, are a defining feature of neurodegenerative disorders,including HD, PD, and AD (36). Neurons thus require energy not onlyfor synaptic neurotransmission and transport of materials back andforth along their dendrites and axons but also for maintaining proteinand organelle quality control.We reasoned that impaired oxidativeme-tabolism in HD likely deprives neurons of the energy required to main-tainmitochondrial quality control and proteostasis, especially given thataltered mitochondrial dynamics and proteostasis are well-establishedcharacteristics of HD pathology. Evidence from patient material, cellculture models, and BAC-HDmice indicates that HD neurons containhighly fragmented mitochondria (37–39). Because excessive mitochon-drial fission occurs in HD, likely because of altered regulation of thefission regulatory proteins Drp1 and Fis1 (38, 40, 41), dysregulationof mitochondrial dynamics in HD implies that maintenance of normalmitochondrial morphology inHD requires an even greater expenditureof energy. Support for this view comes from ultrastructural analysis ofmice that recapitulate HD phenotypes upon expressing dominant-negative PPARd in striatal neurons, whose mitochondria appear highly


fragmented (10). After confirming that excessive mitochondrial frag-mentation occurs in ST-Hdh Q111/Q111 mouse striatal-like cells, wetreated Q111/Q111 cells with bexarotene and observed reductions inmitochondrial fragmentation. We further found that improved energyproduction can counter mitochondrial fragmentation induced by oxi-dative stress in normal striatal-like cells, suggesting that PPARd activa-tion may be capable of supporting mitochondrial quality control indifferent stress situations.

In our bexaroteneHDmouse study and in a previous study of PGC-1a overexpression (9), these interventions yielded reductions in httprotein aggregation in the brains of HD mice. To determine howbexarotene activation of PPARd promotes proteostasis, we testedthe effect of proteasome inhibition or autophagy inhibition on thebexarotene-mediated reduction of htt aggregates and found that auto-phagy inhibition sharply countered htt aggregate reduction in bexarotene-treated cells. We documentedmarkedly increased autophagy flux whenwe transfected cells with PPARd in the presence of an agonist but noteddiminished autophagy flux when we performed PPARd knockdown.Hence, our results indicate that PPARd activation can up-regulate auto-phagy. This finding agrees with previous work where PPARd agonisttreatment of cardiomyocytes yielded increased light chain 3 (LC3)–II,suggestive of autophagy induction, and where an analysis of PPARdknockout mice revealed reductions in autophagy markers in the heart(42). Because PGC-1a can promote increased expression of TFEB (9), amaster regulator of autophagy, we evaluated the effect of PPARdmod-ulation in TFEB knockout cells and observed that PPARd activationyielded increased autophagy flux in the absence of TFEB, indicating thatPPARd up-regulation of autophagy is not TFEB-dependent.

A frequent observation in neurodegenerative diseases is the failureof mitochondrial energy production to keep up with CNS demandfor ATP, coupled with an inability to maintain protein quality controland organellar homeostasis. Energy production and quality controlfunction are inextricably linked because neurons require energy forproteostasis and organelle quality control, and protein and organellequality control must operate efficiently if a neuron is to retain acomplement of fully functional mitochondria to carry out the taskof ATP generation. If one arm of this homeostasis loop is disrupted,the other arm of the loop will inevitably become dysfunctional, andworsening dyshomeostasis will ensue because the altered process willact as a positive feedback loop. InHD, we and others have discovereda central role for impaired mitochondrial energy production andquality control (6, 9, 30), and we have homed in on PPARd as a reg-ulatory factor capable of promoting oxidative metabolism to yieldenergy. Because energy is necessary for promoting mitochondrialquality control at the level of mitochondrial dynamics, we tested whetherPPARd activation could rescuemitochondrial fragmentation inHD andfound that PPARd activation prevented mitochondrial fragmentationin HD cells. In HD, as in other neurodegenerative diseases, there isselective vulnerability of certain types of neurons. In the case of HD,it is striatal medium spiny neurons that are preferentially lost. Althoughthe basis for this selective vulnerability is unknown, certain studies havefound that striatal neurons are prone to mitochondrial fragmentationdue to increased expression of the Drp1 receptor Fis1 (43), a pro-fissioneffect of dopaminergic signaling (40), or increased S-nitrosylation ofDrp-1 (41). Thus, inHD, the energy demands ofmedium spiny neurons,together with a predisposition to fragmentation of the mitochondrialnetwork, may explain why PPARd transcription interference contributestoHDpathogenesis andwhy PPARd promotion ofmitochondrial fusionin HD is neuroprotective.

10 of 13



http://D

ownloaded from

Here, we considered an alternate approach for PPARd activationbased on its formation of permissive heterodimers with RXR and foundthat bexarotene treatment countersmutant htt toxicity both in vitro andin vivo in HD. Our findings thus indicate that bexarotene deserves con-sideration as a potential therapy for HD. However, because this studyonly evaluated bexarotene in an HDmouse model featuring expressionof truncated protein, further study of bexarotene in an HD mousemodel featuring expression of full-length htt protein is warranted. Fur-thermore, although HD mouse models recapitulate many aspects ofthe human disease, their predictive value for gauging the potentialutility of a therapeutic intervention remains uncertain. Performingparallel studies in cell culture, mouse primary neurons, and humanstem cell models is necessary to corroborate evidence for an agent’sneuroprotection. Although this strategy was used here, it is importantto recognize that these systems also have their limitations. Althoughbexarotene is approved for use in humans, its use is associated withside effects that can be dose-limiting (44). We have previously shownthat the PPARd agonist KD3010 is capable of robust neuroprotectionin HD (10). Although KD3010 was found to be safe in humans in aphase 1b clinical trial, its dosage range for chronic, long-term use has yetto be established. Because bexarotene and KD3010 act on differenttranscription factors, one appealing approach would be to usecombinatorial therapy in which the dosages of each compound couldbe reduced to limit their respective side effects while achieving an addi-tive or perhaps even synergistic treatment response in HD patients.


stm.sciencem

ag.org/

MATERIALS AND METHODSStudy designThe primary objective of this study was to determine whether the RXRagonist bexarotene was an effective treatment for HD. Bexarotene wasevaluated in primary mouse neurons and human patient stem cell–derived neurons in experiments that used quantitative real-time PCRand assays of mitochondrial function and cell death. Bexarotene wastested in a preclinical trial in a mouse model of HD, with motorfunction, neuropathology, and survival as the outcome measures. Thepreclinical trial, which was approved by and performed in accordancewith the University of California, San Diego (UCSD) InstitutionalAnimal Care and Use Committee (IACUC), adhered to a protocolwhere we arbitrarily divided littermates and balanced genders betweenexperimental groups, with behavioral testing performedby investigatorsblinded to the treatment group of themice.Webased our group sizes onpower analysis to achieve 80% likelihood of detection of a 30% rescue ofmotor phenotypes. In the second half of this study, we sought themech-anisms by which PPARd agonists achieved neuroprotection in our HDmouse model using experiments that used quantitative real-time PCR,in vitro assays, immunohistochemistry, and Western blotting. For allexperiments, replicate numbers are stated in the figure legends.

Cell culture and primary neuron studiesST-Hdh cells were cultured as described previously (45). Primary cor-tical neurons from BAC-HD andWTmice were prepared as describedpreviously (10, 46). Cotransfection with indicated constructs [as de-scribed previously (10)] was done with Lipofectamine 3000 as per themanufacturer’s protocol (Invitrogen). Lentiviral transduction wasused to induce gene expression or knockdown in primary neurons,with infection achieved by adding 1 × 107 titer units of lentivirus tothe culture media. For reporter assays, cells were drug-treated 24 hoursafter reporter transfection, harvested 24 hours later, and subjected to


analysis using the Dual-Luciferase Reporter Assay System (Promega).Mitochondrial membrane potential was measured via live cell loadingwith a potential-sensitive dye, either JC-1 or tetramethylrhodaminemethyl ester (TMRM), using the Tecan M200 PRO Reader. JC-1 wasused for preliminary tests, followed by more extensive mitochondrialcharacterization by Seahorse metabolism analysis. Analysis of celldeath with immunofluorescence to activated caspase-3 was per-formed as described (47). H2O2 treatment was 25 mM for 4 hours.Glucose starvation was done for 2 hours in Dulbecco’s modified Eagle’smediumwithout glucose and10%dialyzed fetal bovine serum.Treatmentswith PPARd antagonist GSK3787 (GSK), lactacystin (L6785, Sigma-Aldrich), spautin-1 (SML0440, Sigma-Aldrich), and bafilomycinA1 (B1793, Sigma-Aldrich) were as indicated for individual experiments.In all experiments, the investigator was blinded to culture conditions andcell treatments.

Primary mouse cortical neurons were prepared from CD1 mice asdescribed previously (48). Cotransfectionwas performed at days in vitro5 with the N586 fragment of Htt containing either 22Q (N586-Htt22Q) or 82Q (N586-Htt 82Q) and enhanced green fluorescent pro-tein (GFP) (10:1 ratio) with Lipofectamine 2000 according to themanufacturer’s protocol. Bexarotene treatment (1 mM) was done atthe time of transfection. After 48 hours of expression, cells were fixedwith 4% paraformaldehyde (PFA) for 30 min, and nuclei were stainedwith Hoechst 33258 (bis-benzimide, Sigma-Aldrich). Image acquisi-tion was done using the AxioVision imaging software on an Axiovert100 inverted microscope (Carl Zeiss). Analysis and quantificationwere performed using Volocity (PerkinElmer). Nuclear staining in-tensity of GFP-positive cells was measured, and neurons with a nu-clear intensity of up to 200% of the intensity of healthy control wereconsidered viable.

Primary rat cortical and striatal neurons were isolated from E18Sprague-Dawley rat brains as described previously (49). Briefly, Nt-90-8Q and Nt-90-73Q constructs were transfected into separately isolatedcortical and striatal neurons using electroporation. Cortical and striatalneurons were also cotransfected with yellow fluorescent protein (YFP)or mCherry, respectively, as separate viability markers, and then co-cultured on previously established glial cell beds for 5 days beforeautomated counting of YFP or mCherry neurons (49). Total numberscounted for fluorescently transfected neurons for each condition werenormalized to Htt Nt-90-8Q–transfected neurons at baseline, withsurvival arbitrarily set to 1.

Animal studies and preclinical trialAll animal experimentation adhered to National Institutes of Health(NIH) guidelines and was approved by and performed in accordancewith the UCSD IACUC. Cohort sizes were designated on the basis ofpower analysis for threshold effects of at least 25% difference. Aftergenotyping, we performed motor baseline assessment before groupassignments and divided littermates and balanced genders betweenexperimental groups, in accordance with guidelines intended to avoidspurious results (19, 20). After group assignment, we initiatedMonday-Wednesday-Friday intraperitoneal injections of bexarotene (10 mg/kgper day) as a suspension in corn oil (3mg/ml) at 6 weeks of age. Blindedobservers visually inspected mice for obvious neurological signs,examined mice with a composite neurological evaluation tool as de-scribed previously (21), and also examined motor phenotypes byperforming rotarod testing as described previously (10). For neuro-pathology experiments, brains were harvested, and histopathology,volume measurements, and stereology analysis were performed as

11 of 13




http://stm.sciencem

ag.org/D

ownloaded from

described previously (10). In all cases, the scorer was blinded to thegenotype status and treatment condition of the mice.

Real-time RT-PCR analysisRNA samples were isolated using TRIzol (Life Technologies). GenomicDNAwas removed using RNase-free DNase (Ambion). mRNA quan-tification was performed using the 7500 Real-Time PCR System (Ap-plied Biosystems) with Applied Biosystems Assays-on-Demandprimers and TaqMan-based probes (50) or using the SYBR Greensystem (51). Applied Biosystems TaqMan primer and probe set des-ignations are available upon request. 18S or b-actin RNAwas used asinternal controls. Relative expression levels were calculated via theDDCt method.

Western blot analysisProteins were run on 10% bis-tris gels (Invitrogen) and transferredto polyvinylidene difluoride membranes (Millipore) before blockingin Odyssey Blocking Buffer (LI-COR Biosciences). Membranes wereincubated with antibodies as indicated: LC3 (NB100-2220, Fisher);TFEB (4240S, Cell Signaling), p-S6K (9234, Cell Signaling), and p-S6(2215, Cell Signaling); or b-actin (ab8226, Abcam); and imaged onthe Odyssey System (LI-COR Biosciences).

Mitochondrial studiesTheOCR and ECARof primary neurons grown in Seahorse plates weremeasured using an Extracellular Flux Analyzer (Seahorse Bioscience),following the manufacturer’s instructions. The Seahorse values werenormalized by protein mass, which was determined by bicinchoninicacid protein assay (Thermo Fisher Scientific) after the measurement.

ST-Hdh cells were transfected with the indicated constructs [aspreviously described (10)] with Lipofectamine 3000 as per themanufac-turer’s protocol (Invitrogen) and treated with compounds as indicated:H2O2 (25 mM, 4 hours); GW501516was at 100 nM for 24 hours. Cellswere fixed with 4% PFA and stained for translocase of outer mitochon-drial membrane 20 (TOM20) to delineatemitochondria. Cells were im-aged at 63× on a Zeiss 780 confocal microscope. Images were analyzedwith the NIH ImageJ program using a written script from Dickey andStrack (52).

Statistical analysisAll datawere prepared for analysis with a standard spreadsheet software(Microsoft Excel). Statistical analysis was done using Microsoft Excel,Prism 4.0 (GraphPad), or the VassarStats website (http://vassarstats.net/). For analysis of variance (ANOVA), if statistical significance (P <0.05) was achieved, we performed post hoc analysis to account formultiple comparisons. All t tests were two-tailed unless otherwiseindicated, and the level of significance (a) was always set at 0.05.

SUPPLEMENTARY MATERIALSwww.sciencetranslationalmedicine.org/cgi/content/full/9/419/eaal2332/DC1Fig. S1. Bexarotene promotes neuroprotection by activating PPARd.Fig. S2. Bexarotene pharmacodynamics analysis of PPARd target gene activation in CNS yieldsa suitable dosage and delivery scheme for preclinical trial testing.Fig. S3. Bexarotene treatment of HD mice ameliorates motor function decline.Fig. S4. Bexarotene promotes PPARd activation of target genes in skeletal muscle after 1 weekof treatment.Fig. S5. PPARd activation enhances oxidative function in neurons and restores an oxidativegene expression pattern in the CNS of HD mice.Fig. S6. Bexarotene-mediated htt protein aggregate reduction is RXR-dependent.Fig. S7. PPARd activation of autophagy does not require TFEB.


REFERENCES AND NOTES1. M. A. Nance; US Huntington Disease Genetic Testing Group, Genetic testing of children at

risk for Huntington’s disease. Neurology 49, 1048–1053 (1997).2. R. L. Albin, A. Reiner, K. D. Anderson, L. S. Dure IV, B. Handelin, R. Balfour, W. O. Whetsell Jr.,

J. B. Penney, A. B. Young, Preferential loss of striato-external pallidal projection neurons inpresymptomatic Huntington’s disease. Ann. Neurol. 31, 425–430 (1992).

3. A. Reiner, R. L. Albin, K. D. Anderson, C. J. D’Amato, J. B. Penney, A. B. Young, Differentialloss of striatal projection neurons in Huntington disease. Proc. Natl. Acad. Sci. U.S.A. 85,5733–5737 (1988).

4. M. E. MacDonald, C. M. Ambrose, M. P. Duyao, R. H. Myers, C. Lin, L. Srinidhi, G. Barnes,S. A. Taylor, M. James, N. Groot, H. MacFarlane, B. Jenkins, M. A. Anderson, N. S. Wexler,J. F. Gusella, G. P. Bates, S. Baxendale, H. Hummerich, S. Kirby, M. North, S. Youngman,R. Mott, G. Zehetner, Z. Sedlacek, A. Poustka, A.-M. Frischauf, H. Lehrach, A novel genecontaining a trinucleotide repeat that is expanded and unstable on Huntington’s diseasechromosomes. Cell 72, 971–983 (1993).

5. A. R. La Spada, J. P. Taylor, Repeat expansion disease: Progress and puzzles in diseasepathogenesis. Nat. Rev. Genet. 11, 247–258 (2010).

6. L. Cui, H. Jeong, F. Borovecki, C. N. Parkhurst, N. Tanese, D. Krainc, Transcriptionalrepression of PGC-1a by mutant huntingtin leads to mitochondrial dysfunction andneurodegeneration. Cell 127, 59–69 (2006).

7. J. Lin, P.-H. Wu, P. T. Tarr, K. S. Lindenberg, J. St-Pierre, C.-y. Zhang, V. K. Mootha, S. Jäger,C. R. Vianna, R. M. Reznick, L. Cui, M. Manieri, M. X. Donovan, Z. Wu, M. P. Cooper,M. C. Fan, L. M. Rohas, A. M. Zavacki, S. Cinti, G. I. Shulman, B. B. Lowell, D. Krainc,B. M. Spiegelman, Defects in adaptive energy metabolism with CNS-linked hyperactivityin PGC-1a null mice. Cell 119, 121–135 (2004).

8. P. Weydt, V. V. Pineda, A. E. Torrence, R. T. Libby, T. F. Satterfield, E. R. Lazarowski,M. L. Gilbert, G. J. Morton, T. K. Bammler, A. D. Strand, L. Cui, R. P. Beyer, C. N. Easley,A. C. Smith, D. Krainc, S. Luquet, I. R. Sweet, M. W. Schwartz, A. R. La Spada,Thermoregulatory and metabolic defects in Huntington’s disease transgenic mice implicatePGC-1a in Huntington’s disease neurodegeneration. Cell Metab. 4, 349–362 (2006).

9. T. Tsunemi, T. D. Ashe, B. E. Morrison, K. R. Soriano, J. Au, R. A. Vázquez Roque,E. R. Lazarowski, V. A. Damian, E. Masliah, A. R. La Spada, PGC-1a rescues Huntington’sdisease proteotoxicity by preventing oxidative stress and promoting TFEB function.Sci. Transl. Med. 4, 142ra97 (2012).

10. A. S. Dickey, V. V. Pineda, T. Tsunemi, P. P. Liu, H. C. Miranda, S. K. Gilmore-Hall, N. Lomas,K. R. Sampat, A. Buttgereit, M.-J. M. Torres, A. L. Flores, M. Arreola, N. Arbez, S. S. Akimov,T. Gaasterland, E. R. Lazarowski, C. A. Ross, G. W. Yeo, B. L. Sopher, G. K. Magnuson,A. B. Pinkerton, E. Masliah, A. R. La Spada, PPAR-d is repressed in Huntington’s disease,is required for normal neuronal function and can be targeted therapeutically. Nat. Med.22, 37–45 (2016).

11. R. M. Evans, D. J. Mangelsdorf, Nuclear receptors, RXR, and the Big Bang. Cell 157,255–266 (2014).

12. P. E. Cramer, J. R. Cirrito, D. W. Wesson, C. Y. D. Lee, J. C. Karlo, A. E. Zinn, B. T. Casali,J. L. Restivo, W. D. Goebel, M. J. James, K. R. Brunden, D. A. Wilson, G. E. Landreth,ApoE-directed therapeutics rapidly clear b-amyloid and reverse deficits in AD mousemodels. Science 335, 1503–1506 (2012).

13. S. Mandrekar-Colucci, G. E. Landreth, Nuclear receptors as therapeutic targets forAlzheimer’s disease. Expert Opin. Ther. Targets 15, 1085–1097 (2011).

14. E. E. Girroir, H. E. Hollingshead, P. He, B. Zhu, G. H. Perdew, J. M. Peters, Quantitativeexpression patterns of peroxisome proliferator-activated receptor-b/d (PPARb/d) proteinin mice. Biochem. Biophys. Res. Commun. 371, 456–461 (2008).

15. T. Malm, M. Mariani, L. J. Donovan, L. Neilson, G. E. Landreth, Activation of the nuclearreceptor PPARd is neuroprotective in a transgenic mouse model of Alzheimer’s diseasethrough inhibition of inflammation. J. Neuroinflammation 12, 7 (2015).

16. G. Schilling, M. W. Becher, A. H. Sharp, H. A. Jinnah, K. Duan, J. A. Kotzuk, H. H. Slunt,T. Ratovitski, J. K. Cooper, N. A. Jenkins, N. G. Copeland, D. L. Price, C. A. Ross,D. R. Borchelt, Intranuclear inclusions and neuritic aggregates in transgenic miceexpressing a mutant N-terminal fragment of huntingtin. Hum. Mol. Genet. 8, 397–407 (1999).

17. I. Tesseur, A. C. Lo, A. Roberfroid, S. Dietvorst, B. Van Broeck, M. Borgers, H. Gijsen,D. Moechars, M. Mercken, J. Kemp, R. D’Hooge, B. De Strooper, Comment on“ApoE-directed therapeutics rapidly clear b-amyloid and reverse deficits in AD mousemodels”. Science 340, 924-e (2013).

18. S. Wang, P. Wen, S. Wood, Effect of LXR/RXR agonism on brain and CSF Ab40 levels inrats. F1000Res 5, 138 (2016).

19. S. C. Landis, S. G. Amara, K. Asadullah, C. P. Austin, R. Blumenstein, E. W. Bradley,R. G. Crystal, R. B. Darnell, R. J. Ferrante, H. Fillit, R. Finkelstein, M. Fisher,H. E. Gendelman, R. M. Golub, J. L. Goudreau, R. A. Gross, A. K. Gubitz, S. E. Hesterlee,D. W. Howells, J. Huguenard, K. Kelner, W. Koroshetz, D. Krainc, S. E. Lazic,M. S. Levine, M. R. Macleod, J. M. McCall, R. T. Moxley III, K. Narasimhan, L. J. Noble,S. Perrin, J. D. Porter, O. Steward, E. Unger, U. Utz, S. D. Silberberg, A call for transparentreporting to optimize the predictive value of preclinical research. Nature 490, 187–191(2012).

12 of 13

http://vassarstats.net/

http://vassarstats.net/

http://www.sciencetranslationalmedicine.org/cgi/content/full/9/419/eaal2332/DC1




http://stm.sciencem

ag.org/D

ownloaded from

20. S. Perrin, Preclinical research: Make mouse studies work. Nature 507, 423–425(2014).

21. S. J. Guyenet, S. A. Furrer, V. M. Damian, T. D. Baughan, A. R. La Spada, G. A. Garden,A simple composite phenotype scoring system for evaluating mouse models ofcerebellar ataxia. J. Visualized Exp. 2010, 1787 (2010).

22. S. Luquet, J. Lopez-Soriano, D. Holst, A. Fredenrich, J. Melki, M. Rassoulzadegan,P. A. Grimaldi, Peroxisome proliferator-activated receptor d controls muscle developmentand oxidative capability. FASEB J. 17, 2299–2301 (2003).

23. M. Schuler, F. Ali, C. Chambon, D. Duteil, J.-M. Bornert, A. Tardivel, B. Desvergne, W. Wahli,P. Chambon, D. Metzger, PGC1a expression is controlled in skeletal muscles by PPARb,whose ablation results in fiber-type switching, obesity, and type 2 diabetes. Cell Metab. 4,407–414 (2006).

24. M. T. Lin, M. F. Beal, Mitochondrial dysfunction and oxidative stress in neurodegenerativediseases. Nature 443, 787–795 (2006).

25. P. Caro, A. U. Kishan, E. Norberg, I. A. Stanley, B. Chapuy, S. B. Ficarro, K. Polak, D. Tondera,J. Gounarides, H. Yin, F. Zhou, M. R. Green, L. Chen, S. Monti, J. A. Marto, M. A. Shipp,N. N. Danial, Metabolic signatures uncover distinct targets in molecular subsets of diffuselarge B cell lymphoma. Cancer Cell 22, 547–560 (2012).

26. J. Labbadia, R. I. Morimoto, The biology of proteostasis in aging and disease. Annu. Rev.Biochem. 84, 435–464 (2015).

27. F. Trettel, D. Rigamonti, P. Hilditch-Maguire, V. C. Wheeler, A. H. Sharp, F. Persichetti,E. Cattaneo, M. E. MacDonald, Dominant phenotypes produced by the HD mutation inSTHdhQ111 striatal cells. Hum. Mol. Genet. 9, 2799–2809 (2000).

28. M. Frank, S. Duvezin-Caubet, S. Koob, A. Occhipinti, R. Jagasia, A. Petcherski,M. O. Ruonala, M. Priault, B. Salin, A. S. Reichert, Mitophagy is triggered by mild oxidativestress in a mitochondrial fission dependent manner. Biochim. Biophys. Acta 1823,2297–2310 (2012).

29. C. Settembre, C. Di Malta, V. A. Polito, M. Garcia Arencibia, F. Vetrini, S. Erdin, S. U. Erdin,T. Huynh, D. Medina, P. Colella, M. Sardiello, D. C. Rubinsztein, A. Ballabio, TFEB linksautophagy to lysosomal biogenesis. Science 332, 1429–1433 (2011).

30. R. K. Chaturvedi, P. Adhihetty, S. Shukla, T. Hennessy, N. Calingasan, L. Yang, A. Starkov,M. Kiaei, M. Cannella, J. Sassone, A. Ciammola, F. Squitieri, M. F. Beal, ImpairedPGC-1a function in muscle in Huntington’s disease. Hum. Mol. Genet. 18, 3048–3065(2009).

31. N. F. Fitz, A. A. Cronican, I. Lefterov, R. Koldamova, Comment on “ApoE-directedtherapeutics rapidly clear b-amyloid and reverse deficits in AD mouse models”. Science340, 924-c (2013).

32. K. Ghosal, M. Haa, P. B. Verghese, T. West, T. Veenstra, J. B. Braunstein, R. J. Bateman,D. M. Holtzman, G. E. Landreth, A randomized controlled study to evaluate the effect ofbexarotene on amyloid-b and apolipoprotein E metabolism in healthy subjects.Alzheimers Dement. (N.Y.) 2, 110–120 (2016).

33. M. Kuntz, P. Candela, J. Saint-Pol, Y. Lamartinière, M.-C. Boucau, E. Sevin, L. Fenart,F. Gosselet, Bexarotene promotes cholesterol efflux and restricts apical-to-basolateraltransport of amyloid-b peptides in an in vitro model of the human blood-brain barrier.J. Alzheimers Dis. 48, 849–862 (2015).

34. W. Fan, W. Waizenegger, C. S. Lin, V. Sorrentino, M.-X. He, C. E. Wall, H. Li, C. Liddle, R. T. Yu,A. R. Atkins, J. Auwerx, M. Downes, R. M. Evans, PPARd promotes running enduranceby preserving glucose. Cell Metab. 25, 1186–1193.e4 (2017).

35. J.-H. Koh, C. R. Hancock, S. Terada, K. Higashida, J. O. Holloszy, D.-H. Han, PPARb isessential for maintaining normal levels of PGC-1a and mitochondria and for the increasein muscle mitochondria induced by exercise. Cell Metab. 25, 1176–1185.e5 (2017).

36. P. Weydt, A. R. La Spada, Targeting protein aggregation in neurodegeneration—Lessonsfrom polyglutamine disorders. Expert Opin. Ther. Targets 10, 505–513 (2006).

37. U. P. Shirendeb, M. J. Calkins, M. Manczak, V. Anekonda, B. Dufour, J. L. McBride, P. Mao,P. H. Reddy, Mutant huntingtin’s interaction with mitochondrial protein Drp1 impairsmitochondrial biogenesis and causes defective axonal transport and synapticdegeneration in Huntington’s disease. Hum. Mol. Genet. 21, 406–420 (2012).

38. W. Song, J. Chen, A. Petrilli, G. Liot, E. Klinglmayr, Y. Zhou, P. Poquiz, J. Tjong, M. A. Pouladi,M. R. Hayden, E. Masliah, M. Ellisman, I. Rouiller, R. Schwarzenbacher, B. Bossy, G. Perkins,E. Bossy-Wetzel, Mutant huntingtin binds the mitochondrial fission GTPase dynamin-relatedprotein-1 and increases its enzymatic activity. Nat. Med. 17, 377–382 (2011).

39. I. Tellez-Nagel, A. B. Johnson, R. D. Terry, Studies on brain biopsies of patients withHuntington’s chorea. J. Neuropathol. Exp. Neurol. 33, 308–332 (1974).

40. M. Cherubini, M. Puigdellivol, J. Alberch, S. Ginés, Cdk5-mediated mitochondrial fission:A key player in dopaminergic toxicity in Huntington’s disease. Biochim. Biophys. Acta1852, 2145–2160 (2015).

41. F. Haun, T. Nakamura, A. D. Shiu, D.-H. Cho, T. Tsunemi, E. A. Holland, A. R. La Spada,S. A. Lipton, S-nitrosylation of dynamin-related protein 1 mediates mutant huntingtin-


induced mitochondrial fragmentation and neuronal injury in Huntington’s disease.Antioxid. Redox Signal. 19, 1173–1184 (2013).

42. X. Palomer, E. Capdevila-Busquets, G. Botteri, L. Salvadó, E. Barroso, M. M. Davidson,L. Michalik, W. Wahli, M. Vázquez-Carrera, PPARb/d attenuates palmitate-inducedendoplasmic reticulum stress and induces autophagic markers in human cardiac cells.Int. J. Cardiol. 174, 110–118 (2014).

43. P. Guedes-Dias, J. de Proenca, T. R. Soares, A. Leitão-Rocha, B. R. Pinho, M. R. Duchen,J. M. A. Oliveira, HDAC6 inhibition induces mitochondrial fusion, autophagic flux andreduces diffuse mutant huntingtin in striatal neurons. Biochim. Biophys. Acta 1852,2484–2493 (2015).

44. L. Väkevä, A. Ranki, S. Hahtola, Ten-year experience of bexarotene therapy for cutaneousT-cell lymphoma in Finland. Acta Derm. Venereol. 92, 258–263 (2012).

45. S. Gines, I. S. Seong, E. Fossale, E. Ivanova, F. Trettel, J. F. Gusella, V. C. Wheeler,F. Persichetti, M. E. MacDonald, Specific progressive cAMP reduction implicates energydeficit in presymptomatic Huntington’s disease knock-in mice. Hum. Mol. Genet. 12,497–508 (2003).

46. J. E. Young, R. A. Martinez, A. R. La Spada, Nutrient deprivation induces neuronalautophagy and implicates reduced insulin signaling in neuroprotective autophagyactivation. J. Biol. Chem. 284, 2363–2373 (2009).

47. J. E. Young, G. A. Garden, R. A. Martinez, F. Tanaka, C. M. Sandoval, A. C. Smith,B. L. Sopher, A. Lin, K. H. Fischbeck, L. M. Ellerby, R. S. Morrison, J. P. Taylor, A. R. La Spada,Polyglutamine-expanded androgen receptor truncation fragments activate aBax-dependent apoptotic cascade mediated by DP5/Hrk. J. Neurosci. 29, 1987–1997(2009).

48. E. E. Watkin, N. Arbez, E. Waldron-Roby, R. O’Meally, T. Ratovitski, R. N. Cole, C. A. Ross,Phosphorylation of mutant huntingtin at serine 116 modulates neuronal toxicity. PLOS ONE9, e88284 (2014).

49. L. S. Kaltenbach, M. M. Bolton, B. Shah, P. M. Kanju, G. M. Lewis, G. J. Turmel, J. C. Whaley,O. J. Trask Jr., D. C. Lo, Composite primary neuronal high-content screening assay forHuntington’s disease incorporating non-cell-autonomous interactions. J. Biomol. Screen.15, 806–819 (2010).

50. K. J. Livak, S. J. A. Flood, J. Marmaro, W. Giusti, K. Deetz, Oligonucleotides with fluorescentdyes at opposite ends provide a quenched probe system useful for detecting PCRproduct and nucleic acid hybridization. PCR Methods Appl. 4, 357–362 (1995).

51. S. A. Bustin, Quantification of mRNA using real-time reverse transcription PCR (RT-PCR):Trends and problems. J. Mol. Endocrinol. 29, 23–39 (2002).

52. A. S. Dickey, S. Strack, PKA/AKAP1 and PP2A/Bb2 regulate neuronal morphogenesis via Drp1phosphorylation and mitochondrial bioenergetics. J. Neurosci. 31, 15716–15726 (2011).

Acknowledgments: We thank E. Lopez for technical support and Y. Matsuoka for providingcompound GSK3787. Funding: This work was supported by the Hereditary DiseaseFoundation, the Cure Huntington’s Disease Initiative, and grants from the NIH (R01 NS065874and R01 AG033082 to A.R.L.S. and National Research Service Award F32 NS081964 to A.S.D.).R.M.E. is an Investigator of the Howard Hughes Medical Institute at the Salk Institute forBiological Studies and March of Dimes Chair in Molecular and Developmental Biology. Authorcontributions: A.S.D. and A.R.L.S. provided the conceptual framework for the study. A.S.D.and A.R.L.S. designed all the experiments, with assistance from W.F. and R.M.E. for extracellularflux analysis, D.C.L. for the cortico-striatal neuron coculture experiments, C.A.R. for the humanstem cell–derived neuron studies, and E.M. for EM48 immunohistochemistry and neuropathologyanalysis. A.S.D., D.N.S., M.A., K.R.S., N.A., S.A., M.J.V.K., K.O., S.K.G.-H., A.L.F., J.M.N., N.L., C.L.H.,and D.C.L. performed the experiments. A.S.D. and A.R.L.S. wrote the manuscript. Competinginterests: The authors declare that they have no competing interests. Data and materialsavailability: All data needed to evaluate the conclusions in the study are present in the paperand the Supplementary Materials. Requests for HD stem cell materials, available through amaterials transfer agreement, should be directed to A.R.L.S.

Submitted 18 October 2016Accepted 9 August 2017Published 6 December 201710.1126/scitranslmed.aal2332

Citation: A. S. Dickey, D. N. Sanchez, M. Arreola, K. R. Sampat, W. Fan, N. Arbez, S. Akimov,M. J. Van Kanegan, K. Ohnishi, S. K. Gilmore-Hall, A. L. Flores, J. M. Nguyen, N. Lomas, C. L. Hsu,D. C. Lo, C. A. Ross, E. Masliah, R. M. Evans, A. R. La Spada, PPARd activation by bexarotenepromotes neuroprotection by restoring bioenergetic and quality control homeostasis. Sci.Transl. Med. 9, eaal2332 (2017).

13 of 13


quality control homeostasis activation by bexarotene promotes neuroprotection by restoring bioenergetic andδPPAR

Cynthia L. Hsu, Donald C. Lo, Christopher A. Ross, Eliezer Masliah, Ronald M. Evans and Albert R. La SpadaMichael J. Van Kanegan, Kohta Ohnishi, Stephen K. Gilmore-Hall, April L. Flores, Janice M. Nguyen, Nicole Lomas, Audrey S. Dickey, Dafne N. Sanchez, Martin Arreola, Kunal R. Sampat, Weiwei Fan, Nicolas Arbez, Sergey Akimov,

DOI: 10.1126/scitranslmed.aal2332, eaal2332.9Sci Transl Med

activating autophagy.enhanced oxidative metabolism, promoted mitochondrial quality control, and boosted protein homeostasis by

agonistsδ's neuroprotective effect and found that treatment with RXR/PPARδthen examined the basis for PPARderived neurons, and the BAC-HD mouse model. The authors−ouse primary neurons, human HD patient stem cell

HD and in an HD mouse model. They determined that bexarotene was effective at countering HD neurotoxicity in m. evaluated the RXR agonist bexarotene in cellular models ofet alHuntington's disease (HD). In new work, Dickey

target genes contributes to neurodegeneration inδPPARtarget genes. Interference with transcription of is a permissive nuclear receptor that heterodimerizes with the retinoid X receptor (RXR) to activateδPPAR

Defeating neurotoxicity with a repurposed drug

ARTICLE TOOLS http://stm.sciencemag.org/content/9/419/eaal2332

MATERIALSSUPPLEMENTARY http://stm.sciencemag.org/content/suppl/2017/12/04/9.419.eaal2332.DC1

CONTENTRELATED

http://stm.sciencemag.org/content/scitransmed/11/480/eaan0457.fullhttp://stm.sciencemag.org/content/scitransmed/10/461/eaar3959.fullhttp://stke.sciencemag.org/content/sigtrans/11/536/eaam5855.full

REFERENCES

http://stm.sciencemag.org/content/9/419/eaal2332#BIBLThis article cites 52 articles, 10 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

registered trademark of AAAS. is aScience Translational MedicineScience, 1200 New York Avenue NW, Washington, DC 20005. The title

(ISSN 1946-6242) is published by the American Association for the Advancement ofScience Translational Medicine

of Science. No claim to original U.S. Government WorksCopyright © 2017 The Authors, some rights reserved; exclusive licensee American Association for the Advancement


http://stm.sciencem

ag.org/D

ownloaded from

http://stm.sciencemag.org/content/9/419/eaal2332

http://stm.sciencemag.org/content/suppl/2017/12/04/9.419.eaal2332.DC1

http://stke.sciencemag.org/content/sigtrans/11/536/eaam5855.full

http://stm.sciencemag.org/content/scitransmed/10/461/eaar3959.full

http://stm.sciencemag.org/content/scitransmed/11/480/eaan0457.full

http://stm.sciencemag.org/content/9/419/eaal2332#BIBL

http://www.sciencemag.org/help/reprints-and-permissions

http://www.sciencemag.org/about/terms-service


d activation by bexarotene promotes neuroprotection by ...for 3 days/week for 1 week, to 6-week-old...

Documents