Brain-Derived Neurotrophic Factor Induces Excitotoxic Sensitivity in Cultured Embryonic Rat Spinal Motor Neurons Through Activation of the Phosphatidylinositol 3-Kinase Pathway

Download Brain-Derived Neurotrophic Factor Induces Excitotoxic Sensitivity in Cultured Embryonic Rat Spinal Motor Neurons Through Activation of the Phosphatidylinositol 3-Kinase Pathway

Post on 06-Jul-2016




0 download

Embed Size (px)


  • Brain-Derived Neurotrophic Factor Induces ExcitotoxicSensitivity in Cultured Embryonic Rat Spinal Motor Neurons

    Through Activation of the Phosphatidylinositol3-Kinase Pathway

    *Hugh J. L. Fryer, *Daniel H. Wolf, Ronald J. Knox, *Stephen M. Strittmatter,Diane Pennica, Rhona M. OLeary, *David S. Russell, and *Robert G. Kalb

    Departments of *Neurology and Pharmacology and Section of Neurobiology, Yale University School of Medicine, New Haven,Connecticut, and Molecular Oncology Department, Genentech, Inc., South San Francisco, California, U.S.A.

    Abstract: Neurotrophic factors (NTFs) can protectagainst or sensitize neurons to excitotoxicity. We studiedthe role played by various NTFs in the excitotoxic deathof purified embryonic rat motor neurons. Motor neuronscultured in brain-derived neurotrophic factor, but notneurotrophin 3, glial-derived neurotrophic factor, or car-diotrophin 1, were sensitive to excitotoxic insult. BDNFalso induces excitotoxic sensitivity (ES) in motor neuronswhen BDNF is combined with these other NTFs. Theeffect of BDNF depends on de novo protein and mRNAsynthesis. Reagents that either activate or inhibit the75-kDa NTF receptor p75NTR do not affect BDNF-in-duced ES. The low EC50 for BDNF-induced survival andES suggests that TrkB mediates both of these biologicalactivities. BDNF does not alter glutamate-evoked rises ofintracellular Ca21, suggesting BDNF acts downstream.Both wortmannin and LY294002, which specifically blockthe phosphatidylinositol 3-kinase (PI3K) intracellular sig-naling pathway in motor neurons, inhibit BDNF-inducedES. We confirm this finding using a herpes simplex virus(HSV) that expresses the dominant negative p85 subunitof PI3K. Infecting motor neurons with this HSV, but not acontrol HSV, blocks activation of the PI3K pathway andBDNF-induced ES. Through the activation of TrkB andthe PI3K signaling pathway, BDNF renders developingmotor neurons susceptible to glutamate receptor-medi-ated cell death. Key Words: Ionotropic glutamate recep-torsMotor neuronsExcitotoxicityPhosphatidylino-sitol 3-kinaseTrkBLow-affinity neurotrophin receptor.J. Neurochem. 74, 582595 (2000).

    The death of neurons throughout the CNS can beinduced by the prolonged activation of ionotropic gluta-mate receptors (IGRs). This phenomenon, known as ex-citotoxicity, has been implicated in the death of neuronsin vivo in a variety of pathological conditions includingoxygenglucose deprivation, seizures, trauma, and neu-rodegenerative disease (Choi, 1988, 1990; Lees, 1993;Lipton and Rosenberg, 1994; Shaw, 1994). During de-

    velopment, excitotoxicity may contribute to the death ofneurons during naturally occurring neuronal death peri-ods (Caldero et al., 1997; Solum et al., 1997). In vitromodels of excitotoxicity have shown that the excitotoxicdeath of neurons is caused in part by sustained patho-logical rises in intracellular Ca21 concentration ([Ca21]i)(Randall and Thayer, 1992; Frandsen and Schousboe,1993; Hartley et al., 1993; Harman and Maxwell, 1995;Hyrc et al., 1997).

    Neurotrophic factors (NTFs) are required for the sur-vival of developing and adult CNS neurons (Oppenheim,1989; Davies, 1994). Knowledge of the biological ac-tions of NTFs has led to the hypothesis that this class ofgrowth factors might aid in the prevention of neuronaldeath induced by a variety of noxious insults (Hefti,1994; Lindsay, 1994). In fact, a variety of NTFs abrogateIGR-induced death of CNS neurons in vitro and in vivo(Lindholm, 1994; Tatter et al., 1995). Data from in vitrosystems suggest that NTFs may protect neurons by (1)reducing the extent of the sustained rise of [Ca21]i in-duced by chronic IGR activation, (2) reducing cell sur-face-expressed IGRs, (3) inducing proteins that bufferrises of [Ca21]i, or (4) reducing the accumulation ofintracellular superoxides caused by the sustained rise of

    Received August 3, 1999; revised manuscript received September23, 1999; accepted September 28, 1999.

    Address correspondence and reprint requests to Dr. R. G. Kalb atDepartment of Neurology, Yale University School of Medicine, P.O.Box 208018, 333 Cedar St., New Haven, CT 06520-8018, U.S.A.E-mail:

    Abbreviations used: BDNF, brain-derived neurotrophic factor;[Ca21]i, intracellular calcium concentration; CT-1, cardiotrophin 1;ERK, extracellular signal-regulated kinase; ES, excitotoxic sensitivity;GDNF, glial-derived neurotrophic factor; HSV, herpes simplex virus;IGR, ionotropic glutamate receptor; MAP, mitogen-activated protein;NGF, nerve growth factor; NT, neurotrophin; NTF, neurotrophic fac-tor; PI3K, phosphatidylinositol 3-kinase.


    Journal of NeurochemistryLippincott Williams & Wilkins, Inc., Philadelphia 2000 International Society for Neurochemistry

  • [Ca21]i (Mattson et al., 1993, 1995; Brandoli et al., 1998;Klocker et al., 1998).

    NTFs have also been shown, however, to make neu-rons more vulnerable to excitotoxicity. In an in vitrohypoxic/ischemic model, in which the death of corticalneurons is caused by excitotoxicity, incubation of thecells with brain-derived neurotrophic factor (BDNF) andneurotrophin (NT) 3 and 4/5 exacerbates neuronal death(Koh et al., 1995). NTFs potentiate excitotoxicity inculture systems of other neurons as well (Prehn, 1996;Morrison and Mason, 1998). As mature cultures of neu-rons undergo a higher percentage of excitotoxicity-in-duced death than developing neurons, it has been hy-pothesized that NTFs may make neurons more vulnera-ble to excitotoxicity by accelerating their maturation(Samdani et al., 1997). NTFs have also been shown tomake neurons more vulnerable to excitotoxicity in vivo;BDNF, for example, injected into the hippocampus ex-acerbates kainate toxicity (Rudge et al., 1998).

    As a consequence of their responsiveness to a numberof NTFs, cultured embryonic motor neurons have beenan ideal system with which to study NTF actions (Camuand Henderson, 1992; Henderson et al., 1997). In addi-tion, we have previously shown that prolonged activationof IGRs causes the death of a subset of motor neuronspurified from embryonic rats and that this effect is de-pendent on Ca21 influx (Fryer et al., 1999). In this study,the motor neurons were cultured in a mixture of BDNF,NT3, and NT4/5. Considering the contradictory rolesplayed by NTFs in other systems, we wanted to investi-gate the role that various individual NTFs play in theexcitotoxic death of embryonic spinal motor neurons. Inaddition, we wanted to determine which NTF receptorstransduce this signal and the intracellular signaling sys-tems that mediate this response.


    MaterialsTimed pregnant SpragueDawley rats were obtained from

    Charles River (Kingstown, NY, U.S.A.). Leibowitz L15 me-dium, glutamine, penicillin/streptomycin, Ca21/Mg21-freephosphate-buffered saline, horse serum, mouse laminin, andsodium bicarbonate were purchased from GibcoBRL (GrandIsland, NY, U.S.A.). All other culture reagents, glutamate,glycine, and wortmannin were purchased from Sigma (St.Louis, MO, U.S.A.). LY294002 was purchased from Calbio-chem (La Jolla, CA, U.S.A.). Dr. Eugene Johnson (WashingtonUniversity, St. Louis, MO, U.S.A.) kindly provided the hybrid-oma cell line 192 [anti-rat 75-kDa NTF receptor (p75NTR)monoclonal antibody]. Recombinant NTFs were obtained fromthe following sources: Human BDNF was obtained fromCephalon (West Chester, PA, U.S.A.), Genentech (S. San Fran-cisco, CA, U.S.A.), Regeneron (Tarrytown, NY, U.S.A.), andAlomone Labs (Jerusalem, Israel); recombinant NT3 from Re-generon and Alomone Labs; cardiotrophin 1 (CT-1) from Ge-nentech; and glial-derived neurotrophic factor (GDNF) fromAmgen (Thousand Oaks, CA, U.S.A.). For the majority of theexperiments in this study, we used BDNF obtained from Re-generon. Two different antibodies to TrkB, which gave identi-cal results, were used (Santa Cruz Biotechnology, Santa Cruz,

    CA, and Transduction Laboratories, Lexington, KY, U.S.A.).Other antibodies for immunocytochemistry and biochemistrywere obtained from the following sources: p85 (Upstate Bio-technology, Lake Placid, NY, U.S.A.); b-galactosidase (59-39,Inc., Boulder, CO, U.S.A.); phospho-Trk490 and phospho-Akt(New England BioLabs, Beverly, MA, U.S.A.); and phospho-ERK (Promega, Madison, WI, U.S.A.).Motor neuron purification, drug treatments, andquantification of cell survival

    Motor neurons from embryonic rat embryos were purified aspreviously described (Fryer et al., 1999). In brief, dissociatedventral spinal cords from the embryos of 1516-day pregnantrats were first enriched for motor neurons on a cushion ofmetrizamide. These cells were further purified by panning onplates coated with a monoclonal antibody (antibody 192) thatrecognizes the low-affinity neurotrophin receptor p75NTR,which is expressed only on motor neurons of the embryonicventral horn (Yan and Johnson, 1988). Rather than eluting theimmunopanned motor neurons with monoclonal antibody 192,as we had previously reported, the cells for these experimentswere washed from the immunopanning plates with L15 me-dium, which effectively but gently removed the cells from theplate. Purified motor neurons were diluted with L15 mediumsupplemented with 0.63 mg/ml sodium bicarbonate, 100 IU/mlpenicillin, 100 mg/ml streptomycin, 2% (vol/vol) horse serum,20 mM glucose, 5 mg/ml insulin, 0.1 mM putrescine, 20 nMprogesterone, 0.1 mg/ml conalbumin, and 30 nM sodium se-lenite and were seeded at low density (4,500 cells or 4.5cells/mm2) onto 33-mm plates (Nunc, Boston, MA, U.S.A.)that had been coated sequentially with poly-D-ornithine andmouse laminin.

    Excitotoxic sensitivity (ES) assays were performed aftermotor neurons were cultured overnight at 37C and 5% CO2.Stock solutions of glutamate and glycine were prepared inLockes buffer (see below). For ES assays, the culture mediumwas replaced with Lockes buffer (in mM: 134 NaCl, 25 KCl,2.3 CaCl2, 5 dextrose, 4 NaHCO3, and 5 HEPES, pH 7.2)containing vehicle (controls) or 200 mM glutamate 1 20 mMglycine, and the cultures were incubated at 37C in 5% CO2.After 1 h, the Lockes buffer was removed, the plates werewashed three times with fresh Lockes buffer, and the originalor fresh medium was added to the plates.

    Cell survival was quantified as follows. The initial numberof cells per plate was determined by counting cells from at leastthree nonadjacent 3 3 3-mm grids using a Nikon TMS invertedphase-contrast microscope at a magnification of 1003 fromthree to five plates 23 h after plating. Following varioustreatments (see text), cell survival was once again quantified.The percentage of initially plated cells is the survival per platedivided by the number of initially plated cells times 100. Eachdata point is the mean 6 SEM of two to four independentexperiments. Significance was determined using ANOVA withScheffes test.

    Virus studiesReplication-deficient herpes simplex viruses (HSVs) were

    used to introduce genes into cultured motor neurons. Methodsfor the generation and titering of viral stocks have been de-scribed (Neve et al., 1997). The titers of the viruses employedin this study were routinely in the range of 35 3 107 plaque-forming units/ml.

    A dominant negative form of p85 (Hara et al., 1994; Kotaniet al., 1994), the regulatory subunit of phosphatidylinositol3-kinase (PI3K), designated DNp85 (provided by Dr. M. Ka-

    J. Neurochem., Vol. 74, No. 2, 2000


  • suga, Kobe University School of Medicine, Kobe, Japan), wascloned into the amplicon vector pRCUC and used to generaterecombinant virus. HSV engineered to express LacZ served asa control.

    For use in the ES assays, motor neurons were plated inmedium containing 10 ng/ml CT-1 and cultured for 68 h. Onemicroliter of viral stock was added to cell cultured in 33-mmplates. When higher numbers of cells were plated for biochem-ical studies (see below), 5 ml of viral stock was added. Weempirically determined that this concentration of either HSVconstruct gave the highest level of infection (determined im-munocytochemically) without evident toxicity.Immunocytochemistry

    For immunocytological studies, 5,000 cells in 50 ml ofmedium were plated onto the center of 14-mm round glasscoverslips (Assistent, Germany). After a wait of 6 h for cells toattach, more medium was added to the well or dish to com-pletely cover the coverslip. Twenty-four to 48 h after the initialplating, cells were fixed with 4% (wt/vol) paraformaldehydeand 0.1% (vol/vol) glutaraldehyde in 0.1 M phosphate buffer(pH 7.4) for 1020 min at room temperature. Following severalwashes in phosphate buffer, cells were blocked and permeabil-ized in antibody dilution buffer [Dulbeccos modified Eaglesmedium containing 5% (vol/vol) fetal calf serum and 0.2%(wt/vol) sodium azide] with 0.2% (vol/vol) Triton X-100 for1020 min at room temperature. Coverglasses were incubatedat 4C for 13 days with antibodies diluted in antibody dilutionbuffer. Primary antibody was visualized using species-specificfluorescein isothiocyanate- or horseradish peroxidase-conju-gated secondary antibodies (Jackson ImmunoResearch Labora-tories, West Grove, PA, U.S.A.). Stained cells were mountedon glass slides in Vectashield (Vector Labs, Burlingame, CA,U.S.A.) and photographed on a Zeiss Axioscope.Single-cell [Ca21]i imaging

    Single-cell fluorescence ratio [Ca21]i imaging from purifiedembryonic motor neurons was performed as previously de-scribed (Fryer et al., 1999). Neurons that were cultured ontoglass coverslips were loaded with the acetoxymethyl esterCa21 indicator dye fura PE3 (TefLabs, Austin, TX, U.S.A.).Fura PE3 fluorescence was measured using a Nikon Diaphotequipped with a 403 Nikon Plan Fluor objective (NA 5 1.3)and recorded with a Hammamatsu C2400 iCCD camera. Exci-tation (345 and 380 nm) illumination of fura PE3 was per-formed using a 75-W xenon arc lamp and a PC-controlledmonochromator coupled to the microscope by fiberoptic cable(Photon Technology Int., South Brunswick, NJ, U.S.A.).[Ca21]i concentrations were calculated from ratioed images asdescribed previously (Fryer et al., 1999).Biochemistry

    Approximately 50,000 purified motor neurons were culturedovernight in the wells of a 96-well plate in medium containing10 ng/ml CT-1. One hour prior to treatment with drugs and/orNTs (see text), the culture medium was replaced with L15medium containing 0.1% bovine serum albumin, 1% NaHCO3,and 10 ng/ml CT-1. Cells were treated with NTs for 5 min,washed once in phosphate-buffered saline containing 1 mMsodium orthovanadate and 10 mM sodium fluoride, and har-vested with RIPA buffer (Tris-buffered saline containing 1%Triton X100, 0.5% sodium deoxycholate, 1 mM sodium or-thovanadate, 10 mM sodium fluoride, 5 mM EDTA, 1 mMphenylmethylsulfonyl fluoride, 5 mM N-ethylmaleimide, 5 mM-aminocaproic acid, 5 mg/ml leupeptin, and 5 mg/ml pepstatin

    A). Cell lysates from individual wells containing an identicalnumber of plated cells were centrifuged for 10 min at 10,000 gto pellet DNA. Extracts were separated on 7.5% sodium dode-cyl sulfatepolyacrylamide gels and transferred...


View more >