Immunosuppressant FK506 Enhances Phosphorylation of Nitric

Home , Guanylate cyclase

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 9808-9812, November 1993 Neurobiology Immunosuppressant FK506 enhances phosphorylation of nitric oxide synthase and protects against glutamate neurotoxicity (immunophilins/calcineurin/protein phosphatase/excitotoxicity/cyclosporin A) TED M. DAWSON*t, JOSEPH P. STEINER*, VALINA L. DAWSONt, JAY L. DINERMAN*§, GEORGE R. UHL*t*, AND SOLOMON H. SNYDER*¶II** Departments of *Neuroscience, ¶Pharmacology and Molecular Sciences, 1iPsychiatry and Behavioral Sciences, tNeurology, and §Medicine, Division of Cardiology, Johns Hopkins University School of Medicine, 725 North Wolfe Street, Baltimore, MD 21205; and tNational Institute on Drug Abuse, Addiction Research Center, Molecular Neurobiology Branch, P.O. Box 5180, Baltimore, MD 21224 Contributed by Solomon H. Snyder, July 15, 1993

ABSTRACT Immunosuppressants FK506 and cyclosporin calmodulin-dependent protein kinase, and cGMP-dependent A inhibit neurotoxicity of N-methyl-D-aspartate in primary protein kinase (J.L.D., J.P.S., T.M.D., and S.H.S., unpub- cortical cultures, while having no effect on quisqualate- and lished data) all inhibit its catalytic activity. Accordingly, we kainate-mediated neurotoxicity. Rapamycin completely re- wondered whether enhanced phosphorylation associated verses the neuroprotective effect of FK506. Both FK506 and with FK506 treatment might alter neurotoxicity. In the pres- cyclosporin A inhibit NMDA-elicited/nitric oxide-mediated ent study we demonstrate that low concentrations of FK506 increases in cGMP levels in cortical cultures. FK506 has no and cyclosporin A markedly diminish NMDA neurotoxicity, effect on sodium nitroprusside-induced increases in cGMP. In enhance the phosphorylation of NOS, inhibit NOS catalytic a stably transfected human kidney 293 cell line overexpressing activity, and inhibit NO-mediated increases in cGMP. the gene encoding nitric oxide synthase [L-arginine, NADPH:oxygen oxidoreductase (nitric oxide-forming), EC 1.14.13.39], FK506 inhibits the calcium ionophore A23187, MATERIALS AND METHODS stimulated increases in nitrite (a breakdown product of nitric Cell Cultures. Primary neuronal cultures from cortex were oxide), and potentiates phorbol ester-mediated inhibition of prepared from fetal Sprague-Dawley rats, gestation day nitrite formation. FK506-mediated inhibition of nitric oxide 13-14. After dissection, the cells were dissociated by tritu- formation is completely reversed by rapamycin. Calcineurin ration, counted, and plated in 15-mm multiwell (Nunc) plates dephosphorylates protein kinase C-mediated phosphorylation coated with polyornithine at a density of 3-4 x 105 cells per of nitric oxide synthase. FK506 prevents the calcineurin- well. Proliferation of nonneuronal cells was inhibited by mediated dephosphorylation of nitric oxide synthase and applying 10 ug of 5-fluoro-2'-deoxyuridine 4 days after plat- thereby diminishes the enzyme's catalytic activity. These data ing for a total of3 days. Neurons were maintained in modified establish nitric oxide synthase as a calcineurin substrate. Nitric Eagle's medium (MEM) containing 5% (vol/vol) horse serum oxide synthase catalytic activity is regulated by the phosphor- and 2 mM glutamine in humidified 8% C02/92% air at 370C. ylation state ofthe enzyme. Enhanced phosphorylation ofnitric Medium was changed twice a week. Mature neurons (>21 oxide synthase diminishes catalytic activity, and dephosphor- days in culture) were used in all experiments. In these ylation (through activation of calcineurin) enhances catalytic cultures neurons are -.70-90% of the total number of cells as activity. The neuroprotective effect of FK506 and cyclosporin assessed by neuron-specific enolase and glial fibrillary acidic A presumably involves the inhibition of calcineurin, preventing protein immunohistochemistry (unpublished observation). the dephosphorylation of nitric oxide synthase and its subse- Cytotoxicity. Neurotoxicity was determined by exposing quent activation. the neurons to the various test solutions as described (6). Prior to exposure, the cells were washed three times with a Besides their roles in the immune system, the immunophilins Tris-buffered control solution (CSS) containing 120 mM cyclophilin and FK506 binding protein (FKBP) are highly NaCl, 5.4 mM KCI, 1.8 mM CaCl2, 25 mM Tris'HCl, and 15 concentrated in the brain in discrete neuronal structures mM glucose at pH 7.4. FK506 and cyclosporin A were where they are colocalized with the Ca2+-activated phospha- applied 5 min prior to and during the application of the tase calcineurin (1). Liu et al. (2) demonstrated that very low excitatory amino acids. Rapamycin was applied 5 min prior concentrations of FK506 and cyclosporin A, which bind to to and during the application of FK506 and the excitatory FKBP and cyclophilin, respectively, inhibit calcineurin, and amino acids. After immunosuppressant drug pretreatment, we (1) showed that both drugs enhance the phosphorylation NMDA and quisqualate were applied to the cells for 5 min, of a number ofproteins in the brain. Glutamate neurotoxicity and then the cells were washed with CSS and placed in MEM acting via N-methyl-D-aspartate (NMDA) receptors is impli- containing 21 mM glucose for 24 hr in the incubator. Expo- cated in neuronal damage associated with cerebral infarcts sures to kainate were performed in MEM containing 21 mM and neurodegenerative diseases (3-5). In primary cerebral glucose for 24 hr in the incubator. NMDA, quisqualate, and cortical cultures (6, 7), in hippocampal slices (8-10), and in kainate were also applied to the neurons without immuno- animal models of focal ischemia (11), NMDA toxicity is suppressant drug pretreatment. Twenty to 24 hr after expo- mediated at least in part by nitric oxide (NO) because this sure to test solutions, the neurons were exposed to 0.4% toxicity is blocked by inhibitors of NO synthase [NOS; trypan blue in CSS to stain the residue of nonviable cells and L-arginine, NADPH:oxygen oxidoreductase (nitric oxide- to assess toxicity. Viable and nonviable cells were counted forming), EC 1.14.13.391. Phosphorylation ofNOS by protein with -500-1500 cells counted per well. At least two separate kinase C (PKC) (12), cAMP-dependent protein kinase, Ca2+/ Abbreviations: FKBP, FK506 binding protein; NMDA, N-methyl- The publication costs of this article were defrayed in part by page charge D-aspartate; NOS, nitric oxide (NO) synthase; PKC, protein kinase payment. This article must therefore be hereby marked "advertisement" C; TPA, phorbol 12-tetradecanoate 13-acetate. in accordance with 18 U.S.C. §1734 solely to indicate this fact. **To whom reprint requests should be addressed. 9808 Downloaded by guest on September 25, 2021 Neurobiology: Dawson et al. Proc. Natl. Acad. Sci. USA 90 (1993) 9809

experiments utilizing four separate wells were performed for medium containing the calcium ionophore A23187 (10 ,M) in each data point shown. A minimum of 4,000-12,000 neurons the presence ofthe various drugs. Nitrite accumulation in the were counted for each data point. Viable and nonviable medium was determined after a 2-hr incubation by adding 450 neurons in identical fields were counted by an observer ,l of medium to 450 ul of the Greiss reagent [5% (vol/vol) "blinded" to study design and treatment protocol. concentrated H3PO4 as supplied by Mallinckrodt, 1% sulfa- Phosphorylation Assays. NOS, purified from NOS gene- nilic acid, 0.1% N-(1-naphthyl) ethylenediamine]. Nitrite was transfected 293 kidney cells (293.NOS), was phosphorylated quantified at 539 nm against standards in the same buffer. by purified protein kinase C (Promega) in 50 mM Hepes, pH Materials. cGMP RIA kits were obtained from Amersham. 7.5/100 mM NaCl/10 ,uM CaCl2/2 mM dithiothreitol/50 ,g FK506 was obtained from Fujisawa Pharmaceuticals (Osa- of phosphatidylserine per ml/100 nM phorbol 12-tetra- ka), and rapamycin was obtained from Wyeth. All other decanoate 13-acetate (TPA) for 30 min at 25°C. After 30 min, chemicals and supplies were purchased from commercial the reactions were stopped by addition of 5 mM Na2EDTA sources. and quick-freezing in liquid nitrogen. The phosphorylated NOS was thawed, fractioned (100-,l aliquots), and dephosphorylated by purified calcineurin as follows: 100 ,ul of RESULTS phosphorylated NOS was incubated (i) with 10 nM calcineu- NMDA Neurotoxicity is Dependent on Calcineurin Activa- rin (Sigma), 200 nM calmodulin, 10 ,uM CaCl2, and 20 uM tion. We monitored neurotoxicity in primary cerebral cortical MnCl2 in 150 mM NaCl/50 mM Hepes/1 mM DTT, pH 7.5, neuronal cultures in which a 5-min exposure to NMDA or (ii) with calcineurin in the presence of 100 nM FK506/200 results in death of about 80% of the neurons when observed nM FKBP-12-glutathione transferase fusion protein in the 24 hr later (6). Treatment ofthe cultures with FK506 for 5 min same buffer for 1 hr at 25°C. The reactions were stopped by prior to application ofNMDA and during the 5 min ofNMDA addition of SDS gel sample buffer, and the products were application provided marked protection from neurotoxicity fractionated on 3.5-17% linear gradient Laemmli gels and (Table 1). As little as 25 nM FK506 provided significant silver-stained, and autoradiograms were prepared. protection, while 50% protection was evident between 25- For immunoprecipitation experiments 293 kidney cells 100 nM. To ascertain whether FK506 exerted its protective stably transfected with cerebellar NOS DNA (293.NOS) effects by interacting with its receptor FKBP, we examined were incubated with [32P]orthophosphate in phosphorylation the effect ofrapamycin, which binds to FKBP and blocks the buffer (26 mM NaHCO3/124 mM NaCl/5 mM KCl/2 mM effects of FK506 (2, 16-18). Rapamycin at 1 ,uM completely CaCl2/1.3 mM MgCl2/10 mM glucose, pH 7.2) for 90 min at reversed the effects of FK506. The immunosuppressants 37°C. The cells were washed three times in the phosphory- FK506 and cyclosporin A exert a number of different actions lation buffer, and phosphorylation was stimulated by addition but share the ability to inhibit the Ca2+-dependent phospha- of 150 nM TPA, or by addition of TPA after a 5-min tase activity of calcineurin (2, 19-21). Cyclosporin A at 1 ,uM pretreatment with 100 nM FK506. After a 5-min treatment also protected against NMDA neurotoxicity, suggesting that with TPA, the cells were immediately placed at 4°C on ice and the protection involved inhibition of calcineurin (Table 1). washed twice in ice-cold 50 mM Tris.HCl/150 mM NaCl, pH When applied by themselves, in the absence of NMDA, 500 7.4 (TBS). The cells were immediately lysed by addition of nM FK506, 1 ,uM cyclosporin A, and 1 ,M rapamycin had no 1% Triton X-100/0.5% deoxycholate/0.2% SDS/50 mM effect on neuronal viability (data not shown). Moreover, 500 Tris-HCl/150 mM NaCl/1 mM Na2EDTA, pH 7.4 (lysis nM FK506 had no effect on NMDA-elicited Ca2+ currents in buffer) at 4°C. After a 10-min centrifugation at 10,000 x g, the these cultures (data not shown). cell supernatant was mixed with an equal volume of TBS and Earlier we showed that NMDA neurotoxicity in these 5 ,g ofaffinity-purified NOS IgG directed against a synthetic peptide (Cys-Glu-Asn-Thr-Phe-Gly-Val-Gln-Gln-Ile-Gly- cultures is prevented by inhibition of NO formation, suggest- Pro-Asn-Val) corresponding to the N-terminal 2-13 amino ing that NO plays a role in NMDA neurotoxicity (6, 7, 22). acid residues of the NOS protein (13). After a 2-hr incubation Table 1. FK506 attenuates NMDA neurotoxicity with the anti-NOS antibody, formalin-fixed Staphylococcus % cell death aureus bearing protein A was added for an additional 60 min Drug at 4°C. Immunocomplexes were then centrifuged at 10,000 x 500 ,.LM NMDA 82.8 ± 4.3 g for 10 min, and the pellet was washed three times in lysis + 500 pM FK506 78.3 ± 4.7 buffer. Immunoprecipitated NOS was resolved by PAGE on + 1 nM FK506 77.1 ± 4.0 3.5-17% linear gradient gels in Laemmli buffers; duplicate + 10 nM FK506 68.5 ± 11.2 gels were either electroblotted to nitrocellulose or dried, and + 25 nM FK506 67.2 ± 4.1* autoradiograms were prepared. The level of phosphorylation + 50 nM FK506 61.9 ± 5.8* was quantitated by densitometry using the Eagle Eye Imaging + 100 nM FK506 40.0 ± 9.1* System (Stratagene) and National Institutes of Health Image + 500 nM FK506 29.7 ± 5.8t 1.44 software (14, 15). + 1 ,uM FK506 29.2 ± 4.1t cGMP Assay. Primary neuronal cortical cultures were + 500 nM FK506 + 1 ,uM RAPA 82.4 ± 3.3 treated under identical conditions as those used for assess- + 1 ,uM CsA 56.7 ± 4.8* ment of neurotoxicity except for the addition of 100 ,uM 500 ,uM Glutamate 76.0 ± 5.2 isobutylmethylxanthine to all wells to inhibit phosphodi- + 500 nM FK506 40.6 ± 3.9* esterases. Immediately after the 5-min application of NMDA + 500 nM FK506 + 1 ,uM RAPA 61.7 ± 5.4 or sodium nitroprusside with or without the various other 500 uM Quisqualate 85.9 ± 6.0 drugs, the cells were quenched with 15% trichloroacetic acid. + 500 nM FK506 90.5 ± 4.4 Following ether extraction, cGMP was assayed with an 100 MuM Kainate 83.2 ± 6.3 Amersham 125I-labeling assay kit according to the manufac- + 500 nM FK506 88.1 ± 3.8 turer's instructions. Data are means ± SEM (n 2 8). Cell death was determined by 0.4% Nitrite Assay. A stably transfected cell line containing the trypan blue exclusion by viable cells (see text). Significance was NOS gene, line 293.NOS, was grown to confluence in 1.7-cm determined by Student's t test for independent means. RAPA, wells (12). After a 15-min treatment with 100 nM TPA, 500 rapamycin; CsA, cyclosporin A. nM FK506, 100 nM TPA/500 nM FK506, or 100 nM TPA/500 *P < 0.05. nM FK506/1 ,uM rapamycin, the medium was replaced with tp < 0.0001. Downloaded by guest on September 25, 2021 9810 Neurobiology: Dawson et al. Proc. Natl. Acad Sci. USA 90 (1993)

While NMDA neurotoxicity involves NO, neurotoxicity fol- 1 2 3 lowing treatment with the glutamate derivatives quisqualate NOS*' and kainate is not influenced by inhibition of NOS and so presumably involves other mechanisms such as oxygen free * radicals (3, 4, 6, 23). In our cultures 500 ,uM quisqualate and 200- 4 100 ,uM kainate elicited as much cell death as 500 ,uM NOS*--- 3.5 96- ."- 3 glutamate (Table 1). However, whereas glutamate toxicity Cea) v 2.5 was inhibited by 500 nM FK506 with this inhibition reversed ' 2 by 1 ,uM rapamycin, 500 nM FK506 failed to protect against 66- "''X quisqualate or kainate toxicity. 43- NOS is a Calcineurin Substrate. FK506 and cyclosporin A 29- 0.5 enhance the phosphorylation of a variety of protein sub- o TPA - + + CONT TPA TPA+ strates by inhibiting the phosphatase activity of calcineurin FK506 - - + FK506 (2, 19-21). The selectivity of FK506 protection for NMDA- mediated neurotoxicity suggests that the phosphorylation FIG. 2. Phosphorylation of NOS is enhanced by FK506 in intact state and subsequent catalytic activity of NOS might be cells. (Left) Lanes: 1, unstimulated 293.NOS cells; 2, TPA- stimulated 293.NOS cells; 3, FK506-treated/TPA-stimulated involved. First we demonstrated that NOS is a calcineurin 293.NOS cells. Immunoprecipitation of NOS was performed as substrate by showing that NOS, phosphorylated following described in the text. Shown above the autoradiograms are duplicate stimulation of PKC activity with the phorbol ester TPA, was samples stained with affinity-purified NOS antibody. This experi- dephosphorylated in the presence of calcineurin (Fig. 1) and ment was replicated three times with similar results, and a repre- that 100 nM FK506 together with FKBP completely blocked sentative experiment is shown. (Right) Histogram representing the the dephosphorylation (Fig. 1). Furthermore, TPA- mean ± SEM ofthe three experiments. *, P s 0.05; TPA stimulation stimulated phosphorylation of NOS was significantly in- is significantly different from that of control, and stimulation with creased in the presence of 100 nM FK506 (data not shown). TPA and FK506 is significantly different from that of TPA alone. FK506 alone had no effect on the phosphorylation state of Statistical significance was determined by Student's t test. NOS. Second, to examine whether NOS is phosphorylated by PKC and dephosphorylated by calcineurin in intact cells, slices were inconsistent, probably because of the low abun- we utilized a stably NOS gene-transfected human kidney 293 dance of NOS protein in these areas (data not shown). cell line (293.NOS) overexpressing NOS (12). By using an FK506 Inhibits NOS Catalytic Activity. To ascertain anti-peptide antibody against NOS, NOS protein was isolated whether the enhanced phosphorylation of NOS by FK506 by immunoprecipitation from 293.NOS cells prelabeled with diminishes functional NO activity, we monitored cGMP [32P]orthophosphate, revealing that NOS protein has a low levels in the neuronal cultures (Table 2). In brain slices level of basal phosphorylation (Fig. 2). Following treatment (24-26) and neuronal cultures (6), NMDA increased cGMP with TPA, NOS showed a substantial increase in phosphor- levels several fold, and the increase was prevented by NOS ylation. The addition of 100 nM FK506 to TPA significantly inhibitors (6, 24). In these cultures 100 nM FK506 reduced the enhanced TPA-induced phosphorylation, establishing NOS NMDA stimulation ofcGMP levels by z80%. Cyclosporin A as a calcineurin substrate (Fig. 2). Attempts to demonstrate at 1 ,uM also diminished the NMDA stimulation of cGMP. a similar phenomenon in intact cortical cultures or cortical Evidence that FK506 is acting at the level ofNOS rather than blocking the effects of generated NO on guanylyl cyclase 1 2 3 4 came from our experiments showing that the stimulation of NOS* - - cGMP levels by sodium nitroprusside, which generates NO, was not significantly affected by FK506. 200- Since cGMP measurements are only an indirect measure of NOSi 0.b. .i . NO formation and current methods of measuring NO forma- 96- 4 tion and NOS catalytic activity are insensitive in cortical 66- 6 T cultures, we assessed the effects of FK506 on NOS catalytic Z._ 5 in In 43- 4 activity intact 293.NOS cells (Table 3). intact 293.NOS c CCD .- 4- 3 cells, NOS catalytic activity was directly measured by quan- 29- c T 'm X 3- titation of NO-, a breakdown product of NO, after the 18- addition of the calcium ionophore A23187 (10 ,uM) (Table 3). Z <: 2- 1 Treatment of the cells with 500 nM FK506 reduced A23187- PKC + stimulated catalytic activity by 33%, and treatment with 100 TPA - -+,,+ + 0 nM TPA reduced A23187-stimulated activity by 50%. The CN - - + + CONT TPA TPA+CN TPA+CN+ FK506/FKBP - FK506/FKBP combination of FK506 and TPA was additive, reducing A23187-stimulated NOS catalytic activity by 63%. Rapa- FIG. 1. Phosphorylation of NOS is regulated by calcineurin and mycin at 1 uM completely reversed the effect of FK506, FKSO6/FKBP. (Left) Lanes: 1, control phosphorylation (no addition ofCa2+, TPA); 2, Ca2+/TPA-stimulated phosphorylation of NOS by Table 2. FK506 inhibits NO-stimulated cGMP formation in PKC; 3, calcineurin (CN) dephosphorylation of phospho-NOS; 4, cortical cultures CN dephosphorylation of phospho-NOS in the presence of FK506/ FKBP 12. Silver-stained NOS in each sample is shown above the 32p Drug % basal cGMP level autoradiogram. Autophosphorylation of PKC is clearly visible at 77 500 ,IM NMDA 553 ± 96 kDa. Molecular weight markers in kilodaltons are indicated. This + 100 nM FK506 89 ± 25* experiment was replicated three times with similar results; a repre- + 1 IAM CsA 221 ± 32* sentative experiment is shown. (Right) Histogram representing the 300 ,uM SNP 625 ± 85 mean ± SEM ofthe three experiments. *, P 0.05; TPA stimulation is significantly different from that of control. Stimulation with TPA + 100 nM FK506 510 ± 63 and CN is significantly different from that of TPA alone, and that of Data are means ± SEM of 6-12 wells (two or three different TPA with CN and FK506/FKBP is significantly different from that platings of cultures in duplicate). Significance was determined by ofTPA and CN. Statistical significance was determined by Student's Student's t test for independent means. *P c 0.005. CsA, cyclosporin t test. A; SNP, sodium nitroprusside. Downloaded by guest on September 25, 2021 Neurobiology: Dawson et al. Proc. Natl. Acad. Sci. USA 90 (1993) 9811 Table 3. FK506 inhibits NOS catalytic activity in 293.NOS cells Drug % stimulated nitrite level 10 ,IM A23187 100 + 100 nM TPA 49.8 ± 0.1* + 500 nM FK506 66.6 ± 0.4* + 100 nM TPA + 500 nM FK506 37.4 ± 1.8* + 100 nM TPA + 500 nM FK506 75.2 ± 6.5t + 1 ,AM RAPA Data are means ± SEM of three determinations. The experiment was replicated with similar results. Statistical significance was determined by the Student's t test. RAPA, rapamycin. *P., 0.05. tStimulation by 100 nM TPA/500 nM FK506/1 ,uM RAPA is statistically different from that of 100 nM TPA/500 nM FK506 by Student's t test (P < 0.05). establishing that FK506 reduces NOS activity by interacting with FKBP. FIG. 3. Proposed mechanism for the regulation of the phosphorylation state and catalytic activity of NOS. Glutamate (GLUT) DISCUSSION released from nerve terminals from adjacent neurons activates the NMDA subclass of glutamate receptors (NMDA-R) to increase Our findings indicate that the immunophilins may play a role intracellular Ca2+ (33, 34). Ca2+ binds to calmodulin (CAM), acti- in regulating NMDA-mediated glutamate neurotoxicity. Ev- vating NOS. The generated NO diffuses to adjacent cells to activate idence includes the potent neuroprotective effect of FK506, guanylate cyclase and increase intracellular cGMP levels (25). If which is receptor-mediated because it is blocked by rapa- sufficient quantities of NO are produced, adjacent cells die via mycin, and the similar protective effect provided by cyclo- undefined mechanisms (22), whereas neurons that produce NO are uniquely resistant (35). NOS catalytic activity is inhibited by PKC- sporin A. The protective effects of FK506 appear to involve mediated phosphorylation (12). Activation of PKC by glutamate functional inhibition of NOS by blocking its dephosphoryla- acting on metabotropic glutamate receptors (36-38) or by other tion. Thus, NOS is a calcineurin substrate whose PKC- mechanisms would diminish NOS catalytic activity. Ca2+ entry mediated phosphorylation is enhanced by FK506. PKC phos- activates calcineurin, which dephosphorylates and activates NOS. phorylation of NOS does inhibit its enzymatic activity (12) FK506, complexed to FKBP, binds to calcineurin, inhibiting its (J.L.D., J.P.S., T.M.D., and S.H.S., unpublished data), phosphatase activity. This prevents the dephosphorylation of NOS, although others have shown that PKC enhances NOS activity thus decreasing NOS catalytic activity. With lowered NO production (27). FK506 may also inhibit the dephosphorylation of Ca2+/ adjacent neurons remain viable. calmodulin-dependent protein kinase or cyclic nucleotide- dephosphorylated by calcineurin, a Ca2+/calmodulin- dependent phosphorylation sites. The neuroprotective con- dependent phosphatase, while FK506 bound to FKBP inhib- centrations of FK506 diminish NMDA enhancement of its calcineurin activity, thereby increasing NOS phosphory- cGMP levels, which is NO-mediated (22, 25, 26). The extent lation and diminishing its catalytic activity. The resultant of inhibition of cGMP increased by FK506 is about the same decreased formation of NO after FK506 treatment inhibits as the inhibition produced by 100 ,uM nitroarginine (6), which neurotoxicity. correlates with the extent to which these two drugs protect Physiologic stimulation of neuronal NOS by glutamatergic against neurotoxicity (data not shown; ref. 6). The observa- transmission appears to involve direct activation of the tion that FK506 fails to block quisqualate and kainate neu- enzyme by Ca2+ entering NMDA-regulated channels (22, 24, rotoxicity further supports a role for NO in FK506's neuro- 26). Greengard and colleagues (39, 40) have shown that protective effects because quisqualate and kainate neurotox- NMDA stimulates the of MAP-2 icities do not involve NO (6, 22). dephosphorylation (micro- Besides inhibiting calcineurin, FK506 acts at other sites tubule associated protein) and DARPP-32 (dopamine- and including steroid receptors (28, 29), FKBP25 (30, 31), and cAMP-regulated phosphoprotein of Mr 32,000) via calcineu- other unidentified targets such as those related to FKBP13 rin. In a manner analogous to NMDA-stimulated dephos- (32). Cyclosporin A inhibits calcineurin activity but does not phorylation of MAP-2 and DARPP-32, NMDA may activate exert these other actions. The fact that cyclosporin A as well calcineurin to dephosphorylate NOS, thus providing a sec- as FK506 protects against NMDA toxicity reinforces the ond route for glutamatergic enhancement of NO formation. notion that inhibition of calcineurin is responsible for neu- The neuroprotective effect of FK506 may have clinical roprotection. Though our evidence strongly supports the relevance. FK506 and cyclosporin A have both been used hypothesis that FK506 neuroprotection involves inhibition of extensively as immunosuppressants during organ transplant NOS, the fact that FK506 increases phosphorylation of surgery. We have demonstrated that [3H]FK506 penetrates numerous proteins indicates that other mechanisms may also readily into the brain (data not shown), while cyclosporin A play some role. does not (41). Interestingly, in a study of liver transplanta- A model that emerges from these findings is that glutamate, tion, 7 of 14 patients receiving cyclosporin A showed global acting via NMDA receptors, stimulates Ca2+ iniflux into cells cerebral ischemia, while none of the 14 patients receiving (Fig. 3). The Ca2+ binds to calmodulin which activates NOS, FK506 showed such alterations (42). This suggests that drugs generating NO. The NOS-containing cells are resistant to that interact with immunophilins and enter the brain may neurotoxicity (7, 35) but produce NO that diffuses to adjacent have therapeutic potential. cells to elicit toxicity via mechanisms that are currently unknown NOS can be We thank Fujisawa Pharmaceutical for FK506 and Dr. S. Sehgal (22). phosphorylated by PKC, cAMP- at Wyeth for rapamycin. We thank J. Mong and D. A. Bartley for dependent protein kinase, Ca2+/calmodulin-dependent pro- technical assistance. This work was supported by U.S. Public Health tein kinase (12), and cGMP-dependent protein kinase Service Grant DA-00266 and a Research Scientist Award (DA-00074) (J.L.D., J.P.S., T.M.D., and S.H.S., unpublished data), with to S.H.S. T.M.D. is a Pfizer Postdoctoral Fellow and is supported by phosphorylation lowering catalytic activity (12) (J.L.D., grants from the American Academy of Neurology, the French J.P.S., T.M.D., and S.H.S., unpublished data). NOS is Foundation for Alzheimer's Research, and a U.S. Public Health Downloaded by guest on September 25, 2021 9812 Neurobiology: Dawson et al. Proc. Natl. Acad. Sci. USA 90 (1993) Service-Clinical Investigator Development Award (NS-01578). Chang, H. Y., Walsh, C. T. & Rusnak, F. (1992) Proc. Natl. J.P.S. is supported by a postdoctoral fellowship from the U.S. Public Acad. Sci. USA 89, 3741-3745. Health Service (MH-10101). V.L.D. is supported by National Insti- 21. Liu, J., Albers, M. W., Wandless, T. J., Luan, S., Alberg, tute of General Medical Sciences Pharmacology Research Associate D. G., Belshaw, P. J., Cohen, P., MacKintosh, C., Klee, C. B. Training Program and an Intramural Research Training Award from & Schreiber, S. L. (1992) Biochemistry 31, 3896-3901. the National Institutes of Health. S.H.S. was supported by a grant 22. Dawson, T. M., Dawson, V. L. & Snyder, S. H. (1992) Ann. from the W. M. Keck Foundation, the International Life Sciences Neurol. 32, 297-311. Institute, and V.L.D. and G.R.U. were supported by the intramural 23. Puttfarcken, P. S., Lyons, W. E. & Coyle, J. T. (1992) Neu- program of National Institute of Drug Abuse. ropharmacology 31, 565-575. 24. Bredt, D. S. & Snyder, S. H. (1989) Proc. Natl. Acad. Sci. 1. Steiner, J. P., Dawson, T. M., Fotuhi, M., Glatt, C. E., Snow- USA 86, 9030-9033. man, A. M., Cohen, N. & Snyder, S. H. (1992) Nature (Lon- 25. Moncada, S., Palmer, R. M. J. & Higgs, E. A. (1991) Phar- don) 358, 584-587. macol. Rev. 43, 109-142. 2. Liu, J., Farmer, J. D., Jr., Lane, W. S., Friedman, J., Weiss- 26. Garthwaite, J. (1991) Trends NeuroSci. 14, 60-67. man, I. & Schreiber, S. L. (1991) Cell 66, 807-815. 27. Nakane, M., Mitchell, J., Forstermann, U. & Murad, F. (1991) 3. Choi, D. W. (1988) Neuron 1, 623-634. Biochem. Biophys. Res. Commun. 180, 1396-1402. 4. Meldrum, B. & Garthwaite, J. (1990) Trends Pharmacol. Sci. 28. Tai, P.-K. K., Albers, M. W., Chang, H., Faber, L. E. & 11, 379-387. Schreiber, S. L. (1992) Science 256, 1315-1318. 5. Choi, D. W. (1992) Science 258, 241-243. 29. Yem, A. W., Tomasselli, A. G., Heinrikson, R. L., Zurcher- 6. Dawson, V. L., Dawson, T. M., London, E. D., Bredt, D. S. Neely, H., Ruff, V. A., Johnson, R. A. & Deibel, M. R., Jr. & Snyder, S. H. (1991) Proc. Natl. Acad. Sci. USA 88, (1992) J. Biol. Chem. 267, 2868-2871. 6368-6371. 30. Galat, A., Lane, W. S., Standaert, R. F. & Schreiber, S. L. 7. Dawson, V. L., Dawson, T. M., Bartley, D. A., Uhl, G. R. & 31, 2427-2434. Snyder, S. H. (1993) J. Neurosci. 13, 2651-2661. (1992) Biochemistry 8. Izumi, Y., Benz, A. M., Clifford, D. B. & Zorumski, C. F. 31. Jin, Y.-J., Burakoff, S. J. & Bierer, B. E. (1992) J. Biol. Chem. (1992) Neurosci. Lett. 135, 227-230. 267, 10942-10945. 9. Moncada, C., Lekieffre, D., Arvin, B. & Meldrum, B. (1992) 32. Jin, Y.-J., Albers, M. W., Lane, W. S., Bierer, B. E., NeuroReport 3, 530-532. Schreiber, S. L. & Burakoff, S. J. (1991) Proc. Natl. Acad. Sci. 10. Wallis, R. A., Panizzon, K. & Wasterlain, C. G. (1992) Neu- USA 88, 6677-6681. roReport 3, 645-648. 33. Zorumski, C. F. & Thio, L. L. (1992) Prog. Neurobiol. 39, 11. Nowicki, J. P., Duval, D., Poignet, H. & Scatton, B. (1991) 295-336. Eur. J. Pharmacol. 204, 339-340. 34. Mayer, M. L., Benveniste, M., Patneau, D. K. & Vyklicky, L., 12. Bredt, D. S., Ferris, C. D. & Snyder, S. H. (1992) J. Biol. Jr. (1992) Ann. N. Y. Acad. Sci. 648, 194-204. Chem. 267, 10976-10981. 35. Koh, J.-Y., Peters, S. & Choi, D. W. (1986) Science 234,73-76. 13. Bredt, D. S., Hwang, P. H., Glatt, C., Lowenstein, C., Reed, 36. Scholz, W. K. & Palfrey, H. C. (1991) J. Neurosci. 11, 2422- R. R."& Snyder, S. H. (1991) Nature (London) 351, 714-718. 2432. 14. Sutherland, J. C., Sutherland, B. M. & Emrick, A. (1991) 37. Miller, R. J. (1991) Trends Pharmacol. Sci. 12, 365-367. Biotechnology 10, 492-497. 38. Schoepp, D., Bockaert, J. & Sladeczek, F. (1990) Trends 15. Correa-Rotter, R., Mariash, C. N. & Rosenberg, M. E. (1992) Pharmacol. Sci. 11, 508-515. Biotechnology 12, 154-158. 39. Halpain, S. & Greengard, P. (1990) Neuron 5, 237-246. 16. Chang, J. Y., Sehgal, S. N. & Bansbach, C. C. (1991) Trends 40. Halpain, S., Girault, J.-A. & Greengard, P. (1990) Nature Pharmacol. Sci. 12, 218-223. (London) 343, 369-372. 17. Thomas, A. W. (1989) Immunol. Today 10, 6-9. 41. Begley, D. J., Squires, L. K., Zlokovic, B. V., Mitrovic, 18. Schreiber, S. L. (1991) Science 253, 283-287. D. M., Hughes, C. C. W., Revest, P. A. & Greenwood, J. 19. Fruman, D. A., Klee, C. B., Bierer, B. E. & Burakoff, S. J. (1990) J. Neurochem. 55, 1222-1230. (1992) Proc. Natl. Acad. Sci. USA 89, 3686-3690. 42. Lopez, 0. L., Martinez, A. J. & Torre-Cisneros, J. (1991) 20. Swanson, S. K.-H., Born, T., Zydowsky, L. D., Cho, H., Transplant. Proc. 23, 3181-3182. Downloaded by guest on September 25, 2021