- Email: [email protected]
CYTOKINE-MEDIATED NEURONAL APOPTOSIS
Neurochem. Int. Vol.30,Nos 4/5, pp. 427431, 1997
Pergamon
PII: S0197-0186(96)00078-2
Copyright01997 ElsevierScienceLtd Printedin Great Britain.All rightsreserved 0197+186/97$17.00+0.00
CYTOKINE-MEDIATED NEURONAL APOPTOSIS S. HU, P. K. PETERSON and C. C. CHAO* Neuroimmunobiology and Host Defense Laboratory, Minneapolis Medical Research Foundation and the University of Minnesota Medical School, Minneapolis, MN 55404, U.S.A. (Received 11 April 1996; accepted4 June 1996)
Abstract—Cytokines have been reported to induce neuronal injury via the free radical nitric oxide (NO); however, the precise mechanism underlying cytokine-mediated neurotoxicity is unclear. We investigated the hypothesis that cytokine-mediated neurotoxicity in primary cultures of human fetal neurons occurs via an apoptotic mechanism triggered by NO. Treatment of mixed neuronal/glial cell cultures with interferon (IFN)-y plus interleukin (IL)-1~ for 13 days induced a high output of NO accompanied by marked neuronal loss. The NO synthase inhibitor N-monomethyl-L-arginine (NMMA) significantly attenuated cytokineinduced neuronal loss, confirming the involvement of NO. Cytokine-mediated neuronal injury was accompanied by morphologic changes and a DNA fragmentation pattern consistent with apoptosis. Treatment of neuronal cell cultures with NMMA protected against cytokine-mediated apoptotic death. These findings, using primary human neuronal cell cultures, support the hypothesis that cytokine-mediated neurotoxicity involving NO proceeds via an apoptotic mechanism. These findings could lead to the development of new therauies for neurodegenerative diseases involving glia, cytokines, and NO. ~ 1997
Elsevi~rScienceLtd. All rights reserved -
In recent years, intense interest has focused on the participation of immune mediators, i.e. cytokines and free radicals, in the pathogenesis of neurodegenerative diseases (Chao et al., 1995a; Fontana et al., 1996). Histopathological evidence suggests that the appearance of reactive glial cells is a hallmark of immunologically mediated brain injury (Merrill and Chen, 1991; Dickson et al., 1991; Eddleston and Mucke, 1993). Microglial cells (the resident macrophages of the brain) could be a major source of interleukin-1~ (IL-1/?) (Lee et al., 1993a; Chao et al., 1995b), whereas infiltrated lymphocytes and natural killer cells appear to be the key producers of interferon-y (IFN-y) (Lewis and Wilson, 1990). The release of these cytokines within the brain may play a contributory role in the development of neuronal injury. Using rodent brain cell cultures, several laboratories have demonstrated that cytokine-activated microglial cells damage neurons via generation of the toxic free radical nitric oxide (NO) (Boje and Arora, 1992; Chao et al., 1992). However, human microglial cells appear to produce little, if any, NO in response *To whom all correspondence should be addressed at: Minneapolis Medical Research Foundation, 914 South 8th Street, D3, Minneapolis, MN 55404, U.S.A.
to cytokines such as IL-1P, IFN-y, and tumor necrosis factor-u, which explains a lack of microglial cellmediated neuronal damage in human brain cell cultures treated with these cytokines (Peterson et al., 1994). Although human microglia are relatively inefficient in NO production, human astrocytes generate abundant amounts of NO in response to IL-1 stimulation, which is further potentiated by IFN-y (Lee et al., 1993b; Hu et al., 1995). Recently, we have shown that treatment of mixed human fetal neurenal/glial cell cultures with IL-1P plus IFN-y results in the production of NO which is accompanied by neuronal damage (Chao et al., 1996). Little is known, however, about the precise mechanism underlying cytokine-induced neuronal death. Presently, cell death is thought to occur by two distict pathways: programmed cell death (apoptosis) or necrosis. While cells undergoing apoptosis generally die at a relatively slow rate, the necrotic mode of cell loss is relatively rapid. Recent findings suggest that toxic free radicals can precipitate either apoptotic or necrotic neuronal death, depending upon the intensity of the initial insult (Bonfoco et al., 1995). Because cytokine-mediated neurotoxicity appears to be a slow process (Chao et al., 1996), we hypothesized that an apoptotic mode of neuronal loss is induced by NO. 427
428
S. HLI(,Id.
EXPERIMENTALPRO(’ED[JRFX
E.Yp<,/im<,/?/( I[pfoto((]l”
C~ytcjkille-nlcdiiltccineurotoxicity, WM assessed f’ollowirrg treatment of mixed neurona;gliai cell cultures with IL-I/; (10 rrgjml), IFN-; (200 Uml), or a combitl:ition of these cytc>Primary, human fetal brain cell cultures were prepared kines for 13 d:iys. To evaluate the in~olyemcmt of NO, cell using a rcccntly developed technique under a protocol cultures were treated with 500”;tM of the NO synthusc inhihiappl-o~ed by our institutional Human Subjects Research tor N{;-monomethyl-l-argininc (NM MA) silllLllt:tnec>~isiy C’mnmittcc (Chao ond Hu, 1994). Brielly, brain corticcs from when IL- I[~ and 1FN-; were added. Cullurc supernatants 14 1X-week-old abortuses were treated with ().25!L0 trypsin for 30 min. Afler extensive washing. dispersed brain cells were then harvested for nitrite (NO: ) determination and LDH ;Lctivity. WCI-C plated onto collagen-coated 24-well plates at a conTo investigate whether the mechanism cytokine-mecfiatcd centration of 5 x 105 cellsj500 ~~1and cultured in DMEM neurotoxicity involved an apoptotic mode of neuronal death. (Signm, St Louis, M(). U. S.A.) containing 10(YO fetal bovine two approaches were used: lerminal transferase-mediated serum (F fitS) at 37 C in a humidified 10’J/O C’02atrnosphcrc. UTP nick end Iabcl]ing (TUNEL) technique and DNA fragOn d~iy 5. (IICcullurc medium was replaced with DMEM containing I()”/. FBS, uridirrc (33.6 pgjml, Sigma), and ffuc>- mentation, EXJLTl1112Lt]OLl 01’ ncurorral death by these techniques was performed on duys 7, I(). and 13 following rodcoxyuricfine (13.6 ~g~m], Sigma) to prevent overgrowth cytokine tre~tmcnt. of astroglial cells. Culture medicrlm with 10(YOFBS was replaced on day 6 and every 4 days thereafter. On day 12, DNA In situ cnd Itrl)(,llitl,q(Jffiu,y}?!c,tl[c({ the neurcrmd cell cultures contained differentiated neurons ?-Wwltlgon a supporting layer of’astrocytes. The mixed new For i~t .Iitu detection of’nuclear DNA frtigrnentation. the romd and glial cell cultures consisted of about 50–60”/. neuT[JNEL technique was used as described by Gavrieli e/ u/. rons (stainecf with anli-neurou specific enolase antibodies; (1992) with minor modifications, In brief, mixed new Polysclences, Barrington. PA, U.S.A.), 40–50”/0astrocytes ronal~glial cell cultures were fixed for 20 min in 40/0 par(stained with anti-glial fribrillary acicficprotein (GFAP) mltiafornmdehyde at room temperature. C’cllcultures were then body), <2% nlicrogli;L (stained with anti-CD6X antibody; washed f’oul-times with phosphutc-huffered saline. Endogem Dako), und < 1“/0 oligodendrocytes (stained with antious peroxidase was inactivated by incubating cell cultures in galactocerehmside antibodies; Boehringcr Mannheim, 2’Yo1{20, for 5 min. Afler preincub~tion with TDT hufier Indianapolis. IN. U.S.A.). These 12-day mixed necl- (30 mM Trizam base, pH 7.2, 140mM sodium cacodykite, ronai glial cell cultures were used in all experiments in which I mM cobalt chloride) for 5 min at room temper:iture. CCII neumtnxicit) was eva]uuted. cultures were subjected to nick translation procwfurcs f’ollnwed by exposure to TDT hufi’er containing 0.5 LHml ter/f,\//()(,\/[, IL(/lL{/(>,$ minal tmnsferase (Boehringer Mannheim) ~lnd 0.5 ;mol biotinylated 16-dUTP (Boehringer Manuheim) in a humid Human fetal brain tissues were obtained from 16 22-week- atmosphere f’or60 min at 37 C. The reac(ion was terminated nld aborted Ictuses. Primary astrocyte cultures were prepared by incubating in TB buffer (300”mM sodium chlnridc, 30 mM iLs previously described (Chao [,/ a[., 1996). In brief, brain sodium citmtc) for 15min at room tcmpcruturc. After washitissues were cleared of mcninges, minced into small fragng four tirncs with TDT huffcl-.CCIIcultures were incuhated ments, and incubated with 0.25°/0 trypsin (Sigma) in Caz+with the ABC Elite Kit (Vector Laboratories, Burlingame. and Mg2+-free puck’s saline (Gibco Lahoratorics, Grand CA, U.S.A.) for 45 min at rooln temperature followed by Isktnd, NY, U.S.A.) for 30 min at 37 C’with gentle shuking. color development with 3,3’-diaminohezidone. An equal VOILUIIC of mcdiurn containing 10“/. heat-inactivated FBS was :iddcd to iuactivale the trypsin, and tissues DNA (,.~tracfiotl were centrifuged ut 1200rpm for 5 min. After washing twice DNA was extracted from mixed ncuronal’gliul cell culwith Hank’s boffcrcd saline solution (Gibco). tissue fragments were replaced with fresh medium and triturated f’or tures using DNA Isolation Kits (Gcntra Systems, Inc.. Minneapolis, MN, U.S.A.). In brief. cell cultures were Iysed with 15-20 passages thruugb a sterile Pasteur pipette. Cell susLysing huffer followed by ~Nase tre:itment for 60 min at pensions were seeded into 75 cm’ flasks It a density of 80 37’’C’, prntcin prccipitution solution, and centrifuged at 100x 10” cells~flask in a high-glucose DMEM containing 15.()()()gfor 3 min. DNA wus pwpwxl by additicrn 01 1000/” 10!4 FBS, penicillin (100 U~ml), and streptomycin isopropanol followed by ccntrifugatiorr ~t 15,000,< for I min. (l f)OJ[g,’ml).These flmks then were incubtitcd al 37 C in Pellets were wushed with 70(%,ethanol and samples were air a w~iter-s~itu rittecl, I()% C02 incubator. After 24 ~lnd 4Xh dried. incubation. medium was replaced with fresh DMEM containing 10”/. FBS. Culture medium WJSchanged once weekly A,qcw(m,qc[(,[<,cfI(,pl?(II(,.Y/.s thereafter. On day 21, flasks were shaken at 180–200rpm for 16h. followed by washing with Ca2i - and Mg2”+-freeHank’s DNA samp[es ( I()J(g) were loaded ontc 1.X(YO ~garose gel buflered saline solution cont:iining O.125%,trypsin for 30 min contiiining cthidium bromide (0.4 mg~m}) :ind electrm at 37 C. followed by >~ddition of 10“/o FBS-containing phonetically sep~ir:itcd. DNA was kisuali~ed by a UV Imedium. After ccntrifugation. cell suspensions were seeded (302 nM) trtinsilluminator, Jnd the gels were photngr~ipbcd into a new flask in DMEM containing 10“4 FBS, and the with a polaroid c:inwra. culture medium was changed 24 h later. This procedure was repeated [hrce times at weekly intervals, Finally, highly ,Yl(itis{icul(I)u[r$i$ enriched astmcytcs ( > 98°/0 of cells st:iirwd with anti- GF’AP Student’s ~-test(Statview 11,Abacus (’oncepts, Inc., Berkemtihody) were seeded into 24-wc1lplales a~ a density of 105 ley. C-A,U.S.A.) was used to compare the difl’erencesbetween cells ‘well fnr experiments. A’fi. Ye(l Il[,ul’oll(l[,q[i(l[ C(I1cL[/[L1r(,.\’
Cytokine-mediated neuronai apoptosis the means of two groups. Analysis of variance followed by Scheffe’s F-test was used to compare the differences among multiple means. RESULTS
Cytokine-rnediated neuronal death Treatment of mixed neuronal/glial cell cultures with IL-1/l induced a significant (P< O.01) amount of NO reflected by accumulation of the stable metabolize NO; (Fig. l(A)). This was accompanied by a slight release of LDH (Fig. l(B)). IFN-y alone had no effect on NO; production or neuronal LDH release (Fig. 1). Treatment of neuronal/glial cell cultures with IL1~ plus IFN-y for 13 days resulted in elevated (P< O.01)NO; levels (Fig. l(A)), which corresponded with marked neuronal damage (f’< 0.01) as assessed by increased LDH release into culture supernatants (Fig. l(B)). The NO synthase inhibitor NMMA suppressed the elevated NO; levels (Fig. l(A)) as well as LDH release induced by these cytokines (Fig. l(B)),
A
IL-113+ IFN-y
60150 i
i
‘L-“
B
60
[
IL-113+ IFN-y
T
+ - NMMA Fig. 1. Cytokine-mediated NO production and neuronal injury. Mixed human neuronal/glial cell cultures were treated with medium (control), IL-1P (10 rig/ml), IFN-y (200 U/ml) and IL-1P plus IFN-y in the absence or presence of the NO synthase inhibitor NMMA (500 pM). Supernatants were harvested after 13 days of exposure for determination of (A) NO; levels and (B) LDH release. Data are mean~SEM of duplicates and representative of three separate experiments. Control
IFN-y
IL-1 13
429
suggesting the involvement of NO in cytokinemediated neuronal injury within mixed human neurenal/glial cell cultures. In a separate experiment, similar NOq- levels were observed with treatments of highly (>98%) enriched astrocyte cultures with IL1~ or IL-1P plus IFN-y for 13 days; however, little LDH was detected over the control astrocyte cultures, suggesting that astrocytes are more resistant to cytokine-mediated injury (data not shown). Cytokine-mediated neuronal apoptosis To determine the mode of cell death, TUNEL staining for condensed chromatin in pyknotic nuclei was used. Using this technique, 5–70/. of cytokine-treated neuronal cells began to display apoptosis after 7 days of exposure to cytokines. The number of apoptotic neurons increased by 10 days (l&l 50/0).Exposure of neuronal/glial cell cultures to IL-1P and IFN-y for 13 days markedly increased the number of apoptotic neurons (25–300/0), reflected by microscopic examination (Fig. 2(B)) as compared to control cell cultures (Fig. 2(A)). NMMA treatment attenuated (P< O.01) cytokine-induced neuronal apoptosis from 25–30°/0to 3–5~0 (Fig. 2(C)). In highly (> 98Yo)enriched astrocyte cultures, cytokine-induced apoptotic cell death after 10 days of exposure was minor ( < 50/.), suggesting that astrocytes are more resistant to cytokinemediated apoptosis (data not shown). To further confirm that cytokine-mediated neuronal cell death involved an apoptotic mechanism, DNA degradation was assessed by gel electrophoresis (Fig. 3). Exposure of mixed neuronal/glial cell cultures to IL-1P plus IFN-y for 7 days induced a minor increase in large molecular weight fragments (data not shown). After 13 days of treatment with these cytokines, cell cultures displayed. DNA degradation with markedly increased large and small molecular weight fragments (Fig. 3). Treatment of cell cultures with NMMA attenuated cytokine-induced DNA fragmentation (Fig. 3), suggesting an involvement of NO in cytokine-mediated neuronal apoptosis.
DISCUSSION
We demonstrated in the present study that cytokine-induced neurotoxicity, which was partially blocked by the NO synthase inhibitor NMMA, proceeded by an apoptotic pathway as demonstrated by nuclear TUNEL staining technique and DNA fragmentation experiments. These findings strongly support the hypothesis that cytokine-mediated
430
S. Hu e( a[.
.,,
A
#?
‘%
c
4$
‘.
Fig. 2. Morphological assessment of’ apoplosis of neurons induced by cytokines. Mixed neurorral~glial cell cultures were treated with (A) mccfium(Cmrlrol), (B) IL-f~ (10 ng~ml)plus IFN-~ (200 [J/ml) or (C) NMMA (500 ~1M)plus cytokines for 13days prior to nuclear staining with the TUNEL technique. Data arc dm-ivedfrom three separate brain cell cultures.
neurotoxicity involves NO and that this free radical induces programmed Almost a century
cell death of neurons. ago, it was observed that mic-
roglia migrate to, and differentiate and proliferate at, sites of inflammation within the brain (del Rio-Hortega, 1932). It has been proposed that these glial cells play a neuropathogenic role (Chao etal., 1995a). In contrast to studies with murine microglia, cytokineactivated human microglia produce little NO that is incapable of inducing neurotoxicity in vitro (Peterson
Fig. 3. DNA fragmentation. Mixed neuronal~glial cell cultures were treated with medium (control, lane 2), IL-11~ (10 ngml) plus IFN-y (200 Ujml) (lane 3) or NMMA (500 PM) plus cytokincs (kmc 4) for 13 days prior to harvesting of DNA, Molecular marker was loaded into lane 1. DNA (l Opg) from each trcutmcnt group wus loaded into each agarose gel lane prior to clcctrophoresis. Data arc representative of three separate experiments.
et al., 1994). Human neurons are susceptible, however, to NO neurotoxicity as evidenced in human mixed neuronal/glial cell cultures incubated with activated murine microglia (Peterson et al., 1994). Of note, involvement of NO in immunologic brain injury has been implicated in brain lesions of patients with mul-
Cytokine-mediated neuronal apoptosis
tiple sclerosis (Brosnan et al., 1994). Recently, we have found that cytokine-stimulated human astrocytes are capable of generating substantial amounts of neurotoxic NO (Chao et al., 1996). While the precise mechanism underlying cytokine-mediated neuronal death has not been elucidated, in the present study we demonstrated that an apoptotic mode of neuronal death is involved. Because NO generation by cytokine-treated astrocytes is a slow process, it is not surprising that we found an apoptotic mode of neuronal injury in human mixed neuronal/glial cell cultures. Recent studies using rat cortical neuronal cultures have shown that low levels of NO mediate a delayed apoptotic mode of neuronal injury (Bonfoco et al., 1995). The observation that NMMA only partially (approximately 5@60%) attenuated cytokinemediated neuronal injury, suggests that NO is not the only mediator responsible for cytokine-mediated neurotoxicity. Nonetheless, results of the present study may lead to development of new therapeutic strategies for neurodegenerative diseases in which activated glia, cytokines and NO appear to be involved, such as multiple sclerosis, Alzheimer’s disease, and AIDS dementia. Acknowledgements—This study was supported in part by USPHS grants DA-04381, DA-09924 and a grant from the Alzheimer’s Association.
REFERENCES
Boje, K. M. and Arora, P. K. (1992) Microglia-produced nitric oxide and reactive nitrogen oxides mediate neuronal cell death. Brain Res. 587, 25&256. Bonfoco, E., Krainc, D., Ankarcrona, M., Nicotera, P. and Lipton, S. A. (1995) Apoptosis and necrosis: two distinct events induced, respectively, by mild and intense insults with N-methyhD-aspartate or nitric oxide/superoxide in cortical cell cultures. Proc. natn Acad. Sci. U.S.A. 92, 7162-7166. Brosnan, C. F., Battistina, L., Raine, C. S., Dickson, D. W., Casadevall, A. and Lee, S. C. (1994) Reactive nitrogen intermediates in human neuropathology: an overview. Devl Neurosci. 16, 152–161. Chao, C. C., Hu, S., Molitor, T. W., Shaskan, E. G. and Peterson, P. K. (1992) Activated microglia mediate neu-
431
ronal cell injury via a nitric oxide mechanism. J, Immunol. 149, 273&2741. Chao, C. C. and Hu, S. (1994) Tumor necrosis factor-u potentates glutamate neurotoxicity in human fetal brain cell cultures. Devl Neurosci. 16, 172–179. Chao, C. C., Hu, S. and Peterson, P. K. (1995) Glia, cytokines, and neurotoxicity. Crit. Rev. Neurobiol. 9, 189-205. Chao, C, C., Hu, S,, Sheng, W. S. and Peterson, P. K. (1995) Tumor necrosis factor-alpha production by human fetal microglial cells: regulation by other cytokines. Devl Neurosci. 17, 97–105. Chao, C. C., Hu, S., Sheng, W. S., Bu, D-F., Bukrinsky, M. I. and Peterson, P. K, (1996) Cytokine-stimulated astrocytes damage human neurons via a nitric oxide mechanism. Glia 16,276-284, del Rio-Hortega, P. (1932) Microglia. In Cytology and Cellular Pathology of the Nervous System, ed. W. Penfield, Vol. 2, pp. 481–584. Paul P. Hocker, New York. Dickson, D. W., Mattiace, L, A,, Kure, K., Hutchins, K., Lyman, W. D. and Brosnan, C. F. (1991) Microglia in human disease, with an emphasis on acquired immune deficiency syndrome. Lab. Invest. 64, 135–156. Eddleston, M, and Mucke, L, (1993) Molecular profile of reactive astrocytes—implications for their role in necrologic disease. Neuroscience 54, 15–36. Fontana, A., Constam, D., Frei, K., Koedel, U., Pfister, W., and Weller, M. (1996) Cytokines and defense against CNS infection. In Cytokines and the CNS, ed. R. M. Ransohoff and E. N, Benveniste, pp. 187–220. CRC Press, Boca Raton, Florida. Gavrieli, Y., Sherman, Y. and Ben-Sasson, S. A. (1992) Identification of programed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Cell Biol. 119, 493–501. Hu, S., Sheng, W. S., Peterson, P. K. and Chao, C. C. (1995) Differential regulation by cytokines of production of nitric oxide by human astrocvtes. Glia 15, 491494, Lee, S. C:, Liu, W., Dickson, D. W, Brosnan, C. F. and Berman, J. W. (1993) Cytokine production by human fetal microglia and astrocytes: differential induction by lipopolysaccharide and IL-1/J. J.Immunol. 150, 2659–2667. Lee, S. C., Dickson, D. W., Liu, W, and Brosnan, C. F. (1993) Induction of nitric oxide synthaseactivity in human astrocytes by interleukin-1~ and interferon-gamma. J. Neuroimmunol. 46, 19–24. Lewis, D. B. and Wilson, C. B. (1990) Gamma-interferon: an immunoregulatory Iymphokine with immunotherapeutic potential. Pediatr. infect. Dis. J. 9, 642–651. Merrill, J. E. and Chen, I. S. Y. (1991) HIV-1, macrophages,glial cells, and cytokines in AIDS nervous system disease. FASEB J. 5, 2391–2397. Peterson, P. K., Hu, S., Anderson, W. R. and Chao, C. C. (1994) Nitric oxide production andneurotoxicity mediated by activated microglia from human versus mouse brain. J. infect. Dis. 170, 457J160.
Pergamon
PII: S0197-0186(96)00078-2
Copyright01997 ElsevierScienceLtd Printedin Great Britain.All rightsreserved 0197+186/97$17.00+0.00
CYTOKINE-MEDIATED NEURONAL APOPTOSIS S. HU, P. K. PETERSON and C. C. CHAO* Neuroimmunobiology and Host Defense Laboratory, Minneapolis Medical Research Foundation and the University of Minnesota Medical School, Minneapolis, MN 55404, U.S.A. (Received 11 April 1996; accepted4 June 1996)
Abstract—Cytokines have been reported to induce neuronal injury via the free radical nitric oxide (NO); however, the precise mechanism underlying cytokine-mediated neurotoxicity is unclear. We investigated the hypothesis that cytokine-mediated neurotoxicity in primary cultures of human fetal neurons occurs via an apoptotic mechanism triggered by NO. Treatment of mixed neuronal/glial cell cultures with interferon (IFN)-y plus interleukin (IL)-1~ for 13 days induced a high output of NO accompanied by marked neuronal loss. The NO synthase inhibitor N-monomethyl-L-arginine (NMMA) significantly attenuated cytokineinduced neuronal loss, confirming the involvement of NO. Cytokine-mediated neuronal injury was accompanied by morphologic changes and a DNA fragmentation pattern consistent with apoptosis. Treatment of neuronal cell cultures with NMMA protected against cytokine-mediated apoptotic death. These findings, using primary human neuronal cell cultures, support the hypothesis that cytokine-mediated neurotoxicity involving NO proceeds via an apoptotic mechanism. These findings could lead to the development of new therauies for neurodegenerative diseases involving glia, cytokines, and NO. ~ 1997
Elsevi~rScienceLtd. All rights reserved -
In recent years, intense interest has focused on the participation of immune mediators, i.e. cytokines and free radicals, in the pathogenesis of neurodegenerative diseases (Chao et al., 1995a; Fontana et al., 1996). Histopathological evidence suggests that the appearance of reactive glial cells is a hallmark of immunologically mediated brain injury (Merrill and Chen, 1991; Dickson et al., 1991; Eddleston and Mucke, 1993). Microglial cells (the resident macrophages of the brain) could be a major source of interleukin-1~ (IL-1/?) (Lee et al., 1993a; Chao et al., 1995b), whereas infiltrated lymphocytes and natural killer cells appear to be the key producers of interferon-y (IFN-y) (Lewis and Wilson, 1990). The release of these cytokines within the brain may play a contributory role in the development of neuronal injury. Using rodent brain cell cultures, several laboratories have demonstrated that cytokine-activated microglial cells damage neurons via generation of the toxic free radical nitric oxide (NO) (Boje and Arora, 1992; Chao et al., 1992). However, human microglial cells appear to produce little, if any, NO in response *To whom all correspondence should be addressed at: Minneapolis Medical Research Foundation, 914 South 8th Street, D3, Minneapolis, MN 55404, U.S.A.
to cytokines such as IL-1P, IFN-y, and tumor necrosis factor-u, which explains a lack of microglial cellmediated neuronal damage in human brain cell cultures treated with these cytokines (Peterson et al., 1994). Although human microglia are relatively inefficient in NO production, human astrocytes generate abundant amounts of NO in response to IL-1 stimulation, which is further potentiated by IFN-y (Lee et al., 1993b; Hu et al., 1995). Recently, we have shown that treatment of mixed human fetal neurenal/glial cell cultures with IL-1P plus IFN-y results in the production of NO which is accompanied by neuronal damage (Chao et al., 1996). Little is known, however, about the precise mechanism underlying cytokine-induced neuronal death. Presently, cell death is thought to occur by two distict pathways: programmed cell death (apoptosis) or necrosis. While cells undergoing apoptosis generally die at a relatively slow rate, the necrotic mode of cell loss is relatively rapid. Recent findings suggest that toxic free radicals can precipitate either apoptotic or necrotic neuronal death, depending upon the intensity of the initial insult (Bonfoco et al., 1995). Because cytokine-mediated neurotoxicity appears to be a slow process (Chao et al., 1996), we hypothesized that an apoptotic mode of neuronal loss is induced by NO. 427
428
S. HLI(,Id.
EXPERIMENTALPRO(’ED[JRFX
E.Yp<,/im<,/?/( I[pfoto((]l”
C~ytcjkille-nlcdiiltccineurotoxicity, WM assessed f’ollowirrg treatment of mixed neurona;gliai cell cultures with IL-I/; (10 rrgjml), IFN-; (200 Uml), or a combitl:ition of these cytc>Primary, human fetal brain cell cultures were prepared kines for 13 d:iys. To evaluate the in~olyemcmt of NO, cell using a rcccntly developed technique under a protocol cultures were treated with 500”;tM of the NO synthusc inhihiappl-o~ed by our institutional Human Subjects Research tor N{;-monomethyl-l-argininc (NM MA) silllLllt:tnec>~isiy C’mnmittcc (Chao ond Hu, 1994). Brielly, brain corticcs from when IL- I[~ and 1FN-; were added. Cullurc supernatants 14 1X-week-old abortuses were treated with ().25!L0 trypsin for 30 min. Afler extensive washing. dispersed brain cells were then harvested for nitrite (NO: ) determination and LDH ;Lctivity. WCI-C plated onto collagen-coated 24-well plates at a conTo investigate whether the mechanism cytokine-mecfiatcd centration of 5 x 105 cellsj500 ~~1and cultured in DMEM neurotoxicity involved an apoptotic mode of neuronal death. (Signm, St Louis, M(). U. S.A.) containing 10(YO fetal bovine two approaches were used: lerminal transferase-mediated serum (F fitS) at 37 C in a humidified 10’J/O C’02atrnosphcrc. UTP nick end Iabcl]ing (TUNEL) technique and DNA fragOn d~iy 5. (IICcullurc medium was replaced with DMEM containing I()”/. FBS, uridirrc (33.6 pgjml, Sigma), and ffuc>- mentation, EXJLTl1112Lt]OLl 01’ ncurorral death by these techniques was performed on duys 7, I(). and 13 following rodcoxyuricfine (13.6 ~g~m], Sigma) to prevent overgrowth cytokine tre~tmcnt. of astroglial cells. Culture medicrlm with 10(YOFBS was replaced on day 6 and every 4 days thereafter. On day 12, DNA In situ cnd Itrl)(,llitl,q(Jffiu,y}?!c,tl[c({ the neurcrmd cell cultures contained differentiated neurons ?-Wwltlgon a supporting layer of’astrocytes. The mixed new For i~t .Iitu detection of’nuclear DNA frtigrnentation. the romd and glial cell cultures consisted of about 50–60”/. neuT[JNEL technique was used as described by Gavrieli e/ u/. rons (stainecf with anli-neurou specific enolase antibodies; (1992) with minor modifications, In brief, mixed new Polysclences, Barrington. PA, U.S.A.), 40–50”/0astrocytes ronal~glial cell cultures were fixed for 20 min in 40/0 par(stained with anti-glial fribrillary acicficprotein (GFAP) mltiafornmdehyde at room temperature. C’cllcultures were then body), <2% nlicrogli;L (stained with anti-CD6X antibody; washed f’oul-times with phosphutc-huffered saline. Endogem Dako), und < 1“/0 oligodendrocytes (stained with antious peroxidase was inactivated by incubating cell cultures in galactocerehmside antibodies; Boehringcr Mannheim, 2’Yo1{20, for 5 min. Afler preincub~tion with TDT hufier Indianapolis. IN. U.S.A.). These 12-day mixed necl- (30 mM Trizam base, pH 7.2, 140mM sodium cacodykite, ronai glial cell cultures were used in all experiments in which I mM cobalt chloride) for 5 min at room temper:iture. CCII neumtnxicit) was eva]uuted. cultures were subjected to nick translation procwfurcs f’ollnwed by exposure to TDT hufi’er containing 0.5 LHml ter/f,\//()(,\/[, IL(/lL{/(>,$ minal tmnsferase (Boehringer Mannheim) ~lnd 0.5 ;mol biotinylated 16-dUTP (Boehringer Manuheim) in a humid Human fetal brain tissues were obtained from 16 22-week- atmosphere f’or60 min at 37 C. The reac(ion was terminated nld aborted Ictuses. Primary astrocyte cultures were prepared by incubating in TB buffer (300”mM sodium chlnridc, 30 mM iLs previously described (Chao [,/ a[., 1996). In brief, brain sodium citmtc) for 15min at room tcmpcruturc. After washitissues were cleared of mcninges, minced into small fragng four tirncs with TDT huffcl-.CCIIcultures were incuhated ments, and incubated with 0.25°/0 trypsin (Sigma) in Caz+with the ABC Elite Kit (Vector Laboratories, Burlingame. and Mg2+-free puck’s saline (Gibco Lahoratorics, Grand CA, U.S.A.) for 45 min at rooln temperature followed by Isktnd, NY, U.S.A.) for 30 min at 37 C’with gentle shuking. color development with 3,3’-diaminohezidone. An equal VOILUIIC of mcdiurn containing 10“/. heat-inactivated FBS was :iddcd to iuactivale the trypsin, and tissues DNA (,.~tracfiotl were centrifuged ut 1200rpm for 5 min. After washing twice DNA was extracted from mixed ncuronal’gliul cell culwith Hank’s boffcrcd saline solution (Gibco). tissue fragments were replaced with fresh medium and triturated f’or tures using DNA Isolation Kits (Gcntra Systems, Inc.. Minneapolis, MN, U.S.A.). In brief. cell cultures were Iysed with 15-20 passages thruugb a sterile Pasteur pipette. Cell susLysing huffer followed by ~Nase tre:itment for 60 min at pensions were seeded into 75 cm’ flasks It a density of 80 37’’C’, prntcin prccipitution solution, and centrifuged at 100x 10” cells~flask in a high-glucose DMEM containing 15.()()()gfor 3 min. DNA wus pwpwxl by additicrn 01 1000/” 10!4 FBS, penicillin (100 U~ml), and streptomycin isopropanol followed by ccntrifugatiorr ~t 15,000,< for I min. (l f)OJ[g,’ml).These flmks then were incubtitcd al 37 C in Pellets were wushed with 70(%,ethanol and samples were air a w~iter-s~itu rittecl, I()% C02 incubator. After 24 ~lnd 4Xh dried. incubation. medium was replaced with fresh DMEM containing 10”/. FBS. Culture medium WJSchanged once weekly A,qcw(m,qc[(,[<,cfI(,pl?(II(,.Y/.s thereafter. On day 21, flasks were shaken at 180–200rpm for 16h. followed by washing with Ca2i - and Mg2”+-freeHank’s DNA samp[es ( I()J(g) were loaded ontc 1.X(YO ~garose gel buflered saline solution cont:iining O.125%,trypsin for 30 min contiiining cthidium bromide (0.4 mg~m}) :ind electrm at 37 C. followed by >~ddition of 10“/o FBS-containing phonetically sep~ir:itcd. DNA was kisuali~ed by a UV Imedium. After ccntrifugation. cell suspensions were seeded (302 nM) trtinsilluminator, Jnd the gels were photngr~ipbcd into a new flask in DMEM containing 10“4 FBS, and the with a polaroid c:inwra. culture medium was changed 24 h later. This procedure was repeated [hrce times at weekly intervals, Finally, highly ,Yl(itis{icul(I)u[r$i$ enriched astmcytcs ( > 98°/0 of cells st:iirwd with anti- GF’AP Student’s ~-test(Statview 11,Abacus (’oncepts, Inc., Berkemtihody) were seeded into 24-wc1lplales a~ a density of 105 ley. C-A,U.S.A.) was used to compare the difl’erencesbetween cells ‘well fnr experiments. A’fi. Ye(l Il[,ul’oll(l[,q[i(l[ C(I1cL[/[L1r(,.\’
Cytokine-mediated neuronai apoptosis the means of two groups. Analysis of variance followed by Scheffe’s F-test was used to compare the differences among multiple means. RESULTS
Cytokine-rnediated neuronal death Treatment of mixed neuronal/glial cell cultures with IL-1/l induced a significant (P< O.01) amount of NO reflected by accumulation of the stable metabolize NO; (Fig. l(A)). This was accompanied by a slight release of LDH (Fig. l(B)). IFN-y alone had no effect on NO; production or neuronal LDH release (Fig. 1). Treatment of neuronal/glial cell cultures with IL1~ plus IFN-y for 13 days resulted in elevated (P< O.01)NO; levels (Fig. l(A)), which corresponded with marked neuronal damage (f’< 0.01) as assessed by increased LDH release into culture supernatants (Fig. l(B)). The NO synthase inhibitor NMMA suppressed the elevated NO; levels (Fig. l(A)) as well as LDH release induced by these cytokines (Fig. l(B)),
A
IL-113+ IFN-y
60150 i
i
‘L-“
B
60
[
IL-113+ IFN-y
T
+ - NMMA Fig. 1. Cytokine-mediated NO production and neuronal injury. Mixed human neuronal/glial cell cultures were treated with medium (control), IL-1P (10 rig/ml), IFN-y (200 U/ml) and IL-1P plus IFN-y in the absence or presence of the NO synthase inhibitor NMMA (500 pM). Supernatants were harvested after 13 days of exposure for determination of (A) NO; levels and (B) LDH release. Data are mean~SEM of duplicates and representative of three separate experiments. Control
IFN-y
IL-1 13
429
suggesting the involvement of NO in cytokinemediated neuronal injury within mixed human neurenal/glial cell cultures. In a separate experiment, similar NOq- levels were observed with treatments of highly (>98%) enriched astrocyte cultures with IL1~ or IL-1P plus IFN-y for 13 days; however, little LDH was detected over the control astrocyte cultures, suggesting that astrocytes are more resistant to cytokine-mediated injury (data not shown). Cytokine-mediated neuronal apoptosis To determine the mode of cell death, TUNEL staining for condensed chromatin in pyknotic nuclei was used. Using this technique, 5–70/. of cytokine-treated neuronal cells began to display apoptosis after 7 days of exposure to cytokines. The number of apoptotic neurons increased by 10 days (l&l 50/0).Exposure of neuronal/glial cell cultures to IL-1P and IFN-y for 13 days markedly increased the number of apoptotic neurons (25–300/0), reflected by microscopic examination (Fig. 2(B)) as compared to control cell cultures (Fig. 2(A)). NMMA treatment attenuated (P< O.01) cytokine-induced neuronal apoptosis from 25–30°/0to 3–5~0 (Fig. 2(C)). In highly (> 98Yo)enriched astrocyte cultures, cytokine-induced apoptotic cell death after 10 days of exposure was minor ( < 50/.), suggesting that astrocytes are more resistant to cytokinemediated apoptosis (data not shown). To further confirm that cytokine-mediated neuronal cell death involved an apoptotic mechanism, DNA degradation was assessed by gel electrophoresis (Fig. 3). Exposure of mixed neuronal/glial cell cultures to IL-1P plus IFN-y for 7 days induced a minor increase in large molecular weight fragments (data not shown). After 13 days of treatment with these cytokines, cell cultures displayed. DNA degradation with markedly increased large and small molecular weight fragments (Fig. 3). Treatment of cell cultures with NMMA attenuated cytokine-induced DNA fragmentation (Fig. 3), suggesting an involvement of NO in cytokine-mediated neuronal apoptosis.
DISCUSSION
We demonstrated in the present study that cytokine-induced neurotoxicity, which was partially blocked by the NO synthase inhibitor NMMA, proceeded by an apoptotic pathway as demonstrated by nuclear TUNEL staining technique and DNA fragmentation experiments. These findings strongly support the hypothesis that cytokine-mediated
430
S. Hu e( a[.
.,,
A
#?
‘%
c
4$
‘.
Fig. 2. Morphological assessment of’ apoplosis of neurons induced by cytokines. Mixed neurorral~glial cell cultures were treated with (A) mccfium(Cmrlrol), (B) IL-f~ (10 ng~ml)plus IFN-~ (200 [J/ml) or (C) NMMA (500 ~1M)plus cytokines for 13days prior to nuclear staining with the TUNEL technique. Data arc dm-ivedfrom three separate brain cell cultures.
neurotoxicity involves NO and that this free radical induces programmed Almost a century
cell death of neurons. ago, it was observed that mic-
roglia migrate to, and differentiate and proliferate at, sites of inflammation within the brain (del Rio-Hortega, 1932). It has been proposed that these glial cells play a neuropathogenic role (Chao etal., 1995a). In contrast to studies with murine microglia, cytokineactivated human microglia produce little NO that is incapable of inducing neurotoxicity in vitro (Peterson
Fig. 3. DNA fragmentation. Mixed neuronal~glial cell cultures were treated with medium (control, lane 2), IL-11~ (10 ngml) plus IFN-y (200 Ujml) (lane 3) or NMMA (500 PM) plus cytokincs (kmc 4) for 13 days prior to harvesting of DNA, Molecular marker was loaded into lane 1. DNA (l Opg) from each trcutmcnt group wus loaded into each agarose gel lane prior to clcctrophoresis. Data arc representative of three separate experiments.
et al., 1994). Human neurons are susceptible, however, to NO neurotoxicity as evidenced in human mixed neuronal/glial cell cultures incubated with activated murine microglia (Peterson et al., 1994). Of note, involvement of NO in immunologic brain injury has been implicated in brain lesions of patients with mul-
Cytokine-mediated neuronal apoptosis
tiple sclerosis (Brosnan et al., 1994). Recently, we have found that cytokine-stimulated human astrocytes are capable of generating substantial amounts of neurotoxic NO (Chao et al., 1996). While the precise mechanism underlying cytokine-mediated neuronal death has not been elucidated, in the present study we demonstrated that an apoptotic mode of neuronal death is involved. Because NO generation by cytokine-treated astrocytes is a slow process, it is not surprising that we found an apoptotic mode of neuronal injury in human mixed neuronal/glial cell cultures. Recent studies using rat cortical neuronal cultures have shown that low levels of NO mediate a delayed apoptotic mode of neuronal injury (Bonfoco et al., 1995). The observation that NMMA only partially (approximately 5@60%) attenuated cytokinemediated neuronal injury, suggests that NO is not the only mediator responsible for cytokine-mediated neurotoxicity. Nonetheless, results of the present study may lead to development of new therapeutic strategies for neurodegenerative diseases in which activated glia, cytokines and NO appear to be involved, such as multiple sclerosis, Alzheimer’s disease, and AIDS dementia. Acknowledgements—This study was supported in part by USPHS grants DA-04381, DA-09924 and a grant from the Alzheimer’s Association.
REFERENCES
Boje, K. M. and Arora, P. K. (1992) Microglia-produced nitric oxide and reactive nitrogen oxides mediate neuronal cell death. Brain Res. 587, 25&256. Bonfoco, E., Krainc, D., Ankarcrona, M., Nicotera, P. and Lipton, S. A. (1995) Apoptosis and necrosis: two distinct events induced, respectively, by mild and intense insults with N-methyhD-aspartate or nitric oxide/superoxide in cortical cell cultures. Proc. natn Acad. Sci. U.S.A. 92, 7162-7166. Brosnan, C. F., Battistina, L., Raine, C. S., Dickson, D. W., Casadevall, A. and Lee, S. C. (1994) Reactive nitrogen intermediates in human neuropathology: an overview. Devl Neurosci. 16, 152–161. Chao, C. C., Hu, S., Molitor, T. W., Shaskan, E. G. and Peterson, P. K. (1992) Activated microglia mediate neu-
431
ronal cell injury via a nitric oxide mechanism. J, Immunol. 149, 273&2741. Chao, C. C. and Hu, S. (1994) Tumor necrosis factor-u potentates glutamate neurotoxicity in human fetal brain cell cultures. Devl Neurosci. 16, 172–179. Chao, C. C., Hu, S. and Peterson, P. K. (1995) Glia, cytokines, and neurotoxicity. Crit. Rev. Neurobiol. 9, 189-205. Chao, C, C., Hu, S,, Sheng, W. S. and Peterson, P. K. (1995) Tumor necrosis factor-alpha production by human fetal microglial cells: regulation by other cytokines. Devl Neurosci. 17, 97–105. Chao, C. C., Hu, S., Sheng, W. S., Bu, D-F., Bukrinsky, M. I. and Peterson, P. K, (1996) Cytokine-stimulated astrocytes damage human neurons via a nitric oxide mechanism. Glia 16,276-284, del Rio-Hortega, P. (1932) Microglia. In Cytology and Cellular Pathology of the Nervous System, ed. W. Penfield, Vol. 2, pp. 481–584. Paul P. Hocker, New York. Dickson, D. W., Mattiace, L, A,, Kure, K., Hutchins, K., Lyman, W. D. and Brosnan, C. F. (1991) Microglia in human disease, with an emphasis on acquired immune deficiency syndrome. Lab. Invest. 64, 135–156. Eddleston, M, and Mucke, L, (1993) Molecular profile of reactive astrocytes—implications for their role in necrologic disease. Neuroscience 54, 15–36. Fontana, A., Constam, D., Frei, K., Koedel, U., Pfister, W., and Weller, M. (1996) Cytokines and defense against CNS infection. In Cytokines and the CNS, ed. R. M. Ransohoff and E. N, Benveniste, pp. 187–220. CRC Press, Boca Raton, Florida. Gavrieli, Y., Sherman, Y. and Ben-Sasson, S. A. (1992) Identification of programed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Cell Biol. 119, 493–501. Hu, S., Sheng, W. S., Peterson, P. K. and Chao, C. C. (1995) Differential regulation by cytokines of production of nitric oxide by human astrocvtes. Glia 15, 491494, Lee, S. C:, Liu, W., Dickson, D. W, Brosnan, C. F. and Berman, J. W. (1993) Cytokine production by human fetal microglia and astrocytes: differential induction by lipopolysaccharide and IL-1/J. J.Immunol. 150, 2659–2667. Lee, S. C., Dickson, D. W., Liu, W, and Brosnan, C. F. (1993) Induction of nitric oxide synthaseactivity in human astrocytes by interleukin-1~ and interferon-gamma. J. Neuroimmunol. 46, 19–24. Lewis, D. B. and Wilson, C. B. (1990) Gamma-interferon: an immunoregulatory Iymphokine with immunotherapeutic potential. Pediatr. infect. Dis. J. 9, 642–651. Merrill, J. E. and Chen, I. S. Y. (1991) HIV-1, macrophages,glial cells, and cytokines in AIDS nervous system disease. FASEB J. 5, 2391–2397. Peterson, P. K., Hu, S., Anderson, W. R. and Chao, C. C. (1994) Nitric oxide production andneurotoxicity mediated by activated microglia from human versus mouse brain. J. infect. Dis. 170, 457J160.