Abstract
Vaccination with peptides derived from interphotoreceptor retinoid-binding protein (a self-Ag that can cause experimental autoimmune uveoretinitis) resulted in protection of retinal ganglion cells from glutamate-induced death or death as a consequence of optic nerve injury. In the case of glutamate insult, no such protection was obtained by vaccination with myelin Ags (self-Ags associated with an autoimmune disease in the brain and spinal cord that evokes a protective immune response against consequences of injury to myelinated axons). We suggest that protective autoimmunity is the body’s defense mechanism against destructive self-compounds, and an autoimmune disease is the outcome of a failure to properly control such a response. Accordingly, the specific self-Ag (although not necessarily its particular epitopes) used by the body for protection against potentially harmful self-compounds (e.g., glutamate) can be inferred from the specificity of the autoimmune disease associated with the site at which the stress occurs (irrespectively of the type of stress) and is in need of help.
Axonal injury in the CNS leads to an inevitable process of degeneration, not only in the afflicted axons, but also in neighboring axons that escaped the initial insult (1, 2, 3). This secondary degeneration has been attributed to self-destructive compounds that emerge from the degenerating axons into the microenvironment at the lesion site, making it hostile to the remaining tissue (4, 5, 6, 7, 8). We recently discovered that CNS myelinated axons, after suffering a mechanical insult such as a crush injury, can benefit from the activity of autoreactive T cells directed against myelin Ags (9, 10, 11, 12). We further found that the neuroprotective activity exhibited by these autoimmune T cells is not merely the result of an experimental manipulation, but is a physiological way in which the body copes with stressful conditions (13, 14, 15). Accordingly, we proposed that just as the immune system is called upon to defend the body from invading microbes, it is also needed to protect it from self-compounds that under conditions of trauma or stress (not necessarily related to pathogens) become toxic. Interestingly, in the case of damage to myelinated CNS axons, the T cells that induce neuroprotection have the same specificity and phenotype as those known to cause autoimmune disease. Thus, the cells are both potentially protective and potentially destructive, and their actual expression evidently depends on how they are regulated. This might explain the observed correlation between the ability to manifest an autoimmune response with a beneficial outcome and the ability to resist the development of an autoimmune disease (13). Therefore, the ability to protect neuronal tissue apparently does not correspond to a lack of autoimmunity, but, rather, reflects autoimmunity that is well controlled.
We were interested in investigating whether the T cells recruited in the specific environments of different injury sites for the purpose of coping with the local stressful situation have the same or different antigenic specificities. Our previous work indicated that although passive transfer of anti-myelin autoimmune T cells (11) or vaccination with myelin Ags (16) can protect retinal ganglion cells (RGCs)3 after an insult to the optic nerve axons, these procedures are not protective after a direct insult to the RGCs themselves (17). This finding led us to consider the possibility that each tissue has its own specific self-Ags that signal the immune system when the tissue needs help. In the case of axotomy, since the Ags that send signals summoning the immune system to the aid of the stressed neurons are myelin proteins associated not with neurons but with oligodendrocytes, we considered the possibility that the relevant Ags are not necessarily expressed on the cells that require assistance but on other cells in the vicinity. In addition, if an autoimmune disease is indeed the outcome of failure to control an autoimmune response whose original purpose was beneficial, it seems reasonable to postulate that the protection (beneficial response) and the disease (destructive response) share the same antigenic specificity.
Experimental autoimmune uveoretinitis (EAU) is an experimental model for uveitis, a T cell-mediated autoimmune disease of the eye (18, 19, 20, 21). In the present study we tested our working hypothesis (namely, that the protective and the destructive autoimmune response share the same antigenic specificity) by investigating whether a pathogenic, uveitis-related, retinal self-Ag can protect against direct and indirect insults to the RGCs. The results showed that RGCs exposed to a glutamate insult or suffering the secondary consequences of an optic nerve crush injury could be protected by vaccination with a uveitis-associated peptide. In the case of glutamate insult, protection was observed under conditions where no such protective effect could be obtained by vaccination with myelin Ags.
Materials and Methods
Animals
Adult male Sprague Dawley (SPD), Fisher (F344), and Lewis rats (8–12 wk old) and adult female Lewis rats (16–18 wk old) were supplied by the Animal Breeding Center of the Weizmann Institute of Science under germfree conditions. The rats were housed in a light- and temperature-controlled room and were matched for age in each species for each experiment. Animals were handled according to the regulations formulated by the institutional animal care and use committee.
Antigens
The peptides R16 (sequence 1177–1191 of bovine interphotoreceptor retinoid-binding protein (IRBP), ADGSSWEGVGVVPDV), G-8 (sequence 347–354 of human retinal soluble Ag (S-Ag), TSSEVATE), a G-8 analog (TSSEAATE), M-8 (sequence 307–314 of human retinal S-Ag, DTNLASST), and an M-8 analog (DTALASST) were prepared in the Synthesis Unit at The Weizmann Institute. G-8 and M-8 are uveitogenic, while their analogs are immunogenic, but not immunopathogenic (22).
Partial crush injury of the rat optic nerve
The optic nerve was subjected to a well-calibrated crush injury as previously described (3). Briefly, rats were deeply anesthetized by i.p. injection of Rompun (xylazine, 10 mg/kg; VMD, Arendonk, Belgium) and Vetalar (ketamine, 50 mg/kg; Fort Dodge Laboratories, Fort Dodge, IA). Using a binocular operating microscope, lateral canthotomy was performed in the right eye, and the conjunctiva was incised laterally to the cornea. After separation of the retractor bulbi muscles, the optic nerve was exposed intraorbitally by blunt dissection. Using calibrated cross-action forceps, the optic nerve was subjected to a severe crush injury 1–2 mm from the eye. The contralateral nerve was left undisturbed.
Glutamate injection
The right eye of the anesthetized rat was punctured with a 27-gauge needle through the conjunctiva and sclera, anterior to the pars plana so that the retina was untouched, and a 10-μl Hamilton syringe (Reno, NV) with a 30-gauge needle was inserted as far as the vitreal body. Rats were injected with 2 μl (400 nmol) of l-glutamate.
Active immunization
Rats were subjected to optic nerve crush injury and then immediately immunized by s.c. injection at the base of the tail of R16 (30 μg), G-8, G-8 analog, M-8, or M-8 analog (200 or 500 μg) emulsified in CFA supplemented with 2.5 mg/ml Mycobacterium tuberculosis (Difco, Detroit, MI) in a total volume of 0.1 ml. Rats in another group were exposed to a glutamate insult (by intravitreal glutamate injection) and then immediately immunized s.c. at the base of the tail with 30 μg of R16 emulsified in CFA supplemented with 2.5 or 0.5 mg/ml of M. tuberculosis in a total volume of 0.1 ml. Control rats were injected with PBS in CFA. In another set of experiments rats were actively immunized with 30 μg of R16 emulsified in CFA supplemented with 2.5 mg/ml M. tuberculosis 1 wk before the crush injury and given a booster of 30 μg of R16 emulsified in IFA (Difco) immediately after the injury. Control rats were injected with PBS in CFA and boosted with PBS in IFA.
Passive immunization
Male Lewis rats were unilaterally injected in the hind footpads with 30 μg of R16 emulsified in CFA supplemented with 2.5 mg/ml of M. tuberculosis in a total volume of 0.1 ml. Seven days after immunization, spleens from immunized and naive rats were removed and pooled in ice-cold PBS. A single-cell suspension was prepared, and the cells (2 × 106 cells/ml) were cultured with naive thymocytes (2 × 106 cells/ml) in the presence of R16 (20 μg/ml) in proliferation medium containing DMEM supplemented with 2 mM glutamine, 2-ME (5 × 10−5M), sodium pyruvate (1 mM), nonessential amino acids (1 ml/100 ml), 1% fresh autologous rat serum, 100 U/ml of penicillin, and 0.1 mg/ml of streptomycin. After incubation for 72 h the cultures were collected and washed, and the lymphoblasts (1.3 × 107 cells in 3 ml PBS) or PBS alone (3 ml) were injected i.p. into Lewis rats immediately after optic nerve crush injury.
Assessment of secondary degeneration in the rat optic nerve by retrograde labeling of retinal ganglion cells
Secondary degeneration of optic nerve axons and their corresponding RGCs was evaluated after application, 2 wk after injury, of the fluorescent lipophilic dye 4-(4-(didecylamino)styryl)-N-methylpyridinium iodide (4-Di-10-Asp; Molecular Probes Europe, Leiden, The Netherlands), distal to the lesion site. The right optic nerve was exposed for the second time, again without damaging the retinal blood supply. Complete axotomy was performed 1–2 mm distal to the injury site, and solid crystals (0.2–0.4 mm in diameter) of 4-Di-10-Asp were deposited at the site of the new axotomy. Five days after dye application the rats were killed. Retinas were detached from the eyes, prepared as flattened whole mounts in 4% paraformaldehyde solution, and examined for labeled RGCs by fluorescence microscopy and confocal microscopy. Since only intact axons can transport the dye back to their cell bodies, application of the dye distal to the lesion site 2 wk after injury ensures that only axons that survived both primary damage and secondary degeneration will be counted. This approach enables us to differentiate between neurons that are still functional and neurons in which the axons are injured but the cell bodies are still viable.
Spinal cord contusion
Female Lewis rats were anesthetized by i.p. injection of Rompun and Vetalar, and their spinal cords were exposed by laminectomy at the level of T8. One hour after induction of anesthesia, a 10-g rod was dropped onto the laminectomized cord from a height of 50 mm using the NYU impactor, a device shown to inflict a well-calibrated contusive injury of the spinal cord (10, 23, 24, 25).
Active immunization
Rats were immunized s.c. on a random basis with 100 μg of R16 or injected with PBS, each emulsified in CFA supplemented with 0.5 mg/ml M. tuberculosis, in a total volume of 0.1 ml. Rats were immunized within 1 h after contusion.
Animal care
In contused rats bladder expression was assisted by massage at least twice a day (particularly during the first 48 h after injury, when it was performed three times a day) throughout the experiment. All rats were carefully monitored for evidence of urinary tract infection or any other sign of systemic disease. During the first week after contusion and in any case of hematuria after that period, they received a course of sulfamethoxazole (400 mg/ml) and trimethoprim (8 mg/ml; Resprim; Teva Pharmaceutical Industries, Ashdod, Israel), administered orally with a tuberculin syringe (0.3 ml solution/day). Daily inspections included examination of the laminectomy site for evidence of infection and assessment of the hind limbs for signs of autophagia or pressure.
Assessment of recovery from spinal cord contusion
Behavioral recovery was scored in an open field using the locomotor rating scale of Basso, Beattie, and Bresnahan, where a score of 0 registers complete paralysis and a score of 21 indicates complete mobility (23, 26, 27). Blind scoring ensured that observers were not aware of the treatment received by individual rats. Approximately once a week the locomotor activities of the trunk, tail, and hind limbs were evaluated in an open field by placing each rat for 4 min in the center of a circular enclosure (90 cm in diameter, 7-cm wall height) made of molded plastic with a smooth, non-slip floor. Before each evaluation the rats were examined carefully for perineal infection, wounds in the hind limbs, and tail and foot autophagia (24, 28).
Results
Uveitogenic peptide derived from IRBP protects against glutamate-induced RGC loss
To test our working hypothesis, we first investigated whether RGCs can be protected by vaccination with a self-peptide associated with uveitis, an autoimmune disease affecting the eye. The peptide selected for this experiment was R16 (29, 30), an immunodominant sequence within IRBP known to cause uveitis. First we examined whether vaccination with R16 could protect the RGCs of Lewis rats (a strain susceptible to autoimmune disease induction) from glutamate toxicity under conditions where immunization with myelin peptides is not effective (17). Vaccination of Lewis rats with R16 after a glutamate insult indeed resulted in a reduced loss of RGCs (Fig. 1⇓). Relative to normal retinas, the percentage of RGC loss (mean ± SEM) was 14 ± 2% in rats vaccinated with R16 emulsified in CFA compared with 28 ± 4% in rats treated with PBS in CFA (p < 0.04). This finding substantiates our contention that immune protection requires the activity of T cells specific to Ag present within the injured tissue. Since R16 is known to cause uveitis in Lewis rats (but not in SPD or Fisher rats), it was interesting to discover that the RGCs received protection despite massive infiltration of lymphocytes into the eyes of these rats. It is important to mention, however, that the neuroprotective effect of R16 in this model was detected only when the disease in these rats was mild (i.e., when the amount of bacteria in the adjuvant was 0.5 mg/ml). Immunization of glutamate-injected rats with R16 emulsified in CFA at a concentration of 2.5 mg/ml was not protective and even caused additional neuronal loss compared with that seen in rats immunized with PBS in CFA (data not shown). These findings constitute further evidence of the delicate balance between the processes of destruction and protection attributable to these specific autoimmune T cells.
Immunization with R16 protects RGCs from glutamate toxicity in Lewis rats. RGCs of adult Lewis rats were exposed directly to glutamate toxicity by intravitreal injection of l-glutamate (400 nmol). Immediately thereafter the rats were immunized with 30 μg of R16 emulsified in CFA (0.5 mg/ml). Control rats were injected with PBS in CFA. Two weeks later the optic nerves were exposed for the second time, and the fluorescent dye 4-Di-10-Asp was applied distally to the injury site. Five days after dye application the retinas were detached from the eyes and prepared as flattened whole mounts. Labeled RGCs from four randomly selected fields of identical size in each retina (all located at approximately the same distance from the optic disk) were counted under the fluorescence microscope, and the RGC loss was calculated and expressed as mean percentage ± SEM. The percentage of loss was significantly smaller in the R16-immunized rats than in their matched PBS-injected controls (14 ± 2 and 28 ± 4%, respectively; p < 0.04, by two-tailed t test). Each group consisted of five or six rats.
Uveitogenic peptide derived from IRBP protects retinal ganglion cells from the consequences of optic nerve injury
Next we examined the effectiveness of R16 vaccination in protecting RGCs from secondary degeneration after optic nerve crush, an insult known to trigger secondary degeneration in the cell bodies or axons of neurons that escaped direct injury (31). This examination was conducted in the two resistant rat strains (SPD and Fisher) and in the susceptible strain (Lewis). In all three strains vaccination with R16 emulsified in CFA (with the high bacterial content of 2.5 mg/ml) on the day of injury significantly reduced the injury-induced loss of RGCs (Figs. 2⇓ and 3⇓). In SPD rats the number of surviving RGCs per square millimeter (mean ± SEM) was 150 ± 13 in rats immunized with R16 in CFA and 60 ± 14 in rats injected with PBS in CFA (p < 0.01; Fig. 2⇓). The corresponding results were 183 ± 16 and 114 ± 9, respectively, in Fisher rats (p < 0.01; Fig. 2⇓), and 192 ± 8 and 73 ± 10, respectively, in Lewis rats (p < 0.0001; Fig. 3⇓, A–C). Based on the antigenic specificity found in the case of glutamate toxicity, we attributed the dramatic protection of RGCs observed after R16 vaccination in the crush model to protection by T cells that had migrated to the retina and become activated there, rather than to protection adjacent to the lesion site.
Immunization of Fisher and SPD rats with R16 immediately after optic nerve injury protects their retinal ganglion cells from secondary death. Adult Fisher and SPD rats were subjected to partial optic nerve crush injury. Immediately thereafter, the rats were immunized with 30 μg of R16 emulsified in CFA (2.5 mg/ml). Control rats were injected with PBS in CFA. Staining with 4-Di-10-Asp, preparation of retinal slides, and counting of labeled RGCs were as described for Fig. 1⇑. The average number of RGCs per square millimeter was calculated. Significantly more RGCs (mean ± SEM per square millimeter) survived in the R16-immunized injured rats than in their matched PBS-injected controls (150 ± 13 and 60 ± 14, respectively (p < 0.01, two-tailed t test), for SPD rats; 183 ± 16 and 114 ± 9, respectively (p < 0.01, two-tailed t test), for Fisher rats). Each group consisted of five or six rats.
Immunization of Lewis rats with R16 immediately after optic nerve injury protects their retinal ganglion cells from secondary death. Adult Lewis rats were subjected to partial optic nerve crush injury. Immediately thereafter, the rats were immunized with 30 μg R16 emulsified in CFA (2.5 mg/ml). Control rats were injected with PBS in CFA. Staining with 4-Di-10-Asp, preparation of retinal slides, and counting of labeled RGCs were as described for Fig. 1⇑. A, The average number of RGCs per square millimeter was calculated. Significantly more RGCs (mean ± SEM per square millimeter) survived in the R16-immunized injured rats than in their matched PBS-injected controls (192 ± 8 and 73 ± 10, respectively; p < 0.0001, by two-tailed t test). Each group consisted of five or six rats. B and C, Representative fluorescence micrographs of PBS-injected injured Lewis rats (B) and R16-immunized injured Lewis rats (C). D, Survival of RGCs in Lewis rats after optic nerve injury and passive transfer of splenocytes from R16-immunized rats. As controls we used Lewis rats injected with PBS or naive splenocytes after optic nerve injury. Each group consisted of five or six rats.
To verify that the observed protection is mediated by T cells, we transferred R16-activated splenocytes to optic nerve-injured Lewis rats. Passive transfer of splenocytes from R16-immunized Lewis rats to nonimmunized Lewis rats immediately after optic nerve injury resulted in a higher number of surviving RGCs per square millimeter (mean ± SEM) in the recipient rats (92 ± 19 compared with 53 ± 4 in PBS-injected rats and 57 ± 6 in rats injected with naive splenocytes; p < 0.03 for the comparison of recipient rats with pooled controls, by ANOVA; Fig. 3⇑D).
It is interesting to note that the clinical onset of EAU in Lewis rats occurred on day 10 after immunization, and inflammation peaked on day 14. Thus, at the time of assessment of neuronal survival in the crush injury model the EAU disease in Lewis rats was still severe. We were therefore interested in knowing whether under such conditions immunization by itself would have an adverse effect on RGC survival in Lewis rats despite the overall protection. We therefore examined whether R16 immunization in the absence of insult causes any RGC loss in Lewis rats. The percentage of RGC survival (mean ± SEM) was significantly lower in the R16-immunized Lewis rats than in their matched PBS-injected controls (83 ± 5%; p = 0.02, by one-tailed t test). Immunization with R16 did not affect RGC survival in Fisher rats (97 ± 8% survival). Thus, some loss of RGCs was evident 2 wk after R16 vaccination in Lewis, but not in Fisher rats, suggesting that uncontrolled autoimmunity leading to autoimmune disease can indeed be destructive in a susceptible strain, but that even in this strain the beneficial effect of autoimmunity on neuronal survival exceeds its destructive effect, so that the net outcome is favorable. In resistant rats controlled autoimmunity allows the beneficial effect of autoimmunity to be expressed under a wider range of conditions. Support for this suggestion comes from the finding that in the resistant Fisher rats, unlike in the susceptible Lewis rats, vaccination 1 wk before injury resulted in significant protection (Fig. 4⇓).
Immunization of Fisher rats (but not Lewis rats) with R16 1 wk before optic nerve injury protects their RGCs from secondary death. Adult Fisher and Lewis rats were immunized with 30 μg of R16 emulsified in CFA (2.5 mg/ml). Control rats were injected with PBS in CFA. One week later the rats were subjected to partial optic nerve crush injury and immediately thereafter were given a booster injection of 30 μg of R16 emulsified in IFA. Control rats were injected with PBS in IFA. Staining with 4-Di-10-Asp, preparation of retinal slides, and counting of labeled RGCs were as described for Fig. 1⇑. The average number of RGCs per square millimeter was calculated. Significantly more RGCs (mean ± SEM per square millimeter) survived in the R16-immunized injured Fisher rats than in their matched PBS-injected controls (165 ± 22 and 89 ± 10, respectively; p < 0.01, by two-tailed t test). The difference observed between the R16-immunized and PBS-injected Lewis rats (117 ± 21 and 95 ± 22, respectively) was not statistically significant. Each group consisted of five or six rats.
Peptides derived from S-Ag protect against retinal ganglion cell loss as a consequence of optic nerve injury
To gain further support for the idea that the protective response is Ag specific we used two additional uveitogenic epitopes (G-8 and M-8) of another retinal autoantigen, S-Ag, and their immunogenic (but not immunopathogenic) analogs. As with R16, vaccination with the uveitogenic peptides G-8 and M-8 or their immunogenic analogs immediately after optic nerve crush injury resulted in a significant increase in RGC survival in Fisher rats. The numbers of surviving RGCs per square millimeter (mean ± SEM) were 159 ± 5, 153 ± 10, and 159 ± 19 in rats immunized with 200 μg G-8, M-8, or M-8 analog in CFA and 109 ± 12 in rats injected with PBS in CFA (p < 0.01, p < 0.03, and p < 0.04, respectively; Fig. 5⇓A). In the case of the G-8 analog, immunization with 500 μg (but not with 200 μg) of the peptide resulted a significant increase in RGC survival compared with that in rats injected with PBS in CFA (175 ± 15 and 90 ± 11, respectively; p < 0.01; Fig. 5⇓B).
Immunization of Fisher rats with G-8, G-8 analog, M-8, or M-8 analog immediately after optic nerve injury protects their RGCs from secondary degeneration. Adult Fisher rats were subjected to partial optic nerve crush injury and then immunized with peptides emulsified in CFA (2.5 mg/ml). Control injured rats were injected with PBS in CFA. Staining with 4-Di-10-Asp, preparation of retinal slides, and counting of labeled RGCs were as described for Fig. 1⇑. A, The average number of RGCs per square millimeter was calculated. Significantly more RGCs (mean ± SEM per square millimeter) survived in injured rats immunized with 200 μg of G-8, M-8, or M-8 analog than in their matched PBS-injected controls (p < 0.01, p < 0.03, and p < 0.04, respectively, by two-tailed t test). B, Significantly more RGCs survived in injured rats immunized with 500 μg of G-8 analog than in their matched PBS-injected controls (p < 0.01, by two-tailed t test). Each group consisted of five or six rats.
Protection with uveitogenic peptide is restricted to insults residing in the eye
The above results suggest that when a neuronal insult affects the retinal cell bodies directly, immune neuroprotection is restricted to Ags expressed within the retina. This suggests that vaccination with R16 should not protect against injury to the spinal cord, for example, even though spinal cord tissue can benefit from autoimmunity directed to myelin Ags. To examine whether R16 can protect against incomplete spinal cord injury, we subjected Lewis rats to severe spinal cord contusion and then either vaccinated them with R16 in CFA or injected them with PBS in CFA. Recovery was assessed by experimenters who were blinded to the treatment received. At no time were any differences observed in the recovery of motor activity by the two groups (Fig. 6⇓). Under the same experimental conditions in this model, vaccination with a pathogenic peptide derived from myelin basic protein led to better recovery than that seen in nonvaccinated rats (10, 24, 32).
Immunization with R16 has no effect on recovery after spinal cord contusion. Female Lewis rats were subjected to spinal contusion at T8. Immediately after contusion rats in one group (n = 5) were immunized with R16 emulsified in CFA, and rats in the other group (n = 4) were injected with PBS emulsified in CFA. The motor behavior of each rat was assessed weekly in an open field by observers blinded to the treatment received by the rat. Immunization with R16 did not affect spinal cord recovery. Results are the mean value of the motor score ± SEM.
Discussion
The results of this study show that an immunodominant self-Ag causing an autoimmune disease of the eye, EAU, is the same Ag as that inducing protection of RGCs after either mechanical or biochemical insult to the retina or the optic nerve. Until very recently autoimmunity was defined as a destructive attack of the immune system against a tissue(s) of the body. Several observations, however, are apparently inconsistent with this concept. For example, a high incidence of autoimmune T cells is found in healthy individuals, and disease severity is found not to be correlated with the number of autoimmune T cells (33, 34).
Our research group recently demonstrated that after injury to myelinated CNS axons (optic nerve or spinal cord), passive transfer of T cells reactive to myelin-associated self-Ags reduces the injury-induced loss of tissue (9, 10, 11, 35). A substantial body of evidence indicates that any fibers that were directly injured will inevitably degenerate, but that neighboring neurons that escaped the primary injury can be protected (via immune system activity) from the toxicity of injury-evoked self-destructive compounds. On the basis of the experimental evidence, our group suggested that the protective immune response to injury of myelinated axons is manifested in the activity of Th1 cells directed against myelin-associated Ags (16). Moreover, the injury itself initiates a signal to the immune system, recruiting its activity at the injury site to help the tissue cope with the stressful conditions (13, 15). The ability to spontaneously manifest an autoimmune response with a beneficial outcome was found to be correlated with the ability to resist autoimmune disease development (13). It was found in our laboratory that the ability to spontaneously protect neuronal tissue does not correspond to a lack of autoimmunity, but to the operation of an autoimmune response whose times of onset, duration, and intensity, are tightly controlled (I. Shaked, Z. Porat, J. Kipnis, and M. Schwartz, unpublished observations).
The results of this study show that the self-Ag associated with uveitis protects RGCs from both glutamate toxicity and death induced as a consequence of axonal injury. This protective potential is not restricted to the R16 peptide, as two uveitogenic peptides derived from another retinal Ag, S-Ag, exerted a similar protective effect in the rat optic nerve injury model. In addition, analogs of those peptides, designed to evoke an immune response without causing disease, enhanced RGC survival after optic nerve injury, suggesting that retinal Ags can be used to protect RGCs without the risk of developing autoimmune disease. It is important to emphasize that the protection is Ag specific, as the protection of RGCs from death caused by a direct insult (such as glutamate toxicity) is conferred by vaccination with R16, but not with myelin Ags, while the opposite is true for injury to the spinal cord. In the case of injury to the optic nerve, however, vaccination with either myelin Ags (16, 36) or retinal Ags improved RGC survival, presumably by attenuating secondary degeneration at the injury site or in the retina, respectively (37).
Also similar to the case of myelin Ags was the finding that individuals with a limited ability to regulate the autoimmune response are often unable to benefit from the immune response to this peptide if the disease is too severe (24, 36). Protection is obtained, however, if the intensity of autoimmunity is reduced, resulting in a milder disease. Resistant strains benefit from the evoked autoimmune response under a wide range of experimental conditions. An interesting finding of this study was that EAU, an autoimmune disease that affects both the anterior and posterior parts of the eye, can cause loss of RGCs. This loss, however, is minor when weighed against the potential benefit of the autoimmune response. Loss of RGCs recorded when the disease resolved itself showed that the maximal loss measured 2 wk after vaccination in noninjured Lewis rats was ∼17%, whereas the maximal benefit after a neuronal insult was as high as 263% (192 ± 8 surviving RGCs/mm2 in rats immunized with R16 compared with 73 ± 10 in rats injected with PBS). These results show that even if the autoimmune response to the uveitogenic Ag causes some loss of RGCs, this cost is outweighed by the benefit that the neurons derive under injurious conditions.
In our view any tissue uses certain safeguards in its front line of self-defense. We suggest that the Ag that operates evokes an immune response that, in the event of malfunction, induces disease, but not necessarily in the cells that conveyed the stress signal. It thus appears that the tissue endangers some cells for the purpose of saving others. The cells at risk by the disease are neither the RGCs in uveitis nor the myelinated CNS neurons in EAE. Nevertheless, in the absence of appropriate regulation, the intensive autoimmune response against myelin Ags in EAE or against IRBP in uveitis, might eventually lead to neuronal loss as well. We showed here that an anti-IRBP response in uninjured Lewis rats can indeed lead to some RGC loss.
Studies in vitro have shown that activated T cells secrete neurotropins (NTs) such as nerve growth factor, brain-derived neurotropic factor, NT-3, and NT-4/5, suggesting that the neuroprotective effect of autoimmune T cells requires the secretion of these neuronal survival-promoting factors upon Ag-specific reactivation of the T cells at the injury site (38). Other studies have demonstrated in vivo up-regulation of the costimulatory molecule B7-2 on microglia/macrophages and T cells after spinal cord injury in rats treated with T cells reactive to myelin basic protein and suggest that such T cells help prevent cyst formation after spinal cord injury (35). Among the T cells that home to the lesion site, the ones encountering their relevant Ag will be activated. As a consequence they will produce a high level of cytokines. There is evidence to indicate that IFN-γ, being a product of Th1 (the phenotype of the protective T cells) affects glutamate uptake by astrocytes (39, 40). In addition, the T cells and their secreted cytokines can amplify the microglial activity of glutamate uptake (I. Shaked, O. Butovsky, R. Gersner, and M. Schwartz, unpublished observations) and remove the threat as rapidly and as effectively as possible, even at the cost of eliminating some healthy cells (U. Nevo, J. Kipnis, I. Golding, I. Shaked, A. Newmann, S. Akselrod, and M. Schwartz, unpublished observations). Thus, autoimmune T cells may exert their protective effect directly on neurons (41), indirectly (for example, by affecting the state of activation of microglia/macrophages), or both. Further studies are needed to determine whether harnessing of autoimmune T cells as part of a self-defense mechanism also occurs in non-neural tissues and in tissues that, unlike the eye or the CNS, are not immune-privileged.
Acknowledgments
We thank S. Smith for editing the manuscript and A. Shapira for animal maintenance.
Footnotes
-
↵1 The work was supported by Proneuron, Industrial Park, Ness-Ziona, Israel, and in part by grants from the Glaucoma Research Foundation and the Alan Brown Foundation for Spinal Cord Injury (to M.S.), and from the Abe and Kathryn Selsky Foundation (to T.M.). M.S. holds the Maurice and Ilse Katz professional chair in Neuroimmunology.
-
↵2 Address correspondence and reprint requests to Dr. Michal Schwartz, Department of Neurobiology, The Weizmann Institute of Science, 76100 Rehovot, Israel. E-mail address: michal.schwartz{at}weizmann.ac.il
-
↵3 Abbreviations used in this paper: RGC, retinal ganglion cell; 4-Di-10-Asp, 4-(4-(didecylamino)styryl)-N-methylpyridinium iodide; EAU, experimental autoimmune uveoretinitis; IRBP, interphotoreceptor retinoid-binding protein; NT, neurotropin; SPD, Sprague Dawley; S-Ag, soluble Ag.
- Received May 9, 2002.
- Accepted September 3, 2002.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
