Neuroprotective autoimmunity: Naturally occurring CD4⁺CD25⁺ regulatory T cells suppress the ability to withstand injury to the central nervous system

Inbred adult female Lewis and Sprague–Dawley (SPD) rats, adult wild-type BALB/c mice, and nude mice were supplied by the Animal Breeding Center of The Weizmann Institute of Science. Animals were handled according to the regulations formulated by the Institutional Animal Care and Use Committee.

MBP from guinea pig spinal cord and ovalbumin were purchased from Sigma.

Mouse anti-rat CD4 antibodies conjugated to phycoerythrin (PE) and mouse anti-rat CD25 antibodies conjugated to fluorescein isothiocyanate were purchased from Serotec. Rat anti-mouse PE-conjugated CD25 antibody (PC61) was purchased from PharMingen. Mouse recombinant IL-2 and anti-mouse ζ-CD3 were purchased from R & D Systems.

The optic nerve was crushed as described in detail (12). Using a binocular operating microscope, we anesthetized the animals and exposed their right optic nerves. In rats, we used calibrated cross-action forceps to inflict a moderate or severe crush injury on the optic nerve, 1–2 mm from the eye. The severity of the injury determines the number of directly injured neurons. To assess neuroprotection we inflicted a moderate crush injury on the optic nerve in Lewis rats (severe crush in this strain leaves almost no viable retinal ganglion cells (RGCs) because of poor endogenous neuroprotection) and a severe crush in SPD rats. To assess systemic and local inflammatory effects, we inflicted a severe crush in both strains. Mice were subjected to severe crush injury of the intraorbital portion of one optic nerve. The contralateral nerve was left undisturbed.

Secondary degeneration of optic nerve axons was assessed by retrograde labeling of RGCs, which was done by the application, 2 weeks after crush injury, of the fluorescent lipophilic dye 4-[4-(didecylamino)styryl]-N-methylpyridinium iodide (Molecular Probes) distally to the site of lesion, as described (12).

The neurotracer dye FluoroGold (5% solution in saline; Fluorochrome, Denver) was injected (3 days before the injury of the optic nerve) into the anesthetized mouse (1 μl, at a rate of 0.5 μl/min in each hemisphere) by using a Hamilton syringe, at a depth of 2 mm from the exposed brain surface, 2.92 mm posterior to the bregma, and 0.5 mm lateral to the midline. One week after crush injury the mice were killed and their retinas were detached and prepared as flattened whole mounts in 4% paraformaldehyde solution. Labeled cells from four to six selected fields of identical size (0.7 mm²) were counted.

Donor splenocytes from rats (aged up to 10 weeks) were obtained by rupturing the spleen and following conventional procedures. The splenocytes were washed with hypotonic buffer (ACK) to lyse red blood cells.

Cells were immunostained according to manufacturer instructions and were resuspended in 0.4 ml of 1% paraformaldehyde and analyzed by FACSort (Becton Dickinson), with 10,000 events scored. In single-color analysis, positive cells were defined as cells with higher immunofluorescence values, on a logarithmic scale, than those of control cells incubated with isotype antibodies as a control. The cells were scored from a region defined according to physical parameters that indicate the size (forward scatter) and granularity (side scatter) of lymphocytes. CD4⁺ lymphocytes were then gated for CD25⁺ analysis.

Splenocytes obtained from wild-type BALB/c mice were prepared by the standard procedure and then incubated with rat anti-mouse PE-conjugated CD25 antibody followed by incubation with anti-PE beads (Becton Dickinson). After being washed, the splenocytes were subjected to AutoMacs (Miltenyi Biotec, Bergisch Gladbach, Germany) by using the “deplete sensitive” program. Recovered populations were analyzed by FACSort (Becton Dickinson).

Lymph nodes (axillary, inguinal, superficial cervical, mandibular, and mesenteric) and spleens were harvested and mashed. T cells were purified (enriched by negative selection) on T cell columns (R & D Systems). The enriched T cells were incubated with anti-CD8 microbeads (Miltenyi Biotec), and negatively selected CD4⁺ T cells were incubated with PE-conjugated anti-CD25 (30 μg/10⁸ cells) in PBS/2% FCS. They were then washed and incubated with anti-PE microbeads (Miltenyi Biotec) and subjected to magnetic separation with AutoMACS. The retained cells were eluted from the column as purified CD4⁺CD25⁺ cells. The negative fraction consisted of CD4⁺CD25⁻ T cells. Cell purity was checked by FACSort and typically ranged from 88% to 95%. Purified cells were cultured in 24-well plates (1 ml) with T cell-depleted spleen cells as accessory cells (irradiated with 3,000 rads) and 0.5 μg/ml anti-CD3, supplemented with 100 units of mrIL-2 (R & D Systems).

Our first objective was to determine whether the spontaneous T cell-mediated neuroprotective response to injury of myelinated axons in the CNS is autoimmune in nature and is directed against myelin. It seemed reasonable to assume that if the response is indeed specifically evoked by myelin-associated antigens, eliminating the ability to respond to myelin antigens would diminish the animals' capacity to recover from axonal injury. To test this assumption, we immunized rats at birth with whole spinal cord homogenate (WSCH) emulsified in incomplete Freund's adjuvant (IFA). Previous studies have demonstrated strain-related differences in the spontaneous ability of rats to manifest T cell-mediated protection against the consequences of nerve injuries (7). We therefore first immunized neonates of SPD rats (a strain endowed with the spontaneous ability to manifest a T cell-dependent protective response) with myelin-associated antigens. In adult rats that were immunized with these self proteins as neonates, the number of neurons (expressed as the mean number ± SEM of RGCs per square millimeter) that survived a partial optic nerve crush injury (34.8 ± 1.3) was significantly smaller than in control injured rats that were immunized at birth with ovalbumin (48.4 ± 2.1; P < 0.01) or were immunized at birth with IFA without antigen (46.6 ± 3.8; P < 0.001; Fig. 1A). It is important to note that the number of surviving neurons at any time after injury in this model is a reflection, at least in part, of the balance between injury-induced degenerative processes (12) and the ability of T cell-dependent mechanisms to resist them (2, 6, 7). Since the rats were not subjected to any intervention as adults, meaning that other mechanisms that might affect neuroprotection were untouched, the observed decrease of more than 30% in neuronal survival after CNS injury was purely a reflection of nonresponsiveness to myelin proteins and is therefore of great biological significance. Similarly, when the neonatally immunized rats were subjected as adults to a contusive injury of the spinal cord (3, 13), we found that they had lost the relative advantage normally enjoyed by resistant strains in recovering from a spinal cord injury, because their capacity for spontaneous recovery was completely blocked (Fig. 1B). Functional activity after incomplete spinal cord injury is measured by the amount of neural tissue that escaped the primary injury minus the amount of tissue that undergoes delayed degeneration. The latter in turn depends on the rat's ability to withstand the consequences of injury: the more resistant the rat, the better its recovery (7). Our results showed that after spinal cord injury, locomotor activity in an open field was significantly reduced in adult rats that were immunized as neonates with myelin antigens relative to that in rats that were immunized as neonates with ovalbumin with a similar spinal injury. Mean plateau scores (±SEM) of locomotor activity, measured on a scale of 0 (complete paralysis) to 21 (normal mobility) (14), were 3.3 ± 0.4 in the myelin-tolerized rats, reflecting almost complete paralysis, and 6.5 ± 0.8 in the controls, reflecting noncoordinated walking ability (P < 0.01). It is important to note that motor activity on this scale is nonlinear, and that the numbers are arbitrary units assigned to different motor skills (14). Therefore, the difference between values of ≈3 and 7 is of great biological significance. Moreover, once the rats reach plateau values the effect is long-lasting (3).

The induced nonresponsiveness to myelin proteins was verified by assessment of the immune response in rats vaccinated with WSCH as adults. Immunization with spinal cord homogenate is known to evoke a strong response to MBP. Therefore, to verify that the neonatal immunization did indeed eliminate the response to myelin-associated antigens, we assayed the ability of the neonatally immunized rats to respond as adults to myelin antigen by immunizing them as adults with WSCH and assaying the ability of their lymphoid organs to proliferate upon challenge with peptides derived from MBP. The proliferative response to MBP after immunization with WSCH was significantly reduced in the lymphoid organs of rats that were immunized at birth with WSCH compared with control rats that were immunized at birth with ovalbumin (Table 1). We also tested the response to myelin oligodendrocyte glycoprotein in the lymphoid organs of these rats. Because the proliferative response to myelin oligodendrocyte glycoprotein after immunization with WSCH was extremely weak, it was difficult to assess any loss resulting from the neonatal immunization (data not shown).

Lack of a proliferative response in adult rats after neonatal immunization does not necessarily mean that the relevant clones have been deleted. It might be a consequence of an alteration in clonal phenotype (a switch from Th1 to Th2/Th3), resulting in clones that are amenable to activation but only slightly or not at all to proliferation. To verify that the neonatal immunization led to a state of nonresponsiveness and not to a change in phenotype, we compared the activation state of T cells after optic nerve injury in adult rats that had been immunized at birth with WSCH to that in similarly injured rats that had not been immunized as neonates. Activation state was determined by measuring the level of expression of the ergotope IL-2 receptor marker in the lymphoid organs. This marker provides a measure of the expression of the α-chain of the IL-2 receptor CD25⁺. In naïve rats it measures the amount of regulatory T cells, whereas in activated lymphoid organs (after any challenge) it measures the total amount of activated T cells (whether regulatory or not) and naïve regulatory T cells. In rats that were not immunized with myelin at birth, optic nerve injury at the adult stage resulted in an increase in expression of the CD25 ergotope marker relative to matched uninjured controls (Fig. 2). In contrast, no change in CD25 expression was observed in lymphoid organs of similarly injured rats that were immunized with myelin antigens at birth. Taken together, the results shown in Figs. 1 and 2 suggest that the spontaneous ability to withstand the consequences of CNS axonal insult depends on the ability to activate myelin-specific T cells.

As indicated above, the spontaneous ability to manifest a myelin-specific T cell-mediated response is a prerequisite for withstanding the consequences of a CNS axonal insult. We considered the possibility that early onset of the T cell response to self-antigens might be the rate-limiting step in developing a beneficial autoimmune response. We therefore attempted to circumvent the limitation in the physiological ability to manifest a beneficial autoimmune response by up-regulating the myelin-specific autoimmune T cells (effector T cells) by means of active immunization with WSCH emulsified in CFA. Active immunization of Lewis rats with WSCH is often done experimentally to induce an autoimmune response culminating in experimental autoimmune encephalomyelitis. This active immunization, although inducing the monophasic disease, led to improved survival of neurons after optic nerve injury (81 ± 7 compared with 48 ± 4; mean number of RGCs ± SEM; P < 0.001; Fig. 3).

We considered the possibility that one factor limiting the spontaneous ability to manifest a beneficial autoimmune response might be the constitutive presence of CD4⁺CD25⁺ regulatory T cells. These naturally occurring regulatory T cells, known to be powerful inhibitors of effector T cell activation (10), can be eliminated by split-dose γ irradiation after thymectomy in 4-week-old rats (15). The combined treatment is known to disrupt the regulation of autoreactive clones and increase the frequency of spontaneous autoimmune disease development (9, 15). We found that Lewis rats treated in this way became better able to withstand the consequences of a CNS insult, manifested by an increase of almost 4-fold in neuronal survival after optic nerve crush. The number of surviving neurons per square millimeter in the thymectomized irradiated Lewis rats 2 weeks after optic nerve injury was 113 ± 5 (mean ± SEM), compared with 36 ± 1.5 in control rats that were subjected to crush injury only (P < 0.001; Fig. 4A). The number of surviving neurons in thymectomized irradiated SPD rats after optic nerve injury was 53 ± 4, compared with 32 ± 3 in SPD rats subjected to crush injury only (P < 0.01; Fig. 4B).

The activity of naturally occurring regulatory T cells in mice can be eliminated by thymectomy 3 days after birth (11). We found that significantly more neurons survived severe optic nerve crush injury in adult BALB/c mice that had been subjected to thymectomy 3 days after birth than in nonthymectomized adult mice (737 ± 63 compared with 620 ± 63, P < 0.001; Fig. 5A). It is important to note that in the mouse optic nerve injury model the total number of surviving RGCs is counted, whereas in the rat optic nerve injury model only RGCs with still-intact fibers are counted (7, 12, 16).

In seeking to provide direct evidence that the regulatory T cells (CD4⁺CD25⁺) are the cells that limit the individual's ability to manifest a beneficial autoimmune response to CNS injury, we examined neuronal survival after optic nerve injury in animals deprived of regulatory T cells. Nude mice on a background of BALB/c were replenished with a normal splenocyte population obtained from wild-type BALB/c mice or with splenocytes deprived of CD4⁺CD25⁺ regulatory T cells. Naturally occurring regulatory T cells comprise ≈10% of the total CD4⁺ T cell population in naïve spleen. After depletion (see Materials and Methods), the percentage of CD4⁺CD25⁺ T cells was <0.5% (data not shown). Before replenishment, the nude mice, as expected (2, 7), showed significantly fewer surviving neurons after optic nerve crush injury than did their wild-type counterparts (Fig. 5B). The number of surviving neurons in nude mice replenished with a normal splenocyte population was significantly larger than that in the nonreplenished nude mice (643 ± 27 compared with 412 ± 27; P 0.01), and was slightly (though not significantly) higher than to the number of surviving neurons found after crush injury in the wild type (577 ± 58). Neuronal survival after optic nerve crush was found to be best in nude mouse recipients of splenocytes lacking CD4⁺CD25⁺ (877 ± 81; P < 0.001).

To examine the direct effect of regulatory T cells on the ability to resist consequences of CNS injury, we measured neuronal survival in wild-type BALB/c (endowed with spontaneous ability to manifest a T cell-dependent protective mechanism) after optic nerve injury and injection of regulatory CD4⁺CD25⁺ T cells. Three groups of BALB/c mice were subjected to optic nerve crush injury and immediately afterward were injected with activated CD4⁺CD25⁺ regulatory T cells, CD4⁺CD25⁻ effector T cells, or vehicle (PBS). Neuronal survival was determined 2 weeks later. The purity of the isolated CD4⁺CD25⁺ cells ranged between 88% and 95% in different experiments. (Three independent experiments were performed; n = 5–6 in each group in each experiment.) Mice injected with the regulatory CD4⁺CD25⁺ T cells showed significantly worse neuronal survival (705 ± 111; P < 0.01) than PBS-injected mice (1,050 ± 70) or mice injected with effector (CD4⁺CD25⁻) T cells (906 ± 80) (Fig. 6).