Pathologic Role and Temporal Appearance of Newly Emerging Autoepitopes in Relapsing Experimental Autoimmune Encephalomyelitis¹

The mechanism(s) underlying the initiation and progression of autoimmune disease are not well understood. A number of recent studies in both animal models of autoimmune disease and their human counterparts have shown that epitope spreading, i.e., the de novo activation of autoreactive T cells by autoepitopes released secondary to inflammatory tissue damage, occurs during disease progression (1). Epitope spreading has been proposed to contribute to the chronic pathogenesis of T cell-mediated autoimmune diseases, including experimental autoimmune encephalomyelitis (EAE)³ (2, 3) and spontaneous diabetes in the nonobese diabetic mouse (4, 5). Epitope spreading has also been demonstrated in Ab-mediated autoimmune models of systemic lupus erythematosis (6, 7) and autoimmune thyroiditis (8). More recently, epitope spreading to self autoepitopes has also been reported to ensue following virus-induced tissue damage in murine models of multiple sclerosis (MS) (Theiler’s virus-induced demyelinating disease (9)) and diabetes (Coxsackie virus-induced diabetes (10)).

We have been examining the functional role of epitope spreading using relapsing-remitting EAE (R-EAE) in the SJL mouse, a CD4⁺ Th1-mediated demyelinating disease which serves as a model of MS. The R-EAE system is useful for multiple reasons: 1) disease can be actively induced in a peptide-specific manner; 2) numerous encephalitogenic epitopes on multiple myelin proteins including proteolipid protein (PLP), myelin basic protein (MBP), and myelin oligodendrocyte glycoprotein (MOG) have been identified; and 3) the relative encephalitogenic dominance (hierarchy) of these epitopes in R-EAE induction has been characterized in SJL (1, 11, 12) and (SWR × SJL)F₁ (13) mice. PLP_139–151 is the dominant encephalitogenic epitope in the SJL mouse based on priming for active disease (10 μg/mouse), PLP_178–191 is secondary (50 μg/mouse), and MBP_84–104 (200 μg/mouse plus pertussis toxin) is only weakly encephalitogenic. In R-EAE induced by the immunodominant PLP_139–151 epitope, recovery from the acute clinical episode is accompanied by the expansion of PLP_178–191-specific T cells (i.e., intramolecular epitope spreading). In contrast, SJL mice remitting from acute R-EAE induced with the weakly encephalitogenic MBP_84–104 epitope develop PLP_139–151-specific T cell responses concomitant with disease relapse (i.e., intermolecular epitope spreading) (11). Yu et al. (13) have described a similar pattern of epitope spreading in PLP_139–151-induced R-EAE in (SWR × SJL)F₁ mice.

Diversification of autoimmune responses thus appears to be a common sequelae to myelin destruction in murine R-EAE. However, it remains unclear whether T cells specific for endogenous myelin epitopes play a significant pathologic role in tissue damage during the relapsing clinical episodes, or whether this is simply an epiphenomenon without pathologic consequences. Previous studies showing that T cells specific for relapse-associated determinants purified from mice primed with PLP_139–151 adoptively transfer EAE to naive recipients suggest, but do not prove, that these T cells are pathologic (11, 13).

The current study explores the immunopathological consequences and peptide dominance of epitope spreading in R-EAE induced in SJL mice by both the highly encephalitogenic PLP_139–151 epitope and the weakly encephalitogenic MBP_84–104 epitope. Using a number of functional criteria, the results provide strong evidence that responses to relapse-associated endogenous myelin epitopes drive the pathologic progression of R-EAE. We show that splenocytes from mice with MBP_84–104-R-EAE, activated in vitro with PLP_139–151, can serially transfer R-EAE to naive mice demonstrating that these T cells have encephalitogenic potential. Second, we show that T cells specific for the relapse-associated epitopes are demonstrable in the CNS of mice during disease remission. Third, induction of peptide-specific tolerance to the relapse-associated epitope alone, either before disease induction or after the acute phase, blocks disease progression as assessed by a decreased relapse rate. Last, in vivo blockade of the B7-1/CD28 costimulatory pathway with anti-B7-1 F(ab) during remission in R-EAE blocks epitope spreading and clinical relapses when administered either following the acute disease episode or following the first relapse. The pattern of spreading correlates with encephalitogenic dominance and the precursor frequency of T cells in the draining lymph nodes of primed mice specific for the various encephalitogenic epitopes. Collectively, these results indicate that clinically relevant epitope spreading takes place in a hierarchical order of peptide dominance and plays a critical role in disease progression.

Female SJL/J mice, 6–7 wk old, were purchased from Harlan Laboratories (Indianapolis, IN). All mice were housed in the Northwestern animal care facility and were maintained on standard laboratory food and water ad libitum. Paralyzed mice were afforded easier access to food and water.

PLP_139–151 (HSLGKWLGHPDKF), PLP_178–191 (NTWTTCQSIAFPSK), MBP_84–104 (VHFFKNIVTPRTPPPSQGKGR), and OVA_323–339 (RGAENIEAHAAVAQSI) were synthesized using a synergy peptide synthesizer (Applied Biosystems, Foster City, CA). Purity (>97%) of these peptides was confirmed by mass spectroscopy at the University of North Carolina Peptide Synthesis Core.

Peptide binding to cell-surface class II was measured using a competition assay (14, 15). Next, 10 μM of a biotinylated influenza hemagglutinin (HA) indicator peptide (BHA_307–319; PKYVKQNTLKLAT) was incubated with 3 × 10⁵ I-A^s-expressing DAS.15 fibroblasts in the presence or absence of titrated amounts of a competitor peptide for 1 h at pH 7.0, 37°C, in PBS containing 5% FBS. The cells were washed thoroughly in PBS plus 5% FBS and subsequently stained with fluoresceinated-avidin D (Vector Laboratories, Burlingame, CA), followed by biotinylated anti-avidin D (Vector Laboratories), and again with fluoresceinated-avidin D. Fluorescence was analyzed on a FACScan flow cytometer (Becton Dickinson, San Diego, CA) by gating for equivalent forward scatter on the live cell population. The fluorescent signal was in a range that is linear with respect to fluorescein density. The relative binding affinity of a given peptide to I-A^s class II molecules was determined by assessing the concentration of the peptide required to reduce the fluorescence intensity of binding of the biotin-labeled HA peptide alone to DAS.15 cells by 50%. The BHA_307–319 indicator peptide bound specifically to MHC class II as the peptide did not bind to the nontransfected parent fibroblast line DAP/3 (data not shown).

For PLP_139–151-induced R-EAE, each mouse received 100 μl of a CFA emulsion containing 200 μg of Mycobacterium tuberculosis H37Ra (Difco, Detroit, MI) and 40 μg of PLP_139–151 distributed s.c. over three spots on the dorsal flanks on day 0. For MBP_84–104-induced R-EAE, the adjuvant emulsion contained 200 μg of MBP_84–104 distributed over three sites on the dorsal flanks on both days 0 and 7 and 400 ng of pertussis toxin (List Biological Laboratories, Campbell, CA) i.v. in 0.5 ml of PBS on days 7 and 10. Clinical scores were assessed on a 0–5 scale as follows: 1, limp tail or hind limb weakness (legs fall through the cage top); 2, limp tail and hind limb weakness; 3, partial hind limb paralysis; 4, total hind limb paralysis; and 5, moribund. A relapse was defined as an increase in at least one clinical grade sustained for at least two consecutive days after the animal had previous improved at least a full clinical score and had stabilized.

Mice were immunized s.c. with 100 μl of an IFA emulsion containing 200 μg of M. tuberculosis H37Ra (Difco) and 200 μg of MBP_84–104 distributed over three spots on the flank. Seven to 10 days after immunization, the inguinal, brachial, and axillary lymph nodes were removed from sensitized donors, and single-cell suspensions were prepared. The cells were adjusted to 8 × 10⁶/ml in T75 flasks in complete DMEM containing 50 μg/ml MBP_84–104 in a total volume of 30 ml/flask and incubated at 37°C in a humidified atmosphere containing 5% CO₂. Four days later, the cells were washed twice with buffered saline solution, counted, and 50 × 10⁶ viable cells transferred i.p. to naive SJL mice.

Forty-eight days after active immunization with MBP_84–104, mice were sacrificed and spleens were removed. Spleen cells were cultured with 50 μg/ml PLP_139–151 or MBP_84–104 as described above. After 4 days of culture, 4–6 × 10⁷ PLP_139–151-activated splenocytes or 3–5 × 10⁷ MBP_84–104-activated splenocytes were transferred i.p. into naive SJL recipients.

Mice were anesthetized with methoxyflurane and perfused through the left ventricle with 60 ml of PBS. Spinal cords were extruded by flushing the vertebral canal with PBS and rinsed in PBS. Tissues were forced through 100-mesh stainless screens to give a single-cell suspension. The spinal cord homogenate was resuspended in 30% Percoll (Pharmacia, Piscataway, NJ), divided into tubes (equivalent to 4–5 spinal cords per tube), and underlaid with 70% Percoll. The gradients were centrifuged at 500 × g for 20 min at 24°C. CNS mononuclear cells were collected from the 30:70% interface, washed, and incubated on plastic dishes at 37°C in complete DMEM containing 10% FBS for 2 h. Nonadherent cells were eluted, counted, and used in T cell proliferation assays.

T cell proliferative responses were assessed by incorporation of [³H]thymidine. A total of 5 × 10⁵ viable splenocytes recovered from mice with ongoing R-EAE were cultured in triplicate in 96-well flat-bottom microculture plates (Falcon Labware, Oxnard, CA) in 0.2 ml of DMEM (Life Technologies, Grand Island, NY) supplemented with 5% FBS (Sigma, St. Louis, MO), 2 mM l-glutamine (Sigma), 100 μg/ml streptomycin, and 100 U/ml penicillin (Sigma). Alternately, CNS-infiltrating nonadherent mononuclear cells were plated at 5 × 10⁴ per well with 5 × 10⁵ irradiated naive splenocytes as APCs in complete DMEM. A variety of peptide concentrations were tested. Plates were pulsed with 1 μCi [³H]TdR after 72 h of culture and harvested for scintillation counting 24 h thereafter. The results are expressed as Δ cpm = mean cpm of Ag-containing cultures − mean cpm of control cultures, and as stimulation index = mean cpm of Ag-containing cultures/mean cpm of control cultures.

Precursor frequencies of PLP_139–151-, PLP_178–191-, and MBP_84–104-specific T cells were determined using CD4⁺ T cells purified using MACS columns (Miltenyi Biotec, Auburn, CA) from the draining lymph nodes (pooled axillary, brachial, and inguinal) of SJL mice primed 7 days previously with 25 μM of the individual peptides. Frequencies were determined both by limiting dilution analysis (LDA) using [³H]TdR incorporation as a readout and by determining the numbers of IFN-γ-producing ELISPOTs.

For LDA, T cells were plated in U-bottomed microtiter plates (Nunc, Naperville, IL), using 36 replicates per 2-fold dilution, at concentrations ranging from 2.5 × 10⁵ to 1.95 × 10³ cells/well. Peptide (10 μg/well) was added to 27 wells, and PBS was added to the remaining 9 wells along with 4 × 10⁵ irradiated spleen cells in all wells as APCs. Cultures were incubated for 72 h at 37°C, pulsed with [³H]TdR, and harvested 24 h thereafter. Wells were scored positive if the cpm were > 3 SD above the mean of the control wells. Minimal estimates of the peptide-specific cell frequency were calculated by analysis of the Poisson distribution relationship between the number of responder cells added to the cultures and the percentage of replicate wells that failed to show significant proliferation. Frequency calculations were performed by computer using χ² minimization. This analysis gives a minimal frequency estimate and the 95% confidence limits.

For ELISPOT, T cells were cultured in 96-well flat-bottom, nitrocellulose-coated microculture plates (Whatman Polyfiltronics, Clifton, NJ) according to the method of Forsthuber et al. (16). The plates were precoated overnight at 4°C with 100 μl of anti-IFN-γ (R46A2; PharMingen, San Diego, CA) at 4 μg/ml. Plates were then washed four times with sterile PBS, and the wells were blocked with DMEM-1% BSA for 1 h at room temperature. T cells were plated in triplicate at numbers ranging from 2.5 × 10⁵ to 1.95 × 10³ cells/well in HL-1 serum-free media (BioWhittaker, Walkersville, MD) supplemented with 1% l-glutamine ± 10 μg/well of the appropriate peptide and 4 × 10⁵ naive SJL spleen cells as APCs. Cultures were incubated for 24 h at 37°C, washed three times with PBS, and then washed four times with PBS-0.05% Tween 20. Next, 2 μg/ml of biotinylated anti-IFN-γ (XMG1.2; PharMingen, San Diego, CA) diluted in PBS/Tween 20/1% BSA was added, and the plates were incubated at 4°C overnight. The plates were^, washed four times with PBS/Tween 20 and 0.1 ml of anti-biotin Ab (Vector Laboratories) at a 1:1000 dilution in PBS/Tween 20/1% BSA was added and the plates incubated at room temperature for 2 h. Plates were then washed with PBS four times and developed in nitroblue tetrazolium-5-bromo-4-chloro-3-indolyl phosphate substrate solution (Sigma). The developing reaction was quenched after 30–45 min using distilled water. The plates were air dried and spots enumerated under a dissecting microscope.

DTH responses were quantitated using a 24-h ear swelling assay. Prechallenge ear thickness was determined using a Mitutoyo model 7326 engineer’s micrometer (Schlesinger’s Tools, Brooklyn, NY). DTH responses were elicited by injecting 10 μg of peptide (in 10 μl of saline) into the dorsal surface of the ear using a 100 μl Hamilton syringe fitted with a 30-gauge needle. Twenty-four hours after ear challenge, the increase in ear thickness over prechallenge measurements was determined. Results are expressed in units of 10⁻⁴ inches ± SEM. Ear swelling responses were the result of mononuclear cell infiltration and showed typical DTH kinetics (i.e., minimal swelling at 4 h, maximal swelling at 24–48 h).

The following mAbs were employed: hamster control Ig (Parsi12) and the anti-CD80 (B7-1) mAb 16-10A1. Abs were produced in an Acusyst Jr. Bioreactor (Endotronics, Coon Rapids, MN), and F(ab) of anti-B7-1 mAb 16-10A1 were produced and purified as previously described (17, 18). Mice were treated following recovery from the initial paralytic episode of active PLP_139–151-induced R-EAE (20–25 days postimmunization) and/or following recovery of control mice from the first relapse (45–50 days postimmunization). Ab was administered i.p. every other day for five treatments of 25 μg (total of 125 μg), and mice were monitored for development of clinical relapses for an additional 40–60 days.

Peripheral tolerance induction using peptide-coupled splenocytes was performed as previously described (19). Briefly, erythrocyte-free (Tris-NH₄Cl-treated) splenocytes were coupled with PLP_139–151, PLP_178–191, or MBP_84–104 using an ethylene carbodiimide procedure. Saline-washed SJL/J splenocytes were centrifuged in 50-ml tubes and resuspended to a final concentration of 5 × 10⁸ cells/ml in saline containing 2.0 mg/ml peptide, pH 7.0. Control cells were prepared either without peptide in the reaction mixture (sham-coupled) or coupled with OVA_323–339. The coupling reaction was initiated by the addition of 0.5 ml of freshly prepared ethylene carbodiimide (200 mg/ml saline; Calbiochem-Behring, La Jolla, CA) per ml of cell suspension. Following a 1-h incubation with shaking, cells were washed three times with balanced salt solution and maintained at 4°C until use. Coupling efficiency has previously been determined to be ∼30%, yielding 24–55 μg of peptide/5 × 10⁷ splenocytes (19). Tolerance was induced by the i.v. injection of 5 × 10⁷ peptide- or sham-coupled splenocytes into syngeneic recipient mice.

Comparison of the percentage of animals showing clinical relapses between any two groups of mice was done by χ² using Fisher’s exact probability. Comparisons of the mean day of onset of relapse, mean peak disease severity, and DTH responses between any two groups of mice were analyzed by the Student’s t test. Values of p < 0.05 were considered significant.

Relapses in R-EAE in SJL mice initiated either with the highly immunodominant PLP_139–151 epitope or the weakly encephalitogenic MBP_84–104 epitope are associated with development of CD4⁺ T cell responses to endogenous myelin epitopes. We employed several approaches to determine whether T cells specific for these epitopes play a significant pathologic role in tissue damage during the relapsing clinical episodes.

SJL mice with MBP_84–104-R-EAE experience a severe clinical relapse associated with the activation of PLP_139–151-specific T cells (11). Using adoptive MBP_84–104, we first asked if these PLP_139–151-specific T cells were encephalitogenic as assessed by their ability to transfer clinical R-EAE to naive recipient mice. As seen in Fig. 1, splenic T cells recovered 48 days after adoptive MBP_84–104-specific EAE (during the initial clinical relapse) were capable of transferring disease following in vitro activation with either the initiating MBP_84–104 epitope (3 of 3 mice affected with a mean day of onset of 11.7 and mean maximal severity of 3.3) or with the relapse-associated PLP_139–151 epitope (9 of 10 mice affected with a mean day of onset of 10.4 and mean maximal severity of 3.3). Peptide restimulation was required for successful serial transfer. Thus, following a relatively mild course of acute R-EAE initiated by MBP_84–104-specific T cells, PLP_139–151-specific T cells that arise concomitant with the initial severe disease relapse have potent encephalitogenic activity. These results expand our earlier observation that PLP_178–191-specific T cells recovered from mice in remission from severe acute R-EAE initiated by active priming with PLP_139–151/CFA could serially transfer R-EAE (11).

If T cells specific for endogenous myelin epitopes play a major pathologic role in relapsing disease, they should be demonstrable in the CNS of mice before and during clinical relapses. We recovered plastic nonadherent mononuclear cells from the spinal cords of mice remitting from active (day 18 postpriming) PLP_139–151- and adoptive (day 25 posttransfer) MBP_84–104-specific R-EAE. The cells were rested for 7 days in 2 U/ml rIL-2 and then cultured with irradiated splenic APCs from naive mice ± 50 μM of PLP_139–151, PLP_178–191, or MBP_84–104. As seen in Fig. 2, CNS-infiltrating T cells recovered from mice remitting from the acute phase of PLP_139–151-induced R-EAE respond to both the initiating PLP_139–151 peptide and to the relapse-associated PLP_178–191 epitope, but not to MBP_84–104. Similarly, CNS-infiltrating T cells from mice in remission from adoptive MBP_84–104-specific R-EAE responded very strongly to both the initiating MBP_84–104 peptide and to the relapse PLP_139–151 epitope. Interestingly, CNS T cells from these mice also responded, although less vigorously, to a secondary PLP epitope, PLP_178–191, at this time.

To directly determine whether newly recruited PLP_139–151-specific T cells play a significant role in relapsing EAE initiated by the adoptive transfer of MBP_84–104-specific T cells, peptide-specific tolerance was employed. We examined the effects of tolerance to either the initiating and/or relapse-associated epitopes on disease initiation and progression in MBP_84–104- and PLP_139–151-induced R-EAE. As seen in Fig. 3,A, pretolerization with the disease inducing MBP_84–104 epitope 7 days before induction of active EAE with MBP_84–104/CFA totally protected the mice from acute and relapsing disease. Pretolerization with PLP_139–151 had no significant effect on acute clinical disease, but profoundly suppressed disease progression reducing the relapse rate from 0.83 to 0.14 (p = 0.026), indicating a pathologic role for PLP_139–151-specific T cells in disease relapse. This was confirmed by the finding that tolerance induction during the remission period following recovery from acute MBP_84–104-specific adoptive EAE with either the relapse-associated PLP_139–151 epitope alone or a combination of PLP_139–151 plus MBP_84–104 totally suppressed expression of disease relapses from a level of 7 of 10 (70%) in controls to 0 of 10 (0%, p = 0.003). Relapses were delayed, but the rate was not significantly reduced (p = 0.179), in mice tolerized with the disease-inducing MBP_84–104 epitope alone (Fig. 3 B). The mean peak clinical score for mice displaying relapses was also not significantly different (1.33 in MBP_84–104-tolerized compared with 1.66 in the sham-tolerized group).

In active PLP_139–151-induced R-EAE, pretolerization with the disease-inducing PLP_139–151 epitope totally protected the mice from development of both acute and relapsing disease (Fig. 4,A), while pretolerization with the PLP_178–191 had no effect on acute disease, but reduced the primary relapse rate from 1.0 to 0.25 (p = 0.007). These results again suggest that the response to the relapse-associated PLP_178–191 epitope are primarily responsible for the primary disease relapse. In confirmation, induction of tolerance following recovery from acute PLP_139–151-induced EAE (Fig. 4,B) with either the relapse-associated PLP_178–191 epitope alone or with a combination of PLP_178–191 plus PLP_139–151 (RR = 0.2, p = 0.023), but not with PLP_139–151 alone (RR = 0.5, p = 0.349), significantly reduced the relapse rate when compared with sham-tolerized controls (RR = 0.8). The decrease in the clinical symptoms during relapse correlated with inhibition of the PLP_178–191-specific DTH response (Fig. 4,C). Thus, in R-EAE in which the animals have undergone either a relatively mild acute disease (transfer of MBP_84–104-specific T cells in the absence of CFA; Fig. 3) or a relatively severe acute disease (active PLP_139–151-induced EAE; Fig. 4), T cells specific for the relapse-associated epitope play a major pathologic role in mediating disease relapses even though T cell specific for the initiating epitope persist in the periphery. Interestingly, tolerance to the disease initiating MBP_84–104 epitope during remission from MBP_84–104-induced adoptive R-EAE was more effective than PLP_139–151-specific tolerance during remission from active PLP_139–151-induced R-EAE. This may relate to the large number (5 × 10⁷) of MBP_84–104-specific T cells required to initiate adoptive disease.

We have previously reported that treatment of mice with anti-B7-1 F(ab) after the acute phase of PLP_139–151-induced R-EAE blocks relapses and activation of T cells specific for the relapse-associated PLP_178–191 epitope (12). The efficacy of anti-B7-1 F(ab) treatment in inhibiting disease progression and T cell responses to relapse-associated epitopes upon administration following recovery from the acute stage of disease was compared with later treatment following recovery from the primary relapse according to the protocol shown in Fig. 5,A. As seen in Fig. 5,B, SJL mice treated with anti-B7-1 F(ab)s at the time of remission from the acute phase of active PLP_139–151-induced R-EAE displayed significantly fewer clinical relapses (group B, 3 of 18 relapses) over the ensuing 25 days than hamster Ig-treated controls (group A, 17 of 22 relapses, p = 0.0003). This confirmed our earlier studies that blockade of B7-1-mediated costimulation during disease remission was a potent way to ameliorate ongoing disease. In a second experiment, on day 50, after the hamster Ig-treated controls had recovered from the first relapse, mice in both the hamster Ig-treated control and the anti-B7-1 F(ab)-treated groups were further divided into two groups of equal clinical score and relapse history and treated again with either hamster control Ig or anti-B7-1 F(ab)s. This resulted in four treatment groups: C, received the hamster control Ig at the first and second remission; D, received control Ig at the first remission and anti-B7-1 F(ab)s only during the second remission; E, received the anti-B7-1 F(ab)s only at the first remission and hamster control Ig at the second remission; and F, received the anti-B7-1 F(ab)s at both the first and second remissions. As can be seen in Fig. 5 C, mice that received anti-B7-1 F(ab) treatment at the first and/or second remission (groups D, E, and F) exhibited significantly (p ≤ 0.03) fewer secondary relapses (1 of 15, 1 of 13, 1 of 13) during the ensuing 30 days compared with the control group, which received hamster control Ig during both the first and second remission (group C, 7 of 13 relapses). Thus, the ameliorating effect of anti-B7-1 blockade is long-lasting and effective even when given late in disease following recovery from the primary relapse.

Examination of peptide-specific DTH responses in these animals at 80 days postimmunization revealed several interesting points. First, the mice treated with control Ig at both the first and second remission (group C) exhibited significant DTH responses to the initiating PLP_139–151 epitope, to the PLP_178–191 epitope (responses to which first arise following remission from acute disease), and to MBP_84–104 (responses to MBP_84–104 were not observed until remission from the primary relapse). Second, treatment with anti-B7-1 F(ab) only at remission from acute disease (group E) or treatment at both the first and second remissions (group F) blocked the development of DTH to both PLP_178–191 and MBP_84–104. In contrast, only MBP_84–104-specific responses were inhibited in mice treated with anti-B7-1 F(ab)s during the second remission (group D). Interestingly, these mice failed to relapse even though they still displayed significant peripheral DTH responses to PLP_139–151 and PLP_178–191. Collectively, these results indicate that MBP_84–104-specific T cells play an important role in the second relapse; that clinically relevant epitope spreading follows a hierarchical order, likely dependent on the relative encephalitogenic dominance of each epitope (PLP_139–151 > PLP_178–191 > MBP_84–104); and that anti-B7-1 F(ab) treatment specifically blocks the newly expanding T cell response regardless of the peptide specificity.

Our past experience in induction of R-EAE in the SJL mouse has indicated that PLP_139–151 is the most immunodominant epitope in mouse myelin (20) and that disease can be initiated by a single s.c. immunization with 10 μg of peptide (Table I). PLP_178–191 is also highly encephalitogenic, but requires at least 50 μg/mouse to initiate disease. In contrast, MBP_84–104 is weakly encephalitogenic in that two immunizations of 200 μg/mouse plus i.v. administration of pertussis toxin are required to initiate active disease. Current and past (11, 13) data suggest that clinically relevant epitope spreading follows this hierarchical order in that responses proceed from the most to the least dominant encephalitogenic myelin determinants.

Encephalitogenic dominance could be due to the relative ability of a particular myelin peptide to form a stable complex with I-A^s and/or to differences in the precursor frequency of Th1 cells specific for the individual peptides. Therefore, we measured the I-A^s binding affinity and precursor frequency of the of PLP_139–151-, PLP_178–191-, and MBP_84–104-specific T cells in young SJL mice. As can be seen in Table I, the relative encephalitogenicity of these three myelin peptides is likely not due to their relative ability to bind to MHC class II as all three peptides bound well to I-A^s, with MBP_84–104 exhibiting the highest affinity. In contrast, the precursor frequency of T cells in the draining lymph nodes of SJL mice primed 7 days previously with equimolar amounts of the three peptides directly correlates with their relative encephalitogenicity and the order in which responses arise during ongoing R-EAE. As determined by LDA, the precursor frequency of PLP_139–151-specific T cells in the draining lymph nodes was ∼1/50,000, two to three times the frequency of T cells specific for PLP_178–191 and MBP_84–104. Interestingly, the T cell precursor frequency for each peptide was 5- to 10-fold higher when determined by an IFN-γ ELISPOT assay. PLP_139–151-specific Th1 frequency was ∼1/5000, two to six times the frequency of T cells specific for PLP_178–191 and MBP_84–104.

Results from many animal models of autoimmunity indicate that self tissue damage leads to the activation of autoreactive T and/or B lymphocytes specific for autoepitopes distinct from that used to initiate the disease, i.e., epitope spreading (1, 6, 11, 13, 21, 22). Whether these responses have a pathologic role in subsequent disease progression has been difficult to assess, especially in human disease. Autoreactive cells are found in normal individuals (23), but normally do not cause clinical pathology, perhaps due to low precursor frequency and/or overlying regulatory controls. However, in genetically susceptible individuals (as in genetically susceptible animals (24, 25)), the regulatory controls on autoreactive lymphocytes primed during tissue damage may be defective.

R-EAE has proven to be a useful model in which to examine the initiation and development of epitope spreading during a progressive autoimmune disease. It has long been recognized that T cell specificities change during the course of R-EAE (3, 26, 27, 28, 29). The advantages of using R-EAE to study the pathogenesis of epitope spreading include: the ease of disease induction using defined myelin peptide epitopes; knowledge of encephalitogenic epitopes on a variety of myelin proteins in several inbred mouse strains (30, 31, 32, 33, 34); and knowledge of the relative encephalitogenic dominance of these myelin epitopes (1, 11, 13, 35). It has recently been demonstrated in PLP_139–151-induced R-EAE in both SJL (11, 36) and (SJL × SWR)F₁ mice (13) that clinical relapses are associated with the development of T cell responses to newly emerging epitopes on the same (i.e., intramolecular epitope spreading to distinct PLP epitopes) and/or different myelin epitopes (i.e., intermolecular epitope spreading to MBP epitopes). Moreover, these studies have clearly shown that the development of T cell reactivity to these endogenous myelin epitopes correlates with the extent of myelin destruction occurring during acute clinical disease (11). More recent evidence suggests that a similar pattern of epitope focusing and spreading may occur during the transition from isolated monosymptomatic demyelinating syndromes (a group of distinct clinical disorders with variable rates of progression to MS) to clinically defined MS (37). Thus, elucidation of the cellular and molecular mechanisms driving the epitope spreading process are critical to the design of efficient therapies for treating chronic inflammatory autoimmune diseases.

Because the mere existence of self-reactive CD4⁺ T cells does not necessarily initiate pathology in humans or in animal models of autoimmunity, the development of immune reactivity to additional autoepitopes during disease progression in R-EAE may or may not play an inciting and/or regulatory role in chronic disease pathogenesis. In the current study, we used a number of functional criteria to determine the potential pathologic role of epitope spreading in R-EAE disease progression in SJL mice. The results indicate that T cells specific for spread epitopes play a major pathologic role in mediating disease relapses. Clearly, CD4⁺ T cells specific for relapse-associated epitopes have encephalitogenic potential as they can serially transfer R-EAE to naive recipients (Fig. 1). In addition, T cells specific for relapse-associated epitopes are easily demonstrable in the CNS immediately before and during the primary disease relapse (Fig. 2). More significantly, induction of peptide-specific tolerance to relapse-associated epitopes during remission from acute disease in both MBP_84–104- (Fig. 3) and PLP_139–151-induced R-EAE (Fig. 4) blocks disease progression as assessed by the observation that tolerant animals display a significantly reduced relapse rates. Interestingly, tolerance to the relapse-associated epitope by the i.v. injection of peptide-pulsed, ethylene carbodiimide-fixed splenocytes could also be induced before disease initiation, in which case the animals underwent a normal acute clinical disease course, but failed to exhibit the primary relapses (Figs. 3,A and 4A). This indicates that tolerance induced by peptide-coupled splenocytes is exquisitely Ag specific and does not cause immune deviation or TGF-β-mediated bystander suppression as do tolerance protocols previously employed to determine the role of T cell responses to endogenous epitopes in R-EAE relapses (13). In a more clinically relevant context, our results indicate that if the specificity of T cells responsible for ensuing disease relapse is known, peptide-specific peripheral tolerance is a powerful potential method for treating ongoing autoimmune diseases because this approach is effective in blocking disease relapses in animals in remission from acute disease (Figs. 3 B and 4B). The current data thus provide direct and strong evidence that T cells specific for endogenous myelin epitopes play the major pathologic role in the relapsing phase of disease when compared with prior studies, which used tolerance to a crude mouse spinal cord homogenate or to the intact PLP molecule on fixed APCs (3, 11) or high doses of peptides in IFA (13).

These findings have important implications for the design and use of specific forms of immunotherapy for the treatment of autoimmune diseases. Clearly peptide-specific tolerance is an effective method for arresting ongoing disease progression provided the identity of the next epitope in the spreading cascade is known and therapy is initiated before the relapse (Figs. 3 and 4). Interestingly, disease progression is blocked in the mice tolerized to the relapse-associated epitopes although they retain significant peripheral T cell reactivity to the disease-initiating epitope as assessed by DTH reactivity (Fig. 4,C) and the ability of splenic T cells to transfer disease following peptide reactivation in vitro (Fig. 2). This may indicate that peripheral T cells specific for the disease-initiating epitope in animals have down-regulated critical homing receptors required for transmigration across the blood-brain barrier and/or they may be under the control of a specific regulatory T population that prevents their reactivation in situ. Alternately, the peptide-specific cells retained in the periphery may have lower overall avidity than those cells that originally trafficked to the CNS during acute disease. If epitope spreading contributes to disease pathology in MS and other chronic human autoimmune diseases, peptide-specific disease therapy will have to be individualized for every patient due to the myriad of potential organ-specific autoepitopes and extensive MHC diversity. However, the efficient ability to inhibit ongoing R-EAE by blockade of the B7-1/CD28 costimulatory pathway using anti-B7-1 F(ab), even following the initial clinical relapse (Fig. 5), provides a powerful alternative that requires no prior knowledge of the identity of the relapse-associated epitopes. Preliminary experiments indicate that short-term B7-1 blockade during disease remission results in long-term unresponsiveness in T cells specific for relapse-associated epitopes (C. L. Vanderlugt et al., manuscript in preparation) similar to the previously reported ability of CTLA4-Ig to induce long-term, donor-specific tolerance to transplantation Ags (38).

The results also indicate that epitope spreading in SJL mice occurs in predictable hierarchical order similar to published results in EAE in (SJL × SWR)F₁ mice (13). The temporal sequence of anti-myelin peptide responses did not correlate with the affinity of the individual peptides for binding to the I-A^s molecule, but correlated well with the precursor frequency of T cells in the SJL repertoire specific for the individual peptides (PLP_139–151 > PLP_178–191 > MBP_84–104), which was directly related to the encephalitogenic dominance of these determinants (Table I). Interestingly, precursor frequencies of peptide-specific CD4⁺ T cells were considerably higher when assessed by IFN-γ ELISPOT in comparison to conventional LDA paralleling recent data for CD8⁺ CTLs (39, 40). During the natural course of R-EAE, it is also possible that the sequence of epitope spreading may be in part dictated by the efficiency of the various peptides to be processed by relevant APCs in the CNS or the periphery. The immunodominance of the PLP_139–151 response in the SJL mouse is quite remarkable. In our experience, following acute myelin damage in R-EAE initiated by either MBP_84–104 ( Figs. 1–3) or by PLP_178–191, and in (SJL × B10.PL)F₁ mice with R-EAE initiated by MBPAc_1–11 (data not shown), spreading always first involves T cell reactivity to PLP_139–151. PLP_139–151 responses also arise within 2–3 wk after the onset of myelin destruction in SJL mice infected with Theiler’s murine encephalomyelitis virus (9) with autoimmune responses to other encephalitogenic epitopes (including PLP_178–191, PLP_56–70, MOG_92–106, and MBP_84–104) arising progressively later in the chronic disease.

While many questions remain about the initiation of the epitope spreading cascade and its significance in the pathogenesis and regulation of human autoimmune diseases, the results clearly indicate that clinically relevant epitope spreading in the relapsing-remitting SJL mouse model of MS follows a predictable hierarchical order related to the precursor frequency of myelin epitope-specific CD4⁺ T cells and is functionally important in disease progression. Moreover, these results indicate that peptide-specific tolerance is an effective therapy for an ongoing autoimmune disease. Because determining the specificity and hierarchical order of epitope spreading in human disease is not currently feasible, our results suggest that Ag-specific therapies for ongoing treatment of ongoing autoimmune disease may require peripheral tolerance induction strategies using whole tissue extracts (3), mixtures of encephalitogenic proteins/peptides (41), or costimulatory blockade (12).