Monomeric Recombinant TCR Ligand Reduces Relapse Rate and Severity of Experimental Autoimmune Encephalomyelitis in SJL/J Mice through Cytokine Switch¹

One approach for regulating Ag-specific T cell responses to encephalitogenic peptides is to induce nonresponsiveness using TCR ligands containing extracellular domains of class II MHC molecules linked to specific peptide targets. Several such constructs have been developed that involve natural or recombinant α1α2 and β1β2 MHC class II domains in association with various encephalitogenic or other pathogenic peptides that were either linked covalently or bound noncovalently (1, 2, 3, 4, 5). These molecular complexes bind not only to the TCR but also to the CD4 molecule on the T cell surface through the β2 MHC domain (6), and were found to inhibit T cell activation and prevent experimental autoimmune encephalomyelitis (EAE)⁴ in rodents (3, 7, 8).

Our design includes only the minimal TCR interface, which involves the α1 and β1 MHC domains covalently linked to peptide without CD4 binding (9). These constructs signal directly through the TCR as a partial agonist (10), prevented and treated MBP-induced monophasic EAE in Lewis rats (11, 12), inhibited activation but induced IL-10 secretion in human DR2-restricted T cell clones specific for MBP 85-99 or cABL peptides (13, 14), and reversed chronic clinical and histological EAE induced by MOG 35-55 peptide in DR2 transgenic mice (15). To further evaluate the therapeutic properties of recombinant TCR ligands (RTLs), we designed and tested an RTL for use in SJL mice that develop a relapsing form of EAE after injection with PLP 139-151 peptide in CFA. This RTL, comprised of an I-A^s/PLP 139-151 peptide construct (RTL401), prevented relapses and reversed clinical and histological EAE through a mechanism involving cytokine switching that differs strikingly from our previous studies using rat and human RTLs in other models of EAE.

SJL and (C57BL/6 × SJL)F₁ mice were obtained from Jackson Immunoresearch Laboratories (Bar Harbor, ME) at 6–7 wk of age. The mice were housed in the Animal Resource Facility at the Portland Veterans Affairs Medical Center (Portland, OR) in accordance with institutional guidelines.

Mouse PLP 139-151 (HSLGKWLGHPDKF), PLP 178-191 (NTWTTCQSIAFPSK), MOG 35-55 (MEVGWYRSPFSRVVHLYRNGK), and MBP 84-104 (VHFFKNIVTPRTPPPSQGKGR) peptides were synthesized using solid phase techniques and purified by HPLC at Beckman Institute, Stanford University (Palo Alto, CA).

General methods for the design, cloning, and expression of RTLs have been described previously (9, 11, 14). In brief, mRNA was isolated from the splenocytes of SJL mice using an Oligotex Direct mRNA mini-kit (Qiagen,Valencia, CA). cDNA of the Ag binding/TCR recognition domain of murine I-A^s MHC class II β1 and α1 chains was derived from mRNA using two pairs of PCR primers. The two chains were sequentially linked by a 5-aa linker (GGQDD) in a two-step PCR with NcoI and XhoI restriction sites being added to the amino terminus of the β1 chain and to the carboxyl terminus of the α1 chain, respectively, to create RTL400. The PLP 139-151 peptide with a linker (GGGGSLVPRGSGGGG) was covalently linked to the 5′ end of the β1 domain of RTL400 to form RTL401. The murine I-A^s β1α1 insert was then ligated into pET21d(+) vector and transformed into Nova blue Escherichia coli host (Novagen, Madison, WI) for positive colony selection and sequence verification. RTL400 and RTL401 plasmid constructs were then transformed into E. coli strain BL21(DE3) expression host (Novagen). The purification of proteins has been described previously (14). The final yield of purified protein varied between 15 and 30 mg/L bacterial culture.

Light scattering experiments were conducted in a DynaPro molecular sizing instrument (Protein Solutions, Charlottesville, VA). The protein samples, in 20 mM Tris-Cl buffer at pH 8.5, were filtered through 100 nm Anodisc membrane filters (Whatman, Clifton, NJ) at a concentration of 1.0 mg/ml, and 20 μl of filtered sample were loaded into a quartz cuvette and analyzed with a 488-nm laser beam. Fifty spectra were collected at 4°C to get an estimation of the diffusion coefficient and relative polydispersity of the protein in aqueous solution. Data were then analyzed with Dynamics software V.5.25.44 (Protein Solutions) and buffer baselines were subtracted. Data were expressed as the means of hydrodynamic radius of the sample using nanometer as a unit. The m.w. of the RTLs was estimated with Dynamics software V.5.25.44 (Protein Solutions).

CD analyses were performed as previously described (14) using an Aviv Model 215 CD spectrometer (Aviv Associates, Lakewood, NJ), except that the recombinant proteins were in Tris-Cl buffer at pH 8.5. Spectra were averaged and smoothed using built-in algorithms with buffer baselines subtracted. Secondary structure was estimated using the Variable Selection method (16).

SJL mice were inoculated s.c. in the flanks with 0.2 ml of an emulsion containing 150 μg of PLP 139-151 peptide and an equal volume of CFA containing 150 μg of heat-killed Mycobacterium tuberculosis H37RA (M.Tb.; Difco, Detroit, MI) as described previously (17). The (C57BL/6 × SJL)F₁ mice were immunized s.c in the flanks with 0.2 ml of an emulsion containing 200 μg of MOG 35-55 peptide or 150 μg of PLP 139-151 peptide and an equal amount of CFA containing 200 μg of heat-killed M.Tb. In a separate experiment, SJL mice were immunized s.c in the flanks with 0.2 ml of an emulsion containing 150 μg of PLP 139-151 or 150 μg of PLP 178-191 peptides, or 0.1 ml of an emulsion containing 200 μg of MBP 84-104 peptide and an equal volume of CFA containing 200 μg of heat-killed M. tuberculosis. The mice immunized with MBP 84-104 peptide were boosted a week later with the same peptide in CFA. On the day of immunization boost and 2 days after, the mice were injected i.p. with 200 ng of pertussis toxin (Ptx; List Biological Laboratories, Campbell, CA). The mice were assessed daily for signs of EAE according to the following scale; 0, normal; 1, limp tail or mild hindlimb weakness; 2, moderate hindlimb weakness or mild ataxia; 3, moderately severe hindlimb weakness; 4, severe hindlimb weakness or mild forelimb weakness or moderate ataxia; 5, paraplegia with no more than moderate forelimb weakness; and 6, paraplegia with severe forelimb weakness or severe ataxia or moribund condition.

At disease onset, mice were treated with vehicle (20 mM Tris-HCl), 100 μg of RTL400 or RTL401 given i.v. daily for 3 or 4 days, or 8 consecutive days with antihistamine (25 mg/kg) or 100 μg of RTL400 and RTL401 given s.c. for 8 days, or 10 μg free PLP 139-151 peptide given i.v. or s.c. for 8 consecutive days. Groups of control and treated mice were evaluated statistically for differences in disease incidence, day of onset, mortality, and presence or absence of relapse (χ² test), and for differences in Peak Clinical Score and Cumulative Disease Index (sum of daily scores) (Kruskal-Wallis Test). All in vitro data were generated from mice treated at disease onset with vehicle or 100 μg of RTL401 i.v. for 8 days. Mice were sacrificed at different time points following treatment with RTL401 for immunological and histological analyses.

The intact spinal cords were removed from mice at the peak of clinical disease and fixed in 10% formalin. The spinal cords were dissected after fixation and embedded in paraffin before sectioning. The sections were stained with luxol fast blue/periodic acid-Schiff-hematoxylin to assess demyelination and inflammatory lesions, and analyzed by light microscopy. Semiquantitative analysis of inflammation and demyelination was determined by examining at least 10 sections from each mouse.

Draining lymph node (LN) and spleens were harvested from vehicle- and RTL-treated mice at varying time points after immunization as indicated. A single cell suspension was prepared by homogenizing the tissue through a fine mesh screen. Cells were cultured in a 96-well flat-bottom tissue culture plate at 4 × 10⁵ cells/well in stimulation medium either alone (control) or with test Ags (PLP 139-151, PLP 178-191, and MBP 84-104 peptides) at varying concentrations. Cells were incubated for 3 days at 37°C in 7% CO₂. Cells were then pulsed with 0.5 μCi of [methyl-³H]thymidine (PerkinElmer, Boston, MA) for the final 18 h of incubation. The cells were harvested onto glass fiber filters, and tritiated thymidine uptake was measured by a liquid scintillation counter. Means and SDs were calculated from triplicate wells. Net cpm was calculated by subtracting control cpm from Ag-induced cpm.

LN and spleen cells were cultured at 4 × 10⁶ cells/well in a 24-well flat-bottom culture plate in stimulation medium with 2 μg/ml PLP 139-151 peptide for 48 h. Supernatants were then harvested and stored at −80°C until tested for cytokines. The mouse inflammation CBA kit was used to detect IL-12, TNF-α, IFN-γ, MCP-1, IL-10, and IL-6 simultaneously (BD Biosciences, San Diego, CA). Briefly, 50 μl of sample was mixed with 50 μl of the mixed capture beads and 50 μl of the mouse PE detection reagent. The tubes were incubated at room temperature for 2 h in the dark, followed by a wash step. The samples were then resuspended in 300 μl of wash buffer before acquisition on the FACScan. The data were analyzed using the CBA software (BD Biosciences). Standard curves were generated for each cytokine using the mixed bead standard provided in the kit, and the concentration of cytokine in the supernatant was determined by interpolation from the appropriate standard curve.

Mononuclear cells from the brain were isolated on a Percoll density gradient as previously described (18). Cells were then stained with CD3 FITC (BD PharMingen, San Diego, CA) and VLA-4-PE or LFA-1-PE (Southern Biotechnology Associates, Birmingham, AL) expression by adding 1 μl of Ab per 1 × 10⁶ cells. Cells were incubated at 4°C for 20 min, and then washed two times with staining medium (1× PBS, 3% FBS, 0.02% sodium azide) before FACS analysis on a FACScan instrument (BD Biosciences) using CellQuest software (BD Biosciences). Dual positive T cells were calculated as a percentage of total mononuclear cells analyzed.

Total RNA was isolated from spinal cords using the RNeasy mini-kit protocol (Qiagen) and then converted to cDNA using oligo(dT), random hexamers, and Superscript RT II enzyme (Invitrogen, Grand Island, NY). Real-time PCR was performed using Quantitect SYBR Green PCR master mix (Qiagen) and primers (synthesized by Applied Biosystems, Foster City, CA). Reactions were conducted on the ABI Prism 7000 Sequence Detection System (Applied Biosystems) using the listed primer sequences (5′ to 3′) to detect the following genes: L32: (F: GGA AAC CCA GAG GCA TTG AC; R: TCA GGA TCT GGC CCT TGA AC); IFN-γ: (F:TGC TGA TGG GAG GAG ATG TCT; R: TGC TGT CTG GCC TGC TGT TA); TNF-α: (F: CAG CCG ATG GGT TGT ACC TT; R: GGC AGC CTT GTC CCT TGA); IL-10: (F: GAT GCC CCA GGC AGA GAA; R: CAC CCA GGG AAT TCA AAT GC); IL-6: (F: CCA CGG CCT TCC CTA CTT C; R: TGG GAG TGG TAT CCT CTG TGA A); TGF-β3: (F: GGG ACA GAT CTT GAG CAA GC; R: TGC AGC CTT CCT CCC TCT C); RANTES: (F: CCT CAC CAT CAT CCT CAC TGC A; R: TCT TCT CTG GGT TGG CAC ACA C); macrophage-inflammatory protein (MIP)-2: (F: TGG GCT GCT GTC CCT CAA; R: CCC GGG TGC TGT TTG TTT T); IP-10: (F: CGA TGA CGG GCC AGT GA; R:CGC AGG GAT GAT TTC AAG CT); CCR1: (F: GGG CCC TAG CCA TCT TAG CT; R: TCC CAC TGG GCC TTA AAA AA); CCR2: (F: GTG TAC ATA GCA ACA AGC CTC AAA G; R: CCC CCA CAT AGG GAT CAT GA); CCR3: (F: GGG CAC CAC CCT GTG AAA; R: TGG AGG CAG GAG CCA TGA); CCR5: (F: CAA TTT TCC AGC AAG ACA ATC CT; R: TCT CCT GTG GAT CGG GTA TAG AC); CCR6: (F: AAG ATG CCT GGC TTC CTC TGT; R: GGT CTG CCT GGA GAT GTA GCT T); CCR7: (F: CCA GGC ACG CAA CTT TGA G; R: ACT ACC ACC ACG GCA ATG ATC); CCR8: (F: CCA GCG ATC TTC CCA TTC TTC; R: GCC CTG CAC ACT CCC CTT A).

In previous studies, we demonstrated that RTLs could reverse clinical and histological signs of disease in Lewis rats that developed monophasic EAE (9, 11), as well as in Tg DR2 (DRB1*1501) mice that developed chronic EAE (15). In this study, we further explored the efficacy of RTL therapy on relapsing EAE induced by PLP 139-151 peptide in SJL/J mice. Treatment of EAE in SJL mice required mouse MHC class II design modifications and included the α1 and β1 domains of the I-A^s molecule covalently bound to the PLP 139-151 peptide (RTL401) or the RTL without bound peptide (RTL400).

In our previous report (14), CD analysis showed that the human RTLs had a secondary structure composition similar to the TCR recognition/peptide-binding α1β1 domain of native human MHC class II molecule as determined by x-ray crystallography (19, 20). Our CD data showed that murine RTLs shared a similar anti-parallel β-sheet platform, and α helix secondary structure common to all murine MHC class II Ag-binding domains (21, 22, 23). The size exclusion chromatography data (Fig. 1) and hydrodynamic analysis using DLS (not shown) indicated that the purified and refolded RTL400 and RTL401 were monodispersed molecules in Tris-Cl buffer. Fractions of each peak from the size exclusion column were collected and analyzed by CD. Secondary structure analysis using the Variable Selection method (16) indicated that murine RTLs maintain a high order of secondary structure similar to native murine I-A^k and I-A^u MHC class II molecules (21, 22).

In initial preclinical studies, SJL/J mice with established signs of EAE were treated with varying numbers of daily i.v. injections of 100 μg of RTL401 containing PLP 139-151 peptide. As is shown in Fig. 2, control mice typically developed a relapsing EAE disease course, with onset of the initial episode of acute disease occurring on day 11–12 after injection of PLP 139-151 peptide/CFA and peak clinical scores developing on day 15, followed by a clinical improvement that lasted until day 20. The first relapse was evident by day 22 in essentially all the mice, reaching a second peak on days 27–28. The mice generally had subsequent remissions and may have had additional relapses or developed chronic EAE, but these variations in clinical course occurred sporadically in individual mice.

Treatment with 100 μg of RTL401 i.v. beginning on day 12 and continuing for 8 consecutive days had the greatest effect on clinical EAE (Fig. 2; Table I), although fewer daily i.v. injections (3 or 4 consecutive days) were only nominally less effective (Table I). Compared with vehicle-treated controls, all three regimens ameliorated clinical disease within the first 24 h, reduced the peak severity of the first clinical episode, and essentially eliminated relapses (Fig. 2; Table I). RTL401 treatment reduced the daily clinical score to minimally detectable disease that was maintained even after cessation of treatment for nearly 4 wk, and significantly reduced the cumulative disease index (Table I). Mice receiving eight daily i.v. doses of 100 μg of RTL401 were treated with antihistamines to prevent development of allergic responses to RTLs. Treatment of mice with the same regimen of antihistamine alone had no effect on the course of relapsing EAE (Table I). In contrast to mice treated with RTL401, mice treated with the eight daily i.v. doses of 100 μg of empty RTL400 construct or a molar equivalent dose of free PLP 139-151 peptide (10 μg peptide/injection) with antihistamine did not experience significant clinical benefit compared with untreated control mice (Table I).

As is shown in Fig. 2 and Table II, eight daily injections of 100 μg of RTL401 administered by the s.c. route was also effective in treating EAE, nominally reducing the relapse rate, and significantly reducing daily clinical scores and the cumulative disease index in a manner similar to i.v. injections. In contrast, comparable s.c. injections of the empty RTL400 construct or a molar equivalent dose of free PLP 139-151 peptide did not have any effect on the clinical course of EAE in SJL mice (data not shown). These results demonstrate that both i.v. and s.c. administration of RTL401 reduced relapses of EAE and produced long-lasting clinical benefit even after cessation of RTL treatment on day 20.

To evaluate peptide specificity of RTL treatment in vivo, RTL401 was used to treat EAE induced in SJL/J mice with two different encephalitogenic peptides, PLP 178-191 and MBP 84-104, both restricted by I-A^s. As is shown in Table II, eight daily i.v. injections of 100 μg of RTL401 did not significantly affect the overall severity or relapse rate of EAE induced by either peptide compared with vehicle-treated control mice (p > 0.2), although in each case a nominal reduction in the cumulative disease index was observed. Day 42 LN responses in PLP 178-191 and MBP 84-104 peptide-immunized mice with EAE were specific only for the immunizing peptide, and no responses were observed to PLP 139-151 peptide (data not shown), indicating a lack of epitope spreading.

To further evaluate the requirement for MHC and peptide specificity of RTL treatment, RTL401 was used to treat EAE induced by either PLP 139-151 peptide or MOG 35-55 peptide in (C57BL/6 × SJL)F₁ mice. These mice express both I-A^s and I-E^b MHC class II molecules that restrict PLP 139-151 (I-A^s) and MOG 35-55 (I-E^b) peptides, in both cases producing an encephalitogenic response. As is shown in Fig. 3, treatment at disease onset with eight daily i.v. injections of 100 μg of RTL401 significantly reduced the severity of EAE induced by PLP 139-151 peptide, but had no effect on EAE induced by MOG 35-55 peptide. These data demonstrate that RTL treatment of EAE is specific for the cognate combination of MHC and neuroantigen peptide.

LNs and spleen cells from vehicle control and RTL401 treated (eight daily i.v. injections of 100 μg) SJL/J mice with EAE were analyzed during the course of treatment for proliferation and cytokine responses to the immunizing PLP 139-151 peptide. Immune cell responses were assessed just after disease onset but before treatment (day 11), 24 h after initiation of treatment (day 13), at the peak of the initial clinical episode (day 15), at the first remission (day 18), at the beginning of the first relapse (day 22), at the peak of the first relapse (day 28), and at the end of the first relapse (day 42). In contrast to our previously published results in DR2-expressing mice (15), there was no significant inhibitory effect of RTL treatment on proliferation responses at any time during the course of EAE. As exemplified in Fig. 4, treatment with RTL401 nominally inhibited proliferation responses to PLP 139-151 peptide in LN cultures, but significantly enhanced proliferation of splenocyte cultures at several time points, including on day 42 as shown in Fig. 4. In contrast, RTL401 treatment had mixed effects on cytokine secretion from PLP 139-151-stimulated splenocytes (Fig. 5). One day after initiation of RTL401 treatment (day 13), there were no significant changes in cytokine responses compared with control mice. Surprisingly, at the peak of the first episode of EAE (day 15), there was enhanced secretion of both inflammatory (TNF-α, IFN-γ, MCP-1, and IL-6) and anti-inflammatory (IL-10) factors in splenocyte cultures from RTL401-treated vs control mice. However, during remission from the first episode of EAE (day 18), the cytokine picture changed dramatically, with strongly reduced levels of IFN-γ, still enhanced levels of MCP-1, but no significant differences in TNF-α, IL-6, or IL-10 in RTL401-treated mice. At onset of the first relapse (day 22), there was again a significant reduction in secreted IFN-γ in RTL401-treated mice, but no significant differences in the other inflammatory factors (Fig. 5). Of possible importance for systemic regulation, there was a significant increase in secreted IL-10 levels by PLP 139-151-reactive splenocytes from RTL401-treated mice at both the onset and peak of the first relapse (days 22 and 28, respectively). Both IgG1 and IgG2a Abs were detected in serum during the course of EAE, but levels showed only minor fluctuations as a result of RTL401 treatment (data not shown).

To further evaluate the effects of RTL401 therapy on EAE, we obtained histological sections and conducted phenotypic and functional analyses of CNS cells. As is shown in Fig. 6, histological sections of spinal cords taken on day 46 showed slightly reduced inflammatory lesions and decreased demyelination in RTL401-treated vs control mice. This reduction in inflammatory activity found in RTL401-treated mice was reflected by a nominal decrease in the number of inflammatory mononuclear cells obtained from brain and spinal cord tissue over the course of treatment (Fig. 7). The reduction of inflammatory cells was most pronounced at onset and peak of the first clinical episode (days 13 and 15); and at onset of the first relapse (day 22), was marked by an overall decrease of CD4⁺ T cells (from 43 to 23%) but an increase in CD11b⁺ monocytes/macrophages (from 38 to 60%) as determined by FACS analysis (not shown). Moreover, the number of T cells expressing adhesion/homing markers VLA-4 and LFA-1 was consistently reduced in brains and spinal cords from RTL401-treated mice on days 22, 28, and 42 (brain only) after EAE induction (Fig. 8). From day 15 on, RT-PCR analysis of spinal cord tissue from RTL-401-treated mice also showed moderate to strong reduction in expression of mRNA for inflammatory cytokines (IFN-γ, TNF-α, and IL-6) and chemokines (RANTES, MIP-2, and IP-10), but enhanced expression of TGF-β3 (Fig. 9), consistent with our current report suggesting a protective role for this cytokine (24). Expression of IL-10 was very low throughout the EAE disease course in spinal cords from RTL-treated mice, with only a slight enhancement in RTL401-treated mice during the first relapse (day 22; Fig. 9). Interestingly, expression of most chemokine receptors (CCR1, CCR2, CCR5, CCR6, CCR7, and CCR8) was moderately to strongly reduced in spinal cord tissue from RTL401-treated mice beginning at the peak of the first episode (day 15; Fig. 10). In contrast, expression of CCR3 (Th2 associated) appeared to be uniquely enhanced in spinal cord tissue collected from RTL401-treated vs control mice during the first relapse (days 22 and 28, Fig. 10).

Previously, we successfully designed and tested oligomeric RTLs specific for both human and rat T cells that reversed clinical EAE and induced long-term T cell tolerance. In this report, we now describe the design characteristics and therapeutic effects of a monomeric murine RTL401 (I-A^s/PLP 139-151 peptide) on a relapsing model of EAE in SJL/J mice. Generally, RTL401 had very similar structural characteristics and therapeutic effects on EAE compared with previously designed molecules, although some important differences were noted in its effects on the activation and inflammatory properties of targeted encephalitogenic T cells. A similar monomeric form of human DR2 RTL has been produced⁵ and is currently being tested for inhibitory mechanism in HLA-DR2 transgenic mice developing chronic EAE.

Secondary structure analysis from CD spectra of murine RTLs indicated that RTL400 and RTL401 maintained a high order of secondary structure similar to native murine I-A^k and I-A^u MHC class II molecules (21, 22). The recombinant RTL is a relatively small molecule (∼24 kDa) containing a native disulfide bond between cysteine 17 and 79 (RTL401 amino acid numbering, corresponding to murine I-A^s β-chain residues 42 and 104). This disulfide bond was retained upon refolding, demonstrated by comparing mobility during electrophoresis (SDA-PAGE) of the RTL in the presence or absence of the reducing reagent, 2-ME. Both RTL400 and RTL401 showed a higher mobility in the absence of 2-ME, indicative of a more compact structure compared with the reduced RTLs (data not shown). Together, these data represent a primary confirmation of the conformational integrity of the molecule. Unlike the human HLA-DR2 construct and rat I-A constructs that tended to aggregate during the refolding process, the mouse RTL constructs appeared to be monodispersed molecules, based on light scattering and size exclusion chromatography analyses.

Of potential clinical importance, these monodispersed molecules induced specific and significant inhibition of PLP 139-151 peptide-induced EAE, but not EAE induced by other myelin peptides when administered in vivo. Our studies demonstrated the potent activity of this minimal TCR ligand to reverse clinical signs of EAE and prevent relapses for at least 26 days after completion of a single 3-, 4-, or 8-day course of daily RTL injections. Disease expression after RTL treatment was minimal, although persistent, unlike the complete abrogation of clinical signs observed in RTL-treated DR2 Tg mice (15). One explanation for chronic low-level EAE might be epitope spreading (25, 26), but this possibility seems less likely since our RTL-treated mice did not develop T cell responses to other known subdominant encephalitogenic peptides, including PLP 178-191 or MBP 84-104. Although i.v. injections provided the lowest cumulative EAE scores, s.c. injections were also highly effective. This finding may be important for future application of RTL therapy to humans, in whom the s.c route of injection is preferable due to ease of injection and reduced risk of hypersensitivity reactions. Such reactions were noted in i.v. RTL-treated SJL/J mice, but could be controlled by injection of antihistamines.

Mechanistically, the murine RTL401 appeared to possess several differences compared with our human DR2/MOG 35-55 construct that inhibited chronic EAE in DR2 transgenic mice (15) and our rat RT-1B^l/MBP 72-89 construct that inhibited monophasic EAE in Lewis rats (12). Both previous constructs were oligomeric and induced a striking reduction of LN T cell responses, as assessed by proliferation and secretion of inflammatory cytokines including IFN-γ and TNF-α. In contrast, the murine I-A^s/PLP 139-151 construct did not significantly reduce T cell proliferation responses to PLP 139-151 peptide, but instead, enhanced splenocyte proliferation and secretion of both inflammatory (TNF-α and IFN-γ) and anti-inflammatory (IL-10) cytokines during the first 3 days of treatment (Fig. 5). In general, variations in expression of inflammatory cytokines mirrored periods of EAE relapses and remission in control SJL/J mice, with more expression noted on days 15 (peak of initial episode) and 22 (first relapse) than on day 18 (remission). However, continued treatment with RTL401 resulted in strongly decreased levels of IFN-γ, while at the same time maintaining elevated IL-10 levels (Fig. 5). These data suggest that in SJL mice, RTLs induced a cytokine switch rather than anergy or apoptosis in treated T cells that still allowed homing to the target organ (CNS). Interestingly, treatment of human T cell clones in vitro with DR2/MBP 85-99 or DR2/cABL peptide RTLs led to a similar enhancement of IL-10 secretion, raising the possibility of an RTL-induced cytokine switch mechanism in humans as well (13). Other Th2 cytokines such as IL-4 and IL-5 may also be involved, but were not assessed in this study.

The mechanistic differences observed in the periphery apparently resulted in differences in CNS as well. Histological sections of spinal cord tissue from RTL-treated SJL mice showed less demyelination, but only a modest reduction of inflammatory lesions. Moreover, both brain and spinal cord tissue from RTL401-treated mice had only a slight reduction in numbers of infiltrating cells, unlike in RTL312-treated DR2 mice protected from EAE that had a more drastic reduction of infiltrating CNS cells (15). During the first relapse, the RTL-treated SJL/J mice had a significant reduction in the percent of infiltrating cells expressing VLA-4 and LFA-1, adhesion molecules that are known to be important in EAE to direct homing of leukocytes to the perivascular sites of inflammatory lesions in CNS tissue (27, 28). Further analysis of mRNA from CNS tissue also demonstrated a striking reduction in expression of inflammatory cytokines (IFN-γ, TNF-α, and IL-6) and chemokines (RANTES, MIP-2, and IP-10), but enhanced expression of anti-inflammatory cytokines (TGF-β3 and IL-10). IL-10 is known to inhibit IFN-γ production and clinical expression of EAE (29), and we recently described an association with increased expression of TGF-β3 and EAE protection (24). The expression pattern for inflammatory chemokine receptors in CNS appeared to be related to the clinical disease course of EAE, with strongest expression at the peak of the initial episode and/or the beginning of the first relapse. Of interest, CCR1, CCR2, and CCR7 appeared to be expressed preferentially in control mice during the first episode of EAE, whereas CCR5, CCR6, CCR8 were more strongly expressed during the first relapse. Of importance, treatment with RTL401 reduced expression of all these CCRs during both clinical episodes of EAE (Fig. 10). In our previous studies in C57BL/6 mice with EAE, we observed enhanced expression of CCR1, CCR2, and CCR5 in CNS at the peak of EAE (30). Moreover, in vitro treatment of encephalitogenic T cells with IL-12 and IL-18, respectively, enhanced expression of IFN-γ/CCR5 and TNF-α/CCR4/CCR7 and potentiated transfer of EAE (31). CCR5 up-regulation by IL-12 has also been reported to enhance LFA-1-mediated adhesiveness (32) (see above), and CCR7 binding to its ligand, MIP-3b, promotes proliferation of CD4⁺ T cells and progression of autoimmunity (33). Based on their pattern of expression during EAE and their strong down-regulation by RTL401, the current study also implicates CCR6 (34) and CCR8 (35) as inflammatory CCRs that may contribute to EAE. In contrast to its inhibitory effects on inflammatory CCRs, RTL401 treatment strongly enhanced expression of CCR3 that has been associated with Th2 responses (36) during the initiation and peak of the first relapse (Fig. 10). This enhancement of CCR3 in EAE-protected mice is reminiscent of the strong up-regulation of CCR3 in BV8S2 transgenic mice successfully treated with TCR BV8S2 determinants (37). Taken together, these findings indicate that regulation of CCR expression is an important function of the RTL treatment mechanism.

Thus, the systemic effects of RTL therapy that promoted a cytokine switch in response to the encephalitogenic PLP 139-151 peptide apparently produced a nonencephalitogenic T cell phenotype that retained some ability to infiltrate CNS tissue. However, the infiltrating cells from RTL401-treated mice clearly had reduced inflammatory capability, enhanced secretion of anti-inflammatory factors, and enhanced expression of a protective CCR. Thus, replacement of the disease-initiating encephalitogenic T cells in CNS by RTL-altered T cells was associated with partial resolution of inflammatory lesions and reversal of clinical disease. However, the persistent low-level EAE might result from incomplete regulation induced by our postulated T cell cytokine switch mechanism and the residual compact lesions found in the spinal cord sections (Fig. 6). The cytokine switch mechanism postulated here differs from an anergy mechanism reported previously in SJL/J mice by Sharma et al. (3) using purified natural four domain I-A^s molecules loaded with PLP 139-151 peptide, or from an apparent deletional mechanism in HLA-DR2 mice treated with an aggregated form of our two-domain RTL design (15).

In conclusion, our study demonstrates for the first time the potent therapeutic effects of a murine minimal TCR ligand in a relapsing model of EAE in SJL mice. A single course of i.v. or s.c. RTL injections prevented relapses and induced long-term clinical benefits that appeared to be mediated by a cytokine switch mechanism involving IL-10, TGF-β3, and CCR3, leading to a moderation of CNS inflammation and demyelination. These results strongly support the clinical application of this novel class of peptide/MHC class II constructs as treatment for T cell-mediated autoimmune diseases such as multiple sclerosis.

We thank Eva Niehaus for assistance in preparing the manuscript.