HUMAN AUTO EACTIVE T-CELL CLONES

This application is a continuation-in-part of U.S. Patent Application Serial No. 08/426,784, filed April 20, 1995, and a continuation-in-part of U.S. Patent Serial No. 08/545.456 filed October 19, 1995, attorney docket no. 1010/1B107-US1.

Field of the Invention This invention relates to methods and compositions for altering the cytokine pattern of T-cells (particularly inflammation-promoting T-cells of patients afflicted with an autoimmune disease) from a first pattern that promotes immune responses (ThO) to an altered version of the first pattern that downregulates immune responses.

Background of the Invention

Autoimmune diseases are characterized by an abnormal immune response directed against normal autologous (self) tissues.

Based on the type of immune response (or immune reaction) involved, autoimmune diseases in mammals can generally be classified in one of two different categories: cell-mediated (i.e., T-cell-mediated) or antibody-mediated disorders. Non-limiting examples of cell-mediated autoimmune diseases include multiple sclerosis (MS) , rheumatoid arthritis (RA) , autoimmune thyroiditis (AT) , the autoimmune stage of diabetes mellitus (juvenile-onset or Type 1 diabetes) and autoimmune uveoretinitis (AUR) . Antibody-mediated autoimmune

diseases include without limitation myasthenia gravis (MG) and systemic lupus erythematosus (SLE) . The distinction between T-cell-mediated and antibody-mediated autoimmune diseases is not absolute, as many antibody-mediated autoimmune diseases have a T-cell component (i.e. include inflammation of the afflicted organ or tissue) .

Both categories of autoimmune diseases are currently being treated with drugs that suppress immune responses systemically in a non-specific manner, i.e., drugs incapable of selectively suppressing the abnormal immune response. Non- limiting examples of such drugs include methotrexate, cyclophosphamide, Imuran (azathioprine) and cyclosporin A. Steroid compounds such as prednisone and methylprednisolone (also non-specific immunosuppressants) are also employed in many instances. All of these currently employed drugs have limited efficacy against both cell- and antibody-mediated autoimmune diseases. Furthermore, such drugs have significant toxic and other side effects and, more important, eventually induce "global" immunosuppression in the subject being treated. In other words, prolonged treatment with the drugs downregulates the normal protective immune response against pathogens thereby increasing the risk of infection. In addition, patients subjected to prolonged global immuno¬ suppression have an increased risk of developing severe medical complications from the treatment, such as malignancies, kidney failure and diabetes.

In a continuing effort to overcome the drawbacks of conventional treatments for autoimmune disease, various methods and pharmaceutical formulations useful for treating autoimmune diseases (and related T-cell mediated inflammatory disorders such as allograft rejection and retroviral-associated neurological disease) have recently been disclosed. These treatments are based on the concept of inducing tolerance, orally or by inhalation, using as the tolerizers autoantigens or bystander antigens or disease-suppressive fragments or analogs of autoantigens or bystander antigens. This body of work has been described in PCT Patent Applications Nos. PCT/US93/01705 filed February 25, 1993, PCT/US91/01466 filed

March 4, 1991, PCT/US90/07455 filed December 17, 1990, PCT/US90/03989 filed July 16, 1990, PCT/US91/07475 filed October 10, 1991, PCT/US93/07786 filed August 17, 1993, PCT/US93/09113 filed September 24, 1993, PCT/US91/08143 filed October 31, 1991, PCT/US91/02218 filed March 29, 1991, PCT/US93/03708 filed April 20, 1993, PCT/US93/03369 filed April 9, 1993, and PCT/US91/07542 filed October 15, 1991, and PCT US95/04120 filed April 7, 1995 (Attorney Docket No. 1010/27956- O0) and PCT US95/04512 filed April 7, 1995 (Attorney Docket No. 1010/27956- O2) .

The present inventors and their co-workers have developed a method of treatment that uses autoantigens and proceeds by active suppression, a different mechanism than clonal anergy (in which the cell loses its ability to respond to subsequent antigen challenge) . This method, discussed extensively in PCT Application PCT/US93/01705) , involves the oral administration of antigens specific to the tissue under autoimmune attack ("bystander antigens"). As used herein the term autoimmune attack refers to the immune events that lead to destruction by the immune system of self-tissue. Oral administration of these "bystander antigens" causes regulatory

(suppressor) T-cells to be induced in the gu -associated lymphoid tissue (GALT) , or, in the case of by-inhalation administration, mucosa associated lymphoid tissue (MALT) . These regulatory suppressor T-cells are released in the blood or lymphatic tissue and then migrate to the organ or tissue afflicted with the autoimmune disease where they suppress autoimmune attack of the afflicted organ or tissue. The T- cells elicited by the bystander antigen (which recognize at least one antigenic determinant of the bystander antigen used to elicit them) are targeted to the locus of autoimmune attack where they mediate the local release of immunomodulatory cytokines, such as transforming growth factor beta (TGF-S) interleukin-4 (IL-4) or interleukin-10 (IL-10) . Of these, TGF- ^ is an antigen-nonspecific immunosuppressive factor in that it suppresses all immune attack phenomena regardless of the antigen that triggers these phenomena. (However, because oral tolerization with a bystander antigen causes the release of

TGF-/3 only in the vicinity of autoimmune attack, no systemic immunosuppression ensues.) IL-4 and IL-10 are also antigen- nonspecific immunoregulatory cytokines. IL-4 in particular enhances Th2 response, i.e. acts on T-cell precursors and causes them to differentiate preferentially into Th2 cells. IL-4 also indirectly inhibits Thl exacerbation. IL-10 is a direct inhibitor of Thl responses.

After orally tolerizing mammals afflicted with an autoimmune disease conditions with bystander antigens, the present inventors and their co-workers observed increased levels of TGF-S, IL-4 and IL-10 at the locus of autoimmune attack. Chen, Y. et al., Science. 2__L_L:1237-1240, 1994.

Naive T-cells differentiate into distinct populations defined by their cytokine secretion pattern and regulatory function. Thl cells secrete predominantly Interleukin-2 (IL-2) and interferon y (IFN7) and are involved in classic delayed- type hypersensitivity reactions, while Th2 cells, which secrete predominantly IL-4 and IL-10, induce selected immunoglobulin secretion and can down-regulate Thl-mediated immune responses (Mossmann et al., 1986; Romagnan, S., 1994). Thus, cytokines secreted by T-cells after activation qualitatively influence the nature of the immune response. Recent reports have suggested that the dominant factors determining the maturation of naive T-cells into either predominantly Thl or Th2 cells are cytokines, particularly IL-2 and IFNγ for Thl and IL-4 for Th2 (Seder and Paul, 1994) . Other factors include the dose and type of antigen, the type of antigen-presenting cell (APC) , and co-stimulatory molecules involved in activation. Until the present invention, once differentiated into Thl or Th2 type, T-cells were thought to be committed to the production of a given set of lymphokines upon restimulation.

T-cell secretion of certain cytokines is also implicated in the induction and regulation of autoimmune inflammatory disease, which is mediated by activated autoreactive T-cells that recognize self-tissue-specific antigen in the context of major histocompatibility complex

(MHO class II molecules on APCs. In experimental autoimmune encephalomyelitis (EAE) , a model for multiple sclerosis, it is

believed that T-cell secretion of IL-2, IFNγ, and tumor necrosis factor (TNF) mediates inflammation and tissue damage, while the secretion of IL-4, IL-10, and transforming growth factor-3l (TGF-31) by myelin basic protein (MBP) -reactive T- cells is associated with potent suppressor activity and down- regulation of central nervous system inflammation (Miller et al., 1992). The induction of TGF-01 secretion appears to be of particular importance in regulating EAE, as anti-TGF-01 monoclonal antibodies inhibit the suppressor effect of regulatory MBP-reactive T-cells (Chen et al., 1994). Thus, autoreactive T-cells are not necessarily pathologic and can function to down-regulate immune responses associated with tissue inflammation locally. Similar findings have been made in humans affected with multiple sclerosis. TNF has been isolated from central nervous system plaques of MS patients, IFN-γ exacerbates attacks, and TGF-/3 is secreted by T-cells of MS patients tolerized with myelin.

It has recently become clear that T-cell activation in general is not a simple on/off response. Allen and coworkers (Evavold and Allen, 1991; Evavold et al., 1993a) demonstrated that stimulation of a T-cell clone with modified peptide-antigens that trigger a suboptimal T-cell receptor (TCR) signal can induce cytokine secretion or IL-2 receptor expression and cytolysis without thymidine incorporation. Altered peptides have also been shown to act as TCR antagonists or to induce anergy in Thl and Th2 clones (Alexander et al. , 1993; De Magistris et al. , 1992; Jameson et al., 1993; Ostrov et al., 1993; Rupper et al., 1993; Sloan-Lancaster et al., 1993, 1994) . However, none of these publications involved work with an autoimmune disease or an animal model thereof.

There is still a need in the art for methods and compositions for abating immune responses associated with autoimmune inflammatory disease and for treating mammals, including humans afflicted with autoimmune disease.

Objects of the Invention

One object of the invention is to provide additional methods and compositions for combatting autoimmune disease.

Another object is to provide additional methods and compositions for combatting immune responses associated with autoimmune disease.

A third object is to provide methods and compositions for converting T-cells mediating immune response associated with autoimmune disease to a benign or regulatory type, thereby reducing immune response and/or suppressing it.

Summary of the Invention

Disclosed are methods and compositions for altering the cytokine pattern of activated autoreactive T-cells, and for diminishing immune responses associated with autoimmune disease.

Brief Description of the Figures

Figure 1 shows the T-cell receptor (TCR) and major histocompatibility complex (MHC) contact points for native (cognate) peptide ligand MBP p85-99; (arrows up indicate TCR contact points; arrows down indicate MHC contact points) .

Figure 2 is a graph of the ability of cognate and altered peptide ligands to bind to T-cell receptor (counts per minute of tritiated thymidine incorporation) at various peptide concentrations (μg/ml) ; dark triangles: p(91A); dark squares: p(90A); dark diamonds: p(90D); dark circles: p(90K); open circles: p(92A); open triangles: cognate peptide ligand MBP p(85-99); and dark squares: no antigen (negative control).

Figure 3A is an autoradiograph showing phosphorylation of CD4-associated tyrosine kinase pSβ" stimulated by: no antigen (control), MBP p(85-99), and altered peptides p(90A), p(91A) and p(93A) .

Figure 3B is a fluorograph depicting Ca ⁺⁺ flux induction (410nm/480nm fluorescence ratio vs. time in seconds) in T-cells presented with APC pulsed with MBP p(85-99) and p(90A) ; no antigen was used as a negative control and ionomycin and EGTA were used as positive controls.

Figure 3C is a bar diagram showing IL-4 secretion by T-cells presented with: no antigen; MBP p(85-99); no antigen

in the presence of PMA and ionomycin; p(90A); p(90K); no antigen + PMA; p(90A) + PMA; p(90K) + PMA; no antigen + ionomycin; p(90A) + ionomycin; p(90K) + ionomycin.

Figure 4 is a Southern Blot showing modulation of m- RNA encoding TGF-/31 and IL-4 after T-cell stimulation with: no antigen; MBP p(85-99); p(90A); and p(93A).

Figure 5 is a series of bar graphs showing the secretion (in pg/ml) by T-cells of various cytokines (TGF-01,

IL-2, IL-4, IL-10 and γ-IFN) upon stimulation with: no antigen; cognate peptide ligand p(85-99); and altered peptide ligands

(p89K) ; p(90A); p(90D); p(91A); p(92A); p(93A).

Figure 6 contains three graphs depicting the secretion in picograms/ml of IL-4, IFN-γ and TGF-S1 by T-cells stimulated with APCs pulsed with various concentrations of MBP p(85-99) and p(90A) . Levels generated by p(90a) are shown as dark squares; those generated by native Mbp p(85-99) are shown as open circles.

Figure 7 is a graph showing mean disease scores for animals immunized with a native peptide W144 alone (open boxes) or coimmunized with W144 and Q144 at a ratio of 1:5 (open squares) and assessed daily for signs of disease.

Figure 8 consists of three graphs showing the proliferative response of LNC from mice immunized with W144 (a) , Q144(b) or W144 and Q144(c). Figure 9 consists of four graphs showing the proliferative response and cytokine production of T-cell lines specific for native 144( LI) or Q144(QL1) peptides. Graphs a and b show proliferative response while graphs c and d show cytokine production. "W" is 144 (the native peptide) , "Q" is Q144, "LR" is L144/R147, and "A" is A144.

Figure 10 shows cytokine production for T-cells specific for Q144. The amount of cytokine produced for clones of QLl is displayed in terms of IFNγ versus IL-10 and IL-4 versus IL-10 to determine whether the Q144-specific clones were of the THO (IFNγ, IL-10) or TH2 (IL-4, IL-10) phenotype.

Figure 11 is a graph of mean disease scores for SJL mice immunized with W144, Q144 or PLP 190-209. Results shown

are the mean disease scores for each group from three independent experiments.

Detailed Description of the Invention Definitions

"Cytokine pattern" of a T-cell means the cytokines secreted by that T-cell.

"Immune response enhancing cytokine pattern" or "Th-0 cytokine pattern" means cytokines secreted by an activated (immune attack) T-cell and including one or more of IL-2, TNF and IFN-γ, but not substantial amounts of TGF-/S, IL-4 or IL-10.

"Immune response regulating cytokine pattern" means cytokines secreted by a regulatory T-cell including one or more of TGF-S, IL-4, or IL-10 but not substantial amounts of IL-2, TNF or IFN-γ.

"Cognate peptide ligand" means a peptide consisting essentially of an antigenic determinant against which a T-cell has been elicited, and which a T-cell recognizes and proliferates to. "Altered peptide ligand" means a peptide differing in amino acid sequence from a cognate peptide ligand in at least one amino acid residue.

"TCR contact point" or "TCR contact residue" means an amino acid residue within a peptide ligand that participates in binding of the ligand to a T-cell receptor that recognizes it.

"MHC contact point" or "MHC contact residue" means an amino acid residue within a peptide ligand that participates in presentation of the ligand in connection with the major histocompatibility complex (MHC) of an antigen-presenting all

(APC) .

"Cytokine-affecting TCR contact point" is a TCR contact residue which upon conservative or non-conservative substitution with another amino acid residue causes the cytokine pattern of the T-cell to switch towards a regulatory cytokine pattern.

The notation:

MBP p(85-99) or p(85-99) means the cognate peptide ligand MBP peptide having the sequence ENPWHFFKNIVTPR, i.e. amino acid residues 85-99 of myelin basic protein (MBP) .

It is also called the minimum immunodominant domain of MBP p. (82-104) which has the sequence: DENPWHFFKNIVTPRTPP.

MBP p85-99(93A) or p(93A) means an altered peptide ligand wherein the 93d residue of MBP (K in the cognate peptide) has been replaced by A (alanine) .

MBP immunodominant regionpeptide p(143-162) has the amino acid sequence FKGVDAQGTLSKIFKLGGRD.

MBP minimum immunodominant peptide p(148-162) has the sequence AQGTLSKIFKLGGRD.

PLP immunodominant region peptide p(31-50) has the sequence LFCGCGHEALTGTEKLIETY. PLP immunodominant regionpeptide p(181-200) has the sequence WTTCQSIAFPSKTSASIGSL. An altered PLP peptide consisting of p(139-151) (which is immunodominant in mice) with a substitution at residue 144 has suppressed immune responses associated wtih EAE (experimental autoimmune encephalomyelitis) .

The above one-letter amino acid abbreviations correspond to the represented amino acids (or residues) as follows: A«alanine, C=cysteine; D-aspartic acid; E-glutamic acid; F=phenylalanine; G=glycine; H=histidine; I=isoleucine; K=lysine; L=leucine; M«methionine; N=asparagine; P=proline; Q-*glutamine; R«arginine; S-serine; T=threonine; V-valine; =tryptophan; and Y=tyrosine.

The present inventors have found that stimulation of human autoreactive cell clones previously determined to recognize immunodominant epitopes (in humans) of myelin basic protein and to be of the ThO type (i.e. secreting immune response promoting cytokines) with certain altered peptide ligands presented by antigen-presenting cells (i.e. altered- ligand/MHC. complexes) induce a switch in the cytokine pattern of these activated (immune attack) T-cells, and cause them to secrete suppressive cytokines.

The altered peptide ligands that could effect the cytokine pattern switch differ in amino acid sequence from the

cognate peptide ligand of the T-cell clones by one amino acid at a TCR contact point.

This demonstrates that, even after their differentiation, function of human T-cells can be regulated by choice of antigen. More specifically, this demonstrates that autoreactive immune attack mature human T-cells can be converted to benign or regulatory T-cells, thereby abating inflammatory immune response associated with autoimmune disease. Phenylalanine at residue 91 of MBP p(85-99) represents a primary TCR contact residue for MBP p(85-99) .

Altered ligands substituted at this residue do not elicit any response from T-cells recognizing MBP p(85-99). It is accordingly preferred to use as altered peptide ligands analogs of MBP p(85-99) that are not substituted at position 91.

Histidine at residue 90 and lysine at residue 93 are secondary TCR contact residues. Any amino acid substitution at those residues (both conservative and nonconservative) will cause a change in the cytokine pattern of the inflammation- promoting T-cells towards the suppressive cytokine pattern.

The preferred altered peptide ligands will be the ones that will cause patients' T-cells to secrete the highest levels of suppressive cytokines, especially TGF-/3.

It is also preferred that the T-cell clone not proliferate in response to the altered peptide ligand but merely switch its cytokine pattern.

The altered peptide ligands efficiently inhibit antigen-dependent T-cell proliferation and can induce IL-4 or

TGF-31. Thus, the altered peptide ligands appear to engage the TCR with comparable affinity to the cognate peptide but trigger a different set of signals, leading to an altered cytokine pattern, a regulatory cytokine pattern.

A change in cytokine secretion from a Thl (in humans

ThO) type T-cells to a TGF-/J or IL-4 secreting T-cell has important physiological consequences because suppressive cytokines inhibit immune response including immune responses to self-antigens.

Alteredpeptide ligands parenterally administered can thus confer attenuation of autoimmune responses by causing autoreactive T-cells to switch to a suppressive cytokine pattern, and thereby suppress inflammatory immune responses. Effective amounts for administration will range generally from about l to about 50 mg/patient/treatment for any autoimmune disease and any altered peptide ligand. Preferred amounts for altered peptide ligands derived from MBP p(85-99) will be from 5-15 mg/patient/treatment. Treatment frequency can be once a month. Treatment duration can be as long as benefits persist for a period of months or even several years.

The therapy dose, frequency of administration and regimen will be subject to optimization according to the disease, the altered peptide ligand employed, the physical condition of the patient, and the levels of regulatory cytokines elicited in the T-cells with the altered cytokine patterns, as is well-appreciated by those skilled in the art.

More than one altered peptide variant can be administered to a patient simultaneously with or separately from another altered peptide ligand.

The altered peptide ligands are preferably administered in a physiologically acceptable medium or diluent, suitable for injectable preparations. Any parenteral route of administration may be employed, e.g. subcutaneous, intraperitoneal or intramuscular or intravenous injection.

Procedure for Identifying Coσnate and Altered Peptide Liσands The altered peptide ligands of the present invention can be identified by using the following procedure. As the first step in this procedure, T-cells that recognize a known antigen (or a known bystander antigen) , are isolated. If not already known, the immunodominant region(s) of the antigen or bystander antigen are thereafter determined by synthesizing overlapping peptides that span the amino acid sequence of the entire antigen. It is contemplated that such peptides will be of at least about 15 residues and preferably about 20 residues (and more preferably between 18 and 20 residues) in length.

Each of the overlapping peptides is then tested by exposing such peptide to a panel of T-cell clones that are isolated from patients afflicted with the autoimmune disease to be treated and which recognize the antigens. The procedure for T-cell isolation is (as well as for identification of the immune dominant epitopes) disclosed in more detail for MS in PCT US93/03369, filed April 9, 1993 may be employed to create T-cell panels for other autoimmune diseases by appropriate substitution of patients and antigens. The procedure can also be used for individual patients, if desired.

The T-cell panels are used to identify the immunodominant epitope of the antigen, as follows:

Each of the overlapping peptides is exposed to an antigen presenting cell (APC) , i.e. a cell which presents that antigen to the T-cells that recognize such antigen. The T-cell lines from the panel are tested for binding to each overlapping peptide. The overlapping peptide (or peptides) that are recognized by a substantial number of the T-cell panel members are then identified using this procedure. That is, the overlapping peptide or peptides selected are those recognized by the largest number of T-cell panel members.

Once the overlapping peptide containing the immunodominant epitope has been identified, the smallest immunodominant fragment of the overlapping peptide (i.e. the smallest fragment that binds to T-cells which recognize the antigen with nearly the same affinity as the entire overlapping peptide containing the immunodominant epitope) must be located. This is accomplished by creating a series of truncated peptides by sequentially eliminating one amino acid residue from one and then from the other terminal of the original overlapping peptide. Each of the truncated peptides is tested for binding to MHC and for induction of proliferation of T-cells that recognize the overlapping peptide containing the immunodominant epitope. These properties are compared to the same properties of the original untruncated overlapping peptide. Thus is the so-called minimum immunodominant epitope peptide (MIEP) .

Once the MIEP has been identified, those amino acid residues of the MIEP that are (a) MHC contact points or (b) TCR

contact points must be identified. A panel of MIEP analogs each differing from the MIEP by a single amino acid substitution at a different position is then constructed.

Each MIEP analog is thereafter tested for binding affinity to the MHC of an APC (for the antigen or bystander antigen) and for binding (as an analog/MHC complex) to a T-cell clone that recognizes the immunodominant epitope of the antigen. (This has also been described in PCT US93/03369 and the procedure can be easily adapted to other MIEP's from other immunodominant epitopes for MS and other antigens specific to other autoimmune diseases.) The measured binding affinity of each MIEP analog to the APC and to the TCR is compared to the measured binding affinity of the MIEP to the APC (MHC) and the TCR. If substitution of one amino acid residue of the MIEP results in reducing the binding affinity for the MHC, that residue is an MHC contact point. In creating the altered peptide ligand residues that have been identified as MHC contact points should not be substituted.

If the substitution of one amino acid residue of the MIEP results in reducing the binding affinity for the TCR, that residue is a TCR contact point. Each TCR contact point can be a substitution site. Preferred substitution sites are those at which substitution of a non-native amino acid residue does not permit the MIEP recognizing T-cell to proliferate upon encountering the MIEP analog, but merely cause the T-cell to switch to a regulatory cytokine pattern.

Design of Altered Peptide Ligands Useful In the Present Invention

Altered peptide ligands for use in the present invention are created by substitution of a single MIEP amino acid residue with a different amino acid residue at a TCR contact point. The identity (i.e. composition) of the amino acid residue that is used to replace the TCR contact point of the MIEP is not critical and any of the remaining 19 essential amino acids may be employed for this purpose. The substitutions can thus be both conservative and

nonconservative; the substituted amino acid residue can have the same or a different charge as the original (native) MIEP residue (or can have no charge at all) . (For example, as illustrated in Table 1 below substitutions at positions 90, 91 and 93 yield ligands that bind to the T-cells recognizing the MIEP MBP p(85-99) with affinity comparable to that of the native MBP p(85-99) differing by less than one order of magnitude. Substitutions at position 90 and 93 are preferred because the T-cells do not proliferate upon encounter of the altered peptide ligand.)

Using this procedure (in which one TCR contact point at a time is replaced with a different amino acid residue, a panel of altered peptide ligands is assembled. Each panel member is then assayed for binding to a T-cell that recognizes the MIEP, ability to cause such a T-cell to proliferate, and for the cytokine pattern that it elicits in such T-cell.

Altered peptide ligands that bind to the T-cell receptor with comparable affinity to that of the nature MIEP and that cause the T-cell to switch to a regulatory cytokine pattern are selected.

The assay for determination of the T-cell cytokine pattern before and after exposure to an altered peptide ligand is described below with specific reference to T-cell from DR2+ patients that recognize MBP P(85-99) and altered peptide ligands derived from MBP p(85-99) . This procedure can be readily adapted to T-cells of different specificities, different antigens involved in different autoimmune diseases.

Briefly, each altered peptide ligand is contacted with peripheral T-cells isolated from a patient afflicted with the target autoimmune disease in the presence of irradiated mononuclear cells (used as antigen presenting cells) . Preferred candidate peptides for therapy are those which induce the highest levels of release of TGF-01 (or IL-4) without inducing cell division (monitored by thymidine incorporation) or the release of any other immune response-enhancing cytokines.

As indicated above, the techniques described herein, can be used for identification of immunods mirant epitopes of

other antigens involved in other autoimmune diseases. Antigens specific to the afflicted tissue or organ recognized by activated (immune altered) cells from patients suffering from an autoimmune disease can be subjected to the foregoing procedure for identification of immunodominant epitopes, MIEP's and altered peptide ligands that cause activated autoreactive T-cells to switch to a regulatory cytokine pattern. Non- limiting examples of antigens to which there is autoimmuήity in other autoimmune diseases include Type II and Type I collagen for rheumatoid arthritis and glutamic acid decarboxylase for autoimmune (e.g. Type I) diabetes - See PCT application US 93/09113, filed 9/24/93 which discloses collagen fragments and collagen peptides and Yamashita, K. et al. Biochem. Biophys. Res. Comm. 1993, 192: 1347-1352. The invention is illustrated below by various examples which are not however intended to limit its scope.

EXAMPLE 1

Materials and Methods

T-cell clones

T cell clones 0blA12 and 0b3Dl that recognize MBP p(85-99) were generated from peripheral blood mononuclear cells of a MS patient by limiting dilution cloning as described previously (Ota, et al., 1990). The clones were maintained in fetal calf serum, 10% dimethyl sulfoxide (DMSO) in liquid nitrogen. Aliquots of frozen T cell clones were thawed, washed in RPMI 1640, 20% FCS and restimulated with lμg per ml phytohemagglutinin (PHA.P) and allogenic peripheral mononuclear cells (PBMCs) as feeder cells in complete medium (RPMI 1640 supplemented with 10% pooled human A/B serum, 2mM L-glutamine, lOmM Hepes and 100 U/100 μg per ml penicillin/streptomycin) . After 3 days IL-2 (human T cell-Stim, Becton Dickinson & Co», Mountain View, CA) was added to a final concentration of 5%. The medium was changed every 3 to 4 days. Cells were expanded for 8-10 days, 16 hours before the experiments cells were put into IL-2 free medium.

Binding of MBP peptides to MHC

MBP p85-99 (ENPWHFFKNIVTPR) and altered peptides were synthesized in the Biopolymer Laboratory, Harvard Medical School, by automated solid phase methods using FMOC-protected amino acid precursors and purified by reverse-phase HPLC. Peptides were greater then 98% pure on the basis of HPLC analysis. Binding of MBP peptides to purified DRB1 ^* 1501 (human TCR) molecules was determined as previously described (Wucherpfennig, et al., 1994). Briefly, purified DRB1 ^* 1501 molecules (1 to 5 μM) were incubated for 48 h with 100 nM ¹²³ I- radio labeled peptides and various doses of unlabelled MBP analog peptides in the presence of a protease inhibitor cocktail. Relative capacity of MBP peptides to inhibit the binding of MHC molecules and radiolabeled ligands, expressed as a concentration of MBP peptide (μM) yielding half-maximal inhibition (IC 50%) , is the measure of MBP peptide affinity for MHC isotypes.

T cell proliferation assays For T cell stimulation assays, the homozygous B cell line 9010 (DR2/DQ6) was used as APCs. APCs '..'ere pulsed with peptide at l - 100 μg/ml for 2 hours, washed and irradiated with 5000 rads. T cells (5xl0 ⁴ ) and APCs (2xl0 ⁴ ) were then co- cultured in triplicate for 3 days in 200 μl complete medium at 37°C. During the last 18 hours of culture lμCi of ³ H-thymidine was added to each well and T-cell proliferation measured as ³ H- thymidine incorporation. In Table 1, +++represents proliferation (stimulation index >3) at lμg/ml, ++ represents proliferation at lOμg/ml and + proliferation only at lOμg/ml of peptide.

TCR antagonist assay

APCs (B cell line 9010) were prepulsed with a suboptimal dose of 2 μg/ml MBP p85-99 peptide for 2 hours, irradiated and washed, followed by a second pulse with the altered peptide ligands at 0.1, 1, 10 and 100 μg/ml for 2 hours. After another wash the APCs were added to the T-cell clone OblA12 (5x10* APCs and 2xl0 ⁵ T cells per well in

triplicates) and co-cultured for 3 days. Proliferation was measured as thymidine incorporation (in cpm, see Fig. 1) after an 18 hour pulse. Native MBP p85-99 and MBP p85-99(93A) , which are agonist peptides, were used as positive controls.

Phosphorylation of Tyrosine Kinase se^

Immunoprecipitation of CD4 and in vitro kinase assay were performed as previously described (Hollsberg, et al., 1994; Rudd, et al., 1988). In brief, 0blA12 was stimulated with peptide pulsed APCs for 1 hour, washed in PBS and lysed in a buffer containing 1% Triton X-100, 20mM Tris, 150 mM NaCl and 1 mM phenylmethylsulfonyl fluoride (PMSF) . Lysates were precleared with Staphylococcus aureus cells and incubated with OKT4-coated protein A sepharose beads for 1 hour. The immunoprecipitate was washed twice in lysis buffer and incubated in a [μ ³² P] -ATP kinase buffer for 10 minutes. The reaction was stopped by the addition of 2 volumes of sample buffer. Samples were boiled for 5 minutes and separated on a 10% SDS gel. Autoradiography was done for 30 minutes on Kodak XAR 5 film, and the results are shown in Fig. 3A.

Ca ⁺ -flux

T cells (10 ⁷ /ml) were loaded with indo-1 acetoxymethyl ester (Molecular Probes, Junction City, OR) at 2 μg/ml in culture medium for 45 min. at 37°C and then diluted 1:10 in culture medium. APCs were pulsed with peptide at 100 μg/ml for 2 hours, washed in RPMI and resuspended in culture medium. T cells (10 ⁶ ) and APCs (5x10*) were run on an Epics V flow cytometer to establish a baseline. Then cells were spun for 1 min. at 1200 rpms to establish cell to cell contact and rerun on the flow cytometer for up to 600 seconds. After Ca ²⁺ binding the indo-1 dye exhibits changes in fluorescence emission wavelength from 480 to 410 nM. The ratio of 410:480 nM indo-1 fluorescence was recorded vs time and expressed in arbitrary units, one arbitrary unit representing approximately 200 nM increase of Ca ²⁺ . Addition of ionomycin (4 μg/ml) was used as a positive control and EGTA (5μg/ml) as a negative control. The results are shown in Fig. 3B.

Measurement of cytokine mRNA

B cells (9010) were pulsed with the peptides as described above. 10 ⁵ T cells were then co-cultured with 4xl0 ⁴ APCs in 200 μl medium in U-bottomed wells for 4 hours followed by RNA extraction and cDNA synthesis. Total cellular RNA was extracted using the RNAzol B method (Teltest, Inc., Friendswood, TX) . RNA was co-precipitated with 10 μg of transfer RNA in isopropanol overnight. For cDNA synthesis the pellet was resuspended in 6μl of sterile double distilled water, 5μl of oligo-dT (Sigma) and random hexamers (Promega) were added and the samples heated to 70°C for 10 minutes. Then 4μl of 5x buffer, 2μl of 0.1 M DTT and lμl each of lOmM dNTPs, RNAsin and M-MLV reverse transcriptase (all from Promega) were added. cDNA synthesis was carried out at 37°C for 60 minutes and water added to a final volume of 200 μl. Samples were normalized to the jS-actin cDNA concentration as determined by competitive RT-PCR using a non-homologous DNA fragment of known concentration, that has the same primer annealing sites and amplifies with the same efficiency as the internal standard (data not shown) . The PCR for TGF-j8 and IL-4 was semiquantitative with 28 amplification cycles at 94°C denaturation (1 min.), 60°C annealing (1 min.), 72°C extension (90 sec), which was determined to be within the linear range of the PCRs. Concentrations in the PCRs were as previously reported ( ucherpfennig, et al., 1990), except for the addition of 0.6 μCi of ³² P labeled dCTP to each reaction. Forward primer for TGF-01 was 5'-GCC CTG GAC ACC AAC TAT TGC-3', reverse primer 5'-GCT GCA CTT GCA GGA GCG CAC-3'; PCR product length is 336 base pairs. Primers for IL-4 were as follows; forward primer 5' -CTG CTA GCA TGT GCC GGC AAC TTT GTC CAC-3' , reverse primer 5'-GAA GTT TTC CAA CGT ACT CTG GTT GGC TTC-3', length of PCR product is 365 bp.

Radioactive PCR products were separated on 5% polyacrylamide gel and visualized by autoradiography.

Cytokine ELISA

10 ⁶ T cells from OblA12 were estimated in 5 ml polypropylene tubes with 4x10 ^s peptide pulsed APCs (9010 B cells) pulsed with peptide for 2 hours. Culture supernatants for IL-2, IL-4, IL-10 and IFN-μ were taken after 40 hours. To measure secretion of active TGF-01 the cells were then washed carefully 3 times in RPMI in order to remove all serum and resuspended in serum-free medium X-Vivo 20 supplemented with 2U rIL-2/ml and 2U rIL-4/ml (Boehringer Mannheim) . Supernatant for TGF-jSl ELISA was harvested after 72 hours from the start of experiment. The ELISA was done for active TGF-/S1 and was performed as described previously (Miller, et al., 1993). For the detection of IL-4 and IL-10 primary and secondary antibodies were purchased from Pharmingen and used following the manufacturer's directions. IL-2 was assayed using the ELISA kit from Endogen. For IFN-μ the primary antibody was obtained from Genzy e, the secondary from Biosource International. Assays were done in duplicate. To study the effect of ionomycin and PMA on IL-4 secretion T cells (2xl0 ⁵ and APCs (10 ^s ) were co-cultured in triplicate wells for 40 hours in 200 μl complete medium supplemented with ionomycin at

100 ng/ml or PMA at 10 ng/ml. The results are in Fig. 3C.

Results

Binding of Altered Peptide Ligands to Human TCR, DRB1.1501 and Peptide Fine Specificity of T-Cell Clone OblA12

Altered peptide ligands derived from MBP p85-99 were examined for binding to DRBl ^* 150l in an MHC competition assay and for the ability to stimulate MBP p85-99-reactive T-cell clone OblA12 (Table 1) . This clone did not proliferate or incorporate thymidine after stimulation with MBP p85-99 substituted at amino acid positions 90 and 91, while an alanine substitution at position 93 did not significantly alter the T- cell response, indicating that positions 90 and 91 are the critical TCR contact points for this cxone. Whereas substitutions at positions 89 and 92 significantly reduced binding of the altered peptides to DRB1 ^* 1501, substitutions at positions 90, 91 and 93 did not reduce binding, confirming that

positions 89 and 92 in MBP p85-99 are the major MHC contact points.

TCR Antagonist Properties of Altered Peptide Ligands

The TCR antagonist properties of altered peptide ligands were assessed to aid in screening peptide ligands that bound to the TCR. The biologic activity of peptides with substitutions at position 90 was examined in a prepulse TCR- antagonist experiment in which T-cells are stimulated by APCs that were pulsed with a limited concentration of native peptide, followed 2 hours later by stimulation with increasing concentrations of peptide antagonist. This assay was developed to avoid competition between antigen and antagonist at the MHC level and thus allow measurement of events at the TCR level (De Magistris et al., 1992). TCR antagonism has been attributed to active signaling events resulting in the inhibition of early biochemical events required for T-cell activation rather than just competition for TCR binding with the native peptide

(Ruppert et al., 1993). As shown in Figure 1, MBP p85-99 peptides altered at position 90 could inhibit proliferation induced by the native peptide, suggesting that they could act as a TCR ligand and signal in the context of class II MHC. These altered peptide ligands did not induce clonal unresponsiveness as the T-cells but showed a full proliferative response to MBP p85-99 7 days after the initial peptide stimulation (data not shown) . MBP p85-99 with an alanine substitution at position 91 (MBP p85-99[91A]) inhibited proliferation of 0blA12 to a substantially lower degree, even at concentrations of 100 μg/ml, suggesting that this peptide- MHC complex had a low TCR affinity as compared with the peptides with substitutions at position 90 (Figure 1) . As expected, peptide MBP p85-99(93A) acted as an agonist similar to the native peptide, indicating that position 93 is not a critical TCR contact point.

Signaling Properties of Altered Peptide Ligands

Antigen-specific activation of T-cells results in a cascade of events involving Ca ²⁺ influx and activation of the CD4-associated protein tyrosine kinase pδδ" leading to a mobility shift of pSδ ¹ * to p60 ^kk on gel electrophoresis (Barridge, 1993; Rudd et al., 1994). To examine further the signaling properties of the modified peptides and to provide definitive evidence of TCR ligation, we investigated the biochemical consequences of ligating the TCR with MBP p85-99 and the substituted peptides. CD4-associated pSβ" was immunoprecipitated and tested in an in vitro autophosphorylation assay. Whereas MBP p85-99 and MBP (93A) , both of which induced thymidine incorporation in OblA12, were able to induce a mobility shift in p56 ^k * to p60 ^kk , MBP p(90A) and MBP p(91A) did not (Figure 3A) . Likewise, MBP p85-99 induced a strong Ca ⁺⁺ influx in contrast with p(90A) (Figure 3B) . T-cells can be activated by a combination of ionomycin and phorbol myristate acatate (PMA) , which induce a Ca ⁺⁺ flux and direct protein kinase C activation (resulting in pδβ ¹ * activation), respectively. As p(90A) signaled the T-cell without a measurable Ca ⁺⁺ flux, it was of interest to examine

[ ³ H]thymidine incorporation and cytokine secretion in the presence of either ionomycin or PMA with antigen. Stimulation of OblA12 with p(90A) in the presence of ionomycin induced significant amounts of IL-4 (Figure 3C) but no IL-10 of IFNγ secretion or [ ³ H]thymidine incorporation. These findings indicate that the p(90A)DRB1.1501 complex provided a partial TCR signal.

T-Cell Stimulation with Altered Peptide Ligands Specifically Induces TGF-3 mRNA

Direct measurement of cytokine mRNA levels following stimulation with altered peptide ligands substituted at the TCR contact residues were examined by reverse transcriptase- polymerase chain reaction (RT-PCR) . Stimulation of OblA12 with MBP p85-99 or p(93A) induces mRNA expression of IL-2, IL-4, IL- 10, and IFNγ but not TGF-31 as shown for IL-4 and TGF-fll in Figure 4. The altered peptide ligands that failed to induce

thymidine incorporation also failed to induce IL-2, IL-4, IL- 10, and IFNγ mRNA synthesis at different timepoints after T- cell stimulation (data not shown) . However, TGF-/J1 mRNA was induced after stimulation of 0blA12 with MBP p85-99(90A) (Figure 4A) . A second human MBP p85-99-reactive T-cell clone, 0b3Dl, from the same patient with a different peptide fine specificity was examined. This clone also has a ThO cytokine profile and proliferates to MBP p85-99 and (90A) , while substitutions at position 93 result in loss of reactivity, indicating that the lysine at this position is a critical TCR contact residue. Stimulation with MBP p85-99(93A) but not MBP p85-99 or (90A) induced significant MRNA expression of TGF-01 (Figure 4B) .

Switch in Cytokine Secretion after T-Cell Stimulation with Altered Peptide Ligands

The secretion of cytokines into the culture supernatant after stimulation of T-cell clone 0blA12 with altered peptides was measured by enzyme-linked immunosorbent assay (ELISA) (Figure 5) . In accordance with the findings of mRNA expression, IL-2, IL-4, IL-10, and IFNγ were secreted after stimulation with the wild-type peptide. Secretion of these cytokines correlated with T-cell proliferation and was abolished when peptides with substitutions at TCR contact points were used for stimulation. Reduced cytokine secretion after stimulation with MBP p85-99(89K) and p(92A) corresponds to the reduced MHC binding of these peptides as shown in Table 1. Only MBP p85-99(90A) and (90D) induced TGF-01 secretion, corresponding to the induction of mRNA synthesis. Hence, these MBP analog peptides switched cytokine secretion patterns by selectively inducing mRNA synthesis and protein secretion of TGF-31 in the absence of thymidine incorporation or secretion of IL-2, IL-4, IL-10, or IFNγ. TGF-01 secretion could not be detected by ELISA in clone Ob3Dl in spite of specific induction of TGF-S1 mRNA synthesis, which might be due to either postranscriptional regulation or autocrine usage of the secreted TGF-J1 by this cell clone.

To eliminate the possibility that the induction of TGF-S1 secretion by the altered peptide ligand was due to an attenuated T-cell activation, the consequences of lowering the native peptide concentration for the induction of TGF-01 synthesis was examined. 0blA12 was stimulated with either the native MBP peptide or (90A) peptide at lower concentrations and the secretion of IL-4, IFN-γ, and TGF-jSl measured. With decreasing peptide concentrations, T-cell activation was attenuated as measured by IL-4 of IFN-γ secretion (Figure 6) or [ ³ H]thymidine incorporation (data not shown) without the induction of TGF-jSl secretion. This suggests that the induction of TGF-j8l secretion by altered peptide ligands was not due to attenuated T-cell activation. In addition, costimulation with ionomycin and PMA with the altered ligands did not induce TGF-/31 secretion.

The production of active TGF-81 is not exclusively transcriptionally regulated, since a latent form of the cytokine is secreted from the cell and postsecretional activation must occur (Wahl, 1991) . Thus, in these experiments, we measured the production of the active TGF-01 by ELISA. Since TGF-31 can be released from dying cells, cell viability after 3 days in culture was examined by trypan blue uptake and was found to be similar in all conditions, indicating that the altered peptide ligands did not induce cell death. Although we cannot exclude the possibility that peptide analogs are inducing T-cells to signal the APCs that then secondarily secrete TGG, this seems unlikely, since comparable results were obtained using DRBl ⁺ 1501-transfected L-cells as APCs (data not shown) .

EXAMPLE 2 Example of Patient Treatment with Altered Peptide Ligands

A patient (DR2 ⁺ ) is suffering from relapsing remitting multiple sclerosis and has activated autoreactive T- cells that recognize MBP p(85-99) . It is determined that in

vitro exposure of such T-cells from the patient to APCs presenting altered peptide ligand p(90A) induce these cells to secrete substantially increased amounts of TGF-0, without causing them to proliferate. It is also determined that a p93A/MHC complex binds to the TCR of such T-cells with comparable affinity (different by less than an order of magnitude) as the native MBP p(85-99) . This patient is treated parenterally with 10 mg of p(90A) suspended in saline once a month for three months. The number of activated autoreactive T-cells (isolated from the peripheral blood of the patient or from the cerebrospinal fluid) recognizing MBP p(85-99) is observed to decrease significantly and an increased number of T-cells from this patient secrete TGF-/J.

EXAMPLE 3

In the following studies, we confirmed the results using human T-cells and altered peptide ligands described above by showing that altered peptide ligands representing variants of an immunodiminant epitope in rodents produce the same effects, i.e. after the differentiation of T-cells away from the Thl subtype. Further, we showed that altered peptide ligands suppress induced disease in vivo.

Materials and Methods Animals

Female (4 to 6 week-old) SJL mice were purchased from the Jackson Laboratory (Bar Harbor, Maine) and housed under virus-free conditions.

Antigens

Peptide antigens for in vi tro studies were synthesized on a Milligen™ model 9050 synthesizer using Fmoc chemistry. Milligen™ PAL amide resins were used to produce peptides with C terminal amides. Most peptides were >90% pure, as determined by high pressure liquid chromatography (HPLC) , and were not purified further. For disease induction, HPLC-

purified peptide PLP 139-151, obtained from Alkermeres, Incorporated (Cambridge, Massachusetts) was used in some experiments. The peptides used in these experiments were PLP 139-151 designated W144 (HSLGKWLGHPDKF) , Q144 (HSLGKQLGHPDKF) , A144 (HSLGKALGHPDKF) , T144 (HSLGKTLGHPDKF) , L144/R147 (HSLGKLLGRPDKF) , and PLP 190-209 (SKTSASIGSLCADARMYGVL) . Single letter abbreviations for the amino acids are the same as those indicated above.

Induction and Assessment of EAE

Mice were injected subcutaneously at two sites with W144 emulsified in CFA and supplemented with Mycobacterium tuberculosis H37 RA (500 μg/mouse: Difco, Detroit, Michigan) . On day 0 and 3, each mouse was also injected intravenously with 10 ⁹ heated-killed Bordetella pertussis bacilli (pertussis vaccine lot number 264, Massachusetts Public Health Biological Laboratories, Boston, Massachusetts) . The concentration of W144 was titrated in each set of experiments to give optimal disease and was between 50-100 μg/mouse. In the preimmunization experiments, SJL mice were initially immunized at two sites with peptide emulsified in CFA supplemented with M. tuberculosis H37 RA (400 μg/mouse: Difco) or not preimmunized (controls) . In the coimmunization experiments, mice were immunized with the peptide or a mixture of peptides shown, keeping the concentration of disease-inducing peptide constant. Mice were examined daily beginning on day 9 for disease, which was assessed clinically according to the following criteria: 0, no disease; 1, limp tail; 2, hindlimb weakness; 3, hindlimb paralysis; 4, hindlimb plus forelimb paralysis; 5, moribund or dead. When animals were moribund or at the end of the experiment (day 30-40) , mice were sacrificed and brains and spinal cords were fixed, processed for histologic analysis, and evaluated (Kuchroo et al., 1994).

Culture Media

DMEM supplemented with 0.1 mM nonessential amino acids, sodium pyruvate (1 mM) , L-glutamine (2 mM) , MEM essential vitamin mixture (1 x) , penicillin (100 U/ml) , streptomycin (100 U/ml), gentamicin (0.1 mg/ml) , 10% heat- inactivated fetal bovine serum (Bio Whittaker, Incorporated, Walkersville, Maryland), asparagine (0.1 mM) , folic acid (0.1 mg/ml) , and 2-mercaptoethanol (5 x

10 ^'5 M) (Sigma Chemical Corp., St. Louis, Missouri) was used for the culture of lymph node cells (LNCs) . For the expansion of T-cell lines and clones, this medium was supplemented with 0.6% T-cell growth factor (T-Stim™, Collaborative Biomedical Research, Bedford, Massachusetts) and 0.06% recombinant IL-2.

In Vitro Proliferation Assays

Mice were injected subcutaneously at five sites with antigen emulsified in CFA (Difco Laboratories, Inc.) containing a total of 250 μg M. tuberculosis H37 RA. Mice immunized with a single peptide received a total of 100 μg of antigen; mice immunized with a mixture of W144 and Q144 received 100 μg of W144 and either 100 μg or 300 μg of Q144 (i.e. a total of 200 or 400 μg of antigen per mouse) . On day 10, lymph nodes were removed and LNCs prepared from them. LNCs (4 x 10 ^s per well) were cultured in triplicate in 96-well roun -bottomed plates

(Falcon™, Becton Dickinson, Lincoln Park, New Jersey) , in the presence of antigen, for 48 hr and then [ ³ H]thymidine (1 μCl/well) was added for the last 16 hr before harvesting the cells. The [ ³ H]thymidine incorporation was determined in a scintillation counter (model LS 5000; Beckman Instruments Inc., Fullerton, California) . The stimulation index (SI) was calculated as mean cpm with antigen/mean cp with medium plus APCs alone.

In Vi tro Cvtokine Assays

Supernatants were collected from LNCs (4 x 10 ^J per well) . T-cell lines (5 x 10* T-cells plus 5 x 10 ^s syngeneic irradiated spleen cells per well) or T-cell clones 24 or 40 hr after activation in vitro. The concentration of IL-4 in the supernatants of LNC was measured by ELISA or by using CT4S cells, which were maintained in culture in medium supplemented with recombinant IL-4. Prior to assay, CT4S cells were kept overnight in IL-4-free medium, harvested, and washed three times, then resuspended at 1 x 10 ⁵ cells/ml. Aliquots (50 μl) were dispensed into 96-well flat-bottomed plates and incubated either with known concentrations of recombinant mouse IL-4 (Pharmingen, San Diego, California) or supernatants from stimulated LNCs. After 24 hr, the plates were incubated for 15 hr with added [ ³ H]thymidine, harvested, and the radioactivity counted. Standard curves were derived from cells exposed to known concentrations of IL-4.

The concentrations of IFNγ, IL-2, IL-10 (and IL-4 in the supernatants collected from T-cell lines or T-cell clones) were measured by quantitative capture ELISA according to the guidelines of the manufacturers. In brief, purified rat MAb to mouse cytokine IL-2 (clone JES5-1A12) , IL-4 (clone BVD4- 1D11) , IL-10 (clone JES5-2A5) , and IFNγ (clone R4-6A2) were obtained from Pharmingen (San Diego, California) and used to coat ELISA plates (Immulon™ 4, Dynatech Laboratories Inc., Chantilly, Virginia) . Recombinant mouse cytokines (IL-2, IL-4, IL-10 and IFNγ: Pharmingen) were used to construct standard curves and biotinylated rat MAb to mouse cytokine IL-2 (clone JES6-5H4), IL-4 (clone BVD6-24G2) , IL-10 (clone SXC-1) , and IFNγ (clone XMG1.2) (Pharmingen) were used as the second antibody. Assays were developed with TMB Microwell Peroxidase Substrate (Kirkegaard and Perry Laboratories Inc., Maryland) and read after the addition of stop solution at 450 nm using a model 2550 Microplate Reader (Bio-Rad Laboratories, California) .

Derivation of T-Cell Lines and Clones

T-cell lines were generated from LNCs from mice immunized with W144 (WL1) or Q144 (QL1) . LNCs were prepared and cultured in syngeneic serum with the appropriate antigen (20 μg/ml) for 5 days. T-cell blasts were purified over a Ficoll-Hypaque™ gradient and fed with culture medium containing 0.6% T-cell growth factor (T-Stim ™, Collaborative Biomedical Research) and 0.06% recombinant IL-2. Cells were fed every 2-3 days and restimulated every 10-18 days by the immunizing antigen (20 μg/ml) plus irradiated syngeneic spleen cells (5 x 10 ⁶ cells/ml) as a source of APCs. Clones were obtained by culturing cells from QL1 at limiting dilution (Kuchroo et al. 1994) . Cells were fed with culture medium plus T-cell growth factors every 2-3 days and restimulated with a mixture of antigen (20 μg/ml) plus irradiated syngeneic spleen cells (5 x 10* cells/ml) after 10 days. Wells that contained growing cells were identified 4 days later and the cells transferred to 48-well plates (Sumilon™, Sumitomo Bakelite Company, Tokyo, Japan) , fed with medium containing T-cell growth factors every 2-3 days, and expanded by activation with antigen and APCs every 2-4 weeks.

Adoptive Transfer of T-Cell Lines

Mice were immunized using the same protocol as the in vitro proliferation assays. Lymph nodes were removed on day 10 and LNCs were resuspended at a concentration of 6-10 x 10 ⁶ cells per ml in culture medium containing 0.5% syngeneic serum in place of fetal bovine serum. Cells were cultured in the presence of various antigens 920 μg/ml) for 4 days, then harvested and purified over a Ficoll-Hypaque™ gradient. Cells were resuspended in phosphate-buffered saline at 25 x 10 ⁶ cells per ml and injected intravenously into recipient animals (0.2 ml, 5 x 10 ⁶ cells per mouse) , then recipient mice were immunized with the native peptide PLP 139-151 and CFA as described to induce active EAE.

Results

The APL 0144 Protects Mice from EAE

To determine whether altered peptide itself can induce EAE, animals were immunized with Q144 in CFA. No clinical disease was detected up to 42 days following immunization with this peptide, although all animals concurrently immunized with the native peptide, as expected, developed severe EAE. We next investigated whether coimmunization of Q144 peptide with the native PLP 139-151 would effect the induction of EAE in order to determine whether prophylactic treatment with altered peptide is essential to obtaining a suppresive effect. Results from a representative experiment (Figure 7) show that coimmunization of altered peptide with the native peptide reduced the incidence and mean severity of disease. The effect of the altered peptide on EAE was then tested in a larger series of experiments in which Q144 was either given before (preimmunization) or together with

(coimmunization) the encephalitogenic PLP 139-151 peptide. The results show that most of the animals (about 70%) preimmunized with Q144 were protected from disease and in those animals that became sick there was a significant reduction in the clinical disease (Table 2) . The mean day of onset of disease was later in the group preimmunized with Q144 than in the control group

(no preimmunization) or animals immunized with L144 or T144. Coimmunization of Q144, with the native peptide (W144) also protected animals from disease (5 of 10) , considerably decreased disease severity (although it did not delay disease onset) (Fig. 7) , and in the mice that did have clinical signs reduced the amount of CNS inflammation. This shows that preimmunization with the altered peptide ligand is not necessary to obtain a suppressive effect.

The Peptide Q144 is Not an MHC Blocker but Induces T-cells that are Cross-Reactive with the Native Peptide.

We investigated whether Q144 binds to MHC class II molecules more efficiently than autoantigenic peptide, inhibiting its binding and preventing the development of

autoimmune disease. To do so, we tested the LNCs from SJL mice immunized with Q144, the native PLP 139-151 peptide, or a mixture of both for in vi tro proliferative responses against various antigens (Figure 8) . The LNCs from animals immunized with the native peptide proliferated in vitro to the immunogen but not to a control PLP peptide (PLP 190-209) , or to other peptides with a single substitution at position 144 (Q144 or A144) , or a peptide with a double substitution at positions 144 and 147 (L144/R147) . In contrast, LNCs from animals immunized with Q144 showed a much more degenerate response in that they proliferated to the same extent when stimulated in vitro with the native peptide, W144 or Q144, and responded even more vigorously to a mixture of active peptide and Q144. When animals were immunized with the native peptide and Q144 peptides together, the LNCs responded well both to W144 and to a mixture of native peptide and Q144, but the response to Q144 alone was reduced. These data show that the Q144 peptide does not inhibit the generation of T-cells specific for the native PLP peptide. Coimmunization of the native peptide with Q144, however, lowers the T-cell responses specific for Q144. This may be due to competition for I-A binding during the initial T-cell induction, since the native W144 peptide binds to I-A more efficiently than does Q144 (Kuchroo et al., 1994) . In any event, the effect observed with Q144 is not due to MHC binding. This was further confirmed in the following experiment.

Immunization with Q144 Alters the Pattern of Cytokine Production and Therefore Causes an Altered Cell Differentiation Pattern

To compare the type of T-cell response generated in mice immunized with Q144 with that induced by the native peptide, the cytokines elaborated by LNCs taken from these animals 10 days after immunization were measured (Table 3) . LNCs from animals immunized with native W144 peptide proliferated and produced significant amounts of IL-2 and IFNγ in vi tro when stimulated with the native peptide but not when stimulated with Q144 or A144. This Thl pattern of cytokine production is consistent with the pattern of response seen in

T-cell clones derived from mice immunized with W144 (Kuchroo et al., 1993), and with the known effects of CFA upon the outcome of T-cell differentiation (Janeway et al., 1988). In contrast, LNCs from animals immunized with Q144 proliferated quite vigorously when activated with either Q144, native W144, or A144. When these LNCs were activated by Q144, significant levels of IL-2, IFNγ, and IL-10 were detected. IL-4 was measured by enzyme-linked immunosorbent assay (ELISA) or using CT4S cells and was less than 2x background in all assays. The data suggest that immunization with Q144 alters the phenotype of the T-cells induced, leading to cells producing IL-10 in addition to cells producing IL-2 and IFNγ.

To determine the type of cells involved in the antigen-specific proliferation and production of various cytokines following in vivo immunization with the altered peptide Q144, monoclonal antibodies (MAbs) specific for anti- CD4, anti-CD8, and anti-TCR V38.1/V08.2 (as an isotope-matched control antibody) were added to the LNC at the time of in vi tro activation. Data with the control anti-Vj88.1/VJ8.2 MAb and anti-CD4 and antibody are shown (Table 3) , the results obtained with the anti-CD8 antibody were similar to the control MAb (data not shown) . In the LNCs taken from mice immunized with the native PLP peptide, there was a dramatic inhibition (88%) of proliferation and IFNγ production in the presence of anti- CD4 MAb. In the LNCs from mice immunized with the altered peptide Q144 and activated with Q144 peptide, addition of anti- CD4 antibody inhibited proliferation by 65% and IFNγ production by over 70%. IL-2 and IL-10 were completely inhibited by anti- CD4 antibody. These data suggest that the immune deviation (with increase in IL-10 production) mediated by the altered peptide Q144 is dependent upon the antigen-specific activation of CD4+ helper/inducer T-cells.

T-Cells Specific for Q144 Cross-React with the Native Peptide and Produce ThO or Th2 Cvtokines

To analyze the cells induced by Q144 further, T-cell lines were generated from SJL mice immunized with W144 or Q144-

A T-cell line derived from animals immunized with native peptide W144 (WL1) has properties that recapitulate those found in primary LNC cultures, with a dominant proliferative response to W144 and production of the Thl cytokine IFNγ (Figure 9) . The T-cell line WL1 also produced IFNγ upon activation with the altered peptides Q144 and A144 and, in some experiments, low levels of IL-10 were detected when the line was stimulated with Q144. The T-cell line derived from animals immunized with Q144 (QL1) demonstrated greater cross-reactivity in that it showed a significant proliferative response to the Q144, native, and

A144 peptides and lower responses to the L144/R147 peptide

(Figure 9) , but showed no response to an unrelated PLP peptide

(data not shown) . The line produced high levels of both IFNγ and IL-10 in response to these antigens. To analyze whether IL-10 and IFNγ were produced by the same cell, the Q144-specific T-cell line (QL1) was cloned by limiting dilution and screened for cytokine production

(Figure 10) . Of the 48 clones that were established, 26 produced significant levels of cytokine following activation with Q144. IL-10 was the commonest cytokine detected, produced by 22 of 27 (81%) clones. Nine clones produced only IL-10, five produced IL-4 and IL-10, four produced IL-10 and IFNγ, and four clones produced a mixture of three or four cytokines (IL- 10, IL-4, IL-2 and IFNγ) . Of the remaining five clones, three produced only IFNγ, one produced IL-2, and one produced IL-4. These results clearly show that immunization with the APL has resulted in the development of predominantly Th2/Th0 cells producing high levels of IL-4 and IL-10, and in the generation of a population of cells simultaneously secreting IL-10 and IFNγ. A panel of 16 Q144-specific T-cell clones was selected at random and tested for cross-reactivity to the native W144 peptide. As assessed by T-cell proliferation assay 44% (7 of 16) of these Q144-specific T-cell clones cross-reacted with W144 (data not shown) . It is postulated that ThO cells have a particularly desirable cytokine profile because they produce both IL-10 (which is suppressive of Thl responses) and IFNγ (which is suppressive of Th2 responses) .

T-Cell Lines Induced by Q144 Protect Mice From EAE

To determine whether T-cells induced by immunization with Q144 were able to transfer protection, naive SJL mice were immunized with either W144, Q144, or a nonencaphalitogenic peptide PLP 190-209. Short-term T-cell lines were generated from LNCs prepared from mice 10 days after immunization and activated in vitro with immunizing or control peptides. LNCs from mice immunized with W144 and Q144 were also activated with Q144 and W144, respectively, to generate cross-reactive T-cell lines. After 4 days in culture, T-cell blasts were harvested and transferred into naive mice, which were then actively immunized with W144/CFA to induce EAE. The course of the disease was followed for 26 days (Figure 11) . As a positive control, mice preimmunized with Q144 were further immunized with PLP 139-151 and followed for disease progression in the same experiment.

Transfer of T-cell lines generated from the LNCs of mice immunized with native W144 and activated with W144 (data not shown) or Q144 into mice that were then immunized with W144 accelerated the onset of clinical disease by 1-2 days and enhanced the maximum disease severity when compared with the control mice that were injected with a T-cell line specific for an unrelated PLP peptide 190-209. In contrast, the T-cell lines generated from mice immunized with Q144 and activated in vitro with Q144 conferred some protection from clinical disease but the greatest protection was seen after transfer of the T- cell line generated from mice immunized with Q144 and activated in vi tro with W144 (Fig. 11) . This demonstrates that immunization with Q144 induces cells that protect mice from EAE, but that these T-cells are cross-reactive and should be activated by the native peptide to best mediate protection.

We also examined the effect of anti-IL-4 antibody on protection mediated by Q144 in vivo and found that this treatment abrogated protection (data not shown) . This provided further evidence that Q144 mediates protection by inducing regulatory TH2 cells in vivo.

TABLE 1

Altered Peptide Ligand Stimulation of MBP Reactive T-Cell Clone 0blAl2 and Binding to Class II MHC

ac m

Pi

TABLE 2

^• p<0.001 ^b p<0.05 by Fisher's exact probability test (compared with controls) ^c p<0.001 by 1 test (compared with controls)

TABLE 3

REFERENCES

Alexander, J. et al., J. Immunol. 150:1-7. 1993.

Ashton-Rickardt, P. G. et al., Cell 76:651-663. 1994.

Berridge, M. J. Nature 3_6_1:315, 325-325, 1993.

Bertoletti A. et al., Nature 369:407-410. 1994.

Chen, Y. et al., Science 265:1237-1240. 1994.

De Magistris, M. T. et al., Cell 6_&.625-634, 1992.

Evavold, B. D. et al., Science 2≤2.:1308-1310, 1991.

Evavold, B. D. et al., Immunol. Today 14:602-609. 1993a.

Evavold, B. D. et al., J. Immunol. 150:3131-3140. 1993b.

Hogquist, K A. et al., Cell 76_:17-27, 1994.

Hδllsberg, P. et al., J.Immunol. 153:566. 1994.

Jameson, S. C. et al., J Exp. Med. 177:1541-1550. 1993.

Janeway, C.A. et al., Immunol. Rev. 101:39-80. 1988.

Kehrl, J. H. et al., J. Immunol. 137:3855-3860. 1986a.

Kehrl, J. H. et al., J Exp. Med. 163:1037-1050. 1986b.

Khoury, S. J. et al., J. Exp. Med. 176:1355-1364. 1992.

Klenerman P. et al., Nature 369:403-407, 1994.

Kuchroo, V.K. et al., J. Immunol. 151:4371-4382. 1993.

Kuchroo, V. K. et al., J. Immunol. 153:3326-3336. 1994.

Kulkami, A. B. et al., Proc. Natl. Acad. Sci.. USA 90:770-774.

1993.

Miller, A. et al., J. Immunol. 151:7307-7315. 1993.

Miller, A. et al., Proc. Natl. Acad. Sci. USA 89:421. 1992.

Mossmann, T. R. et al., J. Immunol. 136:2348-57. 1986.

Ostrov, D. et al., J. Immunol. 150:4277-4283. 1993.

Ota, K. et al., Nature 346:183. 1990.

Racke, M. K. et al., J. Immunol. 146:3012-3017. 1991.

Rojo, J. M.. et al. , Cold Spring Harbor Symp. Quant. Biol.2:657-

666, 1989.

Romagnani, S., Annu. Rev. Immunol. 12:227-57. 1994.

Rudd, C. E. et al., Immunology Today 15:225-234. 1994.

Rudd, C. E. et al., Proc. Natl. Acad. Sci.. USA 85:5190. 1988.

Ruppert, J. et al., Proc. Natl. Acad. Sci.. USA 9_Q.:2671-2675,

1993.

Seder, R. A. et al., Annu. Rev. Immunol. 12:635-673. 1994.

Shull, M. M. et al., Nature 359:693-699. 1992.

Sloan-Lancaster, J. et al., Nature 363:156-159. 1993.

Sloan-Lancaster, J. et al., J. Exp. Med. 180:1195-1205. 1994.

Smith, L. S. et al., Biochem. and Biophys. Res. Commun. 71:459-

465, 1976.

Wahl, S. M., Immunologic Res. 10:249-54. 1991.

Wold, F., Ann. Rev. Biochem. 50:783-814. 1981.

Wucherpfennig, K. W. , Science 248:1016. 1990.

Wucherpfennig, K. W. et al., J. Exp. Med. 179:1-12. 1994.