Evaluating candidate autoantigens in rheumatoid arthritis
Andrew P. Cope 1, Grete S0nderstrup 2
l The Kennedy Institute of Rheumatology, Hammersmith, London, UK 2 Departments of Microbiology and Immunology, and Medicine, Stanford University School of Medicine, Stanford, California, USA
Introduction
The study of autoimmune disease in animal models has provided strong support for the concept that tissue-specific self antigens drive the chronic inflammatory response in vivo. For example, the non-obese diabetic (NOD) mouse, a spontaneous murine dia- betes model sharing many features with type I insulin-dependent diabetes mellitus (IDDM) in humans [1], and another spontaneous model of diabetes occurring in mice transgenic for a [3-islet cell antigen-specific T cell receptor (TCR) [2], have established a central role for major histocompatibility complex (MHC) class II and T cell immu- nity in the pathogenesis of murine type I diabetes. These animal models have also demonstrated the feasibility of manipulating the autoimmune response and disease in- cidence by injecting mice with immunodominant peptide epitopes, or native autoanti- gen to induce immunological tolerance [3]. For diseases such as rheumatoid arthritis (RA), defining candidate synovial joint antigens as targets of the autoimmune response has become a priority, and an essential step for exploring the prospects of antigen-tar- geted immunotherapy in patients. However, progress has been rather slower for RA, in part because spontaneous arthritis in unmanipulated rodents is relatively rare, and when observed, it differs in many respects from the human disease [4]. In an attempt to develop new animal models to explore the precise nature of immune responses to cartilage antigens in RA, several lines of HLA-DR4 transgenic mice have been gener- ated [5]. These animals have been used to characterise in detail immune responses to human joint-derived proteins restricted to HLA-DR4 molecules associated with suscepti- bility to RA (HLA-DRB 1"0401), or non-associated/resistant to RA (DRB 1"0402).
A molecular basis for susceptibility to RA
RA is one of the most common autoimmune diseases in humans, with a prevalence of approximately 1% in Caucasians in Europe and Northern America [6]. During the
Correspondence to: G. SCnderstrnp
24 A. R Cope, G. SCnderstmp
Table 1. Polymorphic DR~ chain residues that distinguish RA-associated from non-associated HLA-DR4 alleles (RA, Rheumatoid arthritis)
DRB 1 allele CRy3 chain residue Susceptibility to RA
57 67 70 71 74 86
DRB 1 '0401 D L Q K A G Strongly associated DRB 1 *0402 D I D E_ A V Non-associated DRB 1 *0403 D L Q R E_ V Non-associated DRB 1 *0404 D L Q R A V Associated DRB 1 *0405 S L Q R A G Associated
early stages of disease, RA is characterised by localised inflammation of the synovial joints. Cartilage damage may also occur, but is difficult to detect and quantify early in the disease. In later stages, the disease may progress with severe joint destruction, but also extra-articular manifestations such as rheumatoid nodules, vasculitis, and pul- monary complications. Although RA has previously not been considered a disease with significant mortality, the life expectancy of RA patients is about a decade shorter than healthy individuals from the same population [7]. In addition, RA patients have significant morbidity due to progressive loss of physical capabilities, anaemia, fatigue, depression, severe chronic pain, and increased susceptibility to infection, in part due to the disease process, but often as a result of treatment with corticosteroids and other im- munosuppressive drugs [6].
RA has a significant genetic component. However, a concordance rate for RA in monozygotic twin pairs of up to 15% indicates a major influence of environmental fac- tors [8]. Although several genetic loci contribute to disease susceptibility, the most im- portant genetic factor predisposing to RA is associated with the MHC class II locus ([9] and reviewed in [10]). While the association between HLA-DR and RA was first described by Stastny in the 1970s [11], a significant advance in our understanding of these associations was reported more than a decade later, when it was shown that sus- ceptibility to RA across different ethnic populations correlated closely with the ex- pression of a specific consensus amino acid sequence (referred to as the "shared epi- tope") within the HLA-DRI3 chain [12]. This sequence was subsequently shown by several groups to be encoded by several HLA-DRB1 alleles, including HLA-DR4 ( '0401, *0404, *0405, and *0408), but also DR1 (*0101), DR6 ( '1402), and DR10 al- leles (*1001) [12-14].
The precise molecular mechanism for these disease associations is perhaps best il- lustrated by comparing the susceptible HLA-DR4 subtypes DRB 1"0401, *0404, and *0405, which share the consensus shared epitope sequence, with the non-suscepti- ble/resistant DRB 1"0402 allele. Susceptibility or resistance maps to codons 70 and 71 of DRB 1 alleles, as shown in Table 1. The non-associated allele (DRB 1"0402) en- codes for two negatively charged amino acids (D and E) substituting for a polar and a positively charged amino acid (Q and K or R) in susceptible DR4 alleles, a net charge difference of three. The non-susceptible DRB 1"0403 allele is identical to the suscepti- ble DRB 1"0404 allele except at position DR1374, where the negatively charged aspar- tic acid (E) is substituted for an uncharged alanine residue (A). Amino acids 1370, 71, and 74 comprise part of the c~ helix of the DR13 chain that accommodates the peptide side chains of amino acid residue 4 (P4) of 9mer peptides bound in the groove of
Evaluating candidate autoantigens in rheumatoid arthritis 25
HLA-DR molecules [15, 16]. P4 is one of four anchor positions known to influence binding to HLA-DR molecules [17, 18].
A functional basis for the associations between HLA-DR and RA
MHC class II molecules function by selecting and presenting immunogenic peptide fragments of protein antigens to CD4 + T cells in the periphery. They also play an im- portant role in the selection of the TCR repertoire in the thymus. Since CD4 + T lym- phocytes recognise linear stretches of about 9-20 amino acids derived from self or for- eign protein antigens bound in the peptide binding groove of polymorphic MHC class II molecules, it has been suggested that differences in the way that HLA-DR or DQ molecules present selected peptides to T cells could be an important mechanism for susceptibility to autoimmune diseases such as RA and IDDM [19, 20].
On the basis of these findings, two principal models have been proposed to account for the association between RA and the consensus DR~ chain sequence. The models are based on the assumption that the shared epitope is the critical genetic element linked directly to disease. The first model proposes that the shared epitope determines specific peptide binding, and that "pathogenic" peptides bind only to disease-associ- ated HLA class II molecules [19, 20]. According to this model, a gradient of affinities of disease inducing peptide for MHC class II molecules might account for the differ- ences in susceptibility and/or severity conferred by different HLA-DR4 molecules. The second model proposes that the shared epitope influences TCR recognition by binding and selecting autoreactive T cells during thymic development, and expanding these populations in the peripheral T cell compartment [21, 22].
Crystallographic studies of both HLA-DRB 1"0101 and "0401 molecules indicate that these models are not mutually exclusive [16, 23]. The data reveal that residues 1370, 71, and 74 could exert a significant influence on peptide selection, because their side chains form part of the wall of peptide binding pocket 4 of the HLA-DR molecule. Furthermore, the orientation of TCR/peptide-MHC complexes in crystals and based on molecular modeling studies would predict that these same residues are important TCR contacts [24-26], and would thus be capable of influencing the shaping of the TCR repertoire directly during thymic selection, as well as activation and expansion of these T cells in the periphery. Over the last 10 years, much research effort has focussed on trying to understand the relative contribution of the shared epitope to peptide binding and TCR recognition.
Clinical correlates
The clinical implications of these findings have become more apparent, and are based on several observations. The frequency of the shared epitope in a given population cor- relates closely with the prevalence of RA in that population [10]. Secondly, genetic heterogeneity in patients with RA often correlates closely with clinical heterogeneity. For example, in population studies, different HLA-DRB 1 alleles appear to influence the severity of disease, with DRB 1"040 t being found in patients with severe, seropos- itive, erosive disease (often with extra-articular manifestations), while DRB 1"0101 is observed at higher frequency in patients with less severe, seronegative, non-erosive disease [27]. Finally, inheriting two copies of alleles expressing the consensus se- quence greatly increases disease penetrance, time of onset, and severity [28].
26 A. R Cope, G. SOnderstrup
Associations between MHC class II genes and autoimmunity have also been de- scribed in rodent models of experimental autoimmune arthritis such as collagen-in- duced arthritis (CIA). Thus, in CIA, disease susceptibility has been linked to the murine MHC class II genes I-Aq,r [29], the homologue of HLA-DQ in humans. Recent reports of CIA being induced in HLA-DQw8 transgenic mice have proposed that DQ, and not the DR locus was the dominant susceptibility gene mapping to the HLA class II region [30]. However, population genetic studies would not appear to support this hypothesis [10]. Furthermore, experiments from our own laboratory, and those of Kang [31, 32], have shown that HLA-DRB 1"0401 transgenic mic e develop CIA under circumstances in which the HLA-DR4 molecule is the dominant MHC class II restric- tion element. Indeed, HLA-DR1 transgenic mice also develop CIA [33]. These studies provide the first clear evidence that immune responses to cartilage antigens restricted to disease-associated HLA-DR molecules may be involved in the induction of arthri- tis, as well as the perpetuation of the disease, as previously implied from clinical stud- ies [271.
Are immune responses to synovial joint antigens important in RA?
Perhaps the strongest evidence in favour of a major involvement of T cells in arthritis comes from the genetic link with MHC class II, described above, and the finding that CD4 + T cells alone can induce inflammatory arthritis in adoptive transfer experiments in rodents [34]. The observation that IgG antibodies specific for collagen II are in- volved in the pathogenesis of CIA further confirms that B cells as well as T cells are involved [35]. Furthermore, lymphocytic infiltrates in inflamed synovium resemble lymphoid follicles, providing further evidence for chronic immune activation in vivo. What is the evidence for chronic immune responses to cartilage antigens in RA?
Autoantibodies in RA
For many autoimmune diseases, the detection of autoantibodies can have both diag- nostic and prognostic value. For example, in human type I IDDM, close to 90% of re- cent onset IDDM patients have autoantibodies against one or more of the [~-islet cell proteins, preproinsulin, the [3-islet cell tyrosine phosphatase, IA-2, or glutamic acid de- carboxylase, GAD65 [36]. Furthermore, the presence of autoantibodies against any two of these three major autoantigens in pre-diabetic subjects (e.g. siblings of diabet- ics with disease-susceptible HLA genotypes) carries an 80-90% risk of progression to overt diabetes within 5 years [36]. The connective tissue diseases provide further ex- amples of disease states in which the observed autoantibody specificities can have both diagnostic and prognostic implications [37]. Disease activity may also be reflected by the antibody titre in some instances.
In RA, the significance of autoantibodies is less well defined. Thus, while there is some diagnostic and prognostic value in testing for the presence of serum rheumatoid factors (RF), and anti-keratin antibodies (anti-perinuclear factor, fillagrin), the signifi- cance of antibodies to collagen, chrondrocyte antigens, heat shock proteins, HLA anti- gens, endothelial cell and histone antigens is not clear. Somewhat surprisingly, the finding of serum antibodies to cartilage proteoglycans such as aggrecan in RA patients is rare [38]. On the other hand, the tissue specificity of autoantibody responses in RA
Evaluating candidate autoantigens in rheumatoid arthritis 27
is suggested by the finding of antibodies recognising unique epitopes of collagen II that are not expressed on collagens I, III, IV or V [39]. Nevertheless, antibodies to joint antigens such as collagen II are restricted to a subset of patients with RA, and in these patients there is no clear correlation between disease duration, activity or severity [40].
The consistent inability to detect in the majority of patients significant levels of serum antibodies to a dominant candidate antigen, let alone a panel of cartilage anti- gens, is puzzling. One possibility is that these antibodies are rapidly cleared from the circulation, and are deposited in synovial joints. De novo RF production in synovial joints and the finding of antigen/antibody complexes for collagen II in murine arthritic joints and for aggrecan in RA synovial fluid but not serum are consistent with this idea [38, 41, 42]. Another possible explanation for the lack of detectable autoantibody re- sponses relates to the molecular nature of antigens used to screen for responses. For some years it has been recognised that post-translational modifications contribute to the tertiary conformation of antigenic epitopes. One example is the N-linked glycosy- lation of asparagine residues in the C72 and in the V domain of IgG [43]; there exists an inverse relationship between IgG rheumatoid factor (RF) binding and the level of galactosylated IgG Fc. More recently, it has been demonstrated that antibody re- sponses to fillagrin, the target of anti-keratin immunofluorescence responses, are de- pendent on epitopes carrying de-iminated arginine residues, in the form of citrulline [44]. Pools of peptides containing selected epitopes with citrulline residues have been shown by enzyme-linked immunoassay to define a subset of serum autoantibodies in more than 90% RA patients [44]. These data indicate that the nature of antigenic epi- topes expressed in the inflamed joints is complex. Until recently, assays designed to detect specific immune responses to self antigens have not routinely taken these subtle modifications into account.
Autoantigen-specific T cell responses in RA
T cell proliferative responses have been described for a wide range of antigens in pa- tients with RA. Some of those reported in the literature are listed in Table 2. A minor- ity have also been identified as disease-inducing autoantigens in rodent models. In many cases, both the intensity of the response and the frequency of patients respond- ing are highly variable, with perhaps up to 30% of patients responding to a defined antigen. The most trivial explanation for these data could relate simply to the hetero- geneity of the disease, or to the choice of antigens studied. On the other hand, T cell activation can also be influenced by post-translational modifications of epitopes dur- ing processing, in a manner analogous to recognition of native antigenic epitopes by B cells mentioned above. This has been clearly demonstrated for type II collagen T cell epitopes in CIA in mice [56], and has important implications for laboratory evaluations of immune responses to cartilage antigens in patients with arthritis. The molecular na- ture of connective tissue antigenic epitopes presented in the synovial joint may be in- fluenced by other factors, such as the "inflammatory milieu" of the rheumatoid joint, the type of antigen-presenting cell (APC), as well as the proteolytic enzymes in the ex- tracellular compartment which could modify both the size and the fine specificity of peptide fragments available in the joint for presentation to T cells. In some cases, the threshold for activation of cartilage antigen-specific T cells may not be reached in vivo, despite up-regulation of adhesion and accessory molecules. This may be due in part to defective TCR signaling, which is a recently recognised characteristic of chron-
28 A. R Cope, G. SCnderstmp
Table 2. Candidate autoantigens as targets of the immune response in RA (Ag, Antigen)
Immune responses Disease induced Ref. identified in RA Patients in rodent models
Joint Ag
Collagen II + + [45]
Aggrecan + + [46-48]
Chondrocyte Ag + ? [49]
HCgp-39 + + [50]
205-kDa antigen + ? [51 ]
Human hsp60 + + [52]
Microbial Ag
hsp60 + + [52]
EBV transactivators + - [53]
E. coli dnaJ + + [54]
Other Ag
Immunoglobutin + + [55]
Fillagrin + - [44]
ically activated T cells [57, 58]. If the precursor frequency of antigen-specific T cells is also very low, it may not be possible to register a positive response in conventional T cell transformation assays in vitro, even where a specific response may be present and could have pathogenic significance. Accordingly, we have to accept that at present our understanding of the nature of arthritogenic T cell epitopes presented in synovial joints with active inflammation is incomplete.
What are the characteristics of immunogenic T cell epitopes?
An increased understanding of immunogenic peptide epitopes in recent years is de- rived from studies of direct binding of peptides to detergent-solubilised MHC mole- cules [59], and from the sequencing of peptides eluted from peptide/MHC complexes purified from cell lysates [60]. This information, combined with computer modeling, has made it possible to define specific motifs for the amino acid sequences of peptides binding selectively to a range of MHC class I and class II molecule [61]. More recent studies, including those from our own laboratory, indicate that even though a peptide sequence may contain a specific motif, and bind with high affinity to a given MHC molecule, that peptide/MHC complex may not necessarily elicit a T lymphocyte re- sponse in vivo [62]. Many other factors, such as availability of peptide, local tissue de- struction, antigen processing, interaction with invariant chain and HLA-DM, MHC binding affinities, survival time of peptide/MHC complexes on the surface of the APC, and the nature of the APC (dendritic cells, macrophages, or B cells), will determine the immunogenicity of any given epitope (reviewed in [63-65]). While these rules are well established for foreign antigens, they clearly apply for self (auto) antigens as well.
Evaluating candidate autoantigens in rheumatoid arthritis 29
How can we identify potential autoantigens of importance in RA?
Since the target organ in RA is thought to be the synovial joint, the conventional ap- proach has been to focus on cartilage- and matrix-specific antigens, especially those expressed in abundance. Perhaps the best studied of these are collagen type II, and car- tilage proteoglycans such as aggrecan. These proteins are relatively tissue specific. Ex- pression of other antigens, such as heat shock proteins, are not confined to the joint. For example, mycobacterial heat shock protein (hsp)60, which is cross-reactive with autologous human hsp65, has been implicated in driving pathogenic T cell responses in juvenile RA [66], reactive arthritis [67], as well as in rodent models of experimen- tal adjuvant arthritis [52], and type 1 diabetes in NOD mice [68].
Since synovial joint T cells from RA patients are difficult to expand in vitro, sev- eral laboratories have explored alternative methods for studying the specificity of T cells from inflamed joints. For example, Kotzin et al. [69] have studied the nature of antigen specificity by cloning TCRo~ and ~ chains from synovial joint T cells of RA patients known to have oligoclonal expansion of T cells with the same V[3 elements in different joints. The TCRcz and ~ chain constructs were then co-transfected into a murine TCR~x[3-/- thymic tumor cell, a fusion partner commonly used for the genera- tion of murine T cell hybridomas. In this way, stable transfectants could be easily prop- agated and expanded. T cells were then used as tools to probe for antigenic specificity, by stimulating them with pools of fibroblasts transfected with HLA-DR4 engineered to co-express random peptides encoded by a cDNA peptide library. Pools of APC that stimulate T cells could be subcloned, and the peptide cDNA insert sequenced to iden- tify the origin of that peptide (the autoantigenic protein) from protein databases of known human or microbial origin. The prediction would be that this approach would yield a variety of different antigenic specificities, which, while similar in profile in dif- ferent joints from the same patient, could be different between individual patients. An analogous approach has been attempted using serum from RA patients as a tool to probe expression libraries for novel antigenic specificities [70].
An alternative approach has been to target protein antigens which are not expressed in healthy joint tissue, but which are inducible, and expressed in inflammatory joints following stimulation by growth factors or pro-inflammatory cytokines such as IL-1 and TNE Such proteins would be especially attractive as autoantigens, being se- questered from the peripheral immune system, as well as the maturing T cell repertoire in the thymus during early T cell development, thereby escaping central and peripheral tolerance mechanisms. For similar reasons, antigens expressed only in adult cartilage may also prove to be prime targets for the immune system in autoimmune joint dis- ease. A summary of some of the relevant characteristics of autoantigens likely to pro- mote anti-self responses are listed in Table 3.
A novel cartilage antigen with many of these characteristics is human cartilage (HC) gp-39, also known as YKL-40 [71, 72]. HCgp-39 is a member of a family of chitinases, but lacks enzymatic activity. An HCgp-39 homologue has previously been described in mice as Brp39, cloned from murine macrophages derived from infiltrates surrounding inflammatory murine mammary tumours [73]. HCgp-39 turns out to be a good model protein for analysis of the immune response in several respects. It is not expressed in normal cartilage, but its expression is induced by inflammatory cytokines, and expression of HCgp-39 can be down-regulated by TGF-~ [71]. It is relatively small (by connective tissue protein standards) being 383 amino acids, and carries sin- gle potential disulphide bonds and glycosylation sites. A computer search for joint re-
30 A. R Cope, G. SOnderstrup
Table 3. Characteristic features of putative autoantigens in RA
• Not expressed in the thymus • Expression confined to synovial joints • Only expressed in inflamed joints • Expression induced with proinflammatory cytokines • Presented as neo-self epitopes as a result of post-translational modifications • Native antigen processed in extracellular synovial compartment • Primary amino acid sequence contains HLA-DR binding motifs • Epitopes presented as multiple stable MHC/peptide complexes • Complexes recognised by significant proportion of TCR repertoire • Induces experimental marine arthritis following immanisation • T and B cell responses in large proportion of RA patients, but not (or to a lesser degree) in class
II-matched control subjects
lated proteins with prototypic DRBI*0401 binding motifs identified HCgp-39 as a protein having a significant number of potential epitopes [50]*. Four different syn- thetic peptides synthesised from the sequences were found to bind DR*0401 with high affinity in competi t ion binding assays [50]. Three of these four peptides were subse- quently found to stimulate T cells from the peripheral blood of RA patients. In all, about 40% of the first group of RA patients tested demonstrated spontaneous T cell proliferative responses to one or more of the three immunodominant T cell epitopes.
How can we explore immune responses to human joint-related proteins in vivo?
The transgenic mouse model
Several years ago we embarked on a project to produce HLA-DR4 transgenic mice carrying the RA susceptible HLA-DRB 1"0401 allele, and another line carrying the RA non-susceptible HLA-DRB 1"0402 allele. In effect, parts of the mouse immune system were replaced with their human counterparts, for study in vivo [5]. This enabled us to examine the immune response to human synovial joint-related autoantigens in the con- text of RA susceptible or RA non-susceptible HLA-DR4 alleles, encoding human mol- ecules relevant to the pathogenesis of RA.
The studies planned for these transgenic mice would have been difficult to under- take in human subjects for many reasons. In the first instance, it would not be possible to immunise humans with potentially arthritogenic cartilage antigens. Secondly, there are technical constraints associated with the generation of a large panel of human T cell clones in vitro whose specificities are representative of the immune response (i.e. for the full range of immunogenic epitopes), especially when the precursor frequency of antigen-specific T cells in peripheral blood or synovial joint is very low. Finally, our studies focussed on HLA-DR-rest r ic ted responses in mice, while in patients responses restricted to other functional MHC class II molecules must be taken into account. Fur- thermore, a direct comparison of the function of different HLA-DR4 subtypes in ha-
*A software program for searching protein data bases for amino acid sequences with a desired mo- tif is available on the world wide web called pMotif (http://alces.med.umn.edu/motif.thml)
Evaluating candidate autoantigens in rheumatoid arthritis 31
Table 4. Generating the HLA-DR4-restricted peripheral CD4 + TCR repertoire. The number of peripheral CD4 T cells was determined by flow cytometry as described in reference [78]
man subjects in vivo would be restricted by the necessity to evaluate patients whose class II haplotypes differ only at the DRB1 locus. We now describe some of the unique characteristics of this transgenic mouse model, focussing on technical aspects which have allowed us to circumvent some of these problems directly.
CD4 + TCR selection in HLA-DR4 transgenic mice
HLA-DR4 transgenic mice have been genetically engineered using DRc~- and DR~3- specific cDNA introduced into constructs previously shown by the Mathis and Benoist laboratory to drive the expression of foreign cDNA in MHC class II-positive cells in transgenic mice [74]. Both DR~ and DR~ transgene expression are driven by the murine I-Ea promoter. These mice are healthy, and expression of the transgenes has been confirmed by both in situ immunohistochemistry and ex vivo flow cytrometry. The transgenes are functional, as demonstrated by the induction of HLA-DR4-re- stricted T cell responses, and by the influence of DR4 expression on the V~ repertoire of CD4 + T cells [5]. The first generation of HLA-DR4 transgenic mice (DRB 1"0401) were used to determine the immunodominant T cell epitopes of bovine collagen type II [75]. However, the vast majority of antigen-specific CD4 + T cells were restricted to murine MHC class II molecules on immunisation with foreign or self antigens. Subse- quent introduction of the murine MHC class II knock out genotype (using I-A~ -/- mice) of Mathis and Benoist [76], and the human CD4 transgene developed by Dr. D. Littman [77] led to the production of HLA-DR4 transgenic mice selecting close to nor- mal numbers of CD4 + T cells restricted almost exclusively to HLA-DR4. However, there is a "dosage effect", such that mice homozygous for each MHC class II transgene (or heterozygous for two different HLA-DR alleles in double transgenic mice) are re- quired to select in the thymus sufficient numbers of peripheral CD4 + T cells compara- ble to wild-type mice (Table 4). The expression of human CD4 molecules also con- tributes to the selection process through its capacity to bind HLA-DR molecules more efficiently than endogenous murine CD4.
T cell hybridoma production in HLA-DR4 transgenic mice
We immortalised activated T cells within 2 weeks of the primary immunisation by fu- sion with the TCRo~3 + thymoma cell line to provide a "snapshot" of the early immune responses to antigen. This allowed us to define the specificity of a large number of T
32 A. R Cope, G. Sonderstmp
Fig. 1. T cell hybridoma production in HLA- DR4/human CD4 transgenic, murine MHC class II knockout mice (see [78]) (IFA, incom- plete Freund' adjuvant)
Immunize transgenic mice with native autigenflFA
in hind foot pads
(10 days)
Remove draining popliteal and inguinal lymph nodes
Stimulate with native antigen in vitro
(3 days)
Purify cells on density gradient,then propagate in conditioned medium
(l day)
Fuse activated T cells with TCR ~-13 T cell thymoma
(7-14 days)
Expand, propagate and screen hybrids for antigen reactivity
cells, and examine the relative frequency of T cells specific for each epitope. Standard protocols were used for immunisation using intact protein antigen in incomplete Freund's adjuvant (IFA), and for subsequent T cell hybridoma production [78]. The HCgp-39 glycoprotein was chosen for the reasons outlined above. HLA-DR transgenic mice with DR4 alleles associated with either RA susceptibility (DR*0401) or RA re- sistance (DR'0402), expressing human CD4 and lacking routine MHC class II mole- cules, were immunised with HCgp-39 in IFA, and T cell hybrids generated as illus- trated in Fig. 1. This experimental approach allowed us to define large numbers of anti- gen-specific T cells recognising peptides which not only bound to HLA-DR4 mole- cules, but were also immunogenic.
Mapping T cell epitopes
Peptide specificities were defined for 250 T cell hybridomas specific for HCgp-39 and restricted to DR*0401, and more than 150 antigen-specific hybridomas restricted to the RA non-associated DR4 molecule, DR*0402. Fine specificity was subsequently defined with pools of overlapping peptides, and responding T cells were tested further with individual peptides derived from the positive pools. The minimal epitope within each specific peptide was determined using N- and C-terminal truncated synthetic pep- tides for each epitope. For DR*0401, nine immunogenic epitopes were identified, and three of these (100-115, 262-277 and 322-337) accounted for approximately 80% of all antigen-specific DR*0401-restricted T cell hybrids.
Using an identical experimental approach with HLA transgenic mice expressing RA non-associated DR*0402 molecules, T cells from HLA-DR*0402/human CD4 transgenic, I-A~3 -/- mice recognised a completely different set of immunodominant epi- topes of HCgp-39 than T cells from DR*0401 transgenic mice on a similar genetic background. From 151 peptide-specific hybrids, 86% of antigen-reactive DR*0402-re- stricted T cells responded to either of two peptides, 22-37 or 298-313. These results demonstrate for the first time that RA-associated and non-associated HLA-DR4 mole-
Evaluating candidate autoantigens in rheumatoid arthritis 33
cules present completely different sets of peptide epitopes of HCgp-39 to T cells in vivo.
Validating the results obtained in the mouse model
To verify this experimental approach for studying T cell responses in human autoim- mune disease, it was initially shown that peptide-specific T cells responses could be completely inhibited with anti-DR blocking monoclonal antibodies. Secondly, peptide- specific T cell hybrids from transgenic mice were studied for their responses to human DR4-APC pulsed with native HCgp-39. These experiments revealed that T cells re- spond vigorously to either DR4-positive EBV-transformed B lymphoblastoid lines, or to fresh human peripheral blood mononuclear cells from DR4-positive donors. These results suggest that antigen processing in mouse and human APC are quite similar. Furthermore, a significant IL-2 response of DR*0401-restricted T cells from trans- genic mice to peptides presented by closely related RA-associated DR molecules other than DR*0401, such as DR* 0404, 0405, and 0101, indicates that the shared epitope residues common to these alleles are important structural elements for peptide binding and TCR recognition. Significantly, T cell hybridomas did not recognise these epitopes presented by APC expressing non-associated DR*0402 molecules. These data demon- strate that epitope mapping in DR*0401 transgenic mice may identify T cell epitopes presented by a broader range of disease-associated HLA-DR alleles in vivo.
Do these epitopes stimulate T cells from patients with RA?
The three immunodominant T cell epitopes defined in DR*0401 transgenic mice were exactly the same as the three T cell epitopes previously shown to stimulate T cells from RA patients, by Verheijden et al. [50]. Preliminary data indicated that T cells from pa- tients with chronic active RA responded to the three immunodominant epitopes de- fined in HLA-DR*0401 transgenic mice. In a pilot study of patients from the Rheuma- tology Outpatient Clinics at Stanford University, we have also observed responses of peripheral blood mononuclear cells from RA patients not only to the three immun- odominant T cell epitopes, but also to each of the six additional DR*0401 peptide epi- topes defined in this model. In this small study of ten HLA-DR4-positive RA patients (all with active disease treated with various combinations of disease-modifying drugs), we have found that eight out of ten patients responded to at least one of these nine pep- tide epitopes. The two RA non-responders were also unresponsive to tetanus toxoid. Nineteen peptide-specific T cell responses were recorded in this group of eight re- sponders; five (26%) responses were to one of the three immunodominant T cell epi- topes. Significantly, not all these patients expressed DR*0401, but expressed other shared epitope-carrying alleles, providing further evidence that epitopes defined in DR*0401 transgenic mice may also be immunogenic in patients who carry DR*0404 or 0405, for example.
The frequency of peptide-specific responses in transgenic mice and patients is clearly different. There are several possible explanations. Firstly, the heterogeneity of the patient population studied could explain the lower frequency of responses to the three immunodominant epitopes identified in DR*0401 transgenic mice, since there may be preferences for some but not other peptides to be presented by other HLA-DR4
34 A. R Cope, G. SOnderstrup
subtypes, or by HLA-DQ/DP on the same haplotype. Secondly, it is conceivable that the epitopes defined in mice after only 10 days immunisation with native antigen in IFA may be different from those that predominate after persistent antigenic stimulation over periods of months or years in chronic inflammatory disease. Studying the profile of immunogenic peptides in transgenic mice after multiple immunisations over a pro- longed period of time will determine whether the profiles of responses in mice and pa- tients are fundamentally different, or whether those observed in patients are initially similar to mice but after chronic activation the peptide profile changes through epitope spreading. Thirdly, human HCgp-39 shows N80% sequence identity to the mouse ho- mologue [71, 73], and so the human antigen appears "foreign" to the mouse, while be- ing "self" in patients with arthritis. Regardless of these differences, the patient data demonstrate that immunogenic peptide epitopes of a given autoantigenic protein which stimulate T cells from patients with autoimmune disease can be identified using this experimental approach. Further experiments are now underway to explore the possi- bility that synovial joint T cells also recognise these epitopes, and to examine re- sponses in a large population of healthy HLA-DR-matched individuals, as well as in RA patients, to determine whether the selected epitopes also elicit responses in healthy controls, or whether selected peptide epitopes are indeed disease specific.
To address whether the expression of more than one class II molecule could signif- icantly influence the peptide responses observed in vivo, we have examined the rela- tive contribution of HLA-DR and HLA-DQ to the generation of HCgp-39-specific T cell responses. By studying T cell epitopes presented by these molecules in either sin- gle (DR4) or double (DR4 x DQw8) transgenic mice, it has become possible to define the nature of DQ-restricted T cell epitopes, and compare them with those motifs that have already been characterised in single HLA-DR4 transgenic mice. In double trans- genic HLA-DR4 x DQw8 mice we have also studied the relative frequency of re- sponses contributed by each restriction element, and examined whether each class II allele can influence the nature and/or frequency of epitopes presented by the other molecule. Preliminary data from these studies indicate that for HCgp-39-specific T cell responses, the majority o fT cells (>85%) are restricted to HLA-DR and not DQ. While we identified one common epitope presented both by HLA-DR and DQ molecules, we also identified unique epitopes presented by DQw8 molecules not identified previ- ously in mice expressing HLA-DR4 alone. Interestingly, the frequency of DR4-re- stricted responses to immunodominant epitopes did not differ significantly from that defined in mice expressing the HLA-DR4 transgene alone. These studies underscore the value of the transgenic mouse model, since such an approach would not be feasi- ble in man.
W h a t can we learn from the H L A - D R 4 transgenic mouse model?
This transgenic mouse model has provided a unique opportunity to investigate further the molecular nature of antigenic epitopes of relevance to disease pathogenesis, and how they are presented to T cells. Using mice expressing disease-associated (DR*0401) and non-associated DR4 alleles (*0402), this model has demonstrated clearly that one possible basis for susceptibility to RA may involve the presentation of a set of selected peptide epitopes derived from a cartilage antigen to T cells in vivo. Precisely which, if any of these epitopes are pathogenic awaits further investigation. However, the availability of transgenic mice should facilitate such studies.
Evaluating candidate autoantigens in rheumatoid arthritis 35
This transgenic model has also allowed us to investigate in more detail the func- tional consequences of disease-associated HLA-DR4 molecules presenting distinct sets of peptide epitopes. For example, we have recently investigated cytokine produc- tion by T cells from transgenic mice following immunisation with native antigen in IFA. These studies reveal very clear differences between the DR*0401- and DR*0402- restricted T cell response. Thus, DR*0401 responses to either intact HCgp-39 antigen, or pools of allele-specific peptides, induce abundant expression of interferon- 7 (IFN-y) at levels five-fold those observed in cultures of DR*0402-restricted T cells stimulated with the same antigen. Differences were even greater when the responses of estab- lished peptide-specific T cell lines were compared. In addition, DR*0401-restricted T cell lines produced TNF in response to "0401 peptides, while TNF expression was not detectable in DR*0402 cultures. However, the DR*0402 transgenic mice are not defi- cient in IFN-y production, since immunization with the outer surface protein A of Bor- relia burgdo,feri results in abundant 1FN-y production on restimulation with antigen in vitro.
These findings suggest that one possible mechanism by which disease-associated HLA-DRB 1 alleles may contribute to disease severity in RA is by presenting distinct sets of peptides from joint-related self proteins to T cells, and by perpetuating the chronic inflammatory process through the generation of Thl-like T cell subsets which sustain the production of pro-inflammatory cytokines in synovial joints.
Concluding remarks
The studies outlined above provide a novel experimental strategy for evaluating in vivo a potential role for selected autoantigens in human autoimmune disease. The model system has allowed us to dissect the molecular nature of responses of T cells to self peptides presented by disease-associated and non-associated human class II mole- cules. It is anticipated that this approach can be used to investigate how T cell re- sponses to cartilage antigens can be manipulated in a way that might benefit patients. For example, it is now possible to determine which epitopes are the principal inducers of IFN- T production. Based on this information, it will be possible to generate altered peptide ligands for down-modulating the pro-inflammatory Thl-like response, or in- ducing immunoregulatory T cell subsets in a way that has been well documented with other antigens [79]. Along the same lines, it will also be possible to determine whether the response to human cartilage autoantigens can be down-regulated, or even tolerised in vivo by administering either selected peptides or native antigen by a variety of routes. In this way, safer, and more selective and specific forms of immunotherapy can be developed, either for use alone or in combination with current modalities such as disease-modifying anti-rheumatic drugs or new biologicals such as anti-TNF mono- clonal antibody therapy [80]. The importance of selectively targeting both the pro-in- flammatory response, as well as T cell responses is underscored by the observation that overexpression of human TNF in transgenic mice, or expression of a single TCR trans- gene may be sufficient in their own right to induce spontaneous arthritis [81, 82]. These important disease models also indicate that, for reasons that are still unclear, systemic inflammation/autoimmunity targets synovial joints, and suggest that sus- tained immune responses to a single cartilage-derived antigen may turn out not to be essential for the arthritic process in man.
36 A. R Cope, G. Scnderstrup
References
1. Castano L, Eisenbarth GS (1986) Type I diabetes: a chronic autoimmune disease of humans, mouse and rat. Annu Rev Immunol 8:647
2. Katz JD, Benoist C, Mathis D (1995) T helper subsets in insulin dependent diabetes. Science 268:1185 3. Liblau R, Tisch R, Bercovici N, McDevitt HO (1997) Systemic antigen in the treatment of T cell medi-
ated autoimmune disease. Immunol Today 18:599 4. Nordling C, Karlsson-Parra A, Jansson L, Holmdahl R, Klareskog L (1992) Characterization of a spon-
taneously occurring arthritis in male DBA/1 mice. Arthiitis Rheum 35:717 5. Fugger L, Michie SA, Rulifson I, Lock CB, Sonderstrup-McDevitt G (1994) Expression of HLA-DR4
and human CD4 transgenes in mice determines the variable region p-chain T-cell repertoire and medi- ates an HLA-DR-restricted immune response. Proc Natl Acad Sci USA 91: 6151
6. Harris EJ (1997) Clinical features of rheumatoid arthritis. In: Kelley W, Harris E, Ruddy S, Sledge C (eds) Textbook of Rheumatology. Sannders, Philadelphia, pp 898-932
7. Pincus T, Brooks RH, Callahan LF (1994) Prediction of long-term mortality in patients with rheumatoid arthritis according to simple questionnaire and joint count measures. Ann Intern Med 120:26
8. Silman AJ, MacGregor A, Thomson W, Holligan S, Carthy D, Ollier WER (1993) Twin concordance rates for rheumatoid arthritis: results of a nationwide study. Br J Rheumatol 32:903
9. Deighton CM, Walker DJ, Griffiths ID, Roberts DF (1989) The contribution of HLA to rheumatoid arthritis. Clin Genet 36:178
10. Winchester R (1994) The Molecular basis of susceptibility to rheumatoid arthritis Adv Immunol 56:389 11. Stastny P (1978) Association of B-ceil alloantigen DRw4 with rheumatoid arthritis. N Engl J Med 298:
869 12. Gregerson PK, Silver J, Winchester RI (1987) The shared epitope hypothesis: an approach to under-
standing the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis Rheum 30:1205 13. Nepom GT, Byers R Seyfried C, Healey LA, Wilske KR, Stage D, Nepom BS (1989) HLA genes asso-
ciated with rheumatoid arthritis: identification of susceptibility alleles using specific oligonucleotide probes. Arthritis Rheum 32:15
14. Wordsworth BP, Lanchbury JS, Sakkas LI, Welsh KI, Panayi GS, Bell JI (1989) HLA-DR4 subtype fre- quencies in rheumatoid arthritis indicate that DRB1 is the major susceptibility locus within the HLA class II region. Proc Natl Acad Sci USA 86:10049
15. Brown JH, Jardetsky TS, Gorga JC, Stern LJ, Urban RG, Strominger JL, Wiley DC (1993) Three-di- mensional structure of the human class II histocompatibility antigen HLA-DR1. Nature 364:33
16. Stem LJ, Brown JH, Jardetsky TS, Gorga JC, Urban RG, Strominger JL, Wiley DC (1994) Crystal Structure of the human class lI MHC protein HLA-DR1 complexed with an influenza virus peptide. Na- ture 368:215
17. Hammer J, Gallazzi F, Bono E, Karr RW, Guenot J, Valsasnini P, Nagy ZA, Sinigaglia F (1995) Peptide binding specificity of HLA-DR4 molecules: correlation with rheumatoid arthritis association. J Exp Med 181:1847
18. Fu X, Bono CP, Woulfe SL, Swearingen C, Summers NL, Sinigaglia F, Sette A, Schwartz BD, Karr RW (1995) Pocket 4 of the HLA-DR(c~31 "0401) molecule is a major determinant of T cell recognition of peptide. J Exp Med 181:915
19. Todd JA, Acha-Orbea H, Bell JI, Chao N, Fronek Z, Jacob CO, McDermott M, Sinha AA, Timmerman L, Steinman L, McDevitt HO (1988) A molecular basis for MHC class II-associated autoimmunity. Sci- ence 240:1003
20. Wucherpfennig KW, Strominger JL (1995) Selective binding of self peptides to disease-associated ma- jor histocompatibility complex (MHC) molecules: a mechanism for MHC-linked susceptibility to hu- man autoimmune diseases. J Exp Med 181:1597
21. Barber LD, Bal V, Lamb JR, O'Hehir RE, Yendle J, Hancock RJ, Lechler RI (1991) Contribution of T- cell receptor-contacting and peptide-binding residues of the class II molecule HLA-DR4 Dwl 0 to sero- logic and antigen-specific T-cell recognition. Hum Immunol 32:110
22. Walser-Kuntz DR, Weyand CM, Weaver AJ, O'Fallon WM, Goronzy JJ (1995) Mechanisms underlying the formation of the T cell receptor repertoire in rheumatoid arthritis. Immunity 2:597
23. Dessen A, Lawrence CM, Cupo S, Zaller DM, Wiley DC (1997) X-ray crystal structure of HLA-DR4 (DRA*010I, DRB 1 "0401) complexed with a peptide from human collagen II. Immunity 7:473
24. Garboczi DN, Ghosh P, Utz U, Fan QR, Biddison WE, Wiley DC (1996) Structure of the complex be- tween human T-cell receptor, viral peptide and HLA-A2. Nature 384:134
Evaluating candidate autoantigens in rheumatoid arthritis 37
25. Garcia KC, Degano M, Stanfield RL, Brunmark A, Jackson MR, Peterson PA, Teyton L, Wilson IA (1996) An alpha-beta T cell receptor structure at 2.5 A and its orientation in the TCR-MHC complex. Science 274:209
26. Penzotti JE, Doherty D, Lybrand TR Nepom GT (1996) A structural model for TCR recognition of the HLA class II shared epitope sequence implicated in susceptibility to rheumatoid arthritis. J Autoimmun 9:287
27. Weyand CM, McCarthy TG, Goronzy JJ (1995) Correlation between disease phenotype and genetic het- erogeneity in rheumatoid arthritis. J Clin Invest 95:2120
28. Weyand CM, Xie CR Goronzy JJ (1992) Homozygosity for the HLA-DRB1 allele selected for extra-ar- ticular manifestations in rheumatoid arthritis. J Clin Invest 89:2033
29. Griffiths MM, Eichwald EJ, Martin JH, Smith CB, DeWitt CW (1981) Immunogenetic control of ex- perimental type II collagen-induced arthritis. Arthritis Rheum 24:781
30. Nabozny GH, Baisch JM, Cheng S, Cosgrove D, Griffiths MM, Luthra HS, David CS (1996) HLA-DQ8 transgenic mice are highly susceptible to collagen-induced arthritis: a novel model for human pol- yarthritis. J Exp Med 183:27
31. Cope AR Fugger LH, Chu W, S~nderstrup-McDevitt G (1997) Collagen induced arthritis in HLA- DR4/human CD4 transgenic mice. In: Charron D (ed) Genetic diversity of HLA. Functional and med- ical implications. EDK, Paris, pp 634-637
32. Rosloniec EF, Brand, DD, Myers LK, Esaki Y, Whittington KB, Zaller DM, Woods A, Stuart JM, Kang AH (1998) Induction of autoimmune arthritis in HLA-DR4 (DRB 1"0401) transgenic mice by immu- nization with human and bovine type II collagen. J Immunol 160:2573
33. Rosloniec EF, Brand DD, Myers LK, Whittington KB, Gumanovskaya M, Zaller DM, Woods A, Alt- mann DM, Stuart JM, Kang AH (1997) An HLA-DR1 transgene confers susceptibility to collagen-in- duced arthritis elicited with human type II collagen. J Exp Med 185:1113
34. Eden W van, Holoshitz J, Nevo Z, Frenkel A, Klajman A, Cohen IR (1985) Arthritis induced by a T- lymphocyte clone that responds to Mycobacterium tuberculosis and to cartilage proteoglycans. Proc Natl Acad Sci USA 82:5117
35. Taylor PC, Plater-Zyberk C, Maini RN (1995) The role of the B cells in the adoptive transfer of colla- gen-induced arthritis from DBA/I (H-2q) to SCID (H-2d) mice. Eur J Immunol 25:763
36. Hawa M, Rowe R, Lan MS, Notkins AL, Pozzilli R Christie MR, Leslie RD (1997) Value of antibodies to islet protein tyrosine phosphatase-like molecule in predicting type I diabetes. Diabetes 46:1270
37. Elkon KB (1998) Autoantibodies in SLE. In: Klippel JH, Dieppe PA (eds) Rheumatology. Mosby, Lon- don, section 5.1-10
38. Karopoulus C, Rowtey MJ, Ilic MZ, Handley CJ (1996) Presence of antibodies to native GI domain of aggrecan core protein in synovial fluid levels from patients with various joint diseases. Arthritis Rheum 39:1990
39. Holmdahl R, Malmstrom V, Vuorio E (1993) Autoimmune recognition of cartilage collagens. Ann Med 25:251
40. Mottonen T, Hannonen P, Oka M (1988) Antibodies against native collagen II do not precede the clini- cal onset of rheumatoid arthritis. Arthritis Rheum 31: 776
41. Smiley JD, Sachs C, ZiffM (1968). In vitro synthesis of immunoglobulin by rheumatoid synovial mem- brane. J Clin Invest 47:624
42. Holmdahl R, Mo JA, Jonsson R, Karlstrom K, Scheynius A (1991) Multiple epitopes on cartilage type II collagen are accessible for antibody binding in vivo. Autoimmunity 10:27
43. Parekh RB, Dwek RA, Sutton BJ, et al (1985) Association of rheumatoid arthritis and primary os- teoarthritis with changes in the glycosylation pattern of total serum IgG. Nature 316:452
44. Schetlekens GA, Jong BA de, Hoogen FH van den, Putte LB van de, Venrooij WJ van (1998) Citrulline is an essential constituent of antigenic determinants recognized by rheumatoid arthritis-specific autoan- tibodies. J Clin Invest 101:273
45. Londei M, Savill CM, Verhoef A, Brennan F, Leech ZA, Duance V, Maini RN, Feldmann M (1989) Per- sistence of collagen type II-specific T-cell clones in the synovial membrane of a patient with rheuma- toid arthritis. Proc Natl Acad Sci USA 86:636
46. Goodacre JA, Pearson JP (1992) Human cartilage proteoglycans as T-cell autoantigens. Ann Rheum Dis 51:1094
47. Guerassimov A, Zhang Y, Banerjee S, Cartman A, Leroux JY, Rosenberg LC, Esdaile J, Fitzcharles MA, Poole AR (1998) Cellular immunity to the G1 domain of cartilage proteoglycan aggrecan is enhanced in patients with rheumatoid arthritis but only after removal of keratan sulfate. Arthritis Rheum 41 : 1019
38 A. R Cope, G. S0nderstrup
48. Glant TT, Buzas El, Finnegan A, Negroiu G, Cs-Szabo G, Mikecz K (1998) Critical roles of gly- cosaminoglycan side chains of cartilage proteoglycan (aggrecan) in antigen recognition and presenta- tion. J Immunol 160:3812
49. Alsalameh S, Mollenhauer J, Hain N, Stock K-R Kalden JR Burmester GR (1990) Cellular immune re- sponse toward human articular chondrocytes. Arthritis Rheum 33:1477
50. Verheijden GF, Rijnders AWM, Bos E, Coenen-deRoo CJJ, van Stavern CJ, Elewaut D, De Kayser F, Veys E, Boots AMH (1997) Human cartilage glycoprotein-39 is a candidate autoantigen in rheumatoid arthritis. Arthritis Rheum 40:1115
51. Hain NA, Stuhlm~ller B, Hahn GR, Kalden JR, Deutzmann R, Burmester GR (1996) Biochemical char- acterization and microsequencing of a 205-kDa synovial protein stimulatory for T cells and reactive with rheumatoid factor containing sera. J Immunol 157:1773
52. Life PE Bassey EOE, Gaston JSH (1991) T-cell recognition of bacterial heat shock proteins in inflam- matory arthritis. Immunol Rev 121:113
53. Scorer E, David-Ameline J, Peyrat MA, Moreau-Aubry A, Pinczon D, Lim A, Even J, Semana G, Berth- elot JM, Breathnach R, Bonneville M, Houssaint E (I996) T cell response to Epstein-Barr virus trans- activators in chronic rheumatoid arthritis. J Exp Med 184:1791
54. Albani S, Keystone EC, Nelson JL (1995) Positive selection in autoimmunity: abnormal immune re- sponses to a bacterial dnaJ antigenic determinant in patients with early rheumatoid arthritis. Nat Med 1: 448
55. Waaler E (1940) On the occurrence of a factor in human serum activating the specific agglutination of sheep blood corpuscles. Acta Pathol Microbiol Scand 17:172
56. Michaelsson E, Malmstrom V, Reis S, Engstrom A, Burkhardt H, Holmdahl R (1994) T cell recognition of carbohydrates on type II collagen. J Exp Med 180:745
57. Maurice MM, Lankester AC, Bezemer AC, Geertsma MF, Tal P-R Breedveld FC, Lier RAW van, Ver- weij CL (1997) Defective TCR-mediated signaling in synovial T cells in rheumatoid arthritis. J Im- munol 159:2973
58. Cope AR Liblau RS, Yang XD, Congia M, Laudanna C, Schreiber RD, Probert L, Kollias G, McDevitt HO 1997 Chronic tumor necrosis factor alters T cell responses by attenuating T cell receptor signaling. J Exp Med 185:1573
59. Babbitt BE Allen PM, Matsueda G, Haber E, Unanue ER (1985) Binding of immunogenic peptides to Ia histocompatibility molecules. Nature 17:359
60. Friede T, Gnau V, Jung G, Keilholz W, Stevanovic S, Rammensee HG (1996) Natural ligand motifs of closely related HLA-DR4 molecules predict features of rheumatoid arthritis associated peptides. Biochim Biophys Acta 1316:85
61. Rammensee HG, Friede T, Stevanovic S (1995) MHC ligands and peptide motifs: first listing. Im- munogenetics 4:178
62. Moudgil KD, Sekiguchi D, Kim SY, Sercarz EE (1997). Immunodominance is independent of structural constraints. J Immunol 159:2574
63. Sercarz EE, Lehmann PV (1993) Dominance and crypticity ofT cell antigenic determinants. Annu Rev Immunol 11:729
64. Busch R, Mellins ED (1996) Developing and shedding inhibitions: how MHC class II molecules reach maturity. Curr Opin Immunol 8:51
65. Watts C (1997) Capture and processing of exogenous antigens for presentation on MHC moecules. Ann Rev Immunol 15:821
66. Prakken AB, Eden W van, Rijkers GT, Kuis W, Toebes EA, Graeff-Meeder ER de, Zee R van der, Zegers BJ (1996) Autoreactivity to human heat-shock protein 60 predicts disease remission in oligoar- ticular juvenile rheumatoid arthritis. Arthritis Rheum 39:I826
67. Mertz AK, Ugrinovic S, Lauster R, Wu P, Grolms M, Bottcher U, Appel H, Yin Z, Schiltz E, Batsford S, Schauer-Petrowski C, Braun J, Distler A, Sieper J (t998) Characterization of the synovial T ceil re- sponse to various recombinant Yersinia antigens in Yersinia enterocolitica-triggered reactive arthritis. Heat-shock protein 60 drives a major immune response. Arthritis Rheum 41:315
68. Elias D, Markovits D, Reshef T, Zee R van der, Cohen IR (1990) Induction and therapy of autoimmune diabetes the in non-obese diabetic (NOD/LT) mouse by a 65 kDa heat shock protein. Proc Natl Acad Sci USA 87:1576
69. Kotzin BL, Striebich CC, Falta MT, Wang Y, Bill J (1998) Identification of pathogenic T cells in rheumatoid arthritis. Pathogenesis of rheumatoid arthritis: implications for future therapy. In: Proceed- ings of Keystone Symposia on Molecular and Cellular Biology. Abstract, Tamarron, p 18
Evaluating candidate autoantigens in rheumatoid arthritis 39
70. Dybwad A, Forte O, Kjeldsen-Kragh J, Natvig JB, Sioud M (1993) Identification of new B cell epitopes in the sera of rheumatoid arthritis patients using a random nanopeptide phage library. Eur J Immunol 23: 3189
71. Hakala BE, White C, Recklies AD (1993) Human cartilage gp-39, a major secretory product of articu- lar chondrocytes and synovial cells, is a mammalian member of a chitinase protein family. J Biol Chem 268:25803
72. Hu B, Trinh K, Figueira WF, Price PA (1996) Isolation and sequence of a novel human chondrocyte pro- tein related to mammalian members of the chitinase protein family. J Biol Chem 271:19415
73. Morfison BW, Leder P (1994) neu and ras initiate murine inflammatory tumors that share genetic mark- ers generally absent in c-myc and int-2-initiated tumors. Oncogene 9:3417
74. Kouskoff V, Fehling H-J, Lemeur M, Benoist C, Mathis D (1993) A vector driving the expression of for- eign cDNAs in the MHC class II positive cells of transgenic mice. J Immunol Methods 166:287
75. Fugger L, Rothbard JB, S~anderstrup-McDevitt G (1996) The Specificity of the dominant HLA- DRB l*0401-restricted T cell response to type II collagen. Eur J Immunol 26:928
76. Cosgrove D, Gray D, Dierich A, Kaufman J, Lemeur M, Benoit C, Mathis D (1991) Mice lacking MHC class II molecules. Cell 66:1051
77. Killeen N, Sawada S, Littman DR (1993) Regulated expression of human CD4 rescues helper T cell de- velopment in mice lacking expression of endogenous CD4. EMBO J 12:1547
78. Patel SD, Cope AR Congia M, Chen TT, Kim E, Fugger L, Wherret D, SOnderstrup-McDevitt G (1997) Identification of immunodominant T cell epitopes of human glutamic acid decarboxylase 65 using HLA-DR(c~I*0101, [31"0401) transgenic mice. Proc Natl Acad Sci USA 94:8082
79. Gaur A, Boehme SA, Chalmers D, Crowe PD, Pahuja A, Ling N, Brocke S, Steinman L, Conlon PJ (1997) Amelioration of relapsing experimental autoimmune encephalomyelitis with altered myelin ba- sic protein peptides involves different cellular mechanisms. J Neuroimmunol 74:149
80. Elliott MJ, Maini RN, Feldmann M, Long-Fox A, Charles R Bijl H, Woody JN (1994) Repeated ther- apy with monoclonal antibody to tumour necrosis factor alpha (cA2) in patients with rheumatoid arthri- tis. Lancet 344:1125
81. Keffer J, Probert L, Cazlaris H, Georgopoulos S, Kaslaris E, Kioussis D, Kollias G (1991) Transgenic mice expressing human tumour necrosis factor: a predictive genetic model of arthritis. EMBO J 10: 4025
82. Kouskoff V, Korganow AS, Duchatelle V, Degott C, B enoist C, Mathis D (1996) Organ-specific disease provoked by systemic autoimmunity. Cell 87:811
