T cell genetics and rheumatoid arthritis (RA)

References

Chat With PDF

Cite

Share

Abstract

For over 20 years the relevance of the immune system to the aetiology of RA has been well exemplified by its strong MHC class II associations, particularly with certain alleles of the HLA-DRB1 locus. An emerging consensus points towards the involvement of a conserved region (‘the shared epitope’) of the antigen-binding site of these RA-associated variants, but the precise mechanism(s) involved has proved elusive [1–3]. Given the antigen-presenting properties of these HLA molecules, it is reasonable to suppose that the ability to bind and present a limited array of peptides to the immune system could underlie the HLA-DR associations of the disease. Implicit in this theory is the existence not only of ‘arthritogenic peptide(s)’, but also of highly specific interactions with T lymphocytes expressing particular T cell receptor (TCR) gene rearrangements. Thus polymorphisms of the TCR genes could influence susceptibility to RA in a similar way to HLA-DR polymorphisms. However, the extraordinary diversity of the TCR repertoire derives only in part from germ-line polymorphism and is greatly increased by combinatorial rearrangements of the individual TCR gene segments [4]. In recent years both have been the subject of extensive investigation in RA, with varying results. The TCR interacts with both the HLA-DR molecule and its peptide ligand through three complementarity determining regions (CDR) [5]. The CDR3 overlies the peptide and owes much of its diversity to the nature of the TCRBVDJ and TCRAVJ junctional combinations through the random addition of nucleotides. In contrast, the CDR1 and CDR2 are more involved in interactions with the HLA molecule itself, and the lesser degree of variation within them is derived from polymorphisms of the gene segments of individual TCRBV and TCRAV families [6]. These TCR polymorphisms could therefore influence their capacity to interact with the conserved group of HLA-DR molecules associated with RA. Germ-line TCR polymorphisms have now been extensively investigated in RA using the conventional approaches of genetic linkage analysis in affected sib pairs and also in case control studies of coding polymorphisms in variable gene segments. Early studies suggested the possibility of linkage to the TCRB locus [7], but more extensive investigations have suggested that anything other than a minor genetic effect arising from either the TCRA or TCRB locus is unlikely [8]. Nevertheless, the possibility of a small effect has not been totally excluded. Against this background a weak effect arising from the TCRA locus recently described [9] in a very large international collaborative study is tantalizing. The observed risk associated with a TCRAV8S1 allele (1.3, 95% C.I. 1.1–1.8) is small and, of course, will require independent confirmation. One has also to bear in mind the possibility that it represents a linkage marker for another as yet unspecified locus, although linkage disequilibrium across the TCR loci is limited, particularly compared with HLA. Weak associations have been suggested with certain TCRB gene segments [10–12], but these have not been confirmed in other studies [13–16]. In order to obtain a definitive answer it may be necessary to test a large number of coding polymorphisms spanning these regions, particularly since it is possible that susceptibility could arise from more than one gene segment. Alternative approaches to investigating TCR involvement in RA have assessed T cell populations for evidence of oligoclonality in the periphery or in the synovial compartment. Unfortunately, direct systematic study of T cell populations in humans has been problematic for several reasons, not least because of the relatively limited panel of antibodies available. Although Vβ MoAbs now cover about 60% of the T cell population, there is still a dearth of Vα antibodies. Therefore most recent studies in humans have used variations of the polymerase chain reaction (PCR) to infer the frequencies of TCR families. Techniques such as Vβ family-specific PCR [17] can be rapidly applied to relatively large sample sizes, but unfortunately are only semiquantitative and can only really be used to determine the relative frequencies of particular TCR variable gene families at sites of inflammation compared with peripheral blood. In contrast, classical anchor or inverse PCR [18] may allow more accurate absolute measurements of T cell families, but are very labour-intensive, depending on large amounts of DNA sequencing. Numerous studies using a variety of methods have hinted at skewed TCR gene usage variably in both the peripheral blood and the synovium, hinting at superantigen [19] or conventional antigen-driven T cell responses [20,21], but unfortunately there seems to be very little overall consensus (reviewed in [20,22]). This could reflect the generally small sample sizes, differences in the patients (e.g. duration of disease), differences in the handling of clinical samples prior to analysis (e.g. culture with IL-2) or genetic heterogeneity. However, it is equally possible that the relevant T cell populations are so small as to be undetectable by these methods, that their presence could only be detected at a much earlier stage of the disease process, or that they are to be found mainly at other sites such as the regional lymph nodes. In the December issue of Clinical and Experimental Immunology, Mizushima and colleagues [23] described a PCR/ELISA technique modifying the anchor PCR method so that it can theoretically be applied to the analysis of relatively large sample sizes, thereby combining many of the advantages of anchor and TCRBV-specific PCR methods. In this study, identical twins, one of whom had RA, showed larger differences in TCRBV usage than twins who were healthy. However, the strong genetic influence on shaping the TCR repertoire was illustrated by the fact that even these discordant twin pairs were more alike than unrelated individuals. Oligoclonal expansions in the CD8 subset were seen with approximately similar frequencies in both types of twin pair and the authors suggest that the differences in the twins discordant for RA reflect an increased frequency of oligoclonal expansions in the CD4 subset of T cells. This is consistent with the observations of others [24–26]. The study of Mizushima et al. is very limited in scale, but does hint at the possibility of CD4 expansions in the peripheral blood of patients that could be relevant to RA. Previously, Waase et al. have also pointed to the existence of such expansions, and suggested that since they can also be found in the siblings of affected individuals with RA they might indicate a more general defect of immunoregulation [25]. Claims of this kind are not new, but one of the advantages of the technique that Mizushima et al. have employed is that it could be applied to the larger studies necessary to prove the concept that particular T cell populations are over-represented in RA. Ultimately it remains to be seen whether this approach or others based on TCRBV-specific PCR employing analysis of CDR3 length variants (‘spectratyping’) [26–28] analysis using MoAbs, or combinations of these methods will be most effective at answering these questions. What is certain is that these studies are extremely technically demanding, and any method which offers the prospect of analysing larger sample sizes is to be welcomed wholeheartedly. Perhaps it is also salutary to note that while similar oligoclonal expansions of T cell populations can be seen in reactive arthritis [29], we do at least have some idea of the triggering antigens in that disease and these may give us a handle to tease out these T cell responses. Such luxury still awaits us in rheumatoid arthritis.