Article Text


TCRβ spectratyping in RA: evidence of clonal expansions in peripheral blood lymphocytes
  1. F C Halla,
  2. K Thomsona,
  3. J Procterb,
  4. A J McMichaelc,
  5. B P Wordswortha
  1. aNuffield Department of Clinical Medicine, John Radcliffe Hospital, Oxford, bOxford Transplant Centre Tissue Typing Laboratory, Churchill Hospital, Oxford, cInstitute of Molecular Medicine, John Radcliffe Hospital, Oxford
  1. Dr F C Hall, Department of Microbiology and Immunology, Fairchild Building D345, Stanford University, CA 94605–5124, USA.


OBJECTIVE To compare the TCRβ repertoire of peripheral blood CD8 enriched (CD8+) and depleted (CD8−) T cells in rheumatoid arthritis (RA) patients and controls using CDR3 length analysis (spectratyping).

METHODS CD8+ and CD8− T cells were separated from 14 RA patients and 12 controls, using magnetic beads coated with anti-CD8 monoclonal antibodies. cDNA was prepared as the template for amplification with 22 Vβ-Cβ primer pairs. The products were resolved by electrophoresis in an ABI373 sequencer using GENESCAN software. Expansions were identified as dominant CDR3 lengths, where the area underlying the corresponding peak exceeded the sum of the areas of the two adjacent peaks. This method was validated by sequencing 10 samples displaying dominant peaks. The expansion frequencies in RA patients and controls were compared using the χ2 test statistic.

RESULTS Dominant peaks were evident in several Vβ families. They were more frequent in RA patients in both the CD8+ subset (RA normalised frequency 10.6; control normalised frequency 8.0; p=0.03) and the CD8− subset (RA normalised frequency 2.9; control normalised frequency 1.5; p=0.02). Sequencing of 10 samples exhibiting dominant peaks revealed an unequivocal clonal expansion in nine (90%).

CONCLUSIONS RA patients exhibited a significantly increased frequency of T cell expansions both in the CD8+ and CD8− subsets. This phenomenon may reflect the proliferation of autoreactive cells, a non-specific expansion of memory T cells in response to pro-inflammatory cytokines or a defect of T cell regulation that predates the onset of RA and may itself predipose to disease.

  • rheumatoid arthritis
  • T cell
  • clonal expansion

Statistics from

A pivotal role of T cells in the initiation of rheumatoid arthritis (RA) has been inferred from the prominence of CD4+ T cells in the synovium early in disease, the association of RA with particular HLA-DR4 and DR1 antigens, the efficacy of treatments that deplete or suppress CD4+ cells, and animal models of chronic arthritis. The T cell receptor (TCR) endows this cell with its discriminatory properties. In developing T cells, functional genes for the α and β chains of the TCR chain are produced by rearranging selected Vα/Jα or Vβ/Dβ/Jβ segments and joining them to their respective constant regions. Unlike the CDR1 and CDR2 loops, which are encoded in the germline, the CDR3 loops are determined by the VαJα or VβDβJβ junctional region and diversity is further enhanced by trimming the germline segments and insertion of non-germline encoded nucleotides. The resulting variation in length of the CDR3β loop is the basis for the “spectratyping” technique, in which the CDR3β region is amplified and the distribution of CDR3β lengths can be used as a readout of TCRβ repertoire.

It is widely believed that T cells initiate a pathogenic immune response in RA by specific recognition of antigen locally in the joint, leading to synovitis, pannus formation, and joint damage. Although no consistent TCR motif has emerged in RA patients,1repertoire studies using monoclonal antibodies and CDR3 length analyses have identified abnormal patterns of CD4+ clonal expansion in both peripheral blood and synovial T cell populations.2 3 CD8+ T cells are less strongly implicated in RA but expansions of a CD8+ large granular lymphocyte population occur in approximately 33% of cases of Felty’s syndrome.4 Their role is unclear but MHC restricted anti-viral activity, antibody dependent cell mediated cytotoxicity, and suppressor activity have been suggested. CD8+ expansions have also been described in RA by spectratyping5 and using anti-TCR monoclonal antibodies.6

In this study, TCRβ spectratyping was used to compare the repertoire of 22 Vβ families in both the CD8+ and CD8− subsets of peripheral blood lymphocytes in 14 RA patients and 12 controls. The number of CDR3 lengths per Vβ family was determined as an overall measure of repertoire complexity and the frequencies of CD8+ and CD8− expansions in each Vβ family were compared between RA and controls.


Fourteen unrelated RA patients (11 female, 3 male), who fulfilled the ACR revised criteria,7 were recruited from a rheumatology outpatient department. Twelve controls (9 female, 3 male) were recruited from hospital staff. All were white British subjects. The mean (SD) age of patients was 54 (13), whereas, for controls, it was 42 (16). Table 1 shows the demographic characteristics, HLA-DR type, and rheumatoid factor status of subjects.

Table 1

Demographic characteristics and HLA-DR types of RA patients and controls

Peripheral blood mononuclear cells (PBMC) were separated from peripheral venous blood by Ficoll-Hypaque gradient centrifugation (Lymphoprep, Birmingham, UK). A three hour stimulation of PBMC with 1 μg/ml phytohaemagglutinin (PHA) in α modified Eagle’s medium supplemented with 5% AB human serum was performed to increase the yield of mRNA from T cells. The CD8+ and CD8− T cell subsets were separated using magnetic beads coated with anti-CD8 monoclonal antibodies (Dynal, Oslo, Norway), and the extent of separation was assessed by fluorescence analysis on a Becton-Dickinson FACScan with Cellquest software, using phycoerythrin labelled anti-CD4 and fluoroscein isothiocyanate labelled anti-CD8 monoclonal antibodies (Sigma, Poole, UK). Total RNA was prepared from aliquots of 1–5 × 106 CD8+ or CD8− mononuclear cells using the RNAzolB method (Biotecx, Houston, TX, USA). RNA concentration was assessed by optical density and first-strand cDNA synthesis was performed on 5 μg aliquots of RNA using 0.1U oligo(dT)12–18 primer (Pharmacia, Uppsala, Sweden) and AMV reverse transcriptase (Promega, Madison, WI, USA).

The cDNA from 14 RA patients and 12 controls was used as template for 22 Vβ-Cβ amplifications, which were performed in 96 well plates (Costar, High Wycombe, UK). The primer sequences were modified from the set used by Gregersen and coworkers.8 The Vβ10 and Vβ19 primers were excluded, because these BV segments are transcribed pseudogenes.9 The Vβ22 primer described in the Gregersen set is a perfect match for BV23 rather than BV22 and is renamed as the Vβ23 primer in this study. A 10 μl reaction volume contained 1 μl cDNA, 3 pmol of fluorescently labelled Vβ primer, 2 pmol CβR, 16 mM ammonium sulphate, 67 mM TRIS-HCl (pH 8.8), 0.1% TWEEN-20, 0.2 mM dNTPs, 0.5 mM magnesium chloride, and 1.25 units of Taq polymerase (Bioline, London, UK). Omnigene thermocyclers (Hybaid, Teddington, UK) were programmed to provide 35 amplification cycles of 94°C for one minute, 55°C for one minute, 72°C for 45 seconds. The amplification products were separated by electrophoresis on 6% acrylamide gels (Severn Biotech Ltd, Kidderminster, UK) each run at a constant 1000 V in ABI373 sequencers (Applied Biosystems, Warrington, UK) and the gels were collected and analysed using GENESCAN software.10

The GENESCAN GENOTYPER package was programmed to label the 10 highest peaks in each spectratype with the peak area. An expansion was defined as a peak with an area exceeding the sum of the areas of the two adjacent peaks. In cases where an apparently dominant peak was adjacent to a gap in the spectratype, or lay at the edge of the spectratype, an expansion was defined as a peak with an area at least twice that of the single adjacent peak. The frequency of expansions was expressed as the normalised expansion frequency relative to the proportion of successful V-C amplifications in an indivdual subject. Comparative statistics were performed using the χ2 test statistic.

Twelve amplification products (10 containing expansions; two without expansions) were subcloned into the pMosBlue T vector (Amersham International, Amersham, UK) and sequenced using the USB sequencing kit (United States Biochemical Corp, Cleveland, Ohio). The CDR3 loop boundaries were defined by the the conserved amino acid motif, CAS, encoded by the BV gene segment, and the FGXG motif encoded by the BJ gene segment.


The majority of Vβ-Cβ spectratypes consisted of between four and nine peaks representing CDR3 lengths differing by increments of three base pairs. The mean (SD) number of CDR3 lengths per spectratype was 6.0 (1.2) in RA patients and 6.5 (1.2) in controls (NS). In comparison with PBMC, which contained an average of 11% CD8+ cells, CD8− enriched samples consisted of at least 85% CD8+ cells, and CD8− depleted samples of less than 3% CD8+ cells (data not shown). The complexity of CD8+ and CD8− spectratypes, using this measure, was equivalent. The area under the peaks was approximately normally distributed, with a modal CDR3 length of 10 (0.8) inferred amino acid residues. Comparison of spectratypes from unstimulated PBMC or PBMC stimulated for up to 12 days with PHA indicated no change in the repertoire detectable by this technique (results not shown).

The amplification efficiency of the panel of 22 Vβ-Cβ primer pairs was 65% for RA patients and 76% for controls. Expansions were evident as dominant peaks against the background spectratype (fig 1) and the expansion frequencies, normalised for the amplification efficiency of each Vβ family, are displayed in figure 2. RA patients displayed an expansion frequency at least 50% greater than in controls in families Vβ3, 6, 21, and 24 in the CD8− subset and Vβ3, 6, 7, 8, 15, 20, 21, and 23 in the CD8+ compartment. The CDR3 length of expansions ranged between 5 and 18 inferred amino acid residues with means of 10 for the control CD8− expansions, 11 for RA CD8− expansions, and 10 for both control and RA CD8+ expansions. No association was evident between the number of expansions and the age of the person.

Figure 1

Dominant CDR3 lengths are indicated (*) against a background Gaussian distribution of CDR3 length distribution. Each peak in the spectratype is annotated with the area underlying the curve. The y axis represents relative units of fluorescence calculated by the Genescan software. The x axis indicates the length of PCR product (bp).

Figure 2

The frequency of expansions in CD8− and CD8+ subsets of peripheral blood T cells are indicated in both RA patients and controls. Frequencies are normalised according to the proportion of successful Vβ-Cβ amplifications in a given person.

Table 2 displays the sequencing data from 10 expansions in six RA patients. A clonal expansion was demonstrated unequivocally in nine of 10 samples exhibiting dominant peaks. The use of J segments was non-random, with a preponderance for BJ*2S1, BJ*2S3, and BJ*2S7. Two control samples, without dominant peaks, revealed a variety of unique sequences.

Table 2

Sequence data from 10 dominant peaks identifed in RA patients


This spectratyping analysis of peripheral blood lymphocytes from RA patients and controls enables a qualitative assessment of the TCRβ repertoire across 22 Vβ families, covering 80–90% of the expressed TCRβ repertoire. Within a given Vβ spectratype, each peak represents the in-frame rearrangement of TCRβ genes with identical CDR3 length. Clonal T cell expansions can be detected against a polyclonal background with a sensitivity of approximately 1:1000 (results not shown). RA patients and controls had comparable numbers of CDR3 lengths per Vβ family (mean (SD) values 6.5 (1.2) in controls and 6.0 (1.2) in RA), indicating no macroscopic change in the peripheral blood TCRβ repertoire complexity in RA.

Dominant peaks were evident in several Vβ spectratypes, both in RA patients and controls. However, the frequency of expansions was significantly higher in RA patients than controls in both the CD8+ (p=0.03) and CD8− (p=0.02) subsets. The CD8 depleted samples consisted predominantly of CD4+ lymphocytes. However, clonal expansion within the CD4–CD8− compartment cannot be excluded in this study and it is possible that this population includes high affinity autoreactive T cells, which may be relevant in the pathogenesis of autoimmune disease. Sequencing of dominant peaks in this study revealed the presence of a clonal expansion in about 90% of cases. BJ segments *2S1, *2S3, and *2S7, encoding the motif Q(F/Y)FGPG, were expressed preferentially, but their frequency was consistent with the non-random use of BJ segments previously reported in normal subjects.11

The biological significance of non-malignant peripheral blood expansions remains unclear. Initially CD8+ expansions were detected using monoclonal antibodies in elderly human subjects.12These were thought to represent age related impairment of clonal regulation. While there is no correlation between frequency of expansions and age in this study, it remains possible that larger expansions develop in older subjects, and are detectable by less sensitive techniques. The majority of CD8+ expansions in normal and RA subjects have been shown also to express CD57.13 This phenotype has been associated with an MHC independent proliferative response to human fibroblasts infected with cytomegalovirus,14 but the significance of cytomegalovirus in the pathogenesis of RA remains controversial.15 16 The emergence of clonal expansions has been reported after antigenic stimulation,8 and the CD45RO+ phenotype of CD4+3 and CD8+12 17 expansions suggests a memory population. Surprisingly, expansions frequently appear to be CD28−, a phenotype thought to be associated with anergy.18 However, the CD4+CD28− expansions described previously in RA patients apparently retain the capacity to proliferate in response to TCR ligation.3

This study shows that clonal expansions are not merely the province of senescence and that microheterogeneity also exists in young subjects. The increased frequency of expansions evident in RA patients may reflect the proliferation of autoreactive clones, although the lack of CDR3 sequence conservation in the clones studied provides no evidence for a common autoantigen. Alternatively, peripheral blood expansions may reflect a pool of recently or repeatedly stimulated memory cells, recognising common pathogens, such as herpes viruses. Stimulation with interferon α has been shown to expand memory CD8+ cells in the absence of specific antigenic stimulation19 and proinflammatory cytokines in RA patients could result in the proliferation of certain memory cell populations. Finally, the finding that some unaffected relatives of RA display increased frequencies of expansions has prompted the suggestion that an underlying defect in T cell homeostasis predates and may predispose to disease.2


We would like to thank the patients and controls who took part in this study and Ann Burrows for invaluable secretarial assistance. FCH was supported by the Medical Research Council. K T was supported by the Wellcome Trust.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.