Oligoclonal expansion of circulating and tissue-infiltrating CD8⁺ T Cells with killer/effector phenotypes in juvenile dermatomyositis syndrome

Abstract

Although triggering by infectious agents and abnormal immune responses may play some role in the pathogenesis of juvenile dermatomyositis syndrome (JDMS), the precise mechanism of muscle destruction and vascular damage is largely unknown. In this study, we tried to elucidate the role of cytotoxic T cells in two patients with JDMS, who were diagnosed based on the characteristic symptoms, laboratory data, MRI findings and electromyographic patterns. Peripheral blood T cell phenotypes were determined by flow cytometry, using mAbs against specific T cell receptor (TCR) Vβs. Complementarity-determining region3 (CDR3) size analysis was performed by gene scanning of CDR3 polymerase chain reaction (PCR) amplification products specific for each Vβ. Subsequently, CDR3 nucleotide sequences were obtained after cloning of the predominant products. The distribution of lymphocytes infiltrating the muscle tissue was analysed by immunohistochemistry. In both patients examined, a unique combination of TCR Vβ repertoires was increased within the CD8⁺ T cells. These subpopulations expressed a characteristic phenotype, indicating that they are memory/effector T cells with killer functions. At the same time, immunohistological and molecular biological examinations of the biopsied muscle samples revealed that identical CD8⁺ T cell clones with identical phenotypes/TCR Vβ infiltrated within the inflammatory tissue, in particular around vessels. These findings indicate that oligoclonal expansion of CD8⁺ T cells plays a central role in the pathogenesis of muscle injury in the juvenile form of dermatomyositis syndrome and may provide a useful clinical parameter of disease activity and responsiveness to anti-inflammatory therapy.

INTRODUCTION

Patients with juvenile dermatomyositis syndrome (JDMS) present with typical skin manifestations and progressive weakening of muscles. Pathological examination usually reveals characteristic myopathic findings, including perifascicular atrophy of muscle fibres, lymphocyte infiltration and evidence of vasculitis involving small vessels. Triggering by infectious agents in genetically susceptible individuals has been regarded to play some role in the pathogenesis [1–5]. However, precise mechanisms leading to characteristic muscle destruction and vascular damage are largely unknown. Although the clinical manifestations resemble those of the adult form of dermatomyositis, JDMS has been considered to be a distinct clinical entity for several reasons [6–10]. Selective atrophy and/or necrosis of perifascicular fibres is much more common in JDMS patients than in adult dermatomyositis or polymyositis (PM) groups. Furthermore, results of tests for antinuclear antibody frequently are positive in JDMS, but myositis-specific antibodies (MSA) are much less frequently found in JDSM than in adult patients. It is noteworthy that the juvenile patients with positive MSA share similar clinical characteristics with adult patients with identical autoantibody profiles [11]. It is also known that peripheral blood B cells increase and CD8⁺T cells decrease during the acute phase of JDMS, suggesting that some autoimmune mechanism is involved in the pathogenesis of this childhood illness [12–16]. However, the cause of the immune abnormality has not yet been clarified.

Autoimmune diseases are characterized by tissue injury caused by autoantibodies or killer T cells reacting with self antigens. Organ specific or widespread, systemic tissue injury results, depending on the nature of the autoreactive antibodies or T cells. Recent data has shown that polymyositis syndrome is likely caused by autoreactive T cells, because T cell analysis in these patients revealed markedly deviated complementarity-determining region 3 (CDR3) size distribution patterns [17–19]. In contrast, no such T cell abnormality was observed in dermatomyositis syndrome of adults [18,20].

In this study, we investigated the immunopathological features of two cases of JDMS and characterized the nature of expanding CD8⁺ T cells both in the peripheral blood and within muscle tissue.

PATIENTS AND METHODS

Patient profiles

Two patients with a diagnosis of JDMS were focused on in this study. They came to our hospital because of symmetric muscle weakness. Despite apparent muscle involvement, creatinine kinase was not increased, and serum aldolase was only slightly elevated. In contrast, characteristic dermatological manifestations (Gottron papules, heliotrope rash and poikiloderma) were observed. Neopterin concentrations were elevated in both patients. Flowcytometric analysis revealed marked increase of CD20⁺ B cells and relative decrease of CD8⁺ T cells in both cases. The diagnosis was based on these clinical symptoms, laboratory data, myopathic changes on electromyography (fibrillation, polyphasic neuromuscular unit and repetitive complex discharge) and typical histological findings on muscle biopsy. They had received no treatment for JDMS before the muscle biopsy and the initial blood sampling. Neither of the patients had other collagen diseases or side-effects of treatments for JDMS. We used samples from JDMS patients after obtaining informed consent from the parents. The study was approved by the Ethics Committee at our institution.

Immunohistochemistry

The muscle biopsy specimens were taken from quadriceps femoris muscles. Hematoxylin-eosin staining and immunohistochemical staining were performed as described elsewhere [21]. Briefly, sections were fixed in acetone for 15 min at 4°C and dried for 15 min at room temperature. They were blocked with 4% normal goat serum for 20 min after washing three times with tris-buffered saline (TBS, pH 7·4), followed by incubation with primary monoclonal antibodies for 60 min. The antibodies used in this study, anti-CD4 (MT310, mouse IgG1), anti-CD8 (C8/144B, mouse IgG2b) and anti-CD20 (L26, mouse IgG2a) were all purchased from DAKO (Glostrup, Denmark). After further washing three times with PBS, sections were incubated with the second antibody (ENVISION/AP Anti-mouse and anti-rabbit, DAKO) for 30 min, and washed with TBS again. All sections were exposed to substrate solution composed of 10 ml of TBS (pH 8·6) with 10 mg of Fast red TR salt and 2 mg of naphthol-acetate. The substrate solution was prepared immediately before use, and levamisole was added at 0·24 mg/ml to block intrinsic alkaline phosphatase activity. All sections were counterstained with haematoxylin. In patient 1, we counted numbers of infiltrating cells (CD4⁺ T, CD8⁺ T, CD20⁺ cells) within the interfascicular area and within the perivascular area, separately. A total of five sections were examined, and the mean values are shown in the figure. Numbers of Vβ8⁺ cells within CD8⁺ T cells are shown as shaded bars.

Flow cytometry

T cell receptor (TCR) Vβ repertoire distribution among peripheral blood T cells was analysed by 3-colour flow cytometry. After washing twice in PBS, 100 µl peripheral blood samples were incubated with appropriate phycoerythrin-conjugated monoclonal antibodies with specificity for TCR Vβ 1–24 (Coulter Immunotech, Marseille, France) for 30 min on ice, fluorescein isothiocyanate-conjugated anti-CD8 (RPA-T8, mouse IgG1; Becton Dickinson, San Diego, CA, USA) and R-phycoerythrin-Cy5-conjugated anti-CD4 (MT310, mouse IgG1; DAKO) for 15 min on ice. After RBC lysis, the cells were washed twice with PBS again. The resulting cells were analysed with a flow cytometer (FACS Calibur, Becton Dickinson). The data were analysed by CellQuest software (Becton Dickinson). We analysed TCR Vβ repertoire distributions of normal controls (n = 20) and calculated the means ± SD.

Surface antigens expressed on T cells with particular TCR Vβ were also analysed by a flow cytometry. Peripheral blood samples were incubated with PE-conjugated anti-TCR Vβ, RPE-Cy5-conjugated anti-CD8 (DK25, mouse IgG1; DAKO) and FITC-conjugated antibodies against HLA-DR (G46-6, mouse IgG2a; Becton Dickinson), CD25 (M-A251, mouse IgG1; Becton Dickinson), CD57 (NK-1, mouse IgM; Becton Dickinson), CD69 (FN50, mouse IgG1; Becton Dickinson), CD62L (DREG56, mouse IgG1; Coulter Immunotech), CD8β (2ST8·5H7, mouse IgG1; Coulter Immunotech), CD28 (CD28·2, mouse IgG1; Coulter Immunotech) or CD45RO (UCHL1, mouse IgG2a; DAKO) for 15 min on ice. After RBC lysis and washing, 3 colour analysis was performed with a FACS Calibur.

Total RNA extraction and reverse transcriptase – polymerase chain reaction (RT-PCR)

Total RNA was extracted as described previously with slight modification [22]. Briefly, peripheral blood mononuclear cells (PBMC) were separated from peripheral blood by Ficoll/hypaque gradient. T cells were separated from PBMC by E-rosette method [23]. CD4^– and CD8^– T cells were purified from T cells by negative selection using anti-CD8 or CD4 magnetic beads (Dynal ASA, Norway). Fractions of the contaminating CD4⁺ T cells within CD4-depleted cells, or CD8⁺ T cells within CD8-depleted cells were always less than 1% after the negative selection procedures. Total cellular RNA was isolated from CD4^– T cells, CD8^– T cells and muscle tissue with TRIZOL reagent following manufacturer's instructions (Gibco BRL, Bethesda, MD, USA). The RNA was then reverse-transcribed into cDNA in a reaction containing RandamHex Primer (TaKaRa, Otsu, Japan) and RAV-2 (TaKaRa). The concentration of RNA was measured using a GeneQuant proRNA/DNA Calculator (Amersham Pharmacia Biotech, Cambridge, UK).

TCR Vβ CDR3 size analysis

CDR3 spectratyping was pursued as previously described [23]. Briefly, cDNA was PCR-amplified through 35 cycles (94°C for 1 min, 55°C for 1 min, and 72°C for 1 min) with a primer specific to 24 different BV subfamilies (BVs 1–20 [24] and BVs21–24 [25]) and a fluorescent BC primer [24]. The fluorescent PCR products were mixed with formamide and the size standard (GeneScan-500 TAMRA, Applied Biosystems, Foster, CA, USA). After denaturation for 2 min at 90°C, the products were analysed with an automated 310 DNA sequencer (Applied Biosystems), and the data were analysed with GeneScan software (Applied Biosystems). The overall complexity within a Vβ subfamily was determined by counting the numbers of discrete peaks and determining their relative size on the spectratype histogram, as described previously [26]. We used a complexity scoring system with our interpretation, that is complexity score = (sum of all the peak heights/sum of the major peak heights) × (number of the major peaks); major peaks were defined as those peaks on the spectratype histogram whose amplitude was at least 10% of sum of all the peak heights.

TCR CDR3 cloning and sequence

PCR products of selected BV cDNA were electrophoresed on an agarose gel and purified using QIAquick Gel Extraction Kit (QIAGEN, Tokyo, Japan), and then cloned with TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA, USA). Eleven to 19 colonies containing the insert fragment were randomly selected. Purified with QIAprep Spin Miniprep Kit (QIAGEN), the recombinant plasmids were subjected to fluorescence dye terminator cycle sequencing, and the sequence reactions were analysed on a 310 DNA sequencer (Applied Biosystems) after removal of the unincorporated fluorescence dye with Centri-Sep Spin Columns (Applied Biosystems).

RESULTS

Muscle biopsy

In both patients, haematoxylin-eosin staining of the muscle biopsy specimens exhibited characteristic findings. Among these, perifascicular atrophy and lymphocyte infiltration were particularly noteworthy (Fig. 1m). Lymphocyte infiltration was intense around small vessels (v) and adjacent to the necrotic muscle fibres (f). Immunohistochemical staining of muscle specimens from patient 1 (Fig. 1a–f) and patient 2 (Fig. 1g–1) revealed that CD20⁺ lymphocytes (Fig. 1c,i) were predominant within the perivascular area, with small numbers of CD4⁺ (Fig. 1a,g) or CD8⁺ (Fig. 1b,h) lymphocytes. In contrast, CD8⁺ lymphocytes (Fig. 1e,k) and CD4⁺ (Fig. 1d,j) predominated within the interfascicular area and CD20⁺ (Fig. 1f,l) lymphocytes were much fewer. These changes were more prominent in patient 1 than in patient 2. Representative data from patient 1 are shown to illustrate the characteristic distribution of lymphocyte subpopulations within the area of muscle inflammation around the blood vessels (Fig. 1m,v) and interfascicular muscle destruction (Fig. 1m,f). Within the area of muscle destruction, both CD4⁺ T cells and CD8⁺ T cells predominated and CD20⁺ B cells were rare (Fig. 1n). In contrast, most of the lymphocytes infiltrating around the vessels were CD20⁺ B cells (Fig. 1o). It was also found in patient 1 that Vβ8⁺ T cells (Fig. 1n,o, shown by green shading within CD8⁺ T cells) predominated around the vessels but were absent in other areas of the muscle tissue.

Fig. 1.

Immunohistochemical staining of the muscle biopsy specimens from patient 1 (a–f) and 2 (g–l). Tissue infiltrating lymphocytes within the perivascular (a–c, g–i) and interfascicular areas (d–f, j–l) are stained with anti-CD4 (a,d,g,j), anti-CD8 (b,e,h,k) or anti-CD20 (c,f,i,l) monoclonal antibody. Tissue distributions of lymphocyte subpopulations. Numbers of different lymphocyte subpopulations per area were determined by the immunohistochemical stainings. Hematoxylin and eosin staining of the muscle biopsy specimen from patient 1 shows the characteristic perifascicular atrophy and perivascular lymphocyte infiltration (m). Cell density within the interfascicular area (f) and perivascular area (v) were compared, and the number of each cell population per area is shown in (n) and (o), respectively. Coloured bars within CD8⁺ T cells indicate the numbers of Vβ8⁺ cells.

TCR Vβ repertoire distributions

In both patient 1 and 2, the patterns of TCR Vβ repertoire distributions were normal and the percentages of each repertoire remained within two standard deviations of normal control values within CD4⁺ T cells (Fig. 2a). In marked contrast, TCR Vβ8 and TCR Vβ22 were significantly increased in patient 1, and TCR Vβ1 and TCR Vβ20 were increased in patient 2 within CD8⁺ T cells (Fig. 2b).

Fig. 2.

Distributions of TCR Vβ repertoires of peripheral blood T cells in JDMS patients. Open squares indicate the specific repertoires which were significantly increased and were selected for further analysis. (a) CD4⁺ T cells. (b) CD8⁺ T cells. Analysis of surface antigen expression on the expanding TCR Vβ repertoires. Three-colour flow cytometric analysis of various surface antigen expression was performed for Vβ8⁺ T cells from patient 1 (c) and Vβ1⁺ T cells from patient 2 (d) after gating CD8⁺ T cells.

T cell surface antigen phenotype

In patient 1, Vβ8⁺CD8⁺T cells did not express activation antigens, CD25, CD69 or HLA-DR. In contrast to the rest of the cells, Vβ8⁺ cells lacked the expression of CD62L or CD45RO, and expression of CD8β was significantly reduced (Fig. 2c). Similar to patient 1, the expressions of CD45RO, CD62L and HLA-DR were significantly lower on Vβ1⁺CD8⁺T cells than other cells in patient 2. In addition, CD28 expression was selectively reduced in this fraction (Fig. 2d). Most Vβ8⁺CD8⁺T cells in patient 1 and Vβ1⁺CD8⁺T cells in patient 2 expressed significant levels of CD57. Control repertoire, TCR Vβ17⁺CD8 T cells in patient 2 expressed significant levels of CD28 and CD62L but lacked CD57 on their surface (data not shown).

CDR3 size analysis

In peripheral blood of both patient 1 and patient 2, the results revealed that the spectratyping profiles of most Vβ repertoires in CD4⁺ T cells displayed Gaussian distributions (Fig. 3a,d). In contrast, a large fraction of Vβ repertoires within CD8 positive T cells displayed extremely skewed patterns (Fig. 3b,e). Significant skewing of CDR3 size distributions, as shown by a complexity score lower than 4, was seen in TCR Vβ 8, 13·6, 23 in patient 1 and Vβ 1, 7, 24 in patient 2. CDR3 size analysis of the lymphocytes within the muscle tissues was performed at the same time. Most of the profiles were skewed or undetectable in both patients, reflecting partly the paucity of lymphocytes within the analysed tissue compared with in peripheral blood (Fig. 3c,f).

Fig. 3.

CDR3 spectratyping of patient 1 (a–c) and 2 (d–f). CDR3 size analysis for each TCR Vβ was performed using cDNA samples from peripheral blood CD4⁺ T cells (a,d), CD8⁺ T cells (b,e) and muscle infiltrating T cells (c,f). n.d. denotes samples which showed undetectable levels of spectra patterns.

DNA sequence analysis in TCR CDR3 region

In patient 1, clones with identical CDR3 nucleotide sequences were detected both within the peripheral blood and within the muscle tissue. In particular, Vβ8⁺ clone with CDR3 nucleotide sequence, LLKGHE, predominated in the peripheral blood (55·5%), and in the muscle tissue (46·7%) (Table 1). In patient 2, a Vβ1⁺ clone with CDR3 nucleotide sequence, VGSPPTN, was increased both in the peripheral blood (57·1%), and in the muscle (66·7%) (Table 2).

Table 1. TCR CDR3 amino acids sequences of CD8 Vβ8 from patient 1.

Sample	BV	N-D-N	BJ	Frequency
PBMC	CASS	LLKGHE	QYFGP	10/18	55.5%
	CASS	LGASQPQH	QYFGD	1/18	5.6%
	CASS	PLFSDSYE	FGPG	3/18	16.5%
	CASS	LGTGTA	YEQYFG	1/18	5.6%
	CASS	LRNSGANVLT	FGAG	1/18	5.6%
	CASS	LVTGGQETGT	FGPG	1/18	5.6%
	CASS	PIGTSGS	NEQFFGP	1/18	5.6%
Muscle	CASS	LLKGHE	QYFGP	7/15	46.7%
	CASS	PLFSDSYE	QYFGP	1/15	6.7%
	CASS	FGGTGVFEQT	FGPG	1/15	6.7%
	CASS	PEASGRA	DTQYFGP	2/15	13.3%
	CASS	LFGTGGG	GELFFGE	1/15	6.7%
	CASS	PRATGFN	EKLFFGS	1/15	6.7%
	CAS	APTGVFRVD	EQFFGP	1/15	6.7%
	CAS	RARFPGLAGY	NEQFFGP	1/15	6.7%

Table 2.

TCR CDR3 amino acids sequences of CD8 Vβ1 from patient 2

Sample	BV	N-D-N	BJ	Frequency
PBMC	CASS	VGQGYE	QYFGP	1/28	3.6%
	CASS	VVLIT	NEQFFGP	1/28	3.6%
	CASS	VGSPPT	NEQFFGP	16/28	57.1%
	CASS	PTSGSS	YGYTFGS	6/28	21.3%
	CASS	RASAAN	YGYTFGS	1/28	3.6%
	CASS	VGGGRAYE	QYFGP	1/28	3.6%
	CASS	EAFGRGYY	EQFFGP	1/28	3.6%
	CASS	SGSGSFY	NEQFFGP	1/28	3.6%
Muscle	CASS	PS	GELFFGE	1/9	11.1%
	CASS	VGSPPT	NEQFFGP	6/9	66.7%
	CASS	AFYGGPGSET	QYFGP	1/9	11.1%
	CASS	APTEAGVRT	DTQYFGP	1/9	11.1%

DISCUSSION

The pathogenesis of autoimmune diseases is complex. Both humoral and cellular immunity are thought to be involved in the characteristic tissue damage. In some cases, infectious agents such as bacterial or viral pathogens are closely related to the triggering of the disease process [1–5]. Massa et al. [1] identified self T cell epitope (myosin heavy chain) and its homolog (Streptococcus M5 protein) in JDMS patients and indicated T cell cross-recognition of bacterial and human homologs by analysing TCR Vβ gene usage of peptide-specific T cells. Specific types of HLA class I or class II molecules are associated with increased risks of certain autoimmune diseases [27]. Recent studies showed that the abnormal distributions of TCR Vβ repertoires are observed in certain autoimmune diseases or viral infections, reflecting oligoclonal expansion of antigen-specific T cells in response to the precipitating antigens [28–31]. In autoimmune diseases, control of the expanding T cell clones by immunosuppressive agents may lead to reversion of the tissue damage and clinical improvement. Zeng et al. [32] analysed the T cell characteristics of patients with hypoplastic anaemia and showed that those who responded to cyclosporin treatment exhibited marked skewing of TCR Vβ usages, whereas cyclosporin nonresponders exhibited almost normal distributions. The result indicated that active suppression of T cell function is effective when autoreactive T cells are involved in the pathogenesis. Thus, it is important to elucidate the roles of T cells in the pathogenesis of autoimmune diseases before introducing any therapeutic intervention.

JDMS is characterized by a marked decrease in circulating CD8⁺ T cells and significant increase in CD20⁺ B cells, suggesting that immune mechanisms are involved in its pathogenesis [12–16]. Muscle biopsy findings also implicate antibodies and T cells reacting with autoantigens in the muscle injury. Immunoglobulins and complement are deposited on vessels and muscle fibres. Both B cells and T cells infiltrate around the vessels and interfascicular areas. In accord with these findings, it is reported recently that muscle tissue of JDMS patients is infiltrated with T cells containing specific CDR3 Ag-binding region, indicating that oligoclonal expansion of antigen-specific T cells plays significant roles in the pathogenesis of JDMS [1].

In adults, it is recently reported that polymyositis, but not dermatomyositis, patients display severe perturbations of peripheral blood T cell TCR repertoires [17,20]. Both of these groups concluded that polymyositis is characterized by the presence of antigen-specific T cells with killer/effector phenotype and that examination of TCR repertoire offers a useful parameter to discriminate polymyositis and dermatomyositis. However, Benveniste et al. did not separate CD4⁺ and CD8⁺ T cells to perform CDR3 analysis of dermatomyositis patients [20]. Lower levels of repertoire perturbation among CD8⁺ T cells will be easily obliterated in such analysis, especially when circulating CD8⁺ T cell numbers are significantly decreased. Furthermore, adult forms of dermatomyositis may have a distinct aetiology from JDMS and may exhibit different T cell characteristics [6–10].

In this study, we examined TCR structures in the peripheral blood and muscle tissue of two JDMS patients. Oligoclonal expansion of particular T cells was evaluated by flowcytometric analysis of TCR Vβ repertoire distributions and by CDR spectratypings. In both cases, abnormal expansion of CD8⁺ T cells with particular TCR Vβ structures was proved in the peripheral blood. Several pieces of evidence show that the increase of these Vβ repertoires was functionally significant and represent oligoclonal expansion of antigen-specific killer/effector T cells.

First, these cells expressed distinct antigens from other T cells. Most of these antigens are closely associated with killer/effector phenotypes of cytotoxic effector T cells, suggesting that these T cells proliferate in response to certain antigens. Although we could not perform thorough examinations of surface antigen expressions for both patients, we assume that the increased fractions in both patients express identical profiles of surface antigens. In patient 1, Vβ8⁺ T cells expressed a significantly decreased level of CD8β chains. We recently reported that CD8β^low T cells represent memory/effector T cells [23]. They share common surface phenotypes, including negative CD45RO, HLA-DR, CD62L or CD28 [33–36]. In contrast, these T cells express significantly elevated levels of CD57, perforin and granzyme, indicating that these T cells exert potent killer/effector function. Furthermore, CD8β^low T cells have markedly reduced TCR diversity, as determined by CDR3 size distributions, suggesting that oligoclonal T cells are concentrated within these cell fractions. In patient 2, Vβ1⁺ T cells also expressed low levels of CD8β chains and were strongly positive for CD57 [34,35]. We could not evaluate killer activity of these cells by direct cytotoxicity assay. But these data indicate that the proliferating T cells are phenotypically memory/effector T cells which have the potential to damage the muscle tissue as antigen-specific killer T cells.

Second, CDR3 spectratyping revealed that these T cells not only used common Vβ but also showed identical PCR product sizes. Physiologically, CDR3 lengths follow a gaussian-like distribution that reflects the great diversity of polyclonal TCR specificities within a given Vβ family [37]. Significant skewing of CDR3 size distributions, as shown by a complexity score lower than 4, was seen TCR Vβ 8, 13·6, 23 in patient 1 and TCR Vβ 1, 7, 24 in patient 2. But only TCR Vβ 8 in patient 1 and TCR Vβ 1 in patient 2 were expanded in peripheral blood. It is unclear whether TCR Vβ 13·6, 23 in patient 1 and TCR Vβ 7, 24 in patient 2 played any role in the pathogenesis. The skewing of CDR3 size distributions in these repertoires may reflect past infections. Considering the total numbers expanding in the peripheral blood, TCR Vβ 8 in patient 1 and TCR Vβ 1 in patient 2 are thought to have played a major role in the pathogenesis. In addition, CDR3 nucleotide sequence analysis revealed that the predominant clones share identical nucleotide sequences, indicating that these T cells are of a clonal origin. These findings indicate that T cells with the particular TCR Vβ did not expand as a result of polyclonal triggering such as superantigens. The results support the view that antigen-specific T cell clones proliferated in vivo and the numbers increased both in the peripheral blood and within the muscle tissue as memory/effector T cells. In performing CDR3 analysis of the infiltrating cells in muscle, we do not separate CD4⁺ and CD8⁺ T cells. However, CDR3 nucleotide sequences of predominant clones in the patient's peripheral CD8⁺ T cells are identical to those of the infiltrating T cells. So most of the infiltrating T cells in the muscle and predominantly CD8⁺ T cell clone in the peripheral blood are thought to be of the same origin. It is intriguing if we could examine the CDR3 structures of the T cells infiltrating the skin lesion simultaneously, and see if the predominant T cells clones within the muscle lesion and the skin lesion exhibit common CDR3 structures. With such approach, we will further understand the pathogenesis of JDMS.

Third, in both patients, analysis of muscle biopsy samples revealed that CD8⁺ T cells infiltrated within the interfascicular muscle tissue and around the blood vessels. In contrast, CD20⁺ B cells were concentrated around the small vessels, but absent within the interfascicular region. These findings are in accord with previous reports and may indicate distinct roles of CD8⁺ T cells and CD20 B cells. It is intriguing that TCR Vβ8⁺ T cells infiltrated around the small vessels but were absent within the interfascicular region in patient 1. These T cells may be involved in the cell injury of both muscle fibres and blood vessels. We need to analyse more cases to better characterize the functional roles of these T cells.

These findings collectively indicate that CD8⁺ T cells with killer/effector functions proliferated in response to precipitating antigens and triggered muscle injury in JDMS. To further support this contention, patient 1 responded to steroid with marked clinical improvement. In parallel with clinical improvement, TCRVβ8⁺CD8⁺ T cells decreased progressively. In contrast, patient 2 has been resistant to various forms of anti-inflammatory and immunosuppressive therapy, and TCRVβ1 CD8⁺ T cells are still increased in the peripheral blood. The results indicate that oligoclonal expansion of these T cells not only provides useful information as to the pathogenesis of JDMS, but also is a valuable parameter to evaluate the clinical responsiveness to a particular therapy. Analysis of TCR structures may offer an invaluable guide to the therapy to JDMS, and further analysis of similar cases will help clarify the pathogenesis of this childhood autoimmune disease.