Analysis of immune reconstitution after autologous bone marrow transplantation in systemic sclerosis

Authors


Abstract

Objective

To analyze hematopoietic and immune reconstitution after autologous hematopoietic stem cell transplantation (HSCT) in 7 patients with systemic sclerosis (SSc).

Methods

Two groups of patients were retrospectively constituted according to whether they had a favorable clinical response (group A; n = 4) or no response or a relapse of disease (group B; n = 3) after HSCT. Immune reconstitution was analyzed every 3 months using lymphocyte immunophenotyping, α/β T cell receptor (TCR) diversity analysis, and ex vivo thymic function analysis by quantification of TCR rearrangement excision circles (TRECs).

Results

Patients had similar characteristics at study entry, except for a lower modified Rodnan skin thickness score (P = 0.03) and a lower Health Assessment Questionnaire score (P = 0.05) in group A than in group B. The number of reinjected cells and the time to hematopoietic reconstitution were similar in both groups. The absolute numbers of CD19+ and CD20+ B cells were lower in group A than in normal controls (P < 0.05) and within the normal range in group B. Absolute numbers of T and natural killer lymphocytes were normal before HSCT. Numbers of CD3+ cells remained low thereafter. Numbers of CD8+ cells were back to normal 3 months after HSCT in both groups. B cell counts were low until 6 months after HSCT in group A and stayed in the normal range in group B. The CD3+ defect was sustained in group A, with an opposite trend and a faster CD4+ reconstitution profile in group B. The T cell repertoire was skewed before and until 1 year after HSCT, with shared expansions before and after transplant in a given individual. TREC values correlated negatively with C-reactive protein levels (rs = −0.41, P = 0.001) and positively with CD19+ (rs = 0.35, P = 0.001) and CD20+ (rs = 0.34, P = 0.002) lymphocyte counts.

Conclusion

B and T lymphocyte populations remained disturbed for at least 1 year after HSCT in SSc patients, which may reflect the persistence of an underlying disease mechanism.

Systemic sclerosis (SSc) is a heterogeneous autoimmune disease characterized by excessive collagen deposition within the skin and internal organs (1). Although its exact pathogenesis remains unknown, predominant T cell activation (2), production of autoantibodies (including anti–topoisomerase I [anti–Scl-70]), and cytokine release all contribute to increased collagen synthesis and deposition, fibroblast activation, microvascular damage, and vascular injury. Since 1996, the use of high doses of cyclophosphamide (CYC) followed by autologous peripheral blood stem cell or bone marrow transplantation in several phase I–II studies allowed impressive clinical responses in severe diffuse SSc (3–5), with complete or partial remission in two-thirds of patients up to 3 years after hematopoietic stem cell transplantation (HSCT) (6). However, the onset of relapse and the presence of sustained disease activity after longer followup periods in 20–35% of patients may be attributed to the recovery of autoaggressive clones or autoantigen rechallenge (5, 6). Analysis of immune reconstitution in this setting is key to a better understanding of the efficacy of this therapy, taking into account that the reintroduction of immunosuppressive drugs to control relapse may interfere with immune reconstitution.

We analyzed hematopoietic and immune reconstitution in 7 SSc patients who were followed up for at least 1 year after HSCT prior to potential reintroduction of immunosuppression. Assessment of T cell reconstitution used combined approaches of phenotyping, α/β T cell receptor (TCR) diversity analysis, and TCR rearrangement excision circles (TRECs) quantification as a marker of ex vivo thymic function.

PATIENTS AND METHODS

Patients.

The study included 7 patients (3 men and 4 women; mean ± SD age 35 ± 17 years) diagnosed according to the criteria for SSc subsets (1), with severe refractory disease and early visceral involvement, who had been treated with CD34+-selected HSCT in a phase I–II trial (the Intensification et Autogreffe dans les Maladies Auto Immunes Résistantes [ISAMAIR] study; see Appendix A for a list of investigators) (5). The study was approved by the ethics committee, and patients gave their written informed consent. During the month prior to inclusion, patients received either no treatment (n = 2) or low-dose oral steroids (<20 mg/day) (n = 5). Blood was obtained from healthy donors (after informed consent) to determine the reference values of T cell repertoire diversity and TREC content (n = 10 and 41 samples, respectively).

Transplant procedure and followup.

The transplant procedure and followup were performed as described elsewhere (5). Briefly, mobilization and collection of peripheral blood HSCs used CYC at 2 gm/m2/day on 2 consecutive days, followed 4 days later by recombinant human granulocyte colony-stimulating factor (rHuG-CSF) (Lenograstim; Aventis Pharma, Antony, France and Chugai Pharma, Puteaux, France) at 5 μg/kg/day subcutaneously until the last apheresis. Peripheral blood stem cells were collected when there were >20 CD34+ cells/μl in peripheral venous blood, and CD34+ cells were selected using an anti-CD34 monoclonal antibody, immunomagnetic bead technique (Isolex 300i Stem Cell Collection System; Nexell, Irvine, CA). Conditioning was performed at least 4 weeks later using CYC at 50 mg/kg/day from day −5 to day −2 prior to HSC reinjection. All patients received rHuG-CSF after the graft.

Clinical and biologic followup.

Clinical and biologic followup were performed daily until the end of aplasia and then every 3 months during the first year after HSCT, as previously described (5). Primary hematopoietic reconstitution was defined by a neutrophil count >500/mm3 for 3 consecutive days and a platelet count >20,000/mm3 without transfusion. Clinical response to therapy was assessed by the same observer, who categorized responses as major, partial, or no response, or disease progression or relapse according to repeated functional and organ evaluations using previously published criteria (5). Hematopoietic progenitors were assayed on apheresis products. Blood lymphocyte subsets were assessed at inclusion and every 3 months thereafter. Anti–Scl-70 antibodies were measured every 3 months by enzyme-linked immunosorbent assay, and quantified results (7) were expressed in arbitrary units/ml.

Lymphocyte phenotyping.

Lymphocyte immunophenotyping was performed on fresh samples of whole blood preserved in EDTA by direct 4-, 3-, or 2-color immunofluorescence flow cytometry (Becton Dickinson, San Jose, CA) before and 3, 6, and 9 months after HSCT. Measurements of forward and side scatter were combined with CD45 and CD14 determinations to identify lymphocytes and exclude monocytes (gate purity >98%). The following monoclonal antibodies and combinations were used: anti-CD45–peridin chlorophyll protein; anti-CD3–fluorescein isothiocyanate (FITC); anti-CD4–allophycocyanin; anti-CD8–phycoerythrin (PE) and anti-CD45–FITC; anti-CD14–PE and anti-CD4–FITC; anti-CD45RO–PE and anti-CD4–PE; anti-CD45RA–FITC and anti-CD3–FITC; and anti-CD16–PE, anti-CD56–PE, anti-CD19–PE, and anti-CD20–FITC. All antibodies were purchased from Becton Dickinson, except for anti-CD20–FITC, which was purchased from Immunotech (Marseilles, France). Isotype-matched controls were included with each sample. Five thousand gated lymphocytes were analyzed using a FACSCalibur analyzer (Becton Dickinson). Results were expressed as absolute numbers of cells and as percentages of the total numbers of cells at inclusion.

RNA extraction and TCR β-chain spectratyping.

RNA and genomic DNA were purified using TRI Reagent (Molecular Research Center, Cincinnati, OH). Synthesis of complementary DNA, third complementarity-determining region (CDR3) gene amplifications, runoff using an internal β-chain gene constant fluorescent primer, gel running, and Immunoscope software analysis (Antoine Toubert, INSERM, Paris, France) were performed as previously described (8). The definition of Immunoscope profiles as polyclonal, skewed, or negative was based on the identification of peaks that deviate from the normal distribution curve, as described previously (8).

Quantification of TRECs by real-time polymerase chain reaction (PCR).

After extraction using TRI Reagent, DNA samples were adjusted to 250 ng/μl. Quantification of thymic signal-joint TRECs was done by real-time quantitative PCR (ABI Prism 7700; Applied Biosystems, Foster City, CA) according to the method of Douek et al (9). Briefly, a standard was created by cloning the signal-joint fragment in pCR2.1A using the TA cloning kit (Invitrogen, Groningen, The Netherlands). We determined the TREC value by quantitative PCR on genomic DNA from peripheral blood mononuclear cells, using an additional set of primers and a probe for albumin to normalize the genomic copy number. The TREC values were corrected according to the percentage of CD3+ cells in the sample and were then expressed as numbers of TRECs/μg of CD3+ cell DNA. Values were measured before HSCT, then at 6–8 months and 10–12 months after HSCT or at the last followup prior to the reintroduction of immunosuppression.

Statistical analysis.

Results in Tables 1–3 are shown as the mean ± SD. Significant differences between patient groups related to individual conditions were assessed with the Mann-Whitney test, which was a nonparametric rank test appropriate for small sample groups and robust to violations of normality and equality of variance (10). P values less than or equal to 0.05 were considered significant. To reduce intragroup variability induced by differences in individual conditions at study entry, we analyzed the immune reconstitution process using mean percentages of the total numbers of cells at inclusion that remained at 3, 6, and 9 months after HSCT. Correlations between immunologic reconstitution data and clinical and biologic parameters were studied and tested for statistical significance by using a nonparametric Spearman rank test that is both robust to violations of normality of the distributions of the variables and fitted for analysis of small-size samples (10). Global trends of the immune reconstitution throughout followup were graphically expressed using linear regression slope (LRS) analysis (10). All statistical analyses were performed within the R environment for statistical computing and graphics (online at http://www.r-project.org/).

Table 1. Patient characteristics at study inclusion and clinical evolution within the first 12 months after HSCT with positive selection for CD34+ cells*
 Group A patientsGroup B patientsMean ± SD
1234123
  • *

    See Results for description of groups. HSCT = hematopoietic stem cell transplantation; STE = steroids; COL = colchicine; CYC = cyclophosphamide; D-Pen = D-penicillamine; Ig = immunoglobulins; Plasma = plasmaphereses; MTX = methotrexate; AZA = azathioprine; PG = prostaglandins; OCT = octreotide; HAQ = Health Assessment Questionnaire; MRSS = modified Rodnan skin thickness score; FEV1 = forced expiratory volume in 1 second; DLCO = diffusing capacity for carbon monoxide; LVEF = left ventricular ejection fraction; MUGA = multiple gated acquisition scan; CRP = C-reactive protein; GM-CFUs = granulocyte–macrophage colony-forming units; MR = major response; PR = partial response; NR = no response; REL = relapse; PRO = progression; ND = not determined.

  • Calculated since first diagnosis of systemic sclerosis.

  • According to World Health Organization classification (see ref. 11).

  • §

    Quantity of CD34+ progenitor cells contained in the graft. CD34+ cell recovery after cryopreservation was 95% (range 75–100%).

  • GM-CFUs/kg were counted on day 14 using a clonogenic progenitor assay as previously described.

  • #

    See ref. 5 for explanation of numerical gradings.

Patient characteristics at inclusion        
 Age, years1661294027205335 ± 17
 SexFMFMFFM
 Disease duration, months522836653242131 ± 17
 Previous treatmentSTE, COL, CYCSTE, D-Pen, COL, CYCIg, CYC, COL, D-Pen, Plasma, MTX, AZA, STESTE, PG, D-Pen, COLSTE, OCTCYC, MTXSTE
 Steroids at inclusion, mg/day301073411 ± 11
 Performance status score (0–4)0121213
 HAQ score (0–3)01.1251.3522.22.1252.51.55 ± 0.92
 MRSS (0–51)914123228425228 ± 16
 FEV1, % predicted78761125876616976 ± 18
 DLCO, %5828614545485849 ± 11
 LVEF (MUGA, in %)6257616063637062 ± 4
 Serum creatinine, units/liter5085666732596060 ± 16
 CRP level, mg/dl3112062313717 ± 21
Graft characteristics and engraftment duration        
 CD3+ cells × 103 infused/kg4.116.423.63.34.74.12.534.11 ± 1.23
 CD34+ cells × 106 infused/kg§6.77.55.94.99.910.34.87.14 ± 2.23
 GM-CFUs × 104 infused/kg3.7912.3710352.569.9736.33 ± 41.94
 Days to 0.5 × 109 neutrophils/liter1491010991010 ± 2
 Days to 25 × 109 platelets/liter89108712109 ± 2
Clinical evolution after HSCT#        
 3 monthsMR1MR1PR1NRPR1PR1NR
 6 monthsMR1MR1MR1MR1MR1NRPR1
 9 monthsMR1MR1RELMR1MR1NRPR1
 12 monthsMR1MR1MR1MR1RELNRPR0
Anti–Scl-70 antibodies, arbitrary units/ml        
 At inclusion21024.5120013039.1
 3 months14.7015.4ND012927.5
 6 monthsND013.345.5012519.4
 9–12 months15.901043.50120104
Table 2. Phenotype analysis of lymphocyte populations after hematopoietic stem cell transplantation (HSCT)*
Lymphocyte populationNormal rangeAt inclusion3 months after HSCT6 months after HSCT9 months after HSCT
Group AGroup BGroup AGroup BGroup AGroup BGroup AGroup B
  • *

    Values are the mean ± SD counts/ml. See Results for description of groups.

Total1,718–2,6201,703 ± 6622,127 ± 880958 ± 2481,015 ± 5531,203 ± 598957 ± 2901,205 ± 4521,165 ± 1,506
CD3+1,008–1,6471,170 ± 5291,369 ± 960756 ± 467498 ± 338767 ± 446482 ± 274724 ± 208673 ± 263
CD3+,CD4+587–1,009711 ± 278736 ± 433204 ± 59186 ± 110232 ± 28195 ± 36260 ± 41300 ± 98
CD3+,CD8+313–644455 ± 235340 ± 142537 ± 399308 ± 282522 ± 441281 ± 247446 ± 239360 ± 222
CD19+121–26744 ± 26149 ± 10058 ± 68255 ± 272109 ± 168325 ± 452113 ± 135287 ± 337
CD20+121–26756 ± 15149 ± 10062 ± 71255 ± 272109 ± 168304 ± 421112 ± 135281 ± 342
CD4+,CD45RO+333–763591 ± 279540 ± 394183 ± 38321 ± 353168 ± 16170 ± 48196 ± 34195 ± 74
CD4+,CD45RA+161–529219 ± 95434 ± 48163 ± 6434 ± 2770 ± 4457 ± 4278 ± 47106 ± 76
CD3−,CD16+,CD56+82–340192 ± 44298 ± 261196 ± 114158 ± 51167 ± 129142 ± 36155 ± 80201 ± 163
Table 3. T cell repertoire diversity, thymic function analysis by quantification of T cell receptor rearrangement excision circles (TRECs), and phenotypes after hematopoietic stem cell transplantation (HSCT)*
 Normal valuesAt inclusion6–8 months after HSCT10–12 months after HSCT
Group AGroup BGroup AGroup BGroup AGroup B
  • *

    Values are the mean ± SD. See Results for description of groups.

  • Or last followup prior to reintroduction of immunosuppressive treatment.

Polyclonal BV families, %70.30 ± 19.8543.17 ± 7.8459.06 ± 12.0122.50 ± 15.5836.50 ± 6.3636.33 ± 18.8323 ± 90
Skewed BV families, %22.80 ± 20.0239.78 ± 2.2533.33 ± 9.2958.25 ± 13.2045.50 ± 19.0949.25 ± 16.1554.67 ± 19.73
Negative BV families, %6.90 ± 7.8137.53 ± 35.717.33 ± 2.8919.25 ± 14.7018 ± 12.7314 ± 16.1622.33 ± 27.54
TRECs/μg CD3+ cell DNA694 ± 776.8550.75 ± 51.05117.33 ± 19.73112.75 ± 180.68201.50 ± 275.06201.67 ± 208.48105.33 ± 116.45

RESULTS

Clinical data.

All 7 patients with severe refractory SSc had been treated with CD34+ HSCT and enrolled in the same prospective phase I–II trial with common inclusion criteria (5) and similar clinical characteristics at entry (Table 1). According to the observed clinical response as previously defined (5) within the first 12 months after HSCT, 2 groups of patients were retrospectively constituted: group A, consisting of 4 patients with sustained major or partial response, and group B, consisting of 3 patients with no response or with relapse of disease who required reintroduction of immunosuppressive therapy 9–12 months after HSCT. Retrospective analysis of patient clinical characteristics at entry showed that performance status according to the World Health Organization classification (11) (mean ± SD score 1 ± 0.8 versus 2 ± 1 in groups A and B, respectively) and renal function (mean ± SD serum creatinine 67 ± 14.3 units/liter versus 50 ± 15.9 units/liter in groups A and B, respectively) were similar in both groups. Cardiac function was slightly lower in group A than in group B (left ventricular ejection fraction 60 ± 2.2% versus 65 ± 4.0%; P = 0.03), but remained in the normal range for both groups (Table 1). The modified Rodnan skin thickness score (MRSS) (17 ± 10.3 versus 43 ± 7.5; P = 0.03) and the Health Assessment Questionnaire (HAQ) score (0.8 ± 0.72 versus 2.3 ± 0.19; P = 0.05) were significantly lower in group A than in group B (12, 13).

HSCT procedure.

The individual patients' graft characteristics and engraftment durations after CD34+ HSCT are shown in Table 1. There was no significant difference between patients from group A and those from group B. Overall, there was a mean ± SD of 7.14 ± 2.23 × 106 reinjected CD34+ cells/kg and a mean ± SD of 4.11 ± 1.23 × 103 reinjected CD3+ cells/kg. Hematopoietic reconstitution was achieved after times to the appearance of neutrophils (>500/mm3) and platelets (>20,000/mm3) of (mean ± SD) 10 ± 2 days and 9 ± 2 days, respectively.

Phenotype analysis of lymphocyte populations after HSCT.

At inclusion, the absolute numbers of CD3+, CD3+,CD4+, CD3+,CD8+, CD4+,CD45RO+, and CD4+,CD45RA+ T cells and the absolute number of CD3−,CD16+,CD56+ natural killer (NK) cells were in the normal range for all patients and did not differ between groups (Table 2). The absolute numbers of CD19+ and CD20+ B cells were lower in group A than in normal controls (P < 0.05) and within the normal range in group B. During the first 9 months after HSCT, the absolute number of CD3+ T cells, and in particular, the absolute number of CD4+,CD45RA+ T cells, remained below normal, whereas the numbers of CD8+ T cells returned to normal 3 months after HSCT in groups A and B. The absolute number of B cells was lower than normal up to 6 months after HSCT in group A patients, whereas it was in the upper range of normal in group B patients 3 months after HSCT and thereafter. Throughout followup, there was a significant positive correlation between CD19+ and CD20+ cell counts and the presence of high titers of anti–Scl-70 antibodies (Spearman's rank correlation coefficient rs = 0.27, P < 0.05).

When we analyzed the immune reconstitution profiles after HSCT, we observed 2 patterns of evolution (Figure 1). In group A, with sustained clinical responses, there was a small decrease in the relative number of CD3+ cells (LRS = −2.29), whereas group B showed an opposite trend, with an increase in the relative number of CD3+ cells (LRS = 2.87). The percentage of B cells during the first 9 months after HSCT relative to the B cell count at inclusion did not vary in group B (LRS = 1.23), but increased rapidly from 3 to 6 months in group A (LRS = 17.71). Memory CD4+,CD45RO+ T cell profiles were similar in both groups, whereas CD4+,CD45RA+ cell reappearance was much more rapid in group B (LRS = 4.45) than in group A (LRS = −0.87).

Figure 1.

Profiles of reconstitution of lymphocyte phenotypes after hematopoietic stem cell transplantation (HSCT). Patients in group A (red) had sustained major or partial responses to HSCT. Patients in group B (blue) either had no response to HSCT or had relapse of disease requiring reintroduction of immunosuppressive therapy 9–12 months after HSCT. Shown are mean percentages of the total numbers of lymphocytes with a given phenotype at inclusion that remained 3, 6, and 9 months after HSCT. A-LRS = linear regression slope for data for group A patients; B-LRS = linear regression slope for data for group B patients.

T cell repertoire diversity and thymic function.

The α/β T cell repertoire was disturbed before HSCT in the patients compared with the age-matched controls (P not significant), with fewer polyclonal BV families and an overexpression of skewed and/or negative families (Table 3). There was no common expansion of a given BV family in this small group of patients (data not shown). The T cell repertoire remained disturbed in both groups of patients until at least 1 year after HSCT, with at least 1 major shared expansion at the same TCR β-chain CDR3 size before and after transplant in a given individual, as illustrated in Figure 2 and Table 4, whereas other BV family profiles remained or became polyclonal. Although not statistically different because of the small size of the patient groups, there was a trend toward improved T cell diversity in group A with favorable outcome. TREC values at inclusion were significantly lower in all patients than in normal controls (P = 0.009). TREC values increased up to 10–12 months after HSCT in group A. In group B, they increased until 6–8 months after HSCT and then decreased at the 10–12-month time point. Throughout followup, the TREC values correlated negatively with levels of C-reactive protein (CRP), a marker of inflammation (rs = −0.41, P = 0.001), and correlated positively with the CD19+ (rs = 0.35, P = 0.001) and the CD20+ (rs = 0.34, P = 0.002) lymphocyte counts.

Figure 2.

T cell receptor (TCR) β-chain spectratyping. Shown is the third complementarity-determining region size distribution of selected BV families before (Pre) and 12 months after (Post) hematopoietic stem cell transplantation for 2 representative patients with systemic sclerosis (patient 1 from group A and patient 2 from group B). See Figure 1 for description of groups.

Table 4. Individual values of the number of polyclonal, skewed, and negative T cell receptor β-chain families before and after hematopoietic stem cell transplantation*
Group, patient, timeSkewedPolyclonalNegativeNo. of conserved expansions/expanded BV families
  • *

    See Results for description of groups.

Group A    
 Patient 1    
  At inclusion61353/BV3.BV12.BV15
  7 months699 
 Patient 2    
  At inclusion8794/BV5.BV16.BV23.BV24
  8 months1833 
 Patient 3    
  At inclusion31743/BV5.BV20.BV23
  6 months11121 
 Patient 4    
  At inclusion51271/BV6
  9 months5811 
Group B    
 Patient 1    
  At inclusion21841/BV7
  9 months4164 
 Patient 2    
  At inclusion31562/BV3,BV5
  6 months5514 
 Patient 3    
  At inclusion51363/BV3.BV5.BV24
  6 months11103 

DISCUSSION

Immune reconstitution after autologous HSCT has been widely studied in hematologic malignancies or cancer (14), but very few data have yet been reported on patients treated for autoimmune diseases (3–5, 14, 15). Immune reconstitution after autologous HSCT includes reappearance of functional B cells, thymic and extrathymic T cell development, reconstitution of effector cells, including cytotoxic T cells and NK cells, and efficient antigen presentation to reconstitute the pretransplantation immune repertoire (14). Surface markers that characterize T lymphocytes and their functional subsets represent only a part of the overall T cell functional repertoire, which is best studied with T cell diversity analysis through the size of the β-chain CDR3 (2) and by quantitating TRECs for ex vivo evaluation of recent thymic function (8, 9, 15–17). These techniques, plus appropriate statistical analysis (10) regarding our small number of patients (the 7 patients for whom clinical and immune reconstitution data were available, of the 12 patients included in the ISAMAIR prospective phase I–II study [5]), allowed us to point out meaningful differences related to condition at study entry and to global trends of immune reconstitution for at least 1 year after HSCT.

All these patients with common clinical characteristics at study entry underwent the same HSCT procedure with the same hematopoietic reconstitution, but when 2 groups were retrospectively constituted according to clinical response, the MRSS and the HAQ score were significantly lower in group A (with sustained clinical response) than in group B (with no response or with relapse of disease). These indices were shown to be good predictors of survival in several large prospective cohorts of SSc patients (1). Their retrospective significance at the time of entry before HSCT underlines the importance of starting intensive immunosuppressive conditioning early enough in the course of severe SSc when aiming to achieve remission. It also supports the recent enlargement of inclusion criteria implemented in the European phase III Autologous Stemcell Transplantation International Scleroderma (ASTIS) trial (online at www.astistrial.com).

Reports of peripheral blood lymphocyte counts in SSc are contradictory, but these counts consistently exhibit signs of activation and proliferation (1, 2). In our study, the absolute peripheral blood T cell counts, including subsets of CD8+ and CD4+ T cells as well as CD4+,CD45RA+, CD4+,CD45RO+, and NK cells, were in the normal range for all 7 SSc patients at study inclusion. At that time, peripheral B cell counts were significantly higher in group B patients than in group A patients, suggesting that pathogenic B cell clones might preferentially expand in these SSc patients with a less favorable outcome. The correlation of high B cell counts with anti–Scl-70 antibodies could support this hypothesis.

After HSCT, as expected (14), NK cells with a CD3−,CD16+,CD56+ phenotype were among the first cells to recover, having returned to a normal level within 1 month. However, B cell reconstitution in group B patients differed markedly from that found after autologous HSCT performed for nonautoimmune conditions, in which numbers of mature CD19+ and CD20+ B cells usually decrease during the first 3 months and increase thereafter to reach a plateau at 6–9 months, as observed in group A patients. Numbers of B cells in group B patients remained within or even above the normal range throughout the transplant followup. Considering T cells, we observed a delayed recovery of CD3+ T cell reconstitution compared with that found after autologous HSCT performed for solid tumors or myeloma (14). The reinjected T cell dose, after CD34+ selection, was similar to or even lower than that reported for autologous HSCT in autoimmune diseases (3–5). Persistent CD4+ T cell lymphopenia (especially the CD4+,CD45RA+ T cell subset) was predominant in all patients, in contrast with a more rapid reconstitution of CD8+ T cells. In SSc patients in the present study, the numbers of CD4+,CD45RA+ and CD4+,CD45RO+ T cells were profoundly reduced up to 9 months after HSCT, with similar CD4+,CD45RO+ T cell profiles in both groups. Interestingly, group A patients (with sustained clinical response) demonstrated a low slope of reconstitution for CD4+,CD45RA+ T cells, whereas group B patients (with no response or with relapse of disease) had a positive slope of reconstitution, with higher values at 9 months.

The evaluation of T cell reconstitution after HSCT has improved through the development of direct methodologies for T cell diversity analysis using the size of the β-chain CDR3 (2) and for ex vivo evaluation of thymic function by quantitating TRECs (8). Using a molecular approach (Immunoscope or spectratyping), we confirmed the abnormalities of the T cell repertoire previously reported with Vα- and Vβ-specific monoclonal antibodies in SSc patients (2, 18). Despite a severe T cell depletion of the graft, 1–3 oligoclonal T cell expansions were conserved before and after transplant (Table 4). This is clearly different from the allogeneic HSCT setting, in which a number of T cell oligoclonal expansions are seen after transplant, but in which they usually differ from the donor's or the recipient's pretransplant T cell repertoire (ref. 8, and Toubert A, et al: unpublished observations). We cannot exclude the possibility that post-HSCT T cell expansions may originate from residual patient T cells that have survived conditioning or from the small CD3+ population contaminating the graft. However, ours is the first study to compare T cell repertoire abnormalities by spectratyping before and after autologous HSCT in SSc patients. Since the definition of autoantigen(s) in SSc is still elusive (2), functional assessment of such recurrent T cell populations will be the scope of future studies.

TREC content in peripheral T cells is a marker of newly exported antigen-naive T cells, but we should also be aware that TREC values are influenced by the degree of proliferation in the peripheral T cell compartment (15). Indeed, the correlation of low TREC values with the CRP level as a marker of inflammation probably reflects disturbances of T cell homeostasis, which is consistent with data reported in patients with rheumatoid arthritis (16) and multiple sclerosis (17). An original finding in the present study is the positive correlation of B lymphocyte counts with increasing TREC values at 6–8 months in group B patients. We may speculate that a rebound of thymic function could provide help to pathogenic autoantibody-producing B cells soon after transplant. However, TREC values decreased further in these patients at 10–12 months, possibly indicating additional pathogenic mechanisms, including a potential dilution effect due to T cell proliferation in the periphery.

In conclusion, our data indicate that perturbations in T cell homeostasis, as evidenced by T cell repertoire analysis and TREC content, may persist for a long period after autologous HSCT for severe refractory SSc and may reflect the persistence of an underlying disease mechanism in these patients (18). Further studies on a larger scale will help to clarify the respective kinetics of T cell and B cell immune reconstitution profiles according to the present trends and to elucidate whether the maintenance or rapid reintroduction of immunosuppressive therapies after graft, as previously suggested (6), may improve patient outcome in refractory disease.

APPENDIX A

ISAMAIR INVESTIGATORS

The following investigators participated in the ISAMAIR study: study board—D. Farge, MD (chair), E. Gluckman, MD (cochair), M. Breban, MD, P. Brice, MD, J. Cabane, MD, D. Charron, MD, P. Cherin, MD, M. C. Douard, MD, C. Gisselbrecht, MD, L. Guillevin, MD, X. Mariette, MD, Z. Marjanovic, MD, J. P. Marolleau, MD, J. P. Messing, MD, N. Mounier, MD, J. C. Piette, MD, P. Ribaud, MD; cytaphereses—J. M. Miclea, MD, N. Parquet, MD; immunophenotyping—C. Rabian, MD; investigators—S. Arfi, MD, C. Carbon, MD, R. Damade, MD, J. Emmerich, MD, E. Hachulla, MD, J. P. Jouet, MD, C. M'Bappe, MD, S. Menasche, MD, P. Philippe, MD, J. C. Piette, MD, P. Roblot, MD, J. Sibilia, MD; statistics—C. Henegar, MD.

Ancillary