T-cell receptor repertoire usage after allografting differs between CD4+CD25+ regulatory T cells and their CD4+CD25− counterpart

Abstract

Background and Objectives After allogeneic haematopoietic stem cell transplantation (SCT) the whole T-cell receptor (TCR) repertoire shows a markedly skewed pattern for 2–3 years. A small fraction of CD4⁺ T cells is represented by CD25⁺ regulatory lymphocytes (T_reg), which play a crucial role in modulating peripheral tolerance. To investigate their ability to react to the massive antigenic stimulation generated in an allogeneic host, which could significantly affect their pattern of reconstitution, we analyzed the TCR repertoire of T_reg after SCT, focusing on the degree of similarity to CD4⁺CD25⁻ conventional T cells (T_conv).Design and Methods We assessed the TCR Vβ repertoire of T_reg in ten patients who had received allogeneic SCT, by using complementarity determining region 3 (CDR3) spectratyping. We developed a new similarity score for the analysis. This score expresses the proportion of Vβ with similar profile between T_reg and T_conv.Results For up to 3 years after SCT the repertoires of T_reg and T_conv were characterized by several Vβ with different profiles between the two cell subsets, while they were extremely similar in patients more than 3 years post-allografting (similarity score= 0.90 vs. 0.61). The differences observed early after SCT were mainly ascribable to Vβ expressing an oligoclonal profile in T_conv but not in T_reg.Interpretation and Conclusions Our data show that the TCR repertoires of T_reg and T_conv are significantly different early post-SCT, while they tend to become identical with full reconstitution. This difference could reflect either a discrepancy in the in vivo reactivity against common antigenic stimulations or be the result of different post-transplant ontogeny.

The diversity of the T-cell receptor (TCR) repertoire is mainly determined by the complementarity determining region 3 (CDR3), which is one of the key players in the context of antigen recognition and major histocompatibility complex restriction. This diversity is due both to extensive rearrangement between V, D and J segments in the TCRα and TCRβ genes and to random junctional nucleotide insertions and deletions, which generate CDR3 regions of different lengths.1,2 After allogeneic hematopoietic stem cell transplantation (SCT) the overall TCR repertoire is characterized by a lower diversity and a markedly skewed pattern. The TCR repertoire may start to normalize at about 6 months after transplant but most patients continue to show an abnormal profile until 2–3 years after grafting, with different kinetics between CD4 and CD8 cells.3–5

Natural regulatory T cells ( T_reg) are a subpopulation of thymus-derived CD4 T cells which constitutively express the interleukin-2 receptor α chain (CD25).6 They were initially described because of their ability to suppress autoreactive T cells in the periphery.7,8 T_reg play a crucial role in the maintenance of peripheral tolerance to self and foreign antigens and modulate susceptibility to autoimmune,9–11 infective12 and neoplastic diseases.13–16

T_reg are characterized by a few other molecules apart from CD25, including cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4), glucocorticoid-induced tumor necrosis factor receptor (GITR) and the Forkhead box P3 (FoxP3) gene product.6 In contrast to murine T_reg, human T_reg are mainly confined within the fraction expressing high levels of CD25,17 which possesses potent immunosuppressive activity and co-expresses FoxP3, whereas the subset expressing intermediate levels of CD25 contains mainly activated T cells with little or no immunosuppressive function.18,19

The role of T_reg in the induction of immunological tolerance after SCT has not been fully elucidated. Administration of these cells has a protective effect in murine models of acute graft-versus-host disease (GVHD), related to inhibition of the early expansion of alloreactive donor T cells.20 Such an activity does not seem to interfere with stem cell engraftment or the graftversus-leukemia effect.21 Also the studies conducted in humans have suggested a correlation with different outcomes between GVHD and circulating T_reg.22–26 Although, once activated, T seem to perform their activity in a non-antigen-specific way,27 they require activation via their TCR to become suppressive.28 A better knowledge of their TCR repertoire in the post-transplant setting would offer relevant information about their ability to react to the massive antigenic stimulation generated in an allogeneic host, which could significantly affect their pattern of reconstitution. Moreover, the requirement for thymic maturation29 could further contribute to altering the repertoire of T_reg, as has been found for conventional T cells.3–5 Although it is well known that the repertoire of antigen specificities of T_reg in mice is as broad as that of naive T cells,30 only two studies have been conducted in humans.31,32 In normal subjects the TCR repertoire of T_reg appears to be similar to that of the CD25 counterpart, thus suggesting the recognition of a similar spectrum of antigens.31 However, no studies focusing on the analysis of the TCR CDR3 repertoire of this cell subpopulation after allografting have been performed so far.

We examined the TCR Vβ repertoire of T_reg in patients who had received allogeneic SCT for chronic myeloid leukemia (CML), by using CDR3 spectratyping. We concentrated our analysis on the CD4CD25 subset expressing high levels of CD25, which contains the vast majority of regulatory T cells,17 as we confirmed by FoxP3 staining. Purified T_reg were compared to the CD4CD25 conventional T (T_conv) counterpart, analyzing the degree of similarity within each Vβ subfamily.

Design and Methods

Patients

We performed our analysis in ten patients (five males and five females) who had undergone allogeneic SCT for chronic myeloid leukemia. As controls, we analyzed samples from seven adult healthy volunteers (three fresh blood samples and four buffy coats from single blood donations). All patients and donors had given informed consent. The study had been approved by the local ethics committee. Clinical and transplant characteristics are summarized in Table 1. The median age of the patients was 41.5 years (range 25–49). Nine had been transplanted in chronic phase and one in accelerated phase. Nine of them had received a myeloablative conditioning regimen with cyclophosphamide and total body irradiation and one had received a reduced intensity conditioning (RIC) regimen consisting of fludarabine and busulphan. Six patients had undergo in vivo T-cell depletion with campath-1H. All patients had received cyclosporine and methotrexate as GVHD prophylaxis, except two (including the one receiving RIC) who were given cyclosporine alone. Four and six patients had developed acute and chronic GVHD, respectively.

Table 1.Patients’ characteristics.

At the time of analysis, the median time after transplant was 37 months (range 4–171). Five out of the ten patients showed a BCR-ABL/ABL ratio ≤0.02%, indicative of molecular remission. Four of the patients had previously received donor lymphocyte infusions (DLI) for molecular relapse and two of them had achieved molecular remission. Samples were collected an average of 33 months (range 11–70) after the last infusion. Only one patient had active limited GVHD (ocular). None of the patients was or had recently been on systemic steroid treatment. One of the patients was affected by herpes zoster infection and had had a cytomegalovirus (CMV) reactivation 3 months before.

Cell separation

CD4CD25 and CD4CD25-subpopulations were isolated from 40 mL of heparinized peripheral blood samples using a two-step procedure. First, CD4 negative isolation (depletion of CD8, CD16, CD19, CD36, CD56, CD66b and glycophorin A) was performed using Rosette Sep enrichment mixture (Stem Cell Technologies, Vancouver, Canada) according to the manufacturer’s specifications. Next, CD25 positive selection was carried out with magnetic beads (Dynal Biotech, Oslo, Norway). The obtained CD4CD25 cells were kept in RPMI media supplemented with 10% AB serum and 1% penicillin and streptomycin. The CD4CD25 cells were left overnight in supplemented media and cells and, after spontaneous detachment from the beads, were exhaustively washed in order to separate only cells expressing high levels of CD25. The numbers of CD4CD25 and CD4CD25 cells isolated from each patient ranged between 0.3 and 1.5×10, and 1.1 and 6.9×10, respectively. TRIzol (Invitrogen, Paisley, UK) was added to both cell fraction pellets which were kept at −80°C until RNA extraction.

Cytofluorimetric analysis

The purity of the cell separation was assessed by staining with specific monoclonal antibodies for CD4 and CD25 and with the corresponding isotype controls (BD Pharmingen, Oxford, UK) after both the CD4 negative and the CD25 positive selection. Intracellular analysis of FoxP3 expression was performed using anti-FoxP3 antibody (clone PCH101-FITC from Ebioscience, San Diego, CA, USA). Samples were acquired using a FACSCalibur and analyzed with Cell-Quest software (BD Biosciences).

CDR3 spectratyping

The RNA was isolated from CD4CD25 and CD4CD25 cells as described elsewhere.33 For each sample, complementary DNA was synthesized from the same starting amount of total RNA (0.7 μg), by using Superscript III reverse transcriptase and random hexamer primers (Invitrogen, Paisley, UK) according to the manufacturer’s instructions. Polymerase chain reaction (PCR) was performed in a volume of 25 μL comprising 1xPCR buffer, 2.5 mM MgCl₂, 1 U AmpliTaq Gold (Applied Biosystems, Foster City, CA, USA), 200 μM deoxyribonucleoside triphosphate and 500 nM of one of 24 TCR Vβ primers combined with 1 Cβ primer conjugated to the fluorescent dye 6-carboxyfluorescein-amino-hexy (6-FAM).34 The sequences of the Vβ and Cβ primers were described previously.35 The PCR conditions were: 95°C for 10 minutes followed by 36 cycles of 94°C for 20 seconds, 55°C for 40 seconds, 72°C for 40 seconds and a final extension of 72°C for 5 minutes. The PCR fragments were then run on an ABI Prism 3100 Genetic Analyzer and data were collected and analyzed by Gene Mapper ID Software version 3.2 (Applied Biosystems, Foster City, CA, USA).

Spectratyping analysis

Conventional analysis of spectratyping profiles was performed according to the following standard approaches: (i) scoring the number of peaks, in order to determine the overall complexity of the repertoire; (ii) analyzing profiles by peak area and shape, in order to establish the degree of skewing and oligoclonality. Moreover, to quantify the proportion of Vβ subfamilies that were similar between T_reg and T_conv, we developed a novel analysis method described in the result section.

The overall complexity of the TCR Vβ repertoire was determined according to the number of discrete peaks detected per Vβ subfamily, each subfamily being graded on a score from 0 to 5. Spectratypes containing five or more peaks were given a score of 5. The overall spectratype complexity score was calculated for both cell subsets in each subject, as the sum of the scores for each subfamily.36

Each spectratype profile was also assessed according to previously described criteria.36 A profile was defined as normal if it showed a Gaussian bell-shaped distribution, with discrete peaks spaced by three nucleotides. Evidence of oligoclonal expansion or skewing was assessed by calculating the relative fluorescence intensity (RI) of each peak (RI = peak area ÷ total peak area). A profile was defined as skewed if: (i) a dominant peak with a RI greater than 35% of the total peak area, corresponding to an oligoclonal profile, was observed; or (ii) two dominant peaks were present and the RI of each peak was greater than 25% of total peak area; or (iii) there were multiple peaks differing from a Gaussian pattern and the RI of the dominant peaks was greater than 25% of total peak area. The percentages of skewed and oligoclonal Vβ among the total number of Vβ analyzed were calculated.

Statistical analysis

The Student’s t test was used to assess differences in the complexity score, percentage of skewed or oligoclonal Vβ subfamilies and the similarity score between different groups of subjects. Spearman’s correlation coefficient was determined to evaluate the correlation between the similarity score and time after SCT. All quoted p-values are two-sided with values <0.05 considered statistically significant.

Results

Characterization of the selected cell subpopulations

After the first step of separation, the proportion of CD4CD25 cells in all samples was >93% (range 93–99%) (Figure 1A). The analysis of the CD4CD25 subset, performed after the immunomagnetic separation, showed a high proportion of CD4 cells expressing high levels of CD25 between 85% and 92% (Figure 1B). In order to further determine the nature of the selected cell subpopulations, intracellular analysis of FoxP3 expression was performed. The expression of FoxP3 was almost exclusively confined to the isolated CD4CD25 subset (Figure 1C and 1D), thus conclusively confirming that this subset belongs to the regulatory lineage. The percentages of FoxP3 cells within the T_reg and T_conv subpopulations were 81.8±11% and 3.7±2%, respectively.

Figure 1.FoxP3 is selectively expressed on purified CD4⁺CD25⁺ cells. A. Representative profile of the CD4⁺CD25⁻ subpopulation obtained after CD4⁺ negative isolation using the Rosette Sep enrichment mixture (depletion of CD8, CD16, CD19, CD36, CD56, CD66b and glycophorin A); B. Representative profile of CD4⁺ CD25⁺ purified cells after CD25 positive selection with magnetic beads. The plots in the lower panel show the intracellular expression profile of FoxP3 in the CD4CD25⁻ (C) and CD4⁺CD25⁺ (D) subsets (solid lines: FoxP3; dotted line: isotype control).

Treg and Tconv exhibit a similar overall complexity and skewing in the TCR repertoire

The mean number of Vβ profiles available per sample was 21 (range, 16–24) for both T_reg and T_conv in patients as well as in controls. Figure 2 shows representative spectratyping profiles of one of the patients, whilst Table 2 summarizes the data of the TCR repertoire analysis for each patient. The overall spectratype complexity score was not different between T_reg and T_conv in both the patients (mean 106.5 vs. 105.5) and controls (mean 108.5 vs. 112.5). In the latter, T_reg and T_conv showed TCR repertoires with a predominantly Gaussian distribution and shared common patterns also in terms of percentages of skewed and oligoclonal Vβ, thus confirming previous data.31,32

Figure 2.Representative CDR3 profiles of the 24 Vβ subfamilies in the T_reg and T_conv subpopulations. The investigation was carried out 6 months after SCT in a patient who had an active herpes zoster infection and had previously had CMV reactivation. The similarity score was 0.64, with 14 Vβ coincident between the two cell subpopulations out of 22 Vβ pairs analyzed (Vβ 23 and 24 were excluded as not detected in both subsets). Vβ 1, 5, 15, 16 and 18 expressed an oligoclonal profile only in the T_conv subset; Vβ 20 was oligoclonal only in the T_reg subpopulation and Vβ 12 and 14 displayed a skewed but non-oligoclonal profile in only one of the cell subsets.

Table 2.Summary of T-cell receptor repertoire pattern and transplanted related features of the analyzed patients.

It is well documented that the CD4 TCR repertoire is not fully re-established until at least 2–3 years following allogeneic SCT.3–5 Accordingly, we observed a significantly higher percentage of skewed Vβ in both cell subpopulations in patients whose SCT had been performed less than 3 years previously, as compared to those in whom more than 3 years had passed since their SCT (mean 41% vs. 22%, p<0.05, in T_reg and 45% vs. 15%, p<0.05, in T_conv) (Figure 3A). Moreover, the percentage of oligoclonal Vβ tended to be higher in patients closer to the SCT in both cell subpopulations (15% vs. 5%, p=ns, in T_reg and 27% vs. 4%, p<0.05, in T_conv) (Figure 3B). However, when we compared the two cell subsets of each subject following this approach, we did not identify any significant difference in the percentage of skewed and oligoclonal Vβ between T_reg and T_conv, regardless of time elapsed since the SCT. Nevertheless, it is worth noting that in patients who had undergone SCT less than 3 years previously, the percentage of oligoclonal Vβ tended to be higher in T_conv than in T_reg (27% vs. 15%, p=ns). When we looked at the prevalence of skewed or oligoclonal profiles in specific Vβ subfamilies, we did not find any predominant pattern, with the exception of a higher frequency of skewing in Vβ 10 and 19, which are known to be encoded by pseudogenes.

Figure 3.The percentages of both skewed and oligoclonal Vβ subfamilies are higher in patients with a shorter time between SCT and the analysis. The figures report the percentages of skewed (A) and oligoclonal (B) Vβ in the two cell subpopulations, detected in patients less (R) or more (r) than 3 years after transplantation. Horizontal bars indicate the mean for each group.

The “similarity score” detects different TCR repertoires between Treg and Tconv

The conventional methods of analysis evaluate the overall complexity of the repertoire and are, therefore, inadequate for establishing a direct comparison within each Vβ subfamily between different cell subpopulations. To address this issue, we specifically compared the spectratyping profiles of T_reg and T_conv of each subject within each of the 24 Vβ subfamilies. In order to do this, we developed a new score to express similarity. Within each Vβ subfamily, the spectratype profiles of T_reg and T_conv were considered to be similar if: (i) both showed a Gaussian bell-shaped distribution; or (ii) both showed the same skewed pattern, according to the criteria previously defined; or (iii) an oligoclonal peak was observed in the same position in both cell subpopulations. The similarity score was calculated for each subject as the ratio between the number of Vβ with coincident profiles between the two cell subpopulations and the total number of Vβ available in both cell subsets. For example, the patient whose repertoire is represented in Figure 2, who showed 14 Vβ with coincident profiles between the two cell subpopulations out of 22 Vβ available in both cell subsets, had a ratio of 14/22, equivalent to a score of 0.64.

Using this approach, we detected a positive correlation between the similarity score and time elapsed since the SCT (Pearson’s correlation coefficient=0.65) (Figure 4A). A higher score was observed in patients more than 3 years after allografting, as compared to those investigated less than 3 years after SCT (mean 0.90 vs. 0.61, p=0.01) (Figure 4B). Remarkably, the similarity score in patients more 3 years after SCT was close to that detected in fresh blood samples of normal volunteers (mean 0.87).

Figure 4.The TCR repertoires of T_reg and T_conv differ early after transplant while they tend to become identical more than 3 years post-SCT. A. The similarity score was calculated in each subject (?) as a ratio between the number of Vβ ?subfamilies with a profile coincident between the two cell subsets and the total number of subfamilies available in both cell subpopulations. Pearson’s correlation coefficient, calculated for the relation between similarity score and time after SCT, was 0.65. B. Similarity scores in patients less (?) or more (?) than 3 years after transplantation. Horizontal bars indicate the mean for each group.

We found four different patterns among the mismatching Vβ profiles: (i) an oligoclonal profile only in the T_reg; (ii) an oligoclonal profile only in the T_conv subpopulation; (iii) a skewed but not oligoclonal profile only in the T_reg, and (iv) a skewed but not oligoclonal profile only in the T_conv subset. Noticeably, in patients investigated less than 3 years after SCT the mismatches were very often ascribable to the detection in the same Vβ subfamily of an oligoclonal profile in the T_conv but not in the T_reg subpopulation, while the opposite condition was much less frequent (mean 52% vs. 23%, p=0.08). When considering only those patients whose SCT had been performed less than 2 years previously, this trend became statistically significant (mean 56% vs. 12%, p=0.01). Moreover, the presence of oligoclonal profiles confined to the T_conv subset was almost exclusively detected in patients in whom less than 3 years had elapsed since SCT (mean 52% vs. 5%, p=0.002) (Figure 5). No significant differences were shown in the group of patients who had undergone SCT more than 3 years previously, in whom the mismatches were equally distributed between the two cell subsets and more often due to a skewed profile in only one of them (data not shown). A clear example of selective oligoclonal profiles in the T_conv subpopulation early after SCT is represented in Figure 2. This patient was investigated 6 months post-SCT and had an active herpes zoster infection and a previous CMV reactivation. His repertoire was characterized by five Vβ subfamilies expressing an oligoclonal profile only in the T_conv subset.

Figure 5.Oligoclonal profiles selectively confined to T_conv are more frequent in patients with a shorter time between SCT and the analysis. The frequency of oligoclonal profiles confined to either the T_reg (?) or the T_conv (g) subpopulation in patients less or more than 3 years after SCT is represented. The y axis represents the percentage of Vβ showing an oligoclonal profile confined to each cell subpopulation among the overall non-similar Vβ subfamilies. Horizontal bars indicate the mean for each group.

Role of disease status and transplant-related factors

Patients in molecular remission at the time of analysis showed a lower percentage of both oligoclonal and skewed Vβ (data not shown). The percentages of skewed and oligoclonal Vβ in T_reg were 35% and 14% vs. 28% and 6% in the non-remission and remission group, respectively. The percentages in T_conv were 40% and 24% vs. 20% and 8%, respectively. However, none of these differences reached a statistical significance. We did not find any difference in the similarity score between patients in remission or not.

Furthermore, the percentages of oligoclonal and skewed Vβ as well as the degree of similarity between the two cell subpopulations were not influenced by other factors such as stem cell source, donor type, HLA matching, in vivo T-cell depletion and previous acute or chronic GVHD (data non shown).

Effect of DLI

Previous treatment with DLI was associated with an increase in the amplitude of the TCR repertoire, defined on the basis of a lower percentage of oligoclonal and skewed Vβ (data not shown). In the T_reg subpopulation, the percentages of skewed Vβ in the DLI and non-DLI group were 20% and 39% (p=0.02), respectively, whilst the percentages of oligoclonal Vβ in the treated and untreated groups were 7% and 12% (p=ns). In T_conv, the percentages of skewed Vβ in patients who did or did not receive DLI were 14% and 40% (p=ns), respectively, whilst the percentages of oligoclonal Vβ were 3% and 24% (p<0.05). Previous treatment with DLI was strongly associated with a higher similarity score between the two cell subsets (p=0.005).

Discussion

In the last years, T_reg have progressively emerged as an essential component in the control and maintenance of peripheral tolerance. One of their most peculiar functional features is the requirement of activation via TCR to become suppressive,28 notwithstanding their ability to act in a non-antigen-specific way.27 As recent studies provided evidence for their specificity in the modulation of transplant alloreactivity,37–39 a deeper knowledge of their TCR repertoire pattern in the post-transplant setting would offer relevant information about their ability to react to the massive antigenic stimulation generated in an allogeneic host. It is known that this cell subpopulation presents a TCR repertoire profile substantially similar to that of the CD25 counterpart in healthy subjects.31,32 Since the overall T-cell repertoire after allogeneic SCT is greatly impaired3–5 and naturally occurring T_reg require maturation in the thymus,29 their TCR usage may be equally abnormal. We analyzed the CDR3 Vβ repertoire of purified CD4CD25 cells expressing high levels of CD25 in patients who had undergone allografting for chronic myeloid leukemia. This cell subpopulation contains the vast majority of cells with regulatory activity,17 as we confirmed by FoxP3 staining. Our analysis focused on the overall complexity of the repertoire and on the degree of similarity to the T_conv counterpart. Our study is based on the examination of a remarkably large number of Vβ profiles in each cell subpopulation (mean 21; range 16–24), especially considering the profound lymphopenia observed after allogeneic SCT and the exiguity of the T_reg subset. Despite the limited number of patients analyzed, this study is the first CDR3 spectratyping analysis ever performed on the T_reg subset after transplantation. All previous studies have been performed either on unfractionated peripheral blood mononuclear cells4,5,40,41 or on the CD4 and CD8 fractions.3 Moreover, our group of patients was almost completely homogeneous in terms of disease type and conditioning regimen. None of the other transplant-related factors and post-transplant variables, apart from DLI administration, turned out to have a significant influence on the TCR repertoires of the two cell subsets in terms of both overall skewing and degree of similarity between T_reg and T_conv.

The conventional systems for spectratyping analysis, based on the determination of the overall complexity and skewing of a cell subpopulation, showed essentially similar TCR patterns between T_reg and T_conv. However, these methods are inadequate to establish a direct comparison within each Vβ subfamily between two cell subsets. We, therefore, developed a new score system, based on a quantification of the proportion of Vβ subfamilies with coincident profiles between T_reg and T_conv. This analysis revealed significant differences between the two repertoires within the first 3 years after SCT. The differences were mainly due to the frequent presence of Vβ expressing an oligoclonal profile in the T_conv but not in the T_reg subpopulation. The markedly skewed profile detected in the T_conv subpopulation resembled the features described early after allografting in the CD4 subset, in which the repertoire reconstitutes within 3 years.3–5 In fact, after 3 years, the TCR Vβ profiles of the two cell subsets tended to become identical, paralleling the T_conv subset recovery. The much higher degree of TCR similarity between the two cell subsets observed in patients more than 3 years after SCT as compared to those with a more recent transplant could reflect the kinetics of reconstitution of the whole T-cell compartment. On the other hand, the marked differences between the T_conv and T_reg repertoires described in recipients early after SCT seems to be ascribable to a reduced degree of skewing confined to the T_reg subset. The repertoire amplitude and the degree of similarity between the two cell subsets were increased in patients treated with DLI, thus corroborating its possible contribution to the re-establishment of the TCR repertoire.42,43

Although the intrinsic in vitro anergic features of T_reg 44 are fully consistent with the observed discrepancy, these cells have been shown to proliferate actively in vivo after antigenic stimulation,45,46 especially in a lymphopenic environment.22,47 However, from our data it appears that, in the presence of a massive antigenic stimulation, such as the one generated in an allogeneic host, there is selective expansion of T_conv rather than T_reg, producing oligoclonal profiles in their repertoire. On the other hand, the reduced degree of oligoclonality observed in the TCR repertoire of T_reg, when compared to T_conv, does not necessarily mean that T_reg are impaired or that they have not been activated. In fact, the detection of a less restricted repertoire could be an expression of their ability to exert a suppressive effect in a non-antigen-specific manner,27 notwithstanding the requirement of activation via TCR.28 It is worth noting that intereleukin-248 as well as Toll-like receptor 2 ligands of bacterial or fungal origin,49 both of which can be largely present after SCT, are also important mediators of T_reg activation. Another factor that might explain the reduced similarity between the TCR repertoires of the two subpopulations early after allografting is a different origin. T_reg are positively selected in the thymus as a consequence of an interaction between a self-peptide and their TCR.29 The detection of oligoclonal patterns, suggestive of antigen-driven proliferations, in T_conv but not in T_reg could imply that the TCR repertoire is shaped by a peripheral antigen-induced expansion in T_conv and by a broader thymic-dependent selection in T_reg.

Murine models20 as well as clinical studies22 suggest that T_reg of donor origin are primarily involved in transplantation tolerance. Interestingly, on one occasion we had the possibility to compare a patient’s TCR profile 52 months after SCT with that of her donor (data not shown). We observed that the patient’s repertoire was fully reconstituted, with a Gaussian profile in all the Vβ subfamilies, except for an oligoclonal pattern in Vβ 20. The donor’s repertoire was nearly identical, including the Vβ 20, which showed the same oligoclonal profile. Noteworthy, the patient had received DLI 11 months before, which had produced a cytogenetic remission.

In conclusion, our data show that the TCR repertoires of T_reg and T_conv exhibit significant differences early after SCT, which are mainly ascribable to V? subfamilies expressing an oligoclonal profile in the T_conv but not in the T_reg subpopulation. The TCR V? profiles of the two subsets tend to become identical with full reconstitution. These different patterns may reflect a different in vivo reactivity against common antigenic stimulations or result from a disparity in post-transplant T-cell ontogeny.

Footnotes

• Authors Contributions CF performed the research, analyzed the data and wrote the manuscript; EN performed the research and provided intellectual input and critical feedback to the manuscript; ML provided intellectual input and critical feedback to the manuscript; FD designed the research, supervised all aspects of the study and reviewed the manuscript.
• Conflicts of Interest The authors reported no potential conflicts of interest.
• Funding: This work was supported by the Leukaemia Research Fund. Claudio Fozza is the recipient of the fellowship “Master and back TS 07” offered by the Region of Sardinia.

• Received September 15, 2006.
• Accepted January 9, 2007.