Prevalence of Newly Generated Naive Regulatory T Cells (T_reg) Is Critical for T_reg Suppressive Function and Determines T_reg Dysfunction in Multiple Sclerosis¹

Growing evidence suggests a role for natural immunoregulatory T cells of the CD4⁺CD25⁺FOXP3⁺ phenotype (T_reg)³ in the prevention of autoimmunity. Numerous studies have reported numeric or functional deficiencies of T_reg in various human autoimmune diseases including inflammatory demyelinating disorders of the CNS (Refs. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, Restored suppressive capacities of regulatory T cells under therapy with IFN-β and glatiramer acetate in patients with multiple sclerosis, submitted for publication). We and others have shown that T_reg derived from patients with relapsing remitting multiple sclerosis (RRMS) display a functional impairment because they inhibit myelin-specific and Ag-nonspecific T cell proliferation less potently as compared with healthy control donors (Refs. 8, 9, 10 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication). The mechanisms of this T_reg defect are unclear. In particular, T_reg frequencies in the peripheral blood of multiple sclerosis (MS) patients are unaltered as is their susceptibility to undergoing CD95 ligand-mediated apoptosis (Refs. 8 , 10 , and 11 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication).

T_reg mature in the thymus and the majority of the cells rapidly adopt a memory CD45RO phenotype (12) when released into the periphery. However, there is recent evidence that small numbers of T cells with a naive CD4⁺CD25⁺CD45RA⁺ surface profile and immunosuppressive properties are detectable within the peripheral T_reg pool (13, 14, 15, 16). Unlike their Ag-primed memory counterpart, this naive T_reg subset may comprise de novo generated cells that have been recently released from the thymus and have not yet experienced Ag contact. Kimmig et al. have demonstrated that surface expression of CD31 (PECAM-1) on naive CD4⁺ T cells distinguishes recent thymic emigrants (RTEs) from peripherally expanded naive Th cells (RTE-Th) (17). CD4⁺CD45RA⁺RO⁻CD31⁺ RTEs, unlike CD45RA⁺ Th cells lacking CD31, contain high amount of TCR excision circles (TRECs). TRECs are generated as a by-product of the TCR rearrangement process in the thymus and are enriched in newly generated T cells (18). Thus, phenotypic and molecular features define CD4⁺CD45RA⁺CD31⁺ T cells as RTEs and CD4⁺CD45RA⁺CD31⁻ T cells as peripherally expanded naive Th cells. As thymic function declines with age the release of RTEs also decreases. This loss of de novo generated Th cells is compensated by peripheral postthymic expansion, resulting in an increase of CD4⁺CD45RA⁺CD31⁻ T cell numbers with age (17, 19).

This age-related change in the composition of the CD4⁺ T cell pool may also affect the homeostatic balance and, eventually, the functional properties of T_reg. In concordance with this assumption, an age-dependent decline of T_reg suppressive potencies has been recently reported in a pilot study (20). Our own previous work suggests that the thymic export function is impaired in MS as levels of TREC-expressing peripheral T cells are significantly decreased in patients with clinically active RRMS in the absence of an accelerated T cell turnover (21). Therefore, we hypothesized that an altered T_reg homeostasis may contribute to the suppressive deficiency associated with MS.

To elucidate a potential imbalance in T_reg homeostasis in MS, we analyzed the role of naive T_reg that coexpress CD31 and circulate in the periphery as RTE. We measured the numbers of RTE-T_reg within the peripheral blood of patients with RRMS and healthy persons by multicolor flow cytometry analysis and determined their contribution to the suppressive function of the entire T_reg population by in vitro proliferation assays. We also performed CDR3 spectratype analysis to screen T_reg derived from both cohorts for differences in the TCR repertoire.

Blood specimens were obtained from 40 RRMS patients (mean age 33.9 years, range 14–62 years) as well as from 49 healthy controls (HCs) (mean age 36.7 years, range 24–85 years). All patients had definite MS according to McDonald’s or Poser criteria (22, 23) and had experienced an average number of 1.4 previous relapses (range 0–3). Disease duration ranged between 1 and 5 years (median 2). The mean Expanded Disability Status Scale score was 1.7 (range 1–2). All patients had a clinically active disease and had not yet received treatment with corticosteroids or immunomodulatory agents. The protocol was approved by the University of Heidelberg (Heidelberg, Germany) ethics committee and all individuals gave written informed consent.

The mAbs used for flow cytometry were obtained from BD Pharmingen (anti-human CD4, CD45RO, CD45RA, and CD31), Miltenyi Biotech (anti-human CD25), and eBioscience (anti-human FOXP3). FOXP3 staining was performed according to the manufacturer’s protocol. For six-color flow cytometric analysis, cells were first stained with a surface mAb specific for CD4, CD25, CD45RO, CD45RA, or CD31 followed by FOXP3 intracellular staining. FACS acquisition was performed immediately with a FACSCanto cytometer and analyzed with FACSDiva software (BD Biosciences). For the quantification of T_reg and Th subsets within peripheral blood, stained PBMCs were first gated on CD4⁺ cells and than analyzed for the expression of FOXP3 and CD25. Because only CD4⁺ T cells expressing the FOXP3⁺ gene product scurfin exhibit suppressive function (24), CD4⁺CD25⁺FOXP3⁺ cells were defined as T_reg and CD4⁺FOXP3⁻ cells as Th (Fig. 1,A). T_reg and Th cells were further analyzed for their CD45RA/CD45RO surface expression to identify CD45RA⁻CD45RO⁺ memory and CD45RA⁺CD45RO⁻ naive subsets (Fig. 1,B). RTE-T_reg and RTE-Th were identified by coexpression of the CD31 molecule within the naive T_reg and Th subsets (Fig. 1 C).

PBMC were isolated from 50 ml of peripheral blood by density gradient centrifugation using Ficoll/Hypaque (Biochrom). CD4⁺ T cells were isolated from PBMCs by using a negative CD4⁺ T cell isolation kit (Dynal Biotech). CD4⁺CD25^high T_reg and CD4⁺CD25^low/CD4⁺CD25^int effector T cells (T_eff) were isolated from the pure untouched CD4⁺ T cells using CD25 magnetic beads (Dynal Biotech). Magnetic beads were removed from T_reg using Detachabead CD4 solution (Dynal Biotech). T_reg isolated from the PBMC of nine HCs and seven patients were further separated according to their CD31 expression in CD31⁺ T_reg and CD31⁻ T_reg by the use of a FITC-labeled anti-CD31 mAb and an anti-FITC MultiSort kit (Miltenyi Biotec). Immune magnetic separation constantly yielded highly pure T_reg subsets with >90% CD4⁺CD25^high cells as previously described (Ref. 10 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication). The purity and phenotype of random samples were further tested by intracellular FOXP3 staining that revealed >90% FOXP3⁺ cells in preparations of total T_reg (n = 20, mean 94.7%, range 92.2–98.5% FOXP3⁺) as well as in CD31⁺ T_reg (n = 7, mean 94.5%, range 90.9–96.1% FOXP3⁺) and CD31⁻ T_reg (n = 11, mean 93.4%, range 93.0–96.4% FOXP3⁺). Preparations of T_reg and T_reg subsets isolated from patients and HCs showed similar CD25 and FOXP3 expressions (data not shown). The vast majority of isolated CD31⁺ T_reg expressed CD31 on their surface (mean 96.3%, range 92.4–99.5%), whereas only 4.6% (range 0.2–7.7%) of the CD31-depleted T_reg remained CD31⁺. CD31⁻ T_reg exhibited a memory phenotype (mean 82.5%, range 78.3–92.0% CD45RA⁻CD45RO⁺). In contrast, CD31⁺ T_reg were predominantly naive (mean 68.5%, range 61.1–75.4% CD45RA⁺CD45RO⁻) with <10% memory cells (mean 9.7%, range 5.3–12.7% CD45RA⁻CD45RO⁺) (representative examples showing purity and phenotype of isolated T_reg subsets are shown in Fig. 2).

Freshly isolated T_eff (10⁵) were incubated in 96-well plates (Nunc) in 200 μl of culture medium alone or in 4:1 coculture with 2.5 × 10⁴ total T_reg or T_reg subsets. For TCR stimulation, soluble anti-CD3 (1 μg/ml) and anti-CD28 mAbs (1 μg/ml) were added to the culture medium. After 4 days at 37°C in 5% CO₂, 1 μCi of [³H]thymidine per well was added for an additional 16 h. Proliferation was measured using a scintillation counter. Inhibition rate (percentage) of T_reg in coculture experiments was defined as [1 − [³H]thymidine uptake (cpm) within a T_reg plus T_eff coculture ÷ cpm of T_eff alone] × 100. Each experiment was performed in triplicate. The CD3 and CD28 mAbs used for cell culture experiments were purified from hybridoma supernatants by protein A affinity purification as described (25). Cell culture medium and mAbs were endotoxin free (<10 pg/ml) as assessed by a Limulus assay (Sigma-Aldrich).

Genomic DNA was extracted from 2 × 10⁵ T_reg or T_reg subsets by the use of a DNAzol solution (Invitrogen Life Technologies) according to the manufacturer’s recommendations. The numbers of TRECs in T cell subsets were determined by real time PCR as described previously (21). Results were expressed as TRECs per 10⁶ cells.

We performed CDR3 spectratyping to determine the TCR Vβ repertoire in patient and HC T_reg and T_eff. Total RNA was isolated from 2 × 10⁵ T_reg or 10⁶ T_eff using Absolutely RNA Microprep kit (Stratagene) according to the manufacturer’s protocol and converted to cDNA by a SuperScript first-strand synthesis kit (Invitrogen). Amplification of the entire group of 24 TCR V subfamilies was conducted with 24 Vβ-specific primers and one primer recognizing the C region (26, 27). One hundred nanograms of cDNA from each T_eff and T_reg sample was amplified in eight parallel PCRs each containing a set of three different Vβ primers respectively and the C-primer, which was 6-FAM-labeled for CDR3 spectratyping. Size distribution of fluorescent PCR products was determined by laser-induced capillary electrophoresis using an automated DNA analyzer (A310; Applied Biosystems) and GeneScan software (Applied Biosystems). Normal transcript size distribution consists of eight peaks for each V subfamily (28). Complexity within a V subfamily was determined by counting the number of peaks per subfamily. The complexity score (CS) was calculated as the summation of the scores of all 24 subfamilies. The maximum score would be 192.

To determine whether differences and correlations in cell counts, inhibitory capacities, TREC levels, and TCR diversities were statistically significant, we performed the Student’s t test using a two-tailed distribution with unpaired samples and calculated the Pearson’s correlation coefficient. p < 0.05 was considered significant.

To determine the number of naive helper T cells that have just entered the circulation as RTEs (RTE-Th), we performed multicolor flow cytometric staining of PBMCs obtained from 28 MS patients and 28 age-matched HCs. RTE-Th frequencies were calculated as percentage of CD45RA⁺CD45RO⁻CD31⁺ cells within total CD4⁺FOXP3⁻ T cells (Fig. 1). As expected, the proportions of RTE-Th within CD4⁺FOXP3⁻ cells declined with age in HCs (Fig. 3,A). In individuals <30 years of age RTE-Th constituted approximately one-third of CD4⁺ T cells (mean 28.3%, range 20.9–37.5%). RTE-Th numbers were significantly lower in persons aged 30–45 years (mean 22.0%, range 8.9–33.0%, p = 0.048) and lowest in persons >45 years of age (mean 16.6%, range 6.5–33.4%, p = 0.001). RTE-Th frequencies in patients were significantly lower than those in HCs (MS patients: mean 18.1%, range 3.5–34.6%; HCs: mean 22.7%, range 6.5–37.5%; p = 0.025, Fig. 3,A). However, when comparing patients and HCs aged >45 years, the difference was <1% (MS patients: mean 17.1%, range 12.1–29.4%; HCs, mean 16.6%, range 6.5–33.4%; p = 0.905). Similarly, the entire naive Th subset (CD45RA⁺CD45RO⁻ of CD4⁺FOXP3⁻ cells) decreased with age in healthy donors but was constantly low in MS patients (data not shown). Conversely, the proportions of memory cells within CD4⁺ T cells rise with age in HCs (<30 years: mean 38.4%, range 14.6–68.2%; >45 years: mean 68.4%, range 51.0–88.3%; p = 0.002) and are enhanced in MS patients (MS: mean 59.4%, range 39.8–84.3%; HCs: mean 51.0%, range 14.6–88.5%; p = 0.050; Fig. 3 B). CD31 expression on memory CD4⁺ T cells was consistently <10% and did not differ between MS patients and HCs (data not shown). In concordance with numerous studies and our own previous observations, the distribution of total CD4⁺ T cells was independent of age and unaltered within peripheral blood samples of MS patients (data not shown).

Next, we screened peripheral blood samples from 28 MS patients and 28 age-matched HCs for the presence of total T_reg and T_reg subsets by multicolor flow cytometry. Proportions of RTE-T_reg were calculated as percentages of CD45RA⁺CD45RO⁻CD31⁺ cells within total CD4⁺FOXP3⁺CD25⁺ cells (Fig. 1). With regard to total T_reg, we observed slightly reduced percentages of CD25⁺FOXP3⁺ cells among total CD4⁺ T cells in MS patients that were not statistically significant (MS: mean 5.0%, range 1.6–9.2%; HCs: mean 5.7%, range 2.6–12.9%; p = 0.107; not depicted). T_reg frequencies did not change with age in both patients and HCs (data not shown). In contrast, RTE-T_reg frequencies within total FOXP3⁺ cells declined with age in HCs (all: mean 4.8%, range 0.3–19.5%; <30 years of age: mean 7.2%, range 1.5–19.5%; 30–45 years of age: mean 3.1%, range 0.3–6.1%; >45 years of age: mean 3.2%, range 0.7–6.6%; Fig. 3,C). In patients the proportions of RTE-T_reg were significantly reduced and not age-associated (all: mean 2.7%, range 0.0–9.8%; <30 years of age: mean 2.6%, range 0.0–5.7%; 30–45 years of ages: mean 3.1%, range 0.4–9.8%; >45 years of age: mean 2.4%, range 0.4–5.9%). Differences in RTE-T_reg numbers between MS and HCs were statistically significant for younger persons but were no more detectable in older subjects (<30 years of age: p = 0.034; 31–45 years of age: p = 0.799; >45 years of age: p = 0.348; Fig. 3,C). The age-dependent loss of RTE-T_reg in HCs is paralleled by a moderate increase of naive T_reg lacking CD31 expression (<30 years of age: mean 5.7% of total FOXP3⁺ cells, range 1.9–23.2%; >45 years of age: mean 8.6%, range 2.5–14.7%, p = 0.115; not depicted). Percentages of CD45RA⁺CD45RO⁻CD31⁻ within FOXP3⁺ cells were also slightly higher in younger MS patients as compared with HCs (MS patients <30 years of age: mean 9.7%, range 0.7–19.4%; p = 0.107; not depicted). Proportions of CD45RA⁻CD45RO⁺ memory T_reg increased with age in HCs (<30 years of age: mean 60.6%, range 36.4–72.3%; >45 years of age: mean 78.6%, range 64.5–88.5%; p = 0.013; Fig. 3 D) and were enhanced in younger MS patients. CD31 expression on memory T_reg was consistently low in blood samples of both MS patients and HCs (data not shown).

Because TRECs are highly enriched in RTEs (17) as a result of the accomplished TCR gene rearrangement process without further dilution due to proliferation (18), we studied the intracellular levels of TRECs in T_reg and T_eff subsets isolated from four HCs (mean age 29.0 years, range 26–41 years of age).

In concordance with our previous findings (10) total T_reg had a very low TREC content (mean 1.6 × 10³/10⁶ cells, range 0.8–2.5 × 10³/10⁶ cells), suggesting that the T_reg pool in the periphery mainly consists of rapidly dividing memory T_reg (Fig. 4). As expected, TRECs were virtually absent in the purified CD31⁻ T_reg subset predominantly exhibiting a memory phenotype (mean 0.1 × 10³/10⁶ cells, range 0.0–0.1 × 10³/10⁶ cells). In contrast, mean frequencies of TRECs in CD31⁺ T_reg were substantially higher as compared with the entire T_reg population (9.5 × 10³/10⁶ cells, range 7.1–11.6 × 10³/10⁶ cells; p < 0.001), reflecting the more naive status of this distinct T_reg subset.

We evaluated the data of proliferation assays performed with T_reg and T_eff isolated from a total of 89 HC donors and 56 patients. Some were from the present study (HCs: n = 18; MS: n = 14), others from subjects included in earlier studies (HCs: n = 71; MS: n = 42). Average ages of the entire cohort were 33.6 years (range 19–62 years; MS) and 32.4 years (range 19–85 years; HCs). All assays were conducted as a 1:4 coculture of 2.5 × 10⁴ T_reg with 10⁵ TCR-stimulated autologous T_eff and, similarly as in recent reports (Refs. 8, 9, 10 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication), revealed mean T_reg-mediated suppressions of 38.4% in MS and 52.0% in HCs (MS: range 0.0–86.4%; HCs: range 4.1–98.4%; Fig. 5,B). Interestingly, inhibitory capacities in younger HCs were higher as compared with persons of higher age (<30 years of age: mean 58.4%, range 10.7–98.4%; 30–45 years of age: mean 46.0%, range 3.9–89.5%, p = 0.026; >45 years of age: mean 40.6%, range 21.2–59.6%; p = 0.018; Fig. 5,B). Such a decline of suppressive capacity was not seen in patients, who showed the same low suppressive capacity in all age groups (<30 years of age: mean 36.6%, range 0.0–86.4%; 30–45 years of age: mean 41.6%, range 0.0–81.9%; >45 years of age: mean 41.8%, range 8.5–63.7%; Fig. 5, A and B). Differences in T_reg-mediated suppression between MS patients and HCs were statistically significant for younger persons but were no more detectable in older subjects (<30 years of age: p = 0.001; 30–45 years of age: p = 0.369; >45 years of age: p = 0.520). In contrast with the T_reg-mediated suppression rates, the T_eff proliferative responses were not age dependent in HCs and did not differ between MS patients and HCs (Fig. 5, C and D).

To determine whether the T_reg suppressive function correlates with the percentage of cells expressing CD31, we obtained T_reg and T_reg depleted of CD31⁺ cells from seven patients (mean age 30.7 years) and seven HCs (mean age 28.7 years). Total T_reg and CD31⁻ T_reg were tested in parallel in primary in vitro proliferation assays by 1:4 cocultures with 10⁵ TCR-stimulated autologous T_eff. As shown above for the entire cohort, again total T_reg derived from patients exhibited a reduced inhibitory activity in vitro when compared with T_reg isolated from HCs (MS: mean 37.5%, range 19.3–56.1%; HCs, mean 60.0%, range 34.5–83.0%; p = 0.011, Fig. 6,A). CD31⁻ T_reg obtained from HCs exhibited a less potent inhibitory capacity toward activated T_eff with 42.6% (range 8.7–71.8%) suppression only (Fig. 6 A). This inhibition rate was comparable to patient-derived CD31⁻ T_reg, which inhibited T_eff proliferation to the same extent compared with total T_reg (38.7%, range 25.9–59.0%; p = 0.837).

Thus, the difference in inhibitory potencies of patient and HC T_reg cells detectable when using total T_reg in the coculture experiments was neutralized by the depletion of CD31⁺ cells.

From 16 subjects (MS: n = 6; HCs: n = 10) parallel data from both primary in vitro proliferation assays and multicolor flow cytometry were available. When the percentage of inhibition was plotted against the percentage of RTE-T_reg a clear correlation between T_reg suppressive capacities and the prevalence of RTE-T_reg was obtained (r = 0.553; Fig. 6 B).

To analyze, whether an altered de novo generation of T_reg also influences the T_reg clonal distribution, we determined the CDR3 spectratype CS to compare the TCR diversity of T_reg from MS patients and HCs (Fig. 7). The mean CS of T_reg obtained from 22 HCs was 161 (range 138–185). CDR3 spectratyping of patient-T_reg (n = 24) resulted in lower CS values (mean 151, range 101–173). The difference of complexity scores between both study cohorts was statistically significant, with p = 0.041.

In previous work, we and others have demonstrated that T_reg from MS patients are functionally impaired (Refs. 8, 9, 10 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication). In the present study we provide evidence that the diminished suppressive T_reg potencies in MS are related to disequilibrium in the homeostatic composition of circulating T_reg and most likely result from an altered thymic release of newly formed T cells into the periphery. Because T_reg frequencies in peripheral blood are normal in MS patients as reported in a number of independent studies (Refs. 8, 9, 10 and M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication) and the susceptibility of patient-derived T_reg to undergo apoptotic cell death is also unaltered (11), we wondered whether any imbalance in T cell neogenesis contributes to the T_reg defect detectable in MS patients. The rationale to assess this issue in more detail was based on our earlier observation that the thymic export function and the ability to maintain T cell homeostasis is disturbed in MS (21). This may also affect T_reg, which is generated in the thymus as a naturally suppressive T cell subset. Kimmig et al. showed that only naive CD4⁺CD45RA⁺ cells coexpressing CD31 harbor extrachromosomal deletion circles as traceable molecular markers of newly generated T lymphocytes and phenotypically define RTE (17). The prevalence of CD4⁺CD45RA⁺CD31⁺ cells declines with age, reflecting the thymic-dependent neogenesis of this subset (17, 19). In the present work we confirmed these findings, because the proportion of CD45RA⁺CD45RO⁻CD31⁺ cells within the peripheral CD4⁺FOXP3⁻ Th pool decreased with age whereas pure CD4⁺ T cell numbers remained at constant levels. Moreover, the frequencies of CD4⁺FOXP3⁻CD45RA⁺CD45RO⁻CD31⁺ cells were significantly decreased in peripheral blood samples of MS patients, thus establishing the presence of an age-inappropriately reduced neogenesis of CD4⁺FOXP3⁻ Th cells in MS. In concordance with our present and previous results, decreased levels of naive CD4⁺ T cells in MS were also reported by Crucian et al. (29) and recently by Duszczyszyn and coworkers (30). To demonstrate that the altered Th homeostasis also affects the T_reg compartment, we analyzed the CD4⁺FOXP3⁺ cells of patient and control-derived PBMCs. In striking contrast to the total T_reg quantities that were unaltered in MS, the subpopulation of naive T_reg coexpressing CD31 was markedly reduced in patient-derived peripheral blood specimens, with most prominent alterations in younger subjects. This T_reg subset encompasses up to 20% of the entire T_reg pool and is phenotypically defined by the expression of CD4, CD25, FOXP3, CD45RA, and CD31. CD45RA⁺CD45RO⁻CD31⁺ T_reg decline with age in healthy persons and, unlike their CD31⁻ “memory-T_reg” counterparts, harbor TRECs. Therefore, they are most likely naive T_reg recently released from the thymus and are denoted here as “RTE-T_reg.” The reduced de novo generation of RTE-T_reg in MS occurred independently of clinical parameters, including an expanded disability status scale score, disease duration, and number of relapses, and was compensated by higher proportions of CD45RA⁻CD45RO⁺ FOXP3⁺ cells resulting in a stable cell count of the total T_reg population.

This shift from naive toward memory T_reg is also detectable in healthy elder individuals. In MS patients the lower levels of RTE-T_reg coincide with a moderately contracted T_reg TCR Vβ repertoire. Together with our observation that the T_reg inhibitory function seems to be reduced in age, which was also reported previously in a small number of healthy individuals (20), this raised the possibility that the prevalence of RTE-T_reg critically affects total T_reg function.

Accordingly, the comparative functional analysis of total T_reg and T_reg subsets revealed differing suppressive abilities. When assessing immunomagnetically separated total CD25^high T_reg, we found the expected functional defect in T_reg obtained from MS patients. Interestingly, the difference in T_reg-mediated suppression between patient and control-derived T_reg was totally neutralized when T_reg depleted of CD31⁺ cells were used in the coculture assays. The absence of CD31⁺ cells markedly reduced the inhibition of T cell proliferation mediated by donor T_reg but had no effect on the suppressive function of patient T_reg. Overall, there was a clear correlation between T_reg-mediated suppression and the prevalence of RTE-T_reg within the T_reg pool. Taken together, this clearly indicates that CD31-expressing naive T_reg contribute to the functional properties of the entire T_reg population and further illustrates the obvious lack of RTE-T_reg in MS patients.

Fritzsching et al. recently reported that human naive T_reg are less sensitive to apoptotic cell death as compared with conventional T_reg and suggested that naive T_reg are in an earlier state of differentiation whereas “classic” memory T_reg represent fully mature suppressor cells, which can be easily eliminated by apoptotic signals from conventional T cells (16). Moreover, in vitro expansion of human naive T_reg, but not memory T_reg, generates fully suppressive T_reg clones as recently demonstrated (31). Therefore, although it seems counterintuitive, it is possible that also in vivo the naive or especially the CD31⁺CD45RA⁺ T_reg compartment exerts the most effective suppression. In contrast, preactivated T_reg might be exhausted, prone to apoptosis, and less effective in suppression. This idea should be addressed in additional experiments in murine models.

Thus, we claim, that CD45RA⁺CD45RO⁻CD31⁺ T_reg represent a group of immature, natural regulatory T cells recently released from the thymus that are characterized by normal suppressive potential and insensitivity against apoptotic elimination. A constant supply of RTE-T_reg may be essential to compensate exhausted or eliminated memory T_reg and to maintain full inhibitory capacity of the entire T_reg compartment. Any down-regulation of RTE-T_reg release in the context of physiological thymic atrophy or pathological processes may impair T_reg homeostasis in the periphery and is likely to promote alterations in T_reg function in magnitudes that are detectable in elderly individuals and also in MS patients. Concordant with homeostatic shifts within the T_reg subset, Vukmanovic-Stejic et al. recently described a population of short-lived memory T_reg induced from primed/memory CD4⁺CD25⁻ T cells in the periphery that, via rapid peripheral turnover, is responsible for the maintenance of T_reg numbers during aging (32).

The reconstitution of T_reg homeostasis either by adoptive transfer of naive T_reg or by modulation of thymic function should have the potential to restore the impaired T_reg suppressive capacities in MS patients back to normal levels.

We have recently shown that prolonged treatment of MS patients with glatiramer acetate (GA) and IFN-β restores T_reg function via a direct targeting of the T_reg compartment (M. Korporal, M., J. Haas, B. Fritzsching, S. Möller, B. Fritz, C. S. Falk, P. H. Krammer, E. Suri-Payer, and B. Wildemann, submitted for publication). In addition to their anti-inflammatory effects (33, 34), immunomodulatory drugs, including GA, up-regulate numbers of naive T cells in MS patients (35, 36). Our own preliminary data indicate that RTE-T_reg numbers increase under GA treatment (data not shown) and possibly suggest a direct effect of this agent on thymic function. Therefore, we believe that pharmacologic modulation of the thymus and the reconstruction of peripheral T_reg homeostasis may trigger the augmentation of T_reg inhibitory capacities and may be a promising treatment strategy in MS and other autoimmune disorders.

The authors have no financial conflict of interest.