CD4⁺CD25^high Foxp3⁺ regulatory T cells in myelodysplastic syndrome (MDS)

The presence of autoimmune diseases and T-cell–mediated inhibition of hematopoiesis is now a recognized feature of myelodysplastic syndrome (MDS).^1,2  Oligoclonal CD8⁺ T cells occur in up to 95%³  of patients. However, the antigens produced by MDS cells that lead to these T-cell responses are unknown. Recent reports have confirmed that immunosuppressive therapy with antithymocyte globulin and/or cyclosporine A can lead to lasting hematologic responses and abrogation of T-cell clones, which is particularly noticeable in low-risk MDS.^4-6  Regulatory T cells (Tregs) play an important role in the immune surveillance of malignancies.^7,8  We hypothesized that the effect of Tregs in MDS may be 2-fold: first, expansion of Tregs may inhibit effective immune responses against the dysplastic clone, thereby facilitating disease progression; second, low numbers of Tregs may be associated with low-risk MDS, permitting the emergence of autoreactive T-cell clones and secondary bone marrow hypoplasia

We studied the number of CD4⁺/CD8⁺ Tregs and the function and clonality of CD4⁺ Tregs in the peripheral blood of patients with MDS at different disease stages and correlated the results with known prognostic variables. In order to gain a better understanding of the origin of the expanded Tregs, we analyzed the naive/memory subpopulations in both low- and high-risk MDS.

We demonstrate for the first time a significant increase in the number of CD4⁺CD25^high Foxp3⁺ Tregs in high-risk disease. The Tregs are polyclonal, and the naive-memory ratio is significantly higher in the high-risk group.

MDS was defined according to the World Health Organization (WHO) classification⁹  in 52 patients (30 men, 22 women) with a median age of 64.5 years (range,17-83 years). The median age was not different between MDS subgroups (P = .34). All patients were sampled prior to the commencement of any treatment and at least 2 weeks after any blood transfusion. Age-matched controls were obtained from 9 healthy donors. Ethical approval by King's College Hospital Research Ethics Committee (London, United Kingdom) was gained prior to study commencement. Written informed consent was obtained from all patients and controls in accordance with the Declaration of Helsinki.

Mononuclear cells were separated from peripheral blood by density gradient sedimentation (Histopaque; Sigma, St Louis, MO). At least 2 × 10⁶ peripheral blood mononuclear cells (PBMCs) were stained for flow cytometry analysis.

PerCP anti-CD3, FITC anti-CD4 monoclonal antibody (mAb), or FITC anti-CD8 (Becton Dickinson, San Jose, CA) and PE anti-CD25 from eBioscience (San Diego, CA) were used for surface antigen staining. PE-Cy5 anti-human Foxp3 (PCH101) and PE-Cy5 rat IgG2a isotype control from eBioscience were used for intracellular Foxp3 staining according to the manufacturer's instructions. The following antibodies were also used for Treg subpopulation analysis: Pacific Blue anti-CD3, FITC anti-CD27, APC anti-CD45RO, and AmCyan anti-CD4 (BD Biosciences, Palo Alto, CA). Flow cytometry was performed by FACSCantoII (Becton Dickinson), and data were analyzed on a BD FACSDiva (Becton Dickinson). Of the CD3⁺ T cells, the absolute number of CD4⁺CD25^highFoxp3⁺ and CD4⁺/CD8⁺CD25⁺Foxp3⁺ cells was calculated. Simultaneous naive and memory subpopulations of CD4⁺ Tregs were defined by CD25^highFoxp3⁺CD27⁺CD45RO⁻ and CD25^highFoxp3⁺CD27⁺CD45RO⁺, respectively.^10,11 

CD4⁺CD25⁺ Tregs were sorted using a multistep isolation kit (Miltenyi Biotec, Auburn, CA) designed to isolate CD4⁺ cells with high expression of CD25. Sorted cells were consistently greater than 90% Foxp3⁺ (data not shown). Trizol (Invitrogen, Carlsbad, CA) was used for RNA extraction, and first-strand cDNA was generated using the Superscript III kit (Invitrogen). CDR3 of T-cell receptor (TCR) Vβ-chain were amplified using Vβ-specific forward and Cβ reverse primers.¹²  The CDR3 lengths were analyzed using an ABI 3130xl capillary sequencer (Applied Biosystems, Foster City, CA). The overall complexity of Vβ subfamilies was calculated, and the cloning and sequencing of any skewed spectratype was done as previously described.^2,13 

Purified responder T cells (CD4⁺CD25⁻) from patients with MDS were incubated with anti-CD3/CD28 beads with or without a 1:1 ratio of Tregs. Supernatants were analyzed for concentration of IFN-γ by enzyme-linked immunosorbent assay (ELISA).

Statistical analysis was performed using SPSS version 14.0 (Chicago, IL). The nonpaired t test and Mann-Whitney U test were used to compare low- and high-risk groups, and significance was set at a P value less than .05.

A total of 7 (13%) patients had refractory anemia (RA), 16 (31%) patients had refractory cytopenia with multilineage dysplasia (RCMD), 16 (31%) patients had refractory anemia with excess blasts (RAEB), 9 (17%) patients had 5q− syndrome, and 4 (8%) had myelodysplastic/myeloproliferative disease (MDS/MPD).

Cytogenetics were normal in 49% of patients, isolated del(5q) was in 17% of patients, complex was in 16% of patients, and stable single abnormalities were present in 18% of patients. The absolute number of CD4⁺ Tregs was significantly higher in patients with complex cytogenetic abnormalities compared with 5q− syndrome (P = .008; Figure 1D). An International Prognostic Scoring System (IPSS) score of 0 (low risk), 0.5 to 2 (intermediate risk), and 2.5 or higher (high risk) was observed in 18 (35%) of 52 patients, 25 (48%) of 52 patients, and 9 (17%) of 52 of patients, respectively.

The median number of CD4⁺CD25^highFoxp3⁺ Tregs in 5q− syndrome was 0.51 × 10⁷/L (range, 0.2-1.07 × 10⁷/L); in RA, 0.52 × 10⁷/L (range, 0.5-1.29 × 10⁷/L); in RCMD, 1.18 × 10⁷/L (range, 0.24-2.34 × 10⁷/L); in RAEB, 2.11 × 10⁷/L (range, 0.8-7.06 × 10⁷/L); and in MDS/MPD, 3.06 × 10⁷/L (range, 0.8-5.0 × 10⁷/L). Median CD4⁺ Tregs were significantly higher in patients with 5% or more bone marrow blasts compared with less than 5% BM blasts (2.11 × 10⁷/L vs 0.75 × 10⁷/L; P < .001) and in high IPSS scores compared with low/intermediate IPSS scores (1.96 × 10⁷/L vs 0.51 × 10⁷/L; P < .001) despite no difference in the median age between the 2 groups. No significant correlation was observed between the number of Tregs with platelet count (P = .66) or neutrophil count (P = .07). The number of Tregs in the 14 transfusion-dependent patients was slightly lower than in nondependent patients (0.95 × 10⁷/L vs 1.24 × 10⁷/L), but not statistically significant (P = .67).

Although the number of patients in the MDS/MPD group is small, the mean CD4⁺ Treg numbers were higher than that of other subgroups. Similarly, patients studied at the time of disease progression (n = 17) had significantly elevated CD4⁺ Tregs compared with 35 patients with stable disease (2 × 10⁷/L vs 0.69 × 10⁷/L; P < .001; Figure 1). The number of CD4⁺ Tregs was lower in patients with 5q− syndrome, RCMD, and RA, but was not statistically different from healthy controls (P = .6), whereas patients with RAEB and MDS/MPD had significantly higher CD4⁺ Tregs than healthy controls (P < .001 and P = .02, respectively). It is notable that among patients with RCMD, 7 had CD4⁺ Tregs that were in the normal range (0.62% ± 0.78% of CD3⁺ T cells) and 9 had CD4⁺ Tregs in the high-risk range (1.42% ± 2.7%), reflecting the biological heterogeneity of this subgroup.^14-16  There was no difference in the number of CD8⁺ Tregs between MDS subtypes (P = .28), IPSS (P = .19), or disease progression (P = .19).

Subpopulation of Treg CD3⁺CD4⁺CD25^highFoxp3⁺CD27⁺CD45RO⁻ and CD3⁺CD4⁺CD25^highFoxp3⁺CD27⁺CD45RO⁺ subsets were analyzed in 10 patients with the highest and lowest numbers of Tregs and 9 age-matched controls. The percentage of naive Tregs was significantly higher in high-risk patients compared with low-risk and healthy subjects (P = .032; Figure 2A,B). The ratio of naive to memory Tregs was also significantly higher in the high-risk group than in the low-risk (P = .016) or control groups (P = .032).

The clonality of CD4⁺ Tregs was analyzed by the spectratyping of 6 low-risk and 9 high-risk patients. The spectratype of CD4⁺CD25⁺ TCR amplicons showed a polyclonal pattern, and the overall complexity of Vβ spectratypes (confirmed by sequencing) was 100 peaks (range, 77-105 peaks) in the high-IPSS-score group and 102 peaks (range, 75-110 peaks) in the low-IPSS-score group (P = .54; Figure 2C,D). This finding, in addition to increased naive Tregs suggests that in MDS, like in other malignancies, the expanded Tregs are not clonal and may arise by peripheral expansion.⁸  By contrast, the spectratype of CD8⁺ T cells in 10 samples (4 low, 3 intermediate, and 3 high IPSS scores) was skewed on average in 6 of 24 Vβ subfamilies.

The suppressive effect of CD4⁺ Tregs from patients with MDS was demonstrated by a reduced level of IFN-γ in cocultures containing Tregs compared with responder cells alone (Figure 2E).

Our data show that Foxp3⁺ Treg expansion occurs frequently in high-risk MDS as well as at disease progression. The increase is predominantly in the naive subset, as has been reported previously in other hematologic malignancies, suggesting peripheral expansion.⁸  By contrast, in low-risk MDS, the Treg population tends to be lower, thereby permitting the emergence of autoimmune responses, including those directed against the dysplastic clone.

We thank the nursing and medical staff in the department of hematological medicine at King's College Hospital for providing clinical samples and Dr Stephen Devereux and Dr Piers Patten for providing healthy control samples. We also thank Professor T. Hamblin and Dr Ziyi Lim for their critical review of the manuscript.

This work was supported by King's College Hospital Joint Research Committee and King's College London. W.I. is funded by the Leukemia Research Fund. M.W.'s other affiliation is Institute of Immunology, Charite Medical School, Berlin, Germany.

Contribution: S.K. designed and performed research, analyzed and interpreted data, and drafted the manuscript; W.I. provided clinical data and drafted the manuscript; J.H. provided clinical samples; D.D. designed research; L.B. designed and interpreted data; B.A. performed research and analyzed data; G.L. designed research; M.W. provided research tools; J.M. provided research tools and drafted the manuscript; F.F. designed research, interpreted data, and drafted the manuscript; and G.J.M. designed research, interpreted data, and drafted the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Ghulam J. Mufti, Department of Hematological Medicine, Kings College London, Rayne Institute, 123 Coldharbour Lane, London, United Kingdom, SE5 9NU; e-mail: ghulam.mufti@kcl.ac.uk.