Supplementary Figure 1: Activation Status (HLA-DR) of Tregs Is Highly Associated with Disease

Supplementary Figure 1: Activation status (HLA-DR) of Tregs is highly associated with disease progression.

For a subset of patients with different disease status, expression levels of HLA-DR (A/B) and Ki67 (C/D) were measured on non-Tregs and Tregs. Both markers were elevated in patients with progressive disease in both cell populations, indicating generalized immune activation. E/F) Relative and absolute Tregs frequency are significantly associated with generalized immune activation. G/H) Treg activation status was plotted against relative and absolute Treg - number.

Supplementary Figure 2: Functional analysis of sorted CD39+ T cells: CD4+ CD25high CD39+ cells, but not CD4+CD25high CD39- cells suppress PBMC proliferation.

Ca 500 million fresh PBMC of healthy volunteers were life-sorted on a FACSAria machine. Representative plots for 3 independent experiments with PBMC from different healthy donors are depicted. A) Gating strategy for sorting of CD25high, CD39+ (gate P6) and CD25high, CD39- (gate P7) CD4+ cells from PBMC. B) Representative analysis of the cytokine production profile of sorted PBMC regarding IFN-γ and IL-10 production after 12h stimulation with PMA/ionomycin (unstimulated controls not shown): Only the CD25low, CD39- population (=effector T cells) show significant IFN-γ production after stimulation. None of the other populations showed significant IL-10 or IFN-γ production in comparison with the unstimulated control. C) To analyze the suppressive characteristics of CD25high, CD39+ and CD25high, CD39- Tregs, suppression assays were performed as described above. PBMC without additional Tregs showed strong proliferative responses to SEB stimulation (stimulation index (SI)=60,25). When CD25highCD39+ Tregs were added to the PBMC, the proliferation capacity of the cells dropped significantly (SI=27,05, p<0,0086) whereas CD25high CD39- Tregs did not influence the proliferation capacity of the PBMC significantly (SI=68,56 p=0.5).

Supplementary Figure 3: Comparison of relative Treg frequencies in PMBC and LNMC

LNMC and PBMC at corresponding timepoints were available for 10 patients and Treg frequencies were compared. A) FACS plot of PBMC (left) and LNMC (right) of a representative patient. B/C) Significantly higher relative Treg frequencies (10.8% ± 3.1) and FoxP3 MFI (1593 ± 272) were found in lymph nodes compared to PBMC (Tregs: 7.6% ± 3.0; p < 0.0028; FoxP3 MFI: 1095 ± 199; p < 0.0001).