Suppressive activity of human regulatory T cells is maintained in the presence of TNF

To the Editor:

Drugs that block tumor necrosis factor (TNF) have been used with success in the treatment of different autoimmune and inflammatory diseases, such as rheumatoid arthritis, Crohn’s disease and psoriasis. However, some patients do not respond to these treatments, and a few develop autoimmune responses for unknown reasons. Given that regulatory T (T_reg) cells expressing forkhead box P3 (FOXP3) have a major role in regulating autoimmune diseases, and that the production of TNF is increased in these diseases, the effect of TNF on T_reg cells is of interest. In previous studies, we looked at different conditions both in vitro and in vivo and observed that TNF promoted proliferation of T_reg cells while maintaining the cells’ suppressive activity in mice^1–3. However, experiments performed using human cells reported that TNF inhibited the suppressive activity of T_reg cells in vitro, which was associated with either decreased expression or inactivation of FOXP3 (refs. 4–7). Recently, it was shown that TNF, which is produced in the synovial fluid of individuals with rheumatoid arthritis, reduced T_reg cell suppressive function⁷.

To explore whether these conflicting data are due to species differences, we performed experiments with human T_reg cells purified from the blood of healthy donors. We observed a minimal decrease in T_reg cell-mediated suppression in the presence of TNF by using two different assays, in which T cells were activated by either antigen-presenting cells (APCs) and plate-bound CD3-specific antibodies (Fig. 1a) or CD3-specific and CD28-specific antibody coated-beads (bead assay) (Fig. 1b). However, as shown previously^8,9, incubation with TNF increased human effector T (T_eff) cell proliferation, even in the absence of T_reg cells (Supplementary Fig. 1), which confounded the interpretation of the suppressive assay. To examine the direct effect of TNF on T_reg cells, we preincubated T_reg cells with TNF before we assessed their suppressive activity. When T_reg cells were cultivated alone for 3 d with interleukin-2 (IL-2) to maintain their survival and function, the addition of TNF increased CD25 and FOXP3 expression (Fig. 1c). TNF did not increase the expression of CD25 or of FOXP3 in the absence of or at very low doses of IL-2 (Fig. 1d and Supplementary Fig. 2a). T_reg cells that were pretreated with both TNF and IL-2 and then washed twice to eliminate these cytokines before the suppression assay were slightly more efficient at suppressing T_eff cells than were control T_reg cells that had been pretreated with IL-2 alone using either the APC and plate-bound anti-CD3 or bead suppression assays (Fig. 1e,f). Similar findings were obtained using a third assay in which T cells were stimulated by APCs and soluble CD3-specific antibodies (Supplementary Fig. 3). When IL-2 was omitted from the T_reg cell pre-culture, the cells’ suppressive activity was reduced, and when TNF was added to the IL-2-free pre-culture, it did not alter T_reg cell suppressive activity (Supplementary Fig. 4). When T_reg cells were cultured without IL-2 for only 20 h, TNF had no effect on their phenotype (Supplementary Fig. 2b). Additionally, when T_reg cells were cultured for 20 h with or without IL-2 before the suppression assays, the presence of TNF in the pre-culture had no effect on their suppressive function in any assay condition tested (bead; APC and soluble CD3-specific antibodies; or APC and plate-bound CD3-specific antibodies) (Supplementary Fig. 5). Finally, in the presence of IL-2, the effects of TNF on human T_reg cells were also observed in a 3-day culture of whole CD4⁺ T cells, which resulted in an increased proportion of T_reg cells and the upregulation of FOXP3 expression (Supplementary Fig. 6).

Figure 1.

Human T_reg cell suppressive activity is not impaired in the presence of TNF. (a,b) Suppressive activity of purified T_reg cells co-cultured with carboxyfluorescein succinimidyl ester (CFSE)-labeled T_eff cells was measured at day 4 using an APC and plate-bound CD3-specific antibody assay (a) or a bead assay (b) in the presence (+ TNF) or absence of TNF. Representative CFSE profile of T_eff cells at a 1:1 T_eff:T_reg ratio (left) and mean ± s.d. of % of suppression at different T_eff:T_reg cell ratios (right) from 7 (a) and 8 (b) independent experiments using different donors for each experiment. (c,d) CD25 and FOXP3 expression on T_reg cells that were cultivated alone with high-dose IL-2 (c) or increasing doses of IL-2 (d) in the presence (+ TNF) or absence of TNF for 3 d. (c) One representative experiment of 17–20 different donors (left) and cumulative data (17 donors for the CD25 staining and 20 donors for the FOXP3 staining) of mean fluorescence intensity (MFI) of FOXP3 and CD25 (right). (d) MFI ± s.d. of FOXP3 and CD25 at different IL-2 doses from two independent responses using two different donors. The arrow shows values before culture. (e,f) T_reg cells were first cultivated with high-dose IL-2 in the presence (+ TNF) or absence of TNF for 3 d. After this, they were washed twice to remove cytokines and then analyzed for suppressive activity at day 4 in the absence of TNF. Representative CFSE profile of T_eff cells at a 1:1 T_eff:T_reg ratio (left) and ± s.d. of the percentage of suppression at different T_eff:T_reg cell ratios (right) using an APC- and plate-bound CD3-specific antibody assay (e, five different donors) or a bead assay (f, three different donors). *P < 0.05, **P < 0.01, ****P < 0.0001, obtained using nonparametric one-sided Wilcoxon signed-rank tests.

In light of these findings, the effect of TNF on human T_reg cells should be revisited. Because TNF has co-stimulatory effects on T_eff cells, as shown here and previously^8,9, T_eff cells are probably more resistant to T_reg cell–mediated suppression, as we previously reported in mice¹⁰. In previous reports demonstrating that TNF inhibited the suppressive activity of human T_reg cells, the cytokine was present during the suppression assays, and it may have been acting on both T_reg and T_eff cells^5,6. It was recently reported⁷ that human T_reg cells lost their suppressive activity even when they were exposed to TNF before, but not during, the suppressive assay. We were unable to replicate these findings under similar culture conditions (Fig. 1f; Supplementary Figs. 4b and 5c,d). We consistently observed that human T_reg cells do not lose their suppressive activity in the presence of TNF, as assessed in different labs and under different assay conditions (three different suppressive assays, various durations of preincubations, with or without IL-2) using cells from more than 28 different healthy donors. This conclusion is further supported by the fact that TNF increased the expression of CD25 and FOXP3 in human T_reg cells in the presence of IL-2 (Fig. 1c,d); furthermore, T_reg cells that exhibit the greatest suppressive activity are the ones expressing the highest level of tumor necrosis factor receptor 2 (TNFR2) in both mice and humans^2,11. Our findings remain compatible with the results of studies looking at the long-term effects of treatments that block TNF, which reported an increased number and enhanced function of T_reg cells^12,13. Indeed, it has been suggested that these observations were due to an increase in the differentiation of induced T_reg cells, rather than to a direct effect of TNF on natural T_reg cells¹³. Overall, we conclude that TNF does not inhibit the suppressive activity of human T_reg cells.

METHODS

Methods and any associated references are available in the online version of the paper.

ONLINE METHODS

Protocols.

Detailed protocols can be found in the Protocol Exchange at http://dx.doi.org/10.1038/protex.2015.078.

Cell purification and culture.

Peripheral blood mononuclear cells were isolated from 30 healthy donors who had given informed consent, with approval from the Institutional Review Board at the University Pierre et Marie Curie (Paris) and the US National Institutes of Health (NIH) and in accordance with the declaration of Helsinki. After a Ficoll Hypaque gradient centrifugation, T_reg cells were obtained after enrichment for CD25⁺ cells using anti-CD25-coated CliniMACS beads (Miltenyi Biotec, Ref.: 274–01) and LS columns (Miltenyi Biotec, 130-042-401), followed by sorting of CD4⁺CD25^highCD127^low/–CD45RA⁻ cells using a FACSAria (BD Biosciences). Treg cell purity was >99% (Supplementary Fig. 7). The CD25⁻cell fraction (obtained from the above bead purification) was then used to purify T_eff cells (CD4⁺CD25⁻) after the sorting of CD4⁺ cells using anti-CD4-coated beads (Miltenyi Biotec, Ref.: 130-050-301). From the CD25⁻CD4⁻ cell sub-fraction, we purified antigen presenting cells after positive selection of CD 14⁺ and CD19⁺ cells using anti-CD14 and anti-CD19 coated beads (Miltenyi Biotec, Ref.: 130-050-201 and 130-050-301, respectively). For the anti-CD25–coated beads, 4.5 × 10⁸ cells per ml were incubated with 50 μl of beads for 20 min at 4 °C in Baxter medium. For the other beads, 1 × 10⁸ cells per ml were incubated with 10 μl of beads. For antibody staining, 5 × 10⁷ cells per ml were incubated at saturating conditions for 20 min at 4 °C in Baxter medium. All cell culture experiments were performed in RPMI-1640 medium (Gibco or Mediatech, Inc.), supplemented with 10% heat-inactivated FCS (Gibco or Hyclone), 292 μg per ml or 2 mM L-glutamine, 100 IU per ml penicillin and 100 μg per ml streptomycin (Gibco). Pre-cultures were performed in 48-well plates (Corning Costar, 3548) at 1–3 × 10⁵ cells per well (but 4 × 10⁴ in Fig. 1d). Highly purified CD4⁺CD25^highCD127^low/–CD45RA⁻ T_reg cells were used for suppression assays (see below) or were cultivated alone with or without IL-2 (300 IU per ml, or at doses indicated in Fig. 1d, Novartis, Proleukin) and with or without TNF (PeproTech, Ref.: 300-01A, 50 ng per ml). Whole CD4⁺ T cells were sorted using a FACSAria and were then cultivated in 48-well plates at 2 × 10⁵ cells per well with or without IL-2 (100 IU per ml) and with or without TNF (PeproTech, 100 ng per ml).

Supplementary Table 1 lists the antibodies and cell viability dye used in our study. FOXP3 expression, assessed by three different mAbs (PCH101, 236A/E7 and 259D/C7), showed similar findings (Supplementary Fig. 8).

All suppressive assays were performed in U-bottom 96-well plates (Corning Costar, 3799) with 10⁴ T_eff cells and 10⁴ to 1.25 × 10³ T_reg cells in 200 μl of culture medium described above. For the APC suppression assay, CFSE-labeled (Sigma-Aldrich, Ref.: 21888-25MG-F, 1-M, 5 min at room temperature in PBS) T_eff cells, and T_reg cells (10⁴ to 1.25 × 10³) were stimulated by 10⁵ irradiated (50 Grey) APCs and plate-bound CD3-specific monoclonal antibodies (0.5 μg/ml for 2 h at 37 °C followed by two washes with PBS) or soluble CD3-specific monoclonal antibodies (5 μg/ml) for 4 d. For the bead suppressive assay, CFSE-labeled (1 μM) T_eff cells and T_reg cells were stimulated with 10³ anti-CD3/CD28 beads (LifeTechnologies, Ref.: 111.31D) for 4 d. In some cultures, TNF was added during the suppression assay at 50 ng/ml (PeproTech). For each T_eff/T_reg ratio, the percentage of suppression was calculated with the following formula: $[{\log }_2(y){\text{of}(\mathrm{T}}_{\text{eff}}{\text{cells}\text{alone})-\text{log}}_{\text{2}}(y){\text{of}(\mathrm{T}}_{\text{eff}}+{\text{T}}_{\text{reg}}{\text{cells}])/\text{log}}_2(y{)\text{of}(\mathrm{T}}_{\text{eff}}\text{cells alone})\times 100$. The y value corresponds to the mean fluorescence intensity of the CFSE of the whole T_eff cell population, divided by the mean fluorescence intensity of the CFSE of undivided T_eff cells.

Statistical analyses.

Nonparametric paired one-sided Wilcoxon signed-rank tests were used for statistical analyses. No samples were excluded. *P < 0.05, **P < 0.01, ****P < 0.0001.