Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy individuals, provided by the Australian Red Cross Blood Service. CD4⁺ T cells were isolated from PBMCs using the CD4 T cell negative isolation kit and LD columns according to manufacture’s recommendations (Miltenyi Biotec). The purified fraction consistently contained 94–99% CD3⁺CD4⁺ T cells by flow cytometry. CD4⁺CD25⁻ and CD4⁺CD25⁺ cells were obtained by staining CD4 cells with anti-CD25 PE antibody and anti-PE magnetic cell isolation beads as per manufacture’s protocol (BD Pharmingen).

To obtain the induced TNFR2⁺ T cell subsets, the MACS purified T cell populations, either un-fractionated CD4⁺ T cells, or its sub-populations, CD4⁺CD25⁻ and CD4⁺CD25⁺ T cells were cultured. The T cells were suspended in AIM V medium (Invitrogen) containing 5% heat inactivated normal human serum (Sigma). The cells were added (5 × 10⁶ cells/2 mL/well) to 24 well plates, pre-coated with anti-CD3 antibody (2.5 μg/mL; OKT3, Biolegend). This was followed by the addition of soluble anti-CD28 antibody (1.25 μg/mL; CD28.2, BD Pharmingen), and the cells were cultured for 72 h at 37°C with 5% CO₂.

The above un-fractionated CD4⁺ T cell culture was harvested on day 3, and sorted using a FACS ARIA (Becton Dickinson) to isolate the CD25^hiTNFR2⁺, CD25^intTNFR2^int/− and CD25⁻TNFR2⁻ T cell populations.

The following monoclonal antibodies were used for flow cytometry analysis: TNFR2 FITC (R&D systems), CD3 FITC/APC, CD4 APC-Cy7, CD25 PE/PeCy7, CD127 bio-PerCP, CTLA4 APC (BD Pharmingen), and FOXP3 APC/PerCP. Intracellular staining was performed by firstly using the FOXP3 fixation/permeabilization kit (eBioscience) followed by staining the cells intracellular using the FOXP3 antibody. Flow cytometry was performed using BD LSRII, and data were analyzed using FlowJo software (Treestar).

For intracellular cytokine staining, the MACS isolated total CD4⁺ T cells were cultured for 3 days, stimulating with CD3/28. On day 3, PMA (50 ng/mL) and Ionomycin (1 mg/mL) were added for 5 h, with Brefeldin A (ebioscience) supplementation for the final 4 h. After stimulation, the cells were stained with intracellular IFN-γ, IL-2, and FOXP3 staining. Flow cytometry was performed using BD ARIA, and data were analyzed using FlowJo.

For suppression assays, the sorted cells above were irradiated at 40 Gy for use as suppressors. The responder cells consisted of cryopreserved autologous CD4⁺ T cells that are defrosted and washed. The sorted cell subsets, re-suspended at 10⁵ cells/50 μl in AIM V media containing 5% human serum, were mixed with an equal number of the responder cells. The mixture was then added to a 96 U bottom plate (Becton Dickinson) and stimulated for a further 72 h using CD3/28 stimulation as above. On day 3, cells were pulsed overnight at 37°C with 5 μCi/mL per well of TRK 120 titrated thymidine (Amersham, UK). Cells were then harvested and proliferation was determined by thymidine incorporation, measured by a liquid scintillation counter, Topcount NXT (Packard, USA). In some experiments, autologous CD4 depleted (using anti-CD4 microbeads, Miltenyi Biotec) PBMCs were irradiated at 40 Gy, and used as antigen presenting cells. A mixed lymphocyte reaction (MLR) was also used as responders, where PBMCs of three different donors were cultured together.

For proliferation assays, the sorted CD25^hiTNFR2⁺, CD25^intTNFR2^int/− and CD25⁻TNFR2⁻ cells was re-stimulated for 3 days using CD3/28, pulsed with titrated thymidine on day 3 and analyzed as above. Supernatant was removed prior to thymidine addition for cytokine analysis, where the cytokines present in the supernatant were determined using CBA-flex kits (BD Pharmingen) as per the manufacture’s protocol, and data analyzed using the manufacture’s software.

Total RNA was isolated from a minimum of 10⁵ cells of each of the sorted TNFR2 subsets using the RNA isolation kit (Roche, Germany). The first-strand cDNA was synthesized using oligo-dT primers and Superscript II reverse transcriptase (Invitrogen). qPCR for IL-10, TGF-b, IFN-g, T-bet FOXP3, and house keeping control 18SrRNA was performed using commercial primers and SYBR green reagent (Life technologies). PCR was performed using an ABI PRISM 7900 (Applied Biosystems). Results for target genes were normalized to 18SrRNA expression and expressed as fold changes between TNFR2 subsets.

To compare between the induced TNFR2 subsets, one-way ANOVA with Tukey’s multiple comparison test was used if data were normally distributed and Kruskal–Wallis test with Dunn’s multiple comparison test was used if the data were not normally distributed (Graphpad 5.0).

TNFR2 expression on T_regs is believed to be critical for T_reg function (12). It is unknown however, if functional TNFR2⁺ T_regs can be rapidly generated in vivo from circulating human peripheral blood CD4 lymphocytes during an active immune response. To address this question using an in vitro model, we purified CD4⁺CD25⁻ T cells and CD4⁺CD25⁺ T cells and stimulated the cells using anti-CD3 in the presence of CD28 to provide the secondary signal. After 72 h, these in vitro stimulated T cells could be differentiated into distinct CD25⁻TNFR2⁻, CD25^intTNFR2^int/−, and CD25^hiTNFR2⁺ T cells sub-populations (Figure 1A). While substantial numbers of CD25^hiTNFR2⁺ T cells were induced from the CD4⁺CD25⁻ and CD4⁺CD25⁺ T cell fractions, we found that these cells were generated more efficiently from the total un-fractionated CD4⁺ T cells (Figure 1C). It is possible that the interactions between CD4⁺CD25⁻ and CD4⁺CD25⁺ cells are helpful for the optimal induction of CD25^hiTNFR2⁺ T cells under physiological conditions.

These induced CD25^hiTNFR2⁺ T cells (regardless of the starting population) had a typical T_reg phenotype: significantly higher levels of FOXP3, CTLA4, and lower levels of CD127, when compared to the CD25⁻TNFR2⁻ and CD25^intTNFR2^int/− cells within the same culture. Figure 1D is a representative phenotype of CD25^hiTNFR2⁺ T cells induced from CD4⁺CD25⁺ T cell population. Figure 1E represents the quantitative analysis for FOXP3 expression on the induced TNFR2 subsets. The induced TNFR2⁺ T cell subset had the highest level of FOXP3 expression when compared to the other TNFR2^int/− T cell subsets, as shown in both Figures 1B,D. This was further confirmed at the mRNA level using qPCR (Figure 1F). As the phenotype of the induced CD25^hiTNFR2⁺ T cells contained similar percentage of cells positive for CTLA4 and FOXP3 (see Figure A1 in Appendix) across all starting populations (CD4⁺CD25⁻, CD4⁺CD25⁺, or un-fractioned CD4⁺ T cells), in the following sections we used the un-fractioned CD4 T cells as starting population.

As conventional effector human T cells also express low levels of FOXP3, it was important to confirm that the in vitro CD25^hiTNFR2⁺ T cells had the highest level of FOXP3 expression. We compared FOXP3 levels between the induced CD25^hiTNFR2⁺ T cells and ex vivo T_regs (Figure 2). Firstly, consistent with previous studies, we observed that ex vivo T_regs expressing TNFR2 had the higher levels of FOXP3 when compared to TNFR2⁻ T_regs (Figure 2A). Moreover, induced CD25^hiTNFR2⁺ T cells had significantly higher levels of FOXP3 when compared to ex vivo TNFR2⁺ T_regs (Figure 2B). Similar results were obtained even when FOXP3 MFI levels were normalized to the corresponding CD25⁻TNFR2⁻ cells for each donor to avoid any experimental variations and then compared between induced CD25^hiTNFR2⁺ T cells and ex vivo TNFR2⁺ T_regs (Figure A2 in Appendix).

Collectively, our data suggests that induced CD25^hiTNFR2⁺ T cells have a regulatory T cell phenotype and their FOXP3 levels are significantly higher than that of ex vivo T_regs.

Since the induced CD25^hiTNFR2⁺ T cells displayed a typical regulatory T cell phenotype, they would be expected to have regulatory function. Inhibition of proliferative T cell responses is a well-studied suppressor function attributed to T_regs. The CD25^hiTNFR2⁺ and CD25^intTNFR2^int/− cells were isolated using flow cytometry on day 3 from the CD4 T cell starting culture, irradiated, and added at 1:1 ratio to responders, which were autologous CD4⁺ T cells. Surprisingly, the induced CD25^hiTNFR2⁺ T cells did not suppress responder T cell proliferation, but instead, were found to enhance it (Figure 3A). Chen and colleagues demonstrated freshly isolated CD25⁺TNFR2⁺ T cells suppress T cell proliferation in assays that further contain added antigen presenting cells (APCs) (12). We therefore also performed the above suppression assays with the further addition of autologous APCs, to account for any potential indirect suppressor TNFR2⁺ T_reg effects. As shown in Figure 3B, the induced CD25^hiTNFR2⁺ T cells again failed to suppress under these conditions. We speculated that the strong signaling of responder cells by CD3/28 cross linking may not be capable of being suppressed by CD25^hiTNFR2⁺ T_regs, but other, more natural T cell stimulation protocols could be susceptible to suppression. The MLR where T cells from donors with different MHC react to each other is a biologically relevant assay (18). We found that induced CD25^hiTNFR2⁺ T cells were not capable of suppressing MLRs, and instead the addition of these cells into MLR cultures further enhanced proliferative responses (Figure 3C). Therefore, three different independent suppression assays indicated that in vitro induced CD25^hiTNFR2⁺ T cells from healthy CD4 T cells do not have conventional suppressor function.

To further analyze the function of induced CD25^hiTNFR2⁺ T cells, we assessed their proliferative capacity. As shown in Figure 4A, induced CD25^hiTNFR2⁺ T cells had a significantly higher proliferative capacity compared to the TNFR2⁻ T cells when re-stimulated with CD3/28 cross linking. Analyzing the cytokine production capacity of the sorted CD25^hiTNFR2⁺ T cells, we found that intracellular IL-2 production in TNFR2⁺ cells were significantly higher compared to the TNFR2⁻ or TNFR2^int/− subsets (Figure 4B), suggesting a mechanism underlying both their increased proliferative capacity and ability to enhance effector T cell proliferation. The CD25^hiTNFR2⁺ T cells also secreted significantly higher levels of IFN-g into the supernatant compared to the TNFR2⁻ or TNFR2^int/− subsets (Figure 4D), and interestingly, they also secreted IL-10, while the TNFR2^int/− and TNFR2⁻ cells did not secrete this cytokine (Figure 4C). IL-10 secretion, however, was present at a much lower concentration when compared to IFN-g and clearly insufficient to suppress proliferative responses. The phenotype of the induced CD25^hiTNFR2⁺ T cells was further confirmed by mRNA expression level, determined using qPCR. Compared to TNFR2^int/− and TNFR2⁻ cells, the in vitro induced CD25^hiTNFR2⁺ T cells expressed significantly higher mRNA level for IFN-g, IL-10 and the Th1 transcription factor, T-bet (Figures 4C,D). We also analyzed intracellular IFN-g production by these different induced TNFR2 populations from total PBMCs. Consistent with the mRNA levels and the cytokine levels present in the supernatant, we observed that the induced CD25^hiTNFR2⁺ T cells have the highest intracellular production of IFN-g compared to the other induced populations, TNFR2⁻ and TNFR2^int/− T cells (Figure 4E).