Our study population consisted of 19 healthy individuals aged between 26 and 75 (68% female). Peripheral blood of these subjects was collected after informed consent and according to biobank protocol with approval of the local ethics committee (MEC-2010-022). We chose to study healthy individuals to eliminate confounding effects of disease on DNA methylation (30). It is also known that IFNγ DNA methylation is significantly lower in CMV seropositive individuals than in CMV seronegative individuals (31). To compose a homogeneous group and eliminate CMV effects on inter-individual differences in methylation levels, only CMV seropositive individuals were included in the study.

Peripheral blood mononuclear cells (PBMCs) were isolated from the peripheral blood by density gradient centrifugation using Ficoll-Paque (GE Healthcare, Chicago, IL, USA). Isolated PBMCs were stored at −140°C until further use. Total T cells were isolated from the PBMCs by magnetic cell separation on the autoMACS (Miltenyi Biotech, Bergisch Gladbach, Germany) according to the pan T cell protocol using the deplete S settings. Purities were >90% CD3⁺ cells after isolation.

The naive and memory T-cell populations were isolated from the PBMCs using fluorescence-activated cell sorting (FACS) by the BD FACSAria™ II (BD Biosciences, San Jose, CA, USA). The PBMCs were stained with CD3 Brilliant Violet 510 (Biolegend, San Diego, CA, USA), CD4 Pacific Blue (BD Biosciences), CD8 APC-cy7 (BD Biosciences), CD45RO APC (Biolegend), CCR7 PE-cy7 (BD Biosciences), and to exclude non-viable cells the cells were also stained with 7AAD PerCP (BD Biosciences). Naive cells were defined as CCR7⁺CD45RO⁻, central memory cells as CCR7⁺CD45RO⁺, effector memory (EM) as CCR7⁻CD45RO⁺, and the highly differentiated EMRA cells as CCR7⁻CD45RO⁻ (32). After cell sorting, the purities were >95% for each sorted fraction.

The T cells were stimulated for 4 days with α-CD3/CD28 coated Dynabeads^® (Gibco, Waltham, MA, USA) in a bead to cell ratio of 1:1 at day 0. Fifty thousand cells were cultured per well in a 96-well plate. The cells were cultured in the absence or presence of tacrolimus, MPA or 5-aza-2′deoxycytidine (decitabine). Tacrolimus (Prograf^®, Astellas Pharma, Tokyo, Japan) was added to the cells in a concentration of 10 ng/mL which is a clinically relevant concentration that is reached in transplant recipients (33). MPA (Sigma-Aldrich, St. Louis, MO, USA) was added to the cells in a concentration of 0.2 µg/mL, a concentration at which the cells are still able to proliferate. Our positive control, the demethylating agent decitabine (Sigma-Aldrich) (34), was added to the cells in a concentration of 10⁻⁶ M, a concentration at which the cells are still able to proliferate. Each drug-treated sample has a matched negative control (stimulation alone).

The cells were incubated at 37°C in 5% CO₂ and harvested at 4 h, days 1, 3, and 4 for DNA methylation analysis, and at days 1 and 3 for flow cytometry analysis. To assess viability and proliferation, the cells were counted before and after stimulation using conventional light microscopy and Trypan Blue staining (Thermo Fisher Scientific, Waltham, MA, USA).

Flow cytometry was used to determine the phenotype of T cells immediately after isolation and at days 1 and 3 after stimulation. We also measured the percentage of IFNγ producing cells at these time points. The samples were treated with Brefeldin A (GolgiPlug™, BD Biosciences) for 16 h prior to flow cytometry analysis. The monoclonal antibodies used for cell surface staining were the same as previously described for the FACS cell sorting. In addition, the cells were permeabilized using permeabilize solution 2 (BD Biosciences), and stained for intracellular IFNγ with FITC labeled IFNγ (BD Biosciences). The cells were then analyzed on the FACSCanto II (BD Biosciences) with FACSDiva software. All flow cytometry data were analyzed using Kaluza software 1.3 (Beckman Coulter, Brea, CA, USA).

After harvesting, the cells they were pelleted, frozen in liquid nitrogen, and stored at −80°C until bisulfite conversion. The T-cell pellets were digested with proteinase K and bisulfite treatment was performed using the EZ DNA Methylation-Direct kit (Zymo Research, Irvine, CA, USA) according to the manufacturer’s protocol. Bisulfite treatment introduces methylation-dependent changes in the DNA, demethylated cytosines are converted into uracil whereas methylated cytosines remain unchanged. The bisulfite-treated DNA was amplified by PCR. A 230 base pair region of the IFNγ promoter was amplified using the Pyromark PCR kit (Qiagen, Venlo, The Netherlands). A forward primer with the sequence 5′-ATGGTATAGGTGGGTATAATGG-3′ and a biotin-labeled reverse primer with the sequence 5′-CAATATACTACACCTCCTCTAACTAC-3′ (Sigma-Aldrich) were used, both at a concentration of 10 pmol/μL (31). The PCR conditions were 15 min at 95°C, 45 cycles of 30 s 94°C, 30 s 58°C, 30 s 72°C followed by 10 min at 72°C, and final storage at room temperature (21°C). Prior to pyrosequencing, the PCR product was visualized on a 1% agarose gel to verify the size of the amplicon. Two important CpG sites are inside this amplicon, CpG -186 and CpG -54. These sites are within binding domains of transcription factors (26, 31).

Pyrosequencing is an excellent technique to quantitatively measure DNA methylation at single CpG-site resolution, yielding accurate, and reproducible results (35, 36). The IFNγ PCR product was sequenced using a PyroMark Q24 pyrosequencer (Qiagen). Minor adjustments were made to the manufacturer’s protocol: to immobilize the PCR product 1 µL Streptadivin Sepharose High Performance Beads (GE Healthcare) was used per sequence reaction and annealing of the sequence primers was done for 3 min at 80°C. The CpG -186 sequence primer was 5′-GGTGGGTATAATGGG-3′ and the CpG -54 sequence primer was 5′-ATTATTTTATTTTAAAAAATTTGTG-3′, both at a concentration of 10 µM (31). Two DNA methylation standards were used as control, human high, and low methylated DNA (EpigenDx, Hopkinton, MA, USA). Research shows that methylation at adjacent sites is correlated (37) therefore the methylation percentages of the two CpG sites, site -54 and -186, were pooled per individual and the mean DNA methylation percentage is presented in the results.

Statistical analyses were performed with SPSS Statistics version 21.0 (IBM Corp., Armonk, NY, USA). The Mann–Whitney U test was used for unpaired analysis to identify differences between the conditions at a certain time point. The Wilcoxon signed-rank test was used for paired analysis when comparing different time points within a condition. A p-value <0.05 was considered statistically significant.

To exclude complete cell cycle arrest as a cause for methylation differences, we compared cell numbers under the different conditions after stimulation. Cell numbers were lower if cells were cultured with either tacrolimus, MPA, or decitabine than if the cells were cultured without those factors, but due to overlapping ranges this difference was not statistically significant (Figure S1 in Supplementary Material). Our results suggest that the cells were still able to proliferate under the chosen concentrations of the different drugs.

To determine the changes in DNA methylation after T-cell stimulation, we analyzed IFNγ promoter methylation at several time points after stimulation. IFNγ DNA methylation of total T cells increased significantly after stimulation with α-CD3/CD28 (p = 0.002; Figure 1B). Stimulated T cells showed a median DNA methylation percentage of 47% (range: 35–59%) at day 0 and this was significantly increased at day 4 (59%; 46–66%).

DNA methylation of T cells cultured in the presence of tacrolimus increased significantly from 49 (42–59%) to 53% (44–67%) (p = 0.043) and did not differ significantly from the stimulated condition at any of the given time points (Figure 1B). DNA methylation of T cells cultured in the presence of MPA increased from 48 (43–56%) to 61% (46–66%) and also did not differ significantly from the stimulated condition (Figure 1B). Our positive control, T cells cultured in the presence of decitabine, significantly decreased in DNA methylation between day 0 and day 4 (p = 0.028; Figure 1B).

Since our total T-cell population was a heterogeneous mixture of naive and memory T cells with different methylation profiles (29), we continued to study isolated cell populations to infer whether tacrolimus or MPA did influence these cell types individually.

Pure naive (CCR7⁺CD45RO⁻) (Figure 2A) and memory (CD45RO⁺ and CCR7⁻CD45RO⁻) (Figure 2C) T-cell subsets were stimulated separately. IFNγ DNA methylation significantly decreased in the naive start population in the absence of tacrolimus or MPA, from 78 (75–83%) at day 0 to 67% (61–77%) at day 4 (p = 0.011; Figure 2B). The two immunosuppressive drugs had differential effects on this reduction in DNA methylation. While tacrolimus had no effect, MPA neutralized the effect of stimulation significantly and DNA methylation did not decrease (78%; 76–82% at day 0 and 77%; 75–78% at day 4). This differential effect resulted in a significant difference between stimulation only and the addition of MPA on day 3 (p = 0.005) and day 4 (p = 0.014; Figure 2B).

In the total memory start population, IFNγ DNA methylation significantly increased in the absence of tacrolimus or MPA, from 24 (19–31%) at day 0 to 38% (30–46%) at day 4 (p = 0.012; Figure 2D). This increase was not affected by tacrolimus nor MPA, both these conditions were not significantly different from stimulation alone.

As explained in the Section “Introduction,” we expected effector-gene promoters to demethylate after activation to allow transcription of the corresponding effector gene. We observed this in the naive T cells, demethylation of the IFNγ promoter took place after 3 days of stimulation (Figure 2B). However, the IFNγ promoter of the memory T cells did not demethylate after 1, 3, or 4 days after stimulation (Figure 2D). Therefore, we speculated that demethylation occurred in a shorter timeframe than 24 h, to allow memory T cells to produce IFNγ protein. To address this question, we harvested memory T cells at 4 h after stimulation and indeed we observed a significant decrease (3–12%; p = 0.043) in methylation followed by remethylation to base levels after 24 h (Figure 3).

The isolated naive T cells, which were CCR7⁺CD45RO⁻ at day 0, were analyzed for the expression of CD45RO and CCR7 after 1 and 3 days of stimulation in the absence and presence of tacrolimus or MPA. CD4⁺ and CD8⁺ T cells were gated separately (Figure 4), the percentages CD4⁺/CD8⁺ do not differ significantly between the conditions (Figure S2 in Supplementary Material). After 1 day of stimulation, the phenotype did not differ significantly from day 0 in both CD4⁺ and CD8⁺ T cells. On day 3, there was a significant shift toward CD45RO⁺ cells in the stimulated condition (p = 0.008). The shift was observed in all three conditions and in both the CD4⁺ and CD8⁺ T cells (Figures 4B,C). These cells, which were CD45RO⁻ at day 0, upregulated their CD45RO expression showing a central-memory-like phenotype at day 3. When we compared the different conditions with stimulation only at day 3, tacrolimus (p = 0.013) and MPA (p = 0.039) significantly repressed CD4⁺ differentiation and MPA also significantly repressed CD8⁺ differentiation (p = 0.014; Figures 4B,C).

The isolated memory T cells, which were CD45RO⁺ and CCR7⁻CD45RO⁻ at day 0, were also analyzed by flow cytometry after 1 and 3 days of stimulation in the absence or presence of tacrolimus or MPA. CD4⁺ and CD8⁺ T cells were gated separately (Figure 5). The percentage of CD8⁺CD45RO⁺ cells increased significantly after 3 days of stimulation, both in the CCR7⁺ (p = 0.008) and CCR7⁻ (p = 0.021) population (Figure 5C). In the CD4⁺ population, we observed an increase in the CCR7⁺CD45RO⁺ population (p = 0.011) and a decrease in the CCR7⁻ population (p = 0.021) (Figure 5B). When we compared the different conditions with stimulation only at day 3, no significant differences were found.

Interferon gamma protein production was measured using intracellular staining in both the sorted naive T cells and the sorted memory T cells (Figure 6). The sorted naive T cells did not produce IFNγ protein at day 1 after stimulation (data not shown) while 10% (3–19%) of the sorted memory T cells did produce IFNγ. Tacrolimus significantly inhibited IFNγ production, hardly any cells produced IFNγ in the presence of tacrolimus (Figure 6B). MPA did not have a significant effect on IFNγ production and the percentage IFNγ producing cells did not differ from stimulation only. Three days after stimulation of the sorted memory T cells, few cells still produce IFNγ both in the presence and absence of tacrolimus or MPA.