CpG Methylation of the IFNG Gene as a Mechanism to Induce Immunosupression in Tumor-Infiltrating Lymphocytes

Cell purification and culture

Naive CD4⁺ T lymphocytes were isolated by magnetic sorting using a human naive CD4 T cell isolation kit (Miltenyi Biotec) and an AutoMACS Separator. Naive T cells were stimulated with plate-bound anti-CD3 (OKT-3). Th1 skewing cultures contained 5 ng/ml recombinant human IL-12 (cat. no. 407711; Calbiochem) and 5 μg/ml neutralizing anti-IL-4 Ab (R&D Systems, cat. no. MAB204). Th2 skewing conditions included 5 ng/ml recombinant human IL-4 (Calbiochem, cat. no. 407635) and 5 μg/ml neutralizing anti-IFN-γ Ab (R&D Systems, cat. no. MAB285). Both Th1 and Th2 cultures contained 100 units/ml IL-2 and 1 μg/ml soluble anti-CD28 (BD Biosciences, cat. no. 555725). Cells were cultured in HyQ 1640-media (HyClone) with 10% human serum from human male AB plasma (Sigma-Aldrich, cat. no. H4522) and 1% penicillin-streptomycin solution (HyClone). Before DNA and RNA isolation, CD4⁺ T lymphocytes were further purified from culture either by magnetic sorting using human CD4 microbeads (Miltenyi Biotech) or by FACS in the case of naive T cell isolation. CD8⁺ and CD19⁺ lymphocytes were also purified through FACS sorting. All isolated cell populations were confirmed to have a purity of >95%. Tumors from six patients with colon cancer were cut and suspended using a loose fit glass homogenizer to obtain single cell suspensions. During surgery, tumor draining lymph nodes were identified as described previously (16). Tumor-infiltrating CD4⁺ T lymphocytes and CD4⁺ cells from tumor draining lymph node were extracted using a CD4 recognizing magnetically labeled Ab and columns according to manufacturer’s description (Miltenyi Biotec). The study was approved by the local ethical committee and informed consent was obtained by the patients.

Coculture with CCL-221

Naive CD4⁺ T lymphocytes were cultured in Th1 conditions as described above, washed, and put in coculture with the human colon carcinoma cell line CCL221. Th1 and CCL221 cells were grown separated by a cell culture insert with 1.0 μM in pore size (BD Biosciences, cat. no. 353102) in six-well tissue culture plates (BD Biosciences, cat. no. 353046). CCL221 cells were grown on the bottom of the wells and Th1 cells were grown in suspension in the cell culture inserts. After 5 days of coculture in HyQ 1640-media with 10% human serum and 1% penicillin-streptomycin solution, the CD4⁺ containing cell fraction was washed and restimulated with plate-bound anti-CD3 (OKT-3) and soluble anti-CD28 as described above. Cells were removed for RNA isolation before stimulation and after 6 h of stimulation.

5-Azacytidine treatment of tumor-infiltrating lymphocytes

Tumor-infiltrating CD4⁺ T cells were purified as previously described and cultured in HyQ 1640-media with 10% human serum and 1% penicillin-streptomycin solution. The cells were stimulated with anti-CD3 anti-CD28 coated beads (Invitrogen, cat. no. 111.31D) in a cell:bead ratio of 1:1. After 24 h, 5-Azacytidine (5-Aza, Sigma-Aldrich, cat. no. A-1287) was added to the culture at a concentration of 5 μM. After 48 h of culture in 5-Aza containing media the cells were washed and then cultured in 1640-media without 5-Aza for an additional 48 h before RNA isolation.

ELISA

Cells grown in Th1 and Th2 culture for 1 wk were washed and restimulated with plate-bound anti-CD3 and soluble anti-CD28 Abs. After 3 days in culture IFN-γ and IL-4, levels in supernatants were analyzed by ELISA (R&D Systems, DY285 and DY204) according to the manufacturer’s protocol.

RNA isolation and quantitative real-time PCR

RNA isolation was performed using TRIzol reagent (Invitrogen, cat. no. 15596–026). In brief, 100 ng RNA per sample was included in a reverse transcriptase reaction using an iScript cDNA Synthesis Kit according to the manufacturer’s protocol. Quantitative PCR was performed on an iCyclerIQ using 2X IQ SYBR Green Supermix. Cycle thresholds were obtained using iCycler IQ Optical System Software Version 3.1 from Bio-Rad. Expression levels were normalized to RNA polymerase II using the 2^−ΔΔCt method. Primers were purchased from Cybergene AB (Table I⇓).

DNA isolation and methylation analysis

Methylation analysis was conducted using bisulfite sequencing. Genomic DNA was isolated using DNeasy Blood & Tissue Kit (Qiagen) and bisulfite converted using EZ DNA Methylation Kit (Zymo Research) according to the manufacturer’s instruction. Primers shown in Table I⇑ were used in PCR reactions to amplify the IFNG promoter region and the conserved nucleotide sequence 1 (CNS1) region. PCR products were cloned into a pCR4-TOPO vector using a TOPO TA Cloning Kit (Invitrogen). DNA from individual bacterial clones was then included in a sequencing reaction with T3 and T7 primers using BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems) according to manufacturer’s instruction. Samples were run on a 310 Genetic Analyzer (Applied Biosystems) and sequences were analyzed using ABI Prism Sequencing Analysis Software 3.7 (Applied Biosystems). As controls for validity of the bisulfite sequencing reaction M.Sss1 (New England Biolabs) treated DNA from a bacterial artificial chromosome (BAC RPCI11–444B24, Invitrogen) and unmodified BAC DNA were used.

CFSE labeling

Naive CD4⁺ T lymphocytes were resuspended in PBS containing 0.1 μM CFSE solution. Cells were incubated in CFSE solution for 5 min at 37°C. Thereafter, cells were washed in RPMI 1640 medium containing 10% FCS. Cell pellets were resuspended in medium containing FCS and incubated for 30 min. Finally, cells were washed once more before resuspension in medium for culture.

Combined bisulfite restriction enzyme analysis (COBRA)

Genomic DNA was isolated and bisulfite treated as previously described. The CNS1 region was amplified by PCR with primers directed toward the same target sequence as in the case of bisulfite sequencing except with a 6-FAM conjugated forward primer (Primer sequences shown in Table I⇑). PCR products were digested with MaeII in excess (New England Biolabs). Following a phenol chloroform extraction digested products were loaded onto a 310 Genetic Analyzer (Applied Biosystems) and fragment analysis was performed using Gene Scan v3.7 application software (Applied Biosystems).

Methylation status of evolutionarily conserved regulatory elements in sorted cells

To identify evolutionarily conserved regulatory elements within the IFNG region we performed alignment using the VISTA tool between human and mouse DNA sequences. In agreement with previous reports (14, 15), we a found high sequence conservation within the promoter region and the CNS1 region (Fig. 1⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Human naive CD4⁺ T lymphocytes are hypermethylated at the IFNG promoter. A, VISTA plot showing conserved nucleotide sequences relative to the IFNG gene. Conserved nucleotide sequences are defined as stretches of nucleotides more than 100 bp long that show a conservation between the human and mouse genome of more than 70%. Blue regions are conserved nucleotide sequence elements that are located in exons, while red regions are located in noncoding DNA. B, Black bars indicate percentage methylation at each CpG position relative to the IFNG transcription start site. Shown are results from bisulfite sequencing of DNA from naive CD4⁺ T lymphocytes (top panel), CD8⁺ T lymphocytes (second panel from top), and CD19⁺ B lymphocytes (bottom panel) isolated by FACS sorting.

We started by investigating the methylation status of the IFNG promoter region in sorted cell populations (Fig. 1⇑B). Naive CD4⁺CD45RA⁺CD45RO⁻ T lymphocytes, CD8⁺ T lymphocytes, and CD19⁺ B lymphocytes were isolated by FACS sorting (>95% pure) and isolated DNA was subsequently processed for bisulfite sequencing. Naive CD4⁺CD45RA⁺CD45RO⁻ lymphocytes displayed hypermethylation at the promoter region (Fig. 1⇑B, top panel, number of clones analyzed: n = 14), an observation which is in sharp contrast to previously conducted experiments on mice (11, 13) where the Ifng promoter has been shown to be demethylated. CD8⁺ T lymphocytes (Fig. 1⇑B, second panel from top, n = 9) displayed a more pronounced demethylation at the promoter region compared with naive CD4⁺ T lymphocytes (70.4 vs 81% methylation at all positions analyzed). However human CD8⁺ T lymphocytes did not display the complete demethylation reported in mice (11). Surprisingly, demethylation was mostly pronounced in CD19⁺ B lymphocytes (Fig. 1⇑B, bottom panel, n = 10), which do not produce IFN-γ under physiological conditions (46.7% methylation at all positions analyzed). This observation is also contradictory to mouse experiments where CD19⁺ B lymphocytes have been described to have an overall hypermethylated promoter region (11). These results would imply different mechanisms in preventing the production of IFN-γ in human vs mouse CD19⁺ B cells.

Next, we analyzed the methylation status of the conserved upstream region CNS1 (Fig. 2⇓A). This region shows enhancer activity and binds important transcription factors for Th1 cell development such as T-bet and NF-AT (14, 15) and it is differentially methylated between Th1 and Th2 cells in mice (13). To our knowledge, the methylation status of this region in humans has not yet been examined. We found that the CNS1 region was hypermethylated in naive CD4⁺ T lymphocytes where a methylation level of 83.7% was observed at all positions analyzed (Fig. 2⇓B, top panel, n = 21). These findings are in coherence with previous studies in mice (13). Sorted CD8⁺ T lymphocytes cells were more demethylated with a methylation level of 68.3% at all positions analyzed (Fig. 2⇓B, middle panel, n = 9), whereas CD19⁺ B lymphocytes were almost completely methylated (93.7% at all positions analyzed) (Fig. 2⇓B, bottom panel, n = 9). As a control for the validity of the bisulfite sequencing method, we performed analysis on M.Sss1 treated DNA (completely CpG methylated) (Fig. 2⇓C, top panel, n = 10) from a BAC containing the IFNG locus as well as unmodified (unmethylated) BAC DNA (Fig. 2⇓C, bottom panel, n = 10), where 100 and 0% methylation, respectively, was verified.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Methylation status of the enhancer CNS1 in sorted cells. A, VISTA plot showing conserved nucleotide sequenced in the IFNG gene as previously described in Fig. 1⇑A. B, Black bars indicate percent methylation at each CpG position relative to the IFNG transcription start site. Shown are results from bisulfite sequencing of DNA from Naive CD4⁺ T lymphocytes (top panel), CD8⁺ T lymphocytes (second panel from top), and CD19⁺ B lymphocytes (bottom panel, n = 9). C, Validation of the bisulfite sequencing technique by analysis of M.Sss1 treated (top panel) and untreated (bottom panel) bacterial artificial chromosome.

We conclude that the human IFNG promoter displays a high level of methylation in naive CD4⁺ T lymphocytes. This is in sharp contrast to mouse, where the promoter region of naive CD4⁺ T lymphocytes is demethylated. In addition, the human IFNG promoter region in CD19⁺ B lymphocytes is demethylated whereas this region is methylated in mouse suggesting that the silencing of IFN-γ production in mice and men is controlled by different means of regulation.

Methylation pattern of regulatory elements during Th cell differentiation

Establishment of proper gene expression is dependent on chromatin remodeling proteins and transcription factors. The T-box protein T-bet has emerged as a key regulator of Th1 differentiation and Ifng expression. Retroviral expression of T-bet in developed Th2 cells not only induces Ifng expression but also suppresses the expression of Th2 cytokines (17, 18). Furthermore, T-bet induces Ifng expression directly by binding to regulatory elements in the Ifng locus, but also indirectly by interfering with the binding of GATA-3 to regulatory elements in the Th2 locus (19, 20). The importance of T-bet and GATA-3 for the establishment of cytokine expression patterns during Th cell differentiation prompted us to investigate the expression of IFNG, IL-4, T-bet and GATA-3 with quantitative real-time PCR during in vitro differentiation of naive CD4⁺ cells into Th1 cells (Fig. 3⇓A). As expected, IFNG expression was up-regulated while IL-4 expression remained low. T-bet expression also increased during differentiation with a decrement in GATA-3 expression as a consequence. Thus, the skewing protocol used is valid for the induction of relevant regulatory transcription factors.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Demethylation of regulatory elements during Th1 differentiation. A, Quantitative real-time PCR analysis of IFNG (left graph, open bars), IL-4 (left graph, filled bars), T-bet (right graph, open bars), and GATA-3 (right graph, filled bars) transcripts during Th1 differentiation. Naive CD4⁺ T lymphocytes were with plate bound anti-CD3 and anti-CD28 Abs and cultured in Th1 skewing conditions and RNA was isolated by TRIzol extraction before activation (top panel) at day 1 (second panel from top), day 3 (third panel from top), and day 7 (bottom panel). Graphs show analysis of triplicate samples normalized to expression in naive CD4⁺ T lymphocytes using the 2^−ΔΔCt method and RNA polymerase II as housekeeping gene. Error bars indicate mean ± SEM. B, Production of IFN-γ (left graph) and IL-4 (right graph) in Th1 cultures (□) and Th2 cultures (▪). Data represent analysis of triplicate samples and error bars indicate mean ± SEM. C, Methylation pattern of the IFNG promoter and CNS1 region during Th1 differentiation. Naive CD4⁺ T lymphocytes were stimulated and cultured in Th1 skewing conditions and cultures as previously described. DNA was extracted at day 1 (second panel from the top), day 3 (third panel from top), and day 7 (bottom panel) and processed for bisulfite sequencing of the promoter (right column) and the CNS1 region (left column). Black bars indicate percent methylation at each CpG position relative to the IFNG transcription start site. Gray areas in bar graph display clones where methylation status could not be determined during sequencing reaction. D, Methylation status of the IFNG promoter and CNS1 in Th2 differentiated cells. Cells were stimulated in Th2 skewing conditions and DNA was extracted after 7 days of culture and processed for bisulfite sequencing of the promoter (right graph) and the CNS1 region (left graph).

Furthermore, the differentiation into functional Th1 and Th2 effector cells was verified by cytokine production at the protein level. Secretion of IFN-γ and IL-4 in supernatants from Th1 and Th2 cultures was analyzed by ELISA (Fig. 3⇑B). Th1 differentiation was verified by IFN-γ production and low amounts of IL-4, whereas Th2-skewed cells produced high amounts of IL-4 and low levels of IFN-γ.

We next sought to investigate methylation patterns at the promoter and CNS1 regions during in vitro differentiation into Th1 and Th2 effector cells. During in vitro Th1 differentiation, a demethylation was observed at the promoter region, which was mostly pronounced at CpG positions upstream of the transcription start site (Fig. 3⇑C, right panels, number of clones analyzed: day 1, n = 15; day 3, n = 18; day 7, n = 23) whereas the whole promoter region remained hypermethylated during Th2 differentiation (Fig. 3⇑D, right panel, n = 10). These observations agree with the previously conducted experiments by White et al. (12). The upstream enhancer CNS1 became demethylated during Th1 differentiation (Fig. 3⇑C, left panels, day 1, n = 21; day 3, n = 25; day 7, n = 19), but remained methylated in differentiating Th2 cells (Fig. 3⇑C, left panel, n = 16). The demethylation of CNS1 was mostly pronounced at the 3′ end of the enhancer. Interestingly, the demethylation of both regulatory regions was initiated at day 3 of culture (Fig. 3⇑C, third panel from top), when a major increase in IFNG expression was observed (Fig. 3⇑A, third panel from top) suggesting methylation of both regions as a mechanism of limiting IFNG expression in noncommitted Th1 effector cells. Furthermore, increased T-bet expression and decreased GATA-3 expression was observed already at day 1 of culture (Fig. 3⇑A, second panel from top) and thus proceeded the transcription of IFNG. This observation agrees with the concept of T-bet and GATA-3 as determinants of lineage decision during early Th cell differentiation.

To ensure that the demethylation observed in Th1 differentiated cells were not affected by cells that had failed to become activated during differentiation, we performed CFSE staining of naive CD4⁺ cells before activation and subsequently isolated cells that had undergone cell cycle divisions at day 5. DNA from these cells was isolated and subjected to bisulfite sequencing of the promoter and the CNS1 region. The demethylation in dividing cells, 5 days poststimulation, showed a similar pattern to the 1 wk differentiated Th1 cells at both regions analyzed (Fig. 4⇓, B and C). Thus, the differentiation into Th1 effector cells parallels demethylation of the IFNG promoter and CNS1 regions, suggesting demethylation as an important mechanism for the control of IFN-γ production.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Methylation status of conserved regulatory elements in dividing cells. Naive CD4⁺ T lymphocytes were labeled with CFSE before activation and cultured in Th1 skewing conditions. A, At day 5 of culture, cells that had decreased CFSE emission as a consequence of cell cycle divisions were isolated by FACS sorting DNA from dividing cells were isolated and processed for bisulfite sequencing of the CNS1 region (number of clones analyzed: n = 15) (B) and the IFNG promoter region (n = 15) (C).

Methylation status of CD4⁺ tumor-infiltrating lymphocytes (TILs) and CD4⁺ cells from tumor-draining lymph node

Since bisulfite sequencing in itself is both time-consuming and limited by the number of clones analyzed in its ability to reflect the true methylation status of a cell population, we developed a COBRA based method for screening the methylation status of the IFNG gene. The −4229 position in the enhancer CNS1 region was chosen as a reporter position due to its selective demethylation in Th1 cells. Shown in Fig. 5⇓A is a schematic drawing of the COBRA-based method. After bisulfite treatment and PCR amplification, MaeII will selectively digest fragments at the −4229 position which were originally methylated at the cytosine residue and thus protected from deamination to uracil during bisulfite treatment. Fig. 5⇓B shows examples of electropherograms from Th1 and Th2 samples. Th1-skewed lymphocytes displayed a dominant peak of undigested PCR product due to their demethylated status of position −4229 (Fig. 5⇓B, top panel). In contrast when Th2 lymphocytes were subjected to the COBRA method, the major peak was recognized as a short digested PCR fragment demonstrating the presence of a methylated CpG at position −4229 (Fig. 5⇓B, bottom panel). Thus, the COBRA method using MaeII digestion of bisulfite-treated PCR product of position −4229 in the CNS1 region of the IFNG gene is useful for identification of Th1 committed lymphocytes.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Hypermethylation of the IFNG gene in tumor-infiltrating lymphocytes. A, Schematic drawing of the restriction enzyme based analysis of the −4229 reporter position. MaeII will selectively digest PCR products that were methylated at the −4229 CpG before bisulfite treatment and PCR amplification. Digestion is followed by Gene Scan analysis. B, Gene Scan electropherograms of MaeII digested PCR product from Th1 DNA (top panel) and Th2 DNA (bottom panel). Methylation level at the −4229 position is calculated as the relative amount of digested product based upon peak height. C, Graph showing methylation at the −4229 reporter position in Th1 cells, Th2 cells, CD4⁺ TIL, and CD4⁺ cells from sentinel lymph node (SLN). p values show level of significance from t test performed on the different populations. D, Quantitative real-time PCR analysis of IFNG and T-bet transcript from CD4⁺ tumor-infiltrating lymphocytes cultured in media with 5-Azacytidine (□) or without 5-Azacytidine (▪). Shown are results from experiments performed on TILs from two colon cancer patients. Ct values were normalized to expression in TILs grown without 5-Aza using the 2^−ΔΔCt method and RNA polymerase II as housekeeping gene. E, Quantitative real-time PCR analysis of IFNG (left graph) and T-bet (right graph) transcript from Th1 differentiated cells grown in Transwell culture with the colon cancer cell line CCL221 (□) or without CCL221 (▴). RNA was isolated before and after stimulation with plate bound anti-CD3 (OKT-3) and soluble anti-CD28. Graphs show analysis of triplicate measures normalized to expression in the resting population using the 2^−ΔΔCt method and RNA polymerase II as housekeeping gene. Error bars indicate mean ± SEM.

Altered cytokine expression in tumor-infiltrating lymphocytes suggests that the gene expression program of these cells has been shifted to the benefit of tumor survival. We therefore sought to investigate the methylation status of the IFNG gene in tumor-infiltrating CD4⁺ T lymphocytes isolated from colon cancer tumors. We measured the methylation level at the −4229 CNS1 reporter position using the established restriction enzyme-based method in differentiated Th1 and Th2 lymphocytes from four healthy donors, CD4⁺ tumor-infiltrating lymphocytes from six patients with colon cancer (Fig. 5⇑C), and CD4⁺ cells from tumor draining lymph nodes from three patients. The total number of draining lymph nodes analyzed was four since two independently draining lymph nodes were identified in one of the patients. As expected Th1-skewed cells displayed a significantly (p < 0.0001) lower degree of methylation (28.66 ± 0.9651 n = 4) at the −4229 reporter position compared with Th2 cells (71.40 ± 1.156 n = 4). Interestingly, sorted CD4⁺ tumor-infiltrating lymphocytes from six patients with colon cancer displayed hypermethylation at the −4229 CNS1 reporter position indicating a silenced IFNG gene (62.10 ± 1.520 n = 6) (Fig. 5⇑C). The level of methylation of CD4⁺ tumor-infiltrating lymphocytes were similar to those seen among Th2 skewed lymphocytes suggesting inappropriate silencing of the IFNG gene among TILs. The methylation level of CD4⁺ cells from tumor draining lymph nodes was significantly (p = 0.0005) less methylated (45.47 ± 2.934 n = 4) than CD4⁺ cells infiltrating the tumor. Table II⇓ shows detailed description of demethylation levels at the reporter position and detailed results from statistical analysis can be found in Table III⇓.

5-Azacytidine treatment induces IFN-γ expression in TILs

We have previously shown that tumor-infiltrating lymphocytes are poor producers of IFN-γ when stimulated with autologous tumor extract or ConA (16). The observation that TILs are hypermethylated at the IFN-γ (Fig. 5⇑C) region prompted us to investigate if culturing these cells in the presence of demethylating reagents would induce/restore the expression of IFN-γ. CD4⁺ tumor-infiltrating lymphocytes from two colon cancer patients cultured in the presence of 5-Azacytidine displayed 4- and 35-fold induced expression of IFN-γ (Fig. 5⇑D). The expression of T-bet was induced ∼3-fold in cells grown with 5-Aza. These results show that the observed hypermethylation among CD4⁺ TILs is the reason for the limited capacity to produce IFN-γ in these cells.

Suppressed IFN-γ production among Th1 cells in coculture with CCL221

Tumor reactive lymphocytes encounter their Ags in the tumor draining lymph node (16). The limited IFN-γ production (16) and hypermethylation (Fig. 5⇑C) in CD4⁺ TILs compared with CD4⁺ cells from tumor draining lymph nodes suggest tumor microenvironment induced cell fate reprogramming of tumor reactive T cells on the basis of DNA methylation. To investigate whether the presence of colon cancer cells would have any effects on differentiated Th1 cells in vitro we established a transwell culturing system. Differentiated Th1 cells grown in the presence of the colon cancer cell line CCL221 showed no response in IFN-γ production upon restimulation (Fig. 5⇑E). In contrast, cells grown in the absence of CCL221 were able to mount a vigorous IFN-γ response when stimulated, an observation that favors the assumption of an immunosuppressive environment surrounding the tumor.