FOXP3 Expression in GARP-Transduced Helper T Cells Is Not Associated with FOXP3 TSDR Demethylation

Abstract

Aim:

Glycoprotein A repetitions predominant (GARP or LRRC32) represents a human regulatory CD4+ CD25^hi FOXP3+ T (T_reg) cell-specific receptor that controls FOXP3. Ectopic expression of GARP in helper T (T_h) cells has been shown to be sufficient for the induction of FOXP3 and generation of a stable regulatory phenotype. Since expression of FOXP3 in Treg cells is epigenetically controlled by a conserved motif, the so-called T_reg-specific demethylated region (TSDR), we asked whether GARP-mediated upregulation of FOXP3 in Th cells is similarly accompanied by demethylation of the TSDR.

Methods:

DNA methylation of the FOXP3 TSDR was analyzed by direct sequencing of polymerase chain reaction (PCR) products from bisulfite-treated genomic DNA.

Results:

Although GARP-transduced T_h cells exhibit constitutive FOXP3 expression and a regulatory phenotype, the FOXP3 TSDR is completely methylated as in naive T_h cells. GARP-mediated FOXP3 upregulation in T_h cells is not associated with T_reg-specific demethylation of the FOXP3 TSDR.

Conclusion:

Although GARP-engineered T_h cells exhibit stable FOXP3 expression and a phenotypic reprogramming towards T_reg cells in vitro, these cells do not completely mimic the epigenotype of natural T_reg cells. Thus, concepts based on the genetic modification of T_h cells as cellular therapies to treat autoimmune diseases or to control transplantation tolerance should be critically tested before any clinical application.

Introduction

The forkhead box protein P3 (FOXP3) is a transcription factor essential for the development and function of CD4+ CD25^hi regulatory T (T_reg) cells. FOXP3+ T_reg cells are an essential component for the maintenance of tolerance in transplantation and autoimmunity. Thus, isolation and expansion of natural Treg cells or their design via molecular engineering of antigen-specific T cells for regulatory T cell therapies represent a major focus for adoptive T cell transfer concepts in different disease settings [1, 2]. The molecular mechanism maintaining stable FOXP3 expression in Treg cells has been elucidated recently by the identification of an evolutionarily conserved region with CpG motifs within the non-coding part of the gene. This genomic region is completely unmethylated in Treg cells but methylated in naive and effector T cells and thus has been designated the Treg-specific demethylated region (TSDR) [3–6]. The significance of this epigenetic marker is obvious, because the methylation status of the FOXP3 gene is an even better marker of suppressor function than the differential expression of FOXP3 at the mRNA or protein level [7, 8]. Moreover, 4 CpG motifs within the TSDR have been identified as binding sites for essential transcription factors, such as CREB/ATF, NF-kB, and Ets-1. Ets-1 binds only to the unmethylated TSDR, a finding suggesting that restricted binding of a transcription factor complex controls stable FOXP3 expression in T_reg cells [9]. Therefore, the TSDR is an essential control region and marker of natural T_reg cells.

Further insights into T_reg-specific control mechanisms involved in the maintenance of sustained high levels of FOXP3 have been recently gained with the characterization of glycoprotein A repetitions predominant (GARP or LRRC32) [10–14] (reviewed in [2]). GARP is a T_reg-specific protein that is induced after T-cell receptor (TCR) activation and controls FOXP3 via a positive feedback loop [10, 12, 15]. Moreover, GARP functions as a cell surface receptor for latency-associated peptide (LAP)/latent transforming growth factor β (TGF-ß) [13, 14] (reviewed in [16]). T_reg-specific expression of GARP has been corroborated by the identification of a genomic region in intron 1 of GARP with a Treg cell-specific hypomethylation [17]. This finding suggests lineage-specific transcription of GARP in T_reg cells similar to FOXP3. More importantly, retroviral overexpression of GARP in helper T (Th) cells leads to an efficient and stable reprogramming of effector T cells into Treg cells, and this reprogramming is associated with the induction of the constitutive expression of FOXP3 and a regulatory phenotype [10, 12].

In the murine system, partial demethylation of the FOXP3 locus has been observed in CD4+ CD25- FoxP3- T cells activated in the presence of TGF-ß. This partial demethylation is associated with the upregulation of FoxP3, a finding suggesting that the epigenetic control of murine FoxP3 is mediated by external stimuli that manipulate T cell fate [3].

Recently it has been demonstrated that peroxisome proliferator-activated receptor α (PPARα) and PPARγ agonists, in combination with TGF-ß, induce demethylation of the FOXP3 locus in human peripheral CD4+ CD25– Th cells through a downregulation of DNA methyltransferases. This demethylation leads to the stable expression of FOXP3, which results in the generation of functional induced T_reg cells [18].

Since GARP represents a potential tool to engineer disease-associated Th cells to generate FOXP3-expressing regulatory T cells for regulatory T cell therapies [2], the critical risk-benefit assessment for such an approach remains whether stable regulatory functions of GARP-engineered Th cells observed in vitro would persist after adoptive T cell transfer in vivo. Based on the clear association of TSDR demethylation, which permits constitutive expression of FOXP3 and stable regulatory function of natural Treg cells, our hypothesis was that analysis of TSDR would enable such a risk-benefit assessment. For this reason we analyzed TSDR-methylation of GARP-transduced Th cells, characterized by Treg-like FOXP3 expression and stable regulatory function and FOXP3-transduced Th cells with Treg-like FOXP3 expression but only partial regulatory function compared to control-transduced Th cells and natural Treg cells. In the case of Treg-like TSDR epigenotype in GARP-transduced Th cells, the hypothesis of having guaranteed stable regulatory function of such cells in vivo would have been confirmed.

Material and Methods

T Cells

The T cells analyzed have recently been extensively described and characterized [12, 15, 19, 20]. In brief, CD4+ CD25– T_h cells and CD4+ CD25^hi T_reg cells were sorted from a healthy donor and stimulated weekly with the allogeneic Epstein-Barr virus (EBV) B-cell line LG2 EBV to generate alloreactive T_h cells (CD4–39) and T_reg cells (T_regHG and T_regTHU) [19]. Th cells were transduced with a retrovirus expressing either GARP or FOXP3 coupled with an IRES-driven green fluorescent protein (GFP) (T_hGARP and T_hFOXP3, respectively) and the empty IRES-GFP vector (T_hGFP). Transduced cells were sorted for GFP expression, cultured further in the same manner, and extensively characterized by functional, transcriptional, and protein expression analyses, as previously described [12, 15].

Immunostaining

For cell surface immunostaining, PE- and APC-conjugated monoclonal antibodies (mAbs) against CD25, CD27, CD83, and CD33 with respective isotype-control Abs were used (all from BD Bioscience, San Jose, CA, USA). For intracellular immunostaining, Alexa Fluor 467-conju-gated mAb against FOXP3 (206D) and respective isotype control (MOPC; BioLegend Inc., San Diego, CA, USA) were used according to the manufacturer's instructions.

DNA Isolation and Methylation Analysis

DNA isolation was performed with the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's guidelines. Bisulfite modification of DNA was performed according to an established protocol with minor modifications [21]. Genomic DNA (1–2 μg in 50 μl) was denatured for 15 min at 37 °C by adding 5.5 μl of 3 mol/l sodium hydroxide (NaOH). For complete denaturation, samples were incubated at 95 °C for 2 min and immediately cooled on ice. The bisulfite solution was freshly prepared by dissolving 4.25 g of sodium bisulfite (Sigma-Aldrich, St. Louis, MO, USA) in 7.5 ml H₂O. Next, 450 μl of 50 mmol/l hydrochinone solution was added, and the pH was adjusted to 5.15 by the addition of 500 μl of 10 mol/l NaOH. The denatured DNA solution was mixed with 1 ml of the bisulfite solution and incubated in the dark at 50 °C for 16–20 h. The DNA was recovered by using the Wizard DNA Clean-Up System (Promega Corp., Madison, WI, USA) followed by elution in 50 μl H₂O (65 °C). Subsequently, 5.5 μl of 3 mol/l NaOH was added, and the samples were incubated for 15 min at 37 °C. The solution was then neutralized by adding 55 μl of 6 mol/l ammonium acetate (NH₄OAc), pH 7.0. The DNA was precipitated with ethanol, washed in 70% ethanol, dried, and resuspended in 15 to 30 μl of water, depending on the initial DNA input. The regions of interest were analyzed by polymerase chain reaction (PCR) using primers that specifically amplify bisulfite-treated DNA of the TSDR of FOXP3 (Foxp3Amp5fw: CTT GCT TCC TGG CAC GAG TGT TTG GGG GTA GAG GAT TT and Foxp3Amp5rev: TGT AAA ACG ACG GCC AGT TAT CAC CCC ACC TAA ACC AA, modified from Baron et al. [22]. The later direct sequence analysis of PCR products was facilitated by the use of PCR tag primers containing additional nucleotides at the 5’ end that do not bind to the template and thus compensate for the missing base in PCR products from unmethylated samples during sequence analysis. PCR was performed with a GeneAmp 9700 system (Applied Biosystems, Foster City, CA, USA) in a total volume of 25 μl containing 3 μl of bisulfite-treated DNA, 0.2 mmol/l of each dNTP, 0.8 μmol/l of each primer, 2.5 μl of 10 × PCR buffer, 2.5 mmol/l magnesium chloride (MgCl₂), and 0.5 U Taq Polymerase (AmpliTaq Gold, Applied Biosystems). A touchdown protocol was adopted as follows [23]: after denaturation at 95 °C for 5 min, the annealing temperature was decreased by 0.5 °C every cycle from 63 °C to 56 °C, at which temperature 35 cycles were carried out. For all cycles, annealing was performed for 1 min, denaturation at 95 °C for 20 s, and extension at 72 °C for 1 min, followed by a final extension at 72 °C for 5 min. After agarose gel electrophoresis and purification of the PCR products with the MinElute™ Gel Extraction Kit (Qiagen), sequence analysis was performed with an ABI PRISM 3100 automated capillary genetic analyzer (Applied Biosystems, Carlsbad, CA, USA) using Big Dye 1.1 (ABI, Applied Biosystems) and a primer binding to the tag sequence. The sequence files were further analyzed with Geneious Pro 5.3.4. A cytosine was classified as fully methylated if the A signal was absent or very low at scattered positions, as unmethylated in the absence of a G signal, and as partially methylated if both signals were present.

Results and Discussion

We analyzed the methylation status of the FOXP3 TSDR in alloreactive Th cells that had been transduced with GARP (T_hGARP) or FOXP3 (T_hFOXP3) coupled to an IRES-driven GFP or GFP alone (T_hGFP) [12]. These cells differ in many aspects, as described recently [12]. In brief, T_hGARP cells show T_reg-like sustained FOXP3 expression (fig. 1A), an extended Treg-mRNA signature, and Treg-specific function (reviewed in [15]). In contrast, T_h cells transduced with GFP alone exhibited an effector T cell phenotype and limited FOXP3 expression. ThFOXP3 cells lie somewhere between these 2 extreme phenotypes, because they exhibit high levels of FOXP3 expression and an extended Treg-mRNA signature but also unstable or minor regulatory function.

Fig. 1.

A Fluorescence-activated cell sorting (FACS) analysis of alloantigen-specific T_h cells ectopically expressing GARP, FOXP3, or control GFP alone after retroviral transduction, as recently described [12]. Graphs are representative of intracellular expression of FOXP3 and cell surface expression of CD25, CD83, CD33, and CD27 after stimulation with mAbs against CD3/ CD28 and interleukin 2 (IL-2) for 3 days. B Methylation status of the same cells as in A of the complete TSDR of FOXP3 was analyzed. Shown is a representative part of the sequence analysis of bisulfite-treated DNA of the TSDR. CpG positions are highlighted with grey boxes. The G peak represents the methylated C, and the A peak represents the unmethylated C within the CpG. All CpGs are completely methylated in FOXP3- and GARP-transduced T_h cells but are largely unmethylated in the male Treg (TregHG) cells analyzed.

The methylation status of all CpGs of FOXP3 TSDR was analyzed in T_hGARP, T_hFOXP3, and T_hGFP cells and revealed a fully methylated TSDR in all cases (fig. 1 B). In contrast, Treg cells showed a typical largely unmethylated TSDR (fig. 1 B). Thus, neither the stable expression of FOXP3 nor the regulatory function of GARP-transduced Th cells in vitro depends on an unmethylated T_reg-like FOXP3 TSDR despite a phenotypic transdifferentiation towards T_reg cells. Because GARP transduction does not influence the epigenotype at the FOXP3 TSDR, which is regarded as a specific and reliable marker of natural Treg cells and their suppressor function [8], GARP-transduced Th cells do not perfectly mimic Treg cells, and thus our hypothesis that a TSDR epigenotype of GARP-transduced Th cells might guarantee their stable regulatory function in vivo has to be rejected.

Interestingly, in transdifferentiation studies of pluripotency, incomplete demethylation of genes has been shown to be characteristic of incompletely reprogrammed cells [24]. Therefore, FOXP3 TSDR methylation may be regarded as a marker of incomplete transdifferentiation of ThGARP cells. Consequently, concepts of GARP [10, 12] or FOXP3 transduction [25] of T_h cells for cellular therapies to treat autoimmune diseases or control transplantation tolerance [2] should be critically tested before any clinical application to ensure stable regulatory function before adoptive transfer of such engineered T cells and to exclude potential hazardous reversion of a potentially disease-controlling regulatory towards a diseases-aggravating effector phenotype. Despite the encouraging preclinical results in many mouse models using either FOXP3 itself [26, 27] or FOXP3-regulating genes like GPR83 [28, 29] and CD83 [30] to generate regulatory T cells for adoptive transfer, based on the known differences between FOXP3 expression of Treg cells in mice and man [31], critical risk-benefit assessment of such cellular therapies to ensure safe clinical application is of utmost importance.

In summary, ectopic expression of GARP in T_h cells can induce sustained FOXP3 expression and a stable regulatory phenotype in vitro similar to that of T_reg cells. However, this phenotypic transformation is not reflected by a Treg-specific demethylation of the TSDR raising serious questions about the stability of such engineered T cells for cellular therapies to treat autoimmune diseases or control transplantation tolerance in vivo.

Acknowledgments

This work was supported in part by grants from the Mercator-Stiftung.

Disclosure Statement

The authors declared no conflict of interest.