Human Colonic Myofibroblasts Promote Expansion of CD4+ CD25^high Foxp3+ Regulatory T Cells

Abstract

Background & Aims

Regulatory T (Treg) cells (CD4+ CD25^high FoxP3+) regulate mucosal tolerance; their adoptive transfer prevents or reduces symptoms of colitis in mouse models of inflammatory bowel disease (IBD). T-cell functions are regulated by mesenchymal cells. Colonic CD90⁺ mesenchymal myofibroblasts and fibroblasts (CMFs) are abundant, non-professional antigen presenting cells in the normal human colonic mucosa that suppress proliferation of activated CD4⁺ effector T cells. We studied CMF suppressive capacity and evaluated the ability of CMF to induce Treg cells.

Methods

Allogeneic co-cultures of CD4⁺ T cells and CMFs, derived from normal mucosa of patients undergoing colectomy for colon cancer or inflamed colonic tissues from patients with ulcerative colitis or Crohn’s disease, were used to assess activation of the Treg cells.

Results

Co-culture of normal CMF with resting or naive CD4⁺ T cells led to development of cells with a Treg phenotype; it also induced proliferation of a CD25⁺ CD127⁻ FoxP3⁺ T cells, which expressed CTLA-4, interleukin (IL)-10, and transforming growth factor-β and had suppressive activities. In contrast to dendritic cells, normal CMFs required exogenous IL-2 to induce proliferation of naturally occurring Treg cells. Induction of Treg cells in normal CMFs required MHC class II and prostaglandin E2. CMFs from patients with IBDs had reduced capacity to induce active Treg cells and increased capacity to transiently generate CD4⁺CD25^+/− CD127⁺ T cells that express low levels of FoxP3.

Conclusions

CMFs suppress the immune response in normal colon tissue and might therefore help maintain colonic mucosal tolerance. Alterations in CMF induction of Treg cells might promote pathogenesis of IBDs.

Introduction

Gut immune responses are normally regulated to maintain a state of mucosal tolerance, which represents a balance between the need to mount protective immunity toward pathogens while not activating damaging inflammatory responses to innocuous luminal antigens (Ag)¹. Recent studies suggested that a functionally distinct subset of regulatory T cells, CD4⁺ CD25^high FoxP3⁺ cells (T_reg), is actively involved in the maintenance of immunological tolerance in the gastro-intestinal (GI) tract^2–3.

There are at least two types of the T_regs: naturally occurring (_nTreg) and inducible (iT_reg). nT_reg cells develop in the thymus and the majority expresses high levels of cell surface markers associated with an activated/memory T cell phenotype^2–4. iT_reg cells are elicited in the periphery from naïve T cells under low-dose antigenic stimulation or by several cytokines, although the exact mechanism(s) of iT_reg cell induction remains unknown⁴. Forkhead transcription factor 3, Foxp3, is a master switch that controls T_reg cell development and function. Both types of T_regs down regulate immune responses to foreign and self-Ag, and contribute to the suppression of autoimmune disorders^3–5. T_reg exert their effects by multiple mechanisms (e.g., cytokine deprivation, CTLA-4 signaling, IL-10 or TGB-β-production)^4–5. It is not established which of these mechanisms is predominant, but each might have a greater importance under specific physiological/pathological conditions.

Recent studies suggest that T_reg cells play a major role in murine models of inflammatory bowel disease (IBD). Adoptive transfer of CD4⁺CD45RB^high T cells depleted of the T_reg cells into immunodeficient mice resulted in development of IBD-like chronic colitis^7–8. In contrast, transfer of Treg cells into animals with chronic colitis led to the amelioration of the disease^7–8. Adecrease in T_reg number in active IBD in humans has been documented⁹. Despite advances in understanding the role of the T_reg cells in the maintenance of peripheral tolerance in the intestinal mucosa, the mechanisms involved in their regulation in the intestine and what factors contribute to the alteration of the T_reg numbers in IBD remains unclear. It has been recently reported that small intestinal GALT associated CD103⁺ dendritic cells (DCs), as well as small bowel epithelial cells can induce T_regs^10–11. However, the origin and possibility of the expansion of the T_regs in the colon remain obscure.

We recently reported that colonic myofibroblasts/fibroblasts (CMFs) are novel non professional APCs and are abundant in the normal human colonic mucosa¹². CMFs are a distinct population of mesenchymal stromal cells that are positive for CD90. Activated CMFs express α-smooth muscle actin (α-SMA, myofibroblast marker), but are negative for other cell surface markers that define conventional APCs. CMFs form a network throughout the colonic lamina propria and have been implicated in the regulation of mucosal inflammation^12–16. Recent data from our laboratory indicates that CMFs may have a dual regulatory role on CD4⁺ T cell activity. They induce proliferation of resting CD4⁺ T cells in a MHC class II-dependent manner¹², but suppressing proliferation of activated CD4⁺ effector T cells via mechanisms involving the B7-related co-inhibitors PD-L1 and PD-L2¹⁵. Importantly, it has been recently reported that intestinal myofibroblasts taken from chronically inflamed tissues display a fundamentally altered phenotype compared to their counterparts extracted from normal tissues at the same anatomical site^13,16. CMFs are actively involved in the progression of the IBD associated inflammation via altered deposition of matrix, upregulation of proinflammatory cytokine production, and changes in the array of secreted, soluble immunoregulatory molecules and mediators^13,16.

In this study, we have further analyzed interactions between CMFs and CD4⁺ T cells to determine if CMFs stimulate expansion of the T_reg cells. We found that CMFs contribute to the maintenance of FoxP3⁺ phenotype of the nT_reg and induce generation of iT_reg cells from naïve CD4⁺ T cells via mechanisms that involve MHC class II and PGE₂ signals. Moreover, we demonstrated that IBD-derived CMFs, when compared to normal CMFs, have a decreased ability to induce cells bearing the T_reg phenotype. Thus, our data suggests that CMFs might play a prominent role in mucosal tolerance via regulation of the Treg number, and disruption of CMF mediated regulation of the T_reg may contribute to IBD progression.

Materials & Methods

Antibodies and Reagents

Please see Supplemental information online at www.gastrojournal.org

Human colonic tissue and primary CMF culture

For CMF isolation, fresh human colonic mucosal sections were obtained from discarded surgical tissue in compliance with protocols approved by the UTMB Institutional Review Board. Areas of uninvolved colonic tissue from patients undergoing colectomy for colon cancer were used as the source for normal CMFs (N-CMFs). CMFs were also isolated from inflamed colonic tissues from patients with ulcerative colitis (UC) or Crohn’s disease (CD) undergoing colonic resections for IBD (IBD-CMFs, CD-CMFs or UC-CMFs). Primary cultures of CMFs were generated according to the method described by Mahida et al.¹⁷ and routinely used in our laboratory as previously described^12,15–16.

Generation of human DCs from peripheral blood mononuclear cells

Please see Supplemental information online at www.gastrojournal.org

CMF: T cell allogeneic cocultures

Unless otherwise indicated, peripheral blood mononuclear cells (PBMC) were prepared from the blood of healthy donors by density gradient centrifugation over Ficoll-Paque^™ Plus according to the manufacturer’s instructions. Naïve, resting and regulatory human CD4⁺ T cells were purified from these PBMC using commercially available kits that negatively select CD4⁺ CD45RA⁺ T cell, human CD4⁺ resting T cell and positively select CD4⁺ CD25⁺ regulatory T cells, respectively (Miltenyi Biotec., Auburn, CA). The purity of isolated T cells (>98%) was confirmed by flow cytometry. CMF primary cultures were stimulated with IFN-γ (100 U/ml) for 7 days prior to study in order to induce optimal MHC class II expression as previously described¹². Cells were then rested for at least 24 h prior to use in the experiments. CMFs or freshly generated DC (obtained as described above) were cocultured with allogeneic T cells at a ratio 10:1, respectively, and incubated for 5–7 days at 37°C with 5% CO₂. In some CFSE proliferation experiments, human recombinant IL-2 (eBioscience) was added to the CMFs: T co-cultures at a concentration of 10 ng/mL.

Flow cytometry

Please see Supplemental information online at www.gastrojournal.org

IL-10, TGF-β and PGE₂ ELISA

Conditioned media were collected from the CMF: T cell, CMF or T cell culture wells at 24h, 5 and 7 days and analyzed for the production of the IL-10 and TGF-β using ELISA kits (BD Bioscience) and PGE₂ using a R&D Systems kit, according to the manufacturer’s instructions.

Real Time RT-PCR

The real time RT-PCR was carried out as described previously¹⁵. Please see Supplemental information online at www.gastrojournal.org.

CFSE proliferation assays

Please see Supplemental information online at www.gastrojournal.org

T Cell Suppression Assays

The T cell suppression experiments were performed as described previously²⁴. For details, please see Supplemental information online at www.gastrojournal.org.

Statistical analysis

Unless otherwise indicated, the results were expressed as the mean ± SE of data obtained from at least three independent experiments done with triplicate sets in each experiment. Differences between means were evaluated by ANOVA using Student’s t-test for multiple comparisons. Values of P <0.05 were considered statistically significant.

Results

CMFs stabilize FoxP3 expression in nT_reg and induce their proliferation in presence of IL-2

We showed previously that CMFs induce proliferation of resting CD4⁺ T cells isolated from peripheral blood ¹², which is also known to contain CD4⁺ CD25^high FoxP3⁺ nT_reg cells (nT_reg). Thus, we investigated the interaction of the T_reg and CMFs isolated from normal colonic mucosa (N-CMFs).

Previously, we reported that, in culture, MHC class II expression by CMF drastically decreases when compare to that on acutely isolated cells and the high levels demonstrated in situ¹². Thus, in all experiments primary CMFs were stimulated with IFN-γ (100 U/ml) prior to use in order to restore optimal MHC class II expression as described in the Methods. Theability of N-CMFS to induce generation of T_reg in seven day allogeneic co-cultures of the CMFs with CFSE-labeled resting CD4⁺ T cells were studied. A significant increase in the percentage of the CD25^highFoxP3⁺ T cells in the dividing fraction of CD4⁺ T cells co-cultured with N-CMFs was observed (Gate P3, Figure 1A) and represented ~31.4 ± 5.8 % of the dividing T cells (Figure S1, see supplement online at www.gastrojournal.org). This coincides with increased expression of the suppressive cytokines IL-10 and TGF-β1 by T cells co-cultured with N-CMFs (Figure 1B). The majority of proliferating CD4⁺CD25^high T cells derived from CMFs-T cell co-cultures that were positive for FoxP3 did not express CD127, the IL-7 α chain receptor, and, thus, correspond to the true T_reg phenotype (Figure 1B). A moderate increase in of the FoxP3⁺CD127⁺ T cell fraction corresponding to the FoxP3 transiently expressing CD4⁺ effector T cells was also noted in the CMF-T cell co-cultures (Figure 1B). In contrast to T_reg cells the expression of FoxP3 by T effector cells reported to be low, and was not sufficient to suppress expression of CD127 maker and increase the production of suppressive cytokines produced by the T_reg¹⁸.

Figure 1.

Normal (N) CMFs contribute to the maintenance of nT_reg phenotype. CFSE-labeled resting CD4⁺ T cells were cultured without or with allogeneic N-CMFs at a ratio 1:10 for 7 days in 24 well plates. T cell from these co-cultures were subjected to surface CD4 and CD25, CD127 and intracellular FoxP3 staining following by flow cytometry. (A) Live events were gated in P1 and percentage of CD25+FoxP3+ T cells in non dividing (P3) and dividing (P4) fractions of CD4⁺ resting T cells growing in monoculture or cocultured with N-CMFs was evaluated. A representative experiment is shown (n=5 allogeneic donor pair, two experimental replicate each) (B) Distribution of CD127 and FoxP3 (FACS analysis) and expression of TGFβ1 and IL-10 (real-time RT-PCR) in the CD4⁺ resting T cells growing in monoculture or cocultured with N-CMFs was evaluated. A representative experiment is shown (n=4 allogeneic donor pair, two experiment replicate each) (D) The surface CD25 and intracellular FoxP3 expression in the freshly isolated and seven day monoculture of nTreg, and N-CMF primed nTreg was analyzed. A representative experiment is shown (n=5 allogeneic donor pair, two experiment replicate each). (D) Percentage of the dividing CD4⁺CD25^highCd127⁻FoxP3⁺ nT_reg in response to the allogeneic stimulation with N-CMF in presence/absence of IL-2 (10 ng/mL) or BM-derived DC was measured in 7 day co-culture as described above. Results are calculated as the mean value of the percentage of the dividing cells for three independent allogeneic pair of N-CMFs and CD4⁺ T cells healthy donors ± SD. Each assay was conducted in duplicate. * p < 0.05.

Next, we analyzed how N-CMFs affect FoxP3 expression and proliferation of nTreg purified from peripheral mononuclear cells. When purified nT_reg were cultured alone, their FoxP3 expression was reduced, whereas those in co-culture with N-CMFs maintained FoxP3 expression (Figure 1C). Analysis of purified nT_reg induced by CMFs demonstrated that, in contrast to classical APCs such as BM-derived DCs, co-culturing of N-CMFs with nT_reg did not induce significant proliferation of nT_reg cells (Figure 1D). IL-2 is reported to be essential for the physiological expansion of nT_reg in humans and rodents^19–20. Thus, we analyzed whether addition of IL-2 to the N-CMFs-nT_reg co-cultures resulted in proliferation of the nT_reg. Figure 1D demonstrates that addition of IL-2 to these co-cultures resulted in strong proliferation of the nT_reg comparable to that induced by BM-derived DCs.

CMFs induce generation of iT_reg cells from naïve CD4⁺ CD45RA⁺ T cells

Next, we sought to determine the capacity of CMFs to generate iT_reg cells from naïve CD4⁺ CD45RA⁺ T (Th₀) cells. Allogeneic Th₀ cells were incubated with MHC class II expressing CMFs at ratios of 10:1 for up to 12 days. A significant increase in the expression of FoxP3 mRNA was observed at day seven of the co-cultures (Figure 2A) and in the frequency of the iT_reg cells (Figure 2B). The percentage of the induction of FoxP3⁺ T cells in CMF primed Th₀ cells (10.2±3.6%) was comparable to that in the cocultures with DCs (12.3 +− 3.9 %) (Figure 2B). The iT_regs generated from CMFs-primed Th₀ cells did not express surface CD127 (Figure 3A). Since downregulation of the CD127 is associated with acquisition of regulatory function by T cells and inversely correlated with FoxP3 expression^21–23, our data suggests that normal CMFs can contribute to the generation of the iT_reg from naive CD4⁺ T cells at a capacity comparable to that of classical APCs, such as DC.

Figure 2.

CD4⁺ CD25^high FoxP3⁺ T_reg cells (iT_reg) are induced from naive CD4⁺CD45RA⁺ T helper (Th₀) cells primed with N-CMFs. Th₀ cells were cultured without or with allogeneic N-CMFs at a ratio 1:10 for one week. (A) The RNA from the N-CMF primed Th₀ cells was analyzed for FoxP3 mRNA expression using real time RT-PCR. The mRNA levels for FoxP3 was normalized to 18S. Data represent mean of mRNA Δ fold increase ± SEfrom duplicates in four experiments (n=8) *P<0.01. (B) The surface CD25 and intracellular FoxP3 expression in the allogeneic N-CMF or BM-derived DC primed CD4⁺ T cells was analyzed using flow cytometry.. The appropriate isotype controls were included in the experiments. A representative experiment is shown (n=5). *p<0.01.

Figure 3.

CD4⁺ CD25^high FoxP3⁺ iTreg induced by allogeneic N-CMFs exhibit an activated phenotype. Th₀ cells were cultured with allogeneic N-CMFs at ratio 1:10. Harvested T cells (day 7) were stained with anti-CD4 FITC, anti-CTLA-PE, anti-CD25PE-Cy7 and anti-CD127 mAbs or appropriate isotype controls. Live events were gated in P1, then (A) CD4⁺ T cells were gated in P2 and analyzed for CD25 and CD127 expression. The CD4⁺ T cells bearing CD25⁺CD127⁻ phenotype (Q4 quadrant) were (B) analyzed for CTLA-4 expression and sorted for further analysis. A representative experiment is shown (n=5). The RNA was extracted from sorted CD4⁺ CD25⁺CD127⁻ T cells was analyzed for (C) FoxP3, TGF-β1 and IL10 mRNA expression using real time RT-PCR. The mRNA levels for each gene of interest were normalized to 18S. Data represent mRNA Δ fold increase mean ± standard errors from duplicates in five experiments (n=10) *p<0.01.

The T_reg cells induced by CMFs exhibit an activated phenotype

Increased expression of CTLA-4 is associated with an activated phenotype of T_reg cells and is involved in T_reg suppressive function^4–5. Thus, we analyzed whether CMF-induced iT_reg cells express CTLA-4. The CD4⁺ CD25^high CD127⁻ T cells sorted from the CMFs:Th₀ cell co-cultures. (Figure 3A) demonstrated that the cells express both intracellular CTLA-4 (Figure 3B) and FoxP3 mRNA (Figure 3C). Since production of IL-10 and TGF-β1 by T_reg cells is central to their suppressive function^4–5, expression of these cytokines was measured by real-time RT-PCR in CD4⁺ CD25^high CD127⁻ T cells (iT_reg) sorted from the seven day Th₀:CMFs cell co-cultures. CMF-induced iT_reg cells express moderate levels of the IL-10 and high levels of TGF-β1 (Figure 3C), but not TGF-β2 and TGF-β3 (data not shown). Taken together these results indicate that the CMF-induced iT_reg cells have the elements needed to exert suppressive function.

The iT_reg cells induced by CMFs inhibit activated T effector cell proliferation

The main function of T_reg cells is to negatively regulate responses of immune cells, including activated T effector cells^2,4–5. Thus, we analyzed the ability of CMF-induced T_reg to suppress proliferation of CD3/CD28 preactivated syngeneic CD45RA⁺CD4⁺ naïve T cells (T effector, T_eff). T cells from the seven day CMF:Th₀ cell cocultures were sorted into CD4⁺ CD25^high and CD4⁺CD25^low populations (Figure 4A). These populations were incubated separately for 96 h with syngeneic, preactivated T_eff in various T_reg:T_eff ratios. The CMF-induced CD4⁺ CD25^high T cells induced a significant (60%) decrease in the proliferation of preactivated CD45RA⁺CD4⁺ naïve T helper cells (Figure 4B) at a 1: 8 T_reg:T_eff ratio, which was comparable with the percentage of suppression by circulating CD4⁺ CD25^high T cells isolated from PBMC of the same donors (Figure 4C). Thus, our data demonstrated that CMFs induce functionally active iT_reg. Surprisingly, a minor level of suppression by CMF-induced CD4⁺ CD25^low T cell fraction has been also observed (Figure 4C), perhaps due to the production of the immunoregulatory cytokines by CD4⁺CD25^low T cell fraction sorted from CMF-CD4⁺ T cell co-cultures.

Figure 4.

CD4⁺CD25^high iT_reg cells induced by N-CMFs suppress proliferation of syngeneic, activated CD4⁺ T effector cells. (A) CD4⁺CD25^high T cells induced by co-culturing of Th₀ with CMFs for 7 days were sorted by FACSAria sorter and CD4⁺CD25^high or CD4⁺CD25⁻T cells were co-cultured with CD2/CD3/CD28-activated CD45RA⁺CD4⁺CD25⁻ effector (Teff) cells for 4 days. Eighteen hours prior to the end of the co-culture, the cells were pulsed with [3H]-methyl-thymidine. (B) Mean counts per minute (c.p.m) ± standard error (SE) of triplicate cultures of CD4⁺ T cells isolated from one donor are shown here in a representative experiment of three. (C) The percent suppression of Teff is shown and represents a decrease in the proliferative response (c.p.m.) of CD4⁺ Teff cells when compared to activated Teff cells alone. The means are shown as the results of triplicates in 3 experiments, n=9 (* p < 0.05).

De novo generation of CD4⁺ CD25^high FoxP3⁺ T_reg cells by CMFs is MHC class II- and PGE₂- dependent

We have previously shown that MHC class II is essential for CMF-induced resting CD4⁺ T cell proliferation¹². Others have demonstrated that MHC class II-TcR interactions are involved in the development of the FoxP3⁺ polyclonal T_reg¹¹. Thus, we examined the role of MHC class II expression by human CMFs in the induction of iT_reg cells. Allogeneic co-cultures of CMFs with Th₀ cells were established in the presence of the anti-class II MHC antibodies (anti- HLA-DR, -DQ and -DP cocktail, clones L243 and IVA12). A significant decrease in FoxP3 mRNA expression and CMF induction of iT_reg cells in the CMF:T_reg cell co-cultures (up to 50%) was observed in the presence of anti-MHC class II Abs, but not isotype controls (Figure 5A–B). The fact that CMF-induced iT_regs were only partially reduced in the presence of anti-MHC class II Abs suggested that this induction requires other mediators as well.

Figure 5.

Induction of the iT_reg from naïve CD4⁺CD45RA⁺ T cells (Th₀) by allogeneic N-CMFs depends on MHC class II. T cells were cultured without or with allogeneic CMFs at a ratio 1:10 for 7 days in 24 well plates in presence/absence of anti-MHC class II mAb mix (anti- HLA-DR, -DQ and -DP cocktail, clones L243 and IVA12) or isotype mix control. (A) The RNA from the N-CMF primed T cells was analyzed for FoxP3 mRNA expression by using real time RT-PCR. The mRNA level for FoxP3 was normalized to 18S. Data represent mean of mRNA Δ fold increase ± SEfrom duplicates in three experiments (n=6) *P<0.01. (B) The surface CD25 and intracellular FoxP3 expression in the N-CMF primed T cells was analyzed using flow cytometry. Harvested T cells were stained with anti-CD4 FITC, anti-CD25PE-Cy7, and anti-FoxP3-APC mAbs. The appropriate isotype controls were included in the experiments. A representative experiment is shown (n=5). *p<0.01.

Since it has been demonstrated that peripheral induction of regulatory T cells may involve PGE₂²⁴ and because CMFs are avid producers of these molecules ^16–17, we investigated involvement of this mediator in CMF-induced generation of the T_regs. Addition of a PGE₂ synthesis inhibitor indomethacin (INT 10 μM) to the CMF: Th₀ cell co-culture decreased in the induction of iT_regs (5.92±1.8 %), when compared to 10.86±2.6% in the controls (Figure 6A). Additionally, a strong upregulation of indomethacin-inhibitable PGE₂ production was observed in the CMF: naïve T helper cell co-culture (Figure 6B). Moreover, a significant decrease of the FoxP3 mRNA expression in the presence of INT was observed (Figure 6C). Taken together these data suggest the importance of the MHC class II- and PGE₂ mediated signals in the colonic induction of the T_reg cells.

Figure 6.

Induction of iT_reg from naïve CD4⁺CD45RA⁺ T (Th₀) cells by allogeneic N-CMFs involves PGE₂. T cells were cultured without or with allogeneic N-CMFs at a ratio 1:10 for 7 days in 24 well plates in presence/absence of indomethacin, 10 μM. (A) The surface CD25 and intracellular FoxP3 expression in the allogeneic N-CMF primed T cells was analyzed using flow cytometry. The appropriate isotype controls were included in the experiments. A representative experiment is shown (n=6). *P<0.05. (B) The PGE2 production in culture supernatant was measured using ELISA. Data represent mean ± SEfrom duplicates in three experiments (n=6). (C) The RNA from the N-CMF primed T cells were analyzed for FoxP3 mRNA expression using real time RT-PCR. The mRNA level for FoxP3 was normalized to 18S. Data represent mean of mRNA Δ fold increase ± SEfrom duplicates in three experiments (n=6). * P< 0.05.

IBD-derived CMFs have a decreased capacity to induce T_reg cells

CMFs are important contributors to the progression of IBD and, when isolated from chronically inflamed IBD tissues, they display a fundamentally altered phenotype/function compared with cells extracted from normal tissues^13,16. Thus, we compared the capacity of IBD- and N- CMFs to induce a T_reg cell phenotype from Th₀ cells. Allogeneic Th₀ cells were co-cultured with MHC class II expressing CD-, UC- or N-CMFs at a ratio of 10:1 for up to 12 days. No significant difference in the generation of the FoxP3⁺CD25^high cells by CMF derived from IBD vs those derived from normal controls was observed (Figure S2 at www.gastrojournal.org). IL-7 receptor α-chain or CD127 has been shown to be inversely correlated with the suppressive activity of the CD4⁺CD25^high T cells^21–22. Moreover, IL-7/IL-7R signaling has been suggested to be implicated in the development and persistence of murine chronic colitis²³. Thus, we analyzed whether the combination of CD127 with CD4, FoxP3 and CD25 identified by multi-color flow cytometry analysis would lead to the identification of a difference in the generation of the T _reg by CMFs derived from IBD compared to those isolated from normal controls (Figure 7A and B). A decrease in the generation of the CD4⁺CD127⁻FoxP3⁺ T cells and increase CD4⁺CD127⁺FoxP3⁺ T was observed when Th₀ cells were primed with IBD-derived CMFs when compare to N-CMFs (Figure 7A). Up to 50% of IBD-induced FoxP3⁺ T cells were positive for CD127. A significant decrease in the induction of the true T cells bearing the true T_reg phenotype was observed (a.k.a. FoxP3⁺ CD25^high, but negative for the CD127) as shown on Figure 7B. Instead we found increased generation of CD4⁺CD25^+/− FoxP3⁺ cells that were positive for CD127 when Th₀ cells were primed by IBD-derived CMFs. T cells derived from the co-culture of IBD-CMFs with Th₀ cells had reduced expression of TGF-β1 and did not produce IL-10, cytokines associated with active T_reg (Figure 7C). While the mechanisms implicated in this impairment of IBD-derive CMF are currently unknown, our data suggest that in contrast to normal CMFs, IBD-derived CMFs have diminished capacity to induce “true” T_reg phenotype generation from Th₀ cells.

Figure 7.

IBD-derived CMFs have reduced capacity to induce the T_reg cells from CD4⁺CD45RA⁺ T helper (Th₀) cells and promote the expression of FoxP3 by CD127⁺ T cells. Th₀ cells were cultured without or with allogeneic CMFs at a ratio 1:10 up to 12 days in 24 well plates. (A) The surface CD127 and intracellular FoxP3 expression in the N-, CD or UC-CMF primed CD4⁺ T cells was analyzed using flow cytometry. The appropriate isotype controls were included in the experiments. A representative experiment is shown (n=9). (B) Distribution of the iTreg (CD4⁺CD25^highCD127⁻FoxP3⁺) and transitory FoxP3 expressing cells (CD4⁺CD25^+/−CD127⁺FoxP3⁺) in the N-, CD or UC -primed T cells. The results are shown as mean ± standard errors from nine independent experiments. The ANOVA one way variance analysis combined with Bonferroni’s multiple comparison post test was used to analyze the significance of the variation between the experimental groups. *A significant decrease (p<0.05) in the generation of the iTreg (CD4⁺CD25^highCD127⁻FoxP3⁺) by CD- and UC-CMFs when compare to N-CMF was observed. **A significant increase (p<0.05) in the generation of the CD4⁺CD25^+/−CD127⁺FoxP3⁺ by CD- and UC-CMFs when compare to N-CMF was observed. (C) TGF-β1 and IL-10 mRNA expression by T cells derived from the N-, CD or UC-CMF primed T cells was analyzed using relative quantitative real time RT-PCR. The mRNA levels for TGF-β1 and IL-10 was normalized to 18S. Data represent mean of mRNA Δ fold increase ± SEfrom duplicates in four experiments (n=8). The Bonferroni-corrected ANOVA variance analysis revealed a significant increase in TGF-β1 and IL-10 production when Th0 cells were primed by N-CMFs (*p<0.001 and **p<0.05, respectively), but not when the Th0 cells were primed with CD or UC-CMFs.

Discussion

APCs play a critical role in maintaining the balance between tolerance and inflammation in the gut^1,10–12. Role of professional APCs, such as DCs and macrophages, in the differentiation of T cells and regulation of their activity has been well investigated^1,10–12,25. Little is known about the role of intestinal stromal cells in these processes. It has been established that stromal cells are important contributors to immune homeostasis and function^{12–16,26–27}. Despite the fact that the term “stromal” encompasses a broad range of stationary cells with various developmental origins, the CD90⁺ fibroblast-like cells which are replenished in part from bone marrow mesenchymal stem cells form a major stromal component that is ubiquitously found in a close association with T cells in peripheral lymphoid organs, including mesenteric lymph nodes^26,27.

We recently reported that human colonic stromal CD90⁺ myofibroblast/fibroblast cells are abundant in the normal human colonic lamina propria and can act as non professional APCs¹². Low level expression of B7.1/2 molecules and relatively higher expression of B7-related co-inhibitor B7-H1 (PD-L1) and B7-DC(PD_L2)¹⁵ led us to hypothesize that CMFs are local “suppressors” of activated T cell responses in normal colon contributing to mucosal tolerance. The present study supports our hypothesis and suggests that in the normal colon, CMFs can also indirectly contribute to the suppression of active inflammation by supporting expansion of the T_reg cells.

As has been previously reported for the mesenchymal stem cells²⁸ (probable progenitor cells for CMFs), our experiments indicate that N-CMFs may contribute to the maintenance of FoxP3⁺ phenotype of nT_reg. Although the exact mechanisms involved in CMF’s beneficial effect on the maintenance of nTreg phenotype are still to be determined, an effect on both survival and FoxP3 expression stability is likely. Stromal cells are known to produce IL-7, Il-15, retinoic acid and TGFβ^16,29 and these mediators contribute to T_reg maintenance^4,16,29. We demonstrated that CMFs can support nT_reg proliferation, but in contrast to professional APCs such as DCs, N-CMFs require the presence of exogenous IL-2. IL-2 is essential for the physiological expansion the nTreg cells in humans and rodents. TCR engagement on the nT_reg cells in combination with IL-2 was shown to be sufficient to overcome anergic properties of these cells resulting in the nT_reg proliferation^19–20,30. Activated CD4⁺ T effector cells were demonstrated to contribute to the proliferation/survival of nT_reg by secreting IL-2 in vitro and in vivo³¹. Neutralization of the IL-2 reduces nT_reg numbers in vivo²⁰.

While this is the first report demonstrating that human normal colonic CD90⁺ stromal cells support nT_reg expansion, it is not without precedent. Other mesenchymal cells such as dermal fibroblasts³² and rheumatoid synovial fibroblasts²⁹ have been shown to induce proliferation of nT_reg by an IL-15-dependent mechanism. Further, other members of the fibroblast family such as hepatic stellate cells have been shown to be capable of expanding nTreg when supplemented with IL-2³³.

Our data indicate that priming of naïve CD4⁺ T cells with allogeneic N-CMFs can induce generation of iT_reg cells at a rate comparable with BM-derived DCs. Although FoxP3 expression is mainly associated with regulatory T cell phenotype, FoxP3 may possibly be transiently expressed by activated CD4⁺ T cells³⁴. We demonstrate here that FoxP3⁺ iT_reg induced by human CMFs have an activated phenotype with strong expression of CTLA-4, TGF-β1 and IL-10, and are capable of suppressing the proliferation of activated syngeneic effector T cells. In contrast to rodent studies, it has been difficult to convert human peripheral blood derived naïve CD4⁺ T cells into FoxP3⁺ iT_reg cells with potent and stable suppressive ability; only transient FoxP3 expression in the TCR activated human naïve T cells occurred in the presence of TGFβ1³⁵. Further, this type of stimulation was not sufficient to confer to these FoxP3⁺ cells significant anergic and suppressive capacities. More recently it have been reported that the presence of additional factors, such as trans-retinoic acid, PGE₂, PD-L1 signaling, might be required for the generation of active iTreg with suppressive capacity^{10,24,36–387}. All of these factors are known to be produce by human fibroblasts/myofibroblasts^14–16. We demonstrate herein that induction of iT_reg from N-CMF primed naive CD4⁺ T cells depends on both cell-contact mediated interactions (MHC class II-TCR signaling) and production of a soluble factor (PGE₂). PGE₂ reported topromote the conversion of CD4⁺CD25⁻ cells to T_reg cells²⁴. However, to our knowledge, this is the first report suggesting the importance of the CMF-derived PGE₂ in the regulation of T_reg cells and, consequently, effector T cell behavior in the colon. It is not clear which of the various potential iT_reg induction mechanisms is of greater importance under specific physiological conditions, and this will be an important question for future.

It has been suggested that fibroblast-like stromal cells can modify the quality, quantity and duration of inflammatory responses³⁸. The direct immunosuppressive, PD-L1-mediated effect of CMFs has been previously demonstrated by us¹⁵. It is still unclear whether PD-L1 or T_reg mediated immune suppressive mechanisms predominate during a particular inflammatory event. However, it is likely that disruptions in these anti-inflammatory, immunoregulatory functions may prevent the proper transition from acute to resolving inflammation leading to the establishment of chronic inflammation. There is ample substantiation that intestinal fibroblasts/myofibroblasts taken from the chronic disease tissues display a fundamentally altered phenotype compared with these cells from normal tissues at the same anatomical site: changes in TNF-α, TGFβ and MMP production were identified in the IBD-derived myofibroblasts and it is thought that these cells contribute to the IBD-associated fibrosis and disruption of immunoregulation^39–41. Here we present additional evidence that CMFs may play a role in immunopathogenesis of IBD: IBD-derived CMFs have a decreased capacity to induce active T_reg cells capable of producing suppressive cytokines TGFβ1 and IL-10. Moreover, a shift in the balance in IBD-CMF – primed T cells from “true” T_reg phenotype (CD4⁺CD25^highCD127⁻ FOXP3⁺) toward FoxP3 expressing CD127⁺ T cells (CD4⁺CD25^highCD127⁺ FOXP3⁺) suggests that during IBD progression CMFs may promote the persistence of CD127⁺ colitogenic CD4⁺ T cells that only transiently express FoxP3. It is unclear how the disruption of this immunoregulatory function of the CMFs occurs during the switch from acute to chronic inflammation associated with IBD, but in rheumatoid arthritis an alteration in synovial fibroblasts capacity to regulate T responder/T_reg balance has been recently reported to implicate IL-15 expression by the disease derived fibroblasts²⁹. Further studies are clearly necessary to understand the role of the CMF-diminished capacity to induce T_reg in the progression of IBD.

Our study adds human CMFs to the list of mesenchymal stromal cells (e.g., dermal and synovial fibroblasts, hepatic stellate cells and mesenchymal stem cells) ^40–42,45, that may contribute to the regulation of the T_reg cells and, thus, to the maintenance of peripheral tolerance. Further, CMFs appear to be local contributors to the maintenance of mucosal tolerance in the colon via at least two independent mechanisms: (1) expression of co-inhibitors of the B7 family (reference) and (2) support of T_reg cells expansion. These findings taken together with previous reports^12,15 indicate that CMF are relevant to normal inflammatory responses and that alteration of their regulatory function might be involved in IBD immunopathogenesis.