Levels of Foxp3 in Regulatory T Cells Reflect Their Functional Status in Transplantation¹

Immune-mediated rejection remains the single most important cause of graft failure in all forms of tissue transplantation. However, spontaneous development of transplant tolerance in some recipients indicates that, just as rejection is mediated by active immune-mediated events, graft acceptance is also sustained by active immunoregulatory mechanisms. One such immune mechanism is mediated by regulatory T cells (Tregs),³ that have the capacity to suppress host immunity against transplants. Among the various forms of Tregs, CD4⁺CD25⁺ Tregs, which constitute 5–10% of CD4⁺ T cells, play a central role in inducing and maintaining tolerance to both self- and alloantigens (1, 2, 3).

Tregs are exemplified by their expression of transcription factor Foxp3 (4) and have been reported in a number of clinical and animal models of transplantation (5, 6, 7, 8). Despite significant advances in understanding Treg biology, several aspects relevant to transplant settings remain unclear. The majority of previous reports have associated the outcome of the allograft primarily with changes in the frequency of Tregs, either in the lymphoid compartment or at the graft site (9, 10). Moreover, levels of Foxp3 expression have been interpreted primarily to reflect Treg frequency rather than defining the regulatory functions of these cells (11, 12, 13, 14, 15), although it has been shown that Foxp3 is critically important for both the development of Tregs as well as their suppressor function (16, 17). Thus, understanding the significance of Foxp3 expression per se to analyze Treg function and the dynamics of Treg-mediated tolerance in transplantation may afford new opportunities to harness the potential of these cells to promote tolerance.

We herein investigated the kinetics and function of Tregs in allograft acceptors and rejectors using a well-characterized mouse model of corneal transplantation (18, 19, 20). In this model, ∼50% of allografts are rejected within 3 wk of transplantation while the remaining half enjoy indefinite survival, indicating involvement of active regulatory mechanisms that promote allograft longevity. Compared with other transplant systems (e.g., skin, heart, kidney, etc.) where a 100% graft rejection rate is observed in unmanipulated recipients, this model provides an exceptional opportunity to study why some of the allografts are spontaneously accepted and to evaluate the role of Tregs in this high frequency of transplant survival.

In the present study, we clearly demonstrate that contrary to previous reports the functional status of Tregs is more reflective of allograft outcome than their numbers. Moreover, the levels of Foxp3 expression in Tregs are directly associated with their potential to suppress T cell activation and prevent allograft rejections.

The 8- to 10-wk-old BALB/c (H-2^d), C57BL/6 (H-2^b), and C3H (H-2^k) male mice were purchased from Taconic Farms (and used as recipients, donors, and third-party controls, respectively. The study was approved by the Institutional Animal Care and Use Committee, and the mice were treated according to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

Standard protocol for murine orthotopic corneal transplantation was used, as described previously (18, 20). Briefly, donor center corneas (2-mm diameter) were excised from C57BL/6 mice and sutured on to the recipient graft beds prepared by excising a 1.5-mm site in the central cornea of BALB/c mice. Simultaneously, some BALB/c mice received syngeneic (BALB/c) grafts to control for the non-allospecific effects of surgery. The corneal sutures were removed 7 days after surgery. All grafts were evaluated using slit-lamp biomicroscopy at weekly intervals. Grafts were defined as rejected when they became opaque and the iris details could not be recognized clearly using a standardized opacity grading (ranges, 0–5) scheme. This model renders an easy and confirmatory method to directly evaluate graft endpoints in vivo at different time points without sacrificing the animals, which makes it an outstanding system to study immune responses in the transplant setting.

Ipsilateral draining submandibular and cervical lymph nodes (LNs) as well as grafted corneas were harvested and single-cell suspensions were prepared. Single-cell suspensions of corneal samples were prepared by collagenase digestion, as previously described (19). The isolated cells were stained with the following Abs: Anti-CD4/CD3 FITC, anti-CD25 PE, and anti-FoxP3 PECy5 (eBioscience) and analyzed on an EPICS XL flow cytometer (Beckman Coulter).

CD4⁺ T cells and CD4⁺CD25⁺ Tregs from the LN of naive and/or different graft recipients were isolated by magnetic separation using CD4⁺ T cell and Treg isolation kits (Miltenyi Biotec). Purity of sorted cells was >97%, and 95% of CD4⁺CD25⁺ Tregs were Foxp3⁺ in all the groups as confirmed by flow cytometry.

RNA was isolated with RNeasy Micro Kit (Qiagen) and reverse transcribed using Superscript III Kit (Invitrogen). Real-time PCR was performed using TaqMan Universal PCR Mastermix and preformulated primers for FoxP3 (assay ID Mm00475156_ml) and GAPDH (assay ID Mm99999915_gl) (Applied Biosystems). The results were analyzed by the comparative threshold cycle method and normalized by GAPDH as an internal control.

Tregs (n = 5 × 10⁵) isolated from the LN of allograft acceptors and rejectors were dissolved in lysis buffer. The protein concentrations in cell lysates were determined using BCA reagent (Thermo Scientific). Equal amounts of protein were boiled in an equal volume of Laemmli sample buffer and resolved by 10% SDS-PAGE. Proteins were then transferred to polyvinylidene difluoride membrane (Immobilon; Millipore) and probed overnight with primary anti-Foxp3 Ab (Santa Cruz Biotechnology) or anti-β-actin Ab (Abcam). Immunoreactive bands were detected with HRP-conjugated secondary Abs and visualized by ECL. The intensities of immunoreactive bands were analyzed using National Institutes of Health-ImageJ software.

For all the suppression assays, including cytokine and Transwell assays, CD4⁺ T effectors (Teff, total CD4⁺ cells without depleting the CD25⁺ population) and Tregs were isolated from the draining LNs. Teff cells (1 × 10⁵) were cocultured with Tregs (5 × 10⁴), T cell-depleted syngeneic splenocytes (1 × 10⁵), and 1 μg/ml anti-CD3 Ab for 3 days. Proliferation of CD3-stimulated Teff cell without adding Tregs was considered as control proliferation with 0% suppression. Proliferation was measured using the BrdU incorporation assay (Millipore), and percent suppression was calculated using the following formula: % suppression = [(Teff proliferation without Tregs − Teff proliferation with Tregs)/(Teff proliferation without Tregs)] × 100. To measure the allo-specificity of Tregs, Teff and Tregs were cocultured with T cell-depleted allogeneic splenocytes of C57BL/6 or C3H mice without anti-CD3 stimulation. Transwell experiments were performed in 24-well plates using Transwell cell culture inserts (1.0 μm; BD Biosciences). Naive Teff cells were cultured with T cell-depleted syngeneic splenocytes and anti-CD3 Ab, and Tregs were either added directly to the culture or were placed in Transwell inserts.

Cytokine levels in the supernatants of suppression assays were analyzed using commercially available ELISA kits for murine TGF-β1 and IL-10 (eBioscience).

First, Tregs were isolated from the draining LN of syngeneically grafted recipients, and allograft acceptors and rejectors at wk 3 posttransplantation. Subsequently, these Tregs (1 × 10⁵ cells per mouse) were transferred i.v. to three different groups of allograft recipients at 18 h postsurgery. This time point for Treg transfer was selected on the basis of our previous observation of detecting considerable numbers of graft-derived APCs in the draining LNs of recipients at 24 h posttransplantation (18). Allograft survival rate in each group (n = 6 per group) was monitored up to 8 wk posttransplantation.

Student’s t test was used for comparison of mean between the groups. Kaplan-Meier analysis was adopted to construct survival curves, and the log-rank test was used to compare the rates of corneal graft survival. Data are presented as mean ± SEM and considered significant at p < 0.05.

Regional draining LN and grafted corneas were harvested from graft recipients at weekly intervals up to wk 4 posttransplantation. At wk 3 posttransplantation, allograft recipients were separated into acceptors and rejectors. The flow cytometric analysis of draining LN cell suspensions showed no difference in the frequencies of CD4⁺CD25⁺Foxp3⁺ Tregs, either as a proportion of the total CD4⁺ T cells (range: 8.5–10.5%) or of the total LN cells (range: 4–5%) among syngeneic recipients, acceptors, and rejectors at any of the time points (Fig. 1, a and b). Interestingly, in contrast to the Treg frequencies, we observed considerable differences in the levels (mean fluorescence intensity: MFI) of Foxp3 by the Tregs of different groups (Fig. 1,c). The MFI of Foxp3 in Tregs of allograft acceptors was ∼50% higher compared with those in the Tregs of allograft rejectors and syngeneic recipients. Additionally, to confirm these differences in the Foxp3 expression levels, we investigated Foxp3 protein levels by Western blot in sorted Tregs and mRNA levels by real-time PCR in the draining LN as well as in equal number of sorted Tregs of different graft recipient groups. The densities of immunoreactive Foxp3 bands (normalized to densities of β-actin) show that acceptor-Tregs express 2-fold higher Foxp3 protein compared with those by rejector-Tregs (Fig. 1,c). Similarly, there was a ∼2-fold increase in Foxp3 mRNA expression in the LN of acceptors as compared with those of rejectors at wk 3 posttransplantation and thereafter. In addition, the levels of Foxp3 mRNA expression in the LN of rejectors were lower than those of syngeneic recipients, which nearly remained unchanged from baseline levels (Fig. 1,d). Similar to the Foxp3 mRNA expression in total LN, the isolated Tregs from the LN of acceptors expressed 2-fold higher levels of Foxp3 mRNA compared with those from the LN of rejectors (Fig. 1 e) confirming that Tregs in allograft acceptor express high levels of Foxp3.

In addition, we wanted to determine whether Tregs migrate to the graft site and, similar to LN-Tregs, express different levels of Foxp3 in accepted vs rejected allografts. The relative frequencies of Foxp3⁺ Tregs (11–13%) in relation to the total infiltrating T cells (Fig. 1,f) as well as expression levels of Foxp3 (MFI: 33–36) in Tregs (Fig. 1,g) remained the same among the three groups. However, the expression of Foxp3 mRNA was ∼2-fold higher in corneas with rejected grafts than those with accepted grafts (Fig. 1,h). This higher expression of Foxp3 in rejected grafts was principally due to the nearly ∼2-fold increase in the frequencies of infiltrating T cells (and hence also Tregs) in the corneas with rejected as compared with accepted allografts (Fig. 1 f).

Using a coculture suppression assay system, we investigated the regulatory potential of Tregs to suppress the proliferation of CD3 Ab-stimulated naive T cells (isolated from the LN of naive BALB/c mice) and allo-primed T cells (isolated from the LN of allograft rejectors). Tregs isolated from the LN of acceptors (acceptor-Tregs) were significantly more effective in suppressing the proliferation of naive T cells compared with those isolated from the LN of rejectors (rejector-Tregs) and syngeneic recipients (syngeneic-Tregs). Rejector-Tregs were even less effective in suppressing the T cell proliferation than syngeneic-Tregs (Fig. 2,a). However, in contrast to suppression of naive T cell activation, Tregs of all the groups failed to efficiently suppress the proliferation of allo-primed effector T cells (Fig. 2 b).

To delineate the mechanisms by which Tregs suppress the activation of T cells, we measured the cytokine levels in the supernatant of suppression assays and investigated the mechanisms of suppression using ELISA and Transwell assays, respectively. The levels of regulatory cytokines, TGFβ1 and IL-10, were significantly higher (∼2- to 3-fold) in the supernatant of suppression assays with acceptor-Tregs compared with those with rejector-Tregs and syngeneic-Tregs (Fig. 3,a). In Transwell assays, only acceptor-Tregs were capable of significantly suppressing the T cell proliferation, suggesting that these Tregs can use both cell-cell contact-dependent and -independent (cytokine-mediated) mechanisms of suppression (Fig. 3 b). However, rejector-Tregs and syngeneic-Tregs were unable to suppress T cell proliferation in Transwells, corroborating their diminished capacity to produce TGFβ1 and IL-10.

Increased production of TGFβ1 and IL-10 by acceptor-Tregs point toward a subset of induced Tregs which may be alloantigen-specific. To characterize this specificity further, we analyzed the suppression potential of Tregs of the different groups against proliferation of recipient naive T cells (BALB/c) to donor (C57BL6)-specific APCs and third party (C3H) APC, allodisparate to both recipients and donors (Fig. 4). Similar to suppression of CD3 Ab-stimulated T cells, acceptor-Tregs were more effective in suppressing donor-specific Ag (C57BL6-APC)-stimulated T cells compared with syngeneic-Tregs. In contrast, no difference was observed in the potential of acceptor-Tregs and syngeneic-Tregs in suppressing the activation of nonspecific third party Ag (C3H-APC)-stimulated T cells; indeed, this suppression was equal to the suppression of donor-specific Ag-stimulated T cells by syngeneic-Tregs. However, rejector-Tregs showed little to no suppression of any of the Ag-stimulated T cells. These results clearly suggest induction of alloantigen-specific Tregs only in acceptors.

Our in vitro Treg functional data on T cell suppression and cytokine expression clearly demonstrated that Foxp3^high-expressing Tregs from the LN of allograft acceptors are highly functionally effective compared with those of rejectors. Finally, we wanted to confirm this functional efficacy of Foxp3^high-expressing Tregs in vivo by demonstrating there ability to prevent graft rejection. Thus, we performed a Treg adoptive transfer experiment in which Tregs isolated from the LN of syngeneically grafted recipients, allograft acceptors, and rejectors at wk 3 posttransplantation were transferred i.v. to three different groups of allograft recipients (Fig. 5). Only the group that was adoptively transferred with acceptor-Tregs showed significant increase in the allograft survival rate (83%, p = 0.035, Kaplan-Meier analysis); no improvement in graft survival was observed in the other groups of recipients which received syngeneic-Tregs or rejector-Tregs.

In this study, we show that the frequencies of Tregs remain unchanged in allograft-acceptors vs allograft-rejectors in both the LN and grafts. However, Tregs isolated from the LN of allograft acceptors express significantly higher levels of Foxp3 compared with those of allograft rejectors. In addition, these Foxp3^high-expressing Tregs are functionally highly effective in preventing graft rejection as demonstrated by adoptive transfer experiments.

Despite strong evidence about the association of Foxp3 primarily with Treg development and function, care must be taken with any simple interpretation of Foxp3 analysis. Recent human studies have shown that most T cells transiently turn on Foxp3 during early activation stages (21, 22). Thus, it is imperative to characterize Tregs at both the phenotypic and functional levels, as most transplant studies are now using Foxp3 as a marker to define Tregs. Moreover, these reports are associating allograft outcome with Foxp3 PCR analysis or Treg frequencies (9, 10, 11, 12, 13, 14, 15). Our data clearly suggest that Foxp3 PCR analysis does not simply reflect Treg frequencies. Increased Foxp3 expression in the LN of allograft acceptors was mainly due to the high level of Foxp3 expression in Tregs themselves and not simply due to their increased numbers. In contrast to the LN, we did not observe differences either in Foxp3 levels in graft-infiltrating Tregs or their frequencies in relation to total infiltrating T cells in accepted vs rejected grafts. However, the absolute numbers of Tregs in rejected grafts were higher than those in accepted grafts, which is the main reason for the increased Foxp3 mRNA expression in rejected grafts. In addition to Tregs homing into grafts, these findings suggest that evaluation of only a single T cell subset (Tregs or Teffs) either at the graft site or in the lymphoid compartment may potentially lead to incomplete conclusions. Treg migration to the graft site appears to be either a response to down-regulate the ongoing inflammation or nonspecific homing due to a greater propensity of T cells to be recruited to sites of inflammation in the periphery. These findings are in accord with several previous studies demonstrating increased Foxp3 expression in the biopsies of rejected kidney grafts and suggest the presence of intragraft Tregs as a consequence of the normal immune response during rejection (10, 23).

The essential role of Foxp3 in regulating the suppressor function of Tregs has been well documented (16, 17). Our observations of increased Foxp3 levels in the LN-Tregs of allograft acceptors prompted us to hypothesize that higher Foxp3 levels in Tregs are associated with their enhanced regulatory functions. Thus to validate our hypothesis, we analyzed the function of Tregs isolated from the LN of different transplant groups. Our data demonstrate that Foxp3^high Tregs isolated from the LN of allograft acceptors were more effective in suppressing naive-T cell proliferation and in producing higher levels of TGF-β1 and IL-10 compared with those isolated from allograft rejectors and syngeneic recipients. However, Tregs of all the transplant groups failed to efficiently suppress the proliferation of allo-primed effector T cells. These findings suggest that Tregs may be more effective in regulating alloimmune response at the induction phase than at the effector phase. Our previous findings of more numbers of Tregs in rejected corneas with their inability to prevent rejection also lend support to this notion. The inefficiency of Tregs to suppress the allo-primed effector T cells may be due to the intrinsic resistance of Teff cells to suppression, potentially due to Teff cells’ increased capacity to produce cytokines (e.g., IL-6, TNF-α, and IFN-γ), which can inhibit and antagonize Treg function (24). Although, peripherally induced CD4⁺CD25⁺ Tregs are phenotypically indistinguishable from thymus-derived CD4⁺CD25⁺ natural Tregs, higher levels of TGF-β1 and IL-10 secretion by the acceptor-Tregs point toward the development of induced-Tregs in allograft acceptors. In addition, our data demonstrate the induction of alloantigen-specific Tregs only in the LN of allograft acceptors, which corroborates with previous reports suggesting that peripheral development of alloantigen-specific Tregs occurs in the LNs (25) and constitutively high expression of Foxp3 is required for Treg differentiation and function (16, 17). Finally, our findings of Treg adoptive transfer experiment clearly show that only Tregs isolated from the LN of allograft acceptors are capable of preventing allograft rejection.

In summary, these findings provide new insights into the role of Tregs in transplantation immunity. For the first time, the present study reports that the functional status of Tregs is more reflective of the allograft outcome than their absolute frequencies, and that the levels of Foxp3 expression are directly associated with the Treg potential in preventing graft rejection. In addition, this is the first description of Treg localization and function in corneal transplantation, the most common form of tissue grafting. Our data also demonstrate the induction of allospecific Tregs in graft recipients, which clearly are able to suppress the induction of alloimmunity in regional draining LNs but are not equally effective in suppressing the effector phase of the immune response. These observations have at least two important implications in regulating the immune responses to tissue transplants beyond the obvious applications to corneal alloimmunity. First, predicting and/or correlating allograft survival, or studying the effect of therapeutic agents on Treg biology in transplantation, exclusively on the basis of Treg frequencies (either in lymphoid tissues or graft site) may potentially lead to inaccurate conclusions. Second, potentiating Treg function to suppress alloimmunity, particularly at the induction phase, may be more helpful in promoting transplant longevity than at the effector phase.

The authors have no financial conflict of interest.