Targeting the ion channel TRPM7 promotes the thymic development of regulatory T cells by promoting IL-2 signaling

Abstract

The thymic development of regulatory T (T_reg) cells, crucial suppressors of the responses of effector T (T_eff) cells, is governed by the transcription factor FOXP3. Despite the clinical importance of T_reg cells, there is a dearth of druggable molecular targets capable of increasing their numbers in vivo. We found that inhibiting the function of the TRPM7 chanzyme (ion channel and enzyme) potentiated the thymic development of T_reg cells in mice and led to a substantially higher frequency of functional T_reg cells in the periphery. In addition, TRPM7-deficient mice were resistant to T cell–driven hepatitis. Deletion of Trpm7 and inhibition of TRPM7 channel activity by the FDA-approved drug FTY720 increased the sensitivity of T cells to the cytokine interleukin-2 (IL-2) through a positive feed-forward loop involving increased expression of the IL-2 receptor α-subunit and activation of the transcriptional regulator STAT5. Enhanced IL-2 signaling increased the expression of Foxp3 in thymocytes and promoted thymic T_reg (tT_reg) cell development. Thus, these data indicate that inhibiting TRPM7 activity increases T_reg cell numbers, suggesting that it may be a therapeutic target to promote immune tolerance.

Introduction

The immune system has the ability to control the intensity, duration, and scope of inflammatory processes through an elaborate array of checks and balances (1). Regulatory T (T_reg) cells play a salient immunosuppressive role to balance the destructive potential of effector T (T_eff) cells (2). Although T_reg cells develop from the same thymocyte precursors as do the T_eff cells, the distinct T_reg cell lineage is specified through the expression and maintenance of the forkhead box protein P3 (FOXP3) transcription factor (3, 4). Consequently, the genetic loss of FOXP3 in humans and mice prevents the development of T_reg cells and breaks self-tolerance (4-8). Interleukin-2 (IL-2) signaling plays a decisive role in immune tolerance by regulating the development, maintenance, and function of T_reg cells (9-13). The key transcription factor responsible for Foxp3 expression, downstream of IL-2 signaling, is STAT5 (14, 15). Accordingly, ectopic expression of a constitutively active STAT5 variant is sufficient to divert the fate of developing thymocytes toward the T_reg cell lineage (12). The thymic development of T_reg cells is thought to occur through a two-step process (13, 16). First, T cell receptor (TCR) signaling increases the abundance of the IL-2 receptor α-subunit (IL-2Rα, also known as CD25) and other components of the IL-2 signaling pathway, increasing the sensitivity of the cells to IL-2. Second, IL-2-mediated signals increase Foxp3 transcription in a STAT5-dependent manner to finalize the commitment of these CD25⁺ progenitors to the T_reg cell lineage. However, a study indicated an additional developmental program involving Foxp3^lo T_reg progenitor cells (13). Identifying previously uncharacterized “druggable” components in this pathway may enable the pharmacological manipulation of Foxp3 expression and T_reg cell numbers in vivo.

The National Center for Advancing Translational Sciences (NCATS) has identified a druggable genome of ~3000 human genes encoding predominantly three key protein families that are ideal for drug development: nonolfactory GPCRs, ion channels, and protein kinases (49-51). Included in the ion channel family are the transient receptor potential (TRP) channels, a 28-member superfamily of ion channels that constitute an exciting class of drug targets (17). Cation-selective TRP channels mediate context-specific electrical signaling in all organ systems, but the functions of the specific TRP channels prevalent in the immune system remain largely mysterious and thus unexploited for immunomodulation. Three members of the TRP superfamily, TRPM2, TRPM6, and TRPM7 possess ion channel activity and an additional enzyme activity. TRPM7, an ion channel permeable to cations such as Ca²⁺, Na⁺, Zn²⁺, and Mg²⁺, contains a carboxy-terminal serine-threonine kinase domain (18-20) and is highly expressed in T cells (21). Hematopoietic cells exhibit robust TRPM7 currents (I_TRPM7 or simply I_M7), and TRPM7 is an important regulator of both innate and adaptive immunity (21-24). Previously, we generated Trpm7^fl/fl(Lck Cre) mice to delete Trpm7 selectively in the T cell lineage (21). In these mice, thymocyte development is impaired, resulting in a substantial accumulation of CD4⁻CD8⁻ [double-negative 3 (DN3); CD44⁻CD25⁺) thymocytes and reduced thymic cellularity (21). The developmental block is partial, and the residual egress of mature CD4⁺ and CD8⁺ T cells populates the peripheral lymphoid organs normally. Because Trpm7^−/− T cells are also resistant to apoptosis (24), we expected to see a lymphoproliferative phenotype and autoimmunity (25). Surprisingly, the Trpm7^fl/fl (Lck Cre) mice [hereafter referred to as “KO” mice] exhibit T cell lymphopenia (21) and autoimmunity is greatly delayed and mild in its manifestation (24). These observations suggest that deletion of TRPM7 in T cells promotes immunosuppression through an undefined mechanism.

Here, we studied a mouse model of autoimmune hepatitis (AIH), a T cell–driven inflammatory disease that often leads to liver cirrhosis, cancer, and death (26). In mice, AIH can be modeled by intravenous (i.v.) injection of Concanavalin A (Con A), which results in an acute, dose-dependent liver inflammation that is driven by the activity of CD4⁺ T cells (27). We showed that when Trpm7 was deleted in T cells, the mice were resistant to Con A–induced hepatitis. Trpm7^−/− T cells were activated normally and showed only modest differences in their cytokine outputs, but we found that deletion of Trpm7 in the developing thymocytes potentiated the thymic development of T_reg cells, leading to a substantially higher frequency of functional T_reg cells in the periphery. The deletion of Trpm7 or pharmacological inhibition of TRPM7 increased FOXP3 protein abundance in developing thymocytes, steering a larger percentage toward the T_reg cell lineage. Bone marrow transplant experiments indicated that the underlying mechanism was non-cell autonomous and in part depended on increased IL-2 production by the Trpm7^−/− thymocytes. Thus, we have found that blocking TRPM7 channel activity increased IL-2–dependent thymic Treg (tT_reg) development. Through these studies, TRPM7 has emerged as a pharmacological target to potentially increase T_reg cell numbers in vivo and promote immunological tolerance.

Results

Deletion of Trpm7 in T cells protects mice from Con A–induced experimental lethal hepatitis

Autoimmune hepatitis (26), which is modeled by injecting mice i.v. with Con A, results in acute T cell–dependent liver inflammation and lethality (27). We administered Con A and assessed the development of hepatitis over the course of 24 hours (Fig. 1A). Kaplan-Meier survival analysis of Trpm7^fl/fl (WT) and Trpm7^fl/fl(Lck Cre) (KO) mice showed that 50% of the WT mice succumbed within 24 hours (Fig. 1B), whereas The KO mice were resistant to Con A–induced death. To assess liver damage, we measured the concentration of the liver enzyme alanine transaminase (ALT) in the serum 24 hours after Con A injection. The mean serum ALT activity in the WT mice was increased modestly in the surviving mice (Fig. 1C), which is likely an underestimate because it was not possible to measure the ALT concentration in the dead mice. Through histological analysis of liver sections, perivascular inflammation was readily evident in Con A–treated WT mice but not in Con A–treated KO mice (Fig. 1D). We also measured the relative changes in the expression of genes encoding inflammatory cytokines in homogenized liver tissue of the surviving mice 24 hours after Con A injections. In the WT mice, the relative abundance of mRNAs for inflammatory cytokines increased substantially in the Con A–injected mice when compared to saline-injected controls (Fig. 1E). The expression of Il2, Il1b, and Il6 was also increased, but to a lesser extent. In contrast to WT mice, the increase in inflammatory cytokine gene expression was significantly blunted even in the surviving Con A–injected KO mice. The trend was reversed in the case of Tgfb1, which encodes the anti-inflammatory cytokine TGF-β, relative to the saline-treated controls. Together, these results suggest that the TRPM7 in T cells is important for the onset of acute liver inflammation induced by Con A. Because the deletion of Trpm7 in T cells was considerably protective in Con A–induced hepatitis, we tested the hypothesis that TRPM7 was essential for T cell activation. Indeed, the idea that TRPM7 regulates T cell activation has been suggested previously from experiments with T cell lines, albeit without the benefit of gene targeted mice (28).

Fig. 1. Deletion of Trpm7 in T cells protects mice from Con A–induced experimental lethal hepatitis.

(A) Cartoon of the experimental scheme and sample processing. (B) Kaplan-Meier survival analysis of Trpm7^fl/fl (WT) and Trpm7^fl/fl (Lck Cre) (KO) mice injected i.v. with Con A (40 mg/kg) and assessed over 24 hours. Log-rank value = 6.36, P = 0.011; n = 10 mice in each group. (C) Serum alanine transaminase (ALT) concentrations in WT and KO mice 24 hours after the administration of Con A or saline. For the 3 WT mice that died before the 12-hour time-point, ALT concentrations were not recorded (n = 4; data analyzed by two-tailed t test). (D) Hematoxylin and eosin (H&E) staining of liver sections from WT and KO mice injected i.v. with saline or Con A (40 mg/kg). Scale bar, 100 μm. Green arrows indicate regions of perivascular infiltration by immune cells. Images are representative of liver sections from three different WT and KO mice. (E) Gene expression analysis (qPCR) of the indicated cytokine-encoding genes in WT (blue) and KO (red) mouse liver tissue 24 hours after administration of saline or Con A (n = 3; P values were calculated by two-tailed t test).

KO T_eff cells exhibit normal activation and proliferation but have defects in cytokine production

T cell activation is initiated after the stimulation of the TCR complex in conjunction with a costimulatory signal (29). To evaluate T cell activation, we stimulated freshly isolated splenic CD4⁺CD25⁻ T cells, referred to as effector T cells (T_eff cells) with either Con A or a mixture of anti-CD3 and anti-CD28 antibodies. The cells were analyzed for the cell-surface appearance of activation markers by flow cytometry and for cytokine secretion by ELISA (Fig. 2A). After 48 hours of stimulation, both WT and Trpm7^−/− (KO) T_eff cells showed similar increases in CD69 abundance (Fig. 2B). The increase in the cell surface abundance of CD25 were considerably higher in cells from the KO mice than in cells from the WT mice (Fig. 2C). T cell activation results in increased cell size, which is conveniently measured by the forward scatter (FSC) parameter in flow cytometry. We found that both WT and KO T_eff cells increased in size comparably (Fig. 2D). After 72 hours of activation, we measured T cell proliferation using the CFSE-labeling method, which measures the progressive dilution of CFSE during cell divisions. Both WT and KO T_eff cells proliferated comparably in response to stimulation by the anti-CD3 and anti-CD28 antibody cocktail (Fig. 2E). Next, based on another study, we considered the possibility that TRPM7 regulates Orai-mediated Ca²⁺ influx after TCR stimulation (30). We performed live-cell imaging of thymocytes stained with Fura-2-AM upon TCR activation with an anti-CD3e antibody (Fig. 2F). We observed that the deletion of TRPM7 changed the kinetics, but not the amplitude, of store-operated Ca²⁺ entry (SOCE), as determined by peak [Ca²⁺]_i and the area under the curve.

Fig. 2. Activation of KO T_eff cells is normal upon TCR stimulation.

(A) Schematic of the experimental procedures. Splenocytes were magnetically isolated and then activated with a mixture of anti-CD3 and anti-CD28 antibodies (αCD3/CD28) or Con with A. (B) Flow cytometry histograms showing the cell surface abundance of the activation marker CD69 on CD4⁺CD25⁻ primary T cells from WT and KO mice after in vitro activation for 48 hours by Con A (5 μg/ml) (left) or by a mixture of anti-CD3 and anti-CD28 antibodies (right). Blue- and red-filled histograms show CD69 abundance on naïve WT and KO cells, respectively. Unfilled histograms show CD69 after activation. Histograms are representative of three independent experiments. (C) Flow cytometry analysis of the abundance of the activation marker CD25 on WT and KO CD4⁺CD25⁻ T cells after stimulation as described in (B). Histograms are representative of three independent experiments. (D) Representative histograms showing forward scatter (FSC) as a measure of activation for WT (blue) and KO (red) CD4⁺CD25⁻ T cells (n=3). (E) Analysis of the division of CFSE-labeled WT and KO CD4⁺CD25⁻ T cells as measuring the dilution of CFSE 72 hours after activation by anti-CD3 and anti-CD28 antibodies. The percentage of dividing cells (left gate) and undivided cells (right gate) are shown. Data are representative of three experiments. (F) Ca²⁺ responses in WT (blue) and KO (red) thymocytes after TCR stimulation. The intracellular Ca²⁺ increase was observed in response to the binding of streptavidin (SA) to biotinylated anti-CD3e, which was previously added to the cells. Ionomycin (IM) was used as a positive control and denotes maximal Ca²⁺ responses. The number of cells assessed is shown. The inset bar graph illustrates the rise-time (tau) of fluorescence upon streptavidin treatment in WT and KO thymocytes. (G) Quantification of secreted cytokines as detected in the cell culture medium of CD4⁺CD25⁻ T cells activated for 72 hours with Con A (5 μg/ml). The P values were calculated by two-tailed t-test and are denoted on the box charts (n=3-4). (H) Quantification of secreted cytokines as detected in the cell culture medium of CD4⁺CD25⁻ T cells activated for 72 hours with a mixture of anti-CD3 and anti- CD28 antibodies (each at 5 μg/ml). The P values were calculated by two-tailed t-test and are denoted on each box chart (n = three or four samples).

We then wondered whether T cell migration was affected by the absence of TRPM7. We compared the transwell migration of WT and KO T cells in response to chemotactic factors produced by activated macrophages. The KO T cells migrated normally through the membrane (fig. S1A). Because a major function of activated CD4⁺ T cells is the secretion of cytokines that orchestrate inflammation, we measured the concentrations of various inflammatory cytokines in the culture medium of the T cells 72 hours after activation. After activation with Con A, the KO T cells secreted significantly greater amounts of IL-4 but reduced amounts of IL-3 and IL-6 (Fig. 2G). When T cells were activated with a mixture of anti-CD3 and anti-CD28 antibodies, the KO T cells consistently secreted increased amounts of IL-2, IL-4, IL-5, and IL-6 when comparted to the WT T cells (Fig. 2H). Note that the KO T cells also secreted increased amounts of the anti-inflammatory cytokine IL-10, which was due to the increased expression of Il10 (fig. S1B). These results suggest that KO T_eff cells became activated and proliferated normally but had a different cytokine secretion profile to that of WT Teff cells. This likely contributed to the resistance to Con A–induced lethal hepatitis but may not constitute a full explanation. In search for additional explanations, we characterized the composition of hematopoietic infiltration in the livers of Con A–injected mice.

The infiltration of T_reg cells is increased in the livers of Con A–treated KO mice

The liver constitutes a unique immunological environment, housing a large number and diversity of hematopoietic cells (31). To quantify the composition of these cells by flow cytometry, we analyzed single-cell suspensions isolated from freshly excised livers and also evaluated liver sections by immunohistochemistry and scanning electron microscopy (SEM) (Fig. 3A). Based on the cell surface staining of CD45, a marker of hematopoietic cells, the livers of saline-injected WT and KO mice contained an equivalent proportion of hematopoietic cells (~40%), which nearly doubled (~80%) 24 hours after Con A injections in both WT and KO mice (Fig. 3, B and C). Therefore, there was no significant difference between the WT and KO mice in terms of overall infiltration of hematopoietic cells after injection with Con A. Because CD4⁺ T cells are the key mediators of Con A–induced lethality (27), we assessed the proportion of liver-infiltrating CD4⁺ T cells. The liver-resident CD4⁺ T cell population was readily detectable in saline-injected mice (Fig. 3, D and E), and the proportion of CD4⁺ T cells changed slightly upon Con A injection. The percentage of FOXP3⁺ Treg cells was also similar in the livers of saline-injected WT and KO mice (Fig.3, F and G; see fig. S2 for gating strategy). The T_reg cell proportion became significantly increased in Con A-treated KO mice relative to that in their WT counterparts (Fig. 3, F and G). Because FOXP3⁺ T_reg cells are highly immunosuppressive, these data suggest that the increased frequency of T_reg cells in the KO mice contributed to the protection of these mice from Con A-induced liver inflammation. Immunofluorescence microscopy of liver sections suggested that there were increased numbers of CD4+ T cells in the Con A-injected WT mice but that substantially fewer cells were detectable in the livers of the Con A-injected KO mice (Fig. 3H). Analysis of the liver vasculature by SEM (Fig. 3I) revealed the presence of endothelium-adhesive immune cells in the livers of Con A-injected WT mice, but these were less evident in the Con A-injected KO mice. Together, these results led us to hypothesize that the deletion of Trpm7 in T cells results in a significantly increased frequency of T_reg cells in the periphery. The increase in the T_reg:T_eff ratio may be but one important cause of overall immunosuppression, manifesting in this study as an insensitivity to Con A-induced liver inflammation. To gain further insight into the origin of the imbalance in the T_reg:T_eff ratio, we surveyed the T_reg cell frequency in the thymus and spleen.

Fig. 3. Increased infiltration of T_reg cells in the livers of Con A–treated KO mice.

(A) Experimental scheme. (B) Bivariant cytographs showing infiltration of CD45⁺ hematopoietic cells into the livers of WT (blue) and KO (red) mice 24 hours after the administration of saline or Con A (40 mg/kg, i.v.). CD45 FMO refers to the ”fluorescence minus one” control, wherein cells were analyzed without staining with anti-CD45 antibody to enable accurate gating. Plots are representative of four to six experiments. (C) Quantification of the number of CD45+ hematopoietic cells based on the flow cytometry analysis shown in (B). Mean values are indicated by a dash. The data were collected from four independent experiments. (D) Bivariant cytographs showing the infiltration of CD45⁺CD4⁺ T cells into the livers of WT and KO mice 24 hours after administration of saline or Con A (40 mg/kg, i.v.). Plots are representative of four to six experiments. (E) Quantification of the number of CD45⁺CD4⁺ T cells based on the flow cytometry analysis shown in (D). Mean percentage values are indicated by a dash. The data were collected from four independent experiments. (F) Flow cytometry histograms of CD4⁺Foxp3⁺ cells from the livers of WT (blue) and KO (red) mice 24 hours after administration of saline or Con A. The Foxp3 histograms were obtained from a cell population determined by sequential gating for a FSC/SSC profile, CD45⁺ and CD4⁺ (see fig. S2 for the gating scheme). (G) Quantification of Foxp3+ T_reg cells in the livers of WT (blue) and KO (red) mice from the experiments shown in (F). The P value was calculated by t test. The data were collected from four independent experiments. (H) Immunofluorescence microscopy of infiltrating CD4⁺ T cells in liver sections. (I) Representative SEM images of liver sections from the indicated mice at 850X magnification. Scale bar, 100 μm.

KO mice have a higher frequency of CD4+Foxp3+ T_reg cells in the thymus and spleen

We previously showed that the deletion of Trpm7 in T cells results in impaired T cell development and a partial block in the transition from double negative (CD4⁻CD8⁻, DN) to double positive (CD4⁺CD8⁺, DP) cells (21). This results in an accumulation of DN thymocytes and reduces the development of single positive (SP) CD4⁺ and CD8⁺ mature T cells (21). The SP cells egress to the periphery and populate all of the lymphoid organs but the mice exhibit lymphopenia despite these T cells being resistant to apoptosis (24). We measured the frequency of T_reg cells in the thymus (Fig. 4, A to E) using flow cytometry. To assess thymus-derived T_reg cells, we analyzed Thy1.2⁺CD8⁻CD4⁺ thymocytes for FOXP3 expressing cells. Thymi from KO mice showed a consistent ~3-fold increase in the frequency of FOXP3⁺ T_reg cells when compared to thymi from WT mice (Fig. 4C). Because the overall cellularity of the thymus is significantly lower in the KO mice than in the WT mice (21), the ratio of the absolute numbers of T_reg cells was also compared (Fig. 4D). The CD4⁺ thymocytes from KO mice showed modestly increased median fluorescence intensity (MFI) values for FOXP3 staining (Fig. 4E), indicating normal expression of FOXP3 in T cells. These results suggest that the KO thymi generated a T cell population that overrepresents T_reg cells.

Fig. 4. KO mice display increased numbers of CD4+Foxp3+ T_reg cells in the thymus and spleen.

(A) Experimental scheme. (B) Representative flow cytometry contour plots sequentially gated on CD4⁺ (top) and Foxp3⁺ T_reg cells (bottom) in thymi isolated from WT (blue) and KO (red) mice. (C) Frequency of CD4⁺Foxp3⁺ T_reg cells in thymi from WT (blue) and KO (red) mice. The mean value is denoted by an empty square and the median by a horizontal line. The P value was calculated by two-tailed, unpaired Student’s t-test. (D) Absolute numbers of CD4⁺Foxp3⁺ thymocytes in thymi from WT (blue) and KO (red) mice. The empty square denotes the mean and the horizontal line the median. (E) MFI values of FOXP3 in CD4⁺ WT (blue) and KO (red) thymocytes. P value was calculated by two-tailed paired t-test. (F) Contour plots showing the frequencies of CD4⁺ and Foxp3⁺ T_reg cells in the spleens of WT (blue) and KO (red) mice. (G) Box charts show the frequency of CD4⁺Foxp3⁺ T_reg cells in the spleens of WT (blue) and KO (red) mice. (H) Absolute numbers of CD4⁺Foxp3⁺ T_reg cells in splenocytes isolated from WT (blue) and KO (red) mice. (I) Quantification of the MFI for FOXP3 in CD4⁺ splenic T cells from the indicated mice. P value was calculated by two-tailed paired t-test. (J) I-V relationship of I_TRPM7 in CD4⁺CD25⁺ WT T_reg (blue trace) and KO T_reg (red trace) cells obtained by whole cell patch clamp recordings. The patch clamp configuration and the recording conditions are shown in fig. S3F. The box chart shows the I_TRPM7 current densities quantified at 5 min after break-in at 100 mV in WT and KO T_reg cells. See Materials and Methods for a description of the statistical parameters shown in the box chart. (K) The box chart shows the I_TRPM7 current densities quantified at 5 min after break-in at 100 mV in T_eff (CD4⁺CD25⁻) and T_reg (CD4⁺CD25⁺) cells isolated from WT mice. The P value was calculated by two-tailed unpaired t-test. Each filled circle represents data obtained from an individual cell.

To check whether such a high frequency of T_reg cells was also seen in the periphery, we quantified CD4⁺FOXP3⁺ splenocytes (Fig. 4, F to I). The mean percentage of splenic T_reg cells was more than 2-fold higher in KO mice than in WT mice (Fig. 4G). Because the spleens of the KO mice contain a lower number of T cells (21), the mean absolute number of splenic T_reg cells in the KO mice was nearly identical to that quantified in the WT mice (Fig. 4H). Consistent with the CD4⁺ thymocytes, the CD4⁺ splenic T cells from the KO mice showed modestly increased MFI for FOXP3 staining (Fig. 4I). We also analyzed the data by stipulating that T_reg cells be identified as Thy1.2⁺CD4⁺CD25⁺Foxp3⁺ cells and T_eff cells as Thy1.2⁺CD4⁺ cells. The percentage of T_reg cells phenotyped in this manner was also significantly higher in the KO mice thymi (fig. S3B) and showed an increased trend in spleens (fig. S3C). With this analytical approach, the ratio of T_reg:T_eff cells in the KO thymus trended higher than that in the WT mice but was not considered significantly different (fig. S3D). However, the T_reg:T_eff ratio was still significantly higher in the spleens of the KO mice when compared to that in the spleens of the WT mice (fig. S3D). Thus, regardless of how the T_reg cells were immunophenotyped, deletion of Trpm7 appreciably increased the T_reg:T_eff ratio in the periphery. A very small percentage of CD8⁺ T cells is known to express FOXP3 (32). Therefore, we checked if the KO mice also showed an abnormally high percentage of CD8⁺FOXP3⁺ T cells. We found that the percentage of FOXP3⁺CD8⁺ cells was minor and statistically identical in the WT and KO mice (fig. S3A). To rule out a trivial explanation for these findings, we checked whether the increase in the frequency of T_reg cells in the KO mice was due to inefficient deletion of Trpm7 in T_reg cells, giving them an undefined proliferative advantage over T_eff cells. To assess the deletion of Trpm7 in KO T_reg cells, we measured the relative abundance in freshly isolated CD4⁺CD25⁺ T_reg cells of Trpm7 mRNA transcripts that contained loxP-flanked exon 17 and confirmed its efficient deletion in KO T_regs (fig. S3E). We also confirmed the loss of TRPM7 currents in KO Treg cells using patch clamp electrophysiology (see fig. S3F for the recording conditions and parameters). The WT CD4⁺CD25⁺ T_reg cells readily exhibited the characteristic outwardly rectifying I_TRPM7 with a reversal potential at 0 mV (Fig. 4J, left). I_TRPM7 was absent in KO T_regs and only a modest leak current was evident (Fig. 4J, right). These results demonstrate that Trpm7 was deleted efficiently in the T_reg cells. Similar recordings also showed that, when compared to WT CD4⁺CD25⁻ T_eff cells, the WT CD4⁺CD25⁺ T_reg cells elicited significantly higher TRPM7 current densities (Fig. 4K; see fig. S3G). The functional importance of increased TRPM7 currents in T_reg cells, in comparison to that in T_eff cells, is not yet clear. Next, we confirmed that the functional characteristics of KO T_reg cells were identical to those of WT T_reg cells.

KO T_reg cells display normal cell surface and functional characteristics

Immunosuppression by T_reg cells is mediated by multiple mechanisms, including inhibitory cytokines (IL-10 and TGF-β), cytolysis of T_eff cells (granzyme B), disruption of purinergic signaling by ectonucleotidases (CD39 and CD73), and inhibition of dendritic cell maturation (CTLA4) (33). First, we evaluated their functional competency in ex vivo suppression assays wherein the proliferation of WT CD4⁺CD25⁻ T_eff cells was suppressed by the presence of either WT or KO CD4⁺CD25⁺ T_reg cells cocultured at varying ratios (Fig. 5A). The number of CFSE-labeled WT T_eff cells was held constant and the indicated ratios were achieved by varying the number of T_reg cells. The cells were then activated by a mixture of anti-CD3 and anti-CD28 for 3 days, and the cell division of CFSE-labeled T_eff cells was measured by flow cytometry (Fig. 5B). Both WT and KO T_reg cells suppressed the proliferation of WT T_eff cells comparably. Subsequently, we compared the cell surface density of the key functional proteins CD39, CD73, and CTLA4, as well as the expression of genes encoding these proteins, to compare the immunophenotypic features of KO and WT T_reg cells. We found that the cell surface abundance of CD39 was less on CD4⁺FOXP3⁺ thymocytes (tT_reg) compared to that on peripheral splenic T_reg (pT_reg) cells. In the KO mice, this trend was preserved, but we observed a substantially greater percentage of CD39⁺ tT_reg and pT_reg cells than in their WT counterparts (Fig. 5C). A similar result was seen in the case of CD73 and CTLA4. thymic and splenic T_reg The MFIs of CD39 and CD73 on thymic and splenic T_reg cells and of CTLA4 on splenic T_reg cells were significantly higher in KO than in WT mice. Through qPCR analysis, we also showed that Il10 was significantly higher in KO T_reg cells than in KO T_reg cells and that Gzmb displayed a trending increase (fig. S4). The relative amounts of Ctla4, Lag3, Nt5e, and Entpd1 mRNAs were comparable but Tgfb1 mRNA was less abundant in the KO T_reg cells. Based on these data, we conclude that the expression of these genes in KO T_reg cells was comparable, but not identical, to that in WT T_reg cells. The stability of FOXP3 expression is an important feature of T_reg cells, with tT_regs showing the highest stability (34, 35). It has been proposed that tT_regs can be identified on the basis of the cell surface expression of Neuropilin-1 (Nrp-1) because this marker is expressed only at very low amounts in pT_reg cells (36, 37). We found a substantially higher percentage of NRP1⁺ T_reg cells in the thymus, but not spleens of KO mice (Fig. 5, E and F). In conjunction with the analysis of T_reg-cell–specific surface markers, gene expression, and functional competency, we conclude that the KO T_reg cells were functional and, because of their increased frequency, could contribute resistance to experimental hepatitis.

Fig. 5. Characterization of T_reg cells isolated from the thymus and lymph nodes of KO mice.

(A) The experimental scheme used for the suppression assays. (B) Analysis of CFSE-labeled WT T cell proliferation after 3 days of culture in the presence of T_reg cells isolated from WT (blue) and KO (red) mice. The ratios indicate the relative numbers of CD4⁺CD25⁻ T_eff cells to CD4⁺CD25⁺ T_reg cells plated on day 1. Histograms show the dilution of CFSE in T_eff cells as measured by flow cytometry. The percentage of T_eff cells showing dilution of the CFSE dye reflects the proportion of T_eff cells that underwent cell division. Data are representative of four experiments. (C) T_reg cells from the thymi and spleens of WT (blue) and KO (red) mice were identified by flow cytometry (Left: bivariant gating for CD4 and Foxp3) and further analyzed for the expression of the T_reg cell functional markers CD39, CD73, and CTLA4 (histograms). (D) Percentages of positive cells (left) and MFI values (right) for CD39, CD73, and CTLA4 in thymic and splenic CD4⁺Foxp3⁺ T_regs from WT and KO mice. The data were collected from two independent experiments with a total sample size of 4 mice. The P value was calculated by one-tailed unpaired t-test. (E) Cell surface expression of neuropilin-1 (Nrp-1) on thymic and splenic T_reg cells (CD4⁺Foxp3⁺) from WT (blue) and KO (red) mice. (F) Bar graphs showing the percentage of positive cells (left) and MFI values (right) for Nrp-1 in thymic and splenic T cells from WT and KO mice. The data were collected from two independent experiments with a total sample size of four mice. The P value was calculated by one-tailed unpaired t-test.

Deletion of Trpm7 in thymocytes and T cells increases their sensitivity to IL-2

IL-2 plays a crucial role in regulating the development, maintenance, and function of T_reg cells (9-11). The downstream activity of STAT5, a key transcription factor for T_reg cell function, enables stable Foxp3 gene expression (14, 15). To assess the state of IL-2-signaling in developing thymocytes, we first looked at the expression of its receptor subunit IL-2Rα (also known as CD25). We found that the Il2ra transcript abundance in KO thymocytes was five-fold greater than that in WT thymocytes (Fig. 6A). Accordingly, the percentage of CD25^hi thymocytes was also increased (Fig. 6B). Concomitantly, deletion of Trpm7 also increased the amount of IL-2 produced by the KO thymocytes. Analysis of intracellular IL-2 in Thy1.2⁺ thymocytes showed that KO thymocytes exhibited a significant increase in the MFI of intracellular IL-2 (Fig. 6, C and D). Although the KO thymocytes produced increased IL-2 in the thymus, the same was not true for the peripheral Thy1.2⁺ CD4⁺ KO T cells in the spleen and mesenteric lymph nodes (fig. S5A). The concentrations of IL-2 in the sera of WT and KO mice were also identical (fig. S5B). These results show that the deletion of Trpm7 in the T cell lineage increased the expression of IL-2 and IL-2Rα in the developing thymocytes and suggest that this increase in IL-2 signaling promoted the development of FOXP3+ T_reg cells.

Fig. 6. KO mouse thymocytes and splenic T cells show increased phosphorylation of STAT5 in response to IL-2.

(A) RT-qPCR analysis of the relative abundance of Il2ra mRNA in WT and KO thymocytes. The relative fold change was calculated after normalization to the abundance of mRNA encoding beta-2-Microglobulin (b2m). The P value was calculated by two–tailed unpaired t-test (n = 3 experiments). (B) Flow cytometry analysis of the cell surface expression of IL-2Rα (CD25) in WT and KO thymocytes. The histogram is representative of seven independent experiments. (C) Overlaid flow cytometry histograms showing intracellular IL-2 staining in Thy1.2⁺ WT (blue) and KO (red) thymocytes. The gray line denotes an FMO control. The data are representative of three independent experiments. (D) Left: Frequency of IL-2⁺Thy1.2⁺ thymocytes obtained from WT (blue) and KO (red) mice. Right: Bar graph of the MFI of IL-2 staining in Thy1.2⁺ thymocytes from the indicated mice. The P value was calculated by two-tailed unpaired t-test, error bars represent SEM (n = 3). (E) Overlaid histograms show the intracellular p-STAT5 staining in Thy1.2⁺CD4⁺ T cells obtained from WT and KO thymocytes, as indicated. The data are representative of four independent experiments (F) Percentage of p-STAT5+ cells (left) and p-STAT5 MFI values (right) in freshly isolated WT and KO thymocytes. The P value was calculated by two-tailed unpaired t-test (n = 4). (G) MFI values of p-STAT5 in freshly isolated WT and KO thymocytes. The P value was calculated by two–tailed unpaired t-test (n = 4). (H) Representative histogram (of three independent experiments) showing intracellular staining of p-STAT5 in freshly isolated WT and KO splenic T cells. FMO control is shown as a black trace. (I) Representative Bivariant contour plots (of four independent experiments) showing Foxp3 and p-STAT5 staining in CD4⁺ WT and KO thymocytes before and after stimulation with the indicated concentrations of IL-2 at the indicated times. Analysis was confined to CD4+ thymocytes; for the gating strategy used, see fig. S5C. (J) The frequency of p-STAT5+ T_reg cells as determined by the analysis shown in (I). Error bars indicate the SEM (n = 4 experiments); P values were calculated by t-test. (K) Foxp3 and pSTAT5 staining in WT and KO CD4⁺ splenocytes before and 5 min after stimulation with the indicated concentrations of IL-2. (L) The frequency of p-STAT5⁺ splenic T_reg cells quantified from the analysis shown in (K). Error bars indicate the SEM (n = 4 experiments); P values were calculated by t-test.

To confirm that IL-2 signaling was indeed potentiated, we evaluated the phosphorylation of STAT5, a transcription factor downstream of IL-2 signaling. We used flow cytometry to determine the extent of intracellular phosphorylated STAT5 (pSTAT5) staining in thymocytes obtained from WT and KO thymi (Fig. 6E). The frequency of pSTAT5⁺ cells as well as the MFI of pSTAT5, a measure of STAT5 activity per cell, were significantly higher in KO thymocytes than in WT thymocytes (Fig. 6, F and G). The increased pSTAT5 abundance was confined to thymocytes; freshly isolated splenic T cells showed no difference in pSTAT5 staining (Fig. 6H). Because the gene encoding IL-2Rα is also a transcriptional target of STAT5, we thought that KO thymocytes likely maintain high amounts of surface IL-2Rα and increased pSTAT5 through a positive, feed-forward loop. We measured the occupancy of STAT5 on the Il2ra promoter using chromatin immunoprecipitation coupled to qPCR (ChIP-qPCR). Immunoprecipitation of the cross-linked chromatin was performed with a ChIP-grade anti-pSTAT5 or control IgGα antibody. Antibody-bound genomic DNA was incubated with qPCR primers designed to amplify the two promoter sequences of IL-2Rα. A two-fold enrichment of Il2ra promoter sequence in the anti-pSTAT5 chromatin immunoprecipitates was observed in the KO thymocytes as compared to that in the WT thymocytes (fig. S5G). We hypothesized that TRPM7 restrains the JAK-STAT pathway downstream of the Il2ra and thus tested whether KO thymocytes treated with IL-2 ex vivo would exhibit increased activation of STAT5. When WT thymocytes are treated with IL-2, a very small percentage of cells were FOXP3⁺ STAT5⁺ (Fig. 6I; see fig. S5C for gating strategy). In contrast, the KO thymocytes exhibited a substantial increase in the percentage of pSTAT5⁺FOXP3⁺ cells (Fig. 6J). We also quantified the percentage of STAT5⁺ thymocytes and found that these were also increased in KO thymocytes (fig. S5D). Although we did not see a difference in basal activation of STAT5 in freshly isolated splenic T cells from WT and KO mice, we examined their sensitivity to stimulation with IL-2 ex vivo. Splenic T cells from the KO mice responded rapidly to IL-2 stimulation by phosphorylating STAT5 and showed an almost three-fold higher increase in the percentage of pSTAT5⁺FOXP3⁺ cells when compared to splenic T cells from WT mice (Fig. 6, K and L). The increased sensitivity of KO T cells to IL-2 was not limited to FOXP3⁺ cells (fig. S5E). The MFI of p-STAT5 staining was greater in IL-2–treated CD4⁺FOXP3⁺ T cells from the KO mice (fig. S5H), supporting the conclusion that the absence of TRPM7 increases the IL-2 sensitivity of all T-cells. Together, these findings demonstrate that the deletion of Trpm7 potentiates the development of T_reg cells by sensitizing the thymocytes to IL-2 stimulation and promoting the activation of STAT5, a key transcription factor for the regulation of Foxp3 transcription. Based on this model, we evaluated whether the deletion of Trpm7 or pharmacological inhibition of TRPM7 promoted the expression of Foxp3 in IL-2–activated thymocytes ex vivo.

Deletion or pharmacological inhibition of TRPM7 augments the induction of Foxp3 during the ex vivo activation of thymocytes

When WT thymocytes were activated ex vivo by anti-CD3 antibodies in the presence of IL-2, a modest increase in the abundance of Foxp3 transcripts was measured by qRT-PCR analysis (Fig. 7A). The deletion of Trpm7 further increased Foxp3 transcription (Fig. 7A, left). This increase was abrogated in the presence of a highly selective JAK3 inhibitor (CP690550) (Fig. 7A, right). We then activated the thymocytes ex vivo with anti-CD3, IL-2, or a combination of anti-CD3 and IL-2 and evaluated the differentiation of T_reg cells (CD4⁺CD25⁺FOXP3⁺) after 48 hours of stimulation. Activation by anti-CD3 antibodies and IL-2 resulted in a modest increase in the percentage of WT CD4+FOXP3+ thymocytes but the percentage of KO CD4⁺FOXP3⁺ thymocytes almost doubled (Fig. 7, B and C). The augmented differentiation of CD4⁺FOXP3⁺ cells in presence of anti-CD3 alone was likely due to paracrine production of IL-2 that activated JAK3 and STAT5 through the common γ chain.

Fig. 7. Foxp3 gene induction is enhanced by deletion or inhibition of TRPM7.

(A) Left: qRT-PCR analysis measuring the relative expression of Foxp3 in WT and KO thymocytes activated with anti-CD3 and IL-2 at the indicated times. Right: Analysis of Foxp3 induction in KO thymocytes in the presence of 2 nM CP690550 (JAK inhibitor) at the 48-hour time-point. C, untreated control. Both bar graphs denote mean values and the error bars indicate the SEM (n = 3 experiments). The P value was calculated by t-test. (B) Analysis of Foxp3 and CD25 expression on CD4+ thymocytes from WT (blue) and KO (red) mice after activation with anti-CD3, IL-2, or both, as indicated. (C) Quantification of the frequencies of CD4⁺CD25⁺Foxp3⁺ T_reg cells from the experiments shown in (B). Error bars indicate the SEM (WT: n = 3; KO: n = 4). Thymocytes were pooled from at least 2 mice in each experiment. The P values were calculated by two-tailed unpaired Student’s t-test. (D) I-V relationship of I_TRPM7 in WT thymocytes (blue trace) obtained by whole-cell, patch clamp recordings, orange trace shows inhibition of I_TRPM7 after perfusion of 2 μM FTY720 in the bath solution. Inset bar graph shows I_TRPM7 current densities for each condition, obtained 5 min after break-in at 100 mV (n = 3 experiments). (E) I-V relationship of I_TRPM7 in thymocytes during bath perfusion of 2 μM FTY720-P (S1PR agonist; orange trace). The control WT thymocyte current (blue trace) is identical to the I_TRPM7 with perfused FTY720-P. MgCl₂- treated cells (10mM, red trace) display inhibition of TRPM7 current. (F) Quantification by qRT-PCR analysis of the relative expression of Foxp3 in WT thymocytes activated for 48 hours with anti-CD3 antibody in presence of the indicated concentrations of FTY720 (left) or FTY720-P (right). Error bars indicate the SEM (n = 3 experiments). The indicated P values were calculated by unpaired Student’s t-test. (G) Representative histograms of Foxp3 staining in CD4⁺CD25⁺ WT thymocytes activated for 48 hours with anti-CD3 and IL-2 in the absence (blue) or presence of FTY720 (orange). Data are representative of three experiments. (H) Histograms representing intracellular p-STAT5 staining. Thymocytes from xxx mice were pretreated with FTY720 (orange) or left untreated (blue) and then stimulated for 15 min with the indicated concentrations of IL-2. (I) Quantification of pSTAT5⁺ thymocytes from the experiments shown in (H). Data are representative of three independent experiments in which the cells were cultured in triplicate. P values were calculated by Student’s t-test (#P = 0.005).

Because TRPM7 is a bifunctional protein (both an ion channel and a serine-threonine kinase), we investigated whether pharmacological inhibition of the channel pore was sufficient to recapitulate the effects of Trpm7 deletion. To evaluate the role of the TRPM7 channel specifically, we used FTY720, an FDA-approved sphingosine-1-phosphate receptor (S1PR)-targeting prodrug that is phosphorylated in vivo to generate FTY720-P, an analog of S1P (38). FTY720 blocks the TRPM7 channel without being further modified to its phosphorylated form (39). Because the previous report identified FTY720 as a TRPM7 channel blocker in experiments with TRPM7-overexpressing cell lines (39), we first confirmed that it blocked native I_TRPM7 in mouse thymocytes (Fig. 7D). The residual current seen in the presence of FTY720 included leak currents suggesting that we are underestimating the extent of TRPM7 inhibition. The phosphorylated FTY720 derivative (FTY720-P), which binds to S1PR (38), did not block TRPM7, as illustrated and compared to the known TRPM7 blocker Mg²⁺ (Fig. 7E). We activated WT thymocytes in the presence of FTY720 and FTY720-P to assess the effect on Foxp3 gene expression. FTY720-treated, but not FTY720-P–treated thymocytes exhibited an increase in the abundance of Foxp3 mRNA (Fig. 7F). Because FTY720-P, unlike FTY720, is an agonist of S1PR, these results indicate that the increased transcriptional regulation of Foxp3 was not mediated by S1PR. Similar to TRPM7 deletion, inhibition of TRPM7 during ex vivo activation also increased the expression of FOXP3 (Fig. 7G), and there was a concomitant increase in STAT5 phosphorylation (Fig. 7, H and I).

We then investigated whether injecting FTY720 i.p. (3 doses for 6 days at 0.1 μg/kg) could increase the number of T_reg cells in WT mice (fig. S6, C and D). WT mice displayed a ~50% increase in the percentage of T_reg cells in the thymus, which was thus comparable to the T_reg cell proportion in the KO mice (fig. S6D). Together, these data support the model that the TRPM7 channel is a negative regulator of JAK-STAT signaling downstream of IL-2R. In KO thymocytes or in WT thymocytes in which the TRPM7 channel is block, the activation of STAT5 was potentiated, resulting in increased transcription of Foxp3 and increased thymic differentiation of CD4⁺FOXP3⁺ T_reg cells (Fig. 5, E and F). Because IL-2–dependent STAT5 activation also plays a role in the induction of peripheral T_reg cells, we evaluated the ex vivo induction of splenic CD4⁺CD25⁻ cells in the presence of TGF-β1 and IL-2. We found that under these conditions, KO T cells showed a modestly increased propensity to differentiate into CD4⁺CD25⁺FOXP3⁺ T_reg cells (fig. S6, A and B).

Trpm7^fl (Lck Cre) bone marrow chimeras support a non-cell autonomous mechanism for the increased T_reg cell frequency

To gain further mechanistic insight, we performed bone marrow transplantation studies. The bone marrow from either Trpm7^fl Lck Cre or Trpm7^fl control mice was isolated and transplanted into lethally irradiated CD45.1+ WT recipient mice. After 8 weeks of convalescence, during which the transplanted hematopoietic cells repopulated the lymphoid organs, the recipient mice were euthanized for the analysis of thymi and spleens by flow cytometry (Fig.8A). We performed three different bone marrow (BM) transplantations (BMTs), all into CD45.1 WT recipients: Trpm7^fl (WT) BM, Trpm7^fl Lck Cre (KO) BM, and a 1:1 mixture of WT and KO BM (Fig.8A, right). When we analyzed thymic cellularity, it became apparent that the KO BM recipients displayed a significantly reduced number of total thymocytes (on average, a 7-fold reduction, Fig. 8B). This loss of cellularity also manifested itself as thymic involution that was visible during dissection, which mirrored the highly penetrant phenotype of the KO (Trpm7^fl Lck Cre) mice (21). Because the recipient mice were WT for TRMP7, this observation indicates that the deletion of Trpm7 in thymocytes alone contributes to the deterioration of the thymic architecture, possibly through abnormal secretion of various paracrine factors.

Fig. 8. KO BM chimeras suggest that the effects of TRPM7 deletion are both cell-intrinsic and -extrinsic.

(A) Experimental scheme (left) and bone marrow (BM) transplantation scenarios (right). (B) The absolute numbers of thymocytes obtained from WT recipient mice after transplantation of BM from WT (blue), KO (red), or mixed (black) mice. The P value was calculated by one-way ANOVA. (C) Absolute numbers of CD45⁺TCRb⁺CD4⁺CD25⁺Foxp3⁺ T_regs in the thymi obtained from the indicated recipient mice (data analyzed by one-way ANOVA). (D) Absolute numbers of CD45⁺TCRb⁺CD4⁺CD25⁻Foxp3⁻T_effs in the thymi obtained from the indicated recipient mice (data analyzed by one-way ANOVA). (E) The ratio of T_reg:T_eff absolute numbers in the thymi of the recipient mice from each transplant. P values were calculated by one-way ANOVA. (F) Total numbers of splenocytes resulting from the transplantation of WT (blue), KO (red), and mixed (black) BM into WT recipient mice. (G) The absolute numbers of CD45⁺CD4⁺CD25⁺Foxp3⁺ T_reg cells in the spleens obtained from the indicated recipient mice. (H) The absolute numbers of CD45⁺CD4⁺CD25⁻Foxp3⁻ T_eff cells in the spleens obtained from the indicated recipient mice. (I) The ratio of the absolute numbers of splenic T_reg:T_eff cells in mice that received the indicated BM transplants. P values were calculated by one-way ANOVA.

Despite the reduced number of retrieved donor thymocytes, the absolute number of thymic T_reg cells was comparable to those achieved with the WT and MIX donors (Fig. 8C). The number of T_eff cells also trended lower with KO BMT but this difference was not statistically significant with the small sample size of the BMT experiments (Fig. 8D). Unsurprisingly, in comparison to the WT or MIX recipients, the T_reg:T_eff ratio was much higher in the KO BMT recipients (Fig. 8E). To understand the relative contributions of donor and recipient populations in the overall thymic output, we quantified CD45.1⁺ and CD45.2⁺ cells in the WT, KO and MIX BMTs (fig. S7). The reconstitution efficiency was high for the WT cells in both the WT and MIX transplant scenarios (fig. S7B). The KO cells, on the other hand, yielded greatly reduced numbers of thymocytes in the KO BMT and were outcompeted by the WT CD45.1⁺ cells in the MIX BMT. Similarly, the yield of CD4⁺ T cells (fig. S7C) and T_eff cells (fig. S7D) from the KO BMT was also very low. However, when we evaluated the composition of the T_reg cells in the KO BMT, we found that the transplantation of KO BM increased the differentiation of host (CD45.1⁺) T_reg cells (fig. S7E). This result suggests that the Trpm7^−/− thymocytes enhanced the differentiation of WT thymocytes and that non cell-autonomous mechanisms underlay the increased T_reg population in the Trpm7-targeted mice.

We also assessed the corresponding T cell populations in the recipient spleens to further define this phenomenon. The absolute number of splenocytes was similar in the case of the WT, KO and MIX BMTs (Fig. 8F). Thus, as was reported earlier (21), the loss of thymic cellularity caused by the deletion of Trpm7 in thymocytes did not have a major negative effect on peripheral T cell numbers. In comparison to the WT BMT, the number of splenic T_reg cells was modestly higher in the KO and MIX BMTs, but the difference was not statistically significant (Fig. 8G). The number of T_eff cells in the KO BMT was comparable to that in WT cell recipients (Fig. 8H) but the splenic T_reg:T_eff ratio was higher in the case of KO BMT when compared to the WT and MIX chimeras (Fig. 8I). We also examined the splenic cells for their origins by binning the cells into CD45.1⁺ and CD45.2⁺ populations (fig. S7, F and G). Overall, the host population of T_reg cells was significantly increased in the KO BM recipients as compared to that in the WT BM recipients (fig. S7G). These results indicate that the presence of transplanted Trpm7^−/− T cells potentiated the number of T_reg cells and that the underlying mechanism is predominantly non-autonomous in manner. At its simplest level, this is best explained by the IL-2 overproduction by the Trpm7^−/− T cells, but we cannot exclude other paracrine factors. Together, these results are consistent with our proposed model that the deletion of TRPM7 promotes T_reg cell development by increasing IL-2 production and thus promoting the subsequent IL-2–dependent FOXP3 expression in the thymocytes.

Discussion

The major insight derived from this study is the function of the ion channel TRPM7 in controlling the development and frequency of tT_reg cells. This serendipitous discovery emerged from the observation that the deletion of Trpm7 in lineage cells rendered the mice highly resistant to Con A–induced hepatitis, an experimental form of autoimmune hepatitis that is driven by T cells. We propose that the increased generation of tT_reg cells in these mice is an important contributing factor, but not the only contributing factor, that protects these mice in this disease model. Additionally, we also showed that although Ko T_eff cells were activated normally, they showed differences in their cytokine output. We showed that KO thymocytes were competent in immunosuppression assays. Future studies will comprehensively characterize the physiological and immunosuppressive properties of Trpm7^−/− tT_reg cells in the context of other models of T cell–driven autoimmune diseases. Because FTY720 is already an FDA-approved drug, the findings have implications for translational research as well as medicinal chemistry approaches for the development of more selective blockers of TRPM7.

Our results suggest that increased IL-2 signaling in the KO thymocytes underlies the increased output of tT_reg cells. The KO thymocytes produced significantly increased amounts of IL-2 and they also showed increased expression of the IL-2R subunit CD25. Bone marrow transplantation experiments supported the model that the mechanism that drives the increased differentiation of tTreg cells is predominantly cell-extrinsic. The increased production of IL-2 by KO thymocytes and increased cell surface CD25, which is encoded by a STAT5 target gene (40), likely contributes in a feed-forward loop to increase IL-2 sensitivity in the KO T cells. Consistently, during ex vivo activation, the KO thymocytes showed increased sensitivity to IL-2 and showed inducible synthesis of FOXP3. In our conception, this increased sensitivity to IL-2 in ex vivo conditions is primarily a result of the high quantity of CD25 on the cell surface, but it remains possible that TRPM7 also regulates signals downstream of the IL-2R. In the lymph nodes, IL-2 signaling and STAT5 mediate a crucial function in enabling T_reg cells to constrain the local expansion of conventional T cells into damage-causing T_eff cells (41). Future studies will evaluate whether increased sensitivity to IL-2 makes the KO T_reg cells more competent in restraining inflammatory cascades in organ systems such as the liver, gut, and skin.

The mechanism through which the channel activity of TRPM7 influences IL-2 production in the developing thymocytes remains unclear at this point and will be a major focus of our future studies. During T cell activation, Il2 transcription is predominantly controlled by the SOCE-dependent NFAT pathway. In mice in which SOCE, through Orai channels, is abolished selectively in the T cell lineage, the mice fail to develop an adequate number of T_reg cells (42). In essence, the functions of the Orai and TRPM7 channels are designed to be in opposition during the process of T_reg cell development. Hence, the general design principles underlying the checks and balances in immune regulation appear to be preserved even at the scale of rapid electrical signaling (43). From the standpoint of the two-step model (16) of T_reg cell development, Ca²⁺ entry through the Orai channels is likely essential for the strong, TCR-dependent signals that drive NFAT activation and the subsequent increased abundance of IL-2, IL-2R, and possibly other components of the IL-2 signaling pathway in the developing thymocytes. As shown here, deletion of TRPM7 had no discernable effect on SOCE amplitude but did have a modest effect on its kinetics. Whether this is enough to substantially enhance NFAT activity (and increased Il2 expression) is an outstanding question. It is also possible that TRPM7 influences Il2 transcription independently of SOCE. In addition to conducting Ca²⁺, TRPM7 also conducts Zn²⁺, and the role of Zn²⁺-signaling in T cell biology is not yet well understood. TRPM7 channel activity is polymodally regulated by changes in internal Mg²⁺ concentration (44), phospholipids (45), pH (46), and C-terminal proteolysis (24, 47). These regulatory properties of TRPM7 may enable the developing thymocytes to integrate multiple signals in the thymic microenvironment to tune tT_reg cell development.

Because FTY720 is already an FDA-approved drug, our findings have implications for translational research, as well as for medicinal chemistry approaches for more selective blockers of TRPM7. In the context of S1PR pharmacology, FTY720 functions as a prodrug, which is phosphorylated in vivo to generate FTY720-P, which is an S1PR ligand. Binding of S1PR by FTY720-P, but not by FTY720, promotes the sequestration of lymphocytes in lymph nodes, resulting in immunosuppression. Through patch clamp recordings of thymocytes, we confirmed that TRPM7 channel is blocked by FTY720, but not by FTY720-P; this enables the inhibition of TRPM7 without modulating S1PR. In presence of FTY720, but not FTY720-P, stimulation of activated thymocytes with IL-2 increased the transcription of Foxp3 and promoted the ex vivo differentiation of T_reg cells. Note that treating human patients with FTY720 increases the frequency of T_reg cells (48), suggesting that at least some component of FTY720-mediated immunosuppression may be mediated by increasing the frequency and function of T_reg cells through TRPM7 channel inhibition. Although the involvement of the channel activity of TRPM7 was indicated by experiments involving FTY720, the role of the kinase domain of TRPM7 cannot be ruled out. Cleavage of TRPM7 at Asp¹⁵¹⁰ increases the ion channel activity and liberates a functional kinase domain (M7CK-S) (24), whereas cleavage sites N-terminal of Asp¹⁵¹⁰ inactivate the channel and liberate a longer kinase domain (M7CK-L) that translocates to the nucleus and modifies chromatin through histone phosphorylation (23). It remains possible that during T cell development, the kinase domain plays an important role in reprogramming the chromatin landscape. Our study opens the door toward a deeper investigation of TRPM7 and other ion channels in the development, function, and stability of T_reg cells. In future, it may be possible to increase the number of T_reg cells in vivo with TRPM7-targeting drugs and thereby induce tolerance in patients suffering from autoimmunity.

Materials and Methods

Mice

Mice were housed and bred in accordance with policies and protocols approved by the Institutional Animal Care and Use Committee (IACUC) of University of Virginia. Male and female mice aged between 4-12 weeks were used for the experiments. Trpm7^fl mice wherein exon 17 of Trpm7 is flanked by LoxP sites and Trpm7^fl(Lck Cre) mice which deletes Trpm7 selectively in T cells were described by us previously (21).

Genotyping of mice

Tail samples were dissolved at 85°C for 30 min in 75 μl of digestion buffer (25 mM NaOH, 0.2 mM EDTA), and the reaction was stopped by adding 75 μl of neutralization buffer (40 mM Tris-HCl). One microliter was used as a template for PCR amplification with MyTaq Hot Start Polymerase (Bioline; #BIO-21112) and 5× MyTaq Red Reaction Buffer (Bioline; #BIO-37112). PCR products were separated in a 1% agarose gel and visualized with ethidium bromide. The presence of Cre recombinase was determined using the following primers: Forward primer Cre S2F (5’-GATTTCGACCAGGTTCGTTC-3’) and reverse primer Cre S5R (5’-GCTAACCAGCGTTTCGTTC-3’). The presence (fl) or absence (WT) of LoxP sites flanking exon 17 was detected with the following primers: Forward primer Geno 2F* (5’-CAGAGGTACTGGCAATTGTG-3’); Reverse primer Geno 2R* (ACGAGGACTCAGCATATAGC-3’).

Reagents

The following antibodies were used: anti-mouse neuropilin-1-PE (R&D Systems, #FAB5994P), rat anti-mouse CD25 PE-Cy7 (BD Pharmingen, #552880), anti-mouse CD39 PE-Cyanine7 (eBioscience, Inc. #25-0391), anti-mouse CD73 PE (eBioscience, Inc. #12-0731), rat anti-mouse CD4 FITC (BD Biosciences #553046), purified hamster anti-mouse CD28 (BD Biosciences #553294), anti-mouse/rat Foxp3 APC clone FJK-16s (eBioscience), anti-Foxp3 (eBioscience, #17-5773), purified hamster anti-mouse CD3ε chain, clone 145-2C11 (BD Biosciences #553058), anti-mouse CTLA-4 (CD152) PE (eBioscience, #12-1522), and anti-mouse CD16/CD32 antibody (Biolegend,TruStain fcX, #101320). The following sets of magnetic beads were used for T cell isolation: Dynabeads Untouched Mouse T Cells (Life Technologies, #11413D), CD4⁺ CD25⁺ Regulatory T Cell Isolation Kit mouse (Miltenyi Biotech, #130-091-041), and Dynabeads FlowComp Mouse CD4⁺ CD25⁺ T_reg Cells (Life technologies, #11463D). We used the SensiMix SYBR Kit (Bioline USA Inc, #QT605-20), GoScript Reverse Transcription System (Promega Corporation, #A5001), and the ISOLATE II RNA Mini Kit (Bioline, #BIO-52073). Human TGF-β1 was obtained from Cell Signaling Technology (#8915) and recombinant human IL-2 was from BD Biosciences (#BDB554603). CFSE was from Life Tech, γ-irradiated Con A from Canavalia ensiformis was from Sigma-Aldrich, FTY720 was from Cayman Chemical, and 7-AAD was from BD Biosciences (#559925).

Con A–induced hepatitis model

Age-matched and sex-matched WT and KO mice were weighed 2 hours before injection to control for body weight differences. Con A (40 mg/kg) or PBS was administered intravenously (i.v) through tail vein injection. Animals were monitored for 24 hours after injection. Animals were euthanized at the times indicated in the figures or at 24 hours. Blood was collected by cardiac puncture for the measurement of serum ALT in a blinded manner.

Serum ALT measurements

Blood was collected from mice by cardiac puncture and centrifuged at 1500g at 4°C for 10 min to collect serum. Samples with signs of hemolysis were discarded. Sera were preserved at −80°C and analyzed within 48 hours of collection. ALT activity was detected with the Liquid ALT (SGPT) Reagent kit (Pointe Scientific, Inc.) according to the manufacturer’s instructions.

Histology

After perfusion with PBS, a portion of liver tissue was fixed by immersion in 4% paraformaldehyde (PFA) for 24 hours and transferred to 30% sucrose for 24 hours before tissue analysis. Tissue embedding, sectioning, and mounting were performed by the UVA Research Histology Core (University of Virginia, Charlottesville, VA). Tissues were embedded in frozen tissue blocks before sectioning. H&E staining was performed by the UVA Research Histology Core, and slides were photographed and analyzed on an Aperio Scanscope (Leica Biosystems).

Immunofluorescence microscopy

Liver tissue sections were prepared on slides, rehydrated in water, washed with PBS, 0.05% Tween20, and permeabilized for 15 min with PBS containing 0.1% Triton X-100. Slides were then blocked with blocking buffer (PBS with 1% BSA, 0.1% fish gelatin, 0.1% Triton-X-100, 0.05% Tween 20, and 5% animal serum, corresponding to the host species of the secondary antibody). Slides were stained with rat anti-mouse CD4 (1 μg/ml, Clone RM4-5, Biolegend) in blocking buffer overnight at 4°C. Slides were then washed three times in wash buffer and stained with Alexa Fluor–594 conjugated goat anti-rat secondary antibody (0.2 μg/ml, Life Technologies) and solubilized in blocking buffer for 2 hours at room temperature. After being washed three times, the cells were stained with Hoechst 33342 nuclear dye (0.2 μg/ml) for 10 min at room temperature. Finally, the slides were washed twice and coverslips (FischerBrand, 22 mm #1) were mounted over VectaMount mounting medium (Vector Laboratories). Mounted samples were cured overnight at room temperature and imaged within 24 hours. All images were captured on a Zeiss AxioObserver microscope with a Zeiss 40X (1.3 NA) objective lens and Hamamatsu CMOS OrcaFlash 4.0 sCMOS camera. Images were then processed with NIH ImageJ software.

Scanning electron microscopy (SEM)

Livers were perfused through the portal vein with 2.5% glutaraldehyde and 4% PFA in PBS at a constant flow rate (1 ml/min). Liver sections of 5-mm thickness were removed and immersed in 2.5% glutaraldehyde and 4% PFA for 48 hours at 4°C. Samples were then washed in Osmium tetroxide (OsO₄) and dehydrated in an ethanol series before critical point drying. Samples were then mounted on aluminum studs and sputter coated in gold/palladium before being subjected to SEM on a Zeiss Sigma VP HD field emission SEM microscope (Carl Zeiss AG) at 3 kV.

Isolation of intrahepatic immune cells

Livers were perfused with PBS through the portal vein. A portion of the liver was removed for immunohistological, immunofluorescent, and qPCR analysis. PBS-perfused livers were excised, washed in Iscove’s Modified Dulbecco’s Medium (with 10% FBS), and digested in 0.05% collagenase (Collagenase IV, Sigma-Aldrich) for 25 min at 37°C. Mononuclear cells (MNCs) containing immune cells were purified with a 40% Histodenz (Sigma-Aldrich) gradient after centrifugation at 1500g for 20 min at 4°C. Spleens, thymi, and lymph nodes were physically disrupted in cold, sterile RPMI-1640. Cell suspensions were filtered through 70-μm nylon mesh, and contaminating erythrocytes were lysed with ACK buffer [150 mM NH₄Cl, 10 mM KHCO₃, 100 μM Na₂EDTA (pH 7.3)]. These cell suspensions were then subjected to appropriate magnetic separation to isolate different subsets. We followed the protocols included in these kits: Dynabeads Untouched Mouse T Cells (Life Technologies, #11413D), CD4⁺ CD25⁺ Regulatory T Cell Isolation Kit mouse (Miltenyi Biotech, #130-091-041), and the Dynabeads FlowComp™ Mouse CD4⁺ CD25⁺ T_reg Cells (Life technologies, #11463D).

Flow cytometry

Erythrocyte-free, single-cell suspensions isolated from livers, spleens, or thymi were processed in FACS staining buffer composed of 150 mM PBS, 0.5% BSA, and 2 mM EDTA. Fc receptors on the cells were blocked with anti-mouse CD16/CD32 antibody (Biolegend,TruStain fcX, # 101320) for 20 min and then stained with fluorochrome-conjugated antibodies. For intracellular staining of Foxp3, the cells were fixed in 1% PFA and permeabilized with 0.1% saponin. Intracellular staining of p-STAT5 was performed after concomitant fixation and permeabilization with 100% methanol. The stained cells were analyzed with a FACSCalibur or LSRFortessa flow cytometer (Becton Dickinson). The data were acquired by CellQuest Pro Acquisition software or FACSDiva v6 Software (Becton Dickinson), respectively. Every experiment included single-stain and a “fluorescence minus one” (FMO) controls to facilitate accurate compensation and analysis with FlowJo version 10 software (Tree star).

T-cell activation, CFSE proliferation assay, and transwell-migration assay

T cells were cultured at a density of 0.5 X 10⁶ cells/ml in a culture volume of 200 μl/well in 96-well round bottom plates. The cells were activated with Con A (5 μg/ml) for 48 hours or as indicated in the figure legends. Alternatively, cells were activated with a mixture of anti-CD3e (2.5 μg/ml) and anti-CD28 (1 μg/ml) for 48 hours. After activation, cells or the cell culture medium were harvested for analysis. For CFSE proliferation assays, the T cell subsets were cultured at a reduced density (50,000 cells/well) in 96-well round bottom plates. Proliferation was then stimulated with a mixture of anti-CD3e (0.25 μg/ml) and anti-CD28 (0.125 μg/ml) and analysis was performed after 3 days of culture. For transwell-based migration assays, each transwell (5.0 μm pore size membrane) from Corning costar incorporation contained 5X10⁵ splenic T cells from WT or KO mice suspended in 100 μl of complete medium (top chamber). Transwell inserts were placed in a 24-well tray containing 600 μl of variably diluted conditioned medium containing chemotactic signals secreted by LPS-treated bone marrow–derived macrophages (BMDMs). During the incubation of the Transwell plates (16 hours at 37° C, 5% CO₂), the T cells transmigrated to the bottom chamber. They were manually counted using trypan blue. Data are presented as a percentage of the T cells that migrated across the membrane.

Luminex multiplex and ELISA assays

Cell culture supernatants were thawed from −80°C and were analyzed in duplicate to measure cytokine concentrations. The data were acquired and analyzed with the Luminex multiplex assay (Luminex 100 IS). Standards were assayed for each cytokine with QC1 and QC2 in duplicate with beads count of <25. The cytokine concentrations were calculated by curve-fitting five concentrations of standards. ELISA assays were performed according to the manufacturer’s instructions.

Real-time quantitative PCR (qPCR) analysis

The concentrations of total RNA isolated from cells with the ISOLATE II RNA Mini Kit (Bioline, #BIO-52073) were measured with a NanoDrop 2000c spectrophotometer (Thermo Scientific). The cDNA was reverse-transcribed with the GoScript Reverse Transcription System (Promega, #A5001). The qPCR experiments were always set in triplicate with SensiMix SYBR Hi-ROX Kit (Bioline, #QT605-05) in a CFX connect Real-Time system (Bio-Rad). After 40 cycles of PCR amplification, the data were analyzed with CFX manager 3.1 software (Bio-Rad). The primers used for qPCR analysis are shown in table S1. Data were analyzed with the ΔΔCt method (52) using the average Ct values of β2-microglobulin (b2m) to normalize for cDNA input error. Statistical analysis was based on 3 independent experiments.

ChIP assays

Equal numbers of freshly isolated thymocytes from WT and KO mice were cross-linked with 1% formaldehyde for 10 min before the cross-linking reaction was stopped with the addition of glycine. The cytoplasmic contents were extracted and removed with cell lysis buffer containing 1% SDS. The nuclear pellets were resuspended in ChIP buffer (1.1% TritonX-100 containing nuclear lysis buffer). These nuclear pellets were subjected to sonication with a Bioruptor Pico (Diagenode Inc.). The sheared chromatin was then immunoprecipitated for 12 hours at 4°C with ChIP-grade antibodies [anti-pSTAT5 (Tyr⁶⁹⁴) and control-IgG] and ChIP-grade protein G magnetic beads (Cell Signaling Technologies). The immunoprecipitated samples were then eluted with ChIP elution buffer and treated with proteinase K for 12 hours at 65°C to release the DNA fragments. The released DNA was purified with a column-based DNA purification kit (Qiagen) and subjected to qPCR with the following primers. For Il2ra promoter site 1: Forward: TGACAGACTGAGAGGCCTGA; reverse: TGGGTCAACCCCTTTACTGC. For Il2ra promoter site 2: Forward: TTTGACGTCGGGTGTCTTCC; reverse: GTGGAACTCTGGGTTCAGCA. The data were then analyzed to assess the occupancy of STAT5 on Il2ra promoters in KO thymocytes relative to that in WT thymocytes. Briefly, the Ct values obtained from mock IgG ChIPs were first used to calculate the fold-change in anti-STAT5 ChIPs in WT and KO samples. The KO STAT5 ChIPs showed significantly higher fold-change values compared to those in the WT samples. These are shown as bar graphs after normalizing the WT values to 1.

Ca²⁺ imaging in T cells

The cells were loaded with Fura-2 AM and the recording was performed in 2 mM CaCl₂. The baseline was recorded in the presence of biotinylated anti-CD3e antibody, and the Ca²⁺ response was observed by adding streptavidin. Ionomycin (Ca²⁺ ionophore) served as a positive control. The 340/380 nm fluorescence ratio was background-subtracted and normalized to baseline (Δ F/F₀). The rise-time (tau) of fluorescence was derived and plotted. The number of cells assessed is outlined in the figures. Ca²⁺ imaging data was collected on a Zeiss AxioObserver microscope with Zeiss 40X (1.3 NA) objective lens and Hamamatsu CMOS OrcaFlash 4.0 sCMOS camera.

Electrophysiology

The TRPM7 current (I_TRPM7) was measured in the whole-cell configuration (fig. S3F). T cells were activated with a mixture of anti-CD3e and anti-CD28 antibody (or Con A) and cultured in 96-well plates at 37°C, 5% CO₂ for 1 to 4 days before being subjected to patch-clamp electrophysiology. The standard external solution contained: 135 mM sodium methanesulfonate (Na-MeSO₃), 5 mM cesium gluconate, 2.5 mM CaCl₂, 10 mM HEPES (pH 7.3, adjusted with NaOH); Osmolality: 280 to 290 mOsm/Kg. The standard pipette solution contained 110 mM Cs-gluconate, 0.5 mM NaCl, 0.75 mM CaCl₂, 10 mM HEPES, 10 mM HEDTA, 1.8 mM Cs₄-BAPTA, 2 mM Na₂ATP (pH 7.3, adjusted with CsOH). Osmolality was 273 mOsm/Kg and the calculated free [Ca²⁺] = ~100 nM. We used the maxchelator algorithm to calculate free Ca²⁺: https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcS.htm. FTY720 (2 μM) or MgCl₂ (10 mM) was used in the external solution to inhibit TRPM7 currents. All currents were recorded with an Axopatch 200B amplifier (Molecular Devices). The recording protocol consisted of 400-ms ramps from −100 mV to +100 mV and a holding potential (HP) of 0 mV. The signals were low-pass filtered at 5 kHz and sampled at 10 kHz. All electrophysiology experiments were performed at room temperature (~22°C). The average current densities were plotted with the relevant statistical information as a box chart.

T_reg cell-based suppression assay

Splenic cell suspensions were isolated from WT and KO mice. T_eff or T_reg cells were then isolated by negative and positive magnetic selection with Dynabeads Flow Comp Mouse CD4⁺ CD25⁺ T_reg Cells kit (Lifetech). T_eff cells were labeled with 5 μM CellTrace CFSE (Carboxyfluorescein succinimidyl ester) for 20 min, washed and cultured in RPMI-1640 medium supplemented with 2 mM L-glutamine, 10% FBS, 50 μM β-mercaptoethanol, 1% sodium pyruvate, 5 mM HEPES, 100 U/ml penicillin, and 100 μg/ml streptomycin. The CFSE-labeled T_eff cells were cultured in 96-well round bottom plates and stimulated with a mixture of anti-CD3e and anti-CD28 antibodies in 5% CO₂ humidified incubator at 37°C for 3 days. The number of T_eff cells was kept constant at 5 x 10⁴ cells per well, whereas the number of cocultured T_reg cells varied to generate the ratios indicated in the figures. On day 4, the cells were collected and stained for 7-AAD, CD4, and CD25 before undergoing flow cytometry analysis on a FACSCalibur instrument. The data were analyzed with FlowJo software and represented as the percentage of T_eff cells showing proliferation-induced dilution of the CellTrace CFSE dye. This parameter reflects the percentage of cells undergoing cell division.

Ex vivo generation of T_reg cells from thymocytes

Thymocytes were isolated from WT and KO as single-cell suspensions and cultured in X-vivo 15 growth medium (Sartorius Biotech, Cat #04-744Q) supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin. Thymocytes were activated for 48 hours with anti-CD3ε antibody in presence of recombinant hIL-2 (10 ng/ml) in 96-well round bottom plates at 5% CO₂, 37°C. The thymocytes were then collected and were either used to isolate mRNA for Foxp3 mRNA analysis by qPCR or were analyzed by flow cytometry for immunophenotyping and staining for intracellular Foxp3.

Bone marrow transplantation

Bone marrow was harvested from adult mice (12 to 15 weeks of age) and the cells were counted. Equal numbers of the bone marrow hematopoietic cells from the WT and the KO mice were transplanted into the lethally irradiated (1000 cGy) syngeneic recipients, which expressed the CD45.1⁺ marker for the ease of differentiation between the donor and the recipient cells. The recipient mice were allowed to recover for 8 weeks, which also enabled the hematopoietic cells to repopulate the periphery. Mice were then euthanized and the thymi and spleens were collected for further analysis.

Statistical analysis

Box charts were plotted using data analysis and graphing software Origin Pro 9.1.0 (Origin Lab). Statistical box charts in Fig. 4C, Fig. 4G, and fig. S3B are shown as Box (range of 25- to 75-percentile), whisker bars (1 to 99-percentile), and data overlap. Each data point including outliers is shown along with the median (horizontal line) and the arithmetic mean (empty square). Box charts in Fig. 4D, H, K, and J are shown as box (range of 25- to 75-percentile) and whisker bars (SD). All data points including outliers are shown with the median (horizontal line) and mean (empty square) in the box. P values were calculated using a two-tailed unpaired Student’s t-test and as shown in each figure. Sample size equals the number of data points shown in the box chart.

Supplementary Material

Supplementary Material

Fig. S1. Activation and migration of KO T cells.

Fig. S2. Gating scheme used to derive the Foxp3 histograms.

Fig. S3. Quantification of T_reg cells in the thymi and spleens of WT and KO mice.

Fig. S4. Expression of selected mediators in WT and KO T_reg cells.

Fig. S5. Thymocytes and splenic T cells from KO mice show increased IL-2 sensitivity.

Fig. S6. Inhibition of TRPM7 augments Foxp3 expression.

Fig. S7. Gating strategy and quantification of CD45.1⁺ and CD45.2⁺ cells from BM chimeras.

Table S1. Primer sequences used in the real-time qPCR experiments.