Trsp is required by regulatory T cells to prevent lethal autoimmunity in mice

Abstract

Selenoproteins are involved in immune cell metabolism, yet the roles of these proteins in T cell development and function remain largely unknown. The Trsp gene encodes the selenocysteine tRNA (tRNA^Sec) required for translation of all selenoproteins. In this study, we found that Trsp was required for thymopoiesis, with the majority of tRNA^Sec-deficient T cells not progressing beyond double negative 3 stage, with egressed thymocytes undergoing peripheral homeostatic expansion. Trsp-deficient CD4⁺ T cells exhibited impairments in TCR and IL-2 signaling and did not cause inflammation in experimental models. On the other hand, Trsp-deficient regulatory T (Treg) cells exhibited defects in suppressive function ex vivo and Treg-specific Trsp deletion using Trsp^fl/flFoxp3^YFP-Cre (Trsp^ΔTreg) mice caused fatal autoimmunity similar to FOXP3-deficient mice. Reducing oxidative stress via 2-HOBA administration prolonged survival in these Trsp^ΔTreg mice. These findings indicate that tRNA^Sec is required for T cell homeostasis and may be therapeutic targets in inflammation.

One sentence summary

Trsp, a gene required for translation of all selenoproteins, is essential for all T cell development and function, especially regulatory T cells.

Introduction

T cells are a critical component of adaptive immunity and their dysregulation is associated with various pathologies including infection, autoimmunity, and cancer. Both the development and function of T cells are tightly regulated by their metabolic capacity.(1–6) Specifically, degree of oxidative stress has been shown to regulate physiologic activity of T cells. For instance, low concentrations of reactive oxygen species (ROS) can promote T cell activation and proliferation, whereas higher ROS concentrations can inhibit transcription factors essential for T cell activation(7) and subsequent downregulation of transcription of interleukin-2.(8) One protein family that contributes to cellular metabolism and redox state is selenocysteine-containing proteins (selenoproteins), most of which are antioxidants.(9) Translation of the 25 known selenoproteins in humans (24 in mice) requires the selenocysteine tRNA (tRNA^Sec) encoded by TRU-TCA1–1/Trsp along with specialized translation factors that collectively coordinate selenocysteine insertion.(10) Thus, TRU-TCA1–1/Trsp regulates the expression of the entire class of selenoproteins.(11) While human and murine T cells express selenoproteins, there is a preponderance for expression of specific family members, suggesting distinct cell-specific roles.(12)

Selenoproteins are essential for normal growth and development and although there are no reported null mutations in TRU-TCA1–1, individuals with a single nucleotide polymorphism (65C>G) that reduces tRNA^Sec levels and activity in vitro exhibit thyroid hormone imbalances.(13–15) In mice, deletion of Trsp is embryonic lethal.(16) Therefore targeted deletion of Trsp in specific cell types is required to investigate selenoprotein function collectively, which is particularly important as individual selenoproteins can compensate for each other.(17) In T cells, selenoproteins have previously been studied using proximal Lck^Cre-mediated excision of Trsp that revealed a reduction in the frequency of mature CD4⁺/CD8⁺ single positive (SP) T cells in the thymus as well as in the periphery (here defined as extrathymic) and reduced activation/proliferation ex vivo.(18) While selenoproteins are known to be important for thymocyte development, it is currently unknown when selenoproteins are expressed during intrathymic T cell development and how this affects progression through each stage of thymopoiesis. Moreover, the metabolic requirements for T cells differ among various T cell subsets.(19) Although physiological levels of ROS are required for IL-2 production in activated T cells,(20, 21) the regulatory roles of selenoproteins in T cell subtype function are not known. Thus, we sought to determine the effect of en-bloc deletion of murine selenoproteins in T cells on thymopoiesis and T cell function both in vitro and in vivo.

Results

CD2⁺ cells require Trsp for progression from DN3 to DN4 during thymopoiesis

To study the contributions of selenoproteins during thymopoiesis, mice deficient for Trsp in all T cells were generated by crossing Trsp^flox/flox mice(22) with CD2^iCre mice (Trsp^ΔCD2 mice) to allow deletion of selenoproteins in double negative, CD4⁺CD8⁺, and CD4⁺ thymocytes. Trsp^flox/flox mice have loxP sites flanking the Trsp gene that encodes tRNA^Sec (23). Without tRNA^Sec, selenoprotein transcripts are not translated.(23) To explore whether Trsp-deficiency affects T cell development, we examined the thymus in age-matched littermates. Compared to control mice, the thymus of Trsp^ΔCD2 mice was nearly absent (Figure 1A), similar to what was reported in using Lck^Cre-mediated excision of Trsp.(24) Correspondingly, the absolute number of thymocytes was decreased in Trsp^ΔCD2 mice (Figure 1B). Thus, Trsp expression in CD2⁺ cells is required for thymopoiesis. To define which thymopoietic stages require Trsp, developing T cell subsets of control and Trsp^ΔCD2 mice were defined by flow cytometry (FC). In the thymus, T cell development progresses through distinct stages before T cells become double positive (DP) and subsequently single positive (SP) CD4⁺ or CD8⁺ T cells (as reviewed in (25)). In Trsp^ΔCD2 mice, the frequency of double negative (DN) thymocytes was increased with a concomitant decrease in the frequency of DP thymocytes (Figure 1C). This finding indicates that either the progression from DN to DP cells is dependent on Trsp expression, or that progression within the 4 distinct DN stages may be impaired when Trsp is absent in CD2⁺ cells, ultimately resulting in fewer DP cells. Lineage negative (CD19^negLy6C^negTCRγδ^negNK1.1^negCD11c^neg) DN thymocyte stages are distinguished based on cell-surface expression of CD25 and CD44: DN1 (CD44⁺CD25^neg), DN2 (CD44⁺CD25⁺), DN3 (CD44^lowCD25⁺), and DN4 (CD44^negCD25^neg) as reviewed in (26). In Trsp^ΔCD2 mice, the frequency of T cells at the DN3 stage was increased while cells at the DN4 stage were decreased compared to control mice (Figure 1D). Collectively, these data suggest that Trsp expression is required for progression through DN stages of thymocyte development with fewer cells progressing from DN3 to DN4 in Trsp^ΔCD2 mice.

Figure 1: CD2⁺ cells require Trsp for thymopoiesis and T cell homeostasis.

(A) Male Trsp^fl/fl mice were crossed with female Trsp^fl/wtCD2^Cre mice to generate Trsp^fl/flCD2^Cre (Trsp^ΔCD2) mice. Representative photo of the thoracic cavity of a 5-week old control (Trsp^wtCD2^Cre) and Trsp^ΔCD2 mouse (age- and sex-matched). Arrows indicate thymus (black) and thymus remnant (blue).

(B) Absolute cell count of live lymphocytes in the thymus acquired using automatic cell counter. Pooled data from 2 independent experiments. Unpaired t-test.

(C) Representative FC plots (left) and quantification (right) of thymocytes. Pooled data from 2 independent experiments. Mixed model with post-hoc Šídák’s multiple comparisons test. Lineage^neg for thymocytes indicates the following markers: CD19^negLy6C^negTCRγδ^negNK1.1^negCD11c^neg.

(D) Representative FC plots (left) and quantification (right) of DN (CD4^negCD8^neg) thymocyte subsets. Pooled data from 2 independent experiments. Mixed model with post-hoc Šídák’s multiple comparisons test.

(E) FC histogram of CD2 expression in WT (Trsp^wtCD2^wt) mouse CD45⁺Lin^negCD4^negCD8^neg double negative thymocyte subsets. Representative results of 2 independent experiments.

(F) Thymocytes were flow-sorted (DN1: CD45⁺CD4^negCD8^negCD44⁺CD25^neg; DN2: CD45⁺CD4^negCD8^negCD44⁺CD25⁺; DN3: CD45⁺CD4^negCD8^negCD44^negCD25⁺; DN4: CD45⁺CD4^negCD8^negCD44^negCD25^neg) from WT mice for selenoprotein RT-qPCR. Normalized expression is 2^−ΔC(t). Pooled data from 2 independent experiments with total n=6 biological replicates.

(G) CellROX fluorescence in thymocyte subsets from control and Trsp^ΔCD2 mice. Representative histograms (left) and quantification (right). Data from 2 independent experiments performed using the same protocol including the same antibodies, same cytometer (Cytek), and same cytometer settings. Mixed model with post-hoc Šídák’s multiple comparisons test.

(H) Annexin-V⁺ T cells gated and quantified into DN1-DN4 as in (D). Mixed model with post-hoc Šídák’s multiple comparisons test.

(I) Absolute cell count of live lymphocytes in the spleen acquired using automatic cell counter. Pooled data from 2 independent experiments. Unpaired t-test.

(J) Representative FC plots (left) and quantification (right) for CD4 and CD8 in splenocytes from control and Trsp^ΔCD2 mice. Pooled data from 2 independent experiments. Unpaired t tests.

(K) Representative FC plots (left) and quantification (right) of splenic T cells (CD45⁺CD3⁺CD4⁺) from control and Trsp^ΔCD2 mice. Subsets defined as follows: TCM: CD44⁺CD62L⁺; Tnaïves: CD44^negCD62L⁺; TEM: CD44⁺CD62L^neg. Pooled data from 2 independent experiments. Unpaired t tests.

(L-M) Immunoseq was performed on flow-sorted (L) thymic CD4⁺ SP cells (CD45⁺CD4⁺CD8^neg) and (M) splenic naïve T cells. Data from one experiment. Unpaired t-tests.

* p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001.

CD2 and selenoproteins are expressed throughout thymopoiesis

To determine whether CD2 expression temporally coincides with alterations in the frequency of DN3 and DN4 populations in Trsp^ΔCD2 mice, we examined Cd2 expression across thymocyte stages in wild-type (WT) mice. We analyzed a previously published single-cell RNA-seq dataset of mouse DN thymocytes(27) and found that Cd2 expression was readily detected in putative DN1 thymocytes (Figure S01). To confirm, CD2 expression was analyzed on thymic cells from WT mice. A bimodal distribution was observed, with CD2 being high in DN1, low/absent in DN2 and DN3, and high in DN4 (Figure 1E). This CD2 expression pattern is congruent with the concept of CD2 being required for signaling through the pre-T cell receptor (TCR) complex,(28) as DN3 is the first stage at which thymocytes start expressing TCR on their cell surface.(29)

Since CD2 is expressed during DN1, we profiled selenoprotein expression throughout thymopoiesis. To this end, we employed real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR).(30) Several selenoproteins were expressed during all DN stages, including DN1. The selenoproteins with the highest expression levels were Gpx1, Selenow, Selenoh, and Selenof. Interestingly, certain selenoproteins such as Gpx1, Selenow, and Selenof were highly expressed in DN1/DN2 cells, whereas others (e.g., Selenoh) were only detected in later stages (Figure 1F). To examine whether human thymocytes exhibited a similar expression pattern, we analyzed a single-cell RNA-seq dataset of human thymocytes(31) and observed that some of the same selenoproteins were highly expressed in earlier DN stages (GPX1, SELENOH, SELENOF) compared with later stages (Figure S02). Thus, select selenoproteins are expressed by developing thymocytes in mice and humans and may be required for proper thymocyte development.

Thymocytes require Trsp to maintain physiological redox balance and prevent aberrant apoptosis

Having determined that Trsp^ΔCD2 mice exhibit altered thymopoiesis, we next interrogated the potential mechanism underlying this observation. Most selenoproteins function by reducing oxidative stress, which can induce apoptosis (reviewed in (32)). Thymocyte apoptosis is central to thymopoiesis as T cells that do not receive a permissive TCR signal die instead of developing into DP cells. We hypothesized that selenoproteins function to reduce oxidative stress-induced apoptosis during thymopoiesis. First we measured ROS in Trsp^ΔCD2 mice by FC using CellROX. As expected, ROS levels were higher during all DN stages in Trsp^ΔCD2 mice as compared to control mice (Figure 1G). Since the thymi in Trsp^ΔCD2 mice were smaller than in WT controls, we hypothesized that this increase in ROS may drive aberrant apoptosis. As CellROX histograms did not segregate into distinct negative and positive populations (Figure 1G), we were not able to compare the frequency of Annexin-V⁺ cells in cells with high versus low levels of ROS. Therefore, we examined which DN cells undergo apoptosis by measuring the frequency of Annexin-V⁺ cells. In WT mice, the majority of Annexin-V⁺ cells were in the DN4 stage, whereas in Trsp^CD2 mice most of the Annexin-V⁺ cells were observed at the DN3 stage (Figure 1H), consistent with the tRNA^Sec-deficient T cells not progressing beyond DN3. Collectively, these data suggest that CD2⁺ cells require selenoproteins to maintain physiological redox state for appropriate thymocyte development through the DN stages.

tRNA^Sec-deficient T cells that egress from the thymus of Trsp^ΔCD2 mice undergo homeostatic proliferation

Having established that Trsp^ΔCD2 mice have decreased thymic output, we wanted to determine if this leads to lymphopenia. We examined the spleen and mesenteric lymph nodes (mLN) of Trsp^ΔCD2 mice and found a reduction in the absolute number of lymphocytes in the spleen of Trsp^ΔCD2 mice with lower frequencies of CD4⁺ T cells and higher frequencies of CD8⁺ T cells compared to control mice (Figure 1I-J). We then examined CD4⁺ T cell subsets, (i.e., naïve T (TN), T effector memory (TEM) or T central memory (TCM) cells) and found that while TN frequency was reduced, both TEM and TCM frequencies were increased (Figure 1K). Similar findings were observed in the mLNs, with the notable exceptions that there was no change in CD3⁺ T cells, and there were no increases in CD8⁺ T or TCM cell frequencies (Figure S03). This pattern of altered CD4⁺ T cell subsets with increased memory T cells and decreased naïve T cells is consistent with peripheral tRNA^Sec-deficient CD4⁺ T cells undergoing homeostatic proliferation.(33) Furthermore, the TCR repertoire during thymic development and post-egress determined via TCR CDR3 sequencing revealed that the number of productive TCRs in the CD4⁺ SP thymocytes trended lower in tRNA^Sec-deficient T cells (Figure 1L) while the clonality of the splenic naïve T cells was increased (Figure 1M). This is consistent with a TCR repertoire of Trsp^ΔCD2 mice that is less diverse than WT mice. These data indicate that selenoprotein-deficient T cells fail to undergo appropriate positive selection in the thymus and that the few T cells that successfully egress undergo homeostatic proliferation in the periphery.

Pathogenic CD4⁺ T cells require Trsp

Our observations indicate that T cells require Trsp for proper T cell development. Since Trsp^ΔCD2 mice have dysregulated CD4⁺/CD8⁺ T cells, we generated mice that allow for inducible deletion of selenoproteins specifically in CD4⁺ T cells (Cd4^Cre-ERT2;Rosa26^{lox-stop-lox-tdTomato};Trsp^fl/fl mice). Repeated tamoxifen gavage resulted in >90% CD4⁺ T cells expressing tdTomato, indicative of Cre-mediated recombination (Figure S04A). We confirmed the downstream consequences of Trsp deletion by Western blot for the selenoproteins glutathione peroxidase 4 (GPX4) and thioredoxin reductase 1 (TXNRD1), and observed marked reductions in GPX4 and TXNRD1 protein levels in Trsp^ΔCD4 as compared to WT CD4⁺ T cells (Figure S04B). To test whether T cells require Trsp for pathogenicity, we employed the adoptive transfer model of colitis(34) wherein naïve T cells (live, CD4⁺CD45RB^high) from Trsp^ΔCD4 or control mice are injected into lymphocyte-deficient Rag1^−/− recipient mice (Figure 2A). The frequency of injected tdTomato⁺ cells did not differ between tRNA^Sec-deficient or -sufficient naïve T cells (Figure 2B), indicating that a similar frequency of cells underwent inducible Cre-mediated excision of the Rosa26 lox-stop-lox sequence prior to adoptive transfer. At the experimental endpoint, Rag1^−/− mice that received tRNA^Sec-deficient naïve T cells were protected against colitis (Figure 2C). This was not attributed to differences in engraftment or expansion, as the frequency of splenic CD4⁺CD3⁺ T cells did not differ between recipient groups (Figure 2D). Moreover, tRNA^Sec-deficient T cells recovered from recipient mice and restimulated ex vivo with PMA/ionomycin produced IFNγ to a similar degree as tRNA^Sec-sufficient T cells (Figure 2E). Since T cell activation is a critical driver of colitis, this finding suggests that Trsp deficiency may lead to a TCR signaling defect, as cytokine production remained intact when bypassing TCR signaling with PMA/ionomycin. To test this, we utilized a T cell-independent colitis model mediated by dextran-sodium sulphate (DSS) and found no differences in weight loss, colon length, or histological injury between Trsp^ΔCD4 and control mice (Figures 2F-H). Thus, Trsp is required for CD4⁺ T cell pathogenicity in vivo, likely downstream of TCR activation.

Figure 2: T cells require Trsp for TCR activation and pathogenicity.

(A) Cd4^Cre-ERT2;ROSA^{lox-stop-lox-tdTomato};Trsp^wt/wt (control) and Cd4^Cre-ERT2;ROSA^{lox-stop-lox-tdTomato};Trsp^fl/fl (Trsp^ΔCD4) mice were gavaged with 8 mg tamoxifen on days 0, 1 and 3. Naïve T cells (CD45RB^high) were then flow-sorted and intraperitoneally injected into sex-matched Rag1^−/− mice.

(B) Frequency of ROSA⁺ T cells in sorted cells prior to injection into Rag1^−/− recipients. Pooled data from 3 independent experiments. Within each experiment, samples were pooled per genotype prior to sorting; each datapoint represents 3–6 mice. Mann-Whitney U test.

(C) Representative histology (left) and quantification (right) of colitis across 3 independent experiments 7 weeks after injection of naïve T cells. Mann-Whitney U test.

(D) Frequency of splenic T cells in mice that received Trsp-deficient or -sufficient naïve T cells. Measured at experimental endpoint. Mann-Whitney U test.

(E) Representative FC plots (left) and quantification (right) of splenic T cells restimulated with PMA/ionomycin for 5 hours in the presence of Golgi-Stop and stained for IFNγ. Mann-Whitney U test.

(F) Baseline-normalized weight curves for control and Trsp^ΔCD4 mice gavaged with tamoxifen (8 mg x 3 doses) that subsequently received 3.5% DSS in their drinking water for 5 days followed by 5 days regular water, after which mice were euthanized. Mann-Whitney U test.

(G) Colon length at experimental endpoint. Mann-Whitney U test.

(H) Representative histology (left) and colitis score (right). Mann-Whitney U test.

(I) CD4⁺ T cells were magnetically enriched from spleen and mLN of control and Trsp^ΔCD4 mice and stimulated using plate-bound αCD3 and soluble αCD28 for 2 hours, then lysed and subjected to a T cell phosphoprotein assay. Shown are the fold changes comparing Trsp-deficient to -sufficient T cells. n=9 per genotype; one experiment.

(J) Western blot of control and Trsp^ΔCD4 CD4⁺ T cells stimulated as in (I) for 15 minutes for LCK, MEK1, pLCK^Tyr394, pMEK1^Ser298, and vinculin (loading control). Representative data from 2 independent experiments.

(K) Representative FC plots (top) and quantification (bottom) of control and Trsp^ΔCD4 CD4⁺ T cells stimulated with IL-2 for 15 minutes and stained for pSTAT5. Subsets defined as follows: TCM: CD44⁺CD62L⁺; Tnaives: CD44^negCD62L⁺; TEM: CD44⁺CD62L^neg. Representative data from 2 independent experiments. Mann-Whitney U tests.

(L) ELISA for IL-2 on supernatant of control and Trsp^ΔCD4 CD4⁺ T cells after 2 days in culture. Unpaired t test.

* p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001.

Trsp is required for T cell receptor activation

To determine experimentally whether Trsp modulates pathways downstream of TCR activation, Trsp-sufficient/deficient CD4⁺ T cells were stimulated using αCD3 and αCD28 followed by measurement of 213 TCR-relevant antibodies using a phosphoprotein assay. When ranked by fold change, phosphorylation of MEK1 at Ser298 (decreased) and phosphorylation of LCK at Tyr394 (increased) were identified as the two most altered in tRNA^Sec-deficient as compared to control CD4⁺ T cells (Figure 2I). These results were corroborated via Western blot analysis (Figure 2J). In addition, IL-2 production is downstream from TCR activation, and IL-2 is an important regulator of T cell metabolic function.(35) Thus, we assessed whether Trsp deletion in CD4⁺ T cells alters both response to IL-2 and production of IL-2 after activation. Stimulation with exogenous IL-2 in vitro resulted in reduced phosphorylation of signal transducer and activator of transcription 5 (pSTAT5) (Figure 2K). Interestingly, not only was the response to exogenous IL-2 reduced, but IL-2 production was also reduced in Trsp^ΔCD4 as compared to control CD4⁺ T cells following activation (Figure 2L). Altogether, these data show that CD4⁺ T cells require Trsp for activation of several signaling pathways downstream of TCR engagement. Moreover, tRNA^Sec-deficient CD4⁺ T cells secrete less IL-2 after TCR activation and are less responsive to exogenous IL-2. These findings likely explain the inability of Trsp-deficient naïve CD4 T cells to cause colitis in Rag1^−/− mice.

tRNA^Sec-deficient Treg cells exhibit an activated phenotype but are functionally impaired

Since Trsp was required for TCR activation in Trsp^ΔCD4 mice, and Trsp^ΔCD2 mice were lymphopenic, we next sought to identify other CD4⁺ T cell subsets that require Trsp. We hypothesized that the function of Trsp in Treg cells differs from other CD4⁺ T cell subsets as we observed an increase in the frequency of splenic FOXP3⁺ Treg cells in Trsp^ΔCD2 mice (Figure 3A). Trsp^ΔCD2 Treg cells also exhibited a more active and proliferative phenotype based on expression of MHCII, Ki67, and CD69 (Figure 3B). Expression of RORγt was also increased in both CD4⁺FOXP3⁺ (Figure 3B) and CD4⁺FOXP3^neg T cells from Trsp^ΔCD2 mice (Figure S05). Interestingly, CD4⁺FOXP3^neg, but not CD4⁺FOXP3⁺, cells from Trsp^ΔCD2 mice also expressed elevated levels of CD25, the high-affinity component of the IL-2R, which may lead to higher affinity for IL-2 as compared to their WT counterparts (Figure S05). This may, in part, explain the increase in proliferation of CD4⁺FOXP3^neg T cells of Trsp^ΔCD2 mice as measured by Ki67 (Figure S05). Prior work established that RORγt⁺ Treg cells are more suppressive then RORγt^neg Treg cells(36, 37). Therefore, we assessed the suppressive capacity of Treg cells from Trsp^ΔCD2 mice in vitro and found that suppression by tRNA^Sec-deficient Treg cells was modestly impaired compared to WT Treg cells (Figure 3C). Collectively, these data suggest that Trsp plays cell-intrinsic roles in Treg cell function.

Figure 3: Treg cells in Trsp^ΔCD2 mice have an activated phenotype and reduced suppression.

(A) Representative FC plots (top) and quantification (bottom) of splenic T cells from control and Trsp^ΔCD2 mice. Pooled data from 2 independent experiments. Unpaired t-test.

(B) Representative FC gating plots (top) and quantification (bottom) of splenic Treg cells from control and Trsp^ΔCD2 mice. Pooled data from 2 independent experiments. Unpaired t tests.

(C) Schematic (left), representative FC histograms (middle), and quantification (right) of in vitro Treg cell suppression using flow-sorted CD25⁺ T cells from control and Trsp^ΔCD2 mice and naïve T cells from control mice in a 1:1 ratio of Treg: naïve T cells. Each pair of dots (i.e. control and Trsp^ΔCD2-derived Treg cells) represents one individual experiment. One-sided paired t test.

* p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001.

Mice with tRNA^Sec-deficient Treg cells develop fatal autoimmunity

Although Trsp^ΔCD2 mice exhibit impaired thymopoiesis, they do not spontaneously develop overt inflammation. Nevertheless, they do exhibit increased Treg cell frequency, and these Treg cells appear to be activated with increased RORγt⁺ expression, yet they display defects in suppressive function in vitro. To investigate these Treg cell functional defects directly, we generated Trsp^fl/flFoxp3^YFP-Cre (Trsp^ΔTreg) mice. Homozygous deletion of Trsp specifically in FOXP3⁺ cells resulted in an autoimmune phenotype (i.e. scaly tail, early mortality, skin inflammation, and hair loss) similar to FOXP3-deficient Foxp3^scurfy mice (Figure 4A-B).(38) Of note, gross phenotypes of Trsp^ΔTreg mice differed slightly from those of Foxp3^scurfy mice, in that Trsp^ΔTreg mice did not develop the characteristic thickened ears (not shown). Trsp^ΔTreg mice also displayed thymic involution and multiorgan inflammation, as observed in Foxp3^scurfy mice (Figure 4C-D), along with splenomegaly, decreased lymphocytes, and decreased splenic Treg cells (Figure 4E-F). Importantly, this reduction in splenic Treg cells could not be attributed to increased apoptosis, as measured by Annexin-V staining (Figure S06). Additionally, Trsp^ΔTreg and control mice displayed similar frequencies of splenic CD3⁺, CD8⁺, and CD4⁺ cells (Figure 4F). This systemic inflammation was also associated with decreased splenic B cells and increased serum IgE levels (Figure 4G), both characteristics of autoimmune inflammation, and similar to Foxp3^scurfy mice (Figure S07). Of note, Trsp^ΔTreg mice did not develop spontaneous colitis, which is also not reported in Foxp3^scurfy mice(39).

Figure 4: Treg cells require Trsp to prevent lethal autoimmunity.

Trsp^fl/flFoxp3^YFP-Cre mice were generated by crossing Foxp3^YFP-Cre mice with Trsp^fl/fl mice.

(A) Representative photo of control and Trsp^fl/flFoxp3^YFP-Cre mouse at 4.7 weeks of age.

(B) Survival curve of control, Trsp^fl/flFoxp3^YFP-Cre, and Foxp3^scurfy mice with Kaplan-Meier analysis. Pooled data from 5 independent experiments. n=22 control, n=9 Foxp3^scurfy, n=6 Trsp^fl/flFoxp3^YFP-Cre. Log-Rank (Mantel-Cox) test, p<0.0001.

(C) Absolute cell count of live thymocytes from control and Trsp^fl/flFoxp3^YFP-Cre mice acquired using automatic cell counter at 4 weeks of age. Pooled data from 3 independent experiments. Unpaired t-test.

(D) Representative histology of liver, skin and lung of control, Trsp^fl/flFoxp3^YFP-Cre, and Foxp3^scurfy mice.

(E) Photo of spleen (left) and absolute cell count (right) of live splenic lymphocytes from control, Trsp^fl/flFoxp3^YFP-Cre, and Foxp3^scurfy mice. Representative data from 3 independent experiments.

(F-J) FC phenotyping of (F) splenic T cells, (G) splenic B cells and serum IgE, and (H-J) splenic Treg cells. Pooled data from 4 independent experiments. Unpaired t-tests.

(K) RNA-sequencing was performed on FC-sorted splenic Treg cells from control and Trsp^fl/flFoxp3^YFP-Cre mice. DESEQ2 was used to assess differentially expressed genes comparing Trsp^fl/flFoxp3^YFP-Cre versus control Treg cells. n=3 control and n=4 Trsp^fl/flFoxp3^YFP-Cre.

(L) Schematic representing bone marrow chimera experiment. Rag1^−/− mice were sublethally irradiated (450 rads) and transplanted using a 1:1 mixture of lineage-depleted bone marrow cells from Foxp3^YFP-CreCD45.1 and Trsp^fl/flFoxp3^YFP-CreCD45.2 mice. Control mice received a 1:1 mixture of cells from Foxp3^YFP-CreCD45.1 and Foxp3^YFP-CreCD45.2 mice.

(M) Recipient mice were analyzed 6 weeks post-transplantation with the YFP⁺ fraction gated and the CD45.2⁺/CD45.1⁺ ratio quantified. Data representative of 2 independent experiments with N=3–6.

(N-Q) Trsp^fl/flFoxp3^YFP-Cre and control mice received 2-HOBA in the drinking water starting at postnatal day 1–2.

(N) Representative photo of control and Trsp^fl/flFoxp3^YFP-Cre mice at indicated ages.

(O) Survival curve of control and Trsp^fl/flFoxp3^YFP-Cre mice treated without (same data from B) or with 2-HOBA from birth with Kaplan-Meier analysis. Log-rank (Mantel-Cox) test. n=22 control without 2-HOBA, n=6 Foxp3^YFP-Cre; Trsp^fl/fl without 2-HOBA, n=62 control with 2-HOBA, n=12 Foxp3^YFP-Cre; Trsp^fl/fl with 2-HOBA p<0.0001.

(P) Representative histology of liver, skin and lung of control and Foxp3^YFP-Cre; Trsp^fl/fl mice euthanized 58 days after 2-HOBA treatment.

(Q) FC quantification of splenic Treg cells from control and Trsp^fl/flFoxp3^YFP-Cre mice treated without (same data from F) or with 2-HOBA from birth at the end of experiment. Mixed model with post-hoc Šídák’s multiple comparisons test.

p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001.

CD4⁺FOXP3⁺ Treg, but not CD4⁺FOXP3^neg, cells from Trsp^ΔTreg mice also exhibited an activated and proliferative phenotype (i.e. increased Ki67 and MHCII), along with increased ROS as measured via CellRox (Figures 4I-J, S08). To ascertain transcriptomic differences between Trsp-sufficient/deficient cells, we performed RNA-seq on sorted YFP⁺ (FOXP3⁺) cells from Trsp^ΔTreg and control mice. One gene that was significantly downregulated in tRNA^Sec-deficient Treg cells was the WNT transcription factor Lef1 (Figure 4K). Lef1 is both a WNT transcription factor and a WNT target gene and absence of Lef1 renders Treg cells unable to suppress.(40) The decreased expression of Lef1 in tRNA^Sec-deficient Treg cells was confirmed by RT-qPCR (Figure S09). Since the frequency of Treg cells in Trsp^ΔTreg mice was reduced as compared to control mice (Figure 4F), we investigated whether Trsp expression provides a competitive advantage to Treg cells. Using congenically-marked mixed bone marrow chimeras, we found that the majority of Treg cells that persist in conditioned Rag1^−/− mice were Trsp-sufficient (Figures 4L-M).

Given elevated oxidative stress in Treg cells of Trsp^ΔTreg mice, we tested whether the autoimmune phenotypes could be rescued via antioxidant treatment. During ROS-induced lipid peroxidation, toxic isoleuvoglandins are formed that can be scavenged by 2-hydroxybenzylamine (2-HOBA), reducing ROS-induced damage.(41) 2-HOBA is currently under clinical investigation for several autoimmune/inflammatory disorders.(42) Thus, we tested whether 2-HOBA in the drinking water could ameliorate the autoimmune phenotypes of Trsp^ΔTreg mice. At 6 weeks, Trsp^ΔTreg mice that received 2-HOBA were still smaller in size compared to Trsp-sufficient mice but did not develop skin ulcerations, lethargy, or scaly tail (Figure 4N) that were observed in Trsp^ΔTreg mice without 2-HOBA treatment by 5 weeks of age (Figure 4A). However, by 8 weeks Trsp^ΔTreg mice receiving 2-HOBA did begin to develop these phenotypes. Survival of Trsp^ΔTreg mice receiving 2-HOBA was prolonged compared to untreated Trsp^ΔTreg mice (Figure 4O) and inflammation in the skin and lung was less severe in 2-HOBA-treated Trsp^ΔTreg mice compared to untreated (Figure 4P). Splenic Treg cell numbers in Trsp^ΔTreg mice were higher with 2-HOBA treatment compared to Trsp^ΔTreg mice not administered 2-HOBA (Figure 4Q). Importantly, 2-HOBA treatment had no effect on splenic Treg numbers in control mice (Figure 4Q). Altogether, these data indicate that Treg expression of selenoproteins is required to prevent fatal autoimmunity in mice, possibly attributed to impaired Lef1 expression and increased oxidative stress.

Treg cells require Trsp for stability, suppression, and survival in an IL-2-dependent manner

Treg cells can lose Foxp3 expression and consequently their suppressive function to become pathogenic ex-Treg cells.(43, 44) It is possible that the systemic inflammation observed in Trsp^ΔTreg mice is due to conversion of tRNA^Sec-deficient Treg cells to pathogenic ex-Tregs. In female Trsp^fl/flFoxp3^YFP-Cre/WT mice, which have both Trsp-sufficient and -deficient Treg cells due to random X-inactivation(45), the majority of CD4⁺CD25⁺ T cells were YFP^neg (~75%), as compared to Foxp3^YFP-Cre/WT control mice, in which 50% of CD4⁺CD25⁺ T cells were YFP^neg (data not shown). These data suggest that Treg cells lose their identity in Trsp^fl/flFoxp3^YFP-Cre mice. To examine how loss of Trsp affects Treg cells and their propensity towards becoming ex-Treg cells, we generated lineage-tracing Trsp^fl/flFoxp3^{eGFP-Cre-ERT2}Rosa26^{lox-stop-lox-tdTomato} reporter mice and compared them to control Foxp3^{eGFP-Cre-ERT2};Rosa26^{lox-stop-lox-tdTomato} mice. In these mice, FOXP3-expressing cells are constitutively marked by GFP expression, and, upon tamoxifen administration, become irreversibly labeled with tdTomato. Repeated tamoxifen treatment reduced splenic Treg cell frequency in Trsp^fl/flFoxp3^{eGFP-Cre-ERT2}Rosa26^{lox-stop-lox-tdTomato} as compared to control mice (Figures 5A-B). The frequency of ex-Treg cells (tdTomato⁺GFP^neg) trended upwards in Trsp^fl/flFoxp3^{eGFP-Cre-ERT2}Rosa26^{lox-stop-lox-tdTomato} mice, but did not achieve statistical significance (Figure 5C). When we tested GFP⁺tdTomato⁺ Treg cells for suppressive function, we again observed that tRNA^Sec-deficient Treg cells had a trend towards reduced suppressive capacity (Figure 5D). We then determined if loss of Treg cell stability (i.e. conversion to ex-Tregs) associated with Trsp deficiency is sufficient to cause inflammation in lymphopenic mice. To directly test this, GFP⁺tdTomato⁺ Treg cells were sorted and transferred into Rag1^−/− recipient mice (Figure 5E). Both Trsp-deficient and -sufficient Treg cells did not cause weight loss or colitis (Figure 5F and data not shown), indicating that tRNA^Sec-deficient (ex-)Treg cells are not pathogenic in a cell-intrinsic manner. However, when compared to mice that received (WT) naïve T cells, the frequency of Treg cells in mice that received tRNA^Sec-sufficient Treg cells was increased to a greater extent as compared to mice that received tRNA^Sec-deficient Treg cells (Figure 5G), suggesting that in a non-competitive setting, Treg cells need Trsp for sustenance.

Figure 5: Treg cells require Trsp for stability, suppression, and survival in an IL-2-dependent manner.

(A) Foxp3^{eGFP-Cre-ERT2};ROSA^{lox-stop-lox-tdTomato};Trsp^wt/wt (control) and Foxp3^{eGFP-Cre-ERT2};ROSA^{lox-stop-lox-tdTomato};Trsp^fl/fl were gavaged with 8 mg tamoxifen 8 times 4 days apart.

(B) Splenic Treg cell frequency in mice from (A). Unpaired t-test.

(C) Representative FC plots of ex-Treg cells (left) and quantification (right). Pooled data from 3 independent experiments. Unpaired t-test.

(D) In vitro Treg cell suppression using flow-sorted GFP⁺tdTomato⁺ Treg cells from control and Foxp3^{eGFP-Cre-ERT2};ROSA^{lox-stop-lox-tdTomato};Trsp^fl/fl mice and naïve T cells from control mice in a 1:1 ratio of Treg:naïve T cells. Pooled data from 2 independent experiments. Unpaired t-test.

(E-G) Treg cells (ROSA⁺GFP⁺) were adoptively transferred into Rag1^−/− mice. These mice did not develop colitis (not pictured).

(E) Schematic showing experimental design.

(F) Weight curve of Rag1^−/− mice that received control naïve T cells, control Treg cells, or Trsp^ΔTreg Treg cells. n=7–8 per condition.

(G) Splenic Treg cell frequency in Rag1^−/− mice that received control naïve T, control Treg cells, or Trsp^ΔTreg Treg cells at the experimental endpoint. Kruskal-Wallis test.

(H-I) Splenic Treg cells from control and Foxp3^{eGFP-Cre-ERT2}Rosa26 ^{lox-stop-lox-tdTomato}; Trsp^fl/fl mice were FC sorted and cultured in vitro with indicated concentrations of human recombinant IL-2 for 48 hours before (H) FC. One experiment. Mixed model with post-hoc Šídák’s multiple comparisons test.

(I) CellTiter glow quantification of cells in (H). Data was normalized to media only control. Representative data from two independent experiments. Mixed model with post-hoc Šídák’s multiple comparisons test.

* p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001.

As Treg cells are uniquely sensitive to IL-2 due to expression of the trimeric, high-affinity IL-2R, our observations in Trsp^Treg mice raised the question of whether the advantage of Trsp-sufficient Treg cells is IL-2-dependent. To this end, splenic Treg cells were sorted from Trsp^fl/flFoxp3^{eGFP-Cre-ERT2}Rosa26^{lox-stop-lox-tdTomato} mice and Foxp3^{eGFP-Cre-ERT2}Rosa26^{lox-stop-lox-tdTomato} mice and cultured in the presence of increasing doses of IL-2 without TCR stimulation. No differences were observed in the MFI of GFP (FOXP3) in tdTomato⁺ tRNA^Sec-deficient versus - sufficient Treg cells (Figure 5H), indicating that Trsp is not required for IL-2-mediated FOXP3 stability in vitro. Nevertheless, survival of tRNA^Sec-deficient Treg cells was reduced as compared to tRNA^Sec -sufficient Treg cells in the presence of varying concentrations of IL-2 (Figure 5I). In summation, these data show that Trsp is required for Treg cell suppression, Treg cell stability (IL-2-independent), and Treg cell survival (IL-2-dependent).

Discussion

In this report, we demonstrate that Trsp is required for normal thymopoiesis, peripheral CD4⁺ T cell activation, and regulatory T cell survival and function. More specifically, we illustrate that Trsp deletion in CD2⁺ cells impairs thymopoiesis by reducing progression of T cell progenitors from DN3 to DN4, resulting in lymphopenia and T cell exhaustion. In peripheral CD4⁺ T cells, Trsp deficiency impairs TCR activation, IL-2 signaling, and effector T cell function. Finally, we show that Trsp deletion specifically in Treg cells causes Treg cell deficiency and fatal autoimmune inflammation through increased oxidative stress. These phenotypes recapitulate those of mice with loss-of-function mutations in Foxp3.(46–48) As Trsp encodes a tRNA required for translation of all selenoproteins, these data position Trsp and selenoproteins as central to T cell physiology.

An earlier study suggested that Trsp expression is required for Lck-expressing thymocytes, yet did not define the consequences of Trsp deletion on thymopoeisis.(18) We found that Trsp is particularly important in the DN3 to DN4 transition and that selenoproteins (particularly Gpx1, Selenow, and Selenof) are expressed as early as DN1, while expression of Txnrd1 in (all) DN thymocytes is relatively low as compared to other selenoproteins. Others have shown that WT non-Cre-expressing DN1 thymocytes have a competitive advantage over Cre-expressing Txnrd1-deficient DN1 thymocytes after transplantation into WT irradiated mice.(49) Despite increased oxidative stress in DN1 thymocytes, we did not observe this reduced DN1 cellularity in Trsp^fl/flCD2^Cre mice. Though these are seemingly discrepant findings, the advantage of WT DN1 thymocytes as compared with Txnrd1-deficient DN1 thymocytes was measured in a different context (i.e. post-irradiation and transfer as compared with from birth). Nevertheless, our data is consistent with their finding that selenoproteins may be relevant for pre-TCR signaling, as we demonstrated that there is CD2 expression in murine DN thymocytes and found it is highly expressed by DN1 cells, thus supporting a role for CD2 pre-TCR, as has been suggested.(28) Overall, there is sparse literature on how individual selenoproteins affect thymopoeisis, and much more work is required to determine their roles.

The observation that Trsp^ΔCD2 mice have substantial aberrations in thymopoiesis and peripheral T cell proportions, studying the functional consequences of Trsp deletion in conventional CD4⁺ T cells was only possible utilzing inducible Cre based approaches. For example, memory T cells respond differently to external stimuli, i.e. TCR stimulation, then naïve T cells, and Trsp^ΔCD2 mice have significantly more memory T cells than WT mice. Thus, to test the function of Trsp in conventional CD4⁺ T cells, we utilized cells isolated from Cd4^Cre-ERT2;Rosa26^{lox-stop-lox-tdTomato};Trsp^fl/fl mice. To mimic TCR stimulation in vitro, we treated these Trsp-sufficient and -deficient CD4⁺ T cells using αCD3/αCD28, since T cells require signal via the CD3 protein complex and costimulatory signal via the CD28 protein complex, which is typically provided by antigen-presenting cells. We found that phosphorylation of MEK1 on Serine 298 is decreased in Trsp-deficient CD4⁺ T cells, which suggests decreased MEK1 autophosphorylation, decreased ERK activity,(50) and thus decreased T cell function.(51) This finding is consistent with a prior publication which found that ERK signaling was reduced in Trsp-deficient T cells.(18) We also found that phosphorylation of LCK on Tyrosine 394 is increased in Trsp-deficient CD4⁺ T cells, whereas phosphorylation of LCK on Tyrosine 505 is decreased, suggesting that LCK activity is increased in the absence of Trsp.(52) This is in accordance with others who have shown that T cells require Selenok for calcium signaling downstream of TCR signaling.(53) To what extent Trsp-deficient CD4⁺ T cell function can be rescued by restoring individual components of the TCR signaling cascade remains to be determined.

Alterations in TCR signaling lead to functional impairment as the adoptive transfer of Trsp-deficient naïve T cells into Rag1^−/− mice did not cause colitis, likely due to disabling TCR-dependent activation. Consistent with this conclusion, cytokine production itself is intact when TCR is bypassed in these adoptively transferred cells. Yet, Trsp function in CD4⁺ T cells does not seem limited to TCR activation per se. Trsp-deficient CD4⁺ T cells respond to IL-2 stimulation with less phosphorylation of STAT5 then Trsp-sufficient CD4⁺ T cells. This observation is true across all CD4⁺ T cell subsets, including in antigen-inexperienced naïve T cells and in the absence of TCR activation. We did not test whether established colitis in the adoptive transfer model can be (partially) rescued by Trsp deletion, but such an experiment would be important to begin to explore whether modulation of Trsp in CD4⁺ T cells might be a therapeutic option to reduce inflammation in patients with autoimmune diseases.

In immune homeostasis, Treg cells suppress conventional CD4⁺ T cells. Similar to tRNA^Sec-deficient conventional T cells, tRNA^Sec-deficient Treg cells were also dysfunctional and lead to fatal autoimmunity. This finding is significant, as only deletion of CTLA-4(54) and IL-2R(55) subunits have previously been shown to phenocopy the autoimmunity observed in Foxp3^scurfy mice. Understanding the mechanisms behind tRNA^Sec-deficient Treg cell dysfunction is crucial. Our data suggest that increased oxidative stress renders Treg cells incapable of effective suppression, since 2-HOBA partially rescues this phenotype. In vitro suppression data indicate that Treg cells require Trsp during TCR-mediated activation and further investigation into how these processes occur is ongoing. Our findings seem to oppose reports that GPX1-deficient Treg cells are superior in suppressing T cell proliferation in vitro.(56) Yet, tRNA^Sec-deficient Treg cells may be dissimilar from GPX1-deficient Treg cells in several aspects such as levels of oxidative stress and expression of other compensatory selenoproteins. Similar to conventional CD4⁺ T cells, tRNA^Sec-deficient Treg cells also exhibit impaired responses to IL-2. IL-2 signaling is particularly important for Treg cell function, as they require IL-2(55) produced by activated effector T cells. Ex vivo, Trsp-deficient Treg cells demonstrate decreased survival as compared to Trsp-sufficient Treg cells. Based on these observations, Trsp may be a target for reducing the immune suppressive function of Treg cells in cancers, potentially allowing for increased immunosurveillance.

Important next steps in the field will be to determine which selenoproteins are required for differentiation into specific T cell subsets.(57) The extent to which specific selenoproteins affect individual Th subsets and Treg cell function have not been fully defined. While it has been shown that Gpx1^−/− mice displayed enhanced Th1/Th17 polarization with a concomitant decrease in Th2 polarization,(58) and that Gpx1^−/− CD25⁺ Treg cells) suppressed naïve T cell proliferation induced by αCD3 and WT dendritic cells to a greater degree than WT CD25⁺ Treg cells in vitro,(56) there is little data systematically determining which selenoproteins are most important in each Th subset. These are areas of ongoing investigation in our laboratory.

Limitations of study

One limitation of this study is the deletion of all selenoproteins by targeting Trsp. Although this approach prevents confounding from compensatory effects of other selenoproteins, it also hinders the ability to distinguish the contributions of individual selenoproteins. Another limitation is that this study largely focused on the role of selenoproteins in the murine immune system and the translational relevance of these findings remains to be determined. Finally, for part of our studies we used a CD2^Cre mouse. CD2 expression is not limited to T cell progenitors, as it is also expressed by B cells. This is likely important as others have implicated selenoproteins in B cell development.(59, 60)

Material and Methods

Mice

Mice lacking Trsp in specific immune cells were generated by crossing Trsp^flox/flox (22) mice with various Cre lines:

CD2^Cre mice(61) (The Jackson Laboratory, Bar Harbor, ME, JAX stock #008520); Cd4^Cre-ERT2 mice(62) (JAX stock #022356); Foxp3^{eGFP-Cre-ERT2} mice(63) (JAX stock #016961); and Foxp3^YFP-Cre mice(64) (JAX stock #016959). As Foxp3 is located on the X chromosome, we used Foxp3^YFP-Cre hemizygous males and homozygous females for all experiments. To enable lineage tracing, the tamoxifen-inducible Cre mice (i.e. Cd4^Cre-ERT2 and Foxp3^{eGFP-Cre-ERT2}) were crossed with Rosa26^{lox-stop-lox-tdTomato} mice(65) (JAX stock #007914). Mice were administered 8 mg tamoxifen (Millipore Sigma, St. Louis, MO, #T5648) in 200 μL olive oil (Millipore Sigma, O1514) with 1% ethanol (Millipore Sigma, #1085430250) by oral gavage. To prevent bias from tamoxifen gavage and its associated Cre recombination(66), both knockout and control mice received tamoxifen. Trsp^flox/floxFoxp3^YFP-Cre (and control) mice were monitored by independent observers blinded to experiment and genotype, and euthanized when deemed moribund. Macroscopic photos were taken using an iPhone (Apple, Cupertino, CA).

Genotyping for Trsp was done in-house using the following primers (Millipore Sigma) and the following protocol: CKNO2: 5’-GCAACGGCAGGTGTCGCTCTGCG-3’; 8RP: 5’-CGTGCTCTCTCCACTGGCTCA-3’. PCR conditions: 94°C for 5 min; 30 cycles of: 94°C for 30 sec, 58°C for 60 sec, 72°C for 60 sec; 72°C for 3 min. Using these primers and protocol, the floxed Trsp gene yields a ~1.1 kb fragment, while the WT Trsp gene yields a 900 bp fragment. The Cre-excised Trsp gene yields a ~450 bp fragment. Confirmatory genotyping was done by Transnetyx using their proprietary probes after we confirmed these probes yielded the same results as our in-house genotyping.

Foxp3^scurfy (JAX stock #004088)(67) mice have been described previously. All mice except Rag1⁻/⁻ mice were housed in the same room. Rag1^−/− mice(68) (JAX stock #002216) were housed in sterilized cages and provided autoclaved food and water. All experiments were performed using 6–8-week-old sex- and age-matched, co-housed mice unless otherwise indicated. Both female and male mice were used. Littermates were used whenever possible. Experiments were approved by the VA Institutional Animal Care and Use Committee (IACUC) and/or the VUMC IACUC.

Flow Cytometry

Staining

For cell surface staining, cells were blocked using mouse Fcblock (Biolegend, San Diego, CA, #101320) and incubated in the antibody cocktail including a fixable live/dead stain (ThermoFisher Scientific, Waltham, MA, #65–0866-14) for 20 minutes at 4°C in the dark. Red blood cells were lysed prior to staining where applicable using ACK lysing buffer (ThermoFisher Scientific, #A1049201). Intracellular cytokine staining was performed using Cytofix/Cytoperm (BD, #554714) and intranuclear stain was performed using the Foxp3/Transcription Factor Staining Buffer Set (ThermoFisher Scientific, #00–5523-00), the Foxp3/Transcription Factor staining buffer kit (Tonbo/Cytek, #TNB-0607-KIT), or the TrueNuclear^™ Transcription Factor Buffer Set (#424401, Biolegend), according to manufacturer’s instructions. Cells were stained for oxidative stress using CellROX Deep Red (ThermoFisher Scientific, #C10422) for 30 minutes at 37°C with 5% CO₂ in 200 μL T cell media (TCM, RPMI 1640 with 10% heat-inactivated FBS, 100 unitsmL⁻¹ penicillin, 100 μgmL⁻¹ streptomycin (ThermoFisher Scientific, #10378016), GlutaMax (ThermoFisher Scientific, #35050–061), MEM non-essential amino acids (Corning, #25–025-CI), 55 μM beta-mercaptoethanol (ThermoFisher Scientific, #21985–023), and 1 mM sodium pyruvate (Millipore Sigma, #S8636–100ML). Stock vials of CellROX were used a maximum of two times as our observations showed similar results (not shown). Cells were stained for Annexin-V for 15 minutes at room temperature in either 1X Annexin-V binding buffer for Annexin-V antibody staining or PBS for Apo-15 peptide staining (Biolegend). Antibodies used for flow cytometry or cell sorting are listed in Table S1.

IL-2 stimulation and pSTAT5 staining

To measure T cell activation in response to IL-2, CD4⁺ T cells were enriched from spleen and mLN using immunomagnetic negative selection (EasySep^™ Mouse CD4+ T Cell Isolation Kit, STEMCELL Technologies, #19852). T cells were stained for surface markers (Table S1) as above, washed with TCM, incubated for 5 minutes at 37°C, 5% CO₂, then stimulated with 20 ngmL⁻¹ recombinant mouse IL-2 (R&D Systems, #202-IL-010/CF) for 15 minutes at 37° C, 5% CO₂. After stimulation, cells were fixed with warm Phosflow^™ Fix Buffer I (BD, # 557870) for 10 minutes at 37°C. After washing, cells were permeabilized with ice-cold (−20°C) Phosflow^™ Perm Buffer III (BD, # 558050) for 30 minutes on ice. Cells were washed again and stained for pSTAT5 for 1 hour at room temperature, then washed and acquired.

Acquisition and sorting

Flow cytometric analysis was performed using a 4/5-Laser Fortessa or 5-laser LSRII (BD, Franklin Lakes, NJ) with FACSDiva software (BD), or a Cytek^® Aurora (CytekBio, Bethesda, MD). FC-sorting was performed on a Cytek Aurora^™ CS System or on a FACS Aria III (BD).

Adoptive transfer

For adoptive transfer colitis, naïve T cells were sorted from spleen and mLN based on live CD4⁺CD45RB^hi expression following pre-enrichment for CD4⁺ cells as above. Rag1^−/− mice were intraperitoneally injected with 5×10⁵ naïve T cells.(34) As colitis in the adoptive transfer model is dependent on a complex microbiota containing Helicobacter spp., recipients were removed from sterile housing conditions and bedding from other (non-experimental) mice housed in non-autoclaved cages was mixed into the cages of recipient mice at the time of injection.(69) For transfer of Treg cells, Treg cells were sorted from spleen and mLN based on live, CD4⁺FOXP3⁺ROSA^tdTomato+ expression following pre-enrichment for CD4⁺ cells using negative selection as above. Rag1^−/− mice were intraperitoneally injected with 1×10⁵ Treg cells. For both experiments, mice that received Trsp-deficient or -sufficient Tnaïve/Treg cells were co-housed. Mice were weighed at the beginning of each experiment to establish a baseline weight, then weekly thereafter.

In vitro Treg cell suppression

Naïve T cells (Live CD4⁺CD45RB^hi) from spleen and mLN were sorted and labeled with CellTrace Violet (ThermoFisher Scientific, #C34557). Treg cells (live CD4⁺CD25⁺ in CD2^Cre mice and live CD4+FOXP3+ROSA^tdTomato+ in Foxp3^{eGFP-Cre-ERT2;} ROSA26^{lox-stop-lox-tdTomato} mice) were sorted. Cells were stimulated using αCD3αCD28 Dynabeads (ThermoFisher Scientific, #11456D) for approximately 60 hours. Proliferation was assessed via CellTrace Violet dilution in naïve T cells and % suppression determined.

Histology

Hematoxylin and eosin (H&E) staining was performed by the VUMC Translational Pathology Shared Resource on formalin-fixed, paraffin-embedded (FFPE) sections. Histological scoring was performed by a gastrointestinal pathologist (MKW) blinded to genotype and treatment.

Slide imaging

H&E slides were imaged by the VUMC Digital Histology Shared Resource using a high-throughput Leica SCN400 Slide Scanner automated digital image system from Leica Microsystems. Whole slides were imaged at 40X magnification to a resolution of 0.25 μm/pixel. Images were exported from QuPath(70) (version 0.4.3, developed at University of Edinburgh, Edinburgh, Scotland) and if necessary, adjusted for brightness and contrast using Adobe Photoshop CS6 (equally across all images in one figure panel).

Mixed bone marrow chimeras

Rag1^−/− recipient mice were pre-conditioned with 450 rads of 137^Cs source radiation. Conditioned mice were retro-orbitally injected with 5×10⁵ congenically-marked lineage-depleted (ThermoFisher Scientific, #8804–6829-74) bone marrow cells recovered from the tibiae and femora of CD45.1⁺;Foxp3^YFP-Cre and CD45.2⁺;Trsp^fl/fl;Foxp3^YFP-Cre mice in a 1:1 ratio. Similarly, control mice received cells from CD45.1⁺;Foxp3^YFP-Cre and CD45.2⁺;Foxp3^YFP-Cre mice. All recipient mice received 0.4 mgmL⁻¹ enrofloxacin (Baytril, Elanco) in their drinking water from 7 days pre-to 7 days after transplant (14 total days), and 1 mgL⁻¹ neomycin from day 3 to day 10 after transplant. Mice were euthanized 6 weeks post-transplant for evaluation.

Dextran sodium sulfate colitis

To induce colitis, mice received 3.5% DSS (TdB consultancy, Uppsala, Sweden) in their drinking water for 5 days, followed by 5 days of regular autoclaved drinking water, after which mice were euthanized. Mice were weighed at the beginning of each experiment to establish a baseline weight, then daily thereafter.

ELISAs

IgE ELISA

Serum IgE levels were measured with a Mouse IgE ELISA Kit (ThermoFisher Scientific, #EMIGHE) according to the manufacturer’s protocol, with the following modifications. For Foxp3^YFP-Cre;Trsp samples, the precoated anti-mIgE 96-well plate included in the kit was used. For Foxp3^scurfy samples, a Nunc MaxiSorp 96-well plate (ThermoFisher Scientific, #445101) was coated with 1 μgmL⁻¹ anti-mIgE (BioLegend, #406902) overnight at 4° C. The following day, the plate was washed with PBS five times, then blocked with 75% (v/v) PBS with 0.1% Tween 20 (Millipore Sigma, #P9416), 25% (v/v) Block ACE (Bio-Rad, #BUF029), 5% (w/v) α-lactose monohydrate (Millipore Sigma, #L3625), and 5% (w/v) D-sucrose (ThermoFisher Scientific, #BP220–212) for 1 hour at room temperature. All serum samples were diluted 1:10000 in the 1X Assay Diluent included in the kit prior to ELISA. Absorbance at 450 nm was read on a GloMax^® Discover plate reader (Promega, Madison, WI).

IL-2 ELISA

Samples were diluted (1:15–1:50) and used in the ELISA MAX^™ Deluxe Set Mouse IL-2 (Biolegend, #431004) according to manufacturer’s instructions. Absorbance at 450 nm was read on a GloMax^® Discover plate reader (Promega).

Phosphoprotein assay

CD4⁺ T cells were enriched from the spleen and mLN of tamoxifen-treated mice using negative magnetic selection. Cells were stimulated using 2μgmL⁻¹ plate-bound αCD3 (ThermoFisher Scientific, #16–0031-86) and 1 μgmL⁻¹ soluble αCD28 (ThermoFisher Scientific, #16–0281-85) for 2 hours. The T-Cell Receptor Phospho Antibody Array (Full Moon Biosystems, #PTC188) was then performed with these cells according to the manufacturer’s protocol, with 10 μLmL⁻¹ each of phosphatase inhibitor cocktail 2 (Millipore Sigma, #P5726), phosphatase inhibitor cocktail 3 (Millipore Sigma, #P0044), and protease inhibitor cocktail (Millipore Sigma, #P8340) added to the provided Extraction Buffer for cell lysis. Cy3-streptavidin (ThermoFisher Scientific, #SA1010) was used as the detection antibody. The stained arrays were then imaged, quantified, and analyzed by Full Moon Biosystems. Resulting data displayed is the ratio analysis (ratio change, see description that follows) executed and described by the manufacturer: “For each spot on the array, Median Signal Intensity is extracted from array image. Average Signal Intensity of Replicate Spots is the mean value of Median Signal Intensity of the replicate spots for each antibody. Using the Average Signal Intensity of Replicate Spots, for each pair of site-specific antibody and phospho site-specific antibody, Signal Ratio of the paired antibodies is determined. Ratio = (Signal Intensity of Phospho Site-Specific Antibody) / (Signal Intensity of Site-Specific Antibody). Ratio change between samples = Treatment Sample / Control Sample.”

2-HOBA treatment

Foxp3^YFP-Cre;Trsp mice were provided 1.33 gL⁻¹ 2-hydroxybenzylamine acetate (TSI Group Co., Ltd) in the drinking water ad libitum from birth.

Protein extraction

After anti-CD3/28 stimulation, CD4⁺ T cells were collected and centrifuged at 300 g for 5 minutes at 4° C. Cells were resuspended in RIPA Lysis and Extraction Buffer (ThermoFisher Scientific, #89900) with 10 μLmL⁻¹ each of phosphatase inhibitor cocktail 2 (Millipore Sigma, #P5726), phosphatase inhibitor cocktail 3 (Millipore Sigma, #P0044), and protease inhibitor cocktail (Millipore Sigma, #P8340), incubated on ice for 10 minutes, and centrifuged at 16000 g for 10 minutes at 4° C. Supernatant protein concentrations were quantified with a BCA Protein Assay Kit (ThermoFisher Scientific, #23225) per manufacturer’s instructions.

WES^™ Simple Western Analysis

Cell lysates at 0.5 μgμl⁻¹ were denatured at 95°C for 5 minutes and target proteins detected using an automated capillary-based immunodetection system (WES^™ SimpleWestern, Biotechne, Minneapolis, MN) according to the manufacturer’s protocol (ProteinSimple, San Jose, CA). Primary antibodies against GPX4 (abcam, #ab125066), TXNRD1 (abcam, #ab124954) and β-actin (Millipore Sigma, #A5441) were diluted 1:25 in kit-supplied dilution buffer. Wes^™ reagents (biotinylated molecular weight marker, streptavidin–HRP fluorescent standards, luminol-S, hydrogen peroxide, sample buffer, DTT, running buffer, wash buffer, matrix removal buffer, secondary antibodies, antibody diluent, and capillaries) were obtained from the manufacturer and used according to the manufacturer’s instructions in supplied partially pre-filled microplates. “Virtual blots” electrophoretic images were automatically generated by the Compass Software and data analysis was performed using the same software (ProteinSimple).

TCR activation Western blot

Lysates were diluted in 4X Laemmli Sample Buffer (Bio-Rad, #1610747) with 6% (v/v) 2-mercaptoethanol (Millipore Sigma, #M3148), then incubated at 95° C for 5 minutes. 30 μg total protein was loaded into each lane of a 10-well 4–20% Mini-PROTEAN TGX Precast Protein Gel (Bio-Rad, #4561094), alongside Precision Plus Protein Dual Color Standards (Bio-Rad, #1610374) for SDS-PAGE. Separated proteins were then transferred to a 0.45 μm nitrocellulose membrane (ThermoFisher Scientific, #LC2001), blocked with Intercept (TBS) Blocking Buffer (LI-COR, #927–60001) at room temperature for 30 minutes, then probed with primary antibodies diluted in 50% Intercept (TBS) Blocking Buffer/50% TBS with 0.1% (v/v) Tween-20 (Millipore Sigma, P9416) (TBS-T) at 4°C overnight. Primary antibodies used included: rabbit anti-LCK (1:1000, Cell Signaling Technology, #2984), rabbit anti-pLCK (pTyr394) (1:1000, abcam, #ab318960), mouse anti-MEK1 (1:2000, Cell Signaling Technology, #2352), rabbit anti-pMEK1 (pSer298) (1:1000, Cell Signaling Technology, #9128), and as loading control rabbit anti-vinculin (1:1000, Cell Signaling Technology, #13901). Membranes were washed with TBS-T, then probed with IRDye 680LT goat anti-mouse IgG (1:10,000, LI-COR, #92668020) and IRDye 800CW goat anti-rabbit IgG (1:10,000, LI-COR, #92632211) secondary antibodies diluted in TBS-T at room temperature for 30 minutes. Membranes were washed again with TBS-T, imaged with an Odyssey Clx near-infrared fluorescence imaging system (LI-COR), and quantified with Image Studio (LI-COR). Densitometric values for proteins of interest were normalized to those of their corresponding loading controls.

Treg cell survival assay

Treg cells (Live CD4⁺FOXP3⁺ROSA^tdTomato+) were sorted from the spleen and mLN of tamoxifen-treated mice. 4×10⁴ cells were plated in each well of a round-bottom 96-well plate with human IL-2 (Biolegend or R&D Systems) at concentrations indicated in the figure legend. After 48 hours, CellTiter-Glo^® Luminescent Cell Viability Assay (Promega, #G7572) was performed according to manufacturer’s instruction, and results normalized to TCM only negative control.

Human thymus single cell RNA-seq analysis

A public single-cell RNA-seq dataset of human thymocytes(31) was manually transcribed from the web-based interface to Prism (Version 10.1.1).

TCRβ chain immunosequencing

Samples were enriched for CD4⁺ T cells using negative selection as above. Cells were then stained and flow-sorted into PBS with 2% FBS and 0.025 M HEPES followed by DNA extraction using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany, #69504). TCR beta sequencing was performed by Adaptive Biotech using the immunoSEQ Assay. Raw data was analyzed using the immunoSEQ Analyzer 3.0 website (clients.adaptivebiotech.com/login).

RNA isolation, cDNA synthesis, and RT-qPCR

RNA extraction from cells was performed using the RNeasy Mini Kit (Qiagen, #4304437). For low cell numbers the RNAeasy Micro Kit (Qiagen, #74004) was used, and cells were sorted directly into 300 μL RLT buffer (Qiagen, #79216) containing 2-mercapto-ethanol as recommended. On column DNAse treatment was performed as per manufacturer’s recommendations. For cDNA synthesis, 5 μL RNA, 4 μL SuperScript IV VILO Master Mix (ThermoFisher Scientific, #11756050), and 11 μL water were combined in a PCR tube and incubated as follows: 1) at 25° C for 10 minutes, 2) at 50° C for 10 minutes, and 3) at 85° C for 5 minutes. Selenoprotein RT-qPCR was performed in technical duplicate using TaqMan^™ probes described previously by our group(30) and TaqMan^™ Universal PCR Master Mix (ThermoFisher Scientific, #4304437) on a QuantStudio3 (ThermoFisher Scientific). Selenoprotein RT-qPCR results were analyzed by the delta-delta Ct method and normalized to Tbp. For Lef1 RT-qPCR, SYBR primers (Millipore Sigma) (5’ TGTTTATCCCATCACGGGTGG 3’ and 5’ CATGGAAGTGTCGCCTGACAG 3’) and PerfeCTa^® SYBR^® Green SuperMix ROX (Quantabio, #9505502K) were used. Lef1 RT-qPCR results were normalized to Ubc (5’ GCCCAGTGTTACCACCAAGA 3’ and 5’ CCCATCACACCCAAGAACA 3’).

RNA isolation for sequencing

Quality control for RNA was performed by the Vanderbilt Technologies for Advanced Genomics core using RNA 6000 Pico (Agilent, Santa Clara, CA), and cDNA library preparation using a NEB library preparation kit. Paired end 150bp sequencing was performed on a NovaSeq 6000 (Illumina).

RNA-sequencing analysis (bulk)

Briefly, samples were trimmed with fastp (version 0.20.0) using default parameters. Quantification was performed using Salmon (version 1.4.0) against a decoy transcriptome (Mus musculus Gencode version 21). Further analysis was performed in R (version 4.1.2) in R studio (version 2021.09.2 + 382). For differential expression analysis, limma (version 3.50.3) was used on log-CPM transformed counts with prior count set to 3 or DESEq2 (version 1.34.0) was used on non-normalized counts.

T cell polarization

T cells were acquired from spleen and mLN of WT C57/Bl6 mice using negative enrichment. T cells were then activated using 2μgmL⁻¹ plate-bound αCD3 (ThermoFisher Scientific, #16–0031-86) and 1 μgmL⁻¹ soluble αCD28 (ThermoFisher Scientific, #16–0281-85) and the following cytokines: Th0: 20 ngmL⁻¹ IL-2 (Biolegend, #575406). Th1: 20 ngmL⁻¹ IL-2, 10 ngmL⁻¹ IL-12 (R&D Systems, #419-ML-010/CF), and 2 μgmL⁻¹ αIL-4 (ThermoFisher Scientific, #16–7041-85). Th2: 20 ngmL⁻¹ IL-2, 10 ngmL⁻¹ IL-4 (R&D Systems, #404-ML-010/CF), and 1 μgmL⁻¹ αIFNγ (ThermoFisher Scientific, #14–7311-85). Treg: 40 ngmL⁻¹ IL-2, 1 μgmL⁻¹ αIFNγ, 1 μgmL⁻¹ αIL-4, and 2 ngmL⁻¹ human TGFβ (R&D Systems, #240-B-002/CF). Th17: 20 ngmL⁻¹ IL-6 (R&D Systems, #406-ML-005/CF), 1 ngmL⁻¹ human TGFβ, 1 μgmL⁻¹ αIL-12 (BD, #554475), 1 μgmL⁻¹ αIFNγ, and 1 μgmL⁻¹ αIL-4. Th0, Th1, Th2, and Treg cells were cultured in RPMI 1640, while Th17 cells were cultured in IMDM (ThermoFisher Scientific, #12440079) base media. All T cell media contained 10% (v/v) heat-inactivated FBS, 100 unitsmL⁻¹ penicillin, 100 μgmL⁻¹ streptomycin, 1% (v/v) GlutaMax, 1% (v/v) MEM non-essential amino acids, 55 μM beta-mercaptoethanol, and 1 mM sodium pyruvate.

Quantification and statistical analysis

Age-, sex- and, when feasible, littermate-matched mice were used for all experiments. Statistical analysis was performed using GraphPad Prism (9.1.2) unless otherwise described. Sample size was determined empirically. All data points are biological replicates, i.e. each datapoint represents one mouse, unless otherwise indicated, i.e. data from multiple mice were pooled. All figures represent similar results from at least three independent experiments unless otherwise indicated.

Abbreviations

DSS

dextran sodium sulphate

DN

double negative (CD4^neg, CD8^neg)

DP

double positive (CD4⁺, CD8⁺)

FC

flow cytometry

mLN

mesenteric lymph node

RT-qPCR

reverse transcription quantitative polymerase chain reaction

ROS

reactive oxygen species

SI

small intestine

SP

single positive

TCR

T cell receptor

Th

T helper

WT

wild-type

Data and materials availability

Bulk RNA-sequencing data generated as part of this study is deposited in GEO: GSE275428 reviewer password: gpivwoaupzspnyv. Code will be deposited at GitHub prior to publication. Other materials are available upon request from the corresponding author.