Schistosome infection induces naive CD4+ T cell aging in mice
To dissect the effect of severe infection on host naive CD4+ T cells, we employed schistosome infection, a helminthiasis affecting approximately 200 million people19. Naive CD4+ T cells were isolated by magnetic-activated cell sorting (MACS) from the spleen of uninfected mice or schistosome-infected mice and were performed RNA-sequencing (RNA-seq) experiments. RNA-seq analysis revealed that severe schistosome infection upregulated 3,450 genes in naive CD4+ T cells (fold change ≥ 2, q ≤ 0 .05), while it repressed 2,567 genes (Fig. 1a). Notably, naive CD4+ T cells from schistosome-infected mice displayed decreased expression of the double-strand-break repair nuclease MRE11A (encoded by Mre11a) and increased expression of the DNA damage marker γH2aX (encoded by H2ax) (Fig. 1b), suggesting increased DNA damage in these cells8. Naive CD4+ T cells from schistosome-infected mice upregulated the expression of the senescence markers cyclin-dependent kinase inhibitor 1 (CDKN1A/p21, encoded by Cdkn1a) and cyclin-dependent kinase inhibitor 2 A (CDKN2A/P16, encoded by Cdkn2a), downregulated the expression of Cyclin D (encoded by Ccnd1) and Cyclin E (encoded by Ccnd2), and lost the expression of the costimulatory molecules CD27 (encoded by Cd27) and CD28 (encoded by Cd28) (Fig. 1b), supporting that severe schistosome infection induces a senescent-like phenotype in naive CD4+ T cells8. Strikingly, the expression of the mitochondrial transcription factor A (TFAM, encoded by Tfam) was reduced in naive CD4+ T cells from schistosome-infected mice (Fig. 1b), which may cause an immunometabolic dysfunction, and as a consequence, drive T cell aging20. Consistently, gene set enrichment analysis (GSEA) of differentially expressed genes showed significant enrichment of the glycolysis pathway-associated gene signature in naive CD4+ T cells from schistosome-infected mice (Fig. 1c), accompanied by an enhanced senescence-associated secretory phenotype (SASP), as shown by increased expressions of Ccl1, Ccl2, Ccl3, Il6, Tnf, Csf1, Gzmb, and Gzmk (Fig. 1d, e). Meantime, the expression of Foxo1 was substantially reduced in naive CD4+ T cells from schistosome-infected mice (Fig. 1f, g), suggesting a T cell aging-associated loss of proteostasis8. More importantly, the expression levels of CD5 on naive T cells in the spleen and blood, which have been widely used to reflect the strength and/or duration of TCR signaling in naive CD4+ T cells and CD8+ T cells in the thymus and periphery21,22,23,24, were dramatically decreased in schistosome-infected mice (Fig. 1h–j and Fig. S1), indicating that these naive T cells may poorly respond to antigens, such as schistosome specific antigens and bystander antigens. Supportively, results showed that peripheral naive CD4+ T cells from schistosome-infected mice displayed impaired proliferation when polyclonally activated with an anti-CD3 mAb (Fig. 1k, l), which demonstrates that severe schistosome infection induces naive CD4+ T cell aging in mice.
Naive T cell aging is linked to thymic involution during severe schistosome infection
Given that thymic involution is a primary hallmark of T cell aging8 and has been observed in schistosome-infected mice17, we next investigated whether thymic involution contributed to naive T cell aging-associated dysfunction during severe infection. Results showed that the size and weight of the thymus (Fig. 2a, b) and the number of thymocytes (Fig. 2c) were remarkably reduced in mice starting at week 5–6 after infection, accompanied by an aberrant development of T cells (Fig. 2d), reduced expression of CD5 (Fig. S2a–c), and decreased proliferation of naive T cells (Fig. 2e, f). In addition, the apoptosis of thymocytes including CD4+CD8+ double-positive (DP), CD4+ single-positive (CD4SP), CD8+ single-positive (CD8SP), and CD4–CD8– double-negative (DN) thymocytes, was strongly increased in schistosome-infected mice (Fig. S2d, e). In contrast, thymic involution was abrogated in mice 7 weeks after antischistosomal treatment, as shown by the fact that antischistosome-treated mice displayed normal thymic size, weight, cellularity, and thymocyte subsets (Fig. 2g–j). In parallel with abolished thymic involution, the expression of CD5 on naive T cells was restored in infected mice 7 weeks after antischistosomal treatment almost up to normal levels (Fig. S2f–h), which was concomitant with increased naive T cell proliferation (Fig. 2k, l). Taken together, these results suggest that T cell aging-associated dysfunction is linked to thymic involution during severe schistosome infection.
IL-33 results in thymic involution during severe infection
Substantial evidence shows that the alarmin IL-33 was strongly produced by inflammation-damaged tissues in severe infectious diseases such as schistosomiasis and sepsis25,26,27. As expected, our results showed that the level of full-length or short-form of IL-33 was markedly elevated in the liver in schistosome-infected mice (Fig. S3a), where schistosome eggs provoked a vigorous granulomatous inflammatory response. In contrast, we observed that the level of the full-length of IL-33 but not the short-form of IL-33 was substantially increased in the thymus of mice with schistosomiasis (Fig. 3a), which may be due to a low level of cleaved IL-33 or lack of cleaving proteases of IL-33 in the thymus during schistosome infection28. Next, we wondered whether the elevation of IL-33 was involved in thymic involution during severe schistosome infection. Concomitant with reversed thymic phenotype and function in mice 7 weeks after antischistosomal treatment shown in the aforementioned data (Fig. 2), antischistosomal treatment reduced accumulation of IL-33 in the thymus from schistosome-infected mice (Fig. 3a). More importantly, IL-33 deficient mice displayed normal thymic morphology and cellularity (Fig. 3b–d) and thymocyte development (Fig. 3e) after schistosome infection, which was coincident with decreased thymocyte apoptosis (Fig. S3b). IL-33 neutralization (Fig. 3f–i) or ST2 deficiency (Fig. S3c–e) phenocopied IL-33 deficiency, indicating a major role of extracellular IL-33, rather than nuclear IL-33, in thymic involution through the membrane-bound receptor ST2 during schistosome infection. Interestingly, the serum level of sST2 was significantly increased in mice after schistosome infection (Fig. S3f), indicating a potential regulatory role of sST2 in IL-33-mediated thymic involution. In addition, sepsis, a severe systemic infectious disease, also resulted in increased IL-33 level in the thymus (Fig. S3g), and as a consequence, caused thymic involution in mice and humans (Fig. S3h–k), while IL-33 deficient mice displayed the normal morphology, weight, and cellularity of the thymus during sepsis (Fig. S3h–j).
To next determine if IL-33 alone was sufficient to induce thymic involution, we injected recombinant IL-33 or vehicle (PBS) intraperitoneally into normal WT mice. Results showed that administration of IL-33 led to distinct reductions in the size, weight, and cellularity of the thymus (Fig. 3j–l) and abnormal thymocyte development (Fig. 3m), which was coincident with increased thymocyte apoptosis (Fig. S3l), while ST2 deficiency in mice abolished the effect of IL-33 administration on thymic involution (Fig. S3m–o).
To further determine whether IL-33-mediated thymic involution was thymus intrinsic in vivo, we established thymus transplantation under the kidney capsule (Fig. S3p–r). Results showed that the IL-33 receptor (ST2)-sufficient recipient or donor thymus had undergone profound involution, while the ST2-deficient donor thymus was normal in the morphology, weight, and cellularity in IL-33-treated or schistosome-infected recipient mice (Fig. 3n–q). Taken together, these results demonstrate that IL-33 leads to thymic involution in a thymus-intrinsic manner during severe infection.
IL-33-mediated thymic involution induces naive T cell aging and impairs host control of infection
Next, we explored the consequences of IL-33-induced thymic involution on T cell immunity during severe infection. Concomitant with restored thymic phenotype and function, the TCR avidity of naive T cells was markedly higher in Il-33−/− mice with schistosomiasis or sepsis than that in wildtype mice, as shown by a reduced decline in expression levels of CD5 on naive T cells in thymus, spleen, and blood in Il-33-/- mice (Fig. S4a–c) or anti-IL-33-treated mice (Fig. S4d–f) with schistosomiasis, or Il-33-/- mice with sepsis (Fig. S4g–i). Accordingly, naive CD4+ T cells from Il-33-/- or anti-IL-33 antibody-treated infected mice exhibited an increased proliferation than those from wildtype infected mice (Fig. 4a–d). In contrast, administration of IL-33 resulted in naive-memory T cell imbalance (Fig. S5a–d), a decrease in expression levels of CD5 on naive T cells in the thymus, spleen, and blood (Fig. S5e–g), and attenuated responsiveness of peripheral naive CD4+ T cells to the anti-CD3 mAb (Fig. 4e, f).
To further investigate whether IL-33 impaired T cell immunity by inducing thymic involution, ST2 sufficient or deficient thymus was transplanted into wildtype mice treated with IL-33 or infected with schistosome. The results showed that the transplantation of IL-33 receptor (ST2)-deficient thymus strongly restored T-cell proliferative activity from mice treated with IL-33 or infected with schistosome (Fig. 4h, i, k, and l), concomitant with increased expression levels of CD5 on naive T cells in the thymus, spleen, and blood (Fig. 4g, 4j, and Fig. S5h–k). More importantly, the transplantation of ST2-deficient thymus, but not ST2-sufficient thymus, significantly promoted T cell-mediated egg granuloma formation to protect host tissues from egg-derived toxins in recipient mice infected with schistosome (Fig. 4m, n)29, as well as lengthened the survival of recipient mice with sepsis (Fig. 4o). Taken together, these results demonstrate that IL-33-mediated thymic involution impairs host T-cell immune response against pathogens during severe infection, which further suggests that IL-33 and/or its receptor ST2 could be a promising target for intervention to reverse thymic involution and restore T-cell immunity.
IL-33 perturbs the compartment of thymic epithelial cells both in vitro and in mice with IL-33 administration or severe infection
Given that the abnormality in function or compartment of TECs is a leading cause of thymic involution30, we next wondered whether IL-33 had an impact on TEC function or compartment. To dissect the effect of IL-33 on TECs in vitro, we isolated TECs by MACS from thymus cultured in fetal thymic organ cultures (FTOC) treated with IL-33 or PBS (Fig. S6a), and performed RNA-sequencing (RNA-seq) experiments. RNA-seq analysis revealed that IL-33 upregulated 457 genes in TECs (fold change ≥ 2, q ≤ 0.05), while it repressed 1,702 genes (Fig. 5a). Strikingly, IL-33 altered the expression of TEC function-related genes (Fig. 5b), such as Foxn1, Cd83, and prdm1 (required for mTEC or cTEC functions), Aire and Fezf2 (required for tissue-restricted antigen generation), Ccl19, Ccl25, Xcl1, Ccl22, and other chemokines (required for the trafficking of thymic progenitors, thymocytes, or thymic DCs), Ctsh (required for negative selection of T cells), and Tmsb15b2 and Tmsb15a (encoding thymosin β15a and thymosin β15b2, respectively; important regulators of the peripheral immune response), which suggested that IL-33 altered the functions of TECs.
Because of the central role of mTEC/cTEC compartments in thymic involution30, we next determined whether IL-33 disturbed mTEC/cTEC compartments during IL-33-mediated thymic involution. As expected, the mTEC-specific genes, such as Ascl1, Ehf, Ccl19, Pigr, Trpm5, Avil, Gnat3, Plcb2, Chat, Alox5, Ltc4s, Pde2a, Gnb3, and Lrmp, were upregulated in IL-33-treated TECs, while cTEC-specific genes, such as Enpep (encoding LY51), Ccl25, Ackr4 (encoding Ccrl1), Cd83, Foxn1, and Tbeta, were downregulated (Fig. 5c). Of note, Enpep (encoding LY51), a specific marker of cTECs31, was dramatically decreased in IL-33-treated TECs (Fig. 5d), which indicated that IL-33 disrupted mTEC/cTEC compartments. Consistent with RNA-seq analysis, flow cytometry analysis showed an increase in mTEC population, but a decrease in cTEC population in the thymus treated with IL-33 in vitro in FTOC (Figs. 5e–g), while ST2 deficiency abolished IL-33-mediated mTEC/cTEC imbalance (Fig. S6b–d). Furthermore, IL-33 administration also perturbed mTEC/cTEC compartments in vivo (Fig. 5h–k), while ST2 deficiency abolished IL-33-mediated mTEC/cTEC imbalance (Fig. S6e–h). More importantly, schistosome infection or sepsis also resulted in abnormal mTEC/cTEC compartments, as shown by an increased proportion of mTECs but a decreased proportion of cTECs, while IL-33 or ST2 deficiency, or anti-IL-33 treatment restored mTEC/cTEC compartments in mice with schistosomiasis (Fig. 5l–r, S6i–l) or sepsis (Fig. S6m–o). Taken together, these results show that IL-33 disrupts the compartment of thymic epithelial cells, which is a leading cause of thymic involution.
IL-33 disturbs the TEC compartment by inducing excessive mTEC differentiation in a thymocyte-independent manner
To study how IL-33 played a pathogenic role in disturbing the TEC compartment, we examined the expression of ST2 on cells in the thymus. Flow cytometry analysis showed that a higher percentage of TECs expressed ST2 compared with thymocytes, non-TEC stromal cells, or Treg cells (Fig. 6a, b, and Fig. S7a–e), whereas the MFI of ST2 was comparable on these subsets of cells in the thymus during schistosome infection (Fig. S7f). Indeed, no direct effect of IL-33 on thymocyte apoptosis was observed in vitro (Fig. 6c). To further rule out the possibility that thymocytes were involved in IL-33-induced mTEC accumulation, deoxyguanosine (dGUO) was employed to eliminate thymocytes in the thymus. Of note, IL-33 resulted in an excessive increase in the mTEC population in the absence of thymocytes in the dGUO-treated thymus in FTOCs (Fig. 6d–f). These results, in conjunction with our observation that donor dGUO-treated WT thymus, rather than donor dGUO-treated ST2 deficient thymus, displayed severe involution in thymus-transplanted mice treated with IL-33 or infected with schistosome (Fig. 3n–q), demonstrate that IL-33 results in aberrant accumulation of mTECs and subsequent thymic involution independently of thymocytes.
Next, we further wondered how IL-33 led to an excessive increase in the ratio of mTEC/cTEC. Intriguingly, there was no noticeable difference in IL-33-induced apoptosis or proliferation between mTECs and cTECs (Fig. 6g, h), raising the possibility that IL-33-mediated mTEC accumulation may be through affecting the differentiation of TEC into mTECs. As expected, the level of the noncanonical NF-κB transcription factor p100/p52, which plays a key role in mTEC differentiation32, was elevated in TECs treated with IL-33 (Fig. 6i). Inhibition of NF-κB-inducing kinase (NIK), a central component of the noncanonical NF-κB pathway, completely abolished the IL-33-mediated mTEC accumulation (Fig. 6j–l). Taken together, these results suggest that IL-33 disturbs the TEC compartment by inducing aberrant differentiation of TECs into mTECs to alter the proper ratio of mTEC/cTEC, which is a leading cause of thymic involution.
IL-33-induced generation of mTEC IV contributes to an aberrant increase in the mTEC population and subsequent thymic involution
Given the heterogeneity of mTECs, consisting of four major mTEC subpopulations (mTEC I–IV) with distinct transcriptional profiles and lineage regulators33,34, we next wondered whether IL-33 promoted excessive accumulation of the mTEC population by inducing the aberrant generation of a particular subset of mTECs. RNA-seq analysis showed that in TECs from the thymus in FTOCs treated with IL-33 for 4 days, mTEC IV (thymic tuft cell)-specific genes, such as Trpm5, Pou2f3, Avil, Gnat3, Plcb2, Chat, Alox5, Ltc4s, Pde2a, Gnb3, and Lrmp, were globally upregulated (Fig. 7a). Meanwhile, mTEC II-specific genes including Ctsh, Aire, Fezf2, Klk1b16, Apoc2, Cyp24a1, and Muc3, and mTEC III-specific genes including Fgf21, Krt17, Spink5, Plb1, Cd80, Cd86, Krt1, Krt77, and Asprv1, were globally downregulated (Fig. 7a). In contrast, mTEC-I specific genes, such as Six4, Chuk, Bmp4, Itga6, Itgb4, Krt5, Sox4, and Ly6a, displayed no significant change (<2 fold change) (Fig. 7a).
To further confirm that IL-33 promoted excessive accumulation of TECs by inducing the aberrant generation of mTEC IV cells (thymic tuft cells), we detected the expression level of the transcription factor Pou2f3 in TECs, a master regulator of mTEC IV cells33,34, and found a dramatic increase in Pou2f3 expression in TECs treated with IL-33 (Fig. 7b, c). Consistently, flow cytometric analysis showed that mTEC IV cells were indeed induced preferentially in the thymus treated with IL-33 in FTOCs (Fig. 7d, e). Strikingly, the percentage of mTEC I cells was also increased after the accumulation of mTEC IV cells in the thymus in FTOCs with IL-33 treatment for 7 days, while the percentages of mTEC II cells and mTEC III cells were decreased (Fig. 7e). Consistent with these observations in vitro, the percentages of mTEC IV cells and mTEC I cells were also increased in mice injected with IL-33 or infected with schistosome, while the percentages of mTEC II cells and mTEC III cells were decreased (Fig. 7f, g). In contrast, thymic deletion of pou2f3 resulted in reduced accumulation of mTECs by IL-33 treatment (Fig. 7h–j). Meanwhile, Pou2f3 deficient mice displayed reduced accumulation of mTECs (Fig. 7k–m). More importantly, Pou2f3 deficiency alleviated thymic involution (Fig. 8a–c), restored naive T cell proliferation (Fig. 8d, e), and promoted T cell-mediated granuloma formation (Fig. 8f, g) in mice after schistosome infection. Taken together, these results indicate that IL-33-induced excessive generation of mTEC IV cells (thymic tuft cells) contributes to the disruption of TEC compartment by changing the ratio of mTEC/cTEC and subsequent thymus involution-mediated T cell aging during severe infection.