The co-inhibitory receptor TIGIT promotes tissue-protective functions in T cells

Panetti, Camilla; Daetwyler, Rahel; Moncsek, Anja; Patikas, Nikolaos; Agrafiotis, Andreas; Tang, Adelynn; Andreata, Francesco; Fumagalli, Valeria; De Lima, Jean; Wen, Lifen; King, Carolyn G.; Vasanthakumar, Ajithkumar; Iannacone, Matteo; Kallies, Axel; Yermanos, Alexander; Hemberg, Martin; Joller, Nicole

doi:10.1038/s41590-025-02300-w

Download PDF

Article
Open access
Published: 15 October 2025

The co-inhibitory receptor TIGIT promotes tissue-protective functions in T cells

Nature Immunology volume 26, pages 2074–2085 (2025)Cite this article

9797 Accesses
1 Citations
91 Altmetric
Metrics details

Subjects

Abstract

The co-inhibitory receptor TIGIT suppresses excessive immune responses in autoimmune conditions while also restraining antitumor immunity. In viral infections, TIGIT alone does not affect viral control but has been shown to limit tissue pathology. However, the underlying mechanisms are incompletely understood. Here we found TIGIT⁺ T cells to express not only an immunoregulatory gene signature but also a tissue repair gene signature. Specifically, after viral infection, TIGIT directly drives expression of the tissue growth factor amphiregulin (Areg), which is strongly reduced in the absence of TIGIT. We identified regulatory T (T_reg) cells, but not CD8⁺ T cells, as the critical T cell subset mediating these tissue-protective effects. In T_reg cells, TIGIT engagement after T cell antigen receptor stimulation induces the transcription factor Blimp-1, which then promotes Areg production and tissue repair. Thus, we uncovered a nonclassical function of the co-inhibitory receptor TIGIT, wherein it not only limits immune pathology by suppressing the immune response but also actively fosters tissue regeneration by inducing the tissue growth factor Areg in T cells.

TIGIT expression in extrahepatic cholangiocarcinoma and its impact on CD8 + T cell exhaustion: implications for immunotherapy

Article Open access 12 February 2025

Regulators of CD8⁺ T cell exhaustion

Article 01 October 2025

Anti-human-TIGIT agonistic antibody ameliorates autoimmune diseases by inhibiting Tfh and Tph cells and enhancing Treg cells

Article Open access 09 May 2023

Main

Co-inhibitory receptors are essential for maintaining immune balance, preventing excessive immune activation and tissue damage under normal conditions and during inflammation¹. T cell immunoglobulin and ITIM domain (TIGIT) is one of these co-inhibitory receptors, and it exerts its inhibitory function by directly inhibiting effector T cell activation as well as tolerizing dendritic cells through engagement of its ligand CD155 (refs. ^2,3). Moreover, TIGIT is constitutively expressed on regulatory T (T_reg) cells and enhances their suppressive capacity⁴. Loss of TIGIT results in increased susceptibility to autoimmunity, while promoting tumor clearance⁵. Similar to other co-inhibitory receptors, TIGIT is expressed after T cell activation, resulting in its upregulation during infections and sustained expression on exhausted T cells in chronic infections^6,7. Interestingly, while TIGIT modulates the antiviral response following lymphocytic choriomeningitis virus (LCMV) infection, it does not alter viral control but rather limits the tissue pathology resulting from LCMV and influenza infection⁷. However, the underlying mechanism for this is still poorly understood.

In recent years, it has become increasingly evident that T cells, and T_reg cells in particular, play a crucial role in maintaining tissue homeostasis by actively contributing to tissue regeneration and repair⁸. T_reg cells accumulate at the site of tissue injury, and their depletion results in delayed tissue regeneration, exacerbated tissue damage and fibrosis^9,10,11,12. At sites of tissue damage, T_reg cells produce the growth factor amphiregulin (Areg), which binds to the epidermal growth factor receptor (EGFR) to promote tissue repair across a range of settings, including muscle repair, acute lung injury and remyelination of damaged axons^12,13. In the context of viral infections, T_reg cells have been shown to produce Areg following influenza infection and are thereby protective against infection-induced lung damage. Areg production by T_reg cells was critical for limiting tissue damage, as mice lacking Areg specifically in T_reg cells exhibited increased lung pathology following influenza infection¹¹. T_reg cell-derived Areg is sensed by EGFR⁺ mesenchymal cells, which in turn enable epithelial cell regeneration¹⁴. Similarly, T_reg cell-derived Areg mitigates tissue damage following severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection¹⁵. In these settings, alarmins like interleukin-33 (IL-33) and IL-18 can promote Areg production by T_reg cells, while Notch-4 restrains it, thus limiting tissue repair^11,15. However, whether additional signals regulate Areg production through T_reg cells during infection remains unknown.

As TIGIT can limit tissue damage during viral infection, we hypothesized that it may act as an additional signal promoting the repair functions of T cells. To test this, we conducted comprehensive transcriptional profiling of T cells following LCMV infection and found a transcriptional repair signature to be enriched in Tigit-expressing T cells. Notably, Areg was highly coexpressed with Tigit, and we demonstrate that TIGIT is functionally required for Areg production in T cells. Furthermore, we show that TIGIT expression in T_reg cells, but not in CD8⁺ T cells, is essential for limiting tissue damage following infection. Areg-producing T_reg cells present upon infection are clonally expanded, and both TIGIT and TCR signaling are necessary for Areg induction by in vivo-primed T cells. Finally, we found that Blimp-1 acts downstream of TIGIT to induce Areg production. Thus, the co-inhibitory receptor TIGIT not only limits immune pathology by suppressing excessive immune responses but also promotes tissue regeneration by inducing Areg in T_reg cells.

Results

Loss of TIGIT exacerbates pathology following viral infection

We have previously shown that TIGIT engagement limits immune pathology following viral infection⁷. Interestingly, TIGIT engagement reduced infection-induced pathology independent of viral clearance, leading us to hypothesize that TIGIT may exert a tissue-protective function that is independent of its immune-suppressive properties. We first confirmed that Tigit-knockout (KO) mice indeed show increased pathology, as would be expected⁷. To this end, C57BL/6 (B6) wild-type (WT) and Tigit-KO mice (Extended Data Fig. 1a–c) were infected with LCMV clone 13 (Cl13), which induces a high degree of tissue pathology peaking at day 10 after infection (Extended Data Fig. 1d–f). After infection, Tigit-KO mice showed a significant increase in pathology, as assessed by measuring the serum aminotransferases aspartate aminotransferase (AST) and alanine aminotransferase (ALT) as indicators of liver injury (Fig. 1a). Similarly, kidneys and lungs of Tigit-KO mice also showed exacerbated vascular leakage, as indicated by increased levels of Evans blue dye, compared to WT mice (Fig. 1b). Noticeably, increased tissue damage was observed only following LCMV challenge; hence, lack of TIGIT does not affect tissue integrity at steady state but only following infection. Additionally, we observed increased cellular infiltrates in the livers of Tigit-KO mice compared to WT mice, and caspase-3 staining of liver sections revealed increased apoptosis of hepatocytes in Tigit-KO mice after LCMV Cl13 infection (Fig. 1c and Extended Data Fig. 1g). Despite these clear differences in pathology, only a relatively small increase in proinflammatory cytokines was observed in T cells from Tigit-KO mice (Fig. 1d), the response in natural killer cells, which also express TIGIT, was unaltered (Extended Data Fig. 1h), and broad immunophenotyping analysis of infected spleen and lung revealed no major alteration in the immune cell composition between WT and Tigit-KO mice (Fig. 1e and Extended Data Fig. 1i). These results are in line with our previous data^7,16 and confirm that TIGIT is important for limiting tissue pathology following viral infection and that loss of TIGIT exacerbates tissue damage. Nevertheless, TIGIT has limited impact on the proinflammatory response in this context, suggesting that it may have additional functions contributing to tissue protection.

**Fig. 1: *Tigit*-KO mice show increased tissue pathology following infection.**

TIGIT⁺ T cells display transcriptional repair signature

To investigate how TIGIT limits tissue damage after infection, we performed single-cell RNA sequencing (scRNA-seq) of T cells from the spleen and lung of naive and LCMV Cl13-infected WT and Tigit-KO mice, and we enriched for CD4⁺Foxp3⁺ T_reg cells to obtain a solid representation of all T cell subsets (Extended Data Fig. 2a). We identified 15 T cell clusters (Fig. 2a), which were annotated based on commonly used transcriptional profiles (Extended Data Fig. 2b). As expected, comparison of T cells from naive versus LCMV-infected mice revealed the emergence of proliferating T cell clusters marked by Ki67 expression with infection, as well as that of exhausted CD8⁺ T cells and T helper 1 (T_H1)-like CD4⁺ T cells, marked by high expression of Tox and Pdcd1 (Fig. 2a and Extended Data Fig. 2b–d). Also, clusters rich in Cd44, indicative of T cell activation, Tbx21, highlighting the T_H1-dominated immune response after LCMV infection, and Lag3, an early exhaustion marker, were only present after infection. In line with previous reports^6,7, Tigit was expressed at higher levels in activated T cells and was induced upon infection (Extended Data Fig. 2b,c). Given the increasing recognition of T cells as mediators of tissue repair^11,12 and our findings of increased pathology after TIGIT loss, we sought to determine whether Tigit expression might be linked to tissue repair. Indeed, Tigit showed clear coexpression with a curated list of tissue repair genes, including Areg and Batf, characteristic of tissue repair T_reg cells, as well as genes more broadly associated with tissue repair, such as Klf10 and Tgfb1 (ref. ¹²; Fig. 2b and Extended Data Fig. 2e). Although Tigit expression correlated with this repair signature in all T cell subsets, the specific co-regulated genes varied slightly depending on the type of T cell (Fig. 2c). For instance, Ikzf2 (Helios), which marks thymically derived T_reg cells, is part of the tissue repair transcriptome of T_reg cells and CD4⁺, but not CD8⁺, effector T cells. Interestingly, Areg, encoding Areg, a key mediator of T cell-mediated tissue protection following viral infection¹¹, was highly coexpressed with Tigit in all T cell subsets, particularly in the effector T_reg, exhausted-like CD4⁺ and exhausted CD8⁺ T cell clusters (Fig. 2d). These data suggest a module consisting of Areg and Tigit that might be involved in limiting tissue damage upon infection.

**Fig. 2: TIGIT-expressing T cells are enriched for a transcriptional repair signature.**

To address whether Areg induced in T cells is translated into protein to mediate a tissue-protective function during viral infection, we quantified its protein expression. We observed strong Areg induction in spleen and lung T_reg cells after LCMV Cl13 infection (Fig. 2e,f and Extended Data Fig. 3a). While CD4⁺ and CD8⁺ conventional T (T_conv) cells produced lower amounts, Areg was also induced with infection (Fig. 2f and Extended Data Fig. 3a). Furthermore, the kinetics of Areg expression over the course of infection clearly mirrored AST and ALT dynamics and thus liver pathology (Fig. 2e and Extended Data Fig. 1d), hinting at a role for T_reg cell-derived Areg in the resolution of systemic pathology. Importantly, the coexpression of Areg and Tigit observed at the mRNA level was also seen at the protein level in both spleen and lung, where TIGIT⁺ T_reg cells were the main producers of Areg and TIGIT was upregulated following infection (Fig. 2g and Extended Data Fig. 3b-d). Indeed, frequencies of Areg⁺ cells were highest among TIGIT⁺ T_reg cells compared to their TIGIT⁻ counterparts, and the infection-induced increase in Areg was restricted to this population (Fig. 2h). TIGIT also positively correlated with Areg in all T cell populations (Extended Data Fig. 3e). Of note, baseline Areg production observed in naive mice was not restricted to TIGIT⁺ T cells (Fig. 2h), which fits with our finding that loss of TIGIT does not impair tissue integrity at steady state. Overall, our results reveal that TIGIT expression in T cells correlates with a tissue repair signature that includes Areg. Furthermore, infection-induced Areg expression is restricted to the TIGIT⁺ T_reg cell population, suggesting a functional link between TIGIT, Areg and tissue repair in T cells.

TIGIT regulates Areg expression

To determine whether there is a functional link between TIGIT and Areg, we infected WT mice with LCMV Cl13 and treated them with blocking antibody to TIGIT (1B4) or with an isotype control (IgG1). Indeed, TIGIT blockade led to reduced Areg production compared to the control group (Fig. 3a). This was true for T_reg cells and CD4⁺ and CD8⁺ T_conv cells, even though Areg levels were much lower in the latter two populations. Similarly, comparison of scRNA-seq data from infected WT and Tigit-KO mice revealed reduced Areg expression in all three KO T cell populations (Fig. 3b). Similarly, at the protein level, Areg induction was severely impaired in both the spleen and lung of Tigit-KO mice after both LCMV and influenza infection, but not in steady-state T_reg cells (Fig. 3c and Extended Data Fig. 4a–d), demonstrating a functional link between TIGIT and Areg following infection. Importantly, this reduction was seen not only for Areg but also for the whole repair signature, which showed the highest score in WT CD8⁺ T cells as well as T_reg cells, but was reduced in Tigit-KO cells (Fig. 3d). Together, these findings demonstrate that TIGIT can act as a mediator of tissue repair upstream of Areg and that its expression is essential for infection-induced Areg production.

**Fig. 3: Areg expression is TIGIT dependent.**

TIGIT-dependent repair is mediated through Foxp3⁺ T_reg cells

Having established TIGIT’s functional role in inducing Areg, we next addressed through which T cell subtype TIGIT exerts its protective function. As we observed the highest repair scores in CD8⁺ and CD4⁺Foxp3⁺ T cells, we focused on these two T cells subsets. To assess whether TIGIT’s repair function was mediated through CD8⁺ T cells, we used Cd8^Cre × Tigit^fl/fl mice, in which the E8I Cd8 enhancer region drives expression of the Cre recombinase, resulting in Tigit deletion specifically in peripheral CD8⁺, but not CD4⁺, T cells (Extended Data Fig. 4e). We observed a similar induction of Areg after LCMV infection in Cd8^Cre × Tigit^fl/fl mice and littermate control mice (Fig. 3e), demonstrating that TIGIT expression in CD8⁺ T cells is not required for the upregulation of Areg. Furthermore, pathology, determined by vascular permeability in the kidney, lung, liver and spleen, was not altered in Cd8^Cre × Tigit^fl/fl mice compared to control mice (Fig. 3f and Extended Data Fig. 4f). AST levels, as indicators of liver pathology, were also comparable between the two genotypes (Fig. 3g). To determine whether TIGIT expression on CD8⁺ T cells might indirectly influence tissue repair pathways, we analyzed the expression of key transcription factors and cytokines associated with tissue repair but did not observe any change following Tigit deletion in CD8⁺ T cells (Extended Data Fig. 4g). We further examined whether T_reg cells in Cd8^Cre × Tigit^fl/fl mice displayed phenotypic changes that might compensate for a potential loss of repair function of the Tigit-KO CD8⁺ T cell compartment but observed no such alterations (Extended Data Fig. 4h). Thus, we concluded that in the context of systemic viral infection, TIGIT expression on CD8⁺ T cells does not equip them with a specific repair function. Thus, despite the high transcriptional repair score observed in CD8⁺ T cells, TIGIT expression in this subset does not contribute, directly or indirectly, to tissue protection during viral infection.

Next, we tested whether TIGIT expression in T_reg cells was required to limit tissue pathology following LCMV infection. To this end, we generated mice that allow for tamoxifen-induced deletion of Tigit specifically in T_reg cells (Foxp3^GFPcreERT2 × Tigit^fl/fl; Extended Data Fig. 4i,j). Indeed, T_reg cells from Foxp3^GFPcreERT2 × Tigit^fl/fl mice treated with tamoxifen (but not control mice) showed strong impairment in Areg induction following LCMV infection (Fig. 3h), while baseline Areg expression at steady state remained unaffected by Tigit deletion (Extended Data Fig. 4k). To control for the differences in the antiviral immune response elicited in WT mice and mice that lack TIGIT on T_reg cells, we used heterozygous Foxp3^GFPcreERT2 × Tigit^fl/fl females, which harbor both WT and Tigit-KO T_reg cells that can be distinguished based on green fluorescent protein (GFP) expression. Although the WT T_reg cell fraction (Foxp3–GFP⁻) efficiently produced Areg, TIGIT-deficient T_reg cells (Foxp3–GFP⁺) failed to do so (Fig. 3i and Extended Data Fig. 4l). Along with the reduction in T_reg cell-derived Areg in tamoxifen-treated Foxp3^GFPcreERT2 × Tigit^fl/fl mice, vascular leakage in the lung and liver was increased compared to in vehicle-treated control mice (Fig. 3j and Extended Data Fig. 4m). Similarly, AST levels were significantly higher when Tigit was deleted from T_reg cells (Fig. 3k). Thus, in T_reg cells, TIGIT directly induces Areg in a cell-intrinsic manner following infection, and TIGIT expression on T_reg cells is essential for TIGIT-mediated protection from pathology following infection.

TIGIT-mediated Areg induction is regulated by CD155 but not CD226

The TIGIT ligand CD155 is shared with its co-stimulatory counterpart CD226, raising the possibility that CD226 might influence Areg expression in the absence of TIGIT. To test this, we infected Tigit-KO and WT mice with LCMV and treated them with anti-CD226 or isotype control. CD226 blockade did not alter Areg expression in either WT or Tigit-KO T_reg cells, indicating that the reduction in Areg seen in Tigit-KO T_reg cells is not mediated by compensatory CD226 signaling, but rather reflects a direct requirement for TIGIT (Fig. 3l). We next examined the expression dynamics of TIGIT, CD226 and their shared ligand CD155 to understand their interplay in this regulatory pathway. TIGIT expression in T_reg cells was unaffected by CD226 blockade (Extended Data Fig. 5a). CD155, which is highly expressed on T cells, was slightly upregulated in WT T cells following CD226 blockade, but was mildly reduced in Tigit-KO T_reg cells (Extended Data Fig. 5b–d). By contrast, CD155 levels in myeloid cells remained unchanged between WT and Tigit-KO mice (Extended Data Fig. 5e–g). CD226 expression was also mostly unaffected in T cells from both full and conditional Tigit-KO mice (Extended Data Fig. 5h,i).

To address whether CD155 is required for Areg induction downstream of TIGIT, we analyzed Cd155-KO mice (Extended Data Fig. 5j). After LCMV infection, T_reg cells from Cd155-KO mice failed to upregulate Areg to the same extent as WT control mice, despite unaltered TIGIT expression (Fig. 3m and Extended Data Fig. 5k), suggesting that in vivo CD155 engagement promotes TIGIT-driven Areg expression in T_reg cells.

TIGIT induces Areg in T_reg cells in a TCR-dependent manner

Tissue T_reg cells associated with regenerative properties were found to have an oligoclonal TCR repertoire that shows considerable overlap with its precursor population in the spleen¹⁷. We thus wondered whether the T cell antigen receptor (TCR) repertoire of TIGIT⁺ T_reg cells that produce Areg in LCMV infection overlaps with this tissue T_reg cell repertoire. We performed single-cell TCR sequencing of WT and Tigit-KO T cells from the spleens and lungs of both naive and infected mice (Extended Data Fig. 6a–h). The proportion of expanded clones strongly increased after infection in both WT and KO groups, with an overall higher proportion of clonally expanded T cells in the lung than in the spleen (Fig. 4a–c and Extended Data Fig. 6i–p). In WT mice, the majority of the expanded T_reg cell clones after infection wereTigit⁺ (Fig. 4d), which is in line with TIGIT expression marking T_reg cells with a highly activated phenotype⁴. Next, we looked at clonal overlap to determine whether our TCR sequences correspond to those previously reported for tissue T_reg cells. For comparison, we also included TCR repertoires of effector T cells obtained after LCMV infection^{17,18,19,20,21}. When comparing clonotypes within our dataset, we found a partial overlap between mice within the same organ, but not necessarily between different organs, and, overall, only very few clones were shared between mice (Extended Data Fig. 7a,b). Most importantly, no overlap was seen between the previously reported tissue T_reg cell TCR repertoire and our dataset, while both CD4⁺Foxp3⁻ and CD8⁺ effector T cells partially overlapped with existing datasets (Fig. 4e). This suggests that the Tigit-expressing T_reg cells that expand following LCMV infection are not the classical tissue T_reg cells previously associated with tissue repair. To further pursue this notion, we assessed transcriptional and phenotypic differences between classical tissue T_reg cells and TIGIT⁺ repair T_reg cells and found them to have distinct transcriptional profiles and different cluster composition (Fig. 4f,g and Extended Data Fig. 7c–f). Infection-induced TIGIT⁺ repair T_reg cells also did not express the tissue T_reg cell markers ST2, KLRG1 or GATA3 at the transcriptional or protein level (Fig. 4f,g and Extended Data Fig. 7d–f). Additionally, we observed that LCMV infection broadly shifted the T_reg cell compartment toward a higher frequency of peripherally induced T_reg cells, including both TIGIT⁺ and TIGIT⁻ populations, consistent with previous findings²² (Extended Data Fig. 7f). These results confirm that TIGIT⁺ repair T_reg cells represent an infection-induced, phenotypically and transcriptionally distinct population versus classical tissue T_reg cells.

**Fig. 4: TIGIT⁺ repair T_reg cells are clonally expanded and distinct from classical tissue T_reg cells.**

When we further analyzed our dataset for clonal expansion based on the transcriptional expression of Tigit and Areg, we noticed that a higher number of Tigit⁺Areg⁺ T cells was clonally expanded than Tigit⁻Areg⁺ T cells, which was particularly pronounced after infection (Fig. 5a). We thus wondered whether TCR signaling itself might contribute to Areg induction in TIGIT⁺ T_reg cells. To test whether direct TCR stimulation would lead to increased Areg production, we stimulated splenocytes from naive and LCMV-infected WT mice with titrated amounts of anti-CD3 in vitro. Indeed, Areg levels increased proportionally with increasing TCR stimulation, particularly in T_reg cells (Fig. 5b and Extended Data Fig. 8a). Importantly, the fraction of T_reg cells that produced Areg following in vitro stimulation with anti-CD3 was characterized by TIGIT expression, while Areg induction was not seen in TIGIT⁻ T_reg cells (Fig. 5c). Most TIGIT⁺Areg⁺ T_reg cells did not stain with LCMV tetramers (Extended data Fig. 8b,c), suggesting that TIGIT⁺Areg⁺ T_reg cells arising during LCMV infection are mostly not virus specific but likely activated by self-antigens derived from damaged tissue. Interestingly, TCR-dependent induction of Areg was only observed in in vivo-primed T cells, whereas T cells isolated from naive mice did not produce significant levels of Areg following TCR stimulation (Fig. 5b,c), which is in line with previous reports¹¹. Thus, we next asked which infection-induced factors might license naive T_reg cells to express Areg. Co-stimulation with anti-CD28 in addition to anti-CD3 enabled naive T_reg cells to upregulate Areg expression to levels comparable to those observed in infected T_reg cells but did not further enhance Areg expression in in vivo-primed T_reg cells (Fig. 5d and Extended Data Fig. 8d,e). Additional stimulation with cytokines previously reported to induce Areg expression, such as IL-18, TGFβ or Areg itself^11,23, did not further enhance Areg expression beyond anti-CD3/anti-CD28 stimulation (Extended Data Fig. 8e). These results show that TCR signaling promotes Areg production by TIGIT⁺ T_reg cells in vitro and that co-stimulation is required to license naive T_reg cells for Areg expression following TCR stimulation. Importantly, when cells from infected Foxp3^GFPcreERT2 × Tigit^fl/fl mice were stimulated in the same manner, we found that TCR stimulation was unable to induce Areg expression in Tigit-KO T_reg cells, highlighting the requirement of both the TCR as well as TIGIT for Areg induction (Fig. 5e). By contrast, CD8⁺ T cells from infected Cd8^Cre × Tigit^fl/fl mice showed comparable Areg production as that observed in cells from WT mice following in vitro stimulation (Extended Data Fig. 8f), confirming that TIGIT on CD8⁺ T cells is dispensable for Areg production.

**Fig. 5: Areg is induced through TIGIT in a TCR-dependent manner.**

To further explore downstream signaling in TIGIT⁺ T_reg cells, we performed single-cell regulatory network inference and clustering (SCENIC)²⁴ analysis of our scRNA-seq data. Comparison of WT and Tigit-KO cells revealed distinct regulons in the two genotypes (Fig. 5f,g and Extended Data Fig. 8g), suggesting they may contribute to TIGIT-dependent functions. In line with TIGIT⁺ T_reg cells being highly suppressive⁴, we observed increased scores for the IL-2 inhibitors Crem and Ets2 (refs. ^25,26), as well as Helios, Fos and Junb, which are important for T_reg cell development and stability^27,28, in the presence of TIGIT (Fig. 5f,g). TIGIT⁺ T_reg cells also expressed higher levels of JunB protein (Extended Data Fig. 8h,i). Conversely, and in line with TIGIT’s co-inhibitory activity, transcription factors downstream of TCR signaling, such as Nfatc1, Tcf7 and Jun, were upregulated in Tigit-KO mice (Fig. 5f,g). Focusing on T_reg cells, in vitro engagement of TIGIT showed only a minor impact on intermediates of TCR signaling, such as pAKT or pS6 (Extended Data Fig. 8j,k), confirming that TCR signaling in T_reg cells remains intact following TIGIT engagement. Together, we show that TIGIT⁺ T_reg cells, distinct from the classical tissue T_reg cell population but still expressing Areg, are clonally expanded. Areg production by these in vivo-primed, clonally expanded T_reg cells was strictly dependent on both TIGIT expression and TCR stimulation.

TIGIT engagement induces Areg through Blimp-1

To identify candidate transcription factors that could mediate signaling downstream of TIGIT to induce Areg transcription, we looked for transcription factors that were part of our repair signature and co-regulated with Tigit and Areg in our scRNA-seq dataset. Hif1a and Prdm1 were enriched in the effector T_reg cell cluster, in the proliferating and T_H1 CD4⁺ T cell clusters and in the exhausted, effector and proliferating CD8⁺ T cell clusters, which all showed high Tigit expression (Fig. 6a and Extended Data Fig. 2e). HIF1α is upregulated under stressful conditions that cause hypoxia and through STAT3 and TCR signaling²⁹, whereas Prdm1, encoding Blimp-1, is induced by TCR stimulation in conjunction with inflammatory cytokines^30,31. We observed a significant correlation between Areg and both Hif1a and Prdm1 expression after infection, especially in T_reg cells, and thus tested their impact on Areg induction following LCMV infection (Fig. 6b and Extended Data Fig. 9a). To determine whether Hif1α can promote Areg production in T cells, we infected Cd4^Cre × Hif1a^fl/fl mice or Hif1a^fl/fl control mice with LCMV. Cd4^Cre × Hif1a^fl/fl mice failed to induce Areg expression to WT levels in any of the T cell populations addressed (Extended Data Fig. 9b,c), demonstrating that HIF1α is essential for Areg production by T cells. However, the absence of TIGIT did not affect HIF1α expression as WT and Tigit-KO T cells showed comparable Hif1a transcript expression both at the single-cell level and by quantitative real-time PCR (qPCR) of sorted T cell populations (Extended Data Fig. 9a,d,e). Hence, the impact of HIF1α on Areg expression is independent of TIGIT. We next focused on Blimp-1 (encoded by Prdm1), which also showed reduced regulon activity in Tigit-KO cells (Fig. 5f), as a potential downstream mediator of TIGIT-dependent induction of Areg. We observed a substantial correlation between Tigit and Prdm1 expression in the scRNA-seq data (Fig. 6c), with a strong Prdm1 downregulation in Tigit-KO T cells, particularly in T_reg cells and CD8⁺ T cells from infected mice. We thus measured Prdm1 transcript expression in sorted TIGIT⁺ versus TIGIT⁻ T cells and detected higher Prdm1 mRNA expression in the TIGIT⁺ T cell subpopulations (Fig. 6d and Extended Data Fig. 9f). We further validated this dependence by performing qPCR on sorted T cells from WT and Tigit-KO mice, confirming reduced Prdm1 transcription in Tigit-KO T_reg cells (Fig. 6e and Extended Data Fig. 9g). TIGIT and Blimp-1 expression also correlated at the protein level, where Blimp-1 was exclusively expressed in TIGIT⁺ T_reg cells and induced only after infection with LCMV or influenza (Fig. 6f,g and Extended Data Fig. 9h,i). Moreover, Blimp-1 was downregulated in vivo in the absence of TIGIT engagement in Cd155-KO mice (Extended Data Fig. 9j). Furthermore, Blimp-1 downregulation in Tigit-KO mice was a direct effect of Tigit deletion, as its expression was not further altered by CD226 blockade (Extended Data Fig. 9k). Finally, to provide a functional link between TIGIT, Blimp-1 and Areg, we stimulated splenocytes with TIGIT and TCR agonists in vitro and found that the combination of TIGIT plus TCR engagement not only resulted in Areg induction by T_reg cells but also increased Blimp-1 expression (Fig. 6h,i). Hence, TIGIT directly induces Blimp-1 in activated T_reg cells.

**Fig. 6: TIGIT induces Areg through Blimp-1.**

To determine whether Blimp-1 can regulate Areg transcription, we made use of our Blimp-1 chromatin immunoprecipitation with sequencing (ChIP–seq) data³² and found that Blimp-1 directly binds upstream of the Areg promoter and also regulates its expression in tissue T_reg cells³¹ (Fig. 6j). Furthermore, Areg transcription was downregulated in Prdm1-KO T_reg cells isolated from Lck^Cre × Prdm1^fl/fl mice (Fig. 6j). Most importantly, the induction of Areg protein following LCMV infection was impaired in Lck^Cre × Prdm1^fl/fl mice, confirming their functional dependence on the protein level (Fig. 6k,l). Moreover, ALT levels and cellular infiltrates, both characteristic for liver pathology, were increased in Lck^Cre × Prdm1^fl/fl mice compared to in WT control mice, confirming an involvement of Blimp-1 in the repair program we identified (Fig. 6m,n). Thus, our data show that the Areg-dependent tissue repair function of T_reg cells is regulated by both HIF1α and Blimp-1. We further uncover a signaling axis in which the engagement of TIGIT and TCR induces the transcription factor Blimp-1, which then directly drives Areg transcription to facilitate tissue repair following infection.

Discussion

Co-inhibitory receptors are essential for ensuring a controlled immune response during infection, preventing immune-mediated tissue damage. Without this regulation, collateral damage can lead to immune pathology as seen in infections like SARS-CoV-2 and in individuals with cancer receiving checkpoint therapy, where immune-related tissue damage can occur following blockade. In this study, we demonstrate that the co-inhibitory receptor TIGIT not only restrains effector immune responses but also actively promotes tissue repair. Specifically, we show that TIGIT induces Areg in a Blimp-1-dependent manner, activating a tissue repair program in T_reg cells that is essential for limiting pathology during infection.

We have previously shown that TIGIT plays a critical role in limiting immune pathology during LCMV infection by shifting the cytokine balance and decreasing tissue damage⁷. Consistent with these findings, we observed increased systemic pathology in Tigit-KO mice, with no major changes in T cell activation and immune cell composition. Instead, we identify a transcriptional tissue repair signature in TIGIT-expressing T cells, which is strongly impaired following TIGIT loss, revealing an additional protective function of TIGIT in promoting tissue repair and independent from its immune-suppressive capacity.

Over the past decade, growing evidence has highlighted a key role for T cells, particularly T_reg cells, in tissue repair^10,11,33. At steady state, tissue-resident T_reg cells promote homeostasis and regeneration, maintaining tissue integrity and facilitating repair after sterile injuries, such as muscle damage and stroke^8,12. They also promote metabolic homeostasis, where T_reg cells in visceral adipose tissue influence adipogenesis and insulin sensitivity^34,35. These tissue T_reg cells are characterized by IL-33 receptor (ST2) expression and Areg production in response to soluble mediators like IL-33, adipokines and metabolites, enabling tissue–immune cross-talk. However, the mechanisms driving Areg induction in T_reg cells remain poorly understood. Cytokines like IL-18 and IL-33 can promote Areg expression in T_reg cells independently of TCR signaling and unrelated to their suppressive function¹¹. In line with these findings, we did not observe Areg induction in steady-state T_reg cells without additional co-stimulatory signals. However, we found that during an active immune response, such as the one induced following an infection, Areg production by in vivo-primed T_reg cells is driven by combined signals from the TCR and TIGIT. Additionally, we observed clonal expansion of Areg-expressing TIGIT⁺ T_reg cells during infection, emphasizing the antigen-driven expansion of this tissue repair T_reg cell subset. As the Areg-secreting TIGIT⁺ T_reg cells are not LCMV-specific, this is likely driven by self-antigens released from damaged tissues following infection. Notably, the TCR repertoire of Areg-producing TIGIT⁺ T_reg cells that promote tissue repair in this context differed from that of ST2⁺ tissue T_reg cells, as did their transcriptional and proteomic profiles, clearly distinguishing them from this previously described subset^17,36. Furthermore, TIGIT blockade or deletion severely impaired Areg expression and reduced the entire transcriptional repair signature in T cells, highlighting the importance of TIGIT in tissue repair following infection. Differences in the magnitude of Areg induction between the LCMV and influenza infection models are likely due to different pathology and extent of tissue damage. Although LCMV is a noncytolytic virus, the early phase of Cl13 infection is dominated by extensive immune-mediated pathology, whereas the influenza infection model used here caused limited pathology due to the low infectious dose used. Consistent with this, we also observed that the distinct pathological profiles of LCMV WE and Cl13 infections translated into differences in the magnitude of Areg induction, further supporting the notion that the extent of tissue damage is a key determinant of the Areg⁺ T_reg cell response. Interestingly, TCR-dependent Areg expression has also been reported in human T cells, which are unresponsive to IL-18 or IL-33 stimulation, as well as in mouse models of ischemic stroke³³. Similarly, mucosal-associated invariant T cells upregulate Areg expression in a TCR-dependent manner during liver fibrosis. A recent study identified human PD-1⁺TIGIT⁺CD8⁺ T cells that express Areg, suggesting a role in tissue remodeling³⁷. In our model, TIGIT expression in CD8⁺ T cells was not required for Areg production nor to limit pathology, suggesting distinct regulatory mechanisms for tissue repair by CD8⁺ T cells and T_reg cells. Together, these findings indicate that Areg induction in T_reg cells is highly context dependent, with combined TCR and TIGIT signaling being critical during infection.

Despite the progress in understanding the signals that induce Areg, the transcription factors regulating its expression remain largely unknown. Here, we identified HIF1α and Blimp-1 as key regulators of Areg in T_reg cells during infection, with HIF1α acting independently of TIGIT. Interestingly, TIGIT and HIF1α may act synergistically to inhibit tumor growth in preclinical cancer models³⁸, but modulation of Areg expression was not assessed in this study.

The role of Blimp-1 in immune regulation and T cell differentiation is well established^39,40. Blimp-1, known to be induced by TCR signaling and IL-2 (ref. ⁴¹), is crucial for T_reg cell stability in inflamed tissues⁴² and marks highly suppressive effector T_reg cells, playing an essential role in IL-10 production^43,44. Blimp-1 is closely associated with effector T_reg cell differentiation and, together with c-Maf, co-regulates IL-10 expression in effector T cells^43,45. Here, we show that Blimp-1 drives the expression of Areg in T_reg cells, with its induction being dependent on combined signals from the TCR and TIGIT. Notably, TIGIT is also expressed on highly suppressive T_reg cells and directly promotes IL-10 production⁴. This raises the possibility that TIGIT may enhance the suppressive function of T_reg cells and their IL-10 production through Blimp-1 induction, a hypothesis that requires further investigation.

The primary function of most co-inhibitory receptors is to limit immune activation, as seen in alterations in and KO of Ctla4, which leads to severe autoimmunity^46,47, and Pdcd1, which is crucial for controlling the activation of autoreactive T cells in peripheral tissues^48,49. Although TIGIT also dampens T cell activation, it is less potent in this regard than other co-inhibitory receptors⁵. However, TIGIT stands out for its ability to shift the cytokine environment toward a T_H2-like response^4,50, a profile traditionally linked to tissue repair. In this study, we further demonstrate that TIGIT induces the expression of Areg, a key mediator of tissue repair. These findings could potentially serve as a basis for new therapeutic approaches for fibrotic disorders or chronic wounds where targeting TIGIT may promote and improve controlled tissue repair. Thus, beyond merely suppressing immune responses, TIGIT plays a dominant role in actively promoting tissue regeneration. Our findings expand the understanding of co-inhibitory receptors, revealing their capacity to not only limit immune activation but also actively contribute to tissue repair processes.

Methods

Mice

Experiments were conducted according to the institutional policies and national regulations and have been reviewed and approved by the Cantonal Veterinary Office and the University of Melbourne Animal Ethics Committee. B6 mice were purchased from Janvier Labs. Foxp3–GFP.KI and Tigit^fl/fl mice have been described previously^51,52 (from V. K. Kuchroo, Gene Lay Institute of Immunology and Inflammation). B6.129(Cg)-Foxp3^{tm4(YFP/icre)Ayr}/J (Foxp3-cre), Foxp3^{tm9(EGFP/cre/ERT2)Ayr}/J⁵³ (Foxp3^creERT2) and B6-Tg(Cd8a-cre)1Itan/J (Cd8^Cre) strains were obtained from The Jackson Laboratory and crossed with the Tigit^fl/fl strain. B6.Cg-Tg(Cd4-cre)1Cwi/BfluJ (Cd4^Cre) and B6.129-Hif1a^tm3Rsjo/J (Hif1a^fl/fl) mice were provided by the laboratory of N. Aceto (Swiss Federal Institute of Technology Zurich). B6N.129S2-Pvr^tm1Gbn/J (Cd155-KO) mice⁵⁴ were kindly provided by the laboratory of T. Korn (Technical University of Munich). Prdm1^fl/fl mice were previously described⁴⁰. All mice were on a B6 background. Adult 7- to 45-week-old male and female mice were used, age and sex were matched between experimental groups, and animals with the same genotype were randomized to the different treatment groups. Lck^Cre × Prdm1^fl/fl mice were bred and housed at the Peter Doherty Institute, Melbourne, Australia, and all other strains were located at the University of Zurich or the University of Basel, Switzerland, under specific pathogen-free conditions. Animals were housed in individually ventilated cages containing autoclaved bedding and nesting material, with standard diet and water ad libitum on a 12-h light/12-h dark cycle (18–23 °C, 40–60% humidity).

In vivo treatments and infections

To induce Cre-dependent KO, Foxp3^{tm9(EGFP/cre/ERT2)Ayr}/J⁵³ (Foxp3^creERT2) × Tigit^fl/fl mice received intraperitoneal (i.p.) injections of tamoxifen (2.5 mg in oil) or vehicle as a control three times per week throughout the course of infection starting 1 week before infection. For TIGIT blockade, B6 mice were injected i.p. with 100 μg of anti-TIGIT (1B4)¹⁶ or IgG1 isotype control (BioXCell) on days –2, 0, 2, 5 and 8 of infection. For CD226 blockade, B6 WT or Tigit-KO mice were injected i.p. with 100 μg of anti-CD226 (10E5) or IgG2b isotype control.

For LCMV infections, virus was produced as previously described⁷, and mice were infected intravenously (i.v.) with either 2 × 10⁶ focus-forming units of LCMV Cl13 or 200 focus-forming units of LCMV WE. To determine vascular integrity, mice were injected i.v. with 200 μl of 0.5% Evans blue dye (Sigma) in sterile PBS 20 min before euthanasia. Animals were perfused with PBS, organ weights were measured, and organs were placed in 1 ml of formamide (Sigma) and incubated at 56 °C overnight. The amount of extracted Evans blue dye was measured by photometry at 620 nm and quantified against a standard curve using the following formula: (Evans blue concentration (ng ml⁻¹) × volume of formamide (ml)) / total organ weight (g). Serum AST and ALT were measured at the Clinical Chemistry Department of the University Hospital Zurich by determining the catalytic concentration of both enzymes at 37 °C with pyridoxal phosphate activation on a Roche Cobas 8000 (c502) according to International Federation of Clinical Chemistry and Laboratory Medicine.

Immune cell isolation

Single-cell suspensions of spleen tissue were generated by mechanical disruption in RPMI 1640 (Gibco) supplemented with 10% fetal calf serum (Corning or Capri), penicillin and streptomycin (100 U ml⁻¹; Gibco) and 2 mmol l⁻¹ L-glutamine (Gibco). Red blood cells were lysed using ACK lysis buffer (155 mM NH₄Cl, 10 mM KHCO₃ and 0.1 mM Na₂EDTA, pH 7.4) for 3–5 min. Single-cell suspensions of lung tissue were generated by enzymatic tissue digestion at 37 °C with 0.2 mg ml⁻¹ deoxyribonuclease I (Sigma) and 2.4 mg ml⁻¹ collagenase type I (Gibco) and mechanical disruption using a GentleMACS. Lung immune cells were separated from epithelial and parenchymal cells by 30% Percoll gradient centrifugation (GE Healthcare).

Flow cytometry

Single-cell suspensions were obtained from spleen and lung tissue processed in RPMI 1640 (Gibco) supplemented with Brefeldin A (BioLegend) and marimastat (Sigma) and directly stained ex vivo or incubated at 37 °C for 2 h and 45 min with Brefeldin A (BioLegend), monensin (BioLegend) and marimastat (Sigma). Samples were incubated with surface antibodies diluted in PBS for 20 min at 22 °C. For intracellular cytokine staining, cells were permeabilized using a Cytofix/Cytoperm kit (BD Bioscience) for 5–10 min at 22 °C, followed by a 30-min incubation at 22 °C with a mix of intracellular antibodies. For transcription factor staining, cells were permeabilized with a Foxp3/Transcription factor staining buffer set (eBioscience) for 40 min at 22 °C. The Zombie NIR and LD Blue fixable dyes were used to exclude dead cells. Data were acquired on a BD LSR Fortessa, BD FACS Symphony A5 (BD Bioscience) or Cytek Aurora (Cytek Biosciences) and analyzed using SpectroFlo (Cytek Biosciences, v3.0.3) and FlowJo (BD Bioscience, v10.8.2 or 10.10.0) software. For R analysis, unmixed and pregated cells were imported into RStudio using R (v4.4.1) and the flowCore (v2.16.0) and CATALYST (v1.28.0) packages^55,56. Data were transformed using a hyperbolic arcsine transformation. This was followed by dimensionality reduction applying UMAP and FlowSOM algorithms. Clusters were annotated by overlaying the FlowJo workspace using the flowWorkspace (v4.16.0) and CytoML (v2.16.0) packages^57,58. Heat maps and stacked population plots were generated based on the frequency of respective marker expression and visualized using the ggplot2 (v3.5.1) package in R.

The following anti-mouse monoclonal antibodies were used: Areg, Alexa Fluor 647, 1:100, G-4, Santa Cruz; Areg, Biotin, 1:200, R&D; B220, BUV563, 1:200, RA3-6B2, BD; Blimp-1, APC, 1:100, 5E7, BioLegend; Blimp-1, BV421, 1:100, 5E7, BD; CD112, BV711, 1:100, 829038, BD; CD11b, PerCP, 1:400, M1/70, BioLegend; CD11c, BUV496, 1:100, HL3, BD; CD11c, BV605, 1:100, N418, BioLegend; CD134, BV421, 1:100, OX-86, BioLegend; CD155, BUV737, 1:100, TX56, BD; CD155, PE, 1:300, TX56, BioLegend; CD226, APC/Fire750, 1:100, 10E5, BioLegend; CD226, PE-Cy7, 1:100, 10E5, BD; CD3, BUV805, 1:100, 17A2, eBioscience; CD3, BV785, 1:200, 17A2, BioLegend; CD4, BUV737, 1:300, RM4-5, BD; CD4, APC, 1:200, RM4-5, BioLegend; CD4, BUV496, 1:500, RM4-5, BD; CD4, APC-Cy7, 1:300, GK1.5, BD; CD44, APC/Fire750, 1:300, IM7, BioLegend; CD45, BUV805, 1:300, 30-F11, BD; CD62L, BV650, 1:400, MEL-14, BioLegend; CD8, BUV395, 1:500, 53-6.7, BD; CXCR3, Biotin, 1:100, SA051D1, BioLegend; F4/80, APC-Fire810, 1:100, BM8, BioLegend; Foxp3, FITC, 1:200, FJK-16s, eBioscience; Foxp3, PE, 1:200, FJK-16s, Thermo; Foxp3, eF450, 1:200, FJK-16s, eBioscience; GATA3, PE-Cy5, 1:200, TWAJ, eBioscience; GM-CSF, PE-Cy7, 1:100, MP1-22E9, BioLegend; granzyme B, APC, 1:100, GB12, Thermo; GFP tag, Alexa Fluor 488, 1:400, FM264G, BioLegend; Helios, BUV563, 1:100, 22F6, eBioscience; IFNγ, PE, 1:300, XMG1.2, BioLegend; JunB, CoraLitePlus488, 1:125, ProteinTech; Ki-67, BUV737, 1:300, SolA15, Thermo; KLRG1, BUV661, 1:200, 2F1, BD; LD, Zombie NIR, 1:500, BioLegend; LD, Blue, 1:500, Thermo; Ly6G, BV750, 1:200, 1A8, BD; NK1.1, BUV615, 1:200, PK136, BD; NRP-1, BV711, 1:500, V46-1954, BD; pAKT, Alexa Fluor 647, 1:10, M89-61, BD; PD-1, BV785, 1:500, 29F.1A12, BioLegend; pS6, PerCP-eF710, 1:10, cupk43k, eBioscience; ST2, Biotin, 1:100, DIH9, BioLegend; streptavidin, BUV563, 1:800, eBioscience; streptavidin, BV711, 1:300, BioLegend; streptavidin, BV480, 1:500, BD; TCF1/TCF7, R718, 1:100, S33-966, BD; TCRβ, PE-Dazzle594, 1:300, H57-597, BioLegend; TCRγδ, BB700, 1:200, GL3, BD; TIGIT, PE-Dazzle594, 1:50, 1G9, BioLegend; TIGIT, BV421, 1:50, TX99, BD; TIGIT, PerCP-eF710, 1:200, GIGD7, Thermo; TNF, BV421, 1:100, MP6-XT22, BioLegend. PE-labelled tetramers for LCMV GP_66-77, LCMV GP_126-140, LCMV NP_309-328 were used at 1:100 and provided by the NIH Tetramer Core Facility (NIH Contract 75N93020D00005 and RRID:SCR_026557).

Immunohistochemistry

Livers were perfused at the moment the mice were killed, and lobes were collected and fixed overnight at 4 °C in 4% formalin. The tissues were subsequently processed, embedded in paraffin blocks and sectioned for H&E staining and caspase-3 immunohistochemistry. Primary anti-cleaved caspase-3 (polyclonal rabbit, Cell Signaling Technology, clone Asp175, 9661) was applied to the tissue sections at a 1:500 dilution and incubated for 2 h at 22 °C to ensure optimal binding. After primary antibody incubation, sections were treated with horseradish peroxidase-conjugated anti-rabbit secondary antibodies for 30 min at 22 °C. The antigen–antibody complex was visualized using 3,3’,5,5’-tetramethylbenzidine as a substrate. Images were captured using the Axio Scan Z.1 (Zeiss) and analyzed using ImageScope software (Leica Biosystem), following the manufacturer’s instructions.

For image quantification, QuPath software was used to perform automated analysis of tissue samples, ensuring consistent and reliable cell counting across all ROIs. For caspase-3 analysis, multiple ROIs were examined, with cell boundaries delineated using hematoxylin counterstaining. Caspase-3⁺ hepatocytes were automatically identified based on a color threshold. The data were expressed as the percentage of caspase-3⁺ hepatocytes per liver sample. Total cellularity for each sample was calculated using the same ROI-based approach, with QuPath’s segmentation algorithm facilitating the automated cell counts. Analysis was performed blind to the conditions of the experiments.

In vitro stimulation

Isolated splenocytes were stimulated at 37 °C in 10% CO₂ for 2 h and 45 min with titrated amounts of anti-CD3 (clone 145-2C11, BioXCell), anti-CD28 (clone PV-1, BioXCell) and 50 ng ml⁻¹ agonistic anti-TIGIT (4D4) or corresponding IgG1 antibody in the presence of Brefeldin A (BioLegend), monensin (BioLegend) and marimastat (Sigma). Where indicated, IL-18 (10 ng ml⁻¹; BioLegend), TGFβ (10 ng ml⁻¹; BioLegend) or Areg (10 ng ml⁻¹; BioLegend) were added to the culture.

scRNA-seq and TCR-seq

One day before LCMV infection or killing of naive mice, animals were adoptively transferred with MACS-purified LCMV-specific or nonspecific CD4⁺ T cells (Smarta-1 or OTII, 2.5 × 10³ and 10⁵ cells i.v., respectively) and CD8⁺ T cells (P14 or OTI, 5 × 10³ and 2 × 10⁵ cells i.v., respectively). Spleen and lung single-cell suspensions were obtained from naive and LCMV-infected (day 10) WT and Tigit-KO mice. CD3⁺ T cells were purified by MACS using a CD3 Cell Isolation kit (negative selection, BioLegend) and then sorted by FACS using a BD FACSAria III 5L (BD Bioscience). Sorted CD3⁺CD4⁺Foxp3⁺ and CD3⁺Foxp3⁻ cells were mixed at a ratio of approximately 1:3 and labeled before pooling with anti-mouse Hashtag reagents (TotalSeq-C301, 302, 303, 304, 305 anti-mouse Hashtag 1, 2, 3, 4, 5, respectively, BioLegend). Library preparation was performed by the Functional Genomic Center Zurich using 10x Chromium Next GEM Single Cell 5’ Reagent kits, according to manufacturer’s instructions (10x Genomics). scRNA-seq/TCR-seq libraries were sequenced on an Illumina Novaseq 6000 S4 (Illumina).

scRNA-seq analysis

Processing

In total, three to five 10x runs were generated per condition. For each sample, we generated the digital gene expression matrix from the raw reads using Cell Ranger 7.1.0 (ref. ⁵⁹). The unfiltered dataset contained a total of 261,094 cells. Using the digital gene expression matrix from each sample separately, DoubletFinder was applied to filter doublet cells⁶⁰, and subsequent filters were used to remove cells with high mitochondria counts and low RNA content. After quality control, the datasets were merged to a dataset consisting of 198,645 cells with a median of 8,116 reads per cell using Seurat v5.0.1 (ref. ⁶¹). Count normalization was performed using Seurat’s NormalizeData routine to normalize each cell to 10,000 counts and log scale. Highly variable features within normalized counts were selected using the mvp method, which identified 596 highly variable features. Using the highly variable features, scaling of the genes to unit variance was performed using Seurat’s ScaleData routine; dimensionality reduction was performed using principal component analysis. Subsequent batch correction was performed using Harmony v1.2 (ref. ⁶²) and the 10x run identifier. Louvain clustering as implemented through Seurat’s FindClusters() outline was performed with a resolution of 0.5, yielding 15 clusters. Clusters that contained a mixture of Cd4⁺ and Cd8a⁺ cells were subsequently subclustered. Repair scores^12,63,64 were assigned by first normalizing the genes of the repair signature to unit variance without zero centering using Seurat’s ScaleData. Subsequently, a repair score was assigned to all single cells by summing the individual scores of each gene.

Data visualization

Seurat’s built-in functionality was used for UMAPs and dot plots. Gene expression UMAP heat maps were generated setting the order parameter to true. Blend UMAPs were generated using min.cutoff = ‘q01’ and max.cutoff = ‘q99’. For the scatter plots, each genotype and condition (naive and LCMV) were randomly split into ten technical replicates (cell pools), and the average normalized expression value (log₁₀ cpk (process capability index)) was calculated. Each technical replicate was split according to T cell subset. Correlation coefficients were calculated per cell type and condition. Visualizations were performed with ggplot2, and Pearson r correlations were obtained from ggpubr. For repair signature gene correlation heat maps, each value was calculated using single-cell expression values from the subsequent major cell type (T_reg and CD4⁺ and CD8⁺ T_c_onv cells).

Integration with the Delacher et al. T_reg cell dataset

T_reg cells from our filtered dataset (19,100 cells) were integrated with the Delacher et al.³⁶ dataset (10,233 cells) using the Harmony method⁶² (v1.2; covariate and sample ID information was used to correct for batch effects). Seurat’s Louvain clustering routine using a resolution of 0.5 was used to calculate clusters of the integrated dataset. To visualize Delacher T_reg cells were grouped into tissue (fat, liver, lung and skin) and secondary and lymphoid organs (blood, bone marrow, lymph node and spleen).

Identification of statistically significant regulons

To identify regulons that are associated with TIGIT, randomized subsets of 5,000 T_reg and CD8⁺ T cells across infection and genotype conditions were generated. Regulons were generated using the SCENIC method²⁴ with default parameters for promoter and enhancer regions. T_reg, T_conv and CD8⁺ cells were projected onto the identified regulons using SCENIC’s AUCell method. Regulons with more than 70% zeros in all conditions according to their AUCell were removed, and permutation tests (n = 10,000) were performed for each of the cell types and conditions to identify statistically enriched regulons. After calculating the double-sided P values, multiple hypothesis testing corrections were performed using the Holm–Bonferroni method. Effect size is shown as the log (fold change) between WT and KO.

Single-cell TCR-seq analysis

Raw sequencing files arising from Illumina sequencing lanes were supplied as input to the command line program Cell Ranger (v7.0.0) on a high-performance cluster. Raw reads were aligned to the mouse reference genome (mm10) using 10x Genomics Cell Ranger and subsequently supplied into the VDJ_GEX_matrix function of the R package Platypus (v3.1)⁶⁵, which relies on the R package Seurat (v4.0.3)⁶⁶ and generates an integrated transcriptome and repertoire data object. Annotations from GEX were transferred to VDJ and vice versa by matching cellular 10x barcodes. Cells with more than 7% mitochondrial reads were filtered out, and gene expression was log normalized with a scaling factor of 10,000. UMAP was calculated, and feature plots were produced using the FeaturePlot function from Seurat. Cells were assembled into clones based on the default enclone clonotyping strategy within Cell Ranger. Smarta-1 and P14 TCR sequences were excluded from the repertoire analysis. Clonal frequency was determined by counting the number of distinct cell barcodes for each unique clonotype. Those cells in clones supported by only one cell were considered unexpanded clones, whereas those clones supported by two or more cells were considered expanded. Clonal expansion plots were created using the functions VDJ_clonal_expansion and VDJ_clonal_donut from Platypus⁶⁷. Expansion profiles of different T cell subsets were categorized based on the identification of the corresponding Seurat clusters using transcriptional information (Cd4⁺Foxp3⁺, Cd4⁺Foxp3⁻ and Cd8a⁺). All three T cell subsets were further subdivided into Tigit^+/− and Areg^+/− groups based on expression of the respective genes. Overlap matrices were calculated by quantifying the exact CDRb3 amino acid matches across repertoires. No filtering regarding the number of TRB and TRA chains was performed for the repertoire data, with the exception of the repertoire overlap comparison with the public datasets where those clones not containing exactly one TRA and one TRB chain were removed from the analysis. Jaccard indices were calculated by quantifying the intersection between two groups divided by the length of the union of the same groups. Heat maps displaying clonal overlap were produced using the pheatmap function in the pheatmap package (v1.0.12). Figures were imported from R (v4.3.2) and further processed using Adobe illustrator.

qPCR

RNA was extracted using an RNeasy kit (Qiagen) and reverse transcribed to cDNA using a High-Capacity cDNA RT kit (Life Technologies) according to the manufacturer’s instructions. qPCR was performed with a QuantStudio 5 Real-Time PCR system (384-well; Thermo). The following primer–probe mixtures were purchased from Applied Biosystems: Hif1a (Mm00468869_m1), Prdm1 (Mm00476128_m1), Areg (Mm01354339_m1), Il10 (Mm01288386_m1), Tgfb1 (Mm01178820_m1), Batf (Mm00479410_m1) and Nfil3 (Mm00600292_s1). For Tigit, the following primers and probe were used: 5′-CTGATACAGGCTGCCTTCCT-3′ (forward primer), 5′-TGGGTCACTTCAGCTGTGTC-3′ (reverse primer) and 5′-AGGAGCCACAGCAGGCACGA-3′ (probe; FAM, TAMRA). Actb (Thermo) was used as a housekeeping gene (VIC).

Blimp-1 ChIP–seq and RNA-seq

Blimp-1 ChIP–seq and RNA-Seq data were previously reported and analyzed for the Areg locus as previously described³². Briefly, Blimp-1 ChIP–seq reads were mapped to the mouse reference genome, and Blimp-1 binding peaks were called and assigned to their nearest gene. RNA-seq data were generated from Prdm1^fl/flFoxp3^Cre and control T_reg cells, and reads were aligned to the mouse reference genome and assigned to mouse genes.

Statistical analysis

Statistical analyses were performed using GraphPad Prism (GraphPad Software, v10.0.3). Appropriate statistical tests were selected as indicated in the figure legends with significant and relevant nonsignificant differences marked in the figures. Data collection and analysis were not blinded due to technical limitations, with the exception of the histological analysis. No data points were excluded, except when technical issues relating to the execution of the experiment or data acquisition were detected. Sample sizes were chosen based on power calculations using estimated effect size ranges that were based on previous publications^7,16,50. Data distribution was assumed to be normal, but this was not formally tested. To compare more than two groups, ordinary one-way or two-way ANOVA with a Tukey’s or Šídák’s multiple comparisons test was used. Significance was defined as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001 or NS (not significant; P > 0.05).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All scRNA-seq and single-cell TCR-seq data from WT and Tigit-KO T cells have been deposited on the ArrayExpress database at EMBL-EBI (www.ebi.ac.uk/arrayexpress) and is available via accession number E-MTAB-8861. The Seurat object is available at Zenodo at https://doi.org/10.5281/zenodo.14041419 (ref. ⁶⁸). Blimp-1 ChIP–seq and RNA-seq data were previously reported and are available in the Gene Expression Omnibus database under accession numbers GSE79339 and GSE121838. Source data are provided with this paper.

References

Burke, K. P., Chaudhri, A., Freeman, G. J. & Sharpe, A. H. The B7:CD28 family and friends: unraveling coinhibitory interactions. Immunity 57, 223–244 (2024).
CAS PubMed PubMed Central Google Scholar
Joller, N. et al. Cutting edge: TIGIT has T cell-intrinsic inhibitory functions. J. Immunol. 186, 1338–1342 (2011).
CAS PubMed Google Scholar
Yu, X. et al. The surface protein TIGIT suppresses T cell activation by promoting the generation of mature immunoregulatory dendritic cells. Nat. Immunol. 10, 48–57 (2009).
CAS PubMed Google Scholar
Joller, N. et al. T_reg cells expressing the coinhibitory molecule TIGIT selectively inhibit proinflammatory T_H1 and T_H17 cell responses. Immunity 40, 569–581 (2014).
Joller, N., Anderson, A. C. & Kuchroo, V. K. LAG-3, TIM-3, and TIGIT: distinct functions in immune regulation. Immunity 57, 206–222 (2024).
CAS PubMed PubMed Central Google Scholar
Johnston, R. J. et al. The immunoreceptor TIGIT regulates antitumor and antiviral CD8⁺ T cell effector function. Cancer Cell 26, 923–937 (2014).
CAS PubMed Google Scholar
Schorer, M. et al. TIGIT limits immune pathology during viral infections. Nat. Commun. 11, 1288 (2020).
CAS PubMed PubMed Central Google Scholar
Estrada Brull, A., Panetti, C. & Joller, N. Moving to the outskirts: interplay between regulatory T cells and peripheral tissues. Front. Immunol. 13, 864628 (2022).
Haertel, E., Joshi, N., Hiebert, P., Kopf, M. & Werner, S. Regulatory T cells are required for normal and activin-promoted wound repair in mice. Eur. J. Immunol. 48, 1001–1013 (2018).
CAS PubMed Google Scholar
Nayer, B. et al. Local administration of regulatory T cells promotes tissue healing. Nat. Commun. 15, 7863 (2024).
CAS PubMed PubMed Central Google Scholar
Arpaia, N. et al. A distinct function of regulatory T cells in tissue protection. Cell 162, 1078–1089 (2015).
CAS PubMed PubMed Central Google Scholar
Burzyn, D. et al. A special population of regulatory T cells potentiates muscle repair. Cell 155, 1282–1295 (2013).
CAS PubMed PubMed Central Google Scholar
Guan, T., Zhou, X., Zhou, W. & Lin, H. Regulatory T cell and macrophage crosstalk in acute lung injury: future perspectives. Cell Death Discov. 9, 9 (2023).
Kaiser, K. A., Loffredo, L. F., Santos-Alexis, K. L., Ringham, O. R. & Arpaia, N. Regulation of the alveolar regenerative niche by amphiregulin-producing regulatory T cells. J. Exp. Med. 220, e20221462 (2023).
CAS PubMed Google Scholar
Harb, H. et al. Notch4 signaling limits regulatory T-cell-mediated tissue repair and promotes severe lung inflammation in viral infections. Immunity 54, 1186–1199 (2021).
CAS PubMed PubMed Central Google Scholar
Dixon, K. O. et al. Functional anti-TIGIT antibodies regulate development of autoimmunity and antitumor immunity. J. Immunol. 200, 3000–3007 (2018).
CAS PubMed Google Scholar
Delacher, M. et al. Precursors for nonlymphoid-tissue T_reg cells reside in secondary lymphoid organs and are programmed by the transcription factor BATF. Immunity 52, 295–312 (2020).
CAS PubMed PubMed Central Google Scholar
Miragaia, R. J. et al. Single-cell transcriptomics of regulatory T cells reveals trajectories of tissue adaptation. Immunity 50, 493–504 (2019).
CAS PubMed PubMed Central Google Scholar
Shlesinger, D. et al. Single-cell immune repertoire sequencing of B and T cells in murine models of infection and autoimmunity. Genes Immun. 23, 183–195 (2022).
CAS PubMed PubMed Central Google Scholar
Kuhn, R. et al. Clonally expanded virus-specific CD8 T cells acquire diverse transcriptional phenotypes during acute, chronic, and latent infections. Front. Immunol. 13, 782441 (2022).
CAS PubMed PubMed Central Google Scholar
Khatun, A. et al. Single-cell lineage mapping of a diverse virus-specific naive CD4 T cell repertoire. J. Exp. Med. 218, e20200650 (2021).
CAS PubMed Google Scholar
Schorer, M. et al. Rapid expansion of T_reg cells protects from collateral colitis following a viral trigger. Nat. Commun. 11, 1522 (2020).
CAS PubMed PubMed Central Google Scholar
Minutti, C. M. et al. A macrophage–pericyte axis directs tissue restoration via amphiregulin-induced transforming growth factor β activation. Immunity 50, 645–654 (2019).
CAS PubMed PubMed Central Google Scholar
Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017).
CAS PubMed PubMed Central Google Scholar
Bodor, J., Fehervari, Z., Diamond, B. & Sakaguchi, S. ICER/CREM-mediated transcriptional attenuation of IL-2 and its role in suppression by regulatory T cells. Eur. J. Immunol. 37, 884–895 (2007).
CAS PubMed Google Scholar
Panagoulias, I. et al. Transcription factor ETS-2 acts as a preinduction repressor of interleukin-2 (IL-2) transcription in naive T helper lymphocytes. J. Biol. Chem. 291, 26707–26721 (2016).
CAS PubMed PubMed Central Google Scholar
Kim, H. J. et al. Stable inhibitory activity of regulatory T cells requires the transcription factor Helios. Science 350, 334–339 (2015).
CAS PubMed PubMed Central Google Scholar
Katagiri, T. et al. JunB plays a crucial role in development of regulatory T cells by promoting IL-2 signaling. Mucosal Immunol. 12, 1104–1117 (2019).
CAS PubMed Google Scholar
Dang, E. V. et al. Control of T_H17/T_reg balance by hypoxia-inducible factor 1. Cell 146, 772–784 (2011).
CAS PubMed PubMed Central Google Scholar
Xin, A. et al. A molecular threshold for effector CD8⁺ T cell differentiation controlled by transcription factors Blimp-1 and T-bet. Nat. Immunol. 17, 422–432 (2016).
CAS PubMed PubMed Central Google Scholar
Vasanthakumar, A. et al. Sex-specific adipose tissue imprinting of regulatory T cells. Nature 579, 581–585 (2020).
CAS PubMed PubMed Central Google Scholar
Mackay, L. K. et al. Hobit and Blimp1 instruct a universal transcriptional program of tissue residency in lymphocytes. Science 352, 459–463 (2016).
CAS PubMed Google Scholar
Ito, M. et al. Brain regulatory T cells suppress astrogliosis and potentiate neurological recovery. Nature 565, 246–250 (2019).
CAS PubMed Google Scholar
Feuerer, M. et al. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat. Med. 15, 930–939 (2009).
CAS PubMed PubMed Central Google Scholar
Cipolletta, D. et al. PPAR-γ is a major driver of the accumulation and phenotype of adipose tissue T_reg cells. Nature 486, 549–553 (2012).
CAS PubMed PubMed Central Google Scholar
Delacher, M. et al. Single-cell chromatin accessibility landscape identifies tissue repair program in human regulatory T cells. Immunity 54, 702–720 (2021).
CAS PubMed PubMed Central Google Scholar
Delacher, M. et al. The effector program of human CD8 T cells supports tissue remodeling. J. Exp. Med. 221, e20230488 (2024).
Fathi, M. et al. Simultaneous blockade of TIGIT and HIF-1α induces synergistic anti-tumor effect and decreases the growth and development of cancer cells. Int. Immunopharmacol. 101, 108288 (2021).
CAS PubMed Google Scholar
Kallies, A. et al. Initiation of plasma-cell differentiation is independent of the transcription factor Blimp-1. Immunity 26, 555–566 (2007).
CAS PubMed Google Scholar
Kallies, A., Xin, A., Belz, G. T. & Nutt, S. L. Blimp-1 transcription factor is required for the differentiation of effector CD8⁺ T cells and memory responses. Immunity 31, 283–295 (2009).
CAS PubMed Google Scholar
Gong, D. & Malek, T. R. Cytokine-dependent Blimp-1 expression in activated T cells inhibits IL-2 production. J. Immunol. 178, 242–252 (2007).
Garg, G. et al. Blimp1 prevents methylation of Foxp3 and loss of regulatory T cell identity at sites of inflammation. Cell Rep. 26, 1854–1868 (2019).
CAS PubMed PubMed Central Google Scholar
Cretney, E. et al. The transcription factors Blimp-1 and IRF4 jointly control the differentiation and function of effector regulatory T cells. Nat. Immunol. 12, 304–311 (2011).
CAS PubMed Google Scholar
Cretney, E. et al. Characterization of Blimp-1 function in effector regulatory T cells. J. Autoimmun. 91, 73–82 (2018).
CAS PubMed Google Scholar
Alvarez-Martinez, M. et al. Blimp-1 and c-Maf regulate immune gene networks to protect against distinct pathways of pathobiont-induced colitis. Nat. Immunol. 25, 886–901 (2024).
CAS PubMed PubMed Central Google Scholar
Ueda, H. et al. Association of the T-cell regulatory gene CTLA4 with susceptibility to autoimmune disease. Nature 423, 506–511 (2003).
CAS PubMed Google Scholar
Klocke, K., Sakaguchi, S., Holmdahl, R. & Wing, K. Induction of autoimmune disease by deletion of CTLA-4 in mice in adulthood. Proc. Natl Acad. Sci. USA 113, E2383–E2392 (2016).
CAS PubMed PubMed Central Google Scholar
Honda, T. et al. Tuning of antigen sensitivity by T cell receptor-dependent negative feedback controls T cell effector function in inflamed tissues. Immunity 40, 235–247 (2014).
CAS PubMed PubMed Central Google Scholar
Tan, C. L. et al. PD-1 restraint of regulatory T cell suppressive activity is critical for immune tolerance. J. Exp. Med. 218, e20182232 (2021).
CAS PubMed Google Scholar
Lozano, E., Joller, N., Cao, Y. K., Kuchroo, V. & Hafler, D. A. The CD226/CD155 interaction regulates the proinflammatory (T_H1/T_H17)/anti-inflammatory (T_H2) balance in humans. J. Immunol. 191, 3673–3680 (2013).
CAS PubMed Google Scholar
Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogenic effector T_H17 and regulatory T cells. Nature 441, 235–238 (2006).
CAS PubMed Google Scholar
Xiao, S. et al. Checkpoint receptor TIGIT expressed on Tim-1⁺ B cells regulates tissue inflammation. Cell Rep. 32, 107892 (2020).
CAS PubMed PubMed Central Google Scholar
Rubtsov, Y. P. et al. Stability of the regulatory T cell lineage in vivo. Science 329, 1667–1671 (2010).
CAS PubMed PubMed Central Google Scholar
Maier, M. K. et al. The adhesion receptor CD155 determines the magnitude of humoral immune responses against orally ingested antigens. Eur. J. Immunol. 37, 2214–2225 (2007).
CAS PubMed Google Scholar
Hahne, F. et al. flowCore: a Bioconductor package for high throughput flow cytometry. BMC Bioinformatics 10, 106 (2009).
PubMed PubMed Central Google Scholar
Crowell, H. L. et al. An R-based reproducible and user-friendly preprocessing pipeline for CyTOF data. F1000Res. 9, 1263 (2022).
PubMed Central Google Scholar
Finak, G., Jiang, W., Pardo, J., Asare, A. & Gottardo, R. QUAliFiER: an automated pipeline for quality assessment of gated flow cytometry data. BMC Bioinformatics 13, 252 (2012).
PubMed PubMed Central Google Scholar
Finak, G., Jiang, W. & Gottardo, R. CytoML for cross-platform cytometry data sharing. Cytometry A 93, 1189–1196 (2018).
PubMed PubMed Central Google Scholar
Zheng, G. X. Y. et al. Massively parallel digital transcriptional profiling of single cells. Nat. Commun. 8, 14049 (2017).
CAS PubMed PubMed Central Google Scholar
McGinnis, C. S., Murrow, L. M. & Gartner, Z. J. DoubletFinder: doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 8, 329–337 (2019).
CAS PubMed PubMed Central Google Scholar
Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902 (2019).
CAS PubMed PubMed Central Google Scholar
Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods 16, 1289–1296 (2019).
CAS PubMed PubMed Central Google Scholar
Patsalos, A. et al. A growth factor-expressing macrophage subpopulation orchestrates regenerative inflammation via GDF-15. J. Exp. Med. 219, e20210420 (2022).
CAS PubMed Google Scholar
Werner, S. & Grose, R. Regulation of wound healing by growth factors and cytokines. Physiol. Rev. 83, 835–870 (2003).
CAS PubMed Google Scholar
Cotet, T.-S. et al. ePlatypus: an ecosystem for computational analysis of immunogenomics data. Bioinformatics 39, btad553 (2023).
CAS PubMed PubMed Central Google Scholar
Papalexi, E. & Satija, R. Single-cell RNA sequencing to explore immune cell heterogeneity. Nat. Rev. Immunol. 18, 35–45 (2018).
CAS PubMed Google Scholar
Yermanos, A. et al. Platypus: an open-access software for integrating lymphocyte single-cell immune repertoires with transcriptomes. NAR Genom. Bioinform. 3, lqab023 (2021).
PubMed PubMed Central Google Scholar
Joller, N. J., Panetti, C. & Patikas, N. The co-inhibitory receptor TIGIT promotes tissue protective functions in T cells. Zenodo https://doi.org/10.5281/zenodo.14041419 (2024).

Download references

Acknowledgements

We would like to thank the members of the laboratories of N.J. and A. Oxenius for helpful discussions and members of the laboratory of N.J. for technical help. We thank the Functional Genomic Center Zürich and the Clinical Chemistry Department of the University Hospital Zürich for technical support and expertise and T. Korn for the Cd155-KO mice. We thank the NIH Tetramer Core Facility (NIH Contracts 75N93020D00005 and RRID:SCR_026557) for providing LCMV tetramers. We thank Chun-Hsi Su for his help with the infection experiments. This work was supported by the Swiss National Science Foundation (PP00P3_181037, 310030_197590 and 320030-236138 to N.J. and 197683 to C.G.K.), the European Research Council (677200 Immune Regulation to N.J. and 101141363 to M.I.), the EMBO Young Investigator Program (N.J.), the Italian Association for Cancer Research (19891 and 22737 to M.I.), the National Health and Medical Research Council (NHMRC, to A.V. and A.K.) and a fellowship from the University of Melbourne (to A.V.).

Author information

Authors and Affiliations

Department of Quantitative Biomedicine, University of Zurich, Zurich, Switzerland
Camilla Panetti, Rahel Daetwyler, Anja Moncsek & Nicole Joller
The Gene Lay Institute of Immunology and Inflammation of Brigham and Women’s Hospital, Massachusetts General Hospital, and Harvard Medical School, Boston, MA, USA
Nikolaos Patikas & Martin Hemberg
Department of Biosystems Science and Engineering, ETH Zurich, Basel, Switzerland
Andreas Agrafiotis & Alexander Yermanos
Institute of Microbiology, ETH Zurich, Zurich, Switzerland
Andreas Agrafiotis
Department of Microbiology and Immunology, The Peter Doherty Institute for Infection and Immunity, The University of Melbourne, Parkville, Victoria, Australia
Adelynn Tang, Lifen Wen, Ajithkumar Vasanthakumar & Axel Kallies
La Trobe University, Bundoora, Victoria, Australia
Adelynn Tang & Ajithkumar Vasanthakumar
Division of Immunology, Transplantation, and Infectious Diseases, IRCCS San Raffaele Scientific Institute, Milan, Italy
Francesco Andreata, Valeria Fumagalli & Matteo Iannacone
Vita-Salute San Raffaele University, Milan, Italy
Francesco Andreata, Valeria Fumagalli & Matteo Iannacone
Department of Biomedicine, University of Basel, Basel, Switzerland
Jean De Lima & Carolyn G. King
Olivia Newton-John Cancer Research Institute, Heidelberg, Victoria, Australia
Ajithkumar Vasanthakumar
Peter MacCallum Cancer Centre, Melbourne, Victoria, Australia
Ajithkumar Vasanthakumar
Institute of Molecular Medicine & Experimental Immunology, University Hospital Bonn, Bonn, Germany
Axel Kallies
Center for Translational Immunology, University Medical Center Utrecht, Utrecht, the Netherlands
Alexander Yermanos
Botnar Institute of Immune Engineering, Basel, Switzerland
Alexander Yermanos
Center for Human Immunology, University of Zurich, Zurich, Switzerland
Nicole Joller

Authors

Camilla Panetti
View author publications
Search author on:PubMed Google Scholar
Rahel Daetwyler
View author publications
Search author on:PubMed Google Scholar
Anja Moncsek
View author publications
Search author on:PubMed Google Scholar
Nikolaos Patikas
View author publications
Search author on:PubMed Google Scholar
Andreas Agrafiotis
View author publications
Search author on:PubMed Google Scholar
Adelynn Tang
View author publications
Search author on:PubMed Google Scholar
Francesco Andreata
View author publications
Search author on:PubMed Google Scholar
Valeria Fumagalli
View author publications
Search author on:PubMed Google Scholar
Jean De Lima
View author publications
Search author on:PubMed Google Scholar
Lifen Wen
View author publications
Search author on:PubMed Google Scholar
Carolyn G. King
View author publications
Search author on:PubMed Google Scholar
Ajithkumar Vasanthakumar
View author publications
Search author on:PubMed Google Scholar
Matteo Iannacone
View author publications
Search author on:PubMed Google Scholar
Axel Kallies
View author publications
Search author on:PubMed Google Scholar
Alexander Yermanos
View author publications
Search author on:PubMed Google Scholar
Martin Hemberg
View author publications
Search author on:PubMed Google Scholar
Nicole Joller
View author publications
Search author on:PubMed Google Scholar

Contributions

Conceptualization: C.P. and N.J. Experimentation: C.P., R.D., A.M., A.T., V.F., J.D.L. and L.W. Analysis: C.P., R.D., A.T. and F.A. Formal analysis: N.P., A.A. and A.V. Writing, original draft: C.P. and N.J. Review and editing: all authors. Funding acquisition: N.J. Supervision: C.G.K., M.I., A.K., A.Y., M.H. and N.J.

Corresponding author

Correspondence to Nicole Joller.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Immunology thanks Ofer Mandelboim, Kazuko Shibuya and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: L. A. Dempsey, in collaboration with the rest of the Nature Immunology team. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Tissue pathology is increased in LCMV Cl13 versus LCMV WE infection.

a. Heatmap of TIGIT expression in splenocytes from LCMV Cl13 infected (d10) WT and Tigit- KO mice. b., c. Relative Tigit mRNA (b.) and protein (c.) expression in splenic T cells sorted or gated by subset from LCMV Cl13 infected WT and Tigit- KO mice (mean ∓ SD; b. pool of 3 independent experiments, 2way ANOVA, Tukey’s test; c. pool of 2 independent experiments, unpaired two-sided t test). d. Serum AST and ALT levels in LCMV Cl13 and WE infection over time (mean ∓ SD; pool of 2 independent experiments, unpaired two-sided t test) and e. on day 10 p.i. and in naïve controls (mean ∓ SD; pool of 2 independent experiments, ordinary one-way ANOVA, Tukey’s test). f. Quantification of Evans Blue extravasation in liver and lung of naïve and LCMV Cl13 and WE infected mice (mean ∓ SD; pool of 3-8 independent experiments, ordinary one-way ANOVA, Tukey’s test, liver n = 20, 18, 12; lung n = 28, 29, 12). g. Cellular infiltrate in liver from naïve and LCMV Cl13 infected WT and Tigit-KO mice (mean ∓ SD; 1 representative experiment, selected ROI, WT n = 15, 23; KO n = 21, 16, biological replicates n = 2-3, 2way ANOVA, Tukey’s test) and representative hematoxylin and eosin (H&E) staining of infected liver sections. h. IFN-γ, TNF-α and Granzyme B frequencies in NK cells from LCMV Cl13 infected WT and Tigit-KO spleens (mean ∓ SD; 1 representative experiment, unpaired two-sided t test, n = 4). i. UMAPs and relative composition of immune cells in the lung of WT and Tigit-KO mice after LCMV Cl13 infection.

Source data

Extended Data Fig. 2 Transcriptional profiling of WT and Tigit-KO T cells.

a. Experimental scheme for sample preparation for scRNA- and scTCR-seq (Created in BioRender. Joller, N. (2025) https://BioRender.com/pprb5wp). b. UMAP plots showing log-normalized transcript levels of marker genes used for determining the T cell subset in the dataset. UMAPs are split by condition (naïve and LCMV Cl13). c. DotPlot showing standardized gene expression values of genes used for cell type and cluster annotation. d. Barplots showing cell numbers captured in each sample tabulated by cell type in each tissue of the scRNA-seq dataset. Each point shows the cell number of each 10x run, and the barplot shows the median of the cell type in the denoted condition. Within cell types, samples are grouped by genotype (WT and Tigit-KO) and naïve or LCMV Cl13 (technical replicates: WT spleen naïve n = 4; WT spleen LCMV n = 6; WT lung naïve n = 3; WT lung LCMV n = 4; KO spleen naïve n = 5; KO spleen LCMV n = 6; KO lung naïve n = 2; KO lung LCMV n = 4). e. DotPlot showing standardized gene expression values of the repair signature genes and Tigit.

Extended Data Fig. 3 TIGIT and Areg are coexpressed.

a. Areg frequencies in lung Treg cells, conventional CD4⁺ T cells, CD8⁺ T cells after LCMV Cl13, LCMV WE infection and in naïve controls (mean ∓ SD; pool of 2-6 independent experiments, ordinary one-way ANOVA, Tukey’s test, Treg cells n = 7, 20 4; CD4⁺ Tconv n = 10, 14, 4; CD8⁺ n = 4, 4, 4). b. Representative FACS plots for TIGIT and Areg expression in lung T cell subtypes from naïve and LCMV Cl13 infected mice (d10). c.-d. TIGIT frequencies in splenic (c.) and lung (d.) Treg cells, conventional CD4⁺ T cells, CD8⁺ T cells after LCMV Cl13, LCMV WE infection and in naïve controls (mean ∓ SD; pool of 5-10 independent experiments, ordinary one-way ANOVA, Tukey’s test; c. Treg cells n = 16, 32, 6; CD4⁺ Tconv n = 11, 16, 8; CD8⁺ n = 10, 11, 8; d. Treg cells n = 11, 16, 12; CD4⁺ Tconv n = 15, 16, 12; CD8⁺ n = 13, 16, 12). e. Correlation plots for Areg and TIGIT in splenic Treg cells, conventional CD4⁺ T cells and CD8⁺ T cells after LCMV Cl13 infection (pool of 7 independent experiments, simple linear regression and Pearson r correlation; Treg cells p value = 0.0027, CD4⁺ Tconv p value = 0.0029, CD8⁺ p value = <0.0001).

Source data

Extended Data Fig. 4 Areg expression is TIGIT-dependent.

a.-b. Areg in (a.) naïve spleen Treg cells and (b.) LCMV lung T cells (WT and Tigit KO, mean ∓ SD; pool of (a.) 5 and (b.) 2 independent experiments, unpaired two-sided t test). c.-d. Representative plots (c.) and summary (d.) of Areg in lung Treg cells upon influenza infection (WT and Tigit-KO, mean ∓ SD; pool of 3 independent experiments, 2way ANOVA, Šídák’s test). e. TIGIT in splenic T cells (WT and Tigit^ΔCD8, mean ∓ SD; pool of 3 independent experiments, 2way ANOVA, Tukey’s test). f. Evans Blue in WT and Tigit^ΔCD8 mice after LCMV (mean ∓ SD; pool of 2 independent experiments, unpaired two-sided t test). g. Tigit and repair signature gene expression in splenic CD8⁺ T cells after LCMV (mean ∓ SD; pool of 2 independent experiments, unpaired two-sided t test). h. Foxp3 MFI and frequencies of indicated markers on splenic Treg cells after LCMV (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). i. TIGIT in splenic T cells from Foxp3^GFPcreERT2xTigit^fl/fl (Tigit^ΔTreg) mice (+tamoxifen (tam) or vehicle (-tam)) (mean ∓ SD; pool of 4 independent experiments, 2way ANOVA, Tukey’s test). j. Tigit in naïve splenic Foxp3⁻ T cells of Tigit^ΔTreg mice (+tam or -tam) (mean ∓ SD; 1 representative experiment, unpaired two-sided t test, n = 6, 9). k. Areg expression in naïve splenic T cells from Tigit^ΔTreg mice (+tam or -tam) (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). l. TIGIT in splenic WT and Tigit-KO Treg cells from heterozygous Tigit^ΔTreg females (+tam) (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). m. Evans Blue in LCMV infected Tigit^ΔTreg mice (+tam or -tam) (mean ∓ SD; pool of 2 independent experiments, unpaired two-sided t test).

Source data

Extended Data Fig. 5 CD155 and CD226 expression dynamics.

a.-b. WT and Tigit-KO mice infected with LCMV Cl13 were treated with anti-CD226 blocking Ab (or isotype control) and analyzed for expression of TIGIT in Foxp3⁺ Treg cells (a.) and CD155 (b.) in T cells (mean ∓ SD; 1 representative experiment, 2way ANOVA, Tukey’s test, n = 4 for KO aCD226, 5 for all others). c.-d. CD155 frequencies in splenic Foxp3⁺ T cells from naïve and LCMV infected WT and Tigit-KO mice (c.) or Foxp3^GFPcreERT2xTigit^fl/fl (Tigit^ΔTreg) mice treated with tamoxifen or with vehicle as control (d.) (mean + SD; c. 1 representative experiment, naïve n = 4, 2; LCMV n = 9, 5 and d. pool of 2 independent experiments, 2way ANOVA, Tukey’s test). e. Representative FACS plot of myeloid cell gating, pre-gated on CD3⁻ live cells. f.-g. CD155 frequencies in CD11b^hiCD11c⁻, CD11b⁺CD11c⁺ and CD11b⁻CD11c⁺ myeloid cells isolated from spleens of naïve and LCMV infected WT and Tigit-KO mice (f.) or Foxp3^GFPcreERT2xTigit^fl/fl (Tigit^ΔTreg) mice treated with tamoxifen or with vehicle as control (g.) (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). h.-i. CD226 frequencies in T cells isolated from spleens of naïve and LCMV infected WT and Tigit-KO mice (h.) or Foxp3^GFPcreERT2xTigit^fl/fl (Tigit^ΔTreg) mice treated with tamoxifen or with vehicle as control (i.) (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). CD155 (j.) and TIGIT (k.) expression in splenic T cells isolated from LCMV infected WT and Cd155-KO mice (mean ∓ SD; pool of 3 independent experiments, unpaired two-sided t test).

Source data

Extended Data Fig. 6 TCR repertoire analysis.

scRNA/TCR-seq data from lung T cells of naïve and LCMV Cl13 infected WT and Tigit-KO mice. Cells from spleen (a.-d., i.-j., m., o.) and lung (e.-h., k.-l., n.) were analyzed separately. a.-c., e.-g. UMAP based on total gene expression annotated by cluster (a., e.), transcript levels of marker genes (b., f.) and indicating clonal expansion (c., g.). d., h. Proportion of expanded cells and condition distribution by transcriptional clusters. i., k. Clonal proportion of expanded vs non-expanded cells based on VDJ information across individual mice and (j., l.) corresponding expansion clonal rank distributions for repertoires. m. Representative expansion plots for one individual mouse per condition in spleen for the 50 most expanded clones per mouse colored based on T cell subset. n., o. Clonal expansion donut plots for Foxp3⁺ Treg cells, conventional CD4⁺ T cells and CD8⁺ T cells for one representative mouse per group. p. Frequencies of clonal expansion in WT and Tigit-KO mice in LCMV infected spleen (mean ∓ SD; n = 5, unpaired two-sided t test).

Source data

Extended Data Fig. 7 T cell clonality and comparison to classical tissue Treg cells.

a., b. scRNA/TCR-seq data from lung T cells of naïve and LCMV Cl13 infected WT and Tigit-KO mice. Clonal overlap of the VDJ sequences across all clones (a.) and for the 50 most expanded clones (b.) per mouse sorted by T cell subset. c. Integrated RNA-seq data from LCMV infection and published naïve ST2⁺ tissue Treg dataset¹⁷. UMAPs of indicated organs and datasets and cluster composition are shown. d. Representative FACS plots of ST2 and TIGIT expression in splenic Treg cells from naïve and LCMV Cl13 infected mice. e.-f. Frequencies of TIGIT⁺Areg⁺ Treg cells among Klrg1 and Gata3 SP, DP and DN Foxp3⁺ T cells (e.) or among Helios and Nrp1 SP, DP and DN Foxp3⁺ T cells (f.) (mean ∓ SD; one representative experiment, n = 3).

Source data

Extended Data Fig. 8 TCR-dependent stimulation of TIGIT⁺ Treg cells.

a., d., f. Areg frequencies in splenic conventional CD4⁺ T cells and CD8⁺ T cells (a.), Treg cells (d.) and CD8⁺ T cells (f.) stimulated with titrated amounts of αCD3 (a., f.) or αCD28 (d.) in vitro. Comparison between cells from naïve and LCMV Cl13 infected mice (a., d.) (mean ∓ SD; pool of 2 independent experiments, paired two-sided t test) and (f.) LCMV infected WT and CD8^crexTigit^fl/fl (Tigit^ΔCD8) mice (mean ∓ SD; 1 representative experiment, paired t test). b.-c. Frequencies of LCMV tetramer-positive total (left) or TIGIT⁺Areg⁺ (right) Foxp3⁺ T cells (b.) and CD4⁺ conventional T cells (c.) in naïve and LCMV Cl13 infected mice (pool of 2 independent experiments). e. Areg frequencies in total naïve (black) or LCMV infected (grey) splenic Treg cells stimulated with αCD3, αCD28 and cytokines in vitro (mean ∓ SD; 3 pooled experiments, 2way ANOVA, Tukey’s test). g. SCENIC analysis of scRNA-seq of T cells from naïve or LCMV Cl13 infected WT and Tigit-KO mice. Heatmaps show differential regulon activity between WT and Tigit-KO T cells in promoter (left) and enhancer regions (right). h.-i. Representative FACS (h.) and summary plots (i.) of TIGIT and JunB expression on Treg cells from naïve and LCMV infected mice (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). j.-k. pAKT (j.) and pS6 (k.) frequencies in Treg cells from LCMV infected mice treated in vitro with αCD3 in combination with TIGIT agonistic or isotype control Ab (pool of 2 independent experiments, paired t test).

Source data

Extended Data Fig. 9 Hif1α regulates Areg expression in a TIGIT-independent manner.

a. Scatterplots showing the expression pattern of Hif1a and Areg (two-sided p-value for the R ≠ 0 null hypothesis – p-value of pearson correlation, no multiple testing correction was applied). b., c. Representative FACS plots for TIGIT and Areg expression in splenic Treg cells, CD4⁺ Tconv cells and CD8⁺ T cells from LCMV infected Cd4^crexHif1a^fl/fl mice and controls and c. quantification for expression in Treg cells (mean + SD; pool of 2 independent experiments, unpaired two-sided t test). d. Scatterplots showing the expression pattern of Hif1a and Tigit. e. Relative Hif1a mRNA expression in FACS sorted splenic Treg cells, CD4⁺ Tconv cells and CD8⁺ T cells from naïve and LCMV infected WT and Tigit-KO mice (mean ∓ SD; pool of 3 independent experiments, 2way ANOVA, Tukey’s test). f., g. Relative Tigit mRNA expression in FACS sorted splenic TIGIT⁺ and TIGIT⁻ (f.) and WT vs Tigit-KO (g.) Treg cells, CD4⁺ Tconv cells and CD8⁺ T cells from LCMV infected mice (mean ∓ SD; pool of 2 and 3, respectively, independent experiments, 2way ANOVA, Tukey’s test). h., i. Representative FACS (h.) and summary plots (i.) of TIGIT and Blimp-1 expression on Treg cells from control or influenza infected WT and Tigit-KO mice (mean ∓ SD; pool of 2 independent experiments, 2way ANOVA, Tukey’s test). j.-k. Blimp-1 frequencies in splenic Treg cells from LCMV infected (j.) WT or Cd155-KO mice and (k.) WT or Tigit-KO mice treated with anti-CD226 blocking Ab or isotype (mean ∓ SD; j. pool of 3 independent experiments, unpaired two-sided t test; k. one representative experiment, 2way ANOVA, Tukey’s test).

Source data

Extended Data Fig. 10 Gating strategies.

a. Gating strategy of lymphoid and myeloid cells used for immunophenotyping. b. Gating strategy of T lymphocytes used throughout all the experiments. c. Gating strategy of T lymphocytes used for T cell sorting for the scRNA/scTCR-seq experiment.

Supplementary information