The TR1 cell subset has been implicated in the maintenance of peripheral tolerance against immune-mediated pathologies. TR1-like cells were first documented in severe combined immunodeficiency (SCID) patients that did not develop graft-versus-host disease (GvHD) after receiving HLA-mismatched fetal liver hematopoietic stem cell transplants (1). Subsequent work by Groux et al. using antigen-activated CD4⁺ T cells cultured in the presence of IL-10 led to the identification of a distinct T cell subset, thereafter named TR1, that could prevent the development of experimental colitis in an IL-10- and transforming growth factor beta (TGFβ)-dependent manner (2).

Unfortunately, the paucity of information on TR1 cell-specific surface markers or transcription factors have hampered the execution of detailed studies on the role and function of this T cell subset in the maintenance or breakdown of self tolerance. The use of relatively non-specific markers of TR1 cell identity, leading to the implication of this subset in various immunological processes (i.e., sometimes relying exclusively on IL-10 expression), has muddied progress in this area. Fortunately, the last decade has witnessed the discovery of phenotypic and molecular features of ‘TR1-ness’ that have allowed a better definition of TR1-like cells in various experimental settings. These developments, coupled to recent methodological developments in in vitro TR1 cell generation (3–7), and the discovery of pharmacological approaches capable of eliciting the formation and expansion of antigen-specific TR1 cells in vivo (8), have exposed TR1 cells as attractive targets for therapeutic intervention in immune-mediated diseases.

Despite this progress, significant knowledge gaps remain, including a detailed understanding of the developmental biology processes responsible for the genesis of this T cell subset in vivo. The use of novel technologies, including mass cytometry and next-generation-sequencing to address these various gaps are beginning to shed light into these areas of scientific inquiry. In this review, we summarize current knowledge on the phenotypic and molecular hallmarks of TR1 cells and key developmental processes underlying TR1 cell genesis, including recent evidence pointing towards T follicular helper (Tfh) cells as TR1 cell precursors (9, 10).

TR1 cells were initially described as CD4⁺ T cells producing high levels of IL-10 and IL-5, intermediate levels of TGFβ and INFγ and low levels of IL-4 and IL-2, and were capable of suppressing specific immune responses in vitro and in vivo, in an IL-10-dependent manner (2). With rapid IL-10 production kinetics, detectable even 4 hours post-activation, and a peak of production at 24h (11), IL-10 became the hallmark cytokine for the TR1 population and, together with the absence of FoxP3 expression, used to identify TR1 cells in early studies. We now know that these criteria are insufficient, given that other CD4⁺ T cell types such as Th1 (12, 13) or Th2 (14) can produce IL-10 and acquire immunoregulatory properties; such cells do not belong to the TR1 subset. For example, Lönnberg et al. claimed a Th1 origin for TR1 cells in a chronic Plasmodium infection model, solely on the basis of presence of IL-10-expressing cells within the infection-induced Th1 pool, and on the assumption that TR1 cells are simply IL-10/IFNγ-co-expressing cells (15). In fact, further transcriptomic analyses of the IL-10⁺ and IL-10^– Th1 cells of these mice revealed the presence of only two differentially expressed genes between these subsets (Trib2 and BC017643). In another study, also in a chronic Plasmodium infection model, Soon et al. reported a similar outcome; 34% of Th1 lineage cells co-expressed Ifng and Il10 (16). It is thus likely that, based on the evidence provided, the IL-10⁺ cells that arose in these mice were IL-10-expressing Th1 cells, rather than true TR1 cells. This indicates that the assignment of TR1ness cannot merely rely on IL-10 expression.

In light of these challenges, hampering progress in defining the significance of the TR1 cell subset in both physiology and pathology, extensive efforts were made to better describe the molecular hallmarks of the TR1 subset [Reviewed in (17)]. Notwithstanding the fact that markers strictly unique to TR1 cells remain elusive, recent advances have made it possible to more accurately identify such cells in biological samples.

Gagliani and colleagues identified co-expression of CD49b and Lymphocyte-activation gene 3 (LAG-3) as surface markers for both human and murine IL-10-producing TR1-like cell populations (18). Subsequent studies indicated that a significant fraction of these IL-10-producing CD49b⁺LAG-3⁺ T cells are co-inhibitory receptor-rich, expressing Programmed cell death-1 (PD-1), T cell immunoreceptor with Ig and ITIM domains (TIGIT), T-cell immunoglobulin and mucin-domain containing protein-3 (TIM-3), and Cytotoxic T-Lymphocyte antigen 4 (CTLA-4), and co-express the co-stimulatory molecule ICOS (Inducible Costimulator) and the chemokine receptor CCR5, among other molecules (7). In agreement with these observations, intestinal TR1-like cells expressing PD-1 and CCR5 were found to co-express CD49b and LAG-3 by others (19), thus supporting the use of such markers for TR1-like cell identification. Thus, as proposed elsewhere (20), TR1 cell annotation should meet the following four criteria: 1) high IL-10 production competency (co-expression of other cytokines in variable amounts depending of environmental cues is possible); 2) immunoregulatory activity; 3) absence of constitutive FoxP3 expression (expression of FoxP3 upon activation, particularly in human TR1 cells, is not an exclusion criterium); and 4) co-expression of CD49b and LAG-3 in the presence of other co-inhibitory receptors such as PD-1, TIGIT, TIM-3 or CTLA-4 among others.

Indeed, the profoundly immunoregulatory antigen-specific TR1-like cells that arise in vivo in response to systemic delivery of nanoparticles (NPs) coated with mono-specific disease-relevant peptide-Major Histocompatibility Complex class II (pMHCII) molecules (8, 21–23) lack FoxP3 expression and upregulate many of the markers mentioned above, including CD49b and the co-inhibitory receptors LAG-3, PD-1, TIGIT and CTLA-4, the co-stimulator ICOS, the cytokines IL-10, IL-21 and IFNγ and the chemokine receptors CCR5 and CXCR3 (9). We have shown that administration of these compounds can lead to the resolution of inflammation in various organ-specific autoimmune disease models in a disease-specific manner without impairing normal immune responses (8, 21, 22). Adoptive transfer experiments demonstrated that the cognate (pMHCII tetramer⁺) T cells arising in these mice in response to therapy were largely, albeit not exclusively, responsible for the therapeutic properties of these compounds (TR1 cell-induced B-regulatory cells also contributed to disease suppression) (8, 22).

TR1 cells need to be activated to immunoregulate. Upon recognition of their cognate pMHCII complexes on co-stimulation-competent APCs, TR1 cells become productively activated. By actively inhibiting the antigen-presentation and pro-inflammatory properties of these APCs (in addition to direct effects on other target cell types, see below), TR1 cells can suppress both cognate and non-cognate effector T cell activation (a process referred to as ‘bystander immunoregulation’). This biological activity involves the deployment of several mechanisms ( Figure 1 ).

Extensive efforts over the last two decades have sought to define the molecular and transcriptional mechanisms orchestrating TR1 formation. Unlike the case for the FoxP3+ Treg cell subset, where expression of FoxP3 is central to the acquisition of its immunoregulatory properties, there is no known unique master transcriptional regulator of TR1 cell development. Notwithstanding this limitation, experimental evidence has implicated a number of cytokines, kinases and transcription factors in the generation of TR1 cells in vitro. Although TCR engagement in the presence of IL-10 appears to play a major role, other signals are also required.

There is evidence suggesting that superantigens (59, 60) and pMHCII multimers can induce the expression of IL-10 in CD4+ T cells (61–63). It has also been shown that high-avidity TCR–pMHCII interactions favour the production of IL-10 by T cells (64), affecting both the number of cells expressing IL-10 and the immunoregulatory properties of such cells (65). Molecularly, TCR activation leads to the engagement of intracellular signalling pathways that eventually activate the interferon regulatory factor 4 (IRF4) transcription factor via Ras or the inducible tyrosine kinase (ITK) kinases. IRF4 has been shown to promote Il10 gene expression in different CD4+ T cell types, including Th2, Th1 (66), Tfh cells (67) and Tregs (68). Indeed, it has been reported that IRF4 contributes to the development of an IL-10-producing CD4+ T cell that co-expresses LAG-3 and CD49b (69). As noted below, IRF4 is absolutely required for pMHCII-NP-induced TR1 cell formation, albeit through a different mechanism (i.e., it is dissociated from its Il10 transactivating function) (9). The transcription factor Eomes, which can promote Il10 expression in T-bet-expressing cells (70), and the Th17 transcription factor Rorα, which can transactivate the Il10 gene (71), might also be implicated in TR1 formation. However, pMHCII-NP-induced TR1 cells do not upregulate Eomes or Rorα, suggesting that neither of these transcription factors are required for TR1 cell specification.

Although productive TCR ligation is required for TR1 activation, TR1 cell genesis requires additional cues. Early studies by Groux and colleagues using both human and murine CD4+ T cells cultured in presence of IL-10 indicated that these culture conditions promoted the development of an anergic T cell population that included TR1-like cells (2), highlighting a prominent role for this cytokine in TR1 cell generation, at least in vitro. It was subsequently proposed that the IL-10 that contributes to TR1 cell generation in vivo derives from a tolerogenic DC population (72, 73). Indeed, a human DC population expressing high amounts of IL-10 has been identified (27). This DC population, named DC-10, can induce TR1 cells ex vivo with increased efficacy, as compared to other experimental approaches, and such cells have been used to generate and expand TR1 cells for use in clinical trials [Reviewed in (5)].

IL-27 has also been implicated in the generation of murine TR1 cells. IL-27 is an IL-12 family cytokine that is produced by activated APCs (74) and can induce IL-10 expression in murine T cells (75–77), especially in the presence of TGFβ, with which it synergizes (78). Binding of IL-27 to the IL-27 receptor (IL-27R) activates the STAT1 and STAT3 signalling pathways and promotes the expression of the transcription factors c-Maf and AhR, which cooperatively promote Il10 and Il21 expression (78, 79). STAT3-induced Il10 expression also involves the upregulation of Egr2 (encoding the Early Growth Response 2 transcription factor (EGR-2)) and EGR-2’s downstream target Prdm1 (encoding the zinc finger-containing transcription factor Blimp-1) (80). Although Blimp-1 has been primarily implicated in plasma cell differentiation (81), it has also been shown to regulate Il10 gene expression in T cells (82, 83). However, and notwithstanding the fact that the generation of terminally differentiated TR1 cells in response to pMHCII-NPs requires Blimp-1, this role is dissociated from Blimp-1’s Il10 transactivating function (9) (see below). In fact, there is evidence suggesting that IL-27-induced TR1 cell formation does not require IL-27-induced IL-10 (84) and that IL-27 contributes to TR1 cell formation by inducing changes in chromatin accessibility via IRF1 and BAFT (85). Although IL-27 can induce the formation of IL-10-expressing T cells from naïve human CD4+ precursors (86), it remains to be determined whether these cells are bona fide TR1 cells.

Despite all these observations, largely if not exclusively generated in vitro, our in vivo work has demonstrated that IL-27 is not required for pMHCII-NP-induction of TR1-like cells (8). We have proposed that IL-27 and pMHCII-NPs lie upstream and downstream of the TR1 precursors (Tfh cells, see further below); whereas IL-27 would elicit both Tfh and TR1 cell formation from naïve precursors, pMHCII-NPs would just be able to promote the conversion of Tfh cells into TR1-like cells (9).

IL-21 is another cytokine that has been implicated in TR1 cell genesis. This cytokine, produced by antigen-stimulated CD4+ T cells and NKT cells, signals via the IL-21R, composed of the IL-21Rα chain and the common receptor γ_c chain (87), leading to activation of the STAT3 signalling pathway. The transcription factor c-Maf, upregulated by IL-27 among other stimuli, promotes IL-21 expression in TR1 cells (79, 88). In turn, IL-21 promotes the expression of Il10 and cMaf expression in an autocrine manner (79, 89).

In addition to the molecules discussed above, other cytokines and transcription factors have been reported to contribute to IL-10 production by T cells and, consequently, may play a role in TR1 cell specification. For instance, IL-6, which signals through STAT1 and STAT3, can upregulate the transcription factors c-Maf, IRF4 or AhR (90), which are known to participate in Il21 and/or l10 expression in different T cell types (66–69, 78, 79). In fact, IL-6, together with TGFβ, can induce IL-10 production in Th17 cells (91, 92). Type-1 Interferons have also been reported to promote IL-10 expression in CD4+ T cells (93–95) or TR1 cell development in anti-CD3 mAb/IL-10-treated mice (96).

The co-stimulator ICOS may also play an important role in TR1 cell specification, homeostasis or function, perhaps by promoting cMaf and Il10/Il21 expression (88, 97).

In summary, research to date has provided valuable information regarding the transcriptional control of Il10, encoding the hallmark TR1 cytokine, but has not yet been able to define the key transcription factors, co-stimulators and cytokines that control TR1 cell development from their precursors. As summarized below, the recent identification of Tfh cells as precursors of TR1 cells in vivo, coupled with definition of the transcriptional changes that underlie this transdifferentiation process, offer a unique opportunity to carefully map the molecular events responsible for TR1 cell formation.

Given the challenges associated with the lack of TR1 cell-specific markers and our inability to reliably identify this T cell subset in vivo until recently, it is unclear whether the TR1-like cells that have been described to arise in vitro and/or in vivo in response to various cues do so from a single or various precursors [reviewed in (98, 99)].

As noted above, both human and mouse T cells can be differentiated into TR1-like cells in vitro. In vitro-activated naïve T cells from either species can give rise to anergized IL-10-producing CD4+ T cells when cultured in the presence of exogenous IL-10 (2), DC-10 cells (100) or IL-27 (79, 101).

Several other lines of evidence have suggested that TR1 cells arise from memory T cell precursors. Repetitive administration of anti-CD3 monoclonal antibodies (mAb) to mice can induce TR1-like cell formation in vivo (102). In vitro stimulation of memory-like CD4⁺CD44^highFoxP3^– T cells in the absence of polarizing cytokines can also elicit TR1-like cell specification (103). Likewise, extracellular matrix components have been reported to guide the formation of TR1-like cells from human memory CD4⁺ T cells in vitro (104), and others have shown that the precursors of TR1 cells are contained within the memory CD4⁺ T cell pool, in both humans and mice (105).

Th1 and Th2 cells have also been proposed as a source of TR1-like cells. In vitro, Th1 cells can be induced to express IL-10 when stimulated in the presence of CXCL12 (106), but such cells might have just been IL-10-producing Th1 cells rather than full-fledged TR1 cells. As noted earlier in this review, it has been suggested that chronic infection of mice with Plasmodium can trigger the differentiation of Th1 cells into TR1-like cells (15, 16), but the reported TR1-like cells did not appear to be bona fide TR1 cells. It has also been suggested that allergen-specific Th2 cells can be re-programmed into a TR1-like phenotype in vitro (107).

Intestinal Th17 cells can also give rise to anti-colitogenic TR1-like cells in response to anti-CD3 mAb treatment (108). Furthermore, IL-27 and IL-23 promote the up-regulation of Blimp-1 and can elicit the expression of a TR1-like phenotype in Th17 cells (109).

Collectively, the above observations suggest that TR1-like cells might arise from various T cell precursor types, but many of these studies did not use stringent criteria for definition of the TR1 cell state, or did not involve detailed transcriptional studies of the T cell pools used for experimentation or of their progeny; and when such studies were done/reported, the resulting T cell pools were transcriptionally heterogeneous. As a result, it is not possible to unambiguously assign or exclude a specific cell type as a TR1 cell precursor on the basis of these studies.

We have taken advantage of the large pools of antigen-specific TR1 cells that arise in vivo in various animal models of autoimmunity upon systemic delivery of nanoparticles (NPs) coated with disease-relevant pMHCII molecules (8, 21–23), to carefully map the transcriptional events leading to TR1 cell formation in vivo (9, 10) ( Figure 2 ). Early work established that pMHCII-NPs functioned by re-programming cognate antigen-experienced (i.e. memory) CD4⁺ T-cells of unknown identity (excluding a role for naïve T-cells) (8, 23).

The robust and prolongued TCR signaling events that result from sustained assembly of TCR microclusters by pMHCII-NPs on autoantigen-experienced T cells leads to the expression of known TR1-like cell markers, including IL-10, IL-21, c-Maf, LAG-3, CD49b, CTLA-4, PD-1, TIGIT, CCR5, CXCR3, ICOS and OX-40, among others, in a manner that does not require APCs or IL-27 (9). In addition to Maf, these TR1 cells upregulate the transcription factor coding genes Ahr, Egr2, Irf4, Nfil3, Prdm1 and Tbx21 (9), all involved in IL-10 expression (79, 110). In addition, these TR1 cells upregulate three other transcription factors that have been previously implicated in the development, maintenance or function of IL-10-expressing Treg cells (Bhlhe40, Runx2 and Vdr) (9). Whereas the IL-10 produced by these antigen-specific TR1 cells is the direct mediator of some of their immunoregulatory properties, IL-21 contributes to the homeostatic regulation of this T-cell subset and plays a critical role in TR1-induced Breg cell formation (8). This ability of pMHCII-NPs to elicit the formation of large pools of antigen-specific TR1-like cells afforded us a unique opportunity to explore their developmental biology. This work has demonstrated that pMHCII-NP-induced TR1 cells derive from cognate Tfh cells and do so in a Blimp-1-dependent manner (9, 10) ( Figure 2 ).

Initial work indicated that the cognate TR1-like cell pools arising in response to pMHCII-NPs expressed a transcriptional program that shared significant features with Tfh cells, raising the possibility that the latter might function as a source of the former. Subsequent single cell RNA sequencing (scRNAseq) and mass cytometry studies demonstrated that these antigen-specific TR1-like cell pools harboured a cognate Tfh-like cell subcluster, in addition to its TR1-like cell countepart. Importantly, studies of the TCR repertoires of these two cell sub-clusters indicated that they consistently harbored identical clonotypes, thus demonstrating that they were developmentally related (9, 10). This was substantiated with the use of different pMHCII-NP types in different genetic backgrounds and models of autoimmunity (9, 10).

This was further documented by demonstrating that pMHCII-NPs could elicit cognate TR1 cell formation in immunocompromised hosts transfused with purified Tfh cells, and that these compounds lacked pharmacodynamic activity in mice unable to generate Tfh cells (9). Most importantly, T cell-specific deletion of Prdm1 (encoding Blimp-1) revealed that the Tfh-to-TR1 cell conversion evolves through a transitional (TR1-like) subset, and that expression of this transcription factor in these transitional T cells is a sine qua non requirement for full-fledged acquisition of the TR1 transcriptional profile and regulatory function (9). Thus, while specific deletion of Bcl6 or Irf4 in T-cells blunted pMHCII-NP-induced cognate CD4+ T-cell expansion and downstream TR1 generation, deletion of Prdm1 enabled the former but completely abrogated the latter (9).

It could be argued that pMHCII-NP-induced TR1 cells are T-follicular regulatory (TFR) cells, which negatively regulate the germinal center (GC) reaction (111). However, unlike TR1 cells, TFR cells express CXCR5 (but not CCR5), Bcl-6, FoxP3 and CD25, and arise from natural FoxP3+ Treg cell precursors in a Blimp-1-independent manner (9).

Together, the data summarized above conclusively demonstrate that murine TR1 cells can arise from Tfh cells in a Blimp-1-dependent manner. Interestingly, a pool of thymus-derived self-reactive CD4+ T cells that adopt numerous hallmarks of Tfh cell identity in the periphery has been recently discovered (112). This finding raises the possibility that these cells might function as a source of a negative feedback regulatory loop (i.e., formation of autoreactive TR1 cells) to suppress autoimmunity.

Since the discovery of TR1-like cells more than 25 years ago, the last decade has witnessed steady improvements in our ability to identify and phenotype this previously enigmatic CD4+ T cell  subset. While knowledge gaps persist, we have gained detailed new insights into these cells’ transcriptional make-up, mechanisms of action and lineage origin. Further research into the different topics reviewed in this article, as well as other aspects of the TR1 cell biology will undoubtedly help in the translational application of TR1 cells as a therapeutic approach for immune-mediated diseases.

PS: Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing – review and editing. EA: Investigation, Writing – original draft.