The generation of TIM-3 specific monoclonal antibodies has allowed us to probe TIM-3 expression and function on human CD4+ T cells in a more thorough fashion than was previously possible. While TIM-3 has been well studied and characterized in mice, its role in human immune responses is less clear. In the course of our studies we have found similarities and differences between human and murine TIM-3. In agreement with our observations of murine TIM-3, virtually no CD4+ T cells present in the peripheral blood of healthy human subjects express surface TIM-3. In contrast to mouse T cells, which express TIM-3 only after several rounds of Th1 polarization, the majority of in vitro
activated human CD4+ cells become TIM-3+ after a few days of stimulation with anti-CD3/28. Thus it appears TIM-3 is expressed by cells that have been stimulated strongly via TCR plus costimulation through signal 2 (CD28). In support of this we observed TIM-3 expression was correlated to dose of anti-CD3 (data not shown). It is possible that TIM-3 is expressed mainly on activated T cells present in sites other than peripheral blood, such as lymph notes, spleen, skin, or mucosa. In support of this, we have observed TIM-3 expression on activated (CD25+) populations of CD4+ T cells isolated from pancreatic lymph nodes (). This is in agreement with our observations of TIM-3 expressed by mouse T cells in the CNS of mice in the early stages of EAE [11
We observed an apparent paradox in that TIM-3 does not appear to be correlated with cytokine production in CD4+ T cells examined directly ex vivo (), whereas TIM-3 is coexpressed with IFNγ and IL-17 by naïve T cells stimulated in vitro with anti-CD3/28 +/- polarizing cytokines for 7 days (). We believe this is consistent with the hypothesis that TIM-3 is a negative regulator of T cells that have previously undergone activation. Thus ex vivo TIM-3+ cells will be resistant to reactivation, perhaps through activation induced cell death or anergy, and will not proliferate or produce cytokines. Stimulation/polarization of naïve, TIM-3-negative cells in vitro, on the other hand, results in activation, secretion of cytokines, and subsequent expression of TIM-3, which will then render them resistant to further stimulation.
Administration of TIM-3-Ig has been shown to enhance Th1 responses in vivo
and prevent peripheral T cell tolerance [9
]. Extending these observations, we have recently identified Galectin-9 as a ligand of TIM-3, and demonstrated that Galectin-9 induces apoptosis in Tim-3+ Th1 cells but not Th2 cells [8
]. Thus, engagement of Tim-3 on murine Th1 cells negatively regulates Th1-derived IFN-γ secretion by inducing apoptosis in Th1 cells. Our data now show that activated human CD4+ T cells express TIM-3 and Galectin-9, and that addition of TIM-3 antibodies during T cell activation augments secretion of IFNγ, IL-17, IL-6, and IL-2, but not IL-4, IL-10 or TNFα. Thus it appears that human TIM-3 regulates the secretion of specific cytokines (Th1 and Th17 to some extent) but does not does not have a global effect on cytokine production.
Since treatment of murine T cells with Galectin-9 results in reduced production of IFNγ via TIM-3-regulated cell death [8
], we hypothesize that the TIM-3 monoclonal antibodies we describe herein are antagonistic, and are blocking a negative signal leading to enhanced production of cytokines. To clarify this, we attempted blocking experiments with activated CD4+ T cells using TIM-3-Ig, but found no enhancement of cytokines, proliferation, or cell death (not shown). TIM-3 has been shown to be a positive regulator of antigen presenting cells [26
], so we cannot fully rule out the possibility that the TIM-3 antibodies are agonistic, and thus trigger TIM-3 on CD4+ T cells resulting in enhanced cytokine secretion. We believe that this is unlikely, since extensive experimental evidence from multiple studies suggests that TIM-3 acts as a negative, rather than positive, regulator of CD4+ T cells [8
]. Indeed, we have utilized 2E2 to block Galectin-9-induced cytokine secretion by human dendritic cells [24
]. Also, knockdown of TIM-3 in activated CD4+ T cells using siRNA leads to increased IFNγ production [28
], in agreement with our antibody results. Thus it appears that the TIM-3 antibodies block the interaction of TIM-3 with either Galectin-9, or other unknown ligand(s), which results in a release from negative regulation and consequent increase in cytokine production.
Interestingly we found neither protection from, nor induction of, cell death in CD4+ T cells stimulated in the presence of TIM-3 antibodies. In addition, we found that mRNAs of IFNγ, IL-6, and IL-17A were increased in the presence of TIM-3 antibodies, suggesting human TIM-3 negatively regulates cytokine production at the level of transcription, rather than in a cell-death-dependent manner. However, we did not find major differences in expression of T-bet, GATA3, and RORγC, the key transcription factors involved in Th1, Th2, and Th17 differentiation. This is perhaps not surprising as we used total CD4 cells that include naïve and central memory T cells for our functional studies. Thus these data likely reflect memory T cell responses after 48 hours of stimulation as opposed to differentiation of naïve T cells.
In our in vitro system it appears that TIM-3 functions via T-T cell interactions in that TIM-3 may be binding to a ligand expressed on T cells themselves. We depleted regulatory T cells from the total CD4+ population by FACS sorting and observed no loss of cytokine induction resulting from TIM-3 antibody treatment of activated cells (data not shown), suggesting Tregs are not involved in TIM-3-mediated inhibition of cytokines in vitro. We hypothesize that TIM-3 expressed on activated T cells interacts with Galectin-9 or other unknown ligand(s) in either an autocrine or paracrine fashion to inhibit cytokine production. This negative feedback loop may serve as a switch to “turn off” IFNγ and Th1/Th17 responses that could mediate immunopathology if left unchecked.
Collectively, these data suggest TIM-3 is an important regulator of Th1 and Th17 cytokines in human CD4+ T cells. It will be of great interest to use these antibodies in the context of many human inflammatory diseases, where the frequency of TIM-3-expressing T cells can be analyzed and their function in the context of human disease further evaluated.