There is nothing like hearing science told by passionate and talented investigators, and we got a one-two punch on the subject of TIM-3-mediated immunotherapy from Drs Vijay Kuchroo and Ana Carrizosa Anderson at the ImVacS Immunomodulatory Antibodies for Cancer meeting in Boston on Monday, August 11. The image above is modified from this review, written by the speakers.
Having sent a few years tackling this challenging protein family it was refreshing to see the utility of anti-TIM-3 therapy being supported by robust experimentation. This is a complex story and one that’s hard to get right, but I think these investigators from Harvard Medical School are getting very close.
Let’s have a look.
Vijay Kuchroo introduced several unifying hypotheses – first, that TIM-3 is always co-expressed on T cells along with PD-1, thereby marking such T cells as anergized or exhausted (in any case, non-responsive to TCR-stimulation), and second, that TIM-3 functions as a cell death signal for T cells. The first hypothesis is supported by the literature, and likely portrays the consensus view. The second hypothesis was supported by experiments in which isolated T cells were exposed to soluble Galectin-9 (Gal-9) whereupon these T cells “blew apart” as shown in a short video. This latter hypothesis is apt to be more controversial, as the experimental conditions used are probably not physiological, and the described effect is not anticipated in the Gal-9 literature. Regardless, it was impressive data. It is possible that there is some interplay between TIM-3 binding to Gal-9 and TIM-3 binding to phosphatidylserine (PS) that regulates T cell responses. That’s not been worked out yet. Gordon Freeman later offered the additional hypothesis that Gal-9 serves to attenuate or modulate TIM-3 signaling, perhaps in cis (on the same cell) or trans (across cells).
Dr Kuchroo then presented some very interesting new data that tumor stromal cells produce IL-27 that induces T cell suppression and supports T regulatory cells (Treg). IL-27 therefore emerges as a novel regulator of the tumor Treg environment. The “exhaustion module” was further articulated as follows: the tumor stroma secretes IL-27 that signals through the IL-27 receptor to PRDM1. PRDM1 induces the upregulation of TIM-3, LAG-3, PD-1 and IL-10. Additional new targets were identified using this model, and we can look forward to seeing those described in future – that was very nice work. Conversely, IFNy is upregulated when PD-1 and TIM-3 are blocked, e.g. with antibodies. Since interferons, including IFNy, are known to upregulate PD-L1 (the PD-1 ligand) on tumor cells, a particularly attractive therapy might be a combination of anti-TIM-3 and anti-PD-L1 antibodies. Of note, tumor growth and the development of Tregs is deficient in IL-27R gene-deficient (KO) mice, supporting the exhaustion module model. The inhibition of tumor growth in these KO mice is correlated with increased expression of IL-2, TNF and IFNy in the tumor-draining lymph nodes.
One question was raised which addressed how this model could be reconciled with the finding by Steven Rosenberg and others that PD-1 expression marks activated tumor infiltrating lymphocytes (TILs). The difference may be related to the co-expression, or lack of expression, of TIM-3 and perhaps LAG-3. It would appear that we need a way to characterize the diverse variety of T cells found in tumors, beyond the current TILs and Tregs.
Ana Carrizosa Anderson further discussed TIM-3 expression on TILs and showed that TIM-3+ TILs secrete IL-10 and are therefore Tregs. Such TIM-3+ Tregs are preferentially enriched in tumors, pointing to the active induction as described by Dr Kuchroo. Dr Anderson further showed that anti-PD-L1 plus anti-TIM-3 antibodies deprogram Tregs, shutting down their immune suppressive pathways including their own target’s expression – thus anti-TIM-3/anti-PD-1 treatment shuts down expression of TIM-3, PD-1 and a host of other immunosuppressive factors, shining a light on a mechanism of feedback regulation in human TILs.
Dr Anderson then raised an interesting question regarding colorectal cancer which has so far proved resistant to immunotherapy (ipilimumab, nivolumab). In a preclinical CRC model, the combination anti-TIM3 and anti-PD-1 worked very well, This led to the intriguing question of whether TIM-3 is particularly useful here because of the very high expression of Gal9 is high in gut. I like the focus on the local environment, which may be useful in addressing primary tumors.
Speaking of the local environment, this is a good time to mention a paper on TIM-4, another PS-binding protein in the TIM family. Terry Strom along with Vijay Kuchroo and colleagues recently described a “fragile” TIM-4+ tissue macrophage population with immunosuppressive properties (Thornley et al. 2014). While the paper focuses on immunological models, one can imagine that this target might be further investigated in the immunotherapy setting.
Seemed to me there was a fair amount of unpublished data in these talks that we can look forward to seeing published sometime soon, no doubt. In the meantime I’d bet some invention disclosures and preliminary patent filings are in the works!