Background: While results of the TOPAZ-1 trial led to inclusion of anti-PD-1 therapy with standard of care gemcitabine/cisplatin chemotherapy for patients with unresectable biliary tract cancer (BTC), the improvement in overall survival is modest: 12.8 vs 11.5 mos. Identification of novel immunotherapeutic strategies remains a high priority. We discovered Galectin-9 (Gal-9), a carbohydrate-binding protein, is expressed in BTC; it has various immunosuppressive effects, including induction of apoptosis in T cells upon binding to its receptor Tim-3. Defining the role of Gal-9 in the BTC microenvironment will inform how to leverage it as a novel therapeutic target. Methods: Tissue microarrays (TMAs) were constructed from tumors of 66 patients who underwent curative-intent resection of BTC from 2000- 2015 at our institution. These TMAs underwent immunohistochemical staining for CD4, CD8, Gal-9 & Tim-3; the percentage of cells staining positive for each marker was quantified. For each marker, martingale residual plots and bias-adjusted log rank tests were used to identify a cut-off value of “low” versus “high” expression where survival difference was maximal. For in vivo mouse experiments, the syngeneic BTC cell line URCCA4.3 was injected subcutaneously into the flank of C57BL/6 mice, followed by 3 weeks of treatment with anti-Galectin-9 or anti-Tim-3 antibody alone or in combination with anti-PD-1 antibody. Tumor-infiltrating lymphocytes (TILs) were analyzed by flow cytometry. Results: Median overall survival (mOS) was significantly improved in patients with high Gal-9 expression in their tumors, compared to those with low (32.5 mos vs 17.6 mos, p=0.018). However, when patients with high Gal-9 expression were further stratified according to Tim-3, patients with both high Gal-9 & high Tim-3 expression had significantly worse mOS than those with low Tim-3 (22.7 mos vs 113.9 mos, p=0.031), implying that interaction between Gal-9 and Tim-3 wields significant impact on TIL function. There was no difference in OS based on Tim-3 expression alone. When mice bearing BTC tumors were treated with Gal-9 or Tim-3 antibody monotherapy, there was no significant change in tumor growth versus control, but when either was used in combination with PD-1 inhibition, there was a significant reduction in tumor burden at study endpoint. Mice treated with either dual therapy had a significant decrease in CD4⁺ T_reg and increase in CD4⁺ effector memory TIL populations compared to monotherapy (p<0.05). Conclusions: This study provides unique insight into the role for targeting the Gal-9/Tim-3 pathway in BTC. Our clinical data highlights that co-expression of these markers in BTC tumors is associated with significantly worse OS. Our in vivo pre-clinical studies demonstrate that blockade of either Gal-9 or Tim-3 synergizes with PD-1 targeted antibodies for superior anti-tumor efficacy.