Resistance to anti-PD-1/anti-PD-L1: GB1211 reverses galectin-3 induced blockade of pembrolizumab and atezolizumab binding to PD-1/PD-L1

2607 Background: Galectin-3 (Gal-3) is a β-galactoside binding lectin highly expressed across many cancers (Li et al. Biomedicines, 2021) with high expression associated with poor response to the PD-1 inhibitor pembrolizumab (Capalbo et al., Int J Mol Sci, 2019). In addition, it has also been shown preclinically in in vivo mouse non-small cell lung carcinoma (NSCLC) models that inhibition of Gal-3 enhances the performance of checkpoint inhibitors targeting the PD-1/PD-L1 axis (Vuong et al., Cancer Res, 2019; Zhang et al., FEBS Open Bio, 2020). The mechanism by which Gal-3 negatively impacts the effect of PD-1/PD-L1 based checkpoint inhibitors is not fully understood. We investigated the hypothesis that anti-PD1/anti-PDL1 is induced by direct binding and blockade of Gal-3 to the PD-1/PD-L1 complex itself. Methods: SPR was used to investigate the effect of recombinant hGal-3 in combination with a Gal-3 inhibitor (GB1211) on the binding of recombinant hPD-L1 to hPD-1 in the absence and presence of either pembrolizumab or atezolizumab. SPR analysis was performed with a BIAcore (Cytiva, USA). hPD-1 was immobilized onto a CM5 Series S sensorchip via amine coupling and proteins, antibodies or small molecules flowed over in various combinations. Observations made at the protein level were further confirmed in cell binding assays using flow cytometry measuring pembrolizumab and atezolizumab binding to cells lines over expressing high levels of either PD-1 (Jurkat cells) or PD-L1 (RAJI-hPD-L1 cells), respectively. Results: Gal-3 potentiated the binding of PD-L1 to PD-1 with a 3-fold increase in its KD observed and the efficiency of pembrolizumab and atezolizumab binding to and blocking the PD-1/PD-L1 axis was reduced. A 6-fold reduction in KD of pembrolizumab for the PD-1/PD-L1 complex was observed in the presence of Gal-3. In addition, atezolizumab PD-L1 blockade was reversed in the presence of Gal-3 and dramatically improved by GB1211. Similarly, in cell systems expressing high levels of PD-1 or PD-L1, Gal-3 reduced the binding of pembrolizumab and atezolizumab to their respective targets on cell surfaces that was also reversed by GB1211. Conclusions: Gal-3 was shown to vastly reduce the binding of the checkpoint inhibitors pembrolizumab and atezolizumab, by potentiating the interaction between PD-1 and PD-L1. GB1211 restored the binding of the anti-PD1/anti-PDL1 therapeutics and may thus reduce tumor resistance to these agents. These findings harmonize excellently with in vivo and clinical data showing an association between Gal-3 expression and lack of efficacy of PD-1/PD-L1 targeted checkpoint inhibitors. Galecto is now investigating the safety and efficacy of GB1211 in combination with atezolizumab in NSCLC.