Glycosylation of HER2+ human breast cancer cells control sensitivity to Trastuzumab

MARIÑO KV; PERROTTA RM; SALATINO M; CAGNONI A; RABINOVICH GA; DALOTTO-MORENO T

Congreso; Reunion anual de la AACR; 2017

Galectins decode glycan-containing information in various cell receptors adjusting signaling thresholds and modulating cellular functions. Upon specific binding to cell surface glycans they may induce receptor clustering, endocytosis and intracellular signaling, playing roles in many physiological and pathological conditions. In particular, galectin-1 (Gal1), a 14kD prototype galectin, binds to terminal N-acetyllactosamine residues on glycosylated proteins in the absence of α2-6 sialic acid capping (Gal1 permissive glycophenotype). Our lab was pioneer in demonstrating that tumors produce Gal1 to evade immunesurveillance and to promote aberrant angiogenesis mediating resistance to anti-VEGF therapy. The aim of this work is to explore the HER2+ breast cancer glycophenotype and investigate if tumor-produced Gal1 can activates RTKs mediating resistance to Trastuzumab (TZ). We first selected three HER2+ human breast cancer cell lines with different response to TZ: JIMT-1 (resistant-TZR), BT-474 and SK-BR-3 (sensitive-TZS). To identify specific glycan structures, we used a panel of biotinylated lectins (LEL, SNA, PNA, PHA-L and MAA). We found that TZS cell lines exhibited a Gal-1 restrictive glycophenotype characterized by high α2,6 sialic acid capping. Characterization of N-type glycans profile (WAX-HPLC) confirmed that TZSs cell lines exhibits abundant α2,6 sialic acid compared to TZR JIMT-1. This was also confirmed by RT-PCR of specific glycosyltransferases responsible of Gal1 ligands biosynthesis. Our results were further supported by analysis of raw data from public databases arrays (GSE62327) showing that patients who presented complete response to TZ exhibited higher levels of ST6GAL1, the glycosyltransferase responsible for α2-6 sialic acid capping. In accordance with the glycophenotype, TZR cell line binds and expressed higher levels of Gal1 when compared to TZS cell lines by Western Blot, RT-PCR and ELISA of conditioned medium, further suggesting a positive autocrine loop that could modulate cell behavior. Moreover, in silico analysis of raw data from the Long HER Study (GSE44272) revealed that patients with poor response to TZ express higher levels of Gal1 mRNA than long-term responders, reinforcing our hypothesis from a clinical standpoint. Finally, in order to elucidate Gal1 implications in TZ resistance, we knocked down Gal1 in JIMT-1 cell line using shRNA strategies. Interestingly, the absence of Gal1 expression sensitized JIMT-1 cells to TZ-inhibition in vitro, and the resistant phenotype was further reestablished by hrGal1 addition. In summary, we propose that individual HER2+ human breast cancer cells display particular ?glycosylation signatures? which, in association with Gal1 expression pattern, may control resistance to anti-HER2 targeted therapy and predict breast cancer clinical outcome.