Computational drug repositioning of bortezomib to reverse metastatic effect of GALNT14 in lung cancer

Abstract

Although many molecular targets for cancer therapy have been discovered, they often show poor druggability, which is a major obstacle to develop targeted drugs. As an alternative route to drug discovery, we adopted an in silico drug repositioning (in silico DR) approach based on large-scale gene expression signatures, with the goal of identifying inhibitors of lung cancer metastasis. Our analysis of clinicogenomic data identified GALNT14, an enzyme involved in O-linked N-acetyl galactosamine glycosylation, as a putative driver of lung cancer metastasis leading to poor survival. To overcome the poor druggability of GALNT14, we leveraged Connectivity Map approach, an in silico screening for drugs that are likely to revert the metastatic expression patterns. It leads to identification of bortezomib (BTZ) as a potent metastatic inhibitor, bypassing direct inhibition of poorly druggable target, GALNT14. The anti-metastatic effect of BTZ was verified in vitro and in vivo. Notably, both BTZ treatment and GALNT14 knockdown attenuated TGF{beta}-mediated gene expression and suppressed TGF{beta}-dependent metastatic genes, suggesting that BTZ acts by modulating TGF{beta} signalingTaken together, these results demonstrate that our in silico DR approach is a viable strategy to identify a candidate drug for undruggable targets, and to uncover its underlying mechanisms.