Back

TMPRSS2-ERG confers resistance to antiandrogens: mechanism and therapeutic implications

Sekar, A.; Chatterjee, R.; Selvadurai, B. R.; Ray, L.; Verma, A.; Nataraj, N. B.; Garcia, D. D.; Giri, S.; Genna, A.; Karatekin, F.; Gupta, N.; Zerbib, M.; Vinik, Y.; Avioz, T.; Weizman, E.; David, E.; A. Schaffer, A.; Pan, Y.; Huang, H.; M van Weerden, W.; Corey, E.; Hunt, H.; Greenstein, A.; Blecher-Gonen, R.; Oren, R.; Afek, A.; Amit, I.; Lev, S.; Ku, A.; Kartal, S.; R. Bright, J.; T. Lis, R.; L. Dahut, W.; Sowalsky, A. G.; Ruppin, E.; Yarden, Y.

2025-01-20 cancer biology
10.1101/2025.01.17.633582 bioRxiv
Show abstract

Approximately 50% of prostate cancer (PCa) patients harbor fusions involving the TMPRSS2 and ERG genes. Despite this, tailored therapies targeting the fused gene, tERG, remain undeveloped. Our study analyzed biopsy samples from two clinical trials assessing the efficacies of androgen receptor (AR) signaling inhibitors (ARSIs). The results revealed that tERG promotes resistance to ARSIs and is associated with elevated levels of the glucocorticoid receptor (GR). Subsequent assays showed that GR directly interacts with tERG, alleviates allosteric autoinhibition and prevents chemotherapy-induced tERG degradation. In PCa models, either inhibiting GR or lowering cortisol levels suppressed tumor growth in tERG-positive models, but not in fusion-negative models. In addition, patient-derived fusion-positive xenografts displayed enhanced sensitivity to combined GR and AR inhibitors. Collectively, these findings highlight TMPRSS2-ERG as a new biomarker and propose that simultaneous inhibition of GR and AR may specifically benefit tERG-positine patients. However, GR stimulatory corticosteroid therapies may not be advisable for this patient subgroup.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.