Back

AR suppression rewires a SOX4 regulatory program to promote enzalutamide resistance in prostate cancer

Hakkola, S.; Peramaki, A.; Kiviaho, A.; Urbanucci, A.; Nykter, M.; Marttinen, M.

2026-04-27 cancer biology
10.64898/2026.04.27.721035 bioRxiv
Show abstract

Treatment resistance in prostate cancer remains a major clinical challenge. While single-cell transcriptomics enables mapping of cellular composition and resistance-associated states, the gene regulatory programs driving these state transitions remain poorly defined. Here, we integrate single-cell chromatin accessibility and transcriptome sequencing data from an in vitro model to resolve the regulatory program architecture underlying anti-androgen resistance. We identify coordinated remodeling of androgen signaling- and stress response-associated regulatory programs, together with a pronounced gain of a SOX4-driven regulatory program in resistant cells. Analysis of clinical datasets reveals that SOX4 activity is both context-dependent and lineage-specific. In primary tumors, SOX4 cis-regulatory activity is elevated in PTEN-loss tumors, consistent with a PTEN-PI3K/AKT/mTOR-SOX4 axis that primes AR-independent epithelial lineages for subsequent activation of a more extensive SOX4 trans-regulatory program. This program is further amplified following AR-targeted therapy, indicating therapy-induced reinforcement of SOX4-linked regulatory adaptation. TF footprinting additionally implicates Nuclear Factor I family transcription factors as upstream regulators of SOX4, supporting a role in promoting SOX4 activation following AR inhibition. Together, these findings demonstrate that AR inhibition reshapes the gene regulatory landscape to promote engagement of the SOX4 program, facilitating prostate cancer progression under anti-androgen therapy.

Matching journals

The top 3 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.