Back

Loss of cBAF Complex Confers Sorafenib Resistance in Liver Cancer

Brinkman, J.;Dzama-Karels, M.;Bucklan, M.;Manner, C.;Sokolowski, M.;Raab, J.

2026-06-12 Cancer Biology
10.64898/2026.06.11.731725 bioRxiv
Show abstract

Sorafenib resistance limits the clinical benefit of first-line therapy in hepatocellular carcinoma (HCC), yet the epigenetic mechanisms underlying this resistance remain poorly understood. Using a chromatin-focused CRISPR/Cas9 dropout screen in HepG2 cells, we identified subunits of the canonical BAF (cBAF) complex -- ARID1A, ARID1B, and SMARCC1 -- as functional drivers of sorafenib resistance, while loss of the PBAF-specific subunit ARID2 had no effect, implicating cBAF specifically rather than SWI/SNF broadly. ARID1A and ARID1B mutations are associated with significantly worse overall survival in TCGA hepatocellular carcinoma cohorts. To define the underlying mechanism, we performed CUT&RUN profiling and RNA-seq in ARID1B-knockout HepG2 cells under sorafenib treatment. ARID1B loss triggered selective depletion of H3K4me3 and H3K27ac at 74 cBAF-dependent regulatory elements, 72% of which were directly occupied by ARID1B in wild-type cells. These sites were enriched for FOXA1 and HNF4 motifs, consistent with disruption of hepatocyte lineage regulatory chromatin. Direct profiling in HLF hepatocellular carcinoma cells confirmed that 179 FOXA1 binding sites were lost specifically under the combined perturbation of ARID1B loss and sorafenib treatment, with neither condition alone sufficient to drive this effect. Transcriptionally, ARID1B loss induced epithelial-mesenchymal transition programs and suppressed MYC targets, E2F targets, and mTORC1 signaling. These findings define a cBAF-FOXA1 regulatory axis that maintains hepatocyte lineage identity under kinase inhibitor stress, and whose disruption drives epigenetic reprogramming toward a mesenchymal drug-tolerant state. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=92 SRC="FIGDIR/small/731725v1_ufig1.gif" ALT="Figure 1"> View larger version (18K): org.highwire.dtl.DTLVardef@1ce2352org.highwire.dtl.DTLVardef@70f2f5org.highwire.dtl.DTLVardef@15ee96borg.highwire.dtl.DTLVardef@2e7edd_HPS_FORMAT_FIGEXP M_FIG C_FIG

Matching journals

The top 4 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.