Understanding the role of the tumor suppressor gene ARID1A in gastric cancer initiation

Introduction: Gastric cancer remains a major cause of global cancer mortality and disproportionately affects underserved racial and ethnic minority populations in the United States. Latino patients experience a higher incidence and worse outcomes, yet genomic data and patient-derived models remain underrepresented. Whole-exome sequencing (WES) of our Latino GC cohort showed that 30% of tumors harbor pathogenic ARID1A mutations, paralleling TCGA frequencies. ARID1A, a core subunit of the SWI/SNF chromatin-remodeling complex, regulates enhancer activity and transcriptional accessibility. Its loss is linked to poor prognosis and metastasis. We hypothesized that ARID1A loss establishes an early premalignant state through chromatin dysregulation and transcriptional reprogramming. To test this, we developed long-term gastric organoid models derived from Latino patients to define the earliest cellular and molecular consequences of ARID1A loss. Methods: We generated multiple isogenic gastric organoid lines using CRISPR/Cas9 editing in Latino patient-derived normal gastric organoids, including ARID1A/TP53 double-knockout (dKO) organoids and TP53 knockout controls. Organoids were cultured long-term for up to 350 days, with early (~150-day) and late (~350-day) time points analyzed. Multi-omic characterization included H&E histology, RNA sequencing, whole exome sequencing, and CUT&RUN profiling. Comparisons across dKO, TP53 KO, and wild-type controls were used to define the effects of ARID1A loss on early tumorigenesis. Results: Genome editing generated ARID1A and TP53 loss, and dKO organoids developed progressive morphological abnormalities, including increased proliferation and distinct architectural changes compared to controls. RNA-seq analyses showed genotype-specific and time-dependent transcriptional reprogramming. Differential expression analysis revealed downregulation of digestion and iron transport pathways and upregulation of cilium movement and ciliogenesis. Whole exome sequencing analysis showed fourteen variants were shared across long-term dKO clones, while each clone independently acquired 1-6 additional unique mutations. Affected genes were involved in transcriptional regulation, extracellular matrix remodeling, vesicle trafficking, and receptor tyrosine kinase signaling. The presence of shared and clone-specific variants suggests progressive genomic instability and supports early neoplastic evolution driven by dual ARID1A/TP53 loss. Conclusion: Latino-derived gastric organoid models demonstrate that ARID1A and TP53 loss cooperatively drive transcriptional reprogramming, genomic instability, and early dysplastic progression. Ongoing CUT&RUN profiling will define chromatin accessibility changes. This work provides a mechanistic foundation for drivers of gastric cancer and a platform for precision prevention strategies.