PERK-driven ER stress states determine sensitivity to the GRP78 inhibitor BOLD-100 in gastric cancer

Abstract Gastric cancer (GC) has a poor prognosis and limited durable responses to systemic therapy, underscoring the need for new therapeutic strategies. Rapid tumor growth can increase endoplasmic reticulum (ER) stress, and glucose-regulated protein 78 (GRP78)-supported ER proteostasis may help tumor cells tolerate this stress in GC. BOLD-100 is a clinical-stage ruthenium-based compound that disrupts GRP78-associated proteostasis. However, the ER stress response is heterogeneous, and whether baseline protein kinase RNA-like ER kinase (PERK) status influences response to BOLD-100 and supports mechanism-guided combination strategies in GC remains unclear. Across a panel of GC cell lines, baseline PERK expression correlated with BOLD-100 sensitivity across GC cell lines. PERK-high models exhibited marked growth suppression and apoptotic cell death, accompanied by robust ER stress and unfolded protein response activation. PERK knockdown attenuated stress signaling and reduced cytotoxicity. In contrast, PERK-low models preferentially engaged an adaptive ATR-Chk1 checkpoint response, limiting the activity of BOLD-100 monotherapy. ATR inhibition selectively sensitized PERK-low cells to BOLD-100, increasing DNA damage and apoptosis. ER stress and unfolded protein response markers were not further increased, supporting checkpoint abrogation as the primary driver of sensitization. In xenograft models, BOLD-100 suppressed the growth of PERK-high tumors, whereas the addition of an ATR inhibitor produced greater antitumor activity in PERK-low tumors. Together, these findings support baseline PERK expression as a candidate predictive biomarker of BOLD-100 response in GC and suggest a PERK-guided treatment strategy: BOLD-100 monotherapy in PERK-high settings and ATR inhibitor combination therapy in PERK-low settings.