Phase 1 study of WNT pathway Porcupine inhibitor CGX1321 and phase 1b study of CGX1321 + pembrolizumab (pembro) in patients (pts) with advanced gastrointestinal (GI) tumors.

Abstract

3514 Background: Despite the fundamental role of WNT signaling in GI cancers, especially colorectal cancer (CRC), no drug targeting WNT signaling has been successfully developed. CGX1321, a highly-potent and selective inhibitor of O-acyltransferase Porcupine, blocks WNT ligand secretion. Preclinical studies show that CGX1321 inhibits WNT signaling and growth of tumors with RSPO fusions or inactivating RNF43 mutations. Activation of WNT signaling has been associated with cancer immune-suppression and resistance to immunotherapy. Methods: The first-in-human trials (NCT02675946 in U.S., NCT03507998 in China) include phase 1 CGX1321 dose escalation (3 – 18 mg, once daily) and dose expansion (18 mg), and phase 1b CGX1321 + pembro combination (Keynote 596). The primary objectives were safety, tolerability and identification of recommended doses and schedules for further evaluation. Secondary objectives were characterization of the PK profile and PD response and anti-tumor activity. CGX1321 was dosed once daily, orally, for 21 days out of 28-day cycles as single-agent, or for 14 days out of 21-day cycles in combination with pembro. Results: As of January 1, 2023, 77 pts were enrolled, including 38 pts with solid tumors in the CGX1321 dose escalation, 17 pts with CRC or small bowel cancer (SBC) carrying RSPO or RNF43 alterations in the CGX1321 expansion, 17 pts with microsatellite stable (MSS) CRC in the CGX1321 + pembro dose escalation and 5 pts with MSS CRC in the CGX1321 + pembro expansion. Six of the 17 pts with MSS CRC or SBC from the CGX1321 single-agent expansion were rolled over to CGX1321 + pembro upon disease progression. Across all cohorts, treatment was well tolerated with the majority of AEs being Grade 1/2 and not related to CGX1321, while Grade > 3 AEs related to CGX1321 were infrequent (~6%). Dysgeusia, a common AE observed with other WNT inhibitors, was mild (mostly Grade 1). Bone resorption, the main on-target AE, was manageable and preventable by prophylactic administration of denosumab or zoledronic acid. PK and PD analyses showed adequate drug exposure and significant inhibition of WNT signaling as measured by reduced hair follicle axin2 expression from 12 mg dose. Twelve of 17 pts (71%) with confirmed tumor genetic alterations in RSPO or RNF43 had SD with median time to progression of 112 days (range: 56 – 392). Of the 6 pts in the roll-over cohort, 3 pts with RSPO3 fusion tumors achieved PR during CGX1321 + pembro combination therapy. Conclusions: CGX1321 has demonstrated potent inhibition of the WNT pathway with manageable side effects. Promising efficacy signals have been observed in pts whose tumors harbor RSPO fusion, supporting further development of CGX1321 monotherapy and CGX1321 + anti-PD-1/L1 in a defined patient population that is historically known to be refractory to standard therapies and immune checkpoint inhibitors. Clinical trial information: NCT02675946 , NCT03507998 .