RLY-4008, the First Highly Selective FGFR2 Inhibitor with Activity across FGFR2 Alterations and Resistance Mutations

**Background:** Fibroblast growth factor receptor 2 (FGFR2) is a receptor tyrosine kinase that drives oncogenesis in multiple solid tumors, including intrahepatic cholangiocarcinoma (iCCA), endometrial cancer, and gastric cancer. While pan-FGFR inhibitors (pan-FGFRi) have shown clinical activity in FGFR2-altered tumors, their efficacy is limited by on-target off-isoform toxicities—FGFR1-mediated hyperphosphatemia (reported in 55%–91% of patients) and FGFR4-mediated diarrhea (15%–36% of patients)—and by the emergence of acquired FGFR2 resistance mutations, such as FGFR2 V564F/L/I and FGFR2 N549K/H, observed in 50% of iCCA patients at progression. RLY-4008 was designed to overcome these limitations by exploiting differences in conformational dynamics between FGFR2 and other FGFR paralogs.

**Methods:** The paper includes preclinical biochemical, cellular, and in vivo xenograft studies, as well as three case studies from the ongoing phase I/II ReFocus trial (NCT04526106). Biochemical assays measured RLY-4008's covalent binding kinetics, kinase selectivity (against 468 kinases), and IC50 values against FGFR1-4. Cell proliferation was assessed in FGFR2-, FGFR1-, FGFR3-, and FGFR4-dependent cancer cell lines. In vivo efficacy was evaluated in subcutaneous xenograft models: an FGFR2-TTC28 iCCA patient-derived xenograft (PDX), SNU-16 (FGFR2-amplified gastric), AN3CA (FGFR2 K310R;N549K endometrial), and ICC13-7 (FGFR2-OPTN fusion iCCA) with CRISPR-mediated FGFR2 V564F knock-in. Pharmacodynamic (PD) analysis of pFGFR2 and serum phosphate was performed. Clinical cases were from patients with FGFR2-altered solid tumors who received RLY-4008 at 30–70 mg once daily. Responses were assessed per RECIST 1.1, and circulating tumor DNA (ctDNA) was analyzed with Guardant360 CDx.

**Key Results:** Preclinically, RLY-4008 covalently binds to Cys491 of FGFR2 with high potency (biochemical IC50 = 3 nmol/L) and >250-fold selectivity over FGFR1, >80-fold over FGFR3, and >5,000-fold over FGFR4. The covalent modification rate (kinact/KI) for FGFR2 was 3.45 × 10−2 per s/μM, 91 times faster than for FGFR1 (3.79 × 10−4 per s/μM). In cellular proliferation assays, IC50 values were <14 nmol/L in FGFR2-dependent cell lines but ≥212 nmol/L in FGFR1-, FGFR3-, and FGFR4-dependent lines. In xenograft models, RLY-4008 at 30 mg/kg twice daily induced tumor regression across all FGFR2-altered models with ≥90% pFGFR2 inhibition sustained over 12 hours. Serum phosphate levels were not significantly elevated (P > 0.2), unlike pan-FGFR inhibitors, which caused 32%–47% increases (P < 0.0001). Against resistance mutations, RLY-4008 retained or improved potency on FGFR2 V564F (IC50 shift <1) and showed a ~10-fold shift on FGFR2 N549K, comparable to futibatinib but superior to other pan-FGFRi. In ICC13-7-FGFR2 V564F xenografts, RLY-4008 (2 mg/kg once daily) induced rapid regression, while pan-FGFRi were ineffective. Tumors progressing on pemigatinib regressed when switched to RLY-4008.

Clinically, in 3 case studies from the ReFocus trial: Patient A (FGFR2-FLIP1 fusion iCCA, FGFRi-naive) achieved an 84% reduction in target lesions at day 163 without elevated serum phosphate or diarrhea, and underwent surgical resection. Patient B (FGFR2-WAC fusion iCCA with polyclonal resistance mutations FGFR2 N549K, N549D, V564I after infigratinib) showed complete ctDNA clearance by day 30 and a confirmed partial response (PR) with –72% reduction at day 57, deepening to –84% at day 162, without hyperphosphatemia or diarrhea. Patient C (FGFR2 Y375C mutant salivary gland carcinoma) achieved a PR with –62% reduction at day 64 and –67% at day 119, with low-grade self-resolving phosphate elevation. No diarrhea was reported in any case. FGFR2 on-target toxicities (stomatitis, onycholysis, dry eye) were low-grade, reversible, and manageable.

**Clinical Implications:** RLY-4008 represents a paradigm shift in targeting FGFR2-driven cancers by achieving unprecedented isoform selectivity and broad mutational coverage. The preclinical data demonstrate that selective FGFR2 inhibition can induce robust tumor regression while sparing FGFR1-mediated hyperphosphatemia and FGFR4-mediated diarrhea, which have historically limited pan-FGFRi dosing. The clinical cases provide proof of concept that this selectivity translates into meaningful responses in both FGFRi-naive and highly resistant FGFR2-altered tumors, including polyclonal resistance mutations. The ability to clear resistant clones in ctDNA within 30 days suggests that RLY-4008 may overcome acquired resistance, a major cause of pan-FGFRi failure. Furthermore, responses in a FGFR2-mutant salivary gland carcinoma highlight tumor-agnostic potential. The initial safety profile indicates that while on-target FGFR2 toxicities (nail, skin, oral) are common, they are generally low-grade and reversible, contrasting sharply with the dose-limiting off-target toxicities of pan-FGFRi. These findings support further development of RLY-4008 as a potential best-in-class therapy for FGFR2-altered solid tumors. The motion-based drug design approach used to develop RLY-4008 may also represent a novel strategy for designing selective kinase inhibitors across oncology and beyond.