Abstract Activation of RAS and PI3Kα are the most frequent oncogenic events in cancer, playing a key role in many aspects of tumor cell physiology, including growth, survival, differentiation, and migration. While treatment options have recently emerged for a subset of KRAS mutant patients and inhibitors of PI3Kα catalytic activity have been approved, both approaches have been limited by drug resistance and in the case of PI3Kα, hampered by poor tolerability (e.g., hyperglycemia). An alternative therapeutic strategy that addresses drug resistance with improved tolerability is needed. Application of the FrontierTM platform, which integrates chemoproteomics, AI, and covalent fragment-based drug discovery, enabled the discovery of FMC-242, a potent, selective, and orally bioavailable covalent inhibitor of the PI3Kα -RAS family interactions disrupting oncogenic RAS and RTK signaling without impacting the insulin homeostasis. FMC-242 rapidly and selectively forms a covalent bond with cysteine 242 in the RAS Binding Domain (RBD) of PI3Kα resulting in allosteric inhibition of PI3Kα -RAS complex formation. This leads to inhibition of AKT activation in tumors with mutations in KRAS or PI3Kα, and where receptor tyrosine kinases, e.g., HER2, are activated. FMC-242 treatment of CDX and PDX models carrying HER2 amplification and/or KRAS mutation results in potent anti-tumor activity including regressions. FMC-242 is well tolerated in vivo, and inhibition of PI3Kα -RAS interaction does not impact insulin signaling or blood glucose level. Combination of FMC-242 with targeted therapies including EGFR inhibitors, KRASG12C inhibitors such as FMC-376, divarasib, olomorasib, or pan-RAS/KRAS agents results in enhanced efficacy and tumor regressions in vivo. Together, these data demonstrate the potential of FMC-242, a selective covalent inhibitor of PI3Kα -RAS interaction, to deliver improved outcomes for patients as monotherapy and in combination with targeted therapies in the clinic. Citation Format: Kevin R. Webster, Ryan McFadden, Abdul Awol, Koli Basu, Barun Bhhatarai, Yu-Hsin Chao, John Conway, Jay Duffner, Dan Erlanson, Robert Everley, Susan Fong, Sarah Gilfillan, Johannes Hermann, Alessandra Ianari, Lata Jayaraman, Svetlana Kholodar, Nathan Lavey, Tiep Le, Laura Marholz, Bethany Parker, Snahel Patel, Emily Sabbey, Shefali Sabhlok, Shayna Simonstein, Luke Utley, John Vassiliadis, Weiru Wang, Yan Wang. FMC-242, a highly potent and selective covalent inhibitor of the PI3Kα -RAS interaction, demonstrates robust anti-tumor activity as monotherapy and in combination with targeted therapies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 5762.

KRAS G12C is one of the most prevalent oncogenic mutations in nonsmall cell lung cancer. Herein we describe the discovery and optimization of divarasib (GDC-6036), an orally available, highly potent, and selective covalent KRAS G12C inhibitor. We demonstrate a significant noncovalent binding component of divarasib that contributes to its potency and rapid kinetics. Divarasib has greater potency and kinetics of alkylation compared with other KRAS G12C inhibitors in vitro and shows robust tumor growth inhibition in multiple KRAS G12C-positive cell lines.

Mutant KRAS is highly prevalent in human cancer and has been actively pursued as a target for drug discovery. Much progress has been made in drugging KRAS G12C, owing to the ability of inhibitors to covalently target its oncogenic cysteine mutation at codon 12. A number of KRAS G12C inhibitors have advanced to clinical development and are being investigated for the treatment of a variety of solid tumors. Notably, many patients with KRAS G12C-positive non-small cell lung cancer develop brain metastases. Herein, we report the discovery and development of a brain-penetrant inhibitor of KRAS G12C using divarasib as a starting point. Optimization efforts focused on reducing molecular weight and topological polar surface area as well as shielding of hydrogen bond donors. In this manner, active transport by both P-gp and breast cancer resistance protein (BCRP) was attenuated, and high exposure in rodent brain tissue was achieved.

We introduce hydrogen-exchange experimental structure prediction (HX-ESP), a method that integrates hydrogen exchange (HX) data with molecular dynamics (MD) simulations to accurately predict ligand binding modes, even for targets requiring significant conformational changes. Benchmarking HX-ESP by fitting two ligands to PAK1 and four ligands to MAP4K1 (HPK1) and comparing the results to X-ray crystallography structures, demonstrates that HX-ESP can identify binding modes across a range of affinities significantly outperforming flexible docking for ligands necessitating large conformational adjustments. By objectively guiding simulations with experimental HX data, HX-ESP overcomes the long time scales required for binding predictions using traditional MD. This advancement enhances the accuracy of computational modeling in drug discovery and thus will accelerate the development of effective therapeutics.

Targeted protein degradation (TPD) leverages the ubiquitin-proteasome system to eliminate disease-causing proteins via E3 ligases. To date, the field is limited to utilizing a few of the over 600 human E3 ligases. To expand this repertoire, we conducted structural and functional validation of DDB1 (Damage-specific DNA binding protein 1) and Cullin-associated factor (DCAF)2 (DTL/CDT2), a Cullin4-RING ligase substrate adaptor implicated in DNA damage response and cancer, as a novel E3 for TPD. Cryoelectron microscopy (cryo-EM) structures of the DCAF2:DDB1:DDA1 complex (3.3 Å), a ligand bound complex (3.1 Å), and a ternary complex with a covalent proteolysis-targeting chimera (PROTAC) and BRD4 (3.4 Å) reveal PROTAC-mediated substrate recruitment. Using covalent bifunctional tool compounds engaging residue C141 in the WD40 domain, we demonstrate robust ubiquitination in biochemical assays and cellular TPD using the COFFEE (covalent functionalization followed by E3 electroporation) method. These findings position DCAF2 as a promising E3 adaptor for PROTAC strategies and identify C141 as a relevant site for future PROTAC discovery.

Loss of function mutations of the TP53 gene are the most common genetic defects across all human cancers. These mutations have long been intractable to drug development. The p53 Y220C mutation is common, occurring in ∼1% of cancers, leading to destabilization, aggregation and loss of p53 protein function. Application of the FrontierTM platform, which integrates chemoproteomics, machine-learning, and covalent fragment-based drug discovery, enabled the discovery of FMC-220, a potential first-in-class covalent activator of p53 Y220C, which is highly potent and selective in restoring p53 tumor suppressor function. FMC-220 works by forming an irreversible covalent bond with the mutant cysteine at amino acid 220 stabilizing the structure of the central p53 DNA binding domain. The covalent mechanism of action of FMC-220 delivers potent and durable induction of p53 transcriptional response that includes upregulation of known regulators of cell cycle arrest, senescence and cell death. In contrast to reversible Y220C binders including rezatapopt, covalent engagement of Y220C results in persistent activation of p53 following drug washout that leads to enhanced Y220C mutant cancer cell senescence and death. The exquisite selectivity of FMC-220 was confirmed through chemoproteomic analysis and knock-out of the Y220C allele demonstrating drug action is through modification of the Y220C mutant form of p53. Importantly, the action of FMC-220 is specific for the Y220C mutant form of the p53 protein, sparing any effect on wild-type p53 function. FMC-220 inhibits tumor cell viability with ∼100-fold increased potency relative to rezatapopt across a panel of Y220C mutant cell lines including in tumor cells with a KRAS co-mutation and irrespective of histology. This translates into potent pharmacodynamic activity and tumor regression in Y220C mutant CDX and PDX models in vivo. In addition, FMC-220 is very well tolerated in vivo. Together, these data demonstrate the promise of FMC-220, a first-in-class covalent activator of p53 Y220C, with the potential to deliver improved outcomes for patients with Y220C mutation positive tumors. John-Paul Upton, Irina Dotsenko, Vikram Narayan, Marya Liimatta, Jessica Paredes, Mikayla Shanafelt, Brian Tran, Svetlana Kholodar, Kyle Anderson, Tiep Le, Sarah Gilfillan, Shefali Sabhlok, Ziwei Hu, Reed Stein, Evan McMahon, Weiru Wang, Yan Wang, Snahel Patel, Lata Jayaraman, David Sperandio, Kevin R. Webster. Restoring the function of the guardian of the genome: FMC-220 a highly potent and selective covalent activator of p53 Y220C [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3787.

The lack of selective and safe in vivo IRE1α tool molecules has limited the evaluation of IRE1α as a viable target to treat multiple myeloma. Focus on improving the physicochemical properties of a literature compound by decreasing lipophilicity, molecular weight, and basicity allowed the discovery of a novel series with a favorable in vitro safety profile and good oral exposure. These efforts culminated in the identification of a potent and selective in vivo tool compound, G-5758, that was well tolerated following multiday oral administration of doses up to 500 mg/kg. G-5758 demonstrated comparable pharmacodynamic effects to induced IRE1 knockdown as measured by XBP1s levels in a multiple myeloma model (KMS-11).

Hematopoietic progenitor kinase 1 (HPK1) is a negative regulator of T-cell receptor signaling and as such is an attractive target for cancer immunotherapy. Although the role of the HPK1 kinase domain (KD) has been extensively characterized, the function of its citron homology domain (CHD) remains elusive. Through a combination of structural, biochemical, and mechanistic studies, we characterize the structure-function of CHD in relationship to KD. Crystallography and hydrogen-deuterium exchange mass spectrometry reveal that CHD adopts a seven-bladed β-propellor fold that binds to KD. Mutagenesis associated with binding and functional studies show a direct correlation between domain-domain interaction and negative regulation of kinase activity. We further demonstrate that the CHD provides stability to HPK1 protein in cells as well as contributes to the docking of its substrate SLP76. Altogether, this study highlights the importance of the CHD in the direct and indirect regulation of HPK1 function.

Hematopoietic progenitor kinase 1 (HPK1) serves a key immunosuppressive role as a negative regulator of T-cell receptor (TCR) signaling. HPK1 loss-of-function is associated with augmentation of immune function and has demonstrated synergy with immune checkpoint inhibitors in syngeneic mouse cancer models. These data offer compelling evidence for the use of selective small molecule inhibitors of HPK1 in cancer immunotherapy. We identified a novel series of isoquinoline HPK1 inhibitors through fragment-based screening that displayed promising levels of biochemical potency and activity in functional cell-based assays. We used structure-based drug design to introduce key selectivity elements while simultaneously addressing pharmacokinetic liabilities. These efforts culminated in a molecule demonstrating subnanomolar biochemical inhibition of HPK1 and strong in vitro augmentation of TCR signaling in primary human T-cells. Further profiling of this molecule revealed excellent kinase selectivity (347/356 kinases <50% inhibition @ 0.1 μM), a favorable in vitro safety profile, and good projected human pharmacokinetics.

<p>Fig. S1. ERK inhibitors and their effect on MAPK mutant cell lines. A. Structure of ERK inhibitors used in this study. B. Dose response curve of ERK inhibitors on the viability of A375, IPC298, SKMEL30, HCT116, MIA PaCa2 and Panc1 cells. Fig. S2. Western blot analysis of parental, VI-3-R and G994-R HCT116 (A) or MIA PaCa2 (B) cells treated with indicated ERK inhibitors. Fig. S3. Western blot analysis of HCT116 cells (A) and SKMEL30 (B) stably expressing indicated ERK1 or ERK2 mutants. Fig. S4. Copy number and expression analysis of MITF in SKMEL30-ERKi-R cells. A. MITF copy gain in SKMEL30-V11e-R, -VI-3-R and -G994-R. B. MITF copy number vs. expression in SKMEL30-V11e-R, VI-3-R and G994-R cells. Fig. S5. Dose response curve of ERK2 overexpressing IPC298 cells treated with MEK inhibitors GDC-0973 or AZD6244 (A) and ERK2 amplified IPC298-G994-R (B) or MIA PaCa2-S984-R (C) cells treated with indicated ERK-inhibitors from alternate scaffold class. Fig. S6. Dose response curve of parental or ERKi-resistant HCT116 (A) and MIA PaCa2 (B) cells treated with indicated inhibitors. The data shown are same as in Fig. 6B for HCT116 (A) and Fig. 6C for MIA PaCa2 (B) grouped by drug treatment.</p>

PDF file, 869K, Proliferation and apoptosis analysis by IF staining in EL4 tumor sections.

PDF file - 731KB, Supplementary Figure 1. Schematic workflow of MMP-seq. Supplementary Figure 2. Examples of tiling and hotspot amplicon designs. Supplementary Figure 3. Consistent target enrichment and variant allele quantification in FF biological replicates. Supplementary Figure 4. Correlation between called variant frequencies in paired FF and FFPE samples. Supplementary Figure 5. UDG treatment improves sequencing specificity but has no impact on sensitivity. Supplementary Figure 6. PIK3R1 mutation profiles from TCGA endometrial cancer study. Supplementary Table 1. MMP-seq target genes. Supplementary Table 2. Latin Square Design. Supplementary Table 3. Latin square mutation read frequency detected when sequencing input cell lines for Latin square cross-dilutions. Supplementary Table 4. Pretreatment of FFPE DNA samples with uracil-DNA glycosylase (UDG) resulted in markedly reduction of false positives.

PDF file, 28K, Effects of 30D8 on CXCL12 and VEGF serum levels in HM7 tumor bearing mice.

PDF file, 26K, Comparison of PK in Balb c/nude mice of three hamster antibodies with similar binding affinity toward human CXCL12.

KRAS is one of the most frequently mutated genes in cancer and was long considered undruggable until the recent discovery of inhibitors that bind the inactive (GDP-bound) form of KRASG12C. The most clinically advanced of these first-generation molecules have demonstrated clinical response rates of 30-45% and approximately 6-month progression-free survival in lung cancer patients. While significant, a majority of patients fail to achieve a clinical response and acquired resistance is common. Resistance to first-generation inhibitors can be driven by upregulation of the activated (GTP-bound) form of KRASG12C, which remains an undrugged form of the oncoprotein. Here we report the discovery of FMC-376, a novel inhibitor of the activated, GTP-bound, form of KRASG12C, which also potently inhibits the inactive, (GDP-bound), form of KRASG12C. FMC-376 was discovered through the FrontierTM platform, which integrates chemoproteomics, machine-learning, and covalent fragment-based drug discovery. FMC-376 binds KRAS in the switch II pocket, rapidly forming a covalent bond with cysteine 12 in the presence of either GDP or GTP. X-ray crystallography demonstrated that Cys12 adopts a novel confirmation in forming a covalent bond with FMC-376. This results in potent inhibition of RAF1 and PI3Kα effector interactions (IC50 = 0.007 μM for both respectively at 2 h) in contrast to sotorasib or adagrasib (IC50 > 50 and ~ 5 μM respectively). FMC-376 treatment results in potent anti-tumor activity across a panel of KRASG12C mutant tumor cell lines, sparing non- KRASG12C cell lines. To model resistance mediated by activated KRASG12C, a mutation that abrogates GTPase activity (A59G) was introduced into KRASG12C. This upregulation of GTP-bound KRASG12C drives significant (>10-fold) resistance to both adagrasib and sotorasib in tumor cell viability assays whereas FMC-376 remains equipotent in settings where GTP-bound KRASG12C is upregulated. Evaluation of FMC-376 in models where EGFR signaling (a suspected mechanism of clinical resistance) is induced demonstrated rapid and durable target engagement in contrast to both sotorasib and adagrasib which show decreased effectiveness after EGF stimulation. Further evaluation of FMC-376 in vivo has demonstrated rapid and durable KRASG12C target occupancy (>90%) and pathway inhibition in tumors, resulting in regression of CDX/PDX tumor models. FMC-376, an inhibitor of both active and inactive forms of KRASG12C, provides a differentiated mechanism of action with the potential for broader and more durable response in the clinic. Citation Format: Snahel Patel, Barun Bhhatarai, Philamer Calses, Daniel Erlanson, Robert Everley, Susan Fong, Phil Gerken, Johannes C. Hermann, Tiep Le, Li-kai Liu, Evan McMahon, Richard M. Neve, Tony Phan, Allison Roberts, Mikayla Shanafelt, Sophie Siemsgluess, Jocelyn Staunton, Yan Wang, Weiru Wang, Monika Williams, Kevin R. Webster. Discovery of FMC-376 a novel orally bioavailable inhibitor of activated KRASG12C [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1142.

PDF file, 38K, Sequence alignment of 30D8 and humanized 30D8 (hu30D8).

PDF file, 20K, Effects of 30D8 on 4T1 cell migration in vitro.