Discovery of RGT-018: A Potent, Selective, and Orally Bioavailable SOS1 Inhibitor for KRAS-Driven Cancers

KRAS is the most frequently dysregulated oncogene in cancer (1). Activating mutations of KRAS are predominant in some of the most aggressive cancer types, such as non–small cell lung cancer (NSCLC), colorectal cancer (CRC), and pancreatic cancer (PAC; refs. 1, 2). These KRAS mutations are mainly found in three hotspot codons: G12, G13, and Q61, with G12 representing the most common mutation site (3). Despite KRAS being a prime target of oncology drug discovery for decades, it was thought to be undruggable due to the high binding affinity of its nucleotide ligand and the lack of apparent deep pockets to engage a small-molecule inhibitor (4). Currently, the U.S. FDA–approved KRAS inhibitors, sotorasib and adagrasib, are covalently bound small molecules targeting the KRAS^G12C mutant (5, 6). The clinical benefits have been observed for many patients treated with KRAS^G12C inhibitors (7–11). However, acquired resistance to single-agent therapy eventually occurred in most patients, including KRAS amplification and secondary KRAS mutations (12–14).

KRAS is a small GTPase that cycles between the GTP- and GDP-loaded states (1–4). Transition between these states is facilitated by guanine nucleotide exchange factors (GEF) and GTPase-activating proteins that promote the exchange of GDP with GTP and the hydrolysis of GTP, respectively (15, 16). One of the major regulators of this process is the Son of Sevenless (SOS) protein, a GEF that acts as a key activator for KRAS function (15, 17, 18). The binding between KRAS and SOS proteins promotes conversion of inactive GDP-bound KRAS to active GTP-bound KRAS at the catalytic site and then activates the downstream MAPK and PI3K/ AKT pathways (15, 19). There are two homologs of SOS (SOS1 and SOS2) that contribute to the GEF activity (16, 20). Various studies have demonstrated a dominant role for SOS1 over SOS2 in regulating the GDP-GTP cycle of KRAS (21–23). Depletion of SOS1 has been shown to inhibit the growth of KRAS-mutant cancer cells in vitro and in vivo (24, 25). SOS1 deletion attenuated KRAS^G12D-induced myeloproliferative neoplasm and prolonged survival in mice bearing tumors with oncogenic KRAS^G12D-induced leukemogenesis in a genetically engineered mouse model (26). Given its direct interaction with KRAS, inhibition of SOS1 has become an attractive strategy in targeting KRAS-driven tumors (15, 17, 18, 20, 27). The catalytic site of SOS1 has a well-defined binding pocket adjacent to the KRAS:SOS1 interface (28). Therefore, disrupting the KRAS:SOS1 interaction with an SOS1 inhibitor is a potential therapeutic approach to treat KRAS-driven cancers.

Recently, SOS1 inhibitors BI 1701963 and MRTX0902 have entered phase 1/1b clinical trials (29–31). In general, SOS1 inhibitors such as BI-3406 (preclinical tool compound) and MRTX0902 were reported to disrupt KRAS:SOS1 interaction and display antiproliferative activity in various cancer cells (29, 32). SOS1 inhibitors as single agents were less effective in inhibiting KRAS-driven cancer cell growth, and combinational strategies further enhanced antitumor activities (29, 32). SOS1 inhibitors exhibited good in vitro activities and in vivo efficacy, especially when used in combination with MEK or KRAS^G12C inhibitors (5, 6, 29, 32). It was also reported that BI-3406 in combination with trametinib may be a useful strategy to overcome acquired resistance caused by the secondary KRAS^Y96D mutation (33). In addition to the SOS1 inhibitors that have advanced to clinical trials, other small molecules and new modalities such as SOS1 degraders have emerged, suggesting that targeting SOS1 has gained considerable traction across the pharmaceutical industry for KRAS-driven cancer drug discovery (17, 27, 28, 30, 31, 34–38).

In this work, we describe the discovery of RGT-018, a potent and selective SOS1 inhibitor. RGT-018 potently blocked the interaction of KRAS:SOS1 and reduced cell proliferation of a broad spectrum of KRAS-driven cancers as a single agent or in combinational settings with MEK, KRAS^G12C, EGFR, or CDK4/6 inhibitors in vitro. RGT-018 inhibited tumor growth and suppressed KRAS signaling in tumor xenografts. Combination with MEK or KRAS^G12C inhibitors led to more profound tumor regression in vivo. Furthermore, RGT-018 overcame the putative mechanisms of resistance by secondary mutations (such as G13D and Y96D) or amplifications of KRAS as a single agent and in combination with MEK, KRAS^G12C, EGFR, or CDK4/6 inhibitors. RGT-018 demonstrated excellent pharmacokinetics (PK) properties and safety profile in preclinical species, and it is selected as a clinical development drug candidate as a single agent or in combination therapies to treat broad KRAS–driven patients with cancer.

The recombinant protein sequence used covers amino acids 560 to 1,049 of human SOS1 (Q07889), with the ⁵⁶⁰QEEK⁵⁶³ mutated to ⁵⁶⁰GAMA⁵⁶³. The construct was subcloned to a pET28a vector, with an N-His-TEV tag. The SOS1 protein was overexpressed in Escherichia coli BL21(DE3) cells through plasmid transformation of DE3 cells that were cultured at 37°C in Luria broth medium supplemented with 50 μg kanamycin per milliliter of Luria broth medium. Protein expression was induced when the OD₆₀₀ reached a value of ∼ 0.6 to 0.8 by adding isopropyl β-D-1-thiogalactopyranoside to the final concentration of 1 mmol/L, at which point the temperature was lowered to 16°C. The cells were harvested 16 hours later by centrifugation at 6,000g for 10 minutes at 4°C. They were resuspended in lysis buffer [50 mmol/L Tris pH 8.0, 500 mmol/L NaCl, and 1 mmol/L tris(2-carboxyethyl)phosphine (TCEP)] and lysed by sonication. Lysate was clarified by centrifugation at 30,000g for 30 minutes at 4°C. The supernatant was applied to a nickel affinity column (QIAGEN) that was equilibrated with lysis buffer. Protein was eluted using an elution buffer (50 mmol/L Tris pH 8.0, 500 mmol/L NaCl, 1 mmol/L TCEP, and 250 mmol/L imidazole). After cleavage of His-tag by TEV protease, the fraction was applied to a second nickel column and the flow-through was collected. After concentration, the protein fraction was purified with a HiLoad 16/600 Superdex 200 column (GE Healthcare) equilibrated with buffer containing 25 mmol/L Tris (pH 7.5), 50 mmol/L NaCl, and 1 mmol/L dithiothreitol. Peak fractions were collected, concentrated to 100 mg/mL, and stored at −80°C.

To make the SOS1/ RGT-018 complex, SOS1 and RGT-018 were mixed at a 1:5 ratio, in which the final protein concentration was 15 mg/mL. Crystals of the complex were grown by mixing equal volumes of the precipitating solution [0.1 mol/L imidazole pH 8.0, and 6% polyethylene glycol 8000 (PEG8000)]. After going through a cryoprotectant with 30% ethylene glycol, crystals were flash-frozen with liquid nitrogen. Diffraction data were collected at beamline I03 at the Diamond Light Source.

Integration, scaling, and merging of the intensities were carried out with the program iMosflm and SCALA from the CCP4 software suite (39). The structure of SOS1 was solved by molecular replacement with the available SOS1 PDB coordinates (PDB ID 5OVE) as the searching model, using Phaser in the CCP4 software suite (39). Refinement cycles were performed using Refine in Phenix, alternating with iterative manual rebuilding in Coot (40). A summary of the data collection and structure refinement statistics is given in Supplementary Table S1.

Recombinant KRAS proteins were based on KRAS isoform 4B (Uniprot, P01116-2): KRAS^G12D (amino acids 1–169, N-terminal 6×His-tag, TEV cleavage site, Avi-Tag) and KRAS^G12C (amino acids 1–169, N-terminal 6×His-tag, TEV cleavage site, Avi-Tag). Interacting proteins SOS1 (amino acids 564–1,049, N-terminal GST-tag, TEV cleavage site) and SOS2 (amino acids 564–1,043, N-terminal GST-tag, TEV cleavage site) were expressed as GST fusions. All proteins were generated from Viva Biotech (Shanghai) Ltd. Measurements of various protein–protein interactions were performed using the AlphaScreen technology developed by PerkinElmer. The AlphaScreen beads were glutathione-coated AlphaLISA acceptor beads (AL109C, PerkinElmer) and AlphaScreen streptavidin donor beads (6760002S, PerkinElmer). Nucleotide GDP was purchased from Sigma (G7127) and Tween 20 also from Sigma (P7949-100ML).

All assays were carried out in PBS (Gibco, 10010023), containing 0.1% BSA (Sigma, A1933-5G), and 0.05% Tween 20 in white ProxiPlate-384 plus plates (6008280, PerkinElmer) with a final volume of 10 μL. In brief, biotinylated KRAS^G12D or KRAS^G12C proteins (10 nmol/L final assay concentration), GST-SOS1 or GST-SOS2 (5 nmol/L final assay concentration) and 10 μmol/L GDP (final assay concentration) were mixed in assay buffer prior to use and kept at room temperature. RGT-018 was diluted to a final start concentration of 1 μmol/L (SOS1 assay) or 100 μmol/L (SOS2 assay) and tested in duplicate. 100 nL of RGT-018 solution was transferred per well in nine concentrations in duplicate with serial 1:3 dilutions using Tecan D300e Digital Dispenser. 5 μL of KRAS^G12D or KRAS^G12C, SOS1 or SOS2 and GDP mix was added into the assay plate to the 100 nL of compound solution. After a 30 minutes incubation, AlphaLISA glutathione acceptor beads and AlphaScreen streptavidin donor beads were mixed in assay buffer at a concentration of 5 μg/mL (final assay concentration), and 5 μL of bead mix was added into the assay plate. Plates were kept at room temperature in a darkened incubator for 3 hours. After a 3 hours incubation, the signal was determined using Envision (PerkinElmer). Data were analyzed using the GraphPad Prism–based data software.

The kinases selectivity of RGT-018 was investigated in a panel of 330 kinases at the concentration of 1 μmol/L measured by Kinome-Wide Panel service at Nanosyn, Inc. The microfluidic mobility shift assay (Caliper) or ADP-Glo assay was used to determine the inhibitory activity of RGT-018 against kinases. The buffer components and assay conditions differ based on the specific assay. The assay has a total volume of 10 μL (5 μL enzyme buffer and 5 μL substrate buffer). After incubation and termination, substrate and product were separated and quantified electrophoretically using the microfluidic-based LabChip 3000 Drug Discovery System from Caliper Life Sciences. ADP-Glo was another method of detection used in some assays in which data were obtained by quantifying the intensity of luminescence.

NSCLC H358 cells (CRL5807) or PAC MIA PaCa2 cells (CRL1420) from ATCC were grown in cell culture flasks using RPMI1640 or DMEM supplemented with 10% FBS (10099141C, Gibco) and 1% penicillin/streptomycin (15140122, Gibco). Cells were incubated at 37°C and 5% CO₂ in a humidified atmosphere, with subcultivation performed twice a week. Cells were trypsinized, counted, and plated at 50,000 or 16,000 cells/well in 96-well plates. The day after plating, the culture medium was removed and replaced with serum-free medium. The final compound concentration covered a range between 5 μmol/L and 0.25 nmol/L with serial 1:3 dilutions in 10 concentrations. About 0.5 μL serial dilutions of RGT-018 were added in duplicates and incubated with cells at 37°C. The incubation lasted for 1 hour in H358 cells. For MIA PaCa2 cells, after incubation for 2 hours, the cells were stimulated by 10× epidermal growth factor (236-EG, R&D Systems) with 10 ng/mL final concentration for 10 minutes. After incubation, the serum-free medium was removed in both cells and 50 μL freshly prepared 1× lysis buffer was added to the cells in each well. The cell lysates were allowed to mix for 30 minutes on an orbital shaker. About 6 μL of the lysate was transferred to a 384-well OptiPlate (6007290, PerkinElmer), and then 3 μL of acceptor mix was added into the wells. The 384-well assay plate was sealed with an adhesive film, covered with foil, and incubated at room temperature for 1 hour. Then, 3 μL of donor mix was further added into the wells under subdued light and incubated at room temperature for 1 hour in the dark. The assay plate was read by EnVision with standard AlphaLISA settings. The signal was analyzed and quantitated. Data were analyzed using the GraphPad Prism–based data software.

H358, MIA PaCa2, LXF289 (300269, CLS), OCI-AML-5 (ACC247, DSMZ), SW403 cells (CCL230, ATCC), or AsPC1 cells (CRL1682, ATCC) were trypsinized, counted, and plated at 2,500, 400, 1,250, 1,000, 5,000, or 800 cells/well, respectively, in 96-well ultralow attachment plates (7007, Corning) for three-dimensional (3D) cell viability determination. The day after plating, serial 1:3 dilutions of SOS1 inhibitor RGT-018 were made to evaluate the concentration-dependent effect on cell viability. About 0.5 μL serial dilutions of RGT-018 were added in duplicates. After compound treatment for 5 days (SW403 and LXF289 cells), 7 days (H358, OCI-AML-5, and AsPC1 cells), or 14 days (MIA PaCa2), unless specifically mentioned in figure legend, the CellTiter-Glo 3D assay was used to measure cell viability effects of the compound in 3D format. CellTiter-Glo 3D reagent was added to the cell culture medium in each well. All contents were mixed for 5 minutes on an orbital shaker to induce cell lysis. The assay plates were then allowed to incubate at room temperature for an additional 25 minutes to stabilize luminescent signal. The signal was analyzed and quantitated by EnVision. Data were analyzed using the GraphPad Prism–based data software.

For combinational experiments, the combination effect was plotted in a bar graph using representative concentrations of RGT-018 and the combined drug.

Cells were lysed directly in six-well format using cell lysis buffer (9803S, Cell Signaling Technology) and analyzed by Western blot. Proteins were separated by SDS-PAGE and transferred by Trans-Blot Turbo Transfer System (1704150, Bio-Rad) according to standard protocols. Membranes were immunoblotted with antibodies against SOS1 (SC-17793, 1:100, Santa Cruz Biotechnology), SOS2 (SC-393667, 1:100, Santa Cruz Biotechnology), pERK (4370S, 1:1,000, Cell Signaling Technology), tERK (9102S, 1:1,000, Cell Signaling Technology), pAKT (4370S, 1:1,000, Cell Signaling Technology), tAKT (9102S, 1:1,000, Cell Signaling Technology), KRAS (WH0003845M1, 1:500, Sigma-Aldrich), and β-actin (A5441, 1:5,000, Sigma-Aldrich) in 5% BSA in tris-buffered saline with 0.1% Tween 20 blocking buffer. After primary antibody incubation, membranes were incubated with anti-rabbit (7074S, 1:3,000, Cell Signaling Technology) or anti-mouse (7076S, 1:3,000, Cell Signaling Technology) IgG secondary antibody and signals were detected by ChemiDoc MP Imaging System (12003154, Bio-Rad).

The H358 and MIA PaCa2 cells were cultured as described above. The tumor cells were routinely subcultured twice per week by trypsin-EDTA treatment. The cells in the exponential growth phase were harvested and quantitated by using a cell counter before tumor inoculation. After harvesting the cells, a mycoplasma test report was implemented, and cell viability was checked. The mycoplasma test was negative, and the cell viability was ≥90%.

Each mouse was inoculated subcutaneously in the right front flank region with H358 or MIA PaCa2 cells in 0.1 mL of PBS mixed with Matrigel (1:1) for tumor development. Randomization was started when the mean tumor size reached approximately 300 mm³. The date of randomization was denoted as day 0. Each group consisted of 8 to 12 tumor-bearing mice. Dosing was implemented on the next day after randomization, denoted as day 1. The testing articles were administered to the mice according to a predetermined regimen.

After 3 days of treatment according to the dosage regimen, tumor-bearing mice were anesthetized and humanely euthanized at 8 or 24 hours after the last dose on day 3. Harvested tumors were frozen immediately in liquid nitrogen and stored at −80°C for characterization. The snap-frozen tumor samples were weighed and lysed with cell lysis buffer supplemented with phenylmethylsulfonyl fluoride (1 mmol/L), phosphatase inhibitor, and protease inhibitor cocktail (1 tablet/10 mL) into 100 mg/mL. TissueLyser was used to crush the tissues. Protein concentrations of the collected protein samples were quantified by using the bicinchoninic acid assay kit. The tumor tissue lysates were used for Western blot analysis as described above.

The protocols or procedures involving the care and use of animals in all the in vivo studies were reviewed and approved by the Institutional Animal Care and Use Committee of Crown Bioscience (Taicang, Jiangsu, China) prior to execution. During the study, the care and use of animals were conducted in accordance with the regulations of the Association for Assessment and Accreditation of Laboratory Animal Care.

The H358 or MIA PaCa2 tumor cells were maintained in vitro as described above. The murine CRC CT26 cells were cultured in RPMI1640 medium supplemented with 10% FBS. The cells were harvested in the exponential growth phase, and the mycoplasma test was implemented before tumor inoculation as described above.

For tumor inoculation of H358, MIA PaCa2, or CT26 cells, respectively, 6- to 8-week-old female NOD/SCID mice, BALB/c nude mice, or BALB/c mice (Shanghai Lingchang Biotechnology Co., Ltd, Shanghai, China, or GemPharmatech Co., Ltd, Nanjing, China) were used. Each mouse was inoculated subcutaneously in the right front flank region with H358 tumor cells (5 × 10⁶ cells/mouse), MIA PaCa2 tumor cells (3 × 10⁶ cells/mouse), or CT26 tumor cells (3 × 10⁵ cells/mouse) in 0.1 mL of PBS mixed with Matrigel (1:1) for tumor development.

Animals were randomly allocated to appropriate study groups when the mean tumor volume reached a certain size. Randomization was performed based on the “matched distribution” method/“stratified” method using the multitask method (StudyDirector software, version 3.1.399.19)/randomized block design. Dosing was implemented right after randomization. Dosing as well as tumor and body weight measurements were conducted in a laminar flow cabinet. RGT-018 was formulated in 20% PEG400, 10% solutol HS 15, and 70% [0.5% methylcellulose (MC) with distilled water] solution. Trametinib was formulated in 5% DMSO and 95% (0.5% MC with distilled water) solution. Sotorasib was formulated in 1% Tween 80, 2% (hydroxypropyl)methyl cellulose, 97% water; pH 2.4. All formulations were prepared once per week, and dosing solutions were stored protected from light at 4°C.

Mice were orally administered with vehicle, RGT-018, trametinib, and sotorasib orally at the indicated doses and schedules. Tumor volumes were measured three times per week after randomization in two dimensions using a caliper, and the volume was expressed in mm³ using the formula: V = (L × W × W)/2, where V is tumor volume, L is tumor length (the longest tumor dimension), and W is tumor width (the longest tumor dimension perpendicular to L). Body weights and tumor volumes were measured by using Study Director software. Tumor growth inhibition (TGI) analysis or tumor regression was applied at the endpoint. TGI% was an indication of antitumor activity and expressed as TGI (%) = 100% × (1 − ∆T/∆C); ∆T/∆C = (Ti − T0)/(Ci − C0) × 100%, where Ti and Ci are the mean tumor volumes of the treatment and vehicle groups on the measurement day; T0 and C0 are the mean tumor volumes of the treatment and vehicle groups on day 0. Tumor regression (REG) was calculated when the average tumor volume of the final treated tumors was less than the initial treated tumor volumes using the equation: REG% = [(T0 − Ti)/T0] × 100%, if T0 > Ti, where Ti is the mean tumor volume of the treatment group on the measurement day and T0 is the mean tumor volume of the treatment group on day 0.

The protocols or procedures involving the care and use of animals in the in vivo efficacy studies were reviewed and approved by the Institutional Animal Care and Use Committee of Crown Bioscience prior to execution. The care and use of animals were conducted in accordance with the regulations of the Association for Assessment and Accreditation of Laboratory Animal Care.

To knock out the SOS1 or SOS2 gene using the CRISPR/Cas9 approach, SOS1- or SOS2-targeting sgRNAs were designed and assembled according to previous studies. Lenti-CRISPR/Cas9-SOS1-sgRNA 1, 2 and lenti-CRISPR/Cas9-SOS1-sgRNA 1, 2, 3 constructs were generated by cloning each sgRNA into the single guide RNA scaffold after purifying digested pLentiCRISPR v2 vector in GenScript Biotech Corporation (Shanghai). The sequences of all sgRNAs are shown in Supplementary Table S2.

For transfection experiments, HEK293T/17 (CRL11268, ATCC) cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 37°C and 5% CO₂. Cells were transfected with lentiviral backbone constructs (described above), a packaging vector (psPAX2; #12260, Addgene), and an envelope vector (pMD2.G; #12259, Addgene) using Lipofectamine 2000 (11668027, Thermo Fisher Scientific) based on the manufacturer’s protocol. The supernatant was collected 48 and 72 hours after transfection, followed by centrifuging the media-containing virus at 200g for 5 minutes and filtering with a 45-μm filter (SLHVR33RB, Millipore).

MIA PaCa2 cells were cultured as described above. For in vitro transduction, MIA PaCa2 cells were transduced with viral solution for 24 hours. The medium was replaced by fresh DMEM supplemented with 0.3 μg/mL puromycin (A1113803, Gibco) every 2 days after transduction. The transduced cells were selected and established for in vitro validation.

The established MIA PaCa2/Cas9 stable cells were used for validation via sequencing and Western blot. The procedure of Western blot is as described above.

To generate the mutant KRAS-overexpressing (OS) cell lines, KRAS wild-type (WT) or mutants were designed and assembled into the pLVX-CMV-PURO vector in GenScript Biotech Corporation (Shanghai). The sequences of WT or mutant KRAS are shown in Supplementary Table S3.

HEK293T/17 (CRL11268, ATCC) cells were transfected as described above. The supernatant was collected 48 and 72 hours after transfection. Targeted HEK293T/17 or H358 cells were transduced with viral solution. The transduced cells were selected and KRAS WT or mutant stable cell lines established for in vitro validation via Western blot.

Male SD rats (6- to 8-week-old) were randomly divided into two groups for i.v. single dose (2 mg/kg) and oral single dose (5 mg/kg) administration, with three animals in each group. Rats in the i.v. group had free access to food and water, and oral-group rats were fasted overnight and fed at 4 hours following dosing. RGT-018 was prepared in 5% DMSO, 10% solutol HS15, and 85% (20% HP-β-CD in water) solution for i.v. dosing and in 20% PEG400, 10% solutol HS15, and 70% [0.5% methyl cellulose (MC) in deionized water] solution for oral dosing. The animals were restrained manually, and serial bleeding was performed via the jugular vein at 0, 0.033 (i.v. only), 0.083, 0.167 (i.v. only), 0.25, 0.5, 1, 2, 4, 8, and 24 hours. Approximately 150 μL of blood samples was collected in K2EDTA tubes at each time point. The samples were kept in wet ice and then centrifuged to obtain plasma (2,000g, 4°C, 5 minutes) within 15 minutes after sampling. The plasma samples were further processed in LC/MS-MS to determine the concentration of the compounds. Non-compartmental analysis was performed to obtain PK parameters using WinNonlin 7.0 (Certara, USA). Oral bioavailability was calculated using (AUC_oral/dose_oral)/(AUC_i.v./dose_i.v.) × 100.

Male beagle dogs (10–12 kg) were randomly divided into two groups for i.v. single dose (1 mg/kg) and oral single dose (5 mg/kg) administration, with three animals in each group. Dogs in the i.v. group had free access to food and water and oral group dogs were fasted overnight and fed at 4 hours following dosing. RGT-018 was formulated in 5% DMSO, 5% Kolliphor HS15, 90% saline for i.v. dosing and 0.5% MC (400 cP) in water for oral dosing. The animals were restrained manually. Serial bleeding was performed via the cephalic vein at 0, 0.0333 (i.v. only), 0.083, 0.167 (i.v. only), 0.25, 0.5, 1, 2, 4, 8, and 24 hours. Approximately 500 μL of blood samples was collected in K2EDTA tubes at each time point. The samples were kept in wet ice and then centrifuged to obtain plasma (2,000g, 4°C, 5 minutes) within 15 minutes after sampling. The plasma samples were further processed in LC/MS-MS to determine the concentration of the compounds. Non-compartmental analysis was performed to obtain the PK parameters. Oral bioavailability was calculated as mentioned above.

Statistical analyses were performed on the mean values of biological replicates in each group. All results were expressed as mean ± standard deviation or SEM, as indicated in the figure legends. For comparison of two groups, an unpaired t test was used. The significance (P value) of the difference among three or more groups was evaluated using one-way ANOVA for multiple comparisons. The statistical tests used are indicated in the figure legend. Values of P < 0.05 were considered statistically significant.

Atomic coordinates and structure factors for the cocrystal X-ray structure of RGT-018 with SOS1 have been deposited in the PDB with the code 8XJJ.

RGT-018 was identified as a potent and selective SOS1 inhibitor (Fig. 1A). The synthesis of RGT-018 is described in detail in Supplementary Fig. S1A and in the patent WO2022140427A1 (example 75; ref. 41). The cocrystal structure of SOS1/ RGT-018 complex was determined at 2.1 Å resolution, in which RGT-018 was shown to occupy the pocket partially overlapping with the RAS interaction site on SOS1 (Fig. 1B, left), revealing that the inhibition mechanism of RGT-018 is disruption of SOS1:KRAS interaction directly. RGT-018 binds to SOS1 through a very tight and highly ordered H-bond network involving surrounding amino acids and water molecules, among which H905 plays a central role (Fig. 1B, right). Three water molecules (W1, W2, and W3) were involved in mediating the protein–small molecule interactions and an additional water molecule (W4) mediated interactions between H905 and E909 by H-bonds, while H905 formed one more H-bond with N879 (Fig. 1B, right). Several nonpolar contacts were also identified between RGT-018 and SOS1, including interaction between side chain of Y884 and the morpholine moiety of the inhibitor, π-π stacking (F890 to the benzene ring moiety and H905 to RGT-018 core), cation-π (Y884 to RGT-018 core; Fig. 1B, right).

The critical role of H905 observed in the complex structure helped explain the selectivity of RGT-018 for SOS1 over SOS2. Despite SOS1 and SOS2 sharing high sequence similarity, the corresponding residue for SOS1-H905 is V903 in SOS2 (28). The replacement of histidine with hydrophobic amino acid valine not only would greatly weaken the interaction with the core of RGT-018 (Fig. 1B, right) but also would lose the highly organized H-bond network involving the four structured water molecules and the amino acids E909 and N879. As a result, this single amino acid difference would result in disruption of the local conformation of the binding pocket and lack of RGT-018 inhibitory activity for SOS2 (Fig. 1B, right).

Biochemical activity characterization of RGT-018 was carried out using SOS1 and SOS2 proteins with KRAS^G12D or KRAS^G12C mutants. RGT-018 was identified as a potent inhibitor binding to the catalytic site of SOS1, thereby blocking the interaction of SOS1 with KRAS^G12D (IC₅₀ = 8 nmol/L; Fig. 1C). Upon replacement of SOS1 with SOS2, RGT-018 lost its ability to inhibit the binding of SOS2 with KRAS^G12D (>5,000-fold selectivity), indicating that RGT-018 was a highly potent SOS1-specific inhibitor. In the SOS1/KRAS^G12C binding assay, RGT-018 also displayed strong inhibitory activity (IC₅₀ = 19 nmol/L). Based on the KRAS^G12D and KRAS^G12C results above, we believe it is reasonable to extrapolate that RGT-018 can block activities of multiple KRAS-mutant proteins. RGT-018 was tested in a panel of 330 kinases by Kinome-Wide Panel service at 1 μmol/L, and no hits were identified (Fig. 1D; Supplementary Table S4).

The cellular activity of RGT-018 was evaluated in KRAS^G12C cancer cell lines. A dose-dependent reduction of phosphorylation of the ERK (pERK) levels was observed in NSCLC H358 and PAC MIA PaCa2 cells harboring KRAS^G12C mutation, with relative IC₅₀ values of 10 and 9 nmol/L, respectively (Fig. 1E). Consistent with this result, RGT-018 inhibited the 3D growth of H358 and MIA PaCa2, with relative IC₅₀ values of 36 and 44 nmol/L, respectively (Fig. 1F). The data showed a clear correlation between the signaling pathway and growth inhibition by RGT-018 in KRAS-driven cancer cells.

SOS1 is uniformly expressed across all tumor types, thus an SOS1 inhibitor could be broadly applicable in KRAS-driven indications (15, 16, 32). Toward this end, the antiproliferation activity of RGT-018 was further investigated across a broader panel of cancer cell lines driven by different KRAS mutations, SOS1 or EGFR mutations. As expected, RGT-018 caused 3D cell growth inhibition in all KRAS G12, G13, SOS1, and EGFR mutant cell lines with relative IC₅₀ values in the range of 30 to 633 nmol/L (Fig. 1G). Unsurprisingly, RGT-018 did not inhibit the 3D growth of A375 melanoma cells that harbor KRAS WT and the BRAF^V600E mutation and are known to signal through a KRAS-independent pathway (Fig. 1G). Most interestingly, RGT-018 potently suppressed cell growth of LXF289 and OCI-AML-5 cells harboring SOS1^N233Y mutation (Fig. 1H and I), indicating that SOS1 inhibition is an effective therapeutic strategy in SOS1 mutant cancer, in addition to KRAS-driven cancers. Both our findings and that of others suggest that the sensitivity of SOS1 inhibitors is driven by not only KRAS mutations but also other genetic alterations, e.g., EGFR and SOS1 mutations (32, 42). RGT-018 was also tested in the nontumorigenic rat small intestinal epithelial cells (IEC6) and no appreciable general cytotoxicity was observed at 15 μmol/L, suggesting a favorable therapeutic window for RGT-018 (Supplementary Fig. S1B).

To demonstrate that RGT-018 specifically inhibits SOS1 signaling at the cellular level, SOS1 and SOS2 knockout MIA PaCa2 cells was generated using the CRISPR-Cas9 technology (22, 43). In SOS1 knockout cells, SOS1 was pronouncedly suppressed without interfering with SOS2 at protein levels (Supplementary Fig. S1C). The reverse has been true for SOS2 knockout cells. As expected, in SOS1 knockout cells, no effect on proliferation was observed following treatment with RGT-018, suggesting that RGT-018 exhibited its antiproliferation activities by specifically inhibiting SOS1 (Supplementary Fig. S1D). In the SOS2 knockout cells, the antiproliferation effect of RGT-018 was enhanced, consistent with early results that SOS1 and SOS2 double-knockout mice have a more severe phenotype than individual SOS1 and SOS2 knockout mice (21, 24, 25). Collectively, these data supported the notion that RGT-018 is a potent and selective SOS1 inhibitor and showed a broad growth inhibition in KRAS-driven cancer cells in vitro.

We next advanced RGT-018 into in vivo mouse models to investigate drug exposure and pharmacodynamic (PD) properties along with its antitumor activity.

To evaluate the PD response of RGT-018 and to correlate drug exposure with target inhibition, RGT-018 was administered via oral gavage once daily (QD) at doses of 12.5, 25, 50, or 100 mg/kg for 3 days in H358 model. RGT-018 exposure in mouse plasma and the levels of pERK were determined. At 8 hours following dose, RGT-018 single agent at 12.5, 25, 50, and 100 mg/kg doses partially inhibited the level of pERK within a range of 20% to 60% (Fig. 2A). The free plasma concentrations of RGT-018 at 12.5, 25, 50, and 100 mg/kg were 6, 31, 68, 417 ng/mL, respectively (Fig. 2A). At 24 hours following dose, pERK levels returned to baseline in H358 tumors, whereas drug concentrations were minimal in plasma (Supplementary Fig. S1E). In a similar experiment, RGT-018 was administered over a range of doses to MIA PaCa2 xenograft–bearing mice. A reduction of pERK levels in the range of 30% to 55% was observed in MIA PaCa2 tumors (Fig. 2B). In addition, RGT-018 single agent partially inhibited the level of pAKT within a range of 30% to 50% at 8 hours following dose (Supplementary Fig. S1F). Together, these data indicate that RGT-018 partially inhibited pERK and pAKT levels over a period of up to 8 hours, comparable to the ∼50% inhibition reported for a known SOS1 inhibitor BI-3406 (32).

Moreover, the efficacy of RGT-018 was evaluated in H358 and MIA PaCa2 xenograft models. RGT-018 QD treatment for 28 days resulted in prolonged dose-dependent TGI as compared with vehicle control (Fig. 2C and E). In the H358 model, RGT-018 at 25, 50, and 100 mg/kg partially inhibited tumor growth by 20%, 34%, and 66%, respectively. In the MIA PaCa2 model, RGT-018 at 25 and 100 mg/kg partially inhibited tumor growth by 27% and 52%, respectively. In these studies, RGT-018 was well tolerated with minimal body weight changes at all dose levels in mice (Fig. 2D and F). The results were consistent with the effect on cell proliferation in the two cell lines in vitro. Thus, oral administration of RGT-018 demonstrates dose-dependent antitumor efficacy over a well-tolerated dose range.

Considering the overall partial inhibition in pERK levels and tumor growth by RGT-018 both in vitro and in vivo, it is important to explore combinational strategies of RGT-018 with other targeted therapies. Previous work showed that many cancer models develop adaptive resistance to MEK inhibitors, often due to the reactivation of SOS1 (32, 44, 45). Therefore, it has been reported that inhibition of SOS1 and MEK at the same time provided an effective therapeutic approach to treat KRAS-driven tumors (32). Consistent with these findings, the combination of RGT-018 with the MEK inhibitors trametinib or cobimetinib yielded strong combinational antiproliferation effects in H358, MIA PaCa2, and AsPC1 (PAC, KRAS^G12D) cells in vitro (Fig. 3A, C, and E). Meanwhile, pERK levels were further suppressed when RGT-018 was treated together with MEK inhibitor trametinib in both H358 and MIA PaCa2 cells when compared with single-agent treatment of either RGT-018 or trametinib alone (Fig. 3B and D).

Based on these promising cellular data in vitro, combination treatment of RGT-018 with trametinib was assessed in H358 tumor–bearing mice for an in vivo PD study. Treatment groups included RGT-018 only (12.5, 25, 50, or 100 mg/kg, QD), trametinib only [0.1 mg/kg, twice daily (BID)], and the combination of trametinib and RGT-018. The tumor-bearing mice were treated for 3 days, and tumors were collected after 8 hours of the last dose for pERK analysis. As expected, trametinib treatment resulted in a 75% reduction of pERK levels. Combination of RGT-018 and trametinib exhibited deeper pERK suppression, with a range of 85% to 90% reduction in a dose-dependent manner when compared with RGT-018 or trametinib alone (Fig. 3F).

We also evaluated the in vivo efficacy of RGT-018 plus trametinib in both H358 and MIA PaCa2 xenograft models as well as in a CT26 syngeneic model with KRAS^G12D mutation. The combination of RGT-018 50 mg/kg QD with trametinib 0.1 mg/kg BID was well tolerated (Supplementary Fig. S2A) and caused substantial regressions in the H358 tumor–bearing mice (Fig. 3G). The mean tumor volumes of vehicle control, RGT-018 (50 mg/kg, QD), trametinib (0.1 mg/kg, BID), and combination groups were 807, 708, 352, and 133 mm³, respectively, after treatment for 28 days (day 35), with TGI or REG values of 15% (TGI), 69% (TGI), and 9% (REG) (Fig. 3G; Supplementary Fig. S2B). Furthermore, following combination treatment, slower tumor regrowth was observed after drug withdrawal for at least 7 days. Similar results were obtained in MIA PaCa2 xenografts and CT26 syngeneic models. The effect of RGT-018 and trametinib combination therapy was significantly stronger when compared with both monotherapies, with sustained TGI for 7 to 14 days following drug withdrawal in both immunocompromised and immunocompetent mice (Fig. 3H and I; Supplementary Fig. S2C and S2D). All combination treatments were very well tolerated as reflected by minimal impacts on body weight changes (Supplementary Fig. S2E and S2F). The data in the CT26 syngeneic model indicate that there is more to understand about the cross-talk between RGT-018 and the immune system.

In summary, both the in vitro and in vivo results have suggested that the combination of SOS1 and MEK inhibitors enabled more profound pERK inhibition, resulting in tumor regression in KRAS-driven cancer models at well-tolerated doses.

The beneficial effect of RGT-018 described above as a single agent or in combination with MEK inhibitors was further extended to the FDA-approved KRAS^G12C inhibitors sotorasib and adagrasib (5, 6, 9, 10, 46). As expected, the combination of RGT-018 with sotorasib yielded strong combinational antiproliferation effects in H358 and MIA PaCa2 cells in vitro (Fig. 4A; Supplementary Fig. S3A) and produced more robust and more prolonged suppression of pERK as compared with sotorasib treatment alone in these cells (Fig. 4B; Supplementary Fig. S3B). A similar antiproliferation effect was observed when RGT-018 was combined with adagrasib in H358 cells (Fig. 4C).

In addition to the cellular data, the H358 tumor–bearing mice were treated with sotorasib only (10 mg/kg, QD) or with the combination of sotorasib and RGT-018 (12.5, 25, or 50 mg/kg, QD) for 3 days. pERK levels were evaluated. RGT-018 combined with sotorasib resulted in a more prominent reduction of pERK levels over a period of 8 hours compared with sotorasib treatment alone (a range of 75%−95% reduction vs. 60% reduction, Fig. 4D).

Furthermore, the efficacy of RGT-018 in combination with sotorasib was tested in H358 xenograft model. After treatment for 27 days (day 34), the mean tumor volumes of vehicle control, RGT-018 (50 mg/kg, QD), sotorasib (10 mg/kg, QD), and combination groups were 675, 487, 118, and 20 mm³, respectively. The inhibition of tumor growth was 35% (TGI), 17% (REG), and 82% (REG; Fig. 4E and F; Supplementary Fig. S2G). The combination of RGT-018 with sotorasib caused substantial regressions in H358 tumor–bearing mice. Following combination treatment of RGT-018 and sotorasib, slow regrowth of tumors was observed after drug withdrawal for at least 14 days (Fig. 4E). The combination treatment was very well tolerated (Supplementary Fig. S2H).

In conclusion, RGT-018 and KRAS^G12C inhibitors combination led to more prominent pERK inhibition and tumor regression in KRAS^G12C-driven cancer models.

To identify other targeted therapies capable of enhancing the response to RGT-018, compounds targeting key cancer-related signal transduction nodes were investigated in KRAS-driven cell lines to identify combinational effect with RGT-018 (1, 45).

Given the relevance of the EGFR pathway in the pathogenesis of NSCLC, CRC, and PAC, combinations of RGT-018 and different EGFR inhibitors were examined in cancer cell lines with varying sensitivity to RGT-018 as a monotherapy (43, 45). The combination of RGT-018 with EGFR small-molecule inhibitor afatinib or gefitinib improved the antiproliferation effect in H358, LoVo (CRC, KRAS^G13D), and PC9 (NSCLC, EGFR^ex19del) cells (Fig. 5A; Supplementary Fig. S3C–S3E). Additionally, the combination resulted in deeper suppression of pEGFR, pERK, and pAKT when compared with EGFR monotherapy in H358 and PC9 cells (Fig. 5B; Supplementary Fig. S3F). The combination of RGT-018 and EGFR antibody cetuximab improved antiproliferation effect compared with either single agent in SW403 (CRC, KRAS^G12V) cells (Fig. 5C). These data confirmed the participation of EGFR in SOS1/KRAS signaling and suggested a combinatorial strategy to further inhibit KRAS-mediated signaling and enhance antiproliferation effect in EGFR or KRAS-driven cells.

KRAS signaling is known to mediate cell proliferation at least in part through the regulation of cell cycle (47, 48). We tested the combination of RGT-018 with CDK4/6 inhibitor abemaciclib and indeed observed stronger inhibition of proliferation in H358 and MIA PaCa2 cells (Fig. 5D; Supplementary Fig. S3G). Interestingly, pERK and pRb levels were further suppressed by the combination in H358 cells (Fig. 5E), indicating SOS1 inhibition and cell cycle inhibition is an effective combination strategy.

MRTX1133 was reported to be a potent KRAS^G12D-selective small-molecule inhibitor binding to KRAS^G12D in both active and inactive states (49–51). MRTX1133 induces tumor regression in multiple xenograft models (50, 51). Herein, RGT-018 in combination with MRTX1133 was tested in AsPC1 cells and enhanced antiproliferation activity was observed (Fig. 5F). Previously, SOS1 inhibitors were predicted to augment the effect of KRAS inhibitors targeting the inactive state more effectively (15). This result indicated the possibility that the combination of an SOS1 inhibitor with KRAS^G12D active state inhibitor could bring about additional antiproliferation activities.

Altogether, these data suggested that SOS1 inhibitor RGT-018 enhanced the extent and duration of MAPK pathway inhibition upon combination with EGFR, CDK4/6, or KRAS^G12D inhibitors, in addition to combining with MEK or KRAS^G12C inhibitors. This highlighted SOS1 inhibition as a promising combination option for RTK/KRAS/MAPK pathway inhibitors or cell cycle inhibitors with increased antitumor activity through rationally designed combinations.

Small-molecule KRAS^G12C inhibitors sotorasib and adagrasib have received FDA approval for the treatment of patients with advanced KRAS^G12C mutant NSCLC (7–10, 46). Despite the clinical benefit for many patients, acquired resistance to single-agent KRAS^G12C therapy eventually occurred in most patients (13, 14). It was reported that SOS1 inhibitor BI-3406 plus trametinib might be a useful strategy to overcome acquired resistance owing to the secondary Y96D mutation (33). Other new therapeutic strategies are required to delay and overcome the drug resistance.

To confirm that the observed resistance was due to the emergence of the secondary KRAS mutations and not owing to other unidentified mechanisms, KRAS^G12C/G13D, KRAS^G12C/A59S, KRAS^G12C/Y96D double mutants, and KRAS^G12C itself were introduced in HEK293T cells that lack endogenous KRAS^G12C. Western blot revealed that the levels of KRAS and pERK were elevated when compared with the vector control or parental cells (Fig. 6A). Because sotorasib binds covalently to KRAS^G12C, an electrophoretic mobility shift of drug–KRAS^G12C adduct could be observed (Fig. 6B). As expected, this mobility shift was no longer observed when HEK293T cells expressing KRAS^G12C/G13D or KRAS^G12C/Y96D were treated with 100 nmol/L sotorasib (Fig. 6B), suggesting that the G13D and Y96D mutations abrogated sotorasib binding and covalent modification. Furthermore, sotorasib or adagrasib were unable to inhibit pERK levels driven by G12C/G13D and G12C/Y96D in HEK293T cells (Fig. 6B; Supplementary Fig. S4A). Notably, trametinib at either 1 or 3 nmol/L could reduce pERK level in HEK293T cells expressing KRAS^G12C/G13D or KRAS^G12C/Y96D (Supplementary Fig. S4B).

We also introduced KRAS^G12C, KRAS^G12C/G13D, KRAS^G12C/A59S, and KRAS^G12C/Y96D into the H358 cells that originally harbored KRAS^G12C mutation. Higher KRAS and pERK levels were observed in the OE cells (Supplementary Fig. S4C), consistent with the findings in HEK293T cells. In the antiproliferation assay, H358 cells harboring G12C, G12C/G13D, G12C/A59S, or G12C/Y96D had approximately 5−30 times higher IC₅₀ values compared with parental or vector control H358 cells, upon treatment with sotorasib or adagrasib (Supplementary Fig. S4D and S4E). H358 cells OE G12C or harboring G12C/A59S mutation were less resistant to both KRAS^G12C inhibitors than H358 cells harboring mutation of G12C/G13D or G12C/Y96D. However, trametinib was responsive to H358 cells expressing G12C or harboring G12C/G13D, G12C/A59S, or G12C/Y96D mutation (Supplementary Fig. S4F), aligning with the suppression of pERK levels. To assess possible therapeutic strategies to overcome the secondary mutations and KRAS^G12C amplification that mediate resistance to adagrasib and sotorasib, SOS1 inhibitor RGT-018 was investigated in the H358 OE cells using antiproliferation assay. RGT-018 monotherapy inhibited the growth of H358 parental and vector control cells and the growth of H358 cells with G12C/Y96D (Fig. 6C). In G12C/G13D and G12C/A59S mutant H358 cells, RGT-018 was less sensitive compared with parental and G12C/Y96D cells (Fig. 6C).

We next investigated therapeutic strategies to overcome acquired resistance owing to secondary mutations. Because single-agent MEK inhibitor trametinib was responsive to H358 cells harboring the secondary mutations, we firstly combined RGT-018 with trametinib. In H358 OE cells, RGT-018 and trametinib could inhibit cell proliferation and pERK levels (Fig. 6D and E). The result was consistent with previous literature reports that SOS1 inhibitor BI-3406 plus trametinib could be a useful strategy to overcome acquired resistance owing to the secondary Y96D mutation (33). In addition to MEK inhibitor trametinib, other new combinational strategies were evaluated, including CDK4/6, EGFR, and KRAS^G12C inhibitors. As expected, similar results were obtained when RGT-018 was combined with abemaciclib (Fig. 6F and G) or afatinib (Supplementary Fig. S5A and S5B), with more profound suppression of cell proliferation and pERK levels, suggesting that RGT-018 combined with CDK4/6 or EGFR inhibitors could overcome the drug resistance in H358 OE cells. On the other hand, RGT-018 in combination with KRAS^G12C inhibitors was also investigated. Interestingly, further suppression of cell proliferation and the pERK levels were observed when RGT-018 was treated in combination with adagrasib (Fig. 6H and I) or sotorasib (Supplementary Fig. S5C and S5D) in H358 OE cells. These data indicate that dual SOS1 and KRAS^G12C inhibition are possible treatment strategies to overcome the acquired resistance caused by secondary KRAS mutations (bioRxiv 2023.01.23.525210).

Together, RGT-018 in combination with MEK, KRAS^G12C, EGFR, or CDK4/6 inhibitors could overcome clinically acquired mutations after treatment with KRAS^G12C inhibitors.

KRAS is frequently altered in patients with cancer, however, limited treatments are approved for KRAS-driven tumors (7–10, 46). Despite much literature describing the central role of SOS1 in the developmental and oncogenic signaling pathways, only two SOS1 inhibitors BI 1701963 and MRTX0902 have progressed to the clinic in the United States (29, 31, 35). The preliminary result suggested that BI 1701963 was well tolerated, and stable disease was achieved in seven of the 31 treated patients when administered alone (NCT04111458). MRTX0902 also reached the clinical trial stage as a single agent or in combination with KRAS^G12C inhibitor adagrasib (29). Collectively, these suggest that the safety of SOS1 inhibitor BI 1701963 is overall manageable, and there is more to learn of the ongoing clinical development. Special attention will be paid to the safety profiles of SOS1 inhibitors in both clinical and preclinical studies. In addition, multiple SOS1 inhibitors and proteolysis targeting chimera degraders were reported (17, 18, 27, 30, 31, 34–38). Here, we described a novel, potent, and selective small-molecule inhibitor, RGT-018, which binds to SOS1, thereby blocking protein–protein interaction with GDP-bound KRAS.

RGT-018 suppressed the growth of most KRAS-driven tumor cells in vitro. In contrast to the mutant-specific KRAS inhibitors, targeting SOS1 holds promise for efficacy in tumors, carrying the majority of the mutant KRAS alleles, including the most prevalent G12 and G13 variants (Fig. 1G). RGT-018 inhibited tumor growth and suppressed KRAS signaling in tumor xenografts as a single agent and in combination with MEK or KRAS^G12C inhibitors in vivo. Our findings of RGT-018 combination activities with MEK, KRAS, and EGFR inhibitors are consistent with literature reports (29, 32, 43). Interestingly, RGT-018 robustly suppressed the proliferation of cancer cells with SOS1 mutations as a single agent (Fig. 1H and I). SOS1 mutations occur throughout SOS1 with oncogenic hotspots, such as N233Y (15, 16). Activating mutations in SOS1 are relatively rare in human cancers with around 1% in lung adenocarcinoma and uterine carcinoma (15, 16). Collectively, our data suggest that RGT-018 has the opportunity to treat broader KRAS-driven and SOS1 mutation–driven cancers. RGT-018 may provide an effective therapeutic treatment as a single agent for patients with activating SOS1 mutations.

RGT-018 exhibited overall partial inhibition of pERK level and tumor growth in vitro and in vivo; it can enhance the efficacy of multiple targeted therapies in KRAS-driven cancer models. MEK or KRAS^G12C inhibitors were effective combination partners. The combinatorial effect was observed upon combination of SOS1 with MEK or KRAS^G12C inhibitors, leading to tumor regression in KRAS-driven cancer model at well-tolerated doses. Intriguingly, RGT-018 also inhibited tumor growth in CT26 syngeneic model harboring KRAS^G12D mutation as a single agent or in combination with MEK inhibitor in immunocompetent mice (Fig. 3I). The cross-talk between RGT-018 and immune response remains to be further elucidated. Additionally, RGT-018 in combination with EGFR, CDK4/6, or KRAS^G12D inhibitors resulted in a significant increase of antitumor activity in KRAS-driven cancers in vitro. These findings are also in line with recent reports describing a marked synergy in combining SOS1 inhibition with EGFR or KRAS^G12D inhibition in vitro or in vivo (43, 50, 51). Our data highlight SOS1 inhibitor RGT-018 as a promising combination partner for therapeutic agents directly targeting KRAS, RTK, or downstream MAPK and cell cycle pathways.

Moreover, RGT-018 inhibited the growth of H358 cells exogenously expressing KRAS^G12C/Y96D mutation, an acquired mutation observed in patients treated with the inhibitors targeting KRAS^G12C (Fig. 6D). RGT-018 combined with MEK, KRAS^G12C, EGFR, or CDK4/6 inhibitors further suppressed cell proliferation and pERK levels to overcome drug resistance to KRAS^G12C inhibitors in H358 cells harboring G12C/G13D and G12C/A59S, in addition to the G12C/Y96D mutations (Fig. 6). RGT-018 demonstrated efficacy as a single agent and in combination in tumor cells resistant to KRAS^G12C inhibitors.

Our findings position MEK, KRAS^G12C, EGFR, or CDK4/6 inhibition in combination with SOS1 inhibitors as promising strategies for treating both KRAS mutant tumors and for addressing acquired resistance to KRAS^G12C inhibitors. With the recent breakthrough of KRAS^G12C inhibitors approved for the treatment of patients with advanced NSCLC harboring KRAS^G12C mutation, the median progression-free survival is ∼6 months. SOS1 plus KRAS^G12C inhibition could enhance and extend the antitumor response in KRAS^G12C-driven cancers by addressing intrinsic and acquired resistance. Moreover, there are emerging inhibitors targeting other KRAS mutations, e.g., G12D, G12V, and G13C. RGT-018 may also serve as a backbone combination partner with these next-generation KRAS inhibitors to address the unmet needs of cancers with the broad KRAS mutations. It would be interesting to investigate whether these may be translated to patients in the clinic.

Based on the promising pharmacological profiles of RGT-018, the PK properties of RGT-018 were evaluated in SD rats and beagle dogs. RGT-018 exhibited favorable PK profiles with low clearance and desirable oral bioavailability (F%) in preclinical species (Supplementary Table S5). RGT-018 is an attractive candidate for clinical development.

In summary, we report that RGT-018 is a potent, selective, and orally bioavailable SOS1 inhibitor with desirable drug-like properties for combination with targeted agents to treat various KRAS-driven tumors. Based on the characterization data, RGT-018 has not only potent activity but also the pharmaceutical properties as a candidate for clinical development to treat KRAS-driven tumors.

F. Xiao, K. Wang, Z. Hu, W. Huang, T. Feng, and W. Zhong report a patent for WO2022140427A1 pending. No disclosures were reported by the other authors.

F. Xiao: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. K. Wang: Conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft. X. Wang: Data curation, formal analysis, validation, investigation, methodology. H. Li: Data curation, formal analysis, validation, investigation, methodology. Z. Hu: Data curation, formal analysis, investigation, methodology. X. Ren: Data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft. W. Huang: Data curation, formal analysis, validation, investigation. T. Feng: Data curation, formal analysis, validation, investigation, methodology. L. Yao: Data curation, formal analysis, investigation, writing–review and editing. J. Lin: Supervision, investigation, writing–review and editing. C. Li: Data curation, formal analysis, investigation. Z. Zhang: Data curation, formal analysis, investigation. L. Mei: Data curation, formal analysis, investigation. X. Zhu: Supervision, investigation. W. Zhong: Conceptualization, resources, supervision, investigation, writing–review and editing. Z. Xie: Conceptualization, resources, supervision, investigation, writing–review and editing.

This work was supported by Regor Therapeutics Group. The authors would like to thank Xi Chen and Yanny Nie for project management support and Hongying Xiao for assistance with Western blot and combination data analysis.