Introduction

RAS proteins are members of the classical small-molecule GTPase family and function as molecular switches by alternating between the inactive GDP-bound state and active GTP-bound state [1,2,3]. The activation of RAS is regulated by guanine nucleotide exchange factors such as son of sevenless 1 (SOS1), which catalyzes the exchange of GDP for GTP [4, 5]. The RAS family of proteins comprise three isoforms, namely, HRAS, NRAS, and KRAS. KRAS mutations account for nearly 85% of all RAS-driven cancers [6]. Three predominant KRAS mutations at residues, namely, G12, G13 and Q61, are found in human cancers: G12C, G12D, G13D, and Q61H [7, 8]. Single amino acid substitutions at G12, accounting for the majority (83%) of KRAS mutations, are found predominantly in non-small cell lung cancer (NSCLC), comprising ~85% of lung adenocarcinomas, and have been found in 68%–91% of colorectal adenocarcinomas and pancreatic adenocarcinomas. Additionally, mutant KRAS has been shown to drive the initiation and maintenance of various cancer types, representing an important therapeutic target [6, 9].

Currently, the only direct targeting strategy for KRAS-mutant tumors is realized through covalent binding of the mutant cysteine residue in the Switch-II pocket of the KRASG12C protein [10]. Since the first small-molecule inhibitor targeting G12C was identified [11], effective inhibitors have been under continuous development in recent years [12]. To date, AMG510 (sotorasib) has been approved by the US Food and Drug Administration to target the KRASG12C mutant for the treatment of NSCLC [10]. Another inhibitor in development is MRTX849 (adagrasib), which exhibited great potential in preclinical models and clinical responses in patients with KRASG12C- mutant cancers [13]. Both inhibitors covalently bind to the Switch-II pocket of KRASG12C, kee** the mutant largely in an inactive state, which is accompanied by the attenuation of the MAPK/ERK and PI3K/mTOR/S6K pathways in a KRASG12C-inhibition-dependent manner [10, 14, 15]. However, the degree of dependence on the KRASG12C mutation for tumor development varies across different cancer types, and the cooccurring genetic alterations in KRAS-mutant cancers may reduce the clinical response to KRASG12C inhibitors [16]. Furthermore, acquired resistance is inevitable, and patients develop resistance soon after undergoing KRAS inhibitor treatment [17]. An important strategy to overcome drug resistance is the application of combination regimens. Thus, AMG510- or MRTX849-based combination therapies are being explored [10, 13, 18].

Given the above information, the development of novel KRASG12C inhibitors with improved pharmacological characteristics is an area of intense interest, and in this regard, combination therapy is an effective therapeutic option. In this study, a strategy based on the combination of computer design and traditional medicinal chemistry was used to lower the threshold for initial compound screening and thus increase the hit number. 143D, a novel tetrahydronaphthyridine derivative, was characterized as a highly selective and potent KRASG12C inhibitor that exhibited favorable drug-like properties and antitumor activity. Additionally, 143D induced minimal toxicity and enhanced antitumor activity when combined with inhibitors of EGFR/MEK/ERK signaling. Overall, our findings suggest that 143D may be a promising candidate for further clinical development as an effective KRASG12C inhibitor.

Materials and methods

Materials

143D((S)-2-(4-(7-(8-chloronaphthalen-1-yl)-2-((tetrahydro-1H-pyrrolizin-7a(5H)-yl) methoxy)-5,6,7,8-tetrahydro-1,7-naphthyridin-4-yl)-1-(2-fluoroacryloyl) piperazin-2-yl) acetonitrile, WO2022/170947) was provided by Suzhou AlphaMa Biotechnology Co., Ltd. (Suzhou, China). MRTX1133, AMG510, MRTX849, BI3406, trametinib, SCH772984 and afatinib were purchased from MedChemExpress (Monmouth Junction, NJ, USA). Antibodies specific to MEK1/2, phospho-MEK1/2 (Ser217/221), AKT, phospho-AKT (Ser473), phospho-ERK1/2 (Thr202/Tyr204), ERK1/2, phospho-rpS6 (Ser235/236), rpS6, PARP, cleaved-caspase 3, p21, p27, phospho-RB (Ser807/811), RB, CDK2, β-actin, β-tubulin and GAPDH were all purchased from Cell Signaling Technology (Cambridge, MA, USA).

Cell lines and cell culture

MIA PaCa-2, SW1573, SW1463, NCI-H358, Calu-1, NCI-H1373, BxPC-3, and HPAF-II cancer cell lines were purchased from American Type Culture Collection (Manassas, VA, USA) and cultured according to the instructions provided by the manufacturer. Mouse Ba/F3 cells transfected with mutant KRAS (G12C, G12D, or G12V) or wild-type (WT) KRAS were grown in RPMI-1640 with 10% fetal bovine serum or, in the case of parental Ba/F3 cells, supplemented with 10% WEHI-3–conditioned medium as a source of interleukin-3 (IL-3).

Protein expression and purification

The human recombinant protein KRASG12C (isoform 2, residues 1–169) was expressed in Escherichia coli (ER2566) and then purified by affinity chromatography on a Ni-NTA column and by size-exclusion chromatography. Proteins for mass spectrometry were incubated overnight at 4 °C in buffer containing 20 mM HEPES, pH 7.5; 200 mM NaCl; 5 mM β-mercaptoethanol; and 1 mM EDTA with a five-fold excess GDP; the reaction was stopped by adding 10 mM MgCl2 (final concentration). GDP-loaded proteins were exchanged for low-salt buffer in a 5-ml desalting column (Cytiva, Massachusetts, USA). Compounds were added to freshly purified GDP-loaded proteins and incubated for 10 min at room temperature before mass spectrometry identification. The remaining proteins were frozen in liquid nitrogen and stored long-term at −80 °C.

In vitro cell proliferation assay

Cell proliferation was determined by sulforhodamine B (SRB) assay for adherent cells or 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide assay (MTT) for suspended cells. Briefly, ~24 h after plating in 96-well plates, cells were exposed to compounds in serial dilutions for 120 h. The half‐maximal inhibitory concentration (IC50) values were calculated using nonlinear regression analysis with GraphPad Prism 7 software. At least three independent experiments were performed, and the results represent the mean ± SD.

Western blotting

At the end of treatment, cells were collected and lysed in sodium dodecyl sulfate sample buffer (Meilune, Dalian, China). Equal amounts of cell lysate protein were separated by 10% SDS‒PAGE and transferred onto polyvinylidene fluoride membranes. The membranes were probed with specific primary antibodies and then incubated with HRP-conjugated secondary antibodies. Finally, immunoblots were detected using an ECL solution (Thermo Fisher Scientific, MA, USA) on a Tanon 4600 imaging system (Tanon, Shanghai, China).

Colony formation assay

MIA PaCa-2 and NCI-H1373 cells were seeded at 1000 cells per well in a 12-well plate. Compounds were diluted and added to the cells 18 h after cell adherence. Approximately 7 days later, the medium was discarded, and the cells were fixed with 4% paraformaldehyde, stained with crystal violet, and then photographed.

Cell cycle analysis by flow cytometry

At the end of treatment, cells were harvested by trypsinization, washed with cold phosphate-buffered saline (PBS), and fixed in 70% ethanol at −20 °C overnight. Then, the cells were washed twice with cold PBS and stained with 50 μg/mL propidium iodide and 20 μg/mL RNase A at 37 °C for 30 min. The samples were analyzed by flow cytometry (CytoFLEX S, Backman, Germany), and the data were analyzed using ModFit 5.0 software (BD Biosciences, Mississauga, Canada).

Molecular docking

Molecular modeling calculations were based on crystallographic data for a complex comprising KRASG12C (PDB ID 6UT0) and its inhibitor MRTX849. Docking studies were performed using Maestro of Schrödinger Suites (version 2018-1). The docked poses obtained were analyzed with PyMOL (version 2.3.0).

KRASG12C/SOS1 and KRASG12D/SOS1 binding assays

The homogeneous time-resolved fluorescence (HTRF) assay was performed to measure the interaction between KRASG12C or KRASG12D and the SOS1 protein according to the manufacturer’s instructions (Cisbio, France). Compounds were diluted to 1, 3, 10, 30, 100, 300, and 1000 nM, and their effects on the binding between KRAS and SOS1 were assessed. At the end of the incubation period, fluorescence was detected at 665 nm and 620 nm using a SPARK (TECAN) microplate reader. The ratio of acceptor and donor emission signals of each individual well was calculated as follows: Ratio = OD (665 nm)/OD (620 nm) × 104. The data were analyzed using GraphPad Prism 7 software.

Pharmacokinetic (PK) studies with rats

Compound 143D (10 mg/kg) was orally administered to SD rats (males, n = 3/time point) that had been fasted overnight. Serial blood samples were collected in heparinized tubes before dosing and 1.5, 4, and 8 h after dosing. The plasma was obtained by centrifugation (4500 rpm) at 4 °C for 10 min. Twenty microliters of the supernatant was dissolved in 20 μL of ACN/H2O (1/1, v/v) and analyzed by ultraperformance liquid chromatography (UPLC).

Tissue distribution of the compounds in vivo

Compound 143D or MRTX849 was intravenously administered at a dose of 5 mg/kg to 7- to 8-week-old ICR male mice. Tissue samples were collected 0.083, 0.5, 1, and 2 h after injection. Blood was collected from mouse hearts, mixed thoroughly, placed on ice, centrifuged and separated from plasma within 30 min (4 °C, centrifuged at 8000 rpm for 5 min). The brain, lung, colorectum, pancreas, gallbladder, and bile duct were taken from mice on ice, perfused with saline, rinsed, blotted on filter paper, and finally stored at −80 °C until use in assays. Samples at different time points were analyzed by UPLC, and parameters such as the area under the curve (AUC) of the drug–time graph, maximum concentration (Cmax), time to peak concentration (Tmax), and half-life (T1/2) were fitted and calculated using the pharmacokinetic software WinNonlin for sample metabolism analysis.

In vivo antitumor efficacy

Female Balb/cA nude mice (5–6 weeks old) were purchased from Charles River Laboratories (Bei**g, China). All animal experiments were conducted in accordance with the Institutional Animal Care and Use Committee Guidelines of the Shanghai Institute of Materia Medica Chinese Academy of Sciences. Human cancer cells (from 1.0 × 107 to 3.0 × 107 cells/mL) were prepared in NaCl solution, mixed with an equal volume of Matrigel (BD Biosciences), and then transplanted subcutaneously into the right flank of the mice. When tumor volumes reached approximately 100 mm3, the mice were randomly allocated into vehicle and treatment groups (n = 5/group). The treatment groups were administered 143D once daily, BI3406 twice daily, 143D combined with BI3406, or MRTX849 alone once daily. Tumor volumes were monitored in two dimensions and calculated as (length × width2)/2. Two hours after the last dose, tumor tissues were dissected and analyzed by Western blotting.

Statistical analyses

The data are presented as the means ± SDs or the means ± SEMs, and an unpaired, two-tailed Student’s t test was performed to determine the significance of differences. A P value < 0.05 was considered to be statistically significant, and a P value < 0.01 was considered to be markedly significant.

Results

143D shows selective binding to KRASG12C

To explore novel potent and selective KRASG12C inhibitors, we used a structure-based drug design approach and discovered a promising candidate named 143D (Fig. 1a). The scaffold of 143D was tetrahydronaphthyridine, which differed from that of MRTX849 [19], which consists of tetrahydropyridopyrimidin. We first reported this tetrahydronaphthyridine scaffold as a novel KRASG12C inhibitor after using a completely different chemistry synthesis strategy (WO2022170947) than reported herein. As previously reported [19], 143D and MRTX849 showed similar key polar interactions, including covalent bond formation between the acrylamide β-C and Cys12-S moieties, hydrogen bonds between the acrylamide carbonyl and the amino group of Lys16 and between the pyrimidine-N and the imidazole ring-N residues of His95. The pyrrolidine forms a salt bridge with the carboxyl group of Glu62, and the 8-chloronaphthyl of 143D facilitates lipophilic pocket filling and maintains docking in a pattern similar to that of MRTX849 (Fig. 1b). In addition, optimization of the tetrahydropyrrole molecule in the side chain of MRTX849 to a hexahydro-1H-pyrrolizine molecules allows hexahydro-1H-pyrrolizine-N to form a π-cation bond with the imidazole ring of His95, enhancing the tight linkage with His95, Tyr96 and Gln99 and hel** to ensure that the core structure of 143D remains stable at the Switch-II pocket. Furthermore, an extra N in pyrimidine causes the system to exhibit increased electron deficiency, which may explain why the alkalinity of 143D (pKa = 9.6) is higher than that of MRTX849 (pKa = 8.4).

Fig. 1: 143D shows selective binding to KRASG12C.
figure 1

a The chemical structure of 143D and MRTX849. b The docking model of 143D and MRTX849 with the KRASG12C Switch-II pocket. c 143D, MRTX849 or MRTX1133 disrupts the binding between SOS1 and KRASG12C (top) or KRASG12D (bottom). The data represent mean ± SEM. d The 50 μM KRASG12C protein with the addition of 500 μM MRTX849 and 143D, respectively, for primary mass spectrometry detection. KRASG12C-only protein (top), KRASG12C-MRTX849 (middle), KRASG12C-143D (bottom). The peaks indicate the amount of the corresponding peptides.

Full size image

Next, the inhibitory effect of 143D on the interaction of KRASG12C and SOS1 was measured by HTRF assay. 143D inhibited the binding of KRASG12C and SOS1 with an IC50 value of 21.1 nM, which was comparable to that of MRTX849. As expected, 143D exerted little effect on the binding of KRASG12D and SOS1, whereas the selective KRASG12D inhibitor MRTX1133 inhibited this binding with an IC50 value of 11.6 nM (Fig. 1c). Furthermore, a mass spectrometry experiment was performed to confirm the direct binding affinity of 143D with the KRASG12C protein (Fig. 1d). The binding of the compounds resulted in a significant decrease in the content of the peptides containing the Cys12 site. This finding indicates that similar to MRTX849, 143D directly interacts with the KRASG12C protein.

143D displays potent antiproliferative activity by causing cell cycle arrest and apoptosis

Next, the Ba/F3 cell line transformed with KRAS-G12 mutants or WT KRAS was used as an in vitro screening model for assessing the selective cytotoxicity of the tested compounds. 143D and MRTX849 each showed significant antiproliferative effects on the KRASG12C-transformed Ba/F3 cell line with IC50 values of approximately 10 nM (Fig. 2a). However, Ba/F3 cell lines transformed with KRAS G12D or G12V mutants or WT KRAS were insensitive to 143D, AMG510 and MRTX849, with IC50 values higher than 1 μM (Fig. 2a and Table 1).

Fig. 2: 143D displays selective cytotoxicity in KRASG12C-mutant cells.
figure 2

a Effect of 143D, MRTX849, of AMG510 on the viability of the Ba/F3 cell line transfected with KRASG12C, KRASG12D, KRASG12V or KRASWT after 5 days of treatment. b Activity of 143D, MRTX849, or AMG510 in KRASG12C- and non-KRASG12C-mutant tumor cell lines after 5 days of treatment. c NCI-H1373, MIA PaCa-2, NCI-H358, and SW1463 cells were treated with 143D or MRTX849 for 24 h. The cell cycle distribution was analyzed by PI staining via flow cytometry, and the percentages of cells in the G0/G1 phase (black), S phase (light gray), or G2/M phase (dark red) are shown. The numbers on the vertical coordinate indicate the proportion of cells in each phase of the cell cycle. The data represent mean ± SD (n = 3). d Western blot analyses of molecules regulating the MIA PaCa-2 and NCI-H358 cell cycle after treatment with a concentration gradient of 143D or MRTX849 for 24 h. e Effects on Cleaved-caspase3 and PARP in NCI-H358 cells after treatment with a concentration gradient of 143D or MRTX849 for 48 h.

Full size image
Table 1 Anti-proliferative activity of 143D on cultured cell lines.
Full size table

A panel of tumor cell lines that harbor a heterozygous or homozygous KRASG12C mutation, KRASG12D or KRASWT were used to further assess the activity and selectivity of 143D. 143D inhibited all the tested KRASG12C-mutant tumor cell lines (MIA PaCa-2, NCI-H358, NCI-H1373, SW1463 and Calu-1) regardless of their tissue of origin, with IC50 values ranging from 5 nM to 67 nM (Fig. 2b and Table 1). An exception to this finding was that the SW1573 cell line, which has been verified to exhibit intrinsic resistance to KRASG12C inhibitors [10, 20], was insensitive to 143D, which exhibited an IC50 value higher than 1 μM. Furthermore, 143D showed little antiproliferative activity in HPAF-II cell (carrying the KRASG12D mutation) and BxPC-3 cells (carrying KRASWT). These results indicate that 143D shows not only great potential as a KRASG12C inhibitor but also high selectivity for the KRASG12C mutant.

To reveal the mechanism by which 143D inhibits cancer cell proliferation, we first examined its effect on the cell cycle via flow cytometry. Similar to that with MRTX849, treatment with 143D for 24 h induced concentration-dependent G1-phase cell-cycle arrest in KRASG12C mutant cell lines: MIA PaCa-2, NCI-H1373, NCI-H358, and SW1463 (Fig. 2c). The levels of cell cycle-related proteins were measured in MIA PaCa-2 and NCI-H358 cells treated with 143D or MRTX849. Consistent with the flow cytometry data, it showed that the expression of molecules regulating the cell cycle, such as p-RB (Ser807/811) and CDK2, was decreased, while the expression of p21 and p27 was upregulated by either 143D or MRTX849 treatment (Fig. 2d, Fig. S1a). Then, we analyzed the effect of 143D on cell apoptosis. As shown in Fig. 2e, prolonged 143D treatment, for 48 h, induced significant apoptosis in NCI-H358 cells. 143D increased the expression level of apoptosis-related proteins, including cleaved-caspase 3 and cleaved-PARP in a concentration-dependent manner (Fig. 2e, Fig. S1b). Taken together, these data suggest that 143D exerts a highly selective inhibitory effect on KRASG12C-mutant tumor cell proliferation by inducing cell cycle arrest in the G0/G1 phase and apoptosis.

143D selectively inhibits the constitutive activation of KRASG12C signaling pathways

It has been reported that KRAS-RAF-MEK-ERK and KRAS-PI3K-AKT are the two major signaling pathways in KRAS-mutated tumor cells [21], and these pathways regulate cell proliferation, migration, and differentiation [22, 23]. Therefore, the MIA PaCa-2 pancreatic cancer cell line and NCI-H1373 NSCLC cell line, both harboring the KRASG12C mutation, were analyzed to determine the effect of 143D on KRAS signaling pathways. As expected, 143D, similar to AMG510 and to MRTX849, inhibited KRAS signaling in both a concentration-dependent manner and a time-dependent manner (Fig. 3a, b, Supplementary Fig. S2a, S2b). However, consistent with a previous study [13], 143D preferentially inhibited MEK/ERK signaling compared to its effect on PI3K/AKT signaling in both of these KRASG12C-carrying cell lines. The phosphorylation of MEK and ERK, two key factors downstream of KRAS, were significantly inhibited for at least 24 h, whereas minimal evidence indicated significant and/or persistent inhibition of AKT phosphorylation by 143D.

Fig. 3: 143D selectively inhibits the constitutive activation of KRASG12C signaling pathways.
figure 3

a Western blot analysis of KRAS signaling pathway components in MIA PaCa-2 and NCI-H1373 cells treated with 143D, MRTX849 or AMG510 for 12 h. b Western blot analysis of KRAS pathway targets in MIA PaCa-2 and NCI-H1373 cells treated with 30 nM 143D or MRTX849 for the indicated times. c Western blot analysis of KRAS pathway targets in Ba/F3 cells transfected with KRASG12C or KRASG12D after 143D, MRTX849 or AMG510 treatment for 12 h.

Full size image

To determine whether the sensitivity correlated with the ability of 143D to inhibit KRAS-dependent signal transduction, the Ba/F3 cell line transformed with different KRAS mutants was analyzed. Unexpectedly, 143D, AMG510 and MRTX849 exhibited similar effects on MEK/ERK and PI3K/AKT signaling in the KRASG12C-transformed Ba/F3 cells. The phosphorylation of MEK, ERK, AKT, and S6 was significantly and selectively inhibited by 143D in the KRASG12C-transformed Ba/F3 cell line, whereas 143D exerted little effect on either MEK/ERK or AKT signaling in the KRASG12D-transformed Ba/F3 cell line (Fig. 3c, Fig. S2c). Taken together, these results suggest that 143D selectively inhibits cell proliferation and induces cell cycle arrest in the G0/G1-phase and apoptosis through downregulation of a KRASG12C-dependent signaling pathway.

Plasma pharmacokinetics and tissue distribution of 143D

Pharmacokinetic/pharmacodynamics studies were performed to further investigate the preclinical properties of 143D and pave the way to evaluate its in vivo antitumor efficacy. After a single oral dose of 10 mg/kg 143D, plasma samples were collected from SD rats at different time points. As shown in Table 2, 143D exhibited an efficient elimination half-life (T1/2) of 5.2 h and effective plasma exposure with an AUC of 1588 h * ng/g in the rats. In addition, we investigated the distribution of 143D or MRTX849 in the plasma, brain, lung, colorectum, pancreas, and bile duct tissues in rats after intravenous administration of either drug at 5 mg/kg (Table 3). By comparing the two datasets generated with 143D and MRTX849 outcome data, we found that 143D exhibited prolonged T1/2 values in the plasma, brain, and lung. Moreover, in plasma, the colorectum, and the pancreas, 143D exhibited a higher Tmax value than that of MRTX849. Surprisingly, 143D treatment also led to a nearly 1.4-fold higher Cmax value in the lung than that induced by MRTX849. In particular, the lung and pancreas exhibited high exposure levels of 143D, which were 24- and 6.3-fold greater, respectively, than the systemic exposure level. Furthermore, in the lung, the AUClast and AUCINF_obs of 143D were 2.2- and 4.6-fold higher than that of MRTX849, respectively. More importantly, similar to MRTX849 [24], 143D penetrated the mouse brain, with a brain homogenate AUC level of 868.08 h * ng/g. These results indicated that 143D shows promising pharmacokinetic properties and tissue distribution patterns.

Table 2 Pharmacokinetics of 143D in rats.
Full size table
Table 3 Tissue distributions of 143D in mice.
Full size table

143D inhibits the KRAS signaling pathway and growth of KRASG12C-mutant tumors

In view of the strong antitumor potency in vitro and the promising pharmacokinetic properties of 143D in vivo, we next investigated the antitumor efficacy of 143D in vivo. In all the tested human cancer xenograft models, oral administration of 143D caused dose-dependent inhibition of tumor growth in mice, and drug treatment was tolerated well with no loss of body weight at any dose (Fig. 4a, b, c, Table 4). In the SW1463 model of rectal tumors, 143D showed moderate potency with tumor growth inhibition (TGI) values of 20.3% and 76.9% at doses of 5 and 30 mg/kg, respectively (Fig. 4a, Table 4). Administration of 143D at a dose of 3 or 10 mg/kg significantly inhibited the growth of the MIA PaCa-2 xenograft tumors, with TGI values of 80.6% and 96.5%, and leading to partial regression (PR) of tumor in 1 of 5 models (20%) and in 3 of 5 models (60%), respectively. Complete regression (CR) of tumors was observed in 1 of 5 models (20%) at a dose of 10 mg/kg. Moreover, after treatment with 3 or 10 mg/kg MRTX849, the TGI was 51.3% and 97.2%, respectively. Tumor regression was observed in three of five tumors (60%) only at a MRTX849 dose of 10 mg/kg (Fig. 4b, Table 4). In the NCI-H1373 xenograft model, 143D administered at a dose of 5 mg/kg led to a TGI of 51.7%, and tumor growth was completely inhibited, causing tumor regression in all animals (100%) when administered at the high dose of 30 mg/kg (Fig. 4c, Table 4).

Fig. 4: 143D inhibits KRASG12C-mutant tumor growth by targeting the KRAS pathway.
figure 4

Tumor xenografts derived from SW1463 cells (a), MIA PaCa-2 cells (b), or NCI-H1373 cells (c) after oral administration of 143D, MRTX849 or AMG510 as the indicated dosage once daily (n = 5 mice mean ± SEM). Tumor volumes were determined and mouse body weights were measured every 3 days. d The mice were euthanized 2 h after the final dose; then, tumor tissues were excised, the cells were lysed, and the proteins were analyzed by Western blotting.

Full size image
Table 4 In vivo efficacy of 143D and MRTX849 against KRASG12C mutant tumor xenografts.
Full size table

Next, the repression of KRAS-dependent signaling in tumor tissues by 143D was evaluated by Western blotting (Fig. 4d, Fig. S3a). 143D administered at a dose of 5 mg/kg or 30 mg/kg inhibited the phosphorylation of ERK, MEK, and AKT in SW1463 xenograft tumor tissues. Similar results were observed in a MIA PaCa-2 xenograft tumor model, in which 143D and MRTX849 exhibited significant inhibition of MEK/ERK signaling. In conclusion, these results showed that 143D exhibited outstanding antitumor effects in vivo by repressing KRASG12C-dependent signaling pathways.

The combination of 143D and EGFR/MEK/ERK signaling inhibitors exhibits enhanced antitumor effects

Feedback signaling of the MEK/ERK pathway after KRASG12C inhibition is mediated by receptor tyrosine kinases and ERK pathway target genes. Research has revealed that vertical pathway inhibition may block adaptive feedback signaling that drives the MEK/ERK pathway, leading to sustained RAS pathway suppression and enhanced efficacy of KRASG12C inhibitors in multiple KRASG12C-mutant cancer models [17]. Notably, KRASG12C inhibitors enhance tumor-killing activity when administered in combination with other inhibitors of the KRAS/MAPK signaling pathway [10, 18, 25, 26]. As shown in Fig. 5a, 143D combined with BI3406 (SOS1i), trametinib (MEKi), SCH772984 (ERKi), or afatinib (EGFRi) exhibited inhibitory effects on the MAPK signaling pathway and led to enhanced inhibition of colony formation by MIA PaCa-2 and NCI-H1373 cells to a greater extent than any single drug administered alone (Fig. 5a, Fig, S3b). Although BI3406 administered alone showed little antitumor effect on KRAS mutant tumors at the 2D level, the combination treatment of 143D and BI3406 synergistically suppressed the proliferation of MIA PaCa-2 and NCI-H1373 cells. In addition, the combination of 143D with trametinib led to very pronounced effects by almost completely abolishing colony formation, even when administered at very low doses. Accordingly, compared to the inhibition of p-ERK, p-MEK, or p-AKT activity by 143D mono treatment, when combined with BI3406, trametinib, SCH772984, or afatinib (Fig. 5b, Fig. S4a), 143D showed a much greater effect, indicating the potential of vertical pathway inhibition strategies to enhance KRAS pathway suppression.

Fig. 5: Combined 143D with inhibitors of EGFR/MEK/ERK signaling results in enhanced antitumor effects.
figure 5

a MIA PaCa-2 or NCI-H1373 cells were treated with 143D or a combination of drugs and stained with crystal violet for identifying monoclonal cell populations. 143D (10 nM), BI3406 (10 μM), trametinib (5 nM), SCH772984 (200 nM), afatinib (10 nM). b MIAPaCa-2 cells were incubated with 143D or 143D in combination with a MAPK pathway inhibitor for 12 h. Proteins from whole-cell lysates were analyzed by Western blotting using the indicated antibodies. c MIA PaCa-2 tumor-bearing mice received vehicle, 143D (5 mg/kg, once daily), BI3406 (50 mg/kg, twice daily) or a combination of 143D and BI3406; n = 5 animals per group. Tumor volumes and body weights were presented as the means ± SEMs. **P < 0.05, ***P < 0.001 versus the vehicle; ###P < 0.001 versus either single agent alone. d The mice were euthanized 2 h after the final dose; then, tumor tissues were excised, lysed, and proteins were analyzed by Western blotting. e Schematic diagram showing the combination treatment strategy with 143D and KRAS signaling pathway inhibitors.

Full size image

Consistent with the in vitro data, 143D in combination with BI3406 demonstrated significantly greater antitumor efficacy compared with either drug administered alone to MIA PaCa-2 xenograft mice. BI3406 (50 mg/kg) and 143D (5 mg/kg) administered alone showed moderate antitumor activity with TGI values of 62.1% and 62.7%, respectively, whereas cotreatment substantially delayed tumor growth, with a TGI of 88.1% (Fig. 5c). This cotreatment regimen was generally well tolerated and no additional toxicity was observed (Fig. 5c). The combination markedly reduced the activation of MAPK signaling pathway mediated by, for example, ERK phosphorylation, compared with either treatment alone (Fig. 5d, Fig. S4b). Thus, these results indicated that combinations of 143D with the EGFR/MEK/ERK signaling inhibitors largely increased the response and benefit in reducing KRASG12C-mutant tumors (Fig. 5e).

Discussion

KRAS mutants account for 85% of RAS-mutated cancers and are present in 35% of lung adenocarcinomas [27], with the KRASG12C mutation accounting for 44% of these mutations [1]. Currently, the only available and effective antitumor therapies that directly target KRAS mutants are small-molecule inhibitors that covalently bind to the Switch-II pocket of the KRASG12C protein [28]. Two promising inhibitors targeting KRASG12C, namely, AMG510 and MRTX849, have shown positive effects. Although targeting the Switch-II site with covalently binding small-molecule inhibitors indicates a clear advancement in treatment, the design, optimization and application of these molecules are still needed to increase their antitumor potency and improve their pharmaceutical properties. Furthermore, clinical drug resistance is an unavoidable problem, and KRASG12C inhibitor-based combination therapy may be an effective strategy to overcome it [10, 18, 25, 29, 30]. In the present study, we illustrated that 143D is a highly potent and specific small-molecule KRASG12C inhibitor that shows extensive antitumor activity against human xenograft tumors and an enhanced response when combined with other EGFR-/MEK-/ERK-targeting agents.

Tetrahydropyridopyrimidines have been typically used as the bases for designing KRASG12C inhibitors, such as the well-characterized inhibitor MRTX849 [13, 19]. Our innovative designs realized through structure-based computer simulation and medicinal chemistry synthesis provide new ideas for develo** KRASG12C inhibitors. Molecular modeling showed that 143D, a novel tetrahydronaphthyridine scaffold with a tetrahydropyrrolizine sidechain, in contrast to the methylpyrrolidine-based structure of MRTX849, was more stably bound to the Switch-II pocket of KRAS. Furthermore, an extra N in the pyrimidine residue caused the increased electron deficiency in the system, leading to an alkalinity of 143D (pKa = 9.65), which is higher than that of MRTX849 (pKa = 8.46), which may facilitate intestinal absorption. Thus, these structural modifications of 143D may result in improved pharmacokinetics and intrinsic potency.

Ba/F3 mouse pre-B-cells can grow independently of IL-3 when transformed with an oncogene [31]. Herein, a KRAS-transformed Ba/F3 cell line was used in an alternative approach to phenotype screening for selective inhibitors of KRAS, which to some extent, enabled the influence of the genetic background of different tumor cells to be excluded and improved the sensitivity and accuracy of compound screening. Through KRASG12C-dependent proliferation and signaling assays with the Ba/F3 cell model, 143D was characterized as a novel highly potent and selective KRASG12C inhibitor, which was confirmed by analyzing a panel of KRAS-mutant tumor cell lines. 143D selectively inhibited KRASG12C-mutant tumor cell proliferation by inducing G0/G1 cell cycle arrest and apoptosis. However, 143D was ineffective against cell lines with other types of KRAS mutations or wild-type KRAS, which was consistent with the results observed in studies on other specific KRASG12C inhibitors [10, 13, 25, 32].

In addition, according to our preclinical pharmacokinetic data, oral administration of 143D was rapidly absorbed by rats, and the plasma pharmacokinetic parameters exhibited better plasma exposure (AUC = 1558 h * ng/g) and longer T1/2 (5.21 h) than those of MRTX849. Moreover, in a mouse model, we found that 143D crossed the blood‒brain barrier, in contrast to AMG510, as previously reported [33]. Various tissues in mice were exposed to 143D at significantly higher levels than the associated plasma concentrations. Moreover, exposure levels in the lung and pancreas were 24- and 6.3-fold higher, than the systemic exposure level, respectively. As mentioned above, 143D may show promising therapeutic potential for the clinical treatment of lung or pancreatic cancer with the KRASG12C mutation, even in patients with brain metastasis.

In this study, compared to that between MRTX849 and KRASG12C, the enhanced interaction between 143D and KRASG12C and the improvement in the pharmacokinetic and pharmacodynamic properties of 143D may have led to its superior antitumor efficacy in KRASG12C-mutant xenograft models. In our tested models, a response pattern ranging from tumor growth delay to complete regression was observed in the xenograft panel. SW1463 colon cancer, MIA PaCa-2 pancreatic cancer and NCI-H1373 lung cancer cells were selected as KRASG12C inhibitor-responsive tumor models, and they showed clear dose‒response relationships. At all the tested doses, 143D significantly inhibited the growth of the three xenograft tumors and, at higher doses, caused tumor regression. The dependence of tumor growth and survival on KRASG12C mutations can vary across cancers, and the co-occurring genetic alterations observed in KRAS-mutant cancers may lead to differential efficacy by KRASG12C inhibitors. Nevertheless, the KRAS signaling pathway was inhibited in the tumor tissues of all the tested KRASG12C models.

Exposure of KRASG12C-mutant cancer cells to G12C inhibitors suppressed downstream MAPK pathway activation, but long-term treatments result in feedback activation of MEK/ERK or other pathways [19, 34], and same effects were observed with 143D treatment (Fig. S5). These feedback signals are mediated by the ERK pathway with its target genes and receptor tyrosine kinases, such as EGFR [35, 36]. Vertical pathway inhibition led to sustained RAS pathway suppression and enhanced the efficacy of KRASG12C inhibitors in KRASG12C-mutant cancer models [28, 37, 38]. In the current study, 143D combined with inhibitors of MAPK pathway signaling, such as EGFR, SOS1, MEK, or ERK1/2 signaling, led to robust cytotoxicity mediated by sustained and enhanced target inhibition in several KRASG12C cell lines. We also noted that 143D was very effective when administered in combination with the MEK inhibitor trametinib, with the combination showing great antitumor activity than either treatment alone, even at very low doses. Trametinib is a MEK inhibitor with highly effective targeting, which can easily cause compensational reactivation of KRAS signaling pathways [38, 39]. We speculate that the inhibition of KRAS can attenuate this feedback-induced reactivation, leading to better antitumor effects. SOS1 is an important node in the negative feedback regulation of the KRAS pathway [40]. Consistent with data reported in a previous study [8], our data revealed that 143D exhibited synergistic antitumor activity when administered in combination with BI3406, a selective inhibitor of SOS1, to MIA PaCa-2 pancreatic cancer cells.

Secondary mutations within the Switch-II region of the KRAS gene confer acquired resistance to KRASG12C inhibitors [29]. However, it has been reported that the spectrum of secondary mutations is significantly different between AMG510- and MRTX849-treated cells, and in combination with BI3406, these treatments effectively reversed tolerance caused by secondary KRAS mutations [28]. For example, the G13D, A59S, A59T, and R68M mutations conferred resistance to AMG510 but not to MRTX849. Alternatively, the Q99L secondary mutation led to cell resistance to MRTX849 but not to AMG510. Y96D and Y96S secondary mutations induced resistance to both of these treatments. Considering a crystalline model of GDP-KRASG12C bound to 143D, we speculated that the secondary mutations between 143D and MRTX849 may be similar. The results of our analysis suggest that the clinical combination of 143D with SOS1 inhibitors might eliminate bypass or residual signaling that can hamper efficacy or induce resistance. The intertumoral heterogeneity and extensive feedback signaling network in KRAS-mutant cancers may necessitate more strategies to enhance the antitumor response and overcome resistance in concert with KRAS blockade in the future [41, 42]. In addition to direct and indirect targeting, co-targeting strategies suppressing signaling pathways have achieved excellent results in the treatment of KRAS-mutant tumors. Furthermore, KRAS-related glucose metabolism [43], autophagy [44], and immunotherapy [45] are all possible directions for the development of drugs and treatment regimens.

In summary, 143D, a novel, potent, and highly selective small-molecule inhibitor of KRASG12C, shows significant antitumor activity against KRASG12C-mutant cancer both in vitro and in vivo, with evidence indicating its targeted modulating effects. In particular, the combination of 143D with EGFR/MEK/ERK inhibitors demonstrated enhanced responses. The favorable pharmacokinetic/pharmacodynamic properties of 143D and its considerable efficacy constitute a compelling rationale for further clinical evaluation of this drug in single-agent or rationally directed combination therapies. Further studies are needed to identify additional specific KRASG12C inhibitors and to explore how these can be best combined with anticancer agents in different cancer types to achieve better therapeutic outcomes.