Rat sarcoma (RAS) genes have been recognized as the major oncogenes undergoing mutation for several decades (1, 2). Among the three isoforms (KRAS, NRAS, and HRAS), Kirsten rat sarcoma viral oncogene homolog (KRAS) is the common oncogene in a large percentage of cancers, including pancreatic cancer, non-small cell lung cancer (NSCLC), and colorectal cancer (3–6). Mutations in RAS lead to the dysfunction of its small GTPase activity, preventing it from properly breaking down GTP. The molecule remains in a constant active state that triggers downstream pathways, including the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) pathways, leading to oncogenesis.

Attempts to develop effective agents that inhibit RAS mutations have been a failure for a long time (7, 8). In recent years, with the discovery of a new binding site beneath the effector binding switch-II region in RAS protein, several small-molecule agents targeting the KRAS-G12C single-nucleotide mutation (glycine-to-cysteine substitution at codon 12) have been developed and have shown promising efficacy in clinical trials (9–12). Sotorasib (AMG510) has been approved by the FDA as a second-line treatment for patients with KRAS-G12C mutant NSCLC who have received at least one prior systemic therapy (13, 14). Given that several excellent reviews have summarized the role of RAS signaling in oncogenesis and the advances in RAS inhibitors for anti-tumor therapy, we herein focus on KRAS mutations and summarize the promising new treatment options.

RAS mutations may represent the early onset of tumorigenesis and are essential for tumor maintenance, which has been validated by considerable evidence (15–17). The RAS mutation rates in various cancer types are shown in Supplemental Figure 1 . Different single-base missense mutations result in different amino acid substitutions of the RAS oncogene ( Figure 1A ). KRAS, HRAS, and NRAS are the three most commonly mutated RAS isoforms with varying mutation rates in different cancers (18). KRAS mutations, more than 80% of which are G12 mutations, are frequently found in pancreatic ductal adenocarcinoma (> 90%), colorectal adenocarcinoma (> 40%), and lung adenocarcinoma (approximately 30%). NRAS mutations, which occur less frequently than KRAS mutations, mainly occur at codon 61 and are found in nearly 30% of cutaneous skin melanomas. HRAS mutations occurring at codon 12 or 61 are only found in a small subset of bladder urothelial carcinoma, head and neck squamous cell carcinoma, and thyroid carcinoma (19–21). The top ten predominant substitutions and frequencies with which they occur in the three RAS isoforms according to tissue type in common cancers are shown in Figure 1B . For pancreatic ductal adenocarcinoma and colorectal adenocarcinoma, the predominant amino acid substitution is G12D in KRAS. For lung adenocarcinoma, the predominant amino acid substitution is G12C in KRAS. However, for melanoma, the predominant substitution is Q61R in NRAS.

As the most frequently mutated isoform of the RAS family, KRAS has two splice variants, KRAS4A and KRAS4B, which differ in their fourth exon and encode two different proteins that differ only in their C-terminal membrane-targeting regions (22, 23). KRAS4B is the main mutant isoform in human cancer, whereas KRAS4A is commonly expressed in various cancer cell lines and colorectal cancer (24, 25). Certain mutations in the amino acid sequence of KRAS often result in distinct transformation properties and biological behaviors (26). For instance, KRAS-G12V mutations are associated with worse outcomes than KRAS-G12D mutations in patients with lung cancer. Over the last 30 years, the correlation between biological behavior and specific RAS mutations has remained unclear (27–29). KRAS mutations are significantly associated with poor outcomes in patients with lung cancer (30, 31). However, a recent study suggested that for stages I-III, there was no statistical difference in overall survival (OS) between the mutant- and wild-type-carrying patients with NSCLC (32).

The RAS protein, cycling between inactive and active GDP-bound conformations, comprises three major domains: G-domain, C-terminal, and C-terminal CAAX motifs (33, 34). The G-domain is a highly conserved domain that includes switches I and II, which are responsible for the GDP-GTP exchange (33). The C-terminal region containing the CAAX motif varies considerably among different members of the RAS family. However, this motif is essential for post-translational modification (35). RAS is activated by guanine nucleotide exchange factors (GEFs) and transduces signals to downstream pathways.

KRAS encodes a membrane-bound GTPase that is inactive when bound to GDP and active when bound to GTP. The transition of KRAS to its active state is facilitated by GEFs such as SOS1. Once activated by extracellular stimuli, the active form of KRAS acts as a cellular switch, triggering downstream signaling pathways involved in fundamental cellular processes. Mutations in RAS block the binding of GTP to RAS and cause aberrant activation of downstream pathways ( Figure 1C ). RAS mutations may affect the intrinsic GTPase and GDP–GTP exchange rates ( Figure 1D ) (36). Mutations in KRAS at codons 12, 13, and 61 inhibit the ability of GTPase activation protein (GAP) to stimulate GTP hydrolysis. However, KRAS-G13D displays heightened intrinsic exchange activity compared to the wild-type RAS protein (37, 38). Despite the reduced p120 GAP-mediated hydrolysis rate, KRAS-G12C mutant exhibits almost wild-type intrinsic GTPase activity and has been used to develop covalent inhibitors (39).

Despite over three decades of intensive efforts, no effective regimen to inhibit RAS-driven oncogenesis has been developed because of its inaccessible binding surface and picomolar affinity for GTP/GDP (7, 40). The high affinity of the RAS for cytoplasmic GTP renders competitive inhibition difficult to achieve. The absence of a drug-binding groove on the smooth surface of the RAS poses a challenge for targeted inhibitors. Multiple upstream and downstream regulators of RAS pathways contribute to drug resistance mechanisms and bypass signals, further limiting the effectiveness of combination strategies (41).

These complexities underscore the challenges in targeting RAS mutations. In 2013, with the identification of a new covalent pocket of the KRAS-G12C mutation located beneath the effector-binding switch-II region, Shokat et al. reported a novel strategy for overcoming these difficulties in a mutant-specific targeting manner (42). A series of small-molecule agents could irreversibly bind to the KRAS-G12C mutation and disrupt switch-I and switch-II to bind the mutation in the GDP-bound state, thereby blocking the association with Raf and other downstream tyrosine kinases ( Figure 2 ).

Although covalent inhibitors that directly target specific KRAS mutations exhibit promising efficacy, inhibiting other RAS mutations is challenging. New inhibitors have been developed, regardless of the type of RAS mutation or protein. A multivalent small molecular inhibitor compound 3144 was designed to interact with adjacent sites on the KRAS surface and disrupt interactions between RAS proteins and their effectors (56). Preclinical models showed that compound 3144 was capable of binding to HRAS, KRAS, and NRAS and inhibited RAS signaling. Xenograft models also indicated that 3144 could prevent the growth of RAS-mutant mouse cancer xenografts derived from tumor cell lines and patients. Satchell et al. developed a pan-RAS biologic inhibitor by fusing the RAS-RAP1-specific endopeptidase to the diphtheria toxin, which could irreversibly cleave and inactivate intracellular RAS at picomolar concentrations and terminate downstream signaling and induce tumor shrinkage in mouse xenograft models driven by either wild-type or mutant RAS (57). Furthermore, a compound named cmp4 selectively binds to the Switch II pocket of both HRAS and KRAS proteins with different mutations. By interfering with the binding of RAS to GEFs and Raf effectors, cmp4 effectively reduced the intrinsic and GEF-mediated nucleotide dissociation and exchange processes of the Ras protein, ultimately leading to the inhibition of the mitogen-activated protein kinase signaling pathway and a decrease in cell viability. According to a mathematical model of the RAS activation cycle, cmp4 when combined with cetuximab reduces the proliferation of cetuximab-resistant cancer cell lines. However, the affinity of cmp4 for RAS is unsatisfactory, and this limits its application as an ideal clinical drug (58).

Unfortunately, all these compounds that could function as pan-RAS inhibitors have only been tested in preclinical studies. Given the essential role of RAS in normal cell signaling, it is unclear whether pan-RAS inhibitors are tolerated. Previous studies have revealed that homozygous deletion of KRAS is embryonically lethal in mice (59–61). Therefore, the toxicity of pan-RAS inhibitors should be investigated in future studies. In addition, acquired resistance to RAS inhibitors often prevents further clinical benefits. Awad et al. compared the genomic and histological landscapes of pretreatment samples and those obtained after the development of resistance. Acquired KRAS alterations included G12D/R/V/W, Q61H, R68S, and high-level amplification of the KRAS-G12C allele. Bypass mechanisms involve MET amplification, mutations in NRAS and BRAF, and the oncogenic fusion of ALK and RET. Loss-of-function mutations in NF1 and PTEN have been previously reported. Consequently, new therapeutic strategies are necessary to overcome and delay drug resistance in patients with cancer (62).

Specifically, mutant KRAS not only alters the behavior of cancer cells but also affects various cells in the tumor microenvironment (TME). KRAS activation increases the production of the neutrophil chemokines CXCL1, CXCL2, and CXCL5 (63). The upregulation of intercellular adhesion molecule 1 (ICAM1) by KRAS promotes the recruitment of pro-inflammatory M1 macrophages (in contrast, co-activation of KRAS and MYC increases the recruitment of anti-inflammatory M2 macrophages by releasing CCL9 and IL-23). KRAS-mediated secretion of TGFβ and IL-10 leads to the differentiation of immunosuppressive regulatory T cells (Tregs). It also enhances tumor-infiltrating myeloid-derived suppressor cells (MDSCs) through GM-CSF-dependent and IRF2/CXCL3-dependent mechanisms (64).

Moreover, different co-mutation statuses of KRAS can affect the TME and response to immune checkpoint inhibitors (ICIs). For example, tumors with KRAS/STK11 co-mutations often exhibit deficiencies in CD8+ T lymphocytes and a high abundance of T-regulatory cells in the microenvironment. In contrast, tumors with KRAS/p53 co-mutations tend to have an inflamed TME characterized by a higher number of CD8+ T lymphocytes. This can be attributed to p53 mutations, which tend to increase somatic tumor mutations and potentially lead to the development of tumor neoantigens (65).

A detailed understanding of these pleiotropic effects will facilitate the rational design of curative combination therapies. Leidner et al. reported a patient with metastatic pancreatic cancer who received a single infusion of genetically engineered autologous T-cells targeting mutant KRAS-G12D. This led to a 72% partial response at 6 months according to the currently ongoing Response Evaluation Criteria in Solid Tumors version 1.1. Engineered T cells constitute over 2% of the circulating T cells (66). The occurrence of distinct co-mutations affects the clinical efficacy of immunotherapies. In another study involving 536 patients with KRAS-mutant lung adenocarcinoma, both STK11 and KEAP1 mutations in the presence of a KRAS mutation were associated with poor response rates to anti-PD-L1 inhibitors. Median PFS and OS were significantly shorter for KRAS-mutant/STK11-mutant NSCLC (2.0 and 6.2 months, respectively) than that for KRAS-mutant/STK11-wildtype (4.8 and 17.3 months, respectively; HR 2.04, 95% CI 1.66–2.51, p < 0.0001) varieties. Similarly, patients with KRAS-mutant/KEAP1-mutant NSCLC had lower PFS and OS (1.8 and 4.8 months, respectively) than those with KRAS-mutant/KEAP1-wildtype variety (4.6 and 18.4 months, respectively; HR 2.05, 95% CI 1.63–2.59, p < 0.0001) (67).

In preclinical models, combination treatment with AMG510 caused regression of KRAS-G12C-mutant tumors and improved the anti-tumor efficacy of targeted agents and chemotherapy (9). When combined with immunotherapy, AMG510 induces complete and durable tumor regression. The improved efficacy of the combination therapy may be attributed to increased immune cell infiltration and activation. In preclinical models, the AMG510 monotherapy and combination therapy groups demonstrated a notable increase in CD8+ T cell infiltration, which was not observed in the anti-PD-1 monotherapy group. Additionally, AMG510 treatment increased the infiltration of macrophages and CD103+ cross-presenting dendritic cells, which play vital roles in T-cell priming, activation, and recruitment. Furthermore, the combination of AMG510 and anti-PD-1 therapy promoted the establishment of a memory T cell response and enhanced antigen recognition. Phase I/II clinical trials evaluating the efficacy and safety of sotorasib in combination with PD-1/PD-L1 inhibitors in patients with advanced solid tumors and KRAS-G12C mutations are ongoing (CodeBreaK 100/101).

Preclinical models have also demonstrated that human epidermal growth factor receptor (EGFR) family inhibitors, SHP2 inhibitors, mammalian target of rapamycin (mTOR) inhibitors, and inhibition of CDK4/6 could enhance the anti-tumor activity of MRTX849 and inhibit KRAS-dependent signaling pathways (46). Clinical trials were conducted to evaluate the efficacy and safety of combination therapy of adagrasib with pembrolizumab (a PD-1 inhibitor) or afatinib (an HER family inhibitor) in patients with NSCLC, with cetuximab in patients with colorectal cancer, and with TNO-155 in patients with advanced solid tumors. Preliminary results showed that more than 50 patients were treated with adagrasib in combination with either pembrolizumab (a PD-1 inhibitor) for NSCLC, cetuximab (an anti-EGFR antibody) for colorectal cancer, or TNO-155 (an SHP-2 inhibitor) for NSCLC or colorectal cancer. All the combination therapies were well tolerated by patients (48). A phase I-II clinical trial evaluated the efficacy and safety of adagrasib monotherapy or in combination with cetuximab in heavily pretreated patients with metastatic colorectal cancer and mutant KRAS-G12C. The results revealed that 19% of the 43 evaluated patients in the monotherapy group responded, with a median response duration of 4.3 months and a median PFS of 5.6 months. However, the combination therapy group had a higher response rate (46%), with a median response duration of 7.6 months and a median PFS of 6.9 months (72). A phase II clinical trial evaluated the efficacy of adagrasib in patients with KRAS-G12C-mutant NSCLC previously treated with platinum-based chemotherapy and anti-PD-1 or PD-L1 therapy. The results showed that 48 of the 112 enrolled patients had a confirmed objective response, with a median response duration of 8.5 months and a median PFS of 6.5 months. The median OS was 12.6 months (73).

A phase Ib/II clinical trial to characterize the safety and tolerability of JDQ443 in combination with TNO155, spartalizumab (a PD-1 inhibitor), or TNO155 and spartalizumab in patients with advanced solid tumors and KRAS-G12C mutations is ongoing (NCT04699188) (74). Another phase I trial to assess the safety and preliminary activity of GDC-6036 in combination with atezolizumab (a PD-L1 inhibitor) or erlotinib in patients with NSCLC, cetuximab in patients with colorectal cancer, or bevacizumab in patients with advanced solid tumors is underway (NCT04449874). D-1553 is also the regimen used in clinical trials to assess the anti-tumor activity of combination therapy of RAS inhibitors with other treatments (NCT04585035). However, the results of these studies have not been reported. The combination therapies for RAS inhibitors used in clinical development are shown in Table 1 .

KRAS mutations have long been considered attractive targets for cancer therapy. After decades of effort, KRAS mutations are no longer considered undruggable. KRAS-G12C allele-specific inhibitors exhibit promising efficacy in clinical trials and have the potential to alter the treatment status of RAS-mutant cancers. Sotorasib and adagrasib have shown promising results in inhibiting KRAS-G12C and controlling tumor growth. Disease control was observed in a significant percentage of patients, and tumor shrinkage was also noted. However, some patients developed resistance mechanisms, such as mutations activating RAS or the RAS pathway, which rendered the drugs less effective. Combining KRAS-G12C inhibitors with other targeted therapies, like cetuximab or SHP2 inhibitors, has shown enhanced activity in preclinical studies. Resistance mutations were more frequent in patients with lung or colorectal cancer treated with adagrasib. Multiple types of lesions were identified, including mutations preventing drug binding, non-G12C activation of RAS, KRAS amplification, and activation of other pathway components. The presence of multiple and diverse resistance mechanisms poses a challenge to the efficacy of RAS inhibitors. However, similar mechanisms have been observed in resistance to other targeted therapies, indicating the need for further investigation. Despite these challenges, KRAS-G12C inhibitors have demonstrated clinical benefit and are likely to be useful as second-line treatments for lung cancer. Continued research and development are expected to lead to improved drugs and combination therapies that can enhance tumor-cell death and prevent adaptive resistance. Additionally, a new G12C inhibitor that targets active RAS-GTP is being developed and has shown effectiveness against KRAS-G12C tumor cells with resistance to previous inhibitors.

Even though the inhibition of the RAS pathway, including the MAPK and PI3K pathways, showed poor efficacy after monotherapy, a combinational strategy could be useful to improve efficacy. Patients with KRAS-mutant NSCLC can benefit from immunotherapy, and clinical trials evaluating the efficacy of adoptive cell therapy and cancer vaccines are ongoing.

Agents inhibiting RAS post-translational modifications during development have also been researched. Posttranslational modifications of RAS proteins include palmitoylation and depalmitoylation. Palmitoylation attaches palmitic fatty acids to specific amino acid residues, thereby promoting membrane associations and functionality. Depalmitoylation removes these groups and redistributes RAS proteins to the active membrane sites. Inhibition of depalmitoylation has been proposed to hinder RAS membrane binding and functionality. Other modifications such as phosphorylation, nitrosylation, ubiquitination, and acetylation also regulate RAS localization and function. These modifications are potential targets for the development of anti-RAS drugs; however, their mechanisms of action and therapeutic relevance are still controversial. Further research is required to validate their feasibility and specificity for anticancer therapy (8).

Given the encouraging efficacy of KRAS-G12C allele-specific inhibitors, specific inhibitors may be the most promising therapeutic options. However, in addition to KRAS-G12C, other mutations, such as KRAS-G12D and KRAS-G12V, account for a large proportion of KRAS mutations. Therefore, the development of inhibitors targeting specific RAS mutations to provide personalized medicine may be a future direction. However, according to the presented results, the efficacy of sotorasib differs in NSCLC and colorectal cancer and drug resistance is inevitable (10, 52). In addition, combination therapies involving immunotherapy and other targeted therapies or chemotherapies may be worth exploring. The studies discussed in previous sections have shown promising outcomes when KRAS inhibitors were combined with ICIs or other targeted agents. Further investigations should focus on optimizing the treatment regimens, identifying predictive biomarkers, and understanding the mechanisms underlying the synergistic effects observed in preclinical models. Furthermore, understanding the TME and the role of immune cells in KRAS-mutant cancers is crucial. Exploring the factors influencing immune cell infiltration, activation, and recruitment can help in designing strategies to enhance anti-tumor immune responses. Investigating the mechanisms underlying immunotherapy resistance in KRAS-mutant cancers is an important area for future research. This knowledge can guide the development of novel therapeutic approaches to overcome drug resistance and improve patient outcomes. To address these unresolved issues, developing a comprehensive model that integrates the complex interactions between KRAS signaling, the immune system, and the tumor microenvironment would be valuable. Such a model could help explain the observed heterogeneity in treatment responses and potentially predict personalized treatment regimens and responses. This could also guide the design of clinical trials and treatment strategies. Therefore, exploring combination strategies for patients with distinct tumors is vital.

XW drafted the manuscript. WS and CC helped in the literature search. DL provided administrative and financial support, manuscript revision, and final approval of the manuscript. All authors contributed to the article and approved the submitted version.