KRAS proteins are small proteins with a relatively smooth molecular surface without readily accessible binding pockets, with a high affinity for GDP/GTP and complex downstream pathways (5, 43, 44). Therefore, direct targeting KRAS by small molecule inhibitor was a difficult approach until the discovery of a new pocket beneath the effector binding switch II region of KRAS glycine-to-cysteine amino acid substitutions at codon 12 (KRAS G12C), that has allowed the development of direct KRAS G12C inhibitors (45–47). Initially it was thought that mutation of KRAS led to constitutive activation in its GTP-bound state. However, KRAS G12C presents an intrinsic GTPase activity, not presented in other KRAS subtypes, of importance for the activity and efficacy of the direct KRAS G12C inhibitors (8, 46).

Sotorasib (AMG 510) is an oral covalent KRAS G12C (OFF) inhibitor that irreversibly and selectively binds to the cysteine 12, next to pocket (P2) of the switch II region within KRAS mt, keeping it in the inactive GDP-bound state. It was evaluated in a phase I/II study (CodeBreak 100: NCT03600883) in pretreated KRAS G12C mt solid tumors (47, 48). At the 960 mg once-daily dose selected for phase II in patients with metastatic NSCLC (N = 126), the ORR was 37.1% and Disease Control Rate (DCR) was 80.6%. The median duration of response was 11.1 months, the median time to objective response was 1.4 months, with a median PFS 6.8 months (95% confidence interval, 5.1 to 8.2) and a median overall survival of 12.5 months (95% confidence interval, 10.0 to could not be evaluated). The activity of sotorasib was observed across a spectrum of prevalent co-occurring mutations such as STK11, KEAP1 or TP53 as well as different PD-L1 expression or TMB levels. However, these exploratory analyses were not statistically powered, subgroup sample sizes were small, and therefore future prospective studies are warranted to identify subgroups of patients who may benefit differently from sotorasib therapy. Treatment-related adverse events (TRAEs) occurred in 69.8% patients, including grade 3 events in 19.8%. Most common adverse events related to sotorasib were gastrointestinal side effects such as diarrhea (31.7%) and nausea (19%) as well as low-grade hepatic toxicity like alanine aminotransferase (ALT) and aspartate aminotransferase increase (AST) (each 15.1%). No fatal TRAEs were reported. Patients with active brain metastases were ineligible, so the efficacy of sotorasib in the treatment of patients with central nervous system metastases is unknown (48–50). Based on these results, Sotorasib was granted breakthrough designation by the U.S. Food and Drug Administration (FDA) for the treatment of adult patients with KRAS G12C mt locally advanced or metastatic NSCLC who have received at least one prior systemic therapy, becoming the first targeted therapy approved for advanced NSCLC KRAS G12C mt (50). The global phase III trial, CodeBreak 200 (NCT04303780), comparing sotorasib with docetaxel in patients with mt KRAS G12C NSCLC is ongoing, as well as different clinical trials are evaluating sotorasib in combination therapies (CodeBreaK101; NCT04185883) with the aim to identify patients who may benefit from sotorasib regimens in the context of first-line treatment ( Figure 1 ).

Adagrasib (MRTX849) is another oral covalent KRASG12C inhibitor that irreversibly and selectively binds KRAS G12C in its inactive GDP-bound state. It was evaluated in a phase I/II study (KRYSTAL-1; NCT03785249) in pretreated patients with advanced solid tumors. At dose of 600 mg twice daily, of 51 evaluable patients with NSCLC, 45% had ORR and DCR was 96% with 8.2-month median duration of response. Regarding safety for all patients treated at the 600mg twice-daily dose (n = 110), grade 3 or 4 TRAEs occurred in 30% of patients; the most commonly reported (>5%) grade 3 or 4 TRAEs were fatigue (6%) and increased AST/ALT (5%). Two fatal TRAEs due to pneumonitis and cardiac failure were reported (51).

It has been also presented a preliminary analysis examining co-mutations with KRAS G12C. Patients who had KRAS G12C and STK11 co-mutations experienced an ORR of 64%, without apparent trends with KEAP1 or TP53, although the number of samples was small. Moreover, preclinical data and early phase studies indicate that adagrasib can penetrate the brain and cerebrospinal fluid (11, 51–53), although more data are needed to determine the brain activity of adagrasib.

A phase III trial (KRYSTAL-12) of adagrasib versus docetaxel for pretreated patients with KRASG12C-mutated NSCLC (NCT04685135) is ongoing and several combination strategies with adagrasib is under development ( Figure 1 ).

(Similarities and differences among sotorasib and adagrasib (direct KRAS G12C inhibitors) are summarized in Table 1 ).

Other direct KRASG12C (OFF) inhibitors are in the early stages of clinical development as monotherapy and in combination with other therapies, such as GDC-6036, D-1553 JAB-21822, JDQ443 or LY3537982 (NCT04449874, NCT04585035, NCT05009329, NCT04699188, NCT04956640), that will be investigated alone or in combination with other study treatments (11, 13, 54, 55) ( Figure 1 ).

On the other hand, LY3499446 and JNJ-74699157 (ARS 3248) were discontinued, the first one, due to safety issues.

Novel second generation KRAS G12C inhibitors are under development in preclinical models, that consist of tri-complex inhibitors of the oncogenic GTP-bound form of KRAS G12C (ON) that overcome RTK-mediated escape mechanisms and lead to tumor regressions. The covalent tri-complex inhibitor of KRAS G12C (ON) exhibit a preclinical profile that is superior to the leading KRAS G12C (OFF) inhibitors in clinical development (13, 56). RMC-6291 is a first-in-class, potent, oral and selective tri-complex inhibitor of KRAS G12C (ON) and NRAS G12C (ON) that has demonstrated deep and sustained anti-tumor activity in preclinical lung and colorectal cancer models driven by a KRAS G12C mutation (57), and RMC-6236, another first-in-class, potent, oral RAS-selective tri-complex RAS^MULTI(ON) inhibitor, which has demonstrated pronounced anti-tumor activity in preclinical models of human lung, colorectal and pancreatic cancers caused by multiple RAS variants including KRAS G12D and KRAS G12V and also in RAS-dependent wt tumors and RAS-mediated adaptive resistance tumors (13). Both drugs are pending of being tested in early phase trials.

A better understanding of the mechanisms of resistance to direct KRAS G12C inhibitors is crucial to guide combination strategies and the development of new drugs to improve outcomes for patients with KRAS mt NSCLC.

As previously mentioned, an independence of KRAS signaling and an epithelial-mesenchymal phenotype could lead to an intrinsic resistance to therapy based on direct KRAS G12C inhibitors (18–22).

A potential acquired resistance mechanism to direct KRAS G12C inhibitors is the reactivation following KRAS G12C inhibition driven by RTKs. The combination on RTKs inhibitors with direct KRAS G12C inhibitors as well as with Src homology phosphatase 2 (SHP2) inhibitors could reverse this reactivation (58–61).

A study of the possible mechanisms of acquired resistance to adagrasib from KRYSTAL-1 trial has recently been published. This study performed histologic and genomic analyses (NGS on tissue or ctDNA) developing a deep mutational scanning, and compared pretreatment samples of 38 patients (27 with NSCLC) who initially had stable disease for at least 12 weeks or an objective response to therapy followed by subsequent disease progression, with samples obtained after the development of resistance. 41% of patients had more than one concurrent potential resistance mechanism.

The most frequent on-target mechanisms to adagrasib included activating mutations in KRAS (G12D, G12V and G13D), Q61H), secondary KRAS mutations within the adagrasib-binding pocket (R68S, H95D/Q/R or Y96C) and high-level amplification of the KRAS G12C allele (62).

Recently, it has been suggested that the mechanisms of acquired resistance based on the presence of non-KRAS G12C mutations may be present at baseline, and selected by treatment with direct KRAS G12C inhibitors, becoming more prominent during the course of therapy, being potentially also involved in intrinsic resistance (63).

In relation to off-target mechanisms of resistance, the most frequent detected were MET amplification, activating mutations in NRAS, BRAF, MAP2K1, and RET; oncogenic fusions involving ALK, RET, BRAF, RAF1, and FGFR3; and inactivation mutations in NF1 and PTEN. Moreover, it was described histologic transformation to squamous-cell carcinoma in two patients with advanced NSCLC.

In addition, these authors also performed in vitro experimental studies to compare these acquired resistance mechanisms with adagrasib and sotorasib, and it was seen that while R68S and Y96C mutations conferred resistance to both drugs, H95D/Q/R mutations do not confer in vitro resistance to sotorasib, as seen in patients treated with adagrasib (62). Figure 2 .

Other authors studied in vitro model exposed Ba/F3 cells transduced with KRAS G12C, derived resistant cell lines to sotorasib or adagrasib, searching for secondary KRAS mutations, and identified Y96D and Y96S as resistant mutations to both drugs; while G13D, R68M, A59S and A59T were highly resistant to sotorasib but remained sensitive to adagrasib, and Q99L was resistant to adagrasib but sensitive to sotorasib. According to the different resistance mutations and their patterns of sensitivity to the different KRAS inhibitors, these authors proposed a possible treatment sequencing strategy (64).

Src homology phosphatase 2 (SHP2) protein transduces signals from activated RTKs to downstream RAS pathways (65, 66). Recent studies have shown that SHP2 inhibition specifically suppresses the growth capacity of KRAS-mutant, but not wt NSCLC cells in vitro, which is promoted by TKI treatment (8, 11, 65–67).

There are several SHP2 inhibitors in development. RMC-4630 has shown clinical activity with a DCR of 67% for all KRAS mutations, and 75% for KRAS G12C mutations (phase I NCT03989115) (11, 58, 68, 69). TNO155 is a selective, allosteric, oral inhibitor of SHP2 and is being studied in a phase I trial in advanced solid tumors after disease progression following standard therapy (NCT03114319, NCT04330664) (68, 69).

SPH2 inhibition increases KRAS-GDP occupancy what it could increase the effect of direct KRAS G12C (OFF) inhibitors. This also has been demonstrated in preclinical studies of adagrasib combined with SHP2 inhibition (58, 66–69). Based on this, several ongoing clinical studies with both sotorasib and adagrasib are evaluating the combination of a KRAS G12C (OFF) inhibitor and a SHP2 inhibitor (NCT04330664, NCT04699188, NCT04185883).

SOS1 decreases the affinity of RAS proteins for GDP and favors GTP binding, leading to RAS activation. BI-3406 is a potent and selective inhibitor of the SOS1-KRAS interaction that attenuates reactivation by MEK inhibitors and enhances the sensitivity of KRAS-dependent cancers to MEK inhibition in preclinical models. Consequently, the combination of this new drug with a MEK inhibitor could be a good option for future research on KRAS-driven cancers (8, 11, 70).

Another inhibitor of the SOS1-KRAS interaction is BAY-293, which has demonstrated efficacy in KRAS-driven cancers in preclinical studies (71). BI-1701963 is a drug similar to BI-3406 that is being evaluated in a phase 1 trial in combination with a direct KRAS G12C inhibitor (OFF) (NCT04835714, NCT04975256) or trametinib (NCT04111458) in patients with advanced solid tumors with KRAS mutations.

The up-regulation activity of RTKs and, consequently, reactivation of RAS wt is an off-target mechanism of adaptive resistance to direct KRAS G12C (OFF) inhibitors. Therefore, vertical inhibition strategies are being developed to improve the clinical efficacy of KRAS G12C inhibitors (59, 72).

The KRYSTAL-1 study will evaluate adagrasib in combination with afatinib (an EGFR/HER2 inhibitor) or cetuximab (an EGFR monoclonal antibody) among other combinations, and CodeBreak 101 (NCT04185883) also includes an arm combining sotorasib with afatinib and another with panitumumab (an EGFR monoclonal antibody).

The RAS-GTP complex activates several downstream signaling effectors, including the mitogen-activated protein kinase (MAP-K)/ERK and PI3K/AKT/mTORC1 pathways. Over the past decades, several attempts have been made to block these signaling pathways with disappointing results (8, 73).

In contrast to the first generation of RAF inhibitors, which failed due to activation of the RAF/MEK/ERK pathway in BRAF-like tumor cells, pan-RAF inhibitors with a more effective RAS pathway blocking profile are being developed (74, 75). Another potential strategy involves the combination of MEK inhibitors with direct KRAS G12C inhibitors (OFF) (NCT04185883). VS-6766 is a dual RAF/MEK inhibitor that blocks both the kinase activity of MEK and the ability of RAF to phosphorylate MEK (75, 76). The use of VS-6766 in combination with defactinib, an FAK inhibitor, is being investigated in patients with advanced KRAS mt solid tumors (NCT03875820).

The PI3K/AKT/mTORC1 pathway is not dependent on RAS alone for activation (13, 61, 72). In this setting, combination therapies including a direct KRAS G12C inhibitor and PI3K inhibitor could synergistically increase response (77).

Other therapeutic agents being studied in combination with direct KRAS (OFF) inhibitors include cyclin-dependent kinase 4/6 inhibitors (13).

Sotorasib in combination with antiPD-1 has demonstrated complete responses in immunocompetent mice with patient-derived xenografts, and induced in these mice, a durable immune response with increased TILS and antigen presenting cells, greater benefit than that obtained with each agent in monotherapy (11, 43). On the other hand, adagrasib was shown to modulate factors involved in antigen presentation or an immunosuppressive tumor microenvironment in a panel of human xenograft models. Adagrasib was also shown in mice to decrease myeloid-derived immunosuppressive suppressor cells and increase M1-polarized macrophages, dendritic cells, and CD4+ and CD8+ T cells when administered as a single agent, whereas when administered in combination with anti-PD-1 therapy it leads to durable complete regressions through an immune-mediated antitumor response (78).

Early phase clinical trials are evaluating the combination of ICPI with the KRAS G12C inhibitor adagrasib or sotorasib (NCT03785249, NCT04185883, NCT03600883, NCT04613596) (43, 77).

It represents an attractive strategy for those KRAS subgroups less likely to respond to anti-PD1 monotherapy, such as the STK11/KRAS co-mutated subgroup.

The use of combination regimens of IPCI with TKIs has resulted in excess toxicities without additional efficacy in metastatic NSCLC with actionable driver mutations such as EGFR or ALK (79). To date, no increase in grade 3 or higher toxicities, such as interstitial lung disease or liver toxicity, has been reported with sotorasib or adagrasib in phase I/II clinical trials (48–51). However, the impact of KRAS G12C inhibitors on toxicities arising from prior ICPI use is not well understood and remains an important question, as most patients eligible for KRAS G12C inhibitors will have been previously exposed to ICPIs. Recently, a case of severe immune-related hepatitis likely triggered by sotorasib has been reported in a patient with KRAS G12C mt NSCLC who had been previously treated with antiPD-1 (80).

KRAS-targeted degradation might be an important therapeutic approach to KRAS mt tumors regardless of KRAS subtype.

PROteolysis TArgeting Chimeras (PROTACs) or small molecule degrader technology are novel compounds design to induce targeted protein ubiquitination and proteasomal degradation by the cereblon E3 ligase complex (81). These bifunctional molecules simultaneously engage a target protein and an E3 ligase, forming a ternary complex, which allows the E3 ligase to ubiquitinate the target protein at proximal lysine residues that is recognized and degraded by the 26S proteasome (82). Initially, PROTACs targeting KRAS G12C did not degrade endogenous KRAS (83). However, the emergence of covalent inhibitors to target KRAS has enabled the development of PROTACs capable of inducing endogenous KRASG12C degradation in cancer cells, such as LC-2 that couples the covalent KRASG12C inhibitor adagrasib to the von Hippel–Lindau (VHL) ligand (82).