A Splendid New Beginning at the End of a 40-Year Quest: The First KRAS^G12D Inhibitor in Pancreatic Cancer

Pancreatic ductal adenocarcinoma (PDAC) has the highest death rate of the most commonly diagnosed cancers, and with a rising occurrence, it is predicted to become the second leading cause of cancer deaths by 2030 (1). One of the major determinants of the dire outcome of PDAC is the lack of effective targeted therapy. Dubbed the “undruggable” for several decades, mutational activation of the Kirsten rat sarcoma (KRAS) gene occurs in 95% of PDAC and has been unequivocally established as the initiating mutation and genetic driver of this disease (1). The KRAS protein is a GTPase that in response to cellular stimuli cycles between GDP-bound (OFF) and GTP-bound (ON) states to transmit intracellular signals. Under physiologic conditions, this system is controlled by guanine nucleotide exchange factors, which mediate the exchange of GDP for GTP, and GTPase accelerating proteins that accelerate the intrinsic hydrolysis of GTP-bound RAS by several orders of magnitude. Oncogenic mutations of KRAS frequently involve single amino acid substitutions in residues G12, G13, and Q61, which hinder the capacity to hydrolyze GTP, thus favoring the GTP-bound (ON) state of the molecule and uncontrolled signaling. The prevalence and tissue distribution of KRAS mutation subtypes vary greatly. Overall, mutations in KRAS^G12 (D/C/V/R/A/) are the most prominent in cancer (89%), with G12D as the most frequent (36%), followed by G12V (23%) and G12C (14%). The most prevalent KRAS mutations in PDAC are G12D (∼41%) and G12V (∼34%), whereas G12R is less frequent (∼16%; ref. 1). Pioneering work by Shokat and colleagues in 2013 fundamentally changed the 4-decades-long view of KRAS as undruggable and paved the way for the development of the KRAS^G12C inhibitors AMG510 (sotorasib) and MRTX849 (adagrasib; refs. 2–4). These molecules form irreversible covalent bonds with the C12 residue of GDP-bound KRAS and have shown remarkable clinical outcomes in KRAS^G12C non–small cell lung cancer and colorectal cancer (3, 4). Recently, Mirati Therapeutics reported MRTX133, a reversible KRAS^G12D inhibitor with single-digit nanomolar IC₅₀ for KRAS^G12D (5). MRTX133 binds both the GDP and GTP states of KRAS^G12D, locking the protein in its inactive form or providing steric hindrance to block effector binding in the active setting. Building on these prior findings, Kemp and colleagues set out to discover the long-awaited answer to the question of the efficacy of small-molecule inhibition of KRAS^G12D in the treatment of PDAC (6).

The authors used several immunocompetent mouse models of PDAC that recapitulate human disease, including syngeneic orthotopic implantations of clonal cell lines derived from the LSL-KRasG12D/+;LSLTrp53R172H/+;Pdx-1-Cre (KPC) model, the autochthonous KPC/Y (KPC mice expressing a YFP epithelial lineage tag) model, and subcutaneous implantations of clonal cell lines derived KPC/Y tumors (7). Within a short time frame (7 days), treatment with MRTX113 as a monotherapy resulted in unprecedented regressions in all models. Although highly efficacious throughout, the different models revealed variations in the response to MRTX113. As such, treatment of the orthotopic or autochthonous KPC/Y models resulted in substantial regression in 100% of the mice, whereas only ∼50% showed regression in the subcutaneous implantation model after 1 week of treatment. Differences in regression were not due to differential sensitivity of cell lines to MRTX113. In addition, MRTX113 was selective to KRAS^G12D, as it had no impact on RAS downstream signaling and growth of subcutaneous tumors from a syngeneic derivative of a KPC cell line in which D12 had been mutated to a cysteine (C). These outcomes are complemented by another recent study showing that 2 weeks of treatment with MRTX113 induced significant regression (30% or greater) in 11 of 25 KRAS^G12D tumors in xenograft models spanning five tissue types (8). The greatest impact, however, was observed in pancreatic cancer models in which MRTX113 resulted in significant regression in 73% of the mice, supporting the reported critical role of mutant KRAS in PDAC as the earliest oncogenic mutation and bona fide genetic driver. These studies provide compelling evidence for the potential clinical use of MRTX113 in KRAS^G12D-driven PDAC.

From a cancer cell–intrinsic perspective, MRTX113 appears to operate similarly to the KRAS^G12C inhibitors. As with KRAS^G12C inhibitors, MRTX113 treatment led to dose-dependent decreases in phospho-ERK1/2 levels, but not phospho-AKT or phospho-S6 levels, thus identifying the MAPK pathway as a dominant mediator of KRAS-driven tumorigenesis (Fig. 1). Kemp and colleagues identified strong inhibition of phospho-ERK1/2 in the first 2 days of treatment, accompanied by an increase in the apoptotic marker cleaved caspase-3 (CC3) and a decrease in the proliferation marker Ki-67 (6). The antitumor response of MRTX113, therefore, appears to be driven by both the induction of apoptosis and inhibition of growth. After 1 week of treatment, the inhibition of proliferation and phospho-ERK1/2 was maintained; however, the increase in CC3 was lost. The reduction in cell death over time hints at the reactivation of apoptotic-suppressive pathways. On this note, the KRAS^G12C inhibitors tell a tale of several intrinsic and acquired resistance mechanisms, including redundancy in parallel signaling pathways, upregulation of the EGFR pathway, and novel mutations in MYC, BRAF, and the KRAS protein itself. Although MRTX113 led to robust tumor regression, examples of resistance to MRTX113 were observed. Within this framework, the durability of the response of KRAS^G12D-driven PDAC to MRTX113 is an important consideration for patient applicability. The antitumor effect of MRTX113 shows significant durability in the short term, as 100% of mice in the subcutaneous model that reached complete regressions (CR) at 2 weeks of MRTX113 treatment remained disease free for 3 subsequent weeks without therapy. KPC/Y mice with deep regression also remained in remission up to 2 weeks after therapy, but subsequently, tumors progressed in both models. Therefore, a thorough evaluation of biochemical events following KRAS^G12D inhibition is imperative for the identification of mechanisms of intrinsic and emerging resistance and optimal combinatorial therapies that circumvent them. These downstream changes in signaling should also be evaluated in a temporal manner, as it may be critical to pair certain therapies at different time points of treatment. Interestingly, one KPC/Y mouse retained MRTX113 sensitivity even after tumor regrowth; however, further studies with larger cohorts are needed to determine whether this is a widespread property.

The model-dependent differences in tumor regression, although not surprising given the well-established context-dependent differences in oncogenic RAS signaling and the tumor microenvironment (TME) of PDAC, provide important insight into the mechanism(s) of action of MRTX113. A major obstacle in treating KRAS-mutant PDAC stems from its dense, hypovascular, and immune-poor TME. Mutant KRAS itself has been implicated in maintaining this immunosuppressive environment by inhibiting antigen presentation, blocking interferon signaling, and increasing the levels of myeloid chemoattractants (9). Consistent with these findings, KRAS^G12D inhibition by MRTX113 was accompanied by drastic changes in the PDAC TME (Fig. 1). A short treatment with MRTX113 (5 doses, over 60 hours) resulted in a reduction in total myeloid cells, granulocytic myeloid-derived suppressor cells, and dendritic cells, as well as an increase in macrophages and a shift toward an M1 state in the subcutaneous PDAC models. Supporting these observations, MRTX113 induced changes in cytokine profiles that would promote antitumor macrophage infiltration and inhibit myeloid-like macrophage infiltration, such as increases in CCL2 and decreases in GM-CSF. MRTX113 treatment also resulted in a significant increase in tumor-infiltrating T cells (CD3⁺) and CD4⁺ and CD8⁺ T-cell subsets, with CD8⁺ T cells expressing higher levels of proliferation and cytotoxicity markers (Fig. 1). In contrast to macrophages, however, the differences in T cells were not sustained with longer treatment. These data suggest that MRTX113 treatment could help transition immune-cold tumors into tumors with some degree of immune infiltration. However, they also indicate that the effect of KRAS^G12D inhibition on the PDAC immune microenvironment can be dynamic and as such must be carefully considered in the design of approaches that combine immunotherapy with KRAS^G12D inhibition. Interestingly, the degree of regression appears to vary depending on the immune infiltrate status of PDAC. When treated with MRTX113, 100% of immune-hot (high T-cell infiltrates) subcutaneous tumors regressed, with half of these being CR. Meanwhile, only 50% of immune-cold tumors regressed, and no CR was observed. The combination of MRTX113 with T-cell depletion in the immune-hot tumors had lower efficacy than MRTX113 alone, as only 86% of the tumors regressed and no CR was observed, suggesting that T cells play a role in the antitumor effect of MRTX113 in this model. However, the depletion of T cells in the autochthonous KPC model had no impact on regression, as these models experienced a 100% regression rate with or without T-cell depletion. The role of T cells in mediating the activity of MRTX113 may therefore be context (site) specific. Nonetheless, when T cells were continually depleted after MRTX113 therapy was stopped in immune-hot subcutaneous tumors, 100% of tumors showed substantial regrowth, whereas 50% of tumors with functioning T cells remained in remission. Depletion of T cells also exacerbated regrowth in KPC/Y tumors after MRTX113 therapy. Thus, T cells may play an important role in the durability of response to MRTX113. Finally, surviving tumor cells after MRTX113 treatment showed an increase in MHC class I. This reexpression was similarly seen with KRAS^G12C inhibition, leading to newfound sensitivity to immunotherapeutic intervention (10). Alterations in the additional components of the PDAC TME may also contribute to the antitumor effect of MRTX113 and changes in the immune landscape. As such, MRTX113-treated tumors showed elevated collagen deposition paralleled by an increase in myofibroblastic cancer-associated fibroblasts, which have been shown to be tumor-suppressive. In the subcutaneous models, tumor vascularization also increased after MRTX113 treatment, but this was not observed in the orthotopic setting.

Overall, the identification of MRTX113 generates immense potential for the treatment of patients with KRAS^G12D-driven PDAC. Not only is it the first of its kind, but the experimental evidence provided demonstrates that MRTX113 is a potent inhibitor of KRAS^G12D with unprecedented antitumor activity in preclinical models of PDAC. Although key questions regarding the hierarchy and interplay of KRAS downstream signaling pathways involved in the antitumor effect, durability of response, and resistance mechanisms following MRTX113 treatment must be considered for optimal use in patients, these findings are a great leap forward in the treatment of KRAS^G12D-driven PDAC.