KRAS mutations are among the most prevalent tumor drivers, but targeting them pharmacologically has been challenging. Recent landmark studies have demonstrated promising clinical results of KRASG12C inhibition by using small molecules. Bar-Sagi, and Knelson and Sequist provide their distinct perspectives on this recent tour de force in targeting KRASG12C alterations.
From the bench: Dafna Bar-Sagi
KRAS-activating mutations are the most frequent oncogenic alterations in human cancer. Among these mutations, KRASG12C is present in 13% of lung adenocarcinomas, 3% of colorectal cancers and 2% of other solid tumors. KRAS works as a molecular switch, pivoting between active GTP-bound and inactive GDP-bound conformations; KRAS-activating mutations, including the KRASG12C variant, fix the protein in its active GTP-bound form by interfering with this cycling process. Despite RAS being the first oncogene discovered in human cancer, efforts to target RAS family proteins with small-molecule inhibitors have been hampered because this protein activates its downstream signaling through protein–protein interactions and because activating mutations usually derive from the inhibition of its catalytic activity1.
However, the decades-long quest for therapeutic strategies targeting oncogenic RAS-driven cancers has now reached a major milestone. Mirati Therapeutics and Amgen recently unveiled two different KRASG12C inhibitors—MRTX849 and AMG 510, respectively—that exhibit pronounced in vivo anti-tumor effects in mice and clinical activity in people with lung and colon adenocarcinomas2,3. Both inhibitors work by selectively forming a covalent attachment to the mutant cysteine in KRASG12C. Only the GDP-bound, inactive state of KRASG12C is accessible to the inhibitors, and the covalent occupancy of Cys 12 locks the protein in the inactive conformation, thereby blocking oncogenic signaling. These molecular-design attributes stem from the pioneering work of Shokat and colleagues, who were the first to show that the cysteine residue in KRASG12C can be exploited to generate covalent inhibitors with preclinical activity4.
Given their drug-like properties, MRTX849 and AMG 510 have enabled a first-of-its-kind, in-depth interrogation of the mechanisms that affect sensitivity and resistance to KRASG12C inhibition. As predicted, MRTX849 and AMG510 displayed a near-universal mutant-selective inhibition of tumor-cell growth when tested in vitro on 20 cell lines. The inhibition was accompanied by attenuation of KRASG12C downstream signaling, most notably the ERK–MAPK and mTOR–S6K kinase pathways. However, the degree of sensitivity to treatment with the inhibitors varied significantly among cell lines. This variability could not be explained by differences in the binding of the drugs to their target KRASG12C. Instead, in cell lines treated with MRTX849, submaximal responses in tumor viability assays correlated with incomplete inhibition of ERK. Moreover, treatment combination experiments involving inhibitors of critical nodes in mutant KRAS signaling have shown that the combination of AMG 510 with an inhibitor of the kinase MEK is most effective in enhancing sensitivity to AMG 510 across multiple cell lines. Thus, maximal inhibition of MAPK signaling appears to be required to augment the efficacy of KRASG12C inhibitors. Unexpectedly, neither AMG 510 nor MRTX849 affects PI3K signaling, which is one of the major RAS effector pathways5, thus possibly revealing that the extent of engagement of the PI3K pathway might depend on the particular KRAS mutation.
When administered to tumor-bearing mice, AMG 510 and MRTX849 elicit anti-tumor effects in multiple KRASG12C-mutant cell lines and patient-derived-xenograft models. Reminiscent of the variability in drug sensitivity observed in the in vitro studies, the anti-tumor responses were heterogenous, ranging from delayed tumor growth to complete regression. On a cautionary note, no significant correlation between the in vitro and in vivo potencies of MRTX849 was discerned, thus underscoring the importance of experimental context as a determinant of drug responses. Indeed, in the case of AMG 510, the markedly improved anti-tumor activity in immunocompetent mice compared with mice lacking T cells was key to revealing an unanticipated targeting opportunity. In the setting of an intact immune system, AMG 510 increased the infiltration of T cells with tumor-killing capacity, and enhancing this capacity by combining AMG 510 with immune-checkpoint inhibition led to durable cure in 90% of mice. These effects are consistent with the growing evidence of an essential role of mutant KRAS in promoting an immunosuppressive microenvironment6. Identifying precisely how AMG 510 reprograms the immune microenvironment should guide therapeutic approaches to improving immunosurveillance in mutant KRASG12C tumors.
Although the predicted benefits of harnessing the immune system to enhance the efficacy of KRASG12C inhibitors are undisputed, additional combination strategies are anticipated to be needed to realize the full potential of these inhibitors in cancer therapy. Taking initial steps toward fulfilling this aspiration, Canon et al.2 have performed comprehensive molecular profiling of in vitro and in vivo models to identify sensitizers of response to the MRTX849. The top hits emerging from these efforts fall into two broad functional categories: (1) upstream regulators of KRAS G-nucleotide exchange and signaling flux, such as receptor tyrosine kinases and the phosphatase SHP2, and (2) regulators of bypass and feedback pathways downstream of KRAS, such as the mTOR and CDK kinases. Accordingly, combination treatment with inhibitors of SHP2 or the kinases HER2, mTOR, or CDK4 or CDK6 significantly enhances the anti-tumor activity of MRTX849. These findings indicate that the combination strategies identified would probably improve therapeutic outcomes for KRASG12C-mutant tumors with limited sensitivity to target inhibition. Given that the drugs used in these combinations are not tumor-cell specific, understanding their effects on the tumor microenvironment will be essential to optimize the design of combination regimens.
The compelling pharmaceutical properties of AMG 510 and MRTX489 prompted Mirati Therapeutics and Amgen to test their respective inhibitors in clinical settings. Early reports from the Amgen trial show encouraging results for lung cancer. Of 13 patients with non-small-cell lung cancer (NSCLC), 7 achieved partial responses, as evidenced by a decrease in tumor size2,7,8. At least for now, the results are less promising for colorectal cancer (CRC). Of 12 patients with CRC, only one had a partial response. Similar results have recently been reported by Mirati at the 2019 AACR–NCI–EORTC International Conference on Molecular Targets and Cancer Therapeutics: of six patients with NSCLC, three achieved a partial response, and of four patients with CRC, one achieved a partial response. Although the research is still in a nascent stage, these results provide the first documentation that, at least in a subset of patients carrying a KRASG12C alteration, tumor-cell growth and survival depend directly on mutant KRAS. Further insights into why and how this dependence is lost in treatment-refractory patients could inform strategies to broaden the clinical utility of the inhibitors.
The development of AMG 510 and MRTX849 has provided an unprecedented opportunity to selectively target mutant KRAS in patients. By virtue of their design, these inhibitors would work only in the context of the KRASG12C alteration. Nonetheless, lessons learned from testing these inhibitors in the laboratory or the clinic should have broad implications for future targeting efforts as the dream of ‘drugging’ KRAS becomes a reality.
From the clinic: Erik H. Knelson and Lecia V. Sequist
The ability to personalize cancer therapy by targeting driver mutations is one of the most impactful developments in the treatment of lung cancer in the past 20 years. For the fortunate minority of patients with lung cancer who are eligible for targeted therapy because they have activating mutations in EGFR, ALK or ROS1, the median overall survival with advanced disease has improved from less than 1 year up to 4 years or longer9.
KRAS mutations, the largest subset of driver mutations in lung cancer, are found in 30% of lung adenocarcinomas. Other cancers, especially those in the pancreas and colon, also frequently harbor KRAS mutations. In contrast to driver mutations in EGFR, ALK or ROS1, which are susceptible to tyrosine kinase inhibitors (TKIs), KRAS is an intracellular GTPase with a chemical affinity for GTP that is much higher than that of mimetic drugs. A long line of failed strategies have attempted to indirectly target KRAS, such as by inhibiting farnesyltransferase through blocking KRAS post-translational modifications; by inhibiting downstream KRAS effectors, such as the kinases MEK, PI3K, AKT or mTOR; or by inhibiting the proteasome or chaperone proteins such as HSP90 (ref. 10; Fig. 1).
Recent developments in basic and translational research have begun to shed light on the seemingly unsolvable KRAS puzzle. One critical step was considering individual KRAS mutations separately and the discovery of a previously unknown drug-binding pocket1,4 in KRASG12C, representing the most common KRAS-mutation subtype in lung cancer. A recent article on the first small-molecule KRASG12C inhibitor, AMG 510, by Canon et al. in Nature2, has demonstrated impressive pharmacokinetic, pharmacodynamic and preclinical data in models of lung, pancreas and colon cancers. In addition, the clinical relevance of this approach has been illustrated by substantial tumor responses achieved among the initial handful of lung cancer patients treated in the phase 1 clinical trial of AMG 510. Clinical updates for AMG 510 communicated at conferences this year have confirmed that as the number of treated patients with lung cancer increased to greater than 30, the activity level remained encouraging, and multiple patients responded to therapy and/or remained on the treatment for 6 months or more8.
Although preliminary, these results are likely to substantially affect the clinical management of lung cancer. Because KRAS mutations are so common in lung adenocarcinoma, the number of patients with KRASG12C is greater than that with any of the other targetable mutations, such as EGFR and ALK, and because next-generation sequencing is becoming part of standard-of-care diagnostic workups, KRASG12C is already being screened in most lung cancer patients. Just as clinicians have learned over time to recognize specific EGFR mutations when making treatment decisions (for example, EGFR exon 19 deletions are extremely sensitive to available TKIs, whereas EGFR exon 20 insertions are inherently resistant), noting which patients have KRASG12C can be easily incorporated into clinical practice. Although we will continue to gain clarity about the frequency and severity of side effects of AMG 510 from ongoing clinical trials, reassuringly, proteomic profiling by Canon et al. has identified KRASG12C as the only covalent target with substantial engagement2, and preliminary clinical results have shown no dose-limiting toxicities and only infrequent severe adverse events to date8.
After a targeted therapy demonstrates clinical efficacy, there is an outstanding need to foresee potential mechanisms of acquired resistance. Bypass track activation commonly occurs after ALK and EGFR inhibition, and has been implicated in failed attempts to target downstream effectors of KRAS10. The authors have anticipated this challenge with AMG 510 and laid a foundation for future work to circumvent acquired resistance by demonstrating effective pre-clinical combinations of AMG 510 with inhibitors of HER kinases, EGFR, MEK, SHP2, PI3K and AKT2. Whether the strategies of dual vertical pathway inhibition explored in vitro and in mouse models are safe in patients remains to be seen, because of overlapping toxicity profiles, although select TKIs have been successfully combined in lung cancer treatment, for example, dual BRAF–MEK and EGFR–MET inhibition11,12.
Interestingly, AMG 510 has also been found to activate innate immune signaling in KRASG12C mouse models and synergize with anti-PD-1 checkpoint inhibitors2, in line with the finding that lung cancers with KRAS mutations are relatively sensitive to current immunotherapies13. The precise mechanism of how KRASG12C covalent inhibition leads to enhanced anti-tumor immunity remains to be elucidated, although according to the data provided, it correlates with innate immune activation and depends on T cell infiltration2. AMG 510 in combination with anti-PD-1 treatment not only causes tumor regression but also provides protection against subsequent tumor re-challenge in mice, thus pointing to future clinical testing as a first-line therapy for advanced disease and eventual consideration as a neo-adjuvant treatment. However, potential side effects from combining KRASG12C inhibitors with checkpoint inhibitors warrant careful testing, because life-threatening pneumonitis can occur when small molecules are combined with immunotherapy, especially in patients with lung cancer14. Penetrance of the central nervous system will also be of particular interest, given the frequency of brain metastasis in patients with lung cancer.
Data on an exciting second KRASG12C inhibitor have also recently been published by Hallin et al.3 MRTX849 is a different small molecule that similarly demonstrates substantial pre-clinical activity in lung and colon cancer models, and has also shown early activity without concerning safety signals in a small cohort of patients with lung cancer in a phase 1 trial15. Although information on any potential immunological effects of MRTX849 has not yet been made available, the authors also explored potential resistance pathways. Interestingly, they found that KEAP1 or NFE2L2 mutations, which predict poor response to checkpoint immunotherapy13, are associated with resistance to MRTX849, whereas combination strategies of MRTX849 with inhibitors of HER2, CDK4 or CDK6, SHP2 and mTOR pathways proved efficacious, even in MRTX849-resistant models3.
The impressive early clinical data for both AMG 510 and MRTX849 mark the beginning of a new chapter in treating patients with KRASG12C alterations. Although the durability of response and effects on overall survival will determine the ultimate clinical success of KRASG12C covalent inhibition in patients with lung cancer, these breakthroughs inspire hope for continued molecular dissection and targeting of driver alterations, even those previously thought to be impossible to unlock.
Ostrem, J. M. & Shokat, K. M. Nat Rev. Drug Discov. 15, 771–785 (2016).
Canon, J. et al. Nature 575, 217–223 (2019).
Hallin, J. et al. Cancer Discov. https://doi.org/10.1158/2159-8290.CD-19-1167 (2019).
Ostrem, J. M., Peters, U., Sos, M. L., Wells, J. A. & Shokat, K. M. Nature 503, 548–551 (2013).
Castellano, E. & Downward, J. Genes Cancer 2, 261–274 (2011).
Cullis, J., Das, S. & Bar-Sagi, D. Cold Spring Harb. Perspect. Med. 8, a031849 (2018).
Fakih, M., O’Neil, B. & Price, T. J. J. Clin. Oncol. 37 (Suppl.), abstr. 3003 (2019).
Govindan, R. et al. Ann. Oncol. 30 (Suppl. 5), mdz244.008 (2019).
Hosomi, Y. et al. J. Clin. Oncol. https://doi.org/10.1200/JCO.19.01488 (2019).
Matikas, A., Mistriotis, D., Georgoulias, V. & Kotsakis, A. Crit. Rev. Oncol. Hematol. 110, 1–12 (2017).
Planchard, D. et al. Lancet Oncol. 18, 1307–1316 (2017).
Sequist, L. V. et al. Cancer Res. 79, abstr. CT033 (2019).
Arbour, K. C. et al. Clin. Cancer Res. 24, 334–340 (2018).
Oxnard, G. R. et al. J. Clin. Oncol. 33 (Suppl.), abstr. 2509 (2015).
Jänne, P. A. in AACR-NCI-EORTC International Conf. on Molecular Targets and Cancer Therapeutics (2019); https://www.mirati.com/wp-content/uploads/2019/10/AACR-NCI-EORTC-Clinical-Data-Presentation_Janne_October-2019-1-1.pdf
The authors declare no competing interests.
About this article
Cite this article
Bar-Sagi, D., Knelson, E.H. & Sequist, L.V. A bright future for KRAS inhibitors. Nat Cancer 1, 25–27 (2020). https://doi.org/10.1038/s43018-019-0016-8