Jonathan Riess, MD, MS
Paul Paik, MD
Paul Paik, MD, is clinical director of the Thoracic Oncology Service at Memorial Sloan Kettering Cancer Center, where his research focuses on identifying targeted therapies in squamous cell lung cancer. He has recently completed a phase II trial targeting NFE2L2 and KEAP1 mutations in patients with non-small cell lung cancers using the TORC1/2 inhibitor TAK-228, which has provided the impetus for targeting NFE2L2 and KEAP1 in this disease in the phase I trial discussed here.
Jonathan Riess, MD, MS, is medical director of thoracic oncology and associate professor of medicine at the UC Davis Comprehensive Cancer Center, where his research focuses on targeted therapies and immunotherapy in lung cancer. Most recently, he conducted a lung-map substudy S1900A, looking at the PARP inhibitor rucaparib in NSCLC.
Dr. Paik and Dr. Riess are currently conducting a phase I trial investigating the next steps in combination therapy of CB-839 HCl (telaglenastat) and TAK-228 in targeting NFE2L2 and KEAP1 mutations in squamous and KRAS mutant NSCLC.
Q: Can you describe the significance of KEAP1 and NFE2L2 mutations in NSCLC?
Dr. Paik: There are a few different attributes to NRF2 signaling (NRF2 as the gene product for NFE2L2, which is the putative oncogene). Its negative regulator is KEAP1, and both are frequently mutated, particularly in squamous cell lung cancer, where these genes are altered in some 30% of patients.
NRF2 is a transcription factor that functions during times of oxidative and xenobiotic stress. When a cell is very stressed, either from toxins or from oxidation, this transcription factor gets upregulated and plays a whole host of roles in trying to detoxify the cancer cell. That’s one very important biologic function. The other significant role it serves, which we are trying to target in our research, is in metabolism, through mTOR signaling, which connects NRF2 to the Krebs cycle and metabolism through glycolysis in squamous cell lung cancers. NRF2 can thus be targeted using mTOR inhibitors. In the case of my antecedent phase II trial, we used TAK-228, a TORC1/2 inhibitor.
Although there is efficacy in targeting NRF2 in this population as I have shown, both I and Dr. David Shackelford of UCLA have characterized a resistance mechanism that is induced by TAK-228 whereby cancer cells shift to use glutamine as a nutrient to fuel the Krebs cycle. This is something we can target with CB-839, which is a glutaminase inhibitor, which when combined with TAK-228 exhibits synergistic antitumor activity in preclinical models and is thus the combined therapy being used in our phase I trial.
Dr. Riess: These are not uncommon mutations in squamous lung cancer; they occur in approximately 25% to 30% of patients. Some data indicate that patients with these mutations may not do as well with chemotherapy or immunotherapy, so exploration of these targets is critical to help improve outcomes for these patients.
There are no approved targeted therapies for squamous lung cancer, unlike lung adenocarcinoma. Development of new targeted agents for squamous lung cancer in particular is a great unmet need.
Q: Are these alterations more commonly seen in KRAS-mutated NSCLC? If so, what are the implications?
Dr. Riess: There are provocative preclinical data in squamous lung cancer as well as in KRAS-mutant NSCLC from Memorial Sloan Kettering, UCLA, and other institutions. Mutations in the NRF2 pathway are present in about a quarter of all KRAS-mutant NSCLC cases, which are far more likely to be adenocarcinoma. There are more KEAP1 mutations than NFE2L2 mutations in KRAS-mutant lung cancer— again, about 25% of cases. There is some overlap with other important co-mutations as well, such as STK11, in KRAS-mutant lung cancer.
With the aberrations in the NRF2 pathway, cells can shift dependence from the TORC-dependent glycolysis to glutaminolysis. There is some preclinical data suggesting glutaminase inhibitors have antitumor activity in KRAS-mutant, KEAP1–co-mutated NSCLC and that the addition of an mTOR inhibitor may potentiate that activity. Again, there’s a suggestion that patients with the KEAP1 co-mutation may not do as well on standard treatments.
There are data with the direct KRAS G12C inhibitor sotorasib, which was recently approved by the U.S. Food and Drug Administration for advanced KRAS G12C–mutant lung cancer. For the subset of patients who had KEAP1 mutations, the response rate and clinical activity were decreased compared with KRAS G12C lung cancers with different co-mutations such as p53.
The response rate in the KEAP1 group was only around 20%, whereas the overall response rate to sotorasib was about 40%. Because it is a subset analysis, however, it is not definitive. Outcomes may be different with other direct KRAS G12C inhibitors, but there is a suggestion there that those with KEAP1 mutations also may be less responsive to therapies directly targeting KRAS G12C. Our real interest is in trying to identify metabolic targets for these KRAS G12C/KEAP1–mutant lung cancers, as well as for other KRAS-mutant lung cancers with different amino acid substitutions.
Q: Are these mutations routinely discoverable on next-generation sequencing (NGS) panels?
Dr. Paik: We have been routinely performing NGS since around 2012, through an institution-specific platform at Memorial Sloan Kettering. We were seeing relatively high-frequency alterations in both NFE2L2 and KEAP1 in approximately 30% of all squamous cell lung cancer patients. The NFE2L2 mutations were also interesting because they were hotspot mutations, which usually does not happen in squamous cell lung cancer, suggesting that they are functionally active.
KEAP1 and NFE2L2 are now incorporated into most commercial NGS panels that comprehensively test targeted exon sequencing. Some omit NFE2L2, because it was only recently identified as a target of interest, but more and more targets will likely be added to testing panels. However, even when targets are identified and included in the panels, they are not always annotated in a clinically meaningful way. The testing companies are trying to help clinicians understand why each of these targets is important— with respect to variants that have known significance and variants that have unknown significance.
The challenge is identifying the significance of each variant. That will happen as we begin to publish more data, so that the companies have data to cite. Some of the onus is on us as researchers, and we will begin to rectify that, particularly for KEAP1 and NFE2L2 mutations. I am confident we will get to that point soon.
Q: Do you consider these co-mutations to be prognostic overall, or predictive of treatment outcome? If the latter, are they really actionable?
Dr. Paik: There are reports that NRF2 is a poor prognostic biomarker, and we can infer biologically why that is the case based on how it functions. Some of this may have to do with resistance to immunotherapy that emerges when NRF2 is activated, but this is still poorly defined.
The data that will be published in my phase II trial suggest that NFE2L2 mutations arising in squamous cell lung cancers can predict the efficacy of treatment with TAK-228. And so I do think NRF2 can function as a predictive biomarker in this circumstance as well. The key to the actionability is a broader understanding of this disease as one that is metabolically driven. The primary focus of the phase I trial that Jonathan and I are conducting is to leverage this understanding, using dual metabolic inhibition with TAK-228 and CB-839 subsets of NRF2-driven NSCLCs.
Dr. Riess: That is the exciting part: this paradigm shift of targeting a metabolic pathway, as opposed to employing a TKI to target an EGFR mutation, for example. This trial is expanding that paradigm to come up with a targeted therapy for aberrations in the NRF2 pathway in squamous cell cancer. Novel approaches are needed, different from the focus on directly inhibiting tyrosine kinases in adenocarcinoma with EGFR and ALK inhibitors and so forth.
Q: What are some novel therapies targeting the KEAP1/NRF2 axis?
Dr. Riess: Sotorasib is a KRAS G12C inhibitor recently approved for the treatment of KRAS G12C–mutant NSCLC. It directly targets KRAS G12C, whereas KEAP1 co-mutations are present in about a quarter of KRAS lung cancers (NFE2L2 less frequently). They coexist and are co-mutated with KRAS. Patients with KEAP1 and NFE2L2 mutations may not do as well when treated directly by KRAS G12C inhibitors. This is just a suggestion based on a subset of patients with KRAS G12C and KEAP1 mutations treated with sotorasib and is not definitive; but it highlights the need to target the NRF2 pathway in KRAS-mutant lung cancer, whether it be in KRAS G12C (which represents about 40% of all KRAS mutations) or in other KRAS mutations with different amino acid substitutions. It will be important to look at this KEAP1/NFE2L2–mutant subset, as new inhibitors and drugs come up—targeting not just G12C but KRAS-mutant lung cancers with other amino acid substitutions.
Dr. Paik: Novel therapies specifically targeting the NRF2 axis are what we are focusing on in our trial—using TAK-228 as part of it, but then shifting to combination therapy with CB-839 because that is where the data are leading us. There is an inducible metabolic resistance that emerges, focusing on glutamine.
TAK-228 targets TORC1 and metabolism as a downstream effect of activation of the NRF2 axis, but the therapy itself does not specifically target NRF2. That is an important distinction, because we do not yet have a clinical NRF2 inhibitor. It is very difficult to generate agents that inhibit transcription factors, which would more specifically target things that are downstream of NRF2. There may be some additional studies that come down the pipeline showing that targeting NRF2 can synergize with other targeted approaches, but right now all that is available is what Jonathan and I are doing.
Q: There have been recent negative findings with some agents targeting glutaminase. What components of a targeted therapy do you think are necessary for success here? What are the current expectations for glutaminase inhibitors in combination with checkpoint inhibitors?
Dr. Paik: This is correct. There have been some negative findings using glutaminase inhibitors in the recent past. A lot of the initial focus for glutaminase inhibition was in renal cell cancer, because we know that metabolism is functionally important in that disease and there are approved drugs that target mTOR as a result. The idea was that one might be able to generate anti-tumor synergy CB-839 to TORC1 inhibitions in this disease. I think the negative trial data tell us that glutaminase is not a functionally important mechanism in renal cell cancer, not that the approach and rationale is generally incorrect.
It’s important to bear in mind that glutaminase inhibition is a form of targeted therapy. The problem is how to define the target in this case. We are accustomed to defining targets exclusively through somatic genomic variants (i.e., gene mutations). This may be a less robust approach when dealing with metabolism, where distinct metabolic signatures would be more suitable. And so a change in the way we conceptualize biomarkers might be necessary to gain further ground in the targeted therapy space in some diseases. This is built into the phase I trial as an aside, and so we will be able to generate these data in time.
Dr. Riess: There are evolving data looking at glutaminase inhibition with immunotherapy. An ongoing trial called KEAPSAKE is enrolling patients who have these KEAP1/NFE2L2 mutations. They are treating these patients in the first line with chemotherapy and immunotherapy and looking at whether the addition of CB-839 improves clinical outcomes. It will be interesting to see the effect of glutaminase inhibition in combination with chemotherapy and immunotherapy.
We anticipate synergistic activity with dual inhibition of glycolysis and glutaminolysis via combined inhibition of mTOR and glutaminase in these NRF2-aberrant subsets of lung cancer. Single-agent mTOR inhibitors have not shown the results that were anticipated during the long arc of drug development of these drugs in lung cancer, in part because of not matching up with the right target. In addition, dependence on glutamine metabolism as a mechanism of de novo resistance in some patients treated with mTOR inhibition potentially could be overcome via glutaminase inhibition with CB-839.
