Despite the past decade’s accomplishments in identifying the genetic drivers of NSCLC, the complexity of these cancers continues to create barriers to effective targeted treatment.
Finding strategies capable of overcoming those barriers will depend, in part, on the in-depth study of the proteins whose activity causes oncogenicity, according to Ramaswamy Govindan, MD, a medical oncologist with Siteman Cancer Center at the Washington University School of Medicine in Missouri.
That’s why Dr. Govindan hopes the results of a 2021 recent comprehensive analysis of cancer-driving proteins in NSCLC1 —conducted by Dr. Govindan and his colleagues in the National Cancer Institute's Clinical Proteomics Tumor Analysis Consortium—are expected to change the course of research and treatment. The findings further characterized NSCLC tumors and uncovered potentially druggable targets, and their validity is bolstered by the generation of similar conclusions by a collaborating group from Taiwan, which simultaneously published results from its own study in Cell in July.2
“In 1995, if you asked how many genes human beings had, people would have guessed wildly: 20,000, 40,000, 200,000,” Dr. Govindan said. “Until The Human Genome Project came along, we didn't know how many genes we had. Now, roughly, we know it's about 20,000 genes. The same way, we don't know, even today, how many proteins we have, because the same gene can make multiple proteins in different forms. Not only that, the proteins have to then undergo some changes called post-translational modification—most commonly, phosphorylation and acetylation—and those processes make the proteins more functional. These kinds of post-translation modifications are very significant: They suggest how active the protein is.”
The proteomics analysis is exciting from a bench-science standpoint because “we found that you could do this, number one, with a very high level and get detailed information,” Dr. Govindan said. “The second thing is, of course, that there are distinctive differences in how the proteins get altered.”
He expects the research, which identified four subgroups of NSCLC defined by factors including key driver mutations, to help launch an active new branch of investigation into the underpinnings of the disease.
“This proteogenomics dataset represents a unique public resource for researchers and clinicians seeking to better understand and treat lung adenocarcinomas,” Dr. Govindan and his fellow investigators wrote in Cell.
He anticipates that proteomics exploration will continue to uncover new druggable targets, ultimately leading to the ongoing development of novel targeted therapies for patients with NSCLC that harbors KRAS, EGFR, or other common genetic mutations.
Study Rationale and Design
In the study, members of the consortium compared 110 treatment-naive tumors with 101 matched normal adjacent tissues using comprehensive proteogenomic characterization. The study of the tissue considered the interaction between genomics, epigenomics, deep-scale proteomics, phosphoproteomics, and acetyl-proteomics in oncogenic expression.
The team embarked on the study because lung cancer is the leading cause of death in the U.S. and worldwide, with a 19% overall 5-year survival rate in America, despite recent therapeutic advances. They focused on adenocarcinoma, the most common lung malignancy, because it is not only strongly related to tobacco smoking but is also the subtype most frequently found in never-smokers.
Although smoking status, gender, and ethnicity affect the genetics and natural history of lung adenocarcinoma, large-scale sequencing efforts have typically been limited to cohorts of smokers with little ethnic diversity, the authors pointed out. Of the major efforts so far, only The Cancer Genome Atlas (TCGA) measured proteins and phosphopeptides, and that exploration was restricted to a 160-protein array. Yet, proteomic and phosphoproteomic alterations are at the heart of the cellular transformations caused by lung adenocarcinoma’s most frequent genomic aberrations—RAS, RAF, and RTK, the team wrote, suggesting that “global proteogenomic profiling is needed to provide deeper mechanistic insights.” They added that many lung adenocarcinomas may harbor mutations that are not yet recognized as targetable.
Based on these concerns, the study population was selected to reflect diverse demographic and clinical characteristics, including country of origin and smoking status, with the majority of participants from Vietnam, the U.S., and China. The cohort was largely male, with more than half having a history of smoking. The largest proportion of tumors were stage IB. The most common mutations were in TP53 (54%), EGFR (34%), and KRAS (31%).
The patients’ cryopulverized tissue was profiled via whole-exome, RNA, and microRNA sequencing; array-based DNA methylation analysis; and in-depth proteomic, phosphoproteomic, and acetylproteomic characterization. Complete data was collected for 101 tumors and 96 normal adjacent tissues.
Each type of test uncovered distinct information, the authors wrote:
*Through proteomic and phosphoproteomic investigation, the researchers were able to learn about biology downstream of copy-number aberrations, as well as somatic mutations and fusions; they also identified therapeutic vulnerabilities in tumors that were associated with driver events involving KRAS, EGFR, and ALK.
*A complex landscape was revealed through immune subtyping, they determined.
“It’s not just that some immune cells are ‘cold,’ but that immune cells are downregulated in some tumors,” Dr. Govindan explained. “Compared to normal tissue, the immune system is downregulated in a small number of tumors, showing that there is an active immunosuppression going on in some subset of tumors.”
Immune subtyping also reinforced the association of STK11 with immune-cold behavior and underscored neutrophil degranulation as a potential immunosuppressive mechanism in STK11-mutated lung adenocarcinomas that is evident only through proteomics evaluation.
*Smoking-associated lung adenocarcinomas, the authors found, appeared correlated with other environmental exposure signatures and a field effect in normal adjacent tissues.
*Matched normal adjacent tissues allowed the identification of differentially expressed proteins with potential diagnostic and therapeutic utility.
Using multi-omics unsupervised clustering, the investigators found four subgroups defined by key driver mutations, country, and gender (P ˂ 0.01):
Cluster 1: These tumors were of the unfavorable proximal-inflammatory RNA subtype, had mutated TP53 but not STK11, and were CpG island methylator phenotype (CIMP)-high, a characteristic associated with a worse prognosis.3 From an immune perspective, these tumors were generally considered “hot.”
Cluster 2: These cancers were associated with patients who were of western origin, particularly the U.S.; they were of the unfavorable proximal-proliferative RNA subtype, harbored neither TP53 nor EGFR mutations, and were CIMP-intermediate.
Cluster 3: Many patients with these tumors were from Vietnam; their cancers were of the proximal-proliferative RNA subtype and were STK11-mutated. These tumors were considered immunogenically “cold.”
Cluster 4: Frequently occurring in female residents of China, these tumors were of the more favorable terminal respiratory-unit RNA subtype, harbored EGFR but not KRAS or STK11 mutations, and were CIMP-low. This group also included most of the samples that harbored EML4-ALK fusions. These tumors were considered immune-“cold.”
The scientists also found that tumors clustered into five subgroups according to their microRNA expression. Two of those aligned with Clusters 1 and 3, while the others had mixed composition.
The researchers found lung adenocarcinomas to be much more methylated than their matched normal adjacent tissues. And, they found phosphorylation to be common in the studied tumors. The literature reflects that PTPN11/Shp2 inactivation can suppress tumorigenesis, and the team’s data demonstrates that EGFR-mutant and ALK fusion-driven lung adenocarcinomas would be promising targets for existing investigational inhibitors of these proteins.
“That really is quite clinically significant,” Dr. Govindan said. “It shows how certain alterations are not seen at the DNA level, but we see them at the protein level.”
Also found was that SOS1 inhibition could be a valuable strategy against KRAS-mutant tumors.
In further findings that may be applicable in the clinic, the team determined that the risk of venous thromboembolism might be stratified by mutation type, meaning that prophylactic anticoagulation, which poses health risks, could be reserved for patients without EGFR mutations.
By assessing hyperphosphorylation, the team uncovered several driver-specific outlier kinases that were found to have interactions with Food and Drug Administration (FDA)-approved drugs, including EGFR in EGFR mutants, PRKCD in KRAS mutants, BRAF in TP53 mutants, and WEE1 in EML4-ALK fusions. Furthermore, the team identified 27 kinases expected to be druggable with known but as-yet-unapproved inhibitors.
In addition, the team demonstrated that anti-CTLA4 therapy and IDO1 inhibition could be workable strategies against immune-hot tumors.
Overall, analysis of lung adenocarcinoma tumors and matched NATs “provided a powerful dataset to nominate candidate biomarkers,” which “enabled us to capture the impact of cancerization in both tumors and adjacent tissues and highlighted a potential oncogenic mechanism centered on ARGHEF5 in never-smokers,” the researchers wrote.
Targeting Unmet Need
Using the findings to improve the targeted treatment of EGFR-mutated NSCLC is a key goal because, although many targeted drugs are available for the condition, it often develops resistance to those therapies. Meanwhile, with no drugs yet approved by the FDA to target disease driven by KRAS mutations, the identification of new druggable biomarkers for this subgroup will be particularly meaningful. Currently, patients with KRAS mutations are treated with chemotherapy and immunotherapy.
Dr. Govindan explained why mutations such EGFR and ALK, which are driven by mutations that make them sensitive to tyrosine kinase inhibitors, develop resistance to treatment with those drugs.
“When the kinase inhibitors work, they enable the EGFR pathway, and the tumor melts away,” he said. “But the problem is, over time, these cells acquire a mutation in the other part of the kinase domain, or they amplify the EGFR, or they activate the other pathways. For example, they activate the MEK chain or something else, and that leads to resistance.”
Ideally, results from the consortium’s study and future proteomics explorations will shed light on how to prevent or resolve that kind of resistance while providing insights into new strategies for treating NSCLC that harbors EGFR mutations—for instance, as Govindan and his colleagues demonstrated, with PTPN11/Shp2 inhibition.
KRAS-mutated disease, on the other hand, does not possess the tyrosine kinase domain, and hence does not respond to inhibitors of those proteins. Rather, Dr. Govindan said, the mutation that occurs in 13% of NSCLCs is broken into subtypes associated with glitches in various exons, each resulting in a specific aberrant protein, such as D12B, G12B, G12C, T12A, T12C, T12D, and T13D. To target these mutations, a drug must fit into the pocket formed when the cancer-driving protein folds.
Several drugs being developed to treat KRAS mutations target the G12C subset because it has the most accessible pocket. The furthest along in development is sotorasib, or AMG510, which demonstrated a 32.2% objective response rate in a phase 1 study4 reported in September 2020 in the New England Journal of Medicine, noted Dr. Govindan, who was one of the investigators.
The drug is being tested in numerous settings, including in combination with either an EGFR inhibitor, immunotherapy, a MEK inhibitor, or chemotherapy. “It's a little too early to say how this will all play out, but everybody recognizes the fact that a single-agent therapy alone will not cut it,” he said.
If the proteomics study findings translate into the clinic, SOS1 inhibitors may become another strategy in the treatment of KRAS-mutated tumors, as well as the use of appropriate FDA-approved drugs to treat cases that harbor PRKCD.
And Dr. Govindan pointed out some additional good news: that, “by and large, most patients” are undergoing routine genetic sequencing so they—and their doctors—are aware of which mutations their cancers harbor.
“The commercial labs are very good at getting the data from the tumor samples or the blood,” he said. “I think this field is moving forward quite fast.”
Further data have since been released since the time of the writing of this article.
- 1. Gillette MA, Satpathy S, Cao S, et al. Proteogenomic characterization reveals therapeutic vulnerabilities in lung adenocarcinoma. Cell; July 29, 2020; 182(1): 200-225.
- 2. Chen Y-J, Roumeliotis TI, Chang Y-H, et al. Proteogenomics of non-smoking lung cancer in East Asia delineates molecular signatures of pathogenesis and progression. Cell; July 9, 2020; 182(1): 226-244.
- 3. Teodoridis JM, Hardie C, Brown R. CpG island methylator phenotype (CIMP) in cancer: causes and implications. ScienceDirect; September 18, 2008; 268(2): 177-186.
- 4. Hong DS, Fakih MG, Strickler JH, et al. KRASG12C inhibition with sotorasib in advanced solid tumors. NEJM; September 24, 2020; 383: 1207-1217.