Targeting Resistance to Targeted Therapies: Combating a Resilient Foe

In this issue of Clinical Cancer Research, Le and colleagues describe the various mechanisms through which resistance to tyrosine kinase inhibitors (TKI) arises in EGFR-mutated lung adenocarcinomas (1). Lung adenocarcinomas driven by mutations in the kinase domain of EGFR demonstrate a spectrum of alterations that vary in their sensitivity to tyrosine kinase inhibition. Nearly 50% of patients treated upfront with first- or second-generation TKIs become resistant to these drugs by gaining a second mutation in exon 20 (T790M). Osimertinib is active against EGFR T790M clones and has recently been shown to prolong survival in the upfront therapy of EGFR-mutated non–small cell lung cancer (NSCLC) when compared with erlotinib or gefitinib, possibly due its greater potency, better central nervous system penetration, and by precluding the emergence of T790M-mutated subclones. Nevertheless, as is the case with first- and second-generation TKIs, resistance to osimertinib emerges eventually as cancer cells gain additional alterations. There is a clear need to understand the molecular mechanisms underlying acquired resistance to osimertinib to prolong survival in patients with EGFR-mutated lung cancer. Genotyping of tumors and circulating tumor DNA collected from patients prior to osimertinib initiation and following disease progression has offered critical insights into the mechanisms by which cancers become resistant to osimertinib. It is important to keep in mind, however, that the majority of studies on osimertinib resistance to date have included only patients who had acquired or de novo T790M mutations.

Osimertinib inhibits EGFR through covalent binding with the C797 residue. Mutations at this location (such as C797S or C797G) are therefore expected to alter drug binding and promote drug resistance. Along with C797 mutations, other known mutations that confer osimertinib resistance such as L792H that alter drug binding were observed in 26% of samples progressing on osimertinib by Le and colleagues. As reported previously, these additional mutations in EGFR were predominantly observed in tumor samples that showed continued presence of T790M at the time of osimertinib resistance (2). Despite leading to osimertinib resistance, the C797 mutation does not seem to promote resistance to first-generation EGFR TKIs. Cell line studies have shown that the presence of C797 and T790M mutations on the same allele (cis orientation) confers resistance to third- and earlier generation EGFR TKIs, while it is possible that a combination therapy regimen with TKIs that independently target C797 and T790M mutations might be useful when these mutations are present on different alleles (in trans; ref. 3). However, data from studies such as the AURA trial, which examined the role of osimertinib in cancers progressing on first-line EGFR TKIs show that these mutations in EGFR often occur in cis (2).

Several studies have shown that nearly half of the tumors that become resistant to osimertinib demonstrate loss of the acquired T790M mutation at progression (1, 2). Inhibition of subclones containing T790M following osimertinib therapy and the dominance of subclones that have become TKI-resistant independent of T790M is likely explain these observations (Fig. 1). Acquired EGFR mutations such as L718Q and L844V were observed by Le and colleagues in a few samples that lost T790M (1). Some of these mutations, such as L718 mutations, have been shown in cell line studies to confer resistance to osimertinib, gefitinib, and erlotinib, when present in cis with L858R mutations (4). Cell lines with L718 mutations, however, retain sensitivity to afatinib in the absence of T790M. These findings suggest that the context in which various mutations are observed and their allelic orientation are likely to play a critical role in dictating EGFR TKI responses when resistance to osimertinib emerges following its upfront use.

In cancer cells that do not demonstrate additional EGFR mutations, which alter drug binding, TKI resistance often emerges through the activation of bypass signaling pathways (3). Cancer cells typically accomplish this by acquiring alterations such as amplification of MET, ERBB2, or KRAS, mutations in PIK3CA, BRAF, ERBB2, KRAS, and MEK1, or fusions in kinases such as BRAF, RET, and FGFR3, that activate MAPK signaling through alternate routes (2, 3). Le and colleagues also suggest a role for alterations in genes regulating cell-cycle genes (such as CDK4, CDK6, CDKN2A, and CCNE1) and activation of beta-adrenergic signaling as alternate mechanisms of osimertinib resistance (1). These findings will require further independent validation. EGFR-mutated samples also become resistant to TKIs including osimertinib by undergoing histologic transformation to small-cell lung cancer (SCLC). Biallellic inactivation of TP53 and RB1 is a hallmark feature of SCLCs and sequencing studies suggest that EGFR-mutated adenocarcinomas that demonstrate TP53 and RB1 inactivation in pretreatment biopsies are predisposed to SCLC transformation (5). Finally, a clear mechanism of osimertinib resistance is often not identifiable in many samples at progression. These tumors are likely to have become resistant to osimertinib through mechanisms that remain to be characterized including alterations in other less commonly studied genes (not routinely included in clinical assays) or noncoding RNAs. Epigenetic alterations and changes in tumor microenvironment could also certainly play a role in mediating resistance to EGFR TKIs.

A better understanding of the cellular process that fuel the emergence of resistance mechanisms is necessary to develop innovative strategies to treat EGFR-mutated NSCLC. Studies reported by Le and colleagues and others provide a window to study the complex landscape of EGFR-mutated NSCLC following TKI therapy. However, a comprehensive assessment of genomic alterations (whole genome, transcriptome, and methylomes) is necessary to get a clear picture of the alterations and mechanisms leading to treatment resistance to move this field forward.

Selection of preexistent resistant subclones that a tumor acquires stochastically, is one of the routes through which a tumor can become TKI resistant (Fig. 1; ref. 6). Genomic analyses of EGFR-mutated lung cancer prior to initiation of any therapy, demonstrate substantial intratumoral clonal heterogeneity with subclones often containing resistance conferring alterations in genes such as MET, BRAF, ERBB2, CTNNB1, or PIK3CA (7). Combination therapies that concurrently target EGFR and other proteins that drive resistance, such as MET, appear promising in this context.

Gradual acquisition of genetic alterations that mediate drug resistance by “drug tolerant persistent” cells that survive initial therapy is another mechanism through which EGFR-mutated cells become TKI resistant (Fig. 1; ref. 6). Cell line studies suggest that deregulation of apoptosis through activation of signaling through SRC and AKT is likely to play an important role in facilitating innate resistance to EGFR TKIs (8). Mutational processes, such as that driven by APOBEC enzymes, are likely to play an important role in enabling the emergence of resistance in these cells, as shown in tumors that undergo SCLC transformation (5). EGFR-mutated cancers also demonstrate a high degree of genomic instability, which is likely to act as a substrate for selection of treatment resistance (7). The association between TP53 mutations and alterations in genes that regulate the cell cycle in EGFR-mutated tumors and poor outcomes, as observed in some studies, is likely to be explained by the impact of these alterations on genomic stability.

Strategies that target the persistent state or processes that drive genomic instability could potentially prolong patient survival. For instance, experiments have suggested a potential role for agents inhibiting the antiapoptotic protein BCL2, insulin growth factor, or NF-κB signaling in targeting the drug-tolerant state (3). Combination regimens with TKIs and chemotherapy, antiangiogenesis agents, or ablative strategies for residual disease and oligo-progression, can in theory, prolong survival by reducing disease burden and reducing the substrate on which clonal selection can act. Meaningful clinical activity for these strategies has been demonstrated in several small clinical trials, and this study by Le and colleagues, in which osimertinib was continued beyond RECIST progression in selected cases, following radiation to progressing lesions (1). In addition, there could be a role for alternating the use of different EGFR TKIs, or EGFR TKIs with chemotherapy or other targeted therapies. Such strategies could delay the emergence of resistance by constantly varying the nature of selection pressure imposed on cancer cells. Finally, the optimal use of immunotherapeutic agents in EGFR-mutated NSCLC continues to be an active area of research. Exploring combination approaches with personalized vaccines or T-cell therapies that target drug-resistant subclones, and their optimal sequencing with other therapeutic modalities hopefully will yield durable responses in EGFR-mutated NSCLC in the near future.

R. Govindan is a consultant/advisory board member for AbbVie, Adaptimmune, AstraZeneca, Bristol-Myers Squibb, Celgene, Eli Lilly, EMD Sereno, Genentech, GlaxoSmithKline, Janssen, Merck, Nektar, Pfizer, Phillips Gilmore, and Roche. No potential conflicts of interest were disclosed by the other author.

Conception and design: S. Devarakonda, R. Govindan

Development of methodology: S. Devarakonda, R. Govindan

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Govindan

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Govindan

Writing, review, and/or revision of the manuscript: S. Devarakonda, R. Govindan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Govindan

Study supervision: R. Govindan

R. Govindan's work is supported in part by an NCI grant (U54CA224083 Washington University PDX Development and Trial Center).