Targeting the evolution of drug resistance in lung cancer

ABSTRACT

Even in their intended disease subset, Epithelial Growth Factor Receptor tyrosine kinase inhibitors leave behind residual disease eventually resulting in acquired resistance. Our study indicates that this process is driven by Aurora Kinase A. Inhibition of Aurora signaling may prevent the onset of acquired resistance and counteract progressive disease.

The development of Epithelial Growth Factor Receptor tyrosine kinase inhibitors (EGFR-TKIs) has transformed the treatment paradigm for EGFR mutant (EGFRm) non-small-cell lung cancer (NSCLC). Activating mutation in EGFR such as Exon 19 deletion or L858R is present in 15–20% of NSCLC patients.¹ Although EGFR-TKI such as erlotinib, gefitinib, afatinib, and osimertinib are approved as first-line therapy in EGFRm tumors, 98% of patients have an incomplete therapeutic response resulting in residual disease.² There is a direct link between the magnitude of residual disease and shorter progression and limited survival benefits from targeted therapy.³ Acquired resistance to EGFR-TKI occurs through tumor evolution that involves the emergence of distinct genetic alterations as well as non-genetic changes. In approximately 70% of patients genetic resistance mechanisms can be detected upon progression such as EGFR C797S, MET amplification and ERBB2 amplification.⁴ This leaves approximately 30% of patients for which non-genetic mechanisms of resistance cannot be detected. Both in cell-free DNA and in autopsy studies in patients that progress on EGFR inhibitors, multiple potential resistance causing genetic alterations are often found to co-exist in the same patient, undermining the rationale for targeting a single genetic driver of resistance sequentially.⁵ There is compelling rationale and urgent need for approaches that prevent the evolution of resistance, potentially by abrogating the residual disease reservoir.

The origin of genetically defined drug-resistant subclones is not well defined and can occur through selection from clones that pre-existed prior to therapy as well as cells that survive initial treatment and progressively acquire resistance causing genetic alterations.⁶ To understand how tumor cells transition between phases of drug sensitivity, residual disease and eventually acquired resistance, we developed in vitro models of acquired resistance in EGFRm cell lines treated with EGFR-TKIs osimertinib and rociletinib.⁷ Sensitive cells yielded maximum cell death upon EGFR-TKI to give rise to residual cells which are about 0.1–0.3% of total population.⁸ Upon prolonged treatment, these residual cells resume their cell proliferation program and developed acquired resistance (AR). We used these AR models to identify small molecules that could overcome resistance. Aurora kinase inhibitors exhibited strong synergy with EGFR-TKIs in AR cells to abrogate cell proliferation and induce potent apoptosis in vitro and in vivo (Figure 1), indicating their ability to overcome acquired resistance.

Figure 1.

Strategies to combat epidermal growth factor receptor-tyrosine kinase inhibitor resistance with Aurora kinase inhibitors. Schematic representing the role the combination of osimertinib (a third-generation EGFR-TKI) and Aurora Kinase A inhibitors could play up-front in the primary treatment setting, at the point of maximal response to single-agent EGFR-TKIs (residual disease) or upon disease relapse (acquired resistance).

Further interrogation in AR models revealed increased activation of Aurora Kinase A (AURKA) indicated by auto-phosphorylation at Thr288, and an increase in the protein level of TPX2, an activator of AURKA.⁹ Genetic and molecular analysis revealed that AURKA activation is sufficient to engender EGFR TKI resistance and mitigate drug-induced apoptosis through the regulation of the pro-apoptotic factor BIM. Interestingly, this form is drug resistance is accompanied by the emergence of mitotic errors and polyploidy that are phenocopied by TPX2 or AURKA overexpression. Hence, AURKA catalyzes a form of resistance that is associated with chromosomal instability (CIN). Future work can investigate if CIN can be detected as a biomarker of this process and how it might contribute to emergent tumor behaviors such as heterogeneity and interplay with the immune system.

We interrogated TPX2 levels as a biomarker in patients who progressed on EGFR-TKIs. Clinical specimens derived at the time of diagnosis and progression on EGFR-TKI revealed 75% (9/12) patients expressed high levels of TPX2. TPX2 activation was present in patients progressing on erlotinib that also had bona fide genetic resistance mechanisms such as EGFR T790M and MET amplification, indicating that TPX2 upregulation may co-occur with other genetic drivers of acquired resistance. One possibility is that TPX2/AURKA may contribute to such events through structural alterations associated with CIN.

As a non-genetic event in cancer cells, AURKA activation is an adaptive response that emerges shortly within hours after initial EGFR TKI treatment. Interrogation of temporal signaling dynamics in residual drug-tolerant persister cells generated by 9 days of drug treatment revealed significant activation of TPX2/AURKA. Up front pharmacological co-inhibition of EGFR and AURKA can prevent the emergence of resistance, inhibit the growth of residual disease and abrogate acquired resistance in vitro (Figure 1). Moreover, a PDX model of the erlotinib-induced residual disease also indicated TKI induced hyperactivation of TPX2/AURKA and co-inhibition of EGFR and AURKA was able to induce strong anti-tumor responses in this model. These studies highlight the potential for Aurora kinase inhibitors to delay the onset of acquired resistance through up-front combinations, setting the stage for new clinical trials testing this concept.

In summary, our study has identified that non-genetic TPX2/AURKA activation is a molecular driver behind the formation, maintenance, and evolution of drug tolerance and resistance to EGFR TKIs (Figure 1). Overactivation of this axis induced mitotic stress coupled with the loss of apoptotic potential which are hallmarks of genomic instability. This genomic instability may act as fuel to accelerates cellular heterogeneity and drive tumor evolution in response to treatment. In recent years, aurora kinases have emerged as a versatile drug target and a number of small molecule inhibitors of aurora kinases such as alisertib (MLN8237), barasertib (AZD1152), and LY3295668 are under clinical trial in hematological malignancies and solid tumors. The only trial combining these two targets was a phase I trial of alisertib with the EGFR TKI erlotinib in EGFR wild-type NSCLC ¹⁰ which showed that targeting these two kinases was tolerable. Our study provides insight into non-genetic factors that contribute to EGFR TKI resistance and provides a potential new strategy to mitigate resistance by upfront rational polytherapy.

Funding Statement

This work was supported by National Cancer Institute grant nos. U01CA168370, NIGMS R01GM107671, and U54CA224081; National Institute of General Medical Sciences [R01GM107671].

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.