Modeling Genotype-Associated Sensitivity to Receptor Tyrosine Kinase Inhibitors in
Lung Cancer-Derived Cell Lines

Jeffrey Settleman, PhD
Massachusetts General Hospital Cancer Center, Boston, Massachusetts

Non–small cell lung cancer (NSCLC) is the leading cause of cancer deaths worldwide, and is generally refractory to conventional chemotherapy when treated in the advanced stages. Selective small-molecule inhibitors of the epidermal growth factor receptor (EGFR) can yield good clinical responses in 10-30% of treated NSCLC patients. This activity, and the effect of somatic activating mutations on EGFR kinase activity in some tumors, suggested a new paradigm for the therapeutic management of advanced disease.⁸⁴ Recent studies have demonstrated that additional activated receptor tyrosine kinases (RTKs), including HER2, MET, ALK, ROS, and PDGFR, are present in small subsets of NSCLC tumors.^85-88 Pharmacological inhibitors of these kinases may similarly provide clinical benefit for patients whose tumors express them.

Among the challenges facing clinical oncologists in their efforts to evaluate such "personalized" cancer medicine is the fact that activation of particular kinases appears to be typically limited to a relatively small fraction of cases. This highlights the potential importance of genotype-based stratification of patients prior to treatment in order to increase the likelihood of therapeutic benefit. Additionally, the logistical difficulties associated with such clinical studies demand compelling preclinical justification of their value.

Recently, several preclinical studies (including our own) have made use of human NSCLC-derived cell lines to examine the relationship between RTK-associated gene alterations and cellular response to treatment with various kinase inhibitors. These studies have demonstrated, for example, a clear association between the presence of EGFR activating mutations and the cytotoxic activity of the clinically approved EGFR tyrosine kinase inhibitor (TKI) erlotinib in a large panel of NSCLC-derived cell lines.⁸⁵ Of 107 NSCLC-derived cell lines tested for sensitivity to erlotinib, six of the seven most drug-sensitive lines were found to harbor activating EGFR alleles. In contrast, among the remaining 100 erlotinib-insensitive lines, only two demonstrated EGFR activating mutations, including one that additionally harbors the well-documented erlotinib-resistance EGFR mutation T790M. Such findings validate the utility of established NSCLC-derived cell lines for modeling clinically relevant genotype-associated drug responses.

Examining the sensitivity of this large panel of NSCLC cell lines to a variety of other TKIs revealed additional genotypic correlations. For example, analysis of a selective inhibitor of the ALK (anaplastic lymphoma kinase) RTK showed that only a very small fraction of NSCLC lines demonstrated sensitivity.²⁷ The two most sensitive cell lines were found to harbor a chromosomal translocation that produces a fusion protein containing an activated form of ALK. Such translocation-associated ALK fusions have recently been reported to be present in 3-8% of primary NSCLC tumors.^87,89 Taken together, such findings raise the possibility that ALK translocation constitutes a biomarker that may be predictive of clinical response to selective ALK inhibitors. Consequently, recently developed pharmacological ALK inhibitors are currently being evaluated in this disease setting.

Another genotype-correlated drug sensitivity observed by NSCLC cell line profiling was the association of an amplified MET RTK gene and response to selective MET kinase inhibitors. Amplified MET is seen in a small fraction of NSCLC primary tumors and potentially constitutes a biomarker predictive of clinical response to MET kinase inhibition.^85,88 Therefore, MET kinase inhibitors are now being evaluated as therapeutic agents in NSCLC. MET amplification is also seen in about 20% of cases of acquired resistance to EGFR kinase inhibitors,^14,36 suggesting that MET inhibitors could provide clinical benefit as front-line therapy, as well as in the second-line setting upon disease progression following treatment with EGFR-TKIs.

Profiling of 107 NSCLC cell lines with sunitinib, a multikinase inhibitor approved for clinical use in advanced renal cancer and gastrointestinal stromal tumor (GIST), revealed only a single drug-sensitive cell line.⁸⁵ Significantly, this line demonstrated amplification of the gene encoding the PDGFR-αreceptor, which is one of the kinase targets that is very effectively inhibited by sunitinib (J. Settleman, unpublished observation). These cells also were sensitive to other PDGFR kinase inhibitors. Treatment with such agents resulted in the suppression of phospho-PDGFR as well as phosphorylation of downstream survival effectors, including AKT and ERK kinases, suggesting that these cells are "PDGFR-addicted" (J. Settleman, unpublished observation). A recent clinical study of sunitinib in NSCLC revealed an 11% response rate.⁹⁰ It will be of interest to determine whether these responders exhibit mutational activation of PDGFR-α.

Collectively, these NSCLC cell line-based studies have revealed their potential in identifying genotype-correlated sensitivities to a variety of targeted kinase inhibitors (Figure 4). However, it remains to be determined whether any or all of these observed preclinical correlations—other than EGFR—will be substantiated upon clinical evaluation. Genotypes involving activated RTKs have been identified in approximately one third of NSCLC cases. Yet at present it is unclear whether a significant fraction of the remaining cases will similarly be found to harbor activated kinases that may constitute additional therapeutic targets, or if alternative strategies will be required for these patients.

Figure 4

NSCLC-derived cell lines have also been used successfully to model clinically relevant mechanisms of acquired resistance to EGFR-TKI therapy,^14,21 revealing another important use of established cancer cell lines to model clinical response to drug treatment in NSCLC. We have recently begun to use such cell lines to model another aspect of drug treatment. Specifically, we have addressed the question of whether first-line EGFR-TKI therapy is likely to provide greater clinical benefit than second- line treatment. Currently, erlotinib is only approved for clinical use in chemotherapy-refractory NSCLC. However, recent clinical studies in which EGFR-mutant NSCLCs are treated with first-line EGFR-TKIs have revealed impressively high response rates.³³ A standard component of chemotherapy for NSCLC is platinum, which can activate at least one signaling pathway (PI3K kinase/AKT) whose activation has been associated with acquired resistance to EGFR-TKIs. Therefore, we tested the possibility that exposure of EGFR-mutant erlotinib-sensitive NSCLC cells to platinum in vitro would lead to a subsequent reduction in erlotinib sensitivity.⁹¹ This was indeed found to be the case. Although the IC₅₀ values for erlotinib in a cell viability assay comparing cisplatin-exposed cells and cisplatin-naive cells were similar, cisplatin-pretreated cells demonstrated a substantially increased ability to yield TKI-resistant clones in a longer-term clonogenic colony-forming assay. Moreover, this effect seemed to reflect the ability of cisplatin to activate the PI3K kinase/AKT cell survival pathway. These findings suggest that while initial response rates may be similar for EGFR-mutant NSCLCs in the first- or second-line setting, first-line treatment might lead to a longer period of progression-free survival. Considering the numerous clinical studies underway in which first-line EGFR-TKI therapy is being examined, data that address this possibility should be forthcoming.

NSCLC-derived cell lines have been used to model several aspects of the clinical response to selective RTK inhibitors, including genotype-correlated drug sensitivity, mechanisms of acquired TKI resistance, and response to sequential drug treatment. While some of these cell culture observations are supported by similar clinical data, other findings await clinical validation.

< < PREV

NEXT > >