Main

Major advances in molecular biology and high-throughput sequencing techniques have boosted cancer drug treatment strategies over the past decades.1, 2, 3, 4, 5, 6 The envisioned paradigm is the identification of genetic aberrations that are unique for defined subgroups of patients, and the development of drugs that specifically aim at oncogenic targets and their signaling pathways of these patients. In lung cancer, this development has led to the identification of oncogenic EGFR mutations that are present in up to 20% of adenocarcinoma patients7, 8, 9 and confer exquisite sensitivity to EGFR inhibitors (for example, gefitinib, erlotinib).10, 11, 12 Another example is the identification of EML4-ALK, the first fusion gene that is causally linked with the development of lung cancer.13 Encouraging first clinical studies show that in patients with EML4–ALK fusions overall response rates of up to 57% can be observed when treated with crizotinib.14

However, the known ‘druggable’ targets are primarily enriched in the subgroup of adenocarcinomas of never smokers.15 In addition, molecularly targeted therapeutics such as erlotinib, gefitinib and cetuximab, as well as conventional chemotherapeutics such as pemetrexed are poorly active in squamous-cell lung cancer patients.11, 16 Drugs such as bevacizumab or sorafenib that primarily inhibit vascular endothelial growth factor receptor (VEGFR) signaling are also not registered for the treatment of squamous-cell lung cancer patients due to increased risk of fatal hemorrhage.17, 18 Thus, the therapeutic repertoire for squamous-cell lung cancer patients is rather limited.

Several genomic studies tackled this unmet medical need by systematically characterizing the landscape of genetic lesions of squamous-cell lung cancer patients in order to define new therapeutically amenable targets. The first large-scale analysis of significant copy-number changes in squamous-cell lung cancer involving a cohort of 47 specimens revealed recurrent amplification of the 3q26.33 chromosomal locus harboring the transcription factor gene SOX2.19 Bass et al.19 were able to show that in SOX2-amplified lung cancer cells SOX2 functions as an oncogene and that its expression is required for proliferation and anchorage-independent growth. Further studies confirmed recurrent amplification of SOX2 in up to 20% of squamous-cell lung cancer patients and further substantiated its role as a lineage-survival oncogene in these tumors.20, 21 Unfortunately, the amplified SOX2 gene does not represent a chemically ‘druggable’ target and its amplification does not lead to the activation of pathways that are immediate and obvious drug targets. However, studies involving a SOX2-driven mouse tumor model allowed the identification of genes such as CyclinD1 to be specifically overexpressed in these tumors.22 Such findings may represent rationale for the study of therapeutics that inhibit cell cycle progression (for example, polo-like kinase 1 (PLK1) inhibitors, cyclin-dependent kinase (CDK) inhibitors) but also allow the exploration of ‘druggable’ genes that are synthetically lethal with SOX2-amplification in squamous-cell lung cancer.

In an independent study our group analyzed recurrent copy number alterations in a cohort of 155 squamous-cell lung cancer specimens that were collected as part of an international collaborative initiative, the Clinical Lung Cancer Genome Project (CLCGP). Using Affymetrix 6.0 SNP arrays (Affymetrix, Santa Clara, CA, USA) we were able to identify the receptor tyrosine kinase fibroblast growth factor receptor 1 (FGFR1) to be focally and significantly amplified in a subgroup of squamous-cell lung cancer patients.23 Two independent cohorts showed that (a) FGFR1 amplification is highly enriched in the squamous-cell histology and that (b) FGFR1 amplification can be detected by fluorescence in-situ hybridization analysis and that it is present in approximately 15–20% of such patients. On the basis of these findings we used a genetically characterized lung cancer cell line collection24, 25 to analyze the cytotoxic effects of the FGFR inhibitor PD173074 as a function of genetic alterations present in these cells in an unbiased fashion. We observed that the cell lines with the highest sensitivity to the compound were enriched for lines bearing FGFR1 amplifications. In some of these lines FGFR1 inhibition induced apoptosis.23 Consequently, when grafted subcutaneously into nude mice, FGFR1-amplified cells formed tumors that regressed after oral application of the FGFR inhibitor. Further studies of the relevant pathways involved in signaling of FGFR1-amplified cells revealed that the MAPK was primarily inhibited upon inhibition of FGFR1 (Figure 1). This is in line with previous reports that clearly show preferential activation of the MAPK pathway in lung cancer cell lines that were stimulated with the FGFR1 specific ligands FGF2 and FGF9.26 However, further studies will be needed to clarify the role of the PI3K-signaling and other oncogenic pathways in FGFR1-amplified lung cancer cells.

Figure 1
figure 1

Intracellular signaling engaged by amplified FGFR1 and mutant DDR2 in squamous-cell lung cancer. Schematic presentation of signaling cascades that mediate anti-apoptotic and proliferative effects in either FGFR1-amplified (left panel) or DDR2-mutant (right panel) squamous-cell lung cancer cells. As depicted by the size of the arrows amplified FGFR1 primarily engages the MAPK but not the PI3K pathway to execute its oncogenic signaling. Mutant DDR2 is known to engage SRC in order to mediate its oncogenic signaling in squamous-cell lung cancer.

Full size image

Finally, we showed that the viability of FGFR1-amplified cells is significantly reduced when expression of FGFR1 is silenced by short-hairpin RNAs and that the introduction of a resistance allele of the FGFR1 protein abrogates inhibitor-mediated cell death. These results identify FGFR1 as the driving oncogene in the 8p amplicon in squamous-cell lung cancer. Together, these observations introduce the first therapeutically amenable oncogene alteration in squamous-cell lung cancers.23, 27 An independent study recently confirmed our initial observation.28

These findings add to the growing compendium of genetic alterations that can be found in the FGFR gene family across different cancer types. For instance, in subtypes of bladder cancer FGFR3 mutations are found in up to 50% of cases, FGFR2 is mutated in endometrial cancer and amplified in gastric cancer, and breast tumors frequently harbor amplifications of FGFR1 as well.29, 30, 31, 32 One of the challenges of personalized cancer medicine is to prospectively identify patients with such genetic aberrations as potent FGFR inhibitors have been developed and are currently undergoing clinical evaluation in genetically defined subgroups.33, 34, 35 FGFR1 amplifications can be detected in routine clinical setting by fluorescence in-situ hybridization. Initial experience at our own center showed that this assay is robust and accurate enough for clinical screening. Accordingly, prospective molecular screening of squamous-cell lung cancers for the presence of FGFR1 amplifications has been established at the cancer center of the University of Cologne. The goal of these efforts is to include patients with positive tumors into trials assessing the safety and efficacy of FGFR inhibitors.

In a recent study focusing on the identification of novel recurrent kinase mutations in squamous-cell lung cancer a group led by Matthew Meyerson analyzed three different squamous-cell lung cancer cohorts and could show that up to 4% of patients harbor mutations in the discoidin domain receptor 2 (DDR2) gene.36 Unlike most receptor tyrosine kinases the DDR receptor family signals in response to fibrillar collagens that have different specifities for the two known DDR receptors that are primarily expressed in either epithelial cells (DDR1) or interstitial cells (DDR2).37 The authors also found mutant DDR2 in two non-small-cell lung cancer cell lines that were growth-inhibited upon shRNA-mediated knockdown of DDR2. Furthermore, treatment with the FDA- and EMEA-approved src-abl inhibitor dasatinib, which has potent inhibitory activity on DDR2, effectively killed DDR2-mutant lung cancer cells. Furthermore, ectopic expression of mutant DDR2 led to growth factor-independent growth of Ba/F3 cells through activation of SRC signaling (Figure 1). Phosphorylation of wild-type DDR receptors and its downstream targets such as Ras largely depends on SRC activity37 but the importance of SRC signaling in DDR2-mutant cells remains to be clarified. The ectopic expression of mutant DDR2 also resulted in anchorage-independent growth of NIH3T3 in soft agar suggesting that the lung cancer-derived DDR2 mutations are oncogenic.36

The authors went on to show that DDR2-mutant cell lines form tumors in nude mice that regressed upon oral treatment with dasatinib. Finally, a squamous-cell lung cancer patient, who had responded to a combination treatment of dasatinib and the EGFR inhibitor erlotinib as part of a clinical trial,38 was found to bear a DDR2 mutation in the kinase domain. Unfortunately, it could not formally be shown that the mutation was actually somatic and the combination treatment makes it difficult to establish a causal relationship between the mutation and the response to the treatment. However, the tumor did not bear an EGFR mutation and the mutation in DDR2 (S768R) was in close proximity to one of the mutations recurrently found in the sequencing screen. These findings are therefore compatible with a dominant role of the DDR2 mutation in driving tumorigenesis in this patient and in rendering the tumor susceptible to treatment with dasatinib. In summary, these results provide evidence that mutations in DDR2 are important oncogenic events inducing therapeutically tractable oncogene dependency. As dasatinib is an FDA-approved drug these findings have immediate implications for the treatment of squamous-cell lung cancer patients and should rapidly translate into clinical reality.

Overall the discoveries of FGFR1-amplified and DDR2-mutant squamous-cell lung cancers give hope for a subgroup of patients that had previously been excluded from virtually all targeted cancer strategies as they significantly expand the list of druggable targets in squamous-cell lung cancer (Figure 2). An intriguing aspect is the fact that most of these patients were smokers and thus the predominant conception that only lung cancers of never smokers can be driven by monogenic ‘driver’ events such as EGFR mutations may not be correct. Large-scale genomics initiatives with a particular focus on squamous-cell lung cancer such as the TCGA (The Cancer Genome Atlas; http://cancergenome.nih.gov/) and the CLCGP (PI: Roman Thomas) efforts will clarify whether additional driver oncogenes are responsible for the growth of squamous-cell lung cancers and other histologically distinct subtypes such as small-cell lung cancers and others. Additionally, initiatives such as mycancergenome (http://www.mycancergenome.org/) have been launched with the goal of cataloging therapeutically tractable genome alterations in cancer. Such initiatives provide an important tool for the clinical community for making decisions based on the genomic landscape of cancers and are thus indispensable for translation of basic genomic discoveries into clinical practice.

Figure 2
figure 2

Overview of genetic aberrations that are present in squamous-cell lung cancer. A simplified overview of genetic aberrations found in squamous-cell lung cancer patients based on selected manuscripts discussed in the review and the COSMIC database (http://www.sanger.ac.uk/genetics/CGP/cosmic/). The mutations are mutually exclusive except for PIK3CA mutations. amp, amplification; Mut, mutation.

Full size image

If validated in clinical trials, analyses of FGFR1 amplifications and DDR2 mutations will become part of the routine molecular diagnostic workup of squamous-cell lung cancer. It will therefore be important to develop analytical approaches to address these needs in the near future. Lessons learned from the use of erlotinib and other EGFR inhibitors might help to streamline the endeavors of clinicians, pathologists and researchers alike and help to make these exciting findings a success story for the treatment of squamous-cell lung cancer patients.