Final Report Summary - CURELUNG (Determining (epi)genetic therapeutic signatures for improving lung cancer prognosis)
Executive Summary:
Despite continuous advances, lung cancer (LC) is still the leading cause of cancer death worldwide. The extremely poor prognosis for LC patients is partly due to the lack of effective therapies. At present, most patients with pulmonary carcinomas are treated with chemotherapy. This essentially consists of classic cytotoxic drugs which only improve survival in small cohorts of cases. In spite of the rapidly growing understanding of the epigenetic and genetic profile of LC, such knowledge has contributed little to improving therapeutics.
Given the poor success of the traditional treatments but with the promise of novel and targeted drugs, CURELUNG aimed to concentrate on the discovery of novel therapeutic targets and on the study of the (epi)genetic background and biological pathways that will serve the purpose of tailoring patients for improving the efficacy of anticancer treatments.
Therefore, CURELUNG project emerged in January 2011 as the joint effort of leading European groups in exploiting its unique expertise with high-throughput technology and equipment, sharing experimental models and collections of human cancer samples from the different participants. After 42 months, the CURELUNG consortium has provided the cancer research community mainly with:
- The discovery of new tumor-specific gene alterations unveiling novel oncogenes and tumor suppressor genes.
- The establishment of a complete methylation profiling of lung cancer samples, together with the identification of a prognostic methylation signature in early stage lung cancer patients.
- New insights into the histomolecular classification of lung cancers. This will help in the distribution of lung cancer patients into groups with similar therapeutic management.
- The identification of novel target candidates for therapeutic interventions.
- The development of new lung cancer mouse models and the demonstration that patient-derived xenografts have great potential in drug development studies.
- The conduction of a phase III randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in non small cell lung cancer patients.
- The establishment of one of the world´s largest molecular screening network for lung cancer, i.e. the Network Genomic Medicine.
Altogether, although further research is needed, the input from CURELUNG will undoubtedly benefit the research community, allowing us to improve the lung cancer management and prognosis.
Project Context and Objectives:
Summary description of project context and objectives
Lung cancer (LC) is still the most lethal type of cancer worldwide. The extremely bad prognosis of LC patients is partly due to the late diagnosis of the disease, but also to the lack of effective therapies. Chemotherapy is applied in the majority of patients, but so far has only improved survival in very small cohorts. In spite of our rapidly growing understanding of the epigenetic and genetic profile of LC, the knowledge has not been much used to improve therapeutics.
In response to the call HEALTH2010-2.4.1-6: translational research on cancer with poor prognosis, the Curelung consortium has focused on the discovery and implementation of therapeutic approaches to increase LC patient survival. Given the poor success of the traditional treatments and the promise of the novel and targeted drugs, CURELUNG aimed to concentrate on the discovery of novel therapeutic targets and on the study of the (epi)genetic background and biological pathways that will serve the objective of tailoring patients for these upcoming anticancer treatments.
The CURELUNG project has been structured around three general and clear objectives:
OBJECTIVE 1: Discovery of novel (epi)genetically altered genes and characterisation of gene alteration patterns in LC
• ACTIVITY 1. Characterisation of candidate gene patterns and identification of (epi)genetically altered genes in lung cancer (LC) cell lines
Work packages involved: WP1 (Genetic and genomic profiling of LC cell lines), WP2 (Epigenetic profiling of LC cell lines)
The main objective of this activity was to assess the distribution of known (epi)genetic alterations in lung cancer and to identify novel alterations. Here we aimed to study individual genes or genome regions through a combination of traditional (automatic sequencing, FISH, methylation-specific PCR) and novel, high-throughput technologies.
The main goals achieved in this regard have been:
- The generation of a profile of genetic alterations at reported genes, determining their correlation with specific patterns of gene expression and with etiological and pathological characteristics of the lung tumors.
- The discovery of tumor-specific inactivation of genes such as MAX, MYST4 and PARD3 and activation of genes such as FGFR1, SETDB1, etc that constitute tumor suppressor genes and oncogenes, respectively.
- The methylation profiling of nearly 150 LC cell lines with the most up-to-date platform from Illumina, the Infinium 450k array. The analysis of 94 LC cell lines perfectly characterized in terms of response to the targeted drug gefitinib has led to de identification of 124 CpGs associated with response to the drug (reported in Deliverables 2.1 and 2.4).
- The methylation profiling of 444 LC biopsies and 25 non-tumor lung samples with the Infinium 450k array. Further analyses have led to the identification of 338 cancer related-differentially methylated CpGs and, importantly, we have validated 5 of them as biomarkers or prognostic value in stage I LC patients. This subgroup is only treated with surgery and potentially might benefit from an adjuvant therapy (reported in Deliverable 2.3).
• ACTIVITY 2. Verification and pattern of the molecular alterations in lung primary tumours.
Work packages involved: WP3 (Genetic-epigenetic alterations in LC primaries/biopsies), WP4 (Histopathology-immunostaining: in situ validation of (epigenetic) alterations).
The main objective of this activity was to verify the frequency and pattern of alterations at known and new oncogenes, tumour suppressors and others (from ACTIVITY 1) in a cohort of nearly 500 primary lung tumours.
The main goals achieved in this regard have been:
- Gene alterations identified in lung cancer cell-lines have been validated in lung cancer tumour samples, including those in the tumor suppressor PARD3, MYST4 and MAX.
- SME partners have provided sensitive assays for validation of genetic changes and optimised conditions, leading to implementation in clinical samples and publications.
- Different splicing events, miRNA and DNA methylation signatures for diagnosis have been validated
- The application of a histo-molecular classification which allowed to revise the WHO histopathological classification definition, affecting patient assignments in more predictive therapeutic groups.
OBJECTIVE 2: Discovery and preclinical validation of drug targets in LC
• ACTIVITY 3. Generation of predictive models for LC treatment
Work packages involved: WP5 (High-throughput cell line screening to identify therapeutically exploitable synthetic lethality), WP6 (Mouse and cancer cell models)
One of the main challenges in the field of LC is the design of successful strategies to prevent or inhibit the progress of the disease. Here we proposed to perform selective compound screening of the 50-80 lung cell lines in order to test directly the correlation between the (epi)genetic background of the cells and the response to specific therapies. Additionally, as new models for a more rational and optimized study of the drug efficacy are required, we planned to advance in the current models for predicting the efficacy of a new therapeutic tool (mouse models, xenografts).
The main goals achieved in this regard have been:
- 44 LC cell lines have been screened against a library of 267 small organic molecules of diverse chemical classes and targeting several oncogenic signalling pathways and molecular targets. As a result, this work unveiled that sensitivity to FGFR inhibition was enriched in FGFR1 amplified SCLC cell lines.
- A synthetic lethality of MYC-amplified SCLC cell lines with Aurora B kinase was found. Given the possibility that a particular subgroup of SCLC patients might benefit from this targeted therapeutic intervention, preclinical validation experiments with a pharmaceutical company are being planned (reported in Deliverable 5.1).
- Two lung tumor mouse models representing the two most prevalent histological subtypes: adenocarcinoma and squamous cell carcinoma have been developed (reported in Deliverable 6.1).
- We characterized the histopathological differentiation of patient derived lung cancer cells xenografted (PDX). We have studied the correlation between the original tumor and the corresponding PDX by means of: (i) metabolic parameters (SUV), (ii) tumor-stroma interactions and the frequency of tumor-initiating cells. In this regard, we have demonstrated that patient PDXs have great potential in drug development studies.
OBJECTIVE 3: Discovery and validation of (epi)genetic markers predicting responsiveness to specific novel inhibitory treatments (primary sensitivity or resistance to a given therapy)
• ACTIVITY 4. Target assessment in human/clinical trials
Work packages involved: WP1-4 (see titles above), WP7 (Selection of patients in clinical trials).
The work proposed in this objective aimed to involve the use of pre-treatment biopsies/tumours from LC patients in order to validate biomarkers discovered in the course of the previous activities that predict clinical response to a given specific agent.
The main goals achieved in this regard have been:
- The ideation and conduction of a Phase III multicenter randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in completely resected Stage II-IIIA NSCLC. The assessment of ERCC-1 and TS by qRT-PCR allowed the definition of 4 different genetic profiles and a tailored treatment based on pemetrexed, paclitaxel, cisplatin/gemcitabine or cisplatin (pemetrexed. Due to short follow-up, results will be available on a future time.
- In order to investigate the expression of the receptor tyrosine kinases in completely resected non-small cell lung cancer the gene expression of FGFR family of receptors was also investigated in the same case series. However, no significant results were obtained.
- One of the world´s largest molecular screening network for lung cancer, i.e. the Network Genomic Medicine (NGM), has been established. In 2013 the first large evaluation of NGM encompassing the analysis of about 5000 patients has been conducted. NGM helped us to efficiently recruit into personalized trials and to improve the survival of patients with EGFR mutations and ALK fusions among the collaborating NGM partners.
Project Results:
Work Package 1. Genetic and genomic profiling of Lung Cancer cell lines
The work package 1 aimed to identify novel genes, i.e oncogenes and tumour suppressor genes, altered in lung cancer (LC) using the latest genome wide sequencing and DNA copy number determination technologies. To achieve this goal we proposed testing LC cell lines, which are free of normal contaminant DNA, and verified the candidate genes later, in lung primary tumours specimens. In addition, within this work package we aimed to identify genetic and molecular determinants of the acquired resistance to novel targeted therapeutics. In the next pages we will describe all our discoveries that will certainly have an impact our current knowledge of LC biology and, possibly, to the LC clinical management.
Objective 1. Reported genetic alterations in cell lines
First, we selected 200 LC cell lines and study several of their characteristics. We found some differences in the distribution of gender, smoking habit and histopathologies, as compared to lung primary tumours. Particularly surprising, the squamous cell carcinoma (SCCs), which is a subtype of non-small cell lung cancer, is underrepresented. In addition, the non-smokers and the white ethnicity are overrepresented in LC cell lines. The reasons for the different distribution are unknown, but these will need to be taken into consideration if we want to extrapolate data from cancer cell lines to lung primaries. The gene alteration analysis in 160 of these cells confirmed that TP53 mutations are almost universal. The distribution of other major gene alterations in the different types of LC cell lines agrees with the previous general knowledge.
In addition, and within the context of a functional study on the candidate oncogene SOX4, we have observed that some of the genes that confer the gene expression signature of the small cell lung cancer (SCLC) are SOX4 targets. Our results show significant similarities between the SOX4 target genes and the SCLC gene expression profile which attest to the important role of SOX4 as a key element required for the development of SCLC (Castillo et al. 2012). This constitutes the one of the first publications which acknowledges the financial support of the CURELUNG.
Objective 2. Discovery and characterization of novel genetic alterations.
Next, we performed various genome wide analyses to assess for the presence of genetic alterations. Here, we must inform that, in addition to cancer cell lines, we used lung tumorgrafts. Lung tumorgrafts are lung primary tumours that have been implanted in the back of the mice and that, similarly to LC cell lines, have very little amount of normal contaminant cells. These models share the advantages of lung primary tumours and of the LC cell lines while are free of their respective disadvantages. We provided a list of genes that have been found in our study. Most of them have also been validated and studied, functionally.
The tumorgrafts that we have developed here have an important advantage: any possible targetable genetic alteration found in the tumour can be further studied and assayed in the animals. Although we do not report these types of analyses here, the CURELUNG project has been essential to allow for the development of such mice models. More importantly, these models may set the basis for the assay of novel drugs in lung tumours with a completely known genetic background.
Objective 3. Determination of genetic alterations involved in acquired resistance to targeted drugs.
Finally, in this work package we had proposed to work towards the comprehension of the mechanisms that account for the secondary (acquired) resistance to some of the targeted drugs used currently in the clinical settings. This work has been performed lung cancer cell lines. We have generated resistant cell lines from LC cells that were initially exquisitely sensitive to inhibitors of the MET and FGFR1 oncogenes and have studied the cells before and after becoming resistant. Our analysis allowed us to identify various changes that clearly differentiate both types of cells, including genetic and molecular shifts. Among the most interesting findings was the observation that the LC cells that have been made resistant in culture to the inhibitors of MET now have very high levels of the GAS2 growth factor and of its receptor, AXL. More interestingly, these cells had become very sensitive to other inhibitor, erlotinib, which is a molecule that inhibit the activity of another oncogene often activated in lung cancer, EGFR. Furthermore, the LC cell line that had been made resistance to the FGFR1 inhibitor now exhibits high levels of MET and activation of the AKT and mTOR pathway. These observations revealed that acquired resistance to a given inhibitor of a particular growth factor receptor triggers the activation of another growth factor receptor, to bypass the inhibitory effect of the drug.
Work Package 2. Epigenetic profiling of LC cell lines and biopsies
Nowadays, it is well known the high grade of complexity of the biology of cancer, particularly lung cancer (LC). In recent years, epigenetics has emerged as a new layer of information, becoming a key factor in order to complete the whole picture of cancer biology. Breakthroughs in technologies have enabled the possibility of genome-wide epigenetic research and revolutionized the concept of analyzing in an unbiased manner, not limited by previous knowledge. In this sense, the main objective of this WP was the characterization of the epigenetic modifications, mainly DNA methylation and microRNAs (miRNAs) expression, in LC cell lines (which are used as a model of “real” lung tumours), as well as to search new epigenetically regulated genes involved in resistance to specific targeted drugs. In addition, we have also performed the genome-wide methylation profiles of a great amount of LC primary biopsies, that have allowed identifying a potential DNA methylation signature associated with early-stage lung cancer prognosis.
Objectives 1 and 2. Characterize known tumour suppressor genes and identify novel cancer genes and miRNAs epigenetically altered in LC
More specifically, and regarding the epigenetic profile of LC cell lines, we have analyzed 126 cell lines together with 25 non-tumour lung cancer tissue samples. The Infinium 450k Human Methylation Array, which interrogates more than 450,000 CpGs (the main DNA positions where methylation takes place) and 21,000 genes has been used for this purpose. We have identified 17,576 differentially methylated CpGs, including 11,670 hypermethylated and 5,906 hypomethylatated ones in LC cell lines compared to non tumour controls. Hypermethylated events took place mainly in the promoters (regulatory regions) of genes involved in cell cycle and developmental processes. This is consistent with the theory that hypermethylation of the promoter regions of tumour-suppressor genes is a major event in the origin of many cancers. On the other hand, DNA hypomethylation events occur mainly in intergenic areas, where it leads to genomic instability.
MiRNAs are a large family of short non-coding RNAs that play regulatory functions and have been involved in many biological processes, including all the steps in cancer evolution. As the entire epigenetic machinery acts together to ensure the correct regulation of the gene expression, miRNAs may be affected by other epigenetic changes, such as DNA methylation. Thus, our next step was to focus in that specific feature from all the differentially methylated CpGs found in our previous analysis. We observed that, as previously noted, hypermethylated CpGs associated with miRNAs were located in promoters (which is concordant with tumour suppressor miRNA silencing) and hypomethylated ones are located mainly in low-density CpG regions, which could be a profile for onco-miRNAs.
Additionally, we determined the patterns of expression of relevant miRNAs across different lung cancer cell lines and immortalized human bronchial epithelial cells. We first investigated the expression of 24 miRNAs, as we have recently shown that, in plasma, this panel of miRNAs is able to detect lung cancer up to two years earlier than low-dose spiral computed tomography and also identifies tumour aggressiveness. The reciprocal ratios between expression ratios revealed a progressive trend for three miRNA ratios (miR 126/92a: 126/21; 28-3p/92a), recapitulating the deregulation observed in presence of clinical aggressive disease. Additionally we observed that other ratios between miRNAs belonging to signatures of high risk disease could stratify cellular models, highlighting the potential of this approach to identify specific miRNA patterns related to LC progression and offering insights on selection of cellular models to represent different stages during lung cancer progression.
In addition, we were highly interested in expanding our DNA methylation profiles beyond LC cell lines analyzing primary tumour samples from biopsies. We also used the Infinium 450k platform to deeply analyze the methylation profile of 444 tumour samples and 25 normal lung tissues (previously mentioned) (Sandoval J et al. 2013). Apart from generating a rich body of freely available information, we observed that the hierarchical clustering distinguished two groups of patients associated to significant differences in prognosis (estimated in relapse-free survival after surgery). This was also evident for the early-stage lung cancer cases, which undergo potentially curative surgical resection but still are at high risk of dying from recurrent disease, with a 5-year relapse rate of 35% to 50%. This is a particular important subgroup of patients, because although adjuvant platinum-based chemotherapy is beneficial in more advanced resected disease, in which most of the patients have a high risk of recurrence, it has failed to show a survival benefit for patients at early stages. One explanation for these negative data in the early stages could be the lack of biologic factors predicting their recurrence and the fact that, in the absence of useful biomarkers, all these cases are pooled, making it more difficult to draw meaningful clinical conclusions. For this reason, we looked for DNA methylation biomarkers that could distinguish between patients at low risk of relapse and those at high risk for whom adjuvant treatment could be prescribed. If successful, new treatment guidelines could be generated and more patients could benefit from an appropriate treatment.
Thus, we identified and selected those top genes that were more powerful in discriminating between the two groups of patients. As the 450k array is not currently available at hospitals and genome wide DNA methylation profiling cannot be carried out as a daily basis in the clinical setting, we investigated the role of these genes in an independent cohort of early-stage lung cancer patients by pyrosequencing, a more affordable large-scale approach. We ended up with a signature of 5 genes (HIST1H4F, PCDHGB6, NPBWR1, ALX1 and HOXA9) whose methylation can improve prognosis accuracy. Future studies will determine if this signature is also capable of improving treatment efficacy in lung cancer patients.
Objective 3. Characterize those epigenetic alterations involved in primary and acquired resistance to targeted drugs in LC.
Finally, one of our goals was the search of new epigenetically regulated genes involved in resistance to specific targeted drugs. This is relevant because, although one of the main advances in non small lung cancer treatment has been the development of targeted agents, all patients who initially benefit from these targeted therapies eventually develop resistance. Thus, in order to benefit a larger population of patients, it is crucial to develop a more detailed understanding of factors involved in the development of drug resistance and to identify additional therapeutic targets that may arise during the evolution of this process.
As DNA methylation is emerging as a potential driver influencing drug response, we decided to analyze the methylation profiles of 94 lung cancer cell lines perfectly characterized in terms of response to the EGFR targeted drug gefitinib, one of the first and more commonly used targeted agents in lung cancer. These analyses have allowed us to confirm the potential role of DNA methylation in influencing drug efficacy. Particularly, they have led to the identification of potential candidate genes conferring sensitivity or resistance to this particular agent. From those CpGs located in promoter, although some hypermethylated genes show correlation with better response, the great majority of hypermethylation events are related to drug resistance. As DNA methylation is potentially reversible, and DNA demethylating agents are currently available, these data open the possibility of using epigenetic drugs in order to sensitize or resensitize resistant tumours. Additionally, these epigenetic data can be used in the future to gain insight about the role of DNA methylation in the response to other targeted drugs.
Work Package 3. Genetic-Epigenetic alterations in LC primary/biopsies
Objectives 1 and 2. Validation of lung cancer associated alterations (in DNA sequence, DNA methylation, gene splicing or expression of miRNA or proteins) and correlating with etiological, histological and clinical variables. This has included potential implementation of discoveries by validation in clinical samples.
In order for scientific discoveries to make a significant impact in either our understanding of disease or its treatment, they must be validated by confirmation of their significance in a relevant independent study. This requires a co-ordinated effort by multiple clinical and laboratory scientists to identify the relevant samples for validation and to provide a significant number of samples to perform the test in a statistically valid manner. Many gene alterations in DNA sequence, DNA methylation, DNA translocation, gene splicing and gene expression have been validated and are described by other work packages. Here we give some specific examples where validation has strengthened the evidence that such alterations are significant in lung cancer, or has allowed us to determine their utility as biomarkers in clinical settings.
As described in Work Package 2, we have identified differences in DNA methylation (a modification of DNA that helps determine expression of proteins) that in a large series of lung cancers was apparently associated with those that had a poor outcome following surgery. In this discovery phase the DNA methylation changes were particularly associated with outcome in those cancers that were at an early histological stage (i.e. they were small and had not spread outside the lung). We therefore collected a new, independent set of these tumours and retested a representative selection of methylated DNA regions to validate our earlier findings. In doing so, we confirmed that DNA methylation of a set of five gene regions provided a good indication of early stage lung cancers that had a greater incidence of relapse.
Much of our understanding of lung cancer at the molecular level comes from the study of cells that grow in the laboratory after removal from tumours (tissue culture of cancer cell-lines). A large number of such cells have been produced around the world. Having catalogued the known molecular alterations in these, a number of particularly interesting gene alterations were identified (in Work Package 1), often because they appeared to be present mainly in specific types of cancer cell lines that share the same appearance. For example alterations in the gene PARD3 were found only in cell lines with a squamous histology (cancer cells having an elongated appearance) and MAX gene alterations were found in small cell cancer cells. As these DNA lesions were rare in cancers as a whole we performed a focused validation, looking mainly at collections of lung cancers with the same histological appearance. As well as being related to cancer relapse, or to specific groups of cancers, cancer-associated alterations may prove useful for diagnosis. Therefore, through technology validation, we have also demonstrated methods that allow those alterations to be assessed in clinical samples used for diagnosis, such as: formalin-fixed paraffin-embedded (FFPE) samples routinely used to examine histological features of tumours; bronchial lavage samples used in pre-surgical diagnosis of lung cancer; and plasma samples that could be used in early detection or disease monitoring.
MicroRNA (miRNA) molecules are short strands of genetic material that are involved in the regulation of gene expression (i.e. which proteins are made by different cells). Many have a functional role in lung cancer and in order to help identify these we have compared their levels in tumour tissue with that in normal lung tissue. In doing so, we identified 22 miRNA that had the largest and most consistent differences and validated this in an independent set of tumour and normal samples. We then combined multiple miRNA into a miRNA biomarker or signature, which was best able to discriminate between tumour and normal lung samples. This cancer-specific miRNA signature was shown in another validation to have a high accuracy in indicating if samples were derived from tumour material or not. All these studies were performed using samples frozen after being surgically removed from cancer patients, but many samples used clinically are both smaller (e.g. biopsies) and are preserved by being fixed in formalin and embedded in paraffin wax (FFPE). We therefore took a series of small samples from FFPE lung cancers and normal lung tissues and measure the miRNA signature. We were able to validate that the tumour tissues could be identified with a high sensitivity and specificity. Furthermore, this was true even though the tumour samples contained some normal cells (as is normal in cancer samples, where normal and cancer cells are frequently mixed together). We therefore believe that this miRNA signature potentially could be useful in helping clinicians to determine if small samples contain cancer cells, e.g. when examining biopsies form indeterminate nodules frequently discovered by CT screening studies.
Similarly, DNA methylation changes associated with cancer (such as those described above) can be useful in helping identify cancer patients from cytology samples (cell suspensions from fluid or disaggregated samples). For example, bronchial washing samples are a common diagnostic specimen, produced by washing part of the lung suspected of having a tumour and usually examined by cytology (microscopically examining the washing for abnormal cells). We have previously shown that a panel of four DNA methylation alterations found in lung cancer was better able to detect those with cancer than was the current practice of cytological examination. Combining these two techniques produced the highest accuracy. We have now validated additional DNA methylation changes that allow us to improve the accuracy of the DNA methylation biomarker panel and are expanding studies further to include DNA methylation changes discovered in Work Package 2. As these DNA methylation changes occur commonly across many cancers and it has been shown that cancer DNA is present as cell-free DNA (cfDNA) circulating in blood, we may also be able to detect lung cancer patients by detecting DNA methylation alterations in cfDNA from blood plasma. In order to validate DNA methylation biomarkers in plasma we have adapted measurement techniques to allow specific detection of DNA methylation at very low levels.
Objective 3. Deep sequencing forefront genomic platforms in subsets of lung cancer tumours
DNA sequencing studies, using deep sequencing technologies, that have contributed consideration of the lung cancer risk of individuals (determined from lifestyle, family history and medical questionnaire data combined into a lung cancer risk score) and been augmented by consideration of DNA methylation and miRNA in the same context. These studies are in addition to deep sequencing of lung cancer cell lines and primary lung cancers that are described in WP1.
Over recent years DNA sequencing technology has been developed that allows us to examine most of the genome (whole genome sequencing, WGS), those specific parts that code for proteins (exome sequencing) or smaller collections of regions with particular significance to the disease being studied (targeted sequencing). These are sometimes referred to as next generation sequencing (NGS) or deep sequencing; deep sequencing is so called because each base is read multiple times (hundreds or even thousands of times) which allows us to confidently identify differences in DNA sequence even when present as a minor fraction in the sample being analysed. This is important, because not only are cancer cells frequently mixed with normal cells in clinical samples, but different cancer cells can have different gene alterations (tumour heterogeneity).
We have used WGS, exome sequencing on a selected set of squamous cell carcinomas from people with different lung cancer risk and also demonstrated that targeted sequencing can be used to detect mutations in samples from FFPE material.
Lung Cancer Risk
One way forward for improved management and prognosis for lung cancer lies with early detection of this disease, but economics will dictate that screening for lung cancer will not be available for an entire population. Therefore, there is a major need to be able to identify those people at highest risk of developing lung cancer. To this end, lung cancer risk prediction models have been developed to estimate the probability of developing lung cancer over a period of time. The Liverpool Lung Project risk prediction model (LLP risk score) utilises risk factors such as smoking duration, previous diagnosis of pneumonia, prior diagnosis of malignant tumour, occupational exposure to asbestos and family history of lung cancer to estimate an individual’s probability of developing lung cancer within five years. The LLP risk model has been validated in three independent datasets and used to select eligible participants for the United Kingdom Lung Cancer Screening Trial (UKLS).
In one aspect of the DNA sequencing studies of the CURELUNG project, we have examined lung cancers from those people for whom LLP lung cancer risk was high and compared them to those who suffered from lung cancer despite having a low risk score. The aim was to both understand any biological difference between the tumours and potentially to provide molecular biomarkers to improve our lung cancer risk prediction. We also therefore examined differences in DNA methylation and miRNA and looked at the outcome of lung cancer in cases stratified by lung cancer risk score.
In order to validate our findings in an independent set of tumours, we determined a simplified risk score for lung cancer cases publically available datasets of molecular data (The Cancer Genome Atlas, TCGA).
From our studies we concluded that there were only very subtle differences in the gene alterations found in lung cancers that arise in those with very low cancer risk (<2% risk in the next five years) compared to those with higher cancer risk (>8% in the next 5 years). Low risk cases are usually younger patients who have smoked less and/or who have no family or personal history of cancer. It would therefore seem that when lung cancer arises in these individuals it is very similar at the molecular level to that in other cancer patients. This is valuable as it means diagnostic tests validated here should perform equally well in all lung cancers. However, it is interesting to note that despite their relatively young age and lower exposure to smoke-derived mutagens, low risk lung cancer patients also have the same poor outcome as other cancer patients; this is in keeping with the molecular similarities but suggests we do not yet fully understand the dynamics of lung cancer development. We will therefore continue to build on the studies performed as part of CURELUNG in order to gain a better understanding of the relationship between lung cancer risk factors and the molecular characteristics of lung cancer.
Work Package 4. Histopathological-immunostaning: in situ validations of (epi)genetic alterations
The main objective of this WP was the histomolecular classification of large cohorts of patients with lung tumours with histological criteria (histology and phenotype) (histomolecular features ) associated with specific (epi) genetic abnormalities and their prognostic and predictive value .
Objective 1. Histopathological classification of a large cohort of lung tumours obtained by immunostaining.
Histomolecular classification of lung tumours selected and included in this program by the partners using morpho-phenotypical subtypes, including subclassification of adenocarcinoma, squamous cell carcinoma, high grade neuroendocrine tumours (small cell lung carcinoma and large cell neuroendocrine carcinoma) and carcinoids , are derived from recent modifications of the last WHO classification recently revised on the basis of the international guideline recommendations we recently published (Travis WD et al. 2011).
On this basis we have performed histological review of about 1100 tumours from Partner WP5 (R. Thomas) and 1 (IDIBELL) and 4 (ULIV) and 1870 patients blocks from clinical trial patients registered in 2 arms of therapy , thus allowing prognostic and predictive factors( of chemotherapy success) to be evaluated for their influence on survival . We have used not only histological morphological class but differentiation and proliferation antigens TTF1, transcription factor and transcription lineage factor characteristic of adenocarcinoma, P63 transcription factor and its form P40 for squamous cell differentiation lineage, in association with CK5/6, P16, as well as protein markers on the pathway of oncogene driver (EGFR, mutant and wild type, IGFR, and ALK expression, ROS1 expression. The significant result is leading to the reclassification by review (histomolecular reclassification) of 16 % of lung carcinoma. There was a high correlation between cases such reclassified and the genomic- based classification. Histo- molecular reclassification allowed suppression of the class of large cell carcinoma to less than 1% of lung tumours as does genomic- based profiling. This is a strong foreground in the new concept of lung tumour classification which allowed us as WHO editors to thoroughly revise the histopathological classification definition and reassign the previous large cell carcinoma to Squamous or Adenocarcinoma according to histological phenotype. This new classification will affect patients assignments in more predictive therapeutic groups (including all targeted therapy).
Objectives 2 and 3. Histomolecular characterization of tumours with known (epy)genotype according to oncogene, tumour suppressor or copy number alterations and characterization of tumours at recurrence.
New gene fusion can be precisely identified by immunohistochemical markers to preselect and accelerate FISH testing and provide an FDA approved companion test for legitimating of crizotinib therapy. We thus reported for the first time the high sensitivity and specificity (98 %) of 2 specific antibodies directed toward ALK and ROS1 , to rapidly detect cases with one of the 2 new fusions / translocations involving ALK( EML4/ ALK) or ROS1 (CD74 and other partners) .We published the first evaluation of theses 2 antibodies on large series of adenocarcinoma tested for multiple genetic drivers (INCA Grenoble platform of tumour genetics, preventing pathologist to perform FIFH systematically in triple negative adenocarcinoma (neither EGFR , RAS or ALK mutations) and expedite the diagnosis of gene fusion relevant of crizotinib therapy. These 2 antibodies will soon be approved as FDA companion test allowing crizotinib (ALK and Met inhibitor) treatment, especially following the dissemination via the ALK atlas we participated to, disseminated at IASLC World Lung Cancer Conference, Sydney 2013.
We also provided in collaboration with partner 5 (R Thomas) in the identification of the most recent gene fusion described (CD74/NRG1) in japanese patients. We helped with a parallel caucasian french series to delineate the strict histomolecular contours of this new entity: all adenocarcinoma from non smoking women and subclassified as Invasive Mucinous Adenocarcinoma. This new histo-molecular entity will be included now in the next WHO classification (2015). It is highly predictive on simple histology of the presence of Ras mutation (75%) or CD74/NRG1 fusion (% undetermined in Caucasian patients) both in mutual exclusion, prone to be sensitive to Ras or Tyrosine kinase inhibition .
Neuroendocrine lung tumours have been delineated by histomolecular characterization in high correlation with their molecular profile. This is also a huge progress in the community of lung tumour oncology , in providing a genomics profile specific for each subtype of high grade Neuroendocrine tumours and low grade / intermediate grade carcinoid, showing for the first time a wide range of leading oncogenic events in theses tumours, opposing the high grade tumour small cell lung carcinoma to the low grade carcinoids. Identification in SCLC of five oncogene drivers suggests targets for future therapy, presently confined to cytotoxic chemotherapy with no change in the last 20 years . The discovery that carcinoids have no overlapping features with SCLC, implies that they cannot be considered as their precursors Carcinoids are driven as tumour proliferation uniquely by alteration of genes involved in chromatin modification, rather than oncogene mutation (Epigenetic oncogenesis). These are strong advances implying future targeted therapies and therapeutic challenges, that were not handled in the last 20 years. Indeed, the new genetic profile concept of lung neuroendocrine tumours is included in the next WHO classification 2015 (Pathology and genetics of Lung, Pleura, Thymic and Heart tumours).
Basaloid carcinoma has a dismal prognosis among other squamous cell carcinoma: we have under this Curelung program deciphered the molecular traits (gene expression RNA profiling a gene copy number alterations) which explain their poor prognosis and disastrous clinical behaviour. Accounting for 6% of lung tumour, this is a highly aggressive tumour with early recurrence and death, even at early stage I or II, surgically treatable. We have discovered their specific gene expression profile called “basaloid cluster” which identified short living tumour among all external series /databases containing squamous cell carcinoma of the literature (in silico) which tightly recapitulates cases with basaloid histo-molecular classification. This expression cluster is enriched in high cell cycle regulation, high expression of germ cell, embryonic and stem cell gene/pathways, low differentiation and chromatin regulatory genes (PCR) programs. We have published as primary results their pathognomonic marker SOX4 allowing fast and easy recognition of this dismal entity.
Epigenetic alterations promote tumour progression, recurrence and metastases. This was an obvious forefront result in at least 3 levels. A. the methylation profile established in collaboration of partner 1 which allowed a prognostic methylation profile across histological types and stage in lung cancer (J. Sandoval et al. 2013). B. the multiple chromatin modifiers (histone metyl- transferase, polycomb, SWI/SNF, etc), which drive carcinoid tumours, the histone acetylase gene mutations (CREBP/P300) which inactivate proper chromatin functions in SCLC (M. Peifer et al. 2012). C. The signature of 26 genes with ectopic illegitimate expression in lung cancer usually restricted to testis and placenta tissues, which are affected by strong hypomethylation of their CPG sites, in contrast with their methylation in normal testis. The expression of at least 3 of the 26 gene signature predicts short survival (early recurrence and metastases) and allows early detection of aggressive tumours, as well as proposition of a early epigenetic and vaccination therapeutic interventions (S.Rousseaux et al. 2013). Ectopic expression of tissue antigen restricted to certain normal tissues, is a form of identity crisis (deviation) common to highly aggressive tumours such as we demonstrated with an aberrant expression of Prolactin mRNA, which identifies highly malignant tumours including high grade neuroendocrine tumours among which SCLC and large cell neuroendocrine tumours: the prolactin mRNA presence was able to significantly affect prognosis across stage.
Lymphocytic infiltration has prognostic value in lung cancer clinical trials: this is a new discovery we made in analysing the 1576 patients paraffin blocks/sections to evaluate the level (0 or 1) of lymphocytic infiltration whatever phenotypes of theses tumour infiltrating lymphocytes . As a forefront results , this was the first significant factor across histomolecular classification subtypes which allowed stratification of patients for survival, cross -validated among world trials of cisplatin- adjuvant therapy in lung cancer ( LACE trials )(P=0. 009 for high infiltration and longer survival in overall and disease free survival.
Objective 4. Histomolecular characterization of tumours with resistance to therapy with known or new therapeutic agents.
Two patterns of adenocarcinoma are predictive survival benefit from cisplatin adjuvant chemotherapy: Using the largest worldwide patients cohorts randomized in clinical trials which we reviewed for histomolecular classification and survival correlations, we identified two architectural patterns of adenocarcinoma, shown to be of poor prognostic and discovered their strong predictive value. From 533 cases of centrally reviewed adenocarcinoma, we identified 2 of the 5 patterns (micropapillary and solid) conferring to patients receiving cisplatin based adjuvant chemotherapy significant survival benefit (P>001 in DFS and specific DFS) this was the first ever identified factor for stratification of future patients in clinical trials of cisplatin- based adjuvant chemotherapy.
The ERCC1 is not anymore considered as a stratification factor/ biomarker for Cisplatin Adjuvant therapy : we had previously reported that patients registered in cisplatin adjuvant therapy with low ERCC1 level evaluated by immunostainig (IH) with an ERCC1 antibody , derived significant survival benefit from cisplatin adjuvant chemotherapy in great contrast with patient entering the control surgical arm However in contrast with previous factors we could not validate and reproduce this results across world (Lace) clinical trials using the new antibody commercially provided This lead us to undergo further investigations in the frame of this Curelung program d to decipher and elucidate the reasons for this failure. We discover that ERCC1 enzyme displayed 8 different isoforms, variably distributed among normal tissues at tumours, but only one had efficient DNA repair in vitro. Unfortunately ERCC1 antibodies (6 different antibodies tested) were not specific for this isoform epitope thus depriving us from potential biomarkers for patient stratification among cytotoxic therapies. An array of DNA repair IH markers seems more rewarding than only one. In conclusion future qualification of IH markers will necessarily be functionally evaluated and validated on multiple trial series. New antibodies specific for functional isoforms will have to be produced .
Lung Tumors sampled at primary and metastatic site display the same oncogene driver number and profile, if not intended to treatment after surgery. We have provided primary results in comparing primary and metastatic tissue at s different sites of lung cancer .Although diversification in number and nature of passenger (non driver) mutations perceived as increase in genetic instability, we could not identify tumours which change (gain or loss) of oncogene driver ,susceptible to targeted therapy, at the step of metastatic process
Work Package 5. High-throughput cell line screening to identify therapeutically expoitable synthetic lethality
Cancer is a disease of the genome; genetic lesions (gene mutations, gene copy-number changes, structural genetic changes, etc.) lead to irreversible changes in the intracellular signal transduction pathways that the tumour cells become dependent upon. A new class of cancer therapeutics targeting specific signalling pathways activated by genetic lesions has shown clinical success. Due to the molecular heterogeneity of most solid tumours our strategy was to discover new genetic alterations that could act as novel targets for therapies and to define genetic markers that could determine the efficacy response or resistance to targeted therapies.
Objective 1. Performing high-throughput cell line screening in deeply LC cell lines, enriched for certain cancer genotypes.
Within the Curelung project, we have performed a genomic analysis of 60 SCLC cell lines (Sos et al. 2012). We have identified regions of recurrent copy number alterations and compared the pattern of alterations in the cell lines to our complementary sequencing study of primary SCLC tumours (Peifer et al. 2012). Remarkably, SCLC cell lines and primary tumours are highly similar, thus validating the use of this SCLC cell line panel for combined genomic and pharmacological perturbation experiments. Our cell line collection captures hallmark events of SCLC such as recurrent deletions of RB1 and PTEN, but also amplification of genes such as FGFR1. Additionally we identified recurrent and focal amplification of MYCL1, MYCN and MYC. Besides SCLC we systematically studied the molecular alterations of pulmonary carcinoids by copy number analysis, genome/exome and transcriptome sequencing. We observed frequent mutations in chromatin-remodelling genes. In contrast to SCLC and LCNEC (Large cell neuroendocrine lung tumours) TP53 and RB1 mutations are rare events, suggesting that pulmonary carcinoids are not early progenitors but arise through independent cellular mechanisms. These data also suggest that inactivation of chromatin-remodelling genes is sufficient to drive transformation in pulmonary carcinoids.
Objective 2. Linking tumour genetics to individual compound sensitivity in order to identify synthetic interactions of individual hard to target genomic aberrations.
The objective of our study was to use genomic data of different cell lines generated within the Curelung project and to link it with the activity of targeted drugs. This is an important step towards the identification of clinically active drugs that can be used for the treatment of genetically selected cancer patients. We selected LC cell lines based on the presence or absence of individual candidate aberrations. In total forty-four cell lines were screened against a library of 267 small organic molecules of diverse chemical classes and targeting several oncogenic signalling pathways and molecular targets.
One of the findings in this screen was that sensitivity to FGFR inhibition was enriched in FGFR1-amplified SCLC cell lines. However, one of the cell lines was only partly sensitive to FGFR inhibition, possibly due to loss of PTEN protein expression. In our complementary study focusing on the genomics analysis of small cell lung cancer specimen we confirmed FGFR1 amplification to be present in 6% of SCLC cases. These data strongly indicate that FGFR1 might represent an attractive therapeutic target in these tumors. However, we could demonstrate that the 8p12 locus containing the FGFR1 tyrosine kinase gene, shows a genomic heterogeneity and we found that MYC was co-expressed in 40% of FGFR1-amplified tumors. Tumor cells co-expressing MYC were more sensitive to FGFR inhibition, suggesting that patients with FGFR1-amplified and MYC-overexpressing tumors may benefit from FGFR inhibitor therapy, which may have implications for patient selection for treatment with FGFR inhibitors (Malchers et al. 2014). Since we are running a first‐in‐man trial of a novel FGFR inhibitor in FGFR1- amplified squamous cell lung cancers at our center already, we have convinced a pharmaceutical company to include patients with FGFR1‐amplified SCLC into a subsequent expansion phase/ phase II clinical trial of that FGFR inhibitor. Thus, we strive to funnel all potentially clinically relevant results from our basic discovery efforts into immediate clinical application.
As the most important finding of our cell line screening we found synthetic lethality of MYC-amplified SCLC cell lines with Aurora B kinase. Contrasting with findings in neuroblastoma (Otto et al. 2009), where Aurora A inhibition selectively killed MYCN-amplified neuroblastoma cells by depletion of cytoplasmic levels of MYCN protein, MYC levels were not depleted in SCLC cells and we found a predominance of synthetic lethal addiction to Aurora B as opposed to Aurora A. Given the possibility that SCLC patients with MYC-amplified might benefit from targeted therapeutic intervention by exploiting the synthetic lethal interaction with Aurora B, we are currently discussing with a pharmaceutical company to pursue further preclinical validation experiments. The company in question has a compound that exhibits profound selectivity and potency against Aurora B kinase.
In a next step we tested the relevance of biomarkers identified within Curelung to lung cancer carcinogenesis, implication on drug sensitivity or acquired resistance to drugs. Therefore we generated different modified lung cancer cell lines according to alterations of predictive value.
1. FGFR1: We have conducted transcriptome sequencing in FGFR1-amplified lung cancers with the goal of identifying the relevant FGFR1 splice variants and in order to discover possible modifiers of responsiveness to FGFR inhibition. These efforts resulted in the identification of oncogenic mechanisms of amplified FGFR1. We showed marked heterogeneity of the 8p12 amplification event in squamous cell lung cancer, which results from broad genomic rearrangements. Furthermore we observed high expressions levels of MYC associated with FGFR dependency across different cell line models. We have generated several cell lines expressing stably different isoforms of FGFR1 with and without MYC. Additional genotypes were generated to track precisely the oncogenic signalling downstream of amplified FGFR1. Supporting an oncogenic role for amplified FGFR1, we showed that the mesenchymal splice variants are transforming in vitro and in vivo. This effect was strongly enhanced by co-expression of MYC. The impact of high-level expression of MYC in mediating FGFR dependency was further strengthened by clinical observation of two patients with FGFR1 amplification and high MYC-expressing squamous cell lung cancer who responded to FGFR inhibition.
2. CREBBP: We have characterized several SCLC cell line genetically. In particular, we have identified lines with and without mutations in CREBBP or EP300. We have used these cell lines with the goal of functional characterization of the molecular effects of loss of these histone acetyl transferases. Of particular value in this context is a cell line lacking CREBBP mutations, in which we stably silenced this gene. Remarkably, despite the expected dramatic impact of silencing such global transcriptional regulator, we observed that the cells engineered to lack CREBBP expression proliferated more rapidly than the controls. Tumors with mutations and hemizygous deletions in CREBBP and EP300 did not exhibit a significantly different pattern of gene expression compared to wild-type tumours, suggesting that global changes in gene expression are not the predominant mechanism by which loss of HAT activity contributes to SCLC pathogenesis.
3. AURORA B: MYC is known to be involved in the pathogenesis of diverse cancer types. To test the relevance of MYC amplification in SCLC, we silenced expression of MYC and we functionally characterized the dependency of MYC-amplified SCLC tumours on Aurora B. We find that MYC-amplified tumours depend on the kinase activity of Aurora B for their survival. Our data provide a rationale for the testing of Aurora kinase inhibitors in genetically defined SCLC patient groups.
4. NRG1-CD74: By transcriptome sequencing of 25 lung adenocarcinoma of never smokers we have identified a novel CD74-NRG1 gene fusion, which resulted from a somatic genomic rearrangement. To study the impact of this novel gene fusion we transduced different cell lines with retroviruses encoding CD74-NRG1 and performed further experiments (NIH-3T2, H2052, H322 and H1568 lung cancer cell lines). Mechanistically, CD74-NRG1 leads to extracellular expression of the EGF-like domain of NRG1 III-3, thereby providing the ligand for ERBB2-ERBB3 receptor complexes. Ectopic expression of CD74-NRG1 in lung cancer cell lines expressing ERBB2 and ERBB3 activated ERBB3 and the PI3K-AKT pathway, and led to increased colony formation in soft agar. CD74-NRG1 fusions may represent a therapeutic opportunity for invasive mucinous lung adenocarcinomas, which frequently present with multifocal and unresectable disease, and for which no effective treatment exists.
Work Package 6. Mouse and cancer cell models
Objective 1. Testing incidence of biomarkers on lung cancer animal models (chemically induced models, tumorgrafts and genetically modified mice).
The main objective of this WP is to generate and use mouse and cancer cell models in order to test the relevance of biomarkers identified in the present project to lung cancer carcinogenesis, implication on drug sensitivity or acquired resistance to drugs. We first generated three independent clinically relevant mouse models in order to evaluate the implication of histological subtype, inflammation, or tumor heterogeneity on lung tumorigenesis:
1. Lung tumor murine models of pure histological subtype: urethane-induced ADC and NTCU-induced SCC. In these two models, we evaluated by immunohistochemistry the expression of proteins related with the VEGF pathway (VEGF, VEGFR2 and pVEGFR2) in both the urethane-induced ADC and NTCU-induced SCC models.
2. Chemically induced lung cancer mouse model in the presence of silicotic chronic inflammation and the tobacco smoke carcinogen NDMA.
3. Patient-derived xenografts models (PDX).
Objective 2. To test the contribution to response of resistance of a selected therapeutic drug in animal models
The efficacy of some of the more relevant anti-cancer therapies in lung cancer appears to be closely associated with the histological subtype of the tumor. Therefore, we have developed clinically relevant mouse models that resemble the most two most frequent human non-small cell lung cancer (NSCLC) histologies: adenocarcinoma (ADC) and squamous cell carcinoma (SCC). We have used these models to evaluate the effect of antiangiogenic agents (sunitinib and VEGFR2-blocking antibody DC101) on tumor progression in ADC and SCC. We have demonstrated the contrasting responses of NSCLC to antiangiogenic therapies depending on histology. We have shown that two anti-VEGFR2 therapies elicited tumor stabilization in ADC tumors. In contrast, VEGFR2 blockade in SCC caused hyperproliferation of tumor cells and increased the expression of stem cell markers, independently of intratumoral hypoxia. A major new finding of this project is that antiangiogenic therapies induce contrasting responses in NSCLC depending on the histological subtype (Larrayoz et al, 2014, Embo Mol Med). This result has important clinical implications in the design of future antiangiogenic therapeutic trials, and emphasizes the need for precaution when considering the possibility of enrolling SCC patients in clinical trials evaluating antiangiogenic drugs.
Another important aspect is the association between inflammation and lung tumor development. In order to evaluate the role of chronic inflammation on lung cancer initiation and progression, we developed a chemical-induced carcinogenesis mouse model that comprised the administration of a low tumorigenic dose of NDMA, a carcinogen present in tobacco smoke, in the presence of chronic silica induced lung inflammation. Our model is unique in both the type of inflammatory condition we use to promote cancer and the sequential analysis of malignant lesions that allowed us to study separately the effects of inflammation in preneoplastic and full neoplastic lesions. Our findings suggest that chronic inflammation generates a favourable microenvironment for the generation of preneoplastic lesions that potentially contributes to the progression of adenomas to adenocarcinomas. In summary, in this study, we developed a novel mouse model of lung cancer that allows for the study of the interaction of chronic inflammation with lung tumorigenesis. Our study provides a useful new multistep animal model for examining the specific role of chronic inflammation in promoting lung carcinogenesis and identifies for the first time molecular mechanisms by which silica-mediated inflammation creates a favourable microenvironment for tumor progression. Elucidation of the interplay between inflammation and lung carcinogenesis will help in the development of preventive strategies to slow down or avoid the generation of premalignant lesions and the conversion to malignant tumors. The three above mentioned animal models are currently being used for the study of the mechanisms of action, sensitivity and resistance to other molecular targeted drugs apart from the antiangiogenic therapies already described.
Objective 3. To evaluate the effect of selected molecular changes on the phenotype of tumour cell lines.
From the mice models developed in objectives 1 and 2, tumors we also established two new lung adenocarcinoma cell lines which can be successfully transplanted into balb-c syngeneic mice either orthotopically or as metastatic models. This new syngeneic mouse model for lung cancer and metastasis in Balb/c mice present similarities in genetic alterations with the human disease. Hallmarks for this novel model are cancer stem cell properties, mutations in KRAS (G12D) and WWOX deletion, which could faithfully recapitulate human lung carcinogenesis. (Bleau et al. IJC 2014).
We have also developed a valuable tool for preclinical evaluation of novel therapeutic strategies in cancer: the so-called patient-derived xenografts models (PDX). These models, obtained by direct implants of tissue fragments in immunocompromised mice, have great potential in drug development studies because they faithfully reproduce the patient's original tumor for both immunohistochemical markers and genetic alterations as well as in terms of response to common therapeutics. They also maintain the original tumor heterogeneity, allowing studies of specific cellular subpopulations, including their modulation after drug treatment. To evaluate the performance of this model for pharmacological testing we have also started treatment of PDX with standard lung cancer chemotherapeutics starting with cisplatin and antiangiogenic agents.
In the context of this project, PARD3 gene was defined as a potential tumour suppressor gene in squamous cell carcinoma of the lung. We genetically reconstituted the PARD3 gene in H157 cell line in order to evaluate the role of this gene in lung cancer. We concluded that restoration of wild type PAR3 reduces tumor invasiveness and prevents formation of metastasis “in vivo”.
One of the major focuses of this project was the identification of novel genomic alterations in small cell lung cancer and on discovery of novel synthetic lethal interactions by performing chemical vulnerability screens. Part of this effort was a thorough functional validation program in order to test the hypothesis that these alterations play a functional role in the pathogenesis in SCLC. In the context of WP5, we have generated many genetically manipulated since the launch of the project. We have evaluated the effect of the expression of some of the biomarkers described in WP1-3 in these cell lines. Our results demonstrate the existence of a synthetic lethal type of interaction between MAX and BRG1 in small cell lung cancer cells, and raise the possibility of developing a therapeutic strategy for patients with MAX-deficient tumors (Romero et al. 2014 Cancer Discovery).
By screening relevant compounds across SCLC cell lines we showed that Aurora kinase inhibitors are effective in SCLC cell lines bearing MYC amplification, which occur in 3-7 % of SCLC patients. We found the PLK1 inhibitor BI2536, the ROCK1 inhibitor GSK269962A and the pan-Aurora inhibitor VX680 to be specifically active in these cells. Our results show that Aurora dependency in SCLC primarily involved Aurora B, required its kinase activity, and was independent of depletion of cytoplasmic levels of MYC (Sos et al. 2012 PNAS).
Objective 4. To test the LC cells that have acquired resistance to drugs for genetic and epigenetic alterations
The development of acquired resistance to tyrosine kinase inhibitors (TKIs) is a major clinical hurdle for the success of these treatments in cancer patients. Here, we aimed to elucidate novel determinants of the acquired resistance to the MET and FGFR1 inhibitors in lung cancer. To this purpose, we first selected lung cancer cell lines according to the genetic alterations related with the TKIs. Subsequently, we generated multiple clones that have acquired resistance to treatment with TKIs and performed exome and transcriptome sequencing of the resistant clones.
Work Package 7. Selection of patients in clinical trials
Objective 1. Selection of LC patients undergoing clinical trials for targeted therapies.
In agreement with the objective 7.1 the WP 7 contributed to the retrospective collection of tissue samples from patients with stage I non-small cell lung cancer completely resected to generated data about a DNA methylation signature. DNA was extracted from tumour tissue of more than 100 samples and sent to IDI for investigation. The contribution to this task is indicated in the main related publication (Sandoval J et al. 2013). In order to investigate the expression of the receptor tyrosine kinases in completely resected non-small cell lung cancer the gene expression of FGFR family of receptor was also investigated in the same case series without obtaining significant results.
Objective 2. Establish and validate novel predictive models based on gene alterations to help clinicians in the therapeutic decision making in LC.
In agreement with the Objective 7.2 the main task of the WP 7 was the ideation and conduction of a Phase III multicenter randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in completely resected Stage II-IIIA NSCLC (EudraCT Number: 2008-001764-36), already started when the CURELUNG project was financed. This is the largest study never initiated in the area of the molecular predictive markers in the field of completely resected NSCLC. The study was conducted in approximately 22 investigational sites: 10 sites in Italy, 10 sites in Germany and 2 sites in Poland. The primary study objective was to evaluate overall survival in patients with completely resected Stage II-IIIA NSCLC treated with tailored adjuvant chemotherapy driven by the assessment of the pharmacogenomic profile of the resected primary tumour as investigated by quantitative real time polymerase chain reaction (qRT-PCR) of paraffin embedded specimens or with standard cisplatin-based adjuvant chemotherapy selected by the investigator (cisplatin/vinorelbine, cisplatin/docetaxel or cisplatin/gemcitabine). The secondary endpoints of the study included the evaluation recurrence-free survival with tailored adjuvant chemotherapy driven by the assessment of the pharmacogenomic profile of the resected primary tumour as investigated by qRT-PCR of paraffin embedded specimens or with standard adjuvant chemotherapy selected by the investigator, to assess the therapeutic compliance in the different arms, to assess toxicity with tailored adjuvant chemotherapy driven by the assessment of the pharmacogenomic profile of the resected primary tumour as investigated by qRT-PCR of paraffin embedded specimens or with standard adjuvant chemotherapy selected by the investigator and finally to comparatively assess the results of the quantitative assessment of messenger ribonucleic acid for Excision Repair Cross-Complementing 1 (ERCC1) and thymidilate synthase (TS) with the semiquantitative assessment of ERCC1 and TS protein assessment as investigated through immunohistochemistry.
The assessment of ERCC-1 and TS by qRT-PCR allowed the definition of 4 different genetic profiles. Within 30 to 45 days post-surgery, patients in each genetic profile was randomized to receive either a standard adjuvant chemotherapy (control arm) selected by the investigator (cisplatin/vinorelbine, cisplatin/docetaxel or cisplatin/gemcitabine) or an experimental treatment (tailored chemotherapy arm).
The experimental treatment arms are:
Genetic profile ERCC1 and TS assessment Experimental treatment
_____________________________________________________________________
1 High ERCC1 and high TS 4 cycles of single agent paclitaxel
2 High ERCC1 and low TS 4 cycles of single agent pemetrexed
3 Low ERCC1 and high TS 4 cycles of cisplatin/gemcitabine
4 Low ERCC1 and low TS 4 cycles of cisplatin/pemetrexed
The final statistical analysis at the end of the study will group together all the controls in one group (control group) and all tailored chemotherapies in another group (experimental group). The primary efficacy analysis will be performed both on all randomized patients (Intent-to-Treat) and on all evaluable patients. Efficacy analysis will be performed on an intent-to-treat basis. Survival curves will be estimated using the Kaplan- Meier method and compared with the log-rank test. The Cox proportional hazard model will be used for estimating Hazard Ratios after adjusting for relevant variables. Demographics and baseline characteristics will be summarized. Quantitative data will be summarized by arithmetic mean, standard deviation, median, minimum and maximum. Frequency tables will be provided for qualitative data. Adverse event and laboratory data, including laboratory toxicities have been summarized by worst National Cancer Institute Common Terminology Criteria Version 3.0 grade and listed for patients in the safety population.
The expected total number of patients was 684, therefore, it was planned to recruit up to 700 patients (350 per arm) in order to take possible dropouts into account. Due to 10% failures to amplify the RNA from tissues we further increased the sample size of 70 patients and now in July 2014 the study is completed. The study cannot be analyzed because of the low number of events (deaths) due to the limited amount of follow up time. Before the end of the year a publication focusing on distribution of ERCC1 and TS in the four identified genetic profile and correlation studies between biomarker expression and the main clinical-pathological characteristics will be presented in a major meeting and published in a peer reviewed journal. Correlation studies between gene expression and immunohistochemistry are also planned. These analyses are feasible only now to avoid any jeopardization of the study.
For the above reason deliverable 7.2 will be reached definitively at the time of the final analysis of the study, but preliminary information also included in deliverable will be reached at the end of 2015.
Objective 3. Identify/select patients with acquired resistance to a given treatment.
To fit with the objective 7.3 the cell line studies to understand the mechanisms of resistance to cisplatin and potential mechanisms of cisplatin sensitization have been performed using celestrol as potential sensitizer to cisplatin (publication pending).
Additionally, in a close interdisciplinary collaboration between the Lung Cancer Group Cologne (www.lungcancergroup.de) (Chair: Jürgen Wolf) and the Dept. of Pathology of the University Hospital of Cologne (UHC, Chair: Reinhard Büttner) and the Institute for Translational Genomics of the University of Cologne (chair: Roman Thomas) the Network Genomic Medicine (NGM, www.ngml.de) was established supported by funding of WP 5. Within NGM comprehensive genotyping of all known driver mutations in paraffin-embedded biopsy specimens obtained by routine diagnostic procedures was offered free of charge for all interested clinical collaboration partners in the catchment area of our cancer center, the Center for Integrated Oncology (www.cio-koeln-bonn.de). These included large lung cancer centers, smaller regional hospitals and office-based oncologists. In the meantime, more than 70 centers participate in this activity. In parallel, a clinical trial program was established focussing on early proof of concept (PoC) clinical trials to offer each patient with a detected driver mutation a personalized treatment approach (see lungcancergroup.de/clinicaltrials). In addition, a close collaboration was started with the clinical cancer registry of our cancer center as well as with the Epidemiological Cancer Registry of the state North-Rhine Westfalia. The number of collaboration partners increased continuously over time and in parallel the number of tissue samples analysed. The establishment of this molecular screening network enabled us to recruit patients with rare genetic subgroups efficiently in clinical trials in an internationally competitive level (Shaw et al. 2014). Also, within this network we were able to rapidly transfer new results into clinical practice and, for instance, identified and successfully treated the first European patient with a ROS1 fusion (Bos et al, Lung Cancer 2013) followed by the preparation and recent initiation of a European phase II trial led by the Lung Cancer Group Cologne for the evaluation of crizotinib in ROS1 positive lung cancer, which was initiated in May, 2014. In 2013 we conducted and published our first evaluation within NGM (Clinical Lung Cancer Project and Network Genomic Medicine, 2013). Here we presented the analysis of about 5000 lung cancer patients genotyped within NGM, reported the distribution of driver mutations, the correlation with clinicopathological features, the recruitment in clinical trials, and, for the first time, outside clinical trials, a significant overall survival benefit of EGFR mutated patients treated with kinase inhibitors vs. EGFR mutated patients treated with chemotherapy only (32.5 vs. 9.5 months) and of ALK-translocated patients treated with crizotinib vs. ALK-translocated patients treated with chemotherapy only (22 vs. 11 months).
Bibliography (alphabetical order)
- Bos M, Gardizi M, Schildhaus HU, Heukamp LC, Geist T, Kaminsky B, Zander T, Nogova L, Scheffler M, Dietlein M, Kobe C, Holstein A, Maintz D, Büttner R, Wolf J. Complete metabolic response in a patient with repeatedly relapsed non-small cell lung cancer harboring ROS1 gene rearrangement after treatment with crizotinib. Lung Cancer. 2013 Jul;81(1):142-3.
- Castillo SD, Matheu A, Mariani N, Carretero J, Lopez-Rios F, Lovell-Badge R, Sanchez-Cespedes M. Novel transcriptional targets of the SRY-HMG box transcription factor SOX4 link its expression to the development of small cell lung cancer. Cancer Res. 2012 Jan 1;72(1):176-86.
- Clinical Lung Cancer Genome Project (CLCGP); Network Genomic Medicine (NGM). A genomics-based classification of human lung tumors. Sci Transl Med. 2013 Oct 30;5(209):209ra153.
- Larrayoz M, Pio R, Pajares MJ, Zudaire I, Ajona D, Casanovas O, Montuenga LM, Agorreta J. Contrasting responses of non-small cell lung cancer to antiangiogenic therapies depend on histological subtype. EMBO Mol Med. 2014 Apr;6(4):539-50.
- Malchers F1, Dietlein F, Schöttle J, Lu X, Nogova L, Albus K, Fernandez-Cuesta L, Heuckmann JM, Gautschi O, Diebold J, Plenker D, Gardizi M, Scheffler M, Bos M, Seidel D, Leenders F, Richters A, Peifer M, Florin A, Mainkar PS, Karre N, Chandrasekhar S, George J, Silling S, Rauh D, Zander T, Ullrich RT, Reinhardt HC, Ringeisen F, Büttner R, Heukamp LC, Wolf J, Thomas RK. Cell-autonomous and non-cell-autonomous mechanisms of transformation by amplified FGFR1 in lung cancer. Cancer Discov. 2014 Feb;4(2):246-57.
- Otto T, Horn S, Brockmann M, Eilers U, Schüttrumpf L, Popov N, Kenney AM, Schulte JH, Beijersbergen R, Christiansen H, Berwanger B, Eilers M. Stabilization of N-Myc is a critical function of Aurora A in human neuroblastoma. Cancer Cell. 2009 Jan 6;15(1):67-78.
- Peifer M1, Fernández-Cuesta L, Sos ML, George J, Seidel D, Kasper LH, Plenker D, Leenders F, Sun R, Zander T, Menon R, Koker M, Dahmen I, Müller C, Di Cerbo V, Schildhaus HU, Altmüller J, Baessmann I, Becker C, de Wilde B, Vandesompele J, Böhm D, Ansén S, Gabler F, Wilkening I, Heynck S, Heuckmann JM, Lu X, Carter SL, Cibulskis K, Banerji S, Getz G, Park KS, Rauh D, Grütter C, Fischer M, Pasqualucci L, Wright G, Wainer Z, Russell P, Petersen I, Chen Y, Stoelben E, Ludwig C, Schnabel P, Hoffmann H, Muley T, Brockmann M, Engel-Riedel W, Muscarella LA, Fazio VM, Groen H, Timens W, Sietsma H, Thunnissen E, Smit E, Heideman DA, Snijders PJ, Cappuzzo F, Ligorio C, Damiani S, Field J, Solberg S, Brustugun OT, Lund-Iversen M, Sänger J, Clement JH, Soltermann A, Moch H, Weder W, Solomon B, Soria JC, Validire P, Besse B, Brambilla E, Brambilla C, Lantuejoul S, Lorimier P, Schneider PM, Hallek M, Pao W, Meyerson M, Sage J, Shendure J, Schneider R, Büttner R, Wolf J, Nürnberg P, Perner S, Heukamp LC, Brindle PK, Haas S, Thomas RK. Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat Genet. 2012 Oct;44(10):1104-10.
- Rousseaux S, Debernardi A, Jacquiau B, Vitte AL, Vesin A, Nagy-Mignotte H, Moro-Sibilot D, Brichon PY, Lantuejoul S, Hainaut P, Laffaire J, de Reyniès A, Beer DG, Timsit JF, Brambilla C, Brambilla E, Khochbin S. Ectopic activation of germline and placental genes identifies aggressive metastasis-prone lung cancers. Sci Transl Med. 2013 May 22;5(186):186ra66.
- Sandoval J, Mendez-Gonzalez J, Nadal E, Chen G, Carmona FJ, Sayols S, Moran S, Heyn H, Vizoso M, Gomez A, Sanchez-Cespedes M, Assenov Y, Müller F, Bock C, Taron M, Mora J, Muscarella LA, Liloglou T, Davies M, Pollan M, Pajares MJ, Torre W, Montuenga LM, Brambilla E, Field JK, Roz L, Lo Iacono M, Scagliotti GV, Rosell R, Beer DG, Esteller M. A prognostic DNA methylation signature for stage I non-small-cell lung cancer. J Clin Oncol. 2013 Nov 10;31(32):4140-7..
-Shaw AT, Kim DW, Mehra R, Tan DS, Felip E, Chow LQ, Camidge DR, Vansteenkiste J, Sharma S, De Pas T, Riely GJ, Solomon BJ, Wolf J, Thomas M, Schuler M, Liu G, Santoro A, Lau YY, Goldwasser M, Boral AL, Engelman JA. Ceritinib in ALK-rearranged non-small-cell lung cancer. N Engl J Med. 2014 Mar 27;370(13):1189-97.
- Sos ML1, Dietlein F, Peifer M, Schöttle J, Balke-Want H, Müller C, Koker M, Richters A, Heynck S, Malchers F, Heuckmann JM, Seidel D, Eyers PA, Ullrich RT, Antonchick AP, Vintonyak VV, Schneider PM, Ninomiya T, Waldmann H, Büttner R, Rauh D, Heukamp LC, Thomas RK. A framework for identification of actionable cancer genome dependencies in small cell lung cancer. Proc Natl Acad Sci U S A. 2012 Oct 16;109(42):17034-9.
- Travis WD1, Brambilla E, Noguchi M, Nicholson AG, Geisinger KR, Yatabe Y, Beer DG, Powell CA, Riely GJ, Van Schil PE, Garg K, Austin JH, Asamura H, Rusch VW, Hirsch FR, Scagliotti G, Mitsudomi T, Huber RM, Ishikawa Y, Jett J, Sanchez-Cespedes M, Sculier JP, Takahashi T, Tsuboi M, Vansteenkiste J, Wistuba I, Yang PC, Aberle D, Brambilla C, Flieder D, Franklin W, Gazdar A, Gould M, Hasleton P, Henderson D, Johnson B, Johnson D, Kerr K, Kuriyama K, Lee JS, Miller VA, Petersen I, Roggli V, Rosell R, Saijo N, Thunnissen E, Tsao M, Yankelewitz D. International association for the study of lung cancer/american thoracic society/european respiratory society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol. 2011 Feb;6(2):244-85.
Potential Impact:
The main dissemination activities and exploitation of results of the CURELUNG project are mainly scientific publications, conferences and workshops. There have been already published around X publications in scientific journals in the frame of CURELUNG during these three years and a half. Their results have also been exposed various conferences from all the participants, as it is shown in the attached list. With the aim to disseminate the results from the CURELUNG project, as well as to invite and share these results with the prominent researchers in (Epi)genetics profiling and cancer, the CURELUNG consortium organized its “Mid-Term Workshop on Lung Cancer: from Molecular Knowledge to Treatment”. The quality of all publications and conferences offered by all the participants in the CURELUNG project demonstrated the added-value and positive impact of this project in the basic knowledge of lung cancer. The publications and thus the conferences have been related to the main objectives of CURELUNG and all the participants have been contributed.
Potential Impact (WP1)
The work that has been developed within WP1 has provided with a list of novel genes that are altered in lung cancer. The genes are either activated, i.e. oncogenes, or billalically inactivated, i.e. tumour suppressor genes and some of these alterations are specific of a given histopathological type. Thus, our findings benefits the general population and a have strong socio-economic impact. By one hand, the histopathological type-specific appearance of some of the genetic changes found in this work package may contribute to a more precise histopathological classification of the lung cancer. This is important because the type of histopathology in a lung cancer patient guides the therapeutics. However, still more important are the potential clinical interest, especially for the discovery of novel lung cancer treatments. We have found novel oncogenes that are activated in lung cancer and these or their immediate downstream targets could serve as therapeutic targets for the development or the implementation of lung cancer therapeutics. It is also remarkable that we have proposed some strategies to targets some tumour suppressor genes, such as MAX, based on the synthetically lethal concept. This concept is defined as follows: two genes are called synthetic lethal if mutation of either alone is not lethal, but mutation of both leads to death or a significant decrease in organism's fitness. In this regard we proposed that inhibition of another tumour suppressor gene, the so called BRG1 (or SMARCA4) can suppress the growth of the SCLCs that are deficient for MAX. Finally, another important impact of the results obtained within WP1 is the discovery of some mechanisms that account for acquired resistance to MET and FGFR1 inhibitors. Altogether, these discoveries will surely have implications in improving the clinical management of the lung cancer patients.
Potential Impact (WP2)
The work that has been developed within WP2 has provided a great core of information regarding the main objectives initially planned. We have determined the epigenetic profiles of an extensive panel of LC cell lines, identified novel cancer genes and miRNAS with epigenetic alterations and described epigenetic alterations that account for resistance to targeted drugs in LC. Additionally, we have identified a prognostic DNA methylation signature for early-stage NSCLCs.
Particularly, the epigenetic profile of this large set of lung cancer cell lines will serve as a basis for numerous and different studies regarding DNA methylation and lung cancer, including its role in driving resistance to therapies. MiRNAs studies have confirmed in cell lines the potential of these models and their role in delineating tumour aggressiveness and behaviour. Most importantly, the prognostic DNA methylation signature that we have identified in early-stage NSCLCs has a great implication for society and, obviously, a corresponding socio-economic impact. NSCLC is the leading cause of cancer-related death and the poor prognosis of patients with NSCLC is associated with several factors, among which are late disease diagnosis and the small number of effective drugs. But, importantly, the absence of validated prognostic biomarkers could also be relevant, because even patients with stage I NSCLC who undergo potentially curative surgical resection are at high risk of dying from recurrent disease, with a 5-year relapse rate of 35% to 50%. Although adjuvant platinum-based chemotherapy is beneficial in more advanced resected disease, in which most of the patients have a high risk of recurrence, it has failed to show a survival benefit for patients at stage I. One explanation for these negative data in the early stages could be the lack of biologic factors predicting their recurrence and the fact that, in the absence of useful biomarkers, all stage I NSCLCs are pooled, making it more difficult to draw meaningful clinical conclusions. Our DNA methylation signature has shown as an efficient tool to distinguish those patients more prone to recur. Thus, it is likely that adjuvant chemotherapy could improve the prognosis of these particular cases. Importantly, it is expected that, with the extension of lung cancer screening campaigns —mainly based in low dose CT scan—, more patients will present in early stages, increasing the importance of improving the treatment in this particular subgroup of cases. The combination of early diagnosis and more effective therapies will be the key to reduce the lung cancer mortality.
Potential Impact (WP3)
The ultimate impact of the CURELUNG project will be on human health. The aim for this work package 3 has been to improve the outcome of lung cancer through improved understanding of the disease leading to advances in diagnostics and therapy.
The discovery of DNA methylation changes related to lung cancer relapse means that we may be able to develop tests that allow clinicians to identify those who do not respond well to therapy and require additional treatment or monitoring. Moreover, DNA methylation patterns specific for lung cancer can be utilised to improve cytological diagnosis of the disease in bronchial washings and potentially applied to circulating cancer DNA in blood (alongside lung cancer screening). By defining the DNA changes associated with specific groups of lung cancers we can help develop drugs that are specifically targeted to those subtypes with the change, i.e. personalised medicine. Finally, the identification of miRNA signatures specific for lung cancer will aid development of diagnostic tools for use in small biopsies, such as those likely to come from increased use of CT screening for lung cancer.
One key aspect of lung cancer’s lethality is late detection, cancers often being detected when they are at a stage when clinical interventions are limited in scope and in effectiveness. Lung cancer screening not only benefits individual patients by providing earlier detection and improved survival, but society as a whole. Health costs are saved and economic costs associated with loss of work and skills are reduced. CURELUNG discoveries will aid the implementation of lung cancer screening by providing important diagnostic tools required to differentiate cancers from the larger numbers of CT detected non-cancer nodules. Further exploitation opportunities are provided by the development of biomarkers to aid the pre-screening identification of those at greatest risk of having lung cancer, with economic benefit stemming from lower costs associated with better targeted screening programmes.
In order to change clinical practice and have a significant impact on the health of those who suffer from lung cancer, it is important that these studies identifying gene alteration or techniques employed in diagnostic or prognostic tests are validated. We have taken significant steps to validation as part of Work Package 3 and provided opportunities for commercial exploitation of these techniques (including by our SME partners), or implementation in further prospective validation in the healthcare setting. Both these exploitation opportunities provide a positive contribution to the expanding market in biomarkers and personalised, precision and stratified medicine, expanding the range of targeted therapies applicable to lung cancer patients.
Potential Impact (WP4)
The potential impact of our findings is essentially on the management of patient with lung cancer ; application of a histo-molecular classification which recapitulates and correlates , better than ever with the previous WHO classifications , the oncogenic (genetic or epigenetic events ) and is more able to predict patients prognosis ,to stratify patients for choice of therapies modalities ,to facilitate and reduce the time of establishment of the oncogene drivers profile leading to therapeutic choices.
The main dissemination activities is through the wide dissemination of these new concepts and findings in high impact journals offered to the wide community of medical oncologists ,lung biologist, respiratory clinicians, and pharmaceutical laboratories preparing plans for new targeted therapies; the presentation of several indicators and organisers of randomization for patients entering now in clinical trials; the inclusion of these findings and novel findings concerning the new lung cancer histopathological classification. The highest level of dissemination is indeed inclusion of all these discoveries and new concepts in the World Reference book for Pathologists, provided in 2015. Due to worldwide dissemination of theses publication and reference WHO book, exploitation by pathologist and oncologist including patients in therapeutic protocol and evaluating the results based on patients survival. This therapeutical trial will be based on this new histo-molecular classification. This represents a strong guaranty for further exploitation of theses promising findings and their derived patients benefit worldwide.
Potential Impact (WP 5)
Based on our results, we have now launched two First-In-Human dose- escalation studies in Germany together with Novartis and AstraZeneca, which are aimed at testing novel FGFR inhibitors specifically in patients with FGFR1-amplified lung cancer.
AstraZeneca started recently a clinical trial to study the safety and tolerability of AZD4547 at increasing doses in patients with squamous NSCLC. AZD4547 is an orally bioavailable inhibitor of FGFR with potential antineoplastic activity. AZD4547 binds to and inhibits FGFR, which may result in the inhibition of FGFR-related signal transduction pathways, and, so, the inhibition of tumour cell proliferation and tumour cell death. The inclusion criteria beside others were FGFR1 or FGFR2 amplification.
Within the Novartis study, patients with specific FGFR1 amplification in lung cancer are treated with the new potent, selective and orally bioavailable FGFR inhibitor NVP-BGJ398. Within this study 26 patients have been treated, including 10 patients with FGFR1-amplified breast and 3 patients with FGFR1-amplified squamous cell lung cancer. One lung cancer patient with an FGFR1/CEP8 ratio of 2.6 by FISH analysis responded to 100 mg of NVP-BGJ398 with a 33% reduction in target lesions by CT scan at 8 weeks, confirmed at 12 weeks, and a substantial SUV decrease on PET. These observations provide early evidence that inhibition of the FGFR pathway is effective in patients with FGFR dependent cancers (Wolf et al. 2012, Cancer Research). Since we are running the first-in-man trial of the FGFR inhibitor BGJ398 in FGFR1-amplified squamous-cell lung cancers at our centre already, we have convinced Novartis based on our results published in Nature Genetics to include patients with FGFR1-amplified SCLC into a subsequent expansion phase/ phase II clinical trial of that FGFR inhibitor. Thus, we strive to funnel all potentially clinically relevant results from our basic discovery efforts into immediate clinical application and in the personalized treatment of patients.
To systematically subject patients to a screening of therapeutically relevant genomic changes as a prerequisite for participation in the genetic stratified clinical trials the network for molecular diagnosis and therapy of lung cancer (CIO at Cologne – Bonn) was initiated at the University hospital of cologne, one of the world's top five screening networks and unique for Germany. Within this network patients are systematically subjected to a screening of therapeutically relevant genomic changes. Within the past 12 months, over 2000 patients were diagnosed within the network, representing over 70% of regional patients and nearly 5% of all German Lung. All these developments are the result of the funded project.
Potential Impact (WP6)
The impact of our findings is the development of “in vivo” tools to evaluate the mechanisms of action, sensitivity and resistance to molecular targeted drugs in lung cancer.
We have generated three independent clinically relevant mouse models in order to evaluate the implication of histological subtype, inflammation, or tumor heterogeneity on lung tumorigenesis. Firstly, we developed clinically relevant mouse models that resemble the most two most frequent human non-small cell lung cancer (NSCLC) histologies: adenocarcinoma (ADC) and squamous cell carcinoma (SCC). We used these models to evaluate the effect of antiangiogenic agents. Our results have important clinical implications in the design of future antiangiogenic therapeutic trials, and emphasize the need for precaution when considering the possibility of enrolling SCC patients in clinical trials evaluating antiangiogenic drugs.
Secondly, and in order to evaluate the role of chronic inflammation on lung cancer initiation, we developed a chemical-induced carcinogenesis mouse model. Our model is unique in both the type of inflammatory condition we use to promote cancer and the sequential analysis of malignant lesions that allowed us to study separately the effects of inflammation in preneoplastic and full neoplastic lesions. We developed a novel mouse model of lung cancer that allows for the study of the interaction of chronic inflammation with lung tumorigenesis. Our study provides a useful new multistep animal model for examining the specific role of chronic inflammation in promoting lung carcinogenesis and identifies for the first time molecular mechanisms by which silica-mediated inflammation creates a favourable microenvironment for tumor progression.
Finally, we have also developed a valuable tool for preclinical evaluation of novel therapeutic strategies in cancer: the so-called patient-derived xenografts models (PDX). These models, obtained by direct implants of tissue fragments in immunocompromised mice, have great potential in drug development studies because they faithfully reproduce the patient's original tumor for both immunohistochemical markers and genetic alterations as well as in terms of response to common therapeutics. They also maintain the original tumor heterogeneity, allowing studies of specific cellular subpopulations, including their modulation after drug treatment.
The three above mentioned models are currently being used for the study of the mechanisms of action, sensitivity and resistance to other molecular targeted drugs apart from the antiangiogenic therapies already described.
Moreover, we have developed models of acquired resistance in order to determine the molecular aspects related with this major clinical hurdle for the success of these treatments in cancer patients. We have also focused on the study of synthetic lethality in small-cell lung cancer (SCLC) cell lines. Our results raise the possibility of developing a therapeutic strategy for SCLC patients with specific alterations.
Potential Impact (WP7)
We have conducted the largest Phase III multicenter randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in completely resected Stage II-IIIA NSCLC. This is the largest study never initiated in the area of the molecular predictive markers in the field of completely resected NSCLC. Results from this study will have clear socio-economic and societal implications, as the improvement of treatment efficacy and the selection of patients for tailored therapies will ultimately modify treatment guidelines and increase patient survival.
In addition, the establishment of the Network Genomic Medicine, which now has developed to one of the largest international molecular screening networks for lung cancer, paradigmatically shows the implementation of molecular diagnostics and personalized treatment in broad clinical routine within a health care provider network. Of special interest is the rapid transformation of new discoveries (e.g. ROS1 fusions in adenocarcinoma) into clinical routine testing and treatment as well as the demonstration of a pronounced survival benefit by personalized treatment compared to chemotherapy in EGFR-mutated and ALK-translocated NSCLC. Based on these results, the AOK Rheinland/Hamburg, one of the largest German sickness funds reimburses multiplex genotyping of non operable lung cancer within NGM. To our knowledge, this is the first reimbursement model for comprehensive genotyping of lung cancer patients. We will extend this project to other sickness funds as well as to other cancer entities.
In summary, within the CURELUNG project we made different exciting discoveries with immediate clinical impact and we hope that they validate as relevant lung cancer drug targets and thereby improve the survival of lung cancer patients. The success of our project can be evaluated by the implementation of different newly discovered genomic markers into clinical routine as well as by different high impact publications in major internationally visible peer-reviewed biomedical journals. Moreover, our improved understanding of disease and the technological sophistication developed as part of CURELUNG are a further stimulus to the knowledge economy in each of the partner regions.
List of Websites:
Project website: www.curelung.eu
Project coordinator: Dr Manel Esteller, Tel. +34 93260 7140, email: mesteller@idibell.cat
Project manager: Mrs Verónica Padial, Tel. +34 93260 7143, email: vpadial@idibell.cat / pebc@idibell.cat / project@curelung.eu
Despite continuous advances, lung cancer (LC) is still the leading cause of cancer death worldwide. The extremely poor prognosis for LC patients is partly due to the lack of effective therapies. At present, most patients with pulmonary carcinomas are treated with chemotherapy. This essentially consists of classic cytotoxic drugs which only improve survival in small cohorts of cases. In spite of the rapidly growing understanding of the epigenetic and genetic profile of LC, such knowledge has contributed little to improving therapeutics.
Given the poor success of the traditional treatments but with the promise of novel and targeted drugs, CURELUNG aimed to concentrate on the discovery of novel therapeutic targets and on the study of the (epi)genetic background and biological pathways that will serve the purpose of tailoring patients for improving the efficacy of anticancer treatments.
Therefore, CURELUNG project emerged in January 2011 as the joint effort of leading European groups in exploiting its unique expertise with high-throughput technology and equipment, sharing experimental models and collections of human cancer samples from the different participants. After 42 months, the CURELUNG consortium has provided the cancer research community mainly with:
- The discovery of new tumor-specific gene alterations unveiling novel oncogenes and tumor suppressor genes.
- The establishment of a complete methylation profiling of lung cancer samples, together with the identification of a prognostic methylation signature in early stage lung cancer patients.
- New insights into the histomolecular classification of lung cancers. This will help in the distribution of lung cancer patients into groups with similar therapeutic management.
- The identification of novel target candidates for therapeutic interventions.
- The development of new lung cancer mouse models and the demonstration that patient-derived xenografts have great potential in drug development studies.
- The conduction of a phase III randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in non small cell lung cancer patients.
- The establishment of one of the world´s largest molecular screening network for lung cancer, i.e. the Network Genomic Medicine.
Altogether, although further research is needed, the input from CURELUNG will undoubtedly benefit the research community, allowing us to improve the lung cancer management and prognosis.
Project Context and Objectives:
Summary description of project context and objectives
Lung cancer (LC) is still the most lethal type of cancer worldwide. The extremely bad prognosis of LC patients is partly due to the late diagnosis of the disease, but also to the lack of effective therapies. Chemotherapy is applied in the majority of patients, but so far has only improved survival in very small cohorts. In spite of our rapidly growing understanding of the epigenetic and genetic profile of LC, the knowledge has not been much used to improve therapeutics.
In response to the call HEALTH2010-2.4.1-6: translational research on cancer with poor prognosis, the Curelung consortium has focused on the discovery and implementation of therapeutic approaches to increase LC patient survival. Given the poor success of the traditional treatments and the promise of the novel and targeted drugs, CURELUNG aimed to concentrate on the discovery of novel therapeutic targets and on the study of the (epi)genetic background and biological pathways that will serve the objective of tailoring patients for these upcoming anticancer treatments.
The CURELUNG project has been structured around three general and clear objectives:
OBJECTIVE 1: Discovery of novel (epi)genetically altered genes and characterisation of gene alteration patterns in LC
• ACTIVITY 1. Characterisation of candidate gene patterns and identification of (epi)genetically altered genes in lung cancer (LC) cell lines
Work packages involved: WP1 (Genetic and genomic profiling of LC cell lines), WP2 (Epigenetic profiling of LC cell lines)
The main objective of this activity was to assess the distribution of known (epi)genetic alterations in lung cancer and to identify novel alterations. Here we aimed to study individual genes or genome regions through a combination of traditional (automatic sequencing, FISH, methylation-specific PCR) and novel, high-throughput technologies.
The main goals achieved in this regard have been:
- The generation of a profile of genetic alterations at reported genes, determining their correlation with specific patterns of gene expression and with etiological and pathological characteristics of the lung tumors.
- The discovery of tumor-specific inactivation of genes such as MAX, MYST4 and PARD3 and activation of genes such as FGFR1, SETDB1, etc that constitute tumor suppressor genes and oncogenes, respectively.
- The methylation profiling of nearly 150 LC cell lines with the most up-to-date platform from Illumina, the Infinium 450k array. The analysis of 94 LC cell lines perfectly characterized in terms of response to the targeted drug gefitinib has led to de identification of 124 CpGs associated with response to the drug (reported in Deliverables 2.1 and 2.4).
- The methylation profiling of 444 LC biopsies and 25 non-tumor lung samples with the Infinium 450k array. Further analyses have led to the identification of 338 cancer related-differentially methylated CpGs and, importantly, we have validated 5 of them as biomarkers or prognostic value in stage I LC patients. This subgroup is only treated with surgery and potentially might benefit from an adjuvant therapy (reported in Deliverable 2.3).
• ACTIVITY 2. Verification and pattern of the molecular alterations in lung primary tumours.
Work packages involved: WP3 (Genetic-epigenetic alterations in LC primaries/biopsies), WP4 (Histopathology-immunostaining: in situ validation of (epigenetic) alterations).
The main objective of this activity was to verify the frequency and pattern of alterations at known and new oncogenes, tumour suppressors and others (from ACTIVITY 1) in a cohort of nearly 500 primary lung tumours.
The main goals achieved in this regard have been:
- Gene alterations identified in lung cancer cell-lines have been validated in lung cancer tumour samples, including those in the tumor suppressor PARD3, MYST4 and MAX.
- SME partners have provided sensitive assays for validation of genetic changes and optimised conditions, leading to implementation in clinical samples and publications.
- Different splicing events, miRNA and DNA methylation signatures for diagnosis have been validated
- The application of a histo-molecular classification which allowed to revise the WHO histopathological classification definition, affecting patient assignments in more predictive therapeutic groups.
OBJECTIVE 2: Discovery and preclinical validation of drug targets in LC
• ACTIVITY 3. Generation of predictive models for LC treatment
Work packages involved: WP5 (High-throughput cell line screening to identify therapeutically exploitable synthetic lethality), WP6 (Mouse and cancer cell models)
One of the main challenges in the field of LC is the design of successful strategies to prevent or inhibit the progress of the disease. Here we proposed to perform selective compound screening of the 50-80 lung cell lines in order to test directly the correlation between the (epi)genetic background of the cells and the response to specific therapies. Additionally, as new models for a more rational and optimized study of the drug efficacy are required, we planned to advance in the current models for predicting the efficacy of a new therapeutic tool (mouse models, xenografts).
The main goals achieved in this regard have been:
- 44 LC cell lines have been screened against a library of 267 small organic molecules of diverse chemical classes and targeting several oncogenic signalling pathways and molecular targets. As a result, this work unveiled that sensitivity to FGFR inhibition was enriched in FGFR1 amplified SCLC cell lines.
- A synthetic lethality of MYC-amplified SCLC cell lines with Aurora B kinase was found. Given the possibility that a particular subgroup of SCLC patients might benefit from this targeted therapeutic intervention, preclinical validation experiments with a pharmaceutical company are being planned (reported in Deliverable 5.1).
- Two lung tumor mouse models representing the two most prevalent histological subtypes: adenocarcinoma and squamous cell carcinoma have been developed (reported in Deliverable 6.1).
- We characterized the histopathological differentiation of patient derived lung cancer cells xenografted (PDX). We have studied the correlation between the original tumor and the corresponding PDX by means of: (i) metabolic parameters (SUV), (ii) tumor-stroma interactions and the frequency of tumor-initiating cells. In this regard, we have demonstrated that patient PDXs have great potential in drug development studies.
OBJECTIVE 3: Discovery and validation of (epi)genetic markers predicting responsiveness to specific novel inhibitory treatments (primary sensitivity or resistance to a given therapy)
• ACTIVITY 4. Target assessment in human/clinical trials
Work packages involved: WP1-4 (see titles above), WP7 (Selection of patients in clinical trials).
The work proposed in this objective aimed to involve the use of pre-treatment biopsies/tumours from LC patients in order to validate biomarkers discovered in the course of the previous activities that predict clinical response to a given specific agent.
The main goals achieved in this regard have been:
- The ideation and conduction of a Phase III multicenter randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in completely resected Stage II-IIIA NSCLC. The assessment of ERCC-1 and TS by qRT-PCR allowed the definition of 4 different genetic profiles and a tailored treatment based on pemetrexed, paclitaxel, cisplatin/gemcitabine or cisplatin (pemetrexed. Due to short follow-up, results will be available on a future time.
- In order to investigate the expression of the receptor tyrosine kinases in completely resected non-small cell lung cancer the gene expression of FGFR family of receptors was also investigated in the same case series. However, no significant results were obtained.
- One of the world´s largest molecular screening network for lung cancer, i.e. the Network Genomic Medicine (NGM), has been established. In 2013 the first large evaluation of NGM encompassing the analysis of about 5000 patients has been conducted. NGM helped us to efficiently recruit into personalized trials and to improve the survival of patients with EGFR mutations and ALK fusions among the collaborating NGM partners.
Project Results:
Work Package 1. Genetic and genomic profiling of Lung Cancer cell lines
The work package 1 aimed to identify novel genes, i.e oncogenes and tumour suppressor genes, altered in lung cancer (LC) using the latest genome wide sequencing and DNA copy number determination technologies. To achieve this goal we proposed testing LC cell lines, which are free of normal contaminant DNA, and verified the candidate genes later, in lung primary tumours specimens. In addition, within this work package we aimed to identify genetic and molecular determinants of the acquired resistance to novel targeted therapeutics. In the next pages we will describe all our discoveries that will certainly have an impact our current knowledge of LC biology and, possibly, to the LC clinical management.
Objective 1. Reported genetic alterations in cell lines
First, we selected 200 LC cell lines and study several of their characteristics. We found some differences in the distribution of gender, smoking habit and histopathologies, as compared to lung primary tumours. Particularly surprising, the squamous cell carcinoma (SCCs), which is a subtype of non-small cell lung cancer, is underrepresented. In addition, the non-smokers and the white ethnicity are overrepresented in LC cell lines. The reasons for the different distribution are unknown, but these will need to be taken into consideration if we want to extrapolate data from cancer cell lines to lung primaries. The gene alteration analysis in 160 of these cells confirmed that TP53 mutations are almost universal. The distribution of other major gene alterations in the different types of LC cell lines agrees with the previous general knowledge.
In addition, and within the context of a functional study on the candidate oncogene SOX4, we have observed that some of the genes that confer the gene expression signature of the small cell lung cancer (SCLC) are SOX4 targets. Our results show significant similarities between the SOX4 target genes and the SCLC gene expression profile which attest to the important role of SOX4 as a key element required for the development of SCLC (Castillo et al. 2012). This constitutes the one of the first publications which acknowledges the financial support of the CURELUNG.
Objective 2. Discovery and characterization of novel genetic alterations.
Next, we performed various genome wide analyses to assess for the presence of genetic alterations. Here, we must inform that, in addition to cancer cell lines, we used lung tumorgrafts. Lung tumorgrafts are lung primary tumours that have been implanted in the back of the mice and that, similarly to LC cell lines, have very little amount of normal contaminant cells. These models share the advantages of lung primary tumours and of the LC cell lines while are free of their respective disadvantages. We provided a list of genes that have been found in our study. Most of them have also been validated and studied, functionally.
The tumorgrafts that we have developed here have an important advantage: any possible targetable genetic alteration found in the tumour can be further studied and assayed in the animals. Although we do not report these types of analyses here, the CURELUNG project has been essential to allow for the development of such mice models. More importantly, these models may set the basis for the assay of novel drugs in lung tumours with a completely known genetic background.
Objective 3. Determination of genetic alterations involved in acquired resistance to targeted drugs.
Finally, in this work package we had proposed to work towards the comprehension of the mechanisms that account for the secondary (acquired) resistance to some of the targeted drugs used currently in the clinical settings. This work has been performed lung cancer cell lines. We have generated resistant cell lines from LC cells that were initially exquisitely sensitive to inhibitors of the MET and FGFR1 oncogenes and have studied the cells before and after becoming resistant. Our analysis allowed us to identify various changes that clearly differentiate both types of cells, including genetic and molecular shifts. Among the most interesting findings was the observation that the LC cells that have been made resistant in culture to the inhibitors of MET now have very high levels of the GAS2 growth factor and of its receptor, AXL. More interestingly, these cells had become very sensitive to other inhibitor, erlotinib, which is a molecule that inhibit the activity of another oncogene often activated in lung cancer, EGFR. Furthermore, the LC cell line that had been made resistance to the FGFR1 inhibitor now exhibits high levels of MET and activation of the AKT and mTOR pathway. These observations revealed that acquired resistance to a given inhibitor of a particular growth factor receptor triggers the activation of another growth factor receptor, to bypass the inhibitory effect of the drug.
Work Package 2. Epigenetic profiling of LC cell lines and biopsies
Nowadays, it is well known the high grade of complexity of the biology of cancer, particularly lung cancer (LC). In recent years, epigenetics has emerged as a new layer of information, becoming a key factor in order to complete the whole picture of cancer biology. Breakthroughs in technologies have enabled the possibility of genome-wide epigenetic research and revolutionized the concept of analyzing in an unbiased manner, not limited by previous knowledge. In this sense, the main objective of this WP was the characterization of the epigenetic modifications, mainly DNA methylation and microRNAs (miRNAs) expression, in LC cell lines (which are used as a model of “real” lung tumours), as well as to search new epigenetically regulated genes involved in resistance to specific targeted drugs. In addition, we have also performed the genome-wide methylation profiles of a great amount of LC primary biopsies, that have allowed identifying a potential DNA methylation signature associated with early-stage lung cancer prognosis.
Objectives 1 and 2. Characterize known tumour suppressor genes and identify novel cancer genes and miRNAs epigenetically altered in LC
More specifically, and regarding the epigenetic profile of LC cell lines, we have analyzed 126 cell lines together with 25 non-tumour lung cancer tissue samples. The Infinium 450k Human Methylation Array, which interrogates more than 450,000 CpGs (the main DNA positions where methylation takes place) and 21,000 genes has been used for this purpose. We have identified 17,576 differentially methylated CpGs, including 11,670 hypermethylated and 5,906 hypomethylatated ones in LC cell lines compared to non tumour controls. Hypermethylated events took place mainly in the promoters (regulatory regions) of genes involved in cell cycle and developmental processes. This is consistent with the theory that hypermethylation of the promoter regions of tumour-suppressor genes is a major event in the origin of many cancers. On the other hand, DNA hypomethylation events occur mainly in intergenic areas, where it leads to genomic instability.
MiRNAs are a large family of short non-coding RNAs that play regulatory functions and have been involved in many biological processes, including all the steps in cancer evolution. As the entire epigenetic machinery acts together to ensure the correct regulation of the gene expression, miRNAs may be affected by other epigenetic changes, such as DNA methylation. Thus, our next step was to focus in that specific feature from all the differentially methylated CpGs found in our previous analysis. We observed that, as previously noted, hypermethylated CpGs associated with miRNAs were located in promoters (which is concordant with tumour suppressor miRNA silencing) and hypomethylated ones are located mainly in low-density CpG regions, which could be a profile for onco-miRNAs.
Additionally, we determined the patterns of expression of relevant miRNAs across different lung cancer cell lines and immortalized human bronchial epithelial cells. We first investigated the expression of 24 miRNAs, as we have recently shown that, in plasma, this panel of miRNAs is able to detect lung cancer up to two years earlier than low-dose spiral computed tomography and also identifies tumour aggressiveness. The reciprocal ratios between expression ratios revealed a progressive trend for three miRNA ratios (miR 126/92a: 126/21; 28-3p/92a), recapitulating the deregulation observed in presence of clinical aggressive disease. Additionally we observed that other ratios between miRNAs belonging to signatures of high risk disease could stratify cellular models, highlighting the potential of this approach to identify specific miRNA patterns related to LC progression and offering insights on selection of cellular models to represent different stages during lung cancer progression.
In addition, we were highly interested in expanding our DNA methylation profiles beyond LC cell lines analyzing primary tumour samples from biopsies. We also used the Infinium 450k platform to deeply analyze the methylation profile of 444 tumour samples and 25 normal lung tissues (previously mentioned) (Sandoval J et al. 2013). Apart from generating a rich body of freely available information, we observed that the hierarchical clustering distinguished two groups of patients associated to significant differences in prognosis (estimated in relapse-free survival after surgery). This was also evident for the early-stage lung cancer cases, which undergo potentially curative surgical resection but still are at high risk of dying from recurrent disease, with a 5-year relapse rate of 35% to 50%. This is a particular important subgroup of patients, because although adjuvant platinum-based chemotherapy is beneficial in more advanced resected disease, in which most of the patients have a high risk of recurrence, it has failed to show a survival benefit for patients at early stages. One explanation for these negative data in the early stages could be the lack of biologic factors predicting their recurrence and the fact that, in the absence of useful biomarkers, all these cases are pooled, making it more difficult to draw meaningful clinical conclusions. For this reason, we looked for DNA methylation biomarkers that could distinguish between patients at low risk of relapse and those at high risk for whom adjuvant treatment could be prescribed. If successful, new treatment guidelines could be generated and more patients could benefit from an appropriate treatment.
Thus, we identified and selected those top genes that were more powerful in discriminating between the two groups of patients. As the 450k array is not currently available at hospitals and genome wide DNA methylation profiling cannot be carried out as a daily basis in the clinical setting, we investigated the role of these genes in an independent cohort of early-stage lung cancer patients by pyrosequencing, a more affordable large-scale approach. We ended up with a signature of 5 genes (HIST1H4F, PCDHGB6, NPBWR1, ALX1 and HOXA9) whose methylation can improve prognosis accuracy. Future studies will determine if this signature is also capable of improving treatment efficacy in lung cancer patients.
Objective 3. Characterize those epigenetic alterations involved in primary and acquired resistance to targeted drugs in LC.
Finally, one of our goals was the search of new epigenetically regulated genes involved in resistance to specific targeted drugs. This is relevant because, although one of the main advances in non small lung cancer treatment has been the development of targeted agents, all patients who initially benefit from these targeted therapies eventually develop resistance. Thus, in order to benefit a larger population of patients, it is crucial to develop a more detailed understanding of factors involved in the development of drug resistance and to identify additional therapeutic targets that may arise during the evolution of this process.
As DNA methylation is emerging as a potential driver influencing drug response, we decided to analyze the methylation profiles of 94 lung cancer cell lines perfectly characterized in terms of response to the EGFR targeted drug gefitinib, one of the first and more commonly used targeted agents in lung cancer. These analyses have allowed us to confirm the potential role of DNA methylation in influencing drug efficacy. Particularly, they have led to the identification of potential candidate genes conferring sensitivity or resistance to this particular agent. From those CpGs located in promoter, although some hypermethylated genes show correlation with better response, the great majority of hypermethylation events are related to drug resistance. As DNA methylation is potentially reversible, and DNA demethylating agents are currently available, these data open the possibility of using epigenetic drugs in order to sensitize or resensitize resistant tumours. Additionally, these epigenetic data can be used in the future to gain insight about the role of DNA methylation in the response to other targeted drugs.
Work Package 3. Genetic-Epigenetic alterations in LC primary/biopsies
Objectives 1 and 2. Validation of lung cancer associated alterations (in DNA sequence, DNA methylation, gene splicing or expression of miRNA or proteins) and correlating with etiological, histological and clinical variables. This has included potential implementation of discoveries by validation in clinical samples.
In order for scientific discoveries to make a significant impact in either our understanding of disease or its treatment, they must be validated by confirmation of their significance in a relevant independent study. This requires a co-ordinated effort by multiple clinical and laboratory scientists to identify the relevant samples for validation and to provide a significant number of samples to perform the test in a statistically valid manner. Many gene alterations in DNA sequence, DNA methylation, DNA translocation, gene splicing and gene expression have been validated and are described by other work packages. Here we give some specific examples where validation has strengthened the evidence that such alterations are significant in lung cancer, or has allowed us to determine their utility as biomarkers in clinical settings.
As described in Work Package 2, we have identified differences in DNA methylation (a modification of DNA that helps determine expression of proteins) that in a large series of lung cancers was apparently associated with those that had a poor outcome following surgery. In this discovery phase the DNA methylation changes were particularly associated with outcome in those cancers that were at an early histological stage (i.e. they were small and had not spread outside the lung). We therefore collected a new, independent set of these tumours and retested a representative selection of methylated DNA regions to validate our earlier findings. In doing so, we confirmed that DNA methylation of a set of five gene regions provided a good indication of early stage lung cancers that had a greater incidence of relapse.
Much of our understanding of lung cancer at the molecular level comes from the study of cells that grow in the laboratory after removal from tumours (tissue culture of cancer cell-lines). A large number of such cells have been produced around the world. Having catalogued the known molecular alterations in these, a number of particularly interesting gene alterations were identified (in Work Package 1), often because they appeared to be present mainly in specific types of cancer cell lines that share the same appearance. For example alterations in the gene PARD3 were found only in cell lines with a squamous histology (cancer cells having an elongated appearance) and MAX gene alterations were found in small cell cancer cells. As these DNA lesions were rare in cancers as a whole we performed a focused validation, looking mainly at collections of lung cancers with the same histological appearance. As well as being related to cancer relapse, or to specific groups of cancers, cancer-associated alterations may prove useful for diagnosis. Therefore, through technology validation, we have also demonstrated methods that allow those alterations to be assessed in clinical samples used for diagnosis, such as: formalin-fixed paraffin-embedded (FFPE) samples routinely used to examine histological features of tumours; bronchial lavage samples used in pre-surgical diagnosis of lung cancer; and plasma samples that could be used in early detection or disease monitoring.
MicroRNA (miRNA) molecules are short strands of genetic material that are involved in the regulation of gene expression (i.e. which proteins are made by different cells). Many have a functional role in lung cancer and in order to help identify these we have compared their levels in tumour tissue with that in normal lung tissue. In doing so, we identified 22 miRNA that had the largest and most consistent differences and validated this in an independent set of tumour and normal samples. We then combined multiple miRNA into a miRNA biomarker or signature, which was best able to discriminate between tumour and normal lung samples. This cancer-specific miRNA signature was shown in another validation to have a high accuracy in indicating if samples were derived from tumour material or not. All these studies were performed using samples frozen after being surgically removed from cancer patients, but many samples used clinically are both smaller (e.g. biopsies) and are preserved by being fixed in formalin and embedded in paraffin wax (FFPE). We therefore took a series of small samples from FFPE lung cancers and normal lung tissues and measure the miRNA signature. We were able to validate that the tumour tissues could be identified with a high sensitivity and specificity. Furthermore, this was true even though the tumour samples contained some normal cells (as is normal in cancer samples, where normal and cancer cells are frequently mixed together). We therefore believe that this miRNA signature potentially could be useful in helping clinicians to determine if small samples contain cancer cells, e.g. when examining biopsies form indeterminate nodules frequently discovered by CT screening studies.
Similarly, DNA methylation changes associated with cancer (such as those described above) can be useful in helping identify cancer patients from cytology samples (cell suspensions from fluid or disaggregated samples). For example, bronchial washing samples are a common diagnostic specimen, produced by washing part of the lung suspected of having a tumour and usually examined by cytology (microscopically examining the washing for abnormal cells). We have previously shown that a panel of four DNA methylation alterations found in lung cancer was better able to detect those with cancer than was the current practice of cytological examination. Combining these two techniques produced the highest accuracy. We have now validated additional DNA methylation changes that allow us to improve the accuracy of the DNA methylation biomarker panel and are expanding studies further to include DNA methylation changes discovered in Work Package 2. As these DNA methylation changes occur commonly across many cancers and it has been shown that cancer DNA is present as cell-free DNA (cfDNA) circulating in blood, we may also be able to detect lung cancer patients by detecting DNA methylation alterations in cfDNA from blood plasma. In order to validate DNA methylation biomarkers in plasma we have adapted measurement techniques to allow specific detection of DNA methylation at very low levels.
Objective 3. Deep sequencing forefront genomic platforms in subsets of lung cancer tumours
DNA sequencing studies, using deep sequencing technologies, that have contributed consideration of the lung cancer risk of individuals (determined from lifestyle, family history and medical questionnaire data combined into a lung cancer risk score) and been augmented by consideration of DNA methylation and miRNA in the same context. These studies are in addition to deep sequencing of lung cancer cell lines and primary lung cancers that are described in WP1.
Over recent years DNA sequencing technology has been developed that allows us to examine most of the genome (whole genome sequencing, WGS), those specific parts that code for proteins (exome sequencing) or smaller collections of regions with particular significance to the disease being studied (targeted sequencing). These are sometimes referred to as next generation sequencing (NGS) or deep sequencing; deep sequencing is so called because each base is read multiple times (hundreds or even thousands of times) which allows us to confidently identify differences in DNA sequence even when present as a minor fraction in the sample being analysed. This is important, because not only are cancer cells frequently mixed with normal cells in clinical samples, but different cancer cells can have different gene alterations (tumour heterogeneity).
We have used WGS, exome sequencing on a selected set of squamous cell carcinomas from people with different lung cancer risk and also demonstrated that targeted sequencing can be used to detect mutations in samples from FFPE material.
Lung Cancer Risk
One way forward for improved management and prognosis for lung cancer lies with early detection of this disease, but economics will dictate that screening for lung cancer will not be available for an entire population. Therefore, there is a major need to be able to identify those people at highest risk of developing lung cancer. To this end, lung cancer risk prediction models have been developed to estimate the probability of developing lung cancer over a period of time. The Liverpool Lung Project risk prediction model (LLP risk score) utilises risk factors such as smoking duration, previous diagnosis of pneumonia, prior diagnosis of malignant tumour, occupational exposure to asbestos and family history of lung cancer to estimate an individual’s probability of developing lung cancer within five years. The LLP risk model has been validated in three independent datasets and used to select eligible participants for the United Kingdom Lung Cancer Screening Trial (UKLS).
In one aspect of the DNA sequencing studies of the CURELUNG project, we have examined lung cancers from those people for whom LLP lung cancer risk was high and compared them to those who suffered from lung cancer despite having a low risk score. The aim was to both understand any biological difference between the tumours and potentially to provide molecular biomarkers to improve our lung cancer risk prediction. We also therefore examined differences in DNA methylation and miRNA and looked at the outcome of lung cancer in cases stratified by lung cancer risk score.
In order to validate our findings in an independent set of tumours, we determined a simplified risk score for lung cancer cases publically available datasets of molecular data (The Cancer Genome Atlas, TCGA).
From our studies we concluded that there were only very subtle differences in the gene alterations found in lung cancers that arise in those with very low cancer risk (<2% risk in the next five years) compared to those with higher cancer risk (>8% in the next 5 years). Low risk cases are usually younger patients who have smoked less and/or who have no family or personal history of cancer. It would therefore seem that when lung cancer arises in these individuals it is very similar at the molecular level to that in other cancer patients. This is valuable as it means diagnostic tests validated here should perform equally well in all lung cancers. However, it is interesting to note that despite their relatively young age and lower exposure to smoke-derived mutagens, low risk lung cancer patients also have the same poor outcome as other cancer patients; this is in keeping with the molecular similarities but suggests we do not yet fully understand the dynamics of lung cancer development. We will therefore continue to build on the studies performed as part of CURELUNG in order to gain a better understanding of the relationship between lung cancer risk factors and the molecular characteristics of lung cancer.
Work Package 4. Histopathological-immunostaning: in situ validations of (epi)genetic alterations
The main objective of this WP was the histomolecular classification of large cohorts of patients with lung tumours with histological criteria (histology and phenotype) (histomolecular features ) associated with specific (epi) genetic abnormalities and their prognostic and predictive value .
Objective 1. Histopathological classification of a large cohort of lung tumours obtained by immunostaining.
Histomolecular classification of lung tumours selected and included in this program by the partners using morpho-phenotypical subtypes, including subclassification of adenocarcinoma, squamous cell carcinoma, high grade neuroendocrine tumours (small cell lung carcinoma and large cell neuroendocrine carcinoma) and carcinoids , are derived from recent modifications of the last WHO classification recently revised on the basis of the international guideline recommendations we recently published (Travis WD et al. 2011).
On this basis we have performed histological review of about 1100 tumours from Partner WP5 (R. Thomas) and 1 (IDIBELL) and 4 (ULIV) and 1870 patients blocks from clinical trial patients registered in 2 arms of therapy , thus allowing prognostic and predictive factors( of chemotherapy success) to be evaluated for their influence on survival . We have used not only histological morphological class but differentiation and proliferation antigens TTF1, transcription factor and transcription lineage factor characteristic of adenocarcinoma, P63 transcription factor and its form P40 for squamous cell differentiation lineage, in association with CK5/6, P16, as well as protein markers on the pathway of oncogene driver (EGFR, mutant and wild type, IGFR, and ALK expression, ROS1 expression. The significant result is leading to the reclassification by review (histomolecular reclassification) of 16 % of lung carcinoma. There was a high correlation between cases such reclassified and the genomic- based classification. Histo- molecular reclassification allowed suppression of the class of large cell carcinoma to less than 1% of lung tumours as does genomic- based profiling. This is a strong foreground in the new concept of lung tumour classification which allowed us as WHO editors to thoroughly revise the histopathological classification definition and reassign the previous large cell carcinoma to Squamous or Adenocarcinoma according to histological phenotype. This new classification will affect patients assignments in more predictive therapeutic groups (including all targeted therapy).
Objectives 2 and 3. Histomolecular characterization of tumours with known (epy)genotype according to oncogene, tumour suppressor or copy number alterations and characterization of tumours at recurrence.
New gene fusion can be precisely identified by immunohistochemical markers to preselect and accelerate FISH testing and provide an FDA approved companion test for legitimating of crizotinib therapy. We thus reported for the first time the high sensitivity and specificity (98 %) of 2 specific antibodies directed toward ALK and ROS1 , to rapidly detect cases with one of the 2 new fusions / translocations involving ALK( EML4/ ALK) or ROS1 (CD74 and other partners) .We published the first evaluation of theses 2 antibodies on large series of adenocarcinoma tested for multiple genetic drivers (INCA Grenoble platform of tumour genetics, preventing pathologist to perform FIFH systematically in triple negative adenocarcinoma (neither EGFR , RAS or ALK mutations) and expedite the diagnosis of gene fusion relevant of crizotinib therapy. These 2 antibodies will soon be approved as FDA companion test allowing crizotinib (ALK and Met inhibitor) treatment, especially following the dissemination via the ALK atlas we participated to, disseminated at IASLC World Lung Cancer Conference, Sydney 2013.
We also provided in collaboration with partner 5 (R Thomas) in the identification of the most recent gene fusion described (CD74/NRG1) in japanese patients. We helped with a parallel caucasian french series to delineate the strict histomolecular contours of this new entity: all adenocarcinoma from non smoking women and subclassified as Invasive Mucinous Adenocarcinoma. This new histo-molecular entity will be included now in the next WHO classification (2015). It is highly predictive on simple histology of the presence of Ras mutation (75%) or CD74/NRG1 fusion (% undetermined in Caucasian patients) both in mutual exclusion, prone to be sensitive to Ras or Tyrosine kinase inhibition .
Neuroendocrine lung tumours have been delineated by histomolecular characterization in high correlation with their molecular profile. This is also a huge progress in the community of lung tumour oncology , in providing a genomics profile specific for each subtype of high grade Neuroendocrine tumours and low grade / intermediate grade carcinoid, showing for the first time a wide range of leading oncogenic events in theses tumours, opposing the high grade tumour small cell lung carcinoma to the low grade carcinoids. Identification in SCLC of five oncogene drivers suggests targets for future therapy, presently confined to cytotoxic chemotherapy with no change in the last 20 years . The discovery that carcinoids have no overlapping features with SCLC, implies that they cannot be considered as their precursors Carcinoids are driven as tumour proliferation uniquely by alteration of genes involved in chromatin modification, rather than oncogene mutation (Epigenetic oncogenesis). These are strong advances implying future targeted therapies and therapeutic challenges, that were not handled in the last 20 years. Indeed, the new genetic profile concept of lung neuroendocrine tumours is included in the next WHO classification 2015 (Pathology and genetics of Lung, Pleura, Thymic and Heart tumours).
Basaloid carcinoma has a dismal prognosis among other squamous cell carcinoma: we have under this Curelung program deciphered the molecular traits (gene expression RNA profiling a gene copy number alterations) which explain their poor prognosis and disastrous clinical behaviour. Accounting for 6% of lung tumour, this is a highly aggressive tumour with early recurrence and death, even at early stage I or II, surgically treatable. We have discovered their specific gene expression profile called “basaloid cluster” which identified short living tumour among all external series /databases containing squamous cell carcinoma of the literature (in silico) which tightly recapitulates cases with basaloid histo-molecular classification. This expression cluster is enriched in high cell cycle regulation, high expression of germ cell, embryonic and stem cell gene/pathways, low differentiation and chromatin regulatory genes (PCR) programs. We have published as primary results their pathognomonic marker SOX4 allowing fast and easy recognition of this dismal entity.
Epigenetic alterations promote tumour progression, recurrence and metastases. This was an obvious forefront result in at least 3 levels. A. the methylation profile established in collaboration of partner 1 which allowed a prognostic methylation profile across histological types and stage in lung cancer (J. Sandoval et al. 2013). B. the multiple chromatin modifiers (histone metyl- transferase, polycomb, SWI/SNF, etc), which drive carcinoid tumours, the histone acetylase gene mutations (CREBP/P300) which inactivate proper chromatin functions in SCLC (M. Peifer et al. 2012). C. The signature of 26 genes with ectopic illegitimate expression in lung cancer usually restricted to testis and placenta tissues, which are affected by strong hypomethylation of their CPG sites, in contrast with their methylation in normal testis. The expression of at least 3 of the 26 gene signature predicts short survival (early recurrence and metastases) and allows early detection of aggressive tumours, as well as proposition of a early epigenetic and vaccination therapeutic interventions (S.Rousseaux et al. 2013). Ectopic expression of tissue antigen restricted to certain normal tissues, is a form of identity crisis (deviation) common to highly aggressive tumours such as we demonstrated with an aberrant expression of Prolactin mRNA, which identifies highly malignant tumours including high grade neuroendocrine tumours among which SCLC and large cell neuroendocrine tumours: the prolactin mRNA presence was able to significantly affect prognosis across stage.
Lymphocytic infiltration has prognostic value in lung cancer clinical trials: this is a new discovery we made in analysing the 1576 patients paraffin blocks/sections to evaluate the level (0 or 1) of lymphocytic infiltration whatever phenotypes of theses tumour infiltrating lymphocytes . As a forefront results , this was the first significant factor across histomolecular classification subtypes which allowed stratification of patients for survival, cross -validated among world trials of cisplatin- adjuvant therapy in lung cancer ( LACE trials )(P=0. 009 for high infiltration and longer survival in overall and disease free survival.
Objective 4. Histomolecular characterization of tumours with resistance to therapy with known or new therapeutic agents.
Two patterns of adenocarcinoma are predictive survival benefit from cisplatin adjuvant chemotherapy: Using the largest worldwide patients cohorts randomized in clinical trials which we reviewed for histomolecular classification and survival correlations, we identified two architectural patterns of adenocarcinoma, shown to be of poor prognostic and discovered their strong predictive value. From 533 cases of centrally reviewed adenocarcinoma, we identified 2 of the 5 patterns (micropapillary and solid) conferring to patients receiving cisplatin based adjuvant chemotherapy significant survival benefit (P>001 in DFS and specific DFS) this was the first ever identified factor for stratification of future patients in clinical trials of cisplatin- based adjuvant chemotherapy.
The ERCC1 is not anymore considered as a stratification factor/ biomarker for Cisplatin Adjuvant therapy : we had previously reported that patients registered in cisplatin adjuvant therapy with low ERCC1 level evaluated by immunostainig (IH) with an ERCC1 antibody , derived significant survival benefit from cisplatin adjuvant chemotherapy in great contrast with patient entering the control surgical arm However in contrast with previous factors we could not validate and reproduce this results across world (Lace) clinical trials using the new antibody commercially provided This lead us to undergo further investigations in the frame of this Curelung program d to decipher and elucidate the reasons for this failure. We discover that ERCC1 enzyme displayed 8 different isoforms, variably distributed among normal tissues at tumours, but only one had efficient DNA repair in vitro. Unfortunately ERCC1 antibodies (6 different antibodies tested) were not specific for this isoform epitope thus depriving us from potential biomarkers for patient stratification among cytotoxic therapies. An array of DNA repair IH markers seems more rewarding than only one. In conclusion future qualification of IH markers will necessarily be functionally evaluated and validated on multiple trial series. New antibodies specific for functional isoforms will have to be produced .
Lung Tumors sampled at primary and metastatic site display the same oncogene driver number and profile, if not intended to treatment after surgery. We have provided primary results in comparing primary and metastatic tissue at s different sites of lung cancer .Although diversification in number and nature of passenger (non driver) mutations perceived as increase in genetic instability, we could not identify tumours which change (gain or loss) of oncogene driver ,susceptible to targeted therapy, at the step of metastatic process
Work Package 5. High-throughput cell line screening to identify therapeutically expoitable synthetic lethality
Cancer is a disease of the genome; genetic lesions (gene mutations, gene copy-number changes, structural genetic changes, etc.) lead to irreversible changes in the intracellular signal transduction pathways that the tumour cells become dependent upon. A new class of cancer therapeutics targeting specific signalling pathways activated by genetic lesions has shown clinical success. Due to the molecular heterogeneity of most solid tumours our strategy was to discover new genetic alterations that could act as novel targets for therapies and to define genetic markers that could determine the efficacy response or resistance to targeted therapies.
Objective 1. Performing high-throughput cell line screening in deeply LC cell lines, enriched for certain cancer genotypes.
Within the Curelung project, we have performed a genomic analysis of 60 SCLC cell lines (Sos et al. 2012). We have identified regions of recurrent copy number alterations and compared the pattern of alterations in the cell lines to our complementary sequencing study of primary SCLC tumours (Peifer et al. 2012). Remarkably, SCLC cell lines and primary tumours are highly similar, thus validating the use of this SCLC cell line panel for combined genomic and pharmacological perturbation experiments. Our cell line collection captures hallmark events of SCLC such as recurrent deletions of RB1 and PTEN, but also amplification of genes such as FGFR1. Additionally we identified recurrent and focal amplification of MYCL1, MYCN and MYC. Besides SCLC we systematically studied the molecular alterations of pulmonary carcinoids by copy number analysis, genome/exome and transcriptome sequencing. We observed frequent mutations in chromatin-remodelling genes. In contrast to SCLC and LCNEC (Large cell neuroendocrine lung tumours) TP53 and RB1 mutations are rare events, suggesting that pulmonary carcinoids are not early progenitors but arise through independent cellular mechanisms. These data also suggest that inactivation of chromatin-remodelling genes is sufficient to drive transformation in pulmonary carcinoids.
Objective 2. Linking tumour genetics to individual compound sensitivity in order to identify synthetic interactions of individual hard to target genomic aberrations.
The objective of our study was to use genomic data of different cell lines generated within the Curelung project and to link it with the activity of targeted drugs. This is an important step towards the identification of clinically active drugs that can be used for the treatment of genetically selected cancer patients. We selected LC cell lines based on the presence or absence of individual candidate aberrations. In total forty-four cell lines were screened against a library of 267 small organic molecules of diverse chemical classes and targeting several oncogenic signalling pathways and molecular targets.
One of the findings in this screen was that sensitivity to FGFR inhibition was enriched in FGFR1-amplified SCLC cell lines. However, one of the cell lines was only partly sensitive to FGFR inhibition, possibly due to loss of PTEN protein expression. In our complementary study focusing on the genomics analysis of small cell lung cancer specimen we confirmed FGFR1 amplification to be present in 6% of SCLC cases. These data strongly indicate that FGFR1 might represent an attractive therapeutic target in these tumors. However, we could demonstrate that the 8p12 locus containing the FGFR1 tyrosine kinase gene, shows a genomic heterogeneity and we found that MYC was co-expressed in 40% of FGFR1-amplified tumors. Tumor cells co-expressing MYC were more sensitive to FGFR inhibition, suggesting that patients with FGFR1-amplified and MYC-overexpressing tumors may benefit from FGFR inhibitor therapy, which may have implications for patient selection for treatment with FGFR inhibitors (Malchers et al. 2014). Since we are running a first‐in‐man trial of a novel FGFR inhibitor in FGFR1- amplified squamous cell lung cancers at our center already, we have convinced a pharmaceutical company to include patients with FGFR1‐amplified SCLC into a subsequent expansion phase/ phase II clinical trial of that FGFR inhibitor. Thus, we strive to funnel all potentially clinically relevant results from our basic discovery efforts into immediate clinical application.
As the most important finding of our cell line screening we found synthetic lethality of MYC-amplified SCLC cell lines with Aurora B kinase. Contrasting with findings in neuroblastoma (Otto et al. 2009), where Aurora A inhibition selectively killed MYCN-amplified neuroblastoma cells by depletion of cytoplasmic levels of MYCN protein, MYC levels were not depleted in SCLC cells and we found a predominance of synthetic lethal addiction to Aurora B as opposed to Aurora A. Given the possibility that SCLC patients with MYC-amplified might benefit from targeted therapeutic intervention by exploiting the synthetic lethal interaction with Aurora B, we are currently discussing with a pharmaceutical company to pursue further preclinical validation experiments. The company in question has a compound that exhibits profound selectivity and potency against Aurora B kinase.
In a next step we tested the relevance of biomarkers identified within Curelung to lung cancer carcinogenesis, implication on drug sensitivity or acquired resistance to drugs. Therefore we generated different modified lung cancer cell lines according to alterations of predictive value.
1. FGFR1: We have conducted transcriptome sequencing in FGFR1-amplified lung cancers with the goal of identifying the relevant FGFR1 splice variants and in order to discover possible modifiers of responsiveness to FGFR inhibition. These efforts resulted in the identification of oncogenic mechanisms of amplified FGFR1. We showed marked heterogeneity of the 8p12 amplification event in squamous cell lung cancer, which results from broad genomic rearrangements. Furthermore we observed high expressions levels of MYC associated with FGFR dependency across different cell line models. We have generated several cell lines expressing stably different isoforms of FGFR1 with and without MYC. Additional genotypes were generated to track precisely the oncogenic signalling downstream of amplified FGFR1. Supporting an oncogenic role for amplified FGFR1, we showed that the mesenchymal splice variants are transforming in vitro and in vivo. This effect was strongly enhanced by co-expression of MYC. The impact of high-level expression of MYC in mediating FGFR dependency was further strengthened by clinical observation of two patients with FGFR1 amplification and high MYC-expressing squamous cell lung cancer who responded to FGFR inhibition.
2. CREBBP: We have characterized several SCLC cell line genetically. In particular, we have identified lines with and without mutations in CREBBP or EP300. We have used these cell lines with the goal of functional characterization of the molecular effects of loss of these histone acetyl transferases. Of particular value in this context is a cell line lacking CREBBP mutations, in which we stably silenced this gene. Remarkably, despite the expected dramatic impact of silencing such global transcriptional regulator, we observed that the cells engineered to lack CREBBP expression proliferated more rapidly than the controls. Tumors with mutations and hemizygous deletions in CREBBP and EP300 did not exhibit a significantly different pattern of gene expression compared to wild-type tumours, suggesting that global changes in gene expression are not the predominant mechanism by which loss of HAT activity contributes to SCLC pathogenesis.
3. AURORA B: MYC is known to be involved in the pathogenesis of diverse cancer types. To test the relevance of MYC amplification in SCLC, we silenced expression of MYC and we functionally characterized the dependency of MYC-amplified SCLC tumours on Aurora B. We find that MYC-amplified tumours depend on the kinase activity of Aurora B for their survival. Our data provide a rationale for the testing of Aurora kinase inhibitors in genetically defined SCLC patient groups.
4. NRG1-CD74: By transcriptome sequencing of 25 lung adenocarcinoma of never smokers we have identified a novel CD74-NRG1 gene fusion, which resulted from a somatic genomic rearrangement. To study the impact of this novel gene fusion we transduced different cell lines with retroviruses encoding CD74-NRG1 and performed further experiments (NIH-3T2, H2052, H322 and H1568 lung cancer cell lines). Mechanistically, CD74-NRG1 leads to extracellular expression of the EGF-like domain of NRG1 III-3, thereby providing the ligand for ERBB2-ERBB3 receptor complexes. Ectopic expression of CD74-NRG1 in lung cancer cell lines expressing ERBB2 and ERBB3 activated ERBB3 and the PI3K-AKT pathway, and led to increased colony formation in soft agar. CD74-NRG1 fusions may represent a therapeutic opportunity for invasive mucinous lung adenocarcinomas, which frequently present with multifocal and unresectable disease, and for which no effective treatment exists.
Work Package 6. Mouse and cancer cell models
Objective 1. Testing incidence of biomarkers on lung cancer animal models (chemically induced models, tumorgrafts and genetically modified mice).
The main objective of this WP is to generate and use mouse and cancer cell models in order to test the relevance of biomarkers identified in the present project to lung cancer carcinogenesis, implication on drug sensitivity or acquired resistance to drugs. We first generated three independent clinically relevant mouse models in order to evaluate the implication of histological subtype, inflammation, or tumor heterogeneity on lung tumorigenesis:
1. Lung tumor murine models of pure histological subtype: urethane-induced ADC and NTCU-induced SCC. In these two models, we evaluated by immunohistochemistry the expression of proteins related with the VEGF pathway (VEGF, VEGFR2 and pVEGFR2) in both the urethane-induced ADC and NTCU-induced SCC models.
2. Chemically induced lung cancer mouse model in the presence of silicotic chronic inflammation and the tobacco smoke carcinogen NDMA.
3. Patient-derived xenografts models (PDX).
Objective 2. To test the contribution to response of resistance of a selected therapeutic drug in animal models
The efficacy of some of the more relevant anti-cancer therapies in lung cancer appears to be closely associated with the histological subtype of the tumor. Therefore, we have developed clinically relevant mouse models that resemble the most two most frequent human non-small cell lung cancer (NSCLC) histologies: adenocarcinoma (ADC) and squamous cell carcinoma (SCC). We have used these models to evaluate the effect of antiangiogenic agents (sunitinib and VEGFR2-blocking antibody DC101) on tumor progression in ADC and SCC. We have demonstrated the contrasting responses of NSCLC to antiangiogenic therapies depending on histology. We have shown that two anti-VEGFR2 therapies elicited tumor stabilization in ADC tumors. In contrast, VEGFR2 blockade in SCC caused hyperproliferation of tumor cells and increased the expression of stem cell markers, independently of intratumoral hypoxia. A major new finding of this project is that antiangiogenic therapies induce contrasting responses in NSCLC depending on the histological subtype (Larrayoz et al, 2014, Embo Mol Med). This result has important clinical implications in the design of future antiangiogenic therapeutic trials, and emphasizes the need for precaution when considering the possibility of enrolling SCC patients in clinical trials evaluating antiangiogenic drugs.
Another important aspect is the association between inflammation and lung tumor development. In order to evaluate the role of chronic inflammation on lung cancer initiation and progression, we developed a chemical-induced carcinogenesis mouse model that comprised the administration of a low tumorigenic dose of NDMA, a carcinogen present in tobacco smoke, in the presence of chronic silica induced lung inflammation. Our model is unique in both the type of inflammatory condition we use to promote cancer and the sequential analysis of malignant lesions that allowed us to study separately the effects of inflammation in preneoplastic and full neoplastic lesions. Our findings suggest that chronic inflammation generates a favourable microenvironment for the generation of preneoplastic lesions that potentially contributes to the progression of adenomas to adenocarcinomas. In summary, in this study, we developed a novel mouse model of lung cancer that allows for the study of the interaction of chronic inflammation with lung tumorigenesis. Our study provides a useful new multistep animal model for examining the specific role of chronic inflammation in promoting lung carcinogenesis and identifies for the first time molecular mechanisms by which silica-mediated inflammation creates a favourable microenvironment for tumor progression. Elucidation of the interplay between inflammation and lung carcinogenesis will help in the development of preventive strategies to slow down or avoid the generation of premalignant lesions and the conversion to malignant tumors. The three above mentioned animal models are currently being used for the study of the mechanisms of action, sensitivity and resistance to other molecular targeted drugs apart from the antiangiogenic therapies already described.
Objective 3. To evaluate the effect of selected molecular changes on the phenotype of tumour cell lines.
From the mice models developed in objectives 1 and 2, tumors we also established two new lung adenocarcinoma cell lines which can be successfully transplanted into balb-c syngeneic mice either orthotopically or as metastatic models. This new syngeneic mouse model for lung cancer and metastasis in Balb/c mice present similarities in genetic alterations with the human disease. Hallmarks for this novel model are cancer stem cell properties, mutations in KRAS (G12D) and WWOX deletion, which could faithfully recapitulate human lung carcinogenesis. (Bleau et al. IJC 2014).
We have also developed a valuable tool for preclinical evaluation of novel therapeutic strategies in cancer: the so-called patient-derived xenografts models (PDX). These models, obtained by direct implants of tissue fragments in immunocompromised mice, have great potential in drug development studies because they faithfully reproduce the patient's original tumor for both immunohistochemical markers and genetic alterations as well as in terms of response to common therapeutics. They also maintain the original tumor heterogeneity, allowing studies of specific cellular subpopulations, including their modulation after drug treatment. To evaluate the performance of this model for pharmacological testing we have also started treatment of PDX with standard lung cancer chemotherapeutics starting with cisplatin and antiangiogenic agents.
In the context of this project, PARD3 gene was defined as a potential tumour suppressor gene in squamous cell carcinoma of the lung. We genetically reconstituted the PARD3 gene in H157 cell line in order to evaluate the role of this gene in lung cancer. We concluded that restoration of wild type PAR3 reduces tumor invasiveness and prevents formation of metastasis “in vivo”.
One of the major focuses of this project was the identification of novel genomic alterations in small cell lung cancer and on discovery of novel synthetic lethal interactions by performing chemical vulnerability screens. Part of this effort was a thorough functional validation program in order to test the hypothesis that these alterations play a functional role in the pathogenesis in SCLC. In the context of WP5, we have generated many genetically manipulated since the launch of the project. We have evaluated the effect of the expression of some of the biomarkers described in WP1-3 in these cell lines. Our results demonstrate the existence of a synthetic lethal type of interaction between MAX and BRG1 in small cell lung cancer cells, and raise the possibility of developing a therapeutic strategy for patients with MAX-deficient tumors (Romero et al. 2014 Cancer Discovery).
By screening relevant compounds across SCLC cell lines we showed that Aurora kinase inhibitors are effective in SCLC cell lines bearing MYC amplification, which occur in 3-7 % of SCLC patients. We found the PLK1 inhibitor BI2536, the ROCK1 inhibitor GSK269962A and the pan-Aurora inhibitor VX680 to be specifically active in these cells. Our results show that Aurora dependency in SCLC primarily involved Aurora B, required its kinase activity, and was independent of depletion of cytoplasmic levels of MYC (Sos et al. 2012 PNAS).
Objective 4. To test the LC cells that have acquired resistance to drugs for genetic and epigenetic alterations
The development of acquired resistance to tyrosine kinase inhibitors (TKIs) is a major clinical hurdle for the success of these treatments in cancer patients. Here, we aimed to elucidate novel determinants of the acquired resistance to the MET and FGFR1 inhibitors in lung cancer. To this purpose, we first selected lung cancer cell lines according to the genetic alterations related with the TKIs. Subsequently, we generated multiple clones that have acquired resistance to treatment with TKIs and performed exome and transcriptome sequencing of the resistant clones.
Work Package 7. Selection of patients in clinical trials
Objective 1. Selection of LC patients undergoing clinical trials for targeted therapies.
In agreement with the objective 7.1 the WP 7 contributed to the retrospective collection of tissue samples from patients with stage I non-small cell lung cancer completely resected to generated data about a DNA methylation signature. DNA was extracted from tumour tissue of more than 100 samples and sent to IDI for investigation. The contribution to this task is indicated in the main related publication (Sandoval J et al. 2013). In order to investigate the expression of the receptor tyrosine kinases in completely resected non-small cell lung cancer the gene expression of FGFR family of receptor was also investigated in the same case series without obtaining significant results.
Objective 2. Establish and validate novel predictive models based on gene alterations to help clinicians in the therapeutic decision making in LC.
In agreement with the Objective 7.2 the main task of the WP 7 was the ideation and conduction of a Phase III multicenter randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in completely resected Stage II-IIIA NSCLC (EudraCT Number: 2008-001764-36), already started when the CURELUNG project was financed. This is the largest study never initiated in the area of the molecular predictive markers in the field of completely resected NSCLC. The study was conducted in approximately 22 investigational sites: 10 sites in Italy, 10 sites in Germany and 2 sites in Poland. The primary study objective was to evaluate overall survival in patients with completely resected Stage II-IIIA NSCLC treated with tailored adjuvant chemotherapy driven by the assessment of the pharmacogenomic profile of the resected primary tumour as investigated by quantitative real time polymerase chain reaction (qRT-PCR) of paraffin embedded specimens or with standard cisplatin-based adjuvant chemotherapy selected by the investigator (cisplatin/vinorelbine, cisplatin/docetaxel or cisplatin/gemcitabine). The secondary endpoints of the study included the evaluation recurrence-free survival with tailored adjuvant chemotherapy driven by the assessment of the pharmacogenomic profile of the resected primary tumour as investigated by qRT-PCR of paraffin embedded specimens or with standard adjuvant chemotherapy selected by the investigator, to assess the therapeutic compliance in the different arms, to assess toxicity with tailored adjuvant chemotherapy driven by the assessment of the pharmacogenomic profile of the resected primary tumour as investigated by qRT-PCR of paraffin embedded specimens or with standard adjuvant chemotherapy selected by the investigator and finally to comparatively assess the results of the quantitative assessment of messenger ribonucleic acid for Excision Repair Cross-Complementing 1 (ERCC1) and thymidilate synthase (TS) with the semiquantitative assessment of ERCC1 and TS protein assessment as investigated through immunohistochemistry.
The assessment of ERCC-1 and TS by qRT-PCR allowed the definition of 4 different genetic profiles. Within 30 to 45 days post-surgery, patients in each genetic profile was randomized to receive either a standard adjuvant chemotherapy (control arm) selected by the investigator (cisplatin/vinorelbine, cisplatin/docetaxel or cisplatin/gemcitabine) or an experimental treatment (tailored chemotherapy arm).
The experimental treatment arms are:
Genetic profile ERCC1 and TS assessment Experimental treatment
_____________________________________________________________________
1 High ERCC1 and high TS 4 cycles of single agent paclitaxel
2 High ERCC1 and low TS 4 cycles of single agent pemetrexed
3 Low ERCC1 and high TS 4 cycles of cisplatin/gemcitabine
4 Low ERCC1 and low TS 4 cycles of cisplatin/pemetrexed
The final statistical analysis at the end of the study will group together all the controls in one group (control group) and all tailored chemotherapies in another group (experimental group). The primary efficacy analysis will be performed both on all randomized patients (Intent-to-Treat) and on all evaluable patients. Efficacy analysis will be performed on an intent-to-treat basis. Survival curves will be estimated using the Kaplan- Meier method and compared with the log-rank test. The Cox proportional hazard model will be used for estimating Hazard Ratios after adjusting for relevant variables. Demographics and baseline characteristics will be summarized. Quantitative data will be summarized by arithmetic mean, standard deviation, median, minimum and maximum. Frequency tables will be provided for qualitative data. Adverse event and laboratory data, including laboratory toxicities have been summarized by worst National Cancer Institute Common Terminology Criteria Version 3.0 grade and listed for patients in the safety population.
The expected total number of patients was 684, therefore, it was planned to recruit up to 700 patients (350 per arm) in order to take possible dropouts into account. Due to 10% failures to amplify the RNA from tissues we further increased the sample size of 70 patients and now in July 2014 the study is completed. The study cannot be analyzed because of the low number of events (deaths) due to the limited amount of follow up time. Before the end of the year a publication focusing on distribution of ERCC1 and TS in the four identified genetic profile and correlation studies between biomarker expression and the main clinical-pathological characteristics will be presented in a major meeting and published in a peer reviewed journal. Correlation studies between gene expression and immunohistochemistry are also planned. These analyses are feasible only now to avoid any jeopardization of the study.
For the above reason deliverable 7.2 will be reached definitively at the time of the final analysis of the study, but preliminary information also included in deliverable will be reached at the end of 2015.
Objective 3. Identify/select patients with acquired resistance to a given treatment.
To fit with the objective 7.3 the cell line studies to understand the mechanisms of resistance to cisplatin and potential mechanisms of cisplatin sensitization have been performed using celestrol as potential sensitizer to cisplatin (publication pending).
Additionally, in a close interdisciplinary collaboration between the Lung Cancer Group Cologne (www.lungcancergroup.de) (Chair: Jürgen Wolf) and the Dept. of Pathology of the University Hospital of Cologne (UHC, Chair: Reinhard Büttner) and the Institute for Translational Genomics of the University of Cologne (chair: Roman Thomas) the Network Genomic Medicine (NGM, www.ngml.de) was established supported by funding of WP 5. Within NGM comprehensive genotyping of all known driver mutations in paraffin-embedded biopsy specimens obtained by routine diagnostic procedures was offered free of charge for all interested clinical collaboration partners in the catchment area of our cancer center, the Center for Integrated Oncology (www.cio-koeln-bonn.de). These included large lung cancer centers, smaller regional hospitals and office-based oncologists. In the meantime, more than 70 centers participate in this activity. In parallel, a clinical trial program was established focussing on early proof of concept (PoC) clinical trials to offer each patient with a detected driver mutation a personalized treatment approach (see lungcancergroup.de/clinicaltrials). In addition, a close collaboration was started with the clinical cancer registry of our cancer center as well as with the Epidemiological Cancer Registry of the state North-Rhine Westfalia. The number of collaboration partners increased continuously over time and in parallel the number of tissue samples analysed. The establishment of this molecular screening network enabled us to recruit patients with rare genetic subgroups efficiently in clinical trials in an internationally competitive level (Shaw et al. 2014). Also, within this network we were able to rapidly transfer new results into clinical practice and, for instance, identified and successfully treated the first European patient with a ROS1 fusion (Bos et al, Lung Cancer 2013) followed by the preparation and recent initiation of a European phase II trial led by the Lung Cancer Group Cologne for the evaluation of crizotinib in ROS1 positive lung cancer, which was initiated in May, 2014. In 2013 we conducted and published our first evaluation within NGM (Clinical Lung Cancer Project and Network Genomic Medicine, 2013). Here we presented the analysis of about 5000 lung cancer patients genotyped within NGM, reported the distribution of driver mutations, the correlation with clinicopathological features, the recruitment in clinical trials, and, for the first time, outside clinical trials, a significant overall survival benefit of EGFR mutated patients treated with kinase inhibitors vs. EGFR mutated patients treated with chemotherapy only (32.5 vs. 9.5 months) and of ALK-translocated patients treated with crizotinib vs. ALK-translocated patients treated with chemotherapy only (22 vs. 11 months).
Bibliography (alphabetical order)
- Bos M, Gardizi M, Schildhaus HU, Heukamp LC, Geist T, Kaminsky B, Zander T, Nogova L, Scheffler M, Dietlein M, Kobe C, Holstein A, Maintz D, Büttner R, Wolf J. Complete metabolic response in a patient with repeatedly relapsed non-small cell lung cancer harboring ROS1 gene rearrangement after treatment with crizotinib. Lung Cancer. 2013 Jul;81(1):142-3.
- Castillo SD, Matheu A, Mariani N, Carretero J, Lopez-Rios F, Lovell-Badge R, Sanchez-Cespedes M. Novel transcriptional targets of the SRY-HMG box transcription factor SOX4 link its expression to the development of small cell lung cancer. Cancer Res. 2012 Jan 1;72(1):176-86.
- Clinical Lung Cancer Genome Project (CLCGP); Network Genomic Medicine (NGM). A genomics-based classification of human lung tumors. Sci Transl Med. 2013 Oct 30;5(209):209ra153.
- Larrayoz M, Pio R, Pajares MJ, Zudaire I, Ajona D, Casanovas O, Montuenga LM, Agorreta J. Contrasting responses of non-small cell lung cancer to antiangiogenic therapies depend on histological subtype. EMBO Mol Med. 2014 Apr;6(4):539-50.
- Malchers F1, Dietlein F, Schöttle J, Lu X, Nogova L, Albus K, Fernandez-Cuesta L, Heuckmann JM, Gautschi O, Diebold J, Plenker D, Gardizi M, Scheffler M, Bos M, Seidel D, Leenders F, Richters A, Peifer M, Florin A, Mainkar PS, Karre N, Chandrasekhar S, George J, Silling S, Rauh D, Zander T, Ullrich RT, Reinhardt HC, Ringeisen F, Büttner R, Heukamp LC, Wolf J, Thomas RK. Cell-autonomous and non-cell-autonomous mechanisms of transformation by amplified FGFR1 in lung cancer. Cancer Discov. 2014 Feb;4(2):246-57.
- Otto T, Horn S, Brockmann M, Eilers U, Schüttrumpf L, Popov N, Kenney AM, Schulte JH, Beijersbergen R, Christiansen H, Berwanger B, Eilers M. Stabilization of N-Myc is a critical function of Aurora A in human neuroblastoma. Cancer Cell. 2009 Jan 6;15(1):67-78.
- Peifer M1, Fernández-Cuesta L, Sos ML, George J, Seidel D, Kasper LH, Plenker D, Leenders F, Sun R, Zander T, Menon R, Koker M, Dahmen I, Müller C, Di Cerbo V, Schildhaus HU, Altmüller J, Baessmann I, Becker C, de Wilde B, Vandesompele J, Böhm D, Ansén S, Gabler F, Wilkening I, Heynck S, Heuckmann JM, Lu X, Carter SL, Cibulskis K, Banerji S, Getz G, Park KS, Rauh D, Grütter C, Fischer M, Pasqualucci L, Wright G, Wainer Z, Russell P, Petersen I, Chen Y, Stoelben E, Ludwig C, Schnabel P, Hoffmann H, Muley T, Brockmann M, Engel-Riedel W, Muscarella LA, Fazio VM, Groen H, Timens W, Sietsma H, Thunnissen E, Smit E, Heideman DA, Snijders PJ, Cappuzzo F, Ligorio C, Damiani S, Field J, Solberg S, Brustugun OT, Lund-Iversen M, Sänger J, Clement JH, Soltermann A, Moch H, Weder W, Solomon B, Soria JC, Validire P, Besse B, Brambilla E, Brambilla C, Lantuejoul S, Lorimier P, Schneider PM, Hallek M, Pao W, Meyerson M, Sage J, Shendure J, Schneider R, Büttner R, Wolf J, Nürnberg P, Perner S, Heukamp LC, Brindle PK, Haas S, Thomas RK. Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat Genet. 2012 Oct;44(10):1104-10.
- Rousseaux S, Debernardi A, Jacquiau B, Vitte AL, Vesin A, Nagy-Mignotte H, Moro-Sibilot D, Brichon PY, Lantuejoul S, Hainaut P, Laffaire J, de Reyniès A, Beer DG, Timsit JF, Brambilla C, Brambilla E, Khochbin S. Ectopic activation of germline and placental genes identifies aggressive metastasis-prone lung cancers. Sci Transl Med. 2013 May 22;5(186):186ra66.
- Sandoval J, Mendez-Gonzalez J, Nadal E, Chen G, Carmona FJ, Sayols S, Moran S, Heyn H, Vizoso M, Gomez A, Sanchez-Cespedes M, Assenov Y, Müller F, Bock C, Taron M, Mora J, Muscarella LA, Liloglou T, Davies M, Pollan M, Pajares MJ, Torre W, Montuenga LM, Brambilla E, Field JK, Roz L, Lo Iacono M, Scagliotti GV, Rosell R, Beer DG, Esteller M. A prognostic DNA methylation signature for stage I non-small-cell lung cancer. J Clin Oncol. 2013 Nov 10;31(32):4140-7..
-Shaw AT, Kim DW, Mehra R, Tan DS, Felip E, Chow LQ, Camidge DR, Vansteenkiste J, Sharma S, De Pas T, Riely GJ, Solomon BJ, Wolf J, Thomas M, Schuler M, Liu G, Santoro A, Lau YY, Goldwasser M, Boral AL, Engelman JA. Ceritinib in ALK-rearranged non-small-cell lung cancer. N Engl J Med. 2014 Mar 27;370(13):1189-97.
- Sos ML1, Dietlein F, Peifer M, Schöttle J, Balke-Want H, Müller C, Koker M, Richters A, Heynck S, Malchers F, Heuckmann JM, Seidel D, Eyers PA, Ullrich RT, Antonchick AP, Vintonyak VV, Schneider PM, Ninomiya T, Waldmann H, Büttner R, Rauh D, Heukamp LC, Thomas RK. A framework for identification of actionable cancer genome dependencies in small cell lung cancer. Proc Natl Acad Sci U S A. 2012 Oct 16;109(42):17034-9.
- Travis WD1, Brambilla E, Noguchi M, Nicholson AG, Geisinger KR, Yatabe Y, Beer DG, Powell CA, Riely GJ, Van Schil PE, Garg K, Austin JH, Asamura H, Rusch VW, Hirsch FR, Scagliotti G, Mitsudomi T, Huber RM, Ishikawa Y, Jett J, Sanchez-Cespedes M, Sculier JP, Takahashi T, Tsuboi M, Vansteenkiste J, Wistuba I, Yang PC, Aberle D, Brambilla C, Flieder D, Franklin W, Gazdar A, Gould M, Hasleton P, Henderson D, Johnson B, Johnson D, Kerr K, Kuriyama K, Lee JS, Miller VA, Petersen I, Roggli V, Rosell R, Saijo N, Thunnissen E, Tsao M, Yankelewitz D. International association for the study of lung cancer/american thoracic society/european respiratory society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol. 2011 Feb;6(2):244-85.
Potential Impact:
The main dissemination activities and exploitation of results of the CURELUNG project are mainly scientific publications, conferences and workshops. There have been already published around X publications in scientific journals in the frame of CURELUNG during these three years and a half. Their results have also been exposed various conferences from all the participants, as it is shown in the attached list. With the aim to disseminate the results from the CURELUNG project, as well as to invite and share these results with the prominent researchers in (Epi)genetics profiling and cancer, the CURELUNG consortium organized its “Mid-Term Workshop on Lung Cancer: from Molecular Knowledge to Treatment”. The quality of all publications and conferences offered by all the participants in the CURELUNG project demonstrated the added-value and positive impact of this project in the basic knowledge of lung cancer. The publications and thus the conferences have been related to the main objectives of CURELUNG and all the participants have been contributed.
Potential Impact (WP1)
The work that has been developed within WP1 has provided with a list of novel genes that are altered in lung cancer. The genes are either activated, i.e. oncogenes, or billalically inactivated, i.e. tumour suppressor genes and some of these alterations are specific of a given histopathological type. Thus, our findings benefits the general population and a have strong socio-economic impact. By one hand, the histopathological type-specific appearance of some of the genetic changes found in this work package may contribute to a more precise histopathological classification of the lung cancer. This is important because the type of histopathology in a lung cancer patient guides the therapeutics. However, still more important are the potential clinical interest, especially for the discovery of novel lung cancer treatments. We have found novel oncogenes that are activated in lung cancer and these or their immediate downstream targets could serve as therapeutic targets for the development or the implementation of lung cancer therapeutics. It is also remarkable that we have proposed some strategies to targets some tumour suppressor genes, such as MAX, based on the synthetically lethal concept. This concept is defined as follows: two genes are called synthetic lethal if mutation of either alone is not lethal, but mutation of both leads to death or a significant decrease in organism's fitness. In this regard we proposed that inhibition of another tumour suppressor gene, the so called BRG1 (or SMARCA4) can suppress the growth of the SCLCs that are deficient for MAX. Finally, another important impact of the results obtained within WP1 is the discovery of some mechanisms that account for acquired resistance to MET and FGFR1 inhibitors. Altogether, these discoveries will surely have implications in improving the clinical management of the lung cancer patients.
Potential Impact (WP2)
The work that has been developed within WP2 has provided a great core of information regarding the main objectives initially planned. We have determined the epigenetic profiles of an extensive panel of LC cell lines, identified novel cancer genes and miRNAS with epigenetic alterations and described epigenetic alterations that account for resistance to targeted drugs in LC. Additionally, we have identified a prognostic DNA methylation signature for early-stage NSCLCs.
Particularly, the epigenetic profile of this large set of lung cancer cell lines will serve as a basis for numerous and different studies regarding DNA methylation and lung cancer, including its role in driving resistance to therapies. MiRNAs studies have confirmed in cell lines the potential of these models and their role in delineating tumour aggressiveness and behaviour. Most importantly, the prognostic DNA methylation signature that we have identified in early-stage NSCLCs has a great implication for society and, obviously, a corresponding socio-economic impact. NSCLC is the leading cause of cancer-related death and the poor prognosis of patients with NSCLC is associated with several factors, among which are late disease diagnosis and the small number of effective drugs. But, importantly, the absence of validated prognostic biomarkers could also be relevant, because even patients with stage I NSCLC who undergo potentially curative surgical resection are at high risk of dying from recurrent disease, with a 5-year relapse rate of 35% to 50%. Although adjuvant platinum-based chemotherapy is beneficial in more advanced resected disease, in which most of the patients have a high risk of recurrence, it has failed to show a survival benefit for patients at stage I. One explanation for these negative data in the early stages could be the lack of biologic factors predicting their recurrence and the fact that, in the absence of useful biomarkers, all stage I NSCLCs are pooled, making it more difficult to draw meaningful clinical conclusions. Our DNA methylation signature has shown as an efficient tool to distinguish those patients more prone to recur. Thus, it is likely that adjuvant chemotherapy could improve the prognosis of these particular cases. Importantly, it is expected that, with the extension of lung cancer screening campaigns —mainly based in low dose CT scan—, more patients will present in early stages, increasing the importance of improving the treatment in this particular subgroup of cases. The combination of early diagnosis and more effective therapies will be the key to reduce the lung cancer mortality.
Potential Impact (WP3)
The ultimate impact of the CURELUNG project will be on human health. The aim for this work package 3 has been to improve the outcome of lung cancer through improved understanding of the disease leading to advances in diagnostics and therapy.
The discovery of DNA methylation changes related to lung cancer relapse means that we may be able to develop tests that allow clinicians to identify those who do not respond well to therapy and require additional treatment or monitoring. Moreover, DNA methylation patterns specific for lung cancer can be utilised to improve cytological diagnosis of the disease in bronchial washings and potentially applied to circulating cancer DNA in blood (alongside lung cancer screening). By defining the DNA changes associated with specific groups of lung cancers we can help develop drugs that are specifically targeted to those subtypes with the change, i.e. personalised medicine. Finally, the identification of miRNA signatures specific for lung cancer will aid development of diagnostic tools for use in small biopsies, such as those likely to come from increased use of CT screening for lung cancer.
One key aspect of lung cancer’s lethality is late detection, cancers often being detected when they are at a stage when clinical interventions are limited in scope and in effectiveness. Lung cancer screening not only benefits individual patients by providing earlier detection and improved survival, but society as a whole. Health costs are saved and economic costs associated with loss of work and skills are reduced. CURELUNG discoveries will aid the implementation of lung cancer screening by providing important diagnostic tools required to differentiate cancers from the larger numbers of CT detected non-cancer nodules. Further exploitation opportunities are provided by the development of biomarkers to aid the pre-screening identification of those at greatest risk of having lung cancer, with economic benefit stemming from lower costs associated with better targeted screening programmes.
In order to change clinical practice and have a significant impact on the health of those who suffer from lung cancer, it is important that these studies identifying gene alteration or techniques employed in diagnostic or prognostic tests are validated. We have taken significant steps to validation as part of Work Package 3 and provided opportunities for commercial exploitation of these techniques (including by our SME partners), or implementation in further prospective validation in the healthcare setting. Both these exploitation opportunities provide a positive contribution to the expanding market in biomarkers and personalised, precision and stratified medicine, expanding the range of targeted therapies applicable to lung cancer patients.
Potential Impact (WP4)
The potential impact of our findings is essentially on the management of patient with lung cancer ; application of a histo-molecular classification which recapitulates and correlates , better than ever with the previous WHO classifications , the oncogenic (genetic or epigenetic events ) and is more able to predict patients prognosis ,to stratify patients for choice of therapies modalities ,to facilitate and reduce the time of establishment of the oncogene drivers profile leading to therapeutic choices.
The main dissemination activities is through the wide dissemination of these new concepts and findings in high impact journals offered to the wide community of medical oncologists ,lung biologist, respiratory clinicians, and pharmaceutical laboratories preparing plans for new targeted therapies; the presentation of several indicators and organisers of randomization for patients entering now in clinical trials; the inclusion of these findings and novel findings concerning the new lung cancer histopathological classification. The highest level of dissemination is indeed inclusion of all these discoveries and new concepts in the World Reference book for Pathologists, provided in 2015. Due to worldwide dissemination of theses publication and reference WHO book, exploitation by pathologist and oncologist including patients in therapeutic protocol and evaluating the results based on patients survival. This therapeutical trial will be based on this new histo-molecular classification. This represents a strong guaranty for further exploitation of theses promising findings and their derived patients benefit worldwide.
Potential Impact (WP 5)
Based on our results, we have now launched two First-In-Human dose- escalation studies in Germany together with Novartis and AstraZeneca, which are aimed at testing novel FGFR inhibitors specifically in patients with FGFR1-amplified lung cancer.
AstraZeneca started recently a clinical trial to study the safety and tolerability of AZD4547 at increasing doses in patients with squamous NSCLC. AZD4547 is an orally bioavailable inhibitor of FGFR with potential antineoplastic activity. AZD4547 binds to and inhibits FGFR, which may result in the inhibition of FGFR-related signal transduction pathways, and, so, the inhibition of tumour cell proliferation and tumour cell death. The inclusion criteria beside others were FGFR1 or FGFR2 amplification.
Within the Novartis study, patients with specific FGFR1 amplification in lung cancer are treated with the new potent, selective and orally bioavailable FGFR inhibitor NVP-BGJ398. Within this study 26 patients have been treated, including 10 patients with FGFR1-amplified breast and 3 patients with FGFR1-amplified squamous cell lung cancer. One lung cancer patient with an FGFR1/CEP8 ratio of 2.6 by FISH analysis responded to 100 mg of NVP-BGJ398 with a 33% reduction in target lesions by CT scan at 8 weeks, confirmed at 12 weeks, and a substantial SUV decrease on PET. These observations provide early evidence that inhibition of the FGFR pathway is effective in patients with FGFR dependent cancers (Wolf et al. 2012, Cancer Research). Since we are running the first-in-man trial of the FGFR inhibitor BGJ398 in FGFR1-amplified squamous-cell lung cancers at our centre already, we have convinced Novartis based on our results published in Nature Genetics to include patients with FGFR1-amplified SCLC into a subsequent expansion phase/ phase II clinical trial of that FGFR inhibitor. Thus, we strive to funnel all potentially clinically relevant results from our basic discovery efforts into immediate clinical application and in the personalized treatment of patients.
To systematically subject patients to a screening of therapeutically relevant genomic changes as a prerequisite for participation in the genetic stratified clinical trials the network for molecular diagnosis and therapy of lung cancer (CIO at Cologne – Bonn) was initiated at the University hospital of cologne, one of the world's top five screening networks and unique for Germany. Within this network patients are systematically subjected to a screening of therapeutically relevant genomic changes. Within the past 12 months, over 2000 patients were diagnosed within the network, representing over 70% of regional patients and nearly 5% of all German Lung. All these developments are the result of the funded project.
Potential Impact (WP6)
The impact of our findings is the development of “in vivo” tools to evaluate the mechanisms of action, sensitivity and resistance to molecular targeted drugs in lung cancer.
We have generated three independent clinically relevant mouse models in order to evaluate the implication of histological subtype, inflammation, or tumor heterogeneity on lung tumorigenesis. Firstly, we developed clinically relevant mouse models that resemble the most two most frequent human non-small cell lung cancer (NSCLC) histologies: adenocarcinoma (ADC) and squamous cell carcinoma (SCC). We used these models to evaluate the effect of antiangiogenic agents. Our results have important clinical implications in the design of future antiangiogenic therapeutic trials, and emphasize the need for precaution when considering the possibility of enrolling SCC patients in clinical trials evaluating antiangiogenic drugs.
Secondly, and in order to evaluate the role of chronic inflammation on lung cancer initiation, we developed a chemical-induced carcinogenesis mouse model. Our model is unique in both the type of inflammatory condition we use to promote cancer and the sequential analysis of malignant lesions that allowed us to study separately the effects of inflammation in preneoplastic and full neoplastic lesions. We developed a novel mouse model of lung cancer that allows for the study of the interaction of chronic inflammation with lung tumorigenesis. Our study provides a useful new multistep animal model for examining the specific role of chronic inflammation in promoting lung carcinogenesis and identifies for the first time molecular mechanisms by which silica-mediated inflammation creates a favourable microenvironment for tumor progression.
Finally, we have also developed a valuable tool for preclinical evaluation of novel therapeutic strategies in cancer: the so-called patient-derived xenografts models (PDX). These models, obtained by direct implants of tissue fragments in immunocompromised mice, have great potential in drug development studies because they faithfully reproduce the patient's original tumor for both immunohistochemical markers and genetic alterations as well as in terms of response to common therapeutics. They also maintain the original tumor heterogeneity, allowing studies of specific cellular subpopulations, including their modulation after drug treatment.
The three above mentioned models are currently being used for the study of the mechanisms of action, sensitivity and resistance to other molecular targeted drugs apart from the antiangiogenic therapies already described.
Moreover, we have developed models of acquired resistance in order to determine the molecular aspects related with this major clinical hurdle for the success of these treatments in cancer patients. We have also focused on the study of synthetic lethality in small-cell lung cancer (SCLC) cell lines. Our results raise the possibility of developing a therapeutic strategy for SCLC patients with specific alterations.
Potential Impact (WP7)
We have conducted the largest Phase III multicenter randomized trial comparing adjuvant pharmacogenomic-driven chemotherapy versus standard adjuvant chemotherapy in completely resected Stage II-IIIA NSCLC. This is the largest study never initiated in the area of the molecular predictive markers in the field of completely resected NSCLC. Results from this study will have clear socio-economic and societal implications, as the improvement of treatment efficacy and the selection of patients for tailored therapies will ultimately modify treatment guidelines and increase patient survival.
In addition, the establishment of the Network Genomic Medicine, which now has developed to one of the largest international molecular screening networks for lung cancer, paradigmatically shows the implementation of molecular diagnostics and personalized treatment in broad clinical routine within a health care provider network. Of special interest is the rapid transformation of new discoveries (e.g. ROS1 fusions in adenocarcinoma) into clinical routine testing and treatment as well as the demonstration of a pronounced survival benefit by personalized treatment compared to chemotherapy in EGFR-mutated and ALK-translocated NSCLC. Based on these results, the AOK Rheinland/Hamburg, one of the largest German sickness funds reimburses multiplex genotyping of non operable lung cancer within NGM. To our knowledge, this is the first reimbursement model for comprehensive genotyping of lung cancer patients. We will extend this project to other sickness funds as well as to other cancer entities.
In summary, within the CURELUNG project we made different exciting discoveries with immediate clinical impact and we hope that they validate as relevant lung cancer drug targets and thereby improve the survival of lung cancer patients. The success of our project can be evaluated by the implementation of different newly discovered genomic markers into clinical routine as well as by different high impact publications in major internationally visible peer-reviewed biomedical journals. Moreover, our improved understanding of disease and the technological sophistication developed as part of CURELUNG are a further stimulus to the knowledge economy in each of the partner regions.
List of Websites:
Project website: www.curelung.eu
Project coordinator: Dr Manel Esteller, Tel. +34 93260 7140, email: mesteller@idibell.cat
Project manager: Mrs Verónica Padial, Tel. +34 93260 7143, email: vpadial@idibell.cat / pebc@idibell.cat / project@curelung.eu