Over expression of the EGFR gene has been noted in several cancers including lung, head and neck and breast cancers. Mutations of the extracellular part of the gene occur frequently in brain tumors. Recently this gene has received much interest and publicity because mutations in the tyrosine kinase domain of the gene appear to predict for response to small molecule inhibitors ("designer drugs"), such as Iressa (gefitinib). Response rates of 10-30% have been noted, including some spectacular responses, and may relate to mutations in the TK domain.
We have performed mutational analysis of over 600 lung cancers from around the world, and the following are our major conclusions: a) TK domain mutations are more frequent in Pacific Rim countries, especially Hong Kong, and target adenocarcinoma histology, women and never smokers; b) We have initiated four cell lines with the mutant form and they over express the phosphorylated form of EGFR and are more sensitive to growth inhibition by Iressa and c) mutations in the TK domain are very rare in other types of cancers. However, mutations in other parts of the gene have been reported to be frequent in glioblastomas. EGFR mutations are of particular interest as they are the first known molecular change to target cancers arising in never smokers. Thus a study of them may shed light on why these tumors are arising at an ever increasing rate in women, especially in Pacific Rim countries, not exposed to tobacco carcinogens.
We will relate the role of mutations in pathogenesis and prediction of sensitivity to Iressa. In particular, the relationship to smoke exposure and smoke exposure related genetic changes such as RAS mutations will be described. We have utilized fresh cancers and our in vitro models to study the relationship between mutations and response to targeted therapies. Mutations in the TK domain of the gene will be correlated with structural and expression changes in other regions of the gene. In particular we will investigate the relationship between the "glioblastoma" specific and "lung cancer" specific mutations. application to clinical care and selection of patients for optimal response to targeted therapies.

Despite improvements in combined modality treatment, epithelial cancers - and particularly lung cancer - are still major killers throughout the world. Moreover, 85% of lung cancers occur in ever-smokers. The cumulative risk of lung cancer in smokers by age 75 is 16% for men and 9.5% for women.
Although still in an early stage of development, chemopreventionis an approach with tremendous potential to make an impact in the control of these diseases in the next millennium.1 While chemoprevention has been shown to have activity in other areas, it has failed to achieve positive results in many lung cancer chemoprevention trials, indicating the need to develop a model that will enable us to identify a high-risk cohort and target the right population to prevent lung cancer. Building on data from a groundbreaking molecular study of tobacco-induced genetic changes in bronchial epithelial tissue, we developed a translational research program incorporating genetic risk modeling, molecular screening, imaging, and targeted chemoprevention approaches. We are developing a genetic risk model through the integration of genetic susceptibility markers (including mutagen sensitivity, DNA repair capacity, GST-?, GST-µ, and DNA adducts) and biomarkers at the tissue level (RAR-?, genetic instability, DNA methylation, FHIT inactivation, and LOH of 3p, 9p, or 17p). Furthermore, we are using molecular imaging to detect molecular changes in target tissue for use as surrogate markers for targeted therapy.
Many promising new candidate chemopreventive agents that have been developed for lung cancer have not been effective as single-agent treatment for the prevention of primary lung cancer. Therefore, we will be conducting a vanguard chemoprevention clinical trial using Iressa, celecoxib, and farnesyl transferase inhibitor (FTI), with the following objectives:
1) to develop biologically-based treatments to prevent disease recurrence and development of second primary tumors in patients at risk for these events;
2) to understand the molecular events in premalignant tissue that underlie progression to malignancy; and
3) to combine data that is based on treatment outcome with molecular and imaging data to create a risk model for disease recurrence and development of second primary tumors. After completion of these studies, the most promising agent will be forwarded to a clinical trial for primary lung cancer prevention.

Signal transduction, a process leading to changes in the expression pattern of particular genes, occurs via a cascade of protein-protein interactions. It is a very fast, efficient and accurate process, controlled very tightly at the ON and OFF levels. At the ON level, the protein-protein interactions, enhanced by protein modifications, ensure efficient processing of the signal. At the OFF level, several mechanisms operate, one of which is endocytosis. Endocytosis defines the process by which molecules, macromolecules and even cells cross the plasma membrane from the extracellular matrix into cells and depends on a large number of protein-protein interactions, mediated by specific protein modules. A large number of signal transduction pathways as well as endocytosis are abrogated in tumorigenesis. Therefore, intervention with these processes is considered a possible therapeutic approach. In this proposal we aim at characterizing peptides that inhibit lung cancer cell growth and study the EHD proteins, which participate and control endocytosis. Since several signaling pathways are upregulated in lung cancers, a strategy was developed to isolate specific peptides that interfere with these pathways, thus leading to growth inhibition of lung cancer cells. The procedure includes the following steps: a. Construction of cDNA libraries, encoding 10-20 amino acids. b. Transfection of different cells in tissue culture with the libraries, assuming that a large number of the cDNA molecules encode small peptides. c. Screening for transfected cells in which the expressed peptides cause apoptosis. d. Recovering the transfecting plasmids and sequencing the encoding inserts. Several libraries were screened and, according to the cDNA sequences obtained, peptides were designed, with or without modifications. These peptides will be used to test peptideinduced apoptosis of lung cancer cells in tissue culture, IGF-1R fate in peptide treated cells, since IGF-1 induced, IGF-1R mediated signaling is upregulated in lung cancer, the cellular destination of the tested peptides, peptide-induced downregulation of gene activity, toxicity in animals and the influence of peptides on lung cancer tumor growth in animals. We have identified a new family of EH domain containing proteins, designated EHDs, which participate and control different stages of endocytosis. They are located in different intracellular compartments and function via interaction between the different family members and other endocytic proteins. EHD1 is mainly a resident of the endocytic-recycling compartment, EHD2 is a plasma membrane protein, while EHD3 characterizes the tubular structures of the endocytic recycling compartment. Overexpression of EHD1 led to down regulation of IGF-1 induced, IGF1R mediated signaling. In this proposal we wish to characterize biochemically the EHDs, and to create a double EHD knockout model to study the pathophysiology associated with the absence of two of the family members.

Lung cancer is a very common cancer associated with grave prognosis. The majority of the lung cancer cases are associated with exposure to cigarette smoke, either by active smoking or by secondhand passive smoking. Therapy of most cases diagnosed with lung cancer is very disappointing and no significant improvement in outcome was noted in recent years. Small tumors can be treated successfully by surgical removal. Our research is aimed at the identification of polymorphic DNA variants associated with aggressiveness and response to therapy. Comparison of thousands of Single Nucleotide Polymorphisms - SNPs - in isolated tumor cells and peripheral blood cells of the same individuals is expected to identify tumor suppressor genes relevant to lung cancer development and progression. Global gene expression of tumors and isolated malignant and non-malignant cell populations micro dissected from the tumor is expected to identify diagnostic and prognostic genes, gene clusters and gene signatures. This study is expected to lead to the development of tests that will identify those individuals at high risk that may benefit from early detection. It may lead to the development of sensitive and reliable tests for early detection that may result in improved survival. Finally, targets for rational therapy will be characterized and validated.
The approach for the identification of polymorphic variants linked to tumor aggressiveness and response to therapy is based on a large-scale analysis of thousands of SNPs scattered all over the genome. By comparing large number of SNPs in patients and controls it should be possible to find SNPs that are linked to various disease presentations, to response to therapy and to a tendency to develop specific side effects of therapy. We use the GeneChip Human Mapping 10K Array (Affymetrix), which can genotype more than 10,000 SNPs by allele-specific hybridization.
Comparison of DNA extracted from tumor cells with DNA from normal tissue of the same individual can also detect loss of heterozygosity (LOH), a common form of allelic imbalance. The detection of LOH has been used frequently to identify genomic loci that harbor tumor suppressor genes. We compare DNA from lung cancer cells and peripheral blood leukocytes using the GeneChip 10K array in a search for LOH that may harbor tumor suppressor genes relevant to lung cancer.
Finally we plan to combine the global expression analysis of whole lung tumor samples with the analysis of isolated malignant and non-malignant cells found in the tumor mass. The combined processing of the emerged information will also be integrated to information regarding DNA polymorphism and LOH, and crossed with patient characteristics. Such an integrated approach may lead to results of biological, diagnostic and therapeutic relevance.

Epidermal growth factor (EGF) promotes cellular growth by binding to its receptor, EGFR. EGFR is expressed in many cancers and frequently is over-expressed in NSCLC. Binding of EGF to its receptor leads to activation of numerous pathways that are critical for the regulation of cell growth. Over-expression of EGFR has been linked to poorer outcomes in NSCLC.

Phase II trials of erlotinib, a well-tolerated oral inhibitor of EGFR showed promising results in patients with NSCLC. On the basis of the Phase II data, the National Cancer Institute of Canada Clinical Trials Group (NCIC-CTG) conducted an international Phase III, placebo-controlled trial (BR.21) to evaluate erlotinib in patients with advanced NSCLC who had failed standard chemotherapy. The trial did not pre-select patients on the basis of EGFR expression, and only 10% of patients achieved response. However, despite this low response rate, patients treated with erlotinib lived significantly longer than those treated with placebo In addition, treatment with erlotinib was associated with significantly better control of lung cancer-related symptoms and improvement in quality of Life.

Multiple molecular mechanisms have been proposed to influence EGFR signaling and response to EGFR inhibitors. We propose to evaluate biomarkers and genetic markers in the EGFR pathway and to correlate these and other clinical predictive and prognostic markers with outcome. The following markers will be analyzed:
1. EGFR family members/ligands: EGFR, HER-2, HER-3, HER-4, EGFRvIII variant, TGF
2. Signaling markers downstream of EGFR such as phospho-MAPK and phospho-PKB/ Akt
3. Mutations in the EGFR family of receptors

The results of this research have the potential not only to improve the therapeutic ratio of EGFR therapy for patients, but also to provide rational direction for introduction of these agents into national health care delivery systems, whether privately funded, publicly funded or a combination of both. The financial implications of finding a means to select patients who are most likely to benefit will be significant as the current cost of this treatment in the US is ~$25,000-$30,000 per patient per year.

Non small cell lung cancer (NSCLC) is generally thought to be curable, if the primary tumor is surgically resectable, if only lymph nodes on the ipsilateral side of the tumor are involved and if there are no distant metastases. The definition of disease extent is importantly achieved by imaging and mediastinoscopy and currently is based mainly on the insights gained with computed tomography (CT).
With the advent of FDG (Fluorodeoxyglucose)-PET (Positron Emission Tomography) and integrated PET-CT it has become clear, that these imaging modalities are substantially more accurate in defining disease extent in lung cancers than CT. This is due to the fact that the criteria of disease involvement is a molecular biological one, i. e. glucose uptake and metabolism, rather than a morphological one such as lymph node size. It can therefore be hypothesized, that the currently used criteria to define curability of NSCLC are partly outdated, or if the concepts are correct, survival statistics of patient subgroups are substantially influenced by using PET or PET-CT for staging rather than CT alone. PET could influence patient stratification in two ways. First, a more accurate staging by PET would be expected to lead to a more accurate multidisciplinary therapy and better survival of patients, compared to prior statistics obtained in patients staged with CT alone. Second, the specific uptake value (SUV) of lesions observed in PET may correlate with the aggressiveness of the tumor, which in turn could influence treatment and hence survival of the patient as well.
The projects to be worked on is to analyze our vast data base on patients who have been treated at the University Hospital of Zurich in the last years and have received a PET or PET-CT scan (alone last year over 500 patients received a PET for NSCLC at our institution). The aims are to evaluate the two questions raised above, and also to try to identify whether there are PET staged subgroups of patients other than those classified as curable, who have better survival statistics then when staged with CT alone.
The project will also analyze patients with Small Cell Lung Cancer (SLC) and malignant pleural mesothelioma (MPM) who have undergone PET or PET-CT scanning. The aims of this analysis will be similar than those for NSCLC, namely to identify whether PET and particularly PET-CT yield new criteria to provide better prognostic factors regarding the survival of the patients. Specifically, the question will be asked whether PET-CT can provide staging information which permits curable resection of a certain subgroups of patients with MPM.