The primary project of the Biology Division is to elucidate the molecular basis of multistage human lung carcinogenesis. Whole genome scanning of tumor suppressor genes as well as cancer susceptibility genes is under way to achieve this goal. Once genes of interest are identified, functional analyses are performed to elucidate the pathogenetic involvement of the genes in lung carcinogenesis. Mechanisms of DNA double- strand break repair are also being investigated at the Division.

Germline LKB1 mutations cause Peutz- Jeghers syndrome, a hereditary disorder that predisposes to gastrointestinal polyposis and several malignant tumors. Since there are a few reports showing the presence of somatic LKB1 mutations in lung cancers, a large number of lung cancer cell lines and lung adenocarcinoma (AdC) specimens were examined for LKB1 genetic alterations ⁽⁴⁵⁾. LKB1 alterations were frequently detected in the cell lines (21/70), and were significantly more frequent in those with KRAS mutations. In particular, homozygous deletions (HDs) were detected in all histological types. In lung AdC specimens, LKB1 mutations were found in 7/91 (8%) male smokers, but in none of 64 female and/or nonsmokers, and were significantly more frequent in poorly differentiated tumors. The results indicate that LKB1 genetic alterations preferentially occur in a subset of poorly differentiated lung AdCs that appear to be unique to smoking males.

Whole genome scanning of 43 lung cancer cell lines at a 100-kb resolution led to the identification of 51 genomic regions with HDs ⁽⁴⁶⁾. The regions contained 113 genes, including two known tumor suppressor genes, RB1 and CDKN2A, and 8 candidate tumor suppressor genes. Three miRNA genes were also mapped to a region with HD at 21q11-q21. The present study provides a list of protein- and miRNA-encoding genes whose inactivation is possibly involved in lung carcinogenesis.

Several models of cancer progression, including clonal evolution, parallel evolution, and same-gene models, have been proposed. To investigate the validity of these models, whole-genome allelic imbalance (AI) scanning and mutational analysis of the p53, EGFR, and KRAS genes were performed on 8 sets of primary and metastatic lung cancers ⁽⁴⁷⁾. Accumulated genetic alterations were similar between primary and the corresponding metastatic tumors, and the majority (>67%) were commonly detected in both primary and metastatic tumors. On the other hand, in 7 cases, several genetic alterations were accumulated only in the metastases. Among these alterations, AI on chromosome 11p was the most frequently observed (4/7). Likewise, in 4 cases, several genetic alterations were seen only in the primary tumors. The history of each case indicated that all three models are applicable to lung cancer progression. According to the clonal and parallel evolution models, a metastasis suppressor gene(s) for lung cancer was suggested to be present on chromosome 11p. Lung AdC cells with amplification of the mutated EGFR gene or deletion of the wild-type gene were suggested to metastasize more preferentially to the brain and/or to proliferate selectively in the brain ⁽⁴⁸⁾.

Biphasic pulmonary blastoma is a rare lung tumor with epithelial and mesenchymal components. To understand the histogenesis of this tumor, molecular analyses were performed in the epithelial and mesenchymal components of a case of pulmonary blastoma ⁽⁴⁹⁾. AI at chromosome regions 14q and 17p and the b-catenin mutation were commonly detected in both the epithelial and mesenchymal components. On the other hand, AI at chromosome regions 3p and 9p and the p53 mutation were detected only in the mesenchymal component, and AI of chromosome 6 was detected only in the epithelial component. These results indicate that this biphasic tumor is of monoclonal origin and the phenotypic heterogeneity of this tumor is due to the differences in the accumulated genetic alterations in each component of the tumor.

MYO18B was cloned as a candidate tumor suppressor gene at the Division. HOMER2, a Homer/Ves1 family protein, was identified as a binding partner of MYO18B by a yeast two-hybrid screen ⁽⁵⁰⁾. These proteins were co-localized in the regions of membrane protrusion and stress fibers, which are known as regions with filamentous actin-rich structures. Expression of HOMER2 enhanced the ability of MYO18B to suppress anchorage-independent growth. Thus, it was suggested that HOMER2 and MYO18B cooperate to cause tumor suppression. It was also found that restored MYO18B expression in a mesothelioma cell line inhibited the anchorage-independent growth and motility in vitro and the growth in SCID mice in vivo ⁽⁵¹⁾. The result indicates that restoration of MYO18B expression may be a useful strategy for the treatment of not only lung cancers, but also of malignant mesothelioma.

CLASS-11, which consists of genes encoding pro- and anti-inflammatory cytokines, was shown to identify stage I lung AdC patients with a poor prognosis ⁽⁵²⁾. The OCIAD2 gene was shown to be expressed in invasive lung AdCs and to be associated with the favorable prognosis of AdCs with BAC components ⁽⁵³⁾.

The mouse Atf2 gene was shown to act as a tumor susceptibility gene of mammary tumors by activating a group of target genes, including Maspin and Gadd45a ⁽⁵⁴⁾.

Mammalian cells have a mechanism for mutagenic repair of DNA double-strand breaks (DSBs), namely, microhomology-mediated end joining (MMEJ). Recent in vitro studies have indicated the involvement of Ku proteins in MMEJ. To clarify whether Ku proteins are essential for MMEJ in vivo, linearized plasmid DNAs with microhomologous sequences were introduced into Ku80-proficient and Ku-deficient cells, and subjected to MMEJ and non- homologous end joining (NHEJ) ⁽⁵⁵⁾. Ku80 deficiency caused approximately 75% reduction of the MMEJ activity and >90% reduction of the NHEJ activity. These results indicated the existence of a Ku-dependent pathway for MMEJ; however, MMEJ is less dependent on the Ku80 protein than NHEJ. The fraction of MMEJ products increased in proportion to the increase in the amounts of the substrates, suggesting that MMEJ might function as a salvage pathway for DSBs that cannot be repaired by NHEJ.