An NGS based study in a randomized NSCLC population: identification of new potential prognostic biomarkers

Lung cancer is among the most deadly cancers in the world. It is divided into two subgroups: Non-small cell lung cancer (NSCLC, 85%) and Small cell lung cancer (SCLC, 15%). NSCLC can be classified in two major classes: Adenocarcinoma (ADC) and Squamous cell carcinoma (SCC). Each different histology type is characterized by a different pattern of molecular signatures. Since the discovery of EGFR TKIs (EGFR-tyrosine kinase inhibitors), which underline that the presence of a mutation, such as EGFR (Epidermal Growth Factor Receptor, 15% in ADC patients), can be an important predictive factor for drug treatment, it was clear that efforts were necessary for improve the knowledge molecular pathogenesis of lung cancer. Kirsten rat sarcoma viral oncogene homolog (KRAS) gene encodes for membrane bound GTPase proteins, which are mediators of the Mitogen-Activated Protein Kinase (MAPK) pathway. When mutated these proteins continually activate the MAPK pathway, leading to cell proliferation. Although KRAS mutations are the most common recurring molecular events in ADC (30%), little progress has been made from its discovery. Also no direct KRAS inhibitors have proven clinically efficacy. Additionally, the prognostic value of KRAS in patients with NSCLC is controversial in the literature. Our previous works were unable to confirm if mutated KRAS was a prognostic marker in NSCLC. The aims of this project were i) to find prognostic biomarkers associated to KRAS. We genotype 131 samples of NSCLC patients in a TAILOR study. The hypothesis was related to the idea that the association between KRAS and another gene can lead to a strong prognostic value of both together compared to KRAS alone. ii) Finding of other independent prognostic biomarkers. NGS analysis was performed using a wide gene panel of 111 most relevant genes involved in cancer, COX models and Kaplan Meier curves were used to evaluate associations between mutations, clinical characteristics, outcomes. CRISPR-Cas9 technology was used to generate LKB1 deleted clones of human NSCLC cell lines H1299. Difference in sensitivity to drug treatments between clone and parental cell line was determined by MTS assays. Genes obtained by NGS analysis were accounted only if they reach almost 5% of frequency of patients. The three most mutated genes were: TP53, KRAS and LKB1. We analyzed KRAS and LKB1 (an emerging gene nowadays). We had 84 patients KRAS wt, 47 patients KRAS mut, 110 LKB1 wt and 21 LKB1 mut. Comparing KRAS status and LKB1 status in all the patients we obtained 76 patients KRAS-LKB1 wt and 13 patients with both mutations. Over 47 KRAS mutated patients, 13 patients of them were also LKB1 mutated, accounting 27.65% of frequency among the KRAS mutated. When analyzed in combination, KRAS and LKB1 were not related to any statistical association with PFS (Progression Free Survival) or OS (Overall Survival). In contrast, when analyzed alone, LKB1 confers a negative prognosis with an HR for OS of 1.57, 95% CI 0.97-2.55, p-value=0.066. It doesn’t reach the statistically significance, mainly due to limited number of patients. Further studies aimed at validate this finding will be necessary. Because the prognostic value of LKB1 was found only in OS analysis and not in PFS, its role can be most relevant in the first line treatment. For this reason we generated H1299 LKB1 deleted. In this setting, the lack of LKB1 did not sensitize cells to cisplatin and pemetrexed when used alone. LKB1 deleted clone were more sensitive to treatments used in combination when compared to parental cells. This data are in contrast with our hypothesis. Not all the patients received this chemo-regimen, and this can explain the discrepancy. Our results needed to be confirmed in a prospective trial with appropriate multivariate analysis. No specific treatment for LKB1 are available. Future efforts would address the LKB1 role in cancer and exploit its loss of function as possible treatment strategy.