Gene-expression Profiles in CNS-metastatic Non-small Cell Lung Cancer

Brain metastases occur in 30-50% of lung adenocarcinoma (LAC) patients and confer a worse prognosis and quality of life. A better selection of in-risk patients through a more accurate biomarker could improve the benefit of prophylactic therapies. The aim of this prospective study was to determine a gene-expression profile of primary tumor associated with brain metastasis (BM) and to evaluate the overall survival (OS) in patients with advanced LAC.

Introduction.Lung cancer is the first cause of cancer death in the world. Eighty five percent of patients are diagnosed yearly with non-small cell lung cancer (NSCLC). Despite efforts, innovations, and progress in diagnosis and treatment of these patients, overall survival (OS) at 5 years of diagnosis is only 15%. The central nervous system (CNS) is a devastating and frequent site of metastasis development in NSCLC. The reported incidence of CNS metastasis in patients with NSCLC is 54% with an OS of <1 year after diagnosis. Age, clinical stage, gender, and initial treatment period are some of the reported with CNS metastasis development-related factors in patients with NSCLC. We recently described that high carcino-embryonic antigen (CEA) serum levels (>40 ng/dL) at diagnosis and adenocarcinoma histological type are independently associated with higher risk if brain metastasis development. However, due to their lack of specificity of all this reported factors, we are required to detect biomarkers to predict brain metastasis in patients with NSCLC with the objective to prevent their development.Metastasis process is complex it involves well-defined molecular-printed steps, as invasion, vascular intravasation, implantation and growing at the new specific organ. Higher motility, capacity for extracellular matrix degradation, immune system evasion, and adhesion at the new specific organ are some of the tumoral cell characteristics required to metastasize. In particular, molecular events for brain metastasis development from primary NSCLC even if no at al described, are currently well understood. Gene-expression microarray allow analyzing the expression changes of thousands of genes simultaneously, distinguishing the altered expression of neoplastic cell genes from normal tissue. Identification of biological patterns, pharmacological molecular targets, and biomarkers for prognosis and evaluation for therapeutic response, and even a more specific neoplasia classification are some of the results of gene-expression microarray studies in hematological, breast and NSCLC cancer. Determination of tobacco-smoke transcriptional changes in oncogenes and antioncogenes had been determined by the use microarray data. A gene-expression profile of 20 genes differentiates health lung tissue and lung cancer with higher specificity than histopathologic evaluation. Furthermore, gene-expression microarray studies had been developed for predicting survival and recurrence in early NSCLC stages for identification of patients who could have benefit of adjuvant therapy, with promising result.The objective of this study was to identify a gene-expression profile from primary lung adenocarcinoma related with brain metastasis, and to evaluate in a prospective manner their prognostic significance on survival in patients with advanced disease.Experimental Design These study used clinical, longitudinal, prospective, observational, and analytical cohort s with the selection of a nonprobabalistic-type sample.In a prospective manner, from January 2009 to June 2011, patients admitted to our Institute confirmed histologically confirmed stage IIIb and IV of LAC were eligible for study inclusion. Analyzed clinical variables comprised smoking history, gender, age, general condition and brain metastasis development. Primary tumor core-biopsy was performed prior to any treatment and snap-frozen. A single pathologist evaluated all tumors. Standard platinum-based chemotherapy was employed for all patients. All patients were submitted to Magnetic resonance imaging (RM) to evidence the presence or absence of brain metastasis (BM). During follow-up, we carried out Computed tomography (CT) of the brain stem for this purpose. Preliminary statistical analysis was performed utilizing the Student t, the Mann-Whitney U, the χ2, or the Fisher exact test. Once the study ends, we will conduct logistic-regression multivariate analysis. Global survival time will be analyzed with the Kaplan-Meier technique, and comparisons among groups will be performed with the log-rank test. High Risk characteristics as gender, histology, and age related with greater frequency of development of metastasis to CNS were analyzed.The study was carried out according to the principles of ClinicalTrials and accepted by the INCan Bioethical and Scientific Committees (reference numbers INCAN/CC/067/08). A collaboration agreement was signed with the INMEGEN and was approved within the Health Research Sectorial Fund (FSIS) CONACyT-México (SALUD-2009-01-115552).

Selection for obtaining tumor biopsies depended on the characteristics of the patient and of the tumor. The biopsy could be taken by means of CT guided tru-cut needle, open thoracoscopy, or bronchoscopy with optic fiber. Each tumor sample was divided into two; one sample was analyzed by the INCan Pathology Service (by a sole observer blinded to the clinical variables) for their histological clinical diagnosis and quantification of neoplastic cellularity, and the remaining half was immediately stored at -80°C until processing for RNA extraction.RNA Extraction:RNA was extracted from frozen tumor biopsies, weighted and cryofractured in liquid nitrogen. Extraction and purification of total RNA from tissue (up to 5 mg tissue) procedure was done using RNeasy Micro Kit (QIAGEN, Germany) (cat. 217084). RNA was analyzed using the Agilent 6000 Chip (Agilent Technologies, Santa Clara, C.A.). RNA quality consisted on the obtention of RNA Integrity Number (RIN) > 8. RNA samples were chosen for microarray analysis when quality and concentration were obtained for this purpose. Microarray Expression:We are using an Affymetrix platform, which consists of an in situ synthesized oligonucleotide microarray. The GeneChip® that we are employing is the Human Gene 1.0 ST Array, which allows analyzing the expression of 28,869 different genes (each transcript is represented by 26 disperse probes along its entire length, present in the human genome, which covers 99% of the sequences present in the RefSeq data bank of the GenBank. The Two-Cycle Target Labeling protocol is followed as suggested by the manufacturer and described succinctly as follows and shown graphically in the figure at the end of this Materials and Methods section: Total RNA (100 ng/uL) is retro-transcribed using T7-Oligo(dT) Promoter primers in the synthesis reaction of the first complementary DNA chain (cDNA). After treatment with RNAase H, the second cDNA chain is synthesized, which will serve as template in the Transcription reaction in vitro (TVI). The first TVI reaction is carried out with T7 polymerase RNA and unmarked ribonucleotides for the amplification of complementary RNA (cRNA). These cRNAs are retro-transcribed in the synthesis step of the first cDNA chain of the second cycle using random primers. Subsequently, T7-Oligo(dT)-Promoter primers are employed for the synthesis of the second chain of cDNA, generating the double-helix cDNA template that contains the promoter T7 sequence. This cDNA is amplified and marked in a second TVI reaction using the T7 polymerase DNA and a mixture of ribonucleotides/biotinylated nucleotide analog. These target biotinylated cDNAs are cleaned, fragmented, and hybridized to the GeneChip® expression microarrays. After hybridization, fluorescent marking was performed with a streptavidin-phycoerythrin and biotinylated anti-streptavidin antibody amplifier system. Fluorescence is detected with a high-resolution laser scanner.Reading and Analysis of Expression Microarrays:We will utilize the Expression Console of Affymetrix software to evaluate the quality of the microarrays and will correct the background signal with standard methods [49,50]. For signal intensity-level normalization, we will employ "non-supervised" methods, which use the data of all of the experiment's microarrays; this will allow us to render these comparable among themselves, because what a microarray assay evaluates is gene expression by means of fluorescence intensities; thus, it is important to establish cut-off points from which we can consider under- or overexpression. Once the data are normalized, the final stage is to resume the values of all the probes that exist for each gene in the microarray in a sole value, which is the gene's expression level. For detection of differentially expressed genes, we will construct a linear model that includes all of the experiment's microarrays simultaneously; in this manner, we will be able to analyze contrasts between group A (patients with metastasis to CNS during the follow-up period) and group B (patients without metastasis to CNS during the same period). As internal control, we will prove the differential expression of some genes by RT-PCR in real time. Results will be represented in blocks of heat maps and in tables, in which the most active genes are listed along with their functional description. For risk-profile validation, we will perform leave-one-out cross validation method and will postulate a second cohort for this objective.