Sputum MicroRNA Pro ling: A Novel Approach for the Early Detection of Non-Small Cell Lung Cancer

Purpose: MicroRNAs (miRNAs) post-transcriptionally regulate hundreds of gene targets involved in tumorigenesis thereby controlling vital biological processes, including cellular proliferation, di erentiation and apoptosis. MiRNA pro ling is an emerging tool for the potential early detection of a variety of malignancies. is study was conducyed to assess the feasibility and methodological robustness of quantifying sputum miRNAs, employing quantitative real-time polymerase chain reaction (RT-qPCR) and cluster analysis on an optimized miRNA pro le as a novel approach for the early detection of non-small cell lung cancer (NSCLC). Methods: e relative expressions of 11 miRNAs in sputum (miR-21, miR-145, miR-155, miR-205, miR-210, miR-92, miR-17-5p, miR-143, miR-182, miR-372, and let-7a) in addition to U6 were retrospectively assessed in four NSCLC-positive and four negative controls. Subsequently, a set of ve miRNAs (miR-21, miR-143, miR-155, miR-210, miR372) was selected because of degree of relatedness observed in the cluster analysis and tested in the same sputum sample set. e ve optimized miRNAs accurately clustered these eight retrospective patients into NSCLC positive cases and negative controls. e ve miRNA panel was then prospectively quanti ed in the sputum of 30 study patients (24 NSCLC cases and six negative controls) in a double-blind fashion to validate a ve miRNA panel using hierarchical cluster analysis. Results: e optimized ve miRNA panel detected NSCLC (83.3% sensitivity and 100% speci city) in 30 prospectively accrued study patients. Conclusion: Sputum miRNA pro ling using cluster analysis is a promising approach for the early detection of non-small cell lung cancer. Further investigation using this approach is warranted. ORIGINAL RESEARCH © 2012 CIM Clin Invest Med • Vol 35, no 5, October 2012 E271 Correspondence to: Dr. Wilson Roa Department of Radiation Oncology Cross Cancer Institute 11560 University Avenue Edmonton, Alberta, Canada T6G 1Z2 Email: [email protected] Manuscript submitted 20th January, 2012. Manuscript accepted 4th September, 2012. Clin Invest Med 2012; 35 (5): E271-E281. MicroRNAs (miRNAs) are a group of recently discovered sma l l non-prote inc o d ing R NAs . MiR NAs p osttranscriptionally regulate the expression of hundreds of target genes, thereby controlling a wide range of tumorigenic processes including cellular proliferation, di erentiation, and apoptosis [14]. MiRNAs are expressed in a tissue-speci c manner and are, therefore, candidate biomarkers for cancer detection and non-malignant disease processes [15-20]. Furthermore, miRNAs may function as tumor suppressors or tumor promoters (oncomirs), with dysregulated miRNA expression leading to carcinogenesis. [9, 16, 21-22]. MiRNAs have shown potential for the diagnosis and classi cation of human malignancies [14-16]. MiRNA dysregulation (overand under-expression) in surgically resected lung tumor tissues is reliably correlated with the diagnosis and prognosis of the lung cancer patients [14-16, 21-22]. It is becoming evident that miRNA-based detection techniques could improve upon cytology (with its low sensitivity) and of computed tomography (with its high falsepositive rate), and seem to hold promise to develop screening tools for the early detection and monitoring of lung cancers. ere are few available studies that have examined the feasibility of aberrant miRNA expression pro les in sputum as a potential screening tool for lung cancer. In one such study, Xie et al. reported that miR-21 over-expression can be used for the diagnosis of NSCLC with a sensitivity of 69.6% [9]; however, a single miRNA lacks sensitivity and speci city as a cancer biomarker because lung cancer is a heterogeneous entity, with oncogenesis arising through a poorly understood, complex, multi-step process [13, 15, 22-26]. ere is considerable interest in using a combination of speci c miRNAs (a miRNA prole), for use in the early detection of lung cancer [24-26] in patients. Yu et al. reported that four miRNAs (miR-21, miR486, miR-375 and miR-200b) distinguished lung adenocarcinoma patients from normal subjects with a sensitivity and speci city of 80.6% and a 91.7%, respectively [24]. Xing et al. showed that the pro le of miR-205, miR-210, miR-708 detected NSCLC with a relatively low sensitivity (73%). [25]. It is apparent that the selection and optimization of miRNAs and miRNA pro le quantitative analysis methodology requires further investigation prior to its widespread acceptance and implementation as a screening tool [27-29]. Cluster analysis is an analytical mathematical method that identi es groups of samples that perform similarly or possess similar characteristics. Cluster analysis can lead to a diagnosis when a distinct grouping, based on the data, is observed [30]. In this study, hierarchical cluster analysis, a special clustering algorithm that is based on a tree-like structure, was used to study the relationship between the observations of miRNA pro les to di erentiate patients with NSCLC from healthy controls [31]. e objective of this study was to determine the optimum sputum miRNA pro le for use as a future screening test for early NSCLC detection. Eleven miRNAs that have been shown to be dysregulated, and have variable expression strongly correlated with the presence or absence of cancers and NSCLC, were selected [13-27]. An optimized ve miRNA panel was then determined retrospectively on a small subset of eight study patients and further validated prospectively in a larger subset of patients for the diagnosis and early detection of NSCLC. Materials and Methods