Abstract
BACKGROUND: Nasopharyngeal carcinoma (NPC) is a solid tumor of the head and neck. Multimodal therapy is highly effective when NPC is detected early. However, due to the location of the tumor and the absence of clinical signs, early detection is difficult, making a biomarker for the early detection of NPC a priority. The dysregulation of small non-coding RNAs (miRNAs) during carcinogenesis is the focus of much current biomarker research. Herein, we examine several miRNA discovery methods using two sample matrices to identify circulating miRNAs (c-miRNAs) associated with NPC. METHODS: We tested two miRNA discovery workflows on two sample sources for miRNAs associated with NPC. In the first workflow, we assumed that NPC tumor tissue would be enriched for miRNAs, so we compared miRNA expression in FFPE from NPC cases and controls using microarray and RNA-Seq technologies. Candidate miRNAs from both technologies were verified by qPCR in FFPE and sera from an independent NPC sample set. In a second workflow, we directly interrogated NPC case and control sera by RNA-Seq for c-miRNAs associated with NPC, with candidate c-miRNAs verified by qPCR in the sera from the same independent NPC sample set. RESULTS: Both microarray and RNA-Seq narrowed the miRNA signature to 1-5% of the known mature human miRNAs. Moreover, these two methods produced similar results when applied to the same sample type (FFPE), with RNA-Seq additionally indicating "unknown" miRNAs associated with NPC. However, we found different miRNA profiles in NPC sera compared to FFPE using RNA-Seq, with the few overlapping miRNAs found to be significantly up-regulated in FFPE significantly down-regulated in sera (and vice versa). Despite the different miRNA profiles found in FFPE and sera, both profiles strongly associated with NPC, providing two potential sources for biomarker signatures for NPC. CONCLUSIONS: We determined that the direct interrogation of sera by RNA-Seq was the most informative method for identifying a c-miRNA signature associated with NPC. We also showed that there are different miRNA expression profiles associated with NPC for tumor tissue and sera. These results reflect on the methods and meaning of miRNA biomarkers for NPC in tissue and peripheral blood.