Salivary micro RNAs as biomarkers for oropharyngeal cancer

Abstract

Background: Despite the rising incidence, particularly of the human papillomavirus (HPV)-associated fraction of oropharyngeal cancer (OPC), there are no early detection methods for OPC. Considering the close association between saliva and head and neck cancers, this study was designed to investigate salivary micro RNA (miRNAs) associated with OPC, especially focusing on HPV-positive OPC.

Methods: Saliva was collected from OPC patients at diagnosis and patients were clinically followed up ≤5 years. Salivary small RNA isolated from HPV-positive OPC patients (N = 6), and HPV-positive (N = 4) and negative controls (N = 6) were analysed by next-generation sequencing to identify dysregulated miRNAs. Discovered miRNAs were validated by quantitative PCR using two different assays in a separate cohort of patients (OPC = 91, controls = 92). The relative expression was calculated considering SNORD-96A as the normalizer. Candidate miRNAs with diagnostic and prognostic potential were evaluated by generalized logistic regression.

Results: A panel consisting of nine miRNAs was identified to have the best diagnostic performance to discriminate HPV-positive OPC from HPV-positive controls (AUC- validation-1 = 94.8%, validation-2 = 98%). Further, a panel consisting of six miRNAs were identified to discriminate OPC from controls regardless of the HPV status (AUC- validation-1 = 77.2%, validation-2 = 86.7%). In addition, the downregulation of hsa-miR-7-5p was significantly associated with poor overall survival of OPC patients (HR = 0.638). A panel consisting of nine miRNAs were identified for the prediction of the overall survival of the OPC patients (log-rank test-p = 0.0008).

Conclusion: This study highlights that salivary miRNAs can play an essential role in the detection and prognostication of OPC.

Keywords: biomarkers; human papillomavirus; micro RNA; oropharyngeal cancer; saliva.

FIGURE 1

Volcano plot and heat map showing differentially expressed salivary miRNAs (FDR <0.05, logFC >1.5) discovered by small RNA sequencing. (A and B) Control human papillomavirus [HPV]‐positive versus oropharyngeal cancer (OPC) (HPV positive), (C and D) Control HPV‐negative versus OPC (HPV positive), (E and F) Control HPV‐negative versus control HPV‐positive pairwise comparisons.

FIGURE 2

HPV‐positive OPC versus HPV‐positive controls; performance evaluation of miRNA diagnostic panel by ROC analysis and active parameter estimates considered in the panel. (A) qPCR validation 1 by miScript™ primer PCR assays. (B) qPCR validation 2 by miRCURY™ LNA miRNA PCR assays.

FIGURE 3

Human papillomavirus (HPV)‐negative oropharyngeal cancer (OPC) versus HPV‐negative controls; performance evaluation of miRNA diagnostic panel by receiver operating characteristic (ROC) analysis and active parameter estimates considered in the panel, qPCR validation 1 by miScript™ primer PCR assays.

FIGURE 4

Oropharyngeal cancer (Human papillomavirus [HPV] positive and negative) versus controls (HPV positive and negative); performance evaluation of miRNA diagnostic panel by receiver operating characteristic analysis and active parameter estimates considered in the panel. (A) qPCR validation 1 by miScript™ primer PCR assays. (B) qPCR validation 2 by miRCURY™ LNA miRNA PCR assays.

FIGURE 5

Kaplan–Meier estimate of oropharyngeal cancer (OPC) patient survival by salivary expression of a miRNA panel consisting of Hsa‐miR‐194‐5p, Hsa‐miR‐449a, Hsa‐miR‐199a‐5p, Hsa‐miR‐3614‐5p, Hsa‐miR‐07‐5p, Hsa‐miR‐3529‐3p, Hsa‐miR‐99A‐3p, Hsa‐miR‐501‐3p, Hsa‐miR‐1290, Hsa‐miR‐548K and Hsa‐miR‐1246. Median split cut‐off of linear predictor, cut‐off = 2.787.