Identifying epigenetic biomarkers of established prognostic factors and survival in a clinical cohort of individuals with oropharyngeal cancer

Abstract

Background: Smoking status, alcohol consumption and HPV infection (acquired through sexual activity) are the predominant risk factors for oropharyngeal cancer and are thought to alter the prognosis of the disease. Here, we conducted single-site and differentially methylated region (DMR) epigenome-wide association studies (EWAS) of these factors, in addition to ∼ 3-year survival, using Illumina Methylation EPIC DNA methylation profiles from whole blood in 409 individuals as part of the Head and Neck 5000 (HN5000) study. Overlapping sites between each factor and survival were then assessed using two-step Mendelian randomization to assess whether methylation at these positions causally affected survival. Results: Using the MethylationEPIC array in an OPC dataset, we found novel CpG associations with smoking, alcohol consumption and~3-year survival. We found no CpG associations below our multiple testing threshold associated with HPV16 E6 serological response (used as a proxy for HPV infection). CpG site associations below our multiple-testing threshold (P Bonferroni < 0.05) for both a prognostic factor and survival were observed at four gene regions: SPEG (smoking), GFI1 (smoking), PPT2 (smoking) and KHDC3L (alcohol consumption). Evidence for a causal effect of DNA methylation on survival was only observed in the SPEG gene region (HR per SD increase in methylation score 1.28, 95% CI 1.14 to 1.43, P 2.12 × 10 −05). Conclusions: Part of the effect of smoking on survival in those with oropharyngeal cancer may be mediated by methylation at the SPEG gene locus. Replication in data from independent datasets and data from HN5000 with longer follow-up times is needed to confirm these findings.

Epigenetic biomarkers of ageing are predictive of mortality risk in a longitudinal clinical cohort of individuals diagnosed with oropharyngeal cancer

“…BMI, educational attainment, smoking and alcohol intake), which can result in recall bias or misreporting. We utilised a DNAm-derived measure of packyears of smoking in our sensitivity analysis but future studies could implement the use of other methylation scores to index these variables 48 49 . Fourth, there is a disparity in coverage between 450K and 850K Illumina platforms meaning that some of the CpGs included in Horvath’s and Hannum’s clock are missing.…”

Head and Neck Cancers Are Not Alike When Tarred with the Same Brush: An Epigenetic Perspective from the Cancerization Field to Prognosis

“…This model was used to predict all-cause mortality among OPSCC patients, but although it explained 16.5% of phenotypic variance, it showed a similar predictive accuracy to self-reported data [146]. Additionally, in the blood of OPSCC patients, an EWAS identified three new CpG sites and 40 differentially methylated regions associated with alcohol consumption [145]. The gene with the lowest p-value for both analyses was SLC7A11, with the methylation levels of one of its CpG decreasing 0.10% per unit increase in alcohol consumption.…”

“…In OPSCC patients, the blood of 106 never and 303 ever smokers were compared and revealed the differential methylation of 52 CpG sites [145]. Of these, 49 sites had lower methylation levels in smokers, including a CpG mapped to AHRR.…”

Identification and Validation of a Prognostic Signature Based on Methylation Profiles and Methylation-Driven Gene DAB2 as a Prognostic Biomarker in Differentiated Thyroid Carcinoma

“…For instance, Langdon et al identified a panel of DNA methylation biomarkers for early diagnosis of renal cell carcinoma. Besides, DNA methylation of SPEG located at Chr2:220325443-220326041 could potentially regulate oropharyngeal cancer survival [ 17 ]. Shen et al reported that hypermethylation of SSTR2 promoter could be used to predict prognosis of laryngeal squamous cell carcinoma in males [ 18 ].…”