In a recent study published in the Journal of the National Cancer Institute, researchers assessed a novel human papillomavirus (HPV) and host deoxyribonucleic acid (DNA) methylation score for adenocarcinoma in-situ (AIS) and cervical adenocarcinoma (ADC) screening.
HPV persistence and progression (Pap) testing significantly reduced squamous cell carcinoma (SCC) incidence in the 1960s. However, ADC incidence remained high due to the low sensitivity of Pap smears for AIS as well as ADC. These lesions, predominantly within the endocervical canal, are harder to visualize colposcopically or capture in a Pap specimen.
About the study
In the present case-control study, researchers evaluated the specificity and sensitivity of a combined human papillomavirus and human methylation score to screen for adenocarcinoma in situ and adenocarcinoma.
The methylation score was calculated by averaging the percent methylation at 35 different predetermined CpG sites within the L2/L1 open reading frames (ORFs) of HPV-16, 18, and 45, and the promoter regions of two human genetic loci, heparan sulfate glucosamine 3-O-sulfotransferase 2 (HS3ST2) and deleted in colorectal cancer (DCC) on chromosomes 16 and 18, respectively.
The team examined exfoliated cells from the cervicovaginal region of HPV PaP group females who showed HPV-16/18/45 positivity, including (i) 353 cervical intraepithelial neoplasia-3 (CIN-3) or squamous cell carcinoma patients, (ii) 176 AIS or ADC patients, (iii) controls with cleared HPV-16/18/45 infections (HPV-clearers; 579 individuals) or persistent infections (HPV-Persisters; 292 individuals).
Using next-generation methods, the researchers calculated CpG region-specific DNA methylation percentages. DNA methylation scores were calculated by calculating the mean methylation percentage for all the 35 CpG regions analyzed.
From electronic health records, deidentified race, ethnicity, smoking habits, and age-related data, as well as follow-up histopathological and cytological reports, were obtained, with clinical outcomes accessible until 2014.
Lesion identification at Kaiser Permanente Northern California (KPNC) was based on histopathological findings of colposcopically-directed samples. HPV genotyping was performed on banked specimens. Cases were chosen based on HPV-16, 18, and 45 genotyping data. DNA CpG site methylation testing was performed on a sample taken soon before the diagnosis of CIN3/SCC or AIS/ADC. The sequencing reads were mapped to human host and viral reference sequences, which were then bisulfite-converted in silico to determine the methylation status of CpG and non-CpG cytosines.
By calculating the ratio of methylated Cs in the total number of molecules sequenced per sample, the alignment result was converted into methylation percentages for each tested cytosine. The odds ratios (ORs) were computed using multivariable logistic regression modeling. Furthermore, generalized linear modeling (GLM) was used. Surveillance, Epidemiology, and End Results (SEER) population-based data were used to calculate the area under the curve (AUC) values.
Results
AIS/ADC tumors (55%) and CIN3/SCC tumors (54%) had a greater proportion of white females compared to those whose HPC infections cleared (44%). Seven to ten percent of the individuals were current smokers, and 10% to 15% were past cigarette users.
HPV-16 was most commonly detected in CIN3 and SCC cases (66%, 24%, and 10% infections by HPV-16, 18, and 45, respectively), whereas HPV-18 was most commonly detected in AIS and ADC cases (39%, 51%, and 10% infections by HPV-16, 18, and 45, respectively). HPV-16 was likewise more prevalent among individuals with persistent HPV infections (48%) compared to those whose infection cleared (32%).
Each CpG site showed a greater percentage of DNA methylation among exfoliated cells from the cervicovaginal region of AIS/ADC patients and CIN3/SCC patients compared to the control group. The ADC/AIS methylation Score demonstrated 74% sensitivity, 89% specificity, and 88% AUC values.
The comparable proportions for CIN3/SCC were 46%, 82%, and 68%, respectively. Using HPV-clearers as controls, the OR for the DNA methylation score (fourth versus first quartiles) for AIS and ADC was 49. CIN3 and SCC patients showed similar but weaker relationships with the DNA methylation scores.
The OR strengths were equivalent to a human gene or an HPV subtype. The odds ratios for each CpG site were all extremely significant. Similarly, all of the CpG region-specific odds ratios for CIN3 and SCC were very significant. Nevertheless, the CIN3 and SCC relationships were weaker compared to those for adenocarcinoma and adenocarcinoma in situ. The direct comparisons revealed that AIS and ADC were linked with higher levels of methylation at almost all CpG regions.
Individuals with persistent HPV infection showed higher DNA methylation levels than individuals with cleared HPV infection. AIS/ADC showed considerably greater methylation levels than controls for each HPV type and the two human HS3ST2 and DCC CpG sites evaluated. CIN3/SCC connections were found to be similar. Direct case-case comparisons by individual HPV type and both human genes, however, revealed that AIS/ADC had substantially greater methylation levels than CIN3/SCC at all but six of the 35 CpG sites investigated.
Implications
Overall, the study findings showed that HPV 16, 18, or 45-infected females with high methylation scores had high likelihoods of AIS/ADC, indicating the need for careful histopathological examination of the cervical transition zone. Regular screening techniques such as colposcopy and Pap smears have shown low sensitivity in detecting AIS and ADC cases. Compared to SCC, ADC has poorer survival rates, making the findings clinically relevant.
DNA methylation of specific L1/L2 CpG regions in HPV-16,18, and 45 and CpG regions in the promoter sites of the HS3ST2 and DCC genes might be critical molecular determinants for cervical tumors and related biomarkers. L2/L1 DNA methylation could be an intermediate step in cervical tumor formation but may not play biologically significant roles in HPV persistence.