A model for HPV superspreading revealed by long-read sequencing of a rare cervical cancer cell line with episomal and integrated HPV16

Nicole opened the talk with background details on the involvement of HPV in cervical cancer, which is a central focus of her and Michael’s lab based at the National Cancer Institute, USA. The World Health Organisation reports that 300,000 individuals die from cancer each year. Showing a graph displaying side-by-side the numbers of cervical cancer and COVID-19 deaths across the globe, Nicole explained that living in a resource-limited country is the most significant risk factor for cervical cancer mortality; cervical cancer is causing as many deaths in Africa and Asia as is COVID-19. To study cervical cancer in resource-limited settings, Nicole’s team have established a research cohort in Guatemala, and recruited over 700 women, collecting blood, tumour tissue, and clinical data.

The episomal and integrated HPV genome

Nicole introduced the structure of the 7.9 kbp HPV genome, which encodes two oncogenes, E6 and E7, that inhibit host tumour suppressor genes TP53 and RB1. The HPV genome replicates in the host cell nucleus as an episome, and its integration into the host cell genome produces genome instability. Nicole reflected on the Nanopore Community Meeting 2020 presentation in which she shared her findings that the cervical cancer CaSki cell line had complex arrays of both full length and truncated HPV16 genomes integrated at ~30-50 sites in the genomes of the tumour cell line. They termed this phenomenon “superspreading”. Some of these insertions were so long that the entire array could not be sequenced.

To follow up on these findings, her team have recently confirmed this superspreading phenomenon in other cell lines. They further discovered that the cervical cancer cell line SNU-1000 had both episomal and integrated HPV16. Applying PCR amplification of overlapping fragments of the HPV16 genome, they found that the entire genome was intact within this cell line. Using whole-genome sequencing (WGS), they discovered that there were also smaller integrated fragments of HPV in the host genome, in the intron of the CEP126 gene on chromosome 11, including a portion of the HPV E7 gene, although not long enough to encode any proteins. Besides this integration in CEP126, the integration locus on chromosome 11 was also amplified in the SNU-1000 cell line. This amplification involved a key oncogene and two other genes involved in the immune response. Nicole stated that ‘this is a novel mechanism of HPV activity in which the integrated fragment of HPV does not express the E6 or E7 oncogenes’ yet resulted in overamplification and expression of cellular oncogenes.

Displaying the nanopore WGS sequence data, Nicole pointed out the large insertions in the HPV episomal reads, some of which were full-length HPV genome sequences, meaning that there was more than one full-length genome present in some episomes. They coined these concatemeric genomes “multimer episomes”; the longest of which was over 77 kbp in length, others were a mix of deleted and scrambled HPV genomes.

Nicole’s team repeated the whole-genome analyses using the Rapid Sequencing Kit and protocol, combined with adaptive sampling to enrich for HPV16 sequences. With this approach, they obtained an ~8-fold enrichment of HPV16 reads. Similar to their previous findings, they observed multimeric molecules, 634-bp deleted genomes, and scrambled sequences. From these results, they developed a model of HPV16 episomal multimer formation.

HPV transcriptome and epigenetics

Michael introduced the HPV transcriptome and how ‘HPV gene expression is complex’, involving alternative promoters and extensive alternative splicing. Obtaining full-length transcripts has helped the team to better understand transcriptome regulation in cancer cells and tumours. Michael noted a particularly interesting splicing event that removes a portion of the E6 oncogene, perhaps to avoid immune recognition.

Transcriptome data from multiple cervical cancer cell lines has revealed wide variation in their levels of HPV expression, as well as variability in the extent of splicing, and Michael’s team are interested in how this expression regulation is occurring. Cell line transcriptome data has revealed that genes adjacent to HPV integration sites are activated, with at least one gene per cell line highly activated due to integration. It is also known that certain sites, in both episomal and integrated HPV16 genomes, can be methylated. ‘With nanopore direct sequencing we can visualise methylation’, and from full-length HPV16 genome reads, methylation can be assessed across entire HPV16 genomes, both episomal and integrated, including deleted and scrambled forms. Such data was ‘not possible to obtain previously’.

Furthermore, from only 5X to 10X depth of coverage of nanopore sequence data, Michael’s team have called structural variants (SVs) (using the cuteSV tool). Michael displayed an example of an SV they detected, a ~5 kb insertion on chromosome X in the SNU-703 cell line.

Explaining current work in progress, Michael outlined their work on HPV16-positive cervical tumour samples taken from their Guatemalan cohort, on which they have begun using barcoded nanopore sequencing with adaptive sampling, combined with transcriptome sequencing, to study these samples.