by ecancer reporter Clare Sansom

Most cases of the two most common types of cervical cancer – squamous cell carcinoma and adenocarcinoma – arise as a result of infection with the human papilloma virus, HPV.

This virus has many subtypes, at least 15 of which are considered high risk for cervical cancer development, but the most common and most virulent are types 16 and 18, which together cause about 70% of all cervical cancers.

However, HPV infection is very common in sexually active women and most infections are cleared rapidly and spontaneously by the immune system leaving no long-term effects.

There must therefore be other risk factors that determine which cases of HPV infection are cleared and which persist and allow cancer to develop.

Integration of viral DNA into the host genome is considered to be one of the most important risk factors, and it has been suggested that the level of viral integration might be a useful marker for cancer progression. 

Identifying the locations in the genome where this integration occurs and understanding the mechanisms of this process are therefore likely to yield insights into HPV-driven cervical carcinogenesis.

A number of these integration points are already known; the first, between KLF5 and KLF12 on chromosome 13, was identified as long ago as 1987 in SiHa cells.

Since then, a total of 417 host-viral DNA junctions (HPV breakpoints) have been located in approximately 389 human genes, but most of these were in small samples.

A group of researchers led by Ding Ma of Huazhong University of Science and Technology, Hubei, China has now conducted a comprehensive study of HPV integration into the human genome using their own technique, known as high-throughput viral integration detection (HIVID).

Firstly, they compared HIVID with whole-genome sequencing (WGS) on two cervical carcinomas and two HPV-positive cell lines and proved that their technique was a significantly more sensitive method for locating these breakpoints.

They then used HIVID to identify HPV breakpoints in a total of 104 cervical cancers, 26 pre-cancerous cervical intraepithelial neoplasias (CINs) and five cell lines; all were validated using Sanger and RNA sequencing.

A total of 3,666 HPV integration breakpoints were identified, distributed throughout the human genome.

Nine genes with integration sites in at least five of the 135 samples and 33 with sites in at least four of the samples were identified and termed viral integration hot spots.

A total of 1,546 breakpoints (42%) identified were located within genes, many of which encoded proteins in cancer-associated pathways; most of the others were located within 500 kb of a gene.

This suggested to Ma and his co-workers that HPV is integrated within the genome initially at random and that integration recurs at some loci because it provides a selective advantage to host cells.

They also annotated breakpoints on the genome of HPV16, one of the two most carcinogenic types, and discovered that these occur in all parts of the viral genome, more often in gene E1 than in E2 and less often in the viral promoter.

The researchers then investigated the correlation between number of viral integrations and the type and stage of the cervical cancers represented.

They found that both the percentage of lesions containing integrations and the number of integrations per genome increased in the progression from CIN to carcinoma; integration numbers were also significantly higher in squamous cell carcinomas than in adenocarcinomas.

This data suggested that integrations mainly occur in the initiating stage of carcinogenesis, and that integration rate and number can be used to predict disease progression.

Integration hot spots were confirmed in several genes where they had previously been reported: POU5F1B (in 9.7% of samples); FHIT (8.7%); KLF12 (7.8%); KLF5 (6.8%); LRP1B (5.8%) and LEPREL1 (4.9%).

More importantly, hot spots were discovered in three further genes: HMGA2 (in 7.8% of samples), DLG2 (4.9%) and SEMA3D (also 4.9%).

The effect of viral DNA integration into these genes differed according to the location of the breakpoint in the gene structure.

Specifically, integration into the introns of FHIT and LRP1B led to decreases in gene expression, whereas integration into the flanking region of HMGA2 and of MYC (near POU5F1B) led to increases in gene expression.

All these changes in expression may increase the risk of cancer developing.

Finally, the researchers observed that very short regions of sequence similarity between the human and viral genomes – microhomologies – were often found at or near breakpoints, suggesting that viral DNA integration may occur via microhomology-based DNA repair pathways.

In conclusion, Ma and his co-workers suggested that the identification of HPV integration breakpoints using HIVID might provide a useful tool for screening for cervical cancer and for assessing prognosis and selecting treatment options for patients who have already developed the disease.

Reference

Hu et al., (2015). Genome-wide profiling of HPV integration in cervical cancer identifies clustered genomic hot spots and a potential microhomology-mediated integration mechanism. Nature Genetics, published online ahead of print 12 January 2015.