Nearly all cervical cancers are caused by high-risk alpha-papillomavirus (αHPV) infections. HPV infections can also lead to tumors in other areas of the genital tract or the oropharynx and often take decades to develop. The tumors become highly genetically unstable as they progress, and it is thought that much of this instability is caused by two HPV oncogenes, HPV E6 and E7. The viral E6 protein promotes the degradation of the tumor suppressor p53, while E7 inactivates pRB, a key cell cycle regulator, driving the cell into unregulated S-phase entry. Both p53 and pRB are necessary for pausing the cell cycle and repairing DNA damage, so oncogenes E6 and E7 have been surmised to disrupt DNA repair. Furthermore, E6 and E7 have been shown to interact with BRCA1, a key protein in double-stranded break (DSB) repair. However, the mechanism by which E6 and E7 disrupt DNA repair has not been established. In this paper from Dr. Denise Galloway’s Lab (Human Biology Division), lead author Dr. Nick Wallace (now faculty at Kansas State University) and colleagues demonstrate how E6 and E7 impair double-stranded DNA repair through multiple mechanisms.
The authors first showed that HPV16 E6 and E7 oncogenes disrupt DSB repair in human keratinocytes. HPV16 is the most common oncogenic HPV and the type used for this publication; hereafter, E6 and E7 will refer to HPV16 E6 and E7. The researchers expressed E6 and E7 genes simultaneously and separately and found that cells with either or both genes expressed had higher levels of phospho-H2AX (p-H2AX), a marker of DSBs. Furthermore, when they irradiated the cells to cause further DNA damage, the control cells repaired the DSBs far quicker than cells expressing E6 and/or E7. P-H2AX foci resolved in control cells in about 4 hours, while it took the cells expressing E6 and E7 up to 24 hours, and even then, p-H2AX remained.
DSB repair can occur in two ways. If a homologous sequence is present, homologous recombination (HR) can occur using the unbroken strand as a template. If no homologous sequence is present, non-homologous end joining is the alternative, where the strands are fused back together, leaving more room for errors. Using a U2OS cell line (a bone osteosarcoma cell line) modified with a DR-GFP reporter cassette that only expresses GFP if HR repair is functioning properly, the authors measured HR in cells expressing E6 and E7 and found it to be impaired, particularly when E6 was expressed. Expression of E6 alone impaired HR by 50%.
The Galloway Lab previously reported that in the lower risk β-HPV viruses, E6 disrupts HR by decreasing the expression of BRCA1 and BRCA2. The authors tested this possibility, but found that the abundance of HR proteins BRCA1, BRCA2, RAD51, and RPA70 were either not changed or increased in cells expressing either or both of the viral oncogenes. Upregulation of several of the HR proteins after E6 and E7 expression suggests that the HR pathway could be upregulated in patient HPV+ tumors. The authors analyzed mRNA expression of HR repair genes from 86 HPV+ cervical tumor tissue and 58 healthy control samples from the NCBI Gene Expression Omnibus (GEO) database. They found that mRNA expression levels of HR repair genes were significantly higher in the cervical malignancy tissue, confirming an upregulation of the HR pathway in HPV+ tumors.
Since expression of HR proteins was not impaired, the authors next investigated the kinetics of HR repair complex formation to determine if the complexes were active. HR repair complex formation was not impaired after E6 or E7 expression as observed by immunofluorescent (IF) microscopy. To better understand what was happening in these repair complexes, the authors looked at the kinetic relationship among the HR proteins. They compared the timing of recruitment of RPA70, RAD51, BRCA1, and BRCA2 to what is reported in the literature and did not find significant changes indicative of impairment of the repair complex when the HPV oncogenes were expressed.
If expression of HR proteins and repair complex formation were unimpaired, E6 and E7 must disrupt DSB repair by another mechanism. Homologous recombination is dependent upon the availability of a homologous sequence, and therefore can only be completed in S and G2 phases of the cell cycle. The authors hypothesized that expression of E6 and E7 was somehow causing repair foci to form during the wrong part of the cell cycle. They used IF microscopy to examine in which part of the cell cycle RPA foci were forming. Cyclin A was used as a marker of S and G2 phases, and cyclin E for G1. RPA foci formed far more often in cyclin A negative cells when E6 and E7 were expressed suggesting repair foci were forming in G1 when no homologous sequence was present. This was confirmed by flow cytometry.
While investigating the repair complex formation, the researchers noticed sustained RAD51 expression long after irradiation. This suggested RAD51 might be forming unproductive repair complexes by localizing away from DSBs. Using IF microscopy, the authors determined that 50-60% of RAD51 appeared to be localizing away from p-H2AX foci. The authors then created a single persistent DSB in U2OS cells and tracked RAD51 recruitment. In E6 expressing cells, RAD51 was recruited only 50% of the time to the DSB as compared to control cells. In E7 expressing cells, RAD51 was localized correctly in most cells, suggesting that this phenotype is driven by E6.
The viral genome is called the episome. For HPV, it is a circular double-stranded DNA sequence. Integration of this episome is a common rate-limiting step in HPV-associated malignancies. In order for the episome to be integrated into the host genome, both the host DNA and the episome must incur DSBs. The authors found that the viral episome was integrated into the host genome at a higher rate in cells expressing E6 than in control cells. This is likely due to the fact that DSBs were more prevalent in the cells expressing the oncogene and the DSBs provide a greater opportunity for episomal integration.
Graphic courtesy of Dr. Nick Wallace.
This paper provides a new mechanistic framework for HPV16 E6 and E7 oncogene disruption of DSB repair leading to increased genome instability in HPV+ malignancies (Figure 1). “This work was done exclusively in cells expressing oncogenes from HPV 16, but the implications are likely broader,” says first author Dr. Nicholas Wallace. “Most HPV-associated cancers are caused by one of two HPVs (HPV 16 and HPV 18). These HPVs are remarkably similar especially with regard to their dependence on DNA repair machinery to replicate their genomes. We speculate that the deregulation of homologous recombination described here would be shared between HPV 16 and 18... If we are correct, then the majority of HPV+ cancers will have a reduced ability to repair double stranded DNA breaks via the homologous recombination pathway.”
The work described in this paper demonstrates an important therapeutic potential. As Nicholas Wallace explains, “Because breast and ovarian tumors with defects in [the homologous repair] pathway are acutely sensitive to a class of chemotherapeutics known as PARP inhibitors, we are hopeful that our work will support the use of these drugs against HPV-associated cancers.”
Wallace NA,Khanal S,Robinson KL,Wendel SO,Messer JJ,Galloway DA. 2017. High Risk Alpha Papillomavirus Oncogenes Impair the Homologous Recombination Pathway. J Virol. 91(20) e01084-17.
Funding for this work was provided by the National Cancer Institute and the National Institute of General Medical Sciences.