The rationale for repurposing cidofovir to selectively target HPV positive and negative epithelial neoplasias

Human papillomaviruses (HPVs) are double-stranded DNA viruses of which the mucosal alpha-types can induce benign (low-risk HPV types) and malignant (high-risk HPV types) lesions. High-risk HPVs are associated with cervical cancer, other anogenital cancers and a substantial amount of head and neck squamous cell carcinoma (HNSCC). Effective prophylactic vaccines are available but are mostly administered to young women and older women and men remain unvaccinated, although male HPV vaccination programs have been implicated in several countries. Current treatment is based on surgery, radio –and/or chemotherapy, which can be associated with significant side effects. Furthermore, high recurrence rates are described in cases of advanced HPV disease. Thus, the development of a selective therapy capable of inducing regression of the existing HPV lesions, preventing the progression of low-grade disease, and/or preventing recurrences is highly required.

Cidofovir (CDV) is an acyclic nucleoside phosphonate with antiviral activity against a broad range of DNA viruses. CDV proved to have antiproliferative effects against several HPV+ and HPV- cell lines in vitro and in a murine xenograft model of HPV+ tumors and in a cotton tail rabbit PV model. CDV has been demonstrated to be active in different severe clinical presentations of both benign and malignant HPV-associated diseases, including recurrent respiratory papillomatosis, gingival and oral neoplasias, verruca vulgaris, plantar warts, severe laryngeal papillomatosis, anogenital lesions, hypopharyngeal/esophageal tumors, anal intraepithelial neoplasia grade 1-3, vulvar intraepithelial neoplasia grade 3, and cervical intraepithelial neoplasia stages 2+ and 3. CDV also demonstrated activity in vivo against HPV- malignancies such as glioblastoma, hemangiosarcoma and nasopharyngeal carcinoma.

First, we aimed to further unravel the selective mode of action of CDV against both HPV+ and HPV- transformed cells compared to normal cells. Previously, our laboratory performed whole genome gene expression profiling by means of microarrays in two HPV+ cervical carcinoma cells, HPV- immortalized keratinocytes, and normal keratinocytes. The data suggested activation of cell cycle regulation and double-strand break (DSB) repair mechanisms (‘ATM signaling’ and ‘DSB repair by homologous recombination’) following CDV treatment of normal cells in contrast to HPV+ and HPV- transformed cells. Based on these findings, we then compared HPV+ and HPV- cervical cancer and HNSCC cell lines with immortalized keratinocytes (HaCaT) and normal cells (primary human keratinocytes, primary epithelial tonsil cells and human embryonic lung fibroblasts) for their sensitivity to CDV, drug metabolism, incorporation of the active metabolite into cellular DNA and DNA damage induction (Chapter 2). Tumor cells proved to be more sensitive to CDV antiproliferative effects and incorporated more CDV-diphosphate than normal cells. To confirm the induction of DSBs by CDV, ATM activation was evaluated in the different cell types. Upon CDV treatment, the different cell types presented an increase in phospho-ATM levels. Detection of γ-H2AX, a quantitative marker of DSBs, showed a significant increase in DSBs in tumor cells but not in normal cells following CDV exposure. CDV antiproliferative effects correlated with incorporation of the active CDV metabolite into cellular DNA, which was related to induction of DNA damage. However, CDV induced DNA damage and antiproliferative effects did not correlate, indicating that the mode of action of CDV against cancer cells cannot be attributed solely to DNA damage. The effects of CDV on cell growth, incorporation into DNA and induction of DNA damage were independent of tumor type (i.e. originating from cervical or head and neck tissues) and of HPV status. Based on these results, proteins downstream of ATM in the homologous recombination repair pathway (i.e.p16, BRCA1, BRCA2 and RAD51) were investigated (Chapter 3). HPV+ and HPV- tumor cells accumulated CDV induced DSBs as indicated by an accumulation of RAD51 foci while normal cells resolved RAD51 foci, highlighting their capacity to repair CDV-induced DSBs. RAD51 foci formation in tumor cells was not linked to p16 overexpression, which was not altered by CDV treatment. In addition, no difference in accumulation of DSBs was observed between HPV+ and HPV- cell lines. These results indicate that CDV in vitro antitumor activity is not restricted to HPV+ tumor cells but could be expanded to HPV- cervical carcinomas and HNSCC tumors, although further studies including different tumor cell lines with diverse genetic backgrounds are mandatory before CDV can be proposed as a broad antitumor agent.

The second part of this PhD project aimed to investigate tumor-tumor interactions using a double subcutaneous xenograft model in athymic nude mice where the two tumors had no direct contact. We also explored whether intratumoral treatment with CDV would not only have a local antitumor effect, as previously demonstrated by our research group, but also a far-reaching effect leading to retarded growth of a secondary tumor (Chapter 4). Clinical evidence for tumor-tumor interactions within one organism exist for several malignancies but not for cervical cancer. We provided evidence that the presence and the site of a primary HPV+ xenograft had no influence on the growth of a secondary distant HPV+ xenograft. However, CDV in situ treatment of the primary xenograft had a temporary far-reaching antitumor effect, leading to retarded growth of the challenged tumor. In addition, volume reduction of the primary xenograft lowered the mortality and pathology associated with total tumor burden, as indicated by a reduction in splenomegaly, inflammation and an increase in antitumor immune responses. The effects of CDV on a secondary untreated HPV+ xenograft is likely mediated by an indirect activation of an antitumor immune response and inhibition of inflammation. Since the far-reaching effect of CDV was transient, we investigated the use of apoptotic tumor cells as a source of a wide variety of tumor antigens able to induce a more integral immune response, and co-administration of CDV together with vaccine adjuvants (aluminum hydroxyde + monophosphoryl lipid A) in an attempt to boost CDV far-reaching effects (Chapter 4). Apoptotic tumor cells were unable to delay the growth of a secondary xenograft implanted subcutaneously. Adjuvant per se slightly reduced the growth of both the first and the second xenograft but this effect was not synergistic with CDV treatment of the primary xenograft.

In conclusion, this doctoral thesis contributes to a better insight into the selective mechanism of CDV antitumor activity, which is crucial for the development of new strategies to treat HPV-associated neoplasias. Cidofovir is not only able to decrease the volume of a locally treated xenograft but also to reduce inflammation and to direct the immune response in such a way to be favorable to the host. These are important findings in view of the accumulating evidence showing that the immune system has a critical role in the process of tumorigenesis. Our work sets up the basis for further investigations on CDV as an inducer of DNA damage against both HPV+ and HPV- tumors and the potential of combination therapy and warrants the development of improved pharmacological formulations of CDV.