Human papillomavirus (HPV) can drive tumor development in some rare sinonasal squamous cell carcinomas (SNSCCs), according to a new comprehensive study of these tumors from investigators at the Johns Hopkins University School of Medicine and the Johns Hopkins Kimmel Cancer Center. The researchers also identified common mutations among these cancers and a potential combination treatment during the study, which was partially supported by the National Institutes of Health.
The origin of SNSCCs has been poorly described, says senior study author Nyall London Jr., M.D., Ph.D., an associate professor of otolaryngology-head and neck surgery at the Johns Hopkins University School of Medicine. According to a previous study, approximately 25% of SNSCCs are associated with HPV, which has been found at Johns Hopkins and other centers to be a driver of other head and neck cancers. However, it hasn't been clear whether HPV instigated tumor biology in SNSCCs or was merely a neutral bystander, just "hanging out," he says. "It's been hypothesized to be both."
In work published in the June 11 issue of Nature Communications, London and colleagues performed the first known comprehensive genome-wide characterization of HPV-associated and HPV-independent SNSCCs. There are only about 3 per million cases of SNSCCs per year, according to a prior study. However, because they originate near the eyes and behind the nose, where there is some space, masses tend to grow fairly large before they cause symptoms and are diagnosed, London says. The five-year survival rate is about 50%.
"Our findings provide important genomic insights and advance the understanding of these rare sinonasal tumors," says Fernando Zamuner, Ph.D., a faculty research associate in the Department of Otolaryngology-Head and Neck Surgery at Johns Hopkins and co-first author of the study.
During the study, the investigators performed high-throughput DNA sequencing on tissue samples from 56 patients with squamous cell cancers arising from the sinonasal cavity or nasolacrimal duct (tear duct), and some normal DNA as a comparator. Thirty-seven of the 56 were considered HPV-associated. The majority of HPV-associated tumors arose in the nasal cavity while the majority of HPV-independent tumors arose in the maxillary sinus, an air-filled cavity in the upper jaw bone. Patients with HPV-associated disease, on average, were younger at cancer presentation (60.8 years) than those with HPV-independent cancers (66.3 years).
Among samples from HPV-independent SNSCCs, the most frequently mutated cancer driver genes included TP53, NOTCH1 and KRAS, as well as CDKN2A, COL2A1, FAT4, FBXW7 and ROS1. Investigators compared this against a group of 12 additional HPV-independent SNSCCs for which matched normal DNA was not available, finding similar mutation frequencies in TP53, NOTCH1, CDKN2A, COL2A1 and FAT4.
However, investigators found a significantly different mutational profile for HPV-associated SNSCCs, where the most frequently mutated cancer driver genes included KMT2D, FGFR3, KMT2C, GOLGA5, TET1 and ARID1B. The researchers compared this against a group of 25 additional HPV-associated SNSCCs for which matched normal DNA was not available, finding similar mutation frequencies for KMT2D, FGFR3, KMT2C and ARID1B.
Next, the team looked for hotspot mutations, areas of DNA likely to be mutated. HPV-associated SNSCCs have been characterized by hotspot mutations at the E542K/E545K location in PIK3CA and S249C area in FGFR3.
Of the five HPV-associated SNSCCs in the study with PIK3CA missense mutations (alterations in DNA that result in a different amino acid incorporated into the structure of a protein), three tumors had E542K mutations and one had E545K mutations, while none of the HPV-independent SNSCCs demonstrated PIK3CA hotspot mutations. Of the six HPV-associated SNSCCs with FRGR3 missense mutations, four tumors had S249C hotspot mutations, while none of the HPV-independent SNSCCs demonstrated FGFR3 hotspot mutations. Additionally, recurrent mutations not previously observed in other tumors also were noted in HPV-associated SNSCCs, including KMT2C N729D, AP3S1 P158L, UBXN11, MT-ND4 and MT-ND5.
London and colleagues also correlated the presence of frequent mutations in the cancers with clinical outcomes. As previously seen among HPV-independent HNSCCs, the presence of a TP53 mutation correlated significantly with worse overall survival. Among HPV-associated SNSCCs, mutations in KMT2D and FGFR3 were associated with worse overall survival.
Further supporting that HPV-associated SNSCC is driven by HPV, the authors described additional defining characteristics of HPV-driven tumors, including enrichment in an APOBEC mutational signature (a specific pattern of DNA observed in HPV-driven tumors), the integration of viruses at known hotspots, and frequent mutations in epigenetic regulators, which are enzymes that control gene expression by adding or removing chemical "tags" on DNA or histones.
In other experiments, the team looked for chemical pathways that might be involved in tumor development, finding significant activity for HPV-associated SNSCCs in the PI3K pathway and YAP/TAZ pathways, and for HPV-independent SNSCCs in the PI3K, RAS and MYC pathways. Then, they generated an HPV-associated SNSCC cell line from patient samples, and found that administering drugs that block the PI3K and YAP/TAZ pathways, called alpelisib and verteporfin, respectively, synergistically reduced the ability of tumor cells to proliferate.
While the findings require further investigation in a larger study, says London, "The surprising thing is that we identified five new recurrent mutations that are specific for HPV-associated sinonasal squamous cell carcinoma that are not found in other tumors. We are now trying to figure out the biology behind those and their functional importance."
An additional study is looking at behavioral risk factors and some of the epidemiologic aspects of HPV in sinonasal tumors.
The work was supported in part by the NIH's Intramural Research Program, Center for Cancer Research and National Cancer Institute; and by a research grant from the Investigator-Initiated Studies Program of Merck Sharp & Dohme LLC to London. The other authors declared no competing interests.
