Lisa Mirabello: Using Genetic Data to Understand Osteosarcoma Etiology and HPV Carcinogenicity

June 27, 2017, by Victoria A. Fisher, M.P.H.

Tenure-track investigator Lisa Mirabello, Ph.D., M.S., studies genomic and epigenomic alterations as they relate to cancer etiology. She has developed a robust scientific program focused on genetic susceptibility to osteosarcoma, a rare bone cancer, and the genomics of carcinogenic human papillomaviruses (HPV), through highly collaborative projects.

Etiology of Osteosarcoma

Osteosarcoma is the most commonly diagnosed primary malignancy of bone in children and young adults. However, its etiology is poorly characterized. Although cure rates are fairly good for disease caught early, metastatic disease is usually fatal. “We are conducting germline genetic studies of osteosarcoma to better understand the role that genetic variation plays in risk and patient outcomes, and to identify new genes or genomic regions that may be important in osteosarcoma pathogenesis,” Dr. Mirabello said.

Alongside Sharon A. Savage, M.D., Chief of the Clinical Genetics Branch, Stephen J. Chanock, M.D., Division Director, and colleagues, Dr. Mirabello played a key role in the first international, multi-institution genome-wide association study (GWAS) of osteosarcoma. The team also conducted large epidemiologic and meta-analytic studies that clarified osteosarcoma incidence patterns and risk groups. Now, Dr. Mirabello is leading the assembly of additional patients to extend the sample size for a second GWAS. “International collaborations are essential so we can collect enough cases of this rare cancer,” she said. “I’m always working to establish new collaborations to accomplish this important work.”

In addition, Dr. Mirabello led the first GWAS study of osteosarcoma metastasis and identified a single nucleotide polymorphism (SNP) in the NFIB gene that was significantly associated with the presence of metastasis at diagnosis. “The NFIB manuscript is one of my favorite papers, because the data fit together to tell a compelling story,” Dr. Mirabello said. “First, we conducted the GWAS agnostically looking for markers of osteosarcoma metastasis. We found a SNP in NFIB that was strongly associated with metastatic disease at diagnosis in several different ethnic groups. Further, our functional studies showed that the risk SNP was associated with a decrease in NFIB expression, which led to increased osteosarcoma cell migration, proliferation, and colony formation. Together, these findings suggest that NFIB is an osteosarcoma metastasis susceptibility gene.”

Dr. Mirabello is expanding her genomic studies beyond osteosarcoma to identify genetic variants associated with the risk of developing other pediatric solid tumors, including rhabdomyosarcoma and medulloblastoma. As a next step, she is exploring whether this same variation in NFIB is associated with metastasis for other pediatric cancers. She will also look for additional common variants associated with susceptibility and outcomes.

The GWAS data only tell part of the story. Dr. Mirabello and colleagues are also looking at the role of rare variants in osteosarcoma susceptibility. In a recent study, they showed that inherited variations in a known tumor suppressor gene (TP53) among children and adolescents with osteosarcoma are more common than previously thought. In fact, about 1 percent of the osteosarcoma patients in the study had a pathogenic or possibly pathogenic mutation in TP53, compared to the 3 percent expected carrier rate.

To expand the search for rare variants, Dr. Mirabello is using whole-exome sequencing of osteosarcoma patients to look for mutations in other known cancer-causing genes. In addition, she is using the exome data to look for novel genes with an excess of rare variants in the cases. “It is important to determine the extent to which rare germline mutations contribute to osteosarcoma etiology, and to distinguish their effects from those observed in association with more common, lower penetrance SNPs.”

Dr. Mirabello is also expanding upon her previous finding that tall stature is associated with an increased risk of osteosarcoma by leading a study to evaluate genetic risk scores for height and risk of osteosarcoma. “Tall height is one of the strongest risk factors for osteosarcoma,” she said. “We’re trying to determine if genetically determined height is what’s related to risk of osteosarcoma, versus other environmental factors that affect height.”

HPV Carcinogenicity

There are over 200 types of HPV. For reasons largely unknown, only about a dozen HPV types cause virtually all cases of cervical cancer, and a large proportion of other anogenital and oropharyngeal cancers.

“We know a lot about the natural history of an HPV infection and the multiple stages of cervical carcinogenesis,” Dr. Mirabello said. “DCEG scientists have an outstanding record of studying HPV natural history and its translation to screening/vaccination strategies for over 30 years. But, we do not know why HPV types, and closely related variant lineages within a type, profoundly differ in risk of cancer, or why common and typically benign carcinogenic HPV infections only sometimes cause cancer. This indicates that additional risk factors are important for carcinogenesis.”

Researchers know that viral genetic factors are important because HPV types and variant lineages within a type, which differ greatly in carcinogenicity, are defined by differences in their nucleotide sequence. In 2010, Mark Schiffman, M.D., M.P.H., had initially proposed the HPV Genomics Project to compare genetically related HPV genomes and to determine which SNPs or lineages make an HPV type more or less carcinogenic. However, the project never moved forward because even with the relatively simple HPV genome, the only available HPV sequencing methods were slow, labor intensive, and costly.

About three years ago, Dr. Mirabello re-initiated the project after Michael Cullen, Ph.D., and Joseph Boland, M.S., in the Cancer Genomics Research Laboratory, successfully developed a high-throughput assay to whole-genome sequence HPV 16. “Using what we know about human next-generation sequencing, we were able to design and optimize the assay for HPV genome sequencing,” she said. “It had never been done before, so it was kind of like inventing the wheel.”

Dr. Mirabello and team spent the first year designing a sequence analysis pipeline and validating this novel approach. In an 800-sample pilot study, they demonstrated excellent concordance with the gold standard of sequencing and uncovered several novel observations with these next-generation sequence-based data. They were able to annotate numerous novel HPV 16 SNPs that may impact infection outcome, large deleted regions, and HPV 16 variant lineage co-infections.

The assay has now been extended to genome sequence all of the high-risk HPV types. To date, the team has sequenced over 10,000 high-risk HPV whole-genomes. “This high-throughput approach enables the practical sequencing of thousands of HPV-containing specimens from informative populations, and signals a new era in the study of HPV and related cancers,” Dr. Mirabello said.

HPV 16, which causes approximately half of all cervical cancers worldwide, can be grouped into 4 main lineages and at least 10 sublineages. Recently, Dr. Mirabello and colleagues conducted the largest case-control study of HPV 16 variant lineage risk of cervical precancer and cancer and found that different sublineages have very strong and specific risks. Grouping outcomes and lineages together, as previously done, conceals important qualitative heterogeneity in both viral lineages and disease outcomes. Certain lineages were specifically associated with adenocarcinomas with relative risks over 100, and these glandular lesions are very hard to detect with routine screening. “These findings suggest that a finer level of grouping of HPV 16 variant lineage classifications and disease outcomes is warranted,” Dr. Mirabello said. “A highly specific HPV variant test could conceivably have clinical and prevention implications for adenocarcinomas.”

Dr. Mirabello’s latest work (in press at Cell) revealed that HPV16 has tremendous variability below even the level of variant lineages. Within the HPV16 lineages there are thousands of viral isolates, each with a unique genome. Very few women who tested positive for HPV16 were infected with the same isolate. In contrast, the paired HPV16 genomes within a woman at multiple HPV16-infected body sites (oral, anal, and vulvar) were nearly all the same. “This suggests there is only high isolate diversity across different women,” Dr. Mirabello said.

Importantly, she also discovered that women with benign infections had more amino acid changing variants throughout the HPV16 genome. In particular, the HPV16 oncogene, E7, but not the other major oncogene, E6, was strikingly devoid of these variants in cancers from around the world compared to higher levels in the controls. HPV16’s unique risk might be due in part to a particular structure of the E7 protein.

This research is contributing to a paradigm shift in the way scientists think about HPV16. “Historically, the dogma has been that HPV16s are all the same,” Dr. Mirabello said. “We’re beginning to realize that there’s a lot more to the story. You have HPV16, but you also have a specific lineage, and even further you have a specific isolate. This urges a re-evaluation of HPV natural history and the clinical implications of HPV mutation rates, transmission, clearance, and persistence.”

Dr. Mirabello has expanded this approach to look at other high-risk types of HPV. “The ultimate plan is to look across the HPV types and see if we can find something shared that’s associated with progression or clearance,” Dr. Mirabello said. “Either one would be important for understanding viral carcinogensis.”