Researchers Reveal New Subtype of Cervical Cancer

Carolyn E. Banister, MD

A new subtype of cervical cancer has been discovered, according to the results of a study published in Oncotarget.

Like most cervical cancers, this novel subtype is also triggered by HPV, but the virus does not direct the growth of the cancer. Thus, therapies used to target these tumors’ distinct genomic pathways may improve outcomes for patients over standard treatment.

An analysis of data from 255 cervical cancer samples in The Cancer Genome Atlas (TCGA) showed that 2 HPV oncogenes that are thought to be integral to cervical cancer growth were expressed at either high levels (HPV-active disease) or low levels (HPV-inactive disease).

"Physicians managing cervical cancer patients should test for HPV oncogene expression in these tumors and consider personalized treatment depending on HPV activity," lead study author Carolyn E. Banister, MD, assistant professor, College of Pharmacy, University of South Carolina, said in a press release.

OncLive: Please provide some background for studying this new subtype of cervical cancer.

How was your study conducted?

In an interview with OncLive, Banister highlighted the major findings of her study and the significant therapeutic implications that will follow. She also sheds light on some of the other remaining challenges in the treatment of cervical cancer.Banister: It is commonly known that all cervical cancers contain HPV DNA, and we thought they were all the same. But we looked into the TCGA database and discovered that there is a subset of cervical cancers that, although they contain HPV DNA in their genomes, they don’t express the HPV oncogene, and we found that this has therapeutic implications.We used publicly available TCGA data. This was part of a very large, NIH-funded project where large sequencing centers performed lots of analyses on many different types of tumors, and they made the information publicly available so that researchers could go and mine the data for interesting things that, as individual labs, we would not be able to afford to do—things like whole-exome sequencing, DNA methylation analysis, and they did a gene expression analysis.

And what did your findings show?

We used all that data, and combinatorial analyses, to look at cervical cancers. We had a little bit of a hint that HPV silencing may be going on, because some of our coauthors had discovered in head and neck cancer that this subset exists. So we were looking to see if this was also true in cervical cancer.We found a subset of cervical tumors that do not express important HPV genes. These are oncogenes E6 and E7, which are the most important oncogenes, and they’re the ones that drive a lot of cervical cancers. We found that tumors that do not express HPV have a different methylation profile and a different mutational profile, and that comes back around full circle to show that these patients could possibly be treated differently.

Are there trials out there looking at the use of immune checkpoint inhibitor therapy in cervical cancer?

The tumors that do express HPV show increased expression of the inflammatory response genes compared with the tumors that don’t express HPV. This indicates an ongoing stimulant to recruit the immune response to the HPV-active tumors. The HPV-active tumors also express CTLA-4, PD-L1, and TIGIT. So taken together, we think that immune checkpoint inhibitor therapy would work better on the HPV-active than the HPV-inactive tumors.Currently, there are ongoing immune checkpoint inhibitor trials in cervical cancer. There’s some with anti¬—PD-1, anti–PD-L1, and anti–CTLA-4. It would be interesting for these trials to be retrospectively analyzed for the association of response with the expression of the HPV E6 and E7 oncogenes.

One of the cool and very surprising things we found was that about half of the HPV-inactive tumors had mutations in the TP53 gene. Many cervical cancers also demonstrate de novo resistance to cisplatin therapy, and we know that cisplatin kills in a TP53-dependent manner. Therefore, the tumors with the TP53 mutations may be resistant to the DNA-damaging agents, like cisplatin. So it would be important for physicians to test for the TP53 mutation, which would almost exclusively be found in an HPV-inactive tumor, and look for associations with cisplatin resistance. So that is one of the ongoing things that we may be looking into: correlating outcomes with whether or not their tumors are active or inactive.

We also found some good news for patients with HPV-inactive tumors. We found some gene expression evidence that the Src tyrosine kinase inhibitor dasatinib (Sprycel) may work better in those tumors. We also found mutations in the PI3K signaling pathway, which would suggest that the AKT inhibitors might work better in the HPV-inactive tumors.

Is this study the first of its kind to find these correlations?

What is the ultimate takeaway message from this study for community oncologists?

Looking at the overall treatment landscape of cervical cancer, what are the biggest challenges right now?

So there’s a lot of room here for retrospective studies because there are a lot of tumors that we have ongoing clinical data on. We can go back through those samples and look to see if expression of HPV is going to be correlated with what we predict from this database analysis.This is the very first paper out on this phenomenon in cervical cancer. We’re relating the expression profiles that we had with what others have reported with other cancers, etc. that have these 2 different groups. We are hopeful that checkpoint inhibitor trials would be more effective in the HPV-active group.The take-home message is that testing for HPV activity, the E6 and E7 gene expression, may be helpful in designing therapy for patients with cervical cancer.We do know that there’s a difference in incidence and mortality rates between ethnicities, specifically between Caucasians and African American women. African American women tend to get cancer earlier, and they tend to die at a younger age from cervical cancer. For many years, it was thought that a socioeconomic driver accounted for this difference.

I published some previous work on a group of college-age women at the University of South Carolina and showed that, given the same standard of care, and no difference between age of sexual onset and number of partners, that African American women, on average, kept the HPV virus twice as long as European American women. Since they have the exact same standard of care, this really points to the fact that there may be a biological difference. Since this is mediated by a virus, this may be something that has to do with recognition that the virus is there and the ability to clear it. I think that some of the next steps are going to be looking into these genetic differences of what could be causing this difference in disease progression.

Banister CE, Liu C, Pirisi L, et al. Identification and characterization of HPV-independent cervical cancers [published online January 6, 2017]. Oncotarget. doi:10.18632/oncotarget.14533.