Study Reclassifies HPV-Positive Head and Neck Cancer

Iain Morgan, PhD, a member of the Cancer Molecular Genetics research program at VCU Massey Cancer Center

Iain Morgan, PhD

There are now 3 main head and neck cancer tumor types in relation to HPV16 genome status—integrated only, viral episomal only, and viral-human hybrid episomes—according to a report published in Oncotarget. The results suggest that the viral-human hybrid episomes have been previously miscategorized.

Additionally, 73% of HPV16-positive head and neck tumors have episomal versions of the viral genome that are replicating in an E1- and E2-dependent manor, possibly making the direct targeting of E1 and E2 an advantage.

OncLive: Can you please discuss the rationale behind conducting this research?

In an interview with OncLive, co-lead author Iain Morgan, PhD, a member of the Cancer Molecular Genetics research program at VCU Massey Cancer Center, discusses research that redefines HPV-positive head and neck cancer and may also identify new therapeutic targets.Morgan: In cervical cancer, it’s been quite clear that the integration of the lateral genome into the host results in a worse outcome for the patient. In head and neck cancer, there was very little research done on this. For example, if you look at the literature, there’s been an acceleration of research and interest in this field over the last 5 years. Prior to that, there was only a low number of publications but in the last 5 to 10 years, that has increased.

We conducted this research because we realized that HPV-related tumors weren’t well characterized. We received permission from The Cancer Genome Atlas to download the tumor data. One thing we were interested in was what the differences are in a host gene expression with HPV-positive and -negative to determine pathways that might be significantly changed in positive versus negative. That's something that other people are in the process of doing, as well.

Also, it did offer the unique opportunity to look at the viral genomes in a different way. There are several ways clinicians would look at this in the past without full sequencing data. You can look at the presence of viral-human hybrid fragments with a variety of techniques. It was previously concluded that the viral genome was integrated if you could detect this but in a large percentage of these tumors, the virus replicates as a circle with human DNA attached. For example, the technique that looked for virus-human hybrids would have classified that tumor as being integrated when, in fact, as far as we can see, it's not.

Looking at the data, we can separate the tumors into 3 categories. The first, where we think it's integrated; the second, where we think it remains as a circle; and the third, where we think it remains as a circle with human DNA. Looking at the outcomes, the patients who are truly integrated do die earlier.

The reason this has been missed in the past is because people have mixed up the tumors that have a virus in the human sequence replicating it as an episomal circle.

What are the next steps planned for this research?

However, more work needs to be done to prove this research.One of the questions that we're asking ourselves as we look at the literature is what is a simple test that we can do to tell us what is an integrated tumor since they have worse outcomes.

For the other 2 tumor categories, the outcomes for these patients tend to be the same, which needs to be investigated further.

To determine if a virus integrates, it’s important to look for E2 or RNA. Usually when the virus integrates, it loses its target of downstream viral genes, meaning E2, E4, and E5, are not expressed in a truly integrated tumor. The RNA should not be there, even if the DNA is there.

What we've developed in the lab is a fluorescent RNA probe against E2 to detect E2 RNAs in that approach. At VCU, we’re going to go back retrospectively in tumors and do that RNA probe to look for the ones that don’t give us a signal. You can check with one of the viral oncogenes, like E6 and E7, to make sure that the virus is being expressed.

We’re looking for tumors that are E6-positive or E7-positive. We are predicting that patients with E2-negative at an RNA level are going to have worse outcomes.

This research is important because there is a huge number of clinical trials now focused on de-escalation therapy. Patients who have HPV-positive tumors have better outcomes with chemoradiation therapy but it tends to be in the younger patients. The radiation is damaging, leaving patients with no saliva glands and dental problems. This is leading clinicians to want to give these patients a lower dose of chemoradiation if they respond well, which seems to work for most patients

However, not every single person who has an HPV-positive tumor has a good outcome. We’re predicting that the patients with tumors that are truly integrated are not going to respond well to chemoradiation and therefore should not join de-escalation trials, since that runs the risk of letting the disease continue to progress. Right now, the recommended treatment for patients with HPV-positive or -negative is the same.

Can you discuss how these results will affect treatment for head and neck cancer?

Additionally, there are ideas that we can stratify patients and de-escalate the therapy to treat them differently, but at the moment I would say that’s not what happens in the clinical trials. Maybe the clinicians get a few and treat them slightly less aggressively, but the standard of care is no different if you are HPV-positive or -negative.I think, in the end, we will be stratifying these patients. For example, if I went to the clinic with a tumor, I would not want to be part of a de-escalation trial. I would prefer to receive the full treatment, even if I will live with a dry mouth for the rest of my life. This category of patients does not respond well, even if they are HPV-positive.

However, integration is not a total answer. The nodal stage, the metastatic stage, and the tumor size stage, also help to predict the outcome to treatment for patients who are HPV-negative, but it has been difficult for patients with HPV-positive tumors. It’s a whole new genre of classifications. The staging is not the same for patients with HPV-positive tumors.

What are the main takeaways for oncologists from this research?

Is there anything else you would like to highlight?

The ability to predict the outcomes of patients is beginning to be successful, but we would hope that E2 expression analysis could be part of any regimen of prognosis for these patients.I hope this research opens physicians’ minds. There’s a dogma here that if you detect virus human DNA in a patient, they must have an integrated tumor, but that is not necessarily true. We want to be able to provide a simple test to characterize these tumors that might be useful in dictating treatment strategies for these patients. We want to try to get to a situation where actually looking at E2 RNA would be something that we would do routinely for HPV-positive people. We're not curing cancer, but we believe this is something that needs to be studied further.

Additionally, there are many tumor samples that can be looked at but they are characterized in the wrong way. I think if those samples were reanalyzed in light of what we’ve seen, the results would be different.

It’s important to note that our results don’t contradict anybody; it’s the interpretation of the results that we think might be incorrect. We are not saying anybody is wrong in what they’re seeing and concluding, but the interpretation of what we think is integrated from previous data might not be 100% correct.

We have to do certain experiments to show that we're right, which we’re proposing to do in the next couple of years.