Staining of tissue section from the patient for fungus. A Hematoxylin and Eosin (a-c) as well as Grocott Methanamine Silver (d) stain were performed on the tissue sections from an AML patient. Figure (a) shows the low power view of the soft tissue, (b) shows the 20X view, (c) shows the 40X view and (d) shows the 40X view using the silver Grocott stain. Fungus in blood vessels is shown with arrows.
Credit: Erle Robertson, PhD, Perelman School of Medicine, University of Pennsylvania; Cancer Biology & Therapy
The history of viruses causing cancer dates back to 1908, with work by scientists who first investigated chicken leukemia at the University of Copenhagen. Two years later, Peyton Rous at Rockefeller University confirmed that cancers could be transmitted between animals by extracting material from a solid tumor in one chicken and injecting it into a healthy chicken that would later develop cancerous cells.
It would take until the mid-1960s to confirm that a human cancer could be caused by a virus. This “oncovirus” is commonly known as Epstein-Barr virus and is associated with several types of cancer. And, in the mid-1980s, German virologist Harald zur Hausen associated two strains of human papilloma virus (HPV) with cervical cancer. This finding paved the way for seminal work that decades later would lead to the development of the HPV vaccine that is now recommended for teens before they become sexually active.
Now, increasing evidence is showing that a dysregulated human microbiome – changes in the diverse population of microorganisms within every person – may play a key role in either setting the stage for some cancers, or even causing them directly.
"There are a lot of different ways to look at this," said Erle S. Robertson, PhD, a professor and vice chair for research in Otorhinolaryngology, who has been studying the link between viruses and cancer for over two decades. One possibility is that microbes (viruses, bacteria, fungi, parasites) could be adding something harmful to the cellular microenvironment that pushes cells over the edge into uncontrolled growth. On the other hand, certain organisms may simply find tumors a favorable environment, without having any direct involvement with the cancer. "They might just be there because it's a good place to hang out," said James Alwine, PhD, a professor of Cancer Biology and frequent collaborator with Robertson.
Using a new technology developed at Penn called PathoChip, which contains an array of DNA segments from 60,000 different microbes, Robertson, Alwine, and other Penn collaborators from a range of disciplines have so far screened cancer tissue samples from patients with triple-negative breast cancer (TNBC), ovarian cancer, and oral and oropharyngeal squamous cell carcinoma. To date, they have found a unique suite of species in each microbe category that the PathoChip detects associated with each type of cancer.
“Outsmarting cancer takes a village, so to speak,” Robertson said. “The PathoChip technology brings together oncologists who understand a cancer’s pathology, surgeons who treat the patients by removing their cancerous tissue, and pathologists who identify hard-to-treat sub-types of certain cancers. We have a unique environment and culture here at Penn to nurture ideas that help our patients.”
Knowing what microbes are associated with a certain cancer may also have therapeutic implications. Many of the identified microbes can metabolize drugs, so their presence might diminish the effect of some treatments. This knowledge could guide clinical decisions on what therapy to choose. “We are doing this to more specifically identify effective strategies to treat, control, or modify the cancer or treat the tumor-associated microbes associated with a particular group of patients to prevent these malignancies in the first place,” Robertson said.
The team started out testing the PathoChip in 2015, when they identified an association between two different clusters of microbes and TNBC, the most aggressive form of the disease. The team found distinct microbial groups in TNBC tissue compared to non-cancerous samples. The assemblages settled into two broad clusters, one predominantly viral and the other predominantly bacterial, with some fungi and parasites.
In early 2017, Robertson identified a microbiome signature unique to ovarian cancers. Again, they found a distinct group of viral, bacterial, fungal, and parasitic species in ovarian cancer tissue. In addition, they also identified viral integration sites within the cancer patient’s tumor cell genome. Cancer-causing mutations can happen when the DNA of a microbe, usually a virus, inserts into the host organism's preexisting DNA, as part of its replication strategy.
“If we can identify these disease-associated microbe groups earlier in the course of cancer, we could take greater steps to prescribe treatment because ovarian cancer is one of the most deadly and difficult to diagnose, typically when it is already too late,” Robertson said.
Additionally, a better understanding of the link between cancer and infectious microbes may provide additional targets to develop therapeutics. “However, to determine if microbes are ‘direct drivers’ of ovarian cancer, we will need a greater number of samples from a diverse set of patients within a large patient population,” Robertson said. “This approach would have to be similar to what was done for finding the link between cervical cancer and HPV.”
Most recently, the team used PathoChip to home in on a molecular signal unique to oral cancer. They found two distinct microbial populations, one comprised of DNA, and the other RNA viruses, including cancer-causing viruses, bacteria, fungi, and parasites. They surmise that identifying the microbe group and their integration sites may be developed into biomarkers for oral cancer diagnosis and prognosis, as well as if the microbes have a role in the progression of oral cancer.
Last year, the team applied PathoChip in a slightly different way to analyze tissue samples from a patient with relapsed acute myelogenous leukemia (AML) who developed an unknown infection following chemotherapy. “We've run many tests to see if we could identify pathogens in the lab, just to see if the PathoChip can identify a variety of organisms, but this was the first time we actually looked directly at a patient sample to identify a pathogenic agent,” said Robertson.
Patients who are undergoing treatment for cancers often face the added challenge of a compromised immune system, which can be attributed to both their condition and the drugs used to treat it. Many of these infections are not only life-threatening, but caused by rare organisms that are extremely difficult to isolate and identify. However, the sooner an infection is pinned down, the faster and more effectively it can be treated. PathoChip allows for a single sample to be tested simultaneously for thousands of possibilities, greatly reducing the time required for diagnosis.
The Penn team first screened the patient’s sample and analyzed it with a bioinformatics program to narrow the focus to a specific microbial family. Seventeen organisms displaying the highest signal were compared with the signals from a control sample. After all that drilling down, they identified a specific infectious agent – in this case one of the two species of Rhizomucor, a rare fungus known to cause infections in humans.
Whether it is getting a jump on identifying cancer or finding a related opportunistic infection, the team is now working on developing simple blood tests to detect the different microbial groups associated with cancer type. The PathoChip diagnostic tool could work as an early surveillance system for finding patients who might be prone to developing certain cancers.