How Bacteria Could Protect Tumors From Anticancer Drugs

The success of a treatment might depend on a person’s microbes.

Microscope image of bacteria inside a pancreatic cancer cell
Bacteria, in green, inside a pancreatic cancer cell (Leore Geller)

Cancers have unwitting allies: the healthy cells that surround them. Several groups of scientists have now found that normal cells can inadvertently release substances that shield their malignant neighbors from anticancer drugs. That would explain why even targeted therapies—smart drugs that are meant to hit the specific genetic faults behind various cancers—sometimes stumble right out of the gate. When pitted against isolated cancer cells in laboratory tests, they perform as expected. But when pitted against actual tumors, which enjoy a kind of innate resistance because of the healthy cells around them, the drugs can fail.

But at least half of the cells in the human body are not human.

Every person is a seething colony of microbes—a collection of tens of trillions of bacteria and other microscopic organisms that live in and on our bodies. And a team of researchers, led by Ravid Straussman from the Weizmann Institute of Science and Todd Golub from Harvard Medical School, have shown that some of these bacteria can also shield tumors from anticancer drugs.

Back in 2012, Straussman and Golub’s team grew dozens of types of cancer cells together with dozens of types of healthy cells, and found hundreds of combinations where the latter protected the former to some degree against chemotherapy. But one particular interaction was especially dramatic: A lineage of skin cells from one individual could completely protect pancreatic cancer cells from gemcitabine—a frontline drug that’s used to treat this stubborn disease.

“We could pour on more and more gemcitabine—ten times more than was needed to kill the cancers—and the skin cells from this woman were enough to protect them,” Straussman recalls. Even the liquid in which the skin cells had grown was enough to protect cancers from gemcitabine. Clearly, the skin cells were secreting some kind of chemical that neutralized the drug. But what was it? A protein? A piece of DNA? The team spent years trying to identify the mystery molecule, to no avail. “We did tons of experiments and they led us nowhere,” says Straussman. “It didn’t make any sense.”

They finally worked out what was happening when they filtered the liquid—and completely removed its ability to protect tumors. Even filter paper with very large pores, through which most molecules could easily fit, had this effect. That’s when they realized that they weren’t dealing with a molecule at all. They were dealing with a microbe.

The team, including Straussman’s student Leore Geller, showed that a bacterium called Mycoplasma hyorhinis had infected the skin cells. If they killed it off with antibiotics, the cells could no longer rescue cancers from chemo. And if they added the bacterium to mice that were suffering from tumors, the rodents became resistant to gemcitabine.

It turns out that Mycoplasma hyorhinis has an enzyme called CDD that can dull gemcitabine’s sting by converting it into another inactive chemical. And it’s not alone. One in every nine known species of bacteria has a version of CDD with the same ability. And some of these drug-defusing bacteria exist within the actual tumors of people with pancreatic cancer.

The team showed that by checking biopsies from the tumors of 113 pancreatic cancer patients, as well as samples from 20 healthy pancreases taken from organ donors. Just 15 percent of the healthy organs contained traces of bacterial DNA, compared to 76 percent of the tumor samples. And while bacterial DNA could just have come from dead cells, the team also saw whole intact bacteria within the tumors. They treated the samples with glowing antibodies designed to latch onto bacterial molecules, and then looked through a microscope. And there they were: bacteria, nestled among—or sometimes within—the cancer cells.

The bacteria in the tumors were a motley crew of varied species, and Mycoplasma wasn’t among them. But almost all of these microbes had the CDD enzyme, and almost all of them could neutralize gemcitabine.

Straussman thinks that these bacteria could reach the pancreas through two different routes. The pancreas secretes hormones and digestive enzymes into the gut, and microbes could reach it by traveling in the opposite direction. Alternatively, they could travel via the blood. Bacteria enter our bloodstream all the time, leaking through from our guts, mouths, and other organs. Typically, the immune system takes care of these interlopers, but tumors are weird places where the immune system is suppressed and blood vessels are unusually leaky. They could easily act as refuges for wayward bacteria.

“It would be naïve to think that antibiotics would cure these patients,” Straussman says. For a start, microbes would be just one of the many hurdles to successfully treating pancreatic cancer. Courses of antibiotics might also take out bacteria in other parts of the body, and some studies have shown that certain cancer drugs depend on the presence of such microbes. And finally, you’d likely need to administer antibiotics for a long time to kill off the bacteria lurking within a pancreatic tumor. “You’d get antibiotic-resistant bacterial strains, and you really don’t want those in patients,” Straussman says. A better solution, he says, would be to look for molecules that can block the CDD enzyme, and stop the bacteria from neutralizing gemcitabine.

Other scientists will need to replicate this work to see if they can detect the same bacteria in these kinds of tumors. In studies of this kind, there’s always the nagging possibility that the microbes in question could have come from equipment, the researchers’ bodies, or the laboratory environment, rather than from the actual tumors.“Contamination is a hazard in this work; this is an important concern,” says Libusha Kelly from the Albert Einstein College of Medicine.

And even if the microbes are there, it’s not clear if they would neutralize gemcitabine to the degree the researchers saw in the lab, or if they would actually make a difference to a person’s treatment. Microbes or no, gemcitabine or no, pancreatic cancer is still a notoriously difficult disease to treat. “Many different microbial enzymes have the capacity to metabolize drugs,” Kelly says. “What we don’t understand is how often this potential is actually realized in a way that influences treatment outcomes.”

Still, “the study reminds us that the tumor microenvironment consists of more than just human cells,” says Emily Balskus from Harvard University. There are microbes too, and variations in these communities could explain why some people respond well to treatments and others do not. “We can’t fully understand human health and disease without considering these organisms.”

“This work could also have implications for cancer drug development and cancer therapy,” Balskus adds. Doctors might make different decisions about how to treat a patient based on the microbes that are found in or near their tumors. And researchers might develop new ways of treating cancer by targeting these microbes instead of the cancer cells directly.

“It’s not just pancreatic cancer and it’s not just gemcitabine,” Straussman says. “There are bacteria in tumors that we never knew bacteria were present in—and that fact has broad implications. These microbes might influence so many other hallmarks of cancer.” Do they affect a patient’s response to immunotherapy—treatments that stimulate the immune system? Do they influence cancer stem cells—the cells that generate the bulk of a tumor? Do they influence metastasis—the process through which cancers spread through the body?

“Every buzzword you can think of in cancer ... how are they affected by bacteria in tumors?” says Straussman. “My team is exploring many of these directions, but we’re just starting.”