A Gut Feeling: Bacteria in Your Gut May Affect Cancer Treatment


The human gut contains hundreds of species bacteria, which are known to contribute to various bodily functions (such as digestion, of course!) but they also shape our immune system. Now, recent research has revealed how our microbiomes (the abundant bacteria living in our bodies) may affect the efficacy of immune checkpoint blockade (ICB) in cancer treatment.

How it started: about two years ago, an American group of scientists led by Thomas Gajewski of the University of Chicago noticed that melanoma (and some other cancers’) growth in mice was influenced heavily by the type of bacteria found in the mouse gut. They worked with mice purchased from two different vendors, and realized that mice from one vendor had consistently slower-growing tumors. Bifidobacterium bacteria present in the mouse gut were pinpointed to be the culprit, because transfer of Bifidobacterium to mice that did not have it was able to slow down melanoma growth. Treatment with an immune anti-PD-L1 drug was effective in mice that had the bacteria, but not in mice lacking it.

Meanwhile, a French group led by Laurence Zitvogel of Institut Gustave Roussy showed that the gut microbiome, especially a certain species of the Bacteroides genus, heavily influences response to ICB with the anti-CTLA-4 drug ipilimumab (Yervoy, approved by the U.S. Food and Drug Administration). Mice grown in a sterile (germ-free) environment (no gut bacteria) and mice treated with antibiotics did not respond to treatment with ipilimumab. However, transferring the bacterium Bacterioides fragilis into these mice restored the response of their melanomas to ipilimumab. Moreover, transferring fecal matter from human melanoma patients who had responded to ipilimumab into these mice also promoted tumor regression. This strongly suggested that certain gut bacteria are required for the anti-cancer effects of ipilimumab—in mice at least.

This research has, of course, led to efforts to evaluate whether the human microbiome also influences efficacy of ICB in people. A study of the microbiome in melanoma patients treated with various ICB drugs found that a higher percentage of patients responding to treatment (R) had microbiomes with high amounts of three types of bacteria, including species of the Bacteroides, Faecalibacterium, and Holdemania groups. Most non-responders (NR) did not have a high presence of these species in the gut microbiome.

Most recently, in November 2017, the journal Science published two studies that marked a big step forward from just looking for associations. The French scientists led by Dr. Zitvogel analyzed if antibiotic treatment influences the efficacy of ICB in patients with non-small cell lung cancer (NSCLC), melanoma, or renal cancer. Before or during treatment with ICB, close to a third of the patients received antibiotics for various reasons unrelated to their cancers. These patients (NR) received notably less benefit from ICB than those who were not treated with antibiotics (R). The overall survival time of the antibiotic-treated patients was significantly shorter.

Fecal samples from 100 patients from both the R and NR groups were analyzed before and after treatment. The results demonstrated a richer (more variety of species) microbiome in R, and more of them had the bacterium Akkermansia muciniphila present in their gut microbiomes than did the NR patients.

Scientists led by Jennifer Wargo from MD Anderson Cancer Center in Houston, TX, conducted a microbiome study in melanoma patients treated with ICB. They also found that R had a higher variety of gut bacteria compared to NR. They analyzed the patients’ oral microbiomes and found no correlations with their response to ICB. However, analysis of the composition of gut bacteria in R versus NR patients showed that the presence of bacteria in the Ruminococcaceae family, in particular of the Faecalibacterium genus, was associated with prolonged progression-free survival of patients. On the contrary, presence of Bacteroidales was associated with lack of response. Presence of Faecalibacterium was also associated with higher levels of specific T cells (immune system cells), both in the melanoma tumors and in the blood stream.

Both studies (from Zitvogel and Wargo) included experiments with mice. When tumor-carrying germ-free mice (no gut bacteria) received fecal transplants from R patients, they had better treatment responses to ICB than did mice given NR fecal transplants. In the French study, good responses to ICB in germ-free mice could be also promoted by directly feeding mice with the species A. muciniphila.

These two studies are highly significant for two reasons: they involved fairly large numbers of patients and they revealed the important role of the microbiome in shaping immune responses to ICB, confirmed in mice. There is now a move towards initiating clinical trials that will use modulation of the gut microbiome with “good” bacteria in order to increase the efficacy of ICB in patients.

As with all good science, these data also raise a lot more questions. First, why did the French and American studies find different “good” bacterial species to be involved in modulating the response to ICB? Is it because French versus American patients have different diets and therefore different composition of gut microbiomes?

Also, should all patients about to receive ICB have their microbiomes profiled? And if yes, should the results of profiling (for example, absence of A. muciniphila) eventually be used to “correct” the composition of the gut microbiome to make it a positive factor in the efficacy of ICB? Unlike other predictive factors for the efficacy of ICB, such as presence of the protein PD-L1 or tumor mutational burden, it looks like the microbiome can perhaps be modulated.

If there is indeed a solid rationale for modulating patients’ microbiomes, the question is how to do it. There are problems with culturing some of the “good” bacterial species outside of the body. Should then fecal transplant from “verified” donors be practiced? (“Verified” donors would have all the “good” bugs in their guts.) Should dietary changes be made to encourage the growth of “good bacteria” in patients undergoing treatment?

Unlike the findings in the studies described above, other studies have found that there are plenty of “bad” bacteria associated with cancers, in particular gastrointestinal cancers. Gastric (stomach) cancers are known to be associated with the presence of certain “harmful” bacteria, as described in multiple publications including a recent study in Scientific Reports. Recent work has also shown that the presence of certain ”bad” bacteria in pancreatic cancers is associated with lack of response to chemotherapy drugs—apparently some species of bacteria found in pancreatic tumors destroy the chemotherapy drugs before they have a chance to get to the tumor cells.

Analysis of patients with colorectal cancer showed that during metastasis of this cancer to the liver (a common occurrence), a certain type of bacteria (Fusobacterium) travels to the liver along with the cancer cells. Apparently, these metastatic tumors depend on the presence of Fusobacterium because it persists even after multiple transplantations of tumors in mice. Moreover, treatment of mice carrying these tumors with antibiotics slowed down tumor growth.

This would suggest that using antibiotics during chemotherapy treatment of some gastrointestinal cancers is indicated, but this will be contraindicated in patients undergoing ICB treatments for NSCLC, melanoma, or kidney cancer.

The lesson is that our microbiomes are important in more processes than we could have imagined—a rather trivial statement considering all the complexities, both of the microbiome itself, and the understanding of its role in cancer development and treatment.