This article is part of our May 2021 special issue. Download the full issue here.
Davar D, Dzutsev AK, McCulloch JA, et al. Fecal microbiota transplant overcomes resistance to anti-PD-1 therapy in melanoma patients. Science. 2021;371(6529):595-602.
Proof of concept; open-label trial
Between June 2018 and January 2020, researchers enrolled 16 melanoma patients who had not responded to anti–PD-1 (anti–programmed cell death protein 1) immunotherapy and continued to have progressive disease.
Study Medication and Dosage
Participants received fecal microbiota transplant (FMT) from patients who had responded to treatment. Seven donors, including 4 who had experienced a complete response and 3 with partial response to treatment with anti–PD-1 therapy, supplied fecal materials to treat the 16 patients. Each study participant received a single donor–derived FMT followed by additional pembrolizumab therapy every 3 weeks until disease progression or intolerable toxicity.
Progression-free survival (PFS) and overall survival (OS). In addition, researchers conducted and compared detailed analyses of fecal biome in patients and donors. They analyzed a total of 223 fecal samples from 15 of the recipients and from 7 donors.
Fifteen of the enrolled patients received FMT and had at least 1 restaging scan via computed tomography (CT) and were evaluable for response. One patient declined so rapidly as their disease progressed that their response was not included. Researchers noted objective responses (ORs) in 3 out of the 15 patients, a 20% response rate. In addition, 3 of the 15 patients, an additional 20%, had durable sustained disease (SD) that lasted longer than 12 months.
PFS and OS of all patients were 3.0 and 7.0 months, respectively, at a median follow-up of 7 months. In the 6 patients with disease control (ie, OR and SD), median PFS and OS were both 14.0 months, respectively. Among these patients, 1 patient exhibited ongoing partial response after >2 years and is currently on surveillance, and 4 patients still remain on treatment. One patient who had a complete response to treatment died from a complication of an elective surgery unrelated to their cancer or to treatment.
FMT together with pembrolizumab treatment overcame resistance to anti–PD-1 treatment in a subset of refractory melanoma patients.
Note: A similar study by Baruch et al was published in the same issue of Science as this study by Davar et al under discussion. In the Baruch study, 10 patients with refractory metastatic melanoma were treated with FMT from 2 donors who had complete responses to prior immunotherapy treatment. The combination of FMT and, in the Baruch study, nivolumab rather than pembrolizumab resulted in 3 responses, including 1 complete response.1
Both of these drugs, pembrolizumab and nivolumab, are monoclonal antibodies that bind to and block PD-1. The US Food and Drug Administration approved them at about the same time; pembrolizumab was approved in September 2014 and nivolumab that December, for treating advanced melanoma. They are now both approved to treat a range of cancers. PD-1 is a negative regulator of T-cell immune function; that is, it suppresses the immune system’s response. This suppression guards against autoimmune disease, but it also weakens the immune system’s ability to kill cancer cells. Anti–PD-1 drugs, such as these, halt this suppression, allowing the immune system to recognize and kill cancer cells.2
These drugs provide long-term clinical benefit to nearly 40% of patients with advanced melanoma, though only 10% to 20% of patients will have durable complete responses.3 Researchers are directing their efforts toward increasing this response rate. Aware that antibiotic use with these drugs is associated with poorer response rates, and knowing the microbiota influence on the development and function of the immune system,4 investigators have focused on how altering the gut microbiota might improve outcome. Researchers have attempted to identify specific bacteria associated with better treatment outcomes.
For naturopathic doctors, anti–PD-1 drugs are outside our scope of practice, but we should understand the implications for our patients who are undergoing treatment with these or similar drugs.
The term “microbiome” refers to the collective genomes of all the microorganisms in a particular environment, and microbiota are the community of microorganisms themselves. There are about 100 trillion microorganisms, mostly bacteria but also viruses, fungi, and protozoa, living in the human digestive tract. Conceptually we should view the microbiome as a virtual organ, a part of the body. The human genome consists of about 23,000 genes, while the microbiome contains more than 3 million genes, producing thousands of metabolites that augment or replace many of the functions of the host and strongly influence the health of the host.
For several years, we've watched this idea develop—that the gut biome affects immunotherapy drugs. In January 2018, Gopalakrishnan et al reported that after analyzing oral and gut microbiomes of human melanoma patients (N=112), significant differences were seen in diversity and composition in responders versus nonresponders to anti–PD-1 treatment. Fecal samples from the responders showed significantly higher alpha diversity.5
Alpha and beta diversity are 2 terms commonly used in biome studies. They are both higher-level measures used to describe the microbiome in a sample. They do not provide information on changes in abundance of specific taxa. Alpha diversity is a composite of different measurements that estimate the diversity within a single sample. These measures reflect the richness (number) or distribution (evenness) of a microbial sample or aim to communicate a combination of both properties. Beta diversity is a measure of the similarity or dissimilarity of 2 communities.6
In 2018, Bertrand Routy profiled stool samples from patients with lung and kidney cancers and reported that nonresponding patients had low levels of Akkermansia muciniphila. Oral supplementation of these bacteria to antibiotic-treated mice restored their response to immunotherapy.7 Researchers have now created lists of bacteria associated with positive clinical response and which seem to be lacking in nonresponding patients, yet the results have not been definitive enough to yield a recipe for a bacterial combination to give all patients. Instead, both of these 2 new studies used fecal transplants from responders as a way to transfer the entire gut microbiota from 1 patient to another.
Earlier studies had already reported that FMT in mice would transfer benefit from responders to nonresponders, but the question as to whether this could translate into humans persisted. Mice and humans have dissimilar gut microbiota; they share only about 4% of their bacterial metagenome.8 The results from both Davar and Baruch tell us that FMT does work in humans as well as in mice. This is a big deal.
The anti–-PD-1 drugs have had a large impact in treating melanoma over the last few years. Nivolumab and pembrolizumab have achieved an overall response rate (ORR) of 40% to 50% and a 5-year OS rate of 30% to 40% in patients with metastatic melanoma.9 These treatments are expensive. In 2016 treatment with either drug was priced at about $150,000 per year. Combination therapy with both pembrolizumab and nivolumab was a deal at $256,000 per year.10
Let’s pause to do the math on a napkin. Sixty out of 100 patients treated with immunotherapy do not respond. Let’s say that patients in the Davar study were recruited from those 60 nonresponders, and 20% of those 60—that is, another 12 people—responded to additional drug treatment, increasing the number who respond to treatment from 40 to 52. If this proves true, it is remarkable. Even with all the caveats that this was a small study, etc., this still deserves attention as FMT is something that patients might even do on their own, perhaps before they even begin treatment. It’s also something that the drug companies should support, as nonresponders may be justified to attempt further treatment with these drugs after receiving FMT.
As anti–PD-1 drugs are now approved for so many cancers, chances are increasing that any cancer patient you see will eventually be treated with 1 of them. Proactive efforts to improve their response rate by modifying their gut biome may be useful. At this point it is unknown whether similar shifts in the biome that increase response rate to anti–PD-1 drugs affect the natural course of cancer in those not exposed to these drugs. As the general characteristics of a healthy gut biome appear to be shared across a broad range of conditions, it is possible that such measures may prove useful.
We could/should start applying knowledge of how diet and supplements change the gut biome and use that knowledge to inform interventions against a broader range of diseases than just cancer. The basic goal should be to increase alpha diversity. Low diversity is associated with a range of diseases including inflammatory bowel disease, psoriatic arthritis, diabetes, atopic eczema, obesity, and arterial stiffness. Diversity is the “good indicator” being used to define a healthy gut.
Cancer patients, specifically those scheduled for anti–PD-1 treatment, should avoid antibiotics and take steps to encourage alpha diversity.11 They should consider FMT, though at this point, whether this should be done prior to initial treatment is uncertain. We do not yet know of an authorized source for obtaining stool “infusates” from “anti–PD-1 responders.” Certainly, nonresponders to treatment should have FMT before resuming treatment. Davar reports that gut microbiota can change significantly after a single FMT.
The results from both Davar and Baruch tell us that FMT does work in humans as well as in mice. This is a big deal.
In general fiber is considered good for increasing diversity, but recent interventional studies indicate that major increases in a single type of dietary fiber can temporarily reduce diversity, as the microbes that digest the specific type of fiber being supplemented are specifically enriched.12 The obvious work-around is not to rely on 1 specific type of fiber but to consume a range of them and eat a wide variety of foods.
High-intensity sweeteners, the sugar substitutes, are now suspected of reducing alpha diversity. These sweeteners, such as sucralose, aspartame, and saccharin, have been shown to disrupt the balance and diversity of gut microbiota.13,14
Food additives—in particular, emulsifiers found in most processed foods—should be avoided as they are associated with reduced diversity, at least in animals.15 Promoting consumption of less-processed foods and avoiding ultra-processed foods should remain central.
While the focus of our patient advice should be to promote consumption of a healthy diet, many patients are attracted to 1 or more of the popular restrictive diets that limit consumption of specific foods. We need to view these diets with caution in light of how they impact the gut biome and balance that with any symptom relief provided by avoidance. The low FODMAP (fermentable oligosaccharides, disaccharides, monosaccharides, and polyols) diets used to treat irritable bowel syndrome may be the most problematic for obvious reasons. The complex saccharides that this diet hopes to eliminate are food sources for intestinal bacteria, and eliminating these foods will literally starve some of the biome.
Data comparing vegans with omnivores have shown only slight differences in gut bacterial communities between groups. There are significant differences, though, in the metabolomes (the metabolites produced by the bacteria and found in the blood) between the 2 groups.16 Whether these differences affect anti–PD-1 therapy is not yet clear.
Consuming gluten-free bread probably reduces microbiota dysbiosis in people with celiac disease because it appears to do so in monkeys.17 Still, most people who are avoiding gluten these days do not have celiac disease. A large 2017 observational study reported an increased risk of heart disease in gluten avoiders, probably because they avoid most whole grains, which reduces their intake of the fiber required to support bacteria.18
A small clinical trial reported that the gut microbiota changed significantly—and not for the better—in 21 healthy people on gluten-free diets in just 4 weeks, their microbiota showing lower abundance in several key species.19 How this might affect anti–PD-1 therapy also remains unknown.
We naturopaths can’t help ourselves and are always diet-focused, but we should remember that drugs are the largest influence on gut microbiota, not diet, accounting for 10% of the variations seen.20 Antibiotics, proton pump inhibitors, and osmotic laxatives are the major players and are used regularly by large numbers of people.21,22
The gut microbiota can change remarkably fast. A past study reports that major shifts in bacterial population occur within days to weeks of dietary changes.23 In a trial that shifted participants from a plant-based diet to an animal-protein diet, extreme changes occurred in less than a week.24
While the public now assumes probiotics are something of a panacea, they may spell trouble with anti–PD-1 therapy. Many of us recall work done at MD Anderson Cancer Center and reported at a conference in 2018 by Christine Spencer. Spencer and colleagues had gathered data about diet and supplements from 113 melanoma patients, along with stool samples from each. Spencer reported that patients eating high-fiber diets were 5 times more likely to respond to anti–PD-1 therapy than those on a low-fiber diet. (Although these findings are still unpublished, the authors assure me we will see results in print soon.) We might have predicted benefit, as fiber encourages biome diversity, though few would have guessed this much. Surprisingly, taking probiotic supplements was associated with a 70% lower chance of responding to therapy.25 Probiotics do not necessarily increase diversity; rather they discourage growth of competing bacteria. Utility in treating disease varies by species and strain.26
A study published in March 2021 suggests probiotics may still be useful in anti–PD-1 therapy for lung cancer. Kuzuki Takada reported that in patients with non–small cell lung cancer (NSCLC), using probiotics was associated with a significant difference in outcome for the better. In patients (N=294) treated with either pembrolizumab or nivolumab, progression-free survival was significantly longer (HR[95% CI]=1.73[1.42–2.11]) in those who took probiotics. Although overall survival trended longer, it did not reach significance. Those who did not use probiotics were half as likely to experience disease control (OR[95% CI]=0.51[0.35–0.74], P=0.0004). The idea that probiotics are helpful in NSCLC and harmful in melanoma when combined with anti–PD-1 treatments is disturbing. We prefer data that reveal a consistent response across cancer types.27
It may be that probiotics play a distinct role in lung cancer. We’ve seen hints of this before. A large, pooled analysis (N= 627,988 men and 817,862 women) from 2019 linked both yogurt and fiber consumption with significantly lower risk of lung cancer. High versus no yogurt consumption was associated with a 19% reduction in risk (HR[95% CI]=0.81[0.76–0.87]). The combination of high fiber and yogurt consumption was associated with a 30% decrease in risk (HR=0.67[0.61–0.73]).28
Coming back to the Davar and Baruch studies, we see that the utility of FMT as an adjunctive therapy with cancer immunotherapy is certainly of interest. The medical world will want to study this further prior to adoption, but our patients will be impatient to move forward and incorporate FMT into their treatment regimes.