Research Shows Growing Attention Toward Gut Microbiome's Role in Immunotherapy Response

Jennifer A. Wargo, MD, MMSC

Researchers’ understanding of why patients with cancer do or do not respond to treatment with immune checkpoint inhibition is constantly evolving, with new developments in innate and adaptive immunity, the tumor microenvironment, and more changing the way that immunotherapy is viewed and used. Many researchers are now pointing to the effect that gut microbiota have on patients’ response to checkpoint inhibitors and its implications for the treatment of patients receiving immunotherapy.

The first major papers to show that gut microbiota may help determine which patients with cancer respond to checkpoint blockades appeared inScience3 years ago. Researchers from the University of Chicago reported that oralBifidobacteriummade 100% of mice respond to anti—PD-L1 therapy. Researchers from France, writing in the same issue, demonstrated that although few bacteria-free mice responded to anti–CTLA-4 treatment, most did respond after receiving transplants ofBacteroides fragilis(B. fragilis). Both teams concluded that a properly constituted microbiome might be a necessary precondition for immunotherapy response.

That claim may have generated more skepticism than enthusiasm at the time, but subsequent research has lent it significant support. Results from newer studies show that the immunotherapy response against several tumor types strongly correlates both with greater microbiota variety and particular types of bacteria. They also show that microbiota from responding humans usually produce immunotherapy responses in nonresponding mice. They’ve even shown that transfers of good bacteria increase the innate immune response in mice, preventing many transferred cancers from taking root and slowing the growth of those that do, even in the absence of any immunotherapy. Studies have yet to uncover practical strategies for translating such discoveries into better outcomes, but the first trials designed for that purpose are already underway.

“I’ve seen a lot of interesting conference presentations, but the initial mouse studies just floored me. I was pretty awed by the implications and the opportunities for related research,” said Jennifer A. Wargo, MD, MMSc, in an interview withTargeted Therapies in Oncology. Her research team at The University of Texas MD Anderson Cancer Center in Houston has joined the teams from Chicago and France in focusing on the link between microbiota and immune response to cancer.

“While our findings are compelling, we also appreciate that there is a lot of work still left to be done. Our results, and those of others indicate that the microbiome could potentially be modulated to enhance the efficacy of immune checkpoint blockade therapies, though this must be carefully tested in the context of clinical trials. We are working closely with the Parker Institute for Cancer Immunotherapy and Seres Therapeutics to launch a clinical trial to test this hypothesis later this year,” added Wargo, associate professor of genomic medicine and surgical oncology at MD Anderson.

Checkpoint blockades produce very durable responses in a minority of patients with a wide variety of tumor types, but a majority of patients don’t respond. Researchers around the world have sought to discover what separates responders from non-responders and develop either diagnostics to predict response or, better still, mechanisms for increasing response rates. These efforts initially focused on the expression of targeted checkpoints in patient tumors but moved on to tumor heterogeneity, germline genetics, the tumor microenvironment, and many other factors.

Research published before 2015 demonstrates both that the microbiome plays a significant role in immune system function and that patients with cancer who innately produce some immune response against their tumors are far more likely to respond to immunotherapy than other patients. The Chicago team thought these 2 findings, when taken together, implied that changes in microbiome composition might produce changes in immunotherapy response. The team began with 1 group of mice from Jackson Laboratory and 1 group of mice (with different default microbiota) from Taconic Farms. Researchers injected melanoma into mice from both groups and noticed that the disease progressed significantly faster in Taconic mice because they generated less innate immune response. The researchers then transferred fecal matter from Jackson mice into Taconic mice and observed slower melanoma growth in Taconic mice. Analysis of the Jackson microbiota suggested thatBifidobacteriumwas the organism that enhanced immune response, so the researchers tested that finding by feeding oralBifidobacteriumto Taconic mice and noted the expected improvement in their innate immune response to cancer and a near universal response to anti—PD-L1 therapy that normally did little for Taconic mice.¹

The French team tested the CTLA-4 inhibitor ipilimumab (Yervoy) in both specific pathogen—free mice and germ-free mice. The treatment worked as expected in the former but not in the latter. The researchers then wiped out the microbiota of the specific pathogen–free mice by giving them broad spectrum antibiotics and found that treatment no longer worked. To determine which bacteria made treatment work, the researchers recolonized germ-free mice with individual bacteria types and found that efficacy was only restored byBacteroides thetaiotaomicron,B. fragilis, andBurkholderia cepacian. They then performed fecal transplants from human patients into germ-free mice and found that transplants rich in B. fragilis produced the best response in the mice.²

The 2 papers have generated more than 500 citations in less than 3 years, as well as a wave of follow-up that has both confirmed and expanded their initial conclusions.Sciencealone has published 1 paper from each of the 3 major research teams explorin this subject in just the past 6 months.

In a study completed by the Chicago team, investigators collected stool samples from 42 patients with metastatic melanoma before they underwent treatment with checkpoint blockades. Team members used a combination of techniques (16S ribosomal RNA [rRNA] gene amplicon sequencing, species-level identities revealed through metagenomic shotgun sequencing, and database searches) to identify the bacteria in each sample and then compared the prevalence of different bacteria in responders and nonresponders. Only a single bacteria-type was significantly associated with response, but combinations of bacteria that were weakly associated with response collectively proved strongly associated with response; every patient whose “good” bacteria levels were elevated by at least 50% responded. The Chicago team also showed, once again, that these bacterial differences were likely the cause of varying responses by demonstrating that fecal transplants from responding patients into nonresponding, germ-free mice induced responses in those mice.³

A study from Wargo’s team in Houston prospectively collected both oral and fecal bacteria samples from 112 patients with metastatic melanoma who were about to undergo anti—PD-1 therapy. After taxonomic profiling via 16S rRNA gene sequencing, 89 eligible patients were classified after 6 months of treatment as either responders (n = 54) or nonresponders (n = 35) based upon RECIST v1.1 criteria. Responders and nonresponders were comparable in terms of demographics and disease (including the genetics of the 10 tumors that researchers analyzed). Oral bacteria diversity was not significantly associated with response, but fecal bacterial diversity was significantly associated with both response and progression-free survival (PFS; HR for high vs intermediate diversity, 3.60; 95% CI, 1.02-12.74). Responders tended to have a much higher diversity of microbiota, consisting of increased levels ofFaecalibacterium,Ruminococcaceae, andClostridiales, than those who did not respond to treatment with PD-1 inhibition. Further analysis of the bacteria in fecal microbiota found that abundances ofFaecalibacteriumwere associated with increased PFS rates (HR, 2.92; 95% CI, 1.08-7.89). Relative abundances ofBacteroidaleswere associated with a trend of lower response rates (HR, 0.39; 95% CI, 0.15-1.03).⁴

The team at MD Anderson have since collected findings from studies of bacteria’s influence on repsonse to immunotherapy, as well as toxicity from immunotherapy treatment. Their findings indicated which bacteria have been associated with responses and toxicity, and which have not (FIGURE).⁵

“There is a definite consistency [among] data sets and different cancer sites. This is not something that is specific to melanoma. It can be applied to multiple cancer types as well as to other therapies, although that needs to be tested further,” Wargo commented in an earlier interview withTargeted Therapies in Oncology.

The French team followed 249 patients who were given a PD-1 inhibitor for lung, kidney, or urinary cancer and compared outcomes among the roughly 25% of patients who had recently taken an antibiotic and those who had not. Analysis of all patients found that both PFS and overall survival (OS) were significantly shorter for antibiotic users than for other patients. The researchers tested to see whether the different outcomes stemmed from the antibiotics’ effect on patient fecal bacteria by sequencing fecal samples from 152 patients and found that microbiome diversity was positively associated with treatment response and negatively associated with antibiotic usage. The commensal bacteria that was most significantly associated with a favorable clinical outcome wasAkkermansia muciniphila(A. muciniphila) (P= .004 considering all patients;P= .003 excluding antibiotic-treated patients). It was detected in 69% of patients (11/16) who had achieved a partial response from PD-1 inhibition and in 58% (23/40) who had stable disease, but in only 34% (15/44) who had progressed or died (P= .007). The researchers then further validated the importance ofA. muciniphilain PD-1/ PD-L1 by testing it in mice.⁶

A related study the same team performed on confirmation cohorts of patients with advanced renal cell carcinoma (RCC) or non—small cell lung cancer (NSCLC) who were treated with PD-1 inhibitors found significantly different outcomes between patients who did and did not use antibiotics. The median OS was 17.3 months among 16 patients with RCC who received antibiotics and 30.6 months among the 105 who did not (HR, 3.5; 95% CI, 1.1-10.8;P= .03). The median PFS was 1.9 months in the former group and 7.4 months in the latter (HR, 3.1; 95% CI, 1.4-6.9; P <.01). Additionally, the risk of primary progressive disease significantly increased in patients who used antibiotics (75% vs 22%,P<.010). Patients with NSCLC who received antibiotics within 30 days (n = 48) also had a shorter median OS (7.9 vs 24.6 months; HR, 4.4; 95% CI, 2.6-7.7;P<.01) and a shorter median PFS (1.9 vs 3.8 months; HR, 1.5; 95% CI, 1.0-2.2;P= .03) than patients who did not. Multivariate analyses further demonstrated the association of antibiotic use on PFS for patients with RCC and on OS for patients with NSCLC.⁷

&ldquo;These studies obviously argue that you should probably be very cautious about prescribing antibiotics to a [patient who is receiving] immunotherapy. But if a patient has a serious bacterial infection, you really don&rsquo;t have any choice. We are now trying different methods, including pre-biotics and probiotics, to improve microbiota composition and boost immunotherapy effect. As for trying to create a &lsquo;good&rsquo; microbiome before treatment with pro-biotics or kale drinks or anything like that, there&rsquo;s no evidence to support it. Eating a good diet remains the only recommendation we can provide to our patients,&rdquo; said Bertrand Routy, MD, PhD, an assistant professor of hemato-oncology at Centre hospitalier de l&rsquo;Universit&eacute; de Montr&eacute;al.

&ldquo;The next step is validating much of what we&rsquo;ve found in larger studies and doing trials to investigate the many questions that remain unanswered,&rdquo; said Routy, who collaborates with the French microbiome research team and is the corresponding author of the 2 antibiotic studies. &ldquo;Is there 1 bacteria type or 1 bacterial profile that drives immunotherapy response in all cancers? Do different bacteria drive response in different tumor types? Do certain antibiotics preserve response better than others? Can we add bacteria before we start treatment to increase response rates?&rdquo;

Adding bacteria via fecal microbiota transplants is already a standard treatment forClostridium difficilecolitis, but the process is time consuming, costly, and distasteful to many patients. A number of companies are working to produce treatments that deliver oral doses of bacteria that might help fight cancer and other diseases.

Seres Therapeutics, for example, revealed at the 2018 American Association for Cancer Research Annual Meeting that it had increased both adaptive immune responses and antitumor response rates in tumor-bearing germ-free mice by colonizing mice with microbiota from healthy people (rather than from patients with cancer who are already known to respond to immunotherapy). The company also said that colonization with just the spore fraction of each donor&rsquo;s microbiota was also effective, a point that&rsquo;s important because bacterial spores are hardy enough to survive as oral supplements that could end the need for fecal transplants. The latter observations were also made in conventional mice that were colonized following a brief treatment with antibiotics.⁷

&ldquo;The next step for us is to start a clinical phase I trial sometime later this year,&rdquo; said Lata Jayaraman, PhD, the company&rsquo;s head of tumor immunotherapy. &ldquo;The plan is to enroll 60 [patients with] melanoma, all of whom will receive anti—PD-1 treatment. Patients will be randomized to placebo, to fecal material from other patients who were complete responders, or to an oral preparation of bacterial spores from healthy donors.&rdquo;

Many studies have already been initiated to further explore the role of gut microbiota and if it can be manipulated for improved response to immunotherapy and other treatment modalities. A team from Memorial Sloan Kettering Cancer Center (MSK) in New York has already gone a bit further and submitted its first results from a trial of fecal transplants in patients with cancer (NCT02269150). Researchers took stool samples from patients who went on to undergo high-dose chemotherapy or radiation before receiving bone marrow transplants. They then identified patients whose microbiomes were damaged by treatment and randomized them to placebo or doses of their banked fecal samples. The first trial results should show whether the transfers repair microbiome damage. Subsequent results should show if they improve outcomes.

Another research team at MSK hopes to test practical strategies for maximizing immunotherapy response in patients with cancer who must receive antibiotics. Their trial, which has begun recruiting patients, will randomize patients undergoing steam cell transplantation with febrile neutropenia between standard antibiotics, antibiotics designed to target anaerobic bacteria, and antibiotics designed to spare anaerobic bacteria (NCT03078010).

&ldquo;The immunotherapy-boosting bacteria identified in the research to date have mostly been anaerobic. That is why there&rsquo;s not much hope for boosting cancer immunity with store-bought probiotics—which don&rsquo;t usually have anaerobic bacteria&mdash;and why there&rsquo;s hope of preserving immunotherapy efficacy by using antibiotics that don&rsquo;t target anaerobic bacteria,&rdquo; said Jonathan Peled, MD, PhD, an MSK hematologist who studies how the microbiome affects transplant success and graft-versus-host disease.

&ldquo;There are a number of studies like this that are either underway or about to get underway, reasonably short and simple studies that may justify some practical clinical applications for what we have learned,&rdquo; Peled said. &ldquo;Immunotherapy is obviously very exciting on its own, but it&rsquo;s particularly exciting to think that something could substantially increase the percentage of patients that benefit from it.

References:

1. Sivan A, Corrales L, Hubert N, et al. Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy.Science.2015;350(6264):1084- 1089. doi: 10.1126/science.aac4255.
2. V&eacute;tizou M, Pitt JM, Daill&egrave;re R, et al. Anticancer immunotherapy by CTLA-4 blockade relies on the gut microbiota.Science.2015;350(6264):1079-1084. doi: 10.1126/science.aad1329.
3. Matson V, Fessler J, Bao R, et al. The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients.Science. 2018;359(6371):104- 108. doi: 10.1126/science.aao3290.
4. Gopalakrishnan V, Spencer CN, Nezi L, et al. Gut microbiome modulates response to anti—PD-1 immunotherapy in melanoma patients.Science.2018;359(6371):97- 103. doi: 10.1126/science.aan4236.
5. Gopalakrishnan V, Helmink BA, Spencer CN, Reuben A, Wargo JA. The influence of the gut microbiome on cancer, immunity, and cancer immunotherapy.Cancer Cell.2018;33(4):570-580. doi: 10.1016/j.ccell.2018.03.015.
6. Routy B, Le Chatelier E, Derosa L, et al. Gut microbiome influences efficacy of PD- 1-based immunotherapy against epithelial tumors.Science.2018;359(6371):91- 97. doi: 10.1126/science.aan3706.
7. Derosa L, Hellmann MD, Spaziano M, et al. Negative association of antibiotics on clinical activity of immune checkpoint inhibitors in patients with advanced renal cell and non-small cell lung cancer [published online March 30, 2018].Ann Oncol.doi: 10.1093/annonc/mdy103.
8. Sceneay J, Srinivasan S, Halley K, et al. Leveraging gut microbiota networks to impact tumor immunotherapy. Presented at: 2018 American Association for Cancer Research; April 14-18, 2018; Chicago, IL. Abstract LB-283. www.abstractsonline.com/pp8/#!/4562/presentation/10599.