The human intestinal lumen houses a large number of diverse microbes including bacteria, fungi, archaea, viruses, and protozoans. Together these commensal and pathogenic organisms can be collectively referred to as the fecal, or gut, microbiome.
Rapid advances in genomic technology have facilitated a better understanding of each organism and the overall ecology of the gut microbiome, while culture-free methods along with strong bioinformatics tools paved the way for discovery of the important role these microbes play in regulating host immunity. Very quickly, an important and evolving impact has become apparent for the gut microbiome across a variety of human diseases including cancer and cancer therapeutics (N Engl J Med 2016;375:2369-2379).
Early evidence linking the gut microbiome to anti-cancer immunity came from studies investigating the mechanism of action of cytotoxic chemotherapeutic agents such as cyclophosphamide. Translocation of gut microbes to the systemic circulation was shown to play an important role in eliciting immune responses associated with cyclophosphamide in murine models (Science 2013;342:971-976).
More specifically, systemic translocation of gut bacteria increased a pathogenic subset of T helper cells (Th17) and memory T helper cells, and the therapeutic effects of cyclophosphamide are thought to be partially mediated through this anti-tumor immune response (Science 2013;342:971-976).
Beyond chemotherapy, hematopoietic stem cell transplant (HSCT) has long been known to be impacted by clinical factors associated with microbiology, such as antibiotics use. Laboratory models initially suggested an important role for the gut microbiome in the development of graft-versus-host disease (GVHD) after allogeneic stem cell transplant (J Natl Cancer Inst 1974;52(2):401-404, Radiat Res 1971;45(3):577-588), and clinical studies have subsequently demonstrated correlations between the composition and diversity of the gut microbiome at the time of stem cell transplant and patient outcomes.
As examples, a single center trial showed decreased disease relapse among HSCT patients with abundance of Eubacterium limosum in their stool (J Clin Oncol 2017;35(15):1650-1659), and a retrospective analysis of HSCT recipients showed an association between low diversity of the fecal microbiota with significantly increased mortality (52% vs. 8%) (Blood 2014;124:1174-1182). In the later study, increased bacterial diversity and increased amounts of the genus Blautia were both independently shown to be associated with reduced GVHD-related mortality (Biol Blood Marrow Transplant 2015;21(8):1373-1383).
Surrounding antibiotic use, the impact of broad spectrum antimicrobials used for infection prophylaxis and treatment during transplant course have been strongly correlated with outcome. A retrospective analysis of allogeneic HSCT recipients found the use of specific broad spectrum antibiotics to be associated with loss of several commensal gut bacteria including Bacteroidetes, Lactobacillus spp., and Clostridiai, which in turn was associated with increased GVHD-related mortality at 5 years (Sci Transl Med 2016;8:339ra71).
A randomized trial of metronidazole compared with ciprofloxacin for gut decontamination post-HSCT demonstrated a decreased incidence of acute GVHD but no difference in chronic GVHD or overall survival (Blood 1999;93(10):3267-3275).
Immunotherapy Response
The development of immunotherapeutic agents employing checkpoint blockade to illicit anti-tumor immune response, such as anti-PD-1/L1 or CTLA-4 Abs, represented a major breakthrough in cancer therapy. Resistance to immunotherapy remains a challenge however with both primary and secondary resistance seen across tumor types.
The mechanisms involved in resistance to immunotherapy encompass tumor intrinsic factors and systemic factors including germline genetics and environmental factors (Science 2018;359:582-587, Appl Immunohistochem Mol Morphol 2018;26(2):e15-e21) with the effect of the gut microbiota on response to checkpoint blockade being detailed in a series of recent pre-clinical and clinical studies.
Early evidence for a connection between the gut microbiome and immunotherapy response emerged from the observation that mice with solid tumors grown in germ-free conditions or treated with broad spectrum antibiotics had poor response to immunotherapy (Science 2013;342:967-970). The similarly poor response to immunotherapy in these germ-free or antibiotic treated mice was associated with decrease in TNF-[alpha], one of the main mechanisms through which immunotherapy exerted its effects.
Further, the bacterial species Alistipes shahii appeared to be one of several overrepresented species in of mice responding to immunotherapy (Science 2013;342:967-970) and this was correlated with increased TNF-[alpha]. As a therapeutic consideration, oral administration of this bacteria produced a TNF-[alpha] dependent improvement in response to immunotherapy in these mouse models.
Another important observation came from genetically identical mice grown at different animal facilities in which tumor growth rates were divergent despite genetically identical mice and implanted tumors. These differences were ameliorated by cohousing and fecal transfer from responding mice, pointing to the role of the gut microbiota in modulating these differential responses (Science 2015;350:1084-1089).
Enrichment of Bifidobacterium species positively correlated with delayed tumor growth in these mice and with higher CD8+ tumor-infiltrating T cells. Oral administration of Bifidobacterium alone resulted in enhanced tumor control through CD8+ T cells resembling response to anti PD-L1 therapy and combination treatment with Bifidobacterium and anti PD-L1 therapy nearly abolished tumor outgrowth in these mice (Science 2015;350:1084-1089).
In mouse models of sarcoma, melanoma, and colon cancer treated with anti CTLA-4 therapy and subjected to germ-free conditions or broad spectrum antibiotics, similarly poor response to immunotherapy was observed (Science 2015;350:1079-1084). Bacteroides thetaiotamicron and Bacteriodes fragilis were shown to be necessary for anti-CTLA4 antibody efficacy in these mice. These bacteria elicited an interleukin 12-dependent anti-tumor T helper cell 1 (Th-1) response.
Building on Preclinical Models
Clinical studies built on these preclinical models by investigating differential enrichment of several bacterial species in the stool samples of patients with metastatic melanoma responding to PD1 or CTLA-4 based immunotherapy (Science 2018;359:104-108).
Two species of Bifidobacterium were among six other species that were associated with response to immunotherapy in a metastatic melanoma cohort. This data was used to create a model for predicting response to immunotherapy in this cohort based on the ratio of beneficial to non-beneficial gut bacterial species.
Furthermore, "murine avatars" were created through stool transfer from responding and non-responding patients into germ-free mice and these animals were shown to recapitulate the response phenotype of the donors (Science 2018;359:104-108). Similarly, from patients with non-small cell lung cancer (NSCLC) and renal cell carcinoma (RCC), patients responding to immunotherapy demonstrated differential enrichment of specific bacterial species in comparison to non-responders. In this cohort, enrichment of Akkermansia muciniphila at diagnosis correlated positively with response to immunotherapy (Science 2018;359:91-97).
Antibiotics are known to cause dysbiosis, or changes in the gut microbiome, that can persist for months. In a patient cohort of NSCLC, RCC, and urothelial carcinoma patients treated with anti-PD-1 and PD-L1, those who received antibiotics around the time of treatment had significantly shorter progression-free survival (PFS) and overall survival (OS) (Science 2018;359:91-97). Similar results have been recapitulated by others where again the antibiotic treated group had significant reduction in PFS and OS in both tumor types (Ann Oncol 2018;29(6):1437-1444, Oncol Lett 2019;17(3):2946-2952).
Although these studies are limited by their retrospective designs, they provide important clues to the potential adverse effects of antibiotic treatment for patients receiving immunotherapy and argue for the need of prospective data to answer this question.
Immune-related adverse events remain a challenge with the potential for high morbidity in patients receiving checkpoint blocking agents. As the gut microbiome was shown to regulate local and systemic immunity, a number of studies have explored the correlation between specific gut microbiota and immune-related colitis.
Members of the Bacteroidetes phylum have been found to be enriched in colitis-resistant patients while the baseline gut microbiome of patients prone to developing colitis were enriched in Firmicutes (Nat Commun 2016;7:10391, Ann Oncol 2017;28(6):1368-1379). These studies provide evidence for the development of potential biomarkers to predict patients at high risk of developing immune-related colitis.
Ongoing Investigation
The ability to manipulate the microbiome through oral administration of bacteria was shown to be possible in murine model with resultant improvement in anti-tumor immune response. It remains an open question whether a similar improvement in response can be achieved in cancer patients through oral administration of probiotics or through fecal transplants from responders to non-responders.
Several clinical trials were launched around the world to address this question. These trials are currently ongoing and their results are highly anticipated in the field.
Overall the emerging role of the human microbiome in cancer immunity continues to be an exciting and developing story. Strong pre-clinical and clinical evidence supports a future role for these organisms as potential biomarkers for response and toxicities, as well as therapeutic interventions to augment immunotherapeutic agents.
AFAF E. G. OSMAN, MD, is Hematology/Oncology Fellow-PGY5, and JASON J. LUKE, MD, FACP, is Assistant Professor, both from the University of Chicago Medicine.