Precision medicine using microbiota

See allHide authors and affiliations

Science  05 Jan 2018:
Vol. 359, Issue 6371, pp. 32-34
DOI: 10.1126/science.aar2946

Accumulating evidence indicates that dysregulation of microbiota-host interactions associates with various diseases, including inflammatory bowel diseases (IBDs), colorectal cancer, diabetes, and liver cirrhosis (1). Recently, research has generated paradigm shifts in concepts about the interactions between bacteria and cancer therapeutic drugs. For example, bacteria modulate the antitumor efficacy in preclinical models of various chemotherapies (24) and immunotherapeutic agents (5, 6). Conceptually, these findings suggest that bacteria-mediated interactions with the immune system are essential for optimal drug efficacy. However, there is limited information regarding the functional impact of the composition of the human microbiome and therapeutic outcomes in cancer patients. On pages 91, 97, and 104 of this issue, Routy et al. (7), Gopalakrishnan et al. (8), and Matson et al. (9), respectively, address this important issue and demonstrate that patients can be stratified into responders and nonresponders to immunotherapy on the basis of the composition of their intestinal microbiomes, suggesting that microbiota should be considered when assessing therapeutic intervention.

Effector T lymphocytes represent a critical branch of the adaptive immune response to antigens, and controlling the length and strength of this activation is necessary to preserve healthy tissues from the destructive potential of these cells. A series of coinhibitory molecules, called immune checkpoints, expressed by antigen presenting cells are responsible for switching off T cell activation and terminating the immune response. Although this regulatory system is essential for a measured immune response, and thus for host homeostasis, expression of immune checkpoint molecules by tumor cells leads to inactivation of cytotoxic CD8+ T cells (a type of effector T lymphocyte) and, as a result, evasion of the antitumor immune response. Unleashing the power of adaptive immune responses by targeting immune checkpoints, such as the programmed cell death protein 1 (PD-1)–PD-1 ligand 1 (PD-L1) axis has emerged as a promising approach for cancer therapy in solid tumors. However, patient responses to immune checkpoint therapies are heterogeneous and transient, a phenomenon related to limited immune cell infiltration of tumors and an immunosuppressive tumor microenvironment. Because microbiota have a pronounced modulatory effect on the immune system, they may enhance responses to immune checkpoint therapies.

In an effort to identify microbes associated with treatment responsiveness, Gopalakrishnan et al. surveyed the oral and intestinal microbiota of patients with metastatic melanoma undergoing anti–PD-1 therapy. Interestingly, patients responding to this therapy had a high relative abundance of bacteria of the Faecalibacterium genus, whereas nonresponding patients displayed a high relative abundance of bacteria of the order Bacteroidales in their feces (indicating the presence of these bacteria in the intestines). They observed that the strongest fecal microbial predictors of anti–PD-1 therapy response were bacterial diversity and abundance of Faecalibacterium and Bacteroidales. No such microbial correlations were observed in the oral cavity, suggesting that the intestinal community is the source of bacterial-immune synergy for response to anti–PD-1 therapy. By also analyzing patients with metastatic melanoma undergoing this therapy, Matson et al. found responding patients had an increased abundance of eight microbial species, including Bifidobacterium longum. The presence of this species in the intestines of tumor-bearing mice was previously found to improve anti–PD-L1 therapy (6). Interestingly, two species were also associated with nonresponsiveness (Ruminococcus obeum and Roseburia intestinalis).

Routy et al. studied interactions between microbiota and response to anti–PD-1 treatment in patients with non–small cell lung cancer, renal cell carcinoma, and urothelial carcinoma. They observed that antibiotic exposure, taken during the course of cancer therapy to treat various infections, negatively correlates with patients' response to anti–PD-1 treatment. This suggests that disruption of microbial networks and loss of specific bacterial clades interfere with the efficacy of immune checkpoint blockade. Comparing the fecal microbiota of responders to nonresponders revealed increased relative abundance of Akkermansia muciniphila in patients showing favorable outcomes to anti–PD-1 treatment. Routy et al. surmised that microbial diversity and composition are a predictor of anti–PD-1 treatment response for these types of cancer. If the microbiota of responding patients containing immunoregulatory bacteria (for example, Akkermansia, Faecalibacterium, and Bifidobacterium) functionally drive anti–PD-1 efficacy, then it is expected that germ-free mice implanted with human tumor cells and transplanted with feces from these patients [a technique known as fecal microbiota transplantation (FMT)] would display better responses to treatment. Routy et al. and Gopalakrishnan et al. observed an improved response to anti–PD-1 therapy in mice receiving FMT from responsive patients compared to that in mice colonized with feces from nonresponsive patients. Gopalakrishnan et al. showed this improved response correlated with a higher abundance of Faecalibacterium in the feces of the mice. Matson et al. observed that anti–PD-L1 treatment was only effective in mice receiving FMT from responding patients, with five out of the eight strains originally found to be associated with anti–PD-1 response in patients detected in the feces of the mice, including Bifidobacterium. In all three studies, tumors of mice receiving FMT from responsive patients displayed a higher density of CD8+ T cells compared to that in tumors in mice that received nonresponder FMT. This suggests that immune checkpoint therapy targets the repertoire and activity of host immune cells and induces antitumor responses mediated both by CD8+ T cells and a decrease in CD4+ regulatory T (Treg) cells, which are immunosuppressive.

The intestinal microbiota influences the efficacy of PD-1 blockade

The enrichment of specific microbial taxa in intestines correlates with response to PD-1 blockade in cancer patients. FMT from responders into tumor-bearing mice improved responses to anti–PD-1 therapy and correlated with increased antitumor CD8+ T cells in the tumors. Mice receiving FMT from nonresponders did not respond to anti–PD-1 therapy, and tumors had a high density of immunosuppressive CD4+ Treg cells.


An important and clinically relevant issue is whether manipulation of the intestinal microbiome could turn patients that are nonresponsive to immune checkpoint blockade into responders. Routy et al. found that introduction of A. muciniphila to mice receiving human nonresponder FMT reversed the low response to PD-1 blockade, improving antitumor immune cell infiltration and activity in tumors. Overall, these studies report a fascinating interaction between intestinal bacteria and antitumor efficacy of PD-1 blockade in patients, suggesting that precision medicine strategies should include the intestinal microbiota as a potential treatment modifier (see the figure).

What is the evidence that microbiota-derived research could translate to new therapeutics? The most celebrated medical success targeting microbiota came from the field of infectious diseases, with FMT-mediated treatment of recurrent Clostridium difficile infection (CDI), a leading cause of antibiotic-associated diarrhea with an alarming increased incidence in Europe, Asia, and the United States. Patients with recurrent CDI showed a 90% clinical response rate after FMT with feces from healthy donors (10). This resulted in considerable effort from both academic and private enterprises to design synthetic microbial communities to treat various microbiota-associated diseases such as CDI and IBDs. A similar route could be envisioned for cancer treatment, where synthetic microbial communities could be assembled to optimize patient responses to immunotherapy. Intestinal microbial communities have the capacity to modulate health status by engaging various immune and nonimmune cell types through microbial-derived structures (RNA, DNA, and membrane components) and by generating an impressive network of metabolites. In addition to having individualized microbiota, patients responding to immune checkpoint therapy could be predisposed to intestinal bacterial translocation into secondary lymphoid organs, where they could mount a specific antitumor immune response to allow better synergy with treatment. Untangling components that regulate the complex interaction between microbiota and host could also lead to the generation of targeted therapy, rather than nonspecific FMT.

Although these studies collectively agreed on the critical role of bacteria in defining responsiveness to anti–PD-1 therapy, no universal bacterial species define this response. The divergence in microbial communities associated with the same therapeutic agent (anti–PD-1) may be related to the type of cancer (for example, metastatic melanoma with Faecalibacterium and non–small cell lung cancer, renal cell carcinoma, and urothelial carcinoma with Akkermansia) or patient population (U.S. versus European cohorts). However, both Matson et al. and Gopalakrishnan et al. conducted their studies using U.S.-based patient populations with metastatic melanoma, yet observed different PD-1 immunoregulatory bacterial species. The reason for this microorganism specificity is unclear, but it is likely that a convergent mechanism, which is yet to be identified, unifies them. Detailed mechanistic studies into how bacteria reenergize tumor immune microenvironments will be necessary to fully comprehend this phenomenon. Another intriguing observation is that strains of Akkermansia, Faecalibacterium, and Bifidobacterium have been associated with anti-inflammatory responses, a regulatory arm of the immune system that aims to prevent overactivation of the immune response and restores host homeostasis. For example, decreased relative abundance of A. muciniphila in the intestine associates with many diseases, including IBDs, type 2 diabetes, and liver diseases (11). Similarly, Faecalibacterium prausnitzii down-regulates intestinal inflammation, which is associated with the production of specific metabolites, such as butyrate and salicylic acid derived from host cells or bacteria in the intestine and peripheral blood (12). Clearly, molecular characterization of intestinal Akkermansia, Faecalibacterium, and Bifidobacterium strains from cancer patients is needed to fully understand how they influence the tumor microenvironment and synergize with immune checkpoint blockade. These studies could lead to the isolation and characterization of microbial components that are responsible for the beneficial effects. For example, administration of Amuc_1100, a protein isolated from the outer membrane of A. muciniphila, reproduces the beneficial effect of the bacterium on diabetes in preclinical models (13).

The relationship between microbial communities and antitumor drug responses are complex. On the one hand, depletion of selective bacterial taxa by means of antibiotic exposure or other stressor conditions may diminish immunotherapy responses. On the other hand, the presence of specific microorganisms in local or distant sites may interfere with treatment through metabolic activities (14). For example, bacteria of the Enterobacteriaceae family, such as Escherichia coli strains, decrease efficacy of the chemotherapeutic agent gemcitabine by metabolizing and deactivating the active form of the drug, thereby negatively interfering with tumor response (15). Therefore, the presence of specific strains of bacteria may be able to modulate cancer progression and therapeutics, raising the possibility that precision medicine directed at the microbiota could inform physicians about prognosis and therapy. One could view the microbiota as a treasure trove for next-generation medicine, and tapping into this network may produce new therapeutic insights.

References and Notes

Acknowledgments: C.J. is supported by National Institutes of Health grant R01 DK73338 and by the University of Florida, Department of Medicine Gatorade Fund.

Navigate This Article