Fecal Microbial Transplantation For Diseases Beyond Recurrent Clostridium Difficile Infection

Abstract

As microbiome research has moved from associative to mechanistic studies, the activities of specific microbes and their products have been investigated in development of inflammatory bowel diseases, cancer, metabolic syndrome, and neuropsychiatric disorders. Findings from microbiome research have already been applied to the clinic, such as in fecal microbiota transplantation (FMT) for treatment of recurrent Clostridium difficile infection. We review the evidence for associations between alterations in the intestinal microbiome and gastrointestinal diseases and findings from clinical trials of FMT. We discuss opportunities for treatment of other diseases with FMT, based on findings from small clinical and preclinical studies.

Following the completion of human genome project in 2003, scientists set their eyes on the next big genomic challenge—mapping microbial communities throughout the human body. For example, the HMP consortium analyzed bacterial communities of up to 18 body sites of 242 adults (129 men and 113 women, 18–40 years old) using 16S rRNA and whole-genome metagenome sequencing technologies (1,2). Similar microbiome cataloging efforts were pursued throughout the world, with for example a European consortium (Meta-Hit) focusing on intestinal microbiome composition in health and disease (3). Humans are colonized from birth with microorganisms that rapidly assembled into a complex community comprising archaea, bacteria, viruses, and fungi. It was thought that the combination of microbiome and genomic data would provide a powerful holistic view of the human superorganism that narrowed the gap between basic and clinical research. For example, studies of the human microbiome, combined with genetic information, could increase our understanding of susceptibility to immune disorders (4–6). These efforts have produced a large amount of data and contributed significant resources to the scientific community. However, findings from large microbiome analyses are only slowly being applied to the clinic.

The mammalian microbiome is dominated by bacteria, of 6 phyla: Actinobacteria, Bacteroidetes, Cyanobacteria, Firmicutes, Proteobacteria, and Fusobacteria. The relative abundance of these phyla varies not only among body sites (such as intestine vs skin), but also among individuals, depending on age, diet, health status, and location. The intestinal microbiome has a symbiotic relationship with its host, contributing to energy and nutrient extraction from the diet, development of the immune response, intestinal mucosal barrier integrity, and metabolism of xenobiotics (7–10).

Alterations in the microbiome, determined from analyses of fecal and intestinal samples, have been associated with gastrointestinal disorders such as ulcerative colitis (UC), irritable bowel syndrome (IBS), and constipation (11–13) as well as extraintestinal disorders such as cancer, metabolic syndrome, and neuropsychiatric disorders (14–16). Not surprisingly, efforts directed at modulating intestinal microbiota for therapeutic purposes have attracted wide attention from the scientific community.

This interest in fecal microbiota transplantation (FMT) is not recent—it was described by Ge Hong in 4^th century China, who described the oral ingestion of fecal material for treatment of severe diarrhea (17). In 1958, the American surgeon Ben Eiseman reported that 4 patients with antibiotic-induced diarrhea rapidly improved after administration of fecal enemas (18). However, this field of research was put on hold for more than 50 years—the first controlled randomized trial of FMT for recurrent Clostridium difficile-induced colitis was performed by investigators in Amsterdam (19). FMT is now used worldwide to successfully treat patients with recurrent C difficile infection (CDI) that does not respond to antibiotics.

We review the potential for microbiota-based therapies for gastrointestinal disorders, focusing on FMT. For reviews of the clinical effects of probiotics, prebiotics, antibiotics, or bacteriophage on the intestinal microbiota (see ref 20). Research into the human intestinal microbiome and its dysregulation during disease pathogenesis is progressing rapidly. Due to length and topic restriction, we have not addressed all microbiota-associated diseases and excluded interventions in which the microbiota is manipulated with other methods than FMT.

Gastrointestinal Disorders

UC

The most evidence for the benefits of FMT, beyond treatment of CDI, comes studies of patients with inflammatory bowel diseases (IBD). The composition of the microbiota in the feces of patients with UC differs from that of healthy individuals (controls) in diversity is usually lower, with reduced relative abundance of Firmicutes (Clostridium clusters XIVa and IV) and Bacteroidetes. In particular, lower abundance of the butyrate producer Fecalibacterium prausnitzii was associated with UC, as well as an overrepresentation of Proteobacteria and Actinobacteria (21). This microbe imbalance leads reductions in short-chain fatty acids (SCFAs), predominantly butyrate, which are essential nutrients for colonocytes and important immune regulators. Diversion of the fecal stream by stoma can cause a type of colitis that has been treated successfully with topical butyrate preparations (22).

Four controlled trials of the effects of FMT in patients with active UC were published—3 reported positive outcomes (Table 1). Moayyedi et al randomly assigned 70 patients with active UC to groups that received allogeneic FMT or water (placebo), administered in 6 weekly enemas (23). The primary endpoint, a total Mayo score <3 and endoscopic healing (endoscopic Mayo score 0), was achieved by 24% of the patients who received FMT vs 5% of patients who received placebo. Interestingly, most patients who responded to the therapy received FMT from a single super-donor (39% of recipients met the endpoint vs 10% of the other donors), suggesting a donor effect, which has not been observed for CDI treatment. Patients who underwent FMT had greater microbial diversity, based on analyses of fecal samples, than patients who received the placebo.

TABLE 1:

controlled clinical trials with FMT for common gastrointestinal disorders

Study	Indication	Intervention	Control Group	Study size	Effect size
Moayyedi et al. (22)	Ulcerative colitis	allegeneic FMT weekly enemas 6 week	Water enemas	70	19%
Rossen et al. (23)	Ulcerative colitis	duodenal infusion of donor stool	Autologous feces	50	10%
Paramsothy et al. (24)	Ulcerative colitis	colonoscope injection+ 5 enemas (pooled)	Saline	81	19%
Costello et al. (25)	Ulcerative colitis	colonoscope injection+enemas (pooled)	Autologous feces	73	23%
Misra et al. (27)	Ulcerative colitis	Vanco+SER-287		58	40%
Johnsen et al. (36)	Irritable Bowel Syndrome	colonoscope injection (2 mixed donors)	Autologous feces	90	22%
Holvoet et al. (37)	Irritable Bowel Syndrome	colonoscope injection (2 mixed donors)	Autologous feces	64	30%
Halkjær et al. (38)	Irritable Bowel Syndrome	FMT capsules	PLC Capsules	52	NONE
Tian et al. (39)	Slow transit constipation	duodenal infusion of donor stool	PLC	60	33%

A single-center study in Amsterdam studied a different approach for treatment of UC. Based on the reported effects of duodenal infusion of fecal material into patients with Clostridium-induced colitis (19), Rossen et al gave 25 patients with active UC (simple clinical colitis index scores of 4–11) 2 duodenal infusions of 500 ml fresh allogeneic fecal material (or autologous fecal material as control), 3 weeks apart (24). Twenty-five patients received their own feces via the same route as controls. The primary endpoint, clinical response (simple clinical colitis index scores ≤ 2 and mucosal improvement ≥ 1 point), was attained by 30% of patients who received allogeneic fecal material vs 20% of patients who received autologous fecal material, although the difference was not significant. Interestingly, an increase in microbial diversity (Shannon index) was observed in responders to allogeneic or autologous fecal material, but no super donor was identified in this study.

Paramsothy et al administered feces through a colonoscope followed by repeated enemas to patients with mild to moderate UC (25). The patients used colonic lavage prior to FMT and received feces from 3 to 7 pooled donors, rather than from a single donor. Steroid-free clinical remission with endoscopic response or remission was achieved in 1¼1 patients (27%) who received active fecal material and ¾0 (8%) who received the placebo (isotonic saline). Microbial diversity increased in responders and lack of remission was associated with increased relative abundance of Fusobacterium. Costello et al used the same mode of administration with fewer enemas but preserved the donor feces from pooled donors in anaerobic conditions. The outcome was comparable to that of the study by Paramsothy et al (remission in 32% of patients who received fecal material vs 9% who received placebo) (26).

FMT therefore appears to benefit certain patients with UC. However, the net benefit is limited and there is clear need for a better understanding of the kinetics of this treatment. Is duodenal infusion equal or superior to colonic administration? Is pre-treatment with colonic lavage or even antibiotics useful? Are anaerobic conditions necessary to preserve the feces? What types of microbes are included in samples from the super donor, compared with other less effective donors? To what extent is the change in colon flora permanent and are repeated treatments necessary to maintain remission? Which microbes mediate the therapeutic effects of FMT?

Transfer of fecal filtrate, by single nasojejunal administration, to 5 patients with chronic-relapsing CDI restored normal stool habits and eliminate symptoms (27). The fecal filtrate contains bacterial products and other microbes, but not intact bacteria. This finding indicates that the effects of FMT might not require intact bacteria, but instead involved other microorganism, such as bacteriophages, spores, or viruses, or other molecules present in fecal material.

The value of antibiotic pre-treatment was addressed in a randomized controlled trial by Bharat et al (at Seres Therapeutics, in Boston, MA). SER-287 is an investigational product that contains bacterial spores of Firmicutes (28). Patients with ulcerative colitis given vancomycin had higher engraftment of microbes from SER-287 and better outcomes following administration of the product. Antibiotic pretreatment might therefore increase the effects of FMT or other microbe-based therapies. However, further studies of this effect are needed.

Crohn’s disease (CD)

Given the promising effects of FMT in patients with UC, its effects were studied in patients with CD. However, the available evidence of efficacy of FMT for patients CD is weaker than for patients with UC patients and there have been no controlled studies. Only uncontrolled case series studies of 6–30 patients with CD have been published, with poor documentation of outcomes and a great variety of reported effects (29,30). Intriguingly, patients in some of the studies reported high fevers a few hours following FMT. One patient in a trial in Belgium developed aspiration pneumonia as a serious complication. No further trials of the effects of FMT are known to be underway in patients with CD.

Pouchitis

Development of pouchitis following restorative proctocolectomy in patients with refractory IBD is believed to be mediated by bacteria— pouchitis develops only after restoration of the intestinal continuity and usually resolves following administration of antibiotics such as metronidazole and ciprofloxacin (31). Moreover, probiotics have been reported to be effective in treatment of pouchitis, indicating that it responds to changes in the microbiome. (32,33).

In a case series of 8 patients with chronic pouchitis, FMT provided no clear benefit administered via the nasogastric route, although 2 patients regained sensitivity to ciprofloxacin therapy (34). Changes in the composition of the pouch microbiome were observed. Although optimization of the microbiome of the pouch is an attractive approach to reduce morbidities of patients with chronic or recurrent pouchitis, further studies are needed in this difficult population.

IBS

The prevalence of IBS is as high as 20% in areas of the United States (35). IBS is characterized by a variety of symptoms and there is great need for effective treatments. Patients with IBS patients have reduced diversity of the intestinal microbiome, with increased abundance of enterobacteriaceae and relatively lower levels of bifidobacteria and lactobacilli. Lack of butyrate production and increased amounts of acetic and propionic acids have been associated with bloating (36).

IBS is the only disease, beside UC, in which several prospective and controlled clinical trials of FMT have been performed (Table 1). In Norway, 90 patients with IBS with predominant diarrhea or diarrhea and constipation were randomly assigned (2:1) to groups that received FMT or placebo. The response rates at 3 months were 43% in patients who received placebo (quite high) but 65% in patients who received FMT (mixed feces from 2 donors) (a significant increase) (37). Holvoet et al studied 64 patients with IBS with predominant bloating without constipation. Patients were randomly assigned (2:1) to receive FMT (from 2 donors) via colonoscopy or their own feces (controls). The patients who received FMT from the donors had significant reductions in discomfort, abdominal pain, and flatulence, but not patients who received their own fecal samples (38). The microbiome analysis from this study has not been published, but Holvoet et al previously reported that patients with IBS with a response to FMT had higher baseline concentrations of Streptococcus and higher enrichment of the microbiome than non-responders (39).

Halkjær et al studied the effects of fecal microbiota capsules, vs placebo, in patients with IBS. Patients who received the fecal material capsules had an increase in microbial biodiversity, based on analyses of their feces, but, interestingly, symptom improvement was greater in patients who received the placebo (40). In conclusion, although there is some evidence from randomized controlled trials that administration of fecal material is effective for patients with IBS, detailed analyses of microbiome profile are needed before and after administration. Researchers should study changes in composition of microbiomes most associated with symptom improvement.

Other benign GI diseases

Constipation could be associated with dysbiosis of the intestinal microbiome, although there is weaker evidence for this than for other diseases. The effects of conventional constipation treatment were compared with those of 6 sessions of FMT (naso-enteric administration, along with conventional treatment) in 60 adults with slow-transit constipation (Table 1) (41). Patients who received FMT had significant reductions in symptoms and increases in stool consistency and colonic transit time compared to patients given conventional treatment. Further research is warranted in this field. Trials of lactobacilli and bifidobacteria are underway or completed (42).

Doki et al investigated the association between changes in the microbiome and development of graft vs host disease (GVHD) following allogeneic hematopoietic stem cell transplantation. Although the diversity of the intestinal microbiome did not differ among patients who did and did not develop GVHD, patients who did develop GVHD had a significantly higher abundance of Firmicutes and a lower abundance of Bacteroidetes (43). Although this study did not adequately account for diet, the authors concluded that maintenance treatment with Bacteroidetes throughout hematopoietic stem cell transplantation might prevent GVHD. Based on those observations, a number of uncontrolled, small experiments were performed and indicated that FMT is safe and effective for patients undergoing hematopoietic stem cell transplantation, increasing microbial diversity. Higher abundance of Eubacterium limosum reduced risk relapse or progression of GVHD in 541 patients who underwent allogeneic stem cell transplantation and were followed for 2 years, which validated these uncontrolled findings uncontrolled findings (44–46).

Use broad-spectrum antibiotics is often associated with changes in bowel habits, most commonly watery diarrhea, called post-antibiotic colitis. Some researchers have suggested that probiotic preparations might prevent this complication (47). Restitution of the mucosal intestinal microbiome was studied in mice and humans given antibiotics, followed by autologous transplantation of feces (collected prior to the use of antibiotics) or administration of multi-strain probiotics. Whereas autologous feces induced rapid normalization of the microbiome (within days, measured in fecal samples), probiotics delayed normalization (48). FMT might therefore benefit patients with post-antibiotic diarrhea, but larger controlled trials are needed.

Hepatic encephalopathy

End-stage liver cirrhosis often leads to portal hypertension and recurrent hepatic encephalopathy (HE), which often requires hospitalization and has been associated with dysbiosis of the microbiome. In a small open-label randomized trial, 20 male outpatients with cirrhosis received 5 days of broad-spectrum antibiotics followed by FMT, from a single donor and administered via a 1 enema, or the standard of care. Interestingly, the donor was selected based on machine learning data aiming at the highest relative abundances of Lachnospiraceae and Ruminococcaceae among a universal stool donor bank. Encephalopathy recurred in 5/10 patients given the standard of care but in none of 10 patients who underwent FMT; FMT was also reported to improve cognitive function (49). No change in fecal microbiome diversity indices was observed in patients given the standard of care, but changes were observed in patients given FMT (a relative increase in Lactobacillaceae and Bifidobacteriaceae). Further studies of these interesting effects are needed.

Inflammatory Disorders Outside the Gastrointestinal Tract

Psoriasis

Psoriasis is a common inflammatory condition of the skin with a prevalence of approximately 3% worldwide and pathophysiologic similarities with IBD. There is little evidence for the efficacy of FMT in patients with psoriasis. There is evidence that the microbiota of the skin affects the development and severity of psoriasis and possibly also response to therapy (50). The relative abundance of Akkermansia mucinophila appears to be reduced in intestinal microbiota of patients with psoriasis (51) and the ratio of Firmicutes:Bacteroidetes was 3-fold higher in patients with psoriasis compared to controls (52); this disturbance correlated with the psoriasis severity score PASI.

Successful therapies for psoriasis, such as balneotherapy and narrow-band ultraviolet B radiation, have been associated with alterations in the skin microbiota (53). Studies of strategies to alter gut microbiomes of patients with psoriasis are underway. These include a randomized, placebo-controlled trial of FMT into the small intestine of patients with psoriatic arthritis or active peripheral disease that has not responded to methotrexate (54).

Central Nervous System Diseases

Multiple sclerosis

Multiple sclerosis is a chronic autoimmune disease characterized by demyelination and serious neurological disability—few effective treatments available. Many patients with multiple sclerosis have gastrointestinal symptoms and alterations in the intestinal microbiome, compared to healthy individuals (controls) (55). Studies in animal models have established a role of the gut microbiome in disease progression. In mice with autoimmune encephalomyelitis (EAE), inflammatory responses were attenuated under germ-free conditions (56) and reduced by strains butyrate-producing or other bacteria. Butyrate can induce epigenetic modifications such as acetylation of the Foxp3 locus to produce anti-inflammatory effects (57). Multiple sclerosis might be treated with microbes that produce these SCFAs, as in patients with UC. Two prospective trials with FMT are underway.

Parkinson disease

Parkinson disease is an intractable neurodegenerative disorder that has been associated with gastrointestinal conditions such as constipation and IBD. Patients with IBS have an increased risk for developing Parkinson disease (58). The intestinal microbiome of patients with Parkinson disease is characterized by an overabundance of Bacteroidetes, F prausnitzii, Enterococci, Prevotella, and Clostridium species—these alterations are associated with a poor course of Parkinson disease (59). Microbiota transplanted from patients with Parkinson disease into a mouse model of the disease led to worsening of neurological manifestations whereas depletion of gut microbiota in the same model reduced neurologic symptoms (60). Brain derived neurotrophic factor (BDNF) mediates interactions between intestinal cells and the nervous system. BDNF produced by the gut microbiome and is reduced in patients with Parkinson disease, leading to neurodegeneration (61). Changes in the microbiome that increase synthesis of BDNF might reduce symptoms and slow progression of Parkinson disease, but there are few data from controlled studies. Studies to alter the microbiome of Parkinson’s patients are underway.

Autism

Autism is often associated with constipation, bloating, diarrhea, and alterations in the intestinal microbiome. Children with autism had a reduced ratio of Bacteroidetes:Firmicutes (62). Changes in the microbiome might interfere with the tryptophan metabolism to alter behavior. In a pilot trial of FMT for children with autism, preceded by 2 weeks of antibiotics, behavioral symptoms improved in parallel with intestinal symptoms (bloating, constipation, diarrhea); these improvements were maintained for more than 8 weeks after FMT. Overall bacterial diversity increased, with increased abundances of Bifidobacterium, Prevotella, and Desulfovibrio observed for more than 8 weeks after FMT (63). This observation indicates a link between the microbiome and behavior, but results are preliminary. However, in a cohort of simplex families with autism spectrum disorder (ASD) and neurotypical siblings, there was no significant difference detected in diversity or composition of fecal microbiomes of children with ASD vs their siblings without ASD (64). Further studies are needed before FMT can be recommended treatment of autism.

Cancer

Although FMT has not been tested in the patients with cancer, there is great opportunity for ecosystem manipulation for this pathology. A link between carcinogenesis and microorganisms was established decades ago, with for example development of various form of cancers including lymphoma, leukemia, gastric cancer, and hepatocellular carcinoma, following infection with class 1 carcinogenic microorganisms such as Helicobacter pylori, hepatitis B or C viruses, Epstein-Barr virus, or Kaposi sarcoma herpes virus (65). However, over the past decade there has been tremendous progress in our understanding of the role of the entire microbiome in carcinogenesis, as opposed to single microorganisms.

Due to the size and diversity of the intestinal microbiome, it is not surprising that its relationship with colorectal cancer (CRC) has been widely studied (66). Experiments involving transfer of microbial communities from one host to another have demonstrate diseased transmissibility and protection. Fecal microbiota transfer experiments can be performed either passively, through coprophagy (co-housing), or actively, through oral-gastric feeding. For example, mice with disruption of the nucleotide-binding oligomerization domain-containing protein 2 gene (Nod2) gene, which encodes a protein that recognizes bacterial molecules and stimulates the inflammasome pathway, are more susceptible to colitis-associated colorectal cancer than wild-type mice (67). Interestingly, the risk of colitis is increased in wild-type mice passively exposed (co-housed) to fecal microbiota from NOD2-deficient mice. Wild-type mice passively exposed to fecal biota from NLRP6-deficient mice have increased susceptibility to colitis-associated cancer (68). NLRP6 is also part of the inflammasome pathway.

The ability of fecal microbiota from Nod2^−/− mice to induce colitis in wild-type mice was associated with changes in Bacteroides, Butyrivibrio, and Lachnobacterium communities. Since bacteria do not have effects on NLRP6 signaling (69), it is not clear which microorganisms determine susceptibility to colitis. Importantly, transplantation of fecal microbiota from healthy wild-type mice reduced development of colitis in Nod2^−/− mice. The DNA-sensing molecule AIM2 protects against colorectal carcinogenesis (70,71). Passive exposure of Aim2^−/− mice to fecal microbiota from wild-type mice reduced tumor development (70), supporting the concept that a component of the microbiota could prevent colorectal carcinogenesis in animal models.

The importance of the microbial ecosystem in CRC development was demonstrated in an elegant study in which a pool of fecal materials obtained from healthy subjects or patients with CRC was transferred to germ-free wild-type mice or wild-type mice exposed to the carcinogenic compound azoxymethane (72). The proportion of mice with polyps and numbers of colon polyps were significantly higher in mice that received fecal materials from patients with CRC than in mice that received fecal microbiota from healthy subjects. Interestingly, germ-free, wild-type mice that received fecal material from patients with CRC or healthy individuals did not develop polyps, indicating that genetic factors affect the ability of microbes to promote carcinogenesis.

It is important to note that the influence of intestinal microbiota on carcinogenesis extends beyond the intestine. Researchers demonstrated a functional link between intestinal microbiota and the development of pancreatic cancer in mice (73,74). One study showed that development of pancreatic ductal adenocarcinoma (PDAC) was prevented when Pdx1-Cre; LSL-Kras mice were bred under germ-free conditions (74). Moreover, antibiotics reduced development of PDAC in Pdx1Cre;LSL KrasG12D;Trp53R172H (KPC) mice, whereas fecal transferred of KPC-derived feces, but not feces from wild-type mice accelerated tumorigenesis (74). A study of Kras^G12D; PTEN^lox/+ mice showed that PDAC progression was attenuated when the intestinal bacterial community was depleted with antibiotics, compared to microbiota-intact mice (73).

Changes in the intestinal microbiota have also been associated with liver cancer progression, affecting metabolism of bile acid from primary to secondary structures. Secondary bile acids inhibit recruitment of natural killer T cells, which have anti-tumor effects, to the liver (75). Mice with disruption of the tet methylcytosine dioxygenase 2 gene (Tet2^−/− mice) have preleukemic myeloproliferation (PMP). These mice have impaired intestinal barrier function, which causes bacterial translocation in the spleen and mesenteric lymph nodes, resulting in increased plasma levels of IL6 (76). This microbiota-dependent increase in IL6 promotes expansion of IL6Rα⁺ granulocyte-macrophage progenitors—a step in the development of PMP. Importantly, microbiota manipulation through germ-free husbandry conditions or introduction of antibiotics prevented and reversed PMP development inTet2^−/− mice. Overall, these findings indicate that alterations to the intestinal microbiota can promote carcinogenesis, revealing therapeutic opportunities.

Researchers have compared intestinal bacteria of healthy individuals with those of patients with CRC, to identify microbial biomarkers of cancer stage and progression (77–81). Species such as Fusobacterium nucleatum, Bacteroides clarus, Roseburia intestinalis, Clostridium hathewayi, and an undefined species named m7 were detected by quantitative PCR and associated with CRC in 2 Asian cohorts (81). Studies are needed to determine whether these markers can be used to identify patients with CRC in different populations. Furthermore, it is not clear whether any preventive action could be taken after identification of individuals at risk for cancer, based on microbe markers—we don’t know if these markers identify patients with early-stage, treatable neoplasia.

Based on evidence showing differences in the intestinal microbiomes of individuals with vs without CRC, and findings from mice that different microbiomes affect risk of colitis-associated cancers, FMT might be used to prevent CRC or slow its progression (Fig.1). FMT might be included with, or performed before or after, cancer surgery or chemotherapy. Prospective studies are needed to test the effects of FMT in patients with CRC.

Figure 1: Potential Application of FMT to Cancer Therapy.

Intrinsic and extrinsic factors can disrupt healthy intestinal microbiota and increase susceptibility to cancer. Replacement of the intestinal microbiota with FMT might be used to prevent or treat different forms of cancer including colorectal, liver, and pancreatic cancer.

Studies have associated the intestinal microbiome with response to cancer therapy (81–84), expanding this field of microbiome research from promoting to treating cancer. For example, in mice with xenograft tumors grown from P815 mastocytoma or MCA205 sarcoma cells, the anti-tumor effects of cyclophosphamide were reduced if mice were germ-free or given antibiotics (86). Furthermore, orally administered Lactobacillus johnsonii and Enterococcus hirae increased the anti-tumor effects of cyclophosphamide in mice (86). In mice with xenograft tumors grown from EL4 lymphoma or MC38 colon carcinoma cells, antibiotics reduced the cytotoxic effects of oxaliplatin and cisplatin (87). Microbes promoted the antitumor effects of CpG-oligonucleotide immunotherapy in mice with xenograft tumors grown from EL4 lymphoma, MC38 colon carcinoma, or B16 melanoma cells. These findings indicate that specific microbes, or their products, can increase the effects of cancer therapies, and that FMT might have effects in patients with cancer undergoing treatment.

Researchers have also studied the effects of the microbiota on immune checkpoint inhibitor therapy (88). Antibodies against CTLA4 did not inhibit growth of xenograft tumors derived from MCA205 sarcoma, MC38 colon carcinoma, or Ret melanoma cells in germ-free mice or mice given antibiotics compared with mice carrying a complete microbiota (89). Alterations in the intestinal microbiota modified the efficacy of antibodies against PD-L1 in mice with xenograft tumors grown from B16.SIY melanoma cells (90). Interestingly, the modulatory effect of the microbiota on therapeutic efficacy was associated with increased myeloid cell-mediated, T-helper (Th) cell-mediated (Th1 and Th17), and CD8+ T-cell responses. Although these findings were made in studies of mice, they indicate that microbes and their products might synergize with anti-cancer agents to slow tumor development.

How relevant is this research to human cancer? Studies of large cohorts of patients with advanced renal cell carcinoma (RCC) or non-small-cell lung cancer (NSCLC), from 2 different cancer centers, showed that administration of antibiotics within 30 days of anti-PD1, anti-PDL1, or anti-CTLA4 agents (alone or in combination) reduced times of progression-free and overall survival, compared to patients who did not receive antibiotics (91–92). These findings indicate that intestinal bacteria might affect patient responses to immune checkpoint inhibitors. A number of observations support this hypothesis; the microbiomes of patients with advanced RCC, NSCLC, or melanoma who respond to anti-PD1 therapy differs from than those of non-responders (92–94). Remarkably, when fecal samples from patients who responded or did not respond to anti-PD1 therapy were administered orally to germ-free mice with tumors, the mice had the same response (or lack of response) to anti-PD1 treatment (91, 93, 94). Although all 3 studies concluded that the composition of the microbiome is an important determinant of response to immune checkpoint inhibitors, the composition of the microbiomes associated with response varied among the studies. For examples, PD1 responsiveness in patients with advanced melanoma was associated with increased relative abundance of Feacalibacterium species (94) or Bifidobacterium (93), whereas Akkermansia muciniphila was associated with treatment efficacy in patients with RCC or NSCLC (91). Importantly, introduction of A muciniphila was able to reverse unresponsiveness in mice given fecal samples from non-responders (91), so specific microbes might determine the effects of certain immunotherapeutic agents. Interestingly, using datasets from these studies (91–93), researchers found microbe composition to have poor predictive power in defining PD1 responsiveness, whereas microbial gene content had better predictive performance (95).

Side effects are an important concern for strategies to manipulate the microbiomes of patients receiving immune checkpoint inhibitor therapy. For example, an adverse effect of the CTLA4 inhibitor ipilimumab is development of a Crohn’s-like colitis, observed in 8%–30% of patients (96). Dubin et al studied the intestinal microbiota of 34 patients with metastatic melanoma treated with ipilimumab and observed increased proportions of the Bacteriodetes phylum in patients who did not develop colitis (97). Whole-genome metagenome analysis of patients given CTLA4 inhibitors revealed that microbial modules for polyamine transport system and the biosynthesis of thiamine, riboflavin, and pantothenate were associated with risk of colitis (97). In a subsequent study of 26 patients with metastatic melanoma treated with ipilimumab, Chaput et al observed that patients with a high abundance of Bacteroidetes before treatment were resistant to colitis whereas patients who developed colitis had an increased abundance Firmicutes, especially of the Faecalibacterium genus (98). Interestingly, patients with a high abundance of Faecalibacterium had longer progression-free survival, revealing a double-edge sword of microbiota manipulation (efficacy vs toxicity). Not Not surprisingly, trials are underway to investigate the effects of FMT combined with cancer immunotherapy or chemotherapy (99). Interactions among the microbiota, immune response, and cancer treatments should be considered in management of patients with cancer (Fig.2).

Figure 2: Synergy of the Intestinal Microbiota With the Immune System in Cancer.

Treatment Specific microbes or their products could increase the activities of chemotherapeutic or immune checkpoint inhibitor therapy, perhaps through interactions with immune cells.

Metabolic syndrome

Metabolic syndrome comprises obesity, type 2 diabetes, hypertension, and cardiovascular diseases; pathogenesis involves a combination of genetic and environmental factors. Changes in the intestinal microbiome have been associated with development of metabolic syndrome (100). Whole-genome metagenome and 16S rDNA analyses revealed differences in microbiome composition and gene richness between feces of obese subjects and healthy lean subjects (101–103), but these differences were not large enough to distinguish between the groups (104). Higher levels of energy from diet were measured in mice colonized with intestinal microbes from obese vs lean individuals (105). In a twin study, transfer of fecal microbiota from only the obese sibling (not the non-obese sibling) to germ-free mice increased body mass and adiposity (106). A trial of 18 patients who received allogenic (n=9) FMT from lean human donors or obese patients who received (control) autologous (n=9) FMT reported improved insulin sensitivity in the group that underwent allogenic FMT, after 6 weeks (107). A subsequent larger trial (n=38) from the same research group showed no benefit of allogenic FMT (n =26) on insulin sensitivity or weight compared to autologous FMT (n =12) at 18 weeks, which correlated with no changes in the composition of the intestinal microbiome (108). Interestingly, in the same cohort, a modest beneficial effect on insulin sensitivity was observed after 6 weeks in the group given allogenic FMT, due to a subgroup of responders who had low microbial diversity at baseline compared to non-responders. This suggests that patients’ responses to microbiome manipulation might be influenced by their original microbiome

The relationship between the composition of the microbiome and metabolic function is complex and unclear. When severely obese patients underwent bariatric surgery, either with adjustable gastric bands or Roux-en-Y-gastric bypass, most patients who underwent Roux-en-Y-gastric bypass had low microbial gene richness 1 year after surgery, yet these patients had more pronounced improvements in metabolic function than patients who received the gastric bands (109), So clinical and metabolic effects do not always correlate with the composition of the intestinal microbiome. This may be related to the capacity of intestinal microbiota to produce a drastically different set of metabolites, depending on nutrient exposure, without altering their phylogeny. This was also observed in a study that compared in vegan individuals with omnivores (110). This concept is important because changes observed in the microbiomes of individuals with obesity, metabolic syndrome, or hypertension are not consistent, and may be unique to each condition.

Hypertension (when blood pressure exceeds the normal range of 120/80 mm Hg) is risk factor for cardiovascular disease. Although genome-wide association studies identified variants at as many as 120 loci that could affect blood pressure, fewer than 4% of cases of hypertension can be accounted for by all these loci, so environmental factors are likely to be involved (111–113). 16S rDNA sequence analyses revealed lower microbial diversity and richness in 10 patients with high blood pressure compared with subjects with normal blood pressure, with clear differences in principal coordinate analysis (114). In a subsequent metagenomic and metabolomic study of 41 healthy individuals (controls), 56 subjects with pre-hypertension, and 99 individuals with primary hypertension, Li et al observed decreased microbiota diversity and richness in pre-hypertensive and hypertensive patients compared to controls (115). At the genus level, principal coordinate analysis showed a distinctive clustering of microbiota between hypertensive patients and controls, mediated by the presence of Prevotella and Bacteroides, respectively. It is important to note that the microbiomes of prehypertensive vs hypertensive patients did not differ, so changes in microbiome composition might precede disease development.

The functional effects of the intestinal microbiota have been studied in animals. Germ-free mice colonized with feces from hypertensive patients had increases in blood pressure (115). Similarly, transferring feces from hypertensive susceptible SRH rats to normotensive WKR rats increased their blood pressure (114). Interestingly, the antibiotic minocycline was able to decrease blood pressure in rats with angiotensin II-induced hypertension, associated with changes in microbiota composition. It is not clear if reverse alterations in the composition of the microbiota, following therapeutic intervention, associate with functions of specific bacteria; FMT studies might investigate this.

Diet has a large effect on the composition of the intestinal microbiome, its homeostasis (116), and development of metabolic syndrome. Metabolites generated from fiber-rich diets include SCFA (butyrate, propionate, and acetate). Interestingly, a meta-analysis of clinical trials investigating the effect of fiber intake in patients with hypertension reported reduced blood pressures of subjects with high-fiber diets (117). In mice, SCFA receptors such as GPR41 and OLFR78 maintain normal systemic blood pressure (118,119). A high-fiber diet or acetate supplementation decreased high blood pressure in mice with deoxycorticosterone acetate salt-induced hypertension (120) However, not all effects of SCFAs are beneficial. For example, butyrate promotes development of CRC in mice with disruption the DNA repair gene encoding MSH2 (121), whereas acetate promotes insulin resistance and metabolic syndrome in Tlr5^−/− mice (122). Studies are needed to determine how microbe-derived metabolites affect the metabolism and development of metabolic syndrome.

Regardless, trials are underway to evaluate the effects of altering the intestinal microbiota in patients with metabolic syndrome. A phase 3 trial of 44 participants is underway to evaluate whether FMT from lean healthy donors can reduce insulin-resistance more than lifestyle changes alone in patients with metabolic syndrome (NCT02050607). A phase 2 trial (NCT02970877) of 48 participants will test when stool from healthy lean people transplanted into morbidly obese patients will improve insulin resistance and other obesity-related parameters. Manipulation of the intestinal microbiome, with FMT or by administration of specific microbes or groups of microbes, has been tested in patients with an array of medical conditions (Table 2). Analyses of data from these studies will provide important insights into disease pathogenesis and potential treatment strategies.

TABLE 2:

DISEASES FOR WHICH FMT HAS BEEN/IS BEING TESTED

INDICATION	REFERENCE	TYPE OF STUDY	EFFECT
GASTROINTESTINAL DISEASES
Ulcerative Colitis	23-26	RCT	overall positive
Crohn’s disease	29, 30	case series	no effect
Pouchitis	34	case series	no effect
Irritable Bowel Syndrome	37, 38	RCT	suggestive
Graft versus Host Disease	43-46	case series	suggestive
Post-antibiotic diarrhea	47	case series	suggestive
Constipation	41	RCT	suggestive
Hepatic encephalopathy	49	RCT	suggestive
Psoriasis		RCT ongoing	unknown
Multiple sclerosis		2 RCTs ongoing	unknown
Autism	63	uncontrolled pilot trial	suggestive
Metabolic syndrome	104.105	controlled trials	suggestive

Future Directions

Despite advances in studies of the intestinal microbiome, there have been few controlled trials of therapeutic interventions that alter the microbiome. Strategies to alter the intestinal microbiome could be used to treat a variety of gastrointestinal and other diseases, but there are many important questions to answer first. Are alterations in the microbiome associated with certain conditions the cause or consequence? How long can the effect of a microbiome intervention last, since microbiome profiles seems to be highly specific to individuals and are tolerated by the mucosal immune system? To which extent does the residing dysbiotic microbiome need to be destroyed before externally administered microbiota can successfully engraft? And, importantly, are living bacteria (or other microbes) needed for a therapeutic effect or are there other components of the microbiome, such as microbial products, that mediate the effects?

In modern medicine, FMT was first used to treat patients with C difficile-associated colitis following treatment with antibiotics. CDI is an acute/subacute condition in which the natural intestinal microbiome has been completely wiped out, making the mucosa receptive to colonization by externally administered microbiota. In the controlled prospective trials of patients with UC, it has not been clear whether pretreatment with antibiotics, such as vancomycin, increase the effects of FMT. It is important to answer this question soon. UC and IBS are chronic diseases, so we need to determine whether changes in the microbiome following FMT are permanent, or how long they last—repeated treatments are likely to be necessary. In addition, could FMT be used to maintain remission?

Findings from studies of patients with metabolic syndrome or type 2 diabetes are encouraging in light of the expanding obesity pandemic and require further exploration. Further studies are also needed to determine the effects of microbiome alterations on tumor growth and cancer therapies, not only for intestinal but also other types of cancer. Patients with graft vs host disease are often heavily treated with antibiotics, and evidence is mounting that manipulating the intestinal microbiome could improve outcomes.

A series of methodology questions must be answered for FMT. Beside the requirement for antibiotic pre-treatment, route of administration (colonic/nasoenteric), frequency of administration, and volume of fecal material required have varied among trials. The features of superdonors should be determined and will vary among diseases (Fig.3). We must also define the factors in fecal material that mediate its therapeutic effects. Answering these questions will help refine and enhance microbe-based therapies.

Figure 3: Searching for Microbe-based Therapeutic Targets.

Microbe-based therapies could replace the entire intestinal microbiota, such as in FMT (first generation), involve specific combinations of microbes (second generation), or microbe-derived compounds (third generation). Studies are underway identify the microbes or products that are altered during disease development and therefore might be therapeutically targeted, or microbes or molecules with therapeutic effects. Once identified, these require validation and prospective clinical studies.

Microbe-based therapies are likely to eventually involve small molecule compounds derived from microbes identified in mechanistic studies, or complex combinations of microbiota or synthetic cocktails (Fig.3). FMT is not a 1 size fits all strategy, and studies are required to identify components of the microbiota that have specific effects in patients with different diseases. The march toward microbe-based precision medicine is underway. We encourage funding of research in this rapidly expanding field of research