Good bacteria showing promise for the treatment of Crohn’s disease, ulcerative colitis
“A new study published in Nature Communications demonstrates that a consortium of bacteria designed to complement missing or underrepresented functions in the imbalanced microbiome of inflammatory bowel disease (IBD) patients, prevented and treated chronic immune-mediated colitis in humanized mouse models”. The study’s senior author, Balfour Sartor, MD, Midget Distinguished Professor of Medicine, Microbiology and Immunology, Co-Director of the UNC Multidisciplinary IBD Center, said the results are encouraging for future use in treating Crohn’s disease and ulcerative colitis patients.
“The idea with this treatment is to restore the normal function of the protective bacteria in the gut, targeting the source of IBD, instead of treating its symptoms with traditional immunosuppressants that can cause side effects like infections or tumors,” Sartor said.
Introduction
Chronic intestinal inflammation can be induced by multiple exogenous and endogenous signals and is mediated by immune and nonimmune cells in genetically susceptible hosts with defects in epithelial barrier function, immunoregulation, or bacterial killing. Exogenous substances including dietary products, pathogenic microorganisms, xenobiotics including antibiotics, or various combinations thereof, can trigger initial mucosal injury and/or dysbiosis to initiate acute intestinal inflammation that is perpetuated by the antigenic activities of a subset of resident microbiota1,2. Examples of these conditions are inflammatory bowel diseases (IBD), which encompass two main clinical disorders: Crohn’s disease and ulcerative colitis. Current IBD treatments primarily control inflammation through anti-inflammatory and immunosuppressive mechanisms. Some of the most successful drugs for treating IBD include infliximab, adalimumab, vedolizumab, ustekinumab, and tofacitinib, which target specific immune components to control the inflammatory process. However, these and other immune modulating drugs induce sustained, steroid-free remission in only a small subset of patients and can have multiple serious side effects, including an increased risk for serious and potentially life-threatening infections and neoplasia. In addition, these drugs do not correct upstream conditions that contribute to the chronic inflammatory mechanisms, including the leaky mucosal barrier, a pro-inflammatory gut microbiome and immunoregulatory defects.
As an alternative to anti-inflammatory and immunosuppressive therapies, microbiome-inspired live biotherapeutic products (LBPs) are being developed to treat conditions linked to chronic intestinal inflammation and increased permeability. The traditional approach for LBP discovery has been to compare the microbiomes of healthy subjects and patients suffering from a specific condition, such as IBD, to identify microorganisms that are lacking or under-represented in large databases, such as the HMP2 project3. This information, further enforced by the results from microbial association studies, is used to propose a therapeutic formulation to replenish the microorganisms that are lacking or under-represented2,4. In the case of IBD, early efforts have focused on the use of strains belonging to the Clostridium clusters IV and XIVa, which were found to successfully decrease inflammation in rodent IBD models5,6,7. Using germ-free (GF) mice inoculated with healthy human fecal material pretreated with chloroform to enrich for spore-forming bacteria, a stable 17-strain consortium was enriched from a single donor based on their ability to induce colonic regulatory T cells (Tregs)7,8. This consortium was comprised of spore-forming Clostridium cluster IV, XIVa, and XVIII strains that produced butyrate and decreased the severity of several colitis models7.
Open label application of Fecal Microbiome Transplants and enrichment-based approaches have several disadvantages. The outcome is defined by the stool sample used for the enrichment, with different samples representing different consortia with variable efficacy;9 undesirable strains/functions associated with safety risks including virulence factors and transferable antibiotic resistance functions might be present, such as the presence of enteropathogenic and Shigatoxin-producing Escherichia coli strains10 and antibiotic-resistant E. coli strains11 in FMTs, or the presence of transferable vancomycin resistance elements as found in the genome of the VE202 consortium strain Blautia coccoides VE202-06 (GenBank Accession Number Accession: PRJDB525). Furthermore, consortium modeling, as presented in this study, shows that other bacterial species besides spore-forming Clostridium bacteria provide metabolic support and additional therapeutic functions required for optimal engraftment and therapeutic performance of live biotherapeutic products in the hostile gut environment of patients with intestinal inflammation. These shortcomings can be addressed by a bottom-up rational consortium design approach that is rigorously informed by mechanistic modeling and insights from microbiome ecology and disease pathogenesis. We used this approach to combine well-characterized strains isolated from many healthy human stool samples into a consortium of metabolically interdependent strains with a variety of therapeutic functionalities being distributed in a redundant way between strains. Initially a 17-strain consortium, GUT-103, was designed around publicly available strains. GUT-103 rapidly colonized mice, restored normal function to the inflamed colon, and prevented and reversed established experimental colitis in gnotobiotic mice. Based on these proof of concept studies, a refined 11-member consortium, GUT-108, was designed around a panel of proprietary human bacterial strains that strongly engrafted and provided similar redundant protective functions. Therapeutically applied GUT-108 corrected functional dysbiosis of the inflamed gut microbiome and treated established colitis in a humanized mouse colitis model while decreasing opportunistic pathogenic bacteria, increasing resident protective bacterial groups, and restoring immunologic and metabolic homeostasis.
Discussion
Resident bacterial strains of the GUT-103 and GUT-108 consortia, designed to interdependently restore normal function to the inflamed colon, rapidly colonized gnotobiotic mice as well as ex-germ-free humanized mice that had developed a pro-inflammatory gut microbiome. These GUT-103 and 108-colonized mice exhibited the desired functional properties consistent with the rational design of the two consortia.
The two separate rationally designed consortia of intestinal bacterial strains with redundant functionalities that were previously found to be under-represented in the gut microbiome of IBD patients with active disease3 were shown to promote homeostatic immune functions and bacterial metabolism and prevent onset or progression of intestinal inflammation. We believe that this approach has the potential to maintain long term remission in a physiologic and safe manner, which should be tested in a Phase 1 clinical trial.
Testing GUT-103 against EER-induced colitis in gnotobiotic Il10−/− mice is clinically relevant to human Crohn’s disease because adherent-invasive Escherichia coli and Ruminococcus gnavus are increased in active Crohn’s disease3,31,32 and the adherent-invasive E. coli strain used was isolated from the ileum of a Crohn’s disease patient33. GUT-108 showed therapeutic efficacy to reverse established colitis in Il10−/− humanized mice by intervening 2 weeks after onset of moderate-severe intestinal inflammation, supporting a potential role for GUT-108 to treat active IBD. Moreover, the Il10−/− model of chronic pathobiont-driven T-cell-mediated chronic inflammation is more predictive of therapeutic responses in IBD patients than acute epithelial injury models such as dextran sodium sulfate2.
GUT-103 and GUT-108 were designed based on human data showing that protective functions provided by commensal bacteria are under-represented in the gut microbiome of IBD patients. These functions include the synthesis of SCFAs, indole and its derivatives, bile acid deconjugation and conversion, and competition for the critical nutrient iron and the synthesis of antagonistic molecules to control opportunistic pathogens3,34. Several animal studies have highlighted the role for SCFAs, especially propionate and butyrate, in regulatory T cell recruitment and function8,12,13,34,35,36. The recruitment in the colon and extrathymic conditioning of regulatory T cell response by SCFA make these molecules an important link in the crosstalk between the gut microbiome and the immune system. Therefore, commensal bacteria identified to produce propionate and butyrate were selected in the rational design of GUT-103 (Table 1) and GUT-108 (Table 2).
Several studies have shown the role of indole, a metabolite produced from tryptophan, and its metabolites in reducing attachment of pathogenic E. coli to epithelial cells37, strengthening the mucosal barrier and mucin stimulating production38. Therefore, commensal bacteria identified to produce indole and its derivatives were selected in the rational design of GUT-103 (Table 1) and GUT-108 (Table 2).
An inflammatory gut microbiota can result in the inefficient microbial conversion of bile salts into their primary and secondary bile acids3. IBD patients with an unbalanced gut microbiome due to inflammation have lower fecal and circulating concentrations of secondary bile acids and higher conjugated fecal bile acid concentrations than do healthy subjects39. Thus, activities essential for the conversion of primary bile acids, specifically the conversion of CA and TCDCA acid via a multistep process that includes 7-alpha-dehydroxylation by 7-alpha dehydratase (7-α-DH) or 7-alpha-hydroxy steroid dehydrogenase (7-α-HSD) activity15, are included in the strain selection for GUT-103 (Table 1) and GUT-108 (Table 2).
Competition for iron helps drive the competitiveness and establishment of microorganisms40. Therefore, GUT-103 (Table 1) and GUT-108 (Table 2) include several strains that synthesize one or more siderophores under iron-limiting conditions. Ideally, these siderophores are insensitive to inhibition by Lipocalin-2, a peptide that inhibits specific siderophores and their uptake, and is a major colonic defense system triggered by bacterial infections. Lipocalin-2 levels were increased after induction of inflammation in Il10−/− mice colonized with the EER consortium (Fig. 1c) or human fecal microbiota (Fig. 4c).
Bacteriocins, of which lantibiotics are considered a specific class, have shown great promise as new antibiotics for therapeutic application, as reviewed by Field et al.41. Thus, bacteriocin synthesis was included as a key functionality in strain selection as part the rational design process of GUT-103 (Table 1) and GUT-108 (Table 2).
The optimized 11-strain GUT-108 consortium was rationally designed to build on the proof-of-concept results obtained with GUT-103. GUT-108 went beyond members of the Clostridium clusters IV and XIVa strains, including Bacteroides and Akkermansia species. Furthermore, based on their genome analysis, strains with undesirable properties including the presence of transferable antibiotic resistances or putative virulence factors were excluded. We also omitted species that are fastidiously anaerobic, such as Faecalibacterium prausnitzii, from the GUT-108 consortium. Compared to GUT-103, GUT-108 strains provide additional redundancy for the synthesis of the protective secondary bile acids LCA and DCA, plus multiple mechanisms to compete with opportunistic pathogenic Enterobacteriaceae including synthesis of the siderophore yersiniabactin, and lantibiotics. In the inflammatory gut environment of humanized Il10−/− mice, beneficial Lachnospiraceae and Ruminococcaceae family members are decreased while opportunistic Enterobacteriaceae are increased42, as reported for Crohn’s disease patients32,43. Therapeutic application of GUT-108 reduced levels of colitogenic Enterobacteriaceae and increased beneficial resident Clostridium (Clusters IV and XIVa) species, especially Lachnospiraceae including Dorea species and Lachnoclostridium species that are not GUT-108 constituents (Fig. 4b). This altered community composition increased cecal luminal propionate concentrations, but not butyrate levels (Fig. 4f). Previous studies demonstrated that butyrate levels are not necessarily an indicator of a healthy gut microbiome, as butyrate synthesis from fermentation of amino acids such as lysine can contribute to inflammation under conditions associated with mucosal permeability44.
GUT-108 increased expression of Gpr41 and showed an upward trend for Gpr43 in Il10−/− mice humanized with a fecal transplant (Supplementary Fig. 5). As previously reported, SCFA produced by gut bacteria stimulate Tregs8 with propionate’s effect mediated through GPR4336. SCFA also mediate the function of GPR41, a key regulator that controls host energy balance45.
With its functional redundancy and metabolically interdependent auxotrophies, GUT-108 is designed to engraft and perform under a wide range of conditions. When applied to Il10−/− mice humanized with a fecal transplant, all GUT-108 strains except Clostridium scindens GGCC_0168 were established for at least 2 weeks. Clostridium scindens has been previously described as one of the essential strains necessary to convert primary bile acids into LCA and DCA. However, despite the absence of this strain, the established functional multi-strain network produced secondary bile acids, with Extibacter sp. GGCC_0201 providing the 7α-dehydratase activity required to convert CA and CDCA into the therapeutic secondary bile acids DCA and LCA, respectively. Normalizing the intestinal bile acid profile can restore intestinal epithelial stem cell function46, and increase colonic RORγ+ Treg cell counts that ameliorate host susceptibility to colitis47, while LCA stimulates Treg differentiation and inhibits Th17 cells48 consistent with GUT-108’s ability to restore secondary bile acid metabolism (Fig. 4f) and activate inducible IL-10+ RORγ FoxP3+ CD4+ Treg cells (Fig. 3f). GUT-108 stimulated regulatory (protective) immunity by increasing numbers of colonic LP IL-10-producing CD4+ T cells, B cells and DC and numbers and percentages of regulatory T cells, including inducible Tregs (IL-10+ RORγT+ FoxP3+ CD4+ cells) and IL-10+ Tregs (Fig. 3f). We further documented the anti-inflammatory effects of therapeutic GUT-108 in Il10−/− mice humanized with a fecal transplant by demonstrating that GUT-108 decreased IFN-ɣ+, IL-17α+, and IFN-ɣ+ IL-17α+ synthesizing colonic LP CD4+ TH1 and TH17 cells (Fig. 5a) and reduced expression levels of innate and Th1 and Th17 pathway cytokines, including IL-1β, IL-12p40, IL-13, IL-17α, IFNγ, and TNFα (Fig. 5b). Interestingly, GUT-108 treatment increased expression of IL-15 mRNA, a homeostatic cytokine that controls T cell inflammatory responses. Exogenous IL-15 treatment decreases IL-17α expression by Th17 cells in vitro through STAT5 enrichment at the IL-17 locus49, consistent with the ability of GUT-108 therapy to increase IL-15 gene expression and decrease IL-17α mRNA expression in colonic LP cells (Fig. 5b).
Increased intestinal bacterial metabolism of tryptophan, especially indole and its derivatives IAA and IPA, activates the Ahr pathway. AHR acts as a sensor of the microbiota community and, through its established role of modulating immune functions, maintains host-microbe homeostasis50. IPA is also a pregnane X receptor (PXR) agonist mediating its responses through TLR451. GUT-108 therapy increased both IAA and IPA levels in stool (Fig. 4f) and colonic Ahr gene expression (Fig. 5c). AHR is a critical mediator of anti-inflammatory responses to infection by bacterial pathogens and of the differentiation and function of immune cells including T cells, innate lymphoid cells, macrophages and DC28. AHR promotes the expression of the anti-inflammatory cytokine IL-10 and inhibits macrophage apoptosis, decreases the expression of inflammatory cytokines (IL-6 and TNF-α) and inhibits activation of NF-κB. Therefore, the Ahr pathway is critical to protect from excessive inflammatory cytokine expression and septic shock. In addition, Ahr pathway activation protects the mucosa during inflammation52.
Efficacy of GUT-108 in Il10−/− mice is mediated by IL-10-independent mechanisms. Further insights into these possible protective mechanisms include increased expression of metabolite sensors and mediators (Gpr41, Gpr43, Fxr, Pxr, Pparg, Fgf15, Fgf21) and pathways mediating differentiation of immune cells including Treg and Breg cells (cMaf, Il5, April, Aid, Bcl6) (Supplementary Fig. 5). For example, intestinal epithelial FGF15 is activated by bile acids serving as ligands for the nuclear receptor farnesoid X receptor (FXR). FXR/FGF15 signaling regulates bile acid homeostasis and protects against experimental colitis53. APRIL impacts immune regulatory T cells by stimulating their proliferation and survival, and directly contributing to their immune suppression54. Bcl6 induces IL-10 and follicular T helper cells and regulates the balance of innate lymphoid cells subsets55. Decreased expression of Nos2 (Supplementary Fig. 5), a part of the Ace2 -Nos2-IFNɣ biosynthesis gene cascade56, could lead to lower levels of ACE2, a key regulator to control intestinal inflammation induced by epithelial damage57. Certain viruses, including the coronaviruses SARS-CoV-1 and SARS-CoV-2, use the ACE2 protein for infecting respiratory and intestinal epithelial58. Therefore, chronic gut inflammation, as seen in type-2 diabetes and obesity59 might trigger elevated gut epithelial ACE2 levels, and therefore patients suffering from these conditions could be more sensitive for Coronavirus infection and at higher risk for complications, highlighting the importance of therapeutically targeting the pro-inflammatory gut microbiome as the underlying cause of chronic inflammation60.
In addition to analyzing fecal material, we measured cecal microbiota and in some cases cecal luminal metabolites to represent bacterial communities and function within the cecum/colon, since the cecum is one of the sites of most active inflammation (Fig. 4d). Previous rodent studies have shown broadly similar microbial patterns in cecal luminal and fecal samples61. Furthermore, we demonstrate similar cecal (Fig. 2) and fecal (Fig. 4f) secondary bile acid (LCA and DCA) responses to GUT-103 and GUT-108, respectively.
Although no animal model completely replicates all clinical features of human Crohn’s disease or ulcerative colitis, we believe that our use of a human fecal transplant in germ-free mice to initiate chronic TH1/TH17-mediated colitis and our treatment protocol of administering GUT-103 and 108 to mice with established mild to moderate inflammation replicates IBD as closely as feasible for preclinical studies. Proof of efficacy in human IBD will have to await a clinical trial.
GUT-103 and GUT-108 combine multiple modes of action to treat the upstream causes of inflammation by correcting the abnormal microbiome environment, activating various IL-10 synthesizing immune cells, lowering inflammatory responses, and restoring bacterial metabolic profiles to levels found in stool samples of healthy individuals. These overlapping protective mechanisms are predicted to maintain long term remission of IBD in a physiologic and safe manner, in contrast to most biologicals, which block downstream immune effector responses by neutralizing a single cytokine or molecule and induce immunosuppression that can be associated with increased infection and neoplasms. These integrated protective mechanisms make GUT-108 a promising novel therapy to treat a range of conditions whose pathogenesis is characterized by dysbiosis-mediated chronic intestinal inflammation and increased mucosal permeability. Besides IBD this could include graft versus host disease, hepatic encephalopathy, alcoholic liver disease, atherosclerosis, hypertension, obesity, metabolic syndrome, and type-2 diabetes mellitus.
Reference: van der Lelie, D., Oka, A., Taghavi, S. et al. Rationally designed bacterial consortia to treat chronic immune-mediated colitis and restore intestinal homeostasis. Nat Commun 12, 3105 (2021). https://doi.org/10.1038/s41467-021-23460-x
Click here to read more of the study