Mixed-donor faecal microbiota transplantation was associated with increased butyrate-producing bacteria for obesity

We read with interest the recent article by Haifer et al (Gut, 2022, 2022–3 27 742), which reported that donor gut microbiome stability and species evenness were associated with higher donor species engraftment in patients with UC following faecal microbiota transplantation (FMT). This has brought us one step closer towards the selection of optimal FMT donors. However, the high prevalence of extended‐spectrum beta‐lactamase organisms and the COVID‐19 pandemic have restricted the recruitment of FMT donors. An alternative means to increase the stability and species evenness is to pool the stool samples from multiple eligible FMT donors, which has been shown to be associated with higher clinical efficacy in UC. In obesityrelated metabolic disorders, outcomes following FMT have been variable. Although the underlying mechanisms are unclear, the efficacy of FMT is likely to be affected by both donor and recipient microbiota composition. 9–12 Therefore, we performed a pilot study of singledonor faecal microbiota transplantation (sFMT) and compared the findings with our previous mixeddonor faecal microbiota transplantation (mFMT) study, with the aim to determine microbiome factors associated with response to FMT in obese subjects (NCT03789461). Nine obese subjects received faecal infusions from a single lean donor, whereas 38 obese subjects received faecal infusions from two to five lean donors (table 1 and online supplemental table S1). Overall, FMT was safe and well tolerated (online supplemental table S2). After FMT, 13.2% of mFMT recipients (responders) and none of the sFMT recipients achieved ≥10% weight loss (mean±SD, mixeddonor 3.1%±4.8%, singledonor 4.8%±1.7%; online supplemental figure S1). mFMT recipients showed significant reductions of lowdensity lipoprotein and total cholesterol at weeks 24 and 52 since the first FMT infusion (p<0.05, Wilcoxon ranksum test; online supplemental figure S1) and a substantial increase in butyrateproducing bacteria including Eubacterium sp, Roseburia hominis, Anaerostipes hadrus, Faecalibacterium prausnitzii and Collinsella sp (figure 1A and online supplemental figure S2–S4; linear discriminant analysis (LDA) >2, p<0.05). Increase of these organisms is associated with improved glucose and cholesterol metabolism. 14 The Chao1 richness, Shannon Diversity Index and the aggregated abundance of butyrateproducing species were significantly higher after FMT compared with baseline (p<0.01 and p<0.05, Wilcoxon signedrank test; figure 1B,C). In contrast, there was no significant increase in bacteria richness, diversity or abundance of butyrateproducing species in sFMT recipients (figure 1A–C and online supplemental figure S3). Previous sFMT studies also showed an increase in only a few butyrateproducing species. 15 16 Abundance of Bifidobacterium bifidum, which has been shown to interact with butyrateproducing bacteria by crossfeeding interactions, was found to be significantly increased in mFMT recipients (LDA >2, p<0.05). Network analysis showed that the change of major butyrateproducing species showed significant positive correlation with each other at both followup timepoints (figure 1D and online supplemental figure S5), indicating that these species belonged to an ecogroup that maintained covariation following FMT. The richness of overall bacteriome also significantly increased in mFMT recipients (p<0.05, Wilcoxon signedrank test; figure 1E) but not in sFMT recipients. Despite the same selection criteria, the microbiome profile of the lean donors varied largely in terms of the presence and abundance of butyrateproducing species (online supplemental figure S6). Coinfusion of butyrateproducing species by pooling stool samples from multiple donors may therefore promote the colonisation of butyrateproducing species in the recipients’ gut. This was in line with the findings by Haifer et al that donor species evenness was associated with desirable outcomes. Interestingly, the weight loss of the responders lasted at least 6 months since the first FMT infusion. At baseline, no significant difference in clinical aspects was found between responders and the rest of the recipients. Responders had enriched Bacteroides dorei, among other 18 species, and depleted Intestinimonas butyriciproducens, compared with the rest of the recipients (LDA >2, p<0.05; online supplemental figure S7). As obesity is a complex, multifactorial disease with genetic, behavioural and environmental origins, FMT may only be effective for a subset of obese patients that were, at least partially, driven by the gut microbiota. More studies are needed to identify recipient factors associated with favourable outcomes. Letter

Mixed-donor faecal microbiota transplantation was associated with increased butyrateproducing bacteria for obesity We read with interest the recent article by Haifer et al (Gut, 2022(Gut, , 2022, which reported that donor gut microbiome stability and species evenness were associated with higher donor species engraftment in patients with UC following faecal microbiota transplantation (FMT). This has brought us one step closer towards the selection of optimal FMT donors. However, the high prevalence of extended-spectrum beta-lactamase organisms and the COVID-19 pandemic have restricted the recruitment of FMT donors. 1 An alternative means to increase the stability and species evenness is to pool the stool samples from multiple eligible FMT donors, which has been shown to be associated with higher clinical efficacy in UC. 2 In obesity-related metabolic disorders, outcomes following FMT have been variable. [3][4][5][6][7][8] Although the underlying mechanisms are unclear, the efficacy of FMT is likely to be affected by both donor and recipient microbiota composition. 4 9-12 Therefore, we performed a pilot study of single-donor faecal microbiota transplantation (sFMT) and compared the findings with our previous mixed-donor faecal microbiota transplantation (mFMT) study, 8 with the aim to determine microbiome factors associated with response to FMT in obese subjects (NCT03789461).
Nine obese subjects received faecal infusions from a single lean donor, whereas 38 obese subjects received faecal infusions from two to five lean donors (table 1 and  online supplemental table S1). Overall, FMT was safe and well tolerated (online supplemental table S2). After FMT, 13.2% of mFMT recipients (responders) and none of the sFMT recipients achieved ≥10% weight loss (mean±SD, mixeddonor 3.1%±4.8%, single-donor 4.8%±1.7%; online supplemental figure S1). mFMT recipients showed significant reductions of low-density lipoprotein and total cholesterol at weeks 24 and 52 since the first FMT infusion (p<0.05, Wilcoxon rank-sum test; online supplemental figure S1) and a substantial increase in butyrateproducing bacteria including Eubacterium sp, Roseburia hominis, Anaerostipes hadrus, Faecalibacterium prausnitzii and Collinsella sp (figure 1A and online supplemental figure S2-S4; linear discriminant analysis (LDA) >2, p<0.05). Increase of these organisms is associated with improved glucose and cholesterol metabolism. 13 14 The Chao1 richness, Shannon Diversity Index and the aggregated abundance of butyrate-producing species were significantly higher after FMT compared with baseline (p<0.01 and p<0.05, Wilcoxon signed-rank test; figure 1B,C). In contrast, there was no significant increase in bacteria richness, diversity or abundance of butyrate-producing species in sFMT recipients (figure 1A-C and online supplemental figure S3). Previous sFMT studies also showed an increase in only a few butyrate-producing species. 4 15 16 Abundance of Bifidobacterium bifidum, which has been shown to interact with butyrate-producing bacteria by cross-feeding interactions, was found to be significantly increased in mFMT recipients (LDA >2, p<0.05). Network analysis showed that the change of major butyrate-producing species showed significant positive correlation with each other at both follow-up timepoints (figure 1D and online supplemental figure S5), indicating that these species belonged to an ecogroup 17 that maintained covariation following FMT. The richness of overall bacteriome also significantly increased in mFMT recipients (p<0.05, Wilcoxon signed-rank test; figure 1E) but not in sFMT recipients.
Despite the same selection criteria, the microbiome profile of the lean donors varied largely in terms of the presence and abundance of butyrate-producing species (online supplemental figure S6). Coinfusion of butyrate-producing species by pooling stool samples from multiple donors may therefore promote the colonisation of butyrate-producing species in the recipients' gut. This was in line with the findings by Haifer et al that donor species evenness was associated with desirable outcomes.
Interestingly, the weight loss of the responders lasted at least 6 months since the first FMT infusion. At baseline, no significant difference in clinical aspects was found between responders and the rest of the recipients. Responders had enriched Bacteroides dorei, among other 18 species, and depleted Intestinimonas butyriciproducens, compared with the rest of the recipients (LDA >2, p<0.05; online supplemental figure S7). As obesity is a complex, multifactorial disease with genetic, behavioural and environmental origins, FMT may only be effective for a subset of obese patients that were, at least partially, driven by the gut microbiota. More studies are needed to identify recipient factors associated with favourable outcomes. Table 1 Subject baseline characteristics in each study

Letter
One limitation is that data were not a direct comparison; however, both trials included obese recipients of ethnically Chinese origin within a single centre during the same time. Furthermore, recipients underwent FMT from same-pool lean donors, and samples were processed, sequenced and analysed according to the same standardised protocol. These findings suggest that microbiome engraftment after FMT occurred differently using single or multiple donors and shed new light in facilitating the future study design of microbe-based therapies for obesity. (C) Aggregated abundance of butyrate-producing bacteria in both FMT regimens were depicted. (D) Network plot depicting the association of butyrate-producing bacteria following mFMT. (E) Chao1 richness and Shannon Diversity Index in both FMT regimens were depicted. Significance between studies was calculated by Wilcoxon rank-sum test. Significance within the same FMT regime was calculated by Wilcoxon signed-rank test. FMT, faecal microbiota transplantation; mFMT, mixed-donor faecal microbiota transplantation; sFMT, single-donor faecal microbiota transplantation.