Introduction

The intestinal microbiome is increasingly recognized to play major roles in human physiology and health. Indeed, among its numerous functions, it modulates and shapes host immunity, is able to digest some otherwise indigestible dietary-derived nutrients [1] leading to the production of important metabolites for the host [2], including short-chain fatty acids (SCFA), and is involved in biliary acid dehydroxylation and synthesis of several vitamins [1]. Flourishing literature shows that many chronic diseases are associated with gut microbiota dysbiosis which include drastic changes in microbiota composition, function, and bacterial-derived metabolite production. This is particularly described for obesity and type 2 diabetes (T2D) [3,4,5,6].

Among reported features, in both mice [7] and humans [8], obesity is characterized by a decrease in bacterial diversity and low microbial gene richness (MGR: gene count number) in some subjects [5, 6]. In overweight to severe obesity, low MGR is associated with increased body mass index (BMI) as well as metabolic derangements including low-grade inflammation and insulin resistance [5, 6]. The proportion of low MGR concerns 23 to 40% of individuals with overweight or moderate obesity [5, 6] and increases up to 75% in severe obesity [9••]. Furthermore, a low MGR-associated gut microbiota profile shows a switch towards increased bacterial strains with pro-inflammatory properties as well as a decrease in those with anti-inflammatory roles [5, 6]. Likewise, both mice and human studies demonstrated that T2D is also characterized by a modification of gut microbiota composition and function, when compared with healthy controls [10] or with individuals with impaired glucose tolerance [3, 11]. Nevertheless, the role of altered gut microbiota as causality or consequence in metabolic diseases remains a source of discussion [12].

Several types of interventions are able to induce both beneficial metabolic outcomes and partial or complete correction of gut microbiota dysbiosis. For example, a moderately restrictive, fiber-enriched diet proposed to overweight or moderately obese individuals induced a significant increase of MGR especially in patients with baseline low MGR, and associations were found between changes in some bacterial groups and weight loss–improved metabolism [6]. Likewise, bariatric surgery reserved for the most severe forms of obesity leads to significant increases in MGR and modifies some bacterial species concomitantly with improvement in metabolic alteration and major weight loss [9••]. Although based on associations, these studies suggest that the modulation of gut microbiota per se could at least be partly involved in metabolic improvements. The role of gut microbiota modulation after bariatric surgery in the improvement of T2D has recently been reviewed in [13]. Ingestion of dietary prebiotics has also demonstrated potential to induce weight loss and improvement of insulin resistance and overall metabolic health in rodent models [14]. Nevertheless, when translating these interventions to humans, while prebiotics were indeed able to modify the composition of gut microbiota, their effects on metabolic health and weight were either very modest with merely a slight decrease in post-prandial glycemia and no effect on weight loss [15] or completely absent [16]. Similar observations can be made regarding the use of probiotics [17]. Overall these studies suggest that interventions modifying gut microbiota could be considered for therapeutic innovation in the field of T2D and obesity, especially when taking baseline microbiota composition and diversity into account. This concept is further reinforced by fecal transplantation experiments in rodents, highlighting the contribution of gut microbiota in weight reduction at least in mice.

For example, transferring feces from mice [18] or humans [19] who have undergone a bariatric procedure, into germ-free recipients, induces per se a weight reduction in the receivers. While the importance of weight loss is much lower than that induced by bariatric surgery, these studies highlight the important role of gut microbiota in weight reduction at least in mice, through fecal microbiota transplantation (FMT). However whereas FMT is now a recommended procedure in the treatment of Clostridium difficile (CD) infection, as described below, it remains to be demonstrated whether FMT could also be considered in the standard care of common disorders including metabolic diseases or obesity.

In this review, we will first review how FMT is being more broadly used due to its clinical successes and its technological innovation. We will then focus on whether using gut microbiota modulation via FMT has demonstrated any beneficial therapeutic effects on obesity and/or T2D. Finally, we will discuss potential factors impacting and modulating FMT success.

Fecal Microbiota Transplantation in Humans

A Success Story in Clostridium difficile Infection

The main routine care indication for FMT lays in the treatment of CD infection, reserved for cases resistant to antibiotherapy [20,21,22,23]. This disease originates from the major gut microbiota dysbiosis induced by the use of several lines of antibiotic regimens. FMT enhances gut microbiota diversity resulting in an improved clinical response or even the cure of this disease in 80 to 90% of the cases [24] with 60% of complete remission 1 month after FMT [25]. One of the first randomized clinical trials performed by our group was stopped prematurely since it showed, before the end of the trial, that FMT was effective and safe compared with antibiotics [26••], and these results have now been confirmed in several other randomized control trials [22, 27••, 28, 29] as well as in a recent meta-analysis [30]. Also, FMT has now been included in guidelines that recommend FMT in the context of recurrent CD infection to be used as the standard of care [31,32,33]. Most importantly, a consensus expert panel defined key recommendations for FMT use in CD infections [34] as well as the guidelines around technical, regulatory, administrative, and laboratory requirements for optimal FMT usage [35]. FMT is now being tested in clinical trials in other diseases such as inflammatory bowel diseases (IBD) [36, 37] and Crohn’s disease [38], which are also associated with a marked dysbiosis.

Novelty in the Administration Route of FMT

FMT was, and is still in most research studies, performed either using a nasojejunal tube, colonoscopy, or enema, sometimes requiring several administrations to be efficient in some diseases [37]. Most studies have observed good treatment tolerance both in the short and longer term and very rarely adverse events are seen, which mostly seem to be originating from the initial disease severity rather than FMT itself [39, 40]. Importantly, careful selection of the FMT donors [34] ensures part of this treatment safety [37, 39, 41]. Nevertheless, the risks originate mainly from the administration route. While no adverse event was detected using enemas, studies using nasogastric tube administration have reported few cases of intestinal bleeding and peritonitis [41]. In a large meta-analysis including over 1000 patients, colonoscopy administration was also associated with adverse events leading to hospitalization in 0.97% of patients receiving FMT. One death directly due to FMT was reported and the patient died from aspiration during sedation, while undergoing FMT through colonoscopy administration for a chronic CD infection [39]. Another death was reported due to pneumonia potentially caused by regurgitation during FMT using a nasoduodenal tube [42]. Nevertheless the most frequent adverse events are not serious and include intestinal symptoms such as bloating, flatulence, belching and abdominal cramps, abdominal discomfort, irregularity of bowel movements, and vomiting [43, 44]. Thus the FMT procedure appears safe overall; a national FMT registry has been initiated to study potential long-term side effects [45].

Based on the efficacy and (short-term) safety of FMT in CD infection, technologic developments have been made to enhance and facilitate FMT in routine clinical care and decrease adverse events linked to administration route by developing oral capsulized frozen FMT [46] and using frozen FMT stooling treated with glycerol as a cryoprotective procedure [47]. Concerning the efficacy of encapsulated FMT, a recent randomized control trial demonstrated that oral administration was as efficient as classic administration routes to obtain beneficial health outcomes mainly in CD infection [48]. Most importantly, no adverse event was observed post-FMT using this oral capsulized technique. Similar results have recently been reproduced in a study with 15 patients that were treated by 30 frozen capsules taken within 2 days (around 20 g of feces). This treatment enabled 86.6% remission rate after one FMT and 100% remission rate after the second administration, with excellent tolerance to this new administration route [49]. The only adverse events observed were linked to the severity of the initial disease [49]. Overall encapsulated FMT or “poop pills” as they are termed seem to be a relevant and safe alternative to more invasive FMT administration route and will probably increase the feasibility and acceptability of this treatment.

Fecal Microbiota Transplantation in Metabolic Diseases and Obesity

Improved Metabolic Health in Humans?

FMT used as a treatment option in obesity and other metabolic diseases has recently emerged. The reasons why gut microbiota modulation could beneficially affect glucose control have been extensively discussed herein [50] and include its actions on SCFA production, altered bile acids composition, and adipose tissue inflammation [51]. Our first pilot FMT study in metabolic syndrome included nine overweight/moderately obese individuals submitted to donor FMT originating from lean healthy controls that displayed a significant improvement in peripheral insulin sensitivity assessed by the gold standard, which is a stable isotope–based euglycemic-hyperinsulinemic clamp [52]. This beneficial effect was confirmed in a threefold larger follow-up study which also observed a minor yet significant reduction in glycated hemoglobin (HbA1c) at 6 weeks. Importantly, this study also showed that a reduced baseline microbial gene richness before lean donor FMT was associated with a better outcome [53••]. Despite the use of gastroduodenoscopy and nasoduodenal tube for donor FMT administration, no adverse events were reported. Interestingly however, response to donor FMT showed major inter-individual variability among receivers in both studies, with some patients displaying major improvements while others remained stable [52, 53••]. Lean donor FMT induced differential microbiota modifications in good and poor responders [53••]. For example, after FMT, good responders displayed an increase in Akkermansia muciniphila which has been previously associated with metabolic health improvement in mice [54, 55] but also in humans [56]. Another critical question is the long-term effect in FMT. The metabolic syndrome intervention showed that at 18 weeks, there was no remaining effect on insulin sensitivity and gut microbiota composition switched back to the baseline composition [53••]. This underscores the temporary nature of this FMT intervention. Nevertheless, these pioneering studies in treating naive prediabetes (metabolic syndrome) male subjects demonstrate that FMT could be of therapeutic use to improve insulin resistance and decrease HbA1c (Fig. 1) yet with a transient effect suggesting that multiple cures of FMT would probably be necessary.

Fig. 1
figure 1

Effects of fecal microbiota transplantation (FMT) in obese patients with metabolic alterations. Obesity, which represents a low-grade inflammatory disease with increased intestinal inflammation, and type 2 diabetes (T2D) are associated with gut microbiota dysbiosis. The literature has shown that fecal microbiota transplantation from lean donors enables the engraftment of bacterial strains from the donor and thus induces improvement in insulin resistance, for a short period of time. In the future, oral capsulized FMT will have to be tested in patients with T2D to evaluate whether it also improves intestinal inflammation, gut microbiota dysbiosis, and glycemic control

Full size image

Moreover, as T2D is a chronic disease where patients need to take multiple glucose-lowering drugs several times every day, whether FMT has any effect in patients with severe insulin resistance and eventually decreased insulin secretion remains to be studied. Nevertheless, as our study data have shown, subjects with the metabolic syndrome and the most severe insulin resistance do not show any beneficial effect from lean donor FMT on insulin sensitivity. In line with our previously mentioned studies in obesity [52, 53••], it is possible that gut microbiome plays a role in the early stages of insulin resistance and T2D whereas it is less critical in glycemic control of patient with advanced stages of T2D. Nevertheless, since FMT administrated by oral capsule appears safe, it provides an actionable tool to explore the relevance and differential effect of FMT in patients at different stages of glucose dysregulation, paving the way to personalized approaches.

FMT and Weight Modulation

The rationale to use FMT to induce weight loss originates from two sources. Firstly, the gut microbiome from obese individuals was found to increase energy storage as compared with lean individuals with the same caloric intake [57], even if these results have been questioned since. Secondly, since FMT from obese mice or humans into germ-free recipients is able to partly transfer weight gain [58,59,60], it is tempting to speculate that by contrast, using FMT from lean donors into overweight obese individuals could impact the receivers’ weight, thus providing a potential innovative approach in the control of weight. This unproven concept is currently debated. Indeed, a recent study observed that FMT from mice fed a normal chow diet into mice subjected to a high-fat diet induced significantly greater weight gain than that obtained by diet only [61]. Turning to humans, some pieces of evidence accumulate suggesting that FMT could induce weight gain in specific cases. Indeed, FMT from a normal weight individual (BMI = 25) to a patient suffering from anorexia nervosa led the receiver to increase (thus normalize) and stabilize her weight for 36 weeks post-FMT [62]. Furthermore, a brief report observed that a patient suffering from CD infection gained weight and further became obese after receiving FMT from her overweight daughter [63], although she followed strict diet and physical activity. Noteworthy, weight gain in these cited examples could merely reflect that FMT in malnourished patients enables the restoration of a healthy situation with a eubiotic gut microbiota and induces energy storage in patients with previously chronic condition of undernutrition, which thus might be more efficacious than in overweight subjects.

Nevertheless, the effects of FMT from lean donors have also been tested in overweight/obese patients in terms of weight modulation, and no effect on weight was seen [52, 53••]. Thus, to date there is no significant proof that FMT, even when originating from lean donors, is sufficient to induce weight loss. Most importantly, these first sets of data obtained in humans merely stimulate the need to carefully choose the donors both to avoid the transmission of communicable diseases but also to prevent weight gain. In this regard and in line with a study that showed that FMT from both mice [18] or humans [19] that underwent a bariatric procedure into germ-free recipients resulted in weight loss, de Groot et al. recently studied the effects of post-bariatric donor FMT compared with allogenic metabolic syndrome donor FMT in obese insulin-resistant males [64]. Despite the fact that post-bariatric donor FMT affects intestinal transit time in obese humans, in line with a previous animal study [13], we did not observe any effect of post-bariatric surgery donor FMT on either insulin sensitivity or weight modification. Yet and surprisingly, allogenic metabolic syndrome donor FMT worsened insulin sensitivity in all FMT-treated subjects [64]. More insight is now needed to link the correct FMT donor to the recipient in order to gain maximum therapeutic efficacy.

Factors Potentially Modulating FMT Response

Microbiota Engraftment Upon Donor FMT

First, studies have demonstrated that donor FMT induced gut microbiota composition modifications in the receiver as short as 1 week after intervention, switching the receiver’s microbiota composition towards that of the healthy donor [65]. Second, MGR increases post-FMT [52] and can even be restored [66] suggesting that FMT may act via normalizing microbiota dysbiosis, although this observation has not always been reproduced [53••]. Third, with time, gut microbiota composition continues to change after FMT. Using single nucleotide variation analyses, it was observed that whereas it still resembles that of the donor 1 month post-FMT, the dominant bacterial strains in the receiver were not always similar to those of the donor [66]. This result was further confirmed 3 months post-FMT [67]. Importantly, FMT studies have sometimes used antibiotic prior to FMT, potentially inducing confounding effects on the engraftment of the donor’s microbiota into the receiver. Interestingly, a recent study longitudinally analyzed the microbiota engraftment in different receivers without prior antibiotic use and showed a major inter-individual variability in FMT engraftment [68]. FMT as compared with placebo induced a significant modification of the receiver’s microbiota that lasted at least 3 months; however, with time, the receiver’s composition lost similarity with their donors. Furthermore, donor’s specific species only moderately increased post-FMT in all receivers yet were specific to donor-receiver pairs and higher than those observed after placebo FMT [68]. Finally, FMT induced a durable coexistence of bacterial strains present in the receivers and those transferred from the donor as well as a major strain replacement in the receiver yet variable from one receiver to the other [68] which could partially explain FMT variable responses [52, 53••]. Overall, these results suggest that FMT induces the colonization of a new microbiota that can interact with that of the receiver and differs from that of the donor despite an observed beneficial clinical effect of FMT. Furthermore, it suggests that in the future, interventions aiming at improving FMT engraftment according to the receiver or the donor’s microbiota should be undertaken in order to improve and maximize the clinical effect by increased donor bacterial engraftment (i.e., via dietary or antibiotic treatment before donor FMT).

Donor’s Microbiota Composition and Donor FMT Success

Studies evaluating the donor’s characteristics that enable an optimal response are currently ongoing. The notion of super-donor has emerged in the inflammatory bowel diseases (IBD) field. Indeed, seven out of nine patients who entered remission of their ulcerative colitis received their FMT from a single donor [69]. Likewise, an FMT study in patients with IBD using pooled feces from multiple donors finally evidenced that patients who entered remission had benefited from FMT containing the feces of one single donor thus suggested to be a super-donor [70]. Importantly, all subjects who received the FMT were on TNFα blockers, which might have also affected FMT outcome. Whether the notion of a super-donor is also valid in the metabolic field still warrants further investigations. The donor’s microbiota high diversity has been associated with good response in IBD remission [71]. Nevertheless, whether a donor is considered a good donor for every disease type remains an open question. It might well be that to induce a good response in a specific disease, the donor’s microbiota needs to be enriched in specific strains that are lowered or even absent in that specific disease. Further investigations are thus warranted.

Effects of Concomitant Drug Use and Donor FMT Effect

Concomitant presence of drugs and more specifically glucose-lowering drugs taken by the receiver is critical to be examined when using FMT to treat T2D. Metformin, the first line of antidiabetic treatment, induces a change in the composition of gut microbiota with an increase in both beneficial bacterial strains (Akkermansia muciniphila) and known pro-inflammatory strains like E. coli [72, 73]. Forslund et al. observed a shift in microbiota composition in patients with T2D with or without metformin with an increase in E. coli and a decrease in butyrate-producing bacteria, while not impacting MGR [74]. Importantly, Bäckhed’s group identified a strong metformin-related microbiota signature that in itself reproduced some of the beneficial effects of the molecule after FMT experiments in germ-free mice [75]. These data suggest that evaluating the effect of FMT from healthy donors into patients with T2D might be tricky since patients with T2D receive one or associations of glucose-lowering drugs which impact differently the gut microbiota composition as this might interfere with the colonization or engraftment of the donor’s gut microbiota after donor FMT.

Effects of Concomitant Dietary Intervention and Donor FMT Effect

The gut microbiota profile is severely impacted by both short- [76] and longer-term food intake habits, with the latter specifically linked to enterotypes [77], microbial composition, and gene richness [78]. Nevertheless, the acute modification of dietary intake also induces at least partial changes in gut microbiota composition [76, 79]. Dietary intake from receivers might most probably impact the clinical effects of FMT as discussed by Groen and Nieuwdorp [80] and demonstrated in rodent FMT experiments. Indeed, FMT from obese or lean twins into germ-free mice induced differential metabolic and body weight improvement according to the diet administrated to mouse receivers [59]. However, to date there is no standard recommendation regarding the optimal diet that would maximize the FMT therapeutic effects and bacterial strain engraftment. Moreover the donor’s diet might also impact the clinical success of FMT. A recent study using FMT from vegan lean donors enabled a switch in gut microbiota composition in the receiver towards the vegan profile of the donor [81] but with some individual variability in receivers. Importantly, this switch was not accompanied by functional modification as seen with the measured production of trimethylamine. Potentially, the absence of functional effects might originate from inter-individual variability in microbiota composition within the receivers but also by the diet consumed by the receiver, which remained similar before and after FMT [81]. It is tempting to investigate in the future whether modifying the diet after FMT could enhance FMT response yet.

Conclusion and Perspectives

The current literature suggests that FMT could improve metabolic state and in particular insulin sensitivity, yet the impact of FMT on weight loss in obese subjects needs further study. Moreover, to make relevant progress in metabolic disorders, several steps must be completed such as (i) investigating the impact of FMT in patients with various degrees of insulin resistance towards overt T2D, since the evidence to date has been gathered in male subjects with the metabolic syndrome, (ii) evaluating gender effects, (iii) determining the optimal mode of donor selection based on clinical information, (iv) identifying those able to induce a good response, and (v) elucidating which patient profile might benefit the most from FMT. Nevertheless, while donor FMT has undergone technological innovation to simplify and improve its feasibility, it remains a complex and costly procedure, due to the donor selection process which includes a broad infectious analysis as well as complex preparation and administration technics. Furthermore, while FMT has demonstrated its cost-effectiveness in different severe diseases including CD infection [82] and IBD [83], it remains to be proven whether this approach can be extended to T2D and particularly the most severe patients for whom glucose control remains above target, despite intensive medical therapy. A survey on a small study group of obese individuals recently demonstrated that a large number of them would be willing to undergo FMT if the donor was carefully selected [84], thus suggesting that this treatment could be proposed in standard of care for indications other than CD infection. If FMT successfully improves or treats T2D, this could lead to better understanding of pathophysiology and also pave the way towards other microbiota-related therapeutic developments. Indeed, the development of third-generation probiotics based on newly identified beneficial bacterial strains associated with improved metabolism in humans is also an emerging field that calls for future intensive research.