1.1. Acute Graft-versus-Host Disease

Allogeneic hematopoietic stem cell transplantation (HSCT) is a curative procedure for a variety of hematological malignancies. The transplantation process involves administration of a conditioning regimen, which includes chemotherapeutic agents with or without radiation, followed by infusion of hematopoietic progenitor cells from healthy donors [1]. Following conditioning, blood counts go down, and approximately two to three weeks after transplantation the stem cells engraft, and the slow process of immune reconstitution begins. Despite advances in supportive care and transplantation technology, graft-versus-host disease (GVHD) remains a major cause of transplantation-related mortality and morbidity, affecting up to 40–60% of allo-HSCT patients, and accounting for 15–20% of deaths [2]. GVHD is traditionally divided into an acute and chronic disease, and in this review, we will focus on the acute form which involves the skin, gut, or liver. Acute GVHD typically appears in the first 100 days posttransplantation, but can also develop later, especially in patients receiving reduced-intensity conditioning or direct lymphocyte infusions.

Acute GVHD pathophysiology has been reviewed elsewhere [3–5]. Briefly, alloreactive T-cells transplanted from a nonidentical donor recognize the transplant recipient as foreign, thereby initiating an immune reaction damaging the recipient's tissues. GVHD can be considered a three-step process in which the innate and adaptive immune systems interact: (1) tissue damage to the recipient by the radiation/chemotherapy pretransplant conditioning regimen, (2) donor T-cell activation and clonal expansion, and (3) activation of cytotoxic T-cells and natural killer cells, inflicting local tissue damage. There is a growing body of evidence implicating the gut microbiota in the development and propagation of GVHD [6,7].

1.2. The Microbiome in HSCT

The human body is colonized by a multitude of microorganisms. Microbiota refers to the entirety of microorganisms (bacteria, archaea, viruses, fungi, and other eukaryotes) within a specific habitat. Microbiome is defined as the biotic (microorganisms and their genomes) and abiotic (environmental) factors present within a particular habitat. Work led by the groups of Marcel van den Brink and Eric Pamer has revealed intriguing associations between alterations of the gut microbiota and outcomes of allogeneic HSCT, including infections, mortality, and relapse [8–12]. Since the gut microbiota and its metabolites are essential for development andmaturation of the host immune system and modulation of the immune response [13–15], it is unlikely that changes in the microbiome during and following HSCT lack functional implications. As the insult to the gut during conditioning is instrumental for the development of GVHD, and the gut is a target organ in acute GVHD, there is a strong rationale to postulate that the gut microbiome plays a major role in GVHD through its interaction with the immune system. Furthermore, intestinal homeostasis and mucosal integrity are dependent on the microbial community occupying the gut [16,17].

The first studies linking the microbiota to GVHD were performed almost half a century ago. Germ-free mice and antibiotic-treated mice were less likely to develop GVHD and experienced longer survival [18–21]. An attempt to reproduce these findings in humans through gut decontamination, laminar-airflow isolation rooms, and skin cleansing led to mixed results [22–28]. More recently, the advent of high-throughput molecular methods to study the microbiome has provided new insights. Several groups have reported that dysbiosis and a reduction in stool bacterial diversity following allogeneic HSCT are associated with increased risk of acute GVHD and GVHD-related mortality [8,29,30]. Interestingly, depending on the cohort, stool samples of patients who later developed acute GVHD were enriched for specific bacterial taxa; intestinal domination by enterococci and a reduction in Blautia spp., as well as other members of the Clostridia class, were apparent before and during GVHD development [8,30]. The protective role of commensal anaerobes was further supported by an increased risk for GVHD in patients treated with broad-spectrum antibiotics with coverage of anaerobic bacteria [31,32]. Other bacteria implicated in GVHD include members of the Bacteroides genus (B. thetaiotaomicron, B. ovatus, and B. caccae) which were negatively correlated with subsequent severe acute GVHD in stool samples collected at the time of neutrophil engraftment. Rothia mucilaginosa, Solobacterium moorei, and Veillonella parvula were also positively correlated, while several Lachnospiraceae (including B. luti) and a Butyricicoccus species were negatively correlated [29]. The gut microbiome is a complex biosystem and, therefore, assuming that specific species drive GVHD risk might be simplistic. Golob et al. suggested that the gradient between protective and detrimental species is more predictive of acute GVHD [29].

Whether and how the microbiota contributes to the development of GVHD development and propagation remains an open question. Since the immune system and the microbiota have a robust bidirectional relationship [33–37], it is reasonable to postulate that injury to the gastrointestinal (GI) mucosa secondary to conditioning and antibiotics is essential for GVHD evolution. Increased permeability of the intestine leads to bacterial translocation from the gut lumen, which in turn stimulate the innate immune system via toll-like receptors, priming alloreactive T-cells that drive GVHD [6,36,38]. Metabolites of commensal bacteria may also modulate the activation of the immune response. For instance, certain Clostridia species, which might be reduced during transplantation owing to the use of antibiotics with aerobic coverage, exert an anti-inflammatory effect by production of the short chain fatty acid butyrate, thereby relieving or preventing intestinal acute GVHD [39]. Several papers have reviewed the link between GVHD, microbiota, and its metabolites [6,7,17].

Severe acute GVHD often involves the GI system [40]. Patients refractory to treatment with glucocorticosteroids, considered the first-line therapy, have a dismal prognosis with survival rates as low as 5% [40,41]. Currently, there is no standard for second-line therapy [42]. Therefore, novel insights into the prevention and treatment of GI acute GVHD are urgently needed. Knowledge on the role of the microbiome in GVHD pathophysiology have opened up a new avenue of therapeutic targets [7]. Antibiotics, prebiotics (indigestible compounds that are fermented by commensal bacteria to produce protective metabolites such as short-chain fatty acids), probiotics (multiple or selected strains of microorganisms that confer a benefit), postbiotics (bacterial metabolites), and dietary interventions [43,44] are all areas of intense research. In this review, we will focus on fecal microbiota transplantation (FMT), an intervention that has pre-, pro-, and postbiotic elements, and has emerged as a promising live microbial therapy for GI acute GVHD.

2.1. Principle and Current Use

FMT is the transfer of stool from a donor to a recipient that harbors a dysbiotic or perturbed GI microbiome resulting in disease. By introducing a “healthy” microbiota, FMT aims to restore eubiosis and homeostasis in the recipient. Modern use of FMT began in 1958: Eiseman and colleagues described four patients with Staphylococcal pseudomembranous enterocolitis who improved after fecal enemas [45]. FMT was abandoned for many years until its resurgence for treatment of Clostridioides (formerly Clostridium) difficile infection (CDI), first reported by Schwan et al. in 1983 [46]. Since then, multiple studies have demonstrated that FMT is a highly effective treatment for recurrent CDI [47–51]. Mechanisms underlying FMT efficacy have been reviewed by Khoruts and Sadowsky, and include restoration of the colonic microbial community and inhibition of C. difficile by competition for nutrients, direct suppression by antimicrobial peptides, bile-acid-mediated inhibition of spore germination and vegetative growth, and activation of immune-mediated colonization resistance [52].

Fecal donation may be from related or nonrelated donors or precollected autologous stools. Donors undergo screening for infectious disease as well as other comorbid conditions. Centralized stool banks are advocated to ensure safety and permit the use of fecal donation across multiple centers [53]. FMT may be administered via either colonoscopy, nasogastric/duodenal tubes, capsules, or enema; all routes of administration are effective in recurrent CDI [47–51]. Enemas and capsules are less invasive, but the spread through the distal GI tract and the inoculum size in the latter are reduced. Furthermore, there is concern that they may be less effective in repleting certain classes, including Bacteroidia [54], which may have GVHD protective properties. Nasogastric/duodenal tube or colonoscopy allows for administration of large amounts of fecal matter, but are invasive [55]. The process of preparing the fecal inoculum is similar across delivery methods and includes mixing and grinding of the stool with normal saline, removing particulate matter, and storage in an appropriate vehicle for delivery.

FMT is an accepted therapy for recurrent CDI, and has been endorsed by national and professional societies and regulatory authorities across the globe [56]. Thousands of stool transplantation have been performed worldwide with an excellent safetyprofile. Typically, FMT adverse events are mild and include bloating, flatulence, nausea, and abdominal discomfort. Severe adverse events, including infections, are rare and are mainly related to the risks of nasogastric tube insertion, sedation, and endoscopy [57]. There is a theoretical concern of transmission of infectious pathogens not identified in the screening process. Furthermore, data regarding potential long-term effects, including transmission of noninfectious donor disease (e.g., metabolic syndrome) and immune dysregulation are lacking.

2.2. FMT for Treatment of C. Difficile in Immunocompromised Patients

Immunocompromised patients represent a unique population where a greater concern for infectious complications exists. Kellyet al. studied outcomes of 80 immunocompromised patients treated with FMT owing to CDI [58]. Causes for immunosuppression included HIV/AIDS, solid organ transplants, oncologic conditions, immunosuppressive therapy for inflammatory bowel disease, and other medical conditions/medications. Severe complications following FMT were rare and related to aspiration during sedation and mucosal tear during colonoscopy. No infectious adverse events related to FMT were observed. Other studies have reported similar findings with low rates of FMT-related infections [59–65]. Overall, FMT appears safe in immunocompromised patients. However, data are limited and experience with HSCT recipients, which represent the far extreme of the immunosuppressed population, is lacking.

FMT in the HSCT setting was initially reported for treatment of recurrent CDI (Table 1). Webb et al. treated seven allogeneic HSCT recipients with FMT, administered via a nasojejunal tube. CDI resolved in all patients, and the procedure was well tolerated with no serious adverse events or any infectious complications. Interestingly, one patient with acute GI GVHD was able to taper systemic steroids following the FMT. Moss et al., described eight patients who were previously treated with autologous or allogeneic HSCT who developed recurrent CDI. All patients were treated with FMT in capsules and had complete resolution of the infection [64]. Additional case reports and case series have described favorable outcomes in HSCT patients treated with FMT for recurrent CDI [66–69].

2.3. Fecal Microbiota Transplantation for Microbiota Restoration and Eradication of Antibiotic-Resistant Bacteria in HSCT Recipients

FMT has also been explored as a mean of restoring microbiota injury following HSCT (Table 2). The rationale is that FMT may reduce dysbiosis following HSCT, which is associated with poor transplantation outcomes [9]. Furthermore, FMT may diminish hospital-acquired infections and carriage of antibiotic-resistant bacterial strains, which are prevalent in immunocompromised patients [64,70,71]. In a single-center pilot study, 13 patients undergoing allogeneic HSCT received FMT from an unrelated donor via capsules. FMT was performed no later than four weeks after neutrophil engraftment [72]. The procedure was well tolerated without severe adverse events. FMT increased intestinal bacterial diversity. Following FMT 1, 2, and 6 patients developed C. difficile colitis, grades 3–4 GI acute GVHD, and moderate-severe chronic GVHD, respectively. Taur et al. performed a randomized controlled open-label trial of autologous FMT via enema versus no intervention in allo-HSCT recipients; the results of the first 25 of 59 patients have been reported [73]. Stool for FMT was collected before HSCT hospitalization; patients were only considered eligible for the trial if a “healthy” intestinal bacterial profile was observed. FMT was performed following neutrophil engraftment. FMT was safe and boosted microbial diversity, restoring bacterial populations lost during HSCT and reversing the disruptive effects of the broad-spectrum antibiotics. Two studies directly used FMT to eradicate multidrug-resistant bacteria. Bilinski et al. performed FMT via a nasoduodenal tube in 20 patients with various blood disorders, some of which were also HSCT recipients who were colonized with antibiotic-resistant bacteria [74]. Again, FMT was well tolerated without infectious complications; antibiotic-resistant bacteria were successfully eradicated in 75% of the patients. Similar findings were reported by Battipaglia et al., who performed FMT by enema or nasogastric tube in 10 patients before (n = 4) or after (n = 6) allogeneic HSCT [75]. Overall, the bulk of data indicate that FMT is safe in HSCT recipients and its use may extend beyond the treatment of recurrent CDI.

2.4. Fecal Microbiota Transplantation for Treatment of GI Acute GVHD

Given the link between GI acute GVHD and the microbiota, FMT has been investigated as a potential therapeutic intervention (Tables 3 and 4). In their pioneering study, Kakihana et al. performed FMT for treatment of steroid-resistant (n = 3) or steroid-dependent (n = 1) acute GI GVHD [76]. The FMT construct was fresh (i.e., not frozen and then thawed) and donors were relatives. At the time of FMT, patients had at least grade II acute GVHD, were on a minimal dose of 1 mg/kg methylprednisolone and received tacrolimus and beclomethasone. Patients received 1–2 courses of FMT via a nasoduodenal tube. Three patients achieved complete normalization of GI symptoms, and one patient experienced transient relief in diarrhea and was defined as having partial response. Response to FMT correlated with increased peripheral circulating regulatory T-cells and intestinal bacterial diversity, with the appearance of beneficial species such as Bacteroides, Lactobacillus, Bifidobacterium, and Faecalibacterium spp., which are characteristic of a “healthy” microbiota. One patient developed mild hypoxemia, delirium, and lower GI bleeding a few days after the FMT. The same patient experienced transplant-associated thrombotic microangiopathy three days after the second FMT. The authors attributed these events to refractory GVHD rather than the FMT. Three of the four patients in this pilot study eventually died, with disease relapse as the cause of death in two. A theoretical concern that FMT mitigates the graft-versus-tumor effect could be raised. However, these were patients at high risk of relapse without a control arm to provide a valid comparison. Furthermore, prolonged immunosuppression may also increase relapse [77]. Spindelboeck et al. treated three patients with grade IV GI acute GVHD with repeated FMTs via colonoscopy [78]. Two achieved clinical response after 1–6 courses of FMT. Two patients had a complete response, and one a transient reduction in diarrheal symptoms. No severe adverse events were observed. In the most recent study by Qi et al., eight patients with stage IV GI acute GVHD received 1–2 courses of FMT from an unrelated donor via nasoduodenal tube [79]. All patients achieved clinical symptomatic remission after the first FMT. Three patients experienced relapse of GI symptoms. No severe adverse events were noted. Following FMT, intestinal bacterial diversity increased in the recipients, and so did abundance of beneficial bacteria such as Bacteroides. Two abstracts have been published reporting on FMT outcomes in steroid-resistant acute GI GVHD. van Lier et al., performed FMTs from unrelated donors via nasoduodenal infusion [80]. Eleven of 15 patients showed complete response. GI GVHD eventually recurred in five of the responding patients. 16S ribosomal RNA sequencing of fecal samples from the first 13 patients revealed overall low alpha diversity (i.e., microbiota diversity within the samples) in patients pre-FMT. One week after FMT, the fecal microbial composition of patients in complete remission resembled that of the donor. We have also reported our experience with repeated courses of FMT from unrelated donors, administered via frozen capsules, in steroid-resistant (n = 6) and steroid-dependent (n = 1) patients [81]. Two exhibited complete clinical response. Capsules were well tolerated. Notably, two patients developed bacteremia a few days following the FMT. In the first case, using metagenomic sequencing allowing strain-level identification, the Enterococcus Faecium recovered from the patient's blood was already present in the host gut in a sample collected prior to the FMT. In the second case of Pseudomonas aeruginosa bacteremia, 16S rRNA sequencing did not detect the bacteria in the FMT inoculum. With a high degree of certainty, these infectious events were not directly related to the FMT. Possibly, the FMT might have altered intestinal permeability, thereby increasing the risk of infections, as suggested by Quera et al., who also reported a case of bacteremia following FMT in a patient with Crohn's disease [82]. Overall, FMT seems to be a promising approach for GI acute GVHD. Nevertheless, caution should be exerted, since this is a highly immunosuppressed population. In addition, to allow for persistent microbiota engraftment, antibiotics should be withheld. This may prove as a difficult task in steroid-resistant GVHD patients, who are often on antibacterial therapy for prevention or treatment of infections.