Factors affecting microbiota of SCT recipients
In order to understand how changes in the microbiota affect SCT recipients, it is necessary to examine the baseline characteristics of intestinal flora in this population, and the shifts that occur during the transplantation process. Holler et al. used a combination of 16S ribosomal RNA sequencing and urinary indoxyl sulfate measurements to analyze stool samples of patients for changes in bacterial prevalence during transplantation. They found a standard initial distribution of major microbial phyla across hosts before transplant. The most prevalent phylum was
Firmicutes, followed by
Bacteroidetes,
Proteobacteriae, and
Actinobacteriacea [
2]. Major species of
Firmicutes included
Eubacterium rectale, Clostridium phytofermentans, and
Lactobacillus lactis. The difference between the microbiome of non-transplant donors who had recent hospitalization and SCT patients is in the prevalence of
Enterococcus. After SCT, there was a 21 percent increase in prevalence of
Enterococcus, with a notable expansion of
E. faecium and a complementary decrease in
Firmicutes and other commensal phyla. A similar microbiologic shift was noted by Ubeda et al. [
3] who identified dominance of vancomycin-resistant Enterococcus as a predictor of bacteremia in SCT recipients, and by Eriguchi et al. [
4] who observed that active gastrointestinal graft-versus-host disease (GVHD) was associated with a prominent shift towards
E. coli.
Using a murine model, Jenq et al. found a dramatic loss in microbial diversity in the first two weeks after bone marrow transplant. The loss of diversity became even more pronounced in the setting of GVHD. Analysis of bacterial subpopulations demonstrated an expansion of
Lactobacillales and
Enterobacteriales, with a prominent loss of
Clostridiales. Since these changes were irrespective of whether there was GVHD or not, it was concluded that the pre-treatment with radiation and chemotherapy were responsible for the microbiologic shifts observed after SCT [
5].
However, the majority of studies documenting changes in microbiota occurred in the presence of antimicrobial exposure for gut decontamination, which was the standard practice in SCT. This practice came under scrutiny when it was found that the shifts in endogenous flora may have adverse effects on post-transplant complications. Beelen et al. demonstrated that although pre-treatment with the combination of ciprofloxacin and metronidazole resulted in a tenfold decrease in anaerobic bacterial culture growth in recipients of transplants from HLA-identical sibling donors, this was accompanied by an increase in
Enterococcus [
6].
The choice of pre-transplant antimicrobials also plays an important role in shaping shifts in the microbiota. Pre-treatment with metronidazole resulted in the most pronounced Enterococcal predominance in the post-transplant period; conversely fluoroquinolone exposure led to a significant decrease in
Proteobacteria [
7].
It is hypothesized that commensal bacteria induce antimicrobial peptides, which help maintain a healthy, diverse microbiome, and intact intestinal mucosa. With prolonged antibiotic exposure, the loss of commensals is accompanied by decreased induction of microbial-associated molecular patterns (MAMPs), specifically Reg3alpha, the main peptide limiting Enterococcal overgrowth [
8]. These MAMPS also communicate with the host immune system via pattern recognition receptors (PRRs). Activation of PRRs increases major histocompatibility complex expression and co-stimulatory molecules on the antigen-presenting cells and endothelial cells. This cascade ultimately leads to the increased production of tumor necrosis factor, type I interferons, and interleukin (IL)-1 and IL-6. A decrease in the activation of this pathway due to decreased level of MAMPs renders the host more susceptible to intestinally originated infections. The intestinal microbiota is also important in promoting differentiation of regulatory T cells, generation of IgA-secreting B cells, and formation of secondary lymphoid organs phenotypes [
9]. The liver normally maintains tolerance against these harmless antigens and any commensal bacteria that translocate from the gut. However, when immunosuppressants are used, this homeostasis is perturbed. In addition, immunosuppressive agents, including chemotherapy, can lead to a loss of intestinal microvilli, tight junction damage, and a decrease in IgA secretion. Exposure to chemotherapy, even in the absence of antimicrobial use, is associated with a drastic drop in
Faecalibacterium and increase in
Enterococcus [
10].
Clostridium difficile colitis
Hospitalized patients, and immunocompromised hosts in particular, are vulnerable to a wide range of nosocomial pathogens. Among these,
Clostridium difficile is the most common. The emergence of the NAP1/027 strain of
C. difficile over the first decade of the twenty-first century has been associated with more severe infection [
11,
12]. Recent data reveal that
C. difficile has surpassed Methicillin-resistant
Staphylococcus Aureus (MRSA) to become the most prevalent nosocomial infection in community hospitals in certain regions of the United States [
13].
SCT recipients possess many risk factors for
C. difficile infection (CDI), including prolonged hospitalization, exposure to broad-spectrum antimicrobials in the peri-transplant period, and long-term immunosuppression. Recent literature show CDI rates among SCT recipients are up to nine times higher than in the general inpatient population [
11,
12]. A study by Jain et al. found that 12 percent of patients admitted to their institution for SCT were colonized with toxigenic
C. difficile [
14]. Agha et al. reported that 11 percent of SCT recipients developed peri-transplant CDI [
15]. Interestingly, the only factor independently associated with development of CDI was a previous history of the same. Patients with matched-related grafts had a lower likelihood of developing CDI in the study population [
15]. A review by Alonso et al. found an overall CDI rate of 9.2% among SCT recipients, with higher rates among allogeneic transplant recipients compared to those who underwent autologous transplant (12.5 vs. 6.5%) [
11,
12].
Changes in intestinal microbiota are inherently part of the pathogenesis of CDI. The traditionally understood mechanism involves exposure to broad-spectrum antibiotics altering native gut flora, allowing for colonization by toxigenic
C. difficile. The shifts in the intestinal microbiome of SCT recipients is likely intrinsically tied to their unique vulnerability to CDI. Cytotoxic chemotherapy given prior to transplant causes disruption of endogenous flora, enabling
C. difficile colonization. A study by Loo et al. demonstrated that prior chemotherapy increased the likelihood of
C. difficile colonization over twofold [
16]. The eradication of endogenous flora by preparative chemotherapeutic regimens opens a larger microbiological space than exposure to standard antimicrobials, leaving a wider available niche in these patients. The discovery that matched-related hematopoietic grafts were associated with a lower risk of CDI than matched-unrelated grafts suggests that pre-existing microbiome characteristics play a role in determining CDI susceptibility [
15]. Pairing donors and recipients with similar gut microbiota, and probable similar past environmental exposures, may produce less disruption to endogenous flora after transplant.
Graft-versus-host disease
Acute GVHD is a leading cause of morbidity and mortality among SCT recipients. The changes in intestinal microbiome associated with SCT are thought to play a role in the development of GVHD. The pathophysiology of GVHD involves host antigen-presenting cells activating donor T-lymphocytes, leading to a hyper-proliferative inflammatory cascade and tissue damage. There has long been speculation about the role of intestinal environment on GVHD; early studies suggested that antimicrobial prophylaxis may reduce the incidence of GVHD, though these findings were not replicable [
6]. Holler et al. found that at time of admission for SCT, patients showed a similar distribution of dominant commensal strains. With exposure to pre-transplant antimicrobials, all recipients experienced major changes in intestinal microbiota, with an overall loss of bacterial diversity. There was a notable expansion of
Enterococcus species with a decreased in other phyla [
2]. This shift was most pronounced in patients with active GI GVHD.
There is also evidence that specific commensal organisms may play an outsized role in modulating GI inflammation. In one cohort study, decreases in anaerobic bacteria in fecal samples, and in particular
Blautia spp., were correlated with higher rates of acute GVHD and higher mortality [
17]. Studies of intestinal flora in SCT patients with GVHD have shown that a reduction in microbial diversity was associated with higher mortality, and that loss of intestinal bacterial diversity is an independent predictor of mortality in SCT recipient [
10].
Though the role of intestinal microflora in regulating immune response and maintaining host barriers against infection have long been recognized, the specific mechanism by which the gut microbiota may modulate GVHD is not well understood. Using a mouse model, Jenq et al. found a marked loss in microbial diversity in mice with GVHD. They noted expansion of certain bacterial subpopulations, especially
Lactobacillales, in mice with GVHD [
5]. An increased abundance of
Enterobacter came at the expense of other commensals such as
Clostridia and
Bacteroides species. They observed that the expansion of the
Lactobacillales was not the culprit for the development of GVHD since mice depleted of
Lactobacillales developed GVHD that was more severe, pointing to a role of the
Lactobacillales as a disease modulator. Re-introduction of gut commensals in these hosts reduced GVHD mortality. Subsequent analysis in human subjects identified similar dysbiosis in patients with gastrointestinal GVHD, with increase in
Lactobacillales and decrease in
Clostridiales associated with gut inflammation [
18].
Interestingly, several studies have suggested an association between GVHD and
C. difficile infection in allogeneic SCT recipients. Chakrabarti et al. demonstrated an association between CDI and severe GVHD [
19]. Similar results were produced by Dubberke et al. who demonstrated a temporal relationship between CDI and GVHD [
20]. Patients with CDI, including mild infection, had a higher likelihood of subsequently developing new-onset GVHD and new-onset gut GVHD. Trifilio et al. showed that SCT recipients who developed CDI were more likely to be complicated by severe GVHD at day 60 and day 100 after transplant [
21]. The association between CDI and subsequent GVHD may be related to either the direct intestinal inflammation caused by
C. difficile, or to the severe loss of intestinal microflora causing both processes.
Since diet affects the composition of the intestinal microbiome [
22], it follows that certain diet contribute to the development of GVHD. Enteral feeding has been shown to protect against GVHD [
23]. Mice and human subjects receiving total parenteral nutrition exhibited increases in IFNγ and reduction in IL-4, IL-10, gut CD4+ cells [
24,
25], a pattern of immune responses that are pro-inflammatory.
Changes in the intestinal metabolism and microbiota-derived metabolites are likely the mechanisms underlying the role of the intestinal microbiota in regulating GVHD. Butyrate, a short chain fatty acid (SCFA) has been shown to the STAT3-dependent pathways of both the innate immune and allo-stimulatory functions of antigen presenting cells. These are accomplished through upregulating indoleamine-2,3-dioxygenase (IDO) [
26,
27]. High levels of IDO mediate immune suppression through two mechanisms. First, they deplete tryptophan needed for the metabolism of donor T cells, resulting in apoptosis of these allo-reactive T cells. Second, they stimulate Treg generation [
28,
29].
Modulation of the intestinal microbiota may also be exploited to treat acute GVHD. In a small pilot study, Kakihana et al. performed fecal microbiota transplant (FMT) from healthy donors on four patients with steroid-refractory intestinal GVHD [
30]. All four patients responded, three achieved a complete remission and one partial remission with FMT and concurrent corticosteroids.