Chronic Rejection as a Persisting Phantom Menace in Organ Transplantation: A New Hope in the Microbiota?

Abstract

Purpose of review

The microbiota plays an important role in health and disease. During organ transplantation perturbations in microbiota influence transplant outcome. We review recent advances in characterizing microbiota and studies on regulation of intestinal epithelial barrier function and mucosal and systemic immunity by microbiota and their metabolites. We discuss implications of these interactions on transplant outcomes.

Recent findings

Metagenomic approaches have helped the research community identify beneficial and harmful organisms. Microbiota regulates intestinal epithelial functions. Signals released by epithelial cells or microbiota trigger pro- or anti-inflammatory effects on innate and adaptive immune cells, influencing the structure and function of the immune system. Assessment and manipulation of microbiota can be used for biomarkers for diagnosis, prognosis, and therapy.

Summary

The bidirectional dialogue between the microbiota and immune system is a major influence on immunity. It can be targeted for biomarkers or therapy. Recent studies highlight a close association of transplant outcomes with microbiota, suggesting exciting potential avenues for management of host physiology and organ transplantation.

Introduction

Since the first human microbiome study characterizing the gut microbiome in two healthy individuals was published in 2006 (1), plenty has been written and published about microbiome research, its current applications in promoting health, its potential use for novel biomarkers, and its applications for innovative therapeutics targeting chronic and acute diseases. However, as the microbiome research field has now become mature and sequencing has been democratized to a point that “anyone can sequence”, there is still much value in summarizing the latest knowledge about the microbiome and microbiome research for physicians and research scientists, and emphasizing the standards used to measure, assess and characterize the microbiota, with a focus on applying this knowledge to transplantation research. In this review, we also aim to reflect upon the challenges ahead and inform what might be needed for the next phase to move beyond correlations and associations. We will also review the implications for alloimmunity, tolerance, and the use of microbiota and its components as biomarkers, prognostic indicators, or therapeutic interventions. This much-needed shift of microbiome research beyond observational and associative studies is exemplified in this review by the need to more precisely map the critical interactions between the microbiota, the immune system, and the host as a whole, since the precise nature of these interactions – much of which is still currently unknown - will have critical implications for the design of microbiota-targeted therapies, monitoring and interventions, in clinical care in organ transplantation.

Of Microbes and Microbiomes: A Brief History

“Microbes rule the world: it’s that simple”. These words were coined over 13 years ago in a report from the National Academy of Sciences (2) that was the prelude of the NIH-initiated Human Microbiome roadmap initiative, which officially started in 2007. Following this initiative in the United States, other consortia around the world were created to study the role of microbes in human health and diseases, as well as in the environment. Most of these early microbiome projects initially focused on cataloguing – using high-throughput cultivation-independent approaches – the bacterial species in various body sites in a healthy state (3–6), as well as sequencing the reference genomes of isolates from these various body sites (7). Most of these were bacterial genomes, but this sequencing effort also included a smaller number of viruses, archaea and eukaryotic species. The goal was to increase and improve the sequence information in databases used for microbiome data analyses, aiding taxonomic assignment of 16S rRNA gene sequences and the functional annotation of metagenomic sequences from microbiome samples.

Over a decade later, the number of published studies focused on the microbiome has increased exponentially (Figure 1), as interest in all types of microbiome research - host-associated or environmental - has been booming. High profile, and sometimes controversial, studies have also suggested that previously unsuspected niches of the microbiome exists within the human body, such as within the placenta (8), the healthy lungs (9, 10) or even the brain (11). These controversies challenged the classic dogma that some environments within the human body might not be sterile and devoid of bacterial colonization. However, they also highlighted some of the critical limitations of the DNA-based cultivation-independent strategies used for microbiome profiling (discussed below), and that standards including proper controls must be implemented to rule out exogenous contamination during sample collection and processing and biases due to lab reagents and kit contaminations, as well as post hoc in silico processing (12–14). These issues have triggered calls in the research community for more openness and transparency in data generation and analysis (15, 16), as well as a push to establish standards for sample collection and processing, sequencing and analyses (17–19).

Figure 1: Microbiome-related projects funded by the NIH and articles indexed in PubMed published since 2000.

A, B and C: Number of microbiome-related projects (panel A), overall funding amount (panel B) per fiscal year, and types of grant applications (panel C) funded by institutes within the NIH. The search query was run on publicly available information using the Text Search function with “microbiota OR microbiome” as search terms, and “Limit Project search” to Project Title and Project Terms. The search results were then further filtered to only include funding from the NIH. The query was run on 12/25/20 using the initial public release of the Modernized NIH RePORTER, version 2020.9. D: Number of articles indexed in PubMed containing “Microbiome” in keywords. PubMed query was run on 12/25/20.

Despite these controversies, microbiome research strategies represent “an elegant weapon for a more civilized age” [Ref: Star Wars / Obi-Wan Kenobi] in microbiology, compared to traditional cultivation-dependent methods. Because traditional microbiological approaches require isolation and cultivation, their usefulness is inherently limited due to the overwhelming majority of microbial species (less than ~ 5%) that resist cultivation in the laboratory because of yet-unknown culture requirements, such as optimal combination of nutrients, growth temperatures, oxygen levels or co-cultivation with potentially key microbial partners (20). Termed “the great plate-count anomaly”, this observation highlights the wide disparity between the numbers of micro-organisms that can be counted after cultivation, compared to direct microscopic counts (20, 21). This is why cultivation-independent strategies have greatly facilitated our understanding of the microbial communities living in and around us, and have revealed previously unsuspected levels of microbial diversity. Human-associated microbiome studies have also linked shifts in the microbiome, in its composition and functions, to a growing number of diseases; thus suggesting that certain microbiome features could also be used as potential “biomarkers” for these diseases, and eventually as novel therapeutic targets. It is now clear that our microbial partners are not “passive gauges” of diseases, but are active contributors to pathogenesis, sometimes even early initiators of these processes. But with few exceptions, the precise metabolic, immune, and regulatory pathways by which the microbiome influences health remain largely unknown.

Microbiome research has reignited the concept of “beneficial microbes”, i.e. probiotic microbes for improving/boosting human health. This concept is not new, as history of probiotics can be traced back to nearly 10,000 years ago when fermented foods were originally produced (22). As such, the health properties of beneficial bacteria have been recognized long before the discovery of bacteria (23), but how exactly they affect human health remains elusive. After being ignored for decades, microbiome research has generated a renewed interest for characterizing how probiotic bacteria impact human health. For example, a recent study on 61 elderly volunteers showed that a daily dose of L. bulgaricus for 6 months resulted in an increase in the percentage of NK cells, an improvement in the parameters defining the immune risk profile; an increase in T cell subsets that were less differentiated; a decrease in the pro-inflammatory cytokine IL-8; and an increase of the antimicrobial peptide hBD-2 (24). Furthermore, early-stage research performed by our group has linked certain gut bacteria to heart transplant outcomes (25). Further mechanistic characterization of the role of gut bacteria is critical in advancing applications for allograft treatment regimens.

Microbiota Standards and Pitfalls: “Do or do not; there is no try” [Ref: Star Wars / Obi-Wan Kenobi].

Studies of microbial communities have recently been akin to a “fashion phenomenon” fueled by ever-decreasing sequencing costs (26), leading to the democratization of sequencing and the now-widespread perception that “anyone can sequence”. Sequencers, and sequence processing and analyses, have become widely available, cheaper and easier to operate, yielding increasing data output in a shorter amount of time. As a result, microbiome studies that were once too complex and prohibitively expensive can now be achieved in weeks for a few hundred dollars per sample. But despite the ever-increasing number of microbiome studies, standards have been slow to be established (27–29). The absence of such standards has plagued even the basic vocabulary used in microbiome research. For example, the synonymous use of “microbiome” and “microbiota” is common, but these two terms are not equivalent. “Microbiome” refers to the entire habitat and the set of microorganisms living within it, their genomes and the surrounding environmental conditions, whereas “microbiota” refers only to the assemblage of microorganisms present in a defined environment. Another term often misused is “metagenome”, which should be used to describe all the genetic material present in an environmental sample, consisting of the “genomes of many individual organisms”. A corollary is that microbial community surveys that rely on a single marker gene (e.g., the 16S rRNA gene) do not represent metagenomic analyses, since such approaches describe microbial community composition and do not provide information about the collective genomes of these communities, for which the term “metataxonomics” should be employed (30–33).

Thoughtful planning of microbiome experiments is also critical to ensure that the experimental design is tailored to the study hypotheses and research (27, 28, 34) questions. The strategy “sequence now and ask questions later” using previously collected convenience samples has often resulted in studies that are underpowered, confounded and/or lacking appropriate controls, ultimately affecting the conclusions that can be reached (35). A statistical plan along with power analyses should be formulated as part of the study design, before subject enrollment and sample collection (35). Study design planning is critical in order to select the most appropriate sequencing strategy (e.g. 16S rRNA gene sequencing vs. WGS metagenomic sequencing) (28). Another critical planning step is the inclusion of the appropriate negative and positive controls (36). Positive controls are necessary during both extraction and sequencing, to ensure the samples are characterized without introduced errors. The purpose of negative controls is to control for contaminations potentially introduced during sampling, extraction and sequencing, for example originating from lab equipment, kits, or cross-samples contamination. As laboratory techniques have become highly sensitive, the tools used efficiently detect contaminating DNA (exogenous DNA from sources other than the samples studied) and cross-contamination (contamination between samples) (36). While contamination might not be a prominent issue with high bacterial biomass samples (e.g. gut or oral microbiotas); low- to ultra-low bacterial biomass samples, such as placenta or healthy lung microbiota samples, are particularly sensitive to false positives (37). Lists of contaminants (12), best practices for sample collection and processing (27, 35, 38, 39), inclusion of positive and negative controls (12, 40), and post hoc bioinformatic analyses to remove contaminating sequences in silico (41), have been widely adapted by microbiome researchers.

Microbiota related studies have often (and wrongfully) interpreted the detection of microbial DNA as evidence of bacterial colonization, and thus the presence of living bacterial cells. While DNA-based methods form the basis of most microbiome strategies, they are generally not able to distinguish live from dead bacteria, or even cell-free relic DNA (42, 43). In sites previously thought to be sterile and devoid of microbial life, proving colonization and the sustained presence of micro-organisms requires experimental evidence of microbial metabolic activity, and that microbial communities persist over time. A corollary from these observations is that DNA-free and sterility are two different states (Figure 2) and should not be used interchangeably. An environment where the communities of microbes have been killed and rendered sterile, for example through autoclaving in the laboratory or by immune processes in the human body, may still contain bio-active DNA molecules that can be sequenced (Figure 2). In addition, there are many instances where bacterial DNA has been shown to be “cell-free”, or even present inside phagocytic cells (e.g. macrophages or dendritic cells). As such, a bacterial community profile can be established from a “sterile but not DNA-free” environmental sample. In addition, it is also well known that PCR inhibitors, such as hemoglobin or polysaccharides, can lead to the inability to characterize otherwise bio-active DNA molecules. Consequences of these observations are two-fold: 1) microbial community profiles may or may not be sterile, and 2) the inability to obtain a microbial community profile does not imply absence of microbial DNA or live microbial cells.

Figure 2. Dispelling the confusion: why sterility and lack of detection of bacterial DNA in microbiome studies are two different states that tell a fundamentally different story.

DNA-free and sterility are two completely different states and should not be used interchangeably, as is often the case in microbiome studies. Sterility is the state by which a given environment/sample is devoid of live organisms, which is usually assessed by growth on appropriate culture medium or other techniques assessing microbial metabolism. A sample where microbes have been killed and rendered sterile (i.e. through autoclaving in the laboratory or by immune processes in the human body) may still contain bio-active DNA molecules that can be extracted, 16S PCR amplified and sequenced. Therefore, a bacterial community profile can be established from a “sterile but not DNA-free” sample. In addition, PCR inhibitors, such as hemoglobin, humic acids (prominent in soil and stool samples) or polysaccharides, co-extracted with DNA molecules can lead to the inability to characterize otherwise bio-active DNA molecules, leading to the erroneous assumption that the sample was DNA-free. Consequences of these observations are two-fold: 1) environmental samples where microbial community profiles can be obtained through molecular methods may or may not be sterile, and 2) the inability to obtain a microbial community profile through molecular methods does not imply absence of microbial DNA or live microbial cells. 16S rRNA gene profile illustrations were reproduced from (12).

Challenges

Microbiome research involves the lengthy processing of complex biological samples, followed by sophisticated bioinformatic and statistical analyses of high-dimensional datasets in association with the sample-associated metadata (subjects’ demographic data, sample characteristics, experimental conditions, etc). Communicating all the details, from sample handling and processing, library preparation and sequencing, to details about the intricate bioinformatic and statistical analytic workflows, as well as the proper use of controls, is critical to ensure the reproducibility and interpretability, which are essential in microbiome studies (29, 44). Some journals are encouraging researchers to submit, along with their results and detailed protocols, the bioinformatic and biostatistics codes used to process and analyze the datasets, thus providing more complete transparency and reproducibility.

Concerns related to associations and causality are still very much prominent in the microbiome field (45). Studies, generally performed using techniques such as 16S rRNA gene sequencing, typically generate a long list of commensals implicated as “biomarkers” of disease. However, they usually cannot clearly assess how bacterial communities or select species might initiate or contribute to disease pathogenesis, or whether these shifts might just be a consequence of the disease with no specific relevance to it. To move beyond correlations and begin addressing causation, effective strategies and clinically-relevant models will be needed to design mechanistic studies where the precise role of specific bacteria can be thoroughly tested and functionally assessed. The in vitro and in vivo clinically-relevant models (46–49) to study the interactions between bacterial communities and the host are inherently limited in their ability to understand a system where interactions are infinitely more complex than the sum of their individual parts. Animal models can never fully model human diseases; however, they are amenable for the testing of specific organisms as causative factors in pathogenesis. Mice are by far the most used animal model, and two approaches have emerged to characterize the causal effects of the microbiota: germ-free mice and antibiotic treatment regimens (50).

Germ-free models have been critical for assessing the relationships between the gut microbiota and the host (51). Germ-free mice are bred in isolators, preventing exposure to environmental microorganisms in order to keep them free of detectable bacteria, viruses, and eukaryotic microbes (52). Germ-free mice allow physiological studies in the complete absence of microbes, or as the starting point for the generation of gnotobiotic animals in which only known microorganisms are present. An alternative strategy used to manipulate the microbiota has been the use of antibiotics treatment. Treatment with broad-spectrum antibiotics is commonly used to deplete bacterial populations in mice that were normally colonized since birth, and can be applied to any genotype or for any health or disease state. Gnotobiotic mice are generally thought of as the “golden standard” for microbiome manipulation, whereas antibiotics-treated mice model is largely considered a rapid, inexpensive, and more accessible alternative. However, these two approaches are not equivalent, and researchers should carefully weigh the limitations of each strategy. Gnotobiotic animals are broadly impaired in many aspects of development and early immune education, whereas antibiotics treatment in adult mice specifically allows for study of the role of bacteria while maintaining cell functionality and signaling pathways after development. Because of their development in the absence of bacteria, germ-free mice have a wide range of anatomical and physiological abnormalities (53): they generally have a cecum enlarged by 4 to 8-fold, their small intestines are less developed and have a considerably smaller surface area and slower peristalsis, irregular villi and reduced epithelial cell renewal, all of which contribute to deficiencies in nutrient utilization as well as lower weight compared to normally colonized mice (53). More importantly, germ-free mice (and by extent, gnotobiotic mice) have impaired immune responses to certain pathogens and an immature immune system, which includes smaller mesenteric lymph nodes and Peyer’s patches, decreased numbers of IL-17-producing T helper 17 (TH17) cells, and defects in regulatory T (Treg) cells (54). Limitations of antibiotic treatments include the inability to selectively deplete specific bacterial groups, the inability to completely clear bacteria (only significant reductions in bacterial load are possible), and the possibility that the antibiotics themselves trigger shifts in cell populations, signaling pathways, and organ morphology. Therefore, researchers should be careful in their interpretation of the research outcome.

The trifecta in organ transplantation: inter-connection between the immune system, host metabolism and the microbiota

The human microbiota has also been shown to have a critical role in maintaining health and triggering diseases. Microbiota-immune system interactions occur both during homeostasis and inflammation, and this bidirectional dialogue directly affects the diversity of microbial communities and the state of the immune system (Figure 3). It also helps in shaping the gut-epithelial barrier, induces cytokine responses and a broad spectrum of cellular responses, including proinflammatory Th1 and Th17 responses (55, 56), anti-inflammatory Treg responses (57–60) and T cell dependent IgA responses (61). In this section we will discuss the effects of the microbiota and its metabolites on the intestinal epithelium and cells of the immune system. In particular, we will describe the roles of the gut microbiota in lymph node (LN) remodeling and its impact on allograft survival. Finally, we will discuss potential approaches to target the microbiota for therapeutic benefits and the value of the microbiota as a diagnostic or prognostic biomarker for transplantation outcomes.

Figure 3. Interplay between the microbiota, its human host, and the effects from environmental factors, aimed at maintaining the balance between health and diseases.

Health results from a balance of factors affecting the human host, which include genetics, age, immune status, as well as environmental factors (such as hygiene, diet, stress, environmental exposures etc). The human microbiota has also been shown to have a critical role in maintaining health and triggering diseases, and recent studies aimed at characterization of the microbiota has highlighted the critical effects of the microbiota-immune system interactions, occurring both during homeostasis and inflammation. This bi-directional dialogue directly affects the diversity and composition of the microbial communities colonizing the host, but also directly impacts both innate and adaptive arms of the immune system.

Effects of the microbiota and its metabolites on the epithelium

The gut microbiota and its metabolites directly interact with intestinal epithelial cells (IECs) and/or with leukocytes to modulate mucosal barriers and induce signals regulating immune cells in the lamina propria. Ontogeny of the gut microbiota is important in establishing the normal development of intestinal mucosa. Hughes et al. (62) found that, in neonatal mice, microbiota-epithelial interactions regulate and maintain the intestinal barrier and play an important role in shaping the immune system. Disturbances in these communities are linked to dysregulated renewal and replenishment of IECs, and associated with the progressive development of autoimmune diseases such as arthritis (63). Day 14 neonatal mice had increased primary bile acids cholate and taurocholate and a high abundance of Bifidobacterium, Streptococcus, Escherichia, and Enterococcus. In contrast, older day 21 and day 28 mice were dominated by Candidatus Arthromitus, Alistipes, and Lachnoclostridium, and had increased concentrations of the short-chain fatty acid (SCFA) propionate and trimethylamine. Treatment with antibiotics, to disrupt neonatal (day 10–14 mice) microbiota, followed by lipopolysaccharide (LPS) stimulation increased apoptotic cell death in the IEC. While not mechanistically focused, this study linked microbiota composition and its associated metabolites to integrity of the gut epithelial barrier in neonatal mice (62). In a similar approach using the suckling rabbit model to investigate how microbiota colonization drives the ontogeny of gut barrier function, Beaumont and colleagues (64) reported that introduction of solid food induced a shift in the gut microbiota from Bacteroidetes (milk-oriented microbiota) to Firmicutes and Lachnospiraceae, a concomitant shift in metabolites towards the SCFAs acetate and butyrate, and an acceleration of gut epithelial barrier formation. Solid food consumption also downregulated expression of the tight junction proteins occludin and claudin, decreased ex vivo paracellular-permeability to FITC dextran, reduced the expression of Toll-like receptors (TLR)2 and TLR5, IL-4, tumor necrosis factor alpha (TNFα) and transforming growth factor β1 and increased the expression of CCL20 in the cecal mucosa. Thus, targeting the metabolic activity of the gut microbiota during this key developmental window or even at other times might be a promising strategy to promote intestinal barrier function.

Gut microbiota metabolites also regulate gut homeostasis. In patients with ileal Crohn’s disease, gut microbiota-derived succinate induces proliferation of intestinal tuft cells. Tuft cells are involved in responding to parasitic infection by releasing IL-25 (65). In the ileum of mice treated with succinate, tuft cells were shown to reduce intestinal inflammation by increasing Gata3+ cells and type 2 cytokines (IL-22, IL-25, IL-13), and by decreasing Rorγt⁺ cells and type 17 cytokines (IL-23) (66). Scott et al. (67) demonstrated that in mouse and human gut microbiotas, metabolites derived from tryptophan metabolism by Lactobacillus reuteri and Clostridium sporogenes regulate gut barrier functions through aryl hydrocarbon receptor-mediated signaling in the gut epithelium. Nakamura and colleagues (68) studied the polyamine metabolites putrescine and spermidine produced by intestinal E. coli. Spermidine serves as a substrate for hypusine and induces hypusination of eukaryotic translation initiation factor 5A (eIF5A). Hypusinated eIF5A is involved in translation initiation, elongation, termination, and the cell cycle. Polyamine-driven metabolic reprogramming induced colonic epithelial proliferation, and polyamines led to the suppression of dextran sulfate sodium (DSS)-induced colitis in mice.

In addition to regulating the gut epithelium in its immediate vicinity, the microbiota and its metabolites also affect distant organs. A recent study (69) hypothesized that diversity in antibody repertoire is positively correlated with microbiota diversity. Using a B cell antigen receptor-transgenic mouse strain (MD4) with a restricted antibody repertoire against hen egg lysozyme, the autrhors showed that MD4 mice have reduced microbial diversity, and an increased susceptibility to inflammatory disorders. More importantly, gut microbiota metabolism of L-tyrosine in this model reduced CCL20 production by inhibiting epidermal growth factor receptor and TLR4 signaling airway epithelial cells (AECs), demonstrating that microbiota-derived metabolites can act distally in reducing allergic airway responses. In another study, Chaudhari et al. (70) characterized the gut-liver pathway associated with changes in gut metabolites following bariatric surgery used to treat type 2 diabetes. Following sleeve gastrectomy in mice, which reduces the size of the stomach by 70% and induces weight-loss, the investigators found an increase in the transport of the microbial metabolite lithocholic acid (LCA) from the gut to the liver via the portal vein. They dissected the link between increased LCA production in the gut and the secretion of the incretin hormone glucagon-like peptide-1 (GLP-1) through regulation of the LCA-CA7S-TGR5-GLP-1 pathway, thus characterizing the metabolic benefits after sleeve gastrectomy. Microbiota transfer from post-sleeve gastrectomy mice into germ-free diabetic recipients resulted in enhanced glucose tolerance and improved diabetic control, thus establishing causality, but the authors could not identify the specific bacteria responsible for this effect.

Taken together, these investigations characterized the gut microbiota and/or its metabolites as important determinants of intestinal epithelial cell function. Signals released by intestinal epithelial cells or microbial metabolites have far reaching effects in various normal and pathological conditions (Figure 4).

Figure 4. Current knowledge about the pathways linking the gut microbiome to organ inflammation and fibrosis in solid organ transplant outcome.

Our previous studies using a clinically-relevant MHC-mismatched mouse model of chronic allograft rejection (A) have shown that the gut microbiota broadly affects the innate and adaptive immune systems, as well as lymph node (LN) remodeling, with critical downstream consequences on vascularized cardiac allografts outcome (B). In these studies, the murine gut microbiota was altered using FMT of stool samples from colitic mice, representing an inflammatory state, or from pregnant mice, representing an immune suppressed state (A). Transplanted mice that received FMT of colitic stool samples demonstrated poor cardiac graft survival, whereas transplanted mice that received FMT from pregnant mice demonstrated reduced inflammation and fibrosis, and improved graft survival. Differential abundance analyses identified Desulfovibrio bacteria as enriched in colitic gut microbiota and B. pseudolongum enriched in pregnant gut microbiota. Single species gavage with either Desulfovibrio or B. pseudolongum bacteria was sufficient to reproduce the effects observed on transplant outcomes with colitic or pregnant stool FMT, respectively (B). Additional in vitro experiments showed that pro-inflammatory signals in the gut induce strong IL-6 and TNFα responses and inhibits homeostatic IL-10 secretion from DCs and macrophages, resulting in a reduced ratio of stromal laminin α4:α5 fibers in the cortical ridge of the LN, a morphologic change in LNs associated with immunologic activation and graft rejection. In contrast, anti-inflammatory signals in the gut induced anti-inflammatory cytokine responses from DCs and macrophages by maintaining IL-10 secretion and inducing only low-levels of IL-6 and TNFα, resulting in an increased laminin α4:α5 ratio in the cortical ridge of the LN, associated with immunologic suppression and tolerance (B). Question marks denote the steps in these activation pathways that are still under active investigation.

Effects of microbiota and its metabolites on immune cells

The gut microbiota and metabolite-driven signals shape host innate and adaptive immunity. Innate myeloid cells in the intestinal mucosa regulate the activation and differentiation of both innate and adaptive lymphoid cells. Th1 and Th17 adaptive responses provide defense against harmful microbes, while suppressive Tregs induce mucosal tolerance.

The gut microbiota and its metabolites shape macrophage signature and function. Schulthess et al. (71) showed that microbiota-derived butyrate induce antimicrobial activity in macrophages. This activity is associated with a shift in metabolism towards reduced mTOR kinase activity, increased autophagy associated protein LC3-II turnover, and anti-microbial peptide calprotectin production in the absence of an increased inflammatory cytokine IL-1β and TNFα response. Butyrate also drives monocyte to macrophage differentiation through histone deacetylase 3 inhibition. Nakamura et al. (68) demonstrated that commensal E. coli produce putrescine to regulate macrophage differentiation and enhance anti-inflammatory macrophages in the colon that inhibited DSS-induced colitis.

An et al. (72) showed that Bacteroides fragilis modifies the homeostasis of host invariant natural killer T (iNKT) cells by supplementing host endogenous lipids with unique inhibitory sphingolipids. They found that B. fragilis-derived sphingolipids present during neonatal development inhibit IL-2 secretion and iNKT cell proliferation as well as their accumulation in colon lamina propria (72). Oxazolone-induced colitis, a model for ulcerative colitis that is induced by iNKT derived IL-13, was not controlled in germ-free mice monocolonized with B. fragilis lacking serine palmitoyltransferase, an enzyme required for sphingolipid biosynthesis. This indicates a direct link between bacterial sphingolipids and iNKT cell homeostasis and disease susceptibility (72).

Several studies have documented the beneficial effects of the gut microbiota and/or its metabolites in driving Treg induction and function for mucosal tolerance (73, 74). The microbiota mediates the transformation of cholic acid to the secondary bile acid 3β-hydroxydeoxycholic acid (isoDCA) (75), which then acts through the farnesoid X receptor (FXR) on DCs to diminish their immunostimulatory properties, resulting in an FXR-dependent increase in Foxp3 induction in RORγt-expressing Tregs. Tregs have been shown to recognize antigens from Akkermansia muciniphila, a commensal bacteria that induces anergy and drives conversion of naive CD4⁺CD44⁻Foxp3⁻ T cells to the Treg lineage (76). Colonic Tregs highly express GPCR43, encoded by Ffar2, which binds SCFAs (77). Smith et al. (78) demonstrated that the Clostridium clusters XI (Clostridium bifermentans), XIV (ASF 356 and 492) and XVII (C. ramosum) and the Bacteroides species B. fragilis produced SCFAs propionate and acetate which regulate the size and function of the colonic Treg pool. Propionate also increases GPCR43 expression on Tregs (78). Butyrate produced by the colonic microbiota induces the differentiation of Tregs in mice (79), through histone H3 acetylation in the promoter and conserved non-coding sequence regions of the Foxp3 locus. Other bacterial components regulate the mucosal immune system. Verma and colleagues (80) identified Bifidobacterium bifidum (B. bifidum) as a potent inducer of Tregs via its cell surface β-glucan/galactan polysaccharides acting on regulatory DCs and partially through a TLR2-mediated mechanism. Sun and colleagues (81) also identified a Bifidobacterium mixture (B.bifidum, B. longum, B. lactis, and B. breve) that altered the composition of the gut microbiota in a Treg-dependent manner, consequently enhancing both mitochondrial fitness and IL-10–mediated suppressive functions of intestinal Tregs, ultimately ameliorating colitis during immune checkpoint blockade.

Taken together, these studies demonstrate that the microbiota and its derivatives trigger pro- or anti-inflammatory effects on cells of adaptive and innate immune systems (Table 1) and regulate local intestinal and systemic immune homeostasis.

Table 1.

Immunomodulation by microbiota-derived metabolites

Metabolites	Immune cells	Effect on Immune cells	References
Butyrate	Macrophages	Monocyte to macrophage differentiation by HDAC3 inhibition	[71]
Putrescine		Macrophage differentiation and activation	[68]
Indole	Dendritic cells	AhR receptor mediated activation	[108]
Sphingolipids	iNKT cells	Inhibition of iNKT cell proliferation	[72]
isoDCA	Tregs	Increase in RORγT+ Tregs	[75]
SCFA		Control of size and function of colonic Tregs in a GPCR43 dependent manner	[78]
Butyrate		Increase in histone H3 acetylation	[79]
Cell Surface β-glucan/ galactan		Treg induction in a TLR2 dependent manner	[80]

Effects of microbiota and its metabolites on lymph node structure

Gut epithelial cells respond to microbes and their metabolites by modulating mucosal barriers and relaying signals to the immune cells in the lamina propria. Migration of these leukocytes may then induce regional or systemic changes in the immune system, which may be manifested in the LN structure and function in the form of an expanded fibroblastic reticular cell network critical for T cell migration, differentiation, and function (82–84). Acute infections with Yersinia pseudotuberculosis have been shown to cause sustained inflammation, even following clearance, resulting in the migration of DCs to adipose tissues rather than the mesenteric LNs, and compromised mucosal immune tolerance and protective immunity (85). These persistent and pathologic changes in the reticular cell network are termed “immunological scarring”.

We studied the effect of the gut microbiota on vascularized cardiac allografts outcome, in a clinically-relevant MHC-mismatched mouse model of chronic allograft rejection (25). Mice gut microbiota altered using FMT of stool samples from colitic mice, which represent an inflammatory state, demonstrated poor graft survival. The colitic microbiota is enriched with Desulfovibrio desulfuricans, a pro-inflammatory bacterium that induces strong IL-6 and TNFα responses and inhibits homeostatic IL-10 secretion from DCs and macrophages. Colitic FMT also resulted in a reduced ratio of stromal laminin alpha 4: alpha 5 (α4:α5) fibers in the cortical ridge of the LN, a morphologic change in LNs associated with immunologic activation and graft rejection (84, 86) (Figure 4). In contrast, FMT of stool samples from pregnant mice, representing an immune suppressed state, reduced inflammation and fibrosis in the graft and improved graft survival. Microbiota from pregnant mouse stool was enriched in B. pseudolongum. B. pseudolongum induces anti-inflammatory cytokine responses from DCs and macrophages by maintaining IL-10 secretion and inducing only low-levels of IL-6 and TNFα. Pregnant FMT resulted in an increased laminin α4:α5 ratio in the cortical ridge of the LN, associated with immunologic suppression and tolerance (84, 86) (Figure 4). Our study demonstrates that the gut microbiota broadly affects the innate and adaptive immune systems and LN remodeling (25). Moreover, disrupting the microbiota content via antibiotics alone can improve graft survival. Lei et al. (87) assessed the effects of the microbiota on alloimmunity in a male to female, minor antigen–mismatched, skin graft mouse model. Allograft survival was prolonged in both germ-free and antibiotic-pretreated mice. Antibiotic treatment impaired the activation of the type I interferon pathway in DCs and inhibited their capacity to prime alloreactive T cells. We also observed significant changes in the laminin α4:α5 ratio in mice treated with antibiotics (25), suggesting diverse effects on immunity.

Gut microbiota mediated modulation of immunity and its effect on rejection and tolerance in organ transplantation

Graft-restricted microbiota can influence outcomes of the transplanted organ. In a skin allograft model, Lei and colleagues demonstrated that mono-species colonization of the skin with the commensal Staphylococcus epidermidis in germ-free mice was sufficient to accelerate skin graft rejection. S. epidermidis does not alter the ability of DC to prime alloreactive T cells in the LN, but rather enhances the local secretion of TNFα, IL-12A, and IL-18 in the skin graft, indicative of Th1 mediated allograft rejection. This study demonstrates that the microbiota of the transplanted organ can shape the host’s alloreactive immune response (88).

In a prospective study of 134 human lung transplant recipients, Combs et al. (89) analyzed the respiratory microbiome in bronchoalveolar lavage (BAL) fluid one year post lung transplantation. Analysis of bacterial community composition revealed that the lung bacterial burden is predictive of chronic lung allograft dysfunction (CLAD) or death within 500 days. Although no specific bacterial taxa were found to be responsible for CLAD or death, associative data on the differences in microbiota composition indicate a potentially modifiable risk factor for lung allograft dysfunction. In a smaller cohort of 6 single lung transplant recipients, Sharma et al. (90) investigated the host-microbiome interactions of the allograft and native lung, and found enhanced levels of vascular endothelial growth factor (VEGF) and the neutrophil chemoattractant matrikine N-acetyl Proline-Glycine-Proline in the allograft along with an abundance of Acinetobacter and Pseudomonas and sphingosine-like metabolites in the airway. Since allografts have a distinct microbiome and metabolome signature compared to the native lung, these data suggest a functional association between the airway metabolome and local microbiome. Focusing on gene expression of the microbiome, Zinter et al. (91) studied BAL samples from 181 patients with lung injury after pediatric allogeneic hematopoietic cell transplantation. Microbiome gene expression analysis showed that a decrease in microbial load pre-transplant was associated with the activation of epithelial-epidermal differentiation, mucus production, and cellular adhesion in lungs that lead to fatal post-transplant lung injury. This study identifies the value of pre-transplant assessment as a biomarker of high-risk pediatric candidates. Further underscoring the significance of patients’ pre-transplant bacterial status, Das and colleagues (92) studied BAL samples from 64 lung transplant recipients and classified bacterial communities into four states (pneumotypes) associated with various degrees of inflammatory phenotypes and rejection outcomes. Taken together, these studies highlight the importance of allograft microbiota assessment for monitoring allograft health. In fact, Westblade et al. (93) examined the prevalence of gastrointestinal pathogens in 142 non-diarrheal kidney transplant recipients within the first 10 days post-transplantation, which led to the identification of Clostridioides difficile, Escherichia coli, and Norovirus in 43 subjects that were interpreted as biomarkers of gut dysbiosis.

Taken together, these studies demonstrate that commensal microbiota play a significant role in allograft health and disease. The beneficial effects of microbiota-mediated immunomodulation may be manipulated to direct the management of host physiology and organ transplant outcomes.

The Gut microbiome as a therapeutic target

Perturbations in the microbiota have been linked to inflammatory diseases and autoimmune disorders, such as inflammatory bowel disease (94), rheumatoid arthritis (95), and multiple sclerosis (MS) (96), celiac disease (97), among others. An understanding of how changes in the microbiota influence these conditions may help in the development of biomarkers and therapeutic regimens. This has been applied for example in the treatment of MS patients, where beta-interferon or glatiramer acetate leads to an increased abundance of anti-inflammatory Prevotella and Sutterella and reduced numbers of pro-inflammatory Sarcina, compared with untreated patients (96), with these changes correlated with interferon and NFκB signaling in T cells and monocytes and the maturation of DCs (96). This and other studies provide rationales for designing therapeutic interventions targeting microbiota and the characterization of the gut microbiota for diagnostic, prognostic or therapeutic biomarkers.

Among solid organs transplant recipients, the diversity and richness of the gut microbiota may act as a biomarker for transplant outcome. Kidney transplant recipients were found to have reduced bacterial diversity compared to healthy individuals (98, 99). We previously characterized the urinary microbiota associated with spontaneous tolerance in human subjects with stable kidney graft function and identified a highly diverse microbiota with abundance of Proteobacteria among spontaneously tolerant patients, particularly male patients, 6–12 months post-transplantation (100). This study suggests the importance of continuous monitoring and treatment adjustment to maintain microbiota diversity and richness. In the context of liver transplant recipients, acute cellular rejection was found to be associated with an increased abundance of Bacteroides, Bifidobacterium, Streptococcus, Lactococcus, and Blautia and a decrease in the population of Enterococcus, Lactobacillus, Clostridium, Peptostreptococcus, Parabacteroides, Dorea, Megasphaera, Veillonella, and Megamonas (101). Furthermore, in a study of intestinal transplant recipients, an increase in the abundance of Escherichia and Klebsiella was associated with rejection (102). Taken together, these reports indicate that the microbiota in transplant recipients is of diagnostic value for transplant health, and that intervention strategies designed to modulate the gut microbiota might be beneficial to improve transplant outcomes.

Recent investigations were focused on the effects of immunosuppressive medications on the microbiota. Zhang et al. (103) analyzed the effects of high dose of tacrolimus (10 mg/kg; orally) on gut microbiota in a mouse model. After 14 days of treatment, 16S rRNA sequencing of fecal microbiota revealed an increase in the abundance of Allobaculum, Bacteroides, and Lactobacillus. Supplementation with these bacteria has been shown to be beneficial during inflammatory diseases (104). Tacrolimus also enhanced Tregs in the colonic mucosa and in circulation. FMT from mice treated with high dose tacrolimus, which had an abundance of Bacteroides, when given to a recipient receiving only a low dose of tacrolimus inhibited alloimmunity and prolonged the survival of skin allografts. This study demonstrates a strategy where a combination of immunomodulatory drugs and FMT can improve transplant outcome. In other investigations, antibiotics prolonged liver, islet, lung, and skin allograft survival (87, 105–107) and, FMT from antibiotic-treated mice inhibited skin allograft rejection in germ-free recipients (87). We observed prolonged cardiac allograft survival in mice receiving an anti-inflammatory FMT from pregnant mice (25). These studies demonstrated that therapeutic regimens targeting microbiota could lead to improvements in transplant outcomes.

The gut microbiota and its metabolites influence intestinal epithelium integrity (Figure 5). Signals received by intestinal epithelia are relayed locally and systemically to cells of the immune system. These signals then regulate leukocyte migration and function, which subsequently regulate other immune events such as the structure of secondary lymphoid organs. Changes in the microbiome, luminal microbiome-derived molecules, and inflammatory signatures, including LN remodeling, are beginning to be explored as prognostic and diagnostic biomarkers of transplantation outcome or as therapeutic adjuncts. A detailed understanding of the sustainable mutualism between microbiota and the host immune system could provide essential tools for their modulation. On one hand, we need to understand the secretome and/or surface-associated peptides of the microbial population that could be targeted as potential mediators of microbe-host interactions, and molecular cross-talk between bacterial colonization and mucosa-associated immune cells. On the other hand, we need to further understand the essential processes involved in induced productions of peptides in the intestine by epithelial or immune cells, due to FMT or environment perturbation, which are key elements in orchestrating interactions among gut microbial community and host cells. By harnessing the power to manipulate microbial-immune system interactions and mediating the dynamic interplay between microbes and the human host, significant enhancement of organ function, including tolerance, could be achieved following transplantation.

Figure 5. Effect of microbiota and metabolites on gut homeostasis.

The gut microbiota and its metabolites influence gut homeostasis. Signals received by intestinal epithelia are relayed locally and systemically to cells lining the intestine and cells of the immune system. These signals are important in the maintenance of gut barrier function, proliferation of epithelial cells, resolution of intestinal inflammation, and immunomodulation of cells of the immune system.

Conclusions: implications for research and future therapies

Within the last few years, microbiome research has gone from the humble beginnings of just a few projects awarded by the NIH-funded HMP initiative to >10,000 projects funded yearly in the US, and has become an essential tool that any laboratory must possess. This explosion in the number of microbiome projects is an unequivocal testament to the critical importance the microbes living in and on the human body have on health and diseases. Although the field has matured considerably, its applications for prognostics, diagnostics and clinical interventions are still unclear. While microbiome interventions such as fecal microbiota transplants have been transformative and are now standard of care for the treatment of recurrent Clostridium difficile infections, designing interventions beyond FMT will require a better understanding of the precise role the microbiome has on promoting health or initiating and promoting diseases. Developing interventions targeting specific microbiota pathways will undoubtedly take some time, as the interactions between the host and the microbiome are highly complex and highly intertwined. As pathophysiological mechanisms involving the microbiome become more defined, and specific bacteria and/or metabolites are identified, new drug-based interventions could become the next generation of microbiome-informed therapies. Because of the individual nature of the microbiome, it is likely that some interventions will need to be personalized based on a given person’s microbiome characteristics.

Summary.

• Microbiota and its metabolites are key determinants of transplant outcome.

• The gut microbiota relays signals to epithelial cells, which shape the immune system.

• Microbiota could be targeted for immunomodulation and to improve transplant outcomes.

• Microbiota is a next generation medicine and should be explored as a novel therapeutic in transplantation.