Gut associated metabolites and their roles in Clostridioides difficile pathogenesis

ABSTRACT The nosocomial pathogen Clostridioides difficile is a burden to the healthcare system. Gut microbiome disruption, most commonly by broad-spectrum antibiotic treatment, is well established to generate a state that is susceptible to CDI. A variety of metabolites produced by the host and/or gut microbiota have been shown to interact with C. difficile. Certain bile acids promote/inhibit germination while other cholesterol-derived compounds and amino acids used in the Stickland metabolic pathway affect growth and CDI colonization. Short chain fatty acids maintain intestinal barrier integrity and a myriad of other metabolic compounds are used as nutritional sources or used by C. difficile to inhibit or outcompete other bacteria in the gut. As the move toward non-antibiotic CDI treatment takes place, a deeper understanding of interactions between C. difficile and the host’s gut microbiome and metabolites becomes more relevant.


Introduction
Clostridioides difficile is a Gram-positive, pathogenic, spore forming, anaerobic bacterium that is considered the main cause of antibioticassociated diarrhea, pseudomembranous colitis and toxic megacolon. 1 According to the Centers for Disease Control and Prevention (CDC), C. difficile is a major nosocomial pathogen with more than 220,000 infections, 13,000 deaths, and nearly $5 billion in annual treatment associated costs that are predicted to increase in the future. 2,3 This, combined with its inherent natural antibiotic resistance, led to the CDC classifying C. difficile as an 'urgent threat' to the United States healthcare system. The greatest risk factor for C. difficile infection (CDI) is prior treatment with broad-spectrum antibiotics. Antibiotics render the host susceptible to CDI by changing the ecology of the microbiota, which is known to provide 'colonization resistance' against invading pathogens, C. difficileincluded. 4,5 Interestingly, infants and newborns can be C. difficile carriers 6 and C. difficile can be found in approximately 53% of healthy adults. 7 Thus, C. difficile could be considered a part of the microbiome in humans, but this is mostly defined by age. 8 Endospores are metabolically dormant forms of spore-forming bacteria produced in response to stress (e.g., nutrient limitation). [9][10][11] For C. difficile, the spore form is essential for host-to-host transmission due to the strictly anaerobic nature of the vegetative form. [12][13][14] Thus, in susceptible hosts, C. difficile spores must germinate into the active, vegetative form in order to multiply and cause disease. Germination by C. difficile spores is triggered upon recognition of certain bile acids and amino acids by germinant receptors. 11,15,16 Dormant C. difficile spores are considered infectious agents, but the growing C. difficile vegetative cells secrete the TcdA and TcdB toxins that are responsible for the primary symptoms of disease (e.g., diarrhea and colitis); 14,[17][18][19] while some C. difficile strains secrete a third toxin, a binary toxin. 20 The most common treatments for CDI are broad-spectrum antibiotics like vancomycin, although fidaxomicin (a narrower-spectrum antibiotic)is now recommended for initial and recurrent CDI . Unfortunately, the higher cost of fidaxomicin is limiting its use, and the continued alteration to the colonic microbiota by these antibiotics, lead to patients experiencing recurring disease due to the presence of C. difficile spores within the gastrointestinal tract or in the surrounding environment. 21,22 Recently, fecal microbial transplantation (FMT) has emerged as an effective treatment against C. difficile, especially for patients with recurrent CDI. FMT is hypothesized to drive protection against CDI through the restoration of gut microbes, gut associated metabolites [e.g., secondary bile acids and short chain fatty acids (SCFA)] 23 and/or microbial bile acid biotransformation . Nevertheless, the use of FMT is not without risk for the potential introduction of harmful microbes to the recipient. [24][25][26] Thus, effective and safe treatments for CDI that result in the restoration of the protective metabolic effects provided by the microbiome still need to be developed. In this review, we summarize studies focusing on metabolites produced by the host or gut microbiota and how these molecules influence C. difficile pathogenesis.

Bile acid metabolites and their role in C. difficile pathogenesis
Bile acids are cholesterol-based molecules synthesized by hepatocytes in the liver that play key roles in regulating metabolic pathways and aiding in the absorption of fats and cholesterol during digestion. 27,28 There are two pathways for bile acid synthesis: the classical pathway generates both cholic acid (CA) and chenodeoxycholic acid (CDCA) while the alternative bile acid synthesis pathway only forms CDCA. 29 Before being secreted into the intestines, primary bile acids are conjugated with either taurine or glycine at C-24 to generate taurocholic acid (TA)/taurochenodeoxycholic acid (TCDCA) and glycocholic (GCA) acid/ glycochenodeoxycholic acid (GCDCA) ( Figure 1). 30,31 Reabsorption of bile acids takes place throughout the intestines and the reabsorbed bile acids are recycled back to the liver for additional rounds of digestion. 29,31 During this enterohepatic recirculation pathway, bile acid synthesis is regulated through the Farsenoid X receptor (FXR) transcriptional activator 32 . Bile acid binding principally through CDCA to FXR generates a conformational change in FXR that leads to the synthesis of FGF19 in illeal enterocytes 32,33 . FGF19 is secreted by the ileal enterocytes and binds to the FGFR4 receptor on the cell surface of hepatocytes. Binding of FGF19 to the receptor leads to a negative feedback regulation on bile acid synthesis. 32,33 In an FMT model for recurrent CDI patients, levels of FGF19 were significantly increased post-FMT, thus suggesting upregulation of the FXR-FGF pathway after FMT that may aid in C. difficile clearance. 34 Furthermore administration of obeticholic acid, an FXR agonist, to HFD CDI-infected mice resulted in decreases in disease severity, and the use of ursodeoxycholate (UCA) in a CDI mouse model showed increased levels of FXR as well as TGR5 that modulates the immune response against C. difficile. 35,36 The mentioned studies on FXR suggest that FXR modulation may be important for CDI treatment but further research is necessary to move past the correlation observed in FXR that shows protection against CDI.
A small percentahe of bile acids escape enterohepatic recirculation, and can be deconjugated by the bile salt hydrolases that are expressed in many gut microbes. Subsequently, deconjugated bile acids can be 7α-dehydroxylated by a small subset of gut microbial species to generate secondary bile acids (deoxycholic acid (DCA), from CA, and lithocholic acid (LCA) from CDCA) ( Figure 1).
Cholic acid-derived bile acids promote C. difficile germination, whereas bile acids of the CDCA family act as inhibitors of C. difficile germination. 16,37,38 In addition to acting as activators/inhibitors of C. difficile spore germination, bile acids affect C. difficile vegetative growth. The hostderived, primary, bile acids TA and CDCA have different effects on C. difficile growth. Vegetative cell growth seems to be unaffected by TA, however C. difficile growth is strongly inhibited by CDCA. 16 Interestingly, DCA, a product of the 7αdehydroxylation of CA by the colonic microbiota, is also toxic to C. difficile growth in-vitro, suggesting that the 7α-dehydroxylation of CA by the colonic microbiota may prevent C. difficile from colonizing healthy hosts. 16 When the minimal inhibitory concentrations (MIC) of various bile acids were determined, the MIC of CA was 10 mM for 2 different C. difficile strains, a concentration that is not found in the GI tract, thus not physiologically relevant. 39 However, increasing the hydrophobicity of the base CA structure (which has three hydroxyl groups, Figure 1), by removing either the 12α hydroxyl (generating CDCA) or the 7α hydroxyl (generating DCA) resulted in the MIC decreasing from 10-fold to 1 mM. 39 Moreover, the 7β-epimer of CDCA, UCA) ( Figure 1), inhibits growth of C. difficile cells, 40 although over a 24-h time period, C. difficile grew in the presence of 2 mM UCA. 41 Because CDCA completely inhibited the growth of C. difficile vegetative cells over a 24-h period and UCA did not, this could imply that stereochemistry of the bile acid structure plays a role in toxicity. 39,40 Importantly, though these two studies were done using different C. difficile isolates, which may have different MIC values for bile acids as observed for other C. difficile strains. 39 Mice are one of the most common animal models used in CDI studies and their murine bile acid composition and their effects on CDI are well understood. 39,42 In rodents, the presence of alter-native hydroxylating enzymes yields other primary bile acids (α/β/ω-muricholic acids) ( Figure 1). 43 Muricholic acids have a MIC similar to that of CDCA and DCA. 39 Similar to CA, muricholic acids are also trihydroxyl bile acids but the hydroxyls are in the 3, 6, and 7 positions instead of the 3, 7, and 12 positions that are found on CA-derived bile acids (Figure 1). 39 A study found that in addition to the bile acids described above, isodeoxycholic acid (a 7α, 12α-hydroxy bile acid), lithocholic acid (a 3α-hydroxy bile acid), isolithocholic acid (a 3β-hydroxy bile acid), and hyodeoxycholic acid also inhibit the growth of C. difficile vegetative cells. 44 Furthermore, isoallolithocholic acid (isoalloLCA), the isomer of isoLCA had an MIC 90 of 2 µM in C. difficile CD630 (a laboratory strain) and in the highly toxigenic C. difficile VPI10463 strain. 45 Due to the inherent toxicity of secondary bile acids for C. difficile growth, it was hypothesized that their production would toxify an environment and limit C. difficile growth. 16,46 To test this hypothesis, Theriot and Young 5 showed evidence that the presence of secondary bile acids is responsible for protection against CDI. Using a multiomics approach, the authors identified C. difficile resistant and susceptible states in an antibiotictreated murine model. A susceptible state correlated with high levels of conjugated primary bile acids (e.g., TA) whereas a resistance state correlated with higher levels of secondary bile acids (e.g., DCA). Later, Buffie et al. 4 also provided compelling data in support of this hypothesis. Using mouse models and human subjects, the authors found that high levels of DCA or the presence of the genetic operon that is known to generate secondary bile acids (bai), strongly correlated with a protective environment. Additionally, mice treated with different antibiotics resulted in variable metabolic profiles and certain secondary bile acids (i.e., omega-muricholic acid, LCA and, to a lesser extent, hyodeoxycholic acid, and UCA) that could inhibit TA-and DCA-mediated colony formation by C. difficile spores. 16,38,39,47 These landmark studies demonstrated that increased secondary bile acids and the genes encoding proteins responsible for their production strongly correlate with disease-resistant states and increased primary bile acids correlate with diseasesusceptible states. A more recent article 45 analyzing metabolites in centennials collected stool from centennials(~107 year olds), elderly adults (85-89 years old), and young adults (21-55 years old) and bile acids were measured. The authors observed higher levels of isoLCA, 3-oxoLCA, alloLCA, 3-oxoalloLCA, and isoalloLCA in the centennial samples, when compared to the other study population. Additionally, the authors characterized the biosynthetic pathways to generate these CDCA derivatives and suggested that further elucidation of microbial biotransformations may expand our understanding of intestinal homeostasis and, as a result, possible protection effects of additional bile acids against CDI 45 Interestingly, a recent study using a cholatedeficient mouse strain showed that the cholatederived secondary bile acid (DCA) is dispensable for protection against CDI, and that lack of the bile acid transcriptional activator FXR, does not show differences in CDI mice. Also, when germ free mice were monoassociated with bacteria known to generate secondary bile acids, no secondary bile acids were present in fecal samples, but the mice were still protected against CDI. 48 This emerging evidence would suggest that although bile acids are important for C. difficile colonization to be established in the gut, these metabolites may not play such an important role in generating a protective environment against C. difficile. In support of this, a 2018 study 49 showed that the levels of the conjugated primary bile acids TA and TCDCA decreased 2 days post-FMT. In contrast, there was an increase of the secondary bile acids DCA, LCA, and UDCA. 49 Interestingly, the authors note that although levels of secondary bile acids increased, the presence of secondary bile acids was not indicative of FMT success or failure, thus suggesting that although bile acids may play a role in protection, there are other mechanisms driving CDI resistance. 49,50 Regardless, C. difficile encounters bile acid during colonization and its growth in the gut may impact bile acid synthesis in the liver. Taking advantage of MALDI imaging mass spectrometry, Wexler et al. 51 found the levels of primary bile acids in the mouse gut increased significantly as early as one-day post-CDI. Additionally, the authors found that the introduction of cholestyramine, a bile-acid sequestering drug, lead to delayed C. difficile colonization. 51 Their results suggest that primary bile acids are required to efficiently establish C. difficile in a host.
Regarding secondary bile acids, in addition to its toxic effect against C. difficile vegetative cells, the secondary bile salt DCA induces C. difficile biofilm formation at low concentrations (240 µM) in a nonbiofilm-producing C. difficile CD630∆erm strain. 52 Thus, it is possible that a decrease in bile acid levels during late infection stages may generate favorable conditions for biofilm formation. Indeed, the authors hypothesized that the low concentrations of the bile salt encountered during the natural restoration of microbiome may allow the cells to transition toward biofilm formation. This would protect C. difficile vegetative cells against concentrations of antibacterial compounds that may normally inhibit growth (i.e., antibiotics or other bile acids). 52 In addition to promoting germination and inducing biofilm formation, bile acids can alter C. difficile physiology in other ways. Low amounts of LCA (0.08 mM) resulted in more elongated vegetative cells and absence of flagella. Fewer flagella were also observed when vegetative cells were incubated in 0.8 mM DCA and 0.3 mM CDCA. The authors hypothesized that lack of flagella may allow for better adherence to intestinal lining during infection. Additionally, bile acids resulted in up-regulation of the chaperon proteins DnaK, DnaJ, GrpE, GroL, and GroS, but the mechanism of up-regulation for chaperones is still unknown. 53 Finally, bile acids inhibit TcdB toxin activity by binding to the toxin. Both conjugated and deconjugated secondary bile acids (e.g., DCA, LCA, GDCA, TDCA, GLCA, and TLCA) have greater potency in inhibiting toxin activity than their primary bile acid counterparts, but toxin binding by bile acids is reversible. The reversible binding and inhibition of bile acids to C. difficile toxins may thus suggest modulation mechanisms taking place in the gut depending on bile acid composition in the host. 54 The role of bile acids in relation to C. difficile physiology and pathogenesis has been studied and characterized, extensively, since the discovery of their function as germinants for C. difficile spores. 16,46 Because of the many functions bile acids play in the host, it is not surprising that the interaction with C. difficile is complex and multifaceted, with both positive and negative interactions observed -depending on the conformation of, the abundance of, and the location of specific bile acids as well as the C. difficile life cycle stage in the host. More work should be done that move the field from data that correlate what bile acids are present in the host to potential causative interactions between bile acids and the host/C. difficile.

Stickland metabolites, amino acids, and the Wood Ljundahl pathway in C. difficile pathogenesis
Stickland metabolism was first identified in 1934 as the predominant pathway for Clostridium sporogenes energy production. 55 Stickland metabolism couples pairs of amino acids that may act as electron donors or acceptors. 56,57 The donor amino acid is oxidatively deaminated or decarboxylated to produce NADH, whereas the electron accepting amino acid is reduced to regenerate NAD + (Figure 2(a)). Although a myriad of amino acids can be used in the oxidative branch, the reductive branch is fueled only by proline or glycine. In the reductive branch, proline is consumed by proline reductase (PrdB, PrdA), a selenium-containing enzyme (selenoenzyme) that generates 5-aminovalerate and NAD + , and glycine is consumed by glycine reductase (GrdA, GrdB), selenoproteins that generate acetate and NAD + (Figure 2(b,  c)). 58,59 Although Stickland metabolism is wellcharacterized in Clostridium sticklandii, 60 this alternative metabolic pathway in other Clostridial species has just recently been elucidated. Bouillaut et al. 59 analyzed the C. difficile glycine and proline reduction operons and their activity in the reductive pathway of Stickland metabolism. A mutation in the prdB subunit of the proline reductase enzyme resulted in a decrease in growth in rich media, but a mutation in a grdA mutant did not affect growth in rich medium. 59 Additionally, the authors characterize the sigma-54 dependent activator, PrdR that acts as a mediator for PrdB-dependent activation and proline-dependent toxin repression. 59 In addition, given the role the Stickland reductive branch has in regenerating NAD + for the cell, the redox dependent transcriptional repressor, Rex, has a role in prolinedependent regulation and is controlled by PrdR in C. difficile. 61 Using DNA binding assays and qRT-PCR, the authors found that proline seems to be the preferred amino acid for regeneration of NAD + . In presence of excess proline, PrdR stimulates proline reductase expression and simultaneously Rex represses the glycine reductase gene grdE. NADH is then oxidized and, as a result, the ratio of NADH/ NAD + is low. Alternatively, when levels of proline are low, the NADH/NAD + ratio increases and the high levels of NADH prevents Rex from repressing glycine reductase expression (Figure 2(b,c)). 61 Selenium is an essential component of the proline and glycine reductases. Selenium is incorporated into these proteins as selenocysteine. Selenocysteine is generated through a synthesis pathway where the product of the selD gene reacts inorganic phosphate with hydrogen selenide to generate selenophosphate. 58,62,63 A selD mutant is unable to incorporate selenium into proteins and this results in a complete loss of the reductive branch of Stickland metabolism. The loss of selenophosphate generation results in a defect in the ability of C. difficile spores to outgrow following germination in peptide-rich media. 63,64 Also, the absence of selenophosphate altered C. difficile physiology so that other NAD + regeneration pathways were expressed. 64 Several studies have correlated the depletion or increased levels of amino acids important for Stickland metabolism using both in vitro and in vivo approaches. 48,[65][66][67][68] In an in vitro, multiomics approach, the metabolome of multiple C. difficile growth stages found that Stickland metabolites were dramatically depleted upon entry into stationary phase 69 . Furthermore, in stool samples derived from hospitalized patients, amino acids used for Stickland metabolism were depleted in CDI patients, suggesting that these amino acids were consumed by C. difficile vegetative cells. In this study, branched chain amino acids, such as leucine, were hypothesized to be the preferred When excess proline is available in the surrounding environment and the NADH/NAD + ratio in the cell is low, the PrdR activator promotes proline reductase (PR) expression to generate 5-aminovalerate and NAD + . Concurrently, Rex inhibits glycine reductase (GR). C) When proline levels are low and the NADH/NAD + ratio is high, Rex is unable to inhibit glycine reductase allowing acetate formation. Proline reductase activity is not present. Created with BioRender.com. amino acid used by C. difficile; other amino acids, such as proline, tyrosine, and phenylalanine are possible sources of energy for both C. difficile and other gut microbes. 67 Proline availability is important for C. difficile colonization in mice. A mutation in proline reductase (prdB) resulted in decreased colonization in a humanized microbiome mouse model. 65 Moreover, C. difficile can take advantage of inflammation-induced collagen degradation. During this toxin-dependent process, the host cells respond to this inflammation by producing matrix metalloproteases. This results in the degradation of collagena protein rich in proline. 70 Building upon this, in mice monoassociated with C. scindens, C. hiranonis, or C. leptum, proline was greatly depleted and mice were protected against CDI. 48 This work is also supported by in vitro conditions where C. difficile and C. hiranonis compete for nutrients. 71 Interestingly, Battagliogli et al. 65 observed high levels of the amino acids glycine, proline, threonine, and alanine in a dysbiotic colon, whereas Aguirre et. al. 48 observed depletion of proline and glycine in their monoassociated mice (amino acids essential for the reductive pathway of Stickland metabolism). 48,65 Nevertheless, direct comparison should not be made since one study used a humanized microbiome mouse model and the other used germ-free mice that were monoassociated with individual bacteria.
Expanding on the hypothesis that depletion of amino acids is important for Stickland metabolism, and may drive protection against CDI, Girinathan et al. 66 found that the gut commensal bacterium P. bifermentans protects against CDI, at least in part, through depleting nutrient sources used by C. difficile. Using P. bifermentans monoassociated mice, as well as carbon-source enrichment analysis of the gut-metabolomic environment, the authors found alterations in the substrates/products of Stickland metabolism (such as proline, glycine, threonine and 4-hydroxyproline) and production of 5-aminovalerate. 66 Although in vivo FMT studies have reported levels of amino acids in FMT-recipients, such as increasing levels of valine, isoleucine, and leucine (amino acids able to be used in the oxidative branch), the presence of Stickland products in these recipients have not been reported [besides the SCFA (acetate, propionate)]. [72][73][74] In a chemostat model, levels of 5-aminovalerate decreased post-FMT and isobutyrate levels remain constant post-FMT. A plausible explanation for the observed 5-aminovalerate levels could be the lack of Stickland fermenters included in the stable communities introduced into the chemostat, such as P. bifermentans. 75 Expressed at lower levels when other forms of energy production are present, the Wood-Ljungdahl Pathway (WLP) may need examining during CDI. Through the WLP, two molecules of CO 2 are reduced to acetate. 76 The WLP can be coupled with butyrate production to allow for increased efficiency of ATP formation by decreasing nutritional requirements. Thus, the WLP may prove useful when glucose or amino acid levels are low, and C. difficile may need to adapt to low nutrient conditions. 77 Moreover, in a mouse model of CDI, the WLP increased expression twenty-four hours post-infection. 78 This suggests more attention should be given to this pathway in animal models, or in patient cohort samples.
Stickland metabolism by C. difficile during host colonization is well-established. 48,65,68 Nevertheless, the availability of the substrates needed for the oxidative and reductive branches of Stickland metabolism, and required compounds for WLP, during CDI likely shapes the ability of the bacterium to cause disease. Similarly, the presence of Stickland-using microbial competitors in the host microbiome also shapes how C. difficile colonizes a host. 48,66,79 Therefore, a better understanding of specific conditions in which C. difficile takes advantage of alternative metabolic pathways is necessary for the study of novel therapeutics that may modulate these pathways.

Non bile acid metabolites that correlate or show protection against C. difficile pathogenesis
The gut microbiome is complex, with interactions occurring between resident and invading bacteria that result in the generation of secondary metabolites. One such metabolite is coprostanol. Coprostanol is generated through the reduction of the double bond between C5 and C6 of cholesterol. 80 In the metabolome of healthy controls vs. CDI patients, 63 bacterial OTUs were identified that positively correlated with the presence of coprostanol, which in turn negatively correlated with CDI patients. 81 The majority of phylotypes that correlated with coprostanol presence were members of the Lachnospiraceae and Ruminococcaceae families. Coprostanol may enhance resistance to CDI by decreasing the availability of cholesterol which could reduce the abundance of metabolites (i.e., primary bile acids) that are necessary for germination by C. difficile spores (Figure 3(a,b). 81 Three microbial-derived SCFAs (propionate, acetate, and butyrate) have long been associated with CDI resistance. [82][83][84][85] In general, SCFAs are believed to play important roles in maintaining gut homeostasis. 86 Several studies have shown that SCFAs are depleted during CDI 5,81,86 Recently, butyrate, and acetate were found to protect against CDI by aiding in maintaining intestinal barrier integrity. 87,88 Butyrate plays a role in maintaining intestinal barrier integrity through limiting the permeability of intestinal epithelial cells to C. difficile toxins by stabilizing the hypoxia-inducible factor 1 alpha (HIF-1α). 87 Acetate acts during the early stages of C. difficile infection by activating the free fatty acid receptor 2 (FFAR2) signaling pathway by augmentation of IL-1B production from neutrophils and IL-1 R through signaling of ILC3. These two types of immune cells then induce production of IL-22 that is implicated in antimicrobial and repair mechanisms in intestinal epithelial cells. 88 The mechanism by which propionate confers protection against CDI has yet to be elucidated. However, taking from data published on butyrate and acetate, a hypothesis could be that propionate acts upon specific host immune factors (Figure 3a). Interestingly, Gregory et al. 89 suggested that increasing levels of SFCAs as the gut microbiome recovers from disruption, may be used by C. difficile vegetative cells as a triggering signal to upregulate Figure 3. Gut metabolites effect in C. difficile growth and colonization. A) Examples of mechanisms of inhibition or protection against CDI. Aa) Gut microbes produce SCFA that aid in intestinal barrier function. Ab) Bacteria that perform 7α-dehydroxylation generate secondary bile acids that correlate with CDI resistant states. On the other hand Ac) generation of p-cresol by C. difficile may affect barrier function in Gram-negative bacteria, Ad) release of sialic acid from the colon mucus layer is used as a nutrient source by C. difficile. B) In an undisrupted gut environment amino acid, SCFA and secondary bile acid levels are high, creating a resistant state against C. difficile through immune defenses and nutrient limitation. High levels of the cholesterol derivative, coprostanol, are observed indicative of limited availability of primary bile acids. Introduction of antibiotics compromises the gut environment, and C. difficile toxins disrupt the intestinal barrier. C. difficile also promotes increasing levels of p-cresol, conjugated primary bile acids (e.g. TA) and sorbitol as well as use sialic acid as a nutrient source. FMT treatment allows restauration of gut metabolites and restore a resistant state against CDI. Created with BioRender.com. toxin secretion and promote the inflammation that would allow the bacterium to maintain colonization in the gut. 89 A recent study that used 1 H-NMR on stool samples derived from recurring CDI patients that received FMT, through capsule or colonoscopy, found increased levels of acetate, butyrate, and propionate in recipients 12-weeks post-FMT. 85 Interestingly, valerate, another SCFA was depleted in a chemostat C. difficile infection model. 75 Similar to the other SCFA, valerate levels were restored upon FMT. The study also found that valerate inhibited in vitro C. difficile vegetative growth in a dose-dependent manner, and that introducing valerate orally in the form of 15 mM glycerol trivalerate into a C. difficile mouse model decreased total viable counts. 75,85 It is important to note, though, that the glycerol trivalerate given to the mice was above physiologically relevant concentration since the authors found that chemostat FMT cultures only reached 4 mM valerate.
In addition to SCFAs, the microbiome produces antimicrobial compounds. Two Clostridial species, C. scindens and C. sordellii secrete tryptophanderived antibiotic compounds (i.e., 1-acetyl-βcarboline and turbomycin A) that inhibit C. difficile growth. 90 Interestingly, the anti-C. difficile effect of the antibiotics was enhanced by DCA and LCA. These findings suggest that a combination of secondary bile acids as well as antibiotics is at play during prevention of disease. 90 However, a recent study with mice that were monoassociated with C. scindens did not detect 1-acetyl-β-carboline in their metabolomics analysis, 48 suggesting that the activity of 1-acetyl-βcarboline may require specific conditions to be produced.
Dynamic interactions between gut microbes are constantly taking place and involve the production of and the sensing of secondary metabolites. 91 Although the mechanisms by which some of these metabolites protect against CDI are still unclear, 81,87,88,90 their identification provides important data that can be directly tested as potential treatment options for CDI. Moreover, a microbiome-centric approach for CDI treatment will most likely find other metabolites that show protection against CDI and could be explored for therapeutic approaches.

Metabolites and their beneficial roles for C. difficile infection
In recent years, the research focus on the modulation of the microbiome by C. difficile and how the bacterium survives in the host has increased. 92-96 C. difficile can produce products that promote inflammation and that have antibacterial effects. C. difficile generates p-cresol, a phenolic compound that affects the integrity of surface barriers in bacterial cells with a higher effect observed in Gram-negative bacteria. 93 The hpdBCA operon is responsible for the fermentation of tyrosine to p-hydroxyphenylacetate to then generate p-cresol using the 4-hydroxyphenylacetate dehydrogenase enzyme. In a mouse model of recurrent CDI, mice infected with a hpdC mutant strain, had increased microbial diversity compared to mice infected with the wildtype strain, as well as lower C. difficile viable counts. Additionally, when exogenous p-cresol was introduced to healthy human fecal slurries, the number of viable total anaerobes increased, thus suggesting that generation of p-cresol by C. difficile may give the bacterium a competitive advantage over other gut microbial species (Figure 3  (a,c)). 93 Similar to the production of p-cresol, production of sorbitol by C. difficile may enhance colonization. A C. difficile strain that is unable to produce sorbitol is outcompeted 10-fold by its wildtype counterpart in a mouse model. 94 Upon further investigation, the authors found that the inflammation induced by the C. difficile toxins results in an upregulation of aldose reductase which generates host-derived sorbitol. These results show how C. difficile is able to use a diet and host-derived nutrient to expand in the perturbed microbiome environment (Figure 3(b)). 94 C. difficile also can use metabolites produced by the host and other gut microbes during colonization. 92,95,96 The gut symbiont Bacteroides thetatiotaomicron encodes a sialidase enzyme that cleaves and releases sialic acid from mucosal glycoconjugates. C. difficile encodes a sialic acid catabolic operon. Using a transcriptional analysis of germ free and B. thetatiotaomicron monoassociated mice, the data showed that the ability of B. thetatiotaomicron to release sialic acid resulted in increased expression levels of the C. difficile sialic acid operon. 95 Additionally, a spike in free sialic acid 1 day after antibiotic treatment of mice was observed but these levels reduced 3 days after antibiotic treatment. 95 Moreover, mice monoassociated with a B. thetatiotaomicron sialidase mutant strain and then infected with C. difficile had lower C. difficile CFU counts (Figure 3(a-d)). 95 These results suggest an important role of sialic acid during in vivo infection.
Another host-produced metabolite, heme, can be used by C. difficile. 92 The heme-sensing membrane protein system, HsmRA, protects against redox damage generated from antibiotic treatment of CDI. C. difficile HsmA binds to heme that is released from the inflamed GI tract and shields the bacterium from redox-active molecules. Although the specific mechanism of protection by HsmA is still unknown, the authors hypothesized the use of HsmRA by other pathogens for protection against oxidative stress may be taking place. 92 Finally, indole may play a role during C. difficile infections. Supernatant from C. difficile stationary phase cultures can induce the expression of tryptophanase (tnaA) in E. coli. The levels of indole increased in other indole-producing microbes in the gut (e.g., L. reuteri and E. faecalis) in co-culturing assays with C. difficile. Interestingly, the MIC (5 mM) of C. difficile strains were found to be higher than the MIC of multiple gastrointestinal bacteria tested, ranging from 2-4 mM. These results led the authors to the hypothesis that the ability of C. difficile to resist higher gut indole concentrations may provide the bacterium a competitive advantage. 96 Importantly, these experiments were performed under in vitro conditions and future work with animal models is needed to test the proposed hypothesis.
Successful C. difficile colonization, occurs when the host gut environment is disrupted. As a result, the bacterium must adapt and take advantage of resources available to maintain colonization and exclude the reestablishment of competing microbes. 97,98 The ability of the C. difficile bacterium to establish a niche in the gut environment by use of metabolites from intestinal epithelial cells, as well as secreted compounds to outcompete other gut microbes, suggests C. difficile is a bacterial generalist. 68 Although the factors that correlate with an environment that excludes C. difficile are well documented, the factors that C. difficile uses to exclude the microbiome from reestablishing itself are comparatively less understood. By inhibiting these mechanism, the normally-protective microbiome may gain a foothold and return the GI to a colonization resistant environment.

Conclusions and future directions
This review highlights specific metabolic compounds that have been regarded as important during the life cycle and pathogenesis pathway of CDI. As the research field, moves toward a microbiome-centric study of gut diseases, molecules used as nutritional sources for both pathogen and resident gut bacteria are being identified. In Figure 3, we show graphical representations of microbiome stages that may allow or prevent the gut environment to succumb to a C. difficile disease stage, as well as promote recovery.
C. difficile pathogenesis is complex. Although much work has focused on the effect that bile acids and antimicrobials have on the bacterium, considering the physiology of C. difficile for developing new treatments is emerging in the field. Potential treatment options against CDI have slowly progressed toward approaches that focus on restoring the gut microbiome ecosystem as a whole and not what affects C. difficile. Although some progress has been made, there are still many aspects to uncover regarding gut microbiota niche growth and competition in the GI tract. The move from broad-spectrum antibiotic treatment to more specific antibiotics, like ridinilazole that showed fewer CDI recurrences in a phase 2 trial, is happening at present. 99 Targeted molecules such as nanobodies or DARPins have also emerged and, 100 will probably become the norm for treatment against CDI in the future. Alternatively, specific strategies to reintroduce defined microbial communities to combat CDI, in the form of targeted microbiome therapies (e.g., SER-109, composed of purified Firmicutes bacterial spores), 101 might also work as an effective treatment strategy. Unraveling of additional bile salt biotransformations that can be taken advantage of to potentially modulate bile acid composition as well as bile acid analogs that do not undergo enterohepatic recirculation are yet other potential CDI treatments. 45,102 Finally, introduction of probiotics specifically targeting CDI or synbiotics (probiotics combined with prebiotics) that may modify the gut microbiome by increasing engraftment through introduction of specific nutrient sources appear to have potential for treatment in the future. 103,104 It is important to mention though, that as understanding of the complex CDI pathogenesis and life cycle expands, the realization that individual treatments options may not be the way forward and comprehensive treatment plans may provide the best results, especially in regards to relapsing C. difficile episodes.

Disclosure statement
The authors report there are no competing interests to declare.

Data Availability Statement
Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Funding
This project was supported by awards R01AI116895 and R21AI144454 from the National Institute of Allergy and Infectious Diseases. Andrea Martinez Aguirre is a recipient of a CONACYT-COECYT fellowship 2017-2022 scholar/ scholarship 625561/472087. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIAID. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.