The SCFAs are butyrate, acetate, and propionate. In the human colon, acetate is the most abundant SCFA, averaging 54 mmol, while propionate and butyrate represent approximately one third of acetate concentration, thus averaging 18 mmol [14, 15]. Although the main substrates for SCFA production by the intestinal microbiota are dietary fibers and resistant starch [16], several amino acids may contribute as substrates for such production by the intestinal bacteria [17] (Figure 1). These amino acids are released by the bacterial proteases and peptidases from both alimentary and endogenous proteins that have not been fully digested in the small intestine [18]. Of note, in a rodent model, the consumption of a diet containing high amounts of proteins, results in a marked increase of SCFAs in the colonic luminal fluid [19], reinforcing the view that amino acids are significant contributors for microbial SCFA production.

These compounds are transported from the luminal fluid to the intracellular medium of colonocytes by several processes. Specific transporters, notably monocarboxylate transporter 1 (MCT1) which are members of the proton-coupled monocarboxylate electroneutral transporters, and sodium-coupled MCT1 (SMCT1) which are members of sodium-linked electrogenic transporters, represent the predominant transporters for monocarboxylates (SCFAs and lactate) from the luminal fluid to the intestinal epithelial cells [20]. However, butyrate can also enter within the colonocytes, although to a lesser extent, by using the SCFA/HCO₃^– exchanger [21]. In addition, a minor amount of SCFAs in undissociated form can enter colonocytes by diffusion [22]. Butyrate, acetate, and propionate, after their entry inside colonocytes are efficiently oxidized in the mitochondria of colonocytes, thus serving as fuels for ATP synthesis [23]. This explains the micromolar concentrations of these compounds in the portal blood as compared to millimolar concentrations in the colonic luminal fluid [24] (Figure 2). Importantly, the oxidation of butyrate within colonocytes has been proposed as a way to regulate its intracellular concentration, and thus presumably its effect on the level of histone acetylation and consequently on nuclear gene expression [25, 26] (Figure 2). In fact, butyrate has been shown to dose-dependently inhibit histone deacetylase activity in models of epithelial cells originating from human colonic adeno-carcinoma. Such inhibition increases histone acetylation, resulting in alteration of cell cycle regulatory protein expression, and finally reducing markedly cell growth [27]. Of note, among SCFAs, propionate represents also an efficient inhibitor of histone deacetylase activity [28]. In contrast, acetate shows no effects on histone deacetylase activity, and incidentally exerts no effect of colonic epithelial cell growth [26, 29].

The concept that butyrate metabolism within colonocytes regulates its intracellular concentration is indirectly supported by the pioneering work of Roediger [30]. In this work, it was observed that although well oxidized in the mitochondria of colonocytes, butyrate very moderately increases oxygen consumption in these cells. This can be explained by the fact that butyrate suppresses the oxidation of endogenous energy substrates in colonocytes. In other words, when butyrate is available to colonocytes, these cells use preferentially butyrate for ATP synthesis, and thus use less endogenous substrates to maintain energy homeostasis, allowing regulation of its concentration within colonocytes. Of note, and in a similar line of thinking, it has been shown that oxidation of butyrate by the differentiated absorptive colonocytes in the surface epithelium and the upper part of the colonic crypts regulates the concentration of butyrate in vicinity of healthy proliferating epithelial stem/progenitor cells in the lower part of the crypts, thus protecting them from the antiproliferative effects of butyrate at excessive concentrations [31].

Regarding lactate which is generally produced by the intestinal microbiota from amino acids and undigestible saccharide [19, 32, 33], this organic acid is found in the human feces at much variable concentrations among individuals ranging from undetectable amount up to millimolar concentrations [34]. In human colon, lactate concentration averages approximately 3.5 mmol [35]. The absorptive colonocytes have been shown to be characterized by high capacity for L-lactate oxidation [36] (Figure 1).

More recently, H₂S produced by the intestinal microbiota has been shown to represent a mineral fuel for ATP production in colonocytes. H₂S is produced by different bacterial species from different alimentary and endogenous compounds including cysteine, taurine, and sulfite, and by sulfate-reducing bacteria [37–39] (Figure 1). The H₂S-producing bacterial species include Fusobacterium, Clostridium, Escherichia coli, Salmonella, Klebsiella, Streptococcus, Desulfovibrio, and Enterobacter [1]. H₂S which is not bound to luminal components, thus either as a gas pocket or as a gas dissolved in the liquid phase of the colonic luminal fluid, represents the chemical form that diffuses through the mucus layer, and then into the colonocytes [1]. Several dietary compounds like zinc and proanthocyanidins (contained in plants) bind H₂S, thus reducing its concentration in free form [40, 41]. Interestingly, in germ-free mice, H₂S concentrations in its free form are reduced in plasma and intestinal tissues when compared with normal animals equipped with an intestinal microbiota [42]. These data suggest that the microbiota is likely a purveyor of this compound for the host. The concentration of total H₂S, in its free and bound forms, measured in the mammalian colonic luminal fluid is rather divergent, according notably to the specificity of techniques used for measurement, thus ranging from micromolar to low millimolar concentrations [43]. In isolated colonocytes, micromolar concentrations of H₂S (5–65 µmol) increase instantaneously oxygen consumption by entering the mitochondrial respiratory chain [44]. This increased consumption is concomitant with an inner mitochondrial energization and synthesis of ATP [41], thus establishing H₂S as the first mineral energy substrate in human cells [44]. This discovery has challenged the previous concept which considers that mammalian cells are mostly dependent on carbon-based fuels for ATP production. The mitochondrial ATP synthesis through H₂S oxidation in colonocytes is made possible by the catalytic activity of the sulfide oxidation unit which drives the conversion of H₂S into thiosulfate [45] (Figure 3).

Regarding the colonic enteroendocrine cells, a recent study has shown that these cells are equipped with the 3 enzymes constituting the sulfide-oxidizing unit, and that low amounts of the sulfide donor NaHS increase oxygen consumption in the enteroendocrine cell model NCI-h716, thus suggesting that H₂S is a mineral fuel allowing ATP synthesis in these cells [46].

It is of major interest to consider that in IBD, marked lower fecal SCFA concentrations compared to control have been repeatedly measured [77–82]. Although, as presented above, excessive butyrate (in relation to the metabolic capacity of differentiated colonocytes to oxidize them) may affect proliferating epithelial stem/progenitor cells [31], and thus presumably interfering with the normal process of epithelial renewal, it is conceivable that, conversely, insufficient butyrate availability in relationship with the colonocyte requirement may affect energy homeostasis in these cells (Figure 4). In addition to being luminal fuels for absorptive colonocytes, SCFAs are known to regulate electrolyte and water movement across the colonocytes. Indeed, butyrate, and to a minor extent acetate and propionate, have been shown to play a role in the regulation of electrolyte and water absorption [83, 84]. Then, it is tempting to postulate that a shortcoming of butyrate production by the intestinal microbiota, and thus a defect in the supply of one of the major luminal fuels, would diminish the capacity of the colonic epithelium for water absorption, thus participating notably in the IBD-associated diarrheal process. However, despite encouraging results obtained from experimental studies suggesting beneficial effects of butyrate on the diarrheal process [85, 86], the efficiency of butyrate for treating diarrhea in clinical studies remains controversial. Addition of resistant starch to oral rehydration solution has been reported to decrease duration of diarrhea in adults and children hospitalized for acute diarrhea [87, 88]. However, in a randomized, double-blind, placebo-controlled multicenter study, a mixture of non-digestible carbohydrates was found ineffective as an adjunct to oral rehydration solution for reducing the treatment of acute infectious diarrhea in children [89].

Several controlled studies have been performed to evaluate the effects of exogenously administered butyrate on several clinical parameters in patients with IBDs. In ulcerative colitis patients treated with butyrate enemas, the results obtained are much heterogeneous. One study found that clinical improvement was proportionally less frequent in the butyrate-treated group than in the placebo group [90]. Two studies found no therapeutic value of SCFA enemas because of no significant clinical improvements [91, 92], and four studies found some modest improvements of inflammatory parameters after enema administration with butyrate or other SCFAs [93–96]. These discrepancies are likely due to differences in the clinical status among volunteers included in each study, in the number of patients recruited, in the dose of SCFA administered, in the duration of treatment, and finally in the parameters measured. By recapitulating the data obtained in randomized controlled trials, it was concluded that the results obtained do not overall support the application of butyrate enemas for the treatment of ulcerative colitis [97, 98]. Regarding Crohn’s disease, there are no reliable data in support of butyrate enemas for the treatment of this disease [97]. These latter conclusions suggest that treatment with butyrate is not able by itself to reduce the inflammatory flare to any valuable extent in ulcerative colitis. These disappointing results are likely due, in part at least, to a reduction of butyrate uptake by the inflamed mucosa because of downregulation of the MCT1, thus limiting its action on the colonic mucosa [99]. Of note, significant alteration in the expression of the MCT1 has also been detected in experimental colitis model [100].

In deep contrast with the demonstration that luminal sulfide at low micromolecular concentrations serves as a mineral fuel in colonocytes, higher concentrations of H₂S (above 65 µmol) inhibit colonocyte oxygen consumption by inhibiting the mitochondrial cytochrome c oxidase activity (complex IV of the mitochondrial respiratory chain) after binding to this complex with high affinity [101, 102]. Thus, at concentration above the capacity of colonocytes to oxidize this bacterial metabolite, H₂S decreases the capacity of mitochondria within colonocytes to synthesize ATP (Figure 4). In addition, H₂S at concentration above 1 mmol dose-dependently inhibits the oxidation of glutamine and butyrate, thus depriving colonocytes from the utilization of their two main fuels [102, 103]. Of note, the oxidation of H₂S in colonocytes takes precedence over the oxidation of other carbon-based energy substrates [104]. This fact strongly suggests that the oxidation of H₂S by the sulfide-oxidizing unit, likely because of its deleterious effect on mitochondrial ATP synthesis at excessive concentrations, represents a priority. The demonstration that colonocytes appear among the most efficient cells for H₂S disposal is not surprising given that these cells are facing the highest sulfide concentrations in the body [105]. Lastly, when the colonocytes undergo either spontaneous or butyrate-induced differentiation, their capacity to metabolize H₂S is significantly increased [41], raising the view that differentiated colonocytes on the epithelial surface and on the upper part of the colonic crypts may, like in the situation observed with butyrate [31], protect the proliferating epithelial stem/progenitor cells situated at the bottom of crypts from the deleterious effect of H₂S.

As presented above, H₂S, when present in excess, severely inhibits oxygen consumption in colonocytes, thus affecting ATP production in these cells. Consistent results have clearly shown that at concentrations where H₂S inhibits oxygen consumption in colonic epithelial cells, a marked decrease in cell proliferation has been measured. Indeed, the slow-releasing sulfide donor GYY4137, when used at 3 mmol concentration, inhibits markedly the proliferation of HCT116 cancer-derived colonic cells [106]. In the HT-29 adeno-carcinoma colonic epithelial cell model, the rapid H₂S donor NaHS at 0.3 mmol concentration inhibits cell respiration, an effect that coincides with a marked reduction of cell proliferative capacity [102, 107]. Since the colonocytes maintain their ATP cellular content in the presence of NaHS, despite inhibition of mitochondrial oxygen consumption, it has been proposed that by reducing cell growth, and thus ATP-consuming anabolic metabolism, the cells maintain their ATP content and thus their viability under severe restriction of mitochondrial ATP production [102].

The proposition that excessive H₂S concentrations may participate in the pathogenesis of IBDs was proposed more than two decades ago [108]. To further test this hypothesis, a study on patients with new-onset pediatric Crohn’s disease demonstrated that the intestinal microbiota composition of these patients was characterized by higher relative abundance of Atopobium, Fusobacterium, Veillonella, Prevotella, Streptococcus, and Leptotrichia [109]. Several members of those genera are known to be H₂S producers. In this cohort of children, the abundance of H₂S producers from cysteine was correlated with the severity of mucosal inflammation. To identify possible causal links between H₂S production and intestinal inflammation, the authors colonized mice invalidated for the production of the regulatory cytokine interleukin-10 (IL-10) (model of mice susceptible to colitis), with the H₂S producer Atopobium parvulum, and measured a worsening of colonic inflammation [109]. Since H₂S is obviously not the only bacterial metabolite produced by Atopobium parvulum, they used the H₂S scavenger bismuth and found that this latter compound was able to attenuate the worsening of inflammation provoked by Atopobium parvulum. Of major interest, the authors of this latter study found that the colonic mucosal biopsies obtained from the young Crohn’s disease patients displayed decreased expression of the mitochondrial enzymes involved in sulfide detoxification. Overall, the results of this study indicate that diminished capacity for H₂S disposal in the colonic mucosa of Crohn’s disease patients amplifies the pro-inflammatory effect of H₂S overproduction by the intestinal mucosa. Three studies reinforce this latter proposition. In the first one, it has been shown that gene expression and activity of one of the enzymes involved in H₂S detoxification, namely thiosulfur transferase, is downregulated in the mucosa of ulcerative patients [110], while in the second and third studies, decreased capacity for H₂S detoxification was measured in the mucosa of patients with respectively Crohn’s disease [111] and ulcerative colitis [112].

Results from in vivo studies in rats indicate that intra-colonic instillation with phosphate-buffered saline solution containing NaHS at 1.5 mmol concentration increased in colonocytes expression of the gene coding for the hypoxia inducible factor 1 α (Hif-1α) and of the mucosal inflammation-related genes, including inducible nitric oxide synthase (iNOS) and IL-6 [101]. Also, a study has shown that high concentrations of H₂S destabilize the protective mucous layer that covers the intestinal epithelium by reducing disulfide bounds linking the mucin-2 network. This latter process would possibly increase the interactions between bacteria and the epithelium [113]. Then, H₂S in excess may favour mucosal inflammation by inhibiting mitochondrial ATP synthesis, thus affecting proliferation of undifferentiated colonic epithelial cells, and in addition by altering the structure of mucus that protects the colonic epithelium. Interestingly, the study by Jowett and collaborators [114] indirectly suggests a pro-inflammatory effect of excessive H₂S production by the intestinal microbiota. In this study, following for one year patients with ulcerative colitis in remission, it was found that patients with the highest level of consumption of dietary proteins display a threefold increase in the risk of relapse when compared with patients with the lowest protein intake. In addition, high sulfate dietary intake was also associated with a higher risk of relapse in volunteers. Although H₂S concentration in the intestinal fluid was not measured in this study, the fact that high dietary protein consumption is associated with increased fecal H₂S concentration [115], and the fact that sulfate is a precursor for H₂S synthesis by the intestinal microbiota [116], raises the credible possibility that increased H₂S would play a role in the increased risk of relapse in patients in remission. A study in a mouse model deficient in the regulatory cytokine IL-10 shows that a diet with high content in saturated fat increases, in a process modulated by taurine-containing bile acids, the release of H₂S produced by the intestinal microbiota, such event appearing to play a major role in the intestinal inflammation observed in this experimental model [117]. Recently, in a Crohn’s disease case-control study, Marcelino and collaborators [118] identify a lack of bacterial species with the capacity to consume H₂S as the main distinguishing characteristics of the intestinal microbiota in the context of this inflammatory disease. These results reinforce further the view that failure of the intestinal microbiota to metabolize H₂S may contribute to the excessive luminal content of this bacterial metabolite, and thus accordingly to the etiology of Crohn’s disease in predisposed individuals.

However, from experiments performed in animal models in which colitis is induced chemically, and in studies where pharmacological inhibitors of H₂S synthesis or H₂S-releasing compounds are used, it appears that a low minimal amount of H₂S produced endogenously or supplied from the colonic luminal fluid is likely necessary to limit the extent of colonic mucosal inflammation [119–122], thus suggesting that an almost complete depletion of H₂S supplied to colonocytes is counter-productive in situation of intestinal inflammation [123]. In addition to the demonstration that, as explained above, low micromolar concentrations of H₂S are involved in mitochondrial ATP synthesis in colonocytes [124], H₂S have been shown to exert some antioxidant activity, notably by persulfidation of cysteine residues that protects these residues from oxidative damages [45].

Finally, in view of the effects of H₂S on the colonic mucosa, the concept that this compound is a double-edged sword for the intestinal epithelium has been proposed in 2019 [43]. In that concept, the effects of the diffusible bacterial metabolite H₂S on the inflammatory processes in the colonic mucosa depend on both the low endogenous production by the host colonic mucosa, that is protective, and on the exogenous bacterial production, which is deleterious when produced in excess.

The phenolic compound p-cresol (4-methylphenol) is produced from the amino acid tyrosine by the anaerobic bacteria of the large intestine [125]. The p-cresol concentrations measured in the human colonic fluid are in the low millimolar range [126, 127]. The bacterial metabolite p-cresol can markedly affect colonocyte energy metabolism when present in excess in the colonic fluid. Indeed, 0.8 mmol p-cresol diminishes mitochondrial oxygen consumption in human colonocytes when tested in vitro, and increases anion superoxide production, thus indicating alterations of the mitochondrial functions in these cells [128]. This effect is associated with a reduction of the capacity of colonic epithelial cells to proliferate. At 0.8 mmol, p-cresol increases after 1 and 2 days of treatment the proportion of cells in the S-cell cycle phase, and decreases the proportion of cells in the G0/G1 phase. However, the relative number of cells in the subdiploid G1 peak, that is characteristics of cells undergoing nuclear fragmentation, was not modified by the treatment with 0.8 mmol p-cresol [128]. In this latter study, pretreatment of colonocytes for three days with increasing concentrations of p-cresol resulted in dose-dependent decrease of the intracellular ATP content, thus indicating energy deficiency in these cells (Figure 4). Using monolayers of human colonocytes, p-cresol at millimolar concentrations dose-dependently increases the paracellular transport between colonocytes [129], suggesting that this bacterial metabolite at excessive concentrations can alter intestinal barrier function. The capacity of p-cresol in excess to increase anion superoxide production within colonocytes is of interest since reactive oxygen species have been shown to be produced in experimental situations of inhibition of mitochondrial complexes I, II, and III [130–132]. Excessive production of reactive oxygen species (and notably anion superoxide), in regards to the capacity of colonic epithelial cells to eliminate them, is considered as one parameter that is involved in the development of IBDs [133–136].

Indole and indole-related compounds are synthesized from various intestinal bacterial species [137]. The measurement of indole concentration in feces obtained from volunteers has shown huge inter-individual differences ranging from 0.30 mmol to 6.64 mmol [1]. Although these compounds have been shown in numerous experimental studies to exert protective effects against intestinal mucosal inflammation, notably through binding to aryl hydrocarbon receptor [138], indole at excessive concentration (2.5 mmol) affects the respiration of colonocytes by diminishing mitochondrial oxygen consumption [139], and thus mitochondrial ATP production. Such inhibitory effect was associated with a transient oxidative stress in colonocytes as well as an increased secretion of the pro-inflammatory IL-8 and tumor necrosis factor-α. Indoxyl sulfate, the main indole co-metabolite synthesized by host tissues, displayed similar effects as indole on colonocyte respiration and secretion of pro-inflammatory compounds when tested at 2.5 mmol.

Ammonia (considered as the sum of ammonium NH₄⁺ and NH₃) is present at millimolar concentrations mainly in the form of NH₄⁺ in the mammalian large intestinal fluids [140, 141]. Ammonia can be both utilized and produced by the intestinal microbiota. Ammonia provides one of the main sources of nitrogen for bacterial amino acid synthesis, and this compound is notably formed by intestinal microbiota from the conversion of urea to ammonia by the bacterial urease activities and by catabolism of several amino acids [142]. Ammonia used at concentration between 10 mmol and 50 mmol inhibits dose-dependently the mitochondrial oxygen consumption in colonocytes [143], thus affecting colonocyte respiration, and thus energy production, in these cells. In addition, ammonia at 10 mmol concentration inhibits SCFA oxidation in colonic epithelial cells [144] (Figure 4). Of note, at concentrations where ammonia inhibits mitochondrial oxygen consumption in colonocytes, this bacterial metabolite inhibits almost totally the capacity of the human colonic HT-29 cells to proliferate [145]. These latter results suggest that ammonia-dependent alteration of the capacity of colonocyte to produce ATP plays a role in the anti-proliferative effect of this compound.

However, colonocytes are equipped with enzymatic activities that allow to detoxify ammonia, up to a threshold concentration, in the mitochondria of colonocytes. This can be done by converting ammonia into citrulline in two metabolic steps [140], thus limiting the inhibitory effect of excessive ammonia concentration on colonocyte respiration. Interestingly, excessive ammonia concentrations in regards with the capacity of colonocytes to detoxify it, increases the paracellular permeability between colonocytes [146], and ammonia treatment of human colonocytes induces mitochondrial dysfunctions and impairs epithelial barrier integrity as well as the structure of the tight junction proteins [147].

HO-PAA is a bacterial metabolite produced by the intestinal microbiota from tyrosine [148]. This compound when used at concentration equal to 1 mmol decreases in human colonocytes the activity of the complex I of the mitochondrial respiratory chain and the mitochondrial oxygen consumption, thus presumably affecting ATP mitochondria production in these cells [149] (Figure 4). Such effect was paralleled by an increased production of reactive oxygen species in colonocytes.

The intestinal microbiota metabolizes ethanol, either originating from alcohol consumption, or, to a much minor extent, from amino acids, to acetaldehyde [150, 151]. In rodents, by increasing the number of proteins in the diet, the ethanol content in the colonic luminal fluid was increased several folds [19], suggesting capacity of the intestinal microbiota to produce ethanol during amino acid catabolism. Acetaldehyde concentration was increased in the colon of rodents after alcohol administration, and these concentrations were lower in germ-free animals when compared with animal carrying an intestinal microbiota [152], pointing out the role of microbiota for acetaldehyde production from ethanol. Acetaldehyde injected in the lumen of the pig colon is partly metabolized to acetate [153]. Thus, acetaldehyde concentration in the colon appears to depend both on the synthesis from ethanol and on the conversion of acetaldehyde to acetate by the colonic flora.

Acetaldehyde when used between 25 μmol and 100 μmol concentration has been shown to exert deleterious effects on human intestinal goblet cells, inducing cell necrosis. This effect coincided with decreased mitochondrial activity, decreased ATP cellular concentration, and increased production of reactive oxygen species [154]. These findings indicate that acetaldehyde may adversely affect one main component of the intestinal barrier function. In the human colonic mucosa, acetaldehyde in excess reduces the expression of several proteins known to be necessary for normal tight junction and adherent junction functions [155].