Alterations in Bile Acid Metabolism Associated With Inflammatory Bowel Disease

Abstract

Inflammatory bowel disease (IBD) is a chronic relapsing inflammatory disorder closely related to gut dysbiosis, which is associated with alterations in an important bacterial metabolite, bile acids (BAs). Although certain findings pertinent to BA changes in IBD vary among studies owing to the differences in sample type, quantitated BA species, study methodology, and patient characteristics, a specific trend concerning variations of BAs in IBD has been identified. In elaborating on this observation, it was noted that primary BAs and conjugated BAs are augmented in fecal samples but there is a reduction in secondary BAs in fecal samples. It is not entirely clear why patients with IBD manifest these changes and what role these changes play in the onset and development of IBD. Previous studies have shown that IBD-associated BA changes may be caused by alterations in BA absorption, synthesis, and bacterial modification. The complex relationship between bacteria and BAs may provide additional and deeper insight into host-gut microbiota interactions in the pathogenesis of IBD. The characteristic BA changes may generate profound effects in patients with IBD by shaping the gut microbiota community, affecting inflammatory processes, causing BA malabsorption associated with diarrhea, and even leading to intestinal dysplasia and cancer. Thus, therapeutic strategies correcting the alterations in the composition of BAs, including the elimination of excess BAs and the supplementation of deficient BAs, may prove promising in IBD.

Introduction

Inflammatory bowel disease (IBD) is a chronic relapsing inflammatory disorder that comprises 2 subtypes: Crohn’s disease (CD) and ulcerative colitis (UC).¹ Although its exact etiology and pathogenesis have not been completely clarified until recently, it is widely accepted that IBD is mediated by genetic, environmental, and immune factors.^{1, 2} Because of the immediate proximity to intestinal cells, the gut microbiota likely play an important role in the onset and development of IBD. Previous studies have confirmed that bacterial dysbiosis exists in patients with IBD.^3-5 Matsuoka and Kanai⁶ reviewed published studies and proposed that the most consistent observation regarding gut microflora in IBD is reduced bacterial diversity, a decrease of Firmicutes, and an increase of Proteobacteria. The metabolites produced by the human gut microbiome are bioactive molecules that generate profound effects in the host.⁷ Therefore, dysbiosis alters not only the composition of the intestinal microbiota but also its metabolites, which may produce a wide range of effects on the occurrence and development of certain diseases such as IBD. Changes in specific bacterial metabolites have been confirmed in patients with IBD, such as short-chain fatty acids, tryptophan, bile acids (BAs), and polysaccharide A.^8-10 Recent evidence from clinical studies and experimental animal models has shown that BAs act as signaling molecules to influence the host’s susceptibility to intestinal inflammation.^11-14 Therefore, alterations in their composition and metabolism may be associated with IBD.

Research has shown that BAs are synthesized from cholesterol, primarily in the hepatocyte, via 2 main biosynthetic pathways: the classical (neutral) pathway and the alternative (acidic) pathway.¹⁵ In humans, the relative contribution of the classical pathway to total BA production is at least 75%.¹⁶ Cholesterol is converted into BAs by pathways that involve 17 different enzymes, and the rate-limiting enzymes involved in the classical pathway and the alternative pathway are CYP7A1 and CYP27A1, respectively.¹⁷ The immediate products of these pathways are referred to as primary bile acids (PBAs), including cholic acid (CA) and chenodeoxycholic acid (CDCA). Subsequently, these BA molecules are conjugated to glycine or taurine to form conjugated bile acids (CBAs); thereafter, they are discharged into the gallbladder for storage and released into the intestine after meals to promote the digestion and absorption of lipids.

Most CBAs are actively absorbed in the terminal ileum, and their transport across the ileal epithelial apical membrane is mediated by the apical sodium-dependent bile acid transporter (ASBT).¹⁸ The structure of CBAs that escape absorption in the ileum is altered by bacterial enzymes in the distal intestine, and secondary bile acids (SBAs) are produced; bacteria deconjugate CBAs into unconjugated bile acids (UBAs), after which UBAs (CA and CDCA) are dehydroxylated to deoxycholic acid (DCA) and lithocholic acid (LCA), respectively.¹⁹ Furthermore, UBAs are passively absorbed in both the small intestine and the large intestine by diffusion, which is much less efficacious relative to active reabsorption.¹⁸ The BAs are reabsorbed in the intestine and subsequently transported to the liver via the portal venous blood; they are then efficiently taken up by hepatocytes and secreted again to complete the enterohepatic circulation process.²⁰ Depending on the effective enterohepatic circulation, the BA pool is cycled several times a day through the intestine and liver, and only a relatively small amount of BAs are synthesised de novo each day to replace those that escape absorption and are excreted in the feces.²¹

Specific BAs differentially activate 4 nuclear receptors, including the farnesoid X receptor (FXR), the pregnane X receptor, the constitutive androstane receptor, and the vitamin D receptor (VDR), along with a G-protein-coupled receptor; BAs execute their functions in the host by acting via these receptors.²² Generally, the ability of BAs to activate nuclear receptors in cells depends on 2 aspects: the structure of BAs and whether the cells express BA transporters. Studies have shown that activation of FXR by CBAs is only observed in cells that express a BA transporter.²³ The CBAs account for 98% of all BAs in human bile, and therefore CBAs and UBAs are likely to represent natural FXR ligands in tissues that express BA transporters and do not express BA transporters, respectively.²³ The discrepant influence of CBAs and UBAs on the activity of FXR likely applies to the other nuclear receptors.

In addition, different BAs inherently have discrepant activation capacities for the 4 nuclear receptors. BAs are the endogenous ligands for FXR with different potency in the following order: CDCA > DCA > LCA > CA. Different from FXR, BAs are not the main ligands for the other 3 nuclear receptors; LCA is the main BA ligand for the pregnane X receptor; VDR is activated by LCA, but not by other BAs; and the constitutive andtrostane receptor is shown to mainly bind CA and LCA.^{24, 25}

In recent years, in addition to their detergent-like role in the digestion and absorption of lipids, novel functions of BAs as gut hormones modulating physiological and pathological processes, including lipid metabolic homeostasis, mucosal immunity, inflammation, and tumorigenesis, have been elucidated.^{26, 27} For example, the critical regulatory gene in BA synthesis, Cyp7a1, is feedback-regulated by the BAs via the FXR/fibroblast growth factor 19/hepatic fibroblast growth factor receptor 4 signaling pathway in humans.²⁸ Both in vitro and in vivo models have shown that BAs play a crucial immunomodulatory role by regulating the number and function of immune cells.^{13, 22, 29} Moreover, BAs may play an anti-inflammatory role at physiological concentrations, but at high concentrations they play a proinflammatory or even carcinogenic role in vitro and in vivo.^29-31

Regarding the BA metabolic process in vivo, the normal intestinal tract and bacterial flora are essential for the physiological enterohepatic circulation and biotransformation of BAs. Certain intestinal diseases, such as IBD, may cause changes in the composition and metabolism of BAs. In turn, these alterations may promote the maintenance and progression of IBD through signaling pathways mediated by BA receptors and other mechanisms. With the development of metabonomics techniques and the understanding of the potential role of BAs in IBD, alterations of BAs in patients with IBD compared with the BA profile of healthy control patients (HCs), and the association between the changes in BAs and IBD, have attracted increased attention in recent years.^{14, 32}

To date, no study has summarized the consistent changes in BAs in patients with IBD worldwide. Given that BAs exert an important influence on IBD, a review exploring the altered characterization of BAs in patients with IBD is urgently needed. We conducted a PubMed search for English-language articles published up to August 2020, with no date restrictions. The MeSH term and free-text word combinations that we used were the following: “bile acid,” “Bile Acids and Salts,” “Inflammatory Bowel Diseases,” “IBD,” “Crohn’s disease,” “CD,” “ulcerative colitis,” and “UC.” The articles were selected on the basis of certain criteria: observational studies that focused on differences in BA profiles between patients with IBD without concomitant abnormal hepatobiliary function and HCs. We screened titles and abstracts from the databases according to the eligibility criteria. Subsequently, the included articles were subjected to whole-paper reading and the accompanying references were checked to identify additional potentially eligible articles. We summarized the changes of BA profiles in patients with IBD reported in these studies and worked to reveal the most consistent observations regarding the altered BA profile in these patients. We then analyzed the possible causes and effects of these changes and propose promising therapeutic strategies for IBD that involve regulating the composition of BAs.

ALTERATIONS OF BAs IN PATIENTS WITH IBD

In total, 35 studies met our criteria and are included in our review. To determine whether subgroup analysis could be further conducted, the demographics and critical clinical characteristics (eg, disease location, disease activity, and history of enterectomy) of all patients in these studies have been collected and are presented in Table 1. Supplementary Table 1 provides information on the disease duration, medications, and diet of the patients with IBD. According to the clinical characteristics (IBD subtype and history of enterectomy) of patients with IBD and the quantitative objects (BAs or BA pools) of studies, BA alterations are shown in Tables 2-6, respectively. To better exhibit and identify the most important alterations of BAs in patients with IBD, only BAs for which at least 2 studies reported the changes are displayed. In addition, more comprehensive BA changes are summarized and presented in Supplementary Tables 3-7. The methodology for sample handling and analysis in these studies is shown in Supplementary Table 2.

For patients with IBD, as shown in Table 2, CA,^33-37 CDCA,^{34, 36} and CBAs were increased^{14, 36} and SBAs^14,35,37 and DCA^35,37 were decreased in the feces when compared with the corresponding parameters in HCs. In addition, Duboc et al¹⁴ found that there was a reduction in SBAs in the serum, but only in patients with active IBD. However, a subgroup analysis could not be conducted to explore BA alterations in different disease states (active or in remission) for IBD because of the limited number of studies. The BA changes shown in this table may reveal an alternative trend in patients with IBD without considering specific subtypes. Furthermore, an increase in total BAs,³⁶ PBAs,³⁵ glycine-conjugated bile acids (GBAs),³⁶ taurine-conjugated bile acids (TBAs),³⁶ taurochenodeoxycholic acid (TCDCA),³³ UBAs,³⁶ 7-sulfo-CA,³³ 3-OH-sulfated BAs,¹⁴ sulfated CDCA, and 7-keto-DCA,³³ and a decrease in the LCA³⁷ of fecal samples, was found in patients with IBD (Supplementary Table 3). Moreover, Martin et al³⁸ identified metabolic signatures in urine samples from children with IBD in remission that were different from those seen in HCs: Ursocholic acid, ursodeoxycholic acid (UDCA), glycoursodeoxycholic acid (GUDCA), β-CA, chenodeoxycholic acid-3-β-D-glucuronide, and chenodeoxycholic acid 24-Acyl-β-D-glucuronide levels were elevated; conversely, glycolithocholic acid (GLCA) and glycohyodeoxycholic acid levels were decreased (Supplementary Table 3).

For patients with CD, as indicated in Table 3, an increase in total BAs,^{37, 39} CA,^{35, 37, 40} glycocholic acid (GCA),^35,41 glycochenodeoxycholic acid (GCDCA),⁴¹ taurocholic acid (TCA),^{35, 41} and TCDCA levels in fecal samples was identified³⁵; a decrease in DCA^{35, 37, 40} and LCA^{35, 37, 40, 42} levels in fecal samples was determined when compared with the levels seen in HCs. However, the CDCA level in feces differed in the direction of change in 2 studies,^{37, 42} as did the UDCA level.^35,37

In serum, plasma, or fasting duodenal bile, there have been few consistent changes because of limited studies. Wilson et al³² proposed that there was an increase in the GCA level, with a concomitant reduction in the LCA, GCDCA, and TCA levels, in the plasma of patients with CD. In addition, a lower level of TCDCA was found in fasting serum.⁴³ One study showed that the GCA level was elevated in postprandial serum from patients with ileal involvement of less than 30 cm, whereas it was decreased in both fasting serum and postprandial serum from patients with ileal involvement of more than 30 cm; this finding indicated that the serum BA level may be affected by disease extent.⁴⁴

Disease location may be another factor influencing the level of BAs in patients with CD. In fasting duodenal bile, a higher level of CA^45,46 and lower levels of DCA^45,46 and UDCA⁴⁵ were seen in colitis; in addition, a higher level of CDCA and a lower level of DCA have been shown in ileocolitis.⁴⁵ Alterations in other BAs or ratios related to BAs are presented in Supplementary Table 4, including an increase in taurodeoxycholic acid,³⁵ 3α-OH BAs,⁴⁷ and trihydroxy-6β-CA,⁴¹ along with a decrease in taurolithocholic acid (TLCA),⁴² in the feces; an increase in glycodeoxycholic acid and the ratio of CBAs/UBAs in the plasma³²; a decrease in the ratio of DCA/(DCA+CA) in fasting serum⁴⁸; and a decrease in TBAs in postprandial serum and an increase in the ratio of GBAs/TBAs in both fasting and postprandial serum.⁴³ However, because of limited studies, a subgroup analysis could not be conducted to investigate BA alterations in different disease locations for CD.

Finally, significant differences in the mean postprandial increase in serum BAs between patients with CD and HCs were reported in 2 studies. Heuman et al⁴⁷ found that the mean postprandial increase in conjugated CA and CDCA was diminished in patients with CD when compared with that in HCs, and the mean postprandial increase in conjugated CA was decreased in patients with CD with intestinal resection (ileal resection with partial colectomy) when compared with nonoperated patients with CD. Suchy and Balistreri⁴⁹ reported that the mean postprandial increase in total BAs and GCA was significantly affected by disease locations and disease states and only decreased in active CD patients with ileal involvement compared with HCs.

For patients with CD without a history of intestinal resection, an increase in the UDCA pool^{50, 51} and a decline in the total BA pool^{46, 50-52} and CDCA pool^50,52 was found when these levels were compared with those of HCs (Table 4). In addition, the CA pool in ileitis and ileocolitis and the SBAs pool in ileocolitis were reported to be reduced.⁵⁰ The ratio of (GBAs pool size)/(TBAs pool size) was increased in patients with CD (Supplementary Table 5).⁵¹ Rutgeerts et al 1982⁵⁰ reported that the total BA pool size was negatively correlated with disease activity. Different disease locations may produce different effects on the changes in the BA pool. However, owing to the limited number of studies comparing patients with different disease locations with the HCs, no further subgroup analysis could be carried out.

For patients with CD with a history of intestinal resection, total BAs, CA, and CDCA were increased^{40, 53}; DCA and LCA were decreased in the feces when compared with levels in HCs.⁴⁰ In addition, an increase in CDCA⁴⁴ and a decrease in DCA^{45, 54, 55} and LCA⁵⁴ was noted in fasting duodenal bile. However, UDCA differed in the direction of change in fasting duodenal bile (Table 5).^{45, 54, 55} Interestingly, it was found that all patients in the studies listed in Table 5 were patients with CD with a history of intestinal resection in the ileum (Table 1). Furthermore, compared with patients with CD without a history of ileal resection, operated patients with CD also showed higher levels of total BAs, CA, and CDCA,^{40, 53} along with lower levels of DCA and LCA, in the feces (Table 5).⁴⁰ Finally, it was reported that GCA⁴⁴ was decreased in postprandial serum and that TBAs⁴³ were decreased in both the fasting serum and the postprandial serum in patients with CD with a history of intestinal resection (Supplementary Table 6).

For patients with UC, when compared with HCs, an increase in the CA,^{35, 37, 56, 57} GBAs,^{57, 58} TBAs,^{57, 58} and TCA^{35, 58} of fecal samples was observed; a decrease in the DCA^{35, 37, 56-58} and LCA^{37, 42, 56, 57} of fecal samples was found; and an increase in CA in postprandial serum and in CDCA in both fasting serum and postprandial serum was identified (Table 6).⁵⁹ The changing profile of total BAs in the feces was inconsistent between several studies,^{37, 56, 58} as was CDCA^{42, 56} and GCDCA (Table 6).^{42, 58} In addition, in patients with UC, GCA,⁵⁸ GUDCA,³⁵ TCDCA,⁵⁸ nonsulfated BAs,⁵⁸ and keto-DCA³⁵ were increased in the feces; SBAs,⁵⁷ UBAs,⁵⁷ GLCA,⁴² TLCA,⁴² 7-dehydroxylated BA,⁵⁷ and the ratio of DCA/CA were reduced in feces⁵⁸; and there was a decrease in the ratio of DCA/(DCA+CA) in fasting serum in both the remission and exacerbation states of UC (Supplementary Table 7).⁴⁸ However, in 5 articles, the authors found that there was no significant difference in BA levels between patients with UC and HCs.^60-64

In the present review, only BAs of the same sample type, for which 2 separate studies agreed on the direction of difference, were considered consistent alterations in patients. Collectively, there seemed to be a characteristic trend of BA changes in both CD and UC; that is, PBAs and CBAs were augmented and SBAs were diminished in fecal samples. Although the changes in BAs in serum, plasma, urine, or fasting duodenal bile were also observed in IBD, UC, and CD, definite changes could not be completely determined because of the restricted number of studies and limited consistency in results. Based on the findings derived in the present review, certain consistent changes in fasting duodenal bile could be determined, including CD patients with a history of ileal resection had decreased level of DCA, and CD patients with colonic involvement showed an increase in CA and a decrease in DCA. Finally, for BA pool size, an increase in the UDCA pool and a decrease in the total BA pool and CDCA pool were found in patients with CD without a history of intestinal resection.

In addition, we observed differences and similarities in terms of BA profiles between patients with CD and patients with UC (Table 7). When the BA profile was compared between HCs and patients, the changing profiles of patients with CD and patients with UC were different. Weng et al⁴² found that CDCA, LCA, and TLCA were significantly less abundant in the fecal samples of both UC and CD groups, whereas GCDCA and GLCA were decreased only in the fecal samples of UC groups. Another study reported that CA and TCA were found to be at high levels and DCA was found to be at low levels in the fecal samples of both UC and CD groups; an increase in GCA, TCDCA, and taurodeoxycholic acid levels and a decrease in LCA and UDCA levels was noted in the fecal samples of patients with CD; and GUDCA and keto-DCA were increased in the fecal samples of patients with UC.³⁵ Murakami et al⁴⁸ proposed that the ratio of DCA/(DCA+CA) was significantly reduced in the remission state of CD and in both the remission and exacerbation states of UC. Kruis et al³⁷ indicated that CA level was increased and that DCA and LCA levels were decreased in the fecal samples of both patients with UC and patients with CD, total BAs and CDCA and UDCA levels were increased in the feces of patients with CD, and total BAs were decreased in the feces of patients with UC.

Moreover, serum BA profiles in response to different treatment regimens (anti-tumor necrosis factor [TNFs] agents or conventional therapy like 5-aminosalicylates) in patients with CD and patients with UC are different.⁶⁵ In patients with CD treated using conventional therapy, serum BA levels were lower and the ratio of PBAs/SBAs was higher than those in HCs. However, in patients with UC treated using conventional therapy, serum BA profiles were similar to those in HCs. When the BA profile was compared between patients treated using conventional therapy and using anti-TNFs, research has found that patients with CD treated using anti-TNFs had increased serum BA levels and SBA levels, with a decreased ratio of PBAs/SBAs, whereas BA profiles were similar between patients with UC treated using conventional therapy and using anti-TNFs. In addition, when the BA profile was compared between patients with CD and patients with UC, urinary concentrations of DCA, 7-keto-LCA, and β-ursocholic acid in patients with UC were higher than in patients with CD.³⁸ CD and UC may have different effects on alterations of the BA profile, and further studies to confirm these differences may be promising to provide a method to differentiate patients with UC from patients with CD.

POSSIBLE MECHANISMS FOR ALTERATIONS

Alterations in Absorption

The BAs are reabsorbed in the intestine to participate in enterohepatic circulation.²⁰ In addition, ASBT in the ileum plays a vital role in this process; thus, reduced ASBT function impairs intestinal absorption function, leading to bile acid malabsorption (BAM) and increased excretion of BAs in the feces.¹⁸ As reported by Vítek,¹⁸ BAM is prevalent in IBD and is a common symptom manifesting in patients with IBD. The proinflammatory cytokines, such as interleukin (IL)-1β and TNF, have been seen to repress the activity of the ASBT promoter in vitro and in vivo, which may be a reason for the lower expression of ASBT in patients with IBD.^66-68 Compared with its expression profile in HCs, ASBT is downregulated in patients with active UC and downregulated in patients with CD regardless of whether they are in the active disease state or in remission.^{68, 69} This finding indicates that the decisive factor for the decreased ASBT expression in CD may be the disease itself; local inflammation may serve as an auxiliary aggravating factor. Generally, CD involves the ileum, which is the main site of ASBT expression.⁷⁰ For patients with CD with a history of ileal resection, the degree of BAM may be more serious. Lenicek et al⁷¹ reported that in patients with CD, the severity of BAM is associated with resection of the ileum. The most severe BAM occurs in patients with CD after ileal resection; however, BAM can occur in unoperated patients with CD, regardless of disease localization. Furthermore, in patients with active UC, it has been reported that absorption of UBAs was augmented in the diseased colon.⁵⁹ This finding may be one of the factors responsible for the rise in CA and CDCA in the serum.

Alterations in Synthesis

Research has shown that FXR is a physiological BA sensor that is likely to play an essential role in BA homeostasis via the regulation of genes involved in BA metabolism.^{23, 72} In humans, the enterocyte BA pool is detected by FXR, which regulates BA production via the feedback inhibition of CYP7A1 by FXR/fibroblast growth factor 19/hepatic fibroblast growth factor receptor 4 signaling.^{28, 73} Therefore, BAM caused by defective ASBT function in both CD and UC could induce a decreased enterocyte BA pool, decreased inhibition of FXR, and subsequent increased production of BAs.

Alterations in Bacterial Modification

The conversion of PBAs to SBAs is dependent on several factors—that is, availability of the bacteria, adequate bile flow, adequate contact of the bacteria with the bile, and diet.⁷⁴ Dysbiosis of bacteria in patients with IBD may lead to altered availability of the bacteria; furthermore, BAM-associated diarrhea in patients with IBD may result in inadequate contact of bacteria with the bile.

Variations in the capacity for BA modification resulting from dysbiosis of bacteria may act as a significant factor in the onset or progression of IBD.⁷⁵ Modifications in BA such as deconjugation and transformation are carried out by a broad spectrum of intestinal bacteria, and the Firmicutes phylum has occupied a pivotal position in these processes.^{19, 35} The deconjugation, transformation, and desulfation activities of the microbiota have been shown to be impaired in patients with IBD.^{14, 35, 75, 76} It is worth noting that the modification activities of bacteria belonging to the Firmicutes phylum (eg, the Enterococcaceae, Eubacteriaceae, and Ruminococcaceae families) are especially lower in patients with IBD.^{35, 75} The decreased capacity for BA modification in Firmicutes is consistent with the observation of studies focusing on alterations of gut microbiota composition in IBD.⁶ In addition, the 7α-dehydroxylation enzymatic reaction is inhibited below pH 6.5. The pH of stool in patients with CD has been shown to be acidic (5.2 ± 0.5), which may be another reason for the decreased conversion of PBAs to SBAs in patients with CD.³⁷ Alterations in the microbial metabolism of BAs in patients with IBD tend to result in higher levels of CBAs, PBAs, and 3-OH-sulfated BA, with a lower level of SBAs.

With the understanding of the crucial role of intestinal microflora in BA metabolism, several studies have reported a correlation between certain intestinal bacteria and specific BAs in feces. Most of these bacteria are members of the Firmicutes phylum, which underscores the pivotal role of the Firmicutes phylum in BA metabolism. For example, SBAs, DCA, and LCA are positively correlated with multiple bacteria, including the order Clostridiales, the families Lachnospiraceae and Ruminococcaceae, and the following genera: Alistipes, Anaerotruncus, Bacteroides, species belonging to Clostridium IV and XIVa, Coprococcus, Faecalibacterium, Gemmiger, Lachnospira, Odoribacter, Oscillibacter, Roseburia, Ruminococcus, Ruminococcus 2, and Sporobacter. They are negatively correlated with the genera Enterococcus, Klebsiella, Lactobacillus, Proteus, and Veillonella.^{74, 77-80} LCA is positively correlated with Fusobacterium nucleatum.⁴² Vrieze et al⁷⁸ reported that short-chain fatty acid (butyrate)–producing bacteria (Faecalibacterium prausnitzii and Eubacterium hallii) are positively correlated with fecal SBAs and inversely correlated with fecal PBAs. Furthermore, PBAs are positively correlated with the phylum Proteobacteria, the order Clostridiales, the families Enterobacteriaceae and Erysipelotrichaceae, and the genera Enterococcus, Lactobacillus, Streptococcus, Sutterella, Turicibacter, and Lactobacillus plantarum; they are negatively correlated with the family Ruminococcaceae and the genera Anaerotruncus, Butyricicoccus, Coprococcus, Eubacterium, Faecalibacterium, Flavonifractor, Gemmiger, Odoribacter, Oscillibacter, Ruminococcus, and Sporobacter.^78-80 Microbiota abundance of Ruminococcus gnavus and Clostridium clostridioforme is positively correlated with changes in CA and its conjugates, whereas Alistipes finegoldii abundance is negatively correlated.³⁴ In addition, CDCA is positively correlated with the family Enterobacteriaceae and negatively correlated with Alistipes finegoldii.^{34, 74}

Furthermore, several ratios representing 7α- dehydroxylation activity are related to certain bacteria. For instance, DCA/CA is positively correlated with the family Ruminococcaceae and LCA/CDCA is positively correlated with the genus Blautia. In addition, DCA/(DCA+CA) and LCA/(LCA+CDCA) are positively correlated with the genus Clostridium subcluster XIVa.^{48, 74} The correlations between metabolites and bacteria may provide deeper insight into host-gut microbiota interactions in IBD. In addition, certain ratios related to BAs may be useful surrogate markers for the intestinal proportion of certain bacteria, and characteristic changes in BAs may be considered a reliable marker of IBD-associated dysbiosis.^{14, 48}

Collectively, alterations in the absorption, synthesis, and bacterial modification of BAs may lead to distinctive BA changes in patients with IBD. Regarding the changes in BAs identified in the previous section of the current review, ASBT dysfunction may lead to increased fecal excretion of CBAs and/or total BAs in IBD, CD, and UC. In addition, the reduction in the mean postprandial increase of total BAs, conjugated CA and CDCA, and GCA in the serum could also be a result of the impaired absorption function caused by ASBT inhibition. However, the increase in certain UBAs in the serum may be a result of accelerated absorption caused by local intestinal lesions in patients with IBD. The increased production of BAs, the decreased ability to metabolize PBAs to SBAs, and BAM may result in an increase in PBAs and a decrease in SBAs in fecal samples in both CD and UC. In addition, for BA pool size, the increased fecal excretion of total BAs and CDCA and a more rapid conversion of CDCA to UDCA may lead to the decrease of the total BA pool and CDCA pool in addition to the increase of the UDCA pool in patients with CD.⁸¹ The discrepant effects of different disease locations on gut microbiota and the BA absorption function may be the possible explanation for the differences in BA changes in patients with IBD with different disease locations.⁸²

POSSIBLE EFFECTS OF ALTERED BAs ON IBD

The normal intestinal tract and gut microbiota of healthy patients maintain BA homeostasis. They also ensure that the host avoids the harmful effects and utilizes the favorable effects of BAs.⁸³ Bennet⁸⁴ proposed that in genetically susceptible individuals, UC is caused by BA bacterial metabolites. This implies that the BA bacterial metabolites produced by abnormal intestinal flora may prove to be a critical factor in the pathogenesis of IBD. The BAs are cytotoxic when present in abnormally high concentrations.⁸³ Cytotoxicity may occur intracellularly, as seen in the hepatocytes in cholestasis, or extracellularly, as seen in the gut in patients with BAM.⁸³ More important, alterations in the activation of BA receptors and in other signaling pathways resulting from an altered BA profile may exert a significant influence on the host. Abnormally high levels of CBAs and PBAs may lead to an increase in their harmful effects, and the protective effects of SBAs are weakened when their levels are low. A previous study has reported that SBA markers are significantly elevated and highly represented in the serum of patients responding to TNF treatment whereas higher levels of PBAs are associated with patients who do not respond to the anti-TNF treatment in CD.⁸⁵ In addition, Roda et al⁶⁵ have proposed that SBAs may serve as an indirect biomarker of the healing process in CD. The results of these studies seem to indicate the “bad effect” of PBAs and the “good effect” of SBAs in patients with CD from another perspective.

Affecting Gut Microbiome

The BAs shape the intestinal microbiota community and prevent bacterial overgrowth in the small intestine by direct bacteriostatic activities and BA-FXR signaling in intestinal epithelial cells.^{27, 86} Thus, BAs at abnormal concentrations may affect the gut microbiome structure and even induce dysbiosis. In addition, we observed that TBAs could serve as nutritional substrates to promote the growth of possible pathogens in IBD. A Western diet high in saturated (milk-derived) fat has been reported to result in an increase of TCA, thus increasing the availability of organic sulphur used by Bilophila wadsworthia and promoting its expansion. The bloom of this bacteria has been shown to promote a Th1-mediated immune response and the development of colitis in IL-10^-/- mice, which may reveal another mechanism by which alterations in the BA profile lead to changes in gut microbiota and eventually trigger IBD.⁸⁷ We hypothesized that higher levels of TBAs resulting from changes in BA metabolism in patients with IBD may also contribute to the overgrowth of B. wadsworthia.

Adherent-invasive Escherichia coli (AIEC) may contribute to IBD pathogenesis^9,88 and it has a functional lpf operon encoding long polar fimbriae, which allows AIEC bacteria to interact with Peyer’s patches and to translocate across M cells.⁸⁹ In vitro and in vivo experiments have indicated that BAs, especially CDCA, acting as an activator of lpf transcription increase the expression of the lpf operon in a dose-dependent manner, promote the colonization of AIEC in the gut, and thus enhance the pathogenicity of AIEC in CD.⁹⁰

The BAs affect the gut microbiome, and altered gut bacteria further influence BA metabolism. In patients with IBD, dysbiosis of bacteria and alterations of BAs may form a vicious circle and exert a profound influence on the development and progression of IBD.

Direct Cytotoxicity and Other Mechanisms

The harmful effects of CBAs and PBAs may be enhanced owing to their increase in fecal samples of patients with IBD. If BAs are at abnormally high levels, then they are cytotoxic, and their cytotoxicity is strongly related to their concentration and hydrophobicity: The greater the hydrophobicity, the higher the concentration and the greater the cytotoxicity.⁸³ The natural dihydroxy BA CDCA is hydrophobic and highly cytotoxic; CA is moderately hydrophobic, which is noncytotoxic at low concentrations but cytotoxic at very high concentrations.⁸³ Mechanistically, CDCA can damage cellular DNA and rapidly induce cell apoptosis at 500 μM in cell experiments, without first causing growth suppression.^{91, 92} It has been proved in vitro that this effect may be a result of oxidative stress with increased reactive oxygen generation; with longer treatment times with CDCA, apoptosis is followed by secondary necrosis because of impaired mitochondrial activity and ATP depletion.⁹³ Epidermal growth factor receptor phosphorylation has been induced by CA and CDCA, resulting in an increased paracellular permeability via occludin dephosphorylation and cytoskeletal rearrangement at the tight junction level in vitro.⁹⁴ Furthermore, in patients with IBD, BAs could permeate to adjacent epithelial cells from areas of damage caused by inflammation in the physical epithelial barrier, which may play an additional role in exacerbating epithelial barrier function lesions of BAs.⁹⁵ Autophagy also has an important role in the pathogenesis of IBD.⁹⁶ Interestingly, BAs have been known to regulate and modulate autophagy in hepatocytes in vitro.⁹⁷ We hypothesize that the altered BA profiles in patients with IBD may also cause changes in the autophagy function of intestinal cells and have effects on the occurrence and development of IBD, and this is worth discussing in future research.

The protective effects of SBAs may be reduced owing to their decrease in fecal samples of patients with IBD. At physiological concentrations, SBAs are reported to have inhibitory effects on different inflammatory response pathways. For example, an in vitro study indicated that DCA reduced the phagocytosis of monocytes and inhibited the production of TNF through a direct inhibitory effect on the monocytes, independent of endotoxin inactivation or cytolysis.⁹⁸ By inhibiting the activation of nuclear factor-kappa B via VDR in colonic epithelial cells in vitro, LCA could reduce the secretion of nuclear factor-kappa B target genes, such as IL-8.⁹⁹ In addition, in vitro experiments showed that LCA could impede Th1 differentiation of CD4⁺ Th cells and Th1 activation in a VDR-dependent manner, thus inhibiting the production of Th cytokines and proinflammatory responses.²⁹ As a strong endogenous inhibitor of the nucleotide oligomerization domain-like receptor family pyrin domain containing 3 inflammasome activation via the G-protein-coupled receptor-cyclic adenosine monophosphate-protein kinase A axis, LCA has been shown to reduce the secretion of inflammatory cytokines in a dose-dependent manner both in vitro and in vivo.¹⁰⁰ Moreover, antimicrobial peptides contribute to the protective barrier against microbes at epithelial surfaces, and their expression could be induced by LCA in vitro.¹⁰¹ Therefore, IBD-associated impaired BA biotransformation could reduce the level of SBAs, weaken the anti-inflammatory effects of SBAs, and contribute to the inflammatory response in IBD.

BAM and Diarrhea

Patients with IBD are generally characterized by having chronic diarrhea, and it is often believed that inflammatory processes in the intestinal wall are the main contributing factor in IBD-associated diarrhea and related symptoms.^{18, 102} It has been found that BAM also plays a critical role, especially in patients with ileal disease, but it is often underestimated and ignored by clinicians.^{18, 102} When BA diarrhea occurs, it is commonly seen in CD patients with ileal involment or ileal resection.⁷³ Studies have shown that BAM leads to increased BA levels in the colon, which enhance mucosal permeability, activate chloride channels to induce water and electrolyte secretion, accelerate colonic transit partly by stimulating propulsive high-amplitude colonic contractions, and eventually cause diarrhea.^73,103 The BAs such as CDCA could significantly reduce the defecatory urge threshold and the maximum tolerated rectal volume in vivo, which may also contribute to diarrhea in patients with IBD.¹⁰⁴

Intestinal Dysplasia and Carcinoma

A prospective study of a population of patients with extensive UC indicated that high fecal BA concentrations are associated with dysplasia and carcinoma in UC.¹⁰⁵ In addition, it has also been confirmed in an experimental murine UC model that fecal CA and colitis may be intimately related to the development of colorectal neoplasia.³¹

PROMISING THERAPEUTIC APPROACHES

In the present review, a changing trend in the composition of BAs was shown: specifically, that PBAs and CBAs are increased and SBAs are decreased in the feces of patients with IBD. Because these alterations may be crucial to the progression of IBD and could lead to dysplasia and carcinoma of intestinal tissue, further studies that investigate effective treatment measures targeted to BA changes are urgently required. It is feasible that BA therapy is divided into 2 types: displacement and replacement.⁸³ In displacement therapy, for increased PBAs and CBAs, it may be useful to appropriately inhibit the synthesis of BAs and use BA binders, such as cholestyramine, to promote their elimination. In replacement therapy, supplementing SBAs to correct BA deficiency may be meaningful; it has been confirmed that supplementation with SBAs is effective in reducing intestinal inflammation in 3 murine colitis models.¹²

This review has several limitations. First, the composition of BAs in patients with IBD may be affected by diet or medication, which may serve as a source of differences between the individual studies. Second, varying degrees of heterogeneity in the methodology, including sample type, sample storage methods, and quantitative method, may also explain the differences in studies. Third, the heterogeneity in quantified BA species and the sample type in the present study makes it difficult to unify the results. The presentation forms of BA levels also vary among studies, including daily fecal excretion rate, proportion relative to total BAs or total metabolites, and concentrations in feces or serum. These limitations make it impossible to complete a quantitative inference, and a metalevel statistical inference cannot be performed to reveal the influence of multiple covariates on BA levels. Our study only provides a qualitative inference about alterations in BAs to reveal the possible changing trend of BA metabolism in patients with IBD. Finally, the limited number of studies on different clinical characteristics of IBD, such as disease activity, disease location, and disease extent, has led to a lack of subgroup analysis of these critical issues in IBD. Thus, these limitations could result in an incomplete understanding of the compositional and metabolic alterations of BAs in IBD.

Conclusions

This review has summarized BA alterations in different sample types and identified a probable changing trend in patients with IBD. These alterations are closely related to IBD-associated dysbiosis and may reflect another mechanism by which dysbiosis contributes to the onset and development of IBD—that is, by altering bacterial metabolites. The gut microbiota may be a key regulator of BA composition and metabolism, which in turn play a critical role in shaping the gut microbiome structure. Although the causality of the relationship between changes in intestinal flora and BA alterations in IBD cannot be determined, it is certain that they may impact each other and exert significant effects on IBD. Larger patient cohorts are required to confirm the complex relationship between IBD, intestinal flora, and BAs. In addition, future studies should investigate changes in the composition and metabolism of BAs associated with disease activity, disease location, and disease extent. Studies that perform a metalevel statistical inference to reveal the influence of multiple covariates on BA levels are also warranted. Finally, appropriate elimination of excess BAs in the intestinal cavity and inhibition of BAs synthesis, along with supplementation of SBAs, may be promising therapeutic approaches for IBD. Further animal and clinical trials are needed to confirm the protective effects of these therapeutic strategies, and the specific supplementary dosage of SBAs needs to be carefully determined.

Acknowledgments

The data underlying this article are available in the article and in its online supplementary material.

Conflicts of interest: The authors have no conflicts of interest to disclose.

Author contributions: Guarantors of the article: Xiaojun Zhuang and Zhirong Zeng. Xiaojun Zhuang and Zhirong Zeng designed the study. Na Li wrote the manuscript. Na Li, Shukai Zhan, and Caiguang Liu collected the data. Zonglin Xie and Zhenyi Tian analyzed the data. Minhu Chen and Shenghong Zhang revised the manuscript. All authors approved the final version.

References

Author notes