In this study, we conducted an extensive investigation into the causal relationships between 468 blood metabolites and three IBD phenotypes using a robust Mendelian randomization (MR) design. Leveraging a meta-analysis dataset of mGWAS (metabolite Genome-Wide Association Studies) and IBD GWAS, we employed genetic proxies for metabolites to establish associations of 80 potential metabolites with the three IBD phenotypes.Through complementary analysis and sensitivity analysis, including reverse Mendelian randomization, we identified 4 candidate metabolites for IBD, 8 for CD, and 4 for UC with MR evidence. Furthermore, utilizing the Multivariable Mendelian Randomization (MVMR) estimation, we identified metabolites that directly influence IBD, CD, and UC independently of other metabolites.In-depth investigation into the biological roles of these candidate metabolites revealed the enrichment of 8 key metabolic pathways associated with the three phenotypes (IBD, CD, and UC). Additionally, LDSC analysis provided support for the genetic correlation between 1,5-anhydroglucitol and CD.Notably, this study represents the first comprehensive application of blood metabolite GWAS data to explore causal relationships with IBD. It further integrates pathway enrichment and LDSC in MR studies, advancing our understanding of the complex relationships between blood metabolites and IBD phenotypes.
In recent years, the incidence of IBD has been gradually increasing, posing significant challenges to public health worldwide. IBD patients face chronic inflammation, disruption of intestinal function, and a severe decline in quality of life. In this context, the search for new diagnostic and treatment methods has become critically important. The emergence of metabolomics technology in recent years has sparked a strong interest in exploring the value of metabolic products in IBD. Metabolomics is a technique that involves studying the biological state and physiological-pathological processes by analyzing the combination of metabolites within an organism. In disease research, the analysis of blood metabolites is considered to have significant research value. Blood metabolites offer a comprehensive view of biological mechanisms as they can capture changes in both endogenous and exogenous metabolic processes simultaneously. Notably, recent evidence highlights the involvement of Tryptophan and caninurenine in the remission and deterioration of colitis, underscoring their relevance as potential indicators of disease progression. A study of 99 patients with IBD reported that the ratio of kynurenine/Tryptophan in serum samples of Ulcerative colitis patients and the expression of IDO in mucosal samples are positively correlated with endoscopic inflammation[26]. High expression of IDO1 in vivo results in increased levels of kynurenine, which, in turn, activates the AHR (Aryl hydrocarbon receptor) to regulate the immune system's tolerance state. However, prolonged immune system tolerance can hinder the recovery of inflammatory tissue, weaken the ability to defend against and clear pathogens, leading to chronic inflammation[27]. Notably, canine urine activates the AHR to promote the differentiation of immature CD4 + T cells into Treg cells, reducing the level of T17 cells in vivo and ultimately leading to immune tolerance. This highlights the potential of canine urine in inhibiting sustained inflammatory responses in chronic intestinal inflammation, thus offering promising targets for the treatment of such conditions.Although previous research has provided compelling evidence that metabolites are involved in the biological mechanisms of IBD, potentially offering therapeutic benefits, their contribution to early screening and prevention of IBD remains limited due to the unclear causal relationship between the two. Therefore, in this investigation, we conducted a meticulous MR study to unravel the causal connection between blood metabolites and IBD, as well as their associated metabolic pathways. Our objective is to provide valuable insights and reference guidance for the screening and treatment of IBD.
Methionine, an essential amino acid in humans, has been the subject of several studies exploring its potential impact on inflammatory bowel disease (IBD). In a study involving young patients in Finland, significant increases in methionine levels were observed in patients with ulcerative colitis (UC) compared to control individuals with Crohn's disease (CD) or unclassified IBD, along with four other sulfur-containing metabolites[28]. In mouse models of IBD, Yuka et al. found that administering D-methionine or butyrate activated the NLRP3 and Nrf2 signaling pathways, resulting in anti-inflammatory effects and potential therapeutic benefits for IBD. Similarly, Liu et al. investigated the effects of a methionine-restricted diet and observed reduced colitis scores in DSS-treated mice, suggesting that a methionine-restricted diet could serve as a treatment for IBD[29]. In our study, similar to previous findings, methionine was identified as a risk factor for UC. Moreover, homocysteine, a byproduct of the methionine and cysteine metabolism pathway, is known to be a pro-inflammatory factor and plays a critical role in cellular stress and inflammation processes. Homocysteine's association with IBD occurrence was explored in an in vitro study, where treatment of human intestinal microvascular endothelial cells with homocysteine led to the upregulation of vascular cell adhesion molecule-1, monocyte chemoattractant protein-1, and phosphorylation of p38 mitogen-activated protein kinase[30]. Additionally, another in vitro experiment demonstrated that homocysteine exacerbated DSS-induced colitis in mice by stimulating the expression of IL-17 through the p38/cPLA2/COX2/PGE2 signaling pathway[31]. Although the existing evidence indicates potential links between methionine, homocysteine, and IBD, further prospective research is necessary to fully understand the relationship between methionine and UC and their implications for the development and management of IBD.
Phenylacetylglutamine (PAGln) is a prethrombotic metabolite associated with a heightened risk of cardiovascular disease due to its role in enhancing platelet activation[32]. Recent research has revealed a robust association between PAGln and CD. PAGln not only promotes platelet hyperreactivity but also activates the CD40 pathway, triggering inflammatory cascades and contributing to the pathogenesis of CD[33]. This emphasizes the substantial influence of PAGln on the progression of CD and underscores the potential of platelets as a promising therapeutic target for inflammatory bowel disease (IBD).PAGln is a microbial-derived metabolite originating from the gut. The gut microbiota first converts dietary phenylalanine into phenylacetate (PAA), which subsequently combines with host liver enzymes to form PAGln. In a study characterizing 84 serum metabolites using gas chromatography-mass spectrometry, Sitkin et al. identified at least 18 metabolites with a combined origin (human + microbial). Among UC patients, the serum exhibited elevated levels of PAA, 4-hydroxyphenylacetate (4-HPAA), indole-3-acetate (IAA), succinate (SA), and fumarate (FA), suggesting their potential relevance to UC pathology[34]. However, there is still limited research on PAGln blood metabolites, and more prospective studies are needed.
Serine, an amino acid existing in two forms, L-serine and D-serine, has been a subject of interest in understanding its potential impact on inflammatory bowel disease (IBD). In a study by Asakawa et al., administering D-serine to mice before inducing colitis resulted in reduced T cell infiltration in the lamina propria and decreased colonic inflammation. In vitro experiments further demonstrated that D-serine inhibited the proliferation of activated CD4 T cells and limited their differentiation into Th1 and Th17 cells. These findings suggest that D-serine may play a preventive and therapeutic role in IBD, offering potential benefits for the management of the condition[35]. Han et al. conducted a fine mapping study to identify major risk factors for CD in the Asian population, specifically focusing on HLA alleles and amino acids. They found that amino acid position 37 of HLA-DRβ1 significantly influences CD susceptibility and indicated that the presence of serine at this position is protective against CD (P = 3.6×10–67, OR = 0.48 [0.45–0.52])[36]. Our study concludes that serine has a protective effect against CD, which is consistent with our findings.
Glycochenodeoxycholate is a bile acid synthesized in the liver through the conjugation of glycine and chenodeoxycholic acid, commonly found in the form of sodium salt. Its primary function lies in acting as a detergent to facilitate the solubilization of fats for absorption.Furthermore, glycochenodeoxycholate is associated with chenodeoxycholic acid, an essential secondary bile acid known for its significant role in preventing IBD[37]. Bile acids (BAs) are amphipathic molecules synthesized from cholesterol in the liver, acting as physiological detergents that promote the liver-biliary secretion of endogenous and exogenous metabolites. In the intestinal environment, BAs play a vital role in the absorption of dietary fats, fat-soluble vitamins, and other nutrients. Beyond their traditional role as digestive surfactants, BAs have gained recognition as signaling molecules with diverse biological functions. These functions include regulating glucose and lipid metabolism, influencing energy balance, and modulating immune responses[37]. The regulatory functions of BAs primarily involve activating intracellular ligand-activated nuclear receptors (NRs), such as the FX receptor (FXR, NR1H4), and cell surface G protein-coupled receptors (GPCRs), particularly the G protein-coupled BA receptor (TGR5 or GPBAR-1). Bile acids act as high-affinity ligands for FXR, with chenodeoxycholic acid (CDCA) being the most potent BA for FXR activation. FXR plays a crucial role in immune regulation and intestinal barrier function. It contributes to preserving the integrity of the intestinal barrier, modulating inflammatory responses, preventing bacterial translocation in the gut, and alleviating inflammation through diverse mechanisms[38].The role of CDCA in preventing IBD underscores the potential protective effect of glycochenodeoxycholate in IBD, which aligns with the findings of our research. This further suggests the significance of bile acids and their associated metabolites in the context of IBD prevention and the regulation of intestinal health.
Arachidonic acid, an essential fatty acid, is synthesized by three enzymes, namely cyclooxygenase (COX-1, COX-2), lipoxygenase (LOX-5, -8, -12, -15), and cytochrome P450 (CYP450)[39]. These enzymes are sourced from a range of essential fatty acids, including linoleic acid (LA), dihomo-γ-linolenic acid (DGLA), arachidonic acid (ARA), as well as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)[39].Arachidonic acid serves as the primary precursor for the synthesis of prostaglandins, and it also contributes to the production of lipoxins, which play a significant role in inhibiting inflammation[40]. Recent studies have demonstrated that lipoxin analogs have beneficial effects in controlling inflammation[41].Notably, a study by Vong et al. reported higher levels of LXA4 in mucosal biopsy tissues of patients in remission[42],suggesting a potential protective role of arachidonic acid in ulcerative colitis (UC), which aligns with our research findings.However, some authors argue that the inflammatory properties of inflammatory bowel disease (IBD) are associated with elevated levels of arachidonic acid metabolites, such as prostaglandin E, prostaglandin 2, and leukotriene B4[43]. In the initial phases of inflammation, there is an upregulation of pro-inflammatory enzymes, including 5-lipoxygenase (5-LOX) and COX-2. This results in the production of various inflammatory mediators, such as leukotriene B4 (LTB4), leukotriene C4 (LTC4), prostaglandin E2 (PGE2), and prostaglandin D2 (PGD2), which collectively contribute to the progression of the inflammatory response[44]. Further research is needed to explore the precise role of arachidonic acid in UC or IBD and to elucidate the underlying mechanisms responsible for its dual effects in inflammation.
In our study, we discovered two carbohydrates, glucose and 1,5-anhydroglucitol, that showed associations with CD. Interestingly, LD score regression analysis indicated a certain degree of genetic correlation between 1,5-anhydroglucitol and CD. Elevated blood glucose levels are strongly linked to diabetes-related diseases, and serum levels of 1,5-anhydroglucitol have emerged as a novel screening tool for diabetes[45]. Notably, diabetes is recognized as one of the high-risk factors for CD[46]. A recent Danish cohort study reported an increased risk of type 2 diabetes in CD and UC patients, independent of glucocorticoid use. Moreover, Jasser et al. observed an elevated risk of IBD in children and adolescents with type 1 diabetes in German and Austrian populations. These findings imply the presence of common pathogenic pathways between IBD and diabetes and indirectly support the notion that blood metabolites, namely glucose and 1,5-anhydroglucitol, may contribute to the development of CD. Surprisingly, in our analysis, glucose exhibited a protective effect in the univariate analysis of CD but turned into a high-risk factor after adjusting for multiple factors. Conversely, 1,5-anhydroglucitol was identified as a risk factor in the univariate analysis but emerged as a protective factor for CD after accounting for multiple factors. This intriguing contradiction may warrant further prospective research to ascertain the intricate relationship between these metabolites and CD.
Additionally, our results indicate that 2-stearoylglycerophosphocholine, 2-hydroxyisobutyrateuridine, and pyroglutamine are risk factors for the IBD phenotype. Uridine and cortisol are risk factors for the CD phenotype, while gamma-glutamylglutamate and glycerophosphorylcholine are protective factors for the CD phenotype. 2-stearoylglycerophosphocholine is a risk factor for the UC phenotype. However, the mechanisms by which these metabolites influence their respective phenotypes are not yet fully understood and require further experimental exploration.
Among the identified metabolic pathways, Glutathione metabolism stands out as a crucial hub in T cell-mediated inflammatory responses[47].Notably, Sahoo et al. discovered that lipopolysaccharide (LPS) is linked to chronic intestinal inflammation, and the interaction between the LPS/TLR4 signaling pathway and Glutathione metabolism may significantly contribute to driving chronic intestinal inflammation[48]. Moreover, the primary bile acid biosynthesis pathway has also been implicated in the pathogenesis of IBD[49].Notably, our study revealed a shared metabolic pathway, Ether lipid metabolism, among the IBD, UC, and CD phenotypes. However, limited research has been conducted on this specific metabolic pathway to date, and more prospective studies are warranted to gain a comprehensive understanding of its role in the context of these phenotypes.
This MR analysis presents several notable strengths. Firstly, it represents the most comprehensive and systematic investigation to date, encompassing a thorough examination of 486 blood metabolites in relation to IBD. Secondly, we applied rigorous MR analysis methods to overcome previous limitations, such as reverse causation and confounding factors. Through the careful implementation of various techniques, we ensured the removal of any factors that could potentially violate the assumptions of MR, leading to highly reliable estimates. The consistent direction and robustness of the three MR estimates were further supported by sensitivity analyses, bolstering the credibility of our findings. Additionally, we conducted co-localization analysis and LDSC analysis, effectively exploring the genetic correlation between metabolites and IBD.
The current study has certain limitations that should be acknowledged. Firstly, the number of available SNPs for the exposure of interest at the genome-wide level was limited. To address this concern, we employed a slightly relaxed threshold for our MR analysis, a common approach in similar studies. Nonetheless, it is noteworthy that all selected SNPs had F-statistics exceeding 10, ensuring the robustness of our instrumental variables (IVs). Moreover, the consistency of the results in the reverse MR tests provided further support for our relaxed threshold setting.Secondly, in an effort to minimize the influence of population differences, we specifically utilized GWAS data from individuals of European ancestry for our MR analysis. As a result, the generalizability of our findings to other populations requires further exploration and validation.A third limitation lies in the precision of MR estimates, which can be influenced by sample size. Thus, it is imperative to increase the sample size to confirm the reliability of our results.Furthermore, it is essential to recognize that while MR analysis offers valuable insights into etiology, the clinical applications of our findings should be approached cautiously. Rigorous validation through randomized controlled trials and fundamental research is necessary before considering any clinical implications.