Gut microbiota and common gastrointestinal diseases: a bidirectional two-sample Mendelian randomized study

Background Several recent studies have shown an association between gut microbiota and gastrointestinal diseases. However, the causal relationship between gut microbiota and gastrointestinal disorders is unclear. Methods We assessed causal relationships between gut microbiota and eight common gastrointestinal diseases using Mendelian randomization (MR) analyses. IVW results were considered primary results. Cochrane’s Q and MR-Egger tests were used to test for heterogeneity and pleiotropy. Leave-one-out was used to test the stability of the MR results, and Bonferroni correction was used to test the strength of the causal relationship between exposure and outcome. Results MR analyses of 196 gut microbiota and eight common gastrointestinal disease phenotypes showed 62 flora and common gastrointestinal diseases with potential causal relationships. Among these potential causal relationships, after the Bonferroni-corrected test, significant causal relationships remained between Genus Oxalobacter and CD (OR = 1.29, 95% CI: 1.13–1.48, p = 2.5 × 10–4, q = 4.20 × 10–4), and between Family Clostridiaceae1 and IBS (OR = 0.9967, 95% CI: 0.9944–0.9991, p = 1.3 × 10–3, q = 1.56 × 10–3). Cochrane’s Q-test showed no significant heterogeneity among the various single nucleotide polymorphisms (SNPs). In addition, no significant level of pleiotropy was found according to the MR-Egger. Conclusion This study provides new insights into the mechanisms of gut microbiota-mediated gastrointestinal disorders and some guidance for targeting specific gut microbiota for treating gastrointestinal disorders.


Introduction
Gastrointestinal disorders have long been a widespread health problem globally, encompassing a wide range of conditions such as gastroesophageal reflux disease (GERD), ulcerative colitis (UC), Crohn's disease (CD), irritable bowel syndrome (IBS), gastric ulcer (GU), duodenal ulcer (DU), gastric cancer (GC) and colorectal cancer (CRC) (Lanas and Chan, 2017;Qiu et al. 10.3389/fmicb.2023.1273269Frontiers in Microbiology 02 frontiersin.orgClarrett and Hachem, 2018;Narayanan et al., 2018;Seyedian et al., 2019;Patel and Shackelford, 2022).These diseases have a significant impact on the quality of life and health status of patients.Although some progress has been made in the past decades in treating and preventing these gastrointestinal diseases, their pathogenesis is still not fully understood, and the association with gut microbiota, in particular, has not been fully explained (Badillo and Francis, 2014;Lazaridis and Germanidis, 2018;El-Salhy et al., 2019;Rescigno, 2023).
Recently, gut microbiota as a complex microbial community has attracted extensive research interest.These microorganisms live in the human gut and are closely related to our health.It has been shown that gut microbiota is involved in various important physiological functions, including food digestion, immune regulation, and maintenance of the intestinal mucosal barrier (Rowland et al., 2018;Paone and Cani, 2020;Yang and Cong, 2021).Therefore, an in-depth study of the potential relationship between gut microbiota and gastrointestinal diseases is expected to shed light on the pathogenesis of the diseases and provide new ideas for future therapeutic and preventive strategies.
Inflammatory bowel disease (IBD) patients have been found to experience an increase in harmful bacteria, such as Enterobacteriaceae and Bartonellaceae, while beneficial bacteria, like thick-walled and butyrate-producing bacteria, decrease significantly within their intestinal flora (Wang et al., 2014;Schirmer et al., 2019;Lin et al., 2023).A study conducted by Halkjaer et al. showed that antibiotics treatment and fecal flora transplantation had a relieving effect on IBS symptoms, providing evidence for the direct connection between gut flora and IBS (Halkjaer et al., 2018;Fodor et al., 2019).Microbiome and metabolome examination of gastric biopsy tissues using histological techniques revealed a clear correlation between peptic ulcers and flora (Malik et al., 2023;Wang et al., 2023).In patients with GC, the enrichment of microorganisms from the genera Megasphaera, Moryella, and Vibro was observed, and these microorganisms were found to have diagnostic value in differentiating GC patients from healthy individuals (Zhang et al., 2021;Png et al., 2022).Clostridium nucleatum, Porphyromonas fragilis, and Escherichia coli showed a strong association with CRC, according to a study by Tilg et al. (2018).Additionally, Porphyromonas gingivalis and Porphyromonas solanacearum were found to induce butyrate-associated cellular senescence, promoting CRC (Okumura et al., 2021).Although randomized controlled trials are the gold standard for studying causality, they are difficult to implement and design due to constraints such as ethics, subject compliance, and study duration (Zoccali, 2017;Skrivankova et al., 2021).To address this issue, a new method called Mendelian randomization (MR) utilizes genetic tools to assess the causal relationship between exposure and outcome in epidemiological analysis (Lee and Lim, 2019;Richmond and Davey, 2022).By utilizing the random distribution of gametes from parents to offspring, MR studies allow reliable conclusions to be drawn about the impact of risk factors on outcomes unaffected by potential confounders (Li et al., 2023).
To provide more evidence of causality between gut microbiota and gastrointestinal diseases, this paper aims to provide insights into the potential relationship between gut microbiota and various gastrointestinal diseases using a bidirectional two-sample MR analysis.Through this study, we expect to provide new insights into the pathogenesis of gastrointestinal diseases and provide a scientific basis for disease prevention and treatment strategies, thus contributing to improving human health.

Study design and methods
All studies used in our study were based on some publicly summarized data and received ethics approval; all participants had provided informed consent.

Study design
An overview of the study design is shown in Figure 1.Our study is based on the three main hypotheses of the MR study (Davies et al., 2018).The three main hypotheses of MR studies: I: Instrumental variables (IVs) are related to exposure; II: IVs are unrelated to outcome; III: IVs are related to any known or unknown confounders that may mediate from exposure to outcome.

Data sources of gut microbiota
Single nucleotide polymorphisms (SNPs) related to the human gut microbiome composition were selected as IVs from a GWAS dataset of the international consortium MiBioGen (Kurilshikov et al., 2021). 1  This was a multi-ethnic large-scale GWAS that coordinated 16S ribosomal RNA gene sequencing profiles and genotyping data from 18,340 participants from 24 cohorts from the USA, Canada, Israel, South Korea, Germany, Denmark, the Netherlands, Belgium, Sweden, Finland, and the UK to explore the association between autosomal human genetic variants and the gut microbiome.211 taxa (131 genera, 35 families, 20 orders, 16 classes, and 9 phyla) were included.In our study, excluding unknown gut microbiota, we finally included 196 taxa (119 genera, 32 families, 20 orders, 16 classes, and 9 phyla).The gut microbiota GWAS data were adjusted for age, sex, study-specific covariates, and principal components derived from population stratification.

Data sources for gastrointestinal diseases
The pooled data for GERD came from the multi-trait genetic association analysis of GERD by Ong et al. (2022), including 129,080 European ancestry cases and 473,524 European ancestry controls.The pooled data for UC and CD comes from the report of Liu et al. (2015), in which UC included 6,968 cases and 20,464 controls, and CD included 5,956 mixed-ancestry cases and 14,927 controls.The pooled data for GC and CRC comes from the report of Liu et al. (2015), in which GC included 1,029 cases and 475,087 controls, and CRC included 6,581 cases and 463,421 controls.

Instrumental variables
The selection criteria for IVs were as follows: (1) SNPs associated with each genus at the genome-wide significance threshold (p < 1.0 × 10-5) were selected as potential IVs (Sanna et al., 2019); (2) linkage disequilibrium (LD) between SNPs was calculated using the 1,000 Genomes Project European Sample data as a reference panel with an R 2 < 0.001 (Lumped window size = 10,000 kb), only SNPs with the lowest p-value were retained; (3) SNPs with minor allele frequency (MAF) ≤ 0.01 were excluded; (4) when palindromic SNPs were present, the allele frequency information was used to infer the positive-stranded allele; and (5) To satisfy the strong association with exposure, we chose as SNPs with F-statistic values greater than 10.The formula for F is F = Beta 2 /SE 2 .

Statistical analysis
This study used several methods to examine whether a causal relationship exists between gut microbiota and gastrointestinal disorders, including inverse variance weighted (IVW), MR-Egger regression, weighted median, and weighted mode.The IVW approach uses meta-analysis combined with Wald estimates for each SNP to obtain an overall estimate of the impact of gut microbiota on gastrointestinal disorders.Without horizontal pleiotropy, IVW results will be unbiased (Burgess et al., 2016).Therefore, the IVW method  served as the primary method for our analyses.The MR-Egger regression assumes that the instrument strength is independent of the direct effect (InSIDE), which makes it possible to assess the presence of pleiotropy using the intercept term.If the intercept term is equal to zero, it indicates the absence of horizontal pleiotropy, and the results of the MR-Egger regression are consistent with IVW (Bowden et al., 2015).The weighted median approach allows correct causality estimation when up to 50% of the IVs are invalid (Burgess et al., 2016).
If the InSIDE assumption is violated, weighted model estimation is more effective in detecting causal effects than MR-Egger regression, with less bias and lower Type I error rates (Kurilshikov et al., 2021).

Results
In the causal estimation of gut microbes on common gastrointestinal diseases, we obtained a total of 14,587 SNPs that were strongly associated with 196 gut microbes according to the screening criteria (Supplementary Table 2).F-statistics for SNPs ranged from 14.6 to 87.3.
Cochrane's Q test did not show significant heterogeneity (p > 0.05, Table 2).The results of the susceptibility analysis of gut microbiota to CD are displayed in Supplementary Table 3.In addition, leave-one-out analyses showed that any single IV drove none of the identified causal associations.

Causal effects of gut microbiota on GERD
In GERD, causal correlations were found in only five gut microbiota.The higher genetically predictive Class Bacteroidia (OR: 1.11, 95% CI: 1.00-1.22,p = 0.049), Class Mollicutes (OR: 1.11, 95% CI: 1.01-1.26,p = 0.024), and Family Bacteroidaceae (OR: 1.21, 95% CI: 1.03-1.42,p = 0.018) were associated with the occurrence of GERD, whereas the Family Christensenellaceae (OR: 0.91, 95% CI: 0.84-0.99,p = 0.022) was associated with a reduced occurrence of GERD (Figure 2C and Table 2).According to the results of MR-Egger and MR-PRESSO tests (p > 0.05, Table 2), no horizontal pleiotropy and outliers were seen.The results of Cochrane's Q-test showed no significant heterogeneity (p > 0.05, Table 2).The results of the susceptibility analysis of gut microbiota to GERD are displayed in Supplementary Table 5.In addition, leave-one-out analyses showed that any single IV drove none of the identified causal associations.

Causal effects of gut microbiota on DU
Three microbiota genetically predicted to be associated with an increased risk of DU include Genus Eubacteriumeligensgroup (OR: 0.9966, 95% CI: 0.9935-0.9997,p = 0.030) and Genus Enterorhabdus (OR: 0.9965, 95% CI: 0.9938-0.9992,p = 0.010).Of these, the Genus Butyricimonas (OR: 1.0021, 95% CI: 1.0001-1.0041,p = 0.043) was associated with an increased risk of DU (Figure 3D and Table 2).The Genus Eubacteriumeligensgroup and Genus Enterorhabdus were associated with a decreased risk of DU (Figure 3D and Table 2).No significant heterogeneity or horizontal pleiotropy was found according to Cochrane's Q, MR-Egger, and MR-PRESSO tests (Table 2).The results of the susceptibility analysis of gut microbiota to DU are displayed in Supplementary Table 10.

Discussion
Previous research on the link between gut microbiota and gastrointestinal diseases mainly relied on population-based retrospective studies (Gevers et al., 2014;Liu et al., 2021;Salem et al., 2023).These studies typically collected fecal samples from individuals with gastrointestinal disorders and used cross-sectional metabolomics analyses for concluding (Li et al., 2014;Da Silva et al., 2018;Caruso et al., 2020).However, these approaches had limited capacity to establish causal relationships between gut microbiota and gastrointestinal disorders.In contrast, our study employed MR analyses and utilized extensive GWAS data to investigate potential causal connections between gut microbiota and gastrointestinal diseases.This large-scale comprehensive MR investigation represents a pioneering attempt to understand causal associations between gut microbiota and a wide range of prevalent gastrointestinal diseases, operating at the level of gene prediction.Consequently, our study's results possess robust causal explanatory power and provide valuable insights that could guide the targeted treatment of gastrointestinal diseases by identifying specific gut microbiota.
In our study, a total of 62 gut microbiota associated with common gastrointestinal disorders were identified.Among these microbiota, two had strong causal associations.Genus Oxalobacter was associated with a higher risk of CD (OR = 1.29, 95% CI: 1.13-1.48,p = 2.5 × 10-4), while Family Clostridiaceae1 was associated with a lower risk of IBS (OR = 0.9967, 95% CI: 0.9944-0.9991,p = 1.2 × 10-3).Previous studies have shown that Gram-negative bacilli Bacteroides, which are associated with acute exacerbations of CD, indicate the role of gut microbiota metabolites in the development and progression of gastrointestinal diseases.Therefore, this association can be attributed to several reasons.Firstly, intestinal flora produces trimethylamine oxide (TMAO) toxin, which triggers the release of inflammatory mediators and leads to gastrointestinal inflammation (Hosseinkhani et al., 2021).Clinical studies have shown that increased TMAO levels cause an increase in inflammation-associated monocytes that aggravate intestinal inflammation and compromise the intestinal barrier (Wang et al., 2023).Secondly, the gut microbiota affects immune cells and macrophages, leading to immune system activation and increased production of pro-inflammatory cytokines and chemokines.Intestinal flora disturbances may lead to the production of pathogenic immune cells on the surface of the intestinal epithelium or the homing of immune cells to extra-intestinal sites.In patients with IBD, the integrity of intercellular tight junctions is compromised in the intestinal mucosal tissues, disrupting the epithelial barrier and allowing pathogens to enter through the epithelial layer.These pathogens are recognized by pattern recognition receptors (PRRs) on the basolateral membrane of human intestinal epithelial cells (IECs).Consequently, human IECs block the secretion of retinoic acid and TGF-β, while the abundance of pro-inflammatory cytokines in the lamina propria masks the sedative signals secreted by human IECs.Macrophages identify captured antigens as invading pathogens, transforming them into pro-inflammatory phenotypes, thereby preventing immune tolerance and triggering an excessive inflammatory immune response (Kedia et al., 2019;Lee et al., 2022).Examples of such microorganisms include Porphyromonas gingivalis, Actinobacillus, and Chlamydia pneumoniae (Müller et al., 2006;Hills et al., 2019).It should be noted that the accumulation of toxins and the hyperactivation of immune cells can cause damage to organs outside the gastrointestinal system.
Furthermore, the gut flora also produces metabolites such as short-chain fatty acids (SCFAs), with butyrate being the most important one.Butyrate, mainly produced by commensal bacteria like Genus Clostridium, provides protective effects for the gastrointestinal tract (LeBlanc et al., 2017).In addition, SCFAs not only directly provide energy to IECs and maintain the integrity of the intestinal barrier, but also play an antiinflammatory role by participating in the regulation of the body's  significantly improve the lives of the mice.Given this, the protective effect of Family Clostridiaceae1 in patients with IBS can be explained.Therefore, reducing TMAO levels and increasing SCFAs levels in the body could be a potential target for future treatment of patients with gastrointestinal disorders.However, our present study did not confirm the underlying mechanism of gastrointestinal diseases induced by gut microbiota.Instead, our study aimed to explore the casual relationship between the two.Further research is needed to provide a detailed explanation of the mechanisms involved in gastrointestinal diseases.
One thing we need to be aware of in this study is the possibility of false negatives in the Bonferroni correction test.Our findings demonstrated a nominal causal association (q < p < 0.05) between some microbiota and gastrointestinal diseases, but this association was weak.This may be due to the complex association between gut microbiota and gastrointestinal diseases.At the same time gut microbiota are complex microbial communities.Therefore, the role of a single microbial community in developing a disease may be less significant when conducting research analyses leading to a single microbial community.In addition, conducting current research on gut microbiota and gastrointestinal diseases is still challenging.The diversity of gut microbiota is closely related to environmental, regional, and dietary factors, and the composition of the flora varies greatly among different populations.A study reported that the gut microbiota of East Asian populations differed significantly from that of other populations (Kedia et al., 2019).In the future, we need to expand the scope of the study to gain a further in-depth and comprehensive understanding of the occurrence and development of the relationship between gut microbiota and gastrointestinal diseases and to provide guidance for our further development of targeted polymicrobial drugs.
Our study also has some limitations, which should be noted when interpreting the results.Firstly, the data used in our MR analyses were pooled rather than raw, and therefore, subgroup analyses could not be performed to explore the presence of non-linear relationships further.Secondly, gut flora as an exposure phenotype is limitedly explained by genotype, which means that robust calculations of the statistical efficacy of MR analyses are overly stringent.Thirdly, the fact that the smallest category of gut microbiota is the genus prevents us from further exploring causal relationships between gut microbiota and a wide range of gastrointestinal diseases at the species level.Last but not least, since most subjects in the GWAS meta-analysis of gut microbiota data were of European origin, the results of this study may apply to non-European populations.

Conclusion
By performing an MR analysis of causal associations between 196 gut microbiota and eight phenotypes, we identified 62 nominal and two strong causal associations.Family Clostridiaceae1 was strongly associated with lower IBS, whereas Genus Oxalobacter was strongly associated with CD.Our study identified specific microbiota through gene prediction, which may provide useful biomarkers for early disease diagnosis and potential therapeutic targets for gastrointestinal diseases.
immune response through the activation of GPCR receptors.For example, butyric acid reduces TLR4 expression and inflammatory cytokine production in IECs, protecting intestinal health and mucosal integrity.Studies have shown that fecal transplants containing higher levels of SCFA or related bacteria can effectively alleviate intestinal inflammation in mice with IBD and

TABLE 1
Information of GWAS summary data.

TABLE 2
MR analysis between gut microbiota and multiple gastrointestinal diseases with horizontal pleiotropy and heterogeneity tests.95%CI:0.9964-0.9994,p=0.005)wereassociatedwitha reduced risk of IBS (Figure2Dand Table2).According to the results of MR-Egger and MR-PRESSO tests (p > 0.05, Table2), no horizontal pleiotropy and outliers were seen.According to the results of Cochrane's Q-test, we found no heterogeneity (p > 0.05, Table