A Trans-cohort Gut Microbiota Signature For Coronary Artery Disease In Patients With Type 2 Diabetes

Background According to the latest statistics from the International Diabetes Federation（IDF）,there are 537 million adults worldwide with Diabetes,more than 90 percent of whom have Type 2 Diabetes Mellitus（T2DM）.The complications of T2 DM have a great impact on the health,quality of life and life span of patients,among which coronary artery disease（CAD）and other large vessel complications are the main causes of disability and death in T2 DM patients.Patients with T2 DM have atypical CAD symptoms,rapid progression,and high risk of acute cardiovascular events.Therefore,early diagnosis of CAD is of great significance for T2 DM patients.Invasive coronary angiography（ICA）is the gold standard for the diagnosis of coronary heart disease,but it cannot be used as a common screening method because of its invasiveness.In addition,there are many non-invasive tests available,such as electrocardiogram,echocardiography,cardiac radionuclide imaging,coronary CT angiography（CTA）,etc,but their accuracy and applicability are limited.Therefore,a non-invasive CAD diagnostic model is particularly important for T2 DM patients.Gut microbiota is one of the most complex microbial communities and has formed a close symbiotic relationship with the host during the long process of "coevolution".With the deepening of research in recent years,gut microbiota as the "second genome",has received more and more attention for its role in maintaining human health homeostasis,especially in the occurrence and development of metabolic diseases such as T2 DM and CAD.Studies have shown that Firmicutes,Faecalibacterium prausnitzii（F.Prausnitzii）and Bifidobacterium have decreased abundance in T2 DM patients.The abundance of etaproteobacteria,Clostridiales and Lactobacillus gasseri increased,and the increase of Firmicutes-Bacteroidetes（F/B）ratio was significantly correlated with abnormal glucose metabolism.Gut microbiota of CAD patients showed that the abundance of F.Prausnitzii decreased,and the F/B ratio increased in CAD patients,which was consistent with T2 DM patients.Some studies compared the changes of gut microbiota structure between CAD patients with T2 DM and CAD patients with T2 DM.The results showed that there were differences in the composition and diversity of gut microbiota between the two groups.Compared with CAD patients with T2 DM,the abundance of Bacteroidetes in CAD patients with T2 DM decreased.The abundance of Firmicutes and Proteobacteria increased.In patients with CAD combined with T2 DM,the production of beneficial bacteria such as F.Prausnitzii and Bacteroides fragilis decreased,and the production of opportunistic pathogens such as Enterobacteriaceae,Streptococcus and Desulfovibrio increased.However,the focus of existing studies is to observe the changes of gut microbiota structure in T2 DM patients and CAD patients.The changes of gut microbiota in T2 DM patients after CAD have not been reported,and there is currently no non-invasive method to better diagnose CAD patients in T2 DM.To solve the above problems,we analyzed the gut microbiota characteristics of CAD in T2 DM patients with T2 DM alone by metagomic sequencing,and constructed a gut microbiota signature for CAD in T2 DM patients,which was verified in internally and externally in two independent cohorts.Objective1.To investigate the structural changes of gut microbiota and serum metabolites in T2 DM patients with CAD.2.To construct gut microbiota signature for CAD in T2 DM patients.Method Subjects recruit T2 DM patients undergoing invasive angiography（ICA）due to suspected CAD in Henan Provincial People’s Hospital from January 2020 to August 2021 were recruited.CAD was defined when ICA results indicated more than 50% stenosis in at least one coronary vessel.Combined with ICA results,the patients were divided into T2 DM with CAD group（T2DM+CAD）and T2 DM group（T2DM）.To increase the sample size of the T2 DM group,we included T2 DM patients with coronary CTA negative.A total of 66 subjects in the T2DM+CAD group and 33 subjects in the T2 DM group were recruited as the discovery cohort.In addition,we recruited 21 T2DM+CAD patients and 14T2 DM patients as an independent validation cohort.The inclusion and exclusion criteria for the independent validation cohort are the same as those for the discovery queue.All participants signed written informed consent.2 Clinical data Collection Data of all subjects were collected,divided into basic information and clinical data.Basic information includes:Name,age,sex,height,weight,body mass index（BMI）,systolic blood pressure（SBP）,diastolic blood pressure,（DBP）,current smoking history,alcohol consumption history,diabetes mellitus（DM）course,medication history,and the nature of chest pain were recorded for pre-test probability（PTP）calculation.Clinical data included:（1）fasting blood-glucose（FBG）,hemoglobin A1c（Hb A1c）;（2）Myocardial enzyme profile: creatine kinase（CK）,creatine kinase isoenzyme（CKMB）,cardiac troponin I（c Tn I）;（3）Liver function: alanine aminotransferase（ALT）,aspartate aminotransferase（AST）;（4）Lipid metabolism: cholesterol total（TC）,triglyceride（TG）,high density lipoprotein-cholesterol（HDL-C）,low-density lipoprotein-cholesterol（LDL-C）;（5）Renal function: creatinine（CREA）,uric acid（UA）,estimate glomerular filtration rate（e GFR）;（6）ICA and coronary CTA results.Feces samples were collected from all subjects for metagenomic sequencing and blood samples were collected for untargeted metabolomics analysis.3 Statistical analysis SPSS 23.0 software was used for statistical analysis.In clinical characteristics,medication status and laboratory test data,count data were expressed as frequency and percentage,and the difference of qualitative variables between the two groups was tested by x2 test.The measurement data were expressed as x ± s.Shapiro-wilk tests for normality.Mann-whitney U test or independent sample t test was used to test the difference of quantitative variables between the two groups.Result1 Comparison of general clinical data1.1 Comparison of clinical data in the discovery cohort A total of 66 patients with T2DM+CAD and 33 patients with T2 DM were recruited as a discovery cohort.Current smoking rate（28.8 vs 9.1%）,use rate of antiplatelet drugs（63.6 vs 30.0%）,HDL-C（1.01 vs 1.12mmol/L）,CREA（79.17 vs 58.68umol/L）and e GFR（79.53 vs 92.32 m L/（min*1.73m2））between T2DM+CAD and T2 DM groups had statistical significance（P<0.05）.Age,sex ratio,SBP,DBP,BMI,DM course,drinking history,PTP,FBG,Hb A1 c,myocardial enzyme profile（CK,CKMB,c Tn I）,liver function（ALT,AST）,UA,TC,TG and LDL-C had no statistical significance（P>0.05）.1.2 Comparison of clinical data in the independent validation cohort A total of 21 T2DM+CAD patients and 14 T2 DM patients were recruited as part of the independent validation cohort.In internal validation cohort SBP（147.38 vs133.71 mm Hg）,DBP（88.05 vs 77.79 mm Hg）,statin use rate（76.2 vs 35.7%）、PTP（58.91 vs 58.91%）and e GFR（75.34 vs 99.94 m L /（min*1.73m2））between T2DM+CAD group and T2 DM group had statistical significance（P<0.05）.There were no significant differences in age,sex ratio,BMI,current smoking rate,drinking history,DM course,FBG,Hb A1 c,liver function（ALT,AST）,CREA,UA,lipid metabolism（TC,TG,HDL-C,LDL-C）（P>0.05）.2 Gut microbiota characteristics of T2 DM patients after CAD .1 Difference analysis of gut microbiota composition in T2 DM patients after CAD Compared with T2 DM group,α-diversity in T2DM+CAD group was decreased,but the difference was not significant（P>0.05）.Gut microbiota structure of T2DM+CAD group and T2 DM group was significantly separated at the level of PCo A1（P<0.05）.At the genus level,the abundance of Bacteroides and Faecalibacterium decreased in T2DM+CAD group,while the abundance of Prevotella and Alistipes increased.At species level,Prevotella Copri abundance increased significantly after CAD in T2 DM.In T2DM+CAD group Prevotella copri 、 Alistipes_putredinis 、 Bacteroides_stercoris 、Bacteroides_caccae 、 Odoribacter_splanchnicus 、 Barnesiella_intestinihominis 、Alistipes_finegoldii、Bacteroides_faecis、Vicia_cryptic_virus was significantly enriched（P<0.05）,Escherichia_unclassified was significantly enriched in T2 DM group（P<0.05）.2.2 Functional enrichment analysis of gut microbiota in T2 DM patients with CAD Through the functional enrichment analysis of gut microbiota in the T2DM+CAD group and the T2 DM group,we found a total of 415 pathways,32 of which were different between the T2DM+CAD group and the T2 DM group.Ribosome,amino sugar,and nucleotide sugar metabolism pathways were enriched in T2DM+CAD group,pyruvate metabolism,porphyrin metabolism and fatty acid biosynthesis pathways were enriched in T2 DM group.3 Characteristics of serum metabolites in patients with T2 DM after CAD There was significant difference in serum metabolite composition between T2DM+CAD group and T2 DM group（P<0.001）.The abundance of 147 metabolites in serum was quantified.Fexofenadine,medinolol,panaxytriol,ancymidol and beta,beta-Dimethylacrylshikonin were enriched in T2DM+CAD group,while2-methylthiobenzothiazole was enriched in T2 DM group.4 Correlation analysis between gut microbiota and metabolites he correlation between intestinal flora and serum metabolites was observed.The abundance of Tomato_yellow_leaf_curl_China_betasatellite was associated with the decrease of hydroxyphenylbutazone level and the increase of L-proline level.The aminophenol level was positively correlated with the abundance of Fusobacterium_peridonticum and negatively correlated with the abundance of Alistipes_putredinis.Prevotella_copri abundance was positively correlated with fexofenadine levels.The sphingosine level was negatively correlated with the abundance of Bacteroides_caccae and Alistipes_putredinis and positively correlated with the abundance of Fusobacterium_periodonticum.There was a positive correlation between the abundance of Lachnospiraceae_bacterium₁₁₅7FAA and the level of lutein in iris wild.5.1 Construction of gut microbiota signature Random forest combined with logistic regression method was used to construct a diagnostic model for CAD in T2 DM patients.Finally,16 species were selected as biomarkers to establish the diagnostic model.Including Lachnospiraceae_bacterium₁₁₅7FAA 、 Prevotella_copri 、 Bacteroides_caccae 、Bacteroides_stercoris 、 Oscillibacter_unclassified 、 Odoribacter_splanchnicus 、Vicia_cryptic_virus 、 Holdemania_unclassified 、 Klebsiella_unclassified 、Clostridium_citroniae 、 Bilophila_unclassified 、 Rothia_mucilaginosa 、Dasheen_mosaic_virus 、 Bacteroides_eggerthii 、 Escherichia_unclassified 、Burkholderiales_bacterium₁₁₄7.The most important species were Lachnospiraceae_bacterium₁₁₅7FAA,Prevotella_copri,Bacteroides_caccae and Bacteroides_stercoris.The relatively enriched strains in T2DM+CAD group included Prevotella_copri、Bacteroides_stercoris、Bacteroides_caccae、Odoribacter_splanchnicus、Lachnospiraceae_bacterium₁₁₅7FAA.The relatively enriched strains in T2 DM group included Bacteroides_eggerthii 、 Clostridium_citroniae 、 Rothia_mucilaginosa 、Holdemania_unclassified.The signature containing these 16 species could accurately iagnose CAD patients with T2DM（AUC 0.921）,and the diagnostic accuracy was better than that of PTP（AUC 0.620）.5.2 Verification of gut microbiota signature The gut microbiota signature containing these 16 strains can accurately diagnose CAD patients with T2 DM in the independent validation cohort（AUC 0.932）,and the diagnostic accuracy was better than that of PTP（AUC 0.794）.In Jie Z’s study cohort,68 CAD patients with T2 DM and 9 patients with FBG>6.1mmol /L or Hb A1 c > 6.5% were selected as external validation cohort.It was found that the gut microbiota signature constructed in our study,including 16 species,still had good diagnostic accuracy（AUC 0.912）.Conclusion1.In T2 DM patients with CAD,the gut microbiota structure changed,and the abundance of Prevotella_copri and some other bacterial species changed.2.The gut microbiota diagnostic model containing 16 key species can accurately diagnose CAD in T2 DM patients.