Construction And Evaluation Of Vascular Complications Of Type 2 Diabetes Mellitus Auxiliary Diagnosis Model Based On Salivary Glycopatterns

Background: Type 2 Diabetes Mellitus(T2DM)is the group of metabolic chronic diseases with multiple complications.However,it is estimated that more than half of T2 DM patients are not aware of the fact that they already have this disease,so they are more likely to have T2 DM complications.And vascular complications are the main cause of increased mortality of T2 DM.30% of T2 DM patients already have some degree of chronic vascular complications at the time of diagnosis.At present,the treatment of diabetes is to control blood glucose,and there is no radical treatment,so the pre-diagnosis of diabetes and its complications is particularly important.However,there is no ideal in vitro marker for early diagnosis of diabetic complications.A large number of studies have shown that the serum protein glycosylation level changes abnormally with the occurrence and development of T2 DM.In this study,saliva samples from healthy volunteers(HV),T2 DM patients without complications(T2DM-)and T2 DM patients with micro-complications or/and macro-complications(T2DM+)were used to comparative analysis in clinical patients and normal control samples using lectin microarrays that is high sensitivity and high throughput sugar chain detection technology.It is expected to find abnormal salivary protein glycochain structure and explore the possibility of its application as a detection index for early diagnosis of T2 DM and T2DM+ patients.Methods:In this study,salivary glycoproteins of HV(78 cases),T2DM-(72 cases)and T2 DM + patients(80 cases)were detected and analyzed by lectin microarray technology,and then select abnormal expression of glycoprotein of saliva samples and lectins identify the sugar chain structure in three groups,and the lectin imprinting experiment verifies the results.Then,230 samples of lectin microarray normalized data were randomly divided into training set and validation set.The training set data were used to construct the T2 DM diagnostic model(Model T2DM)and T2DM+ diagnostic model(Model T2DM+)respectively by binary stepwise Logistic regression analysis.It is used to differentiate T2 DM and T2DM+ patients.Finally,the receiver operating characteristic curve was used to verify the differential diagnosis ability of the above two disease diagnosis models in the validation set data..Results:(1)The results of lectin microarray in 230 saliva samples showed that there were 4 lectins(Jacalin,GSL-II,UEA-I and WFA)specifically recognized Galβ1-3Gal NAc α-Ser/Thr(T),Glc NAc and agalactosylated Tri/Tetra antennary glycan,Fucα1-2Galβ 1-4Glc NAc,and terminal Gal NAcα/β1-3/6Gal structure of polysaccharides changed differently in T2DM+ group.Compared with T2DM+ group,there’re eight kinds of lectin in HV and T2DM+ group(BS-I,PNA and SJA,etc.)specific recognition of α-Gal NAc,Gal NAcα-Ser/Thr(Tn),Galβ1-3Gal NAcα-Ser/Thr(T),α-D–Man,Siaα2-6Galβ1-4Glc(NAc),etc.glycan chain structures was different,but there was no significant difference between HV and T2DM-groups.Compared with HV group,the structure of Glc NAc,Galβ 1-4Glc NAc,Galα1-3(Fucα1-2)Gal,Siaα2-3Galβ1-4Glc(NAc)/Glc,etc.glycan specifically recognized by 9 lectins(STL,DSA and EEL,etc.)changed significantly in T2DM-and T2DM+ groups.As the disease progresses,a total of 7 lectins(PSA,Con A,LTL,PTL-I,ACA,MPL,PWM)identified the High Mannose,Galβ1-3Gal Nac,Galβ1-3Gal NAc α-Ser/Thr(T),Fucoseα-1,6Glc NAc(core fucose),Fucα1-2Galβ1-4Glc NAC and other glycans were different in HV,T2DM-and T2DM+ groups.In the experiment of lectin imprinting(WFA,MAl-II,Con A and PHA-E are randomly selected),the gray values of lectin WFA and Mal-II bands with molecular weight between 20-25 k Da were significantly lower in T2DM+ group than in HV and T2DM-group.The gray values of lectin Con A and PHA-E bands with molecular weights between 50-70 k Da in HV group were significantly lower than those in T2DM-and T2DM+ groups.The results of lectin imprinting experiment were consistent with those of the lectin microarray.Therefore,the results of lectin imprinting experiment proved the reliability of microarrays data.(2)An auxiliary diagnostic model of disease based on salivary glycoprotein glycochain structure Model T2DM(AUC : 0.997,sensitivity :0.990,specificity : 0.980),Model T2DM+(AUC :0.897,sensitivity : 0.804,specificity : 0.833)were constructed.The model T2 DM could accurately identify 26 out of 27 HV cases and 51 out of 53 T2DM-cases,with AUC of 0.983,sensitivity of 0.962,specificity of 0.926 and accuracy of 96.25%.The model T2DM+ could accurately identify 16 out of 24 cases of T2DM-and 20 out of 29 cases of T2DM+ with AUC of 0.810,sensitivity of 0.690,specificity of 0.750 and accuracy of 71.70%.The AUC values of the area under the ROC curve of the two validation sets were all greater than 0.80,indicating that the two model have high discrimination efficiency.ConclusionThe above experimental conclusions indicate that compared with HV,salivary glycoprotein glycoprotein structure is abnormally expressed in T2 DM and T2DM+ patients,and for T2 DM and T2DM+ patients,Model T2 DM and Model T2DM+ constructed based on salivary glycoprotein structure have good discrimination ability.Therefore,the difference of glycosylation in saliva of T2 DM patients can be used for the early diagnosis of T2 DM and T2DM+.