Association Between TCF7L2Gene Polymorphisms And The Risk Of New-onset Diabetes Mellitus After Liver Transplantation

New-onset diabetes mellitus (NODM) is common in liver transplant recipients and has a negative effect on both graft and patient survival. Although the surgical technique and immunosuppressive management have improved, the prevalence of NODM remains very high, and has been recently reported as17-36.5%. The development of NODM after liver transplantation (LT) has shown to be closely associated with cardiovascular complications, infections, chronic rejection and renal failure, which subsequently lead to a reduced quality of life and high mortality. Identifying patients at high risk of developing NODM is beneficial for preventing disease and improving the long-term prognosis after LT.So far, no definitive risk factors have been clearly established, but clinical parameters such as advanced age, family history, obesity, hepatitis C virus infection, and immunosuppressive agents have been implicated. As NODM shares many similarities with type2diabetes mellitus in both pathophysiology and clinical presentation, genetic variants that are involved in type2diabetes mellitus may also be associated with the development of NODM besides the clinical implications. Recent genome-wide association studies have identified a number of novel type2diabetes-susceptibility genes, such as transcription factor7-like2(TCF7L2), which is regarded as the most important one. By isolating genomic DNA from peripheral blood lymphocytes of kidney transplant recipients, researchers have demonstrated that TCF7L2rs7903146polymorphism is significantly associated with an increased risk of NODM after kidney transplantation.However, unlike kidney transplantation, the liver plays a major role in blood glucose control and hepatic metabolism of insulin. It is well known there is tightly association between advanced liver disease and diabetes mellitus (DM). About20%of cirrhotic patients develop overt DM that requires either oral hypoglycemic agents or insulin administration. Therefore, donors’phenotype and genotype may greatly affect glucose handling and relate to the occurrence of NODM in liver transplant recipients. Increasing evidence has shown that there is a tight link between TCF7L2and hepatic glucose metabolism. Several studies have indicated that the T allele of rs7903146TCF7L2is associated with elevated rates of hepatic glucose production and reduced hepatic insulin sensitivity. In addition, Wnt/b-catenin/TCF signaling pathway plays a key role in driving metabolic zonation in the liver, and liver TCF7L2expression has been reported to strongly associate with diabetes.Part I Impact of new-onset diabetes mellitus on prognosis after liver transplantationAim:We aim to examine the impact of preexisting diabetes mellitus (DM) on post-transplant outcome in patients with hepatitisB virus (HBV)-related liver disease.Methods:The post-transplant morbidities and patient survival were compared between48diabetes patients (DM group) and96non-diabetes patients (Control group) matched for age, gander, primary disease and model for end-stage liver diseases score. DM group was further divided into hyperglycemia patients (fasting blood glucose>8.0mmol/L, n=22) and non-hyperglycemia patients (fasting blood glucose≤8.0mmol/L, n=26).Results:Before transplantation, a higher incidence of hepatic encephalopathy was found in DM group than that in Control group (22.9%vs.10.4%, P=0.045). The blood glucose level was significantly higher in DM group than that in Control group (8.0±0.9mmol/L vs.4.5±0.8mmol/L, P<0.001). The other pre-transplant data were comparable between DM group and Control group. After transplantation, the incidence of post-transplant DM (12/48vs.19/96), cardiac and vascular decease (7/48vs.9/96), sepsis (7/48vs.6/96), fungus infection (9/48vs.10/96), graft failure (1/48vs.0/96), acute kidney injury (11/48vs.17/96), chronic kidney diseases (5/48vs.7/96), acute rejection (4/48vs.6/96), biliary complications (12/48vs.12/96), HBV recurrence (5/48vs.10/96) and HCC recurrence (3/8vs.6/16) did not differ significantly between DM group and controls. Compared to their matched case controls (n=44), hyperglycemia patients showed remarkable higher incidences of sepsis (18.2%vs.2.3%, P=0.039) and biliary complications (31.8%vs.6.8%, P=0.012). The incidence of post-transplant complications did not differ significantly between non-hyperglycemia patients and case controls, or between Insulin-dependent patients and case controls, or between Insulin-independent patients and case controls. Patient cumulative survivals were not found significant difference between DM group and Control group. The1-year,2-year and3-year survival rates did not differ significantly between hyperglycemia patients and case controls, or between non-hyperglycemia patients and case controls, or between insulin-dependent patients and case controls, or between insulin-independent patients and case control.Conclusions:Preexisting DM is not a contraindication for liver transplantation in patients with HBV-related liver disease. A tight control of blood glucose to a level of≤8.0mmol/L was necessary to reduce the risk of complications after liver transplantation.Part II TCF7L2gene polymorphisms are associated with the risk for hepatogenous diabetes in the Chinese Han populationAim:Hepatogenous diabetes (HD) is common in cirrhotic patients but its etiology is still unclear. Genetic variants have been reported to increase the risk of type2diabetes. We aim to evaluate the role of the most significant diabetes susceptibility gene (TCF7L2) in HD in the Chinese Han population. Methods:A total of367adult liver transplant recipients with decompensate cirrhosis were included. Fifteen tag single nucleotide polymorphisms (SNPs) were selected from HapMap CHB database with a minor allele frequency (MAF) of>0.2and r2of>0.8. Another three SNPs were also chosen because of their association with type2diabetes in East Asian. HD was defined as diabetes after cirrhosis.Results:Among all367patients,100(27.2%) had HD,51(13.9%) had IGF, and233(58.9%) showed normal glucose levels. Compared to non-HD group, HD group presented significantly higher MELD score (P=0.022), bilirubin (P=0.008) and international normalized ratio (P=0.017), as well as higher incidence of hepatic encephalopathy (P=0.015) and bleeding esophageal varices (P=0.029). The HOMA-IR index was significantly higher in HD group than non-HD group (P<0.001). Blood glucose levels were significantly correlated with hepatic encephalopathy (r=0.192, P<0.001), MELD score (r=0.161, P=0.003), international normalized ratio (r=0.149, P=0.006) and bilirubin (r=0.134, P=0.013).All18SNPs frequencies were in accordance with Hardy-Weinberg equilibrium with P values of>0.05. Among the18SNPs, TCF7L2rs290487, rs6585194and rs7094463were found to distribute differentially between the two groups. The rs290487minor C-allele, rs6585194major C-allele and rs7094463major A-allele were more frequent in HD group than those in non-HD group. In univariate logistic regression analysis, four SNPs were significantly associated with HD. The genetic variants in TCF7L2rs290481and rs290487increased the risk of HD, with odds ratio ranging from1.38to1.57. In contrast, genetic variants in TCF7L2rs6585194and rs7094463significantly reduced the risk of HD, with odds ratio ranging from0.52to0.75. Multivariate logistic regression analysis revealed that rs290487and rs6585194polymorphisms were independent factors intlueneing the onset of HD. The effect of genotype remained significant after adjustmem of MELD score, bilirubin, international normalized ratio, hepatic encephalopathy and bleeding esophageal varices. According to the result of logistic regression analysis, we calculated risk score for each independent influencing factor. TCF7L2rs290487C/C homozygote presented a high probability of HD (45%). The combination of the two SNPs increased the predictive value. Insulin resistance (IR) was further compared among different genotypes using HOMA-IR index. TCF7L2rs290487SNP was tound to be associated with IR. Compared to the T/T genotype, both C/C (3.8vs.2.2. P=0.004) and T/C (6.3vs.2.2, P=0.001) genotypes showed significantly higher HOMA-IR indexes. The C/C genotype also presented higher HOMA-IR than the T/C genotype (P=0.004).Conclusions:Blood glucose level may depend on the severity of liver disease in cirrhotic patients, and HD could be a marker of end-stage liver disease. TCF7L2gene polymorphisms are associated with the risk for HD. These findings have uncovered the role of genetics in the pathogenesis of diabetes in cirrhotic patients.Part III Association between donor and recipient TCF7L2gene polymorphisms and the risk of new-onset diabetes mellitus after liver transplantation Aim:New-onset diabetes mellitus (NODM) is a frequent and serious complication after liver transplantation (LT). Transcription factor7-like2(TCF7L2) polymorphisms have been reported to strongly associate with type2diabetes. In addition, the donor liver plays a vital role in regulating blood glucose levels. In this study, we aim to evaluate the association between donor and recipient TCF7L2gene polymorphisms with NODM after LT.Methods:A total of125patients undergoing primary LT between November2006and July2009at the First Affiliated Hospital, Zhejiang University School of Medicine, China were enrolled. We excluded patients with a known history of diabetes, less than6-month follow-up time or developing acute rejection. The mean age of recipients at transplantation was47.8±10.5years (range:18-70years) and there were107males and18females. No patient had a family history of diabetes. According to the American Diabetes Association criteria, NODM was defined as a fasting glucose level of at least7mmol/L (126mg/dl), or a non-fasting glucose level of at least11.1mmol/L (200mg/dl) confirmed on at least2occasions or a need for antidiabetic drugs persisting beyond the first month after transplantation. All liver transplant recipients could be divided into NODM group and non-NODM group, respectively. Recipient genotype and donor genotype were determined separately. Within the TCF7L2gene, four single nucleotide polymorphisms (SNPs)(rs290487, rs7903146, rs11196205, and rs12255372) were selected, since they were reported to be significantly associated with type2diabetes in the Eastern Asia population. SNPs were detected using Applied Biosystems SNaPShot and TaqMan technology. Results:(1) Clinical characteristics of NODM patientsThe incidence of NODM was12%(15/125) and17.6%(22/125) at3and6months after LT, respectively. The overall incidence of NODM within the first year after LT was20%(25/125) in this study population. Recipient age (P=0.030) and donor fatty liver (hepatic steatosis>30%, P=0.039) were significantly different between the two groups. Other baseline patient characteristics, including donor parameters (age, gender, and cold ischemia time) and recipient parameters (gender, primary liver disease, co-morbidities, MELD score fasting plasma glucose, and BMI), did not differ significantly between the two groups. The1-and2-year survival rates were84%and76%,92%, and83%in the NODM group and the non-NODM group, respectively. The NODM group showed a lower survival rate but no significant difference was found. After transplantation, the fasting plasma glucose levels were significantly higher in the NODM group than the non-NODM group at both3and6months after LT (P<0.001). Blood tacrolimus levels were significantly higher in the NODM group than the non-NODM group at1(P=0.005) and3months (P=0.043) but not at6months after LT (P=0.511). Although BMI was higher in the NODM group than the non-NODM group at6months after LT, it did not differ significantly (P=0.051).(2) TCF7L2genotype distribution and association with NODMMinor allele frequency (MAF) of donor rs290487, rs7903146, rs11196205, and rs12255372was0.34,0.07,0.02, and0.01, respectively. MAF of recipient rs290487, rs7903146, rs11196205, and rs12255372was0.36,0.06,0.04, and0.03, respectively. Allele frequencies were all in Hardy-Weinberg equilibrium (P>0.05). Relative strong linkage disequilibrium was found between TCF7L2rs12255372and rs11196205(D’=0.905, r2=0.557), but not others (D’<0.8, r2<0.3). Among the four SNPs, rs290487 was found to be significantly differentially distributed between the two groups. The incidences of NODM were remarkably higher in patients with the donor rs290487C/C genotype than in those with the donor rs290487T/T genotype (50.0%vs.11.1%, P=0.004), and in patients with the recipient rs290487C/C genotype than in those with the recipient rs290487T/T genotype (41.2%vs.13.7%, P=0.043). Both donor and recipient rs290487polymorphisms (C/C vs. T/T genotype) were significantly associated with NODM. The incidence of NODM differed remarkably among different donor and recipient rs290487genotypes in an allele dependent manner, and was significantly higher in C allele count>3(donor-recipient:C/C-T/C, T/C-C/C, and C/C-C/C) than<3(donor-recipient:C/C-T/T, T/T-C/C, T/C-T/C, T/T-T/C, T/C-T/T, and T/T-T/T)(70%vs.15.7%, P<0.001).(3) Risk factors of NODM:Multivariate logistic regression analysisClinical parameters that were differentially distributed between the two groups, such as donor fatty liver, recipient age and blood tacrolimus levels, and also donor and recipient TCF7L2rs290487polymorphisms were considered to be potential influencing factors. BMI, hypertension, and hyperlipidemia were considered as potential risk factors and also entered into the logistic regression analysis. Appropriate cut-off levels were selected for their clinical significance. In univariate analysis, NODM was significantly associated with the following factors:recipient age>55years (odds ratio [OR]=2.667[1.074-6.621], P=0.035), donor fatty liver (OR=3.193[1.016-10.04], P=0.047), blood tacrolimus levels at1-month after LT>10ng/ml (OR=3.500[1.413-8.672], P=0.007), donor rs290487genetic variant (OR=2.749[1.387-5.450] per each C allele, p=0.004), and combined donor and recipient rs290487genetic variant (OR=2.627[1.437-4.801]per each C allele, P=0.002). Multivariate analysis showed that advanced recipient age, high blood tacrolimus levels at1month after LT, donor TCF7L2rs290487polymorphism (C allele), and combined donor and recipient TCF7L2rs290487 polymorphism (C allele) were independent risk factors of NODM after LT.(4) Predictive value of risk factors and models on NODMThe predictive abilities of risk factors on NODM were further evaluated according to the results of logistic regression analysis. Donor rs290487C/C genotype with or without recipient rs290487genetic variant presented a high probability (>40%) of onset of NODM. Different predictive models were established by different inclusion criteria. Model1only included clinical parameters, whereas Model2and3contained genetic risk factors. The combined donor and recipient rs290487polymorphism was only entered into Model3but not into Model2. The predictive abilities of the three models on NODM were evaluated by area under receiver operating characteristic curves (AUROCs). Both models containing genetic risks had relative good predictive abilities (AUROC>0.75) and good fits (p>0.8to reject model fit). By comparison of AUROCs, Model2presented significantly better predictive ability than Model1(0.780vs.0.712, P=0.031). But no significant difference was found between Model2and3(0.780vs.0.799, P=0.581). This means that donor rs290487polymorphism with or without the recipient rs290487genetic variant could largely improve the predictive ability of the clinical model.Conclusions:Here we provided the first evidence that the donor liver may contribute to metabolic disorders after LT at the genetic level. We demonstrate that donor TCF7L2rs290487polymorphism (C allele) is associated with an increased risk of NODM after LT and has a potential clinical value for the prediction of NODM. For patients with advanced age and donor liver TCF7L2rs290487genetic variant, blood tacrolimus concentrations should be strictly controlled at relative low levels in the early post-transplant period. The study also indicates that the donor may contribute to the disease-susceptibility genotype and consequently cause poor prognosis after LT. More attention, therefore, should be paid to the donor phenotype and genotype with regard to the outcome after LT.