Plasma Metal Concentrations,serum Complement C3,C4 Levels,and Their Associations With The Risk Of Incident Type 2 Diabetes

Metals and metalloids（henceforth referred to as “metals”）may enter and accumulate in the human body through food,drinking water,skin,and ambient air,and produce various health effects.The immunotoxic effects caused by metals vary according to the type and dose of the metals.Evidence suggested that some metals were critically involved in both the immune sensing of pathogens and the host defense against invading pathogens.However,chronic exposure to some metals often elicits immunotoxic effects.The immunotoxic effects of co-exposure to multiple metals remain to be investigated.Metals could affect the progression of diseases.Studies about the associations between metals exposure and the risk of type 2 diabetes mellitus（T2DM）were increasing in recent years,but the underlying mechanism is not clear yet.The complement system is a highly regulated,complex,and multifunctional network of proteins that are actively involved in the pathogenesis of inflammatory and immune diseases.It’s reported that the pathways which net these proteins might be induced by some metals.Nevertheless,epidemiological association studies on multiple metals and complement levels are still lacking.Meanwhile,the complement system is involved in the pathogenesis of a variety of inflammatory and immune diseases.Mechanism research indicated that complement was an integral regulator of the immune and inflammatory homeostasis,and played an essential role in the pathogenesis of T2DM.Current evidence on the prospective association of complement levels with T2DM is sparse,inconclusive,and mostly limited to Europeans.The third and fourth components of complement（C3,C4）are the most abundant and vital complement proteins,and showed a high level of genetic polymorphism.The ratio of C3 and C4 levels（C3/C4 ratio）is also assessed as a potential measurement to predict several diseases.Thus,investigating the exposure of multiple metals,complement levels,and their associations with the risk of incident T2DM,exploring the potential mediating role of complement in the associations between metals and the risk of T2DM,and investigating the joint effect of metals and complement proteins in relation to the risk of T2DM might further our understanding of the immunological mechanisms of metals,as well as prevention and intervention of related diseases.Based on the study background and the Dongfeng-Tongji（DFTJ）cohort,the aims of this dissertation were as follows: Firstly,we investigated the associations of 17 plasma metals concentrations with serum complement C3,C4 levels and C3/C4 ratio,and examined whether complement related genetic risk score would modulate the associations of metal exposure on serum C3 and C4 levels by performing gene-metal interaction analyses.Secondly,we further investigated the prospective associations of baseline serum C3,C4 levels and C3/C4 ratio with the risk of incident T2DM.Thirdly,we examined the associations between metals and the risk of incident T2DM,explored the potential mediating role of complement in the metal-T2DM associations,and investigated the joint effect of metal exposure and complement in relation the risk of incident T2DM.This study consists of three parts:Part 1 Associations between plasma metal concentrations and serum complement C3,C4 levelsObjectives: To investigate the association between concentrations of 17 plasma metals and serum C3,C4 levels as well as C3/C4 ratio,and further explore the possible interaction between metal exposure and C3 or C4-associated genetic risk score.Methods: We conducted a cross-sectional study with a total of 2977 participants who were available for both sufficient serum and genotyping data from the baseline survey of the DFTJ cohort.We measured concentrations of 23 plasma metals using the inductively coupled plasma mass spectrometry（ICP-MS）,and measured baseline serum C3 and C4 levels by immunoturbidimetry.After excluding the metals with a detection rate below 50% and those in which plasma metals were reported not to be reliable biomarkers of internal exposure,17 plasma metals were included in the present analysis.Both metal and complement levels were naturally log-transformed before analysis,and metals were further standardized to z scores to increase the comparability of estimates.Generalized linear regression models were fitted to evaluate the associations of 17 metals with C3,C4 levels,and C3/C4 ratio,with adjustment for age,sex,smoking status,drinking status,BMI,education,hypertension,diabetes,hyperlipidemia,estimated glomerular filtration rate （eGFR）,and future disease status of CHD.False discovery rate（FDR）adjustment was performed.We conducted the above analyses among participants free of CHD during the follow-up for sensitivity analysis.We further modeled restricted cubic splines to visualize the dose-response relations between the concentrations of metals and C3,C4 levels.The least absolute shrinkage and selection operator（LASSO）was used to select the most C3/C4/C3C4ratio-relevant metals.We applied Bayesian kernel machine regression（BKMR）to estimate the potential joint associations between metals co-exposure on serum C3,C4 levels and C3/C4 ratio.We calculated C3 and C4-associated weighted genetic risk scores（GRSs）using established single nucleotide polymorphisms（SNPs）from studies among Asians.We selected two C3-related SNPs,including rs3753394 in KCNT2-CFH and rs3745567 in C3.We selected nine C4-related SNPs,including rs1052693 in GTF2H4,rs11575839 in NCR3,rs2075799 in HSPA1 L,rs2857009 in TNXB,rs8283 in ATF6 B,rs2071278 in NOTCH4,rs3763317 in BTNL2-HLA-DRA,rs9276606 in HLA-DQB2-HLADOB,and rs241428 in TAP2.We further explored the potential gene-metal interactions via the inclusion of an interaction term of metal concentrations and complement-related GRSs.Results: For the single-metal models,after multivariable adjustment,each increment of one standard deviation in natural log-transformed concentrations of plasma copper was associated with a 0.549 higher level of natural log-transformed serum C3 [β（95% CI）=0.549（0.489,0.608）,FDR<0.0001],a 1.146 higher level of natural log-transformed serum C4 [β（95% CI）=1.146（0.999,1.294）,FDR<0.0001],and a 4.464 lower level of C3/C4 ratio [β（95% CI）=-4.464（-6.104,-2.825）,FDR<0.0001],respectively.Meanwhile,each increment of one standard deviation in natural log-transformed concentrations of plasma zinc was associated with a 0.083 higher level of natural log-transformed serum C3 levels [β（95% CI）=0.083（0.024,0.143）,FDR=0.049] and 0.007 higher level of natural log-transformed serum C4 [β（95% CI）=0.007（-0.138,0.152）,FDR=0.935],respectively.Each increment of one-standard deviation of natural log-transformed concentrations of plasma titanium was associated with a 3.356 increased level of C3/C4 ratio [β（95% CI）=3.356（1.786,4.927）,FDR<0.0001].Plasma copper concentrations remained significantly associated with serum C3,C4 levels and C3/C4 ratio in the sensitivity analyses.The restricted cubic splines demonstrated significant non-linear relations between copper and C3(P_non-linear=0.035),and copper and C4(P_non-linear=0.008).We also found a significant linear relation between zinc and C3(P_non-linear=0.052).Copper was identified as the most important factor of serum C3,C4 levels,and C3/C4 ratio among the 17 metals by the LASSO model.The BKMR analysis showed that metals co-exposure was significantly associated with both serum C3 and C4 levels,but there were no pairwise interactions between metals in relation to serum C3,C4 levels as well as C3/C4 ratio.Participants with higher GRS had higher C3 and C4 levels.Furthermore,we found a significant interaction between plasma arsenic concentrations and C3-related GRS in relation to C3 levels（Pinteraction=0.010）.Conclusions: Concentrations of some plasma metals were significantly associated with serum C3,C4 levels,and C3/C4 ratio,and copper was identified as the strongest factor.C3 genetic predisposition would modify the association of plasma arsenic concentration with serum C3 levels.Part 2 Associations of serum C3,C4 levels with the risk of incident T2DMObjectives: To examine the associations of serum C3,C4 levels,and C3/C4 ratio with the risk of incident T2DM.Methods: We conducted a prospective study with a total of 2662 participants who were available for sufficient serum and free of diabetes at baseline from the CHD nested-case control study.We determined baseline serum C3 and C4 levels by immunoturbidimetry.The diagnosis of T2DM was defined by meeting any of the following criteria: 1)self-report of a physician’s diagnosis of T2DM,2)fasting blood glucose levels of ≥7.0 mmol/L,3)glycated hemoglobin（Hb A1c）level ≥6.5%,and 4)usage of insulin,oral hypoglycemic agent or other diabetes medication.Incident T2DM cases were diagnosed during the 10 years of followup.Cox regression models were employed to examine the incidence of T2DM in relation to C3,C4 levels,and C3/C4 ratio,stratified by age at risk（in 5-year intervals）,and adjusted for sex,BMI,C4 or C3,eGFR,smoking status,drinking status,family history of diabetes,hypertension,hyperlipidemia,education level,physical activity,dietary intake habits,serum CRP levels,and future disease status of CHD.We modeled restricted cubic splines to visualize the dose-response relation of the C3 and C4 levels with the risk of incident T2DM.Results: Overall,711 participants developed T2DM during 23 067 person-years of followup.Higher serum C3 was significantly associated with a higher risk of incident T2DM after multivariable adjustment（HR [95% CI]=1.16 [1.05,1.28];per SD higher）.Higher serum C4 was significantly associated with a lower risk of incident T2DM（HR [95% CI]=0.89 [0.81,0.98];per SD higher）.The association between C3/C4 ratio and incident T2DM was marginally significant（HR [95% CI]=1.057 [0.999,1.119];per SD higher）.The restricted cubic splines demonstrated significant positive relation for C3 and incident T2DM(P_overall=0.003,P_non-linear=0.270),and negative relation for C4 and incident T2DM(P_overall=0.039,P_non-linear=0.096).Conclusion: The present study revealed a positive dose-response relation between serum C3 levels and incident T2DM.Serum C4 levels and C3/C4 ratio were negatively and positively associated with T2DM,respectively.Part 3 The mediating role of complement C3 and C4 in the associations between metals and the risk of incident T2DM,and the joint associations of complement and metals with T2DMObjectives: To explore the potential mediating role of complement C3 and C4 in the associations between metals and T2DM,and investigate the joint effect of metal exposure and complement in relation to the risk of incident T2DM.Methods: The study population was the same as the participants in Part 2.LASSO regression models were conducted to investigate the association between metals coexposure and incident T2DM.The optimal value of λ that gives the minimum mean squared error was determined,at which the metals with nonzero coefficients were selected.Cox proportional hazard regression was employed to examine the incidence of T2DM in relation to the selected metals.We further conducted the above analyses among participants free of CHD during the follow-up for sensitivity analysis.The weighted quantile sum regression（WQS）regressions were used to analyze the associations between co-exposure of the selected metals and incident T2DM.We further calculate the WQS index to reflect coexposure by the weights of the metals in the WQS regression.Moreover,mediation analysis was applied to assess the potential mediating role of complement on the associations of metals exposure with T2DM.Furthermore,the joint effect of complement and metal exposure in relation to the risk of T2DM was investigated.Results: Six metals including antimony,copper,barium,molybdenum,nickel,and selenium were identified as the significant factors of T2DM in the LASSO regression model.After multi-metal adjustment in the Cox regression,plasma antimony concentrations were positively associated with incident T2 Dm,plasma nickel and selenium concentrations were negatively associated with incident T2DM.An increment of one standard deviation in natural log-transformed exposure concentrations of plasma antimony was associated with a higher risk of T2DM(HR [95% CI]=3.58 [1.80,7.10].An increment of one standard deviation in natural log-transformed exposure concentrations of plasma nickel and selenium was associated with a lower risk of T2DM（HR [95% CI]=0.22 [0.09,0.53],HR [95% CI]=0.24 [0.10,0.57],respectively）.Concentrations of plasma antimony,nickel,and selenium remained significantly associated with the risk of incident T2DM in the sensitivity analyses.We observed significant negative association between co-exposure of the abovementioned 6 metals and the risk of incident T2DM.There were no significant mediations by complement C3 or C4 in the associations of single metal or metals co-exposure with incident T2DM.Furthermore,we found significant interactions between C3 and nickel,C3 and selenium,as well as C4 and nickel in relation to the risk of incident T2DM（Pinteraction=0.023,0.044,and 0.010,respectively）.Compared with those with serum C3 levels ≥1.36 g/L and plasma nickel concentrations in the first quartile,participants with C3 levels <1.36 g/L and plasma nickel concentrations in the fourth quartile had a 52% decreased risk of incident T2DM（HR [95% CI]=0.48 [0.34,0.67]）.Compared with those with serum C3 levels ≥1.36 g/L and plasma selenium concentrations in the first quartile,participants with C3 levels <1.36 g/L and selenium concentrations in the fourth quartile had a 47% decreased risk of incident T2DM（HR [95% CI]=0.53 [0.37,0.74]）.Compared with those with serum C4 levels <0.31 g/L and plasma nickel concentrations in the first quartile,participants with C4 levels ≥0.31 g/L and nickel concentrations in the fourth quartile had a 46% decreased risk of incident T2DM（HR [95% CI]=0.54 [0.40,0.73]）.Conclusion: Plasma antimony concentrations were positively associated with incident T2DM.Plasma nickel and selenium concentrations were negatively associated with incident T2DM.No significant mediations of C3 and C4 was observed in the associations of plasma metals with incident T2DM.There were joint associations between C3 and nickel,C3 and selenium,as well as C4 and nickel with the risk of incident T2DM.