Elevated serum complement factors 3 and 4 are strong inflammatory markers of the metabolic syndrome development: a longitudinal cohort study

An epidemiological design, consisting of cross-sectional (n = 2376) and cohort (n = 976) studies, was adopted to investigate the association between complement factors 3 (C3) and 4, and the metabolic syndrome (MetS) development. In the cross-sectional study, the C3 and C4 concentrations in the MetS group were higher than those in the non-MetS group (all P < 0.001), and the levels of immune globulin M (IgM), IgA, IgE, and IgG exhibited no significant differences between MetS and non-MetS (all P > 0.050). After multi-factor adjustment, the odds ratios (ORs) in the highest quartile of C3 and C4 concentrations were 7.047 (4.664, 10.648) and 1.961 (1.349, 2.849), respectively, both Ptrend < 0.050. After a 4 years follow-up, total 166 subjects were diagnosed with MetS, and the complement baseline levels from 2009 were used to predict the MetS risk in 2013. In the adjusted model, the relative risks (RRs) in the highest quartile of C3 and C4 levels were 4.779 (2.854, 8.003) and 2.590 (1.567, 4.280), respectively, both Ptrend < 0.001. Activation of complement factors may be an important part of inflammatory processes, and our results indicated that the elevated C3 and C4 levels were independent risk factors for MetS development.

Moreover, insulin resistance (IR) 16 , 17 , obesity 18 , and hypercholesterolemia 19 as components of MetS are usually independently interrelated to an increase in inflammatory markers. Given these findings, low-grade inflammation exerts significant effects on the development of metabolic dysfunction.
The complement system response to inflammation and infection is significant in innate and adaptive immune mechanisms. C3 and C4, mainly produced by the liver 20 , are the major plasma proteins of the immune system complement pathways 21 . Common genetic variants at the C3 locus have been suggested to be associated with risk of MetS and its components 22 . A cohort study of 1220 adults from a general population proved that the fasting level of C3 was strongly associated with abdominal obesity and blood pressure 23 . C3 is also associated with postprandial triglyceride metabolism, higher homeostasis model assessment of IR (HOMA-IR), and diabetes development [24][25][26] . Furthermore, elevated C3 concentrations show greater predictive value than insulin or C reactive protein for the incidence of fitness and fatness 27 . Adipose tissue may be considered as an immunological organ 28 , and adipose tissue inflammation of obese subjects contributes to the development of IR and metabolic dysfunction 29 . High expressions of C3 and C4 are related to adipose tissue variables and involved in the development of visceral adiposity 30 .
Therefore, the complement system response to inflammation and infection may play an important role in MetS expression. However, to our knowledge, large-scale epidemiological studies that demonstrate the association between C3 and C4, and MetS morbidity are limited. In the present investigation, we conducted a cross-sectional and longitudinal population-based study to elucidate whether C3, C4, or immune globulin concentrations are associated with MetS risk, or predict MetS incidence.

Results
Participants' selection of the study. Figure 1 shows the procedures for selecting participants in the study (n = 4303). Among the 4303 participants aged 17-88 years old, 2376 eligible male participants were selected in the cross-sectional study based on the exclusion criteria. A total of 308 (12.96%) and 2068 subjects diagnosed with MetS and without MetS (non-MetS) were identified, respectively, based on the definition of MetS. After applying the exclusion criteria during the follow-up period, 976 participants were included in the cohort study from the 2068 participants without MetS from 2009 to 2013. After a 4 year follow-up, 166 subjects (17.01%) were diagnosed with MetS. Table 1 shows the general characteristics of male participants. The C3 and C4 levels of participants with MetS were 15% higher than those without MetS (both P < 0.001). However, no significant differences were observed in IgM (P = 0.148), IgA (P = 0.115), IgE (P = 0.699), and IgG (P = 0.269) between the two subjects groups. The smoking status between MetS and non-MetS participants was also not significantly different (P = 0.070).

Correlation coefficients between C3 and C4 concentrations and metabolic variables.
Pearson's correlation coefficient showed the relationship between the C3 and C4 levels and the risk factors of MetS after adjusting the confounding factors (Table 2). C3 levels were significantly positively correlated with body mass index (BMI), waist circumference (WC), systolic blood pressure (SBP), diastolic blood pressure (DBP), triglyceride (TRIG), glucose (GLU), and decreased high-density lipoprotein (HDL) (all P < 0.001). C4 levels were also related to the risk parameters except for TRIG (P = 0.153) and GLU (P = 0.057). In addition, C3 levels were highly correlated with BMI and WC (both R = 0.436), whereas C4 levels were strongly correlated with BMI (R = 0.221). No significant association among IgE, IgG, IgM, IgA, and the components of MetS was found after adjustment of multiple factors (Supplemental Table 1

online).
Association between C3 and C4 levels and MetS in the cross-sectional study. Table 3 presents the association between the concentration quartiles of C3 and C4 and MetS by logistic regression analysis. In the unadjusted model (model 1), the odds radio (ORs) of MetS in the highest quartile of C3 and C4 levels were 6.240 and 2.213, respectively (both P < 0.001). After multi-factor adjustment (model 2), the ORs in the highest quartile of C3 and C4 concentrations were 7.047 (P trend < 0.001) and 1.961(P trend = 0.041). Nevertheless, binary logistic regression analysis showed that the ORs and 95% confidence intervals (CIs) of MetS for IgE, IgG, IgM, and IgA were mostly non-significant (Supplemental Table 2 (Table 4). In unadjusted analysis (model 1), the RRs of MetS for the highest quartile of C3 and C4 concentrations were 4.556 and 2.711, respectively (both P < 0.001). After multi-factor adjustment (model 2), the highest C3 concentrations were associated with 4.7 times higher risk of developing MetS compared with their counterparts. Data showed that the incidence of MetS increased with rising quartiles (P trend < 0.001) (Fig. 2). Similarly, the chance of developing MetS was 2.590 times higher in individuals with the highest quartile of C4 concentrations compared with those with the lowest quartile of C4 concentrations (P < 0.001).

Discussion
Both C3 and C4, but not IgE, IgG, IgM, and IgA, were related to independent components of MetS, and elevated C3 or C4 was positively associated with MetS in the cross-sectional study. We also found that the high baseline C3 and C4 levels indicated increased risk of MetS over a 4-year follow-up study.
Chronic inflammation is associated with MetS 6 . Meanwhile, the activation of complement system is an important part of inflammatory processes that interferes with normal metabolism and insulin signaling 17 . In the present study, baseline C3 and C4 levels, but not IgM, IgA, IgE, and IgG levels, showed significant differences between MetS and non-MetS groups. A previous cross-sectional study reported that a borderline relationship exists between IgM levels and MetS. This finding may be mediated by lipid metabolism disorder in males (P = 0.07) when IgM levels are significantly related to elevated TRIG and decreased HDL 31 . However, in the present study, the IgM, IgA, IgE, and IgG exhibited no significant effects on MetS and its components among the subjects (Supplemental Tables 1  and 2 online). Therefore, we propose that the activation of inflammatory pathways, particularly the complement pathway, mainly affects MetS incidence.
Several investigations, mainly case-control or cross-sectional studies reported consistent results. For example, a case-control study found that C3 concentrations are significantly related to serum TRIG, GLU, and BP levels 32 . Individuals with elevated C3 concentrations exhibit three-fold higher risk of MetS compared with those in the bottom 50th percentile, and higher MetS risk among high dietary fat consumers and smokers 33 . A cross-sectional study of Prospective Investigation of the Vasculature in Uppsala Seniors study also demonstrated that C3 and C4 are significantly related to the occurrence of MetS in 1016 older people (≥ 70 years old) 29 . By contrast, the current study has a larger sample size and relatively young population, with an average age of 42.6 years old. In a cross-sectional population-based survey of Wamba et al. 34 , they proposed that acylation-stimulating protein, but not complement C3, is associated with MetS components in 1603 Chinese children and adolescents (6-18 years old). In another study of Wamba et al. on children aged 2-6 years old, they reported a small but significant increase in plasma C3 in young obese subjects 35 . Thus, these results indicate that C3 and C4 may affect MetS prevalence only in adult population, but not significantly in children.
A cohort study focused on cardiometabolic risk and MetS; by analyzing the association between C3 and the incidence of MetS in 265 women and 230 men, the study reported that C3 predicted the risk of MetS in women but not in men 23 . However, basing on the results from the present cohort study, we propose that elevated C3 is an independent risk factor for the development of MetS in a large sample of men. This relationship still existed in men, even after adjustment for other confounding factors. Moreover, the present study found that increased C4 is another potential factor in predicting MetS.
The pathogenesis of MetS caused by activation of complement system is not clearly elucidated, and several pathways may explain how complement factors affect MetS. C3 and C4, as the major plasma proteins of the complement pathway, play a crucial role in the immune system 21 . High levels of C3 may cause high C3a and C5a, these anaphylatoxins mediate inflammatory processes by acting on their respective receptors (C3aR and C5aR) 36,37 . In vivo, studies showed that C3aR and C5aR knock-out mice present reduced macrophage infiltration and improved insulin sensitivity 32,37,38 . C3a and C5a function as hormones that have insulin-like effects on 3T3-L1 adipocytes through increasing glucose and fatty acid uptake while inhibiting cAMP transmission and lipolysis to promote energy conservation 39 . In addition, C3a-desArg is a degraded product of C3 and can stimulate lipogenesis in adipose cells and triglyceride synthesis 40,41 , and also promote islets secreting glucose-dependent insulin 38,42 . The metabolism of triglyceride was accelerated with increasing C3a-desArg levels. This phenomenon increased the plasma triglyceride levels, thereby explaining the IR in obese people and MetS 29 . Zhou et al. suggested that C3 was involved in mediating epithelial-to-mesenchymal transition and in activating intrarenal RA systems with the influence of renin generation. The finding indicated that high level C3 could increase intrarenal Ang II concentration and BP levels 43 , which could explain the association between C3 and hypertension.

Variable
MetS (n = 308) Non-MetS (n = 2068) P The present work is the first to combine cross-sectional and longitudinal studies to investigate the casual relationship between complement activation and MetS. The strengths of our study include population-based sample, large sample size, and relatively high response rate, which may improve the credibility of the results.
Nevertheless, several limitations should be highlighted. First, this study was conducted only among adult males. The results may lack sufficient evidence to indicate the relationships in children and females. Second, no diet intake data or detailed lifestyles were investigated. Therefore, the effects of diet on MetS were not analyzed in our study.
In conclusion, activation of complement factors, as an important part of inflammatory processes, may directly affect MetS incidence. Elevated C3 and C4 levels were independent risk factors for the development of MetS in Chinese male population. Further studies are urgently needed to explore the molecular mechanisms of activation of complement factors acting on MetS. Ethics. This study was approved by the Ethics and Human Subject Committee of Guangxi Medical University and was conducted in accordance with the Declaration of Helsinki. All participants were given and signed a written informed consent form prior to the study.

Study population. This study is part of Fangchenggang Area Males Health and Examination Survey
Cross-sectional study. The exclusion criteria are as follows: 1 presence of hepatic disease (ALT > 2.0 times upper limit of normal or chronic hepatitis or liver cirrhosis), 2 presence of renal disease (creatinine > 178 μ mol/l), 3 missing data on C3 and C4 concentrations, 4 taking anti-inflammatory medication, and 5 clinical history of hyperthyroidism, rheumatoid arthritis, and cancer. Participants who had signs of high fasting blood glucose levels and

Cohort study.
A total of 2376 participants without MetS at baseline in 2009 were followed up for 4 years. The participants completed the same face-to-face interview and physical examination in 2013. We excluded members who had one or more of the following criteria: 1 loss to follow-up, 2 unwillingness to participate, 3 subsequent diseases that are unsuitable for participation, and 4 missing data on anthropometric measurements or clinical biochemistry assays in 2013. A total of 976 men participated in the follow-up exam.

Data collection. We collected the data from Fangchenggang First People's Hospital Medical Examination
Center. Participants' age, smoking habits, alcohol consumption, and health status were obtained through a face-to-face interview. Subjects who had ever smoked cigarettes for six months or longer were classified as smokers, and others were considered as non-smokers. Drinkers were defined according to alcohol consumption once or more a week, and others were considered as non-drinkers. Subjects who had exercised two hours or more each week were classified as regular physical activity, and otherwise subjects were classified as irregular physical activity 45 . Participants were considered to have a family history of chronic diseases if one of the following is present in participants' parents or siblings: hypertension, type 2 diabetes mellitus, coronary heart disease, and stroke 46 . Anthropometric measurements were conducted by trained personnel using a standard protocol. Standing heights were measured using a vertical telescopic stadiometer with a horizon headboard on the top to the nearest 0.1 cm. Participants were weighed without shoes to the nearest 0.1 kg. BMI was calculated as weight divided by height squared (kg m −2 ). WC was assessed midway between the lower rib margin and the iliac. BP was measured twice on each visit after 5 min of rest with an oscillometric precision blood pressure instrument on the right arm by trained nurses.
Laboratory measurements. Overnight fasting venous blood samples were obtained between 8 a.m. and 10 a.m. The samples were transported on ice in an upward position to the testing center of the Department of Clinical Laboratory at the First Affiliated Hospital of Guangxi Medical University in Nanning within 2 h. Sera were separated using a centrifuge within 15-25 min and stored at − 80 °C until analysis. Blood lipids and serum glucose were determined using the enzymatic assay on a Dimension-RxL chemistry analyzer (Dade Behring, USA). IgE was measured by electro-chemiluminescence immunoassay on COBAS 6000 system E601 (Elecsys module) immunoassay analyzers (Roche Diagnostics, IN, Germany) with the same batch of reagents. C3, C4, IgG, IgA, and IgM were measured with a Hitachi 7600 autoanalyzer (Hitachi Corp., Japan) in the testing center of the First Affiliated Hospital of Guangxi Medical University. The coefficients of variation were 2.16% for C3, 2.51% for C4, 4.67% for IgM, 4.97% for IgG, 3.57% for IgA, and 3.3% for IgE.

Definition of MetS cases. Based on the National Cholesterol Education Program -Adult Treatment Panel
III definition as modified by the American Heart Association/National Heart, Lung and Blood Institute 41 , individuals with MetS were identified having arbitrarily three or more of the following criteria: (1) WC ≥ 90 cm, (2) TRIG ≥ 1.7 mmol/l, (3) HDL < 1.03 mmol/l, (4) SBP ≥ 130 mmHg/DBP ≥ 85 mmHg or current use of hypotensive drugs or have a history of hypertension, and (5) fasting GLU ≥ 5.6 mmol/l or previously diagnosed with type 2 diabetes mellitus or taking medications to control hyperglycemia. Statistical methods. Continuous variables were described as mean and standard deviation (SD) ranges with normal distribution, median, and 25-75 percentile with abnormal distribution, and categorical variables were described as proportions. Two-sided t test and Mann-Whitney U test were used to analyze the differences of variables between the MetS and normal groups. Pearson's coefficient was employed to describe the correlations of C3 and C4 with metabolic variables after multi-factor adjustment (including age, smoking, drinking status, physical activity, and family history of chronic diseases). Estimates of the likelihood of MetS and 95% CIs were obtained using logistic regression analysis in models with or without adjustment of multi-factors. After the 4-year follow-up, Cox proportional hazard regression was conducted to compute RRs and 95% CI for MetS after adjustment or without adjustment for multi-factors. A linear trend across increasing quartiles was tested with a mean value of each quartile as an ordinal variable. All statistical analyses were performed using SPSS (Chicago, IL, USA) for Windows 16.0. Statistical tests were two-sided, and P < 0.05 was considered statistically significant.