Association of Complement-Related Proteins in Subjects With and Without Second Trimester Gestational Diabetes

Introduction Gestational Diabetes Mellitus (GDM) development is related to underlying metabolic syndrome that is associated with elevated complement C3 and C4. Elevated C3 levels have been associated with preeclampsia and the development of macrosomia. Methods This case-control study included 34 pregnant women with GDM and 16 non-diabetic (ND) women in their second trimester. Complement-related proteins were measured and correlated with demographic, biochemical, and pregnancy outcome data. Results GDM women were older with a higher BMI (p<0.001); complement C3, C4 and Factor-H were significantly elevated (p=0.001, p=0.05, p=0.01, respectively). When adjusted for age and BMI, Complement C3 (p=0.04) and Factor-H (p=0.04) remained significant. Partial correlation showed significant correlation between C4 with serum alanine aminotransferase (ALT) (p<0.05) and 2nd term diastolic blood pressure (p<0.05); Factor-H and C-reactive protein (CRP; p<0.05). Pearson bivariate analysis revealed significant correlations between C3, C4, and Factor-H and CRP; p<0.05; C3 and gestational age at delivery (GA; p<0.05); C4 and ALT and second-trimester systolic blood pressure (STBP) (p=0.008 and p<0.05, respectively); Factor-H and glycated hemoglobin (HbA1c) (p<0.05). Regression analysis showed that the elevation of C3 could be accounted for by age, BMI, GA and CRP, with CRP being the most important predictor (p=0.02). C4 elevation could be accounted for by ALT, CRP and STBP. CRP predicted Factor-H elevation. Conclusion The increased C3, C4 and Factor-H during the second trimester of pregnancy in GDM are not independently associated with GDM; inflammation and high BMI may be responsible for their elevation. The elevation of second trimester C3 in GDM is associated with earlier delivery and further work is needed to determine if this is predictive.


INTRODUCTION
Gestational diabetes (GDM) is the most frequent pregnancyassociated metabolic disorder, occurring in 14% of pregnant women (1,2). GDM is usually identified in the latter stage of the second trimester (3). GDM is associated with increased risk of complications for both the mother and the fetus, including preterm delivery, cesarean delivery, preeclampsia, macrosomia, shoulder dystocia, neonatal hypoglycaemia and respiratory distress syndrome (4,5). Women with a history of GDM are at increased risk for cardiovascular disease (6) and type 2 diabetes (T2DM) (7,8) later in life.
During pregnancy, the nutritional demands of the fetus cause a physiological increase in insulin resistance in the mother (9). In a healthy pregnancy, compensatory mechanisms cause an increase in glucose-stimulated insulin release (10) together with an adaptive increase in beta-cell mass (11) to counterbalance the increase in insulin resistance. However, in pregnant women who are overweight or obese, insulin requirements are increased and, if the demand exceeds the insulin-secretory capacity, these conditions can increase risk for GDM (12). Complex, and not yet fully elucidated, mechanisms drive the pregnancy-related insulin resistance; notable factors are placental hormones, obesity, inactivity, poor diet, and genetics/epigenetics (13).
Older maternal age, pre-gravid overweight or obese status (14,15), multiparity, ethnicity, a family history of diabetes, and excessive gestational weight gain (16) are recognized risk factors for the development of GDM.
Biomarkers that could predict GDM and its complications earlier in pregnancy than is currently possible using the standard oral glucose tolerance test would offer practical benefit for patient care and may provide a deeper understanding of the biochemical pathways involved in GDM and whether they parallel, or are divergent from, those in T2DM.
Pregnancy presents a unique immunologic state that initiates in earnest at implantation and usually resolves after delivery. Similarly, GDM is a unique metabolic state that begins and ends with gestation. The immunological changes that occur during normal and in GDM women during pregnancy is not clearly understood. The evolutionarily conserved complement system protects the host against bacterial infection, the individual elements forming a serine-protease cascade that attacks the membranes of invading organisms and induces cell lysis; hence, this system has been most studied in the context of infectious diseases ( Figure 1). Though less well defined, the complement system also plays a role in metabolic disorders, and is essential for homeostasis of host cells, removal of apoptotic cells and priming of the adaptive immune response (17), processes which all occur in normal pregnancy and which may be dysregulated in pregnancies accompanied by metabolic disorders, such as GDM, and preeclampsia (17).
In normal pregnancy, increased activation of the complement system occurs with elevated plasma concentrations of C3a, C4a, and C5a (18), which may be important to counterbalance the normal suppression of adaptive immunity during pregnancy. Regulatory T cells enable fetal tolerance and complement activation controls T cell development (19,20). Notably, preeclampsia has a well-known association with complement activation (21)(22)(23)(24)(25) though which regulatory T cells are decreased (26,27). Factor D concentrations are increased (22). Complement activation, specifically factors C3, C4, C3a, Factor-H and Properdin, are associated with increased incident metabolic syndrome (28); and complement activation was related to adverse pregnancy outcomes, such as intrauterine growth retardation and GDM (24,29). Further, polymorphisms of the mannose-binding lectin gene are associated with a higher risk for GDM (21). At term, patients with GDM have lower levels of C3a, C4a and C5a (30).
Given the elevation of complement in normal pregnancy we hypothesized that there would be further dysregulation of the complement system in GDM. Therefore, in this study, we sought to determine the serum complement protein concentrations in second trimester pregnant women with and without GDM and their relationship to patient demographics, biochemical parameters and pregnancy outcomes.

Study Design
This was a case-control study in 50 pregnant women (34 GDM and 16 non-diabetic (ND)) who were recruited during their second trimester at the antenatal clinic at the Women Wellness and Research Centre (WWRC) of Hamad Medical Corporation (HMC), Doha, Qatar, during 2016-2017. The study protocols were approved by the Institutional Review Boards (IRBs) of HMC (15101/15) and Weill Cornell Medical College in Qatar (WCMQ) (15)(16). Pregnant women between the ages of 18 and 40 without any previous medical history of chronic disease, in the second trimester of pregnancy and willing to comply with all study procedures and be available for the duration of the study were included. Pregnant women who are unable to provide informed consent, who were in first or third trimester of pregnancy, or currently enrolled in other clinical trials were excluded.
Demographics, anthropometrics, and medical history data were collected, including age, ethnicity, socio-economic background, vital signs, height, weight, menstrual cycle, the period of infertility, medications, complications, comorbidities, and family medical history. According to Qatar national guidelines for GDM, all pregnant women are screened at the first antenatal care visit by measuring fasting blood glucose (FBG). If FBG at the first visit was >5.1 mmol/l (92 mg/dl), 75 g OGTT is performed at 24 weeks' gestation. The WHO criteria [FBG ≥5.1 mmol/L (92 mg/dl), 1-h post OGTT ≥10.0 mmol/L (180 mg/dl) or 2-h post OGTT ≥8.5 mmol/L (153 mg/dl)] were used for GDM diagnosis. GDM patients were seen back in the outpatient clinic within 1 week of the laboratory diagnosis of GDM. During this time blood samples were collected for complement proteins measurement. GDM patients were started on a diet for two weeks to achieve an FBG ≤ 5.3 mmol/l (95 mg/dl) and a 2-h post-prandial glucose ≤ 6.8 mmol/l (120 mg/dl) in ≥ 80% of the readings. If more than 20% of the readings were above targets, then metformin therapy was initiated and increased incrementally, followed by insulin supplementation if glucose targets were not achieved.
All patients gave written informed consent, and the conduct of the study was in accordance with ICH GCP and the Declaration of Helsinki. Pregnancy outcomes of gestational age at delivery, birth weight, maternal weight, blood pressure, and fetal outcome were recorded and collated with the apolipoprotein profile for all subjects who participated in the study.

Statistical Analysis
There was no specific study on second trimester complement levels on which to power the study; however, given that GDM has many features of the metabolic syndrome, C3 differences between those with and without metabolic syndrome were used (28). An alpha of 0.05 with 80% power gave an effect size of 1.07, requiring a minimum of 15 subjects per group. (nQuery, Statsol USA). Descriptive statistics and means ± standard deviations (SD) were calculated for all continuous variables in the study. A general linear model was used to compare mean differences between control and GDM groups before and after adjustment for age and BMI. Pearson and partial correlations were performed to understand the associations between complement variables and demographic variables, with the partial correlations adjusted for age and BMI. Linear regression analysis was performed to determine predictors for circulating C3, C4 and Factor-H. All statistical analysis was done using statistical analysis SAS version 9.4 software. A statistical significance level (P-value) of <0.05 was considered as significant.

Demographic and Biochemical Characteristics of Study Participants
Compared to the ND women, the GDM women were older (34.3 ± 4.4 vs 29.7 ± 4.2 years, p=0.001) and had a higher BMI (35.3 ± 5.6 vs 28.4 ± 6.4 kg/m 2 , GDM vs ND, p=0.0003). Baseline systolic and diastolic blood pressure were comparable between the 2 groups. GDM was diagnosed at 22.2 ± 4.1 weeks, and the control women were matched for gestational age (p=0.308). In keeping with the diagnosis, fasting plasma glucose was elevated in the GDM women (5.5 ± 0.9 vs 4.7 ± 0.3, GDM vs ND, p=0.039), though insulin and HbA1c did not differ between groups, indicative of the recent onset of hyperglycemia in the GDM women. Lipids (cholesterol, triglycerides, high-and lowdensity lipoproteins) and CRP did not differ between the cohorts, while ALT was elevated in the GDM women (17.5 ± 11.8 vs 10.2 ± 3.6 U/l, GDM vs ND, p=0.005).There were two premature deliveries in the control and five in the GDM group and no infant was LGA in our cohort. Gestational age at delivery and baby weight were similar between the cohorts ( Table 1).
Using the combined group of women (GDM and ND), partial correlations ( Table 3) and Pearson bivariate analysis (Supplementary Table 1) were used to determine correlations between the complement proteins and the demographic, clinical and biochemical data and between members of the complement system ( Table 4 and Supplementary Table 2).
Using Pearson bivariate analysis, a positive correlation was found for complement C3 with age (r=0. 29 Multiple correlations were found between complement C3, complement C4 and Factor-H and other complement and anti-complement proteins, reflecting their closely integrated regulation ( Table 4 and Supplementary Table 2).
Regression analysis of all women combined showed that the elevation of C3 could be accounted for by age, BMI, GA, and CRP, with CRP being the most important predictor (p=0.024). Complement C3b/iC3b could be accounted for by second trimester SBP. C4 could be accounted for by ALT, CRP, and SBP. CRP predicted Factor-H elevation.

DISCUSSION
The results of this study show an elevation of Complement C3, C4 and Factor-H in GDM versus ND control pregnant women in the second trimester of pregnancy. Of note, the elevation in complement factors in GDM over that of normal pregnancy could largely be accounted for by inflammation, as assessed by CRP, suggesting that C3, C4 and Factor-H are not independently associated with GDM.
Complement C3 has been shown to increase with normal pregnancy (18) and fall at term (30), but their levels in the second trimester have not been described. Also, C3 has been shown to be further elevated in pre-eclampsia (24). Furthermore, increased C3 is associated with obesity, dyslipidaemia, inflammation, insulin resistance and liver dysfunction (32), most of the known factors associated with an increased risk of preeclampsia (33). Many of these features were seen in the women with GDM who were more obese, older, and with higher ALT levels. CRP was the greatest predictor of an elevated C3 and, when taken into account together with age and BMI in the regression model, then C3 was no longer significantly elevated, suggesting that it is reflecting the GDM   rather than being independently associated with it. Consistent with our data, increased C3 and CRP were associated with preterm delivery in other studies (34,35). Complement C4 was also elevated in GDM and associated with CRP as a marker of inflammation but, in the regression model accounting for CRP, ALT and SBP, then C4 was no longer significant, mirroring the findings with C3. ALT levels were higher in GDM, but not above the upper limit of the reference range; however, alterations in ALT are associated with nonalcoholic fatty liver disease which features an increase in insulin resistance, and an elevated ALT is associated with increased risk for GDM (36) though, notably, gamma-glutamyl transferase is a superior predictor of GDM (37). The correlation of C4 with ALT in GDM has not previously been reported.
C4 is associated with systolic blood pressure, and the complement system is associated with arterial hypertension and hypertensive end organ damage (38); however, when adjusted for SBP, then these proteins were not significantly different, suggesting that complement activation was not driving the changes in blood pressure.
Factor H is a serum glycoprotein that accelerates the decay of C3 convertase and is a cofactor for the inactivation of C3b (25). In the absence of factor H, spontaneous activation of the alternative pathway results (23). Factor H, in circumstances of cell stress such as oxidative stress and hypoxia, may be downregulated (39) and that, in turn, would contribute to any underlying inflammatory process (40). Theoretically, the tendency toward increased inflammation in GDM, as seen by the trend toward CRP elevation, is mainly responsible for the increase in C3 found in GDM. This is reflected in the increase in the protective factor H, that correlated with CRP. Increased oxidative stress as occurs in GDM, consequent upon the increase  in insulin resistance (27), may reduce factor H and pathologically increase C3 levels.
To our knowledge this is the first study to investigate all the three complement system cascades (classical, alternative and lectin pathway) during pregnancy in GDM subjects. Limitations of this study include the relatively small numbers of pregnant women in each group, which may have prevented the detection of differences between groups. As a cross-sectional study with serum analysis at a single time point during the second trimester, dynamic changes in complement system proteins levels throughout pregnancy could not be assessed. Further, the study was undertaken in a single homogenous population and, while likely generalizable, these findings should be confirmed in other ethnic populations. Further studies on a suitably powered cohort of women with GDM with and without metformin therapy would address the question whether therapy with metformin could have any additional impact on the inflammatory/complement proteins.

CONCLUSION
In conclusion, inflammation and increased BMI associated with GDM are likely responsible for the increased C3, C4, and Factor-H seen in the second trimester of pregnancy in GDM that are not independently associated with GDM, and the elevation of C3 was negatively associated with GA; further work is needed to determine if this is predictive.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Institutional Review Boards of the Hamad Medical Corporation, Qatar (15101/15) and Weill Cornell Medicine-Qatar (15-00016). The patients/ participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
MR, MA, and JJ performed the complement protein measurements and contributed to the manuscript. MR, LA, MA, AM, ME, and A-BA-S helped with data analysis, preparation of tables, and contributed to manuscript preparation. AB researched the data and wrote the manuscript. SH helped with statistical analysis. AG captured and provided hormonal and biochemistry data. AM researched data and contributed to manuscript preparation. MB and SA were involved in study design, sample collection and data analysis. MR, SA, and A-BA-S designed the experiments, supervised progress, analyzed data, and revised and approved the final version of the article. All authors contributed to the article and approved the submitted version.