A novel association of CCDC80 with gestational diabetes mellitus in pregnant women: a propensity score analysis from a case-control study

Gestational diabetes mellitus (GDM) refers to impaired glucose tolerance with onset or found for the first time in the second or third trimester of gestation, which is a common metabolic disturbance during pregnancy [1, 2]. In global, GDM impacts 3–25% pregnancies and the continuous soared in the incidence of GDM concurs with the increasing prevalence of maternal obesity [1, 3]. With the development of economy and changes in lifestyle, incidences of GDM in China have been rising yearly from 2.3% in 1999 to 6.8% in 2008 and 9.3% in 2012 [4‐6]. Maternal exposure to GDM are regarded to be directly associated with adverse perinatal and late-life complications in offspring [7]. A survey indicated that GDM begot over 50% of pregnant women to suffer from type 2 diabetes among five to 10 years after delivery [8]. Furthermore, it was evaluated that approximately 18.4 million live births worldwide were influenced by GDM in 2017 [9]. Particularly, neonates born to pregnant women with GDM are more vulnerable to obesity, metabolic and cardiovascular diseases during childhood and adulthood [10, 11]. Recently, the most concerning is the reciprocation of the cycle of diabetes between mother and child [12].

Adipocytes not only serve as sites of lipids deposition but can also secrete diverse mediators through autocrine, paracrine or endocrine mechanisms, referred to as adipocyte-secreted proteins or adipokines. These released proteins play pivotal functions in energy homeostasis, insulin sensitivity and systemic inflammation [13]. Coiled-coil domain-containing 80 (CCDC80, also known as DRO1 and URB), a novel adipokine, was originally identified as up-regulated in brown adipose tissue of mildly obese mice [14]. Contrastingly, CCDC80 down-regulation was discovered in white adipose tissue in ob/ob, KKAy and diet-induced obesity mouse models [15]. Previous mechanism studies have provided evidence that CCDC80 plays dual roles in adipogenesis through modulating Wnt/β-catenin signaling, C/EBPα and PPARγ expression [16]. In addition, CCDC80 knockout mice display hyperglycemia, decreased glucose tolerance and impaired insulin secretion in mice fed a high-fat diet [17]. Therefore, CCDC80 might be a novel regulator of energy and glucose metabolism in mice. Meanwhile, a small population-based epidemiological survey has demonstrated that CCDC80 may be a component of the obesity-linked secretome in visceral adipose tissue and whose blood levels are connected to glucose intolerance and low-grade inflammation related chronic complications [18]. And, Li et al. found that serum CCDC80 was negatively correlated with fasting blood glucose (FBG) in overweight and obesity subjects [19]. Nonetheless, whether this molecular is linked to the risk of GDM in pregnant women is less clear.

Hereby, the aims of this study was to estimate the circulatory CCDC80 concentration in pregnant women with GDM in a propensity score matching (PSM) case-control study, to examine the discriminatory power of CCDC80 on GDM, and to elucidate the relationships between the CCDC80 level and clinical cardiometabolic parameters.

In the present study, we determined for the first time that the CCDC80 levels decreased in pregnant women with GDM, compared with normal blood glucose subjects; and this molecule was a strong independent predictor of GDM. So far, the role of CCDC80 in metabolic derangements, especially among pregnant women has been rarely explored. This is one of the first studies to explore the discriminatory power of CCDC80 on the GDM and the relationships between the CCDC80 level and clinical cardiometabolic parameters in pregnant women in a PSM, case-control design study.

CCDC80 has been considered as a multipurpose molecule in vertebrates, mediating diverse developmental processes [23]. Using transcriptional profiling, previous publications have identified that CCDC80 may play a role in tumor inhibition, such as ovarian cancer [24], malignant melanoma [25], thyroid [26] and colorectal carcinoma [27]. CCDC80 is expressed in several types of cells, in particular in preadipocytes and adipocytes, while it is temporarily down-regulated during cell differentiation [16]. Consequently, the role of CCDC80 in metabolism has been gradually recognized in recent years. Tremblay et al. demonstrated that CCDC80 lacking mouse manifested decreased glucose tolerance and insulin sensitivity [17]. Herein, we clarified that the concentration of CCDC80 was lower in pregnancy women with GDM than that of control group suggesting that CCDC80 might have a protective effect on GDM. These results are concordant with previous findings of negative correlation between serum CCDC80 and FBG in adults [19]. Moreover, recent data has indicated that serum level of CCDC80 is linked with glucose clearance and insulin secretion [18]. From the above, we reasoned that circulating CCDC80 may be an effecient biomarker for GDM.

In the present study, conditional multi-logistic regression analyses unveiled that CCDC80 was an independent protective and predictive biomarker for GDM; further, the results of the ROC analysis indicated that CCDC80 displayed the potentiality to identify individuals with GDM (all AUC > 0.5). Taken together, our results collectively suggested that plasma CCDC80 is a candidate clinical indicator for prediction and diagnosis of GDM in pregnant women. Even though the accurate pathophysiological process of CCDC80 on GDM is not yet well-known, findings from mouse models propose that CCDC80 might function to govern glucose equilibrium through Wnt/β-catenin signaling pathway [16, 17]. The role of Wnt/β-catenin pathway signaling in metabolic diseases was well-established. Latest research places WNT signaling pathway in a crucial position in modulating pancreas function, insulin synthesis and secretion [28]. Conjointly, this seems to point out that elevated CCDC80 level is predictive for glucose and insulin sensitivity disturbances, while the specific mechanism pathway needs to be further explored.

Furthermore, univariate and multivariate regression analyses proved that plasma CCDC80 content was negative related with glucose levels at 1 h post- OGTT, but this connection was not remained throughout the OGTT procedure. These findings are in line with prior views of negative correlation between CCDC80 and glucose levels just at 30 min post- OGTT [18]. Besides, our analysis found that CCDC80 was positively correlated with AST and MAO, indicating a positive correlation with liver function. Except for the correlation between serum CCDC80 level and degree of hepatic steatosis observed by Osorio-Conles et al., there is no previous evidence of CCDC80 functioning in liver. Furthermore, component C1q, a marker of innate immune system, was also positively linked to circulating CCDC80 level. Previous studies have illustrated an advantageous effect of component C1q, which triggers activation of the classical pathway, in diabetes mellitus [29, 30]. Finally, we detected that plasma CCDC80 level was positively associated with the dyslipidemia and atherosclerosis marker LDL-C, apoA1 and apoB. These findings are consistent with earlier surveys, which has indicated that blood CCDC80 is positively correlated with atherosclerosis and fasting triglyceride levels [18]. We speculated that circulating CCDC80 might be involved with obesity-related processes. Previous studies have found that CCDC80 promotes the phosphorylation of extracellular regulated protein kinase 1/2 (ERK1/2) [31] that governs lipoprotein lipase, the rate-limiting enzyme for lipid metabolism, expression and activity [32‐34]. A number of studies have disclosed that CCDC80 is driven by estrogen and may process a particular feature in the control of body weight and energy homeostasis [35]. Of note, although a higher expression of CCDC80 in visceral adipose tissue of obese subjects, the circulating CCDC80 was not connected to maternal BMI in the current study, either before or after PSM (r = − 0.036, P = 0.528 and r = 0.044, P = 0.556, respectively). Previous study has yielded similar results [18].

There are also several limitations in our study. First, despite a PSM design has been introduced to minimize the bias, it is impossible all of the relevant confounders were covered in our PSM method. Hence, unidentified potential covariates may still affect the true relations. Second, owing to the cross-sectional data, definite causal relationship between circulating CCDC80 protein and GDM cannot be judged. Further longitudinal studies are required to infer whether this relationship is causal. Third, our blood specimens were collected within the second trimester of gestation, thus we do not have this molecular levels during the first and third trimesters.

Maternal exposure to GDM rises the probability of the development of serious adverse perinatal outcomes and metabolic complications at a later stage of lifespan of the offspring [36], and better understanding the physiopathology and molecular basis of this condition could lead to novel therapeutic strategies. The present study suggests CCDC80 could be vital for the development of therapeutic approaches to battle GDM in pregnant women. It may provide a new target and biomarker for further investigations aimed to explore the longitudinal trend and mechanisms of this molecule during pregnancy.

In summary, this study demonstrates for the first time that the CCDC80 levels depress in pregnant women with GDM, compared with normal blood glucose subjects; and the CCDC80 is a strong independent predictor of GDM. CCDC80 may be a novel predictor, diagnostic and therapeutic target.