RETRACTED ARTICLE: Association of cord blood asprosin concentration with atherogenic lipid profile and anthropometric indices

This study was conducted from June 2018 to August 2018 at the Mobini Hospital in Sabzevar (a town in the Khorasan-Razavi province), Iran. We enrolled 450 healthy pregnant women recruited to the hospital for their delivery based on our inclusion/exclusion criteria. The inclusion criteria were mothers with a normal term of pregnancy (37–42 weeks of gestation), normal singleton pregnancies, normal vaginal delivery, normal lipid profile, self-report of no hypertension, preeclampsia, diabetes, and gestational diabetes. The exclusion criteria included intrauterine fetal growth restriction, fetal structural abnormalities, and drug intake that affects metabolism. It should be noted that the inclusion/exclusion criteria were detected through self-reports and medical records of participants. The project was approved by the Clinical Research Ethical Committee (IR.MEDSAB.REC.1397.012) of Sabzevar University of Medical Sciences and all participants signed the informed consent form before to enrollment.

Cord blood samples were collected from the umbilical cord vein in serum separator tubes with a clot activator at delivery time. Sera were separated, aliquoted and stored at -80 until subsequent analysis. TAG, HDL-C, and TC were measured using an automated analyzer (Biotecnica, BT 1500, Rome, Italy) and commercial kits (P.L:38,231,049, Pars Azmoon, Tehran, Iran) by enzymatic colorimetric methods. LDL-C concentration was calculated using the Friedewald equation. Asprosin concentration was quantified by enzyme-linked immunosorbent assay (ELISA) using a commercially available kit according to manufacture instructions (Cat.No: CK-E91570, EASTBIOPHARM, China). Asprosin concentration was reported in ng/mL. The limit of detection (LOD) was 0.34 ng/mL. The intra-assay and interassay coefficients of variation were < 10% and < 12%, respectively.

The descriptive statistics of the participants are presented in Table 1. The mean (standard deviation (SD)) age of mothers was 27.65 (5.36) years. Of the 450 newborns included in our study, 225 (50%) were male. The median (IQR) of umbilical TAG, TC, LDL-C, HDL-C, were 29.9 (26.9), 60.4 [27], 23.28 (21.48), 28.2 (12.5) mg/dL, respectively. Moreover, the median (IQR) for TAG/ HDL-C, TC /HDL-C, and Non-HDL-C/HDL-C were 1.15 (1.18), 2.01 (0.766) and 1.01 (0.76), respectively. The median (IQR) of umbilical asprosin was 30.44 (19.08) ng/mL. No significant differences were observed in asprosin concentration as well as lipid profile between male and female newborns (P-value > 0.05). The median (IQR) for birth weight (BW), birth length (BL), head circumference (HC) and chest circumference (CC) were 3250 (550) gr, 51 (3) cm, 35 (2) cm and 33 (3) cm, respectively.

The results of the relationship between umbilical cord blood concentration of asprosin and BW, BL, HC and CC are presented in Table 2. In crude and adjusted models, higher levels of cord blood asprosin were positively associated with BW, BL, HC and CC as well as gestational age at delivery; however, these associations were not statistically significant.

The association of asprosin concentration and lipid profile are presented in Table 3. In fully adjusted models increase in umbilical asprosin concentration was associated with increase in umbilical TAG, TC, LDL-C levels as well as TAG/HDL-C, TC/HDL-C, LDL-C/HDL-C and non-HDL-C/HDL-C ratio. Each 1 ng/ml increase in asprosin was associated with 0.19 mg/dL (95% CI 0.06, 0.31, P < 0.01), 0.19 mg/dL (95% CI 0.10, 0.29, P < 0.01), 0.17 mg/dL (95% CI 0.09, 0.25, P < 0.01), 0.17 (95% CI 0.09, 0.25, P < 0.01), 0.01 (95% CI 0.00, 0.01, P < 0.01), 0.01 (95% CI 0.01, 0.01, P < 0.01), 0.008 (95% CI 0.01, 0.01, P < 0.01) and 0.01 (95% CI 0.01, 0.01, P < 0.01) increase in TAG, TC, LDL-C, TAG/HDL-C, TC/HDL-C, LDL-C/HDL-C and non-HDL-C/HDL-C ratio respectively. The association of asprosin concentration and HDL-C concentration was not statistically significant (P-value = 0.43).

The results of adjusted our models for umbilical blood glucose and mother's lipid profile on the association of asprosin concentration with anthropometric indices and umbilical lipid profile showed no notable differences in the main results in terms of significance and direction (data not shown).

We did not find a study on the association of asprosin concentration and lipid profile in cord blood; therefore, it is impossible to compare our results with other studies. However, several studies have examined the association of serum asprosin concentrations and lipid profile in adults. Wang et al. 2018 showed that serum concentrations of asprosin were positively correlated with TAG concentrations. They also reported that asprosin concentrations were negatively correlated with HDL-C concentrations [20]. Moreover, the association of serum concentration of asprosin and HDL-C has been reported in a study by Long et al. [16]. The studies by Whang et al. 2018 and Zhang et al. 2019 reported there was no statistically significant association between serum concentrations of asprosin and LDL-C [15, 20]. However, Li et al. 2018 showed that serum concentrations of asprosin were positively associated with LDL-C, apoB and apoE [21].

We did not observe any significant difference in lipid profile as well as asprosin concentration between male and female newborns. However, a study by Kelishadi et al. 2007 reported that total and HDL cholesterol in girls were significantly higher than in boys [9]. Moreover, a study by Ghiasi et al. 2013 reported that the cord blood from female newborns had higher levels of LDL-C, HDL-C and total cholesterol compared to male newborns [22]. Furthermore, Jahanfar et al. 2016 in a cohort study of twins, reported that male-male twins were heavier than male-females and female-female twin pairs [23].

Our study found a positive but non-significant association between asprosin concentration and anthropometric indices. A study by Hoffmann et al. 2022 on 247 umbilical plasma samples reported that there was no significant correlation between BW and BL with asprosin concentration [24].

Asprosin may affect lipid metabolism by two different mechanisms: the effect on the insulin signaling pathway and proopiomelanocortin (POMC)-positive neurons activity. Some effects of asprosin on lipid profile may be exerted by altering gene expression in liver cells through insulin resistance induction [25‐28]. High insulin concentrations in IR alter the expression of some genes [29‐32]. In the liver, the target tissue of asprosin, insulin induces the expression and activation of enzymes involved in hepatic lipid production [33‐37]. Although the molecular mechanism of asprosin on insulin singling is not well known; however, several studies suggested elevated concentrations of asprosin interfere with insulin function and induces insulin resistance [38‐41]. Romere et al. 2016 showed that single injection of asprosin results in hyperglycemia and hyperinsulinemia [14]. Wang et al. 2018 found a negative relationship between serum asprosin concentrations and indicators regarding the first-phase insulin secretion, such as AUC, AIR, and GDI and HOMA-β, and a positive correlation with HOMA-IR [20]. They suggested that the asprosin-related metabolic pathways dysregulation might be through its role in ꞵ-cell dysfunction and insulin resistance [20]. Jung et al. 2019 showed that treatment of skeletal muscle cells by recombinant asprosin results in impairment of insulin sensitivity through PKCδ-associated ER stress/inflammation pathways [26]. Lee et al. 2019 showed that treatment of primary human islets with recombinant asprosin induced the inflammation response, cellular dysfunction, and apoptosis in a dose-dependent manner [42].

Asprosin exerts some of its effects by changing the function of the CNS system. Duerrschmid et al. 2017 demonstrate that asprosin in the circulation crosses the blood–brain barrier and inhibits the downstream anorexigenic proopiomelanocortin (POMC)-positive neurons. The inhibition of these neurons leads to appetite stimulation and a drive to accumulate adiposity and body weight [19, 43, 44]. Blockade of the CNS melanocortin system increased triglyceride synthesis in the periphery and triglyceride content in the liver [45, 46].

Our study has some limitations that should be considered in future studies. The sample size was relatively limited, and the findings of this study required to confirm from other ethnicities. Also, the cross-sectional nature of this study did not allow us to address the underlying signaling pathways. Moreover, we did not have any data on pre-pregnancy lipid profile, maternal diet, maternal weight gain and other prevalent endocrine disorders like PCOS that can affect our results. Furthermore, we missed data about timing of cord clamping that could affect TG and total cholesterol. Research in cellular models is required to determine underlying mechanisms.