Serum Levels of Asprosin, a Novel Adipokine, Are Significantly Lowered in Patients with Acromegaly

Background Asprosin is a novel identified adipokine secreted mainly by white adipose tissue, which is elevated in metabolic diseases such as diabetes and obesity. Acromegaly is a syndrome caused by pituitary growth hormone (GH) cell adenoma with excessive GH secretion. Serum adipocytokines levels may be involved in abnormal glycolipid metabolism in acromegaly patients. Objectives To investigate serum asprosin levels in acromegaly patients and its correlation with high GH levels and glucolipid metabolic parameters. Methods A retrospective case-control study was conducted and 68 acromegaly patients and 121 controls were included in this study. Clinical information and laboratory examinations were collected and serum asprosin levels were measured by commercial ELISA kits. Results Serum asprosin levels in acromegaly patients were significantly lower than controls (P < 0.001). Serum asprosin levels in patients with the course of acromegaly ≥5 years (compared with <5 years), high area under curve of growth hormone (GH-AUC) after 75 g oral glucose tolerance test (OGTT) (compared with low GH-AUC patients), and high IGF-1 SDS group (compared with low IGF-1 SDS group) were significantly reduced (all P < 0.05). Serum asprosin levels in acromegaly patients were negatively correlated with the course of acromegaly, IGF-1 SDS, nadir growth hormone value (GH-Nadir), and GH-AUC after OGTT. Multiple stepwise linear regression indicated that acromegaly was an independent influencing factor of serum asprosin levels. According to serum asprosin levels tertiles, the risk of acromegaly in the lowest group was 2.67 times higher than the highest group (OR ＝ 3.665, 95% CI 1.677 ∼ 8.007, P=0.001), and the increased risk of the lowest group still existed after adjusting for gender, age, BMI, and TC (Model 2). Conclusions Serum asprosin levels in acromegaly patients are lowered, which may be related to increased blood glucose and reduced body fat mass caused by long-term high GH levels exposure.


Introduction
For many years, adipose tissue has been regarded as the passive energy store organ. In recent years, adipose tissue has been thought to be the largest active endocrine organ in the human body, which can synthesize and secrete a variety of adipokines and participate in glucolipid metabolism through endocrine, autocrine, and paracrine pathways [1]. In 2016, Romere et al. [2] discovered asprosin as a new adipokine by sequencing and mechanically exploring neonatal premature aging syndrome patients with FBN1 mutations. Asprosin is mainly expressed and secreted by white adipose tissue, which can break down liver glycogen and raise blood glucose levels. Pathologically elevated asprosin levels also are observed in insulin-resistant patients and mice. Besides, asprosin leads to islet β-cell secretion dysfunction and apoptosis [3] and reduces skeletal muscle insulin sensitivity [4] through inflammatory pathways. Recently, Li et al. [5] have found that asprosin mediates liver glycogenolysis by binding to the liver OLFR734 receptor using siRNA interference technology and gene knockout animal model. In addition, asprosin is also an anorectic hormone [6]. e latest research performed by Liu et al. [7] have revealed that asprosin binds to the central OLFR734 receptor and promotes appetite by enhancing olfaction and activating AgRP neuron. Moreover, many clinical studies have observed that serum asprosin levels are increased in patients with metabolic diseases such as obesity, type 2 diabetes, and polycystic ovary syndrome (PCOS) [8][9][10][11]. erefore, asprosin may be a potential biomarker for abnormal glucolipid metabolism related diseases [11].
Acromegaly is a syndrome characterized by excessive growth hormone (GH) secretion, usually caused by pituitary growth hormone cell adenoma. Patients with acromegaly have been in a state of high GH levels. As an important metabolic regulating hormone, growth hormone can promote protein synthesis and lipolysis. erefore, patients with acromegaly usually show reduced fat mass and increased lean body mass. Patients with acromegaly often have abnormal glucolipid metabolism, including insulin resistance and diabetes, which may be related to high GH status, resulting in adipose tissue dysfunction and the imbalance of adipocytokines synthesis and secretion [12]. Several studies have reported serum levels of different adipokines in acromegaly patients and the correlations between adipokines and glucolipid metabolism indexes. Silha et al. [13] and Bolanowski et al. [14] revealed that serum leptin levels decreased compared to controls in acromegaly patients. Fukuda et al. [15] showed that, compared with controls, serum adiponectin levels decreased (4.3 ± 1.8 vs 6.7 ± 1.8 g/ ml, P < 0.001) in acromegaly patients, but another study by Ronchi et al. showed no serum adiponectin levels change was observed in acromegaly patients [16]. In general, serum leptin levels in patients with acromegaly are reduced, while other adipocytokines have been assessed in a few studies with inconsistent conclusions [17].
As a novel identified adipocytokine recently, asprosin has been considered to be a potential biomarker for abnormal glucolipid metabolic disorders such as obesity and type 2 diabetes. However, serum asprosin levels and its association with glucolipid metabolism parameters in acromegaly patients have not been investigated and as far as we know, there is no relevant literature. erefore, this study collected clinical data and laboratory examinations of acromegaly patients, measured their serum asprosin levels by ELASA, and investigated serum asprosin levels in acromegaly patients, as well as its correlation with high GH levels and glycolipid metabolism parameters.  acromegaly patients with 13 receiving  somatostatin analogue treatment and 1 receiving bromocriptine treatment. 1 of 68 acromegaly patients had been  treated with drugs for lowering lipid (Lipitor), but the  specific dosage was unknown and the blood lipid data (LDL-C, HDL-C, TC, and TG) were not available and there were 11 acromegaly patients having received the therapy of diabetes. A total of 121 non-acromegaly controls matched with gender, age, and BMI were included from the physical examination center of Peking Union Medical College during the same period.

Subjects and Methods
is study was approved by the Ethics Committee of the Peking Union Medical College Hospital, Chinese Academy of Medical Sciences (ethics number: SJ-1233). All research subjects signed informed consent before the study.

Serum Asprosin Measurements.
e serum of patients was collected from all subjects in the morning with an empty stomach before surgery in our hospital. 3-4 ml peripheral blood was collected and centrifuged at 3000 r for 10 minutes and then stored in a refrigerator at −80°C. Serum asprosin levels were measured by ELISA kits (USCN Life Science Inc., Wuhan, China, Article no. SEA332Hu) according to the instructions in March 2018. e inter-assay and intra-assay coefficients of variation (CVs) were 12.9% and 6.7%, respectively.

Statistical
Analysis. All data were analyzed using SPSS 25.0 statistical analysis, and all graphs were drawn using GraphPad Prism 8.0.2. e Shapiro-Wilk test and the P-P chart were used to test the normality of data, and some nonnormal distributed data were analyzed after ln conversion.
e independent sample t test and non-parametric Mann-Whitney U test were used to compare all variables between acromegaly patients and controls. Chi-square test was used to compare the proportions of gender and metabolic diseases between acromegaly patients and controls. Spearman bivariate correlation analysis was used to explore the correlation between serum asprosin levels and other variables. e Spearman correlation coefficient determined the statistical significance level. Kruskal-Wallis H test and oneway analysis of variance were used to compare serum asprosin levels among multiple groups. Multiple stepwise linear regression was used to explore the influencing factors of serum asprosin levels, with serum asprosin levels as dependent variable, group (acromegaly assigned a value of 1 and controls assigned a value of 0), gender (male assigned a value of 1 and female assigned a value of 0), age, BMI, systolic blood pressure (SBP), diastolic blood pressure (DBP), TC, TG, LDL-C, HDL-C, fasting blood glucose (FBG), two-hour postprandial blood glucose (2h-Glu), ALT, AST, Cr, and UA as independent variables. Binary logistic regression was used to analyze the risk of acromegaly according to serum asprosin levels tertiles.

Basic Characteristics in Acromegaly Patients and Controls.
e basic characteristics of acromegaly patients are shown in Table 1. e ratio of male to female in acromegaly patients was 74% (29/39), the average age was 41.3 ± 15.0 years, and the average BMI was 25.79 ± 3.52 kg/m 2 . In these patients, the median nadir GH (GH-Nadir) after OGTT was 10.7 ng/ ml, the median area under curve of GH (GH-AUC) after OGTT was 2391.0 ng/ml, and the average IGF-1 SDS was 6.56 ± 2.43. ere was no significant difference in sex, age, BMI, and blood pressure between acromegaly patients and controls, suggesting basic characteristics of two groups were consistent.

Serum Asprosin Levels in Acromegaly Patients and
Controls. As shown in Figure 1(a), serum asprosin levels in patients with acromegaly were significantly lower than controls (2.18 ± 0.86 vs 2.71 ± 0.86 ng/ml, P < 0.001). After grouping the patients in terms of the course of acromegaly, serum asprosin levels in patients with the course of acromegaly ≥5 years were significantly lower in comparison with those with the course of acromegaly <5 years as shown in Figure 1(b) (1.97 ± 0.84 vs 2.44 ± 0.83 ng/ml, P � 0.030). To further explore the association between serum asprosin levels and serum GH levels, patients were divided into high GH-AUC group (GH-AUC ≥2391 ng/ ml) and low GH-AUC group (GH-AUC <2391 ng/ml). Compared to patients with low GH-AUC, serum asprosin levels in patients with high GH-AUC were significantly reduced (1.94 ± 0.91 vs 2.44 ± 0.76 ng/ml, P � 0.022, Figure 1(c)). In addition, compared with low IGF-1 SDS group (IGF-1 SDS <6.56), serum asprosin levels in high IGF-1 SDS group (IGF-1 SDS ≥6.56) were significantly decreased (1.94 ± 0.86 vs 2.42 ± 0.80 ng/ml, P � 0.020, Figure 1(d)). Notably, serum asprosin levels in patients with high GH-Nadir (≥10.7 ng/ml) were lower (1.98 ± 0.97 vs 2.40 ± 0.71 ng/ml) than the low GH-Nadir group (<10.7 ng/ml), though there was no statistical difference in serum asprosin levels between two groups (data was not shown).

Subgroups Analysis of Serum Asprosin Levels in
Acromegaly. Next, serum asprosin levels in different subgroups of acromegaly and controls were explored. As shown in Figure 2(a), serum asprosin levels of women were significantly higher than men in acromegaly patients, as well as controls (2.56 ± 0.69 vs 1.66 ± 0.80 ng/ml and 3.15 ± 0.54 vs 1.87 ± 0.71 ng/ml, both P < 0.001). ere was no difference in serum asprosin levels in acromegaly patients stratified into subgroups by BMI (H ＝ 0.834, P � 0.659, Figure 2(b)). However, serum asprosin levels were different among subgroups in controls (H ＝ 10.852, P � 0.04). After further post hoc pairwise comparisons using Bonferroni correction, serum asprosin levels in obesity were significantly higher than normal in controls (3.08 ± 0.64 ng/ml vs 2.42 ± 0.86 ng/ ml, P � 0.004).
Both acromegaly patients and controls were divided into diabetes (DM), abnormal glucose tolerance (IGT), and normal glucose tolerance (NGT); nevertheless, there International Journal of Endocrinology was no significant difference in serum asprosin levels among subgroups both in acromegaly and controls (acromegaly: F ＝ 0.658, P � 0.521, controls: H ＝ 0.045, P � 0.978, Figure 2(c)). Considering the effect of diabetes treatment on serum asprosin levels, acromegaly patients with diabetes treatment (treatment details about DM in controls were unknown) were further excluded, and still there was no difference in serum asprosin levels among subgroups in acromegaly patients (F ＝ 1.211, P � 0.546).

Multiple Regression Analysis of Serum Asprosin Levels.
Moving on to the associations between serum asprosin levels and other variables, multiple stepwise linear regression analysis was used. Multiple stepwise linear regression analysis (Table 4) revealed acromegaly was an independent influencing factor of serum asprosin levels. Serum asprosin levels were negatively correlated with acromegaly (β ＝ −0.273, P < 0.001, acromegaly assigned to 1, controls assigned to 0), which indicated that serum asprosin levels in acromegaly patients were lower than controls. Additionally, sex, age, BMI, and TC were independent influencing factors of serum asprosin levels. Serum asprosin levels were negatively correlated with gender (β � −0.578, P < 0.001, male assigned to 1, female assigned to 0) and age (β ＝ −0.193, P � 0.004), while they were positively correlated with BMI (β ＝ 0.225, P � 0.001) and TC (β ＝ 0.152, P � 0.025).

Serum Asprosin Levels and the Risk of Acromegaly.
In order to further explore serum asprosin levels of patients with acromegaly, patients with acromegaly and controls were stratified into lowest group, median group, and highest group according to serum asprosin tertiles levels (lowest: <2.12 ng/ml; median: 2.12 ∼ 3.06 ng/ml; highest: ≥3.06 ng/ ml). As shown in Table 5  International Journal of Endocrinology serum asprosin levels had a higher risk of developing acromegaly compared with high serum asprosin levels, suggesting lower serum asprosin levels in patients with acromegaly.

Discussion
Our study was the first to demonstrate that serum asprosin levels in acromegaly patients were significantly lower than controls. At the same time, serum asprosin levels in patients with the acromegaly course ≥5 years, high IGF-1 SDS group, and high GH-AUC were significantly decreased. Negative correlations between serum asprosin levels and the course of acromegaly, IGF-1 SDS, GH-Nadir, and GH-AUC in acromegaly patients were also observed. Regardless of acromegaly, serum asprosin levels in women exceeded those in men. Furthermore, serum asprosin levels in acromegaly patients and total subjects were positively correlated with TC, but not BMI in acromegaly patients.
In this research, we preliminarily showed that serum asprosin levels in acromegaly patients were lower than controls. Meanwhile, serum asprosin levels in patients with the acromegaly course ≥5 years, high IGF-1 SDS, and high GH-AUC were significantly decreased. Besides, negative correlations between serum asprosin levels and the course of acromegaly, IGF-1 SDS, GH-Nadir, and GH-AUC in acromegaly patients were observed. e possible reasons for lower asprosin levels in acromegaly were as follows. Firstly, asprosin secretion is intimately related to blood glucose levels. Romere et al. [2] showed that plasma asprosin levels in mice immediately decreased with the increased blood glucose levels after feeding, while they increased with the decreasing blood glucose after fasting overnight, suggesting that blood glucose might be a regulatory factor to asprosin secretion. Interfering vitro cultured mouse hepatocytes with different concentrations of asprosin, release of glycogen increased with increasing asprosin concentrations. Further studies indicated that asprosin could activate the G proteincoupled receptor-cAMP-PKA pathway, induce rapid decomposition and release of liver glycogen, and increase blood glucose. Besides, Wiecek et al. [22] found that, within 30 minutes after anaerobic exercise, blood glucose levels of women gradually decreased, while serum asprosin levels gradually increased, demonstrating that decreased blood glucose may induce increased secretion of asprosin. It was found that patients with acromegaly did have higher fasting blood glucose than controls in this study, though both within normal range. erefore, increased blood glucose in patients with acromegaly may inhibit asprosin secretion. Secondly, decreased adipose tissue by long-term GH exposure in acromegaly patients exerts a negative impact on asprosin synthesis and secretion. To our knowledge, the effect of GH on adipose tissue is to promote lipid mobilization and oxidation, and body fat composition of acromegaly patients after long-term growth hormone exposure has decreased [12]. Previous study of neurosurgery department in our center demonstrated that patients with acromegaly had lower total adipose tissue weight and visceral fat index (8.73 ± 2.19 vs 10.24 ± 2.45) than patients with nonfunctional pituitary tumors matched by gender, age, and BMI [23]. On the contrary, increased waist circumference, fat mass, and body fat rate of adult growth hormone deficient patients about 7 years after discontinuing growth hormone replacement were observed in our previous study [24]. Furthermore, our results found that acromegaly was an independent influential factor for serum asprosin levels, further suggesting that decreased serum asprosin levels in acromegaly patients were related to long-term high GH levels. As we all know, asprosin is mainly secreted by white adipose tissue; thus, we believe that lower serum asprosin levels in acromegaly patients may be associated with longterm effects of high GH levels. However, further studies are needed to explore how GH affects the synthesis and secretion of asprosin.
Serum asprosin levels in both acromegaly patients and controls showed sexually dimorphic and women were significantly higher than men. Further bivariate correlation analysis and multiple linear regression analysis also showed that serum asprosin levels were negatively correlated with gender (male 1, female 0), which indicated serum asprosin levels in female patients were higher. Importantly, gender is a vital factor in distribution of body adipose tissue. In general, women have a higher proportion of adipose tissue than men, and storages of adipose in female are mainly white adipose tissue in visceral gonad and subcutaneous in the groin, while mainly visceral gonad white adipose tissue in  men [25]. us, higher percent of white adipose tissue in women has more asprosin synthesis and secretion, leading to higher serum asprosin levels. Moreover, more remarkable adipose reduction in acromegaly men may be associated with less asprosin synthesis and secretion. Several researches have demonstrated that decreased fat mass was more obvious in men under long-term GH exposure. Guo et al. [23] revealed that total fat tissue mass and visceral fat index were obviously decreased in acromegaly men with an average course of 7 years (89.3 months) compared with nonfunctional pituitary tumors patients matched by age, sex, and BMI, but no significant changes in women. Additionally, a prospective study by Götherström et al. [26] on adult growth hormone deficiency found that fat mass reduction in men after 5 years of growth hormone replacement was more significant than women. erefore, regardless of acromegaly or not, serum asprosin levels in women are higher than men. In addition, our study showed that serum asprosin levels in obesity were higher than normal in controls and consistent with previous studies on serum asprosin levels in obese patients [6,9,27], and serum asprosin levels were positively correlated with BMI in controls and total subjects in our study. However, there was no correlation between serum asprosin levels and BMI in acromegaly patients, and serum asprosin levels in different BMI groups had no significant difference. It is well known that BMI is a rough indicator to assess the ratio of weight to height and cannot objectively reflect body fat composition [26]. As mentioned above, patients with acromegaly have reduced fat mass and changed body fat composition due to high GH levels, which may lead to no correlation between serum asprosin levels and BMI.
Furthermore, serum asprosin levels in acromegaly patients and total subjects were positively correlated with TC, and multiple linear regression analysis showed that TC was an independent influencing factor of serum asprosin levels. To our knowledge, serum asprosin levels in metabolic diseases patients are correlated with lipid metabolism indexes, which support our results [11]. Serum asprosin levels in PCOS patients were positively correlated with TC (r ＝ 0.389, P � 0.019) [10]. At the same time, serum asprosin levels in diabetic patients were positively correlated with TG, and TG was an independent influencing factor of serum asprosin levels in diabetic patients [8,27]. erefore, serum asprosin levels are associated with lipid metabolism indexes, but not with BMI in acromegaly patients, and its relationship with abnormal glycolipid metabolism in acromegaly needs further researches.
Our study preliminarily explored serum asprosin levels in patients with acromegaly and its correlation with high GH levels and glucolipid metabolism indexes. However, this study had some limitations as follows. Firstly, as an observational study, the causal relationship between serum asprosin levels and other variables remained unclear, and further mechanism researches were needed. Secondly, lacking information like body fat distribution, insulin sensitivity, and so on, it was hard to fully understand the decrease of serum asprosin levels in acromegaly patients. In addition, diabetes treatment details of 7 patients (though accounting for only a percent of 5.8) in controls were unknown, and its effect on serum asprosin levels could not be evaluated.

Conclusion
is study for the first time showed that serum asprosin levels in acromegaly patients were lower than controls and related to long-term high GH levels. In addition, our study enriches adipocytokines spectrum of acromegaly and reveals the relationship between serum asprosin and high GH levels. However, the association between serum asprosin levels and glucolipid metabolism in acromegaly patients requires further researches.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.