Serum asprosin level in different subtypes of polycystic ovary syndrome: a cross-sectional study

23 Objective : Polycystic ovary syndrome (PCOS) can be divided into different subtypes, including 24 insulin resistance (IR) and hyperandrogenism (HA). Asprosin is a novel hormone associated with IR; 25 however, the role of asprosin in women with PCOS has not been investigated. Thus, the aim of this 26 study was to investigate the relationship between serum asprosin levels and PCOS subtypes. 27 Methods: Ninety-three women with PCOS and 77 healthy women as controls were selected for this 28 study. Clinical and laboratory data were compared between the PCOS group and the control group. 29 The PCOS group was further divided into subgroups: 1) women with or without HA (PCOS HA and 30 PCOS NHA, respectively); 2) women with or without IR (PCOS IR and PCOS NIR, respectively). 31 Serum asprosin was measured by ELISA. 32 Results: Serum asprosin levels showed no significant difference between the PCOS and control 33 groups. However, it was significantly lower in the PCOS HA and IR groups compared to the 34 respective PCOS NHA and NIR groups (P < .05). In the PCOS group, serum asprosin was negatively 35 correlated with body mass index, luteinizing hormone, testosterone, basal antral follicles, fasting 36 insulin, Homeostatic Model Assessment of Insulin Resistance, and triglycerides. After adjusting for 37 BMI, the correlations were not significant and asprosin was only positively correlated with prolactin 38 (r = 0.426, P < .001). 39 Conclusions: Our study shows that women with PCOS HA or IR exhibit significantly lower serum 40 asprosin levels compared to controls, and the lower asprosin level directly correlated with PRL level. 41

abnormalities, infertility, hyperandrogenism (HA), polycystic ovarian morphology (PCOM), and 48 metabolic abnormalities. Metabolic abnormalities are often manifested as obesity, insulin resistance 49 (IR), and dyslipidemia [2,3]. PCOS increases the risk for type 2 diabetes mellitus (T2DM), gestational 50 diabetes, and other pregnancy-related complications, cardiovascular events, and endometrial 51 cancer [4]. IR is considered as the major risk factor for the onset of PCOS [5] and 70% of patients with 52 PCOS have shown signs of IR [6]. 53 Asprosin, a recently identified hormone, is secreted by the white adipose tissue (WAT) [7]. It is a 54 140-amino-acid fragment from the C-terminal of profibrillin (encoded by FBN1) and induces the 55 liver to increase the levels of plasma glucose. Previous studies showed that asprosin was 56 pathologically elevated in humans and mice with IR or obesity [7]. The olfactory receptor OLFR734 57 specifically binds with asprosin to modulate hepatic glucose production [8]. Several recent studies 58 have shown that asprosin correlated with obesity in children and adults, T2DM and PCOS [9][10][11][12][13][14][15][16]. 59 However, these results have been inconsistent. Thus, the aim of this study was to explore the potential 60 relationship of asprosin with PCOS in women, especially those with HA or IR. contraceptives, insulin sensitizers, glucocorticoids, and ovulation induction agents were also 75 excluded. The threshold for defining PCOM on ultrasound was the presence of 12 or more follicles 76 measuring 2-9 mm in diameter or an increased ovarian volume (>10 mL) in at least one ovary. The Clinical and laboratory data were collected from electronic medical records (EMR) in our hospital.

126
Women in the PCOS group had significantly longer menstrual cycles than those in the control group 127 (50.87 ± 13.68 vs 29.69 ± 2.98; P ＜ .001). In the PCOS group, BMI was higher and basal antral 128 follicle numbers were more than in the control group (P ＜ .05).

129
Serum asprosin levels in distinct groups 130 As shown in Fig.1A Fig.1B-D). Serum asprosin level in the PCOS HA group was notably lower than in the 134 PCOS NHA group 2.52 (2.06-3.19) vs. 4.20 (2.35-5.79) ng/mL, median (25th-75th), P ＜ .05 135 (Fig.1B). Serum asprosin levels in the PCOS IR group were significantly lower than in the PCOS 136 NIR group 2.46 (2.05-4.30) vs. 3.77 (2.47-7.18) ng/mL, median (25th-75th), P ＜ .05 (Fig.1C). In  Table 3). In addition, asprosin 146 was still positively correlated with PRL (r = 0.456, P = .003) in PCOS NHA subjects. Moreover, there 147 was no correlation between asprosin and other characteristics in PCOS HA subjects. These results 148 indicate that obesity rather than PCOS might be responsible for the difference in asprosin levels. The current study showed that serum asprosin levels were similar between women in PCOS and 152 control groups; however, lower levels were seen in PCOS HA and PCOS IR groups. Further analysis 153 demonstrated that asprosin was positively correlated with PRL, independent of BMI. these results are inconsistent [13,15,24]. We would like to further explore the profiles of asprosin in 163 PCOS subtypes, and HA is one of the most important phenotypes of PCOS. 164 We first compared the serum asprosin in women with PCOS and healthy women in the concurrent 165 period. Serum asprosin was comparable between women with or without PCOS, and it was somewhat 166 lower in the PCOS group. The PCOS group was then divided into different subgroups. Serum asprosin 167 levels were lower in both, PCOS HA and IR groups, which was contrary to our expectations. We then 168 analyzed the probable correlations between asprosin and PCOS. The results showed that asprosin was 169 negatively correlated with IR and HOMA-IR, which contradicts previous studies [7,16,23]. However, 170 after adjusting for BMI, there was a positive correlation only between asprosin and PRL. Therefore, 171 it is likely that obesity rather than PCOS might be responsible for the difference in asprosin levels.

172
There might be some explanations for this. One possibility is that the serum asprosin might be . We did not find any further correlations between asprosin levels and T after adjusting 186 for BMI. The specific mechanism might be beyond our cognition in this limited study and more in-187 depth research is needed.

188
Our study provides significant insights about the correlation between asprosin and PCOS. Three 189 different studies about asprosin and PCOS were recently published [13,15,24]. Chia et al reported 190 that asprosin levels in women with PCOS were similar to those in corresponding controls [24]. 191 However, Murat and Li found that circulating asprosin levels were elevated in women with PCOS 192 compared to those in controls [13,15]. Our results were consistent with former, but contrary to the 193 latter. This diversity might be due to the different sample conditions and effects of PRL. Yet,some 194 limitations in this study should be acknowledged. For example, the study is based on EMR from a 195 single hospital and the sample size was relatively limited. 197 This study shows that women with PCOS HA or IR exhibit significantly lower levels of serum  Not applicable.

221
Competing interests 222 The authors declare that there is no conflict of interest regarding the publication of this paper.

224
We thank Rusong Zhao and Wei Zhou for their statistical analysis. We also thank Wolters Kluwer for 225 providing language polish services.