Correlation of plasma adipokines with endometrial atypical hyperplasia and type I/II endometrial cancer

Abstract The aim of the study was to systematically explore the relationships between various adipokines and risks of endometrial atypical hyperplasia (EAH), type I endometrial cancer (EC), and type II EC. We enrolled 219 patients in this study, including 39 EAH, 87 type I EC, 38 type II EC and 55 control individuals. We subsequently explored the association of adipokine levels and the leptin-to-adiponectin (L/A) ratio with EAH, type I EC, and type II EC. The plasma leptin level and L/A ratio were significantly higher in the EAH group than in the control group (p = 0.012). Leptin, resistin, vaspin, and visfatin levels were significantly higher in the type I EC group; however, the adiponectin level was lower in the type I EC, which resulted in a higher L/A ratio. Notably, the L/A ratio and visfatin level in the type II EC group were significantly higher. Multiple logistic regression analysis revealed that a higher leptin level was significantly associated with a higher EAH risk (p = 0.012). Higher leptin level (p = 0.042) and L/A ratio (p = 0.027) were significantly associated with an increased type I EC risk. By contrast, higher leptin (p = 0.059) and visfatin (p = 0.003) levels, higher L/A ratio (p = 0.033), and lower adiponectin level (p = 0.042) were associated with an increased type II EC risk. We suggested that adipokines are potentially correlated with EAH and EC risks. IMPACT STATEMENT What is already known on this subject? EAH and EC are considered significantly correlated with obesity and the related insulin resistance. Studies reported that some of the adipokines mediate obesity-related EC risk. What do the results of this study add? We systematically explored whether the adipokines produced from adipose tissue, including leptin, adiponectin, resistin, vaspin, and visfatin as well as the L/A ratio were associated with increased EAH, type I EC, and type II EC risks; whether they are independent risk factors for EAH and EC. Moreover, we analysed the underlying roles of these adipokines in EC tumorigenesis and development. What are the implications of these findings for clinical practice and/or further research? This study aimed to systematically explore the relationships between various adipokines and risks of EAH, type I EC, and type II EC, to suggest that adipokine levels correlated with EAH and EC risks, and to analyse the underlying molecular mechanisms linking adipokines with endometrial carcinogenesis. It is helpful to improve our understanding of EC tumorigenesis and development.


Introduction
Endometrial cancer (EC), one of the most common gynecological malignancies among women, is traditionally classified as oestrogen (type I) or nonestrogen (type II) dependent.Endometrial atypical hyperplasia (EAH) is considered the premalignant lesion of type I EC (Capozzi et al. 2022, Costales et al. 2014).Currently, EC is considered significantly correlated with obesity (Zhang et al. 2014) and related insulin resistance.Adipose tissue, an active endocrine organ producing various adipokines, including leptin, adiponectin, resistin, vaspin, and visfatin, is crucial in metabolism and related diseases, such as type 2 diabetes mellitus (T2DM) (Feng et al. 2014) and breast cancer (Ando et al. 2019), and may directly mediate obesity-related EC risk (Ilhan et al. 2015, Nergiz Avcioglu et al. 2015).
Leptin, a regulator of food intake, energy balance, glucose, and lipid metabolism, can reduce peripheral tissue sensitivity to insulin.Increased leptin levels may be an independent risk factor for EC.Adiponectin is negatively correlated with obesity, and low adiponectin levels are closely related to insulin resistance, independent of adiposity (Hanley et al. 2007).Thus, the leptin-to-adiponectin (L/A) ratio might be a surrogate marker for EC.Resistin was initially identified in obese mice, and increased resistin levels have been reported in obesity, T2DM, and obesity-related cancers (e.g.breast and colorectal cancer) (Gharibeh et al. 2010, Wang et al. 2018).However, the functions of circulating resistin levels in EC remain unclear.Vaspin, known as a new adipokine, is secreted from the visceral adipose tissue.Significantly higher vaspin levels were observed in obese individuals and patients with T2DM (Feng et al. 2014).Visfatin, initially identified as a pre-B-cell colony-enhancing factor, is secreted from visceral fat and has a role in obesity and cancer.Patients with EC exhibit significantly higher visfatin levels than do controls (Ilhan et al. 2015, Nergiz Avcioglu et al. 2015).
Although several studies have reported adipokine levels in EC, whether they are independent risk factors for EC and whether they occur in EAH as well remains unclear.Moreover, the roles of these adipokines in EC tumorigenesis and development remain uncertain.In addition, few studies have distinguished the differences in the adipokine levels between type I and II EC considering their contradictory pathogeneses.
In this study, we explored whether the leptin, adiponectin, resistin, vaspin, and visfatin levels as well as the L/A ratio were associated with increased EAH, type I EC, and type II EC risks.

Case and control selection
Data from female patients with newly diagnosed, histologically confirmed EAH and EC (including type I EC and type II EC) between July 2014 and October 2017 from Peking University First Hospital, Beijing, China were included.Control data were selected from among patients who underwent a total hysterectomy for uterine myoma (with normal endometrium) or patients with benign endometrial conditions, such as endometrial polyps and atrophic endometrium, at the same hospital during the same study period.These female participants were on a normal diet and without lipid-lowering drugs, they underwent an oral glucose tolerance test (OGTT), insulin release test, and C-peptide test after an overnight (12 h) fasting period.Serum glucose, insulin, and C-peptide levels were all measured at baseline and 30, 60, and 120 min after a 75-g glucose intake.The research protocol was approved by the study hospital's ethical review committee.

Variable definitions
High TG (hypertriglyceridemia), high TC (hypercholesterolemia), high LDL-C, and low HDL-C levels were defined as TG > 1.7 mmol/L, TC > 5.2 mmol/L, LDL-C > 2.6 mmol/L, and HDL-C < 1.1 mmol/L, respectively (Gitt et al. 2012).Dyslipidemia was defined as the presence of any one of the following four factors: high TC, high TG, high LDL-C, or low HDL-C levels (Bhan et al. 2010).Insulin resistance was assessed using the HOMA-IR calculated according to the formula HOMA-IR ¼ FINS (in lIU/mL) � FBG (in mmol/L)/22.5.In addition, we evaluated the serum glucose and insulin levels after the OGTT and insulin release test at different time intervals, and the modified indices were calculated as follows: HOMA-M y ¼ G y (in lIU/mL) � I y (in mmol/L)/22.5 (where y indicates 30, 60, or 120 min glucose and insulin values from the OGTT and insulin release test) (Morciano et al. 2014).

Statistical analysis
SPSS (version 20; SPSS Inc., Chicago, IL, USA) was used for statistical analysis.Comparisons of variables with a normal distribution were made using Student's t test.The Mann-Whitney U test was used to compare parameters with an abnormal distribution.Correlations between the measures were analysed using Pearson's chi-square and Spearman's correlation tests.To develop the logistic regression models, the plasma levels of leptin, adiponectin, resistin, vaspin, and visfatin as well as the L/A ratio were categorised into tertiles based on the distribution of values in the control group.Differences with p values of <0.05 were considered statistically significant.

General characteristics
We included 39, 87, and 38 patients with incident EAH, type I EC, and type II EC, respectively, along with 55 control individuals; their baseline characteristics are summarised in Table 1.Mean ages were significantly higher in type I and II EC groups than in the control group (54.98 ± 8.44 and 59.63 ± 9.50 vs. 49.47 ± 7.70 years, respectively); thus, more patients with menopausal status were present in type I and II EC groups.The proportion of patients with hypertension or diabetes mellitus was relatively high in the type I EC group.No statistically significant differences were found between the type I EC and control groups in terms of BMI, waist-hip ratio, and proportion of patients with obesity.However, the proportion of patients with obesity was lower in the type II EC group than in the control group (10.53% vs. 36.36%,p < 0.01).Lipid metabolism in type I and II EC groups exhibited no significant differences compared with the control group, except for LDL-C levels being higher in the type I EC group than in the control group (2.85 ± 0.76 vs. 2.56 ± 0.80 mmol/L, p < 0.05).In the OGTT, G 30 , and G 60 levels considerably differed between the type I EC and control groups.Furthermore, the EAH and type I EC groups exhibited higher insulin and C-peptide levels.An analogous relationship existed in the HOMA-M y .

Plasma adipokine levels and L/A ratios in the EAH, type I EC, and type II EC groups
In the present study, we found that leptin, resistin, vaspin, and visfatin levels were significantly higher in the type I EC group than in the control group (p < 0.05, Supplementary Table 1), whereas the adiponectin level was lower in the type I EC group than in the control group (p < 0.05), which resulted in a higher L/A ratio (p < 0.001).In addition, the leptin level (p < 0.05) and L/A ratio (p < 0.01)were relatively high in the EAH group.Notably, the L/A ratio (p < 0.05) and visfatin level (p < 0.01) in the type II EC group were significantly higher than in the control group.

Correlations between the adipokine levels and L/A ratio with related variables in the EAH and type I EC groups
As expected, the plasma leptin level was significantly positively correlated with patient BMI, waist-hip ratio, and insulin resistance measures in the EAH and type I EC groups (Table 2).Plasma adiponectin levels were negatively correlated with insulin resistance measures in the type I EC group, but not in the EAH group.By contrast, the L/A ratio was positively correlated with patients' BMI and insulin resistance measures in both the EAH and type I EC groups.Resistin and visfatin levels exhibited some correlation with blood glucose levels, and vapsin levels were positively correlated with I 60 and HOMA-M 60 in the type I EC group.Regarding lipid levels, adiponectin was negatively correlated with TG levels in the both EAH and type I EC groups and positively correlated with HDL-C in the type I EC group.By contrast, the L/A ratio was positively correlated with TG and negatively with HDL-C levels in the type I EC group.In particular, vaspin levels were negatively correlated with HDL-C levels in the EAH group.

Multiple logistic regression analysis of adipokine levels and L/A ratio for EAH, type I EC, and type II EC risks
The multiple logistic regression analysis was adjusted for all confounders, namely age, BMI, obesity, menopause, hypertension, diabetes mellitus, HOMA-IR, HOMA-M 30 , HOMA-M 60 , and HOMA-M 120 (Supplementary table).Table S2 and Table S3 present the risks of EAH and type I EC, respectively, both in relation to plasma adipokine levels and L/A ratios.A higher leptin level was found to be significantly associated with a higher risk of EAH (odds ratio [OR] 5.49, 95% confidence interval [CI] 1.82-31.70;p ¼ 0.012), and a higher resistin level was close to being correlated with the occurrence of EAH (OR 4.55, 95% CI 0.95-21.89;p ¼ 0.059).Table S3 shows that a higher leptin level (OR 4.10, 95% CI 1.05-16.04;p ¼ 0.042) and L/A ratio (OR 5.08, 95% CI 1.20-21.45,p ¼ 0.027) were significantly associated with an increased risk of type I EC.Notably, no significant associations were found between adiponectin and visfatin levels and the risk of type I EC in the multiple logistic regression analysis even though they appeared to be related to the risk of type I EC in the univariate analysis.

Discussion
Adipose tissue is a known source of various bioactive cytokines termed adipokines, such as leptin, adiponectin, resistin, vaspin, and visfatin.Leptin is crucial in reducing insulin sensitivity and secretion and is a predictor of insulin resistance independent of obesity levels.Adiponectin mediates insulinsensitizing effects by binding to its receptors AdipoR1 and AdipoR2, and lower adiponectin levels are negatively associated with insulin sensitivity (Yadav et al. 2013).A study found that the L/A ratio is a sensitive indicator for insulin resistance and a high L/A ratio is associated with an increased EC risk after adjustments for age, BMI, hypertension, and diabetes mellitus (Ashizawa et al. 2010).In the present study, leptin was correlated with patients' BMI, waist-hip ratio, and insulin resistance markers; adiponectin was negatively correlated with patients' insulin resistance markers; and the L/A ratio was correlated with patients' BMI and measures of insulin resistance in the type I EC group.After adjustments for confounders, a higher leptin level and L/A ratio were associated with an increased type I EC risk, and the OR of the L/A ratio was higher than that of leptin alone, consistent with the findings in the literature (Ashizawa et al. 2010).Furthermore, we found that a higher leptin level and L/A ratio were also present in the EAH group, and a higher leptin level remained associated with an increased EAH risk after adjustment for confounders.The independent associations between leptin and EAH and type I EC risk suggest that leptin may directly participate in endometrial carcinogenesis and might be a critical factor in promoting EAH development into type I EC.
A study found that the resistin level is positively correlated with age, insulin, BMI, waist circumference, body-fat content, and HOMA in people with type 2 diabetic mellitus (Gharibeh et al. 2010).Furthermore, researchers reported that patients with EC also have higher resistin levels, but no significant association between higher resistin levels and age and BMI was observed, which is consistent with the results of our study (Hlavna et al. 2011).In addition, we found that after adjustments for confounders, resistin was not an independent risk factor for type I EC.Notably, no significant difference was found in plasma resistin levels between the EAH and control groups, whereas in the multiple logistic regression analysis, the p was on the threshold of significance after adjustments for confounders.This implies that a higher resistin level may be involved in type I EC development.Several mechanisms were reported for pathophysiologic pathway for resistin to cancer progression.Resistin increases cell proliferation by inducing phosphatidylinositol 3-kinase or it could be act as specific receptor inducers (Ilhan et al. 2015, Tarkowski et al. 2010).
In an animal model of abdominal obesity with T2DM, vaspin improved insulin resistance (Hida et al. 2005).In our study, vaspin was negatively correlated with circulating glucose level and positively correlated with insulin resistance measures in the type I EC group; and the level of vaspin increased in the type I EC group compared with the control group, which was consistent with previous reports Table 2. Correlations between leptin, adiponectin and the L/A ratio with related variables in EAH and type I EC.

Correlations in EAH
Correlations in type I EC Leptin(r) Adiponectin(r) L/A ratio(r) resistin(r) vaspin(r) visfatin(r) Leptin(r) Adiponectin(r) L/A ratio(r) resistin(r) vaspin( r ( Cymbaluk-Płoska et al. 2018).Our study suggests that vaspin may have beneficial effects on insulin resistance, and higher levels of vaspin in patients with obesity and T2DM may be a consequence of compensative response to insulin resistance.However, the plasma levels of vaspin in the EAH group did not increase compared with the control group, indicating no strong correlation between vapsin and EAH.Visfatin mainly expresses in visceral adipose tissue, and its level increases in obesity, T2DM, metabolic syndrome (Chang et al. 2011), andEC (Nergiz Avcioglu et al. 2015).In our study, consistent with the literature, the plasma visfatin level was significantly higher in the type I EC group than in the control group.Fukuhara et al. (Fukuhara et al. 2005) found that visfatin lowered the circulating glucose level in mice.Our study revealed that visfatin was negatively correlated with the circulating glucose level in the type I EC group, although no correlation of visfatin with age, BMI, and insulin resistance measures was found.The L/A ratio and visfatin level were significantly higher in the type II EC group, and in the multiple logistic regression analysis, higher leptin and visfatin levels, a higher L/A ratio, and lower adiponectin level were all risk factors for type II EC. (Setiawan et al. 2013) reported that the risk factors for type I EC and type II EC are similar, observing that obesity and diabetes were associated with type II EC.Specifically, for every 2 kg/m 2 increase in BMI, they found that the OR for type II EC was 1.12 (95% CI 1.09-1.14)and the OR of diabetes for type II EC was 1.53 (95% CI 1.19-1.95),which means that both types share common etiologic pathways.These results support that type II EC is to some extent metabolic.
The molecular mechanisms linking leptin and adiponectin with endometrial carcinogenesis remain unclear; emerging evidence suggests that leptin promotes cell growth and invasiveness through activation of the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (also known as AKT), mitogen-activated protein kinase (MAPK)/extracellular regulated protein kinase, and Janus-activated kinase 2 (JAK2)/signal transducers and activators of transcription 3 (STAT3) pathways (Gao et al. 2009, Liu et al. 2011).For example, leptin can enhance the expression of cyclin D1, which is a crucial cell cycle regulator, through signal transducers and activators of STAT3 and cyclic AMP-responsive element binding protein motifs (Catalano et al. 2009), thereby promoting cell proliferation in EC.By contrast, adiponectin induces apoptosis and suppresses cell proliferation by activating AMPK and inhibiting the PI3K/AKT/mTOR and ERK1/2 signalling pathways via AdipoR1/2 (Zhang et al. 2015).Nakatsuka et al. reported that recombinant vaspin activates AKT and AMPK signalling pathways by binding to glucose-regulated protein (GRP) 78 in obesity associated with metabolic dysfunction, thereby improving glucose and lipid metabolism (Nakatsuka A 2012), which is consistent with a report that vaspin exerts an insulin-sensitizing effect in obesity (Hida et al. 2005).Though a higher vaspin level was not an independent risk factor for type I EC in the present study, the plasma level of vaspin in the type I EC group was indeed higher than in the control group.Given that the patients with EC were always accompanied by metabolic syndrome, we posit that the effect of vaspin may be a consequence of compensative response to metabolic dysfunction.Notably, although adiponectin and vaspin were both beneficial to metabolic syndrome, their expressions in the type I EC group were obviously different.We suggest that the effects of adiponectin and vaspin lie in different processes of tumorigenesis.Vaspin may mainly play its role in improving insulin resistance, causing the circulating level of vaspin in patients with EC patients to complementally increase.By contrast, adiponectin can inhibit the development of cancer by regulating the signalling pathway in tumorigenesis, causing the expression of adiponectin to be restrained in patients with EC.In our study, the plasma level of visfatin was significantly higher in the type I EC group; moreover, visfatin level was found to be associated with type II EC.Ilhan et al. suggested that a higher visfatin level is associated with the risk of myometrial invasion in EC (Ilhan et al. 2015), meaning that it may be a prognostic factor in patients with EC.
In conclusion, although several studies have reported the correlation between some plasma adipokines with EC, we systematically explored the relationship between various adipokines produced from adipose tissue and risks of EAH, type I EC, and type II EC.We demonstrated that adipokine levels in patients with EAH or EC differ from those in a control group, and we suggested that adipokine levels are correlated with EAH and EC risks.However, the molecular mechanisms linking adipokines with endometrial carcinogenesis remain unclear, further studies are required to confirm these findings and the mechanisms underlying the effects.Moreover, a larger study sample is required to confirm the present results.

Table 1 .
General characteristics of study subjects.�� p < 0.01 and ��� p < 0.001; the analysis was done by comparing to the control group.FPG, fasting plasma glucose; FINS, fasting insulins; FCP, fasting C-peptide.