The Relationship Between Hedgehog-interacting Protein Levels and Obesity: A Case Control Study

Background Obesity is an independent risk factor for metabolic disorders including diabetes. The Hedgehog-interacting protein (Hhip) is a negative regulator in tissue remodeling, and inhibits the proliferation of adipocytes and promotes their differentiation. In addition, Hhip was positively associated with diabetes. However, relationship between Hhip and obesity in the human body remain unclear. Methods Participants receiving a physical checkup were recruited. Anthropometric and biochemical data were collected. Serum Hhip levels were determined by ELISA. Subjects were classied into normal-weight, overweight, and obese groups based on their body mass index (BMI). The association between Hhip and obesity was examined by a multivariate linear regression analysis. Results In total, 294 subjects who were either of a normal weight (n=166), overweight (n=90), or obese (n=38) were enrolled. Hhip concentrations were 6.51±4.86, 5.79±4.33, and 3.97±3.4 ng/ml in normal-weight, overweight, and obese groups, respectively (p for trend=0.032). Moreover, the regression analysis showed that BMI (β=-0.144, 95% condence interval (CI)=-0.397~-0.046, p=0.013) was negatively associated with Hhip concentrations after adjusting for sex and age. Being overweight (β=-0.181, 95% CI=-3.311~-0.400, p=0.013) and obese (β=-0.311, 95% CI=-6.393~-2.384, p<0.001) were independently associated with Hhip concentrations after adjusting for sex, age, fasting plasma glucose, the insulin level, and other cardiometabolic risk factors. Conclusions Our results showed that overweight and obese subjects had lower Hhip concentrations than those of a normal weight. Being overweight and obese were negatively associated with Hhip concentrations. Hhip might be a link between obesity and diabetes. diastolic HbA1c, hemoglobin; ALT, alanine aminotransferase; AST, aspartate aminotransferase; eGFR, estimated glomeruli ltration rate; hsCRP, high-sensitivity C-reactive protein; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; HOMA-IR, homeostasis model assessment of insulin resistance.


Introduction
Obesity is recognized as an independent risk factor for the development of many diseases, such as diabetes mellitus, cardiovascular diseases, and even cancer (1-3). The World Health Organization de nes obesity as abnormal or excessive fat accumulation that may impair health (4). In clinical practice, the body mass index (BMI) has been used for diagnosing obesity and being overweight (5). An energy imbalance of more calories being consumed than are expended is the most important cause of obesity, and the consequence is storage of excess energy in adipose tissues which increase in size by hypertrophy and hyperplasia (6, 7).
The Hedgehog (Hh) signaling pathway is known to be an important pathway for the growth, development, and homeostasis of many tissues in animals, especially during embryonic development (8). Recently, Hh signaling was proven to be related to adipose tissue differentiation (9). Activation of Hh signaling inhibits adipocyte differentiation in vitro (9). Targeted activation of Hh signaling suppresses high-fat-diet-induced obesity and improves whole-body glucose tolerance and insulin sensitivity in vivo (10). Because the Hh signaling pathway was reported to be involved in adipogenesis, it was proposed as a potential therapeutic target for metabolic diseases such as type 2 diabetes and obesity (11,13).

Page 3/13
The Hh-interacting protein (Hhip), a membrane glycoprotein, is a negative regulator that attenuates Hh signaling by binding to its ligands (14,15). During 8-day adipocyte differentiation, Hhip messenger RNA and protein expressions peaked at day 6 in 3T3-L1 cells (13). In addition, Hhip messenger RNA expression in adipose tissues was higher in 3-day-old than in 180-day-old pigs (13). Recombinant Hhip treatment promoted 3T3-L1 cell differentiation by upregulating the expression of peroxisome proliferatoractivated receptor γ and glucose transporter 4, and downregulating the expression of the Hh signaling transcription factor, Gli1 (13). We previously reported that the Hhip was positively associated with prediabetes and type 2 diabetes (16). As obesity is closely associated with dysglycemia (1), we explored the relationship between Hhip levels and being overweight/obese in humans in this study.

Participants
This study was approved by the Institutional Review Board of National Cheng Kung University Hospital (ER-104-204) (Tainan, Taiwan), and all participants signed an informed consent form before joining the study. All participants in the study were recruited between January and December 2016 from the Health Examination Center of National Cheng Kung University Hospital.
Blood was sampled at 9AM from all participants after they had fasted for 12 h overnight. Subjects without history of diabetes received oral glucose tolerance test. After fasting blood sampling, subjects were instructed to drink 75 g glucose in 300 ml water within 5 minutes. Two hours after drinking glucose solution (11 am), blood sample was collected again to measure blood glucose level. Those who 1) had any acute or chronic in ammatory disease as determined by a leukocyte count of > 10,000/mm 3 or clinical signs of infection; 2) had any other major diseases, including generalized in ammation or advanced malignant diseases contraindicating this study; 3) were pregnant; 4) had a history of diabetes and were receiving insulin therapy, glucagon like-peptide-1, or oral antidiabetic drugs; 5) were taking drugs that affect glucose homeostasis, such as corticosteroids, thiazides, and so on; 6) had experienced an acute coronary syndrome, cerebrovascular accident, or pancreatitis during the past 3 months; or 7) were taking lipid-lowering medications or antihypertensive drugs were excluded.
We grouped all participants into one of three groups according to the recommendations of the Health Promotion Administration of Taiwan based on their BMI: normal weight (18.5 kg/m 2 < BMI < 24 kg/m 2 ), overweight (BMI ≥ 24 kg/m 2 ), and obese (BMI ≥ 27 kg/m 2 ) (17).

Data Collection
We measured every subject's body height and waist circumference to the nearest 0.1 cm and body weight (BW) to the nearest 0.1 kg. The BMI was de ned as the BW (kg) divided by the body height (m) squared. We asked participants to rest in the supine position in a quiet place to measure the blood pressure between 08:00 and 10:00 while in a fasted status. An appropriate-sized cuff was used for the right upper arm, and the pressure was checked twice at an interval of at least 5 min using a DINAMAP vital signs monitor (model 1846SX; Critikon, Irvine, CA, USA). The hexokinase method (Roche Diagnostic, Mannheim, Germany) was used to measure blood glucose. An enzyme-linked immunosorbent assay (ELISA) (Mercodia AB, Uppsala, Sweden) was used to measure serum insulin levels. A highly sensitive ELISA kit (Immunology Consultants Laboratory, Newberg, OR, USA) was used to determine high-sensitivity Creactive protein. A human Hhip ELISA kit (MyBioSource, San Diego, CA, USA) was used for determining serum Hhip concentrations. The intra-assay coe cient of variation of the ELISA was 5.52% and the interassay coe cient of variation of it was 4.9%. An autoanalyzer (Hitachi 747E; Tokyo, Japan) in the central laboratory of National Cheng Kung University was employed to obtain serum alanine aminotransferase, aspartate aminotransferase, total cholesterol, triglycerides, high-density lipoprotein cholesterol, and lowdensity lipoprotein cholesterol. A high-performance liquid chromatographic method (Tosoh Automated Glycohemoglobin Analyzer; Tokyo, Japan) was used to measure glycated hemoglobin (HbA1c). The estimated glomerular ltration rate (eGFR) was calculated by modi cation of the diet in a renal disease equation. The homeostasis model assessment of insulin resistance was de ned by the formula which is equal to fasting insulin (mU/L) multiplied by fasting plasma glucose (mg/dl) divided by 405 to investigate insulin resistance (18).

Statistical analyses
Data were analyzed using SPSS software (vers. 24.0; SPSS, Chicago, IL, USA). Baseline characteristics are expressed as the mean ± standard deviation (SD) for continuous variables or as a percentage for categorical variables. A one-way analysis of variance (ANOVA) was used to determine any difference in variables among the groups. Chi-square tests were used to analyze differences in categorical variables among the groups. The Bonferroni correction was used for a post-hoc study to see if serum Hhip concentrations differed among the groups. A multivariate linear regression analysis was performed to identify independent variables related to serum Hhip concentrations. The criteria for statistical signi cance was a p-value of < 0.05.

Discussion
To the best of our knowledge, this is the rst human study to explore the relationship between obesity and the Hhip. We found that Hhip levels progressively decreased from the normal-weight and overweight groups to the obese group. In addition, the BMI was negatively associated with serum Hhip concentrations. Moreover, being overweight and obese were negatively associated with serum Hhip concentrations.
According to our previous study, the presence of prediabetes and type 2 diabetes was positively associated with serum Hhip concentrations, while the BMI was not (16). However, the average BMI in the previous study was similar among subjects with normal glucose tolerance (BMI = 22.3 kg/m 2 ), impaired fasting glucose (BMI = 23.6 kg/m 2 ), impaired glucose tolerance (BMI = 23.4 kg/m 2 ), and newly diagnosed diabetes (BMI = 23.3 kg/m 2 ), although the difference reached borderline statistical signi cance (p = 0.049), which may be at risk of a type 1 error, and subjects with obesity might not have been included. It is therefore unknown whether or not being overweight/obese is associated with plasma Hhip concentrations. Wei et al. reported that recombinant Hhip can increase adipocyte differentiation, which results in increased accumulation of lipid droplets in adipocytes by inhibiting the Hh signaling pathway in 3T3-L1 cells, and Hhip messenger RNA expression in adipose tissues was lower in 180-day-old than in 3day-old pigs (13). It was suggested that serum Hhip concentrations may be negatively regulated by differentiated adipose tissues. Once one becomes obese, the production of Hhip should decrease to prevent further adipocyte differentiation. However, the mechanism as to how adipose tissues in uence serum Hhip concentrations remains unclear. To address this hypothesis, further human studies are required.
Hh signaling plays an important role in inhibiting fat formation (11). A previous animal study showed the activation of Hh signaling decreased obesity induced by a high-fat diet in adult mice (10), and a de ciency of Hh signaling in myeloid cells increased the BW of mice (19). In a human study, expression of the Hh signaling transcription factor, Gli1, signi cantly decreased in adipose tissues of insulinsensitive obese subjects compared to lean subjects, which may indicate that Hh signaling decreases in obese humans (20). Circulating Hh ligands and expressions of Hh ligands in adipose tissues increased in obese mice. However, serum Hh ligand levels signi cantly decreased in morbidly obese (BMI > 40 kg/m 2 ) people, even in those with HbA1c > 7%, possibly due to the inhibitory effect of metformin on Hh ligand expression in adipose tissues (19). As the Hhip is a negative regulator that attenuates Hh signaling by binding to Hh ligands, further study is needed to clarify the regulatory architecture.
Cholesterol has been shown to be an endogenous Smoothened activator that being a second messenger that activating Hedgehog signaling pathway (21). Exogenously added cholesterol would activate Hh signaling pathway in vitro (22). Cholesterol is not just necessary but also su cient to activate signaling by the Hh pathway (23). Hh signaling plays an important role in inhibiting fat formation (11) which means that elevated cholesterol level might activate Hh signaling to minimize fat formation which in the same time that Hhip should be downregulated to avoid fat formation. However, there has been no human study to discuss the relationship between cholesterol and Hhip so far. In our study, we found no difference in cholesterol level among three groups. Also, in multivariate linear regression analysis, cholesterol was not an independent factor of serum Hhip concentrations. The relationship among cholesterol level, Hh signaling pathway, and serum Hhip concentrations needs to be evaluated in human study.
There were some limitations in this study. First, this study was designed as a cross-sectional study which did not allow for causal inferences between serum Hhip concentrations and the BMI or obesity. Second, although one study revealed that the Hhip was associated with moderate to severe chronic obstructive pulmonary disease, all of our participants were apparently healthy with no airway symptoms (24). Third, we could not directly measure Hhip expression by adipose tissues. Therefore, we could not be sure whether serum Hhip concentrations were representative of those in adipose tissues. Finally, all study subjects were Taiwanese, and thus our ndings might not be applicable to other ethnicities.
In summary, our results demonstrated that serum Hhip concentrations were negatively associated with the BMI, and obese subjects had lower serum Hhip concentrations than normal-weight subjects. Further research is needed to explore the pathophysiological roles and clinical implications of the Hhip in obesity.  Figure 1 Comparisons of serum concentrations of the Hedgehog-interacting protein (Hhip) in normal-weight, overweight, and obese subjects. Box and whisker plot of serum Hhip concentrations in participants with normal-weight (n = 166), overweight (n = 90), and obese subjects (n = 38). The line inside the box