Introduction

Spexin, also referred to as neuro-peptide Q, is a peptide hormone encoded by SPX gene [1, 2]. The protein sequence of spexin is highly conserved from fish to mammals [3]. Spexin has gained considerable attention for its role in regulating metabolism, energy homeostasis, and pancreatic β-cell function [4].

Spexin is involved in glucose regulation. In an obese mouse model with hyperglycemia, spexin treatment not only reduces body weight but also improves glucose tolerance with reduction of insulin resistance and hemoglobin A1c (HbA1c) levels [5], suggesting that spexin might regulate insulin secretion from β cells and insulin sensitivity of peripheral tissues. Subsequent studies demonstrate that spexin increases insulin-induced glucose uptake in skeletal muscle and liver in a diet-induced insulin resistance model [6, 7]. In rodents and humans, pancreatic islets have spexin expression at the transcript and protein level [8, 9]. Spexin reduces insulin secretion response to glucose in the cultured β-cell line, isolated pig islets, rats, and pigs [10, 11]. In addition, spexin increases cell proliferation in cultured β cells [10]. Consistent with the animal studies, numerous clinical studies explore the association of spexin with diabetes. Gu et al found, for the first time, that spexin levels decrease in type 2 diabetes mellitus (T2DM) and are correlated with fasting blood glucose (FBG), HbA1c, triglyceride (TG), and low-density lipoprotein cholesterol (LDL) [9]. Karaca et al. further found that spexin levels are lower in the lean type 1 diabetes mellitus (T1DM) and are not correlated with glycemic parameters, lipids, BMI, cortisol levels, and thyroid-stimulating hormone [12]. However, the correlation between spexin and diabetes might exist only in adults. Spexin is not significantly correlated with body composition, fitness, or blood biochemical measurements in adolescents [13]. Spexin levels are also influenced by the presence of gestational diabetes [14, 15]. Furthermore, exercise training increases spexin levels of T2DM with a concomitant improvement in metabolic profile [16, 17].

Reduced β-cell function and/or mass contributes to T2DM pathogenesis [18]. However, the relationship between serum spexin levels and β-cell function is unknown. Furthermore, both first-degree relatives (FDRs) of patients with T2DM and prediabetes subjects are at high risk of diabetes [19, 20]. Thus, the present study aims to investigate the circulating spexin levels in a group of patients consisting FDRs, prediabetes, newly diagnosed T2DM, and established T2DM, and the association of spexin with β-cell function.

Methods

Study subjects

A total of 462 subjects were recruited in this study, including 106 established T2DM, 80 newly diagnosed T2DM, 115 subjects with impaired glucose regulation (IGR, including both IFG and IGT), and 161 subjects with normal glucose tolerance (NGT, including 106 FDRs and 52 healthy controls). The minimal sample size was calculated according to the report from Peng et al. [6]. We supposed that the mean of spexin levels was significantly different between healthy subjects and one of other four groups. The minimal required sample size in each group was assessed according to the difference of spexin levels between the T2DM and NGT groups (mean difference = 2.13, SD1 = 1.54, SD2 = 1.07) in the reference. The criteria of type I error and type II error were set at α = 0.05/4 = 0.125 and β = 0.10, respectively. We found that at least 12 subjects need to be observed in each group.

The diagnoses of NGT, IGR, and T2DM were based on the WHO 1999 criteria as listed in Table1

Table 1 WHO criteria used to define the different groups in this study
Full size table

All of 106 established T2DM cases were inpatients taking oral hypoglycemic drugs and receiving insulin therapy, who were not suitable for a 75-g oral glucose tolerance test (OGTT). The other 356 subjects without history of diabetes or oral hypotensive, hypolipidemic, anti-diabetic, or other medications known to affect glycolipid metabolism were performed an OGTT. The following exclusion criteria were used in this study: acute or chronic inflammatory disease, heart, liver or renal failure, cancer, or pregnancy. The study was approved by the human research ethics committee of the hospital, which conforms to the provision of the declaration of Helsinki (as revised in Fortaleza, Brazil, October 2013). All participants provided written informed consent before participating in this study.

Clinical measurements

Participants’ weight (in light clothing) and height (without shoes) were measured, and the body mass index (BMI) was calculated as weight in kilograms divided by the square of the height in meters. All blood samples were collected from the antecubital vein following 8–12 h of overnight fasting. An oral glucose tolerance test (OGTT) was conducted on 356 participants (excluding 106 established T2DM cases), and blood samples were obtained after 2 h. After clotting, the serum samples were separated from the blood specimens via centrifugation for 15 min at 1000 g and stored in aliquots at − 80 °C until spexin analysis. The FPG and 2-h postprandial glucose (2hPG) were measured using an automatic biochemical analyzer (DXC 800; Beckman coulter, Brea, CA, USA). The fasting insulin (FINS), fasting c-peptide, and 2-h postprandial insulin (2hINS) levels were determined using a Roche cobas e602 immunoassay analyzer and electrochemiluminescence immunoassay kit (Roche diagnostics, Indianapolis, IN, USA). The HbA1c was measured using high-pressure liquid chromatography (HLC-723G7; Tosoh, Tokyo, Japan).

The homeostasis model assessment of insulin resistance (HOMA2-IR) and β-cell function (HOMA2-β) was calculated from fasting plasma glucose and insulin (c-peptide in established T2DM subjects) concentrations by using the updated HOMA2 calculator (version 2.2.4; https://www.dtu.ox.ac.uk/). The Stumvoll first (1st) phase insulin secretion indices were calculated as 2032 + 4.681 × 6.965 × FINS (mU/L) − 135.0 × 2hPG (mmol/L) + 0.995 × 6.965 × 2hINS (mU/L) + 27.99 × body mass index (kg/m2) − 269.1 × FPG (mmol/L). The Stumvoll second (2nd) phase insulin secretion indices were calculated as 277 + 0.800 × 6.965 × FINS (mU/L) − 42.79 × 2hPG (mmol/L) + 0.321 × 6.965 × 2hINS (mU/L) + 5.338 × body mass index (kg/m2).

Serum spexin measurements

Serum spexin levels were measured using an ELISA kit (Catalog # EH4349, Fine biotech co., Wuhan, Hubei, China) with an intra- and inter-assay CV of < 8% and < 10%, respectively. This assay has high sensitivity and excellent specificity for detection of spexin. No significant cross-reactivity or interference between spexin and galanin, kisspeptin, insulin, and c-peptide (all 25 ng/ml) was observed.

Statistical analysis

The physical parameters and biochemistry indices were compared among the five groups. The distribution of continuous variables was assessed using the Kolmogorov–Smirnov test. Normally distributed data are shown as the mean ± standard deviation and non-normally distributed data are expressed as medians with the interquartile ranges. The right-skewed distributed data were log10-transformed before analysis. Comparisons among five groups of participants were performed by one-way ANOVA with fisher’s post hoc test, and then, multiple comparisons were performed using least-significant-difference t test or Dunnett t test to compare the differences between non-FDRs NGT with other four groups. To adjust covariate such as age, sex, and BMI at group comparisons, analysis of covariance was performed using the general linear model procedure. Pearson’s bivariate correlation and partial correlation analyses were used to explore the correlations between spexin and the clinical parameters. Multivariate linear regression analysis was used to examine the independent association of serum spexin and other parameters. The five groups of subjects which were transformed into dummy variables, as well as the variables that correlated significantly with spexin were selected to enter regression. (According to the collinearity diagnostic in liner regression, Supplementary Tables S1 and S2, the FPG was excluded because of effect of multicollinearity.) IBM SPSS (version 25.0) was used for all statistical analyses. GraphPad prism 7 was used for figures. A two-sided P value of < 0.05 was considered as significant.

Results

Clinical characteristics of the subjects

Table 2 shows the demographic, clinical, and metabolic characteristics of the study population stratified into five groups based on glycemic status. There was no difference in any characteristics between healthy and FDRs subjects. Age, sex, and BMI were comparable among NGT and IGR subjects. Established T2DM subjects were older than other groups (all P < 0.05). Both newly diagnosed and established T2DM subjects had higher BMI (both P < 0.05) compared with the healthy subjects. In addition, OGTT-2hPG, FINS, OGTT-2hINS, and HOMA2-IR index were higher in IGR subjects than those in healthy subjects (all P < 0.05). The newly diagnosed T2DM subjects had lower indices of HOMA2-β, and Stumvoll first- and second-phase insulin secretion, compared with healthy subjects (P < 0.05). Furthermore, established T2DM subjects had a decreased HOMA2-β index compared with healthy subjects (P < 0.05) and further increased HOMA2-IR index compared with newly diagnosed T2DM subjects (P < 0.05).

Table 2 Clinical characteristics of the NGT, FDR, IGR, newly diagnosed T2DM, and established T2DM participants
Full size table

Serum spexin levels

In a total of 462 subjects, serum levels of spexin ranged from 44.32 to 596.38 pg/ml (Fig. 1). There was a difference in spexin levels between male and female in all subjects (P < 0.001), and established T2DM subjects (P < 0.05) (supplementary Fig. S1). Bivariate correlation analysis found that spexin levels had a negative correlation with age (r =  − 0.338, P < 0.001) in all subjects, and a positive correlation with BMI (r = 0.241, P < 0.01) in NGT subjects (Table 3). After adjustment for age and sex, serum spexin levels were positively correlated with BMI in all subjects (r = 0.131, P < 0.01) and NGT subjects (r = 0.223, P < 0.01) (Table 3).

Fig. 1
figure 1

Serum spexin concentrations [Median (min to max)] in subjects with NGT (n = 167) (including 52 healthy subjects and 109 FDRs), IGR (n = 115), newly diagnosed T2DM (n = 80), and established T2DM (n = 106). The analysis was performed using one-way ANOVA and covariance analyses. Further multiple comparisons were performed using least-significant-difference t test. The spexin levels were elevated in the FDR and IGR, and decreased in established T2DM, compared with the healthy subjects (**P < 0.001, unadjusted; $P < 0.05, $$P < 0.001, after age and sex adjustment; #P < 0.05, ##P < 0.05, after age, sex, and BMI adjustment)

Full size image
Table 3 Correlations of serum spexin levels with clinical characteristics were analyzed using Pearson’s bivariate correlation and partial correlation analyses
Full size table

Interestingly, serum spexin levels represented as a bell-shaped curve along the glycemic continuum (Fig. 1). Compared to healthy subjects [200.23 pg/ml (160.32, 275.65)], spexin levels increased in FDRs [235.53 pg/ml (185.28, 293.95)] and IGR [239.79 pg/ml (191.52, 301.69)] subjects, went to the similar level in newly diagnosed T2DM subjects [224.68 pg/ml (187.37, 279.74)], and then decreased significantly in established T2DM subjects [100.11 pg/ml (78.50, 137.34)] (P < 0.001). After age and sex adjustment, the spexin levels were still higher in subjects of FDRs (P < 0.05) and IGT (P < 0.05), and lower in established T2DM subjects (P < 0.001), compared to those of healthy subjects. After further adjustment for age, sex, and BMI, the spexin levels remained higher (P < 0.05) in IGR subjects, and lower in established T2DM subjects (P < 0.001), while it was not significantly higher in FDRs subjects, compared with those of healthy subjects.

Correlation between serum spexin levels and glucose-related variables

A bivariate correlation analysis was conducted between spexin levels and variables related to glucose metabolism in a total of 462 subjects (Table 3 and Fig. 2). The spexin levels were negatively correlated with FPG (r =  − 0.444, P < 0.001), HbA1c (r =  − 0.495, P < 0.001), and HOMA2-IR (r =  − 0.232, P < 0.001) in all subjects. Conversely, the spexin levels were positively correlated with HOMA2-β (r = 0.343, P < 0.001). A partial correlation analysis showed that the spexin levels were still negatively correlated with FPG (r =  − 0.355, P < 0.001), HbA1c (r =  − 0.379, P < 0.001), and HOMA2-IR (r =  − 0.225, P < 0.001), and positively correlated with HOMA2-β (r = 0.245, P < 0.001) after adjusting for age, sex, and BMI.

Fig. 2
figure 2

Correlation between serum spexin levels and glucose-related variables was analyzed using Pearson’s bivariate correlation. A Positive correlations of spexin levels (lg transformed) with HOMA2-β (lg transformed) in 462 subjects. B–C Positive correlation of spexin levels (lg transformed) with HOMA2-β (lg transformed) (B) HOMA2-IR (lg transformed) (C) in established diabetes subjects. D–F Positive correlation of spexin levels (lg transformed) with HOMA2-β (lg transformed) (D), Stumvoll first- E and second-phase F insulin secretion indices in NGT subjects

Full size image

Furthermore, bivariate correlation analysis in each group was conducted. In NGT subjects, spexin levels were positively correlated with HOMA2-β, Stumvoll first- and second-phase insulin secretion indices. This correlation was still significant after adjusting for age and sex. In established T2DM subjects, spexin levels were positively correlated with fasting c-peptide, HOMA2-IR, and HOMA2-β. This correlation remained significant (P < 0.001) after age, sex, and BMI adjustments.

Independent association between serum spexin levels and indices of β-cell function

In correlation matrix analysis, we found a high correlation between FPG and HbA1c (Supplementary Table S1). A multivariate linear regression analysis was performed to determine which indices were independently associated with spexin levels (Table 3). The analysis involved glycemic status and the variables significantly correlated with serum spexin, including age, sex, BMI, FINS, HbA1c, HOMA2-IR, and HOMA2-β. The tolerance and variance inflation factors (VIF) of variables demonstrated that there was little multicollinearity in all variables (Supplementary Table S2). Established diabetes and HOMA2-β were found to be independently associated with serum spexin levels, while FDRs or IGR were not independently associated with elevated spexin levels after HOMA2-β adjustment.

Discussion

In this study, we described a profile of serum spexin levels along the glycemic continuum, from normoglycemia to T2DM. We demonstrated that serum spexin levels were higher in subjects of FDRs and IGR, while lower in subjects of established T2DM, compared with healthy subjects. Correlation analysis showed that serum spexin was correlated with variables related to glucose metabolism. Most importantly, serum spexin was independently associated with β-cell function and T2DM progression on fully adjusted analyses. Thus, serum spexin could provide insight into pathophysiology of diabetes, likely through its implications for β-cell function.

Obesity is a triggering factor for diabetes associated with insulin resistance [21]. A large body of evidence suggests a significant association between spexin levels and body weight status. Spexin-knockout zebrafish show higher food intake than wild-type controls and have altered expression of major appetite-regulating neuro-peptides [22]. Consistently, spexin injection induces anorexia in fish and reduced body weight gain in mice [22,23,24]. Clinical studies further show that circulating spexin levels are lower in obese than non-obese individuals in prepubertal children, adolescents, and adults [24,25,26,27]. Circulating spexin levels increase following weight loss [28]. Thus, in our study, we have fully considered the potential effects of body weight on these associations by adjusting the BMI during the correlation and regression analysis.

We found that spexin levels decreased in established T2DM which is in line with previous studies [9]. However, there is a disparity in the spexin levels in newly diagnosed T2DM [6]. We made a comparison and found that the characteristics of T2DM subjects in Gu’s study were more like those of the established T2DM in our study (shown in Table 1), as obtained from the values of HbA1c, FBG, and HOMA-IR. In our study, the subjects of the newly diagnosed diabetes were recognized from community screening and most of them had no obvious symptom of diabetes. So, one possibility is our newly diagnosed T2DM were freshly onset diabetes. In this regard, this study precisely described the spexin levels of different glycemic status toward T2DM. In addition, some subjects of newly diagnosed T2DM in our study had a family history of diabetes. We found that FDR had higher spexin levels than healthy subjects (shown in Fig. 1). Thus, another possibility is that the genetic background of our newly diagnosed T2DM might contribute to elevating spexin concentration of newly diagnosed diabetes to a level comparable to that of healthy subjects (Table 4).

Table 4 A multivariate linear regression of the variables independently associated with serum spexin levels
Full size table

The mRNA or protein expression levels of spexin in human [9], pigs [10], rats [8], mice [29], and goldfish [30] have been analyzed, and the results show a wild range of tissue distribution of spexin. In rodents and humans, spexin has been detected in tissues, including hypothalamus, hippocampus, cerebral cortex, stomach, small intestine, pancreas, liver, kidney, thyroid, ovary, testis, and adrenal gland [8, 9]. Subsequent studies are trying to clarify the tissue specific enrichment of spexin. Analyses based on integration of various omics technologies show that mRNA expression of spexin is highest in brain, pancreas (http://tiger.bsc.es), and adipose (https://www.proteinatlas.org) [31]. Immunohistochemistry analysis further shows that spexin has a colocalization with insulin in pancreas, suggesting that islet β cells produce spexin [10]. Thus, it can be speculated that the circulating levels of spexin are closely related with the expression of spexin in adipose and/or islet β cells. Actually, in obese human subjects, the circulating levels of spexin decrease and the mRNA levels of spexin are down-regulated in the omental and subcutaneous fat [32, 33]. In this study, we found that serum spexin was independently associated with T2DM. It is possible that the decreased circulating levels of spexin in established T2DM are due to the loss of β-cell mass in these subjects.

This study, for the first time, reported the clinical relevance of serum spexin with β-cell function. The indices of HOMA2 model and Stumvoll are used to measure β-cell function. HOMA is calculated using fasting blood concentrations of glucose and insulin to estimate the steady-state β-cell function. HOMA2 indices are more accurate representation of metabolic process than HOMA1 indices because calculation of HOMA2 indices has considered some physiological adjustments based on HOMA1 model [34]. Stumvoll index is assessed as an oral glucose tolerance test-based measure of β-cell function. Compared with hyperglycemic clamp method (the gold standard), Stumvoll index is a simpler, easier-to-use, and relatively accurate measurement of β-cell function in clinical and large epidemiological studies [35]. We found that spexin levels were independently and positively associated with HOMA-β. In NGT subjects, we also found a positive correlation between spexin levels and Stumvoll first- and second-phase insulin secretion.

There are several limitations of this study, including a relatively limited sample size. One limitation is that this study only detected the spexin levels under fasting condition. The spexin levels might be influenced by glucose ingestion during OGTT [9, 13, 36]. The association between spexin and blood glucose under different glycemic status should be measured in the future. Another limitation is the study that did not address the cause–effect relationship between spexin and β-cell dysfunction as well as diabetes progression. Further prospective studies are warranted to determine whether decreased serum spexin is causally related to obesity and diabetes or a concomitant phenomenon of these diseases.

In summary, we infer that spexin may be an important substance in the development of hyperglycemia and islet β-cell functional failure in patients with T2DM. Therefore, future studies will be critical for understanding the relationship between spexin and T2DM, and studies of spexin should focus on the effect of spexin on β-cell function.