Circulating endotoxin is correlated with fatty acid binding protein 2 and leukocyte nuclear factor-κ

Circulating endotoxin is correlated with fatty acid binding protein 2 and leukocyte nuclear factor-κB activation in menopausal women

Abstract

Circulating endotoxin is mainly derived from gut microbiome and is frequently increased after meals. However, the interplay between endotoxemia and innate immune cells in systemic inflammation remains unclear. The aim was to investigate the role of endotoxemia in systemic inflammation, focusing on the immune responses in innate immune cells. Menopausal women (n = 174) were divided into 2 groups. Blood samples were collected from the participants of group 1 (n = 80), who fasted for at least 8 h, to analyze the correlation between endotoxin levels, fatty acid-binding protein 2 (FABP2), markers of systemic inflammation, and nuclear factor-κB (NF-κB) activation in leukocytes. The participants of group 2 (n = 94) fasted for at least 8 h and received oral glucose tolerance test (OGTT). Their fasting and post-OGTT blood samples were also collected to evaluate the impact of postprandial endotoxemia on the expression of NF-κB target genes in leucocytes. The study revealed that circulating endotoxin level was significantly associated with FABP2 level and NF-κB activation in leukocytes after fasting and within a 2 h-OGTT period. Variation in endotoxin levels (about 1 EU/ml) was sufficient to modulate NF-κB activation in leukocytes. These results provide clues for understanding the effects of postprandial endotoxemia on immune cell activation and inflammation induction. Our findings suggest that an increase in intestinal barrier permeability might be linked to endotoxemia-induced systemic inflammation in menopausal women.

Introduction

Lipopolysaccharides (LPS) are surface glycolipids produced by most Gram-negative bacteria. LPS are endotoxins consisting of three domains: a lipid A, an O-antigen, and a core polysaccharide1. Host innate immune cells express toll-like receptor 4 (TLR4) that recognizes LPS. The binding of LPS to TLR4 activates TLR4/myeloid differentiation factor 2 complex, and in turn, activates NF-κB signaling and its target genes2,3 to trigger subsequent inflammatory responses. Thus, elevated circulating LPS levels is implicated in immune-mediated processes that are detrimental to human health4.

LPS molecules are abundant in the gastrointestinal tract where they are produced by gut bacteria. Aside from absorbing essential nutrients (e.g., electrolytes and water), the complex structure of the intestinal barrier plays an important role in restricting the absorption of intraluminal toxins (e.g., LPSs and foreign antigens) and enteric flora5 into the systemic circulation. Therefore, a compromised barrier integrity may lead to a leaky gut, allowing bacteria-derived molecules to enter the systemic circulation and induce inflammatory signaling cascades6.

Endotoxemia refers to the presence of endotoxins in the systemic circulation and has been strongly associated with meal composition7, gut microbiome, and intestinal barrier function8. Consumption of energy-dense meals (e.g., high-fat diets)7,9,10,11 frequently causes transient increases in circulating LPS by promoting LPS absorption from gut8,12. This has been observed in patients with obesity, diabetes, and other systemic inflammatory diseases13. In innate immunity, activation of the nuclear factor-kappa-B (NF-κB) signaling pathway in inflammatory cells ultimately leads to the release of various inflammatory cytokines14.

Postprandial endotoxemia has been considered as a significant contributor to low-grade inflammation and the development of chronic systemic diseases, including diabetes mellitus14,15. This is supported by the experiments which showed increased plasma levels of inflammatory factors, such as tumor necrosis factor-α (TNF-α), in human volunteers intravenously injected with LPS, with dose as low as 60 pg/kg body weight9,16. However, several studies reported that elevated postprandial endotoxins in humans might not necessarily increase plasma TNF-α9,17,18. For instance, one study reported that postprandial increase in TNF-α and IL-1β was not a result of endotoxemia-induced stimulation of monocytes and dendritic cells, but possibly a result of tissue-level inflammation induced by high-fat meal18. Another study found that most of the inflammatory cytokines were not consistently elevated after a high-fat meal19. These data contradict other studies supporting the inflammatory effects mediated by circulating endotoxins1,2,3,20,21,22,23. Thus, the role of endotoxemia in systemic inflammation needs to be clarified.

The positive association between menopause, gut permeability, and inflammation has been addressed. After menopause, estrogen level drops sharply, becoming a key trigger of the gut microbiome remodeling. The decrease in estrogen level significantly alters the richness and diversity of gut microbes, resulting in an imbalance in the ratio of Firmicutes to Bacteroides. In animal models, menopause leads to gut barrier dysfunction and increased gut permeability24. As a result, microbe enter into the subepithelial space, and activate immune cells to produce proinflammatory cytokines.

Fatty acid binding protein 2 (FABP2) is expressed by enterocytes. Gut barrier dysfunction causes the release of FABP2 into circulation. Therefore, FABP2 is used as an indicator of increased gut permeability. Gut permeability increases during menopause transition, as evidenced by increased blood FABP2 level, and greater gut permeability is associated with more inflammation25. This association occurs particularly in the individuals with glucose metabolic dysfunction. The aim of the present study was to investigate among the menopausal women the role of endotoxemia in systemic inflammation, with a focus on innate immune cells. We hypothesized that endotoxemia dynamically modulates TLR4 responsiveness in leukocytes, contributing to NF-κB activation and inflammation induction.

Results

Demographic and laboratory data in healthy menopausal women

The participants were grouped and the study was carried out as depicted in the Fig. 1. We obtained blood samples from the participants of the first group (80 menopausal women), who fasted for 8 h, and examined the association between endotoxin levels, FABP2 levels, and NF-κB activation profiles. Table 1 showed the baseline demographic and laboratory data of the participants in the first group. The mean age of this group was 55 (55–60) years. The plasma levels of FABP2 and endotoxin were 4.43±1.95 ng/ml and 2.73±3.07 EU/ml, respectively. The average number of WBC was 5.77 × 103/µl.

Table 1 Baseline characteristics of participating healthy menopausal women (n = 80).

Full size table

Correlations between endotoxin and NF-κB activation in the menopausal women of the first group

Circulating endotoxin is mainly derived from the gut microbiome1. FABP2 is a marker of gut permeability26,27. We found that the plasma FABP2 was positively correlated with endotoxin level (r = 0.362, P = 0.001) in the menopausal women of the first group (Fig. 2A). We selected several genes known to be NF-κB’s targets, or have putative κB binding sites28, and examined the correlation between their expression in leucocytes and circulating endotoxin level (Fig. 2B and G). We found that endotoxin level was positively correlated with the expression of TLR4 (r = 0.4718, P < 0.0001), NFKBIA (r = 0.3734, P = 0.0006), and NFBK1 (r = 0.4920, P < 0.0001) mRNA. Moreover, endotoxin level was significantly correlated with transcript levels of TNFA (r = 0.4415, P < 0.0001), IL1B (r = 0.5808, P < 0.0001) and IL6 (r = 0.311, P = 0.005), and plasma levels of TNF-α (r = 0.3484, P = 0.0015), IL-1β (r = 0.3872, P = 0.0004) and IL-6 (r = 0.3379, P = 0.0022). Furthermore, we found significant correlation between the expression of cytokines including TNF-α (r = 0.3060, P = 0.0067), IL-1β (r = 0.4468, P < 0.0001), and IL-6 (r = 0.2319, P = 0.0385). These results suggest that circulating endotoxin was highly correlated to the NF-κB signaling activity in the menopausal women.

Demographic and laboratory data in menopausal women receiving OGTT

Absorption of endotoxin from gut increases with food intake after fasting. Accordingly, we collected blood samples from the second group of participants (94 menopausal women) (Fig. 1), who fasted for at least 8 h. Then, these participants received 75 g glucose for the OGTT and their blood samples were collected again 2 h after. The mean age of this group was 53 ± 6.89 years. Their plasma level of FABP2 and endotoxin before and after OGTT were determined. In the samples before OGTT, the correlation between endotoxin and triglycerides was significant (r = 0.2897, P = 0.0046) (Supplementary Fig. S1). In the samples after OGTT, we also found significant correlation between plasma FABP2 levels and endotoxin (r = 0.3864, P = 0.0001), delta endotoxin (r = 0.3864, P = 0.0001), and triglycerides (r = 0.2794, P = 0.0064), respectively. Glucose uptake seemed not to affect the plasma level of endotoxins, TNF-α, and IL-1β significantly (Supplementary Fig. S2), which was consistent with previous reports7,29,30.

Low dose of LPS activation of NF-κB in leukocytes

To determine the minimal dose of endotoxin that could activate TLR4/NF-κB signaling in leukocytes, the fasting blood samples from 30 participants in the second group were used in the ex vivo LPS-treatment experiment (n = 6/each treatment). The fasting blood samples were treated with various doses of LPS (from 0.01 ng/ml to 1 ng/ml) for two hours. We found that LPS dose-dependently activated NF-κB in leukocytes (Fig. 3). LPS as low as 0.1 ng/ml significantly activated the expression of NFKB1, NFKBIA, TNFA, and IL1B. Because the potency of endotoxin depends on a variety of factors, such as polysaccharide chain length and bacterial source31, endotoxin unit (EU) is preferred over weight in measuring the activity of endotoxin. In general, 10 EU/ml solution contains approximately 1 ng/ml endotoxin32. Accordingly, 0.1 ng/ml of LPS is about 1 EU/ml of LPS. Our data suggested that 1 EU/ml of LPS was the threshold dose for activating NF-κB signaling in leukocytes.

Effect of variation in endotoxemia on NF-κB activation and cytokine expression

The 2-h postprandial endotoxin levels were subtracted by levels during fasting to generate the delta endotoxin values (Δ = 2-h postprandial levels – fasting baseline value) to interpret the impact of dynamic change in plasma endotoxin on NF-kB signaling. Subsequent analyses showed a significant correlation between plasma FABP2 level and delta endotoxin value (r = 0.3864, P = 0.0001) (Fig. 4). Based on the data that 1 EU/ml of LPS was the threshold dose for activating NF-κB signaling in leukocytes, we then stratified the participants, who received OGTT, into 3 groups according to their delta endotoxin values: ≥ 1 EU/ml (Group A), > −1 EU/mL and < 1 EU/ml (Group B), and ≤ − 1 EU/ml (Group C). Table 2 showed the demographic and laboratory data in these three groups. There were no significant differences in their baseline parameters, including age and BMI. Figure 4 showed that the levels of FABP2 and endotoxin were significantly higher in group A and group C compared to those of group B. Meanwhile, the levels of TNF-α and IL-1β were also higher in group A and group C than in group B. Next, we examined the change in NF-κB signaling before and after OGTT of samples from groups A, B, and C using paired t-test. In Group A, NFKB1 (mean ± SD, 0.010 ± 0.006 (fasting) vs. 0.013 ± 0.008 (2 h-OGTT)), TNFA (0.006 ± 0.003 (fasting) vs. 0.008 ± 0.003 (2 h-OGTT)), and IL1B(0.246 ± 0.130 (fasting) vs. 0.304 ± 0.181 (2 h-OGTT)) expression was significantly upregulated (Fig. 5A,C), and the plasma TNF-α level (4.20 ± 2.24 (fasting) vs. 5.01 ± 2.84 (2 h-OGTT) pg/ml) was also significantly increased (Fig. 5D). In contrast, in group C, TNF-α (4.99 ± 2.37 (fasting) vs. 3.72 ± 1.70 (2 h-OGTT) pg/ml) and IL-1β(1.83 ± 1.26 (fasting) vs. 1.08 ± 0.82 (2 h-OGTT) pg/ml) were significantly decreased 2 h after OGTT (Fig. 5E,F), but the expression of NFKB1, TNFA, and IL1B was not significantly downregulated. These results suggested that the absolute delta endotoxin value ≥ 1 EU/ml was required to activate TLR4/NF-κB signaling in leukocytes.

Discussion

The gastrointestinal tract is a large reservoir of gut microbiota, which serves as a main source of endotoxins produced by Gram-negative bacteria33. Intestinal barrier function is critical for normal homeostasis of the gut and systemic circulation. Gut barrier breakdown or dysfunction may lead to increased translocation of LPS into the systemic circulation via transepithelial/transcellular and paracellular transport pathways34,35,36. FABP2 is mainly expressed in intestinal epithelial cells and is released into the circulation following intestinal mucosal injury37,38. After meals, more or less, endotoxins cross the intestinal barrier with nutrition into blood stream, so called endotoxemia. In human, most of the endotoxins that cross the gut barrier are directly transported via portal vein into the liver, where most of the endotoxins are detoxified through deacetylation. Thus, intestinal barrier permeability and hepatic metabolism may determine the circulating endotoxin levels35,36,39. While direct infusion of LPS in human has been shown to induce consistent innate immune system response23, and it is well established that increased circulating endotoxins may trigger systemic inflammation and pathological outcomes, the putative provocative role of postprandial endotoxemia in inflammation remains to be investigated.

In the current study, our analyses reveal the variation in the fasting plasma endotoxin level among 80 menopausal women. Several lines of evidence have revealed that the circulating endotoxin level varies within a narrow range, from less than 1 EU/ml to a few EU/ml40,41. The average plasma endotoxin level of our participants after fasting is about 1.64 EU/ml, which is within this range. The data that endotoxemia was closely associated with plasma FABP2 level (Fig. 2) indicates that the static circulating endotoxin level is highly associated with gut barrier integrity. It is noteworthy that the present study is cross-sectional, our measurement is not sufficient to show if endotoxemia is the cause or the consequence of elevated plasma FABP2. Although our participants showed no signs of systemic inflammation (Table 1), their resting level of endotoxins was still able to trigger NF-B signaling and induce the expression of NF-B target genes in leucocytes (Fig. 2). Informatively, our data disclose that 1 EU/ml is the threshold dose of LPS for activating NF-κB signaling in leukocytes (Fig. 3). Our data support the possibility that low dose of circulating endotoxin can maintain NF-κB signaling in leucocytes at resting state to support the innate immune system. Previous works in animal models have shown that endotoxemia varies diurnally, reaching a zenith at the end of the dark, feeding period and then decreases during the fasting period13. Thus, it is conceivable to assume that circadian rhythms of endotoxin level may elicit a certain degree of NF-κB signaling in leucocytes, which may prime the innate immune system for inflammatory response. The putative pathophysiology of this phenomena and its potential benefit in regulating the immune system warrants further exploration.

In addition to the fasting state, the impact of food consumption on NF-κB signaling was also examined in the menopausal women. Circulating endotoxin level frequently increase after meals, especially with high-fat diets7,13,30 that possibly enhances LPS absorption from the gut. Saturated free fatty acids derived from dietary fat may contribute to systemic inflammation by activating the innate immune cells (e.g., blood monocytes and macrophages)42,43. To avoid the influence of food composition on inflammatory responses, each participant took 75 g glucose for OGTT instead. It has been previously shown that this test exerts a minimal effect on endotoxin level and TLR4 expression7. The plasma FABP2 level was significantly associated with endotoxin level in the subjects after OGTT, which is consistent with the results shown in Fig. 2. However, our data showed no significant difference in the level of either circulating endotoxin or pro-inflammatory cytokines (e.g., TNF-α) between 2-h postprandial and baseline levels (Supplementary Fig. S2). It is worthy to note that several studies have shown that activation of immune cells can still occur even at extremely low doses of LPS22,44. For instance, endotoxin concentrations of 1–4 ng/kg body weight given as an intravenous bolus caused overactive host innate immune responses45. Our ex vivo experiments first reveal that 1 EU/ml of LPS is the threshold dose for activating NF-κB signaling in leukocytes (Fig. 3). Therefore, we suspect that the degree of fluctuation on endotoxemia after 2 h-75 g oral glucose load among those 94 participants might not adequately represent the actual alternation in immune response. Indeed, based on delta endotoxin value, the subgroups (Group A and Group C) had significantly high levels of FABP2, endotoxin and inflammatory cytokines in blood after 2 h-OGTT (Fig. 4) as well as in the expression of NF-κB’s target genes (Fig. 5). Of particular importance is the finding that delta endotoxin value as low as 1 EU/ml was sufficient to trigger NF-B activation in leukocytes and expression of pro-inflammatory cytokines (Fig. 5). These data highlight that regardless of the resting level of circulating endotoxin, the increase of circulating endotoxin as low as 1 EU/ml after food consumption is enough to cause dynamic activation of NF-B signaling in leukocytes and induce low-grade systemic inflammation in the menopausal women. This finding may provide more insight to understand the observed endotoxin-FABP2-NF-κB axis in these menopausal women. However, this may be probably specific to postmenopausal women, that is, its applicability is limited to premenopausal and perimenopausal women due to the lack of comparative data in this study, and its generalizability requires future validation.

Interestingly, our observation may reflect the dynamic relationship between the variation of endotoxin level and NF-κB activation in peripheral leukocytes in vivo (Fig. 5). Indeed, upon repetitive exposures to different doses of LPS, immune cells (e.g., monocytes) may show either suppressed or augmented inflammatory responses, which is different from a single dose of the stimulant46,47,48. It is conceivable that fluctuations in the endotoxin levels may serve as a constant immune insult to the immune system, leading to endotoxin tolerance and priming in innate immune cells. Our data suggest that further investigations are warranted to study the impact of endotoxemia on immune cells, since the tolerance and priming effects of endotoxin are critically involved in the maintenance of immune homeostasis and in the pathogenesis of several inflammatory diseases.

In summary, our results showed that circulating endotoxin level was significantly associated with FABP2 level and NF-κB signaling profiles in leukocytes after fasting in menopausal women. A very low delta concentration of circulating endotoxin (about 1 EU/ml) was sufficient to modulate the activation of NF-κB signaling and inflammatory response in leukocytes. These results provide a deeper insight into the relationship between endotoxemia, gut permeability, leukocyte activation, and regulation of systemic inflammatory responses.

Clinical perspectives

FABP2 is mainly expressed in intestinal epithelial cells and is released into the circulation following intestinal mucosal injury. Circulating endotoxin level is significantly associated with plasma FABP2 level support the notion that FABP2 could be considered as a marker of intestinal barrier impairment, especially in the individuals with marked endotoxemia.

Endotoxemia is significantly associated with the expression of NF-κB target genes such as TLR4, NFKBIA, NFBK1, TNFA, IL1B and IL6 in leukocytes, and with plasma level of TNF-α, IL-1β and IL-6 after fasting in menopausal women.

The increase in circulating endotoxin as low as 1 EU/ml was enough to activate NF-κB signaling in leukocytes to trigger inflammatory responses within 2 h after taking glucose.

In the clinical practice, our findings provide a deeper insight into the relationship between endotoxemia, gut permeability, leukocyte activation, and regulation of systemic inflammatory responses.

Methods

Study design and participants

A total of 174 female participants with age from 45 to 60 years, who visited the Changhua Christian Hospital from January 1, 2019 to December 31, 2022 were included in this study. These participants have experienced at least 12 consecutive months of amenorrhea. Patients receiving hormone replacement therapy, or medications for diabetes, hypertension, hyperthyroidism, or colitis were excluded. Participants were randomly divided into two groups. The first group comprised 80 menopausal women, who had no history of medications for systemic diseases, including diabetes, hypertension, chronic renal diseases, bowel disorders, and thyroid diseases. The second group comprised 94 menopausal women who received OGTT. This study was approved by the Changhua Christian Hospital Institutional Review Board (ID: CCH IRB No. 181019). All participants provided written informed consent regarding their enrollment in the study and the publication of the identifying information. This study has been performed in accordance with the Declaration of Helsinki.

Blood samples and anthropometric measures

Venous blood samples were collected from each participant between 08:00–10:00 a.m. after an overnight fast, and 2 h after taking 75 g glucose for the oral glucose tolerance test (OGTT). Blood samples were fractionated into plasma and whole leukocytes. The plasma was aliquoted and stored at − 80 °C until the assays were performed. Leukocytes were subjected to RNA extraction using NucleoZOL (Düren, Germany). Participants’ height and weight were measured in light clothing without shoes. Body mass index (BMI) was calculated as weight (kg)/height (m2).

The ex vivo blood LPS stimulation in leukocytes

Venous blood samples were collected from 30 participants in the OGTT group after an overnight fast. The fasting blood samples (10 ml/each case) were treated with LPS with the final concentration of 0.01 ng/ml, 0.1 ng/ml, 0.3 ng/ml, and 1 ng/ml, respectively or with vehicle (saline) for two hours. Then, leukocytes were collected for transcript quantification of TLR4/NF-κB signaling parameters, including NFKB1, NFKBIA, TNFA, and IL1B.

Measurements of serum glucose, biochemistry, and WBC count

Fasting and postprandial blood glucose, glycated hemoglobin (HbA1c), total cholesterol, high- and low- density-lipoprotein cholesterol, and triglyceride levels were measured using standard procedures at the Department of Laboratory Medicine, Changhua Christian Hospital. The number of white blood cells (WBC) per µml and the percentage of each type of white blood cell were measured for each blood specimen.

Circulating endotoxin measures

The plasma endotoxin level was measured in triplicate using a chromogenic Limulus Amebocyte Lysate assay (QCL-1000™; Lonza, Walkersville, MD, USA) as previously described46. The endpoint chromogenic LAL test is a quantitative test for detection of Gram-negative bacterial endotoxin. Briefly, plasma samples were diluted 10-fold in 0.1% Tween 80 (Merck Chemicals, Darmstadt, Germany), and heated at 75°C for 8 min. Incubation with LAL reagent was done at 37°C for 10 min and followed by incubation with substrate for 5 min. The reaction was stopped using stop reagent. Absorbance of each sample was determined spectrophotometrically at 405–410 nm. To optimize spectrophotometric measurements, samples were diluted to give the endotoxin readings between 0.1 and 1.0 EU/ml. Samples for endotoxin measurement were stored in endotoxin-free glass tubes and all materials used for the assay were endotoxin-free. The inter-assay and intra-assay coefficients of variation (CV) for endotoxin measurements were 5.5% and 6.2%, respectively.

Measurements of plasma cytokines and chemokines

The plasma levels of tumor necrosis factor-alpha (TNF-α), interleukin 1-beta (IL1-β), and IL6 were determined in triplicate using a Millipore cytokine two-plex panel assay (MILLIPLEX MAP Human High Sensitivity T Cell Magnetic Bead Panel; MILLIPLEX MAP kits, EMD Millipore, Billerica, MA, USA) following the manufacturer’s protocol. Cytokine levels were quantitated using a Luminex 200 system (Luminex, Austin, TX, USA) and reported in pg/ml. Data of cytokine assays were analyzed using an instrument equipped with MILLIPLEX Analyst software (EMD Millipore). For the three cytokines, the intra-assay laboratory coefficients of variation were less than 5.8% and the inter-assay coefficients of variation were less than 6.3%.

Real-time quantitative PCR analysis

Total RNA was prepared from 5 × 107 leukocytes using the RNeasy micro kit (Qiagen, Valencia, CA, USA) and reverse transcribed into cDNA using the Super Script-III First-strand Synthesis System kit (Invitrogen, Carlsbad, CA, USA). To assess the mRNA levels of TLR4, NFKB1, NFKBIA, TNFA, IL1B, and IL6 in leukocytes, real-time quantitative PCR was done using the Rotor-Gene Q cycler (Qiagen, Valencia, CA, and USA) to perform TaqMan Gene Expression assays according to the manufacturer’s protocol. All experiments were conducted in triplicate using TaqMan Gene Expression Master mix (Applied Biosystems), optimized primers, and TaqMan MGB probe sets (Applied Biosystems, Foster City, CA, USA). The PCR primer pairs and TaqMan MGB probes (assay-IDs) (Applied Biosystems) were as follows: Hs01113624_gl (TNFA), Hs01555410_ml (IL1B), Hs00152939_ml (TLR4), Hs00765730_ml (NFKB1), and Hs00355671_g1 (NFKBIA), Hs00174131_ml (IL6), and Hs02758991_g1 (GAPDH). A total of 40 cycles of PCR was performed as follows: activation of AmpliTaq Gold Enzyme (10 min at 95 °C), denaturation (15 s at 95 °C), and annealing/extension (1 min at 60 °C). The threshold cycle (Ct) values of these genes from qPCR data were within the range from 22 to 34. The relative abundance expression (ΔCt) was calculated using the comparative Ct value of each transcript and normalized to the Ct of GAPDH, which serves as a housekeeping gene and a normalizer. The 2–ΔΔCT method was further used to analyze the relative changes in gene expression to test sample side by side and multiple biological sample for each group.

Statistical analysis

Results are expressed as percentage, median (interquartile range), or mean ± standard deviation. Each outcome variable was tested for normal distribution using the Kolmogorov–Smirnov test. Non-normally distributed variables were analyzed using Spearman rank correlation analysis. Multiple comparisons among three groups were performed using ANOVA followed by post hoc Tukey’s test for normally distributed variables or Kruskal–Wallis test followed by the post hoc Dunn test for non-normally distributed variables. All statistical analyses were performed using GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA) and IBM SPSS 20 (SPSS, Inc., Chicago, IL, USA). P-values < 0.05 were considered statistically significant.

name：

Verification code：

My Account

sign in /register

ISO9001：2015

Certificate No.:J22Q28025760R0S