Elevated secreted frizzled-related protein 4 in obesity: A potential role in adipose tissue dysfunction
Disclosure: The authors declared no conflict of interest.
Abstract
Objectives
Rarefaction and inflammation of adipose tissue contributes to insulin resistance in obesity. It was hypothesized that angiostatic secreted frizzled-related protein 4 (SFRP4) causes adipose tissue rarefaction and leads to inflammation and ultimately insulin resistance in obese patients.
Methods
Abdominal subcutaneous adipose tissue (AbdAT), gluteal subcutaneous adipose tissue (GlutAT), and blood from 15 lean and obese subjects were collected. Circulating-SFRP4 was measured by ELISA. Body composition was measured by DEXA and insulin sensitivity by the euglycemic hyperinsulinemic clamp. Adipose tissue was analyzed using qRT-PCR for mRNA gene expression, Luminex system for tissue cytokine release, immunohistochemistry for labeling adipose capillaries, and osmium fixation and Coulter counting for adipocyte sizing.
Results
Circulating-SFRP4 was higher in obese vs. lean subjects (137.8 ± 33.6 ng ml−1 vs. 64.1 ± 23.8 ng ml−1, P < 0.05). Circulating-SFRP4 significantly (P < 0.05) correlated with body fat percentage (R = 0.07), body mass index (R = 0.07), insulin sensitivity (R = −0.66). Circulating-SFRP4 correlated with AbdAT-VEGF (R = −0.67, P < 0.05), AbdAT-capillary density (R = −0.65, P < 0.05), secreted-MIP1α (R = 0.74), and AbdAT-SFRP4 mRNA (R = 0.60). AbdAT-SFRP4 mRNA significantly correlated with AbdAT-capillary density (R = 0.71, P < 0.05), but not with AbdAT mean adipocyte size. There was no difference between AbdAT-SFRP4 and GlutAT-SFRP4 mRNA. Interestingly, GlutAT-SFRP4 correlated with AbdAT mean adipocyte size (P < 0.05).
Conclusions
The results suggested that AbdAT is a major contributor for circulating-SFRP4 and that SFRP4 has an important role in obese adipose tissue pathophysiology.
Introduction
Obesity increases the risk of developing serious health complications including hypertension, steatohepatitis, fatty liver, dyslipidemia and type 2 diabetes mellitus (T2DM) (1).
We have previously shown that human subcutaneous obese adipose tissue has inadequate vascularization, hypoxia, inflammation (2, 3) and fibrosis (4); which is proportional to insulin resistance. Preclinical studies suggest that these relationships are causal with inadequate vascularization being the culprit (5, 6); however what causes impaired adipose tissue vascularization is not known. An interesting circulating antiangiogenic factor called secreted frizzled-related protein 4 (SFRP4) (7) was recently found to predict the development of T2DM 5 years before formal diagnosis is placed (8), therefore having important clinical significance. Limited data is available on SFRP4 at this time (9). SFRP4 acts by decreasing insulin secretion from the pancreatic beta cells (8), by preventing transcription of angiogenesis associated genes (including vascular endothelial growth factor -VEGF) (7), and by modulating Wnt signaling (pathway involved in glucose metabolism) (10). Therefore, SFRP4 could play an important role in the pathogenesis of T2DM. We hypothesize that SFRP4 is causing adipose tissue rarefaction and inflammation and ultimately leads to insulin resistance in obese patients. We conducted a pilot study to evaluate the relationship between SFRP4, obesity, and adipose tissue angiogenesis and inflammation. We showed here for the first time that circulating-SFRP4 is elevated with obesity and adiposity and is significantly correlated with subcutaneous abdominal adipose tissue (AbdAT) vascularization and inflammation and its own gene expression in AbdAT.
Methods
This clinical trial was conducted at Pennington Biomedical Research Center (PBRC) in 8 lean [body mass index (BMI) < 25 kg m−2] and 12 obese subjects [BMI > 30 kg m−2] with (N = 6) or without T2DM (N = 6). Protocol was approved by the PBRC Institutional Review Board and all subjects gave written informed consent. The main study results were published by Pasarica (2-4). Only some of the tissue samples [N = 15; (lean N = 4 and obese N = 11)] were available for the secondary analysis presented in this manuscript.
Clinical measurements
Clinical measurements were previously described by Pasarica et al. (2, 3). Briefly, body composition was measured by dual-energy X-ray absorptiometry (DEXA) on a Hologic Dual Energy X-ray Absorptiometer (Hologic, MA). Maximal aerobic capacity (VO2max) was assessed using a graded treadmill test. Insulin sensitivity was measured during a euglycemic-hyperinsulinemic clamp as the mean rate of exogenous glucose infusion during the 30-min steady-state, corrected for changes in glycemia and divided by fat-free mass. Insulin suppression of lipolysis was assessed as percentage change in rate of appearance of glycerol from the basal to insulin-stimulated state. Subcutaneous adipose tissue biopsies were obtained from the abdominal area (AbdAT) and gluteal area (GlutAT) using a blunt-ended needle.
Laboratory assays
Circulating-SFRP4 was measured in serum using a human SFRP4 ELISA kit (Biomatik). Adipose tissue analysis was previously described by Pasarica (3). Briefly, adipose tissue short-term (3 h) release of proinflammatory cytokines was done using the Luminex system (Millipore) for monocyte chemoattractant protein-1 (MCP1) and macrophage inflammatory protein-1 alpha (MIP1α). VEGF and SFRP4 mRNA gene expression was measured by qRT-PCR using TaqMan (Applied Biosystem) and normalized to the housekeeping gene beta-actin. SFRP4 probe/primer set was purchased from Life Bioscience. VEGF primers/probe set was previously published. Mean adipocyte size was measured by osmium fixation and counting on a Coulter Counter. Adipose tissue capillary density was measured on paraffin-embedded adipose tissue by using tetramethylrhodamine isothiocyanate-conjugated lectin from Ulex europaeus (Sigma–Aldrich) for labeling capillaries and were previously presented (3).
Statistical analysis
Comparison between the lean and obese subjects was performed using an unpaired t test; a Welch correction was applied when the variances between the two groups were significantly different. Statistical significance was defined relative to a nominal two-sided 5% type-1 error rate. Values were presented as means ± standard deviation (SD). Relationships between the SFRP4 and measures of adiposity, angiogenesis or glucose homeostasis were modeled with linear regression. Analyses done in GraphPadv5.
Results
The clinical characteristics and adipose tissue analysis are described in Table 1.
| Lean | Obese | |||
|---|---|---|---|---|
| p value | Anthropometrics | |||
| * | Age (years) | 22.60 ± 3.34 | 38.92 ± 15.83 | a |
| ** | Weight (kg) | 63.49 ± 7.06 | 92.00 ± 12.76 | |
| ** | BMI (kg/m2) | 22.05 ± 1.11 | 31.66 ± 1.94 | |
| ** | Waist-to-hip ratio | 0.78 ± 0.07 | 0.91 ± 0.09 | |
| ** | % Fat | 20.44 ± 8.02 | 34.23 ± 8.24 | |
| Glucose homeostasis | ||||
| ** | GDR (mg/min X kg FFM) | 11.17 ± 3.67 | 6.00 ± 2.20 | |
| ** | VO2max (ml/Kg/min) | 37.10 ± 7.27 | 21.15 ± 2.85 | a |
| AbdAT vascularization | ||||
| * | AbdAT-VEGF mRNA (AU) | 2.46 ± 1.11 | 1.04 ± 0.34 | a |
| * | Capillary density (10−6/μm2) | 307.98 ± 135.09 | 171.79 ± 59.50 | |
| SFRP4 analysis | ||||
| ** | Circulating-SFRP4 (ng/ml) | 64.11 ± 23.82 | 137.77 ± 33.57 | |
| * | AbdAT-SFRP4 mRNA (AU) | 0.27 ± 0.09 | 1.22 ± 0.85 | a |
| * | GlutAT-SFRP4 mRNA (AU) | 0.64 ± 0.23 | 1.79 ± 1.55 | a |
- Comparison between lean and obese was done using an unpaired t-test.
- *P<0.05;
- **P<0.005;
- a: Welch's correction
- Data is presented as means ± SD.
- Abbreviations: BMI: body mass index; FFM: fat-free mass; GDR: glucose disposal rate; VEGF: vascular endothelial growth factor; SFRP4: secreted frizzled-related protein 4.
- Percent body fat was measured by DEXA. GDR was measured by hyperinsulinemic euglycemic clamp to calculate insulin resistance. VO2max was a measure of aerobic exercise capacity. SFRP4 and VEGF mRNA gene expression was measured by qRT-PCR. Capillary density was measured using immunohistochemistry labeling of blood vessels and divided by area labeled.
- The clinical data and tissue angiogenesis were previously described (3). The secondary analysis of SFRP4 was performed in 15 (lean N = 4 and obese N = 11) subjects, as per sample availability.
Circulating-SFRP4
Circulating-SFRP4 is higher in obese vs. lean subjects (P < 0.005) (Figure 1). There is no difference in circulating-SFRP4 between obese subjects with T2DM and the obese subjects without T2DM (P = 0.50), however circulating-SFRP4 is significantly higher in T2DM vs. lean subjects without T2DM (P < 0.005). Circulating-SFRP4 significantly correlates with BMI (R = 0.67, P < 0.01) and body fat percentage (R = 0.70, P < 0.01).
Circulating-SFRP4 in lean and obese subjects with (T2DM) or without T2DM (ND). Circulating-SFRP4 measured by ELISA is (A) elevated in obesity, and correlates with (B) BMI, (C) body fat measured by DEXA and (D) insulin sensitivity measured using hyperinsulinemic euglycemic clamp by glucose disposal rate (GDR). Circulating-SFRP4 correlates with (E) AbdAT inadequate vascularization measured capillary density and (F) inflammation measured by MIP1α secreted in AbdAT conditioned media. Comparison between lean and obese was done using an unpaired t test. Data is presented as Mean ± SD. Regression analysis was used to relate SFRP4 concentration in serum and adiposity markers. P-values and R-values are shown in figure. Abbreviations: ND: subjects without type 2 diabetes; T2DM: subjects with type 2 diabetes; BMI: body mass index; SFRP4: secreted frizzled related protein 4; MIP1α: macrophage inflammatory protein 1 alpha. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Circulating-SFRP4 significantly correlates with insulin sensitivity measured by hyperinsulinemic euglycemic clamp (R = −0.55, P < 0.05), waist-to-hip ratio (R = 0.59, P < 0.05), insulin capacity to suppress lipolysis (R = −0.63, P < 0.05) and aerobic exercise capacity (R = −0.75, P < 0.05).
Interestingly, circulating-SFRP4 significantly correlates with AbdAT vascularization as measured by capillary density (R = −0.65, P < 0.05) and VEGF mRNA (R = −0.67, P < 0.05). Circulating-SFRP4 correlates with AbdAT local inflammation as measured by secreted MIP1α (R = 0.74, P < 0.05), but not with MCP1 (P = 0.22).
AbdAT and GlutAT SFRP4 mRNA
SFRP4 is secreted by multiple tissues (9) and we show here that circulating-SFRP4 significantly correlates with AbdAT-SFRP4 mRNA gene expression (AbdAT-SFRP4) (R = 0.60, P < 0.05) (representative data in Figure 2). As expected, AbdAT-SFRP4 is significantly higher in obese vs. lean (1.22 ± 0.85 AU vs. 0.27 ± 0.09 AU, P < 0.05). AbdAT-SFRP4 is significantly related to AbdAT capillary density (R = −0.71, P < 0.05), but not with AbdAT-VEGF mRNA (P = 0.14). A larger adipocyte size contributes to insulin resistance (11); however we found no relation between AbdAT-SFRP4 and AbdAT mean adipocyte size (P = 0.32).
Abdominal (AbdAT) and gluteal (GlutAT) subcutaneous adipose tissue SFRP4 mRNA expression. AbdAT-SFRP4 mRNA correlated with (A) circulating-SFRP4 and (B) AbdAT capillary density. (C) GlutAT-SFRP4 mRNA correlated with AbdAT mean adipocyte size. SFRP4 mRNA gene expression was measured by qRT-PCR and normalized to internal control. AU: arbitrary units. Circulating-SFRP4 was measured by ELISA. Mean adipocyte size was measured by osmium fixation. Regression analysis was used for these comparisons. P-values and R-values are shown in figure.
AbdAT and GlutAT have different function and gene expression (12); however we found no difference between AbdAT-SFRP4 vs. GlutAT-SFRP4 (1.00 ± 0.85 AU vs. 1.48 ± 1.32 AU, respectively; P > 0.05). GlutAT-SFRP4 does not correlate with circulating-SFRP4 (P = 0.30). Interestingly, there is a significant correlation between GlutAT-SFRP4 and AbdAT mean adipocyte size (R = 0.66, P < 0.05).
Discussion
Adipose tissue rarefaction seems to be the culprit of hypoxia, inflammation, and ultimately insulin resistance. We established here for the first time that angiostatic circulating-SFRP4 is elevated in obesity and is related to adipose vascularization, inflammation and its own gene expression in AbdAT. This suggested that AbdAT-SFRP4 may be a major contributor to the total circulating-SFRP4, and may contribute to adipose tissue pathophysiology in obesity.
Others have shown that circulating-SFRP4 is elevated in T2DM vs. non-diabetic subjects with a BMI ranging from lean to overweight and obese (8); we revealed for the first time in our pilot study that circulating-SFRP4 is doubled in obese compared to lean subjects, and correlates significantly with BMI and body fat percentage. These suggest that in our cohort, circulating-SFRP4 is a measurement of adiposity. T2DM subjects had elevated circulating-SFRP4 vs. lean, but not when compared to obese without T2DM. As the BMI range in our study was different than in the study by Mahdi et al., we only had a small cohort of subjects (15 vs. 74) and we found a strong correlation between circulating-SFRP4 and insulin resistance, we conclude that our results are supporting the previously (8) proposed role of SFRP4 as biomarker for T2DM.
We showed that both circulating-SFRP4 and AbdAT-SFRP4 are significantly correlated with the end effect of angiogenesis, namely capillary density in adipose tissue. VEGF is one of the multiple molecules that stimulate angiogenesis. We found that only circulating-SFRP4 correlates with AbdAT–VEGF, while AbdAT-SFRP4 does not correlate with AbdAT–VEGF. These suggested the hypothesis that SFRP4 may inhibit capillary formation in AbdAT in an endocrine and paracrine manner; however it may not act through VEGF but through a different molecule that we have not yet identified. Circulating-SFRP4 was directly proportional with AbdAT inflammation, suggesting that SFRP4 inhibits AbdAT vascularization, leads to tissue inflammation, and ultimately insulin resistance in an endocrine manner. These hypothesis need to be tested in future ex vivo experiments.
SFRP4 is secreted by multiple tissues including the adipose tissue (9). Obese adipose tissue secrets multiple factors which contribute to insulin resistance (13). We showed here that circulating-SFRP4 significantly correlates with AbdAT-SFRP4, suggesting that AbdAT may be a major contributor for circulating-SFRP4. A larger adipocyte size contributes to insulin resistance (11). Recombinant SFRP4 significantly decreases adiponectin release in the human adipocyte media during differentiation. Inhibiting SFRP4 with small interfering RNA decreased differentiation of human adipocytes (10). SFRP4 inhibits angiogenesis of human umbilical vein endothelial cells (7). These support a role for SFRP4 in human adipocyte physiopathology. AbdAT-SFRP4 was not linked with mean AbdAT adipocyte size, suggesting that SFRP4 is not involved in the energy storage pattern of AbdAT. Surprisingly, GlutAT-SFRP4 significantly correlated with AbdAT mean adipocyte size, which suggests a possible role for SFRP4 in adipose tissue depot distribution. This is an interesting finding that needs to be further explored. We focused here on subcutaneous adipose tissue as our hypothesis revolves around the expandability of this depot; however we acknowledge that visceral adipose tissue is a major contributor for insulin resistance and may also contribute to circulating-SFRP4.
Note this is a cross-sectional clinical trial, with data presented as comparison and correlations. Although other tissues may contribute to circulating-SFRP4, we only studied adipose SFRP4. However, this is a critical pilot study as it opens the door for future interventions, ex vivo and preclinical studies. This is the first time SFRP4 was linked with adipose tissue vascularization and inflammation in obesity. This study was done in humans which limited our capability for extensive exploration; however, it added clinical significance, as SFRP4 may be used in the future for target drug development in patients with obesity and insulin resistance.
Taken together, our results suggest an important role of SFRP4 in pathological events that result from obesity. The mechanism through which SFRP4 exerts its effect in obesity is still unclear, but this study suggests paracrine and endocrine effect.
Acknowledgments
We thank Dr. Steven R. Smith for providing clinical samples.
