Role of Thrombospondin-2 in Murine Adipose Tissue Angiogenesis and Development

Authors


(roger.lijnen@med.kuleuven.be)

Abstract

Expression of thrombospondin-2 (TSP-2), a matricellular protein with anti-angiogenic properties, is modulated in developing adipose tissue. To investigate a potential functional role of TSP-2 in adipose tissue angiogenesis and growth, TSP-2 deficient (TSP-2−/−) and wild-type littermate (TSP-2+/+) mice were kept on normal chow (standard fat diet (SFD)) or on high fat diet (HFD) for 15 weeks. TSP-2−/− mice kept on HFD had a significantly lower total body weight throughout the experimental period. Subcutaneous (SC) and gonadal (GON) fat mass were, however, not different, and their composition in terms of size and density of adipocytes and blood vessels was also comparable in both genotypes. Macrophage infiltration in SC or GON adipose tissues was not affected by TSP-2 deficiency. TSP-2 deficiency had no effect on adipose tissue mRNA expression of gelatinase A (MMP-2), whereas gelatinase B (MMP-9) was downregulated in SC and GON adipose tissues of TSP-2−/− mice on HFD. Glucose tolerance and insulin resistance tests were comparable for TSP-2+/+ and TSP-2−/− mice. TSP-2 deficiency was not compensated by increased expression of TSP-1 in the TSP-2−/− mice. These data suggest that TSP-2, despite its reported anti-angiogenic properties, does not play an important functional role in adipose tissue related angiogenesis or associated fat development in mice.

Introduction

Development of adipose tissue is associated with extracellular matrix proteolysis, adipogenesis and angiogenesis (1). Angiogenesis is generally believed to be ongoing during development of obesity, and inhibition of angiogenesis has the potential to reduce adipose tissue development (2). Thrombospondin-2 (TSP-2) is expressed in adipose tissues, and is upregulated in gonadal adipose tissue of C57Bl/6 mice with nutritionally induced or genetically determined obesity (3). Thrombospondins are large secreted, multimodular, calcium-binding glycoproteins that have complex roles in mediating cellular processes (4). Thrombospondin-1 (TSP-1) has been implicated in cell—matrix and cell—cell interactions important for platelet function, angiogenesis, tumor biology, wound healing, and vascular disease (5). TSP-2 is also believed to have in vivo anti-angiogenic properties (6,7,8,9,10). Enhanced angiogenesis in the context of TSP-2 deficiency was shown to be associated with enhanced activation of gelatinase B (MMP-9) during revascularization (7), and with elevated levels of MMP-2 (gelatinase A) and MMP-9 during wound healing (10). In TSP-2 deficient mice, age-related cardiomyopathy was observed, accompanied by increased MMP-2 activity (11). In this study, we have evaluated the potential contribution of TSP-2 to adipose tissue related angiogenesis and fat development. The rationale for this study was the presumed anti-angiogenic potential of TSP-2 and its effect on gelatinase expression, whereas both angiogenesis and gelatinase A (MMP-2) activity were shown to play functional roles in adipose tissue development (2,12).

Methods and Procedures

Animals and models

TSP-2 deficient (TSP-2−/−) and wild-type (TSP-2+/+) littermate mice (genetic background C57Bl6/129SvJ/EMS + Ter) were generated from heterozygous breeding couples (13). Five-week-old male mice (TSP-2+/+ or TSP-2−/−) were kept on a high fat diet (HFD) for 15 weeks (TD88137; Harlan Taklad, Zeist, The Netherlands, containing 42% kcal as fat with a caloric value of 20.1 kJ per gram) or on a standard fat diet (SFD) (KM-04-k12, Muracon; Carfil, Oud-Turnhout, Belgium, containing 13% kcal as fat with a caloric value of 10.9 kJ per gram).

Mice were kept in microisolation cages on a 12-h day/night cycle and were fed ad libitum. At the end of the experiment, the mice were killed by intraperitoneal injection of 60 mg/kg sodium pentobarbital (Abbott Laboratories, North Chicago, IL). Blood was collected from the retroorbital sinus on trisodium citrate (final concentration 0.01 mol/l) and plasma was stored at −20 °C. Epididymal (gonadal, GON) and inguinal (subcutaneous, SC) fat pads were removed and weighed; portions were immediately frozen at −80 °C for extraction and other portions were used to prepare 10 µm paraffin sections for histology (14). The weight of other organs was also recorded.

Food intake was measured weekly for 4-day periods throughout the experimental period, and expressed as gram per mouse and per day. Physical activity at night (1900–0700 h) was monitored by placing mice in a cage equipped with a turning wheel linked to a computer to record turns/night during at least 3 different weeks.

Glucose tolerance and insulin resistance tests were performed, with a 1-week recovery period, in 12-weeks-old TSP-2+/+ and TSP-2−/− mice (n = 5 each) kept on SFD. Therefore, after an overnight fasting, glucose (3 mg/g body weight) or human insulin (0.5 mU/g) was injected into the peritoneal cavity. Blood was collected via the tail vein for glucose measurement before and at 15–120 min after injection. For detailed analysis, the areas under the curves of glucose level vs. time were determined. All animal experiments were approved by the local ethical committee (KU Leuven, P07123) and performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals (1996) and the guiding principles of the International Society on Thrombosis and Haemostasis (15).

Assays

Adipose tissue sections were stained with hematoxylin/eosin using a standard protocol or with the biotinylated Bandeiraea (Griffonia) Simplicifolia BSI lectin (Sigma-Aldrich, Bornem, Belgium) followed by signal amplification with the Tyramide Signal Amplification Cyanine System (Perkin Elmer, Boston, MA). The number of adipocytes and their mean size were determined by computer-assisted image analysis, using for each animal three to five areas in four different sections; the data were first averaged per section and then per animal. The lectin staining allows to visualize blood vessels (16); blood vessel size was determined by computer-assisted image analysis, and blood vessel density was expressed as the number of vessels per measured surface area (9–12 sections were analyzed per animal and then averaged). In addition, blood vessels were stained with an anti-CD31 monoclonal antibody (clone MEC 13.3; BD Pharmingen, San Jose, CA).

Blood glucose concentrations were determined using Glucocard strips (Menarini Diagnostics, Florence, Italy); triglyceride, total, high-density lipoprotein and low-density lipoprotein cholesterol levels as well as liver enzymes (aspartate aminotransferase, alanine aminotransferase) were evaluated using routine clinical assays. Insulin (Mercodia, Uppsala, Sweden) and leptin (R&D Systems Europe, Lille, France) were measured with commercially available enzyme-linked immunosorbent assays.

Expression of TSP-1, TSP-2, MMP-2, and MMP-9 in extracts of SC or GON adipose tissues was monitored by real-time PCR, as described (Geneassays Mm01335418_m1 for TSP-1, Mm01279236_m1 for TSP-2, Mm00439498_m1 for MMP-2 and Mm00442991_m1 for MMP-9, using GADPH (Geneassay Mm 99999915_gl) as internal control (Applied Biosystems, Carlsbad, CA)). Expression of macrophage markers including tumor necrosis factor-α (Mm00443258_m1), interleukin-1 (Mm01336189_m1), Cd68 (Mm03047340_m1) and monocyte chemotactic protein-1 (Mm00441242_m1) was also evaluated (3). Data were obtained as cycle threshold (Ct) values and expressed as copy number of target mRNA relative to 105 copies of GAPDH. In addition, macrophages were stained with anti Mac-3 antibody (Pharmingen, San Diego, CA), and the positively stained area expressed as a percentage of the section area.

Gelatinase activity was monitored by zymography with extracts of SC and GON adipose tissues and expressed in arbitrary units, as described elsewhere (12). Plasma levels of TSP-1 were determined by enzyme-linked immunosorbent assay (17).

Statistical analysis

Data are expressed as means ± s.e.m. Statistical significance for differences between groups (effect of genotype or diet) is analyzed by two-way ANOVA. Comparison of progress curves is performed by two-way repeated measures ANOVA. Statistical significance is set at P < 0.05.

Results

During the 15-week experimental period, body weight gain was more pronounced on the HFD as compared to the SFD for both TSP-2+/+ and TSP-2−/− mice. Total body weights at the end of the study were lower for TSP-2−/− as compared to TSP-2+/+ mice on HFD, but not on SFD (Figure 1). After overnight fasting, this difference, however, was no longer observed (Table 1). TSP-2 deficiency had no effect on food intake of the mice, either on HFD (2.9 ± 0.17 vs. 2.9 ± 0.08 g/mouse/day for TSP-2+/+), or on SFD (4.0 ± 0.15 vs. 4.0 ± 0.09 g/mouse/day for TSP-2+/+). The feeding efficiency (weight gain normalized to caloric intake) of both genotypes thus was comparable either on HFD or on SFD. Physical activity at night was also comparable for TSP-2−/− and TSP-2+/+ mice on SFD (5,670 ± 710 vs. 6,540 ± 790 turns/12 h) or on HFD (6,550 ± 1,480 vs. 5,440 ± 720 turns/12 h). TSP-2 deficiency did not significantly affect development of SC or GON adipose tissue on SFD or HFD. The weight of other organs was also comparable, indicating that TSP-2 deficiency did not affect general development of the mice (Table 1).

Figure 1.

Evolution of body weight of TSP-2−/− (open symbols) or TSP-2+/+ (closed symbols) mice kept on SFD (squares) or HFD (circles). Data are mean ± s.e.m. of 10–14 mice per group. Two-way repeated measures ANOVA revealed a significant effect of the genotype for mice kept on HFD (P < 0.0001), but not for mice kept on SFD (P = 0.78). HFD, high fat diet; SFD, standard fat diet; TSP, thrombospondin.

Table 1.  Adipose tissue and organ weights of TSP-2+/+ and TSP-2−/− mice kept on SFD or HFD
inline image

Analysis of adipocyte and blood vessel size and density in SC and GON adipose tissues did not reveal a significant effect of TSP-2 deficiency for mice on either diet (Table 2). Furthermore, detailed analysis of the size distribution of blood vessels in SC and GON adipose tissues (in 20 µm2 increments) did not reveal differences in function of genotype for either diet (Figure 2). Staining for CD31 confirmed comparable blood vessel density in SC and GON adipose tissues of TSP-2−/− and TSP-2+/+ mice either on SFD or HFD (expressed as ×10−6/µm2: 742 ± 106 vs. 596 ± 61 for SC fat of TSP-2+/+ and TSP-2−/− mice on SFD, with corresponding values of 442 ± 35 vs. 495 ± 32 on HFD; 514 ± 94 vs. 476 ± 51 for GON fat of TSP-2+/+ and TSP-2−/− mice on SFD, with corresponding values of 362 ± 27 vs. 319 ± 51 on HFD). Overall, the increase in body weight, SC and GON fat mass and adipocyte size observed in TSP-2−/− mice when fed a HFD, was very similar to what was seen in TSP-2+/+ mice. Analysis of expression of macrophage markers revealed no differences between TSP-2−/− and TSP-2+/+ SC or GON adipose tissues for monocyte chemotactic protein-1, tumor necrosis factor-α and Cd68 mRNA, either on SFD or HFD (Table 3). Comparable macrophage content in TSP-2−/− and TSP-2+/+ adipose tissues of mice on either diet was confirmed by staining for Mac-3 (Table 2). Expression of interleukin-1 was about twofold downregulated in SC and GON fat of TSP-2−/− mice on HFD as compared to TSP-2+/+ mice.

Table 2.  Adipocyte and blood vessel size and density in adipose tissues of TSP-2+/+ and TSP-2−/− mice kept on SFD or HFD
inline image
Figure 2.

Blood vessel size distribution in subcutaneous (SC) or gonadal (GON) adipose tissues of TSP-2+/+ or TSP-2−/− mice on SFD or HFD. The data represent average contributions in the total number of blood vessels of blood vessel size in 20 µm2 increments. HFD, high fat diet; SFD, standard fat diet; TSP, thrombospondin.

Table 3.  Gene expression levels in adipose tissues of TSP-2+/+ and TSP-2−/− mice kept on SFD or HFD
inline image

Analysis of plasma metabolic parameters (Table 4) revealed a significant effect of the genotype on glucose levels (P = 0.017 by two-way ANOVA), whereas the diet had no effect (P = 0.87). Other metabolic parameters were not significantly affected by TSP-2 deficiency in mice kept on SFD or HFD. Glucose tolerance tests did not reveal differences between TSP-2+/+ or TSP-2−/− mice (areas under the curves of 13,360 ± 670 vs. 12,150 ± 770); insulin tolerance was also comparable (areas under the curves of 5,710 ± 470 vs. 5,940 ± 610).

Table 4.  Metabolic parameters and liver enzymes of TSP-2+/+ and TSP-2−/− mice kept on SFD or HFD
inline image

Blood cell analysis did not reveal significant effects of TSP-2 deficiency with SFD or HFD on cell counts, including total white and red blood cells, platelets, neutrophils, lymphocytes, monocytes, basophils, eosinophils or hemoglobin or hematocrit levels (data not shown).

TSP-2 mRNA could not be detected in TSP-2−/− mice and was markedly lower in SC as compared to GON adipose tissues of TSP-2+/+ mice. Expression of TSP-1 mRNA in SC or GON adipose tissues was not significantly different for TSP-2+/+ or TSP-2−/− mice, either on SFD or on HFD (Table 3). Plasma levels of TSP-1 were also comparable for TSP-2+/+ and TSP-2−/− mice kept on SFD (879 ± 169 ng/ml vs. 872 ± 260 ng/ml) or on HFD (665 ± 139 ng/ml vs. 511 ± 73 ng/ml). Expression of MMP-2 mRNA in SC or GON adipose tissues was not different for TSP-2+/+ or TSP-2−/− mice, either on SFD or on HFD. Expression of MMP-9 mRNA was downregulated in TSP-2−/− as compared to TSP-2+/+ mice on HFD, in SC (1.8-fold; P = 0.013) but not in GON adipose tissue (P = 0.09). In contrast, on SFD no differences in MMP-9 expression were observed in SC or GON adipose tissues of both genotypes (Table 3). Zymography with extracts of SC and GON adipose tissues (Table 5) confirmed similar levels of latent and active MMP-2 forms for both genotypes on SFD or HFD. Pro MMP-9 levels were also comparable, whereas active MMP-9 was undetectable in all the samples.

Table 5.  Effect of TSP-2 deficiency on gelatinolytic activity in adipose tissues of mice kept on SFD or HFD for 15 weeks
inline image

Discussion

The TSP family currently comprises five members, of which TSP-1 and TSP-2 (subgroup A) are trimeric matricellular proteins that influence cell function by modulating cell—matrix interactions (18). TSP-2 binds to both jagged1 and Notch3 ectodomains, potentially bridging two essential extracellular components of Notch signaling. TSP-2 and LRP1 (low-density lipoprotein receptor-related protein 1) stimulate Notch activity by driving trans-endocytosis of the Notch ectodomain into the signal-sending cell (19,20).

TSP-1 and TSP-2 are expressed in murine adipose tissues (3). Whereas TSP-1 deficiency did not affect development of obesity in murine models (21), no information is available on a potential functional role of TSP-2. We found that TSP-2−/− mice when kept on SFD or HFD developed adipose tissue to a comparable extent as their wild-type littermates, although a trend to reduced GON fat was observed in TSP-2−/− mice on HFD. We have only studied male mice, as it has been suggested previously that TSP-2 may alter adipose tissue regulation by female sex hormones (22). Two-way repeated measures ANOVA indicated a significantly lower body weight over the 15-week experimental period for TSP-2−/− mice on HFD but not on SFD. However, after overnight fasting at the end of the experiment this difference was no longer observed, probably because TSP-2+/+ mice lost more weight. The composition of SC and GON adipose tissues of both genotypes in terms of adipocyte size and density was very similar. In the mixed genetic background of the mice used in this study, weight gain upon HFD feeding for 15 weeks was less pronounced as compared to e.g., a pure C57Bl/6 genetic background (12). Nevertheless, our data allow to conclude that TSP-2 deficiency in mice has no significant effect on adipose tissue, food intake or physical activity. Furthermore, no significant differences were observed in adipose tissue associated angiogenesis between TSP-2−/− and TSP-2+/+ mice kept on SFD or HFD, as revealed by comparable size and density of blood vessels and comparable size distribution. This appears to be in contrast with the previously reported anti-angiogenic effect of TSP-2 in settings of wound healing and neovascularization (6,7,8,9,10), and the higher tissue density of medium and small blood vessels observed in TSP-2−/− mice (13). The absence of an effect of TSP-2 deficiency could not be explained by compensation through enhanced TSP-1 levels, which are comparable for both genotypes on SFD and somewhat lower for TSP-2−/− mice on HFD. The lack of TSP-2 may lead to aberrant extracellular matrix remodeling, increased neovascularization and reduced contraction, due in part to elevated levels of MMP-2 and MMP-9 (10,23). It was also reported that in TSP-2 deficient mice age-related cardiomyopathy was accompanied by increased MMP-2 activity, paralleled by reduced activity of tissue transglutaminase-2, which impairs collagen crosslinking and thus may contribute to cardiac dilation and dysfunction (11). We have previously reported that deficiency of MMP-2, but not MMP-9, impairs adipose tissue development in mice by contributing to adipocyte hypotrophy (17,24). Because of the suggested link between TSP-2 deficiency and MMP-2 and −9 levels, we have monitored expression and activity of gelatinases in the adipose tissues in this study. However, we did not find an effect of TSP-2 deficiency on MMP-2 mRNA levels in SC or GON adipose tissues, whereas MMP-9 was actually downregulated upon HFD feeding. Zymography with protein extracts of SC and GON adipose tissues of both genotypes did, however, not reveal significant differences in MMP-9 or MMP-2 activity. Thus, the functional interaction between TSP-2, angiogenesis and gelatinase expression/activity previously observed in wound healing and neovascularization does not play a relevant role in adipose tissue development. Interestingly, statistical analysis (two-way ANOVA) indicated an effect of the genotype on fasting plasma glucose levels. However, glucose and insulin tolerance tests did not reveal significant differences between both genotypes suggesting that glucose homeostasis is not markedly affected. It is unclear at present why TSP-2 deficiency is associated with enhanced basal glucose levels, both on SFD and HFD.

TSP-2 was also shown to play a role in inflammatory processes (25). Increased macrophage accumulation, associated with increased expression of proinflammatory compounds, was observed in adipose tissues of obese subjects due to an influx of bone marrow—derived macrophages (26,27,28). In our model, however, absence of TSP-2 had no effect on adipose tissue development. Furthermore, analysis of the content and expression of macrophage markers in SC and GON adipose tissues did not reveal significant differences between the two genotypes, except for a twofold downregulation of interleukin-1 mRNA in TSP-2−/− mice on HFD. Overall, this suggests that macrophage infiltration in the adipose tissues is not markedly affected by TSP-2 deficiency. This suggests that recruitment of precursor cells (from bone marrow) to adipose tissue, if any in this model, is not affected by TSP-2 deficiency.

In summary, a nutritionally induced obesity model in TSP-2 deficient mice has not revealed an important role of TSP-2 in adipose tissue related angiogenesis or associated fat development.

Acknowledgment

Thrombospondin-2 deficient mice were originally obtained by courtesy of Dr Bornstein (Departments of Biochemistry and Medicine, Washington University, Seattle, WA). Skilful technical assistance by S. Helsen, C. Vranckx and A. De Wolf is gratefully acknowledged. This study was supported financially by the “Fonds voor Wetenschappelijk Onderzoek-Vlaanderen” (G.0240.07) and by the Interuniversity Attraction Poles (IUAP, P6/30). The Center for Molecular and Vascular Biology is supported by the “Programmafinanciering KULeuven” (PF/10/014).

Disclosure

The authors declared no conflict of interest.

Ancillary