Funding agencies: The study was supported by the Danish Research Council, the Danish Council for Strategic Research (DanORC consortium), the Danish Diabetic Association, and Aarhus University.
Vascular growth is a prerequisite for adipose tissue (AT) development and expansion. Some AT cytokines and hormones have effects on vascular development, like vascular endothelial growth factor (VEGF-A), angiopoietin (ANG-1), ANG-2 and angiopoietin-like protein-4 (ANGPTL-4).
In this study, the independent and combined effects of diet-induced weight loss and exercise on AT gene expression and proteins levels of those angiogenic factors were investigated. Seventy-nine obese males and females were randomized to: 1. Exercise-only (EXO; 12-weeks exercise without diet-restriction), 2. Hypocaloric diet (DIO; 8-weeks very low energy diet (VLED) + 4-weeks weight maintenance diet) and 3. Hypocaloric diet and exercise (DEX; 8-weeks VLED + 4-weeks weight maintenance diet combined with exercise throughout the 12 weeks). Blood samples and fat biopsies were taken before and after the intervention.
Weight loss was 3.5 kg in the EXO group and 12.3 kg in the DIO and DEX groups. VEGF-A protein was non-significantly reduced in the weight loss groups. ANG-1 protein levels were significantly reduced 22-25% after all three interventions (P < 0.01). The ANG-1/ANG-2 ratio was also decreased in all three groups (P < 0.05) by 27-38%. ANGPTL-4 was increased in the EXO group (15%, P < 0.05) and 9% (P < 0.05) in the DIO group. VEGF-A, ANG-1, and ANGPTL-4 were all expressed in human AT, but only ANGPTL-4 was influenced by the interventions.
Our data show that serum VEGF-A, ANG-1, ANG-2, and ANGPTL-4 levels are influenced by weight changes, indicating the involvement of these factors in the obese state. Moreover, it was found that weight loss generally was associated with a reduced angiogenic activity in the circulation.
Excess fat mass increases the risk of multiple conditions, including type II diabetes, hypertension, cardiovascular diseases, and certain types of cancer. In addition, obesity negatively affects physical functioning, vitality, and general quality of life (1). Unfortunately, lifestyle modification with or without medication are relatively ineffective strategies for achieving significant long-term weight loss which is important to obtain the health benefits associated with a lower body weight (2, 3). Thus, a better understanding of the development and consequences of adipose tissue (AT) expansion is necessary to identify better preventive- and treatment strategies for the obese state.
In obesity, excessive expansion of adipose tissue results from adipocyte hypertrophy and adipocyte hyperplasia (1). To supply the growing adipose tissue with nutrients and oxygen, the vasculature responds by increasing the number and/or size of blood vessels (4, 5). This coordinated growth of adipose tissue and blood vessels is thought to be mediated by paracrine interactions between adipocytes and endothelial cells (ECs) (5). Adipocytes produce angiogenic factors that promote EC tube formation and vascularization (5, 6). While ECs also produce factors that stimulate adipogenesis (5). It has been shown that the expansion of adipose tissue in severely obese subjects is associated with active angiogenesis, and inhibition of angiogenesis seems to be able to prevent adipose tissue development and obesity (7). We therefore wanted to test the hypothesis that a decrease in fat mass would lead to a decrease in angiogenic factors. In addition, different strategies to induce weight loss might affect the production of the angiogenic factors differently.
Vascular endothelial growth factor-A (VEGF-A) is believed to be responsible for most of the angiogenic capacity in the adipose tissue, and adipogenesis is dependent on VEGF-A mediated formation of new blood vessels (8, 9). The angiogenic factor VEGF-A exerts its effects after binding to the its receptors, VEGFR-1, VEGFR-2, and VEGFR-3, that are expressed mainly by endothelial cells (10). Intriguingly, the affinity of VEGF-A for VEGFR-1 is much higher than that of VEGFR-2, but the signalling induced by VEGFR-2 is the major way by which VEGF-A regulates endothelial cell migration (10). VEGF-A is expressed in virtually all vascularized tissues, especially in fenestrated and sinusoidal blood vessels in endocrine and secretory organs as well as in large blood vessels, skeletal muscle, and myocardium, suggesting that low physiological VEGF-A levels are needed for the maintenance of general vascular homeostasis (11).
Another signalling system contributing to maintenance, growth, and stabilization of blood vessels, involves angiopoietin-1 (ANG-1) and ANG-2 (4). ANG-1 and ANG-2 are ligands of the endothelial cell–specific TIE-2 receptor. ANG-1 binds to TIE-2 and activates it by inducing dimerization, which results in phosphorylation of the kinase domain of TIE-2. ANG-2 also binds to TIE-2; however, ANG-2 does not induce phosphorylation of TIE-2 at physiological concentration. Therefore, ANG-2 has been suggested to work as a naturally occurring antagonist of ANG-1 (12). Consistent with in vitro results, in vivo studies reveal that ANG-1 seems to act in complementary and coordinated fashion with VEGF-A, having a later role in vascular development by stabilizing blood vessel structure (9, 12-14). ANG-1 is predominantly expressed by perivascular and mural cells, suggesting a paracrine mode of action (15).
One of the recently identified secreted factors from adipose tissue is angiopoietin-like protein 4 (ANGPTL-4). In contrast to ANG-1, angiopoietin-like proteins do not bind to TIE-2 (16). Growing evidence suggest that ANGPTL-4 is a key player in angiogenesis. Recent data from several independent laboratories have demonstrated ANGPTL-4 as a potent antiangiogenic factor (17, 18). This antiangiogenic factor, ANGPTL-4, prevents the formation of focal adhesions by endothelial cells and thereby negatively regulates endothelial cell migration and angiogenesis (17, 18). ANGPTL-4 mRNA is predominantly detected in adipose tissue and liver, followed by thyroid, brain, and small intestine (19).
In general AT mass can be reduced by decreasing energy intake (hypocaloric diet), by increasing energy expenditure (exercise), by a combination of diet and exercise or by surgery (2). A recent study has found reduced levels of VEGF-A after large weight loss induced by bariatric surgery (20). Whether exercise influences the angiogenic factors in adipose tissue is unknown, but it is well-known that exercise stimulates angiogenesis, and VEGF-A production in skeletal muscles (21, 22).
In the present study, we want to investigate whether weight loss induced by diet only, diet in combination with exercise or exercise only induced a different angiogenic profile in blood and adipose tissue gene expression. The study was a 12-week intervention with three groups of obese subjects randomized to regular exercise alone, diet-induced weight loss, and a combination of exercise and diet-induced weight loss. In the last two groups, we intended to obtain similar weight losses to see the possible specific, weight independent, effect of exercise.
Methods and Procedures
Seventy-nine obese but otherwise healthy males and females were recruited via advertisements in local newspapers. The subjects were eligible for inclusion if they were aged 18-45 year, obese (BMI 30-40 kg m−2), physically inactive (<30 min day−1) and weight stable for at least 3 month (±2 kg of current body weight). Exclusion criteria were cardiovascular disease, type 2 diabetes, pregnancy, or orthopedic difficulties causing inability to undertake an exercise program. No subjects received medication that could affect the investigated vascular markers (23). The subjects gave a written informed consent. The study was approved by the local ethics committee in the county of Aarhus and followed the principles outlined in the Declaration of Helsinki.
During a 6-month period, the 79 subjects were randomized into the 12-week intervention study consisting of (1) exercise only (EXO, n = 25), (2) hypocaloric diet (DIO, n = 29), and (3) hypocaloric diet and exercise (DEX, n = 25). Twenty subjects did not complete the study (8 women and 12 men (see (23) for further details).
Subjects in the DIO and DEX groups were prescribed a liquid very low energy diet (VLED; Nupo, Copenhagen, Denmark) of respectively 600 and 800 kcal day−1 (proteins 41 g, carbohydrates 29 g, fat 5.6 g/100 g) for 8 weeks followed by a weight maintenance diet for 4 weeks. In these two groups, we intended the subjects to obtain similar weight losses to see the possible specific, weight independent, effect of exercise. Thus, the subjects in the DEX group were allowed to consume 150-200 kcal more per day as compared with the DIO group, reflecting the estimated extra energy expenditure of 1500 kcal week−1 during exercise activity. To ensure compliance to the diet, subjects in both groups were allowed to consume ad libitum low-energy vegetables and were followed every second week by clinical staff. In the weight maintenance phase, the subjects consumed a diet with the following energy content: 55% from carbohydrates, 15% from protein, and <30% from fat. The daily energy requirement for the subjects during this period was determined by estimating resting energy expenditure multiplied by a factor of 1.5 for subjects in the DIO group and 2.5 in the DEX group. The subjects in the EXO group were advised to maintain an isocaloric diet for the duration of the intervention. Their daily energy expenditure during the intervention was determined by estimating the resting energy expenditure multiplied by a factor of 2.5. All subjects in the three groups were asked to keep dietary intake records over a 2-week period.
The exercise intervention for subjects in the EXO and the DEX group consisted of supervised aerobic exercise three times per week with duration of 60-75 min per training session, with an estimated energy expenditure of 500-600 kcal per session. The subjects could choose between different modes of exercise; stationary bicycling, jogging on a treadmill or stair stepping. The Karvonen method for exercise intensity was used to target an exercise intensity of 70% of heart rate reserve: HR = ((HRmax − HRrest) × % Intensity) + HRrest, where the heart rate maximum was determined during the VO2 max test. The exercise intensity was monitored using heart rate monitors (Polar S810i, Polar Electro Oy, Kempele, Finland). The subjects were required to keep records of training sessions during the intervention.
Maximal rate of oxygen uptake
At baseline and after 12 weeks, each subject completed a progressive maximal exercise test using a stationary cycle ergometer (Monark 828, Monark Exercise AB, Vansbro, Sweden) and standard open-circuit spirometry techniques (AMIS 2001, Innovision, Odense, Denmark).
At baseline and after 12 weeks the body weight was measured to the nearest 0.05 kg with a calibrated scale (Tanita digital scale, Class III NTEP 440 Capacity Scale) and visceral AT (VAT) was measured by whole body MRI (270 equidistant images) with a Philips Gyroscan Achieva 1.5 Tesla MR scanner (Philips Medical systems, Best, The Netherlands) (23). Blood samples were collected after an overnight fast, between 24 and 48 h after the subjects had finished the last exercise session, and before the biopsies were obtained. More details about the intervention have previously been described (23).
Isolation of RNA from adipose tissue
At baseline and after week 12, the AT biopsies were obtained from the abdominal subcutaneous AT. The skin was anesthetized with lidocaine (10 mg ml−1) before a small incision was made and ≈200 mg of AT were removed under sterile conditions using a liposuction needle. Immediately after removal, the AT sample was washed in isotonic NaCl, snap-frozen in liquid nitrogen, and kept at −80°C until RNA extraction. Total RNA was isolated from the biopsies using the TriZol reagent (Gibco BRL, Life Technologies, Roskilde, Denmark); RNA was quantified by measuring absorbance at 260 and 280 nm in TE buffer, and the inclusion criteria was a ratio ≥2. Finally, the integrity of the RNA was checked by visual inspection of the two ribosomal RNAs, 18S and 28S, on an agarose gel.
Real-time reverse transcriptase PCR of angiogenic factors
Complementary DNA was constructed using random hexamer primers (GeneAmpRNAPCR Kit from Perkin-Elmer Cetus, Norwalk, CT). PCR-mastermix, containing the specific primers, Hot star Taq DNA polymerase, and SYBR-Green PCR buffer were used. The following primer pairs were used:
VEGF-A sense primer 5′-GGGGCACACAGGATGGCTTGAAGATGTACT-3′ and antisense primer 5′-GTTCATGGATGTCTATCAGCGCAGCTACTG-3′.
ANG-1 sense primer 5′-ACCGGATTTCTCTTCCCAGA-3′ and antisense primer 5′-CCGACTTCATGTTTTCCACA-3′.
ANGPTL-4 sense primer 5′-GAGATGGCCCAGCCAGTT-3′ and antisense primer 5′-TAGTCCACTCTGCCTCTCCC-3′.
Beta-2 microglobulin sense primer 5′ -TCTCTCTTTCTGGCCTGGAG-3′ and antisense primer 5′-AATGTCGGATGGATGAAACC-3′.
Real-time quantification of genes was performed by SYBR-green real-time reverse transcription PCR assay (Qiagen, Valencia, CA) using an ICycler from Bio-Rad (Bio-Rad Laboratories, Hercules, CA). Human Beta-2 microglobulin was used as an endogenous reference. This housekeeping gene was selected as Beta-2 microglobulin demonstrated no significant changes in expression between the three groups before and after intervention (data not shown). Relative quantification was performed essentially according to the comparative ΔCt method (Perkin Elmer User Bulletin, 1997).
Determination of angiogenic factors in serum
Blood samples were collected after a fast and 24-48 h after last exercise bout. Protein concentration in serum was measured using specific high sensitive human ELISAs; VEGF-A assay (Quantikine DY293B, R&D Systems Europe, Abingdon, UK), ANG-1 assay (Quantikine DY923, R&D Systems Europe, Abingdon, UK), ANG-2 assay (Quantikine DANG20, R&D Systems Europe, Abingdon, UK) and ANGPTL-4 assay (Quantikine DY3485, R&D Systems Europe, Abingdon, UK). Intra-assay reproducibility ranged from 3.4 to 19%.
To determine whether there were differences between the three intervention groups we used two-way repeated measures analysis of variance (ANOVA) with group as one factor and the treatment group (DIO, EXO, DEX) as a second factor. We used the Holm-Sidak method for the post hoc multiple comparison procedure. Moreover, changes in relation to baseline levels were analyzed with a paired t-test in each of the groups. A linear regression model was used to test the association between selected variables. Protein levels and mRNA expression of VEGF-A, ANG-1, ANG-2, and ANGPTL-4 were all log transformed to reduce skewness before statistical calculation. Data presented in figures was not log transformed. P values <0.05 were considered significant. All data are presented as means ± SEM. SigmaStat for Windows Version 3.0 (SPSS, Cary, NC) was used for statistical analysis.
After the 12-week-intervention body weight was reduced by 3.5% in the EXO (P < 0.05) and by 12 and 11% (P < 0.05) in the two diet-induced weight loss groups (DIO and DEX, Table 1). From the MRI-scanning it was found that VAT in the EXO group was significantly reduced by 17% (P < 0.05, Table 1). In the DEX and DIO groups VAT was reduced by 35 and 28%, respectively (P < 0.05) as compared with baseline (Table 1). Subjects in the EXO and DEX groups increased their VO2, expressed as ml/kg/min, max by 23 (P < 0.05) and 28% (P < 0.001) whereas, as expected no changes in VO2 max were observed in the DIO group (P = 0.9) (Table 1).
Table 1. Characteristics of the subjects at baseline and after the 12-week intervention
The subject were divided into three groups; exercise alone (EXO, n = 19), diet-induced weight loss (DIO, n = 19), and the combined diet/exercise group (DEX, n = 21). Protein levels are measured in serum. Data are expressed as mean ± SEM. Asterisks above the bars indicate significant differences (*P < 0.05, **P < 0.001, ***P < 0.001 as compared with baseline).
19 (10♀ 9♂)
19 (9♀ 10♂)
21 (11♀ 10♂)
37.2 ± 1.61
35.6 ± 1.57
37.5 ± 1.66
100.9 ± 2.37
95.6 ± 2.01*
107.8 ± 2.80
95.5 ± 2.62*
105.8 ± 3.32
93.5 ± 2.80*
BMI (kg m−2)
33.3 ± 0.87
32.2 ± 0.84*
35.3 ± 0.86
31.3 ± 0.82*
34.2 ± 0.70
30.3 ± 0.60*
27.8 ± 1.6
34.2 ± 1.8*
26.4 ± 1.8
28.9 ± 1.9
28.4 ± 1.3
36.4 ± 1.8***
Visceral AT (cm3)
2752 ± 235
2284 ± 195*
3173 ± 327
2275 ± 290*
3152 ± 287
2020 ± 217*
VEGF (pg ml−1)
184 ± 32
245 ± 65
249 ± 85
222 ± 59
158 ± 23
124 ± 23
ANG-1 (ng ml−1)
61 ± 5.7
46 ± 3.6**
70 ± 11
55 ± 8.7**
51 ± 4.4
38 ± 3.9**
ANGPTL-4 (ng ml−1)
715 ± 435
818 ± 468*
665 ± 248
722 ± 310*
206 ± 48
211 ± 55
Angiogenic factors in circulation
The serum levels of VEGF-A, ANG-1 and particularly ANGPTL-4 were rather variable from person to person (Table 1).
There was a tendency to a 33% higher circulating VEGF-A levels in the EXO group, 11% lower levels in the DIO group and 21% (P = 0.083) lower levels in the DEX group after the 12-week intervention as compared with baseline levels (Figure 1A, Table 1). ANG-1 protein levels were significantly reduced by 25% (P < 0.01), 23% (P < 0.01) and 22% (P < 0.01) in the EXO, DEX, and DIO groups (Figure 1B, Table 1). The ANG-1/ANG-2 ratio was also significantly reduced in all three groups with a reduction in the exercise groups (EXO) by 38% (P < 0.001), reduced by 31% in the DEX group (P < 0.01), and reduced by 27% in the DIO group (P < 0.05) (Figure 1D). The anti-angiogenic factor, ANGPTL-4, was increased by 15% in the EXO group (P < 0.05) and increased by 9% in the diet-induced weight loss group (DIO; P < 0.05) (Figure 1C, Table 1), whereas no significant changes were observed in the DEX group (P > 0.2). Concerning these angiogenic factors no significantly differences were found between the three intervention groups (analyzed by ANOVA).
Both women and men participated in this intervention but neither baseline levels nor the changes in the angiogenic factors induced by the interventions were affected by gender (data not shown).
Expression of angiogenic factors
VEGF-A, ANG-1, and ANGPTL-4 were all expressed in human AT (Figure 2). The expression of VEGF-A was not affected by exercise or diet (Figure 2A), but we found a tendency of a higher VEGF expression after exercise in muscle tissue (data not shown). ANG-1 expression was not affected by any of the three interventions (Figure 2B). ANGPTL-4 mRNA in the adipose tissue was significantly reduced in the EXO group by 32%, in the DIO group by 29% and in the DEX group by 24% (P < 0.01 for all, Figure 2C).
At baseline circulating levels of VEGF-A and ANGPTL-4 was found positively associated with body weight (both r = 0.3; P < 0.05). Moreover, VEGF-A was found positively associated with the visceral AT (r = 0.27; P < 0.05). No correlations between changes in angiogenic factors and changes in body weight/fat mass were observed (data not shown). No correlations between protein levels in circulation and mRNA expression of ANG-1, VEGF-A, and ANGPTL-4 in AT were found.
In obesity rapid expansion of AT may result in hypoxia and dysfunction of AT with enhanced inflammation and finally insulin resistance. The degree of dysfunction of AT seems to be related to vascularization of AT which is regulated by several angiogenic factors. In the present study in humans we found that several angiogenic factors both in the circulation as well as locally in AT seem to be affected by both exercise and weight loss per se.
We found a positive association between the main proangiogenic factor, VEGF-A, and body weight, and VEGF-A and VAT, which is in agreement with some other studies finding that VEGF positively correlates with BMI and VAT both in obese and lean subjects (20, 24-26). Moreover, in the present study we found VEGF-A levels increased after exercise alone (EXO group) which was opposite to the changes in the diet-induced weight loss groups where VEGF-A decreased. This may be due to the fact that exercise stimulates VEGF-A protein production in skeletal muscle, which has been shown by several groups (21, 22). Thus, the serum concentration of VEGF-A may be a mixture of contributions from several tissues where the local production very well could be regulated in opposite directions (21, 22). That VEGF-A levels decreases in association with weight loss per se has been found by some other groups (25, 26).
We found that ANG-1 in the circulation decreases with weight loss, which is compatible with a recent human in vivo study, where ANG-1 positively correlates with percent fat mass (14). Because both the reduction in ANG-1 and weight loss was rather similar between the three intervention groups and the common factor for all three groups is some degree of weight loss, it is suggested that the reduction in ANG-1 is mainly due to the weight loss and that exercise does not provide any additional effect. These findings indicate that even a modest weight loss is sufficient to reduce ANG-1 levels. ANG-1 is a pro-angiogenic factor whereas ANG-2 may antagonize the effect of ANG-1 (12). Thus, the ratio between these two factors may be of importance for the over-all angiogenic activity of this system and we found that not only ANG-1 was reduced by the interventions but also the ANG-1/ANG-2 ratio in the circulation was reduced by all three interventions indicating a generally reduced angiogenic activity of this system after weight loss.
The anti-angiogenic factor, ANGPTL-4, was found positively associated with body weight and, moreover, was increased after the intervention both in the exercise group (EXO) and in the diet group (DIO), but surprisingly no changes were observed in the combined intervention (DEX) group. This is partly in agreement with some recent studies where ANGPTL-4 has been found to increase in response to fasting, chronic caloric restriction, and endurance exercise (19, 27). However it is unclear why the combination of caloric restriction and exercise did not induce any changes in ANGPTL-4 in the present study. But the results are compatible with the suggestion that both exercise and diet-induced weight loss may reduce the angiogenic activity.
The effect of diet-induced weight loss and exercise on the measured angiogenic factors seems to be independent of gender, since the changes observed in females and males were comparable and the very no significant differences in the angiogenic factors at baseline between sexes.
Concerning the expression of the various angiogenic factors in AT we were able to identify the expression of VEGF-A, ANG-1, and ANGPTL-4 in human AT. Although circulating levels of VEGF-A and ANG- 1 were reduced after diet-induced weight loss no changes in the expression of VEGF-A/ANG-1 in AT was observed after the interventions indicating that AT may not be involved in the observed changes in serum levels of VEGF-A and ANG-1. These results are in agreement with another study where no change in gene expression in the adipose tissue of VEGF-A and ANG-1 after weight loss was detected (28). ANGPTL-4 has recently been shown to be expressed in the adipose tissue (19, 27). In our study, the expression of ANGPTL-4 was significantly reduced after the intervention in all three groups and this is opposite to the changes in circulating levels of ANGPTL-4 which were increased after the interventions—both after exercise and weight loss. ANGPTL-4 is predominantly expressed in the liver (19, 27). Thus, most of the circulating ANGPTL-4 may be released from the liver and the changes observed in association with weight loss could involve the liver as well. Why the gene expression of ANGPTL-4 is reduced in human AT in association with weight loss and the possible biological implications of this reduction are completely unknown.
The present study indicate that weight loss generally was associated by reduced proangiogenic activity in relation to the circulating angiogenic factors since VEGF-A, ANG-1, and ANG-1/ANG-2 ratio were all reduced and the levels of the antiangiogenic factor, ANGPTL-4, is increased after weight loss. On the other hand these changes observed in circulating levels were not reflected by similar concerning gene expression in the AT. Thus, from our gene expression data we found surprisingly no indication of reduced angiogenesis in AT after weight loss in the present study. This could be due to several factors where one of these could be that the gene expression data may not be representative for the protein levels of these factors or that the weight losses were too low to change the angiogenic profile locally in the AT.
It has been discussed whether blocking angiogenesis in AT could reduce high-fat induced obesity and reduce obesity complications (5, 7), on the contrary would enhance local hypoxia, inflammation and metabolic complications. Very recently it has been found that enhancing angiogenesis by overexpressing VEGF-A in AT reduced hypoxia, inflammation and reduced metabolic complications (29) indicating beneficial metabolic effects of enhanced angiogenesis and thereby enhanced vascularization in association with a high fat diet. These authors conclude that “manipulation of AT angiogenesis in various stages of obesity development differentially affect whole AT physiology and energy expenditure.” Thus, whether stimulation of angiogenesis in AT has beneficial metabolic effects or the opposite may strongly depend on the circumstances (i.e., timing in relation to obesity development etc.) (29).
In summary, we see a reduction of VEGF, ANG-1, ANG-1/ANG-2 ratio, and an increment of ANGPTL-4 following diet and exercise induced weight loss indicating a generally reduced angiogenic activity in the circulation. Moreover, these results seem to indicate that VEGF-A, ANG-1, ANG-2, and ANGPTL-4 play a role in the obese state per se and maybe for some of the health complications associated to obesity. Data from different murine obesity models suggests the involvement of these factors in the adipose tissue remodulation that takes place in association with obesity and weight loss. Our gene expression data from human AT were, however, not unequivocal in that respect in the present study. The angiogenic factors in the adipose tissue seem from our present study to be affected by weight loss but not influenced by the methods used to induce weight loss (diet, exercise or both).
We thank Lenette Pedersen and Pia Hornbek for their skilful technical assistance. The study was supported by The danish council for research in health and disease, the Danish Council for Strategic Research and Aarhus University.