Arterial stiffening is a loss of arterial elasticity, progressing with age (Bramwell and Hill, 1922; Hallock, 1934; Eliakim et al., 1971; Tounian et al., 2001; Mitchell et al., 2004; McEniery et al., 2005) and obesity (Sutton-Tyrrell et al., 2001; Safar et al., 2006), therefore affecting a large portion of the population. It is a precursor to isolated systolic hypertension (Najjar et al., 2008; Kaess et al., 2012), which increases the risk of stroke and heart disease (Nielsen et al., 1995; Blacher et al., 1999; Meaume et al., 2001; Boutouyrie et al., 2002; Cruickshank et al., 2002). With a large majority of the United States population expected to be obese by 2030, it is critical to understand the mechanisms by which increased adiposity or changes to adipose tissue composition contribute to vascular dysfunction.

The presence of metabolically active brown adipose tissue (BAT) in human adults has been confirmed (Nedergaard et al., 2010). The quantity of BAT is also negatively correlated with obesity and age (Yoneshiro et al., 2011); in addition, imaging studies show metabolically active BAT along the thoracic, but not the abdominal, spinal area (Cypess et al., 2015). This adipose tissue surrounding the thoracic aorta is similar to interscapular BAT in structure and gene expression pattern (Fitzgibbons et al., 2011). Because of its juxtaposition, thoracic perivascular adipose tissue (PVAT) influences vascular function, mainly by releasing vasoactive, and in the context of obesity, inflammatory molecules (Soltis and Cassis, 1991; Löhn et al., 2002; Verlohren et al., 2004; Gao et al., 2009; Greenstein et al., 2009). High-cholesterol and high-fat diet increase PVAT pro-inflammatory cytokine production in mouse models of atherosclerosis and abdominal aortic aneurysm respectively, further linking obesity to increased risk of cardiovascular disease (Police et al., 2009; Dobrian et al., 2015). In addition, decreased expression of adiponectin, an anti-inflammatory cytokine released by adipocytes, is observed in PVAT under conditions of obesity. Local administration of recombinant adiponectin in mice is able to reverse neointimal thickening observed in obesity (Takaoka et al., 2009), providing evidence that functional adipokines are released by PVAT. However, it should be noted that these studies were performed on adipose tissue depots that are phenotypically similar to white adipose tissue (WAT). The detailed examination of the paracrine function of perivascular brown adipocytes is warranted and important, as it will provide novel targets to modulate cardiovascular pathologies associated with both aging and obesity.

VEGF is a pro-angiogenic molecule known to stimulate vasodilation (Horowitz et al., 1997), while VEGF neutralization is associated with hypertension (Hurwitz et al., 2004; Levine et al., 2004). It is widely known that VEGF produced by smooth muscle cells (SMCs) (Williams et al., 1995) acts through its receptors on endothelial cells to regulate normal and pathological angiogenesis (Ferrara and Davis-Smyth, 1997). VEGF receptors are also expressed on SMCs (Brown et al., 1997; Couper et al., 1997; Wang and Keiser, 1998; Grosskreutz et al., 1999), and their activation critically regulates atherosclerosis and neointimal hyperplasia. Mouse models with overexpression of VEGF in adipose tissue demonstrate increased vascularization within adipose tissue, as well as, protection against systemic metabolic dysfunction (as induced by high-fat diet) (Elias et al., 2012; Sun et al., 2012, 2014). It has been shown that deletion of VEGF in adipose tissue results in increased WAT inflammation, BAT lipid accumulation, and disturbances with glucose tolerance (Sung et al., 2013; Shimizu et al., 2014); however, the role of endogenous VEGF in PVAT function and SMC regulation were not addressed. Therefore, determining the effect of brown adipocyte-derived VEGF on PVAT function, vascular tone, and cardiovascular disease related to aging and obesity has become increasingly important.

Although the capacity of the angiogenic signal provided by VEGF has been shown to drive BAT development (Sun et al., 2014), the critical role of brown adipocyte-derived VEGF in maintaining SMC function has not been demonstrated. Correlative studies using conditioned media demonstrate increased VEGF release from visceral and epicardial PVAT of obese and diabetic patients which induces vascular SMC proliferation (Schlich et al., 2013), however these studies do not assess the mechanical properties of the SMC (Qiu et al., 2010).

We recently published that the vast majority of VEGF in adipose tissue is derived from the adipocyte precursor fraction, with minimal expression of VEGF in other cell populations. As previously shown, and reproduced here, the level of VEGF-A in the BAT is three-fold greater than that of WAT (Mahdaviani et al., 2016). Since mitochondria play an important role in thermogenesis and brown adipocyte function, we hypothesized that the lack of VEGF was responsible for our observations. We differentiated preadipocytes from interscapular BAT of Fabp4cre(+). VEGF^flox/flox and control mice. Protein expression levels of VEGF, UCP1, and porin (VDAC) (used to asses mitochondrial content), and norepinephrine-stimulated uncoupling were all decreased in the adipocytes lacking VEGF. Acute ablation of VEGF by adenoviral-Cre knockdown resulted in decreased VEGF without affecting porin levels and decreased response to norepinephrine albeit to a lesser extent than chronic conditions (Mahdaviani et al., 2016).

The present study was designed to elucidate the effects of capillary rarefaction in PVAT and the subsequent impact on macrovascular function. Here, we demonstrate that PVAT from Fabp4cre(+). VEGF^flox/flox show a decrease in vascularity and increased lipid coalescence. The results of this study demonstrate that VEGF-A deficiency in PVAT is a contributing factor toward several elements of vascular dysfunction. We found that our Fabp4cre(+). VEGF^flox/flox model displayed increased collagen deposition and decreased aortic relaxation in the thoracic aorta in young mice on chow diet, suggesting that there is an early physiological response to decreased VEGF-A within the PVAT.

Paradoxically, the concentration of adiponectin in circulation is inversely proportional to adipose mass, suggesting that adipose quality rather than quantity is more important for expression of adiponectin (Turer et al., 2011). In terms of the anti-fibrotic role of adiponectin, this inverse correlation is also observed, with reduced circulating concentrations associated with increased fibrosis (Marangoni et al., 2017). Although adiponectin expression in PVAT has been previously described (Antoniades et al., 2009), we further demonstrated that VEGF ablation affected the levels of adiponectin expressed in the PVAT. While this decreased adiponectin expression in PVAT may play an important role in this model, we cannot exclude the fact that other adipokines may be contributing to vascular dysfunction.

Endothelial cells are a dynamic paracrine organ and play a crucial role in maintaining vascular tone and mediating inflammatory processes. Endothelial dysfunction from obesity-induced metabolic syndrome is characterized by a decrease in nitric oxide (NO) production (Calles-Escandon and Cipolla, 2001). Our isometric tension measurements revealed that Fabp4cre(+). VEGF^flox/flox aortic rings exhibited a dysfunctional vasomotor response which suggests that VEGF-A deficiency has both a structural and biochemical influence.

Limitations of our study include the possibility that the observed effects are attributed solely to PVAT. As such, the lack of PVAT during aortic ring isometric tension studies prevents conclusively determining the vasomodulatory properties of PVAT. To account for any changes occurring specifically by thoracic PVAT in our model, future studies may include isometric tension experiments with preserved PVAT and denervation of BAT. In addition, we acknowledge that we cannot exclude the possibility that the microvasculature is contributing to any observed effects and would hypothesize that small vessels may also be damaged in the long term, should hypertension be observed. Furthermore, since it is possible that there may be two arms to VEGF signaling – angiogenic and metabolic – one could envision devising a biased agonist to selectively activate the metabolic arm, thereby improving PVAT function without inducing a deleterious angiogenic response.

Taken together, these experiments suggest that the microvascular environment of PVAT has an important effect on the function of its surrounding tissue, specifically, the ability of endothelial cells to modulate vascular tone. With the increasing prevalence of obesity and its association with capillary rarefaction in adipose tissue, further exploring the role of PVAT within this dynamic system is warranted. Such studies can help provide greater insight to various other comorbidities such as hypertension.

Our in vivo data suggest that VEGF produced by adipocytes cells prevents collagen accumulation in the aortic wall, and ex vivo experiments demonstrate a requirement of adipose-derived VEGF in aortic relaxation. Our data also show that impaired VEGF signaling by adipocytes leads to abnormal aortic tone, remodeling, and stiffness, suggesting that paracrine effects of VEGF from adipocytes in the thoracic PVAT maintains adipose tissue and could contribute to effects on vascular integrity. Therefore, cardiovascular complications, as observed in obesity and aging, may benefit from modulation of this critical pathway to improve disease conditions.

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

The Institutional Animal Care and Use Committee of Boston University School of Medicine approved these experiments.

RG and TA designed the study, performed experiments, and wrote the paper. SS provided technical assistance with experiments.