It is increasingly recognized that obesity is a risk factor for microvascular disease, involving both structural and functional changes in the microvasculature. This review aims to describe how obesity impacts the microvasculature of a variety of tissues, including visceral adipose tissue, skeletal muscle, heart, brain, kidneys, and lungs.
These changes involve endothelial dysfunction, which in turn (i) impacts control of vascular tone, (ii) contributes to development of microvascular insulin resistance, (iii) alters secretion of paracrine factors like nitric oxide and endothelin, but (iv) also influences vascular structure and perivascular inflammation. In concert, these changes impair organ perfusion and organ function thereby contributing to altered release and clearance of neurohumoral factors, such as adipokines and inflammatory cytokines. Global microvascular dysfunction in obese subjects is therefore a common pathway that not only explains exercise-intolerance but also predisposes to development of chronic kidney disease, microvascular dementia, coronary microvascular angina, heart failure with preserved ejection fraction, chronic obstructive pulmonary disease, and pulmonary hypertension.
A large body of evidence has accumulated over the years, from both clinical and experimental studies, indicating that obesity is associated with endothelial dysfunction and development of atherosclerosis, and that obesity has become one of the most important risk factors for cardiovascular disease including coronary artery disease, heart failure, and stroke. In addition to atherosclerosis in the larger arteries, obesity is also a risk factor for microvascular disease. Interestingly, a single high fat meal already perturbs endothelial function in the brachial artery, and reduces flow reserve in the coronary vasculature, illustrating how a single exposure to a high circulating lipid load has an impact, albeit transient, on the microvasculature. Regular exposure to high circulating lipid loads, even prior to the onset of overt obesity, leads to an inflammatory response that is accompanied by microvascular dysfunction, the severity of which correlates with the amount of visceral adipose tissue present in the body. Eventually, obesity and the associated inflammation not only impact function, but also structure of the microvasculature (Figure 1).
Proposed mechanisms of obesity-related microvascular dysfunction predisposing to multi-organ disease. High fat diet on a regular basis changes the composition of visceral adipose tissue, and induces a low grade local inflammatory response, which together modify the secretion of adipokines. Simultaneously, high fat diet results in endothelial dysfunction throughout the body, which not only alters vascular tone, and contributes to development of microvascular insulin resistance, but also influences vascular structure and perivascular inflammation. In concert, these microvascular changes impair organ perfusion and organ function thereby further contributing to altered release and clearance of metabolites and neurohumoral factors, like adipokines, inflammatory cytokines as well as (cardio)myokines. Global microvascular dysfunction in obese subjects therefore is a common pathway that contributes to exercise-intolerance and predisposes to development of chronic kidney disease, microvascular dementia, coronary microvascular angina, COPD and pulmonary hypertension. CKD, chronic kidney disease; HFpEF, heart failure with preserved ejection fraction; COPD, chronic obstructive pulmonary disease.
Adipose tissue microcirculation in obesity
Two main types of adipose tissue can be distinguished in the body, which have distinctly different functions both in healthy and obese subjects being white and brown adipose tissue (WAT and BAT respectively). WAT is the primary site of fat accumulation, and not only allows efficient fat storage, but also quick mobilization of fat stores to meet energy demands of the body. WAT comprises both subcutaneous and visceral adipose tissue (VAT). The main role of BAT is thermogenesis. Thermogenesis in BAT is activated by the sympathetic nervous system. High expression of uncoupling protein-1 on the inner membrane of BAT mitochondria results in uncoupling of mitochondrial respiration so that heat is generated instead of ATP. In adult humans, BAT is mainly located in cervical, supraclavicular, mediastinal, paravertebral, suprarenal, and peri-renal areas. In addition, epicardial and perivascular adipose tissue have a phenotype that more closely resembles BAT than WAT. Both WAT and BAT contain dense microvascular networks, but microvascular density is higher in BAT as compared to WAT, with 3 vs. 1 capillary per adipocyte, respectively. The microvasculature in WAT serves as the exchange site for fat deposition and mobilization, while in BAT it is required for both delivery of fuel for and dissipation of heat produced during thermogenesis. Metabolism, perfusion and function of both WAT and BAT are affected by obesity.
Obesity, by definition an excessive accumulation of fat mass, results in expansion of particularly WAT. Ingestion of a single high fat meal induces upregulation of P-selectin on the venular side of the visceral adipose microvasculature, thereby forming an anchoring point for leucocytes. The leucocytes infiltrate the VAT and initiate an inflammatory response. Hence, low grade inflammation particularly in visceral adipose tissue may precede excessive fat accumulation, increase oxidative stress, and cause chronic microvascular dysfunction. Interestingly, blood flow to VAT increases following meal ingestion in lean but not obese subjects. Moreover, diet-induced obesity is accompanied by decreased eNOS-expression and activity, while eNOS overexpression protects against diet-induced obesity . These observations suggest a key role for microvascular function in adipose tissue homeostasis. Besides this paracrine interaction between adipocytes and the microvasculature, it should be noted that fat accumulation in VAT results in an increase in adipocyte-size from 50 μm up to 150–200 μm, which is beyond the diffusion distance for oxygen, while the accompanying reduction in capillary density will further decrease adipose tissue oxygenation. Indeed, chronic hypoxia has been shown to be present in expanded VAT. Similar to WAT, brown adipocytes hypertrophy in obesity. Intriguingly, it has recently been shown that capillary rarefaction, leading to focal hypoxia in BAT, is sufficient to induce ‘whitening’ of BAT, which is associated with reduced beta-adrenergic signaling, mitochondrial dysfunction, loss of thermogenic capacity and further accumulation of lipid droplets.
Hypoxia per se induces a reduction in adiponectin and an increase in leptin release from isolated adipocytes. Moreover, chronic hypoxia results in sustained inflammation thereby further modulating the secretion of adipokines from both WAT and BAT and contributing to metabolic derangement in obesity. In healthy subjects, the secretion of the anti-inflammatory adipokine adiponectin predominates, whereas in obese subjects, there is a shift toward pro-inflammatory adipokines such as leptin, resistin, TNFα, IL-6, and IL-18 (Table 1). Thus, adipose tissue hypoxia and inflammation are centrally involved in the pathophysiology of obesity, and can, through release of vasoactive and/or inflammatory adipokines, modulate microvascular function throughout the body (Figure 1).
Skeletal muscle microvasculature in obesity
Although some studies in young adult humans suggest that skeletal muscle blood flow is relatively well-maintained in obesity even during exercise, others show a reduction in flow normalized for muscle mass both at rest and during exercise. These findings seem to be independent of age and vascular bed, as the reduction in flow is present in children, and adults, both in the forearm and upper leg. Similarly, the exercise-induced increase in systemic vascular conductance is blunted in obese swine as compared to lean swine, consistent with a decrease in flow to exercising muscle. This decrease in flow is compensated by an increase in oxygen extraction to fulfill the oxygen-requirement of skeletal muscle.
In skeletal muscle, close coupling of blood flow to metabolic activity is required and besides substances released from nerve terminals, the endothelium and the contracting muscle, also involves mechanical interaction between the contracting muscle and the vasculature. The nervous system contributes to exercise hyperemia in skeletal muscle via activation of sympathetic vasodilator fibers, vasodilation elicited by acetylcholine spillover from active motor nerves as well as functional sympatholysis in active muscle. Resting muscle sympathetic nerve activity (MSNA) is significantly higher in obese patients with metabolic syndrome, but it does not further increase during exercise. Interestingly, in the presence of β-blockade, exercise resulted in larger increase in forearm blood flow and conductance in obese as compared to lean men. Together with the observation that the exercise-induced increase in skeletal muscle blood flow is reduced or at best preserved in obesity, these data suggest that β-adrenergic vasodilation is reduced in obesity. Furthermore, in the presence of β-blockade, α2– but not α1-stimulation resulted in a larger decrease in forearm vascular conductance in obese vs. lean subjects at rest and during exercise. Similarly, there was a tendency towards a larger increase in conductance with α-blockade with phentolamine at rest in obese vs. lean subjects, but a reduced increase in conductance upon α-blockade during exercise. These data suggest that obesity results in a shift in the balance of neurogenic control of skeletal muscle blood flow, with increased α-adrenergic constriction at rest, that is withdrawn during exercise, thereby compensating for a loss of β-adrenergic vasodilation .
In humans, endothelium-dependent skeletal muscle microvascular vasodilation in response to acetylcholine is either preserved or reduced in obesity, while eNOS expression is unaltered. However, the contribution of both NOS- and cyclooxygenase (COX)- dependent vasodilator mechanisms to acetylcholine-induced vasodilation is reduced, which is compensated by NOS- and COX-independent vasodilator mechanisms, potentially an increase in endothelium-derived hyperpolarizing factors such as EETs and/or H2O2. An increase in H2O2 may result from superoxide dismutase (SOD)-mediated conversion of superoxide, particularly in more active obese individuals. In addition, a reduction in NO-bioavailability has been shown to be counterbalanced by reduced phosphodiesterase 5 (PDE5) activity, as well as by a reduced vasoconstrictor influence of ET. The latter is mediated through a reduction in ET-sensitivity of the skeletal muscle arterioles together with a decrease in circulating ET, reflecting a decrease in local ET production. Interestingly, a study in rodents suggests that basal differences exist in endothelial cell phenotype between arteries perfusing slow-twitch and those perfusing fast-twitch muscle fibers, with the former being less susceptible to endothelial dysfunction.
Alterations in endothelial function may also play a role in insulin-dependent modulation of microvascular tone. In healthy individuals, insulin-induced vasodilation serves to facilitate glucose delivery and uptake in skeletal muscle. As outlined in the introduction, insulin resistance is associated with a shift from insulin-induced vasodilation to vasoconstriction. Indeed, the change in femoral vascular conductance upon glucose ingestion is smaller in obese as compared to lean women, which is associated with impaired body glucose uptake. Moreover, flow to the quadriceps muscle of obese men is lower in the presence of insulin both at rest and during exercise, and insulin-induced vasodilation is converted into vasoconstriction in skeletal muscle arterioles isolated from obese women. The latter is mediated, at least in part, by alterations in perivascular adipose tissue (PVAT), that displays a pro- inflammatory phenotype in obesity.
Although low grade inflammation may lead to an increase in oxidative stress, TBARS (as index of systemic oxidative stress) do not appear to be different between normal and obese young adults. Conversely, ROS production by NADPH-oxidase and Xanthine oxidase in skeletal muscle is increased in overtly obese, but not mildly obese individuals. In accordance with this observation, infusion of the NADPH-oxidase inhibitor and antioxidant apocynin augmented the acetylcholine-induced increase in flow to skeletal muscle in obese but not lean subjects. However, despite a significant inverse correlation between TBARS and vasodilator responses, the antioxidant ascorbic acid augmented acetylcholine-induced vasodilation to a similar extent in normal and obese subjects under resting conditions. Nevertheless, the significant correlation between waist-to-height ratio and TBARS post-exercise, together with a negative correlation between catalase and BMI, suggests that anti-oxidant capacity may fall short during stress in obese subjects.
The obesity-induced decrease in exercise capacity (VO2max) closely correlates with capillary density in skeletal muscle, indicating that besides functional, also structural changes in the skeletal muscle microvasculature are present. A study in rodents suggests that capillary rarefaction in obesity occurs in two phases, of which the first one is mediated by an increase in oxidative stress, and the second one by a decrease in NO-bioavailability.
Taken together, obesity moderately reduces skeletal muscle blood flow at rest. The mechanisms of this reduction are incompletely understood but may involve factors released from perivascular fat that modulate insulin sensitivity and endothelial function as well as an increased vasoconstrictor influence caused by an increase in MSNA. In addition, structural changes in the skeletal muscle microvasculature may contribute to the decreased resting blood flow. Because exercise hyperemia involves many redundant regulatory mechanisms, it is relatively well-maintained. However, the impairments in blood flow and its distribution are likely to become more severe when the metabolic disturbance persists for a longer time or when other co-morbidities such as hypertension, hyperlipidemia, and hyperinsulinemia co-exist in the same patient, which may contribute to exercise intolerance in obese people.
Coronary microcirculation in obesity
Similar to skeletal muscle, coronary microvascular function plays an important role in coupling of myocardial perfusion to cardiac metabolism.54 Clinical studies have shown that in conditions such as obesity and hypercholesterolemia the coupling of coronary blood flow to the myocardial metabolic demand is altered.2 Thus, coronary flow velocity of obese patients during dobutamine stress-echo is impaired, with the impairment becoming even more evident when obesity is associated with other risk factors.55 Also, myocardial blood flow measurements with PET indicate that increases in myocardial blood flow in response to cold-pressor testing are diminished in obese patients.56 Interestingly, female sex and the volume of visceral fat are associated with a reduction in myocardial perfusion at peak dose of dobutamine, as measured by MRI.57 These clinical observations are supported by studies in obese Ossabaw swine with metabolic syndrome and swine with severe familial hypercholesterolemia (FH), demonstrating that coronary blood flow regulation and myocardial oxygen balance are altered in particular during treadmill exercise.58,59
Several studies have reported reduced coronary flow reserve and increased minimal vascular resistance in patients with obesity and hypercholesterolemia as assessed by either PET, Doppler echocardiography or MRI.60–64 Such abnormalities can be the result of either changes in microvascular function, i.e. the regulation of vascular tone, or structural changes within the coronary microcirculation, such as vascular rarefaction or inward remodeling.54,65 Alterations in control of coronary microvascular tone are generally characterized by a loss of endothelial vasodilator influence (such as NO) as well as by increased neurohumoral (angiotensin II) and endothelium-derived vasoconstrictor (ROS, ET, prostanoids) influences resulting in a shift in the vasomotor balance towards increased vasoconstriction (Table 1).2,3,66 Indeed, several studies in humans and animal models have demonstrated that obesity is associated with alterations in the vasodilator-vasoconstrictor balance controlling coronary microvascular tone.2–4,67 Thus, hyperoxia-induced vasoconstriction in the coronary microvasculature is enhanced in obese adolescents,68 while bradykinin-induced vasodilation is impaired in isolated coronary arterioles from obese patients due to increased tissue angiotensin-converting enzyme activity.69 Similarly, in swine with hypercholesterolemia, endothelial dysfunction of isolated small coronary arteries due to impaired NO biovailability is noted early in the disease process (2.5 months after start of the high fat diet).70 In contrast, at 15 months of diet, NO signaling is restored but the constriction to ET-1 is exacerbated, this response being mediated by ETB-mediated vasoconstriction, indicating that disturbances in the balance between vasodilators and vasoconstrictors are modulated during progression of the disease (Table 1).71 The observations that eNOS activity is impaired by adipokines secreted by perivascular fat72 and that coronary microvascular dysfunction correlates with increased inflammation62 suggest that inflammation and fat-derived cytokines in obesity are also important determinants of coronary microvascular dysfunction. This is consistent with the correlation of fat deposits with the decline in coronary microvascular function.57,63
In addition to the functional changes in the control of vascular tone, obesity can also result in structural remodeling of the coronary microcirculation. Histological analysis of left ventricular tissue biopsies obtained during coronary bypass surgery show significantly lower capillary densities in obese patients (Table 1).73 Importantly, both arteriolar remodeling and capillary rarefaction likely contribute to the reduced coronary flow reserve in obese patients as reported in several,60–62 though not all studies.66 Similarly, hypertrophic inward remodeling of coronary arterioles, increased stiffness as well as capillary rarefaction are reported in animal models of obesity and hypercholesterolemia.58,71,74
It is increasingly recognized that the factors released by the coronary endothelium also impact the function of the surrounding cardiomyocytes in a paracrine fashion. Loss of NO, increased oxidative stress and the ensuing tissue inflammatory response are thought to play a key role in development of left ventricular diastolic dysfunction through altering relaxation of the cardiac myocytes and increasing collagen fraction in the extracellular matrix.75,100
In conclusion, both clinical and experimental evidence indicate that obesity is an independent risk factor for coronary microvascular dysfunction, with both functional and structural alterations in the coronary microcirculation contributing to the impairments in coronary flow regulation and having a negative impact on the coronary flow reserve in these patients. Some, but not all, of these changes can be alleviated by weight loss and physical exercise.68,76 However, in order to be able to specifically address the different aspects the obesity-associated coronary microvascular dysfunction, future studies should focus on revealing the underlying mechanisms that drive the obesity-associated coronary microvascular abnormalities.
Cerebral microvasculature in obesity
Similar to the heart, the brain relies on a continuous supply of blood flow, with regional alterations in brain activity requiring corresponding changes in brain flow distribution via metabolic vasodilation, which is referred to in the brain as neurovascular coupling. Obesity-induced changes in brain microvascular structure and function have been proposed to result in disruptions in neurovascular coupling, thereby leading to vascular cognitive impairment.20,77 Importantly, the brain microvasculature is only surrounded by neurons and glia as there is no perivascular adipose tissue. Therefore, the effects of obesity likely manifest as the result of changes in neural/glia-vascular (metabolic) interactions, and hemodynamic (i.e. blood pressure) or circulating factors (hyperglycemia and dyslipidemia) that have a direct vasoactive effect or act indirectly via influencing neurons and glia.
Both pre-clinical78,79 and clinical80,81 obese human populations exhibit reduced brain blood flow and impaired vasodilation during hypercapnia. The mechanisms responsible likely relate to microvascular rarefaction,19,20,82 decreased contribution of NO to basal cerebral microvascular tone control, altered release of vasodilator prostanoids, and/or a direct effect of H+ on vascular smooth muscle ion channels.83 Impaired cerebral vasoreactivity in obesity occurs independently of clinical insulin resistance,78,84 but may also worsen with accompanying hypertension and poor glycemic control.80,81 Obese Zucker rats display impaired endothelium-dependent NO-mediated middle cerebral artery (MCA) vasodilation as well as depressed insulin-stimulated vasodilation, potentially due to increased PKC- and MAPK-activation, combined with eNOS uncoupling resulting in augmented superoxide production.82,85 Indeed, depressed vasodilator responses to hypoxia and NOS-dependent dilators, as well as enhanced constrictor responses to 5-hydroxytryptophan1 reflect potential pathological adaptations that impair neurovascular coupling. Decreased insulin-mediated vasodilation potentially contributes to impaired microvascular insulin delivery in the brain (Table 1).86 The physiological significance of this remains incompletely understood, but similar to skeletal muscle, altered insulin signaling in the brain (i.e. brain insulin resistance) can link microvascular and metabolic dysfunction, thereby leading to cognitive impairment. Collectively, the data from humans81 and rodents82,85 suggest that obesity induces endothelial dysfunction with impaired NO production/bioavailability, resulting in altered cerebral vasoreactivity.
Obesity also affects the structure of small arteries, arterioles and capillaries in the brain, with many of these changes reflecting the development of cerebral microvascular disease.20,77 Indices of cerebral microvascular disease, including cerebral microbleeds, lacunas and microlacunas, increase the vulnerability to neurodegeneration,20,77 and occur more commonly in obese individuals with insulin resistance,87 dyslipidemia,87 and central adiposity.88 In addition, genetic predispositions may also contribute to this relationship.77 Beyond these preliminary observations, the underlying mechanisms responsible for obesity-induced cerebrovascular remodeling in humans remain elusive. However, cortical microvascular density decreases,89 and the MCA undergoes eutrophic inward remodeling and progressive arterial stiffening during the progression of metabolic syndrome in obese Zucker rats.82 In diet-induced obesity in Sprague–Dawley rats, similar MCA adaptations are observed in conjunction with increased MMP-2 activity, collagen I expression, and reduced MMP-13 expression, suggesting that increases in collagen deposition contribute to vascular stiffening (Table 1).90 In the aforementioned studies,82,89,90 inward MCA remodeling coincided with the development of hypertension, while pharmacological treatment of hypertension ameliorated the remodeling.82 This suggests that obesity-induced hypertension, and not metabolic dysfunction,91 serves as the primary stimulus responsible for inducing inward remodeling of the MCA.
Although pharmacological treatment of hypertension ameliorates MCA remodeling, it does not improve cortical microvascular rarefaction in obese Zucker rats.89 Also in Rhesus monkeys, diet-induced obesity causes cortical capillary rarefaction, but without concurrent changes in blood pressure.92 In the latter study, cortical capillary rarefaction occurred alongside decreased VEGF, increased von Hippel-Lindau protein (which degrades HIF-1α) and (paradoxically) increased expression of FOXO3, eNOS, and eNOS uncoupling. In contrast to inward remodeling in the MCA, it appears that obesity-induced metabolic dysfunction and oxidative stress, and not hypertension, is responsible for the observed microvascular/capillary rarefaction (Table 1). Therefore, while inward remodeling may prevent hypertension-induced cerebral hyperperfusion, inward remodeling and microvascular rarefaction may limit cerebrovascular reserve and impair brain blood flow control.
Functional and structural microvascular deficits resulting in impaired neurovascular coupling likely reflect an acquired, and not programmed feature of obesity, suggesting that these abnormalities can be environmentally induced, prevented or even reversed.93–95 Indeed, short-term diet-induced obesity reduces prefrontal cortex blood flow in mini-pigs,94 and impairs metabolic vasodilation and precedes neuronal loss in rodents.95 Furthermore, individuals with an elevated BMI exhibit reduced basal flow79 and post-prandial vasodilation93 in the prefrontal cortex, but formerly obese individuals do not display this defect.93 Reversal of such adaptations is of growing importance as the early signs of cerebral microvascular disease can manifest very early in life, as seen in an obese 2-year-old child.96
In conclusion, the cumulative effect of cerebral inward remodeling, microvascular rarefaction, and impaired vasodilator capacity, likely contribute to obesity-related impairments in brain flow control and neurovascular coupling. Such obesity-induced changes can occur rapidly and in young individuals, but it appears these pathological adaptations are modifiable. Gathering more knowledge about mechanisms of obesity-related changes in brain microvascular structure and function along the (micro)vascular tree is essential in understanding the pathology of disease progression and developing effective prevention and treatment strategies.
Renal microvasculature in obesity
Obesity is associated with an increased risk for chronic renal failure. In a Swedish case-control study, particularly diabetic nephropathy, nephrosclerosis, and glomerulonephritis were associated with obesity,97 suggesting that obesity negatively impacts the renal microvascular bed. Indeed, clinical studies show that in obesity, afferent arteriolar vasodilation results in an increased renal blood flow (RBF), that causes a state of hyperfiltration.98 Results in experimental animals are equivocal with some laboratories showing that RBF and glomerular filtration rate (GFR) are increased 3–4 months99,100 of high fat diet, whereas others show no change in RBF and GFR.101–105 The reported increase in RBF is mostly due to an increase in renal cortical volume, vascular volume fraction and cortical perfusion, whereas filtration fraction, medullary size, and medullary perfusion showed no difference.99,100,106
Obesity is linked to increased peri-renal fat deposition,100,107 which in turn leads to low grade inflammation and renovascular endothelial dysfunction. Indeed, endothelium-dependent vasodilation to acetylcholine was impaired both in vivo98,102,103,108 and in vitro in renal arteries of obese swine,100 while endothelium-independent vasodilation to the NO donor sodium nitroprusside (SNP) was unaltered in vitro (Table 1).100 Although renal eNOS-expression may initially increase,99 prolonged exposure to high fat diet reduces expression of eNOS and promotes eNOS-uncoupling and activation of xanthine oxidase resulting in impaired bioavailability of NO in obese swine.102 Antioxidant capacity in obesity is further reduced by a decrease in SOD activity.103,108 Moreover, increased NAD(P)H-oxidase and LOX-1 expression further contribute to increased oxidative stress, which together with upregulation of iNOS, may have led to increased nitrotyrosine levels.99,102,104,105,108,109 ROS in turn, induce upregulation of renal pre-pro ET-1 and ETA receptors thereby promoting vasoconstriction (Table 1).102–104
Interestingly, incubating renal arteries of lean swine with peri-renal fat of obese animals transferred the impaired endothelial function to those arteries.100 This response appears to be the result of fat derived inflammatory molecules, and not of oxidative stress, as the endothelial dysfunction could only be reversed by neutralizing TNF-α, but not by the free radical scavenger Tempol, in vitro.100 Furthermore, anti-inflammatory treatment with thalidomide in vivo, abolished the renal increase in TNF-α, and improved endothelial function without altering oxidative stress, suggesting that increased levels of TNF-α play a vital role in impaired function of the endothelium.102
In addition to renovascular dysfunction, structural alterations in the kidney produced by obesity have also been reported (Table 1). Interestingly, obesity is associated with a change in the balance of angiogenic and anti-angiogenic factors in the kidney that favors angiogenesis. Thus, VEGF and its receptor Flk-1108 as well as Angpt2110 are increased. Moreover, protein expression of the angiogenesis inhibitor TSP-1 is decreased as a result of increased oxidative stress,99 although this is not a unanimous finding.104 Indeed, both arteriolar and capillary density in the outer cortex of the kidney are increased in animals with obesity.99,108 However, these newly formed vessels are more tortuous and erythrocyte exudation has been shown, suggesting that the newly formed vessels are leaky and immature.99,102,103,108,109 Furthermore, glomerular density is decreased in animals with metabolic derangement,109 while glomerular hypertrophy due to matrix hyperplasia and glomerular swelling are also observed in obese swine.99,109–111 Taken together, these data suggest that the newly-formed microvasculature does enhance glomerular filtration, but is rather damaged and dysfunctional. Vascular remodeling in obesity is facilitated by dynamic processes in the extracellular matrix, as an increased MMP expression, which promotes extracellular matrix (ECM) degradation was noted at 10 weeks,99 but not at 16 weeks104 of high fat diet. Furthermore, in obese swine, renal expression of tissue transglutaminase is increased, causing extracellular matrix crosslinking and vascular remodeling especially in conditions of sustained vasoconstriction.103 Moreover, microvascular media-to-lumen ratio is increased,104,108 and perivascular as well as tubulointerstitial fibrosis is observed in obese swine.104,108 The increased presence of renal M1-macrophages, and increased NF-κB expression, plasma/renal levels of TNF-α as well as activation of the TGFβ-system in the kidneys of obese animals, suggests that inflammatory cells play a central role in these processes.98–100,103,104,106
In conclusion, the increased fat or lipid deposition in and around the kidney acts as a promotor of a pro-inflammatory state with oxidative stress, endothelial dysfunction and microvascular remodeling as a consequence. Although these changes are initially reversible by switching to a healthy life style,104 as well as by interventions that lower lipids and/or oxidative stress,105,108 modulate the immune-system102 or prevent vasoconstriction,98,103 prolonged exposure results in irreversible alterations in the renal microvasculature. Indeed, although diabetes mellitus and hypertension are still the main causes of chronic kidney disease (CKD),112 it has been shown that obesity independently increases the risk of CKD and end-stage renal disease even in the absence of these known cardiovascular risk factors or nephropathy.113 As the kidney is important in clearing waste products and adipokines from the body, renal dysfunction in obesity and the ensuing increase in so-called uremic toxins, can further contribute to microvascular dysfunction in other organs and thereby contribute to development and aggravation of cardiovascular disease.114
Pulmonary microvasculature in obesity
Functional and structural changes in the pulmonary microvasculature as a result of obesity are less well studied as compared to their systemic counterparts. The comparison of the effect of obesity in the pulmonary and systemic microvasculature is of interest as the pulmonary vasculature receives the same cardiac output as the systemic vasculature and is exposed to the same circulating factors, such as glucose, cholesterol, adipokines, and inflammatory factors. Indeed, obesity is also associated with a variety of lung diseases including obstructive sleep apnea, hypoventilation syndrome, chronic bronchitis, asthma, and pulmonary embolism. Many of these diseases have an inflammatory component and it is likely that the change in circulating adipokines with obesity facilitates this pulmonary inflammation. Adiponectin exerts anti-inflammatory and protective effects against inflammatory lung diseases. In contrast, leptin is pro-inflammatory and leptin receptors are present on all inflammatory cell-types in the lung. An increase in leptin primes leucocytes for increased secretion of inflammatory cytokines and reactive oxygen species. However, exactly how a change in adipokine profile impacts pulmonary microvascular structure and function remains to be established.
An autopsy study in 1982 revealed a strong correlation between the size of atherosclerotic plaques in the aorta and the pulmonary artery. In rabbits on a high fat diet, the rate of cholesterol accumulation in the pulmonary artery exceeded that in the aortic arch initially, but at later stages of atherogenesis, the rate of cholesterol accumulation slowed in the pulmonary artery ultimately falling below accumulation rates in the aortic arch. Obesity is also associated with structural changes in the pulmonary microvasculature (Table 1). Increased medial thickness of both pulmonary small arteries and veins, and increased muscularization of pulmonary arterioles were observed in obese humans compared to controls at autopsy. Similarly, an increased wall to lumen ratio was found in pulmonary arterioles of obese rats as compared to lean control rats. Interestingly, adiponectin deficiency has been shown to result in pulmonary microvascular remodeling, with an increased muscularizaton of pulmonary microvessels.
These structural pulmonary microvascular changes in obesity resemble those found in post-capillary pulmonary hypertension. Indeed, the prevalence of pulmonary hypertension (PH) is increased in obese subjects, and is associated not only with sleep apnea and hypoxemia, but also with left ventricular diastolic dysfunction, resulting in increased left atrial pressure. The association between obesity and elevated pulmonary artery pressures has also been found in both obesity prone and overtly obese Zucker rats on a high fat diet and in obese FH-swine. Similar to humans, the elevated pulmonary artery pressure in the obese swine was mainly due to an increase in left atrial pressure as pulmonary vascular resistance was only mildly elevated. However, despite the mildly elevated pulmonary artery pressure and pulmonary vascular resistance at rest, the pulmonary vasodilator response to exercise was preserved in swine with hypercholesterolemia, at a time when systemic vasodilation was reduced.
The effect of obesity and high fat diet on pulmonary vascular function has only been assessed in experimental animals and may depend on the duration of exposure and the stimulus used to assess vascular function. Thus, pulmonary artery vasodilation in response to methacholine is enhanced in rabbits given a 2% cholesterol diet over a period of 2 weeks. Exposure of Zucker rats to high fat diet for 18 weeks has no effect on eNOS expression or acetylcholine-induced NO production in either conductance or resistance pulmonary arteries. Similarly, the increase in pulmonary vascular resistance in response to eNOS inhibition is comparable in healthy swine and in FH-swine on a high fat diet for 6 months, both at rest and during exercise, although ATP-induced NO mediated vasodilation was reduced. Also, the pulmonary vasodilator response to SNP is maintained in both obese Zucker rats and FH-swine exposed to high fat diet. In the latter group, vasodilation to PDE5-inhibition is also preserved. Altogether, these data suggest that pulmonary vasodilation through the NO pathway is well-preserved in obesity (Table 1).
In general, vasoconstrictor responses in the pulmonary microvasculature are reduced in obesity. Thus, vasoconstriction to KCl, phenylephrine, serotonin, and hypoxia are reduced in pulmonary resistance, but not conductance arteries of obese Zucker rats. In FH swine, endothelin receptor blockade did not reduce pulmonary vascular resistance, while it produced pronounced vasodilation in normal swine (Table 1). This loss of ET-mediated vasoconstriction was accompanied by slightly lower plasma ET-levels. These data suggest that reducing the effect of vasoconstrictors may serve as an early compensatory mechanism to maintain pulmonary vascular resistance as low as possible, to limit the workload of the right ventricle.
The term obesity paradox refers to the observation that although obesity is a well-known risk factor for the development of cardiovascular disease, mortality rate in many cardiovascular disorders, once established, is lower in obese patients. Thus, there is evidence suggesting that when CKD is present, mortality is higher in underweight patients and lower in patients with obesity class I, but not with class II or III. Similarly, in patients with established coronary artery disease, heart failure, or stroke, there are large cohort studies suggesting that mortality is reduced in patients with obesity class I, particularly in the short term. Also in patients with established PH, a recent study shows a strong inverse correlation between obesity and mortality, in that both in pre-capillary and out-of proportion post-capillary PH, obese patients had a significantly lower mortality (46% vs. 10% mortality in pre-capillary and 40% vs. 11% in post-capillary PH for lean and obese subjects). Moreover, BMI was the strongest predictor of mortality in a COX hazard analysis, followed by NYHA functional class.
However, there is also concern about selection bias, insufficient control for cardiorespiratory fitness, inadequate determination of body fat localization (i.e. visceral vs. subcutaneous), and underweight vs. normal weight subjects, in these studies. Indeed the overall consensus is that losing weight by for example exercise trainings confers protection against mortality in cardiovascular disease. A large part from these benefits of exercise training can be ascribed to improved microvascular function, and reduced inflammation. Another intriguing possibility is that exercise training results in secretion of ‘myokines’ from skeletal muscle and ‘cardiomyokines’ from cardiac muscle such as FGF21 and irisin, that act in an endo- and/or paracrine fashion on perivascular and epicardial adipose tissue, and induce a ‘browning’ phenotype.
Summary and conclusion
Obesity is a well-established risk factor for microvascular dysfunction throughout the body (Figure 1). This microvascular dysfunction is likely initiated by transient elevations of circulating free fatty acids, and perpetuated by changes in adipokines and inflammatory cytokines released from visceral as well as perivascular adipose tissue. These factors contribute to endothelial dysfunction as well as insulin resistance in the microvasculature, thereby affecting function of different organs not only by impairing tissue perfusion, but also through altering the release of paracrine factors from the endothelial cells. Thus, although the mechanisms can differ between regional vascular beds (Table 1), microvascular dysfunction is a central common pathway that may explain exercise-intolerance as well as the higher prevalence of chronic kidney disease, microvascular dementia, coronary microvascular angina, heart failure with preserved ejection fraction, and pulmonary hypertension in obese subjects.