Pharmacological and non-pharmacological interventions to influence adipose tissue function

Intentional weight loss is associated with lowering blood pressure in hypertensive patients, improvements in the lipid profile and a decreased incidence of diabetes[17‐19]. Intentional weight loss induced by taking dietary measures improves endothelial function[20‐22], lowers systemic markers of inflammation[23] and ameliorates insulin resistance[24]. In patients with and without the metabolic syndrome, 7% weight loss already reduced the prevalence of components of the metabolic syndrome such as systolic blood pressure, plasma glucose levels, triglycerides and high-density lipoprotein (HDL) cholesterol[25]. Although there are no randomized controlled trials (RCT) available showing the benefit of diet induced weight reduction on overall or cardiovascular mortality there is a clear association between obesity and mortality[26‐28]. The currently ongoing Look AHEAD trial compares intensive lifestyle intervention including weight loss with diabetes support and education and has shown promising results on diabetes control and cardiovascular risk factors[18]. Influencing plasma adipokines levels such as adiponectin by a dietary intervention seems especially dependent on a reduction in fat mass[29].

Lowering the daily caloric intake by 500 kilocalories leads to a decrease in bodyweight between 5-10% and a decrease in body mass index (BMI) by 2.0 and 2.9 kg/m2[30‐32]. This weight loss was associated with a 25% fall in leptin plasma concentrations while there was no effect on adiponectin levels[30, 32]. Weight loss of 5-10% increases insulin sensitivity while marginally increasing plasma levels of adiponectin and reducing plasma levels of C-reactive protein (CRP), interleukin-6 (IL-6) and the soluble tumour necrosis factor-α (TNF-α) receptor without an effect on plasma levels of TNF-α, indicating an improvement in adipose tissue function[31, 32].

If caloric intake is further restricted by 600-1000 kcal/day, bodyweight decreases between 7 and 12%[25, 33] resulting in a 46-58% fall in plasma leptin levels and an increase of 42-65% in adiponectin levels[25]. Apparently, diet-induced weight loss is an effective strategy for improving adipose tissue function but at least 10% weight loss is needed to improve plasma concentrations of adiponectin and inflammatory markers such as CRP.[34] Besides the amount of weight loss the duration of the weight loss period might also influence plasma adiponectin levels with adiponectin levels increasing during the weight loss maintenance period after a weight loss of 11-12% in 8 weeks[35].

Salicylates are among the most commonly used nonsteroidal anti-inflammatory drugs and have their main action through cyclooxygenase (COX) inhibition. Studies investigating the effects of salicylates on adipose tissue function have especially implicated specific COX-2 inhibition as the mechanism by which salicylates may improve adipose tissue function[80, 81]. Besides COX inhibition, salicylates also act through inhibition of the activity of inhibitor of nuclear factor kappa-B kinase subunit beta (IKK-β) leading to decreased phosphorylation of inhibitor of NF-κB (IκB) and therefore to a reduction in translocation of Nuclear Transcription factor kappa-B (NF-κB) to the nucleus [82]. Besides the direct anti-inflammatory effects, some data suggests a possible role of PPAR-γ agonism which is of great importance in adipocyte differentiation, function and body fat composition. For example, 5-aminosalicylic acid increases PPAR-gamma expression, promotes its translocation from the cytoplasm to the nucleus, and permits the recruitment of co-activators and the activation of a peroxisome-proliferator response element-driven gene in human epithelial cells[83]

Although high dose acetylsalicylic acid (1 to 1,6 gram) has been shown to reduce fasting and post load glucose levels in patients with type 2 diabetes, the clinical use of high dose acetylsalicylic acid is limited by the increased risk of bleeding)[84]. Low dose acetylsalicylic acid (100 mg and 300 mg) had no effect on IL-6 or CRP levels in patients with type 2 diabetes during 6 weeks[85]. Salsalate at a dose of 3 grams per day however does lower fasting glucose levels and glucose levels after a oral glucose tolerance test in patients with obesity, by increasing insulin levels via an unknown mechanism[86]. In patients with type 2 diabetes, salsalate in doses of 3 and 4,5 grams per day improved insulin resistance as measured during a hyperinsulinemic euglycemic clamp, fasting and post challenge glucose levels, decreased free fatty acid (FFA) levels and increased adiponectin levels by 35-45% without an effect on bodyweight[87]. This effect of salsalate on adipose tissue dysfunction may be mediated by the before mentioned anti-inflammatory effect or by the possible PPAR-γ agonistic action of salicylates, leading to a reduction in insulin resistance.

Although earlier reports have found a possible link between the use of beta blockers and the development of diabetes, some newer beta-blockers are investigated for their beneficial effects on adipose tissue dysfunction[88, 89]. The relation between beta-blockers and diabetes can be explained by β₂ receptor blockade, induced reduction in thermogenesis and subsequent weight gain [90‐92]. A combined β1 and β2-adrenoceptor agonist is capable of down regulating adiponectin and up regulating TNF-α mRNA in murine adipocytes[93].

Indeed, some of the newer beta blockers do have beneficial effects on insulin resistance and adipokines without changes in weight. For example nebivolol (5 mg daily), which has β₂ intrinsic sympaticomimetic action, increases plasma adiponectin levels in overweight patients with hypertension[94] Celiprolol (up to 400 mg daily), a combined β₁ antagonist and β₂ agonist reduces plasma leptin levels without a change in bodyweight in patients with dyslipidemia [95].

Inhibition of mineralcorticosteroid receptor activation by use of aldosterone antagonists is used in the treatment of heart failure and hypertension. Besides an effect on blood pressure, spironolactone is capable of inhibiting TNF-α, IL-6 and Interferon- γ (IFN-γ) production in isolated human mononuclear cells in vitro[96]. Far less is known about the effect of aldosterone antagonists on adipose tissue dysfunction. Adipose tissue is capable of producing an unidentified mineralcorticoid releasing factor that may stimulate aldosterone production[97]. The mineralcorticoid receptor has an important role in adipocyte differentiation as witnessed by diminished differentiation of 3T3-L1 adipocytes in the presence of dexamethason and spironolactone [98]. In obese diabetic mice, blocking the mineralcorticoid receptor reduced the expression of proinflammatory cytokines in adipose tissue while it lead to an increased expression of adiponectin in cardiac and adipose tissue[99]. Further evidence for an important role for the mineralcorticoid receptor in adipose tissue comes from a study in obese mice where blocking the mineralcorticoid receptor with eplererone ameliorated insulin resistance, decreased the number of hypertrophic adipocytes and infiltrating macrophages[100]. Furthermore, eplererone was also capable of blunting aldosterone and H₂O₂ induced radical oxygen species and dysregulated expression of obesity related genes in 3T3-L1 adipocytes. This data from in vitro and murine studies shows that aldosterone might play a relevant role in adipocyte biology. Indeed, although not a strict adipokine, plasma PAI-1 levels were reduced by spironolactone in patients with type 2 diabetes and diabetic nephropathy[101].

Angiotensin Converting Enzyme Inhibitors (ACE-i) are widely used in the treatment of heart failure and hypertension. In a large randomized trial, ramipril was associated with a lower incidence of diabetes, compared to placebo, in patients at high cardiovascular risk [102]. In patients with cardiovascular disease and impaired fasting glucose, ramipril did not reduce the incidence of diabetes mellitus but was associated with regression to normoglycaemia[103].

Circulating levels of angiotensin II are associated with changes in VAT in humans[104]. ACE-inhibitors may affect insulin resistance by reducing plasma concentrations of angiotensin II. Angiotensin II increases serine phosphorylation of the insulin receptor, insulin receptor substrate 1 and phophadidylinositol-3-kinase leading to a state of insulin resistance[105]. Angiotensin II might also influence insulin resistance via a direct pro-inflammatory effect on adipocytes and subsequent changes in MCP-1, IL-6 and IL-8 production via the NF-κB pathway and increased production of leptin via an ERK1/2 dependent pathway in a murine model[106‐108]. Finally, ACE-i decreases total body fat mass and plasma leptin levels in a murine model[109].

Lisinopril binds to PPAR-γ, although with a low binding affinity, suggesting a possible role for a PPAR-γ agonistic action for ACE-i[110]. ACE-i are less effective than angiotensin II type 1 receptor blockers (ARB) in raising plasma adiponectin levels[111‐113], which is likely a consequence of the different effects on PPAR-γ.

The ARB valsartan reduces the risk of developing type 2 diabetes mellitus in patients with hypertension[114]. In clinical studies it is shown that insulin resistance is indeed reduced by the use of ARBs[115, 116]. Apart from blockade of the angiotensin II type 1 receptor, ARBs function as partial agonists of PPAR-γ, even in the absence of a functional AT-II receptor[117, 118]. In a murine model, plasma adiponectin levels were elevated after treatment with irbesartan without a change in adiponectin mRNA levels, suggesting a post-transcriptional mechanism[119]. The effect on PPAR-γ is further shown by studies investigating the effect of ARBs on adipose tissue distribution. Telmisartan decreases VAT by 10%, as measured by CT, without having an effect on subcutaneous fat area[120, 121]. ARBs also have anti-inflammatory effects as seen by lowering effect on plasma TNF-α and IL-6 levels in patients with diabetes and hypertension[120, 122] Telmisartan, but not valsartan, was shown to attenuate TNF-α induced IL-6 production by vascular smooth muscle cells in a PPAR-γ dependent manner[123]. These PPAR-γ agonistic effects of ARBs result in higher plasma levels of adiponectin[120, 121, 124], although no effect was observed on high-molecular weight adiponectin levels[115, 116].

Statins might have a various direct effects on adipose tissue function by inhibiting Toll-like receptor-4 (TLR4) triggered expression of IFN-γ in macrophages, which are abundant in adipose tissue, and through increasing PPAR- γ expression[125, 126]. Besides direct effects on adipose tissue, statins are also capable of reducing inflammation in general as measured by reduced plasma CRP levels[127]. Incubation of murine 3T3-L1 adipocytes with blood samples from patients treated with pravastatin induced adiponectin production[128]. However, pravastatin, a hydrophilic statin[129], did not alter insulin sensitivity, or leptin and adiponectin plasma concentrations in healthy subjects[130]. Pravastatin treatment however did elevate plasma adiponectin levels more in patients with lower baseline levels compared to patients with higher baseline adiponectin plasma concentrations[131, 132].

Due to differences in lipophylicity, statins may have different effects on adipose tissue function. Atorvastatin, which is more lipophylic than pravastatin[129] increases adiponectin levels in patients with coronary artery disease (CAD) or at high risk for CAD while having no effect on adiponectin in patients with diabetes [133‐136]. Simvastatin, the most lipophylic statin, decreases adiponectin [137, 138]. Rosuvastatin, a very hydrophilic statin, was able to lower visfatin levels in patients with increased risk for cardiovascular disease while simvastatin had no effect[139, 140]. This suggests a beneficial effect of hydrophilic statins over lipophylic statins on adipose tissue dysfunction. Initially statins were thought to reduce the incidence of diabetes[141], but two recent meta analyses of statin trials on the incidence of diabetes showed that there is no or even a small increased risk of diabetes due to statin treatment without clear heterogeneity between statins[142, 143].

Fibrates decrease the incidence of type 2 diabetes mellitus[144], by a PPAR-α agonistic effect[145]. The PPAR-α agonistic effects of fibrates also include an anti-inflammatory regulatory action on macrophages by interfering with the NF-κB and AP-1 pathways[146]. Besides the PPAR-α agonistic effect, some fibrates like bezafibrate, can be seen as pan-PPAR agonists and therefore may have effects through PPAR-γ and/or PPAR-β/δ[147, 148] As mentioned before, this could be significant as especially PPAR-γ is of major importance for adipocyte differentiation and function. Fibrates inhibit the expression of PAI-1 in human adipocytes and preadipocytes, an effect which is blunted when cells are co-incubated with a PPAR-α inhibitor[149]. In a murine model, fenofibrate increased adiponectin and visfatin mRNA levels while decreasing the expression of TNF-α in the VAT without an effect on serum TNF-α levels[150]. Short-term treatment effects of fibrates on adipose tissue function are seen by lower TNF-α, IL-6, PAI-1, MCP-1 and RBP-4 plasma levels during treatment[151, 152], and by an increase in high molecular weight adiponectin levels by 12% in patients with hypertriglyceridemia[153].

Peroxisome Proliferator-Activated Receptors or PPARs are ligand-activated transcription factors that belong to the nuclear receptor superfamiliy. While the thiazolidinedione (TZD) rosiglitazone is a selective PPAR-γ agonist, pioglitazone exerts PPAR-γ and -α agonistic activity which may account for the differential metabolic effects of pioglitazone and rosiglitazone. Thiazolidinediones have been investigated as potential drugs in preventing type 2 diabetes. Treatment with rosiglitazone during 3 years lowered the incidence of diabetes mellitus type 2 (HR 0.38, 95% CI 0.33-0.44).[154].

Thiazolidinediones may directly increase insulin sensitivity in the liver and adipose tissue where it is of critical importance for adipocyte differentiation. Indeed, as a consequence of PPAR-γ agonism, thiazolidinediones increase SAT-mass[155]. PPAR-γ agonists are thought to promote free fatty acid uptake and storage in adipocytes and may therefore protect the liver and muscle from excess levels of free fatty acids and their toxic effects, resulting in insulin resistance. Also, PPAR-γ agonists may have indirect effects on insulin resistance by altering adipocytokine production. Pioglitazone increases high molecular weight adiponectin and decreases TNF-α levels and RBP-4 levels in patients with type 2 diabetes[155‐157].In addition, this effect of pioglitazone on plasma levels of adiponectin is highly predictable on baseline levels[158]. Rosiglitazone increases leptin levels as would be expected due to the expansion of the SAT compartment and has effects on adipocytokine production as shown by lowering PAI-1 levels, which is partly dependent on adiponectin, and increasing adiponectin plasma levels [159‐161]. Although PPAR-γ agonists have shown considerable beneficial effects on adipose tissue function, concerns about cardiovascular safety remain. Both thiazolidinediones are associated with a 3-4 kg increase in bodyweight probably due to fluid retention which leads to an increased risk for heart failure[162, 163] Rosiglitazone therapy is associated with an increased risk for the occurrence of myocardial infarctions which has lead to the withdrawal of this drug from the market in 2010[164]. As beneficial vascular effects are seen with pioglitazone, current research is focussing on other dual PPAR-α/γ agonists to improve not only glycemic control but also lipid levels and potentially reducing vascular risk[165].

Metformin reduces the incidence of type 2 diabetes in patients with elevated fasting and post-load glucose concentrations indicating an effect of metformin in reducing insulin resistance[166]. Apart from affecting glucose uptake in the liver and in peripheral tissues, metformin has anti-inflammatory properties by inhibiting NF-κB and blocking the PI3K-Akt pathway in human vascular wall cells[167]. Recent evidence suggests a possible role of metformin on AMP-activated protein kinase dependent lipolysis in adipocytes which may lead to lower plasma levels of fatty acids and therefore to improvement in adipose tissue function[168]. Production of PAI-1 by human subcutaneous adipose tissue (SAT) is inhibited by metformin in vitro, showing a potential direct effect of metformin on adipose tissue function[169]. However in a study with lean and obese patients with and without diabetes, metformin did not result in a reduction of BMI, nor did it affect plasma adiponectin levels after 4 months of treatment[170] Other studies have shown that metformin decreases plasma concentrations of MIF in obese patients and also decreases vaspin while increasing omentin plasma concentrations in overweight women with polycystic ovary syndrome, without an effect on bodyweight[171‐173]. These results indicate a direct effect of metformin on adipose tissue in humans, beyond an effect through weight reduction. Apart from these direct effects on adipose tissue function, metformin may also work through effects on body composition. Metformin does not affect the amount of VAT, but reduces SAT, total body fat percentage, BMI and waist circumference in obese children and adolescents[174]. It can be concluded that metformin, which has an important place in the treatment of type 2 diabetes, also has direct beneficial effects on adipose tissue function.