Does Hypoadiponectinemia Explain the Increased Risk of Diabetes and Cardiovascular Disease in South Asians?

The metabolic and vascular phenotype of the South-Asian patient

Studies to date have documented a plethora of metabolic and vascular risk factors and abnormalities in people of South-Asian descent. First, there are significant differences in body composition between South Asians and Caucasians. South Asians typically have thin limbs (suggestive of smaller muscle mass) combined with increased central obesity, as evidenced by higher waist-to-hip ratio and greater subscapular-to-triceps skinfold ratio (7,17). Importantly, visceral fat mass is higher in South Asians compared with Caucasians at the same BMI (11). Consistent with greater visceral obesity, South Asians exhibit increased insulin resistance, as consistently demonstrated by fasting hyperinsulinemia, reduced insulin sensitivity on oral glucose tolerance tests, and reduced glucose disposal during euglycemic-hyperinsulinemic clamp (11,18–22). Similarly, South-Asian ethnicity is associated with lipid abnormalities that are consistent with increased visceral obesity. Specifically, HDL cholesterol concentration is typically lower in South Asians than in Caucasians (5,6). In addition, HDL particle size may be reduced, reflecting a preponderance of smaller, less cardioprotective HDL particles (23). Indeed, in the Framingham Offspring Study, South-Asian men had 1) lower levels of large, protective HDL cholesterol, 2) increased concentration of small HDL cholesterol, and 3) smaller HDL particle size (23). LDL cholesterol particle size may be unfavorable as well. While total LDL levels are typically comparable with those of other ethnic groups, an increased prevalence of atherogenic, small dense LDL particles has been demonstrated in South Asians (24). The regulation of lipolysis is another process that may be abnormal in South Asians, as evidenced by higher plasma levels of nonesterified fatty acids and impaired insulin-mediated suppression of plasma nonesterified fatty acids compared with Caucasians (25). Excessive lipolysis in adipose tissue, the site of release of nonesterified fatty acids, could contribute to lipid deposition in nonadipose tissues. Indeed, increased intramyocellular lipid content has been documented in South-Asian subjects (26). Finally, the endothelium represents yet another tissue in which abnormalities have been noted in individuals of South-Asian descent. Specifically, significant impairment of flow-mediated endothelium-dependent vasodilation has been shown in healthy South Asians (27). This finding is of interest in the current context, given recent evidence (28,29) that markers of endothelial dysfunction may predict the development of type 2 diabetes and CVD.

It is thus apparent that South-Asian ethnicity is associated with multiple risk factors for type 2 diabetes and CVD. Importantly, however, conventional vascular risk factors, such as diabetes and dyslipidemia, do not fully reconcile the excess cardiovascular risk associated with South-Asian ethnicity (2,6,9). As such, additional factors likely contribute to metabolic and vascular risk within this ethnic group. Nontraditional factors studied to date in this context include inflammatory mediators and prothrombotic factors. Increased serum concentrations of the inflammatory biomarker C-reactive protein, correlating with visceral obesity, have been demonstrated in South Asians (30,31), though prospective study of the association of C-reative protein with incident type 2 diabetes or CVD in South Asians is lacking. The same limitation applies to the prothrombotic factor plasminogen activator inhibitor-1, which has also been shown to be upregulated in healthy South Asians (6,21). Other prothrombotic abnormalities reported in South Asians include increased serum levels of fibrinogen, homocysteine, and lipoprotein(a) (6,32). Importantly, however, none of these nontraditional factors have yet reconciled the increased risk of type 2 diabetes and CVD in South Asians.

Adiponectin and South-Asian ethnicity

One candidate factor of particular interest with respect to both diabetes and vascular risk is adiponectin, an adipokine with putative insulin-sensitizing, antiatherogenic, and anti-inflammatory properties (33). Secreted exclusively by adipose tissue, adiponectin circulates at a relatively high concentration in oligomeric complexes, consisting of trimers, hexamers, and high–molecular weight multimers of 12–18 subunits (34). Total serum adiponectin concentration (i.e., consisting of all multimeric isoforms) is inversely proportional to both intra-abdominal fat and insulin resistance, consistent with the observation of hypoadiponectinemia in patients with type 2 diabetes (35,36). Importantly, low serum concentration of adiponectin predicts the future development of 1) insulin resistance in Pima Indians (37) and 2) incident type 2 diabetes in various populations, including Pima Indian, Caucasian, Japanese, and native South-Asian people (38–41). Thus, hypoadiponectinemia has emerged as a potential factor in the pathophysiology of insulin resistance and type 2 diabetes.

Adiponectin has also been proposed as a factor linking adipose tissue and vascular function (33). Indeed, adiponectin accumulates in the walls of injured blood vessels and displays antiatherogenic bioactivity such as attenuation of tumor necrosis factor-α–mediated inflammation, inhibition of vascular smooth muscle cell proliferation, and suppression of macrophage-to-foam cell transformation (33). Consistent with these effects, hypoadiponectinemia has been documented in patients with CAD (42). Furthermore, low serum levels of adiponectin have been shown to predict incident CVD in 1) healthy men participating in the Health Professionals Follow-up Study (43), 2) men with type 2 diabetes (44), 3) young adults with type 1 diabetes (45), and 4) patients with end-stage renal disease (46). As such, hypoadiponectinemia has been implicated as a possible factor in the development of CVD.

Given the increased risk of type 2 diabetes and CVD in South Asians, it is intriguing to note that hypoadiponectinemia within this ethnic group has been consistently observed in clinical studies (Table 1) (21,25,47–50). Although variability exists in the adiponectin levels reported in these studies (likely reflecting differences between the study populations with respect to factors such as sex and pregnancy status), hypoadiponectinemia in South-Asian subjects is a consistent finding in each study (21,25,47–50). Indeed, hypoadiponectinemia in South Asians has been documented in both sexes (21,25,47,49,50) and during pregnancy (48). Furthermore, these differences in adiponectin concentration are of such a substantial magnitude that they are detectable even in modestly sized studies involving as few as 15–31 South-Asian subjects (21,47,48). In healthy young men, Abate et al. (25) reported that plasma adiponectin concentration remained significantly lower in South Asians (n = 79) compared with Caucasians (n = 61), even after adjustment for differences in total body fat content, waist circumference, and truncal skinfold thickeness. Similarly, in a study of 180 pregnant women, South-Asian ethnicity emerged as the strongest independent and negative determinant of adiponectin concentration on multiple linear regression analysis, after adjustment for potential covariates including glucose intolerance, insulin resistance, family history of diabetes, and prepregnancy BMI (48). These data raise the possibility that hypoadiponectinemia may be a generalized phenomenon in South Asians that could contribute to the excess diabetes and vascular risk within this ethnic group.

Hypoadiponectinemia is consistent with several aspects of the metabolic and vascular phenotype of South-Asian patients (Fig. 1). First, given the inverse relationship between serum adiponectin and intra-abdominal fat (35), hypoadiponectinemia is concordant with the increased central obesity and visceral fat observed in South Asians. Second, low adiponectin in South Asians is consistent with the increased insulin resistance in this ethnic group. Indeed, ethnicity-associated hypoadiponectinemia may relate to the finding of insulin resistance in young healthy South Asians (21,25). Third, given that studies have repeatedly shown that adiponectin is independently associated with HDL cholesterol concentration (35,47), hypoadiponectinemia is consistent with the reduced HDL levels seen in South Asians. Fourth, intramyocellular lipid content, previously reported to be increased in South Asians (26), is inversely related to adiponectin concentration (51). Finally, hypoadiponectinemia is associated with impaired endothelium-dependent vasodilation (33), an aspect of endothelial dysfunction that has been demonstrated by reduced flow-mediated dilatation in healthy South Asians (27). Taken together, these data support a model in which the pathologic effects of visceral obesity in South Asians may be mediated by hypoadiponectinemia, leading to the expression of metabolic and vascular risk factors and subsequent disease.

Consistent with the proposed model, hypoadiponectinemia has been associated with metabolic disease in South Asians. In a study of 200 subjects in Chennai, India, low adiponectin was associated with both type 2 diabetes and the metabolic syndrome (52). Furthermore, in a prospective study from Chennai involving 91 subjects with impaired glucose tolerance, Snehalatha et al. (41) found that baseline hypoadiponectinemia independently predicted the future development of type 2 diabetes, after adjustment for covariates including waist circumference. These data support the notion that hypoadiponectinemia in South Asians is clinically significant. Further prospective study of the associations of adiponectin with incident type 2 diabetes and CVD in this ethnic group is needed.

The vast majority of clinical studies to date have evaluated total circulating adiponectin concentrations, as measured by current commercial assays. It has recently emerged, however, that adiponectin circulates in oligomeric complexes, consisting of trimers, hexamers, and high–molecular weight multimers of 12–18 subunits (34,53). The physiologic activity of adiponectin appears to be primarily influenced by the relative distribution of these isoforms, which activate different signal transduction pathways and hence may mediate different functional properties of the protein (53,54). While conflicting results have been reported (55) regarding the precise pathways affected by specific isoforms, the high–molecular weight complex has emerged as a mediator of both glucose-lowering and antiatherogenic activity (56,57). Indeed, high–molecular weight adiponectin exhibits insulin-sensitizing and vasculature-protective effects (56,57). Furthermore, the relative proportion of adiponectin in high–molecular weight form is decreased in patients with type 2 diabetes and CAD, respectively, compared with control subjects (56,57). Lara-Castro et al. (58) reported that the hypoadiponectinemia observed in patients with type 2 diabetes may be entirely attributable to decreased levels of the high–molecular weight isoform. Interestingly, we have recently demonstrated that pregnant South-Asian women exhibit low serum levels of high–molecular weight adiponectin, thereby documenting ethnic variation in adiponectin isoform composition (59). Indeed, on multivariate analysis, South-Asian ethnicity emerged as an independent and negative determinant of the relative proportion of total adiponectin in high–molecular weight form (59). Thus, the hypoadiponectinemia observed in South Asians may reflect selective reduction of the high–molecular weight isoform. This concept is theoretically consistent with the low levels of HDL cholesterol in South Asians, since high–molecular weight adiponectin strongly correlates with HDL and has even emerged as an independent predictor of the change in HDL over time (55). Furthermore, low serum levels of high–molecular weight adiponectin have recently been shown to correlate with other metabolic abnormalities commonly observed in South Asians, including central obesity, insulin resistance, and small HDL particle size (58). Collectively, these data raise the possibility that deficiency of high–molecular weight adiponectin may be a factor contributing to diabetes and vascular risk in South Asians. Further study is needed to 1) determine whether deficiency of high–molecular weight adiponectin is a generalized phenomenon in South Asians and 2) prospectively evaluate its relevance to incident type 2 diabetes and CVD within this ethnic group.

If further clinical studies establish hypoadiponectinemia as a key contributory factor in the development of type 2 diabetes and CVD in South Asians, then therapeutic strategies to increase adiponectin levels may be particularly relevant in this ethnic group. For instance, weight reduction can increase total adiponectin and has been shown to raise levels of the high–molecular weight and hexameric isoforms, while reducing the trimeric form of the protein (55). Similarly, thiazolidinediones can increase total and high–molecular weight adiponectin in both nondiabetic and diabetic subjects (56,60). In this context, one could hypothesize that there may be a unique role for thiazolidinedione therapy in the South-Asian patient. Further clinical study is needed.

In summary, hypoadiponectinemia in South Asians has been consistently demonstrated in clinical studies to date. As low adiponectin relates to several elements of the metabolic and vascular risk profile of this ethnic group, the possibility emerges that hypoadiponectinemia may contribute to the increased risk of type 2 diabetes and CVD in South Asians. Further study of adiponectin biology within this ethnic group may provide important insights into the pathophysiology of metabolic and vascular disease and ideally may identify therapeutic strategies for optimal care of the South-Asian patient.