The purpose of this study was to assess racial differences in the relationships among IMCL, insulin sensitivity, and adiposity in AA and EA adults. In particular, we (26
) and others (2
) have reported racial differences in the metabolic syndrome trait cluster, however, no data have addressed whether IMCL is a determinant of insulin resistance in AA adults as has been reported in EA. To our knowledge, this is the first study to definitively assess the role of ethnicity on the relationship between IMCL, insulin sensitivity, and body composition in EA and AA adults. We assessed insulin sensitivity as GDR using the gold-standard hyperinsulinemic–euglycemic clamp, IMCL content by proton-magnetic resonance spectroscopy, and total and regional body composition by dual-energy X-ray absorptiometry and circumferences. Mean values for adiposity, body composition, IMCL, and insulin sensitivity were equivalent in our EA and AA subgroups; however, we found marked racial differences in the relationships between IMCL and metabolic and adiposity parameters.
When we analyzed the main effects of IMCL and ethnicity, we found no evidence of an ethnic difference in insulin sensitivity, which is consistent with previous studies using the euglycemic clamp method of measuring insulin sensitivity (22
). When we accounted for the interaction between IMCL and ethnicity, however, our data revealed an ethnic difference in insulin sensitivity (GDR) which is dependent upon IMCL content. We found that, at lower IMCL levels, AAs and EAs have similar levels of insulin sensitivity. At higher levels of IMCL, AAs are respectively more insulin sensitive.
IMCL was significantly and negatively related to insulin sensitivity in EA independent of BMI, supporting previous findings on the relationship between insulin sensitivity and IMCL (19
). This is also consistent with a previous study in lean nondiabetic offspring of patients with type 2 diabetes which showed that insulin-resistant offspring have a substantially higher IMCL content than insulin-sensitive offspring matched for age, sex, BMI, percent fat, waist-to-hip ratio, and physical fitness (6
). In AA, on the other hand, IMCL was not correlated with insulin sensitivity.
In EA, IMCL was extensively correlated with BMI and regional measures of adiposity including waist circumference, trunk fat, and leg fat. IMCL was not related to measures of adiposity in AA. The difference shown in between the IMCL and waist circumference relationships in AA as compared to EA is striking. In EA, IMCL is highly correlated with waist circumference, but this relationship is not present in AA. It is interesting to note that no differences were seen in IMCL content between AA and EA. The finding that IMCL is related to waist circumference and other measures of obesity in EA but not AA, combined with the reduced VAT and liver fat accumulation that is reported in AA (11
), indicates that the shuttling of fat to ectopic stores in response to increased obesity and insulin resistance may not occur in AA, in contrast to that which is observed in EA. It is important to note that the correlations between IMCL and adiposity measures reduced to nonsignificant levels when BMI was controlled (data not shown). This result is expected, given the high degree of collinearity between BMI and other measures of adiposity, and does not change our interpretation of the correlations.
Waist circumference was significantly correlated with insulin sensitivity in both EA and AA, but overall percent fat was not correlated with insulin sensitivity in either group. When controlling for BMI, the correlation between waist circumference and insulin sensitivity diminished, however trunk-to-leg fat ratio in EA and waist-to-hip ratio in AA were significantly correlated with insulin sensitivity after controlling for BMI. These findings indicate that insulin sensitivity has a stronger relationship with central adiposity than it has with general adiposity. Our data are consistent with previous findings that central adiposity plays a substantial role in insulin sensitivity and cardiometabolic disease risk across ethnicities (27
). In 12,814 AA and EA men and women participating in the ARIC study, waist circumference was found to be predictive of developing type 2 diabetes over the 9-year study across both ethnicities and both genders (31
). From the same study, it was reported that BMI and waist-to-hip ratio explain 39.9% of the difference in relative risk of type 2 diabetes between AA and EA (2
). Waist circumference, more so than percent fat, was also found to be highly correlated with several metabolic syndrome measures in both AA and EA adults (32
). Furthermore, a study spanning Europeans and African-Caribbeans found that waist circumference had the highest impact among several metabolic measures on glucose tolerance (33
). A possible explanation for this finding is that waist circumference reflects VAT mass, which is known to relate highly to insulin resistance, even in AA, who have been shown to store less visceral fat than EA (11
); however, subcutaneous abdominal fat has also been strongly and independently correlated to insulin sensitivity in AA (34
). Therefore, while mounting evidence links central adiposity to insulin resistance and the metabolic syndrome trait cluster in AA, it is uncertain whether this relationship is predominantly driven by VAT or subcutaneous adipose tissue or a combination of the two. Since insulin sensitivity was associated with central adiposity, but not with overall percent body fat in both of our groups, it is tempting to argue that the relationship is driven more by VAT than subcutaneous adipose tissue in both EA and AA, although this is merely speculative because VAT was not assessed in our groups.
The reasons that AA and EA exhibit different relationships between IMCL and insulin sensitivity are unknown. One possible explanation is that intramuscular lipid could be compartmentalized differently in AA vs. EA, as has been observed in endurance athletes vs. type 2 diabetic individuals. Indeed, IMCL accumulation occurs in the skeletal muscle of endurance-trained individuals and is associated with insulin sensitivity in this group (35
). Muscle from endurance-trained athletes displays a storage pattern characterized by lipid in droplets adjacent to the mitochondria, presumably providing the athlete with an enhanced ability to utilize the lipid as substrate during training. Skeletal muscle of type 2 diabetics, on the other hand, contains more subsarcolemmal lipid and this accumulation is inversely associated with insulin sensitivity (36
). Whether different compartmentalization of lipids in EA vs. AA could explain the ethnic differences we have observed in the relationship between IMCL and insulin sensitivity remains to be determined.
Another possible explanation is a difference in skeletal muscle fiber type between AA and EA. Insulin resistance has recently been related to a higher IMCL content in type I (oxidative) muscle fibers more so than in type II (glycolytic) muscle fibers (37
). Furthermore, AA women were found to have a higher percentage of type II muscle fibers than their EA counterparts. However, this difference was found to be related more to increased general adiposity in this group than to ethnic differences (38
). Our groups did not differ in adiposity and therefore should not be expected to have differences in muscle fiber type. Nevertheless, more studies analyzing histochemical properties of skeletal muscle in AA vs. EA are needed.
Yet another ethnic difference that may impact insulin sensitivity relates to substrate oxidation. Reduced fatty acid oxidation has been observed in obese AA women, compared to EA women, and is related to reduced insulin sensitivity (39
). Moreover, metabolic inflexibility in substrate use has been reported in healthy premenopausal AA women, compared to EA women (40
). Therefore, the differing relationship between IMCL and insulin sensitivity in EA and AA groups may be due to differences in substrate flux within skeletal muscle.
A limitation of this study is that the subjects were not assessed for aerobic capacity. Maximal aerobic capacity has been shown to be an important determinant of tibialis IMCL. While VO2max
was not significantly related to soleus IMCL, an interaction effect was observed between soleus IMCL, VO2max
, and GDR (17
data was not available on our subjects; however, all subjects were screened to be completely sedentary and involved in no regular physical activity or planned exercise. Therefore, it is unlikely that aerobic fitness could entirely explain the differences we have found in the relationships between soleus IMCL and insulin sensitivity in EA and AA.
Taken together, our results suggest that, in EA, IMCL is a fat depot that closely relates to insulin sensitivity as well as to generalized and central adiposity. In AA, however, central adiposity is more closely related to insulin resistance. These data indicate that IMCL is a determinant of insulin resistance in EA but exists largely independent of insulin resistance in AA. Clearly, skeletal muscle insulin resistance is less dependent upon IMCL accumulation in AA. In both AA and EA, however, central adiposity is associated with insulin resistance and confers increased risk of cardiometabolic disease. These differences in metabolic and body composition traits and their associations with insulin resistance point to potential racial differences in the pathogenesis of the metabolic syndrome. Specifically, IMCL may serve as a less relevant pathophysiological role in the development of insulin resistance and the metabolic syndrome trait cluster in individuals of African descent. Large-scale clinical trials that include analyses for aerobic capacity, VAT, and hepatic fat are needed to further assess the specific contributions of ectopic fat to insulin resistance in individuals of African descent.