In a multiethnic sample of relatively young men and women with stage I obesity and without diabetes, we found striking ethnic differences in the levels of pancreatic TGs and in the relationship between pancreatic TGs and β-cell dysfunction. Black subjects have very low levels of pancreatic TGs and robust compensatory insulin secretion; small increments in pancreatic TGs are accompanied by large linear increases in a compensatory insulin secretion. White subjects have much higher pancreatic TG levels and a flatter positive relationship between pancreatic TGs and compensatory insulin secretion. Hispanic subjects have levels of pancreatic TGs similar to those of whites but the relationship between pancreatic TGs and insulin secretion is paradoxically negative.
The lower level of pancreatic TGs in black compared with Hispanic adults adds to our previous work, showing lower levels of hepatic TGs in black than Hispanic individuals (8
), which also was confirmed in this study. Importantly, hepatic TGs cannot be used as a surrogate for pancreatic TGs; pancreatic TG levels were elevated similarly in Hispanic and white subjects, whereas hepatic TG levels were excessively high only in Hispanics. Analysis of our subject-specific data shows that pancreatic TGs cannot be inferred either from hepatic TGs or from serum TGs, thus necessitating direct organ-specific measurement by 1
It was reported previously that pancreatic TG levels as measured by 1
H MRS are elevated in patients with type 2 diabetes as well as in those with impaired glucose tolerance, impaired fasting glucose, or both (22
); however, no relation of pancreatic TGs to β-cell dysfunction was found for exclusively middle-aged or older white European subjects. The enrichment of our study sample with blacks and Hispanic young adults enhanced our ability to define these relationships. Our spectroscopic data confirm and extend a previous study of an adolescent population that indicated a lower level of pancreatic TGs in black teenagers than in Hispanic teenagers (25
). In that report, the race/ethnicity difference in pancreatic TG levels was small because of the use of a less sensitive method of fat imaging and was related to inflammatory markers and free fatty acids but not to β-cell function. In contrast, we found a striking relationship between pancreatic TG level and compensatory insulin secretion, which differed dramatically by race/ethnicity.
The steep relationship observed between low levels of pancreatic TGs and compensatory insulin secretion in black adults is consistent with glucose-free fatty acid cross-talk, a well-described phenomenon in which free fatty acids stimulate insulin secretion either directly or secondarily by promoting insulin resistance (26
). The weaker positive relationship seen in white adults and the inverse relationship seen in Hispanic adults is consistent with lipotoxicity offsetting, or even overriding, glucose-free fatty acid cross-talk. The stimulatory effect of free fatty acids on insulin secretion also has been shown to be attenuated in persons who are genetically predisposed to develop type 2 diabetes (26
A novel finding of our study is that Hispanic subjects were the only group with both pancreatic and hepatic steatosis, the latter being known to 1
) attenuate insulin-mediated suppression of gluconeogenesis, elevating fasting plasma glucose and basal insulin secretion; and 2
) augment hepatic TG secretion, exposing the pancreas to a higher TG load (29
). Thus, we speculate that hepatic steatosis begets pancreatic steatosis.
The marked racial/ethnic variation in pancreatic steatosis and its relationship to compensatory insulin secretion suggests a major influence of ancestral genes. Mutations related to ancestral genes previously have been implicated as a risk factor for hepatic steatosis in individuals of Hispanic ancestry and a protective factor in those of African ancestry (31
), with the gene products expressed in both hepatocytes and adipocytes. However, the genetic underpinning for racial/ethnic variation in pancreatic steatosis is unknown and will require further exploration.
Our study has several limitations. 1
H MRS cannot selectively measure TGs in β-cells; nevertheless, we previously validated whole pancreas TG level as a valid surrogate for β-cells (32
). Because our study is cross-sectional and correlational, the tight associations found between pancreatic steatosis and β-cell dysfunction are consistent with the lipotoxicity hypothesis but do not prove causal attribution. However, because we excluded patients with type 2 diabetes, the observed pancreatic steatosis constitutes a putative cause rather than the consequence of β-cell dysfunction. Moreover, our human data are consistent with extensive studies in the ZDF rat showing that excess pancreatic TGs are converted to toxic intermediates that cause β-cell apoptosis. Because black adults were resistant to pancreatic steatosis, other mechanisms must mediate the high incidence of type 2 diabetes in the black population. We do not know whether the pancreatic steatosis and associated β-cell dysfunction in white and Hispanic young adults are reversible. However, a prior study (33
) of patients with uncomplicated type 2 diabetes showed that β-cell dysfunction can be rescued by intensive caloric restriction, which also caused a detectable reduction in pancreatic steatosis. Because we studied the three major race/ethnic groups in the U.S., our findings cannot be generalized to other ethnic groups without further study.
This study focused on ethnic differences and did not address potential sex differences. On average, the sex distribution within the groups was balanced (P = 0.4744), although the absolute percentage of Hispanic men was higher compared with white and blacks.
Despite these limitations, the data herein implicate pancreatic TG level measured by 1H MRS as a noninvasive novel biomarker for pancreatic β-cell dysfunction, especially in the at-risk Hispanic population. Pancreatic TGs hold exciting promise as intermediate phenotypes in clinical intervention trials for obesity and as biomarkers to identify mildly obese patients who stand to benefit from early preemptive intervention.