Reduced Insulin Clearance and Insulin-Degrading Enzyme Activity Contribute to Hyperinsulinemia in African Americans

Abstract

Background

African Americans (AAs) are at a higher risk for developing type 2 diabetes compared with non-Hispanic whites (NHWs). The causal role of β-cell glucose sensitivity (β-GS) and insulin clearance in hyperinsulinemia in AA adults is unclear.

Objective

Using a cross-sectional study design, we compared β-cell function and insulin clearance in nondiabetic AAs (n = 36) and NHWs (n = 47) after a mixed meal test (MMT).

Methods

Insulin secretion rate, glucose sensitivity, rate sensitivity, and insulin sensitivity during MMT were derived from a mathematical model. Levels of insulin-degrading enzyme (IDE) and carcinoembryonic antigen-related cell adhesion molecule-1 (CEACAM1), key players in insulin clearance, were measured (by enzyme-linked immunosorbent assay) in hepatic cytosolic fractions from age-, sex-, and body mass index–matched AA and NHW cadaveric donors (n = 10).

Results

Fasting and mean postprandial plasma glucose levels were similar in both ethnic groups. AAs had significantly higher fasting and mean postprandial plasma insulin levels. However, fasting ISR, total insulin output, and insulin sensitivity during MMT were not different between the groups. β-GS and rate sensitivity were higher in AAs. Fasting and meal plasma insulin clearance were lower in AAs. Hepatic levels of IDE and CEACAM-1 were similar in AAs and NHWs. Hepatic IDE activity was significantly lower in AAs.

Conclusions

In this study, lower insulin clearance contributes to higher plasma insulin levels in AAs. Reduced insulin clearance may be explained by lower IDE activity levels in AAs. Further confirmatory studies are needed to investigate diminished insulin clearance in AAs as a result of lower IDE activity levels.

African Americans (AAs) are at a higher risk for the development of type 2 diabetes mellitus (T2DM) compared with non-Hispanic whites (NHWs) (1-3). Differences in diet, body composition, insulin sensitivity, and pancreatic β-cell function have been proposed to contribute to this increased risk (1, 4-6). Insulin sensitivity as measured by the intravenous glucose tolerance test (IVGTT) is reduced in nondiabetic AA adults (7, 8). Acute insulin response (AIR) following an IVGTT is frequently higher in AAs and is often suggested to be a compensatory response to lower insulin sensitivity (7, 8). However, in adolescents, for the same degree of insulin sensitivity, AAs have a higher AIR to intravenous glucose when compared with NHWs (9). Similarly, plasma insulin concentrations in the fasting state and following an oral glucose load is often higher in nondiabetic AA adults (10-14). Diminished insulin clearance may contribute partly to these observed differences in circulating insulin levels (12, 15-18). However, the role of β-cell glucose responsivity following an oral glucose load in adults is unclear.

Insulin clearance rate is an important determinant of and accounts for nearly 25% of the variability in plasma insulin concentrations (19). The liver and kidney are major sites for insulin clearance (20, 21). Following insulin binding, the insulin receptor is internalized, and receptor-bound insulin is degraded by insulin-degrading enzyme (IDE) (22). Although, cellular mechanisms regulating IDE are not clear, IDE polymorphisms are associated with higher risk for developing T2DM (23). In addition to IDE, carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1), a transmembrane protein plays an important role in hepatic insulin clearance (24). Activated insulin receptor, a tyrosine kinase, phosphorylates CEACAM1 to enhance the rate of internalization and degradation of insulin-bound insulin receptor complexes (25). Reduced insulin clearance contributes to hyperinsulinemia in AAs (12, 15-18), but the cellular basis for the lower insulin clearance is unknown. Specifically, there are no studies that have examined ethnic differences in the expression and activity of hepatic IDE and CEACAM1. The purposes of this study were to a) compare β-cell glucose sensitivity (β-GS) and insulin clearance following a meal in nondiabetic AAs and NHWs, and b) examine ethnic differences in the concentrations and activity of hepatic IDE and CEACAM1.

Methods

Study design and participants

Participants in this study were part of an ongoing cross-sectional study conducted at the Clinical Research Center, National Institutes of Health, in Bethesda, Maryland (ClinicalTrials.gov identifier: NCT00428987). The study protocol was approved by the institutional review board of the National Institute of Diabetes and Digestive and Kidney Diseases, and all procedures followed were in accordance with institutional guidelines. Written informed consent was obtained from all participants. Individuals older than 18 years with a body mass index (BMI) greater than 18.5 kg/m² and stable weight over the last 3 months were included in the study. A subset of individuals from the original cohort who were nondiabetic (based on medical history and hemoglobin A_1C < 6.5%) and not on any medications known to modulate insulin sensitivity, glucose metabolism, or weight loss were included in this study. Exclusion criteria were pregnancy, diabetes, liver disease, pulmonary disease, renal insufficiency, coronary heart disease, heart failure, peripheral vascular disease, coagulopathy, or other systemic disease. Three patients were on antihypertensives (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, and thiazide diuretics); 4 were on hypolipidemic agents (statins, n = 3 and niacin, n = 1); and 4 women were on oral contraceptives. All participants were admitted to the Metabolic Clinical Research Unit at the National Institutes of Health Clinical Center for 1 or more days.

Study procedures

Assessment of body composition.

Body weight was measured using a digital balance (Scale-Tronix 5702; Scale-Tronix). Body composition was measured by dual-energy X-ray absorptiometry with a Lunar iDXA scanner (GE Healthcare).

Measurement of β-cell function, insulin sensitivity, and glucagon-like peptide 1 following a mixed meal test.

Each participant underwent a mixed meal test (MMT) after a 12-hour overnight fast. Participants received a liquid meal with an approximate total nutrient content of 360 kcal (50% carbohydrates, 16% protein, and 34% fat). Samples for determination of plasma glucose, insulin, and C-peptide were drawn at –10, 0, 15, 30, 45, 60, 120, and 180 minutes. Insulin and C-peptide concentrations were measured in serum using the Roche Cobas 6000 analyzer (Roche Diagnostics). For analysis of glucagon-like peptide 1 (GLP-1), blood samples were collected in EDTA tubes with dipeptidyl peptidase IV (DPP-IV) and protease inhibitors, immediately centrifuged, and the plasma was stored at −80°C until analysis. In a subset of AAs (n = 22) and NHWs (n = 21) for whom plasma samples were available, intact GLP-1 concentrations were measured with Mesoscale enzyme-linked immunosorbent assay (ELISA) kits.

β-cell function parameters [(β-GS (in pmol.min⁻¹.m⁻². mM⁻¹), rate sensitivity (pmol.m⁻².mM⁻¹), and potentiation] were derived by mathematical modeling of plasma glucose and C-peptide concentrations during the MMT as previously described (26). β-GS is the slope of the relationship between insulin secretion rate and plasma glucose concentration. Rate sensitivity is the magnitude of the β-cell insulin response to the rate of change in plasma glucose concentration. The potentiation factor ratio (PFR) modulates the relationship between glucose concentration and insulin secretion by increasing the sensitivity of the β cells to resulting plasma glucose concentration. The PFR accounts for various potentiating factors such as incretins, neurotransmitters, and secretagogues. PFR was quantified using ratios between mean values at times 160 to 180 and 0 to 20 minutes. Area under the curve (AUC) was calculated with the use of the trapezoidal method. AUC of insulin secretion during the 3-hour MMT represents the total insulin output (expressed in nmol.m⁻²). The AUCs for glucose and insulin concentration were divided by the accumulation time of the MMT to provide a time-averaged value (mmol/L) or pmol/L, respectively. Assuming an absence of hepatic C-peptide extraction, basal and stimulated insulin clearance was calculated from the ratio of insulin secretion AUC/insulin AUC from temporal curves that were followed for 180 minutes, allowing plasma levels to return to baseline. A correction factor [V*(Insulin_{180 min} – Insulin_{0 min})/AUC_Insulin)] was applied to insulin clearance to account for differences in baseline by the end of the sampling period (T = 180).

Insulin sensitivity parameters.

A model-based index of insulin sensitivity (oral glucose insulin sensitivity, OGIS) was calculated using plasma glucose and insulin concentrations during the 3-hour MMT (27). The Quantitative Insulin-sensitivity Check Index (QUICKI) was calculated as previously described (28). Insulin sensitivity index, S_I, was determined from the minimal model using data from the insulin-modified FSIVGTT as previously described (version 6.02; MinMOD Millennium) (29). Briefly, after a 12-hour overnight fast, 2 intravenous catheters were placed, 1 in each arm near or in the antecubital veins. Baseline blood samples were obtained, and at time 0 minutes, dextrose (0.3 g/kg) was administered intravenously. Insulin was given as a bolus at 20 minutes (0.03 U/kg/min). Blood samples for glucose and insulin were obtained at –10, –1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16, 20, 22, 23, 24, 25, 27, 30, 40, 50, 60, 70, 80, 90, 100, 120, 150, and 180 minutes.

Hepatic carcinoembryonic antigen-related cell adhesion molecule-1 and insulin-degrading levels and activity.

IDE is mostly present in the cytosol (~95%) (30) and therefore we measured the levels and activity of IDE in hepatic cytosolic fractions. Protein concentrations of human hepatic cytosolic fractions (Xenotech) from age-, sex-, and BMI-matched AA and NHW cadaveric liver donors (n = 10) were measured using a BCA Protein Assay Reagent Kit (Thermo Scientific). The cause of death of liver donors was cerebral anoxia following head trauma or cerebrovascular accident (CVA). Human hepatic cytosolic fractions were from a commercial source, and additional information regarding comorbidities was unavailable. CEACAM1 levels in individual hepatic cytosolic fractions (n = 6) were measured by ELISA (RayBiotech Life). IDE levels were measured using a human IDE ELISA kit (Fivephoton Biochemicals; catalog No. IDE-ELISA [96T]). IDE activity in liver cytosolic fractions was measured for 30 minutes using the SensoLyte IDE Activity Assay Kit (AnaSpec; catalog No. AS-72231) according to the manufacturer’s instructions.

Statistical analysis

Variables with normal distributions are expressed as mean ± SD. Variables with a non-normal distribution are expressed as median (interquartile range). Comparisons between groups were assessed by the independent unpaired t test or The Wilcoxon-Mann-Whitney test. A P value less than .05 was considered statistically significant. Relationships between parameters were assessed by linear regression analyses using the Pearson correlation coefficient. Ethnic differences in indices of insulin action, β-cell function, and insulin clearance were examined using ANCOVA, with ethnicity as a fixed factor, and relevant covariates. Data were analyzed with JMP version 13.0 (SAS Institute) and GraphPad Prism 7 (GraphPad Software Inc).

Results

Participant characteristics and baseline metabolic profiles

Baseline characteristics and metabolic parameters of the study participants are depicted in Table 1. Fifty-six percent of AAs (n = 20) vs 30% of NHWs (n = 14) were obese (BMI ≥ 30 kg/m²) in this cohort (P = .02). Compared with NHWs, AAs had a significantly higher BMI, waist circumference, and fat-free mass, but a similar percentage of total body and truncal fat. There were a higher proportion of AAs with prediabetes (defined as hemoglobin A_1C, 5.7%-6.4%) compared with NHWs (42 vs 13%, P = .002). The percentage of individuals with impaired fasting glucose (defined as a fasting plasma glucose of 100-125 mg/dL) was not significantly different between the groups (22 vs 19%, P = .73). Plasma triglyceride levels were lower in AAs, and there were no ethnic differences in total, low-density lipoprotein, and high-density lipoprotein cholesterol levels. Fasting plasma insulin and hemoglobin A_1C levels were higher in AAs, whereas fasting glucose and C-peptide were not significantly different. AAs had a lower S_I and QUICKI, but OGIS was similar between the groups. Ethnic differences in S_I and QUICKI remained significantly different even after adjustment for BMI, percentage total body fat, and percentage truncal fat (data not shown). AIR was higher (~2.7-fold) in AAs (Table 1) and remained significant even after adjusting for BMI and S_I (478.1 vs 235.0 μU·ml^–1·min^–1 [ratio of geometric mean 2.0, 95% CI 1.5-2.8]; P < .001). Results were similar when total percentage body or truncal fat was used as a covariate instead of BMI (data not shown). These results suggest that the higher AIR is above the magnitude expected from a compensatory response to reduced insulin sensitivity in AA. Serum creatinine and estimated glomerular filtration rate (eGFR) were significantly higher in AAs.

Table 1.

Clinical and metabolic characteristics in African Americans and non-Hispanic whites

		NHW (n = 47)			AA (n = 36)		P
Age, y	39	±	13	39	±	11	.99
Female, %, n		68 (n = 32)			56 (n = 20)		.24
BMI, kg/m2	26.3 (9.7)			31.2 (12.3)			.02
Fat free mass, kg	49	±	11	57	±	11	.002
Total body fat, %	35	±	11	38	±	12	.24
Waist circumference, cm	86.2	±	16.4	95.8	±	16.8	.02
Truncal fat, %	35.0	±	13.3	39.9	±	12.4	.08
Systolic blood pressure, mm Hg	119	±	12	125	±	12	.02
Diastolic blood pressure, mm Hg	70	±	8	72	±	8	.40
Fasting plasma glucose, mmol/L	4.92	±	0.38	5.03	±	0.39	.49
Fasting plasma insulin, pmol/L	17.4 (49.8)			45 (55.8)			.007
Fasting C-peptide, nmol/	0.49 (0.53)			0.56 (0.42)			.16
Hemoglobin A1c, %	5.3 (0.5)			5.6 (0.8)			.001
QUICKI	0.389 (0.104)			0.353 (0.073)			.003
SI, min–1·μU·ml–1	5.47 (6.37)			2.93 (2.12)			< .001
OGIS, mL.min−1.m−2	360	±	41	347	±	34	.10
AIR, μU·mL–1·min–1		230 (210)			611(632)		< .001
Total cholesterol, mmol/L	4.98	±	0.73	4.67	±	0.70	.06
LDL cholesterol, mmol/L	2.89	±	0.79	2.83	±	0.64	.43
HDL cholesterol, mmol/L	1.50	±	0.43	1.36	±	0.37	.12
Triglycerides, mmol/L	0.95 (0.89)			0.73 (0.53)			.03
Serum creatinine, µmol/L	74.6 ± 14.7			82.8 ± 18.4			.03
eGFR, mL.min−1.1.73 m−2	93 ± 22			103 ± 20			.02

Data are presented as arithmetic mean ± SD or as median (interquartile range); n, No. of participants. P values indicate significance for comparisons between ethnic groups.

Abbreviations: AA, African American; AIR, acute insulin response; BMI, body mass index; eGFR, estimated glomerular filtration rate by the abbreviated Modification of Diet in Renal Disease equation; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NHW, non-Hispanic white; OGIS, oral glucose insulin sensitivity index; QUICKI, quantitative insulin-sensitivity check index; S_I, insulin sensitivity index derived from frequently sampled intravenous glucose tolerance test.

β-cell function and insulin clearance during a mixed meal test

Fasting plasma glucose concentrations as well was postprandial glucose concentrations (AUC) during an MMT were not significantly different between the groups (Fig. 1). Fasting and postprandial insulin levels, but not C-peptide levels, were higher in AAs (Figs. 1 and 2). Differences in mean postprandial insulin concentrations remained significant even after adjustment for mean glucose levels during the meal. Mean insulin was higher in AAs (by ~32%) when compared with NHWs (258.6 ± 136.9 vs 194.7 ± 146.1 pmol/L, P < .001) (Fig. 2D). Basal insulin secretory rate was similar between the groups (Fig. 3B), but early insulin secretion was higher in AAs (time*race interaction, P = .04; Fig. 3A). When total insulin production during the MMT was compared, there was no significant difference between the groups (Fig. 3C). β-GS (main effect of race, P = .004, Fig. 4A) and rate sensitivity (Fig. 4B) were significantly higher in AAs. PFR (Fig. 4C) and the potentiation factor time course were similar in the 2 groups (data not shown). Intact GLP-1 levels during MMT (Fig. 4D) and GLP-1 AUC were comparable between AAs and NHWs (34.6 ± 24.7 vs 25.9 ± 10.3 pM.180 minutes, P = .51). We also compared GLP-1 AUC during the first 45 minutes of the MMT and found no significant differences between the groups (data not shown). A mixed-effects model analysis revealed that intact GLP-1 levels for both groups increased over time (P < .001); the increase appears to be the same for both groups (P = .15), and there was no significant time*race interaction (P = .64) (Fig. 4D). Insulin clearance during fasting and a meal were significantly lower in AAs when compared with NHWs (Fig. 5A and 5B). In the entire group, meal insulin clearance was positively correlated with OGIS (r = 0.44, P < .001) and S_I (r = 0.63, P < .001); slopes of regressions showed no significant ethnic differences. Meal insulin clearance remained lower in AAs even after adjusting for S_I and BMI when compared with NHWs (1.09 vs 1.36 L.min^–1.m², adjusted ratio of geometric means 0.80, 95% CI [0.71-0.89, P = .001]). Results were similar when total percentage body or truncal fat was used as a covariate instead of BMI (data not shown). Higher eGFR may lead to rapid insulin clearance (peripheral component). However, despite the higher eGFR, AA had a lower insulin clearance. In the entire cohort, eGFR was not related to fasting (r = –0.16, P = .13) or meal insulin clearance (r = –0.17, P = .10). There were no ethnic differences in these relationships. Finally, there was no statistically significant interaction between race and prediabetes status on insulin clearance (during fasting and during meal), rate sensitivity, S_I, AIR, and mean meal insulin concentration.

Figure 1.

Plasma levels of A, glucose, B, insulin, and C, C-peptide, in adult African Americans (AA) and non-Hispanic whites (NHW) during a mixed meal test (MMT). Data shown are mean ± SEM. P values for the effects of racial group and time and the race*time interaction were obtained with repeated-measures ANOVA with post hoc Bonferroni test.

Figure 2.

Plasma levels of A, fasting glucose, B, fasting insulin, C, mean postprandial glucose, and D, mean postprandial insulin in adult African Americans (AA) and non-Hispanic whites (NHW) during a mixed meal test (MMT). Data shown are mean ± SEM. Comparisons between groups were assessed by independent unpaired t test or the Wilcoxon-Mann-Whitney test.

Figure 3.

Model-derived A, insulin secretory rate, B, fasting insulin secretory rate, and C, total insulin output, in adult African Americans (AA) and non-Hispanic Whites (NHW) during a mixed meal test (MMT). Data shown are mean ± SEM. In panel A, P values for the effects of racial group and time and the race*time interaction were obtained with repeated-measures ANOVA with post hoc Bonferroni test. Comparisons between groups (in panels B and C) were assessed by independent unpaired t test or the Wilcoxon-Mann-Whitney test.

Figure 4.

Relationship between insulin secretory rate and A, plasma glucose levels, B, rate sensitivity, C, potentiation factor ratio, and intact D, glucagon-like peptide 1 concentrations in adult African Americans (AA) and non-Hispanic whites (NHW) during a mixed meal test (MMT). Data shown are mean ± SEM. In panel A, P values for the effects of racial group and time and the race*time interaction were obtained with repeated-measures ANOVA with post-hoc Bonferroni test. Comparisons between groups in panel B were assessed by independent unpaired t test or the Wilcoxon-Mann-Whitney test.

Figure 5.

A, Fasting insulin clearance and B, meal insulin clearance in adult African Americans (AA) and non-Hispanic whites (NHW) during a mixed meal test (MMT). C, Insulin degrading enzyme levels and activity were measured in hepatic cytosolic fractions from age-, sex-, and body mass index–matched AA and NHW cadaveric donors (n = 10). Data shown are mean ± SEM. Comparisons between groups were assessed by independent unpaired t test or the Wilcoxon-Mann-Whitney test.

Ethnic differences in hepatic carcinoembryonic antigen-related cell adhesion molecule-1 levels and insulin-degrading enzyme levels and activity

Age (AAs vs NHWs: 40.3 ± 10.8 vs 37.8 ± 6.4 y) and BMI (31.5 ± 10.9 vs 31.3 ± 6.2 kg/m²) were not significantly different between the groups (P > .05). IDE concentrations in hepatic cytosol fractions comparable between the groups (Fig. 5C). However, hepatic IDE activity in AA was 29% lower when compared with NHW (Fig. 5D). Hepatic CEACAM1 levels were not different between the 2 groups (AAs vs NHWs: 1.04 ± 0.23 vs 1.17 ± 0.48 ng/µg).

Discussion

In the present study, we found that early insulin secretion, β-GS, and rate sensitivity following a mixed meal was higher, but the cumulative insulin secretion was not different in nondiabetic AAs. Fasting and prandial insulin clearance were lower in AAs. This suggests that the higher postprandial insulin concentration is a result of reduced insulin clearance and not due to higher total insulin secretion. Consistent with this finding, we demonstrate for the first time that hepatic IDE activity, but not levels, was lower in AAs when compared with NHWs.

Ethnic differences in β-cell glucose sensitivity

Increased AIR is a characteristic trait in AAs and may be the result of lower S_I, reduced insulin clearance, and enhanced β-cell function. Dynamic responsivity, a measure of the capacity of β cells to respond to the rate of increase of glucose concentration, can be assessed following an intravenous or oral glucose load and represented by Φ ₁ and Φ _D, respectively (31). Similarly, static responsivity, a parameter that reflects the ability of β cells to deliver new pools of insulin granules in response to continued increases in glucose concentration, can also be derived from these tests and is represented by Φ ₂ and Φ _S, respectively (31). Only a couple of studies have examined differences in β-cell glucose responsivity in AA and NHW women (14, 32). In healthy, normoglycemic AA adult women, first-phase glucose responsivity was significantly higher when compared with their white counterparts during an IVGTT (32). Similarly, Φ _D and dynamic responsivity following a mixed meal were higher in AA women (14). However, there were no ethnic differences in static responsivity in both studies (14, 32). Higher rate sensitivity observed in our study is consistent with these studies, suggesting that β-cell function is enhanced in AAs when compared with NHWs. Whether this feature is intrinsic to the β cell in AAs is unknown and an important area of investigation. Surprisingly, there are no studies examining the racial differences in dynamic insulin secretion in human islet preparations in vitro. In one exploratory study, the effect of race was examined on the pattern of insulin secretion in isolated islet preparations using a perifusion system (33). Basal, glucose-stimulated, and 3′,5′-cyclic AMP (cAMP)-mediated insulin secretion (in the presence of cAMP phosphodiesterase, IBMX) were examined. Islet preparations from AAs had a smaller magnitude of glucose and cAMP-induced insulin secretion (fold-change over baseline) than NHW preparations. However, basal insulin secretion was higher in AAs. In addition to replicating this finding, future studies addressing the cellular basis for this difference are warranted.

Incretin hormones, GLP-1, and glucose-dependent insulinotropic polypeptide (GIP), are known to increase β-GS in humans (34, 35). Prior studies in obese adults have reported higher levels of insulin and intact GLP-1 during an OGTT in AAs (36). But, there were no ethnic differences in the levels of incretins during an MMT in AAs when compared with NHW women (14). In agreement with the latter study, we did not observe any ethnic differences in MMT-induced increases in circulating intact GLP-1 levels (Fig. 4D). This finding was consistent with a lack of race-related differences in PF, a parameter that reflects cumulative effects of changes in incretins, neuronal input to β cells, and glucose concentration (Fig. 4C).

Ethnic differences in insulin clearance

Circulating insulin concentration is a net result of insulin secretion and clearance. In our study, despite similar insulin secretory rates and total insulin secretion during a meal between the groups, AAs had higher levels of postprandial insulin. The relative hyperinsulinemia can be the result of reduced hepatic and/or peripheral insulin clearance. The findings of reduced insulin clearance during fasting and a meal in our study are consistent with other studies in AA adults (12, 14, 18, 32). Reduced insulin clearance may be a compensatory response to reduced insulin sensitivity (37). In our study, OGIS, a surrogate measure of insulin sensitivity, was comparable between the groups. Consistent with other studies, the more widely used measure of insulin sensitivity derived from FSIVGTT (8), S_I was significantly lower in AAs. OGIS is insensitive to detect lower insulin sensitivity in AAs in this cohort because the main differences in insulin and glucose between the groups are in the early part of the meal (Fig. 1) as observed in other studies (14, 36). OGIS was also similar in black and white adolescents (38). QUICKI is a fasting surrogate index of insulin sensitivity based on fasting glucose and insulin levels. Reduced insulin clearance may affect reliability of these indices, especially in AAs (13). Conversely, it is possible that S_I assessed by the minimal model may underestimate insulin sensitivity in AAs who manifest higher peak of insulin concentrations (AIR) compared with NHWs (8). Nevertheless, even after adjusting for the differences in S_I between the groups, meal insulin clearance was significantly lower in AAs. Ethnic difference in insulin clearance, a highly heritable trait (39, 40), may explain hyperinsulinemia frequently observed in AAs.

Ethnic differences in hepatic and peripheral insulin clearance

Insulin secreted into the portal circulation is extracted by the liver as it undergoes first pass before delivery to the systemic circulation. In the periphery, insulin is cleared by skeletal muscle and the kidney. Hepatic IDE and CEACAM1 play an important role in hepatic insulin clearance (22, 24). Insulin is secreted at an equimolar ratio with C-peptide. Because C-peptide is not extracted in the liver, the ratio of insulin secretion to insulin concentration is a measure of endogenous insulin clearance. Recently, using a mathematical model based on serum insulin and C-peptide during an insulin-modified frequently sampled IVGTT, Polidori et al were able to estimate hepatic and peripheral insulin clearance (41). This group subsequently evaluated hepatic and peripheral insulin clearance in 29 NHW and 18 AA healthy, nondiabetic women (15). AAs had a significantly lower (by ~60%) hepatic insulin clearance compared with NHWs. However, peripheral insulin clearance was not different between the groups. These results suggest that lower insulin clearance of insulin in AAs is primarily driven by the reduced extraction of insulin in the liver. We did not measure peripheral insulin clearance in this study. Nevertheless, insulin clearance was not associated with eGFR in AAs and NHWs.

The relationship between hepatic CEACAM1 levels and insulin clearance in humans has not been assessed. In a small study (n = 7), CEACAM1 levels were measured in liver samples in obese, 45- to 50-year-old men and compared with age-, sex-, and race-matched lean individuals (n = 4). Hepatic expression of CEACAM1 was lower in the obese individuals (by 50%) (42). Similarly, IDE expression was lower in patients with T2DM when compared with healthy individuals (43). This finding was supported by a liver microarray data set from the GEO Profiles database (44). Unfortunately, there was no ethnic information provided in any of these studies. To understand the underlying cellular mechanisms mediating the lower endogenous insulin clearance in AAs, we measured the hepatic levels and activity of IDE and CEACAM-1 in age-, sex-, and BMI-matched AA and NHW cadaveric liver donors. The levels of CEACAM-1 and IDE were not significantly different between the groups. However, IDE activity was lower by approximately 30% in AAs, suggesting that differences in posttranslational regulation of the catalytic activity of IDE may contribute to the observed ethnic difference. IDE activity is known to be regulated by cytoplasmic concentrations of calcium, adenosine triphosphate, oxidative stress, fatty acids, and acyl-CoA (45). Consistent with this, hepatic expression of IDE but not CEACAM-1 was lower in dogs that were fed a high-fat diet for 6 weeks (46). Causal factors mediating lower IDE activity in AAs are unknown and remain to be investigated. Nevertheless, this is the first study to report ethnic differences in IDE levels/activity.

Strengths and limitations of the study

There are many strengths of this study. First, using mathematical modeling approaches, we have rigorously examined the contributions of β-cell function and insulin clearance to hyperinsulinemia frequently observed in AA adults. Second, we used MMT, a more physiological test to elicit insulin secretion that accounts for the incretin effect in contrast to the study using intravenous glucose (15). Some, but not all studies, suggest that incretin hormones (GLP-1 and GIP) are known to modulate hepatic insulin clearance in humans (47-52). We did not observe significant ethnic differences in plasma-intact GLP-1 levels during the meal, but a larger cohort that is sufficiently powered may uncover ethnic differences in GLP-1 secretion, especially earlier in the MMT. Third, our study included men and women, unlike other studies that comprised just women (14, 15). Lastly, this is the first to study ethnic differences in IDE/CEACAM-1 levels in human hepatic tissues in AAs and NHWs. Despite these strengths, there are many limitations. We did not directly measure hepatic and peripheral insulin clearance in this study. Assessment of IDE activity in liver biopsies concomitantly performed with studies measuring insulin secretion/clearance would be ideal. Nevertheless, the risk-benefit ratio of liver biopsies or catheterization of hepatic or renal veins in healthy nondiabetic individuals would preclude such studies. It is possible that our study is underpowered to detect differences (<50%) in hepatic CEACAM-1 levels. Higher liver fat content is associated with impaired insulin clearance in nondiabetic individuals and in patients with nonalcoholic fatty liver disease (NAFLD) (53, 54). A recent study did not demonstrate significant ethnic differences in the relationship between hepatic fat and insulin clearance in black and white men (55). Furthermore, the negative relationship between hepatic fat and insulin clearance was observed in white but not in black women (14). It is generally assumed that visceral fat content is correlated with hepatic fat. In AA adults, this is not always the case as reported by Hakim and colleagues (55). We did not measure hepatic fat in our study and absence of NAFLD was elicited by medical history. Therefore, we were unable to examine relationships between insulin clearance, intrahepatic lipid content, and visceral fat.

In conclusion, in this study of nondiabetic men and women, reduced insulin clearance appears to contribute to hyperinsulinemia in AAs when compared with NHWs. To the best of our knowledge, this study is the first to explore cellular mechanisms for the reduced hepatic extraction of insulin observed in AAs. Lower hepatic IDE activity, but not differences in IDE and CEACAM-1 levels, may explicate the ethnic differences in insulin clearance. Reduced insulin clearance could be a proximal cause for hyperinsulinemia, insulin resistance, and T2DM in ethnic groups such as AAs. Future studies examining the mechanisms for lower hepatic IDE activity and higher β-GS are warranted.

Glossary

Abbreviations

β-GS

β-cell glucose sensitivity

AAs

African Americans

AIR

acute insulin response

AUC

area under the curve

BMI

body mass index

CEACAM1

carcinoembryonic antigen-related cell adhesion molecule 1

GIP

glucose-dependent insulinotropic polypeptide

IDE

insulin-degrading enzyme

IVGTT

intravenous glucose tolerance test

MMT

mixed meal test

NHWs

non-Hispanic whites

NAFLD

nonalcoholic fatty liver disease

OGIS

oral glucose insulin sensitivity

PFR

potentiation factor ratio

S_I

insulin sensitivity index

T2DM

type 2 diabetes mellitus

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.