Abstract
Context
High density lipoprotein (HDL) in humans is composed of a heterogeneous group of particles varying in protein composition as well as biological effects.
Objective
We investigated the prospective associations between HDL subspecies containing and lacking apolipoprotein (apo) C-III at baseline and insulin sensitivity at year 3.
Design, Setting, and Participants
A prospective cohort study of 864 healthy volunteers drawn from the relationship between insulin sensitivity and cardiovascular disease (RISC) study, a multicenter European clinical investigation, whose recruitment initiated in 2002, with a follow-up of 3 years.
Main Measures
Insulin sensitivity was estimated from an oral glucose tolerance test at baseline and year 3, and by euglycemic-hyperinsulinemic clamp at baseline only. The apolipoprotein concentrations were measured at baseline by a sandwich enzyme-linked immunosorbent assay (ELISA)-based method.
Results
The 2 HDL subspecies demonstrated significantly opposite associations with insulin sensitivity at year 3 (P-heterogeneity = 0.004). The highest quintile of HDL containing apoC-III was associated with a 1.2% reduction in insulin sensitivity (P-trend = 0.02), while the highest quintile of HDL lacking apoC-III was associated with a 1.3% increase (P-trend = 0.01), compared to the lowest quintile. No significant association was observed for total HDL, and very low density lipoprotein (VLDL) and low density lipoprotein (LDL) containing apoC-III. ApoC-III contained in HDL was associated with a decrease in insulin sensitivity even more strongly than plasma total apoC-III.
Conclusion
Both HDL containing apoC-III and apoC-III in HDL adversely affect the beneficial properties of HDL on insulin response to glucose. Our results support the potential of HDL-associated apoC-III as a promising target for diabetes prevention and treatment.
Keywords: HDL, apolipoprotein C-III, insulin sensitivity, diabetes, cohort study
High density lipoprotein (HDL) is a heterogeneous group of lipoprotein particles. We previously found that HDL existed in subspecies based on the presence of a small protein called apolipoprotein (apo) C-III. In healthy adults, apoC-III was present on approximately 6% to 7% of HDL, as measured by plasma total apoA-I, the main HDL apolipoprotein, whereas the remaining HDL was free of apoC-III (1). ApoC-III provokes inflammatory and atherogenic responses in monocytes and endothelial cells (2, 3), and it impairs lipoprotein metabolism (4, 5). ApoC-III-based HDL subspecies differ in their metabolic properties and downstream biologic interactions, and are differentially related to cardiovascular disease risk; HDL containing apoC-III has been associated with a higher prevalence of obesity, increased carotid intima-media thickness, impaired reverse cholesterol transport through HDL, and an elevated risk of coronary heart disease, opposite to what has been observed for total HDL or HDL lacking apoC-III (1, 6–9).
ApoC-III-based HDL subspecies might similarly relate to the development of insulin resistance and diabetes. Studies have suggested that HDL and apoA-I exert direct beneficial effects on the regulation of glucose metabolism through multiple mechanisms (10, 11). We speculate that these antidiabetic effects of HDL are attributable only to HDL lacking apoC-III, and if apoC-III is present on HDL, apoC-III may diminish the otherwise beneficial effect of HDL on glycemic control. In cross-sectional analyses of 2 prospective cohorts, the Nurses’ Health Study and the Health Professionals Follow-Up Study, concentrations of hemoglobin A1c (HbA1c) and prevalent diabetes were associated with higher levels of HDL containing apoC-III and lower levels of HDL lacking apoC-III (12). Furthermore, we found that only HDL lacking apoC-III was associated with a lower risk of diabetes in 2 prospective cohorts, the Multi-Ethnic Study of Atherosclerosis (13) and the Danish Diet, Cancer, and Health Study (14).
Indeed, apoC-III has a role in the pathogenesis of insulin resistance and diabetes. Transgenic mice with human apoC-III overexpression developed hepatic triglyceride accumulation and diet-induced insulin resistance (15). Mechanistic studies in cells and animal models have demonstrated that apoC-III increased cytokine expression in cultured monocytes and endothelial cells, thereby increasing monocyte adhesion to endothelial cells (2, 3) and pancreatic β-cell apoptosis via hyper activation of β-cell CaV channels (16, 17). These mechanisms link apoC-III to inflammation, pancreatic β-cell function, and insulin resistance. Elevated plasma apoC-III concentration is observed in both type 1 and 2 diabetes (17, 18). Cross-sectional studies (4, 19, 20) and prospective cohort studies (13, 14, 21) have shown that high levels of apoC-III were strongly associated with a risk of diabetes. Further, a group of New York Ashkenazi Jews having genetically reduced plasma apoC-III concentrations maintained greater insulin sensitivity with age and reached exceptional longevity (22).
Therefore, we hypothesized that only HDL lacking apoC-III would be beneficially associated with insulin sensitivity, whereas HDL containing apoC-III would be associated with impaired insulin sensitivity. We also studied the concentration of total apoC-III in whole plasma as well as the concentration of apoC-III itself contained in HDL. In addition, to complete the investigation of lipoproteins carrying apoC-III, we examined apoB-containing lipoproteins, low density lipoprotein (LDL), and very low density lipoprotein (VLDL), containing and lacking apoC-III.
Materials and Methods
Study design and participants
The Relationship between Insulin Sensitivity and Cardiovascular disease (RISC) study is a European multicenter, prospective cohort study with a standardized protocol and a centralized evaluation of measures. Details have been described elsewhere (23). Clinically healthy men and women ages 30 to 60 (n = 1513) were recruited between June 2002 and July 2004 from the local population at 19 centers in 14 European countries. Follow-up examinations were performed after 3 years. Exclusion criteria were treatment for any chronic disease, pregnancy, any cardiovascular event, weight change of ≥5 kg in the last 3 months, cancer in the last 5 years, renal or liver failure, recent major surgery, arterial blood pressure ≥140/90 mmHg, fasting plasma glucose ≥7.0 mmol/L (126 mg/dl), 2-hour plasma glucose (on a standard 75-g oral glucose tolerance test [OGTT]) ≥11.1 mmol/L, total serum cholesterol ≥7.8 mmol/L (300 mg/dl), serum triglycerides ≥4.6 mmol/L, or electrocardiogram abnormalities. Out of 1017 eligible subjects, baseline plasma samples of 987 people were available with sufficient plasma for our analysis and were sent to our laboratory. Written informed consent was obtained from each participant. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. Each recruiting center obtained approval from the local ethics committee.
At the baseline exam, basic characteristics of the participants, including demographics and lifestyle, were collected by interviews and questionnaires. Information on physical activity was collected using the 7-day International Physical Activity Questionnaire (IPAQ), a validated assessment tool for international studies (24). Anthropometric measurements were obtained by trained personnel at baseline (23).
Apolipoprotein measurements
To measure the concentrations of apoA-I in HDL containing and lacking apoC-III, apoB in LDL and VLDL containing and lacking apoC-III, total apoC-III, apoC-III in HDL, and apoC-III in LDL and VLDL, a validated sandwich enzyme-linked immunosorbent assay (ELISA)-based method was used (25).
Plasma samples were removed from -80°C storage and thawed at room temperature. Whole plasma was fractionated into lipoproteins containing apoC-III and those deficient of apoC-III by immunoaffinity separation on a 96-well microplate (MICROLON 600, VWR Cat #82050-734; Greiner Bio-One, Kremsmünster, Austria) coated with rabbit anti-human apoC-III antibody (10 μg/mL in 1 x phosphate-buffered saline (PBS), Cat #33A-R1b; Academy Biomedical Company, Houston, TX, USA (26)). Following overnight incubation at 4°C, the unbound fraction depleted of apoC-III-containing lipoproteins was collected for analysis. After washing with 1 x PBS, ELISA diluent (1 x PBS/2% bovine serum albumin (BSA)/0.05% Tween 20) was added to each well to dissociate the components of the bound apoC-III-containing lipoproteins from the plate during a 2-hour incubation at 37°C and was collected for analysis. HDL was isolated from whole plasma by precipitation of apoB-containing lipoproteins with dextran sulphate and magnesium chloride.
The concentrations of apolipoproteins were measured by sandwich ELISA using polyclonal antibodies (Academy Biomedical Company): apoC-III in whole plasma and HDL (coating antibody Cat# 33A-R1b (26) at 10 μg/mL in 1 x PBS, detection antibody Cat# 33H-G2b (27) at 1 μg/mL in 1 x PBS); apoA-I in whole plasma, apoA-I in HDL containing apoC-III, and apoA-I in HDL lacking apoC-III (coating antibody Cat #11A-G2b (28) at 5 μg/mL in 1 x PBS, detection antibody Cat #11H-G1b (29) at 1 μg/mL in 1 x PBS); and apoB lipoproteins in LDL and VLDL containing apoC-III and apoB lipoproteins in LDL and VLDL lacking apoC-III (coating antibody Cat# 20A-G1b (30) at 5 μg/mL in 1 x PBS, detection antibody Cat# 20H-G1b (31) at 1 μg/mL in 1 x PBS). For all ELISAs, the coating antibody was incubated for 1 hour at 37°C. Plates were washed 3 times with washing buffer (0.1% Tween 20 in 1 x PBS) then blocked (Pierce, Casein in 1 x PBS at 1% w/v, VWR Cat #PI37528) with 1-hour incubation at 37°C, followed by washing 3 times with washing buffer. Three extensively characterized plasma pools whose apoC-III, apoA-I, and apoB concentrations in whole plasma and the apoC-III plasma fractions had been calibrated using commercially available standards and immunoaffinity, column chromatography were used—1 for the calibration curve and the other 2 as known-concentration controls to assess batch validity and between-batch variance. The calibration curve was prepared in dilutions starting at 10 000x and serially 2x further to 640 000x in 1 x PBS containing 0.5% BSA, creating a reliable second-degree polynomial curve fit. The calibration curve, known controls, and unknown samples were diluted in 1 x PBS containing 0.5% BSA and incubated overnight at 4°C. Following incubation, plates were washed 3 times with wash buffer, appropriate detection antibody was added as described, and plates were incubated for 1 hour at 37°C, and then washed 3 times with wash buffer. For detection, antibodies conjugated to biotin, avidin peroxidase was added (Cat #A7419-2ML, 0.01 μg/ml in 1 x PBS; Sigma Aldrich, St. Louis, MO, USA), and plates were incubated for 1 hour at 37°C and then washed 3 times with wash buffer. Finally, o-phenylenediamine (OPD; Cat #P9187-50SET; Sigma Aldrich) was added to all plates to develop color for 1 hour and 20 minutes at room temperature and the absorbance was read at 450 nm.
Each sample was measured in triplicate. Whenever the within-plate coefficient of variation (CV) for the triplicates exceeded 15%, the sample was rerun. Between-batch CVs were 8.6% for total apoC-III, 13.6% for apoA-I in HDL lacking apoC-III, and 8.2% for apoA-I in HDL containing apoC-III. In our method validation study, we found that the CVs of the sandwich ELISA-based method were lower than those of the gold standard method for accuracy of immunoaffinity column chromatography (29.3% for total apoC-III, 15.7% for apoA-I in HDL lacking apoC-III, and 19.9% for apoA-I in HDL containing apoC-III).
It should be noted that apoC-III in HDL is different from apoA-I in HDL containing apoC-III; the former represents the concentration of apoC-III itself contained in HDL, whereas the latter represents the concentration of apoA-I in HDL containing any amount of apoC-III. Because apoA-I is the principal protein in HDL, and is present in the large majority of HDL, the apoA-I concentration represents the concentration of HDL particles. The apoC-III-to-apoA-I molar ratio in apoC-III-containing HDL, which may be interpreted as the density of apoC-III on apoC-III-containing HDL, was calculated as the number of apoC-III molecules per apoA-I molecule on HDL particles containing apoC-III.
Apolipoprotein measurements were corrected for batch variability using methods developed by Rosner et al (32). Age, sex, center, alcohol, smoking, physical activity, and body mass index (BMI) were included in the regression for correction and were chosen based on their associations with the exposure and the outcome.
Outcome assessment
A standard 75-g OGTT was performed in the morning after an overnight fast at baseline and at 3-year follow-up in this study. Blood samples were taken at 0, 30, 60, 90, and 120 minutes after glucose ingestion. The plasma concentrations of glucose and insulin were measured at each timepoint, and insulin sensitivity was quantified using the oral glucose insulin sensitivity (OGIS) index. OGIS is calculated from OGTT measurements and is a validated estimate of the glucose clearance during a euglycemic-hyperinsulinemic clamp (33). The calculation is based on an equation derived from a regression model of the glucose-insulin relationship. In this paper, OGIS will be referred to as insulin sensitivity measured by OGTT (IS-OGTT).
On a separate day within 1 month of the baseline OGTT, a euglycemic-hyperinsulinemic clamp was performed. The clamp test was not repeated at year 3. Insulin was administered as a primed-continuous infusion at a rate of 240 pmol/min/m2 simultaneously with a variable 20% dextrose infusion adjusted every 5 to 10 minutes to maintain plasma glucose level within 0.8 mmol/L (±15%) of the target glucose level (4.5–5.5 mmol/L). Additional blood samples were obtained at 20-minute intervals for insulin determination. For example, the insulin concentrations (mean ± standard deviation [SD]) at 80 and 120 minutes at the end of the clamp were 421 ± 124 and 414 ± 117 pmol/l, respectively. Additionally, endogenous glucose production was measured on a subset of 400 participants. During the last 20 minutes of the basal tracer equilibration period, both plasma glucose concentration and basal (6,6-2H2) glucose tracer-to-tracee ratio were stable in all subjects. Following an overnight fasting, steady state conditions exist and basal endogenous glucose production (EGP_b) equals total glucose rate of appearance. The mean ± SD of EGP_b was 16.3 ± 4.8 μmol.min-1.lbm(kg)-1. During the last 40 minutes of the clamp, endogenous glucose production during clamp (EGP_ss) was calculated by subtracting the mean exogenous glucose infusion rate from the total glucose rate of appearance. The mean ± SD of EGP_ss was 4.19 ± 6.24 μmol.min-1.lbm(kg)-1. Insulin sensitivity was calculated as the ratio of the glucose infusion rate (M value) durnig the clamp metabolized per unit of plasma insulin, standardized by fat-free mass (M-I ratio, in units of mol/min per kgffm/mM). In this paper, the M-I ratio will be referred to as insulin sensitivity measured by clamp (“IS-clamp”). To ensure safety and consistent quality of measurements, data from each clamp study was sent by fax following clamp and assessed against the quality control criteria on receipt in the project office, to which feedback was provided.
At the end of the clamp, a glucose bolus (0.3 g per kilogram body weight) was administered intravenously over 1 minute; plasma glucose and C-peptide concentrations were measured 2, 4, 6, and 8 minutes after the injection. Glucose-induced secretory response (GISR), the acute insulin secretory response to the intravenous glucose bolus, was expressed as the ratio of incremental insulin secretion (calculated by C-peptide deconvolution) to the plasma glucose increment during the same time interval (34).
Statistical analysis
Baseline characteristics of participants were described in men and women separately using medians for continuous variables and percentages for categorical variables. Pearson correlation coefficients were calculated between HDL subspecies and other measured variables.
The top and bottom 1% of the apolipoprotein measurements were trimmed to minimize the influence of outliers (n = 20 removed). Observations that were missing any outcome measurements (baseline or year 3 IS-OGIS, or baseline IS-clamp) were deleted (n = 97 removed). We used the extreme studentized deviate procedure to simultaneously identify multiple outliers in a data set that follows an approximately normal distribution using the number of SDs that an observation is from the mean, to exclude outliers (n = 3 removed) (35, 36). We also excluded observations with residuals larger than 4 SDs (n = 3 removed). The final sample size was 864. For GISR, the sample size was 643, since it was measured on a subset of people.
Linear regression was performed to evaluate the relationships between apolipoprotein variables and outcomes. All outcome variables except GISR were log-transformed to improve normality. Sex-specific quintiles and quintile medians were created for each apolipoprotein variable, as the baseline apolipoprotein levels were different between men and women. Tests for trend were based on quintile medians. Nonlinearity was assessed with likelihood-ratio tests, comparing a model treating quintile medians continuously to a model with quintile indicators. If there was no deviation from linearity (P for nonlinearity >0.05), we also created sex-specific z-scores to analyze associations per SD. For log-transformed outcomes, we back-transformed the coefficients to present results as the percent (%) difference in insulin sensitivity outcome per SD higher or for each quintile compared with the lowest quintile. Potential confounders were selected based on our a priori knowledge as to the associations with the apolipoproteins as well as their effect on insulin sensitivity. Multivariable models were adjusted for age, sex, center, smoking (never; past; current smoker 1–14.9, or ≥15 tobacco grams per day [1 cigarette contains on average one gram of tobacco. ie, approximately tobacco g/day = cigarettes/day]), alcohol consumption (women: 0, 0.1–4.9, 5.0–14.9, or ≥15.0 g/day; men: 0, 0.1–4.9, 5.0–29.9, or ≥30.0 g/day), physical activity (inactive, moderate, or active), and BMI (continuous).
For the main prospective analysis with IS-OGTT at year 3 as the outcome, log-transformed IS-OGTT at baseline was additionally adjusted as a covariate in the model. In this way, the model analyzes IS-OGTT at year 3, adjusting for IS-OGTT at baseline. Because HDL containing and lacking apoC-III represent fractions of total HDL, the 2 HDL subspecies were entered in the model simultaneously. Likewise, both LDL and VLDL containing and lacking apoC-III were included in the models for apoB lipoproteins. Likelihood ratio tests were performed based on linear trends to assess heterogeneity between slopes of the 2 HDL subspecies. A model for apoC-III in HDL and a model for apoC-III in LDL and VLDL were adjusted for total HDL, and total LDL and VLDL, respectively. A sensitivity analysis was conducted to assess the influence of trimming the top and bottom 1% of apolipoprotein measurements. Finally, multivariable models for HDL subspecies were additionally adjusted for triglycerides, C-reactive protein (CRP), and interleukin 6 (IL-6), which were considered mediators between apoC-III and insulin response. To assess whether the associations between HDL subspecies and insulin sensitivity were modified by risk factors for insulin resistance, we first performed analysis stratified by sex, BMI, triglyceride (TG), CRP, IL-6, and HDL-cholesterol. Medians were used as cutoffs. We then created interaction terms to test for effect modification by these factors.
Finally, we examined how the apoC-III-based HDL subspecies were associated with GISR, the ratio of the incremental insulin secretion to the plasma glucose during the same time interval (pmol / [m2·min·mmol/l]) (34).
Results
Study participant characteristics
The median age of study participants was 43 for men and 45 for women. The median BMI was 25.8 kg/m2 for men and 24.1 kg/m2 for women (Table 1). In general, participants had healthy lipid profiles, with high HDL cholesterol (median for men: 1.2 mmol/l; median for women: 1.5 mmol/l), low LDL cholesterol (median for men: 3.0 mmol/l; median for women: 2.8 mmol/l), and low triglycerides (median for men: 1.1 mmol/l; median for women: 0.8 mmol/l) (Table 2). During the 3-year follow-up, only 3 participants developed diabetes.
Table 1.
Baseline (2002–2004) demographics of men and women with apolipoprotein measurements in the RISC study
| Men (N = 389) | Women (N = 475) | |||
|---|---|---|---|---|
| Median | 25th, 75th percentile | Median | 25th, 75th percentile | |
| Age, yrs | 43 | (37, 51) | 45 | (39, 51) |
| BMI, kg/m2 | 25.8 | (23.8, 27.9) | 24.1 | (22.0, 26.7) |
| Systolic blood pressure, mmHg | 123 | (116, 130) | 115 | (105, 123) |
| Diastolic blood pressure, mmHg | 77 | (71, 81) | 73 | (68, 78) |
| Current alcohol consumption ≥1 drink/day, N (%) | 162 (42) | 69 (15) | ||
| IPAQ, N (%) | ||||
| Inactive | 72 (19) | 94 (21) | ||
| Moderate | 148 (40) | 199 (44) | ||
| Active | 150 (41) | 164 (36) | ||
| Smoker, N (%) | ||||
| Never | 182 (48) | 225 (48) | ||
| Current | 100 (26) | 111 (24) | ||
| Ex | 99 (26) | 133 (28) | ||
| Postmenopausal, N (%) | . | 126 (27) |
Abbreviations: BMI, body mass index; IPAQ, international physical activity questionnaire.
Table 2.
Baseline (2002–2004) metabolic characteristics of men and women with apolipoprotein measurements in the RISC study
| Men (N = 389) | Women (N = 475) | |||
|---|---|---|---|---|
| Median | 25th, 75th percentile | Median | 25th, 75th percentile | |
| Fasting glucose, mmol/l | 5.2 | (4.9, 5.5) | 5 | (4.7, 5.3) |
| Fasting insulin, pmol/l | 31 | (22, 45) | 30 | (20, 40) |
| IS-clamp (baseline), µmol/min·kg_FFM·nM | 118 | (86, 155) | 151 | (113, 194) |
| IS-OGTT (baseline), ml/min·m2 | 427 | (392, 460) | 447 | (411, 487) |
| IS-OGTT (year 3), ml/min·m2 | 414 | (379, 451) | 436 | (391, 481) |
| delta_IS-OGTT, ml/min·m2 | -13 | (-42, 18) | -8.3 | (-45.6, 22.7) |
| GISR (baseline), pmol / [min·m2·mmol/l] | 60 | (42, 88) | 55 | (36, 77) |
| Adiponectin, mg/l | 6.3 | (4.8, 7.9) | 9.2 | (7.2, 12) |
| Leptin, ng/ml | 4.4 | (2.2, 7.5) | 15 | (10, 25) |
| Cholesterol, mmol/l | 4.9 | (4.4, 5.5) | 4.8 | (4.23, 5.3) |
| HDL-cholesterol, mmol/l | 1.2 | (1.1, 1.4) | 1.5 | (1.3, 1.8) |
| LDL-cholesterol, mmol/l | 3.0 | (2.6, 3.6) | 2.8 | (2.3, 3.3) |
| Triglyceride, mmol/l | 1.1 | (0.8, 1.5) | 0.8 | (0.6, 1.1) |
| Free fatty acid, mmol/l | 0.43 | (0.35, 0.56) | 0.56 | (0.45, 0.70) |
| ApoA-I in HDL lacking apoC-III, mg/dl | 91 | (79, 103) | 103 | (89, 119) |
| ApoA-I in HDL containing apoC-III, mg/dl | 7.4 | (6.1, 8.9) | 8.4 | (7.1, 10.4) |
| % of apoA-I in HDL containing apoC-III, % | 7.6 | (6.5, 8.9) | 7.7 | (6.6, 8.8) |
| ApoB in LDL and VLDL lacking apoC-III, mg/dl | 71 | (60, 87) | 67 | (55, 80) |
| ApoB in LDL and VLDL containing apoC-III, mg/dl | 4.4 | (3.4, 5.5) | 4.0 | (3.1, 5.1) |
| % of apoB in LDL and VLDL containing apoC-III, % | 5.7 | (4.6, 6.9) | 5.8 | (4.7, 7.0) |
| ApoC-III, mg/dl | 9.9 | (7.6, 13.1) | 9.8 | (7.7, 12.3) |
| ApoC-III in HDL, mg/dl | 5.9 | (4.6, 7.6) | 5.6 | (4.6, 7.0) |
| ApoC-III in LDL and VLDL, mg/dl | 4.2 | (2.8, 6.0) | 4.1 | (2.8, 5.6) |
| ApoC-III to ApoA-I molar ratio | 2.6 | (2.2, 3.1) | 2.2 | (1.9, 2.6) |
| ApoC-III to ApoB molar ratio | 62 | (43, 79) | 62 | (47, 84) |
Abbreviations: ApoC-III to ApoA-I molar ratio, the molar ratio of apoC-III to apoA-I in apoC-III-containing HDL; ApoC-III to ApoB molar ratio, the molar ratio of apoC-III to apoB in apoC-III-containing LDL and VLDL; FFM, fat-free mass; GISR, glucose-induced secretory response; HDL, high density lipoprotein; LDL, low density lipoprotein; OGTT, oral glucose tolerance test; VLDL, very low density lipoprotein.
Associations between apoA-I concentrations and IS-OGTT at year 3
HDL containing and lacking apoC-III measured at baseline demonstrated significantly opposite associations with IS-OGTT at year 3 (P-heterogeneity = 0.004) (Fig. 1A). Each 1 SD higher HDL lacking apoC-III (19 mg/dL for men and 22 mg/dL for women) at baseline was associated with a 1.3% increase in IS-OGTT at year 3 (95% confidence interval [CI]: 0.3, 2.2), while each 1 SD higher HDL containing apoC-III (2.5 mg/dL for men and 2.7 mg/dL for women) was associated with a 1.2% decrease in IS-OGTT (95% CI: -2.1, -0.3). Total HDL was not significantly associated with IS-OGTT at year 3. All the above observations were corroborated by the findings from analysis using the P-trend.
Figure 1.
Associations between apolipoprotein concentrations and insulin sensitivity (IS-OGTT) at year 3. The RISC study (n = 864).
Means and 95% confidence intervals of IS-OGTT across quintiles of apoA-I (A) and apoC-III (B) from multivariable linear regression models adjusted for age, sex, center, alcohol, smoking, physical activity, BMI, and log-transformed IS-OGTT at baseline. Analysis of the apoA-I concentrations of HDL containing and lacking apoC-III simultaneously adjusted. P-heterogeneity in the panels comparing slopes between HDL containing and lacking apoC-III. Analysis of apoC-III in HDL additionally adjusted for total apoA-I concentration. ApoC-III to apoA-I ratio = the molar ratio of apoC-III to apoA-I in apoC-III-containing HDL. Abbreviations: apo, apolipoprotein; BMI, body mass index; HDL, high density lipoprotein; IS-OGTT, insulin sensitivity measured by oral glucose tolerance test; RISC, Relationship between Insulin Sensitivity and Cardiovascular disease.
Associations between apoC-III concentrations and IS-OGTT at year 3
The concentration of apoC-III contained in HDL demonstrated a strong negative association with IS-OGTT at year 3. For every SD higher apoC-III in HDL (2.5 mg/dL for men and 2.0 mg/dL for women), IS-OGTT decreased by 1.3% (95% CI: -2.1, -0.5) (Fig. 1B). Similar inverse associations were present but weaker for total apoC-III and the molar ratio of apoC-III to apoA-I. Overall, these findings were consistent with analysis using the P-trend.
Sensitivity analysis to assess the influence of the top and bottom 1% trimming
The untrimmed and trimmed results were largely the same (Supplementary Table 1 (37)). The P-heterogeneity between HDL lacking apoC-III and HDL containing apoC-III remained statistically significant (P = 0.02).
Additional adjustment for triglycerides, CRP, and IL-6
All the significant associations between apolipoprotein levels and insulin sensitivity at year 3 were attenuated and became null after additionally adjusting for triglycerides, CRP, and IL-6 (Supplementary Table 2 (37)).
Investigation of potential effect modifiers
Stratified analysis showed qualitative effect modifications by TG, CRP, IL-6, HDL-cholesterol, and sex: heterogeneity of associations between the 2 HDL subspecies was significant in groups of participants with higher TG (≥0.93 mmol/l), higher CRP (≥0.61 mg/l), higher IL-6 (≥0.71 pg/ml), higher HDL-cholesterol (≥1.39 mmol/l), and female sex (Fig. 2). However, when these interactions were formally assessed using interaction terms in regression models, they were not statistically significant (P for interaction all >0.1).
Figure 2.
Associations between apoA-I concentrations and insulin sensitivity (IS-OGTT) at year 3 stratified by TG, CRP, IL-6, sex, BMI, and HDL-cholesterol. The RISC study (n = 864). Means and 95% confidence intervals of IS-OGTT across quintiles of apoA-I from multivariable linear regression models stratified by TG (A), CRP (B), IL-6 (C), sex (D), BMI (E), and HDL-cholesterol (F). Analysis of the apoA-I concentrations of HDL containing and lacking apoC-III simultaneously adjusted. P-heterogeneity in the panels comparing slopes between HDL containing and lacking apoC-III. Abbreviations: apo, apolipoprotein; BMI, body mass index; CRP, C-reactive protein; HDL, high density lipoprotein; IL-6, interleukin 6; IS-OGTT, insulin sensitivity measured by oral glucose tolerance test; RISC, Relationship between Insulin Sensitivity and Cardiovascular disease; TG, triglyceride.
Associations between apolipoprotein concentrations and GISR
HDL containing apoC-III was significantly inversely associated with GISR; each 1 SD higher HDL containing apoC-III was associated with 3.66 pmol/[m2·min·mmol/l] lower insulin secretory response to glucose bolus, while HDL lacking apoC-III was not significantly associated with insulin secretory response (Fig. 3). The heterogeneity between the associations of the 2 HDL subspecies with insulin secretory response was borderline significant (P-heterogeneity = 0.05).
Figure 3.
Associations at baseline between apolipoprotein concentrations and glucose-induced secretory response (GISR). The RISC study (n = 643). Means and 95% confidence intervals of GISR across quintiles of apoA-I (A) and apoC-III (B) from multivariable linear regression models adjusted for age, sex, center, alcohol, smoking, physical activity, and BMI. Analysis of the apoA-I concentrations of HDL containing and lacking apoC-III simultaneously adjusted. P-heterogeneity in the panels comparing slopes between HDL containing and lacking apoC-III. Analysis of apoC-III in HDL additionally adjusted for total apoA-I concentration. ApoC-III to apoA-I ratio = the molar ratio of apoC-III to apoA-I in apoC-III-containing HDL. Abbreviations: apo, apolipoprotein; BMI, body mass index; HDL, high density lipoprotein; RISC, Relationship between Insulin Sensitivity and Cardiovascular disease.
Pearson correlation coefficients between HDL subspecies and other measured variables
Plasma total apoC-III, apoC-III in HDL, and apoC-III in LDL and VLDL had moderate to strong correlations with HDL containing apoC-III. Both HDL containing and lacking apoC-III were moderately associated with HDL cholesterol (Table 3).
Table 3.
Pearson correlation coefficients between apoA-I concentration of HDL subspecies and other variables
| HDL Lacking ApoC-III | HDL Containing ApoC-III | |
|---|---|---|
| Age | 0.07 | 0.16 |
| BMI | -0.17 | -0.12 |
| Systolic blood pressure | -0.04 | -0.01 |
| Diastolic blood pressure | -0.03 | -0.04 |
| Fasting glucose | -0.04 | -0.02 |
| Fasting insulin | -0.18 | -0.11 |
| IS-clamp (baseline) | 0.16 | 0.15 |
| IS-OGTT (baseline) | 0.13 | 0.08 |
| IS-OGTT (year 3) | 0.12 | 0.04 |
| delta_IS-OGTT | 0.01 | -0.04 |
| GISR (baseline) | -0.08 | -0.09 |
| IL-6 | -0.05 | -0.06 |
| CRP | 0.02 | -0.03 |
| Adiponectin | 0.28 | 0.26 |
| Leptin | 0.03 | 0.03 |
| Cholesterol | 0.17 | 0.33 |
| HDL-cholesterol | 0.59 | 0.55 |
| LDL-cholesterol | -0.05 | 0.03 |
| Triglyceride | -0.12 | 0.19 |
| Free fatty acid | 0.12 | 0.12 |
| ApoA-I in HDL lacking apoC-III | 1 | 0.60 |
| ApoA-I in HDL containing apoC-III | 0.60 | 1 |
| % of apoA-I in HDL containing apoC-III | -0.11 | 0.69 |
| ApoB in LDL and VLDL lacking apoC-III | -0.01 | 0.07 |
| ApoB in LDL and VLDL containing apoC-III | 0.25 | 0.45 |
| % of apoB in LDL and VLDL containing apoC-III | 0.29 | 0.46 |
| ApoC-III | 0.31 | 0.67 |
| ApoC-III in HDL | 0.30 | 0.67 |
| ApoC-III in LDL and VLDL | 0.26 | 0.57 |
| ApoC-III to ApoA-I molar ratio | -0.30 | -0.28 |
| ApoC-III to ApoB molar ratio | 0.10 | 0.28 |
Abbreviations: ApoC-III to ApoA-I molar ratio, the molar ratio of apoC-III to apoA-I in apoC-III-containing HDL; ApoC-III to ApoB molar ratio, the molar ratio of apoC-III to apoB in apoC-III-containing LDL and VLDL; GISR, glucose-induced secretory response; HDL, high density lipoprotein; IPAQ, international physical activity questionnaire; LDL, low density lipoprotein; OGTT, oral glucose tolerance test; VLDL, very low density lipoprotein.
Concordance between quintiles of apoA-I in HDL containing apoC-III and quintiles of apoA-I in HDL lacking apoC-III
While a large proportion of individuals had concordant levels of apoA-I in HDL containing apoC-III and apoA-I in HDL lacking apoC-III, there were 7.6% of individuals who had high levels of apoA-I in HDL containing apoC-III (4th or 5th quintile) and low levels of apoA-I in HDL lacking apoC-III (1st or 2nd quintile), and 6.7% of individuals had low levels of apoA-I in HDL containing apoC-III (1st or 2nd quintile) and high levels of apoA-I in HDL lacking apoC-III (4th or 5th quintile) (Supplementary Table 3 (37)).
Associations between apoB concentrations and IS-OGTT at year 3
None of the apoB-related measures (total LDL and VLDL, LDL and VLDL containing and lacking apoC-III, and apoC-III contained in LDL and VLDL) were significantly associated with IS-OGTT assessed at year 3 (Supplementary Table 4 (37)).
Discussion
The present study showed that HDL subspecies defined by the presence or absence of apoC-III have significantly opposite associations with insulin sensitivity assessed at year 3. We found that HDL lacking apoC-III was associated with an increase in insulin sensitivity, whereas HDL containing apoC-III was associated with a reduction in insulin sensitivity at year 3. These findings pertaining to apoC-III-containing HDL were supported by our findings on the apoC-III concentration in HDL, which was strongly associated with a reduction in insulin sensitivity. Total HDL, measured by apoA-I, was not associated with insulin sensitivity. In contrast, the apoB concentrations of LDL and VLDL containing or lacking apoC-III, and apoC-III in LDL and VLDL were not associated with insulin sensitivity.
The current findings and our prior research indicate that the measurement of HDL subspecies based on the presence or absence of apoC-III may be important for the assessment of cardiometabolic disease risk. We previously reported the percentage of HDL containing apoC-III to be approximately 13% to 14% in obese individuals, which was double that of individuals with normal weight (1). HDL containing and lacking apoC-III demonstrated heterogeneous associations with diabetes in 2 prospective cohorts (14) and cross-sectional analyses of 2 cohorts (12), risk of coronary heart disease (CHD) in 4 prospective cohorts (7), and subclinical atherosclerosis in 2 cohort studies (6, 8); HDL lacking apoC-III was associated with beneficial outcomes, whereas HDL containing apoC-III was associated with either null or harmful outcomes. Although measuring the concentration of the HDL subspecies is more complex than assessing the apoC-III concentration in HDL, measuring the HDL subspecies is more informative because it shows both beneficial and harmful associations.
Interestingly, total HDL was not associated with insulin sensitivity at year 3, suggesting that the differential associations seen for HDL subspecies containing and lacking apoC-III may be missed if only total HDL is assessed. There are a number of potential mechanisms through which apoC-III might impair an otherwise favorable association of HDL with insulin sensitivity. HDL lacking apoC-III significantly inhibited human monocyte cell adhesion to cultured endothelial cells, in vitro, while HDL containing apoC-III did not reduce adhesion (2), which suggested that the presence of apoC-III on HDL diminished the anti-inflammatory effect of HDL lacking apoC-III. ApoC-III also promotes β-cell apoptosis by hyper activation of β-cell CaV channels in vitro (17) and in vivo (16), and thus enhances insulin resistance. Consistent with these in vitro and in vivo studies, we showed that HDL containing apoC-III was already associated with lower insulin secretory response at baseline.
In a previous analysis in the RISC study, plasma total apoC-III levels measured by multiplex were found to be inversely associated with insulin sensitivity measured as IS-clamp at baseline (38). Interestingly, we found that apoC-III in HDL was even more strongly and inversely associated with insulin sensitivity than total apoC-III was in the prospective analysis. A previous prospective cohort study in Turkish men and women also reported that apoC-III contained in HDL was associated with the risk of diabetes more strongly than total apoC-III was (21). In another study, the concentration of apoC-III in HDL was significantly higher among obese people as compared with people with normal body weight (1). In contrast, apoC-III contained in LDL and VLDL, which comprised about 40% of apoC-III in whole plasma, was not associated with insulin sensitivity.
The results for apoB-lipoprotein subspecies were not significant. This might be because apoB-containing lipoproteins in this study contained both LDL and VLDL. Mendivil et al showed that LDL containing apoC-III was associated with an increased risk of CHD, whereas the association between LDL lacking apoC-III and CHD risk was null (39). The same study showed no differential effects by VLDL subspecies, with both VLDL containing and lacking apoC-III being associated with an increased risk of CHD. In contrast, Lee et al showed that VLDL containing apoC-III was associated with a decreased risk of recurrent coronary events among diabetic patients, while VLDL lacking apoC-III was associated with increased risk (40). Our unpublished data from the Danish Diet, Cancer and Health study similarly demonstrated that LDL and VLDL containing apoC-III was not associated with incident diabetes. These results point to the need to look at LDL and VLDL separately when examining the associations between their apoC-III defined subspecies and the risk of diabetes or CHD.
There are several strengths in our study. First, the RISC study is the largest study to date that used the gold standard technique, the euglycemic-hyperinsulinemic clamp, for measuring insulin sensitivity in healthy nondiabetic people. Second, the sandwich ELISA-based method used for the exposure assessment has low between-assay variability compared with other methods such as immunoaffinity chromatography. Finally, since all assays for apolipoprotein exposures were performed by 1 analyst, there was no inter-rater variability.
There are a number of limitations in our study. Although we were able to perform a prospective analysis, 3 years might not have been long enough for insulin sensitivity to substantially change, particularly because the participants of this cohort were quite healthy at baseline. Other limitations include the variation across study sites due to the multicenter design of the RISC study as well as measurement error of the modified sandwich ELISA and outcome measures. However, the exposure measurement was performed independently of the outcome assessment and blinded to the outcome status; similarly, the outcome measurement was performed independently of the exposure assessment and blinded to the exposure status. Therefore, it is likely that the variation across centers and measurement error was nondifferential and the effect estimates were pulled towards the null.
In conclusion, in this prospective analysis of healthy men and women in the RISC cohort, 2 HDL subspecies had significantly opposite associations with insulin sensitivity at year 3, with HDL containing apoC-III being associated with a decrease in insulin sensitivity and HDL lacking apoC-III associated with an increase in insulin sensitivity. In addition, the concentration of apoC-III contained in HDL had a strong and significant association with a decrease in insulin sensitivity at year 3. Currently, the quest for prevention and treatment for CVD and diabetes, as regards to HDL, focuses on the cholesterol content of HDL or the size of HDL. Our study suggests that if only the total HDL is assessed, the valuable information on differential associations for HDL subspecies would be missed. For example, if we could identify those with discordant levels of HDL subspecies, in particular those with a high level of HDL containing apoC-III but a low level of HDL lacking apoC-III, this may help identify individuals with high risks and prevent further development of insulin resistance in these individuals. Our results support the potential of HDL apoC-III as a promising target for diabetes prevention and treatment, and we believe that further investigation of apoC-III-defined HDL subspecies would provide us with valuable insights into the development of novel diabetes therapeutic strategies.
Acknowledgments
We would like to express our gratitude to the RISC study investigators for generously providing us with the plasma samples and data of the RISC study.
Financial Support: This work was supported by grants from National Institutes of Health (# R01HL095964) and American Diabetes Association (# 1-15-JF-30). The RISC study was supported by grant from the European Union (# QLG1-CT-2001-01252) and AstraZeneca (Sweden).
Author Contributions: R.Y., F.M.S., and M.K.J. designed the research. R.Y., B.R., S.A., F.B.H., and M.K.J. designed the statistical analysis plan and performed statistical analysis. E.F., S.B., B.B., and A.N. gathered clinical data for the RISC cohort and for the endpoints. R.Y., J.D.F., and F.M.S. were responsible for acquiring data on apolipoprotein exposures. R.Y. drafted the manuscript. All authors interpreted the data, gave inputs on the results and reviewed the manuscript.
Additional Information
Disclosures: M.K.J., J.D.F., and F.M.S. are named as co-inventors on a patent for the apoC-III HDL ELISA that has been awarded to Harvard University. F.M.S. is a consultant for Pfizer and AstraZeneca, and an expert witness for Pfizer.
Data Availability
Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.
The RISC Study
RISC recruiting centers:
Amsterdam, The Netherlands: RJ Heine, J Dekker, S de Rooij, G Nijpels, W Boorsma
Athens, Greece: A Mitrakou, S Tournis, K Kyriakopoulou, P Thomakos
Belgrade, Serbia: N Lalic, K Lalic, A Jotic, L Lukic, M Civcic
Dublin, Ireland: J Nolan, TP Yeow, M Murphy, C DeLong, G Neary, MP Colgan, M Hatunic
Frankfurt, Germany: T Konrad, H Böhles, S Fuellert, F Baer, H Zuchhold
Geneva, Switzerland: A Golay, E Harsch Bobbioni,V. Barthassat, V. Makoundou, TNO Lehmann, T Merminod
Glasgow, Scotland: JR Petrie, C Perry, F Neary, C MacDougall, K Shields, L Malcolm
Kuopio, Finland: M Laakso, U Salmenniemi, A Aura, R Raisanen, U Ruotsalainen, T Sistonen, M Laitinen, H Saloranta
London, England: SW Coppack, N McIntosh, J Ross, L Pettersson, P Khadobaksh
Lyon, France: M Laville, F. Bonnet (now Rennes), A Brac de la Perriere, C Louche-Pelissier, C Maitrepierre, J Peyrat, S Beltran, A Serusclat
Madrid, Spain: R. Gabriel, EM Sánchez, R. Carraro, A Friera, B. Novella
Malmö, Sweden (1): P Nilsson, M Persson, G Östling, (2): O Melander, P Burri
Milan, Italy: PM Piatti, LD Monti, E Setola, E Galluccio, F Minicucci, A Colleluori
Newcastle-upon-Tyne, England: M Walker, IM Ibrahim, M Jayapaul, D Carman, C Ryan, K Short, Y McGrady, D Richardson
Odense, Denmark: H Beck-Nielsen, P Staehr, K Hojlund, V Vestergaard, C Olsen, L Hansen
Perugia, Italy: GB Bolli, F Porcellati, C Fanelli, P Lucidi, F Calcinaro, A Saturni
Pisa, Italy: E Ferrannini, A Natali, E Muscelli, S Pinnola, M Kozakova, A Casolaro, BD Astiarraga
Rome, Italy: G Mingrone, C Guidone, A Favuzzi. P Di Rocco
Vienna, Austria: C Anderwald, M Bischof, M Promintzer, M Krebs, M Mandl, A Hofer, A Luger, W Waldhäusl, M Roden
Project Management Board:
B Balkau (Villejuif, France), F Bonnet (Rennes, France), SW Coppack (London, England), JM Dekker (Amsterdam, The Netherlands), E Ferrannini (Pisa, Italy), A Mari (Padova, Italy), A Natali (Pisa, Italy), J Petrie (Glasgow, Scotland), M Walker (Newcastle, England)
Core laboratories and reading centers:
Lipids Dublin, Ireland: P Gaffney, J Nolan, G Boran
Hormones Odense, Denmark: C Olsen, L Hansen, H Beck-Nielsen
Albumin:creatinine Amsterdam, The Netherlands: A Kok, J Dekker
Genetics Newcastle-upon-Tyne, England: S Patel, M Walker
Stable isotope laboratory Pisa, Italy: A Gastaldelli, D Ciociaro
Ultrasound reading center Pisa, Italy: M Kozakova
ECG reading, Villejuif, France: MT Guillanneuf
Actigraph, Villejuif, France: B Balkau, L Mhamdi
Data Management Villejuif, France, Padova, and Pisa, Italy: B Balkau, A Mari, L Mhamdi, L Landucci, S Hills, L Mota
Mathematical modeling and website management, Padova, Italy: A Mari, G Pacini, C Cavaggion, A Tura
Coordinating office, Pisa, Italy: SA Hills, L Landucci. L Mota
Prior presentation
Parts of this study were presented in abstract form at the American Diabetes Association’s 76th Scientific Sessions, New Orleans, LA, 10–14 June 2016.
References
- 1. Talayero B, Wang L, Furtado J, Carey VJ, Bray GA, Sacks FM. Obesity favors apolipoprotein E- and C-III-containing high density lipoprotein subfractions associated with risk of heart disease. J Lipid Res. 2014;55(10):2167–2177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Kawakami A, Aikawa M, Libby P, Alcaide P, Luscinskas FW, Sacks FM. Apolipoprotein CIII in apolipoprotein B lipoproteins enhances the adhesion of human monocytic cells to endothelial cells. Circulation. 2006;113(5):691–700. [DOI] [PubMed] [Google Scholar]
- 3. Zheng C, Azcutia V, Aikawa E, et al. Statins suppress apolipoprotein CIII-induced vascular endothelial cell activation and monocyte adhesion. Eur Heart J. 2013;34(8):615–624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Stalenhoef AF, Demacker PN, Lutterman JA, van’t Laar A. Apolipoprotein C in type 2 (non-insulin-dependent) diabetic patients with hypertriglyceridaemia. Diabetologia. 1982;22(6):489–491. [DOI] [PubMed] [Google Scholar]
- 5. Mendivil CO, Zheng C, Furtado J, Lel J, Sacks FM. Metabolism of very-low-density lipoprotein and low-density lipoprotein containing apolipoprotein C-III and not other small apolipoproteins. Arterioscler Thromb Vasc Biol. 2010;30(2):239–245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Aroner SA, Koch M, Mukamal KJ, et al. High-density lipoprotein subspecies defined by apolipoprotein C-III and subclinical atherosclerosis measures: MESA (The Multi-Ethnic Study of Atherosclerosis). J Am Heart Assoc. 2018;7(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Jensen MK, Aroner SA, Mukamal KJ, et al. High-density lipoprotein subspecies defined by presence of apolipoprotein C-III and incident coronary heart disease in four cohorts. Circulation. 2018;137(13):1364–1373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Yamamoto R, Sacks FM, Hu FB, et al. ; RISC Investigators . High density lipoprotein with apolipoprotein C-III is associated with carotid intima-media thickness among generally healthy individuals. Atherosclerosis. 2018;269:92–99. [DOI] [PubMed] [Google Scholar]
- 9. Morton AM, Koch M, Mendivil CO, et al. Apolipoproteins E and CIII interact to regulate HDL metabolism and coronary heart disease risk. JCI Insight. 2018;3(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. von Eckardstein A, Widmann C. High-density lipoprotein, beta cells, and diabetes. Cardiovasc Res. 2014;103(3):384–394. [DOI] [PubMed] [Google Scholar]
- 11. Rye KA, Barter PJ, Cochran BJ. Apolipoprotein A-I interactions with insulin secretion and production. Curr Opin Lipidol. 2016;27(1):8–13. [DOI] [PubMed] [Google Scholar]
- 12. Jensen MK, Rimm EB, Furtado JD, Sacks FM. Apolipoprotein C-III as a potential modulator of the association between HDL-cholesterol and incident coronary heart disease. J Am Heart Assoc. 2012;1(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Aroner SA, Furtado JD, Sacks FM, et al. Apolipoprotein C-III and its defined lipoprotein subspecies in relation to incident diabetes: the Multi-Ethnic Study of Atherosclerosis. Diabetologia. 2019;62(6):981–992. [DOI] [PubMed] [Google Scholar]
- 14. Aroner SA, Yang M, Li J, et al. Apolipoprotein C-III and high-density lipoprotein subspecies defined by apolipoprotein C-III in relation to diabetes risk. Am J Epidemiol. 2017;186(6):736–744. [DOI] [PubMed] [Google Scholar]
- 15. Lee HY, Birkenfeld AL, Jornayvaz FR, et al. Apolipoprotein CIII overexpressing mice are predisposed to diet-induced hepatic steatosis and hepatic insulin resistance. Hepatology. 2011;54(5):1650–1660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Holmberg R, Refai E, Höög A, et al. Lowering apolipoprotein CIII delays onset of type 1 diabetes. Proc Natl Acad Sci U S A. 2011;108(26):10685–10689. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Juntti-Berggren L, Refai E, Appelskog I, et al. Apolipoprotein CIII promotes Ca2+-dependent beta cell death in type 1 diabetes. Proc Natl Acad Sci U S A. 2004;101(27):10090–10094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Sundsten T, Ostenson CG, Bergsten P. Serum protein patterns in newly diagnosed type 2 diabetes mellitus–influence of diabetic environment and family history of diabetes. Diabetes Metab Res Rev. 2008;24(2):148–154. [DOI] [PubMed] [Google Scholar]
- 19. Onat A, Hergenç G, Sansoy V, et al. Apolipoprotein C-III, a strong discriminant of coronary risk in men and a determinant of the metabolic syndrome in both genders. Atherosclerosis. 2003;168(1):81–89. [DOI] [PubMed] [Google Scholar]
- 20. Koch M, Furtado JD, Jiang GZ, et al. Associations of anthropometry and lifestyle factors with HDL subspecies according to apolipoprotein C-III. J Lipid Res. 2017;58(6):1196–1203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Onat A, Hergenç G, Ayhan E, et al. Serum apolipoprotein C-III in high-density lipoprotein: a key diabetogenic risk factor in Turks. Diabet Med. 2009;26(10):981–988. [DOI] [PubMed] [Google Scholar]
- 22. Atzmon G, Rincon M, Schechter CB, et al. Lipoprotein genotype and conserved pathway for exceptional longevity in humans. Plos Biol. 2006;4(4):e113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Hills SA, Balkau B, Coppack SW, et al. ; Report prepared on behalf of the EGIR-RISC Study Group. The EGIR-RISC STUDY (The European group for the study of insulin resistance: relationship between insulin sensitivity and cardiovascular disease risk): I. Methodology and objectives. Diabetologia. 2004;47(3):566–570. [DOI] [PubMed] [Google Scholar]
- 24. Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381–1395. [DOI] [PubMed] [Google Scholar]
- 25. Furtado JD, Yamamoto R, Melchior JT, et al. Distinct proteomic signatures in 16 HDL (High-Density Lipoprotein) SUBSPECIES. Arterioscler Thromb Vasc Biol. 2018;38(12):2827–2842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Antibody Citation: RRID:AB_1238930.
- 27. Antibody Citation: RRID:AB_1238925.
- 28. Antibody Citation: RRID:AB_1238779.
- 29. Antibody Citation: RRID:AB_1238789.
- 30. Antibody Citation: RRID:AB_1238837.
- 31. Antibody Citation: RRID:AB_1238848.
- 32. Rosner B, Cook N, Portman R, Daniels S, Falkner B. Determination of blood pressure percentiles in normal-weight children: some methodological issues. Am J Epidemiol. 2008;167(6):653–666. [DOI] [PubMed] [Google Scholar]
- 33. Mari A, Pacini G, Murphy E, Ludvik B, Nolan JJ. A model-based method for assessing insulin sensitivity from the oral glucose tolerance test. Diabetes Care. 2001;24(3):539–548. [DOI] [PubMed] [Google Scholar]
- 34. Mari A, Tura A, Natali A, et al. ; RISC Investigators . Influence of hyperinsulinemia and insulin resistance on in vivo β-cell function: their role in human β-cell dysfunction. Diabetes. 2011;60(12):3141–3147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Rosner B. Percentage points for a generalized ESD many outlier procedure. Technometrics. 1983;25:165–172. [Google Scholar]
- 36. Rosner B. Fundamentals of biostatistics. 7th ed.Boston: Brooks/Cole, Cengage Learning; 2011: xvii, 859. [Google Scholar]
- 37. Yamamoto R, Jensen MK, Aroner SA, et al. “Supplemental material: HDL containing apoC-III and insulin sensitivity,” Harvard Dataverse, V2 2021. ProMED-mail website. Accessed March 5, 2021. 10.7910/DVN/L3RCTM. [DOI]
- 38. Baldi S, Bonnet F, Laville M, et al. ; RISC Investigators . Influence of apolipoproteins on the association between lipids and insulin sensitivity: a cross-sectional analysis of the RISC Study. Diabetes Care. 2013;36(12):4125–4131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Mendivil CO, Rimm EB, Furtado J, Chiuve SE, Sacks FM. Low-density lipoproteins containing apolipoprotein C-III and the risk of coronary heart disease. Circulation. 2011;124(19):2065–2072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Lee SJ, Moye LA, Campos H, Williams GH, Sacks FM. Hypertriglyceridemia but not diabetes status is associated with VLDL containing apolipoprotein CIII in patients with coronary heart disease. Atherosclerosis. 2003;167(2):293–302. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.