Lipoprotein Subpopulation Distributions in Lean, Obese, and Type 2 Diabetic Women: A Comparison of African and White Americans

Lipoprotein Subpopulation Distributions in Lean, Obese, and Type 2 Diabetic Women: A Comparison of African and White Americans Paul S. MacLean, Joseph F. Bower, Satyaprasad Vadlamudi, Thomas Green, and Hisham A. Barakat Abstract MACLEAN, PAUL S., JOSEPH F. BOWER, SATYAPRASAD VADLAMUDI, THOMAS GREEN, AND HISHAM A. BARAKAT. Lipoprotein subpopulation distributions in lean, obese, and type 2 diabetic women: a comparison of African and white Americans. Obes Res. 2000:8:62 70. Objective: Abnormal subpopulation distributions of plasma lipoproteins have been reported in white American (WA) women with obesity and type 2 diabetes that explain part of the elevated rate of cardiovascular disease in these patients. This study examined if these perturbations also occur in obese and diabetic African American (AA) women and compared the lipoprotein profiles with WA counterparts. Research Methods and Procedures: We determined the lipoprotein subpopulation distribution in the plasma of 51 lean women (29 WA, 22 AA, body mass index [BMI] 30), 50 obese women (27 WA, 23 AA, BMI 30), and 43 obese women with type 2 diabetes (27 WA, 16 AA), by nuclear magnetic resonance spectroscopy. Results: AA diabetic women, like WA diabetic women, had a larger average very low density lipoprotein (VLDL) size, elevated levels of small low density lipoprotein cholesterol (LDL-C), and lower levels of small high density lipoprotein cholesterol (HDL-C), when compared to lean controls (p 0.05). These differences were accompanied by higher VLDL-triglycerides (TG) and LDL-C in WA (p 0.05), but not in AA. Although the effects of obesity and diabetes on lipoprotein subpopulation were fairly similar for AA and WA, some racial differences, particularly with respect to HDL, were observed. Submitted for publication March 24, 1999. Accepted for publication in final form July 22, 1999. From the Department of Biochemistry, East Carolina University School of Medicine, Greenville, NC 27858. Address correspondence to Hisham Barakat, PhD, Department of Biochemistry, East Carolina University School of Medicine, Greenville, NC 27858. E-mail: BARAKAT@ brody.med.ecu.edu Copyright 2000 NAASO. Discussion: The atherogenic perturbations in lipoprotein profiles of obese AA women, particularly those with diabetes, were relatively similar to those found in WA women and may be contributing to the increased rate of cardiovascular disease (CVD) in AA with obesity and diabetes. The parameters of subpopulation distribution may provide better markers for CVD than lipid concentrations alone, particularly in AA women. Furthermore, subtle racial differences in lipoprotein profiles suggest that race-specific criteria may be needed to screen patients for CVD. Key words: VLDL, LDL, HDL, NMR spectroscopy, CETP Introduction It is well established that the risk and incidence of cardiovascular disease (CVD) is related to alterations in the concentrations of plasma lipids and lipoproteins. The major features of this dyslipidemia shown to be associated with CVD are elevations in total and LDL cholesterol (TC, LDL-C), higher total triglycerides (TG), and a lower HDL cholesterol (HDL-C). In addition to the changes in plasma lipid concentrations, alterations in the chemical and physical characteristics of the lipoproteins, which cause shifts in the subpopulation distribution of these lipoproteins, have been shown to contribute to CVD risk. Small and dense LDL and large very low density lipoprotein (VLDL) particles have been shown to be more prevalent in the plasma of patients with heart disease (1 6). Furthermore, patients with CVD tend to have less of the larger HDL particles and more of the smaller HDL particles (6 9). One recent study has indicated that analysis of the subpopulation distribution of the plasma lipoproteins may be more predictive of occlusive disease than determination of the lipid concentrations (6). Characterizing the relationship between dyslipidemia and CVD has been the focus of a large body of research in recent years, with the goal of identifying those patients who are at a high risk of CVD development and providing a basis for therapeutic approaches. 62 OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000

The risk and incidence of CVD is higher in obese patients with or without diabetes than lean patients of the same age (10,11). Patients with type 2 diabetes have a 4-fold higher risk of CVD than obese non-diabetics (11). This increase in CVD in patients with diabetes could be explained, in part, by changes in plasma lipid concentrations (12 14). We, and others, have also shown that there are differences in the subpopulation distribution of plasma lipoproteins that may predispose patients with diabetes to a higher risk for CVD. Patients with type 2 diabetes have a higher proportion of small and dense LDL and a lower proportion of larger HDL subfractions than non-diabetic controls (13,14). Thus, changes in the subpopulation distribution of plasma lipoproteins may contribute to the increased risk and incidence of CVD in patients with diabetes. African American (AA) women have a higher incidence of obesity (15,16) and diabetes (17), than white women. These elevated rates of obesity and diabetes are associated with an increased risk and incidence of vascular diseases (18,19) as in whites. However, African American women do not exhibit the aberrant lipid concentrations, to the same extent and frequency, that are thought to explain the relationship of obesity and diabetes to CVD, i.e., low HDL-C, high LDL-C, and high TG (20 24). Because the concentrations of the plasma lipids do not explain the higher CVD in obese and diabetic African American women, it is possible that this increase may result, partly, from alterations in the subpopulation distribution of the plasma lipoproteins. To date, there is a paucity of information on the subpopulation distribution of plasma lipoproteins in lean, obese, and diabetic African American women. Thus, this study was undertaken in order to address the following two aims. First: to determine if there are shifts in the subpopulation distribution of the lipoproteins toward more atherogenic profiles in obese and diabetic African American women. We determined the differences in the subpopulation distribution of plasma VLDL, LDL, and HDL, of lean, obese, and diabetic African American women. Second: to determine if the alterations in subpopulation distribution that accompany obesity and diabetes in African American women differ from those found with obesity and diabetes in white American (WA) women. We compared the lipoprotein profiles of lean, obese, and diabetic African American women to white counterparts. Determination of the subpopulation distribution was done by analysis of whole plasma by nuclear magnetic resonance (NMR) spectroscopy. In addition, we examined the activity of plasma cholesteryl ester transfer protein (CETP), a key enzyme in reverse cholesterol transport that has a role in shaping the lipoproteins in the plasma. Methods and Procedures Patient Characteristics and Treatment Fifty-one lean women (29 WA, 22 AA, body mass index [BMI] 30), 50 obese women (27 WA, 23 AA, BMI 30), and 43 obese women with type 2 diabetes (27 WA, 16 AA) participated in this study. Obese and diabetic subjects were recruited from patients examined in the Diabetes and Obesity Center at East Carolina University. Patients were not considered as potential participants if the physician noted complications other than obesity and/or diabetes (severe vascular problems, cancer, extreme emotional distress, etc.). Patients that were taking medications for hyperlipidemia and/or hypercholesterolemia were not included in the study. Lean subjects were recruited from faculty, staff, and students in the School of Medicine and were generally in good health. AA subjects in each group were stratified to subjects in the corresponding WA group, according to BMI. AA women were included in the study only if their parents and grandparents were of African American descent. Diagnosis of type 2 diabetes was based on the criteria of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus (25). Approximately 40% of diabetics were receiving treatment to help control blood glucose (sulfonylureas, insulin, etc.) at the time blood was collected. Body mass and height were recorded to the nearest 0.1 kg and 0.1 cm, respectively, and BMI was calculated (body weight in kg/(height in m) 2 ). Written consent was obtained from the patients after they were informed of the nature and potential risk of the study. The Institutional Review Board for human subject research approved the protocols used in this study. Plasma Analysis Blood was collected from the patients following a 12- hour fast, and a preservative mixture containing sodium azide (50 mg/ml) and aprotinin (1 TIU/mL) were added as preservatives (2 L/mL blood). Plasma was prepared by centrifugation at 1500g for 30 minutes at 4 C and stored at 80 C in aliquots until analyzed. Samples were analyzed spectrophotometrically for glucose (16-UV; Sigma Chemical Co., St. Louis, MO) and by microparticle enzyme immunoassay for insulin (IMx; Abbott Labs, Abbott Park, Il). CETP-specific cholesteryl ester transfer in the plasma was measured as the amount of [ 3 H]cholesteryl esters ( 3 H-CE; Dupont NEN) transferred from 3 H-CE HDL to apoprotein B-containing lipoproteins that could be inhibited by a human CETP-specific monoclonal antibody (TP2), as previously described (26). Lipid concentrations and lipoprotein subpopulation distributions were determined by NMR spectroscopy (6,27,28). As described by Freedman et al. (6), NMR analysis of lipoprotein subclasses is based upon each lipoprotein particle within a given diameter range exhibiting a distinct lipid NMR signal, the intensity of which is proportional to its bulk lipid mass concentration. NMR data are collected in three steps: 1) acquisition of the 250-MHz proton NMR spectra of the plasma sample; 2) deconvolution of the lipid methyl group signal envelope OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000 63

appearing in these spectra at approximately 0.8 ppm, yielding signal amplitudes for 15 lipoprotein subclasses; and 3) conversion of the signal amplitudes into subclass concentrations with experimentally derived factors that are based upon chemically determined isolated subclass standards. The estimated diameters in nanometers for the 15 subclasses (six VLDL: V6, 150 70; V5, 70 10; V4, 50 10; V3, 38 3; V2, 33 2; V1, 29 2; four LDL: IDL, 25 2; L3, 22 0.7; L2, 20.5 0.7; L1, 19 0.7; five HDL: H5, 11.5 1.5; H4, 9.4 0.6; H3, 8.5 0.3; H2, 8.0 0.2; H1, 7.5 0.2) have been previously reported (28). The 15 subclasses were then grouped into nine subclasses based upon the relationship of NMR subclasses to particle size estimates of gradient gel electrophoresis or electron microscopy measurements (28,29). The grouping of the subfractions are as follows: large VLDL (V6 V5), intermediate VLDL (V4), small VLDL (V3 V2 V1), large LDL (IDL L3), intermediate LDL (L2), small LDL (L1), large HDL (H5), intermediate HDL (H4 H3), and small HDL (H2 H1). Plasma lipids were calculated from subclass determinations. LDL particle concentration (LDL #) is calculated from the average size and the distribution data from LDL. HDL diff is a component of HDL lipoprotein subpopulation distribution that has a strong association with occlusive disease (6) and is calculated as large HDL intermediate HDL small HDL. Particle size for each class of lipoprotein was determined by weighting the relative percentage of each subclass by its diameter. Close agreement has been observed between chemically determined lipid concentrations and those determined with NMR (29). We confirmed these findings in a limited number of our patients (data not shown). LDL and HDL subclass determinations by gradient gel electrophoresis and NMR have also been closely related (28). In a limited number of patients, we also found that LDL size determined by NMR was closely correlated to LDL size determined from polyacrylamide gradient gel electrophoresis (PGGE) (r 0.77, p 0.0001). LDL size determined by NMR was approximately 15% lower than for our PGGE determinations, a consistent finding when comparing the two methods. The reason for this methodological difference is unknown. However, LDL size determinations by light scattering techniques also report LDL sizes considerably lower than those determined by PGGE, whereas results from the two methods correlate strongly (30). Oneal et al. (30) discuss some possibilities that may lead to the differences in absolute LDL size values. As a general reference, patients with the PGGE classification of the less atherogenic, type A profile have an NMR LDL size greater than 20.6 nm, whereas those with type B have an NMR LDL size less than 20.6 nm (NMR Lipoprofile; Lipomed, Raleigh, NC). Statistics The first objective of this paper was to determine if the subpopulation distribution in AA women is altered in obese and diabetic states, as has been shown in WA women. To achieve the first objective, data from each race was analyzed in separate one-way ANOVA models. Fisher s least significant difference post hoc test was used to determine differences between lean, obese, and diabetic groups. The second objective of this study was to determine if the alterations in lipoprotein subpopulation distribution found with obesity and diabetes differ between AA and WA women. To achieve the second objective, the data from both races were combined into a two-way ANOVA model to test for the main effects of group and race and the interaction between the two (group*race). Fisher s least significant difference (LSD) post hoc test was used to identify differences in group-related pairwise comparisons (lean vs. obese, obese vs. type 2, and lean vs. type 2). Results for general patient characteristics are expressed as means SEM. In that lean patients were observed to be younger than obese and type 2 groups, results for plasma lipids, lipoprotein subpopulation distribution, and CETP activity are expressed as age-adjusted means SEM, and analyses for these variables were conducted on age-adjusted residuals (Systat/SPSS, SPSS Inc., Chicago, IL). Results Table 1 shows the physical and biochemical characteristics of the subjects in this study. The results were generally similar for African and white Americans. In the combined analysis, the lean subjects were younger and had a lower body weight and BMI than the obese and type 2 groups. Fasting glucose levels in the type 2 were higher than in the obese and lean groups. Insulin levels were higher in the obese patients than in lean controls, and higher in type 2 than in both obese and lean patients. Table 2 shows the plasma lipid concentrations. In AA women, no differences were observed between lean, obese, and type 2 patients with respect to plasma lipid concentrations. In contrast, WA type 2 patients had elevated TC, LDL-C, LDL#, TG, and VLDL-TG, when compared to their lean and obese counterparts. When data from the patients were analyzed in a two-way model, LDL-C and LDL# were both elevated in obese and type 2 patients, when compared to lean. Furthermore, AA women had higher HDL-C, lower TG, and lower VLDL-TG, regardless of the presence of obesity and diabetes. The average lipoprotein particle sizes are also shown in Table 2. In AA, VLDL size was larger in type 2 women than in lean or obese groups. Trends for smaller LDL and HDL particle sizes were also observed in the type 2 group, although these differences did not meet the criteria for statistical significance. In WA, both obese and type 2 patients had larger VLDL and smaller HDL, when compared 64 OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000

Table 1. Physical characteristics African American White American Lean Obese Type 2 Lean Obese Type 2 n 22 23 16 29 27 27 Age (yr) a-1,3 32 1* 40 2 42 3 37 3 42 2 40 2 Weight (kg) a-1,3 63 3 148 7 142 9 64 2 137 6 142 4 BMI (kg/m 2 ) a-1,3 23.5 0.7 51.6 2.1 54.1 3.5 23.8 0.5 51.2 2.3 54.2 1.1 Glucose (mm) a-2,3 4.61 0.11 4.94 0.17 9.66 1.00 4.66 0.10 4.88 0.16 8.44 0.56 Insulin (pm) a-1,2,3 34 4 114 15 139 25 34 4 96 13 163 21.6 * Values are shown as means SEM for each group (lean, obese, and diabetic) and race (AA and WA) combination. To examine the effects of obesity and diabetes in each race, data in each race were analyzed by separate one-way ANOVA models. Fisher s LSD post hoc test was used to test for differences in pairwise comparisons. Significantly different from corresponding lean subset (p 0.05). Type 2 subset significantly different from corresponding obese subset (p 0.05). To examine racial differences, the data from both races were combined into two-way ANOVA model. Statistical significance was inferred with p 0.05 for the main effect of group a. When a group effect a was observed, Fisher s LSD post hoc test was used to identify differences between the three categories (lean vs. obese a-1, obese vs. NIDDM a-2, and lean vs. NIDDM a-3 ). No variable showed a significant main effect of race or an interaction between group and race. to lean. Furthermore, LDL size was significantly smaller in type 2 patients when compared to lean patients. Consistent with these findings, a significant group effect was observed for all three of these parameters when data from the two races were combined in a two-way model. VLDL size was larger in obese patients than in lean, and even larger in type 2 patients, regardless of race. LDL size was smaller in type 2 patients, when compared to both lean and obese patients. HDL size was smaller in obese and type 2 patients, when compared to lean. No effect of race nor interaction between the factors of group and race were observed with the lipoprotein particle sizes. CETP activity (Table 2) was elevated in AA obese women when compared to lean and type 2 patients. This trend was also found in WA women. As expected, in a two-way model, obese women had elevated CETP activity compared to lean and type 2 women. Interestingly, AA women had higher CETP activity than WA women, regardless of the presence of obesity or diabetes. No interaction between the group and race effects was observed. Figure 1 shows the subpopulation distribution of VLDL in AA and WA women. In both races, large VLDL concentrations were higher in type 2 patients than in the lean subjects. In AA women, small VLDL-TG levels were higher in obese subjects when compared to lean group. In WA subjects, the type 2 group had a greater amount of intermediate VLDL-TG than both lean and obese groups. In the two-way model, WA women had higher concentrations of large, intermediate, and small VLDL-TG than AA women, regardless of obesity and diabetes (race effect, p 0.05). A group effect was only observed for large VLDL-TG concentrations (p 0.05), with type 2 patients having higher levels than obese and lean subjects (p 0.05). No interaction between the group and race effects (group*race) was observed for any parameter of VLDL subpopulation distribution in the two-way model. LDL subpopulation distribution is shown in Figure 2. In AA women, small LDL-C levels were higher in type 2 compared to lean and obese groups. In WA women, both intermediate and small LDL-C concentrations were higher in the type 2 patients than in obese and lean women. In the two-way model, a group effect was observed for small LDL-C, with statistical differences reflecting those found in the separate analyses (p 0.05). No race effect nor interaction between effects (group*race) was observed for any of the parameters of LDL subpopulation distribution. Figure 3 shows the subpopulation distribution of HDL. In AA women, large HDL-C concentrations were significantly lower in type 2 when compared to the lean group. Obese women had lower intermediate HDL-C concentrations than lean patients, but no differences were observed in small HDL-C concentrations between any of the three AA groups. In WA women, the concentration of large HDL-C were significantly lower in both obese and type 2 groups, when compared to the lean group. Obese women had higher concentrations of intermediate HDL than in lean, but no differences were observed in small HDL concentrations. HDL diff (Table 2), a parameter of HDL profiles that has a strong negative association with vascular disease, was lower in AA obese and type 2 women than in lean counterparts. OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000 65

Table 2. Lipid concentrations, average lipoprotein particle sizes, and CETP activity African American White American Lean Obese Type 2 Lean Obese Type 2 n 22 23 16 29 27 27 TC (mm) 4.74 0.24* 4.73 0.27 4.73 0.21 4.43 0.21 4.54 0.15 5.25 0.16 LDL-C (mm) a-2,3 3.10 0.21 3.26 0.22 3.29 0.17 2.97 0.19 3.08 0.12 3.68 0.15 LDL# (nm) a-2,3 1283 92 1345 91 1408 77 1211 88 1292 56 1605 69 HDL-C (mm) b 1.36 0.8 1.14 0.06 1.23 0.09 1.07 0.04 1.05 0.05 1.06 0.06 HDL diff (mm) c 0.56 0.09 0.31 0.06 0.32 0.10 0.34 0.06 0.44 0.06 0.37 0.06 TG (mm) b 0.98 0.08 1.12 0.10 1.11 0.10 1.32 0.14 1.37 0.10 1.72 0.15 VLDL-TG (mm) b 0.57 0.07 0.69 0.08 0.66 0.10 0.92 0.13 0.97 0.10 1.27 0.15 VLDL size (nm) a-1,2,3 46.9 1.8 49.2 1.9 55.6 2.2 48.3 1.6 53.0 1.6 56.2 1.6 LDL size (nm) a-2,3 20.98 0.10 20.89 0.08 20.72 0.14 21.07 0.09 20.86 0.07 20.66 0.10 HDL size (nm) a-1,3 9.27 0.09 9.15 0.07 9.02 0.07 9.24 0.05 9.06 0.6 9.00 0.07 CETP (nmol/ml/h) a-1,2;b 73.5 0.30 87.3 3.4 75.7 4.1 72.0 2.7 78.5 3.0 67.6 3.0 * Values are shown as age-adjusted means SEM for each group (lean, obese, and diabetic) and race (AA and WA) combination. LDL# is the concentration of LDL particles. HDL diff, a strong predictor of vascular disease, is calculated as large intermediate small HDL. Age-adjusted residuals were used in the analyses. To examine the effects of obesity and diabetes in each race, data in each race were analyzed by separate one-way ANOVA models. Fisher s LSD post hoc test was used to test for differences in pairwise comparisons. Significantly different from corresponding lean subset (p 0.05). Type 2 subset significantly different from corresponding obese subset (p 0.05). To examine racial differences, the data from both races were combined into a two-way ANOVA model. Data were analyzed by two-way ANOVA with age as a covariate. Statistical significance was inferred with p 0.05 for the main effects of group a, race b, and the interaction between the main effects c. When a group effect a was observed, Fisher s LSD post hoc test was used to identify differences between the three categories (lean vs. obese a-1 ; obese vs. type 2 a-2 ; and lean vs. type 2 a-3 ), using age-adjusted residuals. These differences were not observed in WA women. In the two-way model, both obese and type 2 patients had lower levels of large HDL-C than lean (p 0.05). For both intermediate HDL-C and HDL diff, an interaction between the main effects (group*race) was observed (p 0.05). Furthermore, AA women were observed to have a higher concentration of small HDL-C than WA women (race effect, p 0.05). Interestingly, the atherogenic alterations in the subpopulation distribution of VLDL, LDL, and HDL were closely related to plasma TG levels in WA women, whereas in AA women, these changes were more related to plasma glucose levels (data not shown). Discussion Although the increased rate of CVD in obese and diabetic WA subjects has been attributed, in part, to alterations in lipid concentrations, obese and diabetic AA subjects have elevated CVD without these perturbations. These observations have suggested that dyslipidemia in AA does not play a role in CVD risk. However, abnormal subpopulation distributions of the lipoproteins have also been related to CVD, and there are no reports of subpopulation distribution of lipoproteins in obese and diabetic AAs. Thus, we investigated the subpopulation distribution of the plasma lipoproteins in lean, obese, and diabetic women, selected from both the AA and the WA populations. Given that obesity is more severe in AA than WA (15,16), we stratified AA and WA patients in each group according to BMI, so that we could examine the relative contribution of adiposity to alterations in the subpopulation distribution in each race in the presence and absence of diabetes. As has been shown by other researchers using more traditional methods (21,22,24), AA women in this study tended to have higher HDL-C and lower TG, regardless of the level of obesity or presence of diabetes. Furthermore, although diabetic WA women appeared to exhibit more atherogenic lipid concentrations than their lean and obese counterparts, the lipid concentrations of diabetic AA women were relatively similar to those found in lean and obese AA women. In addition to confirming what has been found with respect to plasma lipid concentrations in AA and WA women, this study also supports our previous studies in WA subjects in which more traditional methods to determine lipoprotein subpopulation distribution were used. In the 66 OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000

Figure 1. VLDL subpopulation distribution in lean, obese, and type 2 patients. Separate comparisons are shown for (A) African Americans and (B) white Americans. Values are shown as ageadjusted means SEM. To examine the effects of obesity and diabetes in each race, data in each race were analyzed by separate one-way ANOVA models. Fisher s LSD post hoc test was used to test for differences in pairwise comparisons. Significantly different from corresponding lean subset (p 0.05). Type 2 subset significantly different from corresponding obese subset (p 0.05). previous reports, we found smaller, more dense LDL and HDL in morbidly obese, diabetic women (13,14). This report extends our finding to include the changes in the subpopulation distribution of VLDL. The results of the present study show that VLDL subpopulation distribution is perturbed in severely obese, diabetic women, in that they have larger VLDL particles. Furthermore, it appears that similar atherogenic alterations in VLDL, LDL, and HDL profiles also occur to some extent in severely obese, WA women without diabetes. The novel findings of this study are that we observed atherogenic perturbations in VLDL, LDL, and HDL subpopulation distributions in the severely obese AA women, particularly those with type 2. AA diabetic women were Figure 2. LDL subpopulation distribution in lean, obese, and type 2 patients. Separate comparisons are shown for (A) African Americans and (B) white Americans. Values are shown as ageadjusted means SEM. To examine the effects of obesity and diabetes in each race, data in each race were analyzed by separate one-way ANOVA models. Fisher s LSD post hoc test was used to test for differences in pairwise comparisons. Significantly different from corresponding lean subset (p 0.05). Type 2 subset significantly different from corresponding obese subset (p 0.05). found to have larger VLDL particles, higher levels of small LDL-C, and lower levels of large HDL-C, than those found in lean women. Furthermore, HDL diff, which was reported to be a strong negative correlate of vascular disease (6), was lower in AA diabetic women than in lean counterparts. All of these alterations in lipoprotein profiles have been associated with vascular disease in other and/or combined populations (1 9) and may be contributing to the elevated disease rates in AA with type 2. In obese AA, potentially atherogenic perturbations in the subpopulation distribution were observed in an elevated small VLDL-TG, as well as in depressed HDL diff and intermediate HDL-C concentrations. OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000 67

Figure 3. HDL subpopulation distribution in lean, obese, and type 2 patients. Separate comparisons are shown for (A) African Americans and (B) white Americans. Values are shown as age-adjusted means SEM. To examine the effects of obesity and diabetes in each race, data in each race were analyzed by separate one-way ANOVA models. Fisher s LSD post hoc test was used to test for differences in pairwise comparisons. Significantly different from corresponding lean subset (p 0.05). Lipoprotein particles tended to be more atherogenic than in the lean group, but these tendencies were not as severe as was found in the type 2 group. These data indicate that alterations in the subpopulation distribution of the lipoproteins in obese and type 2 women of AA descent may be contributing to the elevated risk for vascular disease in these abnormal physiological states. Furthermore, particularly in AA women, the subpopulation distribution of the lipoproteins have better clinical markers to identify high risk patients than the commonly used lipid concentrations. In addition to reporting the alterations in lipoprotein subpopulation distribution in AA women that occur with obesity and diabetes, we were also able to compare the alterations that occur in WA counterparts. Reflecting their lower TG and higher HDL-C concentrations, all subfractions of VLDL were lower, and small HDL-C levels were higher, in AA women. Although the atherogenic perturbations in particle size and subpopulation distribution of the lipoproteins in obese and type 2 women were generally similar for AA and WA women, our data suggest that subtle racial differences may exist. Most striking, the effects of obesity and diabetes regarding intermediate HDL-C and HDL diff were significantly different for AA and WA women. Our data also suggest that there are more subtle racial differences in how obesity and diabetes affected the concentrations of intermediate and small VLDL-TG and intermediate LDL-C concentrations. Although these racial differences are interesting, it is difficult to interpret the significance of these findings without further studies. We are unaware of any large, well-controlled studies that suggest different rates of vascular disease in AA and WA women. From a practical standpoint, however, these racial differences may suggest that race-specific criteria are needed in the clinical setting to better identify patients at high risk for vascular disease. In this respect, studies investigating how these subpopulation distribution parameters relate to vascular disease in race-specific populations are warranted. These studies are particularly important for the AA population, as these parameters of subpopulation distribution may be found to be markedly more predictive of vascular disease than lipid concentrations alone. Although we did not extensively examine the mechanisms behind the racial differences in the subpopulation distribution of the lipoproteins, we speculate that they may be related to the variations in the plasma enzymes known to modulate the composition of lipoproteins. Lipoprotein lipase (LPL), an enzyme that clears TG from plasma to the peripheral tissues, has been shown to be higher in a subset of the AA population when compared to a comparable WA subset (31). If this difference persists between all subset comparisons, the AA population may have lower TG due to an increased ability to clear TG from the plasma. LPL also has been shown to be depressed in diabetics and may account for the elevation in TG in our diabetic WA patients (32). This idea warrants further investigation as we did not measure LPL in our patients. A complimentary explanation could be derived from the comparison of plasma CETP activity in WA and AA observed in each patient group. Although CETP is primarily known for its putative role in at least one arm of reverse cholesterol transport pathway (33), it also clears TG from VLDL by transporting it to HDL. These TG eventually end up in the liver by the action of hepatic triglyceride lipase (HL). Altered expression of all three of these enzymes, LPL, CETP, and HL, contributes to 68 OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000

the remodeling of the lipoproteins and may mediate the effects of obesity and diabetes on lipoprotein subpopulation distribution (32,34,35). It should be noted that the number of patients examined in this study is relatively small for the amount of variability that is usually found in human studies. In that patient inclusion was affected by a number of potentially complicating variables, the confirmation of our findings in a larger, population-based study is warranted. Furthermore, the patients that we examined were severely obese, and observing how obesity and diabetes affect lipoprotein subpopulation distribution, as well as how racial differences are manifested, in overweight and mildly obese individuals would be of interest. Likewise, similar investigations in men, in whom vascular disease rates are higher, would be warranted. Even in light of the limitations and the number of questions raised by our study, this initial look at subpopulation distribution of the lipoproteins in AA women may have clinical relevance now that determining lipoprotein profiles by NMR is commercially available to physicians. In summary, our data suggest that severely obese AA women, particularly those with diabetes, have perturbations in the lipoprotein subpopulation distributions that imply an increased risk for vascular disease. Interestingly, these alterations were not accompanied by drastically perturbed lipid concentrations, as is frequently found in other populations. These findings suggest that, in AA women, lipoprotein subpopulation distribution parameters may be more effective in predicting vascular disease than lipid concentrations alone. Although the alterations in lipoprotein subpopulation distributions found in obesity and type 2 were generally similar for AA and WA women, subtle racial differences did exist, particularly with respect to HDL. These observations suggest that the criteria for predicting vascular disease may need to be race-specific. Given the prevalence of obesity and diabetes in AA and the substantial health care costs that are involved in treating vascular disease in these patients, studies designed to clarify the relationship between lipoprotein subpopulation distribution and vascular disease in the AA population are warranted. Acknowledgments We thank Dr. Jim Otvos in the Department of Biochemistry at North Carolina State University for assistance in the processing and interpretation of NMR data. This work was supported by a grant from the National Institutes of Health (Department of Health and Human Services Public Health Service DK 45029). References 1. Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Low-density lipoprotein subclass patterns and risk of myocardial infarction. JAMA. 1988;260: 1917 21. 2. Campos H, Genest JJ Jr, Blijlevens E, et al. Low density lipoprotein particle size and coronary artery disease. Arterioscler Thromb. 1992;12:187 95. 3. Gardner CD, Fortmann SP, Krauss RM. Association of small low-density lipoprotein particles with the incidence of coronary artery disease in men and women. JAMA. 1996;276: 875 81. 4. Griffin BA, Freeman DJ, Taait GW, et al. Role of plasma triglyceride in the regulation of plasma low density lipoprotein (LDL) subfractions: relative contribution of small, dense LDL to coronary heart disease risk. Atherosclerosis. 1994;106:241 53. 5. Stampfer MJ, Krauss RM, Ma J, et al. A prospective study of triglyceride level, low-density lipoprotein particle diameter, and risk of myocardial infarction. JAMA. 1996;276:882 8. 6. Freedman DS, Otvos JD, Jeyarajah EJ, Barboriak JJ, Anderson AJ, Walker JA. Relation of lipoprotein subclasses as measured by proton nuclear magnetic resonance spectroscopy to coronary artery disease. Arterioscler Thromb Vasc Biol. 1998;18:1046 53. 7. Wilson HM, Patel JC, Russe D, Skinner ER. Alterations in the concentration of an apolipoprotein E-containing subfraction of plasma of high density lipoprotein in coronary heart disease. Clin Chim Acta. 1993;220:175 87. 8. Johansson J, Carlson LA, Landou C, Hamsten A. High density lipoproteins and coronary atherosclerosis: a strong inverse relation with the largest particles is confined to normotriglyceridemic patients. Arterioscler Thromb. 1991; 11:174 82. 9. Laakso M, Voutilainen E, Pyörälä K, Sarlund H. Association of low HDL and HDL 2 cholesterol with coronary heart disease in noninsulin-dependent diabetics. Arteriosclerosis. 1985; 5:653 8. 10. Manson JE, Colditz GA, Stampfer MJ, et al. A prospective study of obesity and risk of coronary heart disease. N Engl J Med. 1990;322:882 9. 11. Abbott RD, Donahue RP, Kannel WB, Wilson WF. The impact of diabetes on survival following myocardial infarction in men vs. women. JAMA. 1988;260:3456 60. 12. Kannel WB, McGee DL. Diabetes and cardiovascular disease: the Framingham Study. JAMA. 1979;24:2035 8. 13. Barakat HA, Carpenter JW, McLendon VD, Khazanie P, Leggett N, Heath J, Marks R. Influence of obesity, impaired glucose tolerance, and NIDDM on LDL structure and composition: possible link between hyperinsulinemia and atherosclerosis. Diabetes. 1990;39:1527 33. 14. Barakat HA, McLendon VD, Marks R, Pories W, Heath J, Carpenter JW. Influence of morbid obesity and non-insulindependent diabetes mellitus on high-density lipoprotein composition and subpopulation distribution. Metabolsim. 1992;41: 37 41. 15. Dawson DA. Ethnic differences in female overweight: data from the 1985 National Health Interview Survey. Am J Public Health. 1988;129:1376 9. 16. Kuczmarski RJ, Flegel KM, Campbell SM, Johnson CL. Increased prevalence of overweight among U.S. adults. The National Health & Nutrition Examination Survey. 1960 to 1961. JAMA. 1994;272:205 11. OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000 69

17. Lipton RB, Liao Y, Cao G, Cooper RS, McGee D. Determinants of incident non-insulin dependent diabetes mellitus among blacks and whites in a national sample: The NHANES I Epidemiological Follow-up Study. Am J Epidemiol. 1993; 138:826 39. 18. Folsom AR, Stevens J, Schreiner PJ, McGovern PG. Body mass index, waist/hip ratio, and coronary heart disease incidence in African Americans and whites. Am J Epidemiol. 1998;148:1187 94. 19. Tull ES, Roseman JM. Diabetes in African Americans. In: Diabetes in America, 2nd ed. Bethesda, MD: National Institutes of Health; 1995:613 25. 20. Dowling HJ, Pi-Sunyer FX. Race-dependent health risks of upper body obesity. Diabetes. 1993;42:537 43. 21. Cowie CC, Howard BV, Harris MI. Serum lipoproteins in African-Americans and whites with non-insulin dependent diabetes in the US population. Circulation. 1994;1690:1185 91. 22. Conway JM, Yanovski SZ, Avila NA, Hubbard VS. Visceral adipose tissue differences in black and white women. Am J Clin Nutr. 1995;61:765 71. 23. Gaillard TR, Schuster DP, Bossetti BM, Green PA, Osei K. The impact of socioeconomic status on cardiovascular risk factors in African-Americans at high risk for type II diabetes. Diabetes Care. 1997;20:745 52. 24. Albu JB, Murphy L, Frager DH, Johnson JA, Pi-Sunyer FX. Visceral fat and race-related health risks in obese nondiabetic premenopausal women. Diabetes. 1997;46:456 62. 25. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care. 1997;20:1183 97. 26. Vadlamudi S, MacLean P, Green T, et al. Role of female sex steroids in regulating cholesteryl ester transfer protein in transgenic mice. Metabolism. 1998;47:1048 51. 27. Otvos JD, Jeyarajah EJ, Bennett DW, Krauss RM. Development of a proton nuclear magnetic resonance spectroscopic method for determining plasma lipoprotein concentrations and subspecies distributions from a single, rapid measurement. Clin Chem. 1992;38:1632 8. 28. Otvos, JD. Measurement of lipoprotein subclass profiles by NMR spectroscopy. In: Rifai et al., eds. Handbook of Lipoprotein Testing. Washington, DC: AACC Press; 1997, pp. 497 508. 29. Otvos J, Jeyarajah E, Bennett D. A spectroscopic approach to lipoprotein subclass analysis. J Clin Ligand Assay. 1996; 19:184 9. 30. O Neal D, Harrip P, Dragicevic G, Rae D, Best JD. A comparison of LDL size determination using gradient gel electrophoresis and light scattering methods. J Lipid Res. 1998;39:2086 90. 31. Ama PFM, Poehlman ET, Simoneau JA, et al. Fat distribution and adipose tissue metabolism in nonobese male black African and Caucasian subjects. Int J Obes. 1986;10:503 10. 32. Taskinen MR, Kuusi T, Helve E, Nikkilä EA, Yki-Järvinen H. Insulin therapy induces antiatherogenic changes of serum lipoproteins in noninsulin dependent diabetes. Arteriosclerosis. 1988;8:168 77. 33. Bruce C, Chouinard RA, Tall AR. Plasma lipid transfer proteins, high-density lipoproteins, and reverse cholesterol transport. Annu Rev Nutr. 1998;18:297 330. 34. Dullaart RPF, Sluiter WJ, Dikkeschei LD, Hoogenberg K, Van Tol A. Effect of adiposity on plasma lipid transfer protein activities: a possible link between insulin resistance and high density lipoprotein metabolism. Eur J Clin Invest. 1994;24: 188 94. 35. Syvanne M, Ahola M, Lahdenpera S, et al. High density lipoprotein subfractions in non-insulin-dependent diabetes mellitus and coronary artery disease. J Lipid Res. 1995;36: 573 82. 70 OBESITY RESEARCH Vol. 8 No. 1 Jan. 2000