ApoE and apoC-III-defined HDL subtypes: A descriptive study of their LCAT and CETP content and activity

High-density lipoproteins (HDL) in plasma are strongly and negatively associated with cardiovascular risk, yet interventions to raise HDL have not improved cardiovascular outcomes. HDL functionality 32 and heterogeneity may hold the clue to this paradox. The apolipoprotein composition of HDL may be an important determinant of their functionality. Lecithin-cholesterol acyl transferase (LCAT) and cholesterol-ester transfer protein (CETP) are key enzymes for HDL-mediated reverse cholesterol transport. We assessed the distribution and activity of LCAT and CETP in HDL subspecies defined by 36 their content of apolipoproteins E (apoE) and C-III (apoC-III) in humans. We isolated in adult humans of both sexes (mean age 55.6, BMI 26.9 Kg/m 2 , HbA1c 5.4%), four subspecies of HDL containing respectively: No apoE and no apoC-III (E-C-), apoE but not apoC-III 40 (E+C-), apoC-III but no apoE (E-C+) and both apoE and apoC-III (E+C+). In each HDL subspecies, 41 we measured LCAT and CETP concentration and activity using immunoenzymatic and fluorometric methods. Additionally, we determined the size distribution of HDL in each apolipoprotein-defined fraction using non-denaturing electrophoresis and anti-ApoA-I western blot.


Determination of enzymatic content and activity in HDL subfractions 131
LCAT concentration was determined using an immunoenzymatic, double-sandwich assay (ALPCO 132 Diagnostics, Salem, NH) in which capture is performed by a first monoclonal antibody against LCAT 133 Determination of HDL size distribution 163 HDL from the four apoC-III and apoE-defined immunofractions were separated by size using non-164 denaturing polyacrylamide gradient gel electrophoresis (NDPAGGE). Twenty-five microliters of each 165 immunofraction plus 25 microliters of sample buffer were loaded into a 4-30% polyacrylamide 166 gradient gel (Jule Inc., Milford, CT). Wells 1 and 10 were loaded with molecular size standards 167 (Amersham HMW Native Marker Kit, GE Healthcare, Little Chalfont, UK) and gels were run for 24 168 hours at constant 70V. Then, contents of the gel were transferred to a 0.45 micrometer pore size 169 polyvinylidene fluoride (PVDF) membrane (Pall Corporation) in a wet transfer apparatus at 30V for 170 24 hours. The lanes containing the MW markers were cut from the rest of the membrane, stained in 171 0.2% amido black solution for 20 minutes and stored for later photographing. The rest of the membrane 172 was blocked with 5% powder low-fat milk, incubated with an HRP-conjugated goat anti-human apoA-173 I antibody (Academy Biomedical, Houston, TX), and revealed using 3,3′,5,5′-tetramethylbenzidine as 174 substrate. Later, the marker lanes and the rest of the membrane were placed side to side and 175 photographed in a Bio-Rad ChemiDoc TM MP gel documenter. Using the molecular size standards as 176 reference, the intensity of the bands/smears in each size range fractions was quantitated in Image Lab TM 177 software. Size fractions were defined as follows: prebeta HDL: <7.1 nm, alpha 3 HDL: 7.1-8.2 nm, 178 alpha 2 HDL: 8.2-9.5 nm and alpha 1 HDL: 9.5-12.2 nm. The concentration of apoA-I in each HDL 179 size subfraction was estimated by multiplying the proportion of apoA-I within that fraction by the 180 directly measured total plasma apoA-I. Laboratory procedures were executed at the Diabetes, Lipids 181 and Metabolism laboratory of Universidad de Los Andes, following current institutional biosafety 182 protocols. 183

Statistical analyses 185
The distribution of plasma apoA-I across HDL subfractions was compared in a 2-way ANOVA model 186 in which apoA-I concentration in each subfraction was the dependent variable, while HDL size and 187 immunofraction were fixed factors. Enzyme masses and activity across were compared using a 1-way 188 ANOVA with HDL immunofraction as fixed factor. When global ANOVA was significant, post-hoc 189 comparisons against the reference fraction (E-C-) were done using Scheffé´s method. Comparisons of 190 quantitative variables rates between groups were done using the Mann-Whitney´s U test. Differences 191 in categorical variables between groups were performed with a Fisher exact test. All tests were done at 192 a 0.05 significance level and all reported p-values are 2-sided. 193 194 195 The study included 8 male and 10 female adults, aged 55.6 +/-11.2 years, body-mass index 26.9 +/-196 4.0 kg/m 2 . Women had higher plasma total cholesterol, HDLc and LDLc. Women also exhibited higher 197 plasma apoA-I (121.4 +/-14.3 mg/dL versus 97.7 +/-23.0 mg/dL in men, p= 0.016). Total plasma 198 LCAT was 6.3 +/-2.2 mg/dL in women and 5.7 +/-2.2 mg/dL in men, while plasma CETP 199 concentration was 1.9 +/-0.7 mg/dL in women and 1.7 +/-0.7 mg/dL in men. Relative plasma CETP 200 activity was very similar in both sexes, 98.7 +/-13.7 AU in women and 101.7 +/-6.1 AU in men. 201

Results
Meanwhile, LCAT activity was 96.1 +/-22.3 AU in women and 96.8 +/-18.4 AU in men (Table 1) ApoC-III containing HDL displayed a higher triglyceride-to-apoA1 molar ratio, while E-C-HDL were 234 poorer in triglycerides relative to the other fractions (p=0.01 for global ANOVA, Table 2). Despite 235 numerically higher phosphatidylcholine-to-apoA1 molar ratios in the apoE or apoC-III containing 236 HDL, this difference did not achieve statistical significance. The cholesterol and cholesteryl ester 237 content of the 4 HDL subtypes were numerically similar, and not statistically different. 238

239
The concentration of some apolipoprotein and size-defined HDL correlates with plasma LCAT and 240

CETP activity 241
The concentration of apoA-I in the E-C+ pre-beta HDL subfraction correlated inversely with total 242 plasma LCAT activity (r=-0.55, p=0.017, Table 3). Similarly, apoA-I in the E-C-alpha 1 HDL 243 subfraction correlated inversely with total plasma LCAT activity (r=-0.49, p=0.041). On the other 244 hand, larger E-C+ HDL (alpha 1) exhibited a positive correlation with plasma CETP activity (r=0.52, 245 p=0.025). The correlation was not significant for smaller HDL of this type, but tended to increase along 246 with HDL size for this fraction. In this study of apolipoprotein-defined human HDL types, we found a different LCAT mass with 253 comparable activity across HDL subfractions. In other words, we encountered a greater ratio of LCAT 254 activity to mass, signaling a greater degree of LCAT activation, in HDL containing apoE . CETP mass 255 and CETP activity on the other hand were not associated with the apoE or apoC-III composition of 256 HDL. HDL containing apoE and/or apoC-III were a minority of total plasma HDL, as reported 257 previously (37), yet they proved to be different from E-C-HDL in their relative enzymatic activity. Concerning the lipid composition of HDL subfractions, we found a numerically higher proportion of 280 phosphatidylcholine in HDL with either apoE or apoC-III. Thus, higher LCAT activity in these HDL 281 subtypes may be due to a larger availability of enzyme substrate. This explanation has biologically 282 plausibility, as the exchange of apolipoproteins between lipoproteins facilitates the simultaneous 283 transfer of some of the surrounding phospholipid-rich membrane by phospholipid transfer protein 284 (PLTP) (46). We did not find statistically significant differences in cholesteryl ester content among the 285 four HDL subtypes. 286

287
We found a significant inverse correlation between total plasma LCAT activity and the presence of 288 pre-beta HDL without apoE and with apo-CIII (E-C+) and alpha 1 HDL without apoE or apoC-III (E-289 C-). The former result suggests that the influence of apoC-III on LCAT activity in this particular HDL 290 size may correlate with less maturation and the accumulation of prebeta HDL (47). Interestingly, a 291 study comparing patients with and without CVD found an inverse correlation between LCAT activity 292 and plasma prebeta HDL concentrations, which was stronger in the CVD group (48). The negative 293 correlation between E-C-alpha 1 HDL and plasma LCAT activity suggests that not all size 294 conversion/maturation of HDL requires LCAT activity. Meanwhile, the positive correlation of plasma 295 CETP activity with the abundance of large E-C+ HDL may just reflect the frequent coexistence of both 296 proteins in this HDL type. 297

298
The main limitations of our study are its limited sample size and the fact that we interrogated only two 299 specific apolipoproteins, out of the many known to be present in HDL. However, this was a proof-of-300 concept study in which the core analyses were performed within-individual, so that enzymatic contents 301 and activities were performed for different HDL subtypes belonging to the same individual. The 302 biological antagonism between apoE and apoC-III (49) has also been proven to modulate HDL 303 metabolism and its association with cardiovascular risk (50), making them particularly worthy of 304 investigation. Also, it is plausible that the employed LCAT activity assay monitors only the fatty acid 305 removal from PC and not its successful transfer to free cholesterol. Nonetheless, given the known 306 nature of the LCAT enzymatic process, its ability to perform this initial cleavage has been widely 307 employed as a proxy for LCAT activity in vitro. 308

309
In conclusion, our results showed specific profiles of LCAT mass and activity in HDL subtypes defined 310 by their content of apoE and apoC-III, and suggest that the presence of apoE protein in HDL may be a 311 correlate of increased LCAT activity. Further studies describing the changes in HDL composition, 312 functionality and associated enzymatic activity in the context of CVD would expand on our findings. 313 Availability of data and materials 342

List of Abbreviations
The datasets used and/or analysed during the current study are available from the corresponding 343 author on reasonable request. 344 345 Competing interests 346 The authors declare that they have no competing interests 347 348 Funding 349 This study was funded by Vicerrectoría de Investigaciones, Universidad de los Andes, Colombia 350