High-Density Lipoprotein Changes in Alzheimer’s Disease Are APOE Genotype-Specific

High-density lipoproteins (HDL) play a critical role in cholesterol homeostasis. Apolipoprotein E (APOE), particularly the E4 allele, is a significant risk factor for Alzheimer’s disease but is also a key HDL-associated protein involved in lipid transport in both the periphery and central nervous systems. The objective was to determine the influence of the APOE genotype on HDL function and size in the context of Alzheimer’s disease. HDL from 194 participants (non-demented controls, mild cognitive impairment, and Alzheimer’s disease dementia) were isolated from the plasma. The HDL cholesterol efflux capacity (CEC), lecithin-cholesterol acyltransferase (LCAT) activity, and particle diameter were measured. Neuropsychological test scores, clinical dementia rating, and magnetic resonance imaging scores were used to determine if cognition is associated with HDL function and size. HDL CEC and LCAT activity were reduced in APOE3E4 carriers compared to APOE3E3 carriers, regardless of diagnosis. In APOE3E3 carriers, CEC and LCAT activity were lower in patients. In APOE3E4 patients, the average particle size was lower. HDL LCAT activity and particle size were positively correlated with the neuropsychological scores and negatively correlated with the clinical dementia rating. We provide evidence for the first time of APOE genotype-specific alterations in HDL particles in Alzheimer’s disease and an association between HDL function, size, and cognitive function.


Introduction
Alzheimer's disease is the leading cause of dementia and poses a considerable economic and public health burden [1]. Alzheimer's disease develops over the course of decades [2], and drugs targeting late-stage processes once dementia and brain volume loss have set in may be "too late". There is an urgent need for effective, long-term treatments that target the underlying pathophysiology of Alzheimer's disease to prevent cognitive function loss and the onset of dementia.
Increasing evidence suggests that high-density lipoprotein (HDL) particles, both in the central nervous system (CNS) and in the periphery, are implicated in Alzheimer's disease pathology. Alzheimer's disease patients typically have lower plasma HDL cholesterol (HDL-C) and apolipoprotein-A1 (ApoA-I) concentrations than controls [3,4], suggesting that a lack of HDL is detrimental. HDL-C concentrations measured during midlife are negatively correlated with the onset of late-life mild cognitive impairment (MCI) and dementia [5], suggesting that HDL are protective. However, the functional capacity of HDL is a stronger predictor of disease risk than HDL-C concentrations, since the measurement of HDL-C does not encompass the complex, destructive changes to HDL particles that can occur in disease states and in aging [6,7]. The main function of HDL particles is to mediate reverse cholesterol transport, though they perform numerous additional protective functions, including reducing inflammation, promoting endothelial function, and antioxidant effects, among others [8].
Although ApoA-I is not expressed in the CNS, the peripheral overexpression of human ApoA-I preserves cognitive function, reduces neuroinflammation, and protects mice from cerebral amyloid angiopathy [9], suggesting a role for peripheral HDL in the clearance of brain amyloid beta (Aβ). In a small study of 39 Alzheimer's disease patients and 20 healthy participants (unknown APOE genotype), HDL isolated from plasma showed a diminished HDL cholesterol efflux capacity (CEC) and lecithin-cholesterol acyltransferase (LCAT) activity-the enzyme responsible for increasing the HDL cholesterol carrying capacity-and the level of LCAT activity was negatively correlated with the cognitive score [10]. Together, these studies suggest that circulating HDL is involved in the pathology of Alzheimer's disease. However, the relationship between the peripheral HDL functional capacity and Alzheimer's disease has not been fully investigated in a larger cohort of Alzheimer's disease patients, and the effects of the APOE genotype on this relationship have not been investigated.
In the current study, we determined whether HDL CEC and LCAT activity are altered in a large APOE genotyped cohort of elderly participants clinically diagnosed as either non-demented, MCI, or Alzheimer's disease dementia (AD). We also determined whether the HDL particle size, which is closely related to the HDL function, was altered in the same cohort. We further explored whether the HDL functional metrics and particle size were associated with the participants' cognitive, functional, and imaging scores.

Participants
This study used plasma samples collected from the University of California, Davis Alzheimer's Disease Research Center (ADRC) Biorepository. The ADRC biorepository is nationally recognized for recruiting an ethnically diverse clinic-based and communitybased elderly cohort [11]. We aimed to select 200 samples from the ADRC biorepository to include non-demented (controls), MCI, and AD patients, with each diagnosis group having as close as possible to equal numbers of participants with the APOE3E3 and APOE3E4 genotypes and with each genotype X diagnosis group having as close as possible to equal numbers of the sexes (males and females), with equal average ages. Only participants with an adequate sample volume available (500 µL) and body mass index (BMI) not greater than 40 kg/m 2 were included. Participants were diagnosed following the ADRC criteria for a clinical diagnosis within one year of a blood draw, as previously described [12]. Dementia patients were included in the AD group if their etiologic diagnosis was classified as "probable" or "possible" AD. Five patients in the AD group had an etiologic diagnosis of frontotemporal dementia. Removing these patients did not change the overall findings; thus, they were included in the analysis. Samples from a total of 194 participants met the above criteria and were included in the final analysis. The study was approved by the Institutional Review Board of the University of California, Davis.

Isolation Method
HDL particles from plasma were isolated by two-step sequential flotation densityultracentrifugation, followed by size exclusion chromatography, to yield highly purified HDL fractions, as previously described [13]. Following the manufacturer's instructions, the total HDL protein was measured using a Micro BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA, catalog 23235).

Cholesterol Efflux Capacity
HDL cholesterol efflux was measured using a commercially available kit (Abcam, Cambridge, UK, catalog ab19685) with modifications, as previously described [14]. J774A.1 (ATCC, Manassas, VA, USA, catalog TIB-67) were seeded at 100,000 cells in 96-well plates for 4 h in Dulbecco's Modified Eagle's Medium containing 10% fetal bovine serum and 100 µg/mL penicillin and streptomycin. The cells were washed and labeled with fluorescent cholesterol, acyl-coenzyme A:cholesterol acyltransferase inhibitor, and cyclic adenosine monophosphate for 4 h, followed by wash and incubation with 10 µg of HDL protein for another 4 h. The cellular supernatant was removed, and the remaining cells were lysed with M-PER cell lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA, catalog 78505). The supernatant and lysed fraction fluorescence were measured at 482/515 nm (emission/excitation) on a Synergy H1 plate reader (BioTek, Winooski, VT, USA). To account for the inter-plate variability, CEC values were normalized, with the CEC value of HDL isolated from pooled plasma collected from healthy volunteers on each plate to calculate a final CEC Index, as previously described [15].

LCAT Activity
A commercially available kit was used to measure the LCAT activity (Roar Biomedical, Millipore Sigma, Burlington, MA, USA, catalog mak107) in 5 µg of HDL (measured by the total protein) following the manufacturer's instructions. Measurements were collected on a Synergy H1 plate reader (BioTek, Winooski, VT, USA) and read at 340-nm excitation and two emission wavelengths at 390 nm and 470 nm, representing the hydrolyzed and intact substrate, respectively. Higher LCAT activity is indicated as increased λ em 390/λ em 470 nm ratios.

HDL Particle Size
The HDL particle size was assessed using negative-stained transmission electron microscopy (TEM) based on the published methods [16], in which over 3000 particles were imaged and sized from each individual participant sample. A schematic workflow of the particle size measurements by TEM is presented in Supplementary Figure S1. Briefly, isolated HDL samples were diluted to 30-100 µg/mL HDL protein concentrations with deionized water. A diluted HDL sample (4 µL) was loaded onto glow-discharged carboncoated grids (TedPella Inc., Redding, CA, USA) and left for sample attachment for 1 min. Extra sample was removed using filter paper. Uranyl formate solution (2%, pH 7.4) was then added onto the grid and removed 5 times to achieve negative staining. Negatively stained sample grids were then air-dried for 5 min and loaded onto a specimen holder for TEM, according to the manufacturer's user manual (JEOL USA 1230 Transmission Electron Microscope, JEOL USA Inc., Peabody, MA, USA). Samples were viewed under the TEM using high tension = 100 kv and 40,000× magnification. Micrograph images were taken at random locations using an attached CCD camera (model) with an exposure time of 300 ms.
The software ImageJ [17] was used to characterize the HDL particle size from the TEM micrographs, following a previously published procedure [13], with minor modifications: Noise in the micrographs was first removed using the "Bandpass Filter" function, with "filter large = 100, filter small = 10, suppress = None, and tolerance = 5 autoscale saturate" parameters. The contrast of the cleaned micrograph was then set to "min = 50, max = 205". The threshold of the micrographs was then set using a premade "Intermodes dark" option. The particle area was then analyzed using the "Analyze Particles" function with "size = 20-7850, circularity = 0.30-1.00, display, exclude, include, and add" parameters. Particles that were captured by the function were then outlined onto the original TEM micrograph and were checked manually for accuracy.

Cognitive Function Analysis
The Spanish English Neuropsychological Assessment Scales (SENAS) evaluated participants' neuropsychological functional assessments, as previously described and validated elsewhere [18,19]. This study uses the SENAS cognitive domains: verbal memory, executive function, spatial, and semantic memory.
The Clinical Dementia Rating Scale (CDR) is a semi-structured interview administered by a clinician as a global measure of independent function [20]. Six cognitive domains in memory, orientation, judgment and problem-solving, community affairs, home and hobbies, and personal care were assessed. The combination of the scores obtained ("sum of boxes") was used for analysis.
The white matter hyperintensities (WMH) measurements were acquired by magnetic resonance imaging (MRI), as previously described [21,22]. The total cranial volume was used to normalize the head size differences among participants.

Statistical Analysis
Data analyses were conducted using statistical software R version 4.1.1. (R Project for Statistical Computing, Vienna, Austria). Differences in the HDL functional metrics and particle sizes between AD patients, MCI patients, and controls were tested with one-way ANOVA. The null hypothesis of no difference between groups was rejected if p < 0.05. A post hoc Tukey's HSD was conducted for pairwise comparisons between groups when a significant difference was observed. For two group comparisons, a two-sample t-test assuming equal variance was used. For all pairwise comparisons, significance values were reported with p < 0.05. Statistically significant findings are indicated as shown: * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. The distribution of the outcome variables and key baseline variables was inspected for normality using the Shapiro-Wilks test. The homogeneity of the variance was examined using Levene's test. Kruskal-Wallis tests were performed on non-normally distributed data between more than two groups. A chi-squared (χ 2 ) test was used to compare the categorical variables.
To examine the association between HDL functional metrics and particle size with cognitive, functional, or MRI assessments, a correlation analysis, adjusting for the APOE genotype, was conducted across all diagnoses using Spearman's correlation. For Spearman's correlation, the rho (r) values and 95% confidence intervals were reported.
We performed complete case analyses to describe the participant characteristics, HDL functional metrics, and particle size and to assess the association between HDL function and size with cognitive, functional, and MRI assessments. For the HDL particle size analysis, one out of the total of 194 samples was removed due to an insufficient number of total particles on the TEM slide at the time of analysis.

HDL CEC Index and LCAT Activity Differences
The mean ± SD CEC index and LCAT activity are shown in Table 2. When participants were not stratified by the APOE genotype, there was no significant difference in the HDL CEC index between the control, MCI, and AD participants by one-way ANOVA (1.10 ± 0.16, 1.06 ± 0.14, and 1.12 ± 0.14, F(2, 191) = 2.14, p = 0.120, Figure 1A). A one-way ANOVA revealed that there was a significant difference in HDL LCAT activity between at least two groups (F(2, 191) = 3.87, p = 0.023). Post hoc comparisons indicated the HDL LCAT activity was significantly higher in the control group vs. MCI (1.05 ± 0.09 vs. 1.01 ± 0.07, p = 0.030) but not significantly higher than the AD group (1.03 ± 0.07, p = 0.118, Figure 1B). There were no significant differences in LCAT activity between MCI and AD (p = 0.661). The CEC index (1.06 ± 0.16 vs. 1.13 ± 0.14, t(192) = 3.00, p = 0.003) and LCAT activity (1.01 ± 0.07 vs. 1.05 ± 0.08, t(192) = 3.53, p < 0.001) were significantly lower in the APOE3E4 carriers relative to the APOE3E3 carriers ( Figure 1C,D).  When the participants were stratified by the APOE genotype, one-way ANOVA showed a significant difference in the HDL CEC index between at least two groups within the APOE3E3 carriers (F(2, 100) = 3.26, p = 0.042) and APOE3E4 carriers (F(2, 88) = 4.38, p = 0.015). In the APOE3E3 group, post hoc comparisons showed the HDL CEC index was lower in MCI patients relative to the controls (1.07 ± 0.13 vs. 1.16 ± 0.13, p = 0.042, Figure   Figure 1. APOE3E4 carriers have reduced HDL CEC index and LCAT activity. HDL (A) CEC index and (B) LCAT activity by diagnosis. One-way ANOVA followed by Tukey's multiple comparison tests was used. Decrease of (C) CEC index and (D) LCAT activity in the APOE3E4 genotype using two-sample t-tests. * p < 0.05, ** p < 0.01, and *** p < 0.001. Abbreviations: AD = Alzheimer's disease dementia, CEC = cholesterol efflux capacity, HDL = high-density lipoproteins, LCAT = lecithin-cholesterol acyltransferase, MCI = mild cognitive impairment. When the participants were stratified by the APOE genotype, one-way ANOVA showed a significant difference in the HDL CEC index between at least two groups within the APOE3E3 carriers (F(2, 100) = 3.26, p = 0.042) and APOE3E4 carriers (F(2, 88) = 4.38, p = 0.015). In the APOE3E3 group, post hoc comparisons showed the HDL CEC index was lower in MCI patients relative to the controls (1.07 ± 0.13 vs. 1.16 ± 0.13, p = 0.042, Figure 2A) but not in AD patients (1.11 ± 0.15) vs. the controls (p = 0.246). There was no significant difference in the HDL CEC index between MCI and AD patients (p = 0.551). For the APOE3E4 carriers, AD patients displayed a higher HDL CEC index relative to the controls (1.12 ± 0.14 vs. 1.02 ± 0.16, p = 0.016, Figure 2B) and a trend toward a higher efflux in AD vs. MCI patients (1.12 ± 0.14 vs. 1.03 ± 0.14, p = 0.114). There was no significant difference in HDL CEC index between the controls and MCI patients (p = 0.964).

HDL Particle Size Differences
The mean ± SD HDL particle size is shown in Table 2. Without stratification for the APOE genotype, one-way ANOVA revealed a significant difference in at least two groups (F(2, 190) = 9.70, p <0.001). Post hoc comparisons showed that the HDL particle sizes in the control group were significantly larger than the MCI (9.06 ± 0.69 nm vs. 8.67 ± 0.63 nm, p = 0.008) and the AD group (8.61 ± 0.68 nm, p < 0.001, Figure 3A). When stratified by the APOE genotype, there were no significant differences among the control, MCI, and AD groups in the APOE3E3 carriers by one-way ANOVA (8.99 ± 0.85 nm, 8.79 ± 0.58 nm, and 8.72 ± 0.64 nm, respectively, F(2, 99) = 1.40, p = 0.252, Figure 3B). In the APOE3E4 carriers, there was a significant difference in HDL particle sizes between at least two groups (F(2, 88) = 13.04, p < 0.001). Post hoc comparisons indicate that the HDL particle size in the control group was larger than the MCI (9.15 ± 0.42 nm vs. 8.53 ± 0.67 nm, p = 0.001) and the AD group (8.50 ± 0.70 nm, p < 0.001, Figure 3C). There was no significant difference in the HDL particle size between APOE3E3 and APOE3E4 carriers when the diagnosis was not taken into account (8.85 ± 0.73 nm vs. 8.78 ± 0.66 nm, t(191) = 0.64, p = 0.520).

Discussion
It is becoming increasingly apparent that disruption in the cholesterol metabolism can influence Alzheimer's disease pathology [23], and HDL particles have a critical role in maintaining the metabolic and homeostatic processes of regulating cellular cholesterol concentrations. HDL particles in the circulation have been recognized for their potential involvement in Alzheimer's disease, and the current evidence has linked peripheral HDL to cerebrovascular health and Alzheimer's disease [6]. However, HDL-C alone is a poor indicator of the disease risk [24], whereas the functionality of the HDL particles may be a better indicator of disease risk, with CEC emerging as useful in identifying individuals at risk for cardiovascular disease across multiple cohorts [25][26][27]. Here, our study revealed APOE genotype-specific alterations in peripheral HDL CEC and LCAT activity, two critical functional metrics in the reverse cholesterol transport pathway.
Limited studies have measured HDL CEC and, to a much lesser degree, LCAT activity in Alzheimer's disease patients. Khalil et al. [10] found impairment in the peripheral HDL CEC and LCAT activity in 39 Alzheimer's disease patients compared to 20 healthy controls, but the APOE genotype was not assessed. Both Yassine et al. [28] and Marchi et al. [29] reported a reduction in cerebrospinal fluid-mediated CEC in Alzheimer's disease patients. Such findings suggest a global decrease in the ability to efflux and remove excess cholesterol in Alzheimer's disease. However, in our study of APOE genotyped controls, as well as MCI and AD patients, we did not find statistically significant alterations in peripheral HDL CEC in the MCI or AD patients without stratifying for the APOE genotype. To our knowledge, we report for the first time that both the CEC and LCAT activity are overall lower in APOE3E4 carriers compared to APOE3E3 carriers regardless of the AD diagnosis, highlighting the importance of the APOE genotype in the ability to efflux and transport cholesterol. APOE3E3 MCI and AD patients exhibited lower LCAT activity, and APOE3E3 MCI patients had significantly lower HDL CEC than the controls, while the HDL particles of APOE3E3 AD patients exhibited a wider variability, and thus, the decrease in HDL CEC did not reach statistical significance.
On the other hand, paradoxically, APOE3E4 AD patients had higher HDL CEC than the controls. There are three plausible explanations for this observation. First, given the fact that APOE4 carriers are known to be at a higher risk for the earlier onset of dementia [30,31], it is possible that the APOE3E4 carriers have higher HDL CEC as a positive compensatory response to a poorer efflux capacity over the entire lifetime. Second, APOE3E4 AD patients had a lower prevalence of metabolic disease (diabetes or hypercholesterolemia), which is known to be linked with poorer HDL function [32,33]. Third, we observed that APOE3E4 AD patients have more small HDL particles, and because we standardized the quantity of HDL used in the efflux experiment based on the HDL protein content, it is likely that a higher number of total HDL particles was applied to the assay compared to controls. Thus, the observed increase in HDL CEC in APOE3E4 AD patients may simply be an observation of more total particles being able to efflux more total cholesterol rather than an actual higher cholesterol efflux capacity per particle. Most experiments measuring HDL CEC do so with apolipoprotein-B (ApoB) precipitated plasma and therefore apply the HDL dose as a percentage by volume in the cell media [25,34]. With this approach, there would also be a higher number of smaller particles applied to the assay if there were more total particles in the plasma. In our study, we isolated and purified HDL particles from plasma and applied equal amounts of HDL as the total protein. The disadvantage of measuring CEC in ApoB precipitated plasma is that this is not a direct measurement of the CEC of the HDL particles per se but of everything that remains in the plasma compartment after the ApoB-containing particles have been removed, thus the CEC measurement includes proteins and other components in plasma that could influence the ability to efflux cholesterol. The advantage of our approach is that we are measuring the CEC of the HDL particles themselves, rather than a combination of factors. One limitation of our approach for measuring the particle size by TEM is that this method does not capture all the particles in plasma and therefore does not enable quantification of the total HDL particle number. In the future, as methods become available to determine the particle count while simultaneously isolating HDL particles, the CEC assay can be dosed by equal particle number to compare the CEC on a per particle basis. Further studies are also needed to determine whether there are differences in the total particle numbers and particle size distributions in MCI and AD patients by the APOE genotype and to determine the HDL function and size in the other APOE genotypes (i.e., APOE3E2, APOE4E4, APOE4E2, and APOE2E2).
We did not find statistically significant differences in LCAT activity among the APOE3E4 carriers. We observed overall lower LCAT activity in APOE3E4 vs. APOE3E3 carriers; thus, the lack of difference in LCAT activity among the APOE3E4 carriers may be due to the fact that all individuals who are carriers of APOE4 may already have diminished LCAT activity compared to non-APOE4 carriers, and this reduction is not further enhanced in patients with dementia compared to the controls. Furthermore, the LCAT abundance in HDL was not measured. Thus, changes in the LCAT activity may be due to a lower abundance of LCAT protein in APOE4 carriers, which could reflect overall lower LCAT activity.
The HDL size reflects the remodeling stage of HDL and is highly associated with the HDL composition and function. For example, it has been documented that certain HDLassociated proteins are either exclusively associated with or enriched in certain size-based HDL subclasses [13,35]. Notably, ApoE tends to associate exclusively with larger HDL particles [13]. Smaller HDL particle sizes are also associated with metabolic syndrome and chronic kidney disease [36]. We found that the HDL particle size was significantly smaller in AD and MCI patients compared to the controls overall and in the APOE3E4 carriers but not in the APOE3E3 carriers. Future studies are needed to better understand the relationships between the overall metabolic status and HDL particle structure and function in the context of Alzheimer's disease and the APOE genotype.
Notably, we observed that HDL LCAT activity and particle size were positively correlated with the verbal memory score and negatively correlated with CDR, while the HDL particle size was additionally positively correlated with the executive function. Together, these findings suggest that disturbances in the HDL structure and function are associated with cognitive function and may be involved in Alzheimer's disease pathology. In elderly participants, verbal memory and expression strongly predict the progression from normal cognitive function to MCI before the appearance of clinical symptoms [37]. Subtle changes in the functional measurements in participants with higher CDR scores are associated with an increased risk of converting from normal cognition to MCI [38].
Although it is not yet clear whether and how peripheral HDL particles cross the bloodbrain barrier, studies have found high correlations between plasma and cerebrospinal fluid concentrations of HDL-associated apoproteins that are known not to be expressed in the CNS [39,40]. Additionally, the peripheral overexpression of human ApoA-I improved the cognitive function, reduced neuroinflammation, and protected mice from cerebral amyloid angiopathy [9]. Future studies are needed to understand the mechanisms by which peripheral HDL particles and/or their components affect the brain.
The study's strengths include the measure of two critical functions of HDL in the reverse cholesterol transport pathway and the measurement of the HDL particle size by TEM in HDL from plasma samples obtained from a large multi-ethnic cohort of APOEgenotyped, well-characterized, clinically diagnosed, and/or pathologically confirmed elderly participants from the UC Davis ADRC biorepository. To our knowledge, we provide evidence for the first time of an APOE genotype-dependent difference in HDL functional capacity, size, and dementia and also an association between HDL functional capacity, size, and cognitive function. However, the use of cross-sectional samples limits our ability to detect changes in HDL function and size across the continuum of Alzheimer's disease and changes in these parameters over time. Furthermore, in this study, HDL CEC was measured using the J774A.1 cell line. Future studies measuring the CEC using neuronal cells, such as neuroblastoma (SK-N-SK) cells, microglia, and other brain-relevant cell types, are needed. Another limitation is that, in this study, information on whether the participants were consuming antidiabetic agents or statins, which are commonly prescribed in AD patients and which may influence the HDL functional capacity and size, was not available, although the effects of statins on CEC remain uncertain, as reviewed elsewhere [41]. Future studies are needed to further investigate the underlying compositional and structural differences that explain the observed differences in the HDL functional capacity, particularly with regards to the APOE genotype. Many modifications of HDL particles, including oxidation, glycation, loss of functional components, and gain of deleterious components, have been found to influence the HDL functional capacity, even beyond the ability to efflux cholesterol [42][43][44].

Conclusions
Our findings further support earlier observations that HDL particles are implicated in Alzheimer's disease pathology and highlight, for the first time, that there is an APOE genotype-dependent relationship that merits further study. Notably, our results suggest that the mechanisms of HDL deficiency in APOE4 carriers vs. non-carriers are different, highlighting the need for further research on HDL metabolism and function in APOE genotyped individuals to elucidate the potential precision medicine-based approaches to improve the HDL functionality in individuals at risk for Alzheimer's disease tailored to their APOE genotype. It will be important to design future studies to determine which HDL compositional and structural changes underlie the loss of function that contributes to Alzheimer's disease pathology and how APOE genotype shapes these processes, so that potential therapeutic strategies to improve HDL functionality can be tested for their effectiveness in the prevention or treatment of cognitive decline.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/biomedicines10071495/s1: Figure S1: Schematic working flow for particle selection and size analysis on sample TEM images. Figure S2: Spearman's correlation between HDL CEC and LCAT activity. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data used in the study are available from the corresponding author upon request.