Beyond Lipidation: CSF APOE4 Protein Burden, Not HDL Subclass, Drives Tau Associations in APOE4 Alzheimer’s Disease

Beyond Lipidation: CSF APOE4 Protein Burden, Not HDL Subclass, Drives Tau Associations in APOE4 Alzheimer’s Disease | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Beyond Lipidation: CSF APOE4 Protein Burden, Not HDL Subclass, Drives Tau Associations in APOE4 Alzheimer’s Disease Zoe E. Tsokolas, Isaac Asante, Amaryllis A. Tsiknia, Juan Pablo Barbosa-Carvajal, and 9 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9349576/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Apolipoprotein E ε4 (APOE4) is the strongest genetic risk factor for late-onset Alzheimer's disease (AD), yet whether its pathogenic effects are driven by HDL lipidation state or by elevated APOE4 levels in the CNS remains unclear. Using ADNI data, we analyzed cerebrospinal fluid (CSF) and plasma small and large HDL particle concentrations, total APOE levels, and isoform-specific APOE3 and APOE4 protein levels in 144 participants with APOE ε3/ε3, ε3/ε4, or ε4/ε4 genotypes, grouped by cognitive state and memory progression over 4 years. Memory loss was defined as ≥ 10% decline on the Rey Auditory Verbal Learning Test (RAVLT) delayed recall between baseline and 48-month follow-up. Cross-sectional associations with CSF Aβ1–42, total tau, and p-tau181 were evaluated using covariate-adjusted linear regression, and longitudinal trajectories were examined using linear mixed-effects models over 6 years. CSF small and large HDL levels were higher in cognitively normal individuals and non-progressors, while APOE4 carriers exhibited reduced CSF small HDL relative to ε3 homozygotes. Importantly, APOE4 status moderated HDL–biomarker associations in opposing directions: in carriers, higher CSF small HDL was associated with lower tau and Aβ1–42, whereas in non-carriers, higher CSF small HDL was associated with higher CSF tau levels. Higher CSF APOE4 protein levels were associated with elevated tau and p-tau181, while plasma APOE measures showed minimal and often opposing associations with CSF biomarkers. Longitudinally, higher baseline CSF APOE4 proportion was associated with greater memory decline over 6 years but did not predict biomarker change. These findings argue against a protective role of APOE4 lipidation in AD and instead support a CNS-compartment model in which APOE4 protein burden aligns more closely with tau pathology and cognitive decline than peripheral APOE or HDL measures. APOE isoforms brain-peripheral APOE4 HDL Alzheimer’s Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Alzheimer's disease (AD) is characterized by altered lipid metabolism in which Apolipoprotein E (APOE) isoform-specific differences are thought to impair Aβ clearance. APOE is the major lipid transport protein in the brain, responsible for maintaining neuronal and vascular health through cholesterol delivery and membrane integrity ( 1 ). The APOE-ε4 allele confers an altered protein conformation that reduces lipid transport efficiency and disrupts lipid homeostasis, heightening neuronal vulnerability and promoting both vascular risk and neuro-inflammation ( 1 ). Carriers of the APOE-ε4 allele additionally show increased amyloid deposition, tau phosphorylation, and accelerated cognitive decline relative to non-carriers ( 2 , 3 ). Despite this strong genetic association, the mechanism by which APOE4 confers AD risk in humans remains unresolved. In particular, it is unclear whether APOE4-associated pathology is driven by altered lipoprotein state or by increased APOE4 protein levels within the central nervous system (CNS). Understanding how APOE4 and its associated HDL particles contribute to AD requires separate evaluation of central and peripheral compartments, as APOE in plasma and cerebrospinal fluid (CSF) may differentially influence lipid homeostasis and Aβ metabolism. Because the blood–brain and blood–CSF barriers restrict lipoprotein exchange, APOE exists in biologically distinct pools, with CSF APOE synthesized largely within the CNS rather than derived from peripheral sources ( 4 , 5 ). In the CSF, APOE is carried on HDL-like particles that support ABCA1-mediated lipid efflux, cholesterol transport, and Aβ trafficking, functions critical for CNS homeostasis. In plasma, HDL mediates reverse cholesterol transport from peripheral tissues to the liver ( 6 ). These HDL particles are heterogeneous, comprising small and large subspecies that differ in composition, function, and tissue distribution, and whose relative abundance in CSF may influence both local and systemic lipid homeostasis ( 7 ). Whether peripheral HDL alterations reflect central APOE biology or whether CNS APOE operates independently of plasma HDL dynamics remains uncertain ( 8 ), and it is unclear whether APOE4-associated risk is mediated primarily through central (CSF) or peripheral (plasma) pathways. Recent studies suggest that APOE4 carriers exhibit altered HDL particle composition and poor lipidation, leading to the hypothesis that reduced APOE lipidation contributes to toxicity ( 9 ). Independent studies have shown that small HDL particles in the CSF, but not HDL measures in plasma, correlate with AD-related biomarkers and cognitive performance ( 7 ). These observations have motivated therapeutic strategies aimed at enhancing APOE lipidation or increasing HDL-associated APOE. Yet direct human evidence testing whether higher HDL subclass abundance mitigates APOE4-associated biomarker abnormalities is lacking. Moreover, few studies have directly quantified isoform-specific APOE protein levels in both CSF and plasma to determine whether APOE4 protein burden itself, rather than HDL subclass distribution, aligns more closely with AD pathology. We fill this gap in the literature by quantifying isoform-specific APOE concentrations in addition to CSF HDL subclasses to examine whether these differentially associate with AD biomarker levels as a function of APOE4 status. In the present study, we address these gaps by integrating CSF and plasma HDL particle measurements with total and isoform-specific APOE protein quantification in a well-characterized cohort of older adults with normal cognition or mild cognitive impairment from the Alzheimer's Disease Neuroimaging Initiative (ADNI). We evaluate cross-sectional associations between CSF HDL subclasses, isoform-specific APOE concentrations, and established AD biomarkers (Aβ1–42, total tau, p-tau181) measured in CSF. We further examine whether quantitative CNS APOE4 burden predicts longitudinal cognitive decline over six years independent of genotype status. By directly comparing central and peripheral APOE measures and testing whether HDL subclass abundance modifies APOE4-associated biomarker relationships, this study tests whether APOE4-associated pathology is better explained by a lipidation-deficiency model or by a CNS APOE4 protein burden model. Methods Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer’s disease (AD). Human Ethics and Consent to Participate Declaration ADNI adheres to ethical standards that include the Declaration of Helsinki, the International Conference on Harmonization Good Clinical Practice (ICH GCP), and the Code of Federal Regulations (CFR). ADNI also obtains informed consent from participants and follows the ethical standards of the institutional review board (IRB). Participants The present study analyzed data from 144 participants (mean age 73.0 ± 6.6) from the ADNI dataset (Table 1 ). Participants with complete baseline measurements of CSF and plasma small and large HDL particle levels, CSF biomarker levels (tau, p-tau181, Aβ1–42), complete measurements of CSF and plasma total APOE, and with available Rey Auditory Verbal Learning Test (RAVLT) scores were included in our study (n = 144), with exclusions for ε2/ε3 genotype (n = 11). Measures of CSF and plasma APOE3 and APOE4 protein levels were available for 85 of the 144 participants in our sample. Individuals were stratified into four clinical groups based on delayed recall test performance (RAVLT) following methods previously described ( 10 ): ( 1 ) cognitively normal (CN) stable (n = 27), with no cognitive impairment at baseline assessment and 10% RAVLT decline between baseline assessment and 48-months follow-up; ( 3 ) Late Mild Cognitive Impairment (LMCI) stable (n = 37), with participants having mild cognitive impairment (MCI) at baseline but no AD progression between baseline and 48-months of follow-up; lastly ( 4 ), LMCI decline (n = 42) with participants having MCI at baseline and AD progression between baseline and 48-months of follow-up. Study data Biomarker data were available from CSF samples collected during annual visits spanning a 6-year period. The concentrations of Aβ1–42, t-tau, and p-tau181 in the CSF were measured using the Luminex platform INNOBIA AlzBio3 RUO test, using a microbead-based multiplex immunoassay as previously described ( 11 ). The CSF Aβ1–42 assay had a range of 200 to 1700 pg/mL (Elecsys CSF Aβ42 immunoassay) and samples with Aβ42 levels that exceeded this range were assigned the maximum value of 1700 pg/mL (n = 27) ( 12 ). For cognitive measures, the RAVLT accuracy score of delayed memory recall by % forgotten on a list-recall paradigm was used ( 13 ). RAVLT scores, used for clinical group categorization, were used in cognitive analyses and were available from each annual visit over the 6-year period. Total APOE Measurement Cerebrospinal fluid (CSF) and plasma levels of APOE were determined using a Meso Scale Discovery (MSD) R-PLEX human APOE singleplex assay (Catalog number K151AMLR) following the manufacturer’s standard recommendations for assay protocol 1. Briefly, 25 µl per well of biotinylated capture antibody diluted in coating diluent 100 (Catalog number R50AA, MSD) was added to each well of MSD small spot streptavidin plates, and adhesive sealed plates were incubated for 1 hour with shaking at 700 rpm at 25 ºC on an orbital shaker (Jitterbug-2, Boekel Scientific, Feasterville, PA, USA). The plates were washed three times with 150 µL/well of wash buffer (0.01 M PBS pH 7.4 with 0.05% Tween-20) and tapped on absorbent Kimwipes (Catalog number 34133, Kimberly-Clark Global Sales, LLC, Roswell, GA, USA). MSD supplied human plasma-purified APOE calibrator standard concentrations (750, 187.50, 46.88, 11.72, 2.93, 0.73, 0.18 ng/mL) were prepared in the assay diluent 57 (Catalog number R50BZ, MSD). The assay diluent was used as the zero-calibrator standard. Human EDTA plasma samples were diluted 100-fold, and CSF samples were diluted 50-fold in the assay diluent 57 in protein LoBind tubes (Eppendorf AG, Hamburg, Germany). Twenty-five microliters of diluent 57 was added to each well of the coated plates prior to the addition of 25 µL of prepared calibrator standards or samples in designated wells. The plates were sealed and incubated for 1 hour with shaking at 700 rpm at 25º C on an orbital shaker. The plates were washed three times with 150 µL/well of wash buffer and tapped on absorbent Kimwipes to remove residual wash buffer. The supplied SULFO-TAGTM-conjugated 100X detection antibody was diluted in diluent 3 (Catalog number R50AP-2, MSD). Fifty microliters of the detection antibody solution were added to each well, and the sealed plates were incubated for 1 hour with shaking at 700 rpm at 25 ºC on an orbital shaker. The plates were washed three times with 150 µL/well of wash buffer and tapped on absorbent Kimwipes to remove residual wash buffer. MSD read buffer T (Catalog number R92TC-2, MSD) was diluted 2-fold in UltraPure™ distilled water (Catalog number 10977-015, Invitrogen, Grand Island, NY, USA), and 150 µL of read buffer was added to each well. The plates were read immediately on an MESO Quickplex® SQ120MM instrument with electrochemiluminescence detection. Calibrator standard curves were used to calculate plasma and CSF APOE concentrations, after correcting for the sample dilution factor. APOE3 and APOE4 Measurement in CSF and Plasma CSF and plasma samples from baseline assessments were analyzed using an APOE mass spectrometric immunoassay (MSIA) as described previously ( 10 ). The MSIA detects 3 APOE isoforms at the intact protein level, with apoE affinity-captured from CSF and plasma using MSIA tips derivatized with APOE-specific antibodies. A MALDI-TOF mass spectrometer (Autoflex III MALDI-TOF, Bruker, Billerica, MA) was used to analyzed eluted and intact APOE in positive ion mode, with a mass range of 7–70 kDa, 700 ns delay, 20.00 kV and 18.45 kV ion source voltages, and < 7000 Da signal suppression. Mass spectra were baseline subtracted (Convex Hull algorithm with 0.8 flatness) and smoothed (Savizky Golay algorithm with 5 m/z width and 1 cycle) using the Bruker Flex Analysis software. Zebra 1.0 software (Intrinsic Bioprobes Inc.) was used to quantify and tabulate the peak intensities of APOE isoforms. A variable representative of relative abundance of ApoE4 to ApoE3 in heterozygotes (ε4/ε3-ratio) was calculated by dividing CSF ApoE4 isoform level by CSF ApoE3 isoform level. An additional variable (percent ε4) was derived by multiplying total CSF ApoE4 isoform proportion by the CSF total ApoE concentration, representative of absolute concentration of ApoE4 protein in CSF (µg/mL). Small and Large HDL Measurement Concentrations of HDL particles were measured by ion mobility (IM) after treatment with dextran sulfate to remove non-lipid-bound proteins such as albumin from 30 µL of plasma or CSF as described previously ( 14 ). Additional details can be found in the Supplementary Materials ( 7 ). Particles in the HDL size range were classified as small (7.0-10.5 nm) and large (10.5–14.5 nm) HDL. The majority of HDL particles in both CSF and plasma were found in the small size range, with a greater proportion of small to large HDLs present in CSF compared to plasma ( 7 ). Statistical Analysis In cross-sectional analyses, linear regression models adjusting for age, sex, education, ε4 positivity, and clinical group were used to examine the associations between baseline independent variables (CSF and plasma small and large HDL/LDL particles, percent ε4, and ε4/ε3-ratio) and AD biomarker outcomes. Pearson’s correlations were used to examine the linear association of these independent variables measured in plasma vs CSF. To evaluate between-group differences on CSF HDL levels and APOE concentrations, we utilized one-way analysis of variance (ANOVA) tests, followed by post-hoc pairwise comparisons using Tukey’s honestly significant difference (HSD) test to control for multiple comparisons. Group comparisons were made on APOE genotype groups (using ANOVA) as well as the two cognitive progression groups (CN vs MCI; Non-progressors vs Progressors) and CN vs MCI (using independent two-sample t-test). Associations between baseline measures of small and large HDL particles and percent ε4 with longitudinally-assessed CSF tau, p-tau181, Aβ1–42 and cognitive performance annually over a 6-year period were tested using linear mixed effects models. HDL particle levels and fluid biomarker measures were log-transformed (log10(x + 1)) to reduce skew and improve normality, and all continuous variables were z-scored with a mean of 0 and SD of 1 to derive standardized parameter estimates. Fixed effects of interest were the baseline HDL and ε4 measures; adjusting covariates were age, sex, education, clinical group, and ε4 positivity (in models that did not include ε4 ratio or amount as an independent variable). A random participant-level intercept was specified to account for the within-participant correlation. We tested for an interaction between time modeled as a continuous variable (calculated as years since baseline assessment) and each of our independent variables to test for associations of baseline measures with longitudinal trajectories of change in outcome variables. Time was modeled as a fixed effect as well as a random participant-specific slope to account for longitudinal individual variation, and an unstructured variance-covariance matrix was specified for random effects. Likelihood ratio tests were conducted to compare models both with and without the addition of random slope to evaluate whether its inclusion was justified. We corrected for multiple comparisons across AD biomarker outcomes (Aβ1–42, tau, p-tau181) using the false discovery rate (FDR) method. All analyses used RStudio (Version 2024.12.1 + 563); the ‘lme4’ package was used for linear mixed effect modeling. Results HDL Subclasses and APOE Levels on Clinical Outcomes CSF small HDL levels were significantly lower in ε3/ε4 genotypes (β = -0.06; p = 0.017) and in ε4/ε4 genotypes (β = -0.06; p = 0.033) compared to ε3/ε3 genotypes (Fig. 1 a). No significant difference was observed between ε3/ε4 and ε4/ε4 genotypes (β = -0.01; p = 0.974). CN individuals had significantly higher levels of CSF small HDL compared to those with MCI (β = 0.05; p = 0.006) (Fig. 1 b), as did cognitive non-progressors compared to progressors (β = 0.05; p = 0.010) (Fig. 1 c). CSF large HDL levels did not significantly differ across APOE genotypes (ε3/ε3 vs ε3/ε4: β = -0.03, p = 0.497; ε3/ε3 vs ε4/ε4: β = -0.05, p = 0.348; ε3/ε4 vs ε4/ε4: β = -0.01, p = 0.907) (Fig. 2 a). CSF large HDL did significantly differ across cognitive diagnoses and progressor status, with higher levels of large HDL in CN (β = 0.08; p = 0.001) compared to MCI (Fig. 2 b), and higher levels in non-progressors compared to progressors (β = 0.11 p < 0.001) (Fig. 2 c). CSF total APOE levels significantly differed by genotype, with lower CSF total APOE levels in ε3/ε4 (β = -0.69; p < 0.001) and ε4/ε4 genotypes (β = -0.80; p = 0.001), compared to ε3/ε3 genotypes (Fig. 3 a). Total APOE levels in CSF did not differ significantly between cognitive diagnoses (β = 0.06; p = 0.730) (Fig. 3 b). Total APOE levels in CSF also did not significantly differ by progressor status in CSF (β = 0.11; p = 0.541) (Fig. 3 c). CSF APOE-ε3 levels were significantly higher in ε3/ε3 genotypes than ε3/ε4 (β = 2.53; p < 0.001) (Fig. 4 a), but APOE-ε3 levels did not significantly differ across cognitive diagnoses (β = -0.03; p = 0.912) (Fig. 4 b) or progressor status groups (β = 0.20; p = 0.463) (Fig. 4 c). CSF APOE-ε4 levels were significantly higher in ε4/ε4 genotypes than ε3/ε4 (β = 0.79; p < 0.001) (Fig. 4 a), but APOE-ε4 levels did not significantly differ across cognitive diagnoses (β = -0.27; p = 0.125) (Fig. 4 b) or progressor status groups (β = -0.33; p = 0.073) (Fig. 4 c). In ε3/ε4 heterozygous carriers, CSF APOE-ε4 levels were significantly higher than CSF APOE-ε3 levels (β = 0.89; p < 0.001) (Fig. 4 a). APOE Isoform Levels, APOE-ε4/ε3 ratio, and APOE-ε4% in CSF There was a trend-level interaction between CSF APOE-ε3 levels and CSF APOE-ε4 levels with baseline CSF Aβ1–42 (β = -0.87; p = 0.076; FDR-p = 0.095) (Fig. 4 d) and CSF p-tau181 (β = -0.75; p = 0.095; FDR-p = 0.095) (Fig. 4 e); however, associations were not sustained following FDR correction. There was a significant interaction between CSF APOE-ε3 levels and CSF APOE-ε4 levels with baseline CSF tau (β = -0.97; p = 0.022; FDR-p = 0.066) that became trend-level following FDR correction (Fig. 4 f). The CSF ε4/ε3 ratio was significantly higher in those with MCI compared to CN (β = 0.16; p = 0.004) (Fig. 5 a) and did not significantly differ between cognitive progressors and non-progressors (β = -0.07; p = 0.256) (Fig. 5 b). Cross-sectional associations between the CSF ε4/ε3 ratio and CSF Aβ1–42 (β = -0.05; p = 0.742), CSF p-tau181 (β = 0.20; p = 0.248), and CSF tau (β = 0.17; p = 0.327) were not significant (Fig. 5 d-f). Clinical Outcome Differences Across APOE Levels and Ratios Total plasma APOE levels (in µg/mL) differed by genotype, with lower plasma total APOE in ε3/ε4 genotypes (β = -7.59; p < 0.001) and ε4/ε4 genotypes (β = -7.16; p < 0.001), relative to ε3/ε3 genotypes ( Supplementary Fig. 1a ). Total plasma APOE levels did not differ significantly by cognitive diagnoses (β = -1.16; p = 0.343) ( Supplementary Fig. 1b ) or progressor status in plasma (β = -0.53; p = 0.653) ( Supplementary Fig. 1c ). Plasma APOE-ε4 levels were significantly higher in homozygous carriers compared to heterozygous carriers (β = 1.47; p = 0.009) ( Supplementary Fig. 2a ) and in MCI compared to CN (β = 1.04; p = 0.023) ( Supplementary Fig. 2b ), but did not significantly differ by progressor groups (β = 0.07; p = 0.890) ( Supplementary Fig. 2c ). Cross-Sectional Associations: CSF HDLs and CSF APOE levels with CSF Biomarkers APOE-ε4 carrier status significantly moderated the relationship between CSF small HDL and CSF Aβ1–42 (interaction β = -0.41; p = 0.025; FDR-p = 0.038), such that small HDL was significantly associated with lower CSF Aβ1–42 in APOE-ε4 carriers (β = -0.45; p = 0.001) and not in non-carriers (β = -0.04; p = 0.687) (Fig. 1 d) (Table 2 ). APOE-ε4 carrier status also moderated the relationship between CSF small HDL and CSF p-tau181 at a trend-level (interaction β = -0.40; p = 0.054; FDR-p = 0.054), such that small HDL was at a trend-level associated with higher CSF p-tau181 among non-carriers (β = 0.21; p = 0.076), whereas no significant association was observed among carriers (β = -0.19; p = 0.229) (Table 2 ). APOE-ε4 carrier status furthermore significantly moderated the relationship between CSF small HDL on CSF tau (β = -0.52, p = 0.015; FDR-p = 0.038), with higher levels of small HDL trend-level associated with higher levels of CSF tau in non-carriers (β = 0.23; p = 0.058) but oppositely associated with lower levels of CSF tau in carriers (β = -0.29; p = 0.073) (Table 2 ). APOE-ε4 significantly moderated the association between CSF large HDL and CSF p-tau181 (β = -0.40; p = 0.024; FDR-p = 0.036), with a non-significant positive trend in APOE-ε4 non carriers (β = 0.19, p = 0.110) and a non-significant negative trend in carriers (β = -0.20, p = 0.143) (Table 2 ). APOE-ε4 significantly moderated the association between CSF large HDL and CSF tau (β = -0.52; p = 0.003; FDR-p = 0.009), such that in APOE-ε4 non-carriers, higher CSF large HDL was significantly associated with higher CSF tau (β = 0.24, p = 0.049) but with lower CSF tau in APOE-ε4 carriers (β = -0.28, p = 0.046) (Table 2 ). We observed a trend-level interaction between APOE-ε4 x CSF large HDL with CSF Aβ1–42 (β = -0.28; p = 0.071; FDR-p = 0.071), with APOE-ε4 non-carriers having no significant association with CSF Aβ1–42 (β = 0.06, p = 0.576) and APOE-ε4 carriers having a trend-level association between higher CSF large HDL with lower CSF Aβ1–42 (β = -0.22, p = 0.075) (Table 2 ). In the whole sample, CSF total APOE was cross-sectionally associated with significantly higher levels of CSF Aβ1–42 (β = 0.27; p = 0.007; FDR-p = 0.007) (Fig. 3 d), CSF p-tau181 (β = 0.66; p < 0.001; FDR-p < 0.001) (Fig. 3 e), and CSF tau (β = 0.74; p < 0.001; FDR-p < 0.001) (Fig. 3 f) (Table 3 ). Plasma total APOE did not significantly associate with CSF Aβ1–42 (β = -0.01; p = 0.890; FDR-p = 0.890) ( Supplementary Fig. 1d ), CSF p-tau181 (β = 0.04; p = 0.711; FDR-p = 0.890) ( Supplementary Fig. 1e ), or CSF tau (β = 0.03; p = 0.788; FDR-p = 0.890) ( Supplementary Fig. 1f ) at baseline (Table 3 ). Total APOE levels were not associated with delayed memory in both CSF (β = 0.08; p = 0.436) and plasma (β = -0.17; p = 0.092) (Table 3 ). Higher levels of CSF APOE-ε4 were not significantly associated with higher levels of CSF Aβ1–42 (β = 0.09; p = 0.422; FDR-p = 0.422) but were significantly associated with higher levels of both CSF p-tau181 (β = 0.48; p < 0.001; FDR-p < 0.001) and CSF tau (β = 0.49; p < 0.001; FDR-p < 0.001) (Table 3 ). CSF APOE-ε4 was not significantly associated with delayed memory (β = 0.15; p = 0.142) (Table 3 ). Higher levels of plasma APOE-ε4 were not significantly associated with CSF Aβ1–42 (β = -0.18; p = 0.124; FDR-p = 0.372) ( Supplementary Fig. 2d ), CSF p-tau181 (β = 0.003; p = 0.978; FDR-p = 0.978) ( Supplementary Fig. 2e ), CSF tau (β = -0.05; p = 0.685; FDR-p = 0.978) ( Supplementary Fig. 2f ) or delayed memory (β = -0.02; p = 0.884) at baseline (Table 3 ). Longitudinal Associations We observed a trend-level longitudinal relationship between baseline CSF percent ε4 and longitudinal changes in delayed memory, such that a higher baseline percentage of ε4 trended with poorer memory recall over a 6-year period (β = 0.02; p = 0.074) (Fig. 5 c). Baseline measurements of CSF percent ε4 were not significantly associated with CSF Aβ1–42 (β = -0.02; p = 0.140), CSF p-tau181 (β = -0.01; p = 0.417), or CSF tau (β = -0.01; p = 0.463) longitudinally. We additionally observed trend-level relationships between baseline measurements of both CSF small HDL and CSF large HDL with CSF p-tau181 longitudinally, but these associations were not maintained following FDR-p correction (β = 0.01; p = 0.069; FDR-p = 0.208 for CSF small HDL; β = 0.01; p = 0.063; FDR-p = 0.190 for CSF large HDL). No associations were observed between baseline measurements of CSF small HDL with longitudinal CSF Aβ1–42 (β = -0.0001; p = 0.985; FDR-p = 0.985) or CSF tau (β = 0.01; p = 0.216; FDR-p = 0.324), or between baseline measurements of CSF large HDL with longitudinal CSF Aβ1–42 (β = 0.001; p = 0.876; FDR-p = 0.876) or CSF tau (β = 0.01; p = 0.307; FDR-p = 0.460). Discussion In this study, we analyzed HDL particle concentrations together with total and isoform-specific APOE protein levels in CSF and plasma and examined how they relate to CSF AD biomarkers in a well-characterized ADNI cohort. Our findings indicate that CNS, rather than peripheral, APOE4 measures align more closely with AD-related biomarkers. Specifically, higher CSF APOE4 protein levels were associated with higher CSF tau and p-tau181 concentrations, whereas plasma APOE4 measures showed minimal and inconsistent relationships with CSF-measured pathology. In addition, the CSF ε4/ε3 ratio was elevated in MCI and in individuals with cognitive progression, while plasma patterns differed in direction, underscoring compartment-specific APOE biology. These results add important nuance to the small HDL hypothesis of AD ( 7 ). While that hypothesis emphasizes the neuroprotective capacity of small HDL particles, through ABCA1-mediated cholesterol efflux, lipid exchange, and synaptic maintenance, our data suggest that this protective biology is insufficient to counteract APOE4 protein burden in the CNS. CSF small HDL correlated positively with ABCA1-mediated cholesterol efflux ex vivo ( 15 ) and with better cognition in an earlier cohort ( 7 ), yet failed to buffer APOE4-biomarker associations in these current analyses. This implies that the pathogenic effects of APOE4 may act through mechanisms at least partially independent of HDL-mediated lipid transport, such as direct promotion of tau seeding, neuroinflammation, or impaired receptor-mediated clearance. The compartment specificity of our findings further aligns with evidence that CSF small HDL formation shares common APOE-dependent mechanisms rather than reflecting passive transport from plasma ( 16 ). Consistent with prior reports ( 7 ), APOE4 carriers exhibited reduced CSF small HDL, yet higher HDL particle levels did not attenuate APOE4-biomarker associations. APOE4 status continued to moderate relationships with Aβ1–42, total tau, and p-tau181 independent of HDL subclass distribution, arguing against a lipidation-deficiency model and instead suggesting that APOE4 protein burden itself drives biomarker associations. The opposing directions of HDL-biomarker associations across genotype groups warrant consideration. In APOE4 non-carriers, the positive association between CSF small HDL and tau may reflect reactive upregulation of HDL production in response to early neurodegeneration, as has been hypothesized for total CSF APOE ( 8 ). In carriers, the inverse association could represent residual ABCA1 function: individuals who maintain higher small HDL despite the ε4 allele may retain better APOE lipidation and less APOE4 aggregation ( 17 ). However, this residual protective signal was overwhelmed by the dominant association between APOE4 protein levels and tau. These genotype-dependent directional differences underscore that HDL particle measurements cannot be interpreted uniformly across APOE genotype groups and suggest that the functional relationship between HDL and neurodegeneration is fundamentally reconfigured by the presence of APOE4. The divergence between plasma and CSF findings further emphasizes the importance of compartmentalization. Plasma APOE measures did not consistently reflect central biomarker status and in some analyses demonstrated opposing patterns, indicating that peripheral APOE dynamics may not mirror CNS APOE biology. Systemic lipid measures are therefore insufficient proxies for central disease mechanisms, and therapeutic strategies aimed at modifying peripheral HDL or systemic APOE may have limited impact on CNS pathology. Analyses within ε3/ε4 heterozygotes reveal that APOE4 burden behaves in a graded manner, scaling with relative ε4 protein proportion rather than genotype status alone. Higher baseline CSF ε4 proportion trended toward greater longitudinal memory decline over six years, supporting a dose-like relationship between CNS APOE4 burden and cognitive vulnerability ( 18 ). Although the association was trend-level and this study does not establish a specific reduction threshold, these graded associations suggest that partial lowering of CNS APOE4 may shift the biomarker profile toward a less pathogenic state. The absence of concurrent associations between CSF percent ε4 and longitudinal AD biomarkers is suggestive of alternate pathways, independent of amyloid and tau, that might mediate APOE4-associated effects on memory. This dose-response pattern provides human evidence aligned with a growing consensus that reducing CNS APOE4 should be a therapeutic goal. An NIA/ADSP Consortium Working Group concluded that 50% reduction in CNS APOE4 is likely tolerable and therapeutic ( 19 ). Several programs are now testing this premise: LX1001, an AAV-delivered APOE2 gene therapy for APOE4 homozygotes, produced dose-dependent APOE2 expression in CSF and reductions in tau biomarkers in Phase 1/2 ( 20 ); preclinical APOE4-to-APOE2 allelic switching in astrocytes reduced amyloid pathology and restored cognition in 5xFAD mice ( 21 ); APOE-targeted ASOs achieved approximately 50% CNS expression reduction with attenuation of tau pathology and microglial activation ( 22 ); and focused ultrasound-mediated CRISPR delivery has been demonstrated for brain-targeted APOE4 knockdown ( 23 ). Our data complement these efforts by demonstrating that quantitative CSF APOE4 protein levels, not genotype status alone, track with tau and p-tau181 in humans, suggesting that CSF APOE4 measurement could serve as a pharmacodynamic biomarker for APOE4-lowering trials. An additional consideration is whether APOE4 modifies the therapeutic window for tau-directed interventions. The anti-tau antibody bepranemab showed reduced tau-PET accumulation preferentially in APOE4 non-carriers with low baseline tau ( 24 ), suggesting that APOE4 may narrow the window during which tau-targeted therapies are effective. This raises the possibility that combination strategies, pairing APOE4 reduction with tau-targeted therapy, may be needed for carriers. This study has several limitations. Although the cohort included 144 participants with comprehensive biomarker characterization, subgroup analyses were performed in smaller subsets, which may limit statistical power and increase the risk of type II error. Moderation analyses examining APOE4-by-HDL interactions are inherently power-sensitive, and some interaction effects did not reach statistical significance in stratified models, warranting cautious interpretation. The ADNI cohort is predominantly White and highly educated, limiting generalizability to more diverse populations. The observational design precludes causal inference. Although HDL particle subclasses were quantified by size, we did not directly measure APOE content within specific HDL particles or assess APOE-containing HDL complexes, and therefore cannot determine whether APOE4 preferentially associates with particular HDL subspecies or whether functional lipidation state, rather than particle size, influences pathogenicity. Finally, we did not evaluate other lipoprotein classes, lipid species, or cell-specific APOE expression patterns, leaving unresolved whether astrocytic, microglial, or other CNS-derived APOE sources differentially contribute to the observed associations. In conclusion, our findings support a CNS-compartment model in which APOE4 protein burden aligns more closely with tau pathology and cognitive decline than peripheral measures or HDL subclass distribution. The absence of a buffering effect of HDL particle abundance argues against lipidation-enhancing strategies as a primary therapeutic approach. Instead, these data identify CNS APOE4 burden as a biologically relevant and potentially modifiable target, and provide a framework for evaluating APOE4-directed interventions, including gene therapies, antisense oligonucleotides, and combination approaches with tau-targeted agents, in future clinical studies. Table 1 . Table 1 Demographic breakdown of participants with complete data for all covariates included at baseline. *NP = Non-Progressors, P=Progressors. N (baseline) Whole Cohort Cognitively Normal, NP Cognitively Normal, P Mild Cognitive Impairment, NP Mild Cognitive Impairment, P 144 27 38 37 42 Age (mean ± SD) 73.0 ± 6.6 73.9 ± 7.0 75.1 ± 5.2 71.2 ± 6.6 72.2 ± 7.1 Sex (N, %) 82 M (56.9%) 15 M (55.6%) 15 M (39.5%) 25 M (67.6%) 27 M (64.3%) APOE genotype (N, %) ɛ3/ɛ3 73 (50.7%) 13 (48.1%) 20 (52.6%) 19 (51.4%) 21 (50.0%) ɛ3/ɛ4 45 (31.3%) 11 (40.7%) 13 (34.2%) 11 (29.7%) 10 (23.8%) ɛ4/ɛ4 26 (18.1%) 3 (11.1%) 5 (13.2%) 7 (18.9%) 11 (26.2%) Race (N, %) Asian 2 (1.4%) 0 (0.0%) 0 (0.0%) 1 (2.7%) 1 (2.4%) Black 7 (4.9%) 2 (7.4%) 1 (2.6%) 2 (5.4%) 2 (4.8%) White 134 (93.1%) 25 (92.6%) 37 (97.4%) 34 (91.9%) 38 (90.5%) Other / Unknown 1 (0.7%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 1 (2.4%) Years of Education (mean ± SD) 16.3 ± 2.7 16.2 ± 2.9 16.2 ± 2.6 16.9 ± 2.5 15.8 ± 2.6 CSF Biomarkers (mean ± SD) Aβ1–42 (pg/mL) 1060.9 ± 466.9 1231.0 ± 406.8 1114.5 ± 499.5 1025.0 ± 507.8 934.8 ± 405.8 t-tau (pg/mL) 263.7 ± 109.4 248.8 ± 67.9 249.9 ± 92.2 255.5 ± 104.4 293.2 ± 142.5 p-tau181 (pg/mL) 24.9 ± 12.0 22.8 ± 7.3 23.4 ± 9.5 24.1 ± 11.2 28.4 ± 16.0 Small HDL (nmol/L) (mean ± SD) 2177.0 ± 515.8 2360.3 ± 518.0 2282.9 ± 531.2 2272.0 ± 496.4 1879.6 ± 394.4 Large HDL (nmol/L) (mean ± SD) 150.2 ± 49.0 178.8 ± 44.9 154.1 ± 46.0 166.5 ± 46.6 114.0 ± 33.7 Table 2 APOE-ε4 carrier status moderation of HDL-biomarker associations. All regression models included covariates for age, sex, education, and clinical group. CSF Small HDL x APOE-ɛ4 on Aβ1–42 Interaction term (β, p-value, FDR-p) APOE-ɛ4 non-carrier slope (β, p-value) APOE-ɛ4 carrier slope (β, p-value) β = -0.41; p = 0.025; FDR-p = 0.038 β = -0.04; p = 0.687 β = -0.45; p = 0.001 CSF Small HDL x APOE-ɛ4 on tau β = -0.52; p = 0.015; FDR-p = 0.038 β = 0.23; p = 0.058 β = -0.29; p = 0.073 CSF Small HDL x APOE-ɛ4 on p-tau181 β = -0.40; p = 0.054; FDR-p = 0.054 β = 0.21; p = 0.076 β = -0.19; p = 0.229 CSF Large HDL x APOE-ɛ4 on Aβ1–42 β = -0.28; p = 0.071; FDR-p = 0.071 β = 0.06; p = 0.576 β = -0.22; p = 0.075 CSF Large HDL x APOE-ɛ4 on tau β = -0.52; p = 0.003; FDR-p = 0.009 β = 0.24; p = 0.049 β = -0.28; p = 0.046 CSF Large HDL x APOE-ɛ4 on p-tau181 β = -0.40; p = 0.024; FDR-p = 0.036 β = 0.19; p = 0.110 β = -0.20; p = 0.143 Table 3 Cross-sectional results of CSF and plasma HDL levels, APOE, and APOE-ε4 measures with clinical biomarker and cognitive outcomes. All regression models included covariates for age, sex, education, clinical group, and ε4 positivity (in models that did not include ε4 ratio or amount as an independent variable). CSF Total APOE CSF Aβ1–42 (β, p-value, FDR-p) CSF total tau (β, p-value, FDR-p) CSF p-tau181 (β, p-value, FDR-p) RAVLT % forgetfulness (β, p-value) β = 0.27; p = 0.007; FDR-p = 0.007 β = 0.74; p < 0.001; FDR-p < 0.001 β = 0.66; p < 0.001; FDR-p < 0.001 β = 0.08; p = 0.436 Plasma Total APOE β = -0.01; p = 0.890; FDR-p = 0.890 β = 0.03; p = 0.788; FDR-p = 0.890 β = 0.04; p = 0.711; FDR-p = 0.890 β = -0.17; p = 0.092 CSF APOE-ɛ4 β = 0.09; p = 0.422; FDR-p = 0.422 β = 0.49; p < 0.001; FDR-p < 0.001 β = 0.48; p < 0.001; FDR-p < 0.001 β = 0.15; p = 0.142 Plasma Total APOE-ɛ4 β = -0.18; p = 0.124; FDR-p = 0.372 β = -0.05; p = 0.685; FDR-p = 0.978 β = 0.003; p = 0.978; FDR-p = 0.978 β = -0.02; p = 0.884 Declarations Funding This study was funded in part by the National Institutes of Health/National Institute on Aging (NIH/NIA), RF1AG076124, R01AG055770, R01AG067063, R01AG054434, R21AG056518, and P30AG066530 to H.N.Y.; Alzheimer’s Drug Discovery Foundation (ADDF) (GC-201711–2014197 to H.N.Y.); and donations from the Vranos and Tiny Foundations and Ms. Lynne Nauss to H.N.Y. M.S.R (AARFD-24-1313939) was supported by the Alzheimer’s Association Research Fellowship. T.A.P was supported by the NIH (#R01AG075336 and #R01AG073267) and Alzheimer’s Association (#AACSF-20-648075). Data collection and sharing for this project was funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd. and its affiliated company, Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions were facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization was the Northern California Institute for Research and Education, and the study was coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California. Author Contribution Z.E.T. analyzed the data and wrote the manuscript; I.A. analyzed the data and wrote the manuscript; A.A.T. reviewed data analysis and revised the manuscript; J.P.B.C wrote the manuscript; L.B. wrote the manuscript; A.S. revised the manuscript; G.P.C.G. revised the manuscript; W.J.M. revised the manuscript and provided data analysis feedback; M.N.B. provided oversight and revised the manuscript; D.N. and R.M.K. contributed to sample processing; H.N.Y. designed the study, obtained ADNI approvals, and wrote the manuscript. References Yassine HN, Hugo C, O'Donovan B, Stephens IO, Johnson LA, Cole G et al. APOE-Targeted Therapeutics for Alzheimer's Disease. J Neurosci. 2025;45(46). Di Battista AM, Heinsinger NM, Rebeck GW. Alzheimer's Disease Genetic Risk Factor APOE-ε4 Also Affects Normal Brain Function. Curr Alzheimer Res. 2016;13(11):1200–7. Baek MS, Cho H, Lee HS, Lee JH, Ryu YH, Lyoo CH. Effect of APOE ε4 genotype on amyloid-β and tau accumulation in Alzheimer's disease. Alzheimers Res Ther. 2020;12(1):140. Liu M, Kuhel DG, Shen L, Hui DY, Woods SC. Apolipoprotein E does not cross the blood-cerebrospinal fluid barrier, as revealed by an improved technique for sampling CSF from mice. Am J Physiol Regul Integr Comp Physiol. 2012;303(9):R903–8. Linton MF, Gish R, Hubl ST, Bütler E, Esquivel C, Bry WI, et al. Phenotypes of apolipoprotein B and apolipoprotein E after liver transplantation. J Clin Invest. 1991;88(1):270–81. Fisher EA, Feig JE, Hewing B, Hazen SL, Smith JD. High-density lipoprotein function, dysfunction, and reverse cholesterol transport. Arterioscler Thromb Vasc Biol. 2012;32(12):2813–20. Martinez AE, Weissberger G, Kuklenyik Z, He X, Meuret C, Parekh T, et al. The small HDL particle hypothesis of Alzheimer's disease. Alzheimers Dement. 2023;19(2):391–404. Martínez-Morillo E, Hansson O, Atagi Y, Bu G, Minthon L, Diamandis EP, et al. Total apolipoprotein E levels and specific isoform composition in cerebrospinal fluid and plasma from Alzheimer's disease patients and controls. Acta Neuropathol. 2014;127(5):633–43. Turri M, Conti E, Pavanello C, Gastoldi F, Palumbo M, Bernini F, et al. Plasma and cerebrospinal fluid cholesterol esterification is hampered in Alzheimer's disease. Alzheimers Res Ther. 2023;15(1):95. Nedelkov D, Tsokolas ZE, Rodrigues MS, Sible I, Han SD, Kerman BE, et al. Increased cerebrospinal fluid and plasma apoE glycosylation is associated with reduced levels of Alzheimer's disease biomarkers. Alzheimers Res Ther. 2025;17(1):151. Olsson A, Vanderstichele H, Andreasen N, De Meyer G, Wallin A, Holmberg B, et al. Simultaneous measurement of beta-amyloid(1–42), total tau, and phosphorylated tau (Thr181) in cerebrospinal fluid by the xMAP technology. Clin Chem. 2005;51(2):336–45. Tosun D, Hausle Z, Thropp P, Concha-Marambio L, Lamoureux J, Lebovitz R, et al. Association of CSF α-synuclein seed amplification assay positivity with disease progression and cognitive decline: A longitudinal Alzheimer's Disease Neuroimaging Initiative study. Alzheimers Dement. 2024;20(12):8444–60. Crane PK, Carle A, Gibbons LE, Insel P, Mackin RS, Gross A, et al. Development and assessment of a composite score for memory in the Alzheimer's Disease Neuroimaging Initiative (ADNI). Brain Imaging Behav. 2012;6(4):502–16. Mora S, Caulfield MP, Wohlgemuth J, Chen Z, Superko HR, Rowland CM, et al. Atherogenic Lipoprotein Subfractions Determined by Ion Mobility and First Cardiovascular Events After Random Allocation to High-Intensity Statin or Placebo: The Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) Trial. Circulation. 2015;132(23):2220–9. Yassine HN, Feng Q, Chiang J, Petrosspour LM, Fonteh AN, Chui HC et al. ABCA1-Mediated Cholesterol Efflux Capacity to Cerebrospinal Fluid Is Reduced in Patients With Mild Cognitive Impairment and Alzheimer's Disease. J Am Heart Assoc. 2016;5(2). Van Valkenburgh J, Meuret C, Martinez AE, Kodancha V, Solomon V, Chen K, et al. Understanding the Exchange of Systemic HDL Particles Into the Brain and Vascular Cells Has Diagnostic and Therapeutic Implications for Neurodegenerative Diseases. Front Physiol. 2021;12:700847. Rawat V, Wang S, Sima J, Bar R, Liraz O, Gundimeda U, et al. ApoE4 Alters ABCA1 Membrane Trafficking in Astrocytes. J Neurosci. 2019;39(48):9611–22. Huang YA, Zhou B, Nabet AM, Wernig M, Südhof TC. Differential Signaling Mediated by ApoE2, ApoE3, and ApoE4 in Human Neurons Parallels Alzheimer's Disease Risk. J Neurosci. 2019;39(37):7408–27. Vance JM, Farrer LA, Huang Y, Cruchaga C, Hyman BT, Pericak-Vance MA, Report of the APOE4 National Institute on Aging/Alzheimer Disease Sequencing Project Consortium Working Group, et al. Reducing APOE4 in Carriers is a Therapeutic Goal for Alzheimer's Disease. Ann Neurol. 2024;95(4):625–34. Johnson KG, Kaplitt M, Kaminsky S, Sondhi D, Amato G, Selvan N, et al. Topline results from Phase 1/2 AAV gene therapy (LX1001) in APOE4/4 homozygotes with Alzheimer’s disease. Alzheimer's Dement. 2025;21(S5):e101538. Golden LR, Siano DS, Stephens IO, MacLean SM, Saito K, Nolt GL, et al. APOE4 to APOE2 allelic switching in mice improves Alzheimer's disease-related metabolic signatures, neuropathology and cognition. Nat Neurosci. 2025;28(12):2461–75. Huynh TV, Liao F, Francis CM, Robinson GO, Serrano JR, Jiang H, et al. Age-Dependent Effects of apoE Reduction Using Antisense Oligonucleotides in a Model of β-amyloidosis. Neuron. 2017;96(5):1013–e234. Zheng K, Tsitsos FN, Batts AJ, Ji R, Nuriel T, Konofagou EE, et al. Focused ultrasound-mediated APOE4 knockdown in mouse brain. Alzheimers Dement. 2025;21(7):e70464. Sidhu JS, Sardu ML, Björnsson MA, Gallais F, Khandelwal A, Byrnes W, et al. Establishing bepranemab posology through exposure-response modeling and simulation for TOGETHER, a double-blind, placebo-controlled Phase II study of bepranemab in prodromal–mild Alzheimer’s disease (AD). Alzheimer's Dement. 2025;21(S5):e099277. Tables Tables are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1Demographics.docx Table2APOE4CarrierStatusModerationofHDLBiomarkerAssociations.docx Table3CrossSectionalAnalyses.docx FinalHDLManuscriptSuppFigure104072026.png Supplementary Figure 1. Plasma Total APOE as a function of Clinical Stage AD Biomarkers FinalHDLManuscriptSuppFigure204072026.png Supplementary Figure 2. Plasma APOE-ε4 as a function of Clinical Stage AD Biomarkers Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 28 Apr, 2026 Reviewers agreed at journal 16 Apr, 2026 Reviewers agreed at journal 15 Apr, 2026 Reviewers invited by journal 15 Apr, 2026 Editor assigned by journal 09 Apr, 2026 Submission checks completed at journal 09 Apr, 2026 First submitted to journal 07 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9349576","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":626305962,"identity":"ce1d6512-868e-4fd0-b61e-db973611e6f5","order_by":0,"name":"Zoe E. Tsokolas","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Zoe","middleName":"E.","lastName":"Tsokolas","suffix":""},{"id":626305963,"identity":"fdd24fc4-d096-46de-b933-a7393a2aaa6e","order_by":1,"name":"Isaac Asante","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Isaac","middleName":"","lastName":"Asante","suffix":""},{"id":626305967,"identity":"dd535dd7-9a9f-4b38-9855-6ca7b76c5fee","order_by":2,"name":"Amaryllis A. Tsiknia","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Amaryllis","middleName":"A.","lastName":"Tsiknia","suffix":""},{"id":626305970,"identity":"bcd5e9d4-9811-4445-b52a-c78f2366b8eb","order_by":3,"name":"Juan Pablo Barbosa-Carvajal","email":"","orcid":"","institution":"Universidad de Antioquia","correspondingAuthor":false,"prefix":"","firstName":"Juan","middleName":"Pablo","lastName":"Barbosa-Carvajal","suffix":""},{"id":626305973,"identity":"eec566e2-df82-4173-9c4b-7e0f1245fe86","order_by":4,"name":"Lailla Burka","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Lailla","middleName":"","lastName":"Burka","suffix":""},{"id":626305975,"identity":"b2e37527-5688-4444-a438-5770bf56168d","order_by":5,"name":"Abhay Sagare","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Abhay","middleName":"","lastName":"Sagare","suffix":""},{"id":626305982,"identity":"52dd2544-341a-4a28-9e1f-3e3644e06d6f","order_by":6,"name":"Gloria Patricia Cardona-Gomez","email":"","orcid":"","institution":"Universidad de Antioquia","correspondingAuthor":false,"prefix":"","firstName":"Gloria","middleName":"Patricia","lastName":"Cardona-Gomez","suffix":""},{"id":626305990,"identity":"d8e0f572-5474-4ef7-a89e-08e18cb182be","order_by":7,"name":"Wendy J. Mack","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Wendy","middleName":"J.","lastName":"Mack","suffix":""},{"id":626305994,"identity":"0a1c556e-c120-4816-bdaf-e52e1062fb70","order_by":8,"name":"Meredith N. Braskie","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Meredith","middleName":"N.","lastName":"Braskie","suffix":""},{"id":626305995,"identity":"144a2d10-7e92-4e9e-9197-015be4b3a118","order_by":9,"name":"Dobrin Nedelkov","email":"","orcid":"","institution":"Isoformix Inc.","correspondingAuthor":false,"prefix":"","firstName":"Dobrin","middleName":"","lastName":"Nedelkov","suffix":""},{"id":626305996,"identity":"0dde23b9-0107-4ed6-99e6-2900d0729704","order_by":10,"name":"Ronald M. Krauss","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Ronald","middleName":"M.","lastName":"Krauss","suffix":""},{"id":626305997,"identity":"19c79d5f-67e7-45e6-9e90-a6e34d081d0b","order_by":11,"name":"Hussein N. Yassine","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAv0lEQVRIiWNgGAWjYBACAxjJD+Ezk6BFsoE0LSDGAWK1mEskP/74o+COnPGN9IcfGCqsExsIabGckWYmzWPwzNjsRo6xBMOZdMJaDG4kmDEzGBxO3HYjh42Bse0wMVrSP3/8YXC4fvOM9GcMjP+I0pJjIMFjcDjBQCLBjIGxgRgtZ96UAf1y2HDGmTfGEgnH0o0Jazmevvnjjz+H5fnbgSH2ocZalqAWBoEEJE4CDkWogP8AUcpGwSgYBaNgJAMAMoZASoYtDyQAAAAASUVORK5CYII=","orcid":"","institution":"University of Southern California","correspondingAuthor":true,"prefix":"","firstName":"Hussein","middleName":"N.","lastName":"Yassine","suffix":""},{"id":626305998,"identity":"7eb8e7bc-1b4e-4fe3-b0d0-459339d5bc56","order_by":12,"name":"for the Alzheimer's Disease Neuroimaging Initiative* ADNI","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"for","middleName":"the Alzheimer's Disease Neuroimaging Initiative*","lastName":"ADNI","suffix":""}],"badges":[],"createdAt":"2026-04-07 22:23:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9349576/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9349576/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107707561,"identity":"1433166f-ad2b-47ea-9489-67b3d360d50d","added_by":"auto","created_at":"2026-04-24 09:20:35","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":868277,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSmall HDL as a function of Clinical Stage AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptFigure104072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/bf212a244258b26d0d3cc7e1.png"},{"id":107675418,"identity":"8e287067-a829-4061-ad34-6086eee35b0e","added_by":"auto","created_at":"2026-04-24 00:42:38","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":861897,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLarge HDL as a function of Clinical Stage AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptFigure204072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/ea88483104475dace7888ded.png"},{"id":107707816,"identity":"aa8b5f94-1d82-41f5-a6c8-537d3d29905f","added_by":"auto","created_at":"2026-04-24 09:21:11","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1224023,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCSF Total APOE as a function of Clinical Stage AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptFigure304072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/8877acae0196f65fbd72eb13.png"},{"id":107706414,"identity":"cabf072b-c1d5-44c8-ba5f-b168a79e1959","added_by":"auto","created_at":"2026-04-24 09:18:04","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1395413,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCSF APOE isoform concentrations as a function of Clinical Stage and AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptFigure404072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/3e08db6e891fbaef74fd4de8.png"},{"id":107675421,"identity":"b0217a61-5078-49e1-b38b-bb91f0a4c204","added_by":"auto","created_at":"2026-04-24 00:42:38","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":939464,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCSF ε4/ε3 Ratio as a function of Clinical Stage and AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptFigure504072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/9573042dbc07f9bff46d3a52.png"},{"id":108490769,"identity":"bc635c99-82bd-4191-b2ee-ea7c951e6b2a","added_by":"auto","created_at":"2026-05-05 09:48:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4944192,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/0ddffd7d-cee7-4ffa-8128-dabdcde3b75e.pdf"},{"id":107675416,"identity":"038dcd2d-6438-4057-a4dd-2a356e5e5072","added_by":"auto","created_at":"2026-04-24 00:42:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":17025,"visible":true,"origin":"","legend":"","description":"","filename":"Table1Demographics.docx","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/2a33ac90b6fb2486f6a53e44.docx"},{"id":107708103,"identity":"4341fe49-38c4-4483-9321-9035a7adb372","added_by":"auto","created_at":"2026-04-24 09:21:56","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":14495,"visible":true,"origin":"","legend":"","description":"","filename":"Table2APOE4CarrierStatusModerationofHDLBiomarkerAssociations.docx","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/12e931ba95121b8aa57a18a2.docx"},{"id":107707841,"identity":"1bfaa722-7c94-40a3-8063-c944eb79c835","added_by":"auto","created_at":"2026-04-24 09:21:15","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":14420,"visible":true,"origin":"","legend":"","description":"","filename":"Table3CrossSectionalAnalyses.docx","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/4206fa0e72237ad59ce03955.docx"},{"id":107706378,"identity":"1318897a-b310-4a05-8986-1b0e8465dc6d","added_by":"auto","created_at":"2026-04-24 09:17:58","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":1140379,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 1. Plasma Total APOE as a function of Clinical Stage AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptSuppFigure104072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/061c7ff52dfbb503d35dadf0.png"},{"id":107675425,"identity":"724438c0-8137-41c0-b027-14accf6b231e","added_by":"auto","created_at":"2026-04-24 00:42:38","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":860114,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Figure 2. Plasma APOE-ε4 as a function of Clinical Stage AD Biomarkers\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"FinalHDLManuscriptSuppFigure204072026.png","url":"https://assets-eu.researchsquare.com/files/rs-9349576/v1/b1818c4ef45b253fd06e1d75.png"}],"financialInterests":"No competing interests reported.","formattedTitle":"Beyond Lipidation: CSF APOE4 Protein Burden, Not HDL Subclass, Drives Tau Associations in APOE4 Alzheimer’s Disease","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAlzheimer's disease (AD) is characterized by altered lipid metabolism in which Apolipoprotein E (APOE) isoform-specific differences are thought to impair Aβ clearance. APOE is the major lipid transport protein in the brain, responsible for maintaining neuronal and vascular health through cholesterol delivery and membrane integrity (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). The APOE-ε4 allele confers an altered protein conformation that reduces lipid transport efficiency and disrupts lipid homeostasis, heightening neuronal vulnerability and promoting both vascular risk and neuro-inflammation (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Carriers of the APOE-ε4 allele additionally show increased amyloid deposition, tau phosphorylation, and accelerated cognitive decline relative to non-carriers (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). Despite this strong genetic association, the mechanism by which APOE4 confers AD risk in humans remains unresolved. In particular, it is unclear whether APOE4-associated pathology is driven by altered lipoprotein state or by increased APOE4 protein levels within the central nervous system (CNS).\u003c/p\u003e \u003cp\u003eUnderstanding how APOE4 and its associated HDL particles contribute to AD requires separate evaluation of central and peripheral compartments, as APOE in plasma and cerebrospinal fluid (CSF) may differentially influence lipid homeostasis and Aβ metabolism. Because the blood\u0026ndash;brain and blood\u0026ndash;CSF barriers restrict lipoprotein exchange, APOE exists in biologically distinct pools, with CSF APOE synthesized largely within the CNS rather than derived from peripheral sources (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). In the CSF, APOE is carried on HDL-like particles that support ABCA1-mediated lipid efflux, cholesterol transport, and Aβ trafficking, functions critical for CNS homeostasis. In plasma, HDL mediates reverse cholesterol transport from peripheral tissues to the liver (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). These HDL particles are heterogeneous, comprising small and large subspecies that differ in composition, function, and tissue distribution, and whose relative abundance in CSF may influence both local and systemic lipid homeostasis (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Whether peripheral HDL alterations reflect central APOE biology or whether CNS APOE operates independently of plasma HDL dynamics remains uncertain (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e), and it is unclear whether APOE4-associated risk is mediated primarily through central (CSF) or peripheral (plasma) pathways.\u003c/p\u003e \u003cp\u003eRecent studies suggest that APOE4 carriers exhibit altered HDL particle composition and poor lipidation, leading to the hypothesis that reduced APOE lipidation contributes to toxicity (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Independent studies have shown that small HDL particles in the CSF, but not HDL measures in plasma, correlate with AD-related biomarkers and cognitive performance (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). These observations have motivated therapeutic strategies aimed at enhancing APOE lipidation or increasing HDL-associated APOE. Yet direct human evidence testing whether higher HDL subclass abundance mitigates APOE4-associated biomarker abnormalities is lacking. Moreover, few studies have directly quantified isoform-specific APOE protein levels in both CSF and plasma to determine whether APOE4 protein burden itself, rather than HDL subclass distribution, aligns more closely with AD pathology. We fill this gap in the literature by quantifying isoform-specific APOE concentrations in addition to CSF HDL subclasses to examine whether these differentially associate with AD biomarker levels as a function of APOE4 status.\u003c/p\u003e \u003cp\u003eIn the present study, we address these gaps by integrating CSF and plasma HDL particle measurements with total and isoform-specific APOE protein quantification in a well-characterized cohort of older adults with normal cognition or mild cognitive impairment from the Alzheimer's Disease Neuroimaging Initiative (ADNI). We evaluate cross-sectional associations between CSF HDL subclasses, isoform-specific APOE concentrations, and established AD biomarkers (Aβ1\u0026ndash;42, total tau, p-tau181) measured in CSF. We further examine whether quantitative CNS APOE4 burden predicts longitudinal cognitive decline over six years independent of genotype status. By directly comparing central and peripheral APOE measures and testing whether HDL subclass abundance modifies APOE4-associated biomarker relationships, this study tests whether APOE4-associated pathology is better explained by a lipidation-deficiency model or by a CNS APOE4 protein burden model.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e \u003cem\u003eData used in the preparation of this article were obtained from the Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer\u0026rsquo;s disease (AD).\u003c/em\u003e \u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eHuman Ethics and Consent to Participate Declaration\u003c/h2\u003e \u003cp\u003eADNI adheres to ethical standards that include the Declaration of Helsinki, the International Conference on Harmonization Good Clinical Practice (ICH GCP), and the Code of Federal Regulations (CFR). ADNI also obtains informed consent from participants and follows the ethical standards of the institutional review board (IRB).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eParticipants\u003c/h3\u003e\n\u003cp\u003eThe present study analyzed data from 144 participants (mean age 73.0\u0026thinsp;\u0026plusmn;\u0026thinsp;6.6) from the ADNI dataset (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Participants with complete baseline measurements of CSF and plasma small and large HDL particle levels, CSF biomarker levels (tau, p-tau181, Aβ1\u0026ndash;42), complete measurements of CSF and plasma total APOE, and with available Rey Auditory Verbal Learning Test (RAVLT) scores were included in our study (n\u0026thinsp;=\u0026thinsp;144), with exclusions for ε2/ε3 genotype (n\u0026thinsp;=\u0026thinsp;11). Measures of CSF and plasma APOE3 and APOE4 protein levels were available for 85 of the 144 participants in our sample. Individuals were stratified into four clinical groups based on delayed recall test performance (RAVLT) following methods previously described (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e): (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) cognitively normal (CN) stable (n\u0026thinsp;=\u0026thinsp;27), with no cognitive impairment at baseline assessment and \u0026lt;\u0026thinsp;10% RAVLT decline between baseline and 48-month follow-up; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) CN decline (n\u0026thinsp;=\u0026thinsp;38), with no cognitive impairment at baseline but \u0026gt;\u0026thinsp;10% RAVLT decline between baseline assessment and 48-months follow-up; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) Late Mild Cognitive Impairment (LMCI) stable (n\u0026thinsp;=\u0026thinsp;37), with participants having mild cognitive impairment (MCI) at baseline but no AD progression between baseline and 48-months of follow-up; lastly (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e), LMCI decline (n\u0026thinsp;=\u0026thinsp;42) with participants having MCI at baseline and AD progression between baseline and 48-months of follow-up.\u003c/p\u003e\n\u003ch3\u003eStudy data\u003c/h3\u003e\n\u003cp\u003eBiomarker data were available from CSF samples collected during annual visits spanning a 6-year period. The concentrations of Aβ1\u0026ndash;42, t-tau, and p-tau181 in the CSF were measured using the Luminex platform INNOBIA AlzBio3 RUO test, using a microbead-based multiplex immunoassay as previously described (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). The CSF Aβ1\u0026ndash;42 assay had a range of 200 to 1700 pg/mL (Elecsys CSF Aβ42 immunoassay) and samples with Aβ42 levels that exceeded this range were assigned the maximum value of 1700 pg/mL (n\u0026thinsp;=\u0026thinsp;27) (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). For cognitive measures, the RAVLT accuracy score of delayed memory recall by % forgotten on a list-recall paradigm was used (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). RAVLT scores, used for clinical group categorization, were used in cognitive analyses and were available from each annual visit over the 6-year period.\u003c/p\u003e\n\u003ch3\u003eTotal APOE Measurement\u003c/h3\u003e\n\u003cp\u003eCerebrospinal fluid (CSF) and plasma levels of APOE were determined using a Meso Scale Discovery (MSD) R-PLEX human APOE singleplex assay (Catalog number K151AMLR) following the manufacturer\u0026rsquo;s standard recommendations for assay protocol 1. Briefly, 25 \u0026micro;l per well of biotinylated capture antibody diluted in coating diluent 100 (Catalog number R50AA, MSD) was added to each well of MSD small spot streptavidin plates, and adhesive sealed plates were incubated for 1 hour with shaking at 700 rpm at 25 \u0026ordm;C on an orbital shaker (Jitterbug-2, Boekel Scientific, Feasterville, PA, USA). The plates were washed three times with 150 \u0026micro;L/well of wash buffer (0.01 M PBS pH 7.4 with 0.05% Tween-20) and tapped on absorbent Kimwipes (Catalog number 34133, Kimberly-Clark Global Sales, LLC, Roswell, GA, USA). MSD supplied human plasma-purified APOE calibrator standard concentrations (750, 187.50, 46.88, 11.72, 2.93, 0.73, 0.18 ng/mL) were prepared in the assay diluent 57 (Catalog number R50BZ, MSD). The assay diluent was used as the zero-calibrator standard. Human EDTA plasma samples were diluted 100-fold, and CSF samples were diluted 50-fold in the assay diluent 57 in protein LoBind tubes (Eppendorf AG, Hamburg, Germany). Twenty-five microliters of diluent 57 was added to each well of the coated plates prior to the addition of 25 \u0026micro;L of prepared calibrator standards or samples in designated wells. The plates were sealed and incubated for 1 hour with shaking at 700 rpm at 25\u0026ordm; C on an orbital shaker. The plates were washed three times with 150 \u0026micro;L/well of wash buffer and tapped on absorbent Kimwipes to remove residual wash buffer. The supplied SULFO-TAGTM-conjugated 100X detection antibody was diluted in diluent 3 (Catalog number R50AP-2, MSD). Fifty microliters of the detection antibody solution were added to each well, and the sealed plates were incubated for 1 hour with shaking at 700 rpm at 25 \u0026ordm;C on an orbital shaker. The plates were washed three times with 150 \u0026micro;L/well of wash buffer and tapped on absorbent Kimwipes to remove residual wash buffer. MSD read buffer T (Catalog number R92TC-2, MSD) was diluted 2-fold in UltraPure\u0026trade; distilled water (Catalog number 10977-015, Invitrogen, Grand Island, NY, USA), and 150 \u0026micro;L of read buffer was added to each well. The plates were read immediately on an MESO Quickplex\u0026reg; SQ120MM instrument with electrochemiluminescence detection. Calibrator standard curves were used to calculate plasma and CSF APOE concentrations, after correcting for the sample dilution factor.\u003c/p\u003e\n\u003ch3\u003eAPOE3 and APOE4 Measurement in CSF and Plasma\u003c/h3\u003e\n\u003cp\u003eCSF and plasma samples from baseline assessments were analyzed using an APOE mass spectrometric immunoassay (MSIA) as described previously (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). The MSIA detects 3 APOE isoforms at the intact protein level, with apoE affinity-captured from CSF and plasma using MSIA tips derivatized with APOE-specific antibodies. A MALDI-TOF mass spectrometer (Autoflex III MALDI-TOF, Bruker, Billerica, MA) was used to analyzed eluted and intact APOE in positive ion mode, with a mass range of 7\u0026ndash;70 kDa, 700 ns delay, 20.00 kV and 18.45 kV ion source voltages, and \u0026lt;\u0026thinsp;7000 Da signal suppression. Mass spectra were baseline subtracted (Convex Hull algorithm with 0.8 flatness) and smoothed (Savizky Golay algorithm with 5 m/z width and 1 cycle) using the Bruker Flex Analysis software. Zebra 1.0 software (Intrinsic Bioprobes Inc.) was used to quantify and tabulate the peak intensities of \u003cem\u003eAPOE\u003c/em\u003e isoforms. A variable representative of relative abundance of ApoE4 to ApoE3 in heterozygotes (ε4/ε3-ratio) was calculated by dividing CSF ApoE4 isoform level by CSF ApoE3 isoform level. An additional variable (percent ε4) was derived by multiplying total CSF ApoE4 isoform proportion by the CSF total ApoE concentration, representative of absolute concentration of ApoE4 protein in CSF (\u0026micro;g/mL).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSmall and Large HDL Measurement\u003c/h2\u003e \u003cp\u003eConcentrations of HDL particles were measured by ion mobility (IM) after treatment with dextran sulfate to remove non-lipid-bound proteins such as albumin from 30 \u0026micro;L of plasma or CSF as described previously (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Additional details can be found in the Supplementary Materials (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Particles in the HDL size range were classified as small (7.0-10.5 nm) and large (10.5\u0026ndash;14.5 nm) HDL. The majority of HDL particles in both CSF and plasma were found in the small size range, with a greater proportion of small to large HDLs present in CSF compared to plasma (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eIn cross-sectional analyses, linear regression models adjusting for age, sex, education, ε4 positivity, and clinical group were used to examine the associations between baseline independent variables (CSF and plasma small and large HDL/LDL particles, percent ε4, and ε4/ε3-ratio) and AD biomarker outcomes. Pearson\u0026rsquo;s correlations were used to examine the linear association of these independent variables measured in plasma vs CSF.\u003c/p\u003e \u003cp\u003eTo evaluate between-group differences on CSF HDL levels and APOE concentrations, we utilized one-way analysis of variance (ANOVA) tests, followed by post-hoc pairwise comparisons using Tukey\u0026rsquo;s honestly significant difference (HSD) test to control for multiple comparisons. Group comparisons were made on \u003cem\u003eAPOE\u003c/em\u003e genotype groups (using ANOVA) as well as the two cognitive progression groups (CN vs MCI; Non-progressors vs Progressors) and CN vs MCI (using independent two-sample t-test).\u003c/p\u003e \u003cp\u003eAssociations between baseline measures of small and large HDL particles and percent ε4 with longitudinally-assessed CSF tau, p-tau181, Aβ1\u0026ndash;42 and cognitive performance annually over a 6-year period were tested using linear mixed effects models. HDL particle levels and fluid biomarker measures were log-transformed (log10(x\u0026thinsp;+\u0026thinsp;1)) to reduce skew and improve normality, and all continuous variables were z-scored with a mean of 0 and SD of 1 to derive standardized parameter estimates. Fixed effects of interest were the baseline HDL and ε4 measures; adjusting covariates were age, sex, education, clinical group, and ε4 positivity (in models that did not include ε4 ratio or amount as an independent variable). A random participant-level intercept was specified to account for the within-participant correlation. We tested for an interaction between time modeled as a continuous variable (calculated as years since baseline assessment) and each of our independent variables to test for associations of baseline measures with longitudinal trajectories of change in outcome variables. Time was modeled as a fixed effect as well as a random participant-specific slope to account for longitudinal individual variation, and an unstructured variance-covariance matrix was specified for random effects. Likelihood ratio tests were conducted to compare models both with and without the addition of random slope to evaluate whether its inclusion was justified. We corrected for multiple comparisons across AD biomarker outcomes (Aβ1\u0026ndash;42, tau, p-tau181) using the false discovery rate (FDR) method. All analyses used RStudio (Version 2024.12.1\u0026thinsp;+\u0026thinsp;563); the \u0026lsquo;lme4\u0026rsquo; package was used for linear mixed effect modeling.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eHDL Subclasses and APOE Levels on Clinical Outcomes\u003c/h2\u003e \u003cp\u003eCSF small HDL levels were significantly lower in ε3/ε4 genotypes (β = -0.06; p\u0026thinsp;=\u0026thinsp;0.017) and in ε4/ε4 genotypes (β = -0.06; p\u0026thinsp;=\u0026thinsp;0.033) compared to ε3/ε3 genotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). No significant difference was observed between ε3/ε4 and ε4/ε4 genotypes (β = -0.01; p\u0026thinsp;=\u0026thinsp;0.974). CN individuals had significantly higher levels of CSF small HDL compared to those with MCI (β\u0026thinsp;=\u0026thinsp;0.05; p\u0026thinsp;=\u0026thinsp;0.006) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e1\u003c/span\u003eb), as did cognitive non-progressors compared to progressors (β\u0026thinsp;=\u0026thinsp;0.05; p\u0026thinsp;=\u0026thinsp;0.010) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e1\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003eCSF large HDL levels did not significantly differ across APOE genotypes (ε3/ε3 vs ε3/ε4: β = -0.03, p\u0026thinsp;=\u0026thinsp;0.497; ε3/ε3 vs ε4/ε4: β = -0.05, p\u0026thinsp;=\u0026thinsp;0.348; ε3/ε4 vs ε4/ε4: β = -0.01, p\u0026thinsp;=\u0026thinsp;0.907) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). CSF large HDL did significantly differ across cognitive diagnoses and progressor status, with higher levels of large HDL in CN (β\u0026thinsp;=\u0026thinsp;0.08; p\u0026thinsp;=\u0026thinsp;0.001) compared to MCI (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e2\u003c/span\u003eb), and higher levels in non-progressors compared to progressors (β\u0026thinsp;=\u0026thinsp;0.11 p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e2\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003eCSF total APOE levels significantly differed by genotype, with lower CSF total APOE levels in ε3/ε4 (β = -0.69; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and ε4/ε4 genotypes (β = -0.80; p\u0026thinsp;=\u0026thinsp;0.001), compared to ε3/ε3 genotypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Total APOE levels in CSF did not differ significantly between cognitive diagnoses (β\u0026thinsp;=\u0026thinsp;0.06; p\u0026thinsp;=\u0026thinsp;0.730) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). Total APOE levels in CSF also did not significantly differ by progressor status in CSF (β\u0026thinsp;=\u0026thinsp;0.11; p\u0026thinsp;=\u0026thinsp;0.541) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e3\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003eCSF APOE-ε3 levels were significantly higher in ε3/ε3 genotypes than ε3/ε4 (β\u0026thinsp;=\u0026thinsp;2.53; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ea), but APOE-ε3 levels did not significantly differ across cognitive diagnoses (β = -0.03; p\u0026thinsp;=\u0026thinsp;0.912) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eb) or progressor status groups (β\u0026thinsp;=\u0026thinsp;0.20; p\u0026thinsp;=\u0026thinsp;0.463) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). CSF APOE-ε4 levels were significantly higher in ε4/ε4 genotypes than ε3/ε4 (β\u0026thinsp;=\u0026thinsp;0.79; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ea), but APOE-ε4 levels did not significantly differ across cognitive diagnoses (β = -0.27; p\u0026thinsp;=\u0026thinsp;0.125) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eb) or progressor status groups (β = -0.33; p\u0026thinsp;=\u0026thinsp;0.073) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). In ε3/ε4 heterozygous carriers, CSF APOE-ε4 levels were significantly higher than CSF APOE-ε3 levels (β\u0026thinsp;=\u0026thinsp;0.89; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ea).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eAPOE Isoform Levels, APOE-ε4/ε3 ratio, and APOE-ε4% in CSF\u003c/h2\u003e \u003cp\u003eThere was a trend-level interaction between CSF APOE-ε3 levels and CSF APOE-ε4 levels with baseline CSF Aβ1\u0026ndash;42 (β = -0.87; p\u0026thinsp;=\u0026thinsp;0.076; FDR-p\u0026thinsp;=\u0026thinsp;0.095) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ed) and CSF p-tau181 (β = -0.75; p\u0026thinsp;=\u0026thinsp;0.095; FDR-p\u0026thinsp;=\u0026thinsp;0.095) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ee); however, associations were not sustained following FDR correction. There was a significant interaction between CSF APOE-ε3 levels and CSF APOE-ε4 levels with baseline CSF tau (β = -0.97; p\u0026thinsp;=\u0026thinsp;0.022; FDR-p\u0026thinsp;=\u0026thinsp;0.066) that became trend-level following FDR correction (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003ef).\u003c/p\u003e \u003cp\u003eThe CSF ε4/ε3 ratio was significantly higher in those with MCI compared to CN (β\u0026thinsp;=\u0026thinsp;0.16; p\u0026thinsp;=\u0026thinsp;0.004) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e5\u003c/span\u003ea) and did not significantly differ between cognitive progressors and non-progressors (β = -0.07; p\u0026thinsp;=\u0026thinsp;0.256) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e5\u003c/span\u003eb). Cross-sectional associations between the CSF ε4/ε3 ratio and CSF Aβ1\u0026ndash;42 (β = -0.05; p\u0026thinsp;=\u0026thinsp;0.742), CSF p-tau181 (β\u0026thinsp;=\u0026thinsp;0.20; p\u0026thinsp;=\u0026thinsp;0.248), and CSF tau (β\u0026thinsp;=\u0026thinsp;0.17; p\u0026thinsp;=\u0026thinsp;0.327) were not significant (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e5\u003c/span\u003ed-f).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eClinical Outcome Differences Across APOE Levels and Ratios\u003c/h2\u003e \u003cp\u003eTotal plasma APOE levels (in \u0026micro;g/mL) differed by genotype, with lower plasma total APOE in ε3/ε4 genotypes (β = -7.59; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and ε4/ε4 genotypes (β = -7.16; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), relative to ε3/ε3 genotypes (\u003cb\u003eSupplementary Fig.\u0026nbsp;1a\u003c/b\u003e). Total plasma APOE levels did not differ significantly by cognitive diagnoses (β = -1.16; p\u0026thinsp;=\u0026thinsp;0.343) (\u003cb\u003eSupplementary Fig.\u0026nbsp;1b\u003c/b\u003e) or progressor status in plasma (β = -0.53; p\u0026thinsp;=\u0026thinsp;0.653) (\u003cb\u003eSupplementary Fig.\u0026nbsp;1c\u003c/b\u003e).\u003c/p\u003e \u003cp\u003ePlasma APOE-ε4 levels were significantly higher in homozygous carriers compared to heterozygous carriers (β\u0026thinsp;=\u0026thinsp;1.47; p\u0026thinsp;=\u0026thinsp;0.009) (\u003cb\u003eSupplementary Fig.\u0026nbsp;2a\u003c/b\u003e) and in MCI compared to CN (β\u0026thinsp;=\u0026thinsp;1.04; p\u0026thinsp;=\u0026thinsp;0.023) (\u003cb\u003eSupplementary Fig.\u0026nbsp;2b\u003c/b\u003e), but did not significantly differ by progressor groups (β\u0026thinsp;=\u0026thinsp;0.07; p\u0026thinsp;=\u0026thinsp;0.890) (\u003cb\u003eSupplementary Fig.\u0026nbsp;2c\u003c/b\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eCross-Sectional Associations: CSF HDLs and CSF APOE levels with CSF Biomarkers\u003c/h2\u003e \u003cp\u003eAPOE-ε4 carrier status significantly moderated the relationship between CSF small HDL and CSF Aβ1\u0026ndash;42 (interaction β = -0.41; p\u0026thinsp;=\u0026thinsp;0.025; FDR-p\u0026thinsp;=\u0026thinsp;0.038), such that small HDL was significantly associated with lower CSF Aβ1\u0026ndash;42 in APOE-ε4 carriers (β = -0.45; p\u0026thinsp;=\u0026thinsp;0.001) and not in non-carriers (β = -0.04; p\u0026thinsp;=\u0026thinsp;0.687) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e1\u003c/span\u003ed) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). APOE-ε4 carrier status also moderated the relationship between CSF small HDL and CSF p-tau181 at a trend-level (interaction β = -0.40; p\u0026thinsp;=\u0026thinsp;0.054; FDR-p\u0026thinsp;=\u0026thinsp;0.054), such that small HDL was at a trend-level associated with higher CSF p-tau181 among non-carriers (β\u0026thinsp;=\u0026thinsp;0.21; p\u0026thinsp;=\u0026thinsp;0.076), whereas no significant association was observed among carriers (β = -0.19; p\u0026thinsp;=\u0026thinsp;0.229) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). APOE-ε4 carrier status furthermore significantly moderated the relationship between CSF small HDL on CSF tau (β = -0.52, p\u0026thinsp;=\u0026thinsp;0.015; FDR-p\u0026thinsp;=\u0026thinsp;0.038), with higher levels of small HDL trend-level associated with higher levels of CSF tau in non-carriers (β\u0026thinsp;=\u0026thinsp;0.23; p\u0026thinsp;=\u0026thinsp;0.058) but oppositely associated with lower levels of CSF tau in carriers (β = -0.29; p\u0026thinsp;=\u0026thinsp;0.073) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAPOE-ε4 significantly moderated the association between CSF large HDL and CSF p-tau181 (β = -0.40; p\u0026thinsp;=\u0026thinsp;0.024; FDR-p\u0026thinsp;=\u0026thinsp;0.036), with a non-significant positive trend in APOE-ε4 non carriers (β\u0026thinsp;=\u0026thinsp;0.19, p\u0026thinsp;=\u0026thinsp;0.110) and a non-significant negative trend in carriers (β = -0.20, p\u0026thinsp;=\u0026thinsp;0.143) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). APOE-ε4 significantly moderated the association between CSF large HDL and CSF tau (β = -0.52; p\u0026thinsp;=\u0026thinsp;0.003; FDR-p\u0026thinsp;=\u0026thinsp;0.009), such that in APOE-ε4 non-carriers, higher CSF large HDL was significantly associated with higher CSF tau (β\u0026thinsp;=\u0026thinsp;0.24, p\u0026thinsp;=\u0026thinsp;0.049) but with lower CSF tau in APOE-ε4 carriers (β = -0.28, p\u0026thinsp;=\u0026thinsp;0.046) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). We observed a trend-level interaction between APOE-ε4 x CSF large HDL with CSF Aβ1\u0026ndash;42 (β = -0.28; p\u0026thinsp;=\u0026thinsp;0.071; FDR-p\u0026thinsp;=\u0026thinsp;0.071), with APOE-ε4 non-carriers having no significant association with CSF Aβ1\u0026ndash;42 (β\u0026thinsp;=\u0026thinsp;0.06, p\u0026thinsp;=\u0026thinsp;0.576) and APOE-ε4 carriers having a trend-level association between higher CSF large HDL with lower CSF Aβ1\u0026ndash;42 (β = -0.22, p\u0026thinsp;=\u0026thinsp;0.075) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn the whole sample, CSF total APOE was cross-sectionally associated with significantly higher levels of CSF Aβ1\u0026ndash;42 (β\u0026thinsp;=\u0026thinsp;0.27; p\u0026thinsp;=\u0026thinsp;0.007; FDR-p\u0026thinsp;=\u0026thinsp;0.007) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e3\u003c/span\u003ed), CSF p-tau181 (β\u0026thinsp;=\u0026thinsp;0.66; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e3\u003c/span\u003ee), and CSF tau (β\u0026thinsp;=\u0026thinsp;0.74; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e3\u003c/span\u003ef) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Plasma total APOE did not significantly associate with CSF Aβ1\u0026ndash;42 (β = -0.01; p\u0026thinsp;=\u0026thinsp;0.890; FDR-p\u0026thinsp;=\u0026thinsp;0.890) (\u003cb\u003eSupplementary Fig.\u0026nbsp;1d\u003c/b\u003e), CSF p-tau181 (β\u0026thinsp;=\u0026thinsp;0.04; p\u0026thinsp;=\u0026thinsp;0.711; FDR-p\u0026thinsp;=\u0026thinsp;0.890) (\u003cb\u003eSupplementary Fig.\u0026nbsp;1e\u003c/b\u003e), or CSF tau (β\u0026thinsp;=\u0026thinsp;0.03; p\u0026thinsp;=\u0026thinsp;0.788; FDR-p\u0026thinsp;=\u0026thinsp;0.890) (\u003cb\u003eSupplementary Fig.\u0026nbsp;1f\u003c/b\u003e) at baseline (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Total APOE levels were not associated with delayed memory in both CSF (β\u0026thinsp;=\u0026thinsp;0.08; p\u0026thinsp;=\u0026thinsp;0.436) and plasma (β = -0.17; p\u0026thinsp;=\u0026thinsp;0.092) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHigher levels of CSF APOE-ε4 were not significantly associated with higher levels of CSF Aβ1\u0026ndash;42 (β\u0026thinsp;=\u0026thinsp;0.09; p\u0026thinsp;=\u0026thinsp;0.422; FDR-p\u0026thinsp;=\u0026thinsp;0.422) but were significantly associated with higher levels of both CSF p-tau181 (β\u0026thinsp;=\u0026thinsp;0.48; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and CSF tau (β\u0026thinsp;=\u0026thinsp;0.49; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). CSF APOE-ε4 was not significantly associated with delayed memory (β\u0026thinsp;=\u0026thinsp;0.15; p\u0026thinsp;=\u0026thinsp;0.142) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Higher levels of plasma APOE-ε4 were not significantly associated with CSF Aβ1\u0026ndash;42 (β = -0.18; p\u0026thinsp;=\u0026thinsp;0.124; FDR-p\u0026thinsp;=\u0026thinsp;0.372) (\u003cb\u003eSupplementary Fig.\u0026nbsp;2d\u003c/b\u003e), CSF p-tau181 (β\u0026thinsp;=\u0026thinsp;0.003; p\u0026thinsp;=\u0026thinsp;0.978; FDR-p\u0026thinsp;=\u0026thinsp;0.978) (\u003cb\u003eSupplementary Fig.\u0026nbsp;2e\u003c/b\u003e), CSF tau (β = -0.05; p\u0026thinsp;=\u0026thinsp;0.685; FDR-p\u0026thinsp;=\u0026thinsp;0.978) (\u003cb\u003eSupplementary Fig.\u0026nbsp;2f\u003c/b\u003e) or delayed memory (β = -0.02; p\u0026thinsp;=\u0026thinsp;0.884) at baseline (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eLongitudinal Associations\u003c/h2\u003e \u003cp\u003eWe observed a trend-level longitudinal relationship between baseline CSF percent ε4 and longitudinal changes in delayed memory, such that a higher baseline percentage of ε4 trended with poorer memory recall over a 6-year period (β\u0026thinsp;=\u0026thinsp;0.02; p\u0026thinsp;=\u0026thinsp;0.074) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e5\u003c/span\u003ec). Baseline measurements of CSF percent ε4 were not significantly associated with CSF Aβ1\u0026ndash;42 (β = -0.02; p\u0026thinsp;=\u0026thinsp;0.140), CSF p-tau181 (β = -0.01; p\u0026thinsp;=\u0026thinsp;0.417), or CSF tau (β = -0.01; p\u0026thinsp;=\u0026thinsp;0.463) longitudinally. We additionally observed trend-level relationships between baseline measurements of both CSF small HDL and CSF large HDL with CSF p-tau181 longitudinally, but these associations were not maintained following FDR-p correction (β\u0026thinsp;=\u0026thinsp;0.01; p\u0026thinsp;=\u0026thinsp;0.069; FDR-p\u0026thinsp;=\u0026thinsp;0.208 for CSF small HDL; β\u0026thinsp;=\u0026thinsp;0.01; p\u0026thinsp;=\u0026thinsp;0.063; FDR-p\u0026thinsp;=\u0026thinsp;0.190 for CSF large HDL). No associations were observed between baseline measurements of CSF small HDL with longitudinal CSF Aβ1\u0026ndash;42 (β = -0.0001; p\u0026thinsp;=\u0026thinsp;0.985; FDR-p\u0026thinsp;=\u0026thinsp;0.985) or CSF tau (β\u0026thinsp;=\u0026thinsp;0.01; p\u0026thinsp;=\u0026thinsp;0.216; FDR-p\u0026thinsp;=\u0026thinsp;0.324), or between baseline measurements of CSF large HDL with longitudinal CSF Aβ1\u0026ndash;42 (β\u0026thinsp;=\u0026thinsp;0.001; p\u0026thinsp;=\u0026thinsp;0.876; FDR-p\u0026thinsp;=\u0026thinsp;0.876) or CSF tau (β\u0026thinsp;=\u0026thinsp;0.01; p\u0026thinsp;=\u0026thinsp;0.307; FDR-p\u0026thinsp;=\u0026thinsp;0.460).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we analyzed HDL particle concentrations together with total and isoform-specific APOE protein levels in CSF and plasma and examined how they relate to CSF AD biomarkers in a well-characterized ADNI cohort. Our findings indicate that CNS, rather than peripheral, APOE4 measures align more closely with AD-related biomarkers. Specifically, higher CSF APOE4 protein levels were associated with higher CSF tau and p-tau181 concentrations, whereas plasma APOE4 measures showed minimal and inconsistent relationships with CSF-measured pathology. In addition, the CSF ε4/ε3 ratio was elevated in MCI and in individuals with cognitive progression, while plasma patterns differed in direction, underscoring compartment-specific APOE biology.\u003c/p\u003e \u003cp\u003eThese results add important nuance to the small HDL hypothesis of AD (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). While that hypothesis emphasizes the neuroprotective capacity of small HDL particles, through ABCA1-mediated cholesterol efflux, lipid exchange, and synaptic maintenance, our data suggest that this protective biology is insufficient to counteract APOE4 protein burden in the CNS. CSF small HDL correlated positively with ABCA1-mediated cholesterol efflux ex vivo (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e) and with better cognition in an earlier cohort (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e), yet failed to buffer APOE4-biomarker associations in these current analyses. This implies that the pathogenic effects of APOE4 may act through mechanisms at least partially independent of HDL-mediated lipid transport, such as direct promotion of tau seeding, neuroinflammation, or impaired receptor-mediated clearance. The compartment specificity of our findings further aligns with evidence that CSF small HDL formation shares common APOE-dependent mechanisms rather than reflecting passive transport from plasma (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). Consistent with prior reports (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e), APOE4 carriers exhibited reduced CSF small HDL, yet higher HDL particle levels did not attenuate APOE4-biomarker associations. APOE4 status continued to moderate relationships with Aβ1\u0026ndash;42, total tau, and p-tau181 independent of HDL subclass distribution, arguing against a lipidation-deficiency model and instead suggesting that APOE4 protein burden itself drives biomarker associations.\u003c/p\u003e \u003cp\u003eThe opposing directions of HDL-biomarker associations across genotype groups warrant consideration. In APOE4 non-carriers, the positive association between CSF small HDL and tau may reflect reactive upregulation of HDL production in response to early neurodegeneration, as has been hypothesized for total CSF APOE (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). In carriers, the inverse association could represent residual ABCA1 function: individuals who maintain higher small HDL despite the ε4 allele may retain better APOE lipidation and less APOE4 aggregation (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). However, this residual protective signal was overwhelmed by the dominant association between APOE4 protein levels and tau. These genotype-dependent directional differences underscore that HDL particle measurements cannot be interpreted uniformly across \u003cem\u003eAPOE\u003c/em\u003e genotype groups and suggest that the functional relationship between HDL and neurodegeneration is fundamentally reconfigured by the presence of APOE4.\u003c/p\u003e \u003cp\u003eThe divergence between plasma and CSF findings further emphasizes the importance of compartmentalization. Plasma APOE measures did not consistently reflect central biomarker status and in some analyses demonstrated opposing patterns, indicating that peripheral APOE dynamics may not mirror CNS APOE biology. Systemic lipid measures are therefore insufficient proxies for central disease mechanisms, and therapeutic strategies aimed at modifying peripheral HDL or systemic APOE may have limited impact on CNS pathology.\u003c/p\u003e \u003cp\u003eAnalyses within ε3/ε4 heterozygotes reveal that APOE4 burden behaves in a graded manner, scaling with relative ε4 protein proportion rather than genotype status alone. Higher baseline CSF ε4 proportion trended toward greater longitudinal memory decline over six years, supporting a dose-like relationship between CNS APOE4 burden and cognitive vulnerability (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). Although the association was trend-level and this study does not establish a specific reduction threshold, these graded associations suggest that partial lowering of CNS APOE4 may shift the biomarker profile toward a less pathogenic state. The absence of concurrent associations between CSF percent ε4 and longitudinal AD biomarkers is suggestive of alternate pathways, independent of amyloid and tau, that might mediate APOE4-associated effects on memory.\u003c/p\u003e \u003cp\u003eThis dose-response pattern provides human evidence aligned with a growing consensus that reducing CNS APOE4 should be a therapeutic goal. An NIA/ADSP Consortium Working Group concluded that 50% reduction in CNS APOE4 is likely tolerable and therapeutic (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). Several programs are now testing this premise: LX1001, an AAV-delivered APOE2 gene therapy for APOE4 homozygotes, produced dose-dependent APOE2 expression in CSF and reductions in tau biomarkers in Phase 1/2 (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e); preclinical APOE4-to-APOE2 allelic switching in astrocytes reduced amyloid pathology and restored cognition in 5xFAD mice (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e); APOE-targeted ASOs achieved approximately 50% CNS expression reduction with attenuation of tau pathology and microglial activation (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e); and focused ultrasound-mediated CRISPR delivery has been demonstrated for brain-targeted APOE4 knockdown (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Our data complement these efforts by demonstrating that quantitative CSF APOE4 protein levels, not genotype status alone, track with tau and p-tau181 in humans, suggesting that CSF APOE4 measurement could serve as a pharmacodynamic biomarker for APOE4-lowering trials. An additional consideration is whether APOE4 modifies the therapeutic window for tau-directed interventions. The anti-tau antibody bepranemab showed reduced tau-PET accumulation preferentially in APOE4 non-carriers with low baseline tau (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e), suggesting that APOE4 may narrow the window during which tau-targeted therapies are effective. This raises the possibility that combination strategies, pairing APOE4 reduction with tau-targeted therapy, may be needed for carriers.\u003c/p\u003e \u003cp\u003eThis study has several limitations. Although the cohort included 144 participants with comprehensive biomarker characterization, subgroup analyses were performed in smaller subsets, which may limit statistical power and increase the risk of type II error. Moderation analyses examining APOE4-by-HDL interactions are inherently power-sensitive, and some interaction effects did not reach statistical significance in stratified models, warranting cautious interpretation. The ADNI cohort is predominantly White and highly educated, limiting generalizability to more diverse populations. The observational design precludes causal inference. Although HDL particle subclasses were quantified by size, we did not directly measure APOE content within specific HDL particles or assess APOE-containing HDL complexes, and therefore cannot determine whether APOE4 preferentially associates with particular HDL subspecies or whether functional lipidation state, rather than particle size, influences pathogenicity. Finally, we did not evaluate other lipoprotein classes, lipid species, or cell-specific APOE expression patterns, leaving unresolved whether astrocytic, microglial, or other CNS-derived APOE sources differentially contribute to the observed associations.\u003c/p\u003e \u003cp\u003eIn conclusion, our findings support a CNS-compartment model in which APOE4 protein burden aligns more closely with tau pathology and cognitive decline than peripheral measures or HDL subclass distribution. The absence of a buffering effect of HDL particle abundance argues against lipidation-enhancing strategies as a primary therapeutic approach. Instead, these data identify CNS APOE4 burden as a biologically relevant and potentially modifiable target, and provide a framework for evaluating APOE4-directed interventions, including gene therapies, antisense oligonucleotides, and combination approaches with tau-targeted agents, in future clinical studies.\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic breakdown of participants with complete data for all covariates included at baseline. *NP\u0026thinsp;=\u0026thinsp;Non-Progressors, P=Progressors.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eN (baseline)\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWhole Cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCognitively Normal, NP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCognitively Normal, P\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMild Cognitive Impairment, NP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003eMild Cognitive Impairment, P\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e144\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e73.0\u0026thinsp;\u0026plusmn;\u0026thinsp;6.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e73.9\u0026thinsp;\u0026plusmn;\u0026thinsp;7.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75.1\u0026thinsp;\u0026plusmn;\u0026thinsp;5.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e71.2\u0026thinsp;\u0026plusmn;\u0026thinsp;6.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e72.2\u0026thinsp;\u0026plusmn;\u0026thinsp;7.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSex (N, %)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e82 M (56.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 M (55.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15 M (39.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e25 M (67.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e27 M (64.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAPOE\u003c/b\u003e \u003cb\u003egenotype (N, %)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eɛ3/ɛ3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e73 (50.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13 (48.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20 (52.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e19 (51.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e21 (50.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eɛ3/ɛ4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45 (31.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (40.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13 (34.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e11 (29.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e10 (23.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eɛ4/ɛ4\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26 (18.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (11.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5 (13.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7 (18.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e11 (26.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c7\" namest=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRace (N, %)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eAsian\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (2.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e1 (2.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eBlack\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7 (4.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (7.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (2.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2 (5.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e2 (4.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eWhite\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e134 (93.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25 (92.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e37 (97.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e34 (91.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e38 (90.5%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eOther / Unknown\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (0.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e1 (2.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eYears of Education (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.3\u0026thinsp;\u0026plusmn;\u0026thinsp;2.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.2\u0026thinsp;\u0026plusmn;\u0026thinsp;2.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16.2\u0026thinsp;\u0026plusmn;\u0026thinsp;2.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e16.9\u0026thinsp;\u0026plusmn;\u0026thinsp;2.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e15.8\u0026thinsp;\u0026plusmn;\u0026thinsp;2.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF Biomarkers (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eAβ1\u0026ndash;42 (pg/mL)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1060.9\u0026thinsp;\u0026plusmn;\u0026thinsp;466.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1231.0\u0026thinsp;\u0026plusmn;\u0026thinsp;406.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1114.5\u0026thinsp;\u0026plusmn;\u0026thinsp;499.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1025.0\u0026thinsp;\u0026plusmn;\u0026thinsp;507.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e934.8\u0026thinsp;\u0026plusmn;\u0026thinsp;405.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003et-tau (pg/mL)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e263.7\u0026thinsp;\u0026plusmn;\u0026thinsp;109.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e248.8\u0026thinsp;\u0026plusmn;\u0026thinsp;67.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e249.9\u0026thinsp;\u0026plusmn;\u0026thinsp;92.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e255.5\u0026thinsp;\u0026plusmn;\u0026thinsp;104.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e293.2\u0026thinsp;\u0026plusmn;\u0026thinsp;142.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003ep-tau181 (pg/mL)\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.9\u0026thinsp;\u0026plusmn;\u0026thinsp;12.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.8\u0026thinsp;\u0026plusmn;\u0026thinsp;7.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e23.4\u0026thinsp;\u0026plusmn;\u0026thinsp;9.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24.1\u0026thinsp;\u0026plusmn;\u0026thinsp;11.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e28.4\u0026thinsp;\u0026plusmn;\u0026thinsp;16.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSmall HDL (nmol/L) (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2177.0\u0026thinsp;\u0026plusmn;\u0026thinsp;515.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2360.3\u0026thinsp;\u0026plusmn;\u0026thinsp;518.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2282.9\u0026thinsp;\u0026plusmn;\u0026thinsp;531.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2272.0\u0026thinsp;\u0026plusmn;\u0026thinsp;496.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e1879.6\u0026thinsp;\u0026plusmn;\u0026thinsp;394.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLarge HDL (nmol/L) (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e150.2\u0026thinsp;\u0026plusmn;\u0026thinsp;49.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e178.8\u0026thinsp;\u0026plusmn;\u0026thinsp;44.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e154.1\u0026thinsp;\u0026plusmn;\u0026thinsp;46.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e166.5\u0026thinsp;\u0026plusmn;\u0026thinsp;46.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003e114.0\u0026thinsp;\u0026plusmn;\u0026thinsp;33.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAPOE-ε4 carrier status moderation of HDL-biomarker associations. All regression models included covariates for age, sex, education, and clinical group.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eCSF Small HDL x APOE-ɛ4 on Aβ1\u0026ndash;42\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInteraction term\u003c/p\u003e \u003cp\u003e(β, p-value, FDR-p)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAPOE-ɛ4 non-carrier slope\u003c/p\u003e \u003cp\u003e(β, p-value)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAPOE-ɛ4 carrier slope\u003c/p\u003e \u003cp\u003e(β, p-value)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.41; p\u0026thinsp;=\u0026thinsp;0.025; FDR-p\u0026thinsp;=\u0026thinsp;0.038\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ = -0.04; p\u0026thinsp;=\u0026thinsp;0.687\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ = -0.45; p\u0026thinsp;=\u0026thinsp;0.001\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF Small HDL x APOE-ɛ4 on tau\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.52; p\u0026thinsp;=\u0026thinsp;0.015; FDR-p\u0026thinsp;=\u0026thinsp;0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.23; p\u0026thinsp;=\u0026thinsp;0.058\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eβ = -0.29; p\u0026thinsp;=\u0026thinsp;0.073\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF Small HDL x APOE-ɛ4 on p-tau181\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.40; p\u0026thinsp;=\u0026thinsp;0.054; FDR-p\u0026thinsp;=\u0026thinsp;0.054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.21; p\u0026thinsp;=\u0026thinsp;0.076\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ = -0.19; p\u0026thinsp;=\u0026thinsp;0.229\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF Large HDL x APOE-ɛ4 on Aβ1\u0026ndash;42\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.28; p\u0026thinsp;=\u0026thinsp;0.071; FDR-p\u0026thinsp;=\u0026thinsp;0.071\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.06; p\u0026thinsp;=\u0026thinsp;0.576\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ = -0.22; p\u0026thinsp;=\u0026thinsp;0.075\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF Large HDL x APOE-ɛ4 on tau\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.52; p\u0026thinsp;=\u0026thinsp;0.003; FDR-p\u0026thinsp;=\u0026thinsp;0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.24; p\u0026thinsp;=\u0026thinsp;0.049\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ = -0.28; p\u0026thinsp;=\u0026thinsp;0.046\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF Large HDL x APOE-ɛ4 on p-tau181\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.40; p\u0026thinsp;=\u0026thinsp;0.024; FDR-p\u0026thinsp;=\u0026thinsp;0.036\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.19; p\u0026thinsp;=\u0026thinsp;0.110\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ = -0.20; p\u0026thinsp;=\u0026thinsp;0.143\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"1\" nameend=\"c5\" namest=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCross-sectional results of CSF and plasma HDL levels, APOE, and APOE-ε4 measures with clinical biomarker and cognitive outcomes. All regression models included covariates for age, sex, education, clinical group, and ε4 positivity (in models that did not include ε4 ratio or amount as an independent variable).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eCSF Total APOE\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCSF Aβ1\u0026ndash;42\u003c/p\u003e \u003cp\u003e(β, p-value, FDR-p)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCSF total tau\u003c/p\u003e \u003cp\u003e(β, p-value, FDR-p)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCSF p-tau181\u003c/p\u003e \u003cp\u003e(β, p-value, FDR-p)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRAVLT % forgetfulness\u003c/p\u003e \u003cp\u003e(β, p-value)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.27; p\u0026thinsp;=\u0026thinsp;0.007; FDR-p\u0026thinsp;=\u0026thinsp;0.007\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.74; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.66; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.08; p\u0026thinsp;=\u0026thinsp;0.436\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePlasma Total APOE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.01; p\u0026thinsp;=\u0026thinsp;0.890; FDR-p\u0026thinsp;=\u0026thinsp;0.890\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.03; p\u0026thinsp;=\u0026thinsp;0.788; FDR-p\u0026thinsp;=\u0026thinsp;0.890\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.04; p\u0026thinsp;=\u0026thinsp;0.711; FDR-p\u0026thinsp;=\u0026thinsp;0.890\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eβ = -0.17; p\u0026thinsp;=\u0026thinsp;0.092\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCSF APOE-ɛ4\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.09; p\u0026thinsp;=\u0026thinsp;0.422; FDR-p\u0026thinsp;=\u0026thinsp;0.422\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.49; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.48; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; FDR-p\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.15; p\u0026thinsp;=\u0026thinsp;0.142\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePlasma Total APOE-ɛ4\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ = -0.18; p\u0026thinsp;=\u0026thinsp;0.124; FDR-p\u0026thinsp;=\u0026thinsp;0.372\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ = -0.05; p\u0026thinsp;=\u0026thinsp;0.685; FDR-p\u0026thinsp;=\u0026thinsp;0.978\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eβ\u0026thinsp;=\u0026thinsp;0.003; p\u0026thinsp;=\u0026thinsp;0.978; FDR-p\u0026thinsp;=\u0026thinsp;0.978\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eβ = -0.02; p\u0026thinsp;=\u0026thinsp;0.884\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eThis study was funded in part by the National Institutes of Health/National Institute on Aging (NIH/NIA), RF1AG076124, R01AG055770, R01AG067063, R01AG054434, R21AG056518, and P30AG066530 to H.N.Y.; Alzheimer\u0026rsquo;s Drug Discovery Foundation (ADDF) (GC-201711\u0026ndash;2014197 to H.N.Y.); and donations from the Vranos and Tiny Foundations and Ms. Lynne Nauss to H.N.Y. M.S.R (AARFD-24-1313939) was supported by the Alzheimer\u0026rsquo;s Association Research Fellowship. T.A.P was supported by the NIH (#R01AG075336 and #R01AG073267) and Alzheimer\u0026rsquo;s Association (#AACSF-20-648075). Data collection and sharing for this project was funded by the Alzheimer\u0026rsquo;s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering and through generous contributions from the following: AbbVie, Alzheimer\u0026rsquo;s Association; Alzheimer\u0026rsquo;s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd. and its affiliated company, Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research \u0026amp; Development, LLC.; Johnson \u0026amp; Johnson Pharmaceutical Research \u0026amp; Development LLC.; Lumosity; Lundbeck; Merck \u0026amp; Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions were facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization was the Northern California Institute for Research and Education, and the study was coordinated by the Alzheimer\u0026rsquo;s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eZ.E.T. analyzed the data and wrote the manuscript; I.A. analyzed the data and wrote the manuscript; A.A.T. reviewed data analysis and revised the manuscript; J.P.B.C wrote the manuscript; L.B. wrote the manuscript; A.S. revised the manuscript; G.P.C.G. revised the manuscript; W.J.M. revised the manuscript and provided data analysis feedback; M.N.B. provided oversight and revised the manuscript; D.N. and R.M.K. contributed to sample processing; H.N.Y. designed the study, obtained ADNI approvals, and wrote the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eYassine HN, Hugo C, O'Donovan B, Stephens IO, Johnson LA, Cole G et al. APOE-Targeted Therapeutics for Alzheimer's Disease. J Neurosci. 2025;45(46).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDi Battista AM, Heinsinger NM, Rebeck GW. Alzheimer's Disease Genetic Risk Factor APOE-ε4 Also Affects Normal Brain Function. Curr Alzheimer Res. 2016;13(11):1200\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaek MS, Cho H, Lee HS, Lee JH, Ryu YH, Lyoo CH. Effect of APOE ε4 genotype on amyloid-β and tau accumulation in Alzheimer's disease. Alzheimers Res Ther. 2020;12(1):140.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu M, Kuhel DG, Shen L, Hui DY, Woods SC. Apolipoprotein E does not cross the blood-cerebrospinal fluid barrier, as revealed by an improved technique for sampling CSF from mice. Am J Physiol Regul Integr Comp Physiol. 2012;303(9):R903\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLinton MF, Gish R, Hubl ST, B\u0026uuml;tler E, Esquivel C, Bry WI, et al. Phenotypes of apolipoprotein B and apolipoprotein E after liver transplantation. J Clin Invest. 1991;88(1):270\u0026ndash;81.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFisher EA, Feig JE, Hewing B, Hazen SL, Smith JD. High-density lipoprotein function, dysfunction, and reverse cholesterol transport. Arterioscler Thromb Vasc Biol. 2012;32(12):2813\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartinez AE, Weissberger G, Kuklenyik Z, He X, Meuret C, Parekh T, et al. The small HDL particle hypothesis of Alzheimer's disease. Alzheimers Dement. 2023;19(2):391\u0026ndash;404.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMart\u0026iacute;nez-Morillo E, Hansson O, Atagi Y, Bu G, Minthon L, Diamandis EP, et al. Total apolipoprotein E levels and specific isoform composition in cerebrospinal fluid and plasma from Alzheimer's disease patients and controls. Acta Neuropathol. 2014;127(5):633\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTurri M, Conti E, Pavanello C, Gastoldi F, Palumbo M, Bernini F, et al. Plasma and cerebrospinal fluid cholesterol esterification is hampered in Alzheimer's disease. Alzheimers Res Ther. 2023;15(1):95.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNedelkov D, Tsokolas ZE, Rodrigues MS, Sible I, Han SD, Kerman BE, et al. Increased cerebrospinal fluid and plasma apoE glycosylation is associated with reduced levels of Alzheimer's disease biomarkers. Alzheimers Res Ther. 2025;17(1):151.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlsson A, Vanderstichele H, Andreasen N, De Meyer G, Wallin A, Holmberg B, et al. Simultaneous measurement of beta-amyloid(1\u0026ndash;42), total tau, and phosphorylated tau (Thr181) in cerebrospinal fluid by the xMAP technology. Clin Chem. 2005;51(2):336\u0026ndash;45.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTosun D, Hausle Z, Thropp P, Concha-Marambio L, Lamoureux J, Lebovitz R, et al. Association of CSF α-synuclein seed amplification assay positivity with disease progression and cognitive decline: A longitudinal Alzheimer's Disease Neuroimaging Initiative study. Alzheimers Dement. 2024;20(12):8444\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrane PK, Carle A, Gibbons LE, Insel P, Mackin RS, Gross A, et al. Development and assessment of a composite score for memory in the Alzheimer's Disease Neuroimaging Initiative (ADNI). Brain Imaging Behav. 2012;6(4):502\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMora S, Caulfield MP, Wohlgemuth J, Chen Z, Superko HR, Rowland CM, et al. Atherogenic Lipoprotein Subfractions Determined by Ion Mobility and First Cardiovascular Events After Random Allocation to High-Intensity Statin or Placebo: The Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) Trial. Circulation. 2015;132(23):2220\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYassine HN, Feng Q, Chiang J, Petrosspour LM, Fonteh AN, Chui HC et al. ABCA1-Mediated Cholesterol Efflux Capacity to Cerebrospinal Fluid Is Reduced in Patients With Mild Cognitive Impairment and Alzheimer's Disease. J Am Heart Assoc. 2016;5(2).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Valkenburgh J, Meuret C, Martinez AE, Kodancha V, Solomon V, Chen K, et al. Understanding the Exchange of Systemic HDL Particles Into the Brain and Vascular Cells Has Diagnostic and Therapeutic Implications for Neurodegenerative Diseases. Front Physiol. 2021;12:700847.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRawat V, Wang S, Sima J, Bar R, Liraz O, Gundimeda U, et al. ApoE4 Alters ABCA1 Membrane Trafficking in Astrocytes. J Neurosci. 2019;39(48):9611\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang YA, Zhou B, Nabet AM, Wernig M, S\u0026uuml;dhof TC. Differential Signaling Mediated by ApoE2, ApoE3, and ApoE4 in Human Neurons Parallels Alzheimer's Disease Risk. J Neurosci. 2019;39(37):7408\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVance JM, Farrer LA, Huang Y, Cruchaga C, Hyman BT, Pericak-Vance MA, Report of the APOE4 National Institute on Aging/Alzheimer Disease Sequencing Project Consortium Working Group, et al. Reducing APOE4 in Carriers is a Therapeutic Goal for Alzheimer's Disease. Ann Neurol. 2024;95(4):625\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson KG, Kaplitt M, Kaminsky S, Sondhi D, Amato G, Selvan N, et al. Topline results from Phase 1/2 AAV gene therapy (LX1001) in APOE4/4 homozygotes with Alzheimer\u0026rsquo;s disease. Alzheimer's Dement. 2025;21(S5):e101538.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGolden LR, Siano DS, Stephens IO, MacLean SM, Saito K, Nolt GL, et al. APOE4 to APOE2 allelic switching in mice improves Alzheimer's disease-related metabolic signatures, neuropathology and cognition. Nat Neurosci. 2025;28(12):2461\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuynh TV, Liao F, Francis CM, Robinson GO, Serrano JR, Jiang H, et al. Age-Dependent Effects of apoE Reduction Using Antisense Oligonucleotides in a Model of β-amyloidosis. Neuron. 2017;96(5):1013\u0026ndash;e234.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng K, Tsitsos FN, Batts AJ, Ji R, Nuriel T, Konofagou EE, et al. Focused ultrasound-mediated APOE4 knockdown in mouse brain. Alzheimers Dement. 2025;21(7):e70464.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSidhu JS, Sardu ML, Bj\u0026ouml;rnsson MA, Gallais F, Khandelwal A, Byrnes W, et al. Establishing bepranemab posology through exposure-response modeling and simulation for TOGETHER, a double-blind, placebo-controlled Phase II study of bepranemab in prodromal\u0026ndash;mild Alzheimer\u0026rsquo;s disease (AD). Alzheimer's Dement. 2025;21(S5):e099277.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"alzheimers-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"azrt","sideBox":"Learn more about [Alzheimer's Research and Therapy](http://alzres.biomedcentral.com/)","snPcode":"13195","submissionUrl":"https://submission.nature.com/new-submission/13195/3","title":"Alzheimer's Research \u0026 Therapy","twitterHandle":"@AlzheimersRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"APOE isoforms, brain-peripheral APOE4, HDL, Alzheimer’s","lastPublishedDoi":"10.21203/rs.3.rs-9349576/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9349576/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eApolipoprotein E ε4 (APOE4) is the strongest genetic risk factor for late-onset Alzheimer's disease (AD), yet whether its pathogenic effects are driven by HDL lipidation state or by elevated APOE4 levels in the CNS remains unclear. Using ADNI data, we analyzed cerebrospinal fluid (CSF) and plasma small and large HDL particle concentrations, total APOE levels, and isoform-specific APOE3 and APOE4 protein levels in 144 participants with APOE ε3/ε3, ε3/ε4, or ε4/ε4 genotypes, grouped by cognitive state and memory progression over 4 years. Memory loss was defined as \u0026ge;\u0026thinsp;10% decline on the Rey Auditory Verbal Learning Test (RAVLT) delayed recall between baseline and 48-month follow-up. Cross-sectional associations with CSF Aβ1\u0026ndash;42, total tau, and p-tau181 were evaluated using covariate-adjusted linear regression, and longitudinal trajectories were examined using linear mixed-effects models over 6 years. CSF small and large HDL levels were higher in cognitively normal individuals and non-progressors, while APOE4 carriers exhibited reduced CSF small HDL relative to ε3 homozygotes. Importantly, APOE4 status moderated HDL\u0026ndash;biomarker associations in opposing directions: in carriers, higher CSF small HDL was associated with lower tau and Aβ1\u0026ndash;42, whereas in non-carriers, higher CSF small HDL was associated with higher CSF tau levels. Higher CSF APOE4 protein levels were associated with elevated tau and p-tau181, while plasma APOE measures showed minimal and often opposing associations with CSF biomarkers. Longitudinally, higher baseline CSF APOE4 proportion was associated with greater memory decline over 6 years but did not predict biomarker change. These findings argue against a protective role of APOE4 lipidation in AD and instead support a CNS-compartment model in which APOE4 protein burden aligns more closely with tau pathology and cognitive decline than peripheral APOE or HDL measures.\u003c/p\u003e","manuscriptTitle":"Beyond Lipidation: CSF APOE4 Protein Burden, Not HDL Subclass, Drives Tau Associations in APOE4 Alzheimer’s Disease","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-24 00:42:32","doi":"10.21203/rs.3.rs-9349576/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-29T03:52:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"314555296090375981661093227664447198922","date":"2026-04-16T16:29:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"143228518977842368199224505571169364531","date":"2026-04-15T13:14:20+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-15T11:46:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-09T08:45:35+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-09T08:45:22+00:00","index":"","fulltext":""},{"type":"submitted","content":"Alzheimer's Research \u0026 Therapy","date":"2026-04-07T22:13:45+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"alzheimers-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"azrt","sideBox":"Learn more about [Alzheimer's Research and Therapy](http://alzres.biomedcentral.com/)","snPcode":"13195","submissionUrl":"https://submission.nature.com/new-submission/13195/3","title":"Alzheimer's Research \u0026 Therapy","twitterHandle":"@AlzheimersRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"484eefc1-9a6d-4450-9050-d6238028c8b4","owner":[],"postedDate":"April 24th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-04-29T03:52:46+00:00","index":25,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-24T00:42:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-24 00:42:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9349576","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9349576","identity":"rs-9349576","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}