ABCA7‐80 moderates vascular stiffness–p‐tau217 association in older African Americans

Abstract

Introduction

Compromised vascular health is increasingly linked to Alzheimer's disease (AD) risk. Estimated pulse wave velocity (ePWV), derived from age and blood pressure, and provides a practical, non‐invasive index of vascular stiffness and overall vascular health. Older African Americans experience a disproportionate burden of vascular disease and AD. Genetic risk factors such as APOE ε4 and ABCA7‐80 (rs115550680) further increase AD susceptibility. However, whether these genetic risks influence vascular stiffness and how that may interact with AD pathology remains unknown, especially in African Americans.

Methods

A total of 143 older African Americans (mean age = 71.10 ± 6.83 years; 109 women) were included. We examined the effects of both ABCA7‐80 and APOE ε4 genotypes on ePWV and plasma phosphorylated tau 217 (p‐tau217). All regression models controlled for sex, education, pulse pressure, waist‐to‐hip ratio, global cognitive status, hypertension status, and APOE genotype (APOE genotype only used for ABCA7‐80 regression models).

Results

ABCA7‐80 risk allele carriers exhibited higher ePWV (F _(1,130) = 8.16, p = 0.005, η² _p = 0.064) and higher p‐tau217 levels (F _(1,130) = 30.11, p < 0.001, η² _p  = 0.201). APOE ε4 allele carriers also showed higher p‐tau217 levels (F _(1,131) = 12.96, p < 0.001, η² _p = 0.092). ABCA7‐80 significantly moderated the relationship between ePWV and p‐tau217 (F_(1,130) = 6.58, p < 0.001), such that higher ePWV was associated with higher p‐tau217 among ABCA7‐80 risk carriers (β = 0.52, t ₍₁₃₀₎ = 2.69, p = 0.008).

Discussion

ABCA7‐80 risk, but not APOE ε4, heightens susceptibility to the tau‐related effects of compromised vascular health among older African Americans. These findings identify a genetically vulnerable subgroup in which vascular stiffness may disproportionately accelerate AD‐related tau pathology and highlight vascular health as a modifiable target for reducing AD risk in African Americans.

Highlights

• African American ABCA7‐80 risk allele carriers show poorer vascular health.

• African American ABCA7‐80 risk allele carriers show higher plasma p‐tau217.

• ABCA7‐80 moderates the vascular health–p‐tau217 relationship.

• APOE ε4 does not modify the vascular health–p‐tau217 association.

• Compromised vascular health may accelerate tau pathology in vulnerable individuals.

1. INTRODUCTION

Alzheimer's disease (AD) is a multifactorial disorder in which vascular aging contributes to early neurodegenerative processes. ¹ Beyond traditional amyloid‐centric models, growing evidence suggests that vascular dysfunction contributes to tau pathology, neuroinflammation, and white matter damage before cognitive impairment emerges. ² These vascular contributions are particularly relevant for older African Americans, who experience a disproportionate burden of cerebrovascular disease, hypertension, and AD. ³

Vascular stiffness is a hallmark of compromised vascular health and reflects cumulative damage to vascular structure and function. Reduced vascular compliance increases transmission of pulsatile forces to cerebral circulation, promoting neurovascular damage. ⁴ Estimated pulse wave velocity (ePWV), derived from age and blood pressure, provides a validated and scalable index of vascular stiffness suitable for community and epidemiological studies, when carotid‐femoral pulse wave velocity (PWV) is not available. ⁵ , ⁶ , ⁷ Accelerated increases in vascular stiffness are common among older African Americans and may represent a key mechanism linking cardiovascular risk to neurodegeneration. ⁸ , ⁹ However, the extent to which vascular stiffness contributes to early AD‐related biological changes in this population remains incompletely understood.

In our prior work, we demonstrated that greater vascular stiffness, indexed by higher ePWV, was robustly associated with higher plasma concentrations of phosphorylated tau 217 (p‐tau217), but not the amyloid beta 42/40 ratio, among cognitively unimpaired older African Americans. ¹⁰ Plasma p‐tau217 has emerged as one of the most sensitive and specific blood‐based biomarkers of AD, showing strong correspondence with tau positron emission tomography (PET), neurodegeneration, and longitudinal cognitive decline, even in preclinical stages. ¹¹ Taken together, these findings suggest that vascular stiffness may preferentially relate to tau‐linked neurodegenerative processes rather than early amyloid‐associated changes.

Despite similar cardiovascular risk profiles, substantial variability exists in vascular health and biomarker expression. ¹² Genetic susceptibility may contribute to this heterogeneity, particularly through AD risk alleles involved in lipid metabolism, vascular integrity, and inflammatory signaling. ¹³ , ¹⁴ Two relevant AD risk variants in African ancestry populations are apolipoprotein ε4 rs429358 and rs7412 (APOE ε4) and ATP‐binding cassette, subfamily A, member 7 rs115550680 (ABCA7‐80). ¹⁵ While APOE ε4 is the most established genetic risk factor for late‐onset AD, its effects appear to be more strongly linked to amyloid‐related pathways. ¹⁶ In contrast, ABCA7 risk variants, especially ABCA7‐80, are associated with a comparatively larger effect on AD risk in populations of African ancestry and have been implicated in lipid transport, immune regulation, and vascular and microglial function, suggesting a potentially distinct mechanistic pathway to neurodegeneration. ¹⁷ Importantly, ABCA7‐80 has been associated with cerebrovascular dysfunction and altered lipid handling, processes that may interact directly with vascular stiffness to influence tau‐related neurodegeneration. ¹⁸ These findings raise the possibility that genetic risk modifies the relationship between vascular dysfunction and neurodegenerative processes rather than exerting only independent effects on biomarker levels. ¹⁹ Identifying such genetic–vascular interactions could help explain why some individuals exhibit heightened biological vulnerability in the context of vascular stiffness while others remain relatively resilient.

Building on our prior findings, we investigated how AD genetic risk influenced vascular stiffness and its association with tau pathology in cognitively unimpaired older African Americans. Specifically, we examined the effects of ABCA7‐80 and APOE ε4 on ePWV and plasma p‐tau217 and tested whether these genetic variants moderated the relationship between vascular stiffness and biomarker levels. We hypothesized that AD risk alleles, particularly ABCA7‐80, would be associated with higher vascular stiffness and higher plasma biomarker concentrations and that genetic risk would strengthen the association between ePWV and tau‐related neurodegeneration. By integrating vascular, genetic, and plasma biomarker data, this work aims to improve understanding of biologically vulnerable subgroups at risk for AD‐related neurodegeneration in African American populations.

2. METHODS

2.1. Participants

Participants were enrolled from an ongoing Rutgers University–Newark observational study of brain aging and AD risk in cognitively unimpaired older African Americans. Recruitment occurred through the Aging & Brain Health Alliance. Eligible participants were ≥60 years old and free of neurodegenerative disease, dementia medication use, recent general anesthesia, substance misuse, or inability to provide biological samples. Cognitively unimpaired status was defined by the absence of mild cognitive impairment or dementia. All participants provided written informed consent, and study procedures were approved by the Rutgers University Institutional Review Board and conducted in accordance with the Declaration of Helsinki.

2.2. Procedure

Eligibility was assessed by telephone. Eligible individuals attended an in‐person visit for written informed consent, color vision screening, and global cognitive assessment, followed by a laboratory session including cognitive testing and saliva and blood collection.

2.3. Measures

2.3.1. Demographics

Participant characteristics included chronological age (years), self‐reported sex, and total years of formal education.

2.3.2. Global cognitive status

Global cognitive status was assessed using the Montreal Cognitive Assessment (MoCA) as a brief screening measure of overall cognitive status. The MoCA assesses multiple cognitive domains, including visuospatial and executive abilities, naming, learning and memory, attention, language, abstraction, and orientation. Higher total scores indicate better global cognitive status. ²⁰ , ²¹

2.3.3. Anthropometrics

Waist and hip circumferences were measured using standardized World Health Organization procedures, ²² and waist‐to‐hip ratio (WHR) was calculated as an index of central adiposity.

WHR was selected as the primary anthropometric indicator because central adiposity is associated with vascular stiffness, cardiometabolic risk, and neurovascular vulnerability in African American populations. ²³

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed the literature using PubMed and recent conference presentations related to vascular stiffness, ePWV, plasma p‐tau217, APOE ε4, and ABCA7 genetic risk in AD, with an emphasis on African ancestry populations. Prior studies linked vascular dysfunction to tau‐related neurodegeneration and identified ABCA7‐80 as an important AD risk factor among African Americans. However, whether genetic risk modifies vascular–tau relationships remains poorly understood.

2. Interpretation: This study demonstrates that ABCA7‐80, but not APOE ε4, moderates the relationship between vascular stiffness (ePWV) and plasma p‐tau217 in older African Americans, identifying a genetically vulnerable subgroup in which vascular dysfunction may accelerate tau‐related neurodegeneration.

3. Future directions: Future longitudinal studies integrating direct vascular stiffness measures, cerebral perfusion imaging, plasma biomarkers, and genetic profiling are needed to clarify causal mechanisms linking vascular aging and tau pathology in African ancestry populations.

2.3.4. Genetic procedures

Saliva samples were collected using DNA Genotek® Oragene (ORG‐600) kits. Samples were transported to an off‐site facility for genotyping. The genotyping process involved quantitative polymerase chain reaction (PCR) on Eppendorf Master cyclers using the TaqMan custom genotyping assay.

Following genotyping, participants were classified as either APOE ε4 allele carriers or non‐carriers, regardless of hetero‐ or homozygosity. ²⁴ All APOE ε4 allele heterozygotes were included in the study, even those with the ε2/ε4 genotype, as the presence of the ε4 allele alone confers an increased risk for AD. ²⁵ Given evidence suggesting that ε2/ε4 carriers may represent a biologically distinct subgroup, sensitivity analyses excluding ε2/ε4 individuals were additionally conducted for all APOE‐related analyses.

The ABCA7 rs115550680 genotyping was performed using a quantitative PCR, identifying “G” allele carriers as ABCA7‐80 risk allele carriers. ²⁶

2.3.5. Blood pressure assessment

Arterial blood pressure was assessed using standardized clinical protocols. ²⁷ Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured using an automated oscillometric device (Yuwell YE660E Electronic Blood Pressure Monitor; Yuwell, Jiangsu, China). Measurements were obtained with participants seated comfortably, feet placed flat on the floor and the back supported, following a brief period of rest. Participants were instructed to remain still and refrain from speaking throughout the assessment. Three standardized resting blood pressure measurements were obtained and averaged for the present analyses, consistent with the parent cohort assessment protocol.

Hypertension status was determined in accordance with American Heart Association criteria, with SBP values ≥ 130 mmHg classified as hypertensive. ²⁸

2.3.6. ePWV

Vascular stiffness was quantified using ePWV, a validated surrogate marker of vascular health derived from chronological age and mean arterial pressure (MAP). ⁵ , ²⁹

MAP was calculated using the formula DBP+0.4×PP rather than the traditional one‐third pulse pressure approximation. Prior methodological work demonstrated that the traditional DBP+0.33×PP formula systematically underestimates brachial MAP, whereas the 0.4 form factor correction more accurately reflects intra‐arterial pressure measurements and shows stronger associations with vascular stiffness and target organ damage. ³⁰ , ³¹ Consistent with the initial ePWV derivation and subsequent validation studies, ⁵ , ³² , ³³ , ³⁴ the present analyses used the validated MAP estimation approach incorporating the 0.4 PP form factor correction.

ePWV provides a non‐invasive index of arterial stiffening, with higher values indicating poorer vascular health and greater cardiometabolic and neurovascular risk. Pulse pressure (PP), MAP, and ePWV were computed using the following equations, where age is expressed in years and MAP in mmHg:

• PP = {SBP–DBP},

• MAP = DBP + 0.40 * {PP},

• ePWV = 9.587–{0.402 * age} + {4.560 * 0.001 * age²}—{2.621* 0.00001 * age² * MAP} + {3.176 * 0.001 * age * MAP}.

2.3.7. Blood‐based biomarkers

Fasting blood samples were obtained to measure circulating plasma biomarkers reflecting AD‐related neuropathological processes, including tau phosphorylation. Blood draws were collected in EDTA tubes (6 to 10 mL). Samples were processed within 1 h of collection by centrifugation, after which plasma was aliquoted into 0.5‐mL polypropylene tubes and stored at −80°C until shipment for biomarker analysis.

Plasma p‐tau217 (Elahi Laboratory)

An additional 0.5‐mL EDTA plasma aliquot per participant was sent on dry ice to the Icahn School of Medicine at Mount Sinai for assessment of plasma p‐tau217. Plasma p‐tau217 levels (pg/mL) were measured using an ultrasensitive single molecule array (Simoa) assay on the LucentAD Diagnostics platform, following previously published analytical procedures. ³⁵

Hemoglobin A1c (HbA1c)

Heparinized blood was collected via a single drop from the EDTA tube and processed via the A1CNow® test kit.

2.4. Statistical analysis

Statistical analyses were performed using Jamovi (version 2.7.31.0). Initial data screening included inspection of descriptive statistics (means, standard deviations) and graphical evaluations of variable distributions to identify skewness and potential outliers. Plasma biomarker variables exhibited positive skew and were natural log‐transformed (LN) prior to inferential analyses to improve distributional properties and satisfy model assumptions. All plasma biomarker analyses were conducted using LN‐transformed values.

Normality assumptions were evaluated through visual inspection of residuals and examination of skewness and kurtosis following transformation. The proportion of missing data was low (<5%), and analyses retained all available observations without listwise deletion.

Prior to primary analyses, Hardy‐Weinberg equilibrium (HWE) ³⁶ was assessed for ABCA7‐80 (rs115550680) and APOE (rs7412 and rs429358) using chi‐squared goodness‐of‐fit and exact tests. Observed and expected genotype frequencies were compared, and inbreeding coefficients (F) were calculated. HWE testing was conducted using a custom Python‐based pipeline (pandas, scipy) applied to the full sample.

Primary analyses used analysis of covariance and multiple linear regression–based moderation models to evaluate the effects of ABCA7‐80 and APOE ε4 on ePWV and plasma p‐tau217 and to test whether genetic risk modified vascular–biomarker associations. Models were adjusted for sex, education, pulse pressure, WHR, MoCA, and hypertension. APOE genotype was additionally included in ABCA7‐80 models. Chronological age was not included as a separate covariate because age is directly incorporated into the validated ePWV calculation formula and constitutes a major component of the composite vascular stiffness metric. Pulse pressure was nevertheless retained as a covariate to account for the independent contribution of pulsatile hemodynamic load beyond the composite ePWV measure. MoCA, hypertension status, WHR, sex, and education were included as relevant cognitive, vascular, metabolic, and demographic covariates. HbA1c was additionally included in sensitivity analyses among participants with available data (N = 103).

Model fit (R ²), effect estimates, and p values were reported. Post hoc comparisons were Bonferroni‐corrected, with significance defined as two‐tailed p < 0.05.

3. RESULTS

3.1. Participant characteristics

Demographic and clinical characteristics are summarized in Table 1. The final sample included 143 cognitively unimpaired older African American adults (mean age = 71.10 ± 6.83 years; 109 women). Participants had a mean of 14.06 ± 4.73 years of education and a mean MoCA score of 27.36 ± 1.90, consistent with intact global cognition. Among all participants, 138 had ABCA7‐80, and 143 of them had APOE ε4 samples. SBP, DBP, pulse pressure, hypertension status, ePWV, WHR, LN‐transformed concentration, raw values of plasma biomarkers (p‐tau217 [N = 143]), and HbA1c (N = 103) are also reported in Table 1.

TABLE 1.

Distribution of demographic data.

		Mean ± SD	Min / Max
Demographics
Age (years)		71.10 ± 6.83	60 / 92
Sex (N [%])
Women		109 (76.2)	—
Men		34 (23.8)	—
Education (years)		14.06 ± 4.73	7 / 20
Cognition
MoCA		27.36 ± 1.90	20 / 30
Genetics
ABCA7‐80 (N [%])
Risk allele carriers		21 (15.2)	—
Non‐risk carriers		117 (84.8)	—
ABCA7‐80 (N [%])
Risk allele carriers	GG	12 (8.7)	—
	AG	9 (6.5)	—
Non‐risk carriers	AA	117 (84.8)	—
APOE ε4 (N [%])
ε4 allele carriers		52 (36.4)	—
ε4 allele non‐carriers		91 (63.6)	—
APOE ε4 subtypes (N [%])
ε4 allele carriers	ε4ε4	29 (20.3)	—
	ε3ε4	14 (9.8)	—
	ε2ε4	9 (6.3)	—
ε4 allele non‐carriers	ε3ε3	66 (46.2)	—
	ε2ε3	24 (16.8)	—
	ε2ε2	1 (0.7)	—
Blood Pressure
Systolic BP (mmHg)		142.62 ± 20.32	99 / 203
Diastolic BP (mmHg)		79.51 ± 10.14	55 / 106
Pulse pressure (mmHg)		63.11 ± 16.83	30 / 115
HTN status (N [%])
Normotensive		71 (49.7)	—
Hypertensive		72 (50.3)	—
Vascular stiffness
ePWV (m/s)		11.99 ± 1.61	4.87/17.43
Anthropometrics
Waist circumference (cm)		100.52 ± 18.60	39/156
Hip circumference (cm)		111.44 ± 18.10	43/147
Waist‐to‐hip ratio		0.90 ± 0.42	0.33 / 1.27
Plasma biomarkers
LN‐transformed concentrations *
p‐tau 217		−1.30 ± 0.59	−2.81 / 0.62
Original (raw) values pg/mL)		0.33 ± 0.26	0.06 / 1.87
p‐tau 217
Glycemic measure
HbA1c (%)		6.03 ± 1.04	4.60 / 11.50

Abbreviations: %, percentage; BP, blood pressure; ePWV, estimated pulse wave velocity; HbA1c, hemoglobin A1c; HTN, hypertension; Max, maximum; Min, minimum; MoCA, Montreal Cognitive Assessment; N, number; p‐tau, phosphorylated tau; SD, standard deviation.

^*Plasma biomarker values are natural log–transformed (LN) unless otherwise indicated.

HWE testing revealed significant deviations for ABCA7‐80 (χ ² = 65.75, p < 0.001, F = 0.69) and APOE (χ ² = 69.18, p < 0.001, F = 0.80), likely reflecting admixture and enrichment for AD genetic risk in this cohort. These findings are consistent with prior reports in African American populations (Figures S1 and S2).

3.2. Effect of ABCA7‐80 genotype on ePWV and plasma biomarkers

Among all participants, 138 had ABCA7‐80 samples, and 21 of them were risk allele carriers. ABCA7‐80 risk allele carriers exhibited higher ePWV (F _(1,130) = 8.16, p = 0.005, η² _p = 0.064, Cohen's d = 0.74; Figure 1), reflecting a moderate effect size, and higher plasma p‐tau217 levels (F _(1,130) = 30.11, p < 0.001, η² _p = 0.201, Cohen's d = 1.43; Figure 2), reflecting a large effect size.

FIGURE 1.

ABCA7‐80 risk allele carriers exhibited higher ePWV than non‐carriers. Note: ePWV values are shown for ABCA7‐80 risk allele carriers (N = 21) and non‐risk allele carriers (N = 117). Violin plots depict the distribution of ePWV within each genotype group, with embedded boxplots indicating the median and interquartile range. Individual points represent single participants. Higher ePWV values reflect poorer vascular health. Group differences reflect covariate‐adjusted analyses controlling for demographic, vascular, and cognitive factors.

FIGURE 2.

ABCA7‐80 risk allele carriers exhibited higher plasma p‐tau 217 levels than non‐carriers. Note: Plasma p‐tau217 values are natural log–transformed. Violin plots illustrate the distribution of p‐tau217 within each genotype group, with embedded boxplots indicating the median and interquartile range. Individual points represent single participants (risk allele carriers, N = 21; non‐risk allele carriers, N = 117). Group differences reflect covariate‐adjusted analyses controlling for demographic, vascular, and cognitive factors.

Sensitivity analyses excluding ε2/ε4 carriers (N = 9), while additional adjustment for APOE status yielded substantially similar findings. ABCA7‐80 risk allele carriers continued to exhibit higher plasma p‐tau217 levels (F _(1,130) = 24.553, p < 0.001, η² _p  = 0.172, Cohen's d = 1.362), reflecting a large effect size, and higher ePWV values (F _(1,130) = 4.890, p = 0.029, η² _p  = 0.040, Cohen's d = 0.608), reflecting a moderate effect size.

Additional sensitivity analyses including HbA1c as a covariate yielded substantially similar findings. ABCA7‐80 risk allele carriers continued to exhibit higher ePWV values (F _(1,93) = 8.230, p = 0.005, η² _p  = 0.081, Cohen's d = 0.837) and higher plasma p‐tau217 levels (F _(1,93) = 26.415, p < 0.001, η² _p  = 0.221, Cohen's d = 1.499) after adjustment for HbA1c and APOE status.

3.3. Effect of APOE ε4 genotype on ePWV and plasma biomarkers

Among all participants, 143 had APOE ε4 allele samples, and 52 of them were risk allele carriers. APOE ε4 allele carriers exhibited significantly higher plasma p‐tau217 levels than non‐carriers after covariate adjustment (F _(1,131) = 12.96, p < 0.001, η² _p  = 0.092; Cohen's d = 0.65; Figure 3), reflecting a moderate effect size. APOE ε4 carrier status was not associated with ePWV after covariate adjustment (F _(1,131) = 0.06, p = 0.812, η² _p  < 0.001, Cohen's d = 0.03), indicating a negligible effect size.

FIGURE 3.

APOE ε4 allele carriers exhibited higher plasma p‐tau217 levels than non‐carriers. Note: Plasma p‐tau217 values are natural log–transformed. Violin plots illustrate the distribution of p‐tau217 for APOE ε4 allele carriers (N = 52) and non‐carriers (N = 91), with embedded boxplots indicating the median and interquartile range. Individual points represent single participants. Group differences reflect covariate‐adjusted analyses controlling for demographic, vascular, and cognitive factors.

Sensitivity analyses excluding ε2/ε4 carriers (N = 9) yielded substantially similar findings. APOE ε4 carrier status remained significantly associated with higher plasma p‐tau217 levels (F _(1,123) = 8.315, p = 0.005, η²p = 0.065, Cohen's d = 0.557), reflecting a moderate effect size, while no significant association with ePWV was observed (F _(1,123) = 0.115, p = 0.735, η²p = 0.001, Cohen's d = ‐0.065), indicating a negligible effect size.

Additional sensitivity analyses including HbA1c as a covariate demonstrated similar results. APOE ε4 carrier status remained significantly associated with higher plasma p‐tau217 levels (F _(1,74) = 4.560, p = 0.036, η² _p  = 0.058, Cohen's d = 0.507), whereas no significant association with ePWV was observed (F _(1,94) = 0.019, p = 0.890, η² _p  < 0.001).

3.4. Moderation effect of ABCA7‐80 genotype on ePWV and plasma biomarker relationship

The overall model was significant (R ² = 0.36, adjusted R ² = 0.30, F _(1,130) = 6.58, p < 0.001). ABCA7‐80 risk status significantly moderated the relationship between ePWV and plasma p‐tau217 (F _(1,130) = 4.40, p = 0.038, η² _p  = 0.036), reflecting a small to moderate interaction effect. Simple effects analyses indicated that higher ePWV was associated with higher plasma p‐tau217 among ABCA7‐80 risk carriers (β = 0.52, t ₍₁₃₀₎ = 2.69, p = 0.008, η² _p  = 0.065), reflecting a moderate effect size, whereas no significant association was observed among non‐carriers (β = 0.10, t₍₁₃₀₎ = 0.88, p = 0.382, η² _p  = 0.010), indicating a negligible effect size. Pulse pressure, WHR, sex, education, MoCA score, and hypertension status were not significantly associated with p‐tau217 (all p > 0.05). The model is shown in Figure 4.

FIGURE 4.

ABCA7‐80 moderates the association between vascular health and plasma p‐tau217. Note: The figure illustrates the interaction between ABCA7‐80 genotype and ePWV in predicting plasma p‐tau217 levels (natural log–transformed). Individual points represent participants (N = 138). Solid lines depict fitted regression slopes for ABCA7‐80 risk allele carriers (N = 21) and non‐risk allele carriers (N = 117). Shaded bands indicate the standard error of the predicted mean. Higher ePWV was associated with higher plasma p‐tau217 levels among ABCA7‐80 risk carriers, whereas no significant association was observed among non‐carriers.

Sensitivity analyses excluding ε2/ε4 carriers (N = 9), while additional adjustment for APOE status demonstrated that ABCA7‐80 risk status continued to significantly moderate the relationship between ePWV and plasma p‐tau217 (F _(1,130) = 4.878, p = 0.029, η² _p = 0.040), reflecting a small to moderate interaction effect.

Additional moderation analyses including HbA1c and APOE status as covariates demonstrated that ABCA7‐80 risk status continued to moderate the relationship between ePWV and plasma p‐tau217 at a trend‐level significance (F_(1,91)  = 3.233, p = 0.076, η² _p = 0.034), reflecting a small interaction effect. Simple effects analyses demonstrated that higher ePWV remained significantly associated with higher plasma p‐tau217 among ABCA7‐80 risk carriers (β = 0.505, t₍₉₁₎  = 2.380, p = 0.019, η² _p  = 0.059), whereas no significant association was observed among non‐carriers (β = 0.088, t₍₉₁₎  = 0.649, p = 0.518, η² _p  = 0.005).

3.5. Moderation effect of APOE ε4 genotypes on ePWV and plasma biomarkers relationship

APOE ε4 did not significantly moderate the association between vascular health (ePWV) and plasma p‐tau217 (F _(1,131) = 0.09, p = 0.759, η² _p  = 0.001), indicating a negligible interaction effect. This interaction effect was small in magnitude, indicating minimal evidence that APOE ε4 modifies vascular–biomarker relationships.

Similarly, sensitivity analyses excluding ε2/ε4 carriers (N = 9) demonstrated that APOE ε4 did not significantly moderate the relationship between ePWV and plasma p‐tau217 (F _(1,121) = 0.118, p = 0.732, η² _p  = 0.001), also indicating a negligible interaction effect.

Additional moderation analyses including HbA1c as a covariate demonstrated that APOE ε4 significantly moderated the relationship between ePWV and plasma p‐tau217 (F _(1,92) = 4.621, p = 0.034, η² _p  = 0.048), reflecting a small interaction effect (Figure S3). Simple effect analyses demonstrated that higher ePWV was significantly associated with higher plasma p‐tau217 among APOE ε4 carriers (β = 0.742, t ₍₉₂₎ = 3.346, p = 0.001, η² _p  = 0.109), whereas the association was not significant among non‐carriers (β = 0.260, t ₍₉₂₎ = 1.920, p = 0.058, η² _p  = 0.039).

4. DISCUSSION

In this study, we extend our prior work demonstrating that vascular stiffness is associated with tau‐related neurodegeneration among cognitively unimpaired older African Americans by showing that genetic risk modifies these vascular–biomarker relationships. ¹⁰ Specifically, ABCA7‐80 risk was associated with higher vascular stiffness, higher plasma p‐tau217, and a stronger ePWV–p‐tau217 relationship. In contrast, APOE ε4 showed associations with p‐tau217 but did not exhibit comparable effects on vascular stiffness. Although primary analyses did not demonstrate significant moderation effects on the relationship between ePWV and plasma p‐tau217, additional sensitivity analyses including HbA1c as a metabolic covariate demonstrated a modest APOE ε4×ePWV interaction effect. These findings suggest that ABCA7‐80 amplifies vulnerability to vascular‐driven tau‐related neurodegeneration.

Our prior findings demonstrated that higher ePWV was associated with plasma p‐tau217, but not with Aβ42/40, ¹⁰ suggesting that vascular stiffness preferentially relates to tau‐related neurodegeneration rather than amyloid pathology. This pattern supports a model in which compromised vascular health contributes to vulnerability to tau‐related neurodegeneration independent of amyloid accumulation. Our findings extend prior work by suggesting that vascular stiffness may contribute to tau‐related neurodegeneration independent of amyloid accumulation ³⁷ and highlight vascular health as a potentially modifiable AD prevention target, particularly in African Americans. ³⁸ Future studies comparing emerging ultra‐sensitive platforms may further improve assay standardization and clinical translation. Consistent with prior work, vascular stiffness appears more strongly associated with tau pathology and neurodegeneration than with amyloid burden in cognitively unimpaired or early‐stage populations. ³⁹ The current results refine this model by demonstrating that genetic risk contributes to heterogeneity in vascular–neurodegenerative relationships. ⁴⁰

Individuals carrying the ABCA7‐80 risk allele exhibited both higher vascular stiffness and stronger associations between ePWV and p‐tau217, indicating heightened biological sensitivity to compromised vascular health. Importantly, sensitivity analyses excluding ε2/ε4 carriers yielded comparable findings, suggesting that the observed ABCA7‐80 effects and vascular–tau interactions were not driven by APOE subgroup classification. Pulse pressure reflects vascular stiffness and pulsatile vascular load. Adjustment for pulse pressure allowed us to determine whether ePWV contributes information beyond conventional hemodynamic measures. The persistence of ABCA7‐80 effects supports the relevance of ePWV as a marker of vascular aging and vascular–brain relationships.

Vascular stiffness is associated with reduced perfusion, endothelial dysfunction, blood–brain barrier disruption, and impaired perivascular clearance. ⁵ , ⁴¹ These processes may promote tau phosphorylation and neurodegeneration. Our findings suggest that ABCA7‐80 carriers may be particularly vulnerable to vascular‐driven neurodegeneration. ⁴²

ABCA7‐80 plays a key role in lipid metabolism, immune regulation, and membrane dynamics, processes closely linked to vascular integrity. ⁴³ In individuals of African ancestry, ABCA7 variants have been associated with cerebrovascular vulnerability, microglial dysregulation, and impaired phagocytic clearance. ¹⁷ , ⁴³ These mechanisms may increase vulnerability to tau‐related neurodegeneration in the context of vascular dysfunction. Our findings support a vascular pathway linking ABCA7‐80 risk and tau‐related neurodegeneration in older African Americans. ⁴⁴ Additional cardiometabolic factors, including diabetes‐related mechanisms and broader cardiovascular risk burden, may further contribute to vascular–tau relationships in older African Americans. Future studies incorporating composite vascular risk indices, such as Framingham Risk Scores, may help refine our understanding of the metabolic and vascular contributors to tau‐related neurodegeneration.

Although APOE ε4 is the most established genetic risk factor for late‐onset AD, its effects are often interpreted within amyloid‐centered frameworks. ⁴⁵ In the present study, APOE ε4 was associated with higher p‐tau217 levels but did not show significant associations with vascular stiffness in the primary analyses. Although APOE ε4 did not significantly moderate the relationship between ePWV and plasma p‐tau217 in the primary models, additional sensitivity analyses including HbA1c as a metabolic covariate demonstrated a modest APOE ε4×ePWV interaction effect. While prior work using carotid–femoral pulse wave velocity (cfPWV) demonstrated increased arterial stiffness among APOE ε4 carriers, ⁴⁶ interactions between APOE ε4 and vascular stiffness were not consistently linked to cognitive decline. ⁴⁷ Moreover, most prior studies were conducted in predominantly White populations, limiting generalizability. These findings suggest that APOE‐related vascular effects may be less prominent in African American populations and highlight alternative pathways, such as ABCA7. The emergence of a modest APOE‐related interaction effect after adjustment for HbA1c further suggests that metabolic factors may partially influence vascular–tau relationships involving APOE‐related risk.

Recent work suggests that APOE ε4 is associated with elevated tau biomarkers independent of large‐artery stiffness, indicating that APOE‐related tau pathology may arise through amyloid‐dependent or lipid‐related mechanisms rather than vascular stiffness. ¹¹ Together, these findings underscore the limitations of extrapolating APOE‐centered models across ancestrally diverse populations and reinforce the importance of considering both genetic and metabolic factors when evaluating vascular contributions to tau‐related neurodegeneration.

Our findings suggest that combined vascular and genetic markers may help improve understanding of heterogeneity in AD‐related neurodegenerative vulnerability among cognitively unimpaired older African Americans. ⁴⁸ However, given the cross‐sectional nature of the present study, future longitudinal investigations are necessary to determine whether these relationships have predictive utility for progression of tau pathology or cognitive decline.

Several limitations should be noted. First, the cross‐sectional design precluded making causal inferences. Second, ePWV is an estimated rather than direct measure of vascular stiffness; future studies should incorporate cfPWV, central pressure pulsatility, and cerebral perfusion measures. ⁴⁹ , ⁵⁰ Third, the modest number of ABCA7‐80 carriers may limit the stability of interaction estimates. Fourth, genetic analyses were limited to selected variants and should be expanded to polygenic approaches. Fifth, broader cardiometabolic measures, including diabetes status and composite cardiovascular risk scores, should be examined. Finally, although p‐tau217 was measured using a sensitive Simoa‐based assay, future studies should evaluate cross‐platform consistency using emerging technologies.

In conclusion, this study demonstrates that ABCA7‐80 risk, but not APOE ε4, modifies the relationship between vascular stiffness and tau‐related neurodegeneration among cognitively unimpaired older African Americans. These findings identify ABCA7‐80 as a genetic factor that may shape vascular–tau associations in cognitively unimpaired older African Americans. Future longitudinal studies integrating vascular, genetic, and plasma biomarker measures will be needed to clarify the temporal and mechanistic relationships underlying these associations.

AUTHOR CONTRIBUTIONS

Conceptualization: Miray Budak, Kevin S. Heffernan, Soodeh Moallemian, Bernadette A. Fausto, and Mark A. Gluck. Methodology: Miray Budak, Kevin S. Heffernan, and Mark A. Gluck. Formal analysis and investigation: Miray Budak, Martina Ishaq, Demiana Abdalla, Soodeh Moallemian, and Victoria Paruzel. Writing – original draft preparation: Miray Budak, and Kevin S. Heffernan. Writing – review and editing: Kevin S. Heffernan, Martina Ishaq, Victoria Paruzel, Demiana Abdalla, Soodeh Moallemian, Bernadette A. Fausto, Fanny M. Elahi, and Mark A. Gluck. Funding acquisition: Mark A. Gluck. Resources: Mark A. Gluck. Supervision: Kevin S. Heffernan, Bernadette A. Fausto, Fanny M. Elahi, and Mark A. Gluck.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the Supporting Information.

CONSENT STATEMENT

All participants provided written informed consent prior to participation in the study. The study protocol was approved by the Institutional Review Board of Rutgers University.

Supporting information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

DATA AVAILABILITY STATEMENT

The data used to support these findings are available from the co‐first authors upon request.