APOE ε4 is associated with decreased synaptic density in cognitively impaired participants

Abstract INTRODUCTION We aimed to investigate the effect of apolipoprotein E4 (APOE) ε4 on synaptic density in cognitively impaired (CI) participants. METHODS One hundred ten CI participants underwent amyloid positron emission tomography (PET) with 18F‐florbetapir and synaptic density PET with 18F‐SynVesT‐1. We evaluated the influence of APOE ε4 allele on synaptic density and investigated the effects of ε4 genotype on the associations of synaptic density with Alzheimer's disease (AD) biomarkers. The mediation effects of AD biomarkers on ε4‐associated synaptic density loss were analyzed. RESULTS Compared with non‐carriers, APOE ε4 allele carriers exhibited significant synaptic loss in the medial temporal lobe. Amyloid beta (Aβ) and tau pathology mediated the effects of APOE ε4 on synaptic density to different extents. The associations between synaptic density and tau pathology were regulated by the APOE ε4 genotype. DISCUSSION The APOE ε4 allele was associated with decreased synaptic density in CI individuals and may be driven by AD biomarkers.


BACKGROUND
The apolipoprotein E (APOE) ε4 allele is the strongest genetic risk factor for sporadic Alzheimer's disease (AD) and has been studied extensively.
ε4 is known to be associated with abnormalities in the AD biomarkers amyloid beta (Aβ), tau pathology (T), and neurodegeneration (N), known as the "A/T/N" framework. 13][4] Several reports have indicated that APOE ε4 also increases tau pathology in AD. 5,6 Among the numerous genetic loci exhibiting sex-specific effects on AD, the intricate interaction between sex and the APOE allele has been investigated in detail. 7,8However, only a limited number of multiomics studies have been conducted to elucidate the interplay between APOE genotype and sex in AD. 8 Among cognitively impaired (CI) individuals, females are more susceptible to APOE ε4-associated accumulation of neurofibrillary tangles than males. 9In opposition to prevailing beliefs, males and females with the APOE ε3/ε4 genotype exhibit nearly equivalent likelihoods of developing AD between the ages of 55 and 85 years, but females exhibit a heightened risk at younger ages. 7Furthermore, decreased hippocampal volume and glucose uptake are both related to APOE ε4. 10,11wever, among these extensively reported findings, the association between the APOE ε4 allele and synaptic function has not been well reported or understood in humans, although this association has been reported in some animal studies, as reviewed in what follows.
Synapses have been reported to play a central role in cognitive performance in AD patients.Damage or loss of synapses is significantly related to neurodegeneration.Therefore, synaptic density is considered a key biomarker of neurodegeneration. 12APOE ε4 has been proposed to aggravate synaptic dysfunction and neuronal loss.However, this phenomenon has been verified only in animal models and cerebrospinal fluid biomarkers.Mechanistically, APOE ε4 is considered to mediate synaptic dysfunction by interfering with Reelin signaling, which is thought to be a modulator of synaptic strength. 13APOE ε4 is also associated with synaptic damage to the cerebrospinal fluid (CSF) at the neurogranin level as well as postsynaptic density protein 95 (PSD95) and synapsin 1 (Syn1) in post mortem human brain tissue. 14,15mulative Aβ, which is influenced by the APOE ε4 allele, can also be toxic to synapses.Synapse loss related to Aβ plaques may be caused by the increase in oligomeric Aβ rather than by the Aβ plaque itself. 16,17naptic vesicle glycoprotein 2A (SV2A) is ubiquitously expressed at synapses in the central nervous system 18 and is specifically localized within synaptic vesicles situated at presynaptic terminals. 19It is a direct biomarker of synaptic density in vivo. 20The development of SV2A radioligands, such as 11 C-UCB-J and 18 F-SynVesT-1, for positron emission tomography (PET) has provided an opportunity to evaluate synaptic density in the brain in various neurodegenerative disorders.
Reduced synaptic density was observed across the hippocampus and neocortex in AD patients by SV2A PET. 21The hippocampus is considered one of the earliest affected brain regions in AD, and dysfunction of this region is believed to be the core feature of disease-related memory impairment; moreover, analysis of this subfield could enhance the predictive value of the hippocampus in AD. 22 Furthermore, hippocam-

RESEARCH IN CONTEXT
1. Systematic review: Apolipoprotein E ε4 (APOE ε4) is the strongest genetic risk factor for sporadic Alzheimer's disease (AD).However, the authors reviewed the literature using PubMed, and their review revealed that the relationship between APOE-associated synaptic loss and cognitive impairment remains incompletely understood.pal synaptic density has also been assessed for its association with the "A/T/N" biomarkers in AD.Hippocampal synaptic density has a significant correlation with global amyloid deposition in participants with amnestic mild cognitive impairments. 23Regional synaptic loss follows tau accumulation after 2 years, which indicates that tau spread might drive synaptic vulnerability. 24 mentioned earlier, several investigations have been performed to explain the influence of the APOE ε4 allele on synaptic density.However, direct evidence about the effect of APOE ε4 on brain synaptic loss is lacking.In this study, we examined the effect of APOE ε4 on synaptic density in individuals with cognitive impairment by using SV2A PET and tested whether APOE ε4 has diverse effects on subfields of the hippocampus.Furthermore, we investigated the effect of APOE ε4 on the associations between synaptic density and AD biomarkers.We hypothesized that APOE ε4 exerts a significant influence on synaptic density throughout the brain, which could be mediated by AD pathology.In addition, we expected a potential interaction between APOE ε4 genotype and sex.

Participants
All participants were aged between 50 and 80 years and were recruited from memory clinics in Shanghai.The exclusion criteria are detailed in the supplementary information.All participants completed comprehensive neuropsychological assessments (described in the supplementary methods), APOE genotyping, 18 F-florbetapir PET/CT (Biograph128.mCT,Siemens, Germany), and 18 F-SynVesT-1 PET/MR (uPMR790 TOF, United Imaging, China).Seventy-four participants underwent 18 F-MK6240 PET/CT (Biograph128.mCT;Siemens, Germany) to survey tau pathology in the brain.Assessment of the amyloid PET images was performed through visual inspection by three experienced raters by consensus. 25All tests and image acquisition were conducted over 1 month.
Patients diagnoses were determined by consensus diagnostic meetings with multiple specialists.The CI participants included individuals with AD dementia and those with mild cognitive impairment (MCI).AD was diagnosed according to the 2011 National Institute on Aging and Alzheimer's Association (NIA-AA) diagnostic criteria after comprehensive neuropsychological assessments. 26Individuals with MCI were identified using the approach proposed by Jak and Bondi, as described in our previous report. 27Cognitively unimpaired participants were defined as those who had no cognitive impairment, specifically those who did not meet the diagnostic criteria for AD or MCI, and were excluded from this study.This study was approved by the Institutional Ethics Review Board of Huashan Hospital.Written informed consent was obtained from all participants.

PET data acquisition and processing
Participants underwent 30-min 18  SPM12 was used to process the 18 F-florbetapir, 18 F-MK6240, and 18 F-SynVesT-1 PET images.The detailed preprocessing steps for the voxelwise and region of interest (ROI) analyses are listed in the supplementary information.The cerebellum crus was used as a reference area to calculate the voxelwise standardized uptake value ratio (SUVR) of 18 F-florbetapir. 28Global SUVR values for 18 F-florbetapir were calculated by weighted averaging of the posterior cingulate, precuneus, and temporal, frontal, and parietal lobes, as described in our previous study. 291][32][33] We chose the ROI for 18 F-SynVesT-1 PET images in this study according to the findings of a previous study; the ROI encompassed the medial temporal lobe, entorhinal region, hippocampus, and parahippocampal gyrus. 34e masks of the hippocampal subfields were created by FreeSurfer (Laboratory for Computational Neuroimaging version 6.0, Boston, MA, USA). 35The hippocampal subfields were further defined as the hippocampal head, hippocampal body, and hippocampal tail, as shown in the supplementary materials (Figure S1).The hippocampal gray matter volume was calculated by CAT12.

Statistical analysis
Group comparisons between ε4 carriers and ε4 non-carriers were

Demographics
In AD group had greater amyloid deposition, lower education levels, and worse cognitive performance (Table S2).

Effect of APOE on synaptic density
We first investigated the effects of APOE ε4 on regional synaptic density.The APOE ε4 carriers displayed more severe synaptic loss across the medial temporal and occipital cortices than the APOE ε4 noncarriers (Figure 1A).This trend was retained in the medial temporal and occipital cortices after adjusting for the global florbetapir SUVR (Figure 1B).Furthermore, compared with those with the ε3ε3 genotype, the individuals with the ε3ε4 genotype also exhibited significant synaptic density loss in the medial temporal and occipital cortices (Figure 1C).These differences were retained after adjusting for the global florbetapir SUVR (Figure 1D).The analysis conducted with partial volume correction (PVC) (Figure S2A-D) yielded similar results as those obtained with non-PVC.

Effect on synaptic density loss in different sex
We observed significant main effects of APOE ε4 on synaptic density based on the ROIs, although no interaction effect was found in any of the ROIs or voxels (Table S3).Sex-stratified voxelwise analysis revealed that, compared with female APOE ε4 non-carriers,  Specifically, ROI-wise analysis revealed that female APOE ε4 carriers exhibited significant synaptic density loss in the hippocampus, parahippocampal gyrus, and hippocampal head, body, and tail (Figure S3G).Compared to male APOE ε4 non-carriers, male APOE ε4 carriers showed a loss of synaptic density in the medial temporal regions according to voxelwise analyses (Figure S3E) and in the hippocampal head (Figure S3H) without controlling for the global florbetapir SUVr.

APOE-stratified effect of association between synaptic density and "A/T/N" biomarkers
We used partial correlations to investigate the association between synaptic density and "A/T/N" biomarkers while controlling for sex, age, and education level.First, in the overall cohort, we found that hippocampal synaptic density was associated with global amyloid deposition (R = −0.446,p < 0.001), medial temporal tau deposition (R = −0.568,p < 0.001), and hippocampal volume (R = 0.699, p < 0.001).

Synaptic density loss mediated by Aβ and tau pathology
We further performed a mediation analysis to investigate the correlation among APOE ε4 genotype, Aβ (or tau) pathology, and synaptic density in the hippocampus and parahippocampal gyrus.We first assessed whether a higher APOE ε4-conferred Aβ burden was associated with severe synaptic density loss.Bootstrapped mediation analyses revealed that the association between the APOE ε4 genotype and synaptic density in the hippocampus or parahippocampal gyrus was partially mediated by the global Aβ PET SUVR (Figure 4A and B

DISCUSSION
The APOE ε4 allele was determined to exert different effects on AD biomarkers, such as amyloid and tau deposition.However, the mechanism underlying synaptic loss in APOE ε4 carriers remains unclear.
Although various studies have shown that APOE ε4 is associated with impaired synaptic integrity, these observations have not been validated in living human brains.In this study, we found that APOE ε4 was associated with a loss of synaptic density in CI participants.Only one copy of APOE ε4 is sufficient to induce synaptic density loss.However, we did not observe a significant sex-APOE interaction effect on synaptic density.Aβ pathology partially mediated the relationship between the APOE ε4 genotype and synaptic density.However, tau pathology fully mediated the relationship between the APOE ε4 genotype and synaptic density.
In this study, we confirmed that APOE ε4 carriers might undergo synaptic damage, which confers additional risk of AD.First, we found that APOE ε4 carriers exhibited significantly reduced synaptic density in the medial temporal lobe and neocortex.The reduced synaptic density remained after controlling for florbetapir SUVR, which may partly explain the synaptic loss associated with the APOE ε4 allele.Some data demonstrate that the soluble amyloid protein is toxic to synapses. 36,37though evidence indicates that the APOE genotype is associated with synaptic function in animal models, investigations regarding the effect of APOE on synaptic function in humans are limited.To our knowledge, this is the first report of significantly reduced synaptic density in the living human brain.Previous studies also revealed a significant decrease in synaptic density in the medial temporal lobe and neocortex in patients with AD compared to that in healthy controls.We found that this reduction focused on the medial temporal lobe in APOE ε4 carriers compared to that in APOE ε4 non-carriers.Synaptic density reduction in the hippocampus is considered the main pathology of AD.Lipidated apoE also regulates neuronal maintenance and repair by regenerating nerve cells to facilitate the repair process after neuronal injury. 38ApoE4 expressed by neuron could resulted in loss of neurons. 39Therefore, APOE ε4 could accelerate synaptic impairment, especially hippocampal synaptic impairment.Pathological analysis revealed that in the early stage of AD, the hippocampus experiences a rapid loss of tissue, which is correlated with the accumulation of amyloid plaques and tau tangles. 40Specifically, a reduction in synaptic density in the hippocampal head was significant.A previous tau PET study demonstrated that the earliest tau accumulation occurs in the entorhinal cortex, Brodmann area 35, and anterior hippocampus and could be associated with the loss of synapse density in the hippocampal head. 41These studies suggested that segmenting the lateral hippocampus into three subregions (head, body, and tail) could be useful for understanding the progressive pathological changes in the hippocampus in AD patients.
However, additional studies are needed to elucidate the underlying mechanism.
The APOE genotype also regulates the association between synaptic density and AD biomarkers.Global amyloid deposition has been associated with hippocampal synaptic density. 42APOE ε4 potentiated the association between synaptic density and tau pathology.apoE immunoreactivity demonstrated that increased expression of apoE in neurons is associated with increased tau phosphorylation. 43u pathology is closely related to neurodegeneration and cognitive performance.APOE ε4 promotes increased tau deposition and brain atrophy in carriers compared to non-carriers, which is proposed to be mediated by glia. 44Astrocytes are the main source of apoE, and signaling by astrocytes results in synaptic and glial transmission. 1Therefore, we observed a close association between synaptic density and tau deposition and hippocampal volume.In addition, analysis of the mediation model suggested that the effects of APOE ε4 on synaptic loss were not independent and could be driven by amyloid/tau pathology.Amyloid pathology partially mediated the effects of APOE ε4 on synaptic loss and may be correlated with increasing oligomeric Aβ levels rather than Aβ plaques.Importantly, tau pathology in the medial temporal lobe fully mediated the effects of APOE ε4 on synaptic loss, which supports the hypothesis that synapses are damaged due to the toxicity of tau. 45,46is study has several limitations.First, we did not have longitudinal data to validate the effect of APOE ε4 on synaptic density loss.Second,

2 . 3 .
Interpretation: APOE ε4 carriers showed a significantly decreased synaptic density compared to non-carriers, and only one copy of APOE ε4 reduced the synaptic density.This effect was partially mediated by amyloid pathology and fully mediated by tau pathology.APOE ε4 also potentiated the association between synaptic density and tau pathology.Future directions: This study emphasized the importance of developing genotype-guided therapies targeting synapses and related protein aggregates in AD.Large sample size and longitudinal data are needed to validate the effect of APOE ε4 on synaptic density loss.
F-SynVesT-1 PET scans at 60 min after injection (6.5 ± 0.65 mCi 18 F-SynVesT-1) with a 3T uPMR790 TOF (United Imaging Healthcare, China).T1-weighted images were collected at the same time (repetition time [TR] = 7.2 ms; echo time [TE] = 3.0 ms).For 18 F-florbetapir PET/CT scans, 10 mCi (± 10%) [ 18 F]florbetapir was injected, and a 20-min scan was performed 50 min after the injection.For 18 F-MK6240 PET/CT scans, 5 mCi (± 10%) [ 18 F]-MK6240 was injected, and a 20-min scan was performed 90 min after the injection.PET images were reconstructed by the filtered back projection (FBP) algorithm.For each PET imaging tracer, there was one time frame consisting of 30 and 20 min, depending on the tracer used. 18F-SynVesT-1 PET attenuation correction was performed using a three-compartment model (bone, soft tissue, air) attenuation map automatically generated from an ultrashort TE MRI sequence.A lowdose CT scan was used to perform 18 F-florbetapir and 18 F-MK6240 PET attenuation correction.
using Fisher's exact test for categorical variables and the two-sample T test for continuous variables such as scores on neuropsychological assessments, age, and education years.The voxelwise and ROI-wise differences in synaptic density among ε4 carriers and ε4 noncarriers were analyzed by a general linear model (GLM) using SPM12 and SPSS (IBM version 26.0), respectively.Age, sex, and educational level were set as covariates, and a value of p < 0.001 was set as the significance threshold for the voxelwise analysis.Voxelwise analyses were repeated using partial volume-corrected data.For ROI-wise analysis, association coefficients were calculated by partial correlation after adjusting for age, sex, and educational level.Fisher's Z test was used to compare coefficients between ε4 carriers and ε4 non-carriers.Statistical significance was defined as an unadjusted two-sided p value less than 0.05.Mediation effects were assessed using the SPSS PROCESS macro (version 3.3).Interaction effects were calculated with the R package stargazer (version 5.2.3).
Demographic and neuropsychological testing data for participants.
total, 110 participants, including 63 AD participants and 47 MCI participants, were enrolled in this study.Among the participants, 64 participants were APOE ε4 carriers (54 heterozygotes and 10 homozygotes) and 46 were APOE ε4 non-carriers (seven individuals with the ε2ε3 genotype and 39 individuals with the ε3ε3 genotype).As shown in Table1, compared to APOE ε4 non-carriers, APOE ε4 caror neuropsychological scores between the APOE ε4 carriers and noncarriers.Furthermore, there were no differences between the APOE ε4 non-carriers and carriers among female and male patients separately, except that APOE ε4 carriers had a higher Aβ PET-positive rate than non-carriers in the female group (TableS1, 82.5% vs 40.0%, p < 0.001).Compared to participants in the MCI group, the participants in the TA B L E 1 Abbreviations: AFT, Animal Fluency Test; AVLT, Auditory Verbal Learning Test; BNT, Boston Naming Test; MMSE, Mini-Mental State Examination; MoCA-B, Montreal Cognitive Assessment-Basic; STT, Shape Trail Test.

Table 2
P value adjusted for age, education, sex, amyloid beta (Aβ) global deposition, and clinical diagnosis (diagnosis).