Blood-brain barrier breakdown is linked to tau pathology and neuronal injury in a differential manner according to amyloid deposition

Abstract

The blood-brain barrier (BBB) breakdown has been suggested as an early marker for Alzheimer's disease (AD); yet the relationship between BBB breakdown and AD-specific biomarkers based on the amyloid/tau/neurodegeneration framework is not clear. This study investigated the relationship between BBB permeability, AD-specific biomarkers, and cognition in patients with cognitive impairment. In this prospective study, we enrolled 62 participants with mild cognitive impairment or dementia between January 2019 and October 2020. All participants were assessed through cognitive tests, amyloid positron emission tomography (PET), dynamic contrast-enhanced magnetic resonance imaging (MRI) for BBB permeability (K_trans), cerebrospinal fluid studies for Aβ42/40 ratio, phosphorylated-tau Thr181 protein (p-tau), total tau protein (t-tau), and structural MRI for neurodegeneration. In amyloid PET (+) group, higher cortical K_trans was associated with lower Aβ40 (r = −0.529 p = 0.003), higher Aβ42/40 ratio (r = 0.533, p = 0.003), lower p-tau (r = −0.452, p = 0.014) and lower hippocampal volume (r = −0.438, p = 0.017). In contrast, cortical K_trans was positively related to t-tau level. (r = 0.489, p = 0.004) in amyloid PET (−) group. Our results suggest that BBB permeability is related to AD-specific biomarkers, but the relationship can vary by the presence of Aβ plaque accumulation.

Keywords

Alzheimer's disease biomarkers blood-brain barrier magnetic resonance imaging permeability

Introduction

Alzheimer's disease (AD) is the most common form of dementia that is characterized by abnormal protein accumulation of amyloid-β (Aβ) and tau and neurodegeneration.¹ In the recent amyloid/tau/neurodegeneration (ATN) framework, the AD continuum is defined by the deposition of Aβ plaques in the brain.¹ Amyloid pathology can be revealed through amyloid- positron emission tomography (PET) or decreased Aβ42/40 ratio (or decreased Aβ42) in cerebrospinal fluid (CSF); tau pathology through tau-PET or increased phosphorylated-tau (p-tau); neurodegeneration through hippocampal atrophy in magnetic resonance imaging (MRI) or increased total-tau in CSF. The introduction of these biomarkers eventually led to the proposed framework to define AD as a biological entity defined by the presence of amyloid, tau, and neurodegeneration.

However, several other potential biomarkers act as crucial players in AD pathogenesis. In particular, blood-brain-barrier (BBB) breakdown is being recognized as another major contributor to AD.^2–6 Experimental studies indicate that dysfunction of BBB, which is implicated in amyloid protein efflux, can exacerbate beta-amyloid accumulation in the brain.^3,6 Conversely, there is also evidence that amyloid accumulation can lead to BBB breakdown, as noted in cerebral amyloid angiopathy. ³ Several recent reports suggest that tau alone can initiate BBB breakdown with tau-induced glial activation and subsequent endothelial damage, and vice versa.^7–9 Besides AD, BBB is also affected in normal aging ¹⁰ and non-Aβ neurodegenerative diseases ^3,11–13 and vascular disease.^5,14 Given the fact that mixed AD with vascular pathology is the most common cause of dementia, BBB dysfunction is often explained as a pathological process linking AD and microvascular pathology.¹⁴

Contrary to the experimental finding of the close relationship between BBB and amyloid/tau pathology,^7,13,15 BBB change was reportedly independent of amyloid or tau pathology in human subjects with cognitive impairment.¹⁶ Recent attempt to open BBB to clear amyloid pathology as a therapeutic trial has revealed that BBB opening and increased BBB permeability was transiently associated with reduction of amyloid pathology in the brain, which confuses further our understanding of the role of BBB in AD pathology.¹⁷

Dynamic contrast-enhanced (DCE)- MRI enables us to measure the BBB dysfunction in vivo in human subjects without an invasive procedure involved and is validated as a tool for BBB function assessment.^18–22 We hypothesized that the relationships between BBB and AD-specific biomarkers differ depending on the accumulation of Aβ plaque, given the experimental evidence of the bi-directional effect in BBB dysfunction and amyloid pathology.¹⁵ Hence, we aimed to evaluate the relationship between BBB permeability and AD-specific biomarker according to the ATN framework. Moreover, we evaluated whether BBB permeability change would affect global cognition in patients with cognitive impairments, using the DCE-MRI-based BBB permeability imaging.

Materials and methods

Participants

We recruited patients with memory complaints who visited the department of neurology and psychiatry of the Konkuk University Medical Center in South Korea between January 2019 and October 2020. We assessed their basic demographic characteristics, including age, sex, years of education, and vascular risk factors, as well as the presence of hypertension, diabetes mellitus, and dyslipidemia. Comprehensive neuropsychological tests, global cognitive assessments (Clinical Dementia Rating–Sum of Boxes [CDR-SOB], and Mini-Mental State Examination [MMSE] scores), and brain imaging were also performed. CDR-SOB and MMSE scores were obtained using standard procedures. The patients who were diagnosed with dementia caused by diseases other than neurodegenerative causes, such as metabolic causes, trauma, tumor, psychiatric disease, or drugs, were excluded. Based on the National Institute of Neurological Disorders and Stroke Association Internationale Pour la Recherche et l’Enseignement en Neurosciences (NINDS-AIREN) criteria,²³ patients diagnosed with vascular dementia as the main form of dementia were also excluded. In addition, patients with severe hearing and vision impairments who could not perform neuropsychological tests were also excluded. The detailed exclusion criteria can be found in our previous studies.^18,24–26

A total of 101 participants with a diagnosis of MCI and dementia were included in this study; the diagnoses of MCI and dementia were based on the criteria suggested by Petersen et al.²⁷ and the Diagnostic and Statistical Manual of Mental Disorders (5th edition), respectively. Patients with MCI performed normal daily living activities; nonetheless, they exhibited an objective memory impairment, implying <1.5 standard deviation from the norm in at least one memory test.²⁸ Patients with an amyloid biomarker were diagnosed as AD with MCI or AD with dementia based on the National Institute on Aging and Alzheimer's Association research framework criteria.¹ Among the 101 participants, 23 and 16 individuals were excluded because of refusal to undergo the CSF study and incomplete MR examination caused by either poor-quality related to motion artifacts or incomplete acquisition of the scans due to poor compliance, respectively. Finally, a total of 62 participants composed of 14 patients of MCI and 48 patients of dementia were included in this analysis.

Apolipoprotein E (APOE) genotyping

Genomic DNA was isolated and purified using the QIA symphony DSP DNA Mini Kit (Qiagen GmbH, Hilden, Germany) on an automated QLA symphony SP system (Qiagen) according to the manufacturer's instructions. APOE genotyping was performed using a Real-Q APOE genotyping kit (Biosewoom, Seoul, Korea) on a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s instructions. Patients with one or two copies of an APOE ε4 allele were defined as APOE ε4 carriers.

MRI acquisition

MRI was performed at the Konkuk University Medical Center using a Magnetom Skyra 3.0 Tesla unit (Siemens Medical Systems, Erlangen, Germany) with a 20-channel high-resolution head coil. The MRI protocol included three-dimensional (3D) T1-weighted images, 3D fluid-attenuated inversion recovery images, 3D susceptibility-weighted images, and coronal DCE imaging, with a 10-min acquisition period and at 10-s resolution using Gadobutrol (1.0 mmol/kg of body weight). The specific parameters and protocols used for structural MRI and DCE MRI are provided in Supplementary Table 1. Our DCE MRI protocol was based on previously recommended parameters and was validated by reliable and reasonable clinical results.^18,19,21,29

MRI analysis

Structural MRI and DCE analysis

Automatic segmentation of distinct brain regions from 3D T1-weighted volumetric images was performed using the InBrain platform (https://ww.inbrain.co.kr/; Seongnam-si, Gyeonggi-do, Republic of Korea; MIDAS Information Technology Co., Ltd.).^30,31 Based on Freesurfer 6.0 platform, InBrain applies deep-learning algorithms to the analysis of failure prediction, brain extraction, white matter segmentation, and analytical quality management. The volume of white matter hyperintensity was extracted based on the Desikan Killiany atlas and subcortical atlases.^30–33

The volume of hippocampus and white matter hyperintensity were used in the following statistical analysis. To correct for differences in head size, volumes were divided by the total intracranial volume, then multiplied by 1000. Hence, hippocampal volume and white matter hyperintensity volume (WMHV) was defined as the ratio of the volume of the hippocampus, and white matter hyperintensity, respectively.

The presence of lobar cortical microbleeds (CMB, <10 mm in diameter with associated blooming on susceptibility-weighted imaging) was evaluated by an expert neuroradiologist and a neurologist, the first author. Presence of cerebral amyloid angiopathy was diagnosed based on the Boston 2.0 criteria.³⁴

The Nordic ICE software (Version 4.1.3) was used to process the DCE imaging data. 3D T1-volume imaging was used for structural imaging reference. We applied a Patlak model for BBB permeability calculation.^20,35 Vascular input function was obtained semi-automatically from the superior sagittal sinus. The analysis was performed by a trained neuroimaging researcher with 3 years of experience under the supervision of an expert neuroradiologist. Both were blinded to the clinical information. Accordingly, we calculated K_trans, which is equivalent to the volume transfer constant between the plasma and the extravascular extracellular space. The unit of K_trans was min⁻¹, which is equivalent to mL/g/min.

The volumes of interest of the whole cortex from T1-weighted volumetric images were coregistered to the BBB permeability map to extract the K_trans values using a mutual information-based algorithm and to search for an optimal rigid transformation on Nordic ICE. Single slices of the K_trans map for each participant that were not overlaid were added in the Supplementary Figure 1 to confirm the exact location of leakage. We checked all images before DCE measurement if there is overlap of brain mask with the CSF spaces and ensured that there is no significant overlap of brain mask and CSF spaces (Figure 1). The mean K_trans value of the whole cortex was used in the statistical analysis.

Figure 1.

Representative cases showing coregistered cortical map on BBB permeability map in patient of amyloid (¹⁸F-florbetaben) PET positive. The volumes of interest of the whole cortex from T1-weighted volumetric images (upper line) were coregistered to the BBB permeability map (lower line) to avoid overlap of brain mask and cerebrospinal spaces and to extract the K_trans values using a mutual information-based algorithm.

CSF collection and AD biomarker analysis

For most participants, CSF samples were collected from a lumbar puncture performed in the L3–4 or L4–5 intervertebral spaces using a 20 G or 22 G needle. Fasting was not required. All CSF samples were collected into 15-mL polypropylene tubes at the time of the tap procedure and were sent to the Hanyang Medical Center laboratory within 30 min after collection. Subsequently, the samples were centrifuged at 2000 × g for 10 min at 4°C, aliquoted into polypropylene cryogenic sterile freestanding conical microtubes, frozen within 30 min of collection, and stored at −80°C until the analysis of cytokine concentrations. According to the manufacturer’s protocol, all cytokines were analyzed using a commercially available Milliplex® Human Amyloid Beta and Tau Magnetic Bead Panel – Multiplex Assay kit [Aβ40, Aβ42, total tau proteins (t-tau), and phosphorylated tau at Thr181 (p-tau)] obtained from Millipore (Billerica, MA).

Aβ PET acquisition and definition of Aβ positivity

PET images were acquired 90 min after intravenous injection of 300 MBq of ¹⁸F-florbetaben to patients in the resting state using a PET/CT scanner (Ingenuity TF 128, Philips Medical Systems, Cleveland OH, USA) in a 20-min 3D dynamic mode (4 × 5-min frames) with appropriate data corrections. One experienced nuclear medicine physician (>20 years of experience in neuro-PET/CT) and one neurologist (>15 years of experience in neuro-PET/CT) evaluated independently amyloid PET images while blinded to clinical information other than the anatomical brain imaging to evaluate morphological change. Any discrepancies were resolved by consensus.

Amyloid PET positivity was determined visually with the help of semiquantitative assessment with measurement of the cerebral-to-whole cerebellar standardized uptake value ratio (SUVR). A visual rating score of 2 or 3 on the brain Aβ plaque load scoring system was categorized as 'amyloid PET-positive'.³⁶ SUVR of the whole cortex was calculated using a FDA-approved software (Veuron-Brain-pAb2, Heuron Inc. Republic of Korea, iheuron.com), by which for the volume of interest analysis, the T1-weighted MRI images were coregistered with the PET images based on the previously suggested methods.³⁷

Markers of amyloid, tau, and neurodegeneration

According to the recently published research framework, we classified participants into two groups based on the presence [amyloid PET positive (+)] or absence [amyloid PET negative (−)] of abnormal deposition of cerebral Aβ plaques. The CSF Aβ40, CSF Aβ42, Aβ42/40 ratio, and CSF p-tau levels were used as a marker of amyloid and tau, respectively. CSF t-tau and the hippocampal volume were used as markers of neurodegeneration or neuronal injury.¹ The CMB. WMHV and CAA were used as one of the radiological covariates that represent microvascular injury and could affect BBB permeability or cognition.

Statistical methods

Age, educational years, MMSE, imaging markers, and CSF markers, except for CDR-SOB, CSF t-tau, and cortical K_trans, were normally distributed, as indicated by the Kolmogorov–Smirnov test. Therefore, we transformed cortical K_trans data using a logarithmic transformation, and cortical K_trans was normally distributed after logarithmic transformation.

The t-test or the Mann–Whitney U test was used for continuous variables, and the chi-squared test or Fisher’s exact test was used for categorical variables. Associations between CSF AD biomarkers, K_trans, and cognition scores were evaluated using Pearson’s or Spearman’s correlation analysis and partial correlation analysis.

Then, we performed the stepwise (or enter) method of multiple linear regression analysis to identify predictive factors for MMSE. As the linear regression model fit was not statistically significant in case of CDR-SOB (R² = 0.391; P-value = 0.318), analysis about CDR-SOB was excluded. CDR-SOB reflects not only cognition, but also the patient's emotions, abnormal behaviors, and muscle strength. Therefore, we selected only MMSE as the outcome to find out the effect of the BBB on cognitive function. After confirming whether the amyloid positivity could be one of the predicting factors for MMSE or not, we re-evaluated the results after correction for multiple comparisons. SPSS (v. 17.0, SPSS Inc., Chicago, IL, USA) was used for statistical analyses, and a p-value < 0.05 was considered as the threshold of significance.

Protocol approvals, registrations, and patient consent

This prospective study was approved by the Institutional Review Board of the Konkuk University Medical Center (no. KUH1170180). All procedures used in this study were carried out according to the ethical standards set by the Helsinki Declaration of 1975 (and as revised in 1983). Written informed consent was obtained from all participants.

Results

Demographic characteristics

The mean age of the participants was 72.9 ± 8.1 years, and 54.8% (n = 34) were female. Mean CDR-SOB was 5.1 ± 3.6. 38.7% (n = 24) of the participants were APOE ε4 carriers.

When stratified according to amyloid PET positivity, 33 patients were classified into the amyloid PET (−), and 29 were in the amyloid PET (+). The two groups did not differ significantly regarding age, sex ratio, educational years, and scores on the MMSE or CDR-SOB. The demographic characteristics of all participants and each group are summarized in Table 1.

Table 1.

Demographic, clinical and imaging data of the study population.

	Total(N = 62)	Amyloid PET(+) group(N = 29)	Amyloid PET(−) group(N = 33)	P-value
Demographic
Age, years	72.9 ± 8.1	73.6 ± 7.0	72.2 ± 8.9	0.507
Female (%)	34 (54.8%)	14 (48.3%)	20 (60.6%)	0.334
Education, years	7.9 ± 4.8	9.1 ± 4.6	6.8 ± 4.9	0.062
Hypertension	32 (51.6%)	14 (48.3%)	18 (54.5%)	0.625
Diabetes mellitus	14 (22.6%)	6 (20.7%)	8 (24.2%)	0.741
Dyslipidemia	14 (22.6%)	8 (27.6%)	6 (18.2%)	0.381
APOE4 carrier (%)	24 (38.7%)	15 (51.7%)	9 (27.3%)	0.050
MMSE	18.7 ± 5.8	17.9 ± 5.2	19.4 ± 6.3	0.302
CDR-SOB	5.1 ± 3.6	5.8 ± 3.1	4.5 ± 3.8	0.130
Amyloid PET SUVR	1.22 ± 3.0	1.47 ± 1.9	0.99 ± 0.2	<0.001
MCI (%)	14 (22.6%)	3 (10.3%)	11 (33.3%)	0.027
MMSE of MCI	25.6 ± 4.0	27.3 ± 4.6	25.2 ± 3.9
MMSE of dementia	16.8 ± 4.7	16.8 ± 4.1	16.6 ± 5.4
p-value	<0.001	0.002	<0.001
CDR-SOB of MCI	25.6 ± 4.0	1.7 ± 2.0	1.5 ± 1.1
CDR-SOB of dementia	16.8 ± 4.7	6.4 ± 2.9	6.0 ± 3.9
p-value	<0.001	0.009	<0.001
CSF biomarker
CSF Aβ40 (pg/mL)	3755.9 ± 1418.4	3609.9 ± 1422.1	3884.3 ± 1424.5	0.452
CSF Aβ42 (pg/mL)	439.4 ± 208.0	291.8 ± 78.2	569.1 ± 200.0	<0.001
CSF Aβ42/40	0.1 ± 0.0	0.1 ± 0.0	0.2 ± 0.0	<0.001
CSF p-tau (pg/mL)^a	65.9 [46.0;135.4]	128.2 [86.3;159.9]	47.4 [36.3;61.7]	<0.001
CSF t-tau (pg/mL)	394.2 ± 270.3	527.9 ± 292.6	276.6 ± 184.2	<0.001
MRI
Hippocampal volume (ratio of ICV×10³)	4.4 ± 0.6	4.2 ± 0.6	4.6 ± 0.6	0.009
WMHV (ratio of ICV×10³)	9.4 ± 6.3	7.9 ± 5.4	10.6 ± 6.8	0.093
CMB (%)	21 (33.9%)	10 (34.5%)	11 (33.3%)	0.925
CMB, number	3.3 ± 16.0	6.3 ± 23.2	0.7 ± 1.6	0.210
CAA (%)	14 (22.6%)	6 (18.2%)	8 (27.6%)	0.381
Cortical K_trans (×10⁻³/min)	0.9 ± 0.8	0.9 ± 0.7	0.8 ± 0.8	0.426
Hippocampal K_trans (×10⁻³/min)	0.5 ± 0.6	0.6 ± 0.7	0.5 ± 0.5	0.330

Value: Mean (Standard deviation) or number (%) or median [interquartile range]. The t-test or chi-squared test was used for continuous variables or categorical variables, respectively. MCI: mild cognitive impairment; Aβ+G: Aβ presence group in PET-CT; Aβ-G, Aβ absence group in PET-CT; CSF: cerebrospinal fluid; Aβ42/40: the ratio of beta-amyloid 1–42 and 1–40; p-tau: phosphorylated tau; t-tau: total tau; CMB: cortical microbleeds; CAA: cerebral amyloid angiopathy. Hippocampal volume ratio= Hippocampal volume/total intracranial volume × 1000; WMHV: white matter hyperintensities volume/total intracranial volume × 1000.

The Mann–Whitney U test was used for CSF p-tau, because it is not normally distributed.

BBB permeability and CSF AD biomarkers depending on amyloid positivity on PET

CSF AD biomarker levels in the two groups are reported in Table 1. The CSF Aβ42/40 ratio was lower, and CSF t-tau was higher in amyloid PET (+) than in amyloid PET (−) (all p < 0.001). The median CSF p-tau levels were higher in amyloid PET (+) than in amyloid PET (−) group (p < 0.001). Hippocampal atrophy was more prominent in amyloid PET (+) than amyloid PET (−) (4.16 ± 0.59 vs 4.57 ± 0.61, ratio of ICV × 10³, p = 0.009, respectively). However, no difference was observed in WMHV (p = 0.093), CMB (p = 0.925) and CAA (p = 0.381) between the two groups. Cortical K_trans did not differ between two groups, either (0.96 ± 0.74 × 10⁻³/min and 0.80 ± 0.80 × 10⁻³/min, in amyloid PET (+) and amyloid PET (−), respectively, p = 0.426) (Table 1, Figure 2).

Figure 2.

Scatter plot of cortical K_trans and CSF AD biomarkers in the amyloid PET (−) group. The solid blue line represents linear regression fit, and the shaded area represents 95% confidence interval.

In the whole group, cortical K_trans was positively correlated with t-tau (r = 0.283, p = 0.026) and negatively correlated with Aβ40 (r = −0.370, p = 0.003, Supplementary Table 2), whereas no correlation between K_trans and the Aβ42 or Aβ42/40 ratio or p-tau or hippocampal atrophy. These relationships remained even after controlling for the effect of WMHV and CMB [K_trans with t-tau (r = 0.308, p = 0.017) and with Aβ40 (r = −0.382, p = 0.003)].

In the amyloid PET (+) group, higher cortical K_trans was associated with lower Aβ40 (r = −0.529 p = 0.003) and higher Aβ42/40 ratio (r = 0.533, p = 0.003). These relationships remained even after controlling for the effect of WMHV and CMB [K_trans with Aβ40 (r = −0.601, p = 0.001) and with Aβ42/40 ratio (r = 0.545, p = 0.003)]. In addition, higher cortical K_trans was associated with lower p-tau (r = −0.452, p = 0.014) and lower hippocampal volume (r = −0.438, p = 0.017). There was no correlation between cortical K_trans and t-tau (p = 0.813). (Figure 3)

Figure 3.

Scatter plot of cortical K_trans and CSF AD biomarkers in the amyloid PET (+) group. The solid blue line represents linear regression fit, and the shaded area represents 95% confidence interval.

In the amyloid PET (−) group, cortical K_trans was also positively related to the t-tau level. (r = 0.489, p = 0.004). However, cortical K_trans was not correlated with any CSF Aβ markers, p-tau, and hippocampal volume.

BBB permeability, CSF AD biomarkers, and cognition

The whole group had no significant correlation between CSF markers and MMSE. Cortical K_trans was not correlated with MMSE (r = 0.054, p = 0.678) when considering educational years as a covariate. Lower MMSE score was correlated with higher WMHV (r = −0.287, p = 0.025).

In amyloid PET (+) group, there was no variable which had correlates with MMSE. Cortical K_trans was not correlated with MMSE (r = 0.279, p = 0.150, Table 2). In amyloid PET (−) group, the only significant correlate of MMSE was WMHV (r = −0.410, p = 0.020, Table 2). Cortical K_trans was not correlated with MMSE (r = 0.023, p = 0.900, Table 2).

Table 2.

Correlation between MMSE and imaging biomarkers including cortical K_trans, and predictors of MMSE in the amyloid PET (+) group and in the amyloid PET (−) group.

	Total(N = 62)		Amyloid PET(+) group(N = 29)		Amyloid PET(−) group(N = 33)
Correlation with MMSE	r	P-value	r	P-value	r	P-value
Age	−0.160	0.218	−0.141	0.475	−0.165	0.366
Cortical K_trans (×10⁻³/min)	0.054	0.678	0.279	0.150	0.023	0.900
Hippocampal volume (ratio of ICA×10³)	0.164	0.208	−0.102	0.606	0.173	0.343
WMHV (ratio of ICV×10³)	−0.287	0.025	−0.108	0.585	−0.410	0.020
Hypertension	−0.108	0.405	−0.021	0.915	−0.150	0.403
Diabetes mellitus	−0.108	0.403	0.026	0.895	−0.208	0.245
Dyslipidemia	−0.116	0.371	−0.032	0.867	−0.194	0.279
Predictors of MMSE	B(SE)	P-value	B(SE)	P-value	B(SE)	P-value
Constant	15.173	0.139	16.367(11.87)	0.366	5.905(16.74)	>0.999
Age	0.024	0.803	−0.102(0.11)	0.766	0.149(0.15)	0.718
Sex	−0.753	0.599	0.731(1.62)	>0.999	−2.582(2.40)	0.586
Educational years	0.400	0.009	0.616(0.20)	0.012	0.275(0.23)	0.492
APOE4	2.048	0.149	3.331(1.59)	0.098	0.292(2.57)	>0.999
Cortical K_trans (×10⁻³/min)	2.124	0.273	3.048(2.62)	0.516	2.294(2.94)	0.886
Hippocampal volume (ratio of ICA×10³)	0.010	0.442	0.005(0.01)	>0.999	0.022(0.02)	0.578
WMHV (ratio of ICV×10³)	−0.003	0.015	−0.001(0.00)	>0.999	−0.004(0.00)	0.078
Amyloid Positivity	−3.806	0.016

Value: Hippocampal volume ratio = Hippocampal volume/total intracranial volume × 10³; WMHV: white matter hyperintensities volume/total intracranial volume × 10³; MMSE: Mini-Mental Status Examination. The model summary was R² = 0.578; P-value = 0.004, R² = 0.553; P-value = 0.009, R² = 0.309; P-value = 0.183 at total group, Amyloid PET (+) group, and Amyloid PET (−) group, respectively.

Clinicodemographic variables (age, sex, educational years, and APOE4 status), cortical K_trans, and the current imaging markers (hippocampal volume and WMHV) were used to forward stepwise regression analysis and revealed that only educational years (B = 0.311, p-value = 0.043) and WMHV (B = −0.003, p-value = 0.044) were significant predictors for impaired cognition in the whole group. In the same analysis for the amyloid PET (+) group, educational years (B = 0.616, p-value = 0.006) was the independent predictors for impaired cognition. For the amyloid PET (−) group, there was no independent predictor for impaired cognition (Table 2).

Discussion

Our study revealed that there was no difference in cortical K_trans between the amyloid PET (+) and the amyloid PET (−) group. However, in both the amyloid PET (−) and amyloid PET (+) group, cortical K_trans is positively correlated with one of the neuronal injury markers (t-tau and hippocampal atrophy, respectively. Further, the amyloid PET (+) group showed that higher cortical K_trans was negatively correlated with amyloid and tau pathologies, as evidenced by CSF biomarkers. Finally, cortical K_trans did not predict cognitive function in patients included in this study. Instead, education, and to a lesser degree, hippocampal atrophy were the predictors of impaired cognition in our study population.

Our observation of no difference in cortical K_trans according to amyloid PET positivity is in line with recent studies.^16,38 According to Nation et al, K_trans of the hippocampus, parahippocampal gyrus, and amygdala did not differ according to the presence of amyloid pathology in subjects with normal cognition or mild cognitive impairment.³⁸ While the previous study focused on the normal and early stages of AD, our study population encompasses a later stage of cognitive impairment. Thus, we speculate that BBB permeability is not different depending on the amyloid positivity irrespective of the cognitive impairment stage. In contrast to our findings, several experimental studies hinted the higher amyloid deposition in the presence of BBB leakage.^15,39,40 The reason for the discrepancy between human imaging results and experimental results is unknown.

We found a clear relationship between whole cortical K_trans and neuronal injury (neurodegeneration) markers in both groups; hippocampal atrophy for the amyloid PET (+) group and t-tau for the amyloid PET (−) group. Increased BBB permeability i.e., BBB leakage, was demonstrated in the disease-specific region such as the hippocampus for MCI^10,16,29,38 and also in the whole cortex for early AD,⁴¹ in the entire cortex for cerebral small vessel disease.⁴² Our result was corroborated by the previous report that BBB permeability measured by albumin index was correlated with the severity of medial temporal lobe atrophy ⁴³ and by the experimental models of BBB leakage that induced neurodegeneration over time.⁴⁴ Furthermore, our result suggests that BBB leakage of the cortex, whether it is an upstream or downstream event, is linked to the final common result of any insult to nervous tissue, neuronal injury/neurodegeneration.¹

However, the discordant relationship of cortical K_trans with t-tau and hippocampal volume depending on amyloid pathology might be explained by the reported less agreement between t-tau and hippocampal atrophy.⁴⁵ First, each neuronal injury marker can deviate from the normal range at a different time point of the AD continuum.⁴⁵ Therefore, the changes in BBB permeability may vary depending on the severity of neurodegeneration and the presence or absence of amyloid pathology. Second, our observation might be related to the innate lower concordance between two neuronal injury markers. Hippocampal atrophy represents the sum of the dendritic loss and neuronal cell death. In contrast, CSF t-tau is reminiscent of neuronal cell death and is less sensitive than dendritic branching loss.⁴⁵

In the amyloid PET (+) group, higher cortical K_trans was associated with higher Aβ42/40 ratio (less amyloid pathology in the brain) and lower p-tau (less tau-pathology). This negative relationship between BBB leakage and amyloid pathology contradicts the previous report using an experimental model of BBB dysfunction in that BBB integrity abnormality impairs the clearance of Aβ42 and Aβ40 from the brain interstitial fluid until cerebral amyloid angiopathy, and brain amyloid accumulation occurs.^15,39,40 Previous DCE studies for early cognitive impairment also reported that BBB leakage was present before the amyloid and tau pathology became evident.¹⁰ BBB leakage is also exacerbated by amyloid accumulation,³ and even by neuroinflammation such as microglia activation.⁴⁶ We speculate that our unexpected findings might be responsible for the dual roles of BBB: one is a barrier against toxic protein from systemic blood, and the other is an interface for transporting nutrients/waste to and from blood and interstitial space.⁴⁷ In that sense, increased BBB permeability can be both a protective effect and aggravating effect on unwanted toxic protein pathology in the brain leading to its non-linear relationship with toxic protein pathology. Furthermore, there is some evidence that brain Aβ has a protective effect on BBB by sealing the BBB leaks.^17,48 This partly explains the lower cortical K_trans observed in our study with more amyloid accumulation. Previous results that ABCG2, a BBB-related gene that is upregulated in AD, may act as a gatekeeper at the BBB to prevent blood Aβ from entering the brain also support our results.⁴⁹ Second, as the lower CSF Aβ40 was associated with the higher cortical K_trans in the amyloid PET (+) group, the seemingly paradoxical relationship between BBB and amyloid pathology (higher Aβ42/40 ratio) in our study may be partly attributed to concomitant amyloid angiopathy in which Aβ40 is accumulated in the blood vessels. Subsequently, CSF Aβ40 can be lower than in individuals not affected by this condition.^50–52

The negative relationship between cortical K_trans and CSF p-tau of amyloid PET (+) group in our study can also be explained in the same context as amyloid. Tau protein could damage the BBB only by mediating the neuroinflammatory process through the activation of glial cells, not directly toxic to brain endothelial cells.^8,9 Even if tau protein accumulates in the brain, its relationship with BBB damage can be either positive or negative, depending on whether it incites a related neuroinflammatory response. Therefore, follow-up studies with neuroinflammation markers as well as ATN markers are considered necessary.

We also did not find a direct relationship between cortical K_trans and general cognition score (MMSE), which is in line with the previous studies that revealed no significant relationship between BBB leakage rate and MMSE in AD⁴¹ and cerebral small vessel disease.⁴² According to van de Haar et al., the fractional leakage volume of the cortex increased with decreasing MMSE score, however, it is unclear whether the cognition is related to the magnitude of BBB leakage or the spatial extent of BBB leakage.⁴¹

When measuring large molecule BBB leakage using the CSF albumin index, one study revealed the magnitude of leakage was related to the CDR-SOB change over time.⁵³ When measuring much smaller molecules such as water using a new water permeability imaging technique, the preliminary result showed that after controlling for age, sex and education, memory score is negatively correlated to the water permeability rate (Kw) but not to larger-molecule leakage (CSF-albumin index).⁵⁴ These results contradict our observation. First, in our study, K_trans measurement using gadolinium contrast enables us to detect medium-sized molecular (approximately 550 kDalton) as compared to the larger-molecule such as albumin (and very small molecule like water, 18 dalton). Therefore, the scale difference used in the various studies may explain the inconsistent and contradictory reports. Second, compared to the previous studies, our study population is more affected by cognitive impairment. Amyloid pathology begins in an upstream event in AD pathogenesis and is plateaued at a certain period of cognitive impairment. In addition, the BBB permeability as an early event, as suggested in many experimental and human studies, cannot predict cognition score at the cross-sectional observation in the advanced stage of cognitive impairment. More recent studies hinted the BBB effect on cognition might be modulated by sex, reporting that female MCI subjects showed the inverse relationship between occipital cortex K_trans and MMSE score. In contrast, male MCI subject did not.¹⁸

Despite the reported disparity of cortex BBB permeability effect on cognition, the increasing shreds of evidence agreed that hippocampal K_trans is related to general cognition score²⁹ or at least global cognition status.³⁸ However, further longitudinal studies using a larger sample would be necessary to better understand the effect of cortex BBB changes.

Our study had several limitations. First, the relatively small sample size may have affected the results. Second, the cross-sectional nature of our study cannot elucidate the exact nature of the inter-relationship of BBB and AD-specific biomarkers in a temporal fashion. A longitudinal study with a larger sample size would be necessary for the future. Third, we only measured CSF-based biomarkers in terms of tau pathology. The use of PET biomarkers for both amyloid and tau pathology might help uncover the relationship between BBB and those A/T/N biomarkers. Further, for future evaluation, it would be necessary to evaluate MR-visible microvascular markers (such as WMHV and microbleeds) and BBB permeability measure (entailing the microstructural and the functional information) on a larger scale.^3,14,55,56 Finally, the acquisition time of our DCE imaging sequence is relatively shorter (10 min) than the recommended time (16 min or longer).²² Longer acquisition time may allow high contrast – to – noise ratio to measure slow subtle BBB leakage but is also susceptible to subject motion especially in older patients.^18,22 In clinical research, 10 min appears to be the reasonable best effort to minimize the patient motion and to maintain the contrast to noise ratio as high as possible.¹⁸

Conclusions

Our results suggest a relationship between cortical K_trans and AD biomarkers. Interestingly, this relationship varied according to the presence of Aβ plaque accumulation. cortical K_trans was correlated with CSF neuronal injury marker in the absence of an Aβ plaque. However, in the presence of an Aβ plaque, cortical K_trans was correlated with amyloid, tau markers, and hippocampal atrophy. This implies that the BBB function works differently depending on the accumulation of Aβ plaques or that, at least, its impact on the pathology of neurodegenerative disease is different. Therefore, to promote a more effective drug-development process or to implement improved interventions against BBB breakdown, further study of region-specific and Aβ-specific strategies is suggested.

Supplemental Material

sj-pdf-1-jcb-10.1177_0271678X231180035 - Supplemental material for Blood-brain barrier breakdown is linked to tau pathology and neuronal injury in a differential manner according to amyloid deposition

Supplemental material, sj-pdf-1-jcb-10.1177_0271678X231180035 for Blood-brain barrier breakdown is linked to tau pathology and neuronal injury in a differential manner according to amyloid deposition by Yeonsil Moon, Hong Jun Jeon, Seol-Heui Han, Noh Min-Young, Hee-Jin Kim, Kyoung Ja Kwon, Won-Jin Moon and Seung Hyun Kim in Journal of Cerebral Blood Flow & Metabolism

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Brain Research Program of the National Research Foundation (NRF), funded by the Ministry of Science & ICT (NRF-2018M3C7A1056571) and he Korea Health Technology R&D Project through the Korea Health Technology R&D Project via the Korea Health Industry Development Institute, funded by the Ministry of Health and Welfare, Republic of Korea (grant HU21C0222). No funding bodies were involved in the design, collection, analysis, interpretation, or writing of the manuscript.

Acknowledgements

The authors thank Hui Jin Ryu, MA, Min Young Kim, MA, Chung-Hwan Kang, RT, Ha-young Kim, Su-ji Kim, and Ja-young Jung and Hyun Woo Chung for their support and guidance in the neuropsychological evaluation of patients, serum and imaging data acquisition, and management of this study. The authors also thank all the participants of this study for their dedication to helping research in dementia.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

Yeonsil Moon: Drafting/revision of the manuscript for content, including medical writing for content; Major role in the acquisition of data; Study concept or design; Analysis or interpretation of data. Hong Jun Jeon: Major role in the acquisition of data. Seol-Heui Han: Drafting/revision of the manuscript for content, including medical writing for content. Noh Min-Young: Major role in the acquisition of data. Hee-Jin Kim: Major role in the acquisition of data. Kyoung Ja Kwon: Major role in the acquisition of data. Won-Jin Moon: Drafting/revision of the manuscript for content, including medical writing for content; Major role in the acquisition of data; Analysis or interpretation of data, Seung Hyun KIM: Drafting/revision of the manuscript for content, including medical writing for content; Major role in the acquisition of data; Analysis or interpretation of data. Additional contributions: W-J Moon and SH Kim; These two corresponding authors contributed equally to this work.

ORCID iD

Won-Jin Moon

Supplementary material

Supplemental material for this article is available online.

References

Jack

Jr. Bennett

Blennow

et al. NIA-AA research framework: toward a biological definition of Alzheimer's disease. Alzheimers Dement 2018; 14: 535–562.

Montagne

Zhao

Zlokovic

BV.

Alzheimer's disease: a matter of blood-brain barrier dysfunction?

J Exp Med 2017; 214: 3151–3169.

Zlokovic

BV.

Neurovascular pathways to neurodegeneration in Alzheimer's disease and other disorders. Nat Rev Neurosci 2011; 12: 723–738.

Sweeney

Sagare

Zlokovic

BV.

Blood-brain barrier breakdown in Alzheimer disease and other neurodegenerative disorders. Nat Rev Neurol 2018; 14: 133–150.

Lin

Sur

Liu

et al. Blood-brain barrier breakdown in relationship to Alzheimer and vascular disease. Ann Neurol 2021; 90: 227–238.

Yamazaki

Kanekiyo

Blood-brain barrier dysfunction and the pathogenesis of Alzheimer's disease. IJMS 2017; 18: 1965.

Blair

Frauen

Zhang

et al. Tau depletion prevents progressive blood-brain barrier damage in a mouse model of tauopathy. Acta Neuropathol Commun 2015; 3: 8.

Majerova

Michalicova

Cente

et al. Trafficking of immune cells across the blood-brain barrier is modulated by neurofibrillary pathology in tauopathies. PLoS One 2019; 14: e0217216.

Michalicova

Majerova

Kovac

Tau protein and its role in blood-brain barrier dysfunction. Front Mol Neurosci 2020; 13: 570045.

10.

Montagne

Barnes

Sweeney

et al. Blood-brain barrier breakdown in the aging human hippocampus. Neuron 2015; 85: 296–302.

11.

Henkel

Beers

Wen

et al. Decreased mRNA expression of tight junction proteins in lumbar spinal cords of patients with ALS. Neurology 2009; 72: 1614–1616.

12.

Alvarez

Cayrol

Prat

Disruption of central nervous system barriers in multiple sclerosis. Biochim Biophys Acta 2011; 1812: 252–264.

13.

Bartels

Willemsen

Kortekaas

et al. Decreased blood-brain barrier P-glycoprotein function in the progression of Parkinson's disease, PSP and MSA. J Neural Transm (Vienna) 2008; 115: 1001–1009.

14.

Freeze

Jacobs

HIL

de Jong

et al. White matter hyperintensities mediate the association between blood-brain barrier leakage and information processing speed. Neurobiol Aging 2020; 85: 113–122.

15.

Abdallah

Al-Shami

Yang

et al. Blood-brain barrier disruption increases amyloid-related pathology in TgSwDI mice. IJMS 2021; 22: 1231.

16.

Montagne

Nation

Sagare

et al. APOE4 leads to blood-brain barrier dysfunction predicting cognitive decline. Nature 2020; 581: 71–76.

17.

Lipsman

Meng

Bethune

et al. Blood-brain barrier opening in Alzheimer's disease using MR-guided focused ultrasound. Nat Commun 2018; 9: 2336.

18.

Moon

Lim

Kim

et al. Sex-related differences in regional blood-brain barrier integrity in non-demented elderly subjects. IJMS 2021; 22: 2860.

19.

Lim

Kim

et al. Regional differences in blood-brain barrier permeability in cognitively normal elderly subjects: a dynamic contrast-enhanced MRI-based study. Korean J Radiol 2021; 22: 1152–1162.

20.

Heye

Thrippleton

Armitage

et al. Tracer kinetic modelling for DCE-MRI quantification of subtle blood-brain barrier permeability. NeuroImage 2016; 125: 446–455.

21.

Thrippleton

Backes

Sourbron

et al. Quantifying blood-brain barrier leakage in small vessel disease: review and consensus recommendations. Alzheimers Dement 2019; 15: 840–858.

22.

Wang

Zhang

Qiu

et al. Dysfunction of the blood-brain barrier in cerebral microbleeds: from bedside to bench. Aging Dis 2021; 12: 1898–1919.

23.

Román

Tatemichi

Erkinjuntti

et al. Vascular dementia: diagnostic criteria for research studies. Report of the NINDS-AIREN international workshop. Neurology 1993; 43: 250–260.

24.

Moon

Kwon

et al. Vitamin D deficiency disrupts neuronal integrity in cognitively impaired patients. J Alzheimers Dis 2015; 45: 1089–1096.

25.

Moon

Kim

et al. Role of muscle profile in Alzheimer's disease: a 3-year longitudinal study. Eur Neurol 2019; 81: 209–215.

26.

Moon

Kim

et al. Regional atrophy of the insular cortex is associated with neuropsychiatric symptoms in Alzheimer's disease patients. Eur Neurol 2014; 71: 223–229.

27.

Petersen

Smith

Waring

et al. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999; 56: 303–308.

28.

Shim

. Clinical Application of Plasma Neurofilament Light Chain in a Memory Clinic: A Pilot Study. Dement Neurocogn Disord 2022; 21: 59–70.

29.

Moon

Lim

et al. Hippocampal blood-brain barrier permeability is related to the APOE4 mutation status of elderly individuals without dementia. J Cereb Blood Flow Metab 2021; 41: 1351–1361.

30.

Min

Moon

Jeon

et al. Diagnostic efficacy of structural MRI in patients with mild-to-moderate Alzheimer disease: automated volumetric assessment versus visual assessment. AJR Am J Roentgenol 2017; 208: 617–623.

31.

Yim

Lee

et al. Comparison of automated brain volume measures by NeuroQuant vs. Freesurfer in patients with mild cognitive impairment: effect of slice thickness. Yonsei Med J 2021; 62: 255–261.

32.

Desikan

Ségonne

Fischl

et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 2006; 31: 968–980.

33.

Fischl

FreeSurfer. Neuroimage 2012; 62: 774–781.

34.

Charidimou

Boulouis

Frosch

et al. The Boston criteria version 2.0 for cerebral amyloid angiopathy: a multicentre, retrospective, MRI-neuropathology diagnostic accuracy study. Lancet Neurol 2022; 21: 714–725.

35.

Patlak

Blasberg

RG.

Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. Generalizations. J Cereb Blood Flow Metab 1985; 5: 584–590.

36.

Barthel

Luthardt

Becker

et al. Individualized quantification of brain β-amyloid burden: results of a proof of mechanism phase 0 florbetaben PET trial in patients with Alzheimer's disease and healthy controls. Eur J Nucl Med Mol Imaging 2011; 38: 1702–1714.

37.

Henschel

Conjeti

Estrada

et al. FastSurfer – a fast and accurate deep learning based neuroimaging pipeline. Neuroimage 2020; 219: 117012.

38.

Nation

Sweeney

Montagne

et al. Blood-brain barrier breakdown is an early biomarker of human cognitive dysfunction. Nat Med 2019; 25: 270–276.

39.

Wang

Chen

Han

et al. Relationship between amyloid-β deposition and blood-brain barrier dysfunction in Alzheimer's disease. Front Cell Neurosci 2021; 15: 695479.

40.

Sagare

Bell

Zhao

et al. Pericyte loss influences Alzheimer-like neurodegeneration in mice. Nat Commun 2013; 4: 2932.

41.

van de Haar

Burgmans

Jansen

et al. Blood-brain barrier leakage in patients with early Alzheimer disease. Radiology 2016; 281: 527–535.

42.

Zhang

Wong

van de Haar

et al. Blood-brain barrier leakage is more widespread in patients with cerebral small vessel disease. Neurology 2017; 88: 426–432.

43.

Matsumoto

Yanase

Noguchi-Shinohara

et al. Cerebrospinal fluid/serum IgG index is correlated with medial temporal lobe atrophy in Alzheimer's disease. Dement Geriatr Cogn Disord 2008; 25: 144–147.

44.

Winkler

Sengillo

Bell

et al. Blood-spinal cord barrier pericyte reductions contribute to increased capillary permeability. J Cereb Blood Flow Metab 2012; 32: 1841–1852.

45.

Vos

SJB

Gordon

et al. NIA -AA staging of preclinical Alzheimer disease: discordance and concordance of CSF and imaging biomarkers. Neurobiol Aging 2016; 44: 1–8.

46.

Edison

Donat

Sastre

In vivo imaging of glial activation in Alzheimer's disease. Front Neurol 2018; 9: 625.

47.

Erickson

Wilson

Banks

WA.

In vitro modeling of blood-brain barrier and interface functions in neuroimmune communication. Fluids Barriers CNS 2020; 17: 26.

48.

Kurz

Walker

Rauchmann

et al. Dysfunction of the blood-brain barrier in Alzheimer's disease: Evidence from human studies. Neuropathol Appl Neurobiol 2022; 48: e1278220220202.

49.

Xiong

Callaghan

Jones

et al. ABCG2 is upregulated in Alzheimer's brain with cerebral amyloid angiopathy and may act as a gatekeeper at the blood-brain barrier for abeta(1–40) peptides. J Neurosci 2009; 29: 5463–5475.

50.

Verbeek

Kremer

Rikkert

et al. Cerebrospinal fluid amyloid beta(40) is decreased in cerebral amyloid angiopathy. Ann Neurol 2009; 66: 245–249.

51.

Catak

Zedde

Malik

et al. Decreased CSF levels of β-Amyloid in patients with cortical superficial siderosis. Front Neurol 2019; 10: 439.

52.

Zhu

Hoos

et al. Reduced levels of cerebrospinal fluid/plasma Aβ40 as an early biomarker for cerebral amyloid angiopathy in RTg-DI rats. Int J Mol Sci 2020; 21: 303.

53.

Bowman

Dayon

Kirkland

et al. Blood-brain barrier breakdown, neuroinflammation, and cognitive decline in older adults. Alzheimers Dement 2018; 14: 1640–1650.

54.

Shao

Jann

et al. Comparison between blood-brain barrier water exchange rate and permeability to gadolinium-based contrast agent in an elderly cohort. Front Neurosci 2020; 14: 571480.

55.

Kim

Park

Cho

et al. Assessment of extent and role of tau in subcortical vascular cognitive impairment using 18F-AV1451 positron emission tomography imaging. JAMA Neurol 2018; 75: 999–1007.

56.

Almutairi

Gong

et al. Factors controlling permeability of the blood-brain barrier. Cell Mol Life Sci 2016; 73: 57–77.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.84 MB