Choroid Plexus Free-Water Correlates with Glymphatic function in Alzheimer Disease: The RJNB-D Study

doi:10.21203/rs.3.rs-4680360/v1

Download PDF

Article

Choroid Plexus Free-Water Correlates with Glymphatic function in Alzheimer Disease: The RJNB-D Study

https://doi.org/10.21203/rs.3.rs-4680360/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The free water imaging of choroid plexus (CP) may improve the evaluation of Alzheimer's disease (AD). Our study investigated the role of free water fraction (FWf) of CP in AD by including 216 participants (133 Aβ + participants and 83 Aβ- controls) continuously enrolled in the NeuroBank-Dementia cohort at Ruijin Hospital (RJNB-D). At baseline, Aβ + participants showed higher CP free water fraction (FWf), increased white matter hyperintensity (WMH) volume, and decreased diffusion tensor image analysis of the perivascular space (DTI-ALPS). In Aβ + participants, DTI-ALPS mediated the association between CP FWf and periventricular WMH. CP FWf was associated with cortical Tau accumulation, synaptic loss, hippocampal and cortical atrophy, and cognitive performance. During follow-up, CP FWf increased faster in Aβ + participants than in controls. The findings suggest that elevated CP FWf may indicate impaired glymphatic function and AD neurodegeneration, potentially serving as a valuable biomarker for AD evaluation and progression.

Health sciences/Neurology/Neurological disorders/Dementia/Alzheimer's disease

Biological sciences/Computational biology and bioinformatics/Image processing

Alzheimer Disease

Free-water mapping

choroid plexus

diffusion tensor image analysis of the perivascular space

white matter hyperintensity

Alzheimer's disease (AD) is a neurodegenerative disorder marked by cognitive decline, β-amyloid plaques, hyperphosphorylated Tau, and cortical atrophy¹. Moreover, dysfunction in the cerebrospinal fluid (CSF) glymphatic clearance and white matter hyperintensity (WMH) have also been identified as contributing factors ^2,3.

The choroid plexus (CP) produces 80% of CSF and regulates its dynamics⁴. Comprising a single-layer epithelium, fenestrated capillaries, connective tissue, and immune cells, CP forms the blood-CSF barrier⁵ and aids immune surveillance^6,7. CP enlargement occurs in neurodegenerative⁸, neuroinflammatory⁵, and cerebral small vessel diseases⁹. While previously linked to aging and reduced function^10,11, the mechanisms behind CP in neurodegeneration are unclear. Despite this, there is a scarcity of research focusing on components of the CP. Given its role in producing CSF and its location within the ventricles, fluid is one of the predominant components within the CP structure.

Free water refers to unconstrained water molecules in fluid. Diffusion MRI techniques can measure free water in brain tissue^12,13. Studies show increased free water fraction (FWf) in white matter tracts¹⁴, cortical and subcortical regions as AD progresses^15,16. Given the choroid plexus's role in CSF regulation, CP FWf may be crucial for brain fluid circulation.

Furthermore, in the brain glymphatic system, CSF enters the brain parenchyma through the perivascular spaces surrounding arteries to clear interstitial fluid waste products¹⁷. Prior studies using diffusion tensor image analysis along the perivascular space (DTI-ALPS) have shown an association between glymphatic dysfunction and the severity of WMH^9,18. Meanwhile, our previous study revealed an association between WMH and remote cortical AD biomarkers¹⁹. However, the relationship between choroid plexus free water fraction (CP FWf) and glymphatic clearance or WMH remains unclear.

We hypothesized that elevated CP FWf in AD correlates with glymphatic function and AD biomarkers across the disease continuum. We undertook this study to: 1) investigate CP FWf changes in AD; 2) examine the relationship between CP FWf and glymphatic clearance function, WMH, as well as positron emission tomography (Aβ-PET, Tau-PET, synaptic vesicle glycoprotein 2A density-PET) and blood markers other than Aβ and Tau (glial fibrillary acidic protein, GFAP; neurofilament light chain, NFL; Neurogranin, NRGN; and Tumor Necrosis Factor-α, TNFα); 3) explore potential mediation factors (such as Tau accumulation, synaptic density) between CP FWF and cognitive tests; and 4) validate the impact of CP FWf in AD using longitudinal data over a 12-month follow-up period.

Demographics and group comparisons

Table 1 summarizes the demographic and imaging characteristics stratified by Aβ status for all patients (n = 216, Fig. 1A) included in the cross-sectional analyses. The average age was 69.20 ± 8.00 years; 128 were females. Table 2 lists the characteristics of the 76 participants included in the longitudinal analyses. Their average age was 70.09 ± 7.29 years; 49 were females.

Table 1

Baseline Demographic and MRI Characteristics (n = 216)
	Aβ+ (n = 133)	Aβ- (n = 83)	p-value
Age	70.59 ± 7.62	66.98 ± 8.13	0.001
Female	77 (58%)	51 (61%)	0.605
Education	11.56 ± 4.08	12.16 ± 3.43	0.266
ApoE 4 carrier^$	67 (62%)	15 (23%)	< 0.001
MMSE	21.07 ± 6.36	27.18 ± 4.02	< 0.001
CP FWf	0.78 ± 0.04	0.76 ± 0.06	0.002
CPV	0.98 ± 0.20	0.94 ± 0.23	0.143
CP FA^#	-1.65 ± 0.25	-1.64 ± 0.23	0.841
CP MD^#	-7.64 ± 0.16	-7.67 ± 0.10	0.091
CP CBF^#	3.80 ± 0.53	3.77 ± 0.85	0.746
DTI-ALPS	1.15 ± 0.17	1.23 ± 0.16	0.001
pWMH^#	-3.83 ± 1.52	-4.78 ± 1.73	< 0.001
dWMH^#	-5.74 ± 1.70	-6.44 ± 1.73	0.005
#, The log-transformed values were used
$, Different sample size for ApoE genotype: Aβ+, n = 108; Aβ-, n = 66.

Table 2

Month 12 Demographic and MRI Characteristics (n = 76)
	Aβ+ (n = 46)	Aβ- (n = 30)	p-value
Age	72.33 ± 6.22	66.67 ± 7.59	0.001
Female	28 (61%)	21 (70%)	0.416
ApoE 4 carrier^$	27(63%)	5 (19%)	< 0.001
Baseline CP FWf	0.77 ± 0.04	0.75 ± 0.05	0.020
Month 12 CP FWf	0.80 ± 0.04	0.76 ± 0.06	0.001
ΔCP FWf	0.03 ± 0.04	0.01 ± 0.04	0.046
Baseline DTI-ALPS	1.20 ± 0.14	1.27 ± 0.15	0.081
Month 12 DTI-ALPS	1.13 ± 0.18	1.23 ± 0.20	0.052
ΔDTI-ALPS	-0.08 ± 0.11	-0.05 ± 0.10	0.226
Baseline pWMH^#	-4.01 ± 1.52	-4.83 ± 1.49	0.022
Month 12 pWMH^#	-4.07 ± 2.20	-4.82 ± 1.96	0.126
ΔpWMH	0.01 ± 0.03	0.00 ± 0.01	0.078
Baseline Tau SUVR^#	0.64 ± 0.46	-0.03 ± 0.14	< 0.001
Month 12 Tau SUVR^#	0.67 ± 0.51	-0.10 ± 0.16	< 0.001
Baseline GFAP	153.05 ± 93.36	76.94 ± 64.01	< 0.001
Month 12 GFAP	167.52 ± 84.90	128.61 ± 71.91	0.146
#, Nature logarithm was taken for normalization.
$, Different sample size for ApoE genotype: Aβ+, n = 43; Aβ-, n = 26.

At baseline, age (70.59 ± 7.62 vs 66.98 ± 8.13, p = 0.001) and ApoE 4 carrier status (62% vs 23%, p < 0.001) differed between Aβ + group and Aβ- controls. Increased FWf of CP (0.78 ± 0.04 vs 0.76 ± 0.06, p = 0.002), increased volume of WMH (pWMH, -3.83 ± 1.52 vs -4.78 ± 1.73, p < 0.001; dWMH, -5.74 ± 1.70 vs -5.74 ± 1.70, p = 0.005, both log-transformed), and decreased DTI-ALPS (1.15 ± 0.17 vs 1.23 ± 0.16, p = 0.001) were observed in Aβ + participants (Table 1, Fig. 2A-C). Conversely, the Aβ + group and Aβ- controls exhibited similar volumes of CP, as well as similar values for FA, MD, and CBF of CP (Table 1, all p > 0.050).

After adjusting for age, sex and ApoE genotype, FWf of CP (β = 10.02, p = 0.015), DTI-ALPS (β = -3.19, p = 0.007), pWMH (β = 0.33, p = 0.007) and dWMH (β = 0.28, p = 0.013) respectively demonstrated an independent association with Aβ positivity (Supplementary Table 1–4).

The relationship between CP FWf, glymphatic clearance and WMH

Across all participants, the CP FWf correlated with DTI-ALPS (r = -0.47, p < 0.001), pWMH (r = 0.46, p < 0.001), and dWMH (r = 0.21, p = 0.003). Mediation analysis revealed a partial mediation effect of DTI-ALPS for the association between CP FWf and pWMH (indirect effect standardized-β = 0.26, p < 0.001, mediated effect = 32.59%; Fig. 2D). However, no mediation effect of DTI-ALPS was observed for the association between CP FWf and dWMH (indirect effect standardized-β = 0.09, p = 0.236).

Within the Aβ + group, CP FWf correlated with DTI-ALPS (r = -0.45, p < 0.001), as well as between CP FWf and pWMH (r = 0.40, p < 0.001) but not dWMH (r = 0.11, p = 0.213). A partial mediation effect of DTI-ALPS for the association between CP FWf and pWMH was observed (indirect effect standardized-β = 0.25, p = 0.003, mediated effect = 35.76%; Fig. 2E).

The Aβ- controls displayed similar associations between CP FWf and DTI-ALPS (r = -0.44, p < 0.001), and pWMH (r = 0.45, p < 0.001). A marginal mediation effect of DTI-ALPS for the association between CP FWf and pWMH was observed (indirect effect standardized-β = 0.25, p = 0.034, mediated effect = 24.67%; Fig. 2F).

The impact of CP FWf across the AD continuum

The relationship between CP FWf, glymphatics and AD imaging markers

We observed significant associations between CP FWf and AD imaging markers (cortical Tau SUVR, cortical SV2A SUVR, hippocampus volume, cortex volume), as well as MMSE score (Fig. 2G, Supplementary Table 5).

After adjusting for age, sex, and ApoE genotype in multivariable linear regression models, an increase in CP FWf was correlated with increased Tau-PET accumulation (Tau SUVR, β = 9.27, p = 0.002, Fig. 3A, Supplementary Table 6) and decreased synaptic density (SV2A SUVR, β = -3.43, p = 0.017, Fig. 3A, Supplementary Table 7). However, CP FWf did not corelated with FBP SUVR (β = 0.98, p = 0.103). The vertex-wise GLM analysis revealed a negative correlation between CP FWf and cortical thickness primarily in the bilateral post cingulate gyrus, precuneus, temporal, and insular lobes. Conversely, CP FWf positively correlated with Tau SUVR in the bilateral insular, temporal regions, and precuneus (Fig. 2H). CP FWf negatively correlated with SV2A SUVR in the right inferior parietal gyrus (Extended Data Fig. 1)

The relationship between CP FWf and blood-based neural biomarkers

Due to restrictions imposed by variability in the availability of adequate blood samples in some patients of the Aβ + cohorts, we specifically marked the sample size in these analyses. Within the Aβ + group, CP FWf was associated with NFL (β = 3.12, p = 0.026, n = 92), GFAP (β = 5.28, p < 0.001, n = 92), NRGN (β = 5.70, p = 0.028, n = 82), and TNF-α (β = 10.83, p = 0.009, n = 81) (Supplementary Table 8–11).

The relationship between CP FWf, glymphatic clearance, WMH and cognition

Increased CP FWf was associated with worse cognitive performance, including overall cognition (MMSE, β = -53.36, p < 0.001), verbal function (AFT, β = -29.83, p = 0.014), and executive function (CDT20, β = -35.29, p = 0.025) within the Aβ + group (Fig. 3A, Supplementary Table 12–14). Similarly, DTI-ALPS and pWMH were also correlated with MMSE (DTI-ALPS, β = 13.17, p < 0.001; pWMH, β = -1.57, p < 0.001), AFT (DTI-ALPS, β = 8.19, p = 0.006; pWMH, β = -0.92, p = 0.003), and CDT20 (DTI-ALPS, β = 13.02, p < 0.001; pWMH, β = -1.60, p < 0.001)

We observed a partial mediation effect of Tau SUVR (indirect effect standardized-β = -0.18, p = 0.043, mediated effect = 40.88%, n = 110, Fig. 3B) and SV2A SUVR (indirect effect standardized-β = -0.21, p = 0.089, mediated effect = 21.73%, n = 64, Fig. 3C) on the association between CP FWf and MMSE. While GFAP exhibited a full mediation effect (mediated effect = 38.69%, n = 111, Fig. 3D), NFL, NRGN, and TNF-α did not significantly mediate the association between CP FWf and global cognitive performance assessed by MMSE (Fig. 3E-G). These findings suggest that the impact of CP FWf on cognitive performance is mediated via Tau- and GFAP-related pathways.

Longitudinal study

CP FWf increased faster in Aβ + than Aβ- participants.

We observed a significant increase in CP FWf during 12-month follow-up period than baseline (F-statistic = 16.673, corrected p < 0.001, Supplementary Table 15) in Aβ + patients. More importantly, Aβ + patients demonstrated a faster increment in CP FWf compared to Aβ- controls (time × group interaction effect: F-statistic = 4.118, corrected p = 0.046, Fig. 4A, Supplementary Table 15).

The rates of increase in CP FWf did not differ between ApoE 4 carriers (n = 27) and non-carriers (n = 16) in the Aβ + group (F-statistic = 0.004, p = 0.949, Fig. 4B).

ΔCP FWf paralleled ΔDTI-ALPS, and was faster than ΔpWMH, ΔTau-PET and ΔGFAP

In Aβ + participants, annual changes in CP FWf, DTI-ALPS, and pWMH volume from baseline to the 12-month follow-up were calculated as ΔCP FWf, ΔDTI-ALPS, and ΔpWMH, respectively. Through Spearman correlation analysis, it was found the ΔCP FWf during the 12-month follow-up period paralleled ΔDTI-ALPS (ρ = -0.42, p = 0.006, Fig. 4C), but not ΔpWMH (ρ = 0.21, p = 0.173). After adjusting for age and sex, the association between ΔCP FWf and ΔDTI-ALPS remained significant (β = -1.02, p = 0.016). The growth rate of CP FWf exceeded that of pWMH (interaction effect, F-statistic = 11.201, corrected p = 0.001), Tau SUVR (interaction effect, F-statistic = 6.804, corrected p = 0.011) and GFAP (interaction effect, F-statistic = 4.430, corrected p = 0.039, Fig. 4D-F).

In this study, we found increased FWf of CP, increased volume of WMH, and decreased DTI-ALPS at baseline in Aβ + participants compared to Aβ- controls. And we noted elevated CP FWf exacerbates pWMH, partially mediated by the glymphatic system (DTI-ALPS), revealing a potential mechanism linking CSF dynamics to white matter lesions. We also observed significant associations between CP FWf and AD imaging markers (cortical Tau SUVR, cortical SV2A SUVR, hippocampus volume, cortex volume) and cognitive performance. In longitudinal analyses, we observed a significant and faster increase in CP FWf during a 12-month follow-up period in Aβ + patients compared to Aβ- controls, and the rate of progression of CP FWf exceeded that of WMH.

The CP is a highly vascularized structure that regulates CSF production and clears neurotoxic proteins, playing a crucial role in the glymphatic system. Recent studies show the glymphatic system's pivotal role in aging and neurodegeneration^20,21. Post-mortem and in vivo studies have revealed the CP plays a role in the pathogenesis of AD²². Other studies have found an association between increased CP volume and greater cognitive impairment in AD.^8,23–25 Additionally, CP volume has been linked to levels of pathological proteins like Aβ and Tau in the CSF, indicating that the CP may participate the clearance of Aβ and Tau²³. Prior results regarding the relationship between CP volume and Aβ positivity were inconsistent^8,24,25. In our study, CP volume did not relate to Aβ positivity. These disagreements may be partially attributed to the contamination of partial volume effect of CSF in calculation CP volume by traditional methods²⁶, and also the morphological nature of CP volume, which may display a relatively delayed response to Aβ deposition. In the current study, we evaluated various novel imaging biomarkers which reflects CP content and microstructure, including CP FWf, CP FA, CP MD, and CP CBF. CP FWf was the only marker that displayed a significant association with Aβ positivity. Free-water mapping is derived from a bi-tensor model that enables the distinction of diffusion properties between water in brain tissue and water in extracellular space (free water), thus offering more comprehensive microstructural insights compared to the traditional single-tensor model^13,27.

The DTI-ALPS index is a reliable parameter for evaluating glymphatic system, reflecting its function in maintaining CSF homeostasis and clearance of toxic proteins^28,29. Previous studies showed that in subjects with WMH or CSVD, CP enlargement and reduced DTI-ALPS were both associated with WMH growth^9,24. Moreover, DTI-ALPS mediated the association between CP and WMH burden, suggesting that the pathogenesis of WMH may result from impaired glymphatic clearance⁹. We demonstrated that CP FWf is a reliable indicator of glymphatic function by examining its relationship with ALPS and WMH. We noticed that in Aβ + population CP FWf was associated with pWMH, but not with dWMH, and there is a partial mediating effect of DTI-ALPS between CP FWf and pWMH.

A recent study by Jeong et al demonstrated a positive correlation between CP enlargement and Aβ burden²⁵. However, subgroup analysis of this study suggested that the correlation was only significant in the AD non-dementia group, but not in the AD dementia group²⁵, which is possibly due to the ceiling effect of Aβ burden, as Aβ burden is saturated in dementia state³⁰. Likewise, although we also found that CP FWf was significantly related to Aβ positivity, we did not detect a correlation with Aβ burden quantified by FBP SUVR. A recent report demonstrated a positive correlation between global Tau burden and CP enlargement in AD³¹, We expanded on these findings by using vertex-wise analysis to provide topographical information and showed that CP FWf changed in parallel with Tau accumulation mainly in bilateral insular, temporal lobes, and precuneus. Furthermore, we found that increased Tau burden mediated the detrimental effects of CP FWf on cognitive performance. This finding adds clinical evidence and support to a recent study indicating that glymphatic dysfunction impairs cognition through poor Tau clearance in AD mouse models³².

In addition to the clearance of neurotoxic proteins, glymphatic system can also modulate neuroinflammation^33,34. We explored the relationship between peripheral neuroinflammatory factors in the blood and CP FWf. GFAP correlated with CP FWf, and substantially mediated the negative impact of CP FWf on cognitive function. Astrocytes are serving as critical components of glymphatic system by ensheathing the brain vasculature with their specialized endfeet.^17,34 GFAP is a structural protein in astrocytes and is a marker of reactive gliosis related to aging³⁵. Our findings, albeit preliminary, are in agreement with the results of another study, showing that blood GFAP is related to impaired glymphatic function and global cognition in community-dwelling older adults³⁶ and provide new evidence for involvement of GFAP-related pathway in glymphatic dysfunction in AD pathology.

Our study has limitations. The sample size of the longitudinal cohort is relatively small, partly due to the cohort being relatively new, resulting in approximately half of the enrolled subjects not reaching the 12-month follow-up time point. Furthermore, the 12-month follow-up duration is relatively short, and certain imaging markers such as the volume of pWMH may not exhibit significant changes within this timeframe. A longer follow-up period may be necessary to capture more pronounced ΔpWMH. However, ΔCP FWf was marked within the short time interval, indicating the sensitivity of CP FWf in monitoring disease progression. Lastly, the number of blood samples and analyses of various blood markers were limited and may have influenced the related results. Therefore, the results of our blood markers analyses should be considered exploratory and hypothesis generating.

In conclusion, our findings suggest that CP FWf emerges as a novel imaging biomarker complementing structural MRI and molecular PET imaging for assessing glymphatic function and CP involvement in AD. By reflecting the multifactorial nature of AD, CP FWf shows promise for early diagnosis and accurate monitoring of AD progression, potentially enabling tailored and timely treatment strategies.

This study was approved by the ethics committee at Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, China. All participants or caregivers provided written informed consent. The study adhered to the 1964 Declaration of Helsinki and its amendments. The study was registered on ClinicalTrials.gov (NCT05623124).

Participants and study design

The data for this study were obtained from the Ruijin NeuroBank of Alzheimer's Disease and Dementia (RJNB-D) cohort study. The participants underwent comprehensive assessments including multimodal MRI, Aβ PET, Tau PET, ¹⁸F-SynVesT-1 PET measuring synaptic density, neuropsychological tests, ApoE genotyping and blood tests for GFAP, NFL, NRGN and TNFα (Fig. 1A). Multimodal MRI included 3D high-resolution T1, fluid-attenuated inversion recovery (FLAIR), multi-shell diffusion MRI (dMRI) and pseudo-continuous arterial spin labelled (pCASL) perfusion MRI. Neuropsychological assessments included Mini-Mental State Examination (MMSE, Chinese Version)³⁷, 20-minute auditory verbal learning test (AVLT)³⁸, Clock-drawing test (CDT)³⁹, Boston naming test (BNT), shape-trail test (STT-A and STT-B)⁴⁰, and animal fluency test (AFT) ⁴⁰.

We collected a 20mL venous blood sample from participants at the visits. Samples were centrifuged to isolate serum and plasma before being stored at -80℃. For blood biomarkers, protein concentrations in ethylenediaminetetraacetic acid (EDTA) plasma were assessed using the ultra-sensitive single-molecule array (Simoa) platform. Specifically, GFAP and NFL levels were quantified utilizing the Neurology 4-Plex E Advantage kit (QTX-103670, Quanterix, Billerica, Massachusetts, United States). Serum concentrations of NRGN and TNFα were determined using the Luminex 200 system (Luminex, Austin, TX, USA) with the Human ProcartaPlex Mix & Match 3-plex kit (PPX-03-MXMFYR9, Thermo Fisher Scientific).

In the cross-sectional analysis, we studied a total of 216 participants. Among them, the 133 participants with cognitive complaints and Aβ positivity were defined as Aβ + group. As across the AD continuum, participants in the Aβ + group were further diagnosed as AD or mild cognitive impairment (MCI). The criteria for diagnosing AD and MCI adhered to the research criteria established by the National Institute on Aging-Alzheimer's Association (NIA-AA) workgroups in 2016⁴¹. Meanwhile, another 83 cognitively normal participants with negative Aβ were included as controls.

For the longitudinal analysis, 76 out of the 216 participants at baseline underwent complete 12-month follow-up assessments, during which they were repeatedly administered neuropsychological tests, multimodal MRI and PET scans. The follow-up assessments were performed 12 months after baseline visit and the assessments (including neuropsychological tests, MRI, PET, etc.) were completed within two weeks.

Acquisition of MRI and PET images

MRI data were acquired on a 3T scanner (uMR 890, United Imaging Healthcare, Shanghai, China) with a dedicated 64-channel head coil. As our previous protocol of RJNB-Dementia dataset^19,42, the T1-weighted structural images were acquired with a 3D fast spoiled gradient-echo sequence, with 0.75 mm isotropic voxels (field of view read: 240 mm, output images of 360 × 439× 480 slices, repetition time (TR) / inversion time (TI) = 7.5 ms /1100 ms, echo time (TE) = 3.4 ms, flip angle = 7 degree). The 3D FLAIR acquisition parameters were: TR = 6000 ms, TE = 408 ms, and TI = 1841ms, with 0.75 mm isotropic voxels. Diffusion images were acquired with an echo-planar imaging (EPI) sequence, and the parameters were: TR/TE = 3500/77.3 ms, voxel size = 1.5 × 1.5 × 1.5 mm, 100 slices, multiple band = 10, diffusion direction = 32/64, b = 1500/3000, phase encoding direction = PA. A pair of four b0 images with reversed phase-encode blips (AP and PA) was acquired to correct susceptibility-induced distortion. The ASL scans were performed using a 5-delay pCASL protocol with background suppressed 3D GRASE (gradient and spin echo) readout (labeling pulse duration = 1.8 s, post label delay = 0.5-1.0-1.5-2.0-2.5 s, matrix = 64 × 64, voxel size = 3.5 × 3.5 × 3.5 mm, 6 pairs of tag and control for each delay).

We utilized PET scans to evaluate the spatial accumulation of Aβ and Tau, as well as synaptic density in the study. The PET scans were conducted using a 3T whole-body PET/MR scanner (uPMR 790, United Imaging, China). Participants were administered an intravenous injection of ¹⁸F Florbetapir (FBP) at an average dose of 3.7 MBq/kg body weight to visualize Aβ accumulation. Static FBP-PET data were acquired in sinogram mode 50 minutes after the injection. Additionally, all participants underwent a static ¹⁸F-MK-6240 PET scan to image Tau. Furthermore, a ¹⁸F-SynVesT-1 PET scan was performed to visualize synaptic vesicle glycoprotein 2A (SV2A) as previously described⁴². Participants refrained from taking drugs targeting SV2A for at least 24 hours before the ¹⁸F-SynVesT-1 PET scan. A 30-minute static PET scan was initiated at 60 minutes post-injection of 18F-SynVesT-1 (~ 3.7 MBq/kg body weight). Each participant had the three PET scans on different days within two weeks; the scans were scheduled at least three days apart from each other.

CP and WMH segmentation

The region of interest (ROI) of CP was segmented from 3D-T1 weighted images. Both baseline and follow-up 3D-T1 images were segmented automatically based on FreeSurfer 7.1.1 platform as previously described^43,44. The segmentation of brain volumes is based on an atlas containing probabilistic information regarding the location of structures. This pre-processing step involves motion correction of 3D T1 images, brain extraction, segmentation of subcortical white matter and deep gray matter volumetric structures, intensity normalization, white matter surface smoothing, topology correction, and optimization of pial and white matter surfaces. The total intracranial volume (TIV) and the CP of the lateral ventricle were determined through this process. In addition, the cortical thickness was determined by measuring the distance between the white matter and pial surfaces. All final automatic segmentations underwent quality checks by a senior neurologist (X.X.) who was blinded to clinical and other image data. The CP ROI were used as mask for analysis after registration with the following MRI modalities. The CP volumes (CPV) were expressed as ratios to TIV.

The WMH was defined as subcortical hyperintensities without cavitation on T2 FLAIR based on the recommendations of Standards for Reporting Vascular Changes on Neuroimaging (STRIVE)⁴⁵. We utilized FSL's Brain Intensity AbNormality Classification Algorithm (BIANCA, FMRIB Software Library, Oxford, UK), version 6.0, a fully automated, supervised k-nearest neighbor (k-NN) algorithm for WMH segmentation. The binary masks of total hyperintensities were generated using BIANCA and then carefully inspected and manually corrected to produce the final WMH binary masks. These derived WMH masks were further categorized into periventricular WMH (pWMH) and deep WMH (dWMH) by applying a 10 mm distance threshold from the ventricles¹⁹. Subsequently, all maps were visually inspected and manually corrected as needed. The volumes of pWMH and dWMH were expressed as ratios to the total volume of white matter.

Diffusion MRI analysis

Pre-processing: The pre-processing pipelines for diffusion MRI generally followed the protocols of an image-processing pipeline developed by UK Biobank⁴⁶. The original diffusion MRI data were corrected for susceptibility-induced distortion by top-up method⁴⁷. Then the data were further corrected for head motion and eddy currents using the Eddy tool in FSL. The output data were computed into diffusion tensor maps using the “DTIFIT” function of FSL.

Fitting diffusion tensors: The DTIFIT estimated a diffusion tensor at each voxel of the brain, and outputs a set of images including fractional anisotropy (FA) and mean diffusivity (MD), as well as voxel-wise diffusivity in the directions of the x-axis (Dx), y-axis (Dy), and z-axis (Dz) images. We linearly aligned individual FA/MD maps with the T1-weighted image. The registration process is a combination of affine transformation⁴⁸. Then, the CP mask in T1-weighted image could be used to extract mean FA and MD value in the CP. Visual examinations were performed to identify any co-registration bias.

DTI-ALPS: The FA map of each subject was registered to the FA map template in the standard MNI space, and the accuracy of the co-registration was visually confirmed. The Dx, Dy, Dz maps of the subject were then warped based on the registration matrix obtained from the FA map. The ROIs corresponding to the left projection fibers (superior and posterior corona radiata) and association fibers (superior longitudinal fasciculus) were extracted using atlas labels. The range of ROIs in the craniocaudal direction was limited to where the x-axis lines were perpendicular to the lateral ventricle bodies. A mask with FA values greater than 0.2 was applied to exclude CSF voxels. Finally, the Dx, Dy, and Dz values in these ROIs on projection fibers and association fibers were determined as Dxproj, Dyproj, Dzproj, Dxassoc, Dyassoc, Dzassoc, respectively. The DTI-ALPS index was automatically computed using the following Eq. 4⁹:

DTI-ALPS index= [(Dxproj + Dxassoc) / (Dyproj + Dzassoc)]

Free-Water image of CP: The FWf can be measured quantitatively by neurite orientation dispersion and density imaging (NODDI) model derived from multi-shell diffusion MRI. NODDI models the voxel contents as neurites (thin pipes, nominally axons), the spaces between them, and freely diffusing water^50,51. NODDI estimation for FWf was done using Accelerated Microstructure Imaging via Convex Optimization⁴⁶. After registered with T1-weighted image, the averaged Fwf of CP was calculated from the left and right CP ROI (Fig. 1B).

Cerebral blood flow in CP

We used FSL BASIL toolbox to estimate perfusion in the CP. The key steps in BASIL pipeline includes motion correction, registration, calibration using M0 image, and Bayesian kinetic modelling to estimate perfusion maps, with spatial regularization and partial volume correction (PVC). Following these steps, the whole brain perfusion map was registered to the T1-weighted image, and the accuracy of the co-registration was visually confirmed. Subsequently, the cerebral blood flow (CBF) of the choroid plexus was extracted from the CP region of interest segmented from the T1-weighted image.

Surface-based PET images for Aβ, Tau and SV2A

An automated pipeline was utilized to derive cortical standardized uptake value ratios (SUVR) using the PETSurfer toolbox in Freesurfer, with the cerebellum cortex serving as the reference region⁵². Initially, high-resolution segmentation was generated from structural T1 images to facilitate PVC techniques. Subsequently, registration of PET and anatomical images was conducted and visually validated. To mitigate partial volume effects resulting from cortical atrophy in AD, the extended Muller–Gartner method was employed as a PVC approach⁵³. Surface-based maps of Aβ, Tau, and SV2A were smoothed on a two-dimensional surface using a Gaussian kernel with a full width at half maximum (FWHM) of 5mm. These biomarker maps were then utilized for vertex-wise analysis and mean cortical SUVRs were extracted for spatial correlation analyses.

Statistical analyses

Group comparisons used t-tests for continuous variables and Chi-square tests for categorical variables. Non-normal distributions were log-transformed. Logistic regression was used to adjust for age, sex, and ApoE genotype.

Within each group, correlations were assessed using Pearson's or Spearman's methods based on variable distribution. Multivariable linear regression analyzed CP FWf and other imaging measures, adjusting for age, sex, and ApoE genotype. Bonferroni correction addressed multiple comparisons. Vertex-wise correlation analysis using generalized linear model (GLM), controlled for age and sex, demonstrated spatial correlations between CP FWf and AD markers. Permutation tests reduced false positivity rates. Mediation analyses were conducted using PROCESS v3.2 for SPSS, adjusted for age. Models tested glymphatic function's mediation between CP FWf and WMH, and blood markers' mediation between CP FWf and cognition.

Longitudinal CP FWf changes were assessed using repeated measures ANOVA. Alteration rates between Aβ + patients and Aβ- controls were compared using a linear mixed model. The relationship between CP FWf and DTI-ALPS changes was explored in the Aβ + group. Rates of longitudinal changes in CP FWf and related metrics were compared after standardizing baseline measures. Analyses were conducted using R (4.3.3) or SPSS, with significance at P < 0.05 (two-tailed).

Data availability

For all main figures with datasets, the extracted traces from the raw image files are available on Figshare (https://figshare.com/articles/dataset/Choroid_Plexus_Free-Water_Correlates_with_Glymphatic_Dysfunction_and_Tracks_Neurodegeneration_in_Alzheimer_s_Disease_The_RJNB-D_Study/26043871).

Code availability

The preprocessing codes for multimodal MR and PET are available as a Github repository (https://github.com/colorfulbrain/CP_FWf). The R/SPSS code used for the main analysis are available from the corresponding author upon request.

Acknowledgements

This study was supported by National Natural Science Foundation of China (82271441, 82171473, 81901180), Shanghai Rising-Star Program (21QA1405800), National Key Research and Development Program of China, Scientific and technological innovation 2030 - Major Projects (2022ZD0213800).

Author contributions

X. X. and B. L. designed the study; X. X., X. Y. M. S. and B. L. wrote the initial draft of the manuscript; X. X., X. Y., J. Z., R. S., F. X. and B. L. performed data analysis; Y. W., Y. G., J. L., and Y. D. consulted on statistical methods; X. X., X. Y. and W. X. generated tables and figures; Y. Z.,Q. H., W., W. performed neurological function evaluation and collected blood samples and imaging data; all authors read the manuscript and provided input and consultation. X. X., X. Y., M. S. and B. L. finalized the manuscript. F. X. and B. L. oversaw the study and provided direction, funding, and resources.

Conflict of interest

The authors declare no conflicts of interest.

Consent Statement

All human subjects provided informed consent.

Jack CR, Bennett DA, Blennow K et al (2016) A/T/N: An unbiased descriptive classification scheme for Alzheimer disease biomarkers. Neurology 87(5):539–547. 10.1212/WNL.0000000000002923
Deane R, Bell RD, Sagare A, Zlokovic BV (2009) Clearance of amyloid-beta peptide across the blood-brain barrier: implication for therapies in Alzheimer’s disease. CNS Neurol Disord Drug Targets 8(1):16–30. 10.2174/187152709787601867
Garnier-Crussard A, Cotton F, Krolak-Salmon P, Chételat G (2023) White matter hyperintensities in Alzheimer’s disease: Beyond vascular contribution. Alzheimers Dement 19(8):3738–3748. 10.1002/alz.13057
Serot JM, Zmudka J, Jouanny P (2012) A possible role for CSF turnover and choroid plexus in the pathogenesis of late onset Alzheimer’s disease. J Alzheimers Dis 30(1):17–26. 10.3233/JAD-2012-111964
Bitanihirwe BKY, Lizano P, Woo TUW (2022) Deconstructing the functional neuroanatomy of the choroid plexus: an ontogenetic perspective for studying neurodevelopmental and neuropsychiatric disorders. Mol Psychiatry 27(9):3573–3582. 10.1038/s41380-022-01623-6
Damkier HH, Brown PD, Praetorius J (2013) Cerebrospinal fluid secretion by the choroid plexus. Physiol Rev 93(4):1847–1892. 10.1152/physrev.00004.2013
Ghersi-Egea JF, Strazielle N, Catala M, Silva-Vargas V, Doetsch F, Engelhardt B (2018) Molecular anatomy and functions of the choroidal blood-cerebrospinal fluid barrier in health and disease. Acta Neuropathol 135(3):337–361. 10.1007/s00401-018-1807-1
Choi JD, Moon Y, Kim HJ, Yim Y, Lee S, Moon WJ (2022) Choroid Plexus Volume and Permeability at Brain MRI within the Alzheimer Disease Clinical Spectrum. Radiology 304(3):635–645. 10.1148/radiol.212400
Li Y, Zhou Y, Zhong W et al (2023) Choroid Plexus Enlargement Exacerbates White Matter Hyperintensity Growth through Glymphatic Impairment. Ann Neurol 94(1):182–195. 10.1002/ana.26648
Bouhrara M, Walker KA, Alisch R (2024) Association of Plasma Markers of Alzheimer’s Disease, Neurodegeneration, and Neuroinflammation with the Choroid Plexus Integrity in Aging. Aging Dis Published online January 8. 10.14336/AD.2023.1226
Scarpetta V, Bodaleo F, Salio C, Agarwal A, Sassoè-Pognetto M, Patrizi A (2023) Morphological and mitochondrial changes in murine choroid plexus epithelial cells during healthy aging. Fluids Barriers CNS 20(1):19. 10.1186/s12987-023-00420-9
Wang Y, Wang Q, Haldar JP et al (2011) Quantification of increased cellularity during inflammatory demyelination. Brain 134(Pt 12):3590–3601. 10.1093/brain/awr307
Pasternak O, Sochen N, Gur Y, Intrator N, Assaf Y (2009) Free water elimination and mapping from diffusion MRI. Magn Reson Med 62(3):717–730. 10.1002/mrm.22055
Li H, Jacob MA, Cai M et al (2024) Meso-cortical pathway damage in cognition, apathy and gait in cerebral small vessel disease. Brain Published online May 6:awae145. 10.1093/brain/awae145
Nakaya M, Sato N, Matsuda H et al (2024) Assessment of Gray Matter Microstructural Alterations in Alzheimer’s Disease by Free Water Imaging. J Alzheimers Dis Published online May 15. 10.3233/JAD-231416
Chiu SY, Chen R, Wang WE et al (2024) Longitudinal Free-Water Changes in Dementia with Lewy Bodies. Mov Disord 39(5):836–846. 10.1002/mds.29763
Hablitz LM, Nedergaard M (2021) The glymphatic system. Curr Biol 31(20):R1371–R1375. 10.1016/j.cub.2021.08.026
Zhang W, Zhou Y, Wang J et al (2021) Glymphatic clearance function in patients with cerebral small vessel disease. NeuroImage 238:118257. 10.1016/j.neuroimage.2021.118257
Zhang J, Chen H, Wang J et al (2024) Linking white matter hyperintensities to regional cortical thinning, amyloid deposition, and synaptic density loss in Alzheimer’s disease. Alzheimers Dement. Published online April 22. 10.1002/alz.13845
Nedergaard M, Goldman SA (2020) Glymphatic failure as a final common pathway to dementia. Science 370(6512):50–56. 10.1126/science.abb8739
Dai Z, Yang Z, Chen X et al (2023) The aging of glymphatic system in human brain and its correlation with brain charts and neuropsychological functioning. Cereb Cortex 33(12):7896–7903. 10.1093/cercor/bhad086
Čarna M, Onyango IG, Katina S et al (2023) Pathogenesis of Alzheimer’s disease: Involvement of the choroid plexus. Alzheimer’s Dement 19(8):3537–3554. 10.1002/alz.12970
Tadayon E, Pascual-Leone A, Press D, Santarnecchi E, Alzheimer’s Disease Neuroimaging Initiative (2020) Choroid plexus volume is associated with levels of CSF proteins: relevance for Alzheimer’s and Parkinson’s disease. Neurobiol Aging 89:108–117. 10.1016/j.neurobiolaging.2020.01.005
Hong H, Hong L, Luo X et al (2024) The relationship between amyloid pathology, cerebral small vessel disease, glymphatic dysfunction, and cognition: a study based on Alzheimer’s disease continuum participants. Alzheimers Res Ther 16(1):43. 10.1186/s13195-024-01407-w
Jeong SH, Park CJ, Cha J et al (2024) Choroid Plexus Volume, Amyloid Burden, and Cognition in the Alzheimer’s Disease Continuum. Aging Dis. Published online January 23. 10.14336/AD.2024.0118
Saito Y, Kamagata K, Uchida W, Takabayashi K, Aoki S (2023) The partial volume effect of choroid plexus in pathogenesis of Alzheimer’s disease. Alzheimer’s Dement 19(10):4756–4757. 10.1002/alz.13123
Zhou L, Li G, Zhang Y et al (2021) Increased free water in the substantia nigra in idiopathic REM sleep behaviour disorder. Brain 144(5):1488–1497. 10.1093/brain/awab039
Butler T, Zhou L, Ozsahin I et al (2023) Glymphatic clearance estimated using diffusion tensor imaging along perivascular spaces is reduced after traumatic brain injury and correlates with plasma neurofilament light, a biomarker of injury severity. Brain Commun 5(3):fcad134. 10.1093/braincomms/fcad134
Kikuta J, Kamagata K, Takabayashi K et al (2022) An Investigation of Water Diffusivity Changes along the Perivascular Space in Elderly Subjects with Hypertension. Am J Neuroradiol 43(1):48–55. 10.3174/ajnr.A7334
Bjorkli C, Sandvig A, Sandvig I (2020) Bridging the Gap Between Fluid Biomarkers for Alzheimer’s Disease, Model Systems, and Patients. Front Aging Neurosci 12:272. 10.3389/fnagi.2020.00272
Ota M, Sato N, Nakaya M et al (2023) Relationship between the tau protein and choroid plexus volume in Alzheimer’s disease. NeuroReport 34(11):546–550. 10.1097/WNR.0000000000001923
Lopes DM, Wells JA, Ma D et al (2024) Glymphatic inhibition exacerbates tau propagation in an Alzheimer’s disease model. Alzheimer’s Res Therapy 16(1):71. 10.1186/s13195-024-01439-2
Szlufik S, Kopeć K, Szleszkowski S, Koziorowski D (2024) Glymphatic System Pathology and Neuroinflammation as Two Risk Factors of Neurodegeneration. Cells 13(3):286. 10.3390/cells13030286
Mogensen FLH, Delle C, Nedergaard M (2021) The Glymphatic System (En)during Inflammation. Int J Mol Sci 22(14):7491. 10.3390/ijms22147491
Sabbatini M, Barili P, Bronzetti E, Zaccheo D, Amenta F (1999) Age-related changes of glial fibrillary acidic protein immunoreactive astrocytes in the rat cerebellar cortex. Mech Ageing Dev 108(2):165–172. 10.1016/s0047-6374(99)00008-1
Wei YC, Hsu CCH, Huang WY et al (2023) Vascular risk factors and astrocytic marker for the glymphatic system activity. Radiol Med 128(9):1148–1161. 10.1007/s11547-023-01675-w
Katzman R, Zhang MY et al (1988) Ouang-Ya-Qu null,. A Chinese version of the Mini-Mental State Examination; impact of illiteracy in a Shanghai dementia survey. J Clin Epidemiol. ;41(10):971–978. 10.1016/0895-4356(88)90034-0
Zhao Q, Lv Y, Zhou Y, Hong Z, Guo Q (2012) Short-term delayed recall of auditory verbal learning test is equivalent to long-term delayed recall for identifying amnestic mild cognitive impairment. PLoS ONE 7(12):e51157. 10.1371/journal.pone.0051157
Mendez MF, Ala T, Underwood KL (1992) Development of scoring criteria for the clock drawing task in Alzheimer’s disease. J Am Geriatr Soc 40(11):1095–1099. 10.1111/j.1532-5415.1992.tb01796.x
Zhao Q, Guo Q, Hong Z (2013) Clustering and switching during a semantic verbal fluency test contribute to differential diagnosis of cognitive impairment. Neurosci Bull 29(1):75–82. 10.1007/s12264-013-1301-7
Cummings J (2019) The National Institute on Aging-Alzheimer’s Association Framework on Alzheimer’s disease: Application to clinical trials. Alzheimers Dement 15(1):172–178. 10.1016/j.jalz.2018.05.006
Zhang J, Wang J, Xu X et al (2023) In vivo synaptic density loss correlates with impaired functional and related structural connectivity in Alzheimer’s disease. J Cereb Blood Flow Metab 43(6):977–988. 10.1177/0271678X231153730
Li B, Jang I, Riphagen J et al (2021) Identifying individuals with Alzheimer’s disease-like brains based on structural imaging in the Human Connectome Project Aging cohort. Hum Brain Mapp 42(17):5535–5546. 10.1002/hbm.25626
Li B, Zhang M, Riphagen J et al (2020) Prediction of clinical and biomarker conformed Alzheimer’s disease and mild cognitive impairment from multi-feature brain structural MRI using age-correction from a large independent lifespan sample. Neuroimage Clin 28:102387. 10.1016/j.nicl.2020.102387
Duering M, Biessels GJ, Brodtmann A et al Neuroimaging standards for research into small vessel disease-advances since 2013. Lancet Neurol. Published online May 23, 2023:S1474-4422(23)00131-X. 10.1016/S1474-4422(23)00131-X
Daducci A, Canales-Rodríguez EJ, Zhang H, Dyrby TB, Alexander DC, Thiran JP (2015) Accelerated Microstructure Imaging via Convex Optimization (AMICO) from diffusion MRI data. NeuroImage 105:32–44. 10.1016/j.neuroimage.2014.10.026
Andersson JLR, Skare S, Ashburner J (2003) How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging. NeuroImage 20(2):870–888. 10.1016/S1053-8119(03)00336-7
Smith SM, Jenkinson M, Woolrich MW et al (2004) Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage 23(Suppl 1):S208–219. 10.1016/j.neuroimage.2004.07.051
Taoka T, Masutani Y, Kawai H et al (2017) Evaluation of glymphatic system activity with the diffusion MR technique: diffusion tensor image analysis along the perivascular space (DTI-ALPS) in Alzheimer’s disease cases. Jpn J Radiol 35(4):172–178. 10.1007/s11604-017-0617-z
Kamiya K, Hori M, Aoki S (2020) NODDI in clinical research. J Neurosci Methods 346:108908. 10.1016/j.jneumeth.2020.108908
Zhang H, Schneider T, Wheeler-Kingshott CA, Alexander DC (2012) NODDI: practical in vivo neurite orientation dispersion and density imaging of the human brain. NeuroImage 61(4):1000–1016. 10.1016/j.neuroimage.2012.03.072
Greve DN, Svarer C, Fisher PM et al (2014) Cortical surface-based analysis reduces bias and variance in kinetic modeling of brain PET data. NeuroImage 92:225–236. 10.1016/j.neuroimage.2013.12.021
Greve DN, Salat DH, Bowen SL et al (2016) Different partial volume correction methods lead to different conclusions: An (18)F-FDG-PET study of aging. NeuroImage 132:334–343. 10.1016/j.neuroimage.2016.02.042

There is NO Competing Interest.

Download PDF

Version 1

posted

You are reading this latest preprint version

Choroid Plexus Free-Water Correlates with Glymphatic function in Alzheimer Disease: The RJNB-D Study

Status:

Version 1

Abstract

Figures

Introduction

Results

Demographics and group comparisons

The relationship between CP FWf, glymphatic clearance and WMH

The impact of CP FWf across the AD continuum

The relationship between CP FWf, glymphatics and AD imaging markers

The relationship between CP FWf and blood-based neural biomarkers

The relationship between CP FWf, glymphatic clearance, WMH and cognition

Longitudinal study

Discussion

Methods

Participants and study design

Acquisition of MRI and PET images

CP and WMH segmentation

Diffusion MRI analysis

Cerebral blood flow in CP

Surface-based PET images for Aβ, Tau and SV2A

Statistical analyses

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1