Aims

Vascular integrity is central to optimal brain function. Reduced cerebral blood flow, impaired cerebrovascular reactivity, and dysfunctional hemodynamic responses are associated with Alzheimer’s disease (AD). A deeper understanding of the angioarchitecture and morphological features of rodent brain vasculature remains essential to investigating AD pathogenesis. This study aims to use light sheet microscopy (LSM) to construct a high-resolution 3D vascular map of the entire murine brain and develop an image analysis workflow for brain-wide vascular morphometric analysis.

Methods

Mice were transcardially perfused with Alexa Fluor 594-conjugated wheat germ agglutin (WGA). To retain native vessel morphology, brain tissue was cleared using SHIELD passive clearing. LSM image analysis was performed using Imaris software.

Results

An optimised clearing and LSM workflow resulted in high spatial resolution and excellent WGA dye signal retention. Maximum intensity projection 3D renderings of volumetric LSM images allowed construction of whole brain vascular tree with labelled large arteries (~240um) as well as small capillaries (~5um). High resolution vascular details were visualised in cortical and subcortical regions such as hippocampus and hypothalamus. A machine learning-based filament tracing pipeline allowed segmentation and tracing of vessels followed by feature extraction including vessel diameter, length, nodes and regional density. Ongoing work using NL-F mice, a knock-in AD model expressing physiological levels of human APP containing the Swedish (NL) and Iberian (F) mutations in the APP gene, will identify AD associated vascular impairments.

Conclusions

We developed a whole-brain 3D imaging and analysis approach to investigate murine neurovascular structure and integrity. Future work using these vascular mapping tools will allow us to understand development of AD associated neurovascular impairments including cerebral amyloid angiopathy.

Aims

We have previously shown that matrix metalloproteinase 9 (MMP9) and tissue inhibitor of proteinases 3 (TIMP3) are specifically expressed in the vasculature of patients with cerebral amyloid angiopathy (CAA), and may be associated with the development of CAA-related intracerebral haemorrhage (ICH). We now extended our studies to investigate whether TIMP4 may have a similar role in CAA.

Methods

Using immunohistochemistry, we studied occipital lobe tissue of 42 controls and 58 CAA patients (40 CAA-nonICH and 18 CAA-ICH). Using ELISA , we measured TIMP4 levels in cerebrospinal fluid (CSF) and serum of 38 CAA patients and 37 controls. For a subset of CSF samples (11 CAA samples and 14 controls), the ratios of MMP2, MMP9, and MMP14 to TIMP4 were calculated. Differences between CAA patients and controls were assessed using (ordinal) linear regression analysis.

Results

In brain tissue, CAA cases had increased vascular TIMP4 expression when compared to controls (p<0.001). In addition, TIMP4 expression was higher in CAA-ICH compared to CAA-nonICH cases (p=0.014). CSF levels of TIMP4 were decreased in CAA patients compared to controls (p=0.031). In addition, CSF TIMP4 levels were lower in CAA patients that had experienced a symptomatic haemorrhage compared to CAA patients who did not (p<0.01). The CSF:serum ratio of TIMP was decreased in CAA patients compared to controls (p=0.003). In addition, the ratios of MMP2 (p=0.038), MMP9 (p=0.005), and MMP14 (p=0.025) to TIMP4 were increased in CAA patients compared to controls.

Conclusions

We demonstrated increased TIMP4 expression in vessel walls of CAA patients compared to controls. In contrast, TIMP4 CSF levels were decreased in CAA patients compared to controls, whereas MMP:TIMP4 ratios were increased, indicating a disbalance of MMP/TIMP4 levels in CAA pathophysiology.

Aims

Cerebral small vessel disease (SVD) is an age-related cerebrovascular condition linked to dementia, stroke and other neurological disorders. SVD manifestations include white matter lesions (WML), microbleeds and infracts, however their molecular mechanisms remain unclear. This study aims to investigate cerebrospinal fluid (CSF) proteomic changes underlying SVD pathogenesis.

Methods

We examined 1331 proteins using Olink proteomics from CSF samples from 959 subjects from the Swedish BioFINDER-2 study, where 319 had WML, 150 had microbleeds and 70 had infarcts. Differential protein abundance was assessed using generalized linear models, using SVD as binary outcomes, and adjusting for age, sex, and CSF dynamics. Functional and cell-type enrichment analyses were performed using Gene Ontology and expression-weighted cell type enrichment (EWCE). SpeakEasy clustering algorithm was employed to identify modules based on protein co-expression.

Results

399, 42 and 25 CSF proteins were dysregulated in WML, microbleeds and infarcts, respectively (pFDR<0.05, Figure 1). Proteins regulating the extracellular matrix, including MMP12 (pFDR<0.01), were upregulated in all SVD, while specific upregulations included CHIT1 in microbleeds and S100A4 in WML (pFDR<0.0001, Figure 2). Upregulated proteins across SVDs were enriched in smooth muscle cells and vascular leptomeningeal cells, while downregulated proteins in WML were enriched in neurons and oligodendrocyte precursor cells (Figure 3). All SVDs were associated with a nitrogen catabolic and nucleobase-containing metabolic processes-related module, while WML and microbleeds were specifically associated with an immune function-related module (Figure 4). These modules were enriched mostly in upregulated proteins.

Conclusions

We unveiled shared and distinct proteomic, cell-type and functional signatures in SVD, with a potential common mechanism involving extracellular matrix dysregulation and contribution of smooth muscle cells. These findings offer insights into the molecular mechanisms underlying SVD and may lead to development of novel diagnostic approaches.

Aims

Malfunction of the blood-brain-barrier (BBB) is a central feature of many neurovascular disorders including small-vessel-disease (SVD). Cerebral SVD accounts for most haemorrhagic strokes, 25% of ischemic strokes, and contributes to over 40% of dementias. Despite this profound impact on brain health, there are no proven treatments. SVD has strong genetic contributors and we recently identified FOXF2 as a major risk factor. But the mechanisms linking FOXF2 deficiency to dysfunction of endothelial cells (EC) are still elusive. Mouse and non-physiological in vitro SVD models display some disease features, but lack complex phenotypes, have limited translatability to humans, and are not well suited for drug discovery.

Methods

We therefore developed a fully iPSC-derived human 3D BBB model by microfluidic 3D tissue engineering using ECs, mural cells, and astrocytes. To validate our model, we compared our data from FOXF2 KO iPSCs to mice with EC- or pericyte-specific inactivation of Foxf2 by applying proteomics, transcriptomics and confocal and electron microscopy in both systems.

Results

Imaging analysis revealed expression of typical cell fate markers and formation of vessel-like tubes in our 3D model, and we also demonstrate controlled perfusion, including with human blood. FOXF2 KO ECs displayed impaired BBB function, including lower cell junction integrity, decreased TEER and elevated caveolae density. Proteomics revealed dysregulation of pathways involved in pericyte-endothelial crosstalk, suggesting an important role of FOXF2 in vascular signalling. Moreover, we could rescue phenotypes by delivery of FOXF2 mRNA using lipid-nanoparticles (LNPs).

Conclusions

Our novel iPSC-derived 3D BBB model recapitulates typical disease features of SVD, and phenocopies data from Foxf2 deficient mouse models, validating the relevance of the iPSC BBB model for disease research. Furthermore, LNP experiments illustrate applicability for developing and testing therapeutic interventions.

Aims

We explored the association of cardiovascular risk factors with CSF markers of amyloid deposition and neurodegeneration stratifying patients with Alzheimer’s Disease (AD) according to age-of-onset: early (< 65,EOAD), classic (65-75,COAD), late (>75,LOAD). Then, we evaluated the effects of vascular comorbidities on cognitive decline and their interplay with albumin quotient (Qalb), accounting for blood-brain-barrier (BBB) integrity, and neurodegeneration

Methods

We enrolled 387 patients with biomarker-confirmed Mild-Cognitive-Impairment due to AD. We computed 8 risk factors in a composite vascular score (VS) and transformed the output to a 0-to-1 percentage. In each group we regressed VS on p-tau/Aβ42 and t-tau, accounting for sex, APOE and Qalb. In a subset of 105 patients, we regressed sex, APOE, Qalb, VS and t-tau on ΔMMSE (negative MMSE-changes after 1 year). A bias-corrected bootstrapped mediation model was used to test the indirect association of VS with ΔMMSE and with t-tau using Qalb as mediator.

Results

VS was not associated with p-tau/Aβ42 in any group, but positively associated with t-tau in EOAD (p=0.016) and LOAD (p<.001), though not in COAD (p=0.580). ΔMMSE was negatively associated with Qalb (p=<.001) and t-tau (p=0.004), but not with VS (p=0.662). The mediation model confirmed the absence of global effects of VS on ΔMMSE, despite the indirect negative Qalb-mediated repercussions (ADE:−1.097,p=0.406; ADME:−0.678,p=0.027; total:−1.775,p=0.183), but a strong direct positive effect on t-tau, with Qalb as a partial negative mediator (ADE:293.98,p=0.001; ADME:−39.25,p=0.044; total:254.74,p=0.006).

Conclusions

Cardiovascular risk factors impact neurodegeneration differently depending on AD age-of-onset. Globally, the VS-induced increase of t-tau levels leads to cognitive worsening, but we found no evidence of a direct role of VS on the progression of cognitive decline.

Aims

Small vessel cerebrovascular disease may converge on downstream neurodegeneration in Alzheimer’s disease (AD) or contribute directly to tau hyperphosphorylation. We hypothesized that there is an amyloid pathway and a vascular pathway through tau to neurodegeneration in a middle-aged sample of community-dwelling adults.

Methods

We analyzed 460 participants (age=63±4, 65% women, 67/19/13% Hispanic/Non-Hispanic Black/Non-Hispanic White), including 98 participants from the Offspring Study of Racial and Ethnic Disparities in Alzheimer’s Disease and 361 participants from the Northern Manhattan Study of Metabolism and Mind who underwent amyloid PET (Florbetaben), tau PET (MK6240), cerebrovascular MRI (white matter hyperintensity (WMH) volume), and structural MRI (cortical thickness in AD signature regions). A serial mediation model estimated two parallel indirect pathways (age-amyloid-tau-neurodegeneration; age-WMH-tau-neurodegeneration). Multiple group models were also estimated across study, amyloid-positivity, race/ethnicity, and sex.

Results

89% were amyloid-negative-tau-negative, 6% were amyloid-positive-tau-negative, and 3% were amyloid-positive-tau-positive, while 2% were amyloid-negative-tau-positive. Older age was associated with greater amyloid, WMH, and neurodegeneration, but not tau. Greater amyloid was associated with greater early Braak tau. Greater occipital WMH was associated with greater occipital tau, even among amyloid-negatives. Greater WMH, but not amyloid or tau, was associated with lower cortical thickness in AD signature regions. Pathways were similar across study, race/ethnicity, and sex.

Conclusions

Small vessel disease, particularly ischemia, may promote tau hyperphosphorylation locally in the occipital lobe, whereas amyloid may promote tau hyperphosphorylation in early Braak regions. Posterior small vessel disease may be directly contributing to AD pathogenesis in addition to downstream neurodegeneration.

Aims

Cerebrovascular dysfunction is an essential and early contributor to Alzheimer’s Disease (AD) and dementia. Endothelial cell (EC) stress, including that induced by vascular amyloid in cerebral amyloid angiopathy (CAA), can result in cerebral blood flow impairment, blood brain barrier (BBB) damage, microhemorrhages and focal ischemia, inducing neurodegeneration. Recent evidence suggests that cardiovascular (CV)/cerebrovascular risk factors, such as hyperhomocysteinemia (Hhcy) contribute to increasing AD/CAA pathology and risk. However, the mechanisms through which Amyloid beta (Aβ) drives EC dysfunction and death remain poorly understood and it is not known whether CV risk factors such as Hhcy activate the same molecular pathways to induce EC dysfunction. Our work is investigating new mechanistic perspectives on the vascular contributions to cognitive impairment and dementia.

Methods

We analyzed mitochondria and death receptor-mediated endothelial cell stress and death pathways. We also identified one specific new target, carbonic anhydrase, whose inhibition mitigates cerebrovascular and neurovascular dysfunction in vitro and in vivo.

Results

Our mechanistic data demonstrates that 1) Aβ promotes mitochondrial metabolic and respiratory changes and activates extrinsic and intrinsic apoptotic pathways in human cerebral microvascular ECs. 2) Hhcy potentiates the Aβ-mediated activation of the death receptors 4/5 (DR4/5)-mediated apoptotic pathway, EC barrier permeability (by deregulating specific junction proteins expression, localization, and phosphorylation), as well as VEGF-A signaling and EC angiogenic capabilities. 3) Carbonic anhydrase inhibitors (both FDA -approved and new mitochondria- specific compounds) prevent cerebrovascular dysfunction and EC stress in vitro and in animal models of CAA and AD.

Conclusions

This work highlights new molecular targets that mediate cerebrovascular pathology in AD/CAA and comorbid cerebrovascular conditions, paving the way to potential preventive or therapeutic strategies.

Aims

In this study, we determined the impact of reactive astrocyte signaling on neurovascular function and glial cell changes in a diet based model of hyperhomocysteinemia (HHcy) and small cerebral vessel disease. Barrel cortex and hippocampal astrocytes were targeted with AAV vectors expressing the NFAT inhibitor, VIVIT. Two photon microscopy was used to assess functional hyperemia in awake mice, while scRNAseq was used to assess molecular phenotypes of microglia and oligodendrocytes.

Methods

Seven-to-eight week-old C57BL/6J mice received intracortical injections of AAV2/5-Gfa2-EGFP (control) or AAV vectors expressing the NFAT inhibitor, VIVIT (AAV2/5-Gfa2-VIVIT-EGFP). Mice were then fed with control chow (CT) or HHcy-inducing chow. Twelve-fifteen weeks later, some mice were implanted with glass cranial windows for two photon imaging. Dilation of penetrating arterioles in barrel cortex and red blood cell velocity in dowstream capillaries were assessed in awake mice during air-puff whisker stimulation . In a second cohort of mice, brains were extracted after 15-weeks of diet treatment, hippocampi were pooled and dissociated for transcriptional analyses using scRNA seq.

Results

2P Imaging: Relative to controls, mice treated with HHcy diet exhibited significant impairments in arteriole dilation and capillary red blood cell velocity in response to whisker stimulation. Both of these deficits were significantly alleviated in HHcy mice treated with AAV-Gfa2-VIVIT.

scRNAseq: The HHcy condition was overrepresented by subpopulations of microglia and oligodendrocytes with corresponding differentially-expressed genes linked to neuroinflammation and disease, and underrepresented by subpoplations linked to homeostatic functions. VIVIT treatment reversed HHcy-related gene expression patterns.

Conclusions

Results suggest that reactive astrocytes disrupt the function of local cerebrovessels and coordinate the molecular makeup of surrounding microglia and oligodendrocytes in the context of vascular inflammation and pathology.