Complex Regulation of Tau Phosphorylation by the Endothelin System in Brain Microvascular Endothelial Cells (BMVECs): Link to Barrier Function

Abstract

Endothelin-1 (ET-1), the most potent vasoconstrictor identified to date, contributes to cerebrovascular dysfunction. ET-1 levels in postmortem brain specimens from individuals diagnosed with Alzheimer’s Disease (AD) and related dementias (ADRD) were shown to be related to cerebral hypoxia and disease severity. ET-1-mediated vascular dysfunction and ensuing cognitive deficits have also been reported in experimental models of AD and ADRD. Moreover, studies also showed that ET-1 secreted from BMVECs can affect neurovascular unit integrity in an autocrine and paracrine manner. Vascular contributions to cognitive impairment and dementia (VCID) is a leading ADRD cause known to be free of neuronal tau pathology, a hallmark of AD. However, a recent study reported cytotoxic hyperphosphorylated tau (p-tau) accumulation, which fails to bind or stabilize microtubules in BMVECs in VCID. Thus, the study aimed to determine the impact of ET-1 on tau pathology, microtubule organization, and barrier function in BMVECs. Cells were stimulated with 1uM ET-1 for 24 hours in the presence/absence of ET_A (BQ123; 20uM) or ET_B (BQ788; 20uM) receptor antagonists. Cell lysates were assayed for an array of phosphorylation site-specific antibodies and microtubule organization/stabilization markers. ET-1 stimulation increased p-tau Thr231 but decreased p-tau Ser199, Ser262, Ser396, and Ser214 levels only in the presence of ET_A or ET_B antagonism. ET-1 also impaired barrier function in the presence of ET_A antagonism. These novel findings suggest that 1) dysregulation of endothelial tau phosphorylation may contribute to cerebral microvascular dysfunction and 2) the ET system may be an early intervention target to prevent hyperphosphorylated tau-mediated disruption of BMVEC barrier function.

INTRODUCTION

Vascular Contributions to Cognitive Impairment and Dementias (VCID) is a generalized definition of cognitive impairment arising from systemic and cerebral vascular diseases (1, 2), and it is a leading cause of Alzheimer’s Disease-Related Dementia (ADRD). While underlying etiologies are multifactorial, aging, microvascular dysfunction, and inflammation appear to be common factors. Indeed, endothelial dysfunction and disruption of the blood-brain barrier (BBB) occur before neuropathologies and cognitive deficits can be detected in all dementias (3, 4). Despite insights into the links of vascular factors to progressive neurodegeneration and memory loss, the underlying pathogenic mechanisms are unclear and there is no disease-modifying treatment to slow the progression (5).

One of the neuropathological hallmarks of AD-related dementia is NFTs (neurofibrillary tangles) made of hyperphosphorylated tau (5). In AD, cognitive decline and weakened behavioral skills accompany brain atrophy and aberrant protein accumulations such as paired helical filaments (PHFs) and NFTs (6). Tau, a microtubule-associated protein discovered in 1975 (7), is important in microtubule assembly and stabilization. Tau consists of multiple splice variants, over 80 potential phosphorylation sites, and domains that have been documented to interact with various proteins and structures. The natively unfolded tau shows little tendency for aggregation. Decades of research have led to the claim that the primary function of tau is to stabilize the axon’s microtubules (8). The dynamic nature of microtubule assembly/disassembly is regulated by phosphorylation/dephosphorylation at several sites of tau protein. However, hyperphosphorylation (p-tau) disrupts its normal function in microtubule dynamics, leading to axonopathy, including impaired axonal microtubule organization and transport, and also increases the propensity for tau oligomerization, aggregation, and NFT and PHF formation (9, 10). Since there are no NFTs in VCID in humans or animal models, VCID is not generally considered a tauopathy, and little is known about the role of tau in the progression from neurovascular insults to neurodegeneration (5, 11, 12). However, a recent study by Qui et al. (5) showed cis p-tau presence in the microvasculature of patients with VCID. Moreover, they reported a robust accumulation of p-tau in a hypoperfusion mouse model of VCID without the formation of tangles and immunotherapy with cis P-tau monoclonal antibody (cis mAb) rescued the cognitive deficits (5), suggesting that regulation of tau phosphorylation in brain endothelial cells may also contribute to the onset and progression of cerebral microvascular dysfunction in VCID. Additionally, it has been shown in the literature that different kinases/isomerases can mediate the phosphorylation of tau protein (10). Pin-1, being one of them, (5, 13, 14) is implicated in the conformational change of phosphorylated tau protein, especially at the proline-rich domain site. It has been well-characterized that Pin1 isomerase mediates the transition from cis p-tau to trans-p-tau, both of which are phosphorylated at Thr231 (14).

Regardless of the initiating event, experts agree that AD, VCID, and mixed dementias all incorporate aspects of cerebrovascular dysfunction, BBB breakdown, and neurovascular uncoupling in a vicious cycle (15). Although the underlying pathology of cerebrovascular dysfunction leading to cognitive impairment is multifactorial, the potential contributions of major vasoactive pathways, including the endothelin (ET) system, to the disruption of cerebrovascular health leading to neurodegenerative diseases like VCID and AD cannot be overlooked (15). The family of ETs, consisting of 3 related vasoactive peptides, ET-1, ET-2, and ET-3, plays important roles in cardiovascular (patho)physiology and embryonic development (16). ET-1 was the first isoform to be isolated as the most potent vasoconstrictor peptide mediating a wide spectrum of physiological functions via autocrine and paracrine mechanisms (15, 17). ET-1 mediates its diverse effects via two distinct G-protein–coupled receptor subtypes, endothelin receptor type A (ET_A) and ET_B. Studies showed that ET-1 is elevated in the brain specimens of AD and VCID patients (18–20) and correlates positively with tissue hypoperfusion (18, 21, 22). More recently, one study showed that Aβ causes the release of ET-1 and mediates potent constriction of pericytes via the activation of ET_A receptors in brain slices from humans with AD (23). Brain vasculature is unique because brain microvascular endothelial cells (BMVECs) express both receptors, whereas endothelial cells express only ET_B receptors in other vascular trees. The contractile effects of ET-1 have been widely studied and implicated as a major mechanism contributing to tissue hypoperfusion in ADRD. However, noncontractile effects and the relative roles and mechanisms by which ET receptors modulate cerebral microvascular function are poorly understood (24). An in vitro study with endothelial cells and pericytes isolated from bovine brain capillaries demonstrated that endogenous ET-1 causes actin cytoskeleton reorganization and morphological changes in the neurovascular unit via regulation of intracellular calcium levels in an autocrine and paracrine manner (15). Along with this data, recent evidence for p-tau accumulation in the cerebral microvasculature (5) led us to question the impact of the ET system on endothelial tau and microtubule stabilization. We hypothesized that ET-1 mediates oligomeric tau phosphorylation without tangle formation at multiple sites, causing microtubule destabilization and loss of barrier function in BMVECs.

METHODS

BMVEC culture

The male-derived human BMVEC line HBEC-5i (American Type Culture Collection-ATCC, CRL 3245) was cultured in 75 cm² culture flasks that were coated with 0.2% w/v gelatin (porcine Type A; Sigma-Aldrich) before cell seeding (25–29). EndoGRO-MV Complete Culture Media Kit (Millipore Sigma SCME001) was used for cell culture. The media was supplied with proper serum and antibiotics according to the manufacturer’s instructions. Cells were seeded with 80% confluency on 100 mm plates which were previously coated with 0.2% w/v gelatin and left in complete media overnight. On the stimulation day, the media was changed to Modification of Eagle’s Medium (DMEM; Corning, Manassas, VA, USA) containing no FBS serum but 1% penicillin-streptomycin. After a 4-hour starvation period, cells were stimulated with 1uM ET-1 for 24 hours in the presence/absence of ETA (BQ123; 20uM) or ETB (BQ788; 20uM) receptor antagonist in DMEM media containing 2% FBS. Cells and supernatants were collected for western blotting. All cell culture experiments were acquired from 3–6 individual experiments done in duplicates (n=3–6).

Western Blot Analysis

p-tau levels were measured by immunoblotting using antibodies selective for various p-tau species as well as total tau levels as depicted in Table 1. Briefly, equivalent amounts of cell lysates (15 μg protein/lane) were loaded onto NuPAGE^™ 8% Bis-Tris 15-well 1.5 mm Mini Protein Gel, and proteins were separated, and transferred to nitrocellulose membranes. The membranes were blocked with 5% non-fat milk for 1 hour followed by overnight incubation at 4°C with various phospho-tau antibodies (Table 1 and Table 2) to detect different phosphorylation sites (Fig 1) at 1:2000 dilution in 5% Bovine Serum Albumin (BSA) or anti-β-actin at 1:10.000 dilution in 5% BSA. After washing, membranes were incubated for 1 hour at room temperature with appropriate secondary antibodies (horseradish peroxidase [HRP]-conjugated; dilution 1:5000). Pre-stained molecular weight markers were run in parallel to identify the molecular weight of proteins of interest. For chemiluminescent detection, the membranes were treated with an enhanced chemiluminescent reagent and the signals were monitored on Amersham imager 680 (GE Healthcare Bio-Sciences Corp., Marlborough, MA). Relative band intensity was determined by densitometry on Quantl and normalized with β-actin protein.

Table 1-.

Commercially Available Antibodies for Various p-tau Species

Antibody Name (Clone)	Species	Clonality	Isotype	Supplier	Catalog No.	Epitope	Species Activity
TNT-1	Mouse	Monoclonal	IgG1κ	Millipore Sigma	MABN471	KLH-conjugated linear peptide corresponding to human Tau	Mouse, Rat, Human
Tau-5	Mouse	Monoclonal	IgG1κ	Abcam	ab80579	aa 210–230 of Tau according to Carmel G doi: 10.1074/jbc.271.51.32789. PMID: 8955115.	Mouse, Rat, Human
p-tauSer214	Rabbit	Polyclonal	IgG	ThermoFisher Scientific	44-742G	Synthetic peptide. The exact sequence is proprietary.	Mouse, Rat, Human
p-tauSer262	Rabbit	Polyclonal	IgG	ThermoFisher Scientific	44-750G	Synthetic peptide. The exact sequence is proprietary.	Mouse, Rat, Human
p-tauSer199	Rabbit	Polyclonal	IgG	Cell Signalling Technology	29957	Synthetic peptide. The exact sequence is proprietary	Mouse, Rat, Human
p-tauSer396	Mouse	Monoclonal	IgG2b	Cell Signalling Technology	9632	Partially purified Human PHF-tau	Mouse, Rat, Human
p-tauThr231	Mouse	Monoclonal	IgG1κ	ThermoFisher Scientific	MN1040	Partially purified Human PHF-tau	Mouse, Rat, Human

Table 2-.

Commercially Available Antibodies for Various p-tau Species and Their Established Cell Types, References, and Effects

Antibody Name (Clone)	Established Cell Types	References	Effect of Phosphorylation
TNT-1	Neurons, Brain Lysate	Fujita Y, doi: 10.1016/j.xpro.2020.100111. PMID: 33377007; PMCID: PMC7756915 Davtyan H,doi: 10.1016/j.vaccine.2017.03.020. Epub 2017 Mar 18. PMID: 28320590	Detection total tau independent of their PTMs
Tau-5	Neurons	Sadick JS, doi: 10.1038/srep33999. PMID: 27666089; PMCID: PMC5036045 Qi ZP, doi: 10.4103/1673-5374.177744. PMID: 27073389	Oligomeric tau
p-tauSer214	HEK293, SHSY5, Pulmonary Endothelial Cells	Wu R. doi: 10.3389/fnmol.2021.631833. PMID: 34054426; PMCID: PMC8155256 Farrell K, doi: 10.1007/s00401-021-02379-z. Epub 2021 Nov 1. PMID: 34719765	Detachment of tau from microtubules leading to endothelial cell barrier disruption
p-tauSer262	HEK293,N2a (Neuronal) Cell line	Liu P. doi: 10.1155/2018/2025914. PMID: 30057671; PMCID: PMC6051032 Sun ZD. doi: 10.4103/1673-5374.336872. PMID: 35259851	Reduced affinity for microtubules and detachment from microtubules
p-tauSer199	Neurons	Johnson GV, doi: 10.1242/jcs.01558. PMID: 15537830 Jin N. doi: 10.1074/jbc.M115.645507. Epub 2015 Apr 27. PMID: 25918155	Convert tau to an inhibitory molecule that sequesters normal microtubule-associated proteins from microtubules
p-tauSer396	Neurons	Di Primio C, doi: 10.1093/pnasnexus/pgad282. PMID: 37731949 Yao SY.doi: 10.1111/bph.16048. Epub 2023 Feb 28. PMID: 36727262	Significantly increased propensity to aggregate
p-tauThr231	HEK293, N2a	Levert S, doi: 10.1007/s12035-022-03121-w. Epub 2022 Nov 18. PMID: 36399251 Jadiya P. doi: 10.1016/j.isci.2023.106296. PMID: 36936788	• PHF-Tau and tangles • trans-to-cis isomerization, leading to a conformational change that reduces the affinity of the protein for microtubules

Fig.1. Phosphorylation sites of tau protein are scattered amongst the proline-rich domain, MT binding domain, and N-Terminus.

Tau protein has ~85 phosphorylation sites within the proline-rich and MT binding domains. The figure shows the phosphorylation sites that have been evaluated throughout this paper. These phosphorylation sites include Ser199, Ser214, and Thr231, all of which lie within the proline-rich domain and Ser 262 and Ser396, which lie within the MT-binding domain. The total tau levels were detected over the proline-rich domain as well as the N-terminus epitope.

Immunocytochemistry

BMVECs were seeded upon four-well plates at a density of 4×10⁴ cells. After they underwent the experimental conditions for 24 hours, they were fixed utilizing 4%PFA for 15 min, permeabilized with 0.1% TritonX-100 in PBS for 10 min and blocked with 5% goat serum in PBS for 1 hour. Cells were incubated overnight with an anti-rabbit β-Tubulin antibody (1:200, Cell Signaling Technology, 2128). Cells were washed and incubated with Alexa Fluor 488 conjugated secondary goat anti-rabbit antibody (1:500; Invitrogen) for 1 hour. Following antibody incubation, BMVECs were washed and mounted utilizing glass coverslips and Vectashield Vibrance mounting media with DAPI. BMVECs were then examined, and representative images were obtained via wide-field fluorescence microscope (Keyence) at 40x magnification.

Trans-Endothelial Electrical Resistance (TEER) Assay

The Agilent xCELLigence Real-Time Cell Analysis (RTCA) DP (dual purpose) instrument was used for TEER experiments. Pre-warmed DMEM media (supplemented with 1% FBS and 1% Pen/Strep) was added to each E-plate (Agilent, REF:300600890) well for background measurement. After the measurement, the wells were seeded with 40.000 cells and left overnight to reach confluency. Then, cells were stimulated with ET-1 along with/without ETA/ETB receptor antagonism for 24 hours. During this experimental procedure, electrical resistance was measured every 2 minutes.

Data Analysis

All group data were analyzed with one-way ANOVA followed by Tukey’s post hoc comparisons. Tukey’s Post hoc comparisons are marked on the graphs with mean ± sem of 3–6 individual experiments (*p<0.05, **p<0.01, ***p<0.0001). GraphPad Prism8 software was used for all analyses.

RESULTS

Evaluation of total tau protein in BMVECs

We used two antibodies for immunoblotting to measure the “total tau” levels independent of their post-translational modification (PTM). The Tau5 antibody recognizes and binds to the proline-rich domain of the tau protein (Fig 2A), and the PAD TNT-1 antibody recognizes and binds to the N-terminus of the tau protein (Fig 2A). There was a significant difference in tau5 levels between the treatment groups (*p<0.05, Fig 2B and C). Further post hoc analysis indicated that ET-1 increased tau5 compared to the control group (*p=0.01) and ET_B antagonism restored tau5 levels (*p=0.02). ET_A blockade did not prevent an ET-1-mediated increase in tau5. The groups had no significant difference in PAD TNT-1 levels (Fig 2D). ET-1 also increased the tau5/TNT-1 ratio (**p<0.01) and ET_B blockade prevented this effect (*p=0.03) (Fig 2E).

Fig. 2. ET-1 stimulation did not affect total tau levels determined by the PAD-TNT-1 antibody.

BMVECs were stimulated with 1uM ET-1 for 24 hours in the presence/absence of ETA (BQ123; 20uM) or ETB (BQ788; 20uM) receptor antagonist. Following the stimulation, the total tau levels were determined using two antibodies (tau5 and PAD TNT-1) that recognize two epitopes including N-terminus and proline-rich domain (A). Representative images of western blots for tau5 and PAD TNT-1 levels (B). Tau5 levels significantly increased with ET-1 stimulation compared to the control group (C and E). PAD-TNT-1 total tau levels did not change (D) with ET-1 stimulation with or without antagonists. Groups were compared using one-way ANOVA followed by Tukey’s post hoc comparisons. *p<0.05

Evaluation of tau phosphorylation

We first measured the change in the AT180 (p-tau^Thr231) levels. This phosphorylation site resides in the proline-rich domain of the tau protein (Fig 3A). One-way ANOVA of p-tau^Thr231 values normalized to actin levels indicated a significant main effect (*p<0.05, Fig 3C). This difference was due to increased p-tau levels in the presence of ET_A blockade with BQ123 compared to the control group (*p=0.03, Fig 3C). A similar pattern was observed when p-tauThr231 values were normalized to total tau to determine whether this increase was due to an increase in total tau levels (*p=0.04, Fig 3D). To further define the p-tau^Thr231 accumulation under ET1-stimulated conditions in BMVECs, we measured Pin-1 levels using the same conditions. One-way ANOVA indicated a significant change in Pin-1 levels (*p<0.05), and further post hoc analysis showed that Pin-1 significantly increased in the presence of ET_A blockade compared to the control group (*p=0.03, Fig. 3B and E).

Fig. 3. p-tau^thr231 and Pin-1 levels were significantly increased with ET-1 stimulation in the presence of an ET_A antagonist.

Following the stimulation, the p-tau^Thr231 levels were determined using the commercially available antibody that recognizes and binds to the epitope within the proline-rich domain (A). Representative images of western blots for p-tau^Thr231 and Pin-1 isomerase levels (B). p-tau^Thr231 levels were significantly increased with ET-1 stimulation in the presence of ET_A antagonist, while there was no change in the presence of ET_B blockade (B, C, and D). Pin1 isomerase levels were also significantly increased with ET-1 stimulation plus ET_A blockade (B, and E). Groups were compared using one-way ANOVA followed by Tukey’s post hoc comparisons, *p<0.05.

Next, we examined Ser214 and Ser199 p-tau levels, which reside within the proline-rich domain (Fig. 4A). One-way ANOVA analysis showed a significant difference among groups in both ser214 and ser199 phosphorylation (*p<0.05, **p<0.01, Fig. 4B–F). Interestingly, ET-1 stimulation decreased ser214 p-tau levels only in the presence of ET_A or ET_B antagonism (*p<0.05, **p<0.01, Fig. 4C and D). There was also a significant difference in Ser199 p-tau levels between the groups and this difference was due to lower Ser199 p-tau in the presence of ET_A or ET_B antagonism (*p<0.05, **p<0.01, Fig. 4E and F). Next, Ser396 and Ser262 phosphorylation were probed. Again, there was no change in p-tau levels when cells were stimulated with ET-1 alone. However, in the presence of ET_A or ET_B blockade, there was a significant decrease in Ser396 and Ser262 phosphorylation (*p<0.05, **p<0.01, Fig. 5 A–F).

Fig. 4. p-tau^Ser199 and p-tau^Ser214 levels were significantly decreased with ET-1 stimulation in the presence of both ET_A and ET_B antagonists.

Following the stimulation, the p-tau^Ser199 and p-tau^Ser214 levels were determined using the commercially available antibodies that recognize and bind to the epitopes within the Proline-rich domain (A). Representative images of western blots for p-tau^Ser199 and p-tau^Ser214 levels (B). The p-tau^Ser199 and p-tau^Ser214 levels were significantly decreased with ET-1 stimulation in the presence of ET_A and ET_B antagonists (B, E, and F). Groups were compared using one-way ANOVA followed by Tukey’s post hoc comparisons. *p<0.05, **p<0.01

Fig. 5. p-tau^Ser262 and p-tau^Ser396 levels were significantly decreased with ET-1 stimulation in the presence of both ET_A and ET_B antagonists.

Following the stimulation, the p-tau^Ser262 and p-tau^Ser396 levels were determined using the commercially available antibodies that recognize and bind to the epitopes within the MT binding domain (A). Representative images of western blots p-tau^Ser262 and p-tau^Ser396 levels (B). The p-tau^Ser262 and p-tau^Ser396 levels were significantly decreased with ET-1 stimulation in the presence of both ET_A and ET_B antagonists (B, C, D, E, and F). Groups were compared using one-way ANOVA followed by Tukey’s post hoc comparisons. *p<0.05, **p<0.01.

Next, we measured the levels of kinases mostly known to phosphorylate tau. We observed that ET-1 decreased DYRK1A only in the presence of ET_B blockade compared to the control group (*p=0.02, Fig. 6A and B), whereas the GSK3β was significantly lower with both ET_A and ET_B blockade (*p<0.05, **p<0.01, Fig. 6A and C).

Fig. 6. Kinase levels were significantly decreased with ET-1 stimulation in the presence of both ET_A and ET_B blockade.

(A) Representative western blot images of DYRK1A and GSK3B kinase levels. (B) DYRK1A levels were significantly decreased with ET-1 stimulation in the presence of ET_B blockade. (C) GSK3B was significantly decreased with ET-1 stimulation in the presence of both ET_A and ET_B blockade. Groups were compared using one-way ANOVA followed by Tukey’s post hoc comparisons. *p<0.05, **p<0.01.

Evaluation of barrier function

To investigate whether these definitive p-tau level changes are associated with any changes in the microtubule assembly, we measured α-tubulin levels using the same conditions. The tubulin levels were significantly improved with ET_B blockade compared to ET-1 stimulation alone (*p=0.02, Fig 7A and B). We also performed β-tubulin immunocytochemistry under the same conditions to further investigate the microtubule assembly and dynamics (Fig 7C). We observed that without intervention, the cells assume a more spherical shape and sprout out to the neighboring cells well. In the presence of antagonists, especially with BQ-123, the staining pattern changed into a fibrillar form, whereas ET-1 stimulation alone did not cause any significant change in cell morphology or microtubule assembly (Fig 7C). When we performed TEER measurements with the same conditions, there was a significant decrease in the resistance only with the ET_A antagonism compared to the control group, suggesting that the rigid fibrillar structure and elongated morphology cause interference with barrier function (Fig 7D and E, ***p<0.0001).

Fig. 7. Endothelial barrier function was impaired when cells were stimulated with ET-1 in the presence of ET_A receptor blockade.

(A). Representative western blot images showing alpha-tubulin levels. (B) The alpha-tubulin levels were higher in cells stimulated with ET-1 in the presence of ET_B blockade as compared to ET-1 stimulation alone. (C) Representative images showing β-tubulin staining in BMVECs under control, ET-1 stimulation alone, or ET-1 stimulation in the presence of BQ-123 or BQ-788. A highly fibrillar structure was observed only in cells stimulated with ET-1 in the presence of BQ-123. (D, and E) Representative tracing of Normalized Cell Index vs Time graph of TEER assay and bar graph representation of cumulative data at 24 h. Groups were compared with one-way ANOVA followed by Tukey’s post hoc comparisons. *p<0.05, **p<0.01, ***p<0.001

DISCUSSION

Dementia is an emerging global healthcare crisis (30–32). Only a small percentage of dementia cases have been confirmed as pure AD and a great majority of patients are believed to have mixed dementia with a vascular etiology (33). Potential cerebrovascular contributions to AD/RD are multifactorial (34), and among these, cerebral hypoperfusion and increased BBB permeability are emerging as major as well as early contributors to cognitive decline and degenerative processes leading to cognitive demise (20). The decrease in cerebral blood flow (CBF) is likely due to changes in the balance of the sensitive constrictor and dilator vasoactive substances, including the ET system (35). Indeed, a series of studies showed that brain ET-1 levels correlate with tissue perfusion status in postmortem brain specimens and disease severity in ADRD patients (18–20). It is also known that ET-1-mediated pericyte constriction via the activation of ET_A receptors lowers blood flow in brain slices from humans with AD (36). However, it is unknown whether the ET system contributes to the onset and progression of neurovascular dysfunction independent of its contractile actions and the relative roles of ET receptors. Given that a) early BBB disruption is another component of vascular dysfunction identified in ADRD biomarker studies, b) brain endothelial cells uniquely express ET_A receptors and c) microtubule-binding tau protein phosphorylation and accumulation is observed in the cerebral microvasculature of patients with vascular dementia as well as in a preclinical model of VCID, we asked the question whether ET-1 stimulation mediates tau phosphorylation via the activation of ET_A receptors and if this is associated with disruption of barrier function. Since BMVECs and pericytes are so closely positioned, to understand how the vasoconstrictive milieu in the neurovascular unit affects the microtubule dynamics, which is vital for BBB integrity, we took a reductionist approach and determined the tau/p-tau profile in response to ET-1 stimulation in BMVECs.

Our novel findings suggest that ET-1 stimulation promotes oligomeric tau formation without tangle detection. We first focused on assessing total tau levels independent of PTMs and used two different antibodies (tau5 and TNT-1) with distinct epitopes to detect total tau. We observed the tau5 signal around ~150 kDa, which has been interpreted as oligomeric tau in the literature (37). The tau5 antibody recognizes and binds to the epitopes within the proline-rich domain of tau, which has been studied the most in terms of hyperphosphorylation with the tauopathies (38, 39). Hence, it has been speculated that the binding of the antibody might be partially inhibited by not only the phosphorylation status, but also other post-translational modifications of the protein (40). Considering this, we concluded that the tau5 antibody doesn’t represent total tau levels but rather represents the oligomeric tau levels independent of posttranslational modifications. Our novel findings show a significant increase in tau5 epitope with ET-1 treatment compared to the control group, and ET_B antagonism showed a significant decrease in oligomeric tau levels. To detect total tau levels in a much more reliable way, we also used the PAD-TNT1 antibody as it recognizes the N-terminal of the protein, which has not been implicated in any PTM under tauopathies (39), and we didn’t observe any significant changes under these conditions. Hence, we used this parameter as our total tau levels for p-tau level normalization which indicated a similar pattern with tau5 epitope. Based on this data, we conclude that ET-1 increases oligomeric tau levels via the ET_B receptors.

Next, we assessed the Thr231 phosphorylation under the same conditions using the AT180 antibody. In the literature, it has been implicated that this antibody recognizes late tangle formation in the neurons (40–42) and is picked up mostly towards the last stages of neurodegenerative diseases like AD (43, 44). However, in our case, we have used this antibody to confirm the accumulation of oligomeric p-tau (Thr231) in brain endothelial cells, as shown before (37). Interestingly, we observed that p-tau^Thr231 levels were significantly increased with ET-1 stimulation only in the presence of ET_A blockade. One possible explanation is that when ET-1 is taken up by the cell via both ET_A and ET_B receptors, activation of downstream pathways is balanced. However, when the ET_A receptor is blocked and it is taken up by mostly the ET_B receptor, it causes a shift in the balance downstream of ET_A or ET_B receptor pathways, leading to oligomeric p-tau formation, which might be mediated by the isomerase Pin-1. It has been well-characterized that Pin1 isomerase mediates the transition from cis p-tau to trans p-tau, and independent of their isomeric form, both are phosphorylated at Thr231 (14). Our results showed that Pin-1 was significantly increased in the presence of ETA blockade compared to the control group. Although we didn’t pick up any further restoration in the levels of Pin1 with the ETB blockade, our results indicate that Pin1 might be responsible for converting cis p-tau to trans-p-tau with ETA blockade. Hence, under ET1-stimulated conditions in the presence of ETA receptor blockade, the significant increase in tau phosphorylation at Thr231 might be mediated via Pin1.

We also assessed both the proline-rich phosphorylation sites Ser199, Ser214, and MT binding domain phosphorylation sites Ser262 and Ser356, all of which have been implicated in inducing cytotoxicity as well as causing microtubule instability in neurons (45–47). Ser396 is mostly implicated in AD progression and NFT formation (9, 44, 48), and the Ser262 site is implicated in stabilizing the microtubules in neurons (8, 10, 38). The significance of these phosphorylation sites in endothelial cells is not known. To our surprise, we observed that independent of whether the phosphorylation site is within the proline-rich or MT binding domain, ET-1 stimulation decreased phosphorylation at these sites only in the presence of ET_A or ET_B antagonism. While we did not measure the activity, we also noted that protein levels of kinases such as GSK3B and DYRK1A which phosphorylate tau are decreased under these conditions. To assess whether these alterations in tau phosphorylation are associated with changes in endothelial function, we assessed barrier function by measuring TEER, which was decreased only in the presence of ET_A receptor antagonism. In this study, we did not test the direct effects of BQ-123 or BQ-788 and used them only in combination with ET-1. Thus, we cannot exclude the possibility of direct effects of the antagonists. However, in our previous studies, we did not observe any changes in barrier function when these antagonists were used independently of ET-1. This in vitro functional outcome along with increased oligomeric tau^Thr231 as well as tau5 suggested that the ET_A receptor blockade unmasks ET_B receptor effects on p-tau in BMVECs.

We acknowledge the limitations of our in vitro reductionist approach. Phosphorylation is a regulatory PTM, and it is very dynamic. So is microtubule organization. We investigated p-tau at a one-time point in cultured BMVECs. Additionally, we didn’t check the soluble/insoluble fractions separately in terms of p-tau. We are also cognizant that phosphorylation of tau protein is not always detrimental but dysregulation of the on/off nature of phosphorylation may lead to hyperphosphorylation that leads to NFT formation. However, this study is the first of its kind in terms of describing the p-tau profile in BMVECs which is historically considered a neuronal pathology. Our findings also show the highly complex nature of tau phosphorylation by ET signaling. Finally, this work is important because it provides proof of concept that in addition to its contractile effects, ET-1 may contribute to neurovascular dysfunction via modulating tau phosphorylation and microtubule dynamics.

CLINICAL PERSPECTIVES.

• Neuronal hyperphosphorylation of tau and neurofibrillary tangles are historically linked to AD as neuropathological hallmarks. However, a recent study identified pathologically toxic p-tau isoforms in brain endothelial cells in patients diagnosed with VCID and in an animal model of VCID.

• Brain ET-1 levels are increased in post-mortem brain specimens from patients diagnosed with AD and VCID.

• This study shows for the first time that brain microvascular endothelial cells harbor the tau protein that undergoes phosphorylation and that ET-1 contributes to this process.

DATA AVAILABILITY STATEMENT

The datasets for the paper are available upon request.