Endothelial Cell Activation by Interleukin-1 and Extracellular Matrix Laminin-10 Occurs via the Yap Signalling Pathway

Background: The extracellular matrix (ECM) plays an important role for normal brain functions and homeostasis, and contributes to the inammatory response and mechanisms of brain repair after acute brain injury. We have previously reported that the ECM laminin-10 (LM-10) is a key regulator of blood-brain barrier (BBB) integrity, and is involved in BBB repair after hypoxic injury and interleukin-1 (IL-1)-induced inammation in vitro. To further investigate the role of LM-10 in BBB inammation and repair, we investigated for the rst time the signalling mechanisms regulated by LM-10 in brain endothelial cells in response to IL-1β-induced inammation in vitro. Methods: Human brain endothelial cell line hCMEC/D3 cultured on Matrigel- or LM-10-coated tissue culture plates were left untreated or were treated with human recombinant IL-1β at various concentrations and/or for various periods of time. In vitro hallmarks of angiogenesis were assessed using a scratch injury model and tube formation assay. Expression of cell adhesion molecules ICAM-1 and VCAM-1, as well as IL-8 was measured using ELISA. Activation of signalling pathways ERK1/2, p38, NF-κB and YAP was assessed by quantitative ELISA or Western blot. Activation of genes downstream of YAP signalling was assessed by quantitative polymerase chain reaction. Results: LM-10 promoted endothelial proliferation and subsequent repair of an endothelial monolayer after scratch injury, induced tube formation, and upregulated IL-1β-induced ICAM-1 and VCAM-1 expression in vitro. Classical IL-1β-induced signalling pathway ERK1/2 and p38 were not modulated by LM-10, whilst LM-10 upregulated IL-1β-induced NF-κB activation. Importantly, we demonstrate for the rst time a role of the YAP signalling pathway in endothelial cell activation, in that LM-10 signicantly downregulates p-YAP (S397) activation without affecting phosphorylation of YAP (S127), leading to differential expression of YAP target genes, ctgf and serpine-1 involved in endothelial cell activation. Conclusion: Our study provides for the rst time evidence that the YAP signalling pathway is an important regulator of endothelial cell activation, and could be a new therapeutic target for the treatment of cerebrovascular inammatory diseases. ImageJ. Phospho-NF-κB p65 and IκBα were normalised to β-actin and the ratio is presented. Data are represented as mean ± SEM of 4 biological replicates, and were analysed by a two-way RM ANOVA followed by Sidak’s post-hoc analysis comparing the means of expression of adhesion molecule/cytokine at corresponding concentrations of IL-1β treatment on Matrigel vs LM-10 and comparing the means of signalling molecules at corresponding time points on Matrigel vs LM-10, # represents signicance within ECM to untreated (A). Signicant main effects of IL-1β treatment and the ECM coating are indicated on the top left of the graphs (A). Data were assessed by a two-way RM ANOVA followed by Sidak’s post-hoc analysis comparing the means of signalling molecules at corresponding time points on Matrigel vs LM-10, # represents signicance within ECM to untreated (B-C). Signicant main effects of IL-1β treatment time and the ECM coating are indicated on the top left of the graphs (B-C). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.


Introduction
In physiological conditions, the extracellular matrix (ECM) of the central nervous system (CNS) provides a structural and functional environment for the cells of the neurovascular unit that is essential for maintenance of blood-brain barrier (BBB) integrity and brain homeostasis. However, after cerebral ischaemia, the BBB undergoes profound changes associated with the breakdown of tight junctions, remodelling of the ECM and enzymatic degradation of ECM proteins [1,2]. The early events of BBB breakdown are also associated with leukocyte in ltration, mediated by cell adhesion molecules and chemokines [3]. The cytokine interleukin(IL)-1 is an established mediator of the pro-in ammatory response associated with BBB dysfunction and subsequent tissue damage [4]. Although IL-1 is known to exert detrimental actions during the acute phase of stroke, increasing evidence suggests a biphasic action of IL-1 that exhibits neuroreparative properties during the sub-acute phase [5]. Furthermore, increasing evidence suggests that the ECM plays a dynamic role during the detrimental phase of stroke, whilst promoting repair during the later stages post-stroke. Interestingly, a novel function of the ECM as a regulator of IL-1-induced signalling in astrocytes and cerebral endothelial activation has been demonstrated in vitro [6,7]. ECM remodelling after CNS injury is associated with BBB repair, and IL-1 has been shown to mediate repair mechanisms, leading to the hypothesis that BBB repair driven by IL-1 could be regulated by ECM remodelling. Indeed, we have recently demonstrated LM-10 as a key ECM molecule involved in BBB repair after hypoxic injury and IL-1-induced in ammation in vitro [8]. However, the role of LM-10 as a regulator of in ammation and angiogenesis had not yet been determined.
The mechanisms underlying the dynamic crosstalk between the ECM and in ammation are poorly understood. Although it has been previously demonstrated that different components of the ECM alter IL-1 signalling pathways in astrocytes and endothelial cells in vitro [6,7], the crosstalk between signalling pathways downstream of the IL-1 receptor and other key signalling pathways has not been fully characterised. The YAP/Hippo pathway has recently gained signi cant interest as an extremely dynamic pathway implicated in ECM remodelling, as seen in cancer and in ammatory diseases [9,10]. It is well established that ECM-integrin signalling is a key step in the initiation of the YAP pathway. Critically, recent evidence has demonstrated a key role of YAP in tumour necrosis factor-α-induced endothelial activation [11]. However, the role of YAP in endothelial cells after IL-1 treatment had not yet been determined, and whether LM-10 could modulate this response is currently unknown. Using in vitro approaches, we demonstrate here that LM-10 modulates key in vitro hallmarks of angiogenesis and modulates endothelial cell activation after IL-1β treatment through the upregulation of key adhesion molecules, and demonstrate a novel mechanism whereby YAP signalling pathway is involved in endothelial activation by IL-1β, providing new signalling mechanisms of cerebrovascular activation regulated by IL-1 and the ECM.

Materials And Methods
Tissue culture plate coating Tissue culture plates (Corning, UK) were coated at 4°C overnight with human recombinant LM-10 (Biolamina, Sweden) diluted in phosphate-buffered saline (PBS) with calcium and magnesium (PBS+) at concentrations of 0.1, 1, 2.5, 5, 10 µg/ml. Control wells were incubated with lter sterilised 0.22 μm pore size lter (Starlab, UK) 0.1 % (w/v) low endotoxin bovine serum albumin (BSA) in PBS + . After overnight incubation, LM-10 solutions were discarded. PBS + was added to control and LM-10 coated wells to prevent drying out. Tissue culture plates were stored at 4 °C. For experiments using Matrigel, neat Matrigel (9.16 mg/ml, Corning) was diluted in Dulbecco's Modi ed Eagle's Medium (DMEM) (Invitrogen, UK) to a concentration of 20 µg/ml, and tissue culture plates were pre-coated at 4°C overnight. Plates were washed in PBSbefore cell seeding.
Concentrations in the samples were calculated by interpolating the values from the standard curve. Levels of all cytokines were corrected for cell number and expressed as pg/ml per µg/ml of total protein (pg/µg of protein), using total protein determined by the BCA assay.
Western blot analyses hCMEC/D3 cells were lysed in RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1% sodium deoxycholate, 0.1% SDS, and 2 mM EDTA) supplemented with 1% protease inhibitor cocktail and phosphatase inhibitors. Lysates were denaturated in Laemmli buffer (2% (w/v) SDS, 5% (v/v) β-mercaptoethanol, 10% (v/v) glycerol in 60mM Tris-HCl, pH 6.8) at 95°C for 5 min, and equal amount of protein were resolved on 10% SDS polyacrylamide gel, then transferred onto polyvinylidene di uoride (Bio-Rad) using a Trans-Blot Turbo Transfer System (Bio-Rad). Non-speci c binding sites were blocked with 5% (w/v) BSA in PBS 0.1% (v/v) Tween 20 (PBST) for 1 h at room temperature (RT). Membranes were then incubated (4 °C) overnight in primary antibody in PBST 5% BSA, as follows; anti IκBα (Cell Signalling UK, 1/1000); anti phospho-p65 (Ser536) (Cell Signalling UK, 1/1000); anti p-YAP127 (Cell Signalling UK, 1/1000); anti p-YAP397 (Cell Signalling UK, 1/1000); anti YAP (Cell Signalling UK, 1/1000); anti β-actin (Abcam UK, 1/1000). Membranes were washed in PBST and incubated with secondary anti rabbit anti-IgG (Agilent UK, 1/1000) antibody in PBST 1 % BSA for 1 h at RT. Membranes were washed and incubated in ECL Prime Western Blotting Detection Reagent (GE Life Sciences, UK) before exposure with G:BOX (Syngene) and Genesys software. Densitometry was determined using ImageJ software, and detected intensities were normalised against β-actin. Finally, ratios of pYAP/YAP were calculated. Scratch wound assay hCMEC/D3 cells were seeded at 40,000 cells/well in 96-well ImageLock plate (Essen BioScience) and left to adhere for 4 h. Scratch wound injury was carried out using a 96-pin IncuCyte WoundMaker Tool (Essen BioScience). Cells were then washed twice with PBS and replaced with fresh media. Phase contrast images were acquired at 2 h intervals for a period of 24 h on an Incucyte Zoom Live Cell Analysis system using a 4x/3.05 Plan Apo OFN25 objective. The 96-well Cell Migration Software Application Module (Essen BioScience) was used to quantify relative wound density. Relative wound density is a measure (%) of the density of the wound region relative to the density of the cell region: Tube formation assay 50 μl of undiluted Matrigel (9.16 mg/ml) was added to individual wells of an ice-cold 96-well plates, and the plate incubated at 37 °C for 45 min to allow gelation. 70 µl of either LM-10 (10 µg/ml) or PBS (control) was added to the Matrigel pre-coated wells, and plates were incubated for a further 2 h at 37°C.
The PBS and LM-10 aqueous layers on top of the Matrigel were aspirated. hCMEC/D3 cells were resuspended in media supplemented with 3 μg/ml bFGF and were seeded at a density of 10,000 cells/well. Phase contrast images were acquired at 0 and 6.5 h on an Incucyte Zoom Live Cell Analysis system using a 4x/3.05 Plan Apo OFN25 objective. The angiogenesis analyser macro in ImageJ was used to quantify total number of branches and total branching length.

Statistical analyses
All data were analysed with GraphPad Prism 8.1.2 (GraphPad Software Inc) using the statistical tests stated in the gure legends. Homoscedasticity of the standard deviations were evaluated with a Brown-Forsythe and Bartlett's test, alongside the use of homoscedasticity plots (predicted vs residual) and QQ plots to assess equal variance and normality. Appropriate transformations were applied where necessary.
We previously reported that LM-10 plays a key role in the maintenance of BBB integrity, and reverses most of the key hallmarks of BBB dysfunction induced by acute hypoxia and IL-1 [8]. In the current study we have used hCMEC/D3 cells, a cell line that has been extensively characterised as a valid brain endothelial phenotype and used as a model of human BBB function [12]. To con rm the cellular effects of LM-10 on hCMEC/D3 cells, we investigated the effect of LM-10 on angiogenic responses using a scratch assay (Fig. 1A) and tube formation assay (Fig. 1B). LM-10 reduced width of scratch in a time-dependent manner compared to Matrigel (Fig.1Ai), re ective of increased endothelial migration into the scratched region.  . 1Aii). In the tube formation assay, LM-10 increased the number of branches (1.35-fold increase, p=0.0263) and total branching length (1.57-fold increase, p=0.0144, Fig. 1Bi and ii) compared to Matrigel, showing that LM-10 increase tube-like structures and their length. These data suggest that LM-10 aids endothelial cell migration, an initial step in the processes of angiogenesis and BBB repair.
Having demonstrated that not all classical IL-1β downstream signalling pathways were modulated by LM-10, we investigated alternative signalling pathways that may be modulated by LM-10 in IL-1-induced activated endothelial cells. The ECM has been recently implicated as a regulator of the YAP/Hippo pathway [9], and evidence suggests that in ammation can modulate this signalling mechanism [11]. Therefore, we hypothesised that IL-1 and/or LM-10 alters YAP signalling in endothelial cells. To address this hypothesis, we initially measured the phosphorylation of YAP at S127 (Fig. 3Ai) and S397 (Fig. 3Aii) after IL-1β treatment. These phosphorylation sites play differential roles in YAP signalling whereby p-YAP(S127) induces cytoplasmic retention of YAP [13,14], and p-YAP(S397) creates a phospho-degron motif for proteasomal degradation [15]. The p-YAP(S127)/YAP and p-YAP(S397)/YAP ratios are an indicator of levels of dephosphorylated YAP in the cell that is able to translocate into the nucleus and activate associated genes.
When analysing the p-YAP(S127)/YAP ratio after IL-1β treatment in hCMEC/D3 cells, we found that IL-1β treatment signi cantly (p<0.0001) affected p-YAP(S127) levels in a time-dependent manner in cells on either Matrigel or LM-10 (Fig. 3Ai); In hCMEC/D3 cells seeded on Matrigel, phosphorylation of YAP(S127) signi cantly decreased at 5 min (54%, p=0.0.164) and 15 min (54%, p=0.0053) compared to baseline levels, to then slowly return to baseline levels at 120 min and 240 min. In cells seeded on LM-10, a signi cant decrease in p-YAP(S127) was observed at 15 min (46%, p=0.0038), which then returned to baseline levels at 120 min and 240 min. No signi cant effect of ECM on p-YAP(S127) levels was determined, and no signi cant differences between Matrigel and LM-10 were observed at each time point after IL-1β treatment. Analysing p-YAP(S397)/YAP ratio after IL-1β treatment found that IL-1β signi cantly (p=0.0002) affected p-YAP(S397) levels in a time-dependent manner both on Matrigel and LM-10 (Fig. 3Aii); When hCMEC/D3 cells were seeded on Matrigel, there was a small, no signi cant, initial reduction in p-YAP(397) levels at 5 min (35% p=0.7288) and 30 min (15%, p=0.9978) after IL-1β treatment compared to baseline. This was followed by a marked increase in p-YAP(397) levels between 30 min and 240 min after IL-1β treatment, and signi cant increases were observed at 120 min (2.7-fold increase, p=0.0063) and 240 min (4.1-fold increase, p<0.0001) compared to baseline. In contrast, the signalling pattern of p-YAP(397) expression in hCMEC/D3 seeded on LM-10 was markedly different to cells on Matrigel, in that, a consistent, signi cant reduction in p-YAP(397) levels in hCMEC/D3 cells on LM-10 was observed between 15 min and 120 min after IL-1β treatment compared to basal levels, with a signi cant reduction at 15 min (55%, p=0.0349), 30 min (72%, p=0.0089) and 60 min (69%, p=0.0073), followed a small increase in levels observed at 240 min (36%, p=0.4310). Interestingly, the ECM had a signi cant To investigate the nuclear activity of YAP as a transcriptional co-activator, we analysed the mRNA levels of its target genes Ctgf (Fig. 3Bi) and Serpine 1 (Fig. 3Bii)

Discussion
Recent research has demonstrated a novel role of the ECM as a regulator of IL-1-induced endothelial in ammation in vitro [7] and LM-10 has been identi ed as a potential mediator of BBB repair in vitro [8]. However, the speci c role of LM-10 as a regulator of angiogenesis and IL-1-driven endothelial in ammation in vitro remained unknown. Furthermore, the signalling crosstalk between IL-1 and YAP/Hippo signalling in endothelial cells has not been investigated. Here we present novel evidence of a dynamic cross-talk between IL-1β and YAP signalling in endothelial cells regulated by LM-10, providing a potential mechanism underlying cerebrovascular in ammatory responses regulated by ECM remodelling in CNS diseases.
Extensive evidence suggests that the ECM plays a dynamic role in CNS repair, providing a supportive framework and contributing to recovery [16]. Indeed, laminins have been shown to be protective against neuronal loss and contribute to neuronal survival [17]. Here we rst show that LM-10 increased endothelial proliferation and migration, as well as formation of tube-like structures of hCMEC/D3 in cultures, all hallmarks that have been previously demonstrated to be indicative of angiogenesis [18][19][20].
However, we further demonstrate a novel function of LM-10 whereby endothelial attachment to LM-10 increases IL-1β-induced ICAM-1 and VCAM-1 expression in hCMEC/D3 cells, further con rming that LM-10 may play an important in ammatory modulatory effect. Indeed, after cerebral ischaemia, ICAM-1 and VCAM-1, key vascular adhesion molecules, are upregulated in brain blood vessels, which aid leukocyte in ltration into the brain and contribute to subsequent tissue injury [4]. Our observation therefore suggest a dual nature of LM-10 during the acute detrimental and subacute reparative phase of injury. Interestingly, VEGF upregulates ICAM-1 brain microvascular endothelial cells [19], mediating their migration, whilst in vivo data show that ICAM-1 knockout mice display attenuated VEGF-mediated angiogenesis [21], suggesting that ICAM-1 plays a key role in pathological angiogenesis. There is similar evidence demonstrating con icting roles of VCAM-1 with the potential angiogenic role evidenced in several studies [22,23]. Interestingly, we saw a strongly potentiated response of VCAM-1 expression compared to ICAM-1 expression in hCMEC/D3 cells, therefore we speculate that they may be regulated by different regulatory mechanisms. In line with this, we did not observe a potentiation of IL-8 on hCMEC/D3 cells seeded on LM-10, which would suggest that it is not just an overall increased in ammatory pro le when cells are seeded on LM-10, but a speci c response dictated by the differential upregulation of selective adhesion molecules.
We next sought to determine the signalling mechanisms by which LM-10 may modulate the in ammatory activation of endothelial cells. Previous studies have shown a novel regulatory mechanism by which the ECM modulates ERK1/2 signalling after IL-1β treatment in rat astrocytes and rat brain endothelial cells in vitro [6,7]. However, IL-1β treatment failed to induce ERK1/2 and p38 MAPK pathways in hCMEC/D3 cells, and therefore no effect of LM-10 was observed. Lack of effect of IL-1β treatment on ERK1/2 in endothelial has already been shown previously [3], however our results are contradictory to those reported by Ni and colleagues [24] who detected an increase in phospho-p38α in response to IL-1β treatment. In contrast, IL-1β treatment transiently activated NF-kB, a response previously observed by us in rat endothelial cell culture [7], and IL-1β-induced NF-kB activation was further potentiated by LM-10, suggesting that NF-kB activation may regulate IL-1β-induced adhesion molecule expression in hCMEC/D3 cells. In contrast our study demonstrates for the rst time that regulation of IL-1-induced endothelial cell activation by LM-10 may be mediated by the YAP pathway, a recently described new signalling pathway shown to have critical roles in vascular systems, contributing to vessel homeostasis, vascular development, angiogenesis [25]. Speci cally, we demonstrate here a novel temporal pattern of YAP phosphorylation at S127 and S397 regulated by IL-1 and LM-10. The subcellular distribution of YAP is controlled by the reversible phosphorylation of S127, resulting in binding and cytoplasmic retention, hence the inability to bind to TEADs [13,14], whilst the phosphorylation of S397 creates a phosphodegron motif for β-TrCP binding resulting in proteasomal degradation, providing an irreversible longerterm mechanism of YAP inhibition [15]. After IL-1β treatment in hCMEC/D3 cells, a similar decrease in pYAP(S127) followed by an increase back to respective basal levels was observed in cells on Matrigel and LM-10, whereas the temporal pattern of pYAP(S397) phosphorylation after IL-1β treatment was signi cantly impaired in cells on LM-10 compared to Matrigel. These ndings suggest that IL-1β -induced phosphorylation of YAP at S127 and S397 is coupled when cells are seeded on Matrigel, whilst the phosphorylation of S127 and S397 is uncoupled in cells on LM-10, and that transient reductions in pYAP(127) and pYAP(397) on LM-10 are indicative of longer dephosphorylation of YAP, and hence translocation into the nucleus. In support of this, we show heavily decreased levels of Ctgf mRNA in hCMEC/D3 cells on matrigel 2 h after IL-1β treatment, re ective of decreased activation of genes downstream of YAP. In comparison, no signi cant change in Ctgf mRNA was observed 2 h after IL-1β treatment in hCMEC/D3 cells on LM-10, supporting our previous ndings that YAP was not sequestered and degraded to the same degree in cells on Matrigel. However, we did not detect any signi cant change in Serpine1 mRNA levels. Since Serpine1 is typically associated with mechano-signalling [26], it is likely that this gene is not regulated via YAP in hCMEC/D3 cells after IL-1β treatment.

Conclusion
In conclusion, we demonstrate here for the rst time a novel signalling mechanism whereby YAP signalling pathway is critically regulated by IL-1β and LM-10 in endothelial cells, providing new signalling mechanisms of cerebrovascular activation regulated by IL-1 and the ECM that could be targeted for the treatment of in ammatory CNS disease.