Factor VII activating protease (FSAP) regulates the expression of inflammatory genes in vascular smooth muscle and endothelial cells

(1)

Accepted Manuscript

Kristina Byskov, Thomas Boettger, Paul F. Ruehle, Nis Valentin Nielsen, Michael Etscheid, Sandip M. Kanse

PII: S0021-9150(17)31251-0

DOI: 10.1016/j.atherosclerosis.2017.08.029 Reference: ATH 15184

To appear in: Atherosclerosis Received Date: 10 April 2017 Revised Date: 22 August 2017 Accepted Date: 23 August 2017

Please cite this article as: Byskov K, Boettger T, Ruehle PF, Nielsen NV, Etscheid M, Kanse SM, Factor VII activating protease (FSAP) regulates the expression of inflammatory genes in vascular smooth muscle and endothelial cells, Atherosclerosis (2017), doi: 10.1016/j.atherosclerosis.2017.08.029.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

(2)

M AN US CR IP T

AC CE PT ED

Factor VII activating protease (FSAP) regulates the expression of inflammatory genes in vascular smooth muscle and endothelial cells

Kristina Byskov¹, Thomas Boettger², Paul F. Ruehle³, Nis Valentin Nielsen¹, Michael Etscheid⁴, Sandip M. Kanse^1,5

1Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway. ²Max Planck Institute, Bad Nauheim, Germany. ³Department of Radiation Oncology,

Universitätsklinikum Erlangen, Erlangen, Germany. ⁴Paul Ehrlich Institute, Langen, Germany. ⁵Oslo University Hospital, Oslo, Norway

Address for correspondence: Institute of Basic Medical Sciences, University of Oslo, 0317 Oslo, Norway.

E-mail: sandip.kanse@medisin.uio.no (S. M. Kanse)

Keywords: FSAP, HABP2, protease activated receptors, inflammation, gene expression, atherosclerosis, thrombosis.

Abbreviations

APC, activated protein C; EC, endothelial cells; EGF, epidermal growth factor; ERK, extracellular signal regulated kinase; FGF, fibroblast growth factor; FSAP, factor VII activating protease; MAPK, mitogen activated protein kinase; MI, Marburg I single nucleotide polymorphism; PAR, protease activated receptor; PDGF, platelet derived

(3)

M AN US CR IP T

AC CE PT ED

2

growth factor; qPCR, quantitative real time polymerase chain reaction; SNP, single nucleotide polymorphism; TFPI, tissue factor pathway inhibitor; VSMC, vascular smooth muscle cells.

(4)

M AN US CR IP T

AC CE PT ED

3 Abstract

Background and aims: The factor VII activating protease (FSAP)knockout mice have a bigger neointima after vascular injury and a larger infarct volume after stroke. The Marburg I (MI) single nucleotide polymorphism (SNP) in the FSAP-encoding gene is associated with an increased risk of stroke and carotid stenosis in humans. We hypothesize that the regulation of gene expression by FSAP in vascular cells accounts for its vasculo-regulatory properties.

Methods: Vascular smooth muscle cells (VSMC) and endothelial cells (EC) were stimulated with FSAP and a microarray-based expression analysis was performed.

Selected genes were further investigated by qPCR. Receptor- and pathway-inhibitors were used to elucidate the mechanisms involved.

Results: Pathways significantly activated by FSAP include those related to

inflammation, apoptosis and cell growth in VSMC and inflammation in EC. The key upregulated genes in VSMC were AREG, PTGS2 and IL6; and in EC these were SELE, VCAM1, and IL8. Secretion of IL6 in VSMC and IL8 in EC was also stimulated by FSAP. Recombinant wild type protease domain of FSAP, but not the MI-isoform, could recapitulate most of these effects. In VSMC, but not EC, gene expression by FSAP was impaired by PAR1 (protease-activated receptor1) receptor antagonists. In VSMC, FSAP-induced expression of AREG and IL6 was blocked by cAMP and MAPK pathway inhibitors indicating that multiple signalling pathways are likely to be involved.

Conclusions: The stimulation of inflammation- and proliferative/apoptosis-related genes in VSMC and EC provides a comprehensive basis for understanding the role of FSAP in vascular diseases.

(5)

M AN US CR IP T

AC CE PT ED

4 Introduction

Factor VII activating protease (FSAP), official gene name is hyaluronic acid binding protein 2 (HABP2), is a circulating plasma serine protease with

multifunctional properties. The cellular effects of FSAP range from the regulation of growth factor activity¹, endothelial permeability², nucleosome release³, histone inactivation⁴, cell survival⁵ and cancerous transformation of cells⁶. In addition to activating factor VII, FSAP has also been shown to activate pro-urokinase as well as inactivate tissue factor pathway inhibitor (TFPI)⁷ indicating a role in coagulation and fibrinolysis⁸. Genetic epidemiological studies have shown that the G534E single nucleotide polymorphism (SNP) in the FSAP gene, also called the Marburg I (MI) polymorphism (rs7080536), is found in approximately 5% of Caucasians and is associated with advanced carotid stenosis⁹ and stroke¹⁰. Interestingly, MI-FSAP has about a 5-fold lower proteolytic activity towards all substrates tested and thus

represents a loss-of-function polymorphism¹¹. In a previous study, we showed that the application of exogenous wild type (WT)-FSAP, but not MI-FSAP, around the dilated artery was able to inhibit neointima formation and SMC proliferation¹².

Conversely, in FSAP^-/- mice, neointima formation, leukocyte accumulation, and SMC proliferation were significantly enhanced¹³.

Although the human genetic data on MI-SNP and studies on FSAP^-/- mice indicate a role for FSAP in vascular proliferative diseases, the mechanism of action of FSAP is still obscure. Does FSAP activate specific receptors and signal transduction mechanisms in cells and are these involved in the pathophysiology of FSAP? In an earlier study it was suggested that protease activated receptors (PARs) were involved in regulating microvascular permeability by FSAP²; the endogenous FSAP expression was inhibited in pulmonary microvascular endothelial cells and this in turn

(6)

M AN US CR IP T

AC CE PT ED

5

was shown to alter endothelial permeability in the presence of hyaluronic acid in a PAR-dependent manner. We have shown that PAR1 is also a receptor for FSAP in astrocytes and neurons using a low MW antagonist of PAR1⁵. Proteolytic inhibition of platelet-derived growth factor (PDGF)-BB is another possible mechanism of action of FSAP on vascular smooth muscle cells (VSMC)¹². Similarly, modulation of growth factor activity of fibroblast growth factor (FGF2) or the activation of the bradykinin- kallikrein system¹ in EC has been proposed.

To further characterize the cellular effects of FSAP on vascular cells, we have performed Affymetrix microarray-based mRNA expression analysis to describe the pattern of gene expression in EC and VSMC. A gene expression profile that is related to inflammation, apoptosis and cell growth was activated in response to FSAP. This regulation of gene expression required the proteolytic activity of FSAP and MI-FSAP did not possess this activity. There were also associated changes in cytokine

secretion in response to stimulation with FSAP. In VSMC a PAR-1-dependent mechanism was evident but this was not the case in EC. These results provide a comprehensive insight into the cellular signal transduction pathway mediated by FSAP in vascular cells and will help to explain the role of FSAP in athero-thrombotic diseases.

(7)

M AN US CR IP T

AC CE PT ED

6 Materials and methods

Materials:

The isolation of FSAP from human plasma and the preparation of D-phenylalanyl- prolyl-arginyl chloromethylketone (PPACK)-FSAP has been described before¹⁴. Human plasma thrombin was from Sigma Aldrich (Oslo, Norway). PAR1 inhibitors Vorapaxar (SCH530348) and SCH79797 were purchased from Axon Medchem (Groningen, The Netherlands), PAR2 inhibitor (ENMD-1068) was from Enzo Life Sciences (Lörrach, Germany), inhibitor of cAMP, RP-8-Br-cAMPS from BIOLOG Life Science Institute (Bremen, Germany), p38 inhibitor SB203580, ERK1/2 inhibitor PD98059 as well as UO126, AKT inhibitor LY29004 and PDGFβR inhibitor AG1296 were all obtained from Merck-Millipore (Oslo, Norway). All inhibitors were prepared in DMSO to the highest possible concentration.

Preparation of serine protease domain (SPD) of FSAP:

Protease domain (amino acids 292-560) of wild type (WT) FSAP as well as MI-FSAP (G534E mutation) was cloned into into the pASK-IBA33plus vector (IBA-Lifesciences, Goettingen, Germany) including a C-terminal 6xHis tag. Expression was in

BL21(DE3) cells (Agilent, La Jolla, CA) and inclusion bodies were prepared by sonification and centrifugation. SPD was purified over Ni-Agarose column (Qiagen, Hilden, Germany). The purified protein was refolded over 48 h and then

characterized by SDS-PAGE and activity assays with chromogenic substrate S2288 (Haemochrome Diagnostica, Molndal, Sweden). Further details will be described in a separate manuscript.

(8)

M AN US CR IP T

AC CE PT ED

7 Cell culture:

Primary human carotid artery vascular smooth muscle cells (VSMC) and human umbilical vein endothelial cells (EC) were obtained from Promocell (Heidelberg, Germany) and cultured in the medium supplied by the manufacturer, Basal Smooth Muscle cell Medium-2 and Endothelial Cell Basal Medium-2, respectively. A total of 4 different isolates of VSMC and 3 different isolates of EC were used for these studies.

Microarray analysis:

VSMC and EC were serum-starved overnight and stimulated with FSAP (10 µg/ml) for 4 or 10 h in duplicate. RNA was isolated and Affymetrix GeneChip Human Gene 1.0 st arrays were used for transcriptome analysis according to the manufacturer’s instructions. Data were analyzed using the RMA algorithm and using DNAStar Arraystar software to calculate mean values for the groups, fold changes and Student’s t-test p-values with or without Bonferroni and Benjamini-Hochberg

correction for multiple testing. Microarray data has been submitted to ArrayExpress (http://www.ebi.ac.uk/arrayexpress) with the following reference number, E-MTAB- 5592.

RNA isolation and qPCR analysis:

Total RNA was extracted using total RNA Miniprep Kit from Sigma-Aldrich. Reverse transcription was performed using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Weiterstadt, Germany). For real-time PCR, SensiFast Hi-ROX SYBR Kit (Bioline GmbH, Luckenwalde, Germany) was used and analysis was performed on a ViiA7 Real-Time PCR System (Applied Biosystems). Amplification (cDNA denaturation 95 °C for 5 sec, primer hybridization/elongation 60 °C for 20 sec)

(9)

M AN US CR IP T

AC CE PT ED

8

plot was monitored over 40 cycles and continuous fluorescence measurement

indicated mRNA expression of analyzed genes. Following amplification, the accuracy of amplicons was confirmed by melting curve analysis and products were run on agarose gels to determine their size. Fluorescent threshold cycles (ct) were set and normalized against ct of the reference gene GAPDH and have been adjusted for individual primer efficiencies as described before¹⁵. Primer sequences are provided in Supplementary Table 1.

IL6 and IL8 ELISA:

Serum-starved VSMC or EC were incubated with FSAP at various concentrations and harvested at different time points. Cell supernatants were collected, centrifuged and kept at -40°C until further use. Supernatants were assayed for secreted IL6 and IL8 using human ELISA kits (ImmunoTools, Friesoythe, Germany) according to manufacturer’s instructions. Streptavidin-HRP (Dako, Glostrup, Denmark) and 1- step^TM Ultra TMB, ELISA (Thermo Scientific, Oslo, Norway) was used for detection.

Statistical analysis:

Statistical significance was analysed by the non-parametric Kurskal-Wallis followed by Dunn’s post-test (Graphpad Prism, SanDiego, CA). Except the microarray experiment, which was performed as a single independent experiment in duplicate, all other experiments were replicated in 3-6 independent experiments. Composite data from all independent replicates or data from a single experiment is shown as indicated in the figure legend. Data is shown as mean + SEM and significance is denoted by * in case p< 0.05.

(10)

M AN US CR IP T

AC CE PT ED

9 Results

FSAP stimulates inflammation, proliferation and apoptosis-related genes in VSMC and EC:

VSMC and EC were stimulated with human plasma-purified FSAP for 4 and 10 h. The mRNA expression analysis using whole genome microarray identified AREG, PTGS2 and IL6 as up-regulated genes in VSMC and in EC these were SELE, IL8 and VCAM1. Lists of the strongly regulated genes can be found in the

Supplementary Tables 2 and 3. These changes were statistically significant using the Student’s t test but not when the Bonferroni and Benjamini-Hochberg corrections for multiple testing were applied. Further, functional annotation analysis using

DAVIDv6.8¹⁶ showed that the up-regulated genes could be categorized into

inflammation, growth regulation, proliferation and intracellular signalling in VSMC and inflammation, chemotaxis and cytokine cell signalling in EC (Supplementary Tables 4 and 5). The terms that described the down-regulated genes in VSMC belonged to the group of aldo/keto reductases that metabolise a variety of endogenous and

exogenous substrates (Supplementary Table 6). Down regulated genes in EC belonged to groups involved in the regulation of cell cycle and cytoskeleton (Supplementary Table 7). Considering the previously known functions of FSAP in athero-thrombosis we centred our further investigations on the up-regulated genes.

These data were further consolidated using real time qPCR. Based on

preliminary studies to determine the optimal time point, dose response analysis was performed at 4 h. FSAP mediated a statistically significant increase in expression of AREG and PTGS2 but not IL6 in VSMC (Fig. 1A). In EC, VCAM1 was induced significantly by FSAP but this was not the case for SELE and IL8 even though the levels of the latter two genes were elevated (Fig. 1B). We next studied if FSAP was

(11)

M AN US CR IP T

AC CE PT ED

10

also able to increase protein levels of some of the induced genes. Using an ELISA we observed that secreted IL6 was increased both in a time (Fig. 1C) and dose dependent (Fig. 1D) manner in VSMC but these changes were not statistically significant. In EC, IL8 secretion was increased is a statistically significant manner by FSAP (Fig. 1E and F). Taken together, we identified significant changes in

expression of genes which, in certain cases, also matched alterations at the protein level.

FSAP-mediated gene expression requires proteolytic activity:

The cellular effects of hemostasis proteins can be dependent or independent of their protease activity. Aprotinin, a known inhibitor of FSAP, diminished the FSAP- induced gene expression of AREG, IL6 and PTGS2 in VSMC (Fig. 2A) as well as VCAM1, SELE and IL8 in EC (Fig. 2B). Aprotinin also reduced the levels of IL6 secretion from VSMC and IL8 secretion from EC in response to FSAP (Fig. 2C and D). Aprotinin did not inhibit the effect of thrombin on AREG expression in VSMC confirming that it did not have a general inhibitory effect on cells (Supplementary data Fig. SI). Likewise, pre-inhibited PPACK-FSAP, did not induce any expression of AREG in VSMC (Supplementary Fig. 1).

The recombinant serine protease domains of WT-FSAP and the MI-isoform were expressed in bacteria and their activity was determined by cleavage of chromogenic peptide substrate and their purity was determined by gel

electrophoresis (Supplementary Fig. 2). The WT protease domain could activate the expression of IL6 and PTGS2, analogous to full length plasma FSAP in VSMC but not of AREG. The MI-isoform, which only differs by one amino acid, did not influence expression of the investigated genes (Fig. 3A). In EC, the WT-isoform increased the

(12)

M AN US CR IP T

AC CE PT ED

11

expression of VCAM1 and IL8, but the 3–fold increase in SELE was not significant.

The MI-isoform had no effect on the expression of SELE and IL8, but it consistently induced VCAM1 in EC (Fig. 3B). In VSMC, IL6 secretion was stimulated by WT but not the MI-isoform and the same pattern was observed in EC for IL8 (Fig. 3C and D).

This demonstrates that only proteolytically active FSAP stimulates a specific gene expression pattern in VSMC and EC whereas this was, generally, not the case for MI- FSAP.

Role of PAR1 in mediating the effects of FSAP on VSMC and EC:

The FSAP-induced expression of AREG and IL6 in VSMC was reversed by the PAR1 inhibitor vorapaxar (Fig. 4A and B). Similar results were obtained with another PAR1 antagonist, SCH79797, but not with the PAR2 antagonist ENMD-1068 (Supplementary Fig. 3). Substantial mRNA levels of PAR1 and PAR3 but lower levels of PAR2 and PAR4 were expressed in VSMC (Supplementary data Fig. SIV). The effect of FSAP on AREG expression was compared to thrombin, which is a prototypic protease agonist known to stimulate gene expression in these cells via PAR1. In VSMC the effect of thrombin (1 U/ml) was comparable to that of FSAP in AREG expression (Fig. 4C and D). Vorapaxar also inhibited the effect of thrombin as would be expected (Fig. 4C and D). These results clearly identify PAR1, in contrast to PAR2, as an essential FSAP receptor in VSMC that mediates the effects of FSAP on gene expression.

EC expressed predominantly PAR1 at the mRNA level (Supplementary Fig. 4).

The effect of FSAP on EC was not inhibited by vorapaxar, even though the effect of thrombin was clearly blocked by this inhibitor (Fig. 5). In EC thrombin was about 40-

(13)

M AN US CR IP T

AC CE PT ED

12

fold more active than FSAP in IL-8 expression thus FSAP was much weaker agonist on EC indicating a different mechanism of action compared to VSMC.

Activation of gene expression by FSAP through the MAPK and cAMP pathway in VSMC:

In order to get some orientation regarding which pathways may be involved in the regulation of gene expression by FSAP in VSMC we screened various pathway inhibitors. FSAP-induced changes in the expression of AREG and IL6 were used as read-outs. LY294002 (Akt inhibitor), AG1296 (PDGFR inhibitor), SB203580 (P38 MAPK inhibitor) had no effect (data not shown). But, RP-8-Br-cAMP (cAMP inhibitor), PD98059, UO126 (both ERK1/2 inhibitors, of which UO126 is not shown) were all inhibitory (Fig. 4). The use of these generic pathway inhibitors indicates that complex signal transduction pathways are responsible for changes in gene expression,

something that needs to be studied in more detail at a mechanistic level. Because of the smaller changes in gene expression by FSAP in EC, and stronger effects of these pathway inhibitors on basal expression, no specific conclusions could be reached about the pathways involved in EC.

Discussion

Pathway analysis of the regulated genes showed that overwhelmingly pro- inflammatory genes (SELE, VCAM1, CXCL1, ICAM1 and IL8) were induced in EC whereas genes related to proliferation (AREG, EREG, HBEGF), apoptosis (BDNF, EDNRB, DUSP1, MEF2C) and inflammation (IL6, CCL2, LIF and PTGS2) were induced in VSMC. These cellular effects of FSAP may be related to the observations in human genetic studies that link an SNP in the FSAP-encoding gene to stroke¹⁷

(14)

M AN US CR IP T

AC CE PT ED

13

and carotid stenosis⁹ as well as investigations in knockout mice indicating a role for endogenous FSAP in stroke⁵ and neointima formation¹³. Thus, these results provide a basis for a further mechanistic understanding of the role of FSAP in vascular biology.

The differences in gene expression were significant when using a Student’s t test; however, when corrections were made for multiple testing, e.g., Benjamini- Hochberg or Bonferroni correction there was a loss of significance due to over- correction. Thus, the microarray screen was used to identify potentially important changes in gene expression and to identify common regulatory mechanisms for these genes. These results were then further consolidated using qPCR and ELISA. In addition, some genes were detected by multiple probe-sets on the arrays and there was very high reproducibility between these internal replicates confirming the technical reliability of these results.

The specificity of these effects of FSAP on gene expression was established in different ways; the effects of FSAP were inhibited by aprotinin and could not be replicated with the pre-inhibited PPACK-FSAP. This was further verified as

recombinant WT FSAP could reproduce these effects, but not the MI-FSAP isoform which had a single amino acid mutation only. Some of the effects of the recombinant proteins did not show the expected pattern for reasons that are not immediately clear at the moment. Thus, a proteolysis-based cellular signalling mechanism of FSAP seems likely, which is unlike some factors e.g., urokinase¹⁸ that can activate cells in a non-proteolytic manner.

We have previously demonstrated that the lack of endogenous FSAP, in FSAP^-/- mice, seems to promote an inflammatory state by recruiting leukocytes into areas of tissue injury and remodelling¹³. Thus, our in vitro results are very surprising

(15)

M AN US CR IP T

AC CE PT ED

14

and counter-intuitive to what has been shown in knock-out mice. FSAP seems to promote, both, pro-inflammatory effects in vitro and anti-inflammatory effects in vivo depending on what system is being investigated. This discrepancy can be reconciled if FSAP should have multiple effects on different cell types; some of which are pro- and some anti-inflammatory. This duality is also true for the receptor PAR1, which is responsible for pro- and anti-inflammatory effects at the same time. Thrombin

stimulates apoptosis in EC via PAR1, but another protease APC is a potent inhibitor of apoptosis through cleavage of PAR1 at a different site¹⁹. Correspondingly,

stimulation of PAR1 can lead to vasodilation or vasoconstriction in the vasculature depending on whether the target cells are EC or VSMC²⁰. This duality is also

observed for some of the genes induced by FSAP, e.g. IL6²¹ and PTGS2²². Although, PTGS2 is considered to be mainly a pro-inflammatory gene, its inhibition leads to more atherosclerosis²². Similarly, blocking IL6 also revealed a role in inhibiting metabolic inflammation²³. Thus, the induction of IL6 and PTGS2 in VSMC in vitro may in fact account for anti-inflammatory effects of FSAP that are evident in vivo.

Thus, the pro- and anti-inflammatory effects of FSAP and the differences in vitro and in vivo need to be evaluated in a context-dependent manner.

In EC, pro-inflammatory genes such as SELE, VCAM1 and IL8 were induced by FSAP, but this effect could not be blocked by PAR1 antagonist. Thrombin’s effect on expression of the same genes was stronger and could be completely inhibited by the PAR1 antagonist indicating the presence of functional PAR1 on these cells and the validity of using the PAR1 antagonist. Further preliminary studies on EC showed that thrombin induced a rapid transient increase in intracellular Ca²⁺ which was completely absent in FSAP-stimulated cells (data not shown). A previous study has shown that PAR1 and PAR3 are involved in mediating the effects of endogenous

(16)

M AN US CR IP T

AC CE PT ED

15

FSAP on hyaluronic acid-dependent permeability in capillary lung endothelial cells². We did not detect any FSAP encoding transcripts in the EC used in the current study (data not shown), so the different conclusions may be related to the use of different cells and read-outs. Although, PAR3 is expressed by VSMC at the mRNA level, we did not investigate their role in further detail because the lack of established

inhibitors²⁴. In an earlier report the activation of bradykinin- and FGF2-dependent signalling pathways by FSAP has been reported in EC¹ and these, rather than PAR1, are more likely to be responsible for the effects of FSAP on gene expression in EC.

In VSMC, mRNA levels of PAR1 were much higher than that of PAR2 and also only the PAR1 inhibitor, but not PAR2 inhibitor, could block the effects of FSAP on gene expression in VSMC. The effects of FSAP and thrombin on gene expression in VSMC were quite comparable but thrombin was far more potent on EC compared to FSAP. The greater effectiveness of FSAP in activating PAR1 in VSMC may be related to the repertoire of co-receptors or receptor dimerization that is specific to VSMC and not found in EC. Thus, VSMC is another cell type, in addition to astrocytes and neurons⁵, where we have observed a PAR1-dependent effect of FSAP. Further investigations are needed to determine where and how PAR1 is cleaved by FSAP and which co-receptors are required for receptor activation.

In the inhibitor screen with various generic pathway inhibitors, a strong inhibitory effect of cAMP antagonist on FSAP-mediated gene expression in VSMC was identified. This seems to be in line with the fact that the expression of IL6 in VSMC was shown to be guided through the cAMP responsive element (CRE)^{25, 26}, and CREB phosphorylation in response to FSAP has been demonstrated before in EC¹. The inhibitory effect of PD98059 is in line with our earlier observations that there is a weak activation of ERK1/2 phosphorylation by FSAP in VSMC¹⁴. Although the

(17)

M AN US CR IP T

AC CE PT ED

16

specificity of the used inhibitors is well characterized further mechanistic studies are needed to understand how FSAP-PAR1 axis can stimulate these multiple pathways.

The hemostasis factors that predominantly activate PARs include thrombin, FXa, FVIIa, APC, kallikrein as well as plasmin. Proteases have differing selectivity towards the various PARs²⁷ and, furthermore, cleavage can take place at many sites within the N-terminal region leading to biased signaling or inactivation of the receptor.

Presence of co-receptors as well as complex formation within the PARs themselves²⁸ mean that every protease agonist has a unique activation profile¹⁹. FSAP can

activate PAR1 on VSMC but not on EC, probably, because an additional cofactor is missing on the latter cell type. Additional information about the exact nature of the cleavage sites is needed to further characterize the effect s of FSAP. The detailed mRNA expression profile in FSAP-stimulated VSMC and EC provides a valuable database to better understand the role of FSAP in vascular pathophysiology.

(18)

M AN US CR IP T

AC CE PT ED

17 Conflict of interest

The authors declared they do not have anything to disclose regarding conflict of interest with respect to this manuscript.

Financial support

This work was supported by grants from Research Council of Norway as well as Helse Sør-Øst to SMK.

Author contributions

KB performed most of the experiments and analysed the data and co-wrote the manuscript with SMK. PR performed some of the experiments with qPCR, analysed the data and edited the manuscript. NVN expressed and characterized recombinant FSAP and edited the manuscript. TB and ME performed and analysed the microarray data and edited the manuscript. SMK designed the study, analysed the data and co- wrote the manuscript together with KB.

Acknowledgements

We would like to thank Kirsten Grundt and Thomas Schmidt Woell for their technical assistance.

(19)

M AN US CR IP T

AC CE PT ED

18 References

1. Kress JA, Seitz R, Dodt J, Etscheid M. Induction of intracellular signalling in human endothelial cells by the hyaluronan-binding protease involves two distinct pathways. Biol Chem. 2006;387:1275-1283

2. Mambetsariev N, Mirzapoiazova T, Mambetsariev B, Sammani S, Lennon FE, Garcia JG, et al. Hyaluronic acid binding protein 2 is a novel regulator of vascular integrity. Arterioscler Thromb Vasc Biol. 2010;30:483-490

3. Zeerleder S, Zwart B, te Velthuis H, Stephan F, Manoe R, Rensink I, et al.

Nucleosome-releasing factor: A new role for factor vii-activating protease (fsap). Faseb J. 2008;22:4077-4084

4. Yamamichi S, Fujiwara Y, Kikuchi T, Nishitani M, Matsushita Y, Hasumi K.

Extracellular histone induces plasma hyaluronan-binding protein (factor vii activating protease) activation in vivo. Biochem Biophys Res Commun.

2011;409:483-488

5. Joshi AU, Orset C, Engelhardt B, Baumgart-Vogt E, Gerriets T, Vivien D, et al.

Deficiency of factor vii activating protease alters the outcome of ischemic stroke in mice. Eur J Neurosci. 2015;41:965-975

6. Gara SK, Jia L, Merino MJ, Agarwal SK, Zhang L, Cam M, et al. Germline habp2 mutation causing familial nonmedullary thyroid cancer. N Engl J Med.

2015;373:448-455

7. Kanse SM, Declerck PJ, Ruf W, Broze G, Etscheid M. Factor vii-activating protease promotes the proteolysis and inhibition of tissue factor pathway inhibitor. Arterioscler Thromb Vasc Biol. 2012;32:427-433

(20)

M AN US CR IP T

AC CE PT ED

19

8. Römisch J, Stöhr H-A, Feussner A, Kleinhenn M, Nakazawa F, Kannemeier C, et al. The contribution of factor vii-activating protease (fsap) to coagulation and fibrinolysis. Annals of Hematology. 2002;Supp. 1, Volume 81:A 69

9. Willeit J, Kiechl S, Weimer T, Mair A, Santer P, Wiedermann CJ, et al.

Marburg i polymorphism of factor vii--activating protease: A prominent risk predictor of carotid stenosis. Circulation. 2003;107:667-670

10. Trompet S, Pons D, Kanse SM, de Craen AJ, Ikram MA, Verschuren JJ, et al.

Factor vii activating protease polymorphism (g534e) is associated with increased risk for stroke and mortality. Stroke Res Treat. 2011;2011:424759 11. Etscheid M, Muhl L, Pons D, Jukema JW, Koenig H, Kanse SM. The marburg i

polymorphism of factor vii activating protease is associated with low proteolytic and low pro-coagulant activity. Thromb Res. 2012;130:935-941

12. Sedding D, Daniel JM, Muhl L, Hersemeyer K, Brunsch H, Kemkes-Matthes B, et al. The g534e polymorphism of the gene encoding the factor vii-activating protease is associated with cardiovascular risk due to increased neointima formation. J Exp Med. 2006;203:2801-2807

13. Daniel JM, Reichel CA, Schmidt-Woell T, Dutzmann J, Zuchtriegel G, Krombach F, et al. Factor vii-activating protease deficiency promotes neointima formation by enhancing leukocyte accumulation. J Thromb Haemost. 2016;14:2058-2067

14. Kannemeier C, Al-Fakhri N, Preissner KT, Kanse SM. Factor vii activating protease (fsap) inhibits growth factor-mediated cell proliferation and migration of vascular smooth muscle cells. Faseb J. 2004;18:728-730

15. Hellemans J, Mortier G, De Paepe A, Speleman F, Vandesompele J. Qbase relative quantification framework and software for management and

(21)

M AN US CR IP T

AC CE PT ED

20

automated analysis of real-time quantitative pcr data. Genome biology.

2007;8:R19

16. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using david bioinformatics resources. Nature protocols.

2009;4:44-57

17. Cheng YC, Stanne TM, Giese AK, Ho WK, Traylor M, Amouyel P, et al.

Genome-wide association analysis of young-onset stroke identifies a locus on chromosome 10q25 near habp2. Stroke. 2016;47:307-316

18. Ferraris GM, Sidenius N. Urokinase plasminogen activator receptor: A functional integrator of extracellular proteolysis, cell adhesion, and signal transduction. Seminars in thrombosis and hemostasis. 2013;39:347-355

19. Griffin JH, Mosnier LO, Fernandez JA, Zlokovic BV. 2016 scientific sessions sol sherry distinguished lecturer in thrombosis: Thrombotic stroke:

Neuroprotective therapy by recombinant-activated protein c. Arterioscler Thromb Vasc Biol. 2016;36:2143-2151

20. Ku DD, Dai J. Expression of thrombin receptors in human atherosclerotic coronary arteries leads to an exaggerated vasoconstrictory response in vitro.

Journal of cardiovascular pharmacology. 1997;30:649-657

21. Hunter CA, Jones SA. Il-6 as a keystone cytokine in health and disease. Nat Immunol. 2015;16:448-457

22. Tang SY, Monslow J, Todd L, Lawson J, Pure E, FitzGerald GA.

Cyclooxygenase-2 in endothelial and vascular smooth muscle cells restrains atherogenesis in hyperlipidemic mice. Circulation. 2014;129:1761-1769

23. Mauer J, Chaurasia B, Goldau J, Vogt MC, Ruud J, Nguyen KD, et al.

Signaling by il-6 promotes alternative activation of macrophages to limit

(22)

M AN US CR IP T

AC CE PT ED

21

endotoxemia and obesity-associated resistance to insulin. Nat Immunol.

2014;15:423-430

24. Hamilton JR, Trejo J. Challenges and opportunities in protease-activated receptor drug development. Annual review of pharmacology and toxicology.

2017;57:349-373

25. Tokunou T, Ichiki T, Takeda K, Funakoshi Y, Iino N, Shimokawa H, et al.

Thrombin induces interleukin-6 expression through the camp response element in vascular smooth muscle cells. Arteriosclerosis, thrombosis, and vascular biology. 2001;21:1759-1763

26. Funakoshi Y, Ichiki T, Ito K, Takeshita A. Induction of interleukin-6 expression by angiotensin ii in rat vascular smooth muscle cells. Hypertension.

1999;34:118-125

27. Zhang C, Srinivasan Y, Arlow DH, Fung JJ, Palmer D, Zheng Y, et al. High- resolution crystal structure of human protease-activated receptor 1. Nature.

2012;492:387-392

28. Nieman MT. Protease-activated receptors in hemostasis. Blood.

2016;128:169-177

(23)

M AN US CR IP T

AC CE PT ED

22 Legends to Figures

Figure 1. FSAP increases mRNA expression and cytokine secretion in VSMC and EC in a time and dose dependent manner.

(A) VSMC and (B) EC were stimulated in a dose dependent manner with FSAP (0- 10 µg/ml) for 4 h. mRNA expression of AREG, IL6 and PTGS2 relative to GAPDH in VSMC and VCAM1, SELE and IL8 relative to GAPDH in EC was determined by qPCR. VSMC were stimulated with FSAP, or the control buffer (HE) in a time course (C) or in a concentration-dependent manner (D) as described above and IL6 secretion in cell supernatants was determined. EC were stimulated with FSAP in a time course (E) or in a concentration-dependent manner (F) as described above and IL8 secretion in cell supernatants was determined. Results represent mean of 3-5 replicates + SEM. and *indicates p<0.05 using Kurskal-Wallis test.

Figure 2. Cellular activation by FSAP requires protease activity.

VSMC (A and C) and EC (B and, D) were stimulated with FSAP (10 µg/ml) or HE buffer for 4 h in the absence or presence of aprotinin (25 µg/ml). In (A) qPCR analysis of AREG, IL6 and PTGS2 mRNA levels relative to GAPDH was performed.

In (B) qPCR analysis of VCAM1, SELE and IL8 relative to GAPDH was performed.

(C) Secreted IL-6 was determined in VSMC using ELISA. (D) Secreted IL-8 was determined in EC using ELISA. Each panel represents results from 3-6 independent experiments; mean ±SEM, *indicate p<0.05 using Kurskal-Wallis followed by Dunn’s post test.

(24)

M AN US CR IP T

AC CE PT ED

23

Figure 3. FSAP WT-SPD but not MI-SPD activates cells.

VSMC (A and C) and EC (B and D) were stimulated with WT-SPD (10 µg/ml) (serine protease domain of FSAP), MI-SPD (Marburg I isoform) or control buffer for 4 h.

qPCR analysis of AREG, IL6 and PTGS2 mRNA levels relative to GAPDH for VSMC (A) and VCAM1, SELE and IL8 mRNA levels relative to GAPDH for EC (B) was performed. Supernatants were analyzed for secreted IL-6 in VSMC (C) and secreted IL-8 for EC (D). Each panel represents results from 3-6 independent experiments;

mean + SEM, *indicate p<0.05 using Kurskal-Wallis followed by Dunn’s post test.

Figure 4. FSAP induced gene expression in VSMC required ERK1/2 and cAMP pathway as well as the PAR1 receptor.

VSMC were pretreated with Vorapaxar (1 µM), PD98059 (10 µM), 8-Br-cAMPS (1 µM) or DMSO for 30 min before stimulation with (A+B) FSAP (10 µg/ml), (C+D) thrombin (1 U/ml) or HE buffer for 4 h. qPCR analysis of (A+C) AREG and (B+D) IL6 mRNA levels relative to GAPDH was performed. Results are mean ± SEM from 4-5 independent experiments, *indicate p<0.05 using Kurskal-Wallis followed by Dunn’s post test.

Figure 5: Effect of PAR1-antagonist on gene expression in EC after stimulation of cells with FSAP or thrombin.

EC were treated for 30 min with Vorapaxar (1µM) or DMSO before induction with either (A) FSAP (10 µg/ml) or control HE buffer or (B) thrombin (1 U/ml) for 4 h.

Expression of IL8 was determined by qPCR and normalized to GAPDH. Results are pooled from 4-5 independent experiments and are mean ± SEM, * denotes p< 0.05 using Kurskal-Wallis followed by Dunn’s post test.

(25)

M AN US CR IP T

AC CE PT ED

24 Graphical abstract

Highlights

-FSAP upregulated inflammation, proliferation and apoptosis-related genes in vascular cells.

-Proteolytically inactive and Marburg I isoform of FSAP had no effect on cells.

-Some, but not all, effect of FSAP were mediated through PAR-1.

(26)

M AN US CR IP T

AC CE PT ED

(A) (B)

(C)

(E)

(D)

(F)

FSAP (µg/ml)

Figure 1. FSAP increases mRNA expression and cytokine secretion in VSMC and EC in a time and dose dependent manner: (A) VSMC and (B) EC were stimulated in a dose dependent manner with FSAP (0-10 µg/ml) for 4 h.

mRNA expression of AREG, IL6 and PTGS2 relative to GAPDH in VSMC and VCAM1, SELE and IL8 relative to GAPDH in EC was determined by qPCR.

VSMC were stimulated with FSAP, or the control buffer (HE) in a time course (C) or in a concentration-dependent manner (D) as described above and IL6 secretion in cell supernatants was determined. EC were stimulated with FSAP in a time course (E) or in a concentration-dependent manner (F) as described above and IL8 secretion in cell supernatants was determined. Results represent mean of 3-5 replicates + SEM and *indicates p<0.05 using Kurskal-Wallis test.

0 2 4 6 8

0 2,5 5 7,5 10

FSAP (µg/ml) 0

2 4 6 8 10 12

0 2,5 5 7,5 10

VCAM1 SELE IL8

mRNA relative toGAPDH

0 10 20 30 40

0 2,5 5 7,5 10 AREG IL6 PTGS2

FSAP (µg/ml)

0 1 2 3 4

0 2,5 5 7,5 10

FSAP (µg/ml) Secreted IL6 (fold change) Secreted IL8 (fold change)

*

0 2 4 6 8

0 4 8 12 16 20 24 HE FSAP

0 2 4 6

0 4 8 12 16 20 24 HE FSAP Time (h)

Time (h)

*

* *

*

FSAP

(27)

M AN US CR IP T

AC CE PT ED

0 1 2 3 4

- Aprotinin AREG HE FSAP

0 2 4 6 8 10

- Aprotinin VCAM1

HE FSAP

Byskov et al., Figure 2

0 1 2 3 4

- Aprotinin IL6

0 1 2 3

- Aprotinin PTGS2

*

Figure 2. Cellular activation by FSAP requires protease activity:

VSMC (A, C) and EC (B, D) were stimulated with FSAP (10 µg/ml) or HE buffer for 4 h in the absence or presence of aprotinin (25 μg/ml).

In (A) qPCR analysis of AREG, IL6 and PTGS2 mRNA levels relative to GAPDH was performed. In (B) qPCR analysis of VCAM1, SELE and IL8 relative to GAPDH was performed. (C) Secreted IL-6 was determined in VSMC using ELISA. (D) Secreted IL-8 was determined in EC using ELISA. Each panel represents results from 3-6 independent experiments; mean + SEM, *indicate p<0.05 using Kurskal-Wallis followed by Dunn’s post test.

(A) (B)

0 2 4 6

- Aprotinin SELE

0 2 4 6

- Aprotinin

* IL8

*

0 1 2 3 4

- Aprotinin HE FSAP

0 1 2 3 4 5

- Aprotinin HE FSAP

Secreted IL-6 (fold change)

(C) (D)

*

Secreted IL-8 (fold change) *

(28)

M AN US CR IP T

AC CE PT ED

0 1 2

- WT-SPD MI-SPD AREG

0 1 2 3

- WT-SPD MI-SPD PTGS2 0

1 2 3

- WT-SPD MI-SPD IL6

0 2 4 6 8

- WT-SPD MI-SPD VCAM1

0 2 4 6

- WT-SPD MI-SPD IL8

0 2 4 6 8

- WT-SPD MI-SPD SELE

* *

*

Figure 3. FSAP WT-SPD but not MI-SPD activates cells: VSMC (A and C) and EC (B and D) were stimulated with WT-SPD (10 µg/ml) (serine protease domain of FSAP), MI-SPD (Marburg I isoform) or control buffer for 4 h. qPCR analysis of AREG, IL6 and PTGS2 mRNA levels relative to GAPDH for VSMC (A) and VCAM1, SELE and IL8 mRNA levels relative to GAPDH for EC (B) was performed.

Supernatant were analyzed for secreted IL-6 in VSMC (C) and secreted IL-8 for EC (D). Each panel represents results from 3-6 independent experiments; mean + SEM, *indicate p<0.05 using Kurskal-Wallis followed by Dunn’s post test.

(A) (B)

Byskov et al., Figure 3

(C) (D)

Secreted IL-8 (fold change) Secreted IL-6 (fold change)

0 1 2 3

- WT-SPD MI-SPD 0 1 2

- WT-SPD MI-SPD

*

* *

(29)

M AN US CR IP T

AC CE PT ED

0 1 2 3

4 AREG

HE FSAP

(A)

(B)

Figure 4. FSAP induced gene expression in VSMC requires ERK1/2 and cAMP pathway as well as the PAR1 receptor:

VSMC were pretreated with Vorapaxar (1 µM), PD98059 (10 µM), 8-Br-cAMPS (1 µM) or DMSO for 30 min before stimulation with (A+B) FSAP (10 µg/ml), (C+D) Thrombin (1 U/ml) or HE buffer for 4 h. qPCR analysis of (A+C) AREG and (B+D) IL6 mRNA levels relative to GAPDH was performed. Results are mean + SEM from 4-5 independent experiments, *indicate p<0.05 using Kurskal-Wallis followed by Dunn’s post test.

Byskov et al., Figure 4

*

(C)

(D)

0 1 2 3 4 5

DMSO Vorapaxar AREG

HE Thrombin

0 1 2 3

4 IL6

0 1 2 3

DMSO Vorapaxar IL6

*

(30)

M AN US CR IP T

AC CE PT ED

Figure 5: Effect of PAR1-antagonist on gene expression in EC after stimulation of cells with FSAP or thrombin: EC were treated for 30 min with Vorapaxar (1µM) or DMSO before induction with either (A) FSAP (10 µg/ml) or control HE buffer or (B) thrombin (1 U/ml) for 4 h. Expression of IL8 was determined by qPCR and normalized to GAPDH. Results are pooled from 4- 5 independent experiments and are mean + SEM, * denotes p<

0.05 using Kurskal-Wallis followed by Dunn’s post test.

Byskov et al., Figure 5

0 1 2 3 4

HE FSAP

*

mRNA relative to GAPDH

0 20 40 60

HE Thrombin

*

(A) (B)