The present invention concerns a method for the modification of a substrate surface by grafting a peptide onto the surface of said substrate, as well as the modified surface substrate obtainable by said method. The present invention also relates to a CD31 -specific biomimetic peptide coating and the method for a standardized and oriented grafting of which fosters the attachment and physiologic function of endothelial cells onto endovascular-suitable prosthetic surfaces, as well as the use of the coated substrates obtainable by said method.

Arterial stents are made of an interconnecting network of solid elements, or struts, made of wires, tubes, or sheets, rolled in cylinder shaped scaffolding conceived to maintain adequate blood flow rate and direction in diseased arteries. Their use has become unavoidable in the management of arterial stenosis, across which balloon-inflatable stents are implanted to keep patent the targeted arterial segment and is gaining increasing importance in the management of saccular aneurysms, from which the arterial flow is diverted by auto-expandable, tightly meshed stents.

Although balloon-expandable stents and flow diverting stents were developed for the treatment of two different arterial pathologies (reopen stenotic arteries to prevent organ ischemia vs diverting the flow from arterial saccular aneurysms to prevent the hemorrhage due to its rupture), and the working mechanisms of these devices is different, both are associated with complications stemming from biocompatibility issues. In particular, the rapid migration, growth to confluence and acquisition of a physiologic (anti-inflammatory and anti-thrombotic) phenotype of the adjacent arterial endothelial cells onto the stent struts is key to the integration and perfect function of the endoprosthesis in the treated arterial segment. No such biocoating exists for flow diverters.

Several strategies aimed at conferring the device with adhesive properties for endothelial cells are already available (capture of circulating endothelial progenitors by anti-CD34 antibodies, use of synthetic peptides derived from the RGD repeated sequence of extracellular matrix components) but none is able to actively and specifically promote endothelial physiologic functions. The latter are critically driven by the engagement of the trans-homophilic CD31 receptor, which is constitutively expressed at a very high density at the surface of endothelial cells (10 ⁶ molecules/cell).

The surfaces of stents does not display the chemical functions required for the conjugation of biomolecules. Thus, their surface must be‘functionalized’ for the subsequent covalent immobilization of a bioactive molecule such as a peptide. A possible approach that allows the direct immobilization of the peptide on an alloy is the plasma glow discharge. However, the use of this technique on certain metallic surfaces, such the nitinol, is challenging, and alternative polymer-based solutions have to be considered.

The use of polymer coatings as intermediate layers for the immobilization of bioactive molecules has several advantages. The first one is that, unlike almost all other types of materials, most polymers either contain functional groups that can react with bioactive molecules, or are easy to functionalize with such groups. The second one is that polymers are generally inexpensive and easy to process into coatings. Finally, there exists a very broad range of polymers, which allows for the fine-tuning of the chemical properties of the coatings. For these reasons, the use of polymer films is generally considered as an optimal strategy for the immobilization of bioactive molecules and has been the preferred system in the design of coated stents.

However, the most used polymer films do have some limitations, especially in their application as biomaterials coating. Their adhesion to the metal, their resistance to stent deployment, and their stability properties must be adapted to the intended use, in order to prevent the deleterious biological effects of delamination (the detachment of the film from its substrate) and uncontrolled degradation. Above all, the biocompatibility of the polymer coatings and of their degradation products is key to the biological performances of coated stents.

Polymer films can be deposited through a multitude of coating methods. Reproducibility, scalability and limited cost are essential requirements for any process developed for an industrial application, and coating methods for medical devices are no exception. In addition, processes used in the manufacturing of medical devices need to be meet some extra criteria regarding the safety and sterility of the end product, stated by the authorities in the“Good Manufacturing Practices”. The aim of the present invention is thus to provide a method for immobilizing a peptide, in particular a CD31 -mimetic peptide on a substrate surface, in particular a stent surface, allowing a strong anchoring of said peptide.

The aim of the present invention is also to provide a method for immobilizing a peptide on a substrate surface, in particular a stent surface, being reproducible, scalable and with limited cost, and also satisfying the requirements regarding the safety and sterility of the end product.

Therefore, the present invention relates to a method for the modification of a substrate surface by grafting a peptide onto the surface of said substrate, comprising the following steps: a) the coating of a polydopamine layer onto the surface of the substrate in order to obtain a polydopamine coated surface; b) the modification of the polydopamine coated surface by the addition of a linker, in particular a biorthogonal linker, comprising at least one reactive moiety chosen from alkyne, in particular cyclooctyne, functions, in order to obtain a modified polydopamine coated surface; and c) the addition of a peptide comprising an azide terminal group and its reaction with the alkyne function of the linker of step b), in order to obtain a polydopamine coated surface grafted by a peptide,

said peptide comprising an azide terminal group being a peptide which is chemically modified with an azide terminal group,

said peptide being not a CD31 -derived peptide consisting of a sequence selected from the group consisting of: SEQ ID NO: 2 to 8, SEQ ID NO: 12 to 79, and SEQ ID NO: 81 when the substrate is a metallic implantable device for interventional neuroradiology.

Thus, the present invention provides a method for immobilizing a peptide, in particular a CD31 -mimetic peptide on a substrate surface, in particular a stent surface, allowing a strong anchoring of said peptide onto a polydopamine polymer functionalized by biorthogonal copper-free chemistry allowing for a standardized density and controlled orientation of said peptide. Step a)

As mentioned above, the method of the invention comprises a step consisting in coating a polydopamine layer onto the surface of the substrate. This step thus leads to a polydopamine coated surface.

The substrate obtained after step a) corresponds to the starting substrate the surface of which is coated with a polydopamine layer.

Polydopamine (PDA) is a bio-inspired, self-assembling polymer which was discovered in 2007 by Messersmith and colleagues. This discovery originated from their investigations on the adhesive proteins of marine mussels. Dopamine, a small molecule previously known for its biological role as a neurotransmitter, which combines an amine and a catechol group (which is converted into quinone by oxidation), when dissolved in an aqueous buffer at a slightly basic pH, self- polymerizes into a very adherent film, on various types of substrates. Besides, PDA exhibits latent reactivity towards amine and thiol groups, which makes it a very attractive substrate for bioactive molecule immobilization.

As for the biocompatibility properties of PDA, they appeared to be very adapted to its application as a stent coating application: it was shown to promote endothelial cell adhesion and proliferation, to decrease platelet adhesion and to reduce SMC proliferation (Ku, S. H., Ryu, J., Hong, S. K., Lee, H. and Park, C. B. 2010. Biomaterials 31 (9): 2535-2541 ; Yang, Z., Tu, Q., Zhu, Y., Luo, R., Li, X., Xie, Y., Maitz, M. F., Wang, J. and Huang, N. 2012. Advanced Healthcare Materials 1 (5): 548-559; Luo, R., Tang, L., Zhong, S., Yang, Z., Wang, J., Weng, Y., Tu, Q., Jiang, C. and Huang, N. 2013. ACS Applied Materials & Interfaces 5(5): 1704-1714).

According to an embodiment, step a) comprises contacting, under stirring, the surface of the substrate with an alkaline (preferably at pH 8.5) solution of dopamine in the air and incubating said substrate and said solution, preferably at a temperature comprised between 18°C and 30 °C, in paticular at room temperature, and preferably for a duration comprised between 18 hours and 30 hours, in particular comprised between 20h and 24h.

Preferably, step a) is followed by a rinsing step of the polydopamine coated surface, in particular with deionized water. According to an embodiment, the polydopamine layer has a thickness comprised between 20 nm and 100 nm, preferably between 30 nm and 50 nm, and more preferably of 45 nm.

Step b)

As mentioned above, step b) consists in modifying the polydopamine coated surface of step a) through the fixation of a biorthogonal, copper-free click chemistry- suitable linker. This step thus leads to a modified polydopamine coated surface, which comprises the polydopamine layer as defined above and a layer comprising the linker onto the surface of the substrate.

The substrate obtained after step b) corresponds to the starting substrate comprising a polydopamine layer on its surface, said polydopamine layer being further coated with a layer comprising the linker as defined above.

According to an embodiment, step b) comprises contacting the polydopamine coated surface of the substrate with a solution of the linker and incubating said substrate and said solution, under stirring, preferably at a temperature comprised between 18°C and 30 °C, in particular at room tempeature, and preferably for a duration comprised between 18 hours and 30 hours, in particular comprised between 20h and 24h.

Preferably, step b) is followed by a rinsing step of the modified polydopamine coated surface, in particular with deionized water.

According to an embodiment, the layer made of the linker is obtained has a thickness comprised between 0.03 nm and 3 nm, preferably between 0.1 nm and 0.2 nm, and more preferably of 0.15 nm.

Linker

The linker according to the invention which is used in step b) is an alkyne derivative, and preferably a cyclooctyne derivative and is thus characterized by the presence of at least one triple bond, especially able to react with an azide group, in particular by click chemistry. According to an embodiment, the linker has the formula (1-1 ):

(1-1 )

wherein R is a radical of formu

. X- _] being chosen from the group consisting of: -CONH-, -CO-, -CS- and - CSNH, X- _| being preferably -CONH- or -CO-, and

. A- _| being an alkylene radical comprising from 2 to 40 carbon atoms, possibly interrupted by at least one oxygen atom.

According to a preferred embodiment, the linker according to the invention comprising at least one alkyne function has the following formula (I):

wherein n is an integer comprised between 2 and 14.

Preferably, the linker according to the invention has the following formula (II):

Step c)

As mentioned above, step c) consists in further modifying the modified polydopamine coated surface of step b) through the addition of a peptide able to react with the linker as mentioned above. This step thus allows the immobilization or grafting of said peptide onto the substrate surface.

Step c) thus leads to a modified polydopamine coated surface grafted by a peptide, which comprises the polydopamine layer as defined above and a layer comprising the linker onto the surface of the substrate.

The substrate obtained after step c) corresponds to the starting substrate comprising a polydopamine layer on its surface, said polydopamine layer being further coated with a layer made of the linker as defined above and the peptide as defined above onto the surface of the substrate.

According to an embodiment, step c) comprises contacting the modified polydopamine coated surface of the substrate with a solution of the peptide in water, at room temperature during 24 hours, at a concentration comprised between 0.001 mM/cm ² and 200 mM/cm ² of surface of the substrate.

Preferably, step c) comprises a step of copper-free click chemistry reaction.

According to an embodiment, the thickness of the layers made of the linker and of the peptide is comprised between 0.5 nm and 15 nm, preferably between 1 nm and 10 nm, and more preferably of 5.6 nm.

According to an embodiment, the thickness of the polydopamine layer and of the layers made of the linker and of the peptide is comprised between 20 nm and 200 nm, preferably between 10 nm and 160 nm, and more preferably between 30 nm and 120 nm.

Peptide comprising an azide terminal group

As mentioned above, the peptide to be grafted or immobilized on the substrate surface comprises an azide terminal group and is thus able to react with the linker comprising at least one triple bond, especially by click chemistry.

The peptide comprising an azide terminal group according to the invention is a peptide as defined below, preferably a CD31 -derived peptide, which is chemically modified with an azide terminal group. The advantages of a CD31 mimetic coating of surfaces such as stents are achieved through the present method of coating endovascular stents based on a simple, reproducible and scalable three dip-coating step procedure, wherein the bioactive agent is a retroinverso CD31 mimetic peptide and its coating is density- and orientation-controlled by the use of biorthogonal copper-free chemistry.

Preferably, the peptide comprising an azide terminal group consists of 3 to 25 amino acids, more preferably of 10 to 19 amino acids, more preferably of 12 to 14 amino acids.

According to a preferred embodiment, the peptide comprising an azide terminal group has the formula:

N =N ⁺ =N-(AA) _n -OH

wherein:

- AA is a residue of an amino acid, and

- n is an integer comprised between 2 and 25, in particular between 4 and 25, preferably between 10 to 19 amino acids, more preferably between 10 and 16, for example 12, 13 or 14 amino acids.

According to a preferred embodiment, the peptide used in the present invention is a CD31 -derived peptide, which is then modified (i) optionally, by adding a spacer at the N-terminus end and (ii) by the grafting of an azide-terminal group. More preferably, such peptide is used with a substrate such as endovascular stents.

According to a preferred embodiment, the peptide comprising an azide terminal group has the formula:

N =N ⁺ =N-(spacer)-(CD31 -derived peptide)-OH wherein:

- the spacer is an amino acid spacer comprising at least 2 amino acids, preferably at least 3 amino acids, and

- the CD31 -derived peptide is as defined below.

The spacer may for example consist of 2 to 14 amino acids, for example 4, 5, 6, 7, 8, 9 or 10 amino acids.

The spacer may for example consist of sequence KGGG (SEQ ID NO: 95), wherein the amino acids are preferably D-enantiomer amino acids. This embodiment allows the promotion of the regulatory functions of CD31 in the cells that directly enter in contact with the stent. Thus, the CD31 coating confers anti-thrombotic and anti-inflammatory properties to the surface, and, above all, it promotes the rapid formation of a functional endothelium on the stent struts. The CD31 agonist P8RI peptide as mentioned below is designed to achieve this goal by targeting the CD31 sequence involved in the cis-homophilic engagement which naturally occurs when endothelial cells, leukocytes, or platelets enter in contact with each other, and which is essential for the intracellular CD31 signaling.

CD31 -derived peptide

As explained above, according to an embodiment, the CD31 -derived peptide (also called CD31 peptide) is used as a peptide (which may also be named initial peptide peptide) and is further chemically modified in order to obtain the peptide comprising an azide terminal group according to the invention.

The initial peptide is preferably a peptide as disclosed in WO2010/000741 or W02013/190014 (before its modification with an azide-terminal group and optionally before adding a spacer).

As mentioned above, according to the invention, when the substrate is a metallic implantable device for interventional neuroradiology, the initial peptide (corresponding to the peptide before its chemical modification as explained above) is not a CD31 -derived peptide consisting of a sequence selected from the group consisting of: SEQ ID NO: 2 to 8, SEQ ID NO: 12 to 79, and SEQ ID NO: 81.

Preferably, the CD31 peptide is a synthetic peptide.

By a“synthetic peptide”, it is intended that the peptide is not present within a living organism, e.g. within human body.

The synthetic peptide may be part of a composition or a kit.

The synthetic peptide is preferably purified.

The initial peptide may be selected in the group consisting of:

(i) a peptide consisting of a fragment of 3 to 15 amino acids of the sequence defined by amino acids 579 to 601 of sequence SEQ ID NO: 1 ,

(ii) a peptide consisting of a fragment of 3 to 15 amino acids of a sequence corresponding to the amino acids 579 to 601 of sequence SEQ ID NO: 1 in a non human mammalian CD31 , (iii) a peptide of 3 to 15 amino acids consisting of a sequence at least 70% identical to the sequence of peptide (i),

(iv) a peptide consisting of a retro-inverso sequence of peptide (i), (ii) or (iii), and

(v) peptide (i), (ii), (iii) or (iv) comprising at least one or at least one further chemical modification, preferably at least one amino acid in the D-enantiomer form.

A“fragment” refers herein to a sequence of consecutive amino acids. For example, a fragment may be a fragment of 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14 or 15 amino acids.

The sequence defined by amino acids 579 to 601 of sequence SEQ ID NO: 1 is sequence SEQ ID NO: 12.

Thus, the peptide may consist of a fragment of 3 to 15 amino acids of sequence SEQ ID NO: 12.

The initial peptide may also consist of a fragment of 3 to 15 amino acids of a sequence corresponding to sequence SEQ ID NO: 12 in a non-human mammalian CD31.

Non-limiting examples of non-human mammalian CD31 are the murine CD31 of sequence SEQ ID NO: 9, the bovine CD31 of sequence SEQ ID NO: 10 and the pig CD31 of sequence SEQ ID NO: 1 1.

The person skilled in the art can easily identify a sequence corresponding to the amino acids 579 to 601 of sequence SEQ ID NO: 1 , i.e. to sequence SEQ ID NO: 12, in a non-human mammalian CD31 protein, for example by performing a sequence alignment between sequence SEQ ID NO: 1 and the sequence of said non-human mammalian CD31 protein, for example with one of sequences SEQ ID NO: 9, SEQ ID NO: 10 and SEQ ID NO: 1 1.

Methods for sequence alignment and determination of sequence identity are well known in the art, for example using publicly available computer software such as BioPerl, BLAST, BLAST-2, CS-BLAST, FASTA, ALIGN, ALIGN-2, LALIGN, Jaligner, matcher or Megalign (DNASTAR) software and alignment algorithms such as the Needleman-Wunsch and Smith-Waterman algorithms.

The sequence of the CD31 peptide according to the invention is preferably derived from the sequence of human CD31 or murine CD31.

The initial peptide may also be a peptide of 3 to 15 amino acids consisting of a sequence at least 70%, at least 75%, at least 80%, at least 85% or at least 90% identical to the sequence of peptide (i), i.e. to the sequence of a fragment of 3 to 15 amino acids of the sequence defined by amino acids 579 to 601 of sequence SEQ ID NO: 1.

A peptide sequence at least 70% identical to a given sequence of 4 to 6 amino acids differs from said given sequence of at most one amino acid.

A peptide sequence at least 70% identical to a given sequence of 7 to 9 amino acids differs from said given sequence of at most two amino acids.

A peptide sequence at least 70% identical to a given sequence of 10 to 13 amino acids differs from said given sequence of at most three amino acids

A peptide sequence at least 70% identical to a given sequence of 14 or 15 amino acids differs from said given sequence of at most four amino acids.

By“a sequence at least x% identical to a reference sequence”, it is intended that the amino acid sequence of the subject peptide is identical to the reference sequence or differ from the reference sequence by up to 100-x amino acid alterations per each 100 amino acids of the reference sequence. In other words, to obtain a polypeptide having an amino acid sequence at least x% identical to a reference amino acid sequence, up to 100-x% of the amino acid residues in the subject sequence may be inserted, deleted or substituted with another amino acid.

Methods for comparing the identity of two or more sequences are well known in the art. For instance, programs available in the Wisconsin Sequence Analysis Package, version 9.1 , for example the programs BESTFIT and GAP, may be used to determine the % identity between two polypeptide sequences. BESTFIT uses the "local homology" algorithm of Smith and Waterman and finds the best single region of similarity between two sequences. Other programs for determining identity between sequences are also known in the art, for instance the Needle program, which is based on the Needleman and Wunsch algorithm, described in Needleman and Wunsch (1970) J. Mol Biol. 48:443-453, with for example the following parameters for polypeptide sequence comparison: comparison matrix: BLOSUM62, gap open penalty: 10 and gap extend penalty: 0.5, end gap penalty: false, end gap open penalty = 10, end gap extend penalty = 0.5; and the following parameters for polynucleotide sequence comparison: comparison matrix: DNAFULL; gap open penalty = 10, gap extend penalty = 0.5, end gap penalty: false, end gap open penalty = 10, end gap extend penalty = 0.5. Peptides consisting of an amino acid sequence“at least 70%, 75%, 80%, 85%, or 90% identical” to a reference sequence may comprise mutations, such as deletions, insertions and/or substitutions compared to the reference sequence.

In case of substitutions, the substitution preferably corresponds to a conservative substitution as indicated in the Table 1 below. In a preferred embodiment, the peptide consisting of an amino acid sequence at least 70%, 75%, 80%, 85% or 90%identical to a reference sequence only differs from the reference sequence by conservative substitutions. Table 1

In another preferred embodiment, the peptide consisting of an amino acid sequence at least 70%, 75%, 80%, 85% or 90% identical to a reference sequence corresponds to a naturally-occurring allelic variant of the reference sequence.

In still another preferred embodiment, the peptide consisting of an amino acid sequence at least 70%, 75%, 80%, 85% or 90% identical to a reference sequence corresponds to a homologous sequence derived from another non-human mammalian species than the reference sequence.

In a preferred embodiment, the peptide consisting of an amino acid sequence at least 70%, 75%, 80%, 85% or 90% identical to a reference sequence differs from the reference sequence by conservative substitutions and/or corresponds to a homologous sequence derived from another non-human mammalian species than the reference sequence.

By the expression“a peptide consisting of a retro-inverso sequence of peptide (i), (ii) or (iii)”, it is herein meant a peptide that differs from the peptide (i), (ii) or (iii) in that its amino acids are in the reverse order by comparison to the sequence of peptide (i), (ii) or (iii), respectively, and consist of D-amino acids instead of the naturally-occurring L-amino acids. D-enantiomers of amino acids (also called D-amino acids) are referred to by the same letter as their corresponding L-enantiomer (also called L-amino acid), but in lower case. Thus, for example, the L-enantiomer of arginine is referred to as‘R’, while the D-enantiomer is referred to as‘r’.

Preferably, the initial peptide is soluble in an organic or nonorganic solvent.

In a preferred embodiment, the initial peptide is soluble in water. More particularly, the peptide is preferably soluble in water and/or in aqueous buffer such as NaCI 9 g/L, PBS, Tris or Tris-phosphate. The solubility in water and aqueous buffers is particularly advantageous on the pharmacological point of view. Thanks to such solubility, the peptide may be dissolved in an aqueous solution, for example at a concentration equal to, at least of or at most of 1 micromolar, 10 micromolar, 50 micromolar, 100 micromolar, 500 micromolar, 1 mM, 50 mM or 100 mM.

A peptide of the invention that is readily soluble in water may be obtained by the presence of at least one charged amino acid (preferably arginine (R) and/or lysine (K)), wherein said charged amino acid is not comprised between two hydrophobic residues.

Thus, in a preferred embodiment, the initial peptide comprises at least one charged amino acid, preferably arginine and/or lysine, wherein said charged amino acid is not comprised between two hydrophobic residues.

In a more preferred embodiment, said charged amino acid is located either at the N- or C-terminal end of the sequence.

For example, the sequence of a preferred initial peptide begins with the motif RV (for example instead of VRV).

In a preferred embodiment, the initial peptide is resistant to peptidase, in particular to eukaryote peptidase.

By “resistant to peptidase”, it is herein meant that the peptide remains undigested, as determined by reverse phase-high-performance liquid chromatography (RP-HPLC) and mass spectroscopy (MS), upon incubation at 37°C with mammalian serum or injection in a living laboratory animal. Laboratory tests aimed at evaluating serum stability of the peptides are well standardized (see for example Jenssen and Aspmo, 2008, Methods Mol Biol 494, 177-186). Highly peptidase-resistant peptides are those that remains undigested for up to 70% of their original mass and/or displaying a half-life longer than 240 minutes in the presence of proteolytic enzymes (see for example Kumarasinghe and Hruby, 2015, In Peptide Chemistry and Drug Design, B.M. Dunn, ed. (Hoboken, New Jersey: Wiley), pp. 247-266).

The peptide is preferably resistant to peptidases present in blood, such as soluble peptidases or peptidases present on cell surface.

The initial peptide may also comprise at least one or at least one further chemical modification, preferably to improve its stability and/or bioavailability.

Such chemical modifications globally aim at obtaining peptides with increased resistance against enzymatic degradation in vivo, thus increasing its half-life and/or maintaining or improving its biological activity. Any chemical modification known in the art can be employed according to the present invention.

The initial peptide may comprise at least one artificial amino acid, said artificial amino acid being preferably selected from the group consisting of a D-enantiomer amino acid, a beta-methyl amino acid, a alpha-substituted alpha-amino acid and an amino acid analog.

By“beta-methyl amino acid”, it is herein meant a derivative of the amino acid alanine with an aminomethyl group on the side chain. This non-proteinogenic amino acid is classified as a polar base.

By“alpha-substituted alpha-amino acid”, it is herein meant that the group on the alpha carbon of an L-amino acid (NH2) has been changed to another, non proteinaceous group, such as a methyl-, aryl- or acyl- group.

By“amino acid analog”, it is herein meant any other artificial analog of a natural amino acid.

The initial peptide may thus comprise at least one amino acid in the D- enantiomer form. For example, 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14 or 15 of the amino acids of the peptide defined above may be in the D-enantiomer form.

In one embodiment, the initial peptide consists of D-amino acids.

The initial peptide may also comprise an inverted sequence, namely an inversion of the amino acid chain (from the C-terminal end to the N-terminal end). The entire amino acid sequence of the peptide may be inverted, or a portion of the amino acid sequence may be inverted. For example, a consecutive sequence of 2, 3, 4, 5, 6, 7, 8, 9, 10, 1 1 , 12, 13, 14 or 15 amino acids may be inverted. Reference herein to‘inverted’ amino acids refers to inversion of the sequence of consecutive amino acids in the sequence.

Other chemical modifications include, but are not limited to:

- modifications to the N-terminal and/or C-terminal ends of the peptides such as e.g. N-terminal acylation (preferably acetylation) or deamination, or modification of the C-terminal carboxyl group into an amide or an alcohol group;

- modifications at the amide bond between two amino acids: acylation (preferably acetylation) or alkylation (preferably methylation) at the nitrogen atom or the alpha carbon of the amide bond linking two amino acids;

- modifications at the alpha carbon of the amide bond linking two amino acids such as e.g. acylation (preferably acetylation) or alkylation (preferably methylation) at the alpha carbon of the amide bond linking two amino acids;

- retro-inversions in which one or more naturally-occurring amino acids (L- enantiomer) are replaced with the corresponding D-enantiomers, together with an inversion of the amino acid chain (from the C-terminal end to the N-terminal end);

- azapeptides, in which one or more alpha carbons are replaced with nitrogen atoms; and/or

- betapeptides, in which the amino group of one or more amino acid is bonded to the b carbon rather than the a carbon.

The peptide includes amino acids modified either by natural processes, such as post-translational processing, or by chemical modification techniques which are well known in the art. Such modifications are well described in basic texts and in more detailed monographs, as well as in a voluminous research literature. Modifications can occur anywhere in a polypeptide, including the peptide backbone, the amino acid side-chains and the amino or carboxyl termini, it will be appreciated that the same type of modification may be present in the same or varying degrees at several sites in a given polypeptide. Also, a given polypeptide may contain many types of modifications. Polypeptides may be branched as a result of ubiquitination, and they may be cyclic, with or without branching. Cyclic, branched and branched cyclic polypeptides may result from natural post-translational processes or may be made by synthetic methods. Modifications include acetylation, acylation, ADP- ribosylation, araidation, covalent attachment of flavin, covalent attachment of a heme moiety, covalent attachment of a nucleotide or nucleotide derivative, covalent attachment of a lipid or lipid derivative, covalent attachment of phosphatidyl inositol, cross-linking, cyclization, disulfide bond formation, demethylation, formation of covalent cross-links, formation of cystine, formation of pyroglutarnate, formylation, gamma-carboxylation, glycosylation, GPI anchor formation, hydroxylation, iodination, methylation, myristoylation, oxidation, proteolytic processing, phosphorylation, prenylation, racemization, selenoyiation, sulfation, transfer-RNA mediated addition of amino acids to proteins such as arginylation, and ubiquitination.

In a preferred embodiment of the invention, the initial peptide is selected in the group consisting of:

(i) a peptide consisting of a fragment of 3 to 15 amino acids of the sequence defined by amino acids 579 to 601 of sequence SEQ ID NO: 1 , said fragment comprising the amino acids 579 to 581 , the amino acids 589 to 591 , the amino acids 599 to 601 and/or the amino acids 593 to 595 of SEQ ID NO: 1 ,

(ii) a peptide consisting of a fragment of 3 to 15 amino acids of a sequence corresponding to the amino acids 579 to 601 of sequence SEQ ID NO: 1 in a non human mammalian CD31 , for example a fragment of 3 to 15 amino acids of the sequence defined by amino acids 568 to 590 of sequence SEQ ID NO: 9, said fragment preferably comprising the amino acids 568 to 570, the amino acids 578 to 580, the amino acids 588 to 590 and/or the amino acids 582 to 584 of SEQ ID NO: 9,

(iii) a peptide of 3 to 15 amino acids consisting of a sequence at least 70% identical, preferably at least 75% identical, preferably at least 80% identical, more preferably at least 85% identical, still more preferably at least 90% identical to the sequence of peptide (i),

(iv) a peptide consisting of a retro-inverso sequence of peptide (i), (ii) or (iii), and

(v) the peptide (i), (ii), (iii) or (iv) comprising at least one or at least one further chemical modification.

Such initial peptide has, for example, a sequence selected from the group consisting of: SSTLAVRVFLAPWKK (SEQ ID NO: 13, amino acids 576 to 590 of

SEQ ID NO: 9), STLAVRVFLAPWKK (SEQ ID NO: 14, amino acids 577 to 590 of

SEQ ID NO: 9), TLAVRVFLAPWKK (SEQ ID NO: 15, amino acids 578 to 590 of

SEQ ID NO: 9), LAVRVFLAPWKK (SEQ ID NO: 16, amino acids 579 to 590 of SEQ

ID NO: 9), AVRVFLAPWKK (SEQ ID NO: 17, amino acids 580 to 590 of SEQ ID NO: 9), VRVFLAPWKK (SEQ ID NO: 3, amino acids 581 to 590 of SEQ ID NO: 9), RVFLAPWKK (SEQ ID NO: 18, amino acids 582 to 590 of SEQ ID NO: 9), VFLAPWKK (SEQ ID NO: 19, amino acids 583 to 590 of SEQ ID NO: 9), FLAPWKK (SEQ ID NO: 20, amino acids 584 to 590 of SEQ ID NO: 9), LAPWKK

(SEQ ID NO: 2, amino acids 585 to 590 of SEQ ID NO: 9), APWKK

(SEQ ID NO: 21 , amino acids 586 to 590 of SEQ ID NO: 9), PWKK (SEQ ID NO: 22, amino acids 587 to 590 of SEQ ID NO: 9), WKK (amino acids 588 to 590 of SEQ ID NO: 9), SKILTVFtV I LAPWKK (SEQ ID NO: 23, amino acids 587 to 601 of SEQ ID NO: 1 ), KILTVFiV I LAPWKK (SEQ ID NO: 24, amino acids 588 to 601 of SEQ ID NO: 1 ), I LTVFiVI LAPWKK (SEQ ID NO: 25, amino acids 589 to 601 of SEQ ID NO: 1 ), LTVR VI LAPWKK (SEQ ID NO: 26, amino acids 590 to 601

TVRVILAPWKK (SEQ ID NO: 27, amino acids 591 to 601

VRVILAPWKK (SEQ ID NO: 4, amino acids 592 to 601

RVILAPWKK (SEQ ID NO: 28, amino acids 593 to 601

VILAPWKK (SEQ ID NO: 29, amino acids 594 to 601 of SEQ

(SEQ ID NO: 30, amino acids 595 to 601 of SEQ ID NO: 1 ), SSMRTSPRSSTLAVR (SEQ ID NO: 31 , amino acids 568 to 582 of SEQ ID NO: 9), SSMRTSPRSSTLAV (SEQ ID NO: 32, amino acids 568 to 581 of SEQ ID NO: 9), SSMRTSPRSSTLA (SEQ ID NO: 33, amino acids 568 to 580 of SEQ ID NO: 9), SSMRTSPRSSTL (SEQ ID NO: 34, amino acids 568 to 579 of SEQ ID NO: 9), SSMRTSPRSST (SEQ ID NO: 35, amino acids 568 to 578 of SEQ ID NO: 9), SSMRTSPRSS (SEQ ID NO: 36, amino acids 568 to 577 of SEQ ID NO: 9), SSMRTSPRS (SEQ ID NO: 37, amino acids 568 to 576 of SEQ ID NO: 9), SSMRTSPR (SEQ ID NO: 38, amino acids 568 to 575 of SEQ ID NO: 9), SSMRTSP (SEQ ID NO: 39, amino acids 568 to 574 of SEQ ID NO: 9), SSMRTS (SEQ ID NO: 40, amino acids 568 to 573 of SEQ ID NO: 9), SSMRT (SEQ ID NO: 41 , amino acids 568 to 572 of SEQ ID NO: 9), SSMR (SEQ ID NO: 42, amino acids 568 to 571 of SEQ ID NO: 9), SSM (amino acids 568 to 570 of SEQ ID NO: 9), NHASSVPRSKILTVRVILAPWKK (SEQ ID NO: 97, amino acids 579 to 601 of SEQ ID NO: 1 ), NHASSVPRSKILTVR (SEQ ID NO: 43, amino acids 579 to 593 of SEQ ID NO: 1 ), NHASSVPRSKILTV (SEQ ID NO: 44, amino acids 579 to 592 of SEQ ID NO: 1 ), NHASSVPRSKILT (SEQ ID NO: 45, amino acids 579 to 591 of SEQ ID NO: 1 ), NHASSVPRSKIL (SEQ ID NO: 46, amino acids 579 to 590 of SEQ ID NO: 1 ), NHASSVPRSKI (SEQ ID NO: 47, amino acids 579 to 589 of SEQ ID NO: 1 ), NHASSVPRSK (SEQ ID NO: 48, amino acids 579 to 588 of SEQ ID NO: 1 ), NHASSVPRS (SEQ ID NO: 49, amino acids 579 to 587 of SEQ ID NO: 1 ), NHASSVPR (SEQ ID NO: 50, amino acids 579 to 586 of SEQ ID NO: 1 ), NHASSVP (SEQ ID NO: 51 , amino acids 579 to 585 of SEQ ID NO: 1 ), NHASSV (SEQ ID NO: 52, amino acids 579 to 584 of SEQ ID NO: 1 ), NHASS (SEQ ID NO: 53, amino acids 579 to 583 of SEQ ID NO: 1 ), NHAS (SEQ ID NO: 54, amino acids 579 to 582 of SEQ ID NO: 1 ), NHA (amino acids 579 to 581 of SEQ ID NO: 1 ), TSPRSSTLAVRVFLA (SEQ ID NO: 55, amino acids 572 to 586 of SEQ ID NO: 9), SPRSSTLAVRVFL (SEQ ID NO: 56, amino acids 573 to 585 of SEQ ID NO: 9), PRSSTLAVRVF (SEQ ID NO: 57, amino acids 574 to 584 of SEQ ID NO: 9), RSSTLAVRV (SEQ ID NO: 58, amino acids 575 to 583 of SEQ ID NO: 9), SSTLAVR (SEQ ID NO: 59, amino acids 576 to 582 of SEQ ID NO: 9), ST LAV (SEQ ID NO: 60, amino acids 577 to 581 of SEQ ID NO: 9), TLA (amino acids 578 to 580 of SEQ ID NO: 9), SVPRSKILTVRVILA (SEQ ID NO: 61 , amino acids 583 to 597 of SEQ ID NO: 1 ), VPRSKILTVRVIL (SEQ ID NO: 62, amino acids 584 to 596 of SEQ ID NO: 1 ), PRSKILTVRVI (SEQ ID NO: 63, amino acids 585 to 595 of SEQ ID NO: 1 ), RSKILTVRV (SEQ ID NO: 64, amino acids 586 to 594 of SEQ ID NO: 1 ), SKILTVR (SEQ ID NO: 65, amino acids 587 to 593 of SEQ ID NO: 1 ), KILTV (SEQ ID NO: 66, amino acids 588 to 562 of SEQ ID NO: 1 ), ILT (amino acids 589 to 591 of SEQ ID NO: 1 ), RVF (amino acids 582 to 584 of SEQ ID NO: 9), RVFL (SEQ ID NO: 67, amino acids 582 to 585 of SEQ ID NO: 9), RVFLA (SEQ ID NO: 68, amino acids 582 to 586 of SEQ ID NO: 9), RVFLAP (SEQ ID NO: 69, amino acids 582 to 587 of SEQ ID NO: 9), RVFLAPW (SEQ ID NO: 70, amino acids 582 to 588 of SEQ ID NO: 9), RVFLAPWK (SEQ ID NO: 5, amino acids 582 to 589 of SEQ ID NO: 9), RVI (amino acids 593 to 595 of SEQ ID NO: 1 ), RVIL (SEQ ID NO: 71 , amino acids 593 to 596 of SEQ ID NO: 1 ), RVI LA (SEQ ID NO: 72, amino acids 593 to 597 of SEQ ID NO: 1 ), RVILAP (SEQ ID NO: 73, amino acids 593 to 598 of SEQ ID NO: 1 ), RVILAPW (SEQ ID NO: 74, amino acids 593 to 599 of SEQ ID NO: 1 ), RVILAPWK (SEQ ID NO: 7, amino acids 593 to 600 of SEQ ID NO: 1 ) RSKILTVRVILAPWK (SEQ ID NO: 75), SKILTVRVILAPWK (SEQ ID NO: 76), KILTVRVILAPWK (SEQ ID NO: 77), ILTVRVILAPWK (SEQ ID NO: 78), LTVRVILAPWK (SEQ ID NO: 79), TVRVILAPWK (SEQ ID NO: 80), VRVILAPWK (SEQ ID NO: 81 ), VILAPWK (SEQ ID NO: 82), ILAPWK (SEQ ID NO: 83), LAPWK (SEQ ID NO: 84), APWK (SEQ ID NO: 85), PWK, RSSTLAVRVFLAPWK (SEQ ID NO: 86), SSTLAVRVFLAPWK (SEQ ID NO: 87), STLAVRVFLAPWK (SEQ ID NO: 88), TLAVRVFLAPWK (SEQ ID NO: 89), LAVRVFLAPWK (SEQ ID NO: 90), AVRVFLAPWK (SEQ ID NO: 91 ), VRVFLAPWK (SEQ ID NO: 92), VFLAPWK (SEQ ID NO: 93), FLAPWK (SEQ ID NO: 94) and a retro-inverso sequence of one of these sequences. Said peptide may comprise at least one or at least one further chemical modification.

In one embodiment, the peptide consists of a sequence selected from the group consist of TVRVILAPWK (SEQ ID NO: 80), VRVILAPWK (SEQ ID NO: 81 ), RVILAPWK (SEQ ID NO: 7), VILAPWK (SEQ ID NO: 82), ILAPWK (SEQ ID NO: 83), AVRVFLAPWK (SEQ ID NO: 91 ), VRVFLAPWK (SEQ ID NO: 92), RVFLAPWK (SEQ ID NO: 5), VFLAPWK (SEQ ID NO: 93), FLAPWK (SEQ ID NO: 94) and a retro-inverso sequence of one of these sequences. Said peptide may comprise at least one or at least one further chemical modification.

In one embodiment, the initial peptide is selected in the group consisting of:

(i) a peptide consisting of a fragment of 3 to 15 amino acids of the sequence defined by amino acids 579 to 601 of sequence SEQ ID NO: 1 , said fragment comprising the amino acids 579 to 582, the amino acids 588 to 592, the amino acids 598 to 601 and/or the amino acids 593 to 595 of SEQ ID NO: 1 ,

(ii) a peptide consisting of a fragment of 3 to 15 amino acids of a sequence corresponding to the amino acids 579 to 601 of sequence SEQ ID NO: 1 in a non human mammalian CD31 , for example a fragment of 3 to 15 amino acids of the sequence defined by amino acids 568 to 590 of sequence SEQ ID NO: 9, said fragment preferably comprising the amino acids 568 to 571 , the amino acids 578 to 580 the amino acids 587 to 590 and/or the amino acids 582 to 584 of SEQ ID NO: 9,

(iii) a peptide of 3 to 15 amino acids consisting of a sequence at least 70% identical, preferably at least 75% identical, preferably at least 80% identical, more preferably at least 85% identical, still more preferably at least 90% identical to the sequence of peptide (i),

(iv) a peptide consisting of a retro-inverso sequence of peptide (i), (ii) or (iii), and

(v) the peptide (i), (ii), (iii) or (iv) comprising at least one or at least one further chemical modification.

In a preferred embodiment, the initial peptide is an 8 amino-acid fragment comprising inversions and/or at least one unnatural amino acid, such as at least one D-amino acids. Such peptides indeed retain the activity of the original peptide or even demonstrate improved activity. Incorporation of unnatural amino acids in peptides intended for therapeutic use is of utility in increasing the stability of the peptide, in particular in vivo stability.

In another preferred embodiment of the invention, the initial peptide is selected in the group consisting of a peptide of sequence SEQ ID NO: 2, a peptide of sequence SEQ ID NO: 3, a peptide of sequence SEQ ID NO: 4, a peptide of sequence SEQ ID NO: 5, a peptide of sequence SEQ ID NO: 6 consisting of D- enantiomer amino acids, a peptide of sequence SEQ ID NO: 7 and a peptide of sequence SEQ ID NO: 8 consisting of D-enantiomer amino acids.

A more preferred initial peptide is a peptide of sequence SEQ ID NO: 5 (also called P8F) or a peptide of sequence SEQ ID NO: 6 consisting of D-enantiomer amino acids (also called P8RI).

The peptide may be prepared by any well-known procedure in the art, such as chemical synthesis, for example solid phase synthesis or liquid phase synthesis, or genetic engineering. As a solid phase synthesis, for example, the amino acid corresponding to the C-terminus of the peptide to be synthesized is bound to a support which is insoluble in organic solvents, and by alternate repetition of reactions, one wherein amino acids with their amino groups and side chain functional groups protected with appropriate protective groups are condensed one by one in order from the C-terminus to the N-terminus, and one where the amino acids bound to the resin or the protective group of the amino groups of the peptides are released, the peptide chain is thus extended in this manner. After synthesis of the desired peptide, it is subjected to the de-protection reaction and cut out from the solid support. Such peptide cutting reaction may be carried with hydrogen fluoride or tri-fluoromethane sulfonic acid for the Boc method, and with TFA for the Fmoc method.

Solid phase synthesis methods are largely classified by the tBoc method and the Fmoc method, depending on the type of protective group used. Typically used protective groups include tBoc (t-butoxycarbonyl), Cl-Z (2-chlorobenzyloxycarbonyl), Br-Z (2-bromobenzyloyycarbonyl), Bzl (benzyl), Fmoc (9-fluorenylmcthoxycarbonyl), Mbh (4,4'-dimethoxydibenzhydryl), Mtr (4-methoxy-2,3,6-trimethylbenzene- sulphonyl), Trt (trityl), Tos (tosyl), Z (benzyloxycarbonyl) and Clz-Bzl (2,6- dichlorobenzyl) for the amino groups; N0 ₂ (nitro) and Pmc (2, 2, 5, 7, 8- pentamethylchromane-6-sulphonyl) for the guanidino groups; and tBu (t-butyl) for the hydroxyl groups.

Alternatively, the CD31 peptide may be synthesized using recombinant techniques.

The method of producing the peptide may optionally comprise the steps of purifying said peptide, chemically modifying said peptide, and/or formulating said peptide into a pharmaceutical composition.

According to a preferred embodiment, the peptide comprising an azide terminal group comprises a peptide of sequence KWPALFVR (SEQ ID NO: 6), wherein the amino acids are D-enantiomer amino acids.

Preferably, the initial peptide consists of sequence KWPALFVR (SEQ ID NO: 6), wherein the amino-acids are D-enantiomer amino acids, and the peptide comprising an azide terminal group thus comprises the sequence KWPALFVR (SEQ ID NO: 6) consisting of D-enantiomer amino acids and an azide terminal group.

A preferred peptide comprising an azide terminal group according to the invention consists of:

(i) a spacer at the N-terminus end, for example a spacer of sequence KGGG (SEQ ID NO: 95) consisting of D-enantiomers amino acids, and

(ii) the sequence KWPALFVR (SEQ ID NO: 6) consisting of D-enantiomers amino acids at the C-terminal end,

the N-terminus of the spacer being an azide group (i.e. presence of an azide group -N=N ⁺ =N instead of the amino group).

According to a preferred embodiment, the peptide comprising an azide terminal group comprises a peptide having the sequence KGGGKWPALFVR (SEQ ID NO: 96), wherein the amino-acids are D-enantiomers amino acids.

Preferably, the peptide comprising an azide terminal group has the sequence KGGGKWPALFVR (SEQ ID NO: 96) with an azide terminal group, said peptide consisting of D-enantiomer amino acids. A preferred peptide comprising an azide terminal group according to the invention has the sequence KGGGKWPALFVR (SEQ ID NO: 96), the N-terminus of which being an azide group (presence of an azide group -N=N ⁺ =N instead of the amino group) and said peptide consisting of D- enantiomer amino acids. For example, the peptide comprising an azide terminal group has the following formula:

According to an embodiment, the surface (or platform) of the substrate as mentioned above is made of metals or metal alloys, preferably a stainless steel, cobalt-chromium (CoCr) alloy, platinum-chromium (PtCr) alloy or a nickel-titanium alloy (such as Nitinol), or polymer-based (bioreseorbable) scaffold, such as those made of poly lactic/glycolic acid.

Preferably, the substrate is chosen from the group consisting of: endovascular stents, such as balloon-expandable stents (stents used for treating atherosclerotic disease of coronary and peripheral medium-sized arteries, covered stents comprising a sheet of fabric or synthetic material and used for aortic and large arteries); self-expanding stents (for use to treat stenotic disease of large arteries and for diverting the flow from saccular aneurysm, the later are called“flow diverting stents”), and percutaneous valved stents (stents used as a platform to allow the percutaneous implantation of endocardiac or endovascular bioprosthetic/synthetic valves).

According to a preferred embodiment, the present invention concerns a method for the modification of a substrate surface by grafting a peptide onto the surface of said substrate, wherein the substrate is a stent, preferably a coronary stent or a flow diverting stent, made of metals of metal alloys, in particular made of a cobalt-chromium (CoCr) alloy.

The present invention also relates to a modified surface substrate, wherein the surface of said substrate is grafted by a peptide, obtainable by the method as mentioned above. Preferably, this substrate has a modified surface comprising a coating made of a layer of polydopamine and a layer made of the linker and the peptide as defined above.

BRIEF DESCRIPTION OF THE SEQUENCES

SEQ ID NO: 1 corresponds to the sequence of human CD31.

SEQ ID NO: 2 corresponds to the sequence LAPWKK of a 6 amino acid peptide derived from human or murine CD31.

SEQ ID NO: 3 corresponds to the sequence VRVFLAPWKK of a 10 amino acid peptide derived from murine CD31 , also called PepReg CD31.

SEQ ID NO: 4 corresponds to the sequence VRVILAPWKK of a 10 amino acid peptide derived from human CD31.

SEQ ID NO: 5 corresponds to the sequence RVFLAPWK of a 8 amino acid peptide derived from murine CD31 , also called P8F.

SEQ ID NO: 6 corresponds to the sequence KWPALFVR of 8 amino acids and also correspond to the inverted sequence of SEQ ID NO: 5. When a peptide of sequence SEQ ID NO: 6 consists of D-amino acids, this peptide kwpalfvr is also called P8RI.

SEQ ID NO: 7 corresponds to the sequence RVILAPWK of a 8 amino acid peptide derived from human CD31.

SEQ ID NO: 8 corresponds to the inverted sequence of SEQ ID NO: 7.

SEQ ID NO: 9 corresponds to the sequence of murine CD31 .

SEQ ID NO: 10 corresponds to the sequence of bovine CD31 .

SEQ ID NO: 1 1 corresponds to the sequence of pig CD31 .

SEQ ID NO: 12 corresponds to the amino acids 579 to 601 of sequence SEQ ID NO: 1.

SEQ ID NO: 13 to 94 and 97 correspond to CD31 -derived sequences.

SEQ ID NO: 95 corresponds to sequence of the spacer KGGG.

SEQ ID NO: 96 corresponds to the sequence of a peptide comprising the spacer of sequence SEQ ID NO: 95 and the CD31 -derived peptide of sequence SEQ ID NO: 6. FIGURES

Figure 1. Elemental composition of PDA-linker-P8RI coated samples after ageing in PBS. A. Surface atomic percentages. B. N/C ratio of the samples. Error bars: standard deviation over 3 replicates.

Figure 2. Hemolysis ratios after 1 h (left) or 24h incubation (right). No significant difference appear between the negative control and the bare and coated disks, after 1 h or 24h of incubation.

Figure 3. Results of a multiplex assay analysis of cell culture supernatants from a representative experiment performed with HUVECs. Absolute concentrations (in pg/mL) are expressed. ^* : p<0.05. ^** : p<0.01. ^*** : p<0.001.

Figure 4. Results of multiplex assay analyses of cell culture supernatants. Each point is a biological replicate. For each analyte and each experiment, concentrations were normalized with respect to“CoCr” group.

Figure 5. SEM images of a“PDA-P8RI” stent in a porcine coronary artery. A. Global view of the internal surface of the half-artery. The stent struts are clearly visible. B. Higher magnification view of stent struts. The struts endothelialization is complete. C-D. A few round-shaped adherent leukocytes can be seen.

Figure 6. SEM images of a“Plasma-P8RI” stent in a porcine coronary artery.

A, B: lower magnification views. C, D: details of adherent cells.

Figure 7. SEM images of a drug eluting stent in a porcine coronary artery. A, B: lower magnification views. C, D: details of adherent cells.

Figure 8. SEM images of a bare metal stent in a porcine coronary artery. A, B: lower magnification views. C, D: details of adherent cells.

EXAMPLES

PREPARATION OF MODIFIED SUBSTRATES

1. Deposition of a polydopamine layer

The first step was the deposition of a self-assembled polydopamine layer on the L-605 CoCr samples.

The CoCr disks were made from mirror-polished L-605 sheets processed by the manufacturer (Goodfellow Cambridge Ltd). Thus, these CoCr disks were free of zinc contaminations, and presented a very flat surface.

The CoCr disks were ultrasound cleaned in three successive 10-minute baths of acetone, ethanol and deionized water. This cleaning procedure ensured the removal of organic contaminants. They were then put in an acid bath (40% HN0 ₃ ) for 40 minutes, in order to remove any ionic deposits from the alloy surface and to passivate it. After extensive washing with deionized water, the disks were sterilized by a 10-minute incubation in 70% ethanol. All subsequent steps were performed in sterile conditions, in a laminar flow hood. Solutions were filtered before use. The samples were rinsed several times with water, and, for the last wash, with the Tris buffer used for the dopamine solution.

The following coating protocol was adapted from the work by Lee and colleagues (Lee, H., Dellatore, S. M., Miller, W. M. and Messersmith, P. B. 2007. Science (New York, N. Y.) 318 (5849): 426-430).

A 2mg/mL dopamine solution was prepared just before use from lyophilized dopamine hydrochloride (Alfa Aesar A1 1 136) dissolved in a 10 mM Tris-HCI aqueous buffer with a pH adjusted to 8.5. The solution was added to the wells of a sterile multi-well plate containing the individual experimental samples. The plate containing the samples and the dopamine solution were thereafter covered with a sterile lid and incubated at room temperature for 22 ± 2h, under orbital shaking and protected from light. This incubation time resulted in a uniform coating, approximately 45 nm thick. The samples were then thoroughly rinsed with deionized water, placed in an ultrasound bath for 5 minutes in order to remove polydopamine aggregates from the surface, and further rinsed with deionized water. The chemical composition of the PDA (polydopamine) coating was analyzed by Fourier Transform Infrared (FTIR) spectroscopy.

Quantitative information on the elemental composition of the PDA coating was obtained by XPS. XPS spectra were recorded on bare and PDA-coated CoCr disks, using a PHI 5600-CI spectrometer (Physical Electronics). Survey spectra were acquired with a monochromatic aluminum X-ray source (300 W) whereas high- resolution spectra were recorded with a monochromatic magnesium X-ray source (300 W). The detection angle was set to 45° . The analysis was done on three spots per disk to assess the homogeneity of the coating. We analyzed the spectra with the software MultiPak (Physical Electronics).

The dopamine solution is originally clear and turns light pink and finally black, as the polymerization process takes place and polydopamine aggregates form. PDA-coated CoCr disks also exhibit a dark brown color.

In the FTIR spectra, several peaks and one band were found on the PDA spectrum and not on the bare CoCr spectrum, all attributed to vibrations of chemical bonds that are present in polydopamine. Those attributions are in agreement with the literature (Xi, Z.-Y., Xu, Y.-Y., Zhu, L.-P., Wang, Y. and Zhu, B.-K. 2009. Journal of Membrane Science 327 (1 ): 244-253; Yu, F., Chen, S., Chen, Y., Li, H., Yang, L., Chen, Y. and Yin, Y. 2010. Journal of Molecular Structure 982(1 ): 152-161 ; Zhou, Y., Weng, Y., Zhang, L., Jing, F., Huang, N. and Chen, J. 201 1. Applied Surface Science 258 (5): 1776-1783; Luo, R., Tang, L., Wang, J., Zhao, Y., Tu, Q., Weng, Y., Shen, R. and Huang, N. 2013. Colloids and Surfaces B-Biointerfaces 106: 66-73) and evidence the successful deposition of a polydopamine coating on the disks.

The below table shows the average surface atomic percentages of bare and PDS-coated CoCr disks (standard deviation between three replicates):

On the bare samples (CoCr), the main elements that compose the L-605 alloy were detected (chromium, cobalt, tungsten and nickel). The presence of oxygen and carbon can be explained by the presence of metal oxides and (to a lesser extent) carbides, as well as residual surface organic contaminants.

On the PDA-coated disks (PDA), no metallic element was detected. That indicates that the PDA coating covered the surface in a uniform and complete manner. The only elements that were detected were oxygen, carbon and nitrogen, which compose the dopamine molecule. They were found in proportions close to those of the dopamine molecule (18% oxygen, 73% carbon, 9% nitrogen). These results therefore confirm the results obtained with the FTIR characterization.

3. Preparation of the peptide P8RI A CD31 -derived peptide having an 8-residue long sequence, included in the original 23-residue long peptide, was chosen because it is highly conserved among species and because it showed high specificity towards CD31 in in vitro assays. The CD31 peptide was designed with D-amino acids. Thus, the peptidic sequence was assembled in reverse order so as to retain the original spatial orientation and the chirality of the side chains. The resulting peptide was named‘P8RI’, where P stands for peptide, 8 is the number of amino acids, and Rl means retro-inverso. The sequence of P8RI is:

(D- Lys> (D-Trp) - (D- Pro) - (D - Ala) - (D - Leu] - (D- Ph e) - (D - Vat) - (D - Arg).

The structure of P8RI is shown below, along with that of the original, non- inverted sequence (called P8F).

’P8F’ designates the original sequence from the CD31 molecule, in forward sense, whereas P8RI is the retro-inverso sequence, where both the order and the chirality of the residues are inversed.

4. Immobilization of the peptide

In order to better control the orientation of the immobilized P8RI and to render it readily accessible to its ligands at the surface of the cells contacting the coated surface, the inventors used a flexible linker as an intermediary between the peptide and the PDA coating.

The linker needs to have: 1 ) either an amine or a thiol function at one extremity, to be able to react with the o-quinone functions of the polydopamine coating, 2) a flexible chain to improve the accessibility of the bound P8RI and 3) a function at the second extremity that would specifically bind to the peptide N- terminus. As regards that last point, the inventors opted for a bioorthogonal reaction so as to avoid any interference from the side chains of the peptide and the amine functions of the polydopamine.

Copper-free click chemistry was used as it allows for a fast reaction in aqueous solution, without the addition of cytotoxic catalysts such as copper.

The strain-promoted azide-alkyne cycloaddition (SPAAC) has been carried out:

This addition has been carried out with an azide on the peptide and a bicyclo[6.1 .0]nonyne (BCN) on the linker.

A“P8RI azide” was custom-synthesized, using a modified lysine with an azide instead of the side chain amine was introduced at the N-terminus of the peptide, and separated from the P8RI sequence by three glycines. The sequence of the“P8RI azide” was therefore:

Lys(N ₃ )-Gly-Gly-Gly-(D-Lys)-(D-Trp)-(D-Pro)-(D-Ala)-(D-Leu)- (D-Phe)-(D-Val)-(D-Arg).

Concerning the linker, we opted for a short PEG chain between the BCN at one end and an amine at the other end. PEG was chosen for its high flexibility and biocompatibility. The resulting linker (Sigma 745073) will be called“BCN-amine” throughout this thesis.

The two-step immobilization of the P8RI azide on the PDA coating through the BCN- amine linker is as follows:

A. Binding of the BCN-amine linker to the PDA coating by Schiff base reaction (left) or Michael addition (right).

B. Binding of the peptide to the immobilized BCN-amine linker by click chemistry. 5. Immobilization of P8RI azide on PDA coatings with the help of the BCN-amine linker

BCN-amine binding was carried out on CoCr disks immediately after the polydopamine coating procedure, in a laminar flow hood. A solution containing 0.1 mg/ml_ BCN-amine (Sigma 745073) diluted in Tris buffer (10 mM Tris, pH adjusted to 8.5) was added to the wells containing the PDA-coated samples. They were incubated for 22 ± 2h under orbital shaking. After thorough rinsing with deionized water, a solution containing 0.2 mg/ml_“P8RI azide” in deionized water was added to the samples. The samples were thenthoroughly rinsed with deionized water.

Fluorescence microscopy and XPS analysis of untagged peptide immobilized on the surface were performed.

For the XPS analysis, unmodified “P8RI azide” was immobilized on CoCr disks as described above. The coated samples were then analyzed by XPS as detailed above. Three points were analyzed on each sample.

For the fluorescence microscopy analysis, a“P8RI azide” with an additional C- terminal lysine conjugated to a FITC fluorophore was custom synthesized by the manufacturer Proteogenix. The pseudo-peptidic sequence of that“P8RI-FITC azide” thus was:

Lys(N ₃ )-Gly-Gly-Gly-(D-Lys)-(D-Trp)-(D-Pro)-(D-Ala)-(D-Leu)- (D-Phe)-(D-Val)-

(D-Arg)-Lys(FITC).

The“P8RI-FITC azide” was immobilized on a BCN-amine linker on PDA- coated CoCr disks as described for the“P8RI azide”. The coated samples were then placed face down in imaging dishes to prepare their observation through an inverted microscope and covered with mounting medium (ProLong Gold Antifade Mountant, Thermo Fisher P36930). This medium has a refractive index close to that of the material through which the samples are observed (glass or plastic), so that light transmission is optimized. The ProLong Gold mounting medium also has antifading properties that minimize photobleaching, thus allowing the preservation of fluorescent samples for longer times. Digital photographs of the samples were acquired on an Axio Observer inverted fluorescence microscope (Zeiss), equipped with the software Zen (Zeiss).

CoCr disks on which the“P8RI-FITC azide” had been immobilized (following the protocol described above) emitted a green fluorescence of much higher intensity than the PDA-coated disks. These results prove the presence of the fluorescent peptide at the surface of the disks after the immobilization protocol. The XPS analysis of the surface of coated CoCr disks revealed only the presence of the three elements nitrogen, oxygen and carbon on the three types of coating: PDA only, PDA + BCN-amine linker, PDA + linker + P8RI azide. This was consistent with the previous analyses of PDA coatings, which had appeared to be homogeneous. Atomic ratios of nitrogen and oxygen over carbon were calculated from the XPS spectra. They are presented in the below table, along with the theoretical values of these ratios for the molecules involved in the coatings:

The O/C ratio of the polydopamine samples is slightly higher, and the N/C ratio slightly lower, than the values of the pure dopamine molecule. The evolution of the ratios with the addition of the linker and the P8RI to the coatings follows the tendency of the theoretical values: a constant decrease of the O/C ratio, and a slight decrease followed by a larger increase of the N/C ratio. Therefore, these results point to the successful immobilization of the BCN-amine linker and of the P8RI azide.

In order to have its full effect in vivo, the“PDA-linker-P8RI” coating needs to maintain its integrity for the time necessary for stent endothelialization (about one week). The ageing behavior of the coating was assessed by an in vitro static ageing test in liquid medium. PBS was chosen as the liquid medium. Each CoCr disk was fitted in a custom sample holder designed to expose only its coated surface to PBS. The floating sample holder was placed in a beaker filled with PBS and stored in an incubator at 37°C (human core body temperature) for 1 or 4 weeks. Before the test, the beakers and sample holders were sterilized by autoclave to prevent bacterial proliferation during the study. The samples were manipulated in sterile conditions, under a laminar flow hood. After ageing, the disks were removed from their holders, thoroughly rinsed with deionized water, and dried with medical- grade compressed air. They were then analyzed by XPS. Three coated disks and one bare CoCr disk were tested for each time point (1 or 4 weeks).

The surface atomic percentages of the coatings measured by XPS are presented on Figure 1A. The N/C ratio was considered as the most reliable indicator of the potential degradation of the coating since it is not influenced by oxidation or water adsorption on the coating. As shown on Figure 2B, the N/C ratio of the“PDA- linker-P8RI” coating was unmodified after 4 weeks of ageing in PBS at 37 °C, thus demonstrating that the coating did not undergo any significant degradation.

IN VITRO EVALUATION OF THE BIOCOMPATIBILITY OF P8RI-COATED

SURFACES

In the subsequent paragraphs, ‘CoCr’ designates bare L-605 CoCr disks, ‘PDA’ means polydopamine-coated CoCr disks, and‘P8RI’ are CoCr disks with a polydopamine-linker-P8RI coating.

The concentration of P8RI used during the last step of the polydopamine- linker-P8RI coating process was chosen on the basis of a dose-effect curve obtained using 4-fold dilutions between 6 and 200 pg/rnL. In preliminary experiments, the concentration of 50 pg/rnL yielded the least pro-inflammatory (as detected by the production of soluble IL-6 and VCAM-1 ) and the most anti- thrombotic (based on the levels of soluble TFPI) phenotype of primary human endothelial cells cultured on the coated surfaces The concentration of 50 pg/mL was therefore used for coating the CoCr samples in the in vitro experiments presented in the following paragraphs, and the stents in the preclinical studies as explained hereafter.

1. Hemocompatibility

1.1. No hemolysis

The hemolysis assay is a required biocompatibility test for all blood-contacting devices. It shows whether a given material causes erythrocyte lysis (either by contact or by the release of toxic molecules). The hemolysis assay protocol we used was adapted from the one used by Bae and coworkers (Bae, l.-H., Park, l.-K., Park, D. S., Lee, H. and Jeong, M. H. 2012. Journal of Materials Science: Materials in Medicine 23 (5): 1259-1269.) as detailed below.

Coated and bare CoCr disks were placed in the wells of a 96-well plate and sterilized by a ten minute incubation in 70° ethand. The disks were rinsed several times in water prior to a final wash in physiological saline solution (0.9% NaCI). Human peripheral whole blood collected in lithium heparin (18 UI/mL) was then gently layered on each disk and the plate was incubated at 37°C during either 1 h or 24h. Heparin was chosen for anticoagulation because it does not interfere with the hemolysis assays.

Disk-free wells were used as negative controls and 1% Triton X-100 (which causes the lysis of most erythrocytes through the dissolution of their plasma cell membrane) was added in blood- containing positive control wells. At the end of the incubation period, the blood was transferred from each well to an individual polypropylene tube (Eppendorf, 1 .5ml) and centrifuged at 1200 g for 15 minutes. 50 pL of platelet-poor plasma were collected at the top of all centrifuged tubes and transferred to a new 96-well plate. Absorbance was measured at 540 nm in a plate reader spectrophotometer (Infinite 200 PRO, TECAN). 540 nm corresponds to the absorbance peak of free hemoglobin, which is released by hemolysis from the erythrocytes. The value of the absorbance at 540 nm is therefore directly proportional to the extent of hemolysis caused by the sample disks in the experimental wells. The experiment was performed in technical quadruplicates and repeated three times using the blood of different healthy blood donors.

According to the photograph of the microplate taken just before the measurement of the absorbance at the end of a hemolysis assay, the platelet-poor plasma from whole blood incubated with Triton X-100 appears dark red, reflecting the effective hemolysis in these wells (positive controls). No macroscopic difference in color is apparent between wells containing the negative control (No disk) and the three experimental groups of plasma.

For each experiment, these observations were quantified by absorbance reading at 540 nm, and the measurements were normalized using the values obtained with the positive and negative controls, according to the following formula: wherein R(x) is the normalized ratio of group x, and A ₅₄ o _nm (x) is the average absorbance of the wells from group x. The measured ratios are presented on Figure 2.

In the three experimental groups (blood incubated with the CoCr disks), after 1 h or 24h or incubation, the hemolysis ratios were not significantly different from the negative controls.

Therefore, it can be concluded that neither the bare CoCr surface nor the coated CoCr surfaces that were used were evidenced to cause hemolysis when contacted with human whole blood.

1.2. Platelet adhesion

As thrombosis is one of the two main complications associated with stenting, evaluating the thrombogenicity of each novel stenting material is essential. To this aim, we performed a platelet adhesion assay.

Coated and bare CoCr disks were incubated in 70° etianol and rinsed, as detailed above. Human whole blood collected in PPack (75 mM PPACK + 0.1% D- mannitol, Haemtech SCAT- 875B) was then deposited on the disks and they were incubated at 37°C for 1 h. PPack (Phe-Pro-Arg-chloronethylketone) is a peptidomimetic thrombin inhibitor that inhibits the coagulation cascade without affecting the physiological concentration of ionized calcium. This is important in a functional test such as platelet adhesion, which is a Ca++ dependent process. At the end of the incubation period, the disks were taken out of the wells using delicate tweezers and rinsed by gently stirring in a beaker full of physiological saline solution at room temperature. They were then placed in a new 96-well microplate and rinsed twice in physiological saline solution, prior to fixation with paraformaldehyde. The disks were then processed for immunocytofluorescence.

In order to visualize all adherent platelets, we chose to immunostain two antigens that are constitutively expressed by platelets: CD41 and von Willebrand Factor (vWF). CD41 , also known as Integrin alpha-lib, is the most abundant platelet adhesion receptor and is thus present on the surface of the platelets. vWF is a glycoprotein which plays a major role in blood coagulation. As it is stored in intracellular compartments of platelets (the a-granules), its immunostaining requires the permeabilization of the platelet membrane. Since this“platelet adhesion assay” was performed with whole blood, leukocyte adhesion was also possible. Hoechst staining of cell nuclei was used to identify leukocytes (which, contrary to platelets, possess a cell nucleus).

The reagents used in the following protocol are presented on the below table.

The experimental samples were fixed in 4% paraformaldehyde at 4°C for 10 minutes, then rinsed 3 times in Dulbecco’s phosphate buffered saline (PBS). Fixation protects biological samples from decay by cross-linking the proteins. The samples were then permeabilized by a 10-minute incubation in a solution containing 100 mM glycine and 0.5% Triton X-100 in PBS. Triton X-100, being a non-ionic detergent, creates pores in the cell membranes without denaturing proteins, while glycine was used to quench the formaldehyde. After PBS rinsing, blocking was then performed by a 30- minute incubation in a solution of 5% bovine serum albumin and 0.1% fish gelatin in PBS, in order to reduce unspecific antibody binding, and thus decrease background noise.

The two primary antibodies were diluted in a solution containing 1% bovine serum albumin (BSA) and 0.02% fish gelatin in PBS, and incubated overnight at 4°C with the samples. The samples were then rinsed with PBS and the secondary antibodies, diluted in a similar fashion, were added and incubated for 1 hour at room temperature. After PBS rinsing, the nuclei of the cells were stained by incubation in Hoechst solution. Finally, the samples were placed face down in imaging dishes (to prepare their observation by inverted microscope) and covered with Prolong Gold mounting medium. Digital photographs of the immunostained samples were then acquired on an Axio Observer inverted fluorescence microscope (Zeiss), equipped with the software Zen (Zeiss). The surface of the disks covered by platelets was identified by positive CD41 and vWF staining and quantified on the digital images using the“Analyze particles” function of the open source software Fiji.

The observation of the resulting photographs showed that very few leukocytes had adhered to the surface of the samples (where they were easily identified by their Hoechst positive nuclei). Globally, the surface density of platelets (identified by positive CD41 and vWF staining) appeared lower on PDA than on bare CoCr, and even lower on “P8RI” disks as compared to the “PDA” ones. However, high variability can be noted between different disks of the same group, and this was reflected by the values obtained by computer-assisted quantification. Therefore, the significance of the differences observed between the experimental groups is questionable. The experiment was repeated several times, but, as intra-group variability was always high, no significant difference in the adhered platelets density was found between the three groups.

It can be concluded from these experiments that the “PDA” and “P8RI” coatings do not increase but rather tend to reduce platelet adhesion on the surface of CoCr samples. 2. Endothelialization

A functional evaluation of coronary artery endothelial cells upon their contact with the experimental (bare and coated) CoCr disks described before was performed. Attachment and survival were evaluated by computer-assisted analysis of digital images after immunofluorescent staining whereas the pro-inflammatory and pro-thrombotic phenotypes were assessed through quantitative analysis of soluble biomarkers in cell culture supernatants.

For practical reasons, the first experiments were performed with human umbilical vein endothelial cells (HUVECs, from PromoCell). Human coronary artery endothelial cells from three different individual donors (HCAECs, purchased from Lonza) were then used, in order to better reproduce the environment of coronary stents. The cells were cultured in Endothelial cells Growth Medium MV2 (Promocell), which contained the nutrients and growth factors needed by ECs. Antibiotic, antifungal and antimycoplasma reagents (penicillin, streptomycin, amphotericin B, plasmocin and primocine) were added to the medium in order to prevent contaminations. The cells were used at passages 3 to 5.

Coated and uncoated CoCr disks were placed in the wells of a 96-well plate and sterilized as described above. 100 000 endothelial cells suspended in their growth medium were seeded on each disk. After a 48h incubation at 37 °C and 5% C02, the cell culture supernatants were collected for multiplex assay analysis, while the adherent cells were processed for immunocytofluorescence.

We chose to immunostain CD31 and VE-Cadherin, because these proteins clusters together at the adherens junction in functional endothelial monolayers. CD31 is known to be truncated and miss the first, most membrane-distal, extracellular domains on stressed endothelial cells. Thus, to detect intact (functional) CD31 molecules we used the mouse monoclonal antibody JC70A (Dako, #M0823) as this antibody typically fails to stain cells that express a truncated CD31 .

The staining protocol was the same as described in the previous paragraphs for adherent platelets, except that the permeabilization step was not performed, and that the following antibodies were used:

- primary antibody: mouse anti-human CD31 (10 pg/rnL, Dako M0823)

- secondary antibody: goat anti-mouse IgG, A488 (5 pg/rnL, Invitrogen A1 1029). The measure of pro-inflammatory and pro-coagulant endothelial biomarkers in the cell culture supernatants was achieved using a custom multiplexed cytometric bead assay (Luminex technology). Each bead contains two different types of fluorophores. This makes it possible to detect several types of beads in the same well: each type of bead is identified by different fluorescence intensities at two emission wavelengths. Each type of capture antibody is associated with one type of bead. The quantity of captured antigen on each bead is measured by the fluorescence intensity of the phycoerythrin (PE) bound to detection antibodies. Thus, the concentration of several antigens can be determined in the same well.

In order to analyze the influence of “PDA” and “P8RI” coatings on the phenotype of cultivated endothelial cells, different types of soluble proteins produced by activated endothelial cells were quantified. A range of soluble molecules considered as markers of EC pro-inflammatory and pro-thrombotic phenotype were quantified in preliminary experiments.

IL-6, a pro-inflammatory cytokine involved in lymphocyte growth and differentiation, acute phase reaction and fever, and IL-8, a chemokine that induces granulocytes migration and phagocytosis, but also promotes angiogenesis, are both known to be released by endothelial cells in response to inflammatory stimuli, such as LPS and TNFa. They were therefore selected as markers of EC pro-inflammatory phenotype. CD62E and VCAM-1 (Vascular Cell Adhesion Molecule-1 ) are transmembrane glycoproteins of the CAM (cell adhesion molecules) family and are responsible for leukocyte adhesion on ECs. Thus, they are not primarily expressed as soluble molecules. However, cytokine activated endothelial cells are known not only to exhibit increased surface expression of CD62E and VCAM-1 , but also to produce soluble forms of these proteins, as a result of unidentified cleavage of shedding processes. For this reason, CD62E and VCAM-1 were also chosen as markers of EC pro-inflammatory phenotype. The pro-thrombotic phenotype of the ECs was assessed by the quantification of PAI-1 (Plasminogen Activator Inhibitor- 1 ), a serine protease inhibitor directed against tissue plasminogen activator (tPA) and urokinase (uPA). Since tPA and uPA activate plasminogen and are therefore the main initiators of fibrinolysis, the action of PAI-1 is pro-thrombotic. These five proteins were quantified in cell culture supernatants in several experiments, performed with cells from different donors. Each experiment included eight technical replicates. The endothelial cell culture supernatants were centrifuged at high speed in order to remove any dead cell and debris, then they were transferred to polypropylene 96-well plates, sealed and stored at - 80° C until the day of the assay. They were then thawed and diluted in staining buffer. The optimal concentration was determined according to previous experiments. The diluted supernatants were transferred to a flat-bottom 96-well plate containing a mix of the capture beads. A standard range was added to the same plate by serial dilution of standard solutions (Bio- Rad), containing a known concentration of each antigen. The plate was incubated on an orbital shaker for 1 h, in order for the capture antibodies to bind their antigens. The plate was then rinsed with wash buffer on an automatic wash station, which rinses the wells while holding the magnetic beads at the bottom of the wells. The biotinylated detection antibodies, diluted according to the manufacturer’s instructions, were then incubated with the beads for 30 minutes. Finally, after rinsing, streptavidin-PE was added to each well and incubated for 10 minutes, so that the streptavidin could bind to the detection antibody’s biotin. After a last rinsing step, the beads were resuspended in assay buffer and analyzed on an assay reader (Bio-Plex 200, Bio-Rad). The statistical analysis of the results was performed with Kruskal-Wallis test. This nonparametric test was chosen as the number of replicates was too low to assess their Gaussian distribution, and hence to fulfill the conditions for an ANOVA test. Differences were considered significant for p<0.05.

The“PDA + P8RI” coating favors rapid and functional endothelialization of CoCr surfaces

In view of representative images of immunostained endothelial cells, the major difference between the three groups (CoCr, PDA and P8RI) is that the adhesion of endothelial cells appears much lesser on bare CoCr surfaces than on“PDA” and “P8RI” surfaces, since the ECs detached only from the bare surfaces. The“PDA” and“P8RI” coatings on experimental CoCr disks both allowed the formation of a functional confluent layer of endothelial cells, as evidenced by the distribution of CD31 at intercellular junctions.

The“PDA + P8RI” coating tends to induce an anti-inflammatory and antithrombotic phenotype in endothelial cells

As can be seen on Figure 3, the variability between the technical replicates of the multiplex assays was generally low, and clear differences between groups (cells cultivated on bare CoCr, “PDA” or “P8RI” surfaces) were found within one experiment. In particular, a tendency consistently appeared in IL-6, IL-8, CD62E and VCAM-1 : concentrations on“PDA” were lower than on“CoCr”, and concentrations on “P8RI” lower than on “PDA”. This experiment thus suggested an anti inflammatory effect of the PDA coating compared to bare CoCr, and a further effect of P8RI immobilization.

The synthesis of the results obtained from all cell culture supernatant analyses is presented on Figure 4. The multiplex assay was performed on five biological replicates (endothelial cells from five different donors). Each point on the graphs of Figure 4 represents the average of the technical replicates for one experiment. As in the experiment presented on Figure 3, the concentrations of IL-6, IL-8, CD62E and VCAM-1 are lower in the“PDA” and“P8RI” groups than in the“CoCr” group. An additional effect of P8RI immobilization is observed on IL-8, but also on PAI-1 concentrations, suggesting a slight anti-thrombotic effect of the P8RI coating. However, given the low number of replicates and high variability between experiments, the differences did not reach the statistical significance according to the Kruskal-Wallis test.

Therefore, the apparent anti-thrombotic and anti-inflammatory effects of“PDA” and“P8RI” coatings can only be expressed in terms of tendency.

IN VIVO EVALUATION OF THE BIOCOMPATIBILITY OF P8RI-COATED

SURFACES

Bare and“PDA+P8RI”-coated coronary stents and FDSs were implanted in animal models in order to assess the effect of the coating on the stents’ in vivo environment. Large animals (pigs) were used for the implantation of coronary stents, since the arteries of these animals are large enough to be compatible with the size of stents designed for human use.

The coating strategy which was developed has been applied to commercially available stents. The biocompatibility of the“PDA+P8RI” coating could hence be evaluated also in vivo, by comparing the performance of the“PDA+P8RI”-coated stents with that of the parent bare metal or active device, made of the same alloy.

The Multilink BMS, made of CoCr, was suitable for both P8RI immobilization strategies developed during this study (the PDA-P8RI coating and the direct immobilization of P8RI on plasma-functionalized CoCr). Thus, BMSs (Multilink, Abbott) and DESs (Xience) were compared to “Plasma-P8RI” and “PDA-P8RI” coronary stents.

P8RI-coated and control coronary stents were implanted in the coronary arteries of farm pigs after creation of elastase-induced saccular aneurysms. Due to the size of the target arteries in these animals, which are large enough to be compatible with the size of stents designed for human use, these models are widely used for the issue of in vivo stent biocompatibility.

Implantation of coronary stents in farm pigs

The primary objective of this first of in vivo experiments was the evaluation of the effect of surface-immobilized P8RI on early stent strut endothelialization and on local inflammation. For this reason, the experimental and control stents were implanted in the three coronary arteries of healthy animals and the stented arteries were evaluated 7 days after stent implantation. According to a study by Perez de Prado and colleagues (Perez de Prado, A., Perez-Martinez, C., Cuellas-Ramon, C., Gonzalo-Orden, J. M., Regueiro-Purrinos, M., Martinez-Fernandez, B., Diego-Nieto, A. and Fernandez- Vazquez, F. 201 1 . Revista Espa oia de Cardiologia (English Edition) 64 (2): 159-162), this period is sufficient for the complete endothelialization of BMSs, but not DESs. Thus, it was deemed to be suitable for the comparison of the performances of BMSs, DESs, and P8RI-coated stents in vivo.

Four groups of Multilink (Abbott) stents, with different coatings, were implanted in the coronary arteries of 9 male farm pigs: bare metal stents (unmodified), everolimus-eluting stents (Xience, Abbott),“PDA-P8RI” stents (that had been coated with the PDA-based strategy) and“Plasma-P8RI” stents, on which P8RI had been immobilized on a PEG linker after plasma amination of the oxide surface.

The four groups of stents were sterilized by beta radiations (25 kGy) by lonisos (Chaumesnil, France). This sterilization technique was chosen because it was reported not to induce changes in irradiated proteins, contrary to the ethylene oxide sterilization technique that is commonly uses for the sterilization of commercial stents.

The protocol for the study of the coronary stents in farm pigs was in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health 1985) and the European Directive 2010/63/EU, and was approved by the local ethic committee (Agence nationale de securite sanitaire de I'alimentation, de I'environnement et du travail, 14 rue Pierre et Marie Curie, 94701 Maisons-Alfort Cedex, n° APAFiS: 20170321 16276884). The pigs were housed at the Centre de Recherche Biomedicale (CRBM) at the veterinary school of Alfort.

After a“pre-load” with ticagrelor (180mg) and aspirin (75mg) and an overnight fast, the animals were sedated, intubated (endotracheal intubation) and connected to a ventilator; anesthesia was maintained with gas anesthetics throughout the procedure. After administration of heparin (intravenous bolus, 5000 Ul) and antibiotic prophylaxis, arterial access was obtained under sterile conditions by femoral or carotid artery cutdown. Thereafter, coronary angiography was performed to select the part of the coronary tree in which to leave the implant as previously described. Briefly, on the basis of the angiograms, at least one segment, with a diameter of 2.5-3.0 mm, was selected (except when the anatomical characteristics of the animal did not allow it) in each of the three epicardial coronary arteries (left anterior interventricular, left circumflex artery and right coronary artery). Thereafter, an angioplasty catheter with the stent crimped on its deflated balloon was advanced to the selected site for implantation over a standard guidewire. The balloon was inflated to a maximal pressure of 8 atm for 30 seconds, deflated, and slowly withdrawn, leaving the stent in place. A total of 5 to 9 stents per group were implanted, in 2 to 3 animals per group. After repeated angiography of the stented coronary arteries to confirm patency, the arteriotomy was repaired, the skin closed, and the animals allowed to recover from anesthesia. Aspirin (75mg/day) and ticagrelor (90mg, twice a day) were pursued until the termination of the study.

The stents were implanted under angiographic monitoring. Seven days after stent implantation, the animals were terminated under anesthesia. The stented arteries were explanted, thoroughly rinsed with PBS and fixed in 2.5% glutaraldehyde at 4°C for at least 48h. Before being processed for scanning electron microscope (SEM) observation, they were cut longitudinally into halves. As electron microscopes operate under vacuum, samples dehydration is indispensable. Just before imaging, they were washed in deionized water and dehydrated in ethanol baths of increasing concentration. They were then coated by gold sputtering. Metallization is a necessary step for biological samples imaging by SEM, as this technique requires an electrically conductive surface in order to avoid the accumulation of electrons on the surface. The metallized samples were then imaged with a scanning electron microscope in secondary electrons imaging mode. These experiments were performed very recently, and the quantitative analysis of the endothelialization on the stented arteries is currently ongoing. Figure 5 shows a representative example of SEM imaging of the luminal side of a circumflex artery implanted with a“PDA+P8RI”-coated stent. The degree of stent strut endothelialization can be evaluated on pictures taken at intermediate magnification (see Figure 5B), and adherent leukocytes can be identified on high magnification images. On this “PDA+P8RI”-coated stent, the endothelialization appears complete, the density of adherent leukocytes is low and their appearance is spheric (non activated), and there are no thrombi. Similar features can be observed on the Plasma-P8RI stent shown in Figure 6.

At variance, the stent struts were so weakly covered by the endothelium in the case of the artery implanted with the DES that the metal structure readily got off the arterial wall at the time of SEM examination, as shown in Figure 7. Furthermore, the morphology of the endothelial cells was altered with loose junctions and the appearance of several microvesicles at their luminal side. The shape of the adherent leukocytes was spheric (non activated), in agreement with the immunosuppressive effect of the eluted drug (everolimus), but platelets and fibrin filaments were readily visible in these specimen (same figure), in spite of the double anti-platelet therapy administered to the pigs during the study period.

The endothelialization was not an issue with BMSs, as shown in Figure 8, but there again striking differences could be detected in terms of platelet/fibrin deposition as compared to P8RI-coated stents. Moreover, the shape of the leukocytes was granular and distorted by the presence of several pseudopods, suggesting an activated phenotype of the adherent leukocytes on these stents.