CA2154166A1

CA2154166A1 - Nonthrombogenic implant surfaces

Info

Publication number: CA2154166A1
Application number: CA002154166A
Authority: CA
Inventors: Roger E. Marchant; Shengmei Yuan; Gyongyl Szakalas-Gratzl
Original assignee: Individual
Current assignee: Case Western Reserve University
Priority date: 1992-11-19
Filing date: 1993-11-18
Publication date: 1994-05-26
Also published as: US5741852A; EP0623150A1; AU5613094A; AU677028B2; WO1994011411A1; US5993890A; EP0623150A4

Abstract

Triblock polymers comprised of polysaccharide, such as heparin or dextran, and a hydrocarbon chain, have been pre-pared. The triblock polymer adsorbs strongly on the surface of hydrophobic polymer substrates such as polyethylene, through hy-drophobic interaction between the polymer and the hydrophobic hydrocarbon chain of the triblock polymer. The surface ad-sorbed with triblock polymer is resistant to protein deposition, which renders the surface nonthrombogenic.

Description

NONTHROMBOGENIC IMPLANT SURFACES

Backqround of the invention The use of synthetic biomaterials to sustain, augment or completely replace diseased human organs has increased tremendously over the past thirty years. Synthetic implants have cardiovascular applications such as vascular grafts, heart valves, and ventricular assist devices;
extra~oL~G~eal systems; and a wide range of invasive treatment and diagnostic systems. Unfortunately, existing biomaterials suffer from well-known problems associated with surface-induced thrombosis or clot formation such as thrombotic occlusion and thromboemboli, and infection.
Synthetic vascular grafts having a diameter less than 6 mm are ~L~e,.~ly impracticable, because of potential thrombotic occlusion, and the artificial heart has been plagued with problems of thromboemboli and infection.
Advances in the development of artificial organs and artificial vascular grafts have resulted in the need for nonthrombogenic materials.
Thrombosis is initiated by the deposition of a plasma protein layer on the surface of the implanted biomaterial.
Thereafter, platelets, fibrin, and possibly leukocytes, adhere to the deposited protein. The interactions between the plasma proteins and the surface of the implant determine the adhesion, the activation and the spreading of platelets, the activation of coagulation, cell attachment and protein deposition. However, at the molecular level, the fundamental forces and interactions of plasma proteins with implants is not well understood.
There have been several attempts to create nonthrombogenic surfaces on polymer implants thereby increasing the blood-biocompatibility of implants.

2~ PCT/US93/11210 ~ar~y a~temPts included precoating the implants with proteins no~ involved in thrombosis, such as albumin, to mask the thrombogenic surface of the implant. However, such implants loose their nonthrombogenic properties within a short time. Attempts have been made to mask the thrombogenic surface by coating gelatin onto implants such as ventricular assist devices. While the gelatin coating reduced the thrombus formation, it did not adhere to the implant and it did not prevent thromboemboli and infection.
Attempts have been made to render implants nonthrombogenic by coating the surface of the implant with polyethylene oxide to mask the thrombogenic surface of the implant; it was discovered that such a coating at times also reduced protein adsorption. While this reduced thrombogenesis, the coupling of polyethylene oxide to the surface of the implant involves very complex procedures, and the coated implants do not consistently exhibit protein resistance.
There have been many attempts to prepare nonthrombogenic surfaces by attaching heparin to biomaterials, because of heparin's potent anticoagulant properties. However, each method requires that the implant surface be first modified by attachment of a coupling molecule before heparin can be attached. For example, the positively charged coupling agent tridodecylmethylammonium chloride, is coated onto an implant, which provides a positively charged surface and allows heparin which has a high negative charge density, to be attached. However, the heparin slowly dissociates from the surface, to expose the positively charged, TDMAC surface which is particularly thrombogenic. The TDMAC attracts platelets and other cells; cells surfaces have a high negative charge density.
Thus the TDMAC heparin coated implant is s~1cceccful only for short term implants such as catheters.
Implants coated with heparin coupled to coupling molecules typically have limited anti-thrombogenic effectiveness because commercial heparin preparations .
contain the protein core and because many heparin molecules which having no anticoagulant activity. As a result, the surfaces soon become covered by adsorbing protein on exposure to blood, thus neutralizing the anticoagulant activity of the active heparin molecule.
It is desirable to have implants which resist plasma protein deposition, and to have a simple procedure for modifying the surface of implants. Nonthrombogenic implants would reduce the need for aggressive anticoagulant therapy, improve the performance of implants, particularly cardiovascular prosthetic devices, and encourage the development of devices not currently feasible.

SummarY of the Invention The present invention provides a triblock polymer which may be easily applied to the surface of a hydrophobic substrate, such as an implant, to provide the substrate with resistance to the deposition of plasma proteins thereby preventing the first step of thrombus formulation.
The triblock polymer contains two hydrophilic segments which are joined via a hydrophobic segment. The h~dL ~hobic segment is a hydrocarbon chain which h~d~o~hobically interacts with the surface of the hydrophobic substrate to provide a means of attaching the triblock polymer to the surface of the substrate. The hydrophilic segments are oligosaccharides or polysaccharides, such as, for example dextran, dextran sulfate, dermatan sulfate, heparin or portions of heparin.
The hydrophobic substrates include biomaterials that are hydrophobic, for example, polyethylene, polypropylene, silicone rubber, Impra0, Gortex0 and Teflon0, and hydrophobic medical polyure~hAnec such as Pellethanes~.
The triblock polymer is easily applied to the substrate. Since the free triblock polymer is water soluble, the triblock polymer is dissolved in water and the implant is then immersed in the a~ueous solution of the W094/11411 21~ ~16 6 4 PCT/US93/11210 triblock polymer for about 24 hours. The triblock polymer spontaneously attaches to the polymeric substrate to provide a protein-resistant, nonthrombogenic surface.

Brief Description of the Drawinqs Figure l is a representation of triblock polymer on a substrate.
Figure 2 is the FTIR spectrum the dextran triblock polymer.
Figure 3A is a GPC chromatogram molecular weight distribution obtained by GPC of dextran on Sephadex G-75 gel chromatography column followed by measurement of the uronic acid content. GPC elution solvent was 20mM Tris +
50mM NaCl (pH=7.0) with flow rate of 2.0 mltminute.
Figure 3B is the molecular weight distribution obt~in~ by GPC of the dextran triblock polymer on Sephadex G-7S gel chromatography column followed by measurement of the uronic acid content. GPC elution solvent was 20mM Tris + 50mM NaCl (pH=7.0) with flow rate of 2.0 ml/minute.
Figure 4A is the 13C-NMR spectra of dextran.
Figure 4B is the l3C-NMR s~e~L~a of the intermediate dextran product.
Figure 4C is the l3C-NMR spectra of the dextran triblock polymer.
Figure 5 is the FTIR/ Attenuated Total Reflectance (ATR) ~e~LLa (1550-900 cm~l region) of: (a) unmodified PE;
(b) PE exposed to dextran solution; (c) dextran triblock polymer adsorbed on PE.
Figure 5A is the FTIR/ Attenuated Total Reflectance (ATR) spectra (1700-900 cm~l region) of: (a) unmodified PE;
(b) PE exposed to heparin solution; (c) heparin triblock polymer adsorbed on PE.
Figure 6 is the FTIR/ Attenuated Total Reflectance (ATR) ~e~LLa (llO0-lO00 cm~l region) of: (a) unmodified Impra0; (b) Impra0 exposed to dextran solution; (c) dextran triblock polymer adsorbed on PE (for comparison).

21~4166 Figure 7 is a FTIR/ATR spectra (1800-900 cm~l region) of unmodified PE exposed to: (a) 5% albumin solution for 24 hours; and (b) dextran triblock polymer adsorbed on PE
exposed to 5% albumin solution for 24 hours.
Figure 8 is a FTIR/ATR spectra (1800-900 cm~l region) of unmodified Impra~ exposed to 5% albumin solution for 24 hours (a), and dextran triblock polymer adsorbed on Impra~
oeD~ to 5% albumin solution for 24 hours(b).
Figure 9 is a W spectra (190-300 nm region) of: (a) unmodified PE ~ros~ 5% albumin solution for 24 hours; and (b) dextran triblock polymer adsorbed PE exposed to 5%
albumin solution for 24 hours.
Figure lO is a FTIR/ATR spectra (1800-900 cm l region) of: (a) unmodified PE exposed to human plasma; and (b) heparin triblock polymer adsorbed PE exposed to human plasma.
Figure ll is a FTIR/ATR spectra (1700-900 cm l region) of dextran triblock polymer adsorbed PE exposed to: (a) PBS
buffer solution, (b) 5% SDS solution, (c) 5% albumin solution, (d) human plasma.
Figure 12 is a FTIR/ATR spectra (1700-900 cm~l region) of heparin triblock polymer adsorbed PE ~Yrose~ to: (a) PBS
buffer solution, (b) 5% SDS solution, (c) 5% albumin solution, (d) human plasma.
Detailed Description of the Invention The present invention provides a triblock polymer which contains two hydrophilic blocks or segments, bridged by a hyd-~hobic block or segment, as shown in Figure l.
The hyd~o~hobic block l is a hydrocarbon chain that attaches spontaneously and iLreveL~ibly to the hydrophobic substrates through hydrophobic interactions to provide a means of attaching the triblock poIymer to the surface of the substrate. The hydrophilic segments 2A and 2B are polysaccharides or oligosaccharides, such as, for example, dextran, dextran sulfate, dermatan sulfate or polysaccharides or oligosaccharides of heparin.

~1 S 416 6 PCT/US93/11210 Hereinafter the term "polysaccharide" includes oligosaccharide. The triblock polymer adsorbs strongly on the surface of hydrophobic substrates, such as for example, polyethylene, polypropylene, silicon rubber, Impra~, Gortex~, Teflon~, and hydrophobic medical polyurethanes such as Pellethanes~.
The free triblock polymer is water soluble; the triblock polymer is dissolved in water and the implant is then immersed in the aqueous solution of the triblock polymer for about 24 hours. The triblock polymer spontaneously attaches to the substrate to provide a nonthrombogenic surface.
Protein adsorption from blood plasma is governed by the surface properties of the plasma proteins and the substrate, and by the process of mass transport in the near-surface layer, which is determined by the hydrodynamic conditions. The three dimensional structure of a protein, and hence its surface, is stabilized, at about 5-15 kcal/mol by intramolecular and intermol~c~ r forces, including attractive van der Waals forces, and strong hyd~o~hobic interactions that result from the attraction between nonpolar species in an aqueous medium. Common biomaterials are all hydrophobic so that attractive interfacial hydrophobic interactions with protein molecules results in a ~ LL Ull~ entropic driving force for the adsorption, of plasma proteins particularly with any hydrophobic domains on the protein. Albumin, in particular, ~Lr 01l~ly interacts with hyd~o~hobic surfaces;
and high albumin adsorption often occurs on implants.
Similarly, proteins present in low coll~e.lLLation in plasma will adsorb in relatively high amounts, if strong attractive forces are present between the substrate and the protein. HYdL o~hobic interactions and binding diminishes with increasing polar character of the polymer substrate, and, depen~ing on the prevailing interfacial force, this may lead to reduced adsorption or to an adsorbed protein WO94/11411 21~ 4 ~ 6 5 PCT/US93/11210 layer of different composition. For example, fibronectin adsorption increases as the hydrophilicity of the substrate is increased.
For highly hydrated biomaterials, such as a substrate having immobilized polyethylene oxide (PE0), entropic repulsive forces (ERF), also referred to as repulsive hydration forces, or steric repulsion, are important in resisting protein deposition. ERF is the long range (0-150 nm) repulsive force that arises from the dynamic motions and segmental interactions of hydrated macromolecular ~hAinc. The strength of the repulsive force increases as the size and mobility of the hydrated chain is increased.
However, chain length, configurational flexibility, surface chain density, and substrate to~oy.aphy and heterogeneity of the macromolecule are all believed to affect the adsorption and strength of the repulsive force. For many macromolecules such as heparin, which have a high charge density, the complex three-dimensional charge distribution constitutes either an additional ele~Llo~Latic repulsive force, or a recogn;tion sequence, such as the pentasaccharide antithrombin bi~;ng sequence in heparin, which provides an attractive force sufficient to overcome ERF at short dist~nc~c. This has permitted the biological macromolecules in blood to achieve highly evolved functional specificity. Interfacial forces are important in both polysaccharide-protein and polysaccharide-cell repulsive and attractive interactions.
It is believed the surface attachment of the triblock polymers of the present invention involves a long range interfacial force. The ERF results from the presence of highly hydrated polymer chains exten~;~g out from the surface of the substrate. The ERF increases with a high radius of gyration, that is, chain length, for the hydrated polysaccharide and with high surface density on the substrate. The triblock polymer structure is designed to maximize the interfacial effects of repulsive forces to increase resistance to protein adsorption and attractive 215~166 8 forces to promote attachment of the triblock polymer. The dextran-hydrocarbon-dextran triblock polymer (hereinafter "dextran triblock polymer") provides hydrophobic substrates with a highly hydrated, neutral, biological molecule at the surface, which maximize ERF. The dextran sulfate-hydlGcarbon-dextran sulfate triblock polymer provides a highly hydrated, negatively charged molecule at the substrate surface that maximizes the ERF with repulsive ele~LL ~a Latic force. Similarly, heparin-hydrocarbon-heparin triblock polymers in which the heparin polysaccharides have very low affinity for antithrombin III, (ATIII), involve ERF with repulsive electrostatic force. The heparin-hydrocarbon-heparin triblock polymer in which the heparin polysaccharides have high affinity ("HA") for ATIII involves ERF with repulsive electrostatic force.
The HA heparin has specific anticoagulant activity achieved through ~L~v~y specific electrostatic attractive force between antithrombin and a unique pentasaccharide binding sequence present in the HA heparin.
The PolYsaccharide The hyd~o~hilic portion of the triblock polymer is a polysaccharide preferably having an average mol~c~ r weight of from about 4,000 to about 500,000, more preferably about 6,000 to about 150,000. The polysaccharide can be a polymer of glucose, such as, for example, dextran or the polysaccharide portion of the glycoprotein, heparin. Good results have been obtained using dextran having a average mol~cll1~r weight of about 8,800 available from Sigma Chemical Company and with polysaccharides of heparin in the molPc~ r weight range 5,000 to 20,000.
The dextran polysaccharide is a neutral hydrated molecule so that plasma proteins are repelled by ERF. The heparin polysaccharides are hydrated and negatively charged which provides an additional electrostatic repulsive force that further repels plasma proteins and cellular elements such as platelets.

WO94/11411 ~ l a 4 ~ 6 ~ PCT/US93/11210 Most, if not all of the heparin polysaccharide molecules in the high affinity heparin and some of the individual heparin polysaccharide molecules of the mixed affinity heparin, contain the unique pentasaccharide seguence that is essential for heparin's anticoagulant activity. The heparin product of deaminative cleavage possesses a terminal 2,5 anhydromanose unit. The terminal aldehyde of the 2,5 anhydromanose binds to one of the terminal diamines on the hydrocarbon chain via reductive amination. Thus two types of heparin triblock polymers are prepared: one cont~i n ing high affinity heparin and the other containing a mixture of heparin polysaccharides having various affinities for ATIII.
Other suitable polysaccharide include the dermatan sulfate, and dextran sulfate, which are hydrated and negatively charged and serves to repulse proteins and platelets.
~he ~vd.v~arbon Chain The hyd~u~hobic segment of the triblock polymer is a h~d~u~arbon chain of sufficient length, at least 5 carbons, so as to provide a sufficient area for hydlG~llobic interaction between the substrate and the triblock polymer thereby serving to bind to triblock polymer to the substrate. The upper limit of the number of carbons in the hyd~û~arbon chain number is determining by the number of carbons that renders the triblock polymer insoluble in the aqueous medium. The solubility of the triblock polymer depends upon the type and molecl11Ar weight of the polysaccharide selected for the triblock polymer.
Preferably the hydrocarbon chain has from 5 to 18 carbons, more preferably from 6-12 carbons. The selection of the hyd~G~arbon chain depends upon the polysaccharide of the triblock polymer. It is believed that the more hydrophilic the polyc~cch~ride, the more it will tend to interact with the ~u~oul-ding plasma and the more the triblock will tend to be pulled from the substrate. In such a case the length of the hyd~ouarbon chain is increased to provide a greater 2154166 i l o surface area for interaction with the substrate. However, hydrocarbon ch~in~ having less than l9 carbons are generally preferred since their smaller size permits a greater number of triblock polymer molecules to bind to the substrate. It is believed that the greater the density of triblock polymer bound to the substrate, the greater the resistance to protein deposition. Good results have been obtained using l,6-hexanediamine and l,12-diamineododecane as the hydrocarbon chain.
The Substrate Any substrate with sufficient hydrophobicity to bind the hydrocarbon chain is a suitable substrate for the triblock polymer; however, for biological implants, the substrate material must also be biocompatible. Such biocompatible materials are known in the art, and include for example, polyethylene available from Abiomed Inc. in Danvers, Mass. and poly tetrafluoroethylene (PTFE) available under the trademark Impra0, from Impee, in Arizona. Other suitable materials include for example:
silicone rubber such as Silastic0 from Dow Corning;
silicone polymers; poly~o~ylene; Impra0; Gortex0; Teflon0;
and h~d~u~hobic polyurethanes. The selection of the substrate material depenAc upon the mechAnical and functional properties re~uired for the implant.

WO94/11411 ~ 1 5 4 1 6 6 PCT/US93/11210 Synthesis of the Triblock Pol~mer The Dextran Triblock PolYmer The first step is to react dextran with epichlorohydrin, to obtain an intermediate product as shown below:
Dextran-OH + CH2-CH-CH2Cl ) Dextran-O-CH2-CH-CH2Cl \/ I
O OH
(Dextran Intermediate Product) Next, the intermediate dextran product is reacted with an amine terminated hydrocarbon, such as, for example, 1,6-hexanediamine, as shown below:
Et3N
2Dextran-o-CH2-cH-CH2Cl ~ NH2-(CH2)m-NH2 Dex LL an-O-CH2-CH-CH2-NH-(CH2)m-NH-CH2-CH-CH2-O-Dextran OH OH

(wherein m preferably = 5-18) Though there are many hydroxyl groups on the dextran polymer chain, the reducing end, that is, the terminal hydlo~yl group, is more reactive than the other hydroxyl groups. As a result, the reaction with the epichlorohydrin OC~UL ~ at the reducing end of the dextran chain. Since the reaction of dextran with epichlorohydrin in the pres~nce of sodium hydroxide is commonly used to produce a three dimensional network of crosslinked chains, sodium hydroxide was specifically excluded from the formulation.

WO94/11411 215 ~16 6 PCT/US93/11210 Exam~le 1 A triblock polymer was prepared by adding 0.44 g (0.00005 mol) dextran, having an average MW 8,8000 from Sigma Chemical Co. to a 100 ml round-bottom flask contAin;ng 4 ml distilled water and 16 ~1 (0.0002 mol) 99%
epichlorohydrin from Aldrich Chemical Co. The mixture was stirred at 80C for 4 hours. A mixture of 1 ml distilled water, 0.3 ml 98% triethylamine from Aldrich Chemical Company, and 5 mg (0.043 mmol) of 98% 1,6-heYAne~iAmine from Aldrich Chemical Company, was added to the flask.
The mixture was stirred for another 3 hours at 80 and then the dextran triblock polymer was precipitated in a large PYcecc of acetone, washed several times with acetone and filtered. After drying overnight in a 70C oven, the dextran triblock polymer was ground into powder using a porcPl A; n mortar and pestle and stored in a clean, dry bottle.

Exam~le 2 A triblock polymer was prepared as in Example 1 except that 10 mg (0.086 mmol) of 1,6-heYA~e~iamine 98%, from Aldrich Chemical Company, was added instead of 5 mg.

Exam~le 3 A triblock polymer was prepared as in Example 1 except that 10 mg (0.05 mm~l) of 1,12-diamino~o~ecAne 98%, from Aldrich Chemical Company, was added instead of 10 mg. 1,6-h~YAnP diamine.

Svnthesis of the HeDarin Triblock PolYmer Heparin is a glycoprotein; its structure contains a unique pentAcAcchAride sequence that is responsible for its anticoagulant activity. The protein core of heparin is removed because it has no anticoagulant activity and its presPn~p on the surface of the substrate would reduce ERF
and increase the adsorption of plasma proteins.

~15~165 First, an aqueous solution of crude heparin having an average MW 5,000-25,000, derived from porcine intestinal mucosa, was obtained from Sigma Chemical Co. St Louis, M0.
The crude heparin and lO.o mg/ml sodium salt solution was - 5 p~c~ through a 3 X 8 cm cation exchange column containing 200-400 mesh Dowex 50W-X8 H+ resin and washed with about lO0 ml water. The pH was monitored during the ion eY~h~nge. The elutate was then mixed with 250 ml ethylene glycol dimethyl ether and cooled to -lO C using an ice-salt bath. Partial deaminative cleavage of crude heparin was carried out by the addition of lO ml isopentyl-nitrite.
The reaction was quenched after 40 minutes by adjusting the pH to 8.0 with 2.0 M Tris buffer. The mixture was concentrated by vacuum distillation at 37 C, and desalted by passage through an Amicon 8400 ultrafiltration membrane having a 500 moler~ r weight cut-off, from Diaflo Ultrafilters Company. Next, 5 ml saturated aqueous sodium acetate solution was added to 20 ml of the mixture and ~o~ed immediately into 2 liters of stirred cold 95%
ethanol. The heparin polysaccharide precipitated and was collected on a glass filter having a pore size 0.45 ~m and washed with 95% ethanol. The resulting heparin polysaccharide was a mixture of polysaccharides having varying affinities for ATIII (hereinafter referred to as "mixed affinity heparin"). This mixed affinity heparin, which po~se~s~s a terminal aldehyde group on the terminal 2-5 anhydromanose unit, was joined to l,6-hexanediamine by reductive amination in the presence of sodium cyanoborohydride.
Exam~le 4 A triblock polymer was prepared by adding 0.185 g of the mixed affinity heparin, (hereinafter "MA-heparin") to a lO0 ml round-bottom flask containing a mixture of 4 ml distilled water, lO0 mg NaBH3CN. Then 80 mg (0.688 mmol) of 98% l,6-hexanediamine from Aldrich Chemical Company was added to the flask. The mixture was stirred for 3 hours at ~lS~lS6 80OC and then the heparin triblock polymer was precipitated in a large ~ysecs of acetone. The precipitate was washed several times with acetone and filtered on the 12.5 cm diameter filter paper. After drying overnight in a 70C
- 5 oven, the MA-heparin triblock polymer was ground into powder using a porcelain mortar and pestle and stored in a clean, dry bottle.
Adsorption of the Triblock PolYmer on Surface of Substrate Ethylene oxide sterilized NHLBI primary reference low density polyethylene film (PE), in sheets having dimensions of 82.5 mm X 27 mm X 0.28 mm, from Abiomed Inc. were rinsed several times with distilled water. Aqueous solutions contAining 0.32% of the triblock polymer of each of the above examples were prepared. Samples of the PE were placed in each of the solutions for 24 hours. The PE was then removed and stirred in distilled water for l hour.
Finally, the PE was washed with distilled water several times and air-dried in a class lO0 clean hood.
For co,.L~ols, unmodified PE was rinsed several times with distilled water and incubated in a 0.32% aqueous solution of either dextran or heparin, in the same manner as above.
Characterization of the Dextran Triblock PolYmer The structure of the dextran triblock polymer was confirmed using four transform infra red (FTIR) ~ -LoScopy, gel permeation chromatography and l3C-nuclear magnetic resonAnce spectroscopy.
FTIR S~e~L~ O~CODY
FTIR transmission spectra of dextran and the dextran triblock polymer samples were obtained using the KBr pellet method on a Diglab FTS-40 FTIR spectrometer equipped with a triglycine sulfate (TGS) detector. Absorption spectra were obtained from rationing 2048 reference and sample scans which were obtained at resolution 8 cm~l. Materials used in the preparation of the KBr pellets were ground up, dried, mixed and reground and then pressed into a pellet under reduced pressure. ATR/FTIR 5~e~ a of various PE

WO94/11411 2~54t 6~ PCT/US93/11210 surfaces were obtained using the same spectrometer equipped with an attenuated total reflectance (ATR) accessary available from Wilkes Scientific and a liquid nitrogen cooled mercury-cadmium-telluride (MCT) detector. A
germanium crystal with dimensions of 50 X 20 X 2 mm and nominal incident angle of 45 was used as an IR internal reflection element (IRE). Spectra were obtained by co-adding 2048 interferogram scans obtained at a resolution of 8 cm~l. All the FTIR/ATR spectra were normalized using CH2 deformation vibration ~ (C-H) at 1460 cm~1 as st~n~rd to eliminate any variation in optical contact.
The FTIR transmission spectrum the dextran triblock polymer is shown in Figure 2. The spectrum shows: a ~L~u~y~ broad absorption band v(0-H) at -3550 cm~1 due to the high hydroxyl group content in dextran; stretching vibration bands vas(C-H) and vs(C-H) in CH2 in the range of 2960-2860 cm~1; and aliphatic bending vibration bands at 1460 cm~1 and 1367 cm~1. The hydroxyl in-plane bending vibration v(0-H) is located at 1410 cm~1. Finally, the C-o-C asymmetric and symmetric stretching vibration bands are observed at 1200-1000 cm~l.
Stre~hing vibrations v(N-H) were not identified above the ~eu~Lal noise in the FTIR spectrum of the dextran triblock polymer because the s~con~ry amine-functional group contributed and extremely small fractions of the long chain dextran.
Gel Permeation ChromatoqraphY
The mol~c?llAr weight distribution of the free dextran and the dextran triblock polymer were determined using a 2.5 x 92.5 cm Sephadex G-75 gel chromatography column. The elution solvent was a mixture of 20 mM tris + 50 mM NaCl, having a pH of 7Ø The solvent flow rate was run at 2.0 ml/min. St~n~rd linear dextrans were purchased from Sigma Chemical Company, St. Louis, M0, and used as reference molecular weight calibration curve. The average molecular weights of these reference dextrans was 39,100, 19,600, 11,000, 8,800 and 5,000. The void volume of the column was WO94/11411 21~ ~ PCT/US93/11210 estimated ~y ~nn-ng of a blue dextran sample (VO=175 mL).
After separation, collected fractions were analyzed by the carbazole reaction for uronic acid content according to T. Bitter et al., Anal. Biochem., 4:330 (1962).
The Kav value was calculated and then Kav-log molec~l Ar weight linear c~lihration curve in which R2=0.983 was obtAi n~--l . The GPC chromatogram of dextran and the dextran triblock polymer are shown in Figures 3A and 3B.
The commercial dextran used to synthesize the dextran triblock polymer, with average molecl~lAr weight of 8,800, is heterogPn~ollc in polysAcchAride chain lengths. The chromatogram of the dextran triblock polymer (Figure 3B) shows that the molec~lAr weight of dextran triblock polymer is much higher than that of the dextran. Kav values and the calculated mol~c~lAr weights for the dextran and the dextran triblock polymer are summarized in Table I.
Table I
Mol~ Ar Weight Distribution of Dextran Fractions Samples Kava Mol~c~lAr Weightb (weight avelL~e) Dextran fractions A

B 0.378 16,200 C 0.498 12,400 D 0.617 9,460 E 0.816 6,030 0.996 4,000 Dextran Triblock Polymer fractions A 0.059 33,700 B 0.159 26,900 C 0.398 15,600 D 0.617 9,500 E 0.697 7,800 F 0.797 6,300 a: Kav ~ def~ned a~ (Ve-Vo)/~Vt-Vo), wh-re Ve- elution volume, Vo=
vo~d volume, and Vt- column volume.
b: ~leoul~r w_~ght~ were ~t~ in~ by comparing the Kav value~ of the e ,le~ w~th tho~e of de~tran ~tandard~ of known molee~
weight.

~15~166 l 7 The molecular weight of the dèxtran triblock polymer is approximately double that of the dextran, indicating the formation of the dextran triblock polymers. These dextran triblock polymers resulted from the using the commercial preparation in which the dextran molecules had a variety of molecular weights. The analysis did not reveal any cross-linked polymer which is consistent with the fact that the dextran triblock polymer is water soluble.
13C-Nuclear Maqnetic Resonance S~ectrosco~Y
Proton-decoupled 13C-NMR spectra were obtained at natural ~hlln~nce~ with a total carbohydrate concentration of about 100 mg/2 mL of deuterium oxide. A Varian XL-200 (200 MHz) spectrometer was employed in the Fourier-transform, data processing mode. The spectra width was 6 kHz; the acquisition time, 1.4 seco~; and the pulse-width, 14 s~co~c. The number of transients was, in general, a function of the desired signal-to-noise ratio for each s~e~LL~m. Chemical shifts were expressed in p.p.m. relative to external tetramethylsilane, and calculated by reference to the lock signal.
The 13C-NMR spectra of the dextran, the intermediate dextran product, and the dextran triblock polymer are shown in Figures 4, 4B, and 4C. The 13C-NMR ~e~L~um of the dextran shown in 4A is located mainly in three regions: The C-2,-3,-4, and -5 chemical shifts are found in the 70-75 p.p.m. region; the anomeric (C-1) carbon atom displays a downfield chemical shift in the 85 to 105 p.p.m region, mainly at 97 to 103 p.p.m., as there is only an infinitesimal proportion of reducing sugar in the polymers;
and upfield chemical shifts in the 60-70 p.p.m region which are associated with bonded C-6 and llnhQn~ed C-6 atoms. The glycoside bond causes the chemical shift of the two carbon atoms involved to be displayed downfield by about 10 p.p.m.
The NMR Spectra for the intermediate dextran product shown in Figure 4B shows an upfield chemical shift at about 31 215416~ l 8 p.p.m. due to the Cl substitution. This is consistent with the measurement of chloride content in the sample obtained to titration.
The NMR Spectra for the dextran triblock polymer is shown in Figure 4C. As a result of the reaction with the amine terminated hydrocarbon, the chemical shift seen in the intermediate dextran product at about 31 p.p.m. was not present. However, the chemical shifts were observed at 55 p.p.m. for the c-carbon, at 53 p.p.m. for the d-carbon, and 24.04 p.p.m. assigned to e and f carbons.
In co"Llast to lH-NMR spectra, the peak area for 13C-NMR spectra does not necessarily reflect the population of atoms present. However, for carbohydrates, including polysaccharides, it has been shown that peak height is, in general, proportional to the number of carbon species present.
HYdroDhobic Interaction of the Dextran Triblock PolYmer with the surface of the PE
Figure 5 shows the ATR/FTIR spectra of unmodified PE, PE eYrose~ to the dextran solution and dextran triblock polymer adsorbed PE. The dextran triblock polymer adsorbed PE had VC-0-C absorption bands at 1200-1000 cm~l. The PE
~Y~oce~ to the dextran solution shows the same spectra as the unmodified PE, which establishes that the dextran does not bind to the surface of the PE.
Figure SA shows the ATR/FTIR spectra of unmodified PE, PE exposed to the heparin solution and heparin triblock polymer adsorbed PE. As shown in Figure 5A, spectra c, the heparin triblock polymer adsorbed PE had VC-0-C absorption bands at 1200-1000 cm~l. The PE exposed to the heparin solution, spectra b, shows the same spectra as the unmodified PE, spectra a, which establishes that the heparin does not bind to the surface of the PE.
Figure 6 shows the ART/FTIR spectra of unmodified Impra~, and dextran triblock polymer adsorbed Impra~. The dextran triblock polymer adsorbed Impra~ had vc-o-c adsorption bands at 1080 cm~1 and 1056 cm 1 (see spectr b) WO94/11411 ~ 1 a 416 6 PCT/US93/11210 which are characteristic of the dextran; compare to spectra c the spectra for dextran triblock polymer adsorbed PE.
The control Impra~ sample lacks such bands.
Water Contact Anqle Measurement Advancing water contact angle(eA) was measured by the sessile-drop method using a Rame-Hart goniometer. The water contact angle measurment is a surface sensitive assay ofthe top 5-lO a~ Loms of the surface of a material. The presence of hydrophillic molecule on the surface should reduce the water contact angle. The advancing water contact angle ~ a H20 were measured by placing a 2 ~l water drop onto the PE surface using a microsyringe attachment.
A second 2 ~L drop was added to the first drop and the new contact angle was measured. The process was repeated three additional times and the results represented the measurement of ~ a H20. Measurement of water contact angle was repeated four times for each sample. All the contact angle measurements were performed at room temperature and about 50% relative humidity.
The unmodified PE surface is hydrophobic and has water contact angle of about 90. The water contact angle of the dextran triblock polymer adsorbed on PE is about 75 and shows very small contact angle hysteresis. This decrease is the water contact angle confirms the presence of the hydlu~1.illic dextran molecule on the substrate.
ESCA Analvsis ESCA analysis is a surface analysis of the top 6 nm of the material. Surface analysis of the unmodified PE and the dextran triblock polymer adsorbed PE was performed using a Perkin Elmer PHI-5400 ESCA system with a 400 W
monochromatized magnesium X-ray source at 44.74 pass energy and 45 take off angle. The results are shown in Table II.
- About 1.3% oxygen, 98.7% carbon and 0% nitrogen were detected on the surface of the unmodified PE control. All the dextran triblock polymer adsorbed PE samples showed a substantial increase in the percent of oxygen and nitrogen present on the surface of the samples. The increased WO94/11411 21 5 ~1 6 S PCT/US93/11210 polymer adsorbed PE, confirms the preQ~nce of the dextran segment of the triblock polymer. The presence of the nitrogen confirms the pres~nce of the amine groups present in the hydrocarbon chain of the triblock polymer. No chloride content was detected on the dextran triblock polymer adsorbed PE. The PE eYpos~ to dextran solution showed results similar to the unmodified PE control.
Table II Atomic Conc~ntration of Dextran Adsorbed PE

Sample Atomic Concentration (%) Unmodified PE l.3 0 98.7 15Example 3 17.9 l.8 80.3 Example 2 21.9 l.l 77.0 Example l 9.4 0.9 89.7 AtomLc co~ Lration wa~ dete in~d by ESCA ~pectrometer operated in ~urvey ~can mode. E~tLmated error for ESCA analy~L~ L~ + 1096.

The ESCA, ATR/FTIR and water contact angle results confirm the pr~æ~nce of the dextran triblock polymer on the PE surface. Since the PE ex~oscd to the dextran solution does not show the pre~nce of dextran, the adsorption of the triblock polymer results from the hydlo~hobic interaction between the PE surface and the hydLo~arbon chain in the dextran triblock polymer.
Characterization of the He~arin Triblock Polvmer Protein Resistance of Triblock PolYmer Adsorbed PE
Albumin is the most ~hlln~nt protein present in blood plasma constituting about 50% of the total protein in plasma. Albumin is often a major constituent of the protein layer that is deposited on polymeric implants.

WO94/11411 2 1 ~ 4 1 fi 6 PCT/US93/11210 Accordingly, the triblock polymer coated substrates were evaluated for resistance to albumin deposition.
Samples of the unmodified PE, unmodified Impra0, the dextran triblock polymer adsorbed PE and the dextran triblock polymer adsorbed Impra0 were rinsed in distilled water and put into 5~ bovine albumin solution from Sigma Chemical Company for 24 hours. The bovine albumin solution, which had a pH of about 7, was sterile filtered, and contained 0.70% NaCl. Then the samples were stirred in distilled water for 2 hours, washed several times with distilled water, and air-dried in a hood for 24 hours at room temperature.
Figure 7 shows normalized ATR/FTIR spectra of unmodified PE (spectra b) and the dextran triblock polymer adsorbed PE (spectra a) after incubation in the 5% albumin solution. Spectra were normalized to the (C-H) band at 1460 cm~l in PE so that the absorbencies are directly comparable. The relative amount of albumin adsorbed on the surface of the PE was calculated from the amide absorption bands at 1650 cm~1 with baseline correction and normalization to polyethylene deformation vibration (in-plane) at 1460 cm~l.
The unmodified PE shows strong amide I and amide II
absorption bands at 1650 and 1550 cm~l~ respectively, which are characteristic of the albumin. The spectrum shows that the albumin ~L~G~I~1Y absorbed onto the unmodified PE
surface. However, the spectra of the dextran triblock polymer adsorbed PE display weak amide I and amide II
absorption bands as compared to unmodified PE. Only about 8.5% +3% albumin was adsorbed on the dextran triblock polymer adsorbed PE surface compared with the unmodified PE.
Figure 8 shows normalized ATR/FTIR spectra of unmodified Impra0 and the dextran triblock polymer adsorbed Impra0 after incubation in the 5% albumin solution.
Spectra were normalized to the (C-H) band at 1150 or 1225 cm~l in the Impra0 so that the absorbencies are directly 21~i4l66 2 2 comparable. The relative amount of albumin adsorbed on the surface of the Impra~ was calculated from the amide I
absorption bands at 1650 cm~1 with baseline correction and normalization to Impra~ deformation vibration.
The unmodified Impra~ shows strong amide I and amide II absorption bands at 1650 and 1550 cm~1~ respectively, which are characteristic of the albumin. Thus, the spectrum shows that the albumin absorbed onto the unmodified Impra~ surface. However, the spectra of the dextran triblock polymer adsorbed Impra~ lack the amide I
and amide II absorption bands.
Samples of the unmodified PE and the mixed affinity heparin triblock polymer adsorbed PE, were exposed to fresh human blood plasma for 1 hours. The the samples were stirred in distilled water for about 1 hour, washed several times with distilled water, and air dried in a hood for about 1 hour at room temperature.
Figure 10 show normalized ATR/FTTR spectra of unmodified PE (spectra a) and the heparin triblock adsorped PE (spectra b) after incubation in the plasma. Spectra were normalized as for Figure 7. The ~e~Lra of the heparin triblock polymer adsorbed PE lacks the characteristic amide I and amide II bands characteristic of albumin, dem~"~L~ating that the heparin triblock adsorbed PE resists plasma protein deposition.
W SDectroscoDy W ~e~LLa of unmodified PE and the triblock polymer adsorbed PE, which were exposed to albumin, were obtained using a W -VIS Sc~nn;ng Spectrophotometer system designated as W -2101PC from Shimadzu Corporation. Sample were fixed on the sample side of the film holder model P/N204-58909, from Shimadzu Corporation. The unmodified PE was fixed on the reference side. Figure 9 is the W spectrum of unmodified PE; it shows a strong absorption band at 208 nm.
In contrast, as shown in Figure 9, spectra b, the dextran WO94/11411 ~ t 6 6 PCT/US93/11210 triblock polymer adsorbed PE does not show such band; the triblock polymer pre-adsorbed PE did not adsorb the albumin.
Rotatinq Disc Ex~eriments ~ 5 To determine the stability of the triblock polymers on the PE surface under high interfacial shear conditions, samples of the triblock polymer adsorbed PE were cut into 17 mm diameter discs and mounted on the disc of spindle assembly of a Model Afasr Analytical Rotator from Pine Instrument Co. The discs were stirred for l hour at 2000 rpm in solutions of either PBS buffer, 5% sodium dodecylsulfate solution (SDS), 5% bovine albumin solution, or fresh human plasma. The samples were then washed several times with distilled water. The shear stress under these conditions was calculated as about 206 dynes/cm2 at the edge of the sample.
The shear stress was determined by:
~ = 0-800 ~ r (~3/v)0 5 where r is the magnitude of shear stress at the disk surface (dynes/ 2), ~ is the absolute viscosity (O.Oll poise), and r is the radial distance from the center of the disk (8 cm), ~ is the angular velocity (209.4 rads/sec), and v is the knematic viscosity (0.0107 stokes). Under the conditions of 5% albumin solution or human plasma, this 2S pro~ e~ a protein flux at the surface of 0.52-0.56 ~g/mm2 sec as calculated from:
~ = 0.62 D2/3 vl/6 ~l/2 C
where j is the mass flux, D is the diffusivity (4.04 x 10-7 cm2/sec), Cm is the bulk concentration (5 g/dL), v and ~ are as above.
All spectra were normalized to the 1460 cm~l band of unmodified PE, so that the spectra may be directly compared. The absorbance scale for Figures 5, 9, and ll are identical.
Figure ll shows ATR/FTIR spectra of the, dextran triblock polymer adsorbed PE after rotating in the solutions. Spectra "a" is the spectra of the sample that WO94/11411 2 ~5 4~ 6 6 2 4 PCT/US93/11210 rotated in PBS buffer solution, "b" is the spectra of the sample that rotated in 5% SDS solution, "c" is the spectra of the sample that rotated in 5% albumin solution, and "d" is the spectra of the sample that rotated in human ~ 5 plasma solution. All the samples show the absorption bands of dextran at 1200-lO00 cm~l; the triblock polymer remained bound to the PE despite the shear conditions. Further, as evidenced by spectra c and d, which revealed weak absorption bands of amide I and amide the dextran triblock polymer adsorbed PE does not significantly adsorb albumin or other plasma proteins.
Figure 12 shows ATR/FTIR spectra of the heparin triblock polymer adsorbed PE surfaces after rotating in the solutions. Spectra "a" is the spectra of the sample that rotated in PBS buffer soIution, "b" is the spectra of the sample that rotated in 5% SDS solution, "c" is the spectra of the sample that rotated in 5% albumin solution, and "d" is the spectra of the sample that rotated in human plasma solution. All the samples show the absorption bands characteristic of the heparin at 1250 cm~l and at 1120-950 cm~l the triblock polymer remained bound to the PE despite the shear conditions. Further, as evi~PncP~ by spectra c and d, which revealed weak absorption bands of amide I and amide II, the heparin triblock polymer adsorbed PE does not significantly adsorb albumin or other plasma proteins.
The results establish that the triblock polymers have been formed, coated and that they tightly bind to the substrates. Indeed, the triblock polymers remained bound to the PE after being subjected to a shear rate of 206 dynes/cm2. This indicates that there is a strong binding force between the hydrophobic hydrocarbon segment of the triblock polymer and substrate surface, to provide a a ~ u~ stable interface between the substrate and the triblock polymer. Moreover, the triblock polymer provides the substrate with resistance to plasma protein binding, including resistance to albumin binding.

WO94/11411 2 ~554 1 6 ~ PCT/US93/11210 Although this invention has been shown and described as a triblock polymer having two polysaccharide segments and one hydrocarbon chain segment, various adaptations and modifications can be made, such as a diblock polymer having one polysaccharide segment and one hydrocarbon chain segment, without departing from the scope of the invention as defined in the appended claims.

Claims

We Claim:

1. A method for reducing the thrombogenicity of a substrate, comprising the steps of:
a. providing a water soluble block polymer comprising at least one hydrophobic block comprising a hydrocarbon chain having from at least 5 carbons and at least one polysaccharide having an weight molecular weight of at least 4,000; and b. applying the block polymer to the surface of the substrate.

2. The invention of Claim 1, wherein the hydrophobic hydrocarbon chain has from 5 to 13 carbons.

3. The invention of Claim 1, wherein the substrate is comprised of polyethylene.

4. The invention of Claim 1, wherein the substrate is comprised of polytetrafluoroethylene.

5. The invention of Claim 1, wherein the hydrophobic hydrocarbon is derived from 1,6-hexanediamine.

6. The invention of Claim 1, wherein the hydrophobic hydrocarbon is derived from 1,12-diaminododecane.

7. The invention of Claim 1, wherein the polysaccharide is dextran.

8. The invention of Claim 1, wherein the polysaccharide is heparin.

9. The invention of Claim 1, wherein the polysaccharide is dermatan sulfate.

10. The invention of Claim 1, wherein the polysaccharide is dextran sulfate.

11. The invention of Claim 1, wherein the polysaccharide is dextran, and the hydrocarbon is derived from 1,6-hexanediamine.

12. The invention of Claim 1, wherein the polysaccharide is dextran, and the hydrocarbon is derived from 1,12-diaminododecane.

13. The invention of Claim 1, wherein the polysaccharide is heparin, and the hydrocarbon is 1,6-hexanediamine.

14. The invention of Claim 1, wherein the polysaccharide is heparin, and the hydrocarbon is 1,12-diaminododecane.

15. A block polymer comprising: at least one hydrophobic hydrocarbon chain having at least 5 carbons; and at least one polysaccharide having an average molecular weight of at least 4,000.

16. A substrate adsorbed with a polymer comprising at least one hydrocarbon chain and at least one polysaccharide, said hydrocarbon chain having at least 5 carbons and said polysaccharide having an average molecular weight of at least 4,000.

17. A triblock polymer comprising a hydrophobic hydrocarbon having from 5 to 13 carbons and at least two polysaccharide molecules, wherein each polysaccharide has an average molecular weight greater than 4,000.

18. A method for making a polymer, comprising the steps of:
a. providing at least one polysaccharide having an average molecular weight of at least 4,000; and providing epichlorohydrin;
b. mixing the polysaccharide and the epichlorohydrin;
c. providing at least one hydrophobic hydrocarbon chain having at least 5 carbons, and having at least one terminal amine group; and d. adding the hydrophobic hydrocarbon chain to the mixture of step e, to provide a water soluble block polymer comprising: at least one hydrophobic hydrocarbon chain having at least 5 carbons; and at least one polysaccharide having an average molecular weight of at least 4,000.