296

Bioconjugate Chem. 1997, 8, 296−303

Synthesis of a New Photoreactive Derivative of Dipyridamole and Its Use in the Manufacture of Artificial Surfaces with Low Thrombogenicity
Yvette B. J. Aldenhoff,† A. Paul Pijpers,‡ and Leo H. Koole*,†
Centre for Biomaterials Research, Maastricht University, P.O. Box 616, 6200 MD Maastricht, The Netherlands, and DSM Research, Geleen, The Netherlands. Received November 15, 1996X

Photoimmobilization of dipyridamole (Persantin) was accomplished through the use of a new synthetic conjugate molecule, 1. Persantin is a powerful inhibitor of platelet activation and aggregation and is widely used as a vasodilator. Conjugate 1 consists of triply protected dipyridamole [three of the four hydroxyl groups carry a tert-butyldimethylsilyl (TBDMS) protective group) and the photoreactive 4-azidobenzoyl group. A short hydrophilic spacer chain, derived from triethylene glycol, separates the protected dipyridamole system and the photoreactive group. Compound 1 was immobilized on polyurethane sheets (Pellethane D-55) through irradiation with ultraviolet (UV) light, and the protective groups were removed afterward. The resulting modified polyurethane surfaces were characterized by different physicochemical techniques: UV extinction, contact angle measurements (captive bubble technique), and X-ray photoelectron spectroscopy (XPS). The UV extinction measurements showed the presence of 13 ( 1 nmol of immobilized dipyridamole/cm2. The contact angle measurements revealed that the modified surface was markedly more hydrophilic than the control (i.e. unmodified polyurethane). XPS measurements clearly established the presence of immobilized dipyridamole in the outermost layers of the modified surface. This was especially clear from the XPS spectra recorded at a low take-off angle (∼6°). Furthermore, the XPS spectra showed that the TBDMS protective groups had been quantitatively removed during the deprotection/washing treatment. The in vitro blood compatibility of the modified surface was studied with the thrombin generation assay as developed in our group, as well as with scanning electron microscopy. The thrombin generation test produced a lag time of 1275 s for the modified surface, as opposed to 569 s for the control. Scanning electron microscopy showed that far fewer platelets adhere to the modified surface (approximately 7 × 103/mm2) as compared to the control (approximately 6 × 102/mm2). Taken together, the experimental data reveal that the modified surface has excellent blood compatibility in vitro. It is discussed that the use of conjugate 1 leads to simultaneous exposure of dipyridamole at the modified surface and to a marked increase of the surface hydrophilicity, which is likely to hamper adsorption of plasma proteins. The combination of these effects is uniquely related to the molecular buildup of 1. Conjugate 1 will be used in future work that is aimed at preparing small-caliber polyurethane vascular grafts with a blood compatible lumenal surface.

INTRODUCTION

Dipyridamole is a nontoxic drug that acts as an inhibitor of the activation and aggregation of blood platelets (1, 2). The drug is also a powerful vasodilator, i.e. a substance which induces widening of the blood vessels after its systemic injection (1, 2). Dipyridamole,

which is marketed as Persantin, is commonly administered to patients before and after percutaneous transluminal coronary angioplasty (“Dottering”), mostly in com† ‡

Maastricht University. DSM Research. X Abstract published in Advance ACS Abstracts, April 1, 1997.

bination with aspirin. We became interested in dipyridamole in the course of our work, which is aimed at the manufacture of new materials with improved blood compatibility. We concentrate both on the preparation and on the testing (in vitro and in vivo) of such materials (3-7). Blood clotting as a result of the contact between blood and an artificial surface remains perhaps the most important problem in the development of improved cardiovascular devices. We know, particularly from the pioneering work of Vroman et al. (8-10), that contact between blood and an artificial surface immediately leads to adsorption of plasma proteins. Adsorbed proteins may be displaced by others, and these processes probably occur within several minutes. Proteins in the adsorbed state adopt a different molecular conformation as compared to the circulating state, and platelets and leukocytes may adhere via receptor sites that match structural patterns of an adsorbate. It has become clear that immunoglobulins, fibronectin, and fibrinogen have specific significance as adsorbates (11). When a platelet is activated through this mechanism, it may in turn trigger other platelets and induce generation of thrombin, leading to deposition of fibrin. These events will result in the formation of a thrombus. Recent work by Bamford
© 1997 American Chemical Society

S1043-1802(97)00020-7 CCC: $14.00

We report consecutively (i) synthesis of 1. It could be demonstrated that such conjugate molecules can be covalently coupled to a polyurethane surface via irradiation with ultraviolet (UV) light (14). (iv) physicochemical characterization of the modified surfaces (contact angles. As the macromolecular vehicles containing dipyridamole cannot cross the platelet’s membrane. fibronectin or fibrinogen).Photoreactive Derivative of Dipyridamole Scheme 1. suggest that the platelet inhibition by immobilized dipyridamole counteracts or overrules platelet activation by adsorbed plasma proteins (e. Then. 13). (iii) deprotection and purification. To this end. was esterified through reaction with 4-azidobenzoyl chloride. implied that dipyridamole could play a role in the development of new blood-compatible materials (12. which was subsequently treated with zinc bromide to cleave the trityl group (20).. taken together. The synthesis route to 1 is outlined in Scheme 1. we could also show that coupling of dipyridamole onto water-insoluble polymers (e. and (v) in vitro tests of the blood compatibility of modified and unmodified surfaces. No. at least in vitro (3). This reaction afforded the desired conjugate structure 1. This reaction afforded compound 4. Immobilization of dipyridamole onto the surface of water-insoluble polymers was accomplished through a novel photoimmobilization technique (3-5. 3. Purity and identity of conjugate 1 and all synthetic intermediates were estab- . 8. In some cases.. Bamford et al.. The preparation started out with tris[tert-butyldimethylsilyl]dipyridamole (2). chemically coupled dipyridamole to a series of water-soluble macromolecules and subsequently showed that the drug retained its platelet-inhibitory activity. which has the characteristic feature that the (protected) dipyridamole unit is linked to 4-azidobenzoyl via a short hydrophilic spacer chain. and this photoreaction generates a highly reactive (electrophilic) didehydroazepine structure (16-18). XPS). Evidence in support of an inhibitory action of immobilized dipyridamole on contacting platelets has accumulated. MATERIALS AND METHODS Preparation of Conjugate 1. a urethane NH group) generates a new covalent bond between the conjugate and the polymer surface. 1. the drug was even slightly potentiated. We report here a comprehensive study on a new conjugate molecule. which was prepared from dipyridamole as described earlier (14). we synthesized a series of novel molecules. 14). in 18% overall yield based on 2. The resulting alcohol. Later. We decided to prepare and test such a system since we anticipated that a hydrophilic spacer chain could facilitate exposure of the deprotected dipyridamole in the aqueous boundary layer at the polymer’s surface. Vol. 1997 297 et al. Compound 2 was reacted with 1-(2′-Otritylethoxy)-2-(2′-bromoethoxy)ethane (3) in the presence of sodium hydride (19).g. it is described how compound 1 can be immobilized on the lumenal surface of a short polyurethane vascular graft. The 4-azidobenzoyl group loses molecular nitrogen (N2) upon UV irradiation. Our data. 5. which will be used in followup in vivo investigations. Preparation of Conjugate Molecule 1 Bioconjugate Chem. (ii) photoimmobilization of 1 on polyurethane sheets. which consist of a triply tert-butyldimethylsilyl (TBDMS)protected (15) dipyridamole connected to a 4-azidobenzoyl group. these experiments revealedsalthough in an indirect mannersthat dipyridamole can exert its function via a putative receptor site on the exterior of the platelet. Moreover. the polyurethane Pellethane D-55) markedly contributes to the blood compatibility of the surface.g.g. reaction with a nucleophilic site at the polymer surface (e.

53 mmol) (19) in 50 mL of anhydrous tetrahydrofuran. CH2CPh3). A solution of 4-azidobenzoic acid (7. respectively. pyridine was removed (last traces were removed by coevaporation with toluene). mp.13 mmol) in 125 mL of anhydrous pyridine was added triphenylmethyl chloride (17. 51. 51. CH2OCPh3).18. 61.50 g. and concentrated to dryness under reduced pressure. 131. t. filtered. d.66. the polymer surface was placed horizontally and consecutively wetted. and concentrated to dryness under reduced pressure. OCH2CH2O).. Subsequently. 45. 7.00 g.507.1 M citric acid. 1H NMR (CDCl3) δ 8. and brine. CH2 of piperidine rings).10 g (25%). Compound 5 (1. stench). Elution with 4:1 petroleum ether/ethyl acetate afforded pure 1 as a yellowbrown oil: Rf (petroleum ether/ethyl acetate 1:1) ) 0. 20.90 mmol) in anhydrous dichloromethane (125 mL).34. 1.76. 4.52 mmol) were dissolved in 50 mL of anhydrous pyridine.70 (6H.35 g. No. A solution of bromine (12.298 Bioconjugate Chem.05 (2H.09. The solid residue was recrystallized from hexane to give pure 4-azidobenzoyl chloride as a yellowish solid: yield. 86. 70. 4. all volatiles were removed under reduced pressure. The resulting mixture was stirred for 1 h under exclusion of moisture.95 (1H. 1.60 g. 61.50-7. 160.74. 138. m.42 g. using deuterated chloroform or deuterated methyl sulfoxide as the solvent. anhydrous triethylamine (10. 118. 131. Silica gel 60 (particle size 0. CH2-N and CH2-O). To a stirred solution of triethylene glycol (75.24 (2H. 22.05. s. 57.28.8 °C.6 MHz. The crude product was applied to a silica gel column. aromatic H). 126. During spraying.95 (27H.29. 127. A compressed-air spayer was used (tubetype. Pyridine and dichloromethane were distilled from calcium hydride and stored over Linde 3Å molecular sieves. 30. CH2OH).79. Crude 5 was applied to a silica gel column. 76. Preparation of Compound 1. m.31. t.59 (2H.15. Preparation of Compound 5. . 25. CH2-N and CH2-O). 127. distilled.4 mm. 500. CH2Br). The residue was taken up in diethyl ether and washed with water. d.00 mmol).05-3. 70. 71.68. and brine.29 g. s.200 mm) was used for column chromatography. 1. we used a medical grade polyurethane foil (Pellethane D-55). 3. obsd m/z ) 1225 (M + H)+.82.34. aromatic H). 0.50 mmol) in 10 portions during 3 h.629-2).43. 25.29. The residue was taken up in dichloromethane and washed with saturated aqueous bicarbonate (three times) and water (three times). Different analytical techniques.69. 59. m.70.87. 106. distilled from calcium hydride. br s. filtered. Aldrich Z12.20. 1. All solvents and starting materials were of the highest available purity or were purified as specified. Tetrahydrofuran was passed through an alumina column. and the solution was magnetically stirred for 3 h.34.27 g.39.75. FABMS. calcd for C55H97N11O8Si3 m/z ) 1224.80 (34H. 48. CH3-Si).04 mmol) in 250 mL of anhydrous tetrahydrofuran was refluxed for 30 min.85 (2H. on a Varian Unity-Plus spectrometer. Elution with dichloromethane afforded 3 as a white solid: Rf (dichloromethane) ) 0. 1H NMR and 13C NMR were recorded at 399.02 (2H. Methanol was predried with iodine and magnesium turnings. br s. 3.59. 26.e. aromatic H). Tetramethylsilane was used as the internal reference (δ ) 0. 70.02 (18H. 3.27 g. 63.6. and stored over Linde 3Å molecular sieves. 64. 1H NMR (CDCl3) δ 8. Then.65 (12H.28 g. t.63. were used.12. 63.43 [2H. Preparation of 4-Azidobenzoyl Chloride.76 g.39 mmol) and 4-azidobenzoyl chloride (0. Vol. 18. 128. The crude product was applied to a silica gel column.94.28. and stored over Linde 3Å molecular sieves.48. and this solution was sprayed onto several specimens of the polyurethane foil (dimensions 50 × 50 mm). high-field nuclear magnetic resonance (NMR. 3.02 (18H. 69. compound 3 (2. 2. 13C NMR (CDCl3) δ 160. 400 MHz for 1H). which was extracted with ethyl acetate.21 (2H. m. 13C NMR (DMSO-d6) δ 166. The mixture was sonicated under argon for 15 min. 3. 1H NMR (CDCl3) δ 7. The residue (crude 4) was taken up in 10 mL of anhydrous dichloromethane and treated with 2 mL of anhydrous methanol and anhydrous zinc bromide (10.38. 126.80.15 (15H. Triethylamine was distilled from calcium hydride and stored over potassium hydroxide pellets.84 (27H. 1. The crude product was applied to a silica gel column. 119. The foil had a smooth surface (as could be verified with scanning electron microscopy) and was transparent. OH).05 (2H. yield. 3. CH3-Si). m. The organic layer was dried on magnesium sulfate.69.76 g (81%). CH2O). s. 1H NMR (CDCl3) δ 4.98. 128.51. 70. 0. 48.85 g (89%). 13C NMR (CDCl3) δ 144. The reaction mixture was poured into saturated aqueous ammonium chloride (150 mL). lished unequivocally. 127. 1H NMR (CDCl3) δ 7. CH2C(O)]. br s. t. Thin layer chromatography was performed on glass plates (3 × 10 cm) with a fluorescent indicator.98 mmol) (21) and thionyl chloride (16.68-1. 1997 Aldenhoff et al.21. Compound 1 was dissolved in 2-propanol (concentration 1 mM).48 (32H. The solvents were evaporated under reduced pressure. 8. and the residue was taken up in ethyl acetate and washed consecutively with a 0.66 (8H.91. -5. Elution with 1:1 petroleum ether/ethyl acetate afforded pure 5 as a yellow viscous oil: yield.02 (2H.49. 0. Elution with 1:1 petroleum ether/ethyl acetate afforded pure 1-(2′-O-triphenylmethylethoxy)-2-(2′-hydroxyethoxy)ethane (i. The addition was stopped as the yellow-brown color of free bromine persisted.60.69. and concentrated to dryness under reduced pressure. The solution was stirred for 16 h under exclusion of moisture.906 g. and fast atom bombardment mass spectrometry. aromatic H). CH2 of piperidine rings). 4. filtered. s. Preparation of 1-(2′-O-Triphenylmethylethoxy)2-(2′-bromoethoxy)ethane (3).84 g. 0. 63. m. including chromatography (thin layer). CPh3).36. 132. 6. Throughout our experiments.65 mmol) was added to a solution of 2 (3.00 mmol) and 4-(dimethylamino)pyridine (0. 70. 7. filtered.00 ppm). filtered.18 (15H. 4. Then.56 g (84%). The organic layer was dried on magnesium sulfate.81 mmol) in anhydrous dichloromethane (60 mL) was added dropwise to a solution of triphenylphosphine (18. 3. 143. t.08. d.9 and 100. and concentrated to dryness under reduced pressure.32. yield. 119. 86. d. The residue was taken up in ethyl acetate and washed with 0. This reaction mixture was magnetically stirred for 2 h.1 M citric acid. 23. Then.70-3.35 mmol) and a solution of 1-(2′-O-triphenylmethylethoxy)-2-(2′-hydroxyethoxy)ethane (20.063-0. 153. yield.38 g.50. its thickness was 0.19 g. Then. t-Bu). Mass spectra were run on a Kratos MS 80 RF instrument. t.12 g (72%). CPh3). saturated bicarbonate.98 mmol) was added and the reaction mixture was magnetically stirred for 16 h. saturated bicarbonate. and concentrated to dryness under reduced pressure.09 g.54. monotritylated triethylene glycol) as a viscous oil: Rf (petroleum ether/ethyl acetate 1:1) ) 0. 13C NMR (CDCl ) δ 144. pyridine was evaporated (last traces were removed by coevaporation with toluene). 69. 3. OCH2CH2O). Sodium hydride (60%. 1. The combined organic layers were dried on magnesium sulfate. All volatiles were removed under reduced pressure (in the hood.76.24. 0. t-Bu). The organic layer was dried on magnesium sulfate.68.28. 70. Preparation of Modified Polyurethane Surfaces. 3. 3 72.08 mmol) in 75 mL of anhydrous dichloromethane were added dropwise. 70.44 (2H. The organic layer was dried on magnesium sulfate.62 (12H. 45.

Besides the usual angle (90°). and Au (24). The distance between the lamp and the foil surface was 30 cm. cut out of the foil. The samples were measured with different take-off angles. which is a relatively passive material. The polyurethane surface was then removed and dried. each foil was irradiated for 15 min with a Philips HPA 1000 high-power UV lamp (Philips Lighting. Physicochemical Characterization of the Modified Surfaces. Assignments: (a) aromatic protons from the azidobenzoyl group. the sample holder was rotated in such a way that a take-off angle of ∼6° was obtained. and a Leybold DS100 data system was used for data acquisition and analysis. Sample areas of 4 × 7 mm2 were analyzed. Technical details on the lamp are as follows: nominal power. 400 MHz 1H NMR spectrum of 1. The UV extinction was measured at 408 nm. dissolved in deuterated chloroform. Note that the blood was taken from a healthy donor. and a linear background was substracted (25). The instrument was controlled by a HP A400 computer. All physicochemical and biochemical experi- . number of transients. and turning was repeated 20 times. size.53 mL of tetrahydrofuran (same concentration as the blank).1 mL was positioned under the surface and the contact angle was measured. Two reference materials were included in the assay: polyethylene (PE). time domain. and sputter-coated with gold. by exposing the polyurethane surface to trifluoroacetic anhydride vapor for 5 min in a sealed vessel (29). No. (f) CH3-Si. (26) and the Scofield ¨ cross sections (27). Biochemical Characterization. (d) CH2 of piperidine rings. UV-C 230. (e) tert-butyl. The spectrometer was calibrated using Ag. Then.Photoreactive Derivative of Dipyridamole Bioconjugate Chem. Cu. and the density of immobilized dipyridamole coupled to the polyurethane was calculated (3). Derivatization of the hydroxyls of the dipyridamole by trifluoroacetic anhydride was achieved Figure 1. Vol. (b) CH2C(O). and a micrometer-adjustable X-Y stage vertically mounted on an optical bench.50 mL). Each foil was then thoroughly washed with 2-propanol and water and stored in dry form. The platelets attached to the polyurethane surfaces were fixed with glutaraldehyde. The box was then filled with doubly distilled water. each foil was thoroughly washed with 2-propanol and immersed for 24 h in a stirred solution of tetrabutylammonium fluoride in nitromethane (22). Fast atom bombardment mass spectrometry provided further evidence for the identity of 1 (vide supra). 7. UV-A 2500. After irradiation. The specimens were then dehydrated with ethanol. For the X-ray photoelectron spectroscopy (XPS) experiments. 3. UV-B 900.95 eV. The molar extinction coefficient at this wavelength is 4397 L mol-1 cm-1. Spectral parameters: spectral width.6 eV) radiation of a Mg/Al double-anode X-ray tube (13 kV. For energy referencing a C 1s binding energy of 285. dissolved in CDCl3. subjected to critical point drying. This was concluded after comparison of different pieces (dimensions 10 × 15 mm).0 eV for aliphatic carbon was used (28). a similar piece of the modified polyurethane (139. radiation output (mW/cm2. 64K. 8. the full width at half-maximum for Ag 3d5/2 was 0. The chemical synthesis of conjugate compound 1 proceeded smoothly. Figure 1 shows the 400 MHz proton NMR spectrum of 1. The UV absorption spectrum of dipyridamole shows a maximum at λ ) 408 nm. Mg KR (1253. (c) CH2-N and CH2-O. Following incubation with platelet-rich plasma. 930 W. the blood plasma was anticoagulated with citrate prior to the assay. 8. (*) solvent signal (CHCl3). The UV extinction spectrum was measured and stored in the digital memory of the spectrometer. RESULTS Physicochemical Characterization. The Netherlands). The contact angle apparatus consists of a traveling microscope with a 40× eyepiece with fine right-angle crosshairs and a longdistance objective. 32K. The procedure of ¨ wetting.00 m distance). The UV extinction of dipyridamole provided a convenient means to obtain an estimate of the surface concentration. using the approach of Noller et al. The spectrum clearly reveals the identity and purity of the product. and the addition of calcium chloride marks the actual start of the experiment. Fourier transformation without application of a spectral window. A bubble of air with a volume of ∼0. Measurements of the surface density of immobilized dipyridamole were performed through UV spectrophotometry. The unmodified and modified surfaces were subjected to the in vitro thrombin generation assay developed by Lindhout et al. the unmodified and modified surfaces were subjected to scanning electron microscopy (SEM) to study the morphology of adhered platelets. The stage contains a plexiglass box. Contact angle measurements were performed according to the captive bubble method (23).. For the quantitative analysis the spectra were corrected for the analyzer transmission function. 1997 299 dried (air fohn). and the plate containing the polyurethane sheet was lowered in the box until the polyurethane sheet was completely immersed. Sensitivity factors were calculated. This reagent is known to effectively cleave the TBDMS groups. The polyurethane sheets were held on the underside of a polycarbonate plate with the modified surface exposed. 6000 Hz. 20 mA) was used. a variable intensity light source. At the used pass energy of 48 eV. The base pressure in the analysis chamber was well below 1 × 10-9 mbar. drying. and poly(vinyl chloride) (PVC). Eindhoven. measured at 1. Then.79 mm2) was dissolved in tetrahydrofuran (4. The modified surfaces (dimensions approximately 50 × 50 mm) showed uniform density of immobilized dipyridamole. polyurethane sheets were mounted on a standard sample holder and inserted via a separately pumped load lock into a Leybold MAX200 XPS instrument. 31). and turned by 90°. which has a high surface thrombogenicity. (30) and described extensively in our previous work (3.67 mm2) was dissolved in 4. A piece of the unmodified polyurethane (138.

This labeling clearly revealed the presence of free hydroxyl groups on the modifed surface. both surfaces were studied with XPS. oxygen. The experimental spectrum can be simulated by adding the simulated spectra i and ii. and carbon species are in good agreement with those of Beamson and Briggs (28). N 1s narrow-scan XPS spectrum of the modified surface measured at a take-off angle of ∼6°. Table 1 compiles the results of the water contact angle measurements on both the unmodified and modified surfaces.3 17 4.3 20.3 Figure 4. For this reason.3 51. which could be derived from changes in the peak shape of the narrow-scan spectra. The sample holder was rotated in such a way that a take-off angle of ∼6° was obtained. using a Gaussian line shape. Figure 2 shows the expansion of the N 1s narrow-scan spectrum measured at a take-off angle of 90°. which equals (i) + (ii).4 82 2. and ref 29). Note that (iii). 8.8 15 5. and O and the elemental ratios C/O and C/N for the unmodified and modified surfaces. Figure 3.. The spectra were analyzed by applying curve deconvolution. see ref 28. we expected the presence of free hydroxyl groups (three OH groups for each immobilized dipyridamole) for the modified surface. As expected.8 79 4. consistent with the presence of another nitrogen functionality in the outermost surface layer (aromatic nitrogen from the dipyridamole). N. Table 2. p 236. The contact angles were measured with the captive-bubble method.5 Aldenhoff et al. The increase of the N 1s signal was not significant. 1997 Table 1.1 mL) were placed. This experiment clearly showed a change in surface chemistry. 3. and carbon were detected. and (ii) corresponds with the N atoms of the polymer urethane groups.7 25. The XPS spectra do not show the very characteristic Si signals at approximately 102 (Si 2p) and 153 (Si 2s) eV. N 1s narrow-scan XPS spectrum of the modified surface measured at a take-off angle of 90°.5 50. nitrogen. but broadened. fits the experimental spectrum. The CF3 tag is easily detected since two extra peaks appear in the XPS spectrum: (i) the F 1s line at approximately 690 eV and (ii) the F Auger line at approximately 600 eV (Mg KR radiation. Spectral simulation was accomplished as described in the legend to Figure 2. Contact Angles Measured with the Captive Bubble Method for the Unmodified and Modified Surfaces unmodified surface 49. The anhydride then reacts with free hydroxyl groups to form the trifluoroacetic ester. The observed binding energies for the oxygen. Overall XPS spectrum of the modified surface measured at a take-off angle of ∼6°. Table 2 gives the atomic percent C. nitrogen. Under each surface. which is consistent with the presence of immobilized and fully deprotected dipyridamole. Vol.2 49. its exposure to tetrafluoroacetic anhydride vapor (takeoff angle of ∼6°). Figure 5 shows the XPS scan of the modified surface after Figure 2.8 22. ments reported here correspond with a surface density of 13 ( 1 nmol/cm2 of immobilized dipyridamole. The contact angle R was then calculated from the formula (23) R ) cos-1[-1 + 2h/d] The data in Table 1 clearly show that the modified surface is more hydrophilic than the unmodified surface. and the nonaromatic Ns of dipyridamole. (i) corresponds with the aromatic N of the dipyridamole ring system. Figure 3 shows the expansion of the N 1s narrow-scan spectrum measured at a take-off angle of ∼6°. . This expectation could be verified by exposure of the modified surface to a vapor of tetrafluoroacetic anhydride in a sealed vessel. Furthermore. Another important conclusion that can be drawn from the XPS spectra of the modified surface is that the Si protective groups were completely removed in the deprotection/washing steps following the irradiation with ultraviolet light. This vapor-phase labeling of hydroxyl groups has been shown to be very successful (29).6 47.300 Bioconjugate Chem. and this inspired us to measure the modified surface also under another take-off angle. five air bubbles (∼0. and height (h) and diameter (d) of the bubbles were measured.3 21. The N 1s narrow-scan subspectrum was used to get more detailed information on the nature of the surface. No contaminations were found on either surface. Elemental Atomic Percentages and Elemental Ratios Measured by XPS for the Unmodified and Modified Surfaces atom %C %N %O C/O C/N modified surface unmodified surface take-off angle 90° take-off angle 90° take-off angle ∼6° 80 3.1 17 4. The N 1s line is less intense. No.5 29. This enables one to look more at the outermost layer of the modified surface.6 18.1 25. Figure 4 shows the overall XPS scan of the modified surface (take-off angle of ∼6°).8 modified surface 19.

. The assay produces a thrombin generation curve for each surface tested. i. The experiments on the modified surface were performed in duplicate. The thrombin generation curves were corrected for the amidolytical activity of the R2m-thrombin complex (32). 75 Figure 8. exposed to human platelet-rich blood plasma. Vol. 3. The results of these experiments with all surfaces are compiled in Table 3.e. No. The unmodified surface was covered to a large extent by activated/spreaded platelets. Figure 6. (9) polyurethane. Figure 7. adhered to the modified surface. the lag time increases from 569 s for the unmodified polyurethane to 1267 and 1282 s (duplicate) for the modified surface. Each thrombin generation curve yields two parameters. This finding is in good agreement with our previous work (3. The maximum free thrombin concentration decreases from 85 nM for the unmodified polyurethane to 68 and 75 nM (duplicate) for the modified surface. Table 3. Biochemical Characterization. a relatively passive material. a highly thrombogenic material. Overview scanning electron micrograph of the modified surface after incubation with platelet-rich plasma. and unmodified polyurethane were used as reference materials. (30). The second parameter is the maximal concentration of free thrombin reached during the experiment. From the scanning electron micrographs. Figure 6 shows the thrombin generation curves for both the unmodified and modified surfaces. The surface immobilization of compound 1 leads to a decrease of the thrombogenicity of the polyurethane. The modified surface showed far less adhered platelets (Figure 8). Overview scanning electron micrograph of the unmodified surface after incubation with platelet-rich plasma. the time elapsing between the moment of recalcification and the moment at which the free thrombin concentration starts to increase. Thrombin generation curves measured for four different surfaces. Results of the Thrombin Generation Test material PE PVC polyurethane modified polyurethane lag time (s) 395 859 569 1267. roughly. (×) modified polyurethane. ([) PVC. Figure 9 shows a detail scanning electron micrograph of an individual platelet. PVC. 1997 301 Figure 5. the scanning electron micrograph indicated a surface density of approximately 6 × 102 platelets/mm2. It should be noted that the maximum thrombin concentrations were obtained after correction of the experimental thrombin generation curves for the residual amidolytic activity of the thrombin-R2-macroglobulin complex according to the method of Hemker et al. The first parameter is the lag time. Assignments: (O) PE. which are directly related to the thrombogenicity of the material. 1 order of magnitude.Photoreactive Derivative of Dipyridamole Bioconjugate Chem. (32). The unmodified and modified surfaces were subjected to the in vitro thrombogenicity test developed by Lindhout et al. PE. 8. A combination of different physical analytical techniques reveals the presence of fully deprotected dipyridamole in the outermost layer of . a surface density of approximately 7 × 103 platelets/mm2 could be estimated. Overall XPS spectrum of the modified surface exposed to trifluoroacetic anhydride vapor measured at a takeoff angle of ∼6°. The thrombogenicity of the different surfaces decreases in the following order: PE > polyurethane > PVC > modified polyurethane. and no spreading. DISCUSSION AND CONCLUDING REMARKS The present results clearly demonstrate that dipyridamole can be immobilized onto polyurethane surfaces through use of conjugate 1. 1282 [thrombin]max (nM) 169 79 85 68. This implies that the surface modification results in a decrease of platelet adhesion by. it should be noted that spread platelets could also be found on the modified surface. 5). The unmodified and modified surfaces were further studied with SEM to look at the adherence and activation of platelets on the surfaces. as is shown in Figure 7. While this particular platelet merely shows the formation of small pseudopodia.

J. Chem..... Orlando. This can be attributed to the short polyethylene glycol spacer chain that bridges dipyridamole to the polymer surface. J. Platz.. Blezer.. Blezer. Upon use of a short hydrophobic spacer chain [-C(O)CH2CH2CH2O-].. (1997) Photo-immobilization of dipyridamole (Persantin) at the surface of polyurethane biomaterials: reduction of in vitro thrombogenicity. On the basis of our experience with photoimmobilization of dipyridamole to polyurethane surfaces.. Scriven.. (10) Vroman. F. 1247-1257. A. 3. and Koole. ¨ Bastiaansen. 243-250. (1992) Chemical methods for improving haemocompatibility of synthetic polymers. To the best of our knowledge. Y. 280-290. van der Veen. (1987) The fate of fibrinogen following adsorption at the blood-biomaterial interface. Koteswar. (16) Leyva. L. (4) Kuijpens. Med. Benzina. H. 99. Am. 8092-8098. G. (3) Aldenhoff. A. Chem. 29. (19) Wu. J. T. R. Kirby. Y. it is known that contact between blood and an artificial surface first leads to adsorption of plasma proteins and then to activation of platelets or leukocytes. K. This graft has an inner diameter of 5 mm and a flexible. 8464-8472. We believe that this combination of effects holds promise with respect to passivation of artificial surfaces. and Koole. Y. George. A. L. prepared with conjugate 1. A.. A. Al-Lamee. Lindhout. 2237-2254. H. T. 21. M. Sci.. Chem. 516. 352-357. D. P. 11 and 15 stereoisomers. FL. J. J.. P. 64. adsorbs far fewer platelets after exposure to human platelet-rich plasma (approximately 7 × 103/mm2) as compared to the control (approximately 6 × 102/mm2). compliant. and Adams.. 110. S. and Koster.. A. M. Reactivity and Utility (E. Ann. as compared to the control (i. 114. J. Sci. 10. J. B.302 Bioconjugate Chem. Bailey. 316. K. (1994) Dipyridamole directly inhibits vascular ´ smooth muscle cell proliferation in vitro and in vivo: implications in the treatment of restenosis after angioplasty. S. J.. From the classic work of Vroman and others (vide supra). Surf. (1984) Azides and Nitrenes. (5) Aldenhoff.. Detailed scanning electron micrograph showing the morphology of a platelet adhered to the modified surface. Bull. Coll.. B. J. 516. 438446. J. G. Gioeli. (17) Li. Chem. Soc. W. N. Am... Bull. B. and Koole. L. J. P.. F.and 3′-carbons in β-Dnucleosides. V. (9) Brash. (8) Vroman. A. Kurtz. the modified surface. (1988) The life of an artificial device in contact with blood: Initial events and their effect on its final state. Soc. 1803-1812. (20) Kohli. J.. P. Biomed. R. (11) Grinell. Pijpers. P.. M. P. Y-Z. and Chattopadhyaya... Soc. (1950) Darstellung aromatischer azidoverbindungen nach der Dutt-Wormallschen reaktion. (1987) Medical intelligence drug therapy dipyridamole. R. L.. Blezer. R. (1969) Findings with the recording ellipsometer suggesting rapid exchange of specific plasma proteins at liquid-solid interfaces. Furthermore. 665-671... 81. A. 7012-7015. N. . Brown. H. and Wirtz. C. 1997 LITERATURE CITED Aldenhoff et al. L. Mater. R.. our thrombin generation assay suggests that photoimmobilization of 1 leads to the best results (15). the modified surface features a markedly increased hydrophilicity.. 970-980. M. (1990) Chemical modification of polymers to increase blood compatibility.. 2683-2686. Kardaun. Ed. J. and Schuster. (ii) the exposure of a platelet inhibitory agent (dipyridamole in our case) suppresses activation of contacting blood platelets. P.. which recognize one or more specific structural patterns of the proteinaceous adsorbates. Chem.. (1) Oates. (1996) Studies on radio-opaque polymeric biomaterials with potential applications to endovascular prostheses. Clin. G. W.. Blocker. Poliakoff. Am.. the present results show for the first time that it is possible to couple a pharmacologically active drug to a polymer surface via a chain of covalent bonds and to simultaneously achieve a substantial increase of the surface hydrophilicity. We did not perform a quantitative comparison of the modified surfaces based on 1 and those to which dipyridamole was coupled either directly or via a short hydrophobic spacer chain. Persy. N. 3783-3790. Belg.. Xi. Ann. Acad. (1995) Immobilization of theophylline on medical-grade polyurethane inhibits surface-induced activation of bloodplatelets. K. G.. unmodified polyurethane). Acad. Biomed. B. J. and Rager. No. (7) Kruft. Vol. (1995) Studies on a new strategy for surface modification of polymeric biomaterials.. and porous wall structure. S. G. Res. However. K. H. the platelets present on the latter surface showed substantial spreading and formation of pseudopodia. R. F. we are currently working on photoimmobilization of conjugate 1 on the lumenal surface of a polyurethane vascular graft. Z. (1992) Total synthesis of (+)-10. 917-928.. R. (1988) 1. 1259-1266. Rothfuss. J. G.e. R.. F. Biomaterials 17. Our results with this new vascular graft (in vitro and in vivo) will be published later. L. L. 8. J. Macromolecules 29. and Koole. (1996) Photochemical coupling of aryl azides to poly(etherurethane) surfaces: studies with a fluorescent model compound. H. N. C. Brune. 28. (15) Witkowski. Tetrahedron Lett. and Koole. Blezer. (1991) Intramolecular cyclization-trapping of carbon radicals by olefins as means to functionalize 2′. Y. Wood. Mater. M. 8691-8697. C. A. Benzina. Lacefield. H. I. Soc. Lindhout. and Fitzgerald. C. Tetrahedron 47.. (1986) Photochemistry of phenyl azide: the role of singlet and triplet phenylnitrene as transient intermediates. H. F. and Koole. G. Formation of small pseudopodia is noted. Bar. Sci.. (6) Kruft. (13) Bamford. H. H. Middleton. Y. B. H. L. 117. J. (14) van der Heiden..... This idea is reinforced by a comparison of the scanning electron micrographs. W.2-Didehydroazepines from the photolysis of substituted aryl azides: analysis of their chemical and physical properties by time-resolved spectroscopic methods. Acad. Mater. (1994) Studies on two new radiopaque polymeric biomaterials.. J. Am. J. Figure 9. and Fried. T. (21) Bretschneider. M.. J. 167-172. Cardiol. 919-925. 108. J.. and Can.. Res.. L. C. (2) Singh. Figure 8 (this work) shows that the modified polyurethane surface. H. H.) Academic Press. and Al-Lamee. Chim. Monatsh. 23. M.. Ramiya. Biomaterials 18. and Dube. (12) Bamford. J. E. J. Nevertheless. A. Soc. B. we previously found that platelet adhesion was also markedly suppressed (in comparison with the control) (3). Wiernicki. (1987) Fibronectin adsorption on material surfaces. Med. Haluska.. H. (1980) The triphen¨ ¨ ylmethyl (trityl) group and its use in nucleotide chemistry. H. H. 206-222. A modified surface as studied in our present work interferes with this mechanism at two different stages: (i) the hydrophilic nature of the surface leads to substantial hydration and hence to a diminished affinity for proteins. A. P. Am. A. W. L. L. V. V. (18) Smith. L. A.10-difluorothromboxane A2 and its 9. 16. Engl.

(1974) ¨ A step towards quantitative electron spectroscopy measurements by improved electron optics.. 94.. Mattern-Klosson. Electron Spectrosc. Andrade.. (1986) Thrombin generation and inactivation in the presence of antithrombin III and heparin. Chem. and Briggs. (27) Scofield. C. (Am. H.. Interface Anal. J. (31) Lindhout. Phenom. G.. M. (1986) A ´ computer assisted method to obtain the prothrombin activation velocity in whole plasma independent of thrombin decay processes. G. P. 8.. M. Relat. H. 341. P. and Schillalies. Ward. Div. Willems. D. Short. Chem. 14. (1989) Post-1989 calibration energies for X-ray photoelectron spectrometers and the 1990 Josephson constant.Photoreactive Derivative of Dipyridamole (22) Corey. R. Haemostasis 56. (1992) High Resolution XPS of Organic Polymers... Chem. D. (24) Seah. M. (1976) Hartree-slater subshell photoionization cross-sections at 1254 and 1487 eV. 1795-1799. K. J. 8. R.. J.. (32) Hemker. J. Biochemistry 25. P. J. Bioconjugate Chem. Wiley. D. Maassen. 9-17. 121-124. H. 5. and Brostrom.K. (28) Beamson. Relat. V. M. A. (1995) Interfacial tensions at acrylic hydrogel-water interfaces. 103. Polaschegg.. Baruch. Prepr. Mater... Beamson. 3.. and Wilmers. and Hemker.. (23) King. Franssen. G. Gregonis. Blezer. U. and Briggs. R. No. and Beguin. A. (25) Berresheim. Med. D. 705-723. Polym. C. 6. H. C. J. Chichester. (26) Noller. and Venkateswarly. H.. T. D. Chem. Soc. Colloid Interface Sci. S.) 34. Vol. 129-137. Schoen. J.. J. C. G. Sci. D. (30) Lindhout.. S.. M. Am. Ma. Thromb. Poly. R. E. Heijnen. Anal. Fresenius’ J. (1991) A standard form spectra for quantitative ESCAanalysis.. E. N. H. 5962-5970. J. P. D. Phenom. 1997 303 (29) Ameen..: Mater.. T. (1993) A high-resolution X-ray photoelectron spectroscopy study of trifluoroacetic anhydride labelling of hydroxyl groups: demonstration of the β shift due to -OC(O)CF3. Surf.. and Reutelingsperger. (1995) Platelet procoagulant surface as an essential parameter for the in vitro evaluation of the blood compatibility of polymers. Soc. J.. 6190-6191. R. BC970020Z . (1972) Protection of hydroxyl groups as tert-butyldimethylsilyl derivatives. L. Electron Spectrosc.. 6275. 367-372. 488.