- Email: [email protected]
Vascular prostheses with controlled release of antibiotics
Biomolecular Engineering 24 (2007) 149–153 www.elsevier.com/locate/geneanabioeng
Vascular prostheses with controlled release of antibiotics Part 1: Surface modification with cyclodextrins of PET prostheses N. Blanchemain a,c, S. Haulon a,b, F. Boschin a,c, E. Marcon-Bachari d, M. Traisnel e, M. Morcellet c, H.F. Hildebrand a,*, B. Martel c a
Groupe de Recherche sur les Biomate´riaux, EA 1049, Faculte´ de Me´decine, 59045 Lille, France b Service Chirurgie Vasculaire, Hoˆpital Cardiologique, CHRU, Lille, France c Laboratoire Chimie Organique et Macromole´culaire, CNRS, UMR 8009, USTL, 59655 Villeneuve D’Ascq, France d Laboratoires Pe´rouse, Route du Manoir, 60173 Ivry-Le-Temple e Laboratoire de Proce´de´s d’Elaboration de Reveˆtements Fonctionnels, UPRES EA 1040, ENSCL, 59655 Villeneuve D’Ascq, France
Abstract Vascular prostheses were functionalised with the aim to obtain a slow release of antibiotics in order to reduce postoperative infections. The original process that we present in this paper is based on the use of a family of cage molecules named cyclodextrins (CD). These compounds have the ability to form reversible inclusion complexes with drugs such as antibiotics. The aim of this work was to graft CD onto the prosthesis, so that an antibiotic can be bound on it by this inclusion phenomenon, and then be progressively released over a prolonged period by a complex dissociation mechanism. This paper presents the first part of this research program and concerns mainly the study of the functionalization parameters. It presents surface characterization results of the modified prostheses. The PET prostheses were immersed into a solution containing a cross linking agent, cyclodextrins (b-CD, g-CD, HP-b-CD and HP-g-CD) and a catalyst and were padded. Grafting occurred by the mean of a thermofixation step at a temperature comprised between 140 and 180 8C. It was observed that the support was permanently modified when the CD polymer that coated the fibres resisted to the final washing process. Grafting rates of 12 wt% in CD polymer could be reached. It was also observed that the fibre coating reaction induced an increase of the permeability of the grafts. # 2006 Elsevier B.V. All rights reserved. Keywords: PET vascular prosthesis; Cyclodextrins; Citric acid; Grafting; Wettability; Permeability
1. Introduction PET vascular prostheses (Dacron1) are used since the early 1960s (Blanchemain et al., 2005). They can replace or bypass damaged arteries in patients with peripheral arterial disease or aneurysms. The main risk factors for vascular disease are smoking and high blood pressure (Krupski et al., 2002). Vascular prosthesis infections are associated with a high mortality rate: 40% for an aortic localisation (Bandyk, 1990), and 10% for an infrainguinal localisation, associated with amputation in 33% of patients (Kieffer et al., 1993). Prostheses infections represent one person per year and per 100,000 residents. It concerns about 350 persons per year in France (O’Brien and Collin, 1992), and represents 6% of the annual surgical procedures (Walkefield et al., 1990).
* Corresponding author. Tel.: +33 320 626975; fax: +33 320 626854. E-mail address: [email protected] (H.F. Hildebrand). 1389-0344/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bioeng.2006.05.012
The infection onset is often secondary to a per-operative contamination, which may also occur late after the initial procedure (redo-operation, puncture, bacteraemia, etc.). The bacterium involved colonizes an empty space around the prostheses in which they develop a bio-film. Protecting the bacterium from antibiotics and immune cells (Chistensen et al., 1990). The two types of infections; early and late, are associated with different types of germs: the former with S. aureus and gram negatives, and the later with S. epidermidis (Bandyk and Esses, 1994). In order to reduce graft infection, we developed a prosthesis that delivers antibiotics during the per-operative risk period mentioned above. This concept is based on the use of cyclodextrins (CDs) that are torus shaped cyclic oligosaccharides issued from enzymatic degradation of starch. They are principally made of six, seven or eight glucopyranosic units and are, respectively, called a-, b- and g-cyclodextrin. The hydrophobic internal cavity and the hydrophilic external part
150
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
of the molecular structure have antagonist characteristics that provide CD their original and interesting properties (Martel et al., 2005a). On the one hand, the interior of the ring shaped molecule has the role of a receptor towards a large variety of guest molecules, to form a ‘‘host–guest’’ compounds. This reaction is based on the equilibrium between the complexed and free form of the molecule (Buschmann et al., 2001). The stability of such supramolecular species mainly depends on the shape and size compatibility between the host and the guest. As a consequence, several different CDs and derivatives of CDs will be used in the present study in order to investigate the most adapted antibiotic-CD couples. In addition, it is known that CDs have the ability to interact with some antiseptics like chlorhexidine (Hong et al., 1994), or antibiotics like ciprofloxacine (Chao et al., 2002). It has been reported that the use of such complexes always results in a delayed drug delivery (Loftsson and Loftson, 1996; Rajewski and Stella, 1996; Tetsumi and Kaneto, 1997). The numerous hydroxyl groups present on the outer surface of the CDs offer water solubility, and possibilities of chemical functionalization such as polymerisation, and anchoring onto various polymeric supports including textile fibres. The applications of CDs in the textile domain have risen since the early 1990s and are at the origin of a new generation of materials with interesting properties in a very large range of applications (Le Thuaut et al., 2000; Martel et al., 2002a,b). The grafting method of CDs onto ‘‘classic’’ textile PET material was developed in our laboratory and was previously reported (Weltrowski et al., 2001; Martel et al., 2002c). The reaction consisted of the coating of the fibres by a CDs crosslinked polymer network that physically adhered to the fibres. The present paper describes the same process adapted to a tubular vascular PET prosthesis, to obtain a new type of drugdelivery system (Martel et al., 2005b). Four different CDs were tested: b and g-CD called ‘‘native CDs’’, and their respective hydoxypropylated derivatives, HP-b-CD and HP-g-CD. a-CD and its derivatives were not considered in the present work because preliminary investigations showed that they were not compatible with the use of bulky antibiotic molecules. 2. Materials and methods Virgin polyester vascular prostheses (POLYTHESE1, Laboratoires PEROUSE SAS, France) were truncated in tubular pieces into length of 25 cm. The diameters of crimped tubes were 10 and 26 mm. Their textile properties were characterized as woven PET yarns with a nominal linear density of 100 dtex and by a mass per unit area of 133 g m2. The characteristic expresses the coarseness of the yarn in units of decitex, i.e., mass per unit length. In the metric system, 10,000 m of 1 dtex yarn weighs 1 g. g-CDs and hydoxylpropyl-g-CDs (HPg-CDs) were provided by Wacker Specialties GmbH (Cavamax1, Burhausen, Germany), b-CD and hydroxylpropyl-b-CD (HPb-CD) by Roquette (Kleptose1, Lestrem, France). Citric acid and sodium dihydrogen hypophosphite were Aldrich chemicals (Milwaukee, Wi, USA). The textile finishing equipment consisted of a padder and a thermofixation oven equipped with a pin frame (Minithermo, Roaches, Leek, UK). The grafting process was based on the pad-dry-cure textile finishing method: prostheses were impregnated by an aqueous solution containing CDs, catalyst and citric acid (CTR) and were roll-squeezed. Grafting occurred
in a thermofixation oven at a curing temperature comprised between 140 and 190 8C, during a variable time. The modified prostheses were finally rinsed with warm water and were then submitted to successive soxhlet extractions with hexane, ethanol and water to eliminate eventual traces of auxiliaries and unreacted products. The amount of grafted CD was measured according to the weight increase of the prostheses by using the following equation: wt% ¼
mi mf 100 mi
mi and mf were the weight of the samples before and after the treatment, respectively. The precision on the weight gain measurements was 1.5 wt%. The hydrophobic character of untreated and modified PET was determined by the drop contact angle method (Digidrop goniometre, GBX): a water drop of 5 ml was positioned onto the surface of the prosthesis; the contact angle was measured after 2, 5, and 10 s. Each measurement was repeated 10 times. The working temperature was 37 8C. The number of millilitres of filtered water that passes through an integral material under a pressure head 130 mmHg during 5 min is measured. The integral water permeability is also defined in ml cm2 min1. This property was measured for each type of grafted prosthesis, in triplicate following the procedure outlined in the ISO 7198. Bursting strength measurements were performed for each prosthesis using a dynamometer (Adamel Lhomargy). The sample prostheses (diameter 16 and 26 mm) are tested for longitudinal and circumferential tensile strengths. To test longitudinal tensile strength, the tubular sample prosthesis is placed with its ends in suitable jaws. It is then stretched at uniform rate of 100 mm min1 until the yield and/or break point is reached. To test circumferential tensile strength, the tubular sample prosthesis is placed onto two round pins. It is then stretched at a uniform rate until the yield and/or break point is reached. The circumferential tensile strength of each sample is expressed as N cm1, by dividing the maximum load (F r) by the original length of the sample. The percentage of dilatation is expressed as followed: dilatation (%) = 100 Ar/pi D/2 with D, diameter of sample; Ar, lengthening. The surface morphologies of the prostheses were observed in a JEOL JSM 5300 scanning electron microscope at an accelerating voltage of 25 kV and a current of 100 mA.
3. Results The investigations started with the assessment of the influence of the temperature applied for the polymerisation and the fixation of CDs onto the PET support. The prostheses were impregnated with a solution containing one of the CDs, the cross linking agent (CTR) and the catalyst (NaH2PO2). They then were cured during 10 min at variable temperatures. As shown in Fig. 1, a threshold temperature of 140 8C had to be applied for both native CDs in order to detect grafting. The threshold grafting temperature occurred below 140 8C with HPb-CD and HP-g-CD. The maximal grafting rate was above
Fig. 1. Grafting rate according to curing temperature for b, g, HP-b, and HP-g CDs.
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
Fig. 2. Grafting rate vs. time of curing for HPb-CD at 140, 150, 160 and 170 8C.
160 8C for native CDs, and close to 150 8C for their hydroxypropyl derivates. Fig. 2 reports the influence of the curing time, with temperatures between 140 and 170 8C. All the curves show that a minimum time of curing had to be applied in order to observe the grafting reaction. This threshold duration generally decreased with increasing temperature. For both native CDs, grafting varied with the time of fixation, but did not reach any maximum plateau value at 60 min with 140 8C. In contrast, at 150 8C and more, a maximum grafting rate was rapidly reached. The plateau was obtained after 20 min at 150 8C, 15 min at 160 8C, and 5 min at 170 8C. Concerning both HPCDs derivatives, the plateau was observed at 140 8C and the influence of the temperature was less remarkable with respect to native CDs. The plateau values were reached within 5– 10 min. Investigating the time and temperature parameters it is necessary to settle the conditions of reaction to be applied in order to reach a chosen grafting rate. Considering the results mentioned above, the best compromise depicted was a fixation time of 20 min, and temperatures of 160 and 150 8C for the native CDs and the HP-CDs, respectively. After the assessment of the optimal time and temperatures of curing, we investigated the influence of the concentration of the impregnating bath. The initial solution was fixed at 16 g L1 in CTR, 1 g L1 in catalyst, and 20 g L1 in CD, and 3C/4, C/2 and C/4 solutions were prepared by dilution. Fig. 3 shows that the grafting rate of the prostheses was almost proportional to the bath concentration.
151
Fig. 3. Grafting rate vs. concentration of CD for b- and g-CD at 160 8C for 20 min and HP-b and HP-g-CD at 150 8C for 20 min.
This linear relationship allows obtaining a desired grafting rate only by adjusting the concentration of the bath, while maintaining the determined curing parameters. This experimental approach should offer a superior repeatability of the method. Scanning electron microscopy (SEM) was used to analyse the textile surface before and after grafting of CD. Fig. 4 shows the woven fibres of PET, before and after grafting, respectively. The fibres of the raw material present a smooth aspect, while sort of irregular shells are visible on the modified ones. As we previously reported, this is due to a CD polymer that adhered to the fibres only by physical interactions (Martel et al., 2002c). The permeability of the woven vascular prosthesis is a fundamental aspect during the implantation procedure. Blood flows inside the prostheses at a certain pressure and should not percolate through the textile structure. In order to avoid haemorrhage, the manufacturers usually impregnate the prostheses with collagen or albumin to reduce the permeability of the textile. This aspect was considered and we measured the permeability of the raw and modified prostheses as reported in Fig. 5. The permeability of the modified prostheses was compared to raw and collagen-impregnated prostheses. The latter was almost impermeable (0.01 mL min1 cm2), while the uncoated prosthesis permeability reached 300 units. Fig. 5 shows that the permeability strongly decreases (40%) when the grafting rate is only 5 wt%. The grafting rate itself had no influence on the permeability. The permeability of the collagen-impregnated prostheses was not equalled, but the
Fig. 4. Woven fibre of PET before grafting (a) and after a 10%-grafting with HP-b-CD (b and c).
152
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
Fig. 6. Water contact angle at 2, 5, and 10 s. Fig. 5. Water permeability of the PET prosthesis according to the grafting rate, compared to collagen coated and uncoated PET prosthesis.
grafting reaction significantly decreased the water permeability of the PET prosthesis. In order to obtain total waterproof of vascular prostheses, one single coat of collagen is sufficient instead of two coatings applied in the usual manufacturing process. The surface analysis of the prostheses was then completed by wettability measurements using the method of the drop contact angle. Fig. 6 reports the evolution of the water drop at 5, 10 and 20 s on untreated and grafted prosthesis with b-CD, gCD, HPb-CD and HPg-CD, applying a grafting rate of 10 wt%. A rapid decrease of the drop contact angle is observed on the untreated PET, so that the drop was almost absorbed within 10 s. In contrast, CD-coated prostheses (10 wt%) presented an initial angle of 1208 that remained constant with time. This result shows that the coating reaction involved a modification of the surface properties of the prostheses that became more hydrophobic and watertight. The cyclodextrin grafting does not affect mechanical properties of vascular prostheses (Table 1). A slightly higher resistance is observed with respect to virgin prostheses for longitudinal traction, respectively, 255 and 220 N cm1 and for circumferential traction, respectively, 250 and 222 N cm1. The grafting rate and the temperature of fixation do not provoke any alteration of the mechanicals properties.
4. Discussion The first part of the study investigated the parameters of the grafting reaction. It was observed that grafting was time and temperature dependent. It has already been reported that the synthesis of CTR–CD polymers obtained from the same reactants could result in water-soluble products or insoluble gels able to swell in water (Martel et al., 2005a). The obtained form also depends on the time and the temperature of the reaction. It is dependant on the degree of polymerisation and on the cross-linking rate of the polymer formed between CD and CTR. The same explanation can be proposed in the present study, where the same reaction occurred directly onto the textile support. High temperatures or long time of curing involved high molecular weights and crosslinking rates of the polymer, and resulted in the permanent coating of the fibres. Under milder conditions, however, the CTR–CD polymer had a low degree of polymerisation and a low cross-linking degree; therefore, it remained water soluble, and was removed by the soxhlet extraction. Different optimal parameters can be settled for the CDs and for the hydroxypropyl derivates. It was observed that the hydroxypropyl CD derivatives reacted more readily than their parent molecules, as they presented lower threshold temperature and time of grafting. In fact, the hydroxypropyl groups are substituents that enhance the reactivity of the CD because they act as spacer arms, the hydroxyl group of which is more accessible to the cross linking agent than the hydroxyl groups [– CH2–CH(OH)–CH3] present directly on the glucopyranosic units of CD.
Table 1 Mechanicals assays on vascular prostheses untreated and grafted by HP-b-CD at 10 and 5 wt% with different thermal temperature Grafting type
HPb-CD 5 wt% 140 8C HPb-CD 5 wt% 150 8C HPb-CD 5 wt% 160 8C HPb-CD 10 wt% 140 8C HPb-CD 10 wt% 150 8C HPb-CD 10 wt% 160 8C Untreated Fr, radial force, Ful, forec per length unit.
Longitudinal traction
Circumferential traction 1
Fr (N)
Fr(pD) (N cm1)
Ful (N cm1)
Dilatation (%)
1183 57 1234 121 1295 85 1337 30 1315 41 1281 34 1173 150
235 11 245 24 258 17 266 6 262 8 255 7 220 28
249 47 248 47 251 48 257 49 278 20 261 29 222 17
47 2 47 3 48 2 49 1 50 3 50 5 43 2
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
This experimental approach allowed us to evaluate the correlation existing between time and temperature of curing. We could determine the optimal compromise to obtain the maximal grafting rate by application of the lowest possible reaction time and temperature. This will avoid the alteration of the mechanical properties of the grafts and is also of economical interest. Moreover, the experiment showed a decrease by four of the permeability of the prostheses after coating. This can be correlated with the SEM pictures that showed that that the grafting reaction increased the diameter of the fibres. The consequence was that the free volume between the yarns in the warp/filling arrangement of the woven structure was partially filled and thus the porosity was significantly reduced. Furthermore, the surface wettability measurements also depicted an increased hydrophobia of the modified prostheses. However, in terms of water permeability, the performances of collagen-impregnated grafts were not reached. The CD coating process is probably not completely blood proof and therefore cannot be used in substitution to collagen. Finally, mechanical tests show a slightly higher resistance of vascular grafted prostheses compared to virgin prostheses, so they are more resistant and may have longer life duration, which is another advantage for the replacement of damaged artery. 5. Conclusion The present work allowed us to confirm the possibility of grafting CDs onto woven PET vascular prostheses using citric acid as a co-reactant, and also allowed to settle the parameters of the process adapted to this specific shape of textile support. It was observed that the reaction induced a change of the physical properties of the prosthesis surface, as the fibres were found to be more hydrophobic and the water permeability significantly decreased, in spite of that, the mechanical properties were not affected. Forthcoming investigations will present the results of the biological tests and the in vitro study of the kinetics of release of several antibiotics.
153
Acknowledgements We are deeply indebted to Laboratoires Pe´rouse (60173 Ivry-Le-Temple) for financial support and the supplying of PET prostheses. We thank Annie Lefe`vre for her skilful and expert technical assistance. This project was also financially supported by the Conseil Re´gional Nord-Pas-de-Calais: ARCir ASBAMED and Fe´de´ration Biomate´riaux. References Bandyk, D.B., 1990. Semin. Vasc. Surg. 3, 122. Bandyk, D.F., Esses, G.E., 1994. Surg. Clin. North Am. 74, 571. Blanchemain, N., Haulon, S., Martel, B., Traisnel, M., Morcellet, M., Hildebrand, H.F., 2005. Eur. J. Vasc. Endovasc. Surg. 29, 628. Buschmann, H.J., Knittel, D., Schollmeyer, E., 2001. J. Incl. Phenom. Macrocycl. Chem. 40, 169. Chao, J.B., Chen, L., Xu, H., Meng, D.P., 2002. Spectrochem. Acta Part A 58, 2809. Chistensen, G.D., Baddour, L.M., Madison, B.M., 1990. J. Infect. Dis. 161, 1153. Hong, K., Nishihata, T., Rytting, J.H., 1994. Pharm. Res. 11, 88. Kieffer, E., Bahnini, A., Koskas, F., 1993. J. Vasc. Surg. 17, 349. Krupski, W.C., Nehler, M.R., Whitehill, T.A., 2002. Cardiovasc. Surg. 10, 415. Le Thuaut, P., Martel, B., Crini, G., Maschke, U., Coqueret, X., Morcellet, M., 2000. J. Appl. Polym. Sci. 77, 2118. Loftsson, T., Loftson, M.E., 1996. J. Pharm. Sci. 85, 1017. Martel, B., Le Thuaut, P., Bertini, S., Crini, G., Bacquet-Torri, M.G., Morcellet, M., 2002a. J. Appl. Polym. Sci. 85, 1771. Martel, B., Morcellet, M., Ruffin, D., Vinet, F., Weltrowski, M., 2002b. J. Inclusion Phenom. Macromol. Chem. 44, 439. Martel, B., Morcellet, M., Ruffin, D., Ducoroy, L., Weltrowski, M., 2002c. J. Inclusion Phenom. Macromol. Chem. 44, 443. Martel, B., Ruffin, D., Weltrowski, M., Lekchiri, Y., Morcellet, M., 2005a. J. Appl. Polym. Sci. 97, 433. Martel, B., Blanchemain, N., Morcellet, M., Hildebrand, H.F., Haulon, S., Boschin, F., Delcourt-Debruyne, E., 2005b. Patent application PCTFR 20050022829. O’Brien, T., Collin, J., 1992. Br. J. Surg. 79, 1262. Rajewski, R.A., Stella, V.J., 1996. J. Pharm. Sci. 85, 1142. Tetsumi, I., Kaneto, U., 1997. J. Pharm. Sci. 86, 147. Walkefield, T.W., Pierson, C.L., Schaberg, D.R., 1990. J. Vasc. Surg. 11, 624. Weltrowski, M., Morcellet, M., Martel, B. Patent PCT/FR00/00377, EP 1157156 (2000); US 09/913,448 (2001); CA 2,362,534; WO 0047811.
Vascular prostheses with controlled release of antibiotics Part 1: Surface modification with cyclodextrins of PET prostheses N. Blanchemain a,c, S. Haulon a,b, F. Boschin a,c, E. Marcon-Bachari d, M. Traisnel e, M. Morcellet c, H.F. Hildebrand a,*, B. Martel c a
Groupe de Recherche sur les Biomate´riaux, EA 1049, Faculte´ de Me´decine, 59045 Lille, France b Service Chirurgie Vasculaire, Hoˆpital Cardiologique, CHRU, Lille, France c Laboratoire Chimie Organique et Macromole´culaire, CNRS, UMR 8009, USTL, 59655 Villeneuve D’Ascq, France d Laboratoires Pe´rouse, Route du Manoir, 60173 Ivry-Le-Temple e Laboratoire de Proce´de´s d’Elaboration de Reveˆtements Fonctionnels, UPRES EA 1040, ENSCL, 59655 Villeneuve D’Ascq, France
Abstract Vascular prostheses were functionalised with the aim to obtain a slow release of antibiotics in order to reduce postoperative infections. The original process that we present in this paper is based on the use of a family of cage molecules named cyclodextrins (CD). These compounds have the ability to form reversible inclusion complexes with drugs such as antibiotics. The aim of this work was to graft CD onto the prosthesis, so that an antibiotic can be bound on it by this inclusion phenomenon, and then be progressively released over a prolonged period by a complex dissociation mechanism. This paper presents the first part of this research program and concerns mainly the study of the functionalization parameters. It presents surface characterization results of the modified prostheses. The PET prostheses were immersed into a solution containing a cross linking agent, cyclodextrins (b-CD, g-CD, HP-b-CD and HP-g-CD) and a catalyst and were padded. Grafting occurred by the mean of a thermofixation step at a temperature comprised between 140 and 180 8C. It was observed that the support was permanently modified when the CD polymer that coated the fibres resisted to the final washing process. Grafting rates of 12 wt% in CD polymer could be reached. It was also observed that the fibre coating reaction induced an increase of the permeability of the grafts. # 2006 Elsevier B.V. All rights reserved. Keywords: PET vascular prosthesis; Cyclodextrins; Citric acid; Grafting; Wettability; Permeability
1. Introduction PET vascular prostheses (Dacron1) are used since the early 1960s (Blanchemain et al., 2005). They can replace or bypass damaged arteries in patients with peripheral arterial disease or aneurysms. The main risk factors for vascular disease are smoking and high blood pressure (Krupski et al., 2002). Vascular prosthesis infections are associated with a high mortality rate: 40% for an aortic localisation (Bandyk, 1990), and 10% for an infrainguinal localisation, associated with amputation in 33% of patients (Kieffer et al., 1993). Prostheses infections represent one person per year and per 100,000 residents. It concerns about 350 persons per year in France (O’Brien and Collin, 1992), and represents 6% of the annual surgical procedures (Walkefield et al., 1990).
* Corresponding author. Tel.: +33 320 626975; fax: +33 320 626854. E-mail address: [email protected] (H.F. Hildebrand). 1389-0344/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bioeng.2006.05.012
The infection onset is often secondary to a per-operative contamination, which may also occur late after the initial procedure (redo-operation, puncture, bacteraemia, etc.). The bacterium involved colonizes an empty space around the prostheses in which they develop a bio-film. Protecting the bacterium from antibiotics and immune cells (Chistensen et al., 1990). The two types of infections; early and late, are associated with different types of germs: the former with S. aureus and gram negatives, and the later with S. epidermidis (Bandyk and Esses, 1994). In order to reduce graft infection, we developed a prosthesis that delivers antibiotics during the per-operative risk period mentioned above. This concept is based on the use of cyclodextrins (CDs) that are torus shaped cyclic oligosaccharides issued from enzymatic degradation of starch. They are principally made of six, seven or eight glucopyranosic units and are, respectively, called a-, b- and g-cyclodextrin. The hydrophobic internal cavity and the hydrophilic external part
150
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
of the molecular structure have antagonist characteristics that provide CD their original and interesting properties (Martel et al., 2005a). On the one hand, the interior of the ring shaped molecule has the role of a receptor towards a large variety of guest molecules, to form a ‘‘host–guest’’ compounds. This reaction is based on the equilibrium between the complexed and free form of the molecule (Buschmann et al., 2001). The stability of such supramolecular species mainly depends on the shape and size compatibility between the host and the guest. As a consequence, several different CDs and derivatives of CDs will be used in the present study in order to investigate the most adapted antibiotic-CD couples. In addition, it is known that CDs have the ability to interact with some antiseptics like chlorhexidine (Hong et al., 1994), or antibiotics like ciprofloxacine (Chao et al., 2002). It has been reported that the use of such complexes always results in a delayed drug delivery (Loftsson and Loftson, 1996; Rajewski and Stella, 1996; Tetsumi and Kaneto, 1997). The numerous hydroxyl groups present on the outer surface of the CDs offer water solubility, and possibilities of chemical functionalization such as polymerisation, and anchoring onto various polymeric supports including textile fibres. The applications of CDs in the textile domain have risen since the early 1990s and are at the origin of a new generation of materials with interesting properties in a very large range of applications (Le Thuaut et al., 2000; Martel et al., 2002a,b). The grafting method of CDs onto ‘‘classic’’ textile PET material was developed in our laboratory and was previously reported (Weltrowski et al., 2001; Martel et al., 2002c). The reaction consisted of the coating of the fibres by a CDs crosslinked polymer network that physically adhered to the fibres. The present paper describes the same process adapted to a tubular vascular PET prosthesis, to obtain a new type of drugdelivery system (Martel et al., 2005b). Four different CDs were tested: b and g-CD called ‘‘native CDs’’, and their respective hydoxypropylated derivatives, HP-b-CD and HP-g-CD. a-CD and its derivatives were not considered in the present work because preliminary investigations showed that they were not compatible with the use of bulky antibiotic molecules. 2. Materials and methods Virgin polyester vascular prostheses (POLYTHESE1, Laboratoires PEROUSE SAS, France) were truncated in tubular pieces into length of 25 cm. The diameters of crimped tubes were 10 and 26 mm. Their textile properties were characterized as woven PET yarns with a nominal linear density of 100 dtex and by a mass per unit area of 133 g m2. The characteristic expresses the coarseness of the yarn in units of decitex, i.e., mass per unit length. In the metric system, 10,000 m of 1 dtex yarn weighs 1 g. g-CDs and hydoxylpropyl-g-CDs (HPg-CDs) were provided by Wacker Specialties GmbH (Cavamax1, Burhausen, Germany), b-CD and hydroxylpropyl-b-CD (HPb-CD) by Roquette (Kleptose1, Lestrem, France). Citric acid and sodium dihydrogen hypophosphite were Aldrich chemicals (Milwaukee, Wi, USA). The textile finishing equipment consisted of a padder and a thermofixation oven equipped with a pin frame (Minithermo, Roaches, Leek, UK). The grafting process was based on the pad-dry-cure textile finishing method: prostheses were impregnated by an aqueous solution containing CDs, catalyst and citric acid (CTR) and were roll-squeezed. Grafting occurred
in a thermofixation oven at a curing temperature comprised between 140 and 190 8C, during a variable time. The modified prostheses were finally rinsed with warm water and were then submitted to successive soxhlet extractions with hexane, ethanol and water to eliminate eventual traces of auxiliaries and unreacted products. The amount of grafted CD was measured according to the weight increase of the prostheses by using the following equation: wt% ¼
mi mf 100 mi
mi and mf were the weight of the samples before and after the treatment, respectively. The precision on the weight gain measurements was 1.5 wt%. The hydrophobic character of untreated and modified PET was determined by the drop contact angle method (Digidrop goniometre, GBX): a water drop of 5 ml was positioned onto the surface of the prosthesis; the contact angle was measured after 2, 5, and 10 s. Each measurement was repeated 10 times. The working temperature was 37 8C. The number of millilitres of filtered water that passes through an integral material under a pressure head 130 mmHg during 5 min is measured. The integral water permeability is also defined in ml cm2 min1. This property was measured for each type of grafted prosthesis, in triplicate following the procedure outlined in the ISO 7198. Bursting strength measurements were performed for each prosthesis using a dynamometer (Adamel Lhomargy). The sample prostheses (diameter 16 and 26 mm) are tested for longitudinal and circumferential tensile strengths. To test longitudinal tensile strength, the tubular sample prosthesis is placed with its ends in suitable jaws. It is then stretched at uniform rate of 100 mm min1 until the yield and/or break point is reached. To test circumferential tensile strength, the tubular sample prosthesis is placed onto two round pins. It is then stretched at a uniform rate until the yield and/or break point is reached. The circumferential tensile strength of each sample is expressed as N cm1, by dividing the maximum load (F r) by the original length of the sample. The percentage of dilatation is expressed as followed: dilatation (%) = 100 Ar/pi D/2 with D, diameter of sample; Ar, lengthening. The surface morphologies of the prostheses were observed in a JEOL JSM 5300 scanning electron microscope at an accelerating voltage of 25 kV and a current of 100 mA.
3. Results The investigations started with the assessment of the influence of the temperature applied for the polymerisation and the fixation of CDs onto the PET support. The prostheses were impregnated with a solution containing one of the CDs, the cross linking agent (CTR) and the catalyst (NaH2PO2). They then were cured during 10 min at variable temperatures. As shown in Fig. 1, a threshold temperature of 140 8C had to be applied for both native CDs in order to detect grafting. The threshold grafting temperature occurred below 140 8C with HPb-CD and HP-g-CD. The maximal grafting rate was above
Fig. 1. Grafting rate according to curing temperature for b, g, HP-b, and HP-g CDs.
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
Fig. 2. Grafting rate vs. time of curing for HPb-CD at 140, 150, 160 and 170 8C.
160 8C for native CDs, and close to 150 8C for their hydroxypropyl derivates. Fig. 2 reports the influence of the curing time, with temperatures between 140 and 170 8C. All the curves show that a minimum time of curing had to be applied in order to observe the grafting reaction. This threshold duration generally decreased with increasing temperature. For both native CDs, grafting varied with the time of fixation, but did not reach any maximum plateau value at 60 min with 140 8C. In contrast, at 150 8C and more, a maximum grafting rate was rapidly reached. The plateau was obtained after 20 min at 150 8C, 15 min at 160 8C, and 5 min at 170 8C. Concerning both HPCDs derivatives, the plateau was observed at 140 8C and the influence of the temperature was less remarkable with respect to native CDs. The plateau values were reached within 5– 10 min. Investigating the time and temperature parameters it is necessary to settle the conditions of reaction to be applied in order to reach a chosen grafting rate. Considering the results mentioned above, the best compromise depicted was a fixation time of 20 min, and temperatures of 160 and 150 8C for the native CDs and the HP-CDs, respectively. After the assessment of the optimal time and temperatures of curing, we investigated the influence of the concentration of the impregnating bath. The initial solution was fixed at 16 g L1 in CTR, 1 g L1 in catalyst, and 20 g L1 in CD, and 3C/4, C/2 and C/4 solutions were prepared by dilution. Fig. 3 shows that the grafting rate of the prostheses was almost proportional to the bath concentration.
151
Fig. 3. Grafting rate vs. concentration of CD for b- and g-CD at 160 8C for 20 min and HP-b and HP-g-CD at 150 8C for 20 min.
This linear relationship allows obtaining a desired grafting rate only by adjusting the concentration of the bath, while maintaining the determined curing parameters. This experimental approach should offer a superior repeatability of the method. Scanning electron microscopy (SEM) was used to analyse the textile surface before and after grafting of CD. Fig. 4 shows the woven fibres of PET, before and after grafting, respectively. The fibres of the raw material present a smooth aspect, while sort of irregular shells are visible on the modified ones. As we previously reported, this is due to a CD polymer that adhered to the fibres only by physical interactions (Martel et al., 2002c). The permeability of the woven vascular prosthesis is a fundamental aspect during the implantation procedure. Blood flows inside the prostheses at a certain pressure and should not percolate through the textile structure. In order to avoid haemorrhage, the manufacturers usually impregnate the prostheses with collagen or albumin to reduce the permeability of the textile. This aspect was considered and we measured the permeability of the raw and modified prostheses as reported in Fig. 5. The permeability of the modified prostheses was compared to raw and collagen-impregnated prostheses. The latter was almost impermeable (0.01 mL min1 cm2), while the uncoated prosthesis permeability reached 300 units. Fig. 5 shows that the permeability strongly decreases (40%) when the grafting rate is only 5 wt%. The grafting rate itself had no influence on the permeability. The permeability of the collagen-impregnated prostheses was not equalled, but the
Fig. 4. Woven fibre of PET before grafting (a) and after a 10%-grafting with HP-b-CD (b and c).
152
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
Fig. 6. Water contact angle at 2, 5, and 10 s. Fig. 5. Water permeability of the PET prosthesis according to the grafting rate, compared to collagen coated and uncoated PET prosthesis.
grafting reaction significantly decreased the water permeability of the PET prosthesis. In order to obtain total waterproof of vascular prostheses, one single coat of collagen is sufficient instead of two coatings applied in the usual manufacturing process. The surface analysis of the prostheses was then completed by wettability measurements using the method of the drop contact angle. Fig. 6 reports the evolution of the water drop at 5, 10 and 20 s on untreated and grafted prosthesis with b-CD, gCD, HPb-CD and HPg-CD, applying a grafting rate of 10 wt%. A rapid decrease of the drop contact angle is observed on the untreated PET, so that the drop was almost absorbed within 10 s. In contrast, CD-coated prostheses (10 wt%) presented an initial angle of 1208 that remained constant with time. This result shows that the coating reaction involved a modification of the surface properties of the prostheses that became more hydrophobic and watertight. The cyclodextrin grafting does not affect mechanical properties of vascular prostheses (Table 1). A slightly higher resistance is observed with respect to virgin prostheses for longitudinal traction, respectively, 255 and 220 N cm1 and for circumferential traction, respectively, 250 and 222 N cm1. The grafting rate and the temperature of fixation do not provoke any alteration of the mechanicals properties.
4. Discussion The first part of the study investigated the parameters of the grafting reaction. It was observed that grafting was time and temperature dependent. It has already been reported that the synthesis of CTR–CD polymers obtained from the same reactants could result in water-soluble products or insoluble gels able to swell in water (Martel et al., 2005a). The obtained form also depends on the time and the temperature of the reaction. It is dependant on the degree of polymerisation and on the cross-linking rate of the polymer formed between CD and CTR. The same explanation can be proposed in the present study, where the same reaction occurred directly onto the textile support. High temperatures or long time of curing involved high molecular weights and crosslinking rates of the polymer, and resulted in the permanent coating of the fibres. Under milder conditions, however, the CTR–CD polymer had a low degree of polymerisation and a low cross-linking degree; therefore, it remained water soluble, and was removed by the soxhlet extraction. Different optimal parameters can be settled for the CDs and for the hydroxypropyl derivates. It was observed that the hydroxypropyl CD derivatives reacted more readily than their parent molecules, as they presented lower threshold temperature and time of grafting. In fact, the hydroxypropyl groups are substituents that enhance the reactivity of the CD because they act as spacer arms, the hydroxyl group of which is more accessible to the cross linking agent than the hydroxyl groups [– CH2–CH(OH)–CH3] present directly on the glucopyranosic units of CD.
Table 1 Mechanicals assays on vascular prostheses untreated and grafted by HP-b-CD at 10 and 5 wt% with different thermal temperature Grafting type
HPb-CD 5 wt% 140 8C HPb-CD 5 wt% 150 8C HPb-CD 5 wt% 160 8C HPb-CD 10 wt% 140 8C HPb-CD 10 wt% 150 8C HPb-CD 10 wt% 160 8C Untreated Fr, radial force, Ful, forec per length unit.
Longitudinal traction
Circumferential traction 1
Fr (N)
Fr(pD) (N cm1)
Ful (N cm1)
Dilatation (%)
1183 57 1234 121 1295 85 1337 30 1315 41 1281 34 1173 150
235 11 245 24 258 17 266 6 262 8 255 7 220 28
249 47 248 47 251 48 257 49 278 20 261 29 222 17
47 2 47 3 48 2 49 1 50 3 50 5 43 2
N. Blanchemain et al. / Biomolecular Engineering 24 (2007) 149–153
This experimental approach allowed us to evaluate the correlation existing between time and temperature of curing. We could determine the optimal compromise to obtain the maximal grafting rate by application of the lowest possible reaction time and temperature. This will avoid the alteration of the mechanical properties of the grafts and is also of economical interest. Moreover, the experiment showed a decrease by four of the permeability of the prostheses after coating. This can be correlated with the SEM pictures that showed that that the grafting reaction increased the diameter of the fibres. The consequence was that the free volume between the yarns in the warp/filling arrangement of the woven structure was partially filled and thus the porosity was significantly reduced. Furthermore, the surface wettability measurements also depicted an increased hydrophobia of the modified prostheses. However, in terms of water permeability, the performances of collagen-impregnated grafts were not reached. The CD coating process is probably not completely blood proof and therefore cannot be used in substitution to collagen. Finally, mechanical tests show a slightly higher resistance of vascular grafted prostheses compared to virgin prostheses, so they are more resistant and may have longer life duration, which is another advantage for the replacement of damaged artery. 5. Conclusion The present work allowed us to confirm the possibility of grafting CDs onto woven PET vascular prostheses using citric acid as a co-reactant, and also allowed to settle the parameters of the process adapted to this specific shape of textile support. It was observed that the reaction induced a change of the physical properties of the prosthesis surface, as the fibres were found to be more hydrophobic and the water permeability significantly decreased, in spite of that, the mechanical properties were not affected. Forthcoming investigations will present the results of the biological tests and the in vitro study of the kinetics of release of several antibiotics.
153
Acknowledgements We are deeply indebted to Laboratoires Pe´rouse (60173 Ivry-Le-Temple) for financial support and the supplying of PET prostheses. We thank Annie Lefe`vre for her skilful and expert technical assistance. This project was also financially supported by the Conseil Re´gional Nord-Pas-de-Calais: ARCir ASBAMED and Fe´de´ration Biomate´riaux. References Bandyk, D.B., 1990. Semin. Vasc. Surg. 3, 122. Bandyk, D.F., Esses, G.E., 1994. Surg. Clin. North Am. 74, 571. Blanchemain, N., Haulon, S., Martel, B., Traisnel, M., Morcellet, M., Hildebrand, H.F., 2005. Eur. J. Vasc. Endovasc. Surg. 29, 628. Buschmann, H.J., Knittel, D., Schollmeyer, E., 2001. J. Incl. Phenom. Macrocycl. Chem. 40, 169. Chao, J.B., Chen, L., Xu, H., Meng, D.P., 2002. Spectrochem. Acta Part A 58, 2809. Chistensen, G.D., Baddour, L.M., Madison, B.M., 1990. J. Infect. Dis. 161, 1153. Hong, K., Nishihata, T., Rytting, J.H., 1994. Pharm. Res. 11, 88. Kieffer, E., Bahnini, A., Koskas, F., 1993. J. Vasc. Surg. 17, 349. Krupski, W.C., Nehler, M.R., Whitehill, T.A., 2002. Cardiovasc. Surg. 10, 415. Le Thuaut, P., Martel, B., Crini, G., Maschke, U., Coqueret, X., Morcellet, M., 2000. J. Appl. Polym. Sci. 77, 2118. Loftsson, T., Loftson, M.E., 1996. J. Pharm. Sci. 85, 1017. Martel, B., Le Thuaut, P., Bertini, S., Crini, G., Bacquet-Torri, M.G., Morcellet, M., 2002a. J. Appl. Polym. Sci. 85, 1771. Martel, B., Morcellet, M., Ruffin, D., Vinet, F., Weltrowski, M., 2002b. J. Inclusion Phenom. Macromol. Chem. 44, 439. Martel, B., Morcellet, M., Ruffin, D., Ducoroy, L., Weltrowski, M., 2002c. J. Inclusion Phenom. Macromol. Chem. 44, 443. Martel, B., Ruffin, D., Weltrowski, M., Lekchiri, Y., Morcellet, M., 2005a. J. Appl. Polym. Sci. 97, 433. Martel, B., Blanchemain, N., Morcellet, M., Hildebrand, H.F., Haulon, S., Boschin, F., Delcourt-Debruyne, E., 2005b. Patent application PCTFR 20050022829. O’Brien, T., Collin, J., 1992. Br. J. Surg. 79, 1262. Rajewski, R.A., Stella, V.J., 1996. J. Pharm. Sci. 85, 1142. Tetsumi, I., Kaneto, U., 1997. J. Pharm. Sci. 86, 147. Walkefield, T.W., Pierson, C.L., Schaberg, D.R., 1990. J. Vasc. Surg. 11, 624. Weltrowski, M., Morcellet, M., Martel, B. Patent PCT/FR00/00377, EP 1157156 (2000); US 09/913,448 (2001); CA 2,362,534; WO 0047811.