Mucoadhesive properties of cross-linked high amylose starch derivatives

International Journal of Biological Macromolecules 40 (2006) 9–14

Mucoadhesive properties of cross-linked high amylose starch derivatives J´erˆome Mulhbacher a , Pompilia Ispas-Szabo a , Mathieu Ouellet b , Serge Alex c , Mircea Alexandru Mateescu a,∗ a

Department of Chemistry and Biochemistry, Centre BioMed, Universit´e du Qu´ebec a` Montr´eal, CP 8888, Succ. Centre-Ville, Montr´eal, Qu´ebec H3C 3P8, Canada b BioSyntech Canada Inc. 475, boul. Armand-Frappier, Laval, Qu´ ebec H7V 4B3, Canada c Institut de Chimie et de Petrochimie, 6220 rue Sherbrooke Est, Montr´ eal, Qu´ebec H1N 1C1, Canada Received 16 October 2005; received in revised form 6 April 2006; accepted 9 May 2006 Available online 19 May 2006

Abstract Acetate (Ac-), aminoethyl (AE-) and carboxymethyl (CM-)derivatives of cross-linked high amylose starch (HASCL-6) were previously shown to control, over more than 20 h, the release of drugs from highly loaded (up to 60% drug) monolithic tablets. It was now of interest to evaluate their mucoadhesive characteristics in view of further utilization in buccal or vaginal transmucosal delivery. The present study shows that ionic AE-HASCL-6 and CM-HASCL-6 derivatives exhibit higher mucoadhesive properties than neutral HASCL-6 and Ac-HASCL-6, suggesting that the ionic groups introduced on cross-linked starch chains play a role in the bioadhesion process. The adhesiveness seemed related to capillary attraction forces. Surface adhesion parameters were calculated for slabs based on the mentioned polymers and corroborated with their swelling behavior at various pH changes. The positively charged AE-derivatives presented a higher adhesion at acidic pH, being thus recommended for vaginal delivery, whereas the negatively charged derivatives (CM-HASCL-6) exhibited a better adhesion at neutral pH, being thus more appropriate for buccal delivery. © 2006 Elsevier B.V. All rights reserved. Keywords: High amylose starch derivatives; Hydrogels; Mucosal adhesion; Force of adhesion; Work of adhesion; Spreading coefficient

1. Introduction Cross-linked high amylose starch (HASCL) was introduced in the nineties as a pharmaceutical excipient (ContramidR ) for oral dosage forms, ensuring drug controlled release over 18–24 h [1,2]. Partial substitution of the hydroxyl groups of HASCL-6 with ionic CM- (carboxymethyl) and AE- (aminoethyl) groups as well as with less polar Ac- (acetate) groups was shown [3] to markedly increase the loading capacity of monolithic matrices (from 20% for HASCL to 60% drug for HASCL derivatives) as excipients. It was also shown that the carboxylic (CMHASCL-6) and amino (AE-HASCL-6) derivatives are able to modulate the release of certain active agents through ionic interactions whereas in the case of acetate groups the release seems modulated by a general enhancing of the hydrophobic character of the Ac-HASCL-6 matrix [3]. Previous studies have

∗

Corresponding author. Tel.: +1 514 987 4319; fax: +1 514 987 4054. E-mail address: [email protected] (M.A. Mateescu).

0141-8130/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ijbiomac.2006.05.003

also showed the key role of hydroxyl groups and of hydrogen bonding [2,4] in the organization of the high amylose starch matrices [5], contributing thus to the control of the drug release. The mechanisms controlling the drug release from matrices based on HASCL-6 or its derivatives, consisting mainly in the polymer swelling and drug diffusion through the swollen matrix were previously studied [6,7]. In the case of HASCL6, Ac-HASCL-6 and AE-HASCL-6 tablets, the swelling was not affected by the ionic strength and the pH of the dissolution media, whereas for the CM-HASCL-6 tablets, swelling depended on both ionic strength and pH of the dissolution medium [6]. The drug diffusion was found to depend on the molecular weight of the drug, whereas the partition coefficient depended on the affinities of the drug for the polymers and for the dissolution medium [7]. In the case of CM-HASCL-6 the drug diffusion was also affected by the tablet swelling, which increased when the ionic strength decreases [6,7]. It was observed that certain tablets based on cross-linked high amylose starch (HASCL) derivatives exhibit sticking properties

10

J. Mulhbacher et al. / International Journal of Biological Macromolecules 40 (2006) 9–14

on wet surfaces (i.e. glass, human skin). This behavior makes these materials attractive for bioadhesive applications, particularly for transmucosal drug delivery. The concept of drug controlled release from high loaded oral dosages based on HASCL derivatives [3] could be of interest for mucoadhesive forms, which require a special attention given to the shape of the dosage form (i.e. for a better acceptability by the subjects, a bioadhesion tablet must be smaller in size than the normal oral dosage forms). Three steps are involved in the process of bioadhesion: (i) the wetting and the swelling of the polymer, (ii) the interpenetration of bioadhesive polymer chains and mucin, and (iii) formation of weak physical bonds between polymer and mucin. Several types of association could be involved such as ionic, Van der Waals and/or hydrogen bonds. These physical interactions, weak when alone, will produce, in their assembly, a strong adhesion through a high number of interaction sites. For this reason, polymers of high molecular weight and exhibiting lots of polar groups, will be potential candidates as mucoadhesive excipients [8]. Several procedures have been developed over the last decades to measure and characterize bioadhesion. Some of them allow the quantification of the adhesion forces [9–12] whereas the others are focused on mechanism of adhesion. For these evaluations, measurements of contact angle used to determine the spreading coefficient [13–15] and the work of adhesion [14] represent useful tools. Several polymers are in use for their adhesive properties in transmucosal formulations, the most common being CarbopolTM (BFGoodrich) of various types [9,10,12,16] and HydroxyPropyl Methyl Cellulose (HPMC) [12,16]. Other polymers such as alginate [16] and N,N-dimethylaminoethyl methacrylate-co-methyl methacrylate [17] are also used. The Mach-1 Micromechanical System (BioSyntech Inc., Canada) was used to evaluate adhesion characteristics. Other research groups previously used the same equipment for mechanical characterization of polymers and biological tissues [18–20]. The aim of this study was to evaluate the impact of the HASCL polymer derivatization, by introducing polar (carboxymethyl, aminoethyl) or non-polar (acetate) groups, on the adhesive properties of tablets, in view to use the starch excipients for transmucosal delivery and to understand the adhesion mechanism of these new materials.

2.1.1. Synthesis of CM-HASCL-6 An amount of 70 g of HAS was first suspended in 170 mL of water, completed with 235 mL of 1.5 M NaOH and then cross-linked by 3.5 mL of epichlorohydrin for 40 min at 50 ◦ C [21]. The reaction medium was then treated with 5 g of monochloroacetic acid at the same temperature, for 1 h [22]. Once the reaction was completed, the CM-HASCL-6 suspension (slurry) was neutralized with acetic acid at room temperature and washed with an equal volume of acetone/water 85/15 (v/v), kept 20 min and filtered. The washing procedure [3] was repeated by resuspending the gel and by two more filtrations with first 1/2 equivalent volume of acetone/water 70/30 (v/v) and then with 85/15 (v/v). The remaining wet gel was dried with pure acetone (a volume equivalent to the half of the final reaction volume, this operation being repeated three times). 2.1.2. Synthesis of AE-HASCL-6 A similar procedure as for CM-HASCL-6 was followed. The same amount of HAS was first cross-linked with epichlorohydrin and then treated with 86 g of chloroethylamine hydrochloride (rapidly solubilized in a minimal volume of water just before the use) for 2 h at 70 ◦ C [23]. The pH was maintained between 9 and 10 during the synthesis (by adding small volumes of 5 M NaOH). The neutralization, washing and drying were done in same conditions as described for CM-HASCL-6. 2.1.3. Synthesis of Ac-HASCL-6 Similar conditions as for CM-HASCL-6 were followed for synthesis of acetate derivatives. A batch of 70 g of HAS was first treated with epichlorohydrin [21] and then with 15 mL of acetic anhydride [24] at room temperature for 1 h, followed by neutralization, washing and drying as for CM-HASCL-6. 2.2. Evaluation of substitution degree of derivatives The carboxylic groups of CM-HASCL-6, were potentiometrically titrated (Corning pH-meter) with 0.1 M NaOH. The amine groups of AE-HASCL-6 were determined [25] with trinitrobenzene sulfonic acid (TNBS). The acetate groups of Ac-HASCL-6 were assayed by 1 H NMR, as previously described [3]. The obtained substitution degrees were in the same range: 0.092 mmol/g for CM-HASCL-6, 0.049 mmol/g for AE-HASCL-6 and 0.029 mmol/g for Ac-HASCL-6. 2.3. Adhesion test

2. Materials and methods High amylose starch (Hylon VII) from National Starch (USA), mucin (Sigma–Aldrich, USA), derivatization agents (monochloroacetic acid, chloroethylamine hydrochloride, and acetic anhydride) and the other chemicals (reagent grade), were used without further purification. 2.1. Synthesis of HAS derivatives The syntheses were realized in conditions previously described by Mulhbacher et al. [3].

Tablets (100 mg) were obtained for each polymer (HASCL-6 and its derivatives) by dry compression at 29.4 kN (4 T/cm2 ) in a Carver hydraulic press using a 9 mm diameter punch. The tablet adhesion properties were tested in triplicate using the Mach-1 Micromechanical System (BioSyntech Inc., Canada). The device (Fig. 1) is a mechanical testing system composed of a load cell (150 mg resolution) and an actuator for the precise control of the displacement (25 nm resolution). Mucin membranes were prepared by dipping a cellulose acetate membrane in 5% mucin solution for at least 30 min, as described by Jain et al. [12]. Mucin membranes were fixed on the membrane holder of the actuator

J. Mulhbacher et al. / International Journal of Biological Macromolecules 40 (2006) 9–14

11

of 5% mucin in simulated gastric fluid (SGF) pH 1.2 or (b) 200 ␮L of 5% mucin in simulated intestinal fluid (SIF) pH 7.5. Both SGF and SIF preparations did not contain enzymes (no pepsin, no pancreatin). The dry tablet of tested polymers were fixed on the tablet holder. 2.3.2. The swelling effect on tablet adhesion Prior to adhesion assays, tablets based on each polymer were swollen in SIF for 24 h and just before the test they were cut at 9 mm diameter, in order to fit in the tablet holder. The tests were conducted as previously described with cellulose acetate membrane coated with mucin and 200 ␮L of SIF. Fig. 1. Schematic representation of the Mach-1 Micromechanical System (Biosynthec).

and spread with 200 ␮L of mucin solution. The tablets were glued with cyanoacrylate adhesive on the metallic tablet holder of the load cell (Fig. 1). The Mach-1 actuator was programmed to move in the direction of the load cell at the rate of 10 ␮m/s, until a weight of 1 g was detected by the load cell. Then, the actuator moves back from the load cell at a rate of 100 ␮m/s. A typical force–time dependency run is presented in Fig. 2. During its course (100 s), the positions of the actuator and the measured weights were all monitored at every 1/10 s. Similar experiments were realized in which the weight (1 g) was maintained for 1 min, obtaining same results. The detachment forces were determined using the data collected during the experiment. 2.3.1. The pH effect on tablet adhesion In order to evaluate the influence of pH on the dry tablets adhesion, tests were conducted using cellulose acetate membranes coated with mucin and two types of solutions: (a) 200 ␮L

2.4. Contact angle measurement Tablets of 100 mg were obtained as for the adhesion test (dry compression, 29.4 kN, 9 mm diameter) and at least three tablets of HASCL-6 and its derivatives were used for each assay. One tablet was stuck to a glass slide, which was then placed on a support and dipped into the appropriate medium (SGF or SIF). Mucin membranes were prepared as for the adhesion tests and stuck on the glass slide. The immerged sample was positioned in the mire of a VCA1000 camera (AST Products Inc., Billerica, MA, USA) by adjusting the goniometer. Octane and air were used to determine the contact angle. A drop of testing liquid was deposed on the sample surface and an image was recorded. 2.4.1. Data analysis The measured contact angles were used to calculate the surface free energy, the spreading coefficient and the works of adhesion with the following equations. The data was compiled using Young’s analysis [8]: γsv = γsl + γlv cos(θ)

Fig. 2. Typical monitoring data during the adhesion test. Weights measured by the load cell (a) and positions of the actuator (b) in function of time. This example corresponds to the adhesion test of Ac-HASCL-6 in SIF medium and is representative for all the polymeric matrices used in this study.

12

J. Mulhbacher et al. / International Journal of Biological Macromolecules 40 (2006) 9–14

where θ is the contact angle, γ sv the interfacial tension between s (solid) and v (vapor), γ sl the interfacial tension between s and l (liquid) and γ lv is the interfacial tension between l and v. The geometric mean equation was used for the calculation of interfacial free energy [10]: 0.5

γij = γi + γj − 2(γiLW γjLW )

0.5

− 2(γiAB γjAB )

where γ ij is the interfacial tension between i and j components, γ i the total surface tension of i, γ j the total surface tension of j, γiLW the Lifshitz–Van der Waals component of the surface tension of i, γjLW the Lifshitz–Van der Waals component of the surface tension of j, γiAB the acid–base component of the surface tension of i and γjAB is the acid–base component of the surface tension of j (i and j are two different substances). The spreading coefficient was calculated as [8]: Scpm = γml − γpm − γpl where Scpm is spreading coefficient of polymer on mucin, γ ml the interfacial tension between m (mucin) and l (liquid), γ pm the interfacial tension between p (polymer) and m, and γ pl is the interfacial tension between p and l. The work adhesion was calculated by Dupre’s equation [26]: Wa = γp + γm − γpm where Wa is the work of adhesion, γ p the total surface tension of p (polymer), γ m the total surface tension of m (mucin) and γ pm is the interfacial tension between p and m. 3. Results and discussion A moderately low cross-linking degree of high amylose starch and derivatives allows enough chains flexibility to establish strong hydrogen bonding between hydroxylic groups [5]. In addition, in certain pH conditions, CM- and AE-groups can also contribute to hydrogen bonding. Table 1 shows the force of adhesion of HASCL-6, Ac-HASCL-6, AE-HASCL-6 and CMHASCL-6 measured at pH 1.2 (SGF) and 7.5 (SIF). The forces of adhesion of HASCL-6, Ac-HASCL-6 and CM-HASCL-6 increased with the pH increase from 1.2 to 7.5, whereas the force of adhesion of AE-HASCL-6 strongly decreased under identical conditions. Similar results were obtained by Quintanar-Guerrero et al. [17] for cationic polymer (N,N-dimethylaminoethyl methacrylate-co-methyl methacrylate). Among the polymers tested in the present study, the CM-HASCL-6 has the highest Table 1 The forces of adhesion to the mucin of dry tablets based on HASCL-6 or its derivatives, at different pH values Tablet composition

Forces of adhesion in SGF (pH 1.2) (mN/cm2 )

HASCL-6 Ac-HASCL-6 AE-HASCL-6 CM-HASCL-6

13.5 9.10 47.03 25.75

± ± ± ±

4.01 2.93 0.31 4.63

Forces of adhesion in SIF (pH 7.5) (mN/cm2 ) 24.52 18.04 11.26 36.70

± ± ± ±

0.77 3.61 4.54 1.08

n ≥ 3. SGF, simulated gastric fluid; SIF, simulated intestinal fluid.

Table 2 The forces of adhesion to the mucin of tablets based on HASCL-6 or its derivatives previously swollen for 24 h in SIF Tablet composition

Forces of adhesion in SIF after 24 h swelling (mN/cm2 )

HASCL-6 Ac-HASCL-6 AE-HASCL-6 CM-HASCL-6

10.18 0.31 17.73 6.32

± ± ± ±

4.38 0.15 2.07 3.01

n ≥ 3. SIF, simulated intestinal fluid.

force of adhesion in SIF. The variation of adhesion forces with pH changes seems to follow the swelling properties of the matrix, and is affected by hydration and by stabilization via hydrogen bonding. The carboxyl groups of CM-HASCL-6 are protonated at pH 1.2 and thus involved in hydrogen bonding whereas at pH 7.5 they will be ionized. On the other hand, the amino groups of AE-HASCL-6 will be fully ionized (protonated) at pH 1.2, whereas at pH 7.5 as a free amine, it will be available for interchain hydrogen bonding. The HASCL-6 and its Ac derivative are less sensitive to pH, with the mention that when the hydroxyl groups will be closer to their pKa , they will be less available for hydrogen bonding. The capillary attraction forces could be considerably high when water from the space between a polymer and mucosa is uptaken by the dry polymer [10]. It was therefore supposed that the adhesion properties of the studied polymers are closely related to their swelling process, which would cause a decrease of the pressure in the space created between the tablet and the mucin. This hypothesis was investigated by testing the adhesion behavior of tablets previously swollen at neutral pH (Table 2). A drastic decrease of the adhesion forces of HASCL-6, AcHASCL-6 and CM-HASCL-6 tablets swollen at pH 7.5 was found in comparison to dry tablets with the same composition. This confirms the role of the swelling in the adhesion process of these derivatives. The AE-HASCL-6 exhibits the lowest tablet adhesion force in dry phase, but, differently to the other derivatives, a moderately higher force of adhesion was obtained when it was previously swollen. In this case, enhanced interactions between opposite charge of cationic AE-HASCL6 and anionic mucin [27] can increase the interpenetration phenomena. The surface characteristics of HASCL-6 and its derivatives were evaluated by contact angle measurements (Table 3), spreading coefficients (Table 4) and the works of adhesion (Table 5) calculations. It is worth to note that measurements of contact angle were made on partially swollen tablets (the swelling started as soon as the tablet was immerged in the testing solution, a gel layer being observed within the first minute). The spreading coefficients at pH 1.2 were negative (−1.07 to −0.09), indicating that there are no spreading between the tablets and the mucin substrate. The lack of spreading in gastric medium can be an advantage for tablets based on HASCL and its derivatives, particularly for formulation aimed to intestinal or colonic absorption. At pH 7.5 the spreading coefficients were higher (4.21–5.09) for HASCL-6 and its Ac-HASCL-6 and AE-HASCL-6 derivatives (indicating stronger interactions between the tablets and

J. Mulhbacher et al. / International Journal of Biological Macromolecules 40 (2006) 9–14

13

Table 3 Surface parameters of mucin and of HASCL-6 or its derivatives, determined by contact angle measurement pH

Compound

γ Total (mJ/m2 )

1.2 SGF

Mucin HASCL-6 Ac-HASCL-6 AE-HASCL-6 CM-HASCL-6

40.96 41.25 36.59 40.49 39.69

± ± ± ± ±

2.22 0.83 3.77 1.50 1.94

12.66 14.18 10.37 11.92 11.24

± ± ± ± ±

0.95 0.37 1.54 0.62 0.79

28.30 27.08 26.22 28.57 28.44

± ± ± ± ±

1.27 0.46 2.23 0.87 1.16

7.5 SIF

Mucin HASCL-6 Ac-HASCL-6 AE-HASCL-6 CM-HASCL-6

38.30 51.49 49.93 51.34 56.22

± ± ± ± ±

2.24 1.44 2.65 2.00 0.56

6.32 20.97 21.63 18.37 27.21

± ± ± ± ±

0.79 0.73 1.38 0.96 0.31

31.97 30.52 28.30 32.97 29.01

± ± ± ± ±

1.44 0.71 1.27 1.03 0.25

γ LW (mJ/m2 )

γ AB (mJ/m2 )

n ≥ 3. SGF, simulated gastric fluid; SIF, simulated intestinal fluid.

Table 4 Spreading coefficient of tablets based on HASCL-6 or its derivatives, on mucin

4. Conclusion

Tablet composition

Spreading coefficient in SGF (pH 1.2)

Spreading coefficient in SIF (pH 7.5)

HASCL-6 Ac-HASCL-6 AE-HASCL-6 CM-HASCL-6

−0.09 −1.07 −0.09 −0.29

4.62 4.21 5.09 2.46

The present study shows that ionic high amylose starch derivatives AE-HASCL-6 and CM-HASCL-6 exhibited higher adhesion forces to the mucin than HASCL-6 and Ac-HASCL6. This could be of interest for bioadhesive formulations for drug delivery in acidic and neutral pH, respectively. Their adhesive properties seem to be due to capillary attraction forces and, by corroboration of all characteristics, CM-HASCL-6 (with its high swelling and high loading capacity) appears as an interesting excipient for buccal delivery applications. These polymers could be of great interest for transmucosal delivery, as they possess adhesive properties needed for tablet adhesion and can also control the drug release.

SGF, simulated gastric fluid; SIF, simulated intestinal fluid.

mucin) whereas for CM-HASCL-6 only a moderate increase (2.46) was found suggesting weaker interactions between these two anionic materials. The works of adhesion values were in the same range for all the studied polymers and they slightly increased with the pH (77.4–82.2 at pH 1.2 and 83.5–87.1 at pH 7.5) showing that the surface energy favors the adhesion of the polymers to the mucin at both pH, with slightly higher capacity at pH 7.5 than at pH 1.2. The spreading coefficients and the works of adhesion obtained for HASCL-6 and its Acand CM-derivatives are in agreement with the increase of the adhesion forces when the pH change from 1.2 to 7.5. The force of adhesion increase seems related to the increase of the spreading coefficients and not to the work of adhesion which remains stable. For AE-HASCL-6, there is a decrease of the force of adhesion at increasing pH (Table 1) despite the fact that the spreading coefficient and the work of adhesion change in the same way as for the other derivatives. This decrease could be due to the loss of the ionic interactions between opposite charges of cationic AE-HASCL-6 and anionic mucin, as previously suggested. Table 5 Works of adhesion of tablets based on HASCL-6 or its derivatives, on mucin Tablet composition

Wa (mJ/m2 ) in SGF (pH 1.2)

Wa (mJ/m2 ) in SIF (pH 7.5)

HASCL-6 Ac-HASCL-6 AE-HASCL-6 CM-HASCL-6

82.2 77.4 81.4 80.6

85.5 83.5 86.5 87.1

SGF, simulated gastric fluid; SIF, simulated intestinal fluid.

Acknowledgements J. Mulhbacher was holder of a University-Industry graduate PhD studentship from Fonds NATEQ (formerly FCAR, Qu´ebec, Canada) with participation of Labopharm Inc. Both contributions are gratefully acknowledged. References [1] V. Lenaerts, Y. Dumoulin, M.A. Mateescu, J. Control. Release 15 (1991) 39–46. [2] M.A. Mateescu, V. Lenaerts, Y. Dumoulin, Canadian Patent 2,041,774, 1994; US Patent 5,456,921, 1995. [3] J. Mulhbacher, P. Ispas-Szabo, V. Lenaerts, M.A. Mateescu, J. Control. Release 76 (2001) 51–58. [4] Y. Dumoulin, S. Alex, P. Szabo, L. Cartilier, M.A. Mateescu, Carbohydr. Polym. 37 (1998) 361–370. [5] P. Ispas-Szabo, F. Ravenelle, I. Hassan, M. Preda, M.A. Mateescu, Carbohydr. Res. 323 (2000) 163–175. [6] J. Mulhbacher, P. Ispas-Szabo, M.A. Mateescu, Int. J. Pharm. 278 (2004) 231–238. [7] J. Mulhbacher, M.A. Mateescu, Int. J. Pharm. 297 (2005) 22–29. [8] D.E. Chickering III, E. Mathiowitz, Bioadhesive Drug Delivery Systems: Fundamentals, Novel Approaches and Development, vol. 98, Marcel Dekker, New York, 1999, pp. 197–221. [9] C.-M. Lehr, J.A. Bouwstra, J.J. Tukker, H.E. Junginger, J. Control. Release 13 (1990) 51–62. [10] C.-M. Lehr, J.A. Bouwstra, H.E. Bodd´e, H.E. Junginger, Pharm. Res. 9 (1992) 70–75.

14

J. Mulhbacher et al. / International Journal of Biological Macromolecules 40 (2006) 9–14

[11] M. Rillosi, G. Buckton, Int. J. Pharm. 117 (1995) 75–84. [12] A.C. Jain, B.J. Aungst, M.C. Adeyeye, J. Pharm. Sci. 91 (2002) 1659–1668. [13] C.-M. Lehr, H.E. Bodd´e, J.A. Bouwstra, H.E. Junginger, Eur. J. Pharm. Sci. 1 (1993) 19–30. [14] E. Oh, P.E. Luner, Int. J. Pharm. 188 (1999) 203–219. [15] H. Khan, J.T. Fell, G.S. Macleod, Int. J. Pharm. 227 (2001) 113–119. [16] G. Ikinci, S. Senel, C.G. Wilson, M. Sumnu, Int. J. Pharm. 277 (2004) 173–178. [17] D. Quintanar-Guerrero, R. Villalobos-Garcia, E. Alvarez-Colin, J.M. Cornejo-Bravo, Biomaterials 22 (2001) 957–961. [18] J. Dumont, M. Ionescu, A. Reiner, A. Poole, N. Tran-Khanh, C.D. Hoemman, M. McKee, M.D. Buschmann, Connect. Tissue Res. 40 (1999) 259–272. [19] M. Garon, A. Legare, R. Guardo, P. Savard, M.D. Buschmann, J. Biomech. 35 (2002) 207–216.

[20] F. Ravenelle, R.H. Marchessault, A. Legare, M.D. Buschmann, Carbohydr. Polym. 47 (2002) 259–266. [21] M.A. Mateescu, Y. Dumoulin, L. Cartilier, V. Lenaerts, US Patent 5,603,956, 1997. [22] H.D. Schell, M. Serban, M.A. Mateescu, T. Bentia, Rev. Roum. Chim. 23 (1978) 1143–1147. [23] M.A. Mateescu, G. Fortier, S. Neidhart, S. Roger, Chromatographia 26 (1988) 110–114. [24] W. Jarowenko, Modified Starches: Properties and Use, CRC Press, Boca Raton, 1986, pp. 55–75. [25] A.F.S.A. Habeeb, Anal. Biochem. 14 (1966) 337–346. [26] P.C. Hiemenz, Undergraduate Chemistry, vol. 4, Marcel Dekker, New York, 1986, p. 815. [27] N.A. Peppas, J.J. Sahlin, Biomaterials 17 (1996) 1553–1561.