Polymer International 42 (1997) 373È379 Preparation, Characterization and Biodegradable Characteristics of Poly(D,L-lactide-co -1,3-trimethylene carbonate) Cai Jie & K. J. Zhu* Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, PeopleÏs Republic of China (Received 23 July 1996 ; accepted 23 October 1996) Abstract : Poly(D,L-lactide-co-1,3-trimethylene carbonate) (PLCA) has been synthesized by ring-opening polymerization of 1,3-trimethylene carbonate (CA) and D,L-lactide (LA) using stannous octoate as catalyst. The copolymers were characterized by 1H nuclear magnetic resonance (NMR), 13C NMR and di†erential scanning calorimetry. Water content and static contact angle of distilled water on the polymer surface were used to evaluate hydrophobicity of the polymers. It was found that the hydrophobicity increased with increasing CA fractions in the copolymers. Biodegradation experiments were conducted in vitro and in vivo. The results indicated that the biodegradation behaviour changed from surface to bulk degradation when the LA content exceeded 30 mol% in the copolymers. These properties of PLCA may be useful in protein delivery systems. Key words : poly(D,L-lactide-co-1,3-trimethylene carbonate), polymer biodegradation, protein-controlled release matrix chemical bonds in the polymer main-chains may be useful in protein delivery systems. When the polymer is mixed with protein and made into a device, the matrix degrades from the surface in vivo, and the protein of the surface layer is released simultaneously. Since there is low permeability of water in the highly hydrophobic polymer, the protein inside the matrix can be kept in a stable solid form. Thus the protein bioactivity may be maintained in the release process. For many polymers, degradation is a very complicated process that depends on many factors, including polymer chain scission, crystallinity, polymer chain length and water di†usivity. Polyanhydrides and polyorthoesters, as surface degradation polymers, have been studied extensively.10h15 Zhu et al.16 reported that poly(1,3-trimethylene carbonate) (PCA) showed surface biodegradable character in vivo, but degradation was very slow. In order to improve this material, a series of CA-containing copolymers have been prepared and the degradation characteristics have been explored in our laboratory recently. In this paper, we report some preliminary results on the INTRODUCTION There have been many papers concerning polymers which can be applied in drug delivery systems.1,2 Among them, biodegradable polymers have received more attention, their main advantage being that no retrieval of the materials is needed.3h7 With the development of biotechnology, many peptides and proteins have been reÐned, which are very useful in livestock farming and human therapy.8 However, these drugs are usually unstable in the body. Maintaining their bioactivities during the release process has become a key problem in protein-controlled release systems.9 Polylactide (PLA) and poly(lactide-co-glycolide) (PLGA) have been studied by many researchers. The bulk degradation characteristics of these materials, however, are unfavourable for maintaining protein bioactivity in the release process. On the other hand, surface biodegradable polymers with high hydrophobicity and unstable * To whom all correspondence should be addressed. 373 Polymer International 0959-8103/97/$09.00 ( 1997 SCI. Printed in Great Britain C. Jie, K. J. Zhu 374 preparation and biodegradation properties poly(lactide-co-1,3-trimethylene carbonate) (PLCA). of EXPERIMENTAL Differential scanning calorimetry (DSC ) measurements Samples of 10È15 mg were heated to melting in aluminium pans with inverted lids on a Perkin Elmer DSC-7 Thermal Analyzer with a heating rate of 20¡C min~1. The instrument was calibrated with cyclohexane and indium prior to use. Monomers 1,3-Trimethylene carbonate (CA) was synthesized according to the literature.17 1,3-Propanediol (60É8 g), diethyl carbonate (114 g) and a small cube of sodium metal were warmed until the sodium had completely dissolved, and then the mixture was reÑuxed at 130¡C for 4 h. Ethanol and residual diethyl carbonate were distilled out and the oil residue was dissolved in benzene, washed with water and then dried over calcium chloride. After removal of the solvent, the crude product was obtained by vacuum distillation (160È165¡C/ 6 mmHg) and then recrystallized from absolute ether (melting point 46È47¡C). D,L-Lactide (LA) was prepared by a method similar to that described in Ref. 18. Five-hundred grams of lactic acid was mixed with 2È5 wt% zinc oxide in a three-necked Ñask, and the temperature raised to 120¡C. As the rate of water elimination fell, the temperature increased to 180¡C and the pressure reduced gradually from 760 to 20 mmHg over a period of 4È6 h. When no more steam was evolved, the vacuum increased to 0É5È 1É0 mmHg. The crude product was distilled out and then recrystallized from ethyl acetate. White crystalline LA with a melting point of 125È127¡C was obtained. Copolymerization CA and LA were copolymerized in bulk. Monomers and catalyst/petroleum ether solution were introduced to a thoroughly cleaned and dried vessel. The solvent was removed in vacuum and the vessel was heated to 140¡C for 36 h. The product was dissolved in methylene chloride and precipitated from methanol. The molecular weight of the polymers was determined by gel permeation chromatography (GPC ; Waters 208 with 103, 104 and 105 ultrastyragel columns). Samples were eluted in tetrahydrofuran (THF) at a Ñow rate of 1É5 ml min~1 at 25¡C. Polymer hydrophobicity evaluation Water content and static contact angle of distilled water on the polymer surface were used to evaluate the hydrophobicity. Water content was deÐned as percentage of water in saturated polymer. It was measured gravimetrically after polymer had been immersed in distilled water to equilibrium at room temperature. Static contact angles were measured by a contact angle meter (JY-82). The polymer Ðlms were made on silanized glass microscopy slides. The coating thickness was about 2 km. Static contact angles were measured at 25¡C on proÐles of sessile drops using a microscope with a Ðxed goniometer eyepiece, magniÐcation 20]. Readings were taken within 10È15 s ; average drop size was 0É05 ml. Angles were measured on six di†erent regions of each polymer surface and the results were averaged. Polymer degradation Samples with cylindrical shape (2É5 mm in diameter, 7 mm in length) were made by compression in a mould at 40¡C for 5 min under a pressure of 100 kg cm~2. Hydrolysis of the polymers in vitro was measured by immersion of samples in phosphate bu†er solution (pH 7É4) at 37¡C. The samples were recovered periodically and the molecular weight changes were determined by GPC. In vivo degradation was tested in rats. Polymer samples were weighed and sterilized by dipping in 70% ethanol solution before implanting subdermally in adult kung ming rats in the scapular area lateral to the dorsal midline. At time intervals animals were sacriÐced and the polymers were recovered. The samples were freed from adhering tissues, then rinsed with distilled water and dried. Weight loss of the matrix was determined gravimetrically and the molecular weight changes of inner bulk and outer surface layer were measured by GPC. Nuclear magnetic resonance (NMR ) measurements A JEOL 90Q NMR instrument was employed for NMR measurements using CDCl as solvent. 1H NMR 3 measurements were conducted in 5 mm o.d. sample tubes with Me Si as internal shift reference at 4 89É55 MHz and in 10 mm o.d. sample tubes at 22É49 MHz for 13C NMR measurements. RESULTS AND DISCUSSION Polymer synthesis The polymerization of aliphatic polyesters, such as polylactide (PLA) and poly(caprolactone) (PCL), has POLYMER INTERNATIONAL VOL. 42, NO. 4, 1997 Properties of PL CA 375 TABLE 1. Some data of 1,3-trimethylene carbonate (CA) and D,Llactide (LA) copolymerizationa Catalyst conc. (M/I, mol) Temperature (¡C) Time (h) Yield (%) M1 n (Ã104) M1 /M1 w n 100 250 500 1000 500 500 500 500 140 140 140 140 140 120 160 140 36 36 36 36 20 36 36 48 84 85 80 75 65 73 86 87 1·85 2·02 1·95 1·73 1·78 1·86 1·72 1·65 2·0 2·1 2·0 1·8 1·9 1·9 2·0 1·9 a CA : LA ¼ 85·15 (by mol) in feed, Sn(oct) as catalyst. M1 and M1 /M1 n w n 2 were measured by GPC using THF as solvent at 25¡C. been studied extensively.19 However, the copolymerization of CA and LA has not been reported so far. We synthesized CA and LA copolymers at 140¡C over 36 h using stannous octoate as catalyst ; some data are shown in Table 1. It can be seen that the yield and molecular weight (MW) increased with increasing reaction time, but over 40 h MW decreased. When the reaction temperature was higher than 140¡C, MW also decreased and the product became brown. The yield and MW decreased slightly with increasing catalyst concentration in the range examined (M/I, molar ratio, from 100 to 1000). All the copolymers (PLCA15, PLCA30, PLCA50 and PLCA60, the numbers representing the approximate molar percentage of LA in the copolymer) were soluble in CH Cl , CHCl , THF 2 2 3 and dioxane, but insoluble in alcohol and ether. The GPC proÐles of the copolymers showed symmetrical single peaks. In comparison with 1H and 13C NMR spectra of PCA and PLA,16,17 the peaks at 2É10 and 4É25 ppm in the 1H NMR spectrum of PLCA can be assigned to wCH w and wOwCH w of CA units, and the peaks 2 2 at 1É65 and 5É20 ppm to wCH and xCHw of LA 3 units. In the 13C NMR spectrum of PLCA, the peaks at 16É7, 70É0 and 169É4 ppm correspond to wCH , 3 xCHw and CO of LA units, respectively, and 28É0, 64É5 and 155É0 ppm to wCH w, wOCH w and CO of 2 2 CA units, respectively. Fig. 1. The relationship between hydrophobicity and composition of PLCA copolymers. (A) Water content, which was measured gravimetrically after samples were immersed in distilled water to equilibrium at room temperature. (B) Static contact angle of distilled water on the polymer surface, which was measured by a contact angle meter at 25¡C ; the average drop size was 0É05 ml and the readings were taken within 10È15 s. POLYMER INTERNATIONAL VOL. 42, NO. 4, 1997 C. Jie, K. J. Zhu 376 Fig. 2. Di†erence in molecular weight between inner bulk and surface layers of PLCA copolymers for in vitro test (M1 and M1 n no were determined by GPC in THF at 25¡C ; M1 \ initial molecular weight of samples). (a) PLCA15 and PLCA30 copolymers ; (b) no PLCA50 and PLCA60 copolymers. TABLE 2. DSC data of poly(D,L-lactide-co -1,3trimethylene) (PLCA)a Sample Composition (LA mol%) T (¡C) g PCA PLCA15 PLCA30 PLCA50 PLCA60 PLA 0 15·3 30·2 49·7 60·4 100 É 26 É 18 É8 9 18 56 a T s were obtained by DSC measurements with a heating g rate of 20¡C minÉ1 using a Perkin Elmer DSC-7 instrument. The compositions were determined by 1H NMR. The reactivity ratios of LA and CA (r and r ) were L C calculated according to the following equation :18 M/N \ (r m/n ] n)/(r n/m ] n) (1) L c where m and n are feed compositions of LA and CA, and M and N are compositions of the copolymers determined by 1H NMR. The values of r and r L C obtained were 1É03 and 0É76, respectively. Polymer hydrophobicity The water content and static contact angles of PLCA copolymers are shown in Fig. 1. Both have good linear relationships with copolymer composition, that is, the POLYMER INTERNATIONAL VOL. 42, NO. 4, 1997 Properties of PL CA 377 Fig. 3. Di†erence in molecular weight between inner bulk and surface layers of PLCA copolymers for in vivo test (M1 and M1 n no were determined by GPC in THF at 25¡C ; M1 \ initial molecular weight of samples). (a) PLCA15 and PLCA30 copolymers ; (b) no PLCA50 and PLCA60 copolymers. hydrophobicity decreases with increasing LA fraction in the copolymers. DSC studies indicate that all of the copolymers are amorphous and have only one glass transition temperature, T (miscible material) (Table 2). g Thus, the change of water content and static contact angle of the copolymers can be attributed to the di†erent hydrophobicity of the LA and CA segments, since the water absorption of PLA is 5É2%, which is higher than that of PCA (1É8%). Polymer degradation The degradation of PLCA copolymers in vitro and in vivo was evaluated by changes in molecular weight of POLYMER INTERNATIONAL VOL. 42, NO. 4, 1997 the inner bulk and outer surface layers, and in the weight loss of the polymers. The low dimensional stability and deformation of the samples recovered (because of their low T ) prevented precise sampling of surface g and bulk. Therefore, the recovered samples were extracted with CHCl for 5 min to selectively separate 3 the surface fraction (about 3 wt% of sample). Control experiments showed that CHCl , a poor solvent for 3 PLCA copolymers, slowly dissolved the surface layer of the polymer and did not selectively leach oligomeric components. It was found that the di†erence in molecular weight between inner bulk and surface layers increased with decreasing LA content in the copolymers both in vitro and in vivo (Figs 2 and 3). There are large C. Jie, K. J. Zhu 378 di†erences in molecular weight for smaller LA content of PLCA, such as PLCA15 and PLCA30, but almost identical results for higher LA content in the copolymers, such as PLCA50 and PLCA60. Thus, the former show surface degradation character, and the latter bulk degradation character over the test period. The critical transition composition for surface to bulk degradation is around 40 mol% LA in the copolymers. The degradation of the polymer matrix can be divided into surface and bulk degradation depending on factors such as matrix hydrophobicity, water penetration rate and polymer chain scission rate.13,19 In the case of PLCA degradation, the copolymers containing more LA have less hydrophobicity and water di†usion is relatively easy. Considering the degradation mechanism of PLA and PCA, as shown in kinetic eqns (2) and (3),16,20 an autocatalysation by degraded acid products existed in PLA degradation, but this e†ect was not important in PCA because of the much lower amount of acid from degraded products (CO ). These factors are 2 favourable for bulk degradation of copolymers containing more LA. PLCA60 (16% for PLCA15, 0É4% for PLCA60). The change of composition of PLCA15 in the degradation process indicates that the CA content decreased from 85 mol% at the initial stage to 80 mol% at 40 days for the in vivo test, but there was no apparent change for the in vitro test. This enzyme attack on the surface of CA segments additionally favours the surface degradation of PLCA copolymers with greater CA content. Figure 4 shows weight loss of PLCA copolymers in the in vivo test. It can be seen that PLCA15 and PLCA30 lost weight gradually and almost linearly. This is another characteristic of surface degradation of polymers. For PLCA50 and PLCA60, the matrix weight changed slightly initially, but after 40 days, weight loss increased rapidly. This phenomenon is consistent with the fact that bulk degradation chain cleavage of the polymer backbone proceeds until a critical MW is reached, after which weight loss occurs by the di†usion of water-soluble cleavage fragments out of the matrix.21 [d[ester]/dt \ k[H O][COOH][ester] (2) 2 [d[carbonate]/dt \ k[H O][carbonate] (3) 2 Each PLCA copolymer has a similar degradation pattern in vitro and in vivo, but the degradation rate in vivo is higher than that in vitro. This may be attributed to enzyme attack on CA segments of the matrix.16 The evidence in our experiments is that by comparing M1 /M1 data of the outer surface layers of the copolyn no mers for in vitro and in vivo degradation at 90 days, the discrepancy for PLCA15 is much larger than that for The copolymer, PLCA, can be synthesized by ringopening polymerization of LA and CA with stannous octoate. Hydrophobicity of the copolymers decreased with increasing LA fraction in the copolymers. It was found that the copolymers containing 30 mol% or less LA showed surface degradation characteristics in both in vitro and in vivo tests. For LA contents exceeding 30 mol% the copolymers had bulk degradation behaviour. Enzyme e†ects may play an important role in the degradation process in vivo for the copolymers containing more CA. CONCLUSIONS Fig. 4. Weight loss of PLCA15, PLCA30, PLCA50 and PLCA60 copolymers for in vivo test, which was determined gravimetrically with recovered samples. POLYMER INTERNATIONAL VOL. 42, NO. 4, 1997 Properties of PL CA ACKNOWLEDGEMENTS We are indebted to the National Natural Science Foundation of China for their support of this work. We also thank Ms Qiu Liyan for her assistance in the animal tests. REFERENCES 1 Langer, R., Science, 249 (1990) 1527. 2 Lenz, R. W., in Advances in Polymer Science 107, Biopolymers 1, eds N. A. Peppas & R. Langer. Spring-Verlay, Berlin, 1993. 3 Brem, H., Mahaley, M. S., Vick, N. A., Black, K. L., Schold, S. C., Burger, P. C., Friedman, A. H., Ciric, T. S., Eller, T. W., Cozzens, J. W. & Kenally, J. N., J. Neurosurg., 74 (1991) 441. 4 Chasin, M. & Langer, R., Biodegradable Polymers as Drug Delivery Systems. Drug and the Pharmaceutical Sciences, Vol. 25, Marcel Dekker, New York, 1990. 5 Asano, M., Yoshida, M., Kaetsu, I., Imai, K., Mashimo, T., Yuasa, H., Yamanaka, H., Suzuki, K. & Yamazaki, I., Makromol. Chem., Rapid Commun., 6 (1985) 509. POLYMER INTERNATIONAL VOL. 42, NO. 4, 1997 379 6 Giacomo, D. C., Biomaterials, 13 (1992) 850. 7 Langer, R., Lund, D., Leong, K. & Folkman, J., J. Contrl. Rel., 2 (1985) 331. 8 Hutchinson, F. G. & Furt, B. J. A., J. Contrl. Rel., 13 (1990) 279. 9 Pitt, C. G., Int. J. Pharm., 59 (1990) 173. 10 Tamada, J. & Langer, R., J. Biomater. Sci. Polymer Edn, 3 (1992) 315. 11 Yasuhiko, T. & Langer, R., Pharm. Res., 10 (1993) 391. 12 Tamada, J. & Langer, R., J. Biomater. Sci. Polymer Edn, 3 :4 (1992) 315. 13. Heller, J., Biomaterials, 11 (1990) 659. 14 Heller, J., J. Contrl. Rel., 8 (1988) 111. 15 Heller, J., Ng, S. Y., Fritzinger, B. K. & Roakos, K. V., Biomaterials, 11 (1990) 235. 16 Zhu, K. J., Hendren, R. W., Jensen, K., & Pitt, C. G., Macromolecules, 24 (1991) 1736. 17 Searles, S., Hummel, D. G., Nukina, S. & Throckmorton, P. E., J. Am. Chem. Soc., 82 (1960) 2928. 18 Gilding, D. K. & Reed, A. M., Polymer, 20 (1979) 1459. 19 Simon, J. H. & Brian, J. T., J. Contrl. Rel., 4 (1986) 155. 20 Pitt, C. G. & Schindler, A., in Progress in Contraceptive Delivery Systems, ed. E. S. E. Hafez. M. T. P. Publishers, Lancaster, 1980, pp. 17È46. 21 Kenley, R. A., Lee, M. O., Mahoney, T. R. & Sanders, L. M., Macromolecules, 20 (1987) 2398.