Antibiotic release from biodegradable PHBV microparticles

Journal of Controlled Release 59 (1999) 207–217

Antibiotic release from biodegradable PHBV microparticles a ¨ Dilek Sendil a , Ihsan Gursel , Donald L. Wise b , Vasıf Hasırcı a ,b , * b

a Middle East Technical University, Department of Biological Sciences, Biotechnology Research Unit, Ankara 06531, Turkey Department of Chemical Engineering and Center for Biotechnology Engineering, Northeastern University, 342 SN, 360 Huntington Avenue, Boston, MA 02115, USA

Received 30 June 1998; received in revised form 6 November 1998; accepted 13 November 1998

Abstract For the treatment of periodontal diseases, design of a controlled release system seemed very appropriate for an effective, long term result. In this study a novel, biodegradable microbial polyester, poly(3-hydroxybutyrate-co-3-hydroxyvalerate), PHBV of various valerate contents containing a well established antibiotic, tetracycline, known to be effective against many of the periodontal disease related microorganisms, was used in the construction of a controlled release system. Tetracycline was loaded in the PHBV microspheres and microcapsules both in its acidic (TC) and in neutral form (TCN). Microcapsules of PHBV were prepared under different conditions using w / o / w double emulsion and their properties such as encapsulation efficiency, loading, release characteristics, and morphological properties were investigated. It was found that concentration of emulsifiers polyvinyl alcohol (PVA) and gelatin (varied between 0–4%) influenced the encapsulation efficiency appreciably. In order to increase encapsulation efficiency (from the obtained range of 18.1–30.1%) and slow down the release of the highly soluble tetracycline.HCl, it was neutralized with NaOH. Encapsulation efficiency of neutralized tetracycline was much higher (51.9–65.3%) due to the insoluble form of the drug used during encapsulation. The release behaviour of neither of the drugs was found to be of zero order. Rather the trends fitted reasonably well to Higuchi’s approach for release from spherical micropheres. Biodegradability was not an appreciable parameter in the release from microcapsules because release was complete before any signs of degradation were observed.  1999 Elsevier Science B.V. All rights reserved. Keywords: Polyhydroxybutyrate-co-hydroxyvalerate (PHBV); Tetracycline; Controlled release; Microencapsulation; Periodontal diseases

1. Introduction The gingival tissue is constantly subjected to mechanical and bacterial attacks. Gingivitis, which is the inflammation of the gingiva, is the most common form of gingival disease. The extension of the inflammation from the marginal gingiva into the supporting periodontal tissues marks the transition *Corresponding author. Tel.: 11-617-3733634; fax: 11-6173732209; e-mail: [email protected]

from gingivitis to periodontitis. Inflammation extends along collagen fiber bundles and goes through alveolar bone and may result in bone destruction [1]. Periodontitis is always preceded by gingivitis but not all gingivitis proceed to periodontitis. The selective removal or inhibition treatment of pathogenic microbes with either systemic or topically applied antibacterial agents combined with scaling and root planing is an effective treatment of advanced cases of periodontitis [2]. Tetracycline is the most abundantly tested and used antibiotic in the treatment of

0168-3659 / 99 / $ – see front matter  1999 Elsevier Science B.V. All rights reserved. PII: S0168-3659( 98 )00195-3

208

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

periodontal diseases. Clinical studies using tetracycline.HCl (TC) have shown it to have an effective spectrum of activity against many of the anaerobic microbes associated with the various periodontal diseases involving both adult and juvenile periodontitis patients [3,4]. TC has several inherent properties which enhance its potential use in the treatment of periodontal disease. These are the substantiality of TC to dentine and cementum surfaces, its ability to etch and / or remove the root surface smear layer and cause surface demineralization (chemical conditioning of the root surface), to delay pellicle and plaque formation, and to exhibit anti-collagenase activity [5]. As the systemic use of antibiotics may cause several side effects (sensitivity, resistant strains, superinfections), the local administration of antibiotics has received considerable attention [5]. Placing TC directly into the pocket avoids the systemic complications which were usually observed in previous studies. Local TC application has been accomplished using biodegradable carrier systems such as acrylic strips, fibers, gels, etc. [6–8]. The solid nondegradable, drug delivery systems require the clinician to remove the carrier at the end of the therapy with the risk of re-infecting. Thus more recent investigations have focused on local drug administration using biodegradable base or support materials or drug carriers [5]. Recently, novel biodegradable and biocompatible polyhydroxyalkanoates of biological origin started to draw the attention of scientists working in several fields of science, including medicine, pharmacy (and cosmetics), and agriculture. [8]. These polymers are ideal for use as biomedical materials due to their unique and interesting physicochemical features (e.g. piezoelectricity) in addition to mechanical properties similar to those of polypropylene and polylactic-coglycolic acids [9]. In this study, polyhydroxybutyrate-co-hydroxyvalerates (PHBV) with varying monomer ratios were used in the construction of controlled antibiotic systems to deliver TC or neutralized TC (TCN) and the properties of the resultant systems were analyzed by in vitro release studies. The morphological changes before and after release were assessed by SEM studies. Also, to determine the extent of retention of the biological activity of the drug after

microcapsule preparation procedures, bioassays were carried out using a tetracycline-sensitive organism, E. coli (C600) (MIC: 0.78).

2. Materials and methods

2.1. Materials Poly(3-hydroxybutyrate-co-3-hydroxyvalerate)s with hydroxyvalerate contents of 7, 14 and 22% (molar) and with molecular weights ranging between 400 000–750 000 Da were purchased from Aldrich (USA). Gelatin (bacteriological grade) was obtained from Oxoid (England). Tetracycline.HCl was a generous gift of Dr I. Yilmaz of FAKO Pharmaceutical Company (Turkey). Tetracycline base form was prepared from this acidic form in our laboratory by neutralizing the HCl with equimolar NaOH. PVA (molecular weight 14 000), acetic acid (glacial), acetone and chloroform (analytical grade) were obtained from Merck AG (Germany). n-Octanol was obtained from BDH (England). Dialysis bag (D9277, molecular weight cut-off 12 400, width 2 cm) was from Sigma (USA).

2.2. Preparation of TC and TCN loaded PHBV microparticles Both forms of the drug were used during experiments. PHBV (7, 14 and 22% HV content) microcapsules were prepared with the TC by using a double emulsion (w / o / w), solvent evaporation technique as described earlier [10,11]. Gelatin solutions (1.5, 2, 2.5 and 4% (w / v)) were prepared in distilled water to be used as the aqueous continuous media together with aqueous PVA solutions (1, 2, 4% (w / v). Gelatin (100 cm 3 ) and PVA (2 cm 3 ) solutions were mixed with an overhead stirrer (Cole Parmer Instruments, USA) for 10 min at the setting of 1 (¯750 rpm). PHBV was dissolved in chloroform (100 mg / cm 3 ). Tetracycline.HCl was dissolved in distilled water (200 mg / cm 3 ) helped by a brief sonication application for 10 s (at setting 20) by Ultrasonic homogenizer (4710 series Cole Parmer, USA). The drug solution (0.3 cm 3 ) was then added to the polymer solution (3 cm 3 ) in a 10-cm 3 glass

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

vial and sonicated again for a more homogeneous mixing (drug to polymer ratio was 1:5). This solution was added into the aqueous gelatin and PVA continuous medium (102 cm 3 total) in a dropwise fashion while stirring with the overhead stirrer at 500 rpm. After all the solution was added, nitrogen atmosphere was created in the flask to help solvent removal and the solution was stirred until all the chloroform was evaporated. The solution was then filtered through a sintered glass filter (Pyrex 2) to collect the microcapsules, which were then washed with distilled water (¯200 cm 3 ) and vacuum dried. In order to encapsulate neutralized tetracyline, it was mixed with equimolar NaOH solution and titrated to pH 7. This neutralized form (TCN) was then used in the preparation of microspheres with a slight modification of the above stated method in which drug was introduced into the PHBV solution in crystalline form rather than in an aqueous solution forming a single emulsion (o / w) leading to microspheres rather than microcapsules.

209

Encapsulation efficiency (%) 5 (Weight of tetracycline found loaded) / (Weight of tetracycline input) 3 100 Loading 5 (Weight of drug (g)) / (Weight of microparticle (g)) 3 100

2.5. Release studies Tetracycline.HCl or neutralized tetracycline loaded microparticles (15 mg) were introduced into dialysis bags and introduced to PBS (0.1 M, pH 7.4, 100 ml) which was continuously stirred by a magnetic stirrer at ambient temperature. At certain time intervals, samples (5.0 ml) were removed from the release medium, tetracycline content was measured spectrophotometrically (at 360 nm) and were returned. Results of triplicate tests were used to calculate the released tetracycline.

2.3. Particle size analysis for microparticles

2.6. Scanning electron microscopy ( SEM)

Analysis of particle size of the microparticles was made in acetone / octanol 18:2 v / v solution. After obtaining a suitable obscuration with a sample population, particles sizes were measured by a particle size analyzer (Mastersizer S Ver. 2.15, Malvern Instruments, Malvern, UK).

Tetracycline.base and tetracycline.HCl loaded microparticles before and after release studies were coated with gold under vacuum and their scanning electron micrographs were obtained using a Jeol (JSM-6400, Japan) scanning electron microscope.

2.4. Determination of encapsulation efficiencies and loading

3. Results and discussion

Microparticles (10 mg) were dissolved in chloroform (1 cm 3 ) and the tetracycline.HCl or TCN in the microparticles was extracted by adding physiological phosphate buffer (PBS) solution (pH 7.4, 0.1 M, 10 cm 3 ) and removal of the chloroform by air provided with an air pump. This process was repeated several times until the polymer obtained as a white, tetracycline free, precipitate. The resultant tetracycline solutions were analyzed with a UV/ Vis spectrophotometer (Shimadzu, Model 2100s, Japan). This extraction process was applied at least to two sets of microparticles and mean of the measurements was used to calculate the encapsulation efficiency and loading:

3.1. Influence of gelatin and polyvinylalcohol ( PVA) concentration on encapsulation It was observed that upon changing of the microcapsule preparation conditions, the extent of loading could be changed (Table 1). In the samples T 1 to T 4 , the only variable reaction parameter was the PVA concentration. It was observed that as its concentration increased, the encapsulation efficiency and the loading increased accordingly. But increase beyond an optimum value led to a gradual decrease. The same ‘bell-curve’ trend was observed with the variation of gelatin concentration, too. The maximum encapsulation efficiency and load-

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

210

Table 1 Properties of PHBV 7 microcapsules loaded with TC in media with varying PVA and gelatin concentration (25628C) Sample code

PVA concentration (% w / v)

Gelatin concentration (% w / v)

Encapsulation efficiency (%)

Loading (mg TC / 100 mg MCs)

Particle size a (mm)

T1 T2 T3 T4 T5 T6 T7 T8

0 1 2 4 1 1 1 1

2 2 2 2 1.5 2 2.5 4

20.260.8 30.161.6 28.060.8 22.360.6 18.160.2 30.161.6 26.860.5 22.661.6

3.8860.17 5.6760.28 5.3360.14 4.2760.12 3.5060.04 5.6760.28 5.0960.09 4.3360.34

399 340 – 357 322 340 538 –

a

Volume mean diameters.

ing were obtained when 2% gelatin and 1% PVA were used. The presence of an optimum value could be explained by considering the role of emulsifier. Below a critical concentration, the amount of emulsifier in the medium is not enough to stabilize all the microcapsules and thus some of the drug is dissolved in the aqueous phase and lost before the evaporation of the solvent (and hardening of microcapsules). On the other hand, when the emulsifier used is more than the critical level it coats all the surfaces excessively and might interact with available free drug during microcapsule formation resulting in its leakage into the aqueous phase. The adhesion and retention of PVA on PHBV surfaces was also shown in a similar study [12], where the use of PVA resulted in increased surface hydrophilicity even after extensive washing while BSA was easily removable.

3.2. In vitro bioassays; retention of antibiotic activity during processing After the extraction of TC from microcapsules (in triplicate), two different concentrations were obtained for each set by dilution (¯200 and 100 mg / cm 3 ) and their UV absorbances were measured. The expected zone sizes were 6 and 7 mm for 100 and 200 mg / cm 3 , respectively. In the study, 6 and 7.03 mm were observed for 103 and 207 mg / cm 3 , respectively. TCN extracted from microspheres also exactly matched with the calibration curve as was observed above. These show that the processing conditions did not adversely affect the bioactivity of the drug.

3.3. Mechanism of drug release Release from microparticles, where the controlling parameters are the diffusion coefficient of the drug in the support and the geometry, can be expressed according to the following equation [13]: ]] Mt Deff ? t | ] 5 ]] Mi p ? r2

œ

(for spheres)

where Mt /Mi are the amounts of drug released at time t and infinity, respectively, and D is the diffusion coefficient of the drug. When the slopes of Higuchi’s fractional release vs square root time curves for the microcapsules prepared with different PVA and gelatin concentrations are examined a decrease is observed in the release of the initial 30% of the drug content when excess PVA (4%) or gelatin (T7–T8) is used (Table 2 and Figs. 1 and 2). This was earlier explained to be due to presence of bound stabilizers on the surface, which must first be dissolved from the surface to let the pores become available for water entrance. These trends in the slopes might also be due to large particle size observed especially for T7 and T8 (volume mean diameters for these data are given in Table 1). In the variable gelatin set (Fig. 2), T5 released all | 105 h while T6 released the tetracycline within 5 60%, T7 released 55% and T8 released 45% within the same period. This is also supported by the slope values. When this is checked against Table 2 it is observed that these release rates are decreased as the gelatin concentration is increased. In all these release graphs direct linearity with respect to square root

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

211

Table 2 Effect of gelatin and PVA concentration on release of TC from PHBV 7 microcapsules Sample code

PVA (%)

Gelatin (%)

Lag Time (h)

Slope (h 21 / 2 )

R2

T1 T2 T3 T4 T5 T6 T7 T8

0 1 2 4 1 1 1 1

2 2 2 2 1.5 2 2.5 4

0.743 0.424 0.406 0.708 0.406 0.424 0.382 0.449

0.105 0.108 0.106 0.050 0.106 0.108 0.073 0.054

0.975 0.989 0.989 0.956 0.989 0.989 0.993 0.984

time was observed for the initial release phase (30%).

3.4. Effect of polymer type and drug form on encapsulation efficiency and loading Effect of polymer type on encapsulation efficiency, loading and release was investigated using PHBV copolymers with different HV contents (7, 14 and 22%) while keeping all other experimental conditions such as PVA, gelatin, and polymer concentration constant. The copolymer ratios presented here are those provided by the supplier and we believe that PHBV 14 does not come from a lot similar to those of the other two because in our earlier studies PHBV 14 behaved in a way un-

expected of its reported composition. In this study also it was the same. However, only these three commercial polymers were available and therefore they were used. The results showed that the highest encapsulation efficiency was obtained with PHBV 14. For the TC loaded microcapsules, it can be deduced that the encapsulation efficiency and loading of PHBV 7 and 22 decreases, as the HV content is increased (Table 3). Gangrade and Price [14] observed a similar trend with PHB, PHBV 9, and PHBV 24 where the drug loading values were 2.24, 1.82, and 1.93%, respectively. This could be explained with the lower crystallinity of PHBV 22 microcapsules due to its higher valerate content compared to PHBV 7. So encapsulation of TC.HCl with PHBV 7 brings out an

Fig. 1. Release of TC from PHBV 7 microcapsules prepared with different concentrations of PVA solution.

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

212

Fig. 2. Release of TC from PHBV 7 microcapsules prepared with different concentrations of gelatin solution.

advantage over PHBV 22 copolymer since higher encapsulation efficiency is a desired goal for controlled drug release studies. In case of TCN loaded microspheres, the encapsulation efficiency was slightly higher for PHBV 22 than PHBV 7, with PHBV 14 behaving unpredictably as in our earlier studies (lower than the rest) (Table 3). It was observed that, during mixing of polymer and drug, TCN solubility was higher in the polymer solution than TC and therefore mixing yielded a more homogeneous PHBV/ TCN solution. This effect is thought to be more distinct in the case of PHBV 22 polymer solution, because its dissolution was much more complete (increasing valerate content increases solubility). It thus formed a better drug dispersion that can hold more TCN during microsphere formation. In TCN loading process, the

decrease of the drug solubility in water and introduction of the drug in powder form, prevented excessive drug loss into the aqueous continuous phase and at the same time drug localization was much more in the polymeric coat than in the core. Higher TCN loading compared to TC and lower release rates are consequences of the above observations.

3.5. Effect of polymer type and drug form on release behaviour It was observed that TC release from PHBV 22 was about 90% complete in 100 h, while it was only 50% from PHBV 7 (Fig. 3). The release with PHBV 7, showed almost a zero-order behaviour while with PHBV 22 this was closer to the Higuchi prediction

Table 3 Properties of TC- and TCN-loaded PHBV microcapsules Microcapsule type

PHBV 7 PHBV 14 PHBV 22 a b

PVA (% w / v)

1 1 1

Gelatin (% w / v)

2 2 2

EE (%)b

Loading a

TC

TCN

TC

TCN

30.161.6 35.961.1 25.860.5

63.360.8 51.961.3 65.363.9

5.760.3 6.760.2 4.960.1

11.060.1 9.360.2 11.660.6

mg drug / 100 mg microcapsules. Encapsulation efficiency (%): percentage of the initially used drug that could be encapsulated within microparticles.

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

213

Fig. 3. Release of TC from PHBV 7, 14, 22 microcapsules.

(see relevant plots in Fig. 3). For PHBV 14 at 100 h only 42% was released. The trend seen in loading and encapsulation efficiencies (Table 3) was also reflected on the release behaviour. The unexpected PHBV 14 performance was thought to be due to the difference in molecular weight of the polymer [14]. To check this PHBV 14 was further characterized by viscosimetry, and NMR. Results indicated that the polymer has a distinctly lower molecular weight than the value stated by the supplier (further confirmed by mechanical tests [15]), possibly causing the deviations from the expected. From the Mt /Mi vs t 1 / 2 plots of the results it is observed that in the case of PHBV 22 the release yields an almost perfect fit (R 2 50.9897). Since a microcapsule with a drug core was aimed while the result implies a behaviour rather like that of a microsphere, it looks as if either most of the drug is entrapped close to the inner surface or within the coat of the PHBV 22 type of microcapsules. Another explanation would be that the membrane has cracks or pores and transfer of TC through the coat is easier due to membrane discontinuity. Possibly, the copolymer coat entrapped fraction of drug is released giving a direct proportionality with square root of time which is obtained for PHBV 7 and 14 during the initial phase of the release. This was also supported by Gangrade and Price [14], who found in

a similar study that a higher valerate content leads to a higher porosity and since the drug is also in the coat, the increased surface area due to the pore walls lead to higher release rates. When Fig. 3 is examined it is seen that even though the results are very close, the differences are still significant and PHBV 14 has the highest initial slope (most rapid release) followed by PHBV 22 and PHBV 7. Normally PHBV 22 should release the fastest due to its chemical composition. Fig. 3 shows that while at later stages PHBV 22 maintains this rapid release, PHBV 7 and 14 do not. It is known that hydroxyvalerate component of the copolymer leads to a more amorphous structure thus decreasing the crystallinity of the molecule. So, it might be possible that the drug could partition into this more amorphous, HV rich regions and is released through HV channels. The partition coefficient of TC (0.22 o / w) also suggests that the drug would prefer the more hydrated, HV rich domains. The biphasic release observed in all cases (except PHBV 22) in this study is nearly a characteristic of PHBV microparticles observed in many studies [16]. This result also suggests that in this short period of time (insufficient for the degradation of PHBV copolymer) release is mainly dependent on the diffusion of the drug through the pores of wall and thus diffusion is the rate determining step, and the initial phase of

214

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

Table 4 Properties of TC- and TCN-loaded microparticles Microcapsule type

EE (%)b

Loading a

Lag time (h)

Slope (h 21 / 2 )

PHBV 71TC1TCN PHBV 71TC PHBV 71TCN

36.261.0 30.161.6 63.360.8

6.860.2 5.060.3 11.060.1

0.689 0.406 0.535

0.033 0.106 0.025

a b

mg drug / 100 mg microparticles. Encapsulation efficiency (%): percentage of the initially used drug that could be encapsulated within microparticles.

release was thought to occur mainly by dissolution and diffusion of drug entrapped close to or at the surface of the microcapsules. The second and slower phase release was thought to involve the diffusion of drug entrapped within the inner part of the polymer matrix by means of aqueous channels of a network of pores [17]. This decline is also due to depletion of the drug content and because of the shape of the microcapsules, within which the surface facing the medium decreases as one goes towards the center from the surface. An interesting observation is the general presence of a lag period (not so apparent in presentation in ‘release vs t 1 / 2 ’ form), which is indicative of the need for solvent penetration into the structure. It appears that this stage is quite significant in these microcapsules.

The original aim of neutralization of TC was to decrease its solubility and thus its release rate from the microcapsules. It also made possible the use of the drug in powder form in o / w encapsulation. When the results (Table 4, Fig. 4) are examined, it is observed that in PHBV 7, the microspheres released ¯20% in 800 h while PHBV 14 released ¯55% and PHBV 22 released only 18% in 480 h. So the TCN looks mostly well entrapped especially within the microspheres of PHBV 7 and 22 for which encapsulation efficiencies are also quite high (Table 3). When the slopes are compared with those of TC released from same polymers, it is observed that the slopes with TCN are between 50 and 80% lower in the case of TCN. Thus much slower release is achieved with the use of TCN.

Fig. 4. Release of TCN from PHBV 7, 14, 22 microspheres.

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

3.6. Release of TC and TCN from PHBV 7 microspheres In order to modulate release kinetics, several approaches are used. Among them are variation of the support material properties (such as using various forms of the polymeric carrier, as it is attempted in the present study) or modification of the drug itself (again as in this work changing of solubility). The release behaviour of two forms of the drug revealed that the TC has a much higher tendency to leach out than its counterpart TCN (see Figs. 3 and 4). This observation prompted us to co-encapsulate the two forms of the drugs (TC and TCN) in one preparation, and thus modify the overall release behaviour (Fig. 5). To test this, TC and TCN (1:1 weight ratio) was co-encapsulated in PHBV 7 microparticles. Encapsulation efficiency of this form was observed to be in between the separate forms, but closer to the TC preparation (Table 4). This might suggest that even a very small amount of aqueous solution within the microparticle decreases the homogeneity of the polymer–drug mixture and thus, microcapsule stability. Initial aim of using this approach was to obtain a bimodal release from the microsphere. As seen in

215

Fig. 5, the release behaviour is not bimodal but more in accordance with Higuchi’s expression. Likewise the value of the slope of the Higuchi curve for co-encapsulated TCN and TC is between the slopes of their individual cases. Release graphs showed that 45% of TC was released within ¯300 h for the two drug loaded case (Fig. 5) but in 50 h for TC microcapsules. Release of drug from TCN microspheres was ¯15% in 300 h. This release data shows that two drug loaded microparticles give a release profile which is between separate loadings of TC and TCN, and closer to that of TC. The reason for this is that although it appears that there were two models of release, the loading process necessitated an encounter of the organic PHBV solution in an organic solvent and TCN granules thus leading to their dissolution. Therefore, encapsulation in powder form was not possible. The results confirm this processing problem.

3.7. SEM results of TC- and TCN-loaded microparticles As can be seen from the SEM micrograph (Fig. 6) of the particle, the form is capsular as expected from the preparation method. The PHBV 7–TC microcap-

Fig. 5. Effect of drug type on release from PHBV 7 microcapsules.

216

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

Fig. 6. SEM micrograph of TC-loaded PHBV 7 microcapsule (3150; bar: 100 mm).

Fig. 8. SEM micrograph of the surface of TCN-loaded PHBV 7 microsphere before release (32300; bar: 10 mm).

sule with the large opening was presented also to give an idea about the thickness of the capsule wall. The wall is quite thin (¯7 mm), which partially explains the rapid release of the TC loaded into these microcapsules. TCN loaded microparticles seemed more like spheres than capsules which probably is a result of the use of single o / w emulsion (Fig. 7). Also because of the incorporation of the drug into the structure, the surface texture is rough (Fig. 8). For PHBV 7–TCN microspheres one of the most important observations was that the surface erosion was very apparent (Fig. 9) causing the formation of large cavities and suggesting the loss of polymer particles or granules (into release medium) and of low molecular weight chains. This micrograph also indicates that release occurs not only by dissolution

and diffusion out of the structure but also as chunks removed from the surface.

Fig. 7. SEM micrograph of the surface of TCN-loaded PHBV 7 microspheres after release (34000; bar: 1 mm).

4. Conclusion This study was initiated to construct a biodegradable drug release system for the treatment of periodontal diseases. It was found that emulsifiers in the continuous medium for microencapsulation of drug (gelatin and PVA) were influential on the encapsulation efficiency in microparticles. Another factor which significantly affected encapsulation efficiency was the drug chemistry. The neutralized form of the originally acidic drug could be loaded more. This was actually a result of the drug form

Fig. 9. SEM micrograph of the surface of TCN-loaded PHBV 7 microspheres after release (31500; bar: 10 mm).

D. Sendil et al. / Journal of Controlled Release 59 (1999) 207 – 217

than its chemistry. The release data for these samples did not fit a zero-order behaviour. They could best be expressed by Higuchi’s drug release equations for spherical matrix systems which depend on diffusion characteristics. The release profiles of the neutralized tetracycline-loaded microspheres gave a release considerably slower than those with the hydrochloride salt. The release kinetic was still not zero order and nearly similar to the unneutralized drug. In conclusion it can be stated that it is possible to construct biodegradable tetracycline release systems which would deliver the drug directly into the periodontal pockets avoiding transport, metabolism and distribution problems, as well as the need for removal. Further tests should be carried to improve the properties and to enhance applicability of the system.

References [1] F.A. Carranza, Jr., F.R. Saglie, Glickman’s Clinical Periodontology, W.B. Saunders, New York, 1990. [2] C.L. Drisko, C.M. Cobb, W.J. Killoy, B.S. Michalowicz, B.L. Pihlstrom, R.A. Lowenguth, J.G. Caton, M. Encarnacion, M. Knowles, J.M. Goodson, Evaluation of periodontal treatments using controlled-release tetracycline fibers: clinical response, J. Periodont. 66 (1995) 692–698. [3] S.R. Smith, D.M. Foyle, I.G. Needleman, N.V. Pandya, The role of antibiotics in the treatment of periodontitis. I. Systemic delivery), Eur. J. Prostodont. Rest. Dent. 3(2) (1994) 79–86. [4] R.M. Palmer, T.L.P. Watts, R.F. Wilson, A double-blind trial of tetracycline in the management of early onset periodontitis, J. Clin. Periodont. 23 (1996) 670–674. ¨ [5] E. Unsal, M. Akkaya, T.F. Walsh, Influence of a single application of subgingival chlorhexidine gel or tetracycline

[6] [7]

[8] [9] [10]

[11] [12]

[13]

[14]

[15]

[16]

[17]

217

paste on the clinical parameters of adult periodontitis patients, J. Clin. Periodont. 21 (1994) 351–355. R. Duncan, L.W. Seymour, Controlled Release Technologies, Elsevier, Amsterdam, 1989. E. Esposito, V. Carotta, A. Scabbia, L. Trombelli, P. D’Antona, E. Menegatti, C. Nastruzzi, Comparative analysis of tetracycline-containing dental gels: polaxomer- and monoglyceride based formulations, Int. J. Pharm. 142 (1996) 9–23. Y. Doi, Microbial Polyesters, VCH, New York, 1990. P.A. Holmes, Applications of PHB: a microbially produced thermoplastic, Phys. Technol. 16 (1985) 32. ¨ I. Gursel, V. Hasirci, Properties and drug release behaviour of poly (3-hydroxybutyric acid) and various poly (3-hydroxybutyrate-hydroxyvalerate) copolymer microcapsules, J. Microencapsul. 12(2) (1995) 185–193. D. Sendil, Antibiotic release from biodegradable microbial polyesters, MSc thesis, METU, Ankara, 1997, F. Boury, H. Marchais, J.P. Benoit, J.E. Proust, Surface characterization of poly(a -hydroxy acid) microspheres prepared by a solvent evaporation / extraction process, Biomaterials 18(2) (1997) 125–136. R. Gurny, E. Doelker, N. A Peppas, Modelling of sustained release of water soluble drugs from porous hydrophobic polymers, Biomaterials 3 (1982) 27–32. N. Gangrade, J.C. Price, Poly(hydroxybutyrate-hydroxyvalerate) microspheres containing progesterone: preparation, morphology and release properties, J. Microencapsul. 8(2) (1991) 185–202. ¨ I. Gursel, C. Balc¸ik, Y. Arica, O. Akkus, N. Akkas, V. Hasirci, Interpenetrating networks of polyhydroxybutyrateco-hydroxyvalerate, and polyhydroxyethyl methacrylate: mechanical properties, Eng. Syst. Des. Anal. 5 (1996) 107. F. Koosha, R.H. Muller, S.S. Davis, Polyhydroxybutyrate as a drug carrier, Crit. Rev. Ther. Drug Carrier Syst. 6 (1989) 117–130. S. Akhtar, W. C Pouton, J.L. Notarianni, The influence of crystalline morphology and copolymer composition on drug release from solution cast and melt-processed P(HB-HV) copolymer matrices, J. Control. Release 17 (1991) 225–234.