Progesterone-loaded albumin microparticles

Journal of Controlled Release, 21 ( 1993 ) I-7 0 1993 Elsevier Science Publishers B.V. All rights reserved

0168-3659/93/$06.00

COREL 00887

Progesterone-loaded

albumin microparticles

Isabella Orienti and Vittorio Zecchi Dipartimento di Scienze Farmaceutlche, Bologna, Italy (Received 4 August 1992; accepted in revised form 7 June 1993 )

Bovine serum albumin (BSA) was used to prepare microspheres and microcapsules containing progesterone. The microparticles were obtained by means of a multiple emulsion method and by thermal or chemical crosslinking. The drug loading was higher for microspheres than microcapsules and varied for the different crosslinking methods: higher with thermal than chemical crosslinking and, among the chemical crosslinkers, loading was higher for formaldehyde than 2,3-butanedione and glutaraldehyde, respectively. Drug release was well differentiated for the different systems: the microspheres released 60% of their content according to an anomalous diffusion mechanism in the shorter periods, whereas microcapsules released 60% of their content according to different kinetic trends and for longer periods of time. The duration of each kinetic period and the relative drug fraction released, depends on the crosslinking method (thermal or chemical) and the nature of the crosslinking agent. Among the systems examined, the glutaraldehyde crosslinked microspheres show the higher release rate while the thermally crosslinked microcapsules had the lowest release rate which is characterized by the longest lasting zero order period. Key words: Bovine serum albumin; Progesterone; Multiple emulsion method; Thermal Chemical crosslinking; Drug loading; Diffusion; Bulk erosion; Release kinetics

Introduction The use of biodegradable polymeric microparticulate systems is an interesting acquisition in the control of drug release and targeting. These systems can be manufactured from a large number of starting materials and by different preparatory methods which generally consist in the utilization of liquid systems from which a solid phase containing the polymer and the drug may be obtained by physical or chemical methods. The yield, drug content and particle size distribution depend on different factors such as the Correspondence to: Isabella Orienti, Dipartimento di Scienze Farmaceutiche, Via S. Donato 19,40 127, Bologna, Italy.

crosslinking;

nature of the polymer and formulative and preparative methods [ l-41. Among the biodegradable polymeric materials employed in the preparation of microparticles, bovine serum albumin (BSA) has met increasing attention as it is suitable for producing non antigenic microparticles whose physico-chemical properties can be widely modulated by the crosslinking methods employed for their production and the nature of the crosslinking agent [ 5- 12 1. The drug may be released from these particles according to kinetics approaching zero order for defined periods of time [ 6,7]. The use of albumins, however, has generally been restricted to the transport of hydrophilic drugs. In this work we describe the use of BSA microparticles as carriers of a lipophilic

2

drug, progesterone. Particular attention has been paid to the formulative and preparative aspects correlated with the main physico-chemical characteristics of the system and with the kinetic aspects of drug release. Experimental

of corn oil as the inner phase of the multiple emulsion, the permanence of a liquid phase within the particles suggests the formation of microcapsules. The dimensions of the particles ranged from a minimum of 5 to a maximum of 50 pm both for spheres and capsules. Micropartitles with a mean geometric diameter of 20& 5 pm were selected for the present study.

Materials Optical microscopic observation of microparticles BSA, gelatin type B, corn oil, formaldehyde, 2,3-butanedione, glutaraldehyde and progesterone were all purchased from Sigma Chemicals (St. Louis, MO, USA). Methylene chloride, diethy1 ether and ethanol were purchased from Carlo Erba Analyticals (Milan, Italy). Microparticles preparation Albumin microspheres and microcapsules were prepared by a multiple emulsion method [ 2 1. Progesterone (78.8 or 28.8 mg) was dispersed in 0.5 ml corn oil or methylene chloride. The dispersion was mixed with 1 ml of an aqueous solution containing 25% BSA and 1% gelatin. The mixture was stirred for 10 min to produce an o/ w emulsion. The emulsion was added to 3 ml of corn oil and the mixture stirred again for 2 min to obtain the corresponding o/w/o multiple emulsion. Thermal or chemical crosslinking were used to harden the albumin dissolved in the aqueous phase. When thermal crosslinking was employed, the multiple emulsion was frozen to 0’ C and subsequently added to 100 ml of corn oil heated to 120°C and stirred for 20 min. Alternatively, when chemical crosslinking was utilized, the multiple emulsion was added to 100 ml corn oil heated at 37 “C, containing 0.1 M of formaldehyde or 2,3-butanedione or glutaraldehyde. The mixture was stirred for 60 min. The microparticles obtained were separated by centrifugation and washed with 100 ml diethyl ether. In the presence of methylene chloride as the inner phase the particles were subsequently warmed at 40’ C under a vacuum to a constant weight to evaporate the methylene chloride and form an internal completely solid phase which suggests microsphere formation. In the presence

Microspheres and microcapsules were observed under an optical microscope (AxiophotZeiss, D) to examine their surface, size and shape. Determination microparticles

of progesterone content in the

The amount of progesterone entrapped in the microspheres and microcapsules was determined by digesting the microparticles in a mixture of ethanol and NaOH 0.5 N aqueous solution and spectrophotometrically detecting the drug in the mixture. Determination of water uptake and weight loss Water uptake and weight loss were followed on unloaded microparticles obtained from the 25% BSA aqueous solution containing 1% gelatin where BSA was crosslinked by the thermal or chemical methods previously described. One hundred mg of the microparticles were immersed in 100 ml of pH 7.4 aqueous buffer at 37°C and at the indicated time intervals, water uptake and weight loss were measured by drying the microparticles to constant weight in vacuum at room temperature [ 131 (I. Orienti, A. Coppola and V. Zecchi, work in progress). In vitro release studies Twenty mg of microspheres or microcapsules were suspended in 200 ml of pH 7.4 buffered aqueous solution. The resulting suspension was thermostated at 37°C. At appropriate time intervals microparticle-free supemate samples were

3

obtained and the drug was detected spectrophotometrically utilizing the corresponding unloaded microparticles as a reference. Results

affinity of the oily inner phase of the o/w/o emulsion (methylene chloride) towards the external lipidic phase (corn oil) and to the consequent higher stability of the double emulsion with respect to the microcapsules which inhibit drug migration.

Drug loading

Water uptake and weight loss

Table 1 reports the loading levels and the yields of BSA microspheres and microcapsules. The table shows that the BSA microspheres and microcapsules obtained by thermal crosslinking show higher drug levels with respect to those obtained by chemical crosslinking. Among the chemical crosslinkers formaldehyde gave a higher drug content with respect to 2,3-butanedione and glutaraldehyde, respectively. Thermal crosslinking probably reduces the time interval in which the drug can diffuse from the oily inner phase through the albumin aqueous solution due to its more rapid establishment compared with chemical crosslinking where the crosslinking agent has to diffuse from the external oily phase towards the aqueous albumin phase to start crosslinking. The differences observed between the different crosslinkers are probably due to their different partition rates between these two phases. The higher drug content observed for microspheres than microcapsules can be attributed to the lower

The water content of the differently crosslinked BSA microparticles and their weight loss are reported as a function of time in Fig. 1. All the systems studied show an increase in water content which tends to a plateau before the start of weight loss. The rates of water uptake and the plateau values are higher for glutaraldehyde crosslinked BSA and decrease for 2,3-butanedione, formaldehyde and thermally crosslinked BSA micropa~iclcs, respectively. This sequence is in accordance with the presence of tighter crosslinks between the BSA macromolecules which probably allow the formation of more packed solid structures characterized by a slower water penetration and a lower amount of water taken up by the polymer for its complete hydra-

50.

/.’

/’

-50

8’

40 -

TABLE I Progesterone content and yield of BSA microspheres and microcapsules the~ally and chemically crosslinked (theoretical progesterone content: 10%) Crosslinking agent or temperature (“C)

Actual drug content (wt%)

Yield (O/O)

Microspheres

120°C formaldehyde 2,3-butanedione glutaraldehyde

8.01 7.45 7.22 5.71

80.27

Microcapsules

120°C formaldehyde 2,3-butanedione glutaraldehyde

4.98 4.51 4.30 3.88

69.50

0

20

SO

40

Time (days) Fig. 1. (---) Increase in the water content of BSA microparticles as a function of time in pH 7.4 aqueous buffer at 37’C: 0, thermal crosslinking; A, chemical crosslinking with formaldehyde; l , chemical crosslinking with 2,3-butanedione; 0, chemical crosslinking with glutaraldehyde. (-----) Weight loss of BSA microparticles as a function of time in pH 7.4 aqueous solution at 37°C: 0, thermal crosslinking; A, chemical crosslinking with formaldehyde; 0, chemical crosslinking with 2,3-butanedione, L], chemical crosslinking with glutaraldehyde.

4

tion. Weight loss started after induction periods varying from 10 days for glutaraldehyde crosslinked BSA to 23.8 days for thermally crosslinked BSA microparticles and increased almost linearly to 18-56% after 40 days. As seen in Fig. 1, weight loss takes place first in the glutaraldehyde crosslinked and then in the 2,3-butanedione, formaldehyde and thermally crosslinked BSA microparticles, respectively. This behaviour may be explained by the wider macromolecular meshes of the ~utar~dehyde crosslinked BSA which allow diffusion of the oligomers formed following erosion even at low erosion levels. Kinetic analysis of release

Progesterone release The progesterone release profiles from BSA microspheres and microcapsuies are reported in Figs. 2 and 3 for the different crosslinking methods, respectively. The profiles show an initial period characterized by a decrease in release rate over time, followed by an almost constant release rate period and a subsequent period of increasing release rate up to drug depletion. In order to define the kinetic characteristics of each period and to determine their extent, the release

20

10

Time

30

40

(days)

Fig. 2. Fractional release of progesterone from BSA microspheres in pH 7.4 aqueous buffer at 37°C. 0, thermal crossfinking; A, chemical crosslinking with formaldehyde; n , chemical crosslinking with 2,3-butanedione; h, chemical crosslinking with ghttaraldehyde.

o,o+ 0

10

20

30

40

TimeYdays)

Fig. 3. Fractional release of progesterone from BSA microcapsules in pH 7.4 aqueous bufferat 37°C. 0, thermal crosslinking; A, chemical crosslinking with formaldehyde; n , chemical crosslinking with 2,3_butanedione; L, chemical crosslinking with glutaraldehyde.

data have been analysed according to the general equation: ~~~~/~~~ =X-tn as tong as 0 < ~~~/~f~ I 0.6

(1)

which correlates the fractional amount released to the diffusional constant k and the diffusional exponent n. For the microspheres, the ln(M,/ M,) vs lnt gave a linear trend up to M,/ M,=O.6. The n values obtained ranged from 0.52 for the thermal crosslinking to 0.69 for the chemical crosslinking with glutaraldehyde, indicating an anomalous release (Table 2). For microcapsules, the In (~~/~~) vs lnt showed three different features: an initial linear trend whose n values range from 0.47 for the thermal crosslinking to 0.60 for the chemical crosslinking with glutaraldehyde, indicating an anomalous release; a second linear trend whose n values ranged from 0.82 for the thermal crosslinking to 0.88 for the chemical crosslinking with glutaraldehyde, indicating a release close to zero order; and a third, non-linear trend up to M,/M, = 0.6 (Table 2). This behaviour may be explained by the influence of polymer hydration and bulk erosion on release. As long as weight loss does not occur, the release from the polymeric structure may be described by the equation: M,=S (2 C,D, exp(k’r)c‘J)“2

(2)

TABLE 2 Release characteristics of progesterone from BSA microspheres and microcapsules: days); related diffusional exponents (n) and drug fraction released (M,/M,) Crosslinking agent or temperature (“C)

120°C

formaldehyde

2,3-butanedione

glutaraldehyde

duration of the different kinetic periods (d,,

Microspheres

Microcapsules

First period

First period

Second period

Third period

6.03 0.52 0.60

9.05 0.47 0.22

1448 0.82 0.16

7.51

4.54 0.56 0.60

8.09 0.40 0.26

13.51 0.86 0.14

3.57 0.65 0.60

6.12 0.57 0.32

12.03 0.86 0.10

2.20 0.69 0.60

4.52 0.60 0.53

5.54 0.88 0.05

where S is the releasing surface, C, is the drug solubility in the polymer, r>, is the drug diffusion coefficient in the polymer before the start of erosion, k’ is the rate constant of the polymeric chain cleavage and C, is the amount of drug initially present in the unit volume of the system. The progressive water absorption in the BSA microparticles is expected to decrease the solubility of the hydrophobic progesterone in the polymer while the erosion process is expected to exponentially increase its diffusion coefficient (D = D, exp( k' t) ) over time. The C, and D relative variations are expected to interfere with the release rate decrease taking place in the diffusion controlled release systems and which is generally attributed to an increase in the drug diffusional pathway over time [ 141, During the anomalous release period a decrease in C,, due to the water content increase in the polymeric structure (Fig. 1 and Table 2 ), and an increase in D may be supposed to allow an incomplete compensation of the release rate decrease, so shifting the release kinetics from Fickian to anomalous. During the following releases period an almost constant C, value, due to a constant water content (Fig. 1 and

0.22 5.20 0.20 4.01 0.18 1.56 0.02

Table 2) in the polymeric structure, and an increase in D may be supposed to allow a complete compensation for periods of time whose length depends on the physico-chemical properties of the system, so shifting the release towards a zero order kinetics. In the third release period, the lack of a linear correlation between In (MJM, ) and lnt, suggests that the release is no longer controlled by diffusion. Probably the erosion has reached sufftciently high levels to allow the establishment of polymer weight loss (Fig. 1 and Table 2), consequently overcoming any diffusive control on release. Comparison between the release from microspheres and microcapsules The length of the different kinetic periods and the corresponding drug fraction released vary for microspheres and microcapsules depending on the different crosslinking methods (Figs. 2,3 and Table 2 ) . The microspheres release 60~ of their content even during the anomalous period, the duration of this period is maximum for the thermal crosslinking and decreases with the chemical crosslinking by passing from fo~aldehyde to

6

2,3-butanedione and ~utaraldehyde, respectively. The release from microcapsules presents significant differences for the different crosslinking methods: the thermally crosslinked microcapsules release approximately the same fraction during the three kinetic periods, while the chemically crosslinked microcapsules release a higher drug fraction during the anomalous than the following two periods. This fraction increases from formaldehyde to 2,3-butanedione to glutaraldehyde, respectively, while the fractions released during the zero order and the erosion controlled periods decrease in the same way. For all the crosslinking methods the zero order period lasts longer than the anomalous and the erosion con-

trolled periods and the length of each period decreases from the thermal to the formaldehyde, 2,3-butanedione and glutaraldehyde crosslinked microcapsules. The higher release rate observed during the anomalous period for microspheres with respect to microcapsules (Figs. 2, 3) may be attributed to the higher porosity of the microspheres as observed from the optical photomicrographs of these systems (Fig. 4). The higher porosity promotes drug diffusion through the water filled pores of the microparticle, increasing the drug permeability through the releasing system. The more rapid release observed, both for microspheres and microcapsules in the presence of chemical rather than thermal crosslinking and, for the chemical crosslinking of glutaraldehyde with respect to 2,3-butanedione and formaldehyde, respectively, may be attributed to the formation of looser crosslinks between the albumin macromolecules in the presence of chemical as opposed to thermal crosslinking and, for the chemical crosslinking, in the presence of ~utaraldehyde with respect to 2,3-butanedione and formaldehyde respectively. The consequent higher water concentration present in the looser crosslinked networks (Fig. 1) allows a more rapid erosion and a consequent more rapid release in each kinetic period.

Conchsions

Fig. 4. Optical micrographs of: (A) microspheres thermally crosslinked; (B) microcapsules thermally crosslinked.

The multiple emulsion method seems suitable for the utilization of albumin in the preparation of microparticles containing lipophilic drugs such as progesterone and to obtain systems characterized by a well differentiated release. Microspheres, in fact, release most of their content according to an anomalous diffusion mechanism, whereas microcapsules release most of their content according to kinetic trends which vary over time. The crosslinking method (thermal or chemical) and the nature of the crosslinking agent, appreciably modify the length of the different kinetic periods and the relative fraction released during these periods. Along the systems examined, the microspheres crosslinked with glutaraldehyde show the most rapid release, while

7

the microcapsules thermally crosslinked show the slowest release which is characterized by the longest lasting zero order period. References E. Doelker, Cinttique et mecanismes de la liberation controlee g partir des systemes polymeriques, in P. Buri, F. Puisieux, E. Doelker and J.P. Benoit (Eds. ), Formes Pharmaceutiques Nouvelles, Lavoisier, Paris, 1985. R. Arshady, Albumin microspheres and microcapsules: methodology of manufacturing techniques, J. ControlledRelease, 14(1990) 111-131. R. Arshady, Preparation ofbiodeg~dable microspheres and microcapsules. 2. PoIylactides and related polyesters, J. Controlled release, I7 ( 199 1) l-22. T. Ishizaka, K. Endo and M. Koishi, Preparation of egg albumin microcapsules and microspheres, J. Pharm. Sci., 70 (1981) 358. S.P. Vyas, S. Bhatnagar, P.J. Gogoi and N.K. Jain, Preparation and characterization of HSA-propanolol microspheres for nasal administration, Int. J. Pharm., 69 (1991) 5-12. P.K. Gupta, C.T. Hung and D.G. Perrier, Albumin microspheres. II. Effect of stabilization temperature on the release ofadriamycin, Int. J. Pharm., 33 ( 1986) 147153.

PK. Gupta, C.T. Hung, F.C. Lam and D.G. Perrier, Albumin microspheres. III. Synthesis and characterization of microspheres containing adriamycin and magnetite, Int. J. Pharm., 43 (1988) 167-177. 8 K. Sugibayashi, Y. Morimoto, T. Nadai, Y. Kato, A. Hasegawa and T. Arita, Chem. Pharm. Bull., 27, 1 ( 1979) 204-209. 9 Y. Nishioka, S. Kyotani, M. Okamura, Y. Mori, M. Miyazaki, K. Okazaki, S. Ohnishi, Y. Yamamoto and K. Ito, Preparation and evaluation of albumin microspheres and microcapsules containing cisplatin, Chem. Pharm. Bull. 37,s (1989) 1399-1400. IO J.J. Burger, E. Tomlinson, E.M.A. Mulder and J.G. McVie, Albumin microspheres for intra-arterial tumor targeting. I. Pharmaceutical aspects, Int. J. Pharm., 23 (1985) 333-344. 11 K. Egbaria and M. Friedman, Sustained release albumin microspheres containing antibacterial drugs: effects of preparation conditions on kinetics of drug release, J. Controlled Release, 14 ( 1990) 79-94. 12 C. Jones, M.A. Burton and B.N. Gray, Albumin microspheres as vehicles for the sustained and controlled release of doxorubicin, J. Pharm. Pharmacol., 41 (1987) 813-816. 13 S.S. Shah, Y. Cha and C.G. Pitt, Poly(glycolic acid-coDL-lactic acid): diffusion or degradation controlled drug delivery? J. Controlled Release, 18 ( 1992) 26 l-270. 14 J.M. Crank, The mathematics of diffusion. Claredon Press, Oxford ( 1975). 7