- Email: [email protected]
Microencapsulation and Dissolution Properties of a Neuroleptic in a Biodegradable Polymer, Poly(d,l-lactide)
Microencapsulation and Dissolution Properties of a Neuroleptic in a Biodegradable Polymer, Poly(d,/-lactide) K. SUZUKI*' AND J. C. PRICE Received September 15, 1983, from the Department of Pharmaceutics, School of Pharmacy, University of Georgia, Athens, GA 30602. Accepted + Present address: School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, NC 27514. for publication May 3, 1984. Abstract 0 Polylactide
was polymerized from dilactide under various conditions to yield polymers in the molecular weight range from 11,000 to 21,000 as determined by osmometry. Chlorpromazine was encapsulated by the polylactide polymers by using an emulsification-solvent evaporation method. Microscopic observation revealed that when drug loading was %18%, the drug was in the form of a solid solution in microspheres of polylactide. At higher drug loadings, crystalline drug was present. In vitro dissolution of the encapsulated drug was followed in hydroalcoholic and aqueous buffer media. Studies with the hydroalcoholic media revealed that the drug release rate decreased as microcapsule size increased. However, dissolution in the hydroalcoholic media did not reflect the effect of molecular weight or percentage loading observed in the aqueous system. Dissolution in aqueous buffer showed that f500,0 increased with molecular weight and that release rates-surface area increased with increasina drua loadina. Long-acting neuroleptics administered by injection are of value in the treatment of chronic psychotic patients-in particular chronic schizophrenics-who have been discharged from a hospital but who rapidly suffer a relapse due to taking their medication incorrectly. Thirty to fifty percent of patients cease medication on their own initiative within 2 months to 1 year following discharge, leading to their readmission to psychiatric hospitals.'-3 The standard treatment of schizophrenia is the use of long-acting injectable esters of fluphenazine, such as fluphenazine decanoate and enanthate. These are injected intramuscularly as a sesame oil suspension once every 2-3 weeks in the dose range of 12.5-25 mg.4-7 The first application of polylactide as an erodible matrix for long-acting controlled release of a drug was reported by Yolles and co-workers in 1971.'Yolles and Sartorigand Sinclair" state that polylactide as a matrix shows, besides biodegradability, the following advantages: ( a ) compatibility with the environment and with many drugs, ( b ) rate of biodegradation compatible with the projected life span of the composite, and (c) ready availability of starting material and cost effectiveness of the polymer itself. The most widely reported uses of polylactide as a matrix for controlled release are in systems containing such compounds as narcotic antagonists, fertility controlling agents, anticancer agents, herbicides, pesticides, and others. However, the system containing neuroleptics has not been reported. In 1975, Thies" first applied the microencapsulation technique to make an injectable sustained-release formulation with polylactide, in which cyclazocine was encapsulated. The present investigation was undertaken to prepare different batches of microcapsules by altering encapsulation factors which influence the rate of drug release, and to determine by in vitro studies whether microcapsules could be made which would be suitable for in vivo testing. The physical factors chosen to be altered were particle size of the microcapsules, percentage of drug loading, and molecular weight of the encapsulating agent. Chlorpromazine was chosen as a model neuroleptic to be encapsulated in a polylactide. 0022-3549/85/0100-002l$O 1 .OO/O 0 1985, American Pharmaceutical Association
Experimental Section Materials-Polylactide was polymerized from both commercially available dilactide (Frinton Laboratories, Vineland, NJ) and dilactide made from lactic acid (J. T. Baker Chemical Co., Phillipsburg, NJ). Tetraphenyltin (Aidrich) was used as a catalyst for polymerization. Chlorpromazine was prepared from chlorpromazine hydrochloride (Sigma). Polyvinyl alcohol type I1 (Sigma) was employed as a suspending agent, sodium lauryl sulfate (Fisher Scientific) as a n ionic surfactant, and sodium sulfite (Fisher Scientific) as an antioxidant. All were USP grade. All other chemicals were analytical reagent grade, unless otherwise indicated. Preparation of Polylactide-The method was based on one reported by Kulkarni et a1.l' The crude dilactide was recrystallized twice using ethyl acetate to obtain purified dilactide with melting points between 124°C and 126°C. The dilactide was placed in a silanized vacuum flask with 0.01-0.02% tetraphenyltin. The contents were exposed to a vacuum pump for 1h to remove air and moisture and then polymerized in an oven at 160-170°C for 3-9 h. The resultingpolymer in the flask was dissolved in acetone using moderate heat and precipitated in a large quantity of water for purification. Intrinsic Viscosity Determinations-A Cannon-Fenske viscometer was used to determine the intrinsic viscosities of polymer-benzene solutions a t 25°C. The average molecular weight was calculated from the intrinsic viscosity by employing the Mark-Houwink equation: 1 = 3.64 X lo4 Mn0.75
(1)
where 1 is the intrinsic viscosity and Mn is the number-average molecular weight. The variables K , 3.64 x lo4, and a, 0.75, were reported by Shindler et al.13 End G r o u p Analysis-The concentrations of the carboxylic acid end-group present in the polymer samples were determined by a dye-interaction techniq~e.'~.''The concentrations of carboxylic acid were calculated using:I3
COOH = 1.77 x 10+ (AE/c~)
(2) where COOH is the concentration of the carboxylic acid group in the polymer in mol/g, A E is the absorbance of sample minus that of blank at 515 nm, and C, is the concentration of polymer solution in g/L. Osmometry-A static-elevation osmometer (constructed by the University of Georgia Instrument Shop) and a regenerated cellulose membrane (grade RC 51, Shleicher & Schuell, Inc., Keene, NH) were employed to determine the molecular weights of the polymer samples at 25°C. The absolute molecular weights determined by osmometry were plotted against the reciprocal of the carboxyl group content, and the number of carboxylic end-groups in each polymer molecule (i.e., one, in the case of polylactide) was estimated from the slope. Journal of PharmaceuticalSciences/ 21 Vol. 74,No. 1, January 1985
Encapsulation Procedure-The basic microencapsulation method was first patented in 193816in Britain and then in the United States in 1939,17and procedures describing the use of polylactic acid as the encapsulating agent were granted patents in 1978'' and 1980.'' These procedures can be classified as an emulsification-solvent evaporation method. Three to six grams of poly(d,l-lactide) were dissolved in 20 mL of methylene chloride, and chlorpromazine was dissolved in the polymer solution. The mixture was dispersed by stirring at 280 rpm in 60 mL of distilled water containing 5% by weight of polyvinyl alcohol and 0.025-0.05% by weight of sodium lauryl sulfate and sodium sulfite. The stirring was carried out under subdued light. After the organic solvent evaporated, the resulting microspheres were separated by filtration. The product was washed with water and hydrochloric acid solution and then dried under reduced pressure at room temperature to yield a white powder. The powder was sized by sieving, and a particle size analysis of each sieve fraction was done microscopically to obtain the mean diameters. Assay P r o c e d u r e f o r Total Chlorpromazine ContentTriplicate samples of -20 mg of the microspheres were accurately weighed. The drug was extracted into ethanol using moderate heat. The spectrophotometric assay method for chlorpromazine was based on those discussed in the U.S.P." and by Salzman and Brodie." I n Vitro Dissolution Studies-Dissolution studies were performed with two fluids: a hydroalcoholic solvent consisting of four parts of 95% ethanol and six parts of phosphate buffer, and an aqueous fluid buffered to pH 7.0. The hydroalcoholic solvent enhanced the extraction rate of chlorpromazine into the dissolution medium. Dissolution studies were carried out using the Levy beaker method discussed in the U.S.P.'' The assembly consisted of the following: a covered, 1000-mL roundbottomed flask; a variable speed motor (G. K. Helker Corp., Bellerose, NY); a stainless steel propeller. While the propeller was rotated at 130 rpm, samples of the dissolution medium were removed periodically and analyzed for chlorpromazine content. The dissolution medium was maintained at constant volume by adding to the flask the volume of dissolution medium which had evaporated during the previous 24 h. Since chlorpromazine is reported to undergo photo-oxidation, all dissolution experiments were done under subdued light.
Table I-Intrinsic Conditions
Conc. of
Tetrapheny,tin,
Viscosities of Polylactide at Various Preparation
Time,
,, Temperature,
YO
0.03 0.02 0.02 0.02 0.01 0.01 None
6 6 6 3 6 3 3
Intrinsic Viscosity,
OC
dL/g
160 170 160 160 160 160 160
0.34 0.28 0.68 0.42 0.58 0.30 0.13'
'The polymer yield was low, 40%.
Table Il-Three Methods Used to Determine the Molecular Weights of Polvmers in the MicroencaDsulation ExDeriments'
Osmornetrv 21,400 20,500 16,300 12,700 10,700
End-Group Analysis 21,448 15,483 15,090 11,272 8,956
Viscornetrv
15,286 12,171 10,032 8,253 3,572
'Determined from five samples which were prepared by varying heating temperature and heating periods of the monomer and by changing the concentration of the catalyst.
Results and Discussion P r e p a r a t i o n of Polylactide-The intrinsic viscosities of polylactide prepared under different conditions are listed in Table I. The highest intrinsic viscosity was noted when dilactide was maintained at 160°C for 6 h in the presence of 0.02%, w/w, of catalyst. Molecular Weight Determination-The molecular weights of polymers employed in the microencapsulation experiments are shown in Table 11. Since there is a direct relationship between the number-average molecular weight and the reciprocal of the carboxyl group content as shown in eq. 3, a, the number of carboxyl end-groups in each polymer molecule, was determined from the slope by plotting these two variables:
-
Mn = a/COOH
(3)
In this present investigation, a n CY value of 1.56 was obtained. Encapsulation-A scanning electron micrograph of chlorpromazine microspheres is shown in Fig. 1. Small pits, observed on many of the particles, increase the release rate of the drug from the microspheres.22Aggregates of small particles were also seen. In Table 111 the yields and drug loss of six different batches of microspheres, prepared in such a way that each batch had an expected loading of 20%, are shown. The reproducibility of microsphere batches prepared under the same conditions was examined with respect to drug loading and disso22 J Journal of Pharmaceutical Sciences Val. 74, No. 1, January 1985
Figure 1-Electron micrograph of chlorpromazine microspheres, 88- 105 pm sieve fraction, containing 28% by weight of chlorpromazine; ~ 8 05kV. ;
lution profiles (Table IV). As shown in Table IV, when the drug to polymer ratio was maintained to give microspheres containing <20% by weight of drug, excellent reproducibility was achieved. In V i t r o Dissolution Studies-The dissolution profile of unencapsulated chlorpromazine indicated that it dissolved rapidly and was stable under the experimental conditions. Figure 2 shows the effect of particle size on the rate of release in hydroalcoholic media. The figure is a plot of the cumulative recovery, calculated as percentage of the original sample on the ordinate, versus time, in hours on the abscissa. It is evident that the greater the microsphere size, the lower the release rate. In all cases, the rate of release was nonlinear and faster initially. The initial rapid release may be due to a greater concentration of drug near the surface of the microspheres and to aggregates of smaller microspheres. The time required for 50% release (t5'%)in hydroalcoholic medium was linearly related to the particle size (r = 0.999) as shown in Fig. 3.
and 32% of the drug revealed that some of the drug was Table Ill-Comparison of Yields of Microspheres Between Batches scattered throughout the polymer as crystals. Figure 7 shows a Loss of drug, % Batch Yield, O h ~~-~ ~
84.0 89.4 91.6 91 .o 87.7 91 .o 89.1 r 2.9
1
2 3 4 5 6 Mean f SD
~
~
18.9 19.5 9.3 14.0 17.2 14.0 15.55 3.8
Table IV-Reproducibility of PercentageDrug Loading and tsO% Oh
Batch
Loading
Expected 40 40 40
1-1 1-2 1-3 Mean k SD
2-1 2-2 2-3
20 20 20
Mean f SD
3-1 3-2 3-3
14 14 14
Mean f SD
YOLoading
Obtained 23.8 26.7 29.6 (26.7+- 2.9) 18.6 19.0 19.4 (19.0f 0.4) 12.6 13.0 13.5 (13.05 0.5)
t50%,'
freeze-etching electron micrograph of the interior of a microsphere in which chlorpromazine is dispersed in the polymeric matrix as crystals. The fractured planes indicate drug crystals. On the other hand, in microspheres containing 14 and 18% of the drug, a polarized light micrograph (Fig. 8) revealed that the drug formed a solid solution with the polymer. In the case of microspheres containing crystals, initial release may be from crystalline material near the surface of the microspheres; as these crystals dissolve they leave channels for rapid diffusion of water into, and drug out of, the interior of the matrix. This dissolving and channeling continues deep into the
2.0 0.9 0.8 (1.25 0.7) 2.5 1 .o 0.8 (1.4f 0.9) 4.5 3.0 2.8 (3.4-1- 0.9)
'
*Values were obtained using a hydroalcoholic fluid as a dissolution medium.
60
100
140
180
LENGTH-NUMBER MEAN DIAMETER, prn
Figure 3-Effect of particle size on tsO%drug release in microspheres containing 3% chlorpromazine; mol. wt., 12,700;r = 0.999,
20 40
60 80 100 120 140 160 180 HOURS
Figure 2-In vitro release of chlorpromazine microspheres containing 3% of drug into 37OC,pH 7.0 hydroalcoholic dissolution medium as a function of particle size; mol. wt., 12,700;Mt is the amount of drug released at time t , and M, is the original weight of the sample. Key: (-) chlorpromazine; (0)>88 pm; (A)105-125 pm; (0) 149-177 pm; (0)
8
.
f 50
r"
10
20
>210 pm.
The influence of molecular weight on the dissolution rate was investigated in aqueous pH 7.0 phosphate buffer (Fig. 4). It was evident from this figure that at least two phases make up the drug-release profile from microspheres. The higher molecular weight microspheres exhibited a lag time of 2 2 4 h at which time the release rate increased rapidly to -540% release. The rate then decreased to about one-sixth of the initial rate and was near zero order. Increasing the molecular weight increased the time required for 50% release, but the difference between microspheres made from 16,300 and 20,500 molecular weight polymers was slight. Figure 5 shows the effect of the percentage of drug loading on the rate of release in an aqueous pH 7.0 phosphate buffer. Figure 6 shows drug release normalized with respect to the surface area of the samples. As shown in Figs. 5 and 6, microspheres containing 23 and 32% of the drug gave higher initial release rates than those containing 14 and 18% of the drug. The daily amount of drug released after the first day was largest in microspheres containing 23 and 32% of drug. This amount could act as a loading dose. Polarized light microscopic observations and electron micrographs of freeze-etched samples of microspheres containing 23
60
100
140
.. 180
HOURS Figure 4-In vitro dissolution profile of chlorpromazine from microspheres containing 18% of drug into 37"C,pH 7.0 aqueous phosphate buffer as a function of polymer molecular weight. Mt is the amount of drug released at time t, and M, is the original weight of the sample. Key: (-) chlorpromazine base; (0)10,700; (0)12,700; (016,300; ) (A)20,500.
"'1 $?
2 50 f 10 HOURS
Figure 5-In vitro dissolution profile of chlorpromazine from microspheres into 37OC,pH 7.0phosphate buffer as a function of drug loading; mol. wt., 12,700;particle size, 88-105 pm. M, is the amount of drug released at time t, and M, is the original weight of the sample. Key: (-) chlorpromazine; (0)32%; (0) 23%; (0) 18%; (A)14%. Journal of Pharmaceutical Sciences J 23 Vol. 74,No. 1, January 1985
20
60
100 HOURS
140
180
Figure 6-In vitro dissolution profile of chlorpromazine from microspheres into 37OC, pH 7.0 aqueous phosphate buffer as a function of drug loading and surface area (S.A.); mol. wt., 12,700; particle size, 88-105 pm. Mt is the amount of drug released at time t, and M, is the original weight of the sample. Key: (0)32%; (0) 23%; (0) 18%; (A)14%.
solid solution is probably caused by the time required for th. dissolution fluid to diffuse into the polymer and cause swelling Once equilibrium is established, the water moving out Carrie: out drug released by the swelling process. The latter (zerc order) part of the curve may represent release of the drug attached to the swollen matrix. The release may be either b, diffusion processes or by breakdown of the matrix or both. The results reported here indicate that a neuroleptic can bc encapsulated in a polylactide polymer and that the release characteristics of these polylactide microspheres can be modified by a t least three methods: ( a ) by employing polymers of different molecular weights, ( b ) by changing the particle size of the microsphere, and ( c ) by controlling the drug loading to obtain different proportions of crystalline drug and drug in solid solution. By employing these methods and in vivo testing it should be possible to optimize the microspheres for various requirements.
References and Notes
Figure 7-freeze-etching electron micrograph of the interior of a microsphere in which chlorpromazine is dispersed in the polymeric matrix as crystals.
1. Retterstol N.; Salvesen, C. Acta. Psychiat. Scand. Suppl. 1973, 246, 18. 2. Ayd, F. J., J r Znt. Drug Ther. Newsletter, 1972, 7, 33. 3. Chien, C.; DiMascio, A. Behao. Neuropsychiatry, 1971,3, 5. 4. Quitkin, F.; Rifkin, A.; Klein, D. F. Psychopharrnacol. Bull., 1977, 13, 27. 5. Hirsch, S. R.; Gaind, R.; Rohde, P. D.; Stevens, B. C.; Wing, J . K. Br. Med. J., 1973, 1, 633. 6. Schooler, N. R.; Levine, J. Ps~chopharrnacol.Bull., 1977, 13, 29. 7. “Martindale, T h e Extra Pharmacopoeia”;27th ed.; Pharmaceutical Press: London, 1977; p 1529. 8. Blake, D. A.; Yolles, S.; Hehrick, M.; Cascorbi, H. F.; Eagan, M. J . “Release of Cyclazocine from Subcutaneously Implanted Polymeric Matrices”; Academy of Pharmaceutical Sciences, San Francisco, CA. 1971. Abstr. 41. 9. Yolles, S.: Sartori, M. P . “Drug Delivery Systems”; Oxford University Press: New York, 1980; p 84. 10. Sinclair, R. G. Enuiron. Sci. Technol. 1973, 7, 955. 11. Thies, C. “Narcotic Antagonists: The Search for Long-Acting Preparations”; National Institute on Drug Abuse: Maryland, 1975; p 19.
Figure 8-Internal section of a microsphere in which chlorpromazine forms a solid solution with the polymer. The clear area around the edge of an internal section is the embedding material, epoxy resin.
matrix of the microsphere. When all crystalline material dissolves, release continues from the remaining drug locked in the matrix. Lag time in the dissolution profile of microspheres containing
24
1 Journal of Pharmaceutical Sciences Vol. 74, No. 1, January 1985
12. Kulkarni, R. K.; Moore, E. G.; Hegyell, A. F.; Leonard, F. J. Biorned. Muter. Res. 1971, 5, 169. 13. Schindler.. A,:. Jeffcoat. R.: Kimmel. G. L.: Pitt. C. G.: Wall. M. E.: Zweidinger, R. “Contemporary Topics’ in ‘Polymer Science”; Plenum Press: New York, 1977; p 251. 14. Palit, S. R.; Mandal, B. M. “Reviews in Macromolecular Chemistry’’; Marcel Dekker: New York, 1968; p 225. 15. Palit, S. R.; Ghosh, P. J . Polyrn. Sci. 1962, 58, 1225. 16. British Patent 490 001, 1938. 17. U.S. Patent 2 183 053, 1939. 18. “U.S. Department of Commerce. National Technical Information Service, Microencapsulation Process”; U.S. Department of Agriculture: Washington, D.C., 1978; pp 1-9. 19. Beck, I,. R.; Flowers, C. F.; Cowsar, D. R.; Tanwuary, A. C. Ger. Offen. 2 940 146. 1980. 20. “ U S . Pharmacopeia”; 20th rev.; U.S. Pharmacopeial Convention: Rockville, MD, 1980; p 142. 21. Salzman, N. P.; Brodie, 13. R. J E’harrnacol Exp. Ther , 1956, 128, AG
1”.
22. Rigshy, W. E.; Suzuki, K.; Humphreys, W. J. Proceedings of the 4 l s t Annual Meeting of the Electron Microscopy Society of America, 1983, Abstr. 620.
was polymerized from dilactide under various conditions to yield polymers in the molecular weight range from 11,000 to 21,000 as determined by osmometry. Chlorpromazine was encapsulated by the polylactide polymers by using an emulsification-solvent evaporation method. Microscopic observation revealed that when drug loading was %18%, the drug was in the form of a solid solution in microspheres of polylactide. At higher drug loadings, crystalline drug was present. In vitro dissolution of the encapsulated drug was followed in hydroalcoholic and aqueous buffer media. Studies with the hydroalcoholic media revealed that the drug release rate decreased as microcapsule size increased. However, dissolution in the hydroalcoholic media did not reflect the effect of molecular weight or percentage loading observed in the aqueous system. Dissolution in aqueous buffer showed that f500,0 increased with molecular weight and that release rates-surface area increased with increasina drua loadina. Long-acting neuroleptics administered by injection are of value in the treatment of chronic psychotic patients-in particular chronic schizophrenics-who have been discharged from a hospital but who rapidly suffer a relapse due to taking their medication incorrectly. Thirty to fifty percent of patients cease medication on their own initiative within 2 months to 1 year following discharge, leading to their readmission to psychiatric hospitals.'-3 The standard treatment of schizophrenia is the use of long-acting injectable esters of fluphenazine, such as fluphenazine decanoate and enanthate. These are injected intramuscularly as a sesame oil suspension once every 2-3 weeks in the dose range of 12.5-25 mg.4-7 The first application of polylactide as an erodible matrix for long-acting controlled release of a drug was reported by Yolles and co-workers in 1971.'Yolles and Sartorigand Sinclair" state that polylactide as a matrix shows, besides biodegradability, the following advantages: ( a ) compatibility with the environment and with many drugs, ( b ) rate of biodegradation compatible with the projected life span of the composite, and (c) ready availability of starting material and cost effectiveness of the polymer itself. The most widely reported uses of polylactide as a matrix for controlled release are in systems containing such compounds as narcotic antagonists, fertility controlling agents, anticancer agents, herbicides, pesticides, and others. However, the system containing neuroleptics has not been reported. In 1975, Thies" first applied the microencapsulation technique to make an injectable sustained-release formulation with polylactide, in which cyclazocine was encapsulated. The present investigation was undertaken to prepare different batches of microcapsules by altering encapsulation factors which influence the rate of drug release, and to determine by in vitro studies whether microcapsules could be made which would be suitable for in vivo testing. The physical factors chosen to be altered were particle size of the microcapsules, percentage of drug loading, and molecular weight of the encapsulating agent. Chlorpromazine was chosen as a model neuroleptic to be encapsulated in a polylactide. 0022-3549/85/0100-002l$O 1 .OO/O 0 1985, American Pharmaceutical Association
Experimental Section Materials-Polylactide was polymerized from both commercially available dilactide (Frinton Laboratories, Vineland, NJ) and dilactide made from lactic acid (J. T. Baker Chemical Co., Phillipsburg, NJ). Tetraphenyltin (Aidrich) was used as a catalyst for polymerization. Chlorpromazine was prepared from chlorpromazine hydrochloride (Sigma). Polyvinyl alcohol type I1 (Sigma) was employed as a suspending agent, sodium lauryl sulfate (Fisher Scientific) as a n ionic surfactant, and sodium sulfite (Fisher Scientific) as an antioxidant. All were USP grade. All other chemicals were analytical reagent grade, unless otherwise indicated. Preparation of Polylactide-The method was based on one reported by Kulkarni et a1.l' The crude dilactide was recrystallized twice using ethyl acetate to obtain purified dilactide with melting points between 124°C and 126°C. The dilactide was placed in a silanized vacuum flask with 0.01-0.02% tetraphenyltin. The contents were exposed to a vacuum pump for 1h to remove air and moisture and then polymerized in an oven at 160-170°C for 3-9 h. The resultingpolymer in the flask was dissolved in acetone using moderate heat and precipitated in a large quantity of water for purification. Intrinsic Viscosity Determinations-A Cannon-Fenske viscometer was used to determine the intrinsic viscosities of polymer-benzene solutions a t 25°C. The average molecular weight was calculated from the intrinsic viscosity by employing the Mark-Houwink equation: 1 = 3.64 X lo4 Mn0.75
(1)
where 1 is the intrinsic viscosity and Mn is the number-average molecular weight. The variables K , 3.64 x lo4, and a, 0.75, were reported by Shindler et al.13 End G r o u p Analysis-The concentrations of the carboxylic acid end-group present in the polymer samples were determined by a dye-interaction techniq~e.'~.''The concentrations of carboxylic acid were calculated using:I3
COOH = 1.77 x 10+ (AE/c~)
(2) where COOH is the concentration of the carboxylic acid group in the polymer in mol/g, A E is the absorbance of sample minus that of blank at 515 nm, and C, is the concentration of polymer solution in g/L. Osmometry-A static-elevation osmometer (constructed by the University of Georgia Instrument Shop) and a regenerated cellulose membrane (grade RC 51, Shleicher & Schuell, Inc., Keene, NH) were employed to determine the molecular weights of the polymer samples at 25°C. The absolute molecular weights determined by osmometry were plotted against the reciprocal of the carboxyl group content, and the number of carboxylic end-groups in each polymer molecule (i.e., one, in the case of polylactide) was estimated from the slope. Journal of PharmaceuticalSciences/ 21 Vol. 74,No. 1, January 1985
Encapsulation Procedure-The basic microencapsulation method was first patented in 193816in Britain and then in the United States in 1939,17and procedures describing the use of polylactic acid as the encapsulating agent were granted patents in 1978'' and 1980.'' These procedures can be classified as an emulsification-solvent evaporation method. Three to six grams of poly(d,l-lactide) were dissolved in 20 mL of methylene chloride, and chlorpromazine was dissolved in the polymer solution. The mixture was dispersed by stirring at 280 rpm in 60 mL of distilled water containing 5% by weight of polyvinyl alcohol and 0.025-0.05% by weight of sodium lauryl sulfate and sodium sulfite. The stirring was carried out under subdued light. After the organic solvent evaporated, the resulting microspheres were separated by filtration. The product was washed with water and hydrochloric acid solution and then dried under reduced pressure at room temperature to yield a white powder. The powder was sized by sieving, and a particle size analysis of each sieve fraction was done microscopically to obtain the mean diameters. Assay P r o c e d u r e f o r Total Chlorpromazine ContentTriplicate samples of -20 mg of the microspheres were accurately weighed. The drug was extracted into ethanol using moderate heat. The spectrophotometric assay method for chlorpromazine was based on those discussed in the U.S.P." and by Salzman and Brodie." I n Vitro Dissolution Studies-Dissolution studies were performed with two fluids: a hydroalcoholic solvent consisting of four parts of 95% ethanol and six parts of phosphate buffer, and an aqueous fluid buffered to pH 7.0. The hydroalcoholic solvent enhanced the extraction rate of chlorpromazine into the dissolution medium. Dissolution studies were carried out using the Levy beaker method discussed in the U.S.P.'' The assembly consisted of the following: a covered, 1000-mL roundbottomed flask; a variable speed motor (G. K. Helker Corp., Bellerose, NY); a stainless steel propeller. While the propeller was rotated at 130 rpm, samples of the dissolution medium were removed periodically and analyzed for chlorpromazine content. The dissolution medium was maintained at constant volume by adding to the flask the volume of dissolution medium which had evaporated during the previous 24 h. Since chlorpromazine is reported to undergo photo-oxidation, all dissolution experiments were done under subdued light.
Table I-Intrinsic Conditions
Conc. of
Tetrapheny,tin,
Viscosities of Polylactide at Various Preparation
Time,
,, Temperature,
YO
0.03 0.02 0.02 0.02 0.01 0.01 None
6 6 6 3 6 3 3
Intrinsic Viscosity,
OC
dL/g
160 170 160 160 160 160 160
0.34 0.28 0.68 0.42 0.58 0.30 0.13'
'The polymer yield was low, 40%.
Table Il-Three Methods Used to Determine the Molecular Weights of Polvmers in the MicroencaDsulation ExDeriments'
Osmornetrv 21,400 20,500 16,300 12,700 10,700
End-Group Analysis 21,448 15,483 15,090 11,272 8,956
Viscornetrv
15,286 12,171 10,032 8,253 3,572
'Determined from five samples which were prepared by varying heating temperature and heating periods of the monomer and by changing the concentration of the catalyst.
Results and Discussion P r e p a r a t i o n of Polylactide-The intrinsic viscosities of polylactide prepared under different conditions are listed in Table I. The highest intrinsic viscosity was noted when dilactide was maintained at 160°C for 6 h in the presence of 0.02%, w/w, of catalyst. Molecular Weight Determination-The molecular weights of polymers employed in the microencapsulation experiments are shown in Table 11. Since there is a direct relationship between the number-average molecular weight and the reciprocal of the carboxyl group content as shown in eq. 3, a, the number of carboxyl end-groups in each polymer molecule, was determined from the slope by plotting these two variables:
-
Mn = a/COOH
(3)
In this present investigation, a n CY value of 1.56 was obtained. Encapsulation-A scanning electron micrograph of chlorpromazine microspheres is shown in Fig. 1. Small pits, observed on many of the particles, increase the release rate of the drug from the microspheres.22Aggregates of small particles were also seen. In Table 111 the yields and drug loss of six different batches of microspheres, prepared in such a way that each batch had an expected loading of 20%, are shown. The reproducibility of microsphere batches prepared under the same conditions was examined with respect to drug loading and disso22 J Journal of Pharmaceutical Sciences Val. 74, No. 1, January 1985
Figure 1-Electron micrograph of chlorpromazine microspheres, 88- 105 pm sieve fraction, containing 28% by weight of chlorpromazine; ~ 8 05kV. ;
lution profiles (Table IV). As shown in Table IV, when the drug to polymer ratio was maintained to give microspheres containing <20% by weight of drug, excellent reproducibility was achieved. In V i t r o Dissolution Studies-The dissolution profile of unencapsulated chlorpromazine indicated that it dissolved rapidly and was stable under the experimental conditions. Figure 2 shows the effect of particle size on the rate of release in hydroalcoholic media. The figure is a plot of the cumulative recovery, calculated as percentage of the original sample on the ordinate, versus time, in hours on the abscissa. It is evident that the greater the microsphere size, the lower the release rate. In all cases, the rate of release was nonlinear and faster initially. The initial rapid release may be due to a greater concentration of drug near the surface of the microspheres and to aggregates of smaller microspheres. The time required for 50% release (t5'%)in hydroalcoholic medium was linearly related to the particle size (r = 0.999) as shown in Fig. 3.
and 32% of the drug revealed that some of the drug was Table Ill-Comparison of Yields of Microspheres Between Batches scattered throughout the polymer as crystals. Figure 7 shows a Loss of drug, % Batch Yield, O h ~~-~ ~
84.0 89.4 91.6 91 .o 87.7 91 .o 89.1 r 2.9
1
2 3 4 5 6 Mean f SD
~
~
18.9 19.5 9.3 14.0 17.2 14.0 15.55 3.8
Table IV-Reproducibility of PercentageDrug Loading and tsO% Oh
Batch
Loading
Expected 40 40 40
1-1 1-2 1-3 Mean k SD
2-1 2-2 2-3
20 20 20
Mean f SD
3-1 3-2 3-3
14 14 14
Mean f SD
YOLoading
Obtained 23.8 26.7 29.6 (26.7+- 2.9) 18.6 19.0 19.4 (19.0f 0.4) 12.6 13.0 13.5 (13.05 0.5)
t50%,'
freeze-etching electron micrograph of the interior of a microsphere in which chlorpromazine is dispersed in the polymeric matrix as crystals. The fractured planes indicate drug crystals. On the other hand, in microspheres containing 14 and 18% of the drug, a polarized light micrograph (Fig. 8) revealed that the drug formed a solid solution with the polymer. In the case of microspheres containing crystals, initial release may be from crystalline material near the surface of the microspheres; as these crystals dissolve they leave channels for rapid diffusion of water into, and drug out of, the interior of the matrix. This dissolving and channeling continues deep into the
2.0 0.9 0.8 (1.25 0.7) 2.5 1 .o 0.8 (1.4f 0.9) 4.5 3.0 2.8 (3.4-1- 0.9)
'
*Values were obtained using a hydroalcoholic fluid as a dissolution medium.
60
100
140
180
LENGTH-NUMBER MEAN DIAMETER, prn
Figure 3-Effect of particle size on tsO%drug release in microspheres containing 3% chlorpromazine; mol. wt., 12,700;r = 0.999,
20 40
60 80 100 120 140 160 180 HOURS
Figure 2-In vitro release of chlorpromazine microspheres containing 3% of drug into 37OC,pH 7.0 hydroalcoholic dissolution medium as a function of particle size; mol. wt., 12,700;Mt is the amount of drug released at time t , and M, is the original weight of the sample. Key: (-) chlorpromazine; (0)>88 pm; (A)105-125 pm; (0) 149-177 pm; (0)
8
.
f 50
r"
10
20
>210 pm.
The influence of molecular weight on the dissolution rate was investigated in aqueous pH 7.0 phosphate buffer (Fig. 4). It was evident from this figure that at least two phases make up the drug-release profile from microspheres. The higher molecular weight microspheres exhibited a lag time of 2 2 4 h at which time the release rate increased rapidly to -540% release. The rate then decreased to about one-sixth of the initial rate and was near zero order. Increasing the molecular weight increased the time required for 50% release, but the difference between microspheres made from 16,300 and 20,500 molecular weight polymers was slight. Figure 5 shows the effect of the percentage of drug loading on the rate of release in an aqueous pH 7.0 phosphate buffer. Figure 6 shows drug release normalized with respect to the surface area of the samples. As shown in Figs. 5 and 6, microspheres containing 23 and 32% of the drug gave higher initial release rates than those containing 14 and 18% of the drug. The daily amount of drug released after the first day was largest in microspheres containing 23 and 32% of drug. This amount could act as a loading dose. Polarized light microscopic observations and electron micrographs of freeze-etched samples of microspheres containing 23
60
100
140
.. 180
HOURS Figure 4-In vitro dissolution profile of chlorpromazine from microspheres containing 18% of drug into 37"C,pH 7.0 aqueous phosphate buffer as a function of polymer molecular weight. Mt is the amount of drug released at time t, and M, is the original weight of the sample. Key: (-) chlorpromazine base; (0)10,700; (0)12,700; (016,300; ) (A)20,500.
"'1 $?
2 50 f 10 HOURS
Figure 5-In vitro dissolution profile of chlorpromazine from microspheres into 37OC,pH 7.0phosphate buffer as a function of drug loading; mol. wt., 12,700;particle size, 88-105 pm. M, is the amount of drug released at time t, and M, is the original weight of the sample. Key: (-) chlorpromazine; (0)32%; (0) 23%; (0) 18%; (A)14%. Journal of Pharmaceutical Sciences J 23 Vol. 74,No. 1, January 1985
20
60
100 HOURS
140
180
Figure 6-In vitro dissolution profile of chlorpromazine from microspheres into 37OC, pH 7.0 aqueous phosphate buffer as a function of drug loading and surface area (S.A.); mol. wt., 12,700; particle size, 88-105 pm. Mt is the amount of drug released at time t, and M, is the original weight of the sample. Key: (0)32%; (0) 23%; (0) 18%; (A)14%.
solid solution is probably caused by the time required for th. dissolution fluid to diffuse into the polymer and cause swelling Once equilibrium is established, the water moving out Carrie: out drug released by the swelling process. The latter (zerc order) part of the curve may represent release of the drug attached to the swollen matrix. The release may be either b, diffusion processes or by breakdown of the matrix or both. The results reported here indicate that a neuroleptic can bc encapsulated in a polylactide polymer and that the release characteristics of these polylactide microspheres can be modified by a t least three methods: ( a ) by employing polymers of different molecular weights, ( b ) by changing the particle size of the microsphere, and ( c ) by controlling the drug loading to obtain different proportions of crystalline drug and drug in solid solution. By employing these methods and in vivo testing it should be possible to optimize the microspheres for various requirements.
References and Notes
Figure 7-freeze-etching electron micrograph of the interior of a microsphere in which chlorpromazine is dispersed in the polymeric matrix as crystals.
1. Retterstol N.; Salvesen, C. Acta. Psychiat. Scand. Suppl. 1973, 246, 18. 2. Ayd, F. J., J r Znt. Drug Ther. Newsletter, 1972, 7, 33. 3. Chien, C.; DiMascio, A. Behao. Neuropsychiatry, 1971,3, 5. 4. Quitkin, F.; Rifkin, A.; Klein, D. F. Psychopharrnacol. Bull., 1977, 13, 27. 5. Hirsch, S. R.; Gaind, R.; Rohde, P. D.; Stevens, B. C.; Wing, J . K. Br. Med. J., 1973, 1, 633. 6. Schooler, N. R.; Levine, J. Ps~chopharrnacol.Bull., 1977, 13, 29. 7. “Martindale, T h e Extra Pharmacopoeia”;27th ed.; Pharmaceutical Press: London, 1977; p 1529. 8. Blake, D. A.; Yolles, S.; Hehrick, M.; Cascorbi, H. F.; Eagan, M. J . “Release of Cyclazocine from Subcutaneously Implanted Polymeric Matrices”; Academy of Pharmaceutical Sciences, San Francisco, CA. 1971. Abstr. 41. 9. Yolles, S.: Sartori, M. P . “Drug Delivery Systems”; Oxford University Press: New York, 1980; p 84. 10. Sinclair, R. G. Enuiron. Sci. Technol. 1973, 7, 955. 11. Thies, C. “Narcotic Antagonists: The Search for Long-Acting Preparations”; National Institute on Drug Abuse: Maryland, 1975; p 19.
Figure 8-Internal section of a microsphere in which chlorpromazine forms a solid solution with the polymer. The clear area around the edge of an internal section is the embedding material, epoxy resin.
matrix of the microsphere. When all crystalline material dissolves, release continues from the remaining drug locked in the matrix. Lag time in the dissolution profile of microspheres containing
24
1 Journal of Pharmaceutical Sciences Vol. 74, No. 1, January 1985
12. Kulkarni, R. K.; Moore, E. G.; Hegyell, A. F.; Leonard, F. J. Biorned. Muter. Res. 1971, 5, 169. 13. Schindler.. A,:. Jeffcoat. R.: Kimmel. G. L.: Pitt. C. G.: Wall. M. E.: Zweidinger, R. “Contemporary Topics’ in ‘Polymer Science”; Plenum Press: New York, 1977; p 251. 14. Palit, S. R.; Mandal, B. M. “Reviews in Macromolecular Chemistry’’; Marcel Dekker: New York, 1968; p 225. 15. Palit, S. R.; Ghosh, P. J . Polyrn. Sci. 1962, 58, 1225. 16. British Patent 490 001, 1938. 17. U.S. Patent 2 183 053, 1939. 18. “U.S. Department of Commerce. National Technical Information Service, Microencapsulation Process”; U.S. Department of Agriculture: Washington, D.C., 1978; pp 1-9. 19. Beck, I,. R.; Flowers, C. F.; Cowsar, D. R.; Tanwuary, A. C. Ger. Offen. 2 940 146. 1980. 20. “ U S . Pharmacopeia”; 20th rev.; U.S. Pharmacopeial Convention: Rockville, MD, 1980; p 142. 21. Salzman, N. P.; Brodie, 13. R. J E’harrnacol Exp. Ther , 1956, 128, AG
1”.
22. Rigshy, W. E.; Suzuki, K.; Humphreys, W. J. Proceedings of the 4 l s t Annual Meeting of the Electron Microscopy Society of America, 1983, Abstr. 620.