- Email: [email protected]
The production of protein-loaded microparticles by supercritical fluid enhanced mixing and spraying
Journal of Controlled Release 101 (2005) 85 – 92 www.elsevier.com/locate/jconrel
The production of protein-loaded microparticles by supercritical fluid enhanced mixing and spraying Martin J. Whitakera, Jianyuan Haob, Owen R. Daviesc, Gulay Serhatkulub, Snow Stolnik-Trenkicc, Steven M. Howdleb, Kevin M. Shakesheff c,* a
Critical Pharmaceuticals, University Park, Nottingham, NG7 2RD, UK School of Chemistry, University of Nottingham, University Park, Nottingham, NG7 2RD, UK c School of Pharmaceutical Sciences, University of Nottingham, University Park, Nottingham, NG7 2RD, UK b
Received 4 April 2004; accepted 12 July 2004 Available online 20 August 2004
Abstract In this study, we use supercritical carbon dioxide as a processing medium for the fabrication of poly(dl-lactic acid) P(DLLA) microparticles that encapsulate a protein material. We have previously demonstrated that this polymer and a dry powder of a protein can be mixed under supercritical carbon dioxide conditions (above 31.1 8C and 73.8 bar) and that the protein component retains its biological activity. In this paper, we progress the work to demonstrate that the plasticized polymer and dry powder protein mixture can be sprayed to form solid polymer particles that encapsulate the protein. Particle size range is between 10 and 300 Am after spraying. Ribonuclease A and lysozyme were encapsulated in the polymer without significant loss of enzymatic activity. Biological assays of insulin and calcitonin confirm retention of activity after fabrication of the microparticles and release of the peptides/proteins. D 2004 Elsevier B.V. All rights reserved. Keywords: Supercritical carbon dioxide; Protein; Polymer; Encapsulation; Activity
1. Introduction Controlled protein delivery from biodegradable polymer microparticles is making a clinical impact in a range of therapeutic areas [1,2]. However, a * Corresponding author. Tel.: +44-115-9515100; fax: +44-1159515102. E-mail address: [email protected] (K.M. Shakesheff ). 0168-3659/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2004.07.017
number of difficulties are still faced when attempting to formulate proteins in a microparticle system. For example, the method of manufacturing of polymer microparticles remains problematic due to the need to expose the polymer phase to heat or solvent to mobilise it. Many therapeutic proteins are denatured by temperature changes or exposure to solvent. Therefore, a number of approaches have been developed to protect protein function throughout a manufacturing route [3,4]. Potentially, supercritical
86
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
fluid mixing could be a useful addition to the current manufacturing options for protein-loaded controlled release devices because it is solvent-free, does not generate solvent/water interfaces and morphology control, via changes in the kinetics of carbon dioxide depressurisation, can generate materials of variable density and surface roughness. We have previously developed a method of mixing proteins and a range of polyesters using supercritical carbon dioxide (scCO2) [5,6]. This method relies on the ability of scCO2 to plasticize amorphous polymer materials, thereby depressing the glass transition temperature. The method can also be employed to produce semi-crystalline polymers [7]. The plasticization effect generates a molten polymer phase that contains a large volume of CO2. This molten phase can then be mixed with a drug and processed into monoliths or particles. We have previously reported the maintenance of protein function after the protein is encapsulated in porous polymer monoliths and released into aqueous medium [5,6] and the preparation of drug-free particles [8]. In this paper, we report on the maintenance of protein function after processing into microparticles.
2. Materials and methods 2.1. Materials PDLLA was synthesized in bulk using stannous octoate (Sn(Oct)2) as the initiator and diethylene glycol (DEG) as the molecular controller. Briefly, dl-lactide/DEG/Sn(Oct)2 were added into a flask previously flamed, purged with nitrogen, and kept at 140 8C for 16 h. The resulting polymer was then dissolved in CHCl3, and recovered by precipitation in an excess of hexane. The purified product was dried under vacuum at room temperature for 48 h. The weight average molecular weight (Mw=8000 Da) and polydispersity (PD=1.30) of the polymer were determined by gel permeation chromatography (GPC). Food grade carbon dioxide (BOC, UK) was supplied to the supercritical fluid system through stainless steel pipe work using a PM101 compressed air-driven pump from New Ways of Analytics, Germany. Ribonuclease A, lysozyme,
recombinant human insulin and salmon calcitonin were all purchased from Sigma and used as received. 2.2. Supercritical fluid enhanced mixing The equipment used to process polymers in supercritical fluids has been described in detail previously [8]. It consists of three main units: a mixing unit, a collecting unit and a cooling unit. The mixing unit contains a mixing vessel wrapped with a heating jacket, a thermocouple, a pressure gauge, a helical stirrer, a CO2 feed pump (PM-101; New Ways of Analytics), a CO2 cylinder and a spraying valve (modified from a commercially available ball valve). The polymer and protein (in dry powder form) were introduced into the mixing vessel and liquefied by an excess of CO2 under controlled conditions for 1 h. The pressure in the autoclave was held at 320 bar and the temperature was held at 35 8C. The helical impeller was rotated at 130 revolutions per minute. Then, the spraying valve was switched on and off repeatedly by a gas pressure motor, allowing the fluid mixture to pass through a cone nozzle and spray into the collecting chamber. The open (0.2 s) and closed (3 s) durations for the spraying valve are programmed through an electronic controller. The collecting unit contains a collecting chamber, a cone nozzle mounted on the top of the chamber, a pressure gauge, nitrogen feeding system, and a backpressure regulator (JASCO BP-1580-81) to maintain the pressure during the spraying process. The collecting chamber was filled with nitrogen to a pressure of 90 bar prior to the spraying. The cooling unit is a container for cooling the bottom of the collecting chamber with liquid nitrogen. Prior to addition of the protein dry powders, the particle size of the protein was reduced by grinding in a pestle and mortar cooled with liquid nitrogen. This dry powder addition method was used for ribonuclease A, lysozyme and insulin. For encapsulation of calcitonin, a modified method of supercritical fluid mixing was employed [5]. The calcitonin was dissolved in 100 Al of phosphatebuffered saline at pH 7.4 and freeze-dried onto the polymer powder (Modulyo Freeze Drier, Edwards, UK) 48 h prior to addition to the high-pressure chamber.
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
87
2.3. Particle size analysis and morphology
2.4. Protein concentration assay
Microparticle size was determined by laser diffraction. A sample of approximately 10 mg of microspheres was dispersed in 0.5 ml of deionised water. The suspension was then added to the sample chamber of a Coulter LS 230 (Beckman Coulter, UK) under moderate stirring to the required concentration (as indicated by the display). Particle size distribution was then determined as a function of the particle diffraction using the Coulter software (version 2.11a), and plotted as a function of volume percentage. For each sample, three sets of readings were performed. Scanning electron microscopy (SEM) was used to determine particle morphology. Microparticles were mounted on aluminium SEM stubs using double-sided carbon tape (Agar Scientific, UK). The samples were sputter coated with gold for 4 min under an argon atmosphere in a Blazers SCD 030 sputter coater unit. Coated samples were examined with a Phillips 505 scanning electron microscope operating at an accelerating voltage of 25 kV. Image analysis was carried out using a Semicaps 2000A (version 8.2) digital imaging system.
To assay concentration and activity of the proteins, the processed microparticles were placed in 2 ml of prewarmed phosphate-buffered saline (pH 7.4) for 15 min at 37 8C. The samples were then vortexed and filtered through a 0.22-Am filter to remove any polymer particles before further analysis. Sample protein content was determined using a colorimetric assay based on bicinchoninic acid (BCA) [9] and was used in a kit form (Sigma, UK). This method measures the protein concentration after 15 min release only. 2.5. Protein function assays The specific activity of ribonuclease was determined by the rate of hydrolysis of cytidine 2V,3Vmonophosphate and adapted from the method of Crook et al. [10]. Briefly, 10 Al aliquots of released sample solution or ribonuclease standard solution were added to 2.5 ml of 0.1 mg/ml cytidine 2V,3Vcyclic monophosphate solution (Sigma) in Tris buffer. Enzyme activity was assayed spectrophotometrically by measuring the rate of conversion of cytidine 2V,3V-
Fig. 1. SEM images of P(DLLA) particles. (A,B) Polymer-only microparticles. (B,C) Polymer and ribonuclease A loaded at 10% w/w.
88
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
Fig. 2. Size distribution of P(DLLA) microparticles (black circles) compared to P(DLLA) loaded with 5% (w/w) ribonuclease (white circles). The mean diameter of P(DLLA) microparticles loaded with ribonuclease=153.7 Am compared to a mean diameter of 157.8 Am for P(DLLA) microparticles without enzyme.
cyclic monophosphate to cytidine-3V-phosphate in a Beckman 640 UV Spectrophotometer at 37F0.1 8C. Samples were mixed for 10 s and readings were started 20 s after the addition of the enzyme and were taken every 2 s (averaged over 1 s) for 60 s thereafter.
The rate of hydrolysis was calculated using the associated software. Lysozyme activity was measured using a standard activity kit (Sigma). Insulin activity was measured following the method of Clark [11]. Calcitonin
Fig. 3. Ribonuclease A activity assay comparing protein activity prior to processing in scCO2 (black circles) and after processing and release from P(DLLA) particles (white circles).
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
89
activity was measured following the method of Chambers and Moore [12].
3. Results and discussion 3.1. Microparticle production encapsulating model enzymes Particles of P(DLLA) and ribonuclease were produced at 35 8C and with a scCO2 pressure of 320 bar. A comparison of particle morphology with and without protein encapsulation is shown in the SEM images shown in Fig. 1. Particle size analysis data are shown in Fig. 2. The particle morphology and size was not influenced by the encapsulation of the ribonuclease. Under the stated production conditions,
Fig. 5. SEM images of P(DLLA) particles loaded with calcitonin 0.00025% w/w.
Fig. 4. SEM images of P(DLLA) particles loaded with lysozyme 5% w/w.
the particles were irregularly shaped with some pores, evidently caused by the rapid loss of CO2 gas bubbles. We have previously reported that particle size and morphology could be controlled, within constraints, by changing the pressure and temperature within the particle formation chamber [8]. Use of an enzyme protein as a model drug in encapsulation studies has the advantage of facilitating the precise quantification of protein function. We have previously reported that the functioning of a range of proteins is unaffected by supercritical fluid mixing during the formation of porous polymer monoliths [6]. The data in Fig. 3 compare the activity of ribonuclease A before processing and after encapsulation and release from P(DLLA) particles. These data demonstrate that the function of the protein is unaffected by the scCO2 conditions. Therefore, the benign conditions present during monolith formation conditions are repeated under particle production conditions. It
90
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
Fig. 6. P(DLLA) particles encapsulating insulin, (A) particle size analysis. (B) protein activity on days 1, 7 and 28 (control powder— black circles; encapsulated insulin—open circles).
should be noted that this experimental method confirms the activity of the proportion of protein released on day 1. To provide a comparison of the potential of different encapsulating processes to decrease ribonuclease A activity, we tested enzymatic activity after formation of P(DLLA) particles using a standard double emulsion method. Protein released from the double emulsion method had lost 33% of its enzymatic activity. This approach of testing enzyme activity before and after processing was repeated using the enzyme lysozyme. Fig. 4 shows SEM images of particle morphology of the encapsulated protein in P(DLLA). The particle morphology and size distribution (data not shown) were similar to data for ribonuclease A encapsulation. Again, rough particle morphologies were generated. Enzyme activity was recorded at 92.4F5.0% compared to unprocessed controls.
Fig. 7. Activity assay for calcitonin. (A) Cell observed in no calcitonin control. Without calcitonin present, the osteoclasts are motile and possess characteristic membrane skirts (shown by black arrows). (B) Cell observed in calcitonin control. Osteoclasts adopt a stellate morphology and are immotile. (C) Cell observed in release media from encapsulated calcitonin.
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
3.2. Microparticle production encapsulating insulin or calcitonin Building on the foundation of the model protein work, we prepared microparticles containing either insulin or calcitonin. For both protein formulations, we maintained the same processing conditions and generated microparticles with rough morphologies and size distributions in the tens to few hundreds of microns. Fig. 5 shows examples of calcitonin encapsulated particles. Due to the low dosage of calcitonin, a modified manufacturing route was performed in which a dilute aqueous solution of the protein is mixed with polymer powder and freeze-dried. This method has been reported by our team for monolith production [5]. The morphology and size range of calcitonin-containing particles are different from the particles for the other three proteins studied in this paper. At present, the cause of this particle character change is unknown. The insulin-encapsulating particles were tested for protein activity using an ELISA assay reported in the literature [11]. The assay was performed immediately after microparticle fabrication and then after microparticle storage at 25 8C for 1 week and 1 month (Fig. 6). Activity of insulin, as measured by this ELISA assay was identical to the control on day 1. On storage of both the microparticles and control insulin powder at 25 8C, there was a decrease in activity for both batches. However, the fall in activity was less pronounced for insulin encapsulated in P(DLLA) than for the control insulin powder. The calcitonin assay is based on the reduction of motility of rat osteoclasts. The concentration of calcitonin released on day 1 was too low to assay by a method independent of protein activity. However, the assay was able to confirm that the calcitonin released on day 1 generated 100% inhibition of osteoclasts. This relates to a released concentration of fully active calcitonin of above 12 pg/ml (Fig. 7).
4. Conclusions By removing the requirement to process P(DLLA) in organic solvent or at high temperatures, the use of supercritical fluid enhanced mixing has the potential to improve the manufacture of protein-encapsulated
91
products. This work has demonstrated that the extra step of microparticle fabrication does not damage protein function when compared to the extensively researched fabrication of protein loaded porous polymer monoliths. The technique requires further improvement and characterisation to prove its utility in controlled protein release. The control of particle size and shape is poor compared to, for example, spray drying or double emulsion methods. In addition, controlled release kinetics may prove difficult to achieve and predictably modify due to the high surface area of the rough particles and the dry powder presentation of the starting protein material. However, the absence of solvent and low temperature environment could prove valuable for certain protein drugs that resist effects at encapsulation in active forms by conventional manufacturing routes.
Acknowledgments The authors acknowledge the EPSRC for funding, Dr. Andrew Lewis for insulin assay performance, Dr. Peter Wilson and Prof. Jim Gallagher for the calcitonin assay performance.
References [1] G. Orive, R.M. Hernandez, A.R. Gascon, A. Dominguez-Gil, J.L. Pedraz, Drug delivery in biotechnology: present and future, Curr. Opin. Biotechnol. 14 (6) (2003) 659 – 664. [2] P. Periti, T. Mazzei, E. Mini, Clinical pharmacokinetics of depot leuprorelin, Clin. Pharmacokinet. 41 (7) (2002) 485 – 504. [3] D.M. Cook, B.M.K. Biller, M.L. Vance, A.R. Hoffman, L.S. Phillips, K.M. Ford, D.P. Benziger, A. Illeperuma, S.L. Blethen, K.M. Attie, L.N. Dao, J.D. Reimann, P.J. Fielder, The pharmacokinetic and pharmacodynamic characteristics of a long-acting growth hormone (GH) preparation (Nutropin Depot) in GH-deficient adults, J. Clin. Endocrinol. Metab. 87 (10) (2002) 4508 – 4514. [4] W.S. Choi, G.G.K. Murthy, D.A. Edwards, R. Langer, K.M. Klibanov, Inhalation delivery of proteins from ethanol suspensions, Proc. Natl. Acad. Sci. U. S. A. 98 (20) (2001) 11103 – 11107. [5] M.S. Watson, M.J. Whitaker, S.M. Howdle, K.M. Shakesheff, Incorporation of proteins into polymer materials by a novel supercritical fluid processing method, Adv. Mater. 14 (24) (2002) 1802 – 1804.
92
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
[6] S.M. Howdle, M.S. Watson, M.J. Whitaker, V.K. Popov, M.C. Davies, F.S. Mandel, J. Don Wang, K.M. Shakesheff, Supercritical fluid mixing: preparation of thermally sensitive polymer composites containing bioactive materials, Chem. Commun. 01 (2001) 109 – 110. [7] S.L. Shenoy, T. Fujiwara, K.J. Wynne, Quantifying plasticization and melting behavior of poly(vinylidene fluoride) in supercritical CO2 utilizing a linear variable differential transformer, Macromolecules 36 (9) (2003) 3380 – 3385. [8] J. Hao, M.J. Whitaker, B. Wong, G. Serhatkulu, K.M. Shakesheff, S.M. Howdle, Plasticization and spraying of poly (dl-lactic acid) using supercritical carbon dioxide: control of particle size, J. Pharm. Sci. 93 (2004) 1083 – 1094.
[9] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Measurement of protein using bicinchoninic acid, Anal. Biochem. 150 (1985) 76 – 85. [10] E.M. Crook, A.P. Mathias, B.R. Rabin, Spectrophotometric assay of bovine pancreatic ribonuclease by the use of cytidine 2V,3V-phosphate, Biochem. J. 74 (1960) 234 – 238. [11] P.M. Clark, Assays for insulin, pro-insulin(s) and C peptide, Ann. Clin. Biochem. 36 (1999) 541 – 564. [12] T.J. Chambers, A. Moore, The sensitivity of isolated osteoclasts to morphological transformation by calcitonin, J. Clin. Endocrinol. Metab. 57 (4) (1983) 819 – 824.
The production of protein-loaded microparticles by supercritical fluid enhanced mixing and spraying Martin J. Whitakera, Jianyuan Haob, Owen R. Daviesc, Gulay Serhatkulub, Snow Stolnik-Trenkicc, Steven M. Howdleb, Kevin M. Shakesheff c,* a
Critical Pharmaceuticals, University Park, Nottingham, NG7 2RD, UK School of Chemistry, University of Nottingham, University Park, Nottingham, NG7 2RD, UK c School of Pharmaceutical Sciences, University of Nottingham, University Park, Nottingham, NG7 2RD, UK b
Received 4 April 2004; accepted 12 July 2004 Available online 20 August 2004
Abstract In this study, we use supercritical carbon dioxide as a processing medium for the fabrication of poly(dl-lactic acid) P(DLLA) microparticles that encapsulate a protein material. We have previously demonstrated that this polymer and a dry powder of a protein can be mixed under supercritical carbon dioxide conditions (above 31.1 8C and 73.8 bar) and that the protein component retains its biological activity. In this paper, we progress the work to demonstrate that the plasticized polymer and dry powder protein mixture can be sprayed to form solid polymer particles that encapsulate the protein. Particle size range is between 10 and 300 Am after spraying. Ribonuclease A and lysozyme were encapsulated in the polymer without significant loss of enzymatic activity. Biological assays of insulin and calcitonin confirm retention of activity after fabrication of the microparticles and release of the peptides/proteins. D 2004 Elsevier B.V. All rights reserved. Keywords: Supercritical carbon dioxide; Protein; Polymer; Encapsulation; Activity
1. Introduction Controlled protein delivery from biodegradable polymer microparticles is making a clinical impact in a range of therapeutic areas [1,2]. However, a * Corresponding author. Tel.: +44-115-9515100; fax: +44-1159515102. E-mail address: [email protected] (K.M. Shakesheff ). 0168-3659/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2004.07.017
number of difficulties are still faced when attempting to formulate proteins in a microparticle system. For example, the method of manufacturing of polymer microparticles remains problematic due to the need to expose the polymer phase to heat or solvent to mobilise it. Many therapeutic proteins are denatured by temperature changes or exposure to solvent. Therefore, a number of approaches have been developed to protect protein function throughout a manufacturing route [3,4]. Potentially, supercritical
86
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
fluid mixing could be a useful addition to the current manufacturing options for protein-loaded controlled release devices because it is solvent-free, does not generate solvent/water interfaces and morphology control, via changes in the kinetics of carbon dioxide depressurisation, can generate materials of variable density and surface roughness. We have previously developed a method of mixing proteins and a range of polyesters using supercritical carbon dioxide (scCO2) [5,6]. This method relies on the ability of scCO2 to plasticize amorphous polymer materials, thereby depressing the glass transition temperature. The method can also be employed to produce semi-crystalline polymers [7]. The plasticization effect generates a molten polymer phase that contains a large volume of CO2. This molten phase can then be mixed with a drug and processed into monoliths or particles. We have previously reported the maintenance of protein function after the protein is encapsulated in porous polymer monoliths and released into aqueous medium [5,6] and the preparation of drug-free particles [8]. In this paper, we report on the maintenance of protein function after processing into microparticles.
2. Materials and methods 2.1. Materials PDLLA was synthesized in bulk using stannous octoate (Sn(Oct)2) as the initiator and diethylene glycol (DEG) as the molecular controller. Briefly, dl-lactide/DEG/Sn(Oct)2 were added into a flask previously flamed, purged with nitrogen, and kept at 140 8C for 16 h. The resulting polymer was then dissolved in CHCl3, and recovered by precipitation in an excess of hexane. The purified product was dried under vacuum at room temperature for 48 h. The weight average molecular weight (Mw=8000 Da) and polydispersity (PD=1.30) of the polymer were determined by gel permeation chromatography (GPC). Food grade carbon dioxide (BOC, UK) was supplied to the supercritical fluid system through stainless steel pipe work using a PM101 compressed air-driven pump from New Ways of Analytics, Germany. Ribonuclease A, lysozyme,
recombinant human insulin and salmon calcitonin were all purchased from Sigma and used as received. 2.2. Supercritical fluid enhanced mixing The equipment used to process polymers in supercritical fluids has been described in detail previously [8]. It consists of three main units: a mixing unit, a collecting unit and a cooling unit. The mixing unit contains a mixing vessel wrapped with a heating jacket, a thermocouple, a pressure gauge, a helical stirrer, a CO2 feed pump (PM-101; New Ways of Analytics), a CO2 cylinder and a spraying valve (modified from a commercially available ball valve). The polymer and protein (in dry powder form) were introduced into the mixing vessel and liquefied by an excess of CO2 under controlled conditions for 1 h. The pressure in the autoclave was held at 320 bar and the temperature was held at 35 8C. The helical impeller was rotated at 130 revolutions per minute. Then, the spraying valve was switched on and off repeatedly by a gas pressure motor, allowing the fluid mixture to pass through a cone nozzle and spray into the collecting chamber. The open (0.2 s) and closed (3 s) durations for the spraying valve are programmed through an electronic controller. The collecting unit contains a collecting chamber, a cone nozzle mounted on the top of the chamber, a pressure gauge, nitrogen feeding system, and a backpressure regulator (JASCO BP-1580-81) to maintain the pressure during the spraying process. The collecting chamber was filled with nitrogen to a pressure of 90 bar prior to the spraying. The cooling unit is a container for cooling the bottom of the collecting chamber with liquid nitrogen. Prior to addition of the protein dry powders, the particle size of the protein was reduced by grinding in a pestle and mortar cooled with liquid nitrogen. This dry powder addition method was used for ribonuclease A, lysozyme and insulin. For encapsulation of calcitonin, a modified method of supercritical fluid mixing was employed [5]. The calcitonin was dissolved in 100 Al of phosphatebuffered saline at pH 7.4 and freeze-dried onto the polymer powder (Modulyo Freeze Drier, Edwards, UK) 48 h prior to addition to the high-pressure chamber.
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
87
2.3. Particle size analysis and morphology
2.4. Protein concentration assay
Microparticle size was determined by laser diffraction. A sample of approximately 10 mg of microspheres was dispersed in 0.5 ml of deionised water. The suspension was then added to the sample chamber of a Coulter LS 230 (Beckman Coulter, UK) under moderate stirring to the required concentration (as indicated by the display). Particle size distribution was then determined as a function of the particle diffraction using the Coulter software (version 2.11a), and plotted as a function of volume percentage. For each sample, three sets of readings were performed. Scanning electron microscopy (SEM) was used to determine particle morphology. Microparticles were mounted on aluminium SEM stubs using double-sided carbon tape (Agar Scientific, UK). The samples were sputter coated with gold for 4 min under an argon atmosphere in a Blazers SCD 030 sputter coater unit. Coated samples were examined with a Phillips 505 scanning electron microscope operating at an accelerating voltage of 25 kV. Image analysis was carried out using a Semicaps 2000A (version 8.2) digital imaging system.
To assay concentration and activity of the proteins, the processed microparticles were placed in 2 ml of prewarmed phosphate-buffered saline (pH 7.4) for 15 min at 37 8C. The samples were then vortexed and filtered through a 0.22-Am filter to remove any polymer particles before further analysis. Sample protein content was determined using a colorimetric assay based on bicinchoninic acid (BCA) [9] and was used in a kit form (Sigma, UK). This method measures the protein concentration after 15 min release only. 2.5. Protein function assays The specific activity of ribonuclease was determined by the rate of hydrolysis of cytidine 2V,3Vmonophosphate and adapted from the method of Crook et al. [10]. Briefly, 10 Al aliquots of released sample solution or ribonuclease standard solution were added to 2.5 ml of 0.1 mg/ml cytidine 2V,3Vcyclic monophosphate solution (Sigma) in Tris buffer. Enzyme activity was assayed spectrophotometrically by measuring the rate of conversion of cytidine 2V,3V-
Fig. 1. SEM images of P(DLLA) particles. (A,B) Polymer-only microparticles. (B,C) Polymer and ribonuclease A loaded at 10% w/w.
88
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
Fig. 2. Size distribution of P(DLLA) microparticles (black circles) compared to P(DLLA) loaded with 5% (w/w) ribonuclease (white circles). The mean diameter of P(DLLA) microparticles loaded with ribonuclease=153.7 Am compared to a mean diameter of 157.8 Am for P(DLLA) microparticles without enzyme.
cyclic monophosphate to cytidine-3V-phosphate in a Beckman 640 UV Spectrophotometer at 37F0.1 8C. Samples were mixed for 10 s and readings were started 20 s after the addition of the enzyme and were taken every 2 s (averaged over 1 s) for 60 s thereafter.
The rate of hydrolysis was calculated using the associated software. Lysozyme activity was measured using a standard activity kit (Sigma). Insulin activity was measured following the method of Clark [11]. Calcitonin
Fig. 3. Ribonuclease A activity assay comparing protein activity prior to processing in scCO2 (black circles) and after processing and release from P(DLLA) particles (white circles).
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
89
activity was measured following the method of Chambers and Moore [12].
3. Results and discussion 3.1. Microparticle production encapsulating model enzymes Particles of P(DLLA) and ribonuclease were produced at 35 8C and with a scCO2 pressure of 320 bar. A comparison of particle morphology with and without protein encapsulation is shown in the SEM images shown in Fig. 1. Particle size analysis data are shown in Fig. 2. The particle morphology and size was not influenced by the encapsulation of the ribonuclease. Under the stated production conditions,
Fig. 5. SEM images of P(DLLA) particles loaded with calcitonin 0.00025% w/w.
Fig. 4. SEM images of P(DLLA) particles loaded with lysozyme 5% w/w.
the particles were irregularly shaped with some pores, evidently caused by the rapid loss of CO2 gas bubbles. We have previously reported that particle size and morphology could be controlled, within constraints, by changing the pressure and temperature within the particle formation chamber [8]. Use of an enzyme protein as a model drug in encapsulation studies has the advantage of facilitating the precise quantification of protein function. We have previously reported that the functioning of a range of proteins is unaffected by supercritical fluid mixing during the formation of porous polymer monoliths [6]. The data in Fig. 3 compare the activity of ribonuclease A before processing and after encapsulation and release from P(DLLA) particles. These data demonstrate that the function of the protein is unaffected by the scCO2 conditions. Therefore, the benign conditions present during monolith formation conditions are repeated under particle production conditions. It
90
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
Fig. 6. P(DLLA) particles encapsulating insulin, (A) particle size analysis. (B) protein activity on days 1, 7 and 28 (control powder— black circles; encapsulated insulin—open circles).
should be noted that this experimental method confirms the activity of the proportion of protein released on day 1. To provide a comparison of the potential of different encapsulating processes to decrease ribonuclease A activity, we tested enzymatic activity after formation of P(DLLA) particles using a standard double emulsion method. Protein released from the double emulsion method had lost 33% of its enzymatic activity. This approach of testing enzyme activity before and after processing was repeated using the enzyme lysozyme. Fig. 4 shows SEM images of particle morphology of the encapsulated protein in P(DLLA). The particle morphology and size distribution (data not shown) were similar to data for ribonuclease A encapsulation. Again, rough particle morphologies were generated. Enzyme activity was recorded at 92.4F5.0% compared to unprocessed controls.
Fig. 7. Activity assay for calcitonin. (A) Cell observed in no calcitonin control. Without calcitonin present, the osteoclasts are motile and possess characteristic membrane skirts (shown by black arrows). (B) Cell observed in calcitonin control. Osteoclasts adopt a stellate morphology and are immotile. (C) Cell observed in release media from encapsulated calcitonin.
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
3.2. Microparticle production encapsulating insulin or calcitonin Building on the foundation of the model protein work, we prepared microparticles containing either insulin or calcitonin. For both protein formulations, we maintained the same processing conditions and generated microparticles with rough morphologies and size distributions in the tens to few hundreds of microns. Fig. 5 shows examples of calcitonin encapsulated particles. Due to the low dosage of calcitonin, a modified manufacturing route was performed in which a dilute aqueous solution of the protein is mixed with polymer powder and freeze-dried. This method has been reported by our team for monolith production [5]. The morphology and size range of calcitonin-containing particles are different from the particles for the other three proteins studied in this paper. At present, the cause of this particle character change is unknown. The insulin-encapsulating particles were tested for protein activity using an ELISA assay reported in the literature [11]. The assay was performed immediately after microparticle fabrication and then after microparticle storage at 25 8C for 1 week and 1 month (Fig. 6). Activity of insulin, as measured by this ELISA assay was identical to the control on day 1. On storage of both the microparticles and control insulin powder at 25 8C, there was a decrease in activity for both batches. However, the fall in activity was less pronounced for insulin encapsulated in P(DLLA) than for the control insulin powder. The calcitonin assay is based on the reduction of motility of rat osteoclasts. The concentration of calcitonin released on day 1 was too low to assay by a method independent of protein activity. However, the assay was able to confirm that the calcitonin released on day 1 generated 100% inhibition of osteoclasts. This relates to a released concentration of fully active calcitonin of above 12 pg/ml (Fig. 7).
4. Conclusions By removing the requirement to process P(DLLA) in organic solvent or at high temperatures, the use of supercritical fluid enhanced mixing has the potential to improve the manufacture of protein-encapsulated
91
products. This work has demonstrated that the extra step of microparticle fabrication does not damage protein function when compared to the extensively researched fabrication of protein loaded porous polymer monoliths. The technique requires further improvement and characterisation to prove its utility in controlled protein release. The control of particle size and shape is poor compared to, for example, spray drying or double emulsion methods. In addition, controlled release kinetics may prove difficult to achieve and predictably modify due to the high surface area of the rough particles and the dry powder presentation of the starting protein material. However, the absence of solvent and low temperature environment could prove valuable for certain protein drugs that resist effects at encapsulation in active forms by conventional manufacturing routes.
Acknowledgments The authors acknowledge the EPSRC for funding, Dr. Andrew Lewis for insulin assay performance, Dr. Peter Wilson and Prof. Jim Gallagher for the calcitonin assay performance.
References [1] G. Orive, R.M. Hernandez, A.R. Gascon, A. Dominguez-Gil, J.L. Pedraz, Drug delivery in biotechnology: present and future, Curr. Opin. Biotechnol. 14 (6) (2003) 659 – 664. [2] P. Periti, T. Mazzei, E. Mini, Clinical pharmacokinetics of depot leuprorelin, Clin. Pharmacokinet. 41 (7) (2002) 485 – 504. [3] D.M. Cook, B.M.K. Biller, M.L. Vance, A.R. Hoffman, L.S. Phillips, K.M. Ford, D.P. Benziger, A. Illeperuma, S.L. Blethen, K.M. Attie, L.N. Dao, J.D. Reimann, P.J. Fielder, The pharmacokinetic and pharmacodynamic characteristics of a long-acting growth hormone (GH) preparation (Nutropin Depot) in GH-deficient adults, J. Clin. Endocrinol. Metab. 87 (10) (2002) 4508 – 4514. [4] W.S. Choi, G.G.K. Murthy, D.A. Edwards, R. Langer, K.M. Klibanov, Inhalation delivery of proteins from ethanol suspensions, Proc. Natl. Acad. Sci. U. S. A. 98 (20) (2001) 11103 – 11107. [5] M.S. Watson, M.J. Whitaker, S.M. Howdle, K.M. Shakesheff, Incorporation of proteins into polymer materials by a novel supercritical fluid processing method, Adv. Mater. 14 (24) (2002) 1802 – 1804.
92
M.J. Whitaker et al. / Journal of Controlled Release 101 (2005) 85–92
[6] S.M. Howdle, M.S. Watson, M.J. Whitaker, V.K. Popov, M.C. Davies, F.S. Mandel, J. Don Wang, K.M. Shakesheff, Supercritical fluid mixing: preparation of thermally sensitive polymer composites containing bioactive materials, Chem. Commun. 01 (2001) 109 – 110. [7] S.L. Shenoy, T. Fujiwara, K.J. Wynne, Quantifying plasticization and melting behavior of poly(vinylidene fluoride) in supercritical CO2 utilizing a linear variable differential transformer, Macromolecules 36 (9) (2003) 3380 – 3385. [8] J. Hao, M.J. Whitaker, B. Wong, G. Serhatkulu, K.M. Shakesheff, S.M. Howdle, Plasticization and spraying of poly (dl-lactic acid) using supercritical carbon dioxide: control of particle size, J. Pharm. Sci. 93 (2004) 1083 – 1094.
[9] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Measurement of protein using bicinchoninic acid, Anal. Biochem. 150 (1985) 76 – 85. [10] E.M. Crook, A.P. Mathias, B.R. Rabin, Spectrophotometric assay of bovine pancreatic ribonuclease by the use of cytidine 2V,3V-phosphate, Biochem. J. 74 (1960) 234 – 238. [11] P.M. Clark, Assays for insulin, pro-insulin(s) and C peptide, Ann. Clin. Biochem. 36 (1999) 541 – 564. [12] T.J. Chambers, A. Moore, The sensitivity of isolated osteoclasts to morphological transformation by calcitonin, J. Clin. Endocrinol. Metab. 57 (4) (1983) 819 – 824.