Drug delivery goes supercritical
by Patrick J. Ginty1,2, Martin J. Whitaker1,2,3, Kevin M. Shakesheff1, and Steven M. Howdle2,*
In the field of drug delivery, the ability to control the size, morphology, and release of drug particles is fundamental to good targeting, but is often hampered by harsh processing conditions or inadequate methods; likewise for the processing of polymeric controlled-release systems. However, the use of supercritical fluids such as supercritical CO2 (scCO2) has provided a ‘clean’ and effective alternative to traditional methods of drug and polymer processing. In particular, scCO2 has a number of unique properties that make it possible to process both bioactive molecules and amorphous polymers without using toxic organic solvents or elevated temperatures. Here, we review the positive impact that supercritical fluids have had on the micronization, encapsulation, and impregnation of molecules of interest to both the pharmaceutical and biotechnology industries.
1School
of Pharmacy, and of Chemistry, University of Nottingham, University Park, Nottingham, NG7 2RD, UK *E-mail:
[email protected] 2School
3Critical Pharmaceuticals Ltd, BioCity Nottingham, Pennyfoot Street, Nottingham, NG1 1GF, UK
42
August 2005
It is known that the efficacy of a drug can be increased if it is delivered to its target selectively and its release profile is controlled1. Most drug delivery strategies focus on the production of drug particles, with approximately two thirds of the products used in the pharmaceutical industry coming in the form of particulate solids2. Furthermore, it has been shown that the incorporation or encapsulation of a drug within a biodegradable polymer can greatly enhance the controlled release of the particles3. The polymer can act to control release through a variety of mechanisms, such as the rate of biodegradation and pore size4. Control over drug particle size is essential for good targeting and efficacy, as a smaller particle size leads to higher rates of dissolution. This increases the bioavailability of the drug, leading to smaller dosages and enhancing controlled release5. Traditional drug micronization techniques, such as milling, grinding, and spray-drying do not provide this level of control and often use toxic solvents, high temperatures, or mechanical stresses that cause degradation of the drug. Furthermore, residual levels of toxic solvents can remain in the product after processing, making them unsuitable for pharmaceutical applications6. Equally, methods traditionally used for the encapsulation of drugs within polymer particles, such as phase separation, spray drying, and double emulsion techniques, are beset with the same problems7. The use of supercritical fluids (SCFs) has provided a more controlled and tunable route to producing a variety of drug
ISSN:1369 7021 © Elsevier Ltd 2005
RESEARCH REPORT
Supercritical region Pressure (bar)
73.8 Liquid
Solid
Pc
5 Gas 216 K
304.1 K (31.1°C)
Temperature Fig. 1 A phase diagram for carbon dioxide, showing the critical point (Pc) and the supercritical region.
delivery systems, while circumventing many of the problems associated with traditional techniques8. Supercritical fluids are substances, the pressure and temperature of which are raised above their critical values. Once above this critical point, they exist as a single phase with unique properties such as liquid-like densities and, hence, solvating characteristics that are similar to those of liquids, yet with mass transfer properties similar to gases. In particular, scCO2 is the most commonly used SCF in drug delivery applications for a number of reasons9. Carbon dioxide is relatively inert and environmentally friendly when compared to organic solvents and has a low critical point of 73.8 bar and 31.1°C, permitting processing in ambient conditions (Fig. 1). A number of SCF techniques have been used for the micronization of drug particles. Many pharmaceuticals can be dissolved or liquefied in scCO2 before being sprayed through a nozzle upon depressurization to produce fine drug particles.
(A)
High pressure vessel
This can be achieved with nonsolvent techniques such as the rapid expansion of supercritical solutions (RESS) and particles from gas saturated solutions (PGSS) (Fig. 2). In addition, a variety of amorphous polymers is either soluble or can become plasticized in scCO2, allowing them to become viscous liquids without the need for organic solvents and elevated temperatures. Once liquefied, the polymer can be either cast into a porous controlled-release matrix for later impregnation, coprecipitated with drugs to form composite-controlled release matrices, or used to produce drug-carrying microparticles. Alternatively, scCO2 can be used as an anti-solvent for the precipitation of drugs already dissolved in organic solvents8. This review will summarize the use of supercritical fluids, and particularly supercritical CO2, to improve the field of drug delivery, and the impact that has been made.
Drug micronization and particle formation with scCO2 Rapid expansion of supercritical solutions The rapid expansion of supercritical solutions (RESS) is the most simplistic SCF technique used to produce fine drug particles. Here, the solute (the drug) is dissolved within the supercritical gas without the need for additional organic solvents10. The mixture is then depressurized through a nozzle resulting in the rapid precipitation of the drug particles into a collection chamber11,12. Two steroid drugs, progesterone and medroxyprogesterone, have successfully been processed using solvent-free RESS to produce particles with sizes that compare
(B) Drug and/or polymer dissolved in SCF
Solute melted in SCF High pressure vessel
Precipitation of particles through nozzle
Microparticles in collection chamber
Precipitation of particles through nozzle
Microparticles in collection chamber
Fig. 2 Schematic of techniques used to produce particles: (A) rapid expansion from supercritical solutions (RESS); and (B) particles from gas saturated solutions (PGSS). These occur by either (A) dissolution of the solute within the SCF or (B) liquefaction of the solute by the SCF. Subsequent release of the pressure through a nozzle results in the expansion of the particles into a collection chamber.
August 2005
43
RESEARCH REPORT
favorably with traditional methods such as jet milling13. Drugs with poor water solubility, such as ibuprofen, have also successfully been micronized using RESS, yielding greater dissolution rates14,15. However, the major drawback with the RESS technique is the requirement for the solute to be completely soluble in the gas, as CO2 is generally considered to be a poor solvent9. This limits the amount of drugs and polymers that can be processed in this way and, for those that can, very high pressures are required in order to do so. For those molecules that are more CO2-phobic, solvents such as acetone16 and various surfactants17 have been used alongside scCO2 to produce drug particles on the micronscale and nanoscale by rapid expansion into aqueous solutions. This technique has also been used in the coprecipitation of drugs with polymers to form composite microparticles capable of controlled release. The formation of drug/polymer particles on the micron scale acts to control both the release of the drug and the ease of administration. For example, polymers have been used to control drug release by diffusion and/or degradation4, but have still been made sufficiently small to allow systemic injection18. A variety of polymers has been used in the production of microparticles, such as poly(lactic acid) (PLA), poly(lactide-co-glycolide) (PLGA) and poly(ethylene glycol) (PEG). These polymers appear most frequently in scCO2-based applications, as they have strong interactions with CO2, are relatively nontoxic and have full FDA approval5. One of the first uses of RESS for this application involved loading lovastatin, an anti-cholesterol drug, into the amorphous form of PLA, poly(DL-lactic acid) (PDLLA)19. Both the drug and the polymer were dissolved in
Fig. 3 A typical SEM image for the formation of spherical particles comprising a large hollow area. Note that the hollow structure is composed of a single internal void and a thin-layer outer crust. (Reproduced from23 by permission of The Royal Society of Chemistry.)
44
August 2005
supercritical CO2 and coprecipitated to form composite microparticles in the 10-100 µm size range. Poly(L-lactic acid) (PLLA) has also been used to encapsulate pharmaceuticals20 and drugs such as flavone have also successfully been embedded in PEG21 using this technique. Particles from gas saturated solutions (PGSS) This process works on the basis that pressurized gas can diffuse into a solute, lowering both its melting point and its viscosity. By continuing to increase the pressure of the gas, the molten solute forms a highly saturated solution3. Particles can then be formed by depressurizing the solution through a nozzle, as the gas is released from the condensed phase. The particle size distribution can be controlled by changing the processing parameters such as temperature, pressure, etc., although nozzle diameter is the single most important factor. Advantages such as these have seen this technique used for the micronization of pharmaceuticals such as nifedipine, an anti-hypertensive drug22. This drug can be liquefied in scCO2, allowing it to be processed into particles when depressurized through a nozzle. These particles were in the 15-30 µm size range with a dissolution rate five times more rapid than nonprocessed particles. A similar study, where the same drug was processed with the crystalline polymer PEG MW4000, showed that the levels of dissolution were yet quicker compared with a nontreated mixture of the same materials3. An anti-angina drug, felodipine, has also been coprecipitated with PEG using scCO2, with dissolution rates that better those of the previous study22. In fact, despite its crystalline structure, PEG has been shown to compare favorably with PDLLA when micronized using the PGSS technique, in terms of the uniform spherical particle morphology and ease of processing23 (Fig. 3). The PGSS technique has several benefits when compared with RESS. Because the PGSS technique does not require the polymer/drug to be soluble in the scCO2, it uses lower pressures than RESS resulting in lower gas consumption. It is also considered to be, a more versatile technique, having been used to process a variety of materials. For example, theophylline (an anti-asthma drug) was encapsulated within a matrix of hydrogenated palm oil (HPO) to form a composite drug carrier system24. This versatility has been further demonstrated by using N2 to control the particle size of PDLLA by creating a backpressure within the collecting chamber25. The backpressure maintains the low viscosity of the CO2/polymer (PDLLA) mixture for sufficient time to allow
RESEARCH REPORT
Fig. 4 SEM images of PDLLA particles produced by a PGSS/N2 backpressure technique to control particle size and incorporate labile proteins: (A), (B) polymer-only microparticles; (C), (D) polymer and ribonuclease A loaded at 10% w/w. (Reprinted with permission from26. © 2005 Elsevier.)
droplet formation, before the particles then solidify by a combination of CO2 escape and N2 displacement. This solvent-free technology has been used to co-precipitate thermally labile proteins with PDLLA to form composite microparticles (10-300 µm) in a single step and under ambient conditions26 (Fig. 4). Here, the protein particles are uniformly distributed throughout the liquefied PDLLA to form a molten mixture that is then sprayed into a collecting chamber, trapping the protein within the polymer particles. Protein functionality and controlled release have been demonstrated with calcitonin and insulin after using this method (Fig. 5). Furthermore, PGSS has been demonstrated on a large scale making it well suited to industrial applications27. Micronization of drugs by this technique is limited by the poor interactions between CO2 and many drugs. However, many of the polymers used in drug delivery applications (PLA, PLGA, etc.) are plasticized under scCO2
conditions allowing a number of pharmaceuticals to be coprecipitated into composite particles. Gas anti-solvent techniques Gas anti-solvent techniques differ from either the RESS or the PGSS process, because as the name suggests, the gas is used as an anti-solvent to reduce the solvating power of the organic solvent in which the solute is contained. Because of this ability for pressurized gases such as CO2 to dissolve and expand organic solvents, they can be used for the precipitation of solids from organic solutions8. The expanded solvent can then be purged from the system, with the particular solid washed by the anti-solvent (CO2). The particles can then be removed from the collection chamber upon depressurization28. Many variations of the anti-solvent technique exist, including the original gas/supercritical antisolvent (G/SAS) technique, the precipitation of compressed anti-solvent (PCA), and the aerosol solvent extraction system (ASES) (Fig. 6). The GAS technique has been used most extensively to produce protein powders such as insulin28,29. Other proteins, including lysozyme and trypsin, have successfully been micronized to within 1-5 µm size range using this technique30. Gas anti-solvent techniques have been used to encapsulate insulin within PEG/PLLA nanoparticles with dichloromethane31 and dimethylsulphoxide32 used as the solvents. In addition to protein micronization, the GAS technique has successfully been used in the micronization of nonproteins such as copper indomethacin, a nonsteroidal anti-inflammatory drug33. Here, particles were formed that showed an eight-fold increase in the dissolution rate in water, compared with the unprocessed drug. Another technique, precipitation from a compressed anti-solvent (PCA) has also been used for the production of biologically active powders or polymeric microparticles. With PCA, a solid (such as a drug or polymer) is dissolved in an
Fig. 5 Results from an assay showing calcitonin activity after being encapsulated using scCO2 (taken from Whitaker et al.26). (A) An osteoclast cell observed in a no-calcitonin control. Without the presence of calcitonin, the osteoclasts are motile and possess characteristic membrane skirts (arrows). (B) Cell observed in calcitonin control. Osteoclasts are immotile. (C) Cell observed in release media from encapsulated calcitonin.
August 2005
45
RESEARCH REPORT
(A)
(B)
Polymer and/or drug dissolved in organic solvent
Nozzle
High pressure collection chamber
High pressure collection chamber
Supercritical anti-solvent
Expanded solution
Supercritical anti-solvent Supercritical anti-solvent and organic solvent
Fig. 6 Schematic of both (A) the gas anti-solvent (GAS) and (B) aerosol solvent extraction system (ASES) techniques. Here, the SCF is used as an anti-solvent to reduce the solvating power of the organic solvent containing the dissolved solute, allowing the precipitation of the particles.
organic solvent, before this liquid is sprayed continuously in small amounts into a vessel filled with scCO2. The solvent dissolves into the scCO2, but the solute does not, resulting in the precipitation of micronized particles28. This technique has been used to form polymer/drug microparticles with PLLA and a wide range of pharmaceuticals (gentamycin, naloxone, and naltrexone) dissolved in methylene chloride34. Continuous optimization of the process, such as the processing conditions and type/size of the nozzle, has lead to the formation of smaller more controllable particles35,36. The use of additional solvents in supercritical anti-solvent techniques provides benefits such as the broad number of compounds and solutes that are soluble in organic solvents. However, the use of toxic solvents such as dichloromethane37 and dimethylsulfoxide38 in the process yields residual solvents in the final product, which detracts from the benefits of such a system for producing bioactive compounds. The aerosol solvent extraction system (ASES) is a variation on the other anti-solvent processing methods, although it still uses the same principles. The organic solution (containing the drug/polymer or both) is sprayed into a vessel pressurized with the anti-solvent (normally scCO2), forming droplets. The nozzle disperses the droplets of liquid uniformly into a flowing anti-solvent stream, causing super-saturation and the rapid precipitation of the solute. Microparticles (1-10 µm) were first produced using this system by Bleich and coworkers39 before drugs such as indomethacin40, parahydroxybenzoic acid38, and lysozyme38 were encapsulated within (PLLA). The model peptide tetracosactide 46
August 2005
has also been encapsulated within PDLLA and PLGA microparticles41, showing favorable results, in terms of polymer degradation properties, compared with traditional methods such as double emulsion42. The ASES technique has also been used in the micronization of steroid drugs intended for pulmonary delivery37 where, in order for drugs to be delivered to the lungs effectively, it is important that the particle size does not exceed 1-5 µm43,44. While this is possible using traditional methods such as jet milling, this technique is problematic because of the insufficient brittleness of the resultant drug crystals. The steroid drugs budesonide and beclomethasone-17,21-dipropionate have successfully been micronized to within the desired size range using the ASES method37, with a residual solvent (dichloromethane) content of less than 350 ppm. Achieving such small particle sizes is one of the major benefits of the ASES technique, as it has been shown to compare favorably with the standard GAS method33. Methods for producing particles using high-pressure gases In addition to the methods mentioned above, several techniques have been developed that use many of the same principles but have added a unique step to improve the particle properties or to gain more control of the particle size. A number of techniques have used the supercritical as an anti-solvent, such as solution-enhanced dispersion by supercritical fluids (SEDS)45 and supercritical assisted atomization (SAA)46. The SEDS method was patented by Hanna and York after research at the university of Bradford and Bradford Particle Design Ltd47. The underlying principle
RESEARCH REPORT
of this technique is based on dispersing an aqueous solution containing the biomaterial with scCO2 and an additional polar solvent in a three-channel coaxial chamber. The scCO2 is used to extract the aqueous phase from the product, although an additional polar solvent is still required because water is not soluble in scCO2, which has been a major problem in previous particle production. Therefore, solvent also acts as a precipitating agent and a modifier, allowing the CO2 to remove the water. Many of the problems seen with other SCF methods, such as agglomeration, long drying times and poor scale up have been reduced or eliminated by using this method48. This technique has been used to produce particles of salmeterol xinafoate (an anti-asthma drug) with a size of 1-5 µm47 that were shown to compare favorably with conventionally produced powders when tested in dry powder inhaler devices49. Another interesting use of this technique is for the production of plasmid DNA-loaded particles with the potential for delivery via inhalation or through the skin (transdermally)50. The delivery of plasmid DNA in powder formulations could potentially play an important role in the treatment of genetic disorders such as cystic fibrosis. In SAA, a thermostated packed contactor is used to solubilize the scCO2 in a liquid solvent containing the solute. This solution is then sprayed into a precipitator at atmospheric pressure and a two-step atomization is obtained: the primary droplets are further divided into secondary droplets owing to scCO2 expansion from the inside of the primary droplets. This process was used successfully to micronize the drug griseofulvin, using N2 to assist in the evaporation of the droplets51. Antibiotics such as ampicillin have also been particulated using the SAA method; a particle size of 1-5 µm makes them suitable for aerosol delivery52.
Incorporating drugs into controlled matrices In addition to particle formation, high-pressure or supercritical CO2 has been used to fabricate porous polymeric controlledrelease matrices. These porous matrices or scaffolds can often have a dual purpose; one example of this is in tissue engineering applications. Here, they act as both supports for cell growth53 and release devices for the delivery of growth factors – molecules known to have an important role in tissue regeneration by affecting a number of cellular processes54. The pores can be formed by either the nucleation of gas bubbles during depressurization55 or the fusion of preformed
polymeric microspheres56. Angiogenic growth factors (those that promote blood vessel formation) such as vascular endothelial growth factor (VEGF)57 and basic fibroblastic growth factor (bFGF)7 have been impregnated successfully within scCO2-fabricated poly(DL-lactic-co-glycolic acid) (PDLLGA) matrices. The former study demonstrated blood vessel formation over seven days and up to 90% activity retention. Hile et al. achieved controlled release of the growth factor, but used an additional solvent (methylene chloride) in the process. What is more, these solvent residues were above the levels deemed safe by the US Pharmacopoeia. New supercritical techniques have been developed that allow the incorporation of thermally labile bioactive molecules into controlled matrices during processing. This is made possible by the ability of high-pressure CO2 to liquefy certain amorphous polymers in ambient conditions, facilitating the incorporation of molecules into the processing step in an a solvent-free environment. A novel one-step supercritical mixing technique has been used successfully to incorporate model proteins into PDLLA matrices, with the proteins trapped within the porous matrix upon depressurization58,59. The subsequent release of the active protein was made controllable by changes in the depressurization speed, as this determines the number and size of the pores. This work was taken a step further by the incorporation of growth factor into PDLLA scaffolds for tissue-engineering applications60. The growth factor bone morphogenetic protein-2 (BMP-2) was incorporated into the polymer matrix during processing, with subsequent controlled release of the molecule inducing the formation of new bone tissue in vivo. Guney and Akgerman used a continuous-flow method to impregnate PLGA with two drugs: 5-fluorouracil (for chemotherapy) and β-estradiol (for estrogen hormone therapy)61. Both of these drugs were previously shown to be soluble in scCO262 and, as a result, were incorporated into PLGA matrices during a continuous flow process. The subcritical CO2 technique first used to produce tissueengineering scaffolds54 has also been used to form porous gene delivery matrices for tissue-engineering applications by fusing together microspheres containing plasmid DNA63. The theory is that DNA that contains specific gene sequence encoding for protein growth factors, could be delivered into the patient, facilitating endogenous growth factor production, and inducing tissue regeneration. Poor transfection data, which is likely to be due to the minimal exposure of the DNA
August 2005
47
RESEARCH REPORT
plasmids to the cells in question, has cast doubt over the use of this method64,65. However, this gas foaming technique has been used to deliver plasmid DNA encoding for plateletderived growth factor (PDGF) into PLGA matrices, with the achievement of controlled release61.
Conclusions and outlook There is a wide range of benefits on offer when using supercritical fluids (SCFs) such as more controllable particle size and high levels of diffusivity in both synthetic polymers and drugs. What is more, the use of supercritical fluids such as scCO2 has provided a clean alternative to traditional techniques that employ toxic organic solvents or elevated
Acknowledgment The authors would like to thank John Barry for contributing Fig. 1.
REFERENCES
33. Warwick, B., et al., Ind. Eng. Chem. Res. (2002) 41 (8), 1993
1. Langer, R., Science (2001) 293, 58
34. Falk, R., et al., J. Controlled Release (1997) 44 (1), 77
2. Roberts, C. J., and Debendetti, P. G., AIChE J. (2002) 48 (6), 1140
35. Jarmer, D. J., et al., J. Supercrit. Fluid (2003) 27 (30), 317
3. Sencar-Bozic, P., et al., Int. J. Pharm. (1997) 148 (2), 123
36. Jarmer, D. J., et al., Langmuir (2004) 20 (17), 7254
4. Heller, J., Controlled drug release from bioerodible polymer implants. In Medical Applications of Controlled Release, Langer, R. S., and Wise, D. L., (eds.), CRC Press, Boca Raton, (1984), 69
37. Steckel, H., et al., Int. J. Pharm. (1997) 152 (1), 99
5. Tom, J. W., and Debendetti, P. G., Biotechnol. Prog. (1991) 7 (5), 403 6. Falk, R. F., and Randolph, T. W., Pharm. Res. (1998) 15 (8), 1233 7. Hile, D. D., et al., J. Controlled Release (2000) 66 (2-3), 177 8. Foster, N., et al., Ind. Eng. Chem. Res. (2003) 42 (25), 6476 9. Woods, H. M., et al., J. Mater. Chem. (2004) 14 (11), 1663 10. Debendetti, P. G., et al., Fluid Phase Equilib. (1993) 82, 303
38. Tu, L. S., et al., Powder Technol. (2002) 126 (2), 134 39. Bleich, J., et al., Int. J. Pharm. (1993) 13, 367 40. Bleich, J., and Muller, B. W., J. Microencapsul. (1996) 13 (2), 131 41. Witschi, C., and Doelker, E., J. Controlled Release (1998) 51 (2-3), 327 42. Witschi, C., and Doelker, E., Int. J. Pharm. (1998) 171 (1), 1 43. York, P., Manufacturing Chemist (2000) 71 (6), 26 44. Hybertson, B. M., et al., J. Aerosol Med. (1993) 6, 275
11. Matson, D. W., et al., J. Mater. Sci. (1987) 22 (6), 1919
45. Hanna, M. H., and York, P., (1995) Method and apparatus for the formulation of particles. Patent WO 96/00 610
12. Matson, D. W., et al., Mater. Lett. (1986) 4 (10), 429
46. Reverchon, E., Ind. Eng. Chem. Res. (2002) 41 (10), 2405
13. Alessi, P., et al., Ind. Eng. Chem. Res. (1996) 35 (12), 4718
47. York, P., Pharm. Sci. Tech. Today (1999) 2 (11), 430
14. Charoenchaitrakool, M., et al., Ind. Eng. Chem. Res. (2000) 39 (12), 4794
48. Beach, S., et al., Org. Proc. Res. Dev. (1999) 5 (5), 370
15. Kayrak, D., et al., J. Supercrit. Fluid (2003) 26, 17
49. Feeley, J. C., et al., In Proc. Millennium World Congress of Pharm. Sci., San Francisco, (2000), 108
16. Hu, G., et al., Chin. J. Chem. Eng. (2003) 11 (4), 403
48
temperatures. This has led to labile bioactive molecules, such as proteins, drugs, and nucleic acids, being introduced at the polymer processing stage – something that was not previously possible when using synthetic polymers. One drawback is the limited number of drugs that can be processed in scCO2, although additional solvents have been used in techniques where the SCF acts as the antisolvent to circumvent this. However, smaller particle sizes, combined with increased versatility and improved rates of dissolution with scCO2 techniques, suggests an important role for this technology in future drug delivery applications. NT
17. Turk, M., and Lietzow, R., AAPS Pharm. Sci. Tech. (2004) 5 (4), 56
50. Tservistas, M., et al., Biotechnol. Bioeng. (2001) 72 (1), 12
18. Juliano, R. L., In Controlled Drug Delivery, Fundamentals and Applications, Robinson, J. R., and Lee, V. H. L., (eds.), Marcel Dekker, New York, (1987), 555
51. Reverchon, E., et al., J. Pharm. Pharmacol. (2004) 56 (11), 1379 52. Reverchon, E., et al., J. Pharm. Pharmacol. (2003) 55 (11), 1465
19. Tom, J. W., et al., ACS Symposium Series (1993) 514, 238
53. Langer, R., and Vacanti, J. P., Science (1993) 260, 920
20. Kim, J.-H., et al., Biotechnol. Prog.(1996) 12 (5), 650 21. Matsuyama, K., et al., J. App. Polym. Sci. (2003) 89, 742
54. Mckay, I. A and Leigh, I., Growth Factors: A Practical Approach, IRL Press, Oxford, (2003)
22. Kerc, J., et al., Int. J. Pharm. (1999) 182 (1), 33
55. Barry, J. J. A., et al., Biomaterials (2004) 25 (17), 3559
23. Hao, J., et al., J. Mater. Chem. (2005) 15 (11), 1148
56. Mooney, D. J., et al., Biomaterials (1996) 17 (14), 1417
24. Rodrigues, M., et al., J. Supercrit. Fluid (2004) 29 (1-2), 175
57. Sheridan, M. H., et al., J. Controlled Release (2000) 64 (1-3), 91
25. Hao, J., et al., J. Pharm. Sci. (2004) 93 (4), 1083
58. Howdle, S. M., et al., Chem. Comm. (2001) (1), 109
26. Whitaker, M. J., et al., J. Controlled Release (2005) 101 (1-3), 85
59. Watson. M. S., et al., Adv. Mater. (2002) 14 (24), 1802
27. Jung, J., and Perrut, M., J. Supercrit. Fluid (2001) 20 (3), 179
60. Yang, X. B. B., et al., Tissue Eng. (2004) 10 (7-8), 1037
28. Yeo, S.-D., et al., Biotechnol. Bioeng. (1993) 41 (3), 341
61. Guney, O., and Akgerman, A., AIChE J. (2002) 48 (4), 856
29. Yeo, S.-D., et al., J. Pharm. Sci. (1994) 83 (12), 1651
62. Guney, O., and Akgerman, A., J. Chem. Eng. Data (2000) 45 (6), 1049
30. Winters, M. A., et al., J. Pharm. Sci. (1996) 85 (6), 586
63. Shea, L. D., et al., Nat. Biotechnol. (1999) 17 (6), 551
31. Elvassore, N., et al., J. Pharm. Sci. (2001) 90 (10), 1628
64. Ledley, F. D., Pharmaceut. Res. (1996) 13 (11), 1595
32. Caliceti, P., et al., J. Controlled Release (2004) 94 (1), 195
65. Jang, J.-H., and Shea, L. D., J. Controlled Release (2003) 86 (1), 157
August 2005
