Smart Microgel Studies. Polyelectrolyte and Drug-Absorbing

Langmuir 2003, 19, 8675-8684

8675

Smart Microgel Studies. Polyelectrolyte and Drug-Absorbing Properties of Microgels from Polyether-Modified Poly(acrylic acid) Lev Bromberg, Marina Temchenko, and T. Alan Hatton* Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received April 30, 2003. In Final Form: July 30, 2003 The study dealt with microgels that consist of loosely cross-linked poly(acrylic acid) onto which poly(ethylene oxide)-b-poly(propylene oxide)-b-poly(ethylene oxide) (PEO-PPO-PEO) copolymers such as Pluronic F127 (average composition EO99PO67EO99) or L92 (EO8PO52EO8) segments are grafted. Microgels based on the more hydrophobic Pluronic L92 exhibited highly porous structure, while the microgels containing Pluronic F127 were generally larger and possessed smooth surfaces, a homogeneous structure, and lower ion-exchange capacity. A good correlation was observed between the microgel surface potential obtained from the potentiometric titration data and the independently measured electrophoretic ζ-potential. The differences in the microgel ion-exchange capacity account for the differences in the capacity of the microgels to absorb weakly basic anticancer drugs such as mitomycin, mitoxantrone, and doxorubicin. The uptake of hydrophobic drugs such as Taxol, estradiol, progesterone, and camptothecin was dictated by the content of PPO in the microgels, which determines their solubilizing capacity. The exclusion of proteins of varying molecular weight by the microgels reveals the effective pore size, which is below 7.5 nm in the F127-based microgels but is on the order of tens of nanometers in the L92-based microgels.

Introduction The delivery of drugs currently used in cancer treatment is performed mostly through either intravenous drug infusions or topical application.1 Due to unfavorable pharmacodynamics, the infusions need to be performed frequently, exacerbating the potential for side effects. Oral treatment with anticancer drugs can be preferred as this delivery route is convenient to patients, reduces administration costs, and facilitates the use of more chronic treatment regimens. Thus far, low oral bioavailability has usually limited the development of cancer treatment by the oral route. We have developed novel temperature- and pH-sensitive gel microparticles as feasible vehicles for oral delivery of anticancer drugs with high bioavailability.2 The microgels consist of covalently cross-linked copolymers of poly(acrylic acid) (PAA) and poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO) PEO-PPO-PEO copolymers (trademark Pluronic). The Pluronic-PAA copolymers are obtained via a unique one-step template polymerization of acrylic acid in the presence of Pluronic, with a chain transfer to the Pluronic.3-6 The synthesis results in graft-comb structures of high molecular weight (above 105 g/mol) and is optimized to yield an extremely low concentration of toxic admixtures, acceptable for medicinal use.3 The copolymers typically consist of about 50 wt % of PAA chains grafted to the polyether backbone of the Pluronic copolymer via C-C bonding.3-6 Introduction of divinyl cross-linkers such as ethylene glycol dimethacrylate (EGDMA) into the synthetic route developed for the Pluronic-PAA enables preparation of submillimeter gel microparticles capable of volume phase (1) DelaFlor-Weiss, E.; Uziely, B.; Muggia, F. M. Ann. Oncol. 1993, 4, 723. (2) Bromberg, L.; Temchenko, M.; Hatton, T. A. Langmuir 2002, 18, 4944. (3) Bromberg, L. Ind. Eng. Chem. Res. 1998, 37, 4267. (4) Bromberg, L. J. Phys. Chem. B 1998, 102, 1956. (5) Bromberg, L. J. Phys. Chem. B 1998, 102, 10736. (6) Bromberg, L. Ind. Eng. Chem. Res. 2001, 40, 2437.

transitions in aqueous solutions.2 Since these transitions can be controlled by changing the quality of aqueous solvent, the microgels are termed smart, or intelligent. When swollen, the size of microgels (tens of microns) will prohibit their passage through gastrointestinal walls. Yet, because of the strong mucoadhesive properties of the Pluronic-PAA,7-10 the microgels can adhere longer onto the mucosal tissues, thus providing potential for sustained release of the loaded drugs by diffusion. The surface activity and structural parameters of the Pluronic-PAA11 provide for their effect as penetration enhancers in drug transport in gastrointestinal models. The Pluronic-PAA polymers have been shown to form micellelike aggregates in aqueous solutions, capable of solubilizing hydrophobic compounds.12-16 The ability of the microgels to undergo volume transitions in response to changing pH2 enables loading of drugs in the expanded state of the microgel, which upon collapse can hold the drug without substantial release. These are properties useful in gastrointestinal applications, since the pH in the stomach is in the 1-2 range, causing the microgels to remain collapsed and thus preventing an undesirable drug release in the stomach. The pH in the intestine is 6.2-7.4, leading to a highly swollen state of the microgels and to a facile drug release from these particles. The favorable characteristics of the Pluronic-PAA microgels motivate our ongoing effort to elucidate the drug-microgel interactions. (7) Bromberg, L. E.; Ron, E. S. Adv. Drug Delivery Rev. 1998, 3, 197. (8) Bromberg, L. E.; Orkisz, M. J.; Ron, E. S. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1997, 38, 626. (9) Bromberg, L. E. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1999, 40, 616. (10) Bromberg, L. J. Pharm. Pharmacol. 2001, 53, 109. (11) Bromberg, L. Langmuir 1998, 14, 5806. (12) Ho, A. K.; Bromberg, L. E.; Huibers, P. D. T.; O’Connor, A. J.; Perera, J. M.; Stevens, G. W.; Hatton, T. A. Langmuir 2002, 18, 3005. (13) Bromberg, L. E.; Barr, D. P. Macromolecules 1999, 32, 3649. (14) Huibers, P. D. T.; Bromberg, L. E.; Robinson, B. H.; Hatton, T. A. Macromolecules 1999, 32, 4889. (15) Bromberg, L.; Magner, E. Langmuir 1999, 15, 6792. (16) Bromberg, L.; Temchenko, M. Langmuir 1999, 15, 8627.

10.1021/la030187i CCC: $25.00 © 2003 American Chemical Society Published on Web 09/18/2003

8676

Langmuir, Vol. 19, No. 21, 2003

Bromberg et al. Table 1. Properties of the Drugs Used in the Study drug

molecular mass

Log P

solubility parameter (MPa1/2)

Taxol camptothecin progesterone β-estradiol mitomycin Ca mitoxantronea doxorubicina

853.3 348.4 314.5 272.4 334.1 444.2 543.5

4b 0.87c 3.77d 3.86d -0.4g 0.24c 1.85g

28.7e 29.2e 17.6f 21.1f 30.5e 27.3e 27.3e

a Data are for free bases and not hydrochloride salts. b Reference 17. c Calculated using ClogP Program (BioByte Corp.). d Data from refs 18. e Calculated using Molecular Modeling Pro. (version 2.14, WindowChem Software, Inc.). f Hildebrand parameter data are taken from ref 19. g Reference 20.

Figure 1. Structures of the anticancer drugs under study.

Herein, we report on the microgels based on Pluronic copolymers representing the most hydrophobic (L92) and relatively hydrophilic (F127) extremes of this class of block copolymers. Since changes in the nature of the Pluronic bonded to the polyelectrolyte appeared to yield different microgel structures, the effective pore size was characterized by exclusion of proteins of varying molecular size. The polyelectrolyte properties of anionic microgels determine their high capacity for small cations.2 The choice of the low molecular weight drugs used in the present study (Figure 1) was dictated by their representative properties such as hydrophobicity (Table 1) and ability to acquire a positive charge in water. Two anthracycline drugs, doxorubicin and mitoxantrone, are available as hydrochloride salts and are weakly basic when dissolved in water. Doxorubicin is used to treat patients with sarcoma, lymphoma, and breast and ovarian carcinoma.21 Unfortunately, its use is limited by potential cardiac toxicity, particularly at high cumulative doses. Liposomal, micellar, or microgel formulations of doxorubicin may improve its toxicity profile. Mitoxantrone is a highly active agent in the treatment of metastatic (17) Niethammer, A.; Gaedicke, G.; Lode, H. N.; Wrasidlo, W. Bioconjugate Chem. 2001, 12, 414. (18) Johnson, M. E.; Blankschtein, D.; Langer, D. J. Pharm. Sci. 1995, 84, 1144. Kamlet, M. J.; Doherty, R. M.; Carr, P. W.; Mackay, D.; Abraham, M. H.; Taft, R. Environ. Sci. Technol. 1988, 22, 503. (19) Michaels, A. S.; Wong, P. S. L.; Pratner, R.; Gale, R. M. AIChE J. 1975, 21, 1073. (20) Hansch, C.; Leo, A. Exploring QSAR; American Chemical Society: Washington, DC, 1995; Vols. 1 and 2. (21) Singal, P. K.; Iliskovic, N. N. Engl. J. Med. 1998, 339, 900.

breast cancer.22 Analogously to the above anthracyclines, mitomycin C behaves as a weak base in water; it is effective against gastrointestinal cancer23 as well as eyelid gland carcinomas when applied topically.24 In contrast to the aforementioned weakly basic drugs, four other compounds studied herein are devoid of ionic groups and are generally water-insoluble, although they vary in hydrophobicity. Two of these compounds, estradiol and progesterone, are steroid hormones used in hormone replacement therapy in women.25 β-Estradiol blocks carcinogenesis and thus is widely used in prevention of mammary cancer.26 Paclitaxel (Taxol) and derivatives are active against various tumors27 and have also been used to treat malignant glioma and brain metastases.28 Camptothecin, a monoterpene-derived indole alkaloid, is a valuable compound as a chemical precursor for the semisynthetic derivatives topotecan and irinotecan, which are used clinically as anticancer agents.29 Camptothecin also possesses an antitumor activity on its own due to an ability to inhibit DNA topoisomerase I.30 The current study revealed differences in the drugmicrogel interactions, depending on the Pluronic used in the microgel synthesis. These findings may provide novel strategies for the design of microgels for drug delivery applications. Experimental Section Materials. Nonionic copolymers Pluronic F127 NF and L92 were obtained from BASF Corp. and used without further treatment. Acrylic acid (99%, vinyl monomer), ethylene glycol dimethacrylate (98%, divinyl cross-linker), dodecane (99+%, solvent), and 4,4′-azobis(4-cyanovaleric acid) (75+%, azo initiator) were purchased from Aldrich Chemical Co. and used as received. Lauroyl peroxide (97%, redox initiator) was obtained from Fluka Chemie AG (Switzerland). Poly(vinylpyrrolidinone-co-1-hexa(22) Konig, E.; Kurbacher, C.; Schwonzen, M.; Breidenbach, M.; Mallmann, P. Anticancer Drugs 2002, 13, 827. (23) Baker, L. H.; Izbicki, R. M.; Vaitkevicius, V. K. Med. Pediatr. Oncol. 1976, 2, 7. (24) Shields, C. L.; Naseripour, M.; Shields, J. A.; Eagle, R. C. Ophthalmology 2002, 109, 2129. (25) Lemay, A. J. Obstet. Gynaecol. Can. 2002, 24, 711. (26) Huggins, C.; Briziarelli, G. Science 1959, 1, 129(3357):1285. (27) Kroger, N.; Achterrath, W.; Hegewisch-Becker, S.; Mross, K.; Zander, A. R. Cancer Treat. Rev. 1999, 25, 279. Fellner, S.; Bauer, B.; Miller, D. S.; Schaffrik, M.; Fankha¨nel, M.; Spruss, T.; Bernhardt, G.; Graeff, C.; Fa¨rber, L.; Gschaidmeier, H.; Buschauer, A.; Fricker, G. J. Clin. Invest. 2002, 110, 1309. Nathan, F. E.; Berd, D.; Sato, T.; Mastrangelo, M. J. Cancer 2000, 88, 79. Papadimitriou, C. A. Cancer 2000, 89, 1547. Belani, C. P. Chest 2000, 117 (Suppl 1), 144S. Bunn, P. A.; Kelly, K. Chest 2000, 117 (Suppl 1), 138S. (28) Glantz, M. J.; Chamberlain, M. C.; Chang, S. M.; Prados, M. D.; Cole, B. F. Semin. Radiat. Oncol. 1999, 9, 27. Brandes, A. A.; Pasetto, L. M.; Monfardini, S. Anticancer Res. 2000, 20, 1913. (29) Baudin, E.; Docao, C.; Gicquel, C.; Vassal, G.; Bachelot, A.; Penfornis, A.; Schlumberger, M. Ann. Oncol. 2002, 13, 1806. (30) Staker, B. L.; Hjerrild, K.; Feese, M. D.; Behnke, C. A.; Burgin, A. B.; Stewart, L. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 15387.

Smart Microgel Studies

Langmuir, Vol. 19, No. 21, 2003 8677

Table 2. Properties of the Proteins Used in the Study

protein insulin chymotrypsin albumin dehydrogenase catalase

hydroisomolecular dynamic electric mass, kDa radius, nm point 5.5 25 66 141 250

1.3a 2.4 3.6 4.9 6.3

5.3 5.2 4.7 5.4 5.4

measured extinction coefficient, mM-1 cm-1 (λ, nm) 5.67 (280) 51.2 (280) 43.9 (280) 47.9 (280) 101 (407)

a Data are for the monomeric form of insulin; a hexameric form has a hydrodynamic radius of ca. 2.6 nm (data from Protein Solutions, Inc., Lakewood, NJ; published on the Web at http:// www.protein-solutions.com/appnotes/app202.pdf).

decene) (Ganex V-216) (dispersion stabilizer) was obtained from International Specialty Products (Wayne, NJ). Doxorubicin hydrochloride and Taxol, both of 99% purity, were obtained from Hande Tech USA (Houston, TX), a subsidiary of Yunnan Hande Technological Development Co. (Kunming, P.R. China). Mitomycin C, camptothecin, and mitoxantrone dihydrochloride (97%) were obtained from Sigma-Aldrich Co. and used as received. β-Chymotrypsin from bovine pancreas (essentially salt-free), alcohol dehydrogenase from Bakers yeast (crystallized and lyophilized), bovine albumin (minimum 99%, essentially fatty acid free), catalase from bovine liver (crystallized), and Zn2+insulin from bovine pancreas (anhydrous) were all obtained from Sigma Chemical Co. (St. Louis, MO) and used as received except for insulin, which was suspended in deionized (DI) water (pH 5.3) and lyophilized. The properties of the proteins used are collected in Table 2. All other chemicals, gases, and organic solvents of the highest purity available were obtained from commercial sources. Microgel Synthesis. Synthesis was carried out on a laboratory scale in an adiabatic mode. Acrylic acid (40 mL) was partially neutralized by addition of 5 M NaOH aqueous solution (0.5 mL). Pluronic (24 g) was dissolved in the resulting solution under nitrogen, and a desired amount of EGDMA was added. The amounts of EGDMA added varied, and the molar ratio of the EGDMA to acrylic acid set in the reaction mixture was used to designate the degree of cross-linking of the microgels [XL, mol % ) 100 × (number of moles of EGDMA/number of moles of acrylic acid)]. Lauroyl peroxide (100 mg) and 4,4′-azobis(4cyanovaleric acid) (100 mg) were dissolved in 2 mL of acrylic acid and added to the solution of Pluronic in acrylic acid. The resulting solution was deaerated by nitrogen bubbling for 0.5 h and added to a three-necked 0.5-mL flask containing 1 wt % solution of Ganex V-216 in dodecane (200 mL). The flask was vigorously stirred by a mechanical stirrer and deaerated by constant nitrogen purge from the bottom. Then the flask was heated to 70 °C using an oil bath and kept at that temperature under stirring and nitrogen purge. After about 1 h, formation of white particles was observed on the flask walls. The reaction was continued at 70 °C for another 4 h. The reactor was disassembled, and the contents of the reactor were filtered using Whatman filter paper (retention size, 10 µm). The microgel particles were extensively washed by hexane and dried under a vacuum. In the control series of experiments, poly(acrylic acid) microgels were synthesized by an analogous procedure, but without addition of Pluronic. Spherical particles were observed under a microscope. Particle sizing was performed in hexane using a ZetaPlus zeta potential analyzer (Brookhaven Instruments Co.). Purification of the microgel particles by Soxhlet extraction and dialysis as well as chemical analysis have been described in detail previously.2 Microgel Titration and ζ-Potential Measurements. To obtain the apparent pKa and ion-exchange capacity of the microgels, bulk titration was performed at 25 ( 1 °C using a 736 GP Titrino potentiometric titration system (Metrohm Ltd., Herisau, Switzerland). Titration of microgels in their acidic or basic form was carried out in a degassed 75 mM NaCl solution with 0.1 M NaOH or 0.1 M HCl as a titrant, at 0.2-0.5 wt % polymer concentrations. The titration was carried out slowly for 8 h to allow for proper equilibration.31 The resulting titration curves yielded clear inflection points (Figure 2), which enabled

Figure 2. Typical potentiometric titration curves of F127PAA-EGDMA microgels (XL ) 1 mol %). The dotted line shows the derivative of the pH vs titrant volume curve. For other details, see the text. calculation of the apparent pKa and ionization degree (R) of the titratable carboxyl groups:32,33

pKa ) pHi + log[Na+] - log[X/2]

(1)

where [Na+] is the bulk sodium concentration, [X/2] is the concentration of carboxyl groups when half of the acrylic acid mers are protonated, and pHi is the pH at the inflection point on the titration curve. The potential at the surface of shear (ζ-potential) of the 0.01 wt % microgel suspensions in 75 mM NaCl, with pH adjusted to a desired value by adding 5 M NaOH or HCl, was measured using a ZetaPals zeta potential analyzer (Brookhaven Instruments Co.) with built-in software employing the Smoluchowski ζ-potential model.34 At the chosen salt concentration, the Debye length (κ-1 ) 0.304/([NaCl])1/2 35) was of the order of 1 nm, while the swollen microgel particles were of 10-100 µm in radius (r). Therefore, we were well within the Helmholtz-Smoluchowski limit (rκ f ∝), where the simple expression of Smoluchowski is valid, and the ζ-potential is roughly equal to the diffuse layer potential.36 Appropriate corrections for the suspension viscosity were made after measurements of the zero-shear viscosity using an oscillatory shear rheometer.37 Each measurement was performed five times. Microscopy. Microgel particles were studied in their dry (as made) and equilibrium swollen form. A 1 wt % microgel suspension (pH 7.0) was equilibrated at room temperature for 2 days; then it was snap-frozen in liquid nitrogen and lyophilized. The dried samples were mounted onto a scanning electron microscopy (SEM) stub with nonconductive glue and sputtercoated with Au/Pd (150 Å). Images of the particles were taken at various magnifications using a JEOL 6320 FE6-SEM microscope. Solute Loading. Water-Soluble Solutes. The maximum loading level of the proteins and of the basic drugs, doxorubicin, mitoxantrone, and mitomycin C, into microgels was measured using a Millipore Ultrafree-MC centrifugal filter device (Millipore Co.). Microgel particles were suspended in Tris buffer (5 mM, pH 7.0), and 50 µL of the suspension (2 mg gel/mL buffer) was equilibrated with either 1 mg/mL protein solution or 3.0 mM stock solution of a drug (450 µL) for 16 h while shaking. Shaking was performed using a KS10 orbital shaker (BEA-Enprotech Corp., Hyde Park, MA) in an environmental chamber at 37 °C. After equilibration, the microgel particles were filtered off by centrifugation (10000g, 0.5 h) and the supernatant was assayed for drug concentration. A Shimadzu model 1600 PC spectrophotometer with a temperature-controlled quartz cuvette (path (31) Kawaguchi, S.; Nishikawa, Y.; Kitano, T.; Ito, K.; Minakata, A. Macromolecules 1990, 23, 2710. (32) Helfferich, F. Ion Exchange; Dover: New York, 1995. (33) Kawaguchi, S.; Yekta, A.; Winnik, M. A. J. Colloid Interface Sci. 1995, 176, 362. (34) Hunter, R. J. Zeta Potential in Colloid Science: Principles and Applications; Academic Press: London, 1986. (35) Larson, R. G. The Structure and Rheology of Complex Fluids; Oxford University Press: London, 1999; p 90. (36) Attard, P.; Antelmi, D.; Larson, I. Langmuir 2000, 16, 1542. (37) Bromberg, L. Macromolecules 1998, 31, 6148.

8678


Bromberg et al.

length, 1 cm) was used for electronic absorption measurements. In the case of the proteins (except for catalase), we found the sensitivity of the absorption measurements at 280 nm to be insufficient, while the direct absorption reading at wavelengths below 200 nm38 was sensitive to the presence of salts and pH. Therefore, the concentration of the proteins in the supernatant was measured, after appropriate dilutions, using the bicinchoninic acid (BCA) assay.39 The extinction coefficients of doxorubicin (λ ) 482 nm) and mitoxantrone (λ ) 614 nm) were determined at pH 7.0 to be 12 200 and 22 100 M-1 cm-1, respectively. The concentration of mitomycin C was assayed by high-performance liquid chromatography (HPLC) using a Capcell Pak MF Ph-1 (100 × 4.6 mm i.d.; particle size, 5 µm) column (Phenomenex, Torrance, CA). HPLC was performed on a Hewlett-Packard 1090 system with an autosampler and a variable wavelength UV detector controlled by the HPLC Chemstation software (Hewlett-Packard). Deionized water was used as the mobile phase (flow rate, 1 mL/min; injection volume, 25 µL), and detection was carried out at 365 nm. The typical retention time of mitomycin C was 4.88 min. The drug uptake was expressed as

U (mmol drug/g gel) ) [(Ac - Ar)/Ac]VCs/Mgel where Ac and Ar are the absorbance or HPLC readings in the appropriately diluted stock solution and in the system equilibrated with microgel, respectively, V ) 0.5 mL is the total volume of the system, Cs ) 3 µmol/mL is the concentration of the stock solution, and Mgel ) 0.1 mg is the microgel mass. The U values were measured in triplicate for each solute and for each gel, respectively. Hydrophobic Solutes. The loading of the hydrophobic solutes such as Taxol, camptothecin, and estradiol into microgels was measured by equilibrating solute adsorbed onto steel beads with the 1 wt % suspension of microgels (pH 7.0). Stainless steel beads (1-3 µm diameter) were soaked in a 10 mM solution of the solute in either acetonitrile (Taxol), dimethyl sulfoxide (DMSO) (camptothecin), or absolute ethanol (estradiol, progesterone), and the solvent was then stripped off under a vacuum. The beads were used in order to enhance the area of contact between the microgel suspension and Taxol. The beads were separated into several fractions. One fraction was added to a polypropylene vial containing the microgel suspension (0.5 mL), and the vial was gently shaken in a horizontal position in an environmental chamber at 20 or 37 °C. Then the beads were recovered from the suspension by using a magnet. The beads were dried under a vacuum and placed into an appropriate solvent (0.5 mL), where the solute was extracted after shaking overnight. The solvent fraction was assayed for the solute concentration using HPLC. The control fraction of loaded beads was subjected to extraction without equilibration with the microgel suspension. The solute concentrations were measured in triplicate using the HPLC system described above. The chromatography assay for Taxol comprised the use of a Capcell Pak C18 UG 120 (150 × 4.6 mm i.d.; particle size, 3 µm) column (Phenomenex), acetonitrile0.1% phosphoric acid in DI water (55:45 v/v, 1.3 mL/min) as a mobile phase, and UV detection at 227 nm. The typical retention time of the Taxol peak was 3.46 min. For camptothecin assay, a solution of the drug in DMSO (50 µL) was injected onto the aforementioned C18 column and eluted at 1 mL/min flow rate and 40 °C using 0.1 M ammonium acetate (pH 5.6) and acetonitrile as the mobile phase. A linear gradient of 45-85% acetonitrile was applied, and the UV/vis detection was carried out at 254 and 370 nm.40 The retention time of the camptothecin peaks was 7.5-8.5 min. The peak area was integrated and used to calculate the camptothecin concentration using standard calibration curves. The HPLC assay of estradiol and progesterone solutions in ethanol was performed as described previously.16 (38) Eichenbaum, G. M.; Kiser, P. F.; Dobrynin, A. V.; Simon, S. A.; Needham, D. Macromolecules 1999, 32, 4867. (39) Smith, P. K.; Krohn, R. I.; Hermanson, G. T.; Mallia, A. K.; Gartner, F. H.; Provenzano, M. D.; Fujimoto, E. K.; Goeke, N. M.; Olson, B. J.; Klenk, D. C. Anal. Biochem. 1985, 150, 76. (40) Greenwald, R. B.; Pendri, A.; Conover, C.; Gilbert, C.; Yang, R.; Xia, J. J. Med. Chem. 1996, 39, 1938.

Figure 3. SEM microphotographs of the F127-PAA-EGDMA (a) and L92-PAA-EGDMA (b) particles, as made (XL ) 1 mol %). Pyrene loading was studied in order to assess the temperaturedependent aggregation within microgels.2 A stock solution of 1 mM pyrene in absolute methanol was prepared, from which 1-3 µL was added to 3.0 mL of an aerated aqueous sample. The sample was then allowed to equilibrate for 24 h at a given temperature, and excitation and emission (λex ) 335 nm) spectra were recorded. Fluorescence spectra were recorded using a 10 mm path length quartz cell in a thermostated cuvette holder using a Shimadzu model RF-5301 PC spectrofluorophotometer under controlled temperature conditions (slit widths, 1.5 nm). The ratio of the intensities of the first (373 nm) to the third (384 nm) vibronic peaks (I1/I3) in the emission spectra of the monomer pyrene were used to indicate the polarity of the pyrene microenvironment. In the control experiments, the I1/I3 values were measured in 1 wt % Pluronic F127 and L92 aqueous solutions, as described elsewhere.12

Results and Discussion SEM Microgel Characterization. Significant differences were observed in the shape and structure of asprepared F127-PAA-EGDMA and L92-PAA-EGDMA particles (Figure 3). The majority of the F127-PAAEGDMA particles were spherical and contained no observable pores. The size distribution was somewhat broad, ranging from about 5 to 100 µm. Particle sizing performed in hexane yielded an effective median diameter of 13 µm and a polydispersity of 1.4 for this type of particle. The surfaces of the spherical F127-PAA-EGDMA particles contained a treelike pattern with submicron depth. The formation of similar characteristic patterns in


droplets of the discontinuous phase caused by the competition between phase separation and chemical reaction has been reported in the literature.41 In our case, such patterns could be due to rapid polymerization of acrylic acid4 followed by phase separation. Note that while the acrylic acid monomer is miscible with the hydrocarbon in which the polymerization takes place, the resulting PAA and Pluronic-PAA are not. The partially polymerized, phase-separated microdroplets that still contain monomer on their surfaces are bridged by further polymerization and cross-linking reactions. The droplet interface becomes diffuse over time, and the droplets eventually fuse together. Unlike the large and smooth F127-PAA-EGDMA particles, the more hydrophobic L92-PAA-EGDMA particles consisted of a spongelike, interconnected and fused network of submicron spherical particles. The resulting “sponge agglomerates” were rough but spherical and were on average smaller than the F127-PAAEGDMA particles (particle sizing in hexane yielded an effective median diameter of 4 µm and a polydispersity of 1.5 for L92-PAA-EGDMA). The profound differences observed in the structures of the F127-PAA-EGDMA and L92-PAA-EGDMA particles (compare structures in Figure 3) are believed to be due to the differences in reaction kinetics and phase separation speed due to the more than 2-fold higher content of PPO in the Pluronic L92 polymer. We have discovered previously4 that a higher content of PPO in the template polymerization of acrylic acid on Pluronic results in a dramatic increase of branching and cross-linking in Pluronic-PAA at the same degree of conversion of acrylic acid. This is because of the much higher susceptibility of PPO (as compared to PEO) to the hydrogen abstraction reactions.4,6 The highly reactive PPO radicals are incorporated into the growing Pluronic-PAA structure during the early stages of the polymerization, resulting in the polymer network containing areas with highly concentrated PPO units. Moreover, the high concentration of PPO is likely to yield multiple hydrogen abstraction sites on one PPO chain. This early incorporation of PPO into the polymer network creates a number of highly cross-linked domains that are concentrated in possible cross-linking points. Once PPO is consumed into the growing copolymeric network, the remaining acrylic acid will continue to react and build relatively linear polymer chains branching away from these cross-linked domains, linking them together. Certainly, this as well as the higher tendency of more PPO-rich Pluronics to phase-separate in solutions can translate into the formation of stable microdroplets that become connected later as the polymerization proceeds (L92-PAA-EGDMA) as compared to the formation of Pluronic-PAA copolymers in relatively large droplets, which become phase-separated as a whole during the later stages of polymerization (F127-PAA-EGDMA). Upon equilibrium swelling in water (Figure 4), the F127-PAA-EGDMA microgels expanded but maintained their spherical shape. The surface pattern developed into 1-2 µm deep “craters” that were not interconnected. We have previously reported that the interior of the F127PAA-EGDMA microgels, imaged by SEM using freezefracturing techniques, appeared to be homogeneous, devoid of any discernible pore structure. In the L92-PAAEGDMA, the gaps between interconnected micro- or nanospheres (resulting from polymerized microdroplets) widened upon swelling, yielding a well-developed porous (41) Tanaka, H.; Suzuki, T.; Hayashi, T.; Nishi, T. Macromolecules 1992, 25, 4453.

Langmuir, Vol. 19, No. 21, 2003 8679

Figure 4. SEM microphotographs of the equilibrium-swollen, snap-frozen, and lyophilized F127-PAA-EGDMA (a,b) and L92-PAA-EGDMA (c) microgel particles.

structure with pores on the order of 100 nm wide. The submicron spheres themselves did not expand significantly in water, presumably because of the high PPO content making them hydrophobic. Characterization of Microgels by the Pyrene Solubilization Technique. Pyrene is a hydrophobic fluorescent molecule often used as a sensitive probe of the polarity of polymeric solutions, including solutions of Pluronic-PAA.12,13 The ratio I1/I3 of the intensities of the first (λ1 ≈ 373 nm) and the third (λ3 ≈ 384 nm) vibronic bands in the emission spectrum of monomeric pyrene decreases with increasing hydrophobicity, and thus pyrene “senses” the presence of the most hydrophobic regions of aggregates and hydrophobic cores of micelles.2,12,13

8680


Figure 5. Temperature dependency of the I1/I3 intensity ratio of pyrene in 1% w/w aqueous solutions of Pluronic F127 and Pluronic L92 and suspensions of Pluronic-PAA cross-linked by ethylene glycol dimethacrylate (XL ) 1 mol %). The pH is 7.0 throughout.

Figure 5 depicts the effect of temperature on the I1/I3 ratio in aqueous solutions of Pluronics F127 and L92 and in the corresponding microgels F127-PAA-EGDMA and L92-PAA-EGDMA. In Pluronic solutions, the I1/I3 ratio is roughly equal to that in water at temperatures below the critical micellization temperature (CMT), indicating a strongly polar environment.12 At temperatures above the CMT, I1/I3 decreases sharply with the partitioning of pyrene to a more hydrophobic environment as micellar aggregates begin to form. When aggregation begins, the pyrene reports to hydrophobic PPO-rich micelle cores. The I1/I3 values in the Pluronic F127 and L92 solutions differ only in the transition range of temperatures (15-25 °C), reflecting the lower CMT for the more hydrophobic L92 than for the hydrophilic F127, and coincide at temperatures where aggregation is complete. Notably, the F127 solutions remain visibly clear above the CMT, while the L92 solutions are cloudy, indicating formation of large aggregates that phase-separate. However, the pyrene probe does not detect differences in its environment above the CMT, irrespective of the Pluronic used, because its environment is primarily PPO. The pyrene behavior in microgel suspensions was drastically different from that in the corresponding Pluronic solutions. The polarity of the Pluronic-PAA solutions at temperatures below the CMT is generally lower than in solutions of the parent Pluronic, because of the presence of sterically constrained, hydrophobic domains of Pluronic onto which several PAA segments are attached.12,14 In F127-PAA solutions, such domains are even more hydrophobic than the cores of the micellelike aggregates that form above the CMT, and thus the I1/I3 values increase above the CMT.12 This trend generally holds for the F127-PAA-EGDMA suspensions (Figure 5). That is, I1/I3 in F127-PAA-EGDMA is below 0.85 at low temperatures, indicating a strongly nonpolar environment, but increases at T > 20 °C up to 1.45 at 25 °C, at which point (corresponding to the formation of micelles, compare with the CMT of Pluronic) the I1/I3 values start decreasing. The temperature effects on I1/I3 indicate redistribution of pyrene between hydrophobic PPO domains, constrained by PAA grafting and cross-linking, and micellar aggregates that form above the CMT due to the rearrangement of polymer segments. Interestingly, there are remarkable differences between temperatureinduced structural rearrangements in suspensions of F127-PAA-EGDMA and L92-PAA-EGDMA. The latter microgels, which possess 2-fold higher PPO content, are sensed as being less hydrophobic at temperatures below the CMT. This observation can be rationalized in terms of the differences in size and effective permeability of the hydrophobic domains. It has been shown that the higher

Bromberg et al.

PPO content in Pluronic copolymer results in an exponential increase in the degree of branching and crosslinking in the synthesis of Pluronic-PAA.4,6 Small and over-cross-linked domains simply fail to accommodate pyrene inside, and hence the higher I1/I3 values in the L92-PAA-EGDMA suspension below the CMT. The I1/ I3 versus T trend in the L92-PAA-EGDMA suspension is congruent with that in the corresponding L92 solution, indicating that a process analogous to micellization in the L92 solution takes place inside the microgels at the same temperatures. Specifically, an assembly of the constrained domains at T > CMT gradually occurs, with these “interdomain” structures sensed as being more hydrophobic by pyrene. The I1/I3 values in L92-PAA-EGDMA at elevated temperatures are lower than in pure PPO and close in value to those in hydrocarbons.12 In summary, the trends observed in the pyrene probing experiments seem to be in accord with the higher PPO content and thus higher hydrophobicity of the microgels based on Pluronic L92. Such microgels are likely to possess a higher content of cross-linked domains and heterogeneity, which is directly observed by SEM. Importantly, by simply replacing one parent Pluronic with another in the synthetic procedure we were able to alter the structure and self-assembling pattern of the resulting microgels significantly. Experiments to accentuate differences that stem from the incorporation of different Pluronics are described below. Polyelectrolyte Properties and Surface Potential of the Microgels. Polyelectrolyte properties of the microgels can be related to their structure. We applied the well-known Katchalsky model42-44 to calculate the surface charge of spherical microgels.33,45 Conventionally, the apparent dissociation constant of a carboxylic acid (Ka ) [-COO-][H+]/[COOH]) can be related to the standard free energy change of the dissociation process (∆G°) as

∆G° RT

pKa ) -log Ka ) 0.4343

(2)

where R is the gas constant, T is the Kelvin temperature, and log e ≈ 0.4343. From the definition of Ka and of the degree of ionization,

R)

[COO-] [COOH] + [COO-]

(3)

we obtain the Henderson-Hasselbach expression,32

pKa ) pH + log

1-R R

(4)

For a polyacid such as poly(acrylic acid), an additional amount of electrostatic work (∆Gel) is required to remove a proton against the electrostatic forces due to the charges already present on the polymer chain.44 Equation 4 thus needs to be modified:

pH ) pK0 - log

∆Gel 1-R + 0.4343 R RT

(5)

(42) Katchalsky, A.; Gillis, J. Recl. Trav. Chim. Pays-Bas 1949, 68, 879. (43) Katchalsky, A.; Lifson, S.; Eisenberg, H. J. Polym. Sci. 1951, 7, 571. (44) Katchalsky, A.; Michaeli, I. J. Polym. Sci. 1955, 15, 69. (45) Zhang, H.; Dubin, P. L.; Kaplan, J.; Moorefield, C. N.; Newkome, G. R. J. Phys. Chem. B 1997, 101, 3494.


Langmuir, Vol. 19, No. 21, 2003 8681

Figure 6. Relationship between the apparent dissociation constant (pKa) and the degree of ionization (R) for PAA microgels that contained Pluronic (L92-PAA-EGDMA and F127-PAAEGDMA) and microgels without Pluronic (PAA-EGDMA). T ) 25 °C; the concentration of the added NaCl is 75 mM. The parameters were calculated from experimental potentiometric titration data using eq 4.

where pK0 is the intrinsic dissociation constant obtained by extrapolating pKa to R ) 0 and/or from the titration and gel swelling data using eq 1, and ∆Gel is expressed in terms of the electrostatic potential (ψa) at the site where the proton originated:

∆Gel ) -eNAψa

(6)

Herein, NA ) 6.02 × 1023 mol-1 and e ) 1.60 × 10-19 C are Avogadro’s number and the elementary charge, respectively. Numerous attempts have been made to calculate the surface potential of the spherical particle.46 We will use only expressions derived from the Katchalsky model (eqs 5 and 6) allowing for the estimate of ψ0 from the titration data:33

pKa - pK0 ) 0.4343eψ0(R)/kBT

(7)

where kB ) 1.36 × 10-23 J/K is the Boltzmann constant and ψ0(R) is the surface potential of a spherical microgel particle. The intrinsic dissociation constants (K0) were found from eq 1 as well as by extrapolating pKa to R ) 0 in the pKa versus R curves (Figure 6). For microgels with XL ) 1 mol %, the pK0 was found to be 6.27 and 4.95 for the L92PAA-EGDMA and F127-PAA-EGDMA, respectively, with no measurable discrepancy between the two methods of pK0 determination. The PAA-EGDMA microgels synthesized under similar conditions, but without the use of PEO-PPO-PEO copolymer as a chain transfer agent, had a pK0 of 4.86, which corresponded well with pK0 values of 4.67 and 4.56 found for un-cross-linked poly(acrylic acid) and acrylic acid, respectively.5,44 In the case of F127PAA-EGDMA gels, which do not have interconnected macropores (Figures 3 and 4) and possess a dry content of the hydrophobic propylene oxide (PO) groups of only 15 wt %, the pKa versus R plot (Figure 6) exhibited a wide plateau in the area 0.3 < R < 0.8, as if the electrostatic free energy of the microgels did not change appreciably after the initial fraction of the carboxyls became ionized and essentially until the titration endpoint. The shape of the curve resembles that predicted by mean-field theories for a smeared charge distribution on a sphere.46 The F127-PAA copolymers exhibit formation of rather large (several nanometers in size), micellelike (46) Ullner, M. In Handbook of Polyelectrolytes and Their Applications; Tripathy, S. K., Kumar, J., Nalwa, H. S., Eds.; American Scientific: Stevenson Ranch, CA, 2002; Vol. 3, Chapter 10, pp 271308.

aggregates in aqueous solutions.14 Such aggregates that form cooperatively at certain critical temperatures due to hydrophobic interactions between PPO segments alter the electrostatic charge density and the availability of the carboxyls on the PAA chains to neutralization. Such micellelike aggregates are weakened at high pH due to the progressive ionization of the PAA segments and increasing osmotic pressure exerted by the mobile counterions, which tend to pull the F127-PAA segments apart.5 The pKa versus R plot for the L92-PAA-EGDMA microgels (PO content, 36 wt % per dry network) was significantly different from that of the more hydrophilic microgels (Figure 8), with pKa systematically shifted to higher values. The 25- and 20-fold higher K0 value in the case of L92-PAA-EGDMA, when compared with values for the PAA-EGDMA and F127-PAA-EGDMA microgels, respectively, reflects the overall compact state of the polyelectrolyte segments neighboring numerous hydrophobic PPO segments. The greater hydrophobicity of the gel requires a higher degree of ionization to destroy the hydrophobic aggregates,47 and hence the shift of pK0 to more alkaline values. The dissociation behavior of carboxyls in polyelectrolytes depends on the electron-withdrawing or inductive effects of neighboring groups, the dielectric properties of the solvent, and the conformation of polyelectrolyte chains.31 The dielectric constant of the solvent in the microgel can change as a result of its swelling and collapse; that is, the dielectric constant depends on the intragel polymer volume fraction.48,49 On the other hand, the dielectric constant significantly affects the number of carboxyl-counterion ion pairs. 49 The number of free (not ion-paired) counterions decreases logarithmically with increasing polymer concentration.50,51 The presence of hydrophobic segments affects the ion pairing in the ionized groups in the vicinity of the hydrophobic groups, which may lead to the local supercollapsed state. Conformational changes in polyelectrolytes modified with hydrophobic segments depend on the ratios of segment sizes and of segment energies.52 The swelling and collapse of the intragel aggregates depend on the balance between Coulombic repulsion and the surface energy of these aggregates.53 The conformational changes, in turn, influence the ionization behavior of the microgels in the present study. In our previous publication,2 large volume changes were reported in the Pluronic-PAA microgels both upon ionization of the carboxyls and upon aggregation of the intraparticle hydrophobic segments. The inductive effect of the neighboring charged groups steeply decreases with the distance between these groups,46 and thus the appearance or disappearance of the hydrophobic domains that contribute to the average distance between carboxyls will increase or lower, respectively, the pK0. Notably, no significant plateau on the pKa versus R curve for the L92-PAA-EGDMA was observed once neutralization started at R > 0.2 (Figure 6), which indicates the absence of a dramatic, cooperative breakup of hydrophobic aggregates above a certain high pH. Analogously, the temperature-induced aggregation reflected in (47) Philippova, O. E.; Hourdet, D.; Audebert, R.; Khokhlov, A. R. Macromolecules 1997, 30, 8278. (48) Khokhlov, A. R.; Kramarenko, E. Yu. Macromol. Theory Simul. 1994, 3, 45. (49) Kramarenko, E. Yu.; Khokhlov, A. R.; Yoshikawa, K. Macromol. Theory Simul. 2000, 9, 249. (50) Solis, F. J.; Olvera de la Cruz, M. Macromolecules 1998, 31, 5502. (51) Dobrynin, A. V.; Rubinstein, M. Macromolecules 2001, 34, 1964. (52) Dobrynin, A. V.; Rubinstein, M. Macromolecules 2000, 33, 8097. (53) Dobrynin, A. V.; Rubinstein, M. Macromolecules 1999, 32, 915.

8682


Bromberg et al.

Figure 8. Equilibrium uptake of weakly basic anticancer drugs into microgels at 37 °C and pH 7.0.

Figure 7. Effect of microgel ionization on the ζ-potential (ζ, filled points) and the surface potential (ψ0, solid line). The ψ0 values were calculated using eq 8 from the titration data.

the pyrene fluorescence pattern (Figure 5) showed less significant changes caused by intraparticle aggregation in L92-PAA-EGDMA than in its F127-PAA-EGDMA analogue. It may be suggested that because these microgels consist of interconnected but hydrophobic and cross-linked clusters, the latter have some steric constraints limiting the “intercluster” assembly. On the other hand, due to the pores between the clusters the microgels exhibit higher overall availability of the carboxyls for ionization. It was of interest to compare the surface potential ψ0 calculated from the titration data using eq 7 and the independently measured ζ-potential (Figure 7). We assigned a negative sign for the ψ0, as is traditional for polyanions. A good correlation was observed between ψ0 and ζ values for both types of microgels under study. There has been much debate about the model-dependent relationship between electrokinetic mobility and the surface charge density of charged particles,36,54-58 with suggestions that counterion condensation in the outer gel layers of the spherical particles would make the titratable charge greater than the net charge that contributes to the ζ-potential.36,54,59 We have, however, found good correspondence between titration and electrophoretic mobility measurements under the simple assumptions inherent in the corresponding experimental techniques. Interestingly, more hydrophobic microgels L92-PAA-EGDMA possessed larger potentials than their F127-PAAEGDMA counterparts, consistent with both the higher effective surface charge and larger ion-exchange capacity. The maximum ion-exchange capacity of COOH groups (54) Huang, Q. R.; Dubin, P. L.; Moorefield, C. N.; Newkome, G. R. J. Phys. Chem. B 2000, 104, 898. (55) O’Brien, R. W.; White, L. R. J. Chem. Soc., Faraday Trans. 2 1978, 93, 3761. (56) Dukhin, S. S.; Derjaguin, B. V. Electrokinetic Phenomena Surface and Colloid Science; Matijevic, E., Ed.; John Wiley & Sons: New York, 1974. (57) Levine, S.; Neale, G. H. J. Colloid Interface Sci. 1974, 47, 520. (58) Fernandez-Nieves, A.; Fernandez-Barbero, A.; de las Nieves, F. J. Langmuir 2000, 16, 4090. (59) Lyklema, J. J. Electroanal. Chem. 1968, 18, 341.

was found by titration to be 6.1 and 7.2 mequiv/g dry gel for F127-PAA-EGDMA and L92-PAA-EGDMA, respectively. These differences can be explained by the welldeveloped pore network in the L92-PAA-EGDMA, absent in the F127-PAA-EGDMA. The equilibrium loading of weakly basic, ionized drugs into microgels at 37 °C is shown in Figure 8. As is seen, the uptake by porous but more hydrophobic L92-PAAEGDMA microgels was consistently higher than for more hydrophilic F127-PAA-EGDMA microgels. This corresponded to the differences in the total ion-exchange capacities of the microgels noted above. The ratio of the drug loading into these two types of microgels was 0.840.91 of the ratio of their respective total ion-exchange capacities. Hence, the loading of the small basic drugs into the microgels is determined by ion-exchange equilibria. The decrease of the equilibrium uptake observed for both types of gels in the sequence mitomycin > mitoxantrone > doxorubicin (Figure 8) corresponded to both the size and hydrophobicity of the drugs, either of which increased in exactly the same order (Table 1). The size of the solute influences the steric “crowding” threshold that limits the equilibrium uptake via ion exchange. Effect of Cross-Linking Degree on Protein Loading into Microgels. As we have seen, our microgels exhibited significantly different polyelectrolyte properties, depending on the type of Pluronic used. It was of interest to investigate the equilibrium uptake of proteins of different molecular sizes by the microgels (Table 2). The protein size exclusion threshold measured on microgels with varying degrees of effective cross-linking can reveal the effective pore size.38 Five mostly globular proteins varying 50-fold in molecular mass were incubated with the microgels at 37 °C and pH 6.5-7.0. A substantial protein loading can be expected,38 provided that the protein effective diameter is smaller than the average pore size in the gel. Since our microgels contain Pluronics, which are known to affect the protein adsorption onto microparticles depending on the length of the PPO block,60,61 the protein uptake in our case could be influenced by specific interactions other than Coulombic attraction and size exclusion effects. Therefore, we term the obtained pore size “effective” in what follows. The results of equilibrium protein uptake are presented in Figure 9. As is seen, the uptake of insulin, the smallest and the most hydrophobic protein62 studied, is 12-24% higher by the L92-PAA-EGDMA microgels. Since no size exclusion effects are expected in the case of insulin, this effect can be attributed to the adsorption via hydrophobic (60) Norman, M. E.; Williams, P.; Illum, L. Biomaterials 1993, 14, 193. (61) Tan, J. S.; Butterfield, D. E.; Voycheck, C. L.; Caldwell, K. D.; Li, J. T. Biomaterials 1993, 14, 823. (62) Bromberg, L. E.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 143.


Langmuir, Vol. 19, No. 21, 2003 8683

Figure 9. Equilibrium protein uptake by microgels of different effective cross-linking density (XL) at 37 °C and pH 6.5-7.0.

interactions. Very minor, if any, statistically significant differences between loadings in the two types of microgels were observed in the uptake of chymotrypsinogen and albumin, indicating an absence of the size exclusion effects. However, the uptake of the largest proteins, dehydrogenase and catalase, was significantly different between the two types of microgels. That is, only very lightly crosslinked (XL ) 0.05%) F127-PAA-EGDMA adsorbed dehydrogenase and catalase to any great extent, with uptake of these proteins rapidly decaying with the increasing degree of cross-linking. Based on the dehydrogenase cutoff at XL ) 0.5%, we conclude that the effective pore size in the F127-PAA-EGDMA with this degree of cross-linking is below 7.5 nm, which is the effective size of catalase based on its crystalline structure.38 No decrease in the dehydrogenase uptake, but a significant reduction in catalase uptake, was observed in L92-PAAEGDMA with XL g 0.5%, which indicates the effective pore size in this type of microgels to be larger than the size of catalase. Uptake of Hydrophobic Drugs. The aforementioned enhancement of the insulin uptake by the more hydrophobic microgels (Figure 9) hinted at the effects of hydrophobic interactions on solute loading. Therefore, the uptake of hydrophobic, water-insoluble drugs was tested to assess the solubilizing capacity of the microgels (Figure 10). The more hydrophobic L92-PAA-EGDMA microgels possessed a slightly higher capacity for the hydrophobic drugs. Interestingly, while the L92-PAA-EGDMA had a 20-60% higher capacity than F127-PAA-EGDMA at 20 °C, where the formation of the hydrophobic aggregates is not complete (compare with Figure 7), only a 4-19% capacity increase with the L92-PAA-EGDMA relative to the F127-PAA-EGDMA was observed at 37 °C. This indicates that the micellelike aggregates in the F127PAA-EGDMA contribute quite significantly to the overall solubilizing capacity of the microgels, while aggregates in L92-PAA-EGDMA are less effective. Generally, the uptake is on the order of µmol/g gel, which is about 1000fold lower than in the case of the bases that are loaded

Figure 10. Equilibrium uptake of hydrophobic drugs by microgels at pH 7.0 at 20 and 37 °C.

via ion-exchange equilibria. Taxol and other hydrophobic drugs are solubilized by the hydrophobic PPO domains in the microgels. The equilibrium solubility of Taxol in water, for instance, is 16-50-fold less than the equilibrium solubility in a 1 wt % microgel suspension;2 similar results hold for the other hydrophobic drugs under study. The solubilizing capacity of the microgels for Taxol and steroid hormones is at least equal to that of the Pluronic-PAA micelles.16 The characteristic increase in the drug loading capacity at higher temperatures (compare results at 20 and 37 °C in Figure 10) above the critical aggregation temperatures provides additional evidence for the suggested mechanism of hydrophobic drug solubilization into micellelike aggregates within the microgels via entropic interactions. The formation of the aggregates within microgel particles due to the aggregation of dangling chains of Pluronic has been demonstrated in our previous work.2 The micelles in Pluronic-PAA solutions typically have a solubilizing capacity similar to that in the Pluronic aqueous solutions.16 The least hydrophobic of the four drugs, camptothecin, showed higher uptake by the microgels (see Log P data in Table 1 and Figure 10). This can be explained by the fact that the loading of the amphiphilic solute can occur both via the solubilization into hydrophobic PPO cores of the aggregates and via entrapment into the more hydrophilic core-swollen gel interface.13 Concluding Remarks We have examined the effect of the Pluronic content on the structure and behavior of Pluronic-PAA microgels. It appears that by simply altering the PPO content of the Pluronic in the emulsion/dispersion polymerization procedure3,6 one can obtain significant variations in the structure and consequent behavior of the resulting microgels. The microgels based on Pluronic L92 have a more

8684


hydrophobic content overall than analogous microgels based on Pluronic F127 and have a heterogeneous, porous structure. The porosity determines the accessibility of the carboxyls and thus the ion-exchange capacity of the microgels. In turn, the hydrophobicity influences the solubilizing capacity of the microgels for hydrophobic solutes. We believe that the novel microgels should add to the arsenal of pharmaceutical technology because of their capacity to change behavior in response to environmental stimuli, their ability to hold both hydrophobic and ionic drugs, and their benign, nonirritating nature. The stabilization of proteins in aqueous solutions by Pluronic-PAA copolymers because of their micellization

Bromberg et al.

and polyelectrolyte properties,63 coupled with the ability of the microgels to effectively load and hold hydrophobic drugs and the high loading of basic drugs and mucoadhesive properties,7-10 provides unusual possibilities for anticancer therapies by oral delivery. Acknowledgment. This work was supported in part by the Singapore-MIT Alliance (SMA) and the Cambridge-MIT Institute (CMI). LA030187I (63) Bromberg, L. J. Pharm. Pharmacol. 2001, 53, 541.

Smart Microgel Studies. Polyelectrolyte and Drug-Absorbing

Recommend Documents