Method for Quantifying the PEGylation of Gelatin Nanoparticle Drug

Chem. , 2007, 79 (12), pp 4574–4580 ... Publication Date (Web): May 17, 2007. Copyright © 2007 American .... Julia Engert , Roman Mathaes , Gerhard Wi...
1 downloads 0 Views 467KB Size
Download PDF
Anal. Chem. 2007, 79, 4574-4580

Method for Quantifying the PEGylation of Gelatin Nanoparticle Drug Carrier Systems Using Asymmetrical Flow Field-Flow Fractionation and Refractive Index Detection Jan C. Zillies, Klaus Zwiorek, Gerhard Winter, and Conrad Coester*

Department of Pharmacy, Pharmaceutical Technology and Biopharmaceutics, Ludwig-Maximilians-University Munich, Butenandtstrasse 5, 81377 Munich, Germany

The PEGylation of colloidal drug carrier systems protects them from a rapid clearance from the blood stream and therefore prolongs their plasma half-lives. This fundamental concept is nowadays widely applied whereas the analytical description, i.e., the quantification of the PEGylation process, is still challenging due to the poor spectrophotometrical properties of PEG. The aim of this work is to quantify the PEGylation process of gelatin nanoparticles by utilizing the combination of asymmetrical flow field-flow fractionation (AF4) and refractive index (RI) detection and to demonstrate the potential of AF4 in the work with colloidal drug carrier systems. An AF4 separation mechanism of gelatin nanoparticles and PEG was developed without further sample preparation. After separation, the PEGylation could be directly quantified from the respective RI data and a threshold of a maximum amount of PEG that can be bound onto the surface of the nanoparticles could be determined. The PEGylation could be further visualized by atomic force microscopy (AFM). In sum, the presented results show the successful application of AF4 in the field of colloidal drug carrier systems, and in combination with AFM, both techniques can be stated as promising tools for the future analysis of colloidal drug carrier systems. In the field of targeted drug delivery, enormous advances were made throughout the last years.1-3 Last but not least based on the knowledge of the complete human genome, a broad variety of possible targets could be defined and is waiting to be addressed by newly developed drug substances. Performing this task numerous colloidal-viral and nonviral-drug carrier systems were developed and are currently under investigation. The growing experience with these systems led to a detailed understanding of their in vivo fate subsequent to intravenous application. Hence, it is well-known that foreign material with a molecular weight higher than the renal threshold, is sequestered by the mononuclear * Corresponding author. Phone: +49 89 2180 77025. Fax: +49 89 2180 77026. E-mail: [email protected]. (1) Langer, R. Nature (London) 1998, 392 (6679, Suppl.), 5-10. (2) Yokoyama, M. Official J. Jpn. Soc. Artif. Organs 2005, 8 (2), 77-84. (3) Vasir, J. K.; Reddy, M. K.; Labhasetwar, V. D. Curr. Nanosci. 2005, 1 (1), 47-64.

4574 Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

phagocytic system (MPS).4 The involved phagocytotic cells recognize colloidal structures like liposomes and nanoparticles due to their rapid opsonization immediately occurring after application.4,5 The resulting low plasma half-life minimizing most therapeutic effects led to intensive research in order to overcome this problem. Ruling out the extensive plasma protein adsorption during opsonization should consequently provide a certain macrophage resistance expressed in a reduced MPS accumulation and a prolonged circulation time. Following from this idea, long circulating liposomes and nanoparticles have been developed.4-6 These systems possess a “molecular cloud” of polymer chains, especially poly(ethylene glycol) (PEG) chains grafted onto their surface, which sterically protect them from opsonization and enable prolonged circulation. Despite the long circulation properties, opsonization even of these systems cannot be completely excluded and is meanwhile critically discussed.5,7 Due to this successful approach, the concept of PEG grafting is widely applied for drug carriers as liposomes,5 nanoparticles based on synthetic4,8 and natural polymers,4,9,10 and polyplexes11,12 as well as for drug molecules on their own, namely, proteins13,14 and in the field of transfusion medicine for red blood cells.15 Besides the grafting process, the quantification of the introduced PEG is a matter of interest as the grade of PEGylation is decisive for a successful shielding5,7,16 and a homogeneous product is a necessary prerequisite during the regulatory process.17 (4) Owens, D. E.; Peppas, N. A. Int. J. Pharm. 2006, 307 (1), 93-102. (5) Yan, X.; Scherphof, G.; Kamps, J. J. Liposome Res. 2005, 15 (1 & 2), 109139. (6) Peracchia, M. T. STP Pharma Sci. 2003, 13 (3), 155-161. (7) Moghimi, S. M.; Szebeni, J. Pro. Lipid Res. 2003, 42 (6), 463-478. (8) Peracchia, M. T.; Fattal, E.; Desmaele, D.; Besnard, M.; Noel, J. P.; Gomis, J. M.; Appel, M.; d’Angelo, J.; Couvreur, P. J. Controlled Release 1999, 60 (1), 121-128. (9) Kaul, G.; Amiji, M. Pharm. Res. 2002, 19 (7), 1061-1067. (10) Lin, W.; Garnett, M. C.; Schacht, E.; Davis, S. S.; Illum, L. Int. J. Pharm. 1999, 189 (2), 161-170. (11) Walker, G. F.; Fella, C.; Pelisek, J.; Fahrmeir, J.; Boeckle, S.; Ogris, M.; Wagner, E. Mol. Ther. 2005, 11, 418-25. (12) Ogris, M.; Brunner, S.; Schuller, S.; Kircheis, R.; Wagner, E. Gene Ther. 1999, 6 (4), 595-605. (13) Fee, C. J.; Van Alstine, J. M. Chem. Eng. Sci. 2005, 61 (3), 924-939. (14) Veronese, F. M.; Pasut, G. Drug Discovery Today 2005, 10 (21), 14511458. (15) Garratty, G. Transfusion Med. Rev. 2004, 18 (4), 245-256. (16) Veronese, F. M. Biomaterials 2001, 22 (5), 405-417. 10.1021/ac062135e CCC: $37.00

© 2007 American Chemical Society Published on Web 05/17/2007

Quantifying PEG is in general an analytical challenge. The lack of chromophors excludes direct spectrophotometrical detection or excitation of any detectable fluorescence. Indirect methods like the determination of unreacted functionalities of the reaction counterpart will not provide a precise evaluation of the number of polymer chains bound. So, due to the various PEGylated systems, numerous analytical approaches have been proposed quantifying the PEGylation. The first successful attempts, utilizing colorimetric assays by complexing the PEG molecule with barium iodide, were described in the 1970s.18,19 The formed PEG barium iodide complex produces an absorption band at 535 nm that can be used for quantitative measurements. This complex formation was transferred to SDS-PAGE20 and is still in use for the evaluation of PEGylation efficiencies.21,22 Another colorimetric assay without initial extraction of PEG could be developed based on the partitioning of a chromophore present in aqueous ammonium ferrothiocyanate from an aqueous to an organic phase in the presence of PEG.23 However, despite rapidness, accuracy, and sensitivity, these methods suffer from exhibiting high blank values, which affects their detection limit. Finally, derivatization techniques have been used to produce UV-active PEG species like PEG dibenzoates.24 Besides these PEG modifying assays, detecting PEG without any changes of the molecule is extensively described in literature: flame ionization detection,25 dynamic surface tension detection,26 nuclear magnetic resonance spectrometry,27,28 mass spectrometry,29-32 and refractive index detection33-36 were applied for the quantitative (and qualitative) evaluation of PEG of different molecular weights. The separation of PEGylated from nonPEGylated compounds and PEG, required prior to any kind of detection, was accomplished by different chromatographic tech(17) Brown, F., Mire-Sluis, A., Eds. An Overview of Scientific and Regulatory Issues for the Immunogenicity of Biological Products. Dev. Biol. 2003, 171 (Proceedings of a Symposium held in Bethesda, MD 31 October-2 November 2001). (18) Childs, C. E. Microchem. J. 1975, 20 (2), 190-192. (19) Skoog, B. Vox Sang. 1979, 37 (6), 345-349. (20) Kurfurst, M. M. Anal. Biochem. 1992, 200, 244-48. (21) Zhang, G.; Wang, X.; Wang, Z.; Zhang, J.; Suggs, L. Tissue Eng. 2006, 12 (1-2), 9-19. (22) Natarajan, A.; Xiong, C.; Albrecht, H.; DeNardo, G. L.; DeNardo, S. J. Bioconjugate Chem. 2005, 16 (1), 113-121. (23) Nag, A.; Mitra, G.; Ghosh, P. C. Anal. Biochem. 1996, 237 (2), 224-231. (24) Murphy, R.; Selden, A. C.; Fisher, M.; Fagan, E. A.; Chadwick, V. S. J. Chromatogr. 1981, 211 (1), 160-165. (25) Kwong, E.; Baert, L.; Bechard, S. J. Pharm. Biomed. Anal. 1995, 13 (1), 77-81. (26) Miller K. E.; Bramanti, E.; Prazen, B. J.; Prezhdo, M.; Skogerboe, K. J.; Synovec, R. E. Anal. Chem. 2000, 72 (18), 4372-4380. (27) Leenheer, J. A.; Wershaw, R. L.; Brown, P. A.; Noyes, T. I. Environ. Sci. Technol. 1991, 25 (1), 161-168. (28) Mazarin, M.; Viel, S.; lard-Breton, B.; Thevand, A.; Charles, L. Anal. Chem. 2006, 78 (8), 2758-2764. (29) Na, D. H.; Lee, K. C. Anal. Biochem. 2004, 331 (2), 322-328. (30) Fakt, C.; Ervik, M. J. Chromatogr., B: Biomed. Sci. Appl. 1997, 700 (1 + 2), 93-100. (31) Crescenzi, C.; Di Corcia, A.; Marcomini, A.; Samperi, R. Environ. Sci. Technol. 1997, 31 (9), 2679-2685. (32) Nielen, M. W. F.; Buijtenhuijs, F. A. Anal. Chem. 1999, 71 (9), 18091814. (33) Delahunty, T.; Hollander, D. Clin. Chem. 1986, 32 (2), 351-353. (34) Oliva, A.; Armas, H.; Farina, J. B. Clin. Chem. 1994, 40 (8), 1571-1574. (35) Kirkland, J. J.; Dilks, C. H., Jr.; Rementer, S. W. Anal. Chem. 1992, 64 (11), 1295-1303. (36) Benincasa, M. A.; Caldwell, K. D. J. Chromatogr., A 2001, 925 (1-2), 159169.

niques as well as flow field-flow fractionation (FFFF). Due to the amphiphilic properties of PEG, RP-HPLC31,34 and organic FFFF37 are likewise applied as conventional SEC-HPLC13,26,33 and aqueous FFFF.35,36 Furthermore, capillary electrophoresis is applied for sample separation.29,38 Working with PEGylated colloidal drug carrier systems, the adoption of field-flow fractionation combined with subsequent PEG determination appears especially interesting. During field-flow fractionation, separation takes place in a hollow channel, allowing the concurrent fractionation of undissolved, i.e., suspended colloidal structures, and dissolved samples like PEG within a broad molecular mass range of ∼104-1018 Da ()1 nm-100 µm).39 Thus, the aim of this study was to utilize this unique characteristic to quantify the PEGylation process of gelatin nanoparticles, i.e., their covalent functionalization without further sample preparation directly from reaction media. The theory behind and the basic mechanisms of asymmetrical flow field-flow fractionation (AF4) are discussed elsewhere.40-43 In brief, FFF is a family of different subtechniques having the basic separation principle in common: A liquid carrier is transported through a separation channel forming a parabolic flow profile with layers of different velocities. Perpendicular to this laminar carrier flow, a field of separation is applied. During AF4, this field is erected by another liquid flow, called cross-flow, which contributes to the distribution of the analytes in the different areas of velocity of the laminar channel flow, finally leading to their fractionated elution. The cross-flow leaves the channel through an ultrafiltration membrane covering the bottom ()accumulation) wall of the channel. The analyte’s diffusion coefficient is thereby the critical number the cross-flow has to compensate. The StokesEinstein equation expresses the relationship between the diffusion coefficient D of a spherical particle and its hydrodynamic radius Rh in a medium with a given viscosity η:

D ) kT/6πηRh

(1)

where k is the Boltzmann constant and T is the absolute temperature. Together the cross-flow Vc and the diffusion coefficient D determine the elution (retention) time tr of the analytes as shown in the following equation:

tr ) t0Vcw2/6DV0

(2)

where t0 is the retention time of an unretained solute, V0 is the volume of the separation channel, and w is the channel height. This means that larger particles with a smaller diffusion coefficient (37) Kirkland, J. J.; Dilks, C. H., Jr. Anal. Chem. 1992, 64 (22), 2836-2840. (38) Na, D. H.; Park, E. J.; Youn, Y. S.; Moon, B. W.; Jo, Y. W.; Lee, S. H.; Kim, W. B.; Sohn, Y.; Lee, K. C. Electrophoresis 2004, 25 (3), 476-479. (39) Giddings, J. C. Science 1993, 260, 1456-65. (40) Colfen, H.; Antonietti, M. Field-flow Fractionation Techniques for Polymer and Colloid Analysis. In New Developments in Polymer Analytics I; Antonietti, M, et al., Eds.; Advances in Polymer Science 150; Springer: Berlin, 2000; pp 67-187. (41) Schimpf, M. E., Caldwell, K., Giddings, J. C., Eds. Field Flow Fractionation Handbook; Wiley-Interscience: New York, 2000. (42) Litzen, A.; Wahlund, K. G. Anal. Chem. 1991, 63 (10), 1001-1007. (43) Wahlund, K. G.; Giddings, J. C. Anal. Chem. 1987, 59 (9), 1332-1339.

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

4575

(eq 1) are stronger influenced by the actual cross-flow and therefore concentrated in areas of the laminar flow closer to the accumulation wall exhibiting lower velocities, which finally leads to prolonged retention (eq 2). The presents study describes an AF4 separation method for PEGylated gelatin nanoparticles and PEG as well as the determination of the PEGylation efficiency of gelatin nanoparticles via the concentration of unreacted PEG residues directly in the reaction media. Beyond the quantitative description of this process, we were interested in visualizing the PEGylation of the nanoparticles. As atomic force microscopy (AFM) offers resolution in the nanometer scale, we therefore investigated plain and PEGylated gelatin nanoparticles in AFM experiments. During AFM experiments, samples are mechanically scanned with a special cantilever resulting in a visualization of nanometer-sized structures like the helical morphology of DNA double strands,44,45 which makes AFM the method of choice in this context. EXPERIMENTAL SECTION Materials. Gelatin type A with Bloom 175, 1-ethyl-3-dimethyl aminopropylcarbodiimide (EDC), and 2-aminoethyltrimethylammoniumchlodride hydrochloride (cholamine) were purchased from Sigma-Aldrich (Taufkirchen, Germany). Glutaraldehyde and all buffer salts were purchased from Fluka (Buchs, Switzerland). Methoxypoly(ethylene glycol)amine (mPEG-NH2) (Mw 5884) was obtained from Nektar (Huntsville, AL). Preparation and Surface Modification of Gelatin Nanoparticles. Gelatin nanoparticles were prepared by a two-step desolvation method as described previously.46 Briefly, gelatin is dissolved in water (5% w/w) under heating and fractionated by a first desolvation with acetone. After adjusting the pH, the remaining sediment containing the high molecular weight fraction of gelatin is transferred into nanoparticles during a second desolvation step. Finally, the in situ formed particles are stabilized by cross-linking with glutaraldehyde. For PEGylation, 50 µL of an aqueous nanoparticle dispersion (20 mg/mL) was incubated for 2 h under constant shaking (800 rpm; 25 °C; Thermomixer comfort, Eppendorf, Hamburg, Germany) with various amounts of an mPEG-NH2 solution in Kolthoff’s borate buffer pH 8.4. Thereby, mPEG-NH2 reacts with residual aldehyde groups on the surface of the gelatin nanoparticles (Figure 1). After incubation, the total volume was completed with highly purified water to 1 mL resulting in a nanoparticle concentration of 1 mg/mL. Finally, an aliquot was transferred to AF4 analysis and the remaining particles were purified by 3-fold centrifugation and redispersion in PBS pH 7.4. Size and size distribution of the nanoparticles were determined by dynamic light scattering (DLS) using a Zetasizer Nano ZS (Malvern Corp., Worcestershire, UK). The size distribution, i.e., the sample quality, is expressed in the polydispersity index (PI) of the samples. The polydispersity index corresponds to the variance of the size distribution of a nanoparticle population and (44) LaVan, D. A.; Lynn, D. M.; Langer, R. Nat. Rev.: Drug Discovery 2002, 1 (1), 77-84. (45) Ravi Kumar, M. N., V.; Bakowsky, U.; Lehr, C. M. Biomaterials 2004, 25 (10), 1771-1777. (46) Coester, C. J.; Langer, K.; Von Briesen, H.; Kreuter, J. J. Microencapsulation 2000, 17 (2), 187-193.

4576

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

Figure 1. PEGylation reaction scheme of gelatin nanoparticles.

is a dimensionless parameter to describe the width of the size distribution:

PI ) σ2

Nanoparticles being used in the AFM experiments demanded, furthermore, a cationic surface charge to enable fixation on the anionic sample grid. Cationization was performed with the quaternary amine cholamine:47 after preparation and purification, the nanoparticles were suspended in highly purified water followed by dissolving cholamine in the resulting suspension. After 5 min of stirring, EDC was added to the reaction vessel in order to activate the free carboxyl groups on the surface of the unmodified nanoparticles for the coupling with cholamine. The resulting increase of the ζ potential was determined under defined ionic conditions in a 10-mmol sodium chloride solution at pH 7.0 using a Zetasizer Nano ZS (Malvern Corp., Worcestershire, UK). Quantification of the PEGylation Reaction. Measurements were conducted using a Wyatt Eclipse2 AF4 system (Wyatt Technology Europe GmbH, Dernbach, Germany). The channel height was 350 µm; the applied ultrafiltration membrane was a regenerated cellulose membrane with 5 kDa cutoff. Isocratic HPLC pump, autosampler, degasser, refractive index (RI), and UV detector are parts of the Agilent 1100 series (Agilent Technologies, Palo Alto, CA). A phosphate-buffered solution pH 7.0 containing 50 mmol of sodium phosphate and 150 mmol of sodium chloride was chosen as running buffer. The channel flow rate was 1 mL/min, the crossflow was 3 mL/min over 11 min (see Figure 2 for complete crossflow profile), and the measurement period amounted to 39 min. The detection was performed UV-spectrophotometrically (gelatin nanoparticles) and via refractive index (PEG 5000). The amount of unreacted PEG was calculated from the area under the curve (AUC) of its respective peaks in the resulting fractograms via a calibration curve. In addition, the recovery rate of PEG was determined to ensure that the applied 5 kDa cutoff ultrafiltration membrane provided sufficient experimental conditions. The PEGylation was quantified by comparing the AUCs of unreacted PEG before and after the PEGylation process. The whole study was repeated three times. Atomic Force Microscopy Analysis of Plain and PEGylated Gelatin Nanoparticles. Size and surface morphology of cationized plain and PEGylated gelatin nanoparticles were ana(47) Coester, C. New Drugs 2003, 1, 14-17.

Figure 2. Cross-flow profile (black line) and resulting RI signal of PEG 5000 (gray line). Dots indicate changes in the instrument’s running mode at 3 (injection /focusing), 8 (elution), 19 (reduction of cross-flow), and 24 min (purging of the injection loop).

lyzed by AFM in cooperation with JPK Instruments (Berlin, Germany). A JPK NanoWizard Life science version (JPK Instruments) was used in intermittent contact mode with a super sharp silicon cantilever (NanoWorld, Schaffhausen, Switzerland). These special cantilevers were chemically etched and end with a slim 200-nm-long and 2-nm-radius tip. The cantilevers had a spring constant of ∼42 N/m measured. Measurements in water were utilized with softer cantilevers having spring constants of ∼0.2 N/m (10-15-nm-radius tip). In both cases, the cantilever was excited close to its resonance frequency (air, ∼300 kHz; water, ∼12 kHz). RESULTS AND DISCUSSION Quantification of the PEGylation Reaction via AF4. For the intended AF4/RI quantification of the PEGylation of gelatin nanoparticles, there are two requirements to be met in order to succeed. At first, the baseline separation of the analytes has to be guaranteed in order to reproducibly determine alterations of the AUC of the PEG detection signal. Second, the linear relation between the concentration of PEG within the respective samples and its detection signal intensity have to be proven to enable the exact calculation of the PEG amount that could be bound onto the surface of gelatin nanoparticles. Baseline separation during AF4 experiments will be achieved by adequately adjusting the separation force, i.e., the cross-flow. Depending on the analytes, the cross-flow profile has to be newly developed for every analytical task. Figure 2 shows the crossflow profile developed for the fractionation of PEG and gelatin nanoparticles and the resulting RI signal of PEG. Refractive index detection was chosen due to the lack of chromophores within the PEG-molecule excluding direct spectrophotometrical detection. Dots indicate changes in the instrument’s running mode and in the cross-flow strength, respectively. The pressure sensitivity of the RI signal40 is thereby reflected by a baseline shift at 3 and 19 min as well as peaks at 8 and 25 min. The chosen cross-flow conditions provide a constant baseline in a time frame between 9 and 19 min, necessary for a reproducible calculation of the AUC and finally the quantification of PEG, as is shown for PEG itself in Figure 3a and water in Figure 3b. Before terminating the crossflow after 10 min of a constant baseline, the analyte is completely eluted. The graph exhibits a shoulder on the descending arm of

Figure 3. Fractogram section between 9 and 19 min showing the peak of PEG 5000, c ) 1 mg/mL (a) and the respective baseline recorded with water (b).

the peak indicating the presence of higher molecular weight compounds than PEG 5000 and, therefore, a certain heterogeneity of the PEG size distribution. These data are confirmed by the specification from the manufacturer. Due to its anionic polymerization, PEG has a certain variance in the number of ethylene oxide units.48 Since the whole amount of unreacted PEG present in the particular reaction medium should be quantified, there was no emphasis laid on a further separation of the PEG sample specimens. The recovery rate of PEG was determined ranging around 90%, which is generally known for AF4 experiments and which was comparable to a bovine serum albumin (MW ∼66 500) standard solution. The separation of unreacted PEG from its reaction partner, the gelatin nanoparticles, is especially important, as the nanoparticles show a certain RI response and are therefore able to disturb the PEG quantification. Due to their size and molecular weight, nanoparticles are affected by very low cross-flow intensities.39,40 After adjusting the cross-flow aiming at a good PEG peak symmetry, the reduced cross-flow therefore enables the controlled elution of the nanoparticles sufficiently separated from PEG. Figure 4 shows the successful separation of PEG and gelatin nanoparticles. Since there is only a weak RI signal arising from the gelatin nanoparticles, an overlay of UV and RI detection is shown. It has to be mentioned that the nanoparticles’ UV signal arises not only from light absorbance but also from the turbidity of the nanoparticle suspension. This has to be taken into consideration when interpreting their UV signal but does not play (48) Nektar Therapeutics. Nektar Advanced PEGylation, Catalog 2005-2006, 2006.

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

4577

Figure 4. Separation of PEG 5000, c ) 1 mg/mL, (black line) and gelatin nanoparticles (gray line) highlighted by the overlay of the concurrently recorded UV260 (gelatin nanoparticles) and RI (PEG 5000) signal.

Figure 5. Calibration curve of PEG 5000 based on the AUCs calculated from the respective fractograms.

any role in this context. Compared with the basic method shown in Figure 2, it was now modified to the point of a prolonged elution step (plus 10 min) after abandoning the cross-flow. This prolonged duration enables a complete elution of the nanoparticles from the AF4 channel subsequent to separation. To achieve a calibration curve describing the correlation of the detection signal intensity with the PEG concentration, a series of PEG-containing solutions with increasing concentrations was investigated. Plotting the resulting AUC of each peak versus the according PEG concentration, the expected linear relation could be proved and a straight calibration curve was obtained (Figure 5). After establishing the analytical basis, the PEGylation reaction was conducted with four different initial PEG concentrations “offered” to a constant amount of gelatin nanoparticles. The idea behind this was to determine a threshold above which a further PEGylation will not take place even with increasing amounts of PEG. PEG was analyzed twice to follow the course of the PEGylation reaction. A defined amount of PEG was at first determined in the absence of gelatin nanoparticles. In a second step, the same amount of PEG was incubated with a given mass of gelatin nanoparticles and the unreacted PEG remaining after incubation was measured directly in the reaction media after separation from the now PEGylated gelatin nanoparticles. Based on the calibration curve, the respective mass of unreacted PEG could be calculated from the obtained AUCs. 4578

Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

Figure 6. Fractograms of PEG 5000 before and after the PEGylation of gelatin nanoparticles. The decrease of the particular AUC reflects the degree of the PEGylation.

Figure 7. Mass of PEG 5000 bound onto 1 mg of nanoparticles in dependency of the initially added PEG.

The decrease of the AUC of the PEG detection signal caused by the consumption of PEG during the PEGylation reaction is visualized in Figure 6. Exemplarily shown for two of the investigated concentrations, the comparison of the fractograms recorded before and after PEGylation reveals this consumption. From the resulting AUCs, the particular amounts of PEG bound onto the surface of the nanoparticles were determined. The amount of bound PEG was calculated by the following equation:

mX ) (AUC0 - AUCX)m0

wjere m0 is the initially added mass of mPEG-NH2, mX is the mass of mPEG-NH2 being loaded onto the nanoparticles,AUC0 is the AUC of the RI signal being detected for the amount m0 of mPEGNH2 before incubation, and AUCX is the AUC of the RI signal being detected for the amount mX of residual free mPEG-NH2 after incubation. Figure 7 shows the mass of bound PEG plotted against the initially added PEG. From the hyperbolic shape of the resulting curve, it is obvious that the amount of PEG bound onto the surface of a given mass of gelatin nanoparticles is limited. We could determine a maximum mass of ∼0.350 mg of PEG/mg of gelatin nanoparticles that can be bound. PEG added in lower concentrations (0.125 and 0.25 mg/mg of nanoparticles) is almost completely bound, whereas the application of higher amounts of PEG did not lead to a further PEGylation of the nanoparticles (data

Figure 8. AFM analysis of plain (A1-A3) and PEGylated (B1-B3) gelatin nanoparticles. A1/B1, error channel image; A2/B2, height channel image; A3/B3, cross section of the height image (see black lines in A2/B2).

not shown). In a comparable context, Xu et al. determined an amount of 0.01-0.02 mg of PEG 5000 monomethyl ether incorporated into 1 mg of silica nanoparticles with a FT-IR-based quantification method. But, due to some technical considerations, they did not chose optimum conditions that may have led to higher degrees of PEGylation.49 Referred to the PEGylation mechanism, as it is shown in Figure 1, a limited number of monofunctionally bound glutraraldehyde is accessible for PEGylation, thus leading to the assumed threshold. Applying a validated process for the manufacturing of the gelatin nanoparticles, the number of free aldehyde groups is held constant for each nanoparticle batch. Finally, we could demonstrate the pH dependency of the PEGylation process. As the coupling of aldehydes with primary (49) Xu, H.; Yan, F.; Monson, E. E.; Kopelman, R. J. Biomed. Mater. Res., Part A 2003, 66A (4), 870-879.

amino groups demands slightly alkaline conditions, we could show a successful PEGylation in borate buffer pH 8.4. When changing the pH toward an acidic value of pH 3.0, the fractograms showed identical peaks for PEG control solutions and unreacted PEG derived from the incubation process (data not shown). These data prove the assumed influence of the pH on the coupling reaction and provide at the same time a successful negative control for the described quantification procedure. In addition, these data indirectly demonstrate the acid lability of the formed Schiff base (see Figure 1). To circumvent its hydrolysis during AF4 experiments, the pH was only slightly lowered to 7.0 compared to reaction conditions, thus enabling reproducible analysis of the formed imine. Visualization of the PEGylation Reaction via AFM. Beside the quantification of the PEGylation process, we were interested Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

4579

Table 1. DLS Size Determination Data of the Applied Nanoparticle Batches plain nanoparticles

non PEGylated nanoparticles PEGylated nanoparticles

cationized nanoparticles

size (nm)

PI

size (nm)

PI

179.7 185.2

0.041 0.072

180.4 184.5

0.064 0.097

in possible morphological changes on the surface of gelatin nanoparticles caused by their PEGylation. Therefore, AFM was applied for the qualitative analysis of PEGylated and nonPEGylated gelatin nanoparticles. In contrast to the AF4 experiments, no quantitative assumption related to AFM data was intended. Prior to measurement, the nanoparticles were fixed on the sample grid by electrostatic interactions. Since the sample grid had a negatively charged coating, the nanoparticles had to be cationized to ensure sufficient fixation. As the cationization had only to be performed prior to AFM experiments, these nanoparticles differ from the ones analyzed via AF4. However, it was assumed that the modification with cholamine, a low molecular weight substance (Mw 175.10), should not have a crucial impact on the particle’s morphology. Otherwise, it could be demonstrated that both modifications (cationization and PEGylation) can be performed on one nanoparticle batch. DLS revealed that neither cationization or PEGylation nor cationization and PEGylation led to changes in the particle homogeneity. All batches exhibited polydispersity indices below 0.1, indicating a monodispers size distribution of the respective nanoparticle populations (Table 1). The images derived from AFM analysis of the cationized versions of the non-PEGylated and the PEGylated nanoparticles are depicted in Figure 8. Overview images being employed in the amplitude mode demonstrated the good homogeneity of the applied samples (Figure 8, A1/B1). No significant differences between both batches were seen. So, previous SEM data, where those gelatin nanoparticles were assessed as homogeneous smooth spheres, were substantiated (data not shown). But, analyzing height images, PEGylated nanoparticles seemed to possess a more uneven surface (Figure 8, A2/B2). Observing the

4580 Analytical Chemistry, Vol. 79, No. 12, June 15, 2007

respective cross sections of these height images finally revealed more detailed information (Figure 8, A3/B3). Converse to the smooth surface of plain gelatin nanoparticles, PEGylated nanoparticles featured a rough surface when analyzing their height profile, which most likely indicates the presence of PEG chains on the nanoparticles’ surface as no other difference but PEGylation existed between the two investigated nanoparticle qualities. Thus, visualization by AFM analysis could contribute to the characterization of the PEGylation process of gelatin nanoparticles. CONCLUSIONS Due to its unique ability to separate suspended and dissolved analytes during one sample run, AF4 was applied in the analysis of the covalent functionalization of the surface of a colloidal drug carrier system. No further sample preparation was necessary to successfully quantify the PEGylation process of gelatin nanoparticles via an AF4 separation followed by a direct RI detection. With regard to the broad separation range AF4 spans, these results provide the basis for the analytical description of colloidal drug carrier systems by AF4 and an application of the described analytical setup in other PEGylation (e.g., the PEGylation of proteins) and PEG quantification (e.g., the release from controlled release devices) scenarios seems to be possible. AFM, as second analytical tool, demonstrated its strength in resolving nanometer-sized structures in the analysis of the nanoparticles’ surface. Especially the combination of both techniques appears to be ideal to track the PEGylation of colloidal carrier systems. The results obtained are of high quality and suggest using both methods as state-of-the-art analysis for various other colloidal systems as the number of existing techniques is rather limited. ACKNOWLEDGMENT JPK Instruments, Berlin, Germany is gratefully acknowledged for performing the AFM studies and providing us with the resulting data; especially Christian Lo¨bbe for the sample analysis.

Received for review November 12, 2006. Accepted April 11, 2007. AC062135E