Bioconjugate Chem. 2005, 16, 785−792
785
PEGylation of Poly(ethylene imine) Affects Stability of Complexes with Plasmid DNA under in Vivo Conditions in a Dose-Dependent Manner after Intravenous Injection into Mice Thomas Merdan,† Klaus Kunath,† Holger Petersen,† Udo Bakowsky,† Karl Heinz Voigt,‡ Jindrich Kopecek,§ and Thomas Kissel*,† Department of Pharmaceutics and Biopharmacy, Philipps-University of Marburg, Ketzerbach 63 , D-35032 Marburg, Germany, Department of Physiology, Philipps-University of Marburg, Pilgrimstein, D-35037 Marburg, Germany, Department of Pharmaceutics and Pharmaceutical Chemistry, The University of Utah, 30 South, 2000 East, Salt Lake City, Utah 84112.. Received October 26, 2004; Revised Manuscript Received April 8, 2005
The influence of PEGylation on polyplex stability from poly(ethylene imine), PEI, and plasmid DNA was investigated both in vitro and after intravenous administration in mice. Polyplexes were characterized with respect to particle size (dynamic light scattering), zeta-potential (laser Doppler anemometry), and morphology (atomic force microscopy). Pharmacokinetics and organ accumulation of both polymers and pDNA were investigated using 125I and 32P radioactive labels, respectively. Furthermore gene expression patterns after 48 h were measured in mice. To elucidate the effect of different doses, all experiments were performed using ca. 1.5 µg and 25 µg of pDNA per mouse. Our studies demonstrated that both PEI and PEG-PEI form stable polyplexes with DNA with similar sizes of 100-130 nm. The zeta potential of PEI/pDNA polyplexes was highly positive, whereas PEGPEI/pDNA showed a neutral surface charge as expected. The pharmacokinetic and organ distribution profiles after 2 h show similarities for both PEI and pDNA blood-level time curves from polyplexes at both doses indicative for significant stability in the bloodstream. A very rapid clearance from the bloodstream was observed and as major organs of accumulation liver and spleen were identified. PEGPEI/pDNA complexes at a dose of ∼25 µg exhibit similar profiles except a significantly lower deposition in the lung. At the lower dose of ∼1.5 µg pDNA, however, for polyplexes from PEG-PEI, significant differences in blood level curves and organ accumulation of polymer and pDNA were found. In this case PEG-PEI shows a greatly enhanced circulation time in the bloodstream. By contrast, pDNA was rapidly cleared from circulation and significant amounts of radioactivity were found in the urine, suggesting a rapid degradation possibly by serum nucleases after complex separation. Regarding in vivo gene expression, no luciferase expression could be detected at ∼1.5 µg dose in any organ using both types of complexes. At 25 µg only in the case of PEI/pDNA complexes were significant levels of the reporter gene detected in lung, liver, and spleen. This coincided with high initial accumulation of pDNA complexed with PEI and a high acute in vivo toxicity. For PEG-PEI, initial accumulation was much lower and no gene expression as well as a low acute toxicity was found. In summary, our data demonstrate that PEG-PEI used in this study is not suitable for low dose gene delivery. At a higher dose of ∼25 µg, however, polyplex stability is similar to PEI/pDNA combined with a more favorable organ deposition and significantly lower acute in vivo toxicity. These findings have consequences for the design of PEG-PEI-based gene delivery systems for in vivo application.
INTRODUCTION
Nonviral gene delivery has proven its efficiency in numerous in vitro studies (1). The use of genes as therapeutic agents offers tremendous potential for the causal treatment of yet incurable diseases, especially viral infections and cancer (2). For systemic gene therapy, however, it is necessary to develop vectors with a low toxicity and sufficient stability and circulation time in the bloodstream in order to reach remote sites of the body, e.g. disseminated metastases. With cationic poly* To whom correspondence should be addressed. Tel.: (0049)6421-282-5881. Fax: (0049)-6421-282-7016. E-mail: kissel@ mailer.uni-marburg.de. † Department of Pharmaceutics and Biopharmacy, PhilippsUniversity of Marburg. ‡ Department of Physiology, Philipps-University of Marburg. § The University of Utah.
mers, such as linear poly(ethylene imine), PEI (3, 4), or methacrylates (5), in vivo transfection experiments have been reported using primarily luciferase encoding reporter genes. Also pharmacokinetic and organ distribution data have been obtained for different nonviral vectors, demonstrating rapid clearance of positively charged polyplexes from circulation (6-8). To extend the blood circulation times of nonviral vectors, hydrophilic polymers, such as poly(ethylene glycol), PEG, or poly(N-(2-hydroxypropyl)methacrylamide) (pHPMA), have been coupled to polycations (911) to achieve hydrophilic shielding (“stealth” effect) similar to PEGylated liposomes (12). PEG-modified polycations have shown promising biophysical properties after complexation with p-DNA, such as a neutral zeta potential, very low cytotoxicity, and little or no tendency for aggregation (9, 11, 13). While the stability of polyplexes under in vitro conditions is well documented (11,
10.1021/bc049743q CCC: $30.25 © 2005 American Chemical Society Published on Web 06/14/2005
786 Bioconjugate Chem., Vol. 16, No. 4, 2005
14), relatively little is known about polyplex stability in an in vivo situation. Promising results have been obtained in animal studies with loco-regional administration of PEG-modified polymer/DNA complexes (15, 16). Also repeated administration of PEG-PEI/DNA polyplexes led to a more prolonged transgene expression in the spinal cord as compared to PEI (17). By contrast, the fate of the constituents of polyplexes in the bloodstream after intravenous injection is incompletely understood, especially the influence of hydrophilic polymers, such as PEG. A comparison of poly(L-lysine) and PEGylated poly(L-lysine)/DNA complexes revealed significant DNA degradation in vivo for the PEGylated polymer indicative of polyplex disintegration (7). Further studies addressed pharmacokinetic profiles and organ distribution of either plasmid DNA or polymer after intravenous injection in mice (9, 11). However, little is known about the quantitative distribution and blood level profiles of both complex constituents determined under the same experimental conditions and to what extent polyplexes retain their integrity in the bloodstream. In this study we investigated blood level profiles and organ distributions of PEI/pDNA and PEG-PEI/pDNA polyplexes after intravenous injection into mice. To trace both complex constituents separately, two different radioactive labels for plasmid (32P) and polymers (125I) were used. To determine the influence of pDNA dose on blood level profiles and organ accumulation, we applied a total dose of 1.5 µg (0.075 mg/kg) and 25 µg (1.25 mg/kg) of plasmid DNA plus the corresponding amount of polymer per animal. Furthermore, we investigated gene expression profiles at both doses for PEI/plasmid and PEGPEI/plasmid complexes, and we compared these with accumulation data of plasmid in major organs 30 min after injection and after 2 h. EXPERIMENTAL SECTION
Materials and Animals. Endotoxin-free luciferaseencoding plasmid (pCMV-luc) was purchased from Plasmid Factory, Bielefeld, Germany. PEI 25 kDa was obtained from Sigma-Aldrich, Germany. PEI(25k)-g-PEG(2k)10 was synthesized as previously described (18). Male balb/c mice were purchased from Charles River, Germany. Preparation of Polyplexes. Polymer solutions for complex formation were prepared in 79 µL of 150 mM NaCl with 10 mM HEPES buffer pH 7.4 and 36 µL glucose 5% pH 7.4. These mixtures were added to solutions of plasmid in 115 µL glucose 5% pH 7.4. All complexes used in this study were prepared at an N/P ratio of six. Dynamic Light Scattering. Hydrodynamic diameters of polymer/pDNA polyplexes were determined by photon correlation spectroscopy. Plasmid (0.5 µg pCMVLuc) in 25 µL of glucose 5%, pH 7.4 were complexed with the appropriate amount of polymer in 25 µL of glucose 5%, pH 7.4 each, as described above. Measurements were performed on a Zetasizer 3000 HS from Malvern Instruments, Germany (10 mW HeNe laser, 633 nm). Scattering light was detected at 90° angle through a 400 µm pinhole. For data analysis, the viscosity (0.88 mPa s) and the refractive index (1.33) of distilled water at 25 °C were used. The instrument was routinely validated using Standard Reference latex particles (AZ 55 Electrophoresis Standard Kit, Malvern Instruments). Each experiment was performed in quintuplet. Measurement of Zeta Potential. Zeta-potential measurements were carried out in the standard capillary
Merdan et al.
electrophoresis cell of the Zetasizer 3000 HS from Malvern Instruments at position 17.0. Measurements were performed in glucose 5%, pH 7.4, and all experiments were performed in quintuplet. Radioactive Labeling of Polymers. Polymers were labeled employing N-succinimidyl-3-(4-hydroxy-3-[125I]iodo-phenyl)propionate (Amersham Pharmacia Biotech, Germany) according to the method of Bolton and Hunter as described earlier (9, 19). Briefly: Polymers were dissolved in 0.1 M borate buffer pH 8.5, and the Bolton Hunter reagent was dissolved in DMSO. The polymer solution was added to the Bolton Hunter reagent solution, and the reaction was carried out for 60 min at room temperature. Purification was performed on a Sephadex G-25 column (PD10, Pharmacia, Germany), using an elution buffer containing 150 mM NaCl and 10 mM HEPES at pH 7.4. Radioactive Labeling of DNA. Plasmid (pCMV-Luc) was radioactively labeled by incorporation of 32P-dCTP (Readivue, Amersham Pharmacia, Germany) using a nick translation kit (Amersham Pharmacia, Germany) following a protocol provided by the manufacturer. Unincorporated nucleotides were carefully removed using Autoseq spin columns containing Sephadex G50 (Amersham Pharmacia, Germany) in two subsequent purification runs. Plasmid purity was verified via size exclusion chromatography using a PD-10 column and via ultracentrifugation using Microcon 10 spin columns (Amicon) via verification of absence of low molecular weight radioactive compounds. No significant amounts of free 32 P-dCTP were detected. Pharmacokinetic Analysis and Organ Distribution. All animal experiments followed the “Principles of Laboratory Animal Care” (NIH publication #85-23, revised 1985) and were approved by an external review committee for laboratory animal care. Male balb/c mice with a body weight of 20-25 g were anaesthetized using Ketamine (Ketavet, Pharmacia & Upjohn, Germany) and Xylazine (Rompun, Bayer AG, Germany). Complexes of either 125I-poylmer/plasmid or polymer/32P-plasmid were injected as a bolus of approximately 100µL through the jugular vein. Two different doses, 1.5 µg of pDNA and 25 µg of pDNA, complexed with the corresponding amount of polymer to give a nitrogen-to-phosphate ratio of 6 were administered. Blood samples were obtained via a catheter in the common carotid artery, and urine was sampled by flushing the bladder with sodium chloride solution through a two-way catheter. After 30 or 120 min, mice were sacrificed and organs (liver, kidneys, lungs, spleen, cortex, heart, fatty tissue, vena jugularis, and cortex) were weighed and assayed for radioactivity. Radioactivity of the 125I-polymer was measured on a 1277 Gammamaster (Perkin-Elmer Wallac, Germany). To assess the radioactivity from 32P-pCMV-Luc, organs and blood samples were dissolved in 1 mL of Soluene 350 (Amersham Pharmacia, Germany). Subsequently, 200 µL of 30% sodium peroxide was incubated with the mixture for 30 min and then added to 15 mL of HionicFluor (Perkin-Elmer, Germany). Measurements were performed using a TriCarb liquid scintillation counter (Perkin-Elmer, Germany) with a counting time of 15 min and 1 min precount delay. Measurements of aliquots of complex solutions were used for both tracers to determine the injected dose of radioactivity. All experiments at low dose were performed in quadruplet; those at high dose were performed using eight animals. In the case of PEI/pDNA complexes at both doses and PEG-PEI/pDNA complexes at ∼25 µg dose, concentration-time curves from polymer and pDNA were fitted
Stability of PEG−PEI Polyplexes in Vitro and in Vivo
Bioconjugate Chem., Vol. 16, No. 4, 2005 787
to a two-compartmental model with the Software Kinetica 1.1 from Simed. The model used was C(t) ) Ae-Rt + Be-βt, and the weighting applied was 1/(ccalcd)2. Polymer concentrations in the samples were then calculated as percent of injected dose (%ID) or %ID/mL, respectively. Unpaired t-test was performed using Microcal Origin version 6.0 to compare blood levels of different polymers at corresponding time points. Differences were considered significant if two-tail P e 0.05. Luciferase Expression in Vivo. Mice were anaesthetized as described above, and complexes containing either 1.5 or 25 µg of pCMV-luc were injected through the jugular vein in the same volume as described above. After 48 h mice were sacrificed and subsequently organs were removed. Liver, lung, spleen, kidneys, vena jugularis, cortex, and heart were dissolved in Lysis Buffer provided by Promega, Germany, and after centrifugation an aliquot of the supernatant was assayed for luciferase using a commercial kit and photon counting with a luminometer (Sirius Berthold, Germany). In vivo gene expression studies were performed in quadruplet. RESULTS AND DISCUSSION
Size and Zeta Potential Measurements. At an N/P ratio of 6, both PEI 25 kDa and PEI(25k)-g-PEG(2k)10 were capable of forming complexes with plasmid DNA. Polyplexes prepared with PEI 25 kDa displayed a size of 106 ((7) nm and a zeta potential of +26 ((2) mV. By contrast, PEG-PEI/DNA complexes reached a size of approximately 127 ((6) nm and a neutral zeta potential (0.5 ( 0.9 mV), indicating an efficient shielding of the cationic charge by grafted linear PEG chains. Compared to measurements obtained in 150 mM sodium chloride (20), complexes were significantly smaller when prepared in 5% glucose. Atomic Force Microscopy (AFM). AFM images shown in Figure 1 as three-dimensional plots illustrate that both PEI and PEG-PEI formed defined, spherical complexes with plasmid DNA. Shape and size for both polyplex types appear to be similar and are in good agreement with previously published data (18, 20). However, PEI/pDNA complexes seemed to be more uniform, more spherical, more compact, and smoother. PEG-PEI/DNA complexes were found to be flatter, and they showed a more irregular shape, suggesting that grafted PEG chains affected condensation of pDNA No free or only loosely complexed pDNA could be observed as described in earlier studies (18). Furthermore, DNA binding studies have been performed which indicated an efficient complexation of pDNA by PEI(25k)-g-PEG(2k)10 (20). These observations suggest that these polyplexes could be suitable for in vivo evaluation. Pharmacokinetic Analysis and Organ Distribution. Blood level profiles of polymers and pDNA are depicted in Figures 2-7. At a dose of approximately 1.5 µg of pDNA, none of the animals died during the experiments and no signs of acute toxicity were observed. Also at the higher dose of approximately 25 µg of pDNA, in the case of PEG-PEI/DNA complexes all animals survived without any signs of acute toxicity. However, after administration of unmodified PEI/pDNA complexes at a dose of 25µg of pDNA, two out of eight animals died within 30 min after injection, confirming previously published observations (21). PEI/plasmid Polyplexes at Low Dose (Figure 2). Pharmacokinetic profiles obtained with PEI and plasmid DNA at a 1.5 µg dose displayed a very rapid clearance from the bloodstream. Curves fitted to a biexponential
Figure 1. Complexes of PEI/DNA (A) and PEG-PEI/DNA (B) visualized by atomic force microscopy at N/P ) 6 in 5% glucose. Images display defined structures of both complex types, and no free DNA could be observed in any image. Some larger aggregates are visible in both preparations.
disposition equation by nonlinear curve fitting (Figure 2A) showed very steep alpha phases with a t1/2 alpha of 4.9 min for the polymer and 3.4 min for pDNA (see Table 1). Subsequent beta elimination phases were very flat with t1/2 beta of 120.0 and 132.6 min, respectively. After 30 min, less than 3% of the injected dose (polymer, as well as pDNA) per milliliter blood remained in blood circulation, approaching levels of approximately 1% ID/ml blood after 2 h. Comparison of pharmacokinetic data from plasmid DNA and PEI at 1.5 µg (Table 1) revealed very similar behavior especially within the first 30 min suggesting that PEI/DNA polyplexes remained intact to a substantial degree during blood circulation. The organ distribution of both PEI and pDNA at a 1.5 µg dose reached similar levels of radioactivity after 2 h in major organs, such as liver, kidney and lung (Figure 2B). These similarities provided additional evidence for polyplex stability in the bloodstream with concomitant uptake of both PEI and pDNA into tissues. Highest levels for polymer and pDNA were detected in the liver. Uptake probably occurred after opsonization of polyplexes as suggested by Planck et al. (22) with a subsequent rapid capture by mononuclear phagocytic cells. Organ deposition in the spleen showed approximately 3-fold higher accumulation of 125I compared to 32P. This fairly high accumulation in the spleen has been reported earlier for PEI 25 kDa polymer at a low dose (9) and was possibly caused by free polymer present at the N/P ratio used (23) or a minor degree of polyplex separation in the bloodstream.
788 Bioconjugate Chem., Vol. 16, No. 4, 2005
Merdan et al.
Figure 2. Blood level curves and organ accumulation after 2 h of PEI/pDNA at ∼1.5 µg of pCMV-luc per mouse. A very rapid clearance of polymers and DNA from the bloodstream can be observed for polymer and pDNA.
Significant levels of radioactivity from 32P were detected in the urine, in contrast to 125I from the polymer. This difference most likely occurred as a result of polyplex separation followed by rapid pDNA degradation as described earlier (24). Uptake into the lungs is fairly low at 1.5 µg dose after 2 h (Figure 2B) and even at a 30 min no elevation in radioactivity were observed (data not shown). Due to the low dose formation of aggregates with e.g. erythrocytes was possibly reduced (18). Other tissues, such as heart, cortex, fatty tissue or the injection site in the jugular vein, did not exhibit significant levels of radioactivity from either polymer or pDNA after application of both PEI/DNA complexes (
