Biomacromolecules 2005, 6, 341-350
341
In Vitro Gene Delivery Using Polyamidoamine Dendrimers with a Trimesyl Core† Xue-Qing Zhang,‡,§ Xu-Li Wang,‡ Shi-Wen Huang,§ Ren-Xi Zhuo,*,§ Zhi-Lan Liu,§ Hai-Quan Mao,‡,| and Kam W. Leong*,‡,⊥ The Division of Biomedical Sciences, Johns Hopkins in Singapore, 31 Biopolis Way, The Nanos, 02-01, Singapore 138669, Repuiblic of Singapore, Key Laboratory of Biomedical Polymers, College of Chemistry and Molecular Science, Wuhan University, Wuhan 430072, People’s Republic of China, Department of Biomedical Engineering, Johns Hopkins University School of Medicine, 726 Ross Building, 720 Rutland Avenue, Baltimore, Maryland 21205, and Department of Materials Science and Engineering, Johns Hopkins University, 102 Maryland Hall, 3400 N. Charles Street, Baltimore, Maryland 21218 Received August 11, 2004
Polyamidoamine (PAMAM) dendrimer represents one of the most efficient polymeric gene carriers. To investigate the effect of the core structure and generation of dendrimers on the complex formation and transfection efficiency, a series of PAMAM dendrimers with a trimesyl core (DT) at different generations (DT4 to DT8) were developed as gene carriers and compared with the PAMAM dendrimers derived from pentaerythritol (DP) and inositol (DI). The minimal generation number of DTs at which the dendrimer has enough amino group density to effectively condense DNA was higher (generation 6) than those of DPs and DIs (generation 5). DTs of generation 6 or higher condensed DNA into complexes with an average diameter ranging from 100 to 300 nm, but the 4th and 5th generations of DT (DT4 and DT5) formed only a severe aggregate with DNA. Interestingly, the DT6/pDNA complex was determined to be much smaller (100-300 nm) than those prepared with DP5 or DI5 (>600 nm) at N/P ratios higher than 15. The optimal generation numbers at which the dendrimers showed the highest transgene expression in COS-7 cells were 5 for DPs and DIs but 6 for DTs. The DT6/pDNAcomplex with smaller size mediated higher transgene expression in COS-7 cells than those prepared with DP5 or DI5. The in vitro transfection efficiency of the DT dendrimers as evaluated in HeLa cells, COS-7 cells, and primary hepatocytes decreased in the order of DT6 > DT7 > DT8 > DT5 > DT4. The transfection mediated by DT6 was significantly inhibited by bafilomycin A1. The acid-base titration curve for DT6 showed high buffer capacity in the pH range from 5.5 to 6.4 (pKa ≈ 6). This permits dendrimers to buffer the pH change in the endosomal compartment. However, the transfection efficiency mediated by DT6 decreased significantly in the presence of serum in both HeLa cells and COS-7 cells. The cytotoxicity of DTs evaluated in HeLa cells using the 3-{4,5-dimethylthiazol-2-yl}-2,5diphenyltetrazolium bromide assay showed a trend of increasing toxicity with the polymer generations. The LD50 values of DT4 through DT8 were 628, 236, 79, 82, and 77 µg/mL, respectively, which were higher than that of poly(ethyleneimide) (18 µg/mL) and poly(L-lysine) (28 µg/mL) in the same assay. With a lower cytotoxicity and versatility for chemical conjugation, these PAMAM dendrimers with a DT core warrant further investigation for nonviral gene delivery. Introduction Polyamidoamine (PAMAM) dendrimers are a class of nanoscopic, spherical, well-defined, highly branched, and monodispersed polymers that carry primary amino groups on the surface.1,2 They were first used as gene carriers in * To whom correspondence and reprint requests should be addressed. Kam W. Leong, 726 Ross Building, 720 Rutland Avenue, Baltimore, MD 21205, U.S.A. Tel: +1-410-614-3741. Fax: +1-410-955-0075. E-mail:
[email protected]. Ren-Xi Zhuo, Department of Chemistry, Wuhan University, Wuhan 430072, P. R. China. Tel & Fax: +86-27-8764-8509. E-mail:
[email protected]. † This paper was presented at the ICMAT 2003 conference, held in Singapore June 29-July 4, 2003. ‡ Johns Hopkins in Singapore. § Wuhan University. | Johns Hopkins University. ⊥ Johns Hopkins University School of Medicine.
1993 by Haensler and Szoka3 and could mediate high efficiency transfection of a variety of suspensions and adherent cultured mammalian cells. Their characteristics of precise control of structure, favorable pKa’s, and relatively low toxicity have fueled more studies on new molecular designs and the characterization of transfection efficiency both in vitro and in vivo.4-8 For example, poly(ethylene glycol) (PEG)-modified dendrimer9 and cyclodextrin-modified dendrimer10,11 have exhibited significant enhancement of gene expression with reduced cytotoxicity. In vivo toxicity study indicated no evidence of immunogenicity. Potential biological complications were observed only with highgeneration dendrimers at a high dose.12 In a murine cardiac transplantation model, dendrimer/plasmid complexes mediate widespread and prolonged gene expression in both myocytes
10.1021/bm040060n CCC: $30.25 © 2005 American Chemical Society Published on Web 12/01/2004
342
Biomacromolecules, Vol. 6, No. 1, 2005
and cells infiltrating into the graft from 7 to 28 days.13 Delivery of the IL-10 gene with these complexes also leads to a significant prolongation of graft survival. Recently it has also been shown that intratumoral delivery of dendrimer/ DNA complexes encoding the Angiostatin or tissue inhibitor of metalloproteinase-2 genes can effectively inhibit tumor growth and tumor-associated vascularization.14;15 Studies have shown that the amino group density on the dendrimer surface and size and generation of the dendrimer are important parameters that determine the gene transfer efficiency.2,16,17 The number of functional groups on the dendrimer surface, diameter, and molecular conformation are dependent on its core. PAMAM dendrimers derived from different cores, although sharing similar chemical structures, will lead to different charge densities on the surface with the same generation number. A core with three, four, and six directions, for example, in generation 6, will result in 96, 128, and 192 primary amine groups on the dendrimer surface, respectively. Previously, we have synthesized PAMAM gene carriers using pentaerythritol (DP) and inositol (DI) as cores, leading to four- and six-directional growth, respectively (data to be published). When these dendrimers grew to generation 5 (DP5 and DI5), they all showed a high transfection efficiency in cell lines as well as relatively low cytotoxicity. However, the DNA/dendrimer complexes prepared from the four- or six-armed core showed relatively large size (>600 nm) even at their respective optimal generation number and charge ratio. Several studies have shown that particle size is one of the most critical parameters affecting the transport and transfection behaviors of the complexes in vivo and in vitro,18-20 such as cellular uptake and passage through the various biological barriers. Smaller particle size (e.g., DT7 > DT8 > DT5 > DT4 (Figure 5A). In addition, transfection efficiency mediated by DT6 continued to increase with the N/P ratio, reaching a level equivalent to that of PEI at an N/P ratio of 10. At an N/P ratio of 20, dendrimer DT6-DT8/DNA complexes showed significant cytotoxicity, resulting in more than 30% decrease in cell number reflected by the measured cell protein level in comparison with the control in which cells were exposed to the same amount of the naked DNA, even though the luciferase expression continued to increase with the N/P ratio above 10 for DT6 and DT7. Transfection efficiencies mediated by DT4 and DT5 were more than 2 orders of magnitude lower than that of DT6 or DT7. This may be due to their poor DNA condensation ability and significant aggregation in a wide range of N/P ratios. Different from the DT series (the optimal transfection efficiency was obtained at the generation number of 6), both DP and DI series mediated the highest transgene expression at the 5th generation (data not shown). This coincided with the minimal generation numbers for efficient DNA binding for these three series of dendrimers. This suggests that there is a balance between the transfection activity and cytotoxicity of the dendrimer gene carriersstransfection efficiency and cytotoxicity both increase with the generation number. In this case, the optimal generation number is the minimal generation number where the dendrimer begins to bind DNA efficiently. The difference in optimal generation numbers among the three series of dendrimers is a result of the core structure (i.e., the branching number), which in turn influences the surface charge density of the dendrimer and DNA binding efficiency. Comparing the optimal carriers of the three series of dendrimers (DT6, DP5, and DI5), the DT6/DNA complex mediated the highest transgene expression with the smallest
348
Biomacromolecules, Vol. 6, No. 1, 2005
Zhang et al.
Figure 5. Transfection efficiencies of dendrimers in COS-7 cells. Cells were incubated with dendrimer/DNA complexes containing 1 µg of VR1255 (A) or 3 µg of pRE-Luciferase plasmid DNA (B) in the absence of serum at the indicated N/P ratios for 4 h at 37 °C. After 4 h, the transfection medium was replaced with fresh complete medium. Luciferase assay was performed 48 h after transfection. Data represent the mean ( standard deviation (n ) 2).
Figure 6. Transfection of rat primary hepatocytes with dendrimer/ DNA complexes. Hepatocytes were incubated with dendrimer (DT4DT8)/DNA complexes containing 3 µg of VR1255 plasmid DNA at the indicated N/P ratios for 4 h at 37 °C. After 4 h, the transfection medium was replaced with fresh medium. Luciferase assay was performed 48 h after transfection. Data represent mean ( standard deviation (n ) 2).
particle size among the complexes prepared with GT6, GP5, and GI5 as shown in Figure 5B. The transfection efficiency followed the order of DT6 > DP5 > DI5, which correlated with the order of average particle sizes at the same charge ratio of 10 (Figure 3). This result indicated that the different flexibilities of the dendrimers caused by different core structures altered the interaction between the dendrimers and DNA and subsequently influenced their transfection activity. The transgene expression mediated by dendrimers with the DT core was also evaluated in the primary rat hepatocytes (Figure 6). A similar trend was observed in hepatocytes: the optimal generation number was 6, followed by 7, 8, 5, and 4, sequentially. The highest transgene expression was mediated by DT6 at N/P ratios between 5 and 10. No obvious cytotoxicity was observed in this N/P ratio range. The transfection efficiency of dendrimers (DT6-DT8) increased with the N/P ratio but gradually attenuated at the higher N/P ratios, which is likely due to the increased cytotoxicity. Effect of Serum on Transfection Efficiency. The presence of serum is known to affect the stability of polymer/ DNA complexes. Therefore, exposure of the cells to DNApolymer complexes in vitro is usually performed with reduced serum or serum-free medium. To investigate the
effect of serum on transfection efficiency, transfection experiments were carried out in a medium containing 5, 10, 20, and 40% FBS, respectively. Figure 7 showed that the presence of serum inhibited the transgene expression significantly. The transfection efficiency decreased with the increase of the serum concentration. In the presence of 40% FBS, DT6-mediated transgene expression showed the same trend in HeLa cells and COS-7 cells. It decreased by 30fold at an N/P ratio of 10. Effect of Chloroquine and Bafilomycin A1 on the Transfection Efficiency of DT6/DNA Complexes. As reportedly previously, the high transfection efficiency of PAMAM dendrimer/DNA complexes correlates with its “proton sponge” effect of the dendrimers,3,26 similar to PEImediated transfection. The multiple tertiary amines in the dendrimers, mostly in the form of free bases at physiological pH, become protonated at the slightly acidic pH of endosome, which disrupt the endosomal vesicle by swelling. To verify the correlation of transfection efficiency dendrimers with its endosome buffering effect, the transfection experiment was performed using DT6 in the presence and absence of chloroquine or bafilomycin A1 (Figure 8), and the results were discussed in relation to the pKa value of DT6 measured by acid-base titration (Figure 9). Chloroquine is known to accumulate in the endosomal compartment, to buffer endosome acidification, and to induce osmotic swelling of the endosome, which eventually results in endosome destabilization and release of internalized polyplex.27 Bafilomycin A1 is an inhibitor of “proton pump”.28 Chloroquine treatment resulted in no change or marginal increase in transfection efficiency mediated by DT6 in HeLa and COS-7 cells. However, the addition of bafilomycin A1 resulted in about 10-fold decrease in transfection efficiency in both cells for DT6 and about 27-fold decrease for PEI. To investigate the buffer capacity of DT6, acidbase titration was performed to measure its pKa value. Figure 9 demonstrated that a relatively high buffer capacity of DT6 occurred at two pH ranges. One was from 8.7 to 9.4 corresponding to a pKa1 of about 9, and the other one was from 5.5 to 6.4 corresponding to a pKa2 of about 6. It is well-known that the endosome matures by further acidifica-
PAMAM Dendrimer Gene Carriers with Trimesyl Core
Biomacromolecules, Vol. 6, No. 1, 2005 349
Figure 7. Effect of serum on the transfection efficiency of DT6/DNA complexes. Transfection was performed with DT6/DNA complexes containing 1 µg of VR1255 plasmid DNA at different N/P ratios under various serum concentrations. Data represent mean ( standard deviation (n ) 2).
Figure 8. Effect of bafilomycin A1 on transgene expression mediated by DT6/DNA complexes. Transfection was performed with DT6/DNA complexes containing 3 µg of pRE luciferase plasmid DNA at different N/P ratios under various conditions (in the absence of 300 nM bafilomycin A1 and 100 µM chloroquine, in the presence of 300 nM bafilomycin A1, and in the presence of 100 µM chloroquine). Data represent mean ( standard deviation (n ) 2).
Figure 9. Acid-base titration curve of DT6. DT solution with 5 mM amino group concentration was titrated with 0.1 N HCl (pH 1.05). pH was recorded on a Corning pH meter 440. d[H]/dpH, which is defined as the buffer capacity, was plotted against the pH value of the solution.
tion from early endosome (pH ≈ 6) to late endosome (pH ≈ 5). The acid-base titration curve suggested that DT6 is able to buffer the endosome lumen in the weak acid environment. This result is consistent with the fact that bafilomycin A1 treatment decreased the transfection efficiency, even though to a less degree than PEI. Our data also suggested that DT6-mediated transfection did not require chloroquine assistance to escape the endolysosomal compartment. In summary, a series of PAMAM dendrimer gene carriers (DTs) was synthesized in successive generations from a DT
core (three-arm), with a defined size, molecular weight, and number of terminal amino groups. Their physicochemical properties and gene transfer activities were compared with those of dendrimers derived from DP (four-arm) and DI (sixarm). The minimal generation number of DTs where the dendrimer has enough amino group density to compact DNA effectively was higher (6) than those of DPs and DIs (5). Interestingly, DT6, DP5, and DI5 have almost the same molecular weight and amino groups, suggesting that the charge density and molecular weight of the dendrimer needs to reach a threshold to completely condense DNA. Among all three series of dendrimers at their respective optimized generation number, the DT/DNA complex showed a much smaller size and mediated the highest transfection efficiency. It highlighted that the difference in the core structure results in different flexibilities of dendrimers, which in turn influence the interaction between dendrimers and DNA and, consequently, their transfection efficiencies. Both cytotoxicity and in vitro transfection efficiency of the DT series increased with the increase of generation number and reached a balance between the cytotoxicity and transfection efficiency at the generation number of 6, at which it showed comparable transfection efficiency but lower cytotoxicity in comparison with PEI. The presence of 300 nM of bafilomycin A1 significantly inhibits the transgene
350
Biomacromolecules, Vol. 6, No. 1, 2005
expression mediated by DT6, suggesting the high transfection efficiency of DT6 might be partly due to its buffer capacity in the endosomal compartment, which was confirmed by its acid-base titration curve. However, the transfection efficiency decreased significantly in the presence of serum. Further modifications of these PAMAM dendrimers, such as conjugation of PEG to avoid the inhibition of serum or attachment of a targeting moiety, are expected to improve their applications. Acknowledgment. This work is supported by Agency for Science, Technology and Research (A*STAR) of Singapore and The Division of Biomedical Sciences, Johns Hopkins in Singapore. The authors thank Dr. Pengchi Zhang for his assistance with primary rat hepatocyte isolation. References and Notes (1) Tomalia, D. A.; et al. Polym. J. 1985, 17, 117-132. (2) Esfand, R.; Tomalia, D. A. Drug DiscoVery Today 2001, 6, 427436. (3) Haensler, J.; Szoka, F. C., Jr. Bioconjugate Chem. 1993, 4, 372379. (4) Aharoni, S. M.; Crosby, C. R.; Walsh, E. K. Macromolecules 1982, 15, 1093-1098. (5) Cheng, H.; Zhou, R.; Liu, L.; Du, B.; Zhuo, R. Genetica 2000, 108, 53-56. (6) Ohsaki, M.; Okuda, T.; Wada, A.; Hirayama, T.; Niidome, T.; Aoyagi, H. Bioconjugate Chem. 2002, 13, 510-517. (7) Sadler, K.; Tam, J. P. J. Biotechnol. 2002, 90, 195-229. (8) Zinselmeyer, B. H.; Mackay, S. P.; Schatzlein, A. G.; Uchegbu, I. F. Pharm. Res. 2002, 19, 960-967. (9) Luo, D.; Haverstick, K.; Belcheva, N.; Han, E.; Saltzman, W. M. Macromolecules 2002, 35, 3456-3462. (10) Kihara, F.; Arima, H.; Tsutsumi, T.; Hirayama, F.; Uekama, K. Bioconjugate Chem. 2002, 13, 1211-1219. (11) Arima, H.; Kihara, F.; Hirayama, F.; Uekama, K. Bioconjugate Chem. 2001, 12, 476-484.
Zhang et al. (12) Roberts, J. C.; Bhalgat, M. K.; Zera, R. T. J. Biomed. Mater. Res. 1996, 30, 53-65. (13) Qin, L.; Pahud, D. R.; Ding, Y.; Bielinska, A. U.; Kukowska-Latallo, J. F.; Baker, J. R., Jr.; Bromberg, J. S. Hum. Gene Ther. 1998, 9, 553-560. (14) Brown, M. D.; Schatzlein, A. G.; Uchegbu, I. F. Int. J. Pharm. 2001, 229, 1-21. (15) Vincent, L.; Varet, J.; Pille, J. Y.; Bompais, H.; Opolon, P.; Maksimenko, A.; Malvy, C.; Mirshahi, M.; Lu, H.; Vannier, J. P.; Soria, C.; Li, H. Int. J. Cancer 2003, 105, 419-429. (16) Malik, N.; Wiwattanapatapee, R.; Klopsch, R.; Lorenz, K.; Frey, H.; Weener, J. W.; Meijer, E. W.; Paulus, W.; Duncan, R. J. Controlled Release 2000, 65, 133-148. (17) Vlasov, G. P.; Korol’kov, V. I.; Pankova, G. A.; Tarasenko, I. I.; Baranov, A. N.; Glazkov, P. B.; Kiselev, A. V.; Stapenko, O. V.; Lesina, E. A.; Baranov, V. S. Russ. J. Bioorg. Chem. 2004, 30, 1220. (18) Bettinger, T.; Remy, J. S.; Erbacher, P. Bioconjugate Chem. 1999, 10, 558-561. (19) Bogdanenko, E. V.; Sviridov, Y. V.; Moskovtsev, A. A.; Zhdanov, R. I. Vopr. Med. Khim. 2000, 46, 226-245. (20) Kircheis, R.; Wightman, L.; Wagner, E. AdV. Drug DeliVery ReV. 2001, 53, 341-358. (21) Fuson, R. C.; McKeever, C. H. J. Am. Chem. Soc. 1940, 62, 20882091. (22) Chia, S. M.; Leong, K. W.; Li, J.; Xu, X.; Zeng, K.; Er, P. N.; Gao, S.; Yu, H. Tissue Eng. 2000, 6, 481-495. (23) Wadhwa, M. S.; Collard, W. T.; Adami, R. C.; McKenzie, D. L.; Rice, K. G. Bioconjugate Chem. 1997, 8, 81-88. (24) Hansen, M.; Nielsen, S. E.; Berg, K. J. Immunol. Methods 1989, 119, 203-210. (25) Tomalia, D. A.; Baker, H.; Dewald, J.; Hall, M.; Kallos, G.; Martin, S.; Roeck, J.; Ryder, J.; Smith, P. Macromolecules 1986, 19, 24662468. (26) De Smedt, S. C.; Demeester, J.; Hennink, W. E. Pharm. Res. 2000, 17, 113-126. (27) Pouton, C. W.; Lucas, P.; Thomas, B. J.; Uduehi, A. N.; Milroy, D. A.; Moss, S. H. J. Controlled Release 1998, 53, 289-299. (28) Kichler, A.; Leborgne, C.; Coeytaux, E.; Danos, O. J. Gene Med. 2001, 3, 135-144.
BM040060N
