In vivo gene delivery by cationic tetraamino fullerene Rui Maeda-Mamiyaa,b, Eisei Noirib,1, Hiroyuki Isobec, Waka Nakanishic, Koji Okamotob, Kent Doib, Takeshi Sugayad, Tetsuro Izumie, Tatsuya Hommaa, and Eiichi Nakamuraa,1 aDepartment of Chemistry and Exploratory Research for Advanced Technology (Japan Science and Technology Agency), University of Tokyo, 7-3-1 Hongo Bunkyo, Tokyo 113-0033, Japan bDepartment of Hemodialysis and Apheresis, University Hospital, University of Tokyo, 7-3-1 Hongo Bunkyo, Tokyo 113-8655, Japan cDepartment of Chemistry, Tohoku University, Aoba-ku, Sendai 980-8578, Japan dCMIC Co. Ltd., 7-10-4 Nishi-gotanda Shinagawa, Tokyo 141-0031, Japan and eLaboratory of Molecular Endocrinology and Metabolism, Institute for Molecular and Cellular Regulation, Gunma University, 3-39-15 Showa-machi, Maebashi 371-8512, Japan Edited by Nicholas J. Turro, Columbia University, New York, NY, and approved January 21, 2010 (received for review August 13, 2009) Application of nanotechnology to medical biology has brought remarkable success. Water-soluble fullerenes are molecules with great potential for biological use because they can endow unique characteristics of amphipathic property and form a self-assembled structure by chemical modification. Effective gene delivery in vitro with tetra(piperazino)fullerene epoxide (TPFE) and its superiority to Lipofectin have been described in a previous report. For this study, we evaluated the efficacy of in vivo gene delivery by TPFE. Delivery of enhanced green fluorescent protein gene (EGFP) by TPFE on pregnant female ICR mice showed distinct organ selectivity compared with Lipofectin moreover, higher gene expression by TPFE was found in liver and spleen, but not in the lung. No acute toxicity of TPFE was found for the liver and kidney, although Lipo- fectin significantly increased liver enzymes and blood urea nitro- gen. In fetal tissues, neither TPFE nor Lipofectin induced EGFP gene expression. Delivery of insulin 2 gene to female C57/BL6 mice increased plasma insulin levels and reduced blood glucose concen- trations, indicating the potential of TPFE-based gene delivery for clinical application. In conclusion, this study demonstrated effec- tive gene delivery in vivo for the first time using a water-soluble fullerene. carbon nanotube ��� gene therapy ��� green fluorescent protein ��� insulin 2 gene ��� toxicity Aand fullerene carbon nanocluster has unique spherical structures high hydrophobicity. These unique structures of the carbon clusters lend it several properties including photosensitiv- ity (1), redox property (2), and high chemical reactivity (3). In particular, water-soluble fullerene derivatives, which can be dis- solved easily in water by introducing proper hydrophilic residues, have received much attention for their possible biological appli- cations. Recently, introduction of amino, carboxyl, and hydroxyl residues and establishment of chemical modification methods on a fullerene structure enabled us to produce highly biocompatible water-soluble fullerenes (4). Amphipathic fullerenes, when che- mically modified with hydrophilic side chains, offer great poten- tial for gene delivery because they can form a complex with DNA effectively. Gene delivery via nonviral routes has become a powerful and popular research tool for elucidating gene structure, regulation, and function (5). Gene delivery will play a pivotal role in devel- oping new therapeutic approaches (e.g., gene therapy and DNA vaccination), which might have a great impact on the develop- ment of clinical medicine. Currently, lipid-based systems are widely used for in vitro and occasionally in vivo experiments be- cause their cationic lipid-DNA complexes can be prepared easily (6, 7). The mechanism of lipid-based DNA delivery systems is to wrap the DNA within a lipid sheath by ionic interaction between the DNA���s phosphate anion and the cationic part of the lipid. The lipid sheath is structurally similar to the cell membrane. There- fore, membrane fusion enables the DNA molecules to penetrate into the cell. To date, several reports have described effective in vivo gene delivery using liposomal gene transfer systems (8���12). However, their inefficiency, which is partly attributable to the instability in serum, and cytotoxicity should be addressed (13). Lipofectin is a representative example of widely used lipid-based transfection reagents (14). Among 182,578 hits of transfection- related data in SciFinder database, Lipofectin appears 2,085 times. Application to in vivo gene delivery is also reported (15). In this study, we compared the efficacy of gene delivery between Lipofectin and a water-soluble fullerene derivative described below. Recently we developed a DNA delivery system using fullerenes that have DNA-binding side chains. Cationic fullerene molecules such as tetraaminofullerene are capable of condensing double- strand DNA into globules smaller than 100 nm. They are there- fore penetrable into the cell (16���18). The more hydrophobic nature of fullerenes than alkyl chains in lipids appears to enable fullerenes to form a stable complex with DNA. Tetraaminofuller- ene is producible in only two steps with fullerene and piperazine derivatives. Moreover, a cationic fullerene can induce gene ex- pression by releasing DNA inside the cell because of its resistance to digestion by nucleases. In an earlier study, we reported the synthesis of tetra(piperazino)fullerene epoxide (TPFE) (Fig. 1A) and its binding abilities to DNA, where protection against endonuclease by TPFE was confirmed (17). In vitro gene delivery efficiency was examined by adjusting various conditions including the fullerene/base pair ratio, the transfection time, and the amount of plasmid DNA. In addition, TPFE showed a 4-fold increase of transfection efficiency compared to Lipofectin. Furthermore, TPFE showed reduced cytotoxicity, indicating that TPFE can be used for in vivo transfection (19, 20). To date, no report in the literature has described in vivo gene delivery using carbon nanomaterials. Because TPFE-DNA complex completely differs from Lipofectin-DNA complex in terms of size, charge, and conformational structure, it is expected that gene delivery by TPFE will have different organ affinity and might be able to overcome the problem of being trapped by the lung in the first pass (21). This study was undertaken to develop an in vivo gene transfer system using a cationic fullurene TPFE. The efficacy of gene delivery to the fetus, organ safety issues, and therapeutic applica- tion by TPFE were also evaluated. Results Size and Stability of TPFE-DNA Complex. Characterization of the TPFE-DNA complex was performed to confirm the optimal pre- Author contributions: E. Noiri, H.I., and E. Nakamura designed research R.M.-M., W.N., K.O., and T.H. performed research H.I., T.I., and E. Nakamura contributed new reagents/analytic tools R.M.-M., E. Noiri, K.O., K.D., T.S., and E. Nakamura analyzed data and R.M.-M., E. Noiri, K.D., and E. Nakamura wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. Freely available online through the PNAS open access option. 1To whom correspondence may be addressed. E-mail: noiri-tky@umin.ac.jp or nakamura@ chem.s.u-tokyo.ac.jp. This article contains supporting information online at www.pnas.org/cgi/content/full/ 0909223107/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.0909223107 PNAS ��� March 23, 2010 ��� vol. 107 ��� no. 12 ��� 5339���5344 APPLIED BIOLOGICAL SCIENCES

paration for in vivo injection. Plasmid DNA encoding enhanced green fluorescent protein (EGFP) (720 bp) was used. The size of the TPFE-DNA complex prepared with the reagent-to-base pair ratio (R) of 5, by which TPFE showed the highest efficiency on gene transfection in vitro, was measured using dynamic light scat- tering (DLS) analysis. The average size of TPFE-DNA complex was approximately 50 nm immediately after preparation it re- mained stable (less than 100 nm) up to 60 min (Fig. 1B). More- over, the TPFE-DNA complex was able to remain stable as long as 120 min when prepared with 10% bovine serum containing buffer (Fig. 1C). Gene Delivery Efficiency of TPFE-DNA Complex in Vivo. The TPFE- DNA complex, Lipofectin-DNA, and naked DNA were injected into the mouse tail vein. The in vivo biodistribution of injected DNA was evaluated by amplifying the EGFP gene encoded in the plasmid DNA. At 24 hr after injection, DNA samples were extracted from the organs of the lung, liver, kidney, and spleen. Comparable localizations of plasmid DNA delivered by TPFE and Lipofectin were found, although it is noteworthy that the amount of DNA in the TPFE group was approximately 2- to 3-fold higher than the Lipofectin group except lung (Fig. 2A). In- jected plasmid DNA was not detected in the whole blood of any group. The naked DNA group showed virtually no amplification of the EGFP gene in any organ. Gene expression of the EGFP gene was examined using real- time PCR for evaluating transfection efficacy and confocal micro- scopy image analysis for detailed localization of gene expression. In the TPFE and the Lipofectin group, mRNA expression was found in the lung, liver, and spleen at 24 hr after injection (Fig. 2B). In the TPFE group, the amount of expressed mRNA was higher than that of the Lipofectin group in the liver and spleen, but lower than that of the Lipofectin group in the lung. This tendency resembled that observed in plasmid DNA distribu- tion described above (Fig. 2A). In accordance with these results, confocal microscopy image analysis showed EGFP signals in the lung, spleen, and liver of the TPFE and the Lipofectin group. Naked DNA showed no EGFP positive signal (Fig. 3). The simi- lar observation was confirmed by immunohistochemical staining of anti-EGFP antibody (SI Text and Fig. S1). The efficacy of gene delivery to the fetus and fetal appendages was also examined. At 24 hr after injection of TPFE-DNA and Lipofectin-DNA to preg- nant mice [8 days post coitum (d.p.c.)], fetuses were removed. Plasmid DNA was found in the placenta and yolk sac. The TPFE group showed a higher number of plasmid DNA copies in placen- ta than the Lipofectin group did (Fig. 4). The mRNA of EGFP was examined using real-time PCR. The mRNA was not consis- tently detected in fetus or fetal appendage in the TPFE or Lipo- fectin group (Table 1). Evaluation of TPFE Toxicity. Acute toxicity of the TPFE and the Lipofectin gene transfer systems was evaluated using serum bio- chemical examinations, which are often used in clinical practice. Blood specimens were collected 24 hr after injections. Liver in- jury was evaluated by measuring aspartate aminotransferase (AST) and alanine aminotransferase (ALT). Increases of AST and ALTwere found only in the Lipofectin group. The Lipofectin group showed modestly but significantly higher blood urea nitro- gen (BUN) levels than the TPFE group did, possibly indicating that lipofection caused acute kidney injury (Table 2). Delivery of Insulin 2 Gene for Therapeutic Application. We evaluated the potential application of gene delivery by TPFE with mouse insulin 2 gene (Ins2) expressing plasmid. As observed for the ex- periment of EGFP gene delivery, insulin expression was detected in the lung by immunohistochemistry. Modest insulin expression was also detected in the liver and spleen (Fig. 5A). At 12 hr after Fig. 1. TPFE. (A) Structure of TPFE is shown. DLS measurements were done at 0, 10, 30, 60, and 120 min after preparation of TPFE-DNA complexes. (B) The size of TPFE-DNA complex was stable for 1 h (n �� 4). *p 0.05 versus 0 min. (C) Adding 10% serum increased the stability of the TPFE-DNA complex (n �� 3). Error bars are SEM. Fig. 2. Distribution of injected plasmid DNA and EGFP mRNA expression in each organ. (A) After 24 hr injection, plasmid DNA was detected at several organs in the TPFE and the Lipofectin group (n �� 6 for each group). (B) mRNA expression of EGFP gene was analyzed using real-time-PCR at 24 h after in- jection. mRNA was detected at the lung, liver, and spleen (n �� 6 for each group). *p 0.05 versus the Lipofectin group. Error bars are SEM. 5340 ��� www.pnas.org/cgi/doi/10.1073/pnas.0909223107 Maeda-Mamiya et al.