HVJ (Sendai virus)-cationic liposomes - Semantic Scholar

Gene Therapy (1997) 4, 631–638  1997 Stockton Press All rights reserved 0969-7128/97 $12.00

HVJ (Sendai virus)-cationic liposomes: a novel and potentially effective liposome-mediated technique for gene transfer to the airway epithelium Y Yonemitsu1,2, Y Kaneda3, A Muraishi4, T Yoshizumi1,2, K Sugimachi 2 and K Sueishi1 Departments of 1Pathology I and 2Surgery II, Faculty of Medicine, Kyushu University, Fukuoka; 3Institute for Molecular and Cellular Biology, Osaka University; and 4Third Department of Internal Medicine, Kurume University School of Medicine, Kurume, Japan

We designed a novel technique for targeted gene transfer into the airway epithelium. This was constructed using multilamellar cationic liposomes, containing N-(atrimethylammonioacetyl)-didodecyl-D-glutamate chloride, phosphatidylcholine and cholesterol and fused with haemagglutinating virus of Japan (HVJ), namely HVJ cationic liposomes. Single aerosol delivery of this novel vector to the airway of rats led to a highly efficient and widespread transduction of fluorescein isothiocyanate-labeled oligonucleotides or lacZ gene into the bronchial epithelium and alveolar macrophages, but not into the alveolar epithelium. The efficiency of gene transfer to the airway epithelium with a single administration of the lacZ gene was about 47.6%

in the trachea, 39.0% in the bronchi and proximal bronchioli, and 2.9% in the terminal bronchioli, respectively (mean value, n = 6). Expression level of the luciferase gene delivered with this novel system was much higher than that without HVJ, in both the trachea and lung tissue. Two pretreatment HVJ-cationic liposome vehicles every other week resulted in minimal inflammatory infiltration in the subepithelial layer with no significant reduction in efficiency of the following gene transfer. We propose that this novel HVJ cationic liposome-mediated gene transfer system may be suitable for clinical gene therapy to treat subjects with lethal lung diseases such as cystic fibrosis.

Keywords: HVJ-cationic liposomes; gene therapy; airway epithelium; cystic fibrosis

Introduction Despite the recent enthusiasm for gene transfer as a therapeutic modality for lethal lung diseases, an appropriate vector for clinical application has been awaited.1 Although a replication-deficient adenoviral vector can achieve highly efficient gene delivery into the bronchial tree, pathogenic immunity induces a severe inflammatory reaction when delivering a gene to the airway in vivo.2 Cationic small lamellar liposome–DNA complexes have also been used for gene transfer into the lung.3 Although this type of gene delivery involved no apparent inflammatory reaction due to vector particles, the low efficiency of gene transduction remains a critical problem for clinical use. Caplen et al4 reported that DCChol/DOPE liposome-mediated cystic fibrosis transmembrane conductance regulator (CFTR) gene transfer to the nasal epithelium of cystic fibrosis patients resulted in approximately 20% correction of the chloride ion transport abnormality. Any method that offered enhanced gene transfer might be expected to result in a more efficient correction of this defect, therefore, novel and more effective viral or nonviral vector constructs need to be established.5,6 Correspondence: Y Yonemitsu, Departments of Pathology I and Surgery II, Faculty of Medicine, Kyushu University 60, 3-1-1 Maidashi, Higashiku, Fukuoka 812-82, Japan Received 20 September 1996; accepted 11 April 1997

Our group reported that haemagglutinating virus of Japan (HVJ) liposomes could achieve efficient in vivo gene transfer into the somatic cells in several organs.7–10 HVJ is a nonharmful virus and will have no untoward effects on humans. The features of this in vivo gene delivery system are as follows: (1) the fusogenic activity of HVJ can carry the plasmid DNA directly into the cytoplasm of target cells, even in the G0 static state; and (2) use of high mobility group 1 (HMG-1) protein brings the transferred nucleic acid promptly to the nuclei of the target cells and stabilizes the exogenous genes.8 Importantly, this is a nonharmful virus for humans and the HVJ liposomes are prepared after ultraviolet irradiation of the virus. These features of HVJ liposomes suggest potential usefulness for human gene therapy, however, this vector system also has problems. There is a relatively low level and short duration of gene expression when it is used for some targeted organs, or with some delivery techniques. This relatively low gene expression level is probably due to low efficiency of attachment to the target cells and gene entrapment into the conventional negative-charged liposomes (5–10% of applied DNA, unpublished data). Our preliminary study revealed that this vector system was ineffective for in vivo gene transfer to lung airway epithelium by noninvasive aerosol delivery, while it is possible to transduce the gene to a small area with a patchy distribution by invasive intratracheal administration (unpublished data). To achieve a more effective gene delivery into the air-

HVJ-cationic liposome-mediated lung gene transfer Y Yonemitsu et al

632

way epithelium using the HVJ-based vector system, we investigated the effect of positively charged liposomes fused with HVJ, namely HVJ cationic liposomes. This method using cationic large multilamellar vesicles differs definitively from the lipofectin method,11 which makes use of small unilamellar vesicles. In the present study, we used N-(a-trimethylammonioacetyl)-didodecyl-dglutamate chloride (TMAG), which is minimally cytotoxic. 12 We assessed (1) quantification of enhanced gene trapping into liposome particles due to TMAG cationic lipid modification; (2) certification of gene transduced cell localization and gene transfer efficiency by single administration using the markers fluorescein isothiocyanate (FITC)-labeled oligonucleotides and the lacZ gene; (3) quantification of enhanced gene expression in vivo in trachea and lung using the luciferase gene; and (4) the effect of pretreatment of vector particles for gene transfer efficiency and the histological inflammatory reaction by repeat transduction. We then discussed the potential usefulness of this novel technique for gene therapy against lethal lung diseases such as cystic fibrosis (CF).

Results Cationic lipid modification enhances gene entrapping A positively charged multilamellar construct of liposomes would be expected to provide advantages for the HVJ liposome-mediated gene transfer system. First, a larger amount of entrapped plasmid DNA would be transferred into cells, and this may result in a stronger gene expression than seen with conventional HVJ liposomes. Cationic modification would increase interactions between liposomes and the negatively charged HVJ particles, and this might reduce the number of viral particles required, without reducing the fusion activity of the HVJ liposomes to the target cells. Positively charged HVJ cationic liposomes could interact with the negatively charged cellular surface of target cells, and this should also lead to a higher recombinant gene expression. To evaluate the electric charge of the materials, we used the low concentrate agarose gel electrophoresis technique for the charge analysis of liposomes; this is usually used for DNA analysis. Low concentrate agarose gel electrophoresis analysis revealed that all conventional liposomes constructed with phosphatidylserine (PS), phosphatidylcholine (PC) and cholesterol (Chol), HVJfused conventional liposomes and HVJ viral particles were negatively charged, as a dome-like appearance of mobility shifts was observed (Figure 1a, left panel). On the other hand, newly formed cationic liposomes containing TMAG, and HVJ cationic liposomes were charged positively (Figure 1a, right panel). The electrophoresed locations of TMAG liposomes were decreased in an almost dose-dependent manner with the amount of fused HVJ (Figure 1a, right panel). We then quantified the amount of pSV-b galactosidase reporter plasmid (6.8 kb) entrapment by conventional liposomes or cationic liposomes (Figure 1b). The quantification study revealed that TMAG liposomes could entrap about five-fold plasmid DNA in comparison with that of the conventional method (63.1 ± 4.23 mg/mg lipid versus 13.0 ± 2.23 mg/mg lipid, n = 3, respectively, P , 0.001).

Figure 1 (a) Analyses of the electric charge of liposomes, with or without haemagglutinating virus of Japan (HVJ) using low concentrated agarose gel (0.5%). −, negative pole; +, positive pole, respectively. The left panel shows electrophoresed locations of conventional liposomes containing phosphatidylserine (PS, left lane), of HVJ-fused PS liposomes (middle lane) and of HVJ viral particle (right lane). All components, visible dome-like appearance, moved to a positive pole indicating that they were negatively charged as in the case of the two-separated marker dye. The right panel shows electrophoresed locations of cationic liposomes containing TMAG, with or without HVJ. The rate of electrophoresis was decreased in an almost dose-dependent manner, but all located at the negative pole, indicating that they were positively charged. (b) Quantitative analysis for efficiency of entrapped plasmid DNA by PS liposomes or TMAG-cationic liposomes. pSV-b galactosidase plasmid (200 mg) was applied to 10 mg of dehydrated liposome film, and liposome solution was treated with 200 U of DNase I in the presence of 5 mmol MgCl2 , overnight at 37°C to digest the untrapped plasmid. After inactivation of DNase I using 0.1% diethyl pyrocarbonate, liposomes were digested with 0.1% Triton-X. The plasmid DNA was collected and quantified spectrophotometrically. These experiments were performed three times.

Cationic lipid modification enhances gene expression efficiency We then administered HVJ cationic liposomes with reporter gene encoding luciferase to the airway of 600– 650 g adult rats, as an aerosol, and we evaluated the levels of gene expression. In the tracheae and right lower lobes exposed to HVJ liposomes with 2.5 mg of lipid with 3750 HAU of HVJ containing pGL2-luciferase reporter plasmid, or with HVJ cationic liposomes without DNA, we detected no significant recombinant luciferase activity, while cationic multilamellar liposome fused with HVJ achieved a higher level of luciferase activity than


that of cationic liposomes without HVJ (1913 ± 610.0 versus 756.7 ± 190.5 c.p.m./mg, n = 5, respectively, P , 0.01) (Figure 2). Furthermore, significant luciferase activity was detected only in the right lower lobes treated with HVJ cationic liposomes with pGL2-luciferase (24.6 ± 4.3 c.p.m./mg, n = 5), except for one of five rats treated with cationic liposome only. The value of luciferase activity in lung tissue seems to be much lower than in the trachea. This major difference is attributed to the considerable amounts of proteins in the lung tissue.

Gene transduction efficiency by HVJ cationic liposomes in the bronchio-bronchilar tree Next, 1 ml of HVJ-cationic liposomes containing 2.5 mg of lipid, 3750 HAU of HVJ and 2 mmol of FITC-labeled oligonucleotides was nebulized as an aerosol to the airway of three rats. The sections demonstrated frequent nuclear localizing fluorescence signals from airway epithelial cells 4 h after aerosol delivery (Figure 3a and b). When HVJ cationic liposomes were used for in vivo gene transfer of lacZ gene to six rats, X-gal-stained cells were widespread and clearly evident on the luminal surface of the trachea, bronchi and bronchioli (Figure 3c–f), while there was no apparent blue stain in bronchiopulmonary tissue exposed to HVJ cationic liposome without DNA (Figure 3g). Cryostat sections also exhibited a frequent

distribution of lacZ-expressing bronchial epithelial cells (Figure 3e). However, there was no evident positive signal in alveolar epithelial cells, except for alveolar macrophages (Figure 3f).

Absence of any significant reduction of gene transfer efficiency and of severe inflammatory response by repeat administration Six rats were used for histological assessment of inflammatory reactions, and the left lung tissues of four animals were evaluated for gene transfer efficiency by repeated administration of HVJ cationic liposomes, in comparison with single treatment. Repeated treatment is a sine qua non due to transient gene expression7–9 and vectorinduced inflammation and reduction of gene transfer efficiency is the main obstacle for clinical application using adenoviral vectors. Sections of the lung which had been exposed three times every other week to HVJ cationic liposomes showed minimal mononuclear cell infiltration around a few bronchi (Figure 4a), however, neither apparent severe inflammatory reaction nor fibrosis was evident in the interstitium and alveolar spaces (Figure 4a). Interestingly, X-gal stain was weak or negative in epithelial cells with chronic inflammatory infiltrates (Figure 4b panel b), while X-gal-positive cells were frequent in the bronchial epithelium without

Figure 2 Measurements of in vivo luciferase activity in gene delivered trachea or lung. Closed triangles indicate each value of recombinant luciferase activity. The value under 0.001 c.p.m. was determined ‘indetectable’. One milliliter of liposome solution (containing 15 mg of DNA and 3750 HAU of HVJ particles) was delivered as an aerosol, using a compressor type nebulizer. This type of nebulizer supplies aerosol 3–15 mm in diameter. The tissue was minced in lysis buffer and the luciferase activity was measured using a luminometer. Protein concentration was also photometrically measured by Bio-Rad Protein Assay System, and all data were expressed as luciferase activity per milligram protein. All values are expressed as a mean ± s.d. Oneway analysis of variance was used to determine the significant differences in each group, and P , 0.05 was considered significant.

633


634

inflammatory infiltrate at the subepithelial layer (Figure 4b panels a and c), similar to that seen in the case of a single exposure. These findings suggest that cellular injury associated with mononuclear cell infiltration reduced transgene expression, but that vector-induced inflammatory reaction seems to be rare. We also noted the X-gal-positive epithelial cell ratio in the trachea, bronchi and bronchioli, and terminal bronchioli in cases of both single (n = 6) and repeated (n = 4) exposure of HVJ cationic liposomes with 15 mg plasmid DNA (Figure 5). To evaluate the immunizing property of this vector in the rat and to avoid overestimation due to consistent gene expression seen with previously administered doses, we planned to expose the rats used on the previous two occasions with an empty vector and for the third exposure to use HVJ cationic liposomes containing the lacZ gene. Tissue sections in the case of a single administration of HVJ cationic liposomes revealed the lacZ gene transfer efficiency to be quantified as 47.6 ± 15.1% in the trachea, 39.0 ± 9.5% in the bronchioli (n = 6, respectively), while the X-gal positive rate was markedly decreased to 2.9 ± 2.1% in the terminal bronchioli (n = 6, P , 0.01). On the other hand, repeated administration of HVJ cationic liposomes did not result in any significant reduction of gene transfer efficiency in each portion, estimated as 43.6 ± 11.5% in the trachea, 38.4 ± 10.1% in the bronchioli, and 4.6 ± 4.4% in the terminal bronchioli (n = 4, respectively).

Discussion We obtained evidence of repeated and effective in vivo gene transfer to the airway epithelium using HVJ cationic liposomes, hence, the potential usefulness of this novel gene transfer technique, for CF gene therapy. Regarding the efficacy of gene transduction in vivo, McLachlan et al13 used a large amount of DNA (266 mg) and repeat transfections (five times) using cationic liposome (DOTAP)–DNA complex were required to achieve sufficient gene expression in vivo, even to small mice, while our gene transfer method requires less than one tenth (15 mg) of DNA per animal even when we used adult rats which had a ten-fold greater weight (600–650 g). Although it is difficult to compare these experiments done under different conditions, gene transduction efficiency of HVJ cationic liposomes may be roughly estimated as 100-fold that of the small lamellar cationic liposome–DNA complex method. Regarding the efficiency of gene transduction of our HVJ cationic liposomes into the airway, an additional point needs attention. In the present system, we can not obtain the HVJ cationic liposome particles without free DNA remaining or DNA attached to particles. Hence, the gene transfection efficiency of our present system contains all states of DNA. However, this may present no great problem since in our preliminary work we detected no significant gene expression of lacZ and luciferase when these were delivered as a naked plasmid outside HVJ cationic liposomes containing BSS in the internal space (unpublished data). The exogenous gene expression level, which is necessary to support the CFTR function in vivo, including electrophysiological correction, as shown by Caplen et al4 is not clear, and has to be clarified using CFTR mutant mice treated with the HVJ cationic liposomes. Induction of an inflammatory reaction is a critical

problem using adenovirus in CF gene therapy. We demonstrated that interval treatment every other week resulted in minimal mononuclear cell infiltration and no significant reduction in gene transfer efficiency. These findings suggest that the host immune response against the HVJ surface protein is relatively low in rats, which are the pathogenic host of HVJ, and a reduction in immune response against HVJ may occur in primates. Moreover, in our system HVJ was inactivated and the viral genome was destroyed by ultraviolet irradiation; hence, the viral protein would be represented on the cellular surface just after gene transfer. Viral genes could not be expressed and the viral proteins would be degraded immediately, unlike the case with an adenoviral vector. On the other hand, gene-transferred cells using an adenoviral vector express viral proteins continuously, which can be a target of cytotoxic T lymphocytes (CTL).2 In the case of the HVJ-based gene transfer system, there have been few reports indicating its pathogenic potential. Therefore, several studies have been focusing on the precise mechanisms of low immunity of HVJ liposome particles. For example, direct and repeat intratracheal administration of HVJ liposomes has not resulted in interruption of gene expression and in inflammatory reaction, despite the elevation of titers of antiHVJ protein antibody (Yoshida M et al, personal communication). Another group has shown that the ultimately low level of CTL induction was recognized with four to six times repeated administrations of HVJ liposomes to the liver via the portal vein (Hirano Y et al, personal communication). In addition, it is possible that a multilamellar structure has additional advantage in comparison with small lamellar liposome–DNA complexes exposing naked DNA, which could be a target of antiDNA antibodies. These features of the HVJ-mediated gene transfer system suggest its safety in its application for clinical use. However, extensive toxicity studies in animals, such as nonhuman primates, are needed before this novel vector can be considered for clinical trials. We also emphasize that this novel system is a highly efficient transduction technique for cell placement in vivo, and has no limitations for a gene expression cassette in liposome particles. This is also an advantage of the liposome-mediated gene transfer method in comparison with viral vectors. Replication-deficient viral vectors have strict limitations regarding recombinant gene expression. Also, the promoter activity of the long terminal repeats of the retrovirus frequently interrupts the inserted foreign promoter, and adenoviral DNA carrying a recombinant

Figure 3 In vivo reporter gene transfer to rat lung tissue. Three, six, and three male adult rats were used for the FITC-oligonucleotides, lacZ and vehicle transfer experiments, respectively. (a, b) Fluorescence microscopic examination for fluorescein isothiocyanate-labeled oligonucleotides transferred airway epithelium of the trachea 5 mm above the carina tracheae. Bright view (a) indicated epithelial layer (epi) and submucosal layer (sm) of the trachea, and dark view (b) demonstrates strong and frequent fluorescence signal in nuclei of the epithelial cells. (c–g) Dissecting (c, d) and light (e, f) microscopic findings of the lacZ gene-transferred and X-galstained lung. Intense blue signal is prominent in the bronchial epithelial layer (c, d), but not evident in pulmonary artery (d; PA) or pulmonary vein (d; PV). There was no apparent X-gal-positive signal in the alveolar epithelium, while bronchial epithelium (e) and alveolar macrophages frequently exhibited an intense blue spot (f). The lung tissue treated HVJ cationic liposomes without plasmid DNA demonstrated no apparent positive blue signals (g).


635


636

Figure 4 Histopathological findings in the rat lung exposed to repeated treatment with HVJ cationic liposomes containing lacZ expression plasmid vector (pSV-b galactosidase). (a) Treatment three times every other week and X-gal-reacted lung tissue (hematoxylin–eosin staining). No inflammation was apparent in the histological sections. (b) High powered view of a demonstrating slight inflammatory cell infiltration, but which was rarely recognized. Morphologically intact cuboidal or tall columnar epithelial cells without mononuclear cell infiltrate in the submucosal layer frequently demonstrated Xgal-positive blue signal (panels a and c; asterisks), but flattened epithelial cells which suggest injury with submucosal lymphocytic infiltrate, rarely express b-galactosidase activity (panel b; asterisk).

gene cannot replicate in host cells. On the other hand, a liposome-based vector system can be used in combination with other gene expression systems, such viral chromosomes as Epstein–Barr virus (EBV) replicon or the genomic integration system of adeno-associated virus.14 We confirmed that the EBV replicon vector co-transfected with EB nuclear antigen-1 (EBNA-1) prolonged gene expression without significant self-replication at least over 1 month in adult rat liver (unpublished data), while recombinant gene expression delivered with a nonreplicable plasmid decreased gradually and was indetectable in 2 weeks.10 These features of HVJ cationic liposomes

may be expected to achieve a relatively stable or more long-term gene expression in vivo. In summary, we demonstrated a repeated and highly efficient reporter gene transfer into the rat airway epithelium using aerosol delivery of the HVJ-based vector construct system. This may possibly be a breakthrough for gene therapy strategy for CF patients.

Materials and methods Materials and animals TMAG is a positively charged lipid obtained from Sogo Pharmaceutical, Tokyo, Japan. Phosphatidylcholine (PC)


Efficiency of DNA entrapment pSV-b galactosidase (200 mg) was entrapped by liposomes of each type, in the same manner as described above. Two hundred units of DNase I (Takara Shuzo, Kyoto, Japan) was added to DNA–liposome complex solution, in the presence of 5 mmol MgCl2 and this solution was incubated overnight at 37°C. Then, DNase I was inactivated with diethyl pyrocarbonate and membranes of liposomes were digested with 0.1% Triton-X. The plasmid DNA was collected after phenol–chloroform– isoamylalcohol treatment and was spectrophotometrically quantified.

Figure 5 Bar graphs demonstrating X-gal-positive cell ratio in the trachea, bronchioli and terminal bronchioli of rats treated with single or three repeated (empty vector; twice and the final; lacZ) administrations. Soon after killing, blood was flushed out with saline via right ventricular cannulation, followed by intratracheal administration of 2% paraformaldehyde with 0.25% glutaraldehyde in 0.1 m phosphate buffered saline for 10 min. The left lung tissue was reacted in X-gal solution to detect bgalactosidase activity in situ. X-gal-positive epithelial cells of two tracheae, four bronchioli, and four terminal bronchioli in the 5 mm sections per animal were noted. No significant reduction in gene transfer efficiency was observed in repeatedly treated animals, while the X-gal-positive ratio was markedly decreased in the peripheral airways in both groups (P , 0.01). All values are expressed as a mean ± s.d. One-way analysis of variance was used to determine the significant differences in each group, and P , 0.05 was considered significant.

and cholesterol (Chol) were obtained from Sigma, St Louis, MO, USA. pSV-b galactosidase and pGL2-luciferase control plasmids (Promega, Madison, WI, USA) were used as reporter genes. Male WKA rats (600–650 g, purchased from Kyudo, Tosu, Japan) were used for the in vivo gene transfer study. For the animal experiments we followed The Law (No. 105) and Notification (No. 6) of the Government and the Principles of Laboratory Animal Care and the Guide for the Care and Use of Laboratory Animals (publication No. NIH 80–23, revised 1985).15,16

Liposome preparation Preparation of HVJ cationic liposomes was almost the same as for conventional HVJ liposomes7–9 but with minor modification. Briefly, 10 mg of mixed lipids (TMAG:PC:Chol = 1:4.8:2) was dried by reversed phase evaporation and propagated with DNA (200 mg) HMG1 (64 mg) complex in balanced salt saline (BSS; 140 mmol/l NaCl, 5.4 mmol/l KCl, 10 mmol/l Tris-HCl, pH 7.6) solution. After 8 × 30 s vortexing and forming multilamellar liposomes, 15 000 haemagglutinating units (HAU) of purified HVJ (Z strain) and 500 ml of 60% sucrose were added to the solution which had been incubated at 37°C for 1 h. This HVJ cationic liposome solution was used for gene transfer with addition of CaCl 2 (final concentration; 2 mmol). Gel electrophoresis of liposomes for charge analysis HVJ (1000 HAU) and several liposomes (0.25 mg) were electrophoresed on 0.5% agarose gel with 2 ml of negatively charged DNA marker dye (0.25% bromophenol blue, 0.25% xylene cyanol, and 30% glycerol) for charge analysis of each liposome.

In vivo gene transfer and tissue harvest procedures Three and six rats were used for histological analysis by FITC-labeled oligonucleotide and lacZ gene transfer by single administration, respectively. Six and four rats given three administrations were also used for histological evaluation and lacZ expression efficiency, respectively. One milliliter of HVJ cationic liposome containing 2.5 mg lipids and 3750 HAU of HVJ, HVJ liposome or cationic liposome without HVJ solutions were delivered as an aerosol via the airway using the clinically available compressor type nebulizer (type NE-C11; Omron Field Engineering, Sapporo, Japan). These animals were individually nebulized in the biohazard draft chamber. Two days after gene delivery, the rats underwent thoracotomy, whole blood was flushed out with phosphatebuffered saline via the right cardiac ventricular and tracheal cannulation and the whole lung was fixed with ice-cooled 2% paraformaldehyde with 0.25% glutaraldehyde for 10 min for X-gal staining or removed immediately for measurements of luciferase activity. After harvest, the lacZ gene delivered lung was cut into 5-mm sections and placed in X-gal solution (5 mm potassium ferrous cyanide, 5 mm ferric cyanide, 2 mm magnesium chloride, 1 mg/ml 5-bromo-4-chloro-3-indolyl-b-dgalactopyranoside) for 6 h at room temperature. The Xgal-stained tissue sections, grossly examined and photographed under the Zeiss Stemi 2000-C dissecting microscope (Zeiss, Oberkochen, Germany) were cut on a cryostat into 5 mm sections and examined under a light microscope. These tissue samples were also embedded in paraffin and cut into 5 mm sections for hematoxylin–eosin staining. All X-gal-positive epithelial cells were microscopically counted in the two sections of trachea 1.0 and 1.5 cm above the carina tracheae, the four bronchi and proximal bronchioli, and four terminal bronchioli of the whole left lung sections in each animal. Measurements of luciferase activity in vivo The harvested tracheae and right lower lobes of the lung were temporarily chilled in cold PBS. The tissue was minced in lysis buffer and the luciferase activity was measured using a luminometer (Aloka, Tokyo, Japan) and a luciferase assay kit (Promega). Protein concentration was also photometrically measured using a BioRad Protein Assay System (Bio-Rad Laboratories, Hercules, CA, USA), and all data were expressed as luciferase activity per milligram protein. Statistical analysis All values are expressed as mean ± s.d. The data were analyzed by a one way analysis of variance and, where appropriate, Student’s t test with Scheffe’s adjustment was used.

637


638

Acknowledgements We thank M Namoto and S Yugawa, Department of Pathology I, Kyushu University, for excellent technical assistance, and R Morishita, Department of Geriatric Medicine, Osaka University, for fruitful discussion. We are also grateful to K Yanaga, Department of Surgery II, Faculty of Medicine, Kyushu University, for critical comments on the manuscript. This work was supported by Grants-inAid from the Ministry of Education, Science, Sports and Culture of Japan (Nos 04454180, 06454186).

References 1 Miller DA. Human gene therapy comes of age. Nature 1992; 357: 455–460. 2 Yang Y et al. Cellular immunity to viral antigens limits E1deleted adenoviruses for gene therapy. Proc Natl Acad Sci USA 1994; 91: 4407–4411. 3 Caplen NJ et al. Gene therapy for cystic fibrosis in humans by liposome-mediated DNA transfer: the production of resources and regulatory process. Gene Therapy 1994; 1: 139–147. 4 Caplen NJ et al. Liposome-mediated CFTR gene transfer to the nasal epithelium of patients with cystic fibrosis. Nature Med 1995; 1: 39–46. 5 Yang Y et al. Inactivation of E2a in recombinant adenoviruses improves the prospect for gene therapy in cystic fibrosis. Nat Genet 1994; 7: 362–369. 6 Curiel DT et al. High-efficiency gene transfer mediated by adenovirus coupled to DNA–polylysine complexes. Hum Gene Ther 1992; 3: 147–154.

7 Kaneda Y et al. The improved efficient method for introducing macromolecules into the cells using HVJ (Sendai virus) liposomes with gangliosides. Exp Cell Res 1987; 173: 56–69. 8 Kaneda Y, Iwai K, Uchida T. Increased expression of DNA cointroduced with nuclear protein in adult rat liver. Science 1989; 243: 375–378. 9 Kaneda Y. Virus (Sendai virus envelopes) mediated gene transfer. Cell Biology: A Laboratory Handbook (vol 3). Academic Press: Orlando, FL, 1994, pp 50–57. 10 Yonemitsu Y et al. Characterization of in vivo gene transfer into the arterial wall mediated by the Sendai virus (HVJ)-liposomes: an effective tool for the in vivo study of arterial diseases. Lab Invest 1996; 75: 313–323. 11 Felgner PL et al. Lipofection: a highly efficient, lipid-mediated DNA transfection procedure. Proc Natl Acad Sci USA 1987; 84: 7413–7417. 12 Yagi K et al. Efficient gene transfer with less cytotoxicity by means of cationic multilamellar liposomes. Biochem Biophys Res Commun 1993; 196: 1042–1048. 13 McLachlan G et al. Evaluation in vitro and in vivo of cationic liposome expression construct complexes for cystic fibrosis gene therapy. Gene Therapy 1995; 2: 614–622. 14 Urabe M et al. Site-specific integration of the gene flanked by AAV-ITR into chromosome 19 is mediated by AAV large Rep protein(s). (Submitted).

References added in proof 15 McPherson C. Regulation of animal care and research? NIH’s opinion. J Anim Sci 1980; 51: 492–496. 16 National Institutes of Health. Guide for the Care of Laboratory Animals. United States Government Printing Office: Washington DC, 1985.