- Email: [email protected]
In vivo imaging of gene transfer to the respiratory tract
Available online at www.sciencedirect.com
Biomaterials 29 (2008) 1533e1540 www.elsevier.com/locate/biomaterials
In vivo imaging of gene transfer to the respiratory tract Uta Griesenbach a,h,*, Cuixiang Meng a,h, Raymond Farley a,h, Seng H. Cheng b, Ronald K. Scheule b, Mark H. Davies c, Paul C. Wolstenholme-Hogg c, Willem ten Hove d, Paul van der Hoeven d, Patrick L. Sinn e, Paul B. McCray Jr. e, Makoto Inoue f, Duncan M. Geddes a,h, Mamoru Hasegawa f, Gad Frankel g, Siouxsie Wiles g, Eric W.F.W. Alton a,h a
Department of Gene Therapy, Imperial College at the National Heart and Lung Institute, Manresa Road, London SW3 6LR, UK b Genzyme Corporation, Framingham, MA, USA c Genzyme Ltd, Haverhill, UK d OctoPlus N.V., Leiden, The Netherlands e Program in Gene Therapy, Department of Pediatrics, University of Iowa, Iowa City, IA, USA f DNAVEC Corporation, Tsukuba, Japan g Division of Cell and Molecular Biology, Imperial College London, London, UK h The UK Cystic Fibrosis Gene Therapy Consortium, London, UK Received 24 September 2007; accepted 17 November 2007 Available online 21 December 2007
Abstract Imaging of in vivo gene expression using luciferase expression in various organs has been used for several years. In contrast to other organs, in vivo imaging of the lung, particularly after non-viral gene transfer has not been extensively studied. The aim of this study was to address several questions: (1) Does in vivo light emission correlate with standard tissue homogenate-based luciferase detection in a dose-dependent manner? Recombinant Sendai virus (SeV) transduces airway epithelial cells very efficiently and was used to address this question, (2) Is the sensitivity of the assay sufficient to detect non-viral gene transfer? We treated mice with SeV-Lux vector using our standard ‘‘sniffing’’ protocol, a method that predominantly results in lung deposition. Dose-related in vivo light emission was visible in all animals. Importantly, there was a significant correlation (r > 0.90, p < 0.0001) between the in vivo and ex vivo assays in both the left and right lung. We next transfected the nasal epithelium via nasal perfusion or the lungs (‘‘sniffing’’) of mice with a luciferase plasmid (pCIKLux) complexed to the cationic lipid GL67 (n ¼ 25e27/group) and imaged luciferase expression in vivo 24 h after transfection. Gene expression was detectable in both organs. Correlation between the in vivo and ex vivo assays was significant (r ¼ 0.52, p < 0.005) in the left, but not the right lung. The correlation in the nose was weaker (r ¼ 0.45, p < 0.05). To our knowledge these studies show for the first time that this non-invasive method of assessing pulmonary gene transfer is viable for evaluating non-viral gene transfer agents. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Lung; Gene transfer; Lipid; In vivo test; Image analysis
1. Introduction * Corresponding author. Department of Gene Therapy, Imperial College at the National Heart and Lung Institute, Manresa Road, London SW3 6LR, UK. Tel.: þ44 (0)207 351 8339; fax: þ44 (0)207 351 8340. E-mail address: [email protected] (U. Griesenbach). 0142-9612/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2007.11.017
Bioluminescent imaging (BLI) of gene expression in living animals has been used for several years in a large range of applications including cancer research [1,2], modelling of infectious diseases [3], studying of proteineprotein interaction
1534
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
[4]. The technique involves the detection by specialised imaging equipment of visible light produced in vivo by luciferasecatalysed reactions which require energy (in the form of FMNH2 and ATP), oxygen and a specific substrate. In the context of gene therapy BLI has been used to study viral and non-viral gene transfer and gene therapy [5,6] using the luciferase gene from the firefly Photinus pyralis which emits lights after addition of the substrate D-luciferin. The technique is rapid, therefore, allowing for high throughput, but more importantly allowing gene expression to be followed over time in the same animal, thereby reducing animal numbers and improving accuracy. In addition, information about the distribution of gene expression can be gained. In contrast to other organs, in vivo imaging of lung transfection, particularly after topical non-viral gene transfer has not been extensively studied. Gene transfer efficiency to the lung after topical administration of viral and non-viral gene transfer agents (GTAs) is vectordependent, but in general comparatively inefficient. This is likely due to potent extra- and intracellular barriers [7] which have evolved to keep foreign particles out of the lung. However, it has previously been shown that transfection efficiency of adenoviral and lentiviral vectors after topical administration is sufficient to enable in vivo imaging [5,8]. In general non-viral GTA-mediated gene transfer to the lung is several log orders lower than viral GTA-mediated transduction [9]. To the best of our knowledge it is unknown whether in vivo BLI is sufficiently sensitive to allow detection of topical non-viral gene transfer to the lung. Cystic fibrosis (CF) is one of the diseases for which lung gene transfer is being optimised by many groups. For this purpose the nasal epithelium is frequently used as a surrogate for the lung in pre-clinical and clinical studies [10e12] and it would, therefore, be of interest to assess if BLI can be detected in the nose after non-viral gene transfer. It is also important to understand the correlation between BLI measurements and tissue protein levels determined using standard homogenization assays. This is a technical report addressing (1) if the sensitivity of BLI was sufficient to detect non-viral gene transfer in lung and nose of mice, which may have implications for the design of future experiments and (2) if in vivo BLI correlated with standard lung homogenate-based luciferase detection in lung and nose after topical administration of gene transfer agents. For these proof-of-principle studies we chose the most efficient viral and non-viral gene transfer agents in our hands for lung transfection, namely recombinant Sendai virus (SeV) [9,13] and the cationic lipid GL67 [14]. SeV, also known as murine parainfluenza virus type 1, is a member of the paramyxovirus family. The virus has a nonsegmented negative-strand RNA genome and replicates only in the cytoplasm. Advances in reverse genetics have enabled the generation of recombinant virus vector carrying reporter or therapeutic genes. We, and others, have previously demonstrated that SeV vector is one of the most efficient vectors for the transduction of airway epithelial cells in vivo, largely because of its ability to overcome some of the extra- and intracellular barriers to airway gene transfer [9,15]. In contrast to most other viral gene transfer vectors the receptors for SeV,
cholesterol and sialic acid, are present at the apical surface of airway epithelial cells and are, therefore, accessible after topical administration. Most of the early work was carried out with transmission-competent SeV virus, which carried reporter or therapeutic genes in addition to all viral genes. These first-generation vectors enter cells, replicate, and virus particles with the capacity to re-infect neighbouring cells bud out. In an attempt to improve the safety potential the fusion (F) protein, essential for virus entry, has been deleted from the viral genome (DF/SeV), but is supplied in trans during virus production in cell culture [16]. DF/SeV enters cells and replicates, but budding virus-like particles do not carry the F protein and, therefore, are unable to transduce neighbouring cells (transmission-incompetent). More recently, we assessed DF/SeV in murine airway epithelium in vivo and primary human nasal epithelium ex vivo and showed that the second generation of DF/SeV transduced airway epithelial cells as efficiently as the first-generation vector [13]. The non-viral gene transfer agent GL67 consists of three components: (1) the cationic lipid GL67 [Cholest-5-en-3-ol (3b)-3-[(3-aminopropyl)[4-[(3-aminopropyl)amino]butyl]carbamate], consisting of an amine (spermine) and a lipid (cholesterol) component linked together via a carbamate linker, (2) DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) a naturally occurring neutral lipid, (3) DMPE-PEG5000 (1,2Dimyristoyl-sn-Glycero-3-Phosphoethanolamine-N-[methoxy (Polyethylene glycol)5000] which is 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine linked to polyethylene glycol monomethylether (average molecular weight ca. 5000) via a carbamate linkage. To generate the gene transfer agent the three lipids are formulated at 1:2:0.05 (GL67:DOPE:DMPEPEG5000) molar ratios and complexed to DNA [14]. GL67 has been used in several lung gene therapy clinical trials [17,18] and is, in our hands the most efficient gene transfer agent for airway epithelial cells. 2. Material and methods 2.1. Transfection of mice with Sendai virus vector The generation and propagation of transmission-incompetent F proteindeleted DF/SeV vector carrying a firefly luciferase reporter gene (DF/SeVLux) was carried out as previously described [19]. The supernatant of LLC-MK2/F7 cells containing infectious particles was subsequently purified, concentrated and stored at 80 C. Virus titre was determined and expressed as transduction units (TU) per ml. All experiments were carried out with approval of appropriate local Ethics Committees and according to Home Office regulations. Balb/C mice (female, 6e10 weeks) were used for all experiments, because the animals are white which limits quenching of light emission by the fur. Briefly, mice were anaesthetised with metophane (Medical Development International Ltd., Springvale, Australia) and a single 100 ml bolus containing DF/SeV-Lux (105e108 TU/mouse, n ¼ 3/group) or a control virus (108 TU/mouse, n ¼ 3) was applied to the nose as a bolus and rapidly ‘sniffed’ into the lung. Reporter gene expression was quantified 48 h after transduction to detect maximal levels of gene expression.
2.2. Transfection of mice with GL67 The cationic lipid GL67 is a mixture of GL67 (Genzyme, Haverhill, UK), DOPE and DMPE-PEG5000 (Avanti Lipids, Alabama, USA) [14]. The lipids
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540 were formulated and freeze-dried by OctoPlus N.V. (Leiden, The Netherlands) and complexed to the eukaryotic expression plasmid pCIKLux carrying a luciferase reporter gene or to an ‘‘empty’’ control plasmid carrying no reporter gene as previously described [14]. 2.2.1. Lung transfection Balb/C mice (female, 6e10 weeks) were anaesthetised with metophane and a single 100 ml bolus containing 80 mg plasmid DNA was applied to the nose and rapidly ‘sniffed’ into the lung. 2.2.2. Nose transfection To maximise transfection efficiency in the nose, carboxymethylcellulose (CMC, Hercules, Inc., Wilmington, Delaware, USA) was added with the final dose containing 0.5% CMC and 80 mg plasmid DNA in 400 ml total volume. Mice were anaesthetised with Ketaset/Domitor (76 mg/kg and 1 mg/kg, respectively, National Veterinary Service, Stoke-on-Trent, UK). Catheters (<0.5 mm outer diameter) were inserted into the nostrils (approximately 2.5 mm). Using a syringe pump (Cole-Palmer, Illinois, USA) the GTA was perfused for 60 min. At the end of the procedure mice were injected intraperitoneally with Antisedan (1 mg/kg, National Veterinary Services) to reverse the anaesthesia. Reporter gene expression was quantified 24 h after transfection to detect maximal levels of gene expression.
2.3. Bioluminescent imaging Mice were injected intraperitoneally with 150 mg/kg of D-Luciferin (Xenogen Corporation, Alameda, CA) 10 min before imaging and were anaesthetised with isofluorane. Bioluminescence (photons s1 cm2 sr1) from living mice was measured using an IVIS50 system (Xenogen Corporation) at a binning of 4, over 1 or 10 min, using the software programme Living Image (Xenogen). For anatomical localisation a pseudocolour image representing light intensity (blue least intense, red most intense) was generated using Living Image software and superimposed over the greyscale reference image. To quantify bioluminescence in the nose and right and left lungs photon emission in a defined area was measured by marking a standardized area for quantification. Importantly the areas were marked using the greyscale reference image to avoid bias.
2.4. Tissue homogenate-based luciferase assay Mice were culled by cervical dislocation. Snout, skin and nasal plate were removed and the septum was dissected and frozen in liquid nitrogen. The mouse lung consists of four right lobes and a single left lobe. Left and right lungs were harvested separately and frozen in liquid nitrogen. To extract luciferase protein the septum and left and right lungs were placed in 100, 150 and 300 ml 1 RLB buffer (Promega, Southampton, UK), respectively. The septum was homogenized by cutting the tissue into small pieces using spring scissors followed by three freeze/thaw cycles and centrifugation at 13,000gav for 10 min. The lung was homogenized using a Fast-Prep homogenizer (Thermo Fisher Scientific, Waltham, MA, USA) set to 40 m/s for 45 s followed by 15 min incubation at room temperature. The supernatant was removed and transferred to a QiaShredder column (Qiagen, Crawley, Sussex) and centrifuged (2 min at 16,000gav). Luciferase activity was measured in the supernatant using a standard luciferase assay kit (Promega) and the TD-20e luminometer (Turner, Sunnyvale, CA, USA). Total protein per sample was determined using the BioRad protein assay kit (BioRad laboratories, Hercules, CA) and luciferase activity was expressed as arbitrary relative light units (RLU)/mg total protein. For comparative purposes, under the luciferase assay conditions used, 1000 RLU equate to approximately 20 ng of recombinant luciferase protein.
2.5. Statistical analysis Statistical analyses were performed by ANOVA after log transformation of not normally distributed data, followed by the Bonferroni post hoc analysis. Non-parametric correlations were analysed using the Spearman correlation. The null hypothesis was rejected at p < 0.05.
1535
3. Results 3.1. Correlation after Sendai virus transfection To assess if in vivo light emission correlated with standard tissue homogenate-based detection of gene expression we first transduced mice with recombinant SeV. Mice were transduced with DF/SeV vector carrying a luciferase gene (DF/ SeV-Lux, 105 to 108 TU/mouse, n ¼ 3/group) or an irrelevant control virus vector (108 TU/mouse, n ¼ 3) using bolus administration to the nose (‘‘sniffing’’). Forty-eight hours after transduction gene expression was imaged in vivo, after which mice were culled, left and right lungs removed, and processed for homogenate-based luciferase detection. Using BLI, dose-related gene expression was detected in the lung and nose of all DF/SeV-Lux vector transduced mice, but not in control animals (Fig. 1A). Gene expression was not equally distributed with 55% of mice showing signal predominantly in the left lung, 27% in the right lung and 18% in both regions. We next quantified photon emission in the left and right lungs separately to allow correlation with the homogenatebased luminescence (HBL) assay. To avoid bias, left and right lungs were marked in each mouse using just the photograph (see red boxes, Fig. 1B). There was a highly significant (r > 0.9, p < 0.001) correlation between BLI and HBL in both the left and right lungs (Fig. 2A,B). We have previously shown that SeV requires only a short contact time with the target cell for viral entry [9]. Thus, it is not surprising that the short contact time (seconds) during ‘‘sniffing’’ was sufficient to transduce the nasal epithelium efficiently. All mice treated (n ¼ 11) showed strong bioluminescence in the nose, but the correlation between BLI and HBL was not as strong as in the lung (r ¼ 0.68, p < 0.05; Fig. 2C). 3.2. Bioluminescence imaging after GL67-mediated lung gene transfer We next transfected mice via ‘‘sniffing’’ with GL67 complexed to the eukaryotic expression plasmid pCIKLux (80 mg pCIKLux/mouse, n ¼ 27) or a negative control plasmid (n ¼ 9). The lungs of all GL67/pCIKLux treated mice were positive for BLI (Fig. 3), but signal intensity and distribution varied, with 26% of mice showing bioluminescence in the left, 37% in the right and 37% in both lungs. As for the SeV-treated mice there was good correlation between BLI and HBL in the left lung (r ¼ 0.53, p < 0.005; Fig. 4A). There was no correlation in the right lung (r ¼ 0.19, p ¼ NS; Fig. 4B). No signal was detected in the nose after the brief contact period inherent in this sniffing technique. 3.3. Bioluminescence imaging after GL67-mediated nose gene transfer In the context of cystic fibrosis gene therapy the nasal epithelium is commonly used as a surrogate for lung transfection.
1536
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
Fig. 1. Bioluminescence imaging in the lungs after Sendai virus (DF/SeV-Lux) transduction. (A) Mice were transduced with DF/SeV-Lux (105 to 108 TU/mouse) or a DF/SeV-GFP control virus (108 TU/mouse) via nasal ‘‘sniffing’’ and bioluminescence was imaged in vivo 48 h after transduction (n ¼ 3/group, a representative images for each group is shown). (B) Images in (A) were generated by overlaying a black and white photographic image and a bioluminescent image. To quantify light emission in nose, left and right lung, these regions were marked (red boxes) in the photographic image to avoid bias.
To increase the contact time we perfused the nasal epithelium of mice with GL67/pCIKLux complexes (80 mg pCIKLux/ mouse, n ¼ 25) or a negative control plasmid (n ¼ 8) over a period of 75 min. BLI was detectable in the nose of 23 out of 25 treated mice (Fig. 5A). As expected gene expression was restricted to the nasal epithelium using this delivery method. Correlation between BLI and HBL was significant, but comparatively weak (r ¼ 0.45, p ¼ 0.01; Fig. 5B). 4. Discussion In vivo detection of gene transfer via bioluminescence imaging has been extensively documented over the last few years. Here, we show that SeV-mediated transduction leads to dose-related light emission and that, importantly, photon
emission and traditional tissue homogenate-based luciferase detection are well correlated in the lung. We also show, for the first time, that topical non-viral gene transfer to the lung generates sufficient gene expression to allow bioluminescence imaging thereby facilitating both intra-animal time course studies and a reduction in animal usage. Photon emission and traditional tissue homogenate-based luciferase detection are correlated, albeit weakly, in the left lung, but not in the right lungs and nose. Efficient viral and non-viral lung gene transfer agents were chosen for these proof-of-principle studies. We have previously shown that transmission-competent and transmissionincompetent F gene-deleted recombinant Sendai virus (DF/ SeV) transduced the lung efficiently [9,13], in part due to appropriate viral receptor being present on airway epithelial
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
1537
Fig. 2. Correlation between photon emission and tissue homogenate-based luciferase detection after transduction with DF/SeV-Lux. Mice were transduced with DF/SeV-Lux (see Fig. 1). Immediately after imaging the mice were culled, left and right lungs and nasal tissues were harvested for homogenate-based luciferase detection. Photon emission and homogenate-based luciferase expression were correlated. (A) Left lungs, (B) right lungs, and (C) nose. Each symbol represents one animal.
cells. In an extensive pre-clinical programme we have compared a number of non-viral gene transfer agents in the lung indicating that cationic lipid GL67 is the most efficient nonviral GTA for in vivo airway epithelial cell transfection
(manuscript in preparation). We are currently preparing to assess the efficacy of GL67 in a large multi-dose clinical trial programme. SeV and GL67 were, therefore, chosen for the proof-of-principle studies described here. Gene expression
Fig. 3. Bioluminescence imaging in the lungs after GL67/pCIKLux transfection. Mice lungs were transfected with pCIKLux (n ¼ 27) or a negative control plasmid (n ¼ 9) complexed to GL67 via bolus administration to the nose (‘‘sniffing’’) and bioluminescence was imaged in vivo 24 h after transduction. Representative images are shown for three treated and one control animal.
1538
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
Fig. 4. Correlation between photon emission and tissue homogenate-based luciferase detection after GL67/pCIKLux-mediated lung transfection. Mice were transfected with GL67/pCIKLux (see Fig. 3). Immediately after imaging the mice were culled, left and right lungs were harvested for homogenate-based luciferase detection. Photon emission and homogenate-based luciferase expression were correlated. (A) Left lungs and (B) right lungs. Each symbol represents one animal.
was quantified 24 and 48 h after GL67- and DF/SeV-mediated transduction, respectively, as these are optimal time-points to ensure maximal gene expression for each vector [13,20]. Dose-related bioluminescence was detected in the lungs of all mice transduced with DF/SeV-Lux vector. Bioluminescence was also detected in the lungs of all mice transfected with GL67/pCIKLux via ‘‘sniffing’’, and 92% of the noses of mice following topical perfusion. Absolute values of light emission (relative light units, RLU) vary greatly from luminometer to luminometer for a given amount of luciferase protein. It is, therefore, difficult to specify universally useful cut-off RLU values over which BLI will be possible for different machines. Using the Turner TD-20e luminometer 10 RLU/mg protein was insufficient, whereas values of 100 RLU/mg protein were high enough to allow BLI (data not shown). In addition to absolute levels of luciferase expression/mg lung protein spatial distribution of gene expression will play a significant role. Due to the comparatively low transfection efficiency of non-viral gene transfer agents we are currently unable to estimate transfection efficiency on a per cell basis and the effect of spacial distribution on BLI can, therefore, not be assessed at this point. Gene expression after DF/SeV or GL67 ‘‘sniffing’’ was not evenly distributed throughout the lungs with distribution varying from animal to animal. This may in part be explained by the asymmetrical bifurcation of the left and right bronchi (i.e. the left and right main stem bronchus radiate a different angles from the trachea) and the position of the animal during gene transfer agent administration and clearly illustrates the additional information that can be gained by this technique. Interestingly, DF/SeV transduced the left lung exclusively in 55% of mice compared to 26% transfected with GL67/ pCIKLux complexes. The reason for this is unknown, but may relate to the different consistency of virus solution and the more viscous GL67/pCIKLux complexes. Bolus administration of gene transfer agents to the nose followed by rapid sniffing into the lung allows only very brief
(seconds) contact time between the nasal epithelium and the gene transfer agent. DF/SeV-Lux but not GL67/pCIKLux ‘‘sniffing’’ led to high levels of bioluminescence in the nose further confirming our previous findings that in contrast to most other vectors SeV requires only very brief contact with the airway epithelium to allow virus entry [9]. Importantly, a viscoelastic gel formulation [8] in combination with slow perfusion of GL67/pCIKLux over the nasal epithelium ensured sufficient contact time for uptake of the liposome/ DNA complexes into the cells. This highlights contact time as a likely important factor to consider in clinical trials, as well as allowing BLI to be used in studies in CF mice, where the nose provides a convenient assay for changes in the underlying bioelectric defect. It is important to understand how this novel technique correlates with standard tissue homogenate-based assays, particularly since tissue depth, bones and fur are known to quench emitted light. Rettig et al. assessed correlations after hydrodynamic liver transfection with plasmid DNA and showed linearity between BLI and HBL expression over five orders of magnitude [21]. To the best of our knowledge correlations have not been assessed for lung and nose gene transfer. The correlation between left and right lungs was very high (r ¼ 0.9, p < 0.0001) after DF/SeV transduction, but weaker in the left lung (r ¼ 0.52, p < 0.005) and absent in the right lung (r ¼ 0.19, p ¼ NS) after GL67-mediated transfection. This implies that the degree of correlation is dependent on the overall level of gene expression, with low-level gene expression being more affected by signal quenching. Importantly, right and left lungs were affected to different degrees, which may be related to differences in anatomy. Thus, in contrast to the left lung, which is a discrete single lobe in mice, the right lung consists of four individual lobes which overlay each other and may, therefore, be more susceptible to tissue quenching of light emission. The correlation between light emission and gene expression in the nose was significant, but not particularly strong after both viral (r ¼ 0.68, p < 0.05) and non-viral (r ¼ 0.45,
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
1539
Fig. 5. Bioluminescence imaging in the nose after GL67/pCIKLux transfection. Mouse nasal tissue was transfected with pCIKLux (n ¼ 27) or a negative control plasmid (n ¼ 8) complexed to GL67 by nasal perfusion. (A) Bioluminescence was imaged in vivo 24 h after transduction. Representative images are shown. (B) Immediately after imaging the mice were culled and nasal tissue was harvested for homogenate-based luciferase detection. Photon emission and homogenate-based luciferase expression were correlated. Each symbol represents one animal.
p ¼ 0.01) gene transfer. Interestingly this was independent of the overall level of gene expression and may be due to (1) incomplete removal of all transfected tissue from the nasal cavity and/or (2) transfection of tissue outside the nose such as the mouth and snout due to overspilling and swallowing of vector during delivery. Firefly luciferase is one of a few proteins that are sensitive to anaesthetics and is partly inhibited by commonly used anaesthetics such as isofluorane, halothane, ketones, alkanes and ethers [22,23]. It has been suggested that anaesthetics compete with binding of luciferin to the protein. We cannot exclude that luciferase expression in our study was somewhat
reduced by the anaesthetic, but clearly all commonly used agents have this potential. 5. Conclusions SeV vector transduction and GL67-mediated lung gene transfer following topical administration via ‘‘sniffing’’ and nasal perfusion can be imaged in living animals and will, therefore, enable us to follow gene expression over time in the same animal. The correlation between light emission and absolute gene expression is dependent on the overall level of luciferase expression, and is better in the left than in the right lung.
1540
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
Acknowledgements We thank Luci Somerton for help with preparing the manuscript. This work was funded by the Cystic Fibrosis Trust and the CF Trust Dr Benjamin Angel Senior Fellowship (UG). References [1] Kutschka I, Kofidis T, Chen IY, von DG, Zwierzchoniewska M, Hoyt G, et al. Adenoviral human BCL-2 transgene expression attenuates early donor cell death after cardiomyoblast transplantation into ischemic rat hearts. Circulation 2006;114(Suppl. 1):I174e80. [2] Wang W, Kim SH, El-Deiry WS. Small-molecule modulators of p53 family signaling and antitumor effects in p53-deficient human colon tumor xenografts. Proc Natl Acad Sci U S A 2006;103(29):11003e8. [3] Wiles S, Pickard KM, Peng K, MacDonald TT, Frankel G. In vivo bioluminescence imaging of the murine pathogen Citrobacter rodentium. Infect Immun 2006;74(9):5391e6. [4] De A, Gambhir SS. Noninvasive imaging of proteineprotein interactions from live cells and living subjects using bioluminescence resonance energy transfer. FASEB J 2005;19(14):2017e9. [5] Sinn PL, Burnight ER, Hickey MA, Blissard GW, McCray Jr PB. Persistent gene expression in mouse nasal epithelia following feline immunodeficiency virus-based vector gene transfer. J Virol 2005;79(20):12818e27. [6] Wilber A, Frandsen JL, Wangensteen KJ, Ekker SC, Wang X, McIvor RS. Dynamic gene expression after systemic delivery of plasmid DNA as determined by in vivo bioluminescence imaging. Hum Gene Ther 2005;16(11):1325e32. [7] Ferrari S, Geddes DM, Alton EW. Barriers to and new approaches for gene therapy and gene delivery in cystic fibrosis. Adv Drug Deliv Rev 2002;54(11):1373e93. [8] Sinn PL, Shah AJ, Donovan MD, McCray Jr PB. Viscoelastic gel formulations enhance airway epithelial gene transfer with viral vectors. Am J Respir Cell Mol Biol 2005;32(5):404e10. [9] Yonemitsu Y, Kitson C, Ferrari S, Farley R, Griesenbach U, Judd D, et al. Efficient gene transfer to airway epithelium using recombinant Sendai virus. Nat Biotechnol 2000;18(9):970e3. [10] Konstan MW, Davis PB, Wagener JS, Hilliard KA, Stern RC, Milgram LJ, et al. Compacted DNA nanoparticles administered to the nasal mucosa of cystic fibrosis subjects are safe and demonstrate partial to complete cystic fibrosis transmembrane regulator reconstitution. Hum Gene Ther 2004;15(12):1255e69.
[11] Limberis MP, Wilson JM. Adeno-associated virus serotype 9 vectors transduce murine alveolar and nasal epithelia and can be readministered. Proc Natl Acad Sci U S A 2006;103(35):12993e8. [12] Sumner-Jones SG, Davies LA, Varathalingam A, Gill DR, Hyde SC. Long-term persistence of gene expression from adeno-associated virus serotype 5 in the mouse airways. Gene Ther 2006;13(24):1703e13. [13] Ferrari S, Griesenbach U, Shiraki-Iida T, Shu T, Hironaka T, Hou X, et al. A defective nontransmissible recombinant Sendai virus mediates efficient gene transfer to airway epithelium in vivo. Gene Ther 2004;11(22): 1659e64. [14] Lee ER, Marshall J, Siegel CS, Jiang C, Yew NS, Nichols MR, et al. Detailed analysis of structures and formulations of cationic lipids for efficient gene transfer to the lung. Hum Gene Ther 1996;7(14):1701e17. [15] Griesenbach U, Cassady RL, Ferrari S, Fukumura M, Muller C, Schmitt E, et al. The nasal epithelium as a factory for systemic protein delivery. Mol Ther 2002;5(2):98e103. [16] Ho TC, Jeng KS, Hu CP, Chang C. Effects of genomic length on translocation of hepatitis B virus polymerase-linked oligomer. J Virol 2000;74(19):9010e8. [17] Alton EW, Stern M, Farley R, Jaffe A, Chadwick SL, Phillips J, et al. Cationic lipid-mediated CFTR gene transfer to the lungs and nose of patients with cystic fibrosis: a double-blind placebo-controlled trial. Lancet 1999;353(9157):947e54. [18] Ruiz FE, Clancy JP, Perricone MA, Bebok Z, Hong JS, Cheng SH, et al. A clinical inflammatory syndrome attributable to aerosolized lipid-DNA administration in cystic fibrosis. Hum Gene Ther 2001;12(7):751e61. [19] Hirata T, Iida A, Shiraki-Iida T, Kitazato K, Kato A, Nagai Y, et al. An improved method for recovery of F-defective Sendai virus expressing foreign genes from cloned cDNA. J Virol Methods 2002;104(2): 125e33. [20] Gill DR, Smyth SE, Goddard CA, Pringle IA, Higgins CF, Colledge WH, et al. Increased persistence of lung gene expression using plasmids containing the ubiquitin C or elongation factor 1alpha promoter. Gene Ther 2001;8(20):1539e46. [21] Rettig GR, McAnuff M, Liu D, Kim JS, Rice KG. Quantitative bioluminescence imaging of transgene expression in vivo. Anal Biochem 2006;355(1):90e4. [22] Franks NP, Lieb WR. Do general anaesthetics act by competitive binding to specific receptors? Nature 1984;310(5978):599e601. [23] Zhang Y, Stabernack CR, Dutton R, Sonner J, Trudell JR, Mihic SJ, et al. Luciferase as a model for the site of inhaled anesthetic action. Anesth Analg 2001;93(5):1246e52.
Biomaterials 29 (2008) 1533e1540 www.elsevier.com/locate/biomaterials
In vivo imaging of gene transfer to the respiratory tract Uta Griesenbach a,h,*, Cuixiang Meng a,h, Raymond Farley a,h, Seng H. Cheng b, Ronald K. Scheule b, Mark H. Davies c, Paul C. Wolstenholme-Hogg c, Willem ten Hove d, Paul van der Hoeven d, Patrick L. Sinn e, Paul B. McCray Jr. e, Makoto Inoue f, Duncan M. Geddes a,h, Mamoru Hasegawa f, Gad Frankel g, Siouxsie Wiles g, Eric W.F.W. Alton a,h a
Department of Gene Therapy, Imperial College at the National Heart and Lung Institute, Manresa Road, London SW3 6LR, UK b Genzyme Corporation, Framingham, MA, USA c Genzyme Ltd, Haverhill, UK d OctoPlus N.V., Leiden, The Netherlands e Program in Gene Therapy, Department of Pediatrics, University of Iowa, Iowa City, IA, USA f DNAVEC Corporation, Tsukuba, Japan g Division of Cell and Molecular Biology, Imperial College London, London, UK h The UK Cystic Fibrosis Gene Therapy Consortium, London, UK Received 24 September 2007; accepted 17 November 2007 Available online 21 December 2007
Abstract Imaging of in vivo gene expression using luciferase expression in various organs has been used for several years. In contrast to other organs, in vivo imaging of the lung, particularly after non-viral gene transfer has not been extensively studied. The aim of this study was to address several questions: (1) Does in vivo light emission correlate with standard tissue homogenate-based luciferase detection in a dose-dependent manner? Recombinant Sendai virus (SeV) transduces airway epithelial cells very efficiently and was used to address this question, (2) Is the sensitivity of the assay sufficient to detect non-viral gene transfer? We treated mice with SeV-Lux vector using our standard ‘‘sniffing’’ protocol, a method that predominantly results in lung deposition. Dose-related in vivo light emission was visible in all animals. Importantly, there was a significant correlation (r > 0.90, p < 0.0001) between the in vivo and ex vivo assays in both the left and right lung. We next transfected the nasal epithelium via nasal perfusion or the lungs (‘‘sniffing’’) of mice with a luciferase plasmid (pCIKLux) complexed to the cationic lipid GL67 (n ¼ 25e27/group) and imaged luciferase expression in vivo 24 h after transfection. Gene expression was detectable in both organs. Correlation between the in vivo and ex vivo assays was significant (r ¼ 0.52, p < 0.005) in the left, but not the right lung. The correlation in the nose was weaker (r ¼ 0.45, p < 0.05). To our knowledge these studies show for the first time that this non-invasive method of assessing pulmonary gene transfer is viable for evaluating non-viral gene transfer agents. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Lung; Gene transfer; Lipid; In vivo test; Image analysis
1. Introduction * Corresponding author. Department of Gene Therapy, Imperial College at the National Heart and Lung Institute, Manresa Road, London SW3 6LR, UK. Tel.: þ44 (0)207 351 8339; fax: þ44 (0)207 351 8340. E-mail address: [email protected] (U. Griesenbach). 0142-9612/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2007.11.017
Bioluminescent imaging (BLI) of gene expression in living animals has been used for several years in a large range of applications including cancer research [1,2], modelling of infectious diseases [3], studying of proteineprotein interaction
1534
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
[4]. The technique involves the detection by specialised imaging equipment of visible light produced in vivo by luciferasecatalysed reactions which require energy (in the form of FMNH2 and ATP), oxygen and a specific substrate. In the context of gene therapy BLI has been used to study viral and non-viral gene transfer and gene therapy [5,6] using the luciferase gene from the firefly Photinus pyralis which emits lights after addition of the substrate D-luciferin. The technique is rapid, therefore, allowing for high throughput, but more importantly allowing gene expression to be followed over time in the same animal, thereby reducing animal numbers and improving accuracy. In addition, information about the distribution of gene expression can be gained. In contrast to other organs, in vivo imaging of lung transfection, particularly after topical non-viral gene transfer has not been extensively studied. Gene transfer efficiency to the lung after topical administration of viral and non-viral gene transfer agents (GTAs) is vectordependent, but in general comparatively inefficient. This is likely due to potent extra- and intracellular barriers [7] which have evolved to keep foreign particles out of the lung. However, it has previously been shown that transfection efficiency of adenoviral and lentiviral vectors after topical administration is sufficient to enable in vivo imaging [5,8]. In general non-viral GTA-mediated gene transfer to the lung is several log orders lower than viral GTA-mediated transduction [9]. To the best of our knowledge it is unknown whether in vivo BLI is sufficiently sensitive to allow detection of topical non-viral gene transfer to the lung. Cystic fibrosis (CF) is one of the diseases for which lung gene transfer is being optimised by many groups. For this purpose the nasal epithelium is frequently used as a surrogate for the lung in pre-clinical and clinical studies [10e12] and it would, therefore, be of interest to assess if BLI can be detected in the nose after non-viral gene transfer. It is also important to understand the correlation between BLI measurements and tissue protein levels determined using standard homogenization assays. This is a technical report addressing (1) if the sensitivity of BLI was sufficient to detect non-viral gene transfer in lung and nose of mice, which may have implications for the design of future experiments and (2) if in vivo BLI correlated with standard lung homogenate-based luciferase detection in lung and nose after topical administration of gene transfer agents. For these proof-of-principle studies we chose the most efficient viral and non-viral gene transfer agents in our hands for lung transfection, namely recombinant Sendai virus (SeV) [9,13] and the cationic lipid GL67 [14]. SeV, also known as murine parainfluenza virus type 1, is a member of the paramyxovirus family. The virus has a nonsegmented negative-strand RNA genome and replicates only in the cytoplasm. Advances in reverse genetics have enabled the generation of recombinant virus vector carrying reporter or therapeutic genes. We, and others, have previously demonstrated that SeV vector is one of the most efficient vectors for the transduction of airway epithelial cells in vivo, largely because of its ability to overcome some of the extra- and intracellular barriers to airway gene transfer [9,15]. In contrast to most other viral gene transfer vectors the receptors for SeV,
cholesterol and sialic acid, are present at the apical surface of airway epithelial cells and are, therefore, accessible after topical administration. Most of the early work was carried out with transmission-competent SeV virus, which carried reporter or therapeutic genes in addition to all viral genes. These first-generation vectors enter cells, replicate, and virus particles with the capacity to re-infect neighbouring cells bud out. In an attempt to improve the safety potential the fusion (F) protein, essential for virus entry, has been deleted from the viral genome (DF/SeV), but is supplied in trans during virus production in cell culture [16]. DF/SeV enters cells and replicates, but budding virus-like particles do not carry the F protein and, therefore, are unable to transduce neighbouring cells (transmission-incompetent). More recently, we assessed DF/SeV in murine airway epithelium in vivo and primary human nasal epithelium ex vivo and showed that the second generation of DF/SeV transduced airway epithelial cells as efficiently as the first-generation vector [13]. The non-viral gene transfer agent GL67 consists of three components: (1) the cationic lipid GL67 [Cholest-5-en-3-ol (3b)-3-[(3-aminopropyl)[4-[(3-aminopropyl)amino]butyl]carbamate], consisting of an amine (spermine) and a lipid (cholesterol) component linked together via a carbamate linker, (2) DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) a naturally occurring neutral lipid, (3) DMPE-PEG5000 (1,2Dimyristoyl-sn-Glycero-3-Phosphoethanolamine-N-[methoxy (Polyethylene glycol)5000] which is 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine linked to polyethylene glycol monomethylether (average molecular weight ca. 5000) via a carbamate linkage. To generate the gene transfer agent the three lipids are formulated at 1:2:0.05 (GL67:DOPE:DMPEPEG5000) molar ratios and complexed to DNA [14]. GL67 has been used in several lung gene therapy clinical trials [17,18] and is, in our hands the most efficient gene transfer agent for airway epithelial cells. 2. Material and methods 2.1. Transfection of mice with Sendai virus vector The generation and propagation of transmission-incompetent F proteindeleted DF/SeV vector carrying a firefly luciferase reporter gene (DF/SeVLux) was carried out as previously described [19]. The supernatant of LLC-MK2/F7 cells containing infectious particles was subsequently purified, concentrated and stored at 80 C. Virus titre was determined and expressed as transduction units (TU) per ml. All experiments were carried out with approval of appropriate local Ethics Committees and according to Home Office regulations. Balb/C mice (female, 6e10 weeks) were used for all experiments, because the animals are white which limits quenching of light emission by the fur. Briefly, mice were anaesthetised with metophane (Medical Development International Ltd., Springvale, Australia) and a single 100 ml bolus containing DF/SeV-Lux (105e108 TU/mouse, n ¼ 3/group) or a control virus (108 TU/mouse, n ¼ 3) was applied to the nose as a bolus and rapidly ‘sniffed’ into the lung. Reporter gene expression was quantified 48 h after transduction to detect maximal levels of gene expression.
2.2. Transfection of mice with GL67 The cationic lipid GL67 is a mixture of GL67 (Genzyme, Haverhill, UK), DOPE and DMPE-PEG5000 (Avanti Lipids, Alabama, USA) [14]. The lipids
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540 were formulated and freeze-dried by OctoPlus N.V. (Leiden, The Netherlands) and complexed to the eukaryotic expression plasmid pCIKLux carrying a luciferase reporter gene or to an ‘‘empty’’ control plasmid carrying no reporter gene as previously described [14]. 2.2.1. Lung transfection Balb/C mice (female, 6e10 weeks) were anaesthetised with metophane and a single 100 ml bolus containing 80 mg plasmid DNA was applied to the nose and rapidly ‘sniffed’ into the lung. 2.2.2. Nose transfection To maximise transfection efficiency in the nose, carboxymethylcellulose (CMC, Hercules, Inc., Wilmington, Delaware, USA) was added with the final dose containing 0.5% CMC and 80 mg plasmid DNA in 400 ml total volume. Mice were anaesthetised with Ketaset/Domitor (76 mg/kg and 1 mg/kg, respectively, National Veterinary Service, Stoke-on-Trent, UK). Catheters (<0.5 mm outer diameter) were inserted into the nostrils (approximately 2.5 mm). Using a syringe pump (Cole-Palmer, Illinois, USA) the GTA was perfused for 60 min. At the end of the procedure mice were injected intraperitoneally with Antisedan (1 mg/kg, National Veterinary Services) to reverse the anaesthesia. Reporter gene expression was quantified 24 h after transfection to detect maximal levels of gene expression.
2.3. Bioluminescent imaging Mice were injected intraperitoneally with 150 mg/kg of D-Luciferin (Xenogen Corporation, Alameda, CA) 10 min before imaging and were anaesthetised with isofluorane. Bioluminescence (photons s1 cm2 sr1) from living mice was measured using an IVIS50 system (Xenogen Corporation) at a binning of 4, over 1 or 10 min, using the software programme Living Image (Xenogen). For anatomical localisation a pseudocolour image representing light intensity (blue least intense, red most intense) was generated using Living Image software and superimposed over the greyscale reference image. To quantify bioluminescence in the nose and right and left lungs photon emission in a defined area was measured by marking a standardized area for quantification. Importantly the areas were marked using the greyscale reference image to avoid bias.
2.4. Tissue homogenate-based luciferase assay Mice were culled by cervical dislocation. Snout, skin and nasal plate were removed and the septum was dissected and frozen in liquid nitrogen. The mouse lung consists of four right lobes and a single left lobe. Left and right lungs were harvested separately and frozen in liquid nitrogen. To extract luciferase protein the septum and left and right lungs were placed in 100, 150 and 300 ml 1 RLB buffer (Promega, Southampton, UK), respectively. The septum was homogenized by cutting the tissue into small pieces using spring scissors followed by three freeze/thaw cycles and centrifugation at 13,000gav for 10 min. The lung was homogenized using a Fast-Prep homogenizer (Thermo Fisher Scientific, Waltham, MA, USA) set to 40 m/s for 45 s followed by 15 min incubation at room temperature. The supernatant was removed and transferred to a QiaShredder column (Qiagen, Crawley, Sussex) and centrifuged (2 min at 16,000gav). Luciferase activity was measured in the supernatant using a standard luciferase assay kit (Promega) and the TD-20e luminometer (Turner, Sunnyvale, CA, USA). Total protein per sample was determined using the BioRad protein assay kit (BioRad laboratories, Hercules, CA) and luciferase activity was expressed as arbitrary relative light units (RLU)/mg total protein. For comparative purposes, under the luciferase assay conditions used, 1000 RLU equate to approximately 20 ng of recombinant luciferase protein.
2.5. Statistical analysis Statistical analyses were performed by ANOVA after log transformation of not normally distributed data, followed by the Bonferroni post hoc analysis. Non-parametric correlations were analysed using the Spearman correlation. The null hypothesis was rejected at p < 0.05.
1535
3. Results 3.1. Correlation after Sendai virus transfection To assess if in vivo light emission correlated with standard tissue homogenate-based detection of gene expression we first transduced mice with recombinant SeV. Mice were transduced with DF/SeV vector carrying a luciferase gene (DF/ SeV-Lux, 105 to 108 TU/mouse, n ¼ 3/group) or an irrelevant control virus vector (108 TU/mouse, n ¼ 3) using bolus administration to the nose (‘‘sniffing’’). Forty-eight hours after transduction gene expression was imaged in vivo, after which mice were culled, left and right lungs removed, and processed for homogenate-based luciferase detection. Using BLI, dose-related gene expression was detected in the lung and nose of all DF/SeV-Lux vector transduced mice, but not in control animals (Fig. 1A). Gene expression was not equally distributed with 55% of mice showing signal predominantly in the left lung, 27% in the right lung and 18% in both regions. We next quantified photon emission in the left and right lungs separately to allow correlation with the homogenatebased luminescence (HBL) assay. To avoid bias, left and right lungs were marked in each mouse using just the photograph (see red boxes, Fig. 1B). There was a highly significant (r > 0.9, p < 0.001) correlation between BLI and HBL in both the left and right lungs (Fig. 2A,B). We have previously shown that SeV requires only a short contact time with the target cell for viral entry [9]. Thus, it is not surprising that the short contact time (seconds) during ‘‘sniffing’’ was sufficient to transduce the nasal epithelium efficiently. All mice treated (n ¼ 11) showed strong bioluminescence in the nose, but the correlation between BLI and HBL was not as strong as in the lung (r ¼ 0.68, p < 0.05; Fig. 2C). 3.2. Bioluminescence imaging after GL67-mediated lung gene transfer We next transfected mice via ‘‘sniffing’’ with GL67 complexed to the eukaryotic expression plasmid pCIKLux (80 mg pCIKLux/mouse, n ¼ 27) or a negative control plasmid (n ¼ 9). The lungs of all GL67/pCIKLux treated mice were positive for BLI (Fig. 3), but signal intensity and distribution varied, with 26% of mice showing bioluminescence in the left, 37% in the right and 37% in both lungs. As for the SeV-treated mice there was good correlation between BLI and HBL in the left lung (r ¼ 0.53, p < 0.005; Fig. 4A). There was no correlation in the right lung (r ¼ 0.19, p ¼ NS; Fig. 4B). No signal was detected in the nose after the brief contact period inherent in this sniffing technique. 3.3. Bioluminescence imaging after GL67-mediated nose gene transfer In the context of cystic fibrosis gene therapy the nasal epithelium is commonly used as a surrogate for lung transfection.
1536
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
Fig. 1. Bioluminescence imaging in the lungs after Sendai virus (DF/SeV-Lux) transduction. (A) Mice were transduced with DF/SeV-Lux (105 to 108 TU/mouse) or a DF/SeV-GFP control virus (108 TU/mouse) via nasal ‘‘sniffing’’ and bioluminescence was imaged in vivo 48 h after transduction (n ¼ 3/group, a representative images for each group is shown). (B) Images in (A) were generated by overlaying a black and white photographic image and a bioluminescent image. To quantify light emission in nose, left and right lung, these regions were marked (red boxes) in the photographic image to avoid bias.
To increase the contact time we perfused the nasal epithelium of mice with GL67/pCIKLux complexes (80 mg pCIKLux/ mouse, n ¼ 25) or a negative control plasmid (n ¼ 8) over a period of 75 min. BLI was detectable in the nose of 23 out of 25 treated mice (Fig. 5A). As expected gene expression was restricted to the nasal epithelium using this delivery method. Correlation between BLI and HBL was significant, but comparatively weak (r ¼ 0.45, p ¼ 0.01; Fig. 5B). 4. Discussion In vivo detection of gene transfer via bioluminescence imaging has been extensively documented over the last few years. Here, we show that SeV-mediated transduction leads to dose-related light emission and that, importantly, photon
emission and traditional tissue homogenate-based luciferase detection are well correlated in the lung. We also show, for the first time, that topical non-viral gene transfer to the lung generates sufficient gene expression to allow bioluminescence imaging thereby facilitating both intra-animal time course studies and a reduction in animal usage. Photon emission and traditional tissue homogenate-based luciferase detection are correlated, albeit weakly, in the left lung, but not in the right lungs and nose. Efficient viral and non-viral lung gene transfer agents were chosen for these proof-of-principle studies. We have previously shown that transmission-competent and transmissionincompetent F gene-deleted recombinant Sendai virus (DF/ SeV) transduced the lung efficiently [9,13], in part due to appropriate viral receptor being present on airway epithelial
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
1537
Fig. 2. Correlation between photon emission and tissue homogenate-based luciferase detection after transduction with DF/SeV-Lux. Mice were transduced with DF/SeV-Lux (see Fig. 1). Immediately after imaging the mice were culled, left and right lungs and nasal tissues were harvested for homogenate-based luciferase detection. Photon emission and homogenate-based luciferase expression were correlated. (A) Left lungs, (B) right lungs, and (C) nose. Each symbol represents one animal.
cells. In an extensive pre-clinical programme we have compared a number of non-viral gene transfer agents in the lung indicating that cationic lipid GL67 is the most efficient nonviral GTA for in vivo airway epithelial cell transfection
(manuscript in preparation). We are currently preparing to assess the efficacy of GL67 in a large multi-dose clinical trial programme. SeV and GL67 were, therefore, chosen for the proof-of-principle studies described here. Gene expression
Fig. 3. Bioluminescence imaging in the lungs after GL67/pCIKLux transfection. Mice lungs were transfected with pCIKLux (n ¼ 27) or a negative control plasmid (n ¼ 9) complexed to GL67 via bolus administration to the nose (‘‘sniffing’’) and bioluminescence was imaged in vivo 24 h after transduction. Representative images are shown for three treated and one control animal.
1538
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
Fig. 4. Correlation between photon emission and tissue homogenate-based luciferase detection after GL67/pCIKLux-mediated lung transfection. Mice were transfected with GL67/pCIKLux (see Fig. 3). Immediately after imaging the mice were culled, left and right lungs were harvested for homogenate-based luciferase detection. Photon emission and homogenate-based luciferase expression were correlated. (A) Left lungs and (B) right lungs. Each symbol represents one animal.
was quantified 24 and 48 h after GL67- and DF/SeV-mediated transduction, respectively, as these are optimal time-points to ensure maximal gene expression for each vector [13,20]. Dose-related bioluminescence was detected in the lungs of all mice transduced with DF/SeV-Lux vector. Bioluminescence was also detected in the lungs of all mice transfected with GL67/pCIKLux via ‘‘sniffing’’, and 92% of the noses of mice following topical perfusion. Absolute values of light emission (relative light units, RLU) vary greatly from luminometer to luminometer for a given amount of luciferase protein. It is, therefore, difficult to specify universally useful cut-off RLU values over which BLI will be possible for different machines. Using the Turner TD-20e luminometer 10 RLU/mg protein was insufficient, whereas values of 100 RLU/mg protein were high enough to allow BLI (data not shown). In addition to absolute levels of luciferase expression/mg lung protein spatial distribution of gene expression will play a significant role. Due to the comparatively low transfection efficiency of non-viral gene transfer agents we are currently unable to estimate transfection efficiency on a per cell basis and the effect of spacial distribution on BLI can, therefore, not be assessed at this point. Gene expression after DF/SeV or GL67 ‘‘sniffing’’ was not evenly distributed throughout the lungs with distribution varying from animal to animal. This may in part be explained by the asymmetrical bifurcation of the left and right bronchi (i.e. the left and right main stem bronchus radiate a different angles from the trachea) and the position of the animal during gene transfer agent administration and clearly illustrates the additional information that can be gained by this technique. Interestingly, DF/SeV transduced the left lung exclusively in 55% of mice compared to 26% transfected with GL67/ pCIKLux complexes. The reason for this is unknown, but may relate to the different consistency of virus solution and the more viscous GL67/pCIKLux complexes. Bolus administration of gene transfer agents to the nose followed by rapid sniffing into the lung allows only very brief
(seconds) contact time between the nasal epithelium and the gene transfer agent. DF/SeV-Lux but not GL67/pCIKLux ‘‘sniffing’’ led to high levels of bioluminescence in the nose further confirming our previous findings that in contrast to most other vectors SeV requires only very brief contact with the airway epithelium to allow virus entry [9]. Importantly, a viscoelastic gel formulation [8] in combination with slow perfusion of GL67/pCIKLux over the nasal epithelium ensured sufficient contact time for uptake of the liposome/ DNA complexes into the cells. This highlights contact time as a likely important factor to consider in clinical trials, as well as allowing BLI to be used in studies in CF mice, where the nose provides a convenient assay for changes in the underlying bioelectric defect. It is important to understand how this novel technique correlates with standard tissue homogenate-based assays, particularly since tissue depth, bones and fur are known to quench emitted light. Rettig et al. assessed correlations after hydrodynamic liver transfection with plasmid DNA and showed linearity between BLI and HBL expression over five orders of magnitude [21]. To the best of our knowledge correlations have not been assessed for lung and nose gene transfer. The correlation between left and right lungs was very high (r ¼ 0.9, p < 0.0001) after DF/SeV transduction, but weaker in the left lung (r ¼ 0.52, p < 0.005) and absent in the right lung (r ¼ 0.19, p ¼ NS) after GL67-mediated transfection. This implies that the degree of correlation is dependent on the overall level of gene expression, with low-level gene expression being more affected by signal quenching. Importantly, right and left lungs were affected to different degrees, which may be related to differences in anatomy. Thus, in contrast to the left lung, which is a discrete single lobe in mice, the right lung consists of four individual lobes which overlay each other and may, therefore, be more susceptible to tissue quenching of light emission. The correlation between light emission and gene expression in the nose was significant, but not particularly strong after both viral (r ¼ 0.68, p < 0.05) and non-viral (r ¼ 0.45,
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
1539
Fig. 5. Bioluminescence imaging in the nose after GL67/pCIKLux transfection. Mouse nasal tissue was transfected with pCIKLux (n ¼ 27) or a negative control plasmid (n ¼ 8) complexed to GL67 by nasal perfusion. (A) Bioluminescence was imaged in vivo 24 h after transduction. Representative images are shown. (B) Immediately after imaging the mice were culled and nasal tissue was harvested for homogenate-based luciferase detection. Photon emission and homogenate-based luciferase expression were correlated. Each symbol represents one animal.
p ¼ 0.01) gene transfer. Interestingly this was independent of the overall level of gene expression and may be due to (1) incomplete removal of all transfected tissue from the nasal cavity and/or (2) transfection of tissue outside the nose such as the mouth and snout due to overspilling and swallowing of vector during delivery. Firefly luciferase is one of a few proteins that are sensitive to anaesthetics and is partly inhibited by commonly used anaesthetics such as isofluorane, halothane, ketones, alkanes and ethers [22,23]. It has been suggested that anaesthetics compete with binding of luciferin to the protein. We cannot exclude that luciferase expression in our study was somewhat
reduced by the anaesthetic, but clearly all commonly used agents have this potential. 5. Conclusions SeV vector transduction and GL67-mediated lung gene transfer following topical administration via ‘‘sniffing’’ and nasal perfusion can be imaged in living animals and will, therefore, enable us to follow gene expression over time in the same animal. The correlation between light emission and absolute gene expression is dependent on the overall level of luciferase expression, and is better in the left than in the right lung.
1540
U. Griesenbach et al. / Biomaterials 29 (2008) 1533e1540
Acknowledgements We thank Luci Somerton for help with preparing the manuscript. This work was funded by the Cystic Fibrosis Trust and the CF Trust Dr Benjamin Angel Senior Fellowship (UG). References [1] Kutschka I, Kofidis T, Chen IY, von DG, Zwierzchoniewska M, Hoyt G, et al. Adenoviral human BCL-2 transgene expression attenuates early donor cell death after cardiomyoblast transplantation into ischemic rat hearts. Circulation 2006;114(Suppl. 1):I174e80. [2] Wang W, Kim SH, El-Deiry WS. Small-molecule modulators of p53 family signaling and antitumor effects in p53-deficient human colon tumor xenografts. Proc Natl Acad Sci U S A 2006;103(29):11003e8. [3] Wiles S, Pickard KM, Peng K, MacDonald TT, Frankel G. In vivo bioluminescence imaging of the murine pathogen Citrobacter rodentium. Infect Immun 2006;74(9):5391e6. [4] De A, Gambhir SS. Noninvasive imaging of proteineprotein interactions from live cells and living subjects using bioluminescence resonance energy transfer. FASEB J 2005;19(14):2017e9. [5] Sinn PL, Burnight ER, Hickey MA, Blissard GW, McCray Jr PB. Persistent gene expression in mouse nasal epithelia following feline immunodeficiency virus-based vector gene transfer. J Virol 2005;79(20):12818e27. [6] Wilber A, Frandsen JL, Wangensteen KJ, Ekker SC, Wang X, McIvor RS. Dynamic gene expression after systemic delivery of plasmid DNA as determined by in vivo bioluminescence imaging. Hum Gene Ther 2005;16(11):1325e32. [7] Ferrari S, Geddes DM, Alton EW. Barriers to and new approaches for gene therapy and gene delivery in cystic fibrosis. Adv Drug Deliv Rev 2002;54(11):1373e93. [8] Sinn PL, Shah AJ, Donovan MD, McCray Jr PB. Viscoelastic gel formulations enhance airway epithelial gene transfer with viral vectors. Am J Respir Cell Mol Biol 2005;32(5):404e10. [9] Yonemitsu Y, Kitson C, Ferrari S, Farley R, Griesenbach U, Judd D, et al. Efficient gene transfer to airway epithelium using recombinant Sendai virus. Nat Biotechnol 2000;18(9):970e3. [10] Konstan MW, Davis PB, Wagener JS, Hilliard KA, Stern RC, Milgram LJ, et al. Compacted DNA nanoparticles administered to the nasal mucosa of cystic fibrosis subjects are safe and demonstrate partial to complete cystic fibrosis transmembrane regulator reconstitution. Hum Gene Ther 2004;15(12):1255e69.
[11] Limberis MP, Wilson JM. Adeno-associated virus serotype 9 vectors transduce murine alveolar and nasal epithelia and can be readministered. Proc Natl Acad Sci U S A 2006;103(35):12993e8. [12] Sumner-Jones SG, Davies LA, Varathalingam A, Gill DR, Hyde SC. Long-term persistence of gene expression from adeno-associated virus serotype 5 in the mouse airways. Gene Ther 2006;13(24):1703e13. [13] Ferrari S, Griesenbach U, Shiraki-Iida T, Shu T, Hironaka T, Hou X, et al. A defective nontransmissible recombinant Sendai virus mediates efficient gene transfer to airway epithelium in vivo. Gene Ther 2004;11(22): 1659e64. [14] Lee ER, Marshall J, Siegel CS, Jiang C, Yew NS, Nichols MR, et al. Detailed analysis of structures and formulations of cationic lipids for efficient gene transfer to the lung. Hum Gene Ther 1996;7(14):1701e17. [15] Griesenbach U, Cassady RL, Ferrari S, Fukumura M, Muller C, Schmitt E, et al. The nasal epithelium as a factory for systemic protein delivery. Mol Ther 2002;5(2):98e103. [16] Ho TC, Jeng KS, Hu CP, Chang C. Effects of genomic length on translocation of hepatitis B virus polymerase-linked oligomer. J Virol 2000;74(19):9010e8. [17] Alton EW, Stern M, Farley R, Jaffe A, Chadwick SL, Phillips J, et al. Cationic lipid-mediated CFTR gene transfer to the lungs and nose of patients with cystic fibrosis: a double-blind placebo-controlled trial. Lancet 1999;353(9157):947e54. [18] Ruiz FE, Clancy JP, Perricone MA, Bebok Z, Hong JS, Cheng SH, et al. A clinical inflammatory syndrome attributable to aerosolized lipid-DNA administration in cystic fibrosis. Hum Gene Ther 2001;12(7):751e61. [19] Hirata T, Iida A, Shiraki-Iida T, Kitazato K, Kato A, Nagai Y, et al. An improved method for recovery of F-defective Sendai virus expressing foreign genes from cloned cDNA. J Virol Methods 2002;104(2): 125e33. [20] Gill DR, Smyth SE, Goddard CA, Pringle IA, Higgins CF, Colledge WH, et al. Increased persistence of lung gene expression using plasmids containing the ubiquitin C or elongation factor 1alpha promoter. Gene Ther 2001;8(20):1539e46. [21] Rettig GR, McAnuff M, Liu D, Kim JS, Rice KG. Quantitative bioluminescence imaging of transgene expression in vivo. Anal Biochem 2006;355(1):90e4. [22] Franks NP, Lieb WR. Do general anaesthetics act by competitive binding to specific receptors? Nature 1984;310(5978):599e601. [23] Zhang Y, Stabernack CR, Dutton R, Sonner J, Trudell JR, Mihic SJ, et al. Luciferase as a model for the site of inhaled anesthetic action. Anesth Analg 2001;93(5):1246e52.