ORIGINAL ARTICLE Pre-clinical evaluation of three non-viral gene transfer agents for cystic fibrosis after aerosol delivery to the ovine lung G McLachlan 1,7 , H Davidson 2,7 , E Holder 2,7 , LA Davies 3,7 , IA Pringle 3,7 , SG Sumner-Jones 3,7 , A Baker 1,7 , P Tennant 1,7 , C Gordon 1,7 , C Vrettou 1,7 , R Blundell 4,7 , L Hyndman 2,7 , B Stevenson 2,7 , A Wilson 2,7 , A Doherty 2,7 , DJ Shaw 1 , RL Coles 7 , H Painter 3,7 , SH Cheng 5 , RK Scheule 5 , JC Davies 6,7 , JA Innes 2,7 , SC Hyde 3,7 , U Griesenbach 6,7 , EWFW Alton 6,7 , AC Boyd 2,7 , DJ Porteous 2,7 , DR Gill 3,7 and DDS Collie 1,7 We use both large and small animal models in our pre-clinical evaluation of gene transfer agents (GTAs) for cystic fibrosis (CF) gene therapy. Here, we report the use of a large animal model to assess three non-viral GTAs: 25kDa-branched polyethyleneimine (PEI), the cationic liposome (GL67A) and compacted DNA nanoparticle formulated with polyethylene glycol-substituted lysine 30-mer. GTAs complexed with plasmids expressing human cystic fibrosis transmembrane conductance regulator (CFTR) complementary DNA were administered to the sheep lung (n¼8 per group) by aerosol. All GTAs gave evidence of gene transfer and expression 1 day after treatment. Vector-derived mRNA was expressed in lung tissues, including epithelial cell-enriched bronchial brushing samples, with median group values reaching 1–10% of endogenous CFTR mRNA levels. GL67A gave the highest levels of expression. Human CFTR protein was detected in small airway epithelial cells in some animals treated with GL67A (two out of eight) and PEI (one out of eight). Bronchoalveolar lavage neutrophilia, lung histology and elevated serum haptoglobin levels indicated that gene delivery was associated with mild local and systemic inflammation. Our conclusion was that GL67A was the best non-viral GTA currently available for aerosol delivery to the sheep lung, led to the selection of GL67A as our lead GTA for clinical trials in CF patients. Gene Therapy (2011) 18, 996–1005; doi:10.1038/gt.2011.55; published online 21 April 2011 Keywords: gene delivery; lung; non-viral; sheep; cystic fibrosis; aerosol INTRODUCTION Proof-of-principle for gene transfer has been demonstrated in a number of clinical trials of gene therapy for cystic fibrosis (CF). 1 The level of gene transfer needed to achieve clinical improvement is unclear from these studies, which use molecular markers as outcome measures. The aim of the UK Cystic Fibrosis Gene Therapy Consortium is to take the optimal currently available non-viral formulation into a multi-dose trial assessing, for the first time, clinical end points. In preparation for this, we wished to identify one or more non-viral gene transfer agents (GTAs) with a safety and efficacy profile suitable for use in future clinical trials. The focus on non-viral GTAs is based on current limitations of most viral GTAs, which show loss of efficacy on repeat administration. Animal models are essential for pre-clinical evaluation of prospec- tive GTAs because in vitro studies are often poorly predictive of in vivo efficacy and fail to model the influence of an intact immune and inflammatory response. Studies in cystic fibrosis transmembrane conductance regulator (Cftr) mutant mice are thought to be useful for testing whether a gene therapy vector can rescue the defects associated with lack of fully functional CFTR in the airway epithelium, such as loss of cyclic adenosine monophosphate-dependent chloride channel activity and/or regulation of epithelial sodium absorption via the sodium channel EnaC. 2–4 However, studies in the mouse alone are likely insufficient to predict clinical efficiency in CF patients because there are significant differences in disease pathology and airway physiology between mouse and human. Although large animal CF models continue to be developed 5–7 these are not yet widely available; thus, we have successfully used wild-type sheep to evaluate the safety and efficacy of lung-directed gene therapy in a clinically relevant way. 8–10 Our choice of species is based on anatomical and functional similarities to human lungs and the observation that similar mechan- isms are initiated in response to lung inflammation that follow a familiar pattern of repair and regeneration. Indeed, the sheep model is proving valuable for assessing gene delivery and efficacy, the localisa- tion of transgene expression and the safety of the gene transfer protocol, all crucial end point measurements in human trials. Practi- calities relating to animal numbers and costs prevent the use of the ovine model for high throughput screening, therefore we adopted a combined strategy for testing potential GTAs in vivo. GTAs that show promise in initial testing in the nasal epithelium of wild-type mice and Received 25 November 2010; revised 4 March 2011; accepted 7 March 2011; published online 21 April 2011 1 Developmental Biology Division, The Roslin Institute and R(D)SVS, EBVC, University of Edinburgh, Roslin, UK; 2 Medical Genetics Section, School of Molecular and Clinical Medicine, University of Edinburgh, Western General Hospital, Edinburgh, UK; 3 Gene Medicine Group, Nuffield Department of Clinical Laboratory Sciences, University of Oxford, John Radcliffe Hospital, Oxford, UK; 4 Department of Veterinary Pathology, College of Medicine and Veterinary Medicine, University of Edinburgh, Edinburgh, UK; 5 Genzyme Corporation, Framingham, MA, USA and 6 Department of Gene Therapy, Faculty of Medicine, Imperial College London, London, UK Correspondence: Dr G McLachlan, The Roslin Institute and R(D)SVS, University of Edinburgh, Easter Bush Veterinary Centre, Roslin EH25 9RG, UK. E-mail: gerry.mclachlan@ed.ac.uk 7 The UK Cystic Fibrosis Gene Therapy Consortium (http://www.cfgenetherapy.org.uk/). Gene Therapy (2011) 18, 996–1005 & 2011 Macmillan Publishers Limited All rights reserved 0969-7128/11 www.nature.com/gt