Published ahead of print on February 23, 2006, doi:10.1165/rcmb.2005-0377OC
© 2006 American Thoracic Society DOI: 10.1165/rcmb.2005-0377OC Human-Specific Cystic Fibrosis Transmembrane Conductance Regulator Antibodies Detect In Vivo Gene Transfer to Ovine AirwaysMedical Sciences (Medical Genetics), University of Edinburgh, Western General Hospital, Edinburgh; The Wellcome Trust Centre in Comparative Respiratory Medicine, Easter Bush Veterinary Centre, University of Edinburgh, Roslin; Gene Medicine Group, Nuffield Department of Clinical Laboratory Sciences, University of Oxford, John Radcliffe Hospital, Oxford; UK Cystic Fibrosis Gene Therapy Consortium, London, Oxford, and Edinburgh, United Kingdom; Centro de Genética Humana, Instituto Nacional Saúde Dr Ricardo Jorge; and Department of Chemistry and Biochemistry, Faculty of Sciences, University of Lisbon, Lisbon, Portugal Correspondence and requests for reprints should be addressed to Heather Davidson, Medical Sciences (Medical Genetics), University of Edinburgh, Molecular Medicine Centre, Western General Hospital, Edinburgh EH4 2XU, United Kingdom. E-mail: H.Davidson{at}ed.ac.uk
A panel of 11 human cystic fibrosis transmembrane conductance regulator (hCFTR) antibodies were tested in ovine nasal, tracheal, and bronchial epithelial brushings. Two of these, G449 (polyclonal) and MATG1104 (monoclonal), recognized hCFTR but did not cross react with endogenous sheep CFTR. This specificity allows immunologic detection of hCFTR expressed in gene transfer studies in sheep against the background of endogenous ovine CFTR, thus enhancing the value of the sheep as a model animal in which to study CFTR gene transfer. Studies on mixed populations of human and sheep nasal epithelial cells showed that detection of hCFTR by these two antibodies was possible even at the lowest proportion of human cells (1:100). The hCFTR gene was delivered in vivo by local instillation using polyethylenimine-mediated gene transfer to the ventral surface of the ovine trachea and hCFTR mRNA and protein levels scored in a blinded fashion. Despite abundant hCFTR mRNA expression, the number of cells expressing hCFTR protein detectable by G449 was low (  0.006–0.05%). Immunohistochemistry for hCFTR in animals treated by whole-lung aerosol demonstrated positive cells in sections of tracheal epithelium and in distal conducting airways. The strategic use of hCFTR-specific antibodies supports the utility of the normal sheep as a model for hCFTR gene transfer studies.
Key Words: CFTR antibodies • cystic fibrosis • gene transfer • nasal brushing cells • sheep model
Cystic fibrosis (CF) is caused by mutations in the gene encoding the CF transmembrane conductance regulator (CFTR) protein. A variety of studies have shown evidence that CFTR functions as a cAMP-regulated chloride channel, whereas the tissue distribution of CFTR expression and localization of the protein in the apical membrane of epithelial cells is consistent with its involvement in transepithelial fluid and electrolyte transport (1–3). The majority of mutations found in CFTR are associated with characteristic airway disease, pancreatic insufficiency, male infertility, and elevated levels of sodium chloride in sweat (4). The most common lethal mutation in white populations is a 3-bp deletion resulting in the loss of F508 (  F508 CFTR) and is found in 70% of patients with CF (5). CFTR has also been implicated in the regulation of other apical membrane conductance pathways through interactions with the amiloride-sensitive epithelial sodium channel and the outwardly rectifying chloride channel (6, 7). CFTR has a highly regulated pattern of expression in the lung (1, 8, 9). This affects the level of protein expression within different cell types and the location of these CFTR-expressing cells in the lung. Until recently, the targets for gene therapy were thought to comprise ciliated, nonciliated, and goblet cells in the airway surface epithelium and epithelial cells lining the submucosal glands within the interstitium of the airways. However, a recent study (1) in lungs from a large cohort of normal subjects found that there was little CFTR expression in the superficial, gland acinar, and alveolar epithelia, whereas apically localized CFTR was found within the superficial epithelium and gland ducts. Although its findings are controversial, this study provides important new data on CFTR expression. With opinion divided regarding CFTR localization, the safest course is for gene therapy protocols to be directed toward targeting all cells implicated in the pathogenesis of CF airway disease (8).
Experiments with CFTR-overexpressing cells and transformed cell lines present strong evidence that wild-type CFTR is directed to the apical membrane of polarized epithelial cells via the Golgi and trans-Golgi networks after polypeptide synthesis (10). This is an inefficient process because 70% of the newly synthesized protein is retained in the endoplasmic reticulum where it is degraded via the ubiquitin/proteasome pathway (11, 12). It has been generally accepted that most, if not all, of the
Data from experiments in vivo have shown mislocalization of The sheep has been proposed as a novel model for lung-directed CFTR gene therapy (22) because the sheep lung is anatomically similar to the human lung (23). Ovine CFTR (ovCFTR) has a similar pattern of expression to the human gene, with 90.7% sequence identity within the coding region (24). There is no CF sheep model. Nevertheless, the similarities to humans in lung physiology and architecture are proving to be invaluable in the assessment of gene therapy efficacy, the localization of transgene expression, and the safety of the gene transfer protocol, all crucially relevant endpoint measurements (22). Delivery of gene therapy agents (GTAs) and vectors to sheep has been restricted to reporter gene constructs. This study focuses on specific immunocytochemical detection of hCFTR performed in parallel on brushings from sheep and human nasal epithelial cells (SNEs and HNEs) to assess the suitability of a panel of anti-hCFTR Abs for ovine gene therapy studies using hCFTR expression constructs. As a prerequisite, we measured and compared the epithelial content and nature of SNEs versus HNEs. We aimed to ascertain whether ovCFTR displays similar patterns of apical localization in SNEs as hCFTR does in HNEs and whether any of the available anti-hCFTR Abs could discriminate between HNEs and SNEs. This would potentially enable in vivo gene transfer studies of hCFTR to sheep without cross reactivity from endogenous ovCFTR. The data presented here demonstrate that specific anti-hCFTR antibodies can detect vector-derived hCFTR in sheep trachea after instillation of plasmid DNA/polyethylenimine (pDNA/PEI) complexes or in cryosections of sheep airways after whole lung aerosol delivery of DNA/PEI complexes.
Nasal Brushings: Collection and Processing Human nasal brushing cells were obtained from healthy individuals. Ovine nasal brushings were obtained from adult greyface ewes. Sheep turbinates were dissected out, transported in PBS, washed several times in PBS, and brushed using an interdental brush (Interdental brush 620; Dent-O-Care, London, UK) applied to the inferior turbinate and lateral nasal wall. Cells were collected into 1 ml of cold PBS and removed from the brush using a cut-off 200-µl pipette tip. Human nasal brushings were obtained after informed consent was obtained. Nasal cavities were cleansed by lavage using 20 applications of 125 µl isotonic saline (0.9%) directly to the inferior turbinate. Nasal brushing was performed using an interdental brush applied to the inferior turbinate and lateral nasal wall and collected as for the sheep. Nasal brushings from human and sheep were harvested, fixed, and adhered to silane glass slides by cytospin as described (21).
Nasal Brushings: Cell Type Composition and Viability A live/dead viability/cytotoxicity kit (L-3224; Molecular Probes, Leiden, The Netherlands) was used to measure cell viability in the epithelium from sheep and human samples. Fresh nasal brushings (50 µl), which had been collected in 1 ml of PBS, were pipetted onto a microscope slide. Fifty microliters of the fluorescence solution (2 ml PBS, 1 µl calcein AM, 4 µl ethidium homodimer-1) was added to the cell solution on the slide and gently covered with a coverslip. After 10 min in the dark, the slides were counted and scored for viability using the Axioskop fluorescence microscope (Zeiss) using FITC (calcein AM = live cells) and Texas Red filters (EthD-1 = dead cells).
Abs
Immunocytochemistry The same immunocytochemistry (ICC) protocol (21) was applied for all Abs used. Briefly, cytospin prepared cells were permeabilized with 0.25% or 1% (vol/vol) Triton X-100 in PBS for 20 min, washed three times in PBS for 5 min each, and antigen blocked with 1% (wt/vol) BSA/PBS for 30 min before overnight incubation at 4°C with anti-CFTR or control Ab. After PBS washes (3 x 10 min), cells were incubated with the relevant secondary Ab for 45 min at room temperature (RT) and washed in PBS (3 x 10 min). For double labeling, the first primary (overnight incubation at 4°C) and secondary Abs (45 min at RT) were added sequentially with PBS washes (3 x 10 min) in between followed by the second primary (1 h at 37°C) and secondary Abs (45 min at RT). The slides were mounted with Vectashield (Vector Laboratories, Peterborough, UK) containing 4,6-diamino-2-phenylindole (DAPI) for nuclear staining and covered with a glass coverslip. Optimization experiments included the following variations in the washing steps: 1% Tween or ASB14 in PBS; in the collection of cells and fixing steps: Dithiothreitol (DTT) concentrations from 1 mM, 10 mM, and 100 mM; and in the permeabilization step: Triton concentrations from 0, 0.2, 0.5, 1, and 2% (see RESULTS). Immunofluorescence staining was observed and collected on an Axioskop fluorescence microscope (Zeiss) with the Photometrics Coolsnap HQ camera and images captured using the Smartcapture2 (Digital Scientific, Cambridge, UK) software.
Animals and Anesthesia All research adhered to the UK Home Office Animals (Scientific Procedures) Act 1986. Female BALB/c mice (8–16 wk old) were placed into a perspex whole-body exposure chamber (24.6 x 24.6 x 13.8 cm) where they remained unrestrained for the duration of the aerosol delivery. Commercially sourced crossbred sheep (body weight, 35–50 kg) were treated with anthelminthic therapy before entry into the study and were judged to be free from significant pulmonary disease on the basis of clinical examination. The study protocol was approved by the UK Home Office Licencing Authority. Sheep were anesthetized and ventilated as described previously (26). Briefly, after intravenous administration of a single bolus of thiopentone sodium (Intraval sodium; Merial Animal Health Ltd., Harlow, Essex, UK), sheep were intubated and maintained using gaseous halothane (2–3%) in oxygen and nitrous oxide. The sheep were placed in sternal recumbency in a whole-body respirator. The proximal end of the endotracheal (E/T) tube was connected to the anesthetic circuit through a connector in the wall of the respirator. Pressure in the box was varied by a large bellows pump (Cuirass; Cape Warwick, Warwick, UK), which induced a sinusoidal tidal respiratory pattern. A tidal volume of 10 ml/kg body weight was maintained, and respiratory rate was adjusted to maintain end-tidal CO2 measurements in the range of 4.5–5.5% (Oxicap Monitor Model 4700; Ohmeda, Louisville, CO).
Gene Delivery to the Trachea
Tracheal Epithelial Cell Sample Collection and Processing
Aerosol Gene Delivery to the Whole Lung For aerosolization of GTA to the sheep lung, 32 mg of pCIKCFTR or pCIKLux (28) was complexed with PEI at a N/P ratio of 10:1 (29) to give a final concentration of 0.4 mg/ml DNA. The final concentration was double that used in the trachea to reduce the volume and delivery time in order to maximize delivered dose. Studies of aerosol delivery to the mouse lung indicated that this formulation was equally active when delivered by aerosol (personal communication, Lee Davies). Four separate aliquots of 10 ml plasmid DNA solution (0.8 mg/ml in sterile endotoxin-free dH2O) were added dropwise to each of four 10-ml aliquots of PEI solution containing 0.3 µl of a stock 4.3 mg/ml solution (0.1 M N) per 1 µg plasmid DNA as described previously. The total volume of GTA was 80 ml.
Anesthetic gas, delivered to the inspiratory limb of a circle system, was entrained in equal proportions through three jet nebulizer devices (Pari LC Plus) mounted on a manifold and connected to the E/T tube. A compressed gas source of medical air (22 psi) and timer-controlled solenoid valve triggered by the start of each breath was used to operate the nebulizers during the first 0.8 s of each inspiration. Exhaled air was passed through a filter (Pari filter set) before removal of CO2 in soda lime. A rebreathing bag and scavenging system was provided for peak inspiratory flow requirements and venting of excess gas. Each nebulizer was charged with 8 ml of GTA and weighed before delivery. The nebulizers were weighed after
Collection of Airway Tissue for Immunohistochemistry
Three lung segments (RA, RC, and LC) for immunohistochemistry (IHC) were inflated with a mix of 30% sucrose: OCT (2:1) and partially frozen on dry ice to improve our ability to slice the tissue transversely into
IHC
Luciferase Assay
TaqMan Quantitative PCR Assay
RNA was heated to 75°C for 5 min and reverse transcribed with TaqMan RT reagents (Applied Biosystems). The RT-reaction mix (5 µl) consisted of 5.5 mM MgCl2, 500 µM of each dNTPs, 0.4 U/µl RNase inhibitor, 1.25 U/µl Multiscribe Reverse Transcriptase, 0.4 µM pCI reverse primer, 0.4 µM reverse rRNA primer, and
Controls included no template control and no-reverse transcriptase control where total RNA or MultiScribe reverse transcriptase and RNase inhibitor were omitted from the reverse transcriptase reaction, respectively. Relative levels of plasmid–derived, CAT-specific mRNA were determined using the
Characterization of Sheep Versus Human Nasal Brushing Samples The May-Grünwald-Giemsa–stained SNE slides were characterized visually for epithelial content and compared with the HNE slides. Morphologically, the appearance of SNE and HNE cells were similar, and a differential count of ciliated, nonciliated, and basal cells was performed for each species (Table 2). The SNE and HNE slides were labeled with anti-CK18, which allows calculation of the epithelial content of the sample, compared with total DAPI-stained cells, before scoring the ciliated, nonciliated, and basal cell content within that epithelial population. Tubulin (anti-tubulin,  and  subunits) staining of ciliary structures showed an identical pattern of expression for both populations of cells. These results indicate that regarding morphology, epithelial content, and expression of relevant control genes, HNEs and SNEs seem to be comparable (Table 2). Tracheal and bronchial sheep cells were also subjected to the same staining and counting procedures. Some differences were observed in the basal cell count and the columnar ciliated cell count between human and sheep (Table 2).
Viability of human and ovine nasal epithelial cells was assessed immediately after collection from the nose using the live/dead cytotoxicity kit. For both species, the percentage of live cells in a brushing at collection was 60% of the sample.
Optimization of CFTR ICH Method Modifications to the standard protocol tested were 1% Tween in the wash steps, replacing 1% Tween with 1% ASB14 (Calbiochem), varying concentrations of Triton (0, 0.2, 0.5, or 1.2%) in the permeabilization step, and varying concentrations of DTT (1, 10, or 100 mM) in the PBS sample collection tube and in the fixation solution. The latter was performed to reduce the nonspecific cilia labeling, which we attributed to Ab binding to mucus adhering to the cilia. Samples were also collected and stored at 37°C in PBS until the fixation stage because correctly localized protein expression may be affected by our usual practice of storage at 4°C. The addition of DTT at any concentration had no discernible effect on Abs that displayed cilia labeling. The collection of samples at 37°C also failed to improve results. Although the addition of Triton at various concentrations in the permeabilization step or the presence of Tween in the washing stages did result in altered labeling patterns, there was no consequent improvement in the variability in the signal of several Abs in HNEs and SNEs (e.g., for MAB3484, see Figures 1A and 1B). Likewise, ASB14 did not alter the variable signal localization observed for several Abs. However, for the Lis Ab, apical localization was evident under all conditions, and an increase in concentration from 0.25–1% Triton did give a sharper apically localized band (Figures 1C and 1D). In summary, the only alteration to the original protocol that we found useful was the increase from 0.25–1% Triton in the permeabilization step.
Comparison of Ovine Versus Human Labeling with Anti-CFTR Abs HNEs and SNEs were centrifuged (Cytospin 2; Shandon, Cheshire, UK) onto the same slide to enable direct comparison of each Ab. Recognition of apically localized CFTR was tested in HNEs and SNEs with MAB25301, MAB1660, MP-CT1, LIS, M3A7, G449, MATG1061, MATG1104, L12B4, MM13-4, and CC24-R (Table 1). M3A7, L12B4, and MM13-4 were discarded after an initial experiment in which they demonstrated intracellular labeling of the SNEs. In a separate study, the apical versus intracellular labeling of human cells from healthy individuals with these Abs was found to vary in an individual-specific manner, making interpretation of results difficult (30). MAB25301, MAB1660, CC24-R, MPCT1, and MATG1104 gave apical signal but displayed inconsistent or weak labeling, which was sometimes accompanied by additional ciliary or intracellular labeling. Lis and MATG1061 gave a clear apically localized signal in HNEs and SNEs (Figures 2A, 2D, 2E, and 2H). G449 and MATG1104 reacted positively with HNEs and negatively for SNEs (Figures 2B, 2C, 2F, and 2G). Experiments with G449, MATG1104, and MATG1061 were repeated using double the normal antibody concentration. G449 and MATG1104 were negative on SNEs and strongly positive in HNEs, and MATG1061 was positive for both species. The pattern of reactivity of these four Abs to sheep tracheal and bronchial cells was the same as for nasal cells (Figures 2Q–2T). Double-labeling of HNEs and SNEs with tubulin (anti-tubulin, and subunits) and G449 shows clear apical membrane and ciliary signal in HNEs (Figures 2I and 2J), whereas in SNEs the only visible signal is ciliary (Figures 2K and 2L).
Alignment of the human R domain peptide sequences used to raise the Abs with the corresponding ovine sequences reveals that the MATG1104 peptide (722-734) has three consecutive amino acid differences (EED in human and DGA in ovine), and the G449 peptide (653-716) has nine differences, three of which involve structure-critical proline residues (Figure 3). These amino acid differences are likely to result in conformationally distinct protein regions and may explain the reduced affinity of these Abs for ovCFTR.
The reliability, consistency of performance, and availability in sufficient quantities of G449 informed our choice of G449 Ab over MATG1104 Ab for ovine studies. Therefore, further characterization of the G449 was performed to compare it with MATG1061 Ab in HNEs. Three separate experiments using three healthy individuals were performed on single (MATG1061 and G449)-labeling of HNEs, and one experiment was performed where the HNEs were double-labeled with MATG1061/G449. The apically localized signal from cells was counted in each case (Table 3, Figures 2M–2O). The counted cells from MATG1061 and G449 were similar and comparable (Table 3). Moreover, the double-labeled cells indicated that the same cells were positive for both antibodies (Figures 2M–2O) and when double-labeled with MATG1061 and Lis (Figure 2P). Because MATG1061 is a mAb raised to the NBD1 domain, G449 is raised to the R domain, and Lis is raised to the C-term, it is unlikely that the signal seen is nonspecific.
Detection of hCFTR in Mixed Populations of HNEs and SNEs To assess the sensitivity of detection of hCFTR-expressing cells, we performed an SNE/HNE mixing experiment. G449 was tested on mixed populations of SNE/HNE containing 50, 10, or 1% HNEs. After optimization experiments, we devised the following protocol. Cell counts were averaged from four independent counts, and cytospins were made of four duplicate samples of each combination containing a total of 1.5 x 104 cells per cytospin. Twenty defined regions of interest per sample were counted using a x40 lens (Table 4). The total and differential cell count per region of interest was determined (see MATERIALS AND METHODS), and the percentage of columnar cells with apically localized FITC or Alexa Green luminescence was calculated. Our HNE samples contained 72–79% columnar epithelial cells, of which 56.8% and 59.8% showed apical localization with the G449 and MATG1104 Ab, respectively. These figures corroborate those of Penque and colleagues (21) using Ab 169b. If only 56.8% (G449, n = 6) or 59.8% (MATG1104, n = 6) of human columnar cells have apically localized signal, then the combination of SNEs mixed with 50%, 10%, or 1% HNEs, should display 28.4%, 5.6%, and 0.56% positive apically localized signal, respectively. Our averaged results were 38%, 2.2%, and 0.77% for G449 (n = 4) and 29.25%, 5.8%, and 1.4% for MATG1104 (n = 4), in good agreement with the estimates. The number of cells counted for each of the parameters for each experiment varied from 500–1500 cells per sample, depending on cell density and sample condition. These results imply that it should be possible to detect hCFTR transfection levels as low as 1% in SNEs (Table 4).
In Vivo Gene Delivery, Expression, and Detection in Ovine Samples Initially G449 was used for detection of hCFTR in vivo after gene transfer to the ventral surface of the sheep trachea only (Table 5). For each experiment, a pretreatment brushing sample was divided into two aliquots; one was the known negative control, and the other was included among the blinded samples. Samples from the first animal were used to establish brushing procedures and were not analyzed in a blinded manner. The low level of positive, apically localized signal indicated very low transfection efficiencies (0.006–0.008%) (Table 5, Experiment 1). Plasmid-derived mRNA was observed in post-treatment samples and was almost undetectable by 1 wk (data not shown). The second animal received double the dose of DNA/PEI (2 mg DNA), and samples from this animal were analyzed under blinded conditions. Transfection efficiencies were very low (0.006–0.012% of cells classed as positive) (Table 5, Experiment 2), although treated samples were positive for transgene RNA (Figure 4A). One week later, the same animal was treated with identical doses of PEI/DNA on two consecutive days, with samples collected 24 h and 48 h after the second dose. Levels of mRNA detected in these were comparable to those seen from the single dose (compare Figures 4A and 4B). The "pretreatment" samples for this set would retain traces of transgene expression from the original treatment, thus explaining the presence of plasmid-derived mRNA in the pretreatment sample (Figure 4B). Immunochemistry samples were analyzed in a blinded manner, and most transfection efficiencies were low. However, the sample collected 24 h after the second dose gave higher staining counts (0.027% highly positive, 0.05% moderately positive) (Table 5, Experiment 3). Examples of positively staining brushed cells are shown in Figures 5A and 5B. The post-treatment samples from a further two blinded experiments using a double dose 24 h apart and sampled at 24 h post-treatment were scored as transfected at comparable levels (Table 5, Experiments 4 and 5). The number of cells assayed for each staining experiment varied between 6 x 104 cells/sample and 2.4 x 105 cells/sample, depending on the number of cells taken from a brushing.
To test the assay under nebulization conditions similar to those envisaged for clinical use, two sheep were treated by whole-lung aerosol via the Pari LC Plus nebulizer. It has been demonstrated that aerosolization of pDNA/PEI complexes (N/P of 10:125 µg/ml DNA) generates an aerosol with an MMAD of 3.4 µm (31). This figure is similar to other published data for the Pari LC Plus, and the droplet size falls into the respirable range for human lung deposition (< 5 µm). Figure 6 shows the data from three separate aerosol delivery experiments with pDNA/PEI (0.2 mg/ml pDNA) in mice. This demonstrates that the pDNA/PEI complexes are stable after aerosolization and that we can achieve reproducible levels of gene expression in mouse lungs using the Pari LC plus nebulizer. In our large animal, negative-pressure system, the aerosolized material enters the lung via an E/T tube. Gravimetric analysis of filters placed at the distal end of the E/T tube suggests that 30% of the dose in the nebulizers reaches the end of the E/T tube (data not shown). Animals were treated with 32 mg pCIKCFTR with PEI or 32 mg pCIKLux with PEI in a total volume of 80 ml over 5 h. IHC on bronchial brushings taken from these sheep 24 h after dosing showed negligible positive signal. However, IHC on tracheal rolls embedded in OCT and sectioned by cryostat onto slides showed a marked increase in hCFTR-positive cells assayed by G449 antibody (Table 5, Experiment 6; Figures 5C–5F). Some positive signal was observed in cells from pretreated samples and in samples from the pCIKLux-treated animal, but these signals occurred at a much lower frequency than the signal observed from pCIKCFTR-treated samples. All tracheal preparations were scored on a x16 lens, and only bright cells that were distinguishable above the general background signal were counted as positive. Some positively stained cells were observed during scoring of cryosections from lung tissue blocks containing smaller airways. Although the absolute number of positive cells was low, they were present at a higher frequency in the pCIKCFTR-treated animal compared with sheep treated with pCIKLux (Table 5, Experiment 6; Figures 5G–5I). Further analysis of the positively transfected lung sections by double-labeling with diluted MATG1061 and normal concentrations of G449 (see MATERIALS AND METHODS and Figures 5J–5L) showed that each antibody gave a positive signal in the same areas in the bronchi of the putative positive sections. Untreated sections were scored as negative. Levels of plasmid derived mRNA were measured by TaqMan analysis. Because the common target sequence for the TaqMan primers is upstream of the coding sequences of pCIKLux and pCIKCFTR, the same assay was used for both plasmids. Significant levels of transgene mRNA were detected in all tissue samples tested from the aerosol-treated animals (Figure 7A).
Tissue lysates and bronchial brushings collected into reporter lysis buffer from the two whole-lung aerosol treated animals were assayed for luciferase activity. Figure 7B shows lux protein results from samples from segment RVD1 of both animals and from pooled bronchial brushings. All the positive lux samples were derived from the animal treated with pCIKLux.
Although several knockout mouse models exist for CF, there is variation in the phenotypes of the different models (32) and also in the degree to which they mimic human disease, particularly in the lung. In addition, the anatomies of human and mouse lungs are divergent, especially with respect to submucosal glands (33). This, together with the significant problem of scale, makes it less attractive to use the mouse as a preclinical model for lung gene delivery. This especially applies to the development of aerosolized gene delivery. Despite the absence of an ovine CF disease model, we believe that the structural and physiologic similarities of the ovine and human lungs allows better assessment of safety and is more predictive of the efficacy of gene therapy formulations than the murine model. We would like to be able to test vectors expressing hCFTR to evaluate whether protein is correctly processed and trafficked to the apical membrane in the correct cell type. However, the background of endogenous ovCFTR expression has meant that studies have been limited to the use of reporter genes to assess gene transfer. The proportions of cell types found in brushings from the three different sites of the sheep airway were similar. These were slightly different from the composition of the HNEs, which had a lower columnar ciliated count and higher basal cell count. Only 60% of the cells in a human or sheep brushing are viable when assessed immediately after capture. The 40% that are not viable may have been damaged during brushing or were dying as a result of normal epithelial cell turnover. Although others have investigated cilia beat and structure of HNEs after brushing, information on cell viability straight after nasal brushing is not available (34). In optimizing the labeling of these brushed cells with the panel of anti-hCFTR antibodies, we have achieved a level of consistency in the performance of each Ab under various experimental conditions. Several of these (including Lis, MATG1061, and G449) performed more consistently than others, giving clear apical localization. A second group (MP-CT1, MAB25301, MAB1660, CC24-R, and MATG1104) give apical localization, although the signal was occasionally weak (with MATG1104, MAB25301, and CC24-R) or accompanied by additional ciliary or intracellular labeling (with MP-CT1, MAB25301, and MAB1660). We found the reproducibility in CFTR localization given by a third group of Abs (M3A7, L12B4, and MM13-4) to be lower despite the good performance of these Abs in western blots and immunoprecipitations (35). In experiments using mixed populations of HNE/SNE, hCFTR was easily detected using G449 when as few as 1% human cells were present. This suggests that hCFTR detection may be possible even with gene therapy protocols that achieve only modest transfection levels. The highest transfection efficiency of 0.065% was obtained in samples from ovine tracheas that had received doses of PEI/pCIKCFTR on two consecutive days and were assayed 24 h after the final treatment, perhaps indicating that increased contact time and increased dose are important factors for airway gene transfer. In whole-lung aerosol experiments, sections from rolled up tracheal epithelium, which allow efficient screening of large numbers of cells, provided the clearest evidence of gene transfer (Figures 5C and 5D). These results demonstrate the inefficiency of transfection of differentiated airway epithelium in vivo with PEI/pDNA but could also imply that IHC is relatively insensitive for the detection of transfected cells that may be expressing low levels of transgene CFTR. In these studies we also observed positive signal in a low percentage of cells in pretreatment samples or those from the pCIKLux treated animal. We speculate that these could represent false positives due to higher background in some samples or rare cells expressing ovCFTR at much higher than mean levels. In these studies we have measured transgene expression at the protein and mRNA level. There were quantitative discrepancies between the amounts of mRNA measured and levels of protein for CFTR by IHC and for luciferase activity, but it is difficult to speculate on the reasons for this when we have n = 1. Figure 4 shows persistence of mRNA at 48 h post-treatment despite the fact that there is little evidence for CFTR protein at 48 h. This may suggest that we need to examine the relationship between plasmid-derived mRNA and fully processed and correctly localized CFTR protein more carefully. There may be more variability in the luciferase activity data due to factors affecting the recovery of luciferase protein in the tissue lysate preparation, particularly when the levels are as low as in our study.
SNEs and ovine tracheal and bronchial cells display the same spectrum of reactivity to the range of anti-hCFTR Abs tested. At standard concentrations, G449 and MATG1104 generate virtually no background staining in ovine airway epithelial samples, whereas Abs consistently stain CFTR in human samples and can detect the lower level of the preexisting hCFTR signal at the apical membrane of
Previously published data indicate that the percentage of cells positive for apically localized CFTR in tall columnar cells varies from 22% for Following the protocols described, we are confident that accurate data for apical localization in experimental samples can be obtained with MATG1061, Lis, and G449. How this increased percentage of apically localized CFTR correlates with correction of CF, at the RNA level or in terms of functional epithelial chloride transport, has not been determined. In our ovine model, with the two human-specific Abs G449 and MATG1104, we can directly measure transfection efficiency as correctly localized CFTR protein and not as an augmented signal as would occur in human gene transfer samples. It is thought that many nonviral formulations achieve low transfection efficiencies in conducting airway epithelial cells due to the lack of cell division of these differentiated cells and the barriers that exist which are peculiar to the lung milieu (e.g., mucociliary clearance, mucus, airway surface liquid, tight junctions, and immune systems). However, in these studies we have demonstrated that even at these low transfection efficiencies we can detect apically localized hCFTR protein in columnar epithelial cells. We have demonstrated that the normal sheep lung can be used for verification of CFTR gene transfer with hCFTR protein as an endpoint. In summary, this study further supports the utility of the normal sheep lung as a model for determining the safety and efficacy of hCFTR gene transfer to the airway epithelium (22) and illustrates how careful selection of antibodies and optimization of IHC techniques can lead to robust methodologies for assessing gene transfer by direct visualization of transfected cells in the lung.
The authors thank Catherine Gordon, Alison Baker, Peter Tennant and Paul Wright for their excellent technical and animal care assistance and Simon Cooper for assistance in manuscript preparation. The authors thank Hugo de Jonge, Erasmus University, Rotterdam, Netherlands; Angus Nairn, Yale University; Robert Dormer, University of Wales, Cardiff, UK; and Transgene, Strasbourg, France for their kind gifts of various Abs.
This work was funded by the UK CF Trust and the UK Medical Research Council and FCT/FEDER 9P/SAU/55/96/Portugal, which financed the production of the Lis antibody. H.D. was a recipient of a travel grant from the European CF Network (EU-QLK3-1999-00241). Originally Published in Press as DOI: 10.1165/rcmb.2005-0377OC on February 23, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form October 7, 2005 Accepted in final form February 6, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
