Cationic liposomes, 1:1 (mol/mol) 1,2-dioleoyldimethylammonium chloride-1,2-dioleoyl-sn-glycero-3-phosphoethanolamine, were used to transfect primary cultures of distal rat fetal lung epithelial cells with pCMV4-based plasmids. A DNA-to-lipid ratio of 1:10 to 1:15 (wt/wt) optimized DNA uptake over a 24-h exposure. At a fixed DNA-to-lipid ratio of 1:15, chloramphenicol acetyltransferase (CAT) reporter gene expression declined at lipid concentrations > 2.5 nmol/cm2 cell surface area, whereas DNA uptake remained concentration dependent. CAT expression peaked 48 h after removal of the liposome-DNA complex, declining thereafter. Reporter gene expression was increased, and supercoiled cDNA degradation was reduced by the addition of 0.2 mM nicotinamide and 10 μM chloroquine. Rat fetal lung epithelial cells transfected with two different expression cassettes had an increased susceptibility to superoxide-mediated cytotoxicity. This could be attributed to a nonspecific delivery of exogenous DNA or some other copurified factor. The DNA-dependent increase in superoxide-mediated cytotoxicity, but not basal levels of cytotoxicity, was inhibited by the addition of 0.2 mM nicotinamide and 10 μM chloroquine.
- deoxyribonucleic acid
- gene transfer
- reactive oxygen species
the lungs of premature infants are at particular risk from pulmonary O2 toxicity. Not only may their lung tissue be subjected to elevated O2 for prolonged periods, which will result in an increased production of superoxide and other partially reduced oxygen species, but the immature lung also has deficient antioxidant enzyme activities at the time of birth (30). Tanswell and colleagues previously showed, both in vivo (29) and in vitro (31), that treatment of fetal or neonatal rat lung cells with antioxidant enzymes is protective against O2-mediated cytotoxicity. Such protein therapy is a useful experimental tool, but there are significant practical limitations to this approach as a therapeutic intervention (30). Theoretically, therefore, this population may eventually become an appropriate candidate for the use of transient antioxidant enzyme gene transfer administered via the airway route. The feasibility of this approach has been demonstrated in adult animals transfected with the human α1-antitrypsin gene (6), whereas successful liposome-mediated transfection of the nasal epithelium of patients suffering from cystic fibrosis has been recently reported (7). As the first step toward studies using liposome-mediated delivery of antioxidant enzyme gene constructs in vivo, we have used reporter gene constructs to establish an in vitro model system for the transfection of primary epithelial cell cultures from the preterm lung.
Transfection of established cell lines with cationic liposomes is a relatively simple and standard laboratory procedure. Primary cell cultures have, however, been more difficult to successfully transfect with standardized protocols. As stated above, the initial objective of the experiments reported herein was simply to establish an in vitro model system for the transfection of primary epithelial cell cultures from the preterm lung. However, during the course of these studies, we made several observations that may be of relevance to the design of in vitro and in vivo gene transfer experiments. First, the degree of reporter gene expression does not have a simple concentration-dependent relationship with the amount of plasmid DNA delivered in that efficiency declines with excess delivery. Second, effective intracellular DNA delivery was much less of a limitation to successful gene transfer than DNA damage occurring after cell uptake. Last, delivery of exogenous DNA can render a cell more susceptible to superoxide-mediated cytotoxicity by an as yet undefined mechanism.
Materials. 35S-dATP was from ICN Biomedicals Canada (Montreal, Quebec). 1,[2-14C]dioleoyl-sn-glycero-3-phosphoethanolamine ([14C]DOPE) was from Amersham Canada (Oakville, Ontario). [8-14C]adenine was from NEN (Boston, MA). Proteinase K and nick translation kits were from Promega (Madison, WI). Sephadex G-50 and other chemicals were from Sigma (St. Louis, MO). Porcine trypsin, heat-inactivated fetal bovine serum (FBS), gentamicin, amphotericin B, topoisomerase I, DMEM, and 2,3-dioleoyloxyl-N-[2(sperminecarboxamido)ethyl]-N,N-dimethyl-1-propanaminium trifluoroacetate were from GibcoCanada (Burlington, Ontario). Collagenase and type I DNase (2,367 U/mg) were from Worthington (Freehold, NJ). Restriction enzymes were from Pharmacia (Baie d’Urfé, Quebec). 1,2-Dioleoyldimethylammonium chloride (DODAC), 1,2-dioleoyl-3-N,N,N-trimethylaminopropane, and 3β-[N-(N′,-N′-dimethylaminoethane)-carbamoyl]-cholesterol hydrochloride were from Inex Pharmaceuticals (Vancouver, British Columbia). 1,2-Dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), and fluorescein-labeled DOPE were from Avanti Polar Lipids (Alabaster, AL). A 10-ml thermobarrel liposome extruder was from Lipex Biomembranes (Vancouver, British Columbia), and extrusion filters were from Nucleopore (Pleasanton, CA). The pCMV4-chloramphenicol acetyltransferase (CAT) construct was a generous gift from Drs. K. Brigham and J. Conary (Vanderbilt University School of Medicine, Nashville, TN). The pCMV4-secreted alkaline phosphatase (SEAP) construct was purchased from Tropix (Bedford, MA).
Cell culture. Primary cultures of 19-day gestation rat distal fetal lung epithelial cells (RFL19Ep) were prepared as previously described (9, 16, 31). Briefly, the lungs of 19-day gestation fetal rats were isolated and separated from major vessels and airways. The lungs were minced and gently vortexed to remove erythrocytes. This step was repeated until the supernatant was clear. The minced lung tissue was then subjected to proteolytic digestion with trypsin and DNase, the process being arrested by the addition of FBS. The resultant cell suspension was eluted through 100-μm mesh nylon bolting cloth. The eluted cells were next subjected to a collagenase digestion, which was also arrested by the addition of FBS, and the fibroblasts were removed by differential adherence. Epithelial cells of >95% purity, as assessed by staining for cell cytoskeleton components, are obtained with this technique (16). For the purposes of these experiments, cells were seeded in DMEM with 10% (vol/vol) FBS at a sufficient cell density to allow near confluence 24–48 h after seeding. Cells were maintained in a humidified gas mixture of 3% O2-5% CO2-92% N2 to maintain the cells at a normal fetal arterial oxygen tension of ≈20 mmHg. The culture medium normally contained 50 μg/ml of gentamicin and 2.5 μg/ml of amphotericin B, but these were omitted from the medium when cells were exposed to the liposome-DNA complexes.
Cationic liposome preparation. An equimolar stock solution of DODAC-DOPE in chloroform was dried down under a stream of N2 until the chloroform had completely evaporated. The dried lipids were then maintained under vacuum for 24–48 h. The resulting lipid film was rehydrated at 1 μg total lipid/μl PBS and sonicated for 2 × 30 s. For calculations of lipid recovery, [14C]DOPE was included in the initial lipid mixture. Liposomes were initially extruded through sequentially smaller gauge syringe needles from 18- to 27-Fr and were then subjected to five freeze-thaw cycles with liquid N2 and extrusion through a 400-nm filter (20). Liposomes were recovered by centrifugation at 165,000g for 1 h. Plasmid DNA was added at selected DNA-to-lipid ratios, and the mixture was vortexed gently, then allowed to stand at room temperature for 20 min before use. This preparation was diluted with culture medium as necessary.
Cytotoxicity index. Cytotoxicity was assessed by the release of [8-14C]adenine from cells preincubated with 0.2 μCi/ml of [8-14C]adenine (23). As reported elsewhere, this assay has an ability to detect cell injury equivalent to measurements of lactate dehydrogenase release (31) or trypan blue exclusion (9). Exposures to liposomes ± DNA for various intervals preceded the addition of [8-14C]adenine, before which the medium containing the liposomes or the liposome-free medium for control conditions was removed and the cell monolayer was washed. The medium containing [8-14C]adenine was added for 2 h before the monolayer was washed two times with fresh medium. The percentage of preincorporated [8-14C]adenine released into the culture medium was assessed after 48 h and is expressed relative to the basal release by cells not exposed to liposomes, which is given an arbitrary value of 1.
DNA uptake. pCMV4-CAT plasmids were nick translated with 35S-dATP, and unincorporated nucleotides were removed by passing the reaction mixture through a small column of Sephadex G-50. For uptake experiments,35S-plasmid DNA was mixed with unlabeled plasmid DNA to give a final specific activity of ≈6.5 × 106 dpm/μg DNA. Integrity of the labeled DNA was confirmed by electrophoresis on 1% (wt/vol) agarose gels with 0.2 μg/ml of ethidium bromide. For studies of liposome-associated plasmid uptake, the cells were harvested at the end of the incubation period by treatment with 0.1% (wt/vol) trypsin and 0.001% (wt/vol) DNase and centrifugation. Trypsin removes cell surface-associated liposomes (5), and DNase was included to degrade any surface-associated DNA not associated with liposomes. Total DNA uptake was calculated from the35S content of the cell pellet. For studies using fluorescence microscopy, plasmid DNA was fluorescently labeled with ethidium monoazide (32).
DNA degradation. After a 24-h exposure to DNA with cationic liposomes, the cells were washed two times with culture medium to remove any free liposome-DNA complexes. Cells were harvested immediately after being washed or after a further 24, 48, 72, or 120 h in culture with treatment with 0.1% (wt/vol) trypsin and 0.001% (wt/vol) DNase and centrifugation to remove cell surface-associated liposomes and DNA (5). After resuspension in PBS, a DNA extract was prepared from the cells with proteinase K (12) and was subjected to electrophoresis on 1% (wt/vol) agarose gels with 0.2 μg/ml of ethidium bromide. Blots were probed with probes specific to the CAT nucleotide sequence. In other studies, separation of certain bands by electrophoresis was enhanced by using 0.8% (wt/vol) agarose gels without ethidium bromide.
Plasmid conformation. To relax supercoiled pCMV4-CAT, 0.5 μg of the plasmid prepared with a cesium chloride gradient was mixed in reaction buffer (50 mM Tris ⋅ HCl, 50 mM KCl, 10 mM MgCl2, 0.5 mM dithiothreitol, and 0.1 mM EDTA, pH 7.5) with 1 U of topoisomerase I in a final volume of 50 μl for 30 min at 37°C. The plasmid was linearized with 100 U of the restriction enzyme Sma I with 100 μg of pCMV4-CAT in 100 μl of buffer (100 mM Tris acetate, 100 mM magnesium acetate, and 500 mM potassium acetate, pH 7.5) at 30°C overnight. After extraction, the plasmids were subjected to electrophoresis in 0.8% (wt/vol) agarose gels. For conformation studies, cell DNA extracts were similarly treated.
Reporter gene activity. The bacterial CAT gene, which is not present in eukaryotic cells, was used as a reporter gene to measure transgene expression in studies designed to optimize the liposome delivery system. The gene product is an enzyme that catalyzes the transfer of acetyl groups to the substrate chloramphenicol from acetyl 1-CoA. Catalytic activity was determined by a radiometric assay as described by Stribling et al. (28).
Superoxide cytotoxicity. Superoxide was generated extracellularly by the addition of hypoxanthine and xanthine oxidase essentially as described by Andreoli (2), except that the release of preincorporated [14C]adenine was used as a marker of cytotoxicity.
Poly(ADP-ribose) polymerase activity.The activity of poly(ADP-ribose) polymerase was determined by the measurement of acid-precipitable radioactivity after incubation of cell lysates with [3H]NAD as previously described (9).
Statistical analysis. All values are means ± SE of each group. Where error bars are not evident, they fall within the plot point. Where figures show data from representative experiments, these experiments have been replicated on two to four occasions. Statistical significance (P< 0.05) was generally determined by ANOVA for repeated measures followed by assessment of differences with Duncan’s multiple range test, although paired t-tests were used for some experiments (24).
As the initial step in optimizing the transfection protocol for primary cultures of RFL19Ep cells, we assessed the cytotoxicity of the cationic liposome preparation (1:1 mol/mol DODAC-DOPE) to be used in our studies. As shown in Fig.1 A, the cationic liposomes were cytotoxic (significantly increased release of [8-14C]adenine) at lipid concentrations ≥ 50 nmol/cm2 after a 2-h exposure. There was a small reduction in the cytotoxicity index at some of the lesser lipid concentrations used, but this did not achieve significance. Based on this finding and the desire to be well outside the cytotoxic range of these lipids, we elected to use DODAC-DOPE at ≤5 nmol/cm2 in subsequent experiments. As expected from the reported experience of others (14a), the presence of serum in the culture medium caused a modest, although significant (P < 0.05), reduction in the uptake of35S-DNA (data not shown), and subsequent transfection experiments utilized serum-free medium. We next examined the effect of DNA-to-lipid ratio on DNA uptake by RFL19Ep cells, using a standard concentration of 35S-DNA (0.1 μg/cm2) and varying the lipid concentrations. As shown in Fig. 1 B, the optimal DNA-to-lipid ratio for DNA uptake by RFL19Ep cells over a 6-h exposure period under serum-free conditions was 1:10–15 (mg/mg). Over a concentration range that bracketed the lipid concentrations selected for study and with a fixed DNA-to-lipid ratio of 1:15 (mg/mg),35S-DNA uptake over 6 h increased in proportion to the lipid concentration applied to the cells (Fig.1 C). Optimal DNA uptake with DODAC-DOPE at 5 nmol/cm2 and a DNA-to-lipid ratio of 1:15 (mg/mg) was seen after a 24-h exposure (Fig.1 D). After a 24-h exposure to the DNA-liposome complex, DNA that had been photolabeled with ethidium monoazide was evident in the majority of cells (data not shown). DODAC-DOPE liposomes, prepared with fluorescein-labeled DOPE, were visualized in all cells after a 24-h exposure (data not shown).
CAT expression after a 24-h exposure to the liposome-DNA complex was maximal 72 h after the start of the transfection period and 48 h after the removal of the liposome-DNA complex, with a significant decline thereafter (Fig.2 A). As shown in Fig. 2 B, the cationic liposomes and plasmid were not cytotoxic at the concentrations used in these experiments. With this more extended exposure protocol and lipid component concentrations of 1–7.5 nmol/cm2, cationic liposome-DNA complexes occasionally reduced the basal level of the cytotoxicity index to below that seen under control conditions, as shown in this particular experiment. Having defined an exposure protocol with DODAC-DOPE that could reliably transfect primary RFL19Ep cells, we wished to ensure that other commercially available cationic liposome preparations would not allow even better transfections with the same exposure protocol. Under the conditions optimized for transfection with DODAC-DOPE, no enhanced CAT activity was seen with any of the commercially available liposome preparations tested (Table 1), and DODAC-DOPE was retained for use in subsequent studies.
In pursuing the relationship between DNA delivery and transgene expression, we observed that transgene expression was only dependent on the amount of DNA delivered, up to a phospholipid concentration of 2.5 nmol/cm2, with transgene expression declining with further increases in DNA delivery (Fig.3). When activity and delivery were compared as a ratio, the reduced efficiency of transfection per microgram of DNA was even more evident. This observation led us to speculate that excess transfected plasmid might adversely affect translational efficiency through an excessive activation of intracellular DNA repair processes after DNA injury after endosomal uptake. A number of interventions that might alter DNA degradation were studied (Table 2), including diphenylphenyldiamine to stabilize lysosomes (22), aurintricarboxylic acid to inhibit DNase activity (3), nicotinamide to inhibit poly(ADP-ribose) polymerase activity (18), and chloroquine to inhibit endosomal acidification (15). The concentrations used were defined in a series of preliminary concentration curves to define their cytotoxicity and optimal effects, if any, on gene expression. Diphenylphenyldiamine was found to have no consistent effect on reporter gene expression. A trend toward enhanced reporter gene expression was observed with both aurintricarboxylic acid and nicotinamide, although significance was not achieved. Enhanced reporter gene expression with chloroquine was significant (P < 0.05). A series of studies was performed with various combinations of these agents to detect the presence of additive effects (data not shown). Although a significant additive effect was not observed, the combination of chloroquine and nicotinamide tended to enhance reported gene expression over that seen with chloroquine alone (Table3), and this combination was used in subsequent studies.
Using a CAT cDNA probe and electrophoretic separation of cell DNA extracts on 1% agarose (wt/vol) gels with ethidium bromide (Fig.4 A), we were able to confirm that cells treated with a combination of chloroquine and nicotinamide had an overall reduction in degradation of delivered DNA. After transfection, cells had three high-molecular-weight bands labeled by the CAT cDNA probe seen at the 48- and 72-h time points. The cellular DNA band that comigrated with linearized plasmid declined in intensity with time after transfection, and there was no obvious effect of interventions on this band. Interventions did have an effect on the band that comigrated with relaxed plasmid. This band represented 6 ± 0.2% (SE) of the total band intensity in untreated samples compared with 13 ± 9% in treated samples (n = 3) at 48 h and 9 ± 3 vs. 42 ± 11% at 72 h. The band that comigrated with supercoiled plasmid was only clearly distinguishable in treated samples at the 48- and 72-h time points, being 2 ± 1% of the total at 48 h and 6 ± 2% at 72 h. These values are derived from vertical scans of each lane, and because the amount of linearized DNA may vary from lane to lane, direct horizontal comparisons cannot be made. Preparation of liposome-DNA complexes had no effect on the proportion of DNA in the supercoiled configuration (Fig. 4 B), but the extraction procedure used to isolate cell DNA, when applied to plasmid alone, did convert a portion (≈10–12%) of the supercoiled DNA to a relaxed DNA configuration as reported by others (10). Enhanced separation of supercoiled from linearized DNA was achieved by using 0.8% (wt/vol) gels without ethidium bromide for electrophoretic separation, although separation of relaxed from linearized DNA became less clear (Fig.5 A). Using this approach, we were able to show that the band that comigrated with supercoiled plasmid was present in DNA extracts of cells treated with nicotinamide and chloroquine at the 24-h time point (Fig.5 B). That the lower band, assumed to be supercoiled cDNA, was indeed supercoiled was confirmed by demonstrating its sensitivity to topoisomerase (Fig.5 B). Treatment with topoisomerase reduced the amount of DNA in the supercoiled configuration from 6% of the total, as assessed by densitometry, to <2% of the total in the example shown. Contrary to our expectations, liposome-DNA complexes containing largely supercoiled cDNA or cDNA that was largely relaxed after treatment with topoisomerase I had equivalent transfection efficiencies, in contrast to linearized plasmid that had no transfection capacity (Fig.6).
Superoxide was generated in the culture medium to establish a RFL19Ep cell cytotoxicity assay (Fig.7 A). RFL19Ep cell cultures were transfected with the CAT expression cassette to determine whether the presence of transfected DNA rendered these cells more susceptible to oxidant injury. The sensitivity of the cells in primary culture to exogenously generated superoxide was significantly increased by prior exposure to the lipid-DNA complex (P< 0.05; Fig. 7 B). To confirm that this effect was not plasmid specific, the experiment was repeated with different constructs, including one containing the SEAP reporter gene (Fig. 7 C), with the same result. To put the cytotoxicity index in perspective, the mean increased release of [14C]adenine with exogenously generated superoxide in this latter experiment represented the total intracellular content of 58% of the cells in the monolayer, which increased to a mean of 74% of the monolayer after transfection with the SEAP construct. The enhanced sensitivity to superoxide-mediated cytotoxicity could not be accounted for by the lipid component alone and could only be attributed to the delivery of plasmid DNA. The addition of nicotinamide and chloroquine (Fig.8) restored superoxide-mediated cytotoxicity to the same level as seen in untransfected cells (P > 0.05). A possible explanation for these findings was the activation of nuclear poly(ADP-ribose) polymerase by short nucleotide strands entering the nucleus after cytosolic DNA degradation, which could amplify any activation initiated by superoxide alone. However, direct measurement of poly(ADP-ribose) polymerase activity (Table4) showed no activation of the enzyme by delivery of either linearized or circular plasmid.
There is increasing evidence that the degree of epithelial integrity maintained after lung injury is a critical determinant of pulmonary fibrosis (1). If true, increased antioxidant gene expression or manipulation of other critical determinant genes in epithelial cells may be a future approach to therapeutic intervention. Adenoviral approaches to gene therapy have, to date, been disappointing due to viral- and reporter gene protein-mediated immune responses that limit the duration of any initial response as well as the efficacy of subsequently administered adenoviral vectors (17). This has led to a renewed interest in the use of cationic liposome-DNA complexes for transfection despite their relative inefficiency compared with viral vectors. Expression cassettes delivered with liposomes via the airway will be mostly taken up by lung epithelium, which also eliminates the problem of minimal lung delivery of cationic liposomes after parenteral injection (29). As a first step toward an in vivo delivery system, we undertook these in vitro studies to define an effective liposome-DNA preparation. Because our primary interest is in neonatal lung injury and lung maturity has a direct effect on the sensitivity of lung tissue to oxidant injury (14), we used primary cultures of immature lung epithelium rather than cell lines that adapt to their culture environment during serial passages.
Our approach to the development of an effective transfection protocol for primary cultures was similar to that taken by Caplen et al. (8) with permanent epithelial cell lines, recognizing that the efficacy of liposome-mediated DNA transfection is markedly dependent on several parameters including the toxicity of the lipids used, the cell type transfected, the amounts of DNA and lipid used, and their ratio. The liposome preparation used in our studies was cytotoxic when sufficient concentrations of lipid were applied to the cells. The amount of DNA taken up by the cells was dependent on the concentrations of lipid and DNA used as well as on their ratio. If a lung cell DNA content of 7 pg/cell (27) and an average cell density of 7.5 × 105cells/cm2 are assumed, the observed DNA delivered with a 24-h exposure increased total cell DNA by ≈0.2%. Reporter gene expression decreased over time in culture. This temporally related to a progressive loss of the delivered plasmid DNA. After uptake across the cell membrane, as much as 99.9% of the delivered plasmid DNA may fail to escape the endosome to be available for transgene delivery (11). Thereafter, most of the DNA that escapes the endosome will not cross the nuclear membrane (25). Any free cytosolic DNA, unprotected by a coating of cationic lipid, will be subject to degradation by a recently recognized cytosolic DNase (19). These factors would account for the rapid and extensive DNA degradation observed in control cell monolayers such that only linearized cDNA was evident 24 h after completion of exposure to the liposome-DNA complex, indicative of endonuclease activity. Treatment of cell monolayers with chloroquine and nicotinamide was partially protective against DNA degradation, preserving some of the DNA in relaxed and supercoiled cDNA configurations up to 72 h from the onset of transfection. The most probable explanation for this finding is a chloroquine-mediated limitation of liposome-DNA complex degradation in the endosome that, on endosomal escape, allowed some cDNA to resist the action of cytosolic DNase. Although simply extending the survival of intracellular plasmid would not necessarily have any beneficial effect on reporter gene expression because any preserved DNA would not automatically increase the traffic of DNA across the nuclear membrane, increased reporter gene expression was seen with cells exposed to chloroquine and nicotinamide. Neither relaxed nor linearized cDNA is believed to be capable of reporter gene expression (21). Although this was true for linearized cDNA, supercoiled cDNA that had been extensively relaxed with topoisomerase I gave equivalent reporter gene expression to that seen with largely supercoiled cDNA. This suggests that either the small amount of supercoiled cDNA that was resistant to topoisomerase I was sufficient to generate the amount of reporter gene activity observed or fetal distal lung epithelial cells have repair mechanisms that allow repair of nicked cDNA to reconstitute a supercoiled configuration (26).
The observation that reporter gene expression declined with excessive DNA delivery was intriguing and suggested to us that the large percentage of delivered DNA that becomes linearized may have adverse effects on the processing of uninjured and functional forms. If this observation is confirmed in vivo, it has significant implications for current approaches to gene therapy, for which it is generally assumed that increased gene delivery will equate to increased gene expression.
Of even greater concern for gene therapy initiatives with liposome-DNA complexes is the potential for enhancing the sensitivity of transfected cells to oxidant injury. Cystic fibrosis has been the major focus for lung gene therapy to date. When limited to studies of nasal electrophysiology, adverse effects secondary to enhanced oxidant injury would not be anticipated. When delivered into an inflammed airway, the risk of enhancing oxidant injury could be considerably increased. Superoxide causes single-strand scissions in plasmid DNA (13), and excessive DNA injury, resulting in poly(ADP-ribose) polymerase activation with secondary adenine nucleotide and ATP depletion, is a well-recognized pathway for lethal oxidant injury (4). A reasonable explanation for the observed effect of DNA delivery on superoxide sensitivity would be DNA degradation, resulting in the release of nucleotide strands, which could cross the nuclear membrane to cause a sublethal activation of poly(ADP-ribose) polymerase. This would allow an additive effect on exposure to superoxide. However, direct measurement of poly(ADP-ribose) polymerase activity showed no activation with delivery of either circular or linearized DNA, and the mechanism of the enhanced sensitivity to superoxide by transfected DNA or some other copurified component remains unclear.
This work was supported by a Group Grant from the Medical Research Council of Canada, a Development Grant from the Hospital for Sick Children (HSC), a Research and Development Programme Grant from the Canadian Cystic Fibrosis Foundation, and an Equipment Grant from the Ontario Thoracic Society.
Address for reprint requests: K. Tanswell, Division of Neonatology, Hospital for Sick Children, 555 University Ave., Toronto, Ontario, Canada M5G 1X8.
A. K. Tanswell is the HSC Women’s Auxiliary Chair in Neonatology. H. M. O’Brodovich was a career investigator of the Heart and Stroke Foundation of Ontario during the tenure of these studies. R. Iles and O. Staub were Fellows of the Canadian Cystic Fibrosis Foundation.
- Copyright © 1998 the American Physiological Society