|Home | About | Journals | Submit | Contact Us | Français|
Responsible for performing experiments: D.A.S., D.K.M., A.A.P., S.R., M.P.R., G.J.D., R.A.H., C.W.L., C.L.D., J.A.B., G.A.N., E.H.C., P.J.T., P.S.L., M.E., E.E.H., J.L.L., M.S., T.R., P.H.K., A.U., T.D.S., K.A.B., J. S., P.B.M., J.Z., and M.J.W.
Data analysis: D.A.S., D.K.M., A.A.P., S.R., C.W.L., C.L.D., K.A.B., S.S.R., P.B.M., J.Z., and M.J.W.
Manuscript preparation: D.A.S., D.K.M., A.A.P., S.R., J.A.B., G.A.N., E.H.C., P.J.T., A.R.H., K.A.B., S.S.R., P.B.M., J.Z., and M.J.W.
Provided important expertise, reagents, technical personnel and advice: R.A.H., C.W.L., C.L.D., T.R., P.H.K., L.S.O., A.U., T.D.S, A.R.H., K.A.B., R.S.P., S.S.R., and J.S. Competing interests: The University of Iowa Research Foundation and the University of Missouri have applied for patents related to this work. MJW is a Co-Founder of Exemplar Genetics, a company that is licensing materials and technology related to this work.
Lung disease causes most of the morbidity and mortality in cystic fibrosis (CF). However, understanding its pathogenesis has been hindered by lack of an animal model with characteristic features of CF. To overcome this problem, we recently generated pigs with targeted CFTR genes. We now report that, within months of birth, CF pigs spontaneously develop hallmark features of CF lung disease including airway inflammation, remodeling, mucus accumulation, and infection. Their lungs contained multiple bacterial species, suggesting an equal opportunity host defense defect. In humans, the temporal and causal relationships between inflammation and infection have remained uncertain. To investigate these processes, we studied newborn pigs. Their lungs showed no inflammation, but were less often sterile than controls. Moreover, after intrapulmonary bacterial challenge, CF pigs failed to eradicate bacteria as effectively as wild-type pigs. These results suggest that impaired bacterial elimination is the pathogenic event that initiates a cascade of inflammation and pathology in CF lungs. Finding that CF pigs have a bacterial host defense defect within hours of birth provides an opportunity to further investigate pathogenesis and to test therapeutic and preventive strategies before secondary consequences develop.
Cystic fibrosis (CF) is an autosomal recessive disease caused by mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR) anion channel (1-4). People with CF develop pancreatic, intestinal, hepatic, vas deferens, and lung disease. Yet, despite much investigation, the understanding of how loss of CFTR causes disease in these organs remains uncertain and controversial. This is especially true for pulmonary disease, which is characterized by airway inflammation and infection (2-7).
Two obstacles have hindered insight into the pathogenesis of CF lung disease. First, it has not been possible to investigate respiratory tract abnormalities at their onset in humans; that is, beginning at birth, before treatments and chronic abnormalities confound investigation. Second, a suitable animal model has not been available; mice with targeted CFTR gene alterations do not develop the clinical manifestations typical of human CF (8). To circumvent these limitations, we recently targeted the porcine CFTR gene to produce null and ΔF508 alleles (9,10). At birth, CFTR−/− pigs displayed intestinal lesions (meconium ileus and microcolon), exocrine pancreatic destruction, gallbladder abnormalities, and early focal biliary cirrhosis similar to that seen in human CF. Like the lungs of human infants (11,12), porcine CF lungs lacked cellular inflammation and submucosal gland changes.
We hypothesized that CF lung disease would not be unique to humans and that CF pigs would develop abnormalities with time. If they did, we wished to know which of the two characteristic features of CF lung disease, inflammation and infection, would appear first. This has been a persistent question in CF (13-20); the answer has important implications for the pathogenesis of lung disease and for preventive and therapeutic strategies.
For CF pigs to survive, it was necessary to circumvent the meconium ileus that occurred with 100% penetrance (10). In humans, surgery is often required to relieve CF intestinal obstruction (21), and this also was true for pigs. We performed an ileostomy or cecostomy on newborn CF and non-CF pigs. Following surgery, animals were raised in a clean, but not germ-free environment and were segregated at birth from other pigs to avoid exposure to porcine viral and bacterial pathogens.
Here we report data from five CF pigs (3 CFTR−/− and 2 CFTR−/ΔF508) that survived ~2 months or more. CFTR−/ΔF508 pigs had similar gastrointestinal and pulmonary histopathological lesions and nasal voltage measurements as CFTR−/− pigs (fig. S1, S2), and hereafter CFTR−/− and CFTR−/ΔF508 pigs are collectively referred to as CF pigs. The CF animals were euthanized because of respiratory disease or due to complications of gastric ulcers, which are common in swine and are exacerbated by pulmonary disease and stress (22,23) (table S1). Necropsy revealed a fatty pancreas, and microscopic examination revealed distended pancreatic ducts with intervening fibrosis closely matching Dorothy Andersen’s descriptions of “cystic fibrosis of the pancreas” (Fig. 1A) (24). The pancreatic parenchymal loss had progressed since birth (Fig. 1B)(25). The microcolon, seen at birth, had reverted to a normal size and structure. The microgallbladder with mucinous changes and the heterogeneous focal biliary cirrhosis had not substantially changed from that of newborn CFTR−/− pigs (25).
We performed sequential bronchoalveolar lavage (BAL) on several animals and found that cell counts and cytokines from CF pigs did not differ from control samples (fig. S3). In contrast, chest X-ray computed tomography showed airway walls with progressive thickening and frequent scattered infiltrates, with the tracheal lobe often showing the most involvement (Fig. 1C, 1D). These changes are also observed in patients with CF (2).
To test for the presence of bacteria, we cultured BAL liquid. BAL liquid from CF pigs was less frequently sterile compared to wild-type littermates (Fig. 2A). It also tended to contain more bacteria (Fig. 2B). Cultures from both CF and non-CF BAL yielded a variety of bacterial species, which differed from animal to animal (Fig. 2C-2E) and included Staphylococcus aureus, which frequently causes CF lung infections (26) (Fig. 2C-2E, S4). The total numbers and species of bacteria recovered from individual CF pigs varied from one BAL to another. Collectively, these data suggest that the CF porcine lung has an impaired ability to eradicate bacteria, consistent with data from humans showing that many different bacteria can infect CF lungs, especially early in the course of disease (26,27).
After removal at necropsy, some regions of CF lungs did not deflate, and lungs from one pig showed areas of atelectasis and pneumonia (Fig. 3A, 3B). These features are characteristic of airway obstruction. Some airways appeared normal, whereas others had striking features. Purulent exudate occluded some bronchi and partially filled the trachea (Fig. 3A). Other airways were filled with tenacious mucus that retained its shape when retracted, and contained neutrophils, some macrophages, and bacteria (Fig. 3B, 3C).
Inflammation, remodeling, and mucus accumulation are three features that characterize human CF airways (24,28,29). We examined the CF lungs for each of these. Microscopic examination revealed a range of inflammation severity within CF pig airways. Inflammation varied from none, to mild mononuclear (Fig. 4A), to severe neutrophilic infiltration that occasionally ulcerated the airway wall with abscess formation or destroyed submucosal glands (fig. S5A, S5B). In addition, some bronchi and small bronchioles contained purulent exudate (Fig. 4B). Inflammation centered on the airways and generally spared alveoli, although at times airway obstruction produced foci of hyperinflation and atelectasis (fig. S5C). Pneumonia was occasionally observed (fig. S5D). In contrast, non-CF lungs showed minimal cellular inflammation (Fig. 4C).
Airway remodeling was heterogeneous. Within individual animals it ranged from none (Fig. 4D, top), to mild (middle), to severe (bottom), with thickened airway walls, goblet cell hyperplasia of surface epithelium, and, uncommonly, expansion of submucosal glands.
Mucus distended airway lumens (Fig. (Fig.4D,4D, ,5A);5A); however, we rarely found mucus filling, obstructing, or dilating submucosal gland ducts; Fig. 5B shows one example. More commonly, the mucus appeared to arise from goblet cells in the surface epithelium (Fig. 5C). Peripheral bronchioles, which lack submucosal glands, also sometimes showed luminal mucus adjacent to proliferative and mucinous epithelia (Fig. 5D). However, we occasionally found mucus in bronchioles lacking mucinous changes, suggesting that the mucus might have migrated from larger airways (Fig. 5E). Consistent with a surface epithelial origin, the mucus sometimes had a distinctive lamellar appearance, reminiscent of growth rings in trees, suggesting progressive deposition over time (Fig. 5C). This striated appearance resembled the gallbladder and intestinal mucus of newborn CF pigs (25). These changes differed strikingly from those in non-CF pigs (fig. S6). Thus, CF pigs spontaneously develop lung disease within the first few months of life, and the abnormalities replicate those in humans with CF (table S2).
The presence of pulmonary disease in CF pigs a few months old raised questions about how the disease begins. We were interested in inflammation because it is a prominent manifestation in the CF pigs as well as in the lungs of children and adults with CF (7,24,28,29). Moreover, a persistent question in CF research has been whether inflammation precedes infection or vice versa (13,14,16-20,30). To answer this question, we studied newborn pigs (6-12h old). Our earlier histopathologic examination of lungs revealed no evidence of inflammation at birth (10). BAL liquid showed no difference in leukocyte cell counts or IL-8 levels between non-CF and CF pigs (Fig. 6A-6C) (10).
We also evaluated mRNA expression profiles in trachea, bronchi, and distal lung. Unsupervised hierarchical clustering segregated samples into two clusters: airways (trachea and bronchus) and distal lung (Fig. 6D), but there was no clustering based on CFTR genotype. Only a small number of differentially expressed genes were identified between non-CF and CF samples (table S3, S4, S5, and Supplementary Material). However, the number of differentially expressed genes decreased as the P-value and fold change stringency were increased, and expression was not consistently altered across the different tissues. We used Gene Set Enrichment Analysis (GSEA) to further query changes in gene expression using 2,775 publicly available, curated gene sets (table S4). This analysis failed to identify any gene sets that significantly correlated with the CF genotype. These results offer no evidence of a pathway, gene network, or signaling cascade consistent with inflammation in the lungs of newborn CF pigs. Moreover, comparison of these expression profiles with previously published data from CF cells did not reveal genes common to the studies that were consistently differentially expressed across the CF pig samples (table S6). Together, the results indicate a lack of inflammation in the lungs of newborn CF pigs.
To learn whether these normal-appearing newborn lungs exhibit a host defense defect, we cultured BAL obtained shortly after birth and found that BAL from CF pigs was less frequently sterile and tended (not statistically significant) to have more bacteria than BAL from wild-type pigs (Fig. (Fig.7A,7A, S7A). Because the BAL procedure might introduce contaminating bacteria or inadequately recover intrapulmonary bacteria, we removed sections of lung under sterile conditions and cultured them. Whereas more than 50% of wild-type lungs were sterile, only 25% of CF lungs had sterile cultures, and more bacteria were present in CF lungs (Fig. 7A, 7B). The species of bacteria were similar in the two groups and may have entered the lung during passage through the birth canal or via aspiration (fig. S7B). These data suggest that CF pigs have a defect in sterilizing their lungs.
To test this hypothesis directly, we challenged newborn pigs by aerosolizing a defined inoculum of S. aureus into the airways. We chose S. aureus because it commonly infects human infants and children (26), and we found it in the lungs of older pigs (fig. S4). We used an organism that we recovered from one of the older CF pigs. Four hours after delivering the S. aureus, non-CF pigs had tracheal washes that were sterile more often than CF pigs, although BAL cultures did not readily segregate by genotype (Fig. (Fig.7C,7C, S7C). Importantly, direct culture of lung showed that CF pigs failed to completely eliminate the S. aureus as frequently as non-CF littermates, and CF pigs had more bacteria in their lungs (Fig. 7C, 7D). Thus within hours of birth, the CF lung manifests a host defense defect in eradicating bacteria.
Our data indicate that, with time, CF pigs spontaneously develop lung disease that largely replicates that in humans, including the hallmark features of inflammation, remodeling, mucus accumulation and infection. Moreover, within hours of birth, the CF pig lung already has a defect in eradicating bacteria. Although it is not possible to determine whether CF pigs or humans have more severe lung disease within the first months of life, the pulmonary changes in these two species differ from findings in mice which lack typical features of human CF lung disease. These observations indicate that the pulmonary consequences of CFTR loss are not unique to primates. Our results also suggest that comparative studies of humans, pigs, mice, and ferrets (31) may aid understanding of pathogenesis.
Inflammation is a key participant in CF lung destruction (7,24,28,29,32). Yet, the chicken and egg conundrum about which comes first, inflammation or bacteria, has remained a persistent question. Our data indicate that shortly after birth the CF lung has an impaired ability to eradicate bacteria in the absence of inflammation. This defect suggests that chronic, perhaps low-level bacterial stimulation in the CF lung generates inflammation, leading to the hallmark features of CF lung disease. How does our study relate to earlier work on the relationship between inflammation and infection? Examinations of BAL liquid from infants and children with CF have yielded conflicting data regarding the presence of inflammation and infection early in CF lung disease (13,14,30). In vitro models of airway epithelial cell lines and primary cultures produced data both for and against inherent CF inflammation (16-19). Studies of human fetal trachea transplanted into mice suggested that inflammation might occur spontaneously in developing CF airways (20). Compared to earlier work, our in vivo studies have the advantages of avoiding culture conditions, serial passage of cells, and expression of recombinant CFTR that might unexpectedly influence results. In addition, our data were collected within a few hours of birth; because newborns display impaired bacterial elimination, studies done even a few days after birth may not allow unambiguous interpretations. Our present work cannot however address the possibility that the CF lung might have an exaggerated inflammatory response to infection (6).
There were few abnormalities in CF BAL liquid despite X-ray and histopathologic evidence of disease. This difference is interesting because BAL liquid from humans with CF can show evidence of inflammation at what appear to be early stages of the disease (13,14,30). The lack of BAL abnormalities in pigs might relate to heterogeneity of the lung disease and/or to inadequate sampling by BAL. However, the relationship between BAL results and histopathology remains unknown in humans with CF. It is also possible that viral infections or environmental factors to which the CF pigs were not exposed might increase evidence of inflammation in BAL liquid of humans. These considerations and the impaired elimination of bacteria at birth suggest a possible benefit of preventive antibacterial strategies initiated very early in babies with CF, perhaps even before evidence of inflammation in BAL liquid.
Our data suggest that CF produces an equal opportunity host defense defect because porcine CF lungs contained a variety of bacterial species. In addition, the numbers and species of bacteria varied over time, suggesting a dynamic interaction between host and bacteria. This may result because the lung has multiple defenses against bacteria; some may be unaffected by CF, and others may be partially disrupted. Thus, residual defenses may provide variable protection. In this regard, it is interesting that we did not culture Pseudomonas aeruginosa, an important pathogen in humans with CF and an ubiquitous environmental organism (2,26). In human CF, many bacteria other than P. aeruginosa infect patients (26,27), and during the first years of life, S. aeureus is at least twice as common as P. aeruginosa (33,34). Thus, the CF host defense defect may allow infection with a diversity of bacterial species, and then time, selective pressures of pharmaceutical antibiotics, and residual host defense mechanisms may constrain the bacterial community to one or a few predominant species. Whether CF pigs will be less or more susceptible to P. aeruginosa lung infection remains unknown.
The observation that the lungs and gastrointestinal organs of CF pigs share many features with those of human CF suggests that these animals might be of value in better understanding the disease and in testing therapeutic and preventive strategies for multiple organs. Finding that CF pigs have a bacterial host defense defect within hours of birth could also provide an opportunity to explore the pathogenesis of the airway disease before secondary consequences develop.
We previously reported production of CFTR+/−, CFTR+/δF508, and CFTR−/− pigs (9,10). For this study, animals were crossed and the progeny studied. Standard procedures for animal husbandry were used throughout. All animal experiments were approved by the Institutional Animal Care and Use Committees of the University of Iowa and the University of Missouri. Within 6-12 h of birth, control and CF piglets underwent surgical placement of an ileostomy or cecostomy to prevent complications from meconium ileus. In the immediate perioperative period, animals were maintained on intravenous fluids, received carprofen (4 mg/kg every 12 h, IV/SC, Vericore Ltd.) for analgesia, and were treated with prophylactic antibiotics: metronidazole (22 mg/kg every 8 h, IV, SCS Pharmaceuticals), gentamicin (3 mg/kg every 24 h, IV, Hospira, Inc.), and ampicillin (10 mg/kg every 8 h, IV, Abraxis Pharmaceuticals). Thereafter, piglets were fed milk replacer and weaned to a solid diet. Animals received: a) Oral pancreatic enzyme replacement therapy (PancreVed, Vedco) with meals (4000 IU lipase/120 mL milk replacer or ~10,000 IU lipase/kg/day divided between meals). b) Oral fat-soluble vitamins (PancreVed, Vedco). c) Either an oral H2 blocker (1 mg/kg) (famotidine) or oral proton pump inhibitor (1 mg/kg) (omeprazole, Sandoz) once a day. d) Oral polyethylene glycol 3350 (Paddock Laboratories, MinneapPharmaceuticals) with each meal titrated to maintain soft stools.
Seven littermates served as controls. There were 6 CFTR+/+ and 1 CFTR+/− animals and 4 males and 3 females. These pigs underwent surgical placement of an ostomy and received similar treatments to those of the CF animals including administration of an H2 blocker or proton pump inhibitor. Their chest computed tomography scans revealed no abnormalities. They were euthanized at 55, 67, 67, 83, 83, 84, and 100 days old. Necropsy revealed normal lungs. We also observed changes in the stomach ranging from gastric erosions to moderate ulcers.
At necropsy, all pigs were examined for gross lesions and findings documented. Cytology specimens were stained with Wright-Giemsa stain. Tissues from pigs were fixed in 10% neutral buffered formalin (48-96 h). Tissues were then routinely processed, embedded, sectioned (4 μm), and stained with hematoxylin and eosin (HE) for general examination. Additional sections were selectively stained with Periodic acid-schiff (PAS) or modified Gram’s stain (GS). All tissues were examined by a pathologist familiar with the model.
For bronchoalveolar lavage (BAL) in the older pigs, animals were anesthetized with ketamine (15-20 mg/kg, IM), xylazine (1-1.5 mg/kg, IM), and intravenous propofol. A flexible fiberoptic bronchoscope (Pentax FB-10X) was inserted orally and passed through the vocal cords. The suction channel was not used until the tip of the bronchoscope was past the vocal cords. The bronchoscope tip was gently inserted and wedged into the tracheal bronchus airway. Three aliquots (10 ml each) of sterile saline were instilled into the airway and lavage liquid was recovered with intermittent suction. Similar volumes of fluid were recovered from non-CF and CF animals. The BAL liquid was pooled and immediately placed on ice, transported to the laboratory for processing (cell counts and microbiology studies) and stored at -80°C for subsequent analysis.
BAL was performed in newborn pigs immediately after euthanasia and aseptic removal of the lung. A 5-ml aliquot of sterile saline was instilled into the mainstem bronchus. This was repeated 3 times. A similar amount of liquid was recovered from both non-CF and CF newborn piglets.
The total number of recovered cells in BAL liquid was quantified with a hemacytometer and morphologic differentiation of cells was performed on cytospin preparations that were stained with Diff-Quick Stain kit (Baxter). Microbiologic studies were performed on collected BAL liquid, and cytokine ELISAs were performed on BAL liquid that was centrifuged for 10 m at 1600 × g.
Standard microbiologic techniques were utilized to identify and quantify bacteria present in BAL liquid and lung homogenate samples. Samples were serially diluted and plated onto blood agar (tryptic soy agar with sheep blood; Remel), Colombia colistin-nalidixic acid agar (Remel), Chocolate agar (Remel), Mannitol Salt Agar (Remel), MacConkey agar (Remel), and Burkholderia cepacia selective agar (Remel). Organisms were identified with standard microbiological procedures. Some identifications were confirmed by API 20E or API 20NE (bioMérieux), Vitek (bioMérieux) or 16S sequencing (University of Iowa Clinical Microbiology Laboratory and Iowa State University Diagnostic Laboratory). Quantification of colony-forming units (CFU) was performed.
IL-8 and TNF-α were assayed with standard sandwich ELISAs (R&D Systems).
Intrapulmonary bacterial challenge experiments were performed on newborn piglets shortly after birth (8-10 h). Staphylococcus aureus 43 (an isolate obtained from the lungs of Case #1) was made chloramphenicol resistant to aid in bacterial recovery from the airways by introduction of plasmid pCM1. The plasmid was moved into isolate 43 by transduction with bacteriophage a80 as described previously (35). After administration of ketamine (15 mg/kg, IM) and xylazine (1.0 mg/kg, IM), S. aureus (1.9 × 105 ± 3.3 × 104 cfu in a total volume of 0.1 ml 0.45% sodium chloride) was delivered into the airways via an atomizer (MADgic laryngo-tracheal atomizer; Wolfe Tory Medical, Inc.) following passage of the atomizer tip just distal to the vocal cords. Four hours later, the animals were euthanized and the trachea and lungs were sterilely removed from the chest. The distal trachea was placed into 1 ml of PBS and vortexed for 20 s to recover bacteria. The tracheal lobe was homogenized in 5 ml of PBS. BAL was performed on the right and left lungs (5 ml × 3 times on each side). BAL liquid recovery was similar between the groups. The tracheal wash, tracheal lobe homogenate, and BAL liquid samples were serially diluted in sterile saline. Each dilution was plated onto tryptic soy agar (TSA, Remel) plates containing 20 μg/ml chloramphenicol (Sigma). Plates were incubated in the dark under aerobic conditions at 37° C for 48 h. Quantification was done by counting the number of CFU on each plate.
The transepithelial voltage (Vt) across the nasal epithelium was measured as described (36,37) under propofol anesthesia with the animal spontaneously breathing. The reference electrode was a small Ringer’s solution-filled catheter inserted into the leg muscle (25 g needle). The exploring electrode was a 6 french Foley catheter filled with Ringer’s solution (148 mM NaCl, 2.4 mM KH2PO4, 0.4 mM K2PO4, 2.25 mM CaCl2, 1.2 mM MgCl2) inserted ~ 4 cm into the nasal cavity, and the nasal mucosa was perfused at a rate of 5 ml/m using Ringer’s solution to obtain baseline Vt. Subsequent perfusion solutions included: Ringer’s solution containing 100 μM amiloride; Cl−-free Ringer’s solution containing gluconate substituted for Cl− plus amiloride; Cl−-free Ringer’s solution containing amiloride plus 10 μM isoproterenol; Cl−-free Ringer’s solution containing amiloride, isoproterenol, plus 100 μM ATP; Cl−-free Ringer’s solution containing amiloride, isoproterenol, ATP, plus 100 μM GlyH-101. Voltage was measured with a voltmeter connected to a strip chart recorder. Each solution was perfused for 3-5 m. Following completion of the measurements, the epithelium was disrupted by brushing, and Vt was measured again to determine the zero value of Vt.
All chest X-ray computed tomographic (CT) imaging was performed on sedated, spontaneously breathing animals with a 64-slice high-resolution CT (HRCT) scanner (SOMATOM 64, Siemens).
Trachea, bronchus, and distal lung tissue samples were dissected from newborn piglets within 12 h of birth. Samples were cut into ~5 mm3 pieces and stored in RNAlater RNA Stabilization Reagent (Ambion) using the manufacturer’s recommended protocols. Total RNA was isolated with TRIzol Reagent (Invitrogen). Only RNA samples attaining a minimum of 7.0 RNA Integrity Number on the Agilent 2100 Bioanalyzer (Agilent Technologies) were processed. 5 μg of total RNA was used to generate biotinylated cRNA using the Affymetrix Gene Chip one-cycle target labeling kit (Affymetrix, Inc.) according to the manufacturer’s recommended protocols, and then hybridized to the Affymetrix Porcine GeneChip (23,937 probe sets that interrogate approximately 23,256 transcripts from 20,201 S. scrofa genes). Each tissue hybridization (trachea, bronchus, and distal lung) was performed on a separate day, with all genotypes represented in each run. The arrays were washed, stained, and scanned with Affymetrix Model 450 Fluidics Station and Affymetrix Model 3000 scanner, and data collected using the GeneChip Operating Software (GCOS) v.1.4, using the manufacturer’s recommended protocols. The following analysis tools were used to analyze the experimental data: Gene Pattern (38) (unsupervised hierarchical clustering), Partek Genomics Suite (39) (oneway ANOVA analysis), and Gene Set Enrichment Analysis (40) (GSEA), Ingenuity Pathway Analysis (41) (IPA) and Database for Annotation, Visualization and Integrated Discovery (42) (DAVID) to aid in identifying differentially expressed genes by pathway analysis.
Data are presented as either mean ± standard error of the mean (SEM) or individual data points and median. Statistical analyses used Mann-Whitney or Chi-square tests. Statistical significance was considered P < 0.05. Prior to log transformation, 1 was added to all data values.
Fig. S1. CFTR−/ΔF508 newborn piglets have similar histopathology as CFTR−/− piglets.
Fig. S2. CFTR−/ΔF508 newborn piglets have similar nasal electrophysiology as CFTR−/− piglets.
Fig. S3. Bronchoalveolar lavage (BAL) neutrophil and IL-8 levels in non-CF and CF animals.
Fig. S4. Microbiology profile of BAL liquid obtained from non-CF and CF animals.
Fig. S5. Airway wall ulceration, obstruction, atelectasis, and pneumonia.
Fig. S6. Lungs from non-CF pigs ≥ 2 months of age.
Fig. S7. Lung microbiology in newborn CF pigs.
Table S1. Summary of CF pigs.
Table S2. Review of pulmonary lesions in human CF infants (less than 6 months) and CF pigs.
Table S3. Number of differentially expressed genes in each tissue, after the implementation of filters.
Table S4. Gene set enrichment analysis of data from CF and non-CF tissues.
Table S5. Differentially expressed genes based on a P-value cut-off <0.05 and a >2-fold change.
Table S6. Comparison of microarray data for porcine tissue with array profiles from other studies.
We thank E. Bagnall, M. Bland, P. Cooper, B. Diestelmeier, K. Dobbs, T. Fagan, N. Gansemer, M. Griffin, K. Heilmann, C. Hochstedler, M. Hudson, J. Kinyon, P. Korsakov, D. Krotz, X. Li, S. Malerich, T. Mayhew, K. Munson, C. Murphy, P. Naumann, M. Parker, M. Parsons, M. Rector, L. Reznikov, A. Rieke, J. Rodgers, P. Ruff, L. Schneider, J. Sieren, K. Singletary, T. Smith, E. Solin, M. Sturm, T. Wallen, and D. Wax for excellent assistance; K. Chaloner, D. Hornick, G. McLennan, and T. Yahr for insightful discussions; CFF Therapeutics and R. Bridges for the gift of GlyH-101. Funding: This work was supported by the NHLBI (HL51670 and HL91842) and the CFF. DAS is a Parker B. Francis Fellow and was supported by NIAID (AI076671). MJW is an HHMI Investigator.
Accession numbers: GSE21071