Disease phenotype of a ferret CFTR-knockout model of cystic fibrosis

Abstract

Cystic fibrosis (CF) is a recessive disease that affects multiple organs. It is caused by mutations in CFTR. Animal modeling of this disease has been challenging, with species- and strain-specific differences in organ biology and CFTR function influencing the emergence of disease pathology. Here, we report the phenotype of a CFTR-knockout ferret model of CF. Neonatal CFTR-knockout ferrets demonstrated many of the characteristics of human CF disease, including defective airway chloride transport and submucosal gland fluid secretion; variably penetrant meconium ileus (MI); pancreatic, liver, and vas deferens disease; and a predisposition to lung infection in the early postnatal period. Severe malabsorption by the gastrointestinal (GI) tract was the primary cause of death in CFTR-knockout kits that escaped MI. Elevated liver function tests in CFTR-knockout kits were corrected by oral administration of ursodeoxycholic acid, and the addition of an oral proton-pump inhibitor improved weight gain and survival. To overcome the limitations imposed by the severe intestinal phenotype, we cloned 4 gut-corrected transgenic CFTR-knockout kits that expressed ferret CFTR specifically in the intestine. One clone passed feces normally and demonstrated no detectable ferret CFTR expression in the lung or liver. The animals described in this study are likely to be useful tools for dissecting CF disease pathogenesis and developing treatments.

Figure 1. A subset of CFTR^–/– kits are born with MI.

(A) The schematic diagram of the targeted neomycin gene insertion into CFTR exon 10 (E10) outlines the approach used to generate the model and indicates the locations of primers and probes used for genotyping. The results of PCR and Southern blot genotyping for a litter of 6 kits, with the genotypes indicated, are shown. The bottom panel depicts the detection of CFTR protein from the intestine by CFTR immunoprecipitation, followed by in vitro phosphorylation in the presence of [γ-³²P]ATP and protein kinase A. The fully glycosylated band-C form of CFTR is shown. (B) Four kits at birth, with genotype indicated. (C) Intestines from kits at 36–48 hours after birth, with genotype indicated, demonstrating the variability in occurrence of MI and microcolon. Arrows mark perforations in the intestine caused by MI. (D) Histological analysis of the distal ileum/colon in H&E- or periodic acid-Schiff–stained (PAS-stained) sections for a CFTR^+/+ kit and CFTR^–/– kit suffering from MI. All kits with MI demonstrated similar histopathology, with enhanced mucous production (i.e., enhanced periodic acid-Schiff staining in purple). Those CFTR^–/– kits without MI demonstrated histology similar to that in the control animals (data not shown). Scale bar: 200 μm.

Figure 2. Primary organ pathologies observed in newborn CFTR^–/– kits.

(A) H&E-stained sections of the pancreas. Eosinophilic zymogen material filled exocrine acini (asterisks) of the CFTR^–/– pancreas. (B) H&E-stained sections of the liver and gall bladder (GB). Arrowheads mark the bile-filled canaliculi infrequently seen in all genotypes. (C) H&E-stained section of the lung demonstrating the type of lung lesions observed in CFTR^–/– newborn kits. The CFTR^–/– examples are from the same animal that passed stool in Figure 1C and died at 48 hours. Boxed regions are enlarged in the bottom row and demonstrate evidence of fibrin deposition, necrosis, bacteria, and/or inflammation. (D) Blood chemistries for ALT, total bilirubin (Tbil), and cholesterol (CHOL) in animals of the indicated genotypes (WT, CFTR^+/+ and CFTR^+/–; CF, CFTR^–/–). CF kits were divided into 2 groups with and without MI. BL, below limits of detection. Blood was drawn at the time of euthanasia. Values depict the mean ± SEM (n = 5–9 animals in each group) (see Supplemental Figure 3 for additional blood chemistry data). Scale bar: 200 μm (A, top row, B, top row, and C, bottom left and right panels); 25 μm (A, bottom row); 50 μm (B, bottom row); 500 μm (C, top row); 100 μm (C, bottom center panel).

Figure 3. CFTR-knockout kits have a degenerate or absent vas deferens at birth.

Panels depict H&E-stained sections of the vas deferens from 2 CFTR^+/+ and 2 CFTR^–/– newborn kits. H&E-stained sections of both vasa deferentia from each animal are shown in the top 2 rows. The bottom row shows enlarged photomicrographs for the boxed regions in the top panels. Arrows in the top 2 rows mark the vas deferens in wild-type animals or the location of the absent or degenerate vas deferens in CF animals. The vas deferens of CFTR^–/– animal 1 was degenerate on side-1 and absent on side-2, while CFTR^–/– animal 2 demonstrated complete bilateral absence of the vas. Arrowheads mark remnant vas deferens epithelial cells in the enlarged image for CFTR^–/– animal 1. See Supplemental Figure 2 for more CFTR^–/– examples of vas deferens morphology. Scale bar: 50 μm (top and middle rows); 25 μm (bottom row).

Figure 4. UDCA treatment lowers plasma ALT levels but does not enhance weight gain.

(A) Weight gain of CFTR^+/+ and CFTR^+/– kits (n = 4) and littermate CFTR^–/– kits that passed stool and did not have MI (n = 4). One of the CFTR^–/– kits survived to 9.5 days but failed to gain weight after the time points shown. Data depict the mean ± SEM. (B) H&E-stained sections of the kidney and perirenal adipose tissue and lung from 5- to 9-day-old kits. Adipocytes demonstrate depletion of fat stores (fewer white spaces) in both CFTR^–/– kits. Three of the four CFTR^–/– kits treated with UDCA demonstrated lung lesions similar to that shown in the right panel of the bottom row (also see Supplemental Figure 4). The middle row shows enlarged images for the boxed regions in the top panels. Scale bar: 200 μm (top and bottom rows); 50 μm (middle row). (C) Blood chemistries for UDCA-treated kits (WT, CFTR^+/+ and CFTR^+/–; CF, CFTR^–/–). Blood was drawn at the time of euthanasia. Values represent the mean ± SEM (n = 3–4 animals in each group). (D) CFTR^–/– and CFTR^+/+ kits treated with UDCA and omeprazole. The left panel demonstrates size at 45 days of age, and the right panel depicts weight gain. The timing of oral pancreatic enzyme replacement is shown by the arrows, with the lipase units administered with each feeding marked above each arrow.

Figure 5. Airway defects observed in CFTR^–/– kits.

(A and B) TEPD measurements in tracheal xenografts generated from CFTR^+/+, CFTR^+/–, and CFTR^–/– kits. (A) Representative tracings of TEPD following sequential addition of the following drugs to the lumen of the airway xenografts: 100 μM amiloride (Amil), Cl^–-free buffer, 10 μM forskolin/200 μM 8-ctp-cAMP (cAMP/Forsk), and 100 μM GlyH-101. (B) The cumulative data for transepithelial voltage responses (ΔVt) to the various buffer changes is shown for the indicated genotypes. Results depict the mean ± SEM transepithelial voltage responses for N measurements in 4 independent xenografts for each genotype (TEPD was evaluated 4 times for each xenograft on different days). (C and D) Analysis of submucosal gland secretion in tracheal xenografts from CFTR^+/+, CFTR^+/–, and CFTR^–/– kits. (C) Representative en face photomicrographs of glandular secretory droplets are shown (several marked by arrows) at baseline (unstimulated), in response to 30-minute stimulation with 3 μM forskolin, and in response to 30-minute stimulation with 1 μM carbachol. Scale bar: 0.5 mm. (D) Averaged data for glandular secretory rates in response to 3 μM forskolin or 1 μM carbachol for the indicated genotypes. Results depict mean ± SEM (n = 10 CFTR^+/+ and CFTR^+/– xenografts and n = 13 CFTR^–/– xenografts). ^†P < 0.05; *P < 0.05, using the Student’s t test. Typically 10–15 glands were measured for each xenograft sample. The average rate of secretion for all glands in a given sample was used to calculate the mean ± SEM. The total number of glands analyzed was 110 for CFTR^+/+ and CFTR^+/– xenografts and 151 for CFTR^–/– xenografts.

Figure 6. Generation of a gut-corrected transgenic CFTR^–/– ferret by SCNT.

(A) Schematic diagram of the FAPBi-HA-CFTR-PGK-Zeocin cassette used to generate transgenic ferrets expressing HA-tagged CFTR under the control of the FABPi promoter (FABP-Pr) and bovine growth hormone (BGH) poly-A. pFABP, plasmid FABP. (B) Primary fibroblasts were transfected with the linear transgenic fragment shown in A, and selected pools were used for SCNT. Four cloned kits were born, and the gross morphology of the intestine is shown. Clone-1 passed stool normally within 24 hours of birth, while clone-2, -3, and -4 suffered from MI and failed to pass stool. St, stomach. (C) PCR genotyping of the 4 transgenic FABP-HA-CFTR/CFTR^–/– cloned kits. Genomic DNA from a CFTR^+/+ kit served as a negative control, while plasmid DNA (pCFTR) harboring the transgene cassette was used as a positive control. The PCR reactions were designed to specifically detect a segment of the HA-tag and CFTR cDNA or the rat FABPi promoter as shown. (D) Detection of CFTR protein levels in intestinal lysates from the 4 FABP-HA-CFTR/CFTR^–/– clones and a CFTR^–/– kit by CFTR immunoprecipitation, followed by in vitro phosphorylation in the presence of [γ-³²P]ATP and protein kinase A. (E) Comparison of CFTR protein levels using in vitro phosphorylation of immunoprecipitated CFTR from the intestine, lung, and liver of FABP-HA-CFTR/CFTR^–/– clone-1. Lanes show results for CFTR^+/+, CFTR^–/–, and FABP-HA-CFTR/CFTR^–/– clone-1 kits. The fully glycosylated band-C and partially processed band-B forms of CFTR are shown (note that migration of transgenic CFTR is slightly slower than that of endogenous CFTR, due to the presence of the HA-tag).