Characterization of two rat models of cystic fibrosis—KO and F508del CFTR—Generated by Crispr‐Cas9

Abstract Background Genetically engineered animals are essential for gaining a proper understanding of the disease mechanisms of cystic fibrosis (CF). The rat is a relevant laboratory model for CF because of its zootechnical capacity, size, and airway characteristics, including the presence of submucosal glands. Methods We describe the generation of a CF rat model (F508del) homozygous for the p.Phe508del mutation in the transmembrane conductance regulator (Cftr) gene. This model was compared to new Cftr −/− rats (CFTR KO). Target organs in CF were examined by histological staining of tissue sections and tooth enamel was quantified by micro‐computed tomography. The activity of CFTR was evaluated by nasal potential difference (NPD) and short‐circuit current measurements. The effect of VX‐809 and VX‐770 was analyzed on nasal epithelial primary cell cultures from F508del rats. Results Both newborn F508del and Knock out (KO) animals developed intestinal obstruction that could be partly compensated by special diet combined with an osmotic laxative. The two rat models exhibited CF phenotypic anomalies such as vas deferens agenesis and tooth enamel defects. Histology of the intestine, pancreas, liver, and lungs was normal. Absence of CFTR function in KO rats was confirmed ex vivo by short‐circuit current measurements on colon mucosae and in vivo by NPD, whereas residual CFTR activity was observed in F508del rats. Exposure of F508del CFTR nasal primary cultures to a combination of VX‐809 and VX‐770 improved CFTR‐mediated Cl− transport. Conclusions The F508del rats reproduce the phenotypes observed in CFTR KO animals and represent a novel resource to advance the development of CF therapeutics.


| INTRODUC TI ON
Genetically engineered animals are essential for gaining a proper understanding of cystic fibrosis (CF) and developing new therapies.
Five animal models (mouse, rat, ferret, pig, and sheep) that lack functional CF transmembrane conductance regulator (CFTR) channels have been developed. 1 The CF mouse has provided a considerable amount of information on the physiopathology of CFTR defects at the organ level, but its main caveat is the lack of a spontaneous lung phenotype. 2,3 One of the reasons for this is the lack of submucosal glands (SMGs), which express CFTR at a high level in humans and other animal models such as the ferret and pig. 4 Indeed, these two animal models better reproduce the CF phenotypes observed in humans, from inflammatory and infectious lung disease to in utero meconium ileus and CF-related diabetes. 5,6 However, the ferret, sheep, and pig models are resource-intensive and pose a number of challenges that limit their widespread use for translational research.
New animal models that develop pulmonary disease and are less challenging in terms of cost and maintenance are required. Due to its zootechnical capacity, size and airway characteristics, the rat may represent a valuable animal model to explore CF. 7 Tuggle et al, using zinc-finger endonuclease technology, recently developed a CFTR −/− -Knock out (KO) -rat model that reproduces many aspects of human disease, including growth failure, tooth enamel abnormalities and bone disease. 8,9 Interestingly, the CFTR −/− rats developed mucus defects with age leading to delayed mucociliary clearance. 10 This is related to the fact that rats have SMGs throughout the cartilaginous tracheal airways that develop postnatally. 11 Furthermore, early reports of CFTR silencing using antisense delivered in utero with adenoviruses resulted in the development of pulmonary disease in newborn rats. CF-associated lung diseases including fibrosis and chronic inflammation became apparent in adult rats at 3.5 months of age following in utero antisense CFTR treatment. 12,13 These data suggest that the rat may represent a useful model to provide further insights into the physiopathology of CF lung disease.
In the expanding area of novel modulators, whose clinical efficiency is based on in vitro assays in human bronchial primary cells, it is crucial to assess the efficacy of mutant CFTR correctors at the organism level in animal models, including rats. Since the most common CFTR mutation in humans is deletion of phenylalanine at position 508, we generated a rat model homozygous for this mutation (p.Phe508del Cftr, Cftr F508del/F508del , F508del thereafter) using the Clustered Regularly Interspaced Short Palindromic Repeats -CRISPR associated protein 9 (CRISPR-Cas9) strategy. This model was compared to new Cftr −/− rats (CFTR KO thereafter) based on the main histological features and electrophysiological characteristics.

| Generation of F508del and CFTR KO rats
All animal care and procedures performed in this study were approved by the Animal Experimentation Ethics Committee of the  14,15 Rats were generated using CRISPR-Cas9 technology and genome editing techniques for the introduction into the genome of single-stranded oligo donor sequences by microinjection into rat zygotes. For the F508del rat, the target sites were based on exon 12; donor DNA generated a codon deletion at F508 and the creation of one NdeI restriction site. For the CFTR KO rat, Results: Both newborn F508del and Knock out (KO) animals developed intestinal obstruction that could be partly compensated by special diet combined with an osmotic laxative. The two rat models exhibited CF phenotypic anomalies such as vas deferens agenesis and tooth enamel defects. Histology of the intestine, pancreas, liver, and lungs was normal. Absence of CFTR function in KO rats was confirmed ex vivo by short-circuit current measurements on colon mucosae and in vivo by NPD, whereas residual CFTR activity was observed in F508del rats. Exposure of F508del CFTR nasal primary cultures to a combination of VX-809 and VX-770 improved CFTR-mediated Cl − transport.

Conclusions:
The F508del rats reproduce the phenotypes observed in CFTR KO animals and represent a novel resource to advance the development of CF therapeutics.

K E Y W O R D S
animal models, CFTR channel activity, CFTR modulators, cystic fibrosis, primary cultures, rat the target sites were based on exon 3 (according to rNO5 6.0 nomenclature); donor DNA generated a frameshift and the creation of one XbaI restriction site and a premature stop codon. This led to generating two CFTR KO founders (referred to as MUKORATs 8.3 and 6.4).
Because no phenotype differences were observed between 8.3 and 6.4 MUKORATs, data obtained from these animals were pooled.

| Characterization of F508del and CFTR KO rats
The colon, ileum, lung, pancreas, and liver from F508del and CFTR KO rats were dissected, fixed, and embedded in paraffin. Blocs were cut into 4-5 µm sections, stained with hematoxylin and eosin (H&E) or periodic acid Schiff (PAS) and examined by light microscopy. Additional staining with Alcian blue to reveal mucus-secreting cells was also performed on intestinal and lung tissue of CFTR KO rats.
Mandibles from 8-to 14-week-old littermate wild-type (WT) and F508del or 14-to 37-week-old CFTR KO rats were scanned by high-resolution X-ray micro-computed tomography (CT) and total enamel volume and density were quantified. The detailed protocol is provided in the Supporting Information.

| Primary nasal cell cultures
The detailed protocol for nasal epithelial cell culture is provided in the Supporting Information. Briefly, the nasal mucosa was scraped and subjected to protease digestion. The cell suspension was then filtered, centrifuged, and the pellet suspended in growth medium for cell expansion on 75-cm 2 flasks. Cells were then trypsinized, seeded on Transwell filters, and differentiated at an air-liquid interface for 20-30 days.

| Nasal potential difference
Transepithelial potential (V TE ) was measured between an Ag/Ag reference electrode and an Ag/AgCl exploring electrode, as described previously with minor adaptations. 16 The following parameters were recorded during V TE measurements: (a) negative stable baseline during Cl − solution perfusion (Baseline V TE ) and (b) sequential V TE change in response to the addition of amiloride (∆Amiloride), low Cl − plus forskolin (∆Low Cl − + FK), and Inh-172 plus GlyH-101 (∆Inh-172 + GlyH-101). The detailed protocol is provided in the Supporting Information.

| Statistics
Data are expressed as means ± SEM calculated using the Prism software package (GraphPad Prism® v6). The Wilcoxon paired signedrank test was used to evaluate treatment (VX-770 or/and VX-809) effect. Between-group comparisons were evaluated using the Mann-Whitney test. N represents the number of rats and n represents the number of filters.

| Generation of F508del and KO rats
Cystic fibrosis transmembrane conductance regulator F508del and KO rats were generated by delivering simultaneously, into one-cell stage zygotes, specific single-guide RNA (sgRNA) and ssODN designed to introduce mutations at codon 508 in exon 12 (insertion of NdeI recognition site) and in exon 3 (insertion of XbaI recognition site) of Cftr gene, respectively ( Figure 1A,B). Sequencing and enzymatic digestion by NdeI for F508del and XbaI for CFTR KO rats were performed to identify gene-edited rats ( Figure 1C,D). As summarized in Table 1, 9 out of 54 pups and 18 out of 53 pups carried the knockin NdeI and XbaI sequences, respectively. One F508del and two CFTR KO founders (MUKORATs 8.3 and 6.4, referred to as CFTR KO) were crossed with a WT partner and the corresponding mutations were transmitted to the offspring.

| Survival and growth
F508del and CFTR KO rat models both showed high mortality at weaning that could not be compensated for by the addition of laxative F I G U R E 1 CRISPR strategy. A, For F508del rats, single-guide RNA (sgRNA) was designed to target exon 12 and the F508 deletion was driven via Knock In (KI) with a ssODN including the ttt deletion, homology arms, and a restriction site to facilitate the genotyping. In blue, a mutation to cancel the sgRNA Protospacer Adjacent Motif (PAM). B, For cystic fibrosis transmembrane conductance regulator (CFTR) KO rats, sgRNA was designed to target exon 3 and the KO was driven via KI with a ssODN including homology arms, a restriction site to facilitate the genotyping and two STOP codons. C and D, Animal genotypes. Representative PCR genotyping for F508del with NdeI digestion (C) and CFTR KO with XbaI digestion (D) rat generation. WT, wild-type  and CFTR KO rats, respectively. All sections from CFTR KO rats were also stained with Alcian blue. Scale bar: 100 µm for F508del rats and 75 µm for CFTR KO rats to drinking water (Figure 2A,B). Autopsies performed on F508del and CFTR KO rats revealed intestinal obstruction, which usually occurred at the level of the cecum with a dilated appearance of the small intestine ( Figure S1A,B). In surviving CF animals, however, no macroscopic abnormalities of the digestive tract were observed when compared to WT rats (Figure S1C-F). Replacement of solid food with DietGel + laxative improved survival to 75% for F508del rats in contrast to CFTR KO rats whose survival was only improved to 25% (Figure 2A,B).
Weight gain after weaning was decreased in F508del and CFTR KO rats compared with their WT littermate ( Figure 2C,D). CFTR KO animals did not display an abnormal fat/lean mass distribution ( Figure 2E). Figure 2F illustrates the weight at 70 days of life of male and female CFTR KO and WT rats fed with DietGel + laxative compared to WT rats fed with normal chow. Although WT rats showed reduced weight gain, it had no effect on animal survival. Figure 2G,H show the growth difference for both F508del and CFTR KO rats compared to their respective WT littermates. Altogether, both models exhibited failure to thrive but CFTR KO rats displayed a more severe nutritional and intestinal phenotype when compared to F508del animals.

| CFTR-dependent phenotypes
Our study reports for the first time the phenotype of F508del

| Nasal potential difference
Nasal potential difference (NPD) was monitored in seven F508del and three CFTR KO rats. As expected, WT controls ( Figure 6A, N = 9) displayed robust Cl − secretion, as indicated by the response to low Cl − -containing medium and the application of forskolin, which was partially inhibited by Inh-172 and GlyH-101 (Table 2). This response was decreased significantly by 64% (P < .05) in F508del rats ( Figure 6B,E; Table 2) and was absent in CFTR KO animals ( Figure 6C,E; Table 2). Quantification of the NPD changes evoked by amiloride and Low Cl − + FK in F508del and CFTR KO rats is shown in Figure 6D and E, respectively. The defect in Cl − secretion was associated with an increase in Na + transport, as revealed by an increased amiloride response in both F508del and CFTR KO rats ( Figure 6D; Table 2). These recordings are consistent with impaired CFTRdependent Cl − secretion in the nasal airways of mutant rats.

| Intestinal bioelectrical measurements
Bioelectrical colon tissue properties from WT, F508del, and CFTR KO rats were evaluated for CFTR by short-circuit current measurements in Ussing chambers ( Figure 7; Table 3). WT tissues exhibited a response to IBMX/FK ( Figure 7A,C,E,G) and Bumetanide ( Figure 7A,D,E,H). These responses were strongly reduced by 66% (P < .01) for 3-isobutyl-1-methylxanthine (IBMX)/FK and by 82% (P < .01) for Bumetanide in F508del rats ( Figure 7B-D), and were virtually absent in CFTR KO animals ( Figure 7F-H). These data are consistent with residual CFTR conductance in the colonic mucosa of F508del rats, which disappeared in CFTR KO animals.

| Primary cultures of nasal airway epithelial cells from F508del rats
Measurements of the short-circuit current were performed on primary nasal cells from WT and F508del rats ( Figure 8). WT primary cells treated with Dimethyl Sulfoxide (DMSO) displayed a mean repolarization of 223.1 µA/cm 2 (±27.5 µA/cm 2 ) in response to IBMX/FK, which was totally inhibited by Inh-172, demonstrating that it was related to CFTR activity ( Figure 8A,C,D; Table 4). These responses were decreased by 72% (P < .0001) for Forskolin/IBMX and by 65% (P < .0001) for Inh-172 in  Table 4). Although there was some variability in the responses, Figure 8E,F show that all individual changes displayed the same profile. Of note, when VX-770 or VX-809 was applied no difference was observed in response to IBMX/FK or Inh-172 (Table 4).

ACK N OWLED G M ENTS
The study was supported by grants from Vaincre la Mucoviscidose to MC and CHC, Swiss CF Foundation (CFCH) and the Swiss National University of Paris) for F508del rat breeding. We also thank Dr Sabrina Noel for the fruitful discussion regarding this manuscript.

CO N FLI C T O F I NTE R E S T S
Dr. Sermet-Gaudelus reports grants from Vertex Therapeutics, personal fees from Vertex Therapeutics, personal fees from Eloxx, nonfinancial support from PTC Therapeutics, outside the submitted work.