Pharmacological Correction of Cystic Fibrosis Manifestations

Home , Linaclotide

PHARMACOLOGICAL CORRECTION OF CYSTIC FIBROSIS MANIFESTATIONS

DANIEL R. MCHUGH

Submitted in partial fulfillment of the requirements

for the degree of Doctor of Philosophy

Dissertation advisor: Dr. Craig A. Hodges

Department of Genetics and Genome Sciences

CASE WESTERN RESERVE UNIVERSITY

May 2019 CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Daniel R. McHugh

Candidate for the degree of Doctor of Philosophy

Committee Chair Mitchell Drumm, PhD

Committee Member Calvin Cotton, PhD

Committee Member Thomas Kelley, PhD

Committee Member Drew Adams, PhD

Committee Member Craig Hodges, PhD (Advisor)

Date of Defense March 15, 2019

*We also certify that written approval has been obtained

for any proprietary material contained therein

Table of Contents List of Figures ...... 6 List of Abbreviations ...... 8 Abstract ...... 10 Chapter 1 – Background and Significance...... 13 1.1 The Disease Cystic Fibrosis ...... 14 1.2 CFTR and Mutation Types ...... 15 1.2.1 Class I...... 16 1.2.2 Class II ...... 17 1.2.3 Class III ...... 18 1.2.4 Class IV ...... 18 1.2.5 Class V ...... 19 1.3 Organ-specific symptoms of CF ...... 19 1.3.1 Lung ...... 19 1.3.2 Pancreas ...... 21 1.3.3 Gastrointestinal System ...... 23 1.3.4 Other ...... 25 1.4 Correction of CF Intestinal Pathology ...... 26 1.5 CFTR-directed Therapies ...... 29 1.6 Therapies for Nonsense Mutations in CFTR ...... 34 1.6.1 Nonsense mutation biology ...... 34 1.6.2 Nonsense-mediated decay ...... 37 1.6.3 Readthrough of premature termination codons ...... 38 1.7 Animal Models of CF ...... 41 1.8 Summary of Research...... 44 Chapter 2 – Linaclotide Improves Gastrointestinal Transit in Cystic Fibrosis Mice by Inhibiting Sodium/Hydrogen Exchanger 3 ...... 51 2.1 Abstract ...... 52 2.2 Introduction ...... 53 2.3 Materials and Methods ...... 55 2.3.1 Mouse Strains ...... 55 2.3.2Measurement of Gastrointestinal Transit ...... 55 2.3.3 Intestinal Short-circuit measurements ...... 56

2.3.4 Intestinal Organoid Harvest and Culture from Mouse Small Intestines ...... 56 2.3.4 Quantification of Intestinal Organoid Swelling ...... 57 2.3.5 Measurement of Intestinal Fluid Absorption ...... 57 2.3.6 Luminal Fluidity Measurements...... 59 2.3.9 Solutions ...... 60 2.3.10 Compounds ...... 60 2.3.11 Intracellular pH recovery assay ...... 60 2.4 Results ...... 63 2.4.1 Linaclotide does not activate F508del CFTR ...... 63 2.4.2 Linaclotide increases gastrointestinal transit in CF mice ...... 63 2.4.3 Linaclotide does not induce fluid secretion in the intestine ...... 64 2.4.4 Linaclotide inhibits fluid absorption in the CF intestine ...... 65 2.4.5 Linaclotide inhibits intestinal sodium absorption...... 67 2.4.6 NHE3 inhibition is sufficient to improve GI transit in CF mice ...... 68 2.5 Discussion ...... 68 Chapter 3 – A G542X Cystic Fibrosis Mouse Model for Examining Nonsense Mutation-Directed Therapies ...... 83 3.1 Abstract ...... 84 3.2 Introduction ...... 84 3.3 Materials and Methods ...... 86 3.3.1 Generation of the G542X allele ...... 86 3.3.2 Mice ...... 87 3.3.3 Expression Analysis ...... 88 3.3.4 Bioelectric Measurements ...... 88 3.3.5 Intestinal Organoid Harvesting and Culture ...... 89 3.3.6 Measurement of Intestinal Organoid Swelling ...... 89 3.3.7 Statistics ...... 90 3.4 Results ...... 90 3.4.1 Generation of the G542X Mutation ...... 90 3.4.2 Cftr expression is reduced and CFTR function is absent in G542X mice ...... 91 3.4.3 G542X mice display characteristic CF manifestations ...... 91 3.4.4 G542X CFTR function is restored following pharmacological nonsense mutation readthrough by G418 but not by PTC124 ...... 92

3.5 Discussion ...... 93 Chapter 4 – Synergy between Readthrough and Nonsense-Mediated Decay Inhibition in a Murine Model of Cystic Fibrosis Nonsense Mutations ...... 107 4.1 Abstract ...... 108 4.2 Introduction ...... 108 4.3 Results ...... 111 4.3.1 Intestinal organoid FIS is a more biologically relevant detector of readthrough than cell-based reporters ...... 111 4.3.2 Inhibition of SMG-1 synergizes with readthrough to restore CFTR function 112 4.3.3 Effectiveness of G418 can be improved by NMD inhibition...... 114 4.3.4 SMG1i improves readthrough for alternative aminoglycosides to G418 ...... 114 4.3.5 Readthrough and NMD inhibition combine to improve CFTR function in primary trachea cells...... 115 4.3.6 Non-aminoglycoside compounds do not cause sufficient readthrough to restore CFTR function ...... 116 4.3.7 G542X-CFTR trafficking can be improved by a CFTR corrector following readthrough with G418 and gentamicin ...... 117 4.4 Discussion ...... 117 4.5 Materials and Methods ...... 122 4.5.1 Mice ...... 122 4.5.2 Compounds...... 123 4.5.3 Crypt harvest and intestinal organoid culture ...... 124 4.5.4 Forskolin-induced swelling assay ...... 124 4.5.5 Toxicity Assessment ...... 125 4.5.6 Murine primary airway cell culture ...... 126 4.5.7 Assessment of CFTR function by short-circuit current measurement...... 126 4.5.8 Expression Analysis ...... 127 4.5.9 Readthrough Reporter ...... 127 4.5.10 Statistical Analysis ...... 128 Chapter 5 – Conclusions and Future Directions...... 150 5.1 Summary ...... 151 5.2 Linaclotide in the CF Intestine ...... 152 5.2.1 GC-C Signaling and CFTR ...... 152 5.2.3 Effects of Chronic Linaclotide Treatment ...... 155

5.2.4 Linaclotide Synergy with CFTR Modulators ...... 156 5.2.5 Clinical Trials on Linaclotide Treatment in Patients with CF ...... 158 5.3 The G542X Mouse Model ...... 159 5.3.1 Restoring CFTR Function in the G542X Mouse ...... 160 5.3.2 Generating Additional CF Nonsense Mutation Models...... 161 5.3.3 Humanized CF Mice ...... 163 5.4 Synergy between Readthrough and NMD Inhibition in CF Tissue ...... 164 5.4.1 Spheroid Culture ...... 165 5.4.2 Benefits of Murine Intestinal Organoids ...... 166 5.4.3 NMDI-14 and Amlexanox ...... 167 5.4.4 In vivo examination of SMG1i ...... 169 5.4.5 High Throughput Screening with Intestinal Organoids ...... 170 5.4.6 Applying Readthrough Therapy to Heritable Cancer Syndromes ...... 173 Concluding Statement ...... 174 References ...... 180

List of Figures

Figure 1.1. Overview of the five CFTR mutation classes...... 47 Figure 1.2. The GI system and common dysfunction...... 48 Figure 1.3. Nonsense-Mediated Decay mechanisms degrade mRNA which carries PTCs ...... 49 Figure 1.4. Ribosomal readthrough by aminoglycosides...... 50 Figure 2.1. Linaclotide does not induce a change in short-circuit current in F508del intestinal tissue ...... 75 Figure 2.2. Linaclotide treatment improves GI transit in F508del and Cftr null mice...... 76 Figure 2.3. Linaclotide does not induce swelling of F508del or null intestinal organoids...... 77 Figure 2.4. Linaclotide-treated CF intestinal segments retain more fluid than untreated...... 78 Figure 2.5. Linaclotide treatment inhibits the S3226-sensitive component pHi recovery after an acid load in Caco-2 cells...... 79 Figure 2.6. Targeted Inhibition of NHE3 improves GI transit in null mice...... 82 Figure 3.1. Generation of the G542X Mutation...... 98 Figure 3.2. Cftr expression in tissues from G542X mice...... 99 Figure 3.3. CFTR function in the airway and intestine of G542X mice...... 100 Figure 3.4. Survival and growth characteristics of G542X mice...... 102 Figure 3.5. Intestinal organoids and tissue from G542X mice are used to test G418- mediated nonsense mutation readthrough...... 105 Figure 3.6. Intestinal organoids from G542X mice are used to test PTC124 mediated readthrough...... 106 Figure 4.1. Readthrough with G418 in a cell based reporter system and G542X intestinal organoids ...... 130 Figure 4.2. SMG1i, but not NMDI-14 or amlexanox, synergizes with G418 to improve CFTR function...... 132 Figure 4.3. Low levels of readthrough are improved by inhibiting NMD...... 133 Figure 4.4. Gentamicin and paromomycin restore detectable levels of CFTR function...... 135 Figure 4.5. SMG1i synergizes with readthrough in trachea tissue...... 136 Figure 4.6. Non-aminoglycoside readthrough agents do not restore CFTR function in G542X intestinal organoids...... 138 Figure 4.7. Folding of readthrough-facilitated CFTR can be improved with VX-661 .. 140

Supplementary Figure 1. Detection of FIS by measurement of the intestinal organoid lumen...... 141 Supplementary Figure 2. FIS increases caused by SMG1i do not occur independently of NMD inhibition...... 142 Supplementary Figure 3. Assessment of NMD inhibitor toxicity...... 143 Supplementary Figure 4. SMG1i toxicity impedes FIS when combined with high doses of G418 ...... 144 Supplementary Figure 5. Lumen measurements of alternative aminoglycosides and SMG1i...... 145 Supplementary Figure 6. Readthrough by non-G418 aminoglycosides measured by pFluc190UGA...... 146 Supplementary Figure 7. Readthrough by non-aminoglycoside readthrough agents measured by pFluc190UGA...... 147 Supplementary Figure 8. VX-661 does not facilitate FIS independent of CFTR...... 148 Supplementary Figure 9. 1% DMSO does not impede intestinal organoid FIS...... 149 Figure 5.1. G542X intestinal organoid FIS in a 384-well format...... 175 Figure 5.2. WT, G551D, and F508del intestinal organoids undergo FIS without CFTR correction...... 176 Figure 5.3. Results of 3,000 bioactive small molecule HTS...... 177 Figure 5.4. Four hits were recorded in a HTS for G542X intestinal organoid FIS...... 178 Figure 5.5. W1282X Intestinal organoids undergo swelling without readthrough...... 179

List of Abbreviations

3’ UTR 3’ Untranslated Region A Site Aminoacyl site ABC ATP-Binding Cassette ATP Adenosine Triphosphate AUC Area under the Curve BAC Bacterial Artificial Chromosome cAMP Cyclic Adenosine Monophosphate CBAVD Congenital Bilateral Absence of the Vas Deferens CF Cystic Fibrosis CFRD Cystic Fibrosis-Related Diabetes CFTR Cystic Fibrosis Transmembrane Conductance Regulator cGMP Cyclic Guanosine Monophosphate CIC Chronic Idiopathic Constipation CRISPR Clustered Regularly Interspersed Palindromic Repeats DECID Decay-Inducing Complex DIOS Distal Obstructive Intestine Syndrome DMD Duchenne Muscular Dystrophy DNA Deoxyribonucleic Acid E Site Exit Site eEF1 Eukaryotic Elongation Factor 1 EJC Exon Junction Complex ENaC Epithelial Sodium Channel ER Endoplasmic Reticulum ERAD Endoplasmic Reticulum-Associated Degradation eRF1 Eukaryotic Release Factor 1 eRF3 Eukaryotic Release Factor 3 FABP Fatty Acid Binding Protein FDA Food and Drug Administration FEV1 Forced Expiratory Volume in 1 Second FIS Forskolin-Induced Swelling FRT Fisher Rat Thyroid Cells GC-C Guanylate Cyclase C GCF Geometric Center of Fluorescence GC-S Soluble Guanylate Cyclase GDP Guanosine Diphosphate GI Gastrointestinal gRNA Guide Ribonucleic Acid GTP Guanosine Triphosphate HTS High-Throughput Screening IBS-CC Irritable Bowel Syndrome with Chronic Constipation MI Meconium Ileus mRNA Messenger Ribonucleic Acid MSD Membrane-Spanning Domain NBD Nucleotide Binding Domain

NHE3 Sodium/Hydrogen Exchanger 3 NMD Nonsense-Mediated Decay NMDI-14 Nonsense-Mediated Decay Inhibitor 14 NPD Nasal Potential Difference NTC Normal Termination Codon OGM Organoid Growth Media P Site Peptidyl Site PABP Poly-A Binding Protein PDE5 Phosphodiesterase 5 PEG Polyethylene Glycol PERT Pancreatic Enzyme Replacement Therapy PKA Protein Kinase A PKG cGMP-Dependent Protein Kinase PTC Premature Termination Codon R Domain Regulatory Domain RFC Release Factor Complex SIBO Small Intestine Bacterial Overgrowth SMA Spinal Muscular Atrophy SMG-1 Suppressor with Morphological Effect on Genitalia 1 SMG1i SMG-1 Inhibitor ssODN Single-Stranded Oligodeoxynucleotides STa Heat-Stable Enterotoxin SURF Surveillance Complex tRNA Transporter Ribonucleic Acid WT Wildtype

Pharmacological Correction of Cystic Fibrosis Manifestations

Abstract

By DANIEL R. MCHUGH

Cystic Fibrosis (CF) is a heritable genetic disorder which is caused by a mutation in the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) gene. Loss of CFTR function leads to dehydration of epithelial tissue and accumulation of viscous mucus which is highly prone to infection. Infected mucus leads to inflammation and eventual destruction of the surrounding tissue. Some of the prominent symptoms of CF include lung failure, pancreatic insufficiency, and intestinal obstruction. A number of therapies have been developed for treating the various symptoms and causes of CF; however currently available therapies are insufficient to ameliorate all manifestations of CF. This indicates that additional interventions are needed to improve the quality of life experienced by patients with CF.

One prominent source of discomfort among patients with CF is gastrointestinal dysfunction. Linaclotide, a guanylate cyclase C (GC-C) receptor agonist, is approved for treatment of chronic constipation, but has not been widely used in CF, as CFTR is the primary mechanism of action. Patients with CF who were prescribed linaclotide reported improvements in their GI symptoms, suggesting utility for linaclotide in CF. Linaclotide treatment improved intestinal transit in mice carrying either F508del or null Cftr mutations, as well as improved luminal fluidity and fluid retention, suggesting that linaclotide works through a CFTR-independent mechanism. Here, we demonstrate that linaclotide promotes

10 these improvements in the CF GI environment by inhibiting Sodium/Hydrogen Exchanger

In addition to symptom-specific therapies for CF, a number of compounds are available which can restore function to CFTR with specific mutations. One mutation type which lacks a safe and effective therapy is CFTR nonsense mutations, which ablate protein synthesis by prematurely terminating translation of the CFTR protein. No currently available animal model of CF accurately recapitulates all the manifestations of a CF nonsense mutation. Therefore, we generated a mouse model of the G542X nonsense mutation using CRISPR/Cas9 gene editing. The G542X mouse model has all expected manifestations of a model of CF nonsense mutations, including reduced Cftr expression, absent CFTR function, and reduced growth. Furthermore, aminoglycoside-facilitated readthrough of the G542X PTC is sufficient to restore CFTR function in intestinal organoids harvested from G542X mice. These results suggest that the G542X mouse model has efficacy for examining novel readthrough therapies.

Readthrough is a rare event, with the highest readthrough efficiency being approximately 5%. One reason for the rarity of readthrough is degradation of mutant transcripts by nonsense-mediated decay (NMD) mechanism. Blocking degradation of mutant transcripts may increase the available substrate for readthrough agents to act on and improve readthrough efficiency. Treatment with NMD inhibitors may be a viable treatment, along with readthrough agents, to increase the restoration of CFTR function in the context of a CF nonsense mutation. Primary cells from the G542X mouse were utilized to examine synergy between NMD inhibitors and currently available readthrough agents.

Strong synergy was observed between a small molecule inhibitor of the NMD component

SMG-1 and several readthrough agents in G542X intestinal organoids and trachea monolayers. These results suggest that NMD inhibitors may be a promising co-therapy to improve the efficiency of readthrough in patients with CF nonsense mutations.

Although great strides have been made in developing therapies to treat CF, further interventions are necessary to improve the quality of life of a CF patient to be similar to that of a non-CF patient. The research described here details examining a novel application of linaclotide to improve CF GI manifestations, generation of a CF G542X mouse model to examine readthrough therapies, and utilizes that mouse model to examine synergy between NMD inhibitors and readthrough agents.

Chapter 1 – Background and Significance

1.1 The Disease Cystic Fibrosis Cystic Fibrosis (CF) is a genetic disorder caused by a mutation in the Cystic

Fibrosis Transmembrane Conductance Regulator (CFTR) gene, which codes for an apical membrane chloride and bicarbonate channel with high degrees of expression in epithelial tissues. CFTR is expressed at lower levels in other cell types, including neurons [1], osteoblasts [2], and dendritic cells of the immune system [3]. Loss of CFTR function leads to reduced chloride and bicarbonate permeability across epithelia, leading to dehydration of epithelial surfaces. CF is the most common lethal genetic disorder in Caucasians, occurring in approximately 1 in 2,500 newborns [4]. However, CF has been found to occur in all ethnicities. Mortality among patients with CF occurs most commonly due to lung failure. Currently, there are ~30,000 patients living in the United States, with ~70,000 worldwide.

The earliest description of CF dates back to 1938, when Dorothy Anderson described a series of pediatric patients with celiac disease who displayed fibrosis and cystic lesions in the pancreas [5]. CF is named for these pancreatic lesions. Despite the first description of CF in 1938, it was not until 1989 that CFTR was identified as the causative gene [6]. The identification of the CFTR gene initiated an exhaustive effort to understand and treat CF, which has led to the development of numerous therapies to improve the quality of life for patients with CF. These therapies have been the driving force behind the increase in predicted lifespan of a patient with CF from 30 in 1986 to approximately 43.6 years in 2017 [7].

Despite advances in therapies, numerous symptoms of CF remain which do not have effective therapies. Therefore, improved therapies are necessary to treat specific

14 individual manifestations of CF. However; the ultimate goal of CF research is to restore function to CFTR by correcting the basic defect in CFTR for all patients. Therefore, continued effort is necessary to identify therapies which will treat all symptoms of CF, until patients with CF can enjoy a quality of life which is indistinguishable from non-CF patients.

1.2 CFTR and Mutation Types

The CFTR gene is large; with 27 exons spread over 230 kb. The gene itself is located on chromosome 7 of the human genome. The CFTR gene codes for a 1480 amino acid ATP-binding cassette transporter (ABC) transmembrane protein which localizes to the apical membrane of epithelial tissues [8-10]. Similar to other ABC transporters, CFTR has two nucleotide binding domains (NBDs) along with two membrane-spanning domains

(MSDs). NBD1 is proximate to the amino terminus of the protein, while NBD2 is proximate the carboxy terminus. The NBDs contain regions which are capable of binding with adenosine triphosphate (ATP), while the transmembrane domains contain hydrophobic sequences of amino acids which allow them to be incorporated into the plasma membrane [11]. NBD phosphorylation induces dimerization of the two NBDs, drawing them closer to one another. This causes a subsequent shift in transmembrane domain conformation, opening a pore through which chloride and bicarbonate can pass [12].

Additionally, CFTR contains a regulatory (R) region which is distinct from other members of the ABC transporter family. This R region contains several protein kinase A (PKA) phosphorylation sites. PKA phosphorylates the NBDs of CFTR. The R region has been proposed to prevent dimerization of the two NBDs, preventing opening of the channel in the absence of phosphorylation [13].

CFTR is essential for maintaining hydration of epithelial surfaces throughout the body, including in the airway, intestine, and pancreatic ducts. CFTR hydrates epithelial surfaces by allowing the flow of anions from the intracellular space to the extracellular surface of the tissue. This flow of anions increases the osmolality of the epithelial surface, which draws in fluid and hydrates the tissue. Additionally, the secretion of HCO3- by CFTR can be essential for regulation of pH. For example, CFTR-mediated HCO3- secretion is essential in the pancreas, as absence of CFTR function in the pig pancreas reduces the pH of pancreatic secretions from 8.4 to 5.7 [14].

A mutation which causes loss of function in both alleles of CFTR is sufficient to produce manifestations of CF. Loss of CFTR function dramatically reduces the ability to hydrate epithelial surfaces. The resultant epithelial dehydration causes accumulation of highly viscous mucus which cannot be cleared by conventional mucus clearance mechanisms [15, 16]. Accumulated mucus is prone to bacterial infection, prompting an inflammatory response and increased mucus secretion [15, 17, 18]. Ultimately, this creates a vicious cycle of infection, inflammation, and mucus secretion resulting in mortality.

There are approximately 1,900 known mutations to CFTR, with 97% of patients having one of 160 mutations [19]. Theses mutations are broken down into five classes, depending upon the type of dysfunction that is caused (Figure 1.1).

1.2.1 Class I

The Class I mutation type is characterized by a failure to synthesize the CFTR protein. Approximately 12% of CF-causing CFTR mutations fall under Class I mutations; making these the second most common mutation class [20, 21]. These mutations are known to confer severe CF manifestations due to the absence of functional CFTR. A number of

16 mutation types can eliminate CFTR synthesis. The most common form is the nonsense mutation, which generates a premature termination codon (PTC) in CFTR. Normal termination codons (NTCs) occur at the end of an mRNA transcript, and end translation, allowing the release of a full-formed polypeptide chain from the ribosome. PTCs are created when a mutation creates a stop codon earlier in the transcript than the NTC, leading to early termination of translation and incomplete protein synthesis. The most common nonsense mutation in CF is G542X, in which 542nd codon, which traditionally codes for glycine, is mutated into a UGA stop codon [22].

Another form of mutation which ablates CFTR synthesis is the frameshift mutation.

A frameshift mutation occurs when a number of nucleotides which is not divisible by three are either added or removed in CFTR mRNA. Such mutations shift the reading frame of the mRNA with codons quite different from the native coding sequence, causing translation with entirely different amino acid. These changes in amino acid sequence eliminate protein function. Furthermore, frameshift mutations often create a series of PTCs which preclude the synthesis of functional protein.

Finally, certain splicing mutations are capable of generating frameshift or PTC outcomes on CFTR mRNA which prevent translation of a completed protein. For example, the c.3140-26A>G mutation creates a splice acceptor site 26 bp prior to exon 20 in CFTR, extending the exon by 25 bp [23]. Addition of 25 bp to an exon creates a downstream frameshift which eventually causes premature termination of translation.

1.2.2 Class II

Class II CFTR mutations are characterized by a mutation which causes improper posttranslational processing of the CFTR protein. Most commonly, these mutations result

17 in an inability to properly fold CFTR. Protein which is unable to be folded properly is retained in the endoplasmic reticulum (ER), until it is eventually degraded by ER- associated degradation (ERAD). Thus, in addition to manifestations caused by an absence of CFTR function, Class II mutations can cause ER stress. The most common Class II mutation is F508del, in which a phenylalanine at the 508th amino acid is deleted. The

F508del mutation is present in approximately 85% of patients with CF [24]. F508del is by far the most common CFTR mutation, and makes Class II the most common class of CFTR mutation overall.

1.2.3 Class III

Class III mutations do not influence CFTR synthesis or processing, but prevent proper gating of the CFTR pore. Thus, CFTR which has a Class III mutation is able to localize to the apical membrane normally, but is unable to open once there. Class III mutations frequently cause alterations which modify the function of one of the NBDs, preventing NBD dimerization and opening of the channel. This is the case for the G551D mutation, the most common Class III mutation [25]. Class III mutations are relatively rare, with approximately 5% of CF patients carrying a Class III mutation in one of their CFTR alleles [26]. With the channel unable to open properly, chloride is retained within epithelial cells and is unable to be exported.

1.2.4 Class IV

Class IV mutations are characterized by poor conductance of anions across the epithelial membrane by CFTR. The most common Class IV mutation is R117H [27], which are found in approximately 0.3% of patients with CF [28]. Similar to Class III mutations,

Class IV mutations undergo proper synthesis and process, but the dysfunction occurs at the

18 apical membrane. Studies into the R117H mutation have found that the channel maintains the ability to open, but the channel is unable to remain open for extended periods of time

[27, 29]. Therefore, a reduced quantity of anions is able to be conducted by R117H-CFTR or other class IV mutants. Although CFTR function is impaired, the manifestations of CF in this mutation type tend to be less severe than other mutations, as some degree of CFTR function is still present [28, 30].

1.2.5 Class V

Class V mutations are defined by a reduced quantity of CFTR produced by the cell.

This can occur due to a number of reasons, but the basis for reduced CFTR production is generally a defect in splicing of CFTR mRNA [31]. A common example of this class of

CFTR is the 3849 + 10kbC→T, which confers mild CF manifestations [32, 33]. As with

Class IV, this mutation type generally has less severe manifestations than Class I-III mutations.

1.3 Organ-specific symptoms of CF CFTR is expressed in a wide range of epithelial tissue types. Considering that epithelial tissues are spread throughout the body, loss of CFTR function adversely affects the function of numerous organ systems. Here, the manifestations of CF in the most prominently affected organs will be reviewed.

1.3.1 Lung

The lung manifestations of CF are considered to be more severe than any other organ. Lung dysfunction is the most common cause of mortality among patients with CF

[34]. Consequently, assessment of lung function, often measured by forced expiratory volume in one second (FEV1), is the primary clinical assessment of patient health [35].

Lung dysfunction in patients with CF occurs early in life, and potentially begins in utero.

Similar to other epithelial tissues, loss of CFTR function leads to dehydration of the lung epithelium. Low fluidity of the lung epithelium causes the accumulation of viscous mucus which is unable to be cleared in the same manner as normal mucus. Normally, the primary function of mucus is to trap potentially infectious bacteria which enter the lung. The mucus containing trapped bacteria is then transported out of the lung using the “mucociliary escalator”. The mucociliary escalator uses the beating movement of cilia which lines the lung epithelium to transport mucus from the alveoli and bronchiole of the lung to the trachea. Coughing can then be used to remove mucus from the trachea, where it passes to the stomach and is digested. In the CF lung, the mucus is too viscous to transported and removed by the mucociliary escalator, greatly increasing the risk of infection [18].

Though lung infection has not been identified at birth, infection occurs very rapidly

(within weeks) after birth [36]. The lungs of patients with CF have been found to carry a complex mixture of infectious bacteria. The most common infectious bacteria which are found in the CF lung include Haemophilus influenzae, Staphylococcus aureus, and

Pseudomonus aeruginosa, [37] but a number of additional infectious bacteria have also been identified. Furthermore, anaerobic areas of the lung are often created by the buildup of viscous mucus, as the oxygen gradient along the mucus is frequently inhibited. These anaerobic zones allow the accumulation of anaerobes which normally would not colonize the lung, particularly obligate anaerobes which colonize the lower lung [38].

Infection of the CF lung stimulates a severe inflammatory response. The overabundance of infectious bacteria in the CF lung stimulates the recruitment of neutrophils to the lung. Neutrophils deposit a number of oxidants and proteases in an

20 attempt to eliminate the infectious bacteria [17, 39]. However, this activity also damages the surrounding tissue and promotes further inflammation. Inflammation accumulates in response to infection, which ultimately ends in lung failure and patient mortality.

Patients with CF experience a range of symptoms associated with the described accumulation of viscous mucus, infection, and inflammation. The most prominent among these is reduced lung function, indicated by frequent coughing and shortness of breath, as well as reduced lung volume. This reduced lung volume is monitored in the form of FEV1, and has become a critical measurement of the long-term health of a patient with CF [35].

Patients with CF frequently experience pulmonary exacerbations, which are characterized by an acute worsening of lung function [40]. Pulmonary exacerbations can induce permanent reductions in lung function, as 25% of patients who experience pulmonary exacerbations never return to 90% of pre-exacerbation FEV1 [35].

1.3.2 Pancreas

CFTR has relatively low expression in the pancreas; however loss of CFTR function has highly deleterious effects on pancreatic function. Loss of pancreatic CFTR function leads to exocrine pancreatic insufficiency (PI) in patients with CF. CFTR is expressed in the pancreatic duct epithelium, where it maintains duct fluidity and pH through secretion of chloride and bicarbonate. Maintenance of these ducts is essential for proper functioning of the exocrine functions of the pancreas, which involves secretion of digestive enzymes into the small intestine to aid with digestion [41]. Loss of CFTR function reduces the volume of fluid in the pancreatic duct, increasing the concentration of secreted enzymes in the duct. This reduced volume of fluid is more acidic due to a lack of bicarbonate secretion [41]. Loss of bicarbonate secretion appears to be particularly

21 damaging, as reduced acidity of the duct can lead to premature activation of proteolytic enzymes such as trypsin [41]. Activated pancreatic trypsin likely contributes to destruction of the CF pancreas. Ultimately, this hyperconcentration of precipitated macromolecules leads to obstruction of the duct. Destruction of exocrine pancreatic function begins in utero, with duct obstruction able to be observed at 17 weeks after gestation [42]. In most cases the exocrine function of the pancreas is destroyed at birth, with up to 85% of newborn patients with CF being pancreatic insufficient [43].

Loss of pancreatic function leads to numerous downstream symptoms for patients with CF. PI causes failure to secrete essential digestive enzymes, including trypsin, lipase, and amylase [44]. Loss of these enzymes impairs digestion and absorption of nutrients, leading to malnutrition, malnourishment, and failure to thrive in CF patients. Furthermore,

PI leads to the inability to digest lipids can lead to steatorrhea, or fatty stool, and diarrhea

[45].

While the exocrine function of the pancreas is rapidly destroyed in CF, the endocrine function of the pancreas degrades slowly. Endocrine functions of the pancreas primarily revolve around maintenance of blood glucose levels. The pancreas does this by secreting insulin, a hormone which promotes blood glucose storage, and glucagon, a hormone which promotes breakdown of glycogen to raise blood glucose [46]. The main symptom of loss of endocrine pancreatic function is CF-related Diabetes (CFRD), which occurs in approximately 50% of adult patients with CF [47]. CFRD is rare in pediatric patients with CF, but the incidence of CFRD increases with age [48]. Similar to type I

Diabetes, CFRD is associated with a reduction in insulin secretion by the pancreas. CFRD exacerbates many of the symptoms of CF, which can be disastrous for patient health. CFRD

22 patients have reduced FEV1, compared to glucose tolerant patients [49]. Furthermore, fewer than 25% of patients with CFRD survive to age 30, compared to 60% of patients with normal glucose tolerance [50].

1.3.3 Gastrointestinal System

The advancement of patient care in CF has significantly increased CF patient lifespan. Due to this increase in lifespan, the intestinal manifestations of CF have become more prominent causes of discomfort in older patients with CF. Therefore, the need to treat

CF intestinal manifestations has increased. CFTR is expressed in all tissues of the gastrointestinal (GI) system; however the level of expression varies from tissue to tissue.

CFTR is weakly expressed in the stomach, but has a much higher level of expression along the epithelium of the small intestine, with more moderate CFTR expression in the large intestine [51]. Furthermore, the cell types of the small intestine epithelium vary in CFTR expression. Enterocytes which make up the majority of the epithelium of intestinal villi have markedly low levels of CFTR expression, while the cells which comprise the intestinal crypt have high CFTR expression [52].

Loss of CFTR function in the intestinal epithelium has similar consequences to other tissue types, primarily resulting in dehydration of the intestinal lumen and accumulation of viscous intestinal mucus. These manifestations have deleterious outcomes on the function of the small intestine. The majority of these manifestations stem from a reduction in GI transit, or ability to properly move fecal matter down the intestine. Impaired

GI transit in the CF intestine can lead to a number of uncomfortable or even fatal disorders if not treated, which will be described here. Figure 1.2 depicts the GI tract, and common

GI dysfunctions exhibited in CF.

GI dysfunction among patients with CF begins in utero. Approximately 15% of newborn patients with CF will be born with an obstruction in their intestine [53-55], which is characterized as Meconium Ileus (MI). The term MI comes from an inability to pass the meconium, or first stool, and this obstruction most frequently occurs in the ileum, which is the segment of the small intestine which is most distal from the stomach. If untreated, MI can lead to a fatal rupture of the small intestine, and only surgical intervention or enema treatment is capable of clearing the obstruction [51].

The threat posed by intestinal obstruction does not cease as a patient with CF grows older. Adult patients with CF are susceptible to distal intestine obstructive syndrome

(DIOS), which is characterized by an accumulation of viscous fecal material which combines with viscous mucus to adhere tightly to the intestinal epithelium. This mass can be permanently affixed to the intestinal epithelium, remaining a threat for obstruction for the remainder of a patient’s life [55]. DIOS in childhood is rare, occurring in approximately

7-8% of patients, but increases in frequency as the patient ages, with an adult lifetime prevalence of 14-16% [56].

Reduced GI transit does not always lead to an obstruction in the intestine, but can still cause persistent discomfort. Patients with CF often experience constipation, which is characterized by reduced stool frequency and increased stool thickness. Constipation is a frequent issue among patients with CF, with 47% of patients with CF having a history of constipation [54]. Often associated with constipation is irritable bowel syndrome, in which patients experience persistent intestinal discomfort [57].

Similar to the lung, the microbiome of the intestine is altered in patients with CF.

As many as 56% of patients with CF experience small intestine bacterial overgrowth

(SIBO) [58]. A patient is considered to have SIBO when the patient has >105 colony- forming units (CFU) of bacteria per mL of proximal jejunum aspiration [59]. A normal value is considered <104 CFU/mL. SIBO can initiate a strong immune response, which induces inflammation of the intestinal epithelium. The clinical presentation of SIBO can vary, from being asymptomatic to causing discomfort, nausea and bloating [58]. More severe cases of SIBO can cause malabsorption of carbohydrates, as the bacteria and associated inflammation often results in damage to absorptive enterocytes [60]. Digestion of certain nutrients can also be impaired, as gut bacteria can deconjugate bile acids, which prevents proper lipid digestion [60]. Furthermore, malnutrition caused by SIBO likely contributes to failure to thrive, defined as reduced height and weight of CF patients.

However, conditional expression of Cftr in the small intestine of CF mice does not improve growth, indicating that malabsorption of nutrients is not entirely responsible for failure to thrive [61]. It is possible that metabolic or hormonal abnormalities also contribute to failure to thrive among patients with CF.

1.3.4 Other

A wide range of other organs are impacted in CF, causing many secondary disorders. One symptom of CF is abnormal sweat chloride, with patients with CF having greatly elevated sweat chloride levels. Measurement of sweat chloride is considered the gold standard as far as diagnosis of CF, and has been utilized since the 1960s [62]. While normal patients will have sweat chloride levels under 40mM, sweat chloride levels over

60mM are extremely indicative of a CF diagnosis [63]. Patients with CF can often have chloride levels as high as 100mM [62]. High sweat chloride content can lead to excessive

25 dehydration, putting patients at risk of overheating on particularly hot days. Patients may also have a “frosting” of accumulated salt on their skin following sweating.

CF-associated Liver Disease (CFLD) is a prominent disorder which impacts 30% of CF patients. A number of symptoms fall under the purview of CFLD, but the most common hepatic steatosis [64], or fatty liver disease. CFLD can progress to the point of cirrhosis; or scarring, of the liver, which can lead to mortality [65]. Loss of CFTR impacts the gallbladder as well, leading to a small or even nonexistent gallbladder. Patients with

CF are also prone to development of gallstones, which is thought to be a result of overacidification of bile [66]. 98% of male patients with CF lack a vas deferens, leading to sterility [67]. This condition, termed congenital bilateral absence of the vas deferens

(CBAVD), arises due to accumulation of mucus leading to blockage of the vas deferens, which ultimately deteriorates the vas deferens before birth [68]. Finally, patients with CF experience low bone mineral density [69, 70], often leading to increased fractures of long bones.

1.4 Correction of CF Intestinal Pathology As stated previously, patients with CF have achieved significantly longer lifespans due to improved therapeutic interventions. Therefore, reducing discomfort caused by various CF manifestations is essential for improving quality of life amongst these patients.

Manifestations of CF in the intestine can result in persistent discomfort induced by SIBO, chronic constipation, and irritable bowel syndrome. These symptoms largely arise due to reduced transit of GI contents. Therefore, a number of therapeutic interventions have been developed with the aim of improving GI transit. Chapter II of this thesis examines a novel

26 therapy to treat CF GI maladies. Common therapies which are currently used to attempt to treat CF GI maladies will be reviewed here.

Pancreatic insufficiency experienced by patients with CF has downstream consequences for GI function. As described previously, many patients with CF lack secretion of pancreatic enzymes, which prevents proper digestion of nutrients. To mediate these symptoms, patients with CF are often placed on pancreatic enzyme replacement therapy (PERT). PERT generally involves oral administration of pancreatic enzymes along with meals to aid in digestion [44]. PERT has been sufficient to improve weight gain among

CF patients following one year of treatment [71].

Improving fluid content of the intestine has been demonstrated to be effective for improving GI transit in both CF and non-CF patients [72-75]. Laxatives are generally the preferred option among clinicians for improving GI motility in patients, both CF and non-

CF. In particular, osmotic laxatives like polyethylene glycol (PEG) have been frequently utilized to treat CF GI discomfort [56, 76]. PEG is a high molecular weight, water-soluble compound capable of forming a high number of hydrogen bonds to water molecules [73].

The high incidence of bonding to water molecules by PEG creates a strong osmotic effect in the intestinal lumen, causing the retention of additional fluid [77]. The increased fluidity creates more fluid intestinal contents which are more readily transported down the intestinal lumen. Additionally, this softens the fecal material, making the material easier to pass as a stool. This combination allows an increase in stool frequency and reduction in consistency, as well as reduces abdominal discomfort. Though PEG treatment has been shown to be effective for improving GI transit, a patient requires approximately 4 liters of

PEG solution in one day to achieve a substantial improvement in GI transit [76].

Consuming this large quantity of PEG daily can be difficult for many patients, making alternative laxatives desirable.

To provide improvements in GI motility to patients who do not respond to PEG treatment or who have difficulty consuming the quantity of PEG necessary to improve GI motility, alternative laxatives have been developed. One such laxative is OligoG, a guluronate oligomer which has been developed as an anti-mucolytic agent [78]. OligoG has been shown to bind mucins directly and disrupt cross-linking within the mucin polypeptide [79-81]. These properties made OligoG an attractive therapy for CF intestinal manifestations due to the overabundance of viscous mucus in the CF intestine. An analysis of OligoG treatment on CF mouse models found that OligoG was sufficient to improve GI transit and reduce the frequency of intestinal obstruction [82]. OligoG may be a viable alternative to PEG treatment in patients with CF, but needs to be further developed and tested.

Additionally, interventions have been developed which manipulate the function of salt transporters in the intestinal epithelium to modify lumen fluidity. One such therapy is tenapanor, an orally available inhibitor of Sodium/Hydrogen Exchanger 3 (NHE3), an antiporter which is the primary mediator of sodium absorption in the small intestine and kidney [83-87]. Tenapanor acts to block sodium absorption by NHE3, most likely by physically obstructing the pore which allows sodium import [88, 89]. Blockage of sodium absorption forced sodium to be retained in the intestinal lumen, increasing luminal salt content. This concept has been demonstrated in rats to increase luminal salt concentration, as tenapanor-treated rats had significantly higher sodium content in their stool than control

[90]. Increased osmolality in the intestine is sufficient to improve luminal fluid retention

28 and increase GI transit. Consequently, tenapanor treatment has been sufficient to increase stool frequency and reduce discomfort from non-CF patients with IBS-CC [91]. Though tenapanor has not been thoroughly tested in patients with CF, it represents a promising

CFTR-independent therapy for improving GI transit.

Although a number of therapies are available for improving GI transit in patients with CF, many patients still experience significant discomfort from their GI symptoms.

Indeed, 60% of children and 36% of adults with CF report chronic abdominal pain [92], presumably in spite of the many laxative treatments available. This indicates that additional treatments are necessary to treat CF GI maladies.

1.5 CFTR-directed Therapies Due to the widespread nature of CF, it is difficult and expensive to treat the symptoms of CF in every organ. Therefore, small molecule therapeutics which can restore function to mutant CFTR have a high probability of improving many CF manifestations simultaneously, making such molecules highly desirable. The concept that a small molecule could improve the function of mutant CFTR has been acknowledged since 1997, when the herbal agent genistein was found to increase the open probability of F508del

CFTR [93]. It appeared that genistein was able to open the CFTR pore through a mechanism that was independent of phosphorylation, proving the concept that compounds can interact directly with CFTR to modify CFTR function. Such molecules have been termed “potentiators”. However, genistein treatment only potentiates mutant CFTR to a small degree, and was not likely to be useful for restoring CFTR function in patients.

It wasn’t until 2009 when the first clinically relevant potentiator of mutant CFTR was identified. The potentiator VX-770 was identified in a high-throughput screen [94] of

228,000 compounds aimed at identifying compounds which potentiate human G551D-

CFTR. G551D mutant CFTR is properly folded and trafficked to the apical membrane, but is unable to open to transmit chloride and bicarbonate. VX-770 demonstrated efficacy at potentiating G551D-CFTR without directly phosphorylating CFTR [95]. In clinical examination, 14 days of 250mg VX-770 treatment in patients with CF who expressed at least one G551D allele was sufficient to improve FEV1 by approximately 12.4 percent, compared to 3.1 percent for the placebo group [96]. VX-770 was approved for clinical use by the FDA in 2012 [97], and is now considered the gold standard for CFTR modulators.

Despite clinical efficacy for potentiation of G551D CFTR, VX-770 is insufficient to improve CF symptoms for all patients with CF. As mentioned previously, the vast majority of CFTR mutations arise from misfolding of the CFTR protein. In mutations such as the F508del mutation, the loss of CFTR function is mostly driven by an absence of

CFTR at the apical membrane, not from an inability to potentiate CFTR. In such circumstances, potentiating the small amount of F508del protein at the apical membrane is not sufficient to reverse CF disease manifestations. Considering that the F508del mutation is present in 85% of patients with CF [98, 99], additional compounds which aid in the folding of F508del CFTR were necessary. Therefore, high-throughput screening for compounds which improve folding of F508del-CFTR has been performed. The first effective molecule identified from these screening efforts was VX-809 [94, 100]. This class of compound which corrects CFTR misfolding has been termed “correctors”. VX-809 was sufficient to increase F508del-CFTR maturation to the apical membrane by approximately

8-fold, and this increased quantity of F508del-CFTR was sufficient to increase CFTR- induced current by approximately 7-fold over control groups [100]. In clinical trials, VX-

809 was sufficient to provide modest improvements in sweat chloride values in patients with CF who were homozygous for the F508del mutation. However, VX-809 was insufficient to improve FEV1 in this patient population [101].

The likely reason why VX-809 did not appear to be as effective for F508del- expressing patients as VX-770 was for G551D-expressing patients is that F508del-CFTR is unable to open properly once it has been trafficked to the apical membrane, similar to

G551D-CFTR. F508del-CFTR can be potentiated effectively by VX-770 [100]; therefore

VX-809 and VX-770 are now prescribed as a combination therapy called Orkambi.

Combination therapy of VX-809 and VX-770 was found to improve FEV1 and reduce incidence of pulmonary exacerbations in patients homozygous for the F508del mutation

[102], indicating that this concept translates into an effective treatment. These results are particularly exciting, as these clinical trials indicate that a combination of VX-770 and VX-

809 are likely to restore CFTR function for 90-95% of patients with CF. However, approximately 25% of patients who receive Orkambi experience adverse respiratory events forcing them to discontinue treatment [103]. Therefore, additional CFTR-directed therapies are necessary.

Due to the success of VX-809 and VX-770 for the correction and potentiation of specific CFTR mutations, additional efforts are being pursued to identify novel compounds which are more effective and are not associated with adverse respiratory events. A promising novel potentiator which appears to potentiate F508del and G551D-CFTR more effectively than VX-770 is GLPG1837, which was identified in 2017 [104]. Initial examination of GLPG1837 in patients with at least one G551D mutation indicates that

GLPG1837 is well tolerated and effective for reducing sweat chloride from 98mmol/L to

66mmol/L [105], but further clinical trials are necessary to determine safety and efficacy.

Additional correctors are also becoming available which may provide improvements over VX-809. The first example is VX-661, which was identified in a similar manner to VX-809. VX-661 has been found to have a similar efficacy for improving

CFTR folding as VX-809. However, patients homozygous for the F508del allele who have taken VX-661 in combination with VX-770 experienced improvements in FEV1 but did not differ in respiratory events from the placebo group [106]. This indicates that VX-661 may be a safe alternative to VX-809 for some patients with F508del mutations. Another corrector with promising preliminary data is ABBV/GLPG-2222, an amide compound which appears to potently aid in F508del-CFTR folding [107]. ABBV/GLPG is particularly exciting, as it differs significantly in chemical structure from correctors such as VX-809 and VX-661. This difference in structure suggests that ABBV/GLGP-2222 may act in a different manner to improve CFTR folding, and may have synergistic effects with VX-809 or VX-661.

Two novel correctors, VX-659 and VX-445, have recently been examined for efficacy in primary bronchial cells from patients who are either homozygous for F508del, or express F508del combined with a mutation which results in minimal CFTR function

[108, 109]. Both drugs were more effective at producing a greater chloride current from

F508del-CFTR than VX-661. More importantly, when combined with VX-661 and VX-

770, both VX-659 and VX-445 produced chloride currents which was significantly greater than a combination of VX-661 and VX-770. These results indicate VX-659 and VX-445 are effective correctors of CFTR misfolding, and can synergize with other correctors to

32 further improve the efficiency of folding. In patients, triple combination of either VX-659 or VX-445 with VX-661 and VX-770 was found to significantly improve FEV1 over VX-

661 and VX-770 treatment, indicating these correctors are effective in vivo.

Recently, a novel class of CFTR modulator has been identified. This class of molecule is characterized by the ability to modulate expression of the CFTR gene such that the transcription of CFTR is increased, and have been termed “amplifiers”. The first molecule from this novel class of CFTR modulator is PTI-428 [110]. PTI-428 was determined to have this property when administration of PTI-428 improved the quantity of

F508del CFTR protein in transgenic HEK293 cells. Despite this increase, PTI-428 alone was insufficient to promote CFTR function as measured by short-circuit current, indicating it does not have potentiator activity. The mechanism of PTI-428 was determined to be specifically increasing CFTR expression when it was found that PTI-428 treatment increased the quantity of CFTR mRNA, but not mRNA of the membrane protein G28V-

PgP [110]. Currently, the mechanism of how PTI-428 increases CFTR expression is unknown. The possibility of increasing mutant CFTR expression is particularly exciting, as it provides a genotype-independent intervention which could aid the treatment of all

CFTR mutations. Furthermore, increasing the quantity of CFTR protein would provide additional substrate for other CFTR modulators to act upon. Accordingly, PTI-428 has been found to improve CFTR function when combined with VX-809 [110] or a combination of VX-809 and VX-770 [111]. These initial results are promising, but further examination of amplifiers before they can be utilized as a viable treatment in CF.

1.6 Therapies for Nonsense Mutations in CFTR CF nonsense mutations are one of the few remaining forms of CFTR mutation which do not have a safe and highly effective intervention which can restore function to mutant CFTR. These nonsense mutations are particularly difficult to develop therapies for, as the mutation prevents the production of the CFTR protein completely. Several preliminary therapies exist, but are not clinically viable. To understand the mechanism of the currently existing therapies, it is important to understand the mechanics underlying nonsense mutations. Therefore, a background on the mechanics of nonsense mutations will be discussed prior to a discussion of the strategies available for treating nonsense mutations.

1.6.1 Nonsense mutation biology

Although many patients with CF have CFTR-directed therapies available to them, one patient population which lacks a viable therapy are patients with nonsense mutations in CFTR. Diseases caused by nonsense mutations comprise approximately 10% of all inherited diseases [112]. In addition to CF, nonsense mutations cause a number of severe genetic diseases, including Duchenne Muscular Dystrophy (DMD) and Spinal Muscular

Atrophy (SMA). Within the CF population, approximately 10% of patients carry a nonsense mutation in at least one CFTR allele [113]. Additionally, a nonsense mutation in tumor suppressor genes such as P53 can cause a predisposition to develop various cancer types [114]. Nonsense mutations ablate gene function, or can even cause deleterious dominant-negative proteins [115], producing particularly severe disease manifestations.

Gene function is eliminated by insertion of a premature termination codon (PTC) in mRNA.

Codons are groups of three nucleotides on mRNA transcripts which code for specific amino acids. At the ribosome, the sequence of codons on mRNA is translated into a polypeptide chain by matching cognate aminoacyl tRNAs to each mRNA codon. The term “cognate” refers to a tRNA which has three nucleotides which correspond to each of the three nucleotides on the mRNA. For example, a CAU codon on mRNA will be cognate to a GUA tRNA anticodon. Aminoacyl tRNAs are charged with an amino acid which is specific to the codon to which the tRNA has cognate homology.

Translation is initiated when an initiator tRNA, which carries a methionine amino acid, combines with the 40S and 60S subunits of the ribosome to form a completed 80S ribosome at the 5’ end of an mRNA transcript [116]. This ribosome has three sites which can accommodate charged tRNAs, the aminoacyl site (A site), where tRNAs are initially incorporated into the ribosome, the peptidyl site (P Site), where amino acids are linked together with peptide bonds, and the exit site (E site), where tRNAs are ejected from the ribosome. The initial Met-tRNA begins in the P site of the ribosome, and the next cognate tRNA is incorporated into the A site. A peptide bond is formed between Met and the next amino acid, and the ribosome translocates three base pairs down the mRNA to enter the next codon into the A site. The uncharged Met-tRNA moves to the E site and is ejected, while the amino acid bound to Met and its tRNA move to the P site. From here, the next amino acid is selected, and the ribosome translocates down the transcript towards the 3’ end.

A subset of codons are known as termination codons. These termination codons consist of the UAG, UGA, and UAA codons. Termination codons are recognized by the release factor complex (RFC) consisting of eRF1 and eRF3, which leads to the termination

35 of translation. Normally, termination codons are located at the end of an mRNA transcript and cause the end of translation. However, a nonsense mutation changes a single nucleotide in a normal codon such that the codon now has near-cognate homology to the canonical tRNA, but cognate homology to the RFC. Thus, the RFC has a stronger affinity than the tRNA, and is stably incorporated during translation. This creates a premature termination codon (PTC) in the transcript. PTCs cause premature termination of translation, leading to the release of a truncated protein.

There are 64 possible codons which code for 20 different amino acids to be incorporated into a polypeptide chain. Considering the wide range of amino acids available to incorporate into a protein, mechanisms have evolved to ensure identification of the appropriate amino acids during translation. These are considered to contain two

“proofreading” steps, which ensure the proper tRNA is selected. These steps are broken down into an initial selection step which is followed by GTP hydrolysis, and then a second kinetic proofreading step.

Selecting the appropriate amino acid begins by a process of sampling available amino acids. Here, available tRNAs enter the A site while bound to the eukaryotic translation elongation factor (eEF1A) and GTP [117]. Interactions between mRNA and tRNA are monitored by the decoding center, which is a section of ribosomal RNA located in the A site. Incorporation of a non-cogante tRNA leads to rapid rejection of the tRNA from the A site, and tRNA sampling continues. If a cognate tRNA enters the A site, a conformational shift occurs in the decoding center which allows a series of hydrogen bonds to form between the mRNA and tRNA, creating a stable bond which allows translation to continue [118]. Additionally, this conformational shift allows GTP hydrolysis by eEF1A

36 to occur, releasing eEF1A from the tRNA. This step is sufficient to exclude non-cognate tRNAs with multiple mismatched nucleotides.

To exclude near-cognate tRNAs with only one mismatched nucleotide, a second proofreading step occurs following GTP hydrolysis. This proofreading step is based on binding constants between correct and incorrect nucleotide pairs. A kinetic proofreading mechanism is used to differentiate between cognate and near-cognate tRNAs based on the binding affinity of each tRNA. This mechanism excludes near-cognate tRNAs, and only incorporation of cognate tRNAs will allow translation to continue. Once a cognate tRNA is incorporated into the A site, a peptide bond is formed to incorporate the amino acid carried by the tRNA into the elongating polypeptide chain. Once this peptide bond is formed, the ribosome translocates three nucleotides down the mRNA and begins the process of tRNA sampling again.

1.6.2 Nonsense-mediated decay

PTCs ultimately prevent the production of a full-length protein product by prematurely terminating translation. The truncated protein product is often quickly targeted for degradation [119]. To prevent spending unnecessary resources producing and then degrading truncated proteins, nonsense-mediated decay (NMD) mechanisms have evolved which detect and degrade PTC-carrying mRNAs prior to multiple rounds of translation.

PTC-carrying mRNA is identified during the first round of translation. Several proteins are essential for identifying PTC-carrying mRNA and recruiting NMD mechanisms to degrade the transcript. A complex of proteins termed the Exon Junction

Complex (EJC) is essential for identifying mRNA which carries a PTC. EJCs are deposited approximately 24 nucleotides upstream of the majority of exon-exon junctions during post-

37 transcriptional mRNA processing [120]. During the first round of translation, the ribosome removes the EJCs as it translocates along the transcript. However, the ribosome pauses when a PTC is encountered, allowing downstream EJCs to remain attached to the transcript. EJCs remaining on the transcript mark the transcript for NMD. NMD is initiated by the UPF1 protein, which binds directly to the RFC, and is then bound by the protein kinase SMG-1. The complex formed by the RFC, UPF-1, and SMG-1 are referred to as the surveillance complex (SURF) [121, 122]. Two additional subunits, SMG-8 and 9, bind to

SURF and regulate SURF through induction of conformational changes. Finally, SURF interacts with UPF2 and 3, as well as a downstream EJC to form the decay-inducing complex (DECID) [122-124]. This formation phosphorylates UPF1, and dissociates the

RFC from the ribosome. Finally, the helicase activity of UPF1 activates, translocating along the mRNA to remove secondary mRNA structures and allowing access of nucleases to degrade the mRNA. This process is depicted in Figure 1.3

1.6.3 Readthrough of premature termination codons

The primary method which is being examined to treat PTCs involves utilizing small molecules which can cause the ribosome to insert a near-cognate amino acid instead of the

RFC at the PTC, allowing translation to finish and producing a full-length protein. This phenomenon is termed “readthough”. The most effective molecules for facilitating readthrough of PTCs have been aminoglycoside antibiotics. Aminoglycosides (such as

G418 or gentamicin) are oligosaccharides with a 2-deoxystreptamine ring which is linked to variable numbers of sugar rings and ammonium groups [112]. Aminoglycosides facilitate readthrough of termination codons by strongly interacting with the decoding center of the prokaryotic ribosome. This interaction causes displacement of the adenine

38 nucleotides at position 1492 and 1493 in the decoding center, such that these adenines assume a conformation similar to the one that is taken during cognate codon accommodation [125]. Adoption of this conformation is likely to be sufficient to reduce the stability required for tRNA accommodation, allowing a near-cognate tRNA to be accommodated in the A site following GTP hydrolysis. Ultimately, this reduces the fidelity of the proofreading process, and creates a greater number of mismatched tRNAs. More importantly, this can cause the incorporation of a charged tRNA instead of the RFC at

PTCs (Figure 1.4). When readthrough occurs, the ribosome continues to translate an mRNA transcript when it would normally be stopped by a termination codon. In prokaryotes, this leads to translation of the 3’ UTR and polyA tail which follows NTCs, which produces a nonfunctional protein. Due to this dysfunction in protein synthesis, bacteria are unable to survive, making aminoglycosides highly effective antibiotic agents.

The readthrough mechanism by aminoglycosides has largely been determined using the prokaryotic ribosome, for which aminoglycosides have a high affinity.

Aminoglycoside antibiotics have been found to act on the eukaryotic ribosome in the same manner as the prokaryotic ribosome [126]. However, aminoglycosides two to three orders of magnitude greater affinity for the prokaryotic ribosome than the eukaryotic ribosome

[127]. Therefore, aminoglycosides do not confer the same degree of toxicity to eukaryotes as they do to prokaryotes. This property gives aminoglycosides utility for inducing readthrough of disease-causing PTCs in humans.

The reduced affinity of aminoglycosides towards the eukaryotic ribosome requires that high doses are utilized to achieve the desired readthrough efficiency. Due to the high dose required, aminoglycosides still display toxicity which has prevented their clinical use

39 to treat PTCs [128]. In particular, aminoglycoside treatment has been associated with damage to renal (nephrotoxicity) and hearing (ototoxicity) in patients [129]. There are several of toxicity which aminoglycosides can confer onto eukaryotic cells to induce dysfunction and death. Initially, aminoglycosides are capable of facilitating readthrough of eukaryotic NTCs. Though this does cause a small degree of toxicity, this is a minimal contributor of toxicity in eukaryotes. At NTCs, additional signals interact with the RFC to increase the efficiency of termination. For example, NTCs are located in close proximity to the 3’ UTR and the polyA tail at the end of an mRNA transcript. A polyA binding protein

(PABP) interacts with eRF3 during termination to facilitate efficient separation of the ribosome from mRNA [130]. This improved efficiency of termination largely prevents readthrough at NTCs. Indeed, examination of proteins produced following readthrough does not reveal any increased size beyond a full-length protein [130, 131]. Therefore, aminoglycosides rarely act strongly enough on the eukaryotic ribosome to cause a large amount of readthrough at normal termination codons.

Rather, aminoglycosides tend to cause toxicity in eukaryotes through a process called phospholipidosis. Aminoglycosides are highly positively charged molecules, and interact strongly with acidic phospholipid bilayers of cell and organelle membranes [132,

133]. This interaction can interfere with proper metabolism and turnover of phospholipids.

At high enough concentrations, this can end in membrane disruption and the contents of an organelle are emptied into the cytosol. Aminoglycosides are actively transported into the cell through a transporter complex known as megalin, and the majority of aminoglycoside molecules accumulate inside liposomes [132]. The aminoglycosides induce

40 phospholipidosis on the plasma membrane of the liposome, causing the toxic contents of the liposome to empty into the cytosol, often resulting in cell death.

The instability of PTC-carrying transcripts due to NMD further complicates treating diseases caused by nonsense mutations. A reduced quantity of mutant mRNA creates reduced amounts of substrate which readthrough agents can act on to create full- length protein. Readthrough of PTCs induced by a readthrough allows the ribosome to translocate along the full length of the transcript, which also causes removal of EJCs. Thus, the transcript is then protected from degradation by NMD. However, the most effective currently available readthrough agents have a readthrough efficiency of less than 5% [134-

136], leaving the majority of mRNA to be subjected to NMD. With reduced mutant mRNA quantity, higher doses of aminoglycosides are necessary to restore functional levels of protein. However, inhibiting NMD may increase the quantity of mutant mRNA which aminoglycosides can act upon, increasing the amount of functional protein produced.

Additionally, an increase in mRNA quantity may also reduce the dose of aminoglycoside necessary to restore protein function. Therefore, pharmacological inhibitors of NMD have been proposed to be an effective method of increasing the quantity of mRNA which can serve as a readthrough substrate. In Chapter 4, this concept is examined in the context of a

CF nonsense mutation.

1.7 Animal Models of CF Animal models have represented an invaluable resource for the CF research community. These models serve as a valuable surrogate to perform experiments on large organ systems in CF which cannot be performed using tissue from patients. A number of

CF animal models have been produced, including the rat [137], mouse [138, 139], pig

[140], sheep [141], rabbit [15], zebrafish [142], and ferret [143]. These models each share several symptoms of CF with patients with CF, though no model fully recapitulates the exact manifestations of CF experienced by humans.

The research described here largely uses mouse models to examine different manifestations of CF. Therefore, additional background on the use of mice as a model for

CF will be described. The mouse was the first animal model of CF to be generated [139].

Mice have been utilized for medical research for over 100 years [144]. Mice are popular model of human disease because their physiology and genetics are similar to those of humans. Additionally, mice are small and have a rapid gestation time of approximately 20 days [145]. These characteristics make mice relatively affordable to maintain and quick to reproduce. Furthermore, since mice are popular animals to research, their genetics are well understood, and genetic manipulation can be performed relatively easily.

CF mice share many symptoms of CF with humans. The most prominent manifestation shared by mice and humans is the pathology of the intestine. Similar to humans, mice experience MI, as well as frequent intestinal obstruction. Due to frequent intestinal obstruction, CF mice have dramatically reduced survival compared to WT littermates, with as few as 29% of CF mice surviving to 40 days old [82]. Furthermore, mice experience failure to thrive, and thus are significantly smaller in size than their WT and heterozygous littermates [146]. Interestingly, this reduced growth is not necessarily due to malnutrition, as mice with conditional restoration of CFTR in the small intestine do not obstruct, but still experience failure to thrive [61]. This points to a hormonal or metabolic deficit as the source of failure to thrive, but additional research is necessary to determine the contribution of either.

Despite the similarities between CF mice and patients, there are a number of symptoms which are not shared. As described earlier, the lung manifestations of CF are very severe in patients, contributing to the main cause of mortality. However, in the CF mouse, the lung manifestations are mild, and rarely cause mortality [139, 147].

Additionally, while the pancreas in patients with CF is rapidly destroyed, the pancreas of the CF mouse is not harmed in a detectable way [148]. The lack of respiratory and pancreatic dysfunction in the CF mouse was the primary motivation to develop the previously mentioned animal models.

The CF Mouse Models Core at Case Western Reserve University is the largest CF mouse model core in the world, with over 50 strains of CF mice available. Currently there are mouse models which represent in one form or another each of the different mutation classes. However, not all mouse models recapitulate their particular CF mutation in a way that is representative of how the mutation would function in a patient with CF. For example; there is a lack of a CF mouse model of Class I mutations in Cftr. Multiple mouse models which express nonsense mutations have been developed, but there are flaws for each model. The very first CF animal model carried a nonsense mutation in Cftr [149]. This model expresses the S489X mutation in murine Cftr. However, this model also carries a neomycin selection cassette following the S489X mutation which prevents examination of

CFTR function following readthrough. Additionally, a mouse model which expresses transgenic human CFTR which carries a G542X nonsense mutation has been developed by

David Bedwell at University of Alabama Birmingham [20]. Unlike the S489X model, this hCFTR-G542X model has been shown to produce functional CFTR following readthrough. However, hCFTR-G542X is driven by a rat fatty acid binding protein

43 promoter (FABP), which confers intestine-specific expression at levels which are significantly higher than endogenous Cftr expression. Finally, this model expresses hCFTR-G542X as cDNA, and thus the transcript is not sensitive to NMD mechanisms.

Considering this lack of models of CF nonsense mutations, Chapter 3 details the generation of a CF nonsense mutation mouse model which recapitulates all aspects of a nonsense mutation in CFTR.

1.8 Summary of Research Although great strides have been made in correcting CFTR mutations as well as treating the symptoms of CF, many mutations to CFTR have no effective therapy to restore

CFTR function. Additionally, the currently existing interventions for many CF symptoms are insufficient to treat these symptoms in all patients with CF. Therefore, a concerted effort to identify novel therapies, or adapt currently existing therapies, are necessary to treat the symptoms of CF. Furthermore, models which faithfully replicate the manifestations of various CF mutations are necessary to examine the efficacy of these novel therapies. The goal of the research described here is to address this lack of CF therapies for nonsense mutations in CFTR, as well as the gastrointestinal manifestations of

CF. The described research addresses these issues in three ways: by examining a novel application of the GC-C agonist linaclotide for the treatment of CF intestinal manifestations; by creating and characterizing a mouse model of the G542X nonsense mutation in Cftr; and by utilizing the G542X mouse model to examine currently available therapies for CF nonsense mutations.

Chapter 1 details our examination of the use of linaclotide for treating the GI manifestations of patients with CF. Our examination into the use of linaclotide for treating

CF intestinal disorders began through anecdotal reports from Dr. Robert Stern and Dr.

Kimberly McBennett of University Hospitals. Drs. Stern and McBennett prescribed linaclotide to CF patients who had complained of intestinal discomfort after the drug was requested by several patients. Surprisingly, linaclotide was effective for reducing the intestinal symptoms, prompting an investigation into what mechanism linaclotide was working through in the CF intestine. By investigating the effects of linaclotide on CF mice, we found that linaclotide improves gastrointestinal transport through inhibition of NHE3, the primary sodium absorption channel in the intestine.

In addition to novel therapies for treating the symptoms of CF, animal models of various CFTR mutations are necessary to test novel CFTR-directed therapies. One form of

CFTR mutation which lacks a viable model is the nonsense mutation. Several models of this mutation type in CFTR do exist, however the currently existing models have drawbacks which prevent them from fully replicating all aspects of CF nonsense mutations.

These drawbacks were discussed under “Animal Models of CF”. As these models are insufficient to accurately replicate all aspects of CF nonsense mutations, we generated a mouse model of the G542X nonsense mutation in mouse Cftr using CRISPR/Cas9 gene editing. Chapter 2 details the process of generating and characterizing this mouse model.

We have characterized this mouse for all manifestations of CF nonsense mutations, including reduced Cftr expression, absent CFTR function, failure to thrive, and poor survival. Furthermore, we demonstrate that CFTR function can be restored utilizing aminoglycoside antibiotics to facilitate readthrough of the G542X mutation.

The aminoglycoside gentamicin and PTC124 are the only two therapies which have been assessed in clinical trials for readthrough efficacy. Unfortunately, both have been

45 insufficient either due to toxicity, as was the case with gentamicin, or ineffectiveness, as was the case for PTC124 [150]. Therefore, further interventions are necessary to treat patients with diseases caused by nonsense mutations. One possible avenue to improve the efficacy of currently existing readthrough agents is to increase the quantity of mutant mRNA using inhibitors of NMD. This strategy may increase the quantity of mRNA which can act as a substrate for readthrough, potentially increasing the quantity of functional protein produced following readthrough. As the nonsense mutation in our G542X mouse model carries the G542X mutation in endogenous Cftr, the transcript is sensitive to NMD degradation. This makes tissue from the G542X mouse ideal for examining a combination of readthrough agents and NMD inhibitors. Therefore, in Chapter 3, we examined combinations of several identified readthrough agents in combination with currently available inhibitors of NMD using primary tissue from G542X mice. We identified significant improvements in CFTR function when readthrough was combined with an inhibitor of the NMD component SMG-1. Furthermore, we demonstrate that supplementing NMD inhibitor may improve the efficacy of improving readthrough therapies.

Figure 1.1. Overview of the five CFTR mutation classes. Several different types of dysfunction can reduce or eliminate CFTR function. The nature of CFTR dysfunction is broken down into five different classes. This figure depicts the five mutation classes and lists the type of dysfunction in CFTR. Figure adapted with permission from Fanen, Wohlhuter-Haddad, and Hinzpeter 2014 [151]

Figure 1.2. The GI system and common dysfunction. GI dysfunction can cause a large degree of discomfort in patient with CF. Depicted here is the organs of the gastrointestinal system, as well as symptoms which are common in patients with CF. Figure modified from [152].

Figure 1.3. Nonsense-Mediated Decay mechanisms degrade mRNA which carries PTCs. NMD is utilized to degrade PTC-carrying CFTR transcripts. This results in a reduction in the quantity of CFTR mRNA. Depicted here are the steps which are taken to identify and degrade PTC-carrying mRNA. Figure adapted with permission from Hug, Longman, and Caceres 2016 [122].

Figure 1.4. Ribosomal readthrough by aminoglycosides. PTCs prematurely terminate translation of mRNA. Aminoglycosides interact with the translating ribosome to reduce ribosomal fidelity, allowing readthrough of the PTC. Figure modified with permission from Bordeira-Carrico et al. 2012 [114].

Chapter 2 – Linaclotide Improves Gastrointestinal Transit in Cystic Fibrosis Mice

by Inhibiting Sodium/Hydrogen Exchanger 3

The research in this chapter is published in: McHugh, D. R., et al. (2018). "Linaclotide Improves Gastrointestinal Transit in Cystic Fibrosis Mice by Inhibiting Sodium/Hydrogen Exchanger 3." Am J Physiol Gastrointest Liver Physiol.

2.1 Abstract

Gastrointestinal dysfunction in Cystic Fibrosis (CF) is a prominent source of pain among patients with CF. Linaclotide, a Guanylate Cyclase C (GCC) receptor agonist, is an FDA-approved drug prescribed for chronic constipation, but has not been widely utilized in CF, as the cystic fibrosis transmembrane conductance regulator (CFTR) is the main mechanism of action. However, anecdotal clinical evidence suggests that linaclotide may be effective for treating some gastrointestinal symptoms in CF. The goal of this study was to determine the effectiveness and mechanism of linaclotide in treating CF gastrointestinal disorders utilizing CF mouse models. Intestinal transit, chloride secretion, and intestinal lumen fluidity was assessed in wildtype and CF mouse models in response to linaclotide. CFTR and Sodium/Hydrogen Exchanger 3 (NHE3) response to linaclotide was also evaluated. Linaclotide treatment improved intestinal transit in mice carrying either

F508del or null Cftr mutations but did not induce detectable Cl- secretion. Linaclotide increased fluid retention and fluidity of CF intestinal contents, suggesting inhibition of fluid absorption. Targeted inhibition of sodium absorption by the NHE3 inhibitor tenapanor produced similar improvements in gastrointestinal transit as linaclotide treatment, suggesting that inhibition of fluid absorption by linaclotide contributes to improved gastrointestinal transit in CF. Our results demonstrate that linaclotide improves gastrointestinal transit in CF mouse models by increasing luminal fluidity through inhibiting NHE3 mediated sodium absorption. Further studies are necessary to assess whether linaclotide could improve CF intestinal pathologies in patients. GCC signaling and

NHE3 inhibition may be therapeutic targets for CF intestinal manifestations.

2.2 Introduction Cystic fibrosis (CF) is an autosomal recessive disorder characterized by a mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. CFTR is an anion channel which is highly expressed at the apical membrane of epithelial tissues, and expressed at lower levels in other tissue types. Patients with CF suffer from a variety of intestinal maladies including pancreatic insufficiency, meconium ileus, distal intestine obstructive syndrome (DIOS), and constipation [54, 56, 57, 153]. Loss of CFTR reduces anion flow into the intestinal lumen, which in turn decreases luminal fluid content, leading to increased amounts of viscous mucus and slowed transit of intestinal contents [51].

Treatments of intestinal disorders in CF patients include laxatives, enema treatments, and surgical intervention. However, CF patients still experience pain arising from intestinal manifestations despite administration of these current therapies, indicating that more effective treatments are necessary.

Linaclotide is a FDA-approved drug frequently prescribed for the treatment of chronic idiopathic constipation (CIC) and irritable bowel syndrome with chronic constipation (IBS-CC). Linaclotide is an agonist for the Guanylate Cyclase C receptor

(GCC) [154], an intestinally-expressed receptor which increases the production of the second messenger cyclic guanosine monophosphate (cGMP). cGMP is involved in regulating a variety of cellular proteins, including cGMP-dependent protein kinase II (PKG

II) and protein kinase A (PKA) [155-157]. Both PKG II and PKA can activate CFTR, increasing luminal chloride and fluid transport, thereby increasing GI transit and relieving intestinal pain [157]. In addition, increased cellular cGMP can lead to inhibition of sodium absorption by the Sodium/Hydrogen Exchanger 3 (NHE3), preventing absorption of fluid from the lumen [158].

Currently, linaclotide is not a widely used treatment for constipation in patients with CF due to CFTR being the primary mediator of its effects. However, anecdotal evidence from linaclotide-treated patients with CF at Rainbow Babies and Children’s

Hospital indicates efficacy for linaclotide in treating CF intestinal manifestations. Patients with CF who have reported improvements in intestinal manifestations on linaclotide had various CFTR mutations including F508del, in which small amounts of CFTR may be present at the apical membrane, as well as CFTR null mutations, which have no CFTR present at the apical membrane. These findings suggest that linaclotide may have a CFTR- dependent and/or CFTR-independent mechanism of action which is effective for the treatment of CF-related GI disorders.

To examine the efficacy of linaclotide for improving CF intestinal manifestations, we utilized CF mouse models that display many of the same intestinal manifestations as patients with CF. Mouse models carrying either an F508del mutation or a null mutation in

Cftr were evaluated to determine whether any potential benefit of linaclotide was dependent on residual CFTR function. In this study, we demonstrated that linaclotide increases intestinal lumen fluidity and improves small intestinal transit. We hypothesize that this improved luminal fluid content and transit occurs through inhibition of Na+ absorption by linaclotide. This study suggests that linaclotide may have therapeutic applications for the treatment of CF-based intestinal manifestations, and that cGMP signaling and/or Na+ absorption inhibition may be therapeutic targets to alleviate CF manifestations in other tissue types.

2.3 Materials and Methods 2.3.1 Mouse Strains

Two CF mouse models were used in these experiments, one carrying the F508del

Cftr mutation (Cftrtm1kth) [159] and one carrying a S489X null Cftr mutation (Cftrtm1Unc)

[160]. Both mutations are congenic on the C57Bl/6J background and wild type littermates were used as controls. All mice were allowed unrestricted access to water and solid chow

(Harlan Teklad 7960; Harlan Teklad Global Diets). The Institutional Animal Care and Use

Committee of Case Western Reserve University approved all animal protocols.

2.3.2Measurement of Gastrointestinal Transit

GI transit was measured as previously described [82, 161]. Briefly, the mice were fasted overnight with free access to water. The next day, drug or vehicle control (100 µl) was administered to mice by oral gavage. Following this treatment, the mice were also given rhodamine labeled dextran (100 µl at 25 mg/ml; Sigma-Aldrich) by oral gavage. The mice were sacrificed 40 minutes after the rhodamine dextran treatment, and the gastrointestinal tract from the stomach to the cecum was removed. The small intestine was divided into 10 segments of equal length, and flushed with 2 mL saline. The flushed contents were centrifuged at 500 rpm for 10 minutes, and 200 µl of the supernatant was placed in a 96-well plate. The fluorescent signal of each segment was quantified using a microplate reader (FLUOstar OMEGA microplate reader; BMG LabTech), which was used to calculate the geometric center of fluorescence (GCF), a measurement of GI transit.

GCF was determined by calculating the fraction of fluorescence per segment multiplied by the segment number and adding all segments together. Linaclotide was administered at a concentration of 50 µg/kg, while tenapanor was administered at 1 mg/kg. Linaclotide and

55 tenapanor doses were based on previous animal studies [88, 90, 162] and are similar to human doses when normalizing for body surface area [163].

2.3.3 Intestinal Short-circuit measurements

Short-circuit measurements on intestinal sections were obtained as previously described [164]. Muscularis propia was not removed from intestinal segments and all experiments were done under short-circuit conditions. The change in short-circuit current was calculated after the addition of 10 µM linaclotide to both the apical and basolateral sides of the intestinal sections.

2.3.4 Intestinal Organoid Harvest and Culture from Mouse Small Intestines

Intestinal organoids were harvested from adult mice as previously described [165,

166]. Briefly, the mouse was sacrificed; the small intestine removed and flushed using

PBS. The intestine was cut longitudinally, villi removed with a razor blade and the remaining intestine was cut into ~0.5 cm segments and suspended in Gentle Cell

Dissociation Reagent (STEMCELL Technologies) in a 50 mL conical tube for 30 minutes under gentle agitation with a shaker. In a sterile tissue culture hood, the intestinal segments were vigorously shaken by hand for 30 seconds, and the supernatant was deposited in a 10 cm dish. The segments were resuspended in ice-cold PBS lacking Mg++ and Ca++, and vigorously shaken again for 30 seconds. This process was repeated until four fractions of supernatant were produced. The fraction which was most enriched for intestinal crypts was filtered using a 100 µm cell strainer, and pelleted at 450xg for 10 minutes. The supernatant was discarded, and the pellet was resuspended in 500 µl of a 1:1 mixture of MatriGel

(Corning) and Intesticult Organoid Growth Media (OGM; STEMCELL Technologies).

Crypts were diluted to a concentration of 10 crypts/µl, and plated into non-tissue culture

56 treated 24-well plates, with 35 µl of MatriGel:OGM mixture deposited in each well. The

MatriGel:OGM was solidified at 37°C for 15 minutes, then 500 µl of OGM was gently deposited in each well. The organoids were stored in a 37°C incubator with 5% CO2. The

OGM was changed once every three days, and the organoids were passaged once every 7 days.

2.3.4 Quantification of Intestinal Organoid Swelling

Swelling of intestinal organoids was performed similar to previously described methods [166-168]. Briefly, intestinal organoids were cultured for 6-8 days, split into non- tissue culture treated 48-well plates (Genesee Scientific) and allowed to grow for 24 hours before drug treatment (10 µM linaclotide or tenapanor). Swelling was assessed by treating organoids and imaging for 50 minutes under live cell conditions with brightfield microscopy on a Lionheart FX Automated Microscope (BioTek). Organoid swelling was quantified by identifying and normalizing organoid area to T=0 with Gen5 ImagePrime software (BioTek). Area under the curve (AUC) at T=50 minutes was calculated to compare statistical significance between treatment groups. All organoids used for swelling quantifications were in culture for less than 1 month.

2.3.5 Measurement of Intestinal Fluid Absorption

The rate of fluid absorption from the intestinal lumen was measured using an in vitro gravimetric method as previously described [169]. WT or null mice were euthanized with CO2 asphyxiation, and the small intestine was removed without disturbing the mesentery by cutting immediately below the stomach and immediately above the cecum.

The entire intestine was placed into ice-cold HEPES-Buffered Ringer (HBR, in mM: 138

NaCl, 10 HEPES, 5 KCl, 2.5 Na2HPO4, 1.8 CaCl2, 1 MgSO4) supplemented with 10mM

57 glucose. The mesentery was carefully cut away with surgical scissors to untangle the intestine. The jejunum and ileum were removed and discarded, while the duodenum was cannulated with a rubber-tipped gavage needle attached to a 50ml syringe. The syringe was used to flush HBR supplemented with 10mM glucose and 1µM indomethacin through intestine to clear any fecal matter. The syringe was then filled with Krebs Bicarbonate

Ringer (KBR, in mM: 115 NaCl, 25 NaHCO3, 5 KCl, 2.5 Na2HPO4, 1.8 CaCl2, 1 MgSO4) supplemented with 10mM mannitol, 0.1µM tetrodotoxin and 1µM indomethacin. For linaclotide-treated segments, 10µM linaclotide was included in the KBR solution. The

KBR was flushed through the duodenum, and during flushing, the distal end of the duodenum was tied off with surgical suture to inflate the duodenum with KBR. Any remaining mesentery tissue was cut away after inflating the duodenum. The duodenum was divided into 2cm segments by sliding a suture thread knot ~2cm up the duodenum, and securing the knot tightly. A second suture thread knot was tied ~0.5cm past the previous knot, and the small segment between the two knots was severed to separate two fluid-filled segments. This process was repeated to produce up to four segments per duodenum. The segments were placed into scintillation vials with 5mL of KBR supplemented with 10mM glucose, 0.1µM tetrodotoxin, and 1µM indomethacin, with one segment per vial. The segments were incubated in a 37°C water bath for 10 minutes prior to beginning measurements. To weigh each segment, the segment was removed from the vial, blotted on filter paper, weighed, and then returned to the vial. The vial was gassed with 95%O2/5%

CO2 for 10 seconds, and returned to the 37°C water bath. Segments were weighed at 10 minute intervals for two hours. Tissue wet weight for each segment was recorded by

58 piercing, draining, and weighing the segment following the experiment. Segment weights were normalized by dividing by tissue wet weight.

2.3.6 Luminal Fluidity Measurements

Fluidity of intestinal contents was measured as previously described [87]. The luminal contents of the small intestine of null mice were collected 1 hour after treatment with 50 µg/kg linaclotide or water (100 µl) by oral gavage. The luminal contents were weighed immediately after collection, placed overnight in a 55°C oven and weighed again in the morning. Fluidity was normalized by dividing wet weight by dry weight.

2.3.8 Cell culture

Human Caco-2 cells with CFTR deletion were utilized to test NHE3 inhibition. A subclonal line of Caco-2 was generated by CRSIPR/Cas9-mediated mutagenesis, in which a guide RNA recognizing exon 11 was used. In this line, there were two alleles of CFTR created, each carrying frame-shifting insertions (c.1432insCC and c1433.insAA), which resulted in no detectable CFTR protein or function as assessed by western blot or short circuit currents, respectively. 5 days prior to performing fluorescence recordings, these non-polarized human Caco-2 cells with CFTR deletion were seeded onto 18 mm poly-D-lysine coated coverslips (#1 thickness, Cat. # GG-18-PDL, Neuvitro corp.

Vancouver, WA), each placed in a 3.5 cm culture dish, at a density of 1 × 105 cells per dish and cultured at 37C in a 5% CO2 incubator. Caco-2 cells were grown in EMEM without L-glutamine (Lonza) supplemented with 10% Fetal Bovine Serum, 1%

GlutaMAX, and 1% Penicillin/Streptomycin in a 37°C cell culture incubator with 5% CO2,

95% O2.

2.3.9 Solutions

The standard HEPES-buffered solution (HBS) contained (mM): 113.6 NaCl, 5 KCl,

2 NaH2PO4, 1 CaCl2, 1.2 MgSO4, 32.5 mM HEPES and 10.5 glucose titrated to pH 7.4 at

+ 37 °C with NaOH. 20 mM NH4Cl was substituted for 20 mM NaCl in the NH3/NH4 - containing solution. Solution osmolalities were measured using a vapor-pressure osmometer (5520, Wescor Inc, Logan, UT), and were adjusted to an osmolality of 300 ± 5 mosmol/kg.

2.3.10 Compounds

Stocks of cariporide (4-isopropyl-3-methylsulfonylbenzoylguanidine methanesulfonate, Cat. # 5358, Tocris Bioscience, Bristol, UK) and S3226 (3-(54)-N- isopropylidene-2-methyl-acrylamide dihydrochloride, Sigma-Aldrich, St. Louis, MO) were dissolved in 100% dimethlysulphoxide (DMSO) and the stock was diluted so that the working solutions contained ≤ 0.1% DMSO. A working concentration of 20 M cariporide was applied to block both NHE1 and NHE2 [170, 171], whereas 10 M S3226 selectively inhibits NHE3 [172]. Linaclotide stocks were dissolved in sterile H2O.

2.3.11 Intracellular pH recovery assay

NHE activity was assayed as the maximum rate of intracellular pH (pHi) recovery after an acid load (d(pHi/dt)max) imposed by a 20 mM NH4Cl pre-pulse followed by return to HBS buffer [173, 174]. pHi recordings were performed using a previously described imaging setup [175, 176]. Briefly, test compounds were added to the culture media of the

Caco-2 cell coverslips 30 min before commencing recording. For the final 10 min of the incubation, media was exchanged for HBS containing 10 M of the pH-sensitive dye precursor (2',7'-bis-2-carboxyethyl)-5(and-6) carboxyfluorescein acetoxymethyl ester

(BCECF-AM; B1170, ThermoFisher Scientific, Waltham, MA) and any appropriate test compound/s. The coverslip was then removed from the incubator and assembled to form the floor of the perfusion chamber, which was then mounted on the automated stage of an

Olympus IX-81 microscope equipped with epi-fluorescence imaging. Flow of HBS at 37

C (± S3326 or cariporide, as appropriate) commenced immediately, along with the excitation of the BCECF, alternately at 440 nm and 490 nm. Emitted light was captured at wavelengths >530 nm, to record changes in pHi. The exposure time for each of the two excitation wavelengths was 100 ms, separated by ~20 ms. A 440-nm (I440) and a 490-nm

(I490) intensity data pair was acquired every 5 seconds. Between the collection of the data pairs, incident light was obstructed. Solutions were delivered to the experimental chamber at 3 ml∙min−1 using syringe pumps (Model 33, Harvard Apparatus, Holliston, MA).

Solutions were selected using a computerized valve system, and maintained at a temperature of 37° C by means of a water-jacket system placed between the valves and the chamber. Slidebook 6.0.14 software (Intelligent Imaging Innovation, Denver, CO) provided data acquisition.

For each cell analyzed, an area of interest (AOI) was selected by using the outline tool in Slidebook to encompass the cell body. The rate constant describing the rate of change of I440 (−k440) was calculated continuously throughout experiments as an index of membrane integrity. Examples of −k440 time courses are shown in Error! Reference

–1 source not found.B (bottom panel). If the absolute value of −k440 is less than 5% min

[177], we regarded the cells as healthy and only included healthy cells in the analyses.

+ pHi values were calculated by using the high-K -nigericin technique [178]. At the end of each recording, Na-free, high-K+ solutions containing 10 M nigericin (N1495,

ThermoFisher Scientific) and 135 mM K+, buffered at five pH values of 5.8, 6.4, 7.0, 7.6, and 8.5 were applied for each coverslip. The I490/I440 ratio data for each cell is described by a pH titration curve:

I 10()pH− pK 490 =+ab I 1+ 10()pH− pK 440 (Equation 1)

Equation 1 was used to calculate pHi from I490/I440 ratios and the fitted values for pK and b for each cell. The initial rate of pHi recoveries from acid loads was analyzed by fitting the pHi vs. time record with a straight line.

2.3.12 Statistics

Results are expressed as the mean +/- S.E.M. Differences between groups were determined using either an ANOVA with post-hoc Tukey test or a two-tailed unpaired t test. A P value of <0.05 was considered significant. The mean d(pHi/dt)max data (Figure

5C) were analyzed by one-way analysis of variance (ANOVA) and we controlled for type

I errors across multiple means comparisons with a Holm-Bonferroni correction [179], setting the familywise error rate (FWER) to α = 0.05. Briefly, the unadjusted p-values for all N comparisons in each dataset were ordered from lowest to highest. For the first test, the lowest unadjusted p-value was compared to the first adjusted α value, α/N. If the null hypothesis is rejected, then the second-lowest p-value was compared to the second adjusted

α value, α/(N–1), and so on, until if at any point the unadjusted p-value is  the adjusted α, the null hypothesis is accepted and all subsequent hypotheses in the test group are considered null.

2.4 Results 2.4.1 Linaclotide does not activate F508del CFTR

Clinical anecdotal reports suggest that linaclotide may decrease intestinal manifestations in patients who carry the F508del mutation. Improvement of CF intestinal manifestations may be attributable to activation of mutant F508del CFTR by linaclotide.

To determine whether anion secretion through F508del CFTR occurs due to linaclotide treatment, we measured short circuit currents across intestinal sections before and after linaclotide treatment. Duodenum and ileum sections of wild type mice displayed a robust change of short-circuit current after the addition of linaclotide, 46.2 and 47.5 µA/cm2, respectively (Figure 2.1). However, the duodenum and ileum sections from F08del mice displayed no change in short-circuit current after linaclotide addition, 0.1 and 1.2 µA/cM2, respectively, indicating a lack of CFTR activation (Figure 2.1). This result suggests that activation of F508del CFTR is not responsible for the improvement in CF intestinal symptoms mediated by linaclotide that has been reported by patients.

2.4.2 Linaclotide increases gastrointestinal transit in CF mice

Transit of intestinal contents is significantly slower in the small bowel of patients with CF compared to non-CF individuals [180]. Some treatments for CF gastrointestinal manifestations, such as osmotic laxatives, are aimed at improving intestinal transit. To test whether observations by CF patients experiencing reduced intestinal pain and improved bowel movement while on linaclotide originate from improved intestinal transit, we assessed rates of GI transit in CF mice carrying the F508del allele or a null allele in response to linaclotide treatment. The effect of linaclotide on intestinal transit in F508del and WT mice was assessed by administering 50 µg/kg linaclotide or vehicle control by oral gavage and assessing the geometric center of fluorescence (GCF). In this assay, a higher

GCF indicates faster transit of intestinal contents. Consistent with previous reports [51, 82,

161], CF mice had significantly slower GI transit than control mice (GCF: F508del- 2.87

± 0.39, null- 2.77 ± 0.42 and control- 6.16 ± 0.32; P<0.001; Figure 2). Linaclotide treatment in both the F508del and null mice increased intestinal transit compared to non- treated F508del and null mice (GCF: treated F508del- 5.73 ± 0.41, treated null- 5.16 ± 0.4;

P<0.001 compared to untreated groups, Figure 2.2). Additionally, there was a significant increase in transit between linaclotide treated and non-treated WT mice (GCF: non-treated

6.16 ± 0.32 vs treated 7.35 ± 0.48; P<0.05) however the magnitude of transit increase was less than that in linaclotide treated CF mice due to the fluorescent dye reaching the end of the small intestine. These data highlight two points regarding linaclotide treatment in CF:

1) linaclotide increases GI transit in CF mice and 2) this improved transit occurs independent of CFTR protein.

2.4.3 Linaclotide does not induce fluid secretion in the intestine

The increased intestinal transit observed in null mice treated with linaclotide suggests a CFTR-independent mechanism of action. CFTR is one of many downstream targets of cGMP signaling, making it possible that a non-CFTR Cl- channel or any other ion channel that can induce fluid secretion could be activated by linaclotide [53, 181-183].

We sought to assess whether fluid secretion through a non-CFTR ion channel is activated by linaclotide. Therefore, we generated intestinal organoids from WT, F508del and Cftr null mouse small intestines. Activation of CFTR in intestinal organoids increases the salt content in the central organoid lumen, drawing fluid into the lumen. The increased fluid content of the lumen results in a prominent swelling phenotype (Figure 2.3A). Activation of Cl- secretion by an alternate Cl- channel or an ion channel that induces fluid secretion is

64 likely to create a similar swelling phenotype. Swelling was observed only in WT organoids following treatment with 10 µM linaclotide, with null and F508del organoids displaying no swelling (Figure 2.3A). The organoid swelling was quantified as a change in area with a significant increase in area of WT organoids versus no change in area of CF organoids

(Normalized percent area at 50 minutes: WT: 146 ± 1.9; F508del: 103 ± 0.4; null: 102 ±

0.9, Figure 2.3B). In an additional analysis, area under the curve (AUC) demonstrated significant differences between WT and both CF intestinal organoid lines as well (WT:

1537.5 ± 85.5, F508del: 138.9 ± 17.1, null: 156.9 ± 29.5, p<0.0001, Figure 2.3C). The lack of swelling in F508del and null organoids indicates that linaclotide does not induce fluid secretion in the CF intestine.

2.4.4 Linaclotide inhibits fluid absorption in the CF intestine

Another mechanism by which linaclotide could improve GI transit is inhibition of fluid absorption. Decreased absorption of fluids may improve fluidity of the intestinal lumen, increasing GI transit. To determine whether inhibition of fluid absorption by linaclotide can improve fluid content in the CF intestine, we assessed the rate of fluid absorption from intestinal segments using an in vitro intestinal fluid absorption assay. In this assay, increased retention of segment weight is indicative of an increased lumen solute content. The weight gain or loss of the segment over time reflects net fluid flux which

+ + - - includes CFTR-dependent fluid secretion and parallel Na /H and Cl /HCO3 exchanger- dependent fluid absorption. Untreated WT segments retained slightly more fluid than untreated CF segments likely due to CFTR dependent basal fluid secretion, but the difference between the two groups was not significant (P=0.074). The weight of WT segments treated with linaclotide stayed consistent during the experiment, indicating

65 activation of CFTR and increased fluid content of the intestinal lumen compared to untreated WT segments (percent final weight normalized to starting weight- WT Untreated:

84 ± 1.4, WT + linaclotide: 98 ± 1.4 Figure 4A P<0.001). CF intestinal segments treated with linaclotide retained more weight than untreated CF intestinal segments (CF Untreated:

80 ± 1.9, CF + linaclotide: 91 ± 0.6, Figure 2.4A, P<0.001). The retention of fluid in linaclotide treated CF segments was also greater than WT untreated segments (WT untreated: 84 ± 1.4, CF + linaclotide: 91 ± 0.6; Figure 2.4A, P<0.05). When assessing the weight lost per minute over the entire experiment, untreated segments had a greater rate of loss than linaclotide-treated WT segments (WT untreated: 0.36 ± 0.04 (mg/min), WT + linaclotide: 0.04 ± 0.03, P<0.005) and untreated CF segments had a greater rate of weight loss than linaclotide treated CF segments (CF untreated: 0.41 ± 0.03, CF + linaclotide: 0.24

± 0.02, P<0.005, Figure 2.4B). In WT segments, linaclotide activates CFTR and fluid secretion occurs. In CF segments, CFTR is absent, so no fluid secretion occurs, but linaclotide still leads to increased fluid retention, likely due to inhibition of fluid absorption. The loss of weight in CF segments in the presence of linaclotide may be due to incomplete inhibition of NHE3 by linaclotide or other sodium coupled absorptive processes not affected by linaclotide. To assess whether these in vitro results were representative of the in vivo intestine, we measured fluidity of intestinal contents of CF mice either treated with linaclotide or water. Linaclotide treatment was sufficient to increase the luminal fluid content in the intestine of CF mice compared to water treatment (wet/dry ratios; untreated:

3.33 ± 0.19, linaclotide treated: 3.92 ± 0.16, Figure 4C; P<0.05). These results suggest that linaclotide increases the fluidity of the CF intestinal lumen, possibly by inhibiting fluid absorption in the absence of CFTR.

2.4.5 Linaclotide inhibits intestinal sodium absorption

In addition to having a stimulatory effect on CFTR, cGMP signaling has been shown to inhibit sodium absorption by blocking intestinal Na+/H+ Exchanger 3 (NHE3)

[158, 184, 185]. NHE3 has been implicated as the primary regulator of sodium absorption in the intestinal epithelium [84]. Knockout of NHE3 in CF mice has been shown to reduce incidence of obstruction [87], indicating that blockage of NHE3-mediated Na+ absorption is sufficient to improve GI pathology. Inhibition of NHE3 by cGMP signaling may be responsible for reduced fluid absorption during linaclotide treatment. The effect of

+ + linaclotide on Na /H exchange was assessed by measuring the d(pHi/dt)max in CFTR-null

+ — Caco-2 cells, following an NH4 -induced acid load in nominally CO2/HCO3 -free HBS.

Caco-2 cells express endogenous NHE1, NHE2 and NHE3 [186]. Thus, in this assay, a reduction in d(pHi/dt)max would primarily reflect decreased NHE activity, which would be accompanied by a decrease in Na+ uptake. The data show linaclotide pre-incubation significantly reduced d(pHi/dt)max compared to non-treated controls (Figure 2.5A, upper panel and B). The magnitude of the linaclotide-mediated inhibition was equivalent to that of S3226, a specific inhibitor of NHE3 [172]. Co-treatment with linaclotide and S3226 caused no significant additional reduction d(pHi/dt)max compared to linaclotide or S3226 alone. 20 M cariporide caused a substantial reduction in d(pHi/dt)max. Cariporide is a specific inhibitor of NHE1 in the sub-micromolar range that inhibits NHE2 at doses 20 times those required to inhibit NHE1 [170, 171], Moreover, co-treatment with linaclotide and cariporide significantly reduced d(pHi/dt)max beyond the effect of cariporide alone which inhibits NHE1 and NHE2 but does not target NHE3. For all experiments, the

–1 absolute value of −k440 was <5% min , confirming cell viability (Figure 5A, lower panel).

The results of Figure 5A and B suggest that linaclotide primarily inhibits the NHE3- mediated component of Na+/H+ exchange. Collectively, the results in Figure 2.5 are consistent with the hypothesis that linaclotide-mediated cGMP signaling inhibits NHE3, leading to reduced Na+/H+ exchange, and, reduced Na+ absorption.

2.4.6 NHE3 inhibition is sufficient to improve GI transit in CF mice

We assessed the effects of NHE3 inhibition on GI transit in the CF mouse intestine using tenapanor, a highly specific NHE3 inhibitor [89, 90]. We hypothesized that if NHE3 inhibition is sufficient to increase GI transit in the CF intestine, tenapanor and linaclotide would increase GI transit in a similar manner. We measured transit of intestinal contents in response to 1 mg/kg tenapanor treatment in null mice. Tenapanor significantly increased the rate of intestinal transit compared to non-treated CF mice (GCF no treatment: 2.79 ±

0.25 vs tenapanor: 5.05 ± 0.48; P=0.0003; Figure 2.6A). The increase in GI transit in response to tenapanor was not significantly different from the increase in response to treatment with 50 µg/kg linaclotide (GCF linaclotide: 5.15 ± 0.42 vs tenapanor 5.05 ± 0.48,

P=0.1). In addition, CF mice treated with both linaclotide and tenapanor displayed a similar

GI transit to those treated with either drug alone (4.39 ± 0.42). This result suggests that

NHE3 inhibition by linaclotide is sufficient to improve GI transit in CF mice. Last, WT and null intestinal organoids were treated with tenapanor and displayed no CFTR- dependent swelling (Figure 2.6B and C) confirming the CFTR-independent mechanism of tenapanor.

2.5 Discussion The intestinal manifestations of CF are a common source of pain and morbidity among CF patients. Current treatments for CF intestinal manifestations do not sufficiently

68 prevent or reduce the symptoms of CF intestinal disorders, as indicated by continued discomfort despite the utilization of these currently available treatments. Thus, new therapies are necessary to increase the quality of life of CF patients. Linaclotide is a GCC agonist which activates CFTR through cGMP signaling to increase luminal fluidity in the intestine [154]. In non-CF patients, linaclotide has been demonstrated to be effective at decreasing symptoms of IBS-CC [187]. As the primary mechanism of action is through

CFTR, linaclotide was not predicted to be effective for patients with CF, and thus was not widely prescribed. However, anecdotal reports from patients with CF who were prescribed linaclotide at Rainbow Babies and Children’s Hospital and reported GI improvement suggest that linaclotide may be beneficial in CF. The goal of this study was to examine this observed improvement in CF related intestinal manifestations by utilizing mouse models with CF which develop similar GI manifestations to patients with CF.

To test the effectiveness of linaclotide, we utilized two independent CF mouse models, one carrying a null Cftr allele and one carrying an F508del allele. Surprisingly, both CF models displayed improved intestinal transit upon treatment with linaclotide. Our data suggest that this improvement was not due to activation of mutant CFTR (in the

F508del model), nor activation of other ion channels that would induce fluid secretion.

This result is in agreement with Pattison et al., who did not observe intestinal organoid swelling following linaclotide treatment with CFTR inhibition [167]. Rather, our data supports a CFTR independent mechanism for linaclotide in CF.

Increasing fluid content in the CF intestine has proven to have significant benefits for decreasing CF intestinal pathologies. Patients with CF are commonly treated with osmotic laxatives to treat constipation and DIOS. Similarly, osmotic laxatives prevent

- - intestinal obstruction in CF mice [72]. Talniflumate, an inhibitor of intestinal Cl /HCO3

- - exchange, has increased survival in CF mice [188]. Inhibition of Cl /HCO3 exchange likely leads to inhibition of Na+/H+ exchange through NHE3 [188]. NHE3 is one of several members of the NHE family, displays exclusive expression in the intestine and kidney [86], and has been demonstrated to have a significant absorptive function in these organs [83].

NHE3 and CFTR work in conjunction with one another to regulate salt and fluid balance in the intestine, as both proteins co-localize at the apical membrane in the same regulatory complex [189]. Interestingly, NHE3 expression is reduced in the absence of CFTR and

CFTR expression is reduced in the absence of NHE3 suggesting a response of the cell to balance electrolyte and water absorption [190]. Deletion of NHE3 in the mouse leads to a significant reduction in Na+ absorption in the intestine, increased intestinal luminal fluidity, and severe diarrhea [83, 191]. Inhibiting NHE3 in CF has been hypothesized as a possible therapeutic role for CF intestinal pathologies given the contrasting roles they serve in intestinal fluid regulation [87, 190]. Bradford et al, tested this hypothesis by creating a double knockout mouse of CFTR and NHE3 [87]. NHE3 knockout in CFTR null mice restored hydration of the intestinal lumen which decreased the incidence of intestinal obstruction and thus increased survival compared to CF mice with function NHE3 [87].

These data suggest that hydration of the intestinal lumen can be preserved by blocking intestinal fluid absorption in the absence of fluid secretion, and are consistent with what we observed in the present study. Treating CF mice with linaclotide increased intestinal fluid retention hydrating the intestinal lumen and improving intestinal transit. The absence of CFTR, the main reported mechanism of action for this drug, did not prevent linaclotide from producing significant improvements in CF mice. While this result was initially

70 surprising, studies have suggested that the water accumulation in the intestine through the

GCC receptor and cGMP signaling is more of a consequence of impaired fluid absorption through NHE rather than the result of secretion of electrolytes through CFTR [192, 193].

This is supported by our treatment of CF mice with tenapanor, a selective NHE3 inhibitor

[88-90], which displayed similarly improved intestinal transit as CF mice on linaclotide independent of CFTR. Though tenapanor is not the focus of this study, the data presented here suggest that tenapanor may also be effective at improving GI transit in CF patients through NHE3 inhibition but further study is necessary to ensure its safety and efficacy.

Whether through linaclotide or tenapanor, blockage of Na+ absorption by inhibition of

NHE3 may be sufficient for treating certain CF small intestinal pathologies. In addition, our data utilizing colon cells suggests that linaclotide mediated NHE3 inhibition occurs in the CF large intestine as well but further studies are necessary to establish if linaclotide improves transit in the CF large intestine as it does in non-CF patients [162, 194]. While our studies suggest NHE3 as the main target of linaclotide, a potential direct effect on other transporters in the small intestine like Slc26a3 (DRA) or Slc26a6 (PAT1), that also effect in fluid absorption [195], cannot be ruled out.

While our data clearly support a CFTR independent mechanism for increasing intestinal luminal fluid, further studies are necessary to explore possible activation of mutant CFTR by linaclotide through the guanylate cyclase pathway. Previous work has shown that mutant F508del CFTR has small levels of activity in the nasal epithelium when stimulated with C-Type Natriuretic Peptide (CNP) and forskolin [196]. CNP increases cellular cGMP levels through interactions with guanylate cyclases, a pathway similar to that used by linaclotide in the intestine. Forskolin increases cAMP levels, also contributing

71 to CFTR activation. This suggests that cGMP signaling alone may be inadequate to activate

F508del CFTR, but a combination of cAMP and cGMP signaling may allow F508del

CFTR activation. In addition, reduced Cftr expression has been noted in the F508del mouse model [197], most likely producing less F508del CFTR to activate at the apical membrane.

Interestingly, linaclotide has been shown to increase WT CFTR trafficking to the cell membrane through the PKGII and possibly PKA pathways [198], raising the possibility that mutant CFTR trafficking could also be modified. Work from Arora et al, support this suggestion as they observed a small but significant increase of forskolin-induced swelling in F508del patient derived organoids after 24 hour exposure to a GCC agonist [199].

Our results highlight the utility of cGMP signaling to increase intestinal lumen fluid content for the treatment of CF intestinal manifestations. cGMP signaling can be activated in a wide range of tissue types, including the airway, where absence of Cl- secretion leads to Na+ hyper absorption, contributing to airway dehydration [200, 201]. Dehydration in the airway leads to infection, inflammation, reduced lung function, and mortality in patients with CF [17, 202-204]. For many CF patients, therapies which restore function to CFTR are not available, thus restoration of Cl- secretion is not a viable therapy. However, our data indicate that blockage of Na+ absorption may increase the fluidity of the airway lumen, helping to alleviate the epithelial tissue dehydration found in CF patients. While the GCC receptor is only expressed in the intestine [205, 206], other guanylate cyclase receptors are present in other tissues, including the airway epithelium [207]. Activation of cGMP signaling has been shown to inhibit Na+ absorption through inhibition of epithelial sodium channels (ENaC) [196, 208]. Similar to NHE3, ENaC has been proposed to be a therapeutic target for CF and ENaC inhibition through cGMP signaling in the airway could have

72 hydrating effects on the airway lumen [201, 209]. While the pharmacological effects of linaclotide are restricted to the intestine, the effect of linaclotide on cGMP signaling in CF demonstrates that blocking Na+ absorption may be effective for hydrating the epithelial surface of the airway, which may alleviate certain CF airway pathologies.

In conclusion, the study presented here indicates that linaclotide may restore luminal fluidity to the dehydrated CF intestine leading to improved GI transit. The mechanism of action in CF is CFTR independent and relies on inhibition of Na+ absorption through NHE3. Constitutive inactivation of NHE3 in the CF mouse has been previously studied and led to similar increased intestinal lumen fluidity, decreased formation of obstructions, and improved survival [87]. These results indicate that chronic treatment of patients with CF with linaclotide, or NHE3 inhibitors, could result in improvement of CF intestinal manifestations. Identifying which CF intestinal pathologies, such as goblet cell hyperplasia, bacterial overgrowth, and inflammation, are alleviated is still necessary. For instance, the clearance of mucus from the lower villus region and crypt region through

NHE3 inhibition may not be as effective, as NHE3 is present in higher amounts in the upper villus compared to the lower villus [87, 210]. A combination of linaclotide and osmotic laxative may be necessary in CF patients to achieve the highest effectiveness in alleviating intestinal discomfort caused by reduced GI transit. A well designed clinical study to assess specific improvements in GI manifestations in CF patients prescribed linaclotide is necessary. These studies suggest that the FDA-approved drug linaclotide may be effective in treating CF intestinal pathologies since an identified mechanism of action is CFTR independent. In addition, these studies suggest that cGMP signaling leading to

73 inhibition of Na+ absorption is a therapeutic target in the CF intestine and perhaps in other organs as well.

Acknowledgments

We thank Alma Wilson, Molly Halligan and Amanda Barabas of the CF Mouse Models

Core at Case Western Reserve University for their assistance in maintaining the mouse colony. We thank Dr. Scott Howell of the Case Western Reserve Microscopy and Digital

Imaging Core for his assistance in imaging organoid swelling. We thank Dr. Patricia

Conrad of the Case Western Reserve School of Medicine Light Microscopy Core for her assistance in quantifying organoid swelling.

Figure 2.1. Linaclotide does not induce a change in short-circuit current in F508del intestinal tissue. A. Representative short-circuit current tracings from two WT and F508del duodenums treated with 10 µM linaclotide (Indicated by the black arrow). B. Average change in short-circuit current measurements from the duodenum and ileum of WT and F508del mice. (Isc was averaged from a minimum of 4 mice per group ± SE; *P<0.05.)

Figure 2.2. Linaclotide treatment improves GI transit in F508del and Cftr null mice. Adult WT, null, or F508del mice were treated with either 50 µg/kg linaclotide or vehicle control. (Bars represent mean ± SE ;*P<0.01 vs WT, ^P<0.05 vs vehicle group of the same genotype, n ≥ 7).

Figure 2.3. Linaclotide does not induce swelling of F508del or null intestinal organoids. A. Representative images at T=0 and T=50 minutes of intestinal organoids harvested from WT, F508del, or null mice following treatment with 10 µM linaclotide. Scale bar = 100 µm. B. Total change in area was measured over 50 minutes and normalized to area at T=0. Each curve represents the average of at least four wells. C. AUC for normalized measurements for each treatment group (Baseline set to 100%; *P<0.01, n ≥ 4 wells).

Figure 2.4. Linaclotide-treated CF intestinal segments retain more fluid than untreated. A. The change in weight over time of duodenal segments from adult null and WT mice which were linaclotide-treated or untreated (*P<0.001 compared to untreated controls for both WT and CF, n ≥ 6 segments from at least 3 mice per group). Treatment groups are represented by best fit lines. B. Rate of weight loss (mg/minute) in each treatment group pictured in A (*P<0.01 vs untreated groups for both WT and CF). C. Fluidity of the intestinal contents in linaclotide-treated and untreated null animals (*P<0.05; n ≥ 5).

Figure 2.5. Linaclotide treatment inhibits the S3226-sensitive component pHi recovery after an acid load in Caco-2 cells. A. Upper panel; Representative traces of pHi recordings from Caco-2 cells pretreated with different combinations of linaclotide and + NHE inhibitors, and then and then subjected to the application and withdrawal of NH4,

79 which imposes an intracellular acid load. Lower panel displays the rate constant describing the loss of the I440 signal (−k440) for each of the cells displayed in the pHi recordings. Cells –1 are considered healthy if the absolute value of −k440 is less than 5% min . Linaclotide, cariporide and S3226 were pre-incubated with the cultures at 10, 20 and 10 M respectively. Cariporide and S3226 were perfused throughout the experiment at the same concentrations at the pre-incubations. B. Summary of maximal rates of pHi recovery from + intracellular acid loads, imposed by NH4 prepulses, for each of the conditions in A. Bars represent mean ± SE; number of replicates are in parentheses at the bottom of each bar. All six mean values are significantly different from each other, except for the 3 comparisons indicated by N.S. above the brackets. Significant differences were determined by one-way ANOVA, with a Holm-Bonferroni correction (see Methods).

Figure 2.6. Targeted Inhibition of NHE3 improves GI transit in null mice. A. Rates of GI transit were measured in adult null mice treated with water, 1 mg/kg tenapanor, 50 µg/kg linaclotide, or both tenapanor and linaclotide. (*P<0.005, n ≥ 5). B. Intestinal organoids were harvested from WT or null and treated with either 10 µM linaclotide or 10 µM tenapanor to assess CFTR dependent swelling. Total change in area was measured over 50 minutes and normalized to area at T=0. Each curve represents the average of at least four wells. C. AUC for normalized measurements for each treatment group (Baseline set to 100%; *P<0.01, n ≥ 4 wells).

Chapter 3 – A G542X Cystic Fibrosis Mouse Model for Examining Nonsense

Mutation-Directed Therapies

The research in this chapter is published in: McHugh, D. R., et al. (2018). "A G542X cystic fibrosis mouse model for examining nonsense mutation directed therapies." PLoS One 13(6): e0199573.

3.1 Abstract Nonsense mutations are present in 10% of patients with CF, produce a premature termination codon in CFTR mRNA causing early termination of translation, and lead to lack of CFTR function. There are no currently available animal models which contain a nonsense mutation in the endogenous Cftr locus that can be utilized to test nonsense mutation therapies. In this study, we create a CF mouse model carrying the G542X nonsense mutation in Cftr using CRISPR/Cas9 gene editing. The G542X mouse model has reduced Cftr mRNA levels, demonstrates absence of CFTR function, and displays characteristic manifestations of CF mice such as reduced growth and intestinal obstruction.

Importantly, CFTR restoration is observed in G542X intestinal organoids treated with

G418, an aminoglycoside with translational readthrough capabilities. The G542X mouse model provides an invaluable resource for the identification of potential therapies of CF nonsense mutations as well as the assessment of in vivo effectiveness of these potential therapies targeting nonsense mutations.

3.2 Introduction Cystic Fibrosis (CF) is an autosomal recessive genetic disorder caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. CFTR is an anion channel expressed throughout the body with highest expression in epithelial tissues.

Absence of CFTR function impairs transepithelial fluid and electrolyte permeation and results in viscous mucus along the epithelial lining of organs leading to a wide range of disease manifestations. Common manifestations of CF include lung failure [17, 201, 211], pancreatic insufficiency [212], intestinal obstruction [51, 55, 153, 213], and reduced growth [214-216]. Over 2000 unique CFTR variants have been identified, with ~300

84 categorized as definitively CF-causing mutations [217, 218]. CF-causing nonsense mutations are the second most common CF mutation type, and are found in approximately

10% of patients with CF [20, 112]. Nonsense mutations create a premature termination codon (PTC) in CFTR, which leads to premature termination of CFTR translation [219].

Recently, there has been progress in the development of CFTR modulators that restore CFTR function in patients with specific CFTR mutations. Examples include VX-

770 (ivacaftor) for gating mutations like G551D [26, 95] and VX-809 (lumacaftor) for the misfolding mutation F508del [220, 221]. However, no therapies are available to CF patients that restore function to nonsense mutations in CFTR. Certain drugs, like gentamicin and PTC124 (ataluren), have been shown to interact with the ribosome to induce readthrough of PTCs by allowing insertion of a near-cognate aminoacyl tRNA and translation of a full-length protein [131, 222, 223]. Both gentamicin and PTC124 have been used in clinical trials for patients with nonsense mutations [224-227]. However, gentamicin treatment is associated with nephrotoxicity and ototoxicity [128] and PTC124 has produced no significant improvements in adult CF patients with nonsense mutations in Phase 3 clinical trials [228]. Therefore, no currently available treatment for CF nonsense mutations is both safe and clinically effective.

Because of the safety issues surrounding PTC readthrough strategies, animal models of CFTR nonsense mutations are essential for in vivo validation of new therapies.

Current CF mouse models with nonsense mutations were created in a way that either does not allow for correction of the nonsense mutation or [149, 160] or utilizes non-endogenous

CFTR with tissue specific expression [20]. Therefore, a model with global expression of the G542X mutation in the endogenous mouse Cftr will have a greater value for examining

85 the efficacy of nonsense mutation therapies on Cftr in multiple tissue types with native expression levels. In this manuscript, we describe the generation of a mouse model with expression of the G542X mutation, the most common nonsense mutation in CF [22], using

CRISPR/Cas9 gene editing technology. Mice homozygous for the G542X mutation have reduced Cftr expression and absence of CFTR function in the airway and intestine. These mice display typical cystic fibrosis manifestations such as poor growth and reduced survival due to intestinal obstruction. Importantly, we demonstrate that pharmacological readthrough of the G542X nonsense mutation in this model allows production of functional

CFTR. The G542X mouse will be a valuable model for the examination of nonsense mutation therapies for CF and other genetic diseases caused by nonsense mutations.

3.3 Materials and Methods 3.3.1 Generation of the G542X allele

To produce the G542X mouse Cftr allele (CftrG542X) using the gene editing system

CRISPR/Cas9, guide RNAs (gRNA) were selected in exon 12 of mouse Cftr using CRISPR design software that identifies optimal gRNAs based on proximity to target as well as off- target cutting predictions [229]. Four of these gRNAs were tested in vitro using guide-it gRNA in vitro transcription and screening kit (Clontech) for the ability to guide Cas9 nuclease activity to the desired DNA sequence. One gRNA (25 ng/ul; PNABio), a 120 bp single stranded oligonucleotide (ssODN) containing the G542X mutation (25 ng/ul; IDT) centered on the cut site and either Cas9 mRNA (25-50 ng/ul; PNABio) or Cas9 protein (25 ng/ul; PNABio) were injected in the pronucleus of C57BL/6 one-cell embryos. Embryos were then placed in pseudo-pregnant females to develop. Ear punches from the 22 resulting pups were sequenced using next generation sequencing. The region of interest was PCR amplified (~300 bp) with flanking primers that were tailed with a universal sequence. A

86 secondary PCR amplification was then performed on the primary PCR product using a set of primers specific for the universal sequence and tailed with I5,I7 Illumina index and specific barcode sequences allowing for the sequencing of multiple samples at the same time. The mixture of barcoded PCR products were analyzed on an Illumina MiSeq instrument using a paired-end 300 base pair kit allowing for 25 million reads per run. Data were de-convoluted and variant analysis in comparison to the original sequence was performed using the Outknocker software program [230].

3.3.2 Mice

In some experiments, the G542X mouse model was compared to a previously published CFTR null model carrying a S489X mutation (Cftrtm1Unc) which is congenic on the C57BL/6J background [160]. Mice homozygous for these mutations were created by breeding heterozygous males and females and wild type littermates were used as controls.

Genotyping was completed by PCR analysis using DNA extracts from ear biopsies. To detect the G542X allele (319 bp) primers P1 (5’- ACAAGACAACACAGTTCTCT -3’) and P2 (5’ TCCATGCACCATAACAACAAGT -3’) were used. To detect the wildtype

(WT) allele (319 bp) P2 and P3 (5’- ACAAGACAACACAGTTCTTG -3’) were used in a separate reaction. PCR reactions were completed for 40 cycles of 95°C for 30 seconds,

58°C for 30 seconds and 72°C for 30 seconds and products were run out on 2% agarose gels. All mice were allowed unrestricted access to water and solid chow (Harlan Teklad

7960; Harlan Teklad Global Diets). All animals were maintained on a 12-h light, 12-h dark schedule at a mean ambient temperature of 22°C and were housed in standard polysulfone microisolator cages in ventilated units with corncob bedding. Animals were monitored on a daily basis, and weight was assessed every 5 days from 10 to 40 days of age. Length of

6-wk-old euthanized mice was assessed from nose to anus by use of digital calipers. For

G418 treatment, mice were intraperitoneal injected on three consecutive days (25mg/kg b.w.). Twenty four hours after last injection, mice were sacrificed and lungs were flash frozen and RNA prepared as described below. The Institutional Animal Care and Use

Committee of Case Western Reserve University approved all animal protocols.

3.3.3 Expression Analysis

One µg of RNA was reversed transcribed in cDNA using QScript cDNA synthesis kit (VWR). Real-time quantitative PCR was performed on a StepOne PCR system

(Applied Biosystems). Cftr expression was assessed using a TaqMan expression assay which used primers spanning exon 17 and 18 (Mm00445197; Applied Biosystems).

Expression was normalized to β-actin as the endogenous control. Each RNA sample was used to make cDNA in duplicate and the expression results were then averaged to yield the final value. The average of each sample was then expressed as a percentage of wildtype expression.

3.3.4 Bioelectric Measurements

Nasal potential difference (NPD) measurements were obtained as previously described [164, 231]. NPD (mV) was assessed after the addition of chloride-free HEPES- buffered saline containing 10 µM forskolin. Short circuit measurements on intestinal sections were obtained as previously described [164]. The change in short-circuit current was calculated after the addition of 10 µM forskolin and 100 µM IBMX to the basolateral side of the intestinal sections.

3.3.5 Intestinal Organoid Harvesting and Culture

Intestinal organoids were harvested from adult mice as described by others [165].

Briefly, the mouse was sacrificed, the small intestine removed and flushed using PBS. The intestine was cut longitudinally, villi removed with a razor blade and the remaining intestine was cut into ~0.5 cm segments and suspended in Gentle Cell Dissociation Reagent

(STEMCELL Technologies) in a 50 mL conical tube for 30 minutes under gentle agitation with a shaker. In a sterile tissue culture hood, the intestinal segments were vigorously shaken by hand for 30 seconds, and the supernatant was deposited in a 10 cm dish. The segments were resuspended in ice-cold PBS lacking Mg++ and Ca++, and vigorously shaken again for 30 seconds. This process was repeated until four fractions of supernatant were produced. The fraction which was most enriched for intestinal crypts was filtered using a

100 µm cell strainer, and pelleted at 450xg for 10 minutes. The supernatant was discarded, and the pellet was resuspended in 500 µl of a 1:1 mixture of MatriGel (Corning) and

Intesticult Organoid Growth Media (OGM; STEMCELL Technologies). Crypts were diluted to a concentration of 10 crypts/µl, and plated into non-tissue culture treated 24-well plates, with 35 µl of MatriGel:OGM mixture deposited in each well. The MatriGel:OGM was solidified at 37°C for 15 minutes, then 500 µl of OGM was gently deposited in each well. The organoids were stored in a 37°C incubator with 5% CO2. The OGM was changed once every three days, and the organoids were passaged once every 7 days.

3.3.6 Measurement of Intestinal Organoid Swelling

Forskolin-induced swelling (FIS) of intestinal organoids was performed similar to previously described methods [167, 168]. Briefly, intestinal organoids which had been cultured for 6-8 days were split into non-tissue culture treated 48-well plates (Genesee

Scientific) and allowed to grow for 24 hours before drug treatment. Organoids were treated with G418 (20-100 µM) or PTC124 (1-20 µM) for 72 hours prior to FIS measurements.

FIS was assessed by treating organoids with 10 µM forskolin and imaging for 5 hours under live cell conditions with brightfield microscopy on a Lionheart FX Automated Microscope

(BioTek). FIS was quantified by identifying and normalizing organoid area to T=0 with

Gen5 ImagePrime software (BioTek). Area under the curve (AUC) at T=300 minutes was calculated to compare statistical significance between treatment groups. All organoids used for swelling quantifications were between passages 1-4.

3.3.7 Statistics

Results are expressed as the mean +/- S.E.M. Differences between groups were determined using either an ANOVA with post-hoc Tukey test or an unpaired t test. A P value of <0.05 was considered significant.

3.4 Results 3.4.1 Generation of the G542X Mutation

To generate the G542X mutation in mouse Cftr, the CRISPR/Cas9 gene editing system was utilized. Guide RNAs (gRNAs) in proximity to the DNA sequence corresponding to glycine at position 542 in mouse Cftr exon 12 were selected and tested in vitro for their ability to support efficient Cas9 nuclease activity to the DNA region as indicated by the reduction of full length DNA amplicon (Figure 3.1A). The gRNA that supported the highest Cas9 nuclease activity was injected in one-cell mouse embryos along with a ssODN containing the G542X mutation and either Cas9 mRNA or protein (Figure

3.1B). DNA from all mice originating from these injected embryos was sequenced. Of 22 mice, 9 (40.9%) had at least one Cftr allele containing the G542X mutation due to homology-directed repair (HDR) with the ssODN, 5 (22.7%) did not have the G542X 90 mutation but had other mutations (insertions/deletions) due to nonhomologous end joining and 8 (34.4%) had no mutation of Cftr in exon 12. One of the founder mice containing the

G542X allele was selected to establish and expand the G542X colony.

3.4.2 Cftr expression is reduced and CFTR function is absent in G542X mice

Cftr expression was assessed using qRTPCR from various tissues of G542X and

WT littermates. Cftr expression levels in tissues from G542X mice were significantly reduced compared to WT expression levels ranging from 2.5±0.5% in the ileum to

28.4±8.6% of WT levels in the lung (Figure 3.2). Assessment of CFTR function in the airway was performed using NPD measurements in the nasal lumen. Activation of CFTR and the subsequent Cl- secretion into the nasal lumen causes a decrease in voltage across the nasal epithelium. Mice with functional CFTR display a decrease in NPD (-14.6±1.6 mV) while G542X mice have a slightly positive NPD (4.38±2.1 mV; Figure 3.3A) indicating no functional CFTR similar to other CF mouse models [164, 196, 231, 232].

CFTR function in the intestinal epithelium was examined by measuring short-circuit current (ΔIsc) in intestinal segments from G542X and WT mice before and after activation of CFTR. In WT intestinal segments, a significant increase in short-circuit current is observed following CFTR stimulation. Corresponding segments from G542X mice displayed significantly reduced changes in short-circuit current (Figure 3.3B), similar to other CF mouse models [138, 159, 164].

3.4.3 G542X mice display characteristic CF manifestations

The most common cause of morbidity in CF mouse models is intestinal obstruction

[51, 61, 139, 164]. G542X homozygous mice displayed a reduced survival rate compared to WT littermates (33.3% vs. 97.9%; Figure 3.4A). Intestinal obstruction was observed in

91 all G542X mice post-mortem but not in WT littermates. Another common disease manifestation in CF mouse models is reduced growth, or failure to thrive [159, 160, 164,

233]. G542X homozygous mice demonstrated significantly reduced length compared to sex-matched 6-week old WT littermates (in cm: males 7.6±0.2 vs. 8.4±0.1; females 6.0±0.3 vs. 7.8±0.1; Figure 3.4B). To assess growth, offspring from G542X heterozygote crosses were weighed at 5 day intervals from 10-40 days of age. G542X homozygous mice displayed significantly reduced weight at all time points compared to sex-matched WT littermates (Figure 3.4 C and D).

3.4.4 G542X CFTR function is restored following pharmacological nonsense mutation readthrough by G418 but not by PTC124

The ability to produce functional CFTR following treatment with potential therapies is crucial for the G542X mouse model to have full utility as a model of CF nonsense mutations. To examine CFTR function following readthrough, we generated intestinal organoids from G542X homozygous mice. Intestinal organoids are a three- dimensional cell culture model produced from LGR5+ stem cells in the intestinal crypt with budding outgrowths and a hollow lumen [168]. Activation of CFTR by forskolin allows flow of Cl- anions into the intestinal organoid lumen, increasing lumen osmolality, drawing in fluid, and causing organoid FIS. FIS is absent in organoids lacking CFTR function, allowing sensitive detection of CFTR activity. G542X organoids were treated for

72 hours with the aminoglycoside G418, a suppressor of nonsense mutations [234, 235].

FIS was examined by imaging organoid swelling over 300 minutes under live cell conditions with brightfield microscopy (Figure 3.5A). We observed a dose-dependent FIS

92 response to G418-induced readthrough indicating restored CFTR function (normalized percent area at 300 minutes: No G418: 114±1.7; 20 µM: 135±2.3; 50 µM: 160±3.0; 100

µM: 181±1.7, Figure 3.5B). Statistical significance between treatment groups was compared by calculating area under the curve (AUC) at 300 minutes for each treatment groups (Figure 5C). AUC of organoids treated with G418 were significantly increased compared to non-treated organoids (Figure 3.5C). S489X organoids, which are non- correctable Cftr-null mutants, did not display FIS following 72 hour treatment with 100

µM G418, confirming that FIS in G542X organoids is due to readthrough of nonsense mutations (Normalized percent area at 300 minutes: 113±0.7, Figure 3.5B-C). Similar to tissues from G542X homozygous mice, cultured intestinal organoids have a reduced amount of Cftr expression compared to those from WT littermates (24.0±1.0%) (Figure

3.5D). However, concordant with the increase in CFTR function in G418 treated organoids from G542X mice, G418 treatment also significantly increased the amount of detectable

Cftr mRNA (Figure 3.5D). Short term treatment of G542X mice with G418 also resulted in a significant increase in the amount of detectable Cftr mRNA in tissue (Figure 3.5E).

Interestingly, PTC124 treatment of the G542X organoid model did not stimulate detectable

FIS at any of the concentrations tested and thus resulted in no significant difference of

AUC between PTC124 treated and non-treated organoids (1-20 µM; Figure 3.6A-C).

3.5 Discussion Currently there are only two FDA approved therapies that directly target CFTR dysfunction, the primary defect of CF [26, 221]. These therapies are approved to modulate aberrant protein that is produced from the CFTR gene in patients with specific mutations that do not include nonsense mutations. Nonsense mutations typically result in little to no protein function. Reduced mRNA levels, through nonsense-mediated decay (NMD),

93 dramatically decreases the amount of protein produced. In addition, any protein that is synthesized is truncated due to the PTC and, depending on severity of the truncation, leads to significantly limited protein function [117, 236]. Current research efforts to increase the amount of full length protein from nonsense mutations include both inducing readthrough of the PTC and targeting the NMD pathway to increase mRNA levels [117, 236].

Readthrough of PTCs through nonsense suppression can be achieved with aminoglycosides that reduce the efficiency of termination and allow for the insertion of near-cognate tRNA [222]. Promising results in cell and mouse models of nonsense mutations have led to the use of the aminoglycoside gentamicin in patients [20, 222, 224,

237]. In CF, functional improvements in CFTR have been described in CF patients receiving topical application to the nasal mucosa [226] but renal and otic toxicities are a concern with long term treatment. Interestingly, a minor component B1 of gentamicin has been shown to provide effective readthrough activity [238], suggesting that an enrichment of this component of gentamicin or modified aminoglycosides with reduced toxicity may hold promise for the future [223]. A high-throughput screen identified PTC124 as a candidate compound with readthrough ability without the toxicity concerns of aminoglycosides [131]. Despite a lack of readthrough efficacy of PTC124 in specific in vitro reporter systems [239], PTC124 treatment of preclinical models of CF have displayed functional improvements [240]. While some initial patient studies suggested PTC124 efficacy in CF [225, 241], larger studies failed to show significant improvements in CF patients with nonsense mutations in Phase 3 clinical trials [228]. While PTC124 will not be used as a therapy for CF, investigation of the compound is still underway for other nonsense mutation genetic disorders such as Duchenne muscular dystrophy.

There has been recent success in the identification of possible nonsense mutation suppressors in CFTR using high-throughput screens [242-245]. The in vitro tools created in each of these screens made the identification of these hits possible. However, a CF animal model to carry out in vivo studies to test the efficacy of these hits is not available to the research community. The first Cftr mutation made in the mouse was a nonsense mutation created by replacing the normal exon sequence with the premature stop codon,

S489X along with selectable markers. The integration of the selectable markers in the exon precludes the production of functional protein [149, 160] and thus serves as a negative control for readthrough strategies. The only current animal model shown to produce functional CFTR following readthrough is a transgenic mouse that expresses human CFTR cDNA containing the G542X mutation driven by a rat Fatty Acid Binding Protein (FABP) promoter that results in intestinal-specific expression [20]. The conditions of CFTR expression in this model differ significantly from endogenous expression and readthrough of G542X can only be assessed in the intestine. In addition, because cDNA was used and there is no intron splicing, the mRNA produced is not susceptible to NMD which limits the interpretation of any in vivo studies. Non-endogenous expression levels and/or the absence of NMD of the resulting mRNA may explain why PTC124 displayed functional improvement in the intestine of hCFTR-G542X [240] model but failed to show similar improvements in patients.

In this study, we created a mouse model of CF containing the G542X mutation in the endogenous Cftr gene. We utilized the CRISPR/Cas9 gene editing system to create the

G542X mutation in one-cell mouse embryos. This is the first report of Cftr CRISPR/Cas9 editing to create a mouse model. As previously reported, the use of CRISPR/Cas9

95 significantly reduces the time of gene editing in embryos compared to embryonic stem cells [246]. In addition, this gene editing system is highly efficient in the creation of the desired mutation. Traditional gene editing in embryonic stem (ES) cells typically produced

1-5% of ES colonies harboring the correctly edited gene and could take 1-2 years to produce the targeting construct and correctly targeted founder mice. With CRISPR/Cas9,

40.9% of founder mice contained the G542X mutation and were produced in as few as 3 months. One disadvantage in utilizing CRISPR/Cas9 gene editing is the possible creation of double-strand breaks in similar sequences to the target which can create off-target effects

[247]. A low incidence of off-target mutations has been identified in mice created using

CRISPR/Cas9 [248, 249]. While off-target mutations can occur, subsequent backcrosses to a parental strain (e.g., C57Bl/6) can remove any unintended consequences from off- target mutations.

Mice homozygous for the G542X mutation demonstrate reduced Cftr expression throughout the body. While our analysis cannot rule out that a truncated form of CFTR protein is produced, there is a clear absence of CFTR function in the airway and intestine of these mice. The G542X mice also display severe CF manifestations, including reduced growth, and reduced survival due to intestinal obstruction. In this model, mRNA levels can be increased and CFTR function can be corrected through readthrough by G418.

Interestingly, the utilization of PTC124 in the intestinal organoid model did not demonstrate any improvement in CFTR function. This is consistent with studies that show no significant improvement in CF patients chronically administered PTC124. All of these findings confirm that this G542X mouse model can be utilized to test the efficacy of nonsense mutation therapies.

The G542X mouse model also provides an unlimited source of primary cells to potentially identify new therapies for nonsense mutations. For example, the prominent intestinal organoid FIS phenotype, which indicates CFTR function, has the potential to be used for high-throughput screens similar to those currently utilizing CF patient samples

[168, 250]. The high degree of similarity between organoids and tissue in vivo may increase the odds of identifying translationally beneficial compounds in a high-throughput screen compared to current screens using transfected cell lines or in vitro reporter assays.

In conclusion, this is the first study to generate a mouse model containing

CRISPR/Cas9 edited Cftr and the first CF animal model that can be utilized to assess whole body efficacy of nonsense mutation directed therapies. With several recent screens identifying potential PTC readthrough compounds, the G542X mouse model can be utilized to prioritize which compounds to test in a clinical trial that may be beneficial to CF patients.

The ability to analyze in vivo effectiveness of current and future potential nonsense mutation therapies will not only increase our understanding of how these therapies work on a basic level but lead to treatments for CF patients with nonsense mutations.

Acknowledgements

We thank Alma Wilson, Molly Schneider and Amanda Barabas of the CF Mouse Models

Core at Case Western Reserve University for their work in maintaining the mouse colony.

The work was supported by the Cystic Fibrosis Foundation (Hodges15XX0).

Figure 3.1. Generation of the G542X Mutation. (A) To validate gRNA efficiency in guiding Cas9 nuclease to the desired site, an in vitro assay was performed. PCR amplified DNA containing exon 12 and surrounding region of mouse Cftr (1056 bp) is displayed on an agarose gel with no gRNAs (-) or with 1 of 4 different gRNAs (1-4). Cas9 nuclease activity results in the cleavage of the DNA into fragments of ~750 bp and ~300 bp. (B) Normal Cftr mouse DNA and amino acid sequence around the desired mutation site is shown with the gRNA sequence (in box) and the protospacer adjacent motif sequence recognized by Cas9 (in blue). A portion of the sequence for the G542X oligo is also shown with the substitution change shown in red. The G to T mutation corresponds to the glycine to stop mutation. A silent T to C mutation was also inserted to assist with genotyping and verify HDR had occurred.

Figure 3.2. Cftr expression in tissues from G542X mice. Cftr expression in tissues from G542X and WT littermates were evaluated using qRTPCR. Airway (nasal epithelium, lung), intestine (duodenum, jejunum, ileum), and kidney Cftr expression as a percentage of WT Cftr expression is displayed. G542X expression was significantly reduced in all tissues compared to WT. (P<0.05 vs. WT. n≥5 per group

Figure 3.3. CFTR function in the airway and intestine of G542X mice. (A) NPD measurements and (B) change in intestinal short circuit current (ΔIsc) measurements from the duodenum, jejunum, and ileum from WT and G542X mice are shown. (*P<0.05 vs. WT. n≥4 per group)

100

101

Figure 3.4. Survival and growth characteristics of G542X mice. (A) Survival of G542X and WT mice up to 40 days of age. (B) Length of G542X and WT mice at 6 weeks of age. (C, D) Weight of G542X and WT mice sex-matched mice up to 40 days of age. Weight at every age was significantly different between G542X and WT littermates. (n≥10 per group. *P<0.05 vs. WT)

102

103

104

Figure 3.5. Intestinal organoids and tissue from G542X mice are used to test G418- mediated nonsense mutation readthrough. Intestinal organoids were incubated with indicated doses of G418, or vehicle for 72 hours prior to swelling with 10 µM forskolin. (A) Representative images at T=0 and T=300 minutes of organoid swelling, conditions as indicated. (Scale bar = 100 µm) (B) Total change in area was measured over 300 minutes, and FIS was quantified by normalizing the total organoid area to T=0. (n=5 wells of organoids for each treatment, ± SE). (C) AUC at T=300 for indicated treatment groups. (*P<0.05 compared to untreated G542X). (D) Expression of Cftr in WT and G542X organoids with and without G418 treatment as a percentage of untreated WT organoids. (E) Expression of Cftr in lung from untreated G542X mice or treated with G418 compared to expression from WT mice. (*P<0.05 compared to untreated WT. ^P<0.05 compared to untreated G542X.)

105

Figure 3.6. Intestinal organoids from G542X mice are used to test PTC124 mediated readthrough. G524X organoids were treated with the indicated doses of PTC124 for 72 hours prior to the FIS assay. (A) Representative images at T=0 and T=300 minutes of organoid swelling, conditions as indicated. (Scale bar = 100 µm). (B) Normalized areas for PTC124 treated organoids are shown, compared to vehicle and 100 µM G418-treated organoids. (C) AUC measurements at 300 minutes. (*P<0.05 compared to untreated G542X.)

106

Chapter 4 – Synergy between Readthrough and Nonsense-Mediated Decay

Inhibition in a Murine Model of Cystic Fibrosis Nonsense Mutations

The research detailed in this chapter has been submitted for publication to Molecular Therapy in January 2019.

107

4.1 Abstract Many heritable genetic disorders arise from nonsense mutations, which generate premature termination codons (PTCs) in transcribed mRNA [112]. PTCs ablate protein synthesis by prematurely terminating translation of mutant mRNA, as well as reduce mutant mRNA quantity through targeted degradation by nonsense-mediated decay (NMD) mechanisms [251]. Therapeutic strategies for nonsense mutations include facilitating ribosomal readthrough of the PTC and/or inhibiting NMD to restore protein function [131,

243, 252, 253]. However, the efficacy of combining readthrough agents and NMD inhibitors has not been thoroughly explored. In this study, we examined combinations of known NMD inhibitors and readthrough agents using functional analysis of the CFTR protein in primary cells from a mouse model carrying a G542X nonsense mutation in Cftr.

We observed synergy between an inhibitor of the NMD component SMG-1 (SMG1i) and the readthrough agents G418, gentamicin, and paromomycin, but did not observe synergy with readthrough caused by amikacin, tobramycin, PTC124, escin, or amlexanox. These results indicate that treatment with NMD inhibitors can increase the quantity of functional protein following readthrough, and that combining NMD inhibitors and readthrough agents represents a potential therapeutic option for treating nonsense mutations.

4.2 Introduction Ten percent of heritable diseases are caused by nonsense mutations [112, 254].

Nonsense mutations are single nucleotide alterations which generate premature UAG,

UAA, or UGA termination codons (PTCs) in mRNA transcripts. During translation, PTCs are decoded by eukaryotic release factor 1 (eRF1), which has a higher affinity for the PTC than the canonical tRNA. Therefore, eRF1 is preferentially incorporated into the A site of the ribosome, causing hydrolysis of the ester bond linking the elongating polypeptide chain

108 to the last tRNA [255]. This results in premature termination of translation and generates truncated proteins which are often nonfunctional or have deleterious dominant-negative properties [115, 256]. Expenditure of energy and resources on protein which is quickly degraded is a wasteful procedure for the cell; therefore, mutant transcripts are degraded prior to translation by nonsense-mediated decay (NMD) mechanisms. Failure to complete the first round of translation signals the recruitment of NMD machinery which degrades the mutant transcripts, preventing future attempts at translation [251, 257, 258].

Aminoglycoside antibiotics have been found to facilitate readthrough of PTCs by interacting with the decoding center of the ribosome, forcing a conformational shift which reduces the energy required to create a stable mRNA to tRNA bond [20, 112, 234, 259,

260]. This allows a near-cognate aminoacyl tRNA to be incorporated in the ribosomal A site, continuing translation. Aminoglycoside treatment at doses required for readthrough is associated with nephrotoxicity and ototoxity in patients [128]; therefore several groups have identified alternative readthrough agents such as PTC124 [131] and the herbal compound escin [245] through high throughput screening. However, these compounds have limited readthrough properties compared to more efficacious aminoglycosides [253,

261, 262]. Furthermore, ribosomal readthrough is rare, with the most potent readthrough agent, G418, achieving approximately 5% efficiency [234, 235, 263]. The rarity of readthrough is due in part to degradation of mutant mRNA by NMD, which reduces the quantity of mRNA that is subject to readthrough. Increasing the quantity of mutant mRNA using inhibitors of NMD is an attractive therapeutic option to make readthrough more effective. To this end, several compounds which inhibit NMD have been identified, including amlexanox [243], NMDI-14 [264], and a small molecule inhibitor of the NMD

109 component SMG-1 (SMG1i) [252]. However, a thorough examination of synergy between currently known readthrough agents and NMD inhibitors has not been performed.

Cystic Fibrosis (CF) is a heritable disease which can be caused by nonsense mutations. CF arises from mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, which expresses an apically localized anion channel in epithelial tissues. Loss of CFTR function causes dehydration of epithelial surfaces, leading to accumulation of viscous mucus. Patients with CF suffer from a wide range of symptoms, including pulmonary dysfunction [203], intestinal maladies [51], and reduced growth

[215]. CF is a fatal disorder, with the most common cause of mortality being lung failure

[265]. Small molecule therapies which restore function to CFTR have been identified for several CFTR mutation types, including VX-770 [26, 95] for specific gating mutations (e.g.

G551D) and VX-809 [220] and VX-661 [266] for specific misfolding mutations (e.g.

F508del). However, an effective therapy which restores function to CFTR containing nonsense mutations is unavailable.

Recently, we generated a mouse model of the G542X nonsense mutation in CFTR

[166], the most common CF-causing nonsense mutation [20, 240]. CFTR function can be examined in intestinal organoids harvested from G542X mice by measuring CFTR- dependent forskolin-induced swelling (FIS) [168, 267]. We have previously demonstrated that G542X organoids are able to undergo FIS following treatment with G418. However, a number of additional aminoglycosides are available which have reduced readthrough potency and toxicity relative to G418 [20, 224, 261, 268-270]. Here, we examine G418, several alternative aminoglycosides, and identified non-aminoglycoside readthrough agents in combination with known NMD inhibitors using functional analyses of CFTR in

110 primary cells from G542X mice. Our results demonstrate that SMG1i treatment can increase the amount of functional CFTR produced in G542X intestinal organoids following readthrough by the aminoglycosides G418, gentamicin and paromomycin, but not with tobramycin, amikacin or the readthrough agents PTC124 or escin. In addition, we observed that inhibition of SMG-1 synergizes with aminoglycoside readthrough to improve CFTR function in primary airway epithelial cell cultures, suggesting that NMD inhibition and readthrough synergy can occur in both intestinal and airway cells.

4.3 Results 4.3.1 Intestinal organoid FIS is a more biologically relevant detector of readthrough than cell-based reporters

Assays which measure ribosomal readthrough are commonly performed using transgenic cell-based reporters which are highly sensitive to readthrough. Such reporters are highly sensitive to readthrough and have been utilized in high throughput screens to identify novel molecules [131, 245, 271]. However, only one of these molecules, PTC124, has undergone clinical trials for CF, and has not been found to have clinical benefit [228].

This suggests that reporter systems may not be optimal for detecting physiologically beneficial readthrough. We hypothesized that G542X intestinal organoids will more accurately recapitulate the effects of readthrough on G542X-Cftr in a physiologically relevant context. Therefore, we compared detection of readthrough with G542X intestinal organoid FIS to a cell-based readthrough reporter. To establish a readthrough reporter assay, we transfected mouse 3T3 fibroblast cells with pFluc190UGA [272] a plasmid which expresses firefly luciferase truncated by a UGA stop codon. Readthrough of the

UGA stop codon allows detection of full-length luciferase. Transfected cells were treated with doses of G418 ranging from 0 to 1mM for 24 hours. We observed increases in

111 luminescence ranging from 8.6 to 20.5 times greater than untreated cells for the G418- treated cells (Figure 4.1A), indicating dose-dependent readthrough of the UGA PTC. By comparison, G542X intestinal organoids were dosed with 50, 100, or 200µM G418 for 24 hours. CFTR function was then assessed by stimulating the organoids with 10µM forskolin, and images of organoid FIS were captured by brightfield kinetic imaging (Figure 4.1B).

Area under the curve (AUC) for each treatment was calculated to compare statistical significance between treatment groups. We observed robust FIS at 100 and 200µM G418, but did not detect FIS at lower G418 doses (Figure 4.1 C and D). An additional method to examine organoid swelling is by measurement of changes in organoid lumen size [199].

However, measurement of organoid lumen expansion did not increase the sensitivity of the

FIS assay (Supplementary Figure 1 A and B). Thus, G542X intestinal organoids required greater doses of G418 than pFluc190UGA to produce a detectable signal. These results suggest that G542X intestinal organoids may be less sensitive detectors of readthrough, but more accurately model the physiological outcomes of readthrough in the context of CF nonsense mutations than cell-based reporter systems.

4.3.2 Inhibition of SMG-1 synergizes with readthrough to restore CFTR function

The G542X PTC is expressed in native Cftr, which makes G542X-Cftr mRNA sensitive to degradation by NMD. We sought to determine whether FIS mediated by readthrough of the G542X PTC could be increased by pharmacological inhibition of NMD.

Intestinal organoids were incubated for 24 hours with three small molecule inhibitors of

NMD, SMG1i [252], NMDI-14 [264], and amlexanox [243] in combination with 100µM

G418. We observed robust synergy between G418 and SMG1i in our intestinal organoid system. FIS was significantly increased beyond G418 alone when several doses of SMG1i

112 were combined with 100µM G418 (Figure 4.2A and B). SMG1i did not allow FIS to occur in the absence of G418 at any tested dose, indicating that NMD inhibition alone is insufficient to restore CFTR function. Additionally, intestinal organoids from non- correctable Cftr-null mutant S489X mice [160, 166] did not undergo FIS following treatment with G418 and SMG1i, indicating that improvements in FIS are due to inhibition of NMD (Supplementary Figure 2). By contrast, incubation with NMDI-14 did not provide any increase in FIS beyond 100µM G418 alone at any tested doses (Figure 4.2B). We also did not observe synergy between amlexanox and G418 in G542X intestinal organoids at several doses (Figure 4.2B). Of note, the highest tested doses of amlexanox and NMDI-14 appeared to reduce swelling. This reduction in swelling appears independent of cellular toxicity, as intestinal organoids treated with highest tested doses of amlexanox and NMDI-

14 did not have significantly greater cell death as indicated by propidium iodide staining than DMSO-treated organoids (Supplementary Figure 3).

To confirm that SMG1i improves FIS by increasing the quantity of G542X-Cftr mRNA, we performed RTqPCR on intestinal organoids treated with SMG1i and G418. We observed a 5.7±0.36 fold increase over DMSO treated organoids with 1µM SMG1i treatment, and an 8.084±0.2 increase when SMG1i was combined with 100µM G418, indicating an increase in mRNA quantity (Figure 4.2C). NMDI-14 and amlexanox did not increase G542X-Cftr quantity, consistent with the results of our functional examination.

Collectively, these results indicate that NMD inhibition by NMDI-14 and amlexanox are insufficient to improve CFTR function following readthrough, but SMG1i robustly inhibits

NMD leading to readthrough and improved CFTR function.

113

4.3.3 Effectiveness of G418 can be improved by NMD inhibition

We next sought to determine if ineffective levels of readthrough could be improved with NMD inhibition. To this end, we treated G542X intestinal organoids with a range of

G418 doses from 0 to 100µM G418, and supplemented each dose with 1µM SMG1i.

Treatment with SMG1i was able to improve FIS when combined with several doses of

G418 (Figure 4.3A and B). Notably, at 12.5, 25, and 50µM G418 alone, no FIS was observed, but addition of 1µM SMG1i was sufficient to significantly increase FIS. A combination of SMG1i and higher concentrations of G418 led to increased propidium iodide staining, indicating increased cell death (Supplementary Figure 4A). At high concentrations of G418 and SMG1i (200µM G418 and 1µM SMG1i), cellular toxicity was sufficient to cause a reduced FIS response (Supplementary Figure 4B). These results indicate that NMD inhibition by SMG1i increases the quantity of G542X-Cftr mRNA, leading to an increased quantity of functional CFTR following G418-mediated readthrough. However, there are toxicity concerns at higher concentrations.

4.3.4 SMG1i improves readthrough for alternative aminoglycosides to G418

We hypothesized that aminoglycosides which cause less readthrough than G418 could be more effective when paired with NMD inhibition. To examine readthrough caused by aminoglycoside alternatives to G418, we treated G542X intestinal organoids with gentamicin, amikacin, paromomycin, and tobramycin at three dilutions (5mM, 2.5mM, and

1.25mM), with and without 1µM SMG1i. We observed FIS only in organoids treated with gentamicin and in organoids treated with a combination of paromomycin and SMG1i, while tobramycin and amikacin did not allow FIS to occur (Figure 4.4A and B). Similar to

G418, measurement of changes in organoid lumen size was not more sensitive to swelling

114 than automated identification of total organoid area (Supplementary Figure 5). Each aminoglycoside tested increased Cftr expression, suggesting readthrough was occurring in response to each aminoglycoside (Figure 4.4C). Increases in Cftr expression were greatest with gentamicin and paromomycin treatment, which is consistent with the observed increases in CFTR function. Furthermore, we examined readthrough caused by gentamicin, paromomycin, tobramycin, and amikacin using the readthrough reporter pFluc190UGA.

We observed significant increases in luciferase expression for doses ranging from 0 to

5mM of each aminoglycoside except for tobramycin, which did not significantly increase readthrough (Supplementary Figure 6A-D). These results indicate that combination treatment of paromomycin and gentamicin with SMG1i can produce small amounts of functional readthrough, with gentamicin producing the greatest readthrough of tested aminoglycosides.

4.3.5 Readthrough and NMD inhibition combine to improve CFTR function in primary trachea cells

To examine synergy of SMG1i with readthrough agents in a non-intestinal tissue type, G542X tracheal and nasal epithelial cells isolated from G542X mice were grown as air-liquid interface cultures and treated with 50µM G418 for 48-72 hours, with or without

1µM SMG1i. CFTR-dependent short-circuit current (Isc) was increased by G418 and

SMG1i treatment; whereas neither compound alone increased CFTR function (Figure 4.5A and B). These experiments indicate synergy between readthrough agents and NMD inhibition in airway tissue.

115

4.3.6 Non-aminoglycoside compounds do not cause sufficient readthrough to restore

CFTR function

Next, we sought to examine synergy of non-aminoglycoside readthrough agents with SMG1i. We examined FIS in G542X intestinal organoids following a 24 hour incubation with PTC124, escin, and amlexanox in conjunction with SMG1i treatment. The dose of each drug was similar to doses used in other publications [131, 150, 243, 245]. No tested dose of any compound allowed FIS to occur, with or without combination treatment with SMG1i (Figure 4.6A). The highest tested doses of PTC124 and escin did not increase

Cftr expression, suggesting minimal readthrough (Figure 4.6B). To test for possible synergy between non-aminoglycoside and aminoglycoside readthrough agents, we examined PTC124, escin or amlexanox in combination with 100µM G418, but found no evidence of synergy as the addition of the non-aminoglycoside did not increase FIS response to G418 (Figures 4.2A and 4.6C). Finally, we utilized the readthrough reporter system pFluc190UGA with each non-aminoglycoside readthrough agent. For amlexanox and escin, we observed a significant increase in luminescence but very small compared to aminoglycosides (Supplementary Figure 7 A and B). We did not use PTC124 in this system as PTC124 has been reported to stabilize firefly luciferase in this system, providing a false positive [272]. Collectively, these results indicate that non-aminoglycoside readthrough agents do not cause sufficient readthrough to restore functional levels of CFTR, even when supplemented with NMD inhibition.

116

4.3.7 G542X-CFTR trafficking can be improved by a CFTR corrector following readthrough with G418 and gentamicin

Readthrough frequently leads to the insertion of a non-optimal amino acid in place of the PTC [273]. This non-optimal amino acid can cause dysfunctional CFTR trafficking

[273]. Several small molecules; known as correctors, are available for correcting abnormal

CFTR trafficking, which is common to mutant CFTR [220, 266]. To examine modification of protein trafficking following readthrough, we incubated intestinal organoids with the corrector VX-661 along with G418, gentamicin, or paromomycin with SMG1i. We found that VX-661 allowed greater FIS than readthrough agents and SMG1i alone (Figure 4.7A-

C), indicating correction of aberrant CFTR trafficking and increased function of CFTR at the apical membrane. We did not observe FIS in G542X intestinal organoids treated with

VX-661, confirming that VX-661 does not allow FIS to occur independent of CFTR function (Supplementary Figure 8). These data indicate that the trafficking of murine

G542X-CFTR can be improved by administration of correctors after readthrough and

NMD inhibition.

4.4 Discussion A number of heritable diseases including CF, Duchenne Muscular Dystrophy

(DMD) [259, 261] and Spinal Muscular Atrophy (SMA) [235] can arise due to nonsense mutations. Additionally, nonsense mutations in a number of genes, including CDH1,

PTEN, BRCA1 and BRCA2, have been linked to a predisposition for cancer development

[114]. Due to the prevalence of diseases caused by nonsense mutations, a therapy which can restore function to genes ablated by nonsense mutations is highly desirable. Although a safe and highly effective readthrough agent has yet to be identified, currently existing readthrough agents may be more effective if administered alongside inhibitors of NMD.

117

Here, we identify several currently available readthrough agents which are more effective in combination with a pharmacological inhibitor of SMG-1, a critical component of NMD, using functional assays of murine G542X-CFTR.

We examined three NMD-inhibiting compounds in this study, NMDI-14 [264], amlexanox, and SMG1i [252]. We did not find NMDI-14 or amlexanox to be effective for improving FIS mediated by G418-facilitated readthrough, indicating that both agents were ineffective for inhibiting NMD in intestinal organoids. However, we found synergy between SMG1i and several readthrough agents to improve G542X-CFTR function. SMG-

1 is a serine/threonine kinase which phosphorylates the NMD component UPF1, allowing the recruitment of mRNA degradation machinery [251]. SMG1i is a pyrimidine derivative which blocks the kinase activity of SMG-1 to inhibit NMD, suggesting that inhibiting

NMD at SMG-1 is stronger than at other locations. Although we found SMG1i to be an effective NMD inhibitor, we also found SMG1i to have a degree of toxicity, particularly when combined with higher doses of aminoglycosides. A degree of toxicity by SMG1i is unsurprising, as SMG-1 is active in a number of cellular processes outside of NMD [274,

275], and interfering with SMG-1 activity likely has deleterious effects on cell health. This raises the possibility that toxicity from SMG1i by itself or in combination with readthrough therapy would preclude SMG1i from use in a clinical setting. A thorough examination on in vivo SMG1i toxicity is necessary. Regardless of clinical efficacy, SMG1i serves as a proof of concept that a compound can inhibit NMD to a degree which is sufficient to improve G542X-CFTR function after readthrough. Considering that toxicity of aminoglycosides at the doses required for readthrough precludes their use in a clinical

118 setting, supplementation with NMD inhibitors may represent a strategy for increasing aminoglycoside-facilitated readthrough with non-toxic aminoglycoside doses.

In our intestinal organoid experiments, we observed varying readthrough efficacy by several different aminoglycosides. Consistent with previous research, G418 provided the strongest readthrough effect [276] and synergized effectively with SMG1i. We also observed this effect in air-liquid interface cultures of primary airway cells from G542X mice. These observations are consistent with Valley and colleagues, who observed synergy between G418 and SMG-1 inhibition in a CRISPR-edited human bronchial epithelial

(HBE) cell line [277]. Gentamicin, though not as effective as G418, did allow a small degree of FIS. When combined with SMG1i, gentamicin facilitated a robust FIS response.

We consider this significant, as gentamicin is the only aminoglycoside which has been utilized in clinical trials to treat nonsense mutations. However, gentamicin was dropped from clinical trials due to causing ototoxicity and nephrotoxicity. The possibility of combining gentamicin with an NMD inhibitor in a clinical setting may improve the efficacy of gentamicin to a clinically relevant level. Additionally, we observed a small degree of

FIS with paromomycin treatment when combined with SMG1i. However, the benefit a patient with CF would receive from an increase in CFTR function is not likely to offset the toxic effects of paromomycin. No FIS was observable in organoids treated with amikacin or tobramycin, even when supplemented with SMG1i. An increase in Cftr expression was detected following treatment with all aminoglycosides. G418 caused the greatest increase in expression followed by gentamicin and then by paromomycin, which was consistent with the efficacy of each aminoglycoside for restoring CFTR function. Small increases in

Cftr expression were observed following tobramycin and amikacin treatment, which

119 indicates that some degree of readthrough was occurring. However, this level of readthrough was insufficient to generate a detectable level of CFTR function.

Although a full-length protein is produced following readthrough, an amino acid may be inserted which is different from that of the wild type protein at the PTC [273]. This alteration in amino acid sequence can have deleterious effects on protein processing and function. Therefore, in the context of CF, and likely other diseases, additional interventions may be necessary to improve protein function. As a proof of concept, we added the CFTR corrector VX-661 to previously examined doses of readthrough agents and SMG1i, and observed increased CFTR function following treatment with VX-661. These results are consistent with what was observed by Xue and colleagues, who observed that the corrector

VX-809 [220] was sufficient to improve folding of human G542X-CFTR following readthrough [273]. Further interventions are being developed which may further improve

CFTR function following readthrough. Recently, a compound named PTI-428 has been identified which selectively increases CFTR expression has been identified by high- throughput screening [110]. PTI-428 represents a novel class of CFTR modulators termed

“amplifiers”, which increase CFTR gene transcription. Increasing CFTR expression in the context of a CFTR nonsense mutation may further improve the quantity of mutant mRNA which can be acted on by readthrough agents.

The non-aminoglycoside readthrough agents escin and PTC124 did not cause enough readthrough to restore CFTR function or to increase Cftr expression. There is abundant published research which suggests that each molecule is able to cause ribosomal readthrough of PTCs [225, 240, 243, 245]. However, many of these experiments were performed using cell-based readthrough reporters which rely on overexpressed transgenic

120 constructs to detect readthrough. As we have shown here, transgenic systems can be sensitive to low levels of readthrough which may be insufficient to translate to physiological functionality. To our knowledge, this is the first examination of the readthrough properties of escin in a system with endogenous expression of mutant Cftr.

However, PTC124 has been more thoroughly examined. In vivo validation experiments of

PTC124 were performed using mouse models of CF which overexpress Cftr carrying a nonsense mutation [20]. Despite success in certain models, PTC124 was ineffective for human nonsense-carrying intestinal organoids [150], and has not improved clinical outcomes in patients with nonsense mutations in CFTR [228, 278]. PTC124 has been sufficient to restore function to the dystrophin gene in a mouse model of Duchenne muscular dystrophy [131]. Therefore, despite ineffectiveness in restoring G542X-CFTR,

PTC124 may be sufficient to restore function to other genes which carry PTCs.

A variety of methods have been utilized to quantify intestinal organoid swelling.

Here, we utilized two methods to quantify swelling of organoids, an automated measurement of total organoid area and manual measurement of the organoid lumen. Total area measurement is the preferred method of quantification among the majority of papers; however lumen measurements are used exclusively by some groups [199]. Lumen measurements were made to quantify small levels of CFTR function which caused an expansion of the organoid lumen, but did not contribute to an overall increase in organoid area. This effect was observable in organoids treated with several doses of gentamicin and paromomycin. However, measuring the organoids in this manner had an increased degree of variability relative to total organoid area. Additionally, manually recording lumen measurements is time consuming, making it inconvenient to measure multiple time points

121 during organoid swelling. Though measuring lumen expansion is not an ideal method of quantifying FIS, it is effective for detecting heterogeneous responses among organoid populations, as well as identifying subtle levels of CFTR function.

In summary, primary cells from a CF nonsense mutation mouse model can effectively assess readthrough, synergy with NMD inhibition, and the resulting CFTR function. Considerable effort is currently being put forth to identify novel readthrough agents which can be both safe and highly effective in a clinical setting. However, even the most effective readthrough agent may still not restore sufficient gene function to reverse disease manifestations caused by nonsense mutations. Therefore, co-treatment with NMD inhibition may be the only viable option for sufficiently restoring gene function. The CF

G542X intestinal organoid model represents a robust model for analyzing the effectiveness of newly identified compounds prior to in vivo treatment studies. Here, we demonstrate that treatments for nonsense mutations which combine NMD inhibition with readthrough, and then further interventions to modify protein function may be a more effective therapeutic option than readthrough alone.

4.5 Materials and Methods 4.5.1 Mice

The creation of the G542X mouse model was previously described [166]. Mice homozygous for these mutations were created by breeding heterozygous males and females. Genotyping was completed by PCR analysis using DNA extracts from ear biopsies. To detect the G542X allele (319 bp) primers P1 (5’-

ACAAGACAACACAGTTCTCT -3’) and P2 (5’ TCCATGCACCATAACAACAAGT -

3’) were used. To detect the wildtype (WT) allele (319 bp) P2 and P3 (5’-

122

ACAAGACAACACAGTTCTTG -3’) were used in a separate reaction. PCR reactions were completed for 40 cycles of 95°C for 30 seconds, 58°C for 30 seconds and 72°C for

30 seconds and products were run out on 2% agarose gels. All mice were allowed unrestricted access to water and solid chow (Harlan Teklad 7960; Harlan Teklad Global

Diets). All animals were maintained on a 12-h light, 12-h dark schedule at a mean ambient temperature of 22°C and were housed in standard polysulfone microisolator cages in ventilated units with corncob bedding. The Institutional Animal Care and Use Committee of Case Western Reserve University approved all animal protocols.

4.5.2 Compounds

Stocks of aminoglycoside antibiotics (All purchased from Sigma-Aldrich) were dissolved in sterile H2O at a 100mg/ml concentration, and stored at 4C. PTC124 (3-[5-

(2-Fluorophenyl)-1,2,4-oxadiazol-3-yl]benzoic acid, MedChemExpress, Cat. #HY-

14832), escin (Sigma-Aldrich, Cat. #E1378), VX-661 (MedChemExpress, Cat. #HY-

15448), NMDI-14 (Ethyl 2-(((6,7-dimethyl-3-oxo-1,2,3,4-tetrahydro-2- quinoxalinyl)acetyl)amino)-4,5-dimethyl-3-thiophenecarboxylate, EMD Millipore, Cat.

#530838,), and amlexanox (MedChemExpress Cat. #HY-B0713) were dissolved in dimethylsulphoxide (DMSO). SMG1i (2-chloro-N,N-diethyl-5-((4-(2-(4-(3- methylureido)phenyl)pyridin-4-yl)pyrimidin-2-yl)amino)benzenesulfonamide) was a kind gift from Dr. Robert Bridges from Rosalind Franklin University, and was dissolved in

DMSO. Working solutions were all 0.1% DMSO, with the exception of NMDI-14, which was 1% DMSO. The increased DMSO did not prevent G418-facilitated FIS

(Supplementary Figure 9). Stocks of forskolin (Sigma-Aldrich, Cat. #F6886) were dissolved in 100% EtOH and stored at -20C.

123

4.5.3 Crypt harvest and intestinal organoid culture

Intestinal organoids were cultured similar to previously described methods [166].

Mice were sacrificed by CO2 asphyxiation, and 20cm of intestine measured from the stomach were removed. Fecal matter was flushed from the intestine with Ca2++ and Mg2++

-free PBS, and the intestine was flayed using dissecting scissors. The villi were scraped from the small intestine using a microscope slide, and the intestine was cut into ~1cm segments, which were suspended in 2mM EDTA in Ca2++ and Mg2++ -free PBS. The intestinal segments were incubated on a shaker for 30 minutes at room temperature. The segments were then vortexed at for 10 seconds, allowed to settle, and then the supernatant was removed and stored in a 10cm dish. This process was repeated until four supernatant fractions were produced. The fractions were inspected under a microscope, and the fraction which was most enriched for crypts was passed through a 70µm cell strainer. The crypts were pelleted at 1,000xG for 10 minutes, then resuspended in 1:1 mixture of Intesticult

Organoid Growth Media (OGM; STEMCELL Technologies) and MatriGel (Corning) at a concentration of 10 crypts/µl. The organoids were seeded to 12-well plates, with 70µl

Matrigel:OGM added to each well in 4-5 droplets. The plate was placed in a 37C/5% CO2 incubator for 15 minutes to allow the MatriGel to harden. The MatriGel domes were then

immersed in 1mL OGM and returned to the 37C/5% CO2 incubator. OGM was changed every 3-4 days, and the organoids were passaged once every 5-7 days.

4.5.4 Forskolin-induced swelling assay

Organoid swelling experiments were carried out similar to previously described methods [166-168], with small modifications. Organoids were collected into a 15ml conical tube using Ca2++ and Mg2++ free PBS and pelleted at 500xG. The organoids were

124 resuspended in 1mL PBS and dissociated by vigorously resuspending the organoids 10-15 times with a P1000 pipette tip covered with a P10 pipette tip. The dissociated organoids were pelleted and resuspended in 1:1 OGM:MatriGel. The organoids were diluted to a concentration of 30-50 organoids per 5µL MatriGel droplet, and plated to 48-well plates with 5µL of MatriGel per well. The plate was then incubated for 15 minutes in a 37C/5%

CO2 incubator to harden the MatriGel. Immediately following hardening of MatriGel, indicated dilutions of readthrough agents and NMD inhibitors were added along with 200µl

OGM. To assess NMD inhibition, we utilized 1µM SMG1i for the majority of experiments.

5µM SMG1i had allowed the largest FIS response of any tested doses when combined with

G418, however, we selected 1µM SMG1i as the optimal dose due to concerns about toxicity. Twenty four hours following addition of readthrough agents, 200µl OGM containing 20µM forskolin was added to each well, creating a 10µM final concentration.

Kinetic brightfield images of FIS were acquired under live cell conditions with a Lionheart

FX Automated Microscope (Biotek Instruments, Winooski, Vermont). FIS was quantified using Gen5 Prime software by normalizing changes in organoid sum area to the sum area in the initial image. Area under the curve (AUC) was calculated using GraphPad Prism.

Expansion of the organoid lumen was measured by marking the organoid lumen at 0 minutes and 180 minutes using the polygon selection tool in ImageJ. Lumen expansion was quantified as the percent change in lumen area between the 0 minutes and 180 minutes.

Ten organoids were measured per treatment group.

4.5.5 Toxicity Assessment

Intestinal organoids were passaged to 96 well plates and treated with indicated compounds for 24 hours. The intestinal organoids were then stained with 10µg/ml Hoechst

125 and 1µg/ml propidium iodide for 30 minutes. Images were recorded using RFP and DAPI channels, and total RFP and DAPI intensity per well were calculated using Gen5 Prime.

RFP intensity was divided by DAPI intensity to normalize propidium iodide signal to the quantity of organoids.

4.5.6 Murine primary airway cell culture

Trachea or nasal epithelium was isolated from mice, cleaned, digested with pronase/DNase and followed by a brief treatment with accutase. Cells are co-cultured with irradiated fibroblasts and allowed to proliferate in a media containing a Rho-kinase inhibitor [279, 280]. After proliferation, cells were seeded into 12-well plate inserts at a density of ~ 2 x 105 cells per filter. Cells were switched to a media that promotes differentiation and grown in an air-liquid interface (ALI). After 3 weeks at ALI, the establishment of electrically tight cultures were determined by Ussing chamber measurements at which time assessment of CFTR function by short-circuit current measurement was completed.

4.5.7 Assessment of CFTR function by short-circuit current measurement

The epithelial monolayers were bathed with symmetrical Krebs bicarbonate ringers solution, and maintained under short-circuit conditions. CFTR function was assessed as previously described [281]. Briefly, after stabilization of baseline, following inhibitors and

+ activators of Isc were sequentially added: sodium (Na )-channel blocker Amiloride (100

μM) to inhibit apical epithelial Na+ channel (ENaC); cAMP agonists Forskolin (10 μM) and 3-isobutyl-1-methylxanthine (IBMX 100 μM) to activate the transepithelial cAMP- dependent current (including Cl− transport through CFTR channels); genistein (30 μM) to potentiate CFTR; and CFTR inhibitor CFTRinh172 (10 μM) to specifically inhibit CFTR.

126

Data were acquired with the software Acquire and Analyze version 2.3.159 (Physiologic

Instruments). CFTR specific function in the cells was calculated as change in Isc (ΔIsc) defined as the difference between the sustained phase of the current response before and after stimulation with forskolin.

4.5.8 Expression Analysis

To examine Cftr expression, G542X intestinal organoids were passaged into 24- well plates, with 35µl of MatriGel per well. One well was considered to be a single experiment. The organoids were grown for five days to the point of being significantly budded, and were then treated for 24 hours with indicated compounds diluted in 500µl

OGM. The organoids were then lysed using a QiaShredder Cell and Tissue Homogenizer

Kit (Qiagen, Cat. #79654). RNA was harvested using an RNeasy Mini Kit (Qiagen, Cat.

#74104). 250ng of RNA was reverse transcribed into cDNA using a QScript cDNA synthesis kit (VWR, Cat. #101414-098). Cftr expression was examined using a TaqMan expression assay which used primers recognizing exons 17 and 18 on Cftr (Mm00445197;

Applied Biosystems). RTqPCR was performed on a StepOne PCR system (Applied

Biosystems) to examine Cftr expression, with ß-actin serving as the endogenous control.

The average of each value was expressed as the fold change difference in Cftr expression from DMSO-treated organoids.

4.5.9 Readthrough Reporter

Mouse 3T3 fibroblast cells (a kind gift from Mitchell Drumm) were grown in

DMEM with 10% fetal calf serum and 1% penicillin/streptomycin. The cells were split to

96-well plates, with 10,000 cells per well. The following day, the cells were transfected with 100ng per well of pFluc190UGA (a kind gift from James Inglese; Addgene plasmid

127

# 41046) [272] using Lipofectamine 2000 (ThermoFisher Scientific, Cat. # 11668030). 48 hours following transfection, the cells were treated with indicated doses of readthrough agents for 24 hours. Readthrough was assessed by quantifying luminescence of firefly luciferase using a Luciferase Assay System (Promega Cat. # E4550). Luminescence was recorded using a GloMax Navigator Microplate Luminometer (Promega, Cat. # GM2000).

4.5.10 Statistical Analysis

All indicated statistical tests were calculated using GraphPad Prism. All error bars are displayed as ±SEM.

Acknowledgments

We thank Alma Wilson, Brittany Burns, Amanda Barabas, and Molly Schneider of the

Cystic Fibrosis Mouse Models Core at Case Western Reserve University for assistance in maintaining the mouse colony. We thank Drs. Jenny Kerschner and Jey Sabith Ebron of

Ann Harris’ Lab at Case Western Reserve University for their assistance in measuring luminescence. We thank Jean Eastman of Mitchell Drumm’s lab for assistance in culturing mouse 3T3 cells.

This work was supported by grants from the Cystic Fibrosis Foundation

(HODGES15XX0 and HODGES18I0 to CAH).

128

129

Figure 4.1. Readthrough with G418 in a cell based reporter system and G542X intestinal organoids A. Firefly luciferase expression in 3T3 cells transfected with pFluc190UGA and treated with the indicated dose of G418. N=3 wells per dose. P*<0.0001 by one way ANOVA with post hoc Tukey test. B. Representative brightfield images of G542X intestinal organoids with indicated treatments at 0 and 180 minutes following stimulation with 10µM forskolin. Scale bar is 100µm. C. FIS curves of intestinal organoids images every 15 minutes following forskolin stimulation. N=3 wells per treatment group. D. AUC measurements recorded from 1C. P****<0.0001, P^<0.0001 vs DMSO by one way ANOVA with post hoc Tukey test.

130

131

Figure 4.2. SMG1i, but not NMDI-14 or amlexanox, synergizes with G418 to improve CFTR function. A. Normalized FIS curves of G542X intestinal organoids treated with indicated doses of SMG1i with or without 100µM G418. N=3. B. AUC values for the indicated doses of SMG1i, NMDI-14, and amlexanox with or without 100µM G418. P*<0.001 vs DMSO: DMSO by ANOVA with a post-hoc Tukey test. C. Cftr expression in organoids treated for 24 hours with the indicated doses of NMDI-14, amlexanox, G418, and SMG1i. Displayed as fold change vs DMSO. P**<0.01, P****<0.0001 by 1 way ANOVA with post-hoc Tukey test. N=3 experiments per group.

132

Figure 4.3. Low levels of readthrough are improved by inhibiting NMD. A. Representative brightfield images of G542X intestinal organoids at 180 minutes following stimulation with 10µM forskolin. Scale bar is 100µm.B. AUC measurements from swelling curves of G542X intestinal organoids treated with the indicated doses of G418 and/or 1µM SMG1i. N=3, P***=0.003, P****<0.0001 vs DMSO by two way ANOVA with post-hoc Tukey test.

133

134

Figure 4.4. Gentamicin and paromomycin restore detectable levels of CFTR function. A. Representative brightfield images of G542X intestinal organoids incubated with 2.5mM of the indicated aminoglycosides 180 minutes after forskolin stimulation. Red arrows indicate expanded organoid lumens. Scale bar is 100µm. B. AUC measurements for four aminoglycosides at indicated doses. Each AUC is representative of 3-6 individual experiments. C. Cftr expression displayed as fold change vs DMSO. N=3 wells per treatment group, P*<0.05; P****<0.0001 vs untreated by one way ANOVA with a Dunnett test.

135

Figure 4.5. SMG1i synergizes with readthrough in trachea tissue. A. Representative Isc traces of G542X trachea monolayers treated with DMSO or a combination of 50µM G418, and 1µM SMG1i for 48 hours prior to treatment with compounds as indicated. Subsequent CFTR stimulation, potentiation and inhibition were achieved through the addition of forskolin, genistein and CFTRinh172 B. Changes in Isc after stimulation with 10µM forskolin for the indicated treatment groups. N=3-11, P*<0.05.

136

137

Figure 4.6. Non-aminoglycoside readthrough agents do not restore CFTR function in G542X intestinal organoids. A. AUC measurements from G542X intestinal organoids treated with the indicated doses of PTC124, amlexanox, and escin with or without SMG1i. N=3. B. Cftr expression after 24 hour incubation with the indicated compounds. C. AUC measurements from intestinal organoids incubated for 24 hours with the indicated compounds. N=3-6 wells, analyzed using two way ANOVA with post-hoc Tukey test. All G418-treated groups are significantly increased over no G418 groups.

138

139

Figure 4.7. Folding of readthrough-facilitated CFTR can be improved with VX-661. A. Representative brightfield images of G542X intestinal organoids treated with 100µM G418, 2.5mM gentamicin, or 2.5mM paromomycin and indicated combinations of 1µM SMG1i and/or 3.3µM VX-661 at 0 minutes and 180 minutes following treatment with 10µM forskolin. Scale bar is 100µm. B. Example FIS curves of G542X intestinal organoids treated with indicated doses. N=3 wells per treatment group. C. AUC measurements from intestinal organoids treated with the indicated doses of compounds as in 6A. P*<0.0001 between indicated treatment groups, P^<0.0001 from DMSO by one way ANOVA with post-hoc Tukey test.

140

Supplementary Figure 1. Detection of FIS by measurement of the intestinal organoid lumen. A. Images of intestinal organoid lumen measurements at 0 and 180 minutes. B. Percent change in lumen area at 180 minutes. N=10 organoids per group. P**=0.0036, P****<0.0001 compared to untreated control by one-way ANOVA with post-hoc Tukey test.

141

Supplementary Figure 2. FIS increases caused by SMG1i do not occur independently of NMD inhibition. S489X intestinal organoids were stimulated with 10µM forskolin following incubation with the indicated compounds. Data were analyzed using a two-way ANOVA with post-hoc Tukey test. N=3.

142

Supplementary Figure 3. Assessment of NMD inhibitor toxicity. Intestinal organoids were treated for 24 hours with the indicated compounds. P*<0.005 vs DMSO by one way ANOVA with a post-hoc Tukey test. N=4-8 wells per treatment group.

143

Supplementary Figure 4. SMG1i toxicity impedes FIS when combined with high doses of G418. A. RFP/DAPI values for intestinal organoids treated for 24 hours with the indicated compounds. N=4 wells per treatment, P****<0.0001 between indicated groups, P^<0.05 vs DMSO by one way ANOVA with post-hoc Tukey test. B. AUC values for intestinal organoids treated with the indicated compounds for 24 hours. P*<0.05 between indicated groups by two-way ANOVA with post-hoc Tukey test. N=3.

144

Supplementary Figure 5. Lumen measurements of alternative aminoglycosides and SMG1i. Percent change in lumen area at 180 minutes for indicated doses of aminoglycoside. Groups treated with 1µM SMG1i are indicated with the + sign. N=10 to 20 per treatment group.

145

Supplementary Figure 6. Readthrough by non-G418 aminoglycosides measured by pFluc190UGA. Firefly luciferase expression in 3T3 fibroblast cells transfected with pFluc190UGA following a 24 hour incubation with the indicated doses of gentamicin (A), paromomycin (B), amikacin (C), or tobramycin (D). N=3 wells per treatment, P*<0.05 vs 0 by one way ANOVA with post-hoc Tukey test.

146

Supplementary Figure 7. Readthrough by non-aminoglycoside readthrough agents measured by pFluc190UGA. Firefly luciferase expression in 3T3 fibroblast cells transfected with pFluc190UGA following a 24 hour incubation with the indicated doses of amlexanox (A) or escin (B). N=3 wells per treatment group, P*<0.05 vs 0 by one way ANOVA with post-hoc Tukey test.

147

Supplementary Figure 8. VX-661 does not facilitate FIS independent of CFTR. AUC measurements from G542X intestinal organoids treated for 24 hours with DMSO or 3.3 µM VX-661. N=3.

148

Supplementary Figure 9. 1% DMSO does not impede intestinal organoid FIS. AUC measurements from G542X intestinal organoids were incubated with the indicated compounds and either 0.1% or 1% DMSO. P*<0.05 between indicated groups by two-way ANOVA with post-hoc Tukey test; N=3.

149

Chapter 5 – Conclusions and Future Directions

150

5.1 Summary

Research into CF has uncovered a wealth of knowledge regarding the mechanics underlying CF pathophysiology. This knowledge has been crucial for developing many therapies which have improved and lengthened the lives of patients with CF. Despite this progress, CF is still a significant source of discomfort which shortens the lives of patients.

The findings presented in this dissertation delve into several potential therapies which could be used to treat CF manifestations in some or all of patients with CF. Chapter 2 describes the investigation of surprising anecdotal claims that linaclotide provides relief from GI symptoms in patients with CF. Our results indicate that linaclotide utilizes an unexpected CFTR-independent mechanism to improve GI transit in the CF intestine.

Chapter 3 details the generation and characterization of a mouse model of the G542X nonsense mutation. The G542X mouse was found to have all expected manifestations of

CF, including reductions in growth, survival, and Cftr expression. We were able to detect restored CFTR function following G418-mediated readthrough of the G542X PTC. We anticipate the G542X mouse to have utility for testing novel CFTR-directed therapies for

CF nonsense mutations. In Chapter 4, we utilized primary tissue from the G542X mouse to examine synergy between readthrough agents and inhibitors of NMD. Robust synergy was observed between SMG1i and multiple readthrough agents in both G542X intestinal organoids and primary trachea monolayers. This result indicates that readthrough agents can be supplemented with NMD inhibitors to improve readthrough efficiency for PTCs in

CFTR. Overall, this dissertation examines novel treatments for CF intestinal manifestations and develops a novel model for examining CF nonsense mutations. The nonsense mutation model has been effective for examining therapies which restore function to CFTR, in

151 addition to NMD inhibitors. The research described in this thesis advances the field of CF closer to the ultimate goal of CF research, which is extend and improve the lives of patients with CF.

5.2 Linaclotide in the CF Intestine Anecdotal reports suggested that linaclotide is effective for relieving GI manifestations in patients with CF. This observation was surprising, as linaclotide was thought to relieve intestinal discomfort by activating CFTR to improve GI transit.

Therefore, linaclotide was predicted to have no effect in patients with CF due to the absence of CFTR function. These anecdotal reports suggested that linaclotide may be working through a CFTR-independent mechanism to improve GI transit in the CF intestine. The research described in Chapter 2 utilized CF mouse models to confirm that linaclotide is effective for improving GI transit in the CF intestine. We determined that linaclotide inhibits NHE3 to improve luminal salt content, increase fluidity, and thereby improve GI transit. Patients with CF still experience GI discomfort despite the availability of a number of interventions which can treat GI symptoms, indicating that additional interventions are necessary. Linaclotide may provide relief from GI discomfort in patients who do not respond well to other interventions. Though linaclotide is likely to be an effective intervention for inhibiting NHE3 in the CF intestine, it is possible that other consequences of linaclotide treatment could be beneficial for patients with CF. Several outcomes of linaclotide treatment on CF patients will be examined in the future.

5.2.1 GC-C Signaling and CFTR

Linaclotide was developed to resemble the heat-stable enterotoxin (STa) which is secreted by Escherichia coli. STa is the primary promoter of diarrhea, dehydration, and

152 death amongst pediatric patients in the developing world [282, 283]. STa is a small 18-19 amino acid peptide which is an agonist to the GC-C receptor in the intestine. Conventional

GC-C agonists include the hormones guanylin and uroguanylin [205, 206, 284]. These hormones regulate intestinal lumen fluidity through GC-C activation, which induces a cGMP second messenger cascade. The consequence of cGMP signaling is ultimately phosphorylation of CFTR, promoting secretion of chloride into the lumen. Additionally, cGMP signaling inhibits NHE3, blocking Na+ absorption from the intestinal lumen. Both actions promote increased fluidity of the intestinal lumen, which is useful for clearance of fecal matter. STa binds GC-C with a greater affinity than guanylin or uroguanylin, likely due to the presence of three disulfide bonds in the peptide chain [285], as opposed to only two in guanylin and uroguanylin [286]. This third disulfide bond confers an affinity for

GC-C which is 10 fold greater than that of uroguanylin and 100 fold greater than that of guanylin [284]. Thus, STa binding to GC-C leads to over-activation of cGMP signaling, causing excessive levels of fluid to enter the small intestine. The consequence of this is diarrhea, dehydration, and potentially mortality. Despite the excessive level of fluid secretion caused by STa, GC-C signaling was recognized as a therapeutic target for treating constipation and other GI disorders. Therefore, linaclotide was developed to target GC-C, while being orally available and non-toxic at appropriate doses.

A functional link between NHE3 and CFTR has been demonstrated by other groups. NHE3 and CFTR have been found to colocalize to one another, and associate in the same protein complex at the apical membrane [287-289]. There is evidence that NHE3 and CFTR influence regulation of one another, as cAMP-mediated inhibition of NHE3 is greater in the presence of CFTR [290]. Inability to properly inhibit NHE3 may lead to

153 hyper-absorption of sodium in the CF intestine, further exacerbating dehydration. This effect has been reported in the CF airway, where loss of CFTR function leads to increased

Na+ absorption by ENaC [291-293]. Other reports have not found Na+ hyper-absorption in the absence of CFTR, so this effect requires further scrutiny [293]. Additionally, reciprocal regulation of CFTR by NHE3 is likely, as Isc by CFTR is reduced in the presence of NHE3

[294]. Furthermore, basal secretion in the duodenum, which is likely to be CFTR- dependent, is increased by inhibition of NHE3 [295]. Based on the evidence of reciprocal regulation of NHE3 by CFTR, we hypothesize that linaclotide inhibits NHE3 to a lesser degree in the context of CF than in non-CF tissue. Differences in NHE3 inhibition by linaclotide between WT and CF could be examined using pHi recovery experiments similar to what was performed in Figure 2.5 using WT and CF Caco-2 cells. Furthermore, ex vivo intestinal sections could be utilized to measure pH recovery [188] in intestinal villi which had been freshly excised from WT or CF mice.

In addition to promoting fluidity of the intestinal lumen, linaclotide was found to increase the percentage of patients with a clinically meaningful improvement in abdominal pain [296, 297]. This analgesic effect was also confirmed in animal models [298]. Though it is likely that some pain relief comes from improvements in GI transit, linaclotide was determined to confer analgesic properties by inhibiting nociceptors in the small intestine

[297]. GC-C activation leads to an increase in cGMP, which is actively transported out of the basolateral side of intestinal enterocytes, where it is able to act on colonic nociceptors to curb pain [297]. Considering that this mechanism of pain relief by linaclotide occurs independently of CFTR, it is likely that this pain-relieving mechanism is still functional in

154 the context of CF. However, research should be performed in the future to confirm that patients with CF would benefit from the analgesic properties of linaclotide.

5.2.3 Effects of Chronic Linaclotide Treatment

We have successfully demonstrated that linaclotide was effective for improving GI transit in the CF intestine. However, it is unknown whether this increase in GI transit is sufficient to reduce the frequency of intestinal obstruction in CF mice. Chronic treatment with other laxatives such as OligoG [82] or PEG [72] in the drinking water of CF mice has reduced obstruction and improved survival of CF mice. Therefore, it is likely that daily treatment of linaclotide would also prevent obstruction in CF mice. In the future, the impact of chronic linaclotide treatment on the survival of CF mice will be examined.

Demonstrating improved survival of CF mice with linaclotide treatment would provide even stronger evidence that linaclotide may be effective for treating GI discomfort in patients with CF.

An increase in GI motility caused by chronic linaclotide treatment may alter other

CF intestinal manifestations besides obstruction. An increase in fluid content caused by laxative treatment has been shown to reduce the mucus content of CF mice [161]. As described in the introduction, accumulated mucus is a fertile bed for infectious bacteria, which can cause inflammation of intestinal tissue. Accumulated bacteria can prompt an innate immune response from intestinal enterocytes, which can be detected by increases in genes such as mast cell protease 2 and the neutrophil marker Lrg [299]. Increases in this innate immune response can cause inflammation of intestinal tissue. Reducing intestinal mucus content with laxatives has been effective for reducing bacterial load and markers of the innate immune response in CF mice [161]. It is likely that GI transit increases caused

155 by linaclotide would produce a similar effect. Therefore, mucus content, bacterial load, and expression of genes associated with innate immunity in CF mice following chronic linaclotide treatment will be examined in the future. Changes in mucus content of the intestine can be measured utilizing stained histological sections from the small intestine

[82], and the quantity of bacteria can be determined by determining the abundance of bacterial 16srRNA [75]. RTqPCR can be performed to determine the expression of genes regulating innate immunity following chronic linaclotide treatment. Examples of innate immune genes which can be tested for changes in expression following chronic linaclotide treatment include Lrg, mast cell protease 2, Muclin, and serum amyloid A3 [299].

Additionally, RNAseq analysis could be performed to determine global changes in gene expression following linaclotide treatment. Confirming these improvements in the intestine of CF mice provides strong evidence that chronic linaclotide treatment would confer similar out comes in patients with CF.

5.2.4 Linaclotide Synergy with CFTR Modulators

Although the research detailed in Chapter 2 focused on a CFTR-independent mechanism for the improvement of GI transit by linaclotide, it is possible that linaclotide also activates mutant CFTR. As demonstrated here and elsewhere, cGMP signaling elicited by linaclotide strongly activates CFTR in the intestine [157, 167]. Considering that several classes of CFTR mutation retain some degree of function, it is possible that patients expressing these mutation types could see additional benefit from linaclotide treatment.

The Hodges Lab is well situated to examine the effect of linaclotide on mutant CFTR, as we have access to mouse models which model mutations from each class of CFTR mutation. The effect of linaclotide on CFTR in each of these mutation types can be tested

156 readily using assays of GI transit and intestinal organoid swelling. Furthermore, examining the interaction between linaclotide and modulators of CFTR (such as VX-661, genistein, or VX-770) would be valuable to determine whether modulators and linaclotide could combine to further CFTR function in the intestine. Intestinal organoid swelling, which has been used extensively in this thesis, would provide an initial assessment of synergy between linaclotide and CFTR modulators. For example, examining organoid swelling in organoids from a G551D mouse model following treatment with linaclotide and the potentiator genistein would examine whether linaclotide activates intestinal G551D-CFTR after potentiation with genistein. Synergy between a CFTR modulator and linaclotide in intestinal organoids would strongly suggest that synergy between the two drugs would improve GI transit beyond either drug alone. Considering that linaclotide likely activates

CFTR through cGMP signaling, as well as cAMP signaling (similar to forskolin), it is likely that linaclotide would synergize with the many or all CFTR modulators.

In addition to activation of CFTR, Ahsan and colleagues have shown that GC-C signaling is sufficient to improve trafficking of CFTR to the apical membrane in both the rat small intestine and Caco-2 colon cell lines [198]. However, Ahsan and colleagues did not indicate whether this increase in CFTR quantity caused a subsequent increase in CFTR- mediated secretion. Furthermore, chronic treatment of STa has been found to improve the quantity of F508del-CFTR in intestinal organoids [199], indicating that GC-C signaling may have utility for increasing trafficking of Class II mutant CFTR to the apical membrane.

Interestingly, several other agents which increase cellular cGMP are currently being examined for their ability to improve F508del-CFTR trafficking. These agents include sildenafil, a phosphodiesterase 5 (PDE5) inhibitor which increases cellular cGMP by

157 preventing cGMP degradation by PDE5, and riociguat, a soluble guanylate cyclase (GC-

S) activator which catalyzes a GTP→cGMP reaction to increase cGMP [300]. Both of these drugs have been found to improve F508del-CFTR trafficking to the apical membrane

[301]. Additionally, PDE5 and GC-S both are expressed in a wide variety of tissues [302,

303], and thus any benefit from these drugs would not be limited to the intestine. Although the mechanism for sildenafil and riociguat for improving F508del-CFTR trafficking is unknown, these results collectively suggest that cGMP signaling may have utility for patients who express F508del-CFTR. In the future, the effects of chronic increases of cGMP on mice which express the F508del mutation will be examined.

5.2.5 Clinical Trials on Linaclotide Treatment in Patients with CF

The initial observation that linaclotide improves the GI symptoms of CF came from anecdotal claims from patients with CF who reported improvements after taking linaclotide. Here, we have produced abundant evidence to indicate that linaclotide is likely to be effective for treating CF GI maladies. However, a double-blind, placebo-controlled clinical trial is necessary to confirm that linaclotide is effective for treating CF GI symptoms in patients with CF. Linaclotide is already FDA-approved for treating IBS and

CC, and has been examined in clinical trials for non-CF patients [296, 304-306]. As linaclotide is FDA-approved, this clinical trial could be performed immediately. Ideally, patients should be recruited who have identical CFTR genotypes and are of similar age. If a clinical trial were to confirm that linaclotide was effective for treating GI maladies in CF, it would provide a much-needed source of relief for patients with CF who suffer from GI dysfunction.

158

The research performed in Chapter 2 has generated valuable information for treating the intestinal pathology of CF. The knowledge that linaclotide has efficacy for treating CF intestinal disorders will provide an immediate benefit to patients with CF, as linaclotide is an FDA-approved drug which could be readily be prescribed to patients for their GI discomfort. Additionally, the examination of the underlying mechanisms of cGMP signaling may provide a way to hydrate the surface of other epithelial tissues and improve

CFTR trafficking in patients with CF. Thus, activation of cGMP signaling, by linaclotide or another mechanism, may be a viable method to treat multiple aspects of CF pathology.

5.3 The G542X Mouse Model The research described in Chapter 3 details the creation of a mouse model of the

G542X nonsense mutation in murine Cftr. The G542X mouse has all the anticipated manifestations of a mouse model of CF nonsense mutations, including reduced growth, low survival, and reduced Cftr expression. Most importantly, this mouse model has utility for examining readthrough of PTCs, as demonstrated by detectable FIS in G418-treated

G542X intestinal organoids. As the G542X mutation has been generated in native Cftr, mRNA from this transcript is sensitive to NMD. We demonstrated that readthrough protects Cftr from NMD in the G542X mouse in both ex vivo intestinal organoids as well as in vivo lung tissue (Figure 3.5 D and E). Though the presented data indicate that the

G542X mouse will have utility for examining CFTR-directed therapies for nonsense mutations, further data are necessary to be able to examine the in vivo effects of readthrough treatment on this mouse model.

159

5.3.1 Restoring CFTR Function in the G542X Mouse

To understand how readthrough treatment in the G542X mouse will translate to patients with CF, further examination of the G542X mouse is necessary. Although functional CFTR is detectable following readthrough in G542X primary cell culture models, detecting restored CFTR function in G542X mouse tissue is essential to understand how effective a readthrough agent would be for a patient with nonsense mutations in CFTR.

We have demonstrated that aminoglycoside injections cause enough readthrough to increase Cftr quantity in the G542X mouse (Figure 3.5 E). However, it is unclear whether the quantity of CFTR restored is sufficient to be detectable. We plan to determine whether a similar aminoglycoside treatment protocol can restore functional levels of CFTR. In chapter 3, we used G418 injections to induce readthrough in G542X mice. However, the aminoglycoside gentamicin has also been found to restore detectable CFTR function in patients with CF [226]. Furthermore, G418 is highly cytotoxic to the mammalian cell which precludes its clinical use, and the reduced cytotoxicity of gentamicin may allow longer treatment periods. Therefore we will use chronic gentamicin injections to attempt to restore

CFTR function in G542X mice. Following several days of gentamicin injections, we will examine CFTR function in the airway and intestinal epithelium with measurements of NPD and Isc, respectively. Establishing the maximum CFTR function detectable using currently available readthrough agents will provide a baseline against which to compare any novel readthrough agents which may be tested in the future.

An ideal compound to compare against conventional aminoglycosides is the

“synthetic” aminoglycoside ELX-2. ELX-2 has been developed by modifying the structure of paromomycin to remove portions which induce cytotoxicity, and add in additional

160 structures which increase readthrough efficiency [234, 307-309]. As ELX-2 is not a naturally occurring compound and must be chemically manufactured, it is referred to as synthetic. These structural changes have been found to confer readthrough levels which are similar to that of G418, but with reduced cytotoxicity [309]. ELX-2 incubation facilitated robust FIS in human intestinal organoids expressing F508del/G542X mutations compared to G418, indicating potent readthrough properties [150, 310]. Due to these properties,

ELX-2 represents an ideal compound to restore CFTR function in an in vivo experiment in

G542X mice. ELX-2 is currently in Phase I clinical trials, and valuable information on the drug can be obtained using the G542X mouse model prior to advancement to Phase II clinical trials. For instance, functional examinations of drug efficacy in tissues such as the intestinal epithelium can be obtained using short-circuit current measurements in the

G542X mouse model. Additionally, determination of which tissues ELX-2 is active in can be determined using mice by examining G542X-Cftr expression across various Cftr- expressing tissues. Gaining insights into the efficacy and safety of ELX-2 may be useful prior to beginning Phase II clinical trials. Furthermore, this form of preclinical examination could be performed on any other novel readthrough agent prior to administration to humans.

5.3.2 Generating Additional CF Nonsense Mutation Models

A number of factors can influence the readthrough efficiency of a PTC. To properly examine all possible manifestations of CF nonsense mutations, more models of nonsense mutations in Cftr are necessary. Research on nonsense mutations has identified differences in readthrough efficiency between the UGA, UAA, and UAG termination codons. UGA has been shown to be the stop codon with the highest readthrough efficiency, followed by

161

UAG, with UAA being the most difficult to read through [311]. The G542X mutation creates a UGA PTC, theoretically making a relatively efficient PTC to read through.

Furthermore, the nucleotide sequence surrounding the PTC can impact the ability to read through the PTC. The nucleotide which immediately follows the PTC (referred to as the

+4 nucleotide) strongly influences the readthrough efficiency. A cytosine at the +4 position is confirmed to promote the greatest readthrough efficiency, but there is conflicting data on the influence of the other three nucleotides on readthrough [270, 312, 313]. Murine

G542X is immediately followed by a guanine nucleotide, and thus models only a few of the possible contexts in which a PTC can occur in CFTR. Therefore, additional models are necessary to examine the outcomes of readthrough on CFTR function in different PTC contexts. A desirable mutation to generate in CFTR to examine several of these possibilities is the Y122X mutation. Y122X is created by a T→A mutation which generates a UAA PTC in CFTR [237]. The +4 nucleotide in this mutation is a C, theoretically creating more efficient readthrough. Indeed, it was found that the Y122X mutation had greater readthrough efficiency than the UGA mutations W1282X, R1162X, and G542X [237].

This result was confirmed in patients who express at least one copy of the Y122X mutation.

These patients saw reduced sweat chloride levels compared to patients with W1282X,

R1162X, or G542X mutations following 15 days of intravenous gentamicin treatment

[237], suggesting restored CFTR function. A mouse model of the Y122X mutation will thus be valuable for determining whether reduced doses of readthrough agents are necessary to restore CFTR function in patients with this genotype.

Additionally, the location of the PTC along the transcript can have functional consequences for the CFTR protein. The region where G542X occurs on Cftr mRNA codes

162 for NBD1, which is located in the first half of the protein. However, PTCs may occur in different locations along the transcript, and the location of the PTC can modify how the

PTC impacts CFTR function. PTCs which occur in the first 50% of the transcript are unlikely to produce CFTR with any function. However, truncated CFTR from PTCs which occur near the end of CFTR have been found to have a small amount of function. For example, the W1282X mutation occurs in a region which codes for the NBD2 on CFTR, which is located near the C terminus of the protein. With only a small portion of the protein truncated, W1282X-CFTR is capable of avoiding immediate degradation [273]. W1282X-

CFTR function is not detectable by itself, but improving trafficking and potentiation of

W1282X-CFTR using VX-809 and VX-770 allows detection of W1282X-CFTR function

[273]. To examine the in vivo consequences of the W1282X mutation, the Hodges lab is in the process of developing a mouse model which expresses the W1282X mutation in murine Cftr. Homozygous W1282X mutants are beginning to be born, and are currently being characterized. Similar to the G542X mouse, W1282X is predicted to be a severe

CFTR mutation with little to no CFTR function detectable. Therefore, we anticipate that

W1282X mice will be reduced in growth and survival, as well as having low W1282X-Cftr expression. Generating and characterizing the W1282X mouse will be critical for fully understanding all aspects of nonsense mutations in CFTR.

5.3.3 Humanized CF Mice

Examining CFTR and CF biology using mouse models has generated invaluable data for the CF community. The mouse remains a very convenient model for examining in vivo effects of specific therapies, particularly therapies directed towards the GI system.

However, the mouse is less useful for examining in vivo effects of certain CFTR

163 modulators, as modulators such as VX-770 have significantly less effect on murine CFTR than human CFTR [95]. As a result, the mouse may not be an ideal model for preclinical examination of new CFTR modulators which may be more effective with human CFTR.

To overcome this shortcoming, the Hodges lab is currently in the process of generating lines of mice which lack murine Cftr, but express human CFTR. A mouse which lacks murine CFTR but expresses human CFTR integrated into the mouse genome as a bacterial artificial chromosome (BAC) transgene has been developed by Dr. Lane Clarke at

University of Missouri. In this mouse, human CFTR is functional, and is sufficient to prevent intestinal obstruction (Data currently submitted for publication). CRISPR/Cas9 gene editing is currently being utilized to insert mutations into these human CFTR transcripts, creating lines of mice which express the same mutated CFTR seen in humans.

Lines of mice will be generated which express representative mutations from Class I, II,

III, and V CFTR mutations. Mice representing the G542X, W1282X, F508del, G551D, and 3849 + 10kbC→T mutations will be created. These mice will be characterized to confirm that they have the expected manifestations of each mutation type. The humanized line of mice will be highly useful for the CF research community, as they combine the convenience of the mouse model with the responsiveness to CFTR modulators of human tissue. Clinical trials of novel drugs are very time consuming and expensive. The humanized CFTR mice will represent an ideal preclinical model to examine and prioritize novel compounds prior to entering a clinical trial.

5.4 Synergy between Readthrough and NMD Inhibition in CF Tissue As discussed previously, generation of the G542X nonsense mutation in native murine Cftr makes G542X-Cftr sensitive to degradation by NMD. Degradation of PTC- carrying mRNA is beneficial for cell health in the context of nonsense mutations, but makes

164 readthrough of PTCs more challenging. Due to low readthrough efficiency, the doses of readthrough agents required to restore gene function are too toxic for long-term administration. Therefore, inhibiting NMD is a potential method to increase the quantity of mutant mRNA substrate and thereby improve the efficiency of readthrough. With more efficient readthrough, the dose of readthrough agent could potentially be lowered to less or even non-toxic levels. In Chapter 4, we examined the utility of inhibiting NMD to improve the efficiency of readthrough using primary cell models from the G542X mouse. We observed robust synergy between SMG1i and several aminoglycoside readthrough agents including G418, gentamicin, and paromomycin (Figure 4.4 B). This synergy was observed in both airway (Figure 4.5 A and B) and intestinal tissue, indicating that NMD inhibition and readthrough synergize in two of the main disease tissue types in CF, and likely many other tissues as well. However, we did not observe NMD inhibition by two other previously reported NMD inhibitors, NMDI-14 and amlexanox. Furthermore, we did not find non- aminoglycoside readthrough agents such as escin or PTC124 to facilitate readthrough of the G542X PTC. Finally, we observed that trafficking of G542X-CFTR to the apical membrane following readthrough can be improved by combining SMG1i, readthrough agents, and VX-661. Collectively, the research in Chapter 4 indicates that NMD inhibitors are a promising treatment to improve the efficiency of readthrough agents when prescribed for nonsense mutations.

5.4.1 Spheroid Culture

Three-dimensional organoid culture has been a rapidly expanding area of research, ever since the developments of the first organoid cultures in 2009 [314]. It was found that by providing intestinal crypts suspended in MatriGel with the growth factors WNT3a,

165

Noggin, and R-spondin3, LGR5+ intestinal stem cells would proliferate in culture similar to in vivo [314]. This proliferation results in a budded three-dimensional structure which more accurately models the intestinal epithelium than conventional two-dimensional cell lines. Since the development of intestinal organoids, organoid cultures from a number of other tissues have been developed, including lung [315], trachea [316], pancreas [317], prostate [318], and brain [319]. For CF research, intestinal organoids have been invaluable for examining CFTR function, as they grow quickly, are relatively easy to maintain, and have a robust CFTR-dependent FIS response [150, 168, 267]. An entire culture of intestinal organoids can be grown from one rectal biopsy from a patient with CF, allowing for personalized examination of CFTR function. This has proven to be particularly valuable, as improvement in FIS response to CFTR modulators in organoids derived from a CF patient correlates with modulator effectiveness in the patient [267]. Therefore, groups in

Europe are currently biobanking intestinal organoid lines from the European CF population in hopes of characterizing modulator responses for a large number of individual patients.

5.4.2 Benefits of Murine Intestinal Organoids

Though many groups utilized intestinal organoids derived from patients with CF, murine intestinal organoids were utilized for the research in this thesis. Murine intestinal organoids do have several drawbacks, the largest being that some CFTR modulators such as VX-770 and VX-809 are less effective on murine CFTR than on human CFTR.

However, there are advantages to examining CFTR function in murine intestinal organoids.

Initially, alterations to organoid gene expression levels may change during longer periods in culture, potentially altering the results of drug screens. Murine intestinal organoids can be replenished more readily than human organoids, making potential genetic changes due

166 to culturing conditions easier to avoid. Additionally, mutations to murine CFTR are produced on a C57BL/6J background, ensuring a genetically homogenous background which is unavailable in human organoids. Finally, the majority of rare CFTR mutations do not frequently occur in a homozygous genotype. Rather, these rare mutations are co- expressed along with a more common mutation, such as F508del. Generating CFTR mutations in a mouse model allows homozygous expression of rare CFTR mutations which may not be available in patients. Furthermore, the research in this thesis largely examines

CFTR function in the context of CF nonsense mutations. Therefore, the defect in CFTR occurs at translation. The machinery governing translation is highly conserved across all eukaryotic species, making it unlikely that a therapy which is effective for facilitating readthrough using a murine ribosome would not be effective with a human ribosome.

Indeed, the majority of aminoglycosides, and even PTC124, have been shown to facilitate readthrough in both human and murine tissue. Therefore, knowledge gained regarding nonsense mutations in murine tissue is likely to translate to human tissue.

5.4.3 NMDI-14 and Amlexanox

We examined three NMD-inhibiting compounds in Chapter 4, NMDI-14 [264], amlexanox, and SMG1i [252]. Though we found SMG1i to effectively inhibit NMD,

NMDI-14 and amlexanox did not. There are several possible reasons why NMDI-14 did not inhibit NMD in intestinal organoids. First, NMDI-14 was specifically developed to prevent interaction between the NMD components UPF-1 and SMG-7. This interaction was targeted because UPF-1 and SMG-7 are only known to interact during NMD, and disrupting this interaction would be unlikely to impact other cellular processes [264].

Inhibiting this segment of the NMD pathway may not sufficiently inhibit NMD in a

167 therapeutic context, and inhibiting an alternate location may have a stronger effect.

Additionally, NMDI-14 was validated in multiple lung and breast carcinoma cell lines. The effect of NMDI-14 was variable between these cell lines, raising the possibility that certain cell types may be more resistant to NMDI-14 than others. Furthermore, NMDI-14 did not increase expression of progranulin in bone-marrow derived macrophages from a mouse model carrying a progranulin nonsense mutation [320], while NMDI-14 appeared to inhibit

NMD in the hippocampus of mice following induced seizures [321]. These conflicting results suggest that the effectiveness of NMDI-14 varies by tissue type, and raise the possibility that NMDI-14 is not effective in intestinal cells.

We also did not find amlexanox to inhibit NMD. This result was consistent with other examinations of NMD inhibition by amlexanox in intestinal organoids [150]. Unlike

NMDI-14, amlexanox was developed initially as an anti-ulcer medication, and was found to have NMD inhibiting properties in a high-throughput screen which aimed to repurpose existing compounds to inhibit NMD [264]. The mechanism of how amlexanox inhibits

NMD has not been determined, making it difficult to hypothesize why amlexanox did not inhibit NMD in G542X intestinal organoids. Additionally, amlexanox has been found to facilitate readthrough of PTCs, which we confirmed using a pFluc190UGA readthrough reporter (Supplementary Figure 7A). However, we did not find increases in G542X-Cftr quantity or CFTR function, suggesting that amlexanox is not causing readthrough in

G542X intestinal organoids. This contrasts with observations by several other groups, who found amlexanox to cause readthrough of PTCs in COL7A1 [322] and AGA [323]. The difference in effectiveness between different genes suggests that PTC context may be essential for readthrough by amlexanox, and that the amount of G542X-Cftr readthrough

168 caused by amlexanox is undetectable. As with NMDI-14, amlexanox may also have tissue- specific effects.

5.4.4 In vivo examination of SMG1i

We demonstrated that primary cells from several tissues in the G542X mouse are excellent models for assessing CFTR function following NMD inhibition and readthrough.

In particular, our experiments highlighted the effectiveness of combining SMG1i with readthrough to dramatically improve readthrough efficiency, resulting in greater CFTR function. However, it is currently unclear what the consequences of in vivo SMG1i administration would be. In the future, we plan to perform in vivo examinations of SMG1i to determine an optimal dose, method of administration, toxicity profile, and effectiveness on different tissue types. Examining the effect of SMG1i on different tissues would be determined using G542X mice, as we have already found that Cftr expression is reduced in the small intestine, colon, lung, kidney, and nasal epithelium due to NMD (Figure 3.2).

G542X mice can be treated with SMG1i, and changes in Cftr expression can be assessed in these tissues to determine NMD inhibition. An increase in Cftr mRNA would be indicative of NMD inhibition.

In addition to assessing restoration of Cftr expression, it will be necessary to determine whether in vivo CFTR function which has been restored by readthrough can be enhanced by SMG1i. As mentioned earlier, we have yet to demonstrate restoration of

CFTR function in mouse tissue using measurements of NPD or Isc, and it is possible that readthrough agents alone will not restore detectable CFTR function. If this occurs, we anticipate that SMG1i co-treatment would increase CFTR activity to detectable levels.

Furthermore, administration of CFTR correctors such as VX-661 may further improve

169

CFTR function, similar to what was seen in G542X intestinal organoids (Figure 4.7).

Although administration of readthrough agents, SMG1i, and VX-661 is a large number of drugs, this type of drug cocktail may be necessary to sufficiently restore CFTR in patients with CF nonsense mutations. Examining combination treatment of readthrough, SMG1i, and correctors in mouse models will provide valuable insight into whether such treatment would be viable for patients.

5.4.5 High Throughput Screening with Intestinal Organoids

High throughput screening (HTS) has been employed by several groups to identify compounds which facilitate readthrough of PTCs in CFTR. One of the most prominent readthrough drugs, PTC124, was discovered by screening using in vitro and cell-line based firefly luciferase reporters which had been disrupted by a PTC [131]. CFTR-directed readthrough therapies had also been screened for in Fisher Rat Thyroid (FRT) cells which had been transfected with exogenous sequences of CFTR linked to either firefly luciferase

[245] or a horseradish peroxidase tag [244] as a reporter. These reporter-based screens were successful in identifying compounds which suppress PTCs, but so far none of these compounds besides PTC124 have been tested in an in vivo context. The lack of translational benefit of these compounds may be because the compounds are effective for

PTC suppression in cell line reporter systems, but less effective in cells with native CFTR expression. As we demonstrated in Figure 4.1, these transgenic reporter systems are sensitive to very low degrees of readthrough. Similar doses of readthrough agents are insufficient to improve CFTR function in G542X organoids, a system with native levels of

Cftr expression (Figure 4.1). Therefore, screening for readthrough agents utilizing G542X

170 intestinal organoid FIS may increase the chances of identifying an agent which is effective for restoring CFTR function in native CFTR.

To utilize the intestinal organoids for HTS, we have adapted the FIS assay from a

48-well format to a 384-well format. Several changes were made to our protocol to ensure efficient FIS in the 384-well plate. The concentration of MatriGel was reduced to 10% from 50% to improve permeability of the MatriGel to compounds. Additionally, we increased the quantity of Matrigel:OGM to 15µl per well to ensure that organoids are distributed evenly across the well instead of accumulating around the edges. Finally, we improved organoid identification by staining with 0.5µg/ml Hoechst nuclear stain (Figure

5.1A). To validate the FIS assay in this format, we utilized G542X intestinal organoids, with G418-treated organoids as a positive control. G542X organoids are an ideal genotype to examine for this type of assay, as CFTR stimulation without G418 allows no swelling, while a robust FIS response occurs after G418 treatment (Figure 5.1B). This creates a large difference in signal between CFTR-corrected and uncorrected groups. Using this setup, we were able to observe robust FIS in G542X intestinal organoids treated with 80µM G418 for 72 hours (Figure 5.1A and B). Other CFTR mutation types (F508del, G551D) have a small degree of organoid swelling without correction of the underlying mutation (Figure

5.2), which would make it more difficult to differentiate between what would represent a

“hit” or no hit in a HTS. These preliminary data suggest that intestinal organoid FIS have applications for identifying novel readthrough agents or CFTR modulators in a HTS format.

We utilized G542X intestinal organoids for a HTS which screened approximately

3,000 bioactive small molecules. Each of these molecules was screened in triplicate,

171 making the screen require thirty 384 well plates to complete. The organoids were incubated with the small molecule library for 72 hours prior to forskolin stimulation. Brightfield and

DAPI images were acquired at 0 and 24 hours following stimulation with forskolin, and the change in area was quantified for each well. The cutoff for what was considered a hit was placed at three times the standard deviation from DMSO-treated organoids. Based on these screening parameters, we identified four compounds which produced a significant increase in organoid area (Figure 5.3). The canonical mechanism of each of these hits is displayed in Figure 5.4A, and a before/after image of one well of the strongest drug hit is displayed in Figure 5.4B. In particular, one of these hits recorded an average of 179% increase in organoid area, which was greater than that of G418. We are currently proceeding with post-screen validation of the strongest drug hit. Provided the compound continues to allow FIS in future experiments, we will proceed to confirm the readthrough mechanism by inspecting readthrough using the pFluc190UGA plasmid, as we did in

Figure 4.1 and Supplementary Figures 6 and 7.

Additionally, we have performed preliminary examination of W1282X intestinal organoids. W128X-CFTR has been found to be potentiated and corrected by VX-770 and

VX-809, respectively [273]. This is possible because the W1282X mutation occurs near the end of CFTR mRNA, allowing the majority of the CFTR protein to be translated prior to termination. This truncated protein does have some residual level of function. Using intestinal organoids from W1282X mice (discussed earlier), we were able to detect FIS following treatment with VX-661 and SMG1i, but without adding any readthrough agents

(Figure 5.5). No FIS was observable without VX-661 or SMG1i, however. This result indicates that W1282X intestinal organoids are capable of undergoing swelling with only

172 a corrector and inhibitor of NMD, and a readthrough agent is not needed. Therefore, it may be possible to use W1282X FIS as an output for NMD inhibition. This observation adds the potential to utilize intestinal organoid FIS in a HTS for inhibitors of NMD, instead of only readthrough agents or CFTR modulators.

5.4.6 Applying Readthrough Therapy to Heritable Cancer Syndromes

As previously mentioned, a large number of heritable diseases can arise as a result of nonsense mutations. However, nonsense mutations can also occur in genes which are not directly disease causing, but predispose a patient to develop certain forms of cancers

[114, 324]. These types of mutations cause what are known as “hereditary cancer syndromes”. Approximately 10-30% of patients who have hereditary cancer syndromes express nonsense mutations [114]. A classic example of a gene which is mutated in hereditary cancer syndromes is the P53 gene, which is the most commonly mutated gene in human cancers [325]. P53 is a tumor suppressor protein which is activated in response to DNA damage and other replication stressors, and acts to coordinate DNA repair [326,

327]. Nonsense mutations in P53 cause approximately 25% of P53 mutations [328]. Loss of P53 function predisposes patients to improper DNA repair and dysfunctional cell cycle regulation, which can eventually lead to tumor development. Restoration of P53 function in tumor cells may be able to slow progression of cancer cell growth. As a proof of concept, aminoglycoside-facilitated readthrough of P53 truncated by a PTC was sufficient to slow the growth of H1299 lung carcinoma cell lines [324]. This suggests that treatment with aminoglycosides could represent a possible strategy to slow the progression of cancer in patients with hereditary cancers syndromes caused by nonsense mutations. Furthermore, administration of readthrough therapy for cancer patients would not need to be lifelong.

173

Readthrough therapy would only be necessary to temporarily stall tumor growth, and the therapy can be stopped once the tumor has been removed. Thus, many of the concerns regarding aminoglycoside toxicity in long-term readthrough treatments are less prominent for treating tumors arising from nonsense mutations.

Concluding Statement

The knowledge gained from research on CF pathophysiology has been invaluable for developing treatments for CF symptoms. These treatments have been the driving force behind the increase in lifespan among patients with CF over the last 30 years. In particular, the development of CFTR modulators has represented a breakthrough in CF treatment, as mutated CFTR function was able to be significantly restored. However, many patients do not respond well to symptom-specific interventions, and patients who express nonsense mutations in CFTR are still without a CFTR-directed therapy. This thesis provides valuable insight into potential interventions which may correct both CF intestinal dysfunction and

CF nonsense mutations. Furthermore, we anticipate that the G542X mouse model will be an invaluable tool for examining nonsense mutation-directed therapies, not just for CF but all diseases arising from nonsense mutations. Ultimately, the research described in this thesis will be utilized to advance therapies which will extend and improve the lives of patients with CF.

174

Figure 5.1. G542X intestinal organoid FIS in a 384-well format. (A) G542X intestinal organoids were seeded to 384 well plates in 15µl of 10% MatriGel. The organoids were incubated for 72 hours with 80µM G418 or no drug prior to stimulation with 10µM forskolin. Images were captured for 24 hours using DAPI and brightfield kinetic imaging under live cell conditions. (B) Organoid FIS was quantified by automated image analysis based on the DAPI channel. Change in organoid area was normalized to T=0, N=5 wells per treatment group, ±SD.

175

Figure 5.2. WT, G551D, and F508del intestinal organoids undergo FIS without CFTR correction. Intestinal organoid from mouse models representing the indicated Cftr mutation were stimulated with 10µM forskolin and imaged for one hour using brightfield kinetic imaging. N=3-4 wells per experiment, ±SD.

176

Figure 5.3. Results of 3,000 bioactive small molecule HTS. Intestinal organoids in 384- well plates were treated with a compound library consisting of approximately 3,000 bioactive small molecules for 72 hours. The organoids were then stimulated with 10µM forskolin and imaged at 0 and 24 hours following stimulation. The percent change in area was calculated to determine hits. The cutoff for a hit was calculated as three times the standard deviation from DMSO treated organoids (indicated by the dashed line). Each dot represents the average of three individual experiments. Hits are displayed as red, non-hits are displayed as blue, and G418 positive control is displayed as black.

177

Figure 5.4. Four hits were recorded in a HTS for G542X intestinal organoid FIS. A. The canonical mechanism and percent change in area recorded by G418 and each hit. B. Sample pre and post FIS images of organoids treated with compound 1 (EZH2 inhibitor).

178

Figure 5.5. W1282X Intestinal organoids undergo swelling without readthrough. Intestinal organoids harvested from W1282X homozygote mice were incubated with 3.3µM VX-661 and 1µM SMG1i for 24 hours. Organoids were imaged at 0 hours and 24 hours following stimulation of CFTR with 10µM forskolin. Scale bar is 100µm. Images courtesy of Zachary Traylor of the Hodges lab.

179

References

1. Guo, Y., et al., Expression and distribution of cystic fibrosis transmembrane conductance regulator in neurons of the human brain. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society, 2009. 57(12): p. 1113-1120. 2. Stalvey, M.S., et al., Osteoblast CFTR inactivation reduces differentiation and osteoprotegerin expression in a mouse model of cystic fibrosis-related bone disease. PloS one, 2013. 8(11): p. e80098-e80098. 3. Xu, Y., et al., Influence of the cystic fibrosis transmembrane conductance regulator on expression of lipid metabolism-related genes in dendritic cells. Respiratory research, 2009. 10(1): p. 26-26. 4. Davies, J.C., E.W. Alton, and A. Bush, Cystic fibrosis. Bmj, 2007. 335(7632): p. 1255-9. 5. Anderson, D.H., Cystic fibrosis of the pancreas and its relation to celiac disease. A clinical and pathologic study. Am J Dis Child, 1938. 56: p. 344-399. 6. Riordan, J.R., et al., Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science, 1989. 245(4922): p. 1066-73. 7. 2017 Cystic Fibrosis Foundation Patient Registry Annual Report. 2017. 8. Zeitlin, P.L., et al., CFTR protein expression in primary and cultured epithelia. Proceedings of the National Academy of Sciences of the United States of America, 1992. 89(1): p. 344-347. 9. Sheppard, D.N. and M.J. Welsh, Structure and function of the CFTR chloride channel. Physiol Rev, 1999. 79(1 Suppl): p. S23-45. 10. Zhang, Z., F. Liu, and J. Chen, Molecular structure of the ATP-bound, phosphorylated human CFTR. Proc Natl Acad Sci U S A, 2018. 115(50): p. 12757-12762. 11. Patrick, A.E., et al., Alteration of CFTR transmembrane span integration by disease-causing mutations. Molecular biology of the cell, 2011. 22(23): p. 4461- 4471. 12. Hwang, T.-C. and K.L. Kirk, The CFTR ion channel: gating, regulation, and anion permeation. Cold Spring Harbor perspectives in medicine. 3(1): p. a009498-a009498. 13. Ma, J., et al., Function of the R domain in the cystic fibrosis transmembrane conductance regulator chloride channel. J Biol Chem, 1997. 272(44): p. 28133- 41. 14. Uc, A., et al., Pancreatic and biliary secretion are both altered in cystic fibrosis pigs. Am J Physiol Gastrointest Liver Physiol, 2012. 303(8): p. G961-8. 15. McCarron, A., M. Donnelley, and D. Parsons, Airway disease phenotypes in animal models of cystic fibrosis. Respir Res, 2018. 19(1): p. 54. 16. Donaldson, S.H., et al., Mucus clearance and lung function in cystic fibrosis with hypertonic saline. N Engl J Med, 2006. 354(3): p. 241-50. 17. Pillarisetti, N., et al., Infection, inflammation, and lung function decline in infants with cystic fibrosis. Am J Respir Crit Care Med, 2011. 184(1): p. 75-81.

180

18. Chmiel, J.F. and P.B. Davis, State of the art: why do the lungs of patients with cystic fibrosis become infected and why can't they clear the infection? Respir Res, 2003. 4: p. 8. 19. Ferec, C. and G.R. Cutting, Assessing the Disease-Liability of Mutations in CFTR. Cold Spring Harb Perspect Med, 2012. 2(12): p. a009480. 20. Du, M., et al., Aminoglycoside suppression of a premature stop mutation in a Cftr-/- mouse carrying a human CFTR-G542X transgene. J Mol Med (Berl), 2002. 80(9): p. 595-604. 21. Zielenski, J. and L.C. Tsui, Cystic fibrosis: genotypic and phenotypic variations. Annu Rev Genet, 1995. 29: p. 777-807. 22. Schloesser, M., et al., A cystic fibrosis patient with the nonsense mutation G542X and the splice site mutation 1717-1. J Med Genet, 1991. 28(12): p. 878-80. 23. Beck, S., et al., Cystic fibrosis patients with the 3272-26A-->G mutation have mild disease, leaky alternative mRNA splicing, and CFTR protein at the cell membrane. Hum Mutat, 1999. 14(2): p. 133-44. 24. Ratjen, F. and G. Doring, Cystic fibrosis. Lancet, 2003. 361(9358): p. 681-9. 25. Bompadre, S.G., et al., G551D and G1349D, two CF-associated mutations in the signature sequences of CFTR, exhibit distinct gating defects. The Journal of general physiology, 2007. 129(4): p. 285-298. 26. Ramsey, B.W., et al., A CFTR potentiator in patients with cystic fibrosis and the G551D mutation. N Engl J Med, 2011. 365(18): p. 1663-72. 27. Sheppard, D.N., et al., Mutations in CFTR associated with mild-disease-form Cl- channels with altered pore properties. Nature, 1993. 362(6416): p. 160-4. 28. Thauvin-Robinet, C., et al., The very low penetrance of cystic fibrosis for the R117H mutation: a reappraisal for genetic counselling and newborn screening. J Med Genet, 2009. 46(11): p. 752-8. 29. Yu, Y.C., Y. Sohma, and T.C. Hwang, On the mechanism of gating defects caused by the R117H mutation in cystic fibrosis transmembrane conductance regulator. J Physiol, 2016. 594(12): p. 3227-44. 30. Curnow, L., R. Savarirayan, and J. Massie, Genetic counselling after carrier detection by newborn screening when one parent carries DeltaF508 and the other R117H. Archives of disease in childhood, 2003. 88(10): p. 886-888. 31. Wilschanski, M., et al., Correlation of sweat chloride concentration with classes of the cystic fibrosis transmembrane conductance regulator gene mutations. J Pediatr, 1995. 127(5): p. 705-10. 32. Stern, R.C., C.F. Doershuk, and M.L. Drumm, 3849+10 kb C-->T mutation and disease severity in cystic fibrosis. Lancet, 1995. 346(8970): p. 274-6. 33. Dugueperoux, I. and M. De Braekeleer, The CFTR 3849+10kbC->T and 2789+5G->A alleles are associated with a mild CF phenotype. Eur Respir J, 2005. 25(3): p. 468-73. 34. Belkin, R.A., et al., Risk factors for death of patients with cystic fibrosis awaiting lung transplantation. American journal of respiratory and critical care medicine, 2006. 173(6): p. 659-666. 35. Morgan, W.J., et al., Forced Expiratory Volume in 1 Second Variability Helps Identify Patients with Cystic Fibrosis at Risk of Greater Loss of Lung Function. J Pediatr, 2016. 169: p. 116-21.e2.

181

36. Lyczak, J.B., C.L. Cannon, and G.B. Pier, Lung infections associated with cystic fibrosis. Clin Microbiol Rev, 2002. 15(2): p. 194-222. 37. Filkins, L.M. and G.A. O'Toole, Cystic Fibrosis Lung Infections: Polymicrobial, Complex, and Hard to Treat. PLoS Pathog, 2015. 11(12): p. e1005258. 38. Surette, M.G., The cystic fibrosis lung microbiome. Ann Am Thorac Soc, 2014. 11 Suppl 1: p. S61-5. 39. Voynow, J.A., B.M. Fischer, and S. Zheng, Proteases and cystic fibrosis. The international journal of biochemistry & cell biology, 2008. 40(6-7): p. 1238-1245. 40. Bhatt, J.M., Treatment of pulmonary exacerbations in cystic fibrosis. Eur Respir Rev, 2013. 22(129): p. 205-16. 41. Wilschanski, M. and I. Novak, The cystic fibrosis of exocrine pancreas. Cold Spring Harb Perspect Med, 2013. 3(5): p. a009746. 42. Gibson-Corley, K.N., D.K. Meyerholz, and J.F. Engelhardt, Pancreatic pathophysiology in cystic fibrosis. J Pathol, 2016. 238(2): p. 311-20. 43. Sinaasappel, M., et al., Nutrition in patients with cystic fibrosis: a European Consensus. J Cyst Fibros, 2002. 1(2): p. 51-75. 44. Sikkens, E.C., et al., Pancreatic enzyme replacement therapy in chronic pancreatitis. Best Pract Res Clin Gastroenterol, 2010. 24(3): p. 337-47. 45. Peretti, N., et al., Mechanisms of lipid malabsorption in Cystic Fibrosis: the impact of essential fatty acids deficiency. Nutr Metab (Lond), 2005. 2(1): p. 11. 46. Huang, W.Q., et al., Abnormal CFTR Affects Glucagon Production by Islet α Cells in Cystic Fibrosis and Polycystic Ovarian Syndrome. Frontiers in physiology, 2017. 8: p. 835-835. 47. Kayani, K., R. Mohammed, and H. Mohiaddin, Cystic Fibrosis-Related Diabetes. Frontiers in endocrinology, 2018. 9: p. 20-20. 48. Moran, A., Cystic fibrosis-related diabetes: an approach to diagnosis and management. Pediatr Diabetes, 2000. 1(1): p. 41-8. 49. Milla, C.E., W.J. Warwick, and A. Moran, Trends in pulmonary function in patients with cystic fibrosis correlate with the degree of glucose intolerance at baseline. Am J Respir Crit Care Med, 2000. 162(3 Pt 1): p. 891-5. 50. Finkelstein, S.M., et al., Diabetes mellitus associated with cystic fibrosis. J Pediatr, 1988. 112(3): p. 373-7. 51. De Lisle, R.C. and D. Borowitz, The cystic fibrosis intestine. Cold Spring Harb Perspect Med, 2013. 3(9): p. a009753. 52. Strong, T.V., K. Boehm, and F.S. Collins, Localization of cystic fibrosis transmembrane conductance regulator mRNA in the human gastrointestinal tract by in situ hybridization. J Clin Invest, 1994. 93(1): p. 347-54. 53. Liu, X., et al., Loss of Slc26a9 anion transporter alters intestinal electrolyte and HCO3(-) transport and reduces survival in CFTR-deficient mice. Pflugers Arch, 2015. 467(6): p. 1261-75. 54. van der Doef, H.P., et al., Constipation in pediatric cystic fibrosis patients: an underestimated medical condition. J Cyst Fibros, 2010. 9(1): p. 59-63. 55. Colombo, C., et al., Guidelines for the diagnosis and management of distal intestinal obstruction syndrome in cystic fibrosis patients. J Cyst Fibros, 2011. 10 Suppl 2: p. S24-8.

182

56. van der Doef, H.P., et al., Intestinal obstruction syndromes in cystic fibrosis: meconium ileus, distal intestinal obstruction syndrome, and constipation. Curr Gastroenterol Rep, 2011. 13(3): p. 265-70. 57. Camilleri, M., K. Lasch, and W. Zhou, Irritable bowel syndrome: methods, mechanisms, and pathophysiology. The confluence of increased permeability, inflammation, and pain in irritable bowel syndrome. Am J Physiol Gastrointest Liver Physiol, 2012. 303(7): p. G775-85. 58. Fridge, J.L., et al., Risk factors for small bowel bacterial overgrowth in cystic fibrosis. J Pediatr Gastroenterol Nutr, 2007. 44(2): p. 212-8. 59. Bures, J., et al., Small intestinal bacterial overgrowth syndrome. World J Gastroenterol, 2010. 16(24): p. 2978-90. 60. Dukowicz, A.C., B.E. Lacy, and G.M. Levine, Small intestinal bacterial overgrowth: a comprehensive review. Gastroenterol Hepatol (N Y), 2007. 3(2): p. 112-22. 61. Hodges, C.A., et al., Cystic fibrosis growth retardation is not correlated with loss of Cftr in the intestinal epithelium. Am J Physiol Gastrointest Liver Physiol, 2011. 301(3): p. G528-36. 62. Di Sant'Agnese, P.A., et al., Abnormal electrolyte composition of sweat in cystic fibrosis of the pancreas; clinical significance and relationship to the disease. Pediatrics, 1953. 12(5): p. 549-63. 63. Rosenstein, B.J. and G.R. Cutting, The diagnosis of cystic fibrosis: a consensus statement. Cystic Fibrosis Foundation Consensus Panel. J Pediatr, 1998. 132(4): p. 589-95. 64. Lindblad, A., H. Glaumann, and B. Strandvik, Natural history of liver disease in cystic fibrosis. Hepatology, 1999. 30(5): p. 1151-8. 65. Kobelska-Dubiel, N., B. Klincewicz, and W. Cichy, Liver disease in cystic fibrosis. Przeglad gastroenterologiczny, 2014. 9(3): p. 136-141. 66. Assis, D.N. and D. Debray, Gallbladder and bile duct disease in Cystic Fibrosis. J Cyst Fibros, 2017. 16 Suppl 2: p. S62-s69. 67. Anguiano, A., et al., Congenital bilateral absence of the vas deferens. A primarily genital form of cystic fibrosis. Jama, 1992. 267(13): p. 1794-7. 68. de Souza, D.A.S., et al., Congenital bilateral absence of the vas deferens as an atypical form of cystic fibrosis: reproductive implications and genetic counseling. Andrology, 2018. 6(1): p. 127-135. 69. Mischler, E.H., et al., Demineralization in cystic fibrosis detected by direct photon absorptiometry. Am J Dis Child, 1979. 133(6): p. 632-5. 70. Hahn, T.J., et al., Reduced serum 25-hydroxyvitamin D concentration and disordered mineral metabolism in patients with cystic fibrosis. J Pediatr, 1979. 94(1): p. 38-42. 71. Dominguez-Munoz, J.E., Pancreatic enzyme replacement therapy: exocrine pancreatic insufficiency after gastrointestinal surgery. HPB (Oxford), 2009. 11 Suppl 3: p. 3-6. 72. Clarke, L.L., et al., Increased survival of CFTR knockout mice with an oral osmotic laxative. Lab Anim Sci, 1996. 46(6): p. 612-8.

183

73. Minguez, M., A. Lopez Higueras, and J. Judez, Use of polyethylene glycol in functional constipation and fecal impaction. Rev Esp Enferm Dig, 2016. 108(12): p. 790-806. 74. DiPalma, J.A., et al., A randomized, placebo-controlled, multicenter study of the safety and efficacy of a new polyethylene glycol laxative. Am J Gastroenterol, 2000. 95(2): p. 446-50. 75. De Lisle, R.C., E. Roach, and K. Jansson, Effects of laxative and N-acetylcysteine on mucus accumulation, bacterial load, transit, and inflammation in the cystic fibrosis mouse small intestine. Am J Physiol Gastrointest Liver Physiol, 2007. 293(3): p. G577-84. 76. Davidson, A.C., et al., Distal intestinal obstruction syndrome in cystic fibrosis treated by oral intestinal lavage, and a case of recurrent obstruction despite normal pancreatic function. Thorax, 1987. 42(7): p. 538-41. 77. Adamcewicz, M., et al., Mechanism of action and toxicities of purgatives used for colonoscopy preparation. Expert Opin Drug Metab Toxicol, 2011. 7(1): p. 89- 101. 78. Khan, S., et al., Overcoming drug resistance with alginate oligosaccharides able to potentiate the action of selected antibiotics. Antimicrob Agents Chemother, 2012. 56(10): p. 5134-41. 79. Nordgard, C.T. and K.I. Draget, Oligosaccharides as modulators of rheology in complex mucous systems. Biomacromolecules, 2011. 12(8): p. 3084-90. 80. Nordgard, C.T., et al., Alterations in mucus barrier function and matrix structure induced by guluronate oligomers. Biomacromolecules, 2014. 15(6): p. 2294-300. 81. Pritchard, M.F., et al., A New Class of Safe Oligosaccharide Polymer Therapy To Modify the Mucus Barrier of Chronic Respiratory Disease. Mol Pharm, 2016. 13(3): p. 863-72. 82. Vitko, M., et al., A novel guluronate oligomer improves intestinal transit and survival in cystic fibrosis mice. J Cyst Fibros, 2016. 15(6): p. 745-751. 83. Gawenis, L.R., et al., Intestinal NaCl transport in NHE2 and NHE3 knockout mice. Am J Physiol Gastrointest Liver Physiol, 2002. 282(5): p. G776-84. 84. He, P. and C.C. Yun, Mechanisms of the regulation of the intestinal Na+/H+ exchanger NHE3. J Biomed Biotechnol, 2010. 2010: p. 238080. 85. Schwark, J.R., et al., S3226, a novel inhibitor of Na+/H+ exchanger subtype 3 in various cell types. Pflugers Arch, 1998. 436(5): p. 797-800. 86. Yun, C.H., et al., Mammalian Na+/H+ exchanger gene family: structure and function studies. Am J Physiol, 1995. 269(1 Pt 1): p. G1-11. 87. Bradford, E.M., et al., Reduced NHE3-mediated Na+ absorption increases survival and decreases the incidence of intestinal obstructions in cystic fibrosis mice. Am J Physiol Gastrointest Liver Physiol, 2009. 296(4): p. G886-98. 88. Zielinska, M., A. Wasilewski, and J. Fichna, Tenapanor hydrochloride for the treatment of constipation-predominant irritable bowel syndrome. Expert Opin Investig Drugs, 2015. 24(8): p. 1093-9. 89. Labonte, E.D., et al., Gastrointestinal Inhibition of Sodium-Hydrogen Exchanger 3 Reduces Phosphorus Absorption and Protects against Vascular Calcification in CKD. J Am Soc Nephrol, 2015. 26(5): p. 1138-49.

184

90. Spencer, A.G., et al., Intestinal inhibition of the Na+/H+ exchanger 3 prevents cardiorenal damage in rats and inhibits Na+ uptake in humans. Sci Transl Med, 2014. 6(227): p. 227ra36. 91. Chey, W.D., A.J. Lembo, and D.P. Rosenbaum, Tenapanor Treatment of Patients With Constipation-Predominant Irritable Bowel Syndrome: A Phase 2, Randomized, Placebo-Controlled Efficacy and Safety Trial. The American Journal Of Gastroenterology, 2017. 112: p. 763. 92. Sermet-Gaudelus, I., et al., Pain in children and adults with cystic fibrosis: a comparative study. J Pain Symptom Manage, 2009. 38(2): p. 281-90. 93. Hwang, T.C., et al., Genistein potentiates wild-type and delta F508-CFTR channel activity. Am J Physiol, 1997. 273(3 Pt 1): p. C988-98. 94. Van Goor, F., et al., Rescue of DeltaF508-CFTR trafficking and gating in human cystic fibrosis airway primary cultures by small molecules. Am J Physiol Lung Cell Mol Physiol, 2006. 290(6): p. L1117-30. 95. Van Goor, F., et al., Rescue of CF airway epithelial cell function in vitro by a CFTR potentiator, VX-770. Proc Natl Acad Sci U S A, 2009. 106(44): p. 18825- 30. 96. Accurso, F.J., et al., Effect of VX-770 in persons with cystic fibrosis and the G551D-CFTR mutation. N Engl J Med, 2010. 363(21): p. 1991-2003. 97. Jih, K.Y. and T.C. Hwang, Vx-770 potentiates CFTR function by promoting decoupling between the gating cycle and ATP hydrolysis cycle. Proc Natl Acad Sci U S A, 2013. 110(11): p. 4404-9. 98. Bobadilla, J.L., et al., Cystic fibrosis: a worldwide analysis of CFTR mutations-- correlation with incidence data and application to screening. Hum Mutat, 2002. 19(6): p. 575-606. 99. Castellani, C., et al., Consensus on the use and interpretation of cystic fibrosis mutation analysis in clinical practice. J Cyst Fibros, 2008. 7(3): p. 179-96. 100. Van Goor, F., et al., Correction of the F508del-CFTR protein processing defect in vitro by the investigational drug VX-809. Proc Natl Acad Sci U S A, 2011. 108(46): p. 18843-8. 101. Clancy, J.P., et al., Results of a phase IIa study of VX-809, an investigational CFTR corrector compound, in subjects with cystic fibrosis homozygous for the F508del-CFTR mutation. Thorax, 2012. 67(1): p. 12-8. 102. Wainwright, C.E., J.S. Elborn, and B.W. Ramsey, Lumacaftor-Ivacaftor in Patients with Cystic Fibrosis Homozygous for Phe508del CFTR. N Engl J Med, 2015. 373(18): p. 1783-4. 103. Hubert, D., et al., Real-life initiation of lumacaftor/ivacaftor combination in adults with cystic fibrosis homozygous for the Phe508del CFTR mutation and severe lung disease. J Cyst Fibros, 2017. 16(3): p. 388-391. 104. Van der Plas, S.E., et al., Discovery of N-(3-Carbamoyl-5,5,7,7-tetramethyl-5,7- dihydro-4H-thieno[2,3-c]pyran-2-yl)-lH-pyr azole-5-carboxamide (GLPG1837), a Novel Potentiator Which Can Open Class III Mutant Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) Channels to a High Extent. J Med Chem, 2018. 61(4): p. 1425-1435.

185

105. Davies, J., et al., WS13.6 GLPG1837 in subjects with cystic fibrosis (CF) and the G551D mutation: results from a phase II study (SAPHIRA1). Journal of Cystic Fibrosis, 2017. 16: p. S24-S25. 106. Taylor-Cousar, J.L., et al., Tezacaftor-Ivacaftor in Patients with Cystic Fibrosis Homozygous for Phe508del. N Engl J Med, 2017. 377(21): p. 2013-2023. 107. Wang, X., et al., Discovery of 4-[(2R,4R)-4-({[1-(2,2-Difluoro-1,3-benzodioxol-5- yl)cyclopropyl]carbonyl}amino)- 7-(difluoromethoxy)-3,4-dihydro-2H-chromen- 2-yl]benzoic Acid (ABBV/GLPG-2222), a Potent Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) Corrector for the Treatment of Cystic Fibrosis. J Med Chem, 2018. 61(4): p. 1436-1449. 108. Davies, J.C., et al., VX-659-Tezacaftor-Ivacaftor in Patients with Cystic Fibrosis and One or Two Phe508del Alleles. N Engl J Med, 2018. 379(17): p. 1599-1611. 109. Keating, D., et al., VX-445-Tezacaftor-Ivacaftor in Patients with Cystic Fibrosis and One or Two Phe508del Alleles. N Engl J Med, 2018. 379(17): p. 1612-1620. 110. Giuliano, K.A., et al., Use of a High-Throughput Phenotypic Screening Strategy to Identify Amplifiers, a Novel Pharmacological Class of Small Molecules That Exhibit Functional Synergy with Potentiators and Correctors. SLAS Discov, 2018. 23(2): p. 111-121. 111. Molinski, S.V., et al., Orkambi(R) and amplifier co-therapy improves function from a rare CFTR mutation in gene-edited cells and patient tissue. EMBO Mol Med, 2017. 9(9): p. 1224-1243. 112. Bidou, L., et al., Sense from nonsense: therapies for premature stop codon diseases. Trends Mol Med, 2012. 18(11): p. 679-88. 113. Population variation of common cystic fibrosis mutations. The Cystic Fibrosis Genetic Analysis Consortium. Hum Mutat, 1994. 4(3): p. 167-77. 114. Bordeira-Carrico, R., et al., Cancer syndromes and therapy by stop-codon readthrough. Trends Mol Med, 2012. 18(11): p. 667-78. 115. Neu-Yilik, G., et al., Mechanism of escape from nonsense-mediated mRNA decay of human beta-globin transcripts with nonsense mutations in the first exon. Rna, 2011. 17(5): p. 843-54. 116. Pestova, T.V., et al., Molecular mechanisms of translation initiation in eukaryotes. Proceedings of the National Academy of Sciences, 2001. 98(13): p. 7029-7036. 117. Keeling, K.M., et al., Therapeutics based on stop codon readthrough. Annu Rev Genomics Hum Genet, 2014. 15: p. 371-94. 118. Prabhakar, A., et al., Dynamic basis of fidelity and speed in translation: Coordinated multistep mechanisms of elongation and termination. Protein Sci, 2017. 26(7): p. 1352-1362. 119. Goldberg, A.L., Protein degradation and protection against misfolded or damaged proteins. Nature, 2003. 426(6968): p. 895-9. 120. Le Hir, H., et al., The exon–exon junction complex provides a binding platform for factors involved in mRNA export and nonsense‐mediated mRNA decay. The EMBO Journal, 2001. 20(17): p. 4987-4997. 121. Kashima, I., et al., Binding of a novel SMG-1-Upf1-eRF1-eRF3 complex (SURF) to the exon junction complex triggers Upf1 phosphorylation and nonsense- mediated mRNA decay. Genes Dev, 2006. 20(3): p. 355-67.

186

122. Hug, N., D. Longman, and J.F. Caceres, Mechanism and regulation of the nonsense-mediated decay pathway. Nucleic Acids Res, 2016. 44(4): p. 1483-95. 123. Melero, R., et al., Structures of SMG1-UPFs complexes: SMG1 contributes to regulate UPF2-dependent activation of UPF1 in NMD. Structure, 2014. 22(8): p. 1105-1119. 124. Deniaud, A., et al., A network of SMG-8, SMG-9 and SMG-1 C-terminal insertion domain regulates UPF1 substrate recruitment and phosphorylation. Nucleic Acids Res, 2015. 43(15): p. 7600-11. 125. Garreau de Loubresse, N., et al., Structural basis for the inhibition of the eukaryotic ribosome. Nature, 2014. 513(7519): p. 517-22. 126. Lynch, S.R. and J.D. Puglisi, Structural origins of aminoglycoside specificity for prokaryotic ribosomes. J Mol Biol, 2001. 306(5): p. 1037-58. 127. Kondo, J., et al., Differential selectivity of natural and synthetic aminoglycosides towards the eukaryotic and prokaryotic decoding A sites. Chembiochem, 2007. 8(14): p. 1700-9. 128. Sandoval, R.M., et al., A non-nephrotoxic gentamicin congener that retains antimicrobial efficacy. J Am Soc Nephrol, 2006. 17(10): p. 2697-705. 129. Lopez-Novoa, J.M., et al., New insights into the mechanism of aminoglycoside nephrotoxicity: an integrative point of view. Kidney Int, 2011. 79(1): p. 33-45. 130. Baradaran-Heravi, A., et al., Gentamicin B1 is a minor gentamicin component with major nonsense mutation suppression activity. Proceedings of the National Academy of Sciences, 2017. 114(13): p. 3479-3484. 131. Welch, E.M., et al., PTC124 targets genetic disorders caused by nonsense mutations. Nature, 2007. 447(7140): p. 87-91. 132. Laurent, G., et al., Mechanism of aminoglycoside-induced lysosomal phospholipidosis: in vitro and in vivo studies with gentamicin and amikacin. Biochem Pharmacol, 1982. 31(23): p. 3861-70. 133. Du, M., et al., Poly-L-aspartic acid enhances and prolongs gentamicin-mediated suppression of the CFTR-G542X mutation in a cystic fibrosis mouse model. J Biol Chem, 2009. 284(11): p. 6885-92. 134. Namy, O., I. Hatin, and J.P. Rousset, Impact of the six nucleotides downstream of the stop codon on translation termination. EMBO Rep, 2001. 2(9): p. 787-93. 135. Baker, S.L. and J.R. Hogg, A system for coordinated analysis of translational readthrough and nonsense-mediated mRNA decay. PLoS One, 2017. 12(3): p. e0173980. 136. Pranke, I., et al., Factors influencing readthrough therapy for frequent cystic fibrosis premature termination codons. ERJ Open Res, 2018. 4(1). 137. Tuggle, K.L., et al., Characterization of defects in ion transport and tissue development in cystic fibrosis transmembrane conductance regulator (CFTR)- knockout rats. PLoS One, 2014. 9(3): p. e91253. 138. Clarke, L.L., et al., Defective epithelial chloride transport in a gene-targeted mouse model of cystic fibrosis. Science, 1992. 257(5073): p. 1125-8. 139. Wilke, M., et al., Mouse models of cystic fibrosis: phenotypic analysis and research applications. J Cyst Fibros, 2011. 10 Suppl 2: p. S152-71. 140. Rogers, C.S., et al., Disruption of the CFTR gene produces a model of cystic fibrosis in newborn pigs. Science, 2008. 321(5897): p. 1837-41.

187

141. Fan, Z., et al., A sheep model of cystic fibrosis generated by CRISPR/Cas9 disruption of the CFTR gene. JCI Insight, 2018. 3(19). 142. Navis, A., L. Marjoram, and M. Bagnat, Cftr controls lumen expansion and function of Kupffer's vesicle in zebrafish. Development, 2013. 140(8): p. 1703-12. 143. Sun, X., et al., Disease phenotype of a ferret CFTR-knockout model of cystic fibrosis. J Clin Invest, 2010. 120(9): p. 3149-60. 144. Nguyen, D. and T. Xu, The expanding role of mouse genetics for understanding human biology and disease. Dis Model Mech, 2008. 1(1): p. 56-66. 145. Perlman, R.L., Mouse models of human disease: An evolutionary perspective. Evol Med Public Health, 2016. 2016(1): p. 170-6. 146. Eckman, E.A., et al., Dietary changes improve survival of CFTR S489X homozygous mutant mouse. Am J Physiol, 1995. 269(5 Pt 1): p. L625-30. 147. Darrah, R.J., et al., Early pulmonary disease manifestations in cystic fibrosis mice. J Cyst Fibros, 2016. 15(6): p. 736-744. 148. Durie, P.R., et al., Characteristic multiorgan pathology of cystic fibrosis in a long-living cystic fibrosis transmembrane regulator knockout murine model. Am J Pathol, 2004. 164(4): p. 1481-93. 149. Koller, B.H., et al., Toward an animal model of cystic fibrosis: targeted interruption of exon 10 of the cystic fibrosis transmembrane regulator gene in embryonic stem cells. Proc Natl Acad Sci U S A, 1991. 88(23): p. 10730-4. 150. Zomer-van Ommen, D.D., et al., Limited premature termination codon suppression by read-through agents in cystic fibrosis intestinal organoids. J Cyst Fibros, 2016. 15(2): p. 158-62. 151. Fanen, P., A. Wohlhuter-Haddad, and A. Hinzpeter, Genetics of cystic fibrosis: CFTR mutation classifications toward genotype-based CF therapies. Int J Biochem Cell Biol, 2014. 52: p. 94-102. 152. Vitko, M., Intestinal Dysfunction in Cystic Fibrosis, in Genetics and Genome Sciences. 2016, Case Western Reserve University: OhioLINK. p. 195. 153. Sun, L., et al., Multiple apical plasma membrane constituents are associated with susceptibility to meconium ileus in individuals with cystic fibrosis. Nat Genet, 2012. 44(5): p. 562-9. 154. Corsetti, M. and J. Tack, Linaclotide: A new drug for the treatment of chronic constipation and irritable bowel syndrome with constipation. United European Gastroenterol J, 2013. 1(1): p. 7-20. 155. Weiglmeier, P.R., P. Rosch, and H. Berkner, Cure and curse: E. coli heat-stable enterotoxin and its receptor guanylyl cyclase C. Toxins (Basel), 2010. 2(9): p. 2213-29. 156. Hannig, G., et al., Guanylate cyclase-C/cGMP: an emerging pathway in the regulation of visceral pain. Front Mol Neurosci, 2014. 7: p. 31. 157. Bryant, A.P., et al., Linaclotide is a potent and selective guanylate cyclase C agonist that elicits pharmacological effects locally in the gastrointestinal tract. Life Sci, 2010. 86(19-20): p. 760-5. 158. Donowitz, M., et al., NHERF family and NHE3 regulation. J Physiol, 2005. 567(Pt 1): p. 3-11. 159. Zeiher, B.G., et al., A mouse model for the delta F508 allele of cystic fibrosis. J Clin Invest, 1995. 96(4): p. 2051-64.

188

160. Snouwaert, J.N., et al., An animal model for cystic fibrosis made by gene targeting. Science, 1992. 257(5073): p. 1083-8. 161. De Lisle, R.C., Altered transit and bacterial overgrowth in the cystic fibrosis mouse small intestine. Am J Physiol Gastrointest Liver Physiol, 2007. 293(1): p. G104-11. 162. Busby, R.W., et al., Pharmacologic properties, metabolism, and disposition of linaclotide, a novel therapeutic peptide approved for the treatment of irritable bowel syndrome with constipation and chronic idiopathic constipation. J Pharmacol Exp Ther, 2013. 344(1): p. 196-206. 163. Reagan-Shaw, S., M. Nihal, and N. Ahmad, Dose translation from animal to human studies revisited. FASEB J, 2008. 22(3): p. 659-61. 164. Hodges, C.A., et al., Generation of a conditional null allele for Cftr in mice. Genesis, 2008. 46(10): p. 546-52. 165. Sato, T., et al., Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett's epithelium. Gastroenterology, 2011. 141(5): p. 1762-72. 166. McHugh, D.R., et al., A G542X cystic fibrosis mouse model for examining nonsense mutation directed therapies. PLoS One, 2018. 13(6): p. e0199573. 167. Pattison, A.M., et al., Intestinal Enteroids Model Guanylate Cyclase C-Dependent Secretion Induced by Heat-Stable Enterotoxins. Infect Immun, 2016. 84(10): p. 3083-91. 168. Dekkers, J.F., et al., A functional CFTR assay using primary cystic fibrosis intestinal organoids. Nat Med, 2013. 19(7): p. 939-45. 169. Cotton, C.U. and L. Reuss, Characterization of epithelial ion transport, in Epithelial Transport: A Guide to Methods and Experimental Analysis. 1996, Chapman and Hall: London. p. 70-92. 170. Bachmann, O., et al., Different acid secretagogues activate different Na⁺/H⁺ exchanger isoforms in rabbit parietal cells. The American Journal of Physiology, 1998. 275(5 Pt 1): p. G1085-1093. 171. Masereel, B., L. Pochet, and D. Laeckmann, An overview of inhibitors of Na(+)/H(+) exchanger. Eur J Med Chem, 2003. 38(6): p. 547-54. 172. Schwark, J.R., et al., S3226, a novel inhibitor of Na⁺/H⁺ exchanger subtype 3 in various cell types. Pflügers Archiv: European Journal of Physiology, 1998. 436(5): p. 797-800. 173. Roos, A. and W.F. Boron, Intracellular pH. Physiological reviews, 1981. 61(2): p. 296-434. 174. Boron, W.F. and P. De Weer, Intracellular pH transients in squid giant axons caused by CO2, NH3, and metabolic inhibitors. J Gen Physiol, 1976. 67(1): p. 91- 112. 175. Bouyer, P., et al., Effect of extracellular acid-base disturbances on the intracellular pH of neurones cultured from rat medullary raphe or hippocampus. The Journal of Physiology, 2004. 559(Pt 1): p. 85-101. 176. Salameh, A.I., V.A. Ruffin, and W.F. Boron, Effects of metabolic acidosis on intracellular pH responses in multiple cell types. American Journal of Physiology. Regulatory, Integrative and Comparative Physiology, 2014. 307(12): p. R1413- 27.

189

177. Bevensee, M.O., C.J. Schwiening, and W.F. Boron, Use of BCECF and propidium iodide to assess membrane integrity of acutely isolated CA1 neurons from rat hippocampus. Journal of neuroscience methods, 1995. 58(1-2): p. 61-75. 178. Thomas, J.A., et al., Intracellular pH measurements in Ehrlich ascites tumor cells utilizing spectroscopic probes generated in situ. Biochemistry, 1979. 18(11): p. 2210-8. 179. Holm, S., A Simple Sequentially Rejective Multiple Test Procedure. Scandinavian Journal of Statistics, 1979. 6(2): p. 65-70. 180. Gelfond, D., et al., Intestinal pH and gastrointestinal transit profiles in cystic fibrosis patients measured by wireless motility capsule. Dig Dis Sci, 2013. 58(8): p. 2275-81. 181. Schreiber, R., et al., Anoctamins support calcium-dependent chloride secretion by facilitating calcium signaling in adult mouse intestine. Pflugers Arch, 2015. 467(6): p. 1203-13. 182. Lipecka, J., et al., Distribution of ClC-2 chloride channel in rat and human epithelial tissues. Am J Physiol Cell Physiol, 2002. 282(4): p. C805-16. 183. Murek, M., S. Kopic, and J. Geibel, Evidence for intestinal chloride secretion. Exp Physiol, 2010. 95(4): p. 471-8. 184. Cha, B., et al., cGMP inhibition of Na+/H+ antiporter 3 (NHE3) requires PDZ domain adapter NHERF2, a broad specificity protein kinase G-anchoring protein. J Biol Chem, 2005. 280(17): p. 16642-50. 185. Chen, T., et al., Cyclic GMP kinase II (cGKII) inhibits NHE3 by altering its trafficking and phosphorylating NHE3 at three required sites: identification of a multifunctional phosphorylation site. J Biol Chem, 2015. 290(4): p. 1952-65. 186. Magro, F., S. Fraga, and P. Soares-da-Silva, Signaling of short- and long-term regulation of intestinal epithelial type 1 Na+/H+ exchanger by interferon- gamma. Br J Pharmacol, 2005. 145(1): p. 93-103. 187. Quigley, E.M., et al., Randomised clinical trials: linaclotide phase 3 studies in IBS-C - a prespecified further analysis based on European Medicines Agency- specified endpoints. Aliment Pharmacol Ther, 2013. 37(1): p. 49-61. 188. Walker, N.M., et al., Talniflumate increases survival in a cystic fibrosis mouse model of distal intestinal obstructive syndrome. J Pharmacol Exp Ther, 2006. 317(1): p. 275-83. 189. Guerra, L., et al., Na+/H+ exchanger regulatory factor isoform 1 overexpression modulates cystic fibrosis transmembrane conductance regulator (CFTR) expression and activity in human airway 16HBE14o- cells and rescues DeltaF508 CFTR functional expression in cystic fibrosis cells. J Biol Chem, 2005. 280(49): p. 40925-33. 190. Gawenis, L.R., et al., Electroneutral sodium absorption and electrogenic anion secretion across murine small intestine are regulated in parallel. Am J Physiol Gastrointest Liver Physiol, 2004. 287(6): p. G1140-9. 191. Schultheis, P.J., et al., Renal and intestinal absorptive defects in mice lacking the NHE3 Na+/H+ exchanger. Nat Genet, 1998. 19(3): p. 282-5. 192. Lucas, M.L., A reconsideration of the evidence for Escherichia coli STa (heat stable) enterotoxin-driven fluid secretion: a new view of STa action and a new paradigm for fluid absorption. J Appl Microbiol, 2001. 90(1): p. 7-26.

190

193. Lucas, M.L., et al., Escherichia coli heat stable (STa) enterotoxin and the upper small intestine: lack of evidence in vivo for net fluid secretion. J Membr Biol, 2005. 206(1): p. 29-42. 194. Andresen, V., et al., Effect of 5 days linaclotide on transit and bowel function in females with constipation-predominant irritable bowel syndrome. Gastroenterology, 2007. 133(3): p. 761-8. 195. Whittamore, J.M. and M. Hatch, Loss of the anion exchanger DRA (Slc26a3), or PAT1 (Slc26a6), alters sulfate transport by the distal ileum and overall sulfate homeostasis. Am J Physiol Gastrointest Liver Physiol, 2017. 313(3): p. G166- G179. 196. Kelley, T.J., C.U. Cotton, and M.L. Drumm, In vivo activation of CFTR- dependent chloride transport in murine airway epithelium by CNP. Am J Physiol, 1997. 273(5 Pt 1): p. L1065-72. 197. Zhang, S., et al., Marked repression of CFTR mRNA in the transgenic Cftr(tm1kth) mouse model. J Cyst Fibros, 2014. 13(3): p. 351-2. 198. Ahsan, M.K., et al., Linaclotide activates guanylate cyclase-C/cGMP/protein kinase-II-dependent trafficking of CFTR in the intestine. Physiol Rep, 2017. 5(11). 199. Arora, K., et al., Guanylate cyclase 2C agonism corrects CFTR mutants. JCI Insight, 2017. 2(19). 200. Chinet, T.C., et al., Mechanism of sodium hyperabsorption in cultured cystic fibrosis nasal epithelium: a patch-clamp study. Am J Physiol, 1994. 266(4 Pt 1): p. C1061-8. 201. Hobbs, C.A., C. Da Tan, and R. Tarran, Does epithelial sodium channel hyperactivity contribute to cystic fibrosis lung disease? J Physiol, 2013. 591(18): p. 4377-87. 202. Matsui, H., et al., A physical linkage between cystic fibrosis airway surface dehydration and Pseudomonas aeruginosa biofilms. Proc Natl Acad Sci U S A, 2006. 103(48): p. 18131-6. 203. Davis, S.D. and F. Ratjen, Reduced lung function in cystic fibrosis: a primary or secondary phenotype? Am J Respir Crit Care Med, 2008. 178(1): p. 2-3. 204. Liou, T.G., et al., Survival effect of lung transplantation among patients with cystic fibrosis. Jama, 2001. 286(21): p. 2683-9. 205. Brierley, S.M., Guanylate cyclase-C receptor activation: unexpected biology. Curr Opin Pharmacol, 2012. 12(6): p. 632-40. 206. Qian, X., et al., Expression of GC-C, a receptor-guanylate cyclase, and its endogenous ligands uroguanylin and guanylin along the rostrocaudal axis of the intestine. Endocrinology, 2000. 141(9): p. 3210-24. 207. Kelley, T.J. and M.L. Drumm, Inducible nitric oxide synthase expression is reduced in cystic fibrosis murine and human airway epithelial cells. J Clin Invest, 1998. 102(6): p. 1200-7. 208. Guo, L.J., et al., ENaC is regulated by natriuretic peptide receptor-dependent cGMP signaling. Am J Physiol Renal Physiol, 2013. 304(7): p. F930-7. 209. Poschet, J.F., et al., Pharmacological modulation of cGMP levels by phosphodiesterase 5 inhibitors as a therapeutic strategy for treatment of

191

respiratory pathology in cystic fibrosis. Am J Physiol Lung Cell Mol Physiol, 2007. 293(3): p. L712-9. 210. Bookstein, C., et al., Na+/H+ exchangers, NHE-1 and NHE-3, of rat intestine. Expression and localization. J Clin Invest, 1994. 93(1): p. 106-13. 211. Livraghi, A. and S.H. Randell, Cystic fibrosis and other respiratory diseases of impaired mucus clearance. Toxicol Pathol, 2007. 35(1): p. 116-29. 212. Singh, V.K. and S.J. Schwarzenberg, Pancreatic insufficiency in Cystic Fibrosis. J Cyst Fibros, 2017. 16 Suppl 2: p. S70-S78. 213. Cleghorn, G.J., et al., Treatment of distal intestinal obstruction syndrome in cystic fibrosis with a balanced intestinal lavage solution. Lancet, 1986. 1(8471): p. 8-11. 214. Lai, H.C., et al., Comparison of growth status of patients with cystic fibrosis between the United States and Canada. Am J Clin Nutr, 1999. 69(3): p. 531-8. 215. Scaparrotta, A., et al., Growth failure in children with cystic fibrosis. J Pediatr Endocrinol Metab, 2012. 25(5-6): p. 393-405. 216. Kyle, U.G., L.S. Shekerdemian, and J.A. Coss-Bu, Growth failure and nutrition considerations in chronic childhood wasting diseases. Nutr Clin Pract, 2015. 30(2): p. 227-38. 217. Hasty, P., et al., Severe phenotype in mice with termination mutation in exon 2 of cystic fibrosis gene. Somat Cell Mol Genet, 1995. 21(3): p. 177-87. 218. O'Neal, W.K., et al., A severe phenotype in mice with a duplication of exon 3 in the cystic fibrosis locus. Hum Mol Genet, 1993. 2(10): p. 1561-9. 219. Rowntree, R.K. and A. Harris, The phenotypic consequences of CFTR mutations. Ann Hum Genet, 2003. 67(Pt 5): p. 471-85. 220. Ren, H.Y., et al., VX-809 corrects folding defects in cystic fibrosis transmembrane conductance regulator protein through action on membrane- spanning domain 1. Mol Biol Cell, 2013. 24(19): p. 3016-24. 221. Wainwright, C.E., et al., Lumacaftor-Ivacaftor in Patients with Cystic Fibrosis Homozygous for Phe508del CFTR. N Engl J Med, 2015. 373(3): p. 220-31. 222. Howard, M., R.A. Frizzell, and D.M. Bedwell, Aminoglycoside antibiotics restore CFTR function by overcoming premature stop mutations. Nat Med, 1996. 2(4): p. 467-9. 223. Xue, X., et al., Synthetic aminoglycosides efficiently suppress cystic fibrosis transmembrane conductance regulator nonsense mutations and are enhanced by ivacaftor. Am J Respir Cell Mol Biol, 2014. 50(4): p. 805-16. 224. Malik, V., et al., Gentamicin-induced readthrough of stop codons in Duchenne muscular dystrophy. Ann Neurol, 2010. 67(6): p. 771-80. 225. Sermet-Gaudelus, I., et al., Ataluren (PTC124) induces cystic fibrosis transmembrane conductance regulator protein expression and activity in children with nonsense mutation cystic fibrosis. Am J Respir Crit Care Med, 2010. 182(10): p. 1262-72. 226. Wilschanski, M., et al., Gentamicin-induced correction of CFTR function in patients with cystic fibrosis and CFTR stop mutations. N Engl J Med, 2003. 349(15): p. 1433-41. 227. McDonald, C.M., et al., Ataluren in patients with nonsense mutation Duchenne muscular dystrophy (ACT DMD): a multicentre, randomised, double-blind, placebo-controlled, phase 3 trial. Lancet, 2017. 390(10101): p. 1489-1498.

192

228. Kerem, E., et al., Ataluren for the treatment of nonsense-mutation cystic fibrosis: a randomised, double-blind, placebo-controlled phase 3 trial. Lancet Respir Med, 2014. 2(7): p. 539-47. 229. Hsu, P.D., et al., DNA targeting specificity of RNA-guided Cas9 nucleases. Nat Biotechnol, 2013. 31(9): p. 827-32. 230. Schmid-Burgk, J.L., et al., OutKnocker: a web tool for rapid and simple genotyping of designer nuclease edited cell lines. Genome Res, 2014. 24(10): p. 1719-23. 231. Brady, K.G., T.J. Kelley, and M.L. Drumm, Examining basal chloride transport using the nasal potential difference response in a murine model. Am J Physiol Lung Cell Mol Physiol, 2001. 281(5): p. L1173-9. 232. van Heeckeren, A.M., et al., Role of Cftr genotype in the response to chronic Pseudomonas aeruginosa lung infection in mice. Am J Physiol Lung Cell Mol Physiol, 2004. 287(5): p. L944-52. 233. Delaney, S.J., et al., Cystic fibrosis mice carrying the missense mutation G551D replicate human genotype-phenotype correlations. Embo J, 1996. 15(5): p. 955- 63. 234. Nudelman, I., et al., Repairing faulty genes by aminoglycosides: development of new derivatives of geneticin (G418) with enhanced suppression of diseases- causing nonsense mutations. Bioorg Med Chem, 2010. 18(11): p. 3735-46. 235. Heier, C.R. and C.J. DiDonato, Translational readthrough by the aminoglycoside geneticin (G418) modulates SMN stability in vitro and improves motor function in SMA mice in vivo. Hum Mol Genet, 2009. 18(7): p. 1310-22. 236. Perez, B., et al., Readthrough strategies for therapeutic suppression of nonsense mutations in inherited metabolic disease. Mol Syndromol, 2012. 3(5): p. 230-6. 237. Sermet-Gaudelus, I., et al., In vitro prediction of stop-codon suppression by intravenous gentamicin in patients with cystic fibrosis: a pilot study. BMC Med, 2007. 5: p. 5. 238. Baradaran-Heravi, A., et al., Gentamicin B1 is a minor gentamicin component with major nonsense mutation suppression activity. Proc Natl Acad Sci U S A, 2017. 114(13): p. 3479-3484. 239. McElroy, S.P., et al., A lack of premature termination codon read-through efficacy of PTC124 (Ataluren) in a diverse array of reporter assays. PLoS Biol, 2013. 11(6): p. e1001593. 240. Du, M., et al., PTC124 is an orally bioavailable compound that promotes suppression of the human CFTR-G542X nonsense allele in a CF mouse model. Proc Natl Acad Sci U S A, 2008. 105(6): p. 2064-9. 241. Kerem, E., et al., Effectiveness of PTC124 treatment of cystic fibrosis caused by nonsense mutations: a prospective phase II trial. Lancet, 2008. 372(9640): p. 719-27. 242. Benhabiles, H., et al., Optimized approach for the identification of highly efficient correctors of nonsense mutations in human diseases. PLoS One, 2017. 12(11): p. e0187930. 243. Gonzalez-Hilarion, S., et al., Rescue of nonsense mutations by amlexanox in human cells. Orphanet J Rare Dis, 2012. 7: p. 58.

193

244. Liang, F., et al., High-Throughput Screening for Readthrough Modulators of CFTR PTC Mutations. SLAS Technol, 2017. 22(3): p. 315-324. 245. Mutyam, V., et al., Discovery of Clinically Approved Agents That Promote Suppression of Cystic Fibrosis Transmembrane Conductance Regulator Nonsense Mutations. Am J Respir Crit Care Med, 2016. 194(9): p. 1092-1103. 246. Wang, H., et al., One-step generation of mice carrying mutations in multiple genes by CRISPR/Cas-mediated genome engineering. Cell, 2013. 153(4): p. 910- 8. 247. Fu, Y., et al., High-frequency off-target mutagenesis induced by CRISPR-Cas nucleases in human cells. Nat Biotechnol, 2013. 31(9): p. 822-6. 248. Iyer, V., et al., Off-target mutations are rare in Cas9-modified mice. Nat Methods, 2015. 12(6): p. 479. 249. Nakajima, K., et al., Exome sequencing in the knockin mice generated using the CRISPR/Cas system. Sci Rep, 2016. 6: p. 34703. 250. Hagemeijer, M.C., et al., Translational research to enable personalized treatment of cystic fibrosis. J Cyst Fibros, 2017. 251. Lykke-Andersen, S. and T.H. Jensen, Nonsense-mediated mRNA decay: an intricate machinery that shapes transcriptomes. Nat Rev Mol Cell Biol, 2015. 16(11): p. 665-77. 252. Gopalsamy, A., et al., Identification of pyrimidine derivatives as hSMG-1 inhibitors. Bioorg Med Chem Lett, 2012. 22(21): p. 6636-41. 253. Yu, H., et al., Comparison of read-through effects of aminoglycosides and PTC124 on rescuing nonsense mutations of HERG gene associated with long QT syndrome. Int J Mol Med, 2014. 33(3): p. 729-35. 254. Mort, M., et al., A meta-analysis of nonsense mutations causing human genetic disease. Hum Mutat, 2008. 29(8): p. 1037-47. 255. Drugeon, G., et al., Eukaryotic release factor 1 (eRF1) abolishes readthrough and competes with suppressor tRNAs at all three termination codons in messenger RNA. Nucleic Acids Res, 1997. 25(12): p. 2254-8. 256. Berger, A.H., A.G. Knudson, and P.P. Pandolfi, A continuum model for tumour suppression. Nature, 2011. 476(7359): p. 163-9. 257. Brogna, S. and J. Wen, Nonsense-mediated mRNA decay (NMD) mechanisms. Nat Struct Mol Biol, 2009. 16(2): p. 107-13. 258. Frischmeyer, P.A. and H.C. Dietz, Nonsense-mediated mRNA decay in health and disease. Hum Mol Genet, 1999. 8(10): p. 1893-900. 259. Finkel, R.S., Read-through strategies for suppression of nonsense mutations in Duchenne/ Becker muscular dystrophy: aminoglycosides and ataluren (PTC124). J Child Neurol, 2010. 25(9): p. 1158-64. 260. Hermann, T., Aminoglycoside antibiotics: old drugs and new therapeutic approaches. Cell Mol Life Sci, 2007. 64(14): p. 1841-52. 261. Howard, M.T., et al., Readthrough of dystrophin stop codon mutations induced by aminoglycosides. Ann Neurol, 2004. 55(3): p. 422-6. 262. Manuvakhova, M., K. Keeling, and D.M. Bedwell, Aminoglycoside antibiotics mediate context-dependent suppression of termination codons in a mammalian translation system. Rna, 2000. 6(7): p. 1044-55.

194

263. Hofhuis, J., et al., The functional readthrough extension of malate dehydrogenase reveals a modification of the genetic code. Open Biol, 2016. 6(11). 264. Martin, L., et al., Identification and characterization of small molecules that inhibit nonsense-mediated RNA decay and suppress nonsense p53 mutations. Cancer Res, 2014. 74(11): p. 3104-13. 265. Corvol, H., et al., Genome-wide association meta-analysis identifies five modifier loci of lung disease severity in cystic fibrosis. Nat Commun, 2015. 6: p. 8382. 266. Pettit, R.S. and C. Fellner, CFTR Modulators for the Treatment of Cystic Fibrosis. P t, 2014. 39(7): p. 500-11. 267. Dekkers, J.F., et al., Characterizing responses to CFTR-modulating drugs using rectal organoids derived from subjects with cystic fibrosis. Sci Transl Med, 2016. 8(344): p. 344ra84. 268. Clancy, J.P., et al., Evidence that systemic gentamicin suppresses premature stop mutations in patients with cystic fibrosis. Am J Respir Crit Care Med, 2001. 163(7): p. 1683-92. 269. Du, M., et al., Clinical doses of amikacin provide more effective suppression of the human CFTR-G542X stop mutation than gentamicin in a transgenic CF mouse model. J Mol Med (Berl), 2006. 84(7): p. 573-82. 270. Beznoskova, P., S. Gunisova, and L.S. Valasek, Rules of UGA-N decoding by near-cognate tRNAs and analysis of readthrough on short uORFs in yeast. Rna, 2016. 22(3): p. 456-66. 271. Loughran, G., et al., Stop codon readthrough generates a C-terminally extended variant of the human vitamin D receptor with reduced calcitriol response. J Biol Chem, 2018. 293(12): p. 4434-4444. 272. Auld, D.S., et al., Mechanism of PTC124 activity in cell-based luciferase assays of nonsense codon suppression. Proc Natl Acad Sci U S A, 2009. 106(9): p. 3585- 90. 273. Xue, X., et al., Identification of the Amino Acids Inserted During Suppression of CFTR Nonsense Mutations and Determination of Their Functional Consequences. Hum Mol Genet, 2017. 274. Azzalin, C.M. and J. Lingner, The human RNA surveillance factor UPF1 is required for S phase progression and genome stability. Curr Biol, 2006. 16(4): p. 433-9. 275. Gewandter, J.S., R.A. Bambara, and M.A. O'Reilly, The RNA surveillance protein SMG1 activates p53 in response to DNA double-strand breaks but not exogenously oxidized mRNA. Cell Cycle, 2011. 10(15): p. 2561-7. 276. Bidou, L., et al., Characterization of new-generation aminoglycoside promoting premature termination codon readthrough in cancer cells. RNA biology, 2017. 14(3): p. 378-388. 277. Valley, H.C., et al., Isogenic cell models of cystic fibrosis-causing variants in natively expressing pulmonary epithelial cells. J Cyst Fibros, 2018. 278. Sermet-Gaudelus, I., et al., Antisense oligonucleotide eluforsen improves CFTR function in F508del cystic fibrosis. J Cyst Fibros, 2018. 279. Liu, X., et al., ROCK inhibitor and feeder cells induce the conditional reprogramming of epithelial cells. Am J Pathol, 2012. 180(2): p. 599-607.

195

280. Gentzsch, M., et al., Pharmacological Rescue of Conditionally Reprogrammed Cystic Fibrosis Bronchial Epithelial Cells. Am J Respir Cell Mol Biol, 2017. 56(5): p. 568-574. 281. Sharma, N., et al., Capitalizing on the heterogeneous effects of CFTR nonsense and frameshift variants to inform therapeutic strategy for cystic fibrosis. PLoS Genet, 2018. 14(11): p. e1007723. 282. Future directions for research on enterotoxigenic Escherichia coli vaccines for developing countries. Wkly Epidemiol Rec, 2006. 81(11): p. 97-104. 283. Schulz, S., et al., Guanylyl cyclase is a heat-stable enterotoxin receptor. Cell, 1990. 63(5): p. 941-8. 284. Hamra, F.K., et al., Uroguanylin: structure and activity of a second endogenous peptide that stimulates intestinal guanylate cyclase. Proc Natl Acad Sci U S A, 1993. 90(22): p. 10464-8. 285. Currie, M.G., et al., Guanylin: an endogenous activator of intestinal guanylate cyclase. Proc Natl Acad Sci U S A, 1992. 89(3): p. 947-51. 286. Kita, T., et al., Characterization of human uroguanylin: a member of the guanylin peptide family. Am J Physiol, 1994. 266(2 Pt 2): p. F342-8. 287. Yun, C.H., et al., cAMP-mediated inhibition of the epithelial brush border Na+/H+ exchanger, NHE3, requires an associated regulatory protein. Proc Natl Acad Sci U S A, 1997. 94(7): p. 3010-5. 288. Sun, F., et al., E3KARP mediates the association of ezrin and protein kinase A with the cystic fibrosis transmembrane conductance regulator in airway cells. J Biol Chem, 2000. 275(38): p. 29539-46. 289. Lamprecht, G., E.J. Weinman, and C.H. Yun, The role of NHERF and E3KARP in the cAMP-mediated inhibition of NHE3. J Biol Chem, 1998. 273(45): p. 29972-8. 290. Ahn, W., et al., Regulatory interaction between the cystic fibrosis transmembrane conductance regulator and HCO3- salvage mechanisms in model systems and the mouse pancreatic duct. J Biol Chem, 2001. 276(20): p. 17236-43. 291. Boucher, R.C., New concepts of the pathogenesis of cystic fibrosis lung disease. Eur Respir J, 2004. 23(1): p. 146-58. 292. Jiang, Q., et al., Epithelial sodium channels regulate cystic fibrosis transmembrane conductance regulator chloride channels in Xenopus oocytes. J Biol Chem, 2000. 275(18): p. 13266-74. 293. Thibodeau, P.H. and M.B. Butterworth, Proteases, cystic fibrosis and the epithelial sodium channel (ENaC). Cell Tissue Res, 2013. 351(2): p. 309-23. 294. Bagorda, A., et al., Reciprocal protein kinase A regulatory interactions between cystic fibrosis transmembrane conductance regulator and Na+/H+ exchanger isoform 3 in a renal polarized epithelial cell model. J Biol Chem, 2002. 277(24): p. 21480-8. 295. Furukawa, O., et al., NHE3 inhibition activates duodenal bicarbonate secretion in the rat. Am J Physiol Gastrointest Liver Physiol, 2004. 286(1): p. G102-9. 296. Macdougall, J.E., et al., An evaluation of the FDA responder endpoint for IBS-C clinical trials: analysis of data from linaclotide Phase 3 clinical trials. Neurogastroenterol Motil, 2013. 25(6): p. 481-6.

196

297. Castro, J., et al., Linaclotide inhibits colonic nociceptors and relieves abdominal pain via guanylate cyclase-C and extracellular cyclic guanosine 3',5'- monophosphate. Gastroenterology, 2013. 145(6): p. 1334-46.e1-11. 298. Eutamene, H., et al., Guanylate cyclase C-mediated antinociceptive effects of linaclotide in rodent models of visceral pain. Neurogastroenterol Motil, 2010. 22(3): p. 312-e84. 299. Norkina, O., et al., Inflammation of the cystic fibrosis mouse small intestine. Am J Physiol Gastrointest Liver Physiol, 2004. 286(6): p. G1032-41. 300. Mijnders, M., B. Kleizen, and I. Braakman, Correcting CFTR folding defects by small-molecule correctors to cure cystic fibrosis. Curr Opin Pharmacol, 2017. 34: p. 83-90. 301. Schmidt, B.Z., et al., Cystic fibrosis transmembrane conductance regulator modulators in cystic fibrosis: current perspectives. Clinical pharmacology : advances and applications, 2016. 8: p. 127-140. 302. Zabel, U., et al., Human soluble guanylate cyclase: functional expression and revised isoenzyme family. The Biochemical journal, 1998. 335 ( Pt 1)(Pt 1): p. 51- 57. 303. Lin, C.S., Tissue expression, distribution, and regulation of PDE5. Int J Impot Res, 2004. 16 Suppl 1: p. S8-s10. 304. Lacy, B.E., et al., Linaclotide in Chronic Idiopathic Constipation Patients with Moderate to Severe Abdominal Bloating: A Randomized, Controlled Trial. PLoS One, 2015. 10(7): p. e0134349. 305. Yang, Y., et al., Linaclotide in irritable bowel syndrome with constipation: A Phase 3 randomized trial in China and other regions. J Gastroenterol Hepatol, 2018. 33(5): p. 980-989. 306. Camilleri, M., et al., Comparison of adequate relief with symptom, global, and responder endpoints in linaclotide phase 3 trials in IBS-C. United European Gastroenterol J, 2015. 3(1): p. 53-62. 307. Nudelman, I., et al., Redesign of aminoglycosides for treatment of human genetic diseases caused by premature stop mutations. Bioorg Med Chem Lett, 2006. 16(24): p. 6310-5. 308. Nudelman, I., et al., Development of novel aminoglycoside (NB54) with reduced toxicity and enhanced suppression of disease-causing premature stop mutations. Journal of medicinal chemistry, 2009. 52(9): p. 2836-2845. 309. Kandasamy, J., et al., Increased selectivity toward cytoplasmic versus mitochondrial ribosome confers improved efficiency of synthetic aminoglycosides in fixing damaged genes: a strategy for treatment of genetic diseases caused by nonsense mutations. J Med Chem, 2012. 55(23): p. 10630-43. 310. Unlocking protein production with translational read-through for rare genetic diseases, in Investor Presentation February 2018, E. Pharmaceuticals, Editor. 2018. 311. Dabrowski, M., Z. Bukowy-Bieryllo, and E. Zietkiewicz, Translational readthrough potential of natural termination codons in eucaryotes--The impact of RNA sequence. RNA biology, 2015. 12(9): p. 950-958.

197

312. Dabrowski, M., Z. Bukowy-Bieryllo, and E. Zietkiewicz, Translational readthrough potential of natural termination codons in eucaryotes--The impact of RNA sequence. RNA Biol, 2015. 12(9): p. 950-8. 313. Beznoskova, P., et al., Translation initiation factor eIF3 promotes programmed stop codon readthrough. Nucleic Acids Res, 2015. 43(10): p. 5099-111. 314. Sato, T., et al., Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature, 2009. 459(7244): p. 262-5. 315. Nadkarni, R.R., S. Abed, and J.S. Draper, Organoids as a model system for studying human lung development and disease. Biochem Biophys Res Commun, 2016. 473(3): p. 675-82. 316. Rock, J.R., et al., Basal cells as stem cells of the mouse trachea and human airway epithelium. Proc Natl Acad Sci U S A, 2009. 106(31): p. 12771-5. 317. Boj, S.F., et al., Organoid models of human and mouse ductal pancreatic cancer. Cell, 2015. 160(1-2): p. 324-38. 318. Drost, J., et al., Organoid culture systems for prostate epithelial and cancer tissue. Nat Protoc, 2016. 11(2): p. 347-58. 319. Quadrato, G., et al., Cell diversity and network dynamics in photosensitive human brain organoids. Nature, 2017. 545(7652): p. 48-53. 320. Nguyen, A.D., et al., Murine knockin model for progranulin-deficient frontotemporal dementia with nonsense-mediated mRNA decay. Proc Natl Acad Sci U S A, 2018. 115(12): p. E2849-e2858. 321. Mooney, C.M., et al., RNA sequencing of synaptic and cytoplasmic Upf1-bound transcripts supports contribution of nonsense-mediated decay to epileptogenesis. Sci Rep, 2017. 7: p. 41517. 322. Atanasova, V.S., et al., Amlexanox Enhances Premature Termination Codon Read-Through in COL7A1 and Expression of Full Length Type VII Collagen: Potential Therapy for Recessive Dystrophic Epidermolysis Bullosa. J Invest Dermatol, 2017. 137(9): p. 1842-1849. 323. Banning, A., M. Schiff, and R. Tikkanen, Amlexanox provides a potential therapy for nonsense mutations in the lysosomal storage disorder Aspartylglucosaminuria. Biochim Biophys Acta Mol Basis Dis, 2018. 1864(3): p. 668-675. 324. Zhang, M., et al., Synergistic Rescue of Nonsense Mutant Tumor Suppressor p53 by Combination Treatment with Aminoglycosides and Mdm2 Inhibitors. Frontiers in oncology, 2018. 7: p. 323-323. 325. Makarov, E.M., et al., The rare nonsense mutation in p53 triggers alternative splicing to produce a protein capable of inducing apoptosis. PLoS One, 2017. 12(9): p. e0185126. 326. Baker, S.J., et al., p53 gene mutations occur in combination with 17p allelic deletions as late events in colorectal tumorigenesis. Cancer Res, 1990. 50(23): p. 7717-22. 327. Kastenhuber, E.R. and S.W. Lowe, Putting p53 in Context. Cell, 2017. 170(6): p. 1062-1078. 328. Shirole, N.H., et al., TP53 exon-6 truncating mutations produce separation of function isoforms with pro-tumorigenic functions. Elife, 2016. 5.

198