The novel aminoglycoside, ELX-02, permits CTNSW138X translational read-through and restores lysosomal cystine efflux in cystinosis

Background Cystinosis is a rare disorder caused by recessive mutations of the CTNS gene. Current therapy decreases cystine accumulation, thus slowing organ deterioration without reversing renal Fanconi syndrome or preventing eventual need for a kidney transplant.15-20% of cystinosis patients harbour at least one nonsense mutation in CTNS, leading to premature end of translation of the transcript. Aminoglycosides have been shown to permit translational read-through but have high toxicity level, especially in the kidney and inner ear. ELX-02, a modified aminoglycoside, retains it read-through ability without the toxicity. Methods and findings We ascertained the toxicity of ELX-02 in cells and in mice as well as the effect of ELX-02 on translational read-through of nonsense mutations in cystinotic mice and human cells. ELX-02 was not toxic in vitro or in vivo, and permitted read-through of nonsense mutations in cystinotic mice and human cells. Conclusions ELX-02 has translational read-through activity and produces a functional CTNS protein, as evidenced by reduced cystine accumulation. This reduction is comparable to cysteamine treatment. ELX-02 accumulates in the kidney but neither cytotoxicity nor nephrotoxicity was observed.


Introduction
Cystinosis is a rare autosomal recessive disorder caused by mutations of the CTNS gene, encoding a cystine-specific transporter (cystinosin) that facilitates cystine efflux through the lysosomal membrane [1]. Homozygous CTNS mutations cause intralysosomal cystine accumulation, disturbance of cellular homeostatic mechanisms and progressive organ dysfunction. At birth, babies with biallelic null CTNS mutations appear normal. However, between 4-6 months of age, they develop severe proximal tubular dysfunction (renal Fanconi Syndrome) and failure to thrive. By 12-14 months of age, microdissection studies show atresia of the proximal tubule (swan-neck lesion). Between 5-10 years of age, there is progressive glomerular podocyte injury, proteinuria and renal insufficiency; dialysis is required by 10-12 years of age [2,3]. Although kidney transplantation resolves renal insufficiency, inexorable deterioration of other organs slowly leads to hypothyroidism, diabetes, brain dysfunction and profound muscle wasting that compromises respiration and swallowing. Without further treatment, life expectancy is less than 30 years.
In the 1970's, Thoene et al. discovered that pathologic intralysosomal accumulation of cystine in mutant cells could be dramatically decreased by the reducing agent, cysteamine [4]. Oral cysteamine (1.3g/m 2 /day) reduces leukocyte cystine content to about 15-20% of baseline and, when therapy is initiated by 24 months, the drug slows deterioration of the kidneys and other organs [5,6]. However, while oral cysteamine therapy has substantially altered the natural history of cystinosis, it does not reverse the renal Fanconi syndrome and most patients eventually require renal transplantation. Muscle wasting remains a significant complication during the third decade and longevity remains compromised. In part, the suboptimal benefit of cysteamine has been attributed to its poor tolerability and there is hope that the advent of a delayed release form will be helpful in older children and adults. On the other hand, CTNS protein (cystinosin) may have cellular functions beyond its role as a cystine-selective channel in the lysosomal membrane. CTNS mutant proximal tubular cells exhibit defective megalindependent endocytosis of luminal proteins in vitro and this is refractory to cysteamine treatment [7]. Furthermore, a CTNS isoform (CTNS LKG ) constituting 10-15% of total cell cystinosin is expressed at the luminal membrane suggesting an alternative function [8]. Other studies show that the CTNS protein is required for vesicular traffic involved in autophagy [9]. Thus, successful therapy of cystinosis may require a complementary strategy to offset the defect in these non-channel functions.
Worldwide, the most common cystinosis mutation is a 57 Kb deletion that eliminates the first 10 exons of the gene and is thought to have arisen in Northern Europe in about 500 AD [10]. In Canada, however, the most common mutant allele is a nonsense mutation (CTNS W138X ), introduced from Ireland into the French Canadian population [11]. The W138X mutation creates a premature STOP codon in the seventh exon of the CTNS gene and accounts for 40-50% of cystinosis alleles in Quebec [11]. Interestingly, Heier and DiDonato reported in 2009 that premature STOP codons can be overcome by aminoglycoside antibiotics (eg. geneticin) [12]. These drugs bind to the mammalian ribosome, relax translational fidelity and allow read-through of premature STOP codons which would otherwise bring translation to a halt and induce transcriptional decay [13,14]. Unfortunately, this observation has not been translated into a useful clinical therapy because of the inherent renal and cochlear toxicity of aminoglycosides in humans. It is conceivable that cochlear toxicity is linked to an interaction between aminoglycosides and the mitochondrial ribosome which is homologous to the ribosome of prokaryotes. Thus, the antibacterial effects of aminoglycosides are accompanied by disturbance of mitochondrial protein synthesis in humans. Individuals with specific therapies for cystinosis), Genomic Applications Partnership Program -Genome Canada (2571: Novel aminoglycoside read-through therapy for nonsense mutations) and an unrestricted research grant (7597: Effect of Eloxx Compounds on CTNS Activity in Cell and Animal Models) from Eloxx Pharmaceutical Inc. EB was the recipient of a graduate studentship award from the Cystinosis Research Foundation. PG is the recipient of a James McGill Research Chair. Eloxx Pharmaceutical scientists (IEO and IA) performed the drug accumulation assay, IEO, IA and PH contributed to experimental design and manuscript preparation. The funder provided support in the form of salaries for authors IEO, IA and PH, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the 'author contributions' section.
Competing interests: Dr. Goodyer was the recipient of an unrestricted grant from Eloxx Pharmaceuticals, Inc. for parts of this work. RIMUHC-affiliated authors were not Eloxx employees nor did they have any commercial affiliation or patent interests related to ELX-02. Idit Eshkar-Oren, Iris Alroy and Pedro Huertas were employees of Eloxx Pharmaceuticals, Inc. This does not alter our adherence to PLOS ONE policies on sharing data and materials. mitochondrial genetic variants that alter the mitochondrial ribosome are particularly susceptible to aminoglycoside-induced renal injury and cochlear dysfunction [14].
In this study, we explore the potential of ELX-02 to serve as a novel therapy for cystinosis caused by CTNS nonsense mutations. In fibroblasts from cystinosis patients, we report that ELX-02 permits translational read-through of the CTNS W138X nonsense mutation without overt cellular toxicity. Furthermore, we show that CTNS protein expression (and its attendant correction of nonsense mutation-mediated transcript decay) is sufficient to reverse pathologic intralysosomal cystine accumulation. In a novel Ctns Y226X nonsense mutant mouse, we demonstrate that subcutaneous ELX-02 accumulates in kidney tissue without overt renal toxicity and that ELX-02 (10mg/kg X2/week for 3 weeks) reduces renal cystine accumulation in vivo.

Cell culture
The human proximal tubule cell line HK-2 (American Type Culture Collection, USA) was maintained as adherent monolayer cultures in DMEM-F12 culture medium (Life Technologies, USA), supplemented with 10% fetal bovine serum (Life Technologies, USA), 1% glutamine and Penicillin-Streptomycin (Life Technologies, USA), 1x Insulin-Transferrin-Selenium (Corning, USA), 36 ng/mL hydrocortisone(Sigma, USA), 10 ng/mL epithelial growth factor (Tonbo Biosciences, USA), and 40 pg/mL tri-iodothyronine (Sigma, USA). HEK293 (American Type Culture Collection, USA) cell line was maintained in DMEM (Life Technologies, USA) supplemented with 10% FBS, 1% glutamine and 1% Penicillin-Streptomycin. Normal fibroblast cell line MCH070 and cystinotic fibroblasts cell lines WG0881, WG1012 and WG2379 were obtained from the Hereditary Renal Disease Cell Repository from the RI-MUHC (IRB protocol # 2018-2922) and maintained as monolayer cultures growing in DMEM (Corning, USA) supplemented with 15% fetal bovine serum. Cells were incubated at 37˚C in an atmosphere of 5% CO 2 . Experiments were performed when cells reached 70-80% confluence. Trypsinization was performed using 0.25% trypsin, 2.21 mM EDTA (Corning, USA), cells were washed 3 times with PBS (Life Technologies, USA) then pellets snap frozen in an ethanol bath. Samples were store at -80˚C until preparation for mRNA, protein or half-cystine analysis.

Cytotoxicity assay
A Cell Counting Kit-8 was obtained from Enzo Life Sciences (Farmingdale, NY). Culture media, fetal bovine serum, and antibiotics were purchased from Invitrogen (Carlsbad, CA). Human Proximal tubule cells (HK-2), Wild type human fibroblast cells (MCH070) and human Fibroblast with W138X homozygous mutation (WG1012) were cultured in DMEM supplemented with 10% fetal bovine serum (FBS). Cells were trypsinized and re-suspended in fresh media at a concentration of 300,000 cells/well in 24-well plates. Cells were allowed to grow overnight at 37˚C in a humidified 5% CO2 environment. G418 and ELX-02 were added to final concentration 0, 100 and 400 μg/mL. Following a 48-72 hours exposure to the drug, 50 μL of the cell counting reagent was added directly to the cell cultures in 24-well plates that have 500 μL of media. Cells were incubated in a humidified environment at 37˚C, 5% CO2 for 1 h. After incubation, the absorbance at 450 nm was determined using Epoch Microplate reader (BioTek Instruments). The following equation was used to calculate the percentage of cell viability: cell viability (%) = (OD sample (mean) )/ (OD control (mean) ) x 100, where OD is the optical density; OD of sample: mean absorbance of treated cells and OD of control: mean absorbance of control cells. To determine the percentage cytotoxicity, calculate the average absorbance of the triplicate samples and controls, the values for each concentration were converted into percentage of control (interpreted as 100% viability).

Transfection and ELX-02 treatment for read-through assay
The transfection of the pcDNA3.1-CTNS-DsRed, pcDNA3.1-CTNS W138X -DsRed, and pcDNA3.1-kozak-CTNS W138X -His plasmids has been previously described [17]. In experiments where an aminoglycoside was used to induce ribosomal read-through, 0-400 μg/mL of ELX-02 (Eloxx Pharmaceuticals, Israel) or G418 (Life Technologies, USA) were added to the culture flasks after 48 h of transfection. Aminoglycosides were dissolved directly at its final concentration in DMEM. Control experiments were performed using cells exposed to the identical media without aminoglycosides. Then the cells were cultured in DMEM media containing 0-400 μg/mL of aminoglycosides for 48 h. Cell pellets were washed and assayed for RNA, protein and half-cystine content as previously described [17]. For the read-through assay in cystinotic fibroblast cells, triplicates of each fibroblast cell lines were seeded in T25 culture flask and grown to confluency. Then the cells were washed with 1x PBS three times and cultured in DMEM media containing 20-400 μg/mL aminoglycosides and/ or 50 μM cysteamine for 24-72 h. Cell pellets were washed and assayed for RNA, protein and half-cystine content as previously described [17].

Confocal microscopy
After aminoglycoside treatment, HEK293 cells transfected with pcDNA3.1-CTNS-DsRed, pcDNA3.1-CTNS W138X -DsRed were seeded in 4-well-chambered coverglass (Nunc) and fixed in 4% paraformaldehyde for 10 min at RT. After washing two times with PBS, the cells were quenched with 20mM ammonia chloride for 10 minutes and wash with PBS. Then the cells were blocked (PBS, 5% normal goat serum, 0.1% Triton X-100, 1% BSA) for 1 h, at RT. After triple washing with PBS, the cells were stained with DAPI and mounted onto slides with Pro-Long Gold antifade reagent (Life Technologies). The cells were observed using a Zeiss LSM780 laser scanning confocal microscope.

Western immunoblotting
Urine was collected from mice in opaque microfuge tubes, centrifuge at 15 000 rpm for 10 min at 4˚C and stored at -80˚C until ready to use. 30 μL of urine was assayed via SDS-PAGE followed by standard immunoblotting method and analysis. Primary antibody overnight incubation with anti-Transferrin (GeneTex, USA), anti-RBP4 (Adipogen, USA) and anti-B2M (GeneTex, USA) at 1:1000 ratio was followed by secondary antibody incubation for 1h with αrabbit IgG-HRP (Life Technologies, USA) or α-mouse IgG-HRP (Life Technologies, USA). Bands were visualized using ECL 2 Western Blotting Substrate (Thermo Scientific Pierce, USA).

RNA extraction and real-time PCR
Total RNA extraction and real-time PCR has been previously described [17].

Intracellular half-cystine measurement
This assay is a derivative of that of Gilfix et al. [18]. Stock solutions of homocystine and cystine were prepared assuming a assuming a molar absorption coefficient of 397 at 244 nm. This is the molar absorption coefficient of the disulfide bond (personal communication from D. Jacobsen, Cleveland Clinic Foundation). Stock solutions of cysteinylglycine were prepared on the basis of weight. The linear range for this assay is: 2-300 μmol/L L-homocysteine, 3-600 μmol/L L-cysteine, and 10-100 μmol/L cysteinylglycine. The within-run CV for cysteine at 3 μmol/L was 1.6% (n = 4). Prior to analysis, samples were supplemented to 30 μmol/L Lhomocysteine as a second internal standard to normalize the cystine values. This was done as in some cases cystamine had been added to the cell cultures and this interfered with the measurement of the usual internal standard, cystamine. An aqueous solution of 10% tris(carboxyethyl) phosphine•HCl (Pierce Chemical Co.) (9 μL) was mixed with 5 μL of 100 μmol/L cystamine. This was added to 90 μL of the sample and mixed immediately. Following incubation for 30 min at room temperature, 90 μL of 10%TCA-1 mmol/L EDTA was added. The sample was centrifuged for 10 min at 15,600 x g and the supernatant recovered. The supernatant (100 μL) was then added to a separate tube containing 20 μL of 1.55 mol/L NaOH and 250 μL of 0.125 mol/L borate buffer-4 mmol/L EDTA, pH 9.5 and mixed immediately. To this, 100 μL ammonium 7-fluorobenzo-2-oxa-1,3-diazole-4-sulfonate (Wako Chemicals USA, Inc.) (1 mg/mL in 0.125 mol/Lborate buffer, pH 9.5) was then added and mixed immediately. The mixture was incubated for 1 h at 60˚C. An aliquot (50 μL) was loaded onto a Hypersil 5 μ C18 ODS (250 x 4.6 mm) column run at 1.5 mL/min. The run buffer was 0.05 mol/L KH2PO4, pH 2.1 containing 4% acetonitrile and the fluorescence monitored (λ excitation = 385 nm; λ emission = 515 nm). The results representing the sum of cysteine, cystine and other mixed form of cysteine in the sample expressed as half-cystine (cysteine). Total protein was measured using the BCA Protein Assay kit (Biovision), as per manufacturer's instructions (96-well plate format). Half-cystine results were corrected to total protein. To estimate pathologic intralysosomal cystine, we subtracted basal cell cystine (15.95 ± 1.44 nmol half-cystine/mg protein). Data was processed to give arithmetic mean for half-cystine levels.

Non-radioactive pulse-chase assay
The Click-iT AHA (L-azidohomoalaine) for newly synthesized protein synthesis kit (Life Technologies, USA) was used for these studies and experiments were performed according to manufacturer's recommendation. HK-2 cells were seeded in T75 flask to reach 80% confluent. Methionine in cells was depleted by culturing in methionine-free DMEM for 1 h. Then 50 μM of AHA (a methionine analog) was added into medium for 1 h. Cells were washed and then incubated with the complete medium. The cells were chased for indicated time. Subsequently, Harvested cells at different time points were lysed and AHA incorporated protein was biotinylated using Click-IT-DIBO Biotin (Life Technologies, USA). Proteins were then immunoprecipitated with Pierce High Capacity Streptavidin slurry agarose (Life Technologies, USA) and further analyzed by Western immunoblotting.

Transgenic mice
All animal experiments were approved by the McGill University Facility Animal Care Committee (FACC). Wildtype C57Bl6 mice were purchased from Charles River Laboratories. A Zinc-finger nuclease (ZFN) pair targeting CTNS exon 8 was identified. We designed a template with homology arms, a premature STOP codon replacing amino acid 226 and a silent XbaI restriction site. The ZFNs and the template were injected into the pronuclei of zygotes that were transferred to pseudo-pregnant females. Two pregnancies were brought to term. Viable F0 pups with a heterozygous CTNS Y226X allele were crossed to CD1 wildtypes to yield F1 progeny for the same CTNS Y226X mutation. Heterozygous F1 littermates were interbred to produce multiple litters of F2 animals, in which germline transmission of CTNS Y226X was confirmed by Sanger sequencing. F2 CTNS Y226X heterozygotes were interbred to generate 4 litters of F3 homozygous CTNS Y226X/Y226X mice. Genotyping was performed at weaning, using ear punch samples taken under aseptic conditions. For subsequent experiments, the animals were sacrificed using CO 2 asphyxiation under isoflurane anesthesia, followed by pneumothorax, according to standard operating procedures. Material acquired for genotyping was done according to standard operating procedures by ear punching. Mice were housed in groups, in solid-bottom cages with contact bedding and nesting material. Animals were monitored daily by a veterinarian, checking for excessive grooming, fighting, wounds, lethargy, activity and respiratory rate. Cages were changed daily, and animals had freely available food and water. Sick or distressed animals were euthanized and excluded from the study. (FACC protocol # 2011-6026).
the following time points: 0.25, 0.5, 1, 2, 4 and 8 h post-dose. Kidneys were collected at the following time points: 0.5, 2, 4 and 8 h post-dose (1 animal/sex for each time point). In Group 2 sixteen (n = 16) Ctns Y226X/Y226X mice were dosed twice weekly with either ELX-02 (10 mg/kg/ dose, free base) or saline for 21 days (total of 7 administrations). Blood was collected after the last dose at the following time points: pre-dose, 0.25, 0.5, 1, 2, 4 and 8 h post-dose. Kidneys were collected at the following time points: pre-dose, 0.5, 2, 4 and 8 h post-dose (1 animal/sex for each timepoint). The collected blood and kidneys were used for quantification of ELX-02 levels, and quantification of half-cystine levels in kidney.

Plasma and tissue processing
After the animals were anaesthetized, blood samples were collected via cardiac puncture into tubes containing anticoagulant (K 3 EDTA), placed in crushed wet-ice and centrifuged (2000 g at 4˚C for 10 minutes). The resultant plasma was separated from the erythrocyte pellet and then transferred to uniquely labelled clear polypropylene tubes and frozen immediately at -80˚C until analysis.
Brain, spleen, heart, kidney, liver, lungs, and cochlea tissues were dissected as follows: once the animal was deeply anaesthetized, the skin was removed, the thorax opened, and the heart exposed. A 5 mL syringe with a 25G needle was inserted into the left ventricle, saline infused to flush any blood out of the organs at a flow rate 10 mL/min, infusion time 0.5 min, and infusion volume 5 mL. An infusion pump, containing saline, was switched on to prime the tubing and needle. The needle, with saline flowing, was inserted into the left ventricle, and saline pumped at flow rate 6.5-8 mL/min to flush any blood out of the tissues. A small incision was made into the right atrium, to allow any blood and /or saline to be released. Once the liquid coming out of the right ventricle was clear, the needle was removed, and the infusion pump switched off. The tissues were then removed and frozen at -80˚C until analysis.

Quantification of ELX-02 in plasma and tissue
Plasma samples were spiked with ELX-02 internal standard (ELX-02-13C, d3), processed by protein precipitation extraction, and analyzed by HPLC-MS/MS using a TurboIonspray interface with positive ion multiple reaction monitoring. The method was qualified over the range 10 to 10000 ng/mL with a lower limit of quantification of 10 ng/mL using a 25μL sample. ELX-02 concentrations in plasma were expressed as ng/mL. ELX-02 was extracted from mouse tissue' homogenate by solid phase extraction after addition of (ELX-02-13C, d3), and analyzed by LC-MS/MS using a TurboIonspray interface with positive ion multiple reaction monitoring. The method was qualified over the range 5 to 5000 ng/mL with a lower limit of quantification of 5 ng/mL using a 50 μL aliquot of mouse tissue homogenate. ELX-02 concentrations in tissue samples were expressed as ng/g.

PK analysis
Plasma and kidney ELX-02 concentration-time data were analysed by non-compartmental analysis using Phoenix WinNonlin. The following parameters for ELX-02 were determined where feasible, using the linear-logarithmic trapezoidal rule:

Statistical analysis
One-way ANOVA followed by multiple comparison tests or t-tests were used for statistical analysis.

Ethics compliance
All primary human cell lines used in this article were subject to IRB approval from the RI-MUHC REB (protocol: 2018-2922). All animal research in this article was subject to Facility Animal Care Committee approval from the RI-MUHC (protocol: 2001-6026).

Absence of cytotoxicity in human fibroblasts
To ascertain whether ELX-02 is toxic for human fibroblasts in vitro, we exposed monolayers of wildtype or homozygous CTNS W138X/W138X mutant fibroblasts to ELX-02 (100, 400 μg/mL) or G418 (100 μg/mL) for 48 hours. Cell viability was assessed by LDH release into the culture medium and expressed as a percent of untreated controls. As seen in Fig 2A and 2B, cell viability was significantly reduced (to 76% of control, p<0.05) by G418 in both wildtype and mutant fibroblasts. In contrast, no significant toxicity was seen in the presence of ELX-02 at 100 or 400 μg/mL. Since aminoglycosides are concentrated by renal proximal tubular cells and exhibit clinical nephrotoxicity at high doses, we also assessed the toxicity of ELX-02 in a human proximal tubular cell line (HK-2). Whereas G418 (100 μg/mL) reduced HK-2 cell viability to 83% of untreated controls (Fig 2C), no reduction in cell viability was noted in the presence of ELX-02 (400 μg/mL) over 72 h (Fig 2D).

Translational read-through of a CTNS W138X mutant expression vector in human embryonic kidney (HEK293) cells and patient fibroblasts
To ascertain whether ELX-02 induces translational read-through of the CTNS W138X nonsense mutation, we transiently transfected HEK293 cells with a pCMV-driven expression vector containing either wildtype CTNS/DsRed or CTNS W138X /DsRed. After 48 hours, DsRed fluorescence was evident in cells transfected with wildtype but not the untreated nonsense mutant vector (Fig 3A). However, strong perinuclear expression of CTNS W138X /DsRed was induced by 48-hour exposure to ELX-02 (400 μg/mL). To confirm this observation, we also transfected CTNS del/del fibroblasts (WG0881) with a CTNS W138X /His vector. As seen in Fig 3B, weak expression of the CTNS/His fusion protein was detectable on immunoblots probed with anti-His antibody. ELX-02 (400 μg/mL) treatment for 48 hours induced strong expression of CTNS W138X /His.

ELX-02 induces translational read-through of endogenous CTNS W138X in nonsense mutant fibroblasts
To confirm that ELX-02 permits expression of endogenous CTNS protein, we treated fibroblasts from a homozygous CTNS W138X/W138X patient with ELX-02 (400 μg/mL) for 72 hours. As seen in Fig 4A, weak endogenous baseline CTNS expression was significantly increased by ELX-02 treatment. We also noted a dose-dependent (20-400 μg/mL) increase of ELX-02 on CTNS transcript level (Fig 4B). At ELX-02 concentrations of 100 μg/mL and above, CTNS transcript levels were greater that the normal transcript level of wildtype fibroblasts.

ELX-02 induces expression of a functional CTNS protein
Fibroblasts from a CTNS W138X/W138X cystinosis patient (WG1012) were treated with ELX-02 (400 μg/mL) for 72 h. As seen in Fig 5A, a time-dependent reduction in cellular cystine is observed; at 72 h, cystine level was 24% (±14% SEM) of untreated baseline. Cells were then treated with increasing concentrations of ELX-02 for 72 h. As seen in Fig 5B, ELX-02 induced a dose-dependent reduction of cystine in the nonsense mutant fibroblasts; at an ELX-02 concentration of 400 μg/mL, cell cystine was reduced to 15% (± 22% SEM) of untreated baseline.
No effect of ELX-02 was noted on fibroblasts from a cystinosis patient harboring homozygous CTNS del/del mutations.

Effect of ELX-02 on nonsense mutant compound heterozygotes
Since the majority of cystinosis patients harbouring a nonsense mutation bear another class of mutation on the second CTNS allele, we studied the effect of ELX-02 on CTNS W138X/del57kb fibroblasts (WG2379). As seen in Fig 6A, reduction of fibroblast cystine at 72 h (53% of We compared the effect of ELX-02(400 μg/mL) to the effect of cysteamine at the peak serum concentration achieved during standard oral therapy (50 μM) in CTNS W138X/W138X fibroblasts (Fig 7). Each drug alone significantly reduced fibroblast cystine levels (60% and 50% respectively) compared to untreated baseline. Addition of ELX-02 to cysteamine appeared to have an additive effect, but reduction of cystine was not significantly different from either drug alone.

Ctns Y226X/Y226X nonsense mutation mouse model
To ascertain whether the nonsense mutation translational read-through effects of ELX-02 are evident in vivo, we generated a Ctns nonsense mutant mouse. Using a zinc-finger nuclease to edit the Ctns gene, we introduced a premature STOP codon (TAG) in Ctns exon 8 ( Fig 8A). Ctns Y226X/Y226X mice showed 8-fold increase in kidney cystine content ( Fig 8B) and reduction of kidney Ctns transcript level to 22% of wildtype controls (Fig 8C). Prior to 6 months, the homozygous nonsense mutant mice appeared normal, but developed low-molecular-weight proteinuria by 9 months (Fig 8D). Kidney sections from mutant mice showed scattered crystals under polarized light ( Fig 9A) and occasional atrophy of columnar cells in Bowman's capsule in male mice (Fig 9B).

ELX-02 reduces cystine accumulation in embryonic fibroblasts derived from Ctns Y226X/Y226X mice
To confirm that ELX-02 induces sufficient Ctns read-through to reduce pathologic cystine accumulation in the Y226X nonsense mutant mouse, we isolated fibroblasts from mutant mouse embryos (MEFs), cultured them in monolayer and exposed the cells to various concentrations of ELX-02. ELX-02 had a dose-dependent effect on restoration of MEF Ctns transcript level ( Fig 10A) and reduction of half-cystine (Fig 10B) at 48 h.

Pharmacokinetics and tissue uptake of ELX-02 in Ctns Y226X/Y226X mice
ELX-02 levels in tissues of Ctns Y226X/Y226X mice were measured following repeat subcutaneous administration (twice weekly, total of 8 doses) at dose levels of 10 and 30 mg/kg/dose. Highest levels of ELX-02 were measured in the kidney, followed by spleen and liver, with lower levels in other tissues (lung, heart, cochlea and brain) (Fig 11A). Accumulation in tissues was dose dependent without gender difference.
The pharmacokinetic profile of ELX-02 was assesed in plasma and kidney tissue of Ctns Y226X/Y226X mice following single and repeat administration (twice weekly total of 7 doses) at a dosesof 10 mg/kg. In plasma ELX-02 was rapidly absorbed with a Tmax of 0.25 h after both single and repeated administration. ELX-02 was rapidly eliminated from plasma in a biphasic manner with the terminal half-life of 0.5 h. Systemic exposure (Cmax and AUC) following single and repeat dosing were similar (Fig 11B).
In the kidney Cmax was reached after 24-41 h and high levels were sustained up to 8 h post dose (Fig 11C). No gender differences were observed for any of the parameters in plasma and kidney.

ELX-02 efficacy and tolerability in CtnsY226X/Y226X mice
The effect of ELX-02 on cystine levels was evaluated in Ctns Y226X/Y226X mice, following repeated administration of ELX-02 (10 mg/kg/dose) or vehicle (Fig 12). ELX-02 reduced  kidney cystine levels by 30% in treated animals compared with untreated animals (Fig 12A). No mice died, exhibited weight loss or changes in fur coat. Serum creatinine was similar in treated vs untreated animals (Fig 12B).
The in vitro effects of ELX-02 on reduction of cystine content in fibroblasts was evident within 24 h of treatment (Fig 4A). To predict how long the ELX-02-induced CTNS protein might last in vivo, we evaluated CTNS protein half-life in proximal tubule (HK-2) cells using a non-radioactive pulse-chase labeling assay [22]. CTNS protein was measured by Western blot and normalized to β-actin. Signal intensities were quantified using the Image Lab software (Bio-Rad).

Discussion
Our studies indicate that ELX-02 restores translational read-through of a functional full-length CTNS. Other investigators have shown that, by relaxing translational fidelity, aminoglycosides permit binding to a near-cognate transfer RNA at the premature STOP codon and most often inserts the wildtype amino acid [13,24]. Our data demonstrate that ELX-02 induces readthrough of an exogenous CTNS W138X construct and of the endogenous CTNS W138X in patient fibroblasts. Furthermore, the magnitude of the ELX-02 (200-400 μg/mL) effect is sufficient to suppress nonsense mutation-induced mRNA decay and reverse pathologic accumulation of cystine. As seen in Fig 7, ELX-02 reduced fibroblast cystine to levels similar to those achieved by cysteamine in vitro (at peak blood levels reported in patients) [20] and in vivo during standard clinical therapy [5]. This benefit is achieved by nonsense mutation read-through, since no effect of ELX-02 is seen in fibroblasts harboring a biallelic CTNS deletion.
The leukocyte cystine level of healthy heterozygous mutant CTNS carriers is only marginally higher than that of normal individuals [25], yet in vitro cystine efflux rate from isolated patient lysosomes is about 50% of normal [26]. Thus, rescue of CTNS function to 50% of normal correlates with clinical benefit. Kalatzis et al estimated that lysosomal cystine efflux rates are about 19% (±10%) of normal in patients with adult-onset ocular cystinosis and about 9-20% in patients with juvenile onset ocular disease [27]. Therefore, the threshold for clinical benefit seems to lie between 20% and 50% residual activity. Our studies did not directly measure the effect of ELX-02 on lysosomal cystine efflux rate, but since cell cystine is significantly reduced, efflux must be at least sufficient to exceed the rate at which cystine is generated within lysosomes.
Current therapy of cystinosis involves oral cysteamine (1.3g/m 2 /day) in four divided doses. At this dose, peak serum cysteamine is about 50 μM [20]. Oral cysteamine delays renal transplantation, improves growth and reduces hypothyroidism in cystinosis [5]. In our study, homozygous CTNS W138X fibroblasts exposed to 400 μg/mL ELX-02 for 48 h achieved reduction of cell cystine to 33% of untreated baseline; this was slightly better than the effect of 50 μM cysteamine (50% of untreated baseline) for 48 h. Thus, the magnitude of the ELX-02 effect on homozygous CTNS W138X/W138X cells appears to be clinically relevant. Interestingly, ELX-02 was only slightly less effective in cells where the W138X mutation was in compound heterozygosity with a CTNS deletion allele. This is important, since the majority of cystinosis patients are not likely to have biallelic nonsense mutations. At least six other CTNS nonsense mutations have been reported and it is likely that the efficacy of ELX-02 varies somewhat, depending on the specific mutation and its nearby sequence context [24,28,29] as in other genetic diseases (Table 1).
To extend our observations, we used zinc-finger nuclease editing to generate a Ctns Y226X nonsense mutant mouse model. The mutation was inserted into Ctns exon 8, to truncate the protein at a position similar to the Ctns neo/neo C57Bl6 knockout mouse [19]. On the CD1 background, the renal phenotype of Ctns Y226X mice was somewhat milder than that of Ctns neo/neo / C57Bl6 knockout mice, but kidney cystine was elevated, Ctns transcript was reduced and cystinosin protein was undetectable in mutant kidney. Between 6-9 months, mutant mice developed low molecular weight proteinuria and scattered atrophy of proximal tubules ("swan neck deformity") with bi-refringent crystals. When the mice were infused with a single dose of ELX-02 (10 mg/kg), plasma levels fell rapidly but the drug was concentrated in kidney 10-15 times the peak level in blood and displayed a prolonged retention phase over 2-3 days. This corresponds to the characteristic uptake and sustained retention of aminoglycosides in association with intracellular megalin, reported by others [30]. In the HK-2 proximal tubular cell line, we estimated the half-life of cystinosin protein to be about 22-30 h. On the basis of these observations, we infused 6-month old Ctns Y226X/Y226X mice with 7 repeated intravenous doses of ELX-02 (10 mg/kg) twice per week and found that kidney cystine was significantly reduced. Thus, ELX-02 displays favorable kidney pharmacokinetics for intermittent dosing and significantly reverses pathologic renal accumulation of cystine in vivo without short-term nephrotoxicity (elevation in serum creatinine).
Although the CTNS W138X allele is prevalent (40-50%) among French Canadian cystinosis patients, it is distributed across North America and Europe at lower frequency [31]. Other CTNS nonsense mutations have been reported but these are uncommon and, in their seminal report, Town et al found that about 15% of cystinosis patients worldwide carried one or more CTNS nonsense mutations [27,32]. Thus, the potential of aminoglycoside read-through therapy would be limited to a specific subset of cystinosis patients. However, if it works in cystinosis, there is no a priori reason why the same strategy cannot be applied to any of the >4000 known genes with phenotype-causing mutations (OMIM).