AON-mediated Exon Skipping Restores Ciliation in Fibroblasts Harboring the Common Leber Congenital Amaurosis CEP290 Mutation.

Leber congenital amaurosis (LCA) is a severe hereditary retinal dystrophy responsible for congenital or early-onset blindness. The most common disease-causing mutation (>10%) is located deep in intron 26 of the CEP290 gene (c.2991+1655A>G). It creates a strong splice donor site that leads to insertion of a cryptic exon encoding a premature stop codon. In the present study, we show that the use of antisense oligonucleotides (AONs) allow an efficient skipping of the mutant cryptic exon and the restoration of ciliation in fibroblasts of affected patients. These data support the feasibility of an AON-mediated exon skipping strategy to correct the aberrant splicing.

Introduction

Retinal dystrophies are the leading cause of incurable blindness in developed countries (prevalence 1/3,000 worldwide). Over the last decade, retinal dystrophies have demonstrated a tremendous clinical, genetic, molecular, and physiopathological heterogeneity.

Leber congenital amaurosis (LCA, MIM204000) is the earliest and most severe of these disorders. With prevalence of 1/60,000 worldwide it is a common cause of blindness in childhood (10%).[1] It is one of the most representative examples of retinal dystrophies heterogeneity. Outcome ranges from severe stationary cone-rod disease with extremely poor visual acuity (VA ≥ light perception; type I) to progressive, yet severe, rod-cone dystrophy with measurable VA in the first decade of life (20/200 ≥ VA ≥ 60/200; type II).[2] Hitherto, 15 disease genes have been identified which account for ~70% of the cases. They are either specifically or preferentially expressed in the retinal pigment epithelium or photoreceptors, or widely expressed.[1,3] This physiopathological heterogeneity challenges the development of therapies with different retinal cell types to be targeted and variable therapeutic windows.

In Western European countries, the centrosomal protein 290 (CEP290, MIM610142) is the most common LCA gene (20%)[4,5] and accounts consistently for the most severe disease phenotype (type I).[5] CEP290 encodes an integral component of the ciliary gate that bridges the transition zone between the cilia and cytoplasm of a broad range of cells including photoreceptors. The protein plays an important role in maintaining the structural integrity of this gate, and thus has a crucial role in maintaining ciliary structure and function.[6] As a result, in addition to LCA, CEP290 mutations cause a wide range of distinct phenotypes,[7] including Senior-Loken syndrome (SLNS6, MIM610189), Joubert syndrome (JBTS5; MIM610188), Bardet-Biedl syndrome (BBS, 209900), and the lethal Meckel-Grüber syndrome (MKS4, MIM611134).

Despite the identification of over 100 unique CEP290 mutations, no clear genotype-phenotype correlations could yet be established[7] with the exception of the LCA-specific c.2991+1655A>G mutation.[4,5] This single mutation accounts for 10% of cases[5] and represents therefore an important therapeutic target. It is located deep in intron 26 where it creates a splice donor downstream of a strong acceptor splice site. As a result, a cryptic 128 bp exon encoding a stop codon is inserted in the CEP290 messenger RNA (mRNA).[4]

Considering the potential of exon-skipping as a means to bypass protein-truncation[8,9,10,11] resulting from the c.2991+1655A>G mutation, we assessed the efficacy of antisense oligonucleotides (AONs) as splice switching oligonucleotides to correct the abnormal splicing in cell lines of LCA patients harboring the mutation.

Results

Expression levels of CEP290 mRNAs

Fibroblasts were derived from skin biopsies of controls (C1–C4), heterozygous carriers (S1–S3), and homozygous (P1, P2, and P4) or compound heterozygous patients (P3). Expression levels of the wild-type and mutant CEP290 mRNAs were measured by quantitative reverse transcription PCR (RT-qPCR) in all cell lines. Compared to controls and healthy carriers, patients' fibroblasts expressed reduced yet measurable levels of wild-type mRNAs (mean relative expression levels of 0.72 ± 0.14, 0.46 ± 0.06, and 0.16 ± 0.02 in control, carrier, and patient cell lines respectively; ). Mutant mRNAs were detected to low levels in patients and heterozygous carriers suggesting degradation through nonsense-mediated mRNA decay (NMD). To assess this hypothesis, we compared the mean level of expression of the mutant and wild-type alleles in the patient P1 cell line before and after treatment with NMD-inhibitor emetine.[12] The significant increase of mutant but not wild-type mRNA levels in cells treated with emetine supported the view that the aberrant transcript is prone to NMD (P < 0.003; ).

Design of AON

Important parameters including guanine-cytosine content influence efficiency of AONs to induce exon skipping.[13] Nevertheless, the design of AON remains empiric. We designed five 2′-O-methyl-phosphorothioate (2′-OMePs) AONs targeting either the donor site (HD26 (+10−11), HD26 (+19−10), and HD26 (+7−18)) or exonic splice enhancer (ESE) sequences predicted using the ESEfinder program (ESE (+50+70) and ESEbis (+90+120); ).

AON-induced correction of the splice defect

AON efficacy was assessed by RT-qPCR by measuring the level of expression of the wild-type and mutant alleles in control, patient, and carrier cell lines transfected for 24 hours with 150 nmol/l of AON in the cationic peptide LAH4-L1[14] (wt/wt 10/1). The use of AONs resulted in a significant increase in wild-type mRNA levels suggesting their ability to skip efficiently the aberrant transcript (Supplementary Figure S1).

The ESE (+50+70) AON was selected for further experiments for its ability to reproducibly induce targeted exon-skipping at high levels (, Supplementary Figure S2 and Supplementary Materials and Methods).

To assess whether the skipping efficacy was dose- dependent, we treated patient, carrier, and control cell lines with increasing amounts of ESE (+50+70) AON. In carrier- and patient-derived cell lines, expression levels of the wild-type mRNA increased with the amount of AON. At an AON concentration of 150 nmol/l, the wild-type mRNA expression in patient cell lines reached the levels measured in controls (). In contrast, the expression of the wild-type CEP290 mRNA was unchanged in control cells whatever the AON concentration indicating that it did not interfere with normal CEP290 splicing ().

Furthermore, we showed that the transfection of a sense version of the ESE (+50+70) AON (ESEsense (+50+70)) in patient, carrier, and control cell lines did not interfere with CEP290 splicing ().

Fluorescently labeled antisense and sense oligonucleotides ESE (+50+70) were delivered to cell lines of patient P1 to make sure that these results were not due to reduced delivery of the sense oligonucleotide. Similar transfection efficiencies (>90%) were observed (Supplementary Figure S3 and Supplementary Materials and Methods).

When, patient P1 fibroblasts were subjected to ESE (+50+70) AON treatment in the presence of emetine, the expression of the wild-type mRNA increased whereas that of the mutant mRNA decreased, supporting the view that restored stable wild-type mRNA resulted from resplicing of the aberrant mutated NMD-prone mRNA (Supplementary Figure S4).

Finally, RT-qPCR carried out on total RNA extracted from cytoplasmic fractions of treated patient P1 cell lines detected significant levels of respliced mRNAs supporting transport of corrected mRNA in the cytoplasm (Supplementary Figure S5 and Supplementary Materials and Methods).

Together, these data support dose-dependent and sequence-dependent ability of the ESE (+50+70) AON to induce skipping.

AON-mediated skipping induces increase in CEP290 concentration and improves ciliogenesis

To assess the effect of AON-induced exon-skipping on the CEP290 protein amount, we performed western blot analyses. A polyclonal anti-CEP290 antibody raised against the carboxy-terminus was used which detects the wild-type but not the mutant CEP290 protein.

Transfection of patient and carrier cells using the ESE (+50+70) AON showed efficient exon-skipping at the mRNA level (). Western blot analyses of the transfected cells showed that ESE (+50+70) AON delivery resulted in an increased amount of the CEP290 protein (). These data give strong support to the view that the increase in the wild-type transcript abundance results in an increase of protein concentration.

Considering the role of CEP290 in cilia assembly and/or maintenance,[6] patient and control cell lines were subjected to a 30 hours-serum starvation to assess their respective ability to build a primary cilium. Significantly reduced primary cilia expression was found in patient cell lines compared to controls (mean 48.6 ± 6.5% versus 83.6 ± 3.2% in controls; P = 0.0097; ).

Interestingly, transfection of the antisense but not the sense oligonucleotide increased cilia expression to control levels, suggesting improved ciliation (mean 75.3 ± 3.5% versus 78.3 ± 3.4% in controls; P = 0.624) ().

Discussion

Mis-splicing is among the most frequent pathogenic mechanisms underlying genetic diseases.[15,16] Strategies have been developed which allow the modification of splice patterns of genes to overcome pathogenic process leading to disease drugs, isoform-specific antibodies, trans-splicing approaches, RNA interference and AONs have been investigated for their ability to modify splicing in various diseases.[15,16,17,18,19] Compared to most other molecules, nuclease-protected phosphorothioate AONs are of easy design[13] and inexpensive. In addition, these splice switching AONs can be widely used to restore a wild-type genetic configuration by allowing the skipping of stray pseudoexons that are relatively common in most genetic diseases;[15] 2′-OMePS AONs have been primarily used to treat acute diseases over a period of weeks or occasionally months.[15] Now, clinical trials aiming at assessing lifelong treatments of various genetic diseases are ongoing.[15] Current ongoing trials in patients affected with Duchenne muscular dystrophy have shown that short-term treatment using 2′-OMePS AONs is well tolerated.[9,20,21]

With respect to diseases affecting vision, it is worth remembering that because the eye is easily accessible, small, confined, and has an efficient blood-retina barrier, it is regarded as an ideal target organ for topical treatment. So far, to our knowledge the use of 2′-OMePS AONs is restricted to the treatment of acute cytomegalovirus retinitis using the FDA approved Vitravene.[22] Considering inherited diseases of the retina, the promising clinical trial phase of gene therapy for LCA due to RPE65 mutations have triggered efforts in developing similar treatments.[23] CEP290 replacement, whose development would benefit from available spontaneous animal models mimicking the human disease (rd16 mouse and rdAc Abbyssinian cat),[24,25] is an attractive option to treat LCA patients harboring CEP290 mutations, including the common c.2991+1655A>G mutation. However, CEP290 may be expressed in at least 11 alternative transcripts most of which are predicted to encode proteins whose tissular expression and function are unknown (http://harvester4.fzk.de/harvester/human/IPI00794/IPI00794668.html). Restoring the wild-type genetic configuration by correcting the aberrant splicing resulting for the c.2991+1655A>G mutation may thus be regarded as a preferable option.

Here, we show that this strategy is efficient to correct abnormal splicing accounted for by the CEP290 c.2991+1655A>G mutation in fibroblasts cell lines obtained from LCA patients. Interestingly, AONs complementary to either the splice donor or exonic splice enhancer sites cause a three to 4.5-fold upregulation of wild-type mRNA in patient-derived fibroblasts to levels equaling that in wild-type cells indicating a significant degree of flexibility in the design of future therapeutic oligonucleotides.

Western blot analyses indicated that despite significant upregulation of wild-type mRNA in patient cells, as measured by RT-qPCR, protein levels were moderately increased (). Discrepancies between have been previously reported.[26,27] Slow synthesis of large proteins, post-transcriptional, and post-translational regulations have been considered to explain these discrepancies.[28,29] Considering that the increase in restored stable wild-type CEP290 mRNA arises from nuclear resplicing of the aberrant mutated NMD-prone mRNA (Supplementary Figure S4) and that significant levels of respliced mRNA were measured in cytoplasmic fractions of AON-treated cells (Supplementary Figure S5), these hypotheses are likely to account at least in part to the discrepancy between mRNA and protein abundance increase.

Recently, we reported that CEP290 mutations, including the c.2991+1655A>G change, are expressed in nasal epithelial cells of LCA patients.[30] In accordance with the function of CEP290 in cilia assembly and/or maintenance,[6] we evidenced consistent ciliary defects in nasal ciliated cells of patients. Most of these cells presented reduced numbers of cilia, most of which were shorter than normal and presented various axonemal defects.[30] We report here, that in contrast to control cells, high numbers of patient cells display no primary cilium even after a 30 hours-serum starvation, suggesting that the CEP290 c.2991+1655A>G mutation alters the ciliation process in the cells. Most interestingly, the results we report here show that, despite a moderate increase in protein abundance compared to respliced mRNAs, the use of AON is able to improve significantly ciliation in patient cell lines supporting the therapeutic potential of this strategy.

Finally, the efficiency and absence of major side-effect of intravitreal injections of FDA-approved Vitravene[22] to treat acute cytomegalovirus retinitis is an encouragement to continue the development of this promising AON-induced exon-skipping strategy to bypass the most common LCA-causing mutation.

Materials and Methods

Transfection agent and AONs. The 26 residues long cationic transfecting peptide LAH4-L1[14] (KKALLAHALHLLALLALHLAHALKKA) was prepared by automated solid-phase synthesis on Millipore 9050 or ABI 431 synthesizers using Fmoc Chemistry (a kind gift by Andrew J. Mason and Burkhard Bechinger). The 2′-O-methyl RNA phosphorothioate oligonucleotides were obtained from Sigma-Aldrich (St Quentin Fallavier, France) ().

Cell culture and transfection of AONs. Skin biopsies were obtained from four LCA patients harboring the c.2991+1655A>G mutation (3/4 homozygous, P1, P2, P4; 1/4 compound heterozygous with the c.5850delT, p.Phe1950LeuFsX14 mutation P3), three heterozygous unaffected carriers (S1–S3), and control individuals (C1–C4). Written informed consent was obtained for each individual and research was approved by institutional review board. Primary fibroblasts were isolated by selective trypsinization and proliferated at 37 °C, 5% CO2 in Opti-MEM Glutamax I medium (Invitrogen, St Aubin, France) supplemented with 10% fetal bovine serum (Invitrogen), 1% ultroser G substitute serum (Pall, Saint-Germain-en-Laye, France), and 1% streptomycin/penicillin (Invitrogen). Fibroblasts between passage 7 and 9 were plated at 4 × 105 cells/well in 6-well plates 24 hours before transfection. Cells at 80% confluence were transfected with 2′-OMePS AONs (150 nmol/l) in Opti-MEM using LAH4-L1 at a 1:10 (wt:wt) AON:peptide ratio. After 3 hours of incubation at 37 °C, the transfection medium was replaced by fresh culture medium. For the inhibition of NMD, 25 µg/ml of emetine dihydrochloride hydrate (Sigma-Aldrich) was added to the fresh culture medium for 12 hours.

RNA extraction and cDNA synthesis. Twenty-four hours after transfection, the transfected and untreated cells were processed. Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Courtaboeuf, France) according to manufacturer's protocol. All samples were DNase treated by the RNase-free DNase set (Qiagen, Courtaboeuf, France). Concentration and purity of total RNA was assessed using the Nanodrop-8000 spectrophotometer (Thermo Fisher Scientific, Illkirch, France) before storage at −80 °C. Qualitative analysis of total RNA was performed using the Bioanalyzer 2220 (RNA 6000 Nano kit; Agilent Technologies, Massy, France) to verify that the RNA integrity number was between 8 and 10. First-stranded cDNA synthesis was performed from 500 ng of total RNA extracted using Verso cDNA kit (Thermo Fisher Scientific) with random hexamer:anchored oligo(dT) primers at a 3:1 (vol:vol) ratio according to the manufacturer's instructions. A non-RT reaction (without enzyme) for one sample was prepared to serve as control in RT-qPCR experiments.

RT-qPCR. To measure the level of expression of CEP290 mRNAs, the wild-type and mutant transcripts were amplified as 93 and 117 bp fragments, respectively. Regions of 132, 80, 84, 101, and 95 bp within the human TATA box-binding protein mRNA (TBP, NM_003194), the human β-2-microglobulin mRNA (B2M, NM_004048.2), the human β-glucuronidase mRNA (GUSB, NM_000181.3), the human hypoxanthine phosphoribosyltransferase 1 mRNA (HPRT1, NM_000194), and the human P0 large ribosomal protein mRNA (RPLP0, NM_001002.3) were used for normalization, respectively. A 99 bp fragment of the human albumin gene (ALB, NM_000477) was used to control the non-contamination of cDNAs by genomic DNA. Primers were designed using the Oligo Primer Analysis Software v.7 available at http://www.oligo.net. The specificity of primer pairs to PCR template sequences was checked against the NCBI database using the Primer-BLAST software (http://www.ncbi.nlm.nih.gov/tools/primer-blast). Primer sequences were as follows: CEP290wt forward, 5′-tgactgctaagtacagggacatcttg-3′ CEP290wt reverse, 5′-aggagatgttttcacactccaggt-3′ CEP290mt forward, 5′-ctggccccagttgtaatttgtga-3′ CEP290mt reverse, 5′-ctgttcccaggcttgttcaatagt-3′ reference genes TBP forward, 5′-tgcacaggagccaagagtgaa-3′ TBP reverse, 5′-caca tcacagctccccacca-3′ B2M forward, 5′-cctggaggctatccagcgt act-3′ B2M reverse 5′-tcaggaaatttgactttccattctct-3′ GUSB forward, 5′-gcggtcgtgatgtggtctgt-3′ GUSB reverse, 5′-gtgagc gatcaccatcttcaagt-3′ HPRT1 forward, 5′-accagtcaacaggggacataaaagta-3′ HPRT1 reverse, 5′-tttgccagtgtcaattatatcttcca-3′ RPLP0 forward, 5′-ttctcgttcctggagggtgt-3′ RPLP0 reverse, 5′-cgttgatgatagaatggggtactgat-3′ ALB forward 5′-atatttccttgtcatcagggttcaga-3′ ALB reverse 5′-accttattggatagcccagatacaga-3′.

cDNAs (5 µl of a 1:25 dilution in nuclease-free water) were subjected to real-time PCR amplification in a buffer (20 µl) containing MESA BLUE qPCR Master Mix Plus for Sybr Assay (Eurogentec, Angers, France) and 300 nmol/l of forward and reverse primers, on a Taqman 7900 HT Fast Real-Time PCR System (Applied Biosystems, Courtaboeuf, France) under the following conditions: Taq polymerase activation and initial denaturation at 95 °C for 5 minutes, followed by 50 cycles for 15 seconds at 95 °C, and 1 minute at 65 °C. The specificity of amplification products was determined from melting curve analysis performed at the end of each run using a cycle at 95 °C for 15 seconds, 65 °C for 15 seconds, and 95 °C for 15 seconds. Data were analyzed using the SDS 2.3 software (Applied Biosystems).

For each cDNA sample, the mean of quantification cycle (Cq) values was calculated from triplicates (SD <0.5 Cq). CEP290 expression levels were normalized to the “normalization factor” obtained from the geNorm software for Microsoft Excel[31] which uses the most stable reference genes and amplification efficiency estimates calculated for each primer-pair using fourfold serial dilution curves (1:5, 1:25, 1:125, 1:625). No reverse transcriptase (non-RT), no template control (NTC) reactions, and non-contamination of cDNAs by genomic DNA (ALBh) were used as negative controls in each run (Cq values NTC = undetermined, non-RT >40 and ALBh >40).

The quantitative data are the means ± SEM of three independent experiments and these are presented as ratio among values for individual mRNAs. The significance of variations among samples was estimated using the protected list significant difference of Fisher according to the significance of analysis of variance test (Statview Software, version 5; SAS Institute, Cary, NC).

Western blot analysis. Cells were harvested 24 hours after transfection and lysed in a Triton/SDS buffer (25 mmol/l Tris-base pH 7.8, 1 mmol/l DTT, 1 mmol/l EDTA, 15% Glycerol, 8 mmol/l MgCl2, 1% Triton, and 1% SDS) containing complete protease inhibitor cocktail (Roche, Boulogne-Billancourt, France) on ice for 30 minutes with repeated mixing. Released DNA was fragmented by 20 seconds of Ultra-turrax homogenizer (Ika-Werke, Staufen, Germany) and the lysates were centrifuged (15,000g at 4 °C for 10 minutes). Protein concentrations were determined from the detergent-soluble fractions using the DC Protein Assay kit according to the manufacturer protocol (Bio-Rad, Marnes la Coquette, France). Proteins (125 µg) were denatured at 90 °C for 10 minutes in 4X premixed protein sample buffer (XT sample Buffer; Bio-Rad) and separated by electrophoresis (50 V for 30 minutes followed by 140 V for 90 minutes at room temperature) on NuPAGE 3–8% Tris Acetate gels (Invitrogen). Proteins were transferred (100 V, 2 hours at 4 °C) to Immobilon-P PVDF membranes (Millipore, Molsheim, France). Membranes were blocked with phosphate-buffered saline (PBS) 0.5% Tween-20/5% dry milk powder and incubated overnight at 4 °C under agitation with polyclonal rabbit anti-human CEP290 (Novus Biologicals, Littletown, CO) or monoclonal mouse anti-α-tubulin (Sigma-Aldrich) primary antibodies in 1:1,800 and 1:500,000 dilutions, respectively. Membranes were washed three times in PBS 0.5% Tween-20 solution (15 minutes) and incubated for 1 hour at room temperature with HRP-conjugated donkey anti-rabbit and sheep anti-mouse immunoglobulins secondary antibodies (Amersham GE Healthcare, Courtaboeuf, France) in 1:10,000 dilutions, respectively. ECL Western Blotting Detection Reagents (Amersham GE Healthcare) was applied according to the manufacturer's instructions and the blot was exposed to Amersham Hyperfilm ECL (Amersham GE Healthcare). The relative expression of the CEP290 protein was estimated by densitometry using α-tubulin as reference on a G:Box from Syngene with the GeneSnap and GeneTool Softwares.

Immunofluorescence microscopy. Fibroblasts were seeded at 2.5 × 105 cells/well on glass coverslips in 12-well plates, 24 hours before transfection with the ESEsense (+50+70) and ESE (+50+70) oligonucleotides, in the conditions described previously. Ten hours after transfection, cells were washed with PBS and incubated for 30 hours in serum-free medium (37 °C, 5% CO2). Untreated fibroblasts were processed in the same conditions. Subsequently, cells were fixed in ice-cold methanol (5 minutes at −20 °C) and washed twice in PBS. Cells were permeabilized in PBS supplemented with 3% bovine serum albumin and 0.1% Triton for 1 hour at room temperature before being incubated overnight at 4 °C in permeabilization buffer containing (rabbit anti-γ-tubulin (1:1,000), mouse monoclonal anti-acetylated tubulin (1:1,000); Sigma-Aldrich) primary antibodies. After three washes with PBS, cells were incubated for 1 hour at room temperature in permeabilization buffer containing secondary antibodies (Alexa-Fluor 594– and Alexa-Fluor 488– conjugated goat anti-rabbit IgG (1:1,000) and goat anti-mouse IgG (1:1,000); Invitrogen, Life Technologies, St Aubin, France) followed by three washes with PBS. A mounting media containing 4′,6-diamidino-2-phenylindole (DAPI) (DAPI Fluoromount G; Southern Biotech, Birmingham, AL) was used to label nuclei. Immunofluorescence images were obtained using a Leica DM IRBE microscope and a MicroPublisher 3.3 RTV camera (Q-Imaging France, Bayonne, France). The final images were generated using the Cartograph version 7.2.3 (Microvision Instruments, Evry, France) and ImageJ (National Institutes of Health, Bethesda, MA). The percentage of ciliated cells was calculated from two independent experiments (n > 200 cells for each cell line). The significance of variations among samples was estimated using the protected list significant difference of Fisher according to the significance of the analysis of variance test (Statview Software, version 5).

Figure S1. Evaluation of exon-skipping efficiency using AONs designed from the CEP290 mutant mRNA sequence (see Figure 2). Figure S2. Visualization of AON-mediated correction of aberrant splicing. Figure S3. AON transfection efficiency. Figure S4. AON treatment of patient P1 fibroblasts in the presence of emetine. Figure S5. Relative expression levels of wild-type and mutant in total extracts and cytoplasmic fractions of AON-treated fibroblasts. Materials and Methods.