Myosin7a Deficiency Results in Reduced Retinal Activity Which Is Improved by Gene Therapy

Mutations in MYO7A cause autosomal recessive Usher syndrome type IB (USH1B), one of the most frequent conditions that combine severe congenital hearing impairment and retinitis pigmentosa. A promising therapeutic strategy for retinitis pigmentosa is gene therapy, however its pre-clinical development is limited by the mild retinal phenotype of the shaker1 (sh1−/−) murine model of USH1B which lacks both retinal functional abnormalities and degeneration. Here we report a significant, early-onset delay of sh1−/− photoreceptor ability to recover from light desensitization as well as a progressive reduction of both b-wave electroretinogram amplitude and light sensitivity, in the absence of significant loss of photoreceptors up to 12 months of age. We additionally show that subretinal delivery to the sh1−/− retina of AAV vectors encoding the large MYO7A protein results in significant improvement of sh1−/− photoreceptor and retinal pigment epithelium ultrastructural anomalies which is associated with improvement of recovery from light desensitization. These findings provide new tools to evaluate the efficacy of experimental therapies for USH1B. In addition, although AAV vectors expressing large genes might have limited clinical applications due to their genome heterogeneity, our data show that AAV-mediated MYO7A gene transfer to the sh1−/− retina is effective.


Introduction
Usher syndrome (USH) is the most common deafness-blindness disorder inherited as autosomal recessive and due to mutations in one of 12 different genes [1]. Mutations in MYO7A that encodes for the unconventional Myosin7a, cause Usher syndrome type IB (USH1B, [MIM 276900]) which accounts for about 50% of USH1 cases [1]. USH1B is one of the most severe forms of USH that leads to severe congenital hearing impairment and progressive vision loss due to retinitis pigmentosa that manifests pre-pubertally [1]. The hearing dysfunction of USH1B patients can be ameliorated by cochlear implants while retinitis pigmentosa is currently untreatable [1]. MYO7A gene replacement may represent a valuable therapeutic strategy for the USH1B retina [2], but its development is hampered by difficulties to find an appropriate animal model. Although various rodent models of USH1B that bear mutations in the Myo7a gene are available and present with clear hearing and vestibular defects [3,4,5,6], none presents with frank retinal dysfunction and degeneration [7,8,9,10]. Among them, the most studied is the shaker1 4626SB/4626SB mouse (referred to as sh1 2/2 ), which is homozygous for an early nonsense mutation in the Myo7a gene. Myo7a encodes for a large 2215 aminoacid actin-based motor protein that is expressed in the cochlear hair cells, retinal photoreceptors (PRs: rods and cones) and retinal pigment epithelium (RPE) [11,12,13]. In rods, Myo7a localizes to the connecting cilium where it cooperates to the transport of Rhodopsin (the rod photopigment) to the rod outer segment (OS) discs [14,15]. In RPE, Myo7a mainly resides at the apical side [11,13] and it is required for both proper localization of melanosomes to the RPE microvilli (which surround the PR OS) [16,17] and phagosome motility [18]. Indeed sh1 2/2 mice show: i. slower transport of Rhodopsin that accumulates at the PR connecting cilium [14]; ii. slower distal migration of PR OSs [14]; iii. mislocalization of melanosomes, that do not enter into the RPE microvilli [16]; and iv. slower phagosome motility that results in slower OS digestion [18]. Notably, abnormal melanosome motility has also been reported in human Myo7a-deficient RPE cell cultures [13]. However, differently from patients which experience severe electroretinogram (ERG) decline and PR degeneration [19,20], sh1 2/2 mice do not present neither PR cell loss up to 24 months of age nor important ERG reduction, similarly to the other USH1B models [7,8]. The base for these interspecies differences remained unknown for long time, however it has been recently attributed to the lack in rodents but not in humans of calyceal processes which are the specific photoreceptor structures where USH1 proteins, including MYO7A, localize [21]. Based on the lack of a frank retinal degeneration in sh1 2/2 mice, rescue of the ultrastructural PR and RPE defects represents, to date, the only useful outcome measure for experimental therapies tested in the sh1 2/2 retina [2]. However, to develop effective and clinically-relevant therapeutic strategies for the USH1B retina, it is critical to establish if these ameliorate retinal function in Myo7adeficient mice. In the present study we identified significant, earlyonset electrophysiological abnormalities of the sh1 2/2 retina that remain stable up to 12 months of age and that are independent of PR degeneration. In addition we show that these abnormalities can be reliably used to assess the efficiency of MYO7A gene transfer to the sh1 2/2 retina.

Statistical Analysis
Data are presented as mean6 standard error of the mean (SEM). Statistical p values,0.05 were considered significant. Twoway ANOVA (factors: genotype and luminance) with post-hoc Multiple Comparison Procedure was used to compare the ERGs recorded from sh1 +/2 vs sh1 2/2 mice (Fig. 1A, S1A and S2A). The interaction p values of the ANOVA are the following: Figure 1A: p = 0.98 (a-wave 6 months), p = 0.5 (a-wave 12 months), p = 0.002 (b-wave 6 months), p = 8.7610 25 (b-wave 12 months); Figure S1A: p = 0.009 (a-wave 3 months), p = 0.96 (bwave 3 months); Figure S2A: p = 0.21 (a-wave 6 months), p = 0.12 (a-wave 12 months), p = 0.09 (b-wave 6 months), p = 0.9 (b-wave 12 months). The statistical significant differences between sh1 +/2 and sh1 2/2 eyes at each specific luminance were determined with the post-hoc Multiple Comparison Procedure and marked by asterisks in Figures 1A, S1A and S2A. Two-way ANOVA (factors: age and genotype) with post-hoc Multiple Comparison Procedure was used to compare light sensitivity of sh1 +/2 and sh1 2/2 mice (Fig. 2). The p values of the ANOVA are the following: genotype p = 1.5610 210 , age p,,2610 216 . The significant comparisons determined with the post-hoc Multiple Comparison Procedure were marked by asterisks in Figure 2. The time-course profile of retinal recovery from light desensitization was compared in sh1 +/2 vs sh1 2/2 mice (Fig. 3) by means of two procedures: Gaussian Processes [22] and Two-way ANOVA with post-hoc Multiple Comparison Procedure. Gaussian Processes enable to quantify the true signal and noise embedded in data profiles over time. In particular, given the observed profiles of recovery from light desensitization, two different hypotheses H1 and H2 were compared: H 1 : the recovery profiles are truly differential; H 2 : the difference in recovery profiles is just the effect of random noise. The plausibility of the two different hypotheses is assessed in terms of Bayes Factor (BF) = Probability (Data|H 1 )/ Probability (Data|H 2 ). The Gaussian Processes Bayes Factors (BF) relative to the recovery profiles depicted in Figure 3 are the following: 85.06 (A), 112.4 (B), 17.6 (C). The retinal recovery from light desensitization was also compared in sh1 +/2 vs sh1 2/2 mice (Fig. 3) performing a Two-way ANOVA (factors: time and genotype) with post-hoc Multiple Comparison Procedure. The interaction p values of the ANOVA are the following: Figure 3: A and B p,,2610 216 , C. p = 5.3610 26 . The significant comparisons determined with the post-hoc Multiple Comparison Procedure were marked by asterisks in Figure 3. A likelihood ratio test for Negative Binomial generalized linear models [23] was used to compare the number of rows of PR nuclei in sh1 +/2 vs sh1 2/2 (Fig. 4). Statistical analyses were not required for data depicted in Figure 5, S3, S4, S5 and S6. One-way ANOVA with post-hoc Multiple Comparison Procedure was used to make comparisons among groups in Figure 6. The p values of the ANOVA are the following: p = 4.6610 211 (albino), p = 0.007 (pigmented). The significant comparisons determined with the post-hoc Multiple Comparison Procedure were marked by asterisks in Figure 6. A likelihood ratio test for Negative Binomial generalized linear models [23] was used to compare the number of correctly localized melanosomes (Fig. 7). The significant comparisons determined with the post-hoc Multiple Comparison Procedure were marked by asterisks in Figure 7. The retinal recovery from light desensitization was compared in hMYO7Avs EGFP-treated sh1 2/2 mice (Fig. 8) performing a Two-way ANOVA (factors: time and treatment) with post-hoc Multiple Comparison Procedure. The interaction p values of the ANOVA is p = 0.0003.The significant comparisons determined with the post-hoc Multiple Comparison Procedure were marked by asterisks and pound keys in Figure 8.

Ethics Statement
This study was carried out in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and with the Italian Ministry of Health regulation for animal procedures. All procedures on mice were submitted to the Italian Ministry of Health; Department of Public Health, Animal Health, Nutrition and Food Safety on October 17th, 2011. The Ministry of Health approved the procedures by silence/consent, as per article 7 of the 116/92 Ministerial Decree. Surgery was performed under anesthesia and all efforts were made to minimize suffering. Mice. Mice were housed at the Institute of Genetics and Biophysics animal house (Naples, Italy) and maintained under 12-hour light/ dark cycle (10-50 lux exposure during the light phase). Albino shaker1 4626SB/4626SB mice (referred as sh1 2/2 ) were imported from the Medical Research Council Institute of Hearing Research (Nottingham, UK) and maintained inbred; breedings were performed crossing heterozygous female with homozygous males. The albino shaker1 mice used in this study were either homozygous shaker1 4626SB/4626SB (sh1 2/2 ) or heterozygous shaker1 +/4626SB (sh1 +/2 ). Pigmented shaker1 4626SB/4626SB mice were imported from the Wellcome Trust Sanger Institute (Cambridge, UK) (a kind gift of Dr. Karen Steel) and back-crossed twice with CBA/Ca mice purchased from Harlan Italy SRL (Udine, Italy) to obtain heterozygous sh1 +/2 mice to expand the colony, then, they were maintained intercrossed; breedings were performed crossing heterozygous female with heterozygous males. The pigmented sh1 mice used in this study were either homozygous shaker1 4626SB/4626SB (sh1 2/2 ) or heterozygous shaker1 +/4626SB (sh1 +/2 ). All homozygous sh1 2/2 mice, whether albino or pigmented, showed at visual inspection hyperactivity, headtossing, and circling behavior caused by vestibular dysfunctions. All mice were genotyped to confirm the visual inspection. The genotype for the Myo7a 4626SB allele was performed by PCR analysis of genomic DNA (extracted from the mouse tip tail) followed by DNA sequencing. The primers used for the PCR amplification are the following: Fw1 (GTGGAGCTTGACATC-TACTTGACC) and Rev3 (AGCTGACCCTCAT-GACTCTGC) which generate a product of 712 bp. Direct sequencing of the PCR product, performed using the Fw1 primer, confirmed that our lines of sh1 mice carry the Myo7a 4626SB allele which results in Gln720X [24]. In addition, since the CBA/Ca mice purchased from Harlan carry the rd1-PDEb mutation we also genotyped the pigmented sh1 mice used in this study for the rd1-PDEb allele. PCR amplification on genomic DNA using the Rd1Fw (CACACCCCCGGCTGAT-CACTG) and Rd1Rev (CTGAAAGTTGAACATTTCATCAG) oligonucleotides gives a product of either 319 or 320 bp for the wild typeor rd1-PDEb allele, respectively. Direct sequencing of the PCR products using the Rd1Rev primer as well as restriction enzyme digestion with DdeI allowed selecting pigmented sh1 mice that did not carry homozygous rd1-PDEb alleles. The RPE65 genotype was performed as previously described [9] and showed that both albino and pigmented sh1 mice were homozygous for the wild type L450 allele.

Electrophysiological Recordings
Before electrophysiological testing, mice were dark reared for three hours and anesthetized [25]. Flash electroretinograms (ERGs) were evoked by 10-ms light flashes generated through a Ganzfeld stimulator (CSO, Costruzione Strumenti Oftalmici, Florence, Italy) and registered as previously described [25]. ERGs and b-wave thresholds were assessed using the following protocol. Eyes were stimulated with light flashes increasing from 25.2 to +1.3 log cd*s/m 2 (which correspond to 1610 25.2 to 20.0 cd*s/m 2 ) in scotopic conditions. The log unit interval between stimuli was 0.3 log from 25.4 to 0.0 log cd*s/m 2 , and 0.6 log from 0.0 to +1.3 log cd*s/m 2 . For ERG analysis in scotopic conditions the responses evoked by 11 stimuli (from 24 to +1.3 log cd*s/m 2 ) with an interval of 0.6 log unit were considered (Fig. 1, S1 and S2). To minimize the noise, three ERG responses were averaged at each 0.6 log unit stimulus from 24 to 0.0 log cd*s/m 2 while one ERG response was considered for higher (0.02+1.3 log cd*s/m 2 ) stimuli. The time interval between stimuli was 10 seconds from 25.4 to 0.7 log cd*s/m 2 , 30 sec from 0.7 to +1 log cd*s/m 2 , or 120 seconds from +1 to +1.3 log cd*s/m 2 . a-and b-waves amplitudes recorded in scotopic conditions were plotted as a function of increasing light intensity (from 24 to +1.3 log cd*s/m 2 , Fig. 1, S1 and S2). The retinal sensitivity to light was assessed in scotopic conditions and measured as the lowest light stimulus able to evoke a b-wave-shaped response in triplicate (only reproducible responses were considered, Fig. 2 and S1B), as previously described [26]. The photopic ERG was recorded after the scotopic session by stimulating the eye with ten 10 ms flashes of 20.0 cd*s/ m 2 over a constant background illumination of 50 cd/m 2 . To assess the recovery from light desensitization (Fig. 3) eyes were stimulated with 3 light flashes of 1 cd*s/m 2 and then desensitized by exposure to constant light (300 cd/m 2 ) for 3 minutes. Then, eyes were stimulated over time using the pre-desensitization flash (1 cd*s/m 2 ) at 0, 5, 15, 30, 45 and 60 minutes post-desensitization. The recovery of rod activity was evaluated by performing the ratio between the b-wave generated post-desensitization (at the different time points) and that generated pre-desensitization. The recovery from light desensitization analysis after AAV injection (Fig. 8) was performed by a blinded operator (different from the one who has injected the animals) who did not know which eyes were injected with AAV-hMYO7A and which with control vectors. For the electrophysiological analysis the number of albino sh1 mice analyzed was the following: Ganzfeld ERG study ( Fig. 1 and S1): 7 sh1 +/2 vs 8 sh1 2/2 at 3 months, 7 sh1 +/2 vs 8 sh1 2/2 at 6 months, 7 sh1 +/2 vs 11 sh1 2/2 at 12 months; light-sensitivity study ( Fig. 2 and S1B): 8 sh1 +/2 vs 8 sh1 2/2 at 3 and 6 months, 7 sh1 +/2 vs 11 sh1 2/2 at 12 months; recovery from light desensitization study (Fig. 3): 4 sh1 +/2 vs 4 sh1 2/2 at 2 months, 7 sh1 +/2 vs 8 sh1 2/2 at 6 months, and 9 sh1 +/2 vs 10 sh1 2/2 at 12 months. The number of pigmented sh1 mice analyzed was the following: Ganzfeld ERG study (Fig. S2): 5 sh1 +/2 vs 4 sh1 2/2 between 6-7 months, 7 sh1 +/2 vs 4 sh1 2/2 at 12 months. The number of eyes varies in Figures 1, 2, 3, S1A and S2A based on whether one or both eyes of the same animal were analyzed or not due to either technical problems (i.e. corneal opacity and dryness) or recording problems.
Generation of the pAAV2.1-human MYO7A Construct, AAV Vector Production and Purification The pAAV2.1-CMV-human MYO7A (pAAV2.1-CMV-hMYO7A) construct was generated as previously described [27]. Briefly, the human MYO7A (hMYO7A) CDS (6648 bp) from isoform 1 (NCBI Reference Sequence NM_000260.3) was obtained by PCR amplification on commercially-available cDNA from human retina and cloned in the pAAV2.1-CMV-EGFP plasmid [28]. The pAAV2.1 plasmid contains the inverted terminal repeats (ITRs) from AAV2 which flank the expression cassette and that are required for AAV vector production [28]. The length of the AAV genome (including ITRs) encoded by the pAAV2.1-CMV-hMYO7A is 8107 bp. This oversize AAV cis-plasmid was used to produce both the AAV2/2 vectors for the in vitro experiments on HEK293 cells (Fig. 5A) as well as the AAV2/5 vectors for the in vivo experiments in sh1 2/2 mice (Fig. 5B, 6-8). The pAAV2.1-CMV-hMYO7A-HA construct was generated by cloning the sequence of the human influenza hemagglutinin tag (HA) in frame with the hMYO7A CDS before the stop codon. The aminoacidic sequence of the HA-tag fused at the hMyo7a Cterminus is the following: MYDVPDYASL. The pAAV2.1-CMV-hMYO7A-HA plasmid was used to produce the AAV2/2 vectors used for the in vitro experiments on HEK293 cells depicted in Figure S5. The pAAV-CMV-CEP290 ( [27], genome length including ITRs: 8930 bp), pAAV2.1-CMV-iPS (sequence available upon request, genome length including ITRs: 7744 bp), and Figure 1. Ganzfeld electroretinograms in albino sh1 mice. A. Mean a-(top panels) and b-(bottom panels) waves from 6-(left panels) and 12-(right panels) month-old sh1 mice. Data are presented as mean6SEM, n indicates the number of eyes analyzed, the arrows point at the photopic ERG. The statistical significant differences between sh1 +/2 and sh1 2/2 mice at each specific luminance are marked by asterisks: *p value,0.05, **p value,0.001. More details on the statistical analysis including specific statistical values can be found in the Statistical analysis paragraph of the Materials and Methods section. B. Representative scotopic ERG waves from one sh1 +/2 and one sh1 2/2 mouse at 12 months of age. doi:10.1371/journal.pone.0072027.g001 pAAV2.1-CMV-Kan (sequence available upon request, genome length including ITRs: 8734 bp) plasmids were used for the production of the oversize AAV2/5 vectors used in Figure S4. AAV vector preparations were produced by the TIGEM AAV Vector Core, by triple transfection of HEK293 cells followed by two rounds of CsCl2 purification [29]. For each viral preparation, physical titers [genome copies (GC)/mL] were determined by averaging the titer achieved by dot-blot analysis [30] and by PCR quantification using TaqMan [29] (Applied Biosystems, Carlsbad, CA, USA).

Southern Blot Analyses of AAV Vector DNA
Alkaline Southern blot was used to analyze the size of the singlestranded DNA packaged in AAV capsids. Three x10 10 GC of viral DNA were extracted from AAV particles. To digest unpackaged genomes, the vector solution was resuspended in 240 mL of PBS pH 7.4 1X (GIBCO, Invitrogen S.R.L., Milan, Italy) and then incubated with 1 U/mL of DNase I (Roche, Milan, Italy) in a total volume of 300 mL containing 40 mM TRIS-HCl, 10 mM NaCl, 6 mM MgCl2, 1 mM CaCl2 pH 7.9 for 2 hours at 37uC. The DNase I was then inactivated with 50 mM EDTA, followed by incubation with proteinase K and 2.5% N-lauroyl-sarcosil solution at 50uC for 45 minutes to lyse the capsids. The DNA was extracted twice with phenol-chloroform and precipitated with 2 volumes of ethanol 100% and 10% sodium acetate (3 M, pH 7). Alkaline agarose gel electrophoresis and blotting were performed as previously described [34]. Ten mL of the 1 kb DNA ladder (N3232L, New England Biolabs, Ipswich, MA, USA) were loaded as molecular weight marker. A 768 bp double strand DNA fragment (corresponding to the CMV promoter from pAAV2.1-CMV-EGFP plasmid) was radio-labeled with [a-32]-CTP using the Amersham Rediprime II DNA labeling System (GE Healthcare Europe, GmbH, Milan, Italy) and used as probe. Prehybridization and hybridization were performed at 65uC in Church buffer [34] for 1 hour and overnight, respectively. Then, the membrane (Whatman Nytran N, charged nylon membrane, Sigma-Aldrich, Milan, Italy) was washed for 30 minutes in SSC 2X-0.1% SDS and for 30 minutes in SSC 0.5X-0.1% SDS at 65uC, and then for 30 minutes in SSC 0.1X-0.1% SDS at 37uC. The membrane was then analyzed by X-ray autoradiography using Amersham Hyperfilm TM MP (GE Healthcare Europe, GmbH, Milan, Italy).

Histology and Light Microscopy
To count the rows of photoreceptor (PR) nuclei in the outer nuclear layer (ONL, Fig. 4) eyes from sh1 mice were enucleated, fixed in 4% paraformaldehyde over-night and infiltrated with 30% sucrose over-night. Then, the cornea and the lens were dissected and the eyecups were embedded in optimal cutting temperature compound (O.C.T. matrix, Kaltek, Padua, Italy). Serial eyecup cryo-sections (10 mm thick) were cut along the horizontal meridian and progressively distributed on slides so that each slide contained representative sections from the whole eye. Sections were stained with hematoxylin and eosin (Richard-Allen Scientific, Kalamazoo, MI, USA) according to standard procedures. The rows of PR nuclei were counted by a blinded observer who did not know which retinas were from sh1 +/2 and which were from sh1 2/2 mice. More specifically, the eyes were collected and the slides were numbered by a person different from the observer who made the measurements. The observer analyzed 3 sections/eye by light microscopy at 40X magnification. Measurements were taken at 500, 1000, and 1500 mm from the optic nerve head both temporally and nasally. The number of sh1 mice analyzed in Figure 4 is the following: 3 sh1 +/2 vs 3 sh1 2/2 (albino, 6 months); 5 sh1 +/2 vs 6 sh1 2/2 (albino, 12 months); 4 sh1 +/2 vs 3 sh1 2/2 (pigmented, 6-7 months); 10 sh1 +/2 vs 3 sh1 2/2 (pigmented, 12 months). To analyze the melanosome localization in the RPE of pigmented sh1 2/2 or sh1 +/2 mice (Fig. 7A-B), eyes were enucleated 4 months following the AAV subretinal injection, fixed in 2% glutaraldehyde-2% paraformaldehyde in 0.1 M phosphate buffer over-night, rinsed in 0.1 M phosphate buffer and dissected Figure 6. Subretinal delivery of AAV encoding human Myo7a rescues Rhodopsin accumulation in sh1 2/2 photoreceptor connecting cilium. A. Quantification of the gold particle density (corresponding to Rhodopsin molecule density) in the connecting cilium of sh1 mice. The quantification is depicted as relative gold density (average gold particle density in sh1/average gold particle density in sh1 +/2 mice). Albino sh1 mice were analyzed at 4 months following the injections; pigmented sh1 mice were analyzed at 4 or 7 months following the injections. Data are presented as mean6SEM. Melanosomes were counted by a blinded observer who did not know which retinas were treated with AAV-hMYO7A and which with control vectors. More specifically, the eyes were collected and the slides were numbered by a person different from the observer who counted the melanosomes. The observer analyzed 10 different fields/eye under a light microscope at 100X magnification. Retinal pictures were captured using a Zeiss Axiocam (Carl Zeiss, Oberkochen, Germany). The number of sh1 mice analyzed in Figure 7A is the following: 3 sh1 +/2 vs 4 sh1 2/2 .

Immunogold and Electron Microscopy
The data showed in Figure 6A were obtained from albino and pigmented sh1 2/2 and sh1 +/2 mice. To quantify Rhodopsin localization to the connecting cilium of sh1 photoreceptors, eyes were enucleated from albino sh1 mice at 4 months following AAV injection and from pigmented sh1 mice at 4 or 7 months following AAV injection. Eyes were fixed in 0.  dissected under a fluorescence or light microscope to select the EGFP-or tyrosinase-positive portions of the eyecups, respectively, which were subsequently embedded in 12% gelatin. Cryo-sections (50 nm) were cut using a Leica Ultramicrotome EM FC7 (Leica Microsystems, Bannockburn, IL, USA) and extreme care was taken to align photoreceptor connecting cilia longitudinally. The anti-Rhodopsin antibody (1:100 1D4, ab5417, Abcam, Cambridge, UK) was used for immunogold labeling. The quantification of gold density of Rhodopsin in the connecting cilia was performed by a blinded observer who did not know which retinas were treated with AAV-hMYO7A and which with control vectors. More specifically, the eyes were collected and the slides were numbered by a person different from the observer who quantified gold density. The analysis was performed using the iTEM software (Olympus SYS, Hamburg, Germany) according to a previously described method [35]. Briefly, a morphometric grid with a 100 nm mesh was placed over images of photoreceptors. Touch counts module of the iTEM software was used to quantify: i. the number of gold particles in the cilium and ii. the number of intersections between grid and cilium membrane. Gold density was expressed in arbitrary units (gold particles per intersection). The number of sh1 mice analyzed in Figure 6A is the following: 2 sh1 +/2 vs 2 sh1 2/2 (albino); 2 sh1 +/2 vs 2 sh1 2/2 (pigmented).

Optical Coherence Tomography (OCT) and and Scanning Laser Ophthalmoscopy (SLO)
Mice were anaesthetized by intraperitoneal injection of a combination of ketamine (Virbac, Carros, France) at 150 mg/kg body weight and xylazine (12.5 mg/kg; KVP Pharma und Veterinaer-Produkte GmbH) at 10 mg/kg body weight. Pupils were dilated with tropicamide eye drops (Mydriaticum Stulln, Pharma Stulln, Stulln, Germany). In vivo imaging was performed as previously described [36]. Briefly, SLO and SD-OCT imaging were performed with a Spectralis TM HRA+OCT (Heidelberg Engineering GmbH, Heidelberg, Germany). This device features a superluminescent diode at 870 nm as low coherent light source. Each A-scan was taken at a speed of 40.000 scans per second and each B-scan contains up to 1536 A-scans. The images were taken with the equipment set of 30u field of view and with the software Heidelberg Eye Explorer (HEYEX version 3.2.1.0, Heidelberg, Germany) and they were processed with Corel Draw X3 (Corel Corporation, Ottawa, ON Canada). Age-matched C57BL/6 mice (purchased from Charles River Laboratories International, Inc.,Wilmington, MA, USA) were kept under a 12 h:12 h light-dark cycle (60 lux) and had food and water ad libitum. All procedures were performed with permission of the local authorities (Regierungspraesidium, Tübingen, Germany).

B-wave Electroretinogram is Decreased in Albino sh1 2/2 Mice
In all our studies we compared sh1 2/2 mice, whether albino or pigmented, to their heterozygous (sh1 +/2 ) littermates. Ganzfeld flash electroretinograms (ERGs) in both photopic and scotopic conditions were recorded in albino sh1 2/2 and sh1 +/2 mice at 3, 6 and 12 months of age ( Fig. 1 and S1). At 3 months, no major abnormalities in the a-and b-wave amplitudes were observed in sh1 2/2 compared to age-matched sh1 +/2 mice (Fig. S1A). However a significant scotopic b-wave reduction was observed in affected sh1 2/2 mice compared to heterozygous controls at 6 To assess statistical significance Two-way ANOVA was used: the statistical significant differences between sh1 +/2 and sh1 2/2 mice from the post-hoc Multiple Comparison Procedure are depicted: *p value,0.05; ** and # p,0.001. Asterisks label significant differences between hMYOA-and EGFPtreated eyes; pound keys label significant differences between EGFP-treated eyes and sh1 +/2 eyes. More details on the statistical analysis can be found in the Statistical analysis paragraph of the Materials and Methods section. AAV-hMYO7A: AAV2/5-CMV-hMYO7A; AAV-EGFP: AAV2/5-CMV-EGFP. doi:10.1371/journal.pone.0072027.g008 months (20% reduction at 1.3 log) and 12 months (27% reduction at 1.3 log, Fig. S1A and Table 1). No a-wave abnormalities were instead observed in albino affected mice up to 12 months compared to heterozygous littermates (Fig. 1A). Notably, in both affected and control albino sh1 mice a-and b-waves declined over time ( Table 1). From 6 to 12 months of age the maximum b-wave amplitude (evoked with 20 cd*s/m 2 in scotopic conditions) decreased of 40% in sh1 +/2 (ANOVA p value,,2610 216 ) and of 45% in sh1 2/2 (ANOVA p value = 1.6610 29 ) ( Table 1) while the maximum a-wave amplitude decreased of 23% in sh1 +/2 (p value = 0.03) and 28% in sh1 2/2 (p value = 0.0075) ( Table 1). In pigmented sh1 2/2 and sh1 +/2 mice, we recorded ERGs between 6 and 7 months and at 12 months of age and did not observe significant differences in affected mice compared to controls ( Fig.  S2 and Table 1). Differently from albino sh1 mice, both pigmented sh1 +/2 and sh1 2/2 mice did not show significant maximum ERG reduction from 6-7 to 12 months of age ( Table 1). Interestingly, the b-wave defects identified by the ERG analysis in albino sh1 2/2 showed that the albino but not the pigmented background negatively influenced retinal function of Myo7a deficient mice (Fig. 1A, S2 and Table 1) as observed in previous studies performed in other pigmented shaker1 lines [7,9].

Light Sensitivity Decreases with Age in Albino sh1 2/2 Mice
Based on the ERG reduction measured in albino sh1 2/2 mice we further investigated their retinal function. Rod sensitivity to light was evaluated in scotopic conditions in dark-adapted albino sh1 2/2 and sh1 +/2 mice by measuring the lowest light stimulus able to evoke a b-wave-shaped response. Rod sensitivity is low when the intensity of the light stimulus (defined as b-wave threshold) that evokes the b-wave is high. Notably, the b-wave thresholds were significantly higher in albino sh1 2/2 mice compared to the corresponding sh1 +/2 controls (Fig. 2). In particular albino sh1 2/2 showed an early onset reduction of rod sensitivity to light ( Fig. 2 and S1B) that further declined with age reaching about 67% of heterozygous controls at 12 months. Therefore, albino sh1 2/2 mice presented with a decrease in rod sensitivity to light.

Recovery from Light Desensitization is Decreased in Albino sh1 2/2 Mice
We then evaluated the ability of rod photoreceptors to recover from light desensitization in albino sh1 2/2 mice and corresponding controls at 2 (Fig. 3A), 6 (Fig. 3B) and 12 (Fig. 3C) months of age. We measured the b-wave amplitude evoked by a light stimulus of 1 cd*s/m 2 before and after rod desensitization. To desensitize rods we exposed the eyes to pre-adapting light for 3 minutes, after desensitization we stimulated rods with a light stimulus of 1 cd*s/m 2 over time (0, 5, 15, 30 45 and 60 minutes) and measured the b-wave amplitude evoked (Fig. 3). Sixty minutes after desensitization albino sh1 +/2 mice generated a bwave similar to the one generated before desensitization, therefore the ratio between the post-and pre-bleaching b-waves is around 1 (Fig. 3). The ability of the albino sh1 +/2 retina to recover from light desensitization decreased with age (Fig. 3): b-wave ratio is 0.9760.01 (mean6SEM) at 2 months and 0.8260.04 (mean6-SEM) at 12 months (ANOVA p value = 0.03). Remarkably, albino sh1 2/2 mice showed an early severe impairment of the rod recovery from light desensitization (Fig. 3), indeed at 2 months of age the affected mice only recovered about 50% of the predesensitization b-wave at 60 minutes (Fig. 3). Notably, the impairment in recovery from light desensitization of the sh1 2/2 retina is stable from 2 to 12 months of age (mean b-wave ratio6SEM = 0.5060.03 at 2 months vs 0.5460.02 at 12 months, ANOVA p value = 0.3) (Fig. 3). This early and marked impairment of recovery from light desensitization by the sh1 2/2 retina is the first important functional defect associated with Myo7a deficiency.

No Significant PR Degeneration in Albino and Pigmented sh1 2/2 Mice
To date no PR degeneration has been observed in shaker1 murine lines that bear various loss-of-function mutations in the Myo7a gene, including the sh1 2/2 mice we used in our studies [8,9,10]. To confirm this in our lines of albino and pigmented sh1 2/2 mice, we counted the rows of PR nuclei in the outer nuclear layer (ONL) using retinal sections from affected and control sh1 mice that were previously subjected to electrophysiological analyses (Fig. 4 and Table 1). The rows of PR nuclei of both albino and pigmented sh1 2/2 mice were not significantly different from those of control sh1 +/2 mice at 6-7 and 12 months of age ( Fig. 4 and Table 1). In addition, retinal optical coherence tomography (OCT) and scanning laser ophthalmoscopy (SLO) imaging (Fig. S3) were performed in a subset of 12 month-old pigmented sh1 mice before they were sacrificed for the histological analysis shown in Figure 4. The OCT and SLO analyses showed that the ONL thickness as well as the retinal architecture was not different between sh1 2/2 and sh1 +/2 mice (Fig. S3), thus confirming the data from the histological analysis (Fig. 4). Therefore the electrophysiological defects we observed in the albino sh1 2/2 retina are not due to significant PR cell loss but are potentially due to functional impairments that result from lack of Myosin7a-mediated motor kinesis.  To rescue the retinal phenotype of adult sh1 2/2 mice we performed MYO7A gene replacement using adeno-associated viral vectors (AAVs), which efficiently transduce both RPE and PRs [37,38]. We generated AAV2/2 and 2/5 vectors by using the pAAV2.1-CMV-hMYO7A cis-plasmid that contains the complete human MYO7A (hMYO7A) coding sequence (6648 bp) under the control of the ubiquitous cytomegalovirus (CMV) promoter, which is active in both PRs and RPE [37,38]. This 8.1 Kb expression cassette exceeds the canonical AAV cargo capacity and results in the generation of the so called ''oversize'' AAV vectors [27,32,33,39,40,41,42,43,44,45]. Oversize AAV vectors have been reported to contain genomes of heterogeneous size [41,42,43,44,45,46,47] which may preclude their clinical application. Indeed, we have analyzed by alkaline Southern Blot the size of DNAse I-resistant viral genomes from oversize AAV2/5 vectors generated by using various pAAV2.1 cis-plasmids containing large transgenes (ranging from 7.7 to 8.9 Kb in size, see Materials and Method section). While we observe a low amount of encapsidated genomes .5 kb, which appear more heterogeneous in size than what we have observed in the past [27], we have found a large population of smaller genomes #5 kb, similarly to what reported by others (Fig. S4) [41,42,43,44,45,46]. Similar results were obtained analyzing both AAV2/2-and 2/5-CMV-hMYO7A vectors (data not shown). Despite this limitation, oversize AAV vectors are useful research tools for large gene expression in photoreceptors and retinal pigment epithelium in vivo [27,48].
To assess the expression of hMyo7a in vitro we used AAV2/2 vectors which infect HEK293 cells more efficiently than AAV2/5 [49]. Oversize AAV2/2 vectors were used to infect HEK293 cells followed by Western blot analyses on cell lysates using anti-Myo7a antibody (Fig. 5A). As negative control we infected cells with AAV2/2 vectors that encode for the enhanced green fluorescent protein (EGFP), while as positive control we transfected HEK293 cells with the same pAAV2.1-CMV-hMYO7A cis-plasmid used for oversize AAV production (Fig. 5A). Notably, infection of HEK293 cells with AAV2/2-CMV-hMYO7A resulted in robust expression of full-length hMyo7a protein (Fig. 5A). We obtained a similar result by Western blot analysis with anti-HA antibodies of cells infected with AAV2/2-CMV-hMYO7A-HA vector which contains the HA tag sequence at the 39end of the hMYO7A coding sequence. This allows us to exclude that the strong AAV-mediated expression of hMyo7a in vitro may be due to the integration of the transgene hMYO7A 59 end containing the CMV promoter at the endogenous HEK293 MYO7A locus (Fig. S5). Then, we injected subretinally AAV2/5-CMV-hMYO7A in one eye and AAV2/5-CMV-EGFP in the contralateral eye of albino sh1 2/2 mice (Fig. 5B). Fifteen months after injection, the eyecups were harvested, lysed and analyzed by Western blot analysis using the anti-Myo7a antibody. We observed by Western blot analysis that the full-length hMyo7a protein was expressed at various levels in the eyecups of sh1 2/2 mice injected with AAV2/5-CMV-hMYO7A, but not with AAV2/5-CMV-EGFP (Fig. 5B). This hMyo7a expression presumably derives from transduction of both RPE and photoreceptors since we have previously reported that oversize AAV2/5 vectors injected subretinally transduce both cell types [27].
We then evaluated the therapeutic efficacy of AAV-mediated hMYO7A transfer to the sh1 2/2 retina (Fig. 6, 7 and 8). We initially analyzed the subcellular localization of Rhodopsin (rho) in sh1 2/2 PRs by immunogold analysis (Fig. 6A-B). Consistently with what previously reported [14], we found a significant accumulation of rho at the connecting cilium of sh1 2/2 mice compared to their heterozygous littermates (Fig. 6A-B). Following AAV2/5-hMYO7A delivery the number of rho molecules at the connecting cilium of PRs in the transduced area was significantly lower than in sh1 2/2 mice injected with AAV2/5-EGFP (Fig. 6A-B). Then, we evaluated the localization of melanosomes in the RPE of pigmented sh1 2/2 and sh1 +/2 mice treated with AAVs by histological analysis of semi-thin retinal sections (Fig. 7A-B). Melanosome counts (Fig. 7B) and representative retinal images (Fig. 7B) show that differently from sh1 +/2 , the melanosomes of sh1 2/2 mice do not enter in the RPE microvilli, and that this is corrected by AAV2/5-CMV-hMYO7A injection (Fig. 7A-B). Figure 7B, shows that in the area treated with AAV2/5-CMV-hMYO7A many RPE cells present correctly localized melanosomes (Fig. 7B). Although the number of AAV-hMYO7A-treated retinas is low, the apical position of melanosomes in the sections from this retina appears to be due to hMYO7A gene replacement as none of the EGFP-treated eyes shows a single correctly-localized melanosome ( Fig. 7A-B). Finally, we measured the ability of the albino sh1 2/2 retina to recover from light desensitization following injection of either AAV2/5-CMV-hMYO7A or AAV2/5-CMV-EGFP as control (Fig. 8). This analysis can not be performed in pigmented sh1 mice which do not have defects in the recovery from light desensitization compared to sh1 +/2 controls (data not shown). At 6 months post-injection (Fig. 8) retinal recovery from light desensitization of sh1 2/2 hMYO7A-treated eyes was significantly different from that of EGFP-treated eyes (ANOVA p value = 9.7610 29 ) but not from those of sh1 +/2 eyes (ANOVA p value = 0.18; Fig. 8). A similar result was observed one month after AAV2/5-CMV-hMYO7A subretinal administration in 7month-old sh1 2/2 mice (data not shown). Thus, recovery from light desensitization is a consistent and sensitive test to assess the outcome on retinal function of experimental therapies for USH1B in sh1 2/2 mice.

Discussion
Shaker1 mice are the best characterized model of USH1B, one of the most common severe forms of syndromic retinitis pigmentosa. While sh1 2/2 mice have clear ultrastructural PR and RPE anomalies, no important functional defects of the sh1 2/2 retina have been identified yet [7,8,9,14,16]. Here we show that Myo7a deficiency causes severe retinal dysfunctions in albino sh1 2/2 mice.
The most significant retinal functional anomaly we found in albino sh1 2/2 mice is the 50% reduction in recovery from light desensitization that is evident from 2 months of age. Light desensitization saturates active Rhodopsin molecules in the PR OSs, therefore the recovery of rod dark adaptation mainly relies on the efficient regeneration of Rhodopsin that depends on: i. transport of new opsin molecules from the PR connecting cilium to their OS; and ii. recycling of 11-cis-retinal (the opsin cofactor) by the visual cycle [50]. It is likely that the slower opsin transport as well as the slower turnover of PR OS reported in sh1 2/2 mice [14] contributes to delayed recovery after light desensitization. RPE defects could additionally contribute to the impairment in the recovery from light desensitization, since Lopes et al. have recently shown that the expression of Myo7a in RPE cells is required for the proper activity of the RPE65 isomerase which is a key enzyme for the generation of the 11-cis-retinal in the visual cycle [9]. Thus, both PR and RPE defects of sh1 2/2 mice could have a functional counterpart in the inability of the sh1 2/2 retina to recover from light desensitization. Independently of the mechanism, reduced levels of Myo7a negatively influence the ability of rods to recover from light desensitization in albino sh1 mice. Whether this is common to other shaker1 lines, that bear alleles other than the 4626SB, remains to be tested. In addition, it remains to be assessed how this defect extrapolates to the human USH1B retina that, differently from the mouse, presents with severe retinitis pigmentosa characterized by photoreceptor degeneration and decreased Ganzfeld ERGs [1,20]. No abnormalities in the dark adaptation kinetics have been recently found in USH1B patients [20]. However, delayed recovery of rod sensitivity after light desensitization has been reported in carrier and affected USH patients [51,52] as well as in individuals with retinitis pigmentosa [53,54,55]. In any case the recovery from light desensitization represents an important functional read-out to test the efficacy of experimental therapies for the USH1B retina at the pre-clinical level. We also found that rod light sensitivity is significantly reduced in affected sh1 mice compared to heterozygous controls. As Rhodopsin transport to the outer segment has been reported to be reduced in the absence of Myo7a [14], it is possible that more photons are required to evoke a b-wave response in presence of a lower number of Rhodopsin molecules. A previous study in pigmented sh1 2/2 mice on a mixed background reported no significant reduction of light sensitivity but attenuated a-and bwave amplitudes at higher light intensities (about 20% reduction at 0.0 log) [7]. We did observe a subtle a-and b-wave reduction at 0.0 log in pigmented sh1 2/2 compared to sh1 +/2 but this was not statistically significant. We think that this can be ascribed to the difference in the sh1 2/2 background and/or to the very subtle entity of the ERG reduction in pigmented sh1 2/2 mice, which can be appreciated only when a very high number of eyes are analyzed.
On the other hand we observed a mild (approximately 20%), although significant, Ganzfeld ERG b-wave reduction in albino sh1 2/2 mice after 6 months of age. However, we consider the entity of the ERG reduction observed in albino sh1 2/2 mice too small to be reliably used to assess rescue of retinal function in gene therapy studies. Notably, the ERG phenotype of sh1 2/2 mice is not due to progressive photoreceptor degeneration since we did not found significant photoreceptor cell loss in affected mice compared to controls. The molecular basis of the ERG reduction in sh1 2/2 mice has not been thoroughly dissected yet however, it is likely that the sh1 2/2 PR and RPE defects, as well as the impaired rod dark adaptation could account for the Ganzfeld ERG phenotype. Recently Peng et al. [10] reported that the translocation of transducin from the OS to the inner segment of PRs is delayed in Myo7a sh1-11J mice. If this holds true for the albino sh1 2/2 mice we used in our studies, it could contribute to retinal dysfunction since transducin is a key component of the phototransduction cascade and its translocation is important for PR adaptation to various light conditions [50,56]. As alternative explanation for the ERG reduction, we excluded the difference in size between sh1 2/2 and sh1 +/2 as this difference in size is present also on the pigmented background which does not present significant ERG reduction.
Interestingly, as previously reported in other studies [8,9] we observed that Myo7a deficiency did not impact on PR survival since the rows of PRs nuclei of affected mice were not significantly different from those of heterozygous controls from both albino and pigmented lines. Notably, we observed that the albino background which is known to exacerbate PR degeneration in other murine models of PR disease [57] did not influence PR cell loss in sh1 2/2 mice. This may be explained by the lower activity of RPE65 found in sh1 2/2 mice [9] which may render Myo7a-deficient PRs more resistant to the damaging effect of ambient light as it makes them extremely resistant to acute light damage [9].
The sh1 2/2 mice are crucial to evaluate the efficiency and safety of experimental therapies for USH1B. Retinitis pigmentosa in USH1B is severe leading to blindness in the first decades of life [1]. Gene therapy with vectors derived from AAV is being evaluated in patients with an inherited form of childhood blindness and the safety and the efficacy results from three independent clinical trials are extremely encouraging [58].
However, AAV vectors have a cargo capacity considered to be limited to 4.7 Kb [41,42,43,44,45,46,47] and this is a major limitation when delivering large genes like MYO7A. This can be overcome by either generating oversize AAV vectors by using a plasmid containing a large gene expression cassette [27,32,33,39,40,41,42,43,44,45] or by splitting a large gene into two halves each contained in a separate, regular size AAV vector (dual-AAV vectors [59]). The potential of dual-AAV vectors in the retina has been underlined using a reporter gene [60]. However no therapeutic applications of this platform in the retina have been reported yet. Gene transfer with oversize AAV vectors has resulted in expression of large proteins and subretinal administration of oversize AAV2/5 vectors has allowed expression in both RPE and photoreceptors of the large MYO7A and ABCA4 [27,48] genes at therapeutic levels in mice.
In a previous study [27] we have shown that oversize AAVs express the full-length hMyo7a protein in RPE cultures from sh1 2/2 mice. In the present study we report that AAV-mediated expression of hMyo7a in the retina of adult sh1 2/2 mice rescues PR Rhodopsin transport and RPE melanosome localization and improves the recovery profile from light desensitization, thus validating this as a sensitive electrophysiological test to assess improvement of retinal function in pre-clinical gene therapy studies for the USH1B retina.
Our current data as well as those published by others [41,42,43,44,45,46,47] show that the size of the genome contained within oversize AAV particles is highly heterogeneous, and mostly smaller than expected. However, after infection of target cells, AAV DNA is found to have an homogenous size which corresponds to the large transgene expression cassette [27,45]. Accordingly, the corresponding mRNA and protein are full size without evidence so far of truncated products [44,45]. Therefore, based on the results presented here and by others the molecular mechanism underlying oversize AAV-mediated transduction remains elusive. Large gene transduction may derive from: i. full-length genomes contained in the capsid [27,32,39,40]; ii. reassembly in transduced cells of truncated genomes [41,43,44,45,46,47] produced during oversize genome encapsidation; iii. a combination of the two previous mechanisms. Independently of the mechanism underlying oversize AAV vector transduction, the heterogeneous size of their genomes may limit their therapeutic applications. However oversize AAVs represents valuable research tools for gene expression in the retina and more specifically in adult photoreceptors where other vector systems may have limited access [61]. Indeed previous gene transfer studies with lentiviral vectors which have higher cargo capacity than AAV but poor PR transduction ability [62,63] have shown that hMYO7A gene delivery to the newborn sh1 2/2 retina results in therapeutic efficacy [2]. However, the ability of lentiviral vectors to efficiently transduce adult murine PRs remains to be demonstrated. In conclusion, the current study provides important tools to assess the rescue of retinal function in sh1 2/2 mice in pre-clinical gene therapy studies and further validates sh1 2/2 as a model of USH1B. In addition, it provides proof-of-principle that gene transfer improves the phenotype of the sh1 2/2 retina including its functional abnormality. Once a vector platform capable of safely and efficiently transfer large genes to photoreceptors and RPE will be made available, our data prompt its testing in the sh1 2/2 retina for further development of gene therapy for USH1B. Figure S1 Ganzfeld electroretinograms (A) and retinal light sensitivity (B) of albino sh1 mice at 3 months of age. A-B. Data are presented as mean6SEM, n indicates the number of eyes analyzed, the arrows point at the photopic ERG. *p value,0.05. More details on the statistical analysis including specific statistical values can be found in the Statistical analysis paragraph of the Materials and Methods section. (TIF) Figure S2 Ganzfeld electroretinograms in pigmented sh1 mice. A. Mean a-(top panels) and b-waves (bottom panels) from 6-7-(left panels) and 12-(right panels) month-old sh1 mice. Data are presented as mean6SEM, n indicates the number of eyes analyzed, the arrows point at the photopic ERG. No statistically significant differences were found between sh1 +/2 and sh1 2/2 mice. More details on the statistical analysis including specific statistical values can be found in the Statistical analysis paragraph of the Materials and Methods section. B. Representative scotopic ERG waves from one sh1 +/2 and one sh1 2/2 mouse at 12 months of age. (TIF) Figure S3 In vivo retinal imaging in pigmented sh1 mice. Retinal optical coherence tomography (OCT) and scanning laser ophthalmoscopy (SLO) imaging were performed in sh1 +/2 control (A-C) and homozygous sh1 2/2 (D-F) mice at 12 months of age. The pictures depicted are representative of four sh1 +/2 mice (4 eyes analyzed) and two sh1 2/2 mice (2 eyes analyzed). The retinal layering was also compared to wild-type C57BL/6 mice (G). There were no fundus abnormalities visible in both sh1 +/2 control and sh1 2/2 affected mice in infrared mode (820 nm; A, D). Western blot analysis of lysate eyecups from albino sh 2/2 and sh1 +/ 2 mice using the anti-Myo7a antibody. The murine Myo7a protein was clearly detected in the eyecups of heterozygous sh1 +/2 mice (n = 2) but was not in the eyecups of sh1 2/2 mice (n = 2). The anti-Dysferlin antibody was used as loading control. One hundred mg of each eyecup lysate were loaded. The arrow points at the Myo7a protein.