Use of neuroglobin agonist for preventing or treating mitochondrial RCCI and/or RCCIII deficiency disease

20220040333 · 2022-02-10

    Inventors

    Cpc classification

    International classification

    Abstract

    The present invention concerns a neuroglobin agonist for use in the treatment or prevention of a mitochondrial disease associated with respiratory chain complex I (RCCI) deficiency and/or respiratory chain complex III (RCCIII) deficiency.

    Claims

    1. A method for preventing or treating a mitochondrial disease associated with respiratory chain complex I (RCCI) deficiency and/or respiratory chain complex III (RCCIII) deficiency in a subject having or at risk of having such a disorder, comprising administration of a therapeutically effective amount of a neuroglobin agonist to the subject.

    2. The method of claim 1, wherein said mitochondrial disease is associated with Apoptosis Inducing Factor (AIF) deficiency.

    3. The method of claim 1, wherein said mitochondrial disease is not associated with a mutation or deletion of neuroglobin gene or is not induced by neuroglobin deficiency.

    4. The method of claim 1, wherein said neuroglobin agonist is a nucleic acid which comprises an expression cassette comprising a polynucleotide encoding neuroglobin protein, said polynucleotide being operatively linked to at least one transcriptional regulatory sequence and wherein said method further comprises the step of expressing said polynucleotide in said subject.

    5. The method of claim 1, wherein said neuroglobin agonist is a polypeptide selected from the group consisting of a dominant activated mutant of neuroglobin, a wild-type neuroglobin protein, a fragment and a peptidomimetic thereof.

    6. The method of claim 1, wherein said method is for restoring or improving RCCI and/or RCCIII function.

    7. The method of claim 1, wherein said mitochondrial disease associated with RCCI deficiency and/or RCCIII deficiency is a neurodegenerative disease or an ocular disease.

    8. The method of claim 1, wherein said mitochondrial disease associated with RCCI deficiency and/or RCCIII deficiency is a retinal disease.

    9. The method of claim 4, wherein said at least one transcriptional regulatory sequence comprises 3′UTR and/or 5′UTR neuroglobin sequence.

    Description

    BRIEF DESCRIPTION OF THE FIGURES

    [0116] FIG. 1: Schematic representation of mitochondrial extraction from rat retinas. Successive steps during the process of purification will lead to a mitochondrial enrichment in the ultimate fraction. Different samples were evaluated by western blot analysis for the presence of Ngb or mitochondrial proteins.

    [0117] FIG. 2: In vivo impact of neuroglobin down regulation on NGB and CYGB mRNA RT-qPCR assays were performed with total RNAs from retinas (isolated from rats 3 months after the eye electroporation, ELP) to determine the steady-state levels of NGB and CYGB mRNAs. 18 control eyes, 8 eyes treated with scrambled shRNA and 10 eyes treated with anti-Ngb shRNA were evaluated (mean±S.E.M presented). Relative fold variations were calculated using the comparative ΔΔCt method and the mitochondrial ATP6 gene as a normalizing gene. Primers used for each gene are shown in Table 2.

    [0118] FIG. 3: In vivo impact of neuroglobin down regulation on SNCG and BRN3A mRNAs RT-qPCR assays were performed with total RNAs from retinas (isolated from rats 3 months after the ELP) to determine the steady-state levels of SNCG and BRN3A mRNAs. 18 control eyes, 8 eyes treated with scrambled shRNA and 10 eyes treated with anti-Ngb shRNA were evaluated (mean±S.E.M presented). Relative fold variations were calculated using the comparative ΔΔCt method and the mitochondrial ATP6 gene as a normalizing gene. Primers used for each gene are shown in Table 2.

    [0119] FIG. 4: Detection of respiratory chain complex activities in optic nerves Illustrative curve for the successive measurements of complex I (EC 1.6.5.3) and complex V activities (EC 3.6.3.14); complex IV (CIV) (EC 1.9.3.1), complex II+III (CII+CIII) (EC 1.3.2.2) and complex III (CIII) (EC 1.10.2.2).

    [0120] FIG. 5: Assessment of visual performance after neuroglobin knockdown Four different groups of rats were evaluated for head tracking movements at an angular speed identical to that of the drum rotation: 22 animals 8 week-old (before the ELP) and 10 control rats 18 week-old. No significant difference was measured in the three grating frequencies monitored in any of the animals examined in both clockwise and counterclockwise directions of motion.

    [0121] FIG. 6: Assessment of visual performance after neuroglobin knockdown 22 rats treated with anti-Ngb shRNA, 20 rats treated with scrambled shRNA. An unequivocal and significant decline in visual performance was measured in rats treated with anti-Ngb shRNA which responded with poor clockwise scores. Rats treated with scrambled shRNA showed no significant difference between clockwise and counterclockwise responses. Mean values±S.E.M are presented.

    [0122] FIG. 7: Cellular distribution and relative amounts of NGB in retinas from Harlequin and control retinas: RT-qPCR assays were performed using total RNAs from retinas (isolated from mice 6 months old) to determine the steady-state levels of NGB mRNA. RNAs purified from 31 control and 37 Hq retinas were evaluated. Relative fold mRNA variations were calculated using the comparative ΔΔCt method and the mitochondrial ATP6 gene as a normalizing gene and are represented relative to the value assessed in RNAs from control retinas considered as 1; each value is the mean of all the assessments±S.E.M. Control and Harlequin NGB mRNA abundance were compared with the unpaired non parametric significance test of Mann-Whitney. Primers used for the NGB gene are shown in Table 7.

    [0123] FIG. 8: Physical Map of the AAV2/2-NGB vector genome (7255 bp), encompassing mouse NGB sequences inserted into the pAAV-IRES-hrGFP plasmid: The NGB ORF (453 bp) encoding 151 amino acids is in frame with three FLAG epitopes and transcribed under the control of the cytomegalovirus promoter (pCMV). The construction contains both UTRs (UnTranslated Regions) at the 5′ (279 bp) and 3′ (895 bp) ends of the mouse NGB mRNA (NM_022414.2). The plasmid possesses also a cassette allowing the expression of the recombinant humanized green fluorescent protein (hrGFP).

    [0124] FIG. 9A-B: AAV2/2-NGB vector generation and administration in Harlequin eyes: RT-qPCR assays were performed with total RNAs extracted from control retinas and 14 pairs of retinas isolated from Hq mice in which one eye was subjected to intravitreal injection of AAV2/2-NGB (mice were euthanized between 5 to 6 months after vector administration). The steady-state levels of NGB and AIF1 mRNAs were compared to those obtained in RNA preparations from 18 age-matched control mice. Two pairs of specific primers were used for NGB mRNA: one recognized the two NGB transcripts (the endogenous and the one issued from the vector) and the second pair recognized exclusively the molecule transcribed from the vector since it is located between the end of the ORF and the Flag epitope sequence (Table 7, NGB-AAV). Histograms show the steady-state levels of AIF1 mRNA and of NGB mRNAs calculated by the comparative ΔΔCt method and the mitochondrial ATP6 gene as a normalizing gene. Values shown are relative to the mRNA amounts measured in control retinas and considered as 1; each value is the mean of all the triplicates obtained from each biological sample±S.E.M. Statistical significance was determined using the paired non parametric test of Wilcoxon.

    [0125] FIG. 10: GFAP expression in retinas from Harlequin and control eyes: RT-qPCR assays were performed with RNAs purified from 14 control retinas and from 9 pairs of retinas isolated from Hq mice in which one eye was subjected to intravitreal injection with AAV2/2-NGB (mice were euthanized between 5 to 6 months after vector administration). Relative GFAP mRNA variations were calculated using the comparative ΔΔCt method and the mitochondrial ATP6 gene as a normalizing gene. Values shown in the histogram are relative to the GFAP mRNA amount measured in control retinas and considered as 1; each value is the mean of all the triplicates obtained from each biological sample±S.E.M. Statistical significance was determined using the paired non parametric test of Wilcoxon. Primers used for the GFAP gene are shown in Table 7.

    [0126] FIG. 11A-B: Retinal ganglion cell evaluation in retinas from Harlequin eyes treated with AAV2/2-NGB vector: (A) RGC numbers were estimated in Hq treated and untreated eyes as well as in age-matched controls by immunolabeling for BRN3A and DAPI staining of retinal sections. BRN3A and DAPI-positive cells in the GCL were counted in 3-4 independent sections per eye for 24 control eyes and 13 Hq pairs of eyes in which only one was subjected to AAV2/2-NGB injection. Harlequin mice were euthanized between 5 to 6 months after vector administration and controls were aged about 7 months when euthanized. Histogram illustrates the results (mean values±S.E.M) corresponding to the overall RGC density (total number of BRN3A-positive cells per millimeter). Statistical significance was determined using the paired non parametric test of Wilcoxon. (B) RT-qPCR assays were performed with total RNAs extracted from control retinas and 14 pairs of retinas isolated from Hq mice in which one eye was subjected to intravitreal injection of AAV2/2-NGB (mice were euthanized between 5 to 6 months after vector administration). The steady-state levels of SNCG mRNA were compared to those obtained in RNA preparations from 18 age-matched control mouse retinas. Histogram shows the steady-state levels of SNCG mRNA calculated by the comparative ΔΔCt method and the mitochondrial ATP6 gene as a normalizing gene. Values shown are relative to the steady-state mRNA levels measured in control retinas and considered as 1; each value is the mean of all the triplicates obtained from each biological sample±S.E.M. Statistical significance was determined using the paired non parametric test of Wilcoxon.

    [0127] FIG. 12A-B: Morphological and functional evaluation of optic nerves from Harlequin mice after ocular AAV2/2-NGB treatment: Specific complex V (CV) and complex I (CI) enzymatic activities were assessed in single optic nerves isolated from 36 control mice, and from 24 Hq mice in which one eye was subjected to AAV2/2-NGB intravitreal injection (treated) and the contralateral one remained untreated. The successive measurements of CI and CV activities were expressed as nanomoles of oxidized NADH/min/mg protein. Histograms illustrate complex V activity (A) and complex I (B) as mean±S.E.M of each assay measured in triplicate. Values obtained in ONs from Hq mice subjected to vector administration, from untreated Hq mice and from age-matched controls were compared unpaired non parametric significance test of Mann-Whitney (*≤0.05, **≤0.01 and ***≤0.005); while data collected from Hq treated eyes and untreated controlateral eyes were compared using the paired non parametric significance test of Wilcoxon (*≤0.05, **≤0.01 and ***≤0.005).

    [0128] FIG. 13A-B: Preservation of nerve fibers in AAV2/2-NGB treated eyes protects Harlequin mouse vision: The Optomotry™ set-up allowed the determination of visual acuity threshold measurements (cycles per degree) for left and right eyes, independently scored (clockwise and couterclockwises responses) under photopic conditions. Visual acuities (right and left eye sensitivities) for Hq and control mice aged 4-8 weeks (n=7) are illustrated in (A) as means±S.E.M of measures performed twice 4-6 days apart. (B) Histogram shows visual acuities (right and left eye sensitivities) for 7 month-old control mice (n=22) and 18 Hq mice subjected to AAV2/2-NGB injection in their left eyes. Hq mice were evaluated 3 and 6 months post-injection (the test was performed twice each time 4-6 days apart), values represented are the means±S.E.M of measures performed 6 months after AAV2/2-NGB administration. Data collected from control and Hq were compared using the unpaired non parametric significance test of Mann-Whitney (*≤0.05, **≤0.01 and ***≤0.005). Data collected from Hq treated eyes and untreated contralateral eyes were compared using the paired non parametric significance test of Wilcoxon (*≤0.05, **≤0.01 and ***≤0.005).

    [0129] FIG. 14A-B: Representative transmission electron microscopy micrographs of optic nerve sections from Harlequin and age-matched control mice: (A) Photomicrographs of longitudinal and transverse optic nerve sections taken from two control mice aged 1 and 12 months showing parallel running myelinated axons (longitudinal sections) and compact amenagement of the myelin lamellae around the axons (transverse sections). Scale bars, in function of images, are equivalent to 1, 2 or 5 μm. (B) Photomicrographs of longitudinal and transverse optic nerve sections from four Hq mice aged 1, 3, 6 and 12 months. Scale bars, in function of images, are equivalent to 1, 2 or 5 μm. The only change seen in the 1-month Hq mouse relative to controls is the presence of few swollen axons; at three months of age the alterations on Hq optic nerves exacerbated. It was noticed fibers undergoing degeneration with axons showing hyperdense axoplasms (dark) and abnormal accumulation of altered organelles. By the age of 6 and 12 months, longitudinal profiles of Hq mice show focal axonal swelling and hyperdense axoplasms. Furthermore, towards the end stage of degeneration nearly collapsed axon structures and the myelin debris phagocytosed by the astrocytes were seen. In the transverse sections from Hq mice after the age of 3 months, hyperdense axoplasms filled with dark material were numerous and many axons showed axoplasms in various stages of dissolution.

    [0130] FIG. 15: Eye fundus imaging of Harlequin and control mice aged between 6 weeks and 4 months: Upper panel: cSLO fundus imaging of three controls mice aged 6 weeks, 2 and 4 months. Different regions of the retina are illustrated: Nasal (N), temporal (T), Superior (S) and inferior (I). In control mice shown, no change was noticed in nerve fiber density with aging. Bottom panel: cSLO fundus imaging of one Hq mice aged 6 weeks, one Hq mice 3 month-old and 2 Hq mice evaluated at 2 and 4 months of age. White discontinued lines show retinal regions with obvious nerve fiber loss; which was evidenced in both eyes of the Hq mice 3-month old. In the two other Hq mice examined, fiber loss was noticed at 4 months of age but in only one of their eyes: nasal region of the right eye for Hq mice #2 and nasal region of the left eye for Hq mice #4. In spite of the interindividual variability, eye fundus highlighted the disappearance of intraocular RGC axons in the majority of Hq mice aged between 4 to 6 months.

    [0131] FIG. 16: Screening of visual function of Harlequin and control mice at various ages: The Optomotry™ set-up allowed the determination of optokinetic tracking (OKT) thresholds (cycles per degree) for left and right eyes, independently (clockwise and counterclockwise responses respectively) under photopic conditions. OKT threshold is considered as an accurate measurement of rodent vision. For the control group, 15 mice aged 4-6 weeks and 25 mice aged about 8 months were evaluated. OKT responses did not change with age; besides scores for right and left eyes were assembled in a unique group since they were almost identical. Visual performance of Hq mice declines with age: it was observed a reduction of 14%, 25% and 56% in 2, 4 and 8 month-old animals relative to Hq mice aged 4-6 weeks. The scores illustrated in the graph, as for control mice, represented the means obtained for right and left eyes. Visual capabilities were also heterogeneous within a same group of age; nevertheless all the mutant animals eventually become severely impaired for responding to the visual stimuli.

    [0132] FIG. 17: Neuroglobin expression in Harlequin retinas: Left panel: The abundance and cellular distribution of NGB was examined by indirect immunofluorescence in retinal sections from one Hq mice which was subjected, at the age of 4 months, to intravitreal administration of AAV2/2-NGB in one eye (treated retina) while its counterpart remained untreated (untreated retina). Immunostaining for NGB was obtained with a specific antibody against NGB (green); nuclei were staining with DAPI (blue). A strong labeling for NGB in the GCL of AAV2/2-NGB treated retina was noticed. Scale bar is equivalent to 20 μm. Abbreviations: ONL, outer nuclear layer; IS, inner segments of photoreceptors INL, inner nuclear layer; GCL, ganglion cell layer. Right panel: RT-qPCR assays were performed using total RNAs from retinas isolated from 3 Hq mice subjected, at the age of 4 months, to intravitreal administration of AAV2/2-NGB in one eye and retinas isolated from 6 Hq mice which eyes were untreated. Since no difference in Ct values for the ATP6 mRNA was observed, it was used as “reference” for the ΔΔCT calculation. Next, steady-state level of NGB mRNA in transduced retinas was estimated after normalization against the mean of the signal obtained for NGB mRNA in retinas isolated from untreated eyes. BRN3A and SNCG mRNA abundances were also evaluated in the same samples; values were normalized against means obtained in untreated eyes. No significant difference was evidenced between treated and untreated retinas; primers used are shown in Table 9.

    [0133] FIG. 18: Effect of gene therapy on retinal ganglion cell integrity: Left panel: Immunofluorescence analysis of retinal sections from one Hq mice, from which one eye was subjected to AAV2/2-NGB administration (treated retina). Immunostaining for BRN3A (red) and nuclei contrasted with DAPI (blue) are shown. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer; OPL, outer plexiform layer; IPL, inner plexiform layer. Scale bar=25 mm. Right panel: RGC numbers were estimated in retinal sections from Hq treated and untreated eyes by counting BRN3A-positive cells in the GCL in 2-4 independent sections per eye for 13 Hq pairs of eyes in which only one was subjected to AAV2/2-NGB injection (7 samples) or AAV2/2-AIF1 injection (6 samples). Histograms illustrate data (mean values±S.E.M) corresponding to: the overall RGC number (BRN3A-positive cells in the GCL) and the total number of cells in the GCL (DAPI-stained nuclei in this layer). Values were compared to the ones obtained for untreated Hq animals aged 8-10 months and age-matched controls.

    [0134] FIG. 19: Effect of gene therapy on optic nerve functional integrity: Left Panel: Proximal optic nerve (ON) transversal sections from a 10 month-old control and the two ONs from one Hq mouse, in which one eye was subjected to intravitreal injection of AAV2/2-NGB were immunostained with an antibody against NF200. An unambiguous and similar diminution in the number of immunopositive dots (each dot revealed a single fiber) was noticed in both ONs from the Hq mouse when compared to the profile detected in the age-matched control. The nuclei were contrasted with DAPI and the scale bar is equivalent to 50 μm. Right Panel: Specific complex I (EC 1.6.5.3) and complex V (EC3.6.3.14) enzymatic activities were assessed in single ONs isolated from Hq and control mice aged about 8-10 months: 30 controls, 30 Hq mice which both eyes were untreated, 8 Hq mice in which one eye was subjected to AAV2/2-NGB intravitreal injection and the contralateral one remained untreated, 7 Hq mice in which one eye was subjected to AAV2/2-AIF1 intravitreal injection and the contralateral one remained untreated. Data collected for ONs isolated from the treated eyes were comparable between the two vectors used; thus they are represented in the bar graph as a single group (n=15). Complex I and V activities are expressed as nanomoles of oxidized NADH/min/mg protein. Histograms illustrate the enzymatic activities as mean±S.E.M of each assay per optic nerve measured in triplicate.

    [0135] FIG. 20: Visual function in Harlequin mice treated with NGB or AIF subsequent to RGC loss onset: Upper panel: cSLO fundus imaging of two Hq mice in which right eyes were subjected to AAV2/2-NGB or AAV2/2-AIF1 administration, images were collected before treatment (T=0) when animals were about 3-4 month-old. Different regions of each retina are illustrated: Nasal (N), temporal (T), Superior (S) and inferior (I). Animals were evaluated monthly after vector administration; images taken 6 months post-injection (just before euthanasia) are also shown. Below each image, the OKT threshold measured with the OptoMotry system was shown as visual acuity (VA). White discontinued lines show retinal regions with obvious nerve fiber loss in untreated eyes and treated eyes. In the four eyes evaluated the temporal region was the most affected; in eyes remained untreated fiber loss was more pronounced before euthanasia (T=6). On the contrary, in treated eyes fiber degeneration appeared almost unchanged, 6 months post-injection. Bottom panel: Twenty-eight mice treated in one of their eyes with either AAV2/2-NGB (14) or AAV2/2-AIF1 (14) were assessed with the OptoMotry system to measure OKT thresholds (cycles per degree) for each eye before vector administration, 3 and 6 months post-injection. Histogram shows the sensitivities for treated and untreated eyes before the treatment and six months later; values presented are the means±S.E.M of measures corresponding to individual eye responses. Each test was performed 3-6 times each time, 2-3 days apart.

    [0136] FIG. 21: Intraocular pressure changes in DBA/2J during glaucoma progression: DBA/2J and C57BL/6J males aged between 2 and 15 months of age were subjected to non-invasive Intraocular Pressure (IOP) using a Tonolab tonometer which makes five individual measurements and gives the mean as one reading displayed in mm Hg. The measurements were performed monthly on the two eyes and collected during daylight; histograms shows the means±SEM per group for evaluated as well as the age in months (m) and the number of eyes per group evaluated. Even though, measures can vary between right and left eyes in some animals data illustrated did not discriminate between both values.

    [0137] FIG. 22A-E: Evaluation of retinal ganglion cell loss, gliosis and microglial activation in DBA/2J mice: (A) RT-qPCR assays were performed using total RNAs from retinas isolated from either C57BL/6J or DBA/2J mice aged between 2 to 15 months to determine the steady-state levels of BRNA3A or SNCG mRNAs. The number of independent RNAs assessed per mouse group is indicated; histograms show the steady-state levels of BRN3A and SNCG mRNAs after normalization of their signals against the mean of the signal obtained for the BRN3A or SNCG mRNA in retinas from 2-month-old C57BL/6J mice. Primers used for are shown in Table 9. (B) The overall number of RGCs and of cells in the GCL estimated in C57BL/6J and DBA/2J mice of different ages by immunolabeling for BRN3A and DAPI staining of retinal sections. BRN3A and DAPI-positive cells in the GCL were counted in 3-4 independent sections per eye; the number of samples assessed per group is indicated. Histograms illustrate data (mean values±S.E.M) corresponding to: the overall RGC number (BRN3A-positive cells in the GCL, upper panel) and the total number of cells in the GCL (DAPI-stained nuclei in the GCL, bottom panel). (C) Immunofluorescence analysis of retinal sections from two C57BL/6J mice aged 2 and 15 months as well as two DBA/2J mice aged 8 and 15 months: Immunostaining for GFAP (bottom panel) and BRN3A (middle panel) of retinas are shown, the nuclei were contrasted with DAPI (upper panel). In control C57BL/6J mice, GFAP expression was restricted to the ganglion cell layer (GCL); while GFAP-labeled glial Muller cell processes which extended across the entire retina in DBA/2J mice. It also appears in the figure that in 15-month-old DBA/2J mouse the INL, IPL and ONL have thinned relative to the younger DBA/2J or to the age-matched control. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer; OPL, outer plexiform layer; IPL, inner plexiform layer. Scale bar=20 mm. (D) RT-qPCR assays were performed with RNAs purified from C57BL/6J and DBA/2J retinas; mice were euthanized at 2, 10 or 15 months of age and the number of independent RNA preparations subjected to the essay is indicated. GFAP mRNA amount variations are represented relative to the mean of the GFAP signal obtained in RNAs from 2 month-old C57BL/6J retinas (upper panel). Primers used for the GFAP gene are shown in Table 9. (E) Representative Western blots from retinas of C57BL/6J and DBA/2J retinas (mice were euthanized at 2 or 15 months of age). Each western blot was performed 3 times with antibodies against GFAP and β-actin. Histogram shows the relative amounts of the GFAP signals obtained from the independent immunoblots quantified with the Quantity One software and normalized against β-actin signals (the number of independent signals evaluated is indicated).

    [0138] FIG. 23A-C: Histopathological changes in DBA/2J optic nerves: Illustrations of immunolabeling for NF200, GFAP, IBA1 and Vimentin of optic nerves sections (near the globe); the nuclei were contrasted with DAPI (blue) and the scale bar is equivalent to 50 μm. (A) Proximal ON transversal sections were immunostained with an antibody against NF200 and GFAP an astrocyte marker. The diminution in the number of immunopositive dots in the ONs from DBA/2J mice noticeable at the age of 8 months and its aggravation with age reflects the progressive disappearance of RGC axons. Conversely the intensity of the immunolabeling for GFAP is higher in all the DBA/2J ONs evaluated relative to ONs from 8 month-old C57BL/6J mice and this independently of the age from which DBA/2J mice were euthanized. Thus, it appears that axon bundles were replaced by GFAP-positive material, probably astrocytes. (B) IBA1 protein is upregulated in all the ON sections from DBA/2J mice reflected by an intense imunolabeling which increased with age while in ON sections from 8-month old C57BL/6J or 2 month-old mice cells labeled appeared more ramified (resting state) and were less intense. (C) Immunochemistry for antibody against vimentin showed a very similar pattern of immunofluorescence relative to GFAP (Panel A), confirming that astrocytes in ONs from DBA/2J mice aged 8 and 12 months exhibited an increase in number and reactivity.

    [0139] FIG. 24A-B: Respiratory chain activity in DBA/2J retinas and optic nerves: Specific activities in single retinas or optic nerves from DBA/2J mice euthanized at various ages were assessed by spectrophotometry. Complex I and complex V activities were expressed as nanomoles of oxidized NADH/min/mg protein; antimycin-sensitive complex III activity was expressed as nanomoles of oxidized decylubiquinone/min/mg protein; Complex IV was expressed as nanomoles of oxidized cytochrome C/min/mg/protein. Values shown in each histogram represent the mean±S.E.M of triplicates per each sample evaluated. In the bottom of the histogram per column is indicated the age in months (m) of the group and number of independent measurements performed from single tissues. (A) Enzymatic activities of complex I, III and V were measured in single retinas isolated from DBA/2J mice aged 2, 8, 10 or 12 months. (B) Enzymatic activities of complex I, III, IV and V were measured in single optic nerves isolated from DBA/2J mice aged 2, 8, 10, 12 or 15 months.

    [0140] FIG. 25A-B: Changes in the abundance of mitochondrial proteins in retinas from DBA/2J mice: (A) Representative Western blots performed with 20 μg of protein extracts from whole retinas; DBA/2J mice were euthanized at the age of 2 or 12 month-old. Experiments were performed 2-3 times with 6-8 independent retinas; the following antibodies sequentially were used: anti-ATPase a (also known as ATP synthase a), anti-NDUFA9, anti-AIF, anti-SOD2. Histogram shows the relative amounts of the signals obtained from the independent immunoblots quantified with the Quantity One software and normalized against β-actin signals. The relative intensities, reflecting protein abundances in retinas, were represented as arbitrary unit±SEM. (B) The abundance and cellular distribution of NGB was examined by indirect immunofluorescence in retinal sections from 2 and 15 month-old DBA/2J mice using a specific antibody against NGB, nuclei were staining with DAPI. Scale bar is equivalent to 20 μm (Upper panel).

    [0141] Abbreviations: ONL, outer nuclear layer; IS, inner segments of photoreceptors INL, inner nuclear layer; GCL, ganglion cell layer.

    [0142] RT-qPCR assays were performed using total RNAs from retinas isolated from DBA/2J mice euthanized at various ages as indicated in months (m). Steady-state levels of NGB mRNA were determined after normalization against the mean of the signals obtained for the NGB mRNA in retinas isolated from 2-month-old DBA/2J mice; the number of independent results for each group was indicated in the histogram (Bottom panel, left).

    [0143] Western blot detection of NGB and β-Actin proteins was performed with 20 μg of whole protein extractions from DBA/2J retinas; mice were euthanized at 2 or 12 months of age. Specific antibodies against NGB and β-Actin recognized proteins of ˜17 and ˜42 kDa apparent molecular masses respectively as expected from their theoretical molecular weight estimations. It is noticeable in the 2 month-old sample two additional signals, with a weaker intensity, for NGB of about 19 and 21 kDa which could correspond to the forms evidenced in enriched mitochondrial fractions both in mouse and rat retinas as inventors previously demonstrated [Lechauve C et al., Biochim Biophys Acta. 2012; 1823: 2261-2273, Lechauve et al., Mol Ther. 2014; 22: 1096-1109 (Bottom panel, center). Histogram shows the relative amount of the NGB protein in retinas from 2 month-old DBA/2J mice (n=6) and 12 month-old DBA/2J mice (n=6). Each western blot was performed three times; signals obtained in the different immunoblots were quantified with the Quantity One software and normalized against β-actin signals (Bottom panel, right). The relative intensities, reflecting NGB protein abundance, were represented as arbitrary unit±SEM.

    [0144] FIG. 26A-C: Effects of AAV2/2-NGB intravitreal administration on DBA/2J retinas: (A) AAV2/2-NGB transduction efficiency was evaluated by immunostaining for NGB. A strong labeling for NGB in the GCL of AAV2/2-NGB treated retinas was noticed. Retinal sections from two DBA/2J mice were shown, AAV2/2 administration was performed in one eye at the age of two months; mice were euthanized eight months later. Nuclei were staining with DAPI and the scale bar is equivalent to 20 μm. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. (B) RGC numbers were estimated in DBA/2J treated and untreated eyes by counting BRN3A-positive cells in the GCL in 2-4 independent sections per eye for seven DBA/2J pairs of eyes in which only one was subjected to AAV2/2-NGB injection. Histograms illustrate data (mean values±S.E.M) corresponding to: the overall RGC number (BRN3A-positive cells in the GCL) and the total number of cells in the GCL (DAPI-stained nuclei in the GCL). Values were compared to the ones obtained for untreated animals aged 2 and 10 months. (C) Immunofluorescence analysis of retinal sections from one DBA/2J mice in which one eye was subjected to AAV2/2-NGB intravitreal injection at 2 month-old; the mice was euthanized at 10 month-old. Immunostaining for GFAP and BRN3A are shown, the nuclei were contrasted with DAPI. In the treated eye it was evidenced a higher number of BRN3A-positive cells while GFAP staining was weak and restricted to the GCL. Scale bars are equivalent to 20 μm. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer.

    [0145] FIG. 27: NGB overexpression effect on respiratory chain and visual function integrity: In the upper-left part of the figure, an illustrative curve is shown for the successive measurements of complex I (EC 1.6.5.3) and complex V enzymatic activities (EC3.6.3.14) in single optic nerves. Abbreviations: ATP, adenosine triphosphate; LDH, lactate dehydrogenase; MgCl2, magnesium chloride; NADH, reduced nicotinamide adenine dinucleotide; PEP, phosphoenol pyruvate; PK, pyruvate kinase.

    [0146] Complex I and V specific activities were measured in optic nerves from seven DBA/2J mice subjected at the age of 2 months to AAV2/2-NGB administration in one eye and euthanized 8 months later. The data for DBA/2J mice aged 10 months and which were not subjected to any treatment in both eyes was also shown. Specific activities are expressed as nanomoles of oxidized NADH/min/mg protein; enzymatic activities for each group were measured in triplicates and illustrated as means±SEM.

    [0147] No difference in complex V activity measurements was observed between ONs from the three groups tested (upper-right). Histograms representing specific Complex I activity and the ratios between Complex I and Complex V in the seven couples of ONs clearly confirm that the reduction of Complex I activity in ONs from untreated DBA/2J mice was efficiently prevented by AAV2/2-NGB administration (lower panel).

    [0148] FIG. 28A-B: Neuronal activity in the visual cortex and impact of NGB overexpression: (A) Photopic ERG responses from controls C57BL/6J and DBA/2J glaucomatous mice as a function of age and compared with DBA/2J mice after treatment with AAV2/2-NGB. In the upper panel ERG traces were illustrated from: two mice C57BL/6J aged 2 and 12 months and two DBA/2J mice aged 2 months and 10 months (treated animal in one eye) respectively. In the bottom panel the plot data correspond to mean±SEM for each group evaluated; a 32% reduction of the wave-b amplitude was noticed in the control mice relative to age while no difference was observed in the three DBA/2J groups assessed. (B) In the upper panel, F-VEP traces of the N1 and P1 waveforms were illustrated from: an 8 month-old C57BL/6J and two DBA/2J mice aged respectively 11 months and 10 months (treated animal in one eye) respectively. In the bottom panel, bar graphs of the peak amplitudes of N1 and P1 waves obtained from C57BL/6J mice (left) is shown; the histogram illustrating the N1 amplitudes (center) the P1 amplitudes (right) obtained from DBA/2J mice is also shown. Data represent means±SEM, statistical analysis was done by Mann-Whitney test (Two-tailed P values); the number of independent responses recorded per group is also indicated (n). There was a statistically significant difference in the N1 peak amplitude between the untreated DBA/2J mice and the AAV2/2-NGB treated mice. Conversely, NGB treatment did not change the intensity of the P1 wave in DBA/2J mice, but a significant decrease was observed in the C57BL/6J mice one year-old relative to their young counterparts.

    [0149] FIG. 29: Electrophysiological activities of the retinas and visual cortices of young C57BL/6J and DBA/2J mice: Left panel: Photopic ERG and F-VEP traces were illustrated from one C57BL/6J and one DBA/2J mouse, both aged 2 months. Right panel: The plot data for the photopic ERGs correspond to mean±SEM for each group evaluated. The number of independent responses recorded per group is also indicated; the wave-b amplitude did not change in the two groups evaluated. The bar graphs of the N1 peak amplitude is shown; data represent means±SEM, statistical analysis was done by Mann-Whitney test (Two-tailed P values); the number of independent responses recorded per group is also indicated. There was a statistically significant difference in the N1 peak amplitude between DBA/2J and C57BL/6J mice 2 month-old.

    [0150] FIG. 30: Histopathological changes in the optic nerves from 2 month-old DBA/2J mice relative to age-matched C57BL/6J mice: Illustrations of immunolabeling for GFAP and Vimentin of optic nerves sections (near the globe); the nuclei were contrasted with DAPI (blue) and the scale bar is equivalent to 50 μm. The intensity of the immunolabeling for GFAP is higher in the optic nerve from the 2 month-old DBA/2J mouse relative to the optic nerve from the 2 month-old C57BL/6J mouse. Immunochemistry for the antibody against vimentin revealed a very similar pattern of immunofluorescence relative to GFAP, confirming that astrocytes in ONs from DBA/2J mice aged 2 months exhibited an increase in number and reactivity relative to age-matched C57BL/6J mice. Thus, it seems that astrocytes in optic nerves from 2 month-old mice could respond by their proliferation/reactivity to the beginning of RGC axon damage.

    EXAMPLES

    Example 1

    1.1 Material and Methods

    1.1.1 Animals

    [0151] Male Long Evans rats were used (Janvier, France). They were housed two per cage in a temperature-controlled environment, 12 h light/dark cycle. All animal studies were conducted in accordance with the guidelines issued by the French Ministry of Agriculture and the Veterinarian Department of Paris (Permit number DF/DF_2010_PA1000298), the French Ministry of Research (Approval number 5575) and the ethics committees of the University Paris 6 and INSERM (Authorization number 75-1710).

    1.1.2 siRNA and shRNA Plasmid Construction

    [0152] Anti-Ngb siRNA (5′ GUGAGUCCCUGCUCUACAU[dt]3′ SEQ ID NO: 22) or unspecific scrambled siRNA (5′GCCACACGAUUGCUGUCUU[dt]3′ SEQ ID NO: 23) were synthesized by Sigma-Aldrich. Rat RGCs were transfected with siRNAs (50 nM) and HiPerfect reagent (Qiagen, Valencia, Calif.) as recommended by the manufacturer. Anti-Ngb shRNA and anti-scrambled shRNA expression vectors targeting the same regions than the siRNAs were constructed in a GFP-expressing shRNA vector (pRNA-U6.1, Genscript, USA).

    1.1.3 Purification of Retinal Ganglion Cells for Primary Cultures

    [0153] Primary cell cultures were derived from 8 weeks-old rat retinas and purified by modifications of the sequential immunopanning described for young rats (B. A. Barres, B. E. et al, Neuron, 1 (1988) 791-803.). RGCs were resuspended in half Neurobasal® medium (LifeTechnologies, Invitrogen) supplemented with B27 (1:50; LifeTechnologies, Invitrogen) and L-glutamine (2 mM; LifeTechnologies, Invitrogen) and half rat-retinal cell-conditioned culture medium (C. Fuchs, et al Invest Ophthalmol Vis Sci, 46 (2005) 2983-2991). RGC were seeded at 25,000 cells/cm.sup.2 into 48-well tissue-culture plates containing glass-coverslips previously coated 1 h with poly-D-lysine (2 μg/cm.sup.2) and then with laminin 1 μg/cm.sup.2 overnight (both from Sigma-Aldrich) (C. Fuchs, et al Invest Ophthalmol Vis Sci, 46 (2005) 2983-2991). siRNAs were added to the cells during the seeding in plates. Cells were incubated at 37° C. in a humidified atmosphere of 5% CO.sub.2. Cells were counted with viability test at one and seven days or fixed before performing immunocytochemistry analyses. RGC viability was assessed with the “live-dead” test (LifeTechnologies, Invitrogen).

    1.1.4 In Vivo Electroporation

    [0154] The electroporation (ELP) procedure was performed in only one eye per rat essentially as described by Ishikawa and colleagues (H. Ishikawa, et al Gene Ther, 12 (2005) 289-298). Under anesthesia with isoflurane (40 mg/kg body weight), 20 μg of shRNA expression vectors were injected into the vitreous body, next the inventors proceeded to the ELP (S. Ellouze, et al, Am J Hum Genet, 83 (2008) 373-387). All the animals were euthanized 12 weeks after ELP.

    1.1.5 Fundus Imaging by Confocal Scanning Laser Ophthalmoscopy (cSLO)

    [0155] A digital cSLO (Heidelberg Engineering, Germany) with green laser illumination was used to examine nerve fiber layer (NFL) in each cardinal area of rat eyes before treatment and every three weeks until euthanasia. Pupil dilation was performed with topical 1% tropicamide (CibaVision, France). Rats were manually held in front of the apparatus, in an upright position. The built-in software was used for post-processing the images, including alignment, adjustment of contrast, construction of a composite image (M. Paques, et al Vision Res, 46 (2006) 1336-1345).

    1.1.6 Optomotor Tests

    [0156] The head-tracking method is based on an optomotor test devised by Cowey and Franzini in 1979 and frequently used since then in both rats and mice (A. Bouaita, et al, Brain, 135 (2012) 35-52). Long Evans rats were placed individually on an elevated horizontal platform surrounded by a motorized drum.

    [0157] The protocol used yields independent measures of the acuities of right and left eyes based on the unequal sensitivities of the two eyes to pattern rotation: right and left eyes are most sensitive to counter-clockwise and clockwise rotations, respectively. A single blinded operator conducted all assessments, and codes were broken upon completion of data acquisition. The operator waited for the animal to settle in the chamber before initiating drum rotation. Vertical black-and-white lines of three varying widths, subtending 0.125, 0.25, and 0.5 cycles/degree (cyc/deg) were presented to the animal and rotated alternatively clockwise and counterclockwise, each for 60 s. This stimulated a subcortical reflex, so that a seeing animal involuntarily turned its head to track the moving lines. Head movements were recorded with a video camera mounted above the apparatus. Animals were scored only when the speed of the head turn corresponded to the speed of rotation of the stripes (12°/s). Light levels were kept constant (240 lux). Each animal was tested at four different time points by a single observer. During the experiments, a second masked grader re-evaluated the recorded videos to confirm the reliability of the scoring system.

    1.1.7 Retinal and Optic Nerve Histology

    [0158] Retinas or optic nerves were fixed in 4% PFA at 4° C., cryoprotected by overnight incubation in PBS containing 30% sucrose at 4° C. Retinas were embedded in OCT (Neg 50; Richard-Allan Scientific) and frozen in liquid nitrogen. Optic nerves were incubated in a 7.5% gelatin solution from porcine skin; Type A (Sigma-Aldrich) and 10% sucrose and frozen in 2-methyl-butane solution. Sections of retinas and optic nerves with a thickness of 10 μm were cut on a cryostat (Microm Microtech) and mounted on SuperFrost Plus slides.

    TABLE-US-00001 TABLE 1 Antibody description Antibody Type Concentration Supplier, reference Ngb (histology) Polyclonal   5 μg/ml Sigma, N-7162 Ngb (histology) Monoclonal   5 μg/ml Abcam, ab37258 Ngb (western Polyclonal   1 μg/ml Biovendor, RD181043050 blot) BRN3A Monoclonal   1 μg/ml Chemicon, MAB1585 ATPsynthase β Monoclonal 0.4 μg/ml Invitrogen, LifeTechnologies, A-  21351 ATPsynthase α Monoclonal 0.4 μg/ml Invitrogen, LifeTechnologies, 459240 TOMM 20 Monoclonal 0.4 μg/ml Abcam, ab56783 Alexa 488 Rabbit   4 μg/ml Invitrogen, LifeTechnologies, A11008 Alexa 594 Mouse   4 μg/ml Invitrogen, LifeTechnologies, A11005

    [0159] For immunochemistry, retinal sections were rinsed with PBS and treated with 1% BSA, 0.1% Triton and 0.05% Tween 20 in PBS for 1 h. They were then incubated with primary antibody overnight at 4° C. Sections were washed in PBS and incubated with the appropriate secondary antibodies and DAPI (2 μg/mL) for 2 h at room temperature. Primary and secondary antibodies used are shown in Table 1. Retinal flat mounts have been performed according to the protocol described by Paques and colleagues (M. Paques, et al Glia, 58 (2010) 1663-1668.).

    1.1.8 Microscopic Observations

    [0160] Fluorescence labeling was monitored with: (i) a confocal laser scanning microscope (Olympus FV1000) Microscope control and image acquisition was conducted by using Olympus Fluoview® software version 3.1. (ii) Retinal sections were also scanned with the Hamamatsu Nanozoomer Digital Pathology (NDP) 2.0 HT, its Fluorescence Unit option (L11600-05) and the NanoZoomer's 3-CCD TDI camera (Hamamatsu Photonics, France). BRN3A-positive cells, as the estimation of overall RGCs, were assessed for each animal by manually counting three entire retinal sections as described earlier (A. Bouaita, et al, Brain, (2011)). A resolution of 0.23 μm per pixel (40×) was used routinely. Finally, all the images were analyzed with Photoshop and Image J.

    1.1.9 RNA Extraction and qRT-PCR Assay

    [0161] Total RNA from rat retinas were extracted using RNeasy Plus Mini kit from Qiagen. One microgram of total RNA was reverse-transcribed with oligo-dT using Superscript® II Reverse Transcriptase (LifeTechnologies, Invitrogen) following the manufacturer's instructions. NGB, Cytoglobin (CYGB), BRN3A, gamma-synuclein (SNCG) and ATP6 primers were customized to be specific for each mRNA species (Table 2) and synthesized by Invitrogen.

    [0162] Quantitative PCR reactions were performed using ABI 7500 Fast (Applied Biosystems). The equivalent of 10 ng and 2 ng of cDNAs (relative to the whole RNA amount used for the reverse transcription) were used per gene as template for qPCR reactions with Power Sybr® green PCR Master Mix (Applied Biosystems) as recommended by the manufacturer. Each biological sample was subjected to the assay in triplicates per gene. Ct values were obtained by using ABI 7500 software (v.2.0.4) and the mitochondrial ATP6 gene was selected to normalize in order to obtain relative mRNA amount quantifications of each studied gene.

    TABLE-US-00002 TABLE 2 Primers for qRT-PCR assays Primer Primer Forward  reverse Gene 5′- 3′ 5′- 3′ Ngb CCAACGATGAA CAGAAATGCC GGAGAGAGG GAACCAAGAG (SEQ ID (SEQ ID  NO: 24) NO: 25) ATP6 CAACCAACCTT GCGGTAAGAA CTAGGGCTTC GTGGGCTAAA (SEQ ID (SEQ ID NO: 26) NO: 27) SNCG GTAACCTCGGT TTCCAAGTCCT GGCTGAGAA CCTTGCGTA (SEQ ID (SEQ ID NO: 28) NO: 29) BRN3A AGGCCTATTTT CGTCTCACACC GCCGTACAA CTCCTCAGT (SEQ ID (SEQ ID NO: 30) NO: 31) Cygb GACTGACTTGC GTCTGAAGTG TCCGGAAAG AGCGGGTGAG (SEQ ID (SEQ ID NO: 32) NO: 33)

    1.1.10 Mitochondrial Purification and Western Blotting Analysis

    [0163] 24 retinas were isolated from 8 week-old rats and washed in PBS at 4° C.; they were then homogenized in extraction buffer (0.32 M sucrose, 30 mM Tris-HCl; pH 7.6, 5 mM MgAc, 100 mM KCl, 0.1% fatty acid-free BSA, 5 mM β-mercaptoethanol, and 1 mM PMFS) and mitochondria were purified as previously described (V. Kaltimbache et alr, Rna, 12 (2006) 1408-1417) and as illustrated in FIG. 1. 30 μg of mitochondrial proteins were treated with 150 or 200 μg/mL proteinase K (PK) in the presence or absence of 1% Triton X-100 (v/v) at 4° C. for 30 min. The reaction was stopped by addition of 1 mM PMFS (Sigma-Aldrich). Samples were collected by centrifugation at 10000 g for 15 min at 4° C. For mitoplast preparation, aliquots of 30 μg of mitochondrial pellets (in 100 μL of extraction buffer) were subjected to osmotic shock by addition of 900 μL of 3 mM HEPES (pH 7.4) containing 1 mM of PMFS. After incubation on ice for 15 min, the suspension was centrifuged at 10000 g for 15 min to yield the mitoplast pellet. All the samples were then resolved in 12% or 15% SDS-PAGE and next transferred to a PVDF membrane. ?Membranes were probed with antibodies against Ngb, ATP synthase-α and TOMM20 (cf. Table 1). Immunoreactive bands were visualized with anti-mouse or anti-chicken coupled to horseradish peroxidase (0.1 mg/mL) followed by ECL Plus detection (Amersham International). Five independent mitochondria purifications were subjected to the analyses.

    1.1.11 Tissue Homogenate Preparation and Respiratory Chain Enzymatic Assays

    [0164] Optic nerves were prepared at 4° C. by homogenization of tissues using a 1 mL hand-driven glass-glass potter in 100 μL of extraction buffer (0.25 mM sucrose, 40 mM KCl, 2 mM EGTA, 1 mg/ml BSA, and 20 mM Tris-HCl, pH 7.2). Large cellular debris were spun down by a low speed centrifugation (1000 g×8 min) and supernatants were used immediately. Respiratory chain complex activities were measured using a Cary 50 spectrophotometer maintained at 37° C. (Varian, Australia) as previously described (P. Benit, et al, Clin Chim Acta, 374 (2006) 81-86.). Each assay was made in duplicate with 20 μL of the homogenates obtained. Complex activity values were converted to specific activities after protein quantification by the Bradford method. All chemicals were of the highest grade from Sigma Chemical Company.

    1.1.12 Statistical Analyses

    [0165] Values are expressed as means±SEM (standard error of the mean).

    [0166] Statistical analyses were performed with the GraphPad Prism5.0 software assuming a confidence interval of 95%. Data collected for all the independent observations were compared using the non parametric significance test of Mann-Whitney U (*≤0.05, **≤0.01 and ***≤0.005).

    1.2 Results

    1.2.1 Neuroglobin Profiles in Adult Rat Retinas

    [0167] RGCs integrity is essential for visual function and their loss is directly involved in optic neuropathies and glaucoma. In an attempt to relate Ngb and mitochondria, the subcellular distribution of the mitochondrial protein ATP synthase-β (a subunit of the respiratory chain complex V) and the Ngb protein by immunostaining of retinal sections have been compared. Consistent with other reports, cells positive for antibodies against Ngb were found in the photoreceptor layer (PL), the inner nuclear layer (INL) and the GCL (Data not shown). It clearly appeared that Ngb is highly expressed in RGCs, specifically labeled with an antibody against the transcription factor BRN3A. To further support the presence of Ngb in RGCs and its possible localization to mitochondria, retinal flat mounts and optic nerve sections from adult rats were examined and processed for immunohistochemistry using anti-Ngb associated to anti ATP synthase-β (Data not shown). Immunolabeled RGCs showed a cytoplasmic, a dendritic, and an axonal distribution of Ngb protein. The immunoreactivity of the Ngb antibody was often revealed as a punctuate distribution of fluorescent dots excluded from the nuclei, similar to the signal obtained with anti-ATP synthase-β antibody; both antibodies revealed an elevated extent of colocalization (Data not shown). Moreover, results clearly show a homogenous signal of Ngb in RGC somas and their dendrites in the IPL as well as in their axons in the Nerve Fiber Layer (NFL). Longitudinal optic nerve sections from adult rats were also performed and processed for immunohistochemistry using anti-Ngb associated with anti-ATP synthase-β. Ngb signals observed in optic nerves not only overlapped with ATP synthase-β signals but also confirmed the presence of mitochondrial labeling of RGC axons which can be distinguished from that of the resident optic nerve cells (Data not shown). Consequently, NGB subcellular localization matches to some extent with the pattern of mitochondrial network revealed by ATP synthase-β in both RGC bodies and their axons.

    1.2.2 Neuroglobin Subcellular Localization in Retinal Neurons

    [0168] To ascertain whether NGB can be detected inside the mitochondria, organelle fractions purified by differential centrifugation (FIG. 1) from rat retinas were subjected to immunoblot analyses. Results show that specific antibodies against Ngb recognized three proteins of ˜17, ˜19 and ˜21 kDa apparent molecular mass, in enriched mitochondria fractions; only the 17 kDa form was found in the cytosol. To compare the proportion of Ngb in the cytosol or the mitochondria relative to the total amount of Ngb in homogenates, four independent experiments of subcellular fractionation (as described in FIG. 1) were compared. The results indicate that: (i) Signals obtained with antibody against Ngb or ATP synthase α were enriched in mitochondrial fractions relative to whole homogenates or cytosols; (ii) when the sum of the NGB signals in mitochondrial and cytosolic factions were calculated it appears that approximately 70% of the overall Ngb signal was revealed in the mitochondrial compartment.

    [0169] To further evaluate if Ngb was completely translocated inside the organelle, mitochondria fractions were treated with PK; significant amounts of the three forms of Ngb and ATP synthase-α are insensitive to PK-induced proteolysis, thus indicating that Ngb forms were truly integrated into the mitochondria and hence remain detectable on immunoblotting (Data not shown). Next, mitochondrial fractions were treated with PK and Triton X-100; theoretically the detergent disrupts both mitochondrial membranes and leads to the entire proteolysis of mitochondrial proteins demonstrating their localization inside the organelle in a protease-sensitive form. Results clearly show that ATP synthase-α and NGB protein signals became diminished as they were digested by PK in the presence of detergent.

    [0170] To determine whether Ngb could be present either in the intermembrane space or the internal side of the inner membrane; mitoplasts were obtained by subjecting the samples to osmotic shock. The outer membrane protein TOMM20 was significantly digested by PK, as expected since a portion of it is exposed to the cytosolic side, moreover the signal entirely disappeared in mitoplasts. By contrast, the three forms of Ngb were mostly preserved as is ATP synthase-α (Data not shown). Hence, the three forms of Ngb identified by Western blotting of mitochondrial fractions were fully integrated inside the organelle and they could be further contained in the matrix or in the inner membrane.

    1.2.3 Neuroglobin Knockdown in Rat Purified Primary Culture of Retinal Ganglion Cells by Small Interfering RNA

    [0171] The inventors examined whether the inhibition of NGB expression influenced RGC survival and neurite outgrowth in primary RGC cultures via the small interfering RNA (siRNA) strategy.

    [0172] To study the impact of NGB knockdown on cell survival and neurite outgrowth, the anti-Ngb siRNA or the scrambled siRNA were associated with transfection agent and added on the 1.sup.st day of RGC cultures in rat retinal cell-conditioned medium. Under this culture condition RGCs survive 7-12 days and developed neuritic processes that may extend for several cell-body diameters and were often branched (Data not shown). After seven days of culture, many viable cells develop neuritic processes and immunoreactivity obtained with the antibody against Ngb was similar to that observed with the antibody against the mitochondrial ATP synthase-β polypeptide, as previously observed in vivo (see point 1.2.1). Indeed, both proteins were distributed along the neuritic processes and in the cytosol of RGC primary culture in control conditions (Data not shown). Furthermore, it clearly appears that Ngb immunostaining was strongly reduced (Data not shown) compared to cells transfected with scrambled siRNA or untreated cells.

    [0173] Cell survival was assessed by counting live RGCs from dead-cell population. In controls, an approximate 50% diminution in the total amount of living cells was observed after 7 days of culture, an expected phenomenon related to the vulnerability of these adult neurons once their axons are disrupted. In fact, previous reports described that after 10 days of culture, the cell density was reduced to 10% in control conditions and to 30% when conditioned medium obtained from retinal cells was used. Despite this effect, the inventors observed a significant and more pronounced decrease (˜3-fold) in cell survival when RGCs were transfected with anti-Ngb siRNA (Table 3) relative to either control conditions or scrambled siRNA treatment. The differences were significant according to the Mann-Whitney non parametric test (p: 0.0007 or 0.02 respectively). This result indicates that NGB expression is essential for RGC survival in vitro.

    TABLE-US-00003 TABLE 3 Cell survival was estimated using the “live-dead” test. Only alive RGCs were counted from ten fields selected identically on each coverslip of three independent RGC cultures at one or seven days. Results were presented as cell survival relative to control conditions (mean values ± S.E.M, standard error for the mean). Day 1 (number of Day 7 (number of Total number of RGCs independent counts) independent counts) Control 85 ± 35 (7)  46.6 ± 18.14 (7) Anti-Ngb siRNA treatment 92 ± 20 (6) 16.0 ± 4.0 (10) P value Ngb/Control 1 0.007 P value Ngb/scrambled 1 0.02 Scrambled siRNA treatment 96 ± 38 (4) 29.3 ± 9.7 (7)  P value Scrambled/Control 0.48 0.08

    1.2.4 In Vivo Knockdown of Neuroglobin Expression in Retinal Ganglion Cells of Adult Rats and its Impact on Nerve Fiber Density

    [0174] The impact of NGB expression attenuation was assessed in vivo by ELP (S. Ellouze et al, Am J Hum Genet, 83 (2008) 373-387), after injection in the vitreous body of one eye of a plasmid DNA leading to the synthesis of a short hairpin RNA (shRNA) anti-Ngb. The in vivo ELP procedure results in a highly efficient gene delivery to the GCL since more than 50% of RGCs express the transgene for at least 2 months. NFL integrity was evaluated using confocal Scanning Laser Ophthalmoscopy (cSLO), which represents a powerful technique for in vivo imaging of rodents eye fundus (A. Bouaita, et al, Brain, 135 (2012) 35-52.). Striations of NFL radiating from the optic disc were clearly visible in each eye from rats subjected to ELP. Each area of the eye fundus was visualized before and at different times after treatment until euthanasia. The inventors analyzed nerve fiber density before and 3 months after scrambled shRNA plasmid administration: eye fundus visualizations did not show any darker or thinner striations, when compared to the ones obtained before ELP or to untreated eyes. In contrast, a noticeable loss of nerve fiber bundles was evidenced in eyes electroporated with anti-NGB shRNA (Data not shown). The NFL striation loss, especially in the superior and inferior retinal areas, reflects RGC axon degeneration and was noticed one month after the ELP. Nerve fiber disappearance was observed in ˜25% of treated anti-NGB shRNA eyes (n=24); while eye fundus from all the eyes electroporated with the scrambled shRNA plasmid (n=22) or untreated eyes (n=46) showed well preserved tracks of axons in all the areas visualized until euthanasia (Data not shown). These results indicated that RGC axons in the NFL of eyes treated with anti-Ngb shRNA have undergone a degenerative process, confirming the deleterious effect of in vivo NGB knockdown.

    1.2.5 Relative Abundance of Neuroglobin and Specific Retinal Ganglion Cells mRNAs

    [0175] To substantiate that the anti-NGB shRNA treatment was efficient in reducing NGB mRNA levels, its abundance in retinas was determined by quantitative real-time PCR of reverse-transcribed mRNAs (RT-qPCR) using the comparative ΔΔCt method and the mitochondrial ATP6 gene as “normalizing” gene. RNA preparations from retinas of 18 control eyes, 8 eyes treated with scrambled shRNA or 10 eyes treated with anti-Ngb shRNA, from rats euthanized 3 months after ELP, were examined. The relative amount of NGB mRNA in anti-NGB shRNA treated retinas was 13.8% less abundant than in retinas isolated from the 18 untreated eyes; this difference was significant according to the Mann-Whitney test (p=0.023) (FIG. 2). The decrease of NGB mRNA may appear small, but RNAs were prepared from the whole retinal cell population. Thus, NGB mRNA levels of RGCs transduced with the anti-NGB shRNA cannot be discriminated from those of untransduced cells which also expressed NGB (bipolar and photoreceptor; cf. point 1.2.1), whose proportion is very high relative to RGCs; indeed, the fraction of RGCs relative to the total cell population was estimated at less than 1% in adult mouse retina (C. J. Jeon, et al, J Neurosci, 18 (1998) 8936-8946). The administration of the plasmid directing the synthesis of scrambled shRNA did not change the relative abundance of NGB mRNA in treated retinas compared to retinas from untreated eyes (p=0.51) (FIG. 2). Additionally, the relative amount of CYGB mRNA encoding another hexacoordinated globin was evaluated (C. Lechauve, et al, Febs J, 277 (2010) 2696-2704); it did not change in any of the retinas examined (p=0.87 and 0.22 for scrambled and anti-Ngb shRNA respectively) confirming that anti-NGB shRNA plasmid administration leads specifically to the diminution of NGB mRNA steady-state levels (FIG. 2).

    [0176] The amount of SNCG and BRN3A mRNAs were also measured since they are highly abundant in adult RGCs, while they are almost undetectable in other retinal neurons. Anti-Ngb shRNA-treated eyes showed a ˜20% reduction of both SNCG and BRN3A mRNA levels relative to values measured in retinas from control eyes (p=0.002 and 0.033 for SNCG and BRN3A respectively) (FIG. 3); Whereas, steady-state levels of SNCG and BRN3A mRNAs did not change in eyes electroporated with the scrambled shRNA relative to controls: p=0.17 and 0.95 for SNCG and BRN3A respectively (FIG. 3). Thus, NGB knockdown was effective on reducing NGB mRNA level which results in a ˜20% reduction of SNCG and BRN3A mRNA amounts. This diminution could reflect RGC loss, since SNCG or BRN3A expression is considered as an index of RGC number (R. Torero lbad, et al J Neurosci, 31 (2011) 5495-5503).

    1.2.6 Deleterious Effect of Neuroglobin Knockdown on Retinal Ganglion Cells Integrity In Vivo

    [0177] To corroborate the negative impact of NGB knockdown on RGC integrity, retinal sections from rats euthanized 3 months after the treatment were examined by immunochemistry using antibodies against NGB and BRN3A proteins. Retinal sections of anti-Ngb shRNA treated animals presenting a noticeably loss of nerve fibers (Data not shown) showed an important diminution of BRN3A-positive cells and the NGB immunostaining signal in the GCL relative to the signals observed in the accompanying untreated eye (Data not shown, rat #1); an additional rat retinal section is shown in which a more subtle reduction of BRN3A-positive cells was evidenced (Data not shown rat #5). Ngb staining in the other retinal layers was similar in treated and control eyes (Data not shown rat #1 and #5). On the other hand, no evident changes in BRN3A-positive cells were noticed in eyes electroporated with scrambled shRNA (Data not shown, rat #4). Cryostat sections of retinas were counted for BRN3A-positive cells in the GCL to estimate the number of RGCs in 8 eyes electroporated with anti-NGB shRNA, 7 eyes electroporated with scrambled shRNA and 10 untreated eyes.

    TABLE-US-00004 TABLE 4 RGC densities were calculated after immunolabeling for BRN3A antibody and DAPI staining; this later allowing estimation of total nuclei in the GCL. BRN3A and DAPI-positive cells in the GCL were counted in three independent retinal sections per animal: 10 control eyes, 7 eyes treated with scrambled shRNA and 8 eyes treated with anti-NGB shRNA. Results were presented as cell density/mm relative to control conditions (mean values ± S.E.M, standard error for the mean, student t test). Cell density per mm Total cells of GCL RGCs Control eyes (n = 10) 121.1 ± 4.2 48.4 ± 3.0 Anti-NGB shRNA treated 116.7 ± 8.7 40.1 ± 4.3 eyes (n = 8) P value NGB/control 0.60 0.001 P value Ngb/scrambled 0.60 0.005 Scrambled shRNA treated 120.9 ± 4.6 47.1 ± 3.0 eyes (n = 6) P value scrambled/control 1 0.25

    [0178] The inventors found that RGC density (overall number of RGCs/mm) was reduced in anti-NGB shRNA-treated retinas relative to control retinas (Table 4), the diminution of about 20% was significant according to the Mann-Whitney t test: p=0.001. Conversely, RGC density in eyes treated with scrambled shRNA was not significantly different to controls: p=0.25 (Table 4). Hence, in vivo ELP with anti-NGB shRNA leads to a significant RGC loss which supports the optic fiber disappearance evidenced by eye fundus imaging (Data not shown).

    1.2.7 Impact of Neuroglobin Knockdown on Respiratory Chain Activity in Optic Nerves

    [0179] Since the number of BRN3A-positive cells and eye fundus imaging indicated that the knockdown of NGB expression leads to RGC degeneration, the inventors evaluated whether respiratory chain function can be hampered in optic nerves isolated from animals sacrificed 3 months after anti-Ngb shRNA treatment. The spectrophotometric method used for assessing enzymatic activities of respiratory chain complexes has been successfully applied to accurately detect isolated defects in small amounts of tissue homogenates (P. Benit, et al, Clin Chim Acta, 374 (2006) 81-86). Two independent spectrophotometric assays were devised to sequentially measure in homogenates of single optic nerves the enzymatic activities of: (1) rotenone-sensitive complex I (CI) and oligomycin-sensitive complex V (CV); (2) complex IV (CIV), malonate-sensitive combined complex II+III (CII+CIII) and antimycin-sensitive complex III (CIII) (FIG. 4). A 30% decrease relative to control values was observed for CI and CIII specific activities in optic nerves of eyes electroporated with anti-Ngb shRNA (n=17) relative to control optic nerves (n=32); the differences were significant according to the Mann-Whitney test; p values were 0.001 and 0.051 for CI and CIII activities respectively (Table 5). Conversely, CI and CIII activities in optic nerves from eyes electroporated with scrambled shRNA (n=15) were not significantly different from those measured in untreated eyes (p values: 0.64 and 0.63 respectively). Besides, CII, CIV and CV activities were similar between anti-shRNA treated eyes and controls (Table 5). Thus, NGB knockdown in retinas interferes in a specific manner on respiratory chain function in optic nerves, since the defect was specifically demonstrated in CI and CIII enzymatic activities.

    TABLE-US-00005 TABLE 5 Specific activities assessed in optic nerves from 32 control eyes, 17 eyes treated with anti-NGB shRNA or 15 treated with scrambled shRNA are shown. Origin of optic nerves Specific activities (nmol/min/mg protein) ± S.E.M and number tested (n) CI CII + CIII CIII CIV CV Control eyes (n = 32) 68.3 ± 14.3 19.4 ± 6.1 82.7 ± 26.4 102.2 ± 32.5 115.3 ± 45.3 Scrambled shRNA 65.8 ± 11.0 19.2 ± 3.6 88.3 ± 25.6 102.8 ± 21.5 120.3 ± 38.5 treated eyes (n = 15) Mann-Whitney test 0.64 0.77 0.63 0.86 0.74 Anti-Ngb shRNA treated 48.8 ± 15.8 16.9 ± 7.3 64.7 ± 22.7  88.0 ± 38.3  99.3 ± 46.4 eyes (n = 17) Mann-Whitney test 0.001 0.34 0.051 0.26 0.40 (/control) Mann-Whitney test 0.016 0.46 0.043 0.30 0.40 (/scrambled) Abbreviations: CI-CV, various complexes of the respiratory chain; Aa, antimycin A; ATP, adenosine triphosphate; Cyt, cytochrome; DCPIP, dichlorophenol indophenol; DQ, DQH2, duroquinone (oxidized), duroquinol (reduced), respectively; EDTA, ethylene diamine tetraacetic acid; KCN, potassium cyanide; LDH, lactate dehydrogenase; MgCl.sub.2, magnesium chloride; NADH, reduced nicotinamide adenine dinucleotide; oligo, oligomycin; ox, oxidized; PEP, phosphoenol pyruvate; PK, pyruvate kinase; red, reduced; rot, rotenone; succ, succinate.
    1.2.8 Functional Assessment of Visual Responses in Animals Treated with Anti-Ngb shRNA

    [0180] The functional consequences of RGC loss due to anti-Ngb shRNA treatment on visual function was assessed by optomotor head-tracking experiments (R. M. Douglas, et al, Vis Neurosci, 22 (2005) 677-684). Tracking capability was examined in both clockwise and counter clockwise drum rotations at three frequencies: 0.5, 0.25 and 0.125 cycles per degree. Because only temporal-to-nasal motion is effective through each eye, clockwise movement will drive tracking through the left eye, whereas counterclockwise motion will activate the right eye. FIG. 5 shows visual responses assessed in 8 and 18 week-old control rats (n=22 and n=10); they mostly presented head tracking scores of similar magnitude for the clockwise or counterclockwise drum rotations. FIG. 6 shows data collected from rats 10 weeks after the treatment with scrambled shRNA or anti-NGB shRNA plasmid. Rats treated with scrambled shRNA (n=20) showed no significant difference between clockwise and counterclockwise responses for all the three spatial frequencies tested (0.125 p=0.30 at 0.125 cycles per degree, p=0.94 at 0.25, and p=0.68 at 0.5. Moreover, their clockwise responses before plasmid administration and 10 weeks later were very similar for instance for the 0.5 cycle per degree frequency the p value calculated was 0.41 (FIG. 5, untreated 8 week-old and FIG. 6 scrambled shRNA). Conversely, an unequivocal decline in visual performance in rats treated with anti-NGB shRNA was measured (n=22) which responded with poor clockwise scores; indeed they spent much less time tracking across the test period for the clockwise drum rotation relative to the counterclockwise responses. The decline in visual performance was statistically different (p<0.0001 at all three frequencies); which represents a 46% decrease of that measured for their accompanying control eyes (FIGS. 5 and 6). Obviously when the clockwise responses of these animals were compared to the one collected from untreated animals of 18 week-old (FIG. 5, untreated 18 week-old) the diminution of about 43% is also significant (p<0.005 at all three frequencies). Thus, RGC loss due to NGB knockdown has a negative impact on head-tracking behavior; indicating the direct connection between NGB expression, respiratory chain integrity and visual function.

    Example 2

    2.1 Material and Methods

    2.1.1 Animals and Diets

    [0181] The Hq strain was B6CBACaAw-J/A-Pdc8Hq/J obtained from Jackson Laboratory (http://jaxmice.jax.org/strain/000501.html). These mice exhibit the main features of human neurodegenerative diseases due to respiratory chain complex I (RCCI) deficiency, such as the degeneration of the cerebellum, retina, optic nerve, thalamic, striatal, and cortical regions. This complex phenotype is caused by the knockdown of the nuclear gene AIF encoding the mitochondrial Apoptosis Inducing Factor, which levels drops to less than 10% of the amount seen in wild-type mice (Klein J A, et al. (2002) Nature 419: 367-374), and leads to RCCI deficiency (Vahsen N, et al. (2004) Embo J 23: 4679-4689). All Hemizygous (Hq/Y) males used in this study were F1 mice bred from founders having a mixed genetic background. Hemizygous (Hq/Y) males were the recipient of evaluations and gene therapy; they were compared exclusively to the littermate males from the colony. The mice were housed from one to four per cage in a temperature-controlled environment, 12-h light/dark cycle and free access to food and water. Studies were conducted in accordance with the statements on the care and use of animals in research of the guidelines issued by the French Ministry of Agriculture and the Veterinarian Department of Paris (Permit number DF/DF_2010_PA1000298), the French Ministry of Research (Approval number 5575) and the ethics committees of the University Paris 6 and the INSERM, Institut National de la Sante et de la Recherche Médicale (Authorization number 75-1710).

    2.1.2 Adeno-Associated Viral Vector and Intravitreal Injections

    [0182] The Mus musculus Neuroglobin (NGB) mRNA sequence of 1630 base pairs (bp) (NM_022414.2, SEQ ID NO: 5) was synthesized by Genscript Corp (Piscataway, N.J. 08854 USA). It encompasses the full-length 5′UTR (279 bp, SEQ ID NO: 6), the entire Open Reading Frame (ORF; SEQ ID NO: 8) encoding 151 amino acid-long protein, and two restriction sites for cloning into the pAAVIRES-hrGFP vector. The hGH (human growth hormone 1) polyadenylation signal was replaced by the full-length 3′UTR of NGB (895 bp, SEQ ID NO: 7). NGB transcription is under the control of the Cytomegalovirus promoter and the β-globin intron for ensuring high levels of expression. The ORF is in frame with the 3× FLAG® sequence at the C-terminus. The pAAV-IRES-hrGFP vector (http://www.genomics.agilent.com/) has a dicistronic expression cassette in which the humanized recombinant green fluorescent protein (hrGFP) is expressed as a second ORF translated from the encephalomyocarditis virus internal ribosome entry site (IRES). The final vector, named AAV2/2-NGB (SEQ ID NO: 9), contains AAV2 inverted terminal repeats (ITRs), which direct viral replication and packaging. The expression cassettes flanked by the two AAV2 ITRs, were encapsidated into AAV2 shells. Vectors were produced by the “Centre de Production de Vecteurs and the INSERM UMR1089, Nantes” (http://www.atlantic-genetherapies.fr/). The rAAV titers were determined by dot blot and expressed as vector genomes (VG) per mL; 1×10.sup.12 VG/mL. For intravitreal injections, after dilatation of the pupil with topical 1% tropicamide (CibaVision, France), mice were subjected to anesthesia with isoflurane (40 mg/kg body weight). The tip of a 33-gauge needle, mounted on a 10 μl Hamilton syringe (Hamilton Bonaduz AG, Bonaduz, Switzerland) was advanced through the sclera and 2 μL of vector suspension (2×10.sup.9 VG) was injected intravitreally, avoiding retinal structure disruption, bleeding or lens injury. Viral particles were extemporaneously mixed with 1/10000 of fluorescein to follow the homogenous dissemination of the suspension into the vitreous body. Forty five mice were subjected to AAV2/2-NGB administration during the course of this study. One animal suffered from a haemorrhage after intravitreal injection; it was euthanized and discarded from the study. Further, two mice died few weeks after the treatment due to technical errors in the animal facility and one died from natural causes.

    2.1.3 Fundus Imaging by Confocal Scanning Laser Ophthalmoscopy

    [0183] A digital confocal Scanning Laser Ophthalmoscope, cSLO (Heidelberg Engineering, Germany) was used to examine nerve fiber layer (NFL) in each cardinal area of mouse eyes before treatment and different times after vector administration as previously described (Paques M, et al. (2006) Vision Res 46: 1336-1345). Briefly, all examinations were carried out in manually restrained conscious animals which were held in front of the cSLO objective after pupil dilation; the overall duration of each examination was 1 minute per eye. Stacks of 30 images (1,500 μm of approximate width and a definition of 512×512 pixels) were acquired at different planes of focus to capture the whole surface of the retina.

    2.1.4 Optomotor Response

    [0184] Visual acuity was measured, under photopic conditions, by observing the optomotor responses of mice to rotating sinusoidal gratings (OptoMotry™) (Prusky G T, et al (2004) Invest Ophthalmol Vis Sci 45: 4611-4616). Mice reflexively respond to rotating vertical gratings by moving their head in the direction of grating rotation. The protocol yields independent measures of right and left eye acuities based on the unequal sensitivities of the two eyes to pattern rotation: right (untreated) and left (treated) eyes are most sensitive to counterclockwise and clockwise rotations, respectively (Prusky G T, et al (2004) Invest Ophthalmol Vis Sci 45: 4611-4616). Each mouse was placed on a pedestal located in the centre of four inward facing LCD computer monitors screens. Once the mouse became accustomed to the pedestal, the test was initiated by presenting the mouse with a sinusoidal striped pattern that rotates either clockwise or counter-clockwise and varying widths. Spatial frequency of the grating was randomly increased by the software until the animal no longer responded. As the mouse moved about the platform, the experimenter followed the mouse's head with a crosshair superimposed on the video image. When a grating perceptible to the mouse was projected on the cylinder wall and the cylinder was rotated (12°/s), the mouse would typically start to track the grating with reflexive head movements in concert with the rotation. The short testing reduced the possibility of adapting to the stimulus and established that each animal was capable of tracking when a changed stimulus was present. The process of changing the spatial frequency of the test grating was repeated a few times until the highest spatial frequency the mouse could track was identified as the threshold which defined the visual acuity. Experimenters were masked to the treatment and to the animal's previously recorded thresholds.

    2.1.5 Retinal and Optic Nerve Histology

    [0185] Retinas and optic nerves (ONs) were carefully collected and fixed in 4% PFA at 4° C., cryoprotected by overnight incubation in PBS containing 30% sucrose at 4° C. Retinas were embedded in Optimal Cutting Temperature compound, OCT (Neg 50; Richard-Allan Scientific), frozen in liquid nitrogen and optic nerves were embedded in a solution of PBS+7.5% gelatin from porcine skin Type A (Sigma-Aldrich) and 10% sucrose and frozen in a 2-methyl-butane solution at −45° C. Sections of retinas and ONs were cut (10 μm thickness) on a cryostat (Microm HM560, Thermo Scientific) at −20° C. and mounted on SuperFrost®Plus slides.

    [0186] For immunochemistry, sections of retinas and ONs were permeabilized with 0.1% Triton X-100 in PBS for 15 minutes at room temperature and treated with 3% BSA, 0.1% Triton and 0.05% Tween 20 in PBS for 1 hour. They were then incubated with primary antibody overnight at 4° C. The next day, sections were washed three times in PBS and incubated with the appropriate secondary antibodies and 2 μg/mL of 4′, 6-diamidino-2-phenylindole (DAPI) for 2 hours at room temperature with 3% BSA, 0.1% Triton and 0.05% Tween 20 in PBS. At last, they were washed 3 times with PBS, rinsed with sterile water and mounted on a glass slide. Primary and secondary antibodies used are shown in table 6.

    TABLE-US-00006 TABLE 6 Antibody description Antibody Type Concentration Supplier, reference NGB (histology) Polyclonal   5 μg/mL Sigma, N-7162 NGB (western blot) Polyclonal   1 μg/mL Biovendor, RD181043050 AIF (western blot) Monoclonal 0.4 μg/mL Millipore, AB16501 ATP synthase β Monoclonal 0.2 μg/mL Invitrogen, LifeTechnologies, A-21351 (western blot) GFP (histology) Polyclonal  20 μg/mL Abcam, ab56783 NDUFB6 (histology) monoclonal  10 μg/mL Invitrogen, LifeTechnologies, A21359 BRN3A (histology) Monoclonal   1 μg/mL Chemicon, MAB1585 GFAP (histology) Polyclonal 2.9 μg/mL Sigma Aldrich, G3893 NF200 (histology) Monoclonal   1 μg/mL Chemicon, MAB1585 FLAG (histology) Monoclonal   5 μg/mL Sigma Aldrich, F1804 Alexa 488 Rabbit   4 μg/mL Invitrogen, LifeTechnologies, A11008 Alexa 594 Mouse   4 μg/mL Invitrogen, LifeTechnologies, A11005

    [0187] Immunofluorescence analyses with antibodies against BRN3A and GFP were performed on retinal sections from Hq mice in which one eye was subjected to intravitreal administration of the AAV2/2-NGB vector while the contralateral eye remained untreated. Mice were sacrificed about 22 weeks after vector administration. Retina from the untreated eye displayed very weak GFP staining and few BRN3A-positive cells in the GCL. Conversely, retinal section from AAV2/2-NGB treated eye showed strong GFP and BRN3A labeling, noticeable throughout the GCL; moreover, GFP staining is also noted in the NFL. Noteworthy the overall number of BRN3A-positive cells was higher in treated retinas than in untreated ones.

    2.1.6 Microscopic Observations

    [0188] Fluorescence labeling was monitored in the Cellular Imaging Facility of the Institute with: (i) a confocal laser scanning microscope (Olympus FV1000) Microscope, image acquisition was conducted by using Olympus Fluoview® software version 3.1. (ii) Retinal sections were also scanned with the Hamamatsu Nanozoomer Digital Pathology (NDP) 2.0 HT, its Fluorescence Unit option (L11600-05) and the NanoZoomer's 3-CCD TDI camera (Hamamatsu Photonics, France). BRN3A-positive cells, as the estimation of overall RGCs, were assessed for each animal by manually counting 2-4 entire retinal sections as described earlier (cf. example 1). Lastly, all the images were analyzed with Photoshop and Image J.

    2.1.7 RNA Extraction and RT-qPCR Assay

    [0189] Total RNA from rat retinas were extracted using RNeasy Plus Mini kit from Qiagen. To ensure the absence of DNA a treatment with RNase-free DNase (Qiagen) and a subsequent cleanup with the RNeasy MinElute cleanup kit (Qiagen) were performed. This was confirmed by subjecting 10 ng of each RNA preparation to qPCR with specific primers for the NGB transgene and the mitochondrial ATP6 gene. One micrograms of total RNA was reverse transcribed with oligo-dT using Superscript® II Reverse Transcriptase (Life Technologies). Quantitative PCR reactions were performed using ABI 7500 Fast (Applied Biosystems) and the specific primers listed on Table 7.

    TABLE-US-00007 TABLE 7 Primers for RT-qPCR assays Primer Primer Forward reverse Gene 5′- 3′ 5′- 3′ NGB CTCAGGCAAG CAGTTAGGTT GGAAGCATAG TCCCCCAAAA (SEQ ID  (SEQ ID NO: 10) NO: 11) NGB-AAV AGGCTATGTC GGGTAACCCT ACGAGGTTGG ATGCAGTCGT (SEQ ID  (SEQ ID NO: 12) NO: 13) ATP6 CGTAATTACA AGCTGTAAGC GGCTTCCGAC CGGACTGCTA A (SEQ ID (SEQ ID NO: 15) NO: 14) SNCG GGAGGCAGCT ACTGTGTTGA GAGAAGACC CGCTGCTGAC (SEQ ID (SEQ ID NO: 16) NO: 17) GFAP CCCGTTCTCT CTTCAGGGCT GGAAGACACT GAGAGCAGTC (SEQ ID (SEQ ID NO: 18) NO: 19) AIF1 GGGGGCAAAA CTGTTTCTCT TGGATAATTC TCTGGGGACAG (SEQ ID (SEQ ID NO: 20) NO: 21)

    [0190] The equivalent of 10 ng and 2 ng of cDNAs were used per gene as template for qPCR reactions with Power Sybr® green PCR Master Mix (Applied Biosystems). Each biological sample was subjected to the assay in triplicates per gene; Ct values were obtained with the ABI 7500 software (v.2.0.6). The comparative ΔΔCt method and the mitochondrial ATP6 gene have been used to determine the relative mRNA amount of each studied gene. The mitochondrial ATP6 gene has been used as normalizing gene since its mRNA steady-state levels remained almost unchanged in all the samples evaluated.

    2.1.8 Mitochondria Extraction and Western Blotting Analysis

    [0191] Thirty six retinas were isolated from 12 week-old mice (B6CBACa strain) and washed in PBS at 4° C.; they were then homogenized in extraction buffer (0.32 M sucrose, 30 mM Tris-HCl; pH 7.6, 5 mM MgAc, 100 mM KCl, 0.1% fatty acid-free BSA, 5 mM β-mercaptoethanol, and 1 mM PMFS) and mitochondria were purified as previously described for rat retinas (cf. FIG. 1 and example 1). 30 μg of mitochondrial proteins were treated with 150 μg/mL of Proteinase K (PK) in the presence or absence of 1% Triton X-100 (v/v) at 4° C. for 30 minutes. The reaction was stopped by addition of 1 mM PMFS (Sigma-Aldrich). Samples were collected by centrifugation at 10,000 g for 15 minutes at 4° C. Three independent mitochondria purifications were subjected to the analyses.

    [0192] For whole proteins extracts, single retinas were homogenized in 50 μL of 20 mM HEPES and 60 mM mannitol (pH 7.2) using a 200 μL micro-hand-driven glass-glass potter at 4° C. Large cellular debris was spun down by a low speed centrifugation (1000 g for 5 minutes at 4° C.) and supernatants were used immediately for Western Blotting. Protein quantification was performed using the Bradford method (Bradford reagent from Sigma-Aldrich). After denaturation at 94° C. for 15 minutes, samples were resolved in 12% or 15% SDSPAGE and next transferred to a PVDF membrane. Membranes were probed with antibodies against NGB, AIF and ATP synthase subunit β (cf. Table 6). Immunoreactive bands were visualized with anti-mouse or anti-chicken coupled to horseradish peroxidase (0.1 mg/mL) followed by detection with Pierce® ECL Plus Western Blotting Substrate (Pierce, Thermo Scientific).

    2.1.9 Tissue Homogenate Preparation and Respiratory Chain Enzyme Assays

    [0193] Optic nerves were rapidly collected and kept frozen (−80° C.). Respiratory chain complex I and V enzymatic activities were measured using a Cary®50 UV-Vis spectrophotometer (Agilent technologies), as described for optic nerves from mice (Bouaita A, et al. (2012) Brain 135: 35-52) and each assay was made in triplicate. Complex I (CI) and Complex V (CV) values were converted to specific activities expressed as nanomoles of oxidized NADH/min/mg protein after protein quantification by the Bradford method. All chemicals were of the highest grade from Sigma-Aldrich.

    2.1.10 Statistical Analyses

    [0194] Values are expressed as means±SEM (Standard Error of the Mean). Statistical analyses were performed with the GraphPad Prism 6.0 software assuming a confidence interval of 95%. Generally, the observations within each group do not fit in a normal distribution, thus nonparametric methods have been applied for evaluating the significance. Data collected from control and Harlequin were compared using the unpaired non parametric significance test of Mann-Whitney (*≤0.05, **≤0.01 and ***≤0.005). Data collected from Harlequin treated eyes and untreated controlateral eyes were compared using the paired non parametric significance test of Wilcoxon (*≤0.05, **≤0.01 and ***≤0.005).

    2.2 Results

    2.2.1 Neuroglobin Expression in Control and Harlequin Mouse Retinas

    [0195] It has been previously shown that NGB localizes to the mitochondria in rat retinas (cf. example 1). To examine the subcellular distribution of the NGB in adult mouse retinas, mitochondrial enriched fractions have been prepared by differential centrifugation and performed Western blot analysis (data not shown). Antibodies against NGB recognized three proteins with apparent molecular masses of about 17, 19 and 21 kDa in homogenates and mitochondria; in this latter the abundance is very high as observed for ATP synthase-β (a subunit of respiratory chain complex V). Interestingly, very discrete NGB signals were evidenced in: (i) the pellet (encompassing nuclei and unbroken cells); (ii) the high speed supernatant (obtained after the 10,000 g centrifugation which spun down mitochondria) indicating that NGB is enriched in the mitochondrial compartment. To further evaluate the amount of NGB which is translocated within the organelle, mild proteolysis with Proteinase K (PK) was performed.

    [0196] The three forms of NGB and the ATP synthase-β gave strong signals indicating that these proteins were integrated into the mitochondria. Next, mitochondrial fractions were treated with PK in the presence of Triton X-100; the detergent disrupts mitochondrial membranes and leads to the entire proteolysis of mitochondrial proteins. ATP synthase-β and NGB protein signals were considerably diminished, confirming their localization inside the organelle in a protease-sensitive form (data not shown).

    [0197] Therefore, NGB localizes to the mitochondria of mouse retinas as the inventors have previously shown in rat retinas.

    [0198] To investigate, the abundance of NGB in RGCs, flat mounted retinas from 6 month-old mice were immunostained for NGB, BRN3A and the mitochondrial NDUFB6 protein, a complex I subunit (data not shown). BRN3A is a nuclear factor exclusively expressed by most of the RGCs in rodent retinas (Nadal-Nicolas et al., PLoS One. 2012; 7(11)); fluorescence microscopy of flat mounted retinas immunodetected for BRN3A showed many stained nuclei distributed throughout the retina. All the BRN3A-positive cells showed an intense NGB labeling as punctuate dots in the cytoplasm. In addition, some NGB positive-cells were not immunostained for BRN3A; they may correspond to displaced amacrin cells or astrocytes (data not shown). When antibodies against NGB and NDUFB6 were combined the majority of cells exhibited similar labeling patterns indicating some extent of colocalization between the two proteins (data not shown). Overall, signals appeared as strong punctuate fluorescent dots in the cytoplasm and apposed to the nuclei; thus distribution of both proteins in RGC bodies is comparable to the one described for mitochondria in mouse cells residing in the outer nuclear layer (ONL) (Johnson J E, et al. (2007) Mol Vis 13: 887-919).

    [0199] NGB distribution in other cell populations was analyzed in radial cryosections of retinas immunostained for NGB in control and Hq mice aged 6 months (data not shown). In normal mice, it appears that the protein was particularly abundant in the ganglion cell layer (GCL) (data not shown) and the inner segments (IS) of photoreceptors (PRs), while very weak labeling with the NGB antibody was observed in PR outer segments (OS). The inner and outer plexiform layers (IPL and OPL) showed strong to moderate punctuate labeling; at the junction between the ONL and the OPL, the staining could represent the mitochondria in the synaptic terminals of PRs and in horizontal cells. The inner nuclear layer (INL) is usually divided in three regions; distal, middle and proximal, NGB immunostaining is strong in the three levels, especially in the distal region, some of the fluorescent cells were located at the very inner margin of the INL, they could be amacrine cells as previously described. Overall, NGB labeling in retinal neurons were consistent with the abundance of mitochondria in the different retinal compartments.

    [0200] On the other hand, retinal sections from Hq mice showed a considerable reduction in the thickness of all the neuron layers with a significant diminution of NGB immunostaining (data not shown). In the GCL, the overall cell number is strongly reduced in Hq retina; as a consequence the number of NGB-positive cells was lower than in control retinas and the pattern of staining within one cell is less strong relative to the NGB-positive cells in control retinas (data not shown). To further investigate the reduction of NGB amount in Hq retinas, NGB mRNA steady-state levels in retinas from Hq mice was determined by RT-qPCR. Total RNA preparations from 37 Hq and 31 control retinas from 6 month-old mice were examined. The relative amount of NGB mRNA was 45% less abundant in Hq retina than in controls; the difference was significant (P<0.0001) (FIG. 7).

    [0201] Whole protein extracts from Hq and control retinas were subjected to Western blotting analysis to corroborate this data. In whole extracts from control retinas the 17 kDa form gave the more intense signal; very weak signals from the 19 and 21 kDa was detected; indeed these NGB forms are enriched in mitochondrial fractions from both rat and mouse retinas (data not shown). Moreover, in the inventors' experience; the abundance of the three NGB forms varies in independent mitochondrial extractions (cf. example 1). This result can be explained by the highly flexible structure of the NGB which leads to great variations in protein conformation depending on exogenous ligand availability, pH changes or rupture/formation of the intra molecular disulfide bond.

    [0202] In Hq retinas, NGB signals were strongly reduced compared to control retinas; as expected AIF was almost undetectable while ATP synthase-β amounts remained unchanged (data not shown). These results confirm the overall reduction of NBG in Hq retinas relative to age-matched controls, both at the level of mRNA and protein, thus the inventors has endeavor to re-establish NGB mRNA levels in Hq retinas to determine the impact on RGC integrity.

    2.2.2 Design of a Gene Therapy using Neuroglobin for Preventing Retinal Damage in Harlequin Mice

    [0203] Ocular administration of recombinant AAV2 vectors did not produce adverse effects in mammals; in addition serotype 2 transduce highly efficiently inner retina, principally RGCs. Thus, a recombinant AAV2/2 possessing the mouse NGB ORF associated with the full-length 5′ and 3′ UTRs of the gene was constructed to ensure mRNA stability and translation capacity (FIG. 8). In an attempt to prevent RGC and optic nerve degeneration, a single intravitreal injection with the AAV2/2-NGB vector (2×10.sup.9 VG per eye) was performed in Hq mice 4-6 week-old; at this age the extent of RGC injury is minimal. Forty mice were evaluated within the course of this study; animals were euthanized between 5-6 months after vector administration. Since NGB sequences were inserted into the pAAV-hrGFP vector the GFP was used to qualitatively assess transduction efficiency (data not shown). The reconstruction of whole retinal sections from injected eye clearly shows a homogenous and intense GFP immunoreactivity mainly restricted to the GCL; On the contrary, the untreated eye displayed very little GFP labelling (data not shown). This data was strengthened when the NGB protein synthesized from the recombinant AAV2/2 vector was seeked in retinal sections using a specific antibody against the Flag epitope. Indeed, as shown in FIG. 8 the NGB ORF is in frame with the 3× FLAG® sequence at the C-terminus. In retinal sections from the AAV2/2-NGB treated eyes, many strong immunoreactive signals were only observed in the GCL while endogenous NGB staining in the other retinal layers was similar in Hq-treated and untreated eyes (data not shown). Hence, intravitreal administration of AAV2/2-NGB results in a highly efficient gene delivery to cells within the GCL, as previously described (Hellstrom M, et al. (2009) Gene Ther 16: 521-532). Next, the relative amount of the transduced NGB mRNA has been evaluated by real-time quantitative PCR in retinas from 14 Hq mice subjected to vector administration about 22 weeks before euthanasia (FIG. 9B). NGB mRNA transcribed from the recombinant AAV2/2 has been evidenced in total RNA preparations from retinas of injected eyes. The steady-state levels of the transduced NGB mRNA contributed to an approximate 3-fold increase of the overall amount of NGB mRNA relative to the one measured in control mouse retinas while no change was evidenced in the relative steady-state levels of AIF1 mRNA between retinas from treated and untreated eyes P=0.3 (FIG. 9A). Thus, AAV2/2-NGB administration to Hq eyes led to NGB overexpression essentially in resident GCL cells without adverse noticeable effects on mouse eyes up to 6 months.

    2.2.3 Neuroglobin Overexpression Limits Gliosis Reaction in Retinas from Harlequin Mice

    [0204] One prominent feature of the progressive retinal degeneration in Hq mice is glial cell activation, hence glial fibrillary acidic protein (GFAP) showed a significant increase, which begins in mice aged about 4 months. Mice older than 7 months exhibited a widespread GFAP immunoreactivity across the entire retinal thickness. To determine whether NGB overexpression could prevent the active growth of Müller cell processes, retinal sections from treated and untreated eyes were subjected to immunochemistry for GFAP and NGB. In control retinas GFAP immunofluorescence was confined exclusively to the GCL, corresponding to the end-feet of Müller cells and astrocytes resident in this cellular layer (data not shown). In Hq mice 6-7 month-old, GFAP immunoreactivity was markedly increased and was not just restricted to the GCL but also found in the Muller cell processes; intense GFAP-stained cell processes extended across the entire thickness of the retina from the untreated eye (data not shown). The steady-state levels of the GFAP mRNA also increased 6.7-fold relative to age-matched controls, confirming that an extensive glial response occurred in Hq retinas concomitantly with RGC degeneration (FIG. 10). In the retinal section from the AAV2/2-NGB treated eye, NGB labeling was noticeably enhanced in the GCL, substantiating the high transduction efficiency (data not shown). In opposition, GFAP immunoreactivity was distinctly less increased relative to controlateral untreated eyes, indeed the end-feet and processes of Muller cells were moderately labeled (data not shown). Moreover, the relative GFAP mRNA abundance between treated and untreated retinas diminished of 16% (FIG. 10), the difference was statistically significant (n=9, P=0.0039). Therefore, NGB overexpression hinders GFAP upregulation in Muller glia; this subtle change may be beneficial for RGC survival in AAV2/2-NGB treated eyes.

    2.2.4 Retinal Ganglion Cell Loss in Harlequin Mice is Prevented by the Intravitreal Administration of AAV2/2-NGB

    [0205] Retinal sections from Hq has been examined by immunochemistry using antibodies against NGB and BRN3A. Retinal sections of treated eye showed a noticeably increase of NGB immunostaining specifically in the GCL relative to the signals observed in the contralateral untreated eyes, many cells were BRN3A-positive (data not shown). More than 95% of the BRN3A-positive cells displayed a strong staining for NGB, hence, confirming the efficiency of AAV2/2-NGB on transducing RGCs. Retinal sections from the same animals were subjected to immunochemistry using antibodies against GFP and BRN3A (data not shown); the labeling obtained confirmed the high efficiency of vector transduction. Remarkably, the higher number of BRN3A-positive cells in the treated eye relative to the untreated one was noticeable as was that many of these cells were intensely GFP-labeled in their somas and processes. Besides, it was also visible GFP staining in the NFL and the INL; this latter can correspond to transduced Muller glial cells (data not shown). Hence, to corroborate the beneficial effect of NGB overexpression on RGC integrity, the quantification of RGC somas in retinal sections from Hq treated and untreated eyes has been proceeded subjected to immunostaining for BRN3A. The inventors have previously shown that by the age of 7-8 months Hq mice loss up to 36% of their overall RGC population (Bouaita A, et al. (2012) Brain 135: 35-52). Here the RGC density has been estimated in whole retinal sections for 24 control mice aged ˜7 months and compared to 13 Hq mice in which one eye was subjected to AAV2/2-NGB administration. The inventors confirmed that RGC somas were significantly reduced in retinas from Hq untreated eyes: 35±3/mm in Hq retinas relative to 58±2/mm in control retinas; a 40% diminution of the total amount of BRN3A-positive cells in the GCL has been observed (FIG. 11A). Uttermost, the number of RGCs in treated eyes per mm was 43±2; 24% higher than in contralateral untreated eyes from the same animals; these difference was statistically significant, P=0.0002 (FIG. 11A) and attained 75% of the number measured in control retinas.

    [0206] Next, the relative amount of γ-synuclein (SNCG) mRNA has been determined by subjecting RNA preparations from 14 couples of Hq mouse retinas and 18 age-matched control retinas to RT-qPCR analysis. SNCG mRNA is considered as a very abundant molecule in adult mouse RGCs and it has been used as a marker of injured RGCs (Bouaita A, et al. (2012) Brain 135: 35-52). As previously shown, the steady-state level of SNCG mRNA was significantly diminished up to 43% of control value in Hq retinas from untreated eyes, P=0.0001 (FIG. 11B). In contrast, a consistent and statistically significant increase of 12.6% in its relative abundance was measured in treated eyes relative to the one measured in their contralateral untreated eyes (P=0.0004); although SNCG mRNA steady-state levels remained statistically different between treated eyes and controls, P=0.0002 (FIG. 11B).

    [0207] Thus, intravitreal administration of AAV2/2-NGB prevented RGC loss and hindered SNCG downregulated expression in Hq retinas.

    2.2.5 Neuroglobine Gene Overexpression Protected Nerve Fiber Integrity in Harlequin Mice

    [0208] To substantiate that NGB overexpression in RGCs from Hq mice impeded, at a non negligible extent, their degeneration, both the amount of their axons as well as RCCI activity in optic nerves (ONs) have been evaluated. First, ON cross-sections were subjected to immunohistochemistry for the heavy chain (200 kDa) subunit of neurofilaments (NF200) to detect RGC axons. An obvious reduction of immunopositive dots in ONs from untreated Hq eyes relative to ONs isolated from age-matched control mice has been observed (data not shown). The axonal profiles detected in ON cross-sections from treated Hq eyes confirmed that they displayed a noticeably increase in NF200-immunopositive signals relative to untreated eyes (data not shown). This indicates that AAV2/2-NGB administration to Hq mouse eyes attenuates RGC axonal damage and corroborates data on their overall number preservation (FIG. 11A). In an attempt to establish a functional link between RGC number preservation and respiratory chain integrity, a spectrophotometric method for assessing respiratory chain complex activities has been utilized. This method has been successfully applied to accurately study respiratory chain in small amounts of tissues from Hq mice (Bouaita A, et al. (2012) Brain 135: 35-52); FIGS. 12A and 12B illustrate CV and CI activity measurements for 3 mouse groups: (1) 36 ONs from control mice aged 6-7 months; (2) 24 ONs from Hq eyes subjected to AAV2/2-NGB intravitreal injection and euthanized between 5 to 6 months after vector administration; (3) 24 ONs from the Hq contralateral untreated eyes. In a previous study on Hq mice, the inventors assessed respiratory chain function in ONs and they demonstrated that AIF depletion leads to a severe CI defect without affecting CV activity (Bouaita A, et al. (2012) Brain 135: 35-52). Here, the inventors show that CV activity was increased of about 22% in ONs from Hq relative to controls independently of AAV2/2-NGB treatment (FIG. 12A). Indeed, the difference between Hq values and controls was statistically significant (P=0.0006 and 0.0002 for untreated and treated eyes respectively) while the difference of CV activities in ONs from untreated and treated eyes was not significant (P=0.9).

    [0209] Nonetheless, it was obvious from the current assessments that ONs from Hq untreated mice ˜7 month-old manifested a diminution in CI activity (expressed as nanomoles of oxidized NADH/min/mg protein) when compared to isogenic age-matched controls (FIG. 12B). The reduction was significant: 8.4±0.6 versus 14.8±1; 43% of the control value (P<0.0001). Noticeably, the ocular administration of AAV2/2-NGB protected efficiently CI function; indeed its enzymatic activity was statistically different from values obtained in ONs from untreated eyes (P<0.0001).

    [0210] The specific activity of CI (11.6±0.6) attained in ONs from treated eyes 78% of the value measured in control mice; further, ONs from NGB-treated eyes exhibited 38% higher CI activity than ONs from their contralateral untreated eyes. Thus, high levels of NGB in RGCs from Hq mice rescued compromised CI activity; the inventors predict that this improvement was involved in the RGC robustness evidenced (FIG. 11).

    2.2.6 Sustained Preservation of Nerve Fibers, due to Neuroglobin Gene Overexpression, Confers Improvements in Harlequin Mouse Vision

    [0211] To make the most trustworthy proof-of-concept on the protective effect of NGB against RGC degeneration besides indications gathered on cell loss prevention and CI activity protection it is also required to confirm the presence of fibers bundles in eyes fundus and their ability to transfer visual inputs to the visual cortex. Hence, a thorough evaluation of eye fundus has been first performed using Confocal Scanning Laser Ophthalmoscopy (cSLO), a reliable method for in vivo cellular imaging in the retina (Paques M, et al. (2006) Vision Res 46: 1336-1345). Imaging of the nerve fiber layer (NFL) was facilitated by the use of pigmented mice and the high contrast between the fiber bundles (RGC axon packages) and the dark background. Striations of NFL radiating from the optic disc were clearly visible in each eye from mouse before AAV2/2-NGB administration for all Hq mice 6 week-old (data not shown). Subsequently, each area of the eye fundus (nasal, temporal, inferior and superior) was visualized monthly until euthanasia to follow over time nerve fiber disappearance and seek for any change related to NGB overexpression.

    [0212] Six months after vector administration, a substantial loss of nerve fiber bundles was evidenced in almost all the untreated eyes; in general we can observe that about half of the entire retinal surface gave the impression of being devoid of fibers that could be either in the temporal or nasal areas. In contrast, images collected for more than the half of eyes that received AAV2/2-NGB revealed significantly well preserved axon tracks in all the areas visualized. Up to date 40 mice have been extensively evaluated using cSLO; more than 80% of the untreated eyes did reveal RGC axon degeneration in mice from the age of 4 months; while 22 eyes out of the 40 subjected to AAV2/2-NGB treatment exhibited high densities of fiber bundles in all the areas examined up to 6 months after vector administration. Thus; NGB overexpression efficiently protected Hq mice against optic nerve degeneration.

    [0213] Ultimately, to address the overriding question of whether AAV2/2-NGB administration confers improvements in vision to Hq mice, visual function of young animals has been assessed at different times after gene therapy by studying their optomotor responses to rotating sinusoidal gratings (OptoMotry™). Visual acuities were meticulously measured in isogenic age-matched controls and Hq mice to gather precise thresholds (highest spatial frequency each mouse could track) under our experimental conditions. Tracking capability was examined in both clockwise and counter clockwise drum rotations at different frequencies because only temporal-to-nasal motion is effective through each eye, clockwise movement will drive tracking through the left eye, whereas counterclockwise motion will activate the right eye. Optomotor responses recorded from Hq mice aged 4-8 weeks (before any ocular intervention) and wild-type mice of the same age were similar: for the clockwise (left eye sensitivity) and the counterclockwise (right eye sensitivity) pattern of rotations; indeed values indicated no significant difference: P=0.06 and 0.44 respectively (FIG. 13A). This data suggests that visual function of Hq mice was not compromised during the first weeks of life and is in agreement with cSLO examinations in which fiber disappearance was noticeable in Hq mice aged 4 months or older (data not shown). Next the evaluation of Hq mice which received in their left eyes AAV2/2-NGB has been proceeded; they were subjected to the test twice, at about 3 months post-injection and just before euthanasia (up to 6 months post-injection). FIG. 13B illustrates clockwise and couterclockwise visual acuities from Hq mice and age-matched control mice. Very reduced head-tracking behavior (counterclockwise responses) was observed in the untreated eyes of Hq mice aged of ˜7 months; indeed Hq mice visual acuity from the left eyes (n=18) was significantly poorer than control mice (n=22): 0.175±0.03 versus 0.44±0.018 cycles per degree (P<0.0001). Impressively, a visual-acuity threshold (clockwise response) of 0.35±0.03 cycles per degree was recorded in Hq treated eyes; the decrease in visual performance in the untreated eyes (counterclockwise responses) was evidenced at the age of 4-5 months; though the responses recorded from the treated-eyes remained stable until euthanasia (FIG. 13B). Visual acuity in Hq treated eyes was significant different from the one measured in untreated eyes (P<0.0001); and represents a value 99% superior to the one from untreated eyes. Thus, this data confirms that NGB overexpression benefit encompassed the long-lasting protection of RGCs and their axons along with the maintenance of their functional integrity.

    Example 3

    3.1 Material and Methods

    3.1.1 Animals and Diet

    [0214] Same animals of example 2 (see paragraph 2.1.1) have been used in this study i.e. Harlequin mice.

    3.1.2 Adeno-Associated Viral vector and Intravitreal Injections

    [0215] The vector named AAV2/2-NGB (SEQ ID NO: 9) (see example 2.1.2) was used in this study together with the vector AAV2/2-AIF1 described below.

    [0216] The entire Mus musculus apoptosis-inducing factor, mitochondrion-associated 1 (Aifm1) mRNA sequence (http://www.ncbi.nlm.nih.gov/nuccore/NM_012019) of 1926 base pairs (bp) was synthesized by Genscript Corp (Piscataway, N.J. 08854 USA), encompassing the original 87 bp of the 5′ UTR, the entire ORF encoding a 612 amino acid-long protein and two restriction sites at the extremities: EcoR1 at the 5′ and XhoI at the 3′ for cloning into the pAAVIRES-hrGFP vector (Stratagene, California, U.S.A) in which we had earlier replaced the hGH (human growth hormone 1 [MIM 139250]) polyadenylation signal with the 176 bp full-length AIF1 3′UTR (http://www.ncbi.nlm.nih.gov/nuccore/NM_012019) by using BgIII and RsrII unique restriction sites.

    [0217] For intravitreal injections Hq mice aged between 3 to 4 months were subjected to anesthesia with isoflurane (40 mg/kg body weight). The tip of a 33-gauge needle, mounted on a 10 μl Hamilton syringe (Hamilton Bonaduz AG, Bonaduz, Switzerland) was advanced through the sclera and 2 μL of vector suspension (2×109 VG for NGB and 5×108 for AIF) was injected intravitreally, avoiding retinal structure disruption, bleeding or lens injury. Fifty-two Hq mice were subjected to intravitreal injection, half of them received in one eye AAV2/2-NGB (2×10.sup.9 VG) and the other half AAV2/2-AIF1 (5×10.sup.8 VG); they were euthanized between 6 to 7 months after vector administration.

    3.1.3 Slit-Lamp Examination and In Vivo Confocal Microscopy Analysis

    [0218] Anterior chamber phenotypes were assayed using a slit-lamp biomicroscope (DC-3, Topcon, Clichy, France) and photodocumented using a digital camera (D100; Nikon, Tokyo, Japan). All ocular exams utilized conscious mice and the anterior chamber was examined for iris stromal atrophy, pigment dispersion, and dark iris appearance. All photographs were taken with identical camera settings, images were collected at 25× magnification. Additionally, a laser-scanning in vivo confocal microscopy (IVCM) Heidelberg Retina Tomograph (HRT) II/Rostock Cornea Module (RCM; Heidelberg Engineering GmbH, Heidelberg, Germany) was used to examine the entire cornea (Pauly A, et al., Invest Ophthalmol Vis Sci. 2007; 48:5473-5483), including superficial epithelium (depth: 0 μm), basal epithelium (8-15 μm), stroma (15-40 μm) and endothelium (65-80 μm). Before proceeding to AAV2/2 administration, Hq mice aged between 1-2 months were extensively evaluated with slit-lamp biomicroscope and laser-scanning in vivo confocal microscopy to discard animals exhibiting corneal dystrophy. In about 10 to 15% of Hq mice, various abnormalities in the cornea appeared early in life such as neovascularisation, oedemas, inflammation, epithelial invaginations, hyper-reflective deposits in the stroma and the epithelium as well as abnormal prominent and tortuous corneal nerves. These mice, as well as the few which developed cataracts (˜1-2% of mouse population in our colony) were not subjected to AAV2/2 administration.

    3.1.4 Fundus Imaging by Confocal Scanning Laser Ophthalmoscopy

    [0219] A digital confocal Scanning Laser Ophthalmoscope (cSLO) was used as described in example 2 (see paragraph 2.1.3). The overall density of nerve fiber bundles was used as a criterion for selecting the eye which will be treated in each mice; fundus imaging was started in 6 week-old animals and repeated each 2-3 weeks to evidence fiber thinning or disappearance. When this first sign of RGC injury was noticed at a different extent in both eyes and in combination with visual performance of each eye; the worst responding eye was selected for AAV2 administration. Usually, the treatment was performed in 12-16 week-old mice.

    3.1.5 Optomotor Responses

    [0220] Optokinetic tracking threshold was measured, as described in example 2 (see paragraph 2.1.4) by observing the optomotor responses of mice to rotating sinusoidal gratings (OptoMotry™).

    [0221] Experiments were performed in animals aged 2 months; they were repeated three times weekly until a clear reduction in visual performance was measured in order to establish which eye could be subjected to the treatment. Then, visual performance was assessed 3 and 6 months post-injection by subjecting mice to the test three times weekly during two weeks. Generally, 2-3 different persons perform the experiments, they were masked to the animal's treatment and previously recorded thresholds.

    3.1.6 Retinal and Optic Nerve Histology

    [0222] Retinas and optic nerves (ONs) and immunochemistry were done as in example 2 (see paragraph 2.1.5). Primary and secondary antibodies used are shown in Table 8.

    TABLE-US-00008 TABLE 8 Antibody description Antibody Type Concentration Supplier, reference NGB (histology) Polyclonal    5 μg/mL Sigma, N-7162 NGB (Western blot) Polyclonal    1 μg/mL Biovendor, RD181043050 AIF (Western blot) Monoclonal  0.4 μg/mL Millipore, AB16501 ATP synthase-subunit α Monoclonal  0.2 μg/mL Invitrogen, LifeTechnologies, (Western blot)   7H10BD4F9 SOD2 (Western blot) Polyclonal  0.4 μg/mL Abcam, 13533 β-Actin (Western blot) Monoclonal  0.2 μg/mL Sigma Aldrich, A5316 NDUFA9 (Western blot) Monoclonal    1 μg/mL LifeTechnologies, 459100 BRN3A (histology) Monoclonal    1 μg/mL Chemicon, MAB1585 GFAP (histology) Polyclonal  2.9 μg/mL Sigma Aldrich, G3893 NF200 (histology) Monoclonal    1 μg/mL Chemicon, MAB1585 IBA1 (histology) Polyclonal  0.5 μg/mL Wako, 019-19741 Alexa 488 Rabbit    4 μg/mL Invitrogen, LifeTechnologies,   A11008 Alexa 594 Mouse    4 μg/mL Invitrogen, LifeTechnologies, A11005 Goat Anti-Rabbit IgG HRP 0.05 μg/mL Jackson ImmunoResearch conjugate Laboratories, 111-035-003 Donkey Anti-Chicken HRP 0.05 μg/mL Jackson ImmunoResearch IgG conjugate Laboratories, 703-035-155 DAPI (4′,6-Diamidino- Nucleic    2 μg/mL Invitrogen, LifeTechnologies, 2-Phenylindole, Acid Stain D1306 Dihydrochloride)

    3.1.7 Microscopic Observations

    [0223] Microscopic observations were done as in example 2 (see paragraph 2.1.6).

    3.1.8 Transmission Electron Microscopy

    [0224] Twelve Hq and nine control mice aged from 6 weeks to 15 months were used for ultrastructural studies. Mice were anesthetized by the intraperitoneal administration of ketamine (100 mg/kg) and xylazine (8 mg/kg) and transcardially perfused with 0.9% NaCl for 30 seconds and then with Karnovsky fixative (paraformaldehyde 2%, Glutaraldehyde 2.5% in 0.1 mol/L phosphate buffer, pH 7.4) for 12 minutes. ONs were removed and postfixed in the same fixative for 1 hour at 4° C. and stored in PBS overnight at 4° C. The samples were then rinsed briefly in water, fixed in 2% aqueous OsO4 for 45 minutes at 4° C., and finally rinsed in water. Samples were dehydrated in a series of graded ethanol solutions (70%, 95%, and 100%, 3×10 minutes for each), then in a mixture 1/1 (v/v) of ethanol and propylene oxide (10 minutes) and finally in pure propylene oxide (3×10 minutes). Next, the specimens were embedded in Epon (Electron Microscopy Sciences, Hatfield, Pa., USA) and propylene oxide at 1:1 for 2 hours and pure Epon for 2×12 hours at room temperature. Ultrathin (80 nm) sections were prepared using a Leica ultracut S microtome fitted with a diamond knife (Diatome histoknife Jumbo or Diatome Ultrathin). Both transverse and longitudinal sections from distal segments of optic nerves were obtained. The sections were contrast-stained with 2% uranyl acetate and lead citrate and photographed in a Jeol S100 transmission electron microscope (Croisy-sur-Seine, France) fitted with an Orius SC200 digital camera (Gatan-Roper Scientific, Evry, France) for image capture.

    3.1.9 RNA Extraction and RT-qPCR Assay

    [0225] RNA extraction and RT-qPCR assay were done as in example 2 (see paragraph 2.1.7), specific primers are listed on Table 9.

    TABLE-US-00009 TABLE 9 Primers for RT-qPCR assays Primer Primer Forward reverse Gene 5′-3′ 5′-3′ NGB CTCAGGCA CAGTTAGG AGGGAAGC TTTCCCCC ATAG AAAA (SEQ ID (SEQ ID NO: 10) NO: 11) BRN3A AGGCCTAT CGTCTCAC TTTGCCGT ACCCTCCT ACAA CAGT (SEQ ID (SEQ ID NO:30) NO: 31) ATP6 CGTAATTA AGCTGTAA CAGGCTTC GCCGGACT CGACA GCTA (SEQ ID (SEQ ID NO: 14) NO: 15) SNCG GGAGGCAG ACTGTGTTG CTGAGAAG ACGCTGCTG ACC AC (SEQ ID (SEQ ID NO: 16) NO: 17) GFAP CCCGTTCT CTTCAGGGC CTGGAAGA TGAGAGCAG CACT TC (SEQ ID (SEQ ID NO: 18) NO: 19)

    3.1.10 Tissue Homogenate Preparation and Respiratory Chain Enzyme Assays

    [0226] Tissue homogenate preparation and respiratory chain enzyme assays were performed as in example 2 (see paragraph 2.1.9). More particularly when the tests were performed tissues were prepared at 4° C. by homogenization with a 1 mL hand-driven glass-glass potter in 100 μL of extraction buffer (0.25 mM sucrose, 40 mM KCl, 2 mM EGTA, 1 mg/mL BSA, and 20 mM Tris-HCl, pH 7.2). Large cellular debris were spun down by a low speed centrifugation (1000 g×8 min) and supernatants were used immediately.

    [0227] The following measurements were performed: (1) rotenone-sensitive complex I or NADH decylubiquinone reductase activity; (2) the ATP hydrolase activity of complex V which is oligomycin-sensitive; the maximal complex V activity and maximal oligomycin effect are only measured after a minute.

    3.1.11 Statistical Analyses

    [0228] Statistical analyses were done as in example 2 (see paragraph 2.1.10).

    3.2 Results

    3.2.1 Time Course of Optic Nerve Degeneration in Harlequin Mice

    [0229] To study the time course of optic nerve damage in Hq mice, ultrastrucutural changes were evaluated in transversal and longitudinal sections using transmission electron microscopy (TEM). Intraorbital unmyelinated axons as well as proximal (unmylinated) and distal (myelinated) axons were evaluated in 1, 3, 6 and 12 month-old Hq and control mice. Very few ultrastructural differences were noticed in Hq and control mice at the age of 4-6 weeks, except for the presence in Hq mice of few swollen axons (FIGS. 14A and B, middle panel). Overall, no alterations were observed in control mice during aging, thus FIG. 14A illustrates data collected from 1 and 12 month-old animals. Alterations reflecting axonal degeneration were noticed in all the 3 month-old Hq mice evaluated (n=4): some axons in the distal part appeared shrunken, cristae within mitochondria disappeared, and vacuoles could be seen between the axolemmas and myelin sheaths, while adjacent to them other axons still appeared normal at this age. These features aggravated in older mice (6 and 12 months; 3 and 6 mice evaluated) since only few scattered fibres showed normal appearance and the overall number of axons left was very small (FIG. 14B). Additionally, the invasion of astrocytes and microglial cells, in the empty space, disorganized further the nerve structure. The progression of the axonal pathology in Hq mice evidenced by TEM (FIG. 14B) shares similarities with Wallerian degeneration in the form of dark (hyperdense axoplasms filled with dark material) and watery (axonal swelling) degeneration with demyelination (vacuolation and splitting of the myelin sheath (Saggu S K et al., BMC Neurosci., 2010, Vol. 11 Issue). Thus, 3 month-old Hq mice exhibit early ultrastructural changes in optic nerve axons which occur prior to axon degeneration.

    [0230] To determine when the ultrastructural changes detected in RGC axons from 3-month old Hq mice could be associated with abnormal eye fundus imaging, Hq mice aged between 6 weeks and 4 months were subjected to thorough evaluations using confocal Scanning Laser Ophthalmoscopy (cSLO) (Paques M, et al. (2006) Vision Res 46: 1336-1345). The technique enables the generation of convenient en face (xy) high-resolution and high-contrast imaging of the retina. Because of the absence of reflected light from the choroid and sclera in pigmented mice the contrast between the fiber bundles (RGC axon packages) composing the RNFL and the dark background is increased (Bouaita A, et al., Brain. 2012; 135: 35-52). FIG. 15 illustrates eye fundus images collected from Hq and control mice at various ages. White striations of fiber bundles radiating from the optic nerve disc were clearly visible for different areas of the eye fundus (N: nasal, T: temporal, I: inferior, S: superior) in 6 week-old Hq mice and age-matched controls. Axon tracks in control mice did not change overtime (FIG. 15, upper panel); while a diffuse loss of nerve fiber bundles was first noticed in some eyes from Hq mice 3-month old (FIG. 15, middle and bottom panels). Despite, that some variability was noticed between Hq mice of the same age and even between eyes from the same mice, generally a substantial loss of fiber bundles was evidenced in the majority of Hq mice aged 4 months (FIG. 15). Thus, loss of RGC soma evidenced from the age of 3 months in Hq retinas (Bouaita A, et al., Brain. 2012; 135: 35-52) is concomitant with the first structural changes in the RNFL and optic nerves. The disappearance of intraocular RGC axons became sufficient, at this age, to allow their highlighting by fundus imaging.

    3.2.2 Correlation between Visual Function Abnormalities and Optic Nerve Degeneration

    [0231] To establish when the continuing process of RGC loss compromises visual function, optokinetic tracking (OKT) thresholds were estimated using the OptoMotry system in Hq mice aged 2, 3 and 4 months (Douglas R M et al., Vis Neurosci. 2005; 22:677-684). The estimation of OKT thresholds (highest spatial frequency that each eye could track) enables the screening of functional vision for right and left eyes independently. Data gathered from 18 Hq mice per group of age was compared to our previously reported data for 6 week-old Hq and control mice as well as Hq and control mice aged between 6 to 8 months. OKT responses (cycles per degree) did not change in control mice with age and very little difference was evidenced between right and left eyes; thus scores for each eye were assembled in a unique group (FIG. 16). Conversely, visual performance of Hq mice declines with age, the 3 month-old group displayed a reduction of 14% (P=0.0031) or of 22% (P=<0.0001) relative to young Hq or control mice respectively. Visual function impairment was further confirmed in the 4 month-old group (FIG. 16) which exhibited a 25% and 30% reduction when compared with young Hq and control mice (P=<0.0001). Functional impairment worsened with age, indeed responses from 8 month-old mice were reduced 56% or 60% when compared to young Hq or age-matched control mice (FIG. 16). As previously described for some phenotypic abnormalities in Hq mouse strain, the onset of visual function impairment varies across individuals. The interindividual phenotypic heterogeneity in the Hq animals regarding visual capability consists of two type of animals: those which showed visual function impairment one month earlier while others displayed a better preserved vision for additional 3-4 weeks. Moreover, the variability was also noticed in some mice for which responses recorded from each eye were different; although, all the mutant animals eventually become severely impaired for responding to the visual stimuli. Hence, it appears that RGC degeneration in Hq mice starts early in life and that 3-4 month-old animals exhibit morphological changes leading to deleterious effects on visual function.

    3.2.3 Effect of Gene Therapy on Retinal Ganglion Cell Integrity

    [0232] AAV2/2-NGB or AAV2/2-AIF administration was performed in mice in which RGC degeneration process and its deleterious effect on visual function already took place. Two criteria were retained before proceeding to vector administration: (1) optic fibers bundles disappearance was noticeable by eye fundus imaging; (2) compromised visual acuity was assessed by OptoMotry (reduction of about 20% in the OKT thresholds). Generally, mice were 4 month-old when subjected to the treatment, whereas in our previous studies they were treated at the age of 4-6 weeks before the onset of optic nerve degeneration (Lechauve et al., Mol Ther. 2014; 22: 1096-1109). We estimated the yield of RGC transduction, six months post-injection, by subjecting retinal sections to immunohistochemistry for NGB and by measuring the relative abundance of NGB mRNA using RT-qPCRs (FIG. 17). Up to date, analyses were performed for 7 and 6 couples of retinas from eyes transduced with AAV2/2-NGB and AAV2/2-AIF respectively. FIG. 17 (left panel) shows the representative staining for NGB of one couple of retinas: the intensity of the fluorescence in the GCL revealed with the antibody against NGB was more prominent in the retinal section from treated eye relative to the untreated eye; besides the pattern of cellular distribution is reminiscent of what has been already shown for mouse retinas (Wei X et al., Am J Pathol. 2011; 179:2788-2797): intense labelling in the inner nuclear layer (INL), the GCL and inner segments (IS) of photoreceptors. Within the GCL of treated retinas, it was noticed many positive and strongly stained RGC somas and axons (FIG. 17, white arrowheads bottom panel). Next, we determined the abundance of the transduced NGB mRNA by real-time quantitative PCR using RNAs prepared from: (1) retinas from 3 Hq mice subjected to AAV2/2-NGB administration, (2) retinas from 6 untreated Hq mice (FIG. 17, right panel). The steady-state levels of the transduced NGB mRNA contributed to a 5.3-fold increase of the total amount of NGB mRNA relative to the one measured in untreated retinas while no change was evidenced in the relative steady-state levels of BRN3A or SNCG (α-synuclein) mRNAs between retinas from treated and untreated eyes (P=0.99 and 0.79 respectively; FIG. 17, right panel). Hence, efficient vector transduction was substantiated by the enhanced NGB labeling in the GCL and the increased abundance of the corresponding transcript in treated retinas.

    [0233] To confirm that gene therapy performed in 4 month-old Hq mice was performed after the initiation of neuron loss in the GCL, we estimated RGC soma number in whole retinal sections from Hq treated and untreated eyes subjected to immunostaining for BRN3A (FIG. 18, left panel). We compared RGC numbers (revealed by BRN3A labeling) and the total number of cells in the GCL (revealed by DAPI staining) in retinas from 7 or 6 Hq mice subjected to intravitreal injection of AAV2/2-NGB or AAV2/2-AIF1 and their untreated counterparts. Values obtained were compared to age-matched controls and Hq mice in which both eyes remained untreated (FIG. 18, right panel). Vector administration did not change the fate of RGCs which were undergone degeneration before the treatment since the number of RGCs was significantly reduced in retinas from Hq indifferently which eye received the vector (132±5 or 131±5 versus 260±5 in control retinas; P<0.0001). The 50% diminution of the total amount of BRN3A-positive cells in the GCL of these samples was in accordance with our previous studies (Bouaita A et al. Brain. 2012; 135:35-52 and Lechauve et al., Mol Ther. 2014; 22: 1096-1109). No difference between treated and untreated eye was noticed (P=0.67; Wilcoxon matched-pairs signed rank test). The reduced number of RGCs in the studied Hq groups corroborated the RT-qPCR data (FIG. 17, right panel); indeed the relative amounts BRN3A and SNCG mRNAs are considered as adult rodent RGC markers (Surgucheva I, et al. Mol Vis. 2008; 14:1540-1548; Bouaita A et al. Brain. 2012; 135:35-52); and they were comparable in all the Hq groups tested. Furthermore, the number of DAPI-stained nuclei in the GCL did not differ in the 13 couples of retinas evaluated. Thus, gene therapy performed in 4 month-old animals was unable to prevent RGC loss; the question arises whether change on the function of the remaining RGCs could be evidenced.

    3.2.4 Respiratory Chain Activity in Optic Nerves does not Correlate Axonal Loss in Treated Harlequin Mice

    [0234] To corroborate the loss of RGCs in Hq mice aged 4 months subjected to gene therapy with NGB or AIF1, we evaluated transversal optic nerves (ONs) sections from animals injected with AAV2/2-NGB or AAV2/2-AIF1 and euthanized six months later. Immunohistochemistry for the heavy chain subunit of neurofilaments (NF200) to detect RGC axonal profiles was performed. We observed a recognizable reduction of immunopositive dots in both ONs from the animal in which one eye was treated with AAV2/2-NGB relative to one ON isolated from an age-matched control (FIG. 19, left panel). Therefore, axon disappearance in this mouse is directly linked to RGC number reduction substantiating that successful gene therapy aimed at preventing RGC loss is only possible in young animals (4-6 week-old) since at this age the degenerative process involving RGC somas and axons was just initiated.

    [0235] In an attempt to establish whether NGB or AIF1 overexpression in RGCs could be beneficial for respiratory chain activity in the residual RGC axons, we sequentially measured rotenone-sensitive NADH decylubiquinone reductase, Complex I (CI) and the oligomycin-sensitive ATP hydrolase, Complex V (CV) activities in single ONs by spectrophotometry (Benit P et al., PLoS One. 2008; 3:e3208; Bouaita A et al. Brain. 2012; 135:35-52; Lechauve et al., Mol Ther. 2014; 22:1096-1109). FIG. 19 illustrates CV and CI activity measurements for ONs isolated from 15 Hq mice which were injected in one of their eyes with AAV2/2-NGB (8 mice) or AAV2/2-AIF1 (7 mice) and euthanized 6 months later; data obtained was compared with activities measured in 30 ONs from control mice aged 6-8 months and 30 ONs from Hq mice aged 8-10 months (which both eyes were untreated). As we have previously shown, we observed an approximate 30% enhancement of CV activities in ONs from Hq mice relative to controls. Nevertheless, the difference of CV activities in ONs from untreated and treated eyes was not significant (P=0.45, Wilcoxon matched-pairs signed rank test). In agreement with our previous studies AIF depletion in Hq mice aged 8-10 months is responsible for a 52% reduction of CI enzymatic activity in ONs relative to age-matched controls (FIG. 19). On the contrary, despite the reduction of RGC axons ONs from treated eyes exhibited 72% higher CI activity (12±0.45) than ONs from their contralateral untreated eyes, 6.96±0.3 (P<0.0001); the specific activity of CI attained in ONs from treated eyes 84% of the value measured in control mice (14.35±0.46). NGB or AIF overexpression driven by the recombinant AAV2/2 vectors, led to a very similar salutary effect on CI activity: for AIF, we measured 7.53±0.34 and 13±0.7 in ONs from untreated and treated eyes respectively; for NGB, we measured 6.5±0.35 and 11.2±0.5 in ONs from untreated and treated eyes respectively. These data suggest that either RGC axons, cells residing within the ONs or both displayed an enhanced CI activity in response to AAV2/2-AIF1 or AAV2/2-NGB treatment.

    3.2.5 Long-Lasting Protection of Visual Function in Harlequin Mice Treated with NGB or AIF despite RGC Loss

    [0236] Hq mice aged four months and subjected to AAV2/2-NGB or AAV2/2-AIF1 administration exhibited similar time course of RGC degeneration than untreated mice regarding the disappearance of RGC somas and axons. Unexpectedly, complex I activity measured in optic nerves from treated mice, six months after vector administration, was particularly high despite the severe reduction of nerve fiber numbers. Eye fundus imaging equally performed in treated mice before the treatment, 3 and 6 months post-injection established the choice of the eye to be treated as exhibiting a diminution of fiber density. Six months post-injection in treated eyes no further deterioration was noticed while in untreated eyes an obvious aggravation was observed since almost the entire temporal areas appeared as being devoid of fiber bundles (FIG. 20, upper and middle panels). To assess whether the robustness of ATP production in optic nerves could have a valuable effect on visual function; visuomotor behavior was quantified in treated mice 3 and 6 months post-injection and responses scored for treated and untreated eyes were compared. Overall, twenty-eight mice treated with either AAV2/2-NGB or AAV2/2-AIF1 were evaluated before vector administration, 3 and 6 months post-injection (FIG. 20, bottom panel). The comparison of OKT thresholds recorded in treated eyes between the time before injection (T=0) and 6 months later allows to draw the following conclusions: 5 eyes did show a significant increase in recorded thresholds, 15 eyes did not respond differently to the test overtime, 6 eyes exhibited a very mild decrease relative to scores recorded before the injection. Only 2 mice out of the 28 assessed displayed in treated eyes visual function deterioration which was similar to the one noticed in 8 month-old Hq untreated mice (FIGS. 16 and 20); however OKT thresholds measured on their counterpart untreated eyes were very low relative to the one measured at the time of vector administration. The mean of OKT thresholds for the 28 treated eyes 6 months post-injection was reduced by only 14% relative to the one measured before AAV2/2 administration (P=0.013) while the mean of OKT thresholds in untreated eyes was reduced by 46% (P<0.0001). Notably, AIF and NGB share the same ability to preserve visual function in treated eyes confirming the functional overlapping between the two proteins in vivo. In conclusion, it emerges from these data that gene therapy conferred functional/structural changes of remaining RGCs that combined with the improved energetic machinery found in ONs ultimately impeded efficiently visual function deterioration.

    [0237] In conclusion, subjecting 4 month-old Hq mice to AAV2/2-AIF or AAV2/2-NGB intravitreal administration permitted the nearly complete and long-lasting protection against vision loss in spite of the severe reduction of nerve fibers in the optic nerves. The unexpected high visual performance of these mice, after gene therapy, could be due to a functional recovery suggesting that RGC properties were changed, via the enhanced activity of complex I in their axons.

    [0238] This demonstrated the efficiency of the intravitreal AAV2/2-AIF or AAV2/2-NGB administration in Hq mice in which the RGC degeneration process and its deleterious effect on visual function already took place, in a curative approach.

    Example 4

    4.1 Material and Methods

    4.1.1 Animals

    [0239] The DBA/2J mouse strain (http://jaxmice.jax.org/strain/000671.html) and C57BL/6J mice were obtained from Charles River Laboratories (L′Arbresle, France). A colony of DBA/2J mice was established from breeders purchased and routinely backcrossed onto new founders from Charles River to reduce genetic drift in the colony. For strain matched-controls we used 2 month-old DBA/2J mice, an age before the onset of glaucoma. C57BL/6J mice were also used as additional controls, although they have a different genetic background than the DBA/2J mouse strain, since their assessment gave insight into how animals without optic nerve degeneration behaved. Only males were the recipient of evaluations and gene therapy; they were compared exclusively to either the males of different ages from the DBA/2J colony or C57BL/6J males. For morphological and functional characterization of optic atrophy a total of 150 DBA/2J mice at 2, 6, 8, 10, and 15 months of age (30 mice per age group) and 75 C57BL/6J mice at 2, 6, 8, 10, and 15 months of age (15 mice per age group) were used in our experiments. Gene therapy experiments involved about 60 DBA/2J mice. The mice were housed from one to four per cage in a temperature-controlled environment, 12-hour light/dark cycle and free access to food and water in a pathogenic-free barrier facility. Studies were conducted in accordance with the statements on the care and use of animals in research of the guidelines issued by the French Ministry of Agriculture and the Veterinarian Department of Paris (Permit number DF/DF_2010_PA1000298), the French Ministry of Research (Approval number 5575) and the ethics committees of the University Paris 6 and the INSERM, Institut National de la Sante et de la Recherche Médicale (Authorization number 75-1710).

    4.1.2 Tonometer Measurements of Intraocular Pressure

    [0240] For noninvasive intraocular pressure (IOP) measurement the Icare® TONOLAB tonometer (Icare, Espo Finland) was used. The assessment is based in a rebound method, which allows IOP to be calculated accurately, rapidly and without a local anaesthetic. The instrument takes five individual measurements and gives the mean as one reading displayed in mm Hg. We studied DBA/2J and C57BL/6J males aged between 1 and 15 months of age; measurements were performed monthly on the two eyes and collected during daylight.

    4.1.3 Slit-Lamp Examination and In Vivo Confocal Microscopy Analysis

    [0241] Slit-Lamp Examination and in vivo confocal microscopy analysis were performed as in example 3 (paragraph 3.1.3). In this study, for the first time, we used the IVCM to detect the iris because of the mice's tiny eyes. The depth of IVCM analysis for iris varied from 80 μm to 160 μm.

    4.1.4 In Vivo Electrophysiology

    [0242] Photopic electroretinogram (ERG) and flash visual evoked potential (VEP) responses were recorded simultaneously from electrodes placed on the cornea and overlying the visual cortex respectively. Photopic ERGs were recorded using two gold loop electrodes with light stimuli (10 cds/m2) applied on a light background (25 cd/m2) as previously described (Jammoul F et al., Ann Neurol. 2009; 65:98-107). Seven days before the recording, deep anesthesia was induced and maintained with sevoflurane (2.5-3%, Sevorane™, Abbott S.p.a. Campoverde, Italy) through a face mask and two stainless steel screws (0.9 mm diameter, length 2.4 mm) were placed using X and Y stereotaxic coordinates into the right and left primary visual cortices from mice; the whole implants were fixed with chirurgical glue. These VEP active electrodes were positioned 2.7 mm posterior to bregma and 2.5 mm lateral to the lambda suture (right and left) and penetrated the cortex to approximately 1 mm. Platinum needles in the forehead and at the base of the tail served as reference and ground electrodes respectively. On the day of recording, the mouse was anesthetized with an intraperitoneal injection of a mixture of ketamine (80 mg/kg) and xylazine (8 mg/kg). Signals were differentially amplified and digitized at a rate of 5 kHz (VEP bandpass filtered 0-100 Hz, ERG 0-300 Hz) using an Espion E2 system (Diagnosys LLC, Cambridge, UK). The amplitude and timing of the major ERG and VEP components was measured with the Espion software (Diagnosys LLC, Cambridge, UK) by placing a cursor at a subjectively determined turning point (i.e. the peak or trough) for each component in individual records (without knowledge of the animal's genotype). The Espion E2 system also generated and controlled the light stimulus. Brief (4 ms) single flash stimuli were delivered in a Ganzfeld dome. All recordings were made in a custom-made, light-tight Faraday cage. VEP responses were elicited by 100 flashes (10 cd/mm2) of white light of 10 μs duration and 1 Hz frequency delivered with the flash photostimulator (intensity 126-231 mJ or 1.0 cds sec/m2) placed 15-20 cm from each eye with band-pass filter 10-80 Hz. The mouse VEP is dominated by a negative polarity component that peaked 50-80 ms following stimulus presentation which is referred to as N1. The implicit time of the N1 component was measured at the negative peak. The amplitude of the VEP was measured from the N1 negative peak to the ensuing positive peak (P1). The responses were averaged per result and the amplitude with respect to baseline and latency from stimulus onset of the two main components of flash-VEPs (N1 and P1) were calculated.

    4.1.5 Adeno-Associated Viral Vector and Intravitreal Injections

    [0243] The vector named AAV2/2-NGB (SEQ ID NO: 9) (see example 2.1.2) was used in this study. For intravitreal injections DBA/2J mice were subjected to anesthesia with isoflurane (40 mg/kg body weight). The tip of a 33-gauge needle, mounted on a 10 μl Hamilton syringe (Hamilton Bonaduz AG, Bonaduz, Switzerland) was advanced through the sclera and 2 μL of vector suspension (2×10.sup.9 VG) was injected intravitreally, avoiding retinal structure disruption, bleeding or lens injury. Fifty-six DBA/2J mice 6-8 week-old were subjected to AAV2/2-NGB intravitreal injection during the course of this study; mice were euthanized between 8 to 10 months after vector administration.

    4.1.6 Retinal and Optic Nerve Histology

    [0244] Retinas and optic nerves (ONs) were carefully collected and fixed in 4% PFA at 4° C., cryoprotected by overnight incubation in PBS containing 30% sucrose at 4° C. Retinas were embedded in OCT (Neg 50; Richard-Allan Scientific), frozen in liquid nitrogen. ONs were embedded in a solution of PBS+7.5% Type A gelatin from porcine skin (Sigma-Aldrich) and 10% sucrose and frozen in a 2-methyl-butane solution at −45° C. Sections of retinas and ONs were cut (10 μm thickness) on a cryostat (Microm HM560, Thermo Scientific) at −20° C. and mounted on SuperFrost®Plus slides.

    [0245] For immunochemistry, sections of retinas and ONs were permeabilized with 0.1% Triton X-100 in PBS for 15 minutes at room temperature and treated with 3% BSA, 0.1% Triton and 0.05% Tween 20 in PBS for 1 hour. They were then incubated with primary antibody overnight at 4° C. The next day, sections were washed three times in PBS and incubated with appropriate secondary antibodies and 2 μg/mL of 4′, 6-diamidino-2-phenylindole (DAPI) for 2 hours at room temperature with 3% BSA, 0.1% Triton and 0.05% Tween 20 in PBS. Finally, they were washed 3 times with PBS, rinsed with sterile water and mounted on glass slides. Primary and secondary antibodies used are shown in Supplementary information (Table 8).

    4.1.7 Microscopic Observations

    [0246] Fluorescence labeling was monitored in the Cellular Imaging Facility of the Institute with: (i) a confocal laser scanning microscope (Olympus FV1000), image acquisition was conducted with Olympus Fluoview® software version 3.1. (ii) Retinal sections were also scanned with the Hamamatsu Nanozoomer Digital Pathology (NDP) 2.0 HT, its Fluorescence Unit option (L11600-05) and the NanoZoomer's 3-CCD TDI camera (Hamamatsu Photonics). BRN3A-positive cells, as the estimation of overall RGCs, were assessed for each animal by manually counting 3-5 entire retinal sections as described earlier (Lechauve C et al., Biochim Biophys Acta. 2012; 1823:2261-2273; Bouaita A, et al., Brain. 2012; 135:35-52).

    4.1.8 RNA Extraction and RT-qPCR Assay

    [0247] Total RNA from mice retinas were extracted using RNeasy Plus Mini kit (Qiagen). To ensure the absence of DNA a treatment with RNase-free DNase (Qiagen) and a subsequent cleanup with the RNeasy MinElute cleanup kit (Qiagen) were performed. Absence of DNA was confirmed by subjecting 10 ng of each RNA preparation to qPCR with specific primers for the NGB transgene and the mitochondrial ATP6 gene. One microgram of total RNA was reverse transcribed with oligo-dT using Superscript® II Reverse Transcriptase (Life Technologies). Quantitative PCR reactions were performed using ABI 7500 Fast (Applied Biosystems) and the specific primers listed on Table 9. The equivalent of 10 ng and 2 ng of cDNAs were used per gene as template for qPCR reactions with Power Sybr® green PCR Master Mix (Applied Biosystems). Each biological sample was subjected to the assay in triplicates per gene; Ct values were obtained with the ABI 7500 software (v.2.0.6). Messenger RNA steady-state levels of the mitochondrial ATP6 gene was the most stable in the 48 independent samples evaluated regardless the eye treatment. Therefore, to determine the relative mRNA amount of each studied gene we used the comparative ΔΔCt method and ATP6 as normalizing gene.

    4.1.9 Western Blotting Analysis

    [0248] Single retinas were homogenized in 50 μL of 20 mM HEPES and 60 mM mannitol (pH 7.2) using a 200 μL micro-hand-driven glass-glass potter at 4° C. Large cellular debris was spun down by a low speed centrifugation (1000 g for 5 minutes at 4° C.) and supernatants were subjected to protein quantification (Bradford reagent from Sigma-Aldrich) before proceeding to Western blotting. After denaturation at 94° C. for 15 minutes, samples were resolved in 15% SDS-PAGE and next transferred to a PVDF membrane. Membranes were probed with antibodies against NGB, AIF, β-actin, NDUFA9, SOD2, GFAP and ATP synthase subunit β (cf. Table 8). Immunoreactive bands were visualized with appropriate secondary antibodies coupled to horseradish peroxidase (0.1 mg/mL) (cf. Table 8) followed by detection with Pierce® ECL Plus Western Blotting Substrate (Pierce, Thermo Scientific). Theoretical molecular mass of each protein was estimated by comparing the electrophoretic properties of each specific signal in the immunoblots with the “PageRuler Plus Prestained Protein Ladder” (Pierce Protein Biology products, ThermoScientific). Signals obtained from different immunoblots were scanned and quantified with the Quantity One Analysis Software (Bio-Rad) to estimate the relative levels of mitochondrial proteins after normalization against β-actin signals. We operated within the linear dynamic range of our detection method and corrected the intra-blot and inter-blot variability by loading two quantities of the protein extracts within the same gels/blots and having in each independent experiment common samples.

    4.1.10 Tissue Homogenate Preparation and Respiratory Chain Enzyme Assays

    [0249] Optic nerves or retinas were rapidly collected and kept frozen (−80° C.). When the tests were performed tissues were prepared at 4° C. by homogenization with a 1 mL hand-driven glass-glass potter in 100 μL of extraction buffer (0.25 mM sucrose, 40 mM KCl, 2 mM EGTA, 1 mg/mL BSA, and 20 mM Tris-HCl, pH 7.2). Large cellular debris were spun down by a low speed centrifugation (1000 g×8 min) and supernatants were used immediately. Respiratory chain enzymatic activities were measured using a Cary®50 UV-Vis spectrophotometer (Agilent technologies), as described for ONs from mice (Bouaita A, et al., Brain. 2012; 135:35-52). The following measurements were performed in two independent assays: (1) rotenone-sensitive complex I or NADH decylubiquinone reductase activity and the ATP hydrolase activity of complex V which is oligomycin-sensitive; the maximal complex V activity and maximal oligomycin effect are only measured after a minute. Each assay was made in triplicate with 20 μL of each homogenate; (2) During the first phase of this assay, cytochrome c oxidase (complex IV) activity is measured by adding reduced cytochrome c and recording the rate of oxidation. The second phase of the assay for succinate cytochrome c reductase (complex II+III) activity is initiated by adding succinate, which triggers reduction of cytochrome c. The addition of the SDH competitive inhibitor, malonate, fully inhibits the SDH-dependent activity. Subsequent addition of G3P initiates the reduction of cytochrome c by G3P dehydrogenase (G3Pdh), thereby providing a measure of G3Pdh+complex III activity. Finally, during the third phase of this assay, after chelation of any metals by the addition of EDTA, decylubiquinol is added to initiate the reduction of cytochrome c by complex III. To discriminate enzymatic reduction from any chemical reduction of cytochrome c antimycin, a specific inhibitor of complex III, is then added. Each assay was made in duplicate with 20 μL of each homogenate.

    [0250] Values were converted to specific activities after protein quantification by the Bradford method. Complex I and complex V activities were expressed as nanomoles of oxidized NADH/min/mg protein; antimycin-sensitive complex III activity was expressed as nanomoles of oxidized decylubiquinone/min/mg protein; Complex IV was expressed as nanomoles of oxidized cytochrome c/min/mg/protein.

    [0251] All chemicals were of the highest grade from Sigma-Aldrich.

    4.1.11 Statistical Analyses

    [0252] Values are expressed as means±SEM (Standard Error of the Mean). Statistical analyses were performed with the GraphPad Prism 6.0 software assuming a confidence interval of 95%. Generally, the observations within each group do not fit in a normal distribution, thus non-parametric methods have been applied for evaluating the significance. Data collected from control and DBA/2J were compared using the unpaired non parametric significance test of Mann-Whitney (*≤0.05, **≤0.01 and ***≤0.005). Data collected from DBA/2J eyes subjected to gene therapy and their untreated contralateral eyes were compared using the paired non parametric significance test of Wilcoxon (*≤0.05, **≤0.01 and ***≤0.005).

    4.2 Results

    [0253] 4.2.1 Anterior segment Eye Pathology and Intraocular Pressure Elevation in DBA/2J during Glaucoma Progression

    [0254] DBA/2J mice develop glaucoma subsequent to anterior segment changes including pigment dispersion and iris stromal atrophy (John S W et al., Invest Ophthalmol Vis Sci. 1998; 39:951-962). We evaluated the apparition of the iris disease using slit-lamp examination and in vivo confocal microscopy.

    [0255] Dispersed pigments and iris atrophy reflected by iris pigment loss were first noticed in some animals aged 5-6 months, subsequently severe iris atrophy is noticed in all the assessed mice up to 8 months old. As the disease progressed in 12-15 month-old mice, corneal neovascularization, iris atrophy, and pupil posterior adhesion prevented the fundus examination. C57BL/6J mice did not exhibit microscopic clinical changes at any of the evaluated ages.

    [0256] To better define iris pathology and changes in the cornea from superficial epithelium to the endothelium, mice between 2 to 12 months were subjected to in vivo confocal microscopy. Four month-old mice generally exhibited normal aspect of superficial epithelium and stroma; conversely, in the endothelium scattered hyperreflective patterns were observed. Additionally, at four months of age the iris presented numerous filamentous and hyperreflective aggregates. The degenerative process in both the cornea and the iris aggravated in 8 month-old mice: activated keratocytes with stellar shape were observed in the stroma, the density of hyperreflective patterns and pigments increased in the endothelial layers, the visualization of the iris was difficult due to the presence of numerous circular hyperreflective partners (anterior part) resembling to pigment networks with mixture of inflammatory cells. One year-old DBA/2J mice presented cornea epitheliopathy with numerous dark microcysts and the absence of epithelial cells with normal morphology. Additionally, the other corneal layers exhibited pathological changes: the basal epithelium layer was abnormal with many dense and hyperreflective polyhedral areas; the stroma presented also pathological aspects with some holes and nearby fibrotic reaction: an increased number of hyperreflective pigments were observed in the endothelial layer when compared to 8-month-old mice. Conversely, the iris of one-year old mice exhibited lower amount of pigments relative to the 8-month-old mice.

    [0257] To correlate the iris pigment dispersion syndrome with ocular hypertension, we measured IOP in DBA/2J mice aged between 2 to 15 months and compared to C57BL/6J mice aged between 2 to 12 months (FIG. 21). The IOPs of C57BL/6J mice are constant overtime; consistent with other studies, the IOP of DBA/2J mice increased from the age of 7 months, prior to the onset of iris disease. The difference was statistically significant in groups of DBA/2J mice aged 8, 9, 10 and 11 months relative to the 2-3 month-old group (P<0.0001). However, ocular hypertension reached a peak in mice aged 11 months and declined progressively in the 12-month old group reaching the baseline in the 13-15 month-old group of mice (P=0.28).

    4.2.2 Evaluation of Retinal Ganglion Cell Loss and Gliosis in Aging DBA/2J Mice

    [0258] To determine the onset of RGC loss in DBA/2J mice we evaluated the relative abundance of BRN3A and SNCG mRNAs in retinas from mice aged between 2 to 15 months (FIG. 22A). BRN3A and SNCG genes are specifically expressed in rodent RGCs; the down regulation of their expression is considered as a marker of RGC injury (Soto I et al., J Neurosci. 2008; 28:548-561). Semi-quantitative RT-qPCRs allowed the determination of BRN3A and SNCG mRNA abundance relative to their amounts in retinas from C57BL/6J mice aged 2 months. Despite, the variable phenotype of DBA/2J mice a significant decrease in both transcripts was noticed in animals aged 10, 12 and 15 months; conversely BRN3A and SNCG mRNA relative amounts remained almost unchanged in retinas from all assessed C57BL/6J groups (FIG. 22A). Next, we evaluated the number of RGCs in retinal sections from DBA/2J mice between 2-15 months of age; additionally we compared with RGC number estimated in C57BL/6J at various ages. RGC loss is observed in about 20% of 8 month-old animals. Almost all the mice evaluated at the age of 10 to15 months possessed less than 65% of RGC number compared with 2 month-old DBA/2J mice or age-matched C57BL/6J mice; as expected the total amount of nuclei in the ganglion cell layer (GCL) was also diminished in DBA/2J. No noticeable change was observed in the estimated amount of RGC or total cell in the GCL in retinas from aged-matched controls (FIG. 22B). Retinal sections from C57BL/6J and DBA/2J mice immunolabeled for BRN3A clearly exemplified the loss of RGC overtime in DBA/2J mice (FIG. 22C). Moreover, it appears that the 15-month-old DBA/2J mouse exhibited a decrease in the thickness of the inner plexiform layer (IPL) and a reduction of the number of cell bodies both in the GCL and the inner nuclear layer (INL), confirming the marked neuron cell loss overtime in DBA/2J mice as previously reported (Jakobs T C et al., J Cell Biol. 2005; 171: 313-325; Schlamp C L et al., BMC Neurosci. 2006; 7: 66). It has been also described that astrocytes in the nerve fiber layer (NFL) and Müller glial cells that span the retina undergo reactive gliosis over time in the DBA/2J mouse strain (Inman D M et al., Glia. 2007; 55: 942-953). Thus, we stained retinal sections for the glial fibrillary acidic protein (GFAP), which is a sensitive marker of glial activation (astrocytes and Müller cells). In C57BL/6J retinas immunofluorescence, was confined exclusively to the GCL, presumably corresponding to the end-feet of Müller cells and astrocytes which reside in this cell layer. Retinas isolated from 8 and 12-month-old DBA2/J mice displayed an intense GFAP-stained which extended across to the outer nuclear layer (ONL), therefore, indicating a constant augmentation in the protein amount in all the cellular compartments: endfeet, somas and apical processes (FIG. 22C). The evaluation of the steady-state levels of GFAP mRNA and protein by RT-qPCRs and Western blotting confirmed GFAP overexpression in 2-month-old DBA/2J mice compared to age-matched C57BL/6J mice: a 2.8-fold increase in the relative mRNA level was measured in 2-month old DBA/2J retinas (p=0.0013); at the protein level the difference between both mouse strains was almost 2-fold after normalization against β-actin (p=0.005). The upregulation of GFAP expression continued in DBA2/J mice 10 and 15 months old; indeed at the mRNA level the increase reached more than 6-fold relative to age-matched C57BL/6J mice (p=0.0004 and 0.0002 when the 10 and 12 month-old groups were compared). Additionally, DBA2/J retinas purified from mice aged 15 months accumulated 2- and 4-fold more protein than retinas from 2-month-old DBA/2J and aged-matched C57BL/6J respectively. Conversely, GFAP mRNA and protein abundances remained stable from 2 to 15 months in C57BL/6J mice (FIGS. 22D and E).

    4.2.3 Determination of Changes in Optic Nerves from DBA/2J Mice with Aging

    [0259] To further confirm RGC death, transversal optic nerves (ONs) sections from DBA/2J and C57BL/6J mice of different ages were subjected to immunohistochemistry for the heavy chain subunit of neurofilaments (NF200) to detect RGC axonal profiles (FIG. 23A). We observed a recognizable reduction of immunopositive dots in the ONs from DBA/2J mice 8 and 12 month-old relative to either an age-matched control or to a 2-month-old DBA/2J mouse indicating a severe axonopathy. Sections from ONs were also subjected to immunostaining for GFAP and IBA1 (a specific calcium binding adaptor protein) since it has been reported that, concomitantly with RGC soma degeneration, optic nerves from DBA/2J mice exhibited reactive gliosis (Inman D M et al., Glia. 2007; 55: 942-953) and microglial activation (Bosco A et al., J Comp Neurol. 2011; 519: 599-620) during the course of the disease. Antibody against GFAP revealed a stronger signal in ONs from 2 month-old DBA/2J mice relative to age matched controls or 8 month-old C57BL/6J mice (FIG. 23A) indicating the initiation of the pathological process in nerve fibers. The intensity of the labeling increased with age suggesting that astrocytes replaced the axon bundles which disappeared progressively during glaucoma development. Besides, a substantial IBA1 immunostaining was noticed in ONs from DBA/2J mice 8 and 12 month-old relative to young DBA/2J mice or 8 month-old C57BL/6J mice; in these latter few scattered Iba1-positive cells were observed and should correspond to resting microglia (FIG. 23B). Hence, IBA1 gene expression was upregulated suggesting the presence of activated microglia (macrophages) in ONs during the course of the glaucomatous disease. We next proceeded to the immunostaining for vimentin, a marker of astrocytes, in the ON; we also noticed a consistent increase of the immunofluorescence in ONs from DBA/2J mice aged from 2 to 12 months when compared with 8 month-old C57BL/6J mice (FIG. 23C). Noteworthy, DAPI staining revealed more nuclei in DBA/2J sections suggesting either cell proliferation, migration from other regions, or both (FIG. 23C). Therefore, coincident with axonal loss, there is a pronounced activation of astrocytes which might have profound impact on the viability of remaining axons.

    4.2.4 Respiratory Chain Activity in DBA/2J Retinas and Optic Nerves during Glaucoma Progression

    [0260] It is largely admitted that RGCs are highly sensitive to mitochondrial dysfunction compared to other neuronal populations. In glaucoma, as in inherited optic neuropathies, RGC death could be caused by mitochondrial failure combining oxidative stress and energy depletion (Yu DY et al., Prog Retin Eye Res. 2013; 36: 217-246). We evaluated respiratory chain function in DBA/2J retinas and optic nerves during the progression of the disease. The spectrophotometric method used for assessing enzymatic activities of respiratory chain complexes has been successfully applied to accurately detect isolated defects in small amounts of tissue homogenates (Befit P et al., Clin Chim Acta. 2006; 374:81-86; Brain. 2012 January; 135(Pt 1): 35-52). Two independent spectrophotometric assays were devised to sequentially measure in homogenates of single retinas or ONs the enzymatic activities of: (1) rotenone-sensitive complex I (CI) and oligomycin-sensitive complex V (CV); (2) complex IV (CIV), malonate-sensitive combined complex II+III (CII+CIII) and antimycin-sensitive complex III (CIII) (FIGS. 24A and B). Surprisingly, the evaluation of single retinas from DBA/2J mice aged between 2 to 12 months indicated a consistent decrease in all the activities measured except for complex V when compared to values assessed in 2-month old DBA/2J retinas (FIG. 24A). Complex I and III enzymatic activities were the most affected with a reduction of ˜50% relative to retinas from 2-month-old mice; the difference was significant accordingly to the non parametric significance test of Mann-Whitney (p<0.0001 for complex I and 0.0031 for complex III). Since RGC population in rodent retina accounts for only 1%, (Salinas-Navarro M et al., Vision Res. 2009; 49:115-126) this data suggest that other neurons in the tissue exhibit a compromised energetic metabolism which could explain the reduction in their overall number such as reported for amacrine cells (Moon J I et al., Cell Tissue Res. 2005; 320:51-59) or the abnormal function/structure of photoreceptors observed in old DBA/2J mice (Heiduschka Pet al., Exp Eye Res. 2010; 91:779-783; Fuchs M et al. PLoS One. 2012; 7:e44645). The evaluation of complex enzymatic activities in ONs was performed in DBA/2J mice from age of 2 to 15 months. We observed a severe deficiency in all the complexes assessed except complex V which remained almost unchanged in all the animals tested: Complex I showed a 70% reduction in the group of mice, aged 12 or 15 months relative to the youngest ones; Complex III and IV activities decreased in the oldest animals almost 2-fold when compared with mice aged 2 months (FIG. 24B). All the differences measured in CI, CIII and CIV activities were significant (P<0.0001). The severe defect in energy supply could exacerbate the degenerative process in DBA/2J mice since it begins just before RGC loss and axonopathy onset. Remarkably, when ONs from 5 month-old mice were assessed for complex I activity a significant decrease relative to two month-old mice was observed: 8.8±0.83 (n=10) instead of 14.8±0.9 (n=24) nanomoles of oxidized NADH/min/mg protein (p=0.002, Mann Whitney test). Hence, bioenergetic failure in optic nerves from DBA/2J mice begins several months before RGC loss onset and could be directly involved in neuron cell degeneration as described in LHON(Yu-Wai-Man et al., Prog Retin Eye Res. 2011 March; 30(2):81-114).

    4.2.5 Mitochondrial Protein Amounts in Retinas from Glaucomatous Mice

    [0261] To determine whether respiratory chain defect could be associated to changes in the steady-state levels of mitochondrial proteins Western blotting analyses were performed with total protein extracts from retinas isolated from DBA/2J mice aged 2 or 12 months (FIG. 25A, left panel). We evaluated the relative abundance of NDUFA9 (a subunit of complex I), ATPase α (a subunit of complex V), Apoptosis Inducing Factor, AIF (a NADH oxidoreductase involved in complex I functional integrity) (Sevrioukova I F., Antioxid Redox Signal. 2011; 14:2545-2579) and Superoxide Dismutase 2, SOD2 (superoxide scavenging enzyme); signals obtained from each protein were normalized against the signal from b-actin. Experiments using 6-8 independent proteins extracts indicated that the abundance of the mitochondrial proteins evaluated decreased about 60% in 12 month-old DBA/2J mice relative to the amount measured in retinas isolated from DBA/2J mice aged 2 months (FIG. 25A, right panel). The difference in the amounts of each mitochondrial protein between young and old DBA/2J was significant: P=0.0286 for NDUFA9 and <0.0001 for all the other proteins evaluated.

    [0262] We assessed whether in the DBA/2J mice the development of the glaucomatous phenotype could be correlate with the downregluation of NGB expression (FIG. 25B). Radial cryosections of retinas were immunostained for NGB in DBA/2J mice aged 2 and 12 months. As previously described (Lechauve C et al., Mol Ther. 2014; 22:1096-1109), in 2 month-old mice the protein was particularly abundant in the ganglion cell layer (GCL), the inner segments (IS) of photoreceptors and the inner nuclear layer (INL) especially at its very inner margin, which could represent amacrine cells (Haverkamp S et al., J Comp Neurol. 2000; 424:1-23). Conversely, retinal sections from 15 month-old DBA/2J mice showed a significant diminution of the overall NGB immunostaining and also a noticeable reduction in the thickness of the neuron cell layers (FIG. 25B, upper panel). NGB mRNA steady-state levels in retinas from DBA/2J mice of various ages were determined by RT-qPCR. The abundance of NGB mRNA in retinas from 8, 10, 12 and 15 month-old mice relative to the amount calculated in retinas from 2 month-old DBA/2J was not significantly different (FIG. 25B, bottom panel, left). Therefore, we estimated the steady-state level of NGB protein by performing Western blotting analysis with identical protein amounts purified from single retinas of 2 and 12 month-old DBA/2J mice; six independent retinas were evaluated three times. Antibody against NGB recognized an intense band of 17 kDa in protein extracts from 2 month-old DBA/2J retinas, weaker signals corresponding to the 19 and 21 kDa NGB forms were also observed (FIG. 25B, bottom panel, right). In 12 month-old DBA/2J retinas, NGB signals were strongly reduced compared to 2 month-old DBA/2J retinas; after normalization against the β-actin signal a 57.4% reduction of NGB amount relative to 2 month-old DBA/2J retinas (FIG. 25B, bottom panel, right). The reduction of the NGB protein amount in retinas from aged DBA/2J mice relative to 2 month-old mice was comparable to the one measured for the other mitochondrial proteins assessed (FIG. 25A). Since RGC viability is strongly dependent on mitochondrial robustness, we attempt to prevent RGC and optic nerve degeneration in DBA/2J mice using gene therapy with NGB.

    4.2.6 Effect of AAV2/2-NGB Intravitreal Administration on Retinal Ganglion Cell Integrity

    [0263] A recombinant AAV2/2 encompassing the mouse NGB ORF associated with the full-length 5′ and 3′ UTRs of the gene was constructed to ensure mRNA stability and translation capacity (Chatterjee S et al Biol Cell. 2009; 101: 251-262). Subsequently, a single intravitreal injection with the AAV2/2-NGB vector (2×10.sup.9 VG per eye) was performed in DBA/2J mice 6-8 week-old. Overall, fifty six mice were euthanized 8 months after vector administration and extensively evaluated to establish the impact of NGB overexpression on RGC viability and respiratory chain integrity in optic nerves. First, we evaluated whether the administration of AAV2/2-NGB into the vitreous body of DBA/2J mice led to an increase in the amount of NGB in the GCL by subjecting retinal sections to immunohistochemistry for NGB. Two treated eyes and their untreated counterparts are illustrated in FIG. 26A; both the immunofluorescence signal was more intense and the number of positive cells was higher in the GCL from treated retinas relative to retinas from untreated eyes when the antibody against NGB was used. Hence, in the retinal section from AAV2/2-NGB treated eyes, NGB labeling was noticeably enhanced in the GCL substantiating vector transduction efficiency. Next, we estimated RGC number by subjecting retinal sections from seven couples of eyes in which one of them was subjected to AAV2/2-NGB administration while the contralateral ones remained untreated to immunostaining for BRN3A. The total number of cells in the GCL was determined by counting the nuclei staining with DAPI. As previously shown in retinas from 10 month-old DBA/2J mice (FIG. 22B) the number of RGCs were significantly reduced in untreated eyes: 97.24±24.6 versus 277.7±11 in retinas from 2 month-old DBA/2J mice; P<0.0001 (FIG. 26B); consequently, the number of DAPI-stained nuclei in the GCL was 35% lower than in retinas from 2 month-old DBA/2J mice (P<0.0001; FIG. 26B). When the untreated eyes from DBA/2J mice were compared with retinas from aged-matched controls (untreated DBA/2J mice) the number of nuclei was not different (P=0.4415); while the 7 untreated eyes exhibited enhanced RGC degeneration relative to aged-matched controls (101±15; p=0.0011). Remarkably, RGC number in AAV2/2-NGB treated eyes (206.6±18) was more than 2-fold higher than in contralateral untreated eyes (P=0.015; FIG. 26B) and attained 74.4% of the value measured in retinas from 2-month old DBA/2J mice. Besides, the overall quantity of cells in the GCL in treated retinas was increased by 32% relative to untreated eyes (P=0.015; FIG. 26B). To evaluate whether NGB overexpression could prevent the active growth of Muller cell processes observed during RGC degeneration (FIG. 22C), retinal sections from treated and untreated eyes were subjected to immunochemistry for BRN3A and GFAP. GFAP immunoreactivity was distinctly less increased relative to contralateral untreated eyes, resulting in the limited labelling of the end-feet of Muller cells; while the number of BRN3A-positive cells in the GCL was noticeable more important in treated eyes than untreated ones (FIG. 26C). Therefore, NGB overexpression hinders the upregulation of GFAP expression in Müller glia; this change may contribute to RGC survival in AAV2/2-NGB treated eyes.

    4.2.7 Neuroglobin Overexpression Protected Respiratory Chain Complex I or III Activity in DBA/2J Optic Nerves

    [0264] To establish a putative link between RGC number preservation and respiratory chain integrity we assessed respiratory chain complex activities in ONs from eleven DBA/2J mice in which one eye was subjected to AAV2/2-NGB administration. Values obtained for Complexes I, III and V were compared to those measured in ONs from either 8 or 10 month-old untreated DBA/2J mice illustrated in FIG. 24B (FIG. 27). Complex V activity did not change in any of the groups evaluated as shown in untreated DBA/2J mice at various ages as shown in FIG. 24B. Conversely, the specific activity of Complex I (10.12±0.6) attained in ONs from treated eyes 68.5% of the value measured in 2 month-old DBA/2J mice; further, ONs from NGB-treated eyes exhibited 86% higher complex I activity than ONs from their contralateral untreated eyes (P<0.0001; n=15). The sequential evaluation of Complex I (CI) and Complex V (CV) enzymatic activities allowed also Cl/CV determination; an accurate parameter for detecting an impairment of respiratory chain activity (Rustin Petal., Lancet. 1991; 338: 60). FIG. 27 shows the values obtained: a 73.5% increase was observed in ONs from treated eyes relative to their untreated counterparts (P=0.0005) which confirms the beneficial effect of AAV2/2-NGB administration for complex I activity. Next, we evaluated complex III activity in 13 mice in which one eye was subjected to AAV2/2-NGB treatment (FIG. 27). We observed an increase of 42% in the value obtained in optic nerves from treated eyes relative to their untreated counterparts (P=0.0005). The value measured in optic nerves from treated eyes reached 79.5% of the value measured in the optic nerves from 2-month old DBA/2J mice. Thus, high levels of NGB in RGCs from treated DBA/2J mice hampered complex I and III activity deficiencies presumably contributing to the improved RGC robustness evidenced (FIG. 26B).

    4.2.8 Preserved Retinal Ganglion Cell in AAV2/2-NGB Treated Mice Elicited Neuronal Activity in the Visual Cortex

    [0265] To assess whether the impediment of complex I and III defects, as a result of NGB overexpression, in DBA/2J mice could preserve the functional integrity of the visual pathway together with RGC enhanced robustness, Flash-Visual Evoked Potential (F-VEP) was recorded in treated DBA/2J mice. F-VEP monitors the communication from the RGC soma, through the axon, to the visual cortex; it has been reported, by F-VEP recording, that young DBA/2J mice produce robust and reproducible signals while by 10 to 24 months of age the signal severely diminishes (Heiduschka Petal., Exp Eye Res. 2010; 91:779-783; Sullivan T A et al., Hum Gene Ther. 2011; 22:1191-1200). Five groups of mice were evaluated: (1) C57BL/6J mice aged 2-3 months (n=10); (2) C57BL/6J mice aged 12 months (n=6); (3) DBA/2J mice aged 2-3 months (n=11); (4) DBA/2J mice aged 10-11 months (n=11); (5) DBA/2J mice aged 10-11 months and treated in one eye with AAV2/2-NGB at the age of 2 months (n=12). First, the electrophysiological activities of the retinas from these mice were assessed by light-adapted electroretinograms (ERGs). Recordings predominantly consisted of a fast, positive b-wave with minor oscillatory potentials and little or no a-wave (FIG. 28A, upper panel panel). It was observed a reduction in the amplitude of the b-wave in 12 month-old C57BL/6J mice relative to 2-3 month-old isogenic mice (FIG. 28A, bottom panel). The progressive degeneration of the iris and the cornea in three DBA/2J mice aged about 11 months and one 2 month-old DBA/2J mouse led to a significant reduction in the ERG recordings; they were excluded from the subsequent analyses. Hence, the results illustrated for the FVEPs corresponded to DBA/2J mice in which the rod and cone visual pathways remained largely unaffected (FIG. 28A, bottom panel). The most consistent components observed in F-VEPs at baseline were a negative N1 and a positive P1 peaks as shown in FIG. 28B, right panel. The means and standard errors of peak latencies, peak to peak amplitudes of F-VEP components in all experimental groups were analyzed by Mann Whitney test. We did not find significant differences in latencies and amplitudes between right and left eyes. Therefore the data from stimulation of both eyes were averaged. The latency of both the N1 and P1 peaks was unchanged in all the mice evaluated (data not shown). Conversely, quantification of the F-VEP components showed a significant reduction in the amplitude of the N1 and P1 waves in 12-month old C57BL/6J mice relative to their younger counterparts (P=0.0004 and <0.0001 respectively) which could be associated to the diminution of the b-wave component of the photopic ERGs (FIG. 28B, left panel). Moreover, there was a significant attenuation of the P1-wave amplitude in one year-old C57BL/6J mice relative to 2-3 month-old isogenic mice (P<0.0001). In all the DBA/2J mice evaluated independently of their age or treatment average amplitude of the P-wave was reduced relative to young C57BL/6J mice. However, their values were slighter higher than the ones found in the one year-old C57BL/6J mouse group (FIG. 28B, right panel).

    [0266] In our hands, 2-3 months old DBA/2J mice exhibited an approximate 2.8-fold diminution in the amplitude of the N1-wave when compared to young C57BL/6J mice (P<0.0001). The decline of the N1-wave amplitude was more pronounced (3.64-fold) in the older DBA/2J untreated group relative to young C57BL/6J mice (P<0.0001); while the 10-11 month old mice subjected to AAV2/2-NGB injection exhibited responses 43% higher than their younger counterparts; values attained 53.2% of the ones recorded in young C57BL/6J mice (FIG. 28B, right panel). The average amplitude of the N1-wave was 93% higher in treated DBA/2J mice relative to their age-matched untreated counterparts; the difference was significant (P=0.0003). Thus, since VEPs measure the cortical activity in response to flash stimuli, we can conclude that the functional alterations of DBA/2J mice leading to vision loss were partially prevented by NGB overexpression, substantiating the salutary effect of NGB for RGC viability and functional integrity.

    [0267] In conclusion, the increased NGB expression in young DBA/2J mice was able to: (1) slow-down the rate by which RGCs dies; (2) protect against optic nerve atrophy and (3) preserve the functional integrity of RGCs and the activity of the visual cortex.

    Example 5

    5.1 Material and Methods

    5.1.1 Animals and Diet

    [0268] The same animals as used in example 4 (see paragraph 4.1.1) have been used in this study i.e. DBA/2J mice.

    5.1.2 In Vivo Electrophysiology

    [0269] Photopic electroretinogram (ERG) and flash visual evoked potential (VEP) responses were recorded as in example 4 (paragraph 4.1.4).

    5.1.3 Retinal and Optic Nerve Histology

    [0270] Retinas and optic nerves (ONs) histology were performed as in example 4 (paragraph 4.1.6).

    5.2 Results

    5.2.1 Glaucoma Onset in DBA/2J Two Three Months Old Mice in the Anterior Segment Eye

    [0271] In our hands, 2-3 months old DBA/2J mice exhibited an approximate 2.8-fold diminution in the amplitude of the N1-wave when compared to young C57BL/6J mice (P<0.0001). Since VEPs measure the cortical activity in response to flash stimuli, we can conclude that DBA/2J mice 2-month old already exhibit functional alterations relative to C57BL/6J mice; while the amplitude of the b wave component of the ERG was unchanged in the two mouse groups evaluated (FIG. 29). Hence, the functional impairment of RGCs took place before the quantifiable RGC loss occurring 6-8 months later.

    5.2.2 Optic Nerves during Glaucoma Onset on DBA/2J Two-Three Months Old Mice

    [0272] The intensity of the immunolabeling for GFAP is higher in the optic nerve from the 2 month-old DBA/2J mouse relative to the optic nerve from the 2 month-old C57BL/6J mouse. Immunochemistry for the antibody against Vimentin revealed a very similar pattern of immunofluorescence relative to GFAP, confirming that astrocytes in ONs from DBA/2J mice aged 2 months exhibited an increase in number and reactivity relative to age-matched C57BL/6J mice (FIG. 30). Thus, it seems that astrocytes in optic nerves from 2 month-old mice could respond by their proliferation/reactivity to the beginning of RGC axon damage. These astrocyte changes may represent a response to early axonal abnormalities that occur prior to axon degeneration but which have functional consequences which are evidenced in these mice by F-VEPs.

    [0273] Putting together the data collected from DBA/2J mice, since F-VEPs measure the cortical activity in response to flash stimuli, we can conclude that the functional alterations of young DBA/2J mice leading to vision loss were efficiently impeded by subjecting to gene therapy with NGB, substantiating the salutary effect of the protein for RGC viability and functional integrity. Moreover, in regards of the data from 2 month-old DBA/2J, it appears that functional impairment of RGCs took place before the quantifiable RGC loss occurring 6-8 months later, thus, we can conclude that NGB treatment led to a very effective functional rescue of RGC dysfunction, in a curative way, via its essential role on the maintenance of respiratory chain complexes I and III activities within normal ranges.