Molecular Vision 2005; 11:542-553 <http://www.molvis.org/molvis/v11/a64/> Received 1 December 2004 | Accepted 5 July 2005 | Published 22 July 2005

Download Reprint

Predominant rod photoreceptor degeneration in Leber congenital amaurosis

Jacqueline van der Spuy,¹ Peter M. G. Munro,² Philip J. Luthert,¹ Markus N. Preising,³ Toke Bek,⁴ Steffen Heegaard,⁵ Michael E. Cheetham¹
 
 

¹Division of Pathology and ²Electron Microscopy Unit, Institute of Ophthalmology, University College London, London, United Kingdom; ³ Department of Pediatric Ophthalmology, Strabismology and Ophthalmogenetics, Klinikum, University of Regensburg, Regensburg, Germany; ⁴Department of Ophthalmology, Århus University Hospital, Århus, Denmark; ⁵Eye Pathology Institute, University of Copenhagen, Copenhagen, Denmark

Correspondence to: Michael E. Cheetham, Division of Pathology, Institute of Ophthalmology, University College London, 11-43 Bath Street, London, EC1V 9EL, United Kingdom; Phone and FAX number: +44 207 608 6944; email: michael.cheetham@ucl.ac.uk

Abstract

Purpose: An unusual retinal vascular morphology in an enucleated eye from a patient with Leber congenital amaurosis (LCA) has been associated with a mutation in AIPL1. The AIPL1 protein is expressed in the pineal gland and retinal photoreceptors. In the retina, AIPL1 is expressed in both developing cone and rod photoreceptors, but it is restricted to rod photoreceptors in the adult human retina. Therefore, this study was conducted to determine the photoreceptor phenotype in this LCA patient to determine if photoreceptors were differentially affected.

Methods: Additional genetic screening was performed and the consequences of the H82Y amino acid substitution characterized in an in vitro assay of NUB1 modulation. The morphology of the photoreceptors was examined by light and electron microscopy. Immunohistochemistry and immunofluorescent confocal microscopy was performed using a range of retinal photoreceptor markers.

Results: Transfection of the H82Y mutant AIPL1 in SK-N-SH cells revealed a normal subcellular localization and solubility but resulted in an increased ability of AIPL1 to redistribute GFP-NUB1 to the cytoplasm and resolve NUB1 fragment inclusion formation. Morphologically, the LCA retina appeared to be cone-dominant with a single layer of cone-like cells remaining in the central retina. Photoreceptor outer segments were absent and the surviving residual inner segments were severely shortened. Severe degeneration of the LCA retina was associated with upregulation of glial fibrillary acidic protein (GFAP). No signal was detected for AIPL1, rhodopsin, or L/M and S cone opsins in the LCA retina. Double labeling with peanut agglutinin (PNA) and wheat germ agglutinin (WGA) supported a cone-dominant phenotype for the surviving photoreceptors in the LCA retina, as did double labeling for cone arrestin, and rod and cone recoverin. The cone arrestin signal was restricted to the residual photoreceptor inner segments and was not detected in the cell bodies, axons, or axon terminals of the surviving photoreceptors. Recoverin immunoreactivity was most intense in the residual photoreceptor inner segments.

Conclusions: The phenotype in this patient suggests that although AIPL1 is required for the development of normal rod and cone photoreceptor function, it might only be essential for rod and not cone survival in the adult.

Introduction

Leber congenital amaurosis (LCA) is the most severe form of retinal dystrophy and is both genetically and clinically heterogeneous. Thus far, disease causing mutations have been reported in seven genes: RetGC1 [1], RPE65 [2], CRX [3], AIPL1 [4], LRAT [5], CRB1 [6], and RPGRIP1 [7]. These genes are also reported in the Retina International Scientific Newsletter Mutation Databases. Two other disease-associated loci have been identified on 14q24 [8] and 6q11-16 [9]. AIPL1 has been shown to interact with and modulate the nuclear translocation of NUB1, a protein that targets substrates implicated in the regulation of the cell cycle or cell growth for proteasomal degradation [10-12]. AIPL1 is also thought to function as a potential chaperone for phosophodiesterase (PDE), as PDE levels were decreased by a post-transcriptional mechanism in mouse models of LCA that both extinguished and reduced AIPL1 levels [13,14]. An H82Y (244C >T) mutation in exon 2 of the AIPL1 gene has been described in an LCA patient with an unusual retinal vascular morphology [15]. A gradual attenuation of the vascular system towards the retinal periphery was described culminating in a total avascular zone along the peripheral circumference.

We have described the expression of AIPL1 in the developing and adult human retina [16,17]. AIPL1 was first detected early in development in the presumptive cone photoreceptors of the central retina. With continued development, AIPL1 expression gradually spread from the central to peripheral retina closely following the spatiotemporal gradient of photoreceptor development and differentiation. AIPL1 was detected in both the presumptive rod and cone photoreceptors, and by the time of birth it was detected in all photoreceptors from central to peripheral retina [16]. These data in combination with the severity of the LCA phenotype suggested that AIPL1 is important for the normal functional development of rod and cone photoreceptors. In the adult human retina, however, the expression of AIPL1 was restricted to the rod photoreceptors with staining throughout the rod cells with the exception of the rod outer segments. No AIPL1 was detected in the cone photoreceptors [17]. This suggested that AIPL1 could also be important for the maintenance of rod photoreceptors but not cone photoreceptors in the adult retina and that there is a developmental switch in AIPL1 function.

Three independent groups have engineered mouse models of LCA with complete or partial inactivation of AIPL1 expression [13,14,18]. In the AIPL1 knock-out models of LCA, retinal lamination and morphology appear essentially normal during development and both cone and rod photoreceptors appear to develop normally [13,18]. Degeneration of the photoreceptors in the outer nuclear layer (ONL) proceeds rapidly after birth. An electroretinographic (ERG) response can not be evoked from AIPL1^-/- mice at any age with any illumination condition (13, 18). Similarly, normal morphological development of the photoreceptors is observed in an AIPL1 hypomorphic (h) mutant expressing significantly reduced levels of AIPL1 (14). The reduction of AIPL1 expression eventually results in photoreceptor degeneration with more than half of the photoreceptors lost by the age of 8 months, however, no evidence for a cone defect is observed by immunofluorescence for cone opsin up to 11 months of age. The photoresponse onset and recovery is delayed in the rod photoreceptors of the AIPL1 hypomorphic mutant. In two of the mouse models, a reduction in the levels of cGMP phosphodiesterase (PDE) was detected before the onset of degeneration, suggesting that AIPL1 is necessary for the biosynthesis, assembly, or stabilization of PDE [13,14].

Histological material from LCA patients is exceedingly rare. Histological material from a 33 week LCA retina with mutations in RPE65 revealed an absence of detectable RPE65 staining and prenatal ocular degeneration [19]. Cell loss and thinning of the outer nuclear (photoreceptor) layer was evident, the outer segments of the fetal rod photoreceptors were stunted and had decreased immunoreactivity for rhodopsin [19]. Labeling of the cone outer segments with peanut agglutinin was sparse and punctuate [19]. This is in contrast to RPE65^-/- mice and in naturally occurring RPE65^-/- Briard dogs, where almost normal retinal histology is observed at birth although no recordable photofunction can be detected [20,21]. A histopathological study of an LCA retina caused by mutant RetGC1 from an 11 and a half year old subject revealed rod and cone photoreceptors in the macular area and far periphery of the retina, but no photoreceptors in the midperipheral retina [22]. The rod and cone photoreceptors did not have outer segments. The cones formed a monolayer of cell bodies, but the rods were clustered and had sprouted neuritis in the periphery. At ages 9 and 10 years, there was no measurable visual field and visual acuity had declined to light perception [22]. An LCA patient who was heterozygous for a putative AIPL1 mutation (H82Y) was diagnosed at age five months with no light perception and had both eyes enucleated at 22 years of age because of pain [15]. Marked changes in the retinal vascular morphology were described [15]. However, AIPL1 is photoreceptor specific in the retina, and while it is expressed in both developing rods and cones in humans, it is present only in rod photoreceptors in adulthood [16,17]. In this LCA patient, the retinal photoreceptors were described as being almost totally absent with remnants of photoreceptor outer segments, but the nature of the surviving photoreceptors was not described further [15].

Therefore, the aim of this study was to examine the outer retina and, in particular, the photoreceptors of this putative AIPL1 LCA patient in greater detail using light microscopy, electron microscopy, immunofluorescent labeling, and scanning confocal microscopy. These results will be compared with our observations for AIPL1 expression in the normal developing and adult human retina in order to enhance understanding of the consequences of AIPL1 mutations.

Methods

Molecular genetics

Patient DNA was screened by SSCP analysis for mutations in AIPL1, RetGC1, RPE65, and CRX genes in an ongoing screen of patients with early onset severe retinal dystrophies [15]. PCR amplimers from SSCPs showing aberrant banding patterns were directly sequenced using the fluorescent dye chain termination method. Mutations were confirmed by restriction endonuclease digestion. Exon 2 amplimers of AIPL1 from 46 controls from a regional Bavarian control set were digested by BseNI to establish the population frequency of D90H. These 46 controls were also screened for the H82Y mutation by SSCP-analysis.

Immunocytochemistry and western blotting

The plasmids used in this assay have been described previously [11]. The pCMV-Tag3C-AIPL1 plasmid was used as the template for generation of the H82Y(244C >T) mutant using site specific primers and the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The pCMV-Tag3C-AIPL1(H82Y) plasmid was sequenced using the Big Dye Terminator (PerkinElmer Life Sciences, Boston, MA) ready reaction sequencing kit in order to confirm the absence of unintentional changes and the presence of the desired mutation. The subcellular distribution of AIPL1(H82Y) in SK-N-SH neuroblastoma cells was examined by immunofluorescent labeling as described previously [11]. The preparation and Western blotting of cell extracts from SK-N-SH cells expressing AIPL1 and AIPL1(H82Y) was performed as described previously [11].

NUB1 modulation assay

The modulation of GFP-NUB1, GFP-NUB1-N, and GFP-NUB1-C by AIPL1(H82Y) was assessed as described previously [11]. Briefly, eight well chamber slides were seeded with 2x10⁴ SK-N-SH neuroblastoma cells per well. 24 h later the cells were transiently transfected using LipofectAMINE Plus Reagent (Invitrogen Life Technologies, Paisley, UK) according to the manufacturer's instructions. For concentration dependent studies, increasing amounts of pCMV-Tag3C-AIPL1 or pCMV-Tag3C-AIPL1(H82Y) were co-transfected with 50 ng of pEGFP-C1-NUB1, pEGFP-C1-NUB1-N, or pEGFP-C2-NUB1-C at a ratio (w/w) of 0:1, 0.5:1, 1:1, and 2:1, and with pCMV-Tag3C stuffer plasmid to maintain an equal amount (150 ng) of total DNA in each transfection. Four fields of about 100 transfected cells each were counted blind using a Leica DM RBE fluorescent microscope in at least three separate experiments. In addition, fluorescence images were captured using a PerkinElmer Ultrapix standard peltier cooled CCD camera and LSR Ultra Plus version 4.1 Spatial Imaging Module. ImageJ (Version 1.32j, NIH, Bethesda, MD) was used to assess the optical density (OD) of GFP-NUB1 fluorescence in the nucleus of about 100 transfected cells as described previously. Mean and standard deviation were calculated for each data set.

Tissue processing

Normal human retina from an adult 65 years of age was fixed immediately post-enucleation with 10% (v/v) neutral-buffered formalin within 2 min of enucleation for at least 24 h and embedded in paraffin wax as described previously [12]. All samples were provided with informed consent, institutional review board approval was obtained, and the tenets of the Declaration of Helsinki were followed. The LCA affected tissue was prepared as described earlier [15]. The histology and immunohistochemistry were compared with normal retina.

Light and electron microscopy

The tissue was processed for light and electron microscopy as described previously [15]. Sections for light and electron microscopy were cut using a Leica ultracut S microtome and semithin (1 μm) sections were stained with 1% toluidine blue (w/v) (Agar Scientific Limited, Stansted, Essex, England) in 1% borax (w/v) (Agar Scientific Limited) and scanned in a Zeiss LSM510 confocal microscope operating in transmission mode. Ultrathin sections (70 nm) were stained sequentially for 5 min with saturated ethanolic uranyl acetate (Agar Scientific Limited) and lead citrate (Agar Scientific Limited), examined and photographed with Kodak 4489 EM film using a JEOL 1010 TEM operating at 80 kV.

Immunohistochemistry of paraffin embedded sections

Immunohistochemistry of paraffin embedded sections from control and LCA retina with rabbit polyclonal antiserum Ab-hAIPL1 (1:500), mouse hybridoma supernatant COS-1 (long/medium wavelength-sensitive (L/M) cone opsin, 1:100), and mouse ascites fluid OS-2 (short wavelength-sensitive (S) cone opsin, 1:10 000) was performed as described previously [16,17]. Immunohistochemistry with mouse monoclonal antibody 1D4 (rhodopsin, 1:200; National Cell Culture Center, Minneapolis, MN) and A233 (GFAP, 1:2000) was performed as described previously [16]. Antigen retrieval was accomplished by treating the sections with 0.1% trypsin (w/v) and 0.1% CaCl₂ (w/v) in TBS (Tris-buffered saline; 100 mM Tris-HCl pH 7.6, 150 mM NaCl) at 37 °C for 15 min. Immunohistochemistry with rabbit polyclonal antiserum LUMIf (1:10000, a generous gift of Dr. Cheryl Craft, Doheny Eye Institute, Los Angeles, CA) [23,24], goat polyclonal antibody Recoverin (N-16, 1:100, Santa Cruz Biotechnology, Santa Cruz, CA), biotinylated peanut agglutinin (bPNA, 1:200, Vector laboratories, Burlingame, CA), and biotinylated wheat germ agglutinin (bWGA, 1:200, Vector laboratories) was performed as described previously [16]. Antigen retrieval was accomplished by microwaving the sections four times for 2.5 min each at 800 W in TBS containing 5% urea (w/v). The sections were double labeled by incubation with LUMIf and Recoverin (N-16) overnight at 4 °C. For double labeling of sections with bPNA and bWGA, the sections were first incubated with bPNA overnight at 4 °C followed by Cy2-Streptavidin conjugate (1:100, Molecular probes, Eugene, OR) for 45 min at room temperature. The sections were then washed in TBS and incubated with a preformed bWGA Cy3-Streptavidin conjugate (ZyMAX Grade 1:100, Zymed laboratories, San Francisco, CA) complex for a further 45 min at room temperature. Secondary antibody incubations were performed with Cy2-conjugated donkey anti-rabbit (Molecular probes), Cy2-conjugated donkey anti-mouse (Molecular probes), and Cy3-conjugated donkey anti-goat (Molecular probes). Retinal sections were incubated with either the primary or secondary antibodies on their own to confirm that the signal observed did not arise nonspecifically from any of the antibodies per se. In addition, rabbit primary antibodies were incubated with mouse secondary antibodies and vice versa to verify that immunosignals did not arise from cross reaction between the antibodies in the double label procedure. The retinal sections were visualized with a Zeiss LSM510 laser scanning confocal microscope. Images were exported from LSM Image Browser as TIFF files. Brightness and contrast were adjusted for clarity prior to conversion from RGB to CMYK in Adobe Photoshop 6.0 (Adobe Systems, Inc., San Jose, CA). Image labeling was performed in Adobe Illustrator 9.0 (Adobe Systems).

Results

Molecular genetics

The LCA eye donor was part of an ongoing genetic screen in 198 patients of mixed ethnic origin with early onset severe retinal dystrophies. In this patient, two sequence variations in the amino acid sequence of AIPL1 were identified, H82Y (244C >T) and D90H (268C >G) [15]. In addition, several noncoding sequence variations were identified in AIPL1 (IVS2-10 A >C, IVS4-33 G >T) and RetGC1 (3418 G >C). The same AIPL1 genotype was also detected in the patient's affected sister. The H82Y and D90H alleles were inherited separately from each parent. The H82Y amino acid substitution was not observed in any other proband or control tested in this study. In contrast, D90H was detected in 10 of the 198 patients. Therefore, a set of 46 unaffected European controls were tested for the D90H change using a BseNI restriction enzyme digest to assess the frequency of this sequence change. The D90H change was detected as homozygous in two individuals and as a heterozygous change in 15 individuals of the 46, suggesting that this is a relatively common allele in the European population.

Analyses of the H82Y amino acid substitution

To test the potential pathogenicity of the H82Y amino acid substitution, the change was created by site-directed mutagenesis of pCMV-Tag3C-AIPL1. Transfection of this construct in SK-N-SH neuroblastoma cells showed that the H82Y mutant protein was indistinguishable from the wildtype protein on the basis of solubility and migration on SDS-PAGE gels (Figure 1A) or subcellular localization (Figure 1B). In contrast, investigation of the ability of the mutant protein to modulate the AIPL1 interacting partner NUB1 revealed that the mutant protein appeared to be hyperactive. The H82Y protein stimulated more full-length GFP-NUB1 to remain in the cytoplasm (24.3±4.3% nuclear) compared to the wildtype protein (42.1±10.3% nuclear; Figure 1C). Furthermore, the H82Y mutant was more efficient than the wildtype protein at suppressing inclusion formation of both the N-terminal (24.7±3.7% compared to 31.9±2.6%; Figure 1D) and C-terminal (0.3±0.2% compared to 3.3±0.2%; Figure 1E) fragments of GFP-NUB1.

Light and electron microscopy of LCA retina

Semithin sections (1 μm) from the peripheral area (Figure 2A), midperipheral area (Figure 2B), and macular area (Figure 2C,D) of the LCA retina were examined by light microscopy. Histological sections of the LCA retina stained with toluidine blue exhibited an abnormal morphology and stratification when compared to that of a normal control. The outer nuclear layer (ONL) was thinner in the peripheral (Figure 2A), midperipheral (Figure 2B), and macular areas (Figure 2C,D), with only a single layer of cells surviving in the ONL in the macular area of the LCA retina (Figure 2C,D). Extensive cell loss and condensed cellular profiles indicating cell death were evident in the ONL (Figure 2A-D). The separation of the ONL and inner nuclear layer (INL) was reduced due to attenuation of the outer plexiform layer (OPL; Figure 2A-D). In all areas, the photoreceptor outer segments were absent and the inner segments were abnormally shortened (Figure 2A-D). The retinal vasculature (V) gradually attenuated towards the periphery with an avascular zone in the far periphery (Figure 2A) and large, dilated vessels in the macular area (Figure 2C) [15]. Ultrathin sections of the ONL from the macular area were examined by transmission electron microscopy (Figure 2E). The ONL was comprised of a single layer of abnormal photoreceptors aligned along a clearly defined outer limiting membrane (OLM). Morphologically, these receptors appeared cone-like with shorter and wider cell bodies than expected for rod photoreceptors. No photoreceptor outer segments could be detected, and only residual and stunted photoreceptor inner segments (IS) that were highly disorganized appeared to remain. Müller cell processes (asterisks) extended between the photoreceptors to the OLM, and in areas appeared to project through the OLM.

AIPL1 and GFAP staining

Paraffin-embedded sections of the normal and LCA retina were treated with a panel of antibodies and lectins to retinal markers (Table 1). In the control retina, Ab-hAIPL1 labeled all rod photoreceptors with the exception of the rod outer segments as previously described [17]. In contrast, no immunosignal was detected with Ab-hAIPL1 in the LCA retina (Table 1). The control and LCA retinas were examined for reactive changes in Müller cells, a common feature in degenerating retina, by detection of glial fibrillary acidic protein (GFAP) expression (Figure 2 and Table 1). Severe gliosis of the LCA retina was evident as the expression of stress-inducible GFAP was upregulated (Figure 3). In the control retina, GFAP staining was most intense in astrocyte processes of the nerve fiber and ganglion cell layers (GCL) and up to the inner limiting membrane (Figure 3). In the LCA retina, the radially oriented processes of Müller cells were strongly positive for GFAP (Figure 3). In the LCA retina, attenuation of the OPL reduced the separation of the ONL and INL when compared to the control, and with fewer surviving cells in the ONL of the LCA retina (Figure 3).

Opsin staining

The antibody 1D4 was used to detect rhodopsin, a marker for rod photoreceptor outer segments, in the control retina as described previously [16,17] (Table 1). COS-1 and OS-2 were used to detect long/medium wavelength-sensitive (L/M) and short wavelength-sensitive (S) cone opsin in the outer segments of L/M and S cones, respectively, in the control retina as described previously [16] (Table 1). The immunosignals for rhodopsin, LM-cone opsin, and S-cone opsin were absent in the LCA retina (Table 1). Additional markers were detected in the LCA retina, but these were abnormal when compared to the normal retina (Table 1 and next two sections).

Lectin staining

Double fluorescent labeling of the control and LCA retina was performed with peanut agglutinin (PNA) and wheat germ agglutinin (WGA; Figure 4 and Table 1). The lectin PNA decorates the interphotoreceptor matrix (IPM) surrounding the inner and outer segments of the cone photoreceptors. The lectin WGA preferentially decorates the IPM surrounding rod photoreceptor inner and outer segments but also associates with the cone matrix. The major WGA-binding glycoprotein in the human IPM is a sialoprotein associated with cones and rods (SPACR) [25]. In the control retina, the lectins PNA and WGA labeled the IPM surrounding the inner and outer segments of the photoreceptors (Figure 4A). At higher magnification, PNA (green) bound to the extracellular matrix sheaths of the cone photoreceptor inner and outer segments while WGA (red) bound to the extramatrix sheaths of both the cone and rod photoreceptor inner and outer segments as previously shown [17] (Figure 4B,C). Colocalization of both lectins surrounding the cones was apparent from the yellow color (asterisks) in the cone photoreceptor matrix compartments (Figure 4D). The cone photoreceptor inner and outer segments (yellow) were clearly interspersed between the inner and outer segments of the rod photoreceptors (red). In the LCA retina, the PNA and WGA label was abnormal and labeled the residual matrix of the stunted inner segments of the photoreceptors (Figure 4E). At higher magnification, the PNA label (green) and WGA label (red) were abnormal (Figure 4F,G). Unlike the control retina, colocalization of the PNA and WGA label shown in yellow (asterisks) was not interspersed by WGA label (red), and all of the photoreceptor sheaths that were labeled with WGA (rod and cone IPM) were also positive for PNA (cone IPM, Figure 4H). In the absence of a strong signal from the rod sheaths the relatively weaker WGA signal from other extracellular matrices could also be detected in the INL.

Recoverin and cone arrestin staining

Double immunofluorescent labeling of the control and LCA retinas was performed with LUMIf to detect cone arrestin and Recoverin (N-16) to detect rod and cone photoreceptor recoverin (Figure 5 and Table 1). In the control retina, Recoverin (N-16, red) labeled all parts of the rod and cone photoreceptors and also stained bipolar cells with less intensity (Figure 5A,B). LUMIf (green) labeled the cone photoreceptors including the cone outer and inner segments, cell bodies, axons, and pedicles in the ONL as shown previously [23,24] (Figure 5A,C). As shown at higher magnification, the recoverin immunoreactivity co-localized (yellow) with the cone arrestin immunosignal only in the cone photoreceptors (Figure 5D, asterisks). In the LCA retina, the immunosignals for cone arrestin and recoverin were very different to the control (Figure 4A,E). The recoverin immunosignal (red) was very reduced compared to the control and was most intense in the residual inner segments of the photoreceptors (Figure 5E,F). At this level of detection the recoverin immunosignal could also be observed in the INL and plexiform layers of the LCA retina (Figure 5E,F) and control (not shown). The immunosignal for cone arrestin (green) was not detected in the cell bodies, axons, or synaptic terminals of the surviving photoreceptors, but was restricted to the residual photoreceptor inner segments (Figure 5E,G). At higher magnification, the surviving photoreceptor inner segments were labeled yellow with markers for both cone arrestin and recoverin (asterisks), with very few residual inner segments expressing only recoverin (Figure 5H).

Discussion

At the age of 22 years, both eyes of an LCA patient were enucleated leading to immediate relief of pain. The initial genetic diagnosis of this patient identified two amino acid substitutions in AIPL1 (H82Y and D90H) that were also inherited by an affected sibling [15]. In the present study, we have shown that the D90H sequence variant is a relatively common sequence variation and homozygosity in two controls argues strongly for D90H to be benign. In contrast, the H82Y variant was not identified in any controls. Furthermore, the H82Y substitution altered the ability of AIPL1 to modify NUB1 function. This sequence variant was hyperactive in modulating NUB1 cytoplasmic localization and suppressing NUB1 fragment inclusion formation. The functional consequences of AIPL1-mediated relocalization and "chaperoning" of NUB1 still remains to be fully determined, however, it has been reported that other AIPL1 pathogenic changes (e.g., G262S) are also more active than wildtype AIPL1 in stimulating the cytoplasmic distribution of NUB1 [11]. These data support the notion that the H82Y sequence variant is a pathogenic change in AIPL1.

The H82Y change was inherited from the patient's unaffected father, whereas the D90H sequence variation was the maternal allele [15]. For AIPL1 to be the causative gene in this patient, the D90H allele must be linked with another change in AIPL1, possibly in the promoter or intronic regions of the gene that renders this maternal allele dysfunctional. Further analyses are needed to identify such a change and unambiguously assign this case as AIPL1-mediated LCA. Nevertheless, the phenotype of this patient and the affected sibling are consistent with mutations in AIPL1, RPGRIP1, or RetGC1 [15,26], yet no coding sequence variations were identified in RetGC1, RPE65, or CRX. Therefore, for the purposes of this discussion we propose that the H82Y AIPL1 allele contributes to disease in this patient with the caveat that until a second "hit" in AIPL1 is identified this is not a definitive genetic diagnosis. It should also be considered that the retinal morphology observed in this patient could be a consequence of the H82Y mutation in particular, rather than a null allele of AIPL1 per se.

The retina of the LCA patient had undergone severe retinal degeneration with abnormal retinal lamination and morphology. The expression of GFAP, a marker of retinal gliosis, was strongly upregulated in the LCA patient retina. Only a single layer of abnormal photoreceptor cells remained in the outer nuclear layer of the macular area. The surviving photoreceptors were morphologically more cone-like than rod-like in appearance. No photoreceptor outer segments could be detected and the residual inner segments were stunted and highly disorganized. In mouse models of LCA with an inactivated AIPL1 gene, retinal degeneration was first detected at postnatal day 12 [13,18] and only a single layer of photoreceptor nuclei remained at postnatal day 18 [13]. Photoreceptor degeneration was complete after four weeks in one model [13] and most of the photoreceptors were lost by eight weeks in another model [18]. At postnatal day 11, photoreceptor outer segments were shorter than normal and disorganized [13]. A different knockdown approach was used to produce a hypomorphic mutant in which AIPL1 expression was reduced [14]. In this mouse model of LCA, the rate of degeneration was slower in that the thickness of the photoreceptor layer was normal at three months of age but the photoreceptor outer segments were disorganized. By eight months of age, more than half of the photoreceptors were lost and the photoreceptor outer and inner segments were shortened [14]. In all the mouse models of LCA, the gross morphological development of the rod and cone photoreceptors appeared to be normal. However, reduced expression of AIPL1 resulted in a delay in photoresponse onset and recovery prior to retinal degeneration [14], and in the absence of AIPL1 a recordable ERG could not be detected at any age [13,18]. In LCA patients with mutations in AIPL1, a recordable ERG cannot be detected within 1 year of birth. This suggests that although the photoreceptors may appear morphologically normal during development, AIPL1 is essential for the normal functional development of the photoreceptors. Hence, in this LCA patient retina, it is possible that both rod and cone photoreceptors, including the photoreceptor inner and outer segments, underwent normal morphological but nonfunctional development, and that photoreceptor degeneration proceeded after birth.

The expression of AIPL1, rhodopsin, L/M-cone opsin, or S-cone opsin could not be detected in the LCA retina. In the normal adult human retina, AIPL1 expression is restricted to the rod photoreceptors with the exception of the rod outer segments, and is not detected in cone photoreceptors [17]. The lack of immunosignal for AIPL1 in the LCA retina may be due to a loss of functional AIPL1 protein through misfolding or degradation of the mutant protein. Alternatively, the nonfunctional AIPL1 may not be detected in the LCA retina due to the physical loss of rod photoreceptors in the degenerate retina.

In the LCA retina, no immunosignal could be detected for rhodopsin, L/M-cone opsin, or S-cone opsin. The absence of immunosignal for rhodopsin in the LCA patient retina may have occurred as a result of massive retinal degeneration and the physical loss of the rod photoreceptors and rod photoreceptor outer segments. In the mouse models of LCA, rhodopsin was expressed and the protein levels were comparable prior to noticeable retinal degeneration [13,14]. However, no rhodopsin protein was detected after the onset of retinal degeneration at postnatal day 18 in AIPL1^-/- mice [13]. In addition, a decrease in the rod specific markers Nrl and Nr2e3 was detected at postnatal day 15 in AIPL1-deficient retina [18]. After 8 weeks, there was a significant reduction in the rod photoreceptor markers recoverin and rod arrestin, and a reduction in cone-arrestin and cone s-opsin was also detected in AIPL1^-/- retina [18]. In the hypomorphic mouse model, the cone opsins were expressed and there was no evidence of a cone defect up to 11 months of age by immunolabeling for cone opsin [14]. The absence of immunosignal for the cone opsins in the LCA patient retina, therefore, was most likely due to the physical loss and absence of cone outer segments.

Markers for the extracellular matrix sheaths of the photoreceptor inner and outer segments were detected in the LCA retina, but their distribution was abnormal when compared to that of a normal adult retina. The residual inner segments of most of the surviving photoreceptors in the LCA retina were labeled with both peanut and wheat germ agglutinin, suggesting that the extramatrix sheaths are cone-derived. The distribution of cone arrestin and recoverin (a marker for rods and cones) in the LCA retina was also abnormal when compared to that of control retina. In the LCA retina, the residual photoreceptor inner segments that were labeled for recoverin were also positive for cone arrestin, suggesting a cone-dominant phenotype for the surviving photoreceptors. In one mouse model of LCA with complete knockdown of AIPL1, the degeneration of rod and cone photoreceptors was observed to occur at a similar rate, with both rods and cones degenerating at postnatal day 18 [13]. In the other knock-out model, by eight weeks there was a near total loss of photoreceptors including the loss of rod photoreceptors and a reduction in cone photoreceptors [18]. In the hypomorphic AIPL1 mutant mouse model, no evidence of a cone defect was found up to 11 months of age [14]. However, shortening of the rod photoreceptor inner and outer segments was observed at eight months of age suggesting that the degeneration of rod photoreceptors was more severe in this LCA model. Our observations suggest that in human AIPL1 LCA patients, severe rod photoreceptor degeneration can occur and that the surviving nonfunctional photoreceptors are more cone-like. Although cone-like photoreceptors appear to develop in the LCA retina, immunolabeling for cone arrestin was not detected in the cell nuclei, axons, or synaptic terminals of the surviving photoreceptors. This suggests that the connectivity of the surviving photoreceptors in the outer nuclear layer with the cells of the inner nuclear layer at the outer plexiform layer is abnormal, leading to a disruption of retinal organization and signal transduction.

In two mouse models of LCA, all three subunits of the cGMP phosphodiesterase (α, β, and γ) holoenzyme were reduced by a post-transcriptional mechanism prior to the onset of retinal degeneration resulting in elevated levels of cGMP [13,14]. However, the severity of disease in LCA is worse than retinitis pigmentosa linked to mutations in the PDE subunits. Taken together with our data, this suggests additional significant functions for AIPL1 in the developing retina, and in particular in the developing cone photoreceptors. The ectopic expression of AIPL1 in retinal progenitor cells suggested that AIPL1 could alter retinal proliferation and cell fate specification [18]. However, abnormalities in cell proliferation were not observed during retinal development in the AIPL1^-/- retina, suggesting that while AIPL1 was not required for cell proliferation or cell fate specification, it might be necessary for complete photoreceptor functional differentiation or synapse formation [18]. Identification of AIPL1-interactors in the developing retina may provide insight into AIPL1 function during retinal development. It is possible that the rescue of photoreceptor function will require therapeutic intervention during photoreceptor development. However, the survival of cone-like photoreceptors in this LCA patient retina at over 20 years of age is encouraging. It is possible that therapies targeted at bypassing or replacing AIPL1 function may be able to rescue transduction in the surviving cone photoreceptors.

Acknowledgements

We are grateful to Dr. Cheryl Craft (Doheny Eye Institute) for the provision of the LUMIf antibody and Dr. Alexander M. Dizhoor (Pennsylvania College of Optometry, Elkins Park, PA) for the provision of P26 anti-recoverin. JvdS and MEC are supported by the Wellcome Trust. Mutation screening was supported by the Deutsche Forschungsgemeinschaft (DFG Lo 457/5-1,2), ReForM C programme of the Medical Faculty, University of Regensburg, and ProRetina Deutschland e.V to MNP. Supported by Wellcome Trust Grant 068579.

References

1. Perrault I, Rozet JM, Calvas P, Gerber S, Camuzat A, Dollfus H, Chatelin S, Souied E, Ghazi I, Leowski C, Bonnemaison M, Le Paslier D, Frezal J, Dufier JL, Pittler S, Munnich A, Kaplan J. Retinal-specific guanylate cyclase gene mutations in Leber's congenital amaurosis. Nat Genet 1996; 14:461-4.

2. Gu SM, Thompson DA, Srikumari CR, Lorenz B, Finckh U, Nicoletti A, Murthy KR, Rathmann M, Kumaramanickavel G, Denton MJ, Gal A. Mutations in RPE65 cause autosomal recessive childhood-onset severe retinal dystrophy. Nat Genet 1997; 17:194-7.

3. Freund CL, Wang QL, Chen S, Muskat BL, Wiles CD, Sheffield VC, Jacobson SG, McInnes RR, Zack DJ, Stone EM. De novo mutations in the CRX homeobox gene associated with Leber congenital amaurosis. Nat Genet 1998; 18:311-2.

4. Sohocki MM, Bowne SJ, Sullivan LS, Blackshaw S, Cepko CL, Payne AM, Bhattacharya SS, Khaliq S, Qasim Mehdi S, Birch DG, Harrison WR, Elder FF, Heckenlively JR, Daiger SP. Mutations in a new photoreceptor-pineal gene on 17p cause Leber congenital amaurosis. Nat Genet 2000; 24:79-83.

5. Thompson DA, Li Y, McHenry CL, Carlson TJ, Ding X, Sieving PA, Apfelstedt-Sylla E, Gal A. Mutations in the gene encoding lecithin retinol acyltransferase are associated with early-onset severe retinal dystrophy. Nat Genet 2001; 28:123-4.

6. Lotery AJ, Jacobson SG, Fishman GA, Weleber RG, Fulton AB, Namperumalsamy P, Heon E, Levin AV, Grover S, Rosenow JR, Kopp KK, Sheffield VC, Stone EM. Mutations in the CRB1 gene cause Leber congenital amaurosis. Arch Ophthalmol 2001; 119:415-20.

7. Gerber S, Perrault I, Hanein S, Barbet F, Ducroq D, Ghazi I, Martin-Coignard D, Leowski C, Homfray T, Dufier JL, Munnich A, Kaplan J, Rozet JM. Complete exon-intron structure of the RPGR-interacting protein (RPGRIP1) gene allows the identification of mutations underlying Leber congenital amaurosis. Eur J Hum Genet 2001; 9:561-71.

8. Stockton DW, Lewis RA, Abboud EB, Al-Rajhi A, Jabak M, Anderson KL, Lupski JR. A novel locus for Leber congenital amaurosis on chromosome 14q24. Hum Genet 1998; 103:328-33.

9. Dharmaraj S, Li Y, Robitaille JM, Silva E, Zhu D, Mitchell TN, Maltby LP, Baffoe-Bonnie AB, Maumenee IH. A novel locus for Leber congenital amaurosis maps to chromosome 6q. Am J Hum Genet 2000; 66:319-26.

10. Akey DT, Zhu X, Dyer M, Li A, Sorensen A, Blackshaw S, Fukuda-Kamitani T, Daiger SP, Craft CM, Kamitani T, Sohocki MM. The inherited blindness associated protein AIPL1 interacts with the cell cycle regulator protein NUB1. Hum Mol Genet 2002; 11:2723-33. Erratum in: Hum Mol Genet 2003; 12:451.

11. van der Spuy J, Cheetham ME. The Leber congenital amaurosis protein AIPL1 modulates the nuclear translocation of NUB1 and suppresses inclusion formation by NUB1 fragments. J Biol Chem 2004; 279:48038-47.

12. van der Spuy J, Cheetham ME. Role of AIP and its homologue the blindness-associated protein AIPL1 in regulating client protein nuclear translocation. Biochem Soc Trans 2004; 32:643-5.

13. Ramamurthy V, Niemi GA, Reh TA, Hurley JB. Leber congenital amaurosis linked to AIPL1: a mouse model reveals destabilization of cGMP phosphodiesterase. Proc Natl Acad Sci U S A 2004; 101:13897-902.

14. Liu X, Bulgakov OV, Wen XH, Woodruff ML, Pawlyk B, Yang J, Fain GL, Sandberg MA, Makino CL, Li T. AIPL1, the protein that is defective in Leber congenital amaurosis, is essential for the biosynthesis of retinal rod cGMP phosphodiesterase. Proc Natl Acad Sci U S A 2004; 101:13903-8. Erratum in: Proc Natl Acad Sci U S A 2005; 102:515.

15. Heegaard S, Rosenberg T, Preising M, Prause JU, Bek T. An unusual retinal vascular morphology in connection with a novel AIPL1 mutation in Leber's congenital amaurosis. Br J Ophthalmol 2003; 87:980-3.

16. van der Spuy J, Kim JH, Yu YS, Szel A, Luthert PJ, Clark BJ, Cheetham ME. The expression of the Leber congenital amaurosis protein AIPL1 coincides with rod and cone photoreceptor development. Invest Ophthalmol Vis Sci 2003; 44:5396-403.

17. van der Spuy J, Chapple JP, Clark BJ, Luthert PJ, Sethi CS, Cheetham ME. The Leber congenital amaurosis gene product AIPL1 is localized exclusively in rod photoreceptors of the adult human retina. Hum Mol Genet 2002; 11:823-31.

18. Dyer MA, Donovan SL, Zhang J, Gray J, Ortiz A, Tenney R, Kong J, Allikmets R, Sohocki MM. Retinal degeneration in Aipl1-deficient mice: a new genetic model of Leber congenital amaurosis. Brain Res Mol Brain Res 2004; 132:208-20.

19. Porto FB, Perrault I, Hicks D, Rozet JM, Hanoteau N, Hanein S, Kaplan J, Sahel JA. Prenatal human ocular degeneration occurs in Leber's congenital amaurosis (LCA2). J Gene Med 2002 Jul-Aug; 4:390-6.

20. Redmond TM, Yu S, Lee E, Bok D, Hamasaki D, Chen N, Goletz P, Ma JX, Crouch RK, Pfeifer K. Rpe65 is necessary for production of 11-cis-vitamin A in the retinal visual cycle. Nat Genet 1998; 20:344-51.

21. Aguirre GD, Baldwin V, Pearce-Kelling S, Narfstrom K, Ray K, Acland GM. Congenital stationary night blindness in the dog: common mutation in the RPE65 gene indicates founder effect. Mol Vis 1998; 4:23 <http://www.molvis.org/molvis/v4/a23/>.

22. Milam AH, Barakat MR, Gupta N, Rose L, Aleman TS, Pianta MJ, Cideciyan AV, Sheffield VC, Stone EM, Jacobson SG. Clinicopathologic effects of mutant GUCY2D in Leber congenital amaurosis. Ophthalmology 2003; 110:549-58.

23. Zhu X, Craft CM. Modulation of CRX transactivation activity by phosducin isoforms. Mol Cell Biol 2000; 20:5216-26.

24. Zhang Y, Li A, Zhu X, Wong CH, Brown B, Craft CM. Cone arrestin expression and induction in retinoblastoma cells. In: Anderson RE, LaVail MM, Hollyfield JG, editors. New insights into retinal degenerative diseases. Proceedings of the IX International Symposium on Retinal Degeneration; 2000 Oct 9-14; Durango, Colorado. New York: Plenum; 2001. p. 309-17.

25. Acharya S, Rayborn ME, Hollyfield JG. Characterization of SPACR, a sialoprotein associated with cones and rods present in the interphotoreceptor matrix of the human retina: immunological and lectin binding analysis. Glycobiology 1998; 8:997-1006.

26. Hanein S, Perrault I, Gerber S, Tanguy G, Barbet F, Ducroq D, Calvas P, Dollfus H, Hamel C, Lopponen T, Munier F, Santos L, Shalev S, Zafeiriou D, Dufier JL, Munnich A, Rozet JM, Kaplan J. Leber congenital amaurosis: comprehensive survey of the genetic heterogeneity, refinement of the clinical definition, and genotype-phenotype correlations as a strategy for molecular diagnosis. Hum Mutat 2004; 23:306-17.