Cone and Rod Loss in Stargardt Disease Revealed by Adaptive Optics Scanning Light Ophthalmoscopy

Importance  Stargardt disease (STGD1) is characterized by macular atrophy and flecks in the retinal pigment epithelium. The causative ABCA4 gene encodes a protein localizing to photoreceptor outer segments. The pathologic steps by which ABCA4 mutations lead to clinically detectable retinal pigment epithelium changes remain unclear. We investigated early STGD1 using adaptive optics scanning light ophthalmoscopy.

Observations  Adaptive optics scanning light ophthalmoscopy imaging of 2 brothers with early STGD1 and their unaffected parents was compared with conventional imaging. Cone and rod spacing were increased in both patients (P < .001) with a dark cone appearance. No foveal cones were detected in the older brother. In the younger brother, foveal cones were enlarged with low density (peak cone density, 48.3 × 10³ cones/mm²). The ratio of cone to rod spacing was increased in both patients, with greater divergence from normal approaching the foveal center, indicating that cone loss predominates centrally and rod loss increases peripherally. Both parents had normal photoreceptor mosaics. Genetic testing revealed 3 disease-causing mutations.

Conclusions and Relevance  This study provides in vivo images of rods and cones in STGD1. Although the primary clinical features of STGD1 are retinal pigment epithelial lesions, adaptive optics scanning light ophthalmoscopy reveals increased cone and rod spacing in areas that appear normal in conventional images, suggesting that photoreceptor loss precedes clinically detectable retinal pigment epithelial disease in STGD1.

Stargardt disease (STGD1) is characterized by macular atrophy and peripheral flecks in the retinal pigment epithelium (RPE). The causative gene, ABCA4 (OMIM 601691),¹ encodes a protein localizing to photoreceptor outer segments² that transports vitamin A byproducts across the disc membrane.³ Loss of ABCA4 function is associated with RPE lipofuscin accumulation⁴ and photoreceptor degeneration in mouse models. A pathogenic sequence of lipofuscin accumulation leading to RPE cell damage and then photoreceptor loss has been proposed.⁵

Mutations of ABCA4 are associated with a spectrum of phenotypes, including cone-rod dystrophy and retinitis pigmentosa.⁶ Several hundred sequence variations in ABCA4 have been identified.^7,8 Assessment of the pathogenic contribution of disease-causing alleles has indicated the presence of non-ABCA4 modifying factors.⁹ Given the phenotypic variability of ABCA4-associated disease, predicting its spatial and temporal progression is critical for the development of potential therapies.¹⁰

Conventional imaging has been useful to characterize STGD1. Fundus autofluorescence (FAF) shows early diffuse lipofuscin accumulation, hyper-AF within flecks, and hypo-AF indicating RPE atrophy.¹¹ Optical coherence tomography (OCT) shows thickening of the foveal external limiting membrane, suggesting early gliosis.¹² These methods remain limited by the eye’s optical irregularities.

Adaptive optics scanning light ophthalmoscopy (AOSLO) enables visualization of single photoreceptor cells in the living human retina, including foveal cones and peripheral rods.¹³ Diseased photoreceptors can be measured quantitatively, and cone loss has been shown¹⁴ to correlate with clinically identified lesions in STGD1. To investigate early STGD1, we used AOSLO to evaluate 2 brothers with macular atrophy, a mild phenotype, and their unaffected parents.

Box Section Ref ID

Complete ophthalmic examinations, fundus photographs (FF450plus; Carl Zeiss Meditec Inc), and spectral-domain OCT and FAF (Spectralis OCT and Heidelberg Regina Angiograph; Heidelberg Engineering) were obtained. The University of Rochester Research Subjects Review Board approved this study. Written informed consent was obtained and participants received financial compensation.

Participants’ DNA was tested for mutations in ABCA4. Sequence variations were evaluated as disease causing by using standard techniques.⁷

Cone and rod photoreceptor images were acquired using the Rochester AOSLO technique (eMethods in the Supplement). Photoreceptors were labeled semiautomatically. A foveal cone density map was generated and peak cone density was measured. The center of the foveal avascular zone and the preferred retinal locus were identified. Photoreceptor spacing was compared with normative AOSLO or histologic data¹⁵ using 1-way analysis of variance with Bonferroni correction.

Two brothers with STGD1 and their unaffected parents, both in their 40s, were examined. The older brother (II-1) had 20/150 Snellen visual acuity with a central scotoma. The younger brother (II-2) had 20/30 Snellen visual acuity. The father (I-1) and mother (I-2) had normal visual acuity. Molecular genetic analysis revealed 3 disease-causing ABCA4 mutations: Gly1961Glu (paternal allele) and Gly863Ala and Arg2030Stop (maternal allele).

Both parents’ fundus, OCT, and FAF images were normal (Figure 1A-C). One son (II-1) showed macular atrophy with no peripheral flecks. Optical coherence tomography confirmed atrophy of the outer retina, RPE at the fovea, and normal layers at 1.7 mm. Fundus autofluorescence indicated central hypo-AF, surrounding hyper-AF at 0.7 mm, and uniform AF at 1.7 mm (Figure 1A-C).

The other son (II-2) had a subtle bull’s-eye maculopathy with no peripheral flecks. Optical coherence tomography showed foveal preservation of the outer segments with a thickened external limiting membrane, perifoveal atrophy of the outer retina and RPE, and normal layers at 1.7 mm. Fundus AF indicated a bull’s-eye with central hyper-AF, and surrounding annular hypo-AF, then hyper-AF at 0.7 mm. Autofluorescence was uniform at 1.7 mm (Figure 1A-C).

Both parents’ photoreceptor mosaics were normal (Figure 1D-F). Cones were continuous and densely packed at the fovea, with increased spacing eccentrically. At 1.7 mm, rods were identified, but rod spacing could not be measured reliably owing to dense packing and the more limited resolution for rods at the wavelength and pinhole settings used.

At the fovea, no cones were identifiable in patient II-1 (Figure 1D). In patient II-2, foveal cones were sparse and enlarged, with a surrounding annulus of no identifiable cones (Figure 1D and Figure 2A). Peak cone density measured 48.3 × 10³ cones/mm² (normal, 199 × 10³ cones/mm² ± 87 × 10³ cones/mm²). The locations of peak cone density, foveal avascular zone center, and preferred retinal locus were within 50 µm (Figure 2B).

Eccentrically, photoreceptors were qualitatively similar for patients II-1 and II-2. At 0.7 mm, cones were sparse, and cone spacing could not be measured reliably owing to the absence of a continuous mosaic (Figure 1E). At 1.7 mm, cones were abnormally dark, enlarged, and sparse; individual rods were identified and quantifiable (Figure 1F; Figure 2C and D).

Both parents’ cone spacing was normal at all locations measured. In both affected brothers, cone spacing was increased and was worse in patient II-1 (P < 10⁻⁶). Rod spacing was also increased and was worse in patient II-1 (P = .048), diverging further from normal with increasing eccentricity. The ratio of cone to rod spacing was increased, again worse in patient II-1 (P < 10⁻⁶), and was most divergent from normal at lower eccentricities (Figure 3).

Although cone loss has been shown¹⁴ to correlate with clinically identified lesions in STGD1, in this study, in vivo AOSLO imaging of both rods and cones revealed several patterns of photoreceptor disease that are not detectable by conventional imaging methods. Although both affected individuals manifested macular atrophy without flecks, photoreceptor spacing was significantly increased peripheral to the clinically detectable lesions. Fundus autofluorescence is homogeneous at these locations, suggesting that a decline in photoreceptors precedes lipofuscin accumulation in the macular atrophy phenotype. The finding that increased photoreceptor spacing was greater in the older brother with more advanced disease lends further credence to the hypothesis that photoreceptor loss represents an early step in the pathogenic sequence. Further investigation, including longitudinal follow-up and fluorescence AOSLO, is needed to confirm this interpretation.

In patient II-2, peak foveal cone density measured approximately 25% of the normal mean, consistent with a previous report¹⁴ describing increased cone spacing near the fovea in STGD1. The foveal cones were continuous and larger than normal. At 1.7 mm in both patients, rods were similarly continuous with increased spacing. Whether photoreceptors enlarge and migrate to fill gaps where cells have died or whether fewer photoreceptors are present at birth in STGD1 remains unknown.

The ratio of cone to rod spacing was increased, with greater divergence from normal approaching the fovea, suggesting that cone loss predominates centrally and rod loss increases peripherally. In patient II-2, the locations of peak cone density, preferred retinal locus, and foveal avascular zone center did not overlap but were within normal range. The divergence of these points may be useful to monitor early foveal disease.

Peripherally, cone reflectance was diminished, producing a dark appearance. The presence of dark cones has been described^16-19 in other retinal degenerations and may represent foreshortened outer segments. Nonconfocal split-detector AOSLO has shown¹⁶ intact inner segments underlying dark cones, which may indicate potential targets for treatment, such as gene therapy.

Two of the 3 disease-causing mutations—Gly1961Glu (paternal) and Gly863Ala (maternal)—have been associated with a milder visual acuity and visual field phenotype.⁹ The additional Arg2030Stop mutation on the maternal allele is uncommon, and its pathogenic contribution has not been well described, but the 2 mutations on the maternal allele were not sufficient to cause disease in the carrier state.

Adaptive optics scanning light ophthalmoscopy reveals increased cone and rod spacing in regions outside of the clinically detectable RPE changes in the macular atrophy phenotype of STGD1. Cone loss predominates closer to the fovea, with a greater contribution from rod loss in the periphery. Dark cones may indicate areas of outer-segment loss with preserved inner segments, suggesting potential therapeutic targets.

Corresponding Author: Mina Chung, MD, Flaum Eye Institute, University of Rochester, 601 Elmwood Ave, PO Box 659, Rochester, NY 14642 (mina_chung@urmc.rochester.edu).

Submitted for Publication: March 6, 2015; final revision received June 2, 2015; accepted June 10, 2015.

Published Online: August 6, 2015. doi:10.1001/jamaophthalmol.2015.2443.

Author Contributions: Drs Song and Chung had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Song, Hunter, Williams, Chung.

Acquisition, analysis, or interpretation of data: Song, Rossi, Latchney, Bessette, Stone, Hunter, Chung.

Drafting of the manuscript: Song, Rossi, Latchney, Chung.

Critical revision of the manuscript for important intellectual content: Song, Rossi, Bessette, Stone, Hunter, Williams, Chung.

Statistical analysis: Song, Chung.

Obtained funding: Stone, Chung.

Administrative, technical, or material support: Song, Latchney, Stone, Williams.

Study supervision: Stone, Hunter, Chung.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Drs Song and Rossi have received grant support from Canon Inc, Dr Hunter has received grant support from Polgenix Inc, Dr Williams has received grant support from Polgenix Inc and Canon Inc, and Dr Chung has served as a consultant for GlaxoSmithKline and received grant support from Canon Inc. No other disclosures are reported.

Funding/Support: This work was supported by National Institutes of Health (NIH) grants EY021786, EY014375, and EY001319; a grant from the Edward N. & Della L. Thome Memorial Foundation; University of Rochester Clinical Trials and Science Institute award number UL1 RR024160 from the National Center for Research Resources and the National Center for Advancing Translational Sciences of the NIH; and an unrestricted departmental grant from Research to Prevent Blindness.

Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Margaret Folwell, MS (Center for Visual Science, University of Rochester), performed image processing. There was no financial compensation.