Baseline 55° fundus autofluorescence images demonstrating the larger autofluorescence area for age of patient 1 (A; mid-30s) and patient 2 (B; early 40s) compared with the control patient (C; early 10s). The scale bars indicate 2 mm.
Patients with a variant affecting the noncanonical splice site of intron 7 demonstrate a milder choroideremia phenotype compared with a cohort of patients with choroideremia. A, Mean autofluorescence island area of each individual (percentage of total 55° field, shown on a logarithmic scale) plotted against age, with patients 1 and 2 identified within a previously published cohort (60 eyes from 30 patients).7 Patient 1 is a significant outlier with a large autofluorescence area for age, and patient 2 has a larger-than-expected autofluorescence area for age, though is less of an outlier. Connecting lines indicate area change over time, the trend line indicates the best-fit curve for all patients (y-axis intercept = 1.975; slope = −0.039; R2 = 0.39), and the shaded area indicates 95% CIs. B, Mean autofluorescence area half-life plotted against baseline age. Patients 1 and 2 have longer degeneration half-lives than the mean autofluorescence area half-life of both eyes of the cohort. The trend line indicates the mean half-life, and the shaded area indicates 95% CIs.
Mildly affected patients with variants affecting the +3 position of the intron 7 donor splice demonstrate residual levels of full-length CHM transcript. A, End point polymerase chain reaction demonstrates that variants in the exon 7 donor splice site result in truncated CHM transcripts consistent with exon 7 skipping and that this is the predominant transcript in patients 1 and 2 and the control patient. Truncated bands (626 base pairs [bp]) are demonstrated in all patients compared with the wild-type phenotype (747 bp). Faint bands consistent with the presence of residual wild-type transcript can be seen in patients 1 and 2 but not in the control patient. No amplification was observed in reverse transcriptase (RT)–deficient reactions. Diagrams indicate primer binding locations within exons in the consensus CHM sequence. B, End point polymerase chain reaction using a primer strategy to only amplify correctly spliced transcripts that contain exon 7. Weak bands in patients 1 and 2 (316 bp) are consistent with amplification of low levels of correctly spliced full-length transcript in the mildly affected patients with a +3 splice site variant. No amplification is seen in the control patient with canonical splice site disruption.
Quantification of levels of residual full-length CHM transcripts. A, Quantitative reverse transcriptase–polymerase chain reaction to assess the relative quantity of all CHM transcripts at different time points. Patients 1 and 2 and the control patient show reduced relative quantities of overall CHM expression normalized to a mean of 6 nonaffected controls (blue line). B, Quantitative reverse transcriptase–polymerase chain reaction results to quantitate the level of full-length transcript produced relative to the predominant exon 7–skipped transcript (del7) in each patient. Patients 1 and 2 have residual levels of full-length transcript, which is not detectable in the control patient. Forward primers for the full-length and del7 transcript spanned either the exon 6/7 or 6/8 junction boundary, respectively, with the same reverse primer in exon 8. Error bars indicate the SEM.
REP1 is undetectable by immunoblot analysis of peripheral blood mononuclear cell lysates in patients 1 and 2 and the control patient compared with wild-type controls. F WT indicates female wild-type control; hREP1, human REP1 protein; kDA, kilodalton; rREP1, recombinant REP1 protein.
eFigure 1. Summary table demonstrating demographic details, Sanger sequencing results across the intron 7 donor splice site, and visual acuity results of all 3 patients at baseline.
eFigure 2. Uncropped gel from Figure 3A.
eFigure 3. Sanger sequencing chromatogram of wild-type RT-PCR product.
eFigure 4. Sanger sequencing of predominant RT-PCR product from the control patient demonstrating exon skipping.
eFigure 5. Uncropped gel from Figure 3B.
eFigure 6. Sanger sequencing of correctly spliced transcripts in patients with a +3 variant in the intron 7 donor splice site.
eFigure 7. Uncropped membrane from Figure 5.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Fry LE, Patrício MI, Williams J, et al. Association of Messenger RNA Level With Phenotype in Patients With Choroideremia: Potential Implications for Gene Therapy Dose. JAMA Ophthalmol. 2020;138(2):128–135. doi:10.1001/jamaophthalmol.2019.5071
Can a mild choroideremia phenotype be explained by genotype?
In this case series study, 2 patients with a splice site variant outside the canonical donor sequence (c.940+3delA) demonstrated a mild phenotype compared with a cohort of 30 patients with choroideremia. Mildly affected patients expressed residual levels of full-length CHM transcript, while a control patient with normal disease progression due to a canonical donor splice site variant (c.940+2T>A) did not express any CHM transcript.
Residual levels of full-length CHM transcript were associated with milder disease, providing evidence of an association of genotype with phenotype in choroideremia and suggesting the transcript level required to slow degeneration with gene therapy.
Gene therapy is a promising treatment for choroideremia, an X-linked retinal degeneration. The required minimum level of gene expression to ameliorate degeneration rate is unknown. This can be interrogated by exploring the association between messenger RNA (mRNA) levels and phenotype in mildly affected patients with choroideremia.
To analyze CHM mRNA splicing outcomes in 2 unrelated patients with the same c.940+3delA CHM splice site variant identified as mildly affected from a previous study of patients with choroideremia.
Design, Setting, and Participants
In this retrospective observational case series, 2 patients with c.940+3delA CHM variants treated at a single tertiary referral center were studied. In addition, a third patient with a c.940+2T>A variant that disrupts the canonical dinucleotide sequence at the same donor site served as a positive control. Data were collected from October 2013 to July 2018.
Main Outcomes and Measures
Central area of residual fundus autofluorescence was used as a biomarker for disease progression. CHM transcript splicing was assessed by both end point and quantitative polymerase chain reaction. Rab escort protein 1 (REP1) expression was assessed by immunoblot.
The 2 mildly affected patients with c.940+3delA variants had large areas of residual autofluorescence for their age and longer degeneration half-lives compared with the previous cohort of patients with choroideremia. The control patient with a c.940+2T>A variant had a residual autofluorescence area within the range expected for his age. Both patients with the c.940+3delA variant expressed residual levels of full-length CHM mRNA transcripts relative to the predominant truncated transcript (mean [SEM] residual level: patient 1, 2.3% [0.3]; patient 2, 4.7% [0.2]), equivalent to approximately less than 1% of the level of full-length CHM expressed in nonaffected individuals. Full-length CHM expression was undetectable in the control patient. REP1 expression was less than the threshold for detection both in patients 1 and 2 and the control patient compared with wild-type controls.
Conclusions and Relevance
These results demonstrate the first genotype-phenotype association in choroideremia. A +3 deletion in intron 7 is sufficient to cause choroideremia in a milder form. If replicated with gene therapy, these findings would suggest that relatively low expression (less than 1%) of the wild-type levels of mRNA would be sufficient to slow disease progression.
Choroideremia is an X-linked inherited retinal degeneration caused by variants in CHM (OMIM, 300390). Affected male patients present with night blindness and constriction of visual fields, with central visual acuity typically preserved until the fourth decade of life, beyond which complete blindness can occur. CHM encodes Rab escort protein 1 (REP1), a ubiquitously expressed protein involved in intracellular trafficking.1 REP1 deficiency leads to a characteristic atrophic fundal appearance due to degeneration of the retinal pigment epithelium (RPE), photoreceptors, and choroid.2
Gene therapy in choroideremia aims to prevent disease progression through the replacement of functional copies of CHM. Recent phase I/II randomized clinical trial results in patients have demonstrated encouraging visual outcomes and safety.3 To slow retinal degeneration with gene therapy, a minimum amount of CHM expression in the retina is required, but in humans, this level is challenging to determine.
Establishing a link between genotype and phenotype may provide insight into factors that affect the rate of degeneration in choroideremia and may help to assess the long-term effects of gene therapy. Many variants have been described in CHM since its identification as the causative gene implicated in choroideremia in 19904; however, an association of genotype with phenotype in choroideremia has yet to be established.5,6
To assess the rate of RPE degeneration in patients with choroideremia, we can follow the rate of RPE degeneration in patients using fundus autofluorescence imaging.7 Retinal pigment epithelium autofluorescence is generated by retinoid by-products of the visual cycle that accumulate in the RPE.7,8 The centripetal degeneration pattern in choroideremia leaves a central area of residual RPE that can be measured to reproducibly assess the rate of RPE degeneration in choroideremia.7,8
From a longitudinal cohort of patients with choroideremia described in 2018,7 we identified 2 unrelated patients with choroideremia with a slower rate of disease progression (patient 1 and patient 2). We were intrigued to discover that both mildly affected patients carried an identical c.940+3delA splice site variant affecting the donor splice site of CHM intron 7. Although some variation of the 5′ splice site can be tolerated, variants in the invariant GT dinucleotide sequence in the first nucleotides of the intron are usually sufficient to abolish splicing.9 Such variants would be expected to cause exon 7 skipping, generating a premature termination codon through open reading frameshift and resulting in nonsense-mediated decay. This study investigated the hypothesis that as c.940+3delA affects the +3 nucleotide, some viable full-length RNA transcripts might be generated in these patients, thus affecting the rate of degeneration.
Patients were seen as part of the screening process for choroideremia gene therapy trials as approved by the UK Research Ethics Committee (GTAC171 and 15/LO/1379). The study was conducted in accord with the Declaration of Helsinki, and written informed consent was obtained.
Patients underwent comprehensive ophthalmological examination at baseline and at 1-year and 2-year follow-up.7 This included measurement of best-corrected visual acuity and dilated bilateral autofluorescence imaging.
Images for each patient were collected at each visit and analyzed, as in a previously published cohort.7 Unlike in that study, all patients with less than 2 years of follow-up were excluded from half-life analysis. BluePeak (488 nm) 55° fundus autofluorescence images were captured using the Spectralis confocal scanning laser ophthalmoscope (Heidelberg Engineering). The central contiguous autofluorescence island of 60 eyes of 30 patients with choroideremia were outlined manually by 2 independent graders using Heidelberg Eye Explorer (HEYEX; Heidelberg Engineering) to obtain a mean, and the outline was registered and normalized to the baseline image to adjust for small differences in focal length between visits. Mean area was expressed as a percentage of the total measured 55° field area. For each patient, the highest-quality image was chosen for analysis at baseline, 1 year, and 2 years. For patient 2, only the right eye was included in all analyses, as the patient received gene therapy in the left eye following baseline assessment (ClinicalTrials.gov identifier: NCT02407678), and this eye was excluded from this analysis. For the control patient, only the left eye was included in analyses, as low quality of autofluorescence imaging of the right eye prevented accurate autofluorescence measurements in this eye. An individual half-life of mean autofluorescence area was calculated for each patient using a linear regression of the logarithm of the autofluorescence area to obtain a decay constant (λ) for each patient from the gradient of the line of best fit for that patient.7 Half-life (t½) was subsequently calculated as N(t) = N(0)e−λt, where t½ = ln(2)/λ.
DNA extraction using the Chemagic Magnetic Separation Module 1 (PerkinElmer Chemagen Technologie) was performed on whole-blood specimens. Bidirectional sequencing was used to explore the 15 CHM gene exons and the intron/exon junctions.10 The BigDye Terminator 3.1 Cycle Sequencing Kit (Thermo Fisher Scientific) was used for sequencing.
RNA was isolated from peripheral blood in ethylenediaminetetraacetic acid using the QIAamp RNA blood mini kit (Qiagen). Approximately 3 μg of RNA was reverse transcribed using random hexamer priming and the SuperScript Reverse Transcriptase kit (Thermo Fisher Scientific). Primer pairs spanning CHM exons 5 to 9 and 7 to 9 were designed to assess the presence of truncated transcripts and correctly spliced transcripts, respectively. End point polymerase chain reaction (PCR) amplification was conducted using KAPA2G Fast ReadyMix PCR Kit (Kapa Biosystems). Products were visualized on a 2% agarose gel treated with ethidium bromide and confirmed via Sanger sequencing. Quantitative PCR was used to quantify gene expression using SYBR Green mastermix (Bio-Rad Laboratories).
To analyze relative quantities of overall CHM transcript expression, a TaqMan-based probe spanning the exon 3/4 junction was used to detect CHM messenger RNA (mRNA) expression (Hs01114157_m1; Thermo Fisher Scientific). Expression level was normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Hs02758991_g1; Thermo Fisher Scientific). TaqMan probes followed a 6-carboxyfluorescein–minor groove binder (FAM-MGB) design. Quantitative reverse transcription PCR reactions were prepared in triplicate using TaqMan Fast Universal PCR Master Mix (Thermo Fisher Scientific) and ran using the QuantStudio 7 Flex Real-Time PCR System (Thermo Fisher Scientific). The CHM expression level for each patient was normalized to their GAPDH level using the ΔCt method. The normalized CHM levels were then plotted relative to the normalized mean CHM levels in 6 nonaffected controls using the ΔΔCt method.
Relative expressions of wild-type transcript and transcript lacking exon 7 were determined using primer pairs specifically targeting each transcript, with forward primers spanning either the exon 6/7 or 6/8 junction, respectively. Primer pair efficiencies as measured by standard curve analysis performed on serial dilution of a PCR product were greater than 90%.
Analysis of REP1 expression of peripheral blood mononuclear cells was assessed as previously described.11 To prepare total cell lysates, cells were disrupted by passage through a 26-gauge needle attached to a 1-mL syringe, then centrifuged for 5 minutes at 1500 relative centrifugal force at 4°C. The supernatant was centrifuged at 100 000 relative centrifugal force for 1 hour at 4°C. The supernatant was kept as total cell lysate for prenylation reactions in vitro. Total protein content was determined using the Bradford method according to the manufacturer’s instructions (Quick Start Bradford 1x Dye Reagent; Bio-Rad Laboratories), and sample values were extrapolated from a standard curve using a sigmoidal 4-parallel line regression.
Reaction products underwent sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 10% precast polyacrylamide gel, transferred to a polyvinylidene fluoride membrane (Bio-Rad Laboratories) and blocked with blocking buffer (phosphate buffered saline [Thermo Fisher Scientific] + 0.1% Tween20 [Sigma-Aldrich] + 3% bovine serum albumin [Sigma-Aldrich]) for 45 minutes. For protein expression, membranes were incubated separately for β-actin monoclonal antibody (AM4302; Thermo Fisher Scientific; 1:20 000 solution) and for anti–REP1 antibody, clone 2F1 (MABN52; Merck KGaA; 1:2500 solution), in blocking buffer for 1 hour under agitation. Membranes were washed 3 times for 7 minutes with phosphate buffered saline + 0.1% Tween20, incubated with fluorescent-labeled secondary antibody for 30 minutse in blocking buffer (1:10 000 solution), washed again as before, and detected using an Odyssey Imaging System (LI-COR Biosciences).
Data are expressed as means with SEMs or 95% CIs. Statistical comparisons were performed with a 1-way analysis of variance. Bonferroni correction was used for adjustment for multiple comparisons, with an adjusted significance level of P less than .05. Statistical analysis was carried out using GraphPad Prism version 8.0 (GraphPad Software).
Within a previously published cohort of 30 patients,7 we identified 2 unrelated patients with significantly milder disease as demonstrated by a larger-than-expected central island of residual autofluorescence for age (Figure 1A and B and Figure 2A). As the rate of autofluorescence loss in this cohort follows an exponential decay function, the half-life of the mean autofluorescence area can be calculated to approximate the rate of degeneration.7 The degeneration half-lives in patient 1 (14.1 years) and patient 2 (10.2 years) are 3.4 and 1.7 SDs greater than the cohort mean, respectively (6.3 years; 95% CI, 5.3-7.2; n = 24), confirming their slower degenerating phenotype (Figure 2B). Patients 1 and 2 both had c.940+3delA variants in the donor splice site of intron 7 (eFigure 1 in the Supplement). The presence of other variants within CHM that might impact splicing efficiency were excluded.
A patient with an alternative variation disrupting the invariant dinucleotide sequence, c.940+2T>A, was identified as a control. Accordingly, this patient demonstrated an area of autofluorescence loss expected for his age and a degeneration half-life of 5.8 years (Figure 1C and Figure 2).
The CHM gene is expressed ubiquitously as a single splice variant, and reduced prenylation activity due to the absence of REP1 can be quantified in the white blood cells of patients affected by choroideremia.2 Hence, mRNA analyses of splicing can be performed on peripheral blood mononuclear cells. Therefore, we used blood samples to explore the association of these variants with CHM splicing using random-primed complementary DNA synthesis on total RNA followed by end point PCR using primer pairs targeting different regions of the consensus CHM transcript (NM_000390.3). Using primers annealing within exons 5 and 9, a truncated product consistent with skipping of the 121–base pair (bp) exon 7 was demonstrated as the predominant transcript both in patients 1 and 2 and in the control patient (626 bp vs 747 bp in the wild-type phenotype) (Figure 3A) (eFigure 2 in the Supplement). The identity of the wild-type and exon-skipped products were confirmed with Sanger sequencing (eFigures 3 and 4 in the Supplement). However, closer inspection of results from patients 1 and 2 identified a faint second band (at 747 bp) consistent with the presence of residual full-length transcript.
To investigate this further, a second PCR was carried out using primers annealing within exons 7 and 9, such that amplification would only occur in the presence of a full-length CHM transcript. A weak band was observed in patients 1 and 2 but not in the control patient (Figure 3B) (eFigure 5 in the Supplement), with wild-type transcript expression confirmed by Sanger sequencing (eFigure 6 in the Supplement). This result confirmed the prediction of some full-length transcript being generated in the 2 patients who showed reduced disease severity.
Having discovered that very low levels of correctly spliced mRNA transcripts appeared to ameliorate the choroideremia disease phenotype considerably, we next sought to determine the levels more precisely. Hence, quantitative reverse transcriptase–PCR was used to quantify the level of CHM transcripts produced in each patient. Probes spanning the exon 3/4 junction were selected to target both the truncated and full-length CHM transcripts. We found that patients 1 and 2 and the control patient expressed reduced mean (SEM) quantities of total CHM mRNA (patient 1, 14.7% [4.7]; P < .001; patient 2, 21.0% [5.7]; P < .001; control patient, 22.3% [6.3]; P < .001) compared with 6 nonaffected controls (100% [1.5]) (Figure 4A). Although nonsense-mediated decay efficiency can be variable in patients with choroideremia,12 no significant differences in total CHM levels were detected between patients. We then assessed the relative abundance of full-length to truncated CHM mRNA in patients 1 and 2. Primer sets specific to the full-length transcript and the predominantly expressed exon 7–skipped transcript were designed with forward primers spanning the exon 6/7 or 6/8 junction, respectively. The results indicate that full-length CHM accounts for a mean (SEM) of 2.3% (0.3) of predominant transcript in patient 1 (from 3 replicates) and 4.7% (0.2) of predominant transcript in patient 2 (from 3 replicates) (Figure 4B). This low level of expression of full-length CHM was not detected in the control patient (from 3 replicates), corroborating the findings of the end point PCR and demonstrating that full-length CHM transcripts are present in patients who show a slow degeneration.
To assess the association of the residual level of CHM transcript with protein expression, REP1 expression was analyzed with immunoblot in peripheral blood mononuclear cells from the patients. Despite the residual full-length mRNA expression detected, REP1 protein was undetectable within the limits of the assay for patients 1 and 2 and the control patient compared with controls without CHM variants (Figure 5) (eFigure 7 in the Supplement).
Establishing a genotype-phenotype association in choroideremia is helpful for counselling patients and assessing the long-term outcomes of gene therapy; however, a link has been challenging to identify. Studying a cohort of patients with choroideremia using fundus autofluorescence area as a reproducible measure of RPE degeneration enabled the identification of mildly affected patients and subsequent interrogation of underlying genetic factors contributing to their slower progressing degeneration. The large residual autofluorescence islands for age observed in patients with the c.940+3delA variant may be caused either by a slower rate or delayed onset of degeneration.7 The longer autofluorescence degeneration half-lives in these patients supports that the progression of autofluorescence area loss is slower in these patients.
Aylward et al7 suggested that patient 2 had a normal rate of degeneration across 1 year (labeled as P27 in that article); however, it should be noted that the current study uses data from patient 2 from 3 points in time over 2 years, which gives a more accurate measurement. Furthermore, the previous study measured the mean of both eyes, and the small area remaining in the left eye in patient 2 would increase the error in the measurements over such a short time frame.7 Since patient 2 underwent gene therapy to the eye with the smaller residual area, we could no longer include this in our analysis, and so the 2-year measurement in the single untreated eye most likely represents a more accurate prediction of his rate of degeneration. Furthermore, the independent analysis of the residual area against age confirmed that the patient was also an outlier.
Donor splice site variants in the +3 position outside the invariant dinucleotide sequence of intron 7 are sufficient to be pathogenic in choroideremia but appear to be associated with a milder form of disease. Splice site variants are common in choroideremia, although they usually affect the canonical positions in the donor or acceptor sites; of 280 variants in the Leiden Open Variation Database13 of patients with choroideremia, 61 involved splice sites and 11 of these occurred outside the canonical donor or acceptor sequences. The consensus sequence of the 5′ splice site starting at the terminal 2 positions of the exon is AG|GTRAGT, where the variable +3 R is the nucleotide A or G. Therefore, the functional effects of +3 variants are likely to be sequence and location specific.14 In choroideremia, +3 disease-causing variants at the 5′ splice site have been reported in intron 1, 7, and 13,13 but the intron 7 splice donor site has a run of 8 T nucleotides immediately downstream of the deletion, which would limit the effects of the single nucleotide frameshift in these 2 patients and may make this variant unique to other noncanonical splice site variants. A variant in the intron 13 splice donor (c.1609+3A>C) was reported to cause full exon 13 skipping15; however, investigations to assess for residual full-length transcripts were not undertaken. We were unable to detect REP1 expression at the low levels of mRNA detected in peripheral tissue (peripheral blood mononuclear cells). Although CHM is ubiquitously expressed, splicing outcomes may differ in retinal tissue.
Unfortunately, there are no known animal models of male choroideremia against which to test the hypothesis. In mice, REP1 deficiency is embryologically lethal, and the hemizygous female Chmnull/+ mouse, which carries a copy of the Chm knockout allele, expresses Chm and has a mild phenotype.16 The conditional knockout Chmflox mouse, which allows the selective knockout of Chm using the Cre-Lox system, overcomes the issue of embryonic lethality; however, this produces a mild degenerative phenotype.17
These results indicate that if similar levels of mRNA expression could be achieved with gene therapy for choroideremia, it may be sufficient to slow disease progression significantly. Taken together, we see that in patients 1 and 2 and the control patient, the mean (SEM) level of total CHM transcript is significantly reduced (range, 14.7% [4.7] to 22.3% [6.3]) compared with nonaffected individuals (Figure 2) and is predominantly a truncated transcript. However, in mildly affected patients, a small proportion of all CHM transcripts are correctly spliced full-length transcripts (mean [SEM]: patient 1, 2.3% [0.3]; patient 2, 4.7% [0.2]) (Figure 2). This relatively low proportion of full-length transcripts in patients 1 and 2 represents a level approximate to less than 1% of the level of full-length CHM expressed in nonaffected individuals. The finding that relatively low wild-type levels of CHM mRNA are associated with slower degeneration provides insight into the minimum dose required for choroideremia gene therapy.
The study is limited by the small sample size (2 patients) and the short follow-up period. Identification of patients with similar intronic and splice site variants in CHM that result in the expression of low levels of full-length CHM mRNA would be useful to validate this observation in future studies.
We demonstrate evidence of a genotype-phenotype association in choroideremia, showing that patients with the c.940+3delA noncanonical splicing variant maintain very low levels of correctly spliced mRNA transcripts, which appears to ameliorate the choroideremia disease phenotype considerably. If replicated with gene therapy, it appears that relatively low wild-type levels of mRNA expression could be sufficient to slow down disease progression.
Accepted for Publication: October 8, 2019.
Corresponding Author: Robert E. MacLaren, MD, PhD, FRCOphth, Nuffield Laboratory of Ophthalmology, Nuffield Department of Clinical Neurosciences, University of Oxford, West Wing, Level 6, John Radcliffe Hospital, Oxford OX3 9DU, United Kingdom (email@example.com).
Published Online: December 19, 2019. doi:10.1001/jamaophthalmol.2019.5071
Author Contributions: Drs Fry and MacLaren had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Fry, Patrício, and Williams contributed equally as joint first authors
Study concept and design: Fry, Patrício, Williams, Hewitt, Xue, MacLaren.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Fry, Williams, Aylward, MacLaren.
Critical revision of the manuscript for important intellectual content: Fry, Patrício, Williams, Hewitt, Clouston, Xue, Barnard, MacLaren.
Statistical analysis: Fry, Patrício, Williams, MacLaren.
Obtained funding: MacLaren.
Administrative, technical, or material support: Patrício, Williams, Aylward, Hewitt, Xue, MacLaren.
Study supervision: Clouston, Xue, Barnard, MacLaren.
Conflict of Interest Disclosures: Drs Patrício, Barnard, and MacLaren have received personal fees for consulting from Nightstar Therapeutics, which is developing gene therapy for choroideremia. Dr MacLaren has received grant funding from Nightstar Therapeutics. No other disclosures were reported.
Funding/Support: This work was supported by grants from the NIHR Oxford Biomedical Research Centre and the Royal College of Surgeons of Edinburgh. Additional support was provided by the Rhodes Trust, the Amar-Franses and Foster-Jenkins Trust, and the North Harbour Club Charitable Trust.
Role of the Funder/Sponsor: The funders 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.
Meeting Presentation: Aspects of this work were presented at the 2018 Association for Research in Vision and Ophthalmology (ARVO) meeting; May 1, 2018; Honolulu, Hawaii.
Additional Contributions: We acknowledge Jasleen K. Jolly, MSc (Nuffield Laboratory of Ophthalmology, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom), for the development of the method for autofluorescence area measurement. She was not compensated for her work.