Increased Plasma cGMP in a Family With Autosomal Recessive Retinitis Pigmentosa Due to Homozygous Mutations in the PDE6A Gene

Abstract

Purpose: To describe genotype and phenotype in a family with autosomal recessive retinitis pigmentosa (arRP) carrying homozygous mutations in the gene for the α-subunit of cyclic guanosine monophosphate (cGMP)–hydrolyzing phosphodiesterase 6 (PDE6A). Moreover, to compare their plasma cGMP levels to controls, exploring the possible role for cGMP in RP diagnostics.

Methods: Seven siblings and their parents were recruited. Microarray, verified by Sanger sequencing, was used for genotyping. Investigations included slit lamp and fundus examination, Goldmann perimetry, full-field and multifocal electroretinography (ERG), and optical coherence tomography (OCT). Cyclic GMP was measured with an immunoassay kit.

Results: All siblings and their father were homozygous, and the mother heterozygous, for IVS6+1G>A in PDE6A. Seven family members also carried c1532G>A in ABCA4. Visual fields were constricted with mere central remnants in older subjects and additional temporal crescents in younger subjects. Visual acuity ranged from 0.8 to amaurosis. Full-field ERGs showed extinguished rod responses and minimal cone responses. Multifocal ERGs were severely reduced. Optical coherence tomography revealed either general attenuation or central macular edema. Mean plasma cGMP in patients was approximately twice that in controls.

Conclusions: To our knowledge, this is the first phenotypic description of arRP due to homozygous IVS6+1G>A mutations in PDE6A and these seem here to be associated with severe RP leading to early extinction of rod responses as well as reduced macular function. Additionally, patients showed increased plasma levels of cGMP, indicating a possible role for cGMP measurements as part of the clinical tests for this and, after further investigations, maybe other forms of RP.

Retinitis pigmentosa (RP) belongs to the group of hereditary retinal degenerations. The prevalence of RP is approximately 1 in 4000 persons worldwide¹ and it is a major cause of progressive deterioration of vision and blindness among younger people.² The disease typically starts with loss of night vision in adolescence, followed by visual field constriction in young adulthood and gradual reduction of central visual acuity later in life.¹ The most common mode of inheritance is autosomal recessive (ar) (50%–60% of cases), but RP can also be inherited in an autosomal dominant (30%–40%) or X-linked (5%–15%) way.^3,4 At the moment, more than 60 genes associated with RP have been identified⁵ (see also: https://sph.uth.edu/retnet/home.htm; provided in the public domain by The University of Texas Health Science Center, Houston, Texas, USA). Many of these genes code for proteins that are involved in the phototransduction cascade, the retinoid metabolism, or in the maintenance of photoreceptor integrity.⁶ To date, RP cannot be cured, although gene therapy has been tested for some genetic variants.^7,8 There are also no established biochemical blood markers yet to contribute to the clinical test possibilities, although a strategy for cases with RP mutations concerning the synthesis of certain lipids has been put forward.⁹ 

Despite the diverse genetic origin of RP, fundus findings most often share a common pattern with pale optic nerve head, attenuated retinal blood vessels, and bone-spicule pigmentations in the midperiphery or toward the far periphery.¹ Those morphologic changes are preceded and accompanied by photoreceptor cell death starting with rod photoreceptors and secondarily also engaging cones, which are considered to get damaged and die from oxidative stress,^10–12 loss of metabolic and trophic support,¹³ or toxicity due to rod cell death.¹⁴ Indeed, alterations in antioxidative status have been demonstrated in peripheral blood of RP patients,¹⁵ with reduced activity of superoxide dismutase 3 and increased levels of nitric oxide (NO). This imbalance in antioxidant status may contribute to a poorer capacity of RP patients to cope with toxic oxygen metabolites.¹⁵ 

The small second messenger molecule cyclic guanosine monophosphate (cGMP) plays an important role in the retinal phototransduction cascade, which is initiated by a conformational change of the photoreceptor opsins after getting hit by the photons of light. Subsequently, transducin levels are elevated, thereby activating the rod photoreceptor–specific cGMP–phosphodiesterase-6 (PDE6), which, in turn, catalyzes the breakdown of cGMP. Decreased cGMP levels lead to closure of cyclic-nucleotide–gated (CNG) cation channels and reduction of Ca²⁺ and Na⁺ influx,¹⁶ and the following hyperpolarization results in signal transmission to second-order neurons.¹⁷ 

Apart from the physiologic importance in phototransduction, PDE and cGMP are likely also involved in pathologic events, since high cGMP levels, and subsequent photoreceptor cell death, are encountered in several animal RP models based on mutations of PDE6, and interestingly also as a consequence of mutations in apparently unrelated genes.^18–21 The PDE6 enzyme is composed of two catalytic subunits called α (A) and β (B) and two identical inhibitory γ subunits.²² Approximately 3% to 4% of arRP cases^21,23 are caused by mutations in the PDE6A gene, coding for the α subunit and resulting in a defective PDE6 enzyme or no enzyme at all. Mutations in the PDE6B gene, coding for the β subunit of PDE6, are likewise frequent and also lead to retinal degenerations.^24–26 The PDE6 mutations, as well as others that lead to increased cGMP levels, may cause an undesired opening of CNG cation channels, allowing continuous and pathologic Ca²⁺ influx with subsequent rod photoreceptor cell death.^27,28 Another possibility is that the increased cGMP levels cause cell death via activation of cGMP-dependent protein kinase.²⁹ Of the PDE6A and PDE6B mutations, the most extensive clinical descriptions so far come from PDE6B families, while descriptions of the phenotype associated with PDE6A mutations are less frequent. 

Another gene that can be associated with arRP is the ABCA4 gene,^30–32 most often encountered in Stargardt disease^33–35 but also seen in other retinal degenerations. The ABCA4 gene codes for the ABCA4 protein, an ATP-binding cassette (ABC) transporter protein located in the rim of the photoreceptor discs.^36–38 ABCA4 promotes the clearance of toxic vitamin A metabolites such as N-retinylidene-phosphatidylethanolamine and phosphatidylethanolamine from the lumen of the outer segment disc membranes during phototransduction.^39,40 Defective function or absence of ABCA4 results in a buildup of the lipofuscin fluorophore N-retinylidene-N-retinylethanolamine, which then accumulates in RPE cells, leading to RPE cell death and probably secondary loss of photoreceptors.^40,41 It has also been proposed that direct photoreceptor cell death may precede RPE alterations.^42,43 

Given the fact that cGMP is an important retinal player in general, and also in RP at least when PDE6A and PDE6B mutations are concerned, clinical measurements of cGMP levels in ocular tissue or in body fluids in various retinal states are few. In the context of retinal detachment, La Heij et al.⁴⁴ have shown reduced cGMP levels in vitreous as well as subretinal fluid in patients with detached retinas, compared to control samples, and similar results have been obtained in experimental retinal detachments in pigs.⁴⁵ The reduced cGMP levels may have been due to lowered production of cGMP in the affected retinal cells.^44,45 Interestingly, with respect to RP, a study by Martinez-Fernandez de la Camara and colleagues¹⁵ reports elevated cGMP blood levels in RP patients. However, the authors do not discuss this in relation with high photoreceptor cGMP, but rather use the blood cGMP as an index of the activity of the NO/cGMP pathway and of oxidative stress as such. Moreover, the report does not describe the RP patients with respect to mutation. 

We thus have a situation where experimental studies suggest a mechanistic cGMP–RP connection, and where Martinez-Fernandez de la Camara et al.¹⁵ have found higher blood cGMP in RP patients; moreover, we know that some patient cohorts indeed carry mutations in PDE6. This warrants further exploration of the possible connections between RP and the cGMP system in compartments outside of the retina, not the least at the clinical level, since the possibility to use blood cGMP as a diagnostic tool for the disease would be most welcome. To this end we here studied retinal function and morphology in a consanguineous family with arRP and identified their genetic errors as homozygote mutations in the PDA6A gene and, in some cases, also heterozygote or homozygote mutations in ABCA4. This permitted us to describe several aspects of the ophthalmologic status of patients with homozygote PDE6A mutations, which has not been done to such an extent previously. In addition, it gave us the opportunity to measure blood cGMP levels and to compare them to those of healthy controls, in a situation where we know that the cGMP system is involved in the disease. 

Methods

Subjects and Controls

In this study, a large consanguineous family of Iraqi origin was investigated. Seven siblings with nonsyndromic arRP and their parents were included (pedigree, Fig. 1). Four of the subjects (II-4–II-7) had also been examined at our Department of Ophthalmology, Skåne University Hospital, Lund, Sweden, at a previous occasion in 2004. One of them, II-4, could not be reexamined for the current study. The mean age of family members was 43 ± 16 years (range, 28–73 years; Table). The study was conducted in accordance with the tenets of the Declaration of Helsinki and it was approved by the Ethical Committee for Medical Research at Lund University. All subjects gave their informed consent to participate. 

Figure 1

 

Pedigree for the family. All family members except for I-2 are homozygous for IVS6+1G>A (c.998+1G>A), a splice site mutation, in the PDE6A gene.

Table

 

Demographic Data on Family Members and a Brief Summary of the Results of Ophthalmologic Examinations and cGMP Measurements

 

Plasma cGMP concentration was also measured in 20 nonsmoking healthy volunteers without any known eye disease or renal problems. Controls were aged 41 ± 6 (range, 30–50) years. 

Genetic Analysis

DNA was extracted from venous blood drawn from the precubital vein in all subjects. The DNA was screened on an arRP genotyping microarray test for 710 mutations in 28 known arRP genes (CERKL, CNGA1, CNGB1, MERTK, PDE6A, PDE6B, PNR, RDH12, RGR, RLBP1, SAG, TULP1, CRB, RPE65, USH2A, USH3A, LRAT, PROML1, PBP3, EYS, ABCA4, AIPL1, CNGA3, CNGB3, GRK1, IMPG2, RHO, and RP1). Testing was performed at Asper Biotech (Asper Ophthalmics, Tartu, Estonia) (http://www.asperbio.com/asper-ophthalmics/autosomal-recessive-retinitis-pigmentosa-genetic-testing/autosomal-recessive-retinitis-pigmentosa-apex-based-test-details; provided in the public domain). Microarray results were verified by Sanger sequencing. 

Ophthalmologic Examination

Slit lamp and fundus examinations were performed in all subjects. Best corrected monocular visual acuity (BCVA) was tested on the Snellen chart at 5 m. Visual fields were mapped by using a Goldmann perimeter with standardized objects V4e and I4e. In one subject (II-5; see pedigree, Fig. 1), the III4e object had to be used instead of I4e at the first visit, since she was not able to identify a smaller object. 

Full-Field Electroretinography

In 2004, full-field electroretinograms (ffERGs) were recorded with a Nicolet Viking analysis system (Nicolet Biomedical Instruments, Madison, WI, USA) as described previously⁴⁶ and in 2014, ffERGs were recorded with an Espion E² analysis system (Diagnosys, Lowell, MA, USA). 

Multifocal Electroretinography

Multifocal ERGs (mfERGs) were registered with a Visual Evoked Response Imaging System (VERIS; EDI, San Mateo, CA, USA) according to previous descriptions.⁴⁶ 

Optical Coherence Tomography

Macular thickness was measured with a Topcon 3D OCT-1000 (Topcon, Inc., Paramus, NJ, USA) as previously described.⁴⁷ 

Cyclic GMP Measurement

All blood sampling was performed at same time of day (noon). The blood was collected in identity-coded heparin tubes and kept on ice until centrifugation at 1750g in a SkySpin CM-6MT centrifuge (ELMI Ltd, Riga, Latvia), to enable the plasma fraction to be collected. The samples were then kept at −80° C until the cGMP measurement. Possible circulating, systemic PDEs were inactivated after the addition of 1/10 volumes of 1 M HCl. If some precipitation occurred, the acidified plasma was centrifuged again, and the supernatant transferred to a clean tube. 

The measurement of the cGMP was carried out by using the cGMP Direct Immunoassay kit (Abcam ab65356; Abcam, Cambridge, MA, USA), following the manufacturer's instructions for the kit, with samples being acetylated before measuring. The measurements were carried out at λ_450nm with an ELISA plate reader (SPECTROStarNano; BMG Labtech, Ortenberg, Germany). The absolute values of the samples were calculated by comparing measured values to a standard curve made up from pure cGMP, as provided in the kit. 

Statistics

Statistical analyses were performed in SPSS 20.0 (IBM Corp., Armonk, NY, USA). Statistical significance was defined as P < 0.05. Owing to the small samples sizes, a nonparametric statistical test (Mann-Whitney U test) was used. 

Results

Ophthalmologic Examination

All siblings had typical RP fundus findings with pale optic discs, narrowed retinal vessels, and bone corpuscle pigmentations in the periphery. In addition, pigmentary changes were found in the macular regions, and II-6 had a macular scar in her left eye. The extent of peripheral pigmentations increased with age as expected. Figure 2 shows fundus photos (Figs. 2a, 2e) from II-3 and II-7, representing one elder and one younger sibling. All subjects except for the youngest, II-7, had some degree of nuclear cataract. II-1 and II-3 had both had cataract surgery and internal limiting membrane peeling in their right eyes in 2010, followed by YAG laser capsulotomy in 2012. These surgical procedures were performed in clinics abroad. II-5, II-6, and II-7 all had slight nuclear lens opacities. 

Figure 2

 

Fundus photos, OCT images, and mfERGs from two representative siblings, one of the older, II-3, and one of the younger, II-7. In both (a, e) the fundus photos show pale optic discs, attenuated retinal vessels, and peripheral bone corpuscle pigmentations, which are more widespread in the older sibling. In II-3 the OCT images show attenuation (b, c) and the mfERGs are severely reduced (d) as compared to normal (i). In II-7 the OCT images reveal central retinal thickening (f) due to macular edema (g), and macular function measured with mfERGs is severely reduced (h) as compared to normal (i).

Best corrected monocular visual acuity ranged from amaurosis in the eldest family member (I-1) to 0.8 Snellen in the youngest (II-7) (Table). In the subjects that had been examined in 2004 (II-4–II-7), a deterioration of BVCA was found (Table). 

Visual fields were constricted to approximately 15° to 10° or in some cases even less, for the V4e object in all siblings (Table). The younger ones (II-4–II-7) also had temporal crescent-shaped remnants of varying extent left (see also Fig. 3 for representative visual fields from siblings II-3 and II-7; Table). In two of the subjects, who were also examined in 2004 (II-5 and II-7), the visual fields showed approximately 10° of further constriction for the V4e object on the last examination compared to the first (V4e approximately 20°–25°). In subject II-6, the central parts of the visual fields were stable between the two investigations, but on the last one, a temporal crescent not previously shown was identified. 

Figure 3

 

Representative Goldmann visual fields from one elder sibling, II-3 (top), and one younger, II-7 (bottom). II-3 has small residual visual fields with V4:e constricted to less than 10° and I4:e to less than 5°. II-7 has somewhat larger central remainders of the visual fields as well as spared temporal crescents for V4:e.

Full-Field Electroretinography

Full-field ERGs were recorded in six of the siblings (II-1, II-3–II-7). All of them showed extinguished rod responses, while they still had minimal residual cone responses measuring approximately 1 μV in amplitude (Table). The same results were found for II-4 to II-7 when they were examined in 2004. 

Multifocal Electroretinography

Multifocal ERGs were obtained in five of the siblings (II-1, II-3, II-5–II-7) and revealed severely reduced cone responses in all rings of the mERG (Table). Data from II-3 and II-7 are shown in Figure 2. 

Optical Coherence Tomography

Optical coherence tomography scans were captured in II-1 to II-3 and II-5 to II-7. The OCT scans showed either severe general attenuation in the posterior pole (II-3, II-6 and II-7 left eye) or central macular edema accompanied with peripheral attenuation (II-1 and II-2, II-5, II-7 right eye). Examples of OCT scans for II-3 and II-7 are shown in Figure 2. 

Genotype

All siblings (II-1–II-7) and their father (I-1) were homozygous for IVS6+1G>A (c.998+1G>A), a splice site mutation, in the PDE6A gene. The mother (I-2) was heterozygous for the same mutation (Fig. 1). Moreover, subjects II-1 and II-3 were homozygous for the c.1532G>A variant in the ABCA4 gene and subjects I-1, I-2, II-4, II-5, and II-7 (Table) were heterozygous for the same mutation, which is a missense mutation leading to a substitution of His for Arg at position 511. It is quite uncommon but it has once, in the compound heterozygote state, been described to lead to a Stargardt phenotype.⁴⁸ Although ABCA4 mutations are known to cause arRP,^{30–32,49–51} in this setting the PDE6A mutation seemed to be the causative mutation considering the fact that all family members with the homozygous IVS6+1G>A (c.998+1G>A) PDE6A mutations had the same severe retinal dysfunction, no matter if they were homozygous, heterozygous, or lacked the ABCA4 mutation (c.1532G>A) (Table). Nor did there seem to be any association with the presence of ABCA4 mutations and macular morphology or function, although ABCA4 mutations most often are related to Stargardt disease.^{33–36,52–55} Macular edema was encountered in a subject without the c.1532G>A ABCA4 mutation (II-2) as well as in family members who were heterozygous (II-5 and II-7) or homozygous (II-1) for the c.1532G>A mutation. The same lack of association to the ABCA4 mutations was shown for macular attenuation with no ABCA4 mutation in subject II-6, a heterozygous c.1532G>A mutation in subject II-7, and homozygous c.1532G>A mutations in subject II-3. 

Cyclic GMP Measurement

Plasma cGMP concentration was measured in six of the family members (I-1, II-1, II-3, II-5–II-7). The mean cGMP level was 166 ± 19 nM (mean ± SD) and the range, 146 to 202 nM. Cyclic GMP in the eldest subject, the father I-1, was 150 nM (Fig. 4a; Table). 

Figure 4

 

(a) Plasma cGMP concentration for controls (n = 20) and RP patients (disease) (n = 6). Boxes show median and interquartile range, while bars illustrate range and circles indicate outliers. (b) The ROC curve (i.e., the blue curve in the figure) shows the sensitivity in relation to the reciprocal of the specificity (i.e., 1-the specificity, sometimes denoted as the false-positive rate). The green diagonal line is a reference line representing a situation where no prediction at all can be made from the presumptive indicator. This ROC curve indicates that the cGMP test has good sensitivity, that is, it can find diseased patients very well, but with a bit lower specificity, that is, carrying a risk that some nondiseased subjects may appear as having the disease (false positive).

Mean plasma cGMP in control subjects was 83 ± 26 (range, 40–157) nM. When analyzed for statistical significance, the values from the PDE6A family members were found to be significantly higher than those of the control group (P = 0.001). 

A receiver operating characteristic (ROC) curve was also established, which revealed that the cGMP test has a good sensitivity, but a somewhat lower specificity (Fig. 4b). 

In the literature, normal plasma/serum cGMP levels, obtained by various methods, have often been reported in the lower nanomolar range, although values up to several hundred nanomolar have also been given.^56–63 Our control cGMP plasma levels at approximately 80 nM therefore compare relatively well with this. 

Discussion

We have studied a consanguineous family with a severe form of arRP and report several ophthalmologic data consistent with extinguished rod responses and markedly reduced cone function by the early twenties. We also present information on a possible disease-related deviation from normal plasma cGMP levels in these patients. 

To the best of our knowledge, this is the first phenotypic description of arRP due to homozygous IVS6+1G>A (c.998+1G>A) mutations in the PDE6A gene and includes the somewhat unique situation where all family members (except the mother) show such homozygosity as well as a rather uniform phenotype. The IVS6+1G>A (c.998+1G>A) mutation has been confirmed as a novel splice site mutation in patients with arRP²³ and it is considered to be a null allele, a fact that may explain the severe phenotype in our subjects, although PDE6A-associated arRP sometimes has been described as a milder form of arRP.^24,26 All of the subjects, regardless of age at examination, also had severely reduced macular function as measured by mfERG, and OCT analyses showed either macular attenuation or edema with, or without, vitreoretinal traction. Macular edema and vitreoretinal traction have been described in arRP patients with other mutations in the PDE6A gene^24,26 and may contribute to the early reduction of central visual function. For example, the youngest sibling in our study had reduced visual acuity (0.25) in her right eye with macular edema, compared to her left eye showing macular attenuation but still quite good visual acuity (0.8). The phenotype was also characterized by presenile nuclear cataract in all but one sibling (II-7, the youngest). 

An aspect that would potentially complicate the interpretation of genotype–phenotype correlation in this family was the additional presence of ABCA4 mutations (c.1532G>A), either heterozygously (I-1, I-2, II-4, II-5, and II-7) or homozygously (II-1 and II-3), in some family members (Table). In this setting, though, the PDE6A mutation seems to be causative, since all family members with the homozygous IVS6+1G>A (c.998+1G>A) PDE6A mutations had the same severe retinal dysfunction, whether or not they carry the ABCA4 mutation (c.1532G>A) (Table). A similar situation with a well-known rhodopsin mutation in all family members and an extra mutation in the gene for cGMP-gated channels in some siblings has been described by Dryja et al.,⁶⁴ who concluded that the shared rhodopsin mutation had to be the cause of retinal degeneration, while the other mutation was carried only by chance and without apparent effects. 

The difference in plasma cGMP levels between RP patients and controls (RP twice that in controls) is very clear, and the corresponding ROC curve suggests that the cGMP test readily finds diseased patients. Still, at this point we cannot claim that the higher level is a direct consequence of the disease. Differential blood cGMP levels have previously been seen in a variety of situations, for instance, in connection with natriuretic drugs for acute heart failure,⁵⁸ blood pressure–altering drugs,⁵⁹ preeclampsia,⁶⁰ migraine,⁶¹ and in relation to the circadian rhythm,⁶³ where the cGMP levels may in turn be dependent on melatonin levels.⁶² With a possible connection to this, the ROC specificity aspect indicates that some healthy subjects may appear as having the disease. Yet, we find it compelling that subjects with a mutation in PDE6A, a gene coding for a cGMP-hydrolyzing enzyme highly expressed in the retinal photoreceptors,⁶⁵ present with high levels of the nonhydrolyzed substrate as well as with reduced photoreceptor function. We are currently analyzing whether retinas from animal RP models do release their high cGMP into culture medium, and note that this seems to be the case (unpublished results, but see Lolley et al.⁶⁶). It is hence possible that the lack of cGMP hydrolysis results in increased extracellular cGMP that manages to find its way to the circulation. One may argue that patients with a loss of measurable retinal functions as in this case should not have any photoreceptors left to produce cGMP. However, results from a PDE6A mutant canine model suggests that ERGs may be severely affected well in advance to structural loss of photoreceptors,⁶⁷ and the present clinical finding does not as such exclude the presence of rod photoreceptors, albeit in reduced numbers. Moreover, our results agree with those of Martinez-Fernandez de la Camara et al.¹⁵ where the serum cGMP levels of a group of RP patients are increased by approximately 65% compared with a control group.¹⁵ In the latter study, the RP mutations in question are not reported, however, and at least some may thus be unrelated to PDE6 gene mutations as such, which would appear inconsistent with our conclusions. On the other hand, as mentioned in the Introduction, several RP models with mutations in genes, including but not only PDE6, display high photoreceptor cGMP, suggesting this is a common denominator for a number of RP types,^18,21,68 although at this point we do not know the underlying mechanisms. Since one may expect high similarity between photoreceptor pathology of human homologue models and that of RP patients, high cGMP in photoreceptors, and possibly in the circulation, might therefore appear also in larger patient groups, independently of the exact mutation. 

It appears unlikely that a plasma cGMP test on its own would be used to screen parts of the general population with respect to detecting RP, not the least because there could be a considerable risk for false positives. The latter may occur because of a cGMP connection to other clinical or nonclinical situations, as mentioned above, and in this context there are reports suggesting PDE6 expression also in nonphotoreceptor tissues and in different developmental stages,⁶⁹ at least in other animals, which may have a bearing on the plasma cGMP levels. Still, the measurement of plasma cGMP is relatively simple and could readily act as a complement to other ophthalmic investigations in order to improve the RP diagnosis, which is not always straightforward. Such cGMP analyses could be envisaged to take place in connection with regular eye examinations of defined or suspected RP patients, including family members. There are of course a number of questions remaining before this can be reached, the most important being whether the blood cGMP–RP connection is real. We therefore plan repeated analysis of cGMP in more RP patients with other causative mutations. 

In conclusion, this is, to our knowledge, the first study that describes the phenotype in arRP due to the IVS6+1G>A (c.998+1G>A) mutation in the PDE6A gene revealing a severe form of RP with early extinguished rod responses and seriously reduced macular dysfunction also at an early stage. Intriguingly, this was associated with elevated plasma cGMP, which may have a direct connection to the disease. 

Acknowledgments

The authors thank Ing-Marie Holst and Boel Nilsson for skillful technical assistance. They also thank Håkan Lövkvist, PhD, biostatistician, for valuable statistical advice. 

Supported by the Medical Faculty, Lund University, and grants from Skåne County Council Research and Development Foundation; The Armec Lindbergs Stiftelse; Stiftelsen för synskadade i f.d. Malmöhus län; Stiftelsen Synfrämjandets Forskningsfond/Ögonfonden; The Swedish Society of Medicine; Skane University Hospital foundations and donations, KMA; and Stiftelsen Olle Engkvist Byggmästare. 

Disclosure: U. Kjellström, None; P. Veiga-Crespo, None; S. Andréasson, None; P. Ekström, None