Oligocone Trichromacy: Clinical and Molecular Genetic Investigations

Abstract

Purpose.: To describe the phenotype and genotype of patients with a diagnosis of oligocone trichromacy (OT).

Methods.: Six unrelated patients had a detailed ophthalmic examination including color vision testing, a Goldmann visual field test, fundus photography, and full-field electroretinography (ffERG). Five patients also underwent multifocal (mf)ERG, autofluorescence recording, and optical coherence tomography (OCT). Genetic analysis included sequencing of all coding regions and flanking introns of CNGA3, CNGB3, GNAT2, KCNV2, and PDE6C.

Results.: All patients had subnormal visual acuity, a history of congenital nystagmus, and subjectively normal or near-normal color vision; five patients reported photophobia. Clinical examinations revealed largely normal fundi, normal Goldmann visual field results with the IV/4e target, and normal color discrimination or mild color vision deficiency. Electrophysiological investigations showed either complete absence of recordable cone responses or severely reduced amplitudes. All retinal layers were identifiable by OCT, which also showed thinning of the peripheral retina. Genetic analysis revealed two causative CNGB3 mutations in one patient and single heterozygous mutations of unknown significance in CNGB3 and PDE6C in two other patients.

Conclusions.: Oligocone trichromacy is a heterogeneous condition with respect to both phenotypic appearance and genetic background. The finding of mutations in genes known to be involved in complete and incomplete achromatopsia supports the notion that some forms of OT is an extreme form of incomplete achromatopsia with preferential loss of peripheral cones.

Oligocone trichromacy (OT) is a rare form of congenital cone dysfunction characterized by a paradoxical combination of severe electrophysiological cone impairment and normal or near-normal color discrimination. OT was delineated as a nosologic entity by van Lith, ¹ who reported on a patient with congenital nystagmus and low vision. The condition was further characterized by a severely reduced photopic electroretinogram (ERG) response and the startling finding of almost normal color vision. Further examinations revealed normal peripheral visual fields, elevated cone thresholds, lowered flicker fusion frequency, normal or moderately reduced rod ERG, and normal electrooculogram (EOG). Based on these observations van Lith hypothesized a very low number of normal functioning cones, explaining both the low ERG cone response and the near normal color vision. Keunen et al. ² corroborated this hypothesis by measuring foveal cone photopigment density in four individuals with OT and found a reduced density difference but a normal photopigment regeneration time constant. 

Taxonomically OT is classified as a cone dysfunction, a very heterogeneous group of disorders including total and incomplete achromatopsia (rod monochromatism), blue cone monochromatism, Bornholm eye disease, and cone monochromatism. (For a recent review, see Michaelides et al. ³ and references therein.) 

In this report, we present multimodal clinical findings and the results of molecular genetic investigations in six patients with a clinical diagnosis of OT. 

Methods

All patients receiving a diagnosis of OT at the National Eye Clinic from 1993 to 2001 were invited to participate in the study. Inclusion criteria were a history of congenital nystagmus, subnormal visual acuity (0.9–0.1), subjectively normal or near-normal color vision, and absent or severely reduced ERG cone responses. Six unrelated patients volunteered and gave their informed consent in compliance with the tenets of The Declaration of Helsinki. Most of the patients reported having photophobia since childhood. Five of the six had been referred as children with congenital nystagmus and reduced visual acuity. Patient B, with near-normal visual acuity, and minimal nystagmus was referred at age 40. 

The study protocol included refraction and assessment of best-corrected visual acuity, Goldman manual kinetic perimetry (I/4e and IV/4e targets), slit lamp biomicroscopy with fundus examination, evaluation of nystagmus, ERG, color vision testing, optical coherence tomography (OCT), fundus autofluorescence imaging, and color fundus photography. Patient F was unavailable for some of the procedures. 

Color vision testing comprised Farnsworth-Munsell's 100-hue test (FM-100), Farnsworth-Munsell's D-15 (FM D-15, saturated), Lanthony's tritan album (LTA), and Ishihara's 38 pseudoisochromatic plates (2001 edition). Two patients were also examined with Nagel's anomaloscope. All color vision tests were performed under a ceiling-mounted lighting panel with four fluorescent tubes (36W/72 Biolux; Osram GmbH, Munich, Germany), color temperature 6500 K. 

Full-field electroretinography (ffERG) was performed in one eye in each of six patients after pupil dilation with cyclopentolate 1% and 30 minutes of dark adaptation according to the standard protocol of the International Society for Clinical Electrophysiology of Vision (ISCEV). ^4,5 Briefly, the recordings where made with a Burian-Allen monopolar electrode after local anesthesia with oxybuprocaine 0.4%, a reference ear clip electrode on the opposite earlobe, and a ground electroencephalography silver cup electrode on the forehead. Stimulations were produced with a xenon flash tube with adjustable intensity mounted in a Ganzfeld globe (Nicolet Biomedical Instruments, Madison, WI). The flash intensity was 0.5 U in all recordings except photopic red, green (both 1.25 U), and blue single flash (1.00 U). Flash duration was 20 to 30 μs. Flashes were attenuated with neutral-density filters. Data processing was performed on workstation (Viking IV; Nicolet Biomedical Instruments). Reference values were obtained from a control group made up of 85 healthy adult volunteers. 

Multifocal electroretinography (mfERG; VERIS Science ver. 5.0; Electro Diagnostic Imaging, Inc., San Mateo, CA) was recorded in five patients, with stimulus 103 hexagons at full contrast. A camera attachment was used for fixation, allowing eccentricity to be scaled. The stimulation pattern extended to 20° eccentricity. Data were processed with the manufacturer's standard software. 

Macular thickness was measured by OCT in both eyes of five patients after pupil dilation (Stratus OCT model 3000, software ver. 4.0.1, and Spectral Domain Cirrus OCT; Carl Zeiss Meditec, Humphrey Division, Dublin, CA). A combination of fast and detailed protocols were used to minimize the effect of nystagmus while obtaining high-resolution tomograms for evaluation of retinal structure, and integral software was used to calculate retinal thickness and macular volume. In addition to reference data supplied by the manufacturer, we compared the patients with 17 healthy adults and 1 child plus published data from children. ⁶ The OCT operator controlled centration on the fovea during the recording. Fixation errors deemed to be larger than 500 μm led to rejection of the scan. Morphologic evaluation was made on the best-centered scans with the thinnest fovea and, if possible, with a visible foveal reflex. 

Fundus autofluorescence photography was performed in five patients after pupil dilation, with a confocal scanning laser ophthalmoscope (Heidelberg Retina Angiograph; Heidelberg Engineering, Heidelberg, Germany) set to align and average at least 16 frames. 

Venous blood was collected from the patients and available family members after informed consent and total genomic DNA was extracted according to standard procedures. The samples were analyzed for mutations in the following genes: CNGA3, CNGB3, GNAT2, KCNV2, and PDE6C ^5,7–12 (SK, BW, unpublished data, 2009; details available on request). Briefly, mutation screening was performed by PCR amplification from genomic DNA and sequencing of the coding exons and flanking intron and UTR sequences. In addition, common alterations of the red/green opsin OPN1MW/OPN1LW gene cluster typically found in blue cone monochromacy were excluded by means of PCR and PCR/RFLP assays in male patients (data not shown). Segregation analysis for the presence and independent inheritance of two mutant alleles was performed for the mutations c.886_896del11insT and c.1148delC in CNGB3 in both parents of patient D by PCR/RFLP analysis and direct sequencing. 

Results

Ophthalmic Examination

All patients had a history of congenital nystagmus and five patients reported photophobia. Ophthalmic examination revealed an unremarkable anterior segment in all six patients, disc drusen in patient B, and temporal peripapillary crescents and unilateral epiretinal fibrosis in patient A. Goldmann perimetry (IV/4e and I/4e) demonstrated no scotomata or constrictions except that with the smaller target (I/4e) in patient D, in whom concentric constriction to 15° to 20° was seen, and in patient B, who was unable to see this target (Table 1). Five patients with a follow-up of eight or more years showed no sign of change in best corrected visual acuity, visual fields, or fundus morphology. All patients had subjectively normal or near-normal color vision. 

Table 1.

 

Summary of Clinical Findings

Patient	Sex, Age at Time of Follow-up	Follow-up Period (y)	BCVA (Right/left)	Refraction (Right/left)	Nystagmus	Photophobia	Goldmann Perimetry
A	M, 46	32	0.4/0.5	+1.00/+0.50	++	No	Normal boundaries for target IV/4e and I/4e, no scotoma
B	F, 47	8	0.6/0.9	+3.00/+2.75	+	Yes	Normal boundaries for target I/V4e, no scotoma, cannot see target I/4e
C	M, 28	26	0.2/0.2	0/0	+++	Yes	Normal boundaries for target IV/4e and I/4e, no scotoma
D	F, 12	12	0.1/0.2	−2.00/−4.25	+	Yes	Normal boundaries for target IV/4e, no scotoma; constricted field with I/4e
E	F, 24	8	<0.6/0.5	+1.25/+0.75	None (history of congenital nystagmus)	Yes	Normal boundaries for target IV/4e and I/4e, no scotoma
F	M, 8	None	0.3/0.3	+1.50/+1.00	++	Yes	Normal boundaries for target IV/4e, no scotoma

Color vision testing was normal for the Lanthony tritan album (n = 5) and Farnsworth-Munsell's D-15 standard test (n = 6; Table 2). Ishihara was read without errors by two patients, and four others had 1, 3, 5, and 8 errors, respectively (C, E, D, and B). Farnsworth-Munsell's 100-hue test showed some nonuniformity, with error scores ranging from a normal score of 36 to a moderately abnormal score of 236 with no clear axis other than a probable tritan axis in patient D. 

Table 2.

 

Summary of Results of Color Vision Evaluation

Patient	F-M 100	F-M D-15	LTA	Ishihara 38-pl.	Nagel Anomaloscope
A	Error score 36, normal (normal value for age: 100)	No errors	No errors	No errors	Normal (41–41/14)
B	Error score 131, diffuse discrimination error (normal value for age: 100)	No errors	No errors	8 errors	—
C	Error score 124, diffuse discrimination error (normal value for age: 78)	No errors	No errors	One error	—
D	Error score 236, diffuse discrimination error, but probable tritan axis (normal value for age: 180)	Few neighbor transpositions	No errors	5 errors	—
E	Error score 67, normal (normal value for age: 78)	No errors	One error	3 errors	Normal with slightly extended setting interval (38–43/14)
F	—	Few neighbor transpositions	—	No errors	—

 —, Unavailable data.

A recordable ffERG cone response, yet with amplitudes below those of the 5th percentile of healthy subjects and normal implicit times was found in patient A whose mfERG responses were barely detectable (Fig. 1, Table 3). Cone responses of lower amplitude but still detectable were found in patients E and F. Cone responses were near the detection threshold in the remaining three patients. Subnormal rod amplitudes were found in patients B, C, E, and F. Nearly absent mfERG cone responses were found in three patients, whereas mfERG failed in patient C and was unavailable in two patients. 

Figure 1.

 

Multifocal electroretinogram with minimal cone responses from patient A. Patients B and E, with a degree of nystagmus that permitted examination, had comparable characteristics.

Table 3.

 

Amplitude and Implicit Time of Responses on ffERG

Patient (Normal Values)	A	B	C	D	E	F
Scotopic
[Median amplitude (μV) (5–95th percentile)/Implicit time (ms) (5–95th percentile)]
Rod response
b-wave [210 (124–302)/103 (81–123)]	162/112	74.5/86.5	89.9/101	204/102	100/132	80.7/142
Rod-cone response
a-wave [205 (111–312)/23.0 (21.0–25.0)]	115/24.0	121/24.0	65.9/22.5	173/23.5	59.7/23.0	115/29.6
b-wave [427 (285–674)/48.0 (41.0–55.0)]	242/45.0	196/60.5	224/51.0	385/57.0	260/59.5	193.1/53.6
Oscillatory potential (OP2 amplitude) [77.0 (29.0–125.0)/25.2 (24.0–27.2)]	27.3/26.4	24.4/34.2	22.2/26.8	36.0/27.0	38.6/31.8	31.2/37.6
Flicker response [94 (53–145)/29.0 (27.0–34.0)]	42.0/27.0	0	0	0	0	37.4/39.2
Photopic
Single white flash response [136 (61–154)/29.0 (27.0–32.0)]	54.0/29.5	0	0	0	4.56/42.0	No data
Flicker response [101 (61–154)/26.0 (25.0–30.0)]	51.7/27.0	0	0	0	0	No data
Long-wave (red) single flash response [105.2 (64.5–172.9)/29.2 (27.0–32.0)]	51.2/29.8	0	0	0	0	38.7/30.4
Middle-wave (green) single flash response [114.6 (62–181)/28.0 (26.0–32.2)]	52.7/28.8	0	0	0	0	35.0/30.0
Short-wave (blue) single flash response [4.4 (0.9–8.9)/43.6 (39.4–51.0)]	0	0	0	0	0	0

 0, Absence of a recordable response.

Optical coherence tomography (n = 4) demonstrated a relative distribution of reflectivity among the retinal layers that was indistinguishable from that in healthy subjects, whereas the overall level of reflectivity was lower than in healthy subjects (Fig. 2). The thickness of the neurosensory retina was normal in the fovea but 6.0% thinner than in healthy subjects in the parafoveal region (P = 0.0003, Table 4) and 5.8% thinner in the perifoveal region (P = 0.0056). The foveal contour was mildly flattened (Fig. 2). One patient was unavailable for OCT examination, and in patient C, nystagmus prevented examination. 

Figure 2.

 

Transfoveal optical coherence tomograms of the retina from a healthy subject (a, aged 11 years) and from two patients with OT (b, patient D; c, patient B). Patients with OT with recordable tomograms were those who had the least nystagmus and these patients all demonstrated diffusely subnormal retinal reflectivity. NFL, nerve fiber layer; GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL&HL, outer nuclear layer and Henle's layer; ELM, external limiting membrane; IS, inner segment of photoreceptors; IS/OS, inner and outer segment junction of photoreceptors; OS, outer segments of photoreceptors; RPE, retinal pigment epithelium; OS/RPE, outer segment of photoreceptors junction with the RPE.

Table 4.

 

OCT Retinal Thickness of the Right Eye

	Macular Volume (mm3)	Foveal Thickness (μm)	Inner Peripheral Thickness (μm)*	Outer Peripheral Thickness (μm)*	Comments
Patient A	6.6	254	260.3	224.5	Slight epiretinal fibrosis
Patient B	6.7	179	244.5	239.3
Patient C	6.3	226	251.8	212.0	Age 12
Patient E	6.7	199	268.5	225.5
Mean value for patients	6.6	214.5	256.3	225.3
Normal value (SD)†	6.7 (1.13)	209.4 (55.15)	272.4 (31.25)	239.1 (45.39)
P-value	0.0122	0.5532	0.0003	0.0056

*  Mean retinal thickness by optical coherence tomography in relation to eccentricity (mean of temporal, superior, nasal, and inferior sectors in the patients' right eyes).

†  Mean of 17 healthy adults.

Autofluorescence fundus images were normal in patients B, C, and E, with little enough nystagmus and photophobia to enable examination. 

Molecular Genetics

We analyzed the genes CNGA3, CNGB3, GNAT2, KCNV2, and PDE6C, which are all known to be involved in autosomal recessive achromatopsia and cone dysfunction disorders, as well as common alterations in the OPN1MW/OPN1LW gene cluster in male patients, typically found in blue cone monochromacy (Table 5). Patient D was a compound heterozygote with a deletion/insertion mutation c.886_896del11insT on the paternal allele and the 1-bp deletion c.1148delC on the maternal allele in the CNGB3 gene (Table 5). The mutation c.886_896del11insT induces a frameshift downstream of Arg296 including a tail of eight novel amino acids followed by a stop codon (p.Thr296TyrfsX9). The mutation c.1148delC results in a frameshift downstream of Thr383 and generates a premature stop codon after 12 altered amino acid residues (p.Thr383IlefsX13). Patients A and B were shown to carry single heterozygous missense mutations. Patient A carried the mutation c.1208G>A p.Arg403Gln in CNGB3 and patient B c.1755G>T p.Lys585Asn in PDE6C, respectively. No other mutation could be identified in the coding exons and flanking intronic sequences of the respective genes in these two patients. The relevance of the mutations to the disease in these patients remains unclear. The single mutation c.1208G>A p.Arg403Gln in CNGB3 is known to be associated with autosomal recessive achromatopsia, but also autosomal recessive cone dystrophy and autosomal recessive macular dystrophy. ¹³ The single mutation c.1755G>T p.Lys585Asn in PDE6C is to date unique to patient B. Mutations in PDE6C are a rare cause of autosomal recessive achromatopsia (SK, BW, unpublished data, 2009). The genetic analysis of the OPN1MW/OPN1LW gene cluster on the X-chromosome in the male patients A, C, and F confirmed its structural integrity, thereby excluding genotypes typically associated with blue cone monochromatism. The molecular genetic analysis in patients C, E, and F did not result in the identification of any mutation or putative pathogenic sequence variants in the analyzed genes CNGA3, CNGB3, GNAT2, KCNV2, and PDE6C. 

Table 5.

 

Genetic Analysis and Results

Patient	CNGA3	CNGB3	GNAT2	KCNV2	PDE6C	BCM
A	Excluded	c.1208G>A p.Arg403Gln heterozygous	Excluded	Excluded	Excluded	Normal OPN1MW/OPN1LW gene cluster
B	Excluded	Excluded	Excluded	Excluded	c.1755G>T p.Lys585Asn heterozygous	NI (female)
C	Excluded	Excluded	Excluded	Excluded	Excluded	Normal OPN1MW/OPN1LW gene cluster, but p.Cys203Arg in a distal gene*
D	NI	1. Allele: c.886_896del11insT p.Arg296TyrfsX8/ 2. Allele: c.1148delC p.Thr383IlefsX12	NI	NI	NI	NI (female)
E	Excluded	Excluded	Excluded	Excluded	Excluded	NI (female)
F	Excluded	Excluded	Excluded	Excluded	Excluded	Normal OPN1MW/OPN1LW gene cluster

 NI, not investigated.

*  The OPN1MW/OPN1LW gene cluster on the X-chromosome is typically composed of one proximal OPN1MW (M, middle-wave-sensitive opsin), and one or more distal OPN1LW (L, long-wave-sensitive opsin) genes. The analysis assesses the structural integrity of this gene cluster. In addition, a common functionally inactivating mutation c.607T>C p.Cys203Arg is tested. Patient C was shown to carry an intact OPN1MW/OPN1LW cluster with a wild-type OPN1MW gene, as determined by OPN1MW-specific long-distance PCR and sequencing, and at least one OPN1LW gene. One of the OPN1LW genes was shown to carry the mutation c.607T>C p.Cys203Arg. This genotype is not suggestive of a clinical diagnosis of blue cone monochromacy.

Discussion

OT (OT) is a rare cone dysfunction with less than 30 cases published so far. The six patients described herein and one described in a previous report ⁵ constitute all cases found at our national low vision center within a limited number of birth cohorts in a population of approximately 6.0 million, leading to a rough prevalence estimate of 1 in 500,000. We agree with other investigators ^1,14 that the condition may be underdiagnosed, as only ERG examinations in individuals with unexplained nystagmus and/or reduced visual acuity will reveal the condition. 

OT is characterized by the triad of low vision, severely reduced ERG cone activity, and normal color vision. Concomitant signs may include nystagmus and photophobia. Complete and incomplete achromatopsia and blue cone monochromacy differ from OT by the severity of color vision impairment (e.g., Nagel anomaloscope identity over a wide range of green-red mixtures with a steep linear sequence). However, it should be noted that the clinical characteristics of OT also may be present in two cone–rod dysfunctional entities, cone dystrophy with supernormal rod response ¹⁵ and Åland eye disease, better known as incomplete congenital essential night blindness. ^16,17 Both conditions are characterized by distinct ERG characteristics that may escape attention: supernormal rod responses, and a reduced b/a-wave proportion of the scotopic mixed cone–rod response with cone flicker abnormalities. A distinguishing feature of OT in relation to cone dystrophies and cone–rod dystrophies is its appearance in early infancy and the lack of progression. Our patients showed some clinical heterogeneity, especially regarding the results of color vision tests and electroretinography. Mild color vision deficiency has been reported with the OT phenotype ^1,5,18 as has variable degrees of residual cone function measured by ERG. ^1,14 Patients B, C, E, and F had subnormal rod specific responses. Similar changes have been reported in OT by others ¹⁴ and also in achromatopsia due to CNGA3 and CNGB3 mutations. ^2,13,19  

Conceptually, OT has been regarded as an extreme variant of incomplete achromatopsia. ²⁰ However, until the molecular basis and the pathophysiology of the condition is better understood, a classification within a hierarchical system remains speculative. 

Based on clinical observations, Van Lith ¹ hypothesized that the condition may be due to a reduced number of normal functioning cones. This view was corroborated by the experiments of Keunen et al. ² who, by foveal densitometry, demonstrated a reduced density difference in the foveal cone photopigments in vivo and a normal time constant of photopigment regeneration. So far, no pathologic report exists, and it is unknown whether the reduced cone responses are due to a reduced number of cones or whether the majority of cones are still present but morphologically and/or functionally silent. The grossly normal structure of the outer retina on transfoveal OCT in this study does not suggest that a significant number of cones are missing in OT. This observation is in contrast to quantitative analysis of OCT characteristics in patients with achromatopsia and blue cone monochromacy. These analyses show absent IS/OS reflectivity in achromatopsia and a reduced reflectivity with a reduced distance from the retinal pigment epithelium in blue cone monochromatism, indicating shortened outer segments of the remaining photoreceptors. ²¹ The significance of our finding that the perifoveal retina is abnormally thin in OT is not known. We suggest that it may be caused by abnormal packing of the photoreceptors, ganglion cells, or some other component of the retina during its development. Others have proposed a hypothetical entity of “peripheral cone disease.” ²²  

The molecular basis of OT is largely unknown. Compound heterozygous mutations in GNAT2 were identified in a male patient with nystagmus and OT. ⁵ One of the mutations, a deep intron variation c.461+24G>A was shown to cause a splicing defect that introduces a premature termination codon. Yet this mutation was “leaky,” so that small amounts of correctly spliced GNAT2 transcripts are formed and can eventually yield some functional protein. Moreover, compound heterozygous CNGA3 mutations (p.Thr224Arg and p.Thr369Ser) were described in two siblings with a mild form of incomplete achromatopsia with moderate color vision deficiency. ²³ Functional in vitro expression experiments of p.Thr224Arg mutant channels showed that it was a functional null mutation, whereas p.Thr369Ser mutant channels were shown to form functional channels, albeit with some altered physiological properties, that were partially compensated by coexpression with CNGB3. ²³  

The compound heterozygous CNGB3 mutations in patient D involve two frameshift mutations that are expected to give rise to truncated proteins. ¹¹ Both mutations have been recurrently observed in patients with (complete) achromatopsia. It remains unclear why patient D expresses an OT phenotype, although she functionally performed the worst in the whole patient series. Patients A and B were shown to carry single heterozygous missense mutations. The relevance of the mutations for the disease in these patients remains unresolved because of the absence of a second mutant allele. Patient A carried the mutation c.1208G>A p.Arg403Gln in CNGB3, which in the homozygous state causes autosomal recessive progressive cone dystrophy. ¹⁴ Patient B carried a single heterozygous mutation c.1755G>T p.Lys585Asn in PDE6C, located in the catalytic domain of the PDE6C polypeptide between the two metal binding motifs, and affecting an amino acid residue conserved between PDE6C proteins of various species. 

Mutations in GNAT2, CNGA3, and CNGB3 are primarily associated with autosomal recessive achromatopsia. ^7,9–11 Although most of these patients were clinically diagnosed with complete achromatopsia, some patients showed residual cone function, either electrophysically or psychophysically. There is evidence that certain mutations or combinations of mutations account for residual function. ^19,23 Indirectly, the involvement of these genes in OT supports the notion that this disorder is an extreme variant of incomplete achromatopsia. The large proportion of cases with an essentially unknown molecular genetic background, however, indicates further genetic heterogeneity, possibly involving different genetic mechanisms. 

In conclusion, this study confirmed the clinical findings overall comparable to that found by other investigators. ^{1,2,14,18,20,24} In addition, nearly absent mfERG responses, grossly normal structure of the outer retina, and reduced perifoveal thickness on transfoveal OCT was found. Causative mutations have so far been identified in only two patients with OT. ⁵ The negative and inconclusive outcome of the mutational analyses in most of our patients gives indirect evidence of still unknown genes or genetic mechanisms being involved in OT. 

Footnotes

 Supported by a scholarship from the Danish Agency for Science Technology and Innovation (MKGA). The molecular genetic analysis was supported by German Research Council (DFG) Grant KFO134-Ko2176/1.

Footnotes

 Disclosure: M.K.G. Andersen, None; N.L.B. Christoffersen, None; B. Sander, None; C. Edmund, None; M. Larsen, None; T. Grau, None; B. Wissinger, None; S. Kohl, None; T. Rosenberg, None

The authors thank the patients and family members for participating and Klaus Kallenbach and Cecilia Rönnbäck for kindly providing the OCT results from normal subjects.