CRB1 Heterozygotes with Regional Retinal Dysfunction: Implications for Genetic Testing of Leber Congenital Amaurosis

Abstract

purpose. To test human CRB1 heterozygotes for possible clinical or functional retinal changes and to evaluate whether a patient with Leber congenital amaurosis (LCA) with CRB1 mutations not consistent with previously described CRB1 phenotypes carried a modifier allele in another LCA gene.

methods. Seven unrelated heterozygous carriers of CRB1 mutations underwent phenotyping by full eye examinations (indirect ophthalmoscopy and slit lamp biomicroscopy) and functional testing (standard full-field electroretinography [ERG] and multifocal ERG). For genotyping of the LCA patients and their parents, denaturing high-performance liquid chromatography (dHPLC) analyses were performed, followed by sequence analysis of CRB1, followed by sequence analysis of the AIPL1 and CRX genes to identify a putative modifier effect in a patient with an atypical CRB1 phenotype.

results. Reduced full-field ERG b-wave amplitudes were observed with scotopic –2 dB flash (140 μV; P < 0.05), normal full-field cone ERGs, and significant regional retinal dysfunction on mfERG in five of seven carriers of CRB1 mutations. A known AIPL1 mutation (p. R302L) was identified as a potential modifier allele in a patient with LCA carrying two CRB1 mutations and with a prominent maculopathy.

conclusions. In human heterozygotes of CRB1 mutations (parents of offspring with LCA), distinctive regional retinal dysfunctions were found by multifocal ERG measurements that were consistent with the focal histologic abnormalities reported for the two CRB1 knockout mice models. This phenotypic finding may identify CRB1 carriers and point to the causal gene defect in affected LCA offspring, significantly facilitating the molecular diagnostic process. Evidence suggests a modifier allele in AIPL1 in a patient with LCA with prominent atrophic macular lesions and homozygous defects in CRB1.

Leber congenital amaurosis (LCA; MIM 204000) is a severe, untreatable, congenital retinal dystrophy that leads to blindness. Theodor Leber defined LCA in 1869 as a congenital form of retinitis pigmentosa (RP) with profound visual loss at or near birth, wandering nystagmus, amaurotic pupils, pigmentary retinopathy, and autosomal recessive inheritance. ¹ Severely reduced findings on electroretinography was later added to this definition because this distinguishes LCA from a complex set of other overlapping congenital retinal disorders. ² LCA creates an appreciable burden on the affected child, the family, and society because the blindness is lifelong and begins at birth. It has a worldwide prevalence of 1 in 100,000 newborns, and it accounts for 5% or more of all inherited retinopathies; approximately 20% of children attending schools for the blind have LCA. ³  

LCA is genetically and clinically heterogeneous. Since 1996, 10 genes/loci participating in a wide variety of retinal functional pathways have been implicated in the disease process. ⁴ LCA-associated proteins participate in phototransduction (GUCY2D), ⁵ vitamin A metabolism (RPE65, ^{6 7} RDH12 ⁸ ), photoreceptor development (CRX), ⁹ biosynthesis of cGMP phosphodiesterase (AIPL1), ^{10 11} photoreceptor cell development and structure (Crumbs homolog-1 [CRB1]), ^{12 13 14} and disk morphogenesis (RPGRIP1). ^{15 16 17} The discoveries of these genes have led to an increased understanding of the molecular determinants of retinal physiology/development by identifying new biochemical and cellular pathways. Therefore, LCA serves as a model for all human retinal dystrophies. The LCA genes at chromosomal loci 1p36, ¹⁸ 6q11, ¹⁹ and 14q24 ²⁰ remain to be identified. The seven known LCA genes account for approximately 40% of cases, ²¹ whereas genes underlying the remaining 60% of cases still await discovery. It is now evident that different LCA animal models respond to gene replacement therapy. ²²  

We previously found that obligate heterozygous parents of offspring with LCA mutations have unanticipated photoreceptor dysfunctions that appear to be specific for the involved gene, which points to the causal gene defect in the child, thereby significantly facilitating the molecular diagnostic process while adding important functional information of the heterozygous mutations themselves. We documented cone ERG abnormalities in parents carrying GUCY2D defects, ²³ rod dysfunction in AIPL1 carriers, ²⁴ normal rod and cone ERG function in the RPE65 carriers, ²⁵ cone ERG dysfunction in RDH12 carriers (Yzer S, et al. IOVS 2005;46:ARVO E-Abstract 528), and cone plus rod dysfunction in RPGRIP1 heterozygotes (Dharmaraj SI, et al. IOVS 2004;45:ARVO E-Abstract 4728). We confirmed many of these clinical findings by performing mutation-specific in vitro biochemical studies (Ortiz A, et al. IOVS 2004;45:ARVO E-Abstract 5109; Koenekoop R, et al. IOVS 2005;46:ARVO E-Abstract 1705). ²⁶ We also performed detailed clinical studies on another 31 obligate heterozygous parents of LCA children with known LCA mutations and found that more than 80% of them have mild gene-specific functional deficits, retinal abnormalities, or both. ²⁵  

The purpose of the present study was to identify a possible distinctive phenotype of heterozygotes for CRB1 mutations. Recessive mutations in the Crumbs homolog-1 gene, CRB1, ^{12 13 14} are associated with a diversity of retinal phenotypes, including autosomal recessive retinitis pigmentosa type 12 (RP12), ¹² Leber congenital amaurosis (LCA) ^{13 14} with or without the two well-known retinal findings, preservation of the para-arteriolar retinal pigment epithelium (PPRPE), and Coats-like exudative vasculopathy. A dominant CRB1 mutation has recently been associated with pigmented paravenous chorioretinal atrophy. ²⁷ The CRB1 protein is involved in photoreceptor development and photoreceptor structure, and the two known Crb1 knockout mice have a distinctive inferior focal retinal dystrophy. ^{28 29}  

We tested the hypothesis that CRB1 heterozygotes have a phenotype distinctive for CRB1 defects. We specifically looked for regional retinal changes by multifocal ERG testing, which would correlate with the mouse CRB1 phenotype. We observed regional areas of cone dysfunction in the carriers of CRB1 mutations. We also identified a child with LCA and prominent maculopathy and postulated an AIPL1 or a CRX genotype. We were surprised to find a homozygous CRB1 mutation, but further analysis revealed a pathogenic AIPL1 mutation. We here present clinical information on the patient’s and the parent’s phenotype, suggesting that the AIPL1 defect contributed to the phenotypes as a modifier allele; the homozygous CRB1 mutation is causal. 

Materials and Methods

Clinical Analyses

Molecular Diagnostic Screening Protocol.

Our laboratory is involved with the molecular diagnostics of more than 400 patients with LCA, and we have set up a protocol by which we postulate molecular hypotheses based on the retinal and disease phenotypes of the patients. For example, a patient with light perception (LP) vision and relatively normal retinal appearance is screened for GUCY2D mutations, whereas a patient with measurable or improving vision, night blindness, and translucency of the RPE might have to undergo initial screening of the RPE65 gene. LCA patients with macular changes (including macular colobomas) are screened for CRX and AIPL1 defects, and LCA patients with PPRPE are thought to harbor CRB1 mutations. 

To investigate a possible regional pattern of retinal dysfunction, we measured local electrophysiologic responses of the retina and chose multifocal ERG (mfERG) because Vajaranant et al. ³⁰ clearly demonstrate by mfERG that in XLRP carriers, patchy areas of retinal dysfunction exist. The mfERGs were performed at two institutions, the Montreal Children’s Hospital at the McGill University Health Center (MUHC) and the Department of Ophthalmology of the University of Illinois at Chicago (UIC), with the use of a specialized system (VERIS Science 5.1; EDI Inc., San Mateo, CA). Both eyes were recorded while the pupil was undilated, and the stimulus matrix consisted of 61 hexagons scaled with eccentricity and was presented with a 7-inch monitor, measuring from 3 to 300 Hz. This particular stimulus allowed us to record 61 localized cone ERGs from the macular and paramacular regions, subtending an area up to 20° from the central fovea. The stimulus consisted of black-and-white hexagonal stimulus elements arranged concentrically around a fixation point. The stimulus was displayed on a digital camera, and each element was reversed pseudorandomly in binary m-sequence at 75 Hz (frame rate). Stimuli were modulated between black (0.45 cd/m²) and white (280 cd/m²). Flash intensity was 1.028 log cd · s⁻¹/m², background intensity was 0.727 log cd · s⁻¹/m², and the fixation light measured 0.13 log cd · s⁻¹/m². Total recording time was 3 minutes 38 seconds. 

Seven obligate CRB1 carriers (from four LCA families) participated in this study. Six carriers were investigated at the MUHC (families 1–3). After we found abnormal focal mfERG measurements in four of the six carriers, we wanted to confirm our results in a second, completely independent laboratory with significant mfERG experience. Therefore, we recruited one participant from a CRB1 pedigree with LCA children at the UIC and did not share with the investigator the details of our mfERG findings. This additional CRB1 carrier was studied in detail at the UIC (family 4). We also repeated mfERG testing in our first subject (family 1:1, Fig. 1 ) to confirm our results. After each subject was counseled, informed consent forms were signed and approved by the Montreal Children’s Hospital and the UIC institutional review boards. The study adhered to the tenets of the Declaration of Helsinki. Venous blood was then drawn for DNA extraction and mutation analysis. 

Detailed ocular and visual histories were obtained, pedigrees were drawn, and detailed eye examinations—including best-corrected visual acuity by projected Snellen charts, cycloplegic refraction, slit lamp biomicroscopy, and dilated indirect ophthalmoscopy—were performed on all seven carriers. Visual fields were measured by Goldmann kinetic perimetry with V4e and l4e test lights and with the target moved from nonseeing to seeing retina. Cone and dark-adapted mixed rod-cone electroretinograms (ERGs) were performed on both eyes, and the a- and b-wave amplitudes and implicit times (peak times) were averaged in accordance with the standards recommended by the International Society for Clinical Electrophysiology of Vision. ³¹ We then tested two hypotheses. The first was that heterozygous carriers of CRB1 mutations have regional retinal dysfunction on multifocal ERG (as suggested by the CRB1 mice models). The second was that a heterozygous AIPL1 mutation acts as a modifier allele in a patient with LCA with severe maculopathy and a homozygous CRB1 mutation. 

Molecular Analyses.

PCR amplification on genomic DNA (of patients with LCA and their parents) was performed with intronic primers designed to flank the splice junctions of each coding amplicon of LCA genes CRB1 (12 exons; 27 amplicons), AIPL1 (6 exons; 5 amplicons), and CRX (3 exons). Heteroduplex formation of the PCR products for denaturing high-performance liquid chromatography (dHPLC) analyses were induced after heat denaturation, and dHPLC analyses were performed with a DNA fragment analysis system (WAVE; Transgenomic Inc., Omaha, NE). The values of the buffer gradient, start and end points of the gradient, and melting temperature prediction were determined (WAVEMAKER software; Transgenomic Inc.). Variants in gel mobility identified by this technique were then subjected to bidirectional automated sequencing (model 3100, ABI Prism; Applied Biosystems, Foster City, CA). Some of the CRB1 mutations in this study were identified by a new LCA genotyping microarray, which now contains more than 300 known LCA mutations from seven LCA genes (for technical details, see Zernant et al. ²¹ ). The LCA disease chip identifies currently known LCA mutations, whereas our conventional mutation screening methods identify and add novel changes. After we determined the CRB1 mutations in the children with LCA, we sequenced the appropriate CRB1 amplicon of the parents’ DNA to assess the parental origin of each mutation. We used the following well-accepted set of five criteria to distinguish polymorphisms from pathogenic mutations: the predicted effect of the base-pair change on the protein product; the relative frequency of the variation in LCA patients compared with ethnically matched controls (greater than 1% in controls assigned as a polymorphism); cosegregation of the mutant allele(s) in the affected families; homozygosity or compound heterozygosity in recessive LCA; and conservation of the codon across different animal species. 

Results

Family 1

In family 1, we identified a child with LCA who exhibited measurable vision and the striking preservation of the para-arteriolar (PP)RPE retinal pattern, which prompted us to screen the CRB1 gene first and documented a c.2843 g to a transition that predicts a p. C948Y mutation in CRB1, which we then determined was on the paternal allele (Fig. 1) . We have not yet identified the maternal mutation, despite sequence analysis of the exons and flanking intronic sequences of the CRB1 gene. The 35-year-old father reported difficulties with the vision in his right eye but not the left. He reported not having nyctalopia or photo-aversion. Acuities were 20/20 OU with –4.00 +3.00 × 120° OD and –1.75 OS. Retinal examinations were positive for marked narrowing of the retinal blood vessels in both eyes. The retina was otherwise normal. Average cone b-wave amplitudes were slightly decreased compared with normal at 92 μV, with a 30-msec implicit time, whereas the averaged mixed rod and cone b-wave amplitudes were markedly decreased at 142 μV (normal, >220 μV), with a normal 50-msec implicit time (normal, 45 ± 3 msec). The mfERG of the right eye showed a striking superior regional pattern of abnormality in the right eye (Fig. 1)that corresponded to inferior retinal dysfunction, but the left eye appeared entirely normal (Fig. 1) . The blue in Figure 1superiorly indicates decreased retinal sensitivity, which was confirmed by the raw data showing the actual waveforms and decreased heights of the waves. The central peak corresponding to the foveal input is also lower in the right eye. We averaged the superior and inferior mfERG field data and used Student t testing to determine statistical differences. We found an average of 13.796 nV/deg² superiorly (5.85 nV SD) and 23.44 nV/deg² (7.08 SD) inferiorly. The two means are significantly different (P < 0.001). We repeated the mfERG on this subject 2 years later and confirmed the same results (data not shown). The carrier mother was asymptomatic; her retinal examination results were normal, as were findings on the standard ERG (cone and rod signals) and the mfERG. No superior mfERG field abnormality was observed in either eye (Fig. 1) . 

Family 2

In the second LCA family, the affected LCA patient also had the PPRPE retinal disease pattern and we again postulated CRB1 involvement. We then determined that the child had a c.493_501del 9 bp that predicts a deletion of three amino acids p. D165_I167del. We also found a c.1360 g to a transition, which predicts a p. G454R missense mutation in CRB1 (Fig. 2) . We then documented that both the p. D165_I67del and the p. G454R mutations were on the paternal allele. The CRB1 mutation on the maternal allele has not yet been identified, despite sequence analysis of the exons and flanking intronic sequences of the CRB1 gene. The 46-year-old mother was asymptomatic and had normal results on retinal examination. We measured averaged cone b-wave amplitudes of 100 μV (N > 120 μV), which is slightly below normal with a 33.5-msec implicit time. Averaged mixed rod-cone b-wave amplitudes were normal at 239 μV (N > 220 μV) with a normal 39.5-msec implicit time. The mfERGs of the mother were abnormal for both eyes and showed a regional pattern of abnormality with marked lowering of the central peak (Fig. 2 ; only the right eye is shown). The father was also asymptomatic and had normal results on retinal examination. His cone b-wave amplitudes averaged 98 μV, which is again slightly below normal with a 32.3-msec implicit time, and his mixed rod-cone b-wave amplitudes averaged 223 μV with 40.5-msec implicit times. mfERGs of both eyes showed regional field abnormalities associated with a marked decrease in central sensitivity, less obvious than in the mother (Fig. 2 ; only the right eye is shown). 

Family 3

In the third family (Fig. 3) , we diagnosed LCA and prominent bilateral maculopathy in a child (see 4 Fig. 5 ) and initially postulated sequence variants in AIPL1 or CRX but not in CRB1. No PPRPE or intraretinal white dots were observed. Our genetic analysis revealed a well-known and previously reported heterozygous c. 905 g to t transversion that predicts a p. R302L AIPL1 mutation, though no mutations were found in CRX. In our comprehensive screening protocol, we were surprised to find a homozygous c. 1345 c to t transition, which predicts a p. Q449X nonsense mutation in CRB1. We then confirmed that both parents carried the p. Q449X CRB1 mutation heterozygously and then determined that the father also carried the p. R302L AIPL1 mutation. Both parents were asymptomatic and had normal results of dilated retinal examinations, and mixed rod-cone and isolated cone signals were normal on conventional ERG (mixed rod-cone b-waves >220 μV; cone b-waves >120 μV). The mother’s mfERG findings were normal (Fig. 3) , but the father had substantially abnormal recordings on the mfERG, with superior, temporal, and inferior field abnormalities, associated with marked decrease in central sensitivities (Fig. 3) . 

Because of the potential importance of our previous findings, we wanted to confirm our results in a parallel study. We identified a fourth CRB1 family at UIC and performed detailed mfERG studies without prior knowledge of the detailed findings. 

Family 4

In the fourth family, we found that three offspring had LCA, with retinal pigmentary changes and white intraretinal dots but without PPRPE. No molecular hypothesis could be postulated, and all LCA genes were systematically screened. The LCA proband was found to have a c. 1463 t to c transition, which predicts a p. F488S and a c. 2258 t to c transition that would lead to a p. L753P missense mutation in CRB1 and a c. 1301 c to t transition that would then lead to a p. A434V mutation in RPE65 (Fig. 4) . We traced the p. L753P change in CRB1 and the p. A434V mutation in RPE65 to the maternal allele and assumed that the p. F488S mutation in CRB1 came from the paternal allele (he died before our studies began). The p. A434V allele is found in 11% of the black population (EM Stone, personal communication); hence, we consider this change a non–disease causing polymorphism. The carrier mother was found to be asymptomatic and had normal findings on dilated retinal examination. Her mfERG recordings were, however, notably abnormal. The right eye was found to have a marked regional mfERG field abnormality, associated with a marked decrease in the sensitivity of the fovea, whereas the left eye was entirely normal (Fig. 4) . The raw data show a similar pattern. 

Child with LCA and Maculopathy

The child with LCA in family 3 was first seen by us at age 3; congenital blindness and nystagmus had been diagnosed elsewhere. He was a term infant from a consanguineous couple from Lebanon. On our first examination we found that the child was able to fix and follow. His cycloplegic refractions were –0.5 OD and +1.50 OS. He was also found to have rotatory nystagmus, paradoxical and amaurotic pupils, a large exotropia, and the oculodigital sign of Franceschetti. He much preferred day vision and protested when the lights were turned off. Averaged cone b-wave amplitudes were 15 μV (N > 120 μV) with 30-msec implicit times, but the mixed rod-cone b-wave was nondetectable. Results of the retinal examination were striking (Figs. 5A 5B) . The optic discs were normal, the retinal vessels were narrow, and there was a lack of pigmentary degeneration and PPRPE. Both maculae were abnormal, and multiple, similarly sized, cystic white lesions with surrounding black rims (total size, approximately four disk diameters) were observed (Figs. 5A 5B) . When he was 5, we found that the cycloplegic refractions had become +2.00 + 1.00 × 90 OD and + 4.50 OS and that the acuities were measurable at 6/200 OU. We postulated involvement of the AIPL1 or CRX gene because these two LCA genes are associated with macular colobomas and maculopathies. DNA analyses revealed a heterozygous p. R302L AIPL1 mutation and an additional homozygous p. Q449X mutation in CRB1. 

Discussion

Human photoreceptors are highly specialized and polarized cells that, in the outer nuclear layer (ONL), are in close contact with Müller glial cells for structural and metabolic support. The adhesion belt, named the outer limiting membrane (OLM), contains multiple adherence junctions that are present between photoreceptors and Müller cells. These adherence junctions consist of multiprotein complexes and are linked to the cytoskeleton of the cell. A core component of these complexes is the transmembrane protein CRB1. ^{28 29}  

A natural Crb1 mutant mouse, called retinal degeneration 8 (rd8), ²⁸ was found to harbor a homozygous single base-pair deletion in Crb1 (c.3481delC) causing a frameshift and an early stop codon and likely leading to nonsense-mediated decay of the encoded RNA. The rd8 mouse develops severe retinal abnormalities, including focal photoreceptor degeneration and irregularities of the OLM. Mehalow et al. ²⁸ found that the rd8 mice developed irregularly shaped large, white subretinal spots detectable at the age of 3 weeks that were more heavily concentrated in the inferonasal quadrant of the retina. These spots were then found to correspond histo-pathologically to retinal folds and pseudorosettes. They also found OLM fragmentation and outer segment shortening but normal inner segments. By 5 months, the outer segments had virtually disappeared, the inner segments were swollen, and the Müller cell processes were unusually prominent. The retinal degeneration was unexpectedly focal in appearance, with nearly normal retina present at the edge of a region with severe degeneration. The reason for this is still unclear. The photoreceptor dysplasia and degeneration reported by Mehalow et al. ²⁸ in the rd8 mouse with the Crb1 mutations strongly vary with genetic background, suggesting modifier effects from other retinal genes. 

Van de Pavert et al. ²⁹ inactivated both Crb1 alleles to produce a complete null and examined the resultant mouse retina. There were marked differences between the Crb1 ^−/− and the rd8 retinal findings. Neither 2-week-old nor 2-month-old Crb1 ^−/− mice had retinal abnormalities. At 3 months, however, the mice developed focal areas of retinal degeneration, the OLM was ruptured, and there was protrusion of single or multiple photoreceptors through the OLM into the subretinal space. In addition, there was an ingress of photoreceptors into the inner retina, namely the outer plexiform layer. One of the most striking histologic findings was a double layer of photoreceptors (half rosettes). These rosettes developed normal inner segments and a full OLM, very much unlike the rd8 model. This finding suggests that CRB1 is not essential for the formation of junctional complexes and OLM but rather for the maintenance of these structures. In 6-month-old Crb1 ^−/− mice, ectopic photoreceptor layers were identified that were so large they resembled a funnel abutting the ganglion cell and inner limiting membrane. The authors suggest that the initial insult of the Crb1 mutation is the loss of the photoreceptor to Müller glial cell adhesion at retinal foci. Light exposure experiments revealed a significant increase in retinal degeneration in the Crb1 ^−/− mouse, especially inferotemporally; therefore, light enhances the retinal degeneration in the Crb1 ^−/− retinas. ²⁹ These histologic abnormalities in the two CRB1 mouse models prompted us to hypothesize that regional retinal abnormalities exist in carriers of CRB1 mutations, and that led to this study and the surprising mfERG abnormalities. 

Given that we have thus far found LCA gene–specific phenotypes in obligate heterozygotes with mutations in GUCY2D ²³ and RDH12 (cone ERG dysfunction) (Yzer S, et al. IOVS 2005;46:ARVO E-Abstract 528), RPE65 ²⁵ (normal ERGs), AIPL1 ²⁴ (rod ERG dysfunction), and RPGRIP1 (rod plus cone ERG dysfunction) (Dharmaraj SI, et al. IOVS 2004;45:ARVO E-Abstract 4728), we wanted to know whether CRB1-specific changes also could occur that may be useful in our classification system, which is designed to use simple clinical tests on carriers to point to the molecular defect in the blind child, or that might in the future help to identify carrier states in family members of LCA patients. We now report that five of seven CRB1 carriers from four tested families were found by multifocal ERG testing to have novel and distinctive retinal dysfunction. We documented superior mfERG abnormalities in one carrier (family 1–1 [Fig. 1 ]), superotemporal mfERG abnormalities in three carriers (family 2–1 and 2–2 [Fig. 2 ]; family 4–2 [Fig. 4 ]), and superotemporal-inferior mfERG field abnormalities in one carrier (family 2–1 [Fig. 3 ]). We also found that all five patients with regional mfERG abnormalities had a coexistent central abnormality that represented the signal from the fovea. 

Abnormalities were detected irrespective of the type or severity of the CRB1 mutation because we found similar changes in the carriers of a complex mutation consisting of a missense mutation (p. G454R) with an in-frame deletion (p. D165_I167del; family 2), a nonsense mutation (family 3), and missense mutations (families 1 and 4). 

We found, however, that not all seven of the CRB1 mutation carriers exhibited a regional retinal dysfunction on mfERG testing. Two of the carriers had normal recordings. We also found striking asymmetries in the eyes of two of the carriers, with mfERG abnormalities in one eye but not in the other. We do not yet know the reason for this, but we postulate that other genetic and environmental factors may contribute to the CRB1 carrier phenotype. We found that most CRB1 carriers in this small series had normal or mildly abnormal rod and cone full-field ERGs but that one carrier exhibited significant rod ERG defects (family 1). This is the same carrier who had obvious narrowing of the retinal blood vessels and who had symptoms. The rest of the carriers were without symptoms and had normal findings on retinal examination, including normal vessel caliber. We postulate that the mfERG changes we have documented in the carriers might have been caused by retinal folds or pseudorosettes (as was found in the mice) as a result of a disrupted OLM and that light toxicity might have played a role in the carrier phenotype, as documented in the Crb1 knockout mouse by van der Pavert et al. ²⁸ It will be informative to perform in vivo high-resolution microscopy (OCT-3) in the CRB1 mutation carriers in future studies to test this hypothesis. Jacobson et al. ³² documented thickened retinas in affected persons with LCA and CRB1 mutations. We also plan to perform further psychophysical testing by automated visual field analysis to extend our mfERG findings in the CRB1 carriers. 

In terms of genetics of the LCA patients and the cosegregation of their CRB1 mutations, we found that only family 3 shows definite autosomal recessive inheritance of the CRB1 mutations, with a possible modifier effect from AIPL1. In family 4, only patient 4.1 follows an autosomal recessive pattern of CRB1 variants. The other two patients in family 4 carry two maternal variants. These patients and the patients of families 2 and 1 could also be digenic. In three of the four families reported here, the segregating CRB1 variants do not fully explain LCA in the younger generations. In families 1 and 2, the maternal CRB1 variants could not be detected; in family 4, two of three patients do not carry maternal and paternal CRB1 variants. Thus, it is possible that defects in other genes cause, or play a major role in, the phenotype of these LCA patients. However, in view of the very low incidence of CRB1 variants in the general population (heterozygosity frequency between 1/350 and 1/500 and assuming an incidence of LCA between 1/50,000 and 1/100,000 and a CRB1 share of 10%), the identification of CRB1 variants in these families (from a total of 250 LCA patients investigated) hardly can be coincidental. Thus, we think the CRB1 variants identified in this study, either in combination with as yet unidentified CRB1 defects or in combination with defects in other genes, are causal for LCA in the patients of families 1 and 2 and for two of three affected siblings in family 4. As such, we cannot exclude the possibility that mfERG abnormalities in parents of LCA patients are not only caused by heterozygous CRB1 variants but, as exemplified by parent 1.1 of family 3 and parent 1.2 of family 4, are caused by defects in more than one gene. 

In the second part of our study, we identified a homozygous CRB1 mutation, p. Q449X, associated with a retinal LCA phenotype consisting of a prominent and unusual maculopathy, which, to the best of our knowledge, has not previously been described for CRB1 patients. ²⁴ We wanted to test the hypothesis that CRX or AIPL1 defects contributed to this phenotype as modifier alleles because both CRX and AIPL1 are well known to be associated with maculopathies in LCA. ²⁴  

In an isolated Dutch population, we previously documented intrafamilial phenotypic variability (both in severity and progression of disease) in patients with early-onset severe retinal dystrophy and determined that 13 of the 14 affected members share a homozygous p. Y368H mutation in RPE65. ³³ Despite their sharing an identical genotype, we made the surprising observation that the members did not share a common phenotype. We carefully documented three disparate disease types in terms of their visual function and found that some patients deteriorated, some improved, and some remained stable. ³³ Identical genotypes with disparate disease phenotypes in this and other families led us to hypothesize the existence of modifier genes/alleles. We tested this idea with a new LCA microarray in a cohort of 200 LCA patients and found that 15% of patients have three sequence variants in two genes instead of the expected two in the same gene. ²¹ We found, in five families available for meaningful functional comparison, that the phenotype was more severe in patients with three sequence variants. 

In our LCA patient with homozygous CRB1 mutations and with an unusual prominent maculopathy, we also identified a heterozygous p. R302L in AIPL1 and postulate that the LCA is caused by the homozygous CRB1 defects and that the AIPL1 mutation causes the altered phenotype (maculopathy), and we provide three lines of suggestive evidence for this in lieu of biochemical analyses that are not yet completed. First, a small diffuse maculopathy has only been reported once in a patient with CRB1 mutations p. C948Y + T745M, but this macular lesion is unlike the maculopathy of our patient. ³⁴ Second, additional support for an effect of the p. R302L mutation in AIPL1 comes from comparing the phenotypes of the two carrier parents, the father with the CRB1 p. Q449X and AIPL1 p. R302L and the mother with only the CRB1 p. Q449X mutation. The mfERG data strongly suggest that the cone dysfunction in the father is significantly worse than in the mother (or any other carrier in this study). Third, p. R302L in AIPL1 has been found to have a biochemical abnormality in in vitro studies with NUB1, the molecular partner of AIPL1. ³⁵ The p. R302L mutation in AIPL1 was initially reported by Sohocki et al. ³⁶ and lies just outside the tetra-tricopeptide region (TPR) 3 domain (3′) of the AIPL1 protein in the proposed NUB1 binding area, which spans codons 181 to 330 of AIPL1. The p. R302L mutation was found to inhibit immunoprecipitation between the mutant AIPL1 and the endogenous NUB1. ³⁵ The p. R302L mutation, however, was found not to affect the interaction between AIPL1 and farnesylated proteins. ¹⁰ Our data suggest that the CRB1-associated phenotype was altered by the p. R302L allele found in AIPL1. 

In summary, we found that the heterozygous phenotype of carrier parents with children who have LCA caused by CRB1 mutations does not have a normal phenotype, as suggested by classical Mendelian inheritance models. Instead, we document a novel, unique retinal phenotype of the obligate carriers with CRB1 mutations consisting of regional retinal dysfunction (found by mfERG testing), which corresponds to abnormalities in the opposite mfERG field and correlates well with both published CRB1 knockout animal models. We also present an LCA retinal phenotype with homozygous nonsense mutations in CRB1 and with a prominent macular lesion and suggest a modifier effect by a heterozygous p. R302L AIPL1 allele. 

 Supported by Foundation Fighting Blindness Canada (RKK, FPMC), Stichting Dondersfonds (SY), Gelderse Blinden Vereniging (SY), Algemene Nederlandse Vereniging ter Voorkoming van Blindheid (SY), Stichting Simonsfonds (SY), the Fonds de la Recherche en Sante du Quebec (RKK), the Grant Healthcare Foundation (Chicago, IL) (GAF), the Knights Templar Foundation (Chicago, IL) (GAF), and the FFB USA (GAF).

 Submitted for publication December 22, 2005; revised March 31, 2006; accepted June 29, 2006.

 Disclosure: S. Yzer, None; G.A. Fishman, None; J. Racine, None; S. Al-Zuhaibi, None; H. Chakor, None; A. Dorfman, None; J. Szlyk, None; P. Lachapelle, None; L.I. van den Born, None; R. Allikmets, None; I. Lopez, None; F.P.M. Cremers, None; R.K. Koenekoop, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Pedigree, CRB1 mutation, and mfERGs of family 1. Patient 2.1 carries a paternal p. C948Y CRB1 mutation (noted as p. C948Y for ease of illustration). The paternal mfERG of the OD is shown on the left with the three-dimensional image at the top and the raw data at the bottom, both showing the superior defects, corresponding to the inferior retina. The maternal mfERG of the OD is on the right. Three-dimensional and raw data show normal values.

Figure 2.

 

Pedigree, CRB1 mutation, and mfERGs of family 2. Patient 2.1 carries the paternal p. D165-I167del and p. G454R CRB1 mutations. The paternal mfERG data of the OD (left) and the maternal mfERG data of the OD (right) show the superior mfERG abnormalities. The abnormalities are especially obvious when comparing the plots to normal data, such as those depicted on the right side of Figure 1 .

Figure 3.

 

Pedigree, CRB1 and AIPL1 mutations, and mfERGs of family 3. Patient 2.1 carries the homozygous p. p. Q449X mutations in CRB1 and the additional paternal p. R302L AIPL1 mutation. The paternal mfERG data of the OD (left) shows striking abnormalities of the entire area, whereas the maternal mfERG of the OD (right) appears entirely normal.

Figure 5.

 

Fundus pictures of family 3 LCA patient 2.1. Shown are the posterior poles of the retinas of the right (A; OD) and left (B; OS) eyes of the affected LCA child with the homozygous p. Q449X mutation in CRB1 and a heterozygous p. R302L mutation in AIPL1. Note the conspicuous maculopathy.

Figure 4.

 

Pedigree, CRB1 and RPE65 mutations, and mfERGs of family 4. Patient 2.1 carries the p. F488S and p. L753P mutations in CRB1 with the additional maternal p. A434V variation in RPE65. Only the maternal mfERG data are shown because the father was unavailable. (left) OS with the three-dimensional image (above) and raw data (below), which appear normal. (right) OD showing abnormal mfERG patterns superiorly on both the three-dimensional image (above) and the raw data (below).

The authors thank all the LCA families involved. They also thank Ed Stone and his laboratory for finding the p. L753P mutation in CRB1 and Claudine Robert and Rene Pigeon for recruiting the patients and helping prepare the manuscript.