Molecular Vision 2017; 23:198-209
Received 18 October 2016 | Accepted 31 March 2017 | Published 03 April 2017
1Institut National de la Santé et de la Recherche Médicale, U1051, Institute for Neurosciences of Montpellier, Montpellier, France; 2University of Montpellier, Montpellier, France; 3CHRU, Genetics of Sensory Diseases, Montpellier, France; 4Aravis Medical Center, Ophthalmology Department, Argonay, France
Correspondence to: Gaël Manes, Inserm U1051, Genetics and Therapy of Retinal and Optic Nerve Blindness, Institute for Neurosciences of Montpellier, Hôpital Saint Eloi, 80 Avenue Augustin Fliche, 34091 Montpellier Cedex 05, France; Phone: +33 499 636 089; FAX: +33 499 636 020; email: firstname.lastname@example.org
Purpose: Sixteen different mutations in the guanylate cyclase activator 1A gene (GUCA1A), have been previously identified to cause autosomal dominant cone dystrophy (adCOD), cone–rod dystrophy (adCORD), macular dystrophy (adMD), and in an isolated patient, retinitis pigmentosa (RP). The purpose of this study is to report on two novel mutations and the patients’ clinical features.
Methods: Clinical investigations included visual acuity and visual field testing, fundus examination, high-resolution spectral-domain optical coherence tomography (OCT), fundus autofluorescence imaging, and full-field and multifocal electroretinogram (ERG) recordings. GUCA1A was screened by Sanger sequencing in a cohort of 12 French families with adCOD, adCORD, and adMD.
Results: We found two novel GUCA1A mutations—one amino acid deletion, c.302_304delTAG (p.Val101del), and one missense mutation, c.444T>A (p.Asp148Glu)—each of which was found in one family. The p.Asp148Glu mutation affected one of the Ca2+-binding amino acids of the EF4 hand, while the p.Val101del mutation resulted in the in-frame deletion of Valine-101, localized between two Ca2+-binding aspartic acid residues at positions 100 and 102 of the EF3 hand. Both families complained of visual acuity loss worsening with age. However, the p.Asp148Glu mutation was present in one family with adCOD involving abnormal cone function and an absence of macular atrophy, whereas p.Val101del mutation was encountered in another family with adMD without a generalized cone defect.
Conclusions: The two novel mutations described in this study are associated with distinct phenotypes, MD for p.Val101del and COD for p.Asp148Glu, with no intrafamilial phenotypic heterogeneity.
Cone and cone–rod dystrophies (CODs and CORDs, respectively) are rare, genetically heterogeneous diseases inherited as a dominant, recessive, or X-linked trait. COD and CORD are rare hereditary retinal dystrophies with an estimated prevalence of 1 in 40,000 individuals . They are primarily characterized by progressive loss of cone photoreceptor function; for CORD, this is followed by the secondary, or sometimes concomitant, loss of rod photoreceptor function . Typical symptoms of COD and CORD include photophobia, reduced central visual acuity, and loss of color vision [1,2]. Funduscopy has revealed a range of macular phenotypes, from normal or mild changes to severe macular atrophy . The preserved peripheral vision distinguishes COD from CORD.
Ten disease-causing genes (AIPL1, CRX, GUCA1A, GUCY2D, PITPNM3, PROM1, PRPH2, RIMS1, SEMA4A, and UNC119) and four loci (CORD4, CORD12, CORD17, and RCD1) are currently known in autosomal dominant (ad) COD and CORD (RetNet). Among these genes, GUCA1A and GUCY2D encode guanylate cyclase-activating protein 1 (GCAP1) and retinal guanylate cyclase 1 (GC1, also called RetGC1), respectively, which are involved in the synthesis of cGMP from GTP, an important step in the recovery of photoreceptors after light stimulation and cGMP hydrolysis [4,5].
Following light stimulation of the photoreceptors, rhodopsin or cone photopigments activate their downstream target, transducin, which in turn binds the effector enzyme PDE6, resulting in the hydrolysis of intracellular cGMP; thus, the concentration of cGMP is lowered and a closure of cGMP-gated channels, a decrease in the concentration of free Ca2+, the hyperpolarization of the photoreceptor outer segment, and a neural response are produced [6-8]. The GCAP proteins, activated by low Ca2+ concentration, then stimulate the photoreceptor guanylate cyclases, which convert GTP to cGMP. The pool of cGMP is thus replenished and the dark-adapted state is restored [9-11].
GUCY2D and GUCY2F encode the two types of membrane-bound guanylate cyclases expressed in the photoreceptors, namely GC1 and GC2, respectively [12-14]. Both guanylate cyclases are expressed in rods and cones, but GC1 is predominantly expressed in cones [15-18]. Three isoforms of GCAP proteins (GCAP1–3), encoded by GUCA1A, GUCA1B and GUCA1C, respectively, stimulate GC1 or GC2 in low [Ca2+]free [19-21]. GCAP1 is present in higher concentrations in cone than in rod outer segments [20,22]. GCAP2 is localized in rods , cone inner segments, somata, synaptic terminals , and inner retinal neurons [22,24]. GCAP3 is expressed exclusively in cones . Thus, with the predominant expression of GC1 and GCAP1 in cones, mutations in GUCY2D and GUCA1A result in more severe degeneration in cones than in rods.
To date, 16 different mutations in GUCA1A have been reported in the literature to cause adCOD, adCORD, ad macular dystrophy (adMD), and in an isolated patient, retinitis pigmentosa (RP; Table 1). In this study, the GUCA1A gene was screened for mutations in 12 probands with adCOD, adCORD, and adMD from unrelated French families (Appendix 1). We report on two novel mutations and clinical data associated with the two families with GUCA1A mutations. We found that one family had MD without a generalized cone defect, while the other had typical COD.
Twelve index patients were included in the study. Informed and written consent was obtained for all patients participating in the study. The study (#2008-A01238–47) received authorization from the Sud méditerranée IV ethical board committee (#08 10 05, 04/11/2008), was approved by the French regulation agency for medication (AFSSAPS #B81319–70), and is registered at Clinical Trials(#NCT01235624). The investigators followed the tenets of the Declaration of Helsinki.
Patients underwent a standard ophthalmologic examination (refractometry, visual acuity, slit-lamp examination, applanation tonometry, funduscopy). Kinetic visual fields were determined with a Goldmann perimeter with targets V4e, III4e, and I4e. Optical coherence tomography (OCT) measurement of the macula was performed using an OCT-3 system (Stratus model 3000, Carl Zeiss Meditec, Dublin, CA) or with a spectral domain OCT (Spectralis, Heidelberg, Germany) with software version 3.0. Autofluorescence measurements were obtained with the HRA2 Heidelberg retinal confocal angiograph (Heidelberg Engineering, Dossenheim, Germany), and fundus pictures were taken. Full-fields ERG was recorded using a ganzfeld apparatus (Metrovision, Pérenchies, France) with a bipolar contact lens electrode on maximally dilated pupils according to the International Society for Clinical Electrophysiology of Vision (ISCEV) protocol. Multifocal ERGs were recorded by pattern stimulations of the central 20° with small corneal electrodes using the Reti-port system (Roland Consult, Brandenburg, Germany).
Genomic DNA was isolated from 10 ml of peripheral blood leucocytes using a standard salting-out procedure . Coding exons and adjacent intronic sequences of GUCA1A gene (NM_000409.3; primer pairs and PCR conditions in Appendix 2) were sequenced with an Applied Biosystems 3130xl genetic analyzer (Applied Biosystems, Foster City, CA) using a BigDye Terminator cycle sequencing ready reaction kit V3.1 (Applied Biosystems) following the manufacturer’s instructions. Sequence analysis and mutation identification were performed using the Collection and Sequence Analysis software package (Applied Biosystems). Sorting Intolerant from Tolerant (SIFT), Polymorphism Phenotyping 2 (PolyPhen2), Protein Variation Effect Analyzer (PROVEAN), and MutationTaster were used to predict the possible effects of missense variants.
A cohort of 12 French families (Appendix 1) with adCOD, adCORD), and adMD was screened for the six exons of GUCA1A (NM_000409.3), including the noncoding exons 1 and 2. We identified two novel mutations (16.7%) in the families, namely MTP434 and MTP1560 (Figure 1A). We found c.302_304delTAG (p.Val101del) in MTP434 and c.444T>A (p.Asp148Glu) in MTP1560 (Figure 1B). Both mutations cosegregated with the disease phenotype in available family members (Figure 1A). They were not found in public human genetic databases, including Ensembl, HapMap, the 1000 Genomes project, dbSNP, Exome Variant Server (EVS; release ESP6500), or Exome Aggregation Consortium (ExAC). The substitution p.Asp148Glu was predicted to be damaging by the PolyPhen2, PROVEAN, and MutationTaster programs, but not by SIFT (Table 1). The c.302_304delTAG mutation led to the in-frame deletion of valine at position 101 (p.Val101del). Both affected amino acids, valine 101 and aspartic acid 148, were conserved in 16 GCAP1 orthologs (Figure 1C). Both mutations were localized in the 12-amino-acid loop of one of the EF-hand domains of the protein. The mutation p.Asp148Glu affected one of the Ca2+-binding amino acids of the EF4 hand, and the mutation p.Val101del resulted in the in-frame deletion of one amino acid, valine 101, localized between two Ca2+-binding aspartic acid residues at positions 100 and 102 of the EF3 hand (Figure 2).
Clinical evaluations of four generations in two families (MTP434 and MTP1560) were performed. They revealed an autosomal dominant pattern of inheritance (Figure 1A).
Family MTP434-- Three patients from the two last generations could be examined. They had all experienced the onset of visual acuity loss in adolescence, with moderate to severe reduction later in life; a perception of color vision impairment since childhood, with a variable axis of confusion (not shown); moderate night blindness in adulthood; and photophobia (Table 2). On visual-field examination, all had a central scotoma with preservation of the periphery (not shown). Fundus examination showed the absence of foveal reflex in the youngest patient (IV:3), perifoveal discoloration in his older sister (IV:1), and a large, round-shaped macular atrophy with sharp limits and few pigment deposits in the lesion in their mother (III:2; Figure 3). Fundus autofluorescence revealed a mild perifoveal hyperautofluorescence in IV:3, whereas IV:1 had a small, round hypoautofluorescent lesion, and III:2 exhibited no autofluorescence in the macular lesion (Figure 4). Importantly, in all patients, the retinal autofluorescence was normal outside the macular lesions. OCT examination confirmed complete (III:2) or near-complete (IV:1) foveal atrophy in two individuals, whereas in IV:3, a thinned ellipsoid zone was still present (Figure 4). Full-field ERG performed on the mother (III:2) and her daughter (IV:1) did not show significant alteration of the scotopic responses (Figure 5A). The photopic responses were not delayed. Their amplitudes were slightly decreased in III:2 and reduced to half the normal value in IV:1. Multifocal ERG could only be performed in two of the children. In IV:1, only the foveal responses had a decreased amplitude, whereas the responses outside the fovea were normal. In IV:3, most responses had a disorganized pattern except those of the outer ring, which showed normal amplitudes (Figure 5B). In summary, family MTP434 exhibited an MD phenotype.
Family MTP1560-- Patients from the three last generations could be examined. They all complained of photophobia and color vision difficulties that had been present since childhood, but none reported night blindness. Visual acuity was variously decreased (Table 2). None of the patients had macular atrophy; the son (IV:1) had a slightly decreased foveal reflex that was absent in his mother (III:2) and his grandfather (II:1; Figure 3). Fundus autofluorescence showed small foveal lesions in the three generations and normal autofluorescence in the peripheral retina; moderate perifoveal hyperautofluorescence was observed in IV:1, whereas the fovea was moderately hypoautofluorescent in III:2 and II:1 (Figure 4). OCT examination of the son (IV:1) revealed a thickened ellipsoid zone in the fovea (Figure 4). His mother (III:2) had an absent ellipsoid zone, with thinning of the outer nuclear layer in the fovea, and a virtually normal outer retina outside the fovea in the temporal area of the macula. The grandfather (II:1) had partial foveal atrophy, with a hyporeflective zone in both eyes. Full-field dim blue scotopic ERG responses were normal in the three family members (Figure 5A), but the scotopic responses for the brightest stimulation were moderately decreased. In the three family members, the photopic responses were dramatically decreased, although they were slightly better in IV:1, and the 30-Hz flickers were delayed. Multifocal ERG, performed for IV:1 and his mother (III:2), showed highly abnormal responses (Figure 5B). In III:2, all responses were disorganized. In IV:1, most responses showed decreased amplitudes. In summary, family MTP1560 exhibited a COD phenotype.
To date, 16 different mutations in GUCA1A that lead to adCOD, adCORD or adMD have been listed on the Human Genome Mutation Database (HGMD), except the unclear p.Thr114Ile substitution associated with atypical RP (Table 1). In this study, we report two novel mutations in GUCA1A and clinical data associated with the two families in which the mutations were identified, where one had adCOD and the other had adMD. The two novel mutations affect evolutionarily conserved amino acids (Figure 1C) and are located in a highly-conserved region, namely the EF-hand domains of the GCAP1 protein, which contain most disease-causing mutations (Table 1, Figure 2). The helix–loop–helix structure, called the EF hand, was first identified in parvalbumin, where the two helices E and F, joined by a calcium-binding loop, resemble a hand . This EF-hand motif is found in many calcium-binding proteins. The GCAP1 loop region of the EF2–4 hands contains 12 residues, and several of these residues are involved in the binding of the Ca2+ (Figure 2, arrowheads). The EF1 hand is not capable of binding Ca2+.
Among the 16 causative mutations in GUCA1A referenced in the literature, six affected one of the amino acids directly involved in the binding of the Ca2+ in the EF loop domains (Figure 2; p.Asp100Glu, p.Asp100Gly, p.Asp102His, p.Asn104Lys, p.Glu155Gly, and p.Glu155Ala). For the mutation p.Asp100Glu in the EF3-hand motif, it was shown that although the exchange of Asp for Glu at position 100 added only a CH2 group to the side chain, it caused an impaired apparent Ca2+ affinity, which was 100-fold lower than that of the wild type [27,28]. The novel mutation p.Asp148Glu, which affects one of the Ca2+-binding EF4 amino acids with a concomitant exchange of Asp for Glu at position 148, probably lowers the affinity of the EF4 to Ca2+. The mutation p.Val101del is located at a position that is not directly involved in complexing Ca2+, although it appears between two Ca2+-binding aspartic acid residues at positions 100 and 102 of the EF3 hand (Figure 2). The in-frame deletion of one valine at position 101 may affect the loop conformation and the Ca2+ binding on the adjacent aspartic acid residues.
All the GUCA1A mutations listed in Table 1 (the two novel mutations of the present study and the 16 referenced mutations) are absent from the public genetic databases except the following two: p.Pro50Leu, found at a high frequency (125/121,374 alleles in ExAC and 13/13,006 alleles in EVS), and p.Thr114Ile (5/121,228 alleles in ExAC). This means that 125 carriers of p.Pro50Leu and 5 carriers of p.Thr114Ile were identified in an ExAC cohort comprising more than 60,000 individuals from the general population. The prevalence of the condition due to p.Pro50Leu and p.Thr114Ile would be 1 in 485 and 1 in 12,122 individuals respectively, which is more than the estimated prevalence for COD and CORD including all inheritance patterns (1 in 40,000 individuals) . The variations p.Thr114Ile and p.Pro50Leu affect amino acids that are not evolutionarily conserved (Figure 1C and data not shown, respectively). Furthermore, the family described with p.Pro50Leu presented intrafamilial variability in expression . The mutation p.Thr114Ile was identified in a simplex patient with an atypical form of retinal degeneration, but no cosegregation analysis was possible to validate the mutation . Unfortunately, the mutated protein was not biochemically characterized in the study. Moreover, p.Pro50Leu and p.Thr114Ile are predicted to be benign by PolyPhen2 and SIFT, and the latter is also predicted benign by PROVEAN and MutationTaster (Table 1). In conclusion, these two variants are unlikely causative mutations of adCOD, adCORD, or adMD.
The present study found the prevalence of GUCA1A to be 16.7% in our small cohort of 12 families. Previous prevalence studies conducted in patients with adCOD and adCORD revealed that among the 10 disease-causing genes, each gene prevalence is low, but together, they explain approximately 20% of cases . The estimated frequency of these genes is 7.7% for GUCY2D, 4.3% for CRX, 3.0% for GUCA1A, 1.8% for PRPH2, 1.4% for PROM1, and 1.2% for PITPNM3. The last four genes (AIPL1, RIMS, SEMA4A, and UNC119) are mutated in less than 1% of patients . In the literature, the prevalence of GUCA1A mutations varies widely from 0.76% to 16.7% for cases of adCOD and adCORD in cohorts of different origins. In a cohort of 131 patients with cone degeneration, cone–rod degeneration, or cone dysfunction, only one patient had a mutation in GUCA1A (0.76%) . In two characterized cohorts of 22 and 24 patients with adCOD and adCORD, respectively, 3/22 (13.6%)  and 4/24 (16.7%)  patients had a causative mutation in GUCA1A; these frequencies are similar to those obtained in our study.
For family MTP1560, the phenotype of the patients with p.Asp148Glu mutation was that of COD on full-field ERG, and none of the patients had macular atrophy. Conversely, the patients in family MTP434 with the p.Val101del mutation had MD. COD and CORD were first associated with GUCA1A mutations [30,33-35]. Next, six referenced GUCA1A mutations were associated with an MD phenotype (Table 1), namely p.Leu84Phe , p.Glu89Lys , Tyr99Cys , p.Asp100Gly , p.Ile107Thr , and p.Leu176Phe . The first report of a macular dysfunction caused by the GUCA1A mutation was published by Michaelides and collaborators . They described six affected individuals from a family with the p.Tyr99Cys mutation displaying isolated MD, COD, and CORD phenotypes. Then, one family with p.Glu89Lys mutation carried by two patients, one with an unknown clinical status and one with COD or MD, was reported . After that, two Spanish families, including a three-generation family with five affected patients carrying the p.Leu84Phe mutation and one index patient with the p.Ile107Thr mutation, were described with intrafamilial heterogeneity that included isolated macular dysfunction, COD, and CORD . Thereafter, a family of three affected individuals with the p.Asp100Gly mutation was described with a heterogeneous presentation of phenotypes and macular atrophy, cone, and cone–rod dysfunction on ERG . Finally, a missense variation, p.Leu176Phe, was identified via whole-exome sequencing in a patient with MD . No information about the clinical status of the patient or about the cosegregation on additional samples were presented. Except for this latter mutation, all the five referenced mutations and the novel p.Val101del associated with MD were localized from the EF2 helix F to the EF3 loop (Figure 2, hashtag). Contrary to what is presented in the literature for the MD phenotype associated with GUCA1A mutations, there was no intrafamilial phenotypic heterogeneity in family MTP434 with the p.Val101del mutation. The three patients examined had MD and no signs of COD or CORD.
Biochemical analysis of GCAP1 function revealed that for most mutations identified in GUCA1A, the Ca2+-sensor GCAP1 exhibits an impaired Ca2+ sensitivity and a dysfunction of its guanylate cyclase–activating properties, as summarized in two reviews [10,40]. Based on these biochemical analysis, we hypothesized that the two novel GUCA1A mutations described in this study, p.Val101del and p.Asp148Glu, result in defects in the GCAP1 EF hands and affect calcium homeostasis. This resulted in MD for p.Val101del and COD for p.Asp148Glu.
Appendix 2. Sequence of the primers (5'-3'), fusion temperatures and concentration of MgCl2 used to amplify GUCA1A (NM_000409.3)
No financial disclosures. The authors acknowledge funding support from the SOS Rétinite Foundation (Montpellier, France) for a fellowship to GM.