Mutations in GRM6 Cause Autosomal Recessive Congenital Stationary Night Blindness with a Distinctive Scotopic 15-Hz Flicker Electroretinogram

Abstract

purpose. Congenital stationary night blindness (CSNB) is a group of nonprogressive retinal disorders characterized by impaired night vision that occurs in autosomal dominant, autosomal recessive, or X-linked forms. Autosomal recessive (ar)CSNB seems to be very rare. Mice lacking the metabotropic glutamate receptor 6 (Grm6) have a defect in signal transmission from the photoreceptors to ON-bipolar cells. In the current study, the human orthologue (GRM6) was screened as a likely candidate for arCSNB.

methods. arCSNB individuals of five families were screened for mutations in GRM6. Subsequently, they were examined with standard and 15-Hz flicker electroretinography (ERG). These recordings were compared with those of patients with X-linked CSNB1.

results. Affected individuals in three of five families carried either compound heterozygous or homozygous mutations in GRM6. Strikingly, all of them displayed a distinctive abnormality of the rod pathway signals on scotopic 15-Hz flicker ERG.

conclusions. The novel profile identified in this study suggests the existence of more than two rod pathways. The distinctive ERG feature was not observed in patients with X-linked CSNB1 and additional affected individuals with unknown molecular defect. These observations will help to discriminate autosomal recessive from X-linked recessive cases by ERG and molecular genetic analysis.

Congenital stationary night blindness (CSNB) comprises a group of nonprogressive retinal disorders characterized by impaired night vision and sometimes other ocular symptoms such as decreased visual acuity, myopia, hyperopia, and nystagmus. X-linked and autosomal recessive (ar)CSNB are characterized by an electronegative electroretinogram (ERG) showing normal or near-normal a-waves, but a selective reduction of the b-wave. ¹ Mutations in two genes, NYX and CACNA1F are responsible for two X-linked types of CSNB designated complete (CSNB1) and incomplete (CSNB2), respectively. ^{2 3 4 5} CSNB1 has been associated with absent rod and normal or subnormal cone b-wave amplitudes, whereas CSNB2 exhibits some rod function, but both rod and cone ERG signals are considerably reduced in amplitude. These types can be distinguished from each other by means of the standard ERG. However, so far, patients with arCSNB cannot be discriminated from those with CSNB1 by clinical features and standard ERG measurements. 

The ERG b-wave reflects the transmission of the signal from the photoreceptors to the second-order neurons. Photoreceptors, when activated by photons, release glutamate as a neurotransmitter, which then stimulates either ON or OFF bipolar cells. Although rods synapse only with rod ON bipolar cells, cones connect with both cone ON and OFF bipolar cells. ⁶ Rod signals in normal human subjects use two different retinal pathways: the rod ON bipolar—amacrine II cell pathway (slow, sensitive rod pathway) and the rod-cone gap junction—cone ON bipolar pathway (fast, insensitive rod pathway). The existence of these two rod pathways has been demonstrated by means of ERG studies using 15-Hz flicker at scotopic intensities. ^{7 8} At scotopic (starlight) luminance levels, the responses to 15-Hz flicker are dominated by the slow rod pathway and at mesopic (twilight) levels by the fast rod pathway. Between these luminance levels, destructive interference produces a psychophysical flicker null and a minimum ERG amplitude with a 180° phase shift, because the signals of both pathways are equal in amplitude but opposite in phase. Electrophysiological studies identified the rod ON pathway as the primary site of the defect in CSNB1. ^{9 10} The ON bipolar dendrites express mainly the high-affinity, sixth subtype metabotropic glutamate receptor (mGluR6). ^{11 12 13} The phenotype of homozygous knockout mice lacking the gene (Grm6) encoding receptor mGluR6 resembles that of the human CSNB: a greatly reduced b-wave under dark-adapted conditions, whereas the a-wave, representing the photoreceptor activity, is normal. Furthermore, the retinal organization in these mice shows no obvious changes on the cellular level. ¹⁴ Because the glutamate receptor plays a critical role in both rod pathways, mutations in this gene may influence both the slow and the fast rod pathway signals. The absence of fundus abnormalities and the electronegative standard ERG renders this gene an attractive candidate for arCSNB. Indeed, most recently, another group identified GRM6 mutations in three patients with arCSNB. ¹⁵ Herein, we report on ERG responses and GRM6 mutation analysis in seven patients with diagnosed arCSNB. 

Subjects and Methods

Patients

Seven patients from five unrelated families had congenital stationary night blindness (CSNB) diagnosed by means of a complete ophthalmic examination, including funduscopy, dark-adapted thresholds and Ganzfeld ERG, according to the standard ISCEV (International Society for Clinical Electrophysiology of Vision) ERG protocol. ¹⁶ The protocol of the study adhered to the tenets of the Declaration of Helsinki. In two male patients, no pathogenic variant was found in NYX or CACNA1F ¹⁷ ; the other five patients were female offspring of unaffected parents. The clinical characteristics of the patients are shown in Table 1 , the standard ERGs of the father of the three affected siblings and one of these siblings are represented in Figure 1together with the responses of a patient with CSNB1. 

Mutation Analysis

All subjects or their parents gave written informed consent before molecular genetic testing. Genomic DNA was isolated from EBV-transformed lymphoblastoid cells or blood by standard techniques. The 10 coding exons of GRM6 were amplified with intronic or exonic primers (Supplementary Table S1). 

Eleven fragments containing the 10 coding exons of GRM6 were amplified with polymerase (HotFirePoly DNA Polymerase; Solis Biodyne, Tartu, Estonia), 1.5 mM MgCl₂, and Q solution (Qiagen, Hombrechtikon, Switzerland). The total volume of the PCR reaction was 25 μL, and 0.8 μL of the PCR product was treated with 5 μL of 1:50 diluted clean-up solution (ExoSAP-IT; USB, Cleveland, OH) at 37°C for 15 minutes, to remove unconsumed primers and dNTPs. After treatment, the enzyme was inactivated by heating to 80°C for 15 minutes. After a short centrifugation step, 1 μL dye-termination and 1.5 μL 5× dye-termination buffer (BigDye Terminator ver. 1.0; Applied Biosystems, Rotkreuz, Switzerland), 2 μL 5× Q solution, and 0.8 μL primer (10 pM/μL) were added to each reaction, and standard sequencing was performed. Sequences were analyzed on an automated DNA sequencer (Prism 3100; Applied Biosystems). 

ERG Stimulation, Recording, and Analysis

The eyes of the subjects were anesthetized with oxybuprocaine 0.4%, and the pupils were dilated with tropicamide 0.5%. Subsequently, DTL electrodes were positioned on top of the lower eyelid, touching the eye. Pupil diameters were determined immediately after the ERG recordings were finished (8–9 mm in all subjects). For the standard ISCEV ERG measurements, Xenon tube flashes (duration <10 μs) were delivered in a custom-made Ganzfeld dome, at one flash every 2 seconds for the low (−2.6 ND), and one flash every 5 seconds for the standard ISCEV intensity (mixed response). Subjects were then light adapted for 10 minutes by exposure to a white 30-cd/m² rod-saturating background, and photopic ERGs were recorded to standard ISCEV intensity and to white 30 flashes per second. 

Preamplification (5.000×) and sampling (1000 Hz) plus multiplexing were performed with a custom-made, battery-driven preamplifier unit. These multiplexed signals were optically transmitted to the recording computer. Signals were digitally filtered online, using a band-pass filter from 0.5 to 200 Hz (3 dB). Sweep length for averaging was 200 ms, with an additional 40-ms prestimulus period. During the averaging process, the raw ERG sweeps were displayed for examination and judged for inclusion or exclusion by the examiner. This process results in a better artifact rejection than is usually obtained with level rejection. Only the averaged signals and the plus–minus averages ^{18 19} were stored for further analysis. 

ERG responses to 15-Hz flicker were recorded in six of the seven patients with arCSNB. Because patient 21973 had Down syndrome, she was unable to comply with the lengthy 15-Hz procedure. The 15-Hz protocol was recorded intermixed with the standard ISCEV ERG at the appropriate intensities. Xenon flash (<10 μs) intensities of 15 flashes per second (sweep length, 135 ms, 500–1000 averages) were attenuated by switching between 0.3 and 0.6 J (Nihon Kohden, Tokyo, Japan) and inserting neutral density (ND) filters (Schott, New York, NY) mounted in a filter wheel. Using maximum attenuation, we obtained a stimulus intensity of −1.56 log scot-td/s, reaching +0.44 log scot-td/s in nine steps. 

The 15-Hz responses were trend corrected off-line and filtered by a moving average filter with a rectangular window of 20 ms (21 data points). The 15-Hz scotopic flicker ERG in normal subjects showed a minimum response at approximately −0.56 log scot-td/s. The ERG signals to flicker intensities between −1.56 and −0.72 log scot-td/s were therefore considered to be dominated by slow rod ERG signals. ²⁰ Amplitudes and phases of the 15-Hz averaged amplitudes and the plus-minus–averaged signals ^{18 19} were obtained by off-line Fourier analysis. Amplitudes were considered significant if the ratio of the amplitudes of the average and the plus–minus average was above 3 (P < 0.01). For those responses, we plotted phases according to a strict criteria, assuming that phases decreases monotonously with stimulus intensity. We subtracted an integer multiple of 360° from the calculated phase (phase is indeterminate for multiples of 360°). An extra 360° was subtracted whenever the phase for that intensity seemed to increase >1 SD of the phase in unaffected subjects. 

Apart from the patients with arCSNB, six patients with CSNB1 (age, 4–14 years) were tested with the 15-Hz protocol. The CSNB1 diagnosis was based on a positive X-linked family history in three consecutive generations, clinical examination, standard ISCEV ERGs, and 30-Hz photopic flicker ERGs. ²¹  

Results

Mutation Analysis

The GRM6 gene consists of 10 exons and encodes a protein with 878 amino acids. We performed mutation analyses in the open reading frame (ORF) of the 10 exons and parts of the flanking introns of GRM6 (Accession NM 000843; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD) in five unrelated families, with at least one member displaying lack of nocturnal vision and an electronegative ERG. In patient 2445 of family 1, we identified two compound heterozygous mutations: one missense mutation (c. 137C→T) leading to a Pro46Leu change in exon 1 and one nucleotide insertion (c.720_721insG) in exon 3, leading to a frame shift at position 243 and a premature stop codon 39 amino acids later. The mother of this patient was heterozygous for the insertion, and the missense mutation was transmitted from the father. The unaffected sister of patient 2445 showed none of the alterations (Fig. 2) . In family 2, a male patient (2496) carried a homozygous missense mutation (c. 1565G→A), leading to a p.Cys522Tyr change in exon 8 of GRM6. The mother and the unaffected brother were heterozygous for this mutation, whereas the unaffected sister was homozygous wild type. The father was not available for testing (Fig. 2) . In family 3, the three affected sisters (26162, 21973, and 21974) carried a heterozygous duplication of 19 nucleotides (c.57_75dupl19) in exon 1, which denotes a frame-shifting exchange of Leucine (Leu) at position 26, as the first affected amino acid, with valine (Val), creating a new ORF ending in a stop at position 168. A second mutation (c.172G→C) was identified in the same exon and leads to a p.Gly58Arg substitution. The mother was heterozygous for the duplication, whereas the father carried the p.Gly58Arg exchange (Fig. 2) . None of the mutations was detected in more than 100 healthy control chromosomes of European descent. Two of the seven analyzed patients (26111, 26151) did not show a mutation in the complete ORF of GRM6. Patient 26111 had one unaffected sister and one unaffected brother. Patient 26151 had one unaffected sister. The parents of both patients, as well as their siblings were unaffected, as were the grandparents of patient 26151 (data not shown). 

Protein Analysis

mGluR6 is a metabotropic glutamate receptor that belongs to a G-protein-coupled receptor family that has been divided into three groups based on their homology, putative signal transduction mechanisms, and pharmacologic properties. ²² mGluR1 and mGluR5 belong to group I and activate phospholipase C. mGluR2 and mGluR3 belong to group II, and group III includes mGluR4, mGluR6, mGluR7, and mGluR8. Group II and III receptors are involved in the inhibition of the cyclic AMP cascade, but differ in their agonist. A multiple sequence alignment of all metabotropic glutamate receptors of group III revealed that all missense mutations (p.Pro46Leu, p.Gly58Arg, and p.Cys522Tyr) are located in residues that are conserved among the other glutamate receptors of this group. In addition, the positions of the p.Gly58Arg and the p.Cys522Tyr exchanges are shared by all eight metabotropic glutamate receptors (data not shown). Comparative sequence analysis of mGluR6 and the crystallized rat mGluR1 revealed that the p.Pro46Leu and the p.Gly58Arg exchanges are located in the first ligand binding domain, whereas the p.Cys522Tyr exchange is predicted to be in a highly conserved cysteine-rich region. The two mutations leading to frame shifts (p.Leu26fs and p.Val243) occur shortly before the first and within the second ligand-binding domain, respectively (Figs. 3A 3B) . 

Slow and Fast Rod ERG Signals

Patients in whom CSNB was diagnosed by means of standard ERG measurements (see the Methods section) were further analyzed with a 15-Hz scotopic ERG to test for a clinical correlation with the identified mutations in GRM6. This method was used because it allows discrimination between the two rod pathways: the rod ON bipolar-amacrine II cell pathway (slow) and the rod-cone gap junction-cone ON bipolar pathway (fast), and both use mGluR6 to signal. 

The ERG responses to 15-Hz flashes at scotopic conditions in a normal subject, a patient with clinically diagnosed CSNB1, and two with arCSNB, without and with the GRM6 mutation, are shown in Figure 4 . In the normal subject, the amplitude of the ERG signal increased slightly with increasing flicker intensity from −1.56 to −1.26 log scot-td/s and decreased thereafter. There was a minimum ERG response at flicker intensities between −0.72 and −0.42 log scot-td/s. The phase of the ERG responses abruptly increased by approximately 180° (corresponding to a half-cycle) as the amplitude minimum was crossed, consistent with destructive interference between the slow and fast rod ERG signals being the cause of the amplitude minimum. The ERG signals rapidly increased in amplitude at higher flicker intensities (from −0.42 log scot-td/s onward). 

In the patient with CSNB1, the ERG signals of the slow rod pathway at flicker intensity between −1.56 and −0.56 log scot-td/s were indistinguishable from noise. An ERG signal became apparent at the higher flicker intensities. However, this signal was reduced in amplitude. The 15-Hz scotopic responses were similar in all six tested patients with CSNB1. 

In patient 21974 of family 3, the amplitude of the ERG response showed a complicated profile: a large response at the lowest intensity and a minimal response at four intensities (−1.42, −0.72, −0.26, and +0.44 log scot-td/s). The respective responses of all our examined patients with GRM6 mutations (2445, 2496, and 26162) had similar characteristics. This behavior of the slow rod pathway signals could not be demonstrated in the two other patients presumed to have arCSNB (26111 and 26151) without GRM6 mutation nor in the patients with CSNB1 analyzed. ²⁰ For the arCSNB patients with GRM6 mutations consistent responses were found for most of the stimuli in the lower range, although some amplitudes were below the normal mean (Fig. 5A) . For these patients the corresponding phase plot is given in Figure 5B . The patients showed a much more rapid phase decrease at the lower intensity range than in unaffected subjects. Finally, in one of the arCSNB subjects an extra phase shift occurred around 0 log scot-td/s. 

In the CSNB1 group the responses at stimulus intensities below −1.0 log scot-td/s were not significant. From −0.72 to −0.26 log scot-td/s some subjects showed significant responses at some intensities; however, with larger interindividual variability. Only at intensities of +0.28 log scot-td/s and higher, response amplitudes were systematically growing. We did not make a phase plot for the patients with CSNB1, because they had so few significant signals at the lower intensities that phases could easily vary 360°. 

Discussion

With a candidate gene approach, we identified, in three of five families with diagnosed CSNB, mutations on both alleles of the GRM6 gene. This gene was an attractive candidate because a mouse model lacking the metabotropic glutamate receptor 6 (Grm6) showed a defect in the transmission of the light signal from the photoreceptors to the ON-bipolar cells, similar as observed in patients with X-linked or arCSNB. Sequence analysis verified our hypothesis. One patient showed a homozygous missense mutation, whereas the others were compound heterozygous. The latter carried different missense mutations on one allele and insertions on the second. All missense mutations affected amino acid residues that are conserved among the other glutamate receptors of group III. In addition, four of the five different mutations occurred in the first part of the receptor, in exons 1 and 3. The third missense mutation affects one of several conserved cysteine residues that are thought to form disulfide bridges outside the cell membrane. Our findings are supported by a very recent report in which also missense mutations in GRM6 were associated with arCSNB. ¹⁵  

The x-ray crystal structure of the rat mGluR1 showed two conformations that undergo a dynamic change between the open and the closed form. On agonist binding the closed conformation is stabilized and thereby activates the receptor. Within the three-dimensional structure of mGluR6, the position of the missense mutations identified in our study is close to the binding region of glutamate (data not shown). Although the direct binding pocket is not affected, our findings suggest that the mutated regions have a role in recruitment or binding of glutamate. This notion is supported by reports showing that amino acids from the upper lobe are predominantly involved in anchoring the ligand and that polarity of these residues is conserved among the different metabotropic glutamate receptor subtypes. ²³ Furthermore, site-directed mutation analyses have shown that in vitro polar or charged amino acids from the upper lobe cleft region of the molecule are key determinants of agonist affinity. ^{24 25} One of the missense mutations reported by Dryja et al. ¹⁵ was also identified in the ligand-binding domain, whereas another one is located in the intracellular loop between two transmembrane domains. Two other mutations, found in exon 8, lead to protein truncation. Together, these findings suggest that the identified mutations lead to the loss of function of the glutamate receptor. Future studies will show how these mutations exactly modulate the functional properties of the receptor. 

With ophthalmic examinations, including standard ERGs, one cannot distinguish arCSNB from CSNB1 in patients. However, rod pathway responses to 15-Hz flashes showed a distinctive profile in patients with mutations in GRM6, different from those of patients without GRM6 mutations and also different from patients with CSNB1. Scholl et al. ²⁰ demonstrated minimal slow rod pathway signals in patients with CSNB1 and reduced fast rod signals with increased phases, by using 15-Hz scotopic ERG measurements. They suggested that defective nyctalopin in patients with CSNB1 leads to a complete blockage of signal transmission within the ON pathway. This hypothesis was recently further supported by investigations in primates. ⁹ In accordance with Scholl’s research, our study has also demonstrated minimal slow rod pathway signals in patients with CSNB1 and reduced fast rod signals with increased phases (Fig. 4) . 

In contrast, the 15-Hz scotopic signals in the patients with arCSNB with GRM6 mutations had remarkable abnormal phase behavior, with several minimum responses. In Scholl’s study of patients with CSNB1, abnormal phase behavior was taken as proof of absence of responses at the lower flicker intensities, because in nonaffected persons, there is a proper alignment of the phases between adjacent ERG responses. In CSNB1, the slow rod pathway signals are very small, and therefore their phases may indeed be substantially influenced by noise. However, in the patients with arCSNB carrying GRM6 mutations the amplitudes of the responses were significantly above the noise level, and the phases were consistent between both eyes of the same patient. Because similar phase shifts occurred in all patients with GRM6 mutations, this abnormal phase behavior may be a characteristic feature of this retinal disorder. 

Large phase shifts between adjacent ERG responses may be caused by abnormal modes of transmission within the rod ON bipolar pathway, but several minima in the amplitude versus intensity plot can only be explained by interference between at least two signals. The first minimum response is apparent at −1.42 log scot-td/s. At this intensity, the largest signals in nonaffected persons are those of the slow rod pathway. To destructively interfere with the slow pathway, fast rod pathway signals should have approximately similar amplitudes. However, at this very low intensity, the fast signals have been shown to be much smaller than the slow rod pathway signals. ²⁶ Previous studies showed that significant interference of slow rod pathway signals with cone signals is also unlikely. ^{7 27}  

The interference pattern in the patients with arCSNB therefore raises the question of the existence of more than two rod pathways, an assumption that was also previously suggested. ^{28 29} The slow rod signal may consist of more than one mode of transmission near absolute threshold, as has been demonstrated in cats. ^{28 30} Apart from mGluR6, ON bipolar dendrites also bear ionotropic glutamate receptor subunits. ³¹ A second slow rod pathway may consist of signals traveling via these receptors. In normal subjects, the interference pattern of two slow rod transmission modes could result in the normal overall slow pathway signal. If a GRM6 mutation leads to abnormal functioning of one of those routes, an interference pattern, as found in our patients, may be conceivable. 

Alternatively, rod signals may be transmitted via a direct rod to OFF cone bipolar pathway. This pathway exists in rodents and cats and may be a common feature of mammalian retinas. ^{31 32 33 34} In a mouse retina genetically modified to be coneless (mimicking human achromatopsia), direct signals via rods to OFF cone bipolar cells have been demonstrated. ³² The residual fast rod signal in CSNB1 could be the result of rod signal transmission through rod–cone gap junctions to OFF cone bipolar cells. ⁹ This third pathway uses the glutamate receptor iGlur2 instead of the mGlur6 transmission of the ON pathways. ³⁴ It is supposed to be less sensitive than the slow rod pathway. ^{28 32} Consequently, signals of this pathway at −1.42 log scot-td/s may be very small, as for the fast rod pathway signals. However, with phase shifts of either the fast rod pathway signals or the third rod pathway signals, or both, interaction could result in a larger combined signal. This could, in turn, destructively interfere with the slow pathway signals. 

In patients with CSNB1, absence or near absence of slow rod pathway responses results in an undetectable rod mediated response in the standard ERG and an absent rod component in dark adaptation. The slow rod pathway responses in the patients with arCSNB with GRM6 mutations seem to indicate that not only amplitude, but also timing of the signals is essential in generating a standard ERG rod response and perception of the stimulus. Proper alignment of consecutive responses appears to be necessary to create an overall signal. 

In a recent study by Dryja et al., ¹⁵ three patients carrying GRM6 mutations showed markedly reduced ON responses but nearly normal OFF responses to a sawtooth flickering light under photopic conditions. The authors suggest the possibility of alternative routes for rod signaling, because the visual functions of these patients could not be fully explained by defective signaling in the ON bipolar pathway. The rod pathway profile found in our study may be the first direct evidence of the existence of such alternative pathways in humans. However, the number of patients in our study was small, due to the rarity of arCSNB. Larger numbers of patients with GRM6 mutations are needed to substantiate our results, and additional experiments, for instance with different scotopic flicker frequencies, are necessary to clarify further the abnormal phase behavior of the slow rod pathway signals. It would also be interesting to see whether other not yet identified gene defects will result in a similar phenotype. Further examinations are necessary to clarify the abnormal phase behavior of the slow rod pathway signals. 

In summary, these studies show that mutations in GRM6 can lead to one form of arCSNB in patients with characteristic 15-Hz flicker ERG responses. This finding may imply the existence of more than the two reported rod pathways for signal transduction from photoreceptors to second-order neurons in the mammalian retina. Genes encoding other components involved in this pathway may provide additional candidates for arCSNB. 

²  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by the Velux Foundation, Switzerland (WB, CZ); Swiss National Science Foundation Grant 3100-067786 (WB); and the Hartmann Müller Foundation, Switzerland (CZ).

 Submitted for publication April 28, 2005; revised June 30, 2005; accepted September 15, 2005.

 Disclosure: C. Zeitz, None; M. van Genderen, None; J. Neidhardt, None; U.F.O. Luhmann, None; F. Hoeben, None; U. Forster, None; K. Wycisk, None; G. Mátyás, None; C. B. Hoyng, None; F. Riemslag, None; F. Meire, None; F.P.M. Cremers, None; W. Berger, 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.

 

Table 1.

 

Clinical Characteristics of the Patients with Presumed arCSNB

Patient	Sex	Age (y)	Visual Acuity			NYST	Refractive Error*			Dark-Adaptation Threshold Elevation (log units)
							OD	OS			OD	OS
21974	F	13	1.0	0.2	−	S − 9.0	S − 10.0	3
21973	F	10	0.1	0.1	+	S − 10	S − 8.0	Not done
26162	F	8	0.5	0.5	−	S − 6.5	S − 6.0	3
26151	F	15	0.2	0.3	+	S − 3.5	S − 2.5	3
26111	F	41	0.8	0.4	−	S − 14.0	S − 13.5	3
2496	M	36	0.4	0.4	+	S − 3.0	S − 3.0	2.5
2445	M	46	0.3	0.1	+	S + 3.5	S + 1.5	2.5

 Patient 21973 had Down syndrome, which also may be associated with nystagmus, high myopia, and low visual acuity. NYST, nystagmus.

*  Spherical equivalent.

Figure 1.

 

ISCEV Standard ERG of the father 26163, a patient with CSNB1, and one with arCSNB (patient 21974). Top to bottom: photopic response, 30-Hz photopic response, scotopic response, mixed response, oscillatory potentials (Gaussian window moving average filter, ς = 1.5 ms). The responses of the father could not be distinguished from our normative data (N = 67). In the subject with CSNB1 and the one with arCSNB, there was no recordable scotopic response, and the mixed response featured a minimal b-wave and showed the characteristic electronegative waveform. A remnant of a single oscillatory potential was found at approximately 33 ms, with an amplitude within normal limits. The photopic response showed a characteristic broad a-wave. The b-waves of the photopic responses were within normal limits in both patient groups. The 30-Hz photopic responses also had normal amplitudes and implicit times in both patient groups.

Figure 2.

 

Mutations in GRM6 cause autosomal recessive CSNB. Direct sequencing of the 10 coding exons of the GRM6 gene revealed five different mutations in three families with arCSNB. (A) Pedigree for arCSNB-bearing families 1, 2, and 3. Circles: females; squares: males; open symbols: unaffected; semifilled symbols: carrier; and filled symbols: affected. Patient identification numbers are indicated above the symbols, mutations below. The father in family 2 was not available for testing. (B) Electropherograms showing the five different mutations. Arrows: site of the mutation; star: common polymorphisms in the investigated sequences. Patient 2445 of family 1 had two compound heterozygous mutations: one missense mutation (c.137C→T) in exon 1 and one nucleotide insertion (c.720_721insG) in exon 3. The mother of this patient was heterozygous for the insertion, whereas the father carried the missense mutation. The unaffected sister of patient 2445 showed none of the alterations. In family 2, the male patient, 2496 revealed a homozygous missense mutation (c.1565G→A) in exon 8. The mother and the unaffected brother were heterozygous for this mutation, whereas the unaffected sister did not show the mutation. In family 3, the three affected siblings 26162, 21973, and 21974 carried a heterozygous (c.172G→C) and 19-nucleotide duplication (c.57_75dupl19). Both mutations were found in exon 1. The mother was also heterozygous for the duplication, whereas the father carries the p.Gly58Arg exchange.

Figure 3.

 

GRM6 mutations in the different regions of the glutamate receptor. The relative positions of the domains were predicted among the crystal structure of the rat Grm1. (A) The ligand-binding domains 1 (LB1; filled boxes) and 2 (LB2; dotted boxes) are shown. Unfilled boxes: protein regions not determined by crystallization; unfilled arrows: missense mutations; black arrows: frameshift mutations. The p.Pro46Leu and the p.Gly58Arg exchanges are located in the first ligand-binding domain, whereas the p.Cys522Tyr exchange is predicted to be in a highly conserved cysteine rich region. One frame shift identified in this study starts shortly before LB1 (Leu26fs), whereas the other is located in LB2 (Valfs243). (B) Schematic representation of the mGluR6 protein illustrating the extracellular bilobed glutamate-binding domains (LB1 and LB2) connected by a cysteine-rich region to the transmembrane domain with an intracellular C terminus. The circular object represents the ligand glutamate.

Figure 4.

 

Rod ERG responses to 15-Hz flicker stimulation obtained from a normal subject, a patient with CSNB1 and two with arCSNB (patient 26151 without and patient 21974 with a GRM6 mutation). Stimulus intensity was −1.56 log scot-td/s at maximum attenuation, reaching +0.44 log scot-td/s in nine steps. ERG signals to flicker intensities between −1.56 and −0.72 log scot-td/s were considered to represent primarily the slow rod pathway. In the patients with arCSNB without mutations in GRM6, the slow rod pathway signals could not be distinguished from noise, whereas in patients with GRM6 mutations the signals had large amplitudes and several reversing minima. In patient 21974, the responses of the right eye are plotted in gray, and those of the left eye in black.

Figure 5.

 

(A) Amplitudes of the 15-Hz component for the unaffected subjects (n = 10; filled symbols; bars ±SD, interindividual). Shaded symbols: amplitudes in the four patients with arCSNB with GRM6 mutations: 2496 S, 2445 R, 21974 E, and 2162 P. Only significant responses were plotted (S-N ratio >3, P < 0.01). Below −0.42 log scot-td/s most of the responses of the GRM6 subjects had an S-N ratio above 3. (B) Phases for the 15-Hz components in the unaffected subjects (filled symbols; bars ±SD, interindividual) and the patients with arCSNB with GRM6 mutations (shaded symbols). At intensities below −1 log scot-td/s, all patients showed an extra phase shift, apart from the phase shift at the intensity range at which unaffected subjects also showed a shift.

Supplementary Materials

Supplementary Table S1 - 59.6 KB (PDF) 

The authors thank the patients and their families for participation in this study and Barbara Kloeckener, Oscar Estevez, and Dick van Norren for critical comments on the manuscript.