Screen of the IMPDH1 Gene among Patients with Dominant Retinitis Pigmentosa and Clinical Features Associated with the Most Common Mutation, Asp226Asn

Abstract

purpose. To determine the frequency of mutations in IMPDH1 among patients with autosomal dominant retinitis pigmentosa (RP), to characterize the clinical features of patients with the Asp226Asn mutation in this gene, and to compare these features with those found among patients with selected dominant mutations in other RP genes.

methods. The coding sequence and the adjacent flanking intron sequences of all 14 coding exons were sequenced in 183 unrelated patients with dominant RP. The clinical findings evaluated included visual acuity, refractive error, visual field area measured with the Goldmann perimeter, final dark-adaptation threshold, full-field electroretinogram (ERG) amplitudes, cataract, and funduscopic bone spicule pigmentation.

results. The mutation Asp226Asn was identified in 6 of the 183 unrelated patients with RP. One patient carried the novel, possibly pathogenic, change Lys238Glu. There was approximately a 100-fold variation in ERG amplitudes among patients of similar age with the Asp226Asn mutation. Patients had similar reductions of rod-plus-cone 0.5-Hz ERG amplitude and cone 30-Hz ERG amplitude. For a given amount of remaining visual field, there was a larger ERG amplitude in IMPDH1-carrying patients (average 0.5-Hz ERG/visual field ratio = 9.5 nV/deg²) compared with groups of patients with the RP1 mutation Arg677End (2.8 nV/deg²), the rhodopsin (RHO) mutation Pro23His (5.1 nV/deg²), or the RHO mutation Pro347Leu (1.7 nV/deg²).

conclusions. IMPDH1 mutations account for approximately 2% of cases of dominant RP in North America. The most frequent mutation, Asp226Asn, appears to cause at least as much loss of rod function as cone function. Patients with this form of RP retain, on average, two to five times more ERG amplitude per unit of remaining visual area than patients with three other forms of dominant RP.

To date, at least 10 genes responsible for autosomal dominant retinitis pigmentosa (RP) have been identified (http://www.sph.uth.tmc.edu/Retnet/disease.htm/ RetNet is provided in the public domain by the University of Texas Houston Health Science Center, Houston, TX). One of these, IMPDH1 or RP10, is a gene on chromosome 7, region q31.3-q32 ¹ that encodes inosine 5′-monophosphate dehydrogenase type I, a ubiquitously expressed enzyme that catalyzes the rate-limiting step in the de novo synthesis of the nucleotide guanine. Guanine is of special importance to photoreceptors, because it can be converted to guanosine monophosphate (GMP) by the enzyme hypoxanthine-guanine phosphoribosyltransferase (HPRT), and rods and cones must rapidly adjust the concentration of cyclic guanosine monophosphate (cGMP) in their outer segments to respond properly to changes in light intensity. However, it is still unknown whether mutations in IMPDH1 cause RP by interfering with cGMP metabolism in photoreceptors. 

To date, there have been only three published investigations of the IMPDH1 gene among families with dominant retinitis pigmentosa. ^{2 3 4} The evidence implicating the IMPDH1 gene as a cause of dominant RP includes the identification of different missense mutations in different families, the observation that none of these mutations is found among normal controls, and the observation that the mutations perfectly cosegregate with RP. There is no reported comparison of the clinical features of patients with mutations in this gene versus those with mutations in other identified RP genes. In this study, we surveyed a large group of patients with dominant RP, mostly from the United States and Canada, to determine the types and frequency of IMPDH1 mutations. We also assessed the clinical features of patients with the most common mutation we identified, Asp226Asn, and compared the features with those in some other genetically defined forms of dominant RP. 

Materials and Methods

This study was performed in accordance with the tenets of the Declaration of Helsinki, was HIPAA compliant, and received institutional review board approval. Informed consent was obtained from the patients and normal control individuals before their donating a blood sample. Index patients came from 318 families from across North America. All families reported two consecutive generations with RP, and most had RP in three or more consecutive generations. The index patients were all unrelated to each other as determined through evaluations of pedigrees of each index patient. Most patients were previously evaluated for mutations in the rhodopsin (RHO), RP1, NRL, HPRP3, PRPF31, and PRPC8 genes, and many patients were evaluated for mutations in the RDS and ROM1 genes (Refs. ^{5 6 7 8 9} and unpublished results from the authors’ laboratory). We excluded 135 patients from the analysis of IMPDH1 because they had previously identified mutations in one of these genes. This left 183 unrelated patients for this study. However, some of the analysis of other dominant RP genes occurred concurrently with the present study, so that at the conclusion of this study, the set of 183 patients included 8 patients with identified mutations in the HPRP3 gene and 14 patients with mutations in the PRPF31 gene. Normal control individuals had no family history of RP and no known blood relatives with RP. 

Exons of the IMPDH1 gene with coding sequence (exons 4–17) were individually amplified using the polymerase chain reaction (PCR). The following primers were used to amplify exon 4 (sense/antisense, 5′ to 3′): GTCAGCAGTAGCAGCAGCAG and GCACCTAGGGGTACGAGACC. For exons 5 to 17, we used the primers reported by Bowne et al. ² Products of the PCR were directly sequenced (sequencer model 3100; Applied Biosystems, Inc., Foster City, CA). The numbering of the bases of the cDNA sequence in this article is according to GenBank accession number J05272 (http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). ¹⁰ The families of some index patients with IMPDH1 sequence variations were contacted, and family members were asked to participate in this study by donating a blood sample and having an ophthalmic examination. 

Ophthalmic examinations were performed as previously reported. ¹¹ Dark-adaptation thresholds after 45 minutes of dark adaptation were measured in a Goldmann-Weekers adaptometer with an 11° white test light projected either centrally or, if the patient’s visual field was sufficiently large, 7° below fixation. Kinetic perimetry was performed with the V4e white test light of the Goldmann perimeter, by moving the test light from nonseeing to seeing areas. Visual field areas were determined after plotting the fields with a desktop planimeter or by scanning images of the visual fields into a computer. Equivalent visual field diameters were calculated as twice the square root of the visual field area divided by π: 2(area/π)^1/2. 

Full-field ERGs were elicited by single flashes (0.5 Hz) of white light (0.2 cd/sec-m²) and 30-Hz white flashes of the same luminance in a Ganzfeld dome. ^{11 12 13} Responses were recorded with or without computer averaging. Amplitudes were measured from the trough of the a-wave (or from the baseline, if the a-wave was absent) to the peak of the b-wave for responses to 0.5-Hz light flashes, and from trough to peak for the responses to 30-Hz flashes. Nondetectable responses, defined as amplitudes <1.0 μV in response to 0.5-Hz light flashes or <0.05 μV in response to 30-Hz light flashes, were coded as 1 or 0.05 μV, respectively. The degree of reproducibility of submicrovolt signals in response to 30-Hz light flashes has been documented in the past. ^{12 14}  

For patients who were examined more than once, data from the initial visit were used unless some tests were not performed or were of poor quality at the first visit, in which case the results from the next available visit were used for those tests. Test results from both eyes were averaged. Visual acuities, visual field areas, and ERG amplitudes were converted to natural logarithms to better normalize their distributions. Clinical findings were compared with those reported previously by our group using identical techniques in sets of patients with the RP1 mutation Arg677End, ⁹ the rhodopsin (RHO) mutation Pro23His, ^{11 13 15} and the RHO mutation Pro347Leu. ^{13 15 16} These forms of RP were selected for this assessment because they are also dominantly inherited and because we had the clinical data available from a suitably large number of patients. To evaluate the effect of a genotype on ocular function, we performed multiple regression with each measure of ocular function as the dependent variable and mutation, age, and (for ERG analyses) refractive error (spherical equivalent) as the independent variables. Age and refractive error were included as covariates in the model because disease expression in RP varies with age and because ERG amplitude varies with refractive error. To test whether the measures of ocular function differed between patients with the mutation IMPDH1-Asp226Asn and patients with mutations in other genes, we adjusted the ocular function values according to the apparent effect of the covariates (age and refractive error) on all groups of patients combined within a multiple regression platform and calculated the statistical significance of any differences in the adjusted means, using the method of linear contrasts. ¹⁷  

Results

Genetic Analyses

Among 183 unrelated patients with dominant RP, 6 carried the mutation Asp226Asn (GAC to AAC; c.1276G→A). Affected and unaffected relatives of three of these index patients were evaluated for this change. In the three families, all affected relatives whose DNA was analyzed carried the change, whereas no unaffected relatives carried it (Fig. 1) . One index patient carried the novel missense change Lys238Glu (AAG to GAG; c.1312A→G). This change was not found among 92 normal control individuals. Segregation analysis in the family of this patient was limited because the parents of this index patient were deceased. The index patient’s three unaffected siblings were evaluated, and none of them carried this change (Fig. 1) . Although the change cosegregated with disease in the relatives whom we could evaluate, the small family size makes it possible that the cosegregation could have occurred by chance (with a likelihood of approximately 0.125), even if the change were nonpathogenic. 

A third missense change, His296Arg (CAC to CGC; c.1487A→G) was found in one index patient. This change did not cosegregate with dominant RP in the index patient’s family (Fig. 1) . Specifically, an affected uncle did not carry this change and an unaffected cousin did carry it. In addition, during a concurrent genetic analysis, this same index patient was found to carry the mutation Thr494Met in the HPRP3 gene, which did cosegregate with disease in his family (data not shown). The IMPDH1 mutation His296Arg was therefore interpreted not to be a cause of dominant RP. 

We also found 12 changes that we interpreted as polymorphisms or nonpathogenic rare variants. Figure 2shows the location of these changes, as well as the missense changes just mentioned, in a schematic intron–exon map of the gene. Two of the nonpathogenic changes were isocoding alterations involving codons Leu244 (CTG to CTC; c.1332G→C) and Ala440 (GCA to GCG; c.1920A→G). Ten were intron changes: IVS5+57G→A, IVS6+67G→A, IVS6+90C→T, IVS13+33C→T, IVS13+54delG, IVS13-10C→G, IVS14-114T→G, IVS15+33C→A, IVS15+43G→A, and IVS16+148(6-bp dupACCCTC). Table 1lists the frequencies of these alleles among our index patients. None of these changes is predicted to affect the amino acid sequence of the encoded protein. None of these intron changes appears to create or destroy any splice donor or acceptor sites, based on analysis with splice-site prediction software available at http://www.fruitfly.org/seq_tools/splice.html. ¹⁸  

Clinical Evaluation of Patients with the Asp226Asn Mutation

We clinically evaluated 23 patients from the 6 families with the Asp226Asn mutation. The ages of these patients ranged from 7 to 57 years. It was apparent that the symptoms of night blindness and subjective visual field loss increased in frequency with age, as did the frequency of cataracts and intraretinal pigment deposits (Table 2) . For example, among the 10 patients aged 21 years or less, none had cataracts, and only 3 had intraretinal pigment deposits in their fundi. In contrast, among the 13 patients older than 21 years, 11 had cataracts and 11 had intraretinal pigment deposits (P < 0.001 and P = 0.01, respectively, calculated by the Fisher exact test). There were no funduscopic features that appeared to distinguish patients with this mutation from those with typical RP. 

All patients who had ERG recordings had reduced but detectable rod-plus-cone ERG amplitudes in response to 0.5-Hz light flashes, and all but one patient (001-117, who was 17 years of age) had reduced but detectable cone ERG amplitudes in response to 30-Hz light flashes (Table 3) . The geometric mean amplitude in response to 0.5-Hz white flashes was 14.3 μV, 96% below the lower limit of normal (350 μV). The geometric mean amplitude in response to 30-Hz flashes was 4.3 μV, or approximately 91% below the lower limit of normal (50 μV). Because the rod-plus-cone responses to the 0.5-Hz flashes are typically dominated by the rod system, and because the responses to the 30-Hz flashes are exclusively derived from the cone system, these amplitude reductions suggest that there is at least as much loss of rod function as cone function. 

Variability in disease severity was observed among patients of comparable age. For example, patients 001-190, 001-117, and 001-170, who were close in age (16–19 years), had 0.5-Hz rod-plus-cone ERG amplitudes ranging from 2.8 to 164 μV and 30-Hz cone ERG amplitudes ranging from 0.5 to 59 μV (Table 3) . 

Comparison of Patients with IMPDH1 to Patients with RP1 or Rhodopsin Mutations

Table 4shows the mean refractive error (spherical equivalent), visual acuity, final dark adaptation threshold elevation, visual field area, 0.5-Hz ERG amplitude, and 30-Hz cone ERG amplitude in the 23 patients with the IMPDH1 mutation Asp226Asn (average age, 27 years) in comparison with 10 patients with the RP1 mutation Arg677End (average age, 41), 39 patients with the rhodopsin mutation Pro23His (average age, 40), and 26 patients with the rhodopsin mutation Pro347Leu (average age, 32). Except for refractive error, all measures have been regressed against age to correct for age differences between groups. The 0.5- and 30-Hz ERG amplitudes have also been regressed against refractive error, since ERG amplitudes decline with increasing axial myopia. ^{19 20} The severity of retinal degeneration in patients bearing IMPDH1-Asp226Asn, in comparison with patients with the other selected forms of dominant RP, varied according to each specific measure of visual function. For example, the age-adjusted elevation in the final dark-adaptation threshold, a measure of rod function, was similar in patients with IMPDH1 (1.6 log units) and RHO-Pro23His patients (1.3 log units; P = 0.32). However, patients with IMPDH1-Asp226Asn patients had mean 30-Hz ERG amplitude, a measure of cone function, significantly lower than that found in RHO-Pro23His patients (2.2 vs. 10.8 μV, respectively; P = 0.002 after adjusting for age and refractive error). In comparison with RHO-Pro23His patients, the IMPDH1-Asp226Asn patients also had worse age-adjusted visual acuity (20/47 vs. 20/26; P = 0.005) and smaller visual field area (633 vs. 4964 deg²; P < 0.001). Overall, IMPDH1-Asp226Asn-bearing patients had adjusted visual acuities comparable to those found in patients with RHO-Pro347Leu, dark-adaptation thresholds comparable to patients with RHO-Pro23His, and 30-Hz ERG amplitudes comparable to RP1-Arg677End-carrying patients (Table 4) . 

Patients with IMPDH1-Asp226Asn had the smallest mean visual field area of any group, yet they had the second highest ERG amplitudes among the four groups. This prompted us to calculate the log of the ratio of the ERG amplitude to the visual field area for each patient and to compare the mean values by mutation. The geometric mean ratios of 0.5-Hz ERG amplitude/field area (in nanovolts per square degree) for each group of patients were IMPDH1-Asp226Asn, 9.5; RP1-Arg677End, 2.8; RHO-Pro23His, 5.1; and RHO-Pro347Leu, 1.7 nV/deg². For comparison, the ratio in normal individuals without RP would be roughly 31 nV/deg², based on the lower limits of normal for the 0.5-Hz ERG amplitude (350 μV = 350,000 nV) and for the visual field area (11,300 deg² for a V4e test light) (350,000/11,300 = 31). The likelihood that the differences between the IMPDH1-Asp226Asn group and the other three groups occurred through chance alone, calculated using the Student’s t-test, were P = 0.02, P = 0.05, and P < 0.001, respectively. The geometric mean ratios of 30-Hz ERG amplitude/field area for each group were 2.8, 0.6, 1.8, and 0.5 nV/deg², respectively; the probabilities, comparing the IMPDH1 group with the other three groups were, respectively, 0.01, 0.22, and < 0.001. For reference, the 30-Hz ERG/visual field area ratio in normal individuals would be roughly 4.4 nV/deg² (lower limit of the 30-Hz amplitude = 50 μV = 50,000 nV; 50,000/11,300 = 4.4). The high ratios for the IMPDH1-Asp226Asn group were found even if we regressed the log ratio on mutation adjusting for age and (in the case of ERG) for refractive error (as listed above) or if we first calculated the ratios and then regressed the ratios against age and refractive error (data not shown). 

Discussion

We estimate that approximately 2% of dominant RP families in North America have an IMPDH1 mutation. This figure is based on our discovery of 6 unrelated patients with Asp226Asn of the 183 screened, and it takes into account the 135 unrelated patients excluded from the study because of previously identified mutations in other dominant RP genes: 6/(183 + 135) = 6/318 = 1.9% ≈ 2%. Although we did not screen patients for IMPDH1 mutations if they had identified mutations in other genes, it is unlikely that they would also have IMPDH1 mutations (i.e., they could be examples of digenic inheritance) because most previously identified dominant RP mutations cosegregate with clinically detectable disease without examples of unaffected carriers that would be expected if a mutation at a second contributory locus was necessary. Our estimate would not substantially change if we counted the index patient with the possibly pathogenic Lys238Glu allele (7/318 = 2.2% ≈ 2%). This proportion is less than a previously reported estimate that 5% to 10% of dominant RP families have IMPDH1 mutations. ² The previous study evaluated 62 North American families and found mutations in 3 (5%) or 4 (6%), depending on whether one counts the possibly pathogenic Val268Ile change that was found in one patient. However, two of the three families with the Asp226Asn mutation in that study were evaluated in part because previous work had shown linkage between the disease gene and markers linked to IMPDH1 in them. This ascertainment bias would tend to produce an overestimate of the prevalence of IMPDH1 mutations. Furthermore, it was not stated whether patients with identified mutations in other RP genes had been excluded from that study. ²  

We found a single family with the novel missense change Lys238Glu. Lys238Glu cosegregated with dominant RP in this family, but the results of the segregation analysis were not definitive because of the small size of the family. We were able to evaluate clinically only the index patient in this family, and it was thus not possible to assess statistically whether his clinical features were consistent with those found in the Asp226Asn patients. We have thus categorized Lys238Glu as an allele of uncertain pathogenicity. Bowne et al. ² reported another missense change, Val268Ile, in another North American family but segregation analysis was not reported, and we have tentatively categorized this change also as being of uncertain pathogenicity. A European study found the mutation Arg224Pro in one Spanish family, and a recent report from another American group found the mutation Arg231Pro in a family of Polish and Irish descent. ^{3 4} The cosegregation of Arg224Pro and Arg231Pro with RP in these large families is strong evidence in support of the authors’ conclusions that these mutations are pathogenic. In summary, our results, in combination with those reported by other groups, demonstrate three convincingly pathogenic mutations (Asp226Asn, Arg224Pro, and Arg231Pro) and two changes that are possibly pathogenic (Lys238Glu and Val268Ile). All five of these changes are in exon 10 (Fig. 2) . 

Our evaluation of 23 patients with the Asp226Asn mutation revealed ocular features typical of RP. Subjective night blindness and loss of visual field were reported mostly by those individuals over age 21, although some of the older individuals recalled that these symptoms began at an earlier age. Cataracts and bone-spicule pigment deposits in the fundus were found more often in those over age 21. The apparently age-related occurrence of symptoms and signs of retinal degeneration and the delayed cone ERG b-wave implicit times in most cases probably reflect a progressive photoreceptor degeneration. ²¹  

Although the subjective symptoms of disease in young patients were few and mostly mild, measures of photoreceptor function indicated that the disease begins at an early age. For example, our youngest patient, age 7, had a 1.5-log-unit elevation of his final dark-adaptation threshold, indicating compromised rod function, despite the lack of the subjective symptom of night blindness. Furthermore, 0.5- and 30-Hz ERG amplitudes indicated a severe reduction of photoreceptor function in some patients under age 20. The ocular abnormalities that we observed in patients with the IMPDH1 mutation Asp226Asn are in accord with those briefly described in patients with the mutation Arg224Pro. ^{3 22} They appear to be less severe than those in patients with the mutation Arg231Pro. ⁴  

Averaging over all cases, there were similar reductions of 0.5- and 30-Hz ERG amplitudes, indicating that both rods and cones are affected in this disease. Similar reductions in rod and cone ERG amplitudes were present in one previously reported patient ² with IMPDH1-Asp226Asn. However, there is considerable variation in the relative reduction of rod versus cone ERG amplitude among patients with various forms of RP, and we doubt that the similar reduction in rod and cone ERG amplitudes by itself would be useful in identifying patients with the mutation IMPDH1-Asp226Asn. However, our comparison of five measures of visual function (visual acuity, dark-adaptation threshold elevation, visual field area, 0.5- and 30-Hz ERG amplitude) in patients with mutations in the IMPDH1, RP1, and RHO genes found a pattern of abnormalities that, on average, appeared to separate IMPDH1-RP from each of the other forms of dominant RP (Table 4) . The IMPDH1-Asp226Asn-bearing patients were closest to those with RHO-Pro347Leu with regard to visual acuity, closest to those with RHO-Pro23His with regard to the final dark-adaptation threshold, and closest to those with RP1-Arg677End with regard to 30-Hz cone ERG amplitude. 

Among the four genetically defined forms of dominant RP compared in this study, there was variation in the ratio of ERG amplitudes to visual field area, with the patients carrying IMPDH1-Asp226Asn having the largest ratios on average. This suggests that the regions of the retina contributing to the residual visual field generate on average a larger ERG in patients with IMPDH1-Asp226Asn than in patients with defects in the other genes. Based on this pattern and the other ocular features found in the patients with IMPDH1-Asp226Asn, this mutation should be suspected as a possible etiology in patients with RP in their 20s who have a moderately reduced visual acuity (e.g., 20/40–20/50), a small remaining visual field (i.e., <700 deg² in area, corresponding to a visual field diameter less than ∼30°), and ERG amplitudes that are unusually high relative to the visual field area. With our ERG techniques, this would correspond to an ERG/visual field ratio greater than approximately 8 nV/deg² for the 0.5-Hz rod-plus-cone ERG and approximately 2.5 nV/deg² for the 30-Hz cone ERG. A determination of the predictive power of these criteria will necessitate clinical analysis of additional genetically defined forms of RP. 

It should also be noted that the pattern of abnormalities in the visual function parameters are based on means. They are not likely to be sufficiently specific to allow by themselves the genetic diagnosis of patients with the IMPDH1-Asp226Asn mutation, primarily because there is also a large variation (up to approximately two orders of magnitude) in the measures of disease severity even among patients of comparable age (Tables 3 4) . This was true, not only for the IMPDH1 patients, but also for the rhodopsin ¹¹ and RP1 ⁹ patients. The considerable variation in disease severity, even among patients with the same primary mutation, indicates that other genetic or environmental factors play a major role in this disease. It remains unknown whether the same modifying factors affect the course of RP regardless of the primary mutation. 

The three definite IMPDH1 mutations (Arg224Pro, Asp226Asn, and Arg231Pro) reported in this article and in previous reports, ^{2 3 4} as well as the two changes that are possibly pathogenic (Val268Ile and Lys238Glu), are all missense changes. They all affect amino acid residues within a 45-residue stretch in the middle of the primary protein sequence. No obvious null alleles (e.g., nonsense, frameshift, or splice-site mutations) have been uncovered in our screen or in the previously reported screens of patients with dominant RP. The apparent clustering of the dominant missense mutations between residues 224 and 268, inclusive, and the lack of obvious null alleles suggest that these alleles cause RP through a gain-of-function mechanism rather than through haploinsufficiency. Transgenic mice lacking a functional IMPDH1 gene have been described. Their retinas show no histologic evidence of retinal degeneration, although ERG amplitudes are slightly decreased at 12 months of age. ^{23 24} Studies of transgenic mice carrying dominant IMPDH1 alleles, such as those associated with RP in humans, have not been reported. 

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

 Supported by NIH Grants EY08683 and EY00169 and the Foundation Fighting Blindness.

 Submitted for publication October 8, 2004; revised January 9, 2005; accepted January 14, 2005.

 Disclosure: Y. Wada, None; M.A. Sandberg, None; T.L. McGee, None; M.A. Stillberger, None; E.L. Berson, None; T.P. Dryja, 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.

 

Schematic pedigrees of families demonstrating transmission of the IMPDH1 alleles Asp226Asn (labeled as the A allele in families 2366, 6457, and 6455), Lys238Glu (the B allele in family 1395), and His296Arg (the C allele in family 6789). At the top left corner of each schematic pedigree is the laboratory’s family identification number. For those family members who were evaluated, the results of IMPDH1 analysis are shown beneath their respective symbols: squares, males; circles, females; solid symbols, affected with RP; open symbols, unaffected; diamond, multiple individuals; +, wild-type sequence at the site of the base change in the index patient in the respective family. Arrow: the index case in each family.

Figure 2.

 

Map of the 17 exons of the IMPDH1 gene showing the locations of the DNA sequence changes. All changes were detected in this study except Arg224Pro, Arg231Pro, and Val268Ile which were reported by other groups. ^{2 3} The changes in boxes are interpreted as causes of dominant RP; those in italics and underlined are of uncertain pathogenicity. All other changes are interpreted as nonpathogenic rare variants or polymorphisms. Regions of exons in black are the 5′ and 3′ noncoding regions. The lengths of the exon boxes are in proportion to the lengths of their DNA sequences. Introns are not drawn to scale. At the top right corner of each box for exons 4 to 17 is the number of the codon at the 3′ end of that exon.

Table 1.

 

Allele Frequencies of DNA Sequence Changes Observed in IMPDH1

Change	Location	Wt:Het:Homo, †	Minor Allele Frequency
IVS5+57G→A	Intron 5	182:1:0	0.003
IVS6+67G→A	Intron 6	182:1:0	0.003
IVS6+90C→T	Intron 6	181:2:0	0.005
Asp226Asn *	Exon 10; c.1276G→A	177:6:0	0.016
Lys238Glu *	Exon 10; c.1312A→G	182:1:0	0.003
Leu244Leu	Exon 10; c.1332G→C	138:42:3	0.131
His296Arg	Exon 11; c.1487A→G	182:1:0	0.003
IVS13+33C→T	Intron 13	137:41:5	0.139
IVS13+54delG	Intron 13	171:12:0	0.033
IVS13-10C→G	Intron 13	182:1:0	0.003
IVS14-114T→G	Intron 14	166:16:1	0.049
Ala440Ala	Exon 15; c.1920A→G	106:61:16	0.254
IVS15+33C→T	Intron 15	177:6:0	0.016
IVS15+43G→A	Intron 15	164:19:0	0.052
IVS16+148(6-bp dupACCCTC)	Intron 16	182:1:0	0.003

*  Asp226Asn is in bold to indicate that it is considered a cause of ADRP. Lys238Glu is in italics because it is of uncertain pathogenicity. All other changes are considered nonpathogenic polymorphisms or rare variants.

†  Wt:Het:Homo, the numbers of index patients who were homozygous for the wild-type sequence, who were heterozygotes, and who were homozygous for the uncommon sequence.

Table 2.

 

Ocular Findings in Patients with the IMPDH Mutation Asp226Asn

Patient	Family	Age	Sex	Age at Onset of Night Blindness	Age at Onset of Field Loss	Lens Opacities*	Bone-spicule Pigment, †
001–275	2366	7	M	, ‡	6	−/−	−/−
15564	2366	8	M	, ‡	, ‡	−/−	−/−
3066	3036	9	M	3	, ‡	−/−	+/+
001–171	2366	11	F	, ‡	, ‡	−/−	−/−
001–540	2366	12	F	, ‡	, ‡	−/−	−/−
001–190	2366	16	M	, ‡	10	−/−	+/+
001–117	2366	17	F	, ‡	30, §	−/−	−/−
001–170	2366	19	F	, ‡	, ‡	−/−	+/+
5033	2366	20	M	, ‡	, ‡	−/−	−/−
001–177	2366	21	M	, ‡	, ‡	−/−	−/−
001–168	6457	25	F	14	18	+/+	+/+
001–127	6455	25	M	12	12	+/+	+/+
001–154	2366	26	F	, ‡	, ‡	−/−	+/+
001–149	2366	29	F	, ‡	6	−/−	+/+
001–005	6787	31	M	22	, ‡	+/+	+/+
3036	3036	34	M	0	15	+/+	+/+
001–163	2366	36	M	, ‡	20	+/+	+/+
001–394	2366	36	M	28	28	+/+	+/+
001–192	2366	37	M	24	24	+/+	+/+
001–273	3923	41	F	, ‡	, ‡	+/+	−/−
038–060	3923	42	M	4	20	+/+	−/−
001–278	3036	54	M	15	15	+/+	+/+
001–288	3923	57	M	55	40	+/+	+/+

*  Coded as positive if posterior subcapsular cataract was present in a given eye; data presented as OD/OS.

†  Coded as positive if patients had bone-spicule pigment in at least one quadrant of a given eye; data presented as OD/OS.

‡  Asymptomatic in this regard.

§  Patient 001–117, first seen at age 17, reported subjective field loss at a later visit.

Table 3.

 

Refractive Error and Ocular Function in Patients with the IMPDH Mutation Asp226Asn

Patient	Family	Age	Sex	Snellen Visual Acuity	Refractive Error (Spherical Equivalent)*	Final Dark Adaptation Threshold Elevation, †	Visual Field Area (deg2), ‡	Full-Field Electroretinograms (ERGs), §
																0.5-Hz Amplitude (μV)	30-Hz Amplitude (μV)	30-Hz Implicit Time (ms)
001–275	2366	7	M	20/25	−0.25	1.5	298	4.6	1.3	36
15564	2366	8	M	20/40	1.62	1.0	8189	44.0	25.0	31
3066	3036	9	M	20/40	2.38	0.7	5783	23.0	12.0	25
001–171	2366	11	F	20/25	0.62	0.5	N/A	88.2	32.3	31
001–540	2366	12	F	20/25	−0.75	1.0	5959	3.9	0.5	40
001–190	2366	16	M	20/35	−1.25	0.5	896	11.2	2.2	40
001–117	2366	17	F	20/25	−0.88	0.0	6566	164.0	59.0	32
001–170	2366	19	F	20/22	−1.56	2.0	228	2.8	0.5	44
5033	2366	20	M	20/30	−0.06	1.8	1525	16.8	7.8	38
001–177	2366	21	M	20/27	−0.62	0.0	6784	79.3	29.6	34
001–168	6457	25	F	20/30	−3.31	3.0	487	1.9	0.5	36
001–127	6455	25	M	20/30	−3.88	2.0	4981	5.0	0.9	44
001–154	2366	26	F	20/25	0.12	0.3	3942	71.0	21.0	35
001–149	2366	29	F	20/30	−1.50	0.5	3260	67.6	27.9	37
001–005	6787	31	M	20/37	−1.06	1.5	499	2.7	1.2	38
3036	3036	34	M	4/400	−0.25	3.7	190	N/A	N/A	N/A
001–163	2366	36	M	20/40	0.19	1.0	N/A	23.5	7.4	37
001–394	2366	36	M	20/45	−2.12	3.5	104	1.0	0.4	57
001–192	2366	37	M	20/30	0.31	2.5	124	3.3	0.5	43
001–273	3923	41	F	20/20	−5.31	0.0	2019	85.0	27.0	30
038–060	3923	42	M	20/30	−2.94	0.0	1878	94.1	27.9	32
001–278	3036	54	M	20/45	−6.62	1.9	394	N/A	N/A	N/A
001–288	3923	57	M	20/37	−4.56	1.0	208	N/A	N/A	N/A

 Data are the averages between the two eyes. N/A, data not available.

*  Spherical equivalent = sphere + one-half cylinder.

†  Log threshold above normal after 45 minutes of dark adaptation.

‡  Equivalent circular diameter derived from visual field area to a V-4e test light expressed in square degrees; normal, ≥11,300 deg² (equivalent to a 120° diameter field).

§  Normal 0.5-Hz rod-plus-cone ERG amplitude ≥350 μV, with the mean in 16 normal adults under age 56 years, 575 μV; normal 30-Hz cone ERG amplitude ≥50 μV, mean in normal adults, 91 μV; 30-Hz cone ERG implicit time ≤32 ms; mean in normal adults, 30 ms.

Table 4.

 

Average Visual Function in Patients with IMPDHI, RP1, and RHO Mutations

Mutation	n	Refractive Error (Spherical Equivalent)	Ln Visual Acuity, # (Snellen Equivalent)	Log Dark-adapted Threshold Elevation, #	Ln Visual Field Area, # , †	Ln 0.5-Hz ERG Amplitude, ‡ , §	Ln 30-Hz ERG Amplitude, ‡ , §	ERG/Field Area, ∥
																0.5 Hz	30 Hz
IMPDH1-Asp226Asn	23	−1.38 ± 0.46	−0.86 ± 0.16	1.6 ± 0.20	6.45 ± 0.29	2.17 ± 0.33	0.78 ± 0.39	9.5 nV/deg2	2.8 nV/deg2
			(20/47)		(633 deg2)	(8.8 μV)	(2.2 μV)	[0.6–50.3]	[0.1–14.8]
RPI-Arg677End	10	−1.23 ± 0.44	−0.11 ± 0.24*	2.4 ± 0.33*	7.33 ± 0.41	1.30 ± 0.46	0.15 ± 0.52	2.8 nV/deg2 *	0.6 nV/deg2 *
			(20/22)		(1525 deg2)	(3.7 μV)	(1.2 μV)	[0.5–18.3]	[0.1–4.8]
RHO-Pro23His	39	−0.60 ± 0.35	−0.25 ± 0.12, **	1.3 ± 0.16	8.51 ± 0.21, ***	3.29 ± 0.23*	2.38 ± 0.27, **	5.1 nV/deg2 *	1.8 nV/deg2
			(20/26)		(4964 deg2)	(26.8 μV)	(10.8 μV)	[0.4–28.7]	[0.1–7.8]
RHO-Pro347Leu	26	−0.33 ± 0.31	−0.99 ± 0.14	3.2 ± 0.20, ***	7.19 ± 0.26	0.87 ± 0.29, **	−0.29 ± 0.34*	1.7 nV/deg2 , ***	0.5 nV/deg2 , ***
			(20/54)		(1326 deg2)	(2.4 μV)	(0.7 μV)	[0.3–18.5]	[0.0–5.2]

 Entries in parentheses are the geometric means (i.e., they are the antilogs of the logarithmic values); the geometric mean of the ln visual acuity is shown as the Snellen fraction. Entries in brackets in the columns of ERG/Field Area are the ranges of values found in each group of patients.

*  0.05 > P > 0.01 in comparison with IMPDH1-Asp226Asn.

**  0.01 ≥ P > 0.001 in comparison with IMPDH1-Asp226Asn.

***   P ≤ 0.001 in comparison with IMPDH1-Asp226Asn.

†  Normal visual field area is ≥11,300 deg².

‡  Data adjusted for age and refractive error.

§  Normal 0.5-Hz and 30-Hz ERG amplitudes are ≥350 μV and ≥50 μV, respectively.

∥  Each patient’s ratio was calculated individually after the clinical values were converted to logarithms to approximate a Gaussian distribution. The values listed are the geometric means of those ratios.

#  Data adjusted for age.