Alward WLM, Kwon YH, Khanna CL, Johnson AT, Hayreh SS, Zimmerman MB, Narkiewicz J, Andorf JL, Moore PA, Fingert JH, Sheffield VC, Stone EM. Variations in the Myocilin Gene in Patients With Open-Angle Glaucoma. Arch Ophthalmol. 2002;120(9):1189-1197. doi:10.1001/archopht.120.9.1189
EDWIN M.STONEMD, PHD
To determine the prevalence and associated phenotype of myocilin (MYOC) coding sequence variations and a specific promoter polymorphism(MYOC.mt1) in patients with glaucoma and glaucoma suspects.
A consecutive, unselected series of 779 patients (652 with open-angle glaucoma and 127 glaucoma suspects) were recruited from a university medical center and clinically characterized. The coding sequences of the MYOC gene and the MYOC.mt1 locus in the promoter region were screened for sequence variations. We determined the prevalence of MYOC coding sequence mutations and the MYOC.mt1 promoter polymorphism. We also compared the clinical features of individuals with and without mutations and the MYOC.mt1 promoter polymorphism.
Plausible disease-causing sequence variations (DCVs) in the MYOC gene were found in 3.0% of the entire group. Such variations were found in patients with most forms of open-angle glaucoma studied. Patients with primary open-angle glaucoma (POAG) who harbored coding sequence DCVs were clinically similar to patients without them. Patients who harbored the rarer allele of the MYOC.mt1 promoter polymorphism were no different in any measure of disease severity from those who harbored the more common allele.
MYOC DCVs were found in approximately 3% of patients with glaucoma and glaucoma suspects. The 2 alleles of the MYOC.mt1 promoter polymorphism were equally distributed among patients with POAG and healthy control subjects. Patients with POAG who harbored the rarer allele of the MYOC.mt1 promoter polymorphism were no different from those with the more common variant in any measure of disease severity.
Testing for the MYOC.mt1 promoter polymorphism appears to be of no value in the evaluation of patients with glaucoma.
GLAUCOMA IS the leading cause of permanent blindness in the world. It is estimated that 66.8 million people have glaucoma and that 6.7 million are bilaterally blind.1 In the United States, glaucoma is the second-leading cause of permanent blindness and the leading cause among African Americans.2 Primary open-angle glaucoma(POAG) is by far the most prevalent form of glaucoma in the United States. A significant fraction of POAG is heritable.3- 5
Despite the high prevalence and heritable nature of adult-onset POAG, identification of the causative genes has been difficult. The late onset of the disease and its often subtle clinical findings result in relatively limited pedigrees of living affected individuals being available for study. However, families with juvenile-onset POAG have provided an avenue to investigate the molecular genetics of adult-onset POAG. Autosomal dominant juvenile-onset POAG is an uncommon form of glaucoma that shares clinical features with the adult-onset version except for having an early age of diagnosis, very high intraocular pressure (IOP), and autosomal dominant inheritance with high penetrance.3,6 The study of a large multigenerational pedigree of a family with autosomal dominant juvenile-onset POAG led to the linkage of this disease to the long arm of chromosome 1 (1q21-1q31).7- 9
In 1997, Stone et al10 identified mutations in a gene in the disease interval that segregated with the disease phenotype. Because the expression of this gene could be induced by applying topical corticosteroids to cultured trabecular meshwork cells, the protein encoded by this gene was called the trabecular meshwork-inducible glucocorticoid response protein (TIGR).11 However, the expression of this gene has also been detected in the photoreceptors of the retina, and these investigators suggested the name myocilin because the protein had a sequence that was homologous to Dictyostelium discoideum myosin.12 The myocilin gene (MYOC)encodes a 57-kd protein. The expression of the gene has been demonstrated in the trabecular meshwork, ciliary body, retina, optic nerve, and iris of the eye.12- 14MYOC production has also been detected in 17 of 23 other organs.13 Despite extensive study, the normal function of the MYOC protein is not yet known.
Stone and colleagues found MYOC mutations not only in families with autosomal dominant juvenile-onset POAG but also in patients with adult-onset POAG.10 Mutations were found in 10 (4.4%) of 227 unrelated individuals with a family history of glaucoma and 3 (2.9%) of 103 unselected patients with adult-onset POAG. Many MYOC mutations have now been found in populations around the world.15- 18 In a large study by Fingert et al,19 the DNA from 1703 patients with POAG from 5 populations was screened for MYOC mutations. These populations included primarily white patients from the United States, Australia, and Canada, African American patients from the United States, and Asian patients from Japan. The prevalence of mutations in these diverse populations was similar (2.6%-4.3%).
Although MYOC was the first open-angle glaucoma gene to be identified in humans, it is no longer the only one. In 2002, Rezaie et al20 discovered another gene, optineurin, in the GLC1E interval on chromosome 10p and identified variations in this gene in patients with POAG and normal-tension glaucoma (NTG).
Our study was designed to evaluate the prevalence of MYOC mutations in a large consecutive, unselected series of patients with a variety of open-angle glaucomas and to determine whether there were any phenotypic differences between patients with adult-onset POAG who harbor MYOC mutations and those who do not. During the analysis of these data, a report appeared by Colomb et al21 suggesting that a common polymorphism located in the promoter region of the MYOC gene was associated with a poor response to glaucoma therapy. This polymorphism is located 1000 base pairs upstream from the start of the gene's coding sequence. Colomb and colleagues found no difference in the prevalence of this polymorphism between patients with POAG and controls(17.6% and 16.0%, respectively). Notwithstanding this lack of association with the actual diagnosis of glaucoma, they hypothesized that the presence of this polymorphism might predict the responsiveness of patients with glaucoma to treatment. Patients with this polymorphism had a higher IOP when they enrolled in the study than those without the polymorphism even though the 2 groups had a similar IOP when they were first diagnosed. Colomb and colleagues claimed that this difference indicated an unresponsiveness to glaucoma treatment. It is important to evaluate this hypothesis critically and independently because a commercial assay for the polymorphism has been advertised to the health care community.
The Human Subjects Review Committee of the University of Iowa (Iowa City) approved this project. Written informed consent was obtained from each study participant. During the course of 9 months, all patients at the clinics of 4 physicians with a special interest in glaucoma (W.L.M.A., Y.H.K., A.T.J., and S.S.H.) were evaluated for inclusion in this study. Patients were considered to have glaucoma if they had a cup-disc ratio greater than 0.7 with glaucomatous visual field loss on either Goldmann or Humphrey perimetry. Patients with smaller cup-disc ratios or optic nerve head changes alone (without visual field loss) could be included if there was documented progression of optic nerve head cupping. Elevated IOP was not required for the diagnosis of glaucoma.
A standard form recorded the clinical characteristics of the patients. Patients were questioned and their records were reviewed to determine age, sex, race, age at diagnosis, family history of glaucoma, medications used, prior laser and incisional surgical procedures, and maximum IOP. The clinical data that were collected included cup-disc ratio (an average of the horizontal and vertical cup-disc ratios was used if the cup was not round), automated static threshold perimeter (Humphrey) mean deviation and corrected pattern SD, and manual kinetic perimeter (Goldmann) visual field loss. A simple 5-point scale was devised for Goldmann visual field loss, which ranged from 0 for a normal field to 4 for severe field loss and 5 for no light perception. Goldmann perimetry may be used more commonly in our institution than in others. In part, this is because we have 6 perimetrists with 10 to 30 years of experience who are experts in Goldmann perimetry. Goldmann perimetry was used for individuals who had received follow-up with this device since before the advent of automated perimetry, who had extensive visual field loss, or who were unable to undergo highly reliable automated testing.
Patients with open-angle glaucoma were categorized into 4 groups: those with POAG, exfoliative glaucoma, pigmentary glaucoma, and corticosteroid-induced glaucoma. Glaucoma suspects were also categorized into 4 groups: those with ocular hypertension (OHT), exfoliation syndrome without glaucoma, pigment dispersion syndrome without glaucoma, and corticosteroid-induced OHT. These categories were selected, and the clinical features to be ascertained were decided on prior to obtaining any clinical or molecular information. The physicians had no knowledge of the molecular genetic status of the patients during the collection of the data.
Patients were considered to have POAG if they had the optic nerve and visual field features of glaucoma with no evidence of an identifiable cause for elevated IOP. Patients with POAG were evaluated as a group and divided into 3 subgroups: those with adult-onset POAG, juvenile-onset POAG, and NTG. Patients with adult-onset POAG were 40 years and older and had an IOP higher than 21 mm Hg. Those with juvenile-onset POAG were younger than 40 years and also had an IOP higher than 21 mm Hg. Patients with NTG had an IOP of 21 mm Hg or lower. Although the division into these 3 subgroups based on IOP and age was arbitrary, it was done for 2 reasons. First, patients are often grouped into these categories in clinical practice. Second, most reported MYOC mutations cause glaucoma that is characterized by a high IOP and an early age of onset. Therefore, it seemed reasonable to separate patients who were young and had an elevated IOP from older patients and those with a normal IOP.
Patients with exfoliative glaucoma met the criteria for POAG but also had typical fibrillar exfoliative material on the lens capsule. Those with pigmentary glaucoma met the criteria for POAG with the addition of characteristic iris transillumination defects and dense black pigment in the trabecular meshwork. The corticosteroid-induced group had normal anterior segments but a history of elevated IOP in response to topical corticosteroid use.
We also obtained DNA from patients with open angles who were at risk for developing glaucoma but did not have optic nerve head damage or visual field loss. This included patients with OHT (IOP > 21 mm Hg with no identifiable cause of elevated IOP), exfoliation syndrome with or without OHT, pigment dispersion syndrome with or without OHT, and corticosteroid-induced OHT.
Patients were excluded if they had any other secondary glaucoma caused by neovascularization of the iris, inflammation, trauma, or surgery prior to the diagnosis of glaucoma. If a known relative was already included in the study, the patient was excluded. All eligible patients were offered participation in this study. When patients met the entry criteria but refused venipuncture or were unable to provide a blood specimen, clinical data were gathered to determine whether the study group was biased. Care was taken to include all eligible patients without biasing the sample toward or against those with a family history. Therefore, if a participant had been tested before, either as a member of a family or as one of the previously described patients with POAG,10,15,19 we filled out the standardized clinical data collection form but did not obtain a second blood sample. We obtained DNA from venous blood as previously described.10
The coding sequence of the MYOC gene was screened for variations using single-strand conformational polymorphism (SSCP) analysis followed by bidirectional automated DNA sequencing as previously described.19 All amplimers were analyzed in standard fashion in a single laboratory. The polymerase chain reaction amplification products were denatured for 3 minutes at 94°C and then underwent electrophoresis on 6% polyacrylamide, 5% glycerol gels at 25 W for approximately 3 hours. The gels were then stained with silver nitrate.22 All SSCP gels were independently scored by a minimum of 2 experienced investigators. Amplimers showing a band shift were reamplified and sequenced bidirectionally using an automated sequencer (ABI 337; Applied Biosystems, Foster City, Calif) and dye terminator chemistry. The MYOC.mt1 polymorphism was assayed using SSCP after polymerase chain reaction amplification with the following primers: forward, TGTGAATTTGAATGAGGAAAAA; reverse, GCAGGAGGTCTAATTTCAA. The annealing temperature was 55°C and lasted for 1 minute. The presence of the least common allele (a guanine residue on the sense strand 1000 base pairs upstream from the start of transcription) was confirmed by digestion with the restriction enzyme AlwNI at 37°C for 2 hours.
In the absence of an in vivo functional assay of a gene's product, the pathogenicity of each observed variation is usually inferred by assessing both the predicted effect of the variation on the structure of the gene product and the distribution of the variant among patient and control populations.13,15,19 For this study, coding sequence changes were assumed to be DCVs if they (1) would be expected to alter the amino acid sequence of the MYOC protein, and (2) were more commonly observed in patients with glaucoma than in controls. To meet the latter criterion, a variant needed to be completely absent from the control population or significantly more common (P<.05 using the Fisher exact test) in the glaucoma population. Sequence changes were considered to be non–disease-causing polymorphisms if they did not result in a change in the MYOC amino acid sequence or if the change was seen equally in the control and glaucoma populations.
There were 92 control subjects without known eye abnormalities recruited from the same population. These subjects were older than 40 years (mean ± SD age, 61.3 ± 13.4 years; range, 43-92 years) with no history of glaucoma, no family history of glaucoma, and an IOP of 21 mm Hg or lower. They had been recruited for a previous study without any knowledge of their MYOC status and were found to be free of DCVs in the MYOC gene.10 Control patients were used in this study to estimate the prevalence of the rarer allele of the MYOC.mt1 promoter polymorphism in the general population.
Patient demographics were compared between those with and without mutation and among the polymorphism types using the Fisher exact test for the categorical variables of sex and family history, the 2-sample ttest or 1-way analysis of variance for age at diagnosis, and the log-rank test for age at first treatment. For the continuous and ordinal variables that were measured for both eyes, the comparison between those with and without mutation and among the polymorphism types was done using mixed-effects model analysis, which accounted for the correlation between eyes from the same individual.23,24 In this study the standard mixed-effects model, which requires a normal data distribution, could not be used because the IOP, cup-disc ratio, and other continuous variables were not normally distributed. As an alternative, a nonparametric rank test for mixed models was used.25 Because these variables had a skewed distribution, the median and interquartile range were computed instead of the mean and SD. For the comparison involving the incidence of trabeculoplasty in the eye, we used logistic regression analysis with the generalized estimating equation method.26 This analysis accounts for the correlation of dichotomous outcomes between eyes from the same person. All statistical analyses were performed using SAS version 8.2 statistical software (SAS Institute Inc, Cary, NC). Statistical significance was defined as P<.05. Statistically significant P values were corrected with the Bonferroni test for multiple comparisons.
There were 809 patients who met the entry criteria. Of these, 23 (2.8%) refused venipuncture, and in an additional 7 (0.9%) we were unable to obtain a satisfactory blood sample. There were no significant differences in diagnosis, age, sex, race, or family history between the patients who failed to enroll in the study and those who did participate (data not shown).
Table 1 indicates the numbers of patients with the various forms of glaucoma and the prevalence of plausible DCVs in the MYOC gene. For the 524 patients with POAG, 17 (3.2%) had a DCV in the MYOC gene. Disease-causing variations were found in 13 (3.3%) of 393 individuals with adult-onset POAG. The juvenile-onset POAG group had roughly twice the prevalence of MYOC DCVs as the adult-onset POAG group, with 3 (6.4%) of 47 patients having a DCV. Patients with NTG had about one third the prevalence of MYOC DCVs of patients with POAG, with 1 (1.2%) of 84 patients exhibiting a DCV. None of these differences reached statistical significance. For every diagnosis except corticosteroid-induced glaucoma and corticosteroid-induced OHT, at least 1 patient harbored a DCV in the MYOC gene. Of the 4 glaucoma suspects with MYOC changes, 3 had elevated IOP. The 1 patient who had exfoliation syndrome without glaucoma and a MYOC DCV was aged 66 years and had never had a recorded IOP higher than 18 mm Hg.
Table 2 lists the DCVs and non–disease-causing polymorphisms found in these subjects. A total of 23 patients had DCVs; 9 different DCVs were identified, 7 of which have been described in prior reports.15,19 The 162Ins163 and Asp208Glu variations have not been previously reported. As in prior studies, the Gln368Stop variant was the most common DCV seen. This change was found in 12 patients: 9 with adult-onset POAG, 1 with juvenile-onset POAG, 1 with OHT, and 1 with pigment dispersion syndrome without glaucoma. Ten coding sequence polymorphisms (non–disease-causing) were identified. All of these have previously been described.15
Only the adult-onset POAG group had enough patients with a DCV to compare the phenotypes of patients with these variations and those without. Table 3 compares the 13 patients with adult-onset POAG and DCVs in the MYOC gene with the 380 patients without such changes. There was no significant difference in age at diagnosis, sex, or family history of glaucoma. Race information is not included in the table. All patients with DCVs were white, as were the vast majority (94.8%) of those without DCVs. This reflects the racial mix of the referral area. There was no difference in age at first treatment, cup-disc ratio, visual field loss, or number of medications used. Patients with DCVs had a higher median peak IOP (P = .02). Those with DCVs were also more likely to have undergone laser trabeculoplasty than those without MYOC sequence variations (P =.04) but were no more likely to have required incisional surgery. Although P<.05 for the individual tests for IOP and trabeculoplasty, considering that 12 variables were evaluated, the Bonferroni-adjusted values for these 2 variables were P = .25 for peak IOP and P = .48 for trabeculoplasty.
The 393 patients with adult-onset POAG and 92 healthy subjects were also examined for the MYOC.mt1 promoter polymorphism (Table 4). None of the healthy subjects exhibited a DCV in the MYOC coding sequence. The rarer allele of the MYOC.mt1 polymorphism was found in 61 (15.5%) of 393 patients with adult-onset POAG and 22 (23.9%) of 92 healthy subjects. This difference was not statistically significant (P = .12).
The demographics and clinical features of patients who had adult-onset POAG with and without the MYOC.mt1 promoter polymorphism are presented in Table 5. There were no statistically significant differences in any of the measured features of disease severity between the 2 populations.
In the study by Colomb et al,21 most of the data indicating an association between glaucoma severity and the MYOC.mt1 promoter polymorphism could be traced to 13 women who had glaucoma with the rarer allele of the polymorphism; therefore, we looked at women separately. Table 6 lists the data for the 208 women with adult-onset POAG in our study. Of these women, 32 (15.4%) harbored the rarer allele of the MYOC.mt1 promoter polymorphism. Two women were homozygous for this change, and 30 were heterozygous. There were no significant differences in any of the disease severity measures between women with and without the rarer allele (P = .44-.98).
MYOC was the first gene found to be associated with open-angle glaucoma.10 It was initially discovered by studying families with autosomal dominant juvenile-onset POAG. Once the gene was identified, MYOC mutations were found to cause approximately 3% of the most prevalent glaucoma, adult-onset POAG, in a series of 103 unselected patients.
Our study was designed to determine the prevalence of MYOC DCVs in a large group of patients with open-angle glaucoma, including a wide variety of glaucoma diagnoses. By sampling all qualifying individuals during a 9-month period, we ascertained a broad representative cross section of a glaucoma population. However, because this study took place in a tertiary referral center, the relative prevalence of glaucoma types may not be representative of the glaucoma population at large. Patients with more difficult treatment problems such as NTG may be overrepresented, and patients with conditions such as OHT are almost certainly underrepresented.
Of the entire group of 779 individuals, 23 (3.0%) had plausible DCVs in 1 of the 3 coding exons of MYOC. Of the 393 patients with adult-onset POAG, 13 (3.3%) had a DCV in the MYOC gene. This was very close to the 2.9% that was reported by Stone et al10 in their original article describing DCVs in this gene. There were no statistically significant differences in MYOC DCV prevalence between groups. This may be due in part to the small numbers of patients in many disease categories.
Disease-causing variations were found in all glaucoma types with the exceptions of corticosteroid-induced glaucoma and corticosteroid-induced OHT, although the numbers in these 2 groups were small. The protein MYOC was identified in the trabecular meshwork beause it was induced by the addition of corticosteroids to trabecular meshwork cells in cell cultures.11 Although patients who respond to corticosteroids might be expected to have a high prevalence of MYOC DCVs, a recent study by Fingert et al27 found no significant association between MYOC DCVs and steroid-induced OHT.
The percentage of DCVs in juvenile-onset POAG was smaller than what might have been expected. Most families with autosomal dominant juvenile-onset POAG had DCVs in MYOC.10 However, most patients in our study were isolated cases and were not part of large families with autosomal dominant glaucoma.
This is the first report to show MYOC DCVs in types of glaucoma other than adult-onset POAG,10 juvenile-onset POAG,10 OHT,28 and NTG.29 There have been no previous reports of DCVs in exfoliative glaucoma, pigmentary glaucoma, exfoliation syndrome without glaucoma, or pigment dispersion syndrome without glaucoma. All of the patients with MYOC DCVs had elevated IOP except for 1 patient with NTG and 1 patient with exfoliation syndrome without glaucoma.
Patients with the Gln368Stop variant represent an interesting subpopulation. Of the 23 patients with DCVs found in this study, 12 (52.2%) had the Gln368Stop. They represent 9 (69.2%) of 13 patients with adult-onset POAG and DCVs, 1(33.3%) of 3 patients with juvenile-onset POAG and DCVs, 1 (50.0%) of 2 patients with OHT and DCVs, and the only case of a DCV seen in pigment dispersion syndrome without glaucoma. In a prior report from our institution, a Gln368Stop variant was found in only 1 (0.2%) of 471 control subjects (91 healthy controls and 380 individuals with other ocular diseases whose glaucoma status was unknown).10 In one study, shared haplotype analysis demonstrated that 27 patients with the Gln368Stop variant were all descended from a common founder.19 None of the patients were aware of a shared ancestry. The 9 individuals with adult-onset POAG and a Gln368Stop variation represented 2.3% of that population. If these numbers held across populations, the Gln368Stop variation would account for thousands of cases of glaucoma in the United States alone.
Patients who have adult-onset POAG and harbor DCVs in the coding regions of the MYOC gene are phenotypically similar to patients with glaucoma who do not harbor MYOC DCVs. The mean age of diagnosis is 60.3 years for those with DCVs compared with 61.5 years for those without. The sex distribution is the same. Both groups had similar numbers of patients with a family history of glaucoma. The median peak IOP was somewhat higher for those with coding sequence DCVs (32 mm Hg OD and 30 mm Hg OS) compared with those without DCVs (26 mm Hg OU); however, the weakly significant difference (P = .02) was lost when corrected for multiple measures (P = .25). Most other measures of glaucoma severity were similar between the 2 groups including cup-disc ratio, visual field loss, number of medications used, and mean number of surgical procedures. Patients with DCVs were somewhat more likely to have undergone laser trabeculoplasty (P = .04), but this significance was also lost with the Bonferroni correction for multiple comparisons (P = .48). The relatively mild nature of the glaucoma in these patients with adult-onset POAG and DCVs is in contrast to that in patients with autosomal dominant juvenile-onset POAG and MYOC DCVs (such as 396Ins397, Tyr437His, and Ile477Asn), who are known to have an aggressive form of glaucoma that is resistant to medical and laser treatment.6,15
We found the prevalence of the rarer allele of the newly described MYOC.mt1 promoter polymorphism to be similar between patients with adult-onset POAG and healthy control patients (15.5% and 23.9%, respectively). There were no differences between patients who had adult-onset POAG with and without the promoter polymorphism in any measure of disease severity examined. Colomb and colleagues performed subgroup analysis of the effect of patient's sex on the response to treatment. They found that most of the effect seen in their study was a result of 13 women who had the MYOC.mt1 polymorphism.21 Because of this, they postulated that the variable response to POAG treatment might be due to the presence of estrogen response elements in the promoter region. We studied 208 women (32 with the rarer allele of the MYOC.mt1 promoter polymorphism) and found no differences between those with and without this allele in any measure of disease severity.
The discrepancy between our results and those of Colomb et al21 may be explained by differences in methods, measures of disease severity, and analysis. Our study evaluated almost 3 times as many patients as theirs (393 vs 142). Our patients were ascertained consecutively and prospectively, whereas their patients were selected retrospectively. The mean age at diagnosis of our patients with POAG was 61.4 years, which is more typical of a POAG population than the 45.1 years reported by Colomb and colleagues. The fact that we did not evaluate the severity of glaucoma using the same measures as their study deserves some explanation; we used several well-established measures of glaucoma severity and reported on every measure for which data had been gathered at the time of patient recruitment. Significant statistical associations were corrected for multiple comparisons.
We did not use the major disease severity measure chosen by Colomb and colleagues (the difference between IOP at diagnosis and IOP at entry into the study) for the following reasons: Because we are a referral center, patients typically come to us after many years of glaucoma treatment, and the IOP at diagnosis cannot be reliably ascertained. More important, we feel that this measure of disease severity is so flawed that it is meaningless. In our opinion, the difference in IOP between diagnosis and study entry would be meaningful only if patients were treated prospectively within a rigidly defined treatment protocol. There is no indication that the patients described by Colomb and colleagues were treated with the same medications or with the same vigor. It seems unreasonable to assume that patients who have a higher IOP at study entry are resistant to medication. In glaucoma practices, patients with the most severe disease often have the lowest treated IOP because they have the lowest target IOP.30 One could make an equally strong assertion for the opposite interpretation of the data of Colomb and colleagues, that those with a higher IOP at entry into the study had milder glaucoma because they had a higher target IOP. In reality, no conclusion can be drawn from these differences in either direction because the treatments the patients received between diagnosis and study entry are unknown and likely variable.
The assertion by Colomb and colleagues that "MYOC.mt1 typing could be a matter of public health"21 implies that clinically important information could be obtained by screening the general population for the MYOC.mt1 promoter polymorphism and coding sequence DCVs. Both our study and that by Colomb and colleagues found no significant difference in the distribution of alleles of the MYOC.mt1 polymorphism between patients with POAG and healthy subjects; therefore, screening for this polymorphism would not be helpful in predicting people at risk for glaucoma. Individuals with coding sequence DCVs are at high risk for developing glaucoma.15 When the 393 patients with adult-onset POAG in our study were screened for all DCVs in the MYOC gene, we identified 3.3% with MYOC DCVs. Assuming that almost 100% of those with DCVs would develop glaucoma, the best we could hope for in a screening test would be 100% specificity and 3.3% sensitivity. This would be a very poor screening device. The ophthalmology community has been searching for glaucoma screening tools that are more sensitive and specific than tonometry.31,32 A test that is nearly 100% specific but almost totally insensitive is not a viable means of screening for glaucoma. Although we are optimistic that molecular genetic screening may eventually help us to identify people at risk for glaucoma within large populations, screening the general population for MYOC variations is not a cost-effective procedure.32
The newly released OcuGene test currently costs $200 (InSite Vision, unpublished data, 2001). This does not include any added charges from the ordering physician. The OcuGene test assays the MYOC.mt1 promoter polymorphism and is also supposed to test for coding sequence DCVs. However, of the dozens of published DCVs,33 OcuGene tests for only 3. If the OcuGene test had been used to study our 393 patients with adult-onset POAG, the tests would have cost at least $78 600. This testing would have found no difference in the distribution of alleles of the MYOC.mt1 promoter polymorphism between patients with glaucoma and the general population. Those with the rarer allele of the promoter polymorphism would have exhibited no difference in their clinical course from those without the polymorphism. Because the test detects only 3 relatively uncommon coding sequence variations, it would have missed all 13 patients in our study who harbored MYOC DCVs.
In summary, DCVs in the MYOC gene are associated with a wide variety of open-angle glaucomas. About 3% of patients with adult-onset POAG have plausible DCVs in the MYOC gene, which is in keeping with a previous report on a smaller group of patients.10 There were no significant differences in phenotype between patients who had adult-onset POAG with and without MYOC coding sequence DCVs. The MYOC.mt1 promoter polymorphism did not provide any information relevant to the clinical course of the patients.
Submitted for publication February 4, 2002; final revision received March 28, 2002; accepted April 16, 2002.
This study was supported in part by grant EY10564 from the National Institutes of Health, Bethesda, Md; the Carver Charitable Trust, Muscatine, Iowa; the Grousbeck Family Foundation, Boston, Mass; the Howard Hughes Medical Institute, Chevy Chase, Md; and unrestricted grants from Research to Prevent Blindness, New York, NY.
We are indebted to the patients for their willing participation in this study. We are also grateful to Rebecca Meyer and Luan Streb, BA, for their excellent technical assistance.
Corresponding author and reprints: Edwin M. Stone, MD, PhD, Department of Ophthalmology, University of Iowa College of Medicine, 200 Hawkins Dr, Iowa City, IA 52242 (e-mail: email@example.com).