Chapitre 3 Mutations fondatrices de TIGR/myocilin associées au glaucome chez la population du Québec (article)

Table des matières

Faucher, M. et coll. (2002) Founder TIGR/myocilin mutations for glaucoma in the Québec population , Hum. Mol. Genet., 11, 2077-2090.

Le glaucome primaire à angle ouvert (GPAO) est une maladie complexe caractérisée par une dégénérescence progressive et traitable du nerf optique. Des mutations du gène TIGR/myocilin (MYOC) sont retrouvées dans approximativement 4% de tous les patients atteints de GPAO. Les populations exhibant des effets fondateurs, comme les Canadiens Français, offrent des avantages uniques à l’implantation de dépistage moléculaire pour la maladie. Pour déterminer le diagnostic moléculaire du GPAO dans cette population, nous avons établi la prévalence des mutations de TIGR/MYOC chez 384 individus non reliés atteints de glaucome, 38 personnes diagnostiquées d’hypertension oculaire et 18 familles affectées (180 patients). Les caractéristiques cliniques associées à ces variations ont été également analysées. Neuf variations de la séquence codante ont été définies comme des mutations causant surtout, mais pas exclusivement, le GPAO. Quatre familles ségréguaient des mutations distinctes: Gly367Arg, Gln368Stop, Lys423Glu et Pro481Leu, alors que 14 patients non reliés atteints de glaucome étaient les porteurs de six mutations connues: Thr293Lys, Glu352Lys, Gly367Arg, Gln368Stop, Lys423Glu et Ala445Val, et deux nouvelles: Ala427Thr et Arg126Trp. Les fréquences de mutations de TIGR/MYOC étaient respectivement de 3,8 et 22,2% dans l’étude d’individus non reliés et des familles. Les variations Gly367Arg et Lys423Glu étaient reliées aux plus jeunes âges d’apparition. Lorsque possible, l’investigation des membres de la famille d’individus non reliés a démontré que les mutations Arg126Trp et Gly367Arg étaient familiales. Des signatures alléliques caractéristiques, démontrant des effets fondateurs spécifiques, ont été observées pour 5 des 6 mutations retrouvées chez au moins deux patients. Des estimations de probabilités de recombinaison suggèrent que la population canadienne française a fort probablement hérité de ces six mutations de 7 à 10 fondateurs différents. Nos données ont démontré qu’un dépistage génétique des mutations de TIGR/MYOC devrait être offert aux familles atteintes de glaucome et aux proches parents des patients non reliés conscients d'antécédents familiaux de glaucome au sein de leur famille.

Primary open-angle glaucoma (POAG) is a complex disorder characterized by a progressive and treatable degeneration of the optic nerve. TIGR/myocilin (MYOC) gene mutations are found in approximately 4 % of all POAG patients. Populations with frequent founder effects, like the French-Canadians, offer unique advantages to implement genetic testing for the disorder. To assess molecular diagnosis for POAG in this population, we determined the prevalence of TIGR/MYOC mutations in 384 unrelated glaucoma patients, 38 ocular hypertensive subjects and in 18 affected families (180 patients). We further analyzed the clinical features associated with these variations. Nine coding sequence variants were defined as mutations causing mostly, but not exclusively, POAG. Four families segregated distinct mutations: Gly367Arg, Gln368Stop, Lys423Glu and Pro481Leu, while 14 unrelated glaucoma patients harbored six known mutations: Thr293Lys, Glu352Lys, Gly367Arg, Gln368Stop, Lys423Glu and Ala445Val, and two novel: Ala427Thr and Arg126Trp. Frequencies of these mutations were respectively 3.8 and 22.2 % in the unrelated and family studies. The Gly367Arg and Lys423Glu variants caused the earliest ages at onset. When achievable, assessment of relatives of unrelated mutation carriers showed the Arg126Trp and Gly367Arg to be familial. Characteristic allele signatures, indicative of specific founder effects, were observed for 5 of the 6 mutations conveyed by at least two patients. Recombination probability estimates suggested that the French-Canadian population had most probably inherited these six mutations from 7 to 10 Québec settlers. Our data demonstrated that genetic screening for TIGR/MYOC mutations should be offered to glaucoma families and to close relatives of unrelated patients aware of a family history for the disorder.

Glaucoma is a complex of ocular disease entities characterized by a progressive degeneration of the optic nerve. The disorder is a major cause of blindness worldwide. Primary glaucomas, for which no causes can be identified, are sub-classified as open- or closed-angle, based upon the structure of the iridocorneal angle of the eye. The most prevalent form of the disorder is adult-onset primary open-angle glaucoma (POAG) that represents between 60 to 70% of all cases of glaucoma. In industrialized countries, POAG affects ~2% of the Caucasian populations > 45 years old (3, 4). In black populations, prevalence of adult-onset POAG is estimated to be three to four times higher than that observed in Caucasians (9). Elevated intra-ocular pressure, also known as ocular hypertension (OHT), represents an important risk factor for POAG.

Genetic factors play a major role in the etiology of POAG (for reviews see (84, 85)). To date, six “ GLC1 ” loci (MIM 137760) have been mapped for the disorder (17-22). The trabecular meshwork-inducible glucocorticoid response ( TIGR ) gene (26), also known as myocilin (MYOC) (25) (MIM 601652) is the first of the “ GLC1 ” disease genes to have been characterized. TIGR/myocilin polypeptide displays two major features: a leucine-zipper-like motif (LZM) in its N-terminal region and a strong homology to members of the olfactomedin protein family in its COOH-terminal region. Its function still remains to be elucidated (25).

More than 40 myocilin mutations have been associated with 2 to 4 % of all POAGs. Almost all of them have been observed in TIGR/MYOC third exon that encodes the olfactomedin homology domain (32, 47). In Western populations, only one mutation, the Arg82Cys variation, has been described in exon 1, N-terminal to the LZM domain (47). Two of these variants have been associated with forms of glaucoma other than POAG (86). In large POAG families carrying TIGR/MYOC mutations, segregation of the disorder was most often autosomal dominant (for example, see [Brezin, 1998 #38]).

POAG displays a slow and insidious process and at least half of the affected individuals remain unaware of it. Palliative therapies, aimed primarily at diminishing elevated intra-ocular pressures, do not always arrest progressive loss of the visual fields (87). Genetic testing for POAG will be thus invaluable to identify persons at-risk for the disorder as these mutation carriers may be offered treatments at an early stage of the disorder before irreversible loss of vision. To optimize the benefits of such testing, it is essential to determine which gene variations cause the disease. It is also worthwhile to estimate the distribution of the mutations within the populations that may be offered the service and understand the clinical features of the glaucomas associated with these mutations. Molecular testing would then be combined with genetic counseling of higher quality.

Population isolates offer major benefits to implement molecular diagnostic procedures. Indeed, it is easier in this type of population to extrapolate the total number of carriers of particular mutations, to standardize phenotypic assessment and to facilitate the definition of the phenotype(s) associated with disease-causing variations. When these populations show founder effects, it is further easier to identify high-risk families and to determine the contribution of the mutated genes to the incidence of the disorder in the population under study (72-75).

It is well recognized that the Québec population exhibits founder effects. Indeed, the more than 6 million French-Canadians now living in the Province of Québec came from about 8500 ancestors who immigrated from France throughout the seventeenth century (78). During the past 300 years and until the middle of the twentieth century, these French settlers maintained an important demographic growth mostly due to a very high birth rate. After the British conquest of Nouvelle-France in 1759, the arrival of new founders to Québec from France virtually stopped, and, for religious reasons, there were very few admixtures between the French and English populations during these three centuries (78). Considering the major advantages offered by the French-Canadians, we exploited this founder population: 1) to assess the prevalence and diversity of TIGR/MYOC mutations in the Province of Québec; 2) to establish genotype/phenotype correlations for patients harboring TIGR/MYOC mutations and 3) to determine how many different founders contributed to the mutations encountered, by characterizing haplotype/allele signatures thereby assessing the potential application of allele signature determination to other heterogeneous genetic diseases. To our knowledge, this is the first study on TIGR/MYOC mutations conducted within a founder population.

We report here that most TIGR/MYOC mutations in the Québec population originated very probably from a few common ancestors. This knowledge should help the design of genetic testing approaches for TIGR/MYOC mutations as well as for other common genetic diseases present in our population.

Eighteen Caucasian kindreds comprising 635 individuals were investigated. In this group of people, 115 persons were diagnosed with POAG, 48 with OHT and 17 were affected by other forms of glaucoma for a total of 180 patients. Another group of unrelated cases comprising 384 glaucoma patient was also investigated. In this second group, 99.0% of the individuals (380/384) were Caucasians, 0.8% (3/384) Black Americans and 0.3% (1/384) American Indians. This second subset of people included 293 POAG, 14 angle-closure glaucoma, 33 mixed-mechanism glaucoma, 31 normal tension glaucoma (NTG), 7 exfoliative glaucoma and 6 pigmentary glaucoma patients. The mean age of the unrelated glaucoma patients was 70.4 years old at time of inclusion in the study and 58.6 years old at diagnosis. Ages at diagnosis ranged from 21 to 92 years old. A third group of 38 unrelated cases of OHT was also investigated. These individuals, all Caucasians, were an average of 66.6 years old at time of inclusion and 57.3 at OHT diagnosis. Clinically normal individuals for glaucoma or any other eye disease included 49 Caucasians and one Black American. Ages of asymptomatic individuals ranged from 43 to 83 years, with a mean of 63.2 years. Criteria for a sequence variation to be considered a mutation were: 1) presence of an amino acid change altering myocilin polypeptide sequence in at least 1 glaucoma patient; 2) presence of this alteration in less than 1% of the general population; 3) absence of the alteration in clinically normal persons and/or 4) report of the variation as a mutation by at least 2 research groups. The first three criteria were obligatory for a novel variation to be considered a probable disease-causing mutation. The fourth criterion was sufficient for a previously identified variation to be considered a mutation when only the first criterion was present.

aProbands having at least one first-degree relative with documented history of glaucoma.

bSubjects recruited as unrelated cases of glaucoma or ocular hypertension (OHT).

cUnrelated spouses in glaucoma pedigrees, not yet investigated.

dInvestigated persons asymptomatic for glaucoma or any other eye disease.

eNovel myocilin mutation.

fNovel myocilin variation, possible mutation.

Genomic DNA from each of the 18 family probands was screened for mutations in all three exons of TIGR/MYOC . Since no mutations have been reported in exon 2, the 422 unrelated glaucoma or OHT individuals were screened only in exons 1 and 3. Our screening revealed 20 different coding sequence variations among all groups. Of these, 13 encoded amino acids change of which nine were considered glaucoma-causing mutations (Table 1 (Tableau 6)). In the 18 families, four previously reported disease-causing mutations were detected: the Gly367Arg (57-59) mutation in the MT kindred, the Gln368Stop (47, 51, 53, 56, 60, 61, 88) in the CT family, the Lys423Glu (66) in the CA pedigree and the Pro481Leu (32) in the VA family (Figure 1). All four families were of French-Canadian ancestry. The Gly367Arg and Lys423Glu mutations were clearly causing glaucoma to segregate as an autosomal trait in the MT and CA kindreds, respectively (Figure 1 (Figure 21)) and see (66)). Among the 422 unrelated cases of glaucoma or OHT, 17 individuals carried eight different disease-causing mutations. Six of these eight mutations have been reported: the Thr293Lys (32, 47, 54), Glu352Lys (28, 32, 54), Gly367Arg (57-59), Gln368Stop (47, 51, 53, 56, 60, 61, 88), Lys423Glu (66) and Ala445Val (47) variants (Table 1 (Tableau 6)). Two variations were considered new glaucoma-causing mutations. An alanine to threonine substitution at position 427 (Ala427Thr) in the olfactomedin homology domain was detected in an unrelated POAG case (UN221) (Table 3 (Tableau 8)). An arginine to tryptophan substitution at position 126 (Arg126Trp) in the LZM motif of the protein was also observed in one unrelated subject with POAG (UN402) and in one patient with mixed-mechanism glaucoma (UN248) (Table 3 (Tableau 8)). Prevalences of TIGR/MYOC mutations were calculated at 22.2% in the family study (4/18) and at 3.8% in unrelated cases of glaucoma (14/384). A mutation frequency of 7.9% (3/38) was also found among unrelated OHT subjects.

The three families segregating the Arg126Trp or the Ala427Thr mutations were extended from individuals recruited from our group of unrelated patients. The phenotypic status of each subject is as described in the box. Family probands are indicated by large vertical arrows.

aProbands having at least one first-degree relative with documented history of glaucoma.

bSubjects recruited as unrelated cases of glaucoma or ocular hypertension (OHT).

cUnrelated spouses in glaucoma pedigrees, not yet investigated.

dInvestigated patients asymptomatic for glaucoma or any other eye disease.

eNovel myocilin variation.

Twelve synonymous codon changes or other sequence variations, previously reported as polymorphisms, were identified in exons 1 and 3 (Table 2 (Tableau 7)). The most common non-synonymous codon change was the Arg76Lys, present in 16.7% (3/18) of the POAG family probands and in 25.0% (96/384) of the unrelated glaucoma cases. A Lys398Arg variation was also observed in 1.8% (7/384) of the unrelated cases of glaucoma or OHT as well as in one person (1/107) of the control population. The most common synonymous codon change: Tyr347Tyr, was detected in 16.7% (3/18) of the family probands, in 5.7% (24/422) of the unrelated glaucoma or OHT individuals and in 7.5% (8/107) of the control population. An aspartic acid to glutamic acid substitution occurring at position 77 (Asp77Glu) in one unrelated POAG individual was considered a new myocilin non-synonymous polymorphism (Table 2 (Tableau 7)).

Our genotype/phenotype correlation studies revealed that three mutations: Gly367Arg, Lys423Glu and Pro481Leu, were associated with POAG diagnosed before 40 years of age (and Table 4 (Tableau 9)). The Gly367Arg mutation was detected in two unrelated individuals as well as in five POAG members of the MT family (Table 3 (Tableau 8)) and Figure 1 (Figure 21)). Median age at diagnosis, calculated from these seven Gly367Arg carriers, was 34.0 years. These carriers all displayed a highest recorded intra-ocular pressure (IOP) of at least 30 mmHg and 4 out of 7 had IOPs measured above 50 mmHg at least once in one eye. The Lys423Glu mutation was identified in one unrelated patient, UN 116, and in one family proband, CA107 (Table 3 (Tableau 8)), a member of the huge CA-001 French-Canadian pedigree encompassing 314 individuals (5). For the present study, ophthalmologic records of 84 glaucoma/OHT patients from the CA family were updated. POAG diagnosis for this mutation ranged from 7 to 63 years old, showing wide phenotypic variability, while median age at diagnosis was of 30.0 years of age (Table 4 (Tableau 9)). The Pro481Leu mutation was identified in one familial proband, VA001, diagnosed with POAG at age 33 with highest recorded IOP at 41/46 (right eye: OD / left eye: OS) mmHg and cup-to-disc ratios of 0.5/0.8 (OD/OS) (Two other family members also harbored this variation, one affected by angle-closure glaucoma at age 48 and the other diagnosed OHT at age 46.

Table 3 (Tableau 8). Clinical features of family probands or unrelated cases carrying glaucoma-causing myocilin mutations

Recruitment

Mutation

Diagnosis

Ethnicity

Proband age

at

diagnosis

(years)

Gonioscopy

(grade)

OD/OSb

Intraocular

pressures

at diagnosis (mmHg)

OD/OSb

Cup to disc ratio

OD/OSb

(age at measure)

Knowledge of

glaucoma

family history

Sporadic (UN402)

Arg126Trpa

POAG

Caucasian

76

IV/IV

34/33

0.8/0.8 (80)

Yes

Sporadic (UN248)

Arg126Trpa

Mixed-mechanism glaucoma

Caucasian

64

I-II/I-II

N/Ac

0.4-0.5/0.6 (71)

Yes

Sporadic (UN391)

Thr293Lys

POAG

Caucasian

67

III/III

22/23

0.8/0.8 (75)

Yes

Sporadic (UN364)

Thr293Lys

OHT

Caucasian

50

IV/IV

25/25

N/Ac

No

Sporadic (UN289)

Glu352Lys

POAG

Black

41

III/III

22/33

0.7/0.8-0.9 (61)

Yes

Sporadic (UN218)

Gly367Arg

POAG

Caucasian

34

IV/IV

30/30

1.0/1.0 (63)

Yes

Sporadic (UN061)

Gly367Arg

POAG

Caucasian

39

IV/IV

47/53

0.9-1.0/0.9-1.0 (45)

Yes

Familial (MT010)

Gly367Arg

POAG

Caucasian

25

IV/IV

42/44

0.6-0.7/0.5-0.6 (25)

Yes

Sporadic (UN190)

Gln368Stop

OHT

Caucasian

43

IV/IV

26/26

0.5/0.5 (52)

Yes

Sporadic (UN191)

Gln368Stop

OHT

Caucasian

42

II/II

33/33

0.3/0.3 (42)

No

Sporadic (UN454)

Gln368Stop

POAG

Caucasian

<86

N/Ac

26/24

0.9-1.0/1.0 (88)

Yes

Sporadic (UN483)

Gln368Stop

Angle-closure glaucoma

Caucasian

46

I-II

25/17

0.9/1.0 (55)

Yes

Sporadic (UN485)

Gln368Stop

POAG

Caucasian

71

III/III

25/24

0.7/0.7 (71)

Yes

Sporadic (UN499)

Gln368Stop

POAG

Caucasian

72

IV/III

50/38

1.0/1.0 (85)

No

Familial (CT003)

Gln368Stop

POAG

Caucasian

62

IV/IV

29/29

0.9/0.8 (71)

Yes

Sporadic (UN116)

Lys423Glu

POAG

Caucasian

38

IV/IV

32/42

1.0/1.0 (49)

No

Familial (CA107)

Lys423Glu

POAG

Caucasian

29

III-IV/III-IV

30/30

0.5/0.8-0.9 (70)

Yes

Sporadic (UN221)

Ala427Thra

POAG

Caucasian

73

IV/IV

28/25

0.8/0.6 (81)

Yes

Sporadic (UN041)

Ala445Val

Pigmentary glaucoma

Caucasian

40

IV/IV

N/Ac

0.6-0.7/0.4 (47)

No

Sporadic (UN159)

Ala445Val

POAG

Caucasian

63

III/III

22/24

0.6/0.8 (68)

Yes

Familial (VA001)

Pro481Leu

POAG

Caucasian

33

IV/IV

41/46

0.5/0.8 (44)

Yes

aNovel myocilin variation.

bOD: right eye, OS : left eye.

cN/A: not available.

The Gln368Stop variation was the most common glaucoma-associated mutation in our sample of unrelated cases (Table 3 (Tableau 8)). Indeed, 6 of the 17 (35.3%) unrelated mutation carriers harbored this variation. Three of them were diagnosed with POAG: UN454, UN485 and UN499, one with angle-closure glaucoma, UN483, and 2 with OHT: UN190 and UN191. The myocilin Gln368Stop mutation was also identified in the CT family proband, CT003 (Figure 1 (Figure 21)). Forty-six members of this family were thus investigated and screened for the mutation. TIGR/MYOCGln368Stop was detected in 15 of them, but only 4 individuals displayed POAG (CT003, CT007, CT066, CT080) while 2 subjects showed OHT (CT029, CT039) (Figure 1 (Figure 21)). Median age at diagnosis of these four POAGs was 57.5 years of age while the four OHT individuals were diagnosed at a median of 42.5 years old. Four POAG (CT093, CT098, CT100, CT101), 2 angle-closure glaucoma (CT083, CT099) and 5 OHT (CT004, CT024, CT026, CT030, CT043) patients in this family did not carry TIGR/MYOCGln368Stop , showing that this mutation did not segregate with the disease phenotype and/or suggesting genetic heterogeneity for glaucoma in this family.

aNovel myocilin mutation.

bDiagnosed with mixed-mechanism glaucoma.

cDiagnosed with angle-closure glaucoma.

dDiagnosed with normal tension glaucoma.

eDiagnosed with pigmentary glaucoma.

fIncludes carriers recruited by extending the families of sporadic cases + total number of carriers in families + controls harboring the mutation.

gMedian calculated when number of individuals > 3.

hPossible myocilin mutation.

iIndividuals with no diagnosis (unknown status).

jFamilies extended from sporadic cases.

Three patients recruited as sporadic: UN221, UN248, UN402, harbored new myocilin variations (Table 3 (Tableau 8)). These carriers were diagnosed with POAG in their sixth or seventh decade of life. Patient UN221, who harbored the novel Ala427Thr variation, was subsequently linked to the KR kindred and thus renamed KR005 (Figure 1 (Figure 21)). She was diagnosed POAG at age 73 with visual field loss, IOPs at 28/25 mmHg (OD/OS), open-angle (grade IV) gonioscopy and a cup-to-disc ratio of 0.8/0.6 (OD/OS). This person was aware of a familial history of glaucoma. To test for co-segregation of POAG with TIGR/MYOCAla427Thr , she collaborated to the recruitment of seven members of her family (Figure 1 (Figure 21)). Five KR members were found to carry the Ala427Thr mutation. Two of them demonstrated variable expressivity of the glaucoma phenotype. Subject KR003 was diagnosed with NTG at age 68 with abnormal visual fields, IOPs at 16/14 mmHg (OD/OS) and a cup-to-disc ratio of 0.35/0.8 (OD/OS). Person KR004 was diagnosed with angle-closure (grade 0 or I gonioscopy) without optic nerve damage at age 77 and stable IOPs at 22/22 mmHg (OD/OS) after laser iridectomy in both eyes. She was considered unaffected. Three other family members with the Ala427Thr variation, KR009, KR010 and KR013, were asymptomatic at ages 41, 44 and 38, respectively. Two other individuals in this family did not carry the Ala427Thr change.

The new Arg126Trp variation was carried by individual UN402 who was diagnosed POAG at age 76 with severe loss of visual fields, IOPs at 34/33 mmHg (OD/OS), grade IV gonioscopy and a cup-to-disc ratio of 0.8/0.8 (OD/OS) (Table 3 (Tableau 8)). Her older brother: GA002, joined our study (Figure 1 (Figure 21)). At age 87, he was under treatment for POAG with IOPs of 15/19 mmHg (OD/OS) and a cup-to-disc ratio of 0.5/0.7 (OD/OS). He also carried the Arg126Trp variation. The other unrelated Arg126Trp carrier, UN248, was diagnosed with mixed-mechanism glaucoma at age 64 with a cup-to-disc ratio of 0.4-0.5/0.6 (OD/OS) and grade I-II gonioscopy (Table 3 (Tableau 8)). Two of his siblings were subsequently recruited to constitute the CC family (Figure 1 (Figure 21)). His sib: CC002, was diagnosed with angle-closure glaucoma at age 58 with abnormal visual fields. This subject did not carry the Arg126Trp variation. The other sib, CC003, carried the variation and was diagnosed with mixed-mechanism glaucoma at age 69, with maximum IOPs at 28/32 mmHg (OD/OS), a cup-to-disc ratio of 0.2/0.3 (OD/OS) and an abnormal visual field in the right eye.

Except for the Ala445Val myocilin change, all missense mutations, including the Arg126Trp and Ala427Thr variations, were at conserved residues between their rat, mouse, bovine and porcine polypeptide counterparts (89-92). The Thr293Lys, Lys423Glu and Pro481Leu changes were also located at highly conserved residues in other human olfactomedin-related proteins (50).

The Arg126Trp, Thr293Lys and Gln368Stop mutations were also identified in our control groups. The Arg126Trp and Thr293Lys variations were detected, respectively, in one and two of the 57 individuals from the general population while the Gln368Stop was observed in 1 of 50 individuals in our clinically normal group. This Gln368Stop carrier was asymptomatic at age 71 with normal visual fields, IOPs at 18/18 mmHg (OD/OS), grade IV gonioscopy and a cup-to-disc ratio of 0.3/0.4 (OD/OS). The Arg126Trp carrier from the general population was investigated normal at age 47. He was the spouse of a previously described homozygous carrier of the Lys423Glu mutation (66). Screening their two children revealed that one of them carried both the Arg126Trp and Lys423Glu variations. This individual was diagnosed POAG at age 11 with intra-ocular pressures of 30/30 mmHg (OD/OS) and complete excavation of the optic nerve heads: 1.0/1.0 (OD/OS). At age 21, this patient displayed severely affected visual fields. The other child, harboring only the Lys423Glu mutation, was diagnosed OHT at age 11 with intra-ocular pressures of 23/27 mmHg. This patient is now 18 years old and has IOPs controlled by medication at 21/18 mmHg (OD/OS) and cup-to-disc ratios of 0.4/0.4 (OD/OS) with normal visual fields.

Allele/haplotype signatures are the alleles/haplotypes surrounding a particular disease susceptibility allele/gene identified among the affected individuals of an isolated population (76). These signatures are of particular importance to rapidly test for founder effects and/or to test specific genes for association with a disorder. To assess founder effects in our glaucoma subjects, carriers of mutations detected in at least two unrelated patients or families were genotyped with 12 polymorphic markers. Ten of these markers were selected from the Généthon human genetic linkage map (93). The sequence assembly of the August 2001 Freeze, build 27, of the Human Genome Project was used to integrate the genetic makers with the physical map and human draft sequence. Usage of the August 2001 Freeze was preferred over the December 2001 and April 2002 versions as positioning of the cluster of markers D1S2851-D1S452-D1S210, estimated at 0.8 Mb centromeric to TIGR/MYOC, was supported by the construction of our yeast artificial chromosome (YAC) contig (94). Indeed, our YAC contig confirmed the physical distance of ~ 800 kb observed between D1S2851-D1S452-D1S210 and TIGR/MYOC in the August 2001 freeze, a distance shorter than either of the two more recent freezes. The 12 markers selected spanned a 7.2 Mb region surrounding the TIGR/MYOC gene at the GLC1A locus on chromosome 1q24-q25 (Figure 2 (Figure 22)).

Data was obtained from the UCSC Genome Browser August 6, 2001 Freeze, build 27 (URL: http://genome.ucsc.edu). Genetic distances (in centiMorgan: cM) between markers were according to the Généthon human genetic linkage map (93). Previously described markers NGA17 and NGA19 (50) (also known as MY5 (NGA17) and MY3 (NGA19) (32)) were respectively located < 1 kb centromeric and telomeric to TIGR/MYOC coding region. Haplotype/allele signatures common between TIGR/MYOC mutation carriers are boxed in black. TIGR/MYOC status is also represented in each black box using the one letter code of the mutated amino acid detected: Arg126Trp (W), Thr293Lys (K), Gly367Arg (R), Gln368Stop (X), Lys423Glu (E), Ala445Val (V). Haplotypes from individuals CA311, BV064, MT010, CT003 and CA310 (in bold) were obtained by genotyping their children and phasing marker alleles. Frequency distributions of the alleles encountered among TIGR/MYOC mutation carriers are indicated for each marker genotyped (number in percentage). Distribution of these alleles in the French-Canadian population was estimated by genotyping at least 100 chromosomes (chr.) obtained from our clinically normal subjects. The number of different alleles (all.) observed in control chromosomes is depicted.

Their allele frequencies were estimated in the French-Canadian population by genotyping at least 100 chromosomes using DNA obtained from our clinically normal individuals. As shown in Figure 2 (Figure 22), allele signatures were established for the Arg126Trp, Thr293Lys, Gly367Arg, Gln368Stop, Lys423Glu and Ala445Val myocilin mutations. Single common allele signatures were detected for 5 of the 6 mutations tested: only the Arg126Trp variation displayed two distinct signatures. Genotyping close relatives of individuals harboring the Gln368Stop (CT003), Thr293Lys (BV064), Gly367Arg (MT010), Lys423Glu (CA310) or Arg126Trp (CA311) variations allowed family phased haplotype signatures to be compared with allele signatures determined in unrelated individuals. Eight of 21 carriers shared with at least one other patient the same signature over the entire 7.2 Mb region (Figure 2 (Figure 22)). All Gln368Stop carriers demonstrated a common signature spanning at least 0.8 Mb, with 5 of the 7 individuals tested sharing a signature of at least 3.4 Mb (Figure 2 (Figure 22)).

The shortest lengths of common inferred haplotypes were observed in individuals UN061 (Gly367Arg) and UN485 (Gln368Stop). Both subjects shared, respectively, an haplotype of three or four polymorphic markers. To assess the likelihood of finding such small haplotype signatures coincidentally in our population, we used the “PHASE” software, a new statistical method to reconstruct haplotypes from an unrelated population sample (79). Using this method, we obtained an estimated distribution of 44 different haplotypes among 46 clinically normal individuals (92 chromosomes) for markers D1S210, NGA17, NGA19 and D1S2815 (data not shown). Among this group of control subjects, 26 individuals were attributed non-ambiguous reconstructed haplotypes while 20 persons were attributed ambiguously called haplotypes. These ambiguous haplotypes were, however, estimated at an average exactitude probability of 0.9095. The 44 different haplotypes reconstructed from the 92 population-matched control chromosomes were compared to the each “core” signature associated with the six TIGR/MYOC mutations tested (Figure 2 (Figure 22)). Frequencies among the control chromosomes for each observed signature were: 1-3-Arg126Trp-1-3 (UN248, UN402) at 0% (0/92), 1-1-Arg126Trp-2-2 (CA311) at 2.2% (2/92), 1-1-Thr293Lys-2-1 (BV064, UN391, UN364, HU019) at 10.9% (10/92), 1-1-Gly367Arg-1-3 (MT010, UN218) at 5.4% (5/92), 1-Gly367Arg-1-3 (UN061) at 13.0% (12/92), 1-1-Gln368Stop-2-3 (CT003, UN190, UN483, UN191, UN454, UN485, UN499) at 0% (0/92), 2-1-Lys423Glu-1-3 (CA310, UN116) at 6.5% (6/92) and 1-3-Ala445Val-1-5 (UN041, UN159) at 1.1% (1/92). None of the signatures associated with the Arg126Trp (1-3-Arg126Trp-1-3) or Gln368Stop (1-1-Gln368Stop-2-3) mutations was observed in any of the normal chromosomes reconstructed, confirming a very small probability for these signatures to be shared coincidentally.

The aim of using a large number of individuals to screen deleterious genes for mutations is to ascertain which variations are true mutations versus those that may be non-pathogenic polymorphisms. Exploiting such strategy in the French-Canadian population, our screen of TIGR/MYOC coding region revealed nine mutations, out of a total of 20 sequence variations observed in 18 families and 422 unrelated patients. To our knowledge, two of these mutations were detected for the first time; they were the Arg126Trp and Ala427Thr variants. Seven mutations were also confirmed in this study; they were the Thr293Lys (32, 47, 54), Glu352Lys (28, 32, 54), Gly367Arg (57-59), Gln368Stop (47, 51, 53, 56, 60, 61, 88), Lys423Glu (66), Ala445Val (47) and Pro481Leu (32) variants. One unreported non-synonymous sequence change was identified; it was the Asp77Glu variant.

The Asp77Glu change, was considered a rare polymorphism as it was recognized only once in 547 individuals and coded for glutamic acid, an amino acid nearly identical to aspartic acid located N-terminal to the LZM region. On the other hand, the Ala427Thr variation was presumed to be a novel disease-causing mutation when encountered in unrelated POAG patient UN221 (renamed KR005) due to its absence in our control population and its conserved nature and position in several mammalian myocilins. Subsequent ascertainment of KR005’s family also revealed a history of hereditary glaucoma (Figure 1 (Figure 21)). Two of KR005’s sisters, who harbored the variant, had been diagnosed with either NTG or angle closure of the anterior chamber of the eye without optic nerve damage requiring laser iridotomy for treatment. The youngest of these two carriers was 68 years old. The other three carriers in the KR family were still asymptomatic (Figure 1 (Figure 21)). As the oldest of these normal carriers was only 44 years old, it is inferred that the Ala427Thr mutation may be associated with a late-onset form of glaucoma.

The new Arg126Trp variation, encountered in two unrelated POAG individuals: GA001 (UN402) and CC001 (UN248) (Figure 1 (Figure 21)), was also assumed to be a mutation as the wild-type arginine residue in the LZM region was conserved between all mammalian species sequenced to date. Both family probands had one affected sibling: GA002 and CC003, who carried the variation (Figure 1 (Figure 21)). One patient in the CC family: CC002 affected by angle-closure glaucoma, however, did not harbor the mutation while one subject in our general population displayed it. This asymptomatic person was the 47 years old spouse of a member of one of our large POAG families. Interestingly, the Arg126Trp variation was also detected in one member of the CA-001 pedigree who had been previously found to carry the Lys423Glu mutation (data not shown). This compound Arg126Trp/Lys423Glu carrier has been diagnosed with an aggressive form of juvenile-onset open-angle glaucoma at 11 years old. These observations therefore suggest that TIGR/MYOCArg126Trp may be a late-onset open-angle disease-causing mutation and/or a genetic modifier accelerating the pathogenesis of the disorder when combined with TIGR/MYOC or other glaucoma mutations.

Although the Lys423Glu mutation displayed a strong founder effect that was traced back to a single affected male born in 1799 in eastern Québec (5), it was detected only once in our unrelated cases demonstrating the efficacy of our earlier recruitment procedures. Indeed, individual UN116, who shared the mutation and the same disease allele signature with all CA patients, had no prior knowledge of being a member of this huge family.

Among other amino acid changes, the Lys398Arg variation was observed in 7 of the 384 glaucoma patients and in one subject of the control groups (1/107). Because this variation was present in less than 1% of our normal individuals, we may have classified it as a pathogenic amino acid change according to our criteria for mutations. However, earlier studies reported the Lys398Arg change as a non-disease-causing polymorphism (32, 51, 53, 54). In transfected human trabecular meshwork cells, Jacobson et al. further demonstrated that TIGR/MYOCLys398Arg polypeptide was secreted outside the cells, in a fashion similar to the release of the wild-type protein, whereas polypeptides encoding confirmed mutations remained sequestered intracellularly (71). Based on these reports, the Lys398Arg variation was thus considered a non-pathogenic amino acid change in the French-Canadian population.

TIGR/MYOC mutation frequencies were estimated at 22.2% (4/18) in families displaying autosomal dominant POAG and at 4.0% (17/422) in unrelated individuals affected by glaucoma or OHT. Among all mutations, the Gln368Stop variant was the most common representing 7 of the 21 (33.3%) mutational events in the affected individuals investigated during our initial screening. Extension of CT003’s family revealed, however, that the Gln368Stop mutation did not segregate with the disease phenotype (Figure 1 (Figure 21)), raising the possibility that this change may not be a disease-causing variation in the pedigree. This observation agreed with other studies describing POAG and OHT individuals not harboring the TIGR/MYOCGln368Stop mutation in Gln368Stop families (28, 61, 88). Although these earlier studies suggested that the mutation may be associated with a late age at onset and/or may be interacting with other uncharacterized factors, none of them proposed that the Gln368Stop may not be pathogenic. With a 71 years old Gln368Stop carrier in our clinically normal group and no French-Canadian families segregating the glaucoma phenotype with TIGR/MYOCGln368Stop , our observations underlined the caution with which future investigations should be undertaken in order to test the pathogenic role of TIGR/MYOCGln368Stop in glaucoma.

In our group of 422 unrelated individuals, 91 patients were diagnosed with forms of glaucoma other than “typical” POAG. Thirty-eight (38) cases of OHT were also investigated. Six mutational events were identified in these 129 individuals. Three of these six carriers were affected by OHT with IOPs ranging between 25 and 33 mmHg at diagnosis (one Thr293Lys and two Gln368Stop carriers). Three other carriers were affected by types of glaucoma that were not POAG. A Gln368Stop carrier was affected by angle-closure glaucoma, an Arg126Trp carrier was suffering from mixed-mechanism glaucoma while an Ala445Val carrier displayed pigmentary glaucoma. Our phenotype/genotype correlations therefore demonstrated that TIGR/MYOC mutations may be associated with forms of glaucoma other than POAG. Our observations also agreed with the study by Vincent et al. who reported one Thr293Lys and one Gly399Val carrier affected, respectively, by pigmentary glaucoma and mixed-mechanism glaucoma (86).

When the group with “typical” POAG was compared with patients showing a broader definition of glaucoma, we calculated a 3.8% (11/293) TIGR/MYOC mutation frequency in the POAG-only affected group and a 4.7% (6/129) frequency in the OHT and other forms of glaucoma subset. These values were thus not significantly different than the frequency of TIGR/MYOC mutations measured when all 422 patients were considered as one group (4.0%). Widening our inclusion criteria allowed us to increase the number of glaucoma/OHT mutation carriers who may be offered close follow-up for diagnosis and/or early treatment, thereby preventing loss of vision. Further studies are also required to assess the proportion of OHT carriers who will develop glaucoma, as this symptom represents a major risk factor for POAG.

The establishment of haplotype signatures in conjunction with our large scale mutational analysis demonstrated the importance of founder effects in the French-Canadian population. Not only did we found that previously characterized mutations and new probable disease-linked variations were present in the affected population, we also observed that almost all TIGR/MYOC alterations showed a unique “core” signature associated with 18 unrelated individuals and 138 additional related individuals comprising 91 patients and 47 asymptomatic mutant carriers. The majority of French-Canadian TIGR/MYOC -linked glaucoma cases may be thus related to a small number of founders that introduced the disease in this population. On the other hand, we also have to consider that smaller signatures may be associated to the same variation linked to an ancestral event pre-dating the founding of the French-Canadian population. For instance, the small signatures detected in some Thr293Lys (HU019), Gly367Arg (UN061) or Gln368Stop (UN485, UN499) carriers may originate from events preceding the French settlement in Québec. Alternatively, these small signatures may be false positive due to lack of marker informativity. The use of a phased disease haplotype as a reference signature for five of the six mutations, however, diminished the chances of finding these false positive signatures. By estimating the haplotype frequencies in our control population, it was possible to evaluate the specificity of these signatures. For the smallest ones, like those observed in individuals UN485 and UN061, we considered that they may be coincidentally shared haplotypes. To assess this likelihood, we used a new statistical method to reconstruct haplotypes in our control population (79). The disease haplotype for individual UN485, a Gln368Stop carrier, was not found in any of the control chromosomes. On the other hand, the haplotype of individual UN061, a Gly367Arg carrier who shared a disease haplotype covering only markers NGA17, NGA19 and D1S2815, was found in 13.0% (12/92) of the reconstructed control haplotypes. These estimates therefore supported that the Gly367Arg signatures found in our population may have originated from two separate events whereas the Gln368Stop haplotypes were most likely related to a common ancestor.

Moreover, if we consider that all common haplotypes observed are non-coincidental, we can estimate the probability (P) for a disease haplotype to be conserved on a specified genetic distance (θ: recombination fraction) for a specified number of generations (g), obtaining P = (1-θ)g. Thus, an estimated probability of observing at least one haplotype originating from a common ancestor shorter than θ after g generations would be less or equal to P, where P ≤ [1 - (1-θ)gn] among n haplotypes. Four mutations displayed a common haplotype of less than 1 cM: Gly367Arg (UN061), Gln368Stop (UN485, UN499), Thr293Lys (HU019) and Arg126Trp (CA311). Considering that the majority of the French-Canadian population originated from ancestors approximately 10 generations old (300 years with average of 30 years/generation (80)) and that the number of haplotypes established for these four mutations were nArg126Trp = 3, nThr293Lys = 4, nGly367Arg = 3 and nGln368Stop = 7, we obtained for θ ≤ 1 cM: PArg126Trp ≤ 0.26, PThr293Lys ≤ 0.33, PGly367Arg ≤ 0.26 and PGln368Stop ≤ 0.51. These calculations therefore showed no overwhelming probability that these small disease haplotypes, sometimes far shorter than 1 cM in length, could be related within 10 generations. On the other hand, the alternative possibility of sharing a common founder is as probable as not sharing such common founder for individual UN485 harboring the Gln368Stop mutation (PGln368Stop ≤ 0.51). We considered, however, that the individuals carrying these small haplotypes may have inherited their mutation from a different founder than other individuals with longer shared haplotypes. Overall, this means that more than one individual or founder may have brought the Arg126Trp, Thr293Lys, Gly367Arg and Gln368Stop mutations in the Québec population. Moreover, we observed that individuals HU019, UN061 and UN485 were demographically isolated from all the others who were carrying the same mutation (data not shown). This observation confirmed that they were more likely to be distantly, rather than closely, related to the other carriers.

Polymorphic markers located less than 1 kb centromeric or telomeric from TIGR/MYOC coding region discriminated two distinct signatures between CA311 and the two other Arg126Trp carriers (UN248, UN402). Two separate individuals or founders most probably brought this alteration in the Québec population as these two different Arg126Trp haplotypes may originate from a recurrent mutational event. Indeed, individual CA311, who was of Irish ancestry, lived in a village more than 500 km away from the two other unrelated individuals with French ancestry who also carried the Arg126Trp alteration (UN248, UN402).

The 7 haplotype/allele signatures for 6 different variations demonstrated strong founder effects in the French-Canadian population. Considering our recombination probability estimates, we then have to specify that the 21 individuals genotyped within the vicinity of the TIGR/MYOC gene probably inherited their mutation from 7 to 10 different Québec settlers: 2 for Arg126Trp, 1 or 2 for Thr293Lys, 1 or 2 for Gly367Arg, 1 or 2 for Gln368Stop, 1 for Lys423Glu and 1 for Ala445Val.

Common haplotypes for markers surrounding TIGR/MYOC have been found among many Gln368Stop carriers in the United-States, Canada and Australia (32, 95), suggesting common signatures in various populations for older mutations. Our genotyping results showed that all Gln368Stop carriers shared a common haplotype signature supporting this worldwide founder effect hypothesis (32). Individuals with smaller haplotypes may have inherited their glaucoma-causing mutation from different founders, these founders sharing an older ancestor. Two other mutations in the Québec population: Thr293Lys and Gly367Arg, have been also identified in various populations (32, 47, 54) (57-59). This observation supported the hypothesis that each one of these two mutations may be linked to an old mutational event also disseminated in several populations. However, as previously discussed, the shorter common signature observed in individual UN061 (Gly367Arg) could very well be coincidental. In this case, more than one mutational event may be responsible for the dispersion of the Gly367Arg mutation in different populations.

On the other hand, the Lys423Glu mutation has been only identified in the French-Canadian population. This variation may be thus, more probably, a recent mutation. Indeed, a few individuals distantly related to the huge Lys423Glu pedigree were also found to carry the TIGR/MYOCLys423Glu disease haplotype. Although these individuals were linked to the family by a common remote ancestor born during the 18th century, they did not harbor the mutation itself (data not shown). As no mutation was identified in the TIGR/MYOC gene in this distant branch, the Lys423Glu mutation may have been introduced in the French-Canadian population by a “de novo” event occurring roughly a century after the settlement of the ancestors of the CA-001 pedigree in the Province of Québec. With such recent mutation, we were thus expecting to find very long linkage desequilibrium around TIGR/MYOC in mutation carriers, as exemplified for the unrelated individual UN116 who shared the entire CA-001 disease haplotype over the 7.2 Mb region. This individual was found to be an adopted child, most probably closely related to the huge Lys423Glu family.

It is well recognized that young population isolates (10-15 generations old) like the French-Canadian population, are ideal to search for haplotype signatures. Since the reduced genetic diversity found in these populations led to expect fewer alleles than in older population, recognizable haplotype signatures should be more likely identified in young populations than in heterogeneous populations (76). However, for common diseases like glaucoma, founder effects may be harder to detect. The observation of nine different mutations in the Québec population and very small common disease haplotypes between some of the carriers supported this assumption. Since the genetic pool of the French-Canadian population is estimated at 8500 founders (78), it was not surprising to find as many different variations and signatures. Yet, it has been proposed that the nonuniform distribution of some hereditary diseases found in the Québec population may be related to a geographical stratification of the founder effect (96). This fragmentation of the founder effects may thus favor homogeneity if the individuals of the Québec population sampled in our study all came from the same region.

Nevertheless, individuals that shared a common signature covering the entire 7.2 Mb tested in the GLC1A region, like patients CT003 with UN190, BV064 with UN391, MT010 with UN218 and CA310 with UN116, were more likely to be close relatives than individuals sharing smaller portions of a signature. In fact, identical signatures observed in individuals MT010 and UN218 were later explained by investigating MT010’s kindred that revealed the unrelated individual UN218 as the grand-uncle of MT010. Such example showed how characterization of haplotype signatures within a population isolate allowed unrelated patients affected by a genetic disease to be linked with affected families participating in the same study. This procedure, combined with ordered and accessible genealogical records, may lead to link virtually every unrelated individuals mutated for a disease gene with already characterized families affected by the same gene. Genetic screening for TIGR/MYOC mutations should be thus offered to glaucoma families and to close relatives of unrelated patients aware of a family history for the disorder.

Haplotype mapping for the localization of new disease loci in the French-Canadian population would hardly be useful as disorders like glaucoma or other multi-gene diseases would greatly dilute potentially interesting signatures. On the other hand, haplotype signatures may be a very useful tool to refine a disease interval once a locus has been mapped. As exemplified by our comparison of previously phased haplotypes in myocilin -mutated families with genotypes of unrelated individuals, the minimum overlap region of a disease locus could be significantly reduced, narrowing the putative disease-causing gene interval and accelerating its discovery.

This research has been approved by the CHUL Research Center Ethics Committee. All participants, affected or not, signed an informed consent document before entering the study. Recruitment was performed through a network of 103 ophthalmologists covering the Province of Québec. Ophthalmologists with knowledge of families affected by glaucoma asked their patients to enter the study. Unrelated cases were recruited at random. Each family proband had at least one first degree relative affected by glaucoma and families were extended to maximize recruitment of glaucoma patients. Clinical assessment comprised complete ophthalmologic evaluation as previously described in our earlier study (5) . Diagnostic criteria for POAG were: 1) characteristic optic disk damage and/or visual field impairment; 2) grade III/IV (open-angle) gonioscopy and 3) intra-ocular pressures ≥ 22 mmHg in at least one eye. In the absence of optic disk damage or visual-field alteration, subjects with intraocular pressures ≥ 22 mmHg in both eyes and grade III or IV gonioscopy were diagnosed OHT. Medical charts were reviewed to detect ages at which intraocular pressures raised ≥ 22 mmHg considering the possibility of late diagnosis. Persons were considered normal when they presented normal optic disks and showed highest intraocular pressures ever recorded at < 22 mmHg. Subjects were diagnosed with normal tension glaucoma (NTG) when they presented optic disk damage, visual field impairment and open-angles with intraocular pressures < 22 mmHg. Subjects were diagnosed with chronic angle-closure glaucoma when there was presence of peripheral anterior synechias (PAS), very narrow angle (grade 0-I), elevated IOP and optic nerve damage. Subjects diagnosed with mixed-mechanism glaucoma had elevated IOP, optic nerve damage and narrow angle (grade 0-II) without PAS. Subjects with deposition of fibrilar protein throughout the anterior segment of the eye were diagnosed with exfoliative glaucoma. Subjects in whom optic nerve degeneration was caused by abnormal amounts of liberated pigment from the posterior surface of the iris that were deposited throughout the anterior and posterior chambers of the eye were diagnosed with pigmentary glaucoma (2). Iridogoniodysgenesis, iris hypoplasia and abnormal tissue in the angle were also carefully assessed. Persons diagnosed with secondary glaucoma (for example: uveitis-, steroid- or trauma-induced glaucoma) or affected by diabetes were excluded from the study.

Control individuals included 57 unrelated spouses sampled in various glaucoma pedigrees and 50 clinically investigated individuals normal for glaucoma or any other eye disease. The unrelated spouses were considered as randomly selected individuals from the general population with an unknown diagnosis. Clinically normal individuals were all examined at the Clinique d’Ophtalmologie de la Cité in Québec City.

Genomic DNA was extracted using the Puregene DNA isolation protocol from 28 ml of whole blood drawn by venipuncture. TIGR/MYOC amplicons were obtained by the polymerase chain reaction (PCR) using the primer pairs described in a previously published study (28). Each familial proband was screened for mutations in all three exons of TIGR/MYOC . As no mutations have been reported in exon 2, unrelated individuals were screened only for exons 1 and 3.

Initial PCRs were performed on a Hybaid Omnigene Temperature Cycling System in a total volume of 50 μl containing 100 ng of genomic DNA, 20 pmol of each primer, 200 μM dNTPs, 50 mM KCl, 10 mM Tris (pH 9), 1.5 mM MgCL2, 0.01% gelatin, 0.1% Triton X-100, and 1 U Taq polymerase (Invitrogen, Burlington ON). Amplifications were carried out using a “hot-start” procedure. Taq polymerase was added after a 5-min denaturation step at 95°C. Samples were then processed through 35 cycles of denaturation (95°C for 30 s) and annealing (55-60°C for 30 s), followed by one last step of elongation (72°C for 50 s). PCR products were diluted in 5 volumes of PB buffer (Qiagen, Mississauga ON), transferred on a Whattman GF/C filter plate, washed twice with 80% ethanol 20 mM Tris pH7.5 and eluted in 50 μl of water. Samples were quantified by the PicoGreen reagent protocol. A second PCR was performed on Applied Biosystems Gene Amp PCR System 9700 (96 wells) or 9700 Viper (384 wells) machines to incorporate the sequencing dyes using a protocol of 25 cycles of denaturation (95°C for 10 s) and annealing (55°C for 5 s), followed by one last step of elongation (59°C for 2 min). PCR products were purified by the ABI ethanol-EDTA precipitation protocol, collected using a Beckman-Coulter Allegra 6R centrifuge and resuspended in a 50% HiDi-formamide solution. Samples were then run on Applied Biosystems Prism 3700 DNA Analyzer automated sequencers. Sequence data was analyzed using the Staden preGap4 and Gap4 programs.

To characterize potential founder effects, carriers of TIGR/MYOC mutations were genotyped at the GLC1A locus between D1S2799 and D1S218 within a region of 7.2 Mb surrounding TIGR/MYOC . Twelve polymorphic markers were genotyped using a standard radioactive protocol and/or an automated fluorescent protocol. For the radioactive protocol, markers were labeled using [35S]α-dATP in a 20-μl PCR reaction mixture containing 50 ng of genomic DNA; 1 x PCR buffer; 0.2 mM each of dCTP, dTTP, and dGTP; 10 mM dATP; 4-20 pmol of each primer; and 1.5 μCi (1.5 pmol) [35S] α−dATP. PCR conditions were 35 cycles at 95°C for 30 s, 55-57°C for 30 s and 72°C for 5-10 s. Taq polymerase was added after a 5 min denaturation step at 95°C. Samples were resolved by electrophoresis on 6% polyacrylamide gels. The gels were then exposed to X-ray film for 12-48 h. For the fluorescent protocol, markers were labeled with Applied Biosystems fluorescent dyes (PET, VIC, NED and 6FAM). In a 15-µl PCR reaction mixture containing 30 ng of genomic DNA; 1 x PCR buffer without MgCl2; 1.5mM of MgCl2; 0.167 mM of dNTP; 2 pmol of each primer; and 0.5 U of Platinum Taq Polymerase (Invitrogen). The forward primer contained the fluorescent molecule and the non-fluorescent reverse primer was modified by the addition of a “pig tail” (5’-GTTTCTT). PCR conditions were: 1 cycle at 95° for 3 min followed by 10 cycles at 94°C for 15 s, 57°C for 15 s and 72°C for 30 s, then 20 cycles at 89°C for 15 s, 57°C for 15 s and 72°C for 30 s completed with 1 cycle at 72°C for 15 min .

Markers were pooled and diluted in water using a Tecan RSP-150 apparatus in ratios ranging from 1:9 to 1:49. From this pool, 2 µl was added to 10 µl of Applied Biosystems Hi-Di formamide containing Genescan 500 LIZ (3:1000 dilution). Samples were denatured for 5 min at 95°C and then resolved on a 3100 Genetic Analyzer from Applied Biosystems using 36 cm capillaries with pop-4 polymer. Data was analyzed using the GeneMapper Version 2 software from Applied Biosytems.

An arbitrary number was assigned for each allele observed. Relative positioning was performed according to the genotype of “Centre d’Études du Polymorphisme Humain” (CEPH) individual 134702 (93). The smallest allele was given a relative length of “0”. The relative alleles associated with the arbitrary numbers of Figure 2 (Figure 22were: D1S2799 (1: +6 nucleotides, 2: 0, 3: -2, 5: +8, 7: +12, 9: -4, 13: +10), D1S2658 (1: +6, 2: 0, 4: +4), D1S2851 (1: +2, 2: 0, 6: +4, 7: -4, 8: -16, 11: -14, D1S452 (1: +2, 2: 0, 3: +4, 7: -6), D1S210 (1: +2, 2: -2, 4: 0), NGA17 (1: 0, 3: +4), NGA19 (1: +4, 2: 0), D1S2815 (1: 0, 2: -4, 3: -2, 5: +6), D1S2790 (1: +6, 2: 0, 3: +2, 4: +4, 5: -2, 6: +8), D1S2814 (2: 0, 6: +2, 7: -4), D1S242 (1: 0, 2: -2, 4: +8, 5: +6, 6: +2), D1S218 (1: 0, 2: -8, 4: +4, 5: -2, 7: -6, 9: -4, 10: +6).

We thank all the family members and unrelated individuals for their enthusiastic participation in this study. We also thank Micheline Plante for her work concerning the analysis of the CA family and the following clinicians who contributed to the recruitment of patients: Yves Asselin, Michelle Barette-Delorme, François Bellefeuille, Maurice Bissada, Claude Boulanger, Jocelyne Brochu, Louise Brossard-Jaimes, Isabelle Brunette, Nicolas Chehade, Jean-René Clément, Gilles J. Cormier, François Demay, Paul-Eugène Demers, Francine Deslauriers, Paule Dubé, Stéphane Dupont, Christian Ferremi, Lise Garand, Gilles Gaudreault, Pierre Gauvin, Guy Gélinas, Julius Gomolin, Jean-Pierre Gravel, Jacques Grégoire, Réjean Grenier, Normand Guilbault, Hachmi Hammami, César Heredia, Barry Kattleman, Élie Khoury, Marvin L. Kwitko, Alain Lachance, Laurent Lamer, André Lapointe, Michel Lefrançois, Céline Legris, Anne-Marie Mathieu, Brian R. Mathieu, Stéphane Morin, Darren Payne, Denis Plante, Gaston Poirier, Gaétan Richard, Jean-François Roberge, Marcel Roberge, Nicole Robillard, Denis Rodrigue, Lise St-Pierre, Marcel Simard, Raymond Simard, Claude Sirois, Richard O. Swieca, Denis Tardif, Paul Thompson, Richard Tourigny, Pierre-Marie Tassot, Charles Weldon. We also thank the nurses Francine Brideau, Claire Blondeau, Francine Larocque, Thérèse Martin, Marie-France Minville and Monique Poirier for their excellent work concerning the recruitment of families, venipunctures and for their gathering of clinical data. This work was supported by InSite Vision Inc., the Canadian Institutes of Health Research (CIHR) (Grant MOP-13428), the Canadian Foundation for Innovation (Grant #548), La Fondation des Maladies de l’Oeil and the Fonds de la Recherche en Santé du Québec (FRSQ) Health Vision Research Network. M.F. was supported by a CIHR K.M. Hunter doctoral studentship. V.R. was an FRSQ National Investigator.