A Locus for Autosomal Dominant Colobomatous Microphthalmia Maps to Chromosome 15q12-q15

Congenital microphthalmia (CMIC [MIM 600165]) is an ocular malformation characterized by a small eye attested by shortened axial length (Weiss et al. 1989; Warburg 1993). The reported prevalence of CMIC at birth is ∼1.5/10,000 in white populations (Warburg 1993; Bermejo et al. 1998; Busby et al. 1998). Severity of the disease often ranges from mild to extreme microphthalmia (anophthalmia) within a single family, suggesting variable gene expression (Warburg 1993). Also, microphthalmia is often asymmetric or unilateral, suggesting that nongenetic factors or somatic genetic events may also contribute to the phenotype. Some other ocular abnormalities—namely, cataracts, microcornea, sclerocornea, retinal coloboma, and optic-nerve coloboma—are considered to be part of the spectrum of anomalies found in isolated CMIC (Warburg 1993). Nonocular abnormalities are associated with microphthalmia in 80% of cases. Chromosomal abnormalities of virtually all chromosomes may result in syndromic microphthalmia, trisomy 13 being the most frequently observed. Moreover, environmental teratogens such as alcohol, rubella, toxoplasma, or cytomegalovirus may also cause microphthalmia (Warburg 1993).

Although most cases of nonsyndromic microphthalmia are sporadic, a few families with autosomal dominant (Pearce 1986; Vingolo et al. 1994; Fryns 1995; Othman et al. 1998), autosomal recessive (Bateman 1984; Kohn et al. 1988; Ghöse et al. 1991; Warburg 1993; Zlotogora et al. 1994; Bessant et al. 1998, 1999) or, more seldom, X-linked recessive inheritance of the trait (Hoefnagel et al. 1963) have been hitherto reported. Human eye development begins as early as the fourth week of gestation and proceeds through the embryonic and fetal life (for reviews, see Jean et al. 1998; van Heyningen 1998). Many transcription factors and growth factors are involved in mammalian eye development (for reviews, see Freund et al. 1996; Graw 1996; Jean et al. 1998). Recently, mutations in the CHX10 homeobox gene were identified in two families presenting with nonsyndromic microphthalmia (Percin et al. 2000), and mutations in both alleles of the RX gene were identified in a patient with severe congenital microphthalmia (Voronina and Mathers 2000). Two other loci for human isolated microphthalmia/anophthalmia were previously identified on chromosome 14q32 (Bessant et al. 1998) and 11p (Othman et al. 1998).

Here, we report a study of a five-generation family of Sephardic Jewish origin, including 38 members (fig. 1). Thirty-five individuals were evaluated by ophthalmologic examination by one of the authors (J.-C.Z.) in most cases and by general examination by four of the authors (J.-C.L., A.R.-R., H.P., and P.E.). Inclusion criteria for microphthalmia was reduction of the total axial length <20 mm in at least one eye, determined by use of ultrasonography (table 1). Seven individuals were considered to be affected, since they had either unilateral or bilateral microphthalmia of variable severity. Five affected individuals had an ocular prosthesis in at least one eye. Two individuals, IV:10 and V:2, were designed as of uncertain status, since they had ocular anomalies but normal axial length of both eyes. The disease trait appears to be inherited in an autosomal dominant manner with incomplete penetrance and variable expression. For example, individual IV:10, who carries the mutant allele, has corneal clouding and iridocorneal synechia (Peters’ anomaly [MIM 604229]) but no microphthalmia, whereas individuals IV:14 and V:4 have unilateral optic-nerve agenesis associated with other ocular abnormalities. This illustrates that virtually any structure of the eye may be involved in the family described here. Neither associated extraocular anomaly nor mental retardation was observed, except in affected individual IV:17, who has cleft lip and palate. Karyotypes performed on blood lymphocytes in individuals IV:10 and IV:14 were normal. Blood sampling was done in each case after informed consent was obtained by clinicians.

Genomic DNA was extracted from blood leukocytes using the QIAamp Maxi Kit (Qiagen). We first analyzed microsatellite DNA markers mapping to seven candidate loci, using the chemoluminescent detection method (Vignal et al. 1993). We studied linkage to the following genes: CHX10 (loci D14S63, D14S74, and D14S68), PAX2 (Sanyanusin et al. 1996; Eccles et al. 1999), PAX6 (Martha et al. 1993; for review, see Prosser and van Heyningen 1998), MITF (locus D3S3633; for review, see Jean et al. 1998), and HRY (loci D3S1265 and D3S1314; for review, see Freund et al. 1996) and the following loci: 11p (markers D11S905, D11S4109, D11S4113, and D11S1314; Othman et al. 1998) and 14q32 (markers D14S62, D14S65, and D14S78; Bessant et al. 1998).

Next, we performed a genome-scan analysis using fluorescent microsatellite markers (ABI prism linkage mapping panels LMSV2). PCR amplifications were performed according to the manufacturer’s recommended protocols (PE Biosystems). PCR products were denaturated and size-fractionated on a 6% denaturing polyacrylamide gel run on an ABI373 sequencer. GENESCAN and GENOTYPER software was used to determine the size of alleles. Additional fluorescent markers were analyzed—namely, D15S975, D15S1019, D15S1040, and D15S118 from the ABI prism linkage mapping set; HD5; and Généthon markers D15S1043, D15S1031, D15S1010, and D15S144 (Dib et al. 1996).

A two-point LOD-score analysis was computed using the LINKAGE program of the LINKAGE package version 5.2 (Lathrop and Lalouel 1984), under the assumption of an autosomal dominant model with an estimated penetrance of 50% and a frequency of the mutant gene of 1/14,000. Allele frequencies were estimated from family data. For haplotype analysis, we only considered affected individuals and obligate carriers of the disease-causing gene, owing to the incomplete penetrance of the disease.

We first excluded the seven analyzed candidate loci. We found LOD scores consistently <−2 at each of these loci, demonstrating absence of linkage to the disease gene in the family presented here (data not shown). We then performed a genome scan study and found a maximum LOD score (Z_max) between the disease locus and microsatellite DNA marker D15S1007 (Z_max=3.77 at recombination fraction [θ] 0; table 2). Additional two-point LOD scores obtained with 11 other surrounding markers are shown in table 2. Another marker gave a positive LOD score >3 at θ=0, and two other markers gave positive LOD scores >1.9 at θ=0.

Haplotype analysis (fig. 1) showed a telomeric recombination event between markers D15S1007 and D15S1040 in individual III:1, placing the disease-causing gene centomeric to the marker D15S1040. Similarly, recombination events between loci D15S1002 and D15S1048 occurred in individuals III:11, IV:8, and IV:17, indicating that the disease gene is telomeric to locus D15S1002. Thus, in this family, the disease gene lies within a 13.8-cM region on chromosome 15q12-q15, bounded proximally by locus D15S1002 and distally by locus D15S1040. Moreover, two recombinations occurred in affected individual V:3, on the disease-associated chromosome inherited from individual IV:6 (fig. 1)—the first one between loci D15S1002 and D15S1048 and the second one between loci D15S1043 and D15S1031. This result allows the exclusion of a 2.2-cM region, flanked by loci D15S1048 and D15S1043 and located within the 13.8-cM candidate interval. Thus, two genetic intervals—encompassing, respectively, 4.5 cM between loci D15S1002 and D15S1048 and 7.1 cM between loci D15S1043 and D15S1040—are likely to contain the disease-causing gene. To date, we have not been able, by using other microsatellite DNA markers (not shown), to exclude the entire D15S1002–D15S1048 interval on the basis of haplotype analysis.

The present study strongly suggests the assignment of a nonsyndromic autosomal dominant microphthalmia gene to chromosomal region 15q12-q15 within a 13.8-cM interval bounded by loci D15S1002 and D15S1040. On the basis of haplotype analysis, a 2.2-cM region defined by loci D15S1048 and D15S1043 and located within the 13.8-cM critical interval was excluded. Indeed, we observed a double recombination in patient V:3 within a 6.8-cM interval defined by loci D15S1002 and D15S1031. The probability of such an event is very low, ∼0.5%. However, it is worth noting that both recombination events occurred close to potential hotspots for recombinations (Amos-Landgraf et al. 1999; Christian et al. 1999).

Clinical expression of the disease was extremely variable in individuals carrying the disease-associated haplotype. For example, patients IV:8 and V:4 had unilateral microphthalmia, and patient V:7 had bilateral anophthalmia. Such an intrafamilial variable expression of the disease has been described elsewhere in families with microphthalmia (Warburg 1993; Zlotogora et al. 1994; Fryns 1995; Bessant et al. 1999). More surprisingly, patient IV:10 had no microphthalmia but did have Peters’ anomaly, including corneal clouding, iridocorneal synechy, and ptosis, whereas patient IV:14 had unilateral anophthalmia and controlateral isolated optic-nerve agenesis, with a normal ocular globe. The latter observation suggests that anterior segment dysgenesis may be allelic to posterior segment anomalies, at least in the family reported here. Incidentally, individual V:2, who inherited part of the disease-associated haplotype, had remnant pupillar membranes, a feature that may be encountered in normal eyes. It is striking that each of the five carriers of the disease-causing gene in generation III has normal eyes, whereas only 2/6 carriers in generation IV are considered to be unaffected, and 0/4 patients with the entire disease-associated haplotype in generation V are unaffected. The apparently increasing penetrance of the disease trait through successive generations may suggest anticipation, a feature usually associated with trinucleotide-repeat expansion. Alternatively, considering the small number of individuals studied, we cannot exclude that this observation may be fortuitous. The phenomenon of anticipation has been proposed previously, to explain a similar observation of increased penetrance in subsequent generations in a family with iris coloboma (Barros-Nunez et al. 1995).

More than forty transcription factors are expressed during mammalian eye development (Freund et al. 1996; Jean et al. 1998). To our knowledge, none of these genes have been mapped to the candidate interval defined by loci D15S1002 and D15S1040 (Genome Database). Since the homozygous knockout of the Hairy Enhancer of Split homologue 1 (HES1) gene in mouse leads to abnormal eye development (Tomita et al. 1996), the TLE3 gene, which belongs to a family of corepressors coacting with HES proteins, should be considered as a candidate gene. However, the precise genetic location of the TLE3 gene with respect to our candidate interval remains to be determined (Richard et al. 1994; Liu et al. 1996). At least six expressed-sequenced tags (ESTs), previously obtained from retinal tissues, map within the D15S1002-D15S1040 candidate interval (GeneMap'99 database). One of these ESTs, stSG29857, is expressed in fetal retina and may contain cDNA sequence of a candidate gene.

It is interesting to note that severe syndromic bilateral microphthalmia was associated with an apparently balanced de novo reciprocal translocation 46, XY, t(1;15) (q41;q21.2), suggesting that either the 1q41 or 15q21.2 regions may contain a gene that is important for eye morphogenesis (Smith et al. 1994). On the other hand, interstitial deletions of chromosome 15q13-q15 (Autio et al. 1988), 15q12-q14 (Tonk et al. 1995), and 15q21.2-q22.1 (Martin et al. 1990) were not associated with eye malformation, suggesting that this disorder is not the result of haploinsufficiency of a gene in this region.

Since a large number of genes appear to be involved in vertebrate eye development, the linkage analysis presented here in a large family with autosomal dominant isolated microphthalmia provides an opportunity to focus on a specific area of the genome, namely the proximal long arm of chromosome 15. Efforts are being undertaken with the aim of identifying the disease gene itself.