Article Text


Potential mapping of corneal dermoids to Xq24-qter

Statistics from

Editor—Corneal dermoids (CND) (MIM 304730) are rare, congenital, benign tumours involving the cornea. Histologically, the dermoids consist of a combination of ectodermal elements including keratinised epithelium, hair, and sebaceous glands and mesodermal elements including fibrous tissue, fat, and blood vessels in different proportions The tumour is by definition a choristoma since it is histologically a normal tissue in an abnormal site. Clinically, these tumours appear as opacification of the cornea at birth and if untreated may result in blindness (fig 1). Treatment includes surgical removal of the tumour with corneal transplantation. When the tumours affect both eyes, surgical intervention is indicated within the first 3 months of life since untreated cases may be irreversible by that time.

Figure 1

Corneal dermoids as seen immediately after birth in subject IV.8.

Most CND cases are sporadic and unilateral. However, three hereditary forms associated with congenital dermoids involving the cornea have been described. The Goldenhar syndrome (oculoauriculovertebral dysplasia, MIM 164210) is a multiple anomaly syndrome that involves the eyes, ears, and the vertebra.1 The dermoids are mostly unilateral, originate in the corneal-scleral border, also known as the limbus, and rarely affect the centre of the cornea. Although most cases are sporadic, familial cases with an autosomal dominant mode of inheritance have been reported. The second form with hereditary dermoids of the cornea is ring dermoids (MIM 180550), an autosomal dominant condition with isolated dermoids at the periphery of the cornea at the limbus.2

A third form of congenital dermoids involving the cornea was initially described by our institution in a family of Puerto-Rican ancestry.3 The three affected males had congenital, bilateral, and central corneal opacification at birth and no additional abnormalities.4 The described pattern of transmission was suggestive of X linked recessive inheritance. To date, this is the only family described with this inherited form of CND. Previous efforts to localise the gene on the X chromosome using RFLPs for linkage analysis5 gave a lod score of 2.4 at θ=0 with the DXS43pD2 probe at Xp22.2. Here, we report the analysis of the extended family to establish linkage.

Case reports

Twenty three members of a four generation family were ascertained in our centre. Clinical data from the 15 subjects who participated in the previous linkage study5 were available to us but blood samples were available for only 10 out of the 15. II.2 and II.10 declined participation in the current study. In addition, we studied seven family members, including three affected males, not previously described (IV.2, IV.3, IV.5-IV.9). This study was approved by the Albert Einstein College of Medicine Institutional Review Board, in accordance with the guidelines of the Office for the Protection from Research Risks, and informed consent was obtained before participation (CCI 91-157).

All family members including carriers and unaffected members underwent full ophthalmic examination in our centre and ophthalmic records and histopathology slides of affected males were reviewed. We identified six males out of 23 family members who were affected with congenital corneal dermoids. The 17 unaffected family members did not exhibit any medical or ophthalmic condition that could be attributed to or associated with CND. The pattern of transmission is consistent with an X linked recessive mode of inheritance as indicated by the presence of only affected men, an absence of male to male transmission, and unaffected obligate carrier females (fig 2). In all cases the dermoids appeared as a bilateral, superficial, greyish layer with irregular, raised, whitish plaques and fine blood vessels that covered the centre of the cornea. The peripheral border of the cornea was intact in all cases. The ophthalmic examination of affected males was consistent with the previously described examination of subjects III.8, III.9, and IV.1.3 4 Briefly, ocular examinations showed normal adnexal structures, eyes were orthophoric to Hirschberg testing, and eye movements were conjugate and normal. Schiotz tonometry ranged from 20 to 22 mmHg bilaterally. Anterior chambers and fundi could not be seen because of the opacification, but intraoperatively were described as normal and clear. The severity of the disease corresponded with the length of time from birth to surgical intervention. All affected subjects had had at least one corneal transplant. Subjects IV.7 and III.8 had had four and five operations, respectively, and both are considered legally blind.

Figure 2

Haplotype analysis of the four generation family segregating X linked CD. The at risk haplotype is blackened to show recombinations. Carrier females are indicated by a black dot within a circle and affected males are indicated by blackened squares. Unblackened squares and circles without a black dot denote unaffected males and females not known to be carriers, respectively. Asterisks show subjects who were analysed in the previous report but for whom blood samples were not available for current linkage analysis.

Microscopic examination of affected corneas was consistent with a dermis-like tissue. It showed a highly vascularised, dense, and irregular collagenous connective tissue in all cases with sparing of the corneal periphery. Hair follicles and hair shafts were present in two cases and sebaceous glands were present in three. The outer layer of stratified squamous epithelium was found to be keratinised in three cases and Bowman's membrane was absent in these cases and in one additional case. The two inner layers, Descemet's membrane and the endothelium, were intact in all cases.

A total of 17 subjects were subjected to genotype analysis. We initially studied the previously reported Xp22.2 locus with the DXS43pD2 probe which detects two PvuII alleles of sizes 6.2 and 5.6 kb on a genomic Southern blot6 and with nine additional microsatellite markers flanking DXS43pD2 at Xp22.1-p22.3. Briefly, polymorphic markers and genetic distances were obtained from the Centre for Medical Genetics, Marshfield7 and the Genome Database (http:\\ For genotyping, the reverse primers were radiolabelled with [32P]-ATP, and the PCR product was amplified under standard reaction conditions.8 The PCR product was denatured by adding 20 μl of stop solution containing 5:1 phosphamide:EDTA and then heating for 10 minutes at 65°C. The radiolabelled PCR products were separated on 6% acrylamide denaturing sequencing gels, and alleles were assigned according to their molecular weight.

Linkage analysis was performed assuming an X linked recessive mode of inheritance with complete penetrance and without phenocopies. The frequency of the disease predisposing allele was set to 0.0001. Marker allele frequencies were estimated from the founders. Disease marker pairwise linkage analysis was performed using MLINK from LINKAGE v 5.29-11 using a range of recombination fractions. Linkage analysis for the DXS43pD2 marker included the results of the five additional subjects previously studied with this probe, as the genotyping results and slides were available to us and could be reviewed and confirmed. Multipoint linkage analysis was performed using only markers which are of known order and intermarker recombination fractions using data from the Marshfield genetic map7( using LINKMAP from LINKAGE v 5.2.9-11 Because of computational limitations, a “shifting” three point analysis was performed.12Analysis of the previously suggested DXS43pD2 probe showed two obligatory meiotic recombination events between the disease locus and the marker in subjects IV.3 and IV.8 not previously tested. Analysis of the additional nine markers flanking DXS43pD2 showed significantly negative lod scores at θ=0.00 and a negative pattern for linkage at Xp22.1-p22.3 (table 1). As disclosed from the new haplotypes, the two point linkage analysis excludes the postulated XLCD locus on Xp22 as the site for the XLCD gene in this family. Thereafter, a systematic genotype analysis was undertaken with 22 microsatellite markers, separated by 5-20 cM, spanning the X chromosome. The disease status in the X linked CND family came into phase as markers from the Xq24-qter region were examined (fig 2). A two point linkage analysis gave positive lod scores of 2.9 for DXS102, DXS1232, and DXS8377 at a recombination fraction of θ=0.00 (table 2). The centromeric boundary was defined by the marker DXS1001 on Xq24 (approximately 45 cM from the telomere). No meiotic crossovers were observed with the 10 markers in the region DXS8057-DXS8377, a ∼38 cM interval. In agreement with the two point linkage results, multipoint analysis gave a very flat lod score curve with a maximum of 2.92 over the non-recombinant interval between DXS1062/DXS8094 and DXS8028, a ∼25 cM interval (fig 3).

Table 1

Two point lod scores between XLCD and markers flanking DXS43pD2 on Xp22.2

Table 2

Two point lod scores between XLCD and markers on Xq21-qter

Figure 3

Multipoint analysis of a 70 cM region encompassing the X linked CD locus on Xq24-Xqter. The multipoint Zmax 2.92 was obtained with an interval between DXS1062 and DXS8028, a 25 cM interval. The physical distance of this interval is 3.7 Mb (The Genome Data Base


In conclusion, we mapped the gene for CND to chromosome Xq24-qter, within a ∼45 cM region, by using haplotype and linkage analyses. Although CND was mapped to a large interval with six markers with lod scores of more than 2.8 at θ=0.00, our findings re-emphasise the caution that is required in interpreting linkage results in a single family. The addition of new members to this family and their inclusion in the study showed two new recombinations in a locus previously suggested to be linked to the disorder.

The unique features of the corneal dermoids in this family were the bilaterality of the tumours and their localisation in the central matrix of the cornea while sparing the periphery of the cornea, Descemet's membrane, and the endothelial membrane. Since corneal dermoids are considered to be an aberrant development of ectopic tissue in the cornea, it can be speculated that it results from a mutation in a gene that has a role in the normal differentiation of the matrix component of the cornea. The interval found to be linked toCND is currently known to contain 119 genes of which 38 have protein products with unknown function or an abnormal function which has not been implicated in any known disorder. Examination of the known data regarding these genes failed to pinpoint any target gene that may have a role in corneal development. Since there are no additional informative recombinations in this interval, the candidate region cannot be further narrowed and only additional families with XLCD can assist in advancing mapping and identification of the XLCD gene.

  • A four generation family with congenital corneal dermoids segregating as an X linked trait was studied.

  • Of 23 family members, bilateral and central dermoid tumours of the cornea were identified at birth in six males.

  • Linkage analysis performed for 17 subjects localised the gene for the disease to a 45 cM interval distal to DXS1001.


Electronic database information. Online Mendelian Inheritance in Man, OMIM (TM). McKusick-Nathans Institute for Genetic Medicine, Johns Hopkins University (Baltimore, MD) and National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD), 2000. Center for Medical Genetics, Marshfield, Wisconsin, Genome database, The Genetic Location Database, Collinset al, 1996,

A D Paterson is supported by an MRC (Canada) Program Grant “The Centre for Applied Genomics”.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.