Statistics from Altmetric.com
Editor—The branchio-oto-renal (BOR) syndrome was first reported in the last century by Heusinger, but not clinically defined until 1976 by Melnick et al.1 The major clinical features associated with the BOR syndrome are hearing loss, branchial defects, ear pits, and renal anomalies.1 Hearing loss can be conductive, sensorineural, or mixed, ranging from mild to profound deafness.2 Renal abnormalities are also variable, including renal aplasia, hypoplasia, and dysplasia, as well as anomalies of the collecting system.3 Branchial defects including fistulas or cysts and ear pits are often observed, and minor features such as external ear abnormalities, ear tags, and lacrimal duct aplasia are sometimes present. The BOR syndrome is inherited in an autosomal dominant manner and penetrance is high, although expressivity can be extremely variable.3-5
The BOR syndrome gene was localised to 8q following the description of a person with an inherited rearrangement, dir ins(8)(q24.11:q13.3:q21.13) presenting with features of both the branchio-oto (BO) syndrome and trichorhinophalangeal syndrome.6 Linkage analysis confirmed that the BOR syndrome mapped to 8q13.3.7 8 Further fine mapping defined the BOR region to be a 450-650 kb interval.9-12In 1997, the gene was cloned by sequencing P1/PAC clones from a contig spanning the region, yielding sequences homologous to theDrosophila developmental geneeyes absent (eya).Point mutations and deletions were subsequently detected in the human homologue, EYA1, in families with the BOR syndrome.13 14
EYA1 consists of 16 coding exons and encodes a 559 amino acid protein. There are two additionalEYA1 isoforms and all show significant homology to the Drosophila eya gene, as well as to the murine homologue, constituting a novel gene family.14 15 The Drosophila eya gene is essential in the formation of the fly compound eye and the pathway has been shown to be conserved between flies and vertebrates.16-18 Preliminary results indicate thatEYA1 is highly expressed in human fetal kidney and brain, as well as in adult heart and skeletal muscle. The murine homologue Eya1 was found to be strongly expressed in the mouse ear and skeletal muscle, but not shown in mouse kidney or branchial tissues.13 The product ofEYA1 contains a highly conserved region called the eyes absent homologous region (eyaHR). This is a 271 amino acid carboxy-terminal region encoded within exons 9-16 and is the site of most mutations identified to date.11 12 14 15
In this study, mutation detection of theEYA1 gene was performed on a collection of cases with a probable diagnosis of the BOR syndrome and clinically overlapping disorders. We aimed to assess whether cases with overlapping phenotypes are allelic to the BOR syndrome at a molecular level.
The cases for the study were ascertained from several sources, with 16 cases (familial and sporadic) recruited for clinical and mapping studies before gene identification.4 13 The remaining 16 cases were identified subsequently through the clinical genetics and nephrourology departments at Great Ormond Street Hospital, London, UK.
Thirty two cases were studied in total, of which 18 (11 familial, seven sporadic) exhibited at least three of the following major features: hearing loss, branchial defects, ear pits, and renal anomalies. These cases were classified as having classical BOR syndrome. Twelve cases (mostly sporadic) exhibited one or two major features, often associated with a minor feature (most commonly external ear anomalies). These were designated as having an atypical form of the disease. Two cases of otofaciocervical (OFC) syndrome were also included. This is an autosomal dominant condition in which there is hearing loss, ear pits, cervical fistulae, hypoplasia of the cervical musculature (sloping shoulders), facial abnormalities, short stature, and mild developmental delay. There is considerable phenotypic overlap between the OFC syndrome and the BOR syndrome.19 A summary of the clinical features of the cases are given in tables 1 and 2.
Exons 1′-3 and 5-16 were amplified from genomic DNA of affected subjects using published intronic primers.14 Exon 4 was amplified using primers EYA1-4F-gtgatgtggttgttaatcggt and EYA1-4R1-agaaggtgacaacacgttctaaatt. All PCRs were carried out under standard conditions. In familial cases, one affected member was initially analysed and then confirmed in other family members where possible.
The PCR products from exons 1′-15 were denatured and run at 15 W overnight at 4°C on 1 × MDE gels (FMC BioProducts) using 0.5 × TBE, with and without 10% glycerol. The gels were then silver stained as previously described.20 Owing to its large size, exon 16 was digested with DdeI and run as above. Samples with SSCP mobility shifts were then sequenced using an ABI 377 DNA sequencer, using the drhodamine terminator cycle sequencing kit (PE Applied Biosystems). Subjects without SSCP shifts were then sequenced for all exons listed above. Exons 11/12 and 13/14 were sequenced together as the introns between them are 100 bp. Mutant alleles in family 3 and subject 5 were sequenced by excising small fragments of the resolved bands from the SSCP gel, adding straight to a PCR premix, and amplifying.
Of the thirty two cases analysed, twenty nine underwent SSCP analysis, yielding 10 mobility shifts that were not present in 50 normal controls. The remaining three cases who did not undergo SSCP analysis were sequenced for the entire gene. All SSCP shifts were then directly sequenced and subjects without mobility shifts were sequenced for the entire coding region of the gene. An eleventh mutation was identified by sequencing in family 10, which was not detected by SSCP. The mutations are listed in table 1. All mutations were confirmed using either a naturally occurring restriction enzyme site or by designing an artificially created restriction site (ACRS). This involved the use of a mismatched primer to introduce a restriction enzyme site in the presence of the mutation. The primers and enzymes used are shown in table 3.
Subject 1 and family 7, who are apparently unrelated, carried the same mutation, 790C→T, in exon 8. This results in the creation of a stop codon, which is predicted to cause premature truncation of the protein. The parents of subject 1 were shown not to carry the mutation and non-paternity was excluded. Only one other relative was available for testing in family 7. Family 7 consisted of six affected subjects with deafness, cervical fistulas, and ear pits, but DNA was only available from one relative who was clinically unaffected and who was shown not to carry the mutation. A summary of the sequence and enzyme data is shown in fig 1.
Another nonsense mutation in exon 8, 732C→A, resulted in the substitution of a tyrosine residue for a termination codon. The parents were clinically unaffected, but unavailable for testing.
A missense mutation in exon 16, 1680A→C, in family 8 is predicted to destroy the final stop codon of the gene. The change creates a tyrosine residue, resulting in the addition of five extra amino acids at the end of the protein.
Another missense mutation in exon 16 was identified in subject 11. The change, 1649 T→C, results in a leucine to proline amino acid substitution at position 549 of the gene. No other mutations were found in subject 11 when the remaining coding exons were sequenced, and the change was not present in 85 normal controls tested. Her parents have yet to be tested.
Family 3 was previously reported as having a complex 5 bp substitution/insertion (1372T→AGAGC) in exon 13.13Resequencing of the mutant allele has shown the insertion to be 6 bp (AGAGAC). This combined with the loss of a thymine nucleotide at the same position resulted in an overall gain of 5 bp. This change is predicted to result in a frameshift leading to premature truncation of the protein. All three affected family members were heterozygous for the mutation, while those unaffected were not. Subject 6 was also heterozygous for a frameshift mutation which was the result of an insertion of a T at position 387 in exon 5.
Family 4, family 10, and subject 9 were heterozygous for mutations which are predicted to cause aberrant splicing in exons 9, 14, and 10 respectively. All affect conserved bases and would be predicted to be disease causing. Computer analysis involving a neural network program was used to analyse DNA sequences containing the mutated splice sites and surrounding sequence (50-100 bp). This predicted that the changes would result in the destruction of the consensus splice site and create no other donor/acceptor sites.21
A fourth splice site mutation was detected in subject 5, a 23 bp inversion starting at position 1042-13, across the intron/exon boundary of exon 11. This is predicted completely to disrupt the consensus acceptor site of exon 11 and was not present in the unaffected parents; non-paternity was excluded. This is shown in fig 2.
There have been seven exonic EYA1polymorphisms published to date. Sequencing of our cohort has confirmed the existence of published polymorphisms 510A→C, 1179C→T, 1233T→C, and 1656T→C.14 We have identified a new polymorphism in exon 7, 714A→G. This sequence variant occurred in 4/32 of the branchial arch syndrome cases and in 3/20 normal controls.
Comprehensive mutation detection of our cohort has identified 11 mutations from 18 cases with classical BOR syndrome. These 11 cases exhibited at least three of the following major clinical features: deafness, branchial defects, ear pits, and renal anomalies, and gives a detection rate of 61% which is higher than in previous studies.13 14 22 23 There were no apparent clinical differences in phenotypic features between the 11 cases with mutations in EYA1 and the seven classical cases without. It is possible that these seven cases may contain major rearrangements of the gene or mutations in the 3′ untranslated or promoter region.
No mutations were detected in any subjects with atypical BOR syndrome or the OFC syndrome. Atypical BOR syndrome cases exhibited only one or two major features of the disease often associated with minor features, most commonly external ear abnormalities. Careful clinical evaluation is therefore essential before determining whether it is appropriate to embark upon mutation screening of theEYA1 gene. However, while sporadic cases of BOR syndrome may not fulfil these strict clinical diagnostic criteria, if similar affected members within large families have between them at least three major features, mutation screening would be deemed appropriate. It is therefore possible that such strict clinical diagnostic criteria would result in failure to screen sporadic cases presenting with one or two major features. However, we found no mutations in 13 such sporadic cases studied.
When the results of our mutation detection are combined with published data, the vast majority of mutations occur in exons within the eyaHR, that is, exons 9-16.13 14 22 23 This region shows 69% identity with the Drosophila eya protein and appears essential for normal function of the gene product,13 as almost all mutations are clustered in exons within or adjacent to it. Exons 8, 13,14, and 15 contain the largest number of mutations and with one exception all are private.13 14 22 23 The missense mutation identified in subject 11 results in an amino acid substitution of leucine to proline in exon 16. The leucine residue is conserved both inC elegans andDrosophila and falls within the eyaHR. Sequencing showed no other changes in this person and the change was not present in 85 normal controls, making it highly likely that it is the disease causing mutation.
A previously reported mutation in a BO syndrome family (no renal anomalies) in exon 4 is the only change reported outside exons 8-16 of EYA1.24 Here, we have identified a frameshift in exon 5 in a classical BOR syndrome patient with a single unilateral kidney. Therefore mutations located outside exons encoding the eyaHR and immediately adjacent exons do not appear to result in different renal manifestations. No mutations have yet been identified in exons 1′-3 and in exons 6 and 7.
The value of molecular testing shown here is that it can confirm a diagnosis of the BOR syndrome and provide genetic recurrence risk information to families or people. However, variable expressivity is a feature of BOR syndrome and it is not possible to predict the severity of the phenotypic features, even when the mutation is identified in the family.3 5 This is highlighted by the occurrence of the same mutation in the following two unrelated families in our study. In family 7, there are no reported renal defects, but a history of deafness, cervical fistulae, and ear pits. However, the sporadic case subject 1 presented with chronic renal failure requiring a kidney transplant at 6 years of age and was noted to have deafness, ear pits, and cupped ears. The identification of the mutation in subject 1 has proven her to be a sporadic case of the condition and greatly reduces the recurrence risk to her parents. Although the possibility of gonadal mosaicism in BOR syndrome cannot be excluded, it has not been reported.
Even with strict clinical criteria, diagnostic uncertainty can still remain as deafness, ear pits, and renal anomalies occur at a relatively high frequency as isolated features in the general population.25 26 In our series, two cases of such uncertainty have been resolved by molecular testing. In subject 9, the donor splice site mutation has been previously reported and was initially thought to be familial, owing to a maternal family history of ear pits.13 We have proved the mutation to be de novo, as it is not present in his parents or sib. In family 8, there is a three generation history of deafness, branchial sinuses and cysts, ear pits, renal abnormalities, facial asymmetry, and abnormal ears. The son of the proband presented with renal abnormalities, but has been shown not to carry the mutation seen in his affected mother. He is therefore very unlikely to be affected with the BOR syndrome and has an unrelated isolated renal anomaly. Molecular testing has allowed confirmation of diagnosis of the BOR syndrome allowing accurate recurrence risks in some of the families in our study.
The absence of renal involvement in some cases withEYA1 mutations and the difference in renal abnormalities in the two cases with the same mutation (family 7 and subject 1) highlights the variable renal manifestations in the BOR syndrome. Two cases of the BO syndrome withEYA1 mutations have been reported,24 but BO syndrome families not mapping to theEYA1 locus have also been reported.27 Although it is likely that the BOR and BO syndromes are allelic mutations of the EYA1gene, other unknown genes can cause branchial arch syndromes. Our study has examined cases with similar phenotypes and is supportive of this hypothesis as no mutations were identified in cases with atypical BOR syndrome (including cases of the OFC syndrome, first and second branchial arch syndrome, and cases of deafness associated with renal defects). Such cases are most unlikely to haveEYA1 mutations and so until mutation detection strategies (including deletion screening and analysis of non-coding regions) yield higher detection rates, screening is probably best limited to cases of classical BOR syndrome. Further research intoEYA1 and its role in branchial arch, ear, and kidney formation is essential in order for us to understand the factors which influence phenotype and variable expressivity of the BOR syndrome.
SR was funded by the Special Trustees of Great Ormond Street Hospital. MB-G is an MRC Clinician Scientist. We would like to thank referring clinicians in clinical genetics and nephrourology and the families for their help. Thanks also to Dr Mary Davis and Professor Marcus Pembrey for collecting many of the families and their initial involvement. Additional thanks to Paul Rutland for his technical assistance and Peter Logan for paternity testing relevant families.
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.