To the Editor,

Pendred syndrome (PDS) is characterized by recessive congenital/pre-lingual sensorineural hearing impairment (HI), variable vestibular dysfunction, temporal bone abnormalities, and euthyroid goitre in late childhood/early adulthood [1]. HI and thyroid disease show considerable inter- and intra-familial variations. PDS is caused by mutations in SLC26A4 on chromosome 7q31.1, which can also cause DFNB4, characterized by recessive non-syndromic sensorineural HI, enlarged vestibular aqueduct (EVA) and variable vestibular dysfunction [1]. SLC26A4 mutations may account for 7–13% of all congenital HI and are the second most common cause of genetic HI. SLC26A4 (21 exons) encodes the anion exchanger pendrin. About 200 different mutations causing PDS/DFNB4 have been reported throughout SLC26A4 and many are private or only detected in few families.

Here, we report the first large study of a Scandinavian PDS/DFNB4 group by SLC26A4 sequencing of 109 unrelated probands mainly of Danish origin with HI and suspected for PDS/DFNB4 (Tables S1–S3 and Appendix S1, Supporting Information). Eighty-five percent of the probands had had a CT scan (abnormal or normal), 58% had a perchlorate discharge test (PDT) result (abnormal or normal) and 95 (87%) had an abnormal result for one or both of these tests. For the large majority of probands, the reason for inclusion was that probands in addition to HI had either an abnormal CT scan (usually EVA or Mondini dysplasia), an abnormal PDT, an abnormal thyroid stimulating hormone (TSH) level, a goitre, and/or a family history of HI. Thus, 71% (78 of 109) had three or more of the criteria fulfilled. All probands were tested for GJB2 mutations prior to SLC26A4 sequencing.

Our study is comprehensive because all probands had sequencing of SLC26A4 performed, and SLC26A4 MLPA analysis [probands who carried only one/no SLC26A4 mutation(s)] as well as sequencing of FOXI1, KJCN10, and SIX1, in this order, when appropriate (i.e. in probands with a single SLC26A4 mutation). The patients without SLC26A4 mutations were also sequenced for SIX1 mutations because of clinical overlap between PDS and disease caused by SIX1 mutations, in particular the malformations of the cochlea. Furthermore, 78 of the probands were analysed in parallel with our establishment of the SLC26A4 sequencing analysis using a microarray test (APEX) for 198 HI-associated mutations in eight genes. When the SLC26A4 sequencing method was fully established this analysis was used as standard method for PDS analysis, and therefore all patients have not been analysed by APEX. Finally, our study is noteworthy because many patients were subjected to a solid clinical pre-selection, which probably contributed to the high detection rate (67%) of probands with SLC26A4 biallelic mutations.

In total, we discovered pathogenic mutations in both or one SLC26A4 allele(s) in 67% and 10%, respectively, of the probands. The most frequent mutations were p.T416P, p.V138F, p.L236P, p.E29Q, c.1001+1G>A, and p.E384G in this order. We found 29 different SLC26A4 mutations, and nine were novel (Table 1). Besides two splice-site mutations and one deletion, the novel mutations included six missense mutations for which all the wild-type amino acids showed evolutionary conservation (Fig. S1). Four of the missense mutations involved addition or omission of a proline (p.P525L, p.A725P, and p.L729P) or a charged amino acid (p.Q383E) residue, alterations previously suggested to functionally impair pendrin [2]. Furthermore, p.G493W and p.S657I were predicted as ‘probably damaging’ by PolyPhen-2. No novel large deletions/duplications were detected by MLPA, suggesting that these mutations are not frequent in SLC26A4. Finally, we identified a p.E125K de novo SIX1 mutation in a patient without SLC26A4 mutations. This is the second case of this mutation, first identified in a Tunisian family with variable HI and pre-auricular pits, reminiscent of the branchiootorenal syndrome [3]. The SIX1 patient identified here had bilateral moderate sensorineural HI and Mondini dysplasia. Over all, our patient and the members of the Tunisian family harbouring the same SIX1 mutation show similar phenotypes. The absence of biallelic SLC26A4 mutation in 11 probands (10%) could suggest causative mutations in the non-coding exon 1 or in regulatory regions of SLC26A4, or mutations in FOXI1 or KCNJ10 [1]. In the above probands, we therefore analysed these exons/genes, but did not identify any disease-causing mutations. These data and the fact that it has generally been difficult to correlate a specific SLC26A4 mutation to clinical symptoms support the idea that other, yet unidentified genetic, epigenetic, and/or environmental factors are involved in SLC26A4 pathogenesis [2]. In the group of probands without any SLC26A4 mutations, it is remarkable that 75% (9 of 12) of probands with both a CT and a PDT result had both these tests abnormal, the latter of which can provide a strong indicator of PDS, thereby further supporting the existence of additional and yet unidentified PDS-causing mechanisms. However, autoimmune thyroid disease could also be a possibility in some of these patients.

Table 1. SLC26A4 mutations detected in 109 PDS/DFNB4 probandsa,b
LocationNucleotide changePredicted effectZygosityPatients (n = 84)Frequency of mutation (%)c
  1. Hetero, heterozygote; Homo, homozygote.

  2. a

    Nomenclature of mutations is based on cDNA sequence (GenBank accession number NM_000441.1). Novel mutations are shown in boldface type. All novel mutations were absent in control individuals.

  3. b

    Numbering of nucleotides: +1 = A of start ATG codon.

  4. c

    The numbers are the percentage of the mutant allele relative to all SLC26A4 mutant alleles identified.

  5. d

    Mutation named del c.305-3465_c766+970 in Ref. [6] in a Lebanese patient. This patient turned out to be a relative to our patient and thus to harbour the same mutation.

  6. e

    p.L597S has been reported with conflicting assumptions regarding its function and pathogenicity.

Exon 2c.85G>Cp.E29QHetero106
Introns 3–6g.8038T_21596Tdeldp.G102DfsX4Hetero01.3
Exon 4c.412G>Tp.V138FHetero2317
Exon 6c.626G>Tp.G209VHetero21.3
Exon 6c. 707T>Cp.L236PHetero2014
Exon 6c.716T>Ap.V239DHetero10.6
Intron 6c.765+3A>T Hetero10.6
Exon 7c.775_779delGAGATp.E259FfsX6Hetero10.6
Intron 7c.918+2T>C Hetero10.6
Intron 8c.1001+1G>A Hetero106
Exon 9c.1003T>Cp.F335LHetero10.6
Exon 9c.1147C>Gp.Q383EHetero10.6
Exon 10c.1151A>Gp.E384GHetero85
Exon 10c.1197delTp.C400VfsX32Hetero11.9
Exon 10c.1226G>Ap.R409HHetero01.3
Exon 10c.1246A>Cp.T416PHetero2820
Exon 11c.1284_1286delTGCp.A429delHetero21.3
Exon 11c.1334T>Gp.L445WHetero43.8
Intron 11c.1341+1G>C Hetero10.6
Exon 13c.1477G>Tp.G493WHetero10.6
Exon 14c.1554G>Ap.W518XHetero01.3
Exon 14c.1574C>Tp.P525LHetero10.6
Intron 141614+1G>A Hetero42.5
Exon 16c.1790T>Cp.L597SeHetero31.9
Intron 16c.1803+2T>C Hetero10.6
Exon 17c.1970G>Tp.S657IHetero10.6
Exon 19c.2127delTp.F709LfsX12Hetero42.5
Exon 19c.2173G>Cp.A725PHetero21.3
continued on next page
Exon 19c.2186T>Cp.L729PHetero53

At present, Sanger sequencing is the standard for SLC26A4 molecular diagnosis, but other efficient methods such as high-resolution melting are available [4]. Development of next-generation sequencing for molecular diagnosis is expected to reduce the utility of methods such as APEX, which here (besides screening for common SLC26A4 mutations) was used to establish whether the HI, alternatively, could be explained by other genes.

The spectrum of SLC26A4 mutations in our Danish/Scandinavian cohort involves mostly missense mutations (59%), which is also observed in several other populations. The three most common mutations among European Caucasian patients, p.T416P, p.L236P, and c.1001+1G>A, constituted 20%, 14%, and 6% of mutated alleles, respectively. Interestingly, in our patient group, the second most frequent mutation is p.V138F, detected in 17% of the mutated alleles. This mutation, which has previously been reported in other ethnic groups, was shown to be a founder mutation in Central Europe and was reported with a frequency of 18% in a Czech cohort [5]. However, apart from this and the present study, no other studies have reported such a high frequency of the p.V138F mutation. In addition, p.E29Q (detected in 6% of mutated alleles) was more common in our population than in other populations. Haplotype studies for both p.V138F and p.E29Q were in favour of a founder origin (Tables S4 and S5). In conclusion, we report that Denmark/Scandinavia may have a distinct mutation spectrum for PDS. Our characterization of the SLC26A4 mutation spectrum in Denmark has streamlined our diagnostic procedure for Danish patients such that we first sequence SLC26A4 exons 2, 4, 6, 8, 10, and 19, because they harbour 80% of all detected mutations, among the other prevalent p.E29Q and p.V138F mutations.


We are grateful to the families who participated, without whom the study would not have been possible. We acknowledge the technical assistance of Elvira Chapka, ICMM. We acknowledge all the physicians (Michael Bille, Anne-Marie Hauch, Bo Walter, Steen Gimsing, Birgitte Bech and others) who referred patients for SLC26A4 testing over many years. The study was financed by the Oticon Foundation, Widex AS, a grant (in 2002) from ‘Ingeborg and Emanuel Jensens Legat’, a grant (in 2002) from ‘Musikforlægger Agnes og Knut Mørk's legat’ (N. D. R., M. L., L. N., and L. T.) and by the Pete and Arlene Harman Pediatric Clinical Scholar Award (I. S. and J. R. P.). The project was hosted by Wilhelm Johannsen Centre for Functional Genome Research, established by the Danish National Research Foundation.

  • ND Rendtorffa

  • I Schrijverb,c

  • M Lodahla

  • J Rodriguez-Parisb

  • T Johnsend

  • EC Hanséene

  • LAA Nickelsena

  • Z Tüumerf

  • T Fagerheimg

  • R Wetkeh

  • L Tranebjærga,i

  • aWilhelm Johannsen Centre for Functional Genome

  • Research, Department of Cellular and Molecular Medicine,

  • The Panum Institute, University of Copenhagen,

  • Copenhagen, Denmark,

  • bDepartment of Pathology,

  • cDepartment of Pediatrics, Stanford University School of

  • Medicine, Stanford, CA, USA,

  • dGl. Kongevej 25, Copenhagen, Denmark,

  • eAudiologiska Klinikken Örebro Universitetssjukhuset,

  • Örebro, Sweden,

  • fCenter for Applied Human Molecular Genetics, Kennedy

  • Center, Glostrup, Denmark,

  • gDivision of Child and Adolescent Health, Department of

  • Medical Genetics, University Hospital of North Norway,

  • TromsøNorway

  • ,

  • hØre-Næse-Hals-Sygdomme, Afdeling H, Århus Sygehus,

  • Århus, Denmark, and

  • iDepartment of Audiology, H:S Bispebjerg Hospital,

  • Copenhagen, Denmark


  1. Top of page
  2. References
  3. Supporting Information
  • 1
    Alasti F, Van Camp G, Smith RJH. Pendred Syndrome/DFNB4. In: Pagon RA, Bird TD, Dolan CR, Stephens K, Adam MP, eds. GeneReviews™ [Internet]. Seattle, WA: University of Washington, 1998, from Accessed on August 30, 2012.
  • 2
    Pera A, Dossena S, Rodighiero S et al. Functional assessment of allelic variants in the SLC26A4 gene involved in Pendred syndrome and nonsyndromic EVA. Proc Natl Acad Sci U S A 2008: 105: 1860818613.
  • 3
    Mosrati MA, Hammami B, Rebeh IB et al. A novel dominant mutation in SIX1, affecting a highly conserved residue, result in only auditory defects in humans. Eur J Med Genet 2011: 54: e484e488.
  • 4
    Chen N, Tranebjaerg L, Rendtorff ND, Schrijver I. Mutation analysis of SLC26A4 for Pendred syndrome and nonsyndromic hearing loss by high-resolution melting. J Mol Diagn 2011: 13: 416426.
  • 5
    Pourova R, Janousek P, Jurovcik M et al. Spectrum and frequency of SLC26A4 mutations among Czech patients with early hearing loss with and without Enlarged Vestibular Aqueduct (EVA). Ann Hum Genet 2010: 74: 299307.
  • 6
    Siem G, Fagerheim T, Jonsrud C et al. Causes of hearing impairment in the Norwegian paediatric cochlear implant program. Int J Audiol 2010: 49: 596605.

Supporting Information

  1. Top of page
  2. References
  3. Supporting Information
cge12074-sup-0001-TableS1.docWord document136KTable S1. Genetic and clinical features of 73 probands with biallelic SLC26A4 mutation and hearing impairment.
cge12074-sup-0002-TableS2.docWord document61KTable S2. Genetic and clinical features of 11 probands with one SLC26A4 mutation identified.
cge12074-sup-0003-TableS3.docWord document75KTable S3. Clinical data of 25 probands with no SLC26A4 mutation identified.
cge12074-sup-0004-TableS4.docWord document56KTable S4. Chromosome 7q31 haplotypes linked to p.V138F.
cge12074-sup-0005-TableS5.docWord document56KTable S5. Chromosome 7q31 haplotypes linked to p.E29Q.
cge12074-sup-0006-TableS6.docWord document55KTable S6. Summary of clinical findings in 109 Pendred probands.
cge12074-sup-0007-TableS7.docWord document55KTable S7. List of all primers used in the study.
cge12074-sup-0008-AppendixS1.docWord document117KAppendix S1. Material, methods and results details.
cge12074-sup-0009-FigureS1.pdfPDF document74KFig. S1. (a) Multiple sequence alignments illustrating evolutionary conservation of pendrin amino acids p.Q383, p.G493, p.P525, p.S657, p.A725, and p.L729. (b) Representative sequence chromatograms for each of the identified SLC26A4 and SIX1 missense mutations. The arrows indicate the nucleotide changes of the heterozygous missense mutations.
cge12074-sup-0010-FigureS2.pdfPDF document98KFig. S2. Identification of a homozygous intragenic deletion in SLC26A4 (patient 5382-05 in Table S1). (a) Sequencing chromatogram of the exons 4–6 deletion junctions in SLC26A4 introns 3 and 6 (located in a L1PA4, family L1 repeat), respectively. The deletion removes a total of 13,558 bp, 53 bp more at the 5′ end and 549 bp less at the 3′ end, when compared to a deletion in a Spanish patient, who also had deletion of exons 4–6. This supports the idea that the genomic regions surrounding the breakpoints contain recombining elements. Coordinates numbering is based on the genomic sequence of SLC26A4 (NT_007933). (b) Confirmation of exons 4–6 deletion of SLC26A4 by MLPA analysis. The probe mix for SLC26A4 includes probes for the 21 SLC26A4 exons and 15 control probes (indicated by asterisks). The patient is homozygous for a deletion covering exons 4–6 in SLC26A4, as revealed by the missing peaks obtained from these three exons when compared to the control individual. To our knowledge, only five large deletions in SLC26A4 have been reported. However, until recently, PDS patients have not been systematically analysed for large deletions/duplications.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.