Clin Genet 2014 Printed in Singapore. All rights reserved

© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd CLINICAL GENETICS doi: 10.1111/cge.12437

Original Article

Mutation update and uncommon phenotypes in a French cohort of 96 patients with WFS1-related disorders Chaussenot A, Rouzier C, Quere M, Plutino M, Ait-El-Mkadem S, Bannwarth S, Barth M, Dollfus H, Charles P, Nicolino M, Chabrol B, Vialettes B, Paquis-Flucklinger V. Mutation update and uncommon phenotypes in a French cohort of 96 patients with WFS1-related disorders. Clin Genet 2014. © John Wiley & Sons A/S. Published by John Wiley & Sons Ltd, 2014 WFS1 mutations are responsible for Wolfram syndrome (WS) characterized by juvenile-onset diabetes mellitus and optic atrophy, and for low-frequency sensorineural hearing loss (LFSNHL). Our aim was to analyze the French cohort of 96 patients with WFS1-related disorders in order (i) to update clinical and molecular data with 37 novel affected individuals, (ii) to describe uncommon phenotypes and, (iii) to precise the frequency of large-scale rearrangements in WFS1. We performed quantitative polymerase chain reaction (PCR) in 13 patients, carrying only one heterozygous variant, to identify large-scale rearrangements in WFS1. Among the 37 novel patients, 15 carried 15 novel deleterious putative mutations, including one large deletion of 17,444 base pairs. The analysis of the cohort revealed unexpected phenotypes including (i) late-onset symptoms in 13.8% of patients with a probable autosomal recessive transmission; (ii) two siblings with recessive optic atrophy without diabetes mellitus and, (iii) six patients from four families with dominantly-inherited deafness and optic atrophy. We highlight the expanding spectrum of WFS1-related disorders and we show that, even if large deletions are rare events, they have to be searched in patients with classical WS carrying only one WFS1 mutation after sequencing. Conflict of interest

The authors have declared no conflicting interests.

A. Chaussenota,b† , C. Rouziera,b† , M. Querea , M. Plutinoa , S. Ait-El-Mkadema,b , S. Bannwartha,b , M. Barthc , H. Dollfusd , P. Charlese , M. Nicolinof , B. Chabrolg , B. Vialettesh and V. Paquis-Flucklingera,b a Department

of Medical Genetics, National Centre for Mitochondrial Diseases, Nice Teaching Hospital, Nice, France, b IRCAN UMR CNRS 7284/INSERM U1081/UNS, School of Medicine, Nice Sophia-Antipolis University, Nice, France, c Department of Biochemistry and Genetics, Angers University Hospital, Angers, France, d Department of Medical Genetics, EA INSERM 3949, Strasbourg, France, e Department of Genetics, Pitié-Salpêtrière hospital, Pierre and Marie Curie University, Paris, France, f Pediatric Endocrinology, Diabetology and Metabolic Diseases Department, Mère-Enfant Hospital, Lyon, France, g Neuropaediatrics and Metabolic Diseases Department, Timone Hospital, Marseille, France, and h Nutrition, Metabolic Diseases and Endocrinology Department, Timone Hospital, Marseille, France † These

authors contributed equally to the

study.

Key words: large-scale rearrangements – uncommon phenotype – WFS1 – Wolfram syndrome Corresponding author: Prof. Véronique Paquis-Flucklinger, IRCAN UMR CNRS 7284/INSERM U1081/UNS, School of Medicine, 28 av de Valombrose, 06107 Nice cedex 2, France.

1

Chaussenot et al. Tel.: +33 4 93 37 77 86; fax: +33 4 93 37 70 33; e-mail: [email protected] Received 20 March 2014, revised and accepted for publication 26 May 2014

The WFS1 gene is mainly involved in Wolfram syndrome (WS) (OMIM #222300), an autosomal recessive neurodegenerative disease, also known as DIDMOAD. The minimum ascertainment criteria of WS are the occurrence of diabetes mellitus (DM) and optic atrophy (OA) before the age of 15 years, that are usually associated with diabetes insipidus (DI), deafness, renal tract abnormalities or neuropsychiatric disorders (1, 2). WFS1 encodes an 890-amino-acid long transmembrane protein (wolframin) localized in the endoplasmic reticulum (3). WFS1 dysfunction induces high levels of endoplasmic reticulum stress, activating the unfolded protein response, and affects insulin secretion and processing. More than 150 different mutations have been identified in WFS1 since its identification as a disease gene in 1998 (https://lovd.euro-wabb.org/home.php? select_db=WFS1). Recently, de Heredia et al. reviewed clinical and genetic data from 412 patients with WFS1-related disorders, extracted from 49 references published since 1998 (4). They listed 178 different mutations that are distributed all along the protein, mainly concentrated in transmembrane domains, N-end of the protein and the 100 last amino acids. Most of these mutations are loss-of-function mutations as stop, frameshift and splice site mutations. Missense mutations are also detected in approximately 35% of cases. de Heredia et al. reported that a single mutation was found in 8.31% of patients pointing to (i) the lack of detection of a second mutation, (ii) the existence of dominant mutations, or (iii) mutations in another gene. To date, only few data are available on the frequency of large-scale rearrangements that are not detected routinely. WFS1 mutations are also responsible for a growing number of phenotypes that differ from ‘classical’ WS, including late-onset WS (5). According to the recent review of de Heredia et al., 6.3% of published patients developed both DM and OA after 15 years old (20/316) (4). WFS1 mutations are also a common cause of autosomal dominant Low Frequency SensoriNeural Hearing Loss (LFSNHL) and more recently, association of OA and deafness secondary to WFS1 dominant mutations was reported (5, 6). Finally, WFS1 was implicated in other atypical phenotypes. Berry et al. reported a four-generation Irish family with isolated congenital cataract transmitted in an autosomal dominant way, associated with a novel missense mutation (p.Glu462Gly) (7). A Finnish family presenting with autosomal dominant non-syndromic adult-onset diabetes, secondary to another missense mutation (p.Trp314Arg), was described by Bonnycastle et al. (8) Lastly, Elli et al. described a complex structural rearrangement in a sporadic case, who presented with a neonatal DI, optic

2

tracts hypoplasia, psychomotor retardation and central hypothyroidism (9). Here, we report a mutation update in our French cohort of 96 patients with 15 novel deleterious putative mutations in WFS1, including a large deletion in a 6 year old patient with WS. We show that late-onset WS represents 13.8% of patients with WFS1-related disorders and a probable autosomal recessive transmission. We also confirm that the clinical spectrum related to WFS1 is constantly expanding. Patients and methods Patients

In the Department of Medical Genetics (Nice, France), we identified by sequencing at least one variant in the WFS1 gene in 96 patients (49 females, 47 males) from 75 families. Fifty-nine patients from 48 families had been described in previous studies (10–12), and 37 novel patients from 27 families have been included for this work (Table 1). When DM or OA began before the age of 15, patients were classified as EOWS (Early-Onset Wolfram Syndrome). In contrast, when DM and OA began at 15 years or more, patients were classified as LOWS (Late-Onset Wolfram Syndrome). Patients with probable autosomal recessive transmission presenting with OA and neurological symptoms, but without diabetes to adulthood, were classified as OA ‘plus’. A patient with multisystemic disorder including DM was in category ‘other’. The study was performed according to French bioethics law. Blood samples were obtained from all available family members after informed consent was obtained. Methods WFS1 and CISD2 genes sequencing

The coding regions of WFS1 (NM_006005.3) and CISD2 (NM_001008388.4) were sequenced as previously described (10, 13). A mutation was considered as new if it was neither present in the LOVD Database (http://lovd.euro-wabb.org/home.php?select_ db=WFS1), in the NCBI (http://www.ncbi.nlm.nih.gov/ sites/), UCSC (http://genome.ucsc.edu/cgi-bin/hgGate way) and EVS databases (http://evs.gs.washington.edu/ EVS/) nor published. To determine the pathogenicity of new variants, we used the following criteria: (i) the evolutionary conservation of the amino acid residue, (ii) the location of the amino acid residue in an important functional domain, (iii) the co-segregation of the variant with the disease within the family, (iv) in silico predictions by

WFS1 mutation update and uncommon phenotype Table 1. Repartition of our cohort according to heredity, phenotype and presence of one or two mutations in WFS1

Heredity Autosomal Recessive

WFS1-related disorders Total EOWS LOWS OA ‘plus’ Other

Dominant LFSNHL HI+/−OA+/−DM

Nb of patients carrying a Nb of already Nb of patients Nb of patient new mutation published with two with only (Nb of families) Nb of patients patients Nb of new patients mutations one mutation = > Nb of (Nb of families) (Nb of families) (Nb of families) (Nb of families) (Nb of families) new mutations 96 (75) 87 (70) 72 (57) 12 (11) 2 (1) 1 (1) 9 (5) 3 (1) 6 (4)

59 (48) 57 (47) 47 (39) 10 (9) 2 (1) − 2 (1) 2 (1)

37 (27) 30 (23) 26 (19) 3 (3) − 1 (1) 7 (4) 3 (1) 4 (3)

78 (61) 78 (61) 66 (51) 10 (9) 2 (1) − − − −

18 (14) 9 (9) 6 (6) 2 (2) 1 (1) 9 (5) 3 (1) 6 (4)

15 (13) = > 15 11 (10) = > 12 2 (2) = > 2

2 (1) = > 1

DM, diabetes mellitus; EOWS, early-onset Wolfram syndrome; HI, hearing impairment; LFSNHL, Low Frequency Sensory-Neural Hearing Loss; LOWS, late-onset Wolfram syndrome; Nb, number; OA, optic atrophy.

PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/) and SIFT (http://sift.jcvi.org/) and, (v) the absence of the variant in 100 normal controls from the Caucasian population. Genotype classification

According to de Heredia et al., mutations were assigned to one of the following types on the basis of their predicted effect on WFS1 expression: (i) type I, which leads to complete depletion of wolframin due to the activation of non-sense-mediated decay; (ii) type II, which leads to complete degradation of wolframin by keeping functional the degron in WFS1; and (iii) type III, leading to the expression of a defective or shorter WFS1 protein (4). Three main genotypic classes were defined: class A, no WFS1 protein produced; class B, reduced expression of a defective WFS1 protein; and, class C, expression of a defective WFS1 protein. In addition, class A was subdivided into three subclasses: class A1, WFS1 depletion due to WFS1 mRNA degradation; class A2, WFS1 depletion due to mRNA and protein degradation; and class A3, WFS1 depletion due to wolframin degradation (4). Real-time quantitative PCR

Because real-time quantitative PCR was performed as a routine analysis in our laboratory, we used this technique instead of Multiplex Litigation-dependent Probe Amplification for detection of large rearrangements. Eight fragments corresponding to the eight exons of WFS1 were amplified. All primers were designed using primer3 software (http://www-genome.wi.mit.edu/ genome_software/other/primer3.html) to amplify 100– 150 base-pair amplicons (Table S1, Supporting Information). The quantitative polymerase chain reaction (qPCR) reactions contained 10 μl of 2× Roche SYBR Green PCR Master Mix Roche (Meylan, France), 5 μl genomic DNA (4 ng/μl) and 1 μl of each primer

(10 pmol/μl) in a total volume of 20 μl. Real-time PCR was run using a Roche 480 Light Cycler Real-Time PCR System, for 45 cycles, consisting of 10 s at 95 ∘ C, 10 s at 60 ∘ C and 10 s at 72 ∘ C. Each sample was amplified in duplicate with primers designed for three control genes assay, ALB (NM_000477.5), SDHA (NM_004168.2) and ADORA2B (NM_000676.2) (Table S1). The data were normalized by setting a pooled genomic DNA to a fold change of 1.0. The pooled genomic DNA consisted in a mix of 10 male and 10 female DNA extracted from 20 healthy controls. If necessary, a more complete protocol is available on request. Identification of deletion breakpoints in patient WS7

To perform the array-CGH, genomic DNA was extracted from peripheral blood using the Gentra Puregene Extraction Kit (Qiagen, Hilden, Germany). Array-CGH analysis was performed using a 1 million 60-mer oligonucleotide probes array (SurePrint G3 Human CGH Microarray 1×1M, Agilent Technologies, Santa Clara, CA). Data were extracted using the Feature Extraction software and analyzed using cytogenomics software (both from Agilent Technologies). Further analysis was performed using the Cartagenia Bench software (Cartagenia Inc, Cambridge). Analysis of the deletion breakpoints was performed by PCR and sequencing of the deletion junction using the following primers: (forward) 5′ -AGC TGA TGT CCG TCG AGT CT-3′ and (reverse) 5′ -AGC AGC AGG AAG GTG GTG-3′ . Results

In our cohort of 96 patients, 72 patients from 57 families presented with a phenotype compatible with EOWS, three from one three-generation family presented with LFSNHL and 21 patients from 17 families presented with uncommon phenotypes which do not fulfill the minimum ascertainment criteria of common WS or

3

Chaussenot et al. LFSNHL (Table 1). Among the 21 patients with unusual presentations, (i) 15 had a probable autosomal recessive transmission including 12 patients with LOWS, 2 with an OA ‘plus’ phenotype and one with a complex phenotype including DM and deafness at 1 year of age and, (ii) six individuals had a probable autosomal dominant transmission. Seventy-eight patients from 61 families carried two variants (81.2%), and 18 patients from 14 families only one variant (18.8%) identified by sequencing. Among the 37 novel patients, 15 carried 15 novel deleterious putative mutations, including 1 large deletion. Their characteristics are listed in Table S2. Among these novel mutations, the pathogenicity of three missense heterozygous variations was difficult to ascertain because we did not find a second abnormality in WFS1. Novel point mutations in patients with EOWS

Using sequencing, we identified 11 novel pathogenic putative mutations in 11 patients with a phenotype suggestive of EOWS (classical or not), including two frameshift and nine missense variants (Table 2 and Table S2). Except the p.Phe247Leufs*40 mutation, all were located in exon 8. Frameshift mutations, p.Phe247Leufs*40 and p.Met781Ilefs*81, are predicted to result in premature translation termination and are therefore interpreted to be deleterious mutations. Each of the described missense variant was defined as a pathogenic mutation based upon the criteria described in the patients and methods section (Table S2). Seven patients from six families had classical EOWS (WS1, WS2, WS3, WS4a, WS4b, WS5 and WS6) (Table 2). The novel p.Ala559Asp and p.Ala806Pro mutations were found with previously described p.Phe354del and p.Trp613* mutations respectively (WS1, WS2). The novel p.Leu432Gln mutation was found in trans position with another novel mutation p.Cys742Gly in patient WS3. Both mutations involve highly conserved amino acids and are predicted to be probably damaging by in silico analysis. Two siblings (WS4a and WS4b) were compound heterozygous for a novel variant (p.Thr641Lys) and a previously described mutation (p.Val503Serfs*15). Although the in silico predictions were not in favor of the pathogenicity of the p.Thr641Lys variant, the threonine residue at position 641 is highly conserved and the familial segregation confirmed the trans position of the two mutations. Four patients with a phenotype compatible with EOWS had incomplete clinical presentation without OA (WS7, WS8 and WS9) and/or unusual early onset of clinical signs (WS9 and WS10) (Table 2). The novel p.Trp639Gly mutation was found homozygous in a 13 year old patient with DM and deafness (WS7), born from healthy consanguineous parents. Although in silico analysis does not predict a damaging effect, the tryptophan residue is located in a transmembrane domain and is highly conserved. As the parents were not available, we performed quantitative PCR to eliminate a large deletion on the second allele. Three novel heterozygous variants, p.Phe439Cys, p.Glu809Lys and p.Ala569Val

4

were identified in patients WS8, WS9 and WS10. WS8 and WS9 developed DM and deafness before 15 years of age. Patient WS9 developed clinical signs at a very early stage. Patient WS10, a 17 year old male, also had an atypical early-onset OA (1 year old) associated with DM and hearing impairment occurring many years later. The three variants have been reported in EVS or dbSNP databases but at a very low frequency (G/?

p.Leu432Val/?

p.Trp867*/? p.Gln520*/?

p.Ala569Val/?

c.1706C>T/?

c.2601G>A/? c.1558C>T/?

p.Trp639Gly homozygous p.Phe439Cys/? p.Glu809Lys/?

p.Ala559Asp/p.Phe354del p.Trp613*/p.Ala806Pro p.Leu432Gln/ p.Cys742Gly p.Val503Serfs*15/ p.Thr641Lys p.Val503Serfs*15/ p.Thr641Lys p.Phe247Leufs*40/ p.Pro504Leu p.Met781Ilefs*81/delc

Amino acid change

c.1915 T>G homozygous c.1316 T>G/? c.2425G>A/?

c.1676C>A/c.1060_1062delTTC c.1839G>A/c.2416G>C c.1295T>A/ c.2224 T>G c.1507_1519del13nt/ c.1922C>A c.1507_1519del13nt/c.1922C>A c.741delT/c.1511C>T c.2343delG/g.6,291,340_6,308, 784del

Nucleotide change

Mutations

II/ ?

II/ ? II/ ?

II/ ?

II/II = A3 II/ ? II/ ?

II/III = B III/II = B II/II = A3 III/II = B III/II = B I/II = A2 III/I = B

Mother

Mother Father

Not done

Not done Not done Not done

Done Done Done Done Done Not done Done

Familial Genotypeb segregation

AD, age at diagnosis; DM, diabetes mellitus; DI, diabetes insipidus; EOWS, early-onset Wolfram syndrome; F, female; GH, growth hormone; HI, hearing impairment; IUGR, intrauterine growth retardation; LOWS, late-onset Wolfram syndrome; M, male; OA, optic atrophy; +, present; −absent; ?, unknown; †, deceased. a Novel mutations are shown in bold; parents of all patients were healthy; done: segregation compatible with in trans mutations; mother: mutation inherited from the mother; father: mutation inherited from the father. b According to de Heredia et al. (4). c Large deletion involving exons 5–8 of WFS1.

M

F M

6 6

13 24

WS11b WS12b, †

10

1

M

17

WS10

16 6 9 + 23 12 6

− − −

8 5 16 14 12 8 3

OA AD (years)

Patients with EOWS compatible phenotype WS7 13 M + WS8 13 M 11 WS9 16 months F 6 months

Patients with classical EOWS WS1 28 F WS2 6 F WS3 17 F WS4a 24 M WS4b 28 F WS5 16 F WS6 8 M

Age

DM AD (years)

Table 2. Clinical and molecular data of patients with possible autosomal recessive transmission and classical clinical presentationsa

WFS1 mutation update and uncommon phenotype

5

Chaussenot et al.

Fig. 1. (a) Oligonucleotide array-CGH analysis, using a 1×1M Agilent array, identified a heterozygous loss of 11.046 bp (minimum deletion size), located on chromosome region 4p16.1, between positions 6,297,228 and 6,308,273 (hg19). Left panel shows a view of the chromosome 4, and right panel shows a magnified view of the deleted region. Each small cross represents an oligonucleotide probe. The red box indicates the minimal deleted region and the red rectangle the deletion. The arrow and the dashed lines indicate the WFS1 gene. (b) PCR performed using primers flanking the deleted probes. The forward primer is in intron 4 of WFS1 and the reverse primer is downstream WFS1. Mw: molecular weight marker; P: patient; F: father; M: mother; bp: base pair. (c) Sequence analysis showing breakpoint junction of the deletion between 6,291,340 and 6,308,784.

WFS1 (13,655 bp) and 3789 bp downstream (Fig. 1). The deletion was inherited from the father and both parents presented no clinical sign. Unusual presentations with possible autosomal recessive transmission

Fifteen patients out of 96 had an unusual presentation with either a late-onset of the disease or a partial phenotype. Twelve patients from 11 families presented with LOWS, defined by the occurrence of DM and OA at 15 years or more, and at least one mutation in the WFS1 gene. Clinical and molecular data are detailed in Table 3. Except three patients (WS14, WS15 and WS16), all developed the first symptoms after 20 years of age. Three homozygous mutations, two of which involving the same amino-acid, were recurrent in LOWS: p.Arg558His was found twice in our cohort and the two others (p.Arg558Cys and p.Arg818Cys) were previously reported (14, 15). Among LOWS, we found two novel mutations in two unrelated patients (WS16 and WS22) (Table 3 and Table S2). Patient WS16, who presented with DM and OA at 15 and 30 years of age, respectively, harbored the previously described mutation p.Pro724Leu in compound heterozygosity with a novel frameshift mutation p.Gly437Alafs*5. Patient WS22, who presented with DM and OA at 27 and 30 years of age respectively, harbored the previously described mutation p.Arg629Trp in compound heterozygosity with a novel

6

missense p.Pro504Arg mutation. This substitution alters the same amino acid as the p.Pro504Leu mutation, previously reported in WS patients, and the in silico predictions are in favor of its pathogenicity. Sequencing identified only one mutation in two patients with LOWS (WS23 and WS24). Both presented with urological and neurological symptoms, hearing impairment being found in WS23 only. Quantitative PCR failed to identify a second mutation in both cases. In a previous study (12), we reported a patient with an OA ‘plus’ phenotype without DM. This 46 year old man (WS25a) presented OA at 10 years of age and cerebellar ataxia associated with neurogenic bladder at 27 years (Table 3). We add the observation of his sister (WS25b) who presented a childhood-onset OA associated with adult-onset urological and balance disorders. Both patients carried the homozygous p.Gly702Ser mutation, which has never been described in a homozygous state, but was reported combined to another missense mutation (p.Leu543Arg) in patients with EOWS (16). Finally, we report a 20 year old patient (WS26) who developed DM and deafness at 1 year of age, associated with glaucoma, bilateral cataract, cerebellar ataxia, areflexia, short stature, hypothyroidism and hypogonadism (Table 3). This patient was heterozygous for a single mutation, p.His860Asp, previously described in classical EOWS (17). No rearrangement was found on the second allele by quantitative PCR.

56 35 48

46

WS21 WS22 WS23c

WS24c,†

−

1

F

M

−

30

−

10

10

20

55 30 40

43 25 35 32

17 25 30 25

OA AD (years)

−

−

−

−

− − −

− − 38 −

17 − − −

DI AD (years)

1

−

−

−

− − 36

− 47 − 40

− − − −

HI AD (years)

Ataxia, areflexia, glaucoma, cataract, short stature, hypothyroidism, hypogonadism

Neurogenic bladder, ND (30 years)

Neurogenic bladder, ND (27 years)

− ND (47 years) UD (33 years) Neurogenic bladder, ND (40 years) ataxia, cognitive disorders UD, Psy Psy Neurogenic bladder, cortical atrophy, pyramidal signs, Psy Neurogenic bladder (35 years), cerebellar ataxia (35 years) , dysphagia, dementia, PN, AN, brain MRI abnormalities

UD (23 years), ND (23 years), hypogonadism Hypothyroidy − ND (20 years)

Other symptoms

p.Gly702Ser homozygous p.His860Asp/?

c.2578C>G/?

p.Gly702Ser homozygous

p.Phe350Val/?

p.Arg558Cys homozygous p.Pro504Arg/p.Lys629Trp p.Arg457Ser/?

p.Arg818Cys homozygous p.Arg818Cys homozygous p.Arg558His homozygous p.Arg558His homozygous

p.Glu776Val/p.Leu723Pro p.Tyr110Asn/p.Ala133Thr p.Gly437Alafs*5/p.Pro724leu p.Gly736Ser homozygous

Amino acid change

c.2104G>A homozygous

c.2104G>A homozygous

c.1048T>G/?

c.1672G>A homozygous c.1511C>G/c.1885C>T c.1371G>T/?

c.2452G>A homozygous c.2452G>A homozygous c.1673G>A homozygous c.1673G>A homozygous

c.2327A>T/c.2168T>C c.328T>A/c.397G>A c.1308delC/c.2171C>T c.2206G>A homozygous

Nucleotide change

II/?

II/II = A3

II/II = A3

II/?

II/II = A3 II/II = A3 II/?

Not done

Not done consanguinity

Father

Done Not done Not done

Not done Done

II/II = A3 II/II = A3

II/II = A3

Done Not done Not done Not done consanguinity Not done

II/II = A3 II/II=A3 III/II = B II/II = A3

Familial Genotypeb segregation

AN, autonomic neuropathy; AD, age at diagnosis; DM, diabetes mellitus; DI, diabetes insipidus; Endoc, endocrine disorder; F, female; HI, hearing impairment; LOWS, late-onset Wolfram syndrome; M, male; MRI, magnetic resonance imaging; ND, neurological disorder; OA, optic atrophy; PN, peripheral neuropathy; Psy, psychiatric disorder; UD, urological disorder; +, present; −, absent; ?, unknown; †, deceased. a Novel mutations are shown in bold; parents of all patients were healthy; done: segregation compatible with in trans mutations; father: mutation inherited from the father. b According to de Heredia et al. (4). c Patients previously described (10, 12); patients with first signs between 15 and 20 years of age are shown in gray.

WS25bc 40 Other patient WS26 20

Patients with OA ‘plus’ WS25ac 46 M

F

F F M

33 27 46

38 32 29 32

F F M M

44 47 38 45

WS18ac WS18bc WS19c WS20c

DM AD (years)

16 15 15 20

Sex

Patients with LOWS WS14c 32 M W15c 25 F WS16 42 M W17c 31 F

Age

Mutations

Table 3. Clinical and molecular data of patients with possible autosomal recessive transmission and unusual clinical presentationsa

WFS1 mutation update and uncommon phenotype

7

Chaussenot et al. Unusual presentations with possible autosomal dominant transmission

We found six patients from four families (WS27-WS30) who carried one WFS1 mutation with a probably autosomal dominant effect. Clinical and molecular data are detailed in Table 4. WS27 and WS28 were sporadic cases with childhood-onset OA and deafness. Both patients harbored a heterozygous mutation, p.Ala684Val, previously reported by Rendtorff et al. in patients with the same phenotype and an autosomal dominant family history (18). The absence of this mutation in asymptomatic parents of WS28 revealed its de novo occurrence. Parents of WS27 were asymptomatic but were not available for analysis. As this mutation has also been reported in compound heterozygosity in EOWS, we performed quantitative PCR analysis in both patients without finding any large rearrangement on the second allele. In the WS29 family previously reported by Valero et al., (11) patients presented with autosomal dominant juvenile-onset deafness predominant on high frequencies and adult-onset DM. The mother (WS29a), who presented with late-diagnosed OA, and his son (WS29b) carried the p.Glu864Lys mutation. This mutation was reported by Eiberg et al. (6) in one family with autosomal dominant deafness and OA associated with either impaired glucose tolerance or DM, and by Fukuoka et al. (19) in two families with LFSNHL. Last, we identified a novel missense mutation (p.Thr321Met) in WS30 family (Table 4 and Table S2). The mother (WS30a) and her daughter (WS30b) presented with childhood-onset OA and high frequency hearing impairment at 42 and 15 year old, respectively. Familial segregation of the variant, high conservation of the amino acid and in silico analysis were in favor of a deleterious effect. Discussion

WS is a rare disease with an estimated prevalence of 1 in 770,000 in the UK (1). This explains the difficulty of studying cohorts that are large enough to determine WS natural history and to elucidate the role of WFS1 mutations. That is why initiatives like that from the European EURO-WABB register (http://www.euro-wabb.org) are fundamental to record high-quality phenotypic and genotypic information in order to address unsolved questions linked to the wide heterogeneity of WFS1-related disorders. Recently, de Heredia et al. (4) analyzed clinical and genetic data of 412 patients with WS published since 1998, date of identification of the WFS1 gene. This review included 59 patients from the French cohort that we update, here, by adding 37 novel patients who carried at least one WFS1 mutation. Uncommon phenotypes associated with WFS1 mutations

From our French cohort of 96 patients, 72 among the 87 with a probable autosomal recessive disease presented an EOWS with DM or OA beginning before the age of 15 years (82.75%). LOWS was observed in 12 patients out of 87 (13.8%) with DM and OA starting at 15 years

8

or more. LOWS has been rarely reported in the literature (Table S3) and its frequency is probably underestimated. Among the 392 patients with age specified for clinical symptoms analyzed in the large review of de Heredia et al., (4) 5% had LOWS (20/392) with almost half of these patients being French (9/20). LOWS was indeed very rare in cohorts of different geographical origin. In the British series of 45 patients with WS, only one patient started DM and OA after 15 years of age (2.2%) (20, 21). In the Brazilian cohort, late-onset was found in one case out of 27 (3.7%) with DM at 27 years and OA at 19 (22). Only one study had a similar rate to that of the French cohort with 2 LOWS out of 20 Spanish cases (10%) (14). Regarding mutations, LOWS was almost exclusively associated with missense variations. Only two stop mutations were previously reported in association with missense mutations: p.Glu273* with DM and OA both starting at 16 years (21) and p.Trp613* with DM and OA beginning at 28 and 42 years, respectively (17). In our series of 12 patients with LOWS, we found one frameshift mutation only (p.Gly437Alafs*5) in association with the p.Pro724Leu mutation in a 30 year old man who started DM at 15 years and OA at 30. All other were missense mutations and some of which were recurrent. The p.Arg818Cys mutation was homozygous in two siblings (WS18a-b) with a disease onset after 25 years. This homozygous mutation had been already reported in three patients from the same family who presented with LOWS (14). It leads to absence of wolframin expression by western blot analysis in brain of affected individuals (4). Another homozygous mutation, p.Arg558His, was found in two unrelated patients with LOWS in our series (WS19 and WS20). This mutation had been previously described associated with the non-sense p.Glu864* mutation in a classical EOWS (17). The arginine residue at position 558 was also affected by the p.Arg558Cys homozygous mutation (WS21). The homozygous p.Arg558Cys mutation had been previously reported in one sporadic LOWS case with DM at 33 years and OA at 53 (15). In these three sporadic cases (WS19-21), disease started around 30 by DM. At this stage, it is not possible to know whether the discrepancy about LOWS frequency between the French cohort and others is linked either to French population specificity or to a country dependent mode of WFS1 screening in atypical phenotypes. An uncommon phenotype was observed in WS25 family with two siblings who are, to date, the only reported cases of recessive OA without DM. Both carried the homozygous p.Gly702Ser mutation. Hereditary optic neuropathies can be associated with any mode of inheritance. Leber’s optic neuropathy and autosomal dominant OA are most frequently sought while autosomal recessive OA are rare and mostly observed in association with multisystemic diseases (23). The involvement of WFS1 in isolated autosomal recessive OA is probably underestimated. When patients present juvenile-onset OA, WFS1 should be analyzed even in absence of DM and deafness when other diagnoses have been excluded. Lastly, 6 patients out of 96 presented atypical phenotypes associated with dominant WFS1 mutations. Clinical presentations were heterogeneous but all

WFS1 mutation update and uncommon phenotype Table 4. Clinical and molecular data of patients with possible autosomal dominant transmissiona Mutation OA Age Sex AD (years)

HI AD (years)

WS27

16

M

15 Moderate

WS28

4

F

WS29ab 85

F

WS29bb 60

M

4 Mild optic disk pallor 60 Mild −

2 Severe with minor anterior labyrinth abnormality 20 months Severe

WS30a

44

F

5 Moderate

42 Moderate (hf)

WS30b

14

F

2 Severe

14 Mild

Childhood (hf) Childhood (hf)

Other symptoms

Family history

Nucleotide change

Amino acid change

−

Sporadic

c.2051C>T

p.Ala684Val

−

de novo

c.2051C>T

p.Ala684Val

Diabetes mellitus (45 years) Diabetes mellitus (45 years) Axonal neuropathy (17 years), ataxia, cerebellar atrophy, neurogenic bladder (40 years), DI (44 years) Learning disabilities, cerebellar atrophy, areflexia

Autosomal dominant Autosomal dominant Autosomal dominant

c.2590 G>A

p.Glu864Lys

c.2590 G>A

p.Glu864Lys

Autosomal dominant

c.962 C>T p.Thr321Met

c.962 C>T p.Thr321Met

AD, age at diagnosis; HI, hearing impairment; hf, high frequency; OA, optic atrophy; −, absent. a Novel mutations are shown in bold. b Patients previously described (11).

patients presented with hearing impairment and five out of six had OA. In all cases, we identified missense mutations among which two had been previously reported in similar phenotypes (OA and deafness), suggesting the existence of recurrent mutations like in LOWS. In WS28, familial study allowed us to prove the de novo occurrence of the p.Ala684Val mutation. Two patients (WS27 and WS28) out of six were sporadic cases and it is likely that WFS1 mutations are underestimated in early-onset deafness with OA especially since the phenotype, described by Eiberg and Rentdorff, can be heterogeneous with absence of OA and presence of DM, like in patient W29b (6, 18). Identification of 15 novel mutations including one large deletion

Among the 37 novel patients, we identified 15 novel mutations. According to the classification proposed by de Heredia et al., (4) we assigned the majority of mutations to type II (69%), types III and I corresponding respectively to 19 and 12%. This result is similar to the one obtained in the review reported by these authors. Among the 10 patients carrying 2 mutated alleles with at least 1 novel mutation, 1 patient was classified as class A2, 3 as class A3 and 6 as class B while de Heredia et al. reported a majority of patients in class A (4). The identification of both WFS1 mutated alleles is not always possible. de Heredia et al. reported only one mutation in 8.31% of cases (28 of 337 patients with described phenotype) (4). This percentage could be explained by

the impossibility to find a second mutation by routine analysis, the existence of a dominant mutation or mutations in another gene. In the French cohort, we identified by sequencing only one mutated allele in 18 patients out of 96 (14 families out of 75). When we consider families with a probable autosomal recessive transmission, 9 out of 70 (12.8%) carried one mutation only. On the hypothesis that the second mutation could be a large deletion on the second allele, we performed quantitative PCR and found a large deletion in one patient only (WS6). This 6 year old boy presented with early onset of both DM and OA (3 and 6 years, respectively). The rearrangement (17,444 bp deletion in size) deleted the 3′ end of WFS1 starting at exon 5. The second mutation was an out of frame deletion (c.2343delG; p.Met781Ilefs*81). Only one large deletion, encompassing most of exon 8, was previously reported by Inoue et al. in siblings with a severe WS phenotype leading to early death in siblings (16 and 21 years) (4, 17). The second complex structural rearrangement previously described in WFS1 concerned a complex phenotype with a neonatal-onset including diabetes insipidus, optic pathway hypoplasia and psychomotor retardation, without DM (9). All together, these results suggest that sequencing does not detect the second mutation in approximately 10% of cases and that large deletions, mainly associated with severe phenotypes, are probably not the majority among intronic or promoter mutations in patients with ‘classical’ WS. Locus heterogeneity involving other functionally related genes could also be suspected but we did not identify any mutation in CISD2.

9

Chaussenot et al. In conclusion, this work highlights the broad spectrum of WFS1-related disorders including late-onset disease, recessive OA without DM and dominant presentations with deafness and OA. It aims helping clinicians to identify situations that require WFS1 testing outside the traditional WS presentation. We also show that in recessive forms, large deletions are probably rare but have to be searched in patients with classical WS carrying only one WFS1 mutation after sequencing analysis. Supporting Information Additional supporting information may be found in the online version of this article at the publisher’s web-site.

Acknowledgements The authors thank Gaëlle Augé, Sabine Mutz and Bernadette Chafino for technical help.

References 1. Barrett TG, Bundey SE, Macleod AF. Neurodegeneration and diabetes: UK nationwide study of Wolfram (DIDMOAD) syndrome. Lancet 1995: 346: 1458–1463. 2. Khanim F, Kirk J, Latif F, Barrett TG. WFS1/wolframin mutations, Wolfram syndrome, and associated diseases. Hum Mutat 2001: 17: 357–367. 3. Inoue H, Tanizawa Y, Wasson J et al. A gene encoding a transmembrane protein is mutated in patients with diabetes mellitus and optic atrophy (Wolfram syndrome). Nat Genet 1998: 20: 143–148. 4. de Heredia ML, Clèries R, Nunes V. Genotypic classification of patients with Wolfram syndrome: insights into the natural history of the disease and correlation with phenotype. Genet Med 2013: 15 (7): 497–506. 5. Tranebjærg L, Barrett T, Rendtorff ND. WFS1-related disorders. In: Pagon RA, Adam MP, Ardinger HH, Bird TD, Dolan CR, Fong CT, Smith RJH, Stephens K, eds. GeneReviews. Seattle, WA: University of Washington, 2009: 1993–2014. 6. Eiberg H, Hansen L, Kjer B et al. Autosomal dominant optic atrophy associated with hearing impairment and glucose regulation caused by a missense mutation in the WFS1 gene. J Med Genet 2006: 43: 435–440. 7. Berry V, Gregory-Evans C, Emmett W et al. Wolfram gene (WFS1) mutation causes autosomal dominant congenital nuclear cataract in humans. Eur J Hum Genet 2013: 21 (12): 1356–1360.

10

8. Bonnycastle LL, Chines PS, Hara T et al. Autosomal dominant diabetes arising from a Wolfram Syndrome 1 mutation. Diabetes 2013: 62 (11): 3943–3950. 9. Elli FM, Ghirardello S, Giavoli C et al. A new structural rearrangement associated to Wolfram syndrome in a child with a partial phenotype. Gene 2012: 509 (1): 168–172. 10. Giuliano F, Bannwarth S, Monnot S et al. Wolfram syndrome in French population: characterization of novel mutations and polymorphisms in the WFS1 gene. Hum Mutat 2005: 25: 99–100. 11. Valero R, Bannwarth S, Roman S et al. Autosomal dominant transmission of diabetes and congenital hearing impairment secondary to a missense mutation in the WFS1 gene. Diabet Med 2008: 25: 657–661. 12. Chaussenot A, Bannwarth S, Rouzier C et al. Neurologic features and genotype-phenotype correlation in Wolfram syndrome. Ann Neurol 2011: 69 (3): 501–508. 13. Amr S, Heisey C, Zhang M et al. A homozygous mutation in a novel zinc-finger protein, ERIS, is responsible for Wolfram syndrome 2. Am J Hum Genet 2007: 81: 673–683. 14. Gómez-Zaera M, Strom TM, Rodríguez B, Estivill X, Meitinger T, Nunes V. Presence of a major WFS1 mutation in Spanish Wolfram syndrome pedigrees. Mol Genet Metab 2001: 72: 72–81. 15. Lieber DS, Vafai SB, Horton LC et al. Atypical case of Wolfram syndrome revealed through targeted exome sequencing in a patient with suspected mitochondrial disease. BMC Med Genet 2012: 13: 3. 16. Aloi C, Salina A, Pasquali L et al. Wolfram syndrome: new mutations, different phenotype. PLoS One 2012: 7 (1): e29150. 17. Smith CJ, Crock PA, King BR, Meldrum CJ, Scott RJ. Phenotype-genotype correlations in a series of Wolfram syndrome families. Diabetes Care 2004: 27: 2003–2009. 18. Rendtorff ND, Lodahl M, Boulahbel H et al. Identification of p.A684V missense mutation in the WFS1 gene as a frequent cause of autosomal dominant optic atrophy and hearing impairment. Am J Med Genet A 2011: 155A (6): 1298–1313. 19. Fukuoka H, Kanda Y, Ohta S, Usami S. Mutations in the WFS1 gene are a frequent cause of autosomal dominant nonsyndromic low-frequency hearing loss in Japanese. J Hum Genet 2007: 52: 510–515. 20. Barrett TG, Bundey SE. Wolfram (DIDMOAD) syndrome. J Med Genet 1997: 34: 838–841. 21. Hardy C, Khanim F, Torres R et al. Clinical and molecular genetic analysis of 19 Wolfram syndrome kindreds demonstrating a wide spectrum of mutations in WFS1. Am J Hum Genet 1999: 65: 1279–1290. 22. Gasparin MR, Crispim F, Paula SL et al. Identification of novel mutations of the WFS1 gene in Brazilian patients with Wolfram syndrome. Eur J Endocrinol 2009: 160 (2): 309–316. 23. Meyer E, Michaelides M, Tee LJ et al. Nonsense mutation in TMEM126A causing autosomal recessive optic atrophy and auditory neuropathy. Mol Vis 2010: 16: 650–664.