Editor—Townes-Brocks syndrome (TBS, MIM 107480) was first described by Townes and Brocks1 in 1972 as an association of imperforate anus, supernumerary thumbs, malformed ears, preauricular tags, and sensorineural hearing loss. Several additional familial as well as isolated cases have been reported.2 TBS is caused by mutations of the putative zinc finger transcription factor geneSALL1.3 AllSALL1 mutations identified to date in TBS patients are located 5′ of the first double zinc finger encoding region.4 Three of these are nonsense mutations at two different positions. The mutation 826C>T was found in three unrelated sporadic cases, and at position 1115 one patient carried an adenine (1115C>A) and another a guanine (1115C>G) instead of a cytosine. All seven other reported mutations are short frameshift deletions of 1, 2, 7, or 10 base pairs.4

SALL1 encodes four double zinc finger domains which are characteristically distributed over the entire protein.5 All known mutations have been predicted, if the mutated transcripts are indeed translated, to result in prematurely truncated proteins lacking all double zinc finger domains presumed to be essential for SALL1 gene function. Since no mutations were found 3′ of the most 5′ located double zinc finger encoding region, it was assumed that only those mutations which remove all double zinc fingers could cause TBS,4 whereas mutations located further 3′ in SALL1 could result in a different phenotype or no abnormal phenotype at all. Here we describe three novel mutations in three independent families illustrating that truncating mutations positioned further 3′ of the previously described hotspot region in SALL1also result in TBS.

All subjects available for investigations were examined forSALL1 mutations after giving informed consent, and all those available for investigations were clinically examined by a clinical geneticist. In all affected subjects chromosomal analysis before DNA studies had shown a normal karyotype.

In the first family (family 1), TBS occurred in a sporadic case. The male patient (patient 1 of this study), aged 44, came to the clinic because of sudden loss of visual acuity owing to optic neuropathy. Townes-Brocks syndrome was diagnosed because he showed bilateral dysplastic ears, bilateral triphalangeal thumbs, and congenital sensorineural deafness. As a child, he had undergone surgery for imperforate anus, and his feet were previously operated on because of toe malposition. X rays showed bilateral partial proximal synostosis of metatarsals IV-V as well as fusion of some tarsal bones (not shown). Unusual features of this patient include acute bilateral optic neuropathy as well as bilateral inguinal hernias and an umbilical hernia. The family history was negative for TBS. His parents were not available for investigation.

In the second family (family 2), TBS was diagnosed in the female index patient (patient 2 of this study) because of a bifid thumb on the left and preaxial polydactyly on the right side, bilateral, small, dysplastic ears, bilateral moderate to severe conductive hearing loss, bilateral renal hypoplasia, acquired hypothyroidism, anteriorly placed and stenotic anus, and hypoplastic third toes with fifth toe clinodactyly. Recently, her mother gave birth to a boy who is also affected. He shows a bifid thumb on the right and complete preaxial polydactyly on the left side. He has bilateral cup shaped microtia with preauricular dimples and is reported not to respond to loud noises. The suspected hearing loss has not yet been fully evaluated. The baby boy has no anal malformation but a prominent midline perineal raphe. Congenital hypothyroidism was also diagnosed. The kidneys were normal on ultrasound. No other abnormalities are reported in these children. Both parents are clinically normal except for an anteriorly placed anus seen in the mother. The mother's family history is unremarkable. Her husband has a great aunt with unilateral preaxial polydactyly whose grandson also has preaxial polydactyly. The aunt's niece had a child with hemifacial microsomia. The parents of the index patient are not consanguineous.

In the third family, at least five persons in three generations are affected (fig 1A). The male index patient (III.2, patient 3) had imperforate anus with a prominent perineal median raphe, bilateral cutaneous syndactyly of the third and fourth toes, bilateral hypoplastic second toes, lop ears, frontal bossing, retrognathia, glandular hypospadias, and bilateral hypoplastic and dysplastic kidneys. Echocardiography showed a secundum type ASD. A mixed hearing loss was diagnosed at 1 year of age, assessed at the age of 2 as 50-60 dB (over the whole tone range). During childhood the boy showed considerable growth retardation (length and weight <3rd centile), probably secondary to chronic renal insufficiency. At the age of 6 haemodialysis was begun because of severe renal insufficiency. Intelligence is normal.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Pedigree and mutation analysis of family 3. (A) Pedigree. Horizontal bars indicate the family members personally examined. The phenotype of II.3 is not known. (B-D) Electrophoretograms of the detected mutations and control. (B) Heterozygous mutation IVS2-19T>A (arrow pointing downwards) detected in all examined affected family members (I.1, II.2, III.2) but not in the unaffected members. The arrow pointing upwards indicates the boundary between intron 2 and exon 3. (C) DNA sequence of the subcloned RT PCR product of the wild type allele spanning from exon 2 to exon 3 (the arrow indicates the boundary). (D) DNA sequence of the subcloned RT PCR product of the mutant allele shows an insertion of 17 bp derived from intron 2 (underlined in top and bottom pictures) between exon 2 and exon 3 derived sequences.

His grandmother (I.1, patient 4) had late onset hearing loss first noticed at the age of 45. At 70 years of age, she had a hearing loss of 60-65 dB. Clinical examination showed lop ears with downfolded scapha helix and hypoplastic anthelix and insufficient rotation of both thumbs on opposition. Anamnestically, there were no anal abnormalities. She had never had any kidney complaints, but her kidneys were not specifically checked by ultrasound or biochemically. Her daughter (II.2, patient 5) showed sensorineural hearing loss (audiometry aged 37: mild perceptive hearing loss, 35 dB over the whole tone range). Her ears were operated on during adolescence (“bat ears”). Her anus was normal on physical examination, and her kidneys were normal on ultrasound. No further abnormalities were seen in I.1 or II.2. No clinical data are available on the brother of II.2 (II.3). He was reported by his spouse to have “strange ears” but he declined to be examined. However, he has two affected children. His son (III.4) had imperforate anus, “bat ears” with downfolded and hypoplastic margin of the scapha helix, perceptive hearing loss of 20 dB on audiometry at the age of 6, and showed insufficient rotation of the thumbs on opposition. His sister (III.3) had imperforate anus with a prominent perineal raphe, lop ears with downfolded upper margin of the scapha helix, and insufficient rotation of the thumbs on opposition.

Mutation analysis of SALL1 was performed by PCR amplification of all exons from genomic DNA (prepared from peripheral lymphocytes) followed by direct sequencing of PCR products as described elsewhere.4 Permanent lymphoblastoid cell lines were prepared as previously described.6 Cells were harvested by centrifugation and total RNA was isolated using Total RNA Reagent (Biomol) according to the manufacturer's instructions. Two μg of total RNA from each stage was reverse transcribed using Ready-To-Go^TM You-Prime First Strand Beads (Amersham Pharmacia) and 20 pmol of primer TR5.5 (5′GGCCACCATAGGTCGCATTC3′), and 1 μl of each first strand reaction was used as a template in a subsequent PCR reaction (conditions: 95°C for four minutes initial denaturation; 35 cycles of 94°C for one minute, 64°C for one minute, 72°C for one minute; 72°C for two minutes final extension) with primers TF4.2 (5′TGGATTTGACATCTAGTCACGC3′) and TR5.8 (5′TGAACAGGAATGAATGCTATGTC3′). A nested amplification was carried out using primers TF4.3 (5′GACACCCCCACCAGTCACG3′) and TR5.7 (5′AGGTGAGCTGTTCCCACTGC3′). Conditions were: 95°C for four minutes initial denaturation; 35 cycles of 94°C for one minute, 64°C for one minute, 72°C for 45 seconds; 72°C for two minutes final extension. PCRs were performed on Primus 25 thermal cyclers (MWG). Amplification products were visualised on agarose gels, gel extracted, and DNA sequences of amplification products were verified by direct sequencing using primers TF4.3 and TR5.7. PCR products were also subcloned in pGEM-Teasy (Promega) and at least four independent clones were sequenced using vector specific primers.

SALL1 mutation analysis in family 1 (patient 1) showed a heterozygous mutation 840delC which is located 5′ of the region encoding the first double zinc finger (fig 2A, fig 3). In family 2, we found a heterozygous nonsense mutation 1509C>A (Y503X) in the affected girl (patient 2, fig 2C, fig 3). This mutation was also detected in her affected newborn brother. However, neither parent showed the mutation in their peripheral lymphocytes. Paternity was confirmed in this family (data not shown). In family 3, molecular analysis was performed on I.1 (patient 4, fig 1A), II.1, II.2 (patient 5, fig 1A), III.1, and III.2 (patient 3, fig 1A). No mutation was found in the whole SALL1 coding region. However, within the intron 2 sequence, a heterozygous transition IVS2-19T>A (fig 1B) was found in III.2, and subsequently in I.1 and II.2, but not in the unaffected family members. This mutation was predicted to create an aberrant splice acceptor site. By comparing the surrounding sequences of the aberrant and the normal intron 2 splice acceptor site to consensus sequences,7 it was estimated that the aberrant splice site was as likely to be functional as the normal site. The mutation was excluded in 200 control alleles.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

(A, C) Electrophoretograms (direct sequencing) of the heterozygous mutations 840delC (A) detected in family 1 and 1509C>A detected in family 2 (C). Corresponding wild type sequences are shown below (B, D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Position of the mutations reported here. 840delC is located within the region where all other mutations reported so far are clustered (small arrows). Note that 1509C>A and IVS2-19T>A are located 3′ of this region. IVS2-19T>A leads to an insertion within the region encoding double zinc finger 4. Zinc finger motifs are indicated as oval symbols.

In order to test if the mutation indeed created a functional splice site, lymphoblastoid cell lines of patients II.2 and III.2 were examined by RT-PCR. In both patients, direct sequencing of RT-PCR products showed an identical pattern of two different overlapping sequences indicating that both the mutated and the normal transcript were present in similar amounts. Sequencing of the subcloned RT-PCR fragments showed the wild type allele (fig 1C) and a mutated allele (fig 1D) carrying an insertion of 17 bp derived from intron 2 sequence between the aberrant and the normal splice acceptor sites. It is placed within the coding region for the most carboxy-terminal double zinc finger between exon 2 and exon 3 sequences (fig 1B-D fig 3). The frameshift resulting from the insertion is further predicted to cause premature termination of the SALL1 protein (1208 amino acids instead of 1324 in the wild type).

In addition to the mutations reported here, the following polymorphisms were detected in SALL1: IVS1+119G>A, IVS1+118C>G, IVS1+36delAC (all intron 1), 2574T>C, 3456C>T (both exon 2), IVS2-31delCT (intron 2), 3872A>G (N1291S), and 3915C>T (exon 3). The intronic polymorphisms occurred in more than 10% of all subjects (affected and unaffected) analysed forSALL1 mutations by our group. The exonic mutation N1291S is thought to be silent since it was found in two unaffected persons. The other exonic variations do not affect translation and are not segregating with the phenotype. Since the exonic sequence variations have only rarely been found, it is as yet unclear if they represent true polymorphisms or rare sequence variants.

All SALL1 mutations previously reported in TBS reside in exon 2, 5′ of the coding region for the first double zinc finger domain (fig 3), and are predicted to result inSALL1 haploinsufficiency.4Nonsense mutations seem to occur less frequently than small deletions, and both known nonsense mutations were found independently in two and three families indicating the existence of two hotspots at nucleotides 826 (mutated in three families) and 1115 (mutated in two families). In contrast, SALL1 small deletions as a group occur more often but they seem to represent private mutations only.4

840delC is yet another short deletion located 5′ of the coding region for the first double zinc finger domain (fig 3). The phenotype of the patient carrying this mutation is typical for TBS, that is, he shows anal, ear, and thumb malformations. Interestingly, this man also shows acute optic neuropathy. While this has not been reported so far in TBS, this might represent another rare feature of the syndrome.

The two other mutations described in this report are the firstSALL1 mutations located 3′ of the region where all previously known mutations cluster. The 1509C>A mutation is neighbouring the 3′ end of the double zinc finger 1 coding region (fig3). While this mutation could result in a prematurely terminated SALL1 protein lacking double zinc finger domains 2-4, it is still unclear if the mutated transcript remains stable and the corresponding protein is indeed expressed. Therefore, this mutation could well result in haploinsufficiency if the mutated transcript is readily degraded. This mutation is the third nonsense mutation detected so far inSALL1, and it is the first one to be found in a family with dominant transmission since all other nonsense mutations known so far have been detected as de novo mutations in patients with severe features of TBS.4 Interestingly, one of the parents carries the mutation in the germline but not in lymphocytes. Yet the parental origin of the mutation needs to be determined in order to explain if the anteriorly placed anus seen in the mother could also reflect the presence of the mutation in other tissues. However, it is clear that the preaxial polydactyly reported in the father's family cannot be related to the mutation which caused TBS in his children.

The most interesting mutation shown here is IVS2-19T>A (fig 1B-D). We were able to show that the predicted transcript resulting from aberrant splicing is indeed expressed. While quantitative PCR has not been performed, direct sequencing of the RT-PCR products indicates that the aberrant transcript is as abundant as the normal one. It seems therefore that splicing of the mutated primary transcript occurs preferentially if not completely at the aberrant splice site. The predicted protein encoded by the mutated transcript contains intact coding sequences for all double zinc finger domains except for the double zinc finger 4 in which the carboxy-terminal finger motif is interrupted. How this mutation might lead to the phenotype remains to be elucidated. It has previously been shown in theDrosophila transcription factorKrüppel that a missense mutation replacing one of the conserved cysteine residues within the second of five tandemly arranged finger motifs results in a null allele.8Therefore, the splice mutation reported here is likely to result in loss of biological function of the most carboxy-terminal double zinc finger domain. It remains unclear if this is sufficient to result inSALL1 haploinsufficiency causing TBS. The predicted mutant protein is 116 amino acids shorter than the wild type protein and contains different carboxy-terminal amino acids because of the frameshift. An alternative explanation for the effect of the mutation could therefore be that the changed three dimensional structure of the mutated protein results in a non-functional SALL1 protein which is unstable or not able to bind to its target sequences.

The phenotype of the severely affected family members reported here is not significantly different from other TBS cases in which classical truncating mutations were found. Therefore, we assume that all mutations shown in this report will lead to haploinsufficiency forSALL1, as suggested to be the common result of all mutations previously reported.4