A novel germ line mutation in SOX9 causes familial campomelic dysplasia and sex reversal
A novel germ line mutation in SOX9 causes familial campomelic dysplasia and sex reversalFergus J. Cameron, Robyn M. Hageman, Claire Cooke-Yarborough1, Cheni Kwok2, Linda L. Goodwin3, David O. Sillence3 and Andrew H. Sinclair*
Department of Paediatrics and Centre for Hormone Research, and 1Department of Anatomical Pathology, University of Melbourne, Royal Children's Hospital, Melbourne, VIC 3052, Australia, 2Department of Genetics, University of Cambridge, Cambridge CB2 3EH, UK and 3Department of Clinical Genetics, The New Children's Hospital, Sydney, NSW 2124, Australia
Received May 10, 1996;Revised and Accepted July 8, 1996
Mutations in the gene SOX9 result in the syndrome of campomelic dysplasia (CD) which includes sex-reversal in 75% of 46,XY affected individuals. These mutations only affect a single allele of SOX9 suggesting a dominant mode of inheritance for this syndrome. Consequently, CD and autosomal sex reversal may result from haploinsufficiency of SOX9. The SOX9 gene maps to the long arm of human chromosome 17 and translocations in this region also result in CD. We report a family in which there were three affected patients, two of whom showed 46,XY sex-reversal. Interestingly, despite all three patients being heterozygous for a familial mutation in SOX9 (insertion of a cytosine residue at nucleotide position 1096), their gonadal phenotypes varied widely. The proband was found to have 46,XY true hermaphroditism with ambiguous genitalia. The other two sibs were 46,XY and 46,XX, and both had bilateral ovaries with normal female genitalia. The somatic cells in both parents revealed wild-type SOX9 nucleotide sequences. However, mutational analysis of the SOX9 gene in the father's germ cells revealed they were mosaic for mutant and wild-type sequences. This family is particularly informative as it demonstrates that the same SOX9 mutation can produce very different 46,XY gonadal phenotypes. The range of gonadal morphologies observed may be explained by several possible mechanisms such as variable penetrance of the mutation, increased activity of the non-mutant SOX9 allele or stochastic environmental factors. These results also demonstrate that paternal germ cell mosaicism of a mutant SOX9 sequence can result in a CD phenotype amongst his offspring.
Campomelic dysplasia (CD) is one of several syndromes that can result in XY gonadal dysgenesis. The syndrome is characterised by the radiological features of bowed femora and tibiae, hypoplastic scapulae and pelvic bones and non-mineralised thoracic pedicles. Clinically, an affected infant exhibits bowed lower limbs with pretibial skin dimpling, dolichocephaly, micrognathia, cleft palate, a flat nasal bridge, low set ears, talipes equinovarus and congenital dislocation of the hips (1 ). Patients usually succumb in the neonatal period due to respiratory insufficiency. Necropsy findings may also demonstrate cardiac, renal and CNS anomalies including absence of the olfactory bulbs. Three-quarters of karyotypic males have external genitalia which lie on a spectrum between those of an unambiguous female to that of hypospadias with a bifid scrotum (1 ). Various combinations of internal Mullerian and Wolffian duct structures have been reported. Gonadal morphology in these patients is similar to that seen in XY gonadal dysgenesis (ranging from dysplastic testicular tissue to poorly differentiated ovarian tissue with a few primordial follicles).
Analysis of patients with CD and de novo chromosome 17 translocations mapped the locus to 17q24.3-q25.1 (2 ). Two groups independently identified the SOX9 gene as responsible for CD (3 ,4 ). The distribution and variety of mutations reported in SOX9 suggest that CD is caused by haploinsufficiency of functional SOX9. Mutations in one allele of SOX9 have been caused by splice acceptor and donor changes; missense and frame-shift mutations and deletions. There has been one reported case of compound heterozygosity, with two separate mutations in the one patient (4 ), and one reported case of an unaffected parent sharing the same mutation as an affected infant (4 ). To date 13 mutations (3 -5 ) and 10 translocations (2 ,5 -10 ) have been described in 24 patients with CD. There have been no reported cases of patients having both a balanced 17q translocation and a mutation (5 ). Nine of these mutations were sex-reversing in 46,XY affected patients. One mutation was seen in two unrelated 46,XY patients with CD, it being sex-reversing in one and not in the other (5 ). Families with two normal parents and more than one CD affected sibling have been described (1 ,11 -13 ) and consequently the condition was previously thought to be transmitted in an autosomal recessive fashion. However, this assumption was never tested as none of the families were described genotypically. This paper describes the genotype, gonadal phenotype and mode of inheritance of three affected sibs with CD.
Gel shifts were seen in all three sibs (II.1; II.2; II.4) for the PCR fragment generated by the SOX9 primers int5 (GTCTGCACAGCCCTTGTTG) and Q (TCAGGTCAGCCTTGCCCGGC) which amplify a 124 bp product in exon 3 (SSCP shift data not shown). In each patient the shift was seen in genomic DNA from both the gonads and liver. No gel shifts were observed for any of the other PCR fragments.
Direct cycle-sequencing of the gel shifted PCR product in patients II.1, II.2, and II.4 revealed the insertion of a cytosine after nucleotide 1096 (Codon 246, Fig. 1 ). This causes a frame-shift mutation which results in a stop codon at position 251 and presumably a truncated SOX9 protein. All three patients were heterozygous for this mutation and the wild-type sequence.
The left gonad of Patient II.1 was situated in the pelvis adjacent to the fimbrial end of an ipsilateral fallopian tube. The right gonad was found in the right inguinal canal. Histologically, both gonads consisted of juxtaposed testicular tissue and dysplastic ovarian tissue (Fig. 2 a).
Patient II.1 possessed a unicornuate uterus that ended blindly proximal to the vagina. There was a left fallopian tube with no evidence of Wolffian duct derivatives (Fig. 2 c,d). Patients II.2 (Fig. 3 c,d) and II.4 both had normal female Mullerian duct structures with normal relationships between fallopian tubes, uterus and vagina.
Patient II.1 had a 1.5 cm phallus with a degree of labioscrotal fusion that was equivalent to Prader 1 (14 ), i.e. two separate visible orifices for urethra and vagina associated with a clitoromegaly (Fig. 2 c,d). Patients II.2 and II.4 had unambiguous female external genitalia consistent with Prader 0 development (Fig. 3 c,d).
The family described in this paper is informative for several reasons. Firstly, they possess a novel mutation in SOX9 (insertion C at position 1096 in exon 3) which causes CD. This brings the total number of different mutations causing CD to 14 (10 of these resulting in 46,XY sex-reversal, see Figure 4 ). Gonad-specific mutations in SRY have been reported in other forms of gonadal dysgenesis (15 ). To eliminate the possibility of an organ specific mutation we analysed SOX9 sequences derived from both a phenotypically affected tissue (gonad) and a tissue phenotypically unaffected by the campomelic syndrome (liver). The SOX9 mutation was present in both tissues.
Figure 4. Mutations in SOX9. Sex reversing mutations are indicated below the diagram while non-sex reversing mutations are shown above. References for the mutations are cited.
Secondly, this study is the first report of true hermaphroditism associated with 46,XY CD. A confounding issue in classifying gonadal dysgenesis histologically is that morphology may be constantly changing. Consequently, the timing of patient examination will determine the gonadal phenotype described. For example, Turners syndrome patients have morphologically normal ovarian stroma and germ cells at 12 weeks gestation but have streak gonads by the time of puberty (16 ). Similarly, we have observed (Cameron et al., unpublished data) an XY individual with partially virilised external genitalia and ovaries at birth. In this instance it suggests there may have been testicular material present at an earlier stage and implies that the `ovaries' may originally have been ovotestes. If our patient had been examined at 8-12 weeks gestation he may have been classified as a true hermaphrodite. Thus the real incidence of true hermaphroditism in this form of 46,XY gonadal dysgenesis may have been previously unrecognised. This underlies the more general concept that in 46,XY gonadal dysgenesis there is a spectrum from dysplastic testes to ovotestis and dysplastic ovaries.
Prior to the isolation and mutation analysis of SOX9, the mode of inheritance for CD was not clear. The observation of affected offspring with normal parents implied that CD may be inherited in an autosomal recessive manner (17 ). With the discovery that CD patients were heterozygous for mutations in SOX9 it was postulated that transmission was autosomal dominant (3 ,4 ). Previous studies have analysed parental genomic DNA obtained from lymphocytes; however, none has examined parental germ cells. While we were not able to obtain maternal oocytes, examination of paternal germ cells revealed mosaicism for the familial SOX9 mutation. This mutation was not seen in paternal lymphocyte DNA. Our study suggests that a mutation in SOX9 arising in a mosaic germ cell line can be transmitted in a dominant fashion and result in familial CD. This finding explains the earlier observations of unaffected parents with CD affected offspring.
While germ cell specific mutations have been implicated in a number of other genetic disorders (18 ) they have only been unequivocally demonstrated in familial von Willebrand disease (19 ), neurofibromatosis type 1 (20 ) and triplet repeat expansion disorders such as Huntington's disease (21 ). Our findings are the first to confirm that CD can be an autosomal dominantly inherited syndrome which may be caused by a mosaic, de novo germ cell mutation. However, as we only examined the father's lymphocyte derived DNA the possibility exists that other somatic tissues may carry the SOX9 mutation.
Finally, this study has shown that the same mutation in SOX9, in two individuals, can cause varying degrees of 46,XY sex-reversal. This finding is consistent with previous studies (5 ). These authors found two unrelated, 46,XY CD patients with identical mutations in SOX9, characterised by insertion of an adenine residue following nucleotide position 1462. One patient had external male genitalia with hypospadias and the other normal female external genitalia. In neither patient was gonadal morphology or internal genital duct structure reported. In our study, the two 46,XY CD siblings had different gonadal morphologies with consequently varying genital phenotypes. The range of gonadal morphologies observed may be explained by several possible mechanisms such as variable penetrance of the SOX9 mutation, increased activity of the non-mutant SOX9 allele or stochastic environmental factors. Studies in mice have also indicated that genetically identical individuals can have varying gonadal phenotypes (22 ,23 ).
It is apparent that both intact SRY and SOX9 are necessary for embryonal testis determination (24 ). Mutations in either of these genes can disrupt testis development and cause a sex-reversed phenotype. While mutations in SRY almost always cause failure of normal testicular differentiation, this is not so with SOX9. Twenty-five per cent of 46,XY patients with CD are not sex-reversed. There are now two examples of the same SOX9 mutations causing variable sexual phenotypes. All but one of the mutations in SRY have been found to lie within the HMG-box domain (25 ), while those of SOX9 span virtually the entire open reading frame. Phenotypic and mutational analyses indicate that there is no portion of SOX9 that is specifically associated with testis or skeletal development.
The ability of the SRY protein to bind and bend the DNA helix appears to be critical to its function. SOX9 has the appearance of a classical transcription factor with the HMG box DNA binding domain and a proline rich region which could act as an activation domain. However, mutation studies indicate that a CD phenotype can still occur even when SOX9 has both these apparently critical regions intact. SRY is thought to be expressed in pre-Sertoli cells and may be responsible for recruiting other cell types necessary for testicular determination. SOX9 is expressed in mesenchymal cells that are the precursors for a number of developing tissues including gonad and bone. In the testis these cells are responsible for testis cord formation. Haploinsufficiency of SOX9 may either prevent migration of these cells from the mesonephros into the developing testis or may cause these cells to be dysfunctional after they arrive.
Genotypic and phenotypic analysis of this family has allowed the mode of inheritence of CD to be established and provided new insights into the differing roles of SOX9 and SRY in mammalian testis determination.
The affected family were referred in 1984 after their first affected infant died early in the neonatal period from respiratory insufficiency. Two subsequent affected pregnancies in 1986 and 1988 were terminated at 19 weeks gestation. The karyotypes and extra-gonadal phenotypic features are listed in Table 1 . The diagnosis of all three affected siblings was confirmed at post mortem. Consequently hepatic and gonadal tissue were available in the form of paraffin-mounted, formalin-fixed blocks for the three sibs. The parents and three unaffected sibs did not display any features of CD. Heparinised blood was obtained from both parents and a semen sample was obtained from the father. The unaffected siblings were not studied.
DNA profiles were generated to confirm maternity and paternity using polymorphic CA repeats near the spino-muscular atrophy gene on human chromosome 5 (26 ). PCR with labelled primers was used to amplify this polymorphic region and the products were run on a 5% polyacrylamide gel. The parental haplotypes were concordant with the children analysed in this family (data not shown).
Formalin fixed, paraffin-mounted blocks of liver and gonadal tissue were available for each patient. These were sectioned by hand with a scalpel blade under sterile conditions to avoid cross-contamination. Genomic DNA was extracted from this material as previously described (27 ). Genomic DNA was extracted from patient blood and semen samples using standard methods (28 ).
The entire open-reading frame of SOX9 was amplified using 18 overlapping primer pairs. The PCR conditions and product sizes have been previously described (5 ). The HMG-box region of the SRY gene was PCR amplified using primers inside the HMG-box region for patients II.1 and II.2. PCR amplification was performed in both 20 [mu]l and 50 [mu]l volumes containing 0.25 mM dNTPs, 0.5 [mu]M each primer, 1 * Taq buffer (Boehringer), 1.0 unit Taq DNA polymerase (Boehringer) and 30-100 ng DNA. The cycling profile consisted of 96oC for 2.5 min, annealing at 62oC for 1 min and extension at 72oC for 1 min (1 cycle), 96oC for 20 s, annealing at 62oC for 30 s and extension at 72oC for 30 s (40 cycles), and 72oC for 5 min (1 cycle).
This was performed according to the method used by Kwok et al. (5 ). SOX9 primers were end-labelled with [[gamma]-33P]ATP and used to amplify the previously described PCR mix (10 [mu]l) under the same cycling conditions. To this was added 10 [mu]l of denaturing buffer (0.2% SDS; 20 mM EDTA) and 10 [mu]l of formamide dye (95% formamide; 20 mM EDTA; 0.05% bromophenol blue; 0.05% xylene cyanol). The reactions were denatured at 100oC for 5 min prior to loading. Two [mu]l of each reaction was loaded on to the gel which consisted of 8% acrylamide/bisacrylamide (37.5:1); 0.5 * TBE; and 5% glycerol. The gel was run at 25 W for 8 h at 4oC in 0.5 * TBE. The gels were dried and exposed to BIOMAX Film (Kodak).
If a gel shift was detected by SSCP then those primer pairs were used to PCR amplify the appropriate region of SOX9. The PCR product was purified using the Wizard Prep kit (Promega) and resuspended in 50 [mu]l H2O. Direct sequencing was performed using [[gamma]-33P]ATP end-labelled primers and the Amplicycle Sequencing kit (Perkin-Elmer). The cycling profile consisted of 95oC for 1 min (1 cycle), 95oC for 30 s, 68oC for 30 s, and 72oC for 1 min (30 cycles). Loading dye from the kit (3.5 [mu]l ) was added and the solutions denatured at 95oC for 5 min. Four [mu]l of each dideoxynucleotide reaction was loaded on to 6% acrylamide; 45% urea; 1 * TBE sequencing gel. Electrophoresis was carried out at 55 W for 4 h, gels were subsequently dried and exposed to BIOMAX film (Kodak). Sequences which appeared to be heterozygous were further examined by cloning the PCR product into the plasmid PCRscript (Stratagene) and these were sequenced as described above.
We would like to thank Jamie Foster and Alan Schafer for supplying materials and information prior to publication and Sue Forrest for the parental haplotyping. Matthijs Smith, Mike O'Neill and Craig Smith are thanked for critical reading of the manuscript. This work was supported by a grant from the National Health and Medical Research Council of Australia awarded to A.H.S.
1 Mansour, S., Hall, C. M., Pembrey, M. E. and Young, I. D. (1995) A clinical and genetic study of campomelic dysplasia. J. Med. Genet., 32, 415-420.MEDLINE Abstract
2 Tommerup, N., Schempp, E., Meinecke, P., Pedersen, S., Bolune, L., Brandt, C., Goodpasture, C., Guldberg, P., Held, K. R., Reinwein, H., Saugstad, O. D., Scherer, G., Skljeldsl, O., Toder, R., Westvik, J., van der Hagen, C. B. and Wolf, U. (1993) Assignment of an autosomal sex reversal locus (SRA1) and campomelic dysplasia (CMPD1) to 17q24.3-q25.1. Nature Genet., 4, 170-174.MEDLINE Abstract
3 Foster, J. W., Dominguez-Steglich, M. A., Guioli, S., Kwok, C., Weller, P. A., Stevanovic, M., Weissenbach, J., Mansour, S., Young, I. D., Goodfellow, P. N., Brook, J. D. and Schafer, A. J. (1994) Campomelic dysplasia and autosomal sex-reversal caused by mutations in an SRY-related gene. Nature, 372, 525-530.MEDLINE Abstract
4 Wagner, T., Wirth, J., Meyer, J., Zabel, B., Held, M., Zimmer, J. J. P., Bricarelli, F. D., Keutel, J., Hustert, E., Wolf, U., Tommerup, N., Schempp, W. and Scherer, G. (1994) Autosomal sex reversal and campomelic dysplasia are caused by mutatuions in and around the SRY-related gene SOX-9. Cell, 79, 1111-1120.MEDLINE Abstract
5 Kwok, C., Weller, P. A., Guioli, S., Foster, J. W., Mansour, S., Zuffardi, O., Punnett, H. H., Dominguez-Steglich, M. A., Brook, J. D., Young, I. D., Goodfellow, P. N. and Schafer, A. J. (1995) Mutations in SOX-9, the gene responsible for campomelic dysplasia and sex reversal. Am. J. Hum. Genet., 57, 1028-1036.MEDLINE Abstract
6 Maraia, R., Saal, H. M. and Wangsa, D. (1991) A chromosome 17q de novo paracentric inversion in a patient with campomelic dysplasia; case report and etiologic hypothesis. Clin. Genet., 39, 401-408.MEDLINE Abstract
7 Young, I. D., Zuccollo, J. M., Maltby, E. L. and Broderick, N. J. (1992) Campomelic dysplasia associated with a de novo 2q-17q reciprocal translocation. J. Med. Genet., 29, 251-252.MEDLINE Abstract
8 Chatters, S. and Whitecross, N. (1994) Campomelic dysplasia with sex-reversal associated with an apparently balanced paracentric inversion within the long arm of chromosome 17. Clin. Cytogenet. Bull., 2, 12-13.
9 Wirth, J., Wagner, T., Meyer, J., Pfeiffer, R. A., Tietze, H. U., Schempp, W. and Scherer, G. (1996) Translocation breakpoints in three patients with campomelic dysplasia and autosomal sex reversal map more than 130 kb from SOX9. Hum. Genet., 97, 186-193.MEDLINE Abstract
10 Ninomiya, S., Isomura, M., Narahara, K., Seino, Y. and Nakamura, Y. (1996) Isolation of a testis-specific cDNA on chromosome 17q from a region adjacent to the breakpoint of t(12;17) observed in a patient with acampomelic campomelic dysplasia and sex reversal. Hum. Mol. Genet., 5, 69-72.
11 Shafai, T. and Schwartz, L. (1976) Campomelic syndrome in siblings. J. Pediatr., 89, 512-513.MEDLINE Abstract
12 Mellows, H. J., Pryse-Davies, J., Bennet, M. J. et al. (1980) The camptomelic syndrome in two female siblings. Clin. Genet., 18, 137-141.MEDLINE Abstract
13 Winter, R., Rosenkranz, W., Hoffmann, H. et al. (1985) Prenatal diagnosis of campomelic dysplasia by ultrasonography. Prenat. Diagn., 5, 1-8.MEDLINE Abstract
14 Prader Von, A. (1954) Der genitalbefund beim Pseudohermaproditismus femininus des kongenitalen adrenogenitalen Syndroms. Morphologie, Hausfigkeit, Entwicklung und Vererbung der verschiedenen Genitalformen. Helv. Pediatr. Acta., 9, 231-248.
15 Braun, A., Kammerer, S., Cleve, H., Lohrs, U., Schwarz, H.-P. and Kuhnle, U. (1993) True hermaphroditism in a 46,XY individual, caused by a postzygotic somatic point mutation in the male gonadal sex-determining locus (SRY): Molecular genetics and histological findings in a sporadic case. Am. J. Hum. Genet., 52, 578-585.MEDLINE Abstract
16 Neely, E. K. and Rosenfeld, R. G. (1996) Turner syndrome. Pediatr. Endocrinol., 267-280.
17 McKusick, V.A. (1992) Camptomelic dwarfism, MIM no. 211970. In Mendelian Inheritance in Man. Johns Hopkins University Press, Baltimore.
18 Hall, J. G. (1988) Somatic mosaicism: observations related to clinical genetics. Am. J. Hum. Genet., 43, 355-363.MEDLINE Abstract
19 Murray, E. W., Giles, A. R. and Lillicrap, D. (1992) Germ-line mosaicism for a valine-to-methionine substitution at residue 553 in the glycoprotein Ib-binding domain of von Willebrand factor, causing type IIb von Willebrand disease. Am. J. Hum. Genet., 50, 199-207.MEDLINE Abstract
20 Lázaro, C., Ravella, A., Gaona, A., Volpini, V. and Estivill, X. (1994) Neurofibromatosis type 1 due to germ-line mosaicism in a clinically normal father. N. Engl. J. Med., 331, 1403-1407.MEDLINE Abstract
21 MacDonald, M. E., Barnes, G., Srinidhi, J., Duyao, M. P., Ambrose, C. M., Myers, R. H., Gray, J., Conneally, P. M., Young, A., Penney, J., Shoulson, I., Hollingsworth, Z., Koroshetz, W., Bird, E., Vonsattel, J. P., Bonilla, E., Moscowitz, C., Penchaszadeh, G., Brzustowicz, J., Alvir, J., Bickham Conde, J., Cha, J.-H., Dure, L., Gomez, F., Ramos-Arroyo, M., Sanchez-Ramos, J., Snodgrass, S. R., de Young, M., Wexler, N. S., MacFarlane, H., Anderson, M. A., Jenkins, B. and Gusella, J. F. (1993) Gametic but not somatic instability of CAG repeat length in Huntington's disease. J. Med. Genet., 30, 982-986.MEDLINE Abstract
22 Eicher, E. M., Washburn, L. L., Whitney, J. B. and Morrow, K. E. (1982) Mus Poschiavinus Y chromosome in the C57BL/6J murine genome causes sex reversal. Science, 217, 535-537.MEDLINE Abstract
23 Washburn, L L., Lee, B. K. and Eicher, E. M. (1990) Inheritance of T-associated sex reversal in mice. Genet. Res., 56, 185-191.MEDLINE Abstract
24 Sinclair, A. H. (1995) New genes for boys. Am. J. Hum. Genet., 57, 998-1001.MEDLINE Abstract
25 Hawkins, J. R. (1994) Sex determination. [Review] Hum. Mol. Genet., 3, 1463-1467.MEDLINE Abstract
26 Morrison, K. E., Daniels, R. J., Suthers, G. K., Flynn, G. A., Francis, M. J., Buckle, V. J. and Davies, K. E. (1992) High-resolution genetic map around the spinal muscular atrophy (SMA) locus on chromosome 5. Am. J. Hum. Genet., 50, 520-527.MEDLINE Abstract
27 Wright, D. K. and Manos, M. M. (1990) Sample preparation from paraffin-embedded tissues. In Innis, M. A., Gelfand, D. H., Sninsky, J. J. and White, T J. (eds). PCR Protocols. A guide to methods and applications. Academic Press, San Diego, pp.153-158.
28 Miller, S. A., Dykes, D. D. and Polesky, H. F. (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res., 16, 1215.MEDLINE Abstract
*To whom correspondence should be addressed
This page is maintained by OUP admin. Last updated Thu Oct 31 15:27:49 GMT 1996. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996