Development 103 Supplement, 233-239 (1988) 233 Printed in Great Britain © The Company of Biologists Limited 1988

The application of molecular genetics to detection of craniofacial abnormality

GUDRUN MOORE1, ALASDAJR JVENS1, JOANNA CHAMBERS', ARNI BJORNSSON2, ALFRED ARNASON3, OLAFUR JENSSON3 and ROBERT WILLIAMSON1

1 Department of Biochemistry and Molecular Genetics, St Mary's Hospital Medical School, University of London, London W2 IPG, UK, 2Department of Plastic Surgery and ^Genetics Division of the Blood Bank, National University Hospital, Reykjavik, Iceland

Summary

Congenital malformations such as secondary cleft human X-chromosome sequences will be isolated and palate can be exclusively monogenic or polygenic, but analysed for overlapping sequences and RFLPs (re- most cases have a multifactorial origin involving both striction fragment length polymorphisms) and the environmental and genetic factors, making genetic regions further defined by pulsed-field gel electrophor- analysis difficult. The new techniques of molecular esis and the identification of coding sequences. This genetics have allowed the successful chromosomal should give data on the location and structure of a gene localization of mutant genes in disorders that show a involved in the craniofacial development of the human simple Mendelian segregation, whether autosomal palatine shelves. This gene, and its protein product, dominant (e.g. Huntington's disease), autosomal re- will identify one component of the pathway that causes cessive (cystic h'brosis) or X-linked (Duchenne muscu- nonfusion of the palate. In the long term, the under- lar dystrophy). Recently, a large Icelandic family standing of the expression of this sex-linked gene for (over 280 members) with X-linked secondary cleft palate and ankyloglossia (tongue-tied) has been used secondary cleft palate and ankyloglossia will provide a as a model to localize the mutant gene associated with model for the molecular identification of other genes this craniofacial clef ting. The gene has been sub- regulating processes in craniofacial development chromosomally localized to Xql3-q21.1, using anony- whose expression is hidden in phenotypic, polygenic mous probe DXYS1; a LOD score of 307 was ob- complexity. tained. We are preparing cosmid libraries from DNA from mouse cell lines containing only the relevant part of Key words: molecular genetics, cleft palate, DNA, X- the human X chromosome, introduced by chromo- chromosome, Y-linked disorders, ankyloglossia, some-mediated gene transfer. Cosmids that contain craniofacial malformation.

Introduction weeks 5-7 and occurs at an incidence of approxi- mately 1 in 1000 births (Thompson & Thompson, The majority of congenital craniofacial malforma- 1986). Although CL is frequently associated with cleft tions occur during the 5-12 weeks of development. palate (CP), CL and CP are different both temporally The earlier part of this period approximates to the and with respect to developmental lineage. time in which teratogens can harm the developing fetus, spanning the embryonic period from 3 to 9 CP alone results from failure of the mesenchymal weeks postconception (Moore, 1982). The most com- masses of the palatine processes to fuse during weeks mon congenital abnormality of the face is cleft lip 7-12. Generally, combined cleft lip and palate (CL), a condition which involves nonfusion of the (CL+P) and CL alone are more frequent in males, upper lip and the anterior part of the maxilla during whereas for isolated CP the reverse is true (Fogh- 234 G. Moore and others Molecular genetics of craniofacial abnormality 235

Table 1. Recurrence risks (%) for cleft lip ± cleft palate

One sib and One sib and Number sibs affected one second- one third- degree relative degree relative 0 affected affected

Neither parent affected 01 3 9 6 4 One parent affected 3 11 19 16 14 Both parents affected 34 40 45 43 44

Data from Bonaiti-Pellie & Smith, 1974.

Andersen, 1942). Due to both the genetic and devel- Multifactorial traits are defined as those that are opmental evidence it is deemed justifiable to treat determined by a combination of factors, genetic and cleft lip with or without cleft palate (CL±CP) as a nongenetic, each with a minor but additive effect separate entity from CP alone (Kernahan & Stark, (such as blood pressure) (Thompson & Thompson, 1958). 1986). It is important to distinguish between multifac- The incidence of CL ± CP is higher than that of CP, torial and polygenic inheritance; the latter term varying from 2-1/1000 in Japan to 0-4/1000 in Nigeria should be used in a more restricted sense with (Leek, 1984), with the geographical variation being reference to conditions determined exclusively by a less important than ethnic differences. However, CP large number of genes, each with a small effect, alone has an average incidence of 0-7/1000 and shows acting additively, such as hair colour (Fraser, 1976). little variation in different racial groups. This may Multifactorial inheritance is more difficult to analyse mean that CP alone will not fit the purely multifac- than other types of inheritance, but is thought to torial model. Such a model includes both polygenic account for much of the normal variation in families, origin and undefined 'environmental' factors that will as well as for many common disorders, including increase the variation in incidence both geographi- cally and to some extent racially. Previous studies congenital malformations. The genetic contribution have shown that CP may include both polygenic and to multifactorial inheritance has traditionally been monogenic types (Bixler, Fogh-Anderson & Con- analysed by comparing the ratio of the concordance neally, 1971; Bear, 1976). To date, there have been rate of the condition in monozygotic and dizygotic only three pedigrees reported in which CP is clearly twin pairs with the incidence of the trait in the general inherited as a single-gene X-linked disorder (Lowry, population. 1970; Rushton, 1979; Bjornsson & Arnason, 1986). It appears that many congenital malformations result from a failure in a specific developmental Fig. 1. The pedigree of 293 individuals in the family process so that subsequent normal stages cannot be studied is shown in Fig. 1; three of us (O.J., A. A. and reached. The normal rate of development can be A.B.) have personally investigated 182 individuals in generations IV-VII. Information on other family thought of as a continuous distribution, resulting members was gained either from written or family from many environmental and genetic factors. If the records. Blood was collected from 82 individuals for normal rate of development is perturbed, a serious RFLP genotyping. These include nine males with malformation may result, dividing the continuous secondary cleft palate and ankyloglossia (CP+A) and ten variable distribution into a normal and abnormal class females with ankyloglossia alone (A). One female had separated by a threshold. Several human congenital CP alone and one male had a patent but high-vaulted malformations show family patterns that fit this palate characteristic of those affected. For purposes of multifactorial-threshold model, CL ± CP and CP be- genetic analysis, each family member was designated affected if they had any of the features of CP or A. The ing two of them. Analysis is made by collection of gene is almost certainly sex-linked, since of the 14 information on the frequency of the malformation in matings where a CP+A father has the opportunity to the general population and in different categories of pass a mutant gene to his son, in no case is a son relatives. This allows the empirical risk of the malfor- affected. This shows that this gene is almost certainly sex- mation to be estimated in subsequent pregnancies. linked as opposed to an autosomal gene with a sex- This risk is based solely on past occurrences and does difference in expression. Phenotypically unaffected not rely on knowledge of the genetic and environ- females who may be assumed to carry the mutant allele mental factors that give rise to the malformation can have affected sons and daughters, verifying that this gene is an X-linked dominant with reduced penetrance. (Table 1 gives recurrence risks). Penetrance was calculated from within the family and was Research on both the genetics and environmental found to be 82 % in the female carriers. factors involved in cleft palate has been carried out 236 G. Moore and others

\ZTTQ shelf force inducing CP in rat fetuses (Ferguson, •r<2> &jQ 1978). Mutant genes in mice have been associated with CP cho, which inhibits jaw growth (Seegmiller & Fraser, 1977) and ur, which reduces palatine shelf width (Fitch, 1957). n n n 12 In man, associations of CP and CL ± CP with HLA o • 0 0 11Kb typings have been excluded (Van Dyke, Goldman, Spielman & Zmijewski, 1983; Van Dyke, Goldman, s s s a a Spielman, Zmijewski & Oka, 1980) despite evidence a of associations of the mouse MHC regions H-2 and m m m H-3 with glucocorticoid- and phenytoin-induced CP P P P 1 I (Gasser, Mele, Lees & Goldman, 1981a; Gasser, e e Mele & Goldman, 19816). It is still not clear exactly how the genes involved in the failure of the palate to fuse, and potential thresholds to teratogenic agents, Ankyloglossia combine; large parts of the complex biochemical mechanisms are still only postulated at the theoretical Cleft 2° palate level. There have been few definitive findings which hold up in more than a single model system, and any Obligate carrier genetic or environmental factor that appears critical in one case can be excluded in another. Unaffected and The analysis of single gene mutations using RFLPs * Y-specific band for linkage studies has had considerable success in determining the chromosomal location of several Fig. 3. To^I-digested DNA samples from the pedigree common inherited disorders. RFLPs are specific are shown hybridized to the anonymous DXYSlmarker. enzyme sites that are polymorphic and segregate in Section of the Icelandic family showing the X-linked Mendelian fashion. These sites can segregate in a inheritance of CP and A. linked or unlinked manner depending on their prox- Methods imity to the disease locus. Both in cases where the DNA extraction affected chromosome is known, as for Duchenne Genomic DNA was prepared from 10 ml of whole muscular dystrophy (Davies et al. 1983) and chronic blood collected in EDTA (Kunkel et al. 1977). granulomatous disease (Royer-Pokura et al. 1986), Restriction enzyme analysis The analysis of inheritance of RFLPs was carried out and when neither the autosome nor the biochemical using standard techniques. 4/zg DNA was digested with defect is identified, as for Huntington's disease the restriction enzyme revealing the polymorphism for (Gusella et al. 1983) and cystic fibrosis (Knowlton et each X-chromosome probe. The digested DNA was al. 1985; Wainwright et al. 1985; White et al. 1985), fractionated by electrophoresis on 1 % agarose gels and linkage to within five centimorgans has been transferred to Hybond-N™ membranes (Amersham achieved. However, the analysis of polygenic dis- International) (Southern, 1975). The probes used were orders by linkage studies with RFLP markers is at labelled to a specific activity of lxl09disintsmin~' /ig"1 present more complex both practically and theoreti- by synthesis using random oligonucleotide primers cally. (Feinberg & Vogelstein, 1983). The filters were washed down to a final salt concentration of OlxSSC (SSC is One approach to dissecting the aetiology of dis- 015M-NaCl, 015M-sodium citrate) in the presence of orders with complex combinations of genetic and 0-1 % SDS at 65°C. Autoradiography was for 24 h at environmental factors is to use as a model a family in -70°C with an intensifying screen. For the linked probe which the phenotype is due to a single gene defect, pDP34 in this figure, three bands are seen. The 13 kb but displays the same features as more common band is Y-chromosome specific and can be seen in all multifactorial and sporadic cases. Such a 'model' for males. The polymorphism for this probe gives two bands midline congenital defects has been found in a large at 12 and 11 kb. The 12 kb band segregates with the Icelandic family (over 280 individuals) (Fig. 1) show- CP+A mutation, i.e. all affected males and female ing Mendelian inheritance of X-linked secondary carriers must show the 12 kb band. cleft palate and ankyloglossia ('tongue-tied') (CP+A) (Fig. 2). Both the large size of this pedigree using rats and different mouse strains, and epidemio- and the availability of many well-localized X-chromo- logical studies followed in man. The nucleotide ana- some probes has made it possible to localize this logue FUDR has been found to inhibit acid muco- defect subchromosomally. The CP+A locus has been polysaccharide synthesis and to decrease palatine found to be closely linked to one anonymous DNA 2A

B

Fig. 2. Photographs of the disorder in affected male children showing the variation in the severity of the cleft palate and ankyloglossia. (A) Partial clefting of the secondary palate; (B) complete clefting of the secondary palate; (C) ankyloglossia or tongue-tied.

Molecular genetics of craniofacial abnormality 237

KY6 D1C 56 22-3 pD2 22-2

22-1

21-3 C7 21-2 211 pERT

11-4 pOTC 11-3 I 11-23 Ll-28 11-22 CpX203 11-21 p581 111 11 I G3-1 121 12-2 pDP34 7b

13 X65H7

211 21-2

21-3 S21/S9 221 22-2 22-3 23 S21/S9

24 4315

25 43-15 A2-7

26 A2-7 27 DX13 DX13 28

-3 -2 -1 +1 +2 +3 LOD score

Fig. 4. Diagramatic description of the X chromosome showing both the in situ localization of the X probes and their maximum LOD scores in the linkage analysis of the CP+A family. Linkage and segregation analysis Combined LOD scores were calculated using the computer program package LINKAGE as described by Lathrop (1984). LOD = decimal logarithm of odds ratio, likelihood of observed recombination frequency to likelihood at 50 % recombination (completely unlinked). The standard significant cut off points for positive linkage and exclusion are +3 and —2 respectively. Multipoint linkage analysis between informative markers was carried out as described in Farrall, Scambler, North & Williamson (1986). Fig. 4 gives the maximum LOD scores at recombination fractions from 0 to 0-5 for all the informative probes. These data locate CP+A to Xql3-Xq21-1. The exclusion map on Fig. 4 is composed of 14 probes that did not show linkage to CP+A. Minimal multipoint linkage was feasible between these probes due to their low information content within this family. However from these data large sections of Xq and Xp have been excluded. 238 G. Moore and others probe DXYS1 (probe name pDP34), which maps to gene provides a model for other genes regulating Xql3-21.1 (Moore et al. 1987) (Fig. 3). processes in embyronic development. Once localization on the X chromosome has been more accurately achieved with finer mapping, cosmid We thank Action Research for the Crippled Child and libraries can be used to jump or walk from specific the Medical Research Council for generous financial sup- restriction enzyme sites to the physical location of the port, Kay Davies, Ken Kidd, Sue Chamberlain and Keith gene. Subsequent cloning and sequencing of the Johnson for many helpful discussions, all those who gener- normal gene and its mutation and analysis of its ously provided probes, and the family for their full coop- eration. functional expression using cDNA libraries and Northern blots will allow one part of the pathway that leads to cleft palate to be clarified. Autosomal se- References quences may either be homologous, or specify other parts of the same pathway, or at least give clues to BEAR, J. C. (1976). A genetic study of facial clefting in potential candidate genes involved in neural crest cell Northern England. Clin. Gen. 9, 277-284. and craniofacial development. BIXLER, D., FOGH-ANDERSEN, P. & CONNEALLY, P. M. (1971). Incidence of cleft lip and palate in the offspring Discussion of cleft parents. Clin. Gen. 2, 155-159. BJORNSSON, A. & ARNASON, A. (1986). X-linked midline defect in an Icelandic family. Abstract. 19th Congress of Localization of the mutation causing cleft palate in Scandinavian Association of Plastic Surgeons, this family to Xql3-21.1 is a first step in understand- Gothenburg. ing the genetic component of congenital neural crest BODMER, W. F. (1983). In Evolution from Molecules to defects. This region of the X chromosome contains an Man (ed. D. S. Bendall), pp. 197-208. Cambridge XY homologous region (Page, Harper, Love & University Press. Botstein, 1982). As the limits of genetic mapping in BONAITI-PELLIE, C. & SMITH, C. (1974). Risk tables for this family are approached, the techniques of cosmid genetic counselling in some common congenital walking and jumping and pulsed-field gel electro- malformation. J. med. Gen. II, 374-377. phoresis (Schwartz & Cantor, 1984; Poustka & Lehr- BURN, J. & GIBBENS, D. (1979). May spina bifida result ach, 1986) will allow the gap between the genetic and from an X-linked defect in a selective abortion physical map to be bridged. mechanism? 'Hypothesis'. J. med. Gen. 16, 210-214. DAVIES, K. E., PEARSON, P. L., HARPER, P. S., MURRAY, The isolation of the gene and studies of its spatial J. H., O'BRIEN, T., SARFARAZI, M. & WILLIAMSON, R. and temporal expression during development of the (1983). Linkage analysis of two cloned DNA sequences secondary palate should lead us to a greater under- flanking the Duchenne muscular dystrophy locus on the standing of mechanisms which may lead to the failure short arm of the human X chromosome. Nucl. Acids of closure of the palate. This mechanism can be Res. 11,2303-2312. compared with those which may cause sporadic cases FARRALL, M., SCAMBLER, P., NORTH, P. & WILLIAMSON, R. in humans, or in animal models. There are also (1986). The analysis of multiple polymorphic loci on a families known in which an X-linked form of spina single human chromosome to exclude linkage to bifida and anencephaly occur (Burn & Gibbens, 1979; inherited disease: cystic fibrosis and chromosome 4. Jensson, Arnason, Gunnarsdottir, Petursdottir, Foss- Am. J. hum. Gen. 38, 75-83. dal & Hreidarsson, 1987); any homology in linkage FEINBERG, A. P. & VOGELSTEIN, B. (1983). A technique between these families and the CP+A Icelandic for radiolabelling DNA restriction endonuclease fragments to high specific activity. Analyt. Biochem. family would be of interest. 132, 6-13. The cloning of this mutation can now be used as a FERGUSON, M. W. J. (1978). The teratogenic effects of 5- starting point for the analysis of other developmental fluoro-2-desoxyuridine (F.U.D.R.) on the Wistar rat abnormalities. For example, this gene may be a fetus, with particular reference to cleft palate. J. Anat. member of a multigene family involved in the embry- 126, 37-49. onic development of other tissues. Bodmer (1983) has FITCH, N. (1957). An embryological analysis of two suggested that genes with related functions whose mutants in the house mouse both producing cleft coordinated expression gives rise to complex pheno- palate. /. exp. Zool. 136, 329-361. types may be clustered on one chromosomal region as FOGH-ANDERSEN, P. (1942). In Inheritance of Harelip and supergenes. In mouse, five related homoeotic genes Cleft Palate. Copenhagen: A. Busck. are known to be clustered in a supergene family at a FRASER, F. C. (1976). The multifactorial/threshold concept - uses and misuses. Teratology 14, 267-280. single genetic locus. Such structure-function re- GASSER, D. L., MELE, L. & GOLDMAN, A. S. (1981a). lationships may allow direct entry into complex Immune response genes and drug-induced genetic systems involved in pattern formation and developmental abnormalities. In MHC and Diseases, ontogeny. Analysis of CP+A as a sex-linked single pp. 320-328. Basel: Karger. Molecular genetics of craniofacial abnormality 239

GASSER, D. L., MELE, L., LEES, D. D. & GOLDMAN, PAGE, D. C, HARPER, M. E., LOVE, J. & BOTSTEIN, D. A. S. (19816). Genes in mice that affect susceptibility (1982). Occurrence of a transposition from the X- to cortisone - induced cleft palate are closely linked to chromosome long arm to the Y-chromosome short arm Ir genes on chromosomes 2 and 17. Proc. natn. Acad. during human evolution. Proc. natn. Acad. Sci. U.S.A. Sci. U.S.A. 78, 3147-3150. 79, 5352-5326. GUSELLA, J. F., WEXLER, N. S., CONNEALLY, P. M., POUSTKA, A. & LEHRACH, H. (1986). Jumping libraries NAYLOR, S. L., ANDERSON, M. A., TANZI, R. E., and linking libraries: the next generation of molecular WATKINS, P. C., OTTINA, K., WALLACE, M. R., tools in mammalian genetics. TIG 2, 174-179. SAKAGUCHI, A. Y., YOUNG, A. B., SHOULSON, I., ROYER-POKURA, B., KUNKEL, L. M., MONACO, A. P., BONILLA, E. & MARTIN, J. B. (1983). A polymorphic GOFF, S. C, NEWBURGER, P. E., BAEHNER, R. L., DNA marker genetically linked to Huntington's COLE, F. S., CURNUTTE, J. T. & ORKIN, S. H. (1986). disease. Nature, Load. 306, 234-238. Cloning the gene for an inherited human disorder - JENSSON, O., AKNASON, A., GUNNARSDOTTIR, H., chronic granulomatous disease - on the basis of its PETURSDOTTIR, I., FOSSDAL, R. & HREIDARSSON, S. chromosomal location. Nature, Lond. 322, 32-38. (1988). A family showing apparent X-linked inheritance RUSHTON, A. R. (1979). Sex-linked inheritance of cleft of both anencephaly and spina bifida. J. med. Gen. (in palate. Hum. Gen. 48, 179-181. press). SCHWARTZ, D. C. & CANTOR, C. R. (1984). Separation of Yeast-Chromosome sized DNAs by pulsed field KERNAHAN, D. A. & STARK, R. B. (1958). A new classification for cleft lip and cleft palate. Plastic Gradient Electrophoresis. Cell 37, 67-75. Reconst. Surg. 22, 435-441. SEEGMILLER, R. E. & FRASER, F. C. (1977). Mandibular growth retardation as a cause of cleft palate in mice KNOWLTON, R. G., COHEN-HAGNENAUER, O., VAN CONG, homozygous for the chondrodysplasia gene. N., FREZAL, J., BROWN, V. A., BARKER, D., BRAMAN, J. Embryol. exp. Morph. 38, 227-238. J. C, SCHUMM, J. W., Tsui, L-C, BUCHWALD, M. & SOUTHERN, E. M. (1975). Detection of specific sequences DONIS-KELLER, H. (1985). A polymorphic DNA marker among DNA fragments separated by gel linked to cystic fibrosis is located on chromosome 7. electrophoresis. /. molec. Biol. 98, 503-517. Nature, bond. 318, 360-382. THOMPSON, J. S. & THOMPSON, M. W. (1986). KUNKEL, L. M., SMITH, K. D., BOYER, S. H., Multifactorial Inheritance. In Genetics in Medicine, 4th BORGAONKER, D. S., WACHTEL, S. S., MILLER, O. J., edn, pp. 210-225. Philadelphia: W. B. Saunders. BREG, W. R., JONES, H. W. & RARY, M. D. (1977). VAN DYKE, D. C, GOLDMAN, A. S., SPIELMAN, R. S. & Analysis of human Y-chromosome specific reiterated ZMIJEWSKI, C. M. (1983). Segregation of HLA in DNA in chromosome variants. Proc. natn. Acad. Sci. families with oral clefts: Evidence against linkage U.S.A. 74, 1245-1249. between isolated cleft palate and HLA. Am. J. med. LATHROP, G. M., LALONEL, J. M., JULIER, C. & OTT, J. Gen. 15, 85-88. (1984). Strategies for multilocus linkage analysis in VAN DYKE, D. C, GOLDMAN, A. S., SPIELMAN, R. S., humans. Proc. natn. Acad. Sci. U.S.A. 81, 3443-3446. ZMUEWSKI, C. M. & OKA, S. W. (1980). Segregation of LECK, I. (1984). The geographical distribution of neural HLA in sibs with cleft lip or cleft lip and palate: tube defects and oral clefts. Br. med. Bull. 40. Evidence against genetic linkage. Cleft Palate J. 17, 390-395. 189-193. LOWRY, R. B. (1970). Sex-linked cleft palate in a British WAINWRIGHT, B. J., SCAMBLER, P. J., SCHMIDTKE, J., Columbian Indian family. Pediatrics 46, 123-128. WATSON, E. A., LAW, H-Y., FARRALL, M., COOKE, H. MOORE, G. E., IVENS, A., CHAMBERS, J., FARRALL, M., J., EIBERG, H. & WILLIAMSON, R. (1985). Localisation WILLIAMSON, R., PAGE, D. C, BJORSSON, A., of cystic fibrosis locus to human chromosome 7cen-q22. ARNASON, A. & JENSSON, O. (1987). Linkage of an X- Nature, Lond. 318, 384-385. chromosome cleft palate gene. Nature, Lond. 326, WHITE, R., WOODWARD, S., LEPPERT, M., O'CONNEL, P., 91-92. HOFF, M., HERBST, J., LALOUEL, J-M., DEAN, M. & MOORE, K. L. (1982). In The Developing Human, 3rd VANDE WOUDE, G. (1985). A closely linked genetic edn, pp. 206-212. Philadelphia: W. B. Saunders. marker for cystic fibrosis. Nature, Lond. 318, 382-384.