Tricho-dento-osseous syndrome (TDO), MIM# 190320, is transmitted as a highly penetrant autosomal dominant trait that is characterized by variable clinical expression. The principal clinical features include kinky/curly hair in infancy, enamel hypoplasia, taurodontism, as well as increased thickness and density of cranial bones. Possible genetic linkage has been reported for TDO with the ABO blood group locus, but the gene defect remains unknown. We have identified four multiplex families (n = 63, 39 affected, 24 unaffected) from North Carolina segregating TDO. We previously have excluded a major locus for TDO in the ABO region for these families. Utilizing a genome-wide search strategy, we obtained conclusive evidence for linkage of the TDO syndrome locus to markers on chromosome 17q21 (D17S791, Zmax = 10.54, [Theta] = 0.00) with no indication of genetic heterogeneity. Multipoint analysis suggests the TDO locus is located in a 7 cM chromosomal segment flanked by D17S932 and D17S941. This finding represents the first step towards isolation and cloning of the TDO gene. Identification of this gene has important implications for understanding normal and abnormal craniofacial development of hair, teeth and bone.
Tricho-dento-osseous syndrome (TDO), MIM # 190320 (1 ), is inherited as a highly penetrant autosomal dominant condition that demonstrates variable expression of curly hair in infancy, enamel hypoplasia, taurodontism (enlarged pulp chambers) and both thickening and increased density of the cortical bones of the skull (2 ,3 ). Other reported features of the TDO phenotype include a high prevalence of dental caries, multiple dental abscesses, fingernail involvement including either splitting of the superficial layers or thick cornification, narrowing of the ear canal and altered craniofacial morphology, principally macrocephaly and dolichocephaly (4 ,5 ). The diversity of the TDO phenotype and the reported familial segregation of specific clinical features of the phenotype led to the proposed subclassification of TDO into three distinct subtypes, based primarily upon differences in osseous and dental findings (6 ,7 ). It remains controversial whether these distinct clinical phenotypes are the result of variable expression of one gene, allelic mutations or non-allelic gene mutations (7 ).
Principal characteristics of the TDO syndrome are alterations of hair, teeth and bone. Many craniofacial syndromes are characterized by alterations of these tissues, suggesting interrelated developmental processes. Identification of the mutation(s) responsible for TDO therefore has broad ramifications for understanding craniofacial growth and development. Additionally, understanding the biological basis for the increased bone density and thickness that occurs in TDO may provide novel therapeutic approaches for the treatment of hereditary as well as acquired osseous diseases.
Possible genetic linkage has been reported for TDO with the ABO blood group locus (8 ), but the gene defect in TDO remains unknown. The purpose of this investigation was to use a genome-wide search strategy to identify genetic linkage as a first step in identifying the gene defect(s) responsible for TDO. We have identified four multiplex families (n = 63, 39 affected, 24 unaffected) from North Carolina segregating TDO. Clinical features associated with all three TDO clinical subtypes have been identified in these four North Carolina families. We previously have excluded a major locus for TDO at the ABO blood group locus in these families (9 ). Utilizing a genome-wide search strategy, we obtained conclusive evidence for linkage of a TDO locus to markers on chromosome 17q21, with no indication of genetic heterogeneity. This is the first conclusive report of genetic linkage for TDO. This finding should facilitate identification of the gene responsible for TDO.
Four multiplex, multigenerational families (n = 63) were ascertained for the study (see Fig. 1 ). Clinical features of TDO identified in these four families included features associated with all three TDO subtypes. Details of these findings are presented elsewhere (9 ). The primary clinical findings present in affected individuals were taurodontism (100%), enamel hypoplasia (100%) and kinky/curly hair at birth (85%). Osseous changes were variably expressed, and included increased thickening of cortical bone (65%), obliterated diploe (68%), lack of frontal sinus pneumatization (49%) and lack of mastoid pneumatization (81%). Taurodontism and enamel hypoplasia, two major diagnostic criteria for TDO, were fully penetrant and were identified in all affected individuals and not identified in any unaffected individuals.
Initial testing for linkage was performed for chromosomes 1, 2, 4, 7, 9, 12, 17, 20, 21 and 22 by genotyping a subset of 40 individuals from families 1, 2 and 3 with a low density short tandem repeat polymorphism (STRP) locus screening panel (Weber Version 6A). Previous simulation studies suggested this subset of families would provide the power to detect linkage for markers within 25 cM. Only the D17S1290 marker suggested linkage to the TDO phenotype (LOD score 2.1, [Theta] = 0.11). All four families were then used to test for linkage with a high density map of additional closely spaced markers that flanked the D17S1290 marker. Two-point LOD scores for markers spanning the interval flanked by D17S1290 and D17S1293 suggested a major locus for TDO. The two-point LOD scores of these markers are summarized in Table 1 .
Table 1
Table 2
A maximum LOD score of 10.54 was obtained for D17S791 at a recombination fraction of [Theta] = 0.00. The co-segregating segment in which recombination was not detected was flanked by the markers D17S932 and D17S809. There was no evidence of genetic heterogeneity, and all four families demonstrated linkage to this region (see Table 2 ). No other genomic regions tested were suggestive for linkage, and outside the 17q21 region the maximum LOD score of 0.55, [Theta] = 0.15 was obtained for D9S925.
In order to define the smallest interval containing the TDO locus, all families were analyzed for recombination events by haplotype reconstruction. The minimum interval containing the TDO locus, defined by recombination between chromosome 17q STRPs and the TDO locus, is shown for each family in Table 3 . Relevant obligatory recombination events could be identified between the TDO locus and marker loci that placed the disease gene locus proximal to D17S932 and distal to D17S809, a distance of 12 cM. Multipoint linkage analysis using 11 markers yielded a maximum multipoint LOD score of 11.41 at D17S941. Using a criterion of LOD -1.0 to determine the 95% confidence interval (11 ), the TDO locus lies in a 7 cM interval flanked by D17S932 and D17S941 (see Fig. 2 ).
All four of the families evaluated are originally from a small geographic area in proximity to Alamance County in Western North Carolina. The common linkage observed for TDO in these four families suggests it is possible that these families are related and have inherited a common allele for TDO. Several of these families are now dispersed throughout North Carolina and the United States. We have determined that the Family 1 reported by Melnick et al. (12 ) is originally from this geographic region in North Carolina and is related to Family 3 reported here. The variable clinical manifestations of TDO found in these four families probably arise from a common TDO gene. The identification of clinical findings characteristic of all three TDO subtypes in these families, together with the possible inheritance of a common TDO gene, does not support the subclassification of TDO as previously proposed (6 ). These findings are supportive of variable clinical expressions of a common gene defect. The use of major and minor diagnostic criteria to identify TDO-affected individuals correlated completely with the co-segregating 12 cM candidate region flanked by D17S932 and D17S809. This observation suggests that TDO is a fully penetrant condition, and that these diagnostic criteria are able to identify accurately all individuals carrying the TDO gene. In order to enhance power to test our linkage hypotheses, we preferentially recruited affected individuals and their descendants. This ascertainment strategy resulted in inclusion of a greater number of affected individuals into the study population. As a result, the affected:unaffected ratio in these four families is >1. This is probably not significant, but reflects an ascertainment bias.
The TDO syndrome is characterized by alterations of hair, teeth and bone. The major diagnostic clinical features of TDO are taurodontism, enamel defects characteristic of amelogenesis imperfecta (2 ), kinky/curly hair and cranial thickening. Individually, the primary clinical findings in TDO have been reported in ~50 genetic conditions, as catalogued in McKusick's Catalogue of Mendelian Inheritance in Man (1 ); amelogenesis imperfecta (20 entries), cranial thickening (eight entries), kinky/curly hair (18 entries) and taurodontism (six entries). These tissues are commonly affected in a variety of craniofacial syndromes, indicating interrelated developmental processes. Clinically, such conditions often display phenotypic variability in the extent and severity of affected tissues, reflecting the interaction of other genetic and environmental factors in these developmental processes. Developmentally important genes including growth factors and transcription factors have been implicated in such processes (13 -16 ). Studies of these complex developmental processes are inherently difficult, particularly because many of the genes involved have yet to be identified. Naturally occurring `experiments of nature' such as TDO can provide significant insight into the growth and development of craniofacial tissues, and may lead to identification of novel developmentally important genes, or to previously unknown genetic interactions. The combination of linkage analysis and haplotype reconstruction allowed us to localize the TDO gene on chromosome 17q21 to the 7 cM genetic interval flanked by D17S932 and D17S941. This represents the first step towards the isolation and cloning of the TDO gene. The dense integrated genetic maps now available in conjunction with the increasing number of mapped genes of known function and expressed sequence tags (ESTs) mapped to this candidate region of chromosome 17 makes a positional candidate gene approach feasible. The manifestation of clinical findings involving distinct yet developmentally interrelated tissues (hair, teeth and bone), is suggestive of alteration of a common factor showing developmental and tissue-specific patterns of expression. Genetic alteration of either a transcription factor or a growth factor could account for the pleiotropic effects observed in TDO. Homeobox-type transcription factors recently have been identified as responsible for developmental defects in teeth (14 ,15 ) and bone (13 ,16 ). A number of such genes have been identified in the chromosome 17q21 co-segregating region, including nuclear transcription factors such as HOX genes (17 ), and growth factor receptors such as the insulin-like growth factor-binding protein-4 (IGFBP4) (18 ). Several of these genes are candidates for the TDO locus. The mammalian HOX gene family contains 38 homeobox gene members located in four independent linkage groups; HOXA, HOXB, HOXC and HOXD, present on human chromosomes 7, 17, 12 and 2 respectively. These are referred to as HOX1, HOX2, HOX3 and HOX4 by the Human Gene Mapping Workshop nomenclature (19 ). The human HOX2 gene family consists of nine genes. The HOX genes are expressed during embryonic development and function as determinants in the body plan organization. Two of the nine HOX2 genes, HOX2.8 and HOX 2.9, are expressed in cephalic tissues that would be consistent with the observed clinical expression observed in TDO (20 ). Craniofacial characteristics of the TDO phenotype are well documented (9 ), but possible involvement of other skeletal areas has not been carefully studied. Hereditary skeletal disorders comprise a large group of human malformation syndromes. The causative mutations can either affect the entire skeleton, or lead to an altered number, size, or shape of particular bones (13 ). It is currently unclear if the osseous changes in TDO are limited to cranial bones. A more complete knowledge of systemic characteristics of the TDO phenotype may be important for identification of analogous murine models. To date, no murine analog for TDO is known. The candidate region on human chromosome 17q21 contains the HOX gene family, and is homologous to the gene first found and so designated in the mouse, where it is located on chromosome 11 (21 ). The hox-2 genes are involved in patterning the hindbrain (22 ) and branchial arches (23 ,24 ). There is a co-linear relationship between the anterior limit of expression of a gene in branchial structures and its position in the hox-2 complex (22 ,24 ). hox2.9 and hox 2.8 are expressed anteriorly and early (temporally) (20 ). Munke and co-workers (25 ) showed that the murine hox-2 contains genes that map to the distal part of chromosome 11, near the tail-short (Ts) locus. The most obvious skeletal abnormalities in heterozygous (Ts/+) mice are short kinky tails, and fused, incompletely developed, missing or extra vertebrae occur with variable frequency (26 ). Munke et al. (25 ) suggested that Ts and hox-2 may be allelic.
Identification of the causative gene for TDO has potentially broad implications. In addition to TDO syndrome, anomalies of hair, teeth and/or bone occur in a number of heritable conditions such as the ectodermal dysplasias (1 ,27 ). Identification and characterization of the TDO gene will permit assessment of other reportedly similar conditions such as amelogenesis imperfecta with taurodontism (MIM 104510) to determine if they are genetically related to TDO (28 ,29 ). Increased understanding of how a single gene mutation leads to clinically pleiotropic effects common to several craniofacial syndromes may help increase our understanding of normal as well as abnormal development of these tissues (30 ). While the hair and dental findings characteristic of TDO occur early in development, osseous changes appear to occur throughout adulthood. The increased bone thickness and density associated with TDO is not accompanied by any evident pathology. Understanding the basis for these osseous changes has implications for treatment of many forms of heritable and acquired osteopathies, including osteoporosis and bone loss associated with periodontal diseases.
Four extended kindreds (n = 63) segregating for TDO were identified through proband ascertainment of individuals presenting for treatment at the University of North Carolina pediatric dental clinics (see Fig. 1 ). Medical and dental histories were obtained from all available family members. Informed consent was provided by all study participants prior to inclusion in the study. All available family members received oral clinical and radiographic (standardized cephalometric and panoramic radiographs) examinations as previously described (5 ). Classification of TDO-affected status was based on clinical and radiographic features. Major diagnostic criteria for inclusion as affected included all three of the following: (ii) a positive family history of TDO; (ii) presence of generalized hypoplastic dental enamel (assessed by clinical and radiographic examination); and (iii) presence of taurodontism in at least two posterior teeth (assessed by radiographic examination). In addition to meeting these major criteria, to be classified as affected the presence of at least one of the following minor features was necessary: (i) curly/kinky hair at birth (verified in baby pictures); (ii) thick and/or dense bone (as assessed by cranial radiographs); or (iii) abnormal nails (e.g. brittle and peeling, cornified). These diagnostic criteria have been proposed to delineate TDO from conditions with similar phenotypes such as amelogenesis imperfecta with taurodontism (28 ).
Peripheral venous blood (7.5 ml) was obtained by standard venipuncture. Genomic DNA was extracted using the QIAamp blood kit (QIAMP, Inc.). An initial genome-wide scan was performed using the Weber version 6A low density markers (Research Genetics, Huntsville, AL) using standard techniques for PCR amplification with [gamma]-32P radioactively labeled primers according to the manufacturer's protocol using a PCR 9600 thermocycler (Applied Biosystems) (31 ). After identification of a linkage relationship, a high density array of DNA markers were selected from the Cooperative Human Linkage Center chromosome 17 version v8c7 integrated marker map (www.chlc.org). Following PCR amplification, individual samples were separated on a 6% PAGE-7 M urea gel (30 W, 1500 V). An M13 sequencing ladder (Sequenase kit, USB) was loaded onto each gel to permit sizing of individual alleles. Following electrophoresis, gels were wrapped in cellophane and exposed in a phosphorimaging cassette for 15 min, and scanned with a scanner (Molecular Dynamics, Inc.). Alleles were scored and genotype data entered into the pedigree file of the LINKAGE computer package (32 ).
LOD scores were generated assuming autosomal dominant inheritance with complete penetrance. The affected allele frequency was taken as 0.0001. Marker allele frequencies were assumed to be distributed uniformly. Calculations were also performed using marker allele frequencies reported in available databases, but these changes had minimal effects on the LOD scores generated.The number of alleles observed at each marker locus are shown in Table 1 .
Two-point linkage analysis was performed by use of the MLINK program version 5.1 from the LINKAGE computer program (32 ). At any given locus, results for the pedigree were used to generate a final LOD score for each marker tested. Precise values for maximum LOD scores were calculated with the ILINK program from the same computer package. Multipoint analyses were performed using FASTLINK (33 ,34 ). A multipoint map (Fig. 2 ) was constructed from several runs with overlapping sets of marker loci. Haplotype analysis was used for error elimination during the linkage scan and for the determination of the critical linkage segment. Haplotype construction was performed using the CRIMAP program with the CHROMPIC option (35 ).
We acknowledge Sue Stonehouse for manuscript preparation, and support from R29 DE10990, Bowman Gray School of Medicine Venture Grant, and the Shared Analytical Imaging Facility, Bowman Gray School of Medicine, supported by grant 2 P30 CA12197-25.
Human Molecular Genetics
Pages
Introduction
Results
Clinical findings
Two-point linkage analysis
Haplotype and multipoint linkage analyses
Discussion
Materials And Methods
Ascertainment and classification of families
DNA marker analysis
Linkage analysis
Acknowledgements
References
Marker
Loc
n
Het
0.00
0.01
0.05
0.10
0.20
0.30
Zmax
[Theta]max
D17S1293
63
8
0.88
-[infinity]
-5.74
-0.71
0.95
1.82
1.60
1.83
0.22
D17S800
69
7
0.75
-[infinity]
-4.47
-0.75
0.46
1.07
0.92
1.08
0.22
D17S932
70
7
0.75
-[infinity]
1.60
3.75
4.13
3.62
2.56
4.13
0.10
D17S791
73
12
0.94
10.54
10.35
9.59
8.60
6.50
4.29
10.52
0.00
D17S806
76
13
0.75
10.07
9.89
9.16
8.22
6.23
4.13
10.06
0.00
D17S943
77
7
0.75
8.33
8.18
7.54
6.70
4.92
3.09
8.32
0.00
D17S941
80
4
0.75
3.48
3.41
3.13
2.76
1.97
1.20
3.47
0.00
D17S809
82
6
0.62
-[infinity]
2.21
3.21
3.26
2.68
1.79
3.29
0.08
D17S787
83
8
0.88
-[infinity]
2.93
3.83
3.79
3.10
2.14
3.87
0.07
D17S957
89
5
0.38
-[infinity]
2.34
3.85
4.01
3.35
2.29
4.02
0.09
D17S923
90
5
0.44
-[infinity]
3.99
4.84
4.73
3.79
2.53
4.86
0.06
D17S1290
90
9
0.92
-[infinity]
3.63
5.02
5.03
4.06
2.70
5.11
0.07
Recombination fraction ([Theta])
0.00
0.01
0.05
0.10
0.20
0.30
Family 1
4.83
4.75
4.42
3.99
3.06
2.06
Family 2
2.44
2.39
2.18
1.92
1.38
0.84
Family 3
2.11
2.07
1.93
1.74
1.34
0.90
Family 4
1.16
1.14
1.05
0.94
0.72
0.49
REFERENCES
This page is maintained by OUP admin. Last updated Sat Oct 18 13:29:50 BST 1997. Part of the OUP Journals World Wide Web service.
Copyright
Oxford University Press, 1997