The Genetics of Holoprosencephaly

THE etiology of holoprosencephaly is heterogeneous and likely can be due to mutations in several different genes. It can be seen in association with chromosomal aberrations, multiple congenital anomaly syndromes, and as nonsyndromic HPE.

Chromosomal Abnormalities

Approximately 24%-45% of live births with HPE have a chromosomal abnormality (Olsen et al., 1997). The craniofacial phenotype does not consistently predict whether there is an associated cytogenic abnormality. Although HPE has been seen in association with a large number of different chromosomal anomalies, certain chromosomal regions display recurrent involvement. Trisomy 13 is the most frequent chromosomal abnormality, and accounts for over one-half of HPE with cytogenetic aberrations. HPE is present in approximately 70% of trisomy 13 cases. Other anomalies affecting 13q, including duplications, deletions, or rings, are also associated with HPE.

We hypothesize that these chromosomal regions which are rearranged may contain genes critical for normal brain development, and abnormalities in these genes could result in HPE. At the Human Gene Mapping 11 conference, four genes which may play a role in HPE were designated: HPE1 at 21q22.3, HPE2 at 2p21, HPE3 at 7q36, and HPE4 at 18p. Based on nonrandom cytogenetic rearrangements, at least 12 chromosomal regions on 11 chromosomes may contain genes involved with HPE.

Familial Holoprosencephaly

HPE is also associated with several multiple congenital anomaly syndromes, some of which are monogenic. In addition to these syndromes, familial instances of nonsyndromic HPE with normal chromosomes have also been described. Pedigrees have been described that support autosomal dominant, autosomal recessive, and possibly X-linked inheritance and provide further evidence of a genetic basis of HPE. The clinical variability can be quite striking within a single pedigree. Individuals with the full range of severity of HPE, microforms, and even clinically normal individuals may be present within a single kindred. Individuals with microforms, or clinically unaffected carriers are all at risk for having a child with HPE. Although precise genotype-phenotype correlations have not yet been performed in families with a known mutation in HPE gene, the penetrance in autosomal dominant HPE is estimated to be 70% (Cohen, 1989).

Reported pedigrees with clinically unaffected parents and multiple affected siblings suggest autosomal recessive inheritance. In addition, there are multiple examples of consanguinity, supporting the concept of autosomal recessive inheritance. However, since abnormal HPE genes are not fully penetrant and there is also the possibility of germline mosaicism, some of these cases may actually be autosomal dominant. The true inheritance pattern of these pedigrees will be determined by molecular analysis once the genes responsible for HPE in these kindreds are identified.

Current Research

Investigators at the Muenke Lab at the Children's Hospital of Philadelphia and the National Institutes of Health have focused on studying the genes involved in HPE. Studies utilizing deletions and translocations involving HPE3 narrowed the critical region to 7q36. Sonic Hedgehog (SHH) was found to map to this region. SHH is a secreted factor expressed early in development in the ventral forebrain that is critical for ventral patterning of the developing neural tube. "Knockout" mice with homozygous null mutations for SHH displayed abnormalities consistent with HPE. Because of itıs chromosomal location, the mouse knockout, and its expression pattern in early embryonic brain development, SHH was considered a strong candidate gene for HPE. The complete coding region and intron-exon boundaries are being screened for mutations in a panel of DNA samples derived from over 300 unrelated individuals with HPE. Ten unrelated individuals were found to have SHH mutation (Roessler et al., 1996). The DNA changes were detected in seven autosomal dominant pedigrees previously linked to 7q36, two additional smaller families with undetermined linkage, and a sporadic case of HPE. Eight new mutations have been found in the coding region of the gene: 4 missense mutations, 2 stop codons, 1 deletion and 1 insertion. Future studies to assess the functional effects of these mutations will be performed.

Other candidate HPE genes which are currently being screened in the Muenke Lab include PATCHED, SIX3, TGIF, and DKK-1. For each gene, the complete coding sequence and intron-exon boundaries are being analyzed for mutations in the panel of over 300 individuals with HPE. In each case where a mutation was detected at least 100 control unaffected individuals were also assessed and do not have the mutation.

PATCHED (PTC) a multipass transmembrane protein binds SHH and likely is its receptor, leading us to think that this is a candidate gene for HPE. In the absence of SHH, PTC inhibits the SHH pathway. When SHH binds to PTC, the pathway is released from inhibition and signaling proceeds (Hammerschmidt et al., 1997). It is hypothesized that mutations in PTC resulting in constitutive inhibition could cause HPE. Thus far, three mutations have been found in PTC. Functional studies of these mutations are in progress. The mutations may cause HPE either by perturbing SHH binding or by inhibiting activation of the SHH pathway, thus preventing SHH signaling.

SIX3 has been identified as a candidate gene based on its murine expression pattern in the ventral forebrain and its chromosomal localization to the HPE2 minimal critical region. It is a DNA binding protein which is a homologue of the Drosophila homeobox gene sine oculis which participates in eye development (Oliver et al., 1995). To date three missense mutations have been identified in the homeodomain of unrelated patients with HPE. These data suggest that mutations in SIX3 are associated with HPE and that haploinsufficiency of SIX3 may result in the HPE phenotype. Further studies will offer insight into the significance of these mutations in the role of HPE.

The minimal critical region in 18p11.3 was defined. Within this region we identified a candidate gene, TGIF. TGIF is expressed during early brain development in mice. The gene product is a transcription factor which competes for DNA binding sites with a retinoid-responsive promoter element (Bertolino et al., 1996). This competition is of particular interest since retinoic acid exposure in mice and hamsters causes malformations consistent with the HPE spectrum. Mutation screening on 240 patients with HPE was completed and four base changes predictive of amino acid substitutions were found. Further studies are being done to study the DNA binding of the abnormal gene product.

The last candidate gene we are currently looking at is Dickkopf-1 (DKK-1). DKK-1 is sufficient and necessary to cause head induction in Xenopus and in some embryos in which DKK-1 activity is blocked they develop cyclopia (Glinka et al., 1998). DKK-1 has been cloned and currently we are screening HPE patients for mutations in this gene.

Summary

The genetic basis is heterogeneous. We are studying several genes which are excellent candidate genes for HPE based on their expression patterns and their biological functions. We have identified mutations in several of these genes in patients with HPE. Our future work will examine the functional effects of these mutations as well as examine the additional genes which play a role in HPE.

Investigators:

Maximilian Muenke
Colleen Campbell
Yangzhu Du
Jeff Ming
Erich Roessler
Luisa Nanni
Karen Gripp
Deeann Wallis

References:

Bertolino et al. (1996) Develop. Dynamics 205:400-420
Cohen (1989) Teratology 40:211-235
Glinka et al. (1998) Nature 391:357-362
Hammerschmidt et al. (1997) Trends in Genetics 13:14-21
Oliver et al. (1995) Development 121:4045-4055
Olson et al. (1997) Am J Med Genet 73:217-226
Roessler et al. (1996) Nature Genetics 14:357-360