New findings for phenotype-genotype correlations in a large European series of holoprosencephaly cases

Abstract

Background: Holoprosencephaly (HPE) is the most common forebrain defect in humans. It results from incomplete midline cleavage of the prosencephalon.

Methods: A large European series of 645 HPE probands (and 699 relatives), consisting of 51% fetuses and 49% liveborn children, is reported.

Results: Mutations in the four main genes involved in HPE (SHH, ZIC2, SIX3, TGIF) were identified in 25% of cases. The SHH, SIX3, and TGIF mutations were inherited in more than 70% of these cases, whereas 70% of the mutations in ZIC2 occurred de novo. Moreover, rearrangements were detected in 22% of the 260 patients screened by array comparative genomic hybridisation. 15 probands had two mutations providing additional support for the 'multiple-hit process' in HPE. There was a positive correlation between the severity of the brain malformation and facial features for SHH, SIX3, and TGIF, but no such correlation was found for ZIC2 mutations. The most severe HPE types were associated with SIX3 and ZIC2 mutations, whereas microforms were associated with SHH mutations. The study focused on the associated brain malformations, including neuronal migration defects, which predominated in individuals with ZIC2 mutations, and neural tube defects, which were frequently associated with ZIC2 (rachischisis) and TGIF mutations. Extracraniofacial features were observed in 27% of the individuals in this series (up to 40% of those with ZIC2 mutations) and a significant correlation was found between renal/urinary defects and mutations of SHH and ZIC2.

Conclusions: An algorithm is proposed based on these new phenotype-genotype correlations, to facilitate molecular analysis and genetic counselling for HPE.

Figure 1. Schematic representation of point mutations and deletions in the SHH, ZIC2, SIX3, TGIF genes

Point mutations are represented in yellow and gene deletions are shown in red, with the corresponding number of probands. The percentage of probands (/645 probands) with a particular point mutation or a deletion is noted below each gene.

Figure 2. Point mutations and inheritance of the SHH, ZIC2, SIX3 and TGIF genes

SHH: 41 different variations (73% missense, 27% nonsense or frameshift mutations); 73% inherited point mutations, predominantly of maternal (67%) rather than paternal (33%) origin. ZIC2: 44 different point mutations (29% missense, 55% nonsense or frameshift, 13% polyalanine tract, 3% splice alterations) with 70% occurring “de novo” and two cases of probable germline mosaicism. SIX3: 18 different point mutations (78% missense, 22% nonsense or frameshift). TGIF: 3 missense mutations and 1 nonsense mutation. The pathogenicity of these alterations has been confirmed by predictive or functional studies.[6, 18, 28, 29, 30] GenBank accession numbers are NM_00193.2 for SHH cDNA, NM_007129.2 for ZIC2 cDNA, NM_005413.2 for SIX3 cDNA and NM_003244.2 for TGIF cDNA.

Figure 3. Face defect category, as a function of HPE type, for the series as a whole, probands with abnormal array-CGH findings and with altered SHH, ZIC2, SIX3 and TGIF genes

The results presented are based on the data for 369 informative probands. Considering a positive gradient of severity from category 1 to category 4 in Spearman’s rank correlation test, we found a significant correlation between the severity of HPE type and facial features for the entire series (p<0.001) and for the SHH (p<0.001), SIX3 (p<0.001) and TGIF (p<0.001) subgroups, but not for the ZIC2 subgroup (p=0.5).

Figure 4. Algorithm for molecular analysis in HPE

In cases of familial HPE, analyses of SHH, SIX3 or TGIF should be given priority over ZIC2 analysis. Very severe forms of HPE, such as alobar and semilobar HPE, should lead to analyses of the SIX3 (particularly in cases of atelencephaly/aprosencephaly) and ZIC2 genes, but facial features may provide useful information. Indeed, probands with category 1 or 2 facial features associated with severe HPE, such as alobar or semilobar HPE, are more likely to display mutations in SIX3 or SHH. By contrast, probands with severe brain malformations and category 4 facial features are likely to have mutations in ZIC2. No particular association with lobar HPE was found. Thus, if associated brain malformations are described, particularly in cases of neuronal migration abnormalities, and in cases of neuronal tube defects (rachischisis in particular), ZIC2 gene analysis should be performed as a matter of priority. Moreover, ZIC2 mutations are the principal cause of extracraniofacial malformations. Renal/urinary malformations are found preferentially in patients with ZIC2 or SHH mutations and coloboma tends to be found in patients with SHH or SIX3 mutations.