Article Text

Download PDFPDF

First occurrence of aprosencephaly/atelencephaly and holoprosencephaly in a family with a SIX3 gene mutation and phenotype/genotype correlation in our series of SIX3 mutations
  1. L Pasquier1,
  2. C Dubourg2,
  3. M Gonzales3,
  4. L Lazaro1,
  5. V David2,
  6. S Odent1,
  7. F Encha-Razavi4
  1. 1Unité de Génétique Médicale, Hôpital SUD, Rennes, France
  2. 2Laboratoire de Génétique Moléculaire, Hôpital Pontchaillou, Rennes, France
  3. 3Laboratoire de Foetopathologie, CHU Saint-Antoine, Paris, France
  4. 4Service d’Histo-Embryologie et Cytogénétique, Hôpital Necker, Paris, France
  1. Correspondence to:
 Dr S Odent
 Unité de Génétique Médicale, Hôpital SUD, 16 Bd de Bulgarie BP 90347, 35203 Rennes cedex 2, France;

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Aprosencephaly/atelencephaly (AP/AT) refers to rudimentary prosencephalon, while holoprosencephaly (HPE) is a failure of telencephalon cleavage.1 The question of whether AP/AT results from a disruptive lesion or a growth failure is still debated.1,2 At least six causal genes, namely SHH, ZIC2, SIX3, TGIF, PATCHED, and CRIPTO (TDGF1), are linked to HPE and to ventral and dorsal patterning defect of the neural tube.3–7

We report the first occurrence of AP/AT and HPE in a non-consanguineous family with three affected sibs, carrying a maternally inherited mutation of the SIX3 gene. A phenotype/genotype correlation in the series of seven SIX3 mutations found in our cohort of 210 HPE probands is provided.


Systematic sequencing of the entire coding region of the SHH, ZIC2, TGIF, and SIX3 genes has been performed in a series of 210 non-chromosomal HPE probands including 56 patients reported previously.8 Probands were included in the study after brain imaging, karyotyping, and detailed clinicopathological studies, referring to the protocol for fetuses.9 Mutational analyses were performed as detailed elsewhere.10


Of the 210 patients screened, seven had mutations in the SIX3 gene but none in other HPE causing genes. The findings are illustrated in figs 1, 2, and 3. HPE was associated with AP/AT in three affected sibs carrying a maternally inherited mutation of SIX3.

Figure 1

SIX3 mutations in HPE: representative pedigrees showing probands and relatives. Filled symbols indicate alobar HPE with synophthalmy or cyclopy. Three quarters filled symbols indicate alobar or semi-lobar HPE without synophthalmy. Half filled symbols indicate minor features of the HPE spectrum with hypotelorism, single upper incisor, and mental retardation. Quarter filled symbols indicate isolated hypotelorism. M, SIX3 mutation carrier; N, no mutation. (A) Familial case with missense mutation in the SIX domain leading to p.Val92Gly. The last fetus showed cyclopia with atelencephaly (hatched symbol). In all cases, pregnancies were terminated because of CNS malformations. The mother had isolated hypotelorism and abnormal palate. Patient 1, the proband, a 23 weeks gestation (WG) male fetus, had alobar HPE with cyclopia, frontal encephalocele, and no visceral malformations. Patient 2, a 13 WG male fetus, had alobar HPE and arrhinia, with exophthalmos and small bilateral gingival clefts. He had no encephalocele or visceral malformations. Patient 3, a 22.5 WG male fetus, had extreme microcephaly with cyclopia, atelencephaly, sloping forehead, and absent frontonasal eminence, but no other external or visceral malformations. Neuropathological examination disclosed a forebrain reduced to a rudimentary single medial vesicle weighing 9 g (normal for age: 90 g) without telencephalic vesicles or olfactory tracts. A single optic nerve was identified as well as a pituitary stalk (fig 2). Histological evaluation (fig 3) of the rudimentary forebrain disclosed a diencephalic pattern with a well defined third ventricle and thalamic and hypothalamic nuclei. In the brain stem, the cortico-spinal tracts were lacking, but the inferior olivaries were well developed as was the cerebellum. In addition, meninges were thick and hypervascularised. The median eye displayed severe retinal dysplasia with posterior retinal coloboma. (B) Familial case with GG duplication in the SIX domain (c.556_557dup) creating a frameshift leading to a nonsense mutation 63 amino acids downstream in the homeodomain region already described.8 The family confirms the recurrence of alobar HPE with synophthalmy (two fused eyes) and proboscis. (C) Familial case with 35 bp duplication in the SIX domain (c.542_576dup) creating a frameshift leading to a nonsense mutation 58 amino acids downstream in the homeodomain region. The female fetus presented synophthalmy and 11 bilateral ribs. The mutation was inherited from the father who did not present any feature of the HPE spectrum. (D) A 2 year old girl had bilateral cleft, infantile spasms, and semi-lobar HPE, without visual disorders. She had a missense, maternally inherited mutation in the SIX domain (c.605C→T) leading to p.Thr202Ile. The mother has no features of the HPE spectrum. (E) Termination of pregnancy at 32 WG of a female fetus because of alobar HPE and median cleft palate. Visceral and ocular evaluation were unremarkable except for hypoplastic optic nerves. In this fetus a missense mutation in the homeodomain (c.692C→G) leading to p.Pro231Arg was found. DNA samples from her parents were not available. (F) The first pregnancy was terminated because of alobar HPE at 15 WG. The female fetus presents a new missense mutation in the homeodomain (c.769C→T) leading to p.Arg257Trp. In this family, recurrence of semi-lobar HPE led to the termination of the next pregnancy at 22 WG. In both cases, fetal autopsy confirmed HPE without eye disorder or visceral malformation. The mutation was inherited from the father who has no features of the HPE spectrum. (G) A 1 year old girl had alobar HPE with no severe ophthalmological abnormalities (detailed phenotypic data were not available for this patient). She had a maternally inherited missense mutation in the SIX domain (c.518A→C) leading to p.His173Pro. The mother had no clinical signs of the HPE spectrum.

Aprosencephaly/atelencephaly (AP/AT) refers to a rare forebrain malformation defined by a lack of prosencephalic derivates.11 In aprosencephaly the prosencephalon is absent, while in atelencephaly it exists as a rudimentary medial vesicle resembling the diencephalon but without the lateral telencephalic vesicles. Holoprosencephaly (HPE) results from incomplete cleavage of telencephalic vesicles. The only constant finding between the three major forms of HPE (alobar, semi-lobar, and lobar) is the continuity of the dorsal telencephalon over the midline (personal observation). Interestingly, craniofacial malformations similar to HPE are found in AP/AT and involve the fronto-nasal eminence (ranging from cyclopia/synophthalmia, cebocephaly, and midline cleft to mild hypotelorism or normal face).11–13

Our neuropathological findings in the probands of family ‘A’ fulfil classical criteria for AP/AT and HPE. Classically in AP/AT and HPE, the midbrain and hindbrain are said to be normal. However, Florell et al12 stress the existence of abnormal cerebellar hemispheres with “a peculiar mesenchymal proliferation” seen only within the central nervous system. These changes have been linked to the “glomeroid vasculopathy” described in an encephaloclastic vasculopathy leading to hydrancephaly.11 We consider thick meninges with persistent embryonic vascular dural plexus to be a common finding in HPE and AP/AT.

Key points

  • Aprosencephaly/atelencephaly (AP/AT) and holoprosencephaly (HPE) belong to a group of severe forebrain malformations. Neither the occurrence of AP/AT and HPE in the same family nor a common causative gene have ever been documented.

  • This is the first report of the occurrence of AP/AT and HPE in a family with three affected sibs carrying a maternally inherited mutation of the SIX3 gene. Phenotype/genotype correlation in this series of SIX3 mutations found a preferential association of HPE with eyes abnormalities.

  • Our neuropathological and mutational analysis with phenotype/genotype evaluation provides strong evidence for a link between AP/AT and HPE and the role of the SIX3 gene in the genesis of AP/AT.

  • In HPE with eye disorders and no other malformations, we suggest that mutational analysis should start with SIX3 gene screening.

  • SIX3 plays an essential role in the patterning of the anterior fate of neuroectoderm and in craniofacial and visual field development.

During CNS development, sequential specification along anteroposterior (rostrocaudal) and dorsoventral axes occurs even before neural tube closure in Carnegie stage 9. This requires a balance between anteriorising and posteriorising signals ending with the differentiation of the forebrain (prosencephalon), midbrain (mesencephalon), and hindbrain (rhombencephalon) following a general scheme common to vertebrates.14 By Carnegie stage 11,15 the rostral most part of the prosencephalon gives rise to the telencephalon medium followed caudally by the diencephalon. At this stage, optic vesicles originate laterally from the diencephalon wall. It is now recognised that the optic primordial appears earlier during development (by Carnegie stage 8) as a single retinal field at the rostral end of the neural plate. The prechordal plate is believed to be responsible for suppressing the median part of the retinal field leading to the emergence of two bilateral optic primordial. Lack of inhibition of the median component leads to a single eye (cyclopia) and to its variant synophthalmia. The neurohypophysial anlage rises from the ventral diencephalon and may be deficient in AP/AT as well as in HPE. The future cerebral hemispheres become identifiable during Carnegie stage 14, rising laterally from the telencephalon medium. In terms of timing, AP/AT may be considered as an early failure of the neural tube in forming a normal prosencephalon, while HPE results from the incapacity of the telencephalon to split into two hemispheric vesicles. One can postulate that in addition to cyclopia, HPE may result from lack of suppression of a presumptive “medial telencephalic field”.

Embedded Image
Figure 2

 Neuropathological findings in patient 3. (A) Coronal sections and (B) basal view of the brain. Note the absence of telencephalic vesicles, replaced by a rudimentary medial brain vesicle.

Embedded Image
Figure 3

 (A) Histological findings of the medial vesicle showing a characteristic pattern of the diencephalon. Note the presence of a third ventricle and thalamic and hypothalamic nuclei and note the thick hypervascularised meninges. (B) Histological findings of the eye showing retinal dysplasia and posterior cystic retinal coloboma.

The aetiology of AP/AT is still debated. Classically, AP/AT has been reported as a sporadic event due to an encephaloclastic process.1,2 AP/AT was associated twice with a chromosomal aberration (46,XY,del(13)(q22q31) and 45,XY,-13/46,XY,r(13)).11 In our family, the mutation of SIX3, an HPE causing gene, supports the hypothesis of a link between AP/AT and HPE. The SIX3 gene has been isolated from the HPE2 region on chromosome 2p21.16,17 This gene belongs to the human SIX family which contains several genes (SIX1/SIX2, SIX3/SIX6, and SIX4/SIX5) exhibiting strong amino acid sequence homology.18 This amino acid sequence of the SIX3 gene has mainly two DNA binding regions called the SIX domain and the homeodomain. Both share a complete amino acid sequence homology with the mouse Six3 gene and almost complete homology with optix, a Drosophila gene (expressed in eye primordia).16,17,19 In addition, the mouse Six3 gene has been found to be expressed in the retina and lens of developing eyes as well as in the hypothalamus and pituitary.20 In human, expression of SIX3 has only been studied in the retina and was detected from embryonic weeks 5–7 until adulthood.19 Interestingly, in Six3−/− mice the prosencephalon is severely truncated, telencephalic vesicles are lacking, and eyes and olfactory placodes are malformed.21

Our mutational analysis of the SIX3 gene in 210 HPE patients found seven SIX3 mutations in both highly conserved regions, two frameshift, and five missense. The new frameshift mutation led to a truncated protein in the homeodomain. This is additional evidence for the involvement of the SIX3 gene in HPE. The five new missense mutations concerned both the SIX domain and the homeodomain. Interestingly, another missense mutation, c.Arg257Pro, has been also reported in association with semi-lobar HPE, microphthalmia, and iris coloboma.17 The missense mutations could be related to HPE for two main reasons: they are localised in highly conserved regions (residues Pro231 and Arg257 which are conserved in the optix gene) and there is adequate segregation between the mutation and the associated malformations in all tested families (the mutation was not found in 100 normal chromosomes from unrelated Caucasian individuals).

Genotype/phenotype correlations in our series of seven SIX3 mutations with HPE found three cases with a spectrum of eye anomalies (from synophthalmia/cyclopia to hypotelorism). Eye abnormalities are consistent with the pattern of SIX3 expression during embryonic development in mice or in humans. Anterior encephalocele found in one case with HPE may be linked to abnormality of ectomenchymal tissue of the cephalic pole. In contrast to other HPE causing genes (SHH, ZIC2, and TGIF),22 in this series no other malformations, usually associated with the limbs and heart, were found. We suggest that mutational analysis starts with screening of the SIX3 gene in cases of HPE associated with severe eye disorders but no other malformations.

In conclusion, the occurrence of HPE and AP/AT and our mutational and neuropathological analysis provides strong evidence of a link between AP/AT and HPE which may form part of a continuum of forebrain patterning defects. These results support the hypothesis that SIX3 is central in the patterning of the anterior fate of the neurectoderm by inhibiting posteriorising (diencephalic) signals, including, but probably not limited to, Wnt signalling.23 Further studies, especially of the timing and pattern of SIX3 gene expression during early human development, should help understanding of the mechanism of AP/AT and HPE.


We are grateful to the families. We thank the French and European physicians who sent us clinical details and DNA samples, in particular Professor Moncla (Marseille, France), Dr Abel (Valence, France), Dr Francannet (Clermont-Ferrand, France), Dr Smithson (Taunton, England), and Dr Coeslier (Lille, France).



  • This work was supported by the GIS - Institut des maladies rares.

  • Competing interests: none declared