Introduction

In 1926, Greig1 first described the condition now known as Greig cephalopolysyndactyly syndrome (GCPS, MIM 175700) in a mother and daughter with digital malformations and an unusual skull shape. The syndrome is inherited in an autosomal dominant pattern and presents with characteristic limb and craniofacial anomalies. It is highly penetrant but there is both inter- and intra-familial variability of the features.2, 3 In the extremities, there may be both pre-axial and post-axial polydactyly with broad or duplicated thumbs, broad or duplicated halluces, post-axial polydactyly of hands and feet and cutaneous syndactyly of the fingers or toes. Craniofacial features include macrocephaly, prominent or broad forehead with frontal bossing, a broad base to the nose and a wide nasal bridge. Some affected individuals have learning difficulties. Developmental anomalies of the corpus callosum and mild cerebral ventriculomegaly are recognised associations. To our knowledge, trigonocephaly was described only twice in possible GCPS cases before molecular confirmation became available.4, 5

Following the recognition that GCPS was associated with rearrangements and translocations of 7p14.1,6, 7, 8, 9 heterozygous GLI3 mutations and deletions were identified as the cause of GCPS by Vortkamp et al in 1991.10, 11 The GLI3 gene encodes a zinc-finger transcription factor that is a downstream mediator of the sonic hedgehog (SHH) pathway. In the presence of SHH, full-length GLI3 functions as a transcriptional activator, whereas in the absence of SHH, GLI3 is cleaved to produce a repressor.12 The phenotype of GCPS is attributable to functional haploinsufficiency of GLI3; when this is caused by heterozygous deletion at 7p14.1, additional developmental delay may be attributable to the deletion of contiguous gene(s).13, 14

In this study we describe five unrelated individuals (four boys and a girl) with GLI3 mutation-proven GCPS and a typical pattern of pre- and post-axial polydactyly, all of whom had trigonocephaly caused by metopic synostosis, confirming that this is a nonrandom association. The key clinical features and results of molecular genetic analyses are summarised in Table 1.

Table 1 Clinical features and results of GLI3 testing in Patients 1–5

Case reports

Patient 1

This boy (Figures 1a–c), the third child of healthy unrelated parents, was tabulated (subject G31) in the report by Johnston et al.15 He was born at term by normal vaginal delivery weighing 4132 g (95th centile) and noted to have a prominent metopic ridge. At 10.5 months his occipito-frontal circumference (OFC) was 48 cm (90th centile) and inner canthal distance was 2.9 cm (95th centile). He had bilateral post-axial polydactyly type B of the fingers and pre-axial polydactyly of the feet, with cutaneous syndactyly between the hallux and second toe and partial cutaneous syndactyly of second and third toes. Cranial computed tomography (CT) scan showed metopic and sagittal synostosis. He attained his developmental milestones normally. Corrective surgery was performed on the hands and feet at 10 months and 1 year, respectively, and trigonocephaly correction at the age of 2 years.

Figure 1
figure 1

Clinical features of patients with trigonocephaly and mutations of GLI3. (a–c) Face (aged 11 weeks), right hand and feet of Patient 1. (d–f) Face (aged 3.1 years), hands (after removal of post-axial digits) and feet of Patient 2. (g–i) Face (aged 6 years), feet and CT scan of Patient 3. (j–l) Face (postoperative, aged 3 years) and feet of Patient 4. (m–o) Face (aged 2 weeks), feet and skull radiograph of Patient 5. Note trigonocephalic appearance (a, d, g, i, m), except in Patient 4, who was postoperative (j). Dysmorphic features (narrow, upslanted palpebral fissures, broad-flat nasal bridge and downturned corners of mouth) are more marked in individuals with deletions (g, j, m). Post-axial polydactyly of the fingers was present only in Patient 1 (b) and 2, and partially duplicated thumbs in Patient 2 (e), but all patients had foot abnormalities (c, f, h, l, n). Note supero-medial extension of orbits on the skull radiograph (o), caused by premature metopic suture fusion.

DNA sequencing showed a heterozygous nonsense mutation in exon 12 of GLI3 (c.1728C>A, p.Y576X; GenBank NM_000168.5),15 which was not present in either parent. This mutation is predicted to truncate GLI3 in the zinc-finger motif and falls within the mutation spectrum previously described as associated with GCPS.15, 16

Patient 2

This boy (Figures 1d–f) was included in the series of patients with metopic synostosis reported by Kini et al.17 He was referred for craniofacial assessment at 3 years with trigonocephaly. He had healthy unrelated white British parents. He was born at term weighing 3260 g (25th centile), and was noted to have limb malformations and a trigonocephalic appearance. He had partially duplicated thumbs, bilateral post-axial polydactyly type B of the hands and bilateral pre-axial polydactyly of the halluces with cutaneous syndactyly between the duplicated halluces and the second and third toes, which have required multiple corrective surgical procedures. At 4.2 years his OFC was 52 cm (75th centile) and developmental milestones were normal. Three-dimensional cranial CT scanning demonstrated metopic synostosis. He underwent fronto-orbital advancement and remodelling at the age of 4.5 years because of parental concerns about his appearance.

DNA sequencing showed a heterozygous de novo 1-bp insertion after nucleotide 1793 in exon 12 of GLI3, causing a frameshift (c.1793dupA, p.N598KfsX7), which was not present in either parent. Similar to patient 1, this mutation is predicted to truncate GLI3 within the zinc-finger motif.

Patient 3

This girl (Figures 1g–i) was born at term to unrelated parents, with a birth weight of 2440 g (<3rd centile), length 48 cm (10th centile) and OFC 32.8 cm (5th centile). She had broad thumbs and halluces, partial cutaneous syndactyly between the first and second toes bilaterally and between the second and third toes on the right, and a trigonocephalic shape to the forehead. A cranial CT scan confirmed metopic synostosis. Additional features were a ventricular septal defect (VSD) that closed spontaneously, moderate developmental delay (see Table 1 for details) and dysmorphic facial features (narrow palpebral fissures, a broad nasal bridge). At 15 years she was 160 cm tall (50th centile) and weight was 68 kg (91st centile). The karyotype was normal (46,XX).

A clinical diagnosis of Carpenter syndrome18, 19 was suggested but DNA sequencing of the open reading frame of RAB23 was normal. Array comparative genomic hybridisation (aCGH; Agilent 244K array) showed an 8.3-Mb deletion at 7p12.3-p14.1 ((40 845 981_40 855 164)_(49 136 714_49 160 830); hg18 assembly), encompassing GLI3 and 59 flanking RefSeq genes, which was not present in either parent (Figure 2a).

Figure 2
figure 2

Array CGH analysis of Patients 3 and 5 with deletions including GLI3. (a) Patient 3 with deletion of 8.3 Mb. (b) Patient 5 with deletion of 6.0 Mb. Both deletions include the GLI3 gene (position shown at top).

Patient 4

This boy (Figures 1j–l) was the first child born at 39 weeks to unrelated parents after a normal pregnancy, with a birth weight of 3560 g (50th centile), length of 50 cm (50th centile), and head circumference of 34 cm (20th centile). The hands were normal. He had bilateral duplication of the halluces and syndactyly between the second and third toes. His head shape was abnormal with a combination of scaphocephaly and trigonocephaly. Radiology confirmed synostosis of the sagittal and metopic sutures, and he had surgical correction at the age of 8 months. Additional features included laryngomalacia, atrial septal defect (ASD), VSD and patent ductus arteriosus (PDA). He developed seizures at the age of 3 years and had severe global developmental delay (Griffiths Quotient 26 aged 4 years). Ophthalmological examination showed horizontal nystagmus and mild optic nerve hypoplasia. His brain magnetic resonance imaging and karyotype (46,XY) were normal.

A clinical diagnosis of Carpenter syndrome was proposed, but in view of the similar clinical features to Patient 3, aCGH was carried out. This showed a heterozygous 6.8-Mb deletion (39 130 081–45 492 392; hg18) of the region 7p13-p14.1, encompassing GLI3 and 51 flanking RefSeq genes, which was not present in either parent (not illustrated).

Patient 5

This boy (Figure 1m–o) was the first child of healthy unrelated parents. He was born at term with normal length (53 cm, 90th centile), weight (3350 g, 50th centile) and OFC (33.5 cm, 10th centile). At the age of 2 weeks he was noticed to have trigonocephaly with a metopic ridge, large anterior fontanelle, low frontal hairline and upslanting palpebral fissures. X-ray confirmed premature synostosis of the metopic suture. Additional dysmorphic facial features included a flat nasal bridge and broad nasal base. His ears had poorly formed and overfolded helices. In his hands, the thumbs were proximally inserted with broadening of the terminal phalanx and he had bilateral pre-axial polydactyly of the halluces with cutaneous syndactyly of the second and third toes. His cranial ultrasound demonstrated mild ventriculomegaly and a small corpus callosum. Fundoscopy was normal. Echocardiography showed a double outlet right ventricle (DORV) with pulmonary stenosis. He had delayed motor development; he did not sit until 2 years of age and at 3.5 years was able to walk with one hand held. He had a vocabulary of about 15 words but had better receptive language skills. The karyotype was normal (46,XY).

Like the two previous cases the initial clinical diagnosis was Carpenter syndrome, but in view of the similar clinical features, aCGH was performed. This demonstrated a 6.0-Mb deletion ((39 013 006_39 213 707)_(45 251 621_45 449 329), hg18) at 7p13-p14.1 encompassing GLI3 and 51 other RefSeq genes, which was not present in either parent (Figure 2b).

Discussion

This series of five patients with GLI3 mutations associated with synostosis of the midline cranial sutures (metopic in all five cases and additional sagittal synostosis in two cases), establish this as an important, low frequency association of GCPS. It confirms and extends the recent reports of McDonald-McGinn et al,20 who described two cases of metopic synostosis associated with intragenic truncating mutations of GLI3 and Johnston et al,15 which tabulated (but did not comment on) three GLI3 mutation-positive cases with craniosynostosis – comprising Patient 1 presented in this study, one subject (G5-2, frameshift mutation) with metopic synostosis and one (G34-2, missense mutation) with sagittal synostosis associated with agenesis of the corpus callosum and mild ventricular prominence.

Mutations in GLI3, which contains 15 exons and spans 280 kb have been described not only in GCPS, but also in Pallister-Hall syndrome (PHS), in pre-axial polydactyly type IV, and in one patient previously diagnosed as acrocallosal syndrome.21, 22 Although GCPS and PHS share some overlapping features (notably polydactyly), they are usually readily distinguished clinically. The distinction between GCPS and pre-axial polydactyly type IV is based on the absence of any craniofacial features or learning difficulties in the latter condition. GCPS is caused by inactivating mutations and deletions13, 14 that lead to haploinsufficiency, whereas PHS is caused by frameshift, nonsense and splicing mutations in the middle third of the gene, corresponding to exon 14 and parts of exons 13 and 15.15, 16 Two of the five children reported here with GCPS had intragenic GLI3 mutations causing premature truncation of the protein; these were both located in exon 12, the region encoding the zinc-finger domain, and these mutations are expected to cause GCPS. Other recently described GLI3 mutations associated with metopic or sagittal synostosis15, 20 were in different exons (7 or 15); both mutations reported by Johnston et al15 were familial, but no other mutation-positive family member (total of four individuals) had craniosynostosis. Together with the identification of three complete deletions reported here, this reinforces the conclusion that midline synostosis is a low-frequency association of GCPS, rather than representing a distinct genotypic or phenotypic subgroup. The single report of GCPS (our case 2) in a consecutive series of 110 individuals with metopic synostosis17 indicates that GLI3 mutations are rare in this group of patients overall.

In children with GCPS due to large chromosomal deletions, additional clinical features may be attributable to haploinsufficiency of contiguous genes in the region of 7p14.1. Multiple genes may contribute to the more severe developmental delay in these children, and deletion of GCK (which occurred in Patients 3–5) is potentially associated with maturity onset diabetes of the young (MODY) type 2.13 In each of Patients 3–5, the combination of craniosynostosis and developmental delay had led to a presumptive diagnosis of Carpenter syndrome and referral for genetic testing. Carpenter syndrome is an autosomal recessive condition caused by biallelic mutations in RAB23.18, 19 Compared with GCPS, craniosynostosis is a more consistent and severe component of Carpenter syndrome, but there is overlap of the polysyndactyly features. Other clinical features that would suggest Carpenter syndrome, rather than GCPS contiguous gene deletion syndrome, are additional fusion of coronal or lambdoid sutures, high birth weight, umbilical hernia and hypogenitalism in males; these were all absent in Patients 3–5. Congenital heart disease, a recognised feature of Carpenter syndrome, was present in all three patients with GLI3 deletions. Guzzetta et al5 considered both GCPS and Carpenter syndrome in the differential diagnosis of a child with polysyndactyly, trigonocephaly and partial agenesis of the corpus callosum; in the light of the new data, a GLI3 deletion may be the most likely cause. Indeed, Patient 5 exhibited hypoplasia of the corpus callosum, a feature that could also cause confusion with acrocallosal syndrome.21

There has been a debate whether craniosynostosis is a feature of GCPS. The early report of Hootnick and Holmes4 described trigonocephaly and metopic synostosis. Most authors describe a broad and or prominent forehead as a consistent part of the facial gestalt, but not metopic synostosis. Craniosynostosis was mentioned as being present in 5% of patients in the first edition of Cohen's monograph on craniosynostosis,23 but in the second edition it was stated that no cases had been molecularly proven.24 The current series of five patients, with the four other recently reported cases,15, 20 now confirms that premature fusion of midline cranial sutures (metopic in 8/9 and sagittal in 3/9), is a feature of GCPS and is found in children both with intragenic mutations and contiguous deletions involving GLI3. The synostosis caused significant trigonocephaly, requiring cranial surgery in four of the five patients presented here. Although metopic synostosis usually leads to significant hypotelorism, this was not evident in some cases probably owing to the counteracting effect of the GLI3 mutation, which is associated with widely spaced eyes and macrocephaly.2, 3

Synostosis affecting the midline sutures (both metopic and sagittal) is heterogeneous in aetiology but share epidemiological characteristics:25 several mechanisms may contribute to their association with GLI3 mutations. Fetal head constraint is a predisposing factor in non-syndromic cases,25, 26 and the macrocephaly normally associated with GCPS would be expected to enhance this effect. A wide variety of chromosome abnormalities are associated with syndromic midline synostosis, suggesting that in cases with megabase-sized deletions, reduced transduction of stretch forces exerted by the growing brain may represent a non-specific contributing factor.27 However gene-specific mechanisms may also have a role, because mice homozygous null for the orthologous Gli3 protein have craniosynostosis, although surprisingly this affects the lambdoid sutures; by contrast the metopic and sagittal sutures are widened, with islands of ossification.28 The excess of males (four cases) over females (one case) in our series is typical of metopic synostosis as a whole.17

In conclusion, two particular clinical management issues arise from our observations. First, all children with GCPS should be examined for evidence of trigonocephaly and/or scaphocephaly to exclude the possibility of a co-existing craniosynostosis. Second, in patients in whom a diagnosis of Carpenter syndrome is being considered, the possibility of a GLI3 deletion should be included in the differential diagnosis and specifically sought using aCGH or comparable technique.