Objective: We have recently shown that de novo mutations in the TUBA1A gene are responsible for a wide spectrum of neuronal migration disorders. To better define the range of these abnormalities, we searched for additional mutations in a cohort of 100 patients with lissencephaly spectrum for whom no mutation was identified in DCX, LIS1 and ARX genes and compared these data to five previously described patients with TUBA1A mutations.
Results: We detected de novo TUBA1A mutations in six patients and highlight the existence of a prominent form of TUBA1A related lissencephaly. In four patients, the mutations identified, c.1190T>C (p.L397P), c.1265G>A (p.R422H), c.1264C>T (p.R422C), c.1306G>T (p.G436R), have not been reported before and in two others, the mutation corresponds to a recurrent missense mutation, c.790C>T (p.R264C), likely to be a hot spot of mutation. All together, it emerges that the TUBA1A related lissencephaly spectrum ranges from perisylvian pachygyria, in the less severe form, to posteriorly predominant pachygyria in the most severe, associated with dysgenesis of the anterior limb of the internal capsule and mild to severe cerebellar hypoplasia. When compared with a large series of lissencephaly of other origins (ILS17, ILSX or unknown origin), these features appear to be specific to TUBA1A related lissencephaly. In addition, TUBA1A mutated patients share a common clinical phenotype that consists of congenital microcephaly, mental retardation and diplegia/tetraplegia.
Conclusions: Our data highlight the presence of consistent and specific abnormalities that should allow the differentiation of TUBA1A related lissencephalies from those related to LIS1, DCX and ARX genes.
Statistics from Altmetric.com
Malformations of the central nervous system, some of which are grouped as neuronal migration disorders or malformations of the cortical development (MCD), are in most cases associated with profound neurodevelopmental disability, including severe mental retardation and epilepsy. MCD encompass many disorders that differ in their genetics basis, structural effects, clinical presentation and associated pathology. Over the past 20 years, our knowledge concerning MCD has increased tremendously. Major contributions have been provided by the advent of modern neuroimaging1–7 and molecular genetics.8–14 A recent update of the classification of MCD has permitted a classification into four groups: (I) malformations due to abnormal neuronal and glial proliferation or apoptosis; (II) malformations due to abnormal migration that include the lissencephaly/subcortical band heterotopia (SBH) spectrum; (III) malformations due to abnormal cortical organisation and late neuronal migration, including polymicrogyria; and (IV) unclassified malformations.7
Studies of the genetic causes of diffuse MCD resulting from neuronal migration abnormalities have shown that there are both X-linked and autosomal dominant and recessive forms. So far mutations in two X-linked genes, DCX and ARX, and three autosomal genes, LIS1, RELN and VLDL receptor, have been associated with distinct lissencephaly syndromes. Extensive genotype–phenotype studies indicate that up to 70% of the lissencephaly spectrum can be attributed to mutations in LIS1 and DCX.15–17 However, this percentage rapidly falls off when considering patients exhibiting lissencephaly with cerebellar hypoplasia (LCH).18 One of the distinctive criteria between ILS17 (isolated lissencephaly sequence, related with LIS1 gene) and XLIS (isolated lissencephaly sequence, X-linked, related with DCX gene) is that the gyral pattern is more severe posteriorly in ILS17 and more severe anteriorly in XLIS.19
Recently, we have shown that de novo missense mutations in the TUBA1A gene are also associated with MCD.20 21 In living patients, mutations in the TUBA1A gene were found to be associated with a large spectrum of lissencephaly related phenotypes often combined with dysgenesis of the corpus callosum, cerebellar and brainstem hypoplasia,21 although further studies are still required to provide a comprehensive definition and description of the clinical and neuroimaging phenotypes of TUBA1A related disorders.
To better define the range of abnormalities caused by TUBA1A mutations and to refine the distinctive criteria of TUBA1A related lissencephaly, we searched for additional mutations in a cohort of 100 sporadic patients with lissencephaly spectrum for whom no mutation was identified in DCX, LIS1 and ARX genes. Based on previous findings, this cohort was enriched with patients with cerebellar hypoplasia as this group of patients represents 50% of the cohort. Although the combination of lissencephaly and cerebellar hypoplasia might suggest a mutation in RELN gene, screening for this gene was not performed, as neither familial cases nor patients from consanguineous families were included.22
We detected a total of six de novo missense mutations, four of which are novel and an additional one that corresponds to a recurrent mutation, c.790C>T (p.R264C), was identified in two unrelated patients. We describe in detail the clinico-radiological features of these six patients, and in order to better define the phenotype of a larger group of TUBA1A mutated patients, we revisited the five patients we previously reported.21
SUBJECTS AND METHODS
Patient recruitment and mutation screening
One hundred sporadic patients with lissencephaly spectrum (that ranges from SBH to agyria/pachygyria, with or without corpus callosum abnormalities, with or without cerebellar hypoplasia or dysplasia) were referred to our laboratory for molecular screening for ARX, DCX and LIS1 mutations. According to the magnetic resonance imaging (MRI) evaluation data, 90 patients have agyria/pachygyria and 10 have SBH. In 50 patients (45 with agyria/pachygyria and five with SBH), mild to severe cerebellar hypoplasia or dysplasia were also pointed out. This population comprised 48 females and 52 males. All patients are followed regularly in different departments of paediatric neurology in France and Belgium. All patient DNA samples were tested for ARX, DCX and LIS1 mutations and no mutation was identified in any of these genes.
For the genetic analysis of the TUBA1A gene, we obtained informed consent from the parents of the affected individuals using protocols approved by local institutional review boards at Cochin Hospital and INSERM. For each patient, the complete TUBA1A coding sequence (accession number NM_006009) and splice sites were amplified in five independent polymerase chain reactions (PCR) using genomic DNA. Primer sequences, PCR conditions and product sizes were as previously reported.21 PCR products were checked by 2% agarose gel electrophoresis before they were subjected either to DHPLC (by the WAVE nucleic acid fragment analysis system (Transgenomic, San Jose, California, USA) and/or direct sequencing using the BigDye dideoxy-terminator chemistry (Applied Biosystem, Foster City, California, USA), and an ABI3700 DNA analyser (Applied Biosystem). For DHPLC analysis, melting profiles and resolution temperatures were as described by Poirier et al.21
Clinical and brain imaging assessments
All the six patients found to harbour a mutation in the TUBA1A gene were re-examined and their clinical and imaging features were re-analysed and compared to those of the five living patients with TUBA1A mutations reported in our previous study.20 21
Clinical follow up and medical records were examined to assess motor and cognitive development, neurological status and the presence of epilepsy. All MRIs were re-evaluated by two investigators (NBB and NB). Because MRI studies were performed in different institutions over a 15 year period, the imaging sequences varied substantially, though all patients had sagittal, coronal and axial T1 weighted as well as axial T2 weighted studies. Three patients had in addition axial FLAIR weight MRI, and two patients had three dimensional reconstruction studies.
The severity of the lissencephaly was scored according to the scale developed by Dobyns (defined as LIS grade): complete agyria (grade 1); partial agyria with few shallow sulci (grade 2); partial agyria–pachygyria (grade 3); generalised or partial pachygyria (grade 4).7
For comparison purpose, we also reviewed MR images from 33 ILSX patients (24 sporadic and nine familial cases) and 27 sporadic ILS17 patients (five with intragenic deletions and 22 with point mutations) (personal data).
Here, we describe the six unrelated patients aged from 19 months to 7 years who have heterozygous de novo missense mutations in the TUBA1A gene.
TUBA1A mutational analysis
Four novel missense mutations are described: c.1190T>C (p.L397P), c.1264C>T (p.R422C), c.1265G>A (p.R422H) and c.1306G>T (p.G436R). In addition, the mutation c.790C>T (p.R264C) was also found in two other unrelated patients of this cohort (fig 1). In view of the current and previous studies,20 21 this latter mutation is likely to correspond to a hot-spot of mutation in the TUBA1A gene, as its occurrence was, so far, identified in four unrelated patients (two in this study that correspond to patients 1 and 2 and two in previous reports, that correspond to cases 1 and 2 in Poirier et al21 (table 1)).
It is of interest to note that all mutations described here and in previous studies20 21 were missense mutations. This suggests that other types of mutations having more severe functional consequences (such as nonsense frameshift or deletions) might lead to either a more drastic phenotype, or to a dramatically different phenotype or to lethality—early in utero death—that were not included in the present population screened in this study.
In all cases, the described mutations were searched in patients’ parents and none of them was found mutated, suggesting the de novo occurrence of these mutations. In addition, to exclude the possibility that substitutions identified here are polymorphisms, we screened the TUBA1A gene in 360 control individuals without identifying the above described mutations.
All patients were born at full term with an uneventful delivery. Motives for specialised medical advice were severe microcephaly with postural delay (5/6 cases) associated with poor sucking during breast feeding (3/6 cases) and early seizures (2/6 cases). At the time of re-evaluation, patients were aged from 19 months to 7 years. Medical records indicated that microcephaly was already present from birth with occipito-frontal circumference (OFC) ranging from 30–32.5 cm (−2.5 to −4 SD), although biparietal diameters were normal in the third trimester of gestation in all cases. Postnatally, OFC remained between −3 and −4 SD in all cases. Motor developmental milestones were delayed in all cases with impaired ambulatory abilities and pronounced truncal hypotonia. In addition, facial diplegia with transient oropharyngoglossal dysfunction that caused transient neonatal feeding problems and later drooling was noted in three patients. Considering eye movements, 4/6 showed early onset strabismus that was either permanent in one patient (case 5) or transient consisting of bilateral divergent strabismus in three others. Adaptative skills23 varied from severe (four required full time aid) to mild (two cases) impairment. All patients had poor communication abilities. Four patients out of six were able to speak a few words and produce some associations at ages ranging from 4.5–7 years. Regarding epilepsy, two patients experienced epileptic seizures that consisted of generalised tonic seizures rapidly controlled with valproate in one case and refractory infantile spasm in the other case. Individual clinical phenotypes are detailed in table 1.
Gyral pattern in TUBA1A related lisssencephaly
In children with TUBA1A mutations, the severity of the gyral malformation varied from perisylvian predominant pachygyria (5/6) (fig 2A–D, fig 3A–C) to posteriorly predominant pachygyria that fit with Dobyns grade 4 grading (that is, generalised or partial pachygyria). In all cases, the thickness of the cortex was increased (>5 mm) in the pachygyric region (fig 3D). In all cases we also found a relative sparing of the temporal and orbito-frontal regions. All TUBA1A mutations patients also demonstrated dysgenesis of the anterior limb of the internal capsule that leads to a dysmorphic aspect of the basal ganglia, in particular of the head of the caudate nucleus combined with a deformation of the frontal horns of the lateral ventricles (fig 2A–D, fig 3B,C). Most patients had corpus callosum abnormalities that consisted in most cases of mild hypoplasia of the rostrum and splenium with flattening of the body of the corpus callosum (four cases) and partial posterior agenesis in two cases (fig 4A,B). The rest of white matter was normal in all cases. Moreover, 4/6 patients showed cerebellar abnormalities with mild vermian hypoplasia in most cases (fig 4C) and severe hypoplasia with an equal involvement of the hemispheres and the vermis (especially the inferior vermis) in one patient (fig 4B). Finally, a mild brain stem hypoplasia (fig 4C) was identified in four of six cases. Individual data are summarised in table 1.
Here we describe the distinct range of clinical and imaging phenotypes in six unrelated patients, aged from 19 months to 7 years, harbouring mutations in the TUBA1A gene. In order to give further insight into the spectrum of the TUBA1A related phenotypes, we compared these data with patients previously20 21 (table 1) and with a large number of patients with lissencephalies of other origins (ILS17, ILSX and ILS without any identified genetic causes). All together, our data highlight the presence of a potentially specific and recognisable combination of developmental features that is not commonly seen in neurodevelopmental disorders resulting from mutations in DCX, LIS1 and ARX genes.24
So far, 11 patients (six in this study and five in previous reports20 21) were found to harbour mutations in TUBA1A gene. It emerges that the most striking radiological hallmark is perisylvian predominant pachygyria that represents one of the common features of TUBA1A related lissencephaly (7/11). This pachygyria is combined with dysgenesis of the anterior limb of internal capsule that results in a dysmorphic aspect of the basal ganglia suggesting a fusion between the caudate nucleus and putamen. To our knowledge, this combination constitutes a unique pattern that seems to be specifically associated with TUBA1A mutations. However, TUBA1A related lissencephaly demonstrates a larger spectrum that ranges from posteriorly predominant agyria-pachygyria (3/11) to subcortical band heterotopia (case 5 from Poirier et al21), and often includes hypoplasia (8/11) to partial posterior agenesis of the corpus callosum (3/11), and mild (6/11) to severe (2/11) cerebellar hypoplasia.
In addition, our data indicate that TUBA1A related clinical phenotype is characterised by severe congenital microcephaly (−3 to −4 SD), moderate to severe mental retardation, axial hypotonia, motor impairment ranging from spastic diplegia to tetraplegia, early onset strabismus, facial diplegia with transient oropharyngoglossal dysfunction and epilepsy.
In order to better define the distinctive criteria of TUBA1 related lissencephaly, we compared the gyral pattern and associated brain abnormalities of the 11 TUBA1A mutations patients described here and previously20 21 with lissencephaly patterns of other origins (DCX and LIS1 mutations and unknown causes, table 2). While the thickness of the cortex is increased in most cases, it appears that one of the main differences is the distribution of gyral abnormalities. In the less severe forms of lissencephaly, including those resulting from DCX and LIS1 mutations, the predominance of pachygyria in perisylvian regions appears to be specific to patients with TUBA1A mutations (table 2). In the most severe form, the TUBA1A related lissencephaly is closer to ILS17 with a posteriorly predominant pachygyria.19 Moreover, this comparison with ILS of other aetiologies emphasises the fact that the dysgenesis of the anterior harm of the internal capsule is also a distinctive feature of TUBA1A related lissencephaly. However, it is of interest to note that 3/84 ILS patients screened in this study showed the same combination of perisylvian pachygyria and dysmorphic basal ganglia, but no TUBA1A mutation was detected. These findings suggest that other types of mutations such as deletion or duplication, or other genes including the tubulin family genes, might account for these unexplained clinico-radiological combinations of phenotypes.
Cerebellar involvement in TUBA1A mutations appear to be more frequent compared with other forms of ILS. Such ILS with prominent cerebellar hypoplasia also referred to as lissencephaly with cerebellar hypoplasia (LCH) are subdivided into six subgroups (LCH type a to e) and designates a wide range of cerebellar abnormalities range from vermian hypoplasia to severe cerebellar dysplasia with variable degrees of cortical malformations.18 Although most of the aetiology of LCH remains to be elucidated, some of them, called LCHa, are within the spectrum of DCX and LIS1 mutations, while LCHb mostly refers to RELN related lissencephalies.18 Because of the combination of cerebellar hemisphere and vermis abnormalities and the severity of microcephaly, TUBA1A related lissencephaly fits with the pattern of LCHd. Indeed, features that distinguish LCHd from LCHa include more pronounced microcephaly (at least −3SD) and a greater involvement of cerebellar hemispheres.18
In conclusion, TUBA1A mutations represent a novel form of lissencephaly that ranges from perisylvian predominant pachygyria to diffuse posteriorly predominant pachygyria, combined with internal capsule dysgenesis, cerebellar dysplasia and callosal hypotrophy. This particular MRI pattern is associated with severe congenital microcephaly, mental retardation, spastic diplegia and eventually epilepsy. This clinico-radiological combination should lead the clinician to direct molecular investigations and search for mutations in the TUBA1A gene.
This work was supported by the FRM (Fondation pour la Recherche Medicale), the ANR project (Neuro 2005 A05183KS), the French National PHRC (2003–32) and INSERM.
The authors wish to thank Professor Francis Brunelle for his helpful discussions. The authors are grateful to Professor Olivier Dulac, Dr Fiona Francis and Dr Tania Attie for their contributions. The authors are grateful to the patients and their parents who contributed in this study as well as to all the colleagues providing clinical information.
Funding: This work was supported by the FRM (Fondation pour la Recherche Medicale), the ANR project (Neuro 2005 A05183KS), the French National PHRC (2003–32), and INSERM.
Competing interests: None declared.
Patient consent: Parental consent obtained.
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.