Abstract

Complex cortical malformations associated with mutations in tubulin genes: TUBA1A, TUBA8, TUBB2B, TUBB3, TUBB5 and TUBG1 commonly referred to as tubulinopathies, are a heterogeneous group of conditions with a wide spectrum of clinical severity. Among the 106 patients selected as having complex cortical malformations, 45 were found to carry mutations in TUBA1A (42.5%), 18 in TUBB2B (16.9%), 11 in TUBB3 (10.4%), three in TUBB5 (2.8%), and three in TUBG1 (2.8%). No mutations were identified in TUBA8. Systematic review of patients’ neuroimaging and neuropathological data allowed us to distinguish at least five cortical malformation syndromes: (i) microlissencephaly (n = 12); (ii) lissencephaly (n = 19); (iii) central pachygyria and polymicrogyria-like cortical dysplasia (n = 24); (iv) generalized polymicrogyria-like cortical dysplasia (n = 6); and (v) a ‘simplified’ gyral pattern with area of focal polymicrogyria (n = 19). Dysmorphic basal ganglia are the hallmark of tubulinopathies (found in 75% of cases) and are present in 100% of central pachygyria and polymicrogyria-like cortical dysplasia and simplified gyral malformation syndromes. Tubulinopathies are also characterized by a high prevalence of corpus callosum agenesis (32/80; 40%), and mild to severe cerebellar hypoplasia and dysplasia (63/80; 78.7%). Foetal cases (n = 25) represent the severe end of the spectrum and show specific abnormalities that provide insights into the underlying pathophysiology. The overall complexity of tubulinopathies reflects the pleiotropic effects of tubulins and their specific spatio-temporal profiles of expression. In line with previous reports, this large cohort further clarifies overlapping phenotypes between tubulinopathies and although current structural data do not allow prediction of mutation-related phenotypes, within each mutated gene there is an associated predominant pattern of cortical dysgenesis allowing some phenotype–genotype correlation. The core phenotype of TUBA1A and TUBG1 tubulinopathies are lissencephalies and microlissencephalies, whereas TUBB2B tubulinopathies show in the majority, centrally predominant polymicrogyria-like cortical dysplasia. By contrast, TUBB3 and TUBB5 mutations cause milder malformations with focal or multifocal polymicrogyria-like cortical dysplasia with abnormal and simplified gyral pattern.

Introduction

The formation of the complex architecture of the mammalian brain requires a coordinated timing of proliferation, migration and layering, as well as differentiation of distinct neuronal populations. The cellular processes and related biological mechanisms underlying their disruption are traditionally used as a basis for the classification of the large spectrum of malformations of cortical development that represent a significant cause of neurological morbidity in children leading to intellectual disability, cerebral palsy, and/or epilepsy. This classification system divides brain malformations into disorders of (i) abnormal neuronal and glial proliferation; (ii) neuronal migration; and (iii) abnormal post-migrational development (Barkovich et al., 2005, 2012).

The division of neuronal progenitors, as well as the striking morphology of neurons as they migrate and establish their dendritic and axonal arbors, implies a strictly regulated process of cytoskeletal structuring and polarization (Higginbotham and Gleeson, 2007; Hoogenraad and Bradke, 2009). Compelling evidence has demonstrated that the nervous system development is highly dependent upon the microtubule cytoskeleton. Microtubules are polymers of tubulin heterodimers, which are themselves assembled from various α- and β-tubulin isotypes encoded by separate genes (Lopata and Cleveland, 1987). Microtubule behaviour varies depending on isotype composition, suggesting that different isotypes may have distinct properties adapted to specific cellular functions (Joshi and Cleveland, 1990; Luduena, 1993).

Recent genetic studies have pointed out critical effects of tubulins and microtubule-associated proteins involvement in malformations of cortical development. Up to now mutations in tubulin genes encoding different α- and β-tubulin isotypes (TUBA1A, TUBA8 and TUBB2B, TUBB3, TUBB5) and more recently γ-tubulin (TUBG1) have been shown to be involved in a large spectrum of developmental brain disorders (Keays et al., 2007; Abdollahi et al., 2009; Jaglin et al., 2009; Tischfield et al., 2010; Breuss et al., 2012; Poirier et al., 2013a). Mutations in TUBA1A, TUBB2B, TUBB3 and TUBB5 are heterozygous missense mutations whereas the unique TUBA8 mutation consists of a homozygous 14 bp intronic deletion upstream of exon 2 (Keays et al., 2007; Abdollahi et al., 2009; Jaglin et al., 2009; Tischfield et al., 2010; Breuss et al., 2012). All mutations in TUBA1A and TUBB5 and TUBB2B are sporadic and occur de novo whereas TUBB3 mutations are a mixture of familial and de novo mutations. The first tubulin gene involved in brain malformation encodes α-tubulin 1A. Initial reports highlighted the presence of specific abnormalities found in TUBA1A-related lissencephalies. These consist of a unique combination of microcephaly, variable cortical malformations that range from perisylvian pachygyria (also named, ‘variant pachygyria’ most prominent in posterior frontal, perisylvian and parietal regions) to posteriorly predominant lissencephalies associated with absence or hypoplasia of the anterior limb of the internal capsule, complete or partial agenesis of the corpus callosum and cerebellar and brain stem hypoplasia (Keays et al., 2007; Poirier et al., 2007; Bahi-Buisson et al., 2008; Morris-Rosendahl et al., 2008; Kumar et al., 2010; Sohal et al., 2012; Amrom et al., 2013; Okumura et al., 2013; Hikita et al., 2014). Subsequently, the spectrum of TUBA1A-related lissencephaly was enlarged with the description of lissencephaly with cerebellar hypoplasia and ‘simplified’ gyral pattern plus complete corpus callosum agenesis and moderate cerebellar hypoplasia (Kumar et al., 2010). More recently, TUBA1A mutations were also described in perisylvian asymmetrical polymicrogyria (Jansen et al., 2011; Poirier et al., 2013; Zanni et al., 2013), and polymicrogyria-like cortical dysplasia (Cushion et al., 2013). In parallel, five foetal cases that probably represent the more severe end of the spectrum of TUBA1A lissencephaly were reported, displaying a combination of complete to posteriorly predominant agyria, corpus callosum agenesis or hypogenesis, severe microcephaly, and variable brainstem and cerebellar defects, consistent with either the phenotype of classic lissencephaly or lissencephaly with cerebellar hypoplasia (Fallet-Bianco et al., 2008; Lecourtois et al., 2010).

Heterozygous missense TUBB2B mutations were initially described in patients with bilateral, asymmetric polymicrogyria, typically more predominant in the frontal and temporal lobes. Similar to TUBA1A mutations, TUBB2B mutation-related cortical dysgenesis was associated with dysmorphic basal ganglia, partial or complete agenesis of the corpus callosum and hypoplasia or dysplasia of the cerebellum (Jaglin et al., 2009; Amrom et al., 2013). Subsequently, TUBB2B mutations were described in symmetrical polymicrogyria as well as schizencephaly, further broadening the spectrum of TUBB2B-related cortical dysgenesis. Remarkably, in some cases the corpus callosum as well as the cerebellum were normal (Guerrini et al., 2012; Romaniello et al., 2012). Recently, Cushion et al. (2013) reported four additional severe cases of TUBB2B mutations showing in all cases polymicrogyria-like cortical dysplasia together with dysmorphic basal ganglia, brainstem and cerebellar vermian hypoplasia. Moreover, one of the TUBB2B mutated cases demonstrated a lissencephaly pattern with an agyric cortex, and bilateral band of heterotopic cortex suggesting that an overlap may also exist in the TUBA1A and TUBB2B related lissencephalies (Cushion et al., 2013).

TUBB3 mutations cause a more diffuse spectrum of brain malformations and neurological disabilities and certain phenotypes often segregate with particular amino acid substitutions (Poirier et al., 2007; Tischfield et al., 2010). Initially, heterozygous familial and de novo TUBB3 missense mutations were reported in a paralytic eye movement disorder (i.e. congenital fibrosis of extraocular muscle, CFEOM3), which results from hypoplasia of the oculomotor nerve(s) combined with a sensorimotor polyneuropathy and mild to moderate intellectual disability, depending upon the specific amino acid substitution. Brain malformations include agenesis or hypoplasia of commissural axon tracts, hypoplasia of the corticospinal tract, and dysmorphic basal ganglia with fusion of the caudate and putamen (Tischfield et al., 2010). Subsequently, the TUBB3 phenotypic spectrum was enlarged with frontal polymicrogyria or simplified and disorganized gyral patterning, brainstem and cerebellar vermian hypoplasia and basal ganglia dysmorphism. Also, one foetal case was reported with microlissencephaly and corpus callosum agenesis, severe brainstem and cerebellar hypoplasia and dysmorphic basal ganglia (Poirier et al., 2010).

More recently, recurrent p.E410K TUBB3 mutations were described, resulting in a specific clinical association that combines intellectual disability, Kallmann syndrome, congenital fibrosis of extraocular muscles, facial weakness, tracheomalacia, vocal cord paralysis and later-onset cyclic vomiting and progressive peripheral neuropathy. Patients show on MRI, no cortical malformations but a thin corpus callosum and anterior commissures, as well as hypoplastic to absent olfactory bulbs, oculomotor and facial nerves, which are consistent with abnormalities in axon guidance and maintenance. This peculiar association led the authors to define it as the ‘TUBB3 E410K syndrome’ (Chew et al., 2013).

Subsequently, TUBB5 mutations were reported in three patients displaying various patterns of brain dysgenesis ranging from gyral simplification to focal polymicrogyria and microcephaly (Breuss et al., 2012). Far less common is the homozygous deletion in TUBA8 that was reported in two consanguineous pedigrees with polymicrogyria and corpus callosum hypoplasia in association with optic nerve hypoplasia (Abdollahi et al., 2009). However, although this phenotype merges with other tubulinopathies, additional mutations are needed to definitively support TUBA8 as a polymicrogyria-associated gene. Finally, most recently the tubulin TUBG1, a gamma tubulin involved in centrosome organization, was also found to be mutated in three patients with lissencephaly with microcephaly (Poirier et al., 2013a).

In the present paper, we report the largest cohort of 80 tubulin-related cortical malformations that result from tubulin mutations in TUBA1A, TUBB2B, TUBB3, TUBB5, or TUBG1 genes. We have attempted to refine the spectrum of cortical dysgenesis in our cohort and identified five distinct malformation subtypes that correspond to agyria-pachygyria (lissencephaly), microlissencephaly, ‘central’ pachygyria and polymicrogyria-like cortical dysplasia, generalized polymicrogyria-like cortical dysplasia, and finally a ‘simplified’ gyral pattern with area of focal polymicrogyria. Findings on all foetal cases are reported in detail elsewhere (Fallet-Bianco et al. submitted for publication). Based on structural data, we discuss the phenotype according to the position of mutated residues and putative consequences of mutations on tubulin assembly and interactions. With five tubulin genes involved, together with previous published cases, the aim of this paper is to propose a diagnostic strategy and to identify genotype–phenotype correlations to further improve molecular testing and diagnosis of malformations of cortical development.

Patients and methods

Patient selection

As part of our increasing collection of patients with malformations of cortical development, comprising to date 540 living patients and 60 foetuses (corresponding to 564 sporadic cases and 36 familial cases) referred to our laboratory for genetic and molecular investigations, we identified mutations in tubulin genes for 80 patients with cortical malformations. Unexplained cases of malformations of cortical development referred in our laboratory for molecular screening encompass the whole range of non-syndromic cortical dysgenesis, agyria-pachygyria (n = 58), typical polymicrogyria (n = 420) or polymicrogyria-like cortical dysplasia (n = 60) or gyral simplification pattern (n = 62).

Of these, 106 patients met the criteria of complex cortical malformation, i.e. combination of cortical dysgenesis, either agyria-pachygyria (n = 23), microlissencephaly (n = 16), polymicrogyria or polymicrogyria-like cortical dysplasia (n = 42), or ‘simplified’ gyral pattern with area of focal polymicrogyria (n = 25) plus partial or complete agenesis of corpus callosum, dysmorphic basal ganglia, brainstem hypoplasia and/or cerebellar hypoplasia or dysplasia and/or microcephaly. All were sporadic cases except for six familial cases from three families. Included patients were from 20 centres in France and Switzerland. All patients were known personally to at least one of the authors. Eighty of 600 (13.3%) patients with tubulinopathies were identified. Of these 80 patients (55 living patients and 25 foetuses), 39 were previously reported by our group and re-evaluated for the purpose of this review, 18 patients with TUBA1A mutations, five patients with TUBB2B mutations, 10 patients with TUBB3 mutations, three with TUBB5 mutations and three with TUBG1 mutations (Keays et al., 2007; Poirier et al., 2007, 2010, 2013a, b; Fallet-Bianco et al., 2008; Jaglin and Chelly, 2009; Lecourtois et al., 2010; Breuss et al., 2012).

Mutation analysis

Blood samples for DNA preparation and genetic investigations were obtained with informed consent from parents and/or patients. DNA was extracted using standard protocols. For each patient, array-CGH was normal and LIS1, DCX and GPR56 mutations and deletions were excluded. Mutation analysis of the coding regions of these six genes (TUBA1A, TUBA8, TUBB2B, TUBB3 and TUBB5 or TUBG1), was carried out on all patients, as described previously (Keays et al., 2007; Abdollahi et al., 2009; Jaglin and Chelly, 2009; Poirier et al., 2010, 2013a, b; Breuss et al., 2012), and sequences of primers and PCR conditions are available upon request. For each patient, the mutation was tested by direct sequencing in both parents’ DNA to assess whether they occurred de novo, or were inherited from one of the parents.

Clinical information, brain magnetic resonance imaging and classification

All patients are followed regularly in various departments of Paediatric Neurology and were known personally to at least one of the authors and were re-examined for the purpose of the study. Both the clinical information and brain MRIs were obtained and centralized by N.B.B. at the Paediatric Neurology unit of Necker Enfants Malades Hospital, Paris Descartes University. Detailed information regarding family history, pre- and perinatal events, motor development, cognitive function, neurological examination including head circumference and ophthalmological features was recorded. With regards to the epilepsy, age of seizure onset, main seizure type and response to antiepileptic drugs was determined.

Brain MRIs were reviewed by at least two investigators (N.B.B and N.B) for the purpose of classifying patients. The scans were reviewed using a proforma in which the following abnormalities (using widely accepted criteria) were recorded. Cortical malformations were classified as follows: polymicrogyria was diagnosed according to the three recognized criteria (Leventer et al., 2010): (i) irregular surface of cortex; (ii) thickened or overfolded cortex aspect; and (iii) irregularity at the grey–white interface. However, as a large proportion of patients did not show all the criteria of typical polymicrogyria, we classified this as atypical polymicrogyria or polymicrogyria-like cortical dysplasia (Cushion et al., 2013). Polymicrogyria and polymicrogyria-like cortical dysplasia were further classified according to the pattern, distribution and topography of the abnormality. Where no specific pattern was identified, they were classified as focal (in association with irregular sulcal pattern but no clear topographic distribution) or multifocal (patchy in both hemispheres without any particular pattern or gradient). Lissencephaly was classified according to the severity (agyria to pachygyria and subcortical band heterotopia) and the maximal location (fronto-parietal i.e. anteriorly predominant and parieto-occipital or central i.e. posteriorly predominant) as described by Dobyns et al. (1999). Microlissencephaly was first reported by Norman et al. (1976) in a patient combining thick and lissencephalic cortex and severe microcephaly. Barkovich et al. (2012) defines microlissencephaly as congenital microcephaly with thick or relatively thick cortex. Here, microlissencephaly refers to severe microcephaly with biometric brain parameters below the third percentile without intrauterine growth retardation and lissencephaly with hemispheric surface completely smooth lacking primary fissures and olfactory sulci.

After this initial classification, the posterior fossa was reviewed with particular attention to the brainstem, the cerebellar vermis and the cerebellar hemispheres. In each case, ‘hypoplasia’ refers to small size but normal anatomy whereas ‘dysplasia’ refers to abnormal foliation or disorientation of fissures. Abnormalities of the corpus callosum were categorized as follows: agenesis (absence of the entire corpus callosum), hypogenesis (absence of the rostrum with small or absent splenium and small or absent inferior genu); hypoplasia (all parts formed but decreased thickness) or dysmorphic (abnormal shape of any part of the corpus callosum). Basal ganglia abnormalities were defined as either hypertrophy of the striatum or thalamus (when the volume is increased based on a consensus from both investigators), or dysmorphy with fusion between caudate and putamen (when the anterior horn of the internal capsule is not visible).

Results

Genotype–phenotype analysis

In this study, TUBA1A, TUBB2B, TUBB3, TUBB5 or TUBG1 mutations were identified in 74 sporadic patients with diffuse cortical malformations (13.1% of sporadic cases) and six familial cases with dominant segregation of the disorders (three families) (16.6% of familial cases). Details concerning distribution of these mutations and associated phenotypes are provided in Fig. 1 and Table 1. Briefly, 45 de novo TUBA1A missense mutations were identified in 27 sporadic living patients and 18 foetuses with malformations of cortical development. Of these, 17 are described here for the first time. Three hot spots were found, the most frequent being missense mutations affecting Arg264 (seven patients and one foetus), Arg402 (resulting in p.R402H in five patients and p.R402C in two patients) and Arg422 (resulting in p.R422H in two patients and p.R422C in one patient). These hot spot mutations represent 40% of all known TUBA1A mutations. Eighteen de novo TUBB2B missense mutations were identified in 12 sporadic living patients and six foetuses with cortical malformations. Among these mutations, 12 are novel and three recurrent mutations p.I202T, p.S172P and p.F265L were found in two patients, respectively. Eight TUBB3 missense mutations, including a novel one, were found in five sporadic patients and three families. One recurrent mutation, p.E205K was found in one family and one sporadic patient. The three patients with TUBB5 mutations and the three patients with TUBG1 mutations have been previously reported but were revisited for the purpose of the study. No mutation was found in the TUBA8 gene (Table 1 and Fig. 1).

Figure 1

Schematic representations of the functional domains of TUBA1A, TUBB2B and TUBB3 tubulin subunits and distribution of mutations associated with malformations of cortical development. Illustrated domains are the N-terminal that contains the guanine nucleotide-binding region, intermediate domain, and C-terminal domains that constitute the binding surface for microtubule-associated proteins and molecular motors such as kinesins and dyneins. In β-tubulin, they correspond to residues 1–229, 230–371, and 372–450, (Lowe et al., 2001) and in α-tubulin, to residues 1–205, 206–381, 382–451 (Nogales et al., 1998), respectively. Mutations associated with a lissencephaly (classic and with cerebellar hypoplasia) phenotype are indicated in red, with microlissencephaly in green, with central pachygyria in purple, central polymicrogyria-like cortical dysplasia in pink with generalized polymicrogyria-like cortical dysplasia in blue, with multifocal polymicrogyria and simplified gyral pattern in black. Grey markers indicate the distribution of the other reported TUBA1A, TUBB2B and TUBB3 mutations. For recurrent variations the number of occurrences is indicated in brackets.

Figure 1

Schematic representations of the functional domains of TUBA1A, TUBB2B and TUBB3 tubulin subunits and distribution of mutations associated with malformations of cortical development. Illustrated domains are the N-terminal that contains the guanine nucleotide-binding region, intermediate domain, and C-terminal domains that constitute the binding surface for microtubule-associated proteins and molecular motors such as kinesins and dyneins. In β-tubulin, they correspond to residues 1–229, 230–371, and 372–450, (Lowe et al., 2001) and in α-tubulin, to residues 1–205, 206–381, 382–451 (Nogales et al., 1998), respectively. Mutations associated with a lissencephaly (classic and with cerebellar hypoplasia) phenotype are indicated in red, with microlissencephaly in green, with central pachygyria in purple, central polymicrogyria-like cortical dysplasia in pink with generalized polymicrogyria-like cortical dysplasia in blue, with multifocal polymicrogyria and simplified gyral pattern in black. Grey markers indicate the distribution of the other reported TUBA1A, TUBB2B and TUBB3 mutations. For recurrent variations the number of occurrences is indicated in brackets.

Table 1

Overview of MRI characteristics in patients with tubulinopathies

Case ID Gene Gender Nucleotidic sequence variation Proteic sequence variation Inheritance Cortical dysgenesis (diagnosis group assignment) Corpus callosum Cerebellum Age or TOP Head size (SD) Tubulin mutations group/structural prediction Reference 
LIS_TUB_005_foetus01 TUBA1A c.959G > A p.R320H De novo MicroLIS p.ACC Sev_Ver_HS_Dys 25 GW  This series 
LIS_TUB_006_fœtus03 TUBA1A c.908T > G p.V303G De novo MicroLIS p.ACC Sev_Ver_HS_Dys 36 GW  Lecourtois et al., 2010 
LIS_TUB_007_fœtus04 TUBA1A c.856C > T p.L286F De novo MicroLIS c.ACC Mod_Ver_Hypo 23 GW  Fallet-Bianco et al., 2008 
LIS_TUB_004_fœtus08 TUBA1A c.978A > C p.K326N De novo MicroLIS c.ACC Sev_Ver_HS_Dys 23 GW  This series 
LIS_TUB_001_fœtus09 TUBA1A c.1285G > C p.E429Q De novo MicroLIS c.ACC Sev_Ver_Hypo 25 GW  This series 
LIS_TUB_008_ fœtus13 TUBB2B c.716G > T p.C239F De novo MicroLIS c.ACC Sev_Ver_HS_Dys 16 GW  This series 
LIS_TUB_010_ fœtus17 TUBB3 c.1162A > G p.M388V De novo MicroLIS c.ACC Sev_Ver_Hypo 27 GW  Poirier et al., 2010 
LIS_TUB_003_ fœtus18 TUBA1A c.167C > T p.T56M De novo MicroLIS c.ACC Sev_Ver_Hypo 24.3 GW  This series 
LIS_TUB_009_ fœtus19 TUBB2B c.745G > C p.D249H De novo MicroLIS c.ACC Sev_Ver_Hypo 27 GW  This series 
LIS_TUB_002_ fœtus20 TUBA1A c.790C > T p.R264H De novo MicroLIS c.ACC Sev_Ver_Hypo 24 GW  This series 
LIS_TUB_080_ fœtus24 TUBA1A c.1112T > A p.V371E De novo MicroLIS c.ACC Sev_Ver_Hypo 23.3 GW  This series 
LIS_TUB_079_ fœtus25 TUBA1A c.302A > G p.N101S De novo MicroLIS c.ACC Sev_Ver_HS_Dys 25 GW  This series 
LIS_TUB_017 fœtus02 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria c.ACC Mild_Ver_Hypo 29 GW  This series 
LIS_TUB_022_ fœtus05 TUBA1A c.712A > G p.I238V De novo LIS_sev_Agyria c.ACC Mod_Ver_Hypo 25 GW  Fallet-Bianco et al., 2008 
LIS_TUB_025 fœtus06 TUBA1A c.787C > A p.P263T De novo LCH_sev c.ACC Sev_Ver_Hypo 26 GW  Fallet-Bianco et al., 2008 
LIS_TUB_021_ fœtus07 TUBA1A c.1204C > T p.R402C De novo LIS_sev_Agyria dysmCC_HypoCC Mild_Ver_Hypo 35 GW  Fallet-Bianco et al., 2008 
LIS_TUB_018_ fœtus10 TUBA1A c.1265G > A p.R422H De novo LIS_sev_Agyria c.ACC Mild_Ver_Hypo 28 GW  This series 
LIS_TUB_013_ fœtus14 TUBB2B c.302G > A p.G98R De novo LIS_sev_Agyria c.ACC Normal 32.8 GW  This series 
LIS_TUB_012_ fœtus22 TUBA1A c.214 C > T p.P72S De novo LIS_sev_Agyria dysmCC_HypoCC Sev_Ver_Hypo 37.8 GW  This series 
LIS_TUB_011_ fœtus23 TUBA1A c.1226T > C p.V409A De novo LCH_sev c.ACC Sev_Ver_Hypo 32 GW  This series 
LIS_TUB_014 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Sev_HS_Dys 4 y −4 This series 
LIS_TUB_015 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Sev_HS_Dys 1 y −4 This series 
LIS_TUB_016 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Ver_Dys 1.5 y −3 This series 
LIS_TUB_019 TUBA1A c.1204C > T p.R402C De novo LIS_mod dysmCC_HypoCC Mild_Ver_Hypo 10 y −3 This series 
LIS_TUB_020 TUBA1A c.1265G > A p.R422H De novo LIS_mod dysmCC_HypoCC Mild_Ver_Dys 7 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_023 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Sev_HS_Dys 12 y −3 Keays et al., 2007; Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_024 TUBA1A c.1256C > T p.S419L De novo LIS_mod hypogeneticCC Normal 18 y −1 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_026 TUBA1A c.562 A > C p.I188L De novo LCH_mod_SBH p.ACC Sev_Ver_Hypo 2.5 y −4 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_027 TUBG1 c.991A > C p.T331P De novo LIS_mod_SBH dysmCC_HypoCC Normal 31 y −1  Poirier et al., 2013a 
LIS_TUB_028 TUBG1 c.1160T > C p.L387P De novo LIS_sev_Agyria dysmCC_HypoCC Normal 21 y −5.5  Poirier et al., 2013a 
LIS_TUB_029 TUBG1 c.275A > G p.Y92C De novo LIS_sev_Agyria dysmCC_HypoCC Normal 1.5 y −4  Poirier et al., 2013a 
LIS_TUB_043_foetus11 TUBA1A c.641G > A p.R214H De novo Central_PMG c.ACC Normal 23 GW  This series 
LIS_TUB_048_foetus16 TUBB2B c.742G > A p.A248T De novo Central_PMG Normal Mild_Ver_Hypo 28.5 GW  This series 
LIS_TUB_030 TUBA1A c.1105G > A p.A369T De novo Central_Pachy dysmCC_HypoCC Mild_Ver_Hypo 11 y −2 This series 
LIS_TUB_031 TUBA1A c.337G > A p.E113K De novo Central_Pachy Normal Normal 11 y −3 This series 
LIS_TUB_032 TUBA1A c.1225G > A p.V409I De novo Central_Pachy dysmCC_HypoCC Normal 10 y −2 This series 
LIS_TUB_033 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 1.5 y −5 This series 
LIS_TUB_034 TUBA1A c.790C > T p.R264C De novo Central_Pachy hypogeneticCC Normal 6.5 y −4 This series 
LIS_TUB_035 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 6 y −3 This series 
LIS_TUB_036 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 2 y −4.5 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_037 TUBA1A c.790C > T p.R264C De novo Central_Pachy hypogeneticCC Mild_Ver_Dys 4.5 y −4 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_038 TUBA1A c.1306G > T p.G436R De novo Central_Pachy dysmCC_HypoCC Normal 7 y −3 Bahi-Buisson et al., 2008 
LIS_TUB_039 TUBA1A c.1190T > C p.L397P De novo Central_Pachy hypogeneticCC Normal 5.5 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_040 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 1.5 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_041 TUBA1A c.790C > T p.R264C De novo Central_Pachy p.ACC Normal 7 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_042 TUBA1A c.1264C > T p.R422C De novo Central_Pachy dysmCC_HypoCC Mild_Ver_Dys 4.5 y −3 Bahi-Buisson et al., 2008 
LIS_TUB_044 TUBA1A c.367C > T p.R123C De novo Central_PMG Normal Ver_Dys 3 y −3 This series 
LIS_TUB_045 TUBA1A c.1168C > T p.R390C De novo Central_PMG dysmCC_HypoCC Ver_Dys 1 y −3 Poirier et al., 2013b 
LIS_TUB_046 TUBA1A c.703 G > T p.V235L De novo Central_PMG dysmCC_HypoCC Normal 7.5 y Poirier et al., 2013b 
LIS_TUB_047 TUBA1A c.481 T > C p.Y161H De novo Central_PMG dysmCC_HypoCC Ver_Dys 11 y − 3 Poirier et al., 2013b 
LIS_TUB_049 TUBB2B c.605T > C p.I202T De novo Central_PMG hypogeneticCC Mild_Ver_Hypo 10 y −2 This series 
LIS_TUB_050 TUBB2B c.605T > C p.I202T De novo Central_PMG hypogeneticCC Mild_Ver_Hypo 15 y This series 
LIS_TUB_051 TUBB2B c.511C > A p.P171T De novo Central_PMG p.ACC Sev_Ver_Dys 17 y −1 This series 
LIS_TUB_052 TUBB2B c.793T > C p.F265L De novo Central_PMG dysmCC_HypoCC Sev_Ver_Dys 1.5 y −3 This series 
LIS_TUB_078 TUBA1A c.1186G > T p.D396Y De novo Central_PMG p.ACC Sev_Ver_Dys 4 y −3 This series 
LIS_TUB_056_foetus12 TUBB2B c.514T > C p.S172P De novo Generalized PMG c.ACC Mild_Ver_Hypo 27 GW  Jaglin et al., 2009 
LIS_TUB_054_foetus15 TUBB2B c.518C > T p.P173L De novo Generalized PMG c.ACC Mild_Ver_Hypo 25 GW  This series 
LIS_TUB_053_foetus21 TUBA1A c.473C > T p.S158L De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 24.5 GW  This series 
LIS_TUB_055 TUBB2B c.514T > C p.S172P De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 3 y − 5 This series 
LIS_TUB_057 TUBB2B c.683T > C p.L228P De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 1 y −5 Jaglin et al., 2009 
LIS_TUB_058 TUBB3 c.533C > T p.T178M De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 5 y −4 Poirier et al., 2010 
LIS_TUB_059 TUBA1A c.1057G > A p.V353I De novo Simp_Gyr p.ACC Normal 4 y −2 This series 
LIS_TUB_060 TUBB2B c.1106G > T p.G369V De novo PMG_multi hypogeneticCC Ver_Dys 9 y −1 This series 
LIS_TUB_061 TUBB2B c.722G > A p.R241H De novo Simp_Gyr Normal Normal 12 y −1 This series 
LIS_TUB_062 TUBB2B c.832A > G p.S278G De novo Simp_Gyr hypogeneticCC Sev_Ver_Dys 0.9 y −2.5 This series 
LIS_TUB_063 TUBB3 c.136C > G p.R46G De novo PMG_multi dysmCC_HypoCC Ver_Dys 2 y −3 This series 
LIS_TUB_064 TUBB2B c.793C > T p.F265L De novo PMG_multi dysmCC_HypoCC Sev_HS_Dys 36 y −4 Jaglin et al., 2009 
LIS_TUB_065 TUBB2B c.629T > C p.I210T De novo PMG_multi hypogeneticCC Sev_HS_Dys 13 y −4 Jaglin et al., 2009 
LIS_TUB_066 TUBB2B c.935C > T p.T312M De novo PMG_multi dysmCC_HypoCC Ver_Dys 5 y −2 Jaglin et al., 2009 
LIS_TUB_067 TUBB3 c.633G > A p.E205K De novo PMG_multi dysmCC_HypoCC Ver_Dys 8 y −2 Poirier et al., 2010 
LIS_TUB_068 TUBB3 c.905C > T p.A302V Inherited Simp_Gyr hypogeneticCC Ver_Dys 3 y −2 Poirier et al., 2010 
LIS_TUB_069 TUBB3 c.633G > A p.E205K Inherited PMG_multi dysmCC_HypoCC Ver_Dys 9 y −1 Poirier et al., 2010 
LIS_TUB_070 TUBB3 c.905C > T p.A302V Inherited Simp_Gyr hypogeneticCC Ver_Dys 40 y −1 Poirier et al., 2010 
LIS_TUB_071 TUBB3 c.967A > G p.M323V Inherited Simp_Gyr hypogeneticCC Ver_Dys 4 y −0.5 Poirier et al., 2010 
LIS_TUB_072 TUBB3 c.633G > A p.E205K Inherited PMG_multi dysmCC_HypoCC Ver_Dys 10 y −0.5 Poirier et al., 2010 
LIS_TUB_073 TUBB3 c.244G > A p.G82R De novo PMG_multi p.ACC Ver_Dys 5 y Poirier et al., 2010 
LIS_TUB_074 TUBB3 c.967A > G p.M323V Inherited Simp_Gyr hypogeneticCC Ver_Dys 38 y Poirier et al., 2010 
LIS_TUB_075 TUBB5 c.1201G > A p.E401K De novo Simp_Gyr p.ACC Normal 3 y −4  Breuss et al., 2012 
LIS_TUB_076 TUBB5 c.1057G > A p.V353I De novo Simp_Gyr dysmCC_HypoCC Normal 4.8 y −4  Breuss et al., 2012 
LIS_TUB_077 TUBB5 c.895A > G p.M299V De novo PMG_multi hypogeneticCC Ver_Dys 3 y −2  Breuss et al., 2012 
Case ID Gene Gender Nucleotidic sequence variation Proteic sequence variation Inheritance Cortical dysgenesis (diagnosis group assignment) Corpus callosum Cerebellum Age or TOP Head size (SD) Tubulin mutations group/structural prediction Reference 
LIS_TUB_005_foetus01 TUBA1A c.959G > A p.R320H De novo MicroLIS p.ACC Sev_Ver_HS_Dys 25 GW  This series 
LIS_TUB_006_fœtus03 TUBA1A c.908T > G p.V303G De novo MicroLIS p.ACC Sev_Ver_HS_Dys 36 GW  Lecourtois et al., 2010 
LIS_TUB_007_fœtus04 TUBA1A c.856C > T p.L286F De novo MicroLIS c.ACC Mod_Ver_Hypo 23 GW  Fallet-Bianco et al., 2008 
LIS_TUB_004_fœtus08 TUBA1A c.978A > C p.K326N De novo MicroLIS c.ACC Sev_Ver_HS_Dys 23 GW  This series 
LIS_TUB_001_fœtus09 TUBA1A c.1285G > C p.E429Q De novo MicroLIS c.ACC Sev_Ver_Hypo 25 GW  This series 
LIS_TUB_008_ fœtus13 TUBB2B c.716G > T p.C239F De novo MicroLIS c.ACC Sev_Ver_HS_Dys 16 GW  This series 
LIS_TUB_010_ fœtus17 TUBB3 c.1162A > G p.M388V De novo MicroLIS c.ACC Sev_Ver_Hypo 27 GW  Poirier et al., 2010 
LIS_TUB_003_ fœtus18 TUBA1A c.167C > T p.T56M De novo MicroLIS c.ACC Sev_Ver_Hypo 24.3 GW  This series 
LIS_TUB_009_ fœtus19 TUBB2B c.745G > C p.D249H De novo MicroLIS c.ACC Sev_Ver_Hypo 27 GW  This series 
LIS_TUB_002_ fœtus20 TUBA1A c.790C > T p.R264H De novo MicroLIS c.ACC Sev_Ver_Hypo 24 GW  This series 
LIS_TUB_080_ fœtus24 TUBA1A c.1112T > A p.V371E De novo MicroLIS c.ACC Sev_Ver_Hypo 23.3 GW  This series 
LIS_TUB_079_ fœtus25 TUBA1A c.302A > G p.N101S De novo MicroLIS c.ACC Sev_Ver_HS_Dys 25 GW  This series 
LIS_TUB_017 fœtus02 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria c.ACC Mild_Ver_Hypo 29 GW  This series 
LIS_TUB_022_ fœtus05 TUBA1A c.712A > G p.I238V De novo LIS_sev_Agyria c.ACC Mod_Ver_Hypo 25 GW  Fallet-Bianco et al., 2008 
LIS_TUB_025 fœtus06 TUBA1A c.787C > A p.P263T De novo LCH_sev c.ACC Sev_Ver_Hypo 26 GW  Fallet-Bianco et al., 2008 
LIS_TUB_021_ fœtus07 TUBA1A c.1204C > T p.R402C De novo LIS_sev_Agyria dysmCC_HypoCC Mild_Ver_Hypo 35 GW  Fallet-Bianco et al., 2008 
LIS_TUB_018_ fœtus10 TUBA1A c.1265G > A p.R422H De novo LIS_sev_Agyria c.ACC Mild_Ver_Hypo 28 GW  This series 
LIS_TUB_013_ fœtus14 TUBB2B c.302G > A p.G98R De novo LIS_sev_Agyria c.ACC Normal 32.8 GW  This series 
LIS_TUB_012_ fœtus22 TUBA1A c.214 C > T p.P72S De novo LIS_sev_Agyria dysmCC_HypoCC Sev_Ver_Hypo 37.8 GW  This series 
LIS_TUB_011_ fœtus23 TUBA1A c.1226T > C p.V409A De novo LCH_sev c.ACC Sev_Ver_Hypo 32 GW  This series 
LIS_TUB_014 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Sev_HS_Dys 4 y −4 This series 
LIS_TUB_015 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Sev_HS_Dys 1 y −4 This series 
LIS_TUB_016 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Ver_Dys 1.5 y −3 This series 
LIS_TUB_019 TUBA1A c.1204C > T p.R402C De novo LIS_mod dysmCC_HypoCC Mild_Ver_Hypo 10 y −3 This series 
LIS_TUB_020 TUBA1A c.1265G > A p.R422H De novo LIS_mod dysmCC_HypoCC Mild_Ver_Dys 7 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_023 TUBA1A c.1205G > A p.R402H De novo LIS_sev_Agyria hypogeneticCC Sev_HS_Dys 12 y −3 Keays et al., 2007; Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_024 TUBA1A c.1256C > T p.S419L De novo LIS_mod hypogeneticCC Normal 18 y −1 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_026 TUBA1A c.562 A > C p.I188L De novo LCH_mod_SBH p.ACC Sev_Ver_Hypo 2.5 y −4 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_027 TUBG1 c.991A > C p.T331P De novo LIS_mod_SBH dysmCC_HypoCC Normal 31 y −1  Poirier et al., 2013a 
LIS_TUB_028 TUBG1 c.1160T > C p.L387P De novo LIS_sev_Agyria dysmCC_HypoCC Normal 21 y −5.5  Poirier et al., 2013a 
LIS_TUB_029 TUBG1 c.275A > G p.Y92C De novo LIS_sev_Agyria dysmCC_HypoCC Normal 1.5 y −4  Poirier et al., 2013a 
LIS_TUB_043_foetus11 TUBA1A c.641G > A p.R214H De novo Central_PMG c.ACC Normal 23 GW  This series 
LIS_TUB_048_foetus16 TUBB2B c.742G > A p.A248T De novo Central_PMG Normal Mild_Ver_Hypo 28.5 GW  This series 
LIS_TUB_030 TUBA1A c.1105G > A p.A369T De novo Central_Pachy dysmCC_HypoCC Mild_Ver_Hypo 11 y −2 This series 
LIS_TUB_031 TUBA1A c.337G > A p.E113K De novo Central_Pachy Normal Normal 11 y −3 This series 
LIS_TUB_032 TUBA1A c.1225G > A p.V409I De novo Central_Pachy dysmCC_HypoCC Normal 10 y −2 This series 
LIS_TUB_033 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 1.5 y −5 This series 
LIS_TUB_034 TUBA1A c.790C > T p.R264C De novo Central_Pachy hypogeneticCC Normal 6.5 y −4 This series 
LIS_TUB_035 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 6 y −3 This series 
LIS_TUB_036 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 2 y −4.5 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_037 TUBA1A c.790C > T p.R264C De novo Central_Pachy hypogeneticCC Mild_Ver_Dys 4.5 y −4 Poirier et al., 2007; Bahi-Buisson et al., 2008 
LIS_TUB_038 TUBA1A c.1306G > T p.G436R De novo Central_Pachy dysmCC_HypoCC Normal 7 y −3 Bahi-Buisson et al., 2008 
LIS_TUB_039 TUBA1A c.1190T > C p.L397P De novo Central_Pachy hypogeneticCC Normal 5.5 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_040 TUBA1A c.790C > T p.R264C De novo Central_Pachy Normal Mild_Ver_Hypo 1.5 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_041 TUBA1A c.790C > T p.R264C De novo Central_Pachy p.ACC Normal 7 y −4 Bahi-Buisson et al., 2008 
LIS_TUB_042 TUBA1A c.1264C > T p.R422C De novo Central_Pachy dysmCC_HypoCC Mild_Ver_Dys 4.5 y −3 Bahi-Buisson et al., 2008 
LIS_TUB_044 TUBA1A c.367C > T p.R123C De novo Central_PMG Normal Ver_Dys 3 y −3 This series 
LIS_TUB_045 TUBA1A c.1168C > T p.R390C De novo Central_PMG dysmCC_HypoCC Ver_Dys 1 y −3 Poirier et al., 2013b 
LIS_TUB_046 TUBA1A c.703 G > T p.V235L De novo Central_PMG dysmCC_HypoCC Normal 7.5 y Poirier et al., 2013b 
LIS_TUB_047 TUBA1A c.481 T > C p.Y161H De novo Central_PMG dysmCC_HypoCC Ver_Dys 11 y − 3 Poirier et al., 2013b 
LIS_TUB_049 TUBB2B c.605T > C p.I202T De novo Central_PMG hypogeneticCC Mild_Ver_Hypo 10 y −2 This series 
LIS_TUB_050 TUBB2B c.605T > C p.I202T De novo Central_PMG hypogeneticCC Mild_Ver_Hypo 15 y This series 
LIS_TUB_051 TUBB2B c.511C > A p.P171T De novo Central_PMG p.ACC Sev_Ver_Dys 17 y −1 This series 
LIS_TUB_052 TUBB2B c.793T > C p.F265L De novo Central_PMG dysmCC_HypoCC Sev_Ver_Dys 1.5 y −3 This series 
LIS_TUB_078 TUBA1A c.1186G > T p.D396Y De novo Central_PMG p.ACC Sev_Ver_Dys 4 y −3 This series 
LIS_TUB_056_foetus12 TUBB2B c.514T > C p.S172P De novo Generalized PMG c.ACC Mild_Ver_Hypo 27 GW  Jaglin et al., 2009 
LIS_TUB_054_foetus15 TUBB2B c.518C > T p.P173L De novo Generalized PMG c.ACC Mild_Ver_Hypo 25 GW  This series 
LIS_TUB_053_foetus21 TUBA1A c.473C > T p.S158L De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 24.5 GW  This series 
LIS_TUB_055 TUBB2B c.514T > C p.S172P De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 3 y − 5 This series 
LIS_TUB_057 TUBB2B c.683T > C p.L228P De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 1 y −5 Jaglin et al., 2009 
LIS_TUB_058 TUBB3 c.533C > T p.T178M De novo Generalized PMG c.ACC Sev_Ver_HS_Dys 5 y −4 Poirier et al., 2010 
LIS_TUB_059 TUBA1A c.1057G > A p.V353I De novo Simp_Gyr p.ACC Normal 4 y −2 This series 
LIS_TUB_060 TUBB2B c.1106G > T p.G369V De novo PMG_multi hypogeneticCC Ver_Dys 9 y −1 This series 
LIS_TUB_061 TUBB2B c.722G > A p.R241H De novo Simp_Gyr Normal Normal 12 y −1 This series 
LIS_TUB_062 TUBB2B c.832A > G p.S278G De novo Simp_Gyr hypogeneticCC Sev_Ver_Dys 0.9 y −2.5 This series 
LIS_TUB_063 TUBB3 c.136C > G p.R46G De novo PMG_multi dysmCC_HypoCC Ver_Dys 2 y −3 This series 
LIS_TUB_064 TUBB2B c.793C > T p.F265L De novo PMG_multi dysmCC_HypoCC Sev_HS_Dys 36 y −4 Jaglin et al., 2009 
LIS_TUB_065 TUBB2B c.629T > C p.I210T De novo PMG_multi hypogeneticCC Sev_HS_Dys 13 y −4 Jaglin et al., 2009 
LIS_TUB_066 TUBB2B c.935C > T p.T312M De novo PMG_multi dysmCC_HypoCC Ver_Dys 5 y −2 Jaglin et al., 2009 
LIS_TUB_067 TUBB3 c.633G > A p.E205K De novo PMG_multi dysmCC_HypoCC Ver_Dys 8 y −2 Poirier et al., 2010 
LIS_TUB_068 TUBB3 c.905C > T p.A302V Inherited Simp_Gyr hypogeneticCC Ver_Dys 3 y −2 Poirier et al., 2010 
LIS_TUB_069 TUBB3 c.633G > A p.E205K Inherited PMG_multi dysmCC_HypoCC Ver_Dys 9 y −1 Poirier et al., 2010 
LIS_TUB_070 TUBB3 c.905C > T p.A302V Inherited Simp_Gyr hypogeneticCC Ver_Dys 40 y −1 Poirier et al., 2010 
LIS_TUB_071 TUBB3 c.967A > G p.M323V Inherited Simp_Gyr hypogeneticCC Ver_Dys 4 y −0.5 Poirier et al., 2010 
LIS_TUB_072 TUBB3 c.633G > A p.E205K Inherited PMG_multi dysmCC_HypoCC Ver_Dys 10 y −0.5 Poirier et al., 2010 
LIS_TUB_073 TUBB3 c.244G > A p.G82R De novo PMG_multi p.ACC Ver_Dys 5 y Poirier et al., 2010 
LIS_TUB_074 TUBB3 c.967A > G p.M323V Inherited Simp_Gyr hypogeneticCC Ver_Dys 38 y Poirier et al., 2010 
LIS_TUB_075 TUBB5 c.1201G > A p.E401K De novo Simp_Gyr p.ACC Normal 3 y −4  Breuss et al., 2012 
LIS_TUB_076 TUBB5 c.1057G > A p.V353I De novo Simp_Gyr dysmCC_HypoCC Normal 4.8 y −4  Breuss et al., 2012 
LIS_TUB_077 TUBB5 c.895A > G p.M299V De novo PMG_multi hypogeneticCC Ver_Dys 3 y −2  Breuss et al., 2012 

In bold are the mutations newly described here.

M = males; F = females; y = years; GW = gestation weeks; TOP = termination of pregnancy; CC = corpus callosum; c.ACC = complete corpus callosum agenesis; p.ACC = partial corpus callosum agenesis; mod = moderate; sev = severe; Ver = vermian; Hypo = hypoplasia; Dys = dysplasia; dysm = dysmorphic; N/A = not available; N = normal. Head size is presented in SD (standard deviation).

Diagnosis group assignment allowed us to define five subgroups: (i) MicroLIS = microlissencephaly; (ii) LIS = lissencephaly, either severe (LIS_sev) or moderate (LIS_mod) with or without subcortical band heterotpia (SBH); LCH = lissencephaly with cerebellar hypoplasia; (iii) Central-PMG = central polymicrogyria-like cortical dysplasia and Central-Pachy = central pachygyria; (iv) Generalized PMG = generalized polymicrogyria and PMG_multi = multifocal polymicrogyria; and (v) Symp_Gyr = simplified gyral pattern with focal polymicrogyria.

Tubulin mutations groups were assigned according to structural modelization and biochemical data, in five subgroups, Group A: GTP binding pocket, nucleotide binding is essential for protein folding, heterodimer stability, and microtubule dynamics; Group B: longitudinal interactions; Group C: lateral protofilament interactions (residues involved in longitudinal and lateral interactions are necessary for both heterodimer and microtubule stability; Group D: microtubule-associated protein and motor protein interactions; and Group E: tubulin folding.

Looking back at their brain MRIs we identified five major subtypes of malformations detailed below: microlissencephaly in 12 patients, agyria-pachygyria (lissencephaly) in 19, central pachygyria and polymicrogyria-like cortical dysplasia in 24, diffuse polymicrogyria-like cortical dysplasia in six, and finally a ‘simplified’ gyral pattern with area of focal polymicrogyria in 19. Clinical and brain imaging features are summarized in Table 1. Representative MRI images are shown in Figs 2–8.

Figure 2

Lissencephaly. In this series of brain magnetic resonance illustrations, multiple images from the same patient are shown: Patient LIS-TUB_015 age 6 months (A–C), Patient LIS-TUB_016 age 4 months (D–F), and Patient LIS-TUB_029 age 4 years (G–I). For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). This figure shows three cases with a lissencephaly pattern with the recurrent p.R402H TUBA1A mutation (A–C) and (D–F) and one with p.Y92C TUBG1 mutation. All have agyria-pachygyria consistent with classic lissencephaly with a posterior more severe than anterior gradient (arrowheads). Other features include a dysmorphic corpus callosum (A, D and G) and non-visible internal capsules (white stars). Both patients with p.R402H TUBA1A mutations (A and D) also had moderate (A) to severe (D–F) vermian hypoplasia (arrow).

Figure 2

Lissencephaly. In this series of brain magnetic resonance illustrations, multiple images from the same patient are shown: Patient LIS-TUB_015 age 6 months (A–C), Patient LIS-TUB_016 age 4 months (D–F), and Patient LIS-TUB_029 age 4 years (G–I). For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). This figure shows three cases with a lissencephaly pattern with the recurrent p.R402H TUBA1A mutation (A–C) and (D–F) and one with p.Y92C TUBG1 mutation. All have agyria-pachygyria consistent with classic lissencephaly with a posterior more severe than anterior gradient (arrowheads). Other features include a dysmorphic corpus callosum (A, D and G) and non-visible internal capsules (white stars). Both patients with p.R402H TUBA1A mutations (A and D) also had moderate (A) to severe (D–F) vermian hypoplasia (arrow).

Figure 3

‘Central’ pachygyria with p.R264C TUBA1A mutations. This figure shows three patients with ‘central’ pachygyria and TUBA1A missense R264C mutations. Patient LIS_TUB-035 age 6 years (A–C), Patient LIS_TUB-034 age 6 years (D–F) and Patient LIS_TUB-033 age 13 months (G–I). For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). All have pachygyria that appears mildly asymmetric and most severe over the central regions rather than over the posterior pole (white arrowheads). The basal ganglia are malformed and dysmorphic appearing as round structures in which the caudate, putamen and globus pallidus cannot be distinguished (white stars). Associated malformations include partial agenesis of the rostrum and the splenium (D), hypoplastic thin corpus callosum, thin brainstem with a flat pons (G), and mild (A–D) vermian hypoplasia with dysplasia (F) (arrow).

Figure 3

‘Central’ pachygyria with p.R264C TUBA1A mutations. This figure shows three patients with ‘central’ pachygyria and TUBA1A missense R264C mutations. Patient LIS_TUB-035 age 6 years (A–C), Patient LIS_TUB-034 age 6 years (D–F) and Patient LIS_TUB-033 age 13 months (G–I). For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). All have pachygyria that appears mildly asymmetric and most severe over the central regions rather than over the posterior pole (white arrowheads). The basal ganglia are malformed and dysmorphic appearing as round structures in which the caudate, putamen and globus pallidus cannot be distinguished (white stars). Associated malformations include partial agenesis of the rostrum and the splenium (D), hypoplastic thin corpus callosum, thin brainstem with a flat pons (G), and mild (A–D) vermian hypoplasia with dysplasia (F) (arrow).

Figure 4

‘Central’ pachygyria. This figure shows three patients with ‘central’ pachygyria and TUBA1A missense mutations: Patient LIS-TUB_032 age 4.5 years (p.V409I, A–C), Patient LIS-TUB_042 age 3 years (p.R422C, D–F) and Patient LIS-TUB_031 age 10 years (p.E113K, G–I). All have pachygyria that appears mildly asymmetrical (E) and most severe over the central regions (white arrowheads). The basal ganglia are dysmorphic, appearing as round structures in which the caudate, putamen and globus pallidus cannot be distinguished (E and H) (white stars). Associated malformations include partial dysmorphism or thinning (G) of the corpus callosum, and vermian dysplasia (C, F and I) (white arrows).

Figure 4

‘Central’ pachygyria. This figure shows three patients with ‘central’ pachygyria and TUBA1A missense mutations: Patient LIS-TUB_032 age 4.5 years (p.V409I, A–C), Patient LIS-TUB_042 age 3 years (p.R422C, D–F) and Patient LIS-TUB_031 age 10 years (p.E113K, G–I). All have pachygyria that appears mildly asymmetrical (E) and most severe over the central regions (white arrowheads). The basal ganglia are dysmorphic, appearing as round structures in which the caudate, putamen and globus pallidus cannot be distinguished (E and H) (white stars). Associated malformations include partial dysmorphism or thinning (G) of the corpus callosum, and vermian dysplasia (C, F and I) (white arrows).

Figure 5

‘Central’ polymicrogyria-like cortical dysplasia. This figure shows three patients with central polymicrogyria-like cortical dysplasia, two with TUBB2B mutations: Patient LIS-TUB_050 age 15 years (A–C), Patient LIS-TUB_046 age 6 years (D–F) and one with TUBA1A mutation: Patient LIS-TUB_051 age 16 years (G–I). For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). All have polymicrogyria that appear typical with an irregular surface of cortex or overfolded cortex aspect and irregularity at the grey-white interface (B and E) or atypical with a coarse aspect (H). The polymicrogyria appear mildly asymmetrical and most severe over the central (mid- and posterior frontal, perisylvian and anterior parietal) regions rather than over the posterior pole (white arrowheads). The frontal horns of both lateral ventricles are dysmorphic with malformed basal ganglia (mostly in E and H) (white stars). The corpus callosum is hypogenetic (D) or thin (G). The superior vermis is dysplastic in all cases (C, F and I) (white arrows).

Figure 5

‘Central’ polymicrogyria-like cortical dysplasia. This figure shows three patients with central polymicrogyria-like cortical dysplasia, two with TUBB2B mutations: Patient LIS-TUB_050 age 15 years (A–C), Patient LIS-TUB_046 age 6 years (D–F) and one with TUBA1A mutation: Patient LIS-TUB_051 age 16 years (G–I). For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). All have polymicrogyria that appear typical with an irregular surface of cortex or overfolded cortex aspect and irregularity at the grey-white interface (B and E) or atypical with a coarse aspect (H). The polymicrogyria appear mildly asymmetrical and most severe over the central (mid- and posterior frontal, perisylvian and anterior parietal) regions rather than over the posterior pole (white arrowheads). The frontal horns of both lateral ventricles are dysmorphic with malformed basal ganglia (mostly in E and H) (white stars). The corpus callosum is hypogenetic (D) or thin (G). The superior vermis is dysplastic in all cases (C, F and I) (white arrows).

Figure 6

‘Central’ polymicrogyria-like cortical dysplasia in a poorly myelinated brain. This figure shows two patients with centrally predominant polymicrogyria-like cortical dysplasia with TUBB2B and TUBA1A mutations. Each row shows multiple images from the same patient; Patient LIS-TUB_052 (A–C), Patient LIS-TUB_044 (D–F) age 4 and 13 months, respectively. For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). All cases show an abnormally thick sylvian and perisylvian cortex with a bumpy appearance at the surface and at the cortical-white matter junction. The frontal horns of both lateral ventricles are dysmorphic with hypertrophic and malformed basal ganglia (white stars). The corpus callosum is dysmorphic (A) or thin (C). The superior vermis is dysplastic in both cases (C and F) (white arrows).

Figure 6

‘Central’ polymicrogyria-like cortical dysplasia in a poorly myelinated brain. This figure shows two patients with centrally predominant polymicrogyria-like cortical dysplasia with TUBB2B and TUBA1A mutations. Each row shows multiple images from the same patient; Patient LIS-TUB_052 (A–C), Patient LIS-TUB_044 (D–F) age 4 and 13 months, respectively. For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C, F and I). All cases show an abnormally thick sylvian and perisylvian cortex with a bumpy appearance at the surface and at the cortical-white matter junction. The frontal horns of both lateral ventricles are dysmorphic with hypertrophic and malformed basal ganglia (white stars). The corpus callosum is dysmorphic (A) or thin (C). The superior vermis is dysplastic in both cases (C and F) (white arrows).

Figure 7

Generalized polymicrogyria-like cortical dysplasia. This figure shows three patients with generalized polymicrogyria-like cortical dysplasia with TUBB3 mutations (A–C) and TUBB2B (D–F and G–I). Each row shows multiple images from the same patient: Patient LIS-TUB_058 (A–C), Patient LIS-TUB_055 (D–F) and Patient LIS-TUB_057 age 3 years, 3 years and 4 months, respectively. For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C and F); where images were not available, coronal images through the cerebellum are shown instead (I). In two cases (B and E) the cortex shows a coarse appearance, with excessively folded gyri (arrowheads). In the last case (H), (the youngest patient aged 4 months), the cortical surface is irregular with undersulcation and a variably thick polymicrogyria-like cortex (arrowheads). The polymicrogyria is extensive in 2/3 cases (B and H) with only few subnormal sulci or generalized without any areas of normal cortex (E) (white arrowheads) and associated with complete agenesis of corpus callosum, dysmorphic basal ganglia (white stars) and vermian cerebellar dysplasia (white arrows).

Figure 7

Generalized polymicrogyria-like cortical dysplasia. This figure shows three patients with generalized polymicrogyria-like cortical dysplasia with TUBB3 mutations (A–C) and TUBB2B (D–F and G–I). Each row shows multiple images from the same patient: Patient LIS-TUB_058 (A–C), Patient LIS-TUB_055 (D–F) and Patient LIS-TUB_057 age 3 years, 3 years and 4 months, respectively. For each patient, we selected midline sagittal images (A, D and G), axial images through the deep nuclei (B, E and H) and cerebellar vermis (C and F); where images were not available, coronal images through the cerebellum are shown instead (I). In two cases (B and E) the cortex shows a coarse appearance, with excessively folded gyri (arrowheads). In the last case (H), (the youngest patient aged 4 months), the cortical surface is irregular with undersulcation and a variably thick polymicrogyria-like cortex (arrowheads). The polymicrogyria is extensive in 2/3 cases (B and H) with only few subnormal sulci or generalized without any areas of normal cortex (E) (white arrowheads) and associated with complete agenesis of corpus callosum, dysmorphic basal ganglia (white stars) and vermian cerebellar dysplasia (white arrows).

Figure 8

Gyral disorganization with multifocal dysplasia. This figure shows four patients with generalized polymicrogyria with TUBB3 (Patient LIS-TUB_068 TUBB3 age 3 years A–C, Patient LIS-TUB_063 TUBB3 age 11 years D–F and Patient LIS-TUB_073 age 3 years G–I, respectively) or TUBB5 mutations (Patient LIS-TUB_077 age 2 years, J–L). For each patient we selected midline sagittal (A, D, G and J) images, axial images through the deep nuclei (B, E, H and K) and cerebellar vermis (C, F, I and L). In all cases, the cortex displays an abnormal appearance with poorly folded gyri, generalized undersulcation (mostly the ternary sulci), with focal variably thick aspect reminiscent of focal dysplasia or atypical polymicrogyria (white arrows). In all cases also, the basal ganglia are malformed and the lateral ventricles are dysmorphic (white stars), and the vermis is dysplasic (white arrows). The corpus callosum is hypogenetic (A–G) or dysmorphic (J).

Figure 8

Gyral disorganization with multifocal dysplasia. This figure shows four patients with generalized polymicrogyria with TUBB3 (Patient LIS-TUB_068 TUBB3 age 3 years A–C, Patient LIS-TUB_063 TUBB3 age 11 years D–F and Patient LIS-TUB_073 age 3 years G–I, respectively) or TUBB5 mutations (Patient LIS-TUB_077 age 2 years, J–L). For each patient we selected midline sagittal (A, D, G and J) images, axial images through the deep nuclei (B, E, H and K) and cerebellar vermis (C, F, I and L). In all cases, the cortex displays an abnormal appearance with poorly folded gyri, generalized undersulcation (mostly the ternary sulci), with focal variably thick aspect reminiscent of focal dysplasia or atypical polymicrogyria (white arrows). In all cases also, the basal ganglia are malformed and the lateral ventricles are dysmorphic (white stars), and the vermis is dysplasic (white arrows). The corpus callosum is hypogenetic (A–G) or dysmorphic (J).

Microlissencephaly

This pattern of severe microcephaly without intrauterine growth retardation and lissencephaly with hemispheres lacking primary fissures and olfactory sulci represents the most severe end of the spectrum of tubulinopathies. It is identified exclusively in foetal cases (eight males, four females). All foetuses had a neuropathological examination—two previously reported (Lecourtois et al., 2010; Poirier et al., 2010). In all cases, the pregnancy was terminated for medical reasons at median gestation of 25 weeks (range 16–36) after a prenatal diagnosis of microcephaly, corpus callosum agenesis and severe brain dysgenesis.

In all cases, brain MRI demonstrated an absent to poorly operculized cortex and virtually no visible gyration together with severe cerebellar hypoplasia and complete (10/12; 83.3%) or partial (2/12; 16.7%) agenesis of the corpus callosum. In this subtype, the basal ganglia were either not visible, virtually absent (5/11; 45.5%) or severely hypoplastic (6/11; 54.5%). Associated malformations included a very thin brainstem with flattening of the pons in all cases.

In patients with microlissencephaly, tubulin mutations included mainly the TUBA1A gene (p.T56M, p.N101S, p.R264H, p.L286F, p.V303G, p.R320H, p.K326N, p.V371E, and p.E429Q), but also TUBB2B (p.D249H, p.C239F) and TUBB3 (p.M388V) in one case. None of these mutations have been found in other tubulinopathies to date. More interestingly, they represent the first dominant aetiology of microlissencephaly, as other known genetic causes, including mutations in NDE1 or WDR62, are of autosomal recessive inheritance (Bilguvar et al., 2010).

Lissencephaly

This pattern was found in 11 living patients and eight foetuses (11 males, eight females). Six out of eight foetuses had a neuropathological examination—three previously reported (Fallet-Bianco et al., 2008), whereas in the remaining two the diagnosis was made on foetal MRI as foetopathological analysis was declined by the parents. In all cases, the pregnancy was terminated for medical reasons at a mean of 30.7 weeks gestation after a prenatal diagnosis of severe brain dysgenesis on ultrasound and MRI.

Two distinct patterns of malformation were observed; lissencephaly with or without moderate to severe cerebellar hypoplasia and dysplasia. The former condition is reminiscent of lissencephaly with cerebellar hypoplasia previously reported (Ross et al., 2001).

Agyria-pachygyria: Fourteen patients (including eight foetuses) had severe lissencephaly corresponding to LIS grade 1–2 with thick cortex with a posterior to anterior gradient of severity and five had moderate lissencephaly corresponding to LIS grade 3–4 (including two with subcortical band heterotopia) or moderate lissencephaly (with subcortical band heterotopia in Patient LIS_TUB_026). In this pattern, the basal ganglia appeared roughly normal in the majority of cases (14/19; 73.7%), although the anterior limb of the internal capsule was not visible. In this subtype, the malformation was always bilateral and showed a generally symmetrical distribution between hemispheres (Fig. 2). The lateral ventricles were often severely enlarged but were regularly shaped and symmetrical, in contrast with perisylvian pachygyria and polymicrogyria-like cortical dysplasia.

Cerebellar and brainstem abnormalities: Malformations of the cerebellum were frequent in this subtype, ranging from severe vermian dysplasia (8/19; 42.1%) with preservation of the hemispheres, to milder vermian hypoplasia (6/19; 31.6%). Five had a normal cerebellum. The brainstem was frequently mildly to moderately hypoplastic (9/19; 47.4%) or severely hypoplastic with a flat pons (5/19; 26.3%) (Fig. 2).

Callosal abnormalities: Malformations of the corpus callosum mainly included a dysmorphic or hypoplastic corpus callosum (5/11; 45.5%) or hypogenesis or partial agenesis of the corpus callosum (6/11; 54.5%) in living patients, whereas all the eight foetal cases showed complete agenesis.

At last follow-up, the 11 living patients of this group were between 1 and 31 years of age (mean age: 9.95 years; median 7). In comparison to the patients with perisylvian pachygyria/polymicrogyria-like and simplified gyral subtypes, patients with the lissencephaly subtype tended to have more severe delay in both cognitive and motor development. All showed severe motor and speech delay, were either wheelchair-bound with no voluntary motor control (4/11; 36.4%) or had head control only (3/11; 27.3%), had no eye contact and no speech development, although three patients were able to walk unaided. All of the patients had severe epilepsy, which typically had an earlier onset and greater severity compared with the other subtypes. All had an early epileptic encephalopathy with infantile spasms, with onset from the neonatal period to 6 months of age (median 1 month). Several different seizure types were observed, with most patients having infantile spasms and tonic seizures except two who also experienced myoclonic seizures and absences. Seizures were refractory in all cases except one (Patient LIS_TUB_026 who showed the milder form of lissencephaly combined with subcortical band heterotopia, and presented early onset infantile spasms that were controlled with oral steroids and valproate). Most patients in this group had congenital microcephaly (9/11; 81.8%) with a head circumference ∼3.2 standard deviations (SD) below the mean at birth. All had severe axial hypotonia, or mixed central hypotonia with limb spasticity. Of note, two also had facial diplegia and strabismus suggestive of pseudobulbar palsy.

At the genetic level, most patients (15/19, 78.9%) with this subtype carried mutations in TUBA1A mutations, whereas one foetus had a TUBB2B mutation (p.G58R). The other three cases had TUBG1 mutations, with a phenotype almost indistinguishable from the others. Of note, however, these three patients had a normal brainstem and cerebellum. As previously reported (Bahi-Buisson et al., 2008; Kumar et al., 2010), the major TUBA1A mutation responsible for this subtype affects the Arg402 residue leading to the recurrent mutation p.R402H (in 5/15) or p.R402C (in 2/15). Another recurrent mutation p.R422H was found in 2/15 patients.

Central pachygyria and polymicrogyria-like cortical dysplasia

This subtype corresponds to a pattern of malformation that is clearly distinguishable by the three following characteristic imaging findings that define this group, comprising 22 living foetuses and two foetuses.

Bilateral ‘central’ cortical malformation more severe in posterior frontal, perisylvian and parietal regions and sparing the occipital cortex: In 13 cases (10 males, three females), this cortical malformation was reminiscent of pachygyria with a thick cortex and cortical infolding (Figs 3 and 4). In these cases, the cortex seemed thick whatever the age of MRI, excluding the misleading pattern of polymicrogyria due to immature myelination (Takanashi and Barkovich, 2003). In 11 other patients (including two foetuses), the malformation was more suggestive of polymicrogyria, although all the characteristics of this were not always present. These cases were defined as polymicrogyria-like cortical dysplasia (Fig. 5). Both pachygyria- and polymicrogyria-like cortical dysplasia were both bilateral in all patients, although asymmetrical in the majority. The lateral ventricles were often dysmorphic, especially anteriorly and were also asymmetrical in most patients.

Basal ganglia dysmorphy: All patients had malformed basal ganglia appearing as large round structures in which the caudate, putamen and globus pallidus could not be distinguished. Lack of clear separation between the caudate and putamen and bilateral hypoplasia of the anterior limbs of the internal capsule was also a major feature (Figs 3–5).

Cerebellar vermis and brainstem abnormalities: Vermian malformations were common while cerebellar hemispheres were always normal. This included mild vermian hypoplasia (9/24; 37.5%) or dysplasia (5/24; 20.8%), and less frequently severe vermian hypoplasia (3/24; 12.5%). The brainstem was commonly mildly hypoplastic with a flat pons (19/24; 79.2%) (Figs 3–5).

Callosal abnormalities: Malformations of the corpus callosum mainly consisted of a dysmorphic thin (7/24; 29.2%) or hypogenetic (5/24; 20.8%) corpus callosum, and less frequently partial (3/24; 12.5%) or complete (1/24; 4.2%) corpus callosum agenesis. Of note, the patient with corpus callosum agenesis was one of the two foetal cases whereas the others had a normal corpus callosum.

For the 22 living patients (of 24 cases) including 14 males and eight females, age at the time of last follow-up ranged from 1 to 17 years (mean 6.72 years; median 6.25 years). Moderate to severe motor delay was observed in all children. All patients had delayed sitting and standing in the first year, although all children older than 2 years were cruising or walking. Motor function in children older than 3 years ranged from assisted walking with a combination of a spastic and ataxic gait to independent toe walking. All children older than 3 years of age had severe speech delay, with a combination of dysarthric speech or expressive language impairment due to facial diplegia. Among the six children under 3 years, three developed babbling with meaningful communication. The majority (16/22; 72.7%) had significant microcephaly (between −3 to −5 SD below the mean). None had a normal neurological examination at the time of last follow-up. They had either isolated axial hypotonia (3/22; 13.6%) or combined with spastic diplegia (17/22; 72.3%) or tetraplegia (3/22; 13.6%). Five patients had epilepsy starting in the first year of life that became refractory in three cases. The three patients that experienced drug-resistant epilepsy had focal daily seizures with repeated status epilepticus, whereas the other two had generalized clonic seizures in one case, and focal motor seizure in the other. Seizure control was achieved with valproate in the first patient and carbamazepine in the second. Three had occasional seizures, and one of these had febrile seizures. In contrast with the agyria-pachygyria subtype, seizures were mainly focal or multifocal and no cases of infantile spasms or myoclonic seizures were reported. Only EEG tracing of patients with controlled epilepsy could be reviewed. In both cases, the EEGs were characterized by non-specific pattern consisting of slow and monotonous background activity (Supplementary Fig. 1).

As far as molecular data are concerned, strikingly, most patients with a pattern of central pachygyria (7/13; 53.8%) carried the same recurrent mutation p.R264C in TUBA1A (Fig. 3). The other mutations were not recurrent (p.E113K, p.R422C, p.A369T, p.D396Y, p.L397P, p.G436R, p.V409I). By contrast, half of the patients (5/11) with ‘central’ polymicrogyria-like cortical dysplasia carried mutations in TUBB2B. Patients with similar neuroimaging findings were found to have the same TUBB2B mutation (p.I202T). The other six patients had TUBA1A mutations (Table 1).

Generalized polymicrogyria-like cortical dysplasia

This subtype (n = 6, including three foetuses) was characterized by generalized polymicrogyria-like cortical dysplasia with complete or near-complete involvement of the entire cerebral cortex, without any region of maximal involvement or gradient of severity (Figs 6 and 7). Enlarged ventricles and dysmorphic basal ganglia with fusion between the caudate and putamen were observed in the majority of patients (5/6; 83.3%). In this subtype, most patients showed severe vermian and hemispheric dysplasia (4/6; 66.7%). The brainstem was severely hypoplastic with a flat pons in all cases. Also, in this subtype, all patients had complete agenesis of corpus callosum which is also a consistent feature in microlissencephaly.

At the last follow-up, the three children (one male and two females) were 1, 3 and 5 years of age, respectively. In comparison to other subtypes, patients with the generalized polymicrogyria-like cortical dysplasia resembled those with the agyria-pachygyria subtype. All patients had no motor or social development. Similarly, all had early onset epileptic seizures with infantile spasms, starting in the neonatal period. Seizures were refractory in all cases. All had congenital microcephaly with a head circumference between 4 and 5 SD below the mean at birth. All had axial hypotonia or mixed central hypotonia with limb spasticity, and two showed abnormal eye movements symptomatic of severe amblyopia.

The majority of patients with diffuse polymicrogyria-like cortical dysplasia were found to carry mutations in TUBB2B (4/6) whereas the remaining patients had TUBA1A and TUBB3 mutations not found in other groups of cortical malformation syndromes (Table 1).

Simplified gyral pattern with area of focal polymicrogyria-like cortical dysplasia

This pattern that represents a milder form of tubulinopathy was found in 19 patients (12 males, seven females), including two inherited cases; one with father to son, and the other with mother to daughter transmission.

In this subtype, the most prominent feature was patchy multifocal atypical polymicrogyria over both hemispheres with a simplified gyral pattern or irregular sulcal pattern (Fig. 8). Although there was no clear specific pattern concerning gyral abnomalities distribution, the focal polymicrogyria-like cortical dysplasia was in all cases predominantly located in perisylvian and fronto-temporal regions. Also, in some cases (7/19; 36.8%) the distribution. was asymmetrical. Moderately enlarged and dysmorphic lateral ventricles were consistent features, as well as dysmorphic basal ganglia with fusion between the caudate and putamen. The corpus callosum was usually hypoplastic with a thin (6/19; 31.6%) or thick (1/19; 5.2%) appearance, whereas 11 had either hypogenesis (8/19; 42.1%) or partial corpus callosum posterior agenesis (3/19; 15.8%).

One patient (Patient LIS_TUB_059) also showed periventricular non-confluent nodular heterotopia. Of note, in this subtype, the cerebellar superior vermis was dysplastic in 15/19 cases (78.9%) and associated with hemispheric dysplasia in three cases. The brainstem was mildly hypoplastic in 5/19 cases (26.3%).

At the last follow-up, the 19 patients (12 males and seven females) in this subtype were 9 months to 40 years of age (median 5 years; mean 11.3 years). Motor delay was observed in all cases although the majority (17/19; 89.5%) were able to walk either with minimal support (10 cases) or unaided (seven cases). Among them, a spastic and ataxic gait was noted in seven (36.8%). All had moderate to severe intellectual disability and attended a special school. Interestingly, the three children with familial disorders presented with minimal intrafamilial variability although at referral, they displayed severe hypotonia and an ataxic gait. Head circumference was within the normal range in the majority (13/19; 68.4%), whereas six patients had significant and severe microcephaly from birth (less than −2.5 SD below the mean). Twelve had subtle neurological signs with mild truncal hypotonia, whereas the remaining had spastic diplegia or tetraplegia and orofacial dyspraxia. Intermittent or permanent strabismus was a frequent finding (15/19; 78.9%). In contrast with the other subtypes, only three patients developed a severe epileptic encephalopathy, and among them only one is refractory. Only five (5/19; 26.3%) had occasional seizures including one case of febrile seizures. EEG tracing could be reviewed in four patients with controlled seizures. In all cases, we identified non-specific EEG patterns consisting of slow and monotonous background activity, unusual asynchrony of sleep pattern or posterior slow theta rhythm with intermixed spikes (Supplementary Figs 2–4).

About half of the patients (9/19; 47.4%) with ‘simplified’ gyral and multifocal polymicrogyria-like cortical dysplasia had TUBB3 mutations. The remaining cases carried de novo mutations either in TUBB2B (6/19; 31.6%) or TUBB5 (3/19; 15.8%). Remarkably, one patient also had a TUBA1A mutation (p.V353I) (Table 1). Of note, all patients with TUBB5 mutations had microcephaly (2/3) compared to others (3/6 in TUBB2B and 1/9 in TUBB3, respectively). Similarly, only patients with TUBB2B mutations had severe cerebellar dysplasia that was not observed in other patients with TUBB3 and TUBB5 mutations.

Discussion

This study describes the largest cohort of patients yet reported with malformations of cortical development related to TUBA1A, TUBB2B, TUBB3, TUBB5 and TUBG1 genes, collectively known as tubulinopathies. Six years after the discovery of TUBA1A, the phenotypes associated with each gene have significantly evolved from clear splitting of each tubulin-associated malformation to a recent lumping concept of overlapping tubulinopathies. In this study, we examined the clinical and brain MRI characteristics of 80 individuals harbouring 61 distinct mainly de novo mutations in the different tubulin genes, associated with cortical malformations, and highlight the following findings.

Firstly, this study together with data from literature indicates the presence of similarities and specific features defining the key features of the ‘tubulinopathy spectrum’. These include dysmorphic basal ganglia, hypoplasia or agenesis of the corpus callosum, brainstem and cerebellar hypoplasia. Also, the emerging concept of polymicrogyria-like cortical dysplasia that is unique in tubulin-related cortical dysgenesis is worth considering when assessing brain magnetic resonance images of patients with diffuse cortical malformations. Secondly, although the range of brain developmental defects is wider than originally described, mutations in each tubulin gene are responsible for some predominant phenotypes. They include a significant proportion of foetal cases with microlissencephaly and at the other end of the spectrum, patients with mild intellectual disability, focal polymicrogyria and abnormal gyral patterning and peripheral neuropathies. Thirdly, there is clear evidence for phenotype–genotype correlation and this gives us some insight into the mechanisms underlying tubulinopathies.

Key features of tubulinopathies

Dysmorphism or unusual orientation of the basal ganglia (60/80; 75%) is one of the pathognomonic features of tubulinopathies. It is suspected to result from abnormal axon guidance of the corticospinal tract through the internal capsule (Bahi-Buisson et al., 2008; Barkovich, 2013; Cushion et al., 2013). This appearance is the result of a combination of a dysgenesis of the anterior limb of internal capsule with a head of the caudate protruding into the ventral horn of the ventricles. Consequently, the lateral ventricles are enlarged with a hooked aspect of the anterior horn. This is an easily recognizable feature on brain MRI and increasing awareness should steer clinicians and neuroradiologists in the direction of specific molecular genetic testing. Of note, the thalamus is usually less enlarged and this sign is particularly visible in association with cortical malformations of mild to moderate severity, and more subtle in lissencephalies.

Corpus callosum anomalies either partial to complete agenesis or hypoplasia (32/80; 40%) are also dominant features in tubulinopathies. We and others proposed that they are the result of the combination of abnormal corticogenesis, either impaired migration or lamination, as well as abnormal growth cone dynamics of callosal axons (Jaglin and Chelly, 2009; Barkovich, 2013). This is further reinforced by the fact that the severity of corpus callosum abnormalities seems to be related to the severity of cortical dysgenesis, either lissencephaly, microlissencephaly or generalized polymicrogyria-like cortical dysplasia (25/37; 67.6% in our series and 39/63; 61.9% including literature data). Conversely, only a few cases with multifocal polymicrogyria or pericentral pachygyria or polymicrogyria-like cortical dysplasia demonstrate complete (1/33; 3% in our series and 3/46; 6.5% including literature data) or partial corpus callosum agenesis (3/33; 9% in our series; 3/46;6.5% including the literature data) (Morris-Rosendahl et al., 2008; Kumar et al., 2010; Jansen et al., 2011; Cederquist et al., 2012; Guerrini et al., 2012; Mokanszki et al., 2012; Sohal et al., 2012; Amrom et al., 2013; Cushion et al., 2013; Okumura et al., 2013; Hikita et al., 2014). Remarkably, other projection neurons, particularly corticofugal projection neurons are also particularly affected in tubulinopathies as shown by the severe brainstem hypoplasia in tubulinopathies. These neurons project away from the cortex but do not cross the midline, and project both subcortically to deeper brain areas, and subcerebrally to brainstem and spinal cord targets.

Polymicrogyria in tubulinopathies

From the seminal description of ‘polymicrogyria’ in tubulinopathies (Jaglin and Chelly, 2009; Jaglin et al., 2009), the imaging features of this cortical malformation show specific and original features that distinguish it from others. On MRI, the consensus criteria for the diagnosis ‘typical’ polymicrogyria combines the presence of regions of apparent cortical thickening with an irregular cortical surface and a ‘stippled’ grey–white junction, usually without associated T2 signal change in patients who are fully myelinated (Barkovich et al., 1999; Takanashi and Barkovich, 2003; Leventer et al., 2010). However, a range of imaging appearances of polymicrogyria (gyral–sulcal dysmorphisms) is seen and can be divided into three main categories; coarse, delicate, and sawtooth polymicrogyria depending on morphologic appearance (Barkovich, 2010). Based on these classifications, TUBA1A, TUBB2B, TUBB3 or TUBB5-related polymicrogyria more closely resemble coarse polymicrogyria with a thick cortex and irregular surfaces on both the pial and grey–white junction sides than other patients with polymicrogyria (Jaglin and Chelly, 2009; Jaglin et al., 2009; Poirier et al., 2010, 2013b; Breuss et al., 2012; Amrom et al., 2013; Cushion et al., 2013). Of note, one of the distinctive signs of atypical polymicrogyria is the absence of deep infolding that is usually highly suggestive of polymicrogyria (Barkovich, 2010). These observations led Cushion et al. (2013) to propose the concept of atypical polymicrogyria or polymicrogyria-like cortical dysplasia. This is further reinforced by neuropathological anomalies of a foetal case of TUBB2B-related polymicrogyria that we have previously described (Jaglin and Chelly, 2009; Jaglin et al., 2009; Cushion et al., 2013). Histological features comprised disorganization of cortical layering, which is consistent with unlayered polymicrogyria, but also additional signs combining neuronal overmigration through breaches in the pial basement membrane and neuronal heterotopias more reminiscent of cobblestone lissencephaly rather than classic polymicrogyria (Jaglin and Chelly, 2009; Cushion et al., 2013).

Importantly, we and others have found a significant number of patients with a milder neurological phenotype and only highly localized areas of irregular cortex, described as ‘focal polymicrogyria’. In these patients, the main features reminiscent of a tubulinopathy, are the abnormal basal ganglia, corpus callosum abnormalities and cerebellar hypoplasia (Cushion et al., 2013). These observations suggest that tubulinopathies may be responsible for a subset of children with corpus callosum abnormalities, either hypogenesis or dysmorphic and cerebellar hypoplasia without obvious cortical malformations. This group of malformations, particularly seen in TUBB3 mutations, but also reported with TUBB2B and TUBA1A mutations (Kumar et al., 2010; Cushion et al., 2013), is complex and closer to phenotypes observed in mutations involved in congenital fibrosis of extraocular muscle disorders (Tischfield et al., 2010, 2011; Chew et al., 2013). Further studies in these latter groups are clearly needed. This group of malformations, is particularly seen in TUBB3 mutations, but is also reported with TUBB2B and TUBA1A mutations (Kumar et al., 2010; Cushion et al., 2013).

The predominant phenotypes associated with mutations in each specific tubulin gene

Our data combined with the literature also emphasize that prominent and prevalent malformations of cortical development are associated with each mutated gene (Table 1).

The core phenotype of TUBA1A-related tubulinopathies (45 described here and 30 reported) consists of lissencephaly, most frequently classic (28/75; 37.3%) or with cerebellar hypoplasia (11/75; 14.7%) (Poirier et al., 2007; Bahi-Buisson et al., 2008; Fallet-Bianco et al., 2008; Morris-Rosendahl et al., 2008; Kumar et al., 2010; Okumura et al., 2012; Sohal et al., 2012; Hikita et al., 2013). Our data also show that a significant proportion of TUBA1A mutations cause microlissencephaly (10/75; 13.3%). More interestingly, they represent the first dominant aetiology of microlissencephaly, as other known genetic disorders, either NDE1 or WDR62 mutations, are of autosomal recessive inheritance (Bilguvar et al., 2010). At the less severe end of the spectrum, ‘central pachygyria’ represents about one-third of our cohort (13/45; 28.9%), whereas this was not reported in other series (Bahi-Buisson et al., 2008). Finally, in addition to the above highlighted phenotypes, the spectrum of TUBA1A-related cortical dysgenesis includes a low proportion of patients with either central (8/75; 10.7%) and less frequently generalized (2/75; 2.7%) polymicrogyria-like cortical dysplasia, or simplified gyral pattern with focal atypical polymicrogyria (3/75; 4%) (Jansen et al., 2011; Cushion et al., 2013; Poirier et al., 2013).

One of the most striking cortical malformations associated with TUBB2B mutations (18 described here and 14 reported elsewhere) is polymicrogyria-like cortical dysplasia (28/32; 87.5%), either centrally predominant (12/28; 42.8%) or generalized (7/28; 25%). Alternatively, a significant proportion of TUBB2B-mutated patients demonstrate a milder cortical malformation consisting of focal polymicrogyria-like cortical dysplasia and simplified gyral pattern (9/28; 32.1%) that is also seen in TUBB3 and TUBB5 mutations. At the other end of the spectrum, TUBB2B tubulinopathies may be associated with microlissencephaly (3/32; 9.4%) or lissencephaly with agenesis of the corpus callosum (1/32; 3.1%) (Cederquist et al., 2012; Guerrini et al., 2012; Romaniello et al., 2012; Amrom et al., 2013; Cushion et al., 2013).

The spectrum of TUBB3-related tubulinopathies is wider than others as it encompasses ophthalmological and peripheral nerve pathologies that are not necessarily associated with cortical malformations (Poirier et al., 2010; Tischfield et al., 2010, 2011; Chew et al., 2013). If we focus our attention on only cortical dysgenesis (n = 11), one of the most recognizable features is their milder severity with a majority of patients demonstrating focal or multifocal polymicrogyria-like cortical dysplasia with an abnormal and simplified gyral pattern (9/11; 81.8%). However, it is noteworthy that the remaining two patients with TUBB3 mutations demonstrated extremely severe cortical dysgeneses, one foetal case showing microlissencephaly and the other, generalized polymicrogyria.

For TUBB5 and TUBG1 mutations, the number of patients with mutations in these genes is limited and conclusions are still difficult to draw. However, on the basis of available data, TUBB5 tubulinopathies are also highly reminiscent of TUBB3 cortical dysgenesis, with one of the most distinctive features of TUBB5 being microcephaly associated with relative mild cortical dysgenesis (Breuss et al., 2012). On the other hand, TUBG1-related cortical dysgenesis consists mainly of a pattern of classic lissencephaly with severe microcephaly that resembles TUBA1A-related lissencephaly (particularly those caused by p.R402C/H substitutions) (Poirier et al., 2013a). One milder case demonstrated posteriorly predominant subcortical band heterotopia, a pattern previously observed in one case of TUBA1A mutation but with corpus callosum agenesis (Bahi-Buisson et al., 2008; Poirier et al., 2013).

Evidence for phenotype–genotype correlations in tubulinopathies

Our data, in combination with the literature (Tischfield et al., 2011) emphasize that prominent and prevalent malformations of cortical development are associated with each of the mutated genes. These predominant tubulins’ related phenotypes, as well as less frequent ones, are summarized in Table 1.

As far as genotype–phenotype correlations are concerned, our data also show that one of the consistent findings in both this study and data in the literature is that recurrent mutations result in similar phenotypes. Amongst TUBA1A mutations, two major recurrent mutations were found. The substitution p.R264C (7/45; 15.6% of TUBA1A mutations) is invariably associated with central pachygyria, and more interestingly is responsible for half of this malformation subtype. The second hot-spot of mutations, p.R402H (5/45; 11.1%) is invariably responsible for classic lissencephalies in our series and in the literature (Kumar et al., 2010; Mokanszki et al., 2012). Similarly, recurrent TUBB2B (p.S172P, p.I202T, p.F265L, p.A248V) and TUBB3 mutations (p.R262H, p.E205K, p.E410K) also result in similar phenotypes, although their occurrence is less common (Tischfield et al., 2010; Amrom et al., 2013; Chew et al., 2013).

Interestingly, mutation-related correlations are reinforced by the observation of distinguishable phenotypes resulting from different substitutions altering the same residue. For example, the most common TUBA1A substitution p.R264C, predominantly results in ‘central pachygyria’, whereas p.R264H causes microlissencephaly with complete agenesis of the corpus callosum, one of the most severe forms of tubulinopathy. Similarly, consequences of recurrent substitutions p.R422H and p.R422C are distinguishable, because the former is responsible for lissencephaly whereas p.R422C causes central pachygyria (Kumar et al., 2010). Similarly also, the TUBB3 substitution p.R262C mostly results in an isolated restriction of eye movements, whereas p.R262H causes a severe ophthalmological disorder combined with a brain malformation. In vitro folding assays and overexpression in mammalian cells has demonstrated that the p.R262H mutation results in the generation of more tubulin heterodimers with higher levels of microtubule incorporation than p.R262C (Tischfield et al., 2010, 2011). Perhaps in a similar manner, the mutation p.R264H that causes lethal microlissencephaly, one of the most severe brain malformations, compared to p.R264C, may also permit higher levels of heterodimer formation and microtubule incorporation. Thus, the consequences of the substituted residue may result in different changes in the function of tubulin, from drastic to milder modifications resulting in various cortical malformations.

Table 2

Clinical features of tubulinopathies subtypes

Phenotypes 45 TUBA1A 18 TUBB2B 11 TUBB3 3 TUBB5 3 TUBG1 
 27 living 12 living 10 living 3 living 3 living 
 18 foetuses 6 foetuses 1 foetus   
Male 29 13 
Female 16 
Age min–max (years) 1–18 0.9–36 2–36 3–4.8 1.5–31 
Motor impairment      
Total 26 (1 N/A) 12 10 
Mild (mild spasticity/ataxia with minor impairment) 9 (34.6%) 4 (30%) 4 (40%) 
Moderate (ability to walk independently or with minimal aid) 9 (34.6%) 4 (30%) 4 (40%) 3 (100%) 1 (33.3%) 
Severe 8 (30.8%) 4 (30%) 2 (20%) 2 (66.7%) 
Abnormal tone      
Total 27 12 10 
Spastic diplegia or tetraplegia 23 (85.2%) 6 (50%) 3 (30%) 3 (100%) 
Hypotonia 4 (14.8%) 3 (25%) 3 (30%) 3 (100%) 
Epilepsy      
Total 27 12 10 
Occasional seizures (or controlled epilepsy) 7 (25.9%) 4 (33.3%) 1 (10%) 1 (33%) 
Refractory 10/27 (37%) 2 (16.7%) 2 (20%) 3 (100%) 
Abnormal head size      
Total 25 12 10 
Microcephaly (HC < −2SD) 22/25 (88%) 6 (50%) 2 (20%) 3 (100%) 2 (66.6%) 
Major characteristics of brain malformations (MRI)      
Total 45 18 11 
Classic lissencephalya 15 (33.3%) 1 (5.6%) 3 (100%) 
Microlissencephalyb 9 (20%) 2 (11.1%) 1 (9%) 
Perisylvian pachygyria or polymicrogyria-like cortical dysplasia 19 (42.2%) 5 (27.8%) 
Generalized polymicrogyria-like cortical dysplasia 1 (2.2%) 4 (22.2%) 1 (9%) 
Multifocal polymicrogyria with or without simplified gyral pattern 1 (2.2%) 6 (33.3%) 9 (81.8%) 3 (100%) 
Moderate to severe cerebellar vermis/hemisphere dysplasia and/or hypoplasia 18/45 (40%) 11/18 (61.1%) 11/11 (100%) 1/3 (33%) 
Agenesis of corpus callosum (partial or complete) 20/45 (44.4%) 8/18 (44.4%) 3/11 (27.3%) 1/3 (33%) 
Phenotypes 45 TUBA1A 18 TUBB2B 11 TUBB3 3 TUBB5 3 TUBG1 
 27 living 12 living 10 living 3 living 3 living 
 18 foetuses 6 foetuses 1 foetus   
Male 29 13 
Female 16 
Age min–max (years) 1–18 0.9–36 2–36 3–4.8 1.5–31 
Motor impairment      
Total 26 (1 N/A) 12 10 
Mild (mild spasticity/ataxia with minor impairment) 9 (34.6%) 4 (30%) 4 (40%) 
Moderate (ability to walk independently or with minimal aid) 9 (34.6%) 4 (30%) 4 (40%) 3 (100%) 1 (33.3%) 
Severe 8 (30.8%) 4 (30%) 2 (20%) 2 (66.7%) 
Abnormal tone      
Total 27 12 10 
Spastic diplegia or tetraplegia 23 (85.2%) 6 (50%) 3 (30%) 3 (100%) 
Hypotonia 4 (14.8%) 3 (25%) 3 (30%) 3 (100%) 
Epilepsy      
Total 27 12 10 
Occasional seizures (or controlled epilepsy) 7 (25.9%) 4 (33.3%) 1 (10%) 1 (33%) 
Refractory 10/27 (37%) 2 (16.7%) 2 (20%) 3 (100%) 
Abnormal head size      
Total 25 12 10 
Microcephaly (HC < −2SD) 22/25 (88%) 6 (50%) 2 (20%) 3 (100%) 2 (66.6%) 
Major characteristics of brain malformations (MRI)      
Total 45 18 11 
Classic lissencephalya 15 (33.3%) 1 (5.6%) 3 (100%) 
Microlissencephalyb 9 (20%) 2 (11.1%) 1 (9%) 
Perisylvian pachygyria or polymicrogyria-like cortical dysplasia 19 (42.2%) 5 (27.8%) 
Generalized polymicrogyria-like cortical dysplasia 1 (2.2%) 4 (22.2%) 1 (9%) 
Multifocal polymicrogyria with or without simplified gyral pattern 1 (2.2%) 6 (33.3%) 9 (81.8%) 3 (100%) 
Moderate to severe cerebellar vermis/hemisphere dysplasia and/or hypoplasia 18/45 (40%) 11/18 (61.1%) 11/11 (100%) 1/3 (33%) 
Agenesis of corpus callosum (partial or complete) 20/45 (44.4%) 8/18 (44.4%) 3/11 (27.3%) 1/3 (33%) 

aLissencephaly includes classic lissencephaly with all degrees of severity, mild, moderate and severe and lissencephaly with cerebellar hypoplasia.

bLissencephaly with cerebellar hypoplasia includes all degrees of severity, mild, moderate and severe.

To gain further insight into the genotype–phenotype relationships, we also performed additional analysis of the relationship between the predicted effects of mutations on TUBA1A, TUBB2B or TUBB3 structure and function, taking into account local or global destabilization in the structure. Depending on their localization in the tertiary structure of tubulin, point mutations can be predicted to affect five distinct biochemical properties: nucleotide exchange and hydrolysis (GTP binding pocket), longitudinal and lateral protofilament interactions and microtubule–protein interactions with kinesin, dynein and other microtubule-associated proteins and heterodimer stability (Lowe et al., 2001; Tischfield et al., 2011) (Figs 9–11 and Table 1). Overall, the large majority of TUBA1A, TUBB2B and TUBB3 mutations are predicted to impair interactions with microtubule-associated proteins and motor proteins (28/75; 37.3%) here, and (64/129; 49.6%) including literature data) or tubulin folding (24/75; 32%) here and (31/129; 24%) including literature data.

Figure 9

Localization of TUBA1A-mutated residues in the microtubule structure. (A) Microtubule structure visualized from the side, plus-end pointing up. The crystal structure of a tubulin dimer (grey ribbons, 1JFF.PDB) was fitted into a subnanometer resolution cryo-EM microtubule reconstruction (EMDB entry 1788), thus generating a pseudo-atomic model of the microtubule. Side-chains of TUBA1A mutated residues reported here are displayed as spheres, and coloured according to the phenotype associated with their mutations: agyria-pachygyria phenotype in red, lissencephaly with cerebellar hypoplasia in green, central pachygyria in purple, polymicrogyria phenotype in blue, simplified gyral pattern in black. GTP and GDP are displayed as grey spheres. The location of mutated residues and the nature of the substitution were used to predict the potential effect of mutations on microtubule assembly and interactions. (B) Microtubule structure visualized from the luminal side, plus end pointing up (180° rotation compared with A). Representation parameters as in A. (C) Single protofilament rotated 90° compared with A, and with plus-end pointing up. Representation parameters as in A. In addition, atomic structures of a kinesin motor domain (red ribbons, 1BG2.PDB) and DCX N-terminal DC domain (yellow ribbons, 1MJD.PDB) were displayed. Mutated residues associated with agyria-pachygyria (red), as well as V409 localize at the interface with kinesin, whereas R264 is in close proximity of DCX.

Figure 9

Localization of TUBA1A-mutated residues in the microtubule structure. (A) Microtubule structure visualized from the side, plus-end pointing up. The crystal structure of a tubulin dimer (grey ribbons, 1JFF.PDB) was fitted into a subnanometer resolution cryo-EM microtubule reconstruction (EMDB entry 1788), thus generating a pseudo-atomic model of the microtubule. Side-chains of TUBA1A mutated residues reported here are displayed as spheres, and coloured according to the phenotype associated with their mutations: agyria-pachygyria phenotype in red, lissencephaly with cerebellar hypoplasia in green, central pachygyria in purple, polymicrogyria phenotype in blue, simplified gyral pattern in black. GTP and GDP are displayed as grey spheres. The location of mutated residues and the nature of the substitution were used to predict the potential effect of mutations on microtubule assembly and interactions. (B) Microtubule structure visualized from the luminal side, plus end pointing up (180° rotation compared with A). Representation parameters as in A. (C) Single protofilament rotated 90° compared with A, and with plus-end pointing up. Representation parameters as in A. In addition, atomic structures of a kinesin motor domain (red ribbons, 1BG2.PDB) and DCX N-terminal DC domain (yellow ribbons, 1MJD.PDB) were displayed. Mutated residues associated with agyria-pachygyria (red), as well as V409 localize at the interface with kinesin, whereas R264 is in close proximity of DCX.

Figure 10

Localization of TUBB2B-mutated residues in the microtubule structure. (A) Localization of TUBB2B mutated residues in the microtubule structure visualised from the side, plus-end pointing up. Representation parameters are as in Fig. 9A. (B) Luminal view. (C) Single protofilament rotated 90° compared to B, and with plus-end pointing up. TUBB2B mutations reported here do not affect the microtubule-kinesin interface. Representation parameters as in Fig. 9C.

Figure 10

Localization of TUBB2B-mutated residues in the microtubule structure. (A) Localization of TUBB2B mutated residues in the microtubule structure visualised from the side, plus-end pointing up. Representation parameters are as in Fig. 9A. (B) Luminal view. (C) Single protofilament rotated 90° compared to B, and with plus-end pointing up. TUBB2B mutations reported here do not affect the microtubule-kinesin interface. Representation parameters as in Fig. 9C.

Figure 11

Localization of TUBB3 mutated residues in the microtubule structure. (A) Localization of TUBB3 mutated residues in the microtubule structure visualized from the side, plus-end pointing up. Representation parameters are as in Fig. 10A. (B) Luminal view. (C) Single protofilament rotated 90° compared to B, and with plus-end pointing up. TUBB3 mutations reported here do not affect the microtubule-kinesin interface. Representation parameters as in Fig. 9C.

Figure 11

Localization of TUBB3 mutated residues in the microtubule structure. (A) Localization of TUBB3 mutated residues in the microtubule structure visualized from the side, plus-end pointing up. Representation parameters are as in Fig. 10A. (B) Luminal view. (C) Single protofilament rotated 90° compared to B, and with plus-end pointing up. TUBB3 mutations reported here do not affect the microtubule-kinesin interface. Representation parameters as in Fig. 9C.

Table 3

Tubulin mutations groups were assigned according to structural data and biochemical data

 Total Group 1 severe Group 2 moderate Group 3 mild 
  This series (n = 34) + literature (n = 64) (+ *3 TUBG1This series (n = 25) + literature (n = 30) This series (n = 15) + literature (n = 36) (+ *3 TUBB5
Gene (number of cases reported) TUBA1A (75) TUBA1A (51) TUBA1A (21) TUBA1A (3) 
TUBB2B (29) TUBB2B (11) TUBB2B (9) TUBB2B (9) 
TUBB3 (27) TUBB3 (2)  TUBB3 (25) 
TUBG1 (3) TUBG1 (3)   
TUBB5 (3)   TUBB5 (3) 
A: GTP binding 7/129 (5.4%) 7 (10.9%) 1 (3.3%) 
B: Longitudinal interactions 12/129 (9.3%) 5 (7.8%) 3 (30%) 4 (12.5%) 
C: Lateral protofilament interactions 15/129 (11.6%) 7 (10.9%) 5 (16.7%) 3 (8.3%) 
D: MAP and motor protein interactions 65/129 (50.4%) 30 (46.8%) 17 (56.7%) 18 (50%) 
E: Tubulin folding 31/129 (24%) 15 (23.4%) 4 (13.3%) 12 (33.3%) 
 Total Group 1 severe Group 2 moderate Group 3 mild 
  This series (n = 34) + literature (n = 64) (+ *3 TUBG1This series (n = 25) + literature (n = 30) This series (n = 15) + literature (n = 36) (+ *3 TUBB5
Gene (number of cases reported) TUBA1A (75) TUBA1A (51) TUBA1A (21) TUBA1A (3) 
TUBB2B (29) TUBB2B (11) TUBB2B (9) TUBB2B (9) 
TUBB3 (27) TUBB3 (2)  TUBB3 (25) 
TUBG1 (3) TUBG1 (3)   
TUBB5 (3)   TUBB5 (3) 
A: GTP binding 7/129 (5.4%) 7 (10.9%) 1 (3.3%) 
B: Longitudinal interactions 12/129 (9.3%) 5 (7.8%) 3 (30%) 4 (12.5%) 
C: Lateral protofilament interactions 15/129 (11.6%) 7 (10.9%) 5 (16.7%) 3 (8.3%) 
D: MAP and motor protein interactions 65/129 (50.4%) 30 (46.8%) 17 (56.7%) 18 (50%) 
E: Tubulin folding 31/129 (24%) 15 (23.4%) 4 (13.3%) 12 (33.3%) 

Three groups of severity:

Group 1 (severe tubulinopathies) included classic lissencephaly and with cerebellar hypoplasia, microlissencephaly, and diffuse polymicrogyria-like subtypes; Group 2 (moderate tubulinopathies) comprises central polymicrogyria and pachygyria; and Group 3 (milder tubulinopathies), simplified gyral pattern and multifocal polymicrogyria.

We also examined whether a specific group of mutations could be related to the severity of cortical malformations. To investigate this question, we identified three groups, Group 1 (severe tubulinopathies) included classic lissencephaly with cerebellar hypoplasia, microlissencephaly and diffuse polymicrogyria-like subtypes (n = 34 in our series, 64 including reported cases), Group 2 (moderate tubulinopathies) comprised central polymicrogyria and pachygyria (n = 25 in our series, 30 including reported cases) and Group 3 (milder tubulinopathies) consisted of simplified gyral pattern and multifocal polymicrogyria subtypes (n = 15 in our series, 36 including reported cases). No significant differences were observed between the different subgroups. However, it is noteworthy that severe tubulinopathies were more often associated with tubulin mutations affecting the GTP binding pocket [7 (10.9%)], although these mutations were never found in milder phenotypes. By contrast, about a third of tubulin mutations leading to milder forms were predicted to impair tubulin folding [12/36 (33.3%), compared to a lower proportion in severe and moderate malformations (15; 23.4%) and (4; 13.3%), respectively (Table 1)].

Because tubulinopathies comprise a defect in axonogenesis and axon tract development as demonstrated by the high frequency of corpus callosum dysgenesis and the abnormal shape of the internal capsule, we looked at whether specific mutation groups could be associated with absent, moderate (defined as hypoplastic, dysmorphic or hypogenetic) or severe (defined as partial or complete agenesis) corpus callosum abnormalities. The distribution of mutation type was similar in each group suggesting that we could not predict the degree of corpus callosum dysgenesis within a mutation group (data not shown). Moreover, among patients with recurrent mutations (i.e. p.R264C and p.R402H), we found a wide range of corpus callosum abnormalities (normal, hypogenesis or complete agenesis), suggesting that the mutation type could not help to determine the degree of alteration of corpus callosum development.

Finally, we attempted to evaluate whether mutation groups could correlate with the degree of cerebellar dysgenesis. Similar to the results obtained for corpus callosum dysgenesis, the distribution of mutation type did not differ among different groups (data not shown). Also, data for recurrent mutations indicate that the resulting cerebellar malformation is variable. For example, the mutation p.S172P was associated with normal cerebellum in one case and severe vermian dysplasia in the other. More interestingly, the mutation p.R264C resulted in either a normal, hypoplastic or severely dysgenetic cerebellum indicating that the mutation type could not help to determine the degree of cerebellar malformation.

Where biochemical and functional data were available, we compared our structure-based predictions to the actual ability of mutant tubulins to fold properly and assemble into microtubules, in vitro and in cells (Supplementary material). Remarkably, most mutants predicted to have an impaired folding (Group E) indeed showed a decreased ability to form heterodimers in vitro, although to variable extents. More surprisingly, several mutations whose locations predicted primarily an alteration of interactions with microtubule-associated proteins were also found to alter heterodimer formation in vitro. In fact, these surface mutations might affect the stability of the protein, either directly by local perturbation of the structure, or indirectly through the interaction with chaperones or tubulin post-translational modification enzymes. In support, the discrepancies found between observations in vitro and in cells indicate that the consequences of mutations not only arise from the intrinsic properties of tubulin but also from its interaction with various cellular factors. In addition, as discussed previously regarding phenotypes, the biochemical properties of mutant tubulins are also differently altered depending on the precise nature of the substituted surface residue: e.g. TUBA1A R402C perturbs heterodimer formation much more dramatically than R402H. This emphasizes the limitations of predictions based on the structure of wild-type tubulin only, and therefore the need for a direct structural characterization of mutant tubulins, either on their own or in complex with relevant microtubule-associated proteins.

Conclusion

Through their structural role as building blocks of microtubules, tubulins play a large part in brain development; through neuronogenesis, neuronal migration, cortical organization–pial basement membrane integrity, and also in axon guidance through axonal leading process (growth cone) and maintenance. Therefore, for each tubulin isotype, the diversity of malformations might arise from the fact that different mutations may variably affect tubulin functions. Moreover, though recurrent mutations are consistently associated with almost identical cortical dysgeneses, at the current level of our understanding of tubulin functions and their cellular roles, predictions using structural data do not show definite correlation between the location of mutations and the degree of phenotype severity. Further structural characterization of tubulin mutants will be required to explain how the precise nature of a substitution can dramatically alter phenotype severity.

Acknowledgements

We would like to thank the patients and their families, as well as their referring physicians and genetic counsellors, without whom this study would not have been possible. We would like to thank Nathalie Carion, Magalie Percevault and Aurelie Toussaint for her help in performing molecular analysis. The authors are grateful to Dr M. Eiserman for their help in the analysis and the selection of EEG. We thank members of the Chelly lab for their thoughtful comments.

Funding

This work was supported by funding from INSERM, Fondation pour la Recherche Médicale (FRM; J Chelly—Equipe FRM 2013: DEQ2000326477), Fondation JED-Belgique, Agence National de Recherche (ANR Blanc 1103 01 - project R11039KK; ANR E-Rare-012-01 – project E10107KP), the EU-FP7 project GENECODYS, grant number 241995 and EU-FP7 project DESIRE, Grant agreement no: 602531.

Supplementary material

Supplementary material is available at Brain online.

References

Abdollahi
MR
Morrison
E
Sirey
T
Molnar
Z
Hayward
BE
Carr
IM
, et al.  . 
Mutation of the variant alpha-tubulin TUBA8 results in polymicrogyria with optic nerve hypoplasia
Am J Hum Genet
 , 
2009
, vol. 
85
 (pg. 
737
-
44
)
Amrom
D
Tanyalcin
I
Verhelst
H
Deconinck
N
Brouhard
G
Decarie
JC
, et al.  . 
Polymicrogyria with dysmorphic basal ganglia? Think tubulin!
Clin Genet
 , 
2014
, vol. 
85
 (pg. 
178
-
83
)
Bahi-Buisson
N
Poirier
K
Boddaert
N
Saillour
Y
Castelnau
L
Philip
N
, et al.  . 
Refinement of cortical dysgeneses spectrum associated with TUBA1A mutations
J Med Genet
 , 
2008
, vol. 
45
 (pg. 
647
-
53
)
Barkovich
AJ
Current concepts of polymicrogyria
Neuroradiology
 , 
2010
, vol. 
52
 (pg. 
479
-
87
)
Barkovich
AJ
Guerrini
R
Kuzniecky
RI
Jackson
GD
Dobyns
WB
A developmental and genetic classification for malformations of cortical development: update 2012
Brain
 , 
2012
, vol. 
135
 
Pt. 5
(pg. 
1348
-
69
)
Barkovich
AJ
Hevner
R
Guerrini
R
Syndromes of bilateral symmetrical polymicrogyria
AJNR Am J Neuroradiol
 , 
1999
, vol. 
20
 (pg. 
1814
-
21
)
Barkovich
AJ
Kuzniecky
RI
Jackson
GD
Guerrini
R
Dobyns
WB
A developmental and genetic classification for malformations of cortical development
Neurology
 , 
2005
, vol. 
65
 (pg. 
1873
-
87
)
Barkovich
J
Complication begets clarification in classification
Brain
 , 
2013
, vol. 
136
 
Pt 2
(pg. 
368
-
73
)
Bilguvar
K
Ozturk
AK
Louvi
A
Kwan
KY
Choi
M
Tatli
B
, et al.  . 
Whole-exome sequencing identifies recessive WDR62 mutations in severe brain malformations
Nature
 , 
2010
, vol. 
467
 (pg. 
207
-
10
)
Breuss
M
Heng
JI
Poirier
K
Tian
G
Jaglin
XH
Qu
Z
, et al.  . 
Mutations in the beta-Tubulin Gene TUBB5 cause microcephaly with structural brain abnormalities
Cell Rep
 , 
2012
, vol. 
2
 (pg. 
1554
-
62
)
Cederquist
GY
Luchniak
A
Tischfield
MA
Peeva
M
Song
Y
Menezes
MP
, et al.  . 
An inherited TUBB2B mutation alters a kinesin-binding site and causes polymicrogyria, CFEOM and axon dysinnervation
Hum Mol Genet
 , 
2012
, vol. 
21
 (pg. 
5484
-
99
)
Chew
S
Balasubramanian
R
Chan
WM
Kang
PB
Andrews
C
Webb
BD
, et al.  . 
A novel syndrome caused by the E410K amino acid substitution in the neuronal beta-tubulin isotype 3
Brain
 , 
2013
, vol. 
136
 
Pt 2
(pg. 
522
-
35
)
Cushion
TD
Dobyns
WB
Mullins
JG
Stoodley
N
Chung
SK
Fry
AE
, et al.  . 
Overlapping cortical malformations and mutations in TUBB2B and TUBA1A
Brain
 , 
2013
, vol. 
136
 
Pt 2
(pg. 
536
-
48
)
Dobyns
WB
Truwit
CL
Ross
ME
Matsumoto
N
Pilz
DT
Ledbetter
DH
, et al.  . 
Differences in the gyral pattern distinguish chromosome 17-linked and X-linked lissencephaly
Neurology
 , 
1999
, vol. 
53
 (pg. 
270
-
7
)
Fallet-Bianco
C
Loeuillet
L
Poirier
K
Loget
P
Chapon
F
Pasquier
L
, et al.  . 
Neuropathological phenotype of a distinct form of lissencephaly associated with mutations in TUBA1A
Brain
 , 
2008
, vol. 
131
 
Pt 9
(pg. 
2304
-
20
)
Guerrini
R
Mei
D
Cordelli
DM
Pucatti
D
Franzoni
E
Parrini
E
Symmetric polymicrogyria and pachygyria associated with TUBB2B gene mutations
Eur J Hum Genet
 , 
2012
, vol. 
20
 (pg. 
995
-
8
)
Higginbotham
HR
Gleeson
JG
The centrosome in neuronal development
Trends Neurosci
 , 
2007
, vol. 
30
 (pg. 
276
-
83
)
Hikita
N
Hattori
H
Kato
M
Sakuma
S
Morotomi
Y
Ishida
H
, et al.  . 
A case of TUBA1A mutation presenting with lissencephaly and Hirschsprung disease
Brain Dev
 , 
2014
, vol. 
36
 (pg. 
159
-
62
)
Hoogenraad
CC
Bradke
F
Control of neuronal polarity and plasticity—a renaissance for microtubules?
Trends Cell Biol
 , 
2009
, vol. 
19
 (pg. 
669
-
76
)
Jaglin
XH
Chelly
J
Tubulin-related cortical dysgeneses: microtubule dysfunction underlying neuronal migration defects
Trends Genet
 , 
2009
, vol. 
25
 (pg. 
555
-
66
)
Jaglin
XH
Poirier
K
Saillour
Y
Buhler
E
Tian
G
Bahi-Buisson
N
, et al.  . 
Mutations in the beta-tubulin gene TUBB2B result in asymmetrical polymicrogyria
Nat Genet
 , 
2009
, vol. 
41
 (pg. 
746
-
52
)
Jansen
AC
Oostra
A
Desprechins
B
De Vlaeminck
Y
Verhelst
H
Regal
L
, et al.  . 
TUBA1A mutations: from isolated lissencephaly to familial polymicrogyria
Neurology
 , 
2011
, vol. 
76
 (pg. 
988
-
92
)
Joshi
HC
Cleveland
DW
Diversity among tubulin subunits: toward what functional end?
Cell Motility Cytoskeleton
 , 
1990
, vol. 
16
 (pg. 
159
-
63
)
Keays
DA
Tian
G
Poirier
K
Huang
GJ
Siebold
C
Cleak
J
, et al.  . 
Mutations in alpha-tubulin cause abnormal neuronal migration in mice and lissencephaly in humans
Cell
 , 
2007
, vol. 
128
 (pg. 
45
-
57
)
Kumar
RA
Pilz
DT
Babatz
TD
Cushion
TD
Harvey
K
Topf
M
, et al.  . 
TUBA1A mutations cause wide spectrum lissencephaly (smooth brain) and suggest that multiple neuronal migration pathways converge on alpha tubulins
Hum Mol Genet
 , 
2010
, vol. 
19
 (pg. 
2817
-
27
)
Lecourtois
M
Poirier
K
Friocourt
G
Jaglin
X
Goldenberg
A
Saugier-Veber
P
, et al.  . 
Human lissencephaly with cerebellar hypoplasia due to mutations in TUBA1A: expansion of the foetal neuropathological phenotype
Acta Neuropathol
 , 
2010
, vol. 
119
 (pg. 
779
-
89
)
Leventer
RJ
Jansen
A
Pilz
DT
Stoodley
N
Marini
C
Dubeau
F
, et al.  . 
Clinical and imaging heterogeneity of polymicrogyria: a study of 328 patients
Brain
 , 
2010
, vol. 
133
 
Pt 5
(pg. 
1415
-
27
)
Lopata
MA
Cleveland
DW
In vivo microtubules are copolymers of available beta-tubulin isotypes: localization of each of six vertebrate beta-tubulin isotypes using polyclonal antibodies elicited by synthetic peptide antigens
J Cell Biol
 , 
1987
, vol. 
105
 (pg. 
1707
-
20
)
Lowe
J
Li
H
Downing
KH
Nogales
E
Refined structure of alpha beta-tubulin at 3.5 A resolution
J Mol Biol
 , 
2001
, vol. 
313
 (pg. 
1045
-
57
)
Luduena
RF
Are tubulin isotypes functionally significant
Mol Biol Cell
 , 
1993
, vol. 
4
 (pg. 
445
-
57
)
Mokanszki
A
Korhegyi
I
Szabo
N
Bereg
E
Gergev
G
Balogh
E
, et al.  . 
Lissencephaly and band heterotopia: LIS1, TUBA1A, and DCX mutations in Hungary
J Child Neurol
 , 
2012
, vol. 
27
 (pg. 
1534
-
40
)
Morris-Rosendahl
DJ
Najm
J
Lachmeijer
AM
Sztriha
L
Martins
M
Kuechler
A
, et al.  . 
Refining the phenotype of alpha-1a Tubulin (TUBA1A) mutation in patients with classical lissencephaly
Clin Genet
 , 
2008
, vol. 
74
 (pg. 
425
-
33
)
Nogales
E
Wolf
SG
Downing
KH
Structure of the alpha beta tubulin dimer by electron crystallography
Nature
 , 
1998
, vol. 
391
 (pg. 
199
-
203
)
Norman
MG
Roberts
M
Sirois
J
Tremblay
LJ
Lissencephaly
Can J Neurol Sci
 , 
1976
, vol. 
3
 (pg. 
39
-
46
)
Okumura
A
Hayashi
M
Tsurui
H
Yamakawa
Y
Abe
S
Kudo
T
, et al.  . 
Lissencephaly with marked ventricular dilation, agenesis of corpus callosum, and cerebellar hypoplasia caused by TUBA1A mutation
Brain Dev
 , 
2013
, vol. 
35
 (pg. 
274
-
9
)
Poirier
K
Keays
DA
Francis
F
Saillour
Y
Bahi
N
Manouvrier
S
, et al.  . 
Large spectrum of lissencephaly and pachygyria phenotypes resulting from de novo missense mutations in tubulin alpha 1A (TUBA1A)
Hum Mutat
 , 
2007
, vol. 
28
 (pg. 
1055
-
64
)
Poirier
K
Lebrun
N
Broix
L
Tian
G
Saillour
Y
Boscheron
C
, et al.  . 
Mutations in TUBG1, DYNC1H1, KIF5C and KIF2A cause malformations of cortical development and microcephaly
Nat Genet
 , 
2013a
, vol. 
45
 (pg. 
639
-
47
)
Poirier
K
Saillour
Y
Bahi-Buisson
N
Jaglin
XH
Fallet-Bianco
C
Nabbout
R
, et al.  . 
Mutations in the neuronal ss-tubulin subunit TUBB3 result in malformation of cortical development and neuronal migration defects
Hum Mol Genet
 , 
2010
, vol. 
19
 (pg. 
4462
-
73
)
Poirier
K
Saillour
Y
Fourniol
F
Francis
F
Souville
I
Valence
S
, et al.  . 
Expanding the spectrum of TUBA1A-related cortical dysgenesis to Polymicrogyria
Eur J Hum Genet
 , 
2013b
, vol. 
21
 (pg. 
381
-
5
)
Romaniello
R
Tonelli
A
Arrigoni
F
Baschirotto
C
Triulzi
F
Bresolin
N
, et al.  . 
A novel mutation in the beta-tubulin gene TUBB2B associated with complex malformation of cortical development and deficits in axonal guidance
Dev Med Child Neurol
 , 
2012
, vol. 
54
 (pg. 
765
-
9
)
Ross
ME
Swanson
K
Dobyns
WB
Lissencephaly with cerebellar hypoplasia (LCH): a heterogeneous group of cortical malformations
Neuropediatrics
 , 
2001
, vol. 
32
 (pg. 
256
-
63
)
Sohal
AP
Montgomery
T
Mitra
D
Ramesh
V
TUBA1A mutation-associated lissencephaly: case report and review of the literature
Pediatric Neurol
 , 
2012
, vol. 
46
 (pg. 
127
-
31
)
Takanashi
J
Barkovich
AJ
The changing MR imaging appearance of polymicrogyria: a consequence of myelination
AJNR Am J Neuroradiol
 , 
2003
, vol. 
24
 (pg. 
788
-
93
)
Tischfield
MA
Baris
HN
Wu
C
Rudolph
G
Van Maldergem
L
He
W
, et al.  . 
Human TUBB3 mutations perturb microtubule dynamics, kinesin interactions, and axon guidance
Cell
 , 
2010
, vol. 
140
 (pg. 
74
-
87
)
Tischfield
MA
Cederquist
GY
Gupta
ML
Jr
Engle
EC
Phenotypic spectrum of the tubulin-related disorders and functional implications of disease-causing mutations
Curr Opin Genet Dev
 , 
2011
, vol. 
21
 (pg. 
286
-
94
)
Zanni
G
Colafati
GS
Barresi
S
Randisi
F
Talamanca
LF
Genovese
E
, et al.  . 
Description of a novel TUBA1A mutation in Arg-390 associated with asymmetrical polymicrogyria and mid-hindbrain dysgenesis
Eur J Paediatric Neurol
 , 
2013
, vol. 
17
 (pg. 
361
-
5
)

Appendix 1

LIS-Tubulinopathies consortium: Jeanne Amiel, Marie-Claude Addor, Edith Andrini, Tania Attie-Bitach, Magalie Barth, Sophie Blesson, Lydie Burglen, Sebastien Cabasson, Claude Cances, Francoise Chapon, Aurore Curie, Alexandre Datta, Isabelle Desguerre, Patricia Dias, Bérénice Doray, Antoinette Gelot, David Genevieve, Brigitte Gilbert-Dussardier, Fabien Guimiot, Benedicte Héron, Delphine Héron, Marie Line Jacquemont, Frédérique Jossic, Pierre Simon Jouk, Annie Laquerrière, Sebastien Lebon, Jean Marie Lepage, Benoit Lhermitte, Philippe Loget, Laurence Loeuillet, Pascale Marcorelles, Mathieu Milh, Marie-Laure Moutard, Philippe Parent, Sandrine Passemard, Lucile Pinson, Chloé Quelin, Vincent des Portes, Ferechté Razavi, Nicole Revencu, Marlene Rio, Francois Rivier, Caroline Rouleau, Joelle Roume, Marc Tardieu.

Author notes

*These authors contributed equally to this work.
See Appendix 1.