Definition, clinical classifications, and clinical features

Microcephaly (MIC) is classically defined as a head circumference more than two standard deviations (SD) below the mean for age and sex. However, most experts define clinically significant (ie, diagnostically useful) MIC as 3 SD or more below the mean, as many children with a head OFC between 2-3 SD are developmentally normal.^19,21,22 About 15% of children referred to child neurologists for evaluation of developmental disabilities have MIC, which is often associated with comorbidities such as epilepsy, autism, and other birth defects.¹⁹ Genetic and non-genetic factors can cause an abnormally small brain size. Genetic causes including a rapidly growing list of single gene disorders, copy number abnormalities (ie, deletions/duplications), and large chromosomal rearrangements (eg, triosomies) (Figure 1, Table I). Non-genetic or environmental insults include in utero or congenital infections (eg, congenital Zika virus, cytomegalovirus, and herpes simple virus, among others), teratogenic exposures (eg, alcohol, drugs such as hydantoin or aminopterin) and vascular insults (eg, placental insufficiency).¹⁹ However, knowledge regarding the underlying causes of MIC remains limited in scope, and the diagnostic and therapeutic approaches for children with MIC are not uniform, as highlighted by the Zika virus epidemic in 2015-2016.^1,2,5,6

A confusing number of terms have been used to clinically classify MIC including “primary microcephaly” which refers to early-onset (ie, typically congenital) MIC with no other syndromic features including (classically) no additional brain abnormalities besides a simplified cortical gyral pattern.²² “Autosomal recessive primary MIC” (or MCPH) was a clinical diagnostic term used for the earliest familial reports and loci of MIC, later defined by several key features including (a) severe congenital MIC (OFC ≤4 SD below the mean at birth), (b) ID in the absence of more severe neurological problems including spasticity, seizures or developmental regression and (c) absence of other syndromic features including body growth abnormalities or additional brain malformations.^23-25 To date, mutations in 23 genes have been identified as causative of primary microcephaly (Table I). However, the remarkable emerging genetic and clinical heterogeneity of MIC syndromes has challenged traditional classifications and uncovered a wide range of syndromes in which individuals can have overlapping features including brain size abnormalities, body growth abnormalities and complex malformations of cortical development.^26,27 Indeed, mutations of the same genes have now been identified in expanding spectra involving these features. Examples include MIC and cancer phenotypes due to mutations in DNA Damage Response genes, such as Nijmegan Breakage syndrome, and isolated MIC and Microcephalic Osteodysplastic Primordial Dwarfism (MOPD) phenotypes due to centrosomal defects.²⁶

When the brain is small, the most common abnormality identified on brain imaging (typically magnetic resonance imaging, MRI) is a simplified cortical gyral pattern that is typically diffuse but can disproportionately affect the frontal lobes (ie, microcephaly with a simplified gyral pattern).²⁸ By definition, this is the most common brain MRI abnormality identified in children with most forms of MIC. However, additional brain abnormalities can co-occur with MIC, partly depending on the underlying etiology, including cortical malformations such pachygyria, polymicrogyria, dysgyria (eg, due to mutations of the tubulin genes), callosal abnormalities (including most notably a small or thin corpus callosum), as well as abnormalities of the cerebellum, brainstem, basal ganglia, white matter among others.²⁸ Therefore, thorough examination of neuroimaging features associated with MIC is often diagnostically very helpful, as shown in Table I.

Genes and pathways underlying human MIC

In examining the large body of literature on genes and pathways underlying human—particularly congenital—MIC, several important functional classes of defects can be clearly discerned.²⁶ The earliest identified MIC-associated genes and loci were critical centrosomal and cell cycle defects that continue to account for the largest fraction of MIC molecular diagnoses worldwide. These include centrosome-specific proteins (eg, CEP135, CENPJ, PCNT, MCHP1), spindle-associated proteins (eg, ASPM, WDR62) and, less commonly, kinetochore associated defects (eg, CENPE). By far, mutations in the majority of these genes are inherited in an autosomal recessive fashion and are, therefore, more widely seen in in-bred or consanguineous populations.²⁹ Mutations of the ASPM (Abnormal Spindle Microtubule Assembly) gene alone are estimated to account for 10%-40% of the causes of autosomal recessive congenital MIC (or MCPH).³⁰ Mutations of this broad class of genes have been well-documented to cause a wide range of cell cycle defects including abnormal centrosome structure and function, abnormal spindle-kinetochore assembly, and abnormal microtubule structure and function.²⁶ More recently, mutations of a critical class of microtubule-associated genes including those belonging to the tubulin family (eg, TUBA1A, TUBB2B, TUBG1, TBCD, among many others) and the kinesin protein (eg, KIF11, KIF14) have been identified to cause complex MIC syndromes characterized by a wide range of brain malformations including cortical malformations, callosal abnormalities, and dysgenesis of the brainstem, basal ganglia, and thaiami, with wide variability among affected individuals.^31,32 In contrast to the most common centrosomal defects underlying MIC, mutations of these genes are typically de novo, with consequently a low-recurrence risk.³²

The second-largest group of human functional defects associated with MIC is defects in the DNA Damage Response (DDR) pathway. The DDR is a signaling cascade critical for repairing DNA strand repairs due to endogenous and exogenous insults. The large number of associated mutations, genes, and syndromes known to be associated with congenital MIC that are related to DDR and centriole/spindle organization support the fundamental roles these pathways play in neuronal development and genomic stability.^33,34 The intimate relationship between these pathways and cancer is further supported by the association of many of these MIC-associated genes with cancer phenotypes in humans, as occurs with mutations in NBN, NHEJ1, XRCC2, XRCC4 among many others, for example.³⁵ Collectively, the overwhelming majority of genes associated with congenital MIC are critical nodes in very early stages of neuronal development and have yielded substantial insights into their fundamental roles.³⁶ In contrast, acquired causes of MIC, such as Zika virus (ZIKV) related MIC, for example, appear to typically affect later stages of neuronal progenitors, inducing apoptosis and defects of differentiation rather than severe neuroproliferative defects, although the mechanisms of ZIKV-related MIC continues to be under intense study.^37-39

Definition, clinical classifications and clinical features

From DeMyer's work, MEG or “large brain” has been defined as an oversized and overweight brain that exceeds the mean by two SD for age and gender, consistent with standard medical practice which uses a normal range of ±2 SD for growth parameters.⁴⁰ Similar to MIC, clinically significant MEG is now defined as 3 or more SD above the mean, as most individuals with mildly large head size (+2-3 SD) have been found to have normal development.⁴⁰ The largest documented brain sizes in human reach more than 10 SD above the mean, reported more recently in individuals with mutations of the phosphatidyl inositol 3-kinase (PI3K)-AKT-MTOR pathway including AKT3.⁴¹ Abnormalities of brain size can occur early on during fetal development and manifest at or shortly after birth (ie, congenital MEG), or evolve more slowly during infancy and early childhood (ie, postnatal MEG).²⁶ Clinically significant MEG has been classified into anatomic and metabolic subtypes, a classification scheme first proposed by DeMyer that continues to be of clinical use as these two large MEG groups differ substantially in their underlying genetic causes, pathomechanisms, clinical features and medical management.⁴⁰ Metabolic (or neurometabolic) syndromes associated with MEG include a wide range of disorders characterized by abnormal accumulation of metabolic substrates, typically associated with consequent neuronal hypertrophy with ballooning of the cytoplasm, and fundamentally less cellular hyperplasia.^40,42 Anatomic MEG includes a wide range of disorders characterized by a multitude of cellular defects including increased cell size and/or number.

From the early literature, brain overgrowth has been proposed to occur in several distinct “non-syndromic” clinical forms. The earliest is familial MEG (also termed “idiopathic MEG”) that has been reported and proposed to be the most common form of MEG seen in children.⁴⁰ Individuals with this clinical “entity” classically had mild MEG (OFC +2-3 SD above the mean) with familial recurrence reported in up to 50% of children in large series.^43-46 Importantly, a subset of these children with familial MEG had neurodevelopmental problems including epilepsy and tone abnormalities, suggesting that brain overgrowth, even if presenting as an isolated feature, can be associated with neurological consequences.^45,47-49 Further, neuroimaging on many children with this reported MEG subtype identified ventriculomegaly, enlarged extra-axial spaces, or hydrocephalus requiring shunting, overlapping with syndromic forms of MEG (such as the PI3K-AKT-MTOR related MEG syndromes) and suggesting a more complex process underlying head overgrowth that may involve both brain overgrowth and cerebrospinal fluid expansion causing hydrocephalus.^47,49-51

MEG has also been reported as a very common clinical feature in children with autism spectrum disorders (ASD), with several studies suggesting that it is the most common physical manifestation seen in ASD.⁵² Mutations of the PTEN gene have been identified in ~10% to 20% of megalencephalic children with ASD, which has led to the delineation of the “macrocephaly-autism syndrome” as a subtype of the PTEN-hamartoma tumor related disorders.^53-56 However, wider scale genomic testing has now identified mutations of many other genes associated with MEG-ASD as well, including most notably MTOR, PPP2R5D, CHD8, among others, strongly suggesting that the etiologies of MEG and ASD are much more genetically heterogeneous but likely linked to some of the same critical cellular pathways.⁵⁷

Genes and pathways underlying human MEG

Disorders causing an abnormally large brain size (ie, true MEG) most often occur as a result of a genetic insult that may cause isolated brain overgrowth (eg, PTEN related hamartoma tumor syndrome) or combined brain and body overgrowth as part of a generalized overgrowth disorder (eg, Sotos syndrome, PIK3CA related overgrowth disorders).^58,59 Most of these disorders are caused by defects of single genes, although a growing number of copy number abnormalities have also been identified in overgrowth syndromes (Table I). Mutations of key nodes in several signaling pathways regulating cellular growth and proliferation cause a wide range of human brain overgrowth disorders including the PI3K-AKT-MTOR, RAS-MAPK-ERK pathways, among others.^60-63 Importantly, a growing number of human overgrowth disorders are caused by mutations in epigenetic regulators such as NSD1, EZH2, DNTM3A, among others.⁶⁴ A broad overview of the involved genes and pathways highlights that most of these are within critical intermediary signaling pathways that are highly pleiotropic in function and, unlike MIC, the cellular effects of these pathway perturbations on neuronal development are not as well-understood (Figure 1).

The most common genetic defects causing human MEG localize to the PI3K-AKT-MTOR pathway. Proper MTOR signaling is critical in key aspects of brain development including neuronal progenitor maintenance, differentiation, migration, synaptogenesis, and regulating protein translation.^65-68 Therefore, mutations of different nodes causing functional upregulation of this pathway are associated with a wide range of neurodevelopmental phenotypes including MEG, ID, ASD, and epilepsy in humans and animal models. These nodes include a growing number of genes located upstream (PIK3CA, PIK3R2, PTEN), midstream (AKT3) and downstream (TSC1, TSC2, MTOR, CCND2, TBC1D7, RHEB, and STRADA) within the pathway.⁶⁹ Besides diffuse brain overgrowth (MEG), a growing number of mosaic (postzygotic) mutations of this pathway are associated with focal malformations of cortical development with intractable epilepsy, including focal cortical dysplasia (FCD), hemimegalencephaly (HMEG) and dysplastic megalencephaly (DMEG).^70-72 The functional endpoint of all mutations associated with diffuse or focal MEG phenotypes is pathway hyperactivation. The proposed mechanisms by which PI3K-AKT-MTOR pathway mutations are believed to cause brain overgrowth include neuronal hypertrophy (eg, PTEN- and MTOR- related brain overgrowth), and increased neuronal proliferation (eg, CCND2- related brain overgrowth).^73-76 Specifically, mutations of the AKT3 gene have been associated with most severe MEG, and further shown to have very strong associations with increased intracranial volume (ICV) in a large meta-analysis of reactome gene sets in relationship to brain size, along with other MTOR pathway genes.^41,69

Recently, mutations of a new family of genes—the PP2A phosphatase family including specifically PPP2R5D—have been identified in children with congenital onset MEG, autism, and hypotonia. This class of phosphatases is believed to have a negative regulatory effect on the PI3K-AKT-MTOR pathway. Therefore, PI3K-AKT-MTOR dysregulation is believed to underlie the brain overgrowth of PPP2R5D-related disorders as well.^77-79 Finally, abnormalities of the evolutionarily-conserved partitioning defective protein complex, including PARD3, PARD6, and atypical protein kinase C (aPKC), that regulate asymmetric cell division of Radial Glial Progenitors (RGPs) have been proposed as a mechanism underlying MEG and heterotopia in humans.^80,81 However, no mutations of these genes have been identified in human MEG, to date. A comprehensive review of additional genes associated with human MEG are further provided in Table I.

Microcephaly

Among the most clinically useful distinguishing features of MIC is its' age of onset (ie, congenital vs postnatal). Congenital MIC, if isolated (ie, not associated with any other major structural anomalies) is typically autosomal recessive in nature (also known as Primary MIC or MCPH) caused by mutations in cell cycle genes such as ASPM most notably, among others. Congenital MIC in association with other structural defects is most often syndromic and likely extremely genetically heterogeneous. Therefore, etiologies falling under this category include microdeletion or microduplication syndromes that can be identified by chromosomal microarrays, or a growing number of single gene disorders, many of which can be delineated based on their associated key clinical features. If clinical delineation is not possible, then consideration for wide-scale genomic testing (eg, exome sequencing) is recommended. It is important to remember the non-genetic causes of congenital MIC including congenital viral infections, vascular insults or injuries and environmental exposures (eg, fetal alcohol syndrome). Postnatal MIC, on the other hand, is extremely genetically heterogeneous, as MIC occurs as a secondary features in numerous neurodevelopmental syndromes characterized by epilepsy, ID and ASD; many of which are caused by de novo mutations. Therefore, clinical delineation of the specific syndrome is recommended. If a specific syndrome is not identified based on the history and physical examination, then consideration of wide scale genomic testing (eg, exome sequencing) is similarly recommended. Detailed recommendations for clinical evaluations of children with MIC have been previously published.¹⁹

Megalencephaly

Similar to MIC, careful review of overall body growth measurements, in conjunction with a detailed neurological examination and/or brain imaging, are very useful in evaluating a child with MEG. Importantly, when a child presents with a large head size, it is imperative to rule out other important causes of macrocephaly such as hydrocephalus or progressive ventriculomegaly. True and isolated brain overgrowth — MEG — is seen most commonly with single gene disorders, most notably PI3K-AKT-MTOR pathway related disorders.^59,72 Given the likelihood of low-frequency postzygotic (mosaic) mutations in MEG disorders, careful selection of molecular diagnostic approaches needs to be made using a method that generates good quality, high depth data (eg, ultra deep targeted capture). Syndromic forms of MEG require a consideration of microdeletion-microduplication syndromes (using chromosomal arrays), followed by careful and thorough clinical assessments to delineate specific overgrowth syndromes. In the presence of diffuse MEG with multiple other congenital anomalies, exome sequencing may be utilized to efficiently and rapidly identify the underlying genetic etiology, given the emerging genetic heterogeneity of MEG.⁵⁸ In spite of wider integration of high-quality and high-depth NGS methods in clinical practice, the genetic causes of a substantial fraction of MEG in children remain unidentified, suggesting the involvement of other genes/loci and/or low-level mosaicism challenging detection using standard molecular methods.⁶⁹