|Home | About | Journals | Submit | Contact Us | Français|
Genes associated with human microcephaly, a condition characterized by a small brain, include critical regulators of proliferation, cell fate, and DNA repair. We describe a syndrome of congenital microcephaly and diverse defects in cerebral cortical architecture. Genome-wide linkage analysis in two families identified a 7.5 Mb locus on chromosome 19q13.12 containing 148 genes. Targeted high throughput sequence analysis of linked genes in each family yielded > 4000 DNA variants and implicated a single gene, WDR62, as harboring potentially deleterious changes. We subsequently identified additional WDR62 mutations in four other families. MRI and postmortem brain analysis supports important roles for WDR62 in proliferation and migration of neuronal precursors. WDR62 is a WD40 repeat-containing protein expressed in neuronal precursors as well as postmitotic neurons in the developing brain and localizes to the spindle poles of dividing cells. The diverse phenotypes of WDR62 suggest central roles in many aspects of cerebral cortical development.
A first pedigree (LIS-900) with microcephaly with simplified gyri (MCSG) from Mexico consists of distantly related parents with three affected children and one healthy sibling (Fig. 1a). The children showed severe microcephaly, developmental delay, and seizures (see Supplementary Note, Clinical Data) but no other syndromic features. A second family (LIS-2600) from Turkey showed similar clinical features and has been described before1 (Fig. 1b). Subsequent to the identification of the responsible gene, four other families were identified with similar clinical features (Fig. 1c–f and Supplementary Note, Clinical Data). Brain MRI of the six families (Fig. 1g–p), revealed small brains, markedly simplified gyral patterns, corpus callosal abnormalities, plus a diversity of additional cortical malformations including polymicrogyria, schizencephaly, and subcortical heterotopia (arrested neurons), sometimes with asymmetry in the same brain (Supplementary Note, Clinical Data and Supplementary Movie 1).
Genome-wide linkage screens of LIS-900 showed a single 11.9 Mb region on chromosome 19q13 (D19S431; D19S217) homozygous in all 3 affected patients and not homozygous in the normal sibling (Supplementary Fig. 1). SNP genotyping of LIS-2600 confirmed the homozygosity and narrowed the interval to ≈ 7.5 Mb (Fig. 2a) (peak combined multipoint LOD score of 3.84, Supplementary Fig. 2). SNP analysis of MC-1400 and MC-1600 confirmed the candidate region but did not reduce it (Fig. 2a).
Because of the high gene density of the linked region, we employed an array capture approach2 followed by high-throughput Illumina sequencing in two affected individuals, LIS-903 and LIS-2602. A custom array was designed with oligonucleotide probes targeting the exons of all 148 predicted UCSC genes in the interval (Fig. 2b) plus non-exonic DNA from a smaller candidate subregion suggested by microsatellite mapping (Supplementary Fig. 2). DNA bound to the array was amplified and used to generate sequencing libraries. 36–40 million paired-end sequence tags (>1.4 Gb of raw sequence) were generated per patient and mapped (hg18) using BWA (Fig. 2c)3. 50–60% of sequence reads were on-target, and average depth of target coverage was 230X and 224X, with 88% and 90% of bases covered by ≥ 10 reads.
Single nucleotide variants and microinsertions/microdeletions were called and filtered for quality and mapping confidence4, revealing >2000 potential sequence variants from each proband (Table 1). Comparison of SNP calls with Affy 5.0 genotype data showed >99.8% concordance. Variants were filtered out if they were present in dbSNP1304 or the 1000 genomes project (1000G), and categorized by their predicted functional effects using GMCC. Of 2310 variants in LIS-903, 262 (11%) were not in dbSNP130 or 1000G, 93 were potentially pathogenic (4.0%); five met both criteria. Of 2570 variants in LIS-2602, 323 (12%) were absent from dbSNP and 1000G, and 99 (3.8%) were potentially pathogenic; seven variants met both criteria. A single gene, WDR62, showed novel, pathogenic mutations in both families, making it a strong candidate (Table 1).
Sanger sequencing confirmed WDR62 mutations in the two families, and identified four additional alleles in four other families (Fig. 3a, b, Supplementary Fig. 3). A homozygous single basepair insertion in exon 30 (c.3936_3937insC) in LIS-2602 produced a frameshift mutation (Fig. 3b). All three affected patients in family LIS-900 showed a homozygous single basepair deletion in exon 4 (c.363delT), creating a frameshift mutation (Fig. 3b). Pedigree analysis confirmed appropriate segregation of WDR62 mutations in both families (Fig. 3b). In LIS-2500, the affected individual showed a homozygous mutation one basepair downstream of exon 8 (c.1043+1G>A), disrupting the highly conserved GT consensus splicing sequence (Fig. 3a, Supplementary Fig. 3). In MC-1400, the affected child was homozygous for a 4 bp deletion after exon 23 (c.2867+4_c2867+7delGGTG) (Fig. 3a, Supplementary Fig. 3) that removes the conserved +4 and +5 positions of the intron and is predicted to disrupt the splice donor5 (reference …CAG|gtgggtgtc…; mutation …CAG|gtg----tc…). In MC-1600, the affected child was homozygous for a deletion of 17 bp in the 30th exon (c.3839_3855delGCCAAGAGCCTGCCCTG), causing a frameshift (Fig. 3a, Supplementary Fig. 3). In PH-16900, all three siblings with microcephaly were homozygous for a missense V65M change (c.193G>A, Fig. 3a, Supplementary Fig. 3) that alters a residue conserved in all vertebrates (Supplementary Fig. 4). An unaffected sibling (PH-16907, with mild speech delay and articulation difficulties but normal head circumference and normal MRI) was homozygous for the wildtype allele (Supplementary Fig. 3). Two additional unaffected siblings were homozygous wildtype and heterozygous, respectively (not shown). None of these mutations were found in over 1000 normal chromosomes.
WDR62 encodes a 1518 amino acid protein with multiple WD40 repeats (Fig. 3a)6, but little is known about its function. A proteomic study identified WDR62 as a binding partner of the centrosomal protein CEP1707, which is of interest given the involvement of other centrosomal proteins in microcephaly8–14. Another report identified WDR62 as a binding partner for c-Jun N-terminal kinase (JNK), suggesting that it potentiates JNK activity15. In situ hybridization with a probe to mouse WDR62 showed widespread expression in the developing brain at embryonic day 14.5 (E14.5) (Fig. 4a), with highest expression in forebrain. Expression was seen in the ventricular zone and the cortical plate (Fig. 4b), consistent with roles in progenitor cells and postmitotic neurons.
WDR62 demonstrates strikingly cell cycle-dependent localization. Immunofluorescence staining of endogenous WDR62 in HeLa cells with an anti-WDR62 antibody revealed punctate, perinuclear expression during interphase, suggesting localization to the Golgi apparatus and confirmed by GM130 immunoreactivity (Fig. 4c, d). In contrast, in HeLa or HEK cells in M phase, WDR62 is found at the spindle poles as demonstrated by double-labeling with gamma-tubulin and dynein (Fig. 4e, f). Close inspection of recombinant HA-tagged WDR62 protein demonstrated subcellular localization closely matching that of CEP170 (Fig. 4g) and surrounding LIS1 (Fig. 4h), a pattern similar to that previously demonstrated for ASPM16,17 and other microcephaly proteins12,13,18,19. Bilguvar and colleagues20 recently described mutations in WDR62 with brain malformations and reported nuclear localization of WDR62 in cortical progenitor cells. While we cannot rule out nuclear localization in some cells, our data, as well as previously published data7 suggest a major centrosomal role.
Postmortem analysis of LIS-2601, an affected sibling in the LIS-2600 family1, suggests roles for WDR62 in neuronal proliferation and migration. The 27-week fetus showed a small cranial cavity enclosing a profoundly small brain (50g, compared to a normal of 127 +/− 20 g for this age and size fetus). The surface of the hemispheres was largely smooth (not shown), with poorly defined Sylvian fissures and few sulci1. The cerebellum appeared remarkably preserved (not shown), whereas the cerebral cortex was severely abnormal (Fig. 5a–g). The outermost cortical layer (layer I) appeared generally normal, but remaining cortical neuronal layers were thinner than normal (Fig. 5b, c). There was an almost complete lack of small- to medium-sized pyramidal neurons in their normal location (layers II and III), suggesting profoundly defective neurogenesis. Neurons located beneath layer I were primarily medium- to large pyramidal neurons, followed by a cell-sparse zone and a third layer of multipolar neurons at the deepest layer of the cortex, an appearance consistent with layers V and VI, respectively (Fig. 5c, d). Neurons in the medium-to-large pyramidal layer retained an immature radial columnar organization and also exhibited abnormal, non-radial clumping (Fig. 5d). In some sections, the pial surface was interrupted by microscopic extrusions of cells into the subarachnoid space, forming heterotopia (Fig. 5e). Heterotopia were also sometimes found in the intermediate zone, arranged in streaks particularly in posterior frontal cortex (Fig. 5f), corresponding to the localization of subcortical heterotopia seen on MRI. There were also clusters (Fig. 5g) of small round cells in the outer subventricular zone (SVZ) suggestive of progenitor cells. These findings support a proliferative defect in brains of patients with WDR62 mutations, potentially affecting the outer SVZ (precursors to upper layer cortical neurons)21 as well as defective neuronal migration, in some ways evocative of the dual roles in proliferation and migration seen for LIS122,23.
Our results suggest that mutations in WDR62 cause microcephaly, simplified gyral pattern, callosal abnormalities, and a wide range of additional cortical abnormalities, including polymicrogyria, schizencephaly, and subcortical heterotopia. WDR62 lies within the MCPH2 locus24–26 a gene-rich locus that has resisted gene identification for > 10 years, and appears to be allelic to MCPH2 (see accompanying paper). The remarkable diversity of cortical malformations associated with WDR62 mutation suggests that it acts at a critical hub of human cerebral development.
All human studies were reviewed and approved by the IRB of Children’s Hospital, Boston, Beth Israel Deaconess Medical Center, and local institutions.
Family LIS-900 underwent a microsatellite genome-wide linkage screen at the Center for Inherited Disease Research (CIDR) using ~400 markers with a 10 cM average density. Genome-wide screens for Family LIS-2600 were performed at Children’s Hospital Boston using ~400 markers in the ABI linkage mapping set MD v2.5 at ~10 cM average density (Applied Biosystems). In addition, families LIS-900, LIS-2600, LIS-2500, MC-1400 and MC-1600 were genotyped on Affymetrix 5.0 SNP chips at the Broad Institute.
Fine mapping was done using polymorphic microsatellite markers from the ABI linkage mapping set HD v2.5 at 5cM average density (Applied Biosystems) along with additional microsatellite markers identified using the UCSC Human Genome Browser. Single and multipoint lod scores were calculated using Allegro, assuming a recessive mode of disease inheritance, full penetrance, and a disease allele frequency of 0.0001. Nucleotide numbers are provided in reference to genomic coordinates (human genome build hg18), coding DNA coordinates (based on RefSeq mRNA NM_001083961.1, where A of the ATG translational start site is designated bp 1 and positioning within an intron are designated relative to the most 3’ base of the nearest upstream coding exon), or protein coordinates (based UniProt accession number O43379), following HGVS guidelines27.
After flanking recombinant markers defined the candidate interval, a tiled microarray (Roche Nimblegen) was designed to target all exons of the 148 genes in the region, as well as most of the nonrepetitive intronic DNA for the 85 genes in the center of the linkage peak. A total of 3.5 Mb of the ~7.5 Mb region was targeted. From each patient, 20ug of genomic DNA was randomly fragmented to 350–400bp, and the ends were repaired and ligated to Nimblegen gSel adaptors. This library was hybridized to the custom array using the manufacturer’s recommended conditions, and nonbound DNA was washed off. Bound DNA was eluted and subjected to PCR amplification. QPCR demonstrated 355X target enrichment for LIS-903 and 1346X for LIS-2601. Nimblegen-captured DNA fragments were then blunted using T4 DNA Polymerase and Klenow fragment using standard conditions, phosphorylated using T4 DNA kinase, and ligated to form concatamers. The concatamers were then randomly fragmented using a Covaris S2 device, and ends were repaired and ligated to Ilumina paired end sequencing adaptors. Ligated products were gel-purified, and PCR-amplified to enrich for adapter-containing fragments. Finally, the purified DNA was subjected to clustering and 40 cycles of paired-end sequencing using the Illumina GA II. Sequencing and subsequent image processing and basecalling were performed per the manufacturer's recommended conditions.
High-throughput sequence analysis was performed according to a customized bioinformatic pipeline for tracking sequence data, aligning reads, calculating coverage, calling variants, annotating variants with respect to functional effect, filtering out benign variation, and flagging candidate rare, pathogenic mutations (available on request). Briefly, BWA version 0.5.73 was employed to align short reads to the human genome (reference build hg18). Consensus and variant basecalls were made with SAMtools version 0.1.7, filtered for quality and mapping confidence, and loaded into a MySQL database for storage and further processing, including annotation of the predicted consequences (noncoding, coding synonymous, coding nonsynonymous, frameshift, splice site) of each variant using GMCC28. Candidate mutations were identified by starting with a list of all variants from the targeted homozygous region, removing those present either in dbSNP130 or a pilot 1 release of the 1000G database, and selecting for coding nonsynonymous, frameshift, or splice site changes. Sequence data were visualized using either the UCSC genome browser29 or the Broad Institute Integrated Genome Viewer. All genomic base positions are presented in reference to human genome NCBI build 36 (hg18).
Coding WDR62 exons and at least 50 basepairs of flanking sequence were PCR amplified and submitted for Sanger capillary electrophoresis (Polymorphic DNA Technologies) in accordance with standard methods. Samples included 384 neurologically normal control samples (Coriell) and 124 unaffected individuals from Middle Eastern families with unrelated disorders.
Evidence for four unvalidated WDR62 variants deposited in dbSNP was also reviewed: rs34840924, rs35055110, rs35938064, and rs34734597 are single basepair insertions in WDR62 coding regions that would be predicted to cause frameshifts, but were unvalidated SNPs ascertained from a single study of five individuals (Submitted batch ID xplore_hum_chr19_2, Submitter Population ID Celera_Donors). Sequencing in 508 controls failed to confirm the existence of these variants.
Plasmid templates for riboprobe synthesis were generated by restriction digestion of mouse WDR62 IMAGE cDNA clone 6405269 to contain a 574 bp sequence corresponding to nucleotide position 4227–4801 of the cDNA. Digoxigenin-labeled sense and antisense RNA probes were synthesized using the DIG RNA Labeling kit (Roche Molecular Biochemicals).
In situ hybridization was performed as described30. Sections were hybridized at 65 °C for 12–16 hours in 1–2 µg/ml of riboprobe and then washed. After blocking for 1 h at room temperature, sections were labeled with anti-DIG Fab-AP antibody (1:2000 dilution, Roche) for 2 h at room temperature and then washed and developed in NBT (nitroblue-tetrazolium-chloride) plus BCIP (5-bromo-4-chloro-indolyl-phosphate) at room temperature until ideal intensity.
HeLa cells were fixed with ice-cold methanol and placed at −20 °C for 5 minutes. The cells were blocked in 1X TBS + 0.15% Tween containing 5% normal donkey serum (v/v) for one hour at room temperature. Cells were then incubated for 3 hours at room temperature in primary antibody diluted in blocking solution. For primary antibodies, we used mouse monoclonal anti-gamma-tubulin (Abcam, ab27074, 1:2000), rabbit polyclonal anti-WDR62 (Bethyl laboratories, A301-560A, 1:500), mouse monoclonal anti-GM130 (BD Transduction Laboratories, 610822), mouse monoclonal anti-dynein heavy chain (Abcam, ab6305, 1:50), rat monoclonal anti-HA (Roche, 3F10, 1:500), rabbit anti-CEP170 (Abcam, ab72505, 1:100), rabbit polyclonal anti-LIS1 (Bethyl laboratories, A300-409A, 1:200). Cells were then washed three times with PBS and incubated with appropriate secondary antibodies against the appropriate species (DyLight, Jackson Labs, 1:200) in blocking solution for 1 hour at room temperature. The cells were then washed twice with blocking solution and incubated with Hoechst 33342 (1:2000, Invitrogen) for 5 minutes at room temperature. Cells were then washed twice with 1X PBS, mounted with Fluoromount-GT, and left at 4 °C overnight to dry. For ICC requiring the use of primary antibodies raised in the same host species, cells were fixed and blocked as normal and then incubated with the first primary antibody for three hours. Cells were then washed three times with blocking solution and incubated with Cy3-conjugated Fab fragments of anti-rabbit IgG (Jackson ImmunoResearch Laboratories) for secondary detection. Cells were then washed four times with blocking solution and incubated with unconjugated Fab fragments of anti-rabbit IgG for two hours. After washing the cells four times with blocking solution, the second primary antibody was applied and detected using the previously described protocol.
A full-length human WDR62 cDNA (IMAGE Consortium clone 4510905) was obtained and used for site-directed mutagenesis (Quikchange, Agilent) to introduce the HA epitope immediately upstream of the stop codon. WDR62-HA was cloned downstream of the CMV promoter, transfected into HEK cells with lipofectamine, and processed for immunocytochemistry at 24 hours.
We thank the many families and researchers who participated in this study, Bernard Chang and Annapurna Poduri for expert review of MRI imaging findings, Sofia Lizarraga for contributions to genetic mapping, Megumi Aita and the In Situ Hybridization Core Facility, University of North Carolina Neuroscience Center, Chapel Hill, North Carolina for performing in situ hybridizations, C. G. Woods for helpful discussions and for sharing data prior to publication, Prof. Dr. Andreas von Deimling from the Institut für Pathologie, Heidelberg, for sectioning of the LIS-2601 brain, Adeline Nicholas and Maryam Khurshid for advice regarding immunohistochemistry, Anthony Hill and Lihong Bu at the MRDDRC (Mental Retardation Developmental Disability Research Center) Imaging Core for assistance with confocal microscopy (Grant: NIH-P30-HD-18655 and Fidelity Foundation), and R. Sean Hill for assistance with linkage analysis and helpful discussions. We are indebted to the members of the Microcephaly Collaborative (MC, Supplementary Table 1) for contributing to the cohort from which these families were drawn. TWY was supported by a NIH T32 grant (T32 NS007484-08), the Clinical Investigator Training Program (CITP) at Harvard-MIT Health Sciences and Technology and Beth Israel Deaconess Medical Center in collaboration with Pfizer, Inc. and Merck and Company, Inc., and the Nancy Lurie Marks Junior Faculty MeRIT Fellowship. GHM was supported by the Young Investigator Award of NARSAD as a NARSAD Lieber Investigator. Research was supported by grants from the NINDS (RO1 NSR01-35129) and the Fogarty International Center (R21 NS061772) to CAW, the Dubai Harvard Foundation for Medical Research, and the Manton Center for Orphan Disease Research. Genotyping services were provided by the Center for Inherited Disease Research (CIDR). CIDR is funded through a federal contract from the National Institutes of Health to The Johns Hopkins University, contract number HHSN268200782096C and NIH N01-HG-65403. Genotyping at Children’s Hospital Boston is supported by the Intellectual and Developmental Disabilities Research Centers (CHB DDRC, P30 HD18655). Genotyping at the Broad Institute is supported by NHGRI. CAW is an Investigator of the Howard Hughes Medical Institute.
Database Accession Numbers
Author ContributionsTWY helped characterize MCSG syndrome, designed and performed targeted capture experiments, generated sequencing libraries, designed bioinformatics pipelines, analyzed sequencing data, identified WDR62 mutations, designed WDR62 in situ expression studies, helped design and analyze WDR62 immunohistochemistry studies, helped analyze the LIS-2601 postmortem specimen, and wrote the manuscript. GHM helped characterize MCSG syndrome, performed initial genome wide linkage studies, and identified the MCSG locus. DJT designed and performed WDR62 immunohistochemistry studies, helped analyze in situ expression studies, and helped analyze the LIS-2601 postmortem specimen. SKS helped characterize the MCSG syndrome, identified additional families, analyzed SNP and microsatellite genotyping, and narrowed the MCSG interval. LF-S helped characterize MCSG syndrome and wrote the initial clinical description of the LIS-900 family. CMS identified the LIS-2600 family and provided clinical information and the LIS-2601 postmortem specimen. MT identified the MC-1400 and MC-1600 families and provided clinical information. MTM identified the PH-16900 family and provided clinical information. BJB organized clinical information and patient samples. JF sequenced WDR62 in families to confirm high throughput sequencing findings and discover additional alleles. CS helped with bioinformatic pipelines, and generated constructs for WDR62 in situ expression studies. WBD referred the PH-16900 family. RDF analyzed the LIS-2601 postmortem specimen. AJB reviewed MRIs for characterization of the MCSG syndrome. CAW directed the overall research, helped analyze the LIS-2601 postmortem specimen, and wrote the manuscript.
Competing financial interests
The authors declare no competing financial interests.
Electronic Database Information
UCSC Human Genome Browser:
1000 Genomes Project:
Human Splicing Finder Version 2.3