Introduction
Primary autosomal recessive microcephaly (MCPH) delineates a genetically heterogeneous and rare subgroup of congenital microcephalies characterized by a pronounced reduction of brain volume, particularly of the neocortex, simplified gyral pattern and intellectual disability [
1,
2]. Homozygous mutations of the Cyclin-dependent kinase 5 regulatory subunit-associated protein 2 gene,
CDK5RAP2 (OMIM*608201), were identified in 2005 as a cause for MCPH type 3 (MCPH3, OMIM#604804) [
3]. To date, three different mutations have been identified: two in three Pakistani families and one mutation in a Somali patient: (i) a nonsense mutation in exon 4 (c.246T > A, p.Y82X), (ii) an A to G transition in intron 26 (c.4005-15A > G, p.R1334SfsX5) introducing a new splice acceptor site, a frame shift and a premature stop codon, and (iii) a nonsense mutation in exon 8 (c.700G > T, p.E234X) [
3‐
5]. All three mutations have been proposed, but not shown, to lead to a truncated protein and a loss of CDK5RAP2 function.
CDK5RAP2 is associated with the centrosome, microtubuli and Golgi apparatus, is enriched in neural progenitors within the ventricular and subventricular zone of the immature brain, can be also detected in glial cells and early neurons, and is strongly downregulated with brain maturation [
6,
7]. One current model for the microcephaly phenotype caused by
CDK5RAP2 mutation invokes a premature shift from symmetric to asymmetric neural progenitor cell divisions with a subsequent depletion of the progenitor pool and a reduction in the final number of neurons, and decreased cell survival [
6,
8]. Underlying mechanisms include a deregulation of the role of CDK5RAP2 in centrosome function, spindle assembly and/or response to DNA damage [
6,
8]. Despite considerable interest in MCPH as a neural stem cell defect and window into the control of neurogenesis in humans, the underlying pathomechanisms have not been definitively established and specifically for MCPH3, no detailed radiological descriptions of patients or functional analyses in patient samples have been reported to date.
In the present study, we report a novel CDK5RAP2 mutation and describe for the first time in detail the clinical, radiological and cellular phenotype in two MCPH3 patients of European descent. We are thereby able to attribute the microcephaly phenotype in MCPH3 at least partially to a mitotic spindle defect and centrosome disorganization.
Material and methods
Patients
Informed consent was obtained from the parents of the patients for the molecular genetic analysis, the publication of clinical data, magnetic resonance images (MRI) and studies on immortalized lymphocytes (LCLs). DNA was extracted from EDTA blood samples using the Illustra BACC2 DNA extraction kit (GE Healthcare, Munich, Germany). Samples from microcephaly patients and controls were used in this study with approval from the local ethics committees of the Charité and the Freiburg University (approval nos. EA1/212/08 and 494/11, respectively).
Haplotype analysis using microsatellite markers
Six microsatellite markers were selected for each of the MCPH1 to 7 and PNKP loci, so that three markers were located on each side of each gene. The markers flanking the CDK5RAP 2 gene were: CHLC.GGAA23B10, D9S258, D9S2152, D9S103, D9S116 and D9S1823. PCR was performed with 1 ng patient DNA and primer pairs in which the forward primer was always labeled with 6-FAM. PCR fragments were resolved by capillary electrophoresis on an ABI 3100 sequencer. Fragment analysis was performed using GeneScan software (Applied Biosystems, Foster City). Haplotypes were constructed in the family by inspection of the microsatellite fragment lengths.
PCR and DNA sequencing
Thirty-eight coding exons of the CDK5RPAP2 gene and at least 50 bp of the intronic, exon-flanking sequence were analyzed by PCR (Taq Polymerase, Qiagen, Hilden, Germany), and cycle sequencing using the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit Version 1.1 (Applied Biosystems, Darmstadt, Germany). Capillary electrophoresis was performed using an ABI 3100 sequencer (Applied Biosystems, Foster City, CA, USA). Sequence data were analyzed using SeqPilot DNA sequence analysis software (JSI, Kippenheim, Germany). The database sequence NM_018249 for the CDK5RAP2 gene was used as reference, and primers were developed in our laboratory (available on request).
Establishment of Ebstein-Barr virus-transformed lymphocytes and culture
Ebstein-Barr virus-transformed lymphocytes (LCLs) were established according to the protocol published by Neitzel et al. 1986 [
9]. Non-adherent LCLs were cultured in RPMI 1640 with L-Glutamine (Invitrogen, Darmstadt, Germany) supplemented with 20% v/v fetal bovine serum (Invitrogen) and 1% v/v penicillin-streptomycin (Sigma-Aldrich, Taufkirchen, Germany).
Immunocytology
For fixation, cells were plated on Poly-L-lysine (Sigma-Aldrich) coated coverslips, cultured for 30 min in standard conditions, and incubated in 37°C PFA 4% for 10 min prior to rinsing with phosphate buffered saline (PBS 1×). Coverslips were further incubated at room temperature (RT) in staining buffer (0,2% gelatin, 0,25% Triton X-100 in PBS 1×) for 20 min and subsequently in 10% donkey normal serum (DNS) in staining buffer for 30 min for blocking. Coverslips were incubated overnight at 4°C with primary antibodies in the staining buffer containing 10% DNS followed by an incubation with the corresponding secondary antibodies for 2 h at RT. Nuclei were labeled with 4’,6-diamidino-2-phenylindole (DAPI, 1:1000, Sigma-Aldrich). Fluorescently labeled cells were analyzed and imaged by a fluorescent Olympus BX51 microscope with the software Magnafire 2.1B (2001) (Olympus, Hamburg, Germany), and all images were processed using Adobe Photoshop. The anti-CDK5RAP2 antibody (HPA035820; 1:200) utilized in this study recognizes amino acids 1307–1390 of CDK5RAP2 which is unique for the human CDK5RAP2 protein sequence (accession no. NP_060719.4, Uniprot Q96SN8). Further primary antibodies were as follows: mouse anti-γ-tubulin (T5326, Sigma-Aldrich, 1:5000), mouse anti-alpha-tubulin (T9026, Sigma-Aldrich; 1:1500), mouse anti-pericentrin (ab28144, Abcam; 1:1000), mouse anti-acetylated alpha-tubulin (T6793, Sigma-Aldrich; 1:1000), mouse anti-GM130 (610823, BD Biosciences; 1:1000).
Protein extracts for Western blots were isolated from LCLs by homogenization in radio-immunoprecipitation assay (RIPA) buffer containing 1 mM phenylmethylsulfonyl fluoride (PMSF; Sigma-Aldrich) and 1 protease inhibitor cocktail tablet per 10 ml RIPA buffer (Complete Mini; Roche Diagnostics, Mannheim, Germany), 20 min incubation on ice and centrifugation at 4°C for 10 min at 3000 g and for 20 min at 16000 g. Protein concentrations were determined using a bicinchoninic acid (BCA) based assay, according to the instructions of the manufacturer (BCA Protein Assay Kit; Pierce Biotechnology, Rockford, IL, USA). Protein extracts (30 μg per sample) were denaturated in Laemmli sample loading buffer at 95°C for 5 min, separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretically transferred in transfer buffer in a semi-dry fashion using Trans-Blot SD Semi-Dry transfer cell (Bio-Rad, Munich, Germany) onto nitrocellulose membrane (Bio-Rad, Munich, Germany). The membranes were incubated for 1 h at RT in blocking buffer (TBS-T 1x with 5% bovine serum albumin (BSA)), rinsed three times with TBS-T (1x) for 8 min each at RT on a shaker and then incubated overnight at 4°C with rabbit anti-CDK5RAP2 (1:200, HPA035820, Sigma-Aldrich; also verified with antibody from Abnova PAB17507, 1:200), mouse anti-γ-tubulin (1:5,000) or mouse anti-CHK1 (1:1000, Sigma-Aldrich) antibodies. After incubation with the corresponding secondary antibodies donkey anti-rabbit (1:2000; Amersham Biosciences, Freiburg, Germany) and goat anti-mouse (1:10,000; Dako, Hamburg, Germany), the immunoreactive proteins were visualized using a technique based on a chemiluminescent reaction. The gel pictures were obtained with a Bio-Rad imager (Bio-Rad laboratories, Munich, Germany). Western blot experiments were run in triplicate.
Discussion
In the present study, we have identified the novel nonsense mutation c.4441C > T (R1481*) in the
CDK5RAP2 gene in a homozygous constellation in two boys of Italian descent with primary microcephaly (Figures
1 and
2). We thereby, for the first time, provide detailed clinical and radiological information on MCPH3 patients of European descent. The siblings suffer from congenital microcephaly, intellectual disability, speech deficit, a tic disorder and severe behavioral problems. Further tests did not reveal any significant hearing impairment or epilepsy as a cause for the speech deficit. Therefore, although sensorineural hearing loss has been reported in two patients with mutations in
CDK5RAP2[
5,
15], this is not a consistent finding in MCPH3. Both patients had microcephaly, simplified gyral pattern of the cerebral cortex, with shallow sulci anteriorly and deep sulci parietally and posteriorly, and corpus callosum hypogensis on cMRI. There was no particular evidence of reduced white matter volume in the patients, despite the fact that
CDK5RAP2 is expressed in glial cells of the developing rodent brain. Why the white matter is not more severely affected in MCPH remains unclear. Future studies will need to address the question as to what extent white matter disease also contributes to brain size reduction in MCPH patients. It is unclear whether these clinical and radiological features are also present in the previously reported three pedigrees of Pakistani descent with MCPH due to homozygous
CDK5RAP2 mutations.
In addition to the brain, we recently reported that CDK5RAP2 is widely expressed in various organs of newborn mice and human fetuses with high CDK5RAP2 mRNA and protein levels in the thymus and the kidney [
7]. Moreover, it has been reported that in the MCPH3 murine model‚ “
Hertwig’s anemia” mice display defects in multiple organs including the thymus and also have a hematopoietic phenotype (hypoproliferative anemia, leucopenia, predisposition to hematopoietic tumors) [
16]. In other MCPH subtypes, individual patients have been reported with short stature (especially in MCPH1 and MCPH5 [
17‐
19]), early puberty, renal agenesis and polycystic kidneys [
17]. As this point warrants further investigation in patients, we investigated the clinical phenotype of our patients in detail with respect to multi-organ involvement. Short stature was detected in both patients up to an age of two years, but thereafter normalized in patient 1 for whom detailed data were available. This normalization of height after infancy (in contrast to the pattern of head growth) is a disease feature that has been reported similarly in patients with
ASPM gene mutations [
18]. There was no evidence of further organ involvement or malignancy, specifically no anemia or leucopenia and no kidney or thymus abnormality.
The three homozygous mutations in the
CDK5RAP2 gene reported so far, 246T > A in exon 4, 700G > T in exon 8 and 4005-15A > G in intron 26, have been proposed, but not shown, to lead to truncated proteins of 82 (Y82*), 234 (E234*) and 1334 (R1334Sfs*5) amino acids, respectively, and a loss of CDK5RAP2 function (full length protein 1893 amino acids; Figure
1). While the first and second mutant protein should lack most of the
CDK5RAP2 transcript except for the N-terminus including a part of the γTuRC-binding domain or the N-terminus including the complete γTuRC-binding domain and a part of the SMC-domain, respectively, the third protein should lack the C-terminus of CDK5RAP2, especially the c-terminal SMC domain as well as the pericentrin and the Golgi binding sites. The homozygous nonsense mutation reported here, 4441C > T in exon 30, is predicted to lead to a truncated protein of 1481 amino acids (R1481*). The resulting CDK5RAP2 protein in our patients should lack parts of the second SMC domain as well as the pericentrin and the Golgi binding sites (Figure
1). No studies on patient specimen exist that shed light on the effect of the reported
CDK5RAP2 gene mutations. We recently reported a high
CDK5RAP2 expression in proliferating progenitors of the germinal matrix and early (not mature) neurons as well as glial cells in the neocortex of murine embryos and human fetuses [
7]. This is in concordance with results of neuroimaging studies in MCPH patients due to non-
CDK5RAP2 mutations demonstrating a reduced brain volume that affects especially the neocortex [
17,
18]. Based on results from
in vivo and
in vitro studies, the human MCPH phenotype is considered to be the result of a premature shift from symmetric to asymmetric neural progenitor-cell divisions (with a subsequent depletion of the progenitor pool) as well as of a reduction in cell survival [
6,
8].
To study the effect of the reported CDK5RAP2 gene mutation on cell proliferation in our patients, we studied EBV-transformed lymphocytes (LCLs) from both of our patients and from controls. Here, CDK5RAP2 localized to the centrosomes during each stage of the cell cycle in controls but was absent from patient cells, when assessed via immunocytology and western blots (Figure
3). The latter finding of CDK5RAP2 levels below detection levels in cells of our patients indicates that very little or no protein is present secondary to nonsense-mediated decay of the mutated transcript. In contrast to
Cdk5rap2 shRNAi studies performed on mouse tissues [
20], we detected a failure of the centrosomal protein γ-tubulin to localize properly at the centrosome, while total γ-tubulin protein levels were normal in patient cells (Figure
3, Additional file
1: Figure S1). Pericentrin, which interacts with CDK5RAP2 through defined protein domains [
21], was not altered in its localization in patient LCLs (Figure
5). This result is in line with those of Buchman et al. 2010 [
21] who concluded from their studies in murine tissues that the centrosomal recruitment of pericentrin is not dependent upon Cdk5rap2, while the converse is true. Despite the predicted loss of the C-terminal Golgi domain in mutant CDK5RAP2, the Golgi apparatus could be visualized normally using immunostaining with GM130. However, Golgi fragmentation appeared to occur earlier during mitosis and had disappeared by prophase (Figure
4). Further, we observed an unfocused and disorganized mitotic microtubule assembly, a decrease in spindle pole distance and a trend towards more multipolar spindle poles as well as chromosome misalignment in patient cells (Figure
6). These results, although generated in lymphocytes and not neural progenitors, suggest that spindle defects and a disruption of centrosome integrity could play a role in the development of microcephaly in MCPH. On the other hand, they also underline the fact that, at least in the cells studied, despite a lack of normal CDK5RAP2, a centrosomal structure can still be formed, microtubuli can still be nucleated to the centrosomes and cells can still divide. Since microcephalin and pericentrin regulate mitotic entry via centrosome-associated Chk1 [
22] and Chk1-downregulation has been demonstrated in mutant Cdk5rap2 cells [
14], we analyzed CHK1 levels in CDK5RAP2 mutant and control cells. Although slightly reduced in both patients, there was not a significant decrease in total CHK1 levels in patients cells (Figure
7).
Brain size at birth is largely determined by the relative rates of proliferation and cell death. By highlighting the clinical, radiological and cellular phenotype of MCPH3 patients, we offer a further glimpse into how a disruption of the CDK5RAP2 gene may impact on the development in humans. Further analysis of patient samples provides a means to investigate processes that cause MCPH and to verify mechanisms described in other model systems and in settings where animal models are neither sufficient nor satisfactory.
Acknowledgements
The authors thank Julia König, Horst von Bernuth, Margret Oberreit, Angela Hübner, Katrin Köhler, Marcus Lettau, Ottmar Janssen, Angela Steiert, Christine Zeschnigk, Ulrike Schneider, Elke Jantz-Schuble and Bernd Rösler for discussions and technical assistance. We are grateful to Dr. Wolfgang Brunk and Prof. Peter Winkler for supplying MR images of the patients.
Funding
This work was supported by the German Research Foundation (SFB665), the Sonnenfeld Stiftung and the Berliner Krebsgesellschaft e.V.
Competing interests
The authors declare no competing interests in the preparation or publication of the data in this manuscript.
Authors’ contributions
AM and DMR were responsible for the project conception and wrote the manuscript. LI, NK and ON performed the lymphocyte analysis, generated figures and proofread the manuscript. KM and KS performed genetic analysis and compiled clinical data. HR and MB attended the patients and provided clinical data. All authors read and approved the final manuscript.