Mutations of CASK cause an
X-linked brain malformation
phenotype with microcephaly
and hypoplasia of the brainstem
Juliane Najm1,13, Denise Horn2,13, Isabella Wimplinger1,
Jeffrey A Golden3, Victor V Chizhikov4, Jyotsna Sudi4,
Susan L Christian4, Reinhard Ullmann5, Alma Kuechler6,
Carola A Haas7, Armin Flubacher7, Lawrence R Charnas8,
Go ¨khan Uyanik9, Ulrich Frank10, Eva Klopocki2,
William B Dobyns4,11,12& Kerstin Kutsche1
CASK is a multi-domain scaffolding protein that interacts with
the transcription factor TBR1 and regulates expression of genes
involved in cortical development such as RELN. Here we
describe a previously unreported X-linked brain malformation
syndrome caused by mutations of CASK. All five affected
individuals with CASK mutations had congenital or postnatal
microcephaly, disproportionate brainstem and cerebellar
hypoplasia, and severe mental retardation.
Human microcephaly has previously been associated with defects in
mitosis and DNA repair1. Recently, homozygous inactivation of the
T-box family transcription factor TBR2, which is encoded by EOMES
and putatively regulates neural identity and cortical neurogenesis2, was
associated with microcephaly and other brain malformations in four
individuals from an inbred family3. These data indicate that
disturbances in transcriptional regulation may also be responsible
To further elucidate causes of human microcephaly, we analyzed an
individual (individual 1) referred at 4 years because of congenital and
marked postnatal microcephaly, severe mental retardation and sensor-
ineural hearing loss. Her brain MRI showed reduced number and
complexity of gyri, thin brainstem and severe cerebellar hypoplasia
(Supplementary Table 1 and Supplementary Fig. 1a–c online).
Chromosome analysis showed a paracentric inversion of one X
chromosome: 46,X,inv(X)(p11.4p22.3). This was not inherited from
her mother, and her father could not be studied. We constructed a
physical map using fluorescence in situ hybridization (FISH) for both
breakpoint regions (Supplementary Methods online). The Xp22.33
breakpoint was narrowed to a B20-kb gene-poor region (Supplemen-
tary Table 2 online), and the Xp11.4 breakpoint interrupted the CASK
gene (NM_003688) (Fig. 1a,b and Supplementary Table 2), and poss-
ibly the GPR34 (NM_005300, NM_001033513 and NM_001033514) or
GPR82 (NM_080817) genes located in reverse orientation in CASK
intron 5. CASK encodes a calcium/calmodulin-dependent serine pro-
tein kinase that belongs to the membrane-associated guanylate kinase
(MAGUK) family. Members of this family target to neuronal synapses
and regulate trafficking, targeting and signaling of ion channels. CASK
has been proposed to be a ‘pseudokinase’ and functions as part of large
signaling complexes in both pre- and postsynaptic sites4. However,
CASK also translocates to the nucleus and interacts with the brain-
specific T-box family member TBR1 (ref. 5).
Recently, a heterozygous B3.2-Mb deletion resulting in loss of eight
annotated genes, including EFHC2, NDP and CASK exons 1 and 2 but
not GPR34 or GPR82, was detected by X-chromosome array compara-
tive genomic hybridization (aCGH) in a girl with multiple abnormalities
that partly overlap with those of individual 1 (Supplementary Table 1)6.
Available functional data make CASK an excellent candidate gene for
this phenotype, as Cask mouse mutants have a small brain, abnormal
cranial shape and cleft palate7,8. Further, CASK enhances transcriptional
activity of TBR1, which regulates expression of the extracellular matrix
protein Reelin (Reln), a key player in neuronal migration and lamina-
tion9. Tbr1–/–mice have severe microcephaly and defective cortical
connectivity10, and Reln–/–mice have severe cerebellar hypoplasia9. Both
mutants have a forebrain malformation consisting of abnormal disper-
sion of neurons within and below the cortex, and inverted position of
neurons within the cortex9,10. In humans, mutations in RELN are
associated with the neuronal migration disorder lissencephaly as well
as severe cerebellar and hippocampal hypoplasia11. These data suggest
that CASK has a role in embryonic brain development.
The core phenotype for individual 1 consists of severe mental
retardation, microcephaly and disproportionate pontine and cerebellar
hypoplasia (MIC-PCH). We obtained genome-wide aCGH data (Sup-
plementary Methods) on a clinical basis for two girls with this
phenotype and found copy number losses in Xp11.4 in both (Supple-
mentary Table 1 and Supplementary Fig. 1d–i). In individual 2, we
detected a B740-kb heterozygous deletion encompassing CASK,
GPR34 and GPR82, which we confirmed by FISH (Supplementary
Methods and Supplementary Figs. 2a and 3 online). Neither
parent carried the deletion (data not shown; paternity confirmed). In
individual 3, we found two separate deletions of CASK, including
Received 2 May; accepted 9 June; published online 10 August 2008; doi:10.1038/ng.194
1Institut fu ¨r Humangenetik, Universita ¨tsklinikum Hamburg-Eppendorf, 20246 Hamburg, Germany.2Institut fu ¨r Medizinische Genetik, Charite ´ Universita ¨tsmedizin,
13353 Berlin, Germany.3Department of Pathology, Children’s Hospital of Philadelphia and the University of Pennsylvania School of Medicine, Philadelphia,
Pennsylvania 19104, USA.4Department of Human Genetics, University of Chicago, Chicago, Illinois 60637, USA.5Max-Planck-Institut fu ¨r Molekulare Genetik,
14195 Berlin, Germany.6Institut fu ¨r Humangenetik, Universita ¨tsklinikum Essen, 45122 Essen, Germany.7Experimental Epilepsy Research Group, Neurocenter,
University of Freiburg, 79106 Freiburg, Germany.8Division of Clinical Neuroscience, Department of Pediatrics, University of Minnesota, Minneapolis, Minnesota
55455, USA.9Department of Neurology, University of Regensburg, 93053 Regensburg, Germany.10Sozialpa ¨diatrisches Zentrum, Sta ¨dtisches Klinikum Braunschweig,
38118 Braunschweig, Germany.11Departments of Neurology and12Pediatrics, University of Chicago, Chicago, Illinois 60637, USA.13These authors contributed
equally to this work. Correspondence should be addressed to K.K. (email@example.com).
NATURE GENETICS VOLUME 40 [ NUMBER 9 [ SEPTEMBER 2008 1065
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
B170 kb covering the 3¢ region and B150 kb encompassing the 5¢
region that seem to be interrupted by a stretch of B190 kb of normal
copy number containing GPR34 and GPR82 (Supplementary Fig. 2b
and data not shown). These data suggest a complex rearrangement
such as an inversion-deletion. Notably, the distal deletion breakpoint of
individual 2 and the most telomeric breakpoint in individual 3 are
located in the same region (Fig. 1a and data not shown), suggesting
that nonallelic recombination may be a common mutational mechan-
ism. Quantitative PCR (Supplementary Methods) in individual 3 con-
firmed loss of one copy of CASK exons 2 and 24, and normal copy
number of a probe B92 kb centromeric to CASK (Supplementary
Table 3 online). We also identified several heterozygous SNPs in the
middle of CASK (introns 4–12) as well as in intron 1 of the nearby NYX
gene (data not shown). The parents of individual 3 have normal copy
numbers in this region (Supplementary Table 3; paternity confirmed).
We next selected 46 individuals (33 males and 13 females, including
individuals 2 and 3 discussed above) with MIC-PCH ascertained from
our existing database and colleagues, and analyzed CASK for intragenic
mutations (Supplementary Methods). We identified the heterozygous
transition 1915C4T in exon 21 in individual 4, and the hemizygous
transition 915G4A in exon 9 in individual 5, a boy who died at
2 weeks (Fig. 1c). Both had brain malformations resembling those of
individuals 1–3, although these were more severe in the boy (Fig. 2a,b,
Supplementary Table 1, and Supplementary Fig. 1j–o). The former
mutation was not detected in the parents of individual 4 or in 150
control X chromosomes, and the latter was not found in the mother of
individual 5 or in DNA samples from 515 healthy males (data not
shown; all parental identities confirmed). The 1915C4T mutation
must be pathogenic, as it results in a premature stop codon (R639X).
However, the 915G4A transition is a synonymous mutation (K305)
located in the last nucleotide of exon 9, where it could result in altered
splicing12. We investigated the effect of this sequence change on splicing
using three splice site prediction programs (Supplementary Methods)
that all detected the wild-type donor site. Only one recognized the
mutant sequence as a donor site with reduced splicing efficiency
(Supplementary Table 4 online). Because no RNA was available from
individual 5, we carried out in vitro splicing analyses using ‘minigene’
constructs (Supplementary Methods) and identified skipping of exon
9 in about 20% of the mutant transcripts, suggesting a defect in splicing
(Supplementary Note and Supplementary Table 4 online).
Our finding of heterozygous loss-of-function mutations of CASK in
four girls and a partly penetrant splice mutation in a severely affected boy
suggests that the CASK-associated phenotype belongs to the group of
X-linked disorders with reduced male viabilityoreven in utero lethality13.
However, ‘mild’ (hypomorphic) mutations, such as the synonymous
mutation in CASK exon 9, may be compatible with livebirth in affected
males. We investigated the X chromosome inactivation (XCI) pattern in
genomic DNA extracted from lymphocytes (Supplementary Methods)
in our four female subjects with a heterozygous CASK mutation and
found random XCI (Fig. 1d and Supplementary Note).
Individuals with mutation of CASK show a previously undescribed
and recognizable disease phenotype mainly characterized by severe or
profound mental retardation and distinct structural brain anomalies
(Supplementary Note, Supplementary Table 1 and Supplementary
Fig. 4 online). We examined the brain of the deceased individual 5 in
detail and identified a severely hypoplastic cerebellum, a small
brainstem and abnormalities in cortical and cerebellar layering
(Fig. 2c–h and Supplementary Note). We also analyzed the Cask
knock-in mouse (Cask-KI), a hypomorphic Cask mutant7(Supple-
mentary Note and Supplementary Fig. 5 online). These mice showed
disproportional cerebellar hypoplasia. Taken together, our data suggest
that CASK is required for cerebellar (and forebrain) development in
both human and mouse.
CASK has an important function during neuronal development. In
embryonic neurons of the cerebral cortex, B20% of CASK protein
is present in the nucleus, where it regulates gene expression by
interacting with the TATA-binding protein associated factor TAF9
(CINAP) and the transcription factor TBR1. These proteins form a
Figure 1 Mutations of CASK are associated with
a previously unreported X-linked brain
malformation phenotype. (a) Physical map of
Xp11.4. BAC and fosmid (F–) clones used for
mapping the Xp11.4 breakpoint of the individual
with the Xp inversion (individual 1) and the distal
deletion breakpoint of individual 2 are indicated
by colored bars and names are given. Color code
of BACs: black, mapped distal to both the
deletion and inversion breakpoint; green, spanned
the deletion breakpoint and mapped distal to the
inversion breakpoint; dark blue, located within
the deleted interval and distal to the inversion;
red, spanned the inversion breakpoint; light blue,
mapped proximal to the inversion. Color code of
fosmids: black, mapped distal to the inversion
breakpoint; red, spanned the breakpoint; light
blue, proximal to the breakpoint. Exons of CASK,
GPR34, GPR82 and NYX are indicated by
vertical bars, and the 5¢-3¢ orientation is given.
Breakpoint regions are indicated. (b) FISH with
fosmid clone F-6949F6 on metaphase spread
from lymphocytes of individual 1 yielded split
signals. Wild-type (WT) and derivative (der)
X chromosomes are indicated by arrows.
(c) Partial electropherograms depicting the heterozygous CASK mutation 1915C4T in individual 4 and the hemizygous 915G4A mutation in individual 5,
as well as the respective wild-type sequence. (d) X chromosome inactivation data on genomic DNA from the four female subjects with a CASK mutation
using the androgen receptor assay. Segments from the genotype plots after HpaII restriction enzyme digestion are shown. The study was approved by all
institutional review boards of the participating institutions, and written informed consent was obtained from all participants or their legal guardians.
250 260 270 280
F-3076E8 F-6097D8 F-1393B1
Distal breakpoint region
C TA A A AA A A
A A AGAA AA T TC T
AT TCT T
NGA T TC
1066VOLUME 40 [ NUMBER 9 [ SEPTEMBER 2008 NATURE GENETICS
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
complex that induces transcription of genes containing TBR1 binding Download full-text
sequences, such as RELN4,5. The potential importance of the CASK-
TBR1-RELN signaling cascade is supported by reports of similar
malformations in Tbr1 and Reln mouse mutants in brain regions
where both are expressed10,14. Our data show that this resemblance
applies to CASK as well, especially for brain size, where loss of CASK
in humans and Cask or Tbr1 in mouse results in microcephaly, and
for cerebellar cortex and brainstem, where loss of CASK in humans
and Cask or Reln in mouse results in brainstem hypoplasia and
defective inward migration of granule cells9. However, the cerebellar
hypoplasia observed with loss of CASK in human or mouse cannot be
mediated by TBR1 as it is not expressed in cerebellar cortex2. The
CASK-interacting protein BCL11A (also known as Evi9 or CTIP1)4, a
zinc-finger transcription factor, is a good candidate to substitute for
TBR1 in the cerebellum, as it is expressed in part of the cerebellum
anlage during mouse development15.
The cerebral cortical malformation found in individual 5 consists of
abnormal dispersion of neurons only in the deeper layers V and VI
(Fig. 2d), which are the primary expression domains of TBR1 in
cortex10,14. This malformation might also result from disruption of
CASK-TBR1 interactions, although it is less severe than that seen in
Tbr1 or Reln knockout mice.
A crucial role for CASK during synaptogenesis and a possible
involvement of CASK in stereocilia and retinal function has recently
been proposed (Supplementary Note). Identification and study of
additional transcription factors that bind to CASK may help explain
the pathogenesis of the CASK-associated phenotype, and identify
candidate genes for related disorders.
Note: Supplementary information is available on the Nature Genetics website.
We are grateful to the study participants and their parents. We thank I. Jantke,
L. Schroedter and F. Trotier for skillful technical assistance, S. Fuchs and
K. Ziegler for chromosome analysis, and A. Nowka for help with the in vitro
splicing assay. We also thank T. Su ¨dhof (Howard Hughes Medical Institute) for
providing Cask knock-in mice. This work was supported by a grant from the
Deutsche Forschungsgemeinschaft (KU 1240/3-2 to K.K.), and a grant from the
US National Institutes of Health/National Institute of Neurological Disorders and
Stroke (R01-NS050375 to W.B.D.).
J.N. contributed to mutation analysis, X chromosome inactivation studies, in vitro
splicing assay and manuscript writing. D.H. contributed to analysis of clinical data
and manuscript writing and evaluated individual 2. I.W. contributed to FISH and
mutation analysis and X chromosome inactivation studies. J.A.G. contributed to
brain pathology of individual 5 and manuscript writing. V.V.C. contributed to
analysis of Cask-KI mice. J.S. and S.L.C. interpreted the microarray, designed
quantitative PCR assays and performed and interpreted the results in individual 3.
R.U. contributed to design and production of BAC arrays. A.K. referred and evalu-
ated individual 4. C.A.H. and A.F. contributed to obtaining and building up the
Cask-KI mouse colony, genotyping of mice and preparing brains for morphological
and histological analysis. L.R.C. referred and evaluated individual 5 and arranged
for the brain to be obtained. G.U. interpreted brain scans of individual 4. U.F.
referred and evaluated individual 1. E.K. performed and interpreted array CGH
analysis for individuals 2 and 3 and FISH analysis for individual 2. W.B.D.
contributed to patient ascertainment, clinical evaluation including interpretation of
all brain scans, delineation of the phenotype, obtainment of the brain of individual
5, supervision of the mouse cerebellar studies and manuscript writing. K.K.
designed the study, performed data analysis and contributed to manuscript writing.
Published online at http://www.nature.com/naturegenetics/
Reprints and permissions information is available online at http://npg.nature.com/
1. Cox, J. et al. Trends Mol. Med. 12, 358–366 (2006).
2. Hevner, R.F. et al. Neurosci. Res. 55, 223–233 (2006).
3. Baala, L. et al. Nat. Genet. 39, 454–456 (2007).
4. Hsueh, Y.P. Curr. Med. Chem. 13, 1915–1927 (2006).
5. Hsueh, Y.P. et al. Nature 404, 298–302 (2000).
6. Froyen, G. et al. Hum. Mutat. 28, 1034–1042 (2007).
7. Atasoy, D. et al. Proc. Natl. Acad. Sci. USA 104, 2525–2530 (2007).
8. Laverty, H.G. et al. Genomics 53, 29–41 (1998).
9. D’Arcangelo, G. et al. Bioessays 20, 235–244 (1998).
10. Hevner, R.F. et al. Neuron 29, 353–366 (2001).
11. Hong, S.E. et al. Nat. Genet. 26, 93–96 (2000).
12. Cartegni, L. et al. Nat. Rev. Genet. 3, 285–298 (2002).
13. Franco, B. et al. Curr. Opin. Genet. Dev. 16, 254–259 (2006).
14. Bulfone, A. et al. Neuron 21, 1273–1282 (1998).
15. Leid, M. et al. Gene Expr. Patterns 4, 733–739 (2004).
V / VI /WM
ML EGLML EGL
Figure 2 Brain imaging and pathologic features in a boy with the CASK
915G4A mutation (individual 5). (a,b) Compared to a normal control
brain (a), magnetic resonance images in individual 5 (b) show a thin and
unmyelinated corpus callosum (angled arrows) and severe hypoplasia of the
brainstem (horizontal arrow) and cerebellum (vertical arrow). (c,d) The frontal
cortex (d) is moderately disorganized and mildly thickened compared to an
age-matched control (c). Cortical laminae I-IV appear normal, whereas layers
V and VI merge together and show a vaguely nodular organization. These
layers subtly merge into the white matter (WM), which differs from the
well-defined border between layer VI and WM in the control cortex.
(e,f) The pons is markedly reduced in size in the affected boy (f) compared
to a control (e) because of loss of neurons in the basis pontis (borders
defined by black lines). (g,h) The cerebellum (h) shows poorly formed,
shallow and unbranched folia when compared to an age-matched control (g).
In contrast to the normal external granular cell layer (EGL) in the age-
matched control cerebellum (g, bottom panel), the EGL is abnormally thick
in the cerebellum of individual 5 (h, bottom panel), and the internal granular
cell layer (IGL) virtually absent (compare g and h). Purkinje cells are evenly
spaced, forming a single layer in normal brain (arrows in g, bottom panel).
In contrast, Purkinje cells in individual 5 are not seen in a single row (arrows
in h, bottom panel). The bottom panels of g and h also show hypercellularity
in the molecular layer (ML) and increased thickness of the EGL in individual
5 (h, bottom panel). Scale bar in c (lower right corner): 250 mm (c,d),
500 mm (e–h), 50 mm (bottom panels of g and h).
NATURE GENETICS VOLUME 40 [ NUMBER 9 [ SEPTEMBER 2008 1067
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics