REPORT
Mutations in NOTCH1 Cause Adams-Oliver Syndrome
Anna-Barbara Stittrich,1,9 Anna Lehman,2,9 Dale L. Bodian,3,9 Justin Ashworth,1 Zheyuan Zong,2
Hong Li,1 Patricia Lam,2 Alina Khromykh,3 Ramaswamy K. Iyer,3 Joseph G. Vockley,3 Rajiv Baveja,4
Ermelinda Santos Silva,5 Joanne Dixon,6 Eyby L. Leon,7 Benjamin D. Solomon,3,8 Gustavo Glusman,1
John E. Niederhuber,3,10,* Jared C. Roach,1,10 and Millan S. Patel2,10,*
Notch signaling determines and reinforces cell fate in bilaterally symmetric multicellular eukaryotes. Despite the involvement of Notch
in many key developmental systems, human mutations in Notch signaling components have mainly been described in disorders with
vascular and bone effects. Here, we report five heterozygous NOTCH1 variants in unrelated individuals with Adams-Oliver syndrome
(AOS), a rare disease with major features of aplasia cutis of the scalp and terminal transverse limb defects. Using whole-genome
sequencing in a cohort of 11 families lacking mutations in the four genes with known roles in AOS pathology (ARHGAP31, RBPJ,
DOCK6, and EOGT), we found a heterozygous de novo 85 kb deletion spanning the NOTCH1 50 region and three coding variants
(c.1285T>C [p.Cys429Arg], c.4487G>A [p.Cys1496Tyr], and c.5965G>A [p.Asp1989Asn]), two of which are de novo, in four unrelated
probands. In a fifth family, we identified a heterozygous canonical splice-site variant (c.743�1 G>T) in an affected father and daughter.
These variants were not present in 5,077 in-house control genomes or in public databases. In keeping with the prominent develop-
mental role described for Notch1 in mouse vasculature, we observed cardiac and multiple vascular defects in four of the five families.
We propose that the limb and scalp defects might also be due to a vasculopathy in NOTCH1-related AOS. Our results suggest that
mutations in NOTCH1 are the most common cause of AOS and add to a growing list of human diseases that have a vascular and/or
bony component and are caused by alterations in the Notch signaling pathway.
Adams-Oliver syndrome (AOS [MIM 100300]) is a rare
developmental disorder with an incidence of approxi-
mately 1 in 225,000 individuals and is defined by the
combination of aplasia cutis congenita of the scalp vertex
and terminal transverse limb defects (e.g., amputations,
syndactyly, brachydactyly, or oligodactyly).1,2 In addition,
vascular anomalies, such as cutis marmorata telangiecta-
tica congenita, pulmonary hypertension, portal hyper-
tension, and retinal hypovascularization, are recurrently
observed.3 Congenital heart defects have been estimated
to be present in 20% of individuals with AOS; reported
malformations include ventricular septal defects, anoma-
lies of the great arteries and their valves, and tetralogy
of Fallot.4–6 Both familial and sporadic AOS occurrences
have been described, and genetic heterogeneity is evident
given that AOS has been shown to be caused by muta-
tions in four different genes: heterozygous mutations
in Rho GTPase activating protein 31 (ARHGAP31 [MIM
610911]) or recombination signal binding protein for
immunoglobulin kappa J region (RBPJ [MIM 147183]) or
biallelic mutations in dedicator of cytokinesis 6 (DOCK6
[MIM 614194]) or EGF-domain-specific O-linked N-
acetylglucosamine transferase (EOGT [MIM 614789]).7–10
Collectively, mutations in these four genes have not
been shown to account for more than 10% of individuals
with AOS.
1Institute for Systems Biology, Seattle, WA 98109, USA; 2Department of Medi
Columbia, Vancouver, BC V6H 3N1, Canada; 3Inova Translational Medicin
Neonatal Associates, Inova Health System, Falls Church, VA 22042, USA;
4050-111, Portugal; 6Genetic Services, Wellington Hospital, Capital & Coast D
and Metabolism, Children’s National Medical Center, Washington, DC 20010
System, Falls Church, VA 22042, USA9These authors contributed equally to this work10These authors contributed equally to this work
*Correspondence: [email protected] (J.E.N.), [email protected] (M.S
http://dx.doi.org/10.1016/j.ajhg.2014.07.011. �2014 by The American Societ
The American
We identified heterozygous variants in NOTCH1 (MIM
190198) as an additional cause of AOS in multiple families.
Phenotypic analysis of affected individuals with NOTCH1
mutations indicates a high rate of vascular anomalies
(Table 1); in contrast, neither RBPJ nor ARHGAP31 muta-
tions have yet been shown to associate with abnormal
vascularization (Table S1, available online). Both of the
known recessive genes, however, have shown such an
association: homozygous EOGT mutations have been
found in AOS individuals with septal defects, patent duc-
tus arteriosus, and brain infarcts, and homozygous or com-
pound-heterozygous DOCK6 mutations have been found
in AOS individuals with congenital heart defects andmark-
edly abnormal blood vessels (data not shown).7,9,11
We used whole-genome sequencing (WGS) to analyze
14 AOS-affected individuals from 12 unrelated families,
11 of which are of European descent and one of which is
of mixed European-Asian descent. In addition to analyzing
the entire genome, we also screened for variants in genes
with a known role in AOS and found one individual with
compound-heterozygous mutations in DOCK6 (data not
shown). Here, we report on five families in which we found
mutations in NOTCH1 (Figure 2). Four of these families
were recruited with informed consent through a study
protocol (H08-02077) that was approved by the institu-
tional review board (IRB) at the University of British
cal Genetics and Child and Family Research Institute, University of British
e Institute, Inova Health System, Falls Church, VA 22042, USA; 4Fairfax5Pediatric Gastroenterology Service, Centro Hospitalar do Porto, Porto
istrict Health Board, Wellington 6242, New Zealand; 7Division of Genetics
, USA; 8Department of Pediatrics, Inova Children’s Hospital, Inova Health
.P.)
y of Human Genetics. All rights reserved.
Journal of Human Genetics 95, 275–284, September 4, 2014 275
Table 1. Clinical Characteristics of the AOS-Affected Individuals
Clinical Characteristic
Individual
1-II-3 2-II-1 2-III-2 3-II-1 4-II-1 5-II-1
Aplasia cutis of the scalp þ þ þ þ þ forme fruste
Terminal transverse limb defects þ þ þ þ þ þ
Cutis marmorata þ � � þ þ þ
Intracranial vascular lesions � � � þ þ þ
Pulmonary hypertension � � � � þ �
Cardiac malformation narrow pulmonaryarteries
? ? pulmonary valve stenosis distal narrowingof the aortic arch
?
Otherfeatures
NA NA NA thrombosis of the portal veinand splenorenal shunt, spasticdiplegia, intellectual disability
thrombosis of thesagittal sinus
NA
The following abbreviations are used: ?, not assessed; and NA, not applicable.
Columbia. A fifth family was recruited at the Inova
Translational Medicine Institute through an IRB-approved
protocol with informed consent (20121680). The clinical
phenotypes of affected individuals are summarized in
Table 1.
The proband of family 1 (1-II-3) has aplasia cutis conge-
nita affecting the occiput and marked cutis marmorata.
At birth, white vesicles (or areas of focal calcinosis cutis)
were present at the tips of the fingers, which were other-
wise well formed. The toenails were hypoplastic and
dystrophic bilaterally (Figure 1), some vesicles were pre-
sent, and subtle, semicircumferential constriction of the
skin and soft tissue was present at the bases of several
toes. An echocardiogram was normal apart from mild
narrowing of the branch pulmonary arteries. Cardiac
assessment was repeated in the first year, and there
appeared to be no hemodynamic consequence to this
finding. Renal ultrasound and brain MRI were normal.
The vesicles on the hands and feet would occasionally
rupture to release a thick chalky substance; this substance
was radio-opaque, suggesting that it was of bony origin.
Cutis marmorata persisted into early childhood, and at
the last follow-up at age 6 years, height was on the tenth
percentile, the branch pulmonary arteries were normal,
the main pulmonary artery was dilated but stable for years,
the vesicles or calcinosis cutis were persistent but nonerup-
tive on the feet, and moderate micrognathia with a
retruded tongue caused intermittent airway obstruction
during sleep. The family history is negative for aplasia
cutis and terminal transverse limb defects.
In family 2, a father and daughter are affected by AOS.
The daughter (2-III-1), age 20 months at enrollment, was
born with severe aplasia cutis of the scalp (Figure 1), which
was complicated by recurrent hemorrhage during a
lengthy healing process. She has hypoplastic toes on the
left foot and nail hypoplasia of the second and third
toes. Neurodevelopmental milestones have been age
appropriate. Her father (2-II-1) was also born with a
cutaneous and bony defect affecting two-thirds of his
276 The American Journal of Human Genetics 95, 275–284, Septemb
cranium, brachydactyly of the right hand, and terminal
transverse limb defects of both feet, including soft-tissue
syndactyly of hypoplastic toes (Figure 1). Bony in-growth
never fully bridged the cranial defect. There is no
other family history of aplasia cutis or limb defects. Two
maternal first cousins of the father reportedly had cardiac
septal defects but were not available for enrollment.
The clinical phenotype of 3-II-1 from family 3 was
described in detail by Silva et al.12 In summary, this now
14-year-old boy was affected by aplasia of the scalp, under-
lying cranial defect, cutis marmorata, brachysyndactyly
of the toes, hypoplastic fingernails, inguinal and umbilical
hernias, mild pulmonary stenosis, and hypoplasia of
the intrahepatic portal venous tree. He developed portal
venous thrombosis in infancy, which led to symptomatic
portal hypertension; a mesenteroportal shunt was placed,
but this was also complicated by thrombosis. On the sec-
ond day following this surgery, he suffered an ischemic
stroke. Since then, no further thrombotic events have
occurred. Brain imaging demonstrated evidence of water-
shed infarctions, and he has mild intellectual disability
and spastic diplegia. The family history is negative for evi-
dence of AOS.
The family 4 proband (4-II-1) was born at term with
severe aplasia cutis affecting most of the scalp superior to
the ears, as well as the posterior neck. She had bilateral
prominent, tortuous scalp vessels, truncal cutis marmor-
ata, and bilateral toe hypoplasia with absent toenails. Neu-
roimaging (brainMRI and intracranial magnetic resonance
angiogram and venogram) on day of life (DOL) 1 showed
small focal areas of bilateral parietal and left frontal
white-matter acute infarction; the superior sagittal sinus
was patent but had a focal abnormality felt to be consistent
with a partial superior sagittal sinus thrombosis with
recanalization. Repeat neuroimaging 1 week later revealed
evolving biparietal and left frontal lobe infarcts, near-
complete sagittal sinus thrombosis, and biparietal cortical
venous thromboses (Figure 1). Serial neuroimaging
showed stabilization and improvement of the thromboses
er 4, 2014
Figure 1. Clinical Vignettes for AOS Indi-viduals from Families 1, 2, and 4(A–C) Family 1 proband (1-II-3): scarredaplasia cutis lesion affecting the scalp atage 5 years (A), calcific deposits in thesubcutaneous tissue of the distal firsttoe (arrow, B), and terminal transversedefect of the toes and cutis marmorata ininfancy (C).(D–F) Family 2 proband (2-III-1): distalhypoplasia of the digits of the left hand(D) and toes of the feet (E) and aplasia cutisof the scalp (F).(G and H) Family 2 father of the proband(2-II-1): aplasia cutis of the scalp (G) andterminal transverse defects of both feet (H).(I–K) Family 4 proband (4-II-1): MRI of thebrain on day of life (DOL) 9 shows areasof infarcts (solid arrows) and an area ofpartial thrombus (outlined arrow) fromaxial sections on diffusion weighted MRI(I and J) and axial T2-weighted MRI (K).
over the next several months. She did not undergo anti-
coagulation therapy, and there was no clinical or labora-
tory-based evidence for thrombophilia. Echocardiograms
showed pulmonary hypertension, estimated to be almost
half of systemic pressure, on DOL 1, and mild mitral valve
annulus hypoplasia (annulus diameter of 8.7 mm) with no
stenosis on DOL 4. A trileaflet aortic valve with mild distal
aortic arch narrowing felt to be of no hemodynamic conse-
quence and multiperforate patent foramen ovale with
insignificant shunting were observed on DOL 9, and the
pulmonary hypertension resolved by DOL 10. No other
congenital heart defects were detected. Renal ultrasound
was normal except for mildly echogenic kidneys felt to
be secondary to dehydration. The family pedigree, having
both European and Asian ancestry, is negative for any
similar anomalies.
The clinical phenotype of the proband of family 5
(5-II-1) at age 24 years has been previously described by
The American Journal of Human Gen
Vandersteen and Dixon.13 She has a
family history consistent with auto-
somal-dominant AOS. Her deceased
sister had severe aplasia cutis, cutis
marmorata, nail aplasia of the toes,
brachydactyly, tricuspid valve incom-
petence, absence of the inferior
medullary velum of the cerebellum,
and hypoplasia of the dentate nuclei.
She died of pulmonary hypertension
at 3 years of age. Their deceased father
had aplasia cutis, cutis marmorata,
myopathy, and epilepsy. The proband
was noted at birth to have short
digits and toes and a macular heman-
gioma around the circumference
of her anterior fontanelle. Other
features, including generalized cutis
marmorata telangiectatica congenita and periventricular
and gray-white-matter-junction hyperintense lesions on
brain MRI, became apparent over time. Her intelligence
is normal.
For WGS, two different platforms and analysis pipelines
were used. For families 1, 2, 3, and 5, genomic DNA was
extracted from peripheral blood or saliva (DNA Genotek
kit, Orasure Technologies). Paired-end library preparation,
WGS, alignment to the reference genome (NCBI human
genome assembly build 37), and variant calling were per-
formed by Complete Genomics Inc. (CGI). Further variant
annotation and analysis were performed with Ingenuity
Variant Analysis software (QIAGEN) and the Family Geno-
mics Toolkit, including Kaviar, a software program that
estimates variant frequencies by taking into account not
only the 1000 Genomes data set14 and the NHLBI Exome
Sequencing Project Exome Variant Server (ESP6500) but
also publically available personal genomes and exomes
etics 95, 275–284, September 4, 2014 277
and in-house genomes from the Institute for Systems
Biology (ISB) and combines all available data to compute
integrated variant frequencies.15 To identify candidate
variants, we filtered for single-nucleotide variants (SNVs)
and small indels that had CGI sequencing quality scores
R 35 and had no reported variants in any data encom-
passed by Kaviar. We considered both recessive (hemizy-
gous, homozygous, or compound heterozygous) and
dominant (heterozygous) models. We searched genome-
wide for individual variants or an increased gene burden
of rare disrupting variants (frameshift, splice site, or stop
gain or loss) in all or a subset of families. No variants or
genes were found to be significantly enriched. Therefore,
we relaxed our thresholds to include missense variants
and restricted our analysis to the genes with known roles
in AOS—ARHGAP31, DOCK6, EOGT, and RBPJ—as well as
a set of 273 genes reported by Ingenuity Knowledge Base
to be in direct interaction or relationship with one of the
four AOS-associated genes. We also assessed copy-number
variations (CNVs) and structural variants. For CNVs, we
not only used the CGI annotation but also developed
an algorithm that normalizes the depth of sequencing
coverage in 1 kb bins to the median coverage depth of
comparable in-house genomes and uses a hidden Markov
model to segment the genome by observed ploidy. We
identified sequences that differ from the expected diploid
state and computed frequencies of such events by using
ISB’s in-house database of 5,077 genomes (including
3,816 CGI-sequenced genomes and 1,261 Illumina-
sequenced genomes).
WGS of family 4 was performed on an Illumina HiSeq
2000. Genomic DNA of the three study participants
was isolated from peripheral blood (QIAGEN DNAeasy
kit, QiaSymphony DNA). Paired-end libraries were gener-
ated, quantified, and quality checked with the protocols
recommended by Illumina. WGS, alignment to the
reference genome (NCBI human genome assembly build
37), and variant calling were performed by Illumina’s
FastTrack WGS service. SNVs and indels with ‘‘PASS’’
annotations and quality scoresR 30 in proband, maternal,
and paternal genomes were merged with gVCFtools
v.0.16, annotated with snpEff v.3.3,16 and then queried
with GEMINI v.0.6.4.17 We used modified versions of the
GEMINI tools both to identify variants in known AOS-
associated genes in the proband and to conduct genome-
wide analysis for variants on the basis of possible inheri-
tance patterns. Only variants that were fully called in
the trio and that had minor allele frequencies < 0.5%
(for de novo variant queries) or < 1% (for other queries)
in ESP6500 and 1000 Genomes were considered. Variants
were filtered for those predicted by snpEff to be of
medium or high impact or those annotated as possibly
disease associated for any disease by ClinVar (v.12/30/13)
or HGMD Professional (v.2013.2, BIOBASE). Variants pre-
sent with the same inheritance pattern in Inova’s internal
database of 659 healthy families were excluded. Parent-
offspring relationships were confirmed by GRAB.18
278 The American Journal of Human Genetics 95, 275–284, Septemb
Candidate variants were confirmed with bidirectional
Sanger sequencing, and where available, additional family
members were screened by Sanger sequencing.
We identified candidate variants in NOTCH1 (on the
basis of the RefSeq transcript NM_017617.3 and NCBI hu-
man genome assembly build 37) in five families (Figure 2).
The proband of family 1 (1-II-3) was found to have a
heterozygous 85 kb deletion of the 50 region of NOTCH1;
it includes part of the promoter and the entire first
exon (chr9: 139,439,620–139,524,480; ClinVar accession
number SCV000172278). We identified this deletion via
sequence-coverage analysis at 20 bp resolution. This
finding is supported by an analysis of heterozygous sites
around the NOTCH1 50 end in family 1, which showed a
run of homozygosity in the proband, but not in other fam-
ily members (Figure S1), and by quantitative real-time PCR
within the deleted region, which revealed half the amount
of template in the proband (Figure S2). Because the dele-
tion is not present in the unaffected parents or two unaf-
fected siblings, de novo occurrence is apparent. In the
two affected members from family 2 (2-II-1 and 2-III-1),
we identified a heterozygous NOTCH1 canonical splice-
site variant, c.743�1G>T, at chr9: 139,414,018 (ClinVar
SCV000172277). Sanger sequencing revealed that the
variant is present in neither the unaffected mother (2-II-
2) nor the unaffected brother (2-III-2) of the proband.
The unaffected paternal grandparents of the proband
were unavailable for genetic testing. In the proband of
family 3 (3-II-1), we found a heterozygous NOTCH1
missense variant, c.1285T>C (p.Cys429Arg), at chr9:
139,412,360 (ClinVar SCV000172279). It has a SIFT score
of 0.0 (damaging) and is rated as ‘‘probably damaging’’
by PolyPhen-2. Sanger sequencing of the two unaffected
parents revealed that the variant occurred de novo. Anal-
ysis of family 4 revealed a heterozygous de novo NOTCH1
missense variant, c.4487G>A (p.Cys1496Tyr), at chr9:
139,399,861 (ClinVar SCV000172281) in the proband
(4-II-1). The variant has a SIFT score of 0 (damaging) and
a PolyPhen-2 score of 1 (probably damaging). The
family 5 proband (5-II-1) has a heterozygous NOTCH1
missense mutation, c.5965G>A (p.Asp1989Asn), at chr9:
139,393,681 (ClinVar SCV000172280). It has a SIFT score
of 0.25 (tolerated) and is rated ‘‘probably damaging’’ by
PolyPhen-2. No DNA was available from the deceased
affected father and sister.
The Notch signaling pathway is highly conserved in
Bilateria and governs cell-fate decisions by maintaining
progenitors in an undifferentiated state, initiating dif-
ferentiation, or maintaining quiescence of differentiated
cells.19,20 The mammalian family of Notch single-pass
transmembrane receptors includes NOTCH1, NOTCH2,
NOTCH3, and NOTCH4. Signaling is initiated by binding
of one of the Jagged or Delta-like ligands—JAG1, JAG2,
DLL1, DLL3, or DLL4—to the extracellular domain of
one of the Notch receptors. NOTCH1 has an extracellular
domain with 36 epidermal-growth-factor (EGF)-like re-
peats, 21 of which are potentially calcium binding, and
er 4, 2014
Figure 2. Pedigrees of the AOS-Affected Families and the Respective NOTCH1 Variants(A) Affected males are represented by solid squares, affected females are represented by solid circles, and index individuals are indicatedby arrows. For individuals tested byWGS and/or Sanger sequencing, the genotypes are noted below in boxes; WT indicates the wild-typegenotype, and the five different mutations observed in NOTCH1 (RefSeq accession number NM_017617.3) are numbered as follows: (1)the 85 kb deletion, (2) c.743�1G>T (splice site), (3) c.1285T>C (p.Cys429Arg), (4) c.4487G>A (p.Cys1496Tyr), and (5) c.5965G>A(p.Asp1989Asn).(B) For family 1, analysis of sequencing coverage is shown and indicates a 85 kb deletion (chr9: 139,439,621–139,524,478) that overlapswith the promoter region and coding sequence of NOTCH1 (chr9: 139,388,896–139,440,238) in the proband. For families 2–5, Sangersequencing chromatograms show the respective variants (red asterisks).(C) The exon-intron structure ofNOTCH1 is shown. The red triangles indicate our reportedmutations inNOTCH1: (1) the 85 kb deletion,(2) c.743�1G>T (splice site), (3) c.1285T>C (p.Cys429Arg), (4) c.4487G>A (p.Cys1496Tyr), and (5) c.5965G>A (p.Asp1989Asn). Bluedots indicate previously reported NOTCH1 variants associated with isolated cardiac defects.
three Lin-12 NOTCH repeats (LNRs). The Notch intracel-
lular domain (NICD) contains an RBPJ-associationmodule,
seven ankyrin repeats, and a C-terminal proline-, glutamic-
acid-, serine-, and threonine-rich domain. Ligand binding
leads to two proteolytic cleavage steps—one of which is
mediated by ADAM-family metalloproteases and the other
of which is mediated by the g-secretase complex—that
release NICD to translocate to the nucleus. There, NICD
activates target-gene transcription in cooperation with
the DNA-binding protein CSL (named after RBPJ, Su(H),
and LAG-1) and a coactivator of the Mastermind-like fam-
ily (MAML1–MAML3).19,20 Noncanonical Notch signaling
also occurs and contributes to the pleiotropic effects of this
pathway.
The 85 kb deletion found in the proband of family 1 (1-
II-3) removes the promoter and the first exon, suggesting
that it causes disease through haploinsufficiency. We did
not find any other largeNOTCH1 deletions in our database
of 5,077 genomes. In a 2010 computational study, the
probability that NOTCH1 function is susceptible to hap-
loinsufficiency was predicted to be 0.957.21 The canonical
splice-site variant (c.743�1G>T) in family 2 (2-II-1 and
2-III-1) disrupts the exon 5 acceptor splice site, probably
ablating exon 5 and leading to a 41 amino acid in-frame
deletion affecting the EGF-like repeats 6 and 7. These two
EGF repeats have not been implicated with ligand binding
but might be involved in glycosylation-mediated regula-
tion of NOTCH1 activity.22 Another possible consequence
The American
of this mutation is that a cryptic splice site might be re-
cruited. Because this would most likely introduce a stop
codon toward the 50 end of the gene, nonsense-mediated
decay would probably lead to loss of function.23 For the
three missense variants in families 3–5, we aimed to esti-
mate the functional impact on NOTCH1 folding and
stability by using the macromolecular modeling software
Rosetta.24–26 We modeled these variants on the following
template structures from the Protein Data Bank (PDB):27
p.Cys429Arg (PDB ID 2VJ3),28 p.Cys1496Tyr (PDB ID
3L95),29 and p.Asp1989Asn (PDB ID 3V79).30 For both
wild-type and altered proteins, we sampled the side-chain
and backbone conformations of each protein domain at
least 100 times by using a Monte Carlo procedure to mini-
mize the estimated total free energy of folding for the pro-
tein (Figure 3). We calculated the change in the free energy
upon mutation (DDG) as the estimated free energy of
folding of the altered protein minus that of the wild-type
protein. Positive DDG values indicate destabilization of
the protein upon substitution, and values exceeding
þ3.0 kcal/mol indicate significantly reduced protein stabil-
ity.31 The two cysteine substitutions, p.Cys429Arg (3-II-1)
and p.Cys1496Tyr (4-II-1), are predicted to strongly desta-
bilize the protein domains in which they occur, primarily
as a result of the disruption of disulfide bridges in both
cases. The p.Cys429Arg variant (3-II-1) resides in the
calcium-binding EGF-like repeat 11. EGF-like repeats 11
and 12 were previously shown to be required for Delta
Journal of Human Genetics 95, 275–284, September 4, 2014 279
D EF
A B C
HG I
Figure 3. Modeling the Effect of the NOTCH1 Variants p.Cys429Arg, p.Cys1496Tyr, and p.Asp1989Asn on Protein Structureand EnergeticsFor each variant, the left and middle panels show 20 energy-minimized structures from Rosetta modeling (A, D, and G represent thewild-type proteins, and B, E, and H represent the simulated substitutions). In (C), (F), and (I), the predicted change in the free energyof protein folding and stability upon substitution (DDG) is shown for each variant (blue, wild-type; pink, variant protein). Higher valuesindicate decreased domain stability.
ligand binding in Drosophila Notch.32 The variant could
impair calcium binding, which is a prerequisite for ligand
binding.33 The p.Cys1496Tyr variant (4-II-1) occurs within
the negative regulatory region (NRR) of the extracellular
region of NOTCH1 in the second LNR domain. The NRR
sterically inhibits processing of NOTCH1 in the absence
of ligand stimulation. Thus, destabilization of this domain
could increase constitutive Notch signaling and lead to a
gain of function. Modeling of the third missense substitu-
tion, p.Asp1989Asn (5-II-1), suggests a more subtle effect
280 The American Journal of Human Genetics 95, 275–284, Septemb
on protein stability. The affected aspartic acid is involved
in a bipartite-charged hydrogen-bonding interaction with
the backbone nitrogen-hydrogen atoms of p.Asp2020,
which is located in a neighboring ankyrin repeat. The
asparagine substitution is predicted to be capable of main-
taining an uncharged hydrogen bond here. However, the
different strength and geometry of an uncharged versus
charged hydrogen bond34 could destabilize or perturb the
regional domain structure. Although p.Asp1989 itself
does not appear to directly interact with Notch binding
er 4, 2014
partners, the ankyrin repeat domain is involved in the
binding of NOTCH1 to RBPJ and transcriptional coacti-
vators of the Mastermind-like family.30 Alignment of
the three missense-variant positions with the use of
MUSCLE35 and Jalview36 shows that they are strictly
conserved in all vertebrates and that the two cysteines
are conserved in Bilateria (Figure S3).
Here, we report that NOTCH1 mutations are a frequent
and possibly the largest single genetic cause of AOS. We
found five NOTCH1 mutations in 5 of 12 unrelated
families affected by AOS (42%). None of these variants
were found in over 10,000 control genomes or exomes.
For three of the five families, unaffected parents were
available for analysis, and in all three families, the
NOTCH1 variants occurred de novo (Figure 2). In one fam-
ily affected by autosomal-dominant AOS, a canonical
splice-site variant in NOTCH1 segregated with the disease
in the father and daughter. The missense variants each
alter an amino acid that is strictly conserved in vertebrates.
Further supporting our contention that mutations in
NOTCH1 cause AOS, previous reports of other individuals
with AOS showed autosomal-dominant mutations in the
gene encoding the main NOTCH1-binding partner, RBPJ,
and recessive mutations in EOGT, which encodes a puta-
tive NOTCH1 glycosylase.7,8,11 Our findings confirm that
Notch signaling alterations play a prominent role in the
development of AOS.
The majority of our reported AOS-affected individuals
with NOTCH1 variants show cardiac or vascular defects.
In addition to observations that Notch1 loss-of-function37
or gain-of-function38 mutations in mice cause major
vascular defects, several hereditary cardiovascular disorders
have unveiled the pivotal role that Notch signaling plays
in cardiovascular development and homeostasis. Examples
include bicuspid aortic valve with calcification and/or
thoracic aortic aneurysm (MIM 109730; caused by
NOTCH1mutations); defects of the left ventricular outflow
tract (caused by NOTCH1 mutations); Alagille syndrome
(ALGS), involving predominantly right-sided congenital
cardiovascular malformations (ALGS1 [MIM 118450],
caused by JAG1 mutations; ALGS2 [MIM 610205], caused
by NOTCH2 mutations); Hajdu-Cheney syndrome, fea-
turing cardiac defects (MIM 102500; caused by NOTCH2
mutations); and cerebral autosomal-dominant arteriopa-
thy with subcortical infarcts and leukoencephalopathy
(MIM 125310; caused byNOTCH3mutations).39–41 Our re-
sults add to this growing list a wide variety of AOS-related
vascular defects, including right- and left-sided congenital
heart defects, pulmonary hypertension, portal hyperten-
sion, cutis marmorata, venous ectasia, and thrombophilia.
Still, the AOS phenotype qualitatively differs from
those previously reported with NOTCH1 variants, and it
is unclear at present how the mutations we report cause
AOS. The identified variants might lead to loss of function
by preventing transcription (the 85 kb deletion), pre-
venting translation (the splice-site variant c.743�1G>T),
or destabilizing NOTCH1 (the three missense variants
The American
c.1285T>C [p.Cys429Arg], c.4487G>A [p.Cys1496Tyr],
and c.5965G>A [p.Asp1989Asn]). Given that Notch recep-
tors bind ligands as tetramers,33 it is also possible that some
AOS-associated NOTCH1 missense mutations exert domi-
nant-negative effects by destabilizing normal tetramer
formation and thus disrupt ligand binding. It is also
conceivable that some of the NOTCH1 mutations induce
a gain of function; for example, c.4487G>A (p.Cys1496-
Tyr), which resides in the negative regulatory region,
might reduce the restraint on ligand-independent sig-
naling. There is also a possibility that the 85 kb deletion
induces utilization of an alternative downstream promoter
and transcription start site that skips the extracellular
domain of NOTCH1 and produces a constitutively active
NICD.42 The possibility that both loss- and gain-of-func-
tion mutations in NOTCH1 cause AOS could be explained
if Notch-signaling strength is the critical element in
determining normal development of the vasculature.43
An alternate hypothesis for the different phenotypes
observed with AOS-related NOTCH1 mutations is that
other genetic, epigenetic, or environmental factors shape
the NOTCH1 mutant phenotype.
Several mechanisms have been proposed to explain the
limb reduction and scalp defects in AOS. Vascular obstruc-
tion by thromboses during embryonic development
was suggested as a cause of terminal transverse limb
defects,44,45 and we report that two of our five probands
experienced multiple thromboses. Examination of blood
vessels from an affected individual with hepatoportal
sclerosis led to the suggestion of endothelial cell dysfunc-
tion as the root cause of AOS.46 On the basis of the wide-
spread appearance on autopsy of denuded or reduplicated
internal elastic laminae in AOS blood vessels (both signs
of intimal damage), aswell as blood vessel stenosis or dilata-
tion in association with exuberant or poor, respectively,
vascular smooth muscle cell coverage, we suggest pericyte
dysfunction as the basic pathophysiologic mechanism in
AOS.47 Consistent with this suggestion, a critical role for
Notch signaling in the vascular smooth muscle cell precur-
sors of developing mouse limbs was shown by Chang and
colleagues.48 Using a tetracycline-inducible dominant-
negativeMastermind transgenicmodel, theydemonstrated
that inhibition ofNotch signaling during a specificwindow
in late gestation leads to defective vasculogenesis and hem-
orrhage at the scalp and tips of all developing limbs in
mouse embryos. Necrotic, hemorrhagic digits have also
been observed in a preterm neonate with AOS, in keeping
with the notion that some of the limb defects in AOSmight
be due to aNOTCH1-related vasculopathy.49 Underbranch-
ing of vasculature trees, irregular vascular smooth muscle
cell coverage, tortuous ectatic vessels, and a general paucity
of small to medium blood vessels are key features of the
AOS vasculopathy12,47 and might also explain other symp-
toms observed with AOS, including CNS microbleeds,
retinopathy and visual impairment due to failed retinal
vascularization, portal hypertension, ischemic bowel dis-
ease, placental insufficiency, and pulmonary hypertension.
Journal of Human Genetics 95, 275–284, September 4, 2014 281
Our data clearly implicate NOTCH1 variants as a cause
of AOS. There are numerous reported instances of the
co-occurrence of aplasia cutis congenita with congenital
cardiovascular anomalies, which seem to form a contin-
uum within the variability in clinical expression of AOS,
and NOTCH1 was previously predicted as a candidate
gene for this spectrum of disorders.50 Some of the most
serious manifestations of AOS, such as pulmonary hyper-
tension, can be progressive. Revealing the molecular and
cellular mechanisms that underlie AOS pathogenesis
might facilitate the development of rational therapies to
interrupt or slow this sometimes lethal complication.
Supplemental Data
Supplemental Data include three figures and one table and can be
found with this article online at http://dx.doi.org/10.1016/j.ajhg.
2014.07.011.
Acknowledgments
We are very grateful to the families who took part in this study
and the Adams-Oliver syndrome Facebook support group. This
work was funded in part by the Rare Disease Foundation, the BC
Children’s Hospital Foundation, the University of Luxembourg
Institute for Systems Biology Program, the Center for Systems
Biology (2P50GM075647), the Inova Health System, and a gift
from the Odeen family. A.B.S. was supported by the German
Academic Exchange Service (Deutscher Akademischer Austausch
Dienst) and the German Research Foundation (Deutsche For-
schungsgemeinschaf). J.A. is a Gordon and Betty Moore Founda-
tion Fellow of the Life Sciences Research Foundation. We thank
Edward Greenberg and John H. Lee for help with radiologic
interpretation for the proband from family 4.
Received: July 2, 2014
Accepted: July 22, 2014
Published: August 14, 2014
Web Resources
The URLs for data provided herein are as follows:
1000 Genomes, http://browser.1000genomes.org
ClinVar, http://www.ncbi.nlm.nih.gov/clinvar/
Family Genomics Group, http://familygenomics.systemsbiology.
net/software
HGMD Professional, http://www.biobase-international.com/
product/hgmd
Kaviar2, http://db.systemsbiology.net/kaviar/cgi-pub/Kaviar2.pl
NHLBI Exome Sequencing Project (ESP) Exome Variant Server,
http://evs.gs.washington.edu/EVS/
Online Mendelian Inheritance in Man (OMIM), http://www.
omim.org/
RefSeq, http://www.ncbi.nlm.nih.gov/RefSeq
Accession Numbers
The ClinVar accession numbers for the variants reported in this
paper are SCV000172277, SCV000172278, SCV000172279,
SCV000172280, and SCV000172281.
282 The American Journal of Human Genetics 95, 275–284, Septemb
References
1. Adams, F., and Oliver, C. (1945). Hereditary deformities in
man due to arrested development. J. Hered. 36, 3–7.
2. Martınez-Frıas, M.L., Arroyo Carrera, I., Munoz-Delgado, N.J.,
Nieto Conde, C., Rodrıguez-Pinilla, E., Urioste Azcorra, M.,
Omenaca Teres, F., and Garcıa Alix, A. (1996). [The Adams-
Oliver syndrome in Spain: the epidemiological aspects]. An.
Esp. Pediatr. 45, 57–61.
3. Snape, K.M.G., Ruddy, D., Zenker, M., Wuyts, W., Whiteford,
M., Johnson, D., Lam, W., and Trembath, R.C. (2009). The
spectra of clinical phenotypes in aplasia cutis congenita and
terminal transverse limb defects. Am. J. Med. Genet. A.
149A, 1860–1881.
4. Algaze, C., Esplin, E.D., Lowenthal, A., Hudgins, L., Tacy, T.A.,
and Selamet Tierney, E.S. (2013). Expanding the phenotype of
cardiovascular malformations in Adams-Oliver syndrome.
Am. J. Med. Genet. A. 161A, 1386–1389.
5. Lin, A.E., Westgate, M.N., van der Velde, M.E., Lacro, R.V., and
Holmes, L.B. (1998). Adams-Oliver syndrome associated with
cardiovascular malformations. Clin. Dysmorphol. 7, 235–241.
6. Maniscalco,M., Zedda, A., Faraone, S., de Laurentiis, G., Verde,
R., Molese, V., Lapiccirella, G., and Sofia, M. (2005). Associa-
tion of Adams-Oliver syndrome with pulmonary arterio-
venous malformation in the same family: a further support
to the vascular hypothesis. Am. J. Med. Genet. A. 136,
269–274.
7. Shaheen, R., Aglan, M., Keppler-Noreuil, K., Faqeih, E., Ansari,
S., Horton, K., Ashour, A., Zaki, M.S., Al-Zahrani, F., Cueto-
Gonzalez, A.M., et al. (2013). Mutations in EOGT confirm
the genetic heterogeneity of autosomal-recessive Adams-
Oliver syndrome. Am. J. Hum. Genet. 92, 598–604.
8. Hassed, S.J., Wiley, G.B., Wang, S., Lee, J.-Y., Li, S., Xu, W.,
Zhao, Z.J., Mulvihill, J.J., Robertson, J., Warner, J., and Gaff-
ney, P.M. (2012). RBPJ mutations identified in two families
affected by Adams-Oliver syndrome. Am. J. Hum. Genet. 91,
391–395.
9. Shaheen, R., Faqeih, E., Sunker, A., Morsy, H., Al-Sheddi, T.,
Shamseldin, H.E., Adly, N., Hashem, M., and Alkuraya, F.S.
(2011). Recessive mutations in DOCK6, encoding the guani-
dine nucleotide exchange factor DOCK6, lead to abnormal
actin cytoskeleton organization and Adams-Oliver syndrome.
Am. J. Hum. Genet. 89, 328–333.
10. Southgate, L., Machado, R.D., Snape, K.M., Primeau, M.,
Dafou, D., Ruddy, D.M., Branney, P.A., Fisher, M., Lee, G.J.,
Simpson, M.A., et al. (2011). Gain-of-function mutations of
ARHGAP31, a Cdc42/Rac1 GTPase regulator, cause syndromic
cutis aplasia and limb anomalies. Am. J. Hum. Genet. 88,
574–585.
11. Cohen, I., Silberstein, E., Perez, Y., Landau, D., Elbedour, K.,
Langer, Y., Kadir, R., Volodarsky, M., Sivan, S., Narkis, G.,
and Birk, O.S. (2014). Autosomal recessive Adams-Oliver
syndrome caused by homozygous mutation in EOGT, encod-
ing an EGF domain-specific O-GlcNAc transferase. Eur. J.
Hum. Genet. 22, 374–378.
12. Silva, G., Braga, A., Leitao, B., Mesquita, A., Reis, A., Duarte, C.,
Barbot, J., and Silva, E.S. (2012). Adams-Oliver syndrome
and portal hypertension: fortuitous association or common
mechanism? Am. J. Med. Genet. A. 158A, 648–651.
13. Vandersteen, A.M., and Dixon, J.W. (2011). Adams-Oliver
syndrome, a family with dominant inheritance and a severe
phenotype. Clin. Dysmorphol. 20, 210–213.
er 4, 2014
14. Abecasis, G.R., Auton, A., Brooks, L.D., DePristo, M.A., Dur-
bin, R.M., Handsaker, R.E., Kang, H.M., Marth, G.T., McVean,
G.A., Gabriel, S.B., et al.; 1000 Genomes Project Consortium
(2012). An integrated map of genetic variation from 1,092
human genomes. Nature 491, 56–65.
15. Glusman, G., Caballero, J., Mauldin, D.E., Hood, L., and
Roach, J.C. (2011). Kaviar: an accessible system for testing
SNV novelty. Bioinformatics 27, 3216–3217.
16. Cingolani, P., Platts, A., Wang, L., Coon, M., Nguyen, T.,
Wang, L., Land, S.J., Lu, X., and Ruden, D.M. (2012).
A program for annotating and predicting the effects of single
nucleotide polymorphisms, SnpEff: SNPs in the genome
of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly
(Austin) 6, 80–92.
17. Paila, U., Chapman, B.A., Kirchner, R., and Quinlan, A.R.
(2013). GEMINI: integrative exploration of genetic variation
and genome annotations. PLoS Comput. Biol. 9, e1003153.
18. Li, H., Glusman, G., Huff, C., Caballero, J., and Roach, J.C.
(2014). Accurate and robust prediction of genetic relationship
from whole-genome sequences. PLoS ONE 9, e85437.
19. Bray, S.J. (2006). Notch signalling: a simple pathway becomes
complex. Nat. Rev. Mol. Cell Biol. 7, 678–689.
20. Jarriault, S., Brou, C., Logeat, F., Schroeter, E.H., Kopan, R., and
Israel, A. (1995). Signalling downstream of activated mamma-
lian Notch. Nature 377, 355–358.
21. Huang, N., Lee, I., Marcotte, E.M., and Hurles, M.E. (2010).
Characterising and predicting haploinsufficiency in the
human genome. PLoS Genet. 6, e1001154.
22. Haines, N., and Irvine, K.D. (2003). Glycosylation regulates
Notch signalling. Nat. Rev. Mol. Cell Biol. 4, 786–797.
23. Frischmeyer, P.A., van Hoof, A., O’Donnell, K., Guerrerio, A.L.,
Parker, R., and Dietz, H.C. (2002). An mRNA surveillance
mechanism that eliminates transcripts lacking termination
codons. Science 295, 2258–2261.
24. Ashworth, J., and Baker, D. (2009). Assessment of the optimi-
zation of affinity and specificity at protein-DNA interfaces.
Nucleic Acids Res. 37, e73.
25. Fleishman, S.J., Leaver-Fay, A., Corn, J.E., Strauch, E.M., Khare,
S.D., Koga, N., Ashworth, J., Murphy, P., Richter, F., Lemmon,
G., et al. (2011). RosettaScripts: a scripting language interface
to the Rosetta macromolecular modeling suite. PLoS ONE 6,
e20161.
26. Leaver-Fay, A., Tyka, M., Lewis, S.M., Lange, O.F., Thompson,
J., Jacak, R., Kaufman, K., Renfrew, P.D., Smith, C.A., Sheffler,
W., et al. (2011). ROSETTA3: an object-oriented software suite
for the simulation and design of macromolecules. Methods
Enzymol. 487, 545–574.
27. Berman, H.M., Battistuz, T., Bhat, T.N., Bluhm, W.F., Bourne,
P.E., Burkhardt, K., Feng, Z., Gilliland, G.L., Iype, L., Jain, S.,
et al. (2002). The Protein Data Bank. Acta Crystallogr. D
Biol. Crystallogr. 58, 899–907.
28. Cordle, J., Johnson, S., Tay, J.Z.Y., Roversi, P., Wilkin, M.B., de
Madrid, B.H., Shimizu, H., Jensen, S., Whiteman, P., Jin, B.,
et al. (2008). A conserved face of the Jagged/Serrate DSL
domain is involved in Notch trans-activation and cis-inhibi-
tion. Nat. Struct. Mol. Biol. 15, 849–857.
29. Wu, Y., Cain-Hom, C., Choy, L., Hagenbeek, T.J., de Leon, G.P.,
Chen, Y., Finkle, D., Venook, R., Wu, X., Ridgway, J., et al.
(2010). Therapeutic antibody targeting of individual Notch
receptors. Nature 464, 1052–1057.
30. Choi, S.H., Wales, T.E., Nam, Y., O’Donovan, D.J., Sliz, P.,
Engen, J.R., and Blacklow, S.C. (2012). Conformational lock-
The American
ing upon cooperative assembly of notch transcription com-
plexes. Structure 20, 340–349.
31. Kellogg, E.H., Leaver-Fay, A., and Baker, D. (2011). Role of
conformational sampling in computing mutation-induced
changes in protein structure and stability. Proteins 79, 830–
838.
32. Rebay, I., Fleming, R.J., Fehon, R.G., Cherbas, L., Cherbas, P.,
and Artavanis-Tsakonas, S. (1991). Specific EGF repeats of
Notch mediate interactions with Delta and Serrate: impli-
cations for Notch as a multifunctional receptor. Cell 67,
687–699.
33. Hambleton, S., Valeyev, N.V., Muranyi, A., Knott, V., Werner,
J.M., McMichael, A.J., Handford, P.A., and Downing, A.K.
(2004). Structural and functional properties of the human
notch-1 ligand binding region. Structure 12, 2173–2183.
34. Kortemme, T., Morozov, A.V., and Baker, D. (2003). An
orientation-dependent hydrogen bonding potential improves
prediction of specificity and structure for proteins and pro-
tein-protein complexes. J. Mol. Biol. 326, 1239–1259.
35. Edgar, R.C. (2004). MUSCLE: multiple sequence alignment
with high accuracy and high throughput. Nucleic Acids Res.
32, 1792–1797.
36. Clamp, M., Cuff, J., Searle, S.M., and Barton, G.J. (2004). The
Jalview Java alignment editor. Bioinformatics 20, 426–427.
37. Krebs, L.T., Xue, Y., Norton, C.R., Shutter, J.R., Maguire, M.,
Sundberg, J.P., Gallahan, D., Closson, V., Kitajewski, J., Calla-
han, R., et al. (2000). Notch signaling is essential for vascular
morphogenesis in mice. Genes Dev. 14, 1343–1352.
38. Krebs, L.T., Starling, C., Chervonsky, A.V., and Gridley, T.
(2010). Notch1 activation in mice causes arteriovenous mal-
formations phenocopied by ephrinB2 and EphB4 mutants.
Genesis 48, 146–150.
39. Joutel, A., Corpechot, C., Ducros, A., Vahedi, K., Chabriat, H.,
Mouton, P., Alamowitch, S., Domenga, V., Cecillion, M.,
Marechal, E., et al. (1996). Notch3 mutations in CADASIL, a
hereditary adult-onset condition causing stroke and demen-
tia. Nature 383, 707–710.
40. McDaniell, R., Warthen, D.M., Sanchez-Lara, P.A., Pai, A.,
Krantz, I.D., Piccoli, D.A., and Spinner, N.B. (2006). NOTCH2
mutations cause Alagille syndrome, a heterogeneous disorder
of the notch signaling pathway. Am. J. Hum. Genet. 79,
169–173.
41. Roca, C., and Adams, R.H. (2007). Regulation of vascular
morphogenesis by Notch signaling. Genes Dev. 21, 2511–
2524.
42. Tsuji, H., Ishii-Ohba, H., Ukai, H., Katsube, T., and Ogiu, T.
(2003). Radiation-induced deletions in the 50 end region of
Notch1 lead to the formation of truncated proteins and are
involved in the development of mouse thymic lymphomas.
Carcinogenesis 24, 1257–1268.
43. Petrovic, J., Formosa-Jordan, P., Luna-Escalante, J.C., Abello,
G., Ibanes, M., Neves, J., and Giraldez, F. (2014). Ligand-
dependent Notch signaling strength orchestrates lateral
induction and lateral inhibition in the developing inner ear.
Development 141, 2313–2324.
44. Hoyme, H.E., Jones, K.L., Van Allen, M.I., Saunders, B.S., and
Benirschke, K. (1982). Vascular pathogenesis of transverse
limb reduction defects. J. Pediatr. 101, 839–843.
45. Toriello, H.V., Graff, R.G., Florentine, M.F., Lacina, S., and
Moore, W.D. (1988). Scalp and limb defects with cutis
marmorata telangiectatica congenita: Adams-Oliver syn-
drome? Am. J. Med. Genet. 29, 269–276.
Journal of Human Genetics 95, 275–284, September 4, 2014 283
46. Swartz, E.N., Sanatani, S., Sandor, G.G., and Schreiber, R.A.
(1999). Vascular abnormalities in Adams-Oliver syndrome:
cause or effect? Am. J. Med. Genet. 82, 49–52.
47. Patel, M.S., Taylor, G.P., Bharya, S., Al-Sanna’a, N., Adatia, I.,
Chitayat, D., Suzanne Lewis, M.E., and Human, D.G. (2004).
Abnormal pericyte recruitment as a cause for pulmonary
hypertension in Adams-Oliver syndrome. Am. J. Med. Genet.
A. 129A, 294–299.
48. Chang, L., Noseda, M., Higginson, M., Ly, M., Patenaude, A.,
Fuller, M., Kyle, A.H., Minchinton, A.I., Puri, M.C., Dumont,
D.J., and Karsan, A. (2012). Differentiation of vascular smooth
284 The American Journal of Human Genetics 95, 275–284, Septemb
muscle cells from local precursors during embryonic and adult
arteriogenesis requires Notch signaling. Proc. Natl. Acad. Sci.
USA 109, 6993–6998.
49. Pereira-Da-Silva, L., Leal, F., Santos, G.C., Videira Amaral, J.M.,
and Feijoo, M.J. (2000). Clinical evidence of vascular abnor-
malities at birth in Adams-Oliver syndrome: report of two
further cases. Am. J. Med. Genet. 94, 75–76.
50. Digilio, M.C., Marino, B., and Dallapiccola, B. (2008). Auto-
somal dominant inheritance of aplasia cutis congenita and
congenital heart defect: a possible link to the Adams-Oliver
syndrome. Am. J. Med. Genet. A. 146A, 2842–2844.
er 4, 2014