Whole-exome sequencing combined with functional
genomics reveals novel candidate driver cancer
genes in endometrial cancer
Han Liang,1,9,11Lydia W.T. Cheung,2,9Jie Li,2Zhenlin Ju,1Shuangxing Yu,2
Katherine Stemke-Hale,2Turgut Dogruluk,3Yiling Lu,2Xiuping Liu,4Chao Gu,2
Wei Guo,2Steven E. Scherer,5Hannah Carter,6Shannon N. Westin,7Mary D. Dyer,2
Roeland G.W. Verhaak,1Fan Zhang,2Rachel Karchin,6Chang-Gong Liu,4Karen H. Lu,7
Russell R. Broaddus,8Kenneth L. Scott,3Bryan T. Hennessy,2,10and Gordon B. Mills2
1Department of Bioinformatics and Computational Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas
77030, USA;2Department of Systems Biology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA;
3Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas 77030, USA;4Department of
Experimental Therapeutics, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA;5Human Genome
Sequencing Center, Baylor College of Medicine, Houston, Texas 77030, USA;6Department of Biomedical Engineering and Institute
for Computational Medicine, Johns Hopkins University, Baltimore, Maryland 21205, USA;7Department of Gynecologic Oncology,
The University of Texas MD Anderson Cancer Center, Houston, Texas 77030, USA;8Department of Pathology, The University of Texas
MD Anderson Cancer Center, Houston, Texas 77030, USA
Endometrial cancer is the most common gynecological malignancy, with more than 280,000 cases occurring annually
worldwide. Although previous studies have identified important common somatic mutations in endometrial cancer, they
have primarily focused on a small set of known cancer genes and have thus provided a limited view of the molecular basis
underlying this disease. Here we have developed an integrated systems-biology approach to identifying novel cancer
genes contributing to endometrial tumorigenesis. We first performed whole-exome sequencing on 13 endometrial cancers
and matched normal samples, systematically identifying somatic alterations with high precision and sensitivity. We then
combined bioinformatics prioritization with high-throughput screening (including both shRNA-mediated knockdown
and expression of wild-type and mutant constructs) in a highly sensitive cell viability assay. Our results revealed 12
potential driver cancer genes including 10 tumor-suppressor candidates (ARID1A, INHBA, KMO, TTLL5, GRM8, IGFBP3, AKTIP,
PHKA2, TRPS1, and WNT11) and two oncogene candidates (ERBB3 and RPS6KC1). The results in the ‘‘sensor’’ cell line were
profiles with functional proteomics in 222 endometrial cancer samples, demonstrating that ARID1A mutations frequently
co-occur with mutations in the phosphatidylinositol 3-kinase (PI3K) pathway and are associated with PI3K pathway ac-
tivation. siRNA knockdown in endometrial cancer cell lines increased AKT phosphorylation supporting ARID1A as a novel
regulator ofPI3Kpathway activity. Ourstudy presents the first unbiased view ofsomatic codingmutations inendometrial
cancer and provides functional evidence for diverse driver genes and mutations in this disease.
[Supplemental material is available for this article.]
Endometrial cancer is the most common gynecological malignancy
and the fourth most common cancer among women in Western
countries, with more than 280,000 cases occurring annually
worldwide (Ferlay et al. 2010). Although the prognosis is favorable
for patients identified with early-stage disease, the outcomes for
patients with metastatic or recurrent endometrial cancer remain
abysmal, with median survival of <1 yr. For effective cancer pre-
vention and treatment, it is necessary to identify critical genetic
changes that initiate endometrial cancer and contribute to its
progression. Over the last decade, multiple groups have identified
common somatic mutations of several important genes in endo-
metrial cancer (Dedes et al. 2011). For example, PIK3CA mutations
occur in ;30% of endometrial cancers, with the mutation fre-
quency varying according to histological grade (Oda et al. 2005;
Velasco et al. 2006; Catasus et al. 2008; Miyake et al. 2008; Hayes
et al. 2009; Cheung et al. 2011). PTEN, a major negative regulator
of the phosphatidylinositol 3-kinase (PI3K) pathway, is frequently
mutated in endometrial tumors (Risinger et al. 1998; Mutter et al.
2000; Sun et al. 2001; Salvesen et al. 2004). Other cancer genes
in this disease (Lax et al. 2000; Schlosshauer et al. 2000; Moreno-
Buenoetal.2002;Stefanssonet al.2004;Byronetal. 2008;Jiaetal.
2008; Catasuset al. 2009;Shoji et al.2009). Morerecently, we have
demonstrated that PI3K pathway aberrations occur in >80% of
9These authors contributed equally to this work.
10Present address: Department of Medical Oncology, Beaumont
Hospital, Dublin 9, Ireland.
Article published online before print. Article, supplemental material, and pub-
lication date are at http://www.genome.org/cgi/doi/10.1101/gr.137596.112.
2120 Genome Research
22:2120–2129 ? 2012, Published by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/12; www.genome.org
members (Cheung et al. 2011). Although these studies provide
important insights into the molecular basis of endometrial cancer,
they have primarily focused on a limited set of well-known cancer
genes. To date, a comprehensive view of the mutation spectrum in
this disease has not been achieved.
Recent advances in next-generation sequencing technology
have enabled the unprecedented characterization of a full spec-
trum of somatic alterations in cancer genomes (Mardis 2011). In
particular, through target enrichment, whole-exome sequencing
represents a cost-effective strategy to identify mutations in pro-
tein-coding exons in the human genome. In the past several years,
this approach has been successfully applied to several human
cancers (Cancer Genome Atlas Research Network 2008, 2011;
Jones etal. 2010;Gui etal. 2011;Wang etal. 2011).Giventhe large
a key challenge in the downstream analysis is to distinguish
‘‘drivers’’ that functionally contribute to tumorigenesis from
‘‘passengers’’ that occur as the consequence of genomic instability.
So far, most of the published studies have characterized at most
one or two genes of interest, leaving the vast majority of mutated
Here we implemented a systems-biology approach to iden-
tify driver cancer genes in endometrial cancers. We first per-
formed whole-exome sequencing on 13 endometrial cancers
along with matched normal samples to detect somatic alter-
ations with high precision and sensitivity. For validated mutated
genes, we combined bioinformatics prioritization with high-
throughput screening of short-hairpin RNA (shRNA) and over-
expression constructs in a highly sensitive cell viability assay to
identify candidate driver cancer genes. The impact of candidate
cancer cell lines by small interfering RNA (siRNA) knockdown.
Finally, we focused on one candidate driver gene, ARID1A, for
mechanistic investigation. Figure 1 shows an overview of our ap-
proach. Our study thus provides an integrated approach to identi-
fying and characterizing novel cancer genes that have not been
implicated in endometrial tumorigenesis.
Landscape of somatic mutations in the exomes
of endometrial cancer
To gain an unbiased view of somatic alterations that contribute to
the pathogenesis of endometrial cancer, we performed whole-
exome sequencing of 14 endometrial tumor samples (one tumor
with an unusually high mutation frequency and low estimated
tumor content was excluded from further analysis) and matched
normal DNA samples using the SOLiD platform. Supplemental
Table 1 provides the histologic information and other clinical
characteristics of these samples. On average, we obtained 185 mil-
lion 50-nt single reads for tumor samples and 62 million reads for
normal samples (Table 1). The Agilent SureSelect capture reagent
used for enrichment targets 170,843 exons (;1.2% of the human
genome); for the targeted regions, the average coverage was 943
and 423 for tumor and normal samples, respectively. We de-
veloped a computational pipeline to detect somatic mutations by
comparing single-nucleotide variations in tumor and normal
samples (Supplemental Figs. 1, 2; Supplemental Methods). At the
depth of sequencing used and with the majority of tumors being
low grade (grade 2), we identified a mean of 89.5 somatic point
mutations (range 34–263) per tumor corresponding to an average
of 3.7 mutations per megabase (range 0.98–12.0) (Table 1). As
shown in Figure 2, the most common mutations were transitions
in the CpG context (that is, G>A/C>T), which parallels several
other cancer lineages (Ding et al. 2010; Chapman et al. 2011;
Puente et al. 2011). In addition, the tumors contained a mean of
14.2 somatic coding small insertion/deletions (indels) (Table 1;
Supplemental Table 2; Supplemental Methods).
function, across the 13 tumor samples, we identified 576 unique
nonsynonymous mutations and 24 nonsense mutations in 566
genes (Supplemental Table 2). The most perturbed biological
pathways were integrin, angiopoietin, complement system, and
PTEN signaling (Fisher exact test, P < 2.8 3 10?3, false discovery
rate [FDR] < 0.05, Supplemental Table 3). Based on potential bi-
ological interest, we selected 97 mutation sites for Sequenom
MassARRAY validation and obtained a true-positive rate of 81%,
leading to 69 genes with validated mutations and a prediction of
487 nonsynonymous mutations and 18 nonsense mutations in
the 13 tumors (Supplemental Table 2; Supplemental Methods). To
estimate the sensitivity of our mutation calling algorithm, we ex-
amined full-length sequences of nine genes with concurrent
Sanger sequencing and Sequenom MassARRAY detection. The es-
timated sensitivity was 80%, with false negatives occurring pri-
marily due to low coverage (Supplemental Methods). These results
indicate that the identified mutated gene set largely characterizes
the landscape of somatic coding mutations in the 13 endometrial
Novel candidate driver cancer genes revealed by Ba/F3
Several of the validated mutated genes have previously been
shown to be targeted in endometrial cancers, including TP53,
PTEN, CTNNB1, PIK3CA, PIK3R1, and PIK3R2 (Lax et al. 2000; Sun
et al. 2001; Moreno-Bueno et al. 2002; Oda et al. 2005; Cheung
et al. 2011). We recently reported their mutation profiles in a large
collection of well-characterized endometrial tumors, demonstrat-
ing that aberrations, including frequent co-mutations, of the PI3K
pathway occur in ;80% of endometrial cancers (Cheung et al.
cancer driver genes in endometrial cancer.
Overview of our systems-biology approach to identifying
Novel driver cancer genes in endometrial cancer
2011). However, the functional roles of the remaining mutated
genes in endometrial cancer are largely unknown. To identify po-
tential driver cancer genes for further characterization, we per-
formed bioinformatics analyses and selected 30 genes using the
following criteria: (1) if the mutation in a gene is predicted to have
a high impact on the protein function by CHASM (FDR of impact
score <30%) (Carter et al. 2010) or MutationAssessor (high impact)
(Reva et al. 2011); or (2) a gene contains multiple or recurrent mu-
tations (Methods). Supplemental Table 4 provides details on selec-
tion of candidates for further analysis. We then evaluated the func-
becomes IL3-independent in the presence of an oncogene or onco-
genic event. Thus, the Ba/F3 cells represent a sensitive tool for
measuring the effect of an introduced perturbation on cell pro-
liferation and survival, not only for kinase
2000; Warmuth et al. 2007; Yoda et al.
2010; Cheung et al. 2011).
We first performed a ‘‘loss-of-
function’’ analysis on the 30 candidate
genes using shRNA gene silencing. As
shown in Figure 3A, compared with pa-
rental cells and cells transfected with
empty vector or nonspecific shRNAs, in-
hibiting the expression of eight genes
(INHBA, KMO, PHKA2, TTLL5, AKTIP,
IGFBP3, GRM8, and ARID1A) with two
independent shRNAswas sufficient
to promote the IL3-independent survival
of Ba/F3, compatible with these genes
having tumor-suppressor-like activity. In
contrast, the introduction of ERBB3 and
RPS6KC1 shRNAs remarkably inhibited
IL3-independent cell survival, compati-
ble with oncogene-like activity. These
observations were confirmed by at least
one of two additional independent
shRNAs in a secondary screen (data not
shown). We confirmedshRNAknockdown
efficacy by Western blotting for eight pro-
teins where appropriate antibodies were
available (Fig. 3C). Nine genes with efficient shRNA knockdown
had no effect in the Ba/F3 cells with four (NOS1AP, CDK17,
C1orf198, and PHKG1) shown in Figure 3A and Supplemental
Figure 3a. One of two shRNAs in the initial screen for two genes
(TRPS1 and WNT11) promoted IL3-independent Ba/F3 survival.
The activity was confirmed by two additional independent shRNAs
(Fig. 3B; Supplemental Fig. 3b). Significant knockdown was not
observed or antibodies were unavailable for the remaining nine
genes. Taken together, the shRNA-mediated knockdown screen
revealed 12 potential driver cancer genes including 10 tumor
suppressors and two oncogene candidates from the 30 genes
assessed in the Ba/F3 survival assay.
To complement the shRNA screen, we performed a ‘‘gain-of-
function’’ analysis on 17 out of the 30 genes by overexpressing the
wild-type or mutated gene in the Ba/F3 cells. Among the 17 genes
examined, six genes were scored positive in the Ba/F3 shRNA
Landscape of somatic mutations in the exomes of 13 endometrial tumors
9.75 3 109
8.84 3 109
8.75 3 109
9.17 3 109
9.55 3 109
9.76 3 109
9.22 3 109
8.84 3 109
9.13 3 109
8.97 3 109
9.21 3 109
9.10 3 109
9.67 3 109
4.90 3 109
4.18 3 109
4.60 3 109
4.69 3 109
2.72 3 109
4.35 3 109
4.01 3 109
4.74 3 109
4.60 3 109
4.24 3 109
4.22 3 109
4.09 3 109
3.53 3 109
30.9 3 106
25.1 3 106
29.0 3 106
29.1 3 106
21.9 3 106
20.7 3 106
31.2 3 106
29.9 3 106
31.3 3 106
25.9 3 106
18.7 3 106
28.1 3 106
28.3 3 106
The mutation rate was calculated by dividing the total number of somatic mutations by the total number of callable nucleotide positions ($153 in tumor
and $83 in matched normal samples).
of six classes of mutations is shown for all mutations in the exomes, nonsilent coding mutations, and
silent mutations, respectively.
Mutation profile of somatic mutations in the exomes of endometrial cancer. The frequency
Liang et al.
screen. As shown in Figure 4A, overexpression of wild-type ERBB3
significantly increased cell survival, consistent with its proposed
relative to the wild-type, suggesting that this mutation is not
with controls, overexpression of wild-type AKTIP and INHBA sig-
nificantly reduced cell survival, consistent with their potential roles
as tumor suppressors. Interestingly, their corresponding mutations
(Q281K in AKITP and R310Q in INHBA) significantly increased the
Ba/F3 survival compared to the wild-type constructs (P < 0.05),
suggesting a critical inactivating effect of these mutations and
strongly supporting AKTIP and INHBA as bona fide tumor sup-
pressors. In contrast, upon overexpression, wild-type or mutants of
the GRM8, PHKA2, and WNT11 genes which scored positive in
the shRNA screen showed no significant effect on survival of Ba/F3,
suggesting that the effects of these genes may be dosage-in-
screen, there was no significant change in Ba/F3 survival with
overexpressionofthe remaining 11genesortheir mutants(Fig.4B).
Effect of candidate driver cancer genes by siRNA knockdown
in endometrial cancer cell lines
To extend our observations in Ba/F3 cells to human endometrial
cancer cells, we introduced siRNAs targeting the 12 candidate
driver cancer genes into four human endometrial cancer cell lines
and assessed effects on cell viability. We included an siRNA tar-
geting the established PTEN tumor-suppressor gene to validate the
approach. As shown in Figure 5 and
Supplemental Figure 4, decreased PTEN
protein and increased AKT phosphory-
lation was accompanied by significant
increased cell viability in PTEN siRNA-
transfected cells carryingawild-typePTEN
gene and expressing PTEN protein, but not
in PTEN protein-negative cell lines, sup-
porting the utility of the approach. Im-
portantly, KLE, in which all genes assessed
were wild-type, exhibited the same re-
F3 except for IGFBP3 siRNA (Fig. 5B). The
responses of EFE184 (all genes tested
are wild-type except ARID1A with un-
detectable protein expression), SK-UT-2
(ARID1A is mutated with detectable
protein; mutation status of others is un-
known), and SNG-II (ARID1A is mu-
tated with undetectable protein; the mu-
tation status of the others is unknown)
were more variable (Supplemental Fig. 4).
viability of KLE and SK-UT-2 where the
ARID1A protein is expressed (Fig. 5B;
Supplemental Fig. 4). Importantly, AKTIP
siRNA significantly increased viability in
all cell lines, consistent with its potential
tumor-suppressor role suggested in the
Ba/F3 system. The siRNAs for WNT11
and INHBA increased viability in three
cell lines, whereas IGFBP3 siRNA in-
creased cell viability in two cell lines. In
contrast, ERBB3 and RPS6KC1 siRNAs inhibited cell viability in all
cell lines at least at one time point. siRNAs targeting NOS1AP,
PHKG1, and C1orf198, which had no effect in Ba/F3 survival, did
not alter cell viability in all of the endometrial cell lines we ex-
amined, providing additional support for the validity of the Ba/F3
model and the pipeline of characterization of mutations. Collec-
tively, the effect of silencing candidate driver genes in endometrial
cancer cell lines largely recapitulated the results obtained in Ba/F3
Regulation of PI3K pathway activation by ARID1A
ARID1A (AT-rich interactive domain 1A; also known as BAF250) is
a key member of the SWI/SNF chromatin-modeling complex, and
the gene has recently been reported to be frequently mutated in
a wide variety of cancer types (Jones et al. 2010; Wiegand et al.
2010, 2011; Guan et al. 2011b; Gui et al. 2011; Wang et al. 2011;
Mamo et al. 2012). While previous studies and our present results
suggest its potential role as a tumor suppressor, the molecular
mechanism underlying its functional role in cancer is largely un-
endometrial cancer, we first resequenced this gene in the same
et al. 2011) and observed a (nonsilent) mutation frequency of
co-occur with mutations in PTEN (Fisher exact test, P < 1.2 3 10?5,
Bonferroni corrected P < 1.0 3 10?4) and PIK3CA (P < 2.3 3 10?3,
Bonferroni corrected P < 1.8 3 10?2) (Fig. 6A; Supplemental Table
5), as well as with overall PI3K pathway aberration (P < 1.1 3 10?3,
(A) Ba/F3 cells were transfected with short-hairpin RNAs (shRNA) targeting indicated genes. Empty
vector (pGIPZ) and nonspecific shRNA served as the control. Cells were cultured without IL-3 for 4 wk
and harvested for viability assay. Cell viability relative to Ba/F3 parental cells was shown. (*) P < 0.05,
compared with Ba/F3 parental control. (B) Whole cell lysates were also collected for Western blotting
with indicated antibodies, and ERK2 was used as the loading control.
Novel candidate driver cancer genes identified by shRNA screening in Ba/F3 viability assay.
Novel driver cancer genes in endometrial cancer
Bonferroni corrected P < 8.6 3 10?3). Since coordinate PI3K
pathway mutations are common in endometrial cancer relative to
other cancer lineages, the concordance of aberrations in the PI3K
pathway and ARID1A could represent coordinate targeting of the
PI3K pathway. We thus examined the effect of ARID1A mutations
onproteinandphosphoproteinlevelsofcore members in thePI3K
pathway using reverse-phase protein arrays and found that the
phosphorylationofseveraldownstream targets(PDK1,AKT, GSK3,
TSC2, p70S6K, and ACC) are significantly up-regulated in tumors
with ARID1A mutations (two-sided t-test, P < 0.05, FDR < 0.1; Fig.
6B). Since (1) PTEN loss has a dominant effect on the activation of
mutation patterns with PTEN and PIK3CA, we then focused on
a subset of tumor samples in which both PTEN and PIK3CA genes
were wild-type, and also PTEN expression was retained (n = 47).
Strikingly, in these samples, phosphorylation of AKTpS473 and
p70S6KpS371, two key PI3K pathway proteins, remained signifi-
cantly up-regulated in ARID1A-mutated samples (Fig. 6C) as com-
pared with samples where ARID1A, PTEN, and PIK3CA are wild-
type, indicating that the activation of PI3K pathway by ARID1A
mutations is not due to co-occurrence of ARID1A mutations with
aberrations in PTEN or PIK3CA.
We determined whether ARID1A regulated PI3K pathway
activity in endometrial cancer cell lines. Consistent with RPPA
analysis from the large endometrial cancer sample cohort, knock-
down of ARID1A significantly elevated AKT phosphorylation
levels in three cell lines (KLE, ESS1, and
MFE280) expressing wild-type ARID1A
(Fig. 6D). In contrast, up-regulation of
AKT phosphorylation was not observed
in the EFE184 cell line in which ARID1A
protein is not present (Fig. 6D). These
results are consistent with inhibition of
the PI3K pathway contributing to the
tumor-suppressor activity of ARID1A.
Our study represents the first unbiased
view of somatic mutations in endome-
trialcancer, which strongly complements
previous gene- and pathway-focused stud-
ies. More importantly, our results provide
substantial functional evidence for a di-
versity of novel candidate drivers, sug-
of endometrial cancer with implications
for the development of targeted therapy.
Next-generation sequencing tech-
nology has facilitated characterization of
the full spectrum of aberrations in cancer
genomes in a cost-effective and timely
manner (Mardis 2011). However, there is
stillagreatgap betweencreatinga catalog
of mutations and alterations and identi-
fying a short list of ‘‘actionable’’ elements
(Chin et al. 2011). Here we developed
a systems-biology approach to filling
this gap: We combined computation-
prediction-based prioritization with func-
tional screening in a highly sensitive cell
viability assay. This approach allows the identification of a large
number of candidate driver cancer genes in an efficient way. Fo-
cusing on a candidate driver gene of high interest, we further ex-
amined the underlying molecular mechanisms through an in-
tegrative analysis of mutation profile and protein expression on a
large, well-characterized sample set and ‘‘hypothesis-driven’’ func-
tional studies in endometrial cancer cell lines. These methods and
the related experimental systems can be readily applied to other
Through the combination of gene silencing and over-
revealed 12 potential driver cancer genes in endometrial cancer.
Among these genes, ARID1A has attracted wide interest recently.
Frequent mutations throughout the gene sequence including
multiple truncation mutations have been reported in ovarian
(Jones et al. 2010; Wiegand et al. 2010), gynecological (Guan et al.
2011b), bladder (Gui et al. 2011), gastric (Wang et al. 2011), breast
(Mamo et al. 2012), and endometrial cancers (Guan et al. 2011a;
Wiegand et al. 2011), suggesting a role as a tumor suppressor.
ARID1A encodes BAF250a, a nuclear protein and a key component
of the SWI/SNF chromatin-remodeling complex that functions as
a regulator of gene expression and chromatin dynamics (Wu et al.
2009). However, to date,the mechanisms by whichloss of ARID1A
function contributes to cancer pathophysiology remains poorly
understood. ARID1A has been suggested to suppress cell prolifer-
ation of ovarian and endometrial cancer cell lines through physi-
cally interacting with p53 to coordinately regulate the transcription
in Ba/F3 viability assay. Ba/F3 cells were transfected with wild-type (WT) or corresponding mutant(s)
(mutation sites indicated) of (A) six positive genes in the shRNA screen and (B) 11 genes inactive in the
screen. Cells transfected with shRNA were included in the assay as reference. pGIPZ vector is the empty
vector carrying shRNA while LacZ corresponds to b-galactosidase in the pLenti6.3 vector. Cells were
cultured without IL-3 for 4 wk and harvested for viability assay. Cell viability relative to Ba/F3 parental
cells was shown. (*) P < 0.05, compared with Ba/F3 parental control. (#) A significant difference in cell
viability between WT and mutant-transfected cells (P < 0.05).
Candidate driver cancer genes confirmed by overexpression of wild-type genes or mutants
Liang et al.
2124 Genome Research
of cell cycle–related genes (Guan et al. 2011b), linking its function
to nuclear localization. Here we not only confirmed a high muta-
the first time, demonstrated that ARID1A can regulate PI3K path-
way activity, consistent with inhibition of the PI3K pathway
contributing to ARID1A tumor-suppressor activity. Since the PI3K
pathway represents a promising target for therapy (Hennessy et al.
2005), these results have direct implications for clinical trans-
lation. At present, it remains unclear which molecules in the
pathway are targeted by ARID1A, and further studies are required
to elucidate mechanistic details.
Among the other candidate tumor suppressors identified,
AKTIP, INHBA, and WNT11 appear to be the most promising
ones. AKTIP and INHBA are particularly likely to represent tumor
suppressors because knockdown and overexpression demon-
strated opposite effects in Ba/F3, and, critically, patient-derived
mutations abrogated the effects of the wild-type expression
construct (Figs. 3, 4). AKTIP was first identified as an AKT bind-
ing partner (Remy and Michnick 2004). Exogenous over-
expression of AKTIP enhanced AKT phosphorylation, but this
activation induced apoptosis for unknown reasons (Remy and
Michnick 2004). Based on data from the Cancer Genome Atlas,
homologous deletion of AKTIP occurs in multiple cancer line-
ages including breast (2%), ovary (0.9%), and prostate (1%),
consistent with a tumor-suppressor role. However, the func-
tional role of AKTIP could be tumor type-specific, since it has
been proposed as a putative oncoprotein in cervical cancer
but a tumor suppressor in ovarian cancer (Cinghu et al. 2011;
Notaridou et al. 2011). INHBA encodes inhibin beta A, a subunit
of both the activin and inhibin receptors of the transforming
growth factor (TGF-beta) superfamily (Risbridger et al. 2001).
Inhibin beta A, like TGF-beta, can inhibit or stimulate cell
growth dependent on the cellular con-
text. For example, INHBA substantially
inhibitedtumorgrowth and angiogenesis
in in vivo gastric cancer and neuroblas-
toma models (Schramm et al. 2005;
Kaneda et al. 2011). Meanwhile, INHBA
mRNA was up-regulated in lung cancer
and may result in the promotion of cell
proliferation (Seder et al. 2009). In endo-
metrial cancer cell lines, the role of
INHBA is controversial (Di Simone et al.
2002; Tanaka et al. 2003). It is possible
that the contribution of INHBA to the
tumorigenesis is determined by the rela-
tive expression levels of other receptor
subunits and their interacting partners.
WNT11 is a key member of the Wingless-
type (Wnt) signaling pathway whose
deregulation has been implicated in endo-
metrioid endometrial tumors as evidence
by beta-catenin mutations and aberrant
nuclear accumulation (Ikeda et al. 2000;
Moreno-Bueno et al. 2002). The canoni-
cal Wnt cascade is mediated by nuclear
beta-catenin binding to T-cell factor tran-
scription factors to activate genes relevant
to tumorigenesis, while noncanonical
Wnt signaling is beta-catenin independent
(Bejsovec 2005). Interestingly, WNT11 is
down-regulated in hepatocellular carci-
noma, and it can modulate both canonical and noncanonical Wnt
pathways to execute its tumor-suppressor actions (Toyama et al.
2010). Thus, the role of WNT11 in Wnt signaling in endometrial
cancer warrants further investigation.
ERBB3 and RPS6KC1 are potential oncogenes identified in
our study. ERBB3 (HER3) belongs to the epidermal growth factor
receptor family and has been implicated in cancer, previously.
ERBB2, a partner of ERBB3, as well as ERBB3 itself is often am-
plified and overexpressed in breast, ovarian, prostate, and lung
cancers (for review, see Sithanandam and Anderson 2008).
Overexpression of ERBB3 promotes tumorigenesis through mul-
tiple mechanisms including cell cycle progression, stimulation of
cell migration, and invasion primarily via activation of the PI3K
pathway (Sithanandam and Anderson 2008). ERBB3 is highly
expressed in endometrial cancer, but the functional role remains
unclear (Srinivasan et al. 1999; Ejskjaer et al. 2007). RPS6KC1
(encoding ribosomal protein S6 kinase polypeptide 1) is another
potential oncogene. RPS6KC1 mutations have been previously
found in breast, ovary, and lung cancers (Davies et al. 2005;
Stephens et al. 2005; Cancer Genome Atlas Research Network
kinase activity (Hayashi et al. 2002; Liu et al. 2005), but, instead,
likely functions as an adaptor molecule to recruit binding part-
ners (sphingosine kinase-1 and peroxiredoxin-3) to early endo-
somes (Hayashi et al. 2002; Liu et al. 2005). These trafficking
pathways within the endosomal system play a fundamental role
in regulating protein degradation, recycling, secretion, and
compartmentalization; and, indeed, defective vesicular traffick-
ing is a hallmark of malignant transformation (Mosesson et al.
Our study has several limitations. First, somatic alter-
ations in noncoding regions, genome rearrangements, and copy
endometrial cancer cell line. KLE cells were transfected with siRNAs targeting the indicated genes.
Mock, risc-free siRNA, and nonspecific siRNA served as controls. (A) Efficacy of PTEN siRNA on AKT
phosphorylation was determined by Western blotting. Cells transfected with indicated siRNAs were
assayed for cell viability (B) 7 d or (C) 5 d post-transfection. Cell viability relative to mock-transfected
cells was shown. (*) P < 0.05, compared with mock control.
Functional effect of candidate driver cancer genes by siRNA-mediated gene silencing in KLE
Novel driver cancer genes in endometrial cancer
number variations would have been missed. Second, the num-
ber of samples sequenced is relatively small, thus sequencing
of additional samples will be necessary to establish a compre-
hensive genomic landscape for endometrial cancer. Finally,
additional functional and clinical studies will be required
before the insights from the genomic analysis can be trans-
lated into effective personalized therapy for endometrial
related to the PI3K pathway. (Upper panel) Mutation diagram in the full set of endometrial tumor samples (n = 222). Each column represents a tumor,
andeachrowcorresponds toasinglegene.(Lowerpanel)Mutationorco-mutationfrequencies areexpressedasapercentage of all thesamples, and the
co-mutation frequencies from random expectation are shown in parentheses for comparison. (Dark red) Genes with statistically significant co-mu-
tations, accompanied with Bonferroni-corrected P-values. (B) The functional effect of ARID1A mutationon protein expressionof the PI3K pathway. Each
arrow represents a protein marker with significant differential expression between ARID1A wild-type and mutated samples: (red arrows) phosphory-
lated proteins and (green arrows) total proteins with P < 0.05 (two-sided t-test, FDR < 0.1); (orange arrows) phosphorylated proteins with marginal
significance P < 0.07 (FDR < 0.13). (Solid red) Activated genes; (solid gray) genes without available protein expression data. (C) The functional effect of
ARID1A mutation on the phosphorylation of AKT and p70S6K in tumor samples in which both PTEN and PIK3CA genes are wild-type, and also PTEN
expression is retained (n = 47). P-values were calculated based on a two-sided t-test. (Boxes) The distribution of individual values from the lower 25th
percentile to the upper 75th percentile; (solid line in the middle) median values; (lower and upper whisker) fifth and 95th percentiles; (small circles)
outlier data points. (D) Four endometrial cancer cell lines were transfected with 20 nM ARID1A siRNA or nonspecific siRNA and harvested after 72 h for
Western blotting with the indicated antibodies. Numerical values below each lane of the immunoblots represent the quantification of the relative
protein level by densitometry.
Mutational and functional analysis of ARID1A on the activation of the PI3K pathway. (A) Co-mutation patterns of ARID1A and key genes
Liang et al.
2126 Genome Research
All studies of 222 patients diagnosed at the University of Texas MD
Anderson Cancer Center (Houston, TX, USA) from 1998 to 2009
were approved by the Institutional Review Board. Tumor content
($80%), histological classification, grade, and stage were reviewed
by two independent pathologists. Genomic DNA from frozen tu-
mor resections was extracted by the MD Anderson Bioextraction
core using the QIAamp DNA Mini Kit (QIAGEN); normal DNAwas
extracted from peripheral blood leukocytes using the QIAamp
Blood kit (QIAGEN).
Exome sequencing and detection of somatic alterations
Whole-exome libraries (captured by Agilent SureSelect All Exon
Kit) were constructed and sequenced on a SOLiD V3.0 Genome
Analyzer using 50-nt single reads. Sequencing reads were pro-
cessed, mapped to the human genome (hg19), and analyzed using
Bioscope (version 1.21) combined with in-house scripts. Supple-
mental Figure 1 shows the overall computational pipeline for
detecting somatic mutations. Mutation validation was performed
by Sequenom MassARRAY genotyping, and whole-gene resequenc-
ing on the large sample cohort was performed by Sanger sequencing.
A complete description is provided in the Supplemental Methods.
Mutation data of ARID1A in 222 endometrial tumors were
obtained as previously described (Cheung et al. 2011). A two-sided
Fisher exact test was used to assess whether (nonsilent) mutations
in ARID1Atend toco-occur withmutations in othergenes (orPI3K
pathway aberrations) in the tumor samples. PI3K pathway aber-
rations were defined as PTEN loss or mutations in any pathway
genes (including PTEN, PIK3CA, PIK3R1, PIK3R2, and AKT1). A
Bonferroni correction was used to control multiple testing.
RPPA data analysis
To examine the effect of ARID1A mutations on the PI3K pathway,
high-throughput RPPA data for 24 PI3K-pathway-related proteins
in the 222 samples were obtained from our previous study
(Cheung et al. 2011). RPPA data were processed and normalized as
previously described (Park et al. 2010). A two-sided t-test was used
to test whether the expression of a given protein in ARID1A wild-
type samples was significantly different from that in ARID1A
mutated samples, and a false discovery rate (Benjamini and
Hochberg 1995) was used to control multiple testing.
Pathway enrichment analysis was performed with Ingenuity
Pathways Analysis (IPA) (version 8.0). The status of PTEN was de-
Fisher exact test was used to assess whether ARID1A mutations
were correlated with the status of PTEN (loss or retained). To select
mutated genes for shRNA studies, functional effects of somatic
mutations were predicted with two programs: (1) CHASM (Carter
et al. 2010), mutations with FDR < 30% were selected; (2)
MutationAssessor (Reva et al. 2011), mutations with high func-
tional effect were selected.
Ba/F3 viability assays
The interleukin-3 (IL-3)–dependent prolymphoid cell line Ba/F3
was maintained in RPMI1640 medium containing 10% fetal bo-
vine serum supplemented with 5 ng/mL IL-3 at 37°C in a 5% CO2
atmosphere. Constructs containing wild-type genes or their cor-
responding mutants cloned into pLenti6.3 (Invitrogen) and
shRNAs in pGIPZ vector (Open Biosystems) were transfected into
Ba/F3 cells using the Neon electroporation system according to the
manufacturer’s instructions (Invitrogen). At 96 h post-transfection,
the cells were resuspended in medium without IL-3. Cells (5 3 103)
were plated in 96-well plates and cultured for 4 wk. Cell viability
was evaluated using Cell Titer Blue (Promega) for mitochondrial
dehydrogenase activity. Statistical analysis was performed using
ANOVA followed by Tukey’s post hoc test (GraphPad Software).
P < 0.05 was considered statistically significant.
siRNA transfection in endometrial cancer cell lines
from DSMZ-German Collection of Microorganisms and Cell Cul-
tures (Braunschweig, Germany). KLE was provided by Dr. Russell
was kindly provided by Dr. Bo R. Rueda (Massachusetts General
Hospital, Boston). All cell lines were cultured in media according
to the suppliers’ instructions at 37°C in a 5% CO2atmosphere.
Cells were transfected with 20 nM siRNAs (Dharmacon) using
Lipofectamine RNAiMAX reagent (Invitrogen) according to
the manufacturers’ instructions. The cells were harvested at
time points as indicated in the figure legends and processed for
5 mM EDTA, 1 mM sodium orthovanadate, 1% phenyl-
methylsulphonyl fluoride, and complete protease inhibitor
cocktail). The protein concentrations were determined with DC
Protein Assay Reagent (Bio-Rad Laboratories). Cell lysates (25 mg)
were loaded onto SDS/PAGE and transferred to a Hybond-ECL ni-
trocellulose membrane (Amersham Biosciences). The membrane
was blocked with 5% nonfat milk and incubated with primary
antibody overnight at 4°C. Protein expression was visualized with
an ECL plus kit (Amersham Biosciences).
The sequencing data from this study can be accessed at the Euro-
pean Genome-phenome Archive (EGA) (http://www.ebi.ac.uk/
ega/) under accession number EGAS00001000318.
This work was supported by a Stand Up to Cancer Dream Team
Translational Research Grant, a Program of the Entertainment
Industry Foundation (SU2C-AACR-DT0209) to G.B.M.; a Uterine
Cancer SPORE grant (NIH/NCI P50 CA098258) to R.R.B, K.H.L.,
and G.B.M.; U01 CA168394 to G.B.M.; a Career Development
Award from the Conquer Cancer Foundation of the American
Society of Clinical Oncology to B.T.H.; NIH/NCI R21CA152432 to
R.K.; a training fellowship from the Keck Center Computational
Cancer Biology Training Program of the Gulf Coast Consortia
(CPRIT Grant No. RP101489) to L.W.T.C.; grants from the G.S.
Hogan Gastrointestinal Research Fund and the Lorraine Dell Pro-
gram in Bioinformatics for Personalization of Cancer Medicine to
H.L.; and an Institutional Core grant (NIH/NCI P30 CA016672).
Novel driver cancer genes in endometrial cancer
Author contributions: H.L., G.B.M., and B.T.H. designed the
research; L.W.T.C., J.L., S.Y., K.S.H., T.D., Y.L., C.G., W.G., S.E.S.,
and K.L.S. performed the experimental work, including Sequenom,
Sanger sequencing, Ba/F3 viability assays, and Western blots; X.L.
and C.G.L. performed library preparation and SOLiD sequencing;
H.L., L.W.T.C., J.L., Z.J., H.C., S.N.W., M.D., R.G.W.V., F.Z., R.K.,
and G.B.M. performed data analyses; K.H.L., R.R.B., and B.T.H.
provided patient samples; H.L., L.W.T.C., J.L., and G.B.M. pro-
the project. All authors contributed to the final manuscript.
Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate—a
practical and powerful approach to multiple testing. J R Stat Soc Ser B
Methodol 57: 289–300.
Byron SA, Gartside MG, Wellens CL, Mallon MA, Keenan JB, Powell MA,
Goodfellow PJ, Pollock PM. 2008. Inhibition of activated fibroblast
growth factor receptor 2 in endometrial cancer cells induces cell death
despite PTEN abrogation. Cancer Res 68: 6902–6907.
Cancer Genome Atlas Research Network. 2008. Comprehensive genomic
characterization defines human glioblastoma genes and core pathways.
Nature 455: 1061–1068.
of ovarian carcinoma. Nature 474: 609–615.
Carter H, Samayoa J, Hruban RH, Karchin R. 2010. Prioritization of driver
mutations in pancreatic cancer using cancer-specific high-throughput
annotation of somatic mutations (CHASM). Cancer Biol Ther 10: 582–
Catasus L, Gallardo A, Cuatrecasas M, Prat J. 2008. PIK3CA mutations in the
kinase domain (exon 20) of uterine endometrial adenocarcinomas are
associated with adverse prognostic parameters. Mod Pathol 21: 131–139.
Catasus L, Gallardo A, Cuatrecasas M, Prat J. 2009. Concomitant PI3K-AKT
and p53 alterations in endometrial carcinomas are associated with poor
prognosis. Mod Pathol 22: 522–529.
Chapman MA, Lawrence MS, Keats JJ, Cibulskis K, Sougnez C, Schinzel AC,
Harview CL, Brunet JP, Ahmann GJ, Adli M, et al. 2011. Initial genome
sequencing and analysis of multiple myeloma. Nature 471: 467–472.
Cheung LW, Hennessy BT, Li J, Yu S, Myers AP, Djordjevic B, Lu Y, Stemke-
Hale K, Zhang F, Ju Z, et al. 2011. High frequency of PIK3R1 and PIK3R2
mutations in endometrial cancer elucidates a novel mechanism for
regulation of PTEN protein stability. Cancer Discov 1: 170–185.
Chin L, Hahn WC, Getz G, Meyerson M. 2011. Making sense of cancer
genomic data. Genes Dev 25: 534–555.
Cinghu S, Anandharaj A, Lee HC, Yu JR, Park WY. 2011. FTS (fused toes
homolog) a novel oncoprotein involved in uterine cervical
carcinogenesis and a potential diagnostic marker for cervical cancer.
J Cell Physiol 226: 1564–1572.
Davies H, Hunter C, Smith R, Stephens P, Greenman C, Bignell G, Teague J,
Butler A, Edkins S, Stevens C, et al. 2005. Somatic mutations of the
protein kinase gene family in human lung cancer. Cancer Res 65: 7591–
Dedes KJ, Wetterskog D, Ashworth A, Kaye SB, Reis-Filho JS. 2011.
Emerging therapeutic targets in endometrial cancer. Nat Rev Clin Oncol
Di Simone N, Schneyer AL, Caliandro D, Castellani R, Caruso A. 2002.
Regulation of endometrial adenocarcinoma cell proliferation by Activin-A
and its modulation by 17b-estradiol. Mol Cell Endocrinol 192: 187–195.
Ding L, Ellis MJ, Li S, Larson DE, Chen K, Wallis JW, Harris CC, McLellan
MD, Fulton RS, Fulton LL, et al. 2010. Genome remodelling in a basal-
like breast cancer metastasis and xenograft. Nature 464: 999–1005.
Ejskjaer K, Sorensen BS, Poulsen SS, Forman A, Nexo E, Mogensen O. 2007.
Expression of the epidermal growth factor system in endometrioid
endometrial cancer. Gynecol Oncol 104: 158–167.
Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. 2010.
GLOBOCAN 2008, Cancer incidence and mortality worldwide: IARC
CancerBase No 10. International Agency for Research on Cancer, Lyon,
Guan B, Mao TL, Panuganti PK, Kuhn E, Kurman RJ, Maeda D, Chen E, Jeng
YM, Wang TL, Shih IM. 2011a. Mutation and loss of expression of
ARID1A in uterine low-grade endometrioid carcinoma. Am J Surg Pathol
Guan B, Wang TL, Shih IM. 2011b. ARID1A, a factor that promotes
formation of SWI/SNF-mediated chromatin remodeling, is a tumor
suppressor in gynecologic cancers. Cancer Res 71: 6718–6727.
Gui Y, Guo G, Huang Y, Hu X, Tang A, Gao S, Wu R, Chen C, Li X, Zhou L,
et al. 2011. Frequent mutations of chromatin remodeling genes in
transitional cell carcinoma of the bladder. Nat Genet 43: 875–878.
Hayashi S, Okada T, Igarashi N, Fujita T, Jahangeer S, Nakamura S. 2002.
Identification and characterization of RPK118, a novel sphingosine
kinase-1-binding protein. J Biol Chem 277: 33319–33324.
PIK3CA in uterine serous carcinoma. Gynecol Oncol 113: 370–373.
Hennessy BT, Smith DL, Ram PT, Lu Y, Mills GB. 2005. Exploiting the PI3K/
AKT pathway for cancer drug discovery. Nat Rev Drug Discov 4: 988–
Hollander MC, Blumenthal GM, Dennis PA. 2011. PTEN loss in the
continuum of common cancers, rare syndromes and mouse models. Nat
Rev Cancer 11: 289–301.
Ikeda T, Yoshinaga K, Semba S, Kondo E, Ohmori H, Horii A. 2000.
Mutational analysis of the CTNNB1 (b-catenin) gene in human
endometrial cancer: Frequent mutations at codon 34 that cause nuclear
accumulation. Oncol Rep 7: 323–326.
Jia L, Liu YJ, Yi XF, Miron A, Crum CP, Kong B, Zheng WX. 2008.
Endometrial glandular dysplasia with frequent p53 gene mutation: A
genetic evidence supporting its precancer nature for endometrial serous
carcinoma. Clin Cancer Res 14: 2263–2269.
Jones S, Wang TL, Shih Ie M, Mao TL, Nakayama K, Roden R, Glas R, Slamon
D, Diaz LA Jr, Vogelstein B, et al. 2010. Frequent mutations of chromatin
remodeling gene ARID1A in ovarian clear cell carcinoma. Science 330:
Kaneda H, Arao T, Matsumoto K, De Velasco MA, Tamura D, Aomatsu K,
Kudo K, Sakai K, Nagai T, Fujita Y, et al. 2011. Activin A inhibits vascular
endothelial cell growth and suppresses tumour angiogenesis in gastric
cancer. Br J Cancer 105: 1210–1217.
Lax SF, Kendall B, Tashiro H, Slebos RJ, Hedrick L. 2000. The frequency of
p53, K-ras mutations, and microsatellite instability differs in uterine
endometrioid and serous carcinoma: Evidence of distinct molecular
genetic pathways. Cancer 88: 814–824.
LiuLL,YangCY,YuanJ,ChenXJ, XuJN,WeiYH, YangJC,Lin G,Yu L.2005.
through pseudo-kinase domains. Mol Cells 19: 39–45.
Lutz PG, Moog-Lutz C, Coumau-Gatbois E, Kobari L, Di Gioia Y, Cayre YE.
2000. Myeloblastin is a granulocyte colony-stimulating factor-
responsive gene conferring factor-independent growth to
hematopoietic cells. Proc Natl Acad Sci 97: 1601–1606.
Mamo A, Cavallone L, Tuzmen S, Chabot C, Ferrario C, Hassan S, Edgren H,
genomic approach identifies ARID1A as a candidate tumor-suppressor
gene in breast cancer. Oncogene 31: 2090–2100.
Mardis ER. 2011. A decade’s perspective on DNA sequencing technology.
Nature 470: 198–203.
Miyake T, Yoshino K, Enomoto T, Takata T, Ugaki H, Kim A, Fujiwara K,
Miyatake T, Fujita M, Kimura T. 2008. PIK3CA gene mutations and
amplifications in uterine cancers, identified by methods that avoid
confounding by PIK3CA pseudogene sequences. Cancer Lett 261:
Moreno-BuenoG,HardissonD,Sanchez C,Sarrio D,CassiaR,Garcia-Rostan
G, Prat J, Guo M, Herman JG, Matias-Guiu X, et al. 2002. Abnormalities
of the APC/b-catenin pathway in endometrial cancer. Oncogene 21:
Mosesson Y, Mills GB, Yarden Y. 2008. Derailed endocytosis: An emerging
feature of cancer. Nat Rev Cancer 8: 835–850.
MutterGL, LinMC, FitzgeraldJT,KumJB,Baak JPA,LeesJA,Weng LP,Eng C.
2000. Altered PTEN expression as a diagnostic marker for the earliest
endometrial precancers. J Natl Cancer Inst 92: 924–931.
Notaridou M, Quaye L, Dafou D, Jones C, Song H, Hogdall E, Kjaer SK,
Christensen L, Hogdall C, Blaakaer J, et al. 2011. Common alleles in
candidate susceptibility genes associated with risk and development of
epithelial ovarian cancer. Int J Cancer 128: 2063–2074.
Oda K, Stokoe D, Taketani Y, McCormick F. 2005. High frequency of
coexistent mutations of PIK3CA and PTEN genes in endometrial
carcinoma. Cancer Res 65: 10669–10673.
Park ES, Rabinovsky R, Carey M, Hennessy BT, Agarwal R, Liu W, Ju Z, Deng
W, Lu Y, Woo HG, et al. 2010. Integrative analysis of proteomic
line set. Mol Cancer Ther 9: 257–267.
Puente XS, Pinyol M, Quesada V, Conde L, Ordonez GR, Villamor N,
Escaramis G, Jares P, Bea S, Gonzalez-Diaz M, et al. 2011. Whole-genome
sequencing identifies recurrent mutations in chronic lymphocytic
leukaemia. Nature 475: 101–105.
Remy I, Michnick SW. 2004. Regulation of apoptosis by the Ft1 protein,
a new modulator of protein kinase B/Akt. Mol Cell Biol 24: 1493–1504.
Reva B, Antipin Y, Sander C. 2011. Predicting the functional impact of
protein mutations: Application to cancer genomics. Nucleic Acids Res
e118. doi: 10.1093/nar/gkr407.
Liang et al.
2128 Genome Research
Risbridger GP, Schmitt JF, Robertson DM. 2001. Activins and inhibins in
endocrine and other tumors. Endocr Rev 22: 836–858.
Risinger JI, Hayes K, Maxwell GL, Carney ME, Dodge RK, Barrett JC,
Berchuck A. 1998. PTEN mutation in endometrial cancers is associated
with favorable clinical and pathologic characteristics. Clin Cancer Res 4:
Salvesen HB, Stefansson I, Kretzschmar EI, Gruber P, MacDonald ND, Ryan
A, Jacobs IJ, Akslen LA, Das S. 2004. Significance of PTEN alterations in
endometrial carcinoma: A population-based study of mutations,
promoter methylation and PTEN protein expression. Int J Oncol 25:
Schlosshauer PW, Pirog EC, Levine RL, Ellenson LH. 2000. Mutational
analysis of the CTNNB1 and APC genes in uterine endometrioid
carcinoma. Modern Pathol 13: 1066–1071.
Schramm A, von Schuetz V, Christiansen H, Havers W, Papoutsi M,
Wilting J, Schweigerer L. 2005. High activin A-expression in
human neuroblastoma: Suppression of malignant potential and
correlation with favourable clinical outcome. Oncogene 24: 680–
SederCW,HartojoW,Lin L, Silvers AL, Wang ZW,Thomas DG, Giordano TJ,
Chen GA, Chang AC, Orringer MB, et al. 2009. Upregulated INHBA
expression may promote cell proliferation and is associated with poor
survival in lung adenocarcinoma. Neoplasia 11: 388–396.
Shoji K, Oda K, Nakagawa S, Hosokawa S, Nagae G, Uehara Y, Sone K,
Miyamoto Y, Hiraike H, Hiraike-Wada O, et al. 2009. The oncogenic
mutation in the pleckstrin homology domain of AKT1 in endometrial
carcinomas. Br J Cancer 101: 145–148.
Sithanandam G, Anderson LM. 2008. The ERBB3 receptor in cancer and
cancer gene therapy. Cancer Gene Ther 15: 413–448.
Srinivasan R, Benton E, McCormick F, Thomas H, Gullick WJ. 1999.
Expression of the c-erbB-3/HER-3 and c-erbB-4/HER-4 growth factor
receptors and their ligands, neuregulin-1 a, neuregulin-1 b, and
betacellulin, in normal endometrium and endometrial cancer. Clin
Cancer Res 5: 2877–2883.
Stefansson IM, Salvesen HB, Akslen LA. 2004. Prognostic impact of
endometrial cancer. J Clin Oncol 22: 1242–1252.
Stephens P, Edkins S, Davies H, Greenman C, Cox C, Hunter C, Bignell G,
Teague J, Smith R, Stevens C, et al. 2005. A screen of the complete
protein kinase gene family identifies diverse patterns of somatic
mutations in human breast cancer. Nat Genet 37: 590–592.
Y. 2001. Mutational analysis of the PTEN gene in endometrial carcinoma
and hyperplasia. Am J Clin Pathol 115: 32–38.
Tanaka T, Toujima S, Otani T, Minami S, Yamoto M, Umesaki N. 2003.
Expression and function of activin receptors in human endometrial
adenocarcinoma cells. Int J Oncol 23: 657–663.
Toyama T, Lee HC, Koga H, Wands JR, Kim M. 2010. Noncanonical Wnt11
inhibits hepatocellular carcinoma cell proliferation and migration. Mol
Cancer Res 8: 254–265.
Velasco A, Bussaglia E, Pallares J, Dolcet X, Llobet D, Encinas M, Llecha N,
Palacios J, Prat J, Matias-Guiu X. 2006. PIK3CA gene mutations in
endometrial carcinoma: Correlation with PTEN and K-RAS alterations.
Hum Pathol 37: 1465–1472.
Wang K, Kan J, Yuen ST, Shi ST, Chu KM, Law S, Chan TL, Kan Z, Chan AS,
Tsui WY, et al. 2011. Exome sequencing identifies frequent mutation of
ARID1A in molecular subtypes of gastric cancer. Nat Genet 43: 1219–
WarmuthM,Kim S, GuXJ, XiaG, Adrian F.2007. Ba/F3 cells and theiruse in
kinase drug discovery. Curr Opin Oncol 19: 55–60.
Wiegand KC, Shah SP, Al-Agha OM, Zhao Y, Tse K, Zeng T, Senz J,
McConechy MK, Anglesio MS, Kalloger SE, et al. 2010. ARID1A
mutations in endometriosis-associated ovarian carcinomas. N Engl J Med
Wiegand KC, Lee AF, Al-Agha OM, Chow C, Kalloger SE, Scott DW, Steidl C,
Wiseman SM, Gascoyne RD, Gilks B, et al. 2011. Loss of BAF250a
(ARID1A) is frequent in high-grade endometrial carcinomas. J Pathol
Wu JI, Lessard J, Crabtree GR. 2009. Understanding the words of chromatin
regulation. Cell 136: 200–206.
Yoda A, Yoda Y, Chiaretti S, Bar-Natan M, Mani K, Rodig SJ, West N, Xiao Y,
Brown JR, Mitsiades C, et al. 2010. Functional screening identifies
CRLF2 in precursor B-cell acute lymphoblastic leukemia. Proc Natl Acad
Sci 107: 252–257.
Received January 11, 2012; accepted in revised form July 2, 2012.
Novel driver cancer genes in endometrial cancer