doi:10.

1093/jmcb/mjv015
Published online March 10, 2015

Journal of Molecular Cell Biology (2015), 7(2), 105 118

| 105

Article

p53 deficiency-induced Smad1 upregulation

suppresses tumorigenesis and causes
chemoresistance in colorectal cancers
Xinsen Ruan1, , Qiao Zuo2, , Hao Jia1, , Jenny Chau3, Jinlin Lin1, Junping Ao4, Xuechun Xia1, Huijuan Liu1,
Samy L. Habib5, Chuangang Fu2, *, and Baojie Li1,*
1

The DNA damage response helps to maintain genome integrity, suppress tumorigenesis, and mediate the effects of radiotherapy and
chemotherapy. Our previous studies have shown that Smad1 is upregulated and activated by Atm in DNA damage response, which can
further bind to p53 and promote p53 stabilization. Here we report another aspect of the interplay between p53 and Smad1. Comparison
of rectal tumor against paired paraneoplastic specimens and analysis of >500 colorectal tumors revealed that Smad1 was upregulated
in tumor samples, which was attributable to p53 defects. Using MEFs as a model, we found that knockdown of the elevated Smad1 in
p532/2 MEFs promoted cell proliferation, E1A/Ras-induced cell transformation, and tumorigenesis. Mechanistic studies suggest that
elevated Smad1 and momentary activation inhibit cell proliferation by upregulating p57Kip2 and enhancing Atm Chk2 activation.
Surprisingly, elevated Smad1 appears to have a negative effect on chemotherapy, as colorectal tumors, primary cancer cells, and
cell lines with Smad1 knockdown all showed an increase in chemosensitivity, which could be attributable to elevated p57Kip2.
These findings underscore the significance of Smad1 p53 interaction in tumor suppression and reveal an unexpected role for
Smad1 in chemoresistance of colorectal cancers.

Keywords: Smad1, p53, p57Kip2, chemoresistance, colorectal cancer

Introduction
Tumor suppressor p53 is regarded as the guardian of the
genome. Activation of p53 by genotoxic stress or oncogene activation turns on its target genes such as p21, Puma, and Bax to induce
cell cycle arrest, apoptosis, and/or senescence, thus maintaining
genome integrity and suppressing tumorigenesis caused by accumulation of mutations in oncogenes and tumor suppressor genes
(Jackson and Bartek, 2009; Lord and Ashworth, 2012; Reinhardt
and Schumacher, 2012). As such, p53 gene is mutated in .50%
of the primary tumors, which express mutant p53 molecules that
either lose the normal function or display dominant-negative
effects (Vogelstein et al., 2000; Goh et al., 2011; Muller and
Vousden, 2013). The lack of p53 function automatically leads to

Received June 10, 2014. Revised November 4, 2014. Accepted November 21, 2014.
# The Author (2015). Published by Oxford University Press on behalf of Journal of
Molecular Cell Biology, IBCB, SIBS, CAS. All rights reserved.

escape of cell senescence and immortalization and promotes cell

transformation. In addition, p53 deficiency has been shown to
affect cell differentiation in several cell types, including neuron
and osteoblast (Ma et al., 2012; Liu et al., 2013), although the
mechanisms by which p53 regulates cell differentiation remain
under-explored.
Due to its critical roles in cell proliferation, differentiation, and
death, the p53 expression needs to be tightly regulated. A
Mdm2 p53 loop provides such a mechanism for fine-tuning p53
expression (Brooks and Gu, 2006). Moreover, the levels of p53
are also influenced by the energy level, nutrition, and other
growth conditions of the cell, in addition to the severity of DNA
damage. For example, it has been shown that growth factorsactivated mTOR S6K1 pathway has an influence on p53 induction
in response to DNA damage (Lai et al., 2010). In addition, p53
deficiency leads to positive feedback regulation of the expression
of genes that have redundant function as p53, e.g. p16INK4a

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Bio-X Institutes, Key Laboratory for the Genetics of Developmental and Neuropsychiatric Disorders, Ministry of Education, Shanghai Jiao Tong University,
Shanghai 200240, China
2
Department of Colorectal Surgery, Changhai Hospital, Second Military Medical University, Shanghai 200433, China
3
Institute of Molecular and Cell Biology, Proteos, 61 Biopolis Drive, Singapore 138673, Singapore
4
State Key Laboratory of Oncogenes and Related Genes, Shanghai Cancer Institute, Renji Hospital, Shanghai Jiao Tong University, School of Medicine,
Shanghai 200032, China
5
Department of Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX 78229, USA

These authors contributed equally to this work.

* Correspondence to: Baojie Li, E-mail: libj@sjtu.edu.cn; Chuangang Fu, E-mail: fugang416@126.com

106

| Ruan et al.

Results
Rectal tumor samples show Smad1 upregulation compared with
paraneoplastic tissues
Our previous studies have shown that Atm-mediated Smad1
Ser239 phosphorylation promotes Smad1 upregulation and activation and leads to p53 stabilization. To test the clinical relevance of
this finding, we collected 22 rectal tumors (5 stage I samples, 6
stage II samples, 8 stage III samples, and 3 stage IV samples; for
patient information, see Supplementary Table S1, sample 9
being discarded due to protein degradation) and paraneoplastic
specimens and analyzed Smad1 expression. We found that both
protein (Figure 1A and B) and mRNA (Figure 1C and D) levels of
Smad1 were significantly upregulated in the tumor samples compared with matched paraneoplastic samples. However, the
protein levels of Smad4, an important tumor suppressor
(Derynck et al., 2001), and Smad5 were not significantly altered
(Supplementary Figure S1). Surprisingly, we found that Smad1/
5/8 activation was reduced in most of the tumor samples even in
the presence of elevated Smad1 expression (Figure 1A). This is consistent with previous findings that Smad1/5/8 activation is inhibited in colorectal tumor samples due to mutations in bone

morphogenetic protein (BMP) receptors (Sancho et al., 2004;

Kodach et al., 2008), and suggests that BMP Smad1 signaling
plays complex roles in tumor development and progression.
Smad1 upregulation is common in human colorectal cancer
samples
To validate the finding of Smad1 upregulation in tumors, we collected 542 human colorectal tumor specimens (stages I to IV) and
53 normal tissues, which were obtained from Chinese Han population (for patient information, see Supplementary Table S2). These
tissues were used to generate tissue arrays, which were immunohistochemically stained for Smad1, and the levels of Smad1 were
scored (Supplementary Figure S2A). We found that Smad1 was
expressed at low levels in normal mucosa samples, yet tumor
patient samples expressed increased levels of Smad1 (from 17%
to 64%) (Table 1). At later stages, Smad1 levels appeared to go
down. In the tumor samples with increased Smad1 expression,
immunohistochemical staining showed that Smad1 was detectable
in both the cytoplasm and the nucleus, just like in normal tissues
(Supplementary Figure S2B). These results indicate that colorectal
tumors tend to upregulate Smad1 expression. The function of
Smad1 upregulation is the focus of present study.
Smad1 upregulation is attributable to p53 defects
Previous studies have shown that primary p532/2 osteoblasts
and neural stem cells (NSCs) displayed elevated Smad1 expression, which have an impact on osteogenic and neural differentiation, respectively (Ma et al., 2012; Liu et al., 2013). To
determine whether Smad1 upregulation is related to p53 in
tumor samples, we analyzed the protein levels of p53 in 22 pairs
of samples by western blot and found that p53 was either downregulated or expressed in truncated forms in .70% tumor
samples (Figure 1A). Sequencing the p53 cDNA obtained from 22
samples confirmed the existence of various p53 mutations
(Supplementary Table S1). Five of the tumor samples carry
R175H mutations, and two carry G245R mutation. Moreover, 16
of them carry P72R, a common polymorphism of p53 gene that is
associated with tumorigenesis (Aaltonen et al., 2001; Olivier
et al., 2010), two of which also carry R248W and R282W mutations.
Some of the tumors that do not carry mutations showed a decrease
in p53 protein level. These results suggest that p53 defects are
common in colorectal cancers.
Comparison of p53 down-regulation/mutations with Smad1
upregulation revealed that there exists a significant correlation
between p53 defects and Smad1 upregulation (Figure 2A), suggesting that p53 suppresses the expression of Smad1 in human
colorectal tumors. We then isolated primary rectal cancer cells
from three patient samples, which showed a great reduction in
p53 and an elevation in Smad1 compared with its paired paraneoplastic sample (Figure 2B and Supplementary Figure S3A). The
sample shown in Figure 2B carries a R175H mutation in p53
gene. To test whether the increase in Smad1 is a result of p53 deficiency, we ectopically expressed p53 using a retroviral vector in
the tumor cells and found that p53, but not p53-R273H mutant,
reduced the protein level of Smad1 (Figure 2C and D, and

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

and p73 (Kravchenko et al., 2008; Leong et al., 2009; Lunghi et al.,
2009; Hong et al., 2014). It is understandable that an important
protein such as p53 needs to be tightly regulated and when
mutated, compensatory mechanisms need to kick in.
Our previous studies showed that Samd1 is upregulated and
further activated in response to DNA damage. DNA lesions especially double-stranded DNA breaks activate the Atm p53 and
Atm Chk2 pathways to induce cell cycle arrest and apoptosis
(Jackson and Bartek, 2009). DNA damage-induced Smad1 activation requires Atm and Atm-mediated phosphorylation of Smad1
on Ser239, which also promotes Smad1 p53 interaction and inhibits Mdm2-mediated p53 ubiquitination, leading to p53 upregulation (Chau et al., 2012). This provides a mechanistic explanation
how BMP Smad1 signaling suppresses tumor development
(Howe et al., 2001; Hardwick et al., 2008; Thawani et al., 2010;
Walsh et al., 2010; Tomlinson et al., 2011; Gao et al., 2012;
Lubbe et al., 2012). In this study, we uncovered a novel aspect of
the functional interaction between p53 and Smad1. We found
that Smad1 is upregulated in human colorectal tumor specimens
compared with normal tissues or paired paraneoplastic samples,
which is attributable to p53 deficiency. This elevation helps to
curb cell proliferation, cell transformation, and tumor formation,
by increasing the expression of cyclin-dependent kinase inhibitor
p57Kip2, a Smad1 target gene (Jia et al., 2014), and potentiating
Atm Chk2 activation. In addition, elevated Smad1 appears to
render chemoresistance in tumor models and human rectal
cancer cells, which could be explained by increased p57Kip2 expression. These findings suggest that Smad1 elevation serves as
a compensatory mechanism for p53 deficiency by potentiating
the activation of p53 parallel pathways, and that Smad1 plays critical roles not only in tumor suppression but also in chemosensitivity. In addition, parallel studies show that the functions of Smad1 in
tumor suppression and chemosensitivity are not shared by Smad5.

Smad1 compensates for p53 deficiency | 107

Table 1 Expression of Smad1 in normal mucosa and stage I IV colorectal cancer.

Characteristic

Normal mucosa
Stage I
Stage II
Stage III
Stage IV

Smad1 immunostaining distribution (%)

2

11

1 11

0
0
0
0
0

81.1
77.8
31.1
45.6
60.4

17.0
22.2
64.0
50.6
39.6

1.9
0
4.9
3.8
0

Total patients

Positive rate

Mean rank*

P-value**

53
45
286
158
53

1
1
1
1
1

222.43
229.83
383.36
336.49
284.12

0.002a
,0.001b
0.003c
0.043d

n 595.
*KruskalWallis test was used to determine significances among the six groups (P , 0.001).
**Mann Whitney U-test was used to determine significances between each two groups.
a
Compared with the Normal mucosa group.
b
Compared with the Stage I cancer group.
c
Compared with the Stage II cancer group.
d
Compared with the Stage III cancer group.

Supplementary Figure S3B). These results, taken together, suggest

that p53 defects contribute to Smad1 elevation.
Knockdown of Smad1 promotes proliferation of p532/2 MEFs
p53 mutation is a frequent and early event in some tumor types
(Rivlin et al., 2011; Vinall et al., 2012). Our previous studies showed

that p53 is often mutated during MEF immortalization, a process

that is required for cell transformation (Zhang et al., 2013a). We
used primary MEFs to test the function of Smad1 elevation in relation to p53 deficiency. We found that primary p532/2 MEFs showed
elevated levels of Smad1 but not Samd5 (Figure 3A). It has been
shown that p53 deficiency induced Smad1 upregulation in

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 1 Smad1 is upregulated in clinical rectal tumor samples. (A) Total proteins of tumor (T) and paraneoplastic (P) samples from 22 rectal cancer
patients were extracted, and the levels of Smad1, p53, and p-Smad1/5/8 were determined by western blot. (B) Quantitation data for A. n 22.
* P , 0.05 vs. paraneoplastic tissues. (C) Real-time PCR was used to determine the mRNA levels of Smad1, showing upregulation in human rectal
tumors. (D) Quantitation data for C. n 22. *P , 0.05 vs. paraneoplastic tissues.

108

| Ruan et al.

osteoblasts and NSCs during differentiation via E2F1 (Ma et al.,

2012). To confirm that p53 represses Smad1 expression under
normal growth conditions via the same mechanisms, we knocked
down E2F1 with siRNA in p532/2 MEFs, which also led to a decrease in Smad1 (Figure 3A). Importantly, Smad1 elevation is accompanied by an increase in Smad1/5/8 activation in p532/2
MEFs. The findings that primary MEFs, osteoblasts, and NSCs
show co-elevation of Smad1 expression and Smad1/5/8 activation
in response to p53 deficiency suggest that p53 defects initially lead
to elevated Smad1 expression and activation, which is later deactivated due to alteration in upstream regulators of BMP Smad1
signaling during tumorigenesis.
To test the function of Smad1 upregulation in p532/2 cells, we
knocked down Smad1 in p532/2 MEFs using short hairpin RNA
(shRNA) expressed by a retroviral vector, with WT MEFs as a
control (Figure 3B). Four shRNA hairpins were tested and two of
them (37928 and 30824) produced similar degrees of Smad1
knockdown as siRNA (Figure 3B and Supplementary Figure S4),
and thus were used further in this study. We found that p532/2
MEFs with Smad1 knockdown showed enhanced proliferation
(Figure 3C), evidenced by an increase in the number of cells after
the same number of cells were plated. Similar results were
obtained from Smad1-knockdown WT MEFs (Figure 3C). In

addition, analysis of the cell cycle profiles revealed that Smad1

knockdown led to an increase in S phase especially in p532/2
cells (Figure 3D). These findings suggest that Smad1 may regulate
cell proliferation in p53-independent manners.
Previous studies have demonstrated functional redundancy
between Smad1 and Smad5 and the doses of Smad1 and Smad5
alleles are critical in embryonic development (Arnold et al., 2006;
Eivers et al., 2008). To test a possible dose effect on cell proliferation, we knocked down Smad5 and found that this had no significant effect on cell proliferation of either p532/2 or WT MEFs,
whereas knockdown of both Smad1 and Smad5 showed similar
results as Smad1 knockdown (Figure 3C), suggesting that Smad1
and Smad5 play different roles in MEF proliferation. There is
reported evidence that Smad1 and Smad5 have distinct or even opposite functions in vivo (Dick et al., 1999; Liu et al., 2003;
McReynolds et al., 2007). Alignment of mouse Smad1 and Smad5
protein sequences reveals that the identity between these two proteins is 88%. The divergent region is the linker located between
the MH1 and MH2 domains (Supplementary Figure S5), which
contains residues that are phosphorylated by various kinases,
e.g. mitogen-activated protein kinases (MAPKs), glycogen synthase kinase (GSK), and Atm (Fuentealba et al., 2007; Sapkota
et al., 2007; Eivers et al., 2008; Chau et al., 2012). A few potential

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 2 Smad1 upregulation is correlated with and attributable to p53 defects in clinical rectal tumor samples. (A) Pearsons correlation analysis
shows that there exists a significant correlation between Smad1 upregulation and p53 defects in rectal tumor samples. (B) Primary tumor cells were
isolated from samples of rectal cancer patients, and Smad1 and p53 levels were analyzed by western blot. (C) p53 or p53-R273H mutant was
expressed in the primary tumor cells by retroviruses, with pMSCV-vector as control, and the protein levels of p53 and Smad1 were determined
by western blot. (D) Quantification of p53 and Smad1 (normalized to b-Actin). n 3. *P , 0.05 vs. retroviral vector-infected cells.

Smad1 compensates for p53 deficiency | 109

Smad1 protein level, which could be brought down by E2F1 knockdown. Primary p53+/+ and p532/2 MEFs were transfected with control or E2F1
siRNA, and the levels of Smad1, Smad5, and p-Smad1/5/8 were analyzed by western blot after 2 days post-transfection. (B) Western blot analysis
of Smad1 and Smad5 after the cells were infected with control, Smad1 (shRNA 37928), or Smad5 shRNA-expressing retrovirus. (C) Viable cells were
determined by trypan blue staining of p53+/+ and p532/2 MEFs with Smad1, Smad5, or Smad1/5 knockdown. n 6. *P , 0.05 vs. vectorinfected cells of the same genotype. (D) Smad1 knockdown led to an increase in S phase and a decrease in G1 phase in p532/2 MEFs. n 6.
*P , 0.05 vs. vector-infected cells. (E) Comparison of the potential phosphorylation sites in the linker regions of Smad1 and Smad5, with different
Ser or Thr residues framed.

phosphorylation sites are not shared by Smad1 and Smad5

(Figure 3E), e.g. Ser181, Ser191, and Thr235. Yet, how the linker
region including the phosphorylation sites determines distinct
functions of Smad1 and Smad5 warrants further investigation.
Knockdown of Smad1 promotes transformation of p532/2 cells
We next tested whether Smad1 elevation in p532/2 MEFs has
any effect on cell transformation. Using a standard cell transformation procedure, we expressed E1A and RasV12 in p532/2

primary MEFs and cell transformation rates were scored. WT

MEFs were used as a control in these experiments. It was found
that Smad1 knockdown enhanced transformation of p532/2
MEFs (Figure 4A and B). Our previous studies have shown that overexpression of Smad1 inhibits cell transformation in p532/2 or WT
MEFs (Chau et al., 2012). These results suggest that Smad1 elevation has an inhibitory effect on cell transformation in p532/2 cells.
Similar to the cell proliferation results (Figure 3B), Smad5 knockdown showed little effect on cell transformation rates whereas

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 3 Knockdown of Smad1, but not Smad5, further promotes proliferation of p532/2 MEFs. (A) Primary p532/2 MEFs showed an increase in

110

| Ruan et al.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 4 Knockdown of Smad1 enhanced Ras/E1A-induced transformation and tumorigenesis of p532/2 MEFs. (A) Primary p53+/+ and p532/2
MEFs with Smad1, Smad5, or Smad1/5 knockdown were infected with Ras/E1AV12-expressing retroviruses, and the colony forming units were
stained with giemsa. (B) Statistics analysis of the numbers of colony forming units. n 6. *P , 0.05 vs. control shRNA-infected cells of the
same genotype. (C) p532/2 MEFs with Smad1, Smad5, or Smad1/5 knockdown were immortalized with E1A/RasV12 and then were transplanted
into nude mice. Tumor volumes were measured twice on Days 12 and 14 after the subcutaneous injection. (D) Same experiments were performed
as in C except that p53+/+ MEFs were used and that tumor volumes were measured on Days 15 and 17 after the subcutaneous injection. n 6.
*P , 0.05 vs. control shRNA-infected cells of the same genotype.

knockdown of both Smad1 and Smad5 gave rise to similar results

as Smad1 knockdown (Figure 4A and B). These results demonstrate
an inhibitory role for Smad1 in cell proliferation and transformation, a function not shared by Smad5.

Knockdown of Smad1 promotes tumor formation of p532/2 cells

We then tested whether Smad1 elevation in p532/2 MEFs has
any effect on tumorigenesis. The transformed MEFs described
in Figure 4A and B were transplanted into nude mice. Smad1

Smad1 compensates for p53 deficiency | 111

knockdown in p532/2 MEFs accelerated tumorigenesis (Figure 4C),

which was similarly observed in WT MEFs (Figure 4D). Histological
analysis of the tumors revealed that knockdown of Smad1 increased
the cellularity of tumor (Supplementary Figure S6). These findings
suggest that Smad1 elevation in p532/2 cells helps to repress cell
transformation and tumor formation via suppressing cell proliferation, and thus is likely to act as a compensatory mechanism for
p53 deficiency, indicating that Smad1 must repress cell proliferation
via p53-independent mechanisms.
Although Smad5 knockdown did not significantly alter cell proliferation and transformation, it inhibited tumorigenesis of p532/2
cells but not WT cells. Knockdown of both Smad1 and Smad5
gave rise to results in between Smad1 knockdown and Smad5
knockdown (Figure 4C and D), suggesting that Smad5 plays a
role opposite to Smad1 in p532/2 tumor growth.

Smad1 knockdown sensitizes colorectal cells and cell lines to

genotoxic stress-induced cell death
To confirm the role for Smad1 in chemosensitivity, we tested
three colorectal tumor cell lines HCT116, Caco-2, and HT29.

Smad1 elevation/activation increases p57Kip2 expression and

Atm Chk2 activation
We then wanted to understand how elevated Smad1 expression/
activation inhibits cell proliferation and transformation in
p53-deficient cells? Our recent study indicates that CDK inhibitor
p57Kip2 is a Smad1 target gene in DDR and it renders chemoresistance by suppressing cell death (Jia et al., 2014). In consistent with
the increase in Smad1 expression and activation, we found that the
levels of p57Kip2 were increased at basal level and in response to
Dox in p532/2 MEFs, which was diminished by Smad1 knockdown
(Figure 7A). We have previously shown that p57Kip2 was upregulated in 22 rectal tumors compared with the matched paraneoplastic specimens (Jia et al., 2014), which were the same samples used
in this study (Figure 1A and Supplementary Figure S1). We compared Smad1 upregulation and p57Kip2 upregulation in these
tumor samples and found that there exists a significant association between these two events (Supplementary Figure S9A).
Knockdown of Smad1 in human rectal tumor cells also led to a
decrease in p57Kip2 (Supplementary Figure S9B), suggesting
that p57Kip2 elevation in colorectal tumors is mediated by
Smad1. Moreover, ectopic expression of Smad1 in WT MEFs led
to an increase in p57Kip2, which was not further increased by
Dox treatment (Figure 7B). In addition, Dox-induced Atm Chk2 activation was enhanced in p532/2 MEFs, which could be impeded
by Smad1 knockdown (Figure 7A), and elevated expression of
Smad1 was able to potentiate Atm Chk2 activation in response
to Dox (Figure 7B). These results suggest that the BMP Smad1
pathway, in addition to the Atm p53 pathway, also plays a positive
role in Atm Chk2 activation in DDR. The increase in p57Kip2 expression and Atm Chk2 activation may explain how elevated
Smad1 suppresses cell proliferation and transformation, whereas
increased expression of p57Kip2 may mediate the effects of
Smad1 on chemoresistance.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Smad1 knockdown sensitizes tumors to the chemotherapeutic

effects of Dox
We then asked whether elevated levels of Smad1 in tumor
cells have any effect on chemotherapeutic efficacy. We took advantage of the tumor models developed in nude mice and
treated them with Dox for a week and monitored the size of
tumors. We started Dox treatment when the tumor sizes reached
200 mm3. It was found that tumors derived from p532/2 cells
with Smad1 knockdown showed increased sensitivity to Dox treatment (Figure 5A). Moreover, Smad1 knockdown also increased
the sensitivity to Dox treatment in tumors derived from WT cells
(Figure 5B).
We found that Smad1 could be activated by Dox in transformed
MEFs, cancer cell lines, and tumors, which may regulate cell proliferation and/or death [(Chau et al., 2012) and data not shown]. We
stained the cancer samples for cell proliferation marker Ki67 and
found that Smad1 knockdown tumors showed a greater reduction
in the number of S phase cells than control tumors in response to
Dox treatment (Supplementary Figure S7). Similarly, transformed
MEFs with Smad1 knockdown also showed a modest decrease in
cell proliferation, judged by BrdU incorporation, and a significant
decrease in cell survival rate in response to Dox treatment
(Figure 5C and D). These cells with Smad1 knockdown also
showed a decrease in cell survival rate in response to two genotoxic
drugs 5-iodotubercidin and oxaliplatin (Figure 5E and F) (Zhang
et al., 2013b). These results suggest that Smad1 knockdown
might improve chemosensitivity by inhibiting cell proliferation
and survival.
We found that tumors derived from p532/2 cells with Smad5
knockdown showed decreased sensitivity to Dox treatment, and
knockdown of both Smad1 and Smad5 showed sensitivity similar
to control cells, in between Smad1 knockdown and Smad5 knockdown (Figure 5A), suggesting that Smad5 might also play a role opposite to Smad1 in chemosensitivity.

Smad1 knockdown with siRNA in these cell lines resulted in a decrease in BrdU incorporation, although to modest extents, and sensitized these cells to Dox-induced cell death, to a greater extent
than in MEFs (Figure 6A C). Primary human rectal tumor cells
with Smad1 knockdown also showed a significant decrease in
cell proliferation and survival in response to Dox treatment
(Figure 6D and E). Smad1 knockdown also increased the sensitivity
of primary cancer cells to anti-cancer drug oxaliplatin (Figure 6F).
Further analysis showed that Dox-induced cell death mainly occurred by apoptosis, with a small percentage of cells undergoing
necrosis in cells with Smad1 knockdown (Supplementary Figure
S8A). It has been reported that Dox might kill cancer cells via generating reactive oxygen species (ROS). We found that pre-treating
the cells with 5 mM N-Acetylcysteine (NAC), a ROS scavenger,
only modestly inhibited Dox-induced cell death (Supplementary
Figure S8B), suggesting that ROS may play a minor role in
Dox-induced cell death under this setting. Nevertheless, the
results derived from in vivo tumor model, colorectal cancer lines,
and primary rectal cancer cells all indicate that reducing Smad1
levels can increase chemosensitivity by inhibiting cell proliferation
and survival, with the inhibition on survival more pronounced.

112

| Ruan et al.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 5 Smad1 knockdown sensitizes tumors and transformed MEFs to chemotherapy. (A) p532/2 MEFs with Smad1, Smad5, or Smad1/5 knockdown were immortalized with E1A/RasV12 and then were transplanted into nude mice. After 14 days, the tumor volumes were measured while nude
mice were treated with 5 mg/kg Dox every other day for five times. n 6. *P , 0.05 when Smad1 shRNA-expressing cells were compared with
control shRNA-infected cells. (B) Same experiments were carried out as in A except that p53+/+ MEFs were used and the tumor volumes were
measured after 17 days. n 6. *P , 0.05 when Smad1 shRNA-expressing cells were compared with control shRNA-infected cells. (C) Smad1
knockdown in transformed p532/2 or p53+/+ MEFs resulted in decreased proliferation in response to Dox. p53+/+ cells were treated with
0.1 mM Dox for 24 h, while p532/2 MEFs were treated with 0.5 mM Dox for 24 h. Cell proliferation rates were determined with BrdU assay, and
the percentage of BrdU-positive cells of each group was shown. n 6. *P , 0.05 vs. control shRNA-infected cells of the same genotype under
Dox treatment. (D) Smad1 knockdown in transformed p532/2 or p53+/+ MEFs resulted in a decrease in cell survival in response to Dox treatment.
Cell viability was determined with WST-1 assay. n 6. *P , 0.05 vs. control shRNA-infected cells of the same genotype under Dox treatment. (E)
Smad1 knockdown in transformed p532/2 MEFs resulted in a decrease in cell survival in response to oxaliplatin. n 6. *P , 0.05 vs. untreated
cells. (F) Smad1 knockdown in transformed p532/2 MEFs resulted in a decrease in cell survival in response to 5-iodotubercidin (ITU) treatment.
n 6. *P , 0.05 vs. untreated cells.

Smad1 compensates for p53 deficiency | 113

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 6 Smad1 knockdown leads to enhanced chemosensitivity in colorectal tumor cell lines and primary human rectal cancer cells. (A) Western
blot result shows that Smad1 levels were reduced by siRNA in HCT116, Caco-2, and HT29 cells. (B) Smad1 knockdown led to a modest decrease in
BrdU incorporation in response to Dox treatment in HCT116, Caco-2, and HT29 cells compared with control siRNA-transfected cells. These cells
were treated with 0.1 mM Dox for 24 h. Cell proliferation rates were determined with BrdU assay, and the percentage of BrdU-positive cells of
each group was shown. n 3. *P , 0.05 vs. control siRNA-transfected cells under Dox treatment. (C) Colorectal cell lines with Smad1 knockdown
were treated with Dox for 48 h, and cell viability was determined with WST-1 assay. n 3. *P , 0.05 vs. control siRNA-transfected cells under Dox
treatment. (D) Primary rectal cancer tumor cells with Smad1 knockdown were treated with 0.1 mM Dox for 24 h. Cell proliferation rates were determined with BrdU assay, and the percentage of BrdU-positive cells was shown. Inset: western blot results showing knockdown of Smad1. n 3.
*P , 0.05 vs. control siRNA-transfected cells under Dox treatment. (E) Primary rectal tumor cells with Smad1 knockdown were treated with Dox for
48 h and cell viability was determined with WST-1 assay. n 3. *P , 0.05 vs. control siRNA-transfected cells under Dox treatment. (F) Smad1
knockdown in human rectal cancer cells resulted in a decrease in cell survival in response to oxaliplatin. n 3. *P , 0.05 vs. control
siRNA-transfected cells under oxaliplatin treatment.

114

| Ruan et al.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Figure 7 Elevated Smad1 expression/activation promotes p57Kip2 expression and potentiates Atm Chk2 activation. (A) Primary p532/2 MEFs
showed an increase in p57Kip2 expression and Atm Chk2 activation, which was diminished by Smad1 knockdown. Primary p53+/+ and p532/2
MEFs were transfected with control or Smad1 siRNA for 2 days, and were then challenged with Dox. The protein levels of Atm, Chk2, p57Kip2, Actin,
Smad1, p-Atm, and p-Chk2 were analyzed by western blot. Right panels show quantitation data. n 3. *P , 0.05 vs. p53+/+ cells under the same
treatment. **P , 0.05 vs. control siRNA-transfected cells of the same genotype under the same treatment (color matched). (B) Ectopic expression
of Smad1 in WT MEFs increased p57Kip2 expression and Atm Chk2 activation. Primary WT MEFs were infected with Smad1-expressing retrovirus
or empty retrovirus for 2 days, and were then challenged with Dox. The protein levels of Atm, Chk2, p57Kip2, Actin, Smad1, p-Atm, and p-Chk2 were
analyzed by western blot. Right panels show quantitation data. n 3. *P , 0.05 vs. cells infected with empty vector under the same treatment. (C)
A diagram showing that elevated Smad1 expression/activation inhibits cell proliferation/transformation via increasing p57Kip2 expression and
Atm Chk2 activation, and causes chemoresistance via increasing p57Kip2 expression.

Smad1 compensates for p53 deficiency | 115

a dynamic change of Smad1 expression and activation in colorectal

cancer, which appears to play critical roles in tumor suppression.
This study also provides evidence to support the concept that
there is a division of labor between Smad1 and Smad5 in tumorigenesis. It is generally believed that Smad1, 5, and 8 have redundant functions. Here we show that only Smad1 is upregulated in
the absence of p53, and only Smad1 knockdown appears to
enhance cell transformation and tumor formation and sensitize
tumor cells to Dox treatment. Moreover, only Smad1 is activated
and upregulated in response to DNA damage (Chau et al., 2012).
Functionally, while Smad5 knockdown did not significantly affect
cell proliferation and cell transformation, it seems to play a role opposite to Smad1 in tumor growth and chemosensitivity to Dox.
Mechanistically, the different functions of Smad1 and Smad5
could be caused by targeting different sets of genes or by responding to different ligands, as revealed during embryonic hematopoiesis (Liu et al., 2003; McReynolds et al., 2007).
While elevated Smad1 helps to inhibit cell proliferation and
tumorigenesis in p53-deficient cells, our findings suggest that elevated Smad1 also renders chemoresistance in tumors and tumor
cells. This may be explored to enhance chemosensitivity in treatment of colorectal cancer (Lee et al., 2011; Langenfeld et al.,
2013). Our present and previous studies suggest that p57Kip2
may be an important mediator of the effects of elevated Smad1
on chemoresistance (Figure 7C), based on the following observations. Firstly, p57Kip2 is co-elevated with Smad1 in colorectal
tumor samples. Secondly, p57Kip2 is a target gene of Smad1 and
is upregulated in response to chemotherapeutic drugs. Thirdly,
knockdown of Smad1 or p57Kip2 enhances chemosensitivity in
p53-proficient or p53-deficient tumors.
In summary, analysis of Smad1 expression in .500 patient
tumor samples uncovered that Smad1 is often upregulated in colorectal cancers, which was attributable to p53 defects. Smad1 elevation appears to act as a feedback compensatory mechanism in
p532/2 cells to help curb cell proliferation, cell transformation,
and tumor formation, via increasing p57Kip2 expression and
Atm Chk2 activation. Moreover, elevated Smad1 also causes chemoresistance, which may involve p57Kip2. Thus, Smad1 might be a
target to improve chemotherapeutic efficacy. Moreover, our study
also reveals that Smad5 plays roles distinct from Smad1 in tumor
growth and chemosensitivity.
Materials and methods
Patients and tissue samples
A total of 595 patients, who received the operation at the
Department of Colorectal Surgery, Changhai Hospital, Second
Military Medical University, Shanghai, China, between December
1999 and December 2009, were collected in this study
(Supplementary Materials and methods and Table S2). Informed
consent had been obtained from all patients and the project had
been approved by the local Ethics Committee.
Mice
p532/2 and Balb/c nude mice were bred and used, following the
guidelines of mouse breeding and cage density expectations for

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Discussion
The TGFb superfamily plays critical roles in development and
tissue homeostasis. The TGFb subfamily and its downstream
Smad2/3 have long been known to have anti-proliferative activities. Smad2/3 can interact with p53 to induce the expression
of target genes such as p21Cip1 to inhibit cell proliferation
(Cordenonsi et al., 2003; Kortlever et al., 2006; Shirai et al.,
2011; Samarakoon et al., 2013; Overstreet et al., 2014). The BMP
subfamily also has tumor suppressive activities, especially in colorectal tissues (Ming Kwan et al., 2004; Pangas et al., 2008;
Neumann et al., 2011), although BMPs have been reported to
have pro-proliferative activity in certain cell types, which can
be antagonized by TGFb (Goumans et al., 2003). Our previous
studies have shown that the BMP Smad1 pathway is activated
by Atm in response to DNA damage. Smad1 then interacts with
p53 and stabilizes p53. On top of that, this study reveals that
when p53 is mutated or expressed at low levels, Smad1 is upregulated, which helps to suppress cell proliferation and oncogenesis,
by increasing p57Kip2 expression and enhancing Atm Chk2
activation (Figure 7C). Thus, Smad1 elevation may act as a compensation for p53 mutation/loss. These findings further highlight
the significance of Smad1 p53 interplay and suggest that smad1
suppresses tumorigenesis with both p53-dependent and p53independent mechanisms.
While our previous study identified Smad1 as a DDR effector
molecule that interacts with and stabilizes p53, this study suggests that Smad1 can influence Atm Chk2 activation as well,
especially in p53-deficient cells (Figure 7C). Smad1 knockdown
compromises Atm Chk2 activation, while elevated expression of
Smad1 potentiates Atm Chk2 activation. Atm activation mainly
occurs at the double-stranded DNA breaks, yet the activation
mechanism is currently unknown (Lord and Ashworth, 2012). We
found that Smad1 does not interact with Atm and is not localized
to the DNA damage-induced foci (Chau et al., 2012). It is possible
that BMP Smad1 may affect the expression of proteins that
help Atm foci recruitment and/or activation. Interestingly, recent
studies revealed that the TGFb Smad2/3 pathway is also involved
in DDR (Wang et al., 2013; Barcellos-Hoff and Cucinotta,
2014). TGFb receptor activation is required for Atm activation,
and Smad2 and Smad7 are localized on DNA damage-induced
nuclear foci (Kirshner et al., 2006; Zhang et al., 2006; Park et al.,
2015). Yet, how BMP Smad1 signaling regulates Atm activation
and whether Smad7 mediates this effect of Smad1 await further
investigation.
It has been reported that BMP Smad1 signaling is inhibited in
some human tumor types (Kodach et al., 2008; Chau et al.,
2012). However, our present study shows that Smad1 protein
level is upregulated in colorectal cancer samples. Based upon
the findings that p53 deficiency leads to elevated Smad1 expression in primary MEFs, NSCs, and osteoblasts, which is accompanied
by elevated Smad1/5/8 activation, we conceive that during tumorigenesis, p53 defects lead to a compensatory Smad1 upregulation
and momentarily activation, which helps to suppress tumor formation; late deactivation of BMP receptor due to further mutations
facilitates tumor progression. Thus, our results and others reveal

116

| Ruan et al.

animal colonies at Bio-X Institutes, Shanghai Jiao Tong University.

Mouse embryonic fibroblasts were isolated from E13.5 embryos
following a standard protocol.

synthesized with 0.5 mg of total RNA using iScript cDNA Synthesis

Kit (Fermentas). The detection and quantification of target mRNA
were performed with real-time PCR.

Isolation of primary rectal tumor cells and cell cultures

A portion of rectal cancer tissue was cut into small pieces and
enzymatically digested with 0.5% Trypsin EDTA (Gibco). The dispersed tissue was filtered through a 100-mm cell strainer (BD
Falcon), and the cells were washed with PBS and cultured. The
primary tumor cells were cultured in DMEM/F12 (1:1), whereas
MEFs, HCT116, Caco-2, and HT-29 were cultured in DMEM, which
were supplemented with 10% FBS.

Western blot analysis

Western blot analysis was carried out as previously described (Jia
et al., 2014). Anti-Smad1 (9743), Smad5 (9517), p-Smad1/5/8
(9511), p53 (2524), p-Atm (S1981) (4526S), p-Chk2 (Thr68) (2661S),
Chk2 (2662), and E2F1 (3742) antibodies were purchased from Cell
Signaling Technology. Anti-Actin (SC81178) antibody was purchased
from Santa Cruz Biotechnology, Inc. Antibodies against Atm
(GTX70103) were from Genetex. The protein bands were quantitated
using the software provided by FluorChem M system (ProteinSimple)
and the average of three repeated experiments was shown.

Immunohistochemistry and estimation of Smad1 expression

Expression of Smad1 in normal and tumor tissues from the patients
was examined by immunohistochemistry using a monoclonal antibody to Smad1 (ab108994, abcam, dilution 1:200) and DAB-based
staining technique (Dako ChemMateTM EnvisionTM Kit). The sections
were counterstained with Mayers hematoxylin (see Supplementary
Materials and methods for scoring Smad1 protein levels).
BrdU labeling assay
Cell proliferation rate was determined with the BrdU assay as
previously described (Jia et al., 2014).
Trypan blue exclusion assay
Cells treated with Dox were trypsinized from culture plates,
pooled with floating cells from medium, and centrifuged at
1000 g for 10 min at 48C. Trypan blue dissolved in buffered
phosphate-buffered saline (pH 7.2) was added to the cell cultures
to a final concentration of 0.04%. The number of live cells was
counted using hemocytometer under a light microscope. The
results were expressed as percentage of live cells.
Cell viability assay
Cell viability was measured with the water-soluble tetrazolium
salt (WST-1) assay (Roche Diagnostics), as previously described
(Jia et al., 2014). IC50 was calculated from the cell survival curves.
When IC50 could not be obtained from the survival curves, especially of control cells, additional experiments using higher concentrations of Dox were carried out to determine the IC50 values.
Real-time PCR analysis
Total RNA was extracted from cells with TRIzol reagent (Invitrogen)
following the manufacturers protocol. Complementary DNAs were

Cell transformation and tumorigenesis

Cell transformation and tumorigenesis assays were carried out
as previously described (Jia et al., 2014).
Statistical analysis
Associations between expression of Smad1 and clinicopathological variables were analyzed by non-parametric analysis,
using Mann Whitney U-test for dichotomization variables and
Kruskal Wallis test for the others. All statistical analyses were conducted using SPSS 17.0 statistical software. For other studies, statistical analysis was performed using an unpaired t-test. Significant
association was defined when P , 0.05 compared with control.
Pearsons correlation analysis was used to determine the correlation of the expression levels of Smad1 and p53 using SPSS 17.0
software. R 0.6 is deemed significant association.
Supplementary material
Supplementary material is available at Journal of Molecular Cell
Biology online.
Acknowledgements
We would like to thank Lina Gao (Shanghai Jiao Tong University)
for technical assistance.
Funding
This work was supported by grants from the National Natural
Science Foundation of China (81130039, 31300684, and 81421
061), the National Key Basic Research Program of China
(2012CB966901 and 2014CB942900), and Program of Shanghai
Subject Chief Scientist (13XD1401900).
Conflict of interest: none declared.
References
Aaltonen, L.M., Chen, R.W., Roth, S., et al. (2001). Role of TP53 P72R polymorphism in human papillomavirus associated premalignant laryngeal neoplasm.
J. Med. Genet. 38, 327.
Arnold, S.J., Maretto, S., Islam, A., et al. (2006). Dose-dependent Smad1,
Smad5 and Smad8 signaling in the early mouse embryo. Dev. Biol. 296,
104 118.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Smad1 knockdown
For transient Smad1 knockdown, siRNAs were used (Thermo
Scientific and Santa Cruz Biotechnology, Inc.). For stable knockdown of Smad1, four shRNA hairpins expressed by a retrovirus
vector (37928, 31427, 32418, and 30824, Dharmacon, Inc.) were
tested. Two of them (37928 and 30824) could efficiently knock
down Smad1 and were used further in this study. The experiments
that involve stable Smad1 knockdown were carried out using the
two shRNA hairpins with three repeats for each of them.

Smad1 compensates for p53 deficiency | 117

Liu, H., Jia, D., Li, A., et al. (2013). p53 regulates neural stem cell proliferation and
differentiation via BMP-Smad1 signaling and Id1. Stem Cells Dev. 22,
913 927.
Lord, C.J., and Ashworth, A. (2012). The DNA damage response and cancer
therapy. Nature 481, 287 294.
Lubbe, S.J., Pittman, A.M., Olver, B., et al. (2012). The 14q22.2 colorectal cancer
variant rs4444235 shows cis-acting regulation of BMP4. Oncogene 31,
3777 3784.
Lunghi, P., Costanzo, A., Mazzera, L., et al. (2009). The p53 family protein p73
provides new insights into cancer chemosensitivity and targeting. Clin.
Cancer Res. 15, 6495 6502.
Ma, G., Li, L., Hu, Y., et al. (2012). Atypical Atm-p53 genetic interaction
in osteogenesis is mediated by Smad1 signaling. J. Mol. Cell Biol. 4,
118 120.
McReynolds, L.J., Gupta, S., Figueroa, M.E., et al. (2007). Smad1 and
Smad5 differentially regulate embryonic hematopoiesis. Blood 110,
3881 3890.
Ming Kwan, K., Li, A.G., Wang, X.J., et al. (2004). Essential roles of BMPR-IA signaling in differentiation and growth of hair follicles and in skin tumorigenesis.
Genesis 39, 10 25.
Muller, P.A., and Vousden, K.H. (2013). p53 mutations in cancer. Nat. Cell Biol.
15, 2 8.
Neumann, J.C., Chandler, G.L., Damoulis, V.A., et al. (2011). Mutation in the type
IB bone morphogenetic protein receptor Alk6b impairs germ-cell differentiation and causes germ-cell tumors in zebrafish. Proc. Natl Acad. Sci. USA
108, 13153 13158.
Olivier, M., Hollstein, M., and Hainaut, P. (2010). TP53 mutations in human
cancers: origins, consequences, and clinical use. Cold Spring Harb.
Perspect. Biol. 2, a001008.
Overstreet, J.M., Samarakoon, R., Meldrum, K.K., et al. (2014). Redox control of
p53 in the transcriptional regulation of TGF-b1 target genes through SMAD
cooperativity. Cell. Signal. 26, 1427 1436.
Pangas, S.A., Li, X., Umans, L., et al. (2008). Conditional deletion of Smad1 and
Smad5 in somatic cells of male and female gonads leads to metastatic tumor
development in mice. Mol. Cell. Biol. 28, 248 257.
Park, S., Kang, J.M., Kim, S.J., et al. (2015). Smad7 enhances ATM
activity by facilitating the interaction between ATM and Mre11-Rad50-Nbs1
complex in DNA double-strand break repair. Cell. Mol. Life Sci. 72,
583 596.
Reinhardt, H.C., and Schumacher, B. (2012). The p53 network: cellular and
systemic DNA damage responses in aging and cancer. Trends Genet. 28,
128 136.
Rivlin, N., Brosh, R., Oren, M., et al. (2011). Mutations in the p53 tumor suppressor gene: important milestones at the various steps of tumorigenesis. Genes
Cancer 2, 466 474.
Samarakoon, R., Dobberfuhl, A.D., Cooley, C., et al. (2013). Induction of renal
fibrotic genes by TGF-b1 requires EGFR activation, p53 and reactive oxygen
species. Cell. Signal. 25, 2198 2209.
Sancho, E., Batlle, E., and Clevers, H. (2004). Signaling pathways in intestinal
development and cancer. Annu. Rev. Cell Dev. Biol. 20, 695 723.
Sapkota, G., Alarcon, C., Spagnoli, F.M., et al. (2007). Balancing BMP
signaling through integrated inputs into the Smad1 linker. Mol. Cell 25,
441 454.
Shirai, Y.T., Ehata, S., Yashiro, M., et al. (2011). Bone morphogenetic protein-2
and -4 play tumor suppressive roles in human diffuse-type gastric carcinoma.
Am. J. Pathol. 179, 2920 2930.
Thawani, J.P., Wang, A.C., Than, K.D., et al. (2010). Bone morphogenetic proteins
and cancer: review of the literature. Neurosurgery 66, 233 246.
Tomlinson, I.P., Carvajal-Carmona, L.G., Dobbins, S.E., et al. (2011). Multiple
common susceptibility variants near BMP pathway loci GREM1, BMP4, and
BMP2 explain part of the missing heritability of colorectal cancer. PLoS
Genet. 7, e1002105.
Vinall, R.L., Chen, J.Q., Hubbard, N.E., et al. (2012). Initiation of prostate cancer in
mice by Tp53R270H: evidence for an alternative molecular progression. Dis.
Model. Mech. 5, 914 920.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015

Barcellos-Hoff, M.H., and Cucinotta, F.A. (2014). New tricks for an old fox: impact
of TGFb on the DNA damage response and genomic stability. Sci. Signal. 7,
re5.
Brooks, C.L., and Gu, W. (2006). p53 ubiquitination: Mdm2 and beyond. Mol. Cell
21, 307 315.
Chau, J.F., Jia, D., Wang, Z., et al. (2012). A crucial role for bone morphogenetic
protein-Smad1 signalling in the DNA damage response. Nat. Commun. 3, 836.
Cordenonsi, M., Dupont, S., Maretto, S., et al. (2003). Links between tumor suppressors: p53 is required for TGF-b gene responses by cooperating with
Smads. Cell 113, 301 314.
Derynck, R., Akhurst, R.J., and Balmain, A. (2001). TGF-b signaling in tumor suppression and cancer progression. Nat. Genet. 29, 117 129.
Dick, A., Meier, A., and Hammerschmidt, M. (1999). Smad1 and Smad5 have distinct roles during dorsoventral patterning of the zebrafish embryo. Dev. Dyn.
216, 285 298.
Eivers, E., Fuentealba, L.C., and De Robertis, E.M. (2008). Integrating positional
information at the level of Smad1/5/8. Curr. Opin. Genet. Dev. 18, 304 310.
Fuentealba, L.C., Eivers, E., Ikeda, A., et al. (2007). Integrating patterning
signals: Wnt/GSK3 regulates the duration of the BMP/Smad1 signal. Cell
131, 980 993.
Gao, H., Chakraborty, G., Lee-Lim, A.P., et al. (2012). The BMP inhibitor Coco
reactivates breast cancer cells at lung metastatic sites. Cell 150, 764 779.
Goh, A.M., Coffill, C.R., and Lane, D.P. (2011). The role of mutant p53 in human
cancer. J. Pathol. 223, 116 126.
Goumans, M.J., Valdimarsdottir, G., Itoh, S., et al. (2003). Activin receptor-like
kinase (ALK)1 is an antagonistic mediator of lateral TGFb/ALK5 signaling.
Mol. Cell 12, 817 828.
Hardwick, J.C., Kodach, L.L., Offerhaus, G.J., et al. (2008). Bone morphogenetic
protein signalling in colorectal cancer. Nat. Rev. Cancer 8, 806 812.
Hong, B., Prabhu, V.V., Zhang, S., et al. (2014). Prodigiosin rescues deficient p53
signaling and antitumor effects via upregulating p73 and disrupting its interaction with mutant p53. Cancer Res. 74, 1153 1165.
Howe, J.R., Bair, J.L., Sayed, M.G., et al. (2001). Germline mutations of the gene
encoding bone morphogenetic protein receptor 1A in juvenile polyposis. Nat.
Genet. 28, 184 187.
Jackson, S.P., and Bartek, J. (2009). The DNA-damage response in human biology
and disease. Nature 461, 1071 1078.
Jia, H., Cong, Q., Chua, J.F., et al. (2014). p57Kip2 is an unrecognized DNA
damage response effector molecule that functions in tumor suppression
and chemoresistance. Oncogene, doi: 10.1038/onc.2014.287.
Kirshner, J., Jobling, M.F., Pajares, M.J., et al. (2006). Inhibition of transforming
growth factor-b1 signaling attenuates ataxia telangiectasia mutated activity
in response to genotoxic stress. Cancer Res. 66, 10861 10869.
Kodach, L.L., Wiercinska, E., de Miranda, N.F., et al. (2008). The bone morphogenetic protein pathway is inactivated in the majority of sporadic colorectal
cancers. Gastroenterology 134, 1332 1341.
Kortlever, R.M., Higgins, P.J., and Bernards, R. (2006). Plasminogen activator
inhibitor-1 is a critical downstream target of p53 in the induction of replicative
senescence. Nat. Cell Biol. 8, 877 884.
Kravchenko, J.E., Ilyinskaya, G.V., Komarov, P.G., et al. (2008). Small-molecule
RETRA suppresses mutant p53-bearing cancer cells through a p73-dependent
salvage pathway. Proc. Natl Acad. Sci. USA 105, 6302 6307.
Lai, K.P., Leong, W.F., Chau, J.F., et al. (2010). S6K1 is a multifaceted regulator of
Mdm2 that connects nutrient status and DNA damage response. EMBO J. 29,
2994 3006.
Langenfeld, E., Hong, C.C., Lanke, G., et al. (2013). Bone morphogenetic protein
type I receptor antagonists decrease growth and induce cell death of lung
cancer cell lines. PLoS One 8, e61256.
Lee, Y.C., Cheng, C.J., Bilen, M.A., et al. (2011). BMP4 promotes prostate tumor
growth in bone through osteogenesis. Cancer Res. 71, 5194 5203.
Leong, W.F., Chau, J.F., and Li, B. (2009). p53 Deficiency leads to compensatory
up-regulation of p16INK4a. Mol. Cancer Res. 7, 354 360.
Liu, B., Sun, Y., Jiang, F., et al. (2003). Disruption of Smad5 gene leads to
enhanced proliferation of high-proliferative potential precursors during embryonic hematopoiesis. Blood 101, 124 133.

118

| Ruan et al.

Vogelstein, B., Lane, D., and Levine, A.J. (2000). Surfing the p53 network. Nature
408, 307 310.
Walsh, D.W., Godson, C., Brazil, D.P., et al. (2010). Extracellular BMP-antagonist regulation in development and disease: tied up in knots. Trends Cell Biol. 20, 244256.
Wang, M., Saha, J., Hada, M., et al. (2013). Novel Smad proteins localize to
IR-induced double-strand breaks: interplay between TGFb and ATM pathways.
Nucleic Acids Res. 41, 933 942.

Zhang, S., Ekman, M., Thakur, N., et al. (2006). TGFb1-induced activation of ATM
and p53 mediates apoptosis in a Smad7-dependent manner. Cell Cycle 5,
2787 2795.
Zhang, M., Li, L., Wang, Z., et al. (2013a). A role for c-Abl in cell senescence and
spontaneous immortalization. Age (Dordr) 35, 1251 1262.
Zhang, X., Jia, D., Liu, H., et al. (2013b). Identification of 5-Iodotubercidin as a
genotoxic drug with anti-cancer potential. PLoS One 8, e62527.

Downloaded from http://jmcb.oxfordjournals.org/ by guest on June 7, 2015