Oncogene (2010) 29, 305–312

& 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00
www.nature.com/onc

SHORT COMMUNICATION

HDAC inhibitors regulate claudin-1 expression in colon cancer cells

through modulation of mRNA stability

M Krishnan1, AB Singh1, JJ Smith1,2, A Sharma1, X Chen3, S Eschrich4, TJ Yeatman4,

RD Beauchamp1,2,5 and P Dhawan1,5
1
Department of Surgery, Vanderbilt University Medical Center, Nashville, TN, USA; 2Department of Cell and Developmental
Biology, Vanderbilt University Medical Center, Nashville, TN, USA; 3Department of Biostatistics, Vanderbilt University Medical
Center, Nashville, TN, USA; 4H Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA and 5Department of Cancer
Biology, Vanderbilt University Medical Center, Nashville, TN, USA

Expression and cellular distribution of claudin-1, a tight Introduction

junction protein, is dysregulated in colon cancer and its
overexpression in colon cancer cells induced dedifferentia-
tion and increased invasion. However, the molecular
mechanism(s) underlying dysregulated claudin-1 expres-
sion in colon cancer remains poorly understood. Histone
deacetylase (HDAC)-dependent histone acetylation is an
important mechanism of the regulation of cancer-related
genes and inhibition of HDACs induces epithelial
differentiation and decreased invasion. Therefore, in this
study, we examined the role of HDAC-dependent
epigenetic regulation of claudin-1 in colon cancer. In this
study, we show that sodium butyrate and Trichostatin A
(TSA), two structurally different and widely used HDAC
inhibitors, inhibited claudin-1 expression in multiple colon
cancer cell lines. Further studies revealed modulation of
claudin-1 mRNA stability by its 30 -UTR as the major
mechanism underlying HDAC-dependent claudin-1
expression. In addition, overexpression of claudin-1
abrogated the TSA-induced inhibition of invasion in colon
cancer cells suggesting functional crosstalk. Analysis of
mRNA expression in colon cancer patients, showed a
similar pattern of increase in claudin-1 and HDAC-2
mRNA expression throughout all stages of colon cancer.
Inhibition of claudin-1 expression by HDAC-2-specific
small interfering RNA further supported the role of
HDAC-2 in this regulation. Taken together, we report a
novel post-transcriptional regulation of claudin-1 expres-
sion in colon cancer cells and further show a functional
correlation between claudin-1 expression and TSA-mediated
regulation of invasion. As HDAC inhibitors are considered
to be promising anticancer drugs, these new findings will
have implications in both laboratory and clinical settings.
Oncogene (2010) 29, 305–312; doi:10.1038/onc.2009.324;
published online 2 November 2009
Keywords: tight junction; claudin-1; invasion; TSA; HDAC
Correspondence: Dr P Dhawan, Division of Surgical Oncology,
Department of Surgery, Vanderbilt University Medical Center,
MCN, B-2211, Nashville, TN, USA.
E-mail: punita.dhawan@vanderbilt.edu
Received 8 November 2008; revised 31 August 2009; accepted 2
September 2009; published online 2 November 2009
Efficient transcription requires concerted actions of
multiple protein factors and post-translational modifi-
cations of histones including acetylation, methylation,
ubiquitination and adenosine diphosphate ribosylation.
The best understood histone modification is acetylation
of core histones and is maintained by the opposing
activities of histone acetylases and histone deacetylases
(HDACs). Histone deacetylase inhibitors (HDACIs)
have been noted for their ability to induce cell cycle
arrest, differentiation and apoptosis in numerous cancer
cell types including colon cancer (Dokmanovic and
Marks, 2005). In addition, HDACIs have also shown
signs of efficacy in phase I and phase II clinical trials
(Minucci and Pelicci, 2006). However, the molecular
mechanism(s) underlying HDACI actions remain in-
completely understood. Therefore, it is imperative that
we gain a more thorough understanding of the effect of
these agents on intracellular targets.
Tumor invasion and metastasis with loss of functional
tight junctions are key characteristics of the aggressive
cancer phenotype and metastasis is a major cause of
cancer-related death. The claudin family of proteins are
the major constituent of tight junctions, are expressed in
a tissue-specific manner and appears to be controlled
differently in cancers from distinct tissues.
We have earlier observed dysregulated claudin-1
expression in colon cancer. We have also shown a
positive correlation of claudin-1 expression with colon
tumor progression, invasion and metastasis (Dhawan
et al., 2005). Other studies have shown epigenetic
regulation of specific claudin family members (Honda
et al., 2007; Nishikiori et al., 2008) in addition to
occludin and its correlation with tumorigenicity of
cancer cells (Osanai et al., 2007b). HDACI suppress
activities of multiple HDACs that lead to increased
histone acetylation and thus modulate expression of
specific genes involved in growth arrest, differentiation
and apoptosis. Inhibition of claudin-1 expression in
colon cancer cells induces similar morphological and
metabolic changes (Dhawan et al., 2005; Shiou et al.,
2007). Therefore, we hypothesized a role for HDACs in
the regulation of claudin-1 in colon cancer.
 Epigenetic regulation of claudin-1 expression
M Krishnan et al
306
In this study, we report HDAC-dependent regulation
of claudin-1 expression in colon cancer with change in
mRNA stability as the principal mechanism underlying
this regulation. We further provide data supporting a
functional correlation between claudin-1 expression and
the anti-invasive effects of HDACI supporting claudin-1
as one of the target molecules affected on HDAC
inhibition.
Results and discussion
HDACIs decreased claudin-1 steady-state mRNA
and protein levels in colon cancer cells
To test our hypothesis, we exposed multiple claudin-1
expressing colon cancer cells (SW480, SW620 and DLD-
1) to increasing concentrations of sodium butyrate (NaB;
0, 1, 2.5, 5 and 10 mM), a known HDACI. Untreated and
DMSO-treated cells served as controls. As shown in
Figure 1a, upper panel immunoblot analysis using total
cell lysates showed a dose-dependent decrease in claudin-
1 expression in response to increasing concentrations of
NaB in SW480 and SW620 cells. In contrast, E-cadherin
expression, known to be increased on HDAC inhibition
(Wu et al., 2007) was used as positive control. We further
determined effects of increasing concentration of Tricos-
tatin-A (TSA, 0, 50, 100, 200 and 500 ng/ml), a robust
and specific inhibitor of class I and II HDACs. Similar to
NaB, TSA treatment resulted in a dose-dependent
decrease in claudin-1 expression in SW480 and SW620
cells (Figure 1a, lower panel). In the same samples,
claudin-4 expression was largely unaltered (Figure 1a)
whereas expression of claudin-2, yet another claudin
family member, increased in a dose-dependent manner
similar to E-cadherin expression thus confirming the
specificity of regulation (data not shown; communicated
elsewhere). Similar results were obtained in DLD-1 cells
(data not shown).
To confirm that the HDACIs indeed functioned as
expected, we determined expression of acetyl histone
H3. Indeed, exposure to either NaB or TSA resulted in
accumulation of acetylated histones H3 in SW480 and
SW620 cell lines (Figure 1a, lower panel). Together,
these findings suggested a role of HDACs in the
regulation of claudin-1 expression in colon cancer.
We further examined whether this HDACI-dependent
decrease in claudin-1 expression was through the
regulation of mRNA expression. Quantitative reverse
transcription–PCR (Figure 1b) and real-time quantita-
tive PCR (Figure 1c) were carried out using gene-specific
primers. As shown in Figures 1b and c, similar to the
protein levels, claudin-1 mRNA expression was markedly
decreased in both NaB- and TSA-treated samples. In
contrast, mRNA levels of E-cadherin (Figure 1b) in-
creased on treatment with HDACI whereas claudin-4
mRNA expression remains largely unchanged (Figure 1c).
A similar decrease in claudin-1 mRNA and protein
expression suggested a possible HDACI-dependent tran-
scriptional downregulation of claudin-1 expression.
HDACI-dependent decrease in claudin-1 steady-state
mRNA levels involved changes in claudin-1 transcription
and mRNA stability
A role of HDAC-dependent deacetylation in the
regulation of mRNA transcription is reported (Chavey
et al., 2008; Liu et al., 2008). Furthermore, transcription
of certain claudin family members including claudin-3, -
4 and -6 is regulated through epigenetic processes (for
example, acetylation of critical promoter region (Abuaz-
za et al., 2006; Honda et al., 2006, 2007; Osanai et al.,
2007a)). To determine a role of HDACs in the transcrip-
tional regulation of claudin-1 expression, we examined
effects of HDACI on claudin-1 promoter activity. A
human claudin-1 promoter (–1160 to þ 160)-based luci-
ferase reporter construct described earlier (Shiou et al.,
2007) and referred to as pGL3/–1.2Cld-1 was used. The
SW480 and SW620 cells were transiently transfected
with pGL3/–1.2Cld-1 or pGL3-Basic (vector alone)
constructs, along with a reference reporter plasmid.
Four hours after transfection, cells were exposed to
200 ng/ml of TSA for 36 h. Samples were collected post-
TSA treatment and luciferase reporter activity was
determined. Findings from multiple experiments showed
a modest decrease of 15–20% as compared with control
in claudin-1 promoter activity after TSA treatment
(Figure 2a). An increase (2.5-fold versus control) in E-

Figure 1 Histone deacetylase (HDAC) inhibitors decrease the expression of claudin-1. (a) Equal amounts of total protein (25 mg)
from SW480 and SW620 cell lines (treated with increasing concentration of sodium butyrate (NaB) or Trichostatin A (TSA) (Sigma-
Aldrich Inc.) for 36 h) were immunoblotted with claudin-1, claudin-4 (Invitrogen), E-cadherin (BD Biosciences) or actin antibody
(upper and middle panel). Same samples were also immunoblotted with acetyl histone H3 and histone H3 (Cell Signaling) to confirm
the specificity of inhibitors (lower panel). Blots shown are representative of three independent experiments. (b) HDAC inhibitor
decreases the claudin-1 levels at the RNA levels. The SW480 and SW620 cells were treated with HDAC inhibitors for 36 h. Total RNA
were extracted using TRIzol according to manufacturer’s instructions and subjected to semiquantitative reverse transcription–PCR
(RT–PCR). Primers for actin were used as internal control in the same reaction. Data represent from four independently replicated
experiments. (c) Total RNA was extracted from cells using Trizol (Invitrogen, Carlsbad, CA, USA) according to the protocol provided
by the manufacturer. Total RNA (300 ng) of each sample (group) was reverse transcribed using the SuperScript III reverse
transcriptase (RT) with gene-specific primers. The primers used in the experiments were sense 50 -TCACTCCCAGGAGGATGC-30 and
anti-sense 50 -GGCAGATCCAGTGCAAAGTC-30 for claudin-1; sense 50 -ATAAAGCCAGTCCTGATGCG-30 and anti-sense
50 -TAACTGCTCAACCTGTCCCC-30 for claudin-4. Reverse transcription was performed by using dNTPs with specific primers at
65 1C for 5 min, 50 1C for 20 min, followed by 10 min at 95 1C to inactivate enzymes. For real-time PCR complementary DNA was
synthesized and PCR was carried out using SYBR PCR master kit (AB Biosystems). PCR was carried out in triplicates for each gene
being validated. For each individual well, fluorescence curves were log transformed and the slope of the logarithmic portion of the
reaction was extracted to determine efficiency. Ct values and efficiency were then used according to the methods of Schefe et al. (2006)
to calculate the fold change.

Oncogene

cadherin promoter activity, used as positive control,
under similar conditions supported the efficacy of our
luciferase reporter assay. However, at present (1)
Epigenetic regulation of claudin-1 expression
M Krishnan et al
307
presence of a HDAC-response element beyond the tested
region of claudin-1 promoter cannot be ruled out and (2)
transient transfections might not be enough to rule out

SW480 SW620
Claudin-1 Claudin-1

E-cadherin E-cadherin
Actin Actin
U DMSO 1 2.5 5 10 NaB (mM) U DMSO 1 2.5 5 10 NaB (mM)

Claudin-1 Claudin-1
E-cadherin E-cadherin
Claudin-4 Claudin-4
Actin Actin
Acetyl H3
Acetyl H3
H3 H3
U M 50 100 200 500 TSA(ng/ml) U M 50 100 200 500 TSA(ng/ml)

SW480 SW620

Claudin-1 Claudin-1

E-cadherin E-cadherin

Actin Actin

U DMSO 1 2.5 5 10 NaB (mM) U DMSO 1 2.5 5 10 NaB (mM)

Claudin-1 Claudin-1

E-cadherin E-cadherin

Actin Actin
U M 50 100 200 500 TSA(ng/ml) U M 50 100 200 500 TSA(ng/ml)

SW480 SW620
5 Claudin-1 5 Claudin-1
FOLDCHANGE

0
FOLDCHANGE

-5
-5
-10
-10
-15
U

50

0
U

50

10

20

50
10

20

50

-15
TSA treatment (200ng/ml) TSA treatment (200ng/ml)

5 Claudin-4 5 Claudin-4
FOLDCHANGE

0
FOLDCHANGE

-5 -5

-10
-10
U

50

0
10

20

50

50

-15
10

20

50

TSA treatment (200ng/ml) -15

TSA treatment (200ng/ml)

Oncogene
 Epigenetic regulation of claudin-1 expression
M Krishnan et al
308
18
1.2Claudin-1 1.2Claudin-1 E-cadherin Promoter
1.2 1.2 16
1.2Claudin-1/TSA 1.2Claudin-1/TSA
1 1 14

Fold change
12

Fold change
Fold change
0.8 0.8
10
0.6 0.6 8
0.4 6
0.4
4
0.2 0.2 2
0 0 0
SW480 SW620 TSA U 100ng/ml 200ng/ml

SW480 SW620
1.1 1.1
1 1
0.9 TSA 0.9 TSA
% mRNA remaining

%mRNA remaining
0.8 Act.D 0.8 Act.D
0.7 Act.D +TSA 0.7 Act.D +TSA
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0hr 4hrs 8hrs 16hrs 24hrs 36hrs 0hr 4hrs 8hrs 16hrs 24hrs 36hrs

Figure 2 Effect of Trichostatin A (TSA) treatment on claudin-1 promoter activity and mRNA stability. (a) SW480, SW620 on
six-well plates was transiently transfected with pGL3/–1.2Cld-1, pGL3/–1.5 E-cadherin or pGL3-Basic, along with reference reporter
(phRL-TK). Four hours later cells were treated with TSA (200 ng/ml) for 36 h and control (pGL3-Basic) was treated with methanol.
The cells were lysed, and luciferase activity in lysates was measured. Each bar represents the mean±s.d. of three experiments. (b and c)
Claudin-1 mRNA half-life was determined using actinomycin D treatment. SW480 (b) and SW620 (c) cells at 70–80% confluence on
six-well plates were treated with either TSA alone or actinomycin D alone for different time points, or with TSA (200 ng/ml) for 4 h
followed by treatment with 10 mg/ml of actinomycin D, transcription inhibitor. The samples for RNA analysis were then collected at
0, 4, 8, 16, 24 and 36 h after actinomycin D treatment. Total RNA was extracted using the Trizol reagent and real time qPCR was
performed as described in Figure 1c. Data are representative of three separate experiments.

the transcriptional regulation because they could not be half-life after combined treatment of TSA and actino-
well suited for the identification of regulatory elements mycin D was B7.5 h. Similar findings were obtained
mediating the TSA effects as the chromatin configura- from the use of SW620 cells wherein the half-life of
tion might not be physiological. Furthermore, histone claudin-1 mRNA was B20 h after actinomycin D
acetylation is critical component of chromatin remodel- treatment whereas combined exposure to TSA and
ing and transcriptional regulation. Studies are currently actinomycin D decreased it to B9 h (Figure 2c). Taken
underway to define the mechanism of potential tran- together, these findings suggested change in mRNA
scriptional regulation of claudin-1 by HDAC inhibitors. stability as the principal mechanism underlying
HDACI-dependent decreases in claudin-1 expression
in colon cancer cells.
HDACI decreases claudin-1 expression through
modulation of mRNA stability
In addition to being transcriptional activators, HDACIs The 30 -UTR of claudin-1 is important
are also involved in the post-transcriptional regulation for its mRNA stability
(Yamamichi et al., 2005; Pryzbylkowski et al., 2008) and An important role of 30 -UTR in the regulation of
acetylation of cytoplasmic proteins. Therefore, we mRNA stability is reported. This regulation primarily
further examined effects of HDACI on possible changes involves interaction of cis-elements in the 30 -UTR of
in claudin-1 mRNA stability. a gene with specific trans-acting factors. In addition,
To determine changes in claudin-1 mRNA stability, the presence of a long 30 -UTR is frequently indicative
we blocked de novo mRNA transcription using actino- of post-translational regulation of gene expression
mycin D (10 mg/ml), an inhibitor of mRNA transcrip- through modulation of mRNA stability (Pesole et al.,
tion. SW480 or SW620 cells were exposed to either 2001). It should be noted that claudin-1 possesses a
actinomycin D (10 mg/ml) or TSA or actinomycin long (B2.3 kb) 30 -UTR. Therefore, we further deter-
D þ TSA in which actinomycin D was added 4 h after mined the importance of the claudin-1 30 -UTR in TSA-
TSA treatment. Samples were collected at 0, 4, 8, 16, 24 dependent mRNA degradation. We used SW480claudin1
and 36 h after actinomycin D treatment. The mRNA cells for this purpose. SW480claudin1 cells stably over-
expression levels were determined using gene-specific express the human claudin-1 coding sequence without the
primers and real-time quantitative PCR. 30 -UTR. Simultaneous use of SW480 cells that express
As shown in Figure 2b, results from the cells exposed similar levels of endogenous claudin-1 as SW480claudin1
to actinomycin D alone showed half-life of claudin-1 cells served as control. It is noteworthy that contrary to
mRNA in SW480 cells to be B18 h, whereas the the control cells, no appreciable decrease in claudin-1
Oncogene

mRNA expression was observed in SW480claudin1 cells in
response to TSA treatment (Figure 3a). In addition, the
rate of mRNA decay in SW480claudin1 cells (t1/2 ¼ 19 h)
was relatively slower compared with the parental SW480
cells (t1/2 ¼ 8 h) on combined treatment of TSA and
actinomycin D.
To further confirm the role of 30 -UTR of claudin-1
mRNA, we used luciferase reporter constructs in which
claudin-1 30 -UTR regions (2650–3370, fragment 1 or
3121–3370, fragment 2) were inserted into the 30 -UTR of
luciferase gene. Transfection using different concentra-
tions (0.5 and 1.0 mg of above constructs) was carried
out in SW480 and SW620 cells wherein pGL3-vector
construct served as a control. As shown in Figure 3b, a
significant reduction in a dose-dependent manner in the
promoter activity was observed as compared with the
parental pGL3 control. Together, these results showed
that the 30 -UTR of claudin-1 mRNA inhibits the
expression of a heterologous reporter gene in colon
cancer. The above studies provide strong support for a
putative role of 30 -UTR in the regulation of claudin-1
mRNA stability, however, more definitive studies
including identification and characterization of crucial
claudin-1 stabilization sequences/proteins in the 30 -UTR
are required and are currently under way.
Overexpression of claudin-1 attenuated Trichostatin
A- induced decreases in invasion
We further determined functional significance of HDACI-
dependent decreases in claudin-1 expression. In colon
cancer cells, claudin-1 expression is positively associated
with invasion and metastasis (Dhawan et al., 2005). By
contrast, exposure to HDACI including TSA or NaB
decreases invasion (Wu et al., 2007). We first confirmed
that this is also true in our cell model by treating SW620
cells with TSA (Figure 3c). Indeed as shown in Figure 3c,
TSA treatment resulted in a dose-dependent decrease in
invasion. Therefore, we next asked whether overexpression
of claudin-1 expression would effectively abrogate or signif-
icantly reduce the effect of TSA on cell invasion.
As SW620 cells express high levels of claudin-1, we
overexpressed Claudin-1 in SW480 cells with low levels
of endogenous claudin-1. SW480 and two different
clones of SW480claudin1 cells were subjected to identical
treatments of TSA, 200 ng/ml (36 h) as before and
effects on cell invasion were observed. As expected,
TSA treatment resulted in significant decreases in cell
invasion in control cells. However, TSA-treatment failed
to induce a similar level of decreased invasion in
SW480claudin1 cells (Figure 3d). These findings support
a functional role of claudin-1 expression in TSA-induced
cell invasion and suggest a pivotal role of claudin-1
protein in the anticancer effects induced by HDAC
inhibitors.
Correlation between HDAC-2 and claudin-1 expression
in colon cancer
Similar to claudin-1, HDAC-1, -2 and -3 are upregu-
lated in colon cancer. Therefore, we used two clinical
and genomic databases (representing patients from the
Epigenetic regulation of claudin-1 expression
M Krishnan et al
309
Moffitt Cancer Center in Tampa, FL, USA and the
Vanderbilt Medical Center, Nashville, TN, USA) to
examine a potential correlation in the expression of
HDACs and claudin-1 in colorectal cancer. On pre-
liminary analysis, comparing colon cancer tissues versus
adenoma tissues in the Vanderbilt Medical Center data
set, we noted increased expression of claudin-1 and
HDAC-2 in colon cancer patients while we could not
observe a similar trend of claudin-1 expression with
HDAC-1 or HDAC-3 expression (data not shown).
Among HDACs, the specific expression of HDAC2 is
higher in human colon cancer samples and cell lines
(Zhu et al., 2004; Godman et al., 2008). In addition,
HDAC-2 levels are upregulated in ApcMin/ þ mice
(Zhu et al., 2004; Godman et al., 2008). It should be
noted that claudin-1 expression is markedly increased
and mislocalized in adenomas from ApcMin/ þ mice
compared with normal controls (Dhawan et al., 2005).
Indeed, when we analysed the microarray data from the
larger Moffitt Cancer Center data set that represents a
total of 205 patient samples (n ¼ 10 normal adjacent
colonic tissues and n ¼ 195 colorectal cancer tissues
(stages I–IV)), a significant increase in claudin-1
expression across all stages compared with normal
adjacent colonic tissue was observed (Po0.001,
Figure 4a). In addition, HDAC-2 expression was
significantly upregulated across all stages of colorectal
cancer compared with normal colonic tissue (Po0.001,
Figure 4a). Together, our data support the potential
coordinate regulation of claudin-1 and HDAC-2 ex-
pression in colorectal cancer progression.
To further test the direct dependence of claudin-1
expression on HDAC-2, we silenced HDAC-2 expres-
sion in cell models of our study. Both SW480 and
SW620 cell lines were transfected using commercially
available human HDAC-2-specific small interfering
RNAs (siRNAs) or control siRNA and effect on
claudin-1 expression was determined. As shown in
Figure 4b, immunoblot analysis confirmed effective
silencing of HDAC-2 expression in both cell lines after
transient transfections with respective siRNAs while
there was no change in HDAC-3 expression. In
addition, in both cell lines, inhibition of HDAC-2
expression inhibited claudin-1 expression. As expected,
claudin-4 expression remained unaltered in the same cell
lysates. To further confirm the specificity of this HDAC-
2-mediated claudin-1 inhibition, we inhibited HDAC-6
in the same cells. As shown in Figure 4c, inhibition of
HDAC-6 using two different siRNAs had no effect on
claudin-1 expression. Taken together, these findings
further supported our initial hypothesis regarding a role
of HDACs in the regulation of claudin-1 expression in
colon cancer cells and confirmed the central role of
HDAC-2 in particular in this regulation.
Overall, in this report, we show a novel role of HDAC
inhibitors in the regulation of claudin-1 expression in
colon cancer cells through the regulation of mRNA
stability by its 30 -UTR. In addition, our findings
support a functional cross-talk between HDAC and
claudin-1 expression in the regulation of cell invasion.
Further investigations are needed to clearly understand
Oncogene
 Epigenetic regulation of claudin-1 expression
M Krishnan et al
310
the details of HDACI effects on claudin-1 regulation Materials and methods
and are currently underway. Findings from our Plasmids and reagents
studies have the potential to provide the necessary The antibodies against claudin-1 and -4, were from Invitrogen
insight into the mechanism of anticancer effects as well (San Francisco, CA, USA), and E-cadherin was from BD
as to improve the therapeutic efficacy of HDAC Biosciences (San Jose, CA, USA). Anti-HDAC -2, -3 anti-
inhibitors. bodies and NaB were purchased from Upstate (NY, USA).

SW480 SW480Claudin-1
120
120
Act.D
100
TSA 100
% mRNA remaining

% mRNA remaining
TSA+Act.D
80 80

60 Act.D
60
TSA
40 40 TSA+Act.D

20 20

0 0
0hr 4hrs 8hrs 16hrs 24hrs 36hrs 0hr 4hrs 8hrs 16hrs 24hrs 36hrs
Treatment Treatment

120 PGL3 120 PGL3

Fragment 1 Fragment 1
Fragment 2 Fragment 2
100 PGL3 100 PGL3
* Fragment 1
Fragment 1
* * * Fragment 2
80 Fragment 2 80 *
% Change
% Change

* *
*
60 60

40 40

20 20

0 0
SW480 SW620

SW620
120

100
% of cells invading

80 *

60 *

40
*
20

0
U M 50 100 200 500ng/ml TSA

140 SW480
SW480Claudin-1(1)
120 SW480Claudin-1(2)

100
% of cells invading

80

60 *

*
40
*
20

Oncogene
 Epigenetic regulation of claudin-1 expression
M Krishnan et al
311

11

Claudin 1 Normalized expression

11.0

HDAC 2 Normalized expression

10

10.5
9

8
10.0

9.5
6
Normal Stage1 Stage 2 stage 3 Stage 4 Normal Stage1 Stage 2 stage 3 Stage 4

SW480 SW620
SW480
HDAC-2 HDAC-2
HDAC-6

HDAC-3 HDAC-3
Claudin-1

Claudin-1 Claudin-1
Actin
Claudin-4
Claudin-4
HDAC-2
Actin
Actin

ControlsiRNA
Untransfected

HDAC6siRNA

HDAC6siRNA
ControlsiRNA
Untransfected

HDAC2siRNA
ControlsiRNA
Untransfected

HDAC2siRNA

Figure 4 Correlation of expression of HDAC-2 and claudin-1. (a) Human colorectal tissue collection and processing: The protocols and
procedures used have been approved by the respective Institutional Review Boards (Birmingham, Nashville, TN, USA), and informed
consent was obtained for each patient. Fresh tissue specimens from all surgically resected colorectal cancers and endoscopically biopsied rectal
cancers were submitted to Surgical Pathology on removal and a representative section of tumor was immediately flash frozen in liquid
nitrogen, transported to the laboratory and stored at 80 1C until used for RNA isolation. RNA was purified using the RNeasy kit from
Qiagen (Valencia, CA, USA) according to manufacturer protocol. RNA was eluted with 10 m MTris/DEPC H2O at pH 8.0, and samples
were submitted to the Vanderbilt Microarray Shared Resource. The raw expression data (cel files) were obtained and converted to expression
values using the Affy function in R (www.bioconductor.org). The normalized expression data were analysed for adenomas, normal adjacent
colon tissue and American Joint Commission on Cancer stages 1–4 adenocarcinoma in the Vanderbilt and Moffitt data sets. To compare the
expression level of each probe between the adenoma, normal tissue and different stages of tumors, the Mann–Whitney tests were conducted
and exact P-values were reported because of small sample size and data from the Moffitt patients is shown. (b) Effect of HDAC-2 silencing in
colon cancer cells. The SW480 and SW620 cells were seeded into six-well plates and grown overnight followed by transfection with HDAC-2
along with control small interfering RNA (siRNA) (Dharmacon Biotechnology) under serum free condition. After 24 h of transfection, 3 
serum was added and kept for another 36 h. The cells were lysed and used for immunoblot analysis. Protein levels of HDAC-2, HDAC-3,
claudin-1 and claudin-4 were determined by western blot. Blots shown are representative of three independent experiments. (c) Effect of
HDAC-6 silencing in SW480 colon cancer cells. The SW480 cells were seeded into six-well plates and grown overnight followed by
transfection with HDAC-6 siRNA along with control siRNA (Dharmacon Biotechnology ) under serum free condition. After 24 h of
transfection, 3  serum was added and kept for another 36 h. The cells were lysed and used for immunoblot analysis. Protein levels of
HDAC-6, HDAC-2, claudin-1 and actin were determined by western blot. Blots shown are representative of three independent experiments.

Figure 3 The 30 -UTR of claudin-1 has a role in claudin-1 mRNA stability. (a) Claudin-1 mRNA half-life was determined using actinomycin
D treatment. SW480 and SW480claudin1 cells at 70–80% confluence on six-well plates were treated with either Trichostatin A (TSA) alone or
actinomycin D alone for different time points, or with TSA (200 ng/ml) for 4 h followed by treatment with 10 mg/ml of actinomycin D,
transcription inhibitor. The samples for RNA analysis were then collected at 0, 4, 8, 16, 24 and 36 h after actinomycin D treatment. Total
RNA was extracted using the Trizol reagent. The densitometric analysis was carried out and actin was used for normalization. Data are
representative of three separate experiments. (b) SW480 and SW620 parental cells were transiently cotransfected with pGL3, pGL3/-fragment
1 (2650–3370) and fragment 2 (3121–3370), constructs as indicated concentration along with a reference reporter construct (phRL–TK).
Forty-eight hours later, cells were lysed, and luciferase activity in lysates was measured. Mean of three independent experiments with s.d.
and shown in percentage change (pGL3 taken as 100%). (c) Effect of claudin-1 on TSA-mediated invasion. SW620 cells were plated in
collagen coated (100 mg/ml) transwells (100 000 cells per transwell) and left for overnight. TSA was added in increasing concentration for 48 h
after which invasive cell numbers were scored. Each invasion assay was performed in triplicate and each bar represents the mean±s.d. of
three experiments. (d) SW480 cells and SW480claudin1 cells were plated in collagen coated (100 mg/ml) transwells (100 000 cells per transwell)
and left for overnight. TSA was added to the lower chamber in increasing concentration for 48 h after which invasive cell numbers were
scored. Each invasion assay was performed in triplicate and each bar represents the mean±s.d. of three experiments.

Oncogene
 Epigenetic regulation of claudin-1 expression
M Krishnan et al
312
Anti-acetyl H3 antibody was from Cell Signaling (Danvers, MA,
USA). The anti-actin antibody, Trichostastin A and actinomycin
D were from Sigma-Aldrich, Inc., (St Louis, MO, USA). HDAC-
2 specific as well as respective control siRNA were obtained from
Dharmacon Biotechnology (Chicago, IL, USA) and HDAC-6
siRNA were obtained from Sigma-Aldrich. Human colorectal
tissue collection and processing was carried out as approved by
the respective Institutional Review Boards (University of
Alabama-Birmingham, Vanderbilt Medical Center and Veterans
Affairs Medical Center, Nashville, TN, USA and the Moffitt
Cancer Center). Informed consent was obtained for each patient.
RNA isolation, semiquantitative reverse transcription–PCR
and real-time PCR
Total RNA isolation, reverse transcription–PCR was performed
by standard protocols as described earlier (Zhang et al., 2000;
Dhawan et al., 2005). For real-time PCR complementary DNA
was synthesized and PCR was carried out using SYBR PCR
master kit (Applied Biosystems Inc., Foster City, CA, USA).
PCR was carried out in triplicates for each gene being validated.
Analysis of mRNA stability
Actinomycin D (10 mg/ml, Sigma-Aldrich) was used to inhibit
nascent RNA synthesis. SW620 or SW480 cells (70–80%
confluence) were pretreated with TSA 200 ng/ml for 4 h and
then actinomycin D (10 mg/ml) for 0, 2, 4, 8, 16, 24 and 36 h
and collected for RNA isolation.
References
Abuazza G, Becker A, Williams SS, Chakravarty S, Truong HT, Lin F
et al. (2006). Claudins 6, 9, and 13 are developmentally expressed
renal tight junction proteins. Am J Physiol Renal Physiol 291:
F1132–F1141.
Chavey C, Muhlbauer M, Bossard C, Freund A, Durand S, Jorgensen
C et al. (2008). Interleukin-8 expression is regulated by histone
deacetylases through the NF-{kappa}B pathway in breast cancer.
Mol Pharmacol 74: 1359–1366.
Dhawan P, Singh AB, Deane NG, No Y, Shiou S-R, Schmidt C et al.
(2005). Claudin-1 regulates cellular transformation and metastatic
behavior in colon cancer. J Clin Invest 115: 1765–1776.
Dokmanovic M, Marks PA. (2005). Prospects: histone deacetylase
inhibitors. J Cell Biochem 96: 293–304.
Godman CA, Joshi R, Tierney BR, Greenspan E, Rasmussen TP,
Wang HW et al. (2008). HDAC3 impacts multiple oncogenic
pathways in colon cancer cells with effects on Wnt and vitamin D
signaling. Cancer Biol Ther 7: 1570–1580.
Honda H, Pazin MJ, D’Souza T, Ji H, Morin PJ. (2007). Regulation
of the CLDN3 gene in ovarian cancer cells. Cancer Biol Ther 6:
1733–1742.
Honda H, Pazin MJ, Ji H, Wernyj RP, Morin PJ. (2006). Crucial
roles of Sp1 and epigenetic modifications in the regulation of the
CLDN4 promoter in ovarian cancer cells. J Biol Chem 281: 21433–
21444.
Liu P-Y, Chan JY-H, Lin H-C, Wang S-L, Liu S-T, Ho C-L et al. (2008).
Modulation of the cyclin-dependent kinase inhibitor p21WAF1/Cip1
gene by Zac1 through the antagonistic regulators p53 and histone
deacetylase 1 in HeLa cells. Mol Cancer Res 6: 1204–1214.
Minucci S, Pelicci PG. (2006). Histone deacetylase inhibitors and the
promise of epigenetic (and more) treatments for cancer. Nat Rev
Cancer 6: 38–51.
Nishikiori N, Sawada N, Ohguro H. (2008). Prevention of murine
experimental corneal trauma by epigenetic events regulating claudin
6 and claudin 9. Jpn J Ophthalmol 52: 195–203.
Osanai M, Murata M, Chiba H, Kojima T, Sawada N. (2007a).
Epigenetic silencing of claudin-6 promotes anchorage-
Oncogene
Invasion assay
Assays were performed as we have described earlier (Shiou
et al., 2007).
Statistical analysis
To compare the expression level of each probe between the
adenoma, normal tissue and different stages of tumors, the Mann–
Whitney tests were conducted and exact P-values were reported.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements
This work was supported by NIH grant CA119005, CA124977
(P Dhawan), and AHA Grant 0435471N, 5P50DK044757 and
P30DK058406 Pilot project (AB Singh), the Society of Uni-
versity Surgeons-Ethicon Scholarship Award (JJS) CA112215
(TJ Yeatman), DK052334, CA 069457, the GI Cancer SPORE
grant CA95103 (RD Beauchamp), the Vanderbilt-Ingram Cancer
Center P30CA68485 and the NIH grant supporting the Digestive
Diseases Center DK58404. We thank Christian Kis for the help
in real-time quantitative reverse transcription–PCR analysis.
independent growth of breast carcinoma cells. Cancer Sci 98:
1557–1562.
Osanai M, Murata M, Nishikiori N, Chiba H, Kojima T, Sawada N.
(2007b). Occludin-mediated premature senescence is a fail-safe
mechanism against tumorigenesis in breast carcinoma cells. Cancer
Sci 98: 1027–1034.
Pesole G, Mignone F, Gissi C, Grillo G, Licciulli F, Liuni S. (2001).
Structural and functional features of eukaryotic mRNA untrans-
lated regions. Gene 276: 73–81.
Pryzbylkowski P, Obajimi O, Keen J. (2008). Trichostatin A
and 5 Aza-20 deoxycytidine decrease estrogen receptor mRNA
stability in ER positive MCF7 cells through modulation of HuR.
Breast Cancer Res Treat 111: 15–25.
Schefe JH, Lehmann KE, Buschmann IR, Unger T, Funke-Kaiser H.
(2006). Quantitative real-time RT-PCR data analysis: current
concepts and the novel ‘gene expression’s CT difference’ formula.
J Mol Med 84: 901–910.
Shiou SR, Singh AB, Moorthy K, Datta PK, Washington MK,
Beauchamp RD et al. (2007). Smad4 regulates claudin-1
expression in a transforming growth factor-beta-independent
manner in colon cancer cells. Cancer Res 67: 1571–1579.
Wu Y, Starzinski-Powitz A, Guo S-W. (2007). Trichostatin A, a
histone deacetylase inhibitor, attenuates invasiveness and reactivates
E-cadherin expression in immortalized endometriotic cells. Reprod
Sci 14: 374–382.
Yamamichi N, Yamamichi-Nishina M, Mizutani T, Watanabe H,
Minoguchi S, Kobayashi N et al. (2005). The Brm gene suppressed
at the post-transcriptional level in various human cell lines is
inducible by transient HDAC inhibitor treatment, which exhibits
antioncogenic potential. Oncogene 24: 5471–5481.
Zhang Z, Sheng H, Shao J, Beauchamp RD, DuBois RN. (2000).
Posttranscriptional regulation of cyclooxygenase-2 in rat intestinal
epithelial cells. Neoplasia 2: 523–530.
Zhu P, Martin E, Mengwasser J, Schlag P, Janssen K-P,
Göttlicher M. (2004). Induction of HDAC2 expression upon loss
of APC in colorectal tumorigenesis. Cancer Cell 5: 455–463.