Báo cáo khoa học: PKA independent and cell type specific activation of the expression of caudal homeobox gene Cdx-2 by cyclic AM pptx

PKA independent and cell type specific activation of the
expression of caudal homeobox gene Cdx-2 by cyclic AMP
´
Liang Chen1,2, Peixiang Wang1,2, Cristiano F. Andrade1,2, Ian Y. Zhao1,2, Philip E. Dube3,
Patricia L. Brubaker3,4, Mingyao Liu1,2,3 and Tianru Jin1,2,4,5
1
2
3
4
5

Division of Cell and Molecular Biology, Toronto General Research Institute, University Health Network
Institute of Medical Science, University of Toronto, Canada
Department of Physiology, University of Toronto, Canada
Department of Medicine, University of Toronto, Canada
Department of Laboratory Medicine and Pathobiology, University of Toronto, Canada

Keywords
Cdx-2; cAMP; Epac; ERK1 ⁄ 2; proglucagon
Correspondence
T. Jin, Division of Cell and Molecular
Biology, Toronto General Research Institute,
University Health Network. 67 College St.,
Toronto, Ontario, M5G 2M1
Fax: +1 416 340 3453
Tel: +1 416 340 4800, ext. 4768
E-mail:
(Received 4 March 2005, accepted
31 March 2005)
doi:10.1111/j.1742-4658.2005.04694.x

Cdx-2 is a transactivator for the proglucagon gene in pancreatic and intestinal endocrine cells. Cdx-2 is also expressed in differentiated intestinal epithelia of nonendocrine origin. Cdx-2– ⁄ – mice are embryonic lethal, while
Cdx-2+ ⁄ – mutants show multiple malfunctions including the formation of
intestinal polyps. Within the polyps, the remaining wild type Cdx-2 allele
ceases its expression, while the expression of both Cdx-2 and proglucagon
in the endocrine cells remains unaltered, indicating that Cdx-2 could be
haplo-insufficient for nonendocrine cells, but not for proglucagon producing endocrine cells. We propose that mechanisms underlying Cdx-2
expression and auto-regulation [Xu F, Li H & Jin T (1999), J Biol Chem
274, 34310–34316] differ in these two types of cells. We show here that
forskolin and cAMP upregulate Cdx-2 expression in proglucagon producing cells, but not in colon cancer cells and primary intestinal cell cultures.
It is unlikely that the activation is mainly mediated by PKA, because the
activation was observed in a PKA deficient cell line. Cotransfecting a
dominant negative Ras expression plasmid substantially repressed the
Cdx-2 promoter, in contrast to a previous finding that Ras is a negative
factor for Cdx-2 expression in colon cancer cells. Furthermore, forskolin
activated ERK1 ⁄ 2 phosphorylation in the endocrine cells, and attenuation
of ERK1 ⁄ 2 phosphorylation by its inhibitor is associated with attenuated
Cdx-2 expression. Finally, an Epac pathway specific cAMP analogue
stimulated both ERK1 ⁄ 2 phosphorylation and Cdx-2 expression. Taken
together, our observations suggest that Cdx-2 expression is regulated by
the second messenger cAMP, cell-type specifically, via the Epac pathway.

Homeodomain (HD) proteins, encoded by homeobox
genes, are tissue or cell-type specific transcription factors. They are involved in embryogenesis; cell growth,
differentiation and apoptosis; hormone synthesis; and
many other biological and physiological cellular
events. In the adult, the caudal HD protein Cdx-2 is

expressed in differentiated intestinal epithelia, including the proglucagon producing endocrine L cells [1–5].
It is also expressed in the pancreatic islets, pancreatic
insulin producing endocrine B cell lines, and proglucagon producing endocrine A cell lines [4–7]. Using

‘knock out’ approaches, two research laboratories have

Abbreviations
CRE, cAMP response element; Epac, Exchange protein directly activated by cyclic AMP; FRIC, fetal rat intestinal cell; HD, homeodomain;
IBMX, 3-isobutyl-1-methylxanthine; LUC, luciferase; MAPK, mitogen activated protein kinase; PKA, protein Kinase A; PKAc, catalytic subunit
of PKA.

2746

FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

independently demonstrated that Cdx-2– ⁄ – mice die
during 3.5–5.5 day post coitus (dpc) [8,9]. Interestingly,
deleting one wild type Cdx-2 allele leads to multiple
malfunctions, including the development of polyp-like
lesions in the proximal colon [8,9]. Consistently,
extensive in vitro and ex vivo studies by a number
of laboratories have identified more than two-dozen
potential downstream target genes of Cdx-2 [3–7,
10–25], including the genes that encode proglucagon
and insulin in pancreatic islets and intestinal endocrine
cells [4–6].
Within the intestinal polyps in the Cdx-2+ ⁄ – mice,
the remaining wild type Cdx-2 allele ceases its expression, while its expression in the surrounding normal
intestinal epithelia continues [8,9]. However, the
expression of both Cdx-2 and proglucagon in the
endocrine cells of both pancreatic and intestinal origin

appears to be unaltered in the Cdx-2+ ⁄ – mice. We
hypothesized that Cdx-2 expression could be haploinsufficient in selected types of cells [20].
To understand molecular mechanisms underlying this
intriguing cell-type specific event, we have isolated
the mouse Cdx-2 gene promoter, and initiated an examination of transcription factors and signaling molecules that regulate the expression of this promoter. We
found that cotransfection of the Cdx-2 cDNA led to
upregulation of the expression of the Cdx-2 gene promoter in cell lines that express endogenous Cdx-2 [20]. The
activation, however, was not observed when the naive
fibroblast cell lines were utilized [20]. We also demonstrated that the POU HD protein Oct-1 is able to bind
to the Cdx-2 gene promoter and is implicated in regulating Cdx-2 promoter expression and auto-expression
[21]. Such an auto-regulatory mechanism would provide
an explanation as to why one functional Cdx-2 allele is
sufficient for maintaining its own expression and for
regulating proglucagon gene expression in the pancreatic A and intestinal L endocrine cells. It, however,
raises an even more intriguing question: why could one
functional Cdx-2 allele be insufficient in maintaining its
own expression in the intestinal nonendocrine cells, and
in preventing the formation of intestinal polyps for the
nonendocrine intestinal epithelia?
We propose that molecular mechanisms underlying
Cdx-2 expression, including its auto-expression, differ
in the endocrine cells from those in the nonendocrine
intestinal epithelia. In this study, we demonstrate
that forskolin ⁄ 3-isobutyl-1-methylxanthine (IBMX)
and the second messenger cAMP upregulate Cdx-2
promoter and endogenous Cdx-2 gene expression,
specifically in the proglucagon producing endocrine
cells. It is unlikely that the activation is mainly
mediated by the protein kinase A (PKA) signaling
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

Activation of Cdx-2 expression by cAMP

pathway. In contrast, forskolin ⁄ IBMX does not activate Cdx-2 promoter and ⁄ or endogenous Cdx-2 gene
expression in nonendocrine colon cancer cell lines
and in primary fetal rat nonendocrine intestinal cell
cultures. Our results also suggest that the exchange
protein directly activated by cAMP (Epac) signaling
pathway is involved in this cell type specific activation event.

Results
Forskolin/IBMX upregulates Cdx-2 gene promoter
in proglucagon producing endocrine cell lines
We started our investigation by seeking for chemicals
or signaling molecules that upregulate Cdx-2 promoter
expression in the proglucagon producing cell lines
only. The )769 Cdx-2 ⁄ LUC fusion gene plasmid
[20,21] was transfected into the proglucagon producing
GLUTag cell line. After the transfection, RA (1 lm),
TPA (1 lm), or forskolin ⁄ IBMX (F ⁄ I, 10 lm each)
was added to the medium. The cells continued to grow
for 20 h before harvested for LUC reporter gene
analysis. As can be seen in Fig. 1A, RA and TPA had
virtually no effect on Cdx-2 promoter expression
(panel i), while forskolin ⁄ IBMX treatment caused
approximately 2.5-fold activation (panel ii). Similar
results were obtained for two other endocrine cell lines,
InR1-G9 and STC-1 (see below, and data not shown).
On the other hand, when the same )769 Cdx-2 ⁄ LUC
fusion reporter construct was transfected into the three

nonendocrine colon cancer cell lines, we did not
observe any appreciable activation by forskolin ⁄ IBMX
treatment. A representative result on the HT-29 cell
line is shown in Fig. 1A, panel iii. Thus, it seems that
forskolin ⁄ IBMX has different effects on Cdx-2 promoter expression between endocrine cells and nonendocrine cells. We then further examined the effect of
forskolin ⁄ IBMX treatment for different time lengths
on Cdx-2 promoter expression in the GLUTag cell
line. It was found that the activation appeared at 4 h
and gradually increased during the 20 h experimental
period (Fig. 1A, panel iv). Panel v shows that forskolin ⁄ IBMX generated no substantial effect on the
expression of pBLUC, the promoter-less plasmid utilized in the construction of both Cdx-2 ⁄ LUC and
GLU ⁄ LUC [4,20].
Forskolin/ IBMX activates endogenous Cdx-2
mRNA expression in endocrine cell lines
We then asked the question whether forskolin ⁄ IBMX
treatment would stimulate endogenous Cdx-2 mRNA
2747


Activation of Cdx-2 expression by cAMP

A

ii) GLUTag

i) GLUTag
2
Relative LUC Activity

L. Chen et al.

3

iii) HT-29

iv) GLUTag

1.2

**

2

3

0.8

*

v) GLUTag
2

**

2

**

1

1
1

0
Reporter

0

Treatment V TPA RA

0
-769 Cdx-LUC
F/I
V

V

0
F/I

0
BLUC
F/I

V

B

F/I

V

20

4 8 20

20

4 8 20

4

20

Time (h)

6

24

24

C
InR1-G9

GLUTag
C

1

0.4

C

8

8

24

C

Time (h)

24

C

2

2

4

6

12

12

Cdx-2

Cdx-2

Tub

1

D

3.1

Time (h)

Tub

2.9

1

1.4

2.7

2.6

3.7

4.3

Caco-2
C

C

2

2

4

4

6

6

12

12

24

24

Time (h)
Cdx-2

Tub

1

0.9

0.9

1.0

0.9

1.0

Fig. 1. Forskolin ⁄ IBMX activates Cdx-2 promoter and Cdx-2 mRNA expression in the GLUTag cell line. (A) Forskolin ⁄ IBMX (panels ii, iv), but not
TPA or RA (panel i), activated Cdx-2 promoter expression in GLUTag cells. The activation was not observed for the colon cancer cell line HT-29
(panel iii), nor for the promoter-less control LUC reporter pBLUC (panel v). Indicated cell lines were transfected with 3 lg )769 Cdx-2 ⁄ LUC
fusion gene plasmid ()769 Cdx-LUC), or pBLUC. All trans retinoic acid (RA, 1 lM), TPA (1 lM), or forskolin ⁄ IBMX (F ⁄ I, 10 lM each), or ethanol
(vehicle, V) was added 20 h before the cells were harvested for LUC reporter gene analysis. Relative LUC activity was calculated as the fold
increase with the activity in the vehicle treated cells defined as onefold (mean ± SE, n ¼ 3). (B–D). GLUTag (B), InR1-G9 (C), and Caco-2 (D) cell
lines were treated with either control medium (with ethanol as the vehicle, V), or medium with 10 lM forskolin plus 10 lM IMBX (F ⁄ I) at indicated h before harvesting. Total RNA was extracted for northern blot analysis using cDNA probes for hamster Cdx-2 (Cdx-2) or mouse tubulin
(Tub) as the loading control. F ⁄ I, forskolin ⁄ IBMX.

expression. After the endocrine cell lines were treated
for the indicated period of time, total cell RNA was
extracted and analyzed by northern blotting. We found
that, consistent with the data from our LUC reporter
gene assay, Cdx-2 mRNA expression was notably
2748

activated in the GLUTag cell line by 8–24 h treatment
with forskolin ⁄ IBMX (Fig. 1B).

A previous study has shown that the expression of
proglucagon mRNA in the InR1-G9 cell line cannot
be activated by forskolin ⁄ IBMX [26], indicating that
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

Activation of Cdx-2 expression by cAMP

A

B

HT-29/Cdx-2

GLUTag/Cdx-2

4

15

**
Cdx-2/GAPDH

Cdx-2/GAPDH

3

2

10

5

1

0
Treatment V
Time (h) 16

2

F/I
4 6

0
Treatment V

C

F/I

Time (h) 8

16

8

D

FRIC/CDX-2

5

*

4
3
2

Proglucagon/GAPDH

4

*

2

1
0
Treatment V F/I
Time (h) 2 2

this cell line is PKA deficient. For this cell line, we
therefore included additional time points to examine
the effects of forskolin ⁄ IBMX. As shown in Fig. 1C,
2 h after forskolin ⁄ IBMX treatment, Cdx-2 mRNA
expression started to arise. Substantial activation was
observed 4 h after the treatment, and the elevated
Cdx-2 mRNA expression was maintained during the

whole 24 h experimental procedure. Considering the
PKA deficient nature of this cell line, the above observation would suggest that the activation by forskolin ⁄ IBMX on Cdx-2 expression in the proglucagon
producing endocrine cell lines is not a PKA dependent
event (see further examination below).
In an effort to examine Cdx-2 expression profiles in
different cell lines, we were unable to detect Cdx-2
mRNA expression by northern blotting in two colon
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

*

6

Cdx-2/GAPDH

Fig. 2. Examination of the effect of
forskolin ⁄ IBMX on Cdx-2 mRNA expression
in HT-29 cell line by real time RT-PCR. The
nonendocrine colon cancer cell line HT-29
(A), or the endocrine cell line GLUTag (B), or
the primary cell culture FRIC (C,D) were
treated with either the control medium (with
ethanol as the vehicle, V), or the same medium plus 10 lM forskolin and 10 lM IBMX
(F ⁄ I) at indicated h before harvesting. Total
RNA was extracted and real time RT-PCR
experiments were conducted against Cdx-2
genes (A–C) or proglucagon gene (D), as
described in the Experimental procedures
section. Results obtained were normalized using GAPDH as the control (Cdx-2 ⁄
GAPDH). N ¼ 4. *P < 0.01; **P < 0.005.

FRIC/Proglucagon

V F/I
6 6

V F/I
24 24

0
Treatment V F/I
Time (h) 2 2

V F/I
6 6

V F/I
24 24

cancer cell lines, HT-29 and SW480 (data not shown).
We, however, could detect Cdx-2 mRNA expression in
the human colon cancer Caco-2 cell line. As shown in
Fig. 1D, forskolin ⁄ IBMX treatment generated no substantial effect on Cdx-2 mRNA expression in the
Caco-2 cell line.
We then developed a real time RT-PCR approach
to examine the effect of forskolin ⁄ IBMX on Cdx-2
mRNA expression in the HT-29 cell line, while the gut
endocrine GLUTag cell line was utilized as the positive
control. As shown in Fig. 2A, in HT-29 cells there is
no significant effect on Cdx-2 mRNA expression by

forskolin ⁄ IBMX treatment for 2, 4, 6, and 16 h. In
contrast, Cdx-2 mRNA expression in the GLUTag cell
line was significantly activated by an 8- h treatment
(Fig. 2B), indicating that the real time RT-PCR
2749


Activation of Cdx-2 expression by cAMP

approach utilized here could detect elevated Cdx-2
mRNA expression.
The nonendocrine cell lines utilized in this study are
colon cancer cells of human origin. To investigate whether the lack of response to forskolin ⁄ IBMX on Cdx-2
expression in these cell lines is associated with their
cancerous status, we conducted further examination
using the primary fetal rat intestinal cell (FRIC) cultures. It has been reported that 99% of FRIC cells
represent the nonendocrine intestinal cells, while 1%
population represents the proglucagon producing
endocrine L cells [27]. Proglucagon mRNA expression
in the endocrine L cell lines and in the FRIC cultures
can be activated by forskolin ⁄ IBMX [28–31]. As
shown in Fig. 2C, Cdx-2 mRNA expression in the
FRIC cultures was not apparently affected by forskolin ⁄ IBMX treatment for 2 and 6 h, and the 24-h treatment repressed Cdx-2 mRNA expression. On the other
hand, proglucagon mRNA expression in the FRIC cultures was significantly stimulated by forskolin ⁄ IBMX
treatment for 6 and 24 h (Fig. 2D). The above real
time RT-PCR results further supported our suggestion
that forskolin ⁄ IBMX specifically upregulates Cdx-2
expression in the proglucagon producing endocrine
cells. Forskolin ⁄ IBMX may repress, or at least not
activate, Cdx-2 expression in nonendocrine intestinal

cells.
It should be pointed out that when FRIC cells were
examined using the real time RT-PCR approach, large
S.D. values were generated in each set of assay. This
is understandable because of the heterogeneity of the
FRIC cultures. When, however, HT-29 was examined,
the S.D. values in each set of experiments were also
relatively high. We have no proper explanation for this
observation at this time.
Forskolin ⁄ IBMX or cAMP stimulates Cdx-2
protein expression in the proglucagon producing
endocrine cells
We next examined whether forskolin ⁄ IBMX activates
Cdx-2 protein expression. The GLUTag cell line was
examined first. As shown in Fig. 3A, after a 3 h treatment, forskolin ⁄ IBMX substantially activated Cdx-2
protein expression. To our surprise, virtually no activation was observed after 6 or 12 h treatment (Fig. 3A).
This is in contrast with the activation profile by forskolin ⁄ IBMX treatment at the Cdx-2 mRNA level
(Fig. 1B,C). We then focused on assessing the activation within 4 h period. Figure 3B shows the substantial
activation by forskolin ⁄ IBMX treatment on Cdx-2
protein expression in the GLUTag cell line at 1, 2, 3,
and 4 h.
2750

L. Chen et al.

To investigate whether the activation of Cdx-2 protein expression by forskolin ⁄ IBMX is caused by
increasing intracellular levels of cAMP, the cell membrane permeable cAMP analogue, 8-Br-cAMP, was
utilized for the InR1-G9 cell line (Fig. 3C). After the
cells were treated with forskolin ⁄ IBMX or 8-Br-cAMP
for 2 or 4 h, Cdx-2 protein expression was elevated

substantially, suggesting that the intracellular level of
cAMP plays a role in the regulation of Cdx-2 expression. To our surprise, the cell membrane permeable
cGMP analogue, 8-Br-cGMP, considered as a negative
control in our experimental design, also activated
Cdx-2 protein expression substantially (Fig. 3C).
The above observations also indicated that the vehicle
utilized in this study (from a 1 to 12 h period) had virtually no effect on Cdx-2 protein expression. Our further
examinations were then conducted using the vehicle as
the control for the longest time point. Figure 3D shows
the activation of forskolin ⁄ IBMX treatment on Cdx-2
protein expression in the gut endocrine STC-1 cell line
at 1, 2 and 4 h, but not at 6 h. We then examined the
effect of forskolin ⁄ IBMX on Cdx-2 protein expression
in the Caco-2 cell line. As shown in Fig. 3E, no activation was observed. In conducting this examination, we
included the Epac pathway specific cAMP analogue 8pMeOPT-2¢O-Me-cAMP. This analogue also generated
no stimulatory effect on both Cdx-2 protein expression
and ERK1 ⁄ 2 phosphorylation (see below).
The activation may not be mediated by PKA
As indicated above, the InR1-G9 cell line may carry a
defect in its PKA signaling pathway. Forskolin ⁄ IBMX
failed to stimulate the expression of proglucagon gene
mRNA and its promoter in this cell line, in contrast
to significant activation by forskolin ⁄ IBMX in primary
pancreatic islet cell cultures or in intestinal endocrine
L cell lines [26,29–31]. Based on those observations, we
further investigated the involvement of PKA in Cdx-2
expression. First, we compared the effects of PKAc
cotransfection on the expression of the Cdx-2 promoter vs. the proglucagon gene promoter. PKAc cotransfection generated no significant effect on Cdx-2
promoter expression in InR1-G9 (Fig. 4A, left panel)
and GLUTag cells (data not shown). PKAc cotransfection, however, significantly activated the expression of

the proglucagon gene promoter in InR1-G9 (Fig. 4A,
right panel) and GLUTag cells (data not shown).
Similar to what has been observed for the GLUTag
cell line shown in Fig. 1A, forskolin ⁄ IBMX treatment
activated Cdx-2 gene promoter, approximately 2.5fold, in the InR1-G9 cell line (Fig. 4B). Further LUC
reporter gene analyses were conducted using different
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

A

Activation of Cdx-2 expression by cAMP

B

GLUTag
3

12

6

Time (h)

12

GLUTag
1

1

2

2

3

3

4

4

Time (h)

Cdx-2

Cdx-2

Actin
Actin
-

-

+

+

-

+

+

-

+

+

F/I

-

+

-

-

+

-

-

+

-

-

V

1

.92 3.1

2.9 1.1

-

-

+

-

+

-

+

F/I

+

1 .94 1.2 .99 1.2 fold

+
-

+

-

+

-

+

-

V

1 3.6 1 2.9 1 3.7 1 3.5
InR1-G9

C

STC-1

D

2h

C

4h

Time (h)

V

1

2

4

6

Cdx-2

E

+ - - + - - +
- - 1.9 2.9 2.7

+
+
.89 2.4

- + - +
- 2.3 2.7

F/I
cGMP
cAMP
V
fold

Time (h)
Cdx-2

Actin
- - - - +
1 .97

fold

Actin
1

3.9 2.7 1.4 0.7

fold

Caco-2
1

V
2 4

1

F/I
2 4

Epac
1 2 4
Cdx-2

Actin
pERK

ERK
Fig. 3. Comparison of the effect of forkolin ⁄ IBMX on Cdx-2 protein expression in endocrine cell lines vs. the nonendocrine HT-29 cell line.
(A,B) GLUTag cells were grown in the presence of ethanol (vehicle, V) or 10 lM forskolin plus 10 lM IBMX (F ⁄ I) for the indicated h before
the cells were harvested for examination of Cdx-2 protein expression by western blotting. The same membranes were stripped and followed
by hybridization with an anti-(b-actin) Ig (loading control). (C) InR1-G9 cells were grown in the presence of ethanol (vehicle, V) or 10 lM forskolin plus 10 lM IBMX (F ⁄ I), or 0.25 mM 8-Br-cGMP (cGMP), or 0.25 mM 8-Br-cAMP (cGMP), for the indicated h before the cells were harvested for examination of Cdx-2 protein expression by western blotting. The same membranes were stripped and hybridized with an anti-(bactin) Ig. (D) The response to forskolin ⁄ IBMX treatment on Cdx-2 expression in the STC-1 cell line. (E) Caco-2 cells were treated with 10 lM
forskolin and IBMX, or 10 lM Epac pathway specific cAMP analogue (8-pMeOPT-2¢O-Me-cAMP) and their effects on Cdx-2 expression and
ERK1 ⁄ 2 phosphorylation were examined. Antibodies against b-actin and total ERK1 ⁄ 2 were utilized to ensure equal loading.

sized Cdx-2 ⁄ LUC fusion gene constructs [20] for both
InR1-G9 and GLUTag cell lines for the identification
of cis-element(s) that mediates forskolin ⁄ IBMX treatFEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

ment, and our results indicate that it is unlikely that a
putative CRE element in the Cdx-2 gene promoter is
involved (data not shown).
2751


Activation of Cdx-2 expression by cAMP

InR1-G9

InR1-G9

NS

1

0
Rep

-769 Cdx-LUC
0 .5 1.0
PKAc
0
pCDNA3 1.0 .5
InR1-G9

B
Relative LUC Activity

3

*

2

1

0
Rep -769 Cdx-LUC
+
F/I

Relative LUC Activity

NS

**

6

**
4

2

0
Rep

-472 GLU-LUC

0
PKAc
pCDNA3 1.0

.5

1.0

.5

0

C
Relative PKA Activity (%)

Relative LUC Activity

2

PKAc

300

200

InR1-G9
GLUTag

A

L. Chen et al.

100

0
-

-

+

Fig. 4. PKAc cDNA cotransfection has different effects on Cdx-2
promoter vs. proglucagon gene promoter expression in the PKA
deficient InR1-G9 cell line. (A) InR1-G9 cells were cotransfected
with 3.0 lg )769 Cdx-LUC (left panel) or 3.0 lg )472 GLU-LUC
(right panel), plus the indicated amount of PKAc and ⁄ or pCDNA3
(vector for PKAc). Cells were harvested 20 h after the transfection,
and relative LUC reporter gene activity was calculated as the fold
increase with the activity in the cells received no PKAc transfection, defined as onefold (mean ± SE, n ¼ 3). (B) InR1-G9 cells were
transfected with 3 lg )769 Cdx-2 ⁄ LUC fusion gene plasmid ()769
Cdx-LUC). Forskolin ⁄ IBMX (F ⁄ I, 10 lM each), or ethanol (vehicle, V)
was added 20 h before the cells were harvested for LUC reporter
gene analysis. Relative LUC activity was calculated as the fold
increase with the activity in the vehicle treated cells defined as
onefold (mean ± SE, n ¼ 3). (C) PKA activities in GLUTag and InR1G9 cells, and PKAc transfected InR1-G9 cells were assayed.
Approximately 3.0 mg total cell lysates were utilized for immunoprecipitation, and one third of the precipitate was used for the PKA
assay. Relative PKA activity in untransfected and PKAc transfected
InR1-G9 cells was calculated as the percentage of that obtained
from GLUTag cells (mean ± SE, n ¼ 5).

We then investigated whether the InR1-G9 cell line
is indeed PKA deficient, and whether PKAc cotransfection would trigger the PKA pathway in this cell
line. We examined the PKA kinase activity in PKAc
2752

transfected InR1-G9 cells along with untransfected
InR1-G9 and GLUTag cells. A representative result is

shown in Fig. 4C. Without the PKAc transfection,
PKA kinase activity in the InR1-G9 cell line was
barely detectable, compared with that detected in the
GLUTag cell line. Nevertheless, the PKA protein
expression for both cell lines was readily detected by
Western blotting (data not shown). However, when the
InR1-G9 cells were transfected with the PKAc expression plasmid, the PKA kinase activity was significantly
elevated. These results confirm that the InR1-G9 cell
line is defective in PKA activity and cotransfection of
PKAc restores its PKA activity.
Finally we tested the effect of H-89, a known inhibitor of PKA, on the expression of both the proglucagon gene promoter and the Cdx-2 gene promoter in
the PKA active GLUTag cell line. As shown in Fig. 5,
the effects of H-89 on these two promoters are apparently different. At the concentrations of either 0.1 or
1 lm, H-89 significantly repressed the basal expression
of the proglucagon gene promoter, although forskolin ⁄
IBMX stimulated expression cannot be completely
blocked (left panel). However, at the same concentrations, H-89 failed to repress the Cdx-2 promoter. Thus,
H-89 inhibits the proglucagon gene promoter presumably by blocking the PKA signaling pathway. That the
expression of the Cdx-2 gene promoter was not affected
by H-89, is probably due to the fact that pathway(s)
other than PKA is ⁄ are responsible for mediating the
activation by forskolin ⁄ IBMX and membrane permeable cAMP treatment.
Dominant negative Ras represses the Cdx-2 gene
promoter
Numerous recent reports suggested that the second
messenger cAMP may use the cAMP-Epac-Ras ⁄ RapMEK-MAPK signaling pathway to regulate gene
expression and other cellular functions [32,33]. In
addition, it has been suggested that in colon cancer
cell lines, Cdx-2 expression could be down regulated
by oncogenic Ras [34]. We therefore investigated the

effect of a dominant negative Ras molecule on Cdx2 promoter expression in the proglucagon producing
endocrine cell lines. Figure 6 shows that the dominant negative Ras significantly inhibited Cdx-2 promoter in the InR1-G9 cell line (right panel). In
contrast, significant repression was not observed
when either the TK promoter (data not shown) or
the proglucagon gene promoter (left panel) was tested. These results indicate that Ras could be a positive factor for Cdx-2 expression in proglucagon
producing endocrine cells.
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

Activation of Cdx-2 expression by cAMP

4

Fig. 5. Comparison of the effect of H-89 on
expression of the proglucagon and Cdx-2
gene promoters. Three micrograms )476
GLU-LUC (left panel) or )769 Cdx-2-LUC
reporter gene plasmid was transfected into
the GLUTag cell line. Ten micromolar forskolin plus 10 lM IBMX (F ⁄ I) was added to the
cells 20 h before harvesting, with or without
indicated amount of H-89 (added 45 min
before the addition of forskolin ⁄ IBMX).
Relative LUC reporter gene activity was
calculated as the fold increase with the
activity in the untreated cells defined as
onefold (mean SE, n ẳ 3).

1.8

1.4

Relative LUC Activity

NS
Relative LUC Activity

2

1

0
-

F /I
H-89(àM)

-

+

-

-

1

1

1.0

0.5

0

**

0
+

+

Cdx-2/LUC

+

+

+

-

-

+

PeDHA3

+

-

PeDHA3

Ras-PRSv

-

+

Ras-PRSv

Fig. 6. The effect of a dominant negative ras cDNA cotransfection
on Cdx-2 promoter expression. InR1-G9 cells were cotransfected
with 3 lg )476 GLU ⁄ LUC (left panel) or )769 Cdx-2 ⁄ LUC, and
1.5 lg pCDNA3 (vector), or 1.5 lg dominant negative ras cDNA.
Cells were harvested 20 h later for LUC reporter gene analysis. Relative LUC reporter gene activity was calculated as the fold increase
with the activity in the untreated cells defined as onefold (mean ±
SE, n ¼ 3).

Forskolin/IBMX stimulates ERK1/2 phosphorylation in proglucagon producing cell lines
As a small GTPase, Ras may utilize MAP kinase to
carry out its signaling and biological functions. We
then investigated the effect of forskolin ⁄ IBMX treatment on the phosphorylation status of one of the
MAP kinases, ERK1 ⁄ 2 in proglucagon producing cell
lines.
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

3

2

1

0
Glu/LUC
+
+

1.5

1.2

Glu/LUC

3

Relative LUC Activity

Relative LUC Activity

4

0.1 0.1

-

Cdx-2/LUC
+ - + - +

H-89(µM) -

- 0.1 0.1 1 1

F /I

Forskolin ⁄ IBMX treatment was found to stimulate
ERK1 ⁄ 2 phosphorylation in GLUTag, STC-1, and
InR1-G9 cell lines. A representative result from the
InR1-G9 cell line is shown in Fig. 7A. The expression
level of phosphorylated ERK1 ⁄ 2 was elevated 5 min
after forskolin ⁄ IBMX treatment, while the effect on
total ERK1 ⁄ 2 expression was not appreciable. In this
particular experiment, forskolin ⁄ IMBX treatment for
120 min did not activate ERK1 ⁄ 2 phosphorylaton
(lane 5). We, however, observed substantial activation
for this time course for three other experiments (data
not shown). The activation of ERK1 ⁄ 2 phosphorylation by forskolin ⁄ IBMX treatment was associated with
elevated Cdx-2 protein expression, consistent with the
result shown in Fig. 3B. When PD98059, a MEK-1
inhibitor, was included, forskolin ⁄ IBMX mediated
ERK1 ⁄ 2 phosphorylation was significantly inhibited,
starting from 30 min (lanes 9–12). PD98059 also inhibited the basal expression of phosphorylated ERK1 ⁄ 2
(comparing lane 1 with lane 7).
Although the basal expression of phosphorylated
ERK1 ⁄ 2 was inhibited by PD98059, the basal Cdx-2
expression was not affected (comparing lane 7 with
lane 1). However, PD98059 inhibited activated Cdx-2
protein expression by forskolin ⁄ IBMX treatment for

4 h (comparing lane 12 with lane 6), despite the
absence of the repression at the other time courses. It
should be pointed out that if phosphorylated ERK1 ⁄ 2
mediates forskolin ⁄ IBMX stimulated Cdx-2 transcription, further phosphorylation events on transcription
factor(s) should be involved. Inhibition of ERK1 ⁄ 2
phosphorylation by PD98059, may not affect the factors ⁄ mediators that have already been phosphorylated
2753


Activation of Cdx-2 expression by cAMP

A

L. Chen et al.

InR1-G9
1

2 3

4

5 6 7

8

9 10 11 12
Cdx-2
Actin

phosphorylation to that induced by forskolin ⁄ IBMX
treatment. As we have already presented in Fig. 3E,
such stimulatory effects were not observed when the
colon cancer cell line Caco-2 was examined.

pERK1/2
ERK1/2
F/I (min) PD98059 -

B

- 5
- + +

5

30 60 120 240

-

Several groups have investigated signaling molecules or
pathways that may up or down regulate Cdx-2 expression in the nonendocrine intestinal epithelia [34,37–41].
Cdx-2 expression was found to be down regulated by
the oncogenic Ras [34], and up regulated by butyrate
[39]. It has also been reported that Cdx-2 expression
cannot be detected in the adnomatous polyposis coli
(APC) mutated colon cancer cell lines, while introducing
wild type APC cDNA into an APC mutated cell line
rendered it to re-express Cdx-2 mRNA [37]. However,
a recent immunohistochemistry study indicated that

Cdx-2 expression in the mouse gut was not altered by
APC or Ras status, or by butyrate treatment [40].
Another study by Kim et al. linked Cdx-2 expression
with the tumor suppressor PTEN (phosphatase and
tension homologue deleted from chromosome 10), and
the phosphatidylinositol 3-kinase (PI3K) signaling pathway [38]. However, little is known about signaling molecules and pathways that may regulate Cdx-2 expression
in the proglucagon producing endocrine cells.
In this study, we examined the effects of forskolin ⁄ IBMX on Cdx-2 expression, and investigated possible signaling pathways that mediate such expression.
We found that Cdx-2 expression could be activated by
forskolin ⁄ IBMX. Using a reporter gene assay and northern blotting, we demonstrated that this activation
occurs at the transcriptional ⁄ mRNA level. Additionally, Cdx-2 protein expression was also elevated by
forskolin ⁄ IBMX as detected by western blotting.
However, the activation only takes place in the proglucagon producing endocrine cell lines, but not in the
nonendocrine colon cancer cell lines. Those data collectively suggested that activation of Cdx-2 expression by
forskolin ⁄ IBMX occurs in a cell type specific manner.
The InR1-G9 cell line may be PKA deficient [26].
We found that indeed PKA kinase activity in this cell
line is significantly low, compared with that in the
GLUTag cell line (Fig. 4C). The observations that
forskolin ⁄ IBMX activates the Cdx-2 promoter with or
without a putative CRE element (data not shown),
and that Cdx-2 promoter and endogenous Cdx-2
mRNA and protein expression in the InR1-G9 cell line
were also activated by forskolin ⁄ IBMX treatment
(Figs 1C, 3C and 4B), led to the hypothesis that the
activation is not mediated by PKA. We found that in
three proglucagon producing cell lines, cotransfection

30 60 120 240

-

Discussion

+ + + +

-

-

InR1-G9
V
F/I
1 2 4 1 2 4

Epac
1 2 4
Cdx-2
Actin

pERK
ERK

Fig. 7. The effect of forskolin ⁄ IBMX and Epac pathway specific
cAMP analogue on the phosphorylation status of ERK1 ⁄ 2 and Cdx2 protein expression. A. InR1-G9 cells were grown in the presence
of ethanol (vehicle for forskolin ⁄ IBMX, V) or 10 lM forskolin plus
10 lM IBMX (F ⁄ I) for the indicated time before the cells were
harvested for examination of Cdx-2 and phosphorylated ERK ⁄ 1 ⁄ 2
expression by western blotting. For one set of cells, the MEK inhibitor PD98059 (50 lM) was added 45 min before the addition of F ⁄ I.
The same membranes were stripped and followed by hybridization

with an anti-(b-actin) Ig (loading control), and the antibody against
total ERK1 ⁄ 2. B. InR1-G9 cells were grown in the presence of
10 lM 8-pMeOPT-2¢O-Me-cAMP [F ⁄ I as the positive control, and
vehicle (V) as the negative control]. Cdx-2 expression and ERK1 ⁄ 2
phosphorylation were then examined. Antibodies against b-actin
and total ERK1 ⁄ 2 were utilized to ensure equal loading.

by active ERK1 ⁄ 2. Therefore, it is reasonable to accept
a 2- to 3-h time delay to attenuate Cdx-2 protein
expression in response to the inhibition on ERK1 ⁄ 2
phosphorylation (see the Discussion section for our
further interpretation).
Epac pathway specific cAMP analogue stimulates
Cdx-2 protein expression and ERK 1/2
phosphorylation
Finally, we initiated an examination of whether the
Epac signaling pathway [35,36] specific cAMP analogue would also stimulate Cdx-2 protein expression
in the InR1-G9 cell line. As shown in Fig 7B,
10 lm 8-pMeOPT-2¢O-Me-cAMP generated comparable
effects on both Cdx-2 protein expression and ERK1 ⁄ 2
2754

FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

of a dominant negative Ras expression plasmid
significantly repressed Cdx-2 expression, but not the
proglucagon gene promoter, indicating that Ras is a

positive factor. This is in contrast with the previous
report that Ras is a negative factor for Cdx-2 expression in colon cancer cell lines [34]. In addition, we
found that forskolin ⁄ IBMX specifically stimulates
ERK1 ⁄ 2 phosphorylation in proglucagon producing
cell lines, and MEK inhibition attenuated forskolin ⁄ IBMX activated Cdx-2 protein expression at 4 h
(Fig. 7). These observations collectively suggest the
involvement of cAMP-Epac-Ras ⁄ Rap-MEK-MAPK
signaling pathway (see below).
We noticed the discrepancy regarding the effects of
forskolin ⁄ IBMX on Cdx-2 expression at the mRNA
and protein levels. First, the activation of Cdx-2 protein expression could be observed as early as 30 min.
Such an early response would suggest that, in addition
to stimulating Cdx-2 mRNA expression, forskolin ⁄
IBMX might also play a role in stabilizing Cdx-2
protein. Consistently, we found that PD98059 repressed
forskolin ⁄ IBMX activated Cdx-2 protein expression
only at 4 h, but not within the first 2 h. One may speculate that forskolin ⁄ IBMX may stabilize Cdx-2 protein
via a yet to be identified mechanism, and this effect
cannot be blocked (or immediately blocked) by MEK
inhibition. Activated ERK1 ⁄ 2, however, may stimulate
Cdx-2 transcription via phosphorylating its transcriptional activators, and this event would take longer time,
and it could be blocked by MEK inhibition. Second,
activation by forskolin ⁄ IBMX at the protein level was
not observed beyond 4 h. However, the activation at
the mRNA level was detectable by northern blot analysis during the whole 2–24 h experimental period for the
InR1-G9 cell line (Fig. 1C). One may postulate the
existence of a negative feedback loop at the Cdx-2 protein expression level to explain such a difference. To
our surprise, both 8-Br-cAMP and 8-Br-cGMP were
found to activate Cdx-2 protein expression in InR1-G9
cells. A possible explanation would be the involvement

of a nucleotide gated ion channel [32,42,43] in these
endocrine cell lines. We made an attempt to identify
the existence of such channel in the InR1-G9 cell line
without success. However, we did observe that treating
InR1-G9 cells with 45 mm potassium chloride (inducing membrane depolarization) led to enhanced Cdx-2
protein expression. The discrepancy between the
responses of Cdx-2 protein and mRNA expression to
forskolin ⁄ IBMX treatment further indicated the complexity of the corresponding regulatory networks. One
may suggest that this discrepancy and the presence of
negative feedback loops may partially explain why
opposite results have been obtained by different research
FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

Activation of Cdx-2 expression by cAMP

groups in assessing the role of APC, Ras status, and
butyrate treatment on Cdx-2 expression [34,37–41].
An important observation in this study is that the
activation of Cdx-2 promoter by forskolin ⁄ IBMX took
place only in the proglucagon producing endocrine cells.
Such a finding supported our overall hypothesis that
signaling molecules and ⁄ or pathways involved in Cdx-2
expression differ in the endocrine cells from that in the
nonendocrine intestinal epithelia. However, as the nonendocrine cell lines utilized in this study are cancerous
cells of human origin, we employed the primary FRIC
culture system for conducting real time RT-PCR examinations (Fig. 2). A majority of epithelial cells (99%) in
FRIC cultures are nonendocrine [28,29]. Such cultures
therefore would represent primary nonendocrine intestinal cells. Lack of the response to forskolin ⁄ IBMX
treatment in the FRIC cultures on Cdx-2 mRNA
expression supported our overall hypothesis that the

second messenger cAMP activates Cdx-2 expression
only in the proglucagon producing endocrine cells.
Although the second messenger cAMP has long been
shown to mediate specific intracellular signaling events
through PKA [44,45], more recent observations have
suggested that PKA does not account for all of the
intracellular targets of cAMP, especially in endocrine
cells [32,33,42,43]. Furthermore, the identification of
novel cAMP binding proteins that exhibit guanine
nucleotide exchange (GEF) activities (cAMP-GEFs, or
Epac) has opened a new research direction for understanding the function of the second messenger cAMP
[32,35,36,40]. A GEF molecule may serve as a bridge
between cAMP and small GTPases, including Ras,
Rap1 and Rap2, leading to the activation of Raf and the
mitogen activated protein kinase (MAPK) signaling
pathway [32]. The observation that ras may serve as a
positive factor for Cdx-2 expression (Fig. 6), and both
the Epac pathway specific cAMP analogue and forskolin ⁄ IBMX stimulate ERK1 ⁄ 2 phosphorylation and
Cdx-2 expression collectively suggest the involvement of
the Epac-Ras ⁄ Rap-ERK pathway in regulating Cdx-2
expression. To examine the expression profile of Epac
molecules in pancreatic and intestinal proglucagon producing cells, and to identify whether Ras and ⁄ or Rap
are indeed involved in regulating Cdx-2 expression
would further our understanding on this cell type specific gene expression event.

Experimental procedures
Materials
Tissue culture medium, calf serum and oligonucleotides
were purchased from Invitrogen Life Technology Inc

2755


Activation of Cdx-2 expression by cAMP

(Burlington, Ontario, Canada). Radioisotopes were obtained
from Amersham Pharmacia Biotech (Baie d’Urfe, Quebec,
Canada). Forskolin, dideoxyforskolin, 3-isobutyl-1-methylxanthine (IBMX), 8-Bromoadenosine 3¢, 5¢-cyclin monophoshate (8-Br-cAMP), 8-Bromoguampsome 3¢, 5¢-cyclin
monophoshate (8-Br-cGMP), all trans retinoic acid (ATRA,
RA), and 12-O-tetraadecanoylphorbol 13-acetate (TPA)
were purchased from Sigma-Aldrich (Oakville, Ontario,
Canada). The mitogen activated protein kinase (MAPK)
kinase (MEK) inhibitor PD98059 was purchased from Calbiochem (EMD Biosciences, Inc., San Diego, California).
The Epac pathway specific cAMP analogue 8-pMeOPT2¢O-Me-cAMP was purchased from BIOLOG life Science
Institute (Bremen, Germany).

Plasmids
The Cdx-2 ⁄ Luciferase (LUC) fusion gene constructs, and
the proglucagon ⁄ LUC (GLU ⁄ LUC) fusion gene constructs
were generated in previous studies, using the promoterless
LUC reporter gene plasmid pBLUC as the cloning vector
[4,20,46]. The catalytic subunit of PKA expression plasmid
(PKAc), in which the expression of the rat PKAc is driven
by a CMV promoter, and the dominant negative Ras
expression plasmid, in which the V12 mutant of H-Ras is
driven by the RSV promoter were provided by X. Fang
(MD Andersen Cancer Institute, Houston, TX, USA) [47].

Cell culture, plasmid DNA transient transfection,
and LUC reporter gene analysis

The hamster pancreatic islet cell line InR1-G9, the mouse
intestinal proglucagon producing endocrine cell lines GLUTag and STC-1, and the human colon cancer cell lines HT29, Caco-2 and SW480, were maintained in the Dulbecco’s
modified Eagle’s medium (DMEM), supplemented with
appropriate serum as described previously [4,20,21,46].
InR1-G9, HT29, Caco-2, and SW480 cells were transfected
by the method of calcium phosphate precipitation [4,20].
GLUTag and STC-1 cell lines were transfected by the procedure using LipofectAMINE (Invitrogen Life Technology
Inc). Cells were harvested for LUC reporter gene analysis,
20 h after the transfection procedure [4,46].

L. Chen et al.

The method for preparing the primary fetal rat intestinal
cell (FRIC) cultures has been described in our previous
studies [27–29], using 19–21 day gestation fetal Wistar rats
(Charles River Canada, Saint Constant, Quebec, Canada).
For this study, FRIC cells were grown in the absence or
presence of 10 lm forskolin plus 10 lm IBMX for the indicated period of time. Total RNA was extracted for real
time RT-PCR analysis against the Cdx-2 and proglucagon
genes. In FRIC cultures, the proglucagon producing endocrine L cells account for approximately 1% of the cell numbers [28].

Northern blot analysis
RNA from cultivated cell lines was extracted using the
TRIzol reagent (Invitrogen Life Technology Inc), per
manufacturer’s instruction. Methods for northern blot analysis were described previously, using the hamster Cdx-2
cDNA as the probe to detect Cdx-2 mRNA expression,
and the mouse tubulin cDNA as the probe to detect tubulin
mRNA expression [4].

Western blot analysis

Anti-Cdx-2 Ig was generated in our previous studies
[20,48]. For western blot analysis, whole cell lysate containing approximately 20 lg proteins from each of the cultivated cell lines were size-fractionated by 10% SDS ⁄ PAGE
and transferred to a nitrocellulose membrane (Protran,
Schleicher & Schuell). Cdx-2 immunoreactive protein was
detected with an ECL western blot analysis system (Amersham Pharmacia Biotech), per manufacturer’s instruction,
with the peroxidase-linked anti-(rabbit IgG) Ig as the second antibody. Anti-actin Ig was purchased from SigmaAldrich, while antibodies against phosphorylated ERK1 ⁄ 2
(sc-7383) and total ERK1 ⁄ 2 (sc-94) were purchased from
Santa Cruz Biotechnology, Inc (Santa Cruz, CA).

Reverse transcription of RNA templates
and real-time PCR
Single-strand cDNA synthesis was carried out using the
Superscript II kit, purchased from Invitrogen Life Technol-

Table 1. Primers utilized in real time RT-PCR. Orientation is 5¢)3¢.
Gene

Forward primer

Reverse primer

Rat Cdx-2
Mouse Cdx-2
Human CDX2
Rat GAPDH
Mouse GAPDH
Human GAPDH
Rat proglucagon

AAACCAGGACGAAAGACAAATACC

GGACGTGAGCATGTATCCTAGCT
CCTCGGCAGCCAAGTGAA
TGATTCTACCCACGGCAAGT
AACGACCCCTTCATTGAC
TGCACCACCAACTGCTTAG
GCCATTCACAGGGCACATTC

CCTCCTGATGGTGATGTATCGA
TAACCACCGTAGTCCGGGTACT
AGCGACTGTAGTGAAACTCCTTCTC
AGCATCACCCATTTGATGT
TCCACGACATACTCAGCAC
GACGCAGGGATGATGTTC
CCGGTTCCTCTTGGTGTTCA

2756

FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

ogy Inc (Burlington, Ontario, Canada) per instruction. Real
time PCR analyses were conducted using the machine by
Applied Biosystems (ABI PrismTM 7900HT, Foster City,
CA). Each assay included a standard curve of five serial
dilutions of a known concentration of cDNA and a nontemplate control. All assays were performed in triplicates.
The expression levels of the genes were normalized with the
house keeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). PCR primers were designed using the Primer Express 2.0 software (Applied Biosystems, Foster City,
CA) and purchased from ACGT Corp. (Toronto, Ontario,

Canada). All primers utilized in conducting the real time
PCR are shown in Table 1.

PKA activity assay
A rabbit polyclonal antibody (SC903, C20) against PKAc
and the PKA substrate were purchased from Santa Cruz
Biotechnology, Inc (Santa Cruz, CA). PKA assay was conducted on pancreatic InR1-G9 and intestinal GLUTag cell
lines, along with InR1-G9 cells transfected with PKAc
cDNA, according to manufacture’s instruction. Briefly, cells
were harvested in cold NaCl ⁄ Pi and lysed in a RIPA buffer.
Testing sample was prepared by immunoprecipitation using
the anti-PKA Ig. The kinase assay was carried out in a kinase buffer containing 0.3 mm ATP, 4 mm MgCl2, 0.5 lCi
[32P]ATP[cP], 200 ng PKA substrate, and the testing sample. After an incubation at 30 °C for 60 min, the reaction
was stopped and spotted on the P81 phosphor-cellulose
paper. Unutilized [32P]ATP[cP] was washed away with 1%
phosphoric acid and the radioactivity remained on the P81
paper was measured. The data in Fig. 4C represent 5 independent experiments.

Data analysis
All data (relative luciferase activity and relative PKA activity) are expressed as mean ± SD, n > ⁄ ¼ 3. Statistic differences between samples were assessed using Student
T-test. Significance was assumed at P < 0.05.

Acknowledgements
We thank Drs Donald Branch, Denize Belsham, Harry
Elsholtz, David Irwin and Weiyang Lu for useful
discussions, Dr Xianjun Fang for providing the PKAc
and dominant negative Ras cDNAs, and Dr Zuyao Ni
for technical assistance. This work was supported
by operating grants from the Canadian Institute of
Health Research (CIHR, MOP-62745G), and Canadian

Diabetes Association (CDA, 1198) to TJ, and CIHR
and CDA operating grants to PLB. PLB is a Canadian
Research Chair, LC is a recipient of Banting and Best
Diabetes Center Novartis Graduate Award, and PD is

FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

Activation of Cdx-2 expression by cAMP

a recipient of the CIHR Canadian Digestive Health
Foundation and Ontario Graduate Scholarship Award.

References
1 James R & Kazenwadel J (1991) Homeobox gene
expression in the intestinal epithelium of adult mice.
J Biol Chem 266, 3246–3251.
2 James R, Erler T & Kazenwadel J (1994) Structure of
the murine homeobox gene cdx-2. Expression in
embryonic and adult intestinal epithelium. J Biol Chem
269, 15229–15237.
3 Suh E, Chen L, Taylor J & Traber PG (1994) A homeodomain protein related to caudal regulates intestine-specific gene transcription. Mol Cell Biol 14, 7340–7351.
4 Jin T & Drucker DJ (1996) Activation of proglucagon
gene transcription through a novel promoter element by
the caudal-related homeodomain protein Cdx-2 ⁄ 3. Mol
Cell Biol 16, 19–28.
5 German MS, Wang J, Chadwick RB & Rutter WJ
(1992) Synergistic activation of the insulin gene by a
LIM-homeo domain protein and a basic helix-loop-helix
protein: building a functional insulin minienhancer complex. Genes Dev 6, 2165–2176.
6 Laser B, Meda P, Constant I & Philippe J (1996) The

caudal-related homeodomain protein Cdx-2 ⁄ 3 regulates
glucagon gene expression in islet cells. J Biol Chem 271,
28984–28994.
7 Hussain MA & Habener JF (1999) Glucagon gene transcription activation mediated by synergistic interactions
of pax-6 and cdx-2 with the p300 co-activator. J Biol
Chem 274, 28950–28957.
8 Chawengsaksophak K, James R, Hammond VE, Kontgen F & Beck F (1997) Homeosis and intestinal
tumours in Cdx2 mutant mice. Nature 386, 84–87.
9 Tamai Y, Nakajima R, Ishikawa T, Takaku K, Seldin
MF & Taketo MM (1999) Colonic hamartoma development by anomalous duplication in Cdx2 knockout mice.
Cancer Res 59, 2965–2970.
10 Drummond F, Sowden J, Morrison K & Edwards YH
(1996) The caudal-type homeobox protein Cdx-2 binds
to the colon promoter of the carbonic anhydrase 1 gene.
Eur J Biochem 236, 670–681.
11 Lambert M, Colnot S, Suh ER, L’Horset F, Blin C,
Calliot ME, Raymondjean M, Thomasset M, Traber
PG & Perret C (1996) Cis-Acting elements and transcription factors involved in the intestinal specific
expression of the rat calbindin-D9K gene: binding of
the intestine-specific transcription factor Cdx-2 to the
TATA box. Eur J Biochem 236, 778–788.
12 Troelsen JT, Mitchelmore C, Spodsberg N, Jensen AM,
Noren O & Sjostrom H (1997) Regulation of lactasephlorizin hydrolase gene expression by the caudal-related
homoeodomain protein Cdx-2. Biochem J 322, 833–838.

2757


Activation of Cdx-2 expression by cAMP

13 Taylor JK, Levy T, Suh ER & Traber PG (1997) Activation of enhancer elements by the homeobox gene
Cdx2 is cell line specific. Nucl Acids Res 12, 2293–2300.
14 Mitchelmore C, Troelsen JT, Spodsberg N, Sjostrom H
& Noren O (2000) Interaction between the homeodomain proteins Cdx2 and HNF1alpha mediates expression of the lactase-phlorizin hydrolase gene. Biochem
J 346, 529–535.
15 Park J, Schulz S & Waldman SA (2000) Intestine-specific activity of the human guanylyl cyclase C promoter is
regulated by Cdx2. Gastroenterol 19, 89–96.
16 Yamamoto H, Miyamoto K, Li B, Taketani Y, Kitano
M, Inoue Y, Morita K, Pike JW & Takeda EJ (1999)
The caudal-related homeodomain protein Cdx-2 regulates vitamin D receptor gene expression in the small
intestine. Bone Miner Res 14, 240–247.
17 Hinoi T, Lucas PC, Kuick R, Hanash S, Cho KR
& Fearon ER (2002) CDX2 regulates liver intestinecadherin expression in normal and malignant colon
epithelium and intestinal metaplasia. Gastroenterol 123,
1565–1577.
18 Yamamoto H, Bai YQ & Yuasa Y (2003) CDX2 regulates liver intestine-cadherin expression in normal and
malignant colon epithelium and intestinal metaplasia.
Biochem Biophys Res Commun 300, 813–818.
19 Isshiki S, Kudo T, Nishihara S, Ikehara Y, Togayachi
A, Furuya A, Shitara K, Kubota T, Watanabe M,
Kitajima M & Narimatsu H (2003) Lewis type 1 antigen
synthase (beta3Gal-T5) is transcriptionally regulated by
homeoproteins. J Biol Chem 278, 36611–36620.
20 Xu F, Li H & Jin T (1999) Cell type-specific autoregulation of the Caudal-related homeobox gene Cdx-2 ⁄ 3.
J Biol Chem 274, 34310–34316.
21 Jin T & Li H (2001) POU homeodomain protein OCT1
is implicated in the expression of the caudal-related
homeobox gene Cdx-2. J Biol Chem 276, 14752–14758.
22 Gautier-Stein A, Domon-Dell C, Calon A, Bady I,
Freund JN, Milthieux G & Rajas F (2003) Differential

regulation of the glucose-6-phosphatase TATA box by
intestine-specific homeodomain proteins CDX1 and
CDX2. Nucleic Acids Res 31, 5238–5246.
23 Bai YQ, Miyake S, Iwai T & Yuasa Y (2003) CDX2, a
homeobox transcription factor, upregulates transcription of the p21 ⁄ WAF1 ⁄ CIP1 gene. Oncogene 22,
7942–7949.
24 Kim SP, Park JW, Lee SH, Lim JH, Jang BC, Lee SH,
Jang IH, Freund JN, Suh SI, Mun KC, Song DK, Ha
EM, Lee WJ & Kwon TK (2004) Homeodomain protein CDX2 regulates COX-2 expression in colorectal
cancer. Biochem Biophys Res Commun 315, 93–99.
25 Uesaka T, Kageyama N & Watanabe H (2004) Identifying target genes regulated downstream of Cdx2 by
microarray analysis. J Mol Biol 337, 647–669.
26 Drucker DJ, Campos R, Reynolds R, Stobie K & Brubaker PL (1991) The rat glucagon gene is regulated by a

2758

L. Chen et al.

27

28

29

30

31

32

33

34

35

36

37

38

39

protein kinase A-dependent pathway in pancreatic islet
cells. Endocrinol 128, 394–400.
Anini Y, Hansotia T & Brubaker PL (2002) Muscarinic
receptors control postprandial release of glucagon-like
peptide-1: in vivo and in vitro studies in rats. Endocrinol
143, 2420–2426.
Ni Z, Anini Y, Fang X, Mills G, Brubaker PL & Jin T
(2003) Transcriptional activation of the proglucagon
gene by lithium and beta-catenin in intestinal endocrine
L cells. J Biol Chem 278, 1380–1387.
Drucker DJ & Brubaker PL (1989) Proglucagon gene
expression is regulated by a cyclic AMP-dependent
pathway in rat intestine. Proc Natl Acad Sci USA 86,
3953–3957.
Drucker DJ, Jin T, Asa SL, Young TA & Brubaker PL
(1994) Activation of proglucagon gene transcription by

protein kinase-A in a novel mouse enteroendocrine cell
line. Mol Endocrinol 8, 1646–1655.
Gajic D & Drucker DJ (1993) Multiple cis-acting
domains mediate basal and adenosine 3¢,5¢-monophosphate-dependent glucagon gene transcription in a mouse
neuroendocrine cell line. Endocrinol 132, 1055–1062.
Richards JS (2001) New signaling pathways for hormones and cyclic adenosine 3¢,5¢-monophosphate action
in endocrine cells. Mol Endocrinol 15, 209–218.
Kupperman E, Wen W & Meinkoth JL (1993) Inhibition
of thyrotropin-stimulated DNA synthesis by microinjection of inhibitors of cellular Ras and cyclic AMP-dependent protein kinase. Mol Cell Biol 13, 4477–4484.
Lorentz O, Cadoret A, Duluc I, Capeau J, Gespach C,
Cherqui G & Freund JN (1999) Downregulation of the
colon tumour-suppressor homeobox gene Cdx-2 by
oncogenic ras. Oncogene 18, 87–92.
de Rooij J, Zwartkruis FJ, Verheijen MH, Cool RH,
Nijman SM, Wittinghofer A & Bos JL (1998) Epac is a
Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP. Nature 396, 474–477.
Ebinu JO, Bottorff DA, Chan EY, Stang SL, Dunn RJ
& Stone JC (1998) RasGRP, a Ras guanyl nucleotidereleasing protein with calcium- and diacylglycerol-binding motifs. Science 280, 1082–1086.
da Costa LT & He TC., YuJ, Sparks AB, Morin PJ,
Polyak K, Laken S, Vogelstein B & Kinzler KW (1999)
CDX2 is mutated in a colorectal cancer with normal
APC ⁄ beta-catenin signaling. Oncogene 18, 5010–5014.
Kim S, Domon-Dell C, Wang Q, Chung DH, Di Cristofano A, Pandolfi PP, Freund JN & Evers BM (Year?)
PTEN and TNF-alpha regulation of the intestinal-specific Cdx-2 homeobox gene through a PI3K, PKB ⁄ Akt,
and NF-kappaB-dependent path. Gastroenterol 123,
1163–1178.
Domon-Dell C, Wang Q, Kim S, Kedinger M, Evers
BM & Freund JN (2002) Stimulation of the intestinal
Cdx2 homeobox gene by butyrate in colon cancer cells.
Gut 50, 525–529.

FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS


L. Chen et al.

40 Christensen AE, Selheim F, de Rooij J, Dremier S,
Schwede F, Dao KK, Martinez A, Maenhaut C, Bos
JL, Genieser HG & Døskeland SO (2003) cAMP analog
mapping of Epac1 and cAMP kinase. Discriminating
analogs demonstrate that Epac and cAMP kinase act
synergistically to promote PC-12 cell neurite extension.
J Biol Chem 278, 35394–35402.
41 Qualtrough D, Hinoi T, Fearon E & Paraskeva C
(2002) Expression of CDX2 in normal and neoplastic
human colon tissue and during differentiation of an in
vitro model system. Gut 51, 184–190.
42 Kaupp UB & Seifert R (2002) Cyclic nucleotide-gated
ion channels. Physiol Rev 82, 769–824.
43 Dremier S, Kopperud R, Doskeland SO, Dumont JE &
Maenhaut C (2003) Search for new cyclic AMP-binding
proteins. FEBS Lett 546, 103–107.
44 Drucker DJ, Philippe J, Mojsov S, Chick WL & Habener. JF (1987) Glucagon-like peptide I stimulates insulin gene expression and increases cyclic AMP levels in a

FEBS Journal 272 (2005) 2746–2759 ª 2005 FEBS

Activation of Cdx-2 expression by cAMP

45

46

47

48

rat islet cell line. Proc Natl Acad Sci USA 84, 3434–
3438.
Vallejo M, Penchuk L & Habener JF (1992) Somatostatin gene upstream enhancer element activated by a
protein complex consisting of CREB, Isl-1-like, and
alpha-CBF-like transcription factors. J Biol Chem 267,
12876–12884.
Jin T & Drucker DJ (1995) The proglucagon gene
upstream enhancer contains positive and negative
domains important for tissue-specific proglucagon gene
transcription. Mol Endocrinol 9, 1306–1320.
Fang X, Yu SX, Lu Y, Bast RC Jr, Woodgett JR &
Mills GB (2000) Phosphorylation and inactivation of
glycogen synthase kinase 3 by protein kinase A. Proc
Natl Acad Sci USA 97, 11960–11965.
Trinh KY, Jin T & Drucker DJ (1999) Identification of
domains mediating transcriptional activation and cytoplasmic export in the caudal homeobox protein Cdx-3.
J Biol Chem 274, 6011–6019.

2759