EphrinA5 suppresses colon cancer development by
negatively regulating epidermal growth factor receptor
stability
Tong-Hong Wang
1,
*, Junn-Liang Chang
2,3,
*, Jar-Yi Ho
4,
*, Hsiao-Chun Wu
1
and Tse-Ching Chen
1,5
1 Tissue Bank, Chang Gung Memorial Hospital, Tao-Yuan, Taiwan
2 Department of Pathology & Laboratory Medicine, Taoyuan Armed Forces General Hospital, Tao-Yuan, Taiwan
3 Biomedical Engineering Department, Ming Chuang University, Tao-Yuan, Taiwan
4 Graduate Institute of Life Sciences, National Defense Medical Center, Taipei, Taiwan
5 Department of Anatomic Pathology, Chang Gung Memorial Hospital, Chang Gung University School of Medicine, Tao-Yuan, Taiwan
Keywords
carcinogenesis; colon cancer; ephrinA5;
epidermal growth factor receptor (EGFR);
tumor suppressor
Correspondence
T.-C. Chen, Department of Anatomic
Pathology, Chang Gung Memorial Hospital,
5, Fu-Shin St, Kwei-Shan, Tao-Yuan 333,
Taiwan
Fax: +886 3 3280147
Tel: +886 3 3281200 ext. 2730
E-mail: ctc323@adm.cgmh.org.tw
*These authors contributed equally to this
work
(Received 6 July 2011, revised
4 October 2011, accepted 26 October
2011)
doi:10.1111/j.1742-4658.2011.08419.x
Colon cancer is one of the most common human cancers worldwide. Owing
to its aggressiveness and lethality, it is necessary to determine the mecha-
nisms regulating the carcinogenesis of colon cancer. EphrinA5 has been
reported to act as a putative tumor suppressor in glioma; however, little is
known concerning the role of this protein in the context of colon cancer. To
elucidate the biological significance of ephrinA5 in colon cancer, we exam-
ined ephrinA5 and epidermal growth factor receptor (EGFR) expression
profiles in both colon cancer and normal tissues, using immunohistochemis-
try on a 96-spot tissue microarray. Gain-of-function and loss-of-function
experiments were performed on the human colon cancer cell lines SW480
and WiDr to determine the biological effects of ephrinA5 in relation to cell
proliferation, survival, and migration. It was found that ephrinA5 mRNA
and protein levels were significantly reduced in colon cancer as compared
with normal colon tissue specimens. EphrinA5 expression was also nega-
tively associated with tumor differentiation and clinical stage. In colon
cancer cell line models, ephrinA5 exerted an inhibitory effect on EGFR by
promoting c-Cbl-mediated EGFR ubiquitination and degradation. Ephri-
nA5 did not affect the transcriptional regulation of EGFR mRNA expres-
sion in colon cancer cells. Expression of ephrinA5 suppressed colon cancer
cell proliferation, migration, and chemotherapeutic resistance. In conclusion,
ephrinA5 inhibited colon cancer progression by promoting c-Cbl-mediated
EGFR degradation. Our findings identify a novel mechanism that could be
utilized to improve the therapeutic efficiency of EGFR-targeting strategies.
Introduction
Colon cancer is the third most common human
cancer diagnosed in developed countries [1]. Colon
cancer also ranks third among life-threatening and
fatal malignancies in Taiwan, with a 5-year survival
rate of 53–57% over the past decade. This survival
rate, however, declines to < 10% in patients with
distant metastases at the time of diagnosis [2]. The
underlying pathogenesis of this disease remains
poorly understood [1,2]. Previous clinical studies have
shown that epidermal growth factor receptor
(EGFR) overexpression occurs in the invasive front
of high-stage colorectal cancers, indicating that
EGFR is a critical risk factor for colon cancer pro-
gression [3,4].
Abbreviations
ACP, acid phosphatase; EGFR, epidermal growth factor receptor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase;
siRNA, small interfering RNA.
FEBS Journal 279 (2012) 251–263 ª2011 The Authors Journal compilation ª2011 FEBS 251
The Eph receptors constitute the largest family of
receptor protein tyrosine kinases, and interact with their
ephrin ligands to form a bidirectional, cell-to-cell signal-
ing communication system designated as ‘forward’
receptor signaling and ‘reverse’ ephrin signaling. The
Eph receptors and ephrins are divided into two subclass-
es (A and B), which are encoded by nine EphA, five
EphB, five ephrinA and three ephrinB genes in the
human genome. EphrinA ligands are tethered to the
outer leaflet of the plasma membrane by a glycosyl-
phosphatidylinositol anchor, and ephrinB ligands pos-
sess a transmembrane domain and a short cytoplasmic
tail [5,6]. Intercellular Eph–ephrin signaling is funda-
mentally involved in developmental processes that
require organized patterning and movement of cells,
especially in the development of the central nervous sys-
tem [7–9] and in the remodeling of blood vessels [10–12].
Eph–ephrin signaling is also essential for the correct
formation of crypts and villi in the intestinal epithelium
[13–15]. Several lines of evidence also suggest that
Eph–ephrin signaling is involved in tumorigenesis,
invasiveness, metastasis, and angiogenesis [10,16–18].
EphrinA5 was recently identified as a tumor suppres-
sor that is significantly downregulated in several types of
human carcinoma [19–23]; however, ephrinA5 was also
reported to be overexpressed in breast cancer carcinoma
cell lines [24], and was associated with poor prognosis in
both ovarian cancer [25] and pancreatic ductal adenocar-
cinoma [26]. Recently, ephrinA5 has been shown to regu-
late EGFR in gliomas by promoting c-Cbl-mediated
EGFR degradation [21]. Dysregulated EGFR signaling
in various types of malignancy, either through overex-
pression or constitutive activation, promotes cancer cell
proliferation, survival, invasiveness, metastasis, and
angiogenesis. Misregulation of EGFR signaling is asso-
ciated with poor prognosis in many human malignan-
cies, and several therapeutic remedies have been
developed to suppress EGFR signaling [27–33].
In this study, we investigated the association
between ephrinA5 and colon cancer progression. Our
data identified mutually exclusive expression profiles of
ephrinA5 and EGFR in colon cancer tissues. We also
found that ephrinA5 can suppress EGFR function
through c-Cbl-mediated EGFR ubiquitination and
degradation, and reduce colon cancer cell proliferation,
migration, and chemotherapeutic resistance.
Results
EphrinA5 is downregulated in colon cancer
In order to elucidate the biological significance of
ephrinA5 in the progression of colon cancer, we
analyzed ephrinA5 expression levels with immunohis-
tochemistry on a 96-spot tissue microarray slide com-
posed of 10 normal colon tissues, 14 benign colon
lesions (adenomas and polyps), and 72 colon malig-
nancies (68 colon adenocarcinomas and four non-
Hodgkin lymphomas). EphrinA5 expression levels were
significantly reduced in cancer tissue as compared with
normal tissue or benign lesions (P< 0.001). Although
ephrinA5 was slightly downregulated in benign lesions,
there was no significant difference between normal
colon tissues and benign lesions (Table 1). Figure 1A
is representative of immunohistochemical staining for
ephrinA5 and EGFR in a colon adenocarcinoma. In
contrast, EGFR was significantly upregulated in colon
cancer as compared with normal tissue or benign
lesions (P= 0.001). We also examined the mRNA
expression levels of ephrinA5 in 14 paired frozen colon
adenocarcinoma specimens. As shown in Fig. 1B,
ephrinA5 mRNA levels were significantly reduced by
up to 50–90% in tumor tissues as compared with
normal colon tissue (P= 0.04).
EphrinA5 downregulation and EGFR upregulation
are correlated with tumor differentiation and
clinical stage
To investigate the clinical significance of ephrinA5
downregulation, we analyzed the association between
ephrinA5 and the clinicopathological parameters of the
10 normal colon tissues, 14 benign colon lesions (ade-
nomas and polyps), and 72 colon cancer samples. Uni-
variate and multivariate analyses were performed with
Cox’s hazard regression model (Table 1). EphrinA5
downregulation, EGFR upregulation and older age
were positively associated with colon cancers; however,
only ephrinA5 and age were identified as independent
risk factors by the multivariate regression model. This
indicated that ephrinA5 was significantly downregulat-
ed during the genesis of colon cancer. EGFR levels,
however, were not statistically significant according
to multivariate analysis, indicating that EGFR may
interact with other risk factors during the process of
carcinogenesis.
To determine the role of ephrinA5 downregulation in
the process of cancer progression, we further analyzed
the correlation between ephrinA5 and cancer prognostic
factors in 68 colon adenocarcinomas, including 24
grade 1, 26 grade 2 and six grade 3 adenocarcinomas.
Of these tumors, 38 were at stage I, four were at
stage II, and 26 were at stage III. EphrinA5 downregu-
lation was significantly correlated with higher stage
(P= 0.047) and poorer differentiation (P= 0.012) by
Cox regression model analysis (Table 2). Additionally,
EphrinA5 represses EGFR in colon cancer T.-H. Wang et al.
252 FEBS Journal 279 (2012) 251–263 ª2011 The Authors Journal compilation ª2011 FEBS
Table 1. Association between ephrinA5, EGFR expression and clinicopathological parameters. CI, confidence interval; HR, hazard ratio; IHC, immunohistochemistry. P-values: benign
lesion or colon cancer versus normal. The statistical significance (P< 0.05) is shown in bold.
Subject
characteristics Normal (n= 10) Benign lesion (n= 14)
Univariate
a
Multivariate
a
Cancer (n= 72)
Univariate
b
Multivariate
b
HR
(95% CI) P
HR
(95% CI) P
HR
(95% CI) P
HR
(95% CI) P
Age
(years) (mean)
49.6 ± 18.5 41.1 ± 17.1 0.97
(0.92–1.02)
0.242 0.98
(0.92–1.03)
0.381 60.3 ± 13.2 1.05
(1.01–
1.10)
0.028 1.070
(1.01–
1.13)
0.019
Sex Male (8) Female (2) Male (10) Female (4) 1.60
(0.23–11.08)
0.634 1.43
(0.19–10.99)
0.734 Male (36) Female
(36)
4.00
(0.79–
20.15)
0.093 4.25
(0.72–
25.27)
0.111
EphrinA5
IHC intensity
Positive
(2+, 3+)
Negative
(0, 1+)
Positive
(2+, 3+)
Negative
(0, 1+)
Positive
(2+, 3+)
Negative
(0, 1+)
8 2 6 8 5.33
(0.82–34.83)
0.080 4.00
(0.56–28.46)
0.166 24 48 8.00
(1.575–
40.63)
0.012 6.09
(1.02–
36.26)
0.047
EGFR IHC
intensity
Negative
(0, 1+)
Positive
(2+, 3+)
Negative
(0, 1+)
Positive
(2+, 3+)
8 2 8 6 3.00
(0.46–19.59)
0.251 2.28
(0.31–16.72)
0.418 28 44 6.29
(1.24–
31.78)
0.026 3.902
(0.66–
22.95)
0.132
a
Cancer versus normal.
b
IHC data were analyzed as categorized variables on the basis of Cox’s hazard regression model. Comparisons were performed among three groups: (a) colon
cancer (n= 72); (b) benign lesions (adenomas ⁄polyps) (n= 14); (c) normal (n= 10).
T.-H. Wang et al. EphrinA5 represses EGFR in colon cancer
FEBS Journal 279 (2012) 251–263 ª2011 The Authors Journal compilation ª2011 FEBS 253
most (24 ⁄26) of the stage III samples exhibited reduced
ephrinA5 expression as compared with stage I + II
samples (23 ⁄42). As expected, ephrinA5 levels were not
associated with patient age or gender (Table 2). EGFR
overexpression was not observed to be associated with
colon cancer differentiation (P= 0.206) or clinical
stage (P= 0.756) in these samples (Cox’s regression
model).
EphrinA5 reduces EGFR protein expression and
phosphorylation
EphrinA5 has been reported to be downregulated in
glioma cells and to function as a tumor suppressor by
negatively regulating EGFR expression [21]. To deter-
mine whether a similar regulatory mechanism between
ephrinA5 and EGFR exists in colon cancer, we examined
the protein expression levels of ephrinA5 and EGFR with
immunohistochemistry. As shown in Fig. 1A, ephrinA5
was highly expressed in normal colon tissues and down-
regulated in cancerous lesions. In the same regions,
EGFR was observed to be overexpressed. We found
there were 45.5% (31 ⁄68) colon cancer specimens exhib-
iting a mutually exclusive expression pattern of ephrinA5
and EGFR. Additionally, ephrinA5 expression was also
significantly negatively correlated with EGFR levels in
those samples (Spearman’s rho correlation = )0.367,
P< 0.0001). We therefore utilized a gain-of-function
A
EphrinA5 EGFR
B
10
1
0.1
EphrinA5 expression in tumor as compared with normal tissue
0.01 P = 0.04 (one-sample t-test)
Fig. 1. (A) Representative immunohistochemical results for ephrinA5 and EGFR from tissue microarray analysis of colon cancer. EphrinA5
and EGFR were expressed with mutually exclusive distributions within colon cancer tissue and normal colon tissue, respectively. White
arrows indicate normal colonic mucosa. (B) Relative ephrinA5 mRNA levels of 14 pairs of colon adenocarcinoma tissues normalized with
GAPDH. EphrinA5 was also transcriptionally downregulated in most colon cancer specimens to a greater extent than in normal colon tissues;
only sample CO001 had modestly higher ephrinA5 expression in tumor specimens.
EphrinA5 represses EGFR in colon cancer T.-H. Wang et al.
254 FEBS Journal 279 (2012) 251–263 ª2011 The Authors Journal compilation ª2011 FEBS
and a loss-of-function experimental strategy to verify the
clinical observation of ephrinA5–EGFR interaction
within colon cancer cell line models. Human ephrinA5
was cloned into the pIRESneo mammalian expression
vector for transfection of ectopic human ephrinA5 into
the colon cancer cell lines. Forced ectopic ephrinA5
expression significantly reduced endogenous EGFR pro-
tein and phosphorylation levels in cancer cell lines as
compared with a vector control, as detected by both wes-
tern blot and immunofluorescence staining (Fig. 2A,B).
We further examined the influence of ephrinA5 on
EGFR expression at the transcriptional level. Forced
ectopic ephrinA5 expression did not alter endogenous
EGFR mRNA expression (Fig. 2C). We also observed
that ephrinA5 knockdown by specific small interfering
RNA (siRNA) did not influence EGFR mRNA expres-
sion in colon cancer cell lines (Fig. 2C). These results
indicate that ephrinA5 downregulates EGFR expression
at the protein level and not at the mRNA level.
EphrinA5 reduces EGFR expression by promoting
c-Cbl-mediated EGFR degradation
The E3 ubiquitin ligase c-Cbl is required for the ubiqu-
itin-dependent degradation of EGFR, and ephrinA5
may enhance c-Cbl binding to EGFR to promote
ubiquitination and degradation of the receptor. We
therefore examined the regulatory effects of ephrinA5
on c-Cbl in colon cancer cell lines. Forced ectopic eph-
rinA5 expression resulted in a greater than two-thirds
reduction in endogenous EGFR protein levels. In the
reverse experiment, c-Cbl knockdown by siRNA signif-
icantly rescued EGFR protein levels (Fig. 3A,B). Eph-
rinA5 overexpression did not alter c-Cbl mRNA levels
(Fig. 3C,D), indicating that ephrinA5 promotes c-Cbl
binding to EGFR to facilitate EGFR ubiquitination
and degradation in colon adenocarcinomas.
EphrinA5 inhibits cancer cell proliferation and
migration
EGFR overexpression was observed in several types of
cancer and was associated with cancer cell growth. We
also found that EGFR expression was significantly
increased in colon cancers. EphrinA5 reduced EGFR
protein levels by facilitating its degradation, indicating
that ephrinA5 may function to repress cancer cell
growth. Here, we examined the inhibitory effects of
ephrinA5 on colon cancer cells. First, we analyzed the
effect of ephrinA5 expression on colon cancer cell
growth with the acid phosphatase (ACP) assay. Forced
expression of ephrinA5 significantly reduced cellular
proliferation in the colon cancer cell lines SW480 and
WiDr (Fig. 4A). We observed similar suppressive
effects in HEK cells (data not shown). Ectopic expres-
sion of ephrinA5 also increased doxorubicin sensitivity
in SW480 cells, and flow cytometric analysis revealed a
sub-G
1
increment (Fig. 4B). A reduction in the expres-
sion of ephrinA5 could therefore increase cancer cell
survival under conditions of cytotoxic stress.
Next, we analyzed the expression of EphB2, EphA2,
EphA3, and EphA5, the preferred receptors for ephri-
nA5, in the SW480 and WiDr cell lines. Expression of
EphB2, EphA2, and EphA3, but not of EphA5, was
detected in both cell lines. EphA2 expression was also
found in all tumor and paratumoral colon tissues
(Fig. 4C). EphB2 and EphA3 were present in the major-
ity of colon cancers. This suggests that ephrinA5 is able
to exhibit its suppressive effect through its Eph receptors
in colon cancer.
Table 2. Association between ephrinA5 expression and cancer prognostic factors (for carcinoma patients only). On the basis of Cox regres-
sion model analysis, the statistical significance (P< 0.05) is shown in bold. CI, confidence interval; HR, hazard ratio; IHC, immunohistochem-
istry.
EphrinA5 IHC intensity
Negative (0, 1+)
(n= 47)
Positive (2+, 3+)
(n= 21)
Univariate Multivariate
HR (95% CI) PHR (95% CI) P
Age (years) (mean) 60.66 ± 14.03 60.71 ± 12.31 1.000 (0.962–1.039) 0.988 1.011 (0.959–1.066) 0.680
Sex
Male 22 10 1 (ref.)
Female 25 11 1.033 (0.369–2.895) 0.951 3.858 (0.655–22.720) 0.136
Grade
a
I 12 12 1 (ref.)
II + III 25 7 1.890 (1.059–3.373) 0.031 3.006 (1.274–7.091) 0.012
Stage
I + II 23 19 1 (ref.) –
III 24 2 3.148 (1.440–6.886) 0.004 2.438 (1.012–5.871) 0.047
a
There were 12 mucinous adenocarcinomas without grade records.
T.-H. Wang et al. EphrinA5 represses EGFR in colon cancer
FEBS Journal 279 (2012) 251–263 ª2011 The Authors Journal compilation ª2011 FEBS 255