Experimental Cell Research 361 (2017) 163–169

Contents lists available at ScienceDirect

Experimental Cell Research

journal homepage: www.elsevier.com/locate/yexcr

CCN1 sensitizes esophageal cancer cells to TRAIL-mediated apoptosis T

a b a a b a,b,c,⁎
Tong Dang , Cristina Modak , Xiemei Meng , Jinbao Wu , Reinier Narvaez , Jianyuan Chai
a
The Second Affiliated Hospital of Baotou Medical College, Inner Mongolia University of Science and Technology, 30 Hudemulin Rd, Baotou 014030, China
b
Laboratory of Gastrointestinal Injury and Cancer, VA Long Beach Healthcare System, Long Beach, CA 90822, USA
c
College of Medicine, University of California Irvine, CA 92697, USA

A R T I C L E I N F O A B S T R A C T

Keywords: TRAIL is one of the best anti-cancer molecules in our body. It kills a variety of cancer cells that are resistant to
TRAIL conventional chemotherapy, without causing much negative impact on normal cells, because its death receptors
CCN1 are almost exclusively found on cancer cells. However, some cancer cells are not sensitive to TRAIL treatment,
Apoptosis even though they express its death receptors. A second molecule is needed to help TRAIL to complete its mission.
Esophageal cancer
Finding such molecules now becomes a top priority in cancer research. Our study shows that CCN1 is such a
Receptor
molecule. CCN1 was highly expressed in the esophageal epithelium of the patients suffering from gastro-
esophageal reflux disease, but faded away as the situation worsened towards adenocarcinoma. Treating the
tumor cells with CCN1 resulted in apoptosis, while the same treatment to the normal cells only nourished cell
growth. It was TRAIL that mediated this process. Apparently, CCN1 altered the expression profile of TRAIL and
its receptors in tumor cells, namely, activating TRAIL and its death receptors and shutting down its decoy
receptors. CCN1 and TRAIL worked as a team to put the cancer cells to death, as elimination of either one failed
apoptosis.

1. Introduction squamous epithelium to intestinal columnar phenotype, known as

Esophageal cancer is one of the most common and deadliest ma-
lignancies in the world. It is globally ranked as number nine by its
prevalence and number six by its mortality. For example, in 2013, there
were 442,000 new cases of esophageal cancer diagnosed worldwide,
while the related death was as high as 440,000 [1]. There are two types
of esophageal cancer commonly seen, esophageal squamous cell carci-
noma (ESCC) and esophageal adenocarcinoma (EAC). The former has
been the dominant one historically. However, recent literature [1,2]
indicates that EAC has been rising by six fold annually in western
countries, and now it becomes the fastest growing cancer in the world.
A similar trend is developing in many other countries as well. The
possible reason can be several, but the leading cause has been attrib-
uted to gastroesophageal reflux disease (GERD). A recent study showed
that GERD increases the risk of EAC by 8.6 fold [3].
GERD occurs when the esophageal sphincter weakens and allows
stomach acid (often mixed with duodenal contents) to back up into the
esophagus. The acidic refluxate erodes the epithelial lining of the lower
esophagus and causes esophagitis. Overtime, these repeated episodes
can lead to a mucosal metaplastic transformation from esophageal
Barrett's Esophagus (BE). This change is supposed to be a self-defense
mechanism against the acidic insult, because the columnar epithelium
is more tolerant to acid erosion than the squamous mucosa. However,
this metaplasia confers an increased risk of malignancy. According to
recent literature [4,5], BE patients can have as high as 400-fold more
likelihood to develop EAC than normal people.
The rise of GERD is largely associated with the fast growing obese/
overweight population. The excess fat in the abdominal area puts a
constant pressure on the stomach, and creates a high frequency of
esophageal acid exposure. A recent study showed that global obesity
rates have doubled since 1980 [6]. It is predicted that by the year 2020,
77.6% of American men and 71.1% of American women will be over-
weight [7]. A similar trend is also seen in China, the most populated
state. According to the statistics in 2013 [8], at least 46 million of
Chinese adults were suffering from obesity, and another 300 million
were considered as overweight. Therefore, the health issues associated
with body weight, like EAC, will become one of the top global concerns
in the near future.
There has never been a special drug to treat EAC. Current GERD
treatment primarily relies on acid suppressive medications and repair

Abbreviations: TRAIL, Tumor necrosis factor-Related Apoptosis-Inducing Ligand; EAC, esophageal adenocarcinoma; BE, Barrett's Esophagus; GERD, gastroesophageal reflux disease;
TNF, tumor necrosis factor; FADD, Fas-Associated protein with Death Domain
⁎
Corresponding author at: College of Medicine, University of California Irvine, CA 92697, USA.
E-mail address: Jianyuan.chai@gmail.com (J. Chai).

http://dx.doi.org/10.1016/j.yexcr.2017.10.015
Received 7 September 2017; Received in revised form 11 October 2017; Accepted 17 October 2017
Available online 18 October 2017
0014-4827/ © 2017 Elsevier Inc. All rights reserved.
T. Dang et al.
surgery, but neither is a clear winner thus far. In fact, increasing evi-
dence shows multiple side-effects associated with this line of drugs,
such as decreased absorption of vitamins/minerals [9], susceptibility to
infections [10], bone fracture [11], and even elevated risk of devel-
oping cancer [12]. For these reasons, the Food and Drug Administration
of the United States has repeatedly issued warnings on the use of acid
suppressive drugs. For people who have responded to medication but
continue to experience GERD symptoms, surgery to reconstruct the
esophageal sphincter is usually an option. However, only 5% of GERD
patients undergo surgery and follow-up study found that almost two-
thirds of the surgery patients were back on medication [13].
A good anti-cancer drug should not only be able to kill cancer cells,
but also be harmless to the normal tissue. TRAIL (Tumor necrosis
factor-Related Apoptosis-Inducing Ligand) is such an example. As a
member of TNF family, TRAIL can selectively kill a variety of cancer
cells that are resistant to conventional chemotherapy while leaving
normal cells unharmed, because its death receptors are almost ex-
clusively found on cancer cells [14]. When an adequate amount of
TRAIL is in present, cancer cells will usually go into apoptosis. How-
ever, more and more studies found that some cancer cells are not sen-
sitive to TRAIL treatment, despite expression of the death receptors, and
require a second agent to precondition them in order for TRAIL to do its
job [15,16]. Therefore, finding the right sensitizers for TRAIL now be-
comes a top priority in cancer research. Based on our study, CCN1
appears to be such a molecule, at least for EAC.
CCN1 is a matricellular protein that essentially supports cell adhe-
sion, migration and survival. It is one of the targets of serum response
factor (SRF), a transcription factor that we have been studying for
nearly two decades [17,18]. Cells attaching to CCN1 usually activate
cyto-protective pathways against apoptosis. However, a number of
studies have revealed the other edge of this molecule [19–21], i.e.,
triggering apoptosis sometimes. The mechanisms are still not fully un-
derstood, but recent data have drawn a line to TNF cytokines. It was
shown that incubation of fibroblasts with either CCN1 or TNFα alone
promoted cell proliferation, but in the presence of both, cells underwent
apoptosis [22]. Using CCN1 with other TNF molecules like FasL or
TRAIL produced a similar effect [19,23]. In prostate cancer cells, CCN1
was found to be able to do both protective and destructive work, all
depending on the availability of TRAIL [23]. Yet, these cells express
CCN1 naturally and therefore, do not need a sensitizer to engage in
TRAIL-mediated apoptosis.
The TNF superfamily now includes 19 identified members which
function through one or more of the 29 receptors. Some of the receptors
(e.g. Fas, TRAILR1) contain death domains in their cytoplasmic tails,
which can trigger apoptosis upon ligand binding, while others (e.g.
TRAILR3, TRAILR4) do not have such a structure and therefore cannot
provide intracellular signals for apoptosis. However, this second group
of receptors can effectively compete with the first group for their cor-
responding ligands and abort the apoptotic effort of the first group, and
for this reason, they are often referred to as the decoy receptors in
contrast to the first group, the death receptors.
2. Materials and methods
2.1. Cell culture, transfection and treatment
Human EAC cells OE33, OE19 and FLO-1 (Sigma-Aldrich, St Louis)
were cultured in RPMI medium plus 10% fetal bovine serum, while
normal human esophageal epithelial cells Het1A (American Type
Culture Collection, Manassas) were maintained in culture in KBM-2
medium with supplements (Lonza, Walkersville).
The following plasmids were used for transfection: pCMV6 carrying
one of the following open reading frames: CCN1, TRAIL, TRAILR2,
TRAILR3, TRAILR4, and OPG; and pRS with shRNA against each one of
them. pRS with negative control shRNA or pCMV6 without an insert
were used as controls (Origene, Rockville). All cell transfections were
164
Experimental Cell Research 361 (2017) 163–169
done in Lipofectamine-2000 (Invitrogen, Carlsbad) and selected using
Neomycin or Puromycin.
For cell treatment, cells were plated either in Petri-dishes (for RNA
or protein isolation), or on collagen-coated coverslips (for staining).
Cells were synchronized and treated accordingly with recombinant
human CCN1, or TRAIL (Peprotech, Rocky Hill) or combinations as
indicated.
2.2. Apoptosis assays
Cells were cultured on coverslips and treated as indicated.
Following the manufacturer's protocol, cells were then incubated in a
binding buffer containing FITC-conjugated Annexin V and propidium
iodide (Abcam, Cambridge) for 5 min. The number of apoptotic cells in
5 microscopic fields from each of 3 coverslips per treatment was re-
corded under a Nikon fluorescence microscope (n=15). The apoptotic
index (%) was calculated by dividing the number of apoptotic cells by
the total number of cells in control (verified by DAPI staining) under
the same magnification.
2.3. Real-time RT-PCR and array analysis
Total RNA was extracted using RNeasy kit (Qiagen, Valencia) fol-
lowing the manufacturer's protocol. Reverse transcription was done
following the procedure: 25 °C/10 min – 55 °C/30 min – 85 °C/5 min –
4 °C/∞. 96-well plates of human gastroesophageal cancer tissue cDNA
array, apoptosis array and TNF array were purchased from Origene.
Real-time PCR was performed following the two-step program using
SYBR Green master mix (SABiosciences, Frederick). Data were gener-
ated from at least 5 independent experiments and analyzed according to
the ΔΔCt method. Briefly, ΔCt was calculated by subtracting the Ct value
of GAPDH from the Ct value for each gene; and then ΔΔCt was calcu-
lated by subtracting the ΔCt of the control from the ΔCt of the treat-
ment; and finally the fold change was calculated using the formula:
Fold Change = 2(-ΔΔCt).
2.4. Western blot analysis
Western blot analyses were done as described in our earlier studies
[24,25]. Antibodies included CCN1 and GAPDH (Santa Cruz Bio-
technology, Santa Cruz). Signals were quantified with the TotalLab
TL100 software based on at least five replicates.
2.5. Immunofluorescence and immunohistochemistry
Cells were cultured till desired confluence on coverslips that have
been pre-coated with type I collagen. After serum starvation and in-
dicated treatment, cells were fixed for 10 min in 4% paraformaldehyde.
After incubation with a primary antibody for 2 h, cells were washed in
PBS and incubated with a FITC-conjugated secondary antibody
(Abcam) for an hour. Nuclei were counter-stained with Propidium
Iodide (Invitrogen).
For immunohistochemistry, tissue specimens were fixed in 10%
Formalin and embedded in paraffin. Tissue sections were de-paraffi-
nized, re-hydrated, and retrieved in a microwave oven by pause-
heating. Slides were incubated with the primary antibody of interest for
2 h in a humidified chamber. After 3 × 10 min wash in PBS, signals
were detected using a LSAB+ kit (Agilent) containing a universal sec-
ondary linker and HRP-conjugated streptavidin with AEC chromogen as
substrate. Nuclei were counter-stained with hematoxylin.
2.6. Statistical analysis
Numerical data were analyzed by single classification one-way
ANOVA and P < 0.05 was considered as significant.
T. Dang et al.
3. Results
3.1. CCN1 up in esophagitis and then down as esophageal malignancy
advances
CCN1 essentially supports normal cell survival, but in cancer cells it
acts like a double-edged sword. Its elevation has been found in breast
cancer [26], ovarian carcinoma [27] and pancreatic carcinoma [28],
where it facilitates tumor cell proliferation and metastasis; while in
others like hepatocyte carcinoma [29], endometrial adenocarcinoma
[30], and non-small-cell lung cancer [31], CCN1 is either down-
regulated or completely vanished. Forced expression of CCN1 in latter
cells causes apoptosis.
Based on our previous study that CCN1 highly elevated in ESCC
[32], we were wondering how CCN1 was expressed in EAC. So we
examined CCN1 expression at different stages of EAC development by
immunostaining human tissue slides. CCN1 was found heavily ex-
pressed in GERD patients with esophagitis (Fig. 1B), in contrast to the
normal tissue (Fig. 1A). Esophagitis was histologically characterized
with markedly thickened epithelium, elongation of the lamina propria
papillae into the epithelium, and basal cell hyperplasia. When the
normal esophageal squamous epithelium transformed into intestinal
columnar phenotype (BE), CCN1 expression was still higher (Fig. 1C)
than its normal level, but a lot less compared to it in the earlier in-
flammatory phase (Fig. 1B). BE was distinguished by the appearance of
goblet cells in the esophageal epithelium. However, once the situation
turned into EAC, CCN1 became barely detectable in the transformed
cells (Fig. 1D).
At the time, we obtained a gastroesophageal cancer tissue array
which contains cDNA products derived from 13 EAC patients, 3 healthy
individuals, and the rest were either from ESCC or from gastric tissue.
165
Experimental Cell Research 361 (2017) 163–169
Fig. 1. CCN1 expression is down in EAC (im-
munohistochemistry and Western blotting). A.
normal esophageal epithelium. B. esophagitis. C. BE.
D. EAC. Magnification 160x. E. Western blot analysis
of CCN1 expression in esophageal tissues including
normal (1), esophagitis (2), BE (3) and EAC (4). F.
Western blot analysis of CCN1 expression in cell lines
including Het1A (1), FLO-1 (2), OE19 (3) and OE33
(4). GAPDH was used as a loading control. (in color).
In consistent with the results obtained from immunohistochemistry,
tissue array analysis showed that CCN1 mRNA expression was down in
EAC tissue by 61% (N = 13 × 3, P < 0.05) compared to its level in a
healthy esophagus (N = 3 × 3). A similar downregulation was found in
three commonly used EAC cell lines as well, including OE33, OE19 and
FLO-1, where CCN1 was averagely lower by 65% (N = 17, P < 0.01)
than it in normal esophageal epithelial cell lines such as Het1A (N =
35). Western blot analyses generated similar results (Figs. 1E and 1F).
All of these data made us think that the presence of CCN1 in eso-
phageal epithelium might be hostile to the survival of EAC tumor cells.
3.2. CCN1 induces apoptosis in esophageal adenocarcinoma
To find out whether CCN1 was really an enemy of the EAC cells, we
seeded equal number of tumor cell OE33 and normal cell Het1A on the
coverslips. After synchronization, cells were incubated with either BSA
(control) or recombinant human CCN1 protein at 1 μg/ml for 6 h.
Apoptosis was examined by Annexin V FITC staining. As shown in
Fig. 2, a great number of OE33 cells underwent apoptosis in the pre-
sence of CCN1 (78% in CCN1-treated vs. 6% in control, P < 0.01),
while no significant change was noticed in Het1A cells with or without
CCN1, supporting our original speculation that CCN1 can kill EAC cells
without causing recognizable damage to the normal cells. Apoptotic
cells were characterized as a green fluorescent membrane staining
(early stage) or a red fluorescent nuclear staining (later stage).
To elucidate the apoptotic pathways involved in this event, we
treated both OE33 and Het1A cells with the recombinant CCN1 for 2 h
at the same dosage as mentioned above. Total RNA was extracted from
the cells to probe a RT_PCR array containing 54 pro-apoptotic and 37
anti-apoptotic genes. A clear distinction was revealed between these
two types of cells. In OE33 cells, CCN1 treatment downregulated 9 anti-
T. Dang et al. Experimental Cell Research 361 (2017) 163–169

Fig. 2. CCN1 induces apoptosis in EAC cells (Annexin V binding

assay), shuts down TRAILR4 expression in EAC cells but activates
it in normal cells (immunofluorescence). Equal number of OE33
and Het1A cells were seeded on the coverslips and treated with
CCN1 for 6 h. A-D, apoptosis was assayed in the presence and
absence of CCN1. Control of OE33 (A) and Het1A (B); CCN1
treatment in OE33 (C) and Het1A (D). E-H, cells were stained with
a rabbit antibody against TRAILR4 and a FITC-conjugated mouse-
anti-rabbit secondary antibody. Nuclei were counterstained with
propidium iodide. Control of OE33 (E) and Het1A (G); CCN1
treatment in OE33 (F) and Het1A (H). (in color).

apoptotic genes by > 50%, and at meantime, upregulated 15 pro-

Table 1
apoptotic genes by > 2 fold; in contrast, the same treatment in Het1A
The top 5 genes upregulated by CCN1 in EAC cells (OE33) versus normal esophageal cells only activated one pro-apoptotic gene (CDKN2A), the remaining
epithelial cells (Het1A), based on apoptosis array analysis. The numbers within par- effect was the upregulation of 9 anti-apoptotic genes by > 2 fold.
enthesis represent fold change. Table 1 listed the top 5 genes upregulated by CCN1 in OE33 versus
Het1A cells, along with their known roles related to apoptosis. The most
OE33 Apoptosis Het1A apoptosis
prominent effect in CCN1-treated Het1A cells was the huge induction of
ERN2 (60) Pro BIRC5 (137) Anti BIRC5, the gene coding for the anti-apoptotic protein Survivin. In OE33
TRADD (30) Pro HIPK3 (30) Anti cells, on the other hand, the top activated genes by CCN1 were mostly
BAD (20) Pro IRF4 (24) Anti
related to the apoptotic activity of TNF family, such as TRADD, FADD,
FADD (7) Pro TRAF1 (7) Anti
TP53 (5) Pro CDKN2A (5) Pro BAD, and TP53, among which, FADD is the common link between the
death receptors of TNF cytokines and caspase-8/10 activation.

166
T. Dang et al. Experimental Cell Research 361 (2017) 163–169

Table 2 induced TRAIL expression, but also activated TRAILR2, the second
The effects of CCN1 transfections on the expression of BIRC5 and TP53 in EAC cells death receptor, pushing the cancer cells more towards death. In addi-
(OE33) versus normal esophageal epithelial cells (Het1A), based on apoptosis array
tion, all the three decoy receptors were shut down in the presence of
analysis. The numbers represent relative levels of the gene after CCN1 manipulation in
comparison to its normal expression. CCN1, meaning that the EAC cells doomed to death. In Het1A cells, on
the other hand, CCN1 treatment stopped the expression of TRAIL li-
Cell lines CCN1 BIRC5 TP53 gand, but activated two of the three decoy receptors, TRAILR4 and
OPG, while the two death receptors remained in silence, that is to say,
OE33 1.00 1.00 1.00
OE33CCN1- 0.04 3.13 × 107 0.18 the normal cells were completely immune to TRAIL apoptotic attacks as
OE33CCN1+ 29.86 0.15 3.06 × 104 long as CCN1 is around. To verify this result, we stained all three types
Het1A 1.00 1.00 1.00 of AEC cells and Het1A cells before and after CCN1 treatment for all the
Het1ACCN1- 0.09 0.26 2.10 × 103 TRAIL receptors. In agreement with the RT_PCR outcomes, the decoy
Het1ACCN1+ 64.00 6.81 × 106 0.93
receptors were found activated in Het1A cells but shut down in EAC
cells in the presence of CCN1. Fig. 2E-H show the immunofluorescence
To verify the results obtained from using recombinant CCN1, we staining for TRAILR4.
transfected the cells with plasmids carrying either CCN1 open reading To follow this trend of thought, next, we wanted to know how im-
frame (CCN1+) or shRNA against CCN1 (CCN1-), and then assessed portant TRAIL was to CCN1-induced EAC cell apoptosis. We incubated
apoptosis using the same array analysis. Overexpression of CCN1 both OE33 and Het1A cells on coverslips for 6 h with one of the fol-
caused a similar effect in gene expression as using recombinant CCN1 lowing combinations: (1) recombinant TRAIL at 30 ng/ml, (2) re-
protein, while knockdown of CCN1 promoted survival in OE33 cells but combinant TRAIL at 30 ng/ml plus recombinant CCN1 at 1 μg/ml, (3)
suppressed it in Het1A cells. Table 2 shows the effects on two re- recombinant CCN1 alone at 1 μg/ml, (4) recombinant CCN1 at 1 μg/ml
presentative genes, BIRC5 and TP53. plus TRAIL antibody at 50 μg/ml, or (5) BSA. Apoptosis was then as-
sayed using Annexin V. As shown in Fig. 3A, using recombinant TRAIL
alone did not cause a significant change in terms of apoptosis compared
3.3. CCN1 requires TRAIL to induce apoptosis in EAC cells to control (BSA), but addition of CCN1 to this experiment (TRAIL+C-
CN1) raised the number of apoptotic OE33 cells by 7 fold (P < 0.01),
As mentioned above, CCN1 is commonly known for supporting cell indicating that, TRAIL cannot kill EAC cells without the presence of
survival, therefore, its impact on Het1A cells was expected. However, CCN1. Using recombinant CCN1 alone, on the other hand, induced
its apoptotic effect on OE33 cells made us think that EAC cells might apoptosis at almost the same magnitude as TRAIL and CCN1 combi-
express something that could convert CCN1 from pro-survival to pro- nation, suggesting that extrinsic TRAIL was unnecessary for CCN1 to
death. According to Lau's experiments [19–23], who discovered CCN1, induce EAC cell death. However, when we added TRAIL antibody to
CCN1 cannot cause apoptosis in its own, and it requires the presence of keep any available TRAIL neutralized (CCN1+TRAIL.Ab), CCN1 lost all
TNF cytokines to become a cell killer. In his experiment, combining its apoptotic power, indicating that CCN1-induced TRAIL expression is
500 ng/ml CCN1 with 100 ng/ml FASLG doubled the number of an essential requirement for EAC cell apoptosis.
apoptotic cells caused by FASLG alone; while increasing CCN1 to 2 μg/
ml tripled the effect. Lau's work led us to speculate that EAC cells might 3.4. The shutdown of TRAIL decoy receptors is the key for CCN1-induced
have a different expression profile of TNF and/or TNFR family members EAC cell apoptosis
that makes them more sensitive to CCN1 apoptotic potential.
To test this hypothesis, at first, we screened these two families using Our previous experiments indicated that CCN1 and TRAIL need
a RT_PCR array containing gene sequences for 18 TNF ligands and 27 each other to kill EAC cells, as eliminating either one failed to cause
their receptors. EAC cells were found to express more of these genes apoptosis. Our next question was why TRAIL itself could not kill the
than the normal cells, namely, 7 ligands and 16 receptors in OE33 cells, cancer cells even when all its death receptors were available. We
8 ligands and 16 receptors in OE19 cells, 8 ligands and 15 receptors in thought that the expressions of TRAIL decoy receptors might be the
FLO-1 cells, versus 6 ligands and 8 receptors in Het1A cells, suggesting reason, because the decoy receptors have a dominant effect over the
that EAC cells were more vulnerable to TNF apoptotic attacks than the death receptors; they can form either homo-oligomers with themselves
normal cells. What even more interesting was the expressions of TRAIL or hetero-oligomers with the death receptors to disable the apoptotic
and its receptors: OE33 cells expressed four out of five TRAIL receptors, power of TRAIL.
including the death receptor TRAILR1 and three decoy receptors To test this hypothesis, we evaluated the contribution of each in-
TRAILR3, TRAILR4 and OPG, but not TRAIL ligand; while Het1A cells dividual TRAIL receptor to CCN1-induced apoptosis. At first, we
expressed TRAIL ligand but none of its receptors (Table 3). These results transfected OE33 cells with various plasmid combinations to establish
prompted us to shift our attention immediately to TRAIL and its re- cell lines with the following expression profiles: (1) OE33TRAILR2+, (2)
ceptors. OE33TRAILR3-, (3) OE33TRAILR4-, (4) OE33OPG-, (5) OE33DCR-, (6)
Next, we probed the same type of array using RNA extracts after OE33TRAILR2+/DCR-, and (7) OE33. “DCR-” represents shutdown of all
CCN1 treatment. As shown in Table 3, in OE33 cells, CCN1 not only three decoy receptors. The knockdowns were created using shRNA
vectors and selected for at least 80% drop in the expression of the
Table 3 targeted gene. All the cells were then incubated with recombinant
The effect of CCN1 on the expression of TRAIL and its receptors in EAC cells (OE33)
versus normal esophageal epithelial cells (Het1A), based on TNF/TNFR array analysis. In
TRAIL at 30 ng/ml for 6 h. Apoptosis was examined using Annexin V.
controls, OE33 and Het1A, expression levels were all set to be 1 arbitrarily, the values in As shown in Fig. 3B, expression of TRAILR2 without changing anything
CCN1-treated groups are fold changes (P < 0.01) relative to their corresponding controls. else (OE33TRAILR2+) approximately doubled the rate of apoptosis
caused by TRAIL alone, but the total number of apoptotic cells was still
Gene Protein OE33 OE33+CCN1 Het1A Het1A+CCN1
below 25%. Knockdown of each individual decoy receptor increased
TNSF10 TRAIL 1 6.21 × 10 3
1 0 the rate to 25–45%. Among them, knockdown of TRAILR4 had the best
TNSRSF10A TRAILR1 1 1.25 1 0 effect. Simultaneous knockdown of all three decoy receptors (OE33DCR-)
TNFRSF10B TRAILR2 1 14.93 1 0 raised the apoptosis rate to 65%, while expressing TRAILR2 at the same
TNFRSF10C TRAILR3 1 0 1 0
time caused a massive cell death, almost equivalent to the effect of
TNFRSF10D TRAILR4 1 0 1 3.10 × 103
TNFRSF11B OPG 1 0 1 1.46 × 107 TRAIL and CCN1 combination as in the previous experiment. Taken
together, this experiment demonstrated that each receptor contributed

167
T. Dang et al.
Fig. 3. TRAIL and CCN1 need each other to induce EAC cell apoptosis (Annexin V
binding assay). A. Equal number of OE33 and Het1A cells were seeded on the coverslips
and treated for 6 h with one of the following combinations: 1-[BSA], 2-[TRAIL], 3-[TRAIL
+ CCN1], 4-[CCN1], 5-[CCN1 + TRAIL antibody]. B. OE33 cells were transfected with
various plasmid combinations to establish cell lines with the following expression pro-
files: (1) OE33TRAILR2+, (2) OE33TRAILR3-, (3) OE33TRAILR4-, (4) OE33OPG-, (5) OE33DCR-,
(6) OE33TRAILR2+/DCR-, and (7) OE33. All cells were incubated with TRAIL for 6 h. C.
OE33 cells were transfected with various plasmid combinations to establish cell lines with
the following expression profiles: (1) OE33TRAILR3+, (2) OE33TRAILR4+, (3) OE33OPG+, (4)
OE33DCR+, and (5) OE33. All cells were incubated with CCN1 for 6 h.
to TRAIL-mediated apoptosis at some degree, either pro or against, with
TRAILR2 and TRAILR4 a bit more powerful on each side; only ex-
pression of all the death receptors without any interference from the
decoy receptors could give TRAIL the full power.
We also performed similar experiments using CCN1 to replace
TRAIL, to evaluate the contribution of the shutdown of each decoy
receptor to CCN1-induced EAC cell apoptosis. This time, we transfected
OE33 cells with a vector that expresses one of the decoy receptors one
by one or all three (DCR+) together, and then we treated the cells with
the recombinant CCN1 protein as in the previous experiments. As
shown in Fig. 3C, expression of anyone of the three decoy receptors
caused some reduction in CCN1-induced apoptosis, especially when
TRAILR4 was expressed (OE33TRAILR4+); when all three decoy re-
ceptors (OE33DCR+) were expressed, even CCN1 could not cause a
significant number of cell death, indicating that shutting down all three
decoy receptors is the main mechanism for CCN1 to induce EAC cell
apoptosis.
168
Experimental Cell Research 361 (2017) 163–169
4. Discussion
Up to date, five TRAIL receptors have been identified, and two of
them, TRAILR1 and TRAILR2, have the ability to trigger apoptosis [33].
Normally, upon binding to TRAILR1 or TRAILR2, TRAIL causes the
receptor reconfiguration in its intracellular death domain, so that an
adaptor protein FADD (Fas-Associated protein with Death Domain) can
hook up with the death receptor via the affinity of their matching death
domains. FADD also contains a death effector domain, which just fits to
the same structure in pro-caspase-8/−10. Once caspase-8/−10 is re-
cruited to FADD, it becomes activated to initiate caspase cascade and/
or mitochondrial leakage through Bid truncation/Bax activation, and
thereby leads to apoptotic execution [34].
TRAIL has been used to treat a variety of cancers and has gained
remarkable effects, because its death receptors are most likely ex-
pressed in tumor cells, not in healthy cells. However, more and more
cancer cells have been found resistant to TRAIL treatment for various
reasons that are not yet fully understood [35]. Several possible me-
chanisms have been postulated to explain TRAIL apoptotic failure. On
top of the list is the expression of TRAIL decoy receptors, namely,
TRAILR3, TRAILR4 and OPG. Anyone of them can effectively compete
with the death receptors for the ligand. Since the decoy receptors have
the ability to form both homo-oligomers and hetero-oligomers, their
dominant negative power can paralyze the death receptors completely
when they are expressed [36,37], resulting in TRAIL dysfunction. This
has been found true in breast cancer [38], lung cancer [39], leukemia
[40], and ovarian cancer [41]. Secondly, while TRAIL triggers apoptotic
pathways, it can also activate NF-κB, a transcription factor that pro-
motes several anti-apoptotic proteins including c-FLIP, Bcl-xL and IAPs.
Among them, c-FLIP has the same death effector domain as pro-caspase-
8/−10, therefore it can directly compete with pro-caspase-8/−10 for
its binding site on FADD, and thereby blocks the apoptotic signal to go
further [42]. In addition, the members of IAP family, such as XIAP and
survivin, have the ability to inhibit multiple caspases and have been
found to contribute to TRAIL failure in a number of tumor cells [43].
Other downstream molecules, such as Bcl-xL and Bcl-2, can also par-
tially abort TRAIL's effort through protecting mitochondrial integrity
[44].
In a given situation, which mechanisms responsible for TRAIL
failure is the key. Once we have the key, we can solve the problem case
by case. In EAC, our experiments showed that the expression of the
decoy receptors is the main reason, if not only, for TRAIL incapability.
We found that although the EAC cells express the death receptor
TRAILR1, they also express all three decoy receptors. Without the
presence of a sensitizer like CCN1, these decoy receptors are powerful
enough to beat TRAILR1 and keep the tumor cells alive, even if a plenty
of TRAIL is given. On the other hand, we were also startled by the fact
how tricky the EAC cells are. They express the death receptor TRAILR1,
but they do not express TRAIL ligand, having cunningly avoided a
possible suicide caused by the ligand-receptor interaction. In addition,
the expression of all three decoy receptors forms the second line of
defense for the tumor cells against apoptosis. In a way, this explains the
toughness of finding a cure for cancers. After so many years of effort,
our battle against cancers still continues.
CCN is a group of matricellular proteins with unique features in
between soluble growth factors and structural extracellular matrix
proteins. They use a variety of integrin combinations or heparan sulfate
proteoglycans as receptors to transmit signals into cells. Depending on
which receptor they choose, the result can be totally opposite. For in-
stance, when CCN1 attaches to integrin αvβ3 in fibroblast, it induces
cell proliferation [45], but if it sits on integrin αvβ5, the cells will
commit suicide [22]. Integrins do not possess an enzymatic domain, but
upon ligand binding, their cytoplasmic tails can recruit specific kinases
(e.g., JNK, ERK) and thereby induce specific cellular activities. Their
high specificity gives them the edge as targets for interventions. Recent
studies have generated a number of integrin targets for cancer therapy,
T. Dang et al.
and some of them are already under development for chemotherapeutic
drugs [46].
Acknowledgements
This study was supported by the VA Research Foundation in the
United States and Baotou Medical College in China, and did not receive
any specific grant from funding agencies in the public, commercial, or
not-for-profit sectors.
Conflict of interest
None.
References
[1] Global Burden of Disease Cancer Collaboration, C. Fitzmaurice, D. Dicker, A. Pain,
H. Hamavid, M. Moradi-Lakeh, M.F. MacIntyre, C. Allen, G. Hansen, R. Woodbrook,
et al., The Global burden of cancer 2013, JAMA Oncology 1 (2015) 505–527.
[2] J. Chai, M.M. Jamal, Esophageal malignancy – a growing concern, World J.
Gastroenterol. 18 (2012) 6521–6526.
[3] N. Pandeya, P.M. Webb, S. Sadeghi, A.C. Green, D.C. Whiteman, Australian cancer
study. gastro-oesophageal reflux symptoms and the risks of oesophageal cancer: are
the effects modified by smoking, NSAIDs or acid suppressants? Gut 59 (2010)
31–38.
[4] J. Lagergren, Adenocarcinoma of oesophagus: what exactly is the size of the pro-
blem and who is at risk? Gut 54 (2005) (Suppl 1: i1-5).
[5] D.A. Corley, Obesity and the rising incidence of oesophageal and gastric adeno-
carcinoma: what is the link? Gut 56 (11) (2007) 1493–1494.
[6] M.M. Finucane, G.A. Stevens, M.J. Cowan, et al., National, regional, and global
trends in body-mass index since 1980: systematic analysis of health examination
surveys and epidemiological studies with 960 country-years and 9.1 million parti-
cipants, Lancet 377 (2011) 557–567.
[7] C.J. Ruhm, Current and future prevalence of obesity and severe obesity in the
United States. Forum for Health Economics & Policy; 10 (2), article 6, 2007.
[8] M. Ng, T. Fleming, M. Robinson, et al., Global, regional, and national prevalence of
overweight and obesity in children and adults during 1980–2013: a systematic
analysis for the Global burden of disease study 2013, Lancet 384 (2014) 766–781.
[9] T. Ito, R.T. Jensen, Association of long-term proton pump inhibitor therapy with
bone fractures and effects on absorption of calcium, vitamin B12, iron, and mag-
nesium, Curr. Gastroenterol. Rep. 12 (2010) 448–457.
[10] J.S. Bajaj, S.M. Ratliff, D.M. Heuman, K.L. Lapane, Proton pump inhibitors are as-
sociated with a high rate of serious infections in veterans with decompensated
cirrhosis, Aliment. Pharmacol. Ther. 36 (2012) 866–874.
[11] L.A. Fraser, W.D. Leslie, L.E. Targownik, A. Papaioannou, J.D. Adachi, CaMos re-
search Group. The effect of proton pump inhibitors on fracture risk: report from the
Canadian multicenter osteoporosis study, Osteoporos. Int. 24 (2013) 1161–1168.
[12] D.A. Corley, A. Kubo, Body mass index and gastroesophageal reflux disease: a
systematic review and meta-analysis, Am. J. Gastroenterol. 101 (2006) 2619–2628.
[13] A.H. Merry, L.J. Schouten, R.A. Goldbohm, P.A. van den Brandt, Body mass index,
height and risk of adenocarcinoma of the oesophagus and gastric cardia: a pro-
spective cohort study, Gut 56 (2007) 1503–1511.
[14] F.C. Kimberley, G.R. Screaton, Following a TRAIL: update on a ligand and its five
receptors, Cell Res. 14 (2004) 359–372.
[15] M.A. Hall, J.L. Cleveland, Clearing the TRAIL for cancer therapy, Cancer Cell 12
(2007) 4–6.
[16] F.A. Kruyt, TRAIL and cancer therapy, Cancer Lett. 263 (2008) 14–25.
[17] C. Modak, J. Chai, Serum response factor: look into the gut, World J. Gastroenterol.
16 (2010) 2195–2201.
[18] J. Chai, Roles of SRF in endothelial cells during hypoxia, in: J. Chai (Ed.), Research
Directions in Tumor Angiogenesis, 2013 Intech Publisher, Rijeka, 2013, pp. 29–45.
[19] V. Juric, C.C. Chen, F. Lau, Fas-mediated apoptosis is regulated by the extracellular
matrix protein CCN1 in vitro and in vivo, Mol. Cell Biol. 29 (2009) 3266–3279.
[20] C.C. Chen, L.F. Lau, Deadly liaisons: fatal attraction between CCN matricellular
proteins and the tumor necrosis factor family of cytokines, J. Cell Commun. Signal.
4 (2010) 63–69.
[21] V. Todorovic, C.C. Chen, N. Hay, L.F. Lau, The matrix protein CCN1 (CYR61) in-
duces apoptosis in fibroblasts, J. Cell Biol. 171 (2005) 559–568.
[22] C.C. Chen, J.L. Young, R.I. Monzon, N. Chen, V. Todorović, L.F. Lau, Cytotoxicity of
TNFalpha is regulated by integrin-mediated matrix signaling, EMBO J. 26 (2007)
169
Experimental Cell Research 361 (2017) 163–169
1257–1267.
[23] C.A. Franzen, C.C. Chen, V. Todorović, V. Juric, R.I. Monzon, L.F. Lau, Matrix
protein CCN1 is critical for prostate carcinoma cell proliferation and TRAIL-induced
apoptosis, Mol. Cancer Res. 7 (2009) 1045–1055.
[24] J. Chai, M. Norng, C. Modak, K. Reavis, W. Mouazzen, J. Pham, CCN1 Induces a
reversible epithelial-mesenchymal transition in gastric epithelial cells, Lab.
Investig. 90 (8) (2010) 1140–1151.
[25] J. Chai, M. Norng, A.S. Tarnawski, J. Chow, A critical role of serum response factor
in myofibroblast differentiation during experimental oesophageal ulcer healing in
rat, Gut 56 (5) (2007) 621–630.
[26] M.T. Lin, C.C. Chang, S.T. Chen, H.L. Chang, J.L. Su, Y.P. Chau, et al., Cyr61 ex-
pression confers resistance to apoptosis in breast cancer MCF-7 cells by a me-
chanism of NF-kappaB-dependent XIAP up-regulation, J. Biol. Chem. 279 (2004)
24015–24023.
[27] S. Gery, D. Xie, H. Gabra, C. Miller, H. Wang, D. Scott, W.S. Yi, M.L. Popoviciu,
J.W. Said, H.P. Koeffer, Ovarian carcinomas: CCN genes are aberrantly expressed
and CCN1 promotes proliferation of these cells, Clin. Cancer Res. 11 (2005)
7243–7254.
[28] S.E. Holloway, A.W. Beck, L. Girard, M.R. Jaber, C.C. Jr Barnett, R.A. Brekken,
J.B. Fleming, Increased expression of Cyr61 (CCN1) identified in peritoneal me-
tastases from human pancreatic cancer, J. Am. Coll. Surg. 200 (2005) 371–377.
[29] C.C. Chen, K.H. Kim, L.F. Lau, The matricellular protein CCN1 suppresses hepato-
carcinogenesis by inhibiting compensatory proliferation, Oncogene 35 (2016)
1314–1323.
[30] W. Chien, T. Kumagai, C.W. Miller, J.C. Desmond, J.M. Frank, J.W. Said, et al.,
Cyr61 suppresses growth of human endometrial cancer cells, J. Biol. Chem. 279
(2004) 53087–53096.
[31] X. Tong, D. Xie, J. O'Kelly, C.W. Miller, C. Muller-Tidow, H.P. Koeffler, Cyr61, a
member of CCN family, is a tumor suppressor in non-small cell lung cancer, J. Biol.
Chem. 276 (2001) 47709–47714.
[32] J. Chai, C. Modak, Y. Ouyang, S.Y. Wu, M.M. Jamal, CCN1 induces β-catenin
translocation in esophageal squamous cell carcinoma through integrin α11, ISRN
Gastroenterol. (2012) 207235.
[33] H.N. LeBlanc, A. Ashkenazi, Apo2L/TRAIL and its death and decoy receptors, Cell
Death Differ. 10 (2003) 66–75.
[34] F.C. Kischkel, D.A. Lawrence, A. Chuntharapai, P. Schow, K.J. Kim, A. Ashkenazi,
Apo2L/TRAIL-dependent recruitment of endogenous FADD and caspase-8 to death
receptors 4 and 5, Immunity 12 (2000) 611–620.
[35] M.J. Dyer, M. MacFarlane, G.M. Cohen, Barriers to effective TRAIL-targeted therapy
of malignancy, J. Clin. Oncol. 25 (2007) 4505–4506.
[36] J.P. Sheridan, S.A. Marsters, R.M. Pitti, et al., Control of TRAIL-induced apoptosis
by a family of signaling and decoy receptors, Science 277 (1997) 818–821.
[37] S. Neumann, J. Hasenauer, N. Pollak, P. Scheurich, Dominant negative effects of
tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) receptor 4
on TRAIL receptor 1 signaling by formation of heteromeric complexes, J. Biol.
Chem. 289 (2014) 16576–16587.
[38] A.D. Sanlioglu, E. Dirice, C. Aydin, N. Erin, S. Koksoy, S. Sanlioglu, Surface TRAIL
decoy receptor-4 expression is correlated with TRAIL resistance in MCF7 breast
cancer cells, BMC Cancer 5 (2005) 54.
[39] C. Aydin, A.D. Sanlioglu, B. Karacay, G. Ozbilim, L. Dertsiz, O. Ozbudak, C.A. Akdis,
S. Sanlioglu, Decoy receptor-2 small interfering RNA (siRNA) strategy employing
three different siRNA constructs in combination defeats adenovirus-transferred
tumor necrosis_factor-related apoptosis-inducingligand resistance in lung cancer
cells, Hum. Gene Ther. 18 (2007) 39–50.
[40] R. Riccioni, L. Pasquini, G. Mariani, E. Saulle, A. Rossini, D. Diverio, E. Pelosi,
A. Vitale, A. Chierichini, M. Cedrone, R. Foà, F. Lo Coco, C. Peschle, U. Testa, TRAIL
decoy receptors mediate resistance of acute myeloid leukemia cells to TRAIL,
Haematologica 90 (2005) 612–624.
[41] D. Lane, I. Matte, C. Laplante, P. Garde-Granger, C. Rancourt, A. Piché,
Osteoprotegerin (OPG) activates integrin, focal adhesion kinase (FAK), and Akt
signaling in ovarian cancer cells to attenuate TRAIL-induced apoptosis, J. Ovarian
Res. 6 (2013) 82.
[42] M. Irmler, M. Thome, M. Hahne, et al., Inhibition of death receptor signals by
cellular FLIP, Nature 388 (1997) 190–195.
[43] A.D. Schimmer, K. Welsh, C. Pinilla, et al., Small-molecule antagonists of apoptosis
suppressor XIAP exhibit broad antitumor activity, Cancer Cell 5 (2004) 25–35.
[44] S. Hinz, A. Trauzold, L. Boenicke, et al., Bcl-XL protects pancreatic adenocarcinoma
cells against CD95- and TRAIL-receptor-mediated apoptosis, Oncogene 19 (2000)
5477–5486.
[45] T.M. Grzeszkiewicz, D.J. Kirschling, N. Chen, L.F. Lau, CYR61 stimulates human
skin fibroblast migration through integrin αVβ5 and enhances mitogenesis through
integrin αVβ3, independent of its carboxyl-terminal domain, J. Biol. Chem. 276
(2001) 21943–21950.
[46] D. Cox, M. Brennan, N. Moran, Integrins as therapeutic targets: lessons and op-
portunities, Nat. Rev. Drug Discov. 9 (10) (2010) 804–820.