• Open Access

Connective tissue growth factor enhances the migration of gastric cancer through downregulation of E-cadherin via the NF-κB pathway


To whom correspondence should be addressed.
E-mail: dy_jin@yahoo.cn; jianxzhang@yahoo.cn


Local invasion and distant metastasis are difficult problems for surgical intervention and treatment in gastric cancer. Connective tissue growth factor (CTGF/CCN2) was considered to have an important role in this process. In this study, we demonstrated that expression of CTGF was significantly upregulated in clinical tissue samples of gastric carcinoma (GC) samples. Forced expression of CTGF in AGS GC cells promoted their migration in culture and significantly increased tumor metastasis in nude mice, whereas RNA interference-mediated knockdown of CTGF in GC cells significantly inhibited cell migration in vitro. We disclose that CTGF downregulated the expression of E-cadherin through activation of the nuclear factor-κappa B (NF-κB) pathway. The effects of CTGF in GC cells were abolished by dominant negative IκappaB. Collectively, these data reported here demonstrate CTGF could modulate the NF-κappaB pathway and perhaps be a promising therapeutic target for gastric cancer invasion and metastasis. (Cancer Sci 2011; 102: 104–110)

Gastric cancer (GC) is one of the most common malignancies in the world, ranking as the second most common cause of death worldwide.(1) The only way for cure is surgery combined with chemotherapy. However, approximately 65% of patients with gastric carcinoma have regional or distant metastases at the time of diagnosis, limiting the chance for complete excision of these tumors and thus making treatment extremely difficult. The 5-year survival rate remains poor for this type of cancer,(2,3) especially in patients with metastatic and advanced stages, ranging 5–15%.(4)

Although we have made great progress in perioperation, recurrence and metastasis of disease are almost uniform and contribute to the main cause of death. Failure of traditional therapeutic strategy to treat this disease might be due to our limited understanding of the underlying mechanism facilitating the migration and invasion of gastric cancer cells.

Connective tissue growth factor (CTGF) is a cysteine-rich protein. It belongs to the CCN family, which consists of six members: Cyr61 (cysteine-rich protein 61, CCN1), CTGF/CCN2, Nov (nephroblastoma overexpressed gene, CCN3), WISP-1 (Wnt-1-induced secreted protein 1, CCN4), WISP-2 (CCN5) and WISP-3 (CCN6).(5,6) Connective tissue growth factor has versatile functions associated with the modular domain structure of itself. It has four structural domains as follow: NH2-terminal signal peptide and four conserved domains with sequence similarities to insulin-like growth factor-binding proteins; von Willebrand type C factor; thrombospondin 1; and a cysteine knot characteristic of other growth factors, including platelet-derived growth factor, nerve growth factor and transforming growth factor (TGF)-β,(7) each of which is thought to have a distinct biological function.(8)

Although the specific receptor of CTGF has not yet been identified, CTGF is reported to interact with several growth factor signaling pathways, including TGF-β, bone morphogenetic protein (BMP) and Wnt.(9) It can also bind to integrins on the cell surface. It is thought to be involved in extracellular matrix production, desmoplasia, tumor cell proliferation, adhesion, migration, angiogenesis and metastasis.(5,10) Increasing evidence suggests that CTGF plays a role in tumorigenesis. Its overexpression is observed in prostate cancers,(11) gliomas,(12) breast cancers(13) and esophageal adenocarcinoma.(14) Conversely, downregulation of CTGF was found in lung cancer(15) and colon cancer,(16) in which forced expression of CTGF inhibits invasion and metastasis of cancer cells both in vitro and in vivo. Also, CCN2 expression levels correlated with increased survival in chondrosarcoma patients.(17) It is proposed that this disparity in expression in different types of tumors implies that regulation of CTGF expression might be cell-type specific and CTGF might have versatile roles in different cellular contexts.

In this study, we demonstrated that the expression of CTGF was significantly increased in GC. Furthermore, we demonstrated that the CTGF activated the NF-κB signaling pathway in GC cells, which enhanced the migration and thus facilitated the metastasis of cancer cells. Our study strongly highlights the significance of CTGF in the migration and invasion of GC, and therefore provides a potential drug target in gastric cancer therapy.

Material and Methods

Cell culture, cDNA and tissue samples.  The GC cell line AGS was obtained from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai Institute of Cell Biology, Chinese Academy of Sciences) and cultured in RPMI 1640 medium (Invitrogen, Carlsbad, CA, USA) with 10% fetal bovine serum (FBS; PAA Laboratories, Pasching, Austria), 10 units/mL penicillin-G and 10 mg/mL streptomycin. 293T cells were cultured in DMEM (Invitrogen) supplemented with 10% fetal bovine serum, 10 units/mL penicillin-G and 10 mg/mL streptomycin. All cells were incubated at 37°C in 5% CO2 humidified air.

Seventy-three pairs of primary GC and their corresponding adjacent normal tissues, which were at least 5 cm away from the cancer, were obtained from GC patients treated at Zhongshan Hospital of Fudan University (Shanghai, China) from 2007 to 2008. The entire specimen was stored in liquid nitrogen. The GC cDNA was prepared from the samples collected above. Whole cDNA samples were stored at −80°C until analysis. Our work was approved by the Institutional Review Board of Zhongshan Hospital of Fudan University and informed consent was obtained for every specimen examined. The characteristics of all cases of GC are summarized in Table 1.

Table 1.   Association between CTGF expression and clinicopatho-logical characteristics of patients with GC (n = 73)
CharacteristicNo. patientsCTGF expressionχ2 test
  1. *P < 0.05. CTGF, connective tissue growth factor; GC, gastric carcinoma; TNM, tumor–node–metastasis.

 Male452322P = 0.325
Age (years)
 ≤62301515P = 0.624
Size (cm)
 <5211110P = 0.527
 Well1073P = 0.029*
Lauren type
 Intestinal342212P = 0.04*
TNM stage
 I + II301614P = 0.334
 III + IV431825
LN metastasis
 Absent24168P = 0.016*
 Absent633033P = 0.654

Plasmid construction and transfection.  Construction of plasmids encoding CTGF and the selection of positive clones expressing CTGF have been described previously.(14) Luc2 gene was cloned into FG12 expressing construct and prepared for lentiviral expression. Both AGS vector control and CTGF expressing cells were then labeled with FG12-Luc2 for metastasis assay. The dominant negative form of IκappaB expression construct (DN IκappaB) was a gift from Professor Xiangjun Tong.

Real-time PCR (RT-PCR) analysis.  Primers for the genes tested in the present experiments were designed using software PRIMER5. The CTGF and E-cadherin mRNA were assayed using real-time reverse transcriptase-PCR and the data were analyzed by SPSS as previously described.(14,18)

Western blot analysis.  Western blot analysis was performed as previously described.(14) Primary antibodies to CTGF, E-cadherin, β-catenin, p65 and nucleoporin 62 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), and antibody to β-actin was purchased from Sigma (St. Louis, MO, USA). Mouse anti-Akt polyclonal antibody and rabbit anti-phosphorylated Akt polyclonal antibody were from Cell Signaling (Boston, MA, USA). Secondary antibodies, rabbit anti-mouse IgG (Sigma) and goat anti-rabbit IgG (Cell Signaling Technology), were used at a dilution of 1:1500. Primary antibodies were diluted in 0.1 mol/L Tris-HCl, 1.5 mol/L NaCl, 0.05% Tween-20 (v/v) (TBST) containing 1% BSA and NaN3. The immunoreactive protein bands were visualized using an ECL kit (Pierce, Rockford, IL, USA).

MTT assay. In vitro cell growth was measured using an MTT assay as previously described.(18) The measurement process was performed every 48 h for 5 or 7 days to generate a cell growth curve.

In vitro migration assays.  Cell migration was determined using a modified Boyden chamber assay as previously described.(19)

Luciferase activity assay.  Luciferase activity of the NF-kappa B reporter was measured 24 h after transfection using a Dual-Luciferase Reporter Assay System following the product instructions.

Immunofluorescence.  Immunofluorescence was carried out in the same way as the procedure previously described.(18)

Nuclear protein extraction.  Nuclear extracts were prepared by the mini-extraction method as previously described.(14) Nuclear extracts were immediately used for western blot.

ILK kinase assay.  The integrin-linked kinase (ILK) kinase assay was determined as previously described.(20)

RNAi-mediated knockdown of CTGF.  The CTGF RNAi sequences were designed with the siRNA Design Tools online (http://www.Ambion.com). FG12 RNAi vector was used to produce small double-stranded RNA (small interfering RNA) to inhibit target gene expression in gastric cancer cells.

In vivo metastasis assay.  The in vivo metastasis assay was detected by inspecting the luminescence with an IVIS imaging system as previously described.(18,21) There were five mice for vector control and five mice for those overexpressing CTGF. All mice were treated in accordance with the American Association for the Accreditation of Laboratory Animal Care guidelines.

Statistical analysis.  All analyses were performed using the statistical software SPSS (Chicago, IL, USA, version 16.0 for Windows). All data were expressed as mean ± SD. For comparison of two different groups, Student’s t-test was used. Differences between groups were considered significant at P < 0.05.


CTGF is upregulated in clinical tissue samples of gastric carcinoma.  To study the expression pattern of the CTGF gene in gastric cancers, we first examined CTGF mRNA expression levels in 42 pairs of matched gastric cancer tissue samples by real-time polymerase chain reaction (RT-PCR) (Fig. 1A). Expression levels were expressed as a ratio between CTGF and the reference gene β-actin to correct for the variation in the amounts of RNA. Upregulation of the CTGF gene occurred in 33 of 42 (78.6%) gastric cancers compared with the paired normal gastric tissues. Univariate analysis showed that the mRNA levels of CTGF in the gastric cancer samples was significantly higher than paired normal samples (P < 0.05).

Figure 1.

 Levels of connective tissue growth factor (CTGF) are increased in gastric carcinoma (GC) clinical samples. (A) Quantitative real-time PCR results of the relative expression level of CTGF in 42 pairs of GC and normal gastric tissue samples. All of the reactions were performed in triplicate and results represent the mean ± SD (P < 0.05). The expression of CTGF is normalized to beta-actin. (B) Expression of CTGF and p65 protein in four randomly selected paired GC samples was analyzed using western blot. N, normal; T, primary GC. (C) immunohistochemistry (IHC) results of CTGF expression in paired GC tissue samples.

Elevated levels of CTGF protein were also found in four randomly picked pair GC samples with different expression levels of CTGF mRNA, as shown by western blot analysis (Fig. 1B). In addition, in order to further confirm the result demonstrated above, 73 paired samples were used for examination with immunochemical staining. The primary antibody used, IgG, was considered as a negative control (data not shown). As shown in Figure 1C, elevated levels of CTGF protein were also found in human GC tissues compared with the paired normal tissue from the patients, as shown by immunochemical staining.

Forced expression of CTGF in AGS GC cell line stimulated their migration.  To further examine the effects of CTGF in gastric cancer cells, AGS cells were stably transfected with either a pcDNA/CTGF containing full length CTGF or an empty vector pcDNA3.1 as a control. Six G418-resistant clones were screened for CTGF expression by western blot analysis. Two pcDNA/CTGF stable transfected clones with high expression of CTGF (AGS/CTGF 3 and 5), which were used for further study, are shown in Figure 2A. In addition, because CTGF has been reported to be able to secrete into the medium, we also measured the level of CTGF in the media of these CTGF stably-transfected cell lines by western blot, which was coincided with the level of CTGF in the lysate of each cell line.

Figure 2.

 Effects of connective tissue growth factor (CTGF) on cell migration. (A) Expression of CTGF in AGS/CTGF stable cell lines was evaluated by western blot. Equal loading was ascertained using β-actin as an internal control. (B) AGS cell migration detected with Boyden chamber assay. Each experiment was performed in triplicate, and results represent the mean ± SD (P < 0.05). (C) Cell growth rates of AGS/V, AGS/CTGF3# and -5# in 1% FBS were measured using an MTT assay. Results are representative of three independent experiments (P > 0.05). (D) The level of E-cadherin in AGS/pcDNA3.1, AGS/CTGF3# and AGS/CTGF5# transfected cells was examined by western blot. (E) Expression of the E-cadherin mRNA level was examined by RT-PCR.

The effect of CTGF on cell migration was evaluated by collagen-I-coated Boyden chamber analysis. AGS/CTGF-3, AGS/CTGF-5 and AGS/V cells were seeded into the upper compartment of the chamber and their ability to migrate was evaluated. The results showed that overexpression of CTGF dramatically increased AGS migration ability compared with the control cells (P < 0.05) (Fig. 2B). In the MTT assays, no significant difference was observed in the rate of growth between AGS/CTGF3, -5 and AGS/V control cells (P > 0.05) (Fig. 2C).

Furthermore, real-time PCR and western analysis showed that cell adhesion-related protein E-cadherin was downregulated in the two stable clones compared with the AGS/V control cells (Fig. 2D,E).

These results indicate that the forced expression of CTGF in GC cells enhanced their cell migration.

Knockdown of CTGF inhibited cell migration.  To further determine the effect of CTGF on cell migration, we used RNAi mediated knockdown lentiviral vector to decrease the basal level of CTGF in MKN45 cells. After being infected with lentiviral vector for 8 h in condition media, the cells were cultured in DMEM containing 10% FBS for 72 h and then harvested. Cell sorting was performed with the FACS and the GFP-positive cells were collected for further study. Western blot results showed that two sequences markedly decreased the expression of CTGF compared with the control sequence (Fig. 3A). The RNAi dramatically inhibited cell migration of these cells in liquid culture (< 0.05) (Fig. 3B).

Figure 3.

 Inhibition of MKN45 cell migration by connective tissue growth factor (CTGF) knockdown. MKN45 cells were infected with either CTGF siRNA or CTGF siRNA control (con) for 8 h in condition media, cultured in DMEM containing 10% FBS for 72 h, then harvested for further study. Cell sorting was performed with the FACS and the GFP-positive cells were collected for further study. (A) Western blot analysis to show RNAi-mediated knockdown of CTGF. (B) Migration assay detected with Boyden chamber showed the effect of CTGF gene silencing on cell migration of MKN45 cells. The results are expressed as mean number (±SD) of migrated cells within 6 h of three separate experiments (P < 0.05). (C) Knockdown of CTGF enhanced the expression of E-cadherin in MKN45.

To investigate whether E-cadherin also contributes to the decreased cell migration induced by silencing of CTGF, we detected the expression of E-cadherin in MKN45 cells. As expected, E-cadherin was markedly increased after knocking down CTGF (Fig. 3C). These results suggest that downregulation of CTGF in GC cells inhibited the migration of MKN45 cells.

Involvement of NF-kB in CTGF-induced cell migration via E-cadherin suppression.  NF-κB, a well-known mediator in cell migration, has been implicated in cancer carcinogenesis.(22,23) We assumed that CTGF could enhance cell migration by activating NF-κB. We first examined the activation of NF-κB by using the luciferase reporter assay. These reporter constructs were transfected into either AGS or 293T cells together with either the CTGF expression construct or pcDNA3.1, respectively, and luciferase activity was determined. In both AGS cells and 293T cells, the cellular transcriptional activity of NF-κB was significantly increased on CTGF overexpression (Fig. 4A).

Figure 4.

 Connective tissue growth factor (CTGF) stimulated NF-κ B signaling pathway. (A) NF-κB reporter activity in AGS and 293T cells. Cells were co-transfected with NF-κB reporter plasmid as well as either pcDNA3.1/CTGF or pcDNA3.1. Reporter activities were normalized to internal control Renilla. Data represent results from triplicate dishes in two separate experiments (P < 0.05). (B) NF-κB nuclear translocation enhanced by CTGF was evaluated by immunofluorescence of p65. (C) The expression of nuclear p65 in AGS/V and AGS/CTGF was examined by specific antibodies and nucleoporin 62 served as a loading control. Dominant negative IκappaB (DN IκappaB) blocked downregulation of E-cadherin in AGS/CTGF cells at the mRNA level (D) and protein level (E) and inhibited the translocation of p65 to the nucleus (F). (G) ILK kinase activity and active Akt were examined in AGS/V and AGS/CTGF stable cell lines. (H) Western blot assay demonstrated that knockdown of ILK by siRNA abolished the up-regulation of AKT phosphorylation induced by CTGF overexpression. (I) AGS vector control and CTGF overexpressing cells were treated as follow: no treat (control), LY294002 for 24 h, or transfected with dominant negative Akt-expressing constructs, and then the NF-κB activity was tested by dual luciferase assay.

In most cell types, NF-κB is found in the cytoplasm as an inactive dimer bound to one of the nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor (IκB) proteins that mask its nuclear localization signal. However, as assessed by immunofluorescence, marked nuclear localization of p65 was observed in AGS/CTGF cells compared with control cells (Fig. 4B).

To investigate whether overexpression of CTGF could alter NF-κB signaling, western blot was performed to examine the expression of nuclear p65. We found that the expression levels of nuclear p65 were elevated significantly in AGS/CTGF stables cells compared with AGS/V control cells (Fig. 4C). Meanwhile, expression of p65 coincided with the expression of CTGF in the GC samples (Fig. 1B).

To further examine these results, AGS/CTGF cells were transfected transiently with a dominant negative IκappaB (DN IκappaB). Real-time PCR and western analysis showed that the upregulation of nuclear p65 and downregulation of E-cadherin in AGS/CTGF cells was abolished by DN -IκappaB (Fig. 4D–F).

Expression and subcellular localization of E-cadherin, NF-κB p65 subunit with CTGF expression in the sample were compared. There was more nucleus p65 localization and less membrane E-cadherin with CTGF expression in diffuse-type adenocarcinomas than in intestinal-type adenocarcinomas (P < 0.05) (Table 2). These results suggested that CTGF stimulated NF-κB signaling causing downregulation of the E-cadherin.

Table 2.   Subcellular localization of p65 and E-cadherin with CTGF in the histological type in patients with GC (n = 73)
  1. P < 0.05. CTGF, connective tissue growth factor; GC, gastric carcinoma.

  2. ---, incidence of nucleus p65 and abberrant E-cadherin with high CTGF in intestinal versus diffuse GC.


It has been demonstrated that CTGF could affect cell function through ILK/PI3K/AKT signaling,(24,25) so we determined whether overexpression of CTGF in AGS/CTGF cells had an effect on this pathway. We measured the level of ILK, Akt and activated Akt by western blot analysis and ILK activity by ILK kinase assay. As shown in Figure 4G, ILK activity and phosphorylation Akt in both AGS/CTGF sublines were increased dramatically compared with those in the AGS/V control cells, while protein levels of ILK and total Akt in AGS/CTGF cells had no significant difference. Furthermore, downregulated expression of ILK with siRNA dramatically decreased the level of activated Akt in CTGF-expressing cells (Fig. 4H).

It is reported that NFκB can be regulated by AKT signaling pathway, we asked if AKT pathway is also involved in CTGF-dependent regulation of NF-κB. Indeed, treatment of AKT inhibitor LY294002, or overexpression of dominant-negative AKT, abolished CTGF-induced activation of NF-κB, respectively. CTGF-induced activation of NF-κB was blocked by LY294002, dominant-negative Akt, respectively (Fig. 4I). These results indicate that the ILK/Akt pathway is involved in the modulation of NF-κB activity stimulated by CTGF.

CTGF promotes metastasis of AGS cells in vivo.  To examine the role of CTGF in metastasis in vivo, AGS/V cells and AGS/CTGF cells were injected into the heart of 6-week-old nude mice and monitored by in vivo luminescence imaging every week. Once the same amount of cancer cells were injected, luminescence intensity could be detected immediately in the two groups after the luciferin was injected intraperitoneally and there was no difference. Four weeks post-injection, the luminescence intensity detected in the AGS/CTGF tumor group was higher than the luminescence in the AGS tumor group. In 6 weeks, the luminescence intensity in AGS/CTGF was more significantly higher than in the control. This suggested that CTGF dramatically enhanced the capability of AGS cells to develop tumors in distant organs (Fig. 5A).

Figure 5.

 Overexpression of connective tissue growth factor (CTGF) promotes gastric cell growth in distant organs. (A) AGS vector control and CTGF overexpressing cells, each 2 × 105, both labeled with luciferase, were injected into the left ventricle of nude mice, and metastasis was measured by luciferase activity through intraperitoneal injection of the luciferin every week (five mice for each group). The metastasis events were detected weekly by inspecting the intensity of luminescence with imaging via IVIS systems, which is demonstrated as the color depth in the image. (B) The metastasis lesion was picked out to confirm that the bioluminescence signaling was not artificial. (C) Metastasis lesions were taken out after the mice were killed. CTGF, p65 and E-cadherin expression levels were detected by western blot.

To exclude the possibility that the signal examined might be artificial, after the mice were killed we found the metastasis lesions according to the previously obtained data (Fig. 5B). Proteins from the metastasis were extracted and the expression of CTGF, p65 and E-cadherin was examined. As shown in Figure 5C, the level of CTGF and p65 in the AGS/CTGF metastasis was higher than that in the AGS/V metastasis, while E-cadherin was lower (Fig. 5C). These data thus suggest that CTGF upregulation also enhanced the ability of metastasis of AGS cells in vivo.


In the present study, our results showed that CTGF was highly expressed in GC tissues compared with matched normal gastric tissue from patients. We subsequently evaluated the function of CTGF in AGS, a GC cell line. Forced expression of CTGF in AGS promoted cell migration in vitro and stimulated tumor metastasis in nude mice, whereas knockdown of CTGF in AGS inhibited cell migration in vitro. These results suggest that CTGF might be implicated in the metastasis of GC. We also found that expression of E-cadherin decreased with overexpression of CTGF. E-cadherin is one of the important elements involved in adhesion, which explains in part how CTGF enhances migration of the GC cells.

The involvement of CTGF in migration has been mentioned in several previous investigations. Early studies demonstrated that overexpression of CTGF promoted migration in chondrosarcoma cancer(26) and stimulated drug resistance and metastasis in glioblastoma.(27) The level of CTGF showed a significant correlation with tumor stage and survival in breast cancers.(28)

However, some other studies found that CTGF inhibited tumor metastasis and invasion in ovarian carcinoma(29) and lung cancers.(24) Altogether, the data suggest that the role of CTGF in tumor invasion is dependent on the tumor type and microenvironment.

An important finding of the present study is the characterization of CTGF as a metastatic enhancer by stimulating the NF-κB signaling pathway then downregulated expression of E-cadherin. NF-κB is one member of the NF-κB/Rel family, and p65 is the most common subunit of NF-κB translocated into the nuclei, which stimulates the transcription of a set of target genes involved in cell migration.(30) The nuclear localization of p65 is controlled by the bind state regulated by IκB, and the latter is inhibited by IKK through direct phosphorylation.

Previous studies have documented that CTGF binds to various integrins such as α6β1(31) and α5β1.(32) Moreover, Cyr61, another CCN member, has been shown to bind with integrins stimulating the activity of integrin-linked kinase and AKT;(33) the latter is considered to directly activate the NF-κB pathway. Because the p65 signaling pathway has been reported to be involved in the development of human cancers,(34) and dysfunction of NF-kappaB signaling in cancer of the stomach has been recently reported,(35) regulation of the p65 pathway by CTGF might further explain the mechanism by which CTGF stimulates the migration of GC cells.

Besides CTGF, several other CCN molecules function through the p65 pathway in cancers. For example, Cyr61 promotes breast cancer cell growth through activating NF-κB signaling(36). Considering the similarity of the two proteins, we speculate that they may function in similar signaling pathways in the development and progression of tumors.

In summary, our results suggest that CTGF can stimulate NF-κB signaling and thus contribute to GC malignancy. These findings recapitulated the involvement of the NF-κB pathway in cancer metastasis. Given the diverse function of CTGF, the NF-κB pathway seems to be involved in more important physiological and pathophysiological processes than earlier conceived. Taken together, our work lays the foundation for novel exploration of the mechanisms and functions of CTGF and the NF-κB signaling pathway in the metastasis and invasion of GC.


This work was supported by The Foundation for Young Scientists of Zhongshan Hospital, Fudan University (Grant numbers: 2008-S-61). The authors thank Dr Yuezhen Deng, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences for expert technical assistance and Dr YeHe, Department of Neuroscience, Mount Sinai School of Medicine for helpful discussions, suggestions and critical reading of this manuscript.

Disclosure Statement

The authors have no conflict of interest.


connective tissue growth factor


gastric carcinoma