Hepatology 50(5): 1453-1463

PMC: PMC3328302

PMID: 19676131

doi: 10.1002/hep.23143

Targeting Cadherin-17 Inactivates Wnt Signaling and Inhibits Tumor Growth in Liver Carcinoma

Xiao Wang+9 authors

Conclusion

Our results identify CDH17 as a novel oncogene in HCC and suggest that CDH17 is a biomarker and attractive therapeutic target for this aggressive malignancy.

Materials and Methods

HCC Cell Lines and CDH17 Monoclonal Antibodies

Hepatoma cell lines PLC/PRF/5, HuH-7, Hep3B, and HepG2 were obtained from the American Type Culture Collection (Manassas, VA) or Japanese Cancer Research Resources Bank (Tokyo, Japan). Primary and metastatic HCC cell lines H2-P, H2-M,¹⁸ MHCC97H, and MHCC97L¹⁹ and the immortalized human hepatocyte cell line MIHA²⁰ were obtained and used as described. A mouse monoclonal anti-CDH17 antibody (Lic-3, IgG2a) was established using the recombinant amino-terminal domain 1–2 (amino acid residues: 30–244) of human CDH17 as an antigen according to previously reported procedures.²¹

MicroRNA-Based CDH17 Short Hairpin RNA (CDH17 shRNAmir)

The procedure for suppressing the expression of CDH17 using a lentiviral-mediated approach was performed as described.²² Several target regions spanning exon 3 (nt 244–262), 4 (nt 414–432), 12 (nt 1558–1576), and 18 (nt 2506–2524) of human CDH17 ({"type":"entrez-nucleotide","attrs":{"text":"NM_004063.2","term_id":"16507959","term_text":"NM_004063.2"}}NM_004063.2) were chosen. These target sequences were cloned into an Expression Arrest pSHAG-MAGIC2 (pSM2) vector (Open Biosystems, Huntsville, AL). Testing showed that the construct targeting exon 3 yielded the best suppression efficiency (data not shown); therefore, this construct was chosen. The construct was cloned into a LUNIG vector,²² which was then transfected into 293T/17 (American Type Culture Collection). The lentivirus was used to transduce MHCC97H cells, and the stable transfectants were selected using G418 sulfate for 2 weeks. Two controls were included, MHCC97H cells that received either no treatment (97H) or a nontargeted RNAi vector (Mock). Construction of the CDH17 short hairpin RNA (shRNA) vector is described in the Supporting Methods.

Luciferase Reporter Assay

Tumor cells were seeded onto 24-well plates and cultured for 24 hours. TOPFlash (T cell factor reporter plasmid) luciferase construct (Upstate-Millipore, Billerica, MA) and pRL-TK-Luc vector (Promega) for normalization were used to cotransfect cells using FuGENE-6 reagent (Roche). After 48 hours, cells were lysed and assayed for luciferase activities using the Dual-Luciferase Reporter Assay System (Promega).

Cell Cycle and Apoptotic Assays

Cells were fixed in 1% paraformaldehyde and stained with 0.5 mL PI/RNase Staining Buffer (BD Biosciences, San Jose, CA) for 15 minutes at room temperature. Stained cells were then analyzed using flow cytometry (Beckman Coulter Cytomics FC500). Apoptotic cells were detected using the ApopTag Red In Situ Apoptosis Detection Kit (Chemicon-Millipore, Billerica, MA) as described.²³

In Vitro Assays to Assess Tumor Phenotypes

The experimental procedures for cell proliferation, colony formation, cell adhesion, wound healing, cell invasion, and cell migration are described in the Supporting Methods.

In Vivo Mouse Models of HCC

Tumor xenograft and genetically defined mouse models for HCC are described in the Supporting Methods. For the liver tumor–bearing mouse model, subcutaneous tumors were induced in nude mice using MHCC97H cells as mentioned. One week later, mice with inoculated tumors were subjected to different experimental treatments: recombinant adeno-associated virus expressing TP53 (rAAV-TP53), lentivirus-mediated suppression using CDH17 shRNAmir (CDH17 shRNAmir), or a combined treatment of both rAAV-TP53 and CDH17 shRNAmir. Five mice were used in each group. For each group, mice received 100 µL therapeutic agents by way of intratumoral injections. For the treatment groups, 1 × 10^{copies of the corresponding virus for each gene therapy were injected twice weekly for 2 weeks. For the 97H control and Mock groups, intratumoral injections of 1 × Tris-buffered saline and 1 × 10}^{copies of lentivirus with the nontargeted RNAi vector were performed. Tumor formation and growth were monitored daily. To monitor virus infection in each group, the expression levels of genes associated with virus infection and neutrophilic granulocyte-related inflammation, such as 2’–5’-oligoadenylate synthetase 1, interleukin-8, and protein-kinase, interferon-inducible double-stranded RNA dependent inhibitor, repressor of (P58 repressor) (PRKRIR), in the subcutaneous tumors were assessed by way of quantitative polymerase chain reaction.}

Statistical Analysis

Statistical analyses were performed using PRISM version 4.0 software (GraphPad, San Diego, CA). A Student t test and χ^{test were used for calculating the significance between different groups. Values are expressed as the mean ± standard error of the mean. A}P value <0.05 was considered significant. Copy number variation (CNV) was measured using the Ilumina platform that contains 650,000 single-nucleotide polymorphisms, and the results are expressed as the logR ratio (that is, the intensity ratio of the studied sample to a number of reference samples). The genomic locations of CDH17 single-nucleotide polymorphism probes were identified. CNV of CDH17 was obtained by averaging the logR ratio from those single-nucleotide polymorphism probes. The same procedure was used to study the CNV of all ≈40,000 transcripts. In addition to the continuous CNV data, the hidden Markov model (HMM) was applied to determine if there was copy number gain or loss.²⁴

HCC Cell Lines and CDH17 Monoclonal Antibodies

MicroRNA-Based CDH17 Short Hairpin RNA (CDH17 shRNAmir)

Luciferase Reporter Assay

Cell Cycle and Apoptotic Assays

In Vitro Assays to Assess Tumor Phenotypes

The experimental procedures for cell proliferation, colony formation, cell adhesion, wound healing, cell invasion, and cell migration are described in the Supporting Methods.

In Vivo Mouse Models of HCC

Statistical Analysis

Results

Tumorigenic and Metastatic Properties of CDH17 in HCC

We used a genetically defined mouse model in which the CDH17 gene was subcloned into the murine stem cell virus SV40–green fluorescent protein recipient vector, then constitutively induced in immortalized premalignant liver progenitor cells.⁴,6,25 The CDH17-overexpressing cells gave rise to subcutaneous tumors by 18 days after implantation in mice, whereas the vector-carrying cells did not (Fig. 1A). This is the first direct demonstration of tumorigenesis by CDH17 in the HCC context.

An external file that holds a picture, illustration, etc.
Object name is nihms353854f1.jpg

Fig. 1

(A) Tumor induction of CDH17 in a genetically engineered mouse model of HCC. Mouse embryonic liver progenitor cells derived from p53−/−; Myc cells, transfected with a vector containing CDH17 (CDH17) or an empty vector (Mock), were seeded in recipient mice. Tumor advent and development were monitored using a fluorescent imaging technique for 26 days. The photograph shows the tumors observed at 18 days after cell seeding. The tumor volume was quantified for these two groups (n = 4). (B) The expression of CDH17 transcript in metastatic HCC cell lines (MHCC97H [97H], MHCC97L [97L], and H2-M) and primary HCC cell lines (H2-P, PLC, HuH-7, Hep3B, and HepG2) measured using quantitative polymerase chain reaction and represented as the ratio compared with expression in the immortalized normal hepatocyte cell line (MIHA). (C) Genomic amplification of CDH17 in HCC. CDH17 showed a 49% amplification rate in tumor tissues (TU) (about half of patients have DNA copy number gain according to the hidden Markov model) that is 1.7e-4 by chance according to the amplification rate at adjacent normal liver (AN) samples. Its deletion rate was 5.3% and considered not significant.

Next, we measured the CDH17 messenger RNA level in a panel of human HCC cell lines with different metastatic potential by way of quantitative polymerase chain reaction analysis. The control cell line MIHA is an immortalized normal human hepatocyte that expressed very little or undetectable CDH17. Primary HCC cell lines expressed some CDH17 transcript, whereas strong expression was seen in their metastatic counterparts (e.g. MHCC97H, MHCC97L, and H2-M) (Fig. 1B).

These cell line data were supported by results from 46 pairs of tumor and adjacent liver tissues from HCC patients who received curative surgery. Overexpression of CDH17 was strongly associated with advanced tumor stages (pathological tumor–node–metastasis III and IV) (P = 0.022) and tumor venous invasion (P = 0.022). No significant correlation was found for other clinico-pathologic parameters (Table 1). Most strikingly, copy number variation analysis revealed genomic amplification of the CDH17 gene in the tumor compared with adjacent nontumor tissues in 49% of HCC cases that were analyzed (n = 231) (Fig. 1C).

Table 1

Clinical Correlation Between CDH17 Messenger RNA Expression Level and Clinico-pathological Parameters of HCC Patients (n = 46)

Variables	Frequency (%)	CDH17 Overexpression		P Value

		−	+
Sex				0.116
Male	37 (80.4)	19	18
Female	9 (19.6)	2	7
Age				0.305
<60	34 (73.9)	14	20
≥60	12 (26.1)	7	5
Tumor size, cm				0.786
<5	8 (17.4)	4	4
≥5	38 (82.6)	17	21
Alpha fetoprotein, ng/mL				0.226
<250	24 (52.2)	13	11
≥250	22 (47.8)	8	14
Hepatitis B surface antigen				0.126
Negative	6 (1.0)	1	5
Positive	40 (87.0)	20	20
Histological differentiation^*				0.261
Well	16 (38.1)	9	7
Moderate/Poor	26 (61.9)	10	16
Pathological tumor–node–metastasis stage				0.022^†
Early (I, II)	18 (39.1)	12	6
Late (III, IV)	28 (60.9)	9	19
Venous infiltration				0.022^†
Absent	18 (39.1)	12	6
Present	28 (60.9)	9	19
Recurrence				0.161
Absent	35 (76.1)	18	17
Present	11 (23.9)	3	8
Cirrhosis				0.375
Negative	23 (50.0)	9	14
Positive	23 (50.0)	12	11

^{A total of 42 HCC tissues with defined clinical data were chosen for this correlation study.}

^P < 0.05.

Taken together, these results suggest that CDH17 is a candidate target for intervening in the initiation and metastasis of HCC. To test this principle, we used primary and metastatic HCC cell lines HuH-7, PLC, and MHCC97H, which express CDH17 most strongly and have high metastatic properties, to examine the clinical potential of reducing the expression of CDH17 RNA.

Targeting CDH17 Expression Alleviated Malignant Phenotypes in HCC Cells

shRNA in pcDNA and lentiviral vectors was employed to test the knockdown efficiency at different sites in the CDH17 open-reading frame. The target sequences in exon 3 and exon 5 yielded >50% reduction in both messenger RNA and protein levels (Supporting Fig. 1 and 2A). We then assessed the tumorigenic and metastatic properties (cell proliferation, colony formation, adhesion, invasion, and apoptosis) of the HCC cells either transduced with CDH17 shRNAmir (CDH17 shRNAmir cells) or transfected with CDH17 shRNA (CDH17 shRNA cells) compared with the control cells—the parental line transfected with vector (vector) or scrambled shRNA (Mock). CDH17 shRNAmir MHCC97H cells in culture showed a reduction in cell proliferation (by 49.6% at 6 days), substrate-adhering ability (by 62.7% at 45 minutes), cell migration ability (by 90.2% at 12 hours), colony-forming ability (by 93.1% at 14 days), and cell invasion ability (by 88.3% at 36 hours) (Fig. 2A–E). We also observed an increase in the number of apoptotic cells and in the amount of cleaved caspase-3 fragment (Fig. 2F). More CDH17 shRNAmir cells were arrested in G0/G1 phase, with fewer cells in S phase, whereas there was no significant alteration in cell numbers in G2/M-phase (Fig. 2G). A similar reduction in tumorigenicity was observed when we employed the shRNA knockdown method (Supporting Fig. 2B–E). The antitumor effect of CDH17 shRNA was also seen in two other primary HCC cell lines: PLC and HuH-7. Knockdown of CDH17 expression remarkably impaired their migration activities (Supporting Fig. 3). These results provide strong evidence that in vitro delivery of CDH17 shRNA into hepatic carcinoma cells could reduce the tumorigenicity of HCC.

An external file that holds a picture, illustration, etc.
Object name is nihms353854f2.jpg

Fig. 2

Association of CDH17 with tumorigenic and malignant phenotypes of MHCC97H cells. The expression of CDH17 was suppressed by lentiviral-mediated transduction of CDH17 shRNAmir. MHCC97H cells without any treatment (97H) and nontargeted RNAi vector-transduced MHCC97H cells (Mock) were used as controls. Different assays were performed to investigate the effects of CDH17 suppression as described in supplementary Methods. (A) Cell proliferation (P < 0.05). (B) Cell adhesion to laminin-1 substrate. (C) Cell migration. (D) Colony formation in soft agar. (E) Cell invasion. (F) Apoptosis and cleaved caspase-3 by way of immunoblotting. (G) Cell cycle analysis by way of flow cytometry. Knockdown CDH17 significantly impaired the tumorigenic and invasive properties of MHCC97H cells, while inducing cell cycle arrest and apoptosis. *P < 0.05. **P < 0.005. ***P < 0.005. Scale bars: 100 µm (C,E), 20 µm (D).

CDH17 as an In Vivo Target for Liver Cancer Therapy

We further examined the effect of CDH17 shRNA in liver tumors in vivo using a xenograft mouse model. Our initial observation showed that the tumorigenic potential of the primary HCC line PLC was significantly hampered when the CDH17 gene was silenced. We then moved to test the CDH17 knockdown effect on MHCC97H cells, which expresses the highest transcript level of CDH17. At 8 weeks after tumor cell inoculation, large tumors were seen in the control groups, but the tumor volume was still minimal in those mice transplanted with the CDH17 shRNAmir (Fig. 3A) or CDH17 shRNA (Fig. 3B) MHCC97H cells. Of great interest, all control animals (Mock group) developed metastasis in the lungs as shown by the presence of green fluorescent protein–positive tumor cells at 8 weeks regardless of whether the MHCC97H tumor cells were transplanted subcutaneously or injected systemically. By contrast, no lung metastasis was observed in animals from the CDH17 shRNAmir treatment group (Fig. 4). Thus, knockdown of CDH17 not only reduced tumor growth but also diminished the metastatic potential of hepatic carcinoma. This has a great clinical implication, because most of the HCC patients die from tumor recurrence due to intrahepatic or extrahepatic metastasis.

An external file that holds a picture, illustration, etc.
Object name is nihms353854f3.jpg

Fig. 3

CDH17 knockdown in MHCC97H cells reduced their tumorigenicity in vivo. (A) MHCC97H cells transduced with CDH17 shRNAmir (shRNAmir) or nontargeted RNAi vector (Mock) and MHCC97H cells without treatment (97H) were transplanted into nude mice in a tumor xenograft model. Whole-body fluorescent imaging was performed to reveal the tumors (left panel). Once a week, four mice were killed and the subcutaneous tumors were removed. Tumor size and volume were photographed, and tumor volume was measured at each time point. A line chart illustrates the changes in volume of the subcutaneous tumors (right panel). (B) Effect of CDH17 shRNA suppression on the growth of subcutaneous tumors in nude mice. Subcutaneous tumors were induced in nude mice by injecting native (97H), scramble clone-transfected (Mock), or CDH17 shRNA-transfected (shRNA) MHCC97H cells (n = 3). Photographs illustrate the subcutaneous tumor produced by Mock and shRNA cells in nude mice. The volume of tumor induced in nude mice was measured among the three experimental groups after 35 days. *P = 0.0002.

An external file that holds a picture, illustration, etc.
Object name is nihms353854f4.jpg

Fig. 4

Reduced lung metastatic potentials of MHCC97H cells after transduction with CDH17 shRNAmir. MHCC97H cells transduced with CDH17 shRNAmir (shRNAmir) or nontargeted RNAi vector (Mock), were either (A) inoculated subcutaneously or (B) injected through the tail vein in nude mice (n =5). Eight weeks after treatment, frozen sections were prepared from the lungs of these mice and stained with DAPI. Lung metastasis was determined based on the presence of green fluorescent protein–positive tumor cells in the lungs of the nude mice. Scale bar: 150 µm.

We then investigated whether in vivo delivery of CDH17 shRNAmir could impede the growth of an established tumor xenograft derived from parental MHCC97H cells in nude mice. At 1 week after tumor inoculation, when the tumor had reached approximately 0.5 cm in diameter, the animals were injected at the tumor site with placebo (Tris-buffered saline), mock reagents, rAAV-TP53, CDH17 shRNAmir, or combined rAAV-TP53 and CDH17 shRNAmir regimen. Mice in the control groups (97H/placebo and mock) grew large, solid, vascularized tumors, whereas the groups treated with rAAV-TP53 or CDH17 shRNAmir developed much smaller subcutaneous tumors. A further reduction in tumor size was observed when CDH17 shRNAmir and rAAV-TP53 were used in combination (Fig. 5; Supporting Fig. 4). No extensive coverage of blood vessels was observed in the tumors. This shows that knockdown of CDH17 is an effective means of shrinking tumor growth of HCC in a tumor-bearing mouse model. On the other hand, no distortion or rupture of tissue architecture was noticed in other organs, including the brain, heart, kidney, liver, lungs, and stomach of the animals treated as above (Supporting Fig. 5A). There was no significant difference in the expression levels of three genes (2’–5’-oligoadenylate synthetase 1, interleukin-8, protein-kinase, interferon-inducible double-stranded RNA–dependent inhibitor, repressor of [P58 repressor]) associated with virus infection and neutrophilic granulocyte-related inflammation due to the use of lentiviral vector for delivery in the subcutaneous tumors among the five experimental groups (Supporting Fig. 5B).

An external file that holds a picture, illustration, etc.
Object name is nihms353854f5.jpg

Fig. 5

Targeting CDH17 for liver cancer therapy. Subcutaneous tumors were induced in nude mice by injecting MHCC97H cells. After 1 week, different treatments were applied directly to the tumors: injections of Tris-buffered saline (97H), lentivirus carrying a nontargeted RNAi vector (Mock), recombinant adeno-associated virus expressing TP53 (rAAV-TP53), lentivirus carrying CDH17 shRNAmir (shRNAmir), and combined treatment with shRNAmir and rAAV-TP53. Mice received the specific treatment twice weekly for 2 weeks. Five mice were used for each group. (A) Representative photograph showing the morphology of the subcutaneous tumors derived from different groups. (B) The changes in volume of the subcutaneous tumors of different groups.

Knockdown of CDH17 Inactivated Wnt/Catenin Signaling Pathway

Our preliminary observation indicates CDH17-mediated oncogenic signaling associated with the Wnt/catenin pathway. In a TOPFlash reporter luciferase assay (Fig. 6A), knockdown of CDH17 by shRNA in MHCC97H cells significantly decreased the strength of TCF/LEF signals compared with the vector and mock controls. Furthermore, as shown by way of immunoblot analysis of CDH17 shRNAmir cells, we detected a significant reduction of both the total and nuclear β-catenin protein, but no change in the p53 level, when compared with the 97H and Mock controls (Fig. 6B). Then we evaluated the involvement of other Wnt pathway signaling components. Knockdown of CDH17 resulted in down-regulation of the phosphorylated glycogen synthase kinase (GSK)-3β, β-catenin, and cyclin D1 proteins, but the tumor suppressor retinoblastoma (Rb) level was increased compared with the rAAV-TP53, 97H, and mock controls (Fig. 6C). In addition, rAAV-TP53 enhanced Bax expression but suppressed Bcl-XL expression, whereas CDH17 shRNAmir did not. Nevertheless, both rAAV-TP53 and CDH17 shRNAmir enhanced caspase-3 cleavage, and there was an additive effect on the pro-apoptotic mechanism (caspase 3, Bax, and Bcl-XL levels) in the combined shRNAmir & rAAV-TP53 regimen.

An external file that holds a picture, illustration, etc.
Object name is nihms353854f6.jpg

Fig. 6

Suppression of the β-catenin/Wnt pathway in MHCC97H cells in vitro and in vivo by knockdown of CDH17. (A) A TOPFlash reporter luciferase assay was performed using MHCC97H cells transfected with (shRNA) or without (Vector and Mock) CDH17 shRNA. Luciferase intensities for each experimental group were quantified using a luminometer. The relative luciferase activities were calculated when the luciferase intensities of the pRL-TK-Luc vector were taken into account. **P < 0.0001. (B) Western blots of β-catenin and p53 in MHCC97H cells infected with (shRNAmir) or without (97H and Mock) CDH17 shRNAmir. (C) For the in vivo study, subcutaneous tumors were induced in nude mice and CDH17- and p53-associated gene therapies were applied to the tumors directly. The effects on the expression of a panel of proteins associated with the Wnt pathway and apoptosis were revealed by way of western blotting for each treatment group. (D) Tumors were subsequently dissected, then stained immuno-histochemically using antibodies specific for CDH17, β-catenin, cyclin D1, and Rb proteins. Immunoglobulin G was used to substitute the antibody to perform the immunohistochemistry to reveal the authenticity of the staining. (Original magnification ×400.)

To confirm the effect of CDH17 knockdown on the Wnt pathway in vivo, we dissected MHCC97H xenograft tumors from the animals and conducted immunohistochemical analyses on the tissues. Consistent with the above findings, we observed relocalization of β-catenin to the cytoplasm, with concomitant reduction in cyclin D1 and increase in Rb in HCC tumors treated with CDH17 shRNA (Fig. 6D).

Tumorigenic and Metastatic Properties of CDH17 in HCC

Fig. 1

Table 1

Clinical Correlation Between CDH17 Messenger RNA Expression Level and Clinico-pathological Parameters of HCC Patients (n = 46)

Variables	Frequency (%)	CDH17 Overexpression		P Value

		−	+
Sex				0.116
Male	37 (80.4)	19	18
Female	9 (19.6)	2	7
Age				0.305
<60	34 (73.9)	14	20
≥60	12 (26.1)	7	5
Tumor size, cm				0.786
<5	8 (17.4)	4	4
≥5	38 (82.6)	17	21
Alpha fetoprotein, ng/mL				0.226
<250	24 (52.2)	13	11
≥250	22 (47.8)	8	14
Hepatitis B surface antigen				0.126
Negative	6 (1.0)	1	5
Positive	40 (87.0)	20	20
Histological differentiation^*				0.261
Well	16 (38.1)	9	7
Moderate/Poor	26 (61.9)	10	16
Pathological tumor–node–metastasis stage				0.022^†
Early (I, II)	18 (39.1)	12	6
Late (III, IV)	28 (60.9)	9	19
Venous infiltration				0.022^†
Absent	18 (39.1)	12	6
Present	28 (60.9)	9	19
Recurrence				0.161
Absent	35 (76.1)	18	17
Present	11 (23.9)	3	8
Cirrhosis				0.375
Negative	23 (50.0)	9	14
Positive	23 (50.0)	12	11

^{A total of 42 HCC tissues with defined clinical data were chosen for this correlation study.}

^P < 0.05.

Discussion

The CDH17 oncogene is an attractive therapeutic target for HCC, because it is highly expressed in tumor tissues but not in the normal liver. Over 80% of HCCs are CDH17-positive, and half of these patients reveal genomic copy gain of this gene. Our previous work first demonstrated the overexpression and prognostic significance of CDH17 associated with poor overall survival and disease-free survival times in HCC patients.⁷,11 A high CDH17 level has also been reported in a stem-like progenitor HCC subpopulation driven by c-Met, which exhibited high metastasis of liver tumors.²⁶ In the present study, we provide new evidence correlating high expression of CDH17 and HCC tumorigenesis, showing the oncogenic properties of CDH17 in a genetically defined mouse model. The expression of CDH17 in HCC cell lines is positively correlated with metastatic potential, and in HCC patients a high level of CDH17 clinically correlates with tumor venous invasion, a strong risk factor for cancer metastasis. Furthermore, HCC patients at advanced stages usually have high expression of CDH17 in the tumors, which is often associated with aggressive and malignant phenotypes of the cancer.

Today, there is accumulating evidence that CDH17 is a disease marker for other gastrointestinal malignancies.²⁷,28 For instance, high expression of CDH17 is associated with high metastatic potential, positive lymph node metastasis and short overall survival in gastric cancer patients.¹⁶,29 However, down-regulation of CDH17 is found in human colorectal cancers and its reduced expression indicates the presence of lymph node metastasis.¹² The difference in the expression pattern of CDH17 between HCC and colon cancer are perhaps context-dependent. For instance, CDH17 may support normal physiology and/or epithelial integrity in colon tissue, while aberrant expression of CDH17 may drive oncogenesis in liver and gastric cancers. Furthermore, we identify different isoforms of CDH17 in HCC, but there is no information on whether the wild-type or splice variants are present in the colon.

The present study demonstrates the use of shRNA vectors to knock down CDH17 expression in a highly tumorigenic and metastatic HCC cell line, MHCC97H, rendering the cells less tumorigenic in vitro in terms of cell proliferation, cell adhesion, cell migration, cell invasion, and anchorage-independent colony formation. This was confirmed in vivo when constitutive knockdown of CDH17 expression in MHCC97H cells markedly impaired growth of a tumor xenograft. In tumor-bearing mice, local delivery of lentiviral-based CDH17 shRNAmir vector inhibited tumor growth, and further reduction was seen after combination with rAAV-TP53 treatment, showing an additive antitumor effect on HCC.

Although we focus on targeting CDH17 as a novel therapeutic approach in HCC, deficiency or loss of tumor suppressor p53 is common in various cancers, including HCC, for which restoration of the TP53 gene may induce tumor apoptosis.³⁰,31 Furthermore, a 249(ser) TP53 mutation has been found in ≈50% of Asian patients associated with poor clinical outcomes.³² This mutation is also found in the metastatic MHCC97H cell line used in this study (data not shown). Our data revealed no direct link between CDH17 and TP53, because there was no obvious change in the level or localization of p53 when CDH17 expression was suppressed, or vice versa. They may be involved in distinct pathways during the development and progression of HCC, although the loss of p53 and deregulation of the Wnt pathway (as reflected by β-catenin nuclear localization) frequently occur in HCC. Therefore, targeting these two molecules could be used to develop a personalized treatment strategy for those patients with high CDH17 and deficient p53 in their tumors, particularly at the late advanced stages for which there are no effective treatments.

Altered expression levels of phosphorylated GSK, cyclin D1, and Rb were found in tumor xenografts with reduced CDH17 expression. GSK is a protein kinase associated with the destruction complex of β-catenin, consisting of at least adenomatous polyposis coli and axin, whereas cyclin D1 is a Wnt-responsive target gene. Loss or down-regulation of Rb expression in HCC was shown to be associated with poor cell differentiation and metastasis.³³ Herein, CDH17 knockdown in HCC decreased GSK-3β phosphorylation, accompanied by a concomitant reduction of cyclin D1 and induction of Rb. Moreover, CDH17 knockdown in MHCC97H cells led to cytoplasmic sequestration (or nuclear extravasation) and potentially degradation of β-catenin, which subsequently reduced TCF/LEF transactivation activity. In the normal hepatocyte cell line MIHA, our studies revealed that overexpression of CDH17 induced epithelial-to-mesenchymal transition and enhanced cyclooxygenase 2 activities (unpublished data). In agreement with our observations, prostaglandin 2 was recently shown to regulate the Wnt signaling in vivo.³⁴ Furthermore, a recent report showed a trans-interaction between E-cadherin and CDH17 in enterocytes during development of the intestinal epithelium,³⁵,36 suggesting that CDH17 might intersect with the Wnt pathway through its coordination with E-cadherin and/or associated partners. Together, these observations point to a potential oncogenic role for CDH17 in HCC, eventually leading to stabilization, nuclear shuffling, and localization of β-catenin (Fig. 7). In the proposed model, we hypothesize that CDH17 may (1) destabilize the cadherin–catenin network through interaction with E-cadherin, unleashing the β-catenin molecule from cytoplasmic membrane, (2) inactivate the GSK-3β activities and prevent cytosolic β-catenin from degradation, and (3) up-regulate the cyclooxygenase 2 pathway, thereby activating Wnt signaling.

An external file that holds a picture, illustration, etc.
Object name is nihms353854f7.jpg

Fig. 7

Schematic diagram illustrating the proposed interaction between CDH17 and Wnt/β-catenin pathway during HCC progression. The Wnt/β-catenin pathway is frequently activated during tumorigenesis. Genomic amplification and/or overexpression of CDH17 induces cadherin switch and initiates epithelial-mesenchymal transition, probably through destabilization of the E-cadherin/catenin complex, inhibition of GSK-3β/Axin-based proteosome activity, and activation of cyclooxygenase 2/cyclic adenosine monophosphate/protein kinase A signaling pathway. Thus, nuclear shuffling and accumulation of β-catenin would result in LEF/TCF-mediated transcription of Wnt-responsive genes, such as cyclin D1 leading to HCC cell proliferation and metastasis. Knockdown CDH17 expression by shRNA both in vitro and in vivo could inhibit tumor growth and enhance cell death, and the underlying antitumor mechanism involves relocalization of β-catenin to the cytoplasm and a concomitant reduction in cyclin D1 and an increase in Rb.

In conclusion, the present study presents a novel therapeutic approach to liver cancer therapy by targeting CDH17 alone or in combination with p53, and suggests beneficial effects of targeting multiple pathways. This study also provides the first direct evidence of CDH17 as a bona fide oncogene, with tumorigenic properties giving rise to aggressive phenotypes of HCC.

Supplementary Material

Supp Information

Click here to view.^{(52K, doc)}

02

Click here to view.^{(1.0M, pdf)}

Supp Information

Click here to view.^{(52K, doc)}

02

Click here to view.^{(1.0M, pdf)}

Acknowledgments

Supported by General Research Fund Grant HKU771607M from the Research Grants Council of Hong Kong (to J. M. L.). I. O. N. is Loke Yew Professor in Pathology and is supported by Collaborative Research Fund Grant HKU1/06C. S. W. L. is a Howard Hughes Medical Institute investigator and is supported by National Cancer Institute Grant CA13106.

^{Department of Surgery and Center for Cancer Research, The University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong}

^{Department of Pathology, The University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong}

^{Department of Radiology, Zhongshan Hospital, Fudan University, Shanghai, China}

^{Howard Hughes Medical Institute and Cold Spring Harbor Laboratory, Cold Spring Harbor, NY}

^{Rosetta Inpharmatics, LLC, Seattle, WA}

^{Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Shatin, Hong Kong}

^{State Key Laboratory of Oncology in Southern China and Stanley Ho Center for Emerging Infectious Diseases, The Chinese University of Hong Kong, Shatin, Hong Kong}

Address reprint requests to: John Luk, Department of Surgery, The University of Hong Kong, L9-52, 21 Sassoon Road, Pokfulam, Hong Kong. kh.ukh@kulmj; fax: (852)-2974-1389; or Jianhua Wang, Department of Radiology, Zhongshan Hospital, Fudan University, Shanghai, 200032 China; nc.hs.latipsoh-sz@auhnaij.gnaw

Current address for Mao Mao: Oncology Research, Pfizer Inc., San Diego, CA.

^{These authors contributed equally to this work.}

Abstract

Hepatocellular carcinoma (HCC) is a lethal malignancy for which there are no effective therapies. To develop rational therapeutic approaches for treating this disease, we are performing proof-of-principle studies targeting molecules crucial for the development of HCC. Here, we show that cadherin-17 (CDH17) adhesion molecule is up-regulated in human liver cancers and can transform premalignant liver progenitor cells to produce liver carcinomas in mice. RNA interference–mediated knockdown of CDH17 inhibited proliferation of both primary and highly metastatic HCC cell lines in vitro and in vivo. The antitumor mechanisms underlying CDH17 inhibition involve inactivation of Wnt signaling, because growth inhibition and cell death were accompanied by relocalization of β-catenin to the cytoplasm and a concomitant reduction in cyclin D1 and an increase in retinoblastoma.

Conclusion

Our results identify CDH17 as a novel oncogene in HCC and suggest that CDH17 is a biomarker and attractive therapeutic target for this aggressive malignancy.

Abstract

Hepatocellular carcinoma (HCC) is one of the most prevalent and lethal malignancies. It is the fifth most common cancer and ranks as the third leading cause of cancer-related deaths worldwide.¹ The number of incident cases is over 600,000 per year, which is almost the same as the number of deaths, owing to the difficulty in early detection and high postsurgical recurrence rate. The prognosis of HCC is extremely poor, and only ≈5% of patients survive more than 5 years.² Patients afflicted with HCC are often asymptomatic, and the lack of sensitive and reliable biomarkers for early detection of HCC and cancer surveillance in at-risk populations (hepatitis B or C virus carriers and those with cirrhosis) means that diagnosis normally occurs late, when surgical intervention is not an option.³ Today, there are no effective drugs for curing liver cancer.

Using integrative genomic and proteomic approaches in mouse models,⁴–6 we have begun to identify novel oncogenes and tumor suppressors in liver cancer. Our previous studies using clinical cohorts identified a cell surface adhesion molecule, cadherin-17 (CDH17), also known as liver-intestine cadherin, as a potential disease marker for HCC.⁷ It consists of seven cadherin-like ectodomains and a short cytoplasmic tail of 24 amino acid residues.⁸ Expression of CDH17 is restricted to the colon, intestine, and pancreas in humans; it is not found in the healthy adult liver and stomach.⁷,9–12 It plays an important role during embryonic gastrointestinal development and also functions as a peptide transporter.¹³,14 Most, if not all, cadherin molecules interact with the cytosolic β-catenin network and thereby regulate the Wnt signaling pathway, but there are no published reports that CDH17 does so.

Increased CDH17 expression has been reported in liver and stomach cancer.⁷,9,15,16 In our earlier studies, we identified an isoform that lacks exon 7 and correlates with poor prognostic outcomes in HCC patients.¹¹ Furthermore, this specific CDH17 haplotype is associated with increased risk of HCC in Chinese subjects.¹⁷ Despite these significant clinical findings, the molecular pathogenesis of CDH17 remains unknown, and its tumorigenic role in HCC has not yet been confirmed. Here, we aimed to dissect the oncogenic signaling mechanisms of CDH17 in the HCC context and evaluated the feasibility of targeting CDH17 using RNA interference (RNAi) as a potential therapeutic approach for HCC.

Abbreviations

CDH17	cadherin-17
CNV	copy number variation
GSK	glycogen synthase kinase
HCC	hepatocellular carcinoma
Rb	retinoblastoma
RNAi	RNA interference
shRNA	short hairpin RNA

Abbreviations

Footnotes

Potential conflict of interest: Nothing to report.

Additional Supporting Information may be found in the online version of this article.

Footnotes

References

1. Parkin DM, Bray F, Ferlay J, Pisani PGlobal cancer statistics, 2002. CA Cancer J Clin. 2005;55:74–108.[PubMed][Google Scholar]
2. Poon RT, Fan STHepatectomy for hepatocellular carcinoma: patient selection and postoperative outcome. Liver Transpl. 2004;10(Suppl):S39–S45.[PubMed][Google Scholar]
3. El-Serag HB, Marrero JA, Rudolph L, Reddy KRDiagnosis and treatment of hepatocellular carcinoma. Gastroenterology. 2008;134:1752–1763.[PubMed][Google Scholar]
4. Zender L, Xue W, Zuber J, Semighini CP, Krasnitz A, Ma B, et al An oncogenomics-based in vivo RNAi screen identifies tumor suppressors in liver cancer. Cell. 2008;135:852–864.[Google Scholar]
5. Lee NP, Leung KW, Cheung N, Lam BY, Xu MZ, Sham PC, et al Comparative proteomic analysis of mouse livers from embryo to adult reveals an association with progression of hepatocellular carcinoma. Proteomics. 2008;8:2136–2149.[PubMed][Google Scholar]
6. Zender L, Spector MS, Xue W, Flemming P, Cordon-Cardo C, Silke J, et al Identification and validation of oncogenes in liver cancer using an integrative oncogenomic approach. Cell. 2006;125:1253–1267.[Google Scholar]
7. Wong BW, Luk JM, Ng IO, Hu MY, Liu KD, Fan STIdentification of liver-intestine cadherin in hepatocellular carcinoma—a potential disease marker. Biochem Biophys Res Commun. 2003;311:618–624.[PubMed][Google Scholar]
8. Wendeler MW, Jung R, Himmelbauer H, Gessner RUnique gene structure and paralogy define the 7D-cadherin family. Cell Mol Life Sci. 2006;63:1564–1573.[PubMed][Google Scholar]
9. Grotzinger C, Kneifel J, Patschan D, Schnoy N, Anagnostopoulos I, Faiss S, et al LI-cadherin: a marker of gastric metaplasia and neoplasia. Gut. 2001;49:73–81.[Google Scholar]
10. Takamura M, Sakamoto M, Ino Y, Shimamura T, Ichida T, Asakura H, et al Expression of liver-intestine cadherin and its possible interaction with galectin-3 in ductal adenocarcinoma of the pancreas. Cancer Sci. 2003;94:425–430.[PubMed][Google Scholar]
11. Wang XQ, Luk JM, Leung PP, Wong BW, Stanbridge EJ, Fan STAlternative mRNA splicing of liver intestine-cadherin in hepatocellular carcinoma. Clin Cancer Res. 2005;11:483–489.[PubMed][Google Scholar]
12. Takamura M, Ichida T, Matsuda Y, Kobayashi M, Yamagiwa S, Genda T, et al Reduced expression of liver-intestine cadherin is associated with progression and lymph node metastasis of human colorectal carcinoma. Cancer Lett. 2004;212:253–259.[PubMed][Google Scholar]
13. Berndorff D, Gessner R, Kreft B, Schnoy N, Lajous-Petter AM, Loch N, et al Liver-intestine cadherin: molecular cloning and characterization of a novel Ca^{-dependent cell adhesion molecule expressed in liver and intestine.}J Cell Biol. 1994;125:1353–1369.[Google Scholar]
14. Dantzig AH, Hoskins JA, Tabas LB, Bright S, Shepard RL, Jenkins IL, et al Association of intestinal peptide transport with a protein related to the cadherin superfamily. Science. 1994;264:430–433.[PubMed][Google Scholar]
15. Park SS, Kang SH, Park JM, Kim JH, Oh SC, Lee JH, et al Expression of liver-intestine cadherin and its correlation with lymph node metastasis in gastric cancer: can it predict N stage preoperatively? Ann Surg Oncol. 2007;14:94–99.[PubMed][Google Scholar]
16. Ko S, Chu KM, Luk JM, Wong BW, Yuen ST, Leung SY, et al Overexpression of LI-cadherin in gastric cancer is associated with lymph node metastasis. Biochem Biophys Res Commun. 2004;319:562–568.[PubMed][Google Scholar]
17. Wang XQ, Luk JM, Garcia-Barcelo M, Miao X, Leung PP, Ho DW, et al Liver intestine-cadherin (CDH17) haplotype is associated with increased risk of hepatocellular carcinoma. Clin Cancer Res. 2006;12:5248–5252.[PubMed][Google Scholar]
18. Hu L, Wen JM, Sham JS, Wang W, Xie D, Tjia WM, et al Establishment of cell lines from a primary hepatocellular carcinoma and its metastatis. Cancer Genet Cytogenet. 2004;148:80–84.[PubMed][Google Scholar]
19. Li Y, Tang ZY, Ye SL, Liu YK, Chen J, Xue Q, et al Establishment of cell clones with different metastatic potential from the metastatic hepatocellular carcinoma cell line MHCC97. World J Gastroenterol. 2001;7:630–636.[Google Scholar]
20. Brown JJ, Parashar B, Moshage H, Tanaka KE, Engelhardt D, Rabbani E, et al A long-term hepatitis B viremia model generated by transplanting nontumorigenic immortalized human hepatocytes in Rag-2-deficient mice. Hepatology. 2000;31:173–181.[PubMed][Google Scholar]
21. Luk MC, Tsang RS, Ng MHMurine monoclonal antibody specific for lipopolysaccharide of Salmonella serogroup A. J Clin Microbiol. 1987;25:2140–2144.[Google Scholar]
22. Chen Y, Lin MC, Yao H, Wang H, Zhang AQ, Yu J, et al Lentivirus-mediated RNA interference targeting enhancer of zeste homolog 2 inhibits hepatocellular carcinoma growth through down-regulation of stathmin. HEPATOLOGY. 2007;46:200–208.[PubMed][Google Scholar]
23. Lee NP, Leung KW, Wo JY, Tam PC, Yeung WS, Luk JMBlockage of testicular connexins induced apoptosis in rat seminiferous epithelium. Apoptosis. 2006;11:1215–1229.[PubMed][Google Scholar]
24. Shah SP, Lam WL, Ng RT, Murphy KPModeling recurrent DNA copy number alterations in array CGH data. Bioinformatics. 2007;23:i450–i458.[PubMed][Google Scholar]
25. Zender L, Xue W, Cordon-Cardo C, Hannon GJ, Lucito R, Powers S, et al Generation and analysis of genetically defined liver carcinomas derived from bipotential liver progenitors. Cold Spring Harb Symp Quant Biol. 2005;70:251–261.[Google Scholar]
26. Kaposi-Novak P, Lee JS, Gomez-Quiroz L, Coulouarn C, Factor VM, Thorgeirsson SSMet-regulated expression signature defines a subset of human hepatocellular carcinomas with poor prognosis and aggressive phenotype. J Clin Invest. 2006;116:1582–1595.[Google Scholar]
27. Kwak JM, Min BW, Lee JH, Choi JS, Lee SI, Park SS, et al The prognostic significance of E-cadherin and liver intestine-cadherin expression in colorectal cancer. Dis Colon Rectum. 2007;50:1873–1880.[PubMed][Google Scholar]
28. Ge J, Chen Z, Wu S, Yuan W, Hu BA clinicopathological study on the expression of cadherin-17 and caudal-related homeobox transcription factor (CDX2) in human gastric carcinoma. Clin Oncol (R Coll Radiol) 2008;20:275–283.[PubMed][Google Scholar]
29. Ito R, Oue N, Yoshida K, Kunimitsu K, Nakayama H, Nakachi K, et al Clinicopathological significant and prognostic influence of cadherin-17 expression in gastric cancer. Virchows Arch. 2005;447:717–722.[PubMed][Google Scholar]
30. Dey A, Verma CS, Lane DPUpdates on p53: modulation of p53 degradation as a therapeutic approach. Br J Cancer. 2008;98:4–8.[Google Scholar]
31. Wang W, El-Deiry WSRestoration of p53 to limit tumor growth. Curr Opin Oncol. 2008;20:90–96.[PubMed][Google Scholar]
32. Pineau P, Marchio A, Battiston C, Cordina E, Russo A, Terris B, et al Chromosome instability in human hepatocellular carcinoma depends on p53 status and aflatoxin exposure. Mutat Res. 2008;653:6–13.[PubMed][Google Scholar]
33. Edamoto Y, Hara A, Biernat W, Terracciano L, Cathomas G, Riehle HM, et al Alterations of RB1, p53 and Wnt pathways in hepatocellular carcinomas associated with hepatitis C, hepatitis B and alcoholic liver cirrhosis. Int J Cancer. 2003;106:334–341.[PubMed][Google Scholar]
34. Goessling W, North TE, Loewer S, Lord AM, Lee S, Stoick-Cooper CL, et al Genetic interaction of PGE2 and Wnt signaling regulates developmental specification of stem cells and regeneration. Cell. 2009;136:1136–1147.[Google Scholar]
35. Baumgartner W, Wendeler MW, Weth A, Koob R, Drenckhahn D, Gessner RHeterotypic trans-interaction of LI- and E-cadherin and their localization in plasmalemmal microdomains. J Mol Biol. 2008;378:44–54.[PubMed][Google Scholar]
36. Wendeler MW, Drenckhahn D, Gessner R, Baumgartner WIntestinal LI-cadherin acts as a Ca2+-dependent adhesion switch. J Mol Biol. 2007;370:220–230.[PubMed][Google Scholar]