High expression levels of IKKα and IKKβ are necessary for the malignant properties of liver cancer

Abstract

IKK-NF-κB signaling is regarded as an important factor in hepatocarcinogenesis and a potential target for liver cancer therapy. Therefore, in this study, we analyzed the expression of mRNAs encoding components and targets of NF-κB signaling including IKKα, IKKβ, RANK, RANKL, OPG, CyclinD3, mammary serine protease inhibitor (Maspin), CyclinD1, c-FLIP, Bcl-xl, Stat3, Cip1 and Cip2 by real-time PCR in 40 patients with liver cancer. After statistical analysis, 7 indices including IKKα, IKKβ, RANK, Maspin, c-FLIP, Cip2 and cyclinD1 were found to show significant differences between tumor tissue and its corresponding adjacent tissue. When IKKα and IKKβ were downregulated in the hepatocellular carcinoma (HCC) cell lines of MHCC-97L and MHCC-97H in vitro, the numbers of BrdU positive cells were decreased in both IKKα and IKKβ knockdown cells. Levels of apoptosis were also investigated in IKKα and IKKβ knockdown cells. The growth of HCC was inhibited in the subcutaneous implantation model, and lung metastatogenesis was also significantly inhibited in the kidney capsule transplantation model. Downregulation of IKKα and IKKβ in HCC cultured in vitro revealed that increased Maspin, OPG and RANKL expression was associated with metastasis of HCC. These findings were associated with downregulation of Bcl-XL and c-FLIP, which may be the reason for increased apoptosis. The therapeutic effect of IKKα and IKKβ downregulation depends on extent of NF-κB inhibition and the malignant nature of the HCC. We anticipate that IKK-targeted gene therapy can be used in the treatment of HCC, a cancer that is notoriously resistant to radiation and chemotherapy.

Hepatocellular carcinoma (HCC) is the final outcome of chronic hepatitis and is a major focus of epidemiological and clinical research. Recently, studies using mouse models have clarified some of the mechanisms involved in the development of HCC. The transcription nuclear factor (NF)-κB acts as a potential master control of tumorigenesis and HCC development. In the early stages of tumorigenesis, the cytoprotective effect of NF-κB prevents hepatocyte cell death and inhibits compensatory hepatocyte proliferation, which may promote tumor development in the late stages of tumorigenesis. NF-κB supports malignancy by promoting survival of transformed hepatocytes. In the nonparenchymal cells of the liver, NF-κB activation during hepatocarcinogenesis is detrimental, because it provides the tumor cells with essential growth and survival factors, allowing them to continue proliferating.1–5

NF-κB can function as both a promoter and suppressor of hepatocellular carcinogenesis. When acting as a suppressor, inhibiting activation of NF-κB by a specific knockout of IKKβ in hepatocytes results in increased number and size of hepatocarcinomas induced by diethylnitrosamine.3 In addition, mice with a specific deletion of IKKγ/Nemo, the regulatory subunit of the IKK complex in hepatocytes and cholangiocytes, develops spontaneous chronic hepatitis resembling human nonalcoholic steatohepatitis, which eventually progresses to HCC. Both carcinogenic processes were shown to be caused by increased oxidative stress in NF-κB deficient hepatocytes and were consequently inhibited by antioxidants.1 However, NF-κB also acts as a tumor promoter in hepatocellular carcinogenesis. Switching off NF-κB in mice from birth to 7 months of age, using a hepatocyte-specific inducible IkappaB-super-repressor transgene, had no effect on the course of hepatitis, nor did it affect the early phases of hepatocyte transformation. By contrast, suppressing NF-κB through anti–TNF-alpha treatment or induction of IkappaB-super-repressor in the later stages of tumor development resulted in apoptosis of transformed hepatocytes and failure to progress to HCC. The studies, thus, indicated that NF-κB was essential for promoting inflammation-associated cancer and was, therefore, a potential target for cancer prevention in chronic inflammatory diseases.6

Compared with IKKβ and IKKγ, the role of IKKα seems to be less important in protecting against liver injury and tumorigenesis.7, 8 Furthermore, there is only limited information about the role of IKKα in tumorigenesis, other than its involvement in the metastatic progression of prostate cancer.9 Previous studies have shown that activation of IKKα by set unknown factors resulted in repression of the metastasis suppressor Maspin, thereby allowing metastatic spread of prostate cancer. IKKα is also involved in mammary gland development where it may be activated by the TNF family member, RANKL.10 Not surprisingly, IKKα is also important for the renewal capacity of tumor progenitors in mammary carcinoma.11

Intervention with NF-κB signaling has been considered as a potential therapeutic approach in cancer.12–15 Recently, a novel drug, KINK-1, a small-molecule inhibitor of IKKβ, has shown promise in the treatment of melanoma.16 This drug increases tumor susceptibility to chemotherapy by inhibiting IKKβ and NF-κB activities. Studies have also shown that osteoprotegerin-FC, a receptor that activates NF-κB signaling and an antagonist of RNAKL, may be an effective inhibitor of bone metastasis in prostate cancer.17 We found that both IKKα and IKKβ are highly expressed in human HCC compared with adjacent normal tissue. Therefore, we examined the effect of IKKα and IKKβ downregulation on the growth of 2 HCC cell lines, MHCC-97L and MHCC-97H.

Material and Methods

Lentivirus production and transduction

The cassette of H1-unrelated-SiRNA, H1-IKKα-SiRNA or H1-IKKβ-SiRNA was excised from pSUPER-SK-unrelated-SiRNA, pSUPER-SK-IKKα-SiRNA and pSUPER-SK-IKKβ-SiRNA plasmids (Gift from Dr. Michael Karin, University of California, San Diego, La Jolla, CA), respectively, by restrictive digestion using SalI and BamHI (New England Biolab, United Kingdom) and subcloned into the site of SalI and BamHI in an HIV type-1 construct, pWPTS-GFP (gift from D. Trono, University of Geneva, Switzerland),18 replacing the GFP gene. Subsequently, an elongation factor-1α promoter and GFP cassette were inserted into the SalI site and named pLenti-GFP-control-SiRNA, pLenti-GFP-IKKα-SiRNA or pLenti-GFP-IKKβ-SiRNA. Recombinant lentivirus was generated from 293T cells.19 Human HCC cell lines MHCC97L and MHCC97H were provided by Dr. Z.Y. Tang, Liver Cancer Institute of Fudan University (Shanghai, China), and grown in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum, 100 U/ml penicillin and 100 μg/ml streptomycin (Gibco, CA), which was tranduced with lentivirus using polybrene (8 μg/ml). A pure population of transduced cells was sorted for GFP-expression by flow cytometry; more than 98% of the cells were fluorescent after sorting.

Animal model

CBYJ.CG-FOXNL^NU/J nude mice and B6.129S7-RAG1^tm1Mom/J RAG1 mice were purchased from Nanjing University (Nanjing, China). Animals (both males and females) at least 6 weeks of age were used in these experiments. Procedures were performed in accordance with the guidelines established by the Institutional Animal Care and Use Committee at Nanjing Medical University. For the subcutaneous injection model, 30 nude mice were randomly divided into 6 groups (5 mice in each group). Lentiviral modified cells (5 × 10⁶) including MHCC97L-pLenti-control-SiRNA, MHCC97L-IKKα-SiRNA, MHCC97L-IKKβ-SiRNA, MHCC97H-pLenti-control-SiRNA, MHCC97H-IKKα-SiRNA and MHCC97H-IKKβ-SiRNA were subcutaneously injected into the nude mice. The mice were examined twice a week, and tumor growth was evaluated by measuring the length and width of tumor mass until natural death or 9 weeks after first injection, and the tumor tissues were analyzed using immunohistochemistry, histology and a whole body Lightool fluorescent system. For the kidney capsule transplantation assay, 18 RAG1 mice were randomly divided into 3 groups, lentiviral modified cells (2 × 10⁶) including MHCC97H-pLenti-control-SiRNA, MHCC97H-IKKα-SiRNA and MHCC97H-IKKβ-SiRNA were transplanted beneath the right kidney capsule of RAG1 mice using a protocol described previously.20 Animals were sacrificed 6 weeks after transplantation. Tumors and lungs were visualized by fluorescence using a 470-nm light source (Lightools Research, Encinitas, CA). Micro-metastases in the lungs were detected using fluorescent microscopy (Axiovert 200; Zeiss, Stuttgart, Germany).

Quantitative real-time PCR

HCC patients gave written informed consent or assent, respectively, and the study was approved by the local ethics committee. Reverse transcription reactions were performed using the SuperScript First-Strand Synthesis System (Invitrogen, CA). To determine the number of cDNA molecules in the reverse transcribed samples, real-time PCR analyses were performed using the LightCycler system (Roche, Indianapolis, IN). PCR was performed using 10 μl 2× Master Mix SYBR Green I (Takara, Japan), 0.25 μm of each 5′ and 3′ primer, and 2 μl samples or H₂O to a final volume of 20 μl. Samples were denatured at 94°C for 5 min. Amplification and fluorescence determination were carried out in 3 steps: denaturation at 94°C for 10 sec, annealing at 60°C for 15 sec, extension at 72°C for 20 sec; and at the end of extension, detection of SYBR green fluorescence, which reflects the amount of double-stranded DNA. The amplification cycle number was 35. To discriminate specific from nonspecific cDNA products, a melting curve was obtained at the end of each run. Products were denatured at 95°C for 3 sec, and the temperature was then decreased to 58°C for 15 sec and raised slowly from 58 to 95°C using a temperature transition rate of 0.1°C/sec. Data were normalized with GAPDH levels in the samples. The primer sequences used for the real-time PCR are listed in Table 1.

Table 1. Primer sequences for amplification of target genes

Western blot

Proteins were extracted from cultured cells or tumor tissues, and concentrations were determined by Bradford assay (Bio-Rad, CA). Nuclear protein extracts were prepared according to the manufacturer's protocol (NE-PER Nuclear and cytoplasmic extraction reagents; Thermo Scientific, IL). For analysis of IκB phosphorylation, cytoplasmic proteins were extracted according to following protocol. MHCC-97H and their gene transduced cells were quickly washed with ice-cold phosphate-buffered saline (PBS: 150 mM NaCl, 10 mM Na₂HPO₄/NaH₂PO₄, pH 7.4) and lysed in 0.5 ml ice-cold lysis buffer [25 mM Hepes, pH 7.5, 150 mM NaCl, 1% Igepal CA-630, 10 mM MgCl₂, 1 mM EDTA, 2% glycerol, 1 mM sodium orthovanadate, 25 mM sodium luoride, protease inhibitor cocktail [complete Mini EDTA free; Roche Diagnostics, Basel, Switzerland]). Protein samples (30 μg) were fractionated by SDS-PAGE and transferred to a nitrocellulose membrane. Immunoblotting was carried out using antibodies against IKKα, IKKβ, cyclinD1, E-cadherin (Cell Signaling Technology, MA), c-FLIPs, Bcl-xl, Maspin, IκB, p IκB, P65, GAPDH and β-actin (Santa Cruz Biotechnology, CA). The results were visualized via a chemiluminescent detection system (ECL Substrate Western blot detection system, Pierce, IL) and exposure to autoradiography film (Kodak XAR film).

Histology and immunohistochemistry

Tumor tissues were removed and fixed in 4% paraformaldehyde overnight at 4°C and processed, sectioned at 5 μm and stained with hematoxylin and eosin using standard techniques. The sectioned slides were stained immunohistochemically for IKKα, IKKβ, cyclinD1 (Cell Signaling Technology), c-FLIPs and Bcl-xl (Santa Cruz Biotechnology) using techniques described previously.20

Serum-free induced apoptosis and flow cytometry analysis

MHCC97H and MHCC97L and their IKKα SiRNA and IKKβ SiRNA-modified cells were cultured in serum-free DMEM for 36 hr. Then, the cells were harvested and fixed by 70% cold EtOH at −20°C overnight and further analyzed by flow cytometry (FACSCalibur™; BD Biosciences, NJ) using PI/RNase staining buffer (BD Biosciences) protocol.

BrdU incorporation assay and immunocytochemistry

MHCC97H and their IKKα SiRNA and IKKβ SiRNA-modified cells were cultured on coverslips such that they were rapidly dividing. Cells were incubated with 100 μM BrdU for 2 hr, the medium aspirated and immediately fixed and permeabilized in cold methanol:acetone (1:1), then blocked with PBS/3% BSA (Sigma-Aldrich, MO), and incubated with primary BrdU monoclonal antibody (Sigma-Aldrich) diluted in 3% BSA/PBST (0.2% Triton X-100). Following incubation with rabbit anti-mouse HRP-conjugated secondary antibody (Sigma-Aldrich), samples were colorized by DAB, mounted in neutral gum and examined by microscopy (Axiovert 200; Zeiss, Stuttgart, Germany).

Cell growth curves and growth inhibition rates

Cell growth curve was measured by 3-(4,5-dimethy thiazol-2-yl)-2,5-diphenyl terazolium-bromide (MTT) assay. MHCC-97H and their IKKα SiRNA and IKKβ SiRNA-modified cells (5000) were seeded in 96-well plates in DMEM containing 10% FCS and processed at the indicated days of culture following the manufacturer's instructions. Each condition was tested in quadruplicate, and results were confirmed in 3 independent experiments. The absorbance rate of each well was measured using a full automatic enzyme standardized instrument (Tokyo, Japan). The growth curve of each group was calculated by averaging A.

Statistical analyses

Data were calculated as the mean values with 95% confidence intervals. Statistical comparisons were performed using EXCEL2003 (Microsoft, CA). Results were compared using the Student's t test, and p values less than 0.05 (95% CI) were considered statistically significant. All statistical tests were two sided.

Results

High expression of IKKα and IKKβ in human HCC

As discussed above, the IKK-NF-κB signaling pathway plays an important role in the development of HCC at least in mice. We, therefore, decided to examine the expression of IKKα and IKKβ, 2 essential molecules, which control canonical and noncanonical NF-κB signaling, respectively. We also examined the expression of other upstream regulators of NF-κB signaling, as well as NF-κB target genes. RNA was extracted from 40 pairs of tumor tissue and corresponding adjacent tissue, and gene expression was analyzed by real-time PCR. After statistical analysis, 8 indices were found to exhibit significant differences between tumor tissue and their corresponding adjacent tissue. Both IKKβ and cyclinD1 mRNA were significantly higher in tumor tissue (p = 0.0053 and 0.0049, p < 0.01). IKKα, RANK, OPG, c-FLIP and cip2 were also significantly higher in tumor tissue (p = 0.011, 0.038, 0.022, 0.0348 and 0.048, p < 0.05). However, Maspin mRNA was significantly lower in tumor tissue (p = 0.0027, p < 0.01). Other mRNAs including cyclinD3, RANKL, Bcl-xl, stat3 and Cip1 did not show significant differences between tumor and adjacent tissue (Fig. 1a). The upregulation of IKKα and IKKβ suggested that NF-κB signaling might be activated in human HCC. To confirm constitutive activation of NF-κB signaling in HCC, we performed western blot to detect nuclear P65, an activated form of NF-κB, the result indicated that all 8 tumor tissues had nuclear P65 expression, although the expression levels were different (Fig. 1b). Compared with tumor tissues, there were very weak nuclear P65 expression in the corresponding adjacent tissues, and almost negative expression of nuclear P65 in normal liver tissue (Fig. 1c). NF-κB activation is usually considered to be a marker in tumor development and metastasis, and therefore we designed this experiment to investigate the antitumor and antimetastasis effect elicited by inhibition of IKK-NF-κB signaling.

Figure 1. (a) IKKα and IKKβ are highly expressed in human HCC. Transcriptional levels of IKKα, IKKβ, cyclinD1, cyclinD3, OPG, RANK, RANKL, Maspin, c-FLIP, Bcl-xl, Stat3, Cip1 and Cip2 are detected in human HCC and corresponding adjacent tissue from 40 HCC patients. Regions represent the values of relative expression of mRNA compared with GAPDH. IKKβ and cyclinD1 mRNA levels are significant higher in tumor tissue. IKKα, RANK, OPG, c-FLIP and Bcl-xl are also significant higher in tumor tissue, however, Maspin is significantly lower in tumor tissue. There were no significant differences in CyclinD3, Bcl-xl, Stat3, Cip1 and RANKL between tumor and adjacent tissue. The group shows the mean ± SD. *p < 0.05; **p < 0.01 by Student's t test. (b) Western blot showing total P65 and nuclear P65 in tumor tissues from 8 HCC patients (Line 1–8), Line 9 is positive control from A375 melaloma cells. (c) Western blot showing total P65 and nuclear P65 in HCC adjacent tissues (Line 1–7), Line 8 is from normal liver tissue.

Effects of NF-κB signaling-related molecules on tumor proliferation in vitro

To investigate whether the downregulation of IKKα and IKKβ can inhibit tumor cell proliferation and increase apoptosis, we designed a small interference RNA (SiRNA) to block expression of IKKα and IKKβ in the Hepatocellular carcinoma cell line MHCC97H in vitro. When compared with MHCC-97H-pLenti-control-SiRNA cells, the number of positive cells for BrdU was significantly decreased when IKKβ was knocked down. Whereas the number of positive cells was slightly decreased in IKKα knockdown cells (Fig. 2a).

Figure 2. The effect of NF-κB signaling related molecules on tumor proliferation in vitro. (a) BrdU incorporation assay for MHCC-97H-control-SiRNA (a1), MHCC-97H-IKKα-SiRNA (a2) and MHCC-97H-IKKβ-SiRNA (a3; ×200). Positive cells are stained in brown granules. (b) Detection of transcription of CyclinD1 by real-time PCR for 97L (MHCC97L-control-SiRNA), 97H (MHCC97H-control-SiRNA) and their IKKα, IKKβ SiRNA-treated cells; 1–6 are respectively for 97L (MHCC-97L-control SiRNA), MHCC-97L-IKKα-SiRNA, MHCC-97L-IKKβ-SiRNA, 97H (MHCC-97H-control SiRNA), MHCC-97H-IKKα-SiRNA and MHCC-97H-IKKβ-SiRNA. The group shows the mean ± SD. **p < 0.01 by student's t test. (c) Detection of transcription of RANK and RANKL by real-time PCR for MHCC-97L, MHCC-97H and their IKKα, IKKβ SiRNA-treated cells; 1–6 respectively for 97L (MHCC-97L-control-SiRNA), MHCC-97L-IKKα-SiRNA, MHCC-97L-IKKβ- SiRNA, 97H (MHCC-97H-control-SiRNA), MHCC-97H-IKKα-SiRNA and MHCC-97H- IKKβ-SiRNA. The group shows the mean ± SD. **p < 0.01 by student's t test. (d) Western blot analysis of IKKα, IKKβ and CyclinD1 protein contents in unrelated SiRNA-treated MHCC-97L, MHCC-97H and their IKKα, IKKβ SiRNA-treated cells. (e) Cell growth curve from different cells. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

To further interpret this result, we analyzed some molecules related to NF-κB signaling by real-time quantitative PCR and Western blot. CyclinD1 significantly increased in both MHCC97L-IKKβ-SiRNA and MHCC97H-IKKβ-SiRNA groups (Figs. 2b and 2d). Another molecule, RANK, which is a NF-κB signaling-activating receptor and key regulator in tumor metastasis was also analyzed. The expression of RANK was also significantly increased in the MHCC97L-IKKα-SiRNA and MHCC97L-IKKβ-SiRNA groups. Moreover, the expression was higher in MHCC97L-IKKβ-SiRNA than MHCC97L-IKKα-SiRNA. On the other hand, RANK expression was different in that its expression was significantly lower in MHCC97H-IKKα-SiRNA cells than in 97H (MHCC-97H-pLenti-control-SiRNA), and there was almost no difference between MHCC97H-IKKβ-SiRNA and 97H cells (Fig. 2c).

Increased apoptosis in IKKα and IKKβ down-regulated MHCC97L and MHCC97H cells

In addition to the effect of IKKα and IKKβ knockdown on tumor cell proliferation, we hypothesized that there could still be increased apoptosis in tumor cells with SiRNA targeting both genes. Therefore, a serum-free induced apoptosis assay was performed in IKKα SiRNA and IKKβ SiRNA-modified cells. As predicted, increased apoptosis occurred when IKKα and IKKβ were downregulated. The percentage of apoptotic cells increased 17.83 and 21.43% in the MHCC97L-IKKα-SiRNA and MHCC97H-IKKα-SIRNA cells, respectively when compared with their corresponding unrelated SiRNA-modified tumor cells (Fig. 3a). In addition, there was slightly more apoptosis in IKKβ knockdown cells. The levels of apoptosis were increased 25.76 and 25.79% in MHCC97L-IKKβ-SiRNA and MHCC97H-IKKβ-SiRNA cells, respectively (Fig. 3a). A TUNEL assay also confirmed that there were more apoptotic cells in both IKKα and IKKβ SiRNA-modified MHCC97H cells compared with MHCC97H-pLenti-control-SiRNA. In addition, there were more apoptotic cells in IKKβ transduced cells compared with IKKα SiRNA-transduced MHCC97H cells (data not shown).

Figure 3. Increased apoptosis in IKKα and IKKβ downregulated MHCC-97L and MHCC-97H cells. (a) cell cycle analysis for 97L (MHCC97L-control-SiRNA), 97H (MHCC97H-control-SiRNA) and their IKKα, IKKβ SiRNA-treated cells by flow cytometry. (b) Detection of transcription of c-FLIP and Bcl-xl by real-time PCR for MHCC-97L, MHCC-97H and their IKKα, IKKβ SiRNA-treated cells; The group shows the mean ± SD. **p < 0.01 by student's t test. (c) Western blot analysis for c-FLIP and Bcl-xl for MHCC-97L, MHCC-97H and their IKKα, IKKβ SiRNA-treated cells. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

To further interpret the mechanism of apoptosis, we screened a series of apoptosis related molecules and eventually focused on 2 NF-κB regulated antiapoptosis genes: c-FLIP and Bcl-xl. Real-time PCR and Western blot results indicated a downregulation of both genes when IKKα and IKKβ were blocked (Figs. 3b and 3c). This would be expected to trigger the apoptosis pathway, contributing to increased apoptosis of the tumor cells.

Significant inhibition of HCC growth by SiRNA targeting IKKα and IKKβ in a subcutaneous injection model

At first, we compared the proliferation rates of MHCC-97H and their IKKα-SiRNA and IKKβ-SiRNA modified cells by using MTT assay in vitro. The cell growth curve indicated that both IKKα and IKKβ SiRNA had suppressing effect on tumor cells growth (Fig. 2e). To determine if IKKα and IKKβ played the same roles in tumorigenesis and apoptosis in vivo, we subsequently evaluated the growth of subcutaneous tumors formed in nude mice. Thirty nude mice were divided into 6 groups: MHCC97L-pLenti-control-SiRNA, MHCC97L-IKKα-SiRNA, MHCC97L-IKKβ-SiRNA, MHCC97H-pLenti-control-SiRNA, MHCC97H-IKKα-SiRNA, and MHCC97H-IKKβ-SiRNA. A tumorigenesis assay was conducted in each group by subcutaneously injecting HCC cells treated with lentivirus-mediated SiRNA-targeting IKKα and IKKβ and unrelated SiRNA. Two and a half weeks after injection, tumors were detected in the MHCC97L-pLenti-control-SiRNA and MHCC97H-pLenti-control-SiRNA group animals. The volume of the tumors increased to a maximum of approximately 1200 mm³ and 1000 mm³ for the MHCC97L and MHCC97H group animals, respectively until their death (Figs. 4a and 4d). The tumor growth induced by SiRNA-treated cells was retarded. The tumors in the groups treated with SiRNA were visible about three and a half weeks after injection, and gradually disappeared in MHCC97H cells treated with SiRNA 5 weeks after injection. All of the mice in the SiRNA-treated group survived for 9 weeks after injection and were sacrificed for further experiments. The volume of the tumors in the MHCC97L-IKKα-SiRNA and MHCC97L-IKKβ-SiRNA groups only reached maximum values of approximately 700 mm³ and 600 mm³, which is about half the volume of the MHCC97L-pLenti-control-SiRNA induced tumors. This effect was even greater in the MHCC97H-IKKα-SiRNA and MHCC97H-IKKβ-SiRNA groups where the tumors increased in size very slowly to a maximum of less than 200 mm³ five weeks after injection. Interestingly, the tumor volume then decreased and some were no longer visible 9 weeks after injection. Figures 4c and 4d graphically describe the information pertaining to tumor growth. Because all of the injected tumor cells were labeled with green fluorescence protein (GFP), tumor growth was easily monitored using the whole body Lightool fluorescent system.

Figure 4. SiRNA targeting IKKα and IKKβ can inhibit IκB phosphorylation and NF-κB activation in HCC cell lines in vitro and suppress tumor growth of HCC in the subcutaneous injection model. (a) Western blot showing expression levels of total P65, nuclear P65, IκB and phosphorylated IκB in different cells. Line 1, MHCC-97H-control-SiRNA; Line 2, MHCC-97H-IKKα-SiRNA; Line 3, MHCC-97H-IKKβ-SiRNA; Line 4, MHCC-97L-control-SiRNA; Line 5, MHCC-97L-IKKα-SiRNA; and Line 6, MHCC-97L-IKKβ-SiRNA. The results indicated that NF-κB signaling was consequently activated in both cell lines of MHCC-97L and MHCC-97H, there was clear IκB phosphorylation in both HCC cell lines (Line 1 and Line 4). (b) Both activated NF-κB signaling and IκB phosphorylation were significantly inhibited by IKKα (*p < 0.05) and IKKβ SiRNA (**p < 0.01). The data are the result of 3 measurements in 3 independent experiments. (c) Growth inhibition of MHCC-97L. There is no difference among 97L (MHCC-97L control-SiRNA), MHCC-97LIKKα-SiRNA and MHCC-97L-IKKβ-SiRNA. (d) Growth inhibition of MHCC-97H. 97H (MHCC-97H control-SiRNA) grows much faster than the groups of IKK α SiRNA and IKK β SiRNA-treated cells. However, no lung metastasis was observed in both 97L (MHCC-97L-control-SiRNA) and 97H (MHCC-97H- control-SiRNA) cells.

To determine the molecular mechanism of the antitumor effects of IKKα and IKKβ SiRNA, some related indices were investigated. First, we did western blot to detect nuclear P65 and phosphorylated IκB in MHCC-97L, MHCC-97H cells and their IKKα and IKKβ SiRNA-modified cells, respectively, the results indicated that NF-κB signaling was consequently activated in both cell lines of MHCC-97L and MHCC-97H, there was clear IκB phosphorylation in both HCC cell lines. The results also showed that both activated NF-κB signaling and IκB phosphorylation were inhibited by IKKα and IKKβ SiRNA (Figs. 4a and 4b). The actual expression of IKKα and IKKβ in tumor tissues from each group was analyzed by immunohistochemistry. The sections were stained for IKKα and IKKβ. After IKKα and IKKβ SiRNA lentivirus infection, the expression of IKKα and IKKβ was significantly reduced in the tumor tissues (Figs. 5a2, 5a5, 5b3 and 5b6) compared with the tumor tissue caused by MHCC97L-pLenti-control-SiRNA and MHCC97H-pLenti-control-SiRNA (Figs. 5a1, 5b1, 5a4 and 5b4).

Figure 5. Histoimmunochemistry analysis for tumor tissue in subcutaneous injection model. The histoimmunochemistry analysis for IKKα, IKKβ, CyclinD1, Bcl-xl and c-FLIP for tumor tissue induced by MHCC-97L, MHCC-97H and their IKKα and IKKβ SiRNA-treated cells are shown in a1–a6, b1–b6, c1–c6, d1–d6 and e1–e6, respectively (×200). The results of integrated optical density analyzed for each graph by image-pro plus software (ver 5.0) are shown from f–j.

CyclinD1 is considered to be an index of tumor proliferation, and as expected, it was highly expressed in tumors caused by the HCC cell lines MHCC97L and MHCC97H (Figs. 5c1 and 5c4). When IKKα was knocked down, tumor cyclinD1 expression was significantly reduced (Figs. 5c2 and 5c5). However, when IKKβ was blocked in the tissue, the expression of CyclinD1 dramatically increased (Figs. 5c3 and 5c6). Bcl-xl and c-FLIP were both decreased in tumor tissue when IKKα or IKKβ was knocked down compared with tumor tissue caused by the unrelated SiRNA-infected HCC cell lines (Figs. 5d1–5d6 and 5e1–5e6). Figures 5f–5j shows the results obtained using image-pro plus software (Version 5.0, Media Cybernetics, MD), which calculated the integrated optical density of the same reference indices.

Significant inhibition of lung metastatogenesis by SiRNA-targeting IKKα and IKKβ in a kidney capsule transplantation model

Although MHCC97H cells have excellent lung metastasis potential, in the subcutaneous injection model, we did not detect any evidence of lung metastasis. Combined with data from previous experiments,20 we concluded that the microenvironment in the subcutaneous injection model was not as good as the kidney capsule model. Therefore, to investigate lung metastatogenesis of HCC, we transplanted MHCC97H modified cells beneath the right kidney capsule of RAG-1 mice. Six weeks after transplantation, the tumors in the kidney and lung metastases of transplanted RAG-1 mice were detected using the whole body Lightool fluorescent system. Our data showed that the tumor volumes were not significant different (Figs. 6a–6c), which was a contrast to the subcutaneous injection model. Significant differences in the pulmonary metastases were noted during the study. Lung metastases from the kidney tumor induced by MHCC97H-pLenti-control-SiRNA cells were extensive, and many large metastases were detected (Fig. 6a1 and 6a2). Pathological examination confirmed that metastatic tumors occurred in the lungs of RAG1 mice (Fig. 6a3). The pulmonary metastases caused by MHCC97H-IKKα-SiRNA were less extensive than those in the MHCC97H-pLenti-control-SiRNA group. Only one large metastasis and a little green spot were detected (Fig. 6b1) when we used the Lightool fluorescent whole body system. Even more encouraging, the pulmonary metastasis in total 6 mice of 97H-IKKβ-SiRNA group was almost invisible; we were unable to find any evidence of metastasis using the Lightool whole body fluorescent system. However, when we used fluorescent microscopy, we observed very small green spots on the surface of lung (Figs. 6c1 and 6c2). These findings indicate a major effect of IKK SiRNA transduction, especially in the animals that were administered cells transduced with pLenti-GFP-IKKβ-SiRNA.

Figure 6. Lung metastatogenesis is significantly inhibited by SiRNA targeting IKKα and IKKβ in kidney capsule transplantation model. (a) MHCC-97H-control-SiRNA transplanted into kidney capsule of RAG-1 mice, 6 weeks later, with lung metastasis (a1, a2, viewed by Lightool whole body Fluorescent system) and hematoxylin and eosin stain of section (a3; ×40). (b) MHCC-97H-IKKa-SiRNA transduced cells transplanted into kidney capsule of RAG-1 mice, 8 weeks later, micro-metastasis was detected (b1, viewed by fluorescent microscopy); (c) MHCC-97H-IKKβ-SiRNA transduced cells transplant into kidney capsule of RAG-1 mice, 8 weeks later, no lung metastasis grossly (c1), only detected micro-metastasis (c2, viewed by fluorescent microscopy). (d) Real-time analysis for transcription of Maspin, RNAKL and OPG for 97H (MHCC-97H-control-SiRNA), MHCC-97H-IKKα-SiRNA and MHCC-97H-IKKβ-SiRNA. The group shows the mean ± SD. **, ##, §§, p < 0.01 by Student's t test. (e) Western blot for expression of Maspin and E-Cadherin for 97H (MHCC-97H-control-SiRNA), MHCC-97H-IKKα-SiRNA and MHCC-97H-IKKβ-SiRNA.

Maspin, RANKL and osteoprotegerin (OPG) are critical molecular targets that control metastatogenesis

The characteristics of metastatogenesis between the 2 HCC cell lines are quite different in that MHCC97L had no lung metastatic potential. However, MHCC97H cells grew easily in lungs and readily metastasized. Experiments were designed to determine why the antimetastasic effect was so apparent after the treatment of SiRNA for IKKα and IKKβ. It is interesting to note that the expression of Maspin, which is a negative regulator of tumor growth and metastasis, also increased significantly when both IKKα and IKKβ were knocked down in MHCC97H cells (Figs. 6d and 6e). As a common index for cell adhesion and movement, we observed that the expression of E-cadherin was increased in IKKα and IKKβ knockdown MHCC97H cells (Fig. 6e). Furthermore, real-time PCR analyses of these 2 cell lines indicated that OPG, RANK and RANKL expression levels were much higher in MHCC97H than MHCC97L cells. When we compared the expression levels before and after IKKα or IKKβ knockdown in MHCC97H cells, we found that OPG and RANKL expression was significantly decreased (Fig. 6d). There was no difference in the expression of RANK before and after IKKα or IKKβ knockdown in MHCC97L (Fig. 2c). For this reason, we concluded that OPG and RANKL exert some control on the metastatic characteristics of HCC cells. Because OPG is an antagonist to RANKL, when it is consumed, one would also expect a decrease in RANKL expression. However, further studies are required to confirm this hypothesis.

Discussion

It is well accepted that hepatitis is one of the important causes of HCC. Some clinical research has indicated that NF-κB signaling, which plays an important role in inflammation, is activated during human hepatitis and HCC.21 However, these reports have not provided a complete picture of the complicated role of NF-κB signaling in human HCC. Moreover, a series of new targets such RANK, RANKL and Maspin have been reported to have an important role in tumorigenesis.9, 10, 22–25 In the present communication, we have undertaken a study of the expression of several NF-κB related molecules, such as IKKα, IKKβ, RANK and Maspin in human liver cancer and in adjacent noncancerous tissue. The results have shown that almost all of the indices were significantly different in cancerous tissues compared with noncancerous tissues (Fig. 1a). IKKα and IKKβ, which are 2 essential molecules in the IKK-NF-κB axis, were expressed at significantly higher levels in human HCC tissue accompany with NF-κB signaling activation (Fig. 1b). Based on this observation, we designed experiments to knockdown IKKα and IKKβ expression using lentivirus-mediated SiRNA and found that inhibition of NF-κB signaling and downstream targets in HCC cell lines had antitumor and antimetastasis effects.

SiRNA was used to knockdown the expression of IKKα and IKKβ in two HCC cell lines, MHCC97L and MHCC97H. When these cells were subcutaneously injected into nude mice, we observed a dramatic decrease in tumor size and an increase in survival time in the SiRNA-treated carcinoma cell lines compared with the unrelated SiRNA-treated cells. Because the loss of IKK can efficiently inhibit activation of NF-κB signaling (Figs. 4a and 4b),7, 8, 26, 27 it is expected that NF-κB-mediated cell survival would be attenuated in the HCC cell line. Although the immune response is completely absent in nude mice, some apoptosis inducing molecules such as TNF α still exist in the dermis and could induce apoptosis of the injected tumor cells. This, in turn, would inhibit the growth of the implanted cells. A further in vitro study using a serum-free apoptosis inducing assay demonstrated a similar increase in apoptosis after blocking IKKα and IKKβ. We subsequently examined the molecular mechanism of apoptosis and screened 2 NF-κB signaling regulated antiapoptosis genes c-FLIP and Bcl-xl. The results showed that these genes were downregulated when IKKα and IKKβ were knocked down in both in vivo and in vitro assays. The c-FLIP and Bcl-xl were highly expressed in human malignancy,28–32 and we found that the transcription level of c-FLIP was significantly increased in human HCC tissue, whereas there was no difference in Bcl-xl transcription level. Nevertheless, Bcl-xl is a significant prognostic factor for disease progression in human HCC, because the actual effect of this gene is not determined by its expression level but rather by the rate of deamidation.33 Previous studies in NF-κB signaling-inactivated tumor cells have shown that downregulation of both IKKα and IKKβ, using a knockdown strategy, results in increased apoptosis.34, 35

When we performed BrdU incorporation assays, we found the proliferative rate was significantly decreased in 97H-IKKβ SiRNA cells compared with 97H-plenti-control-SiRNA. Although the proliferative rate was slightly decreased in 97H-IKKα SiRNA, there was no significant difference between 97H-IKKβ SiRNA and 97H-plenti-control-SiRNA cells. In this experiment, the extent of proliferation inhibition in the MHCC97L-IKKα-SiRNA and MHCC97L-IKKβ-SiRNA groups was not as dramatic as the MHCC97H-IKKα-SiRNA and MHCC97H-IKKβ-SiRNA groups when NF-κB was inhibited in both cell types. The reason for this difference could be explained by the different characteristics of different tumor cells; however, future investigations are required to address this question.

In the gene expression analysis for IKKα or IKKβ knockdown HCC cells, we noticed that a receptor activator of NF-κB ligand (RANK) is notably upregulated in the MHCC97L-IKKα-SiRNA and MHCC97L-IKKβ-SiRNA groups. This may be an important reason to explain the differences between MHCC97H and MHCC97L. RANK is an integral membrane protein with an intracellular signaling domain. It can directly bind to RANKL, which in turn triggers multiple downstream signaling pathways including Akt, MAPKs and NF-κB. It has been reported that RANK can also induce the production of reactive oxygen species, which are essential for activation of the IKK upstream component and NF-κB activation in MHCC97L.36–39 When expression of IKKα and IKKβ is blocked in MHCC97L cells' according to the results from previous experiments, activation of NF-κB will be effectively inhibited.7, 8, 26, 27 We hypothesize that an increased amount of RANK and sustained expression of RANKL (see Fig. 2c) may sensitize downstream signaling pathways, maybe including NF-κB, to maintain cell proliferative ability. This may serve as a compensatory mechanism for the deficiency of IKKs. We consider the upregulation of RANK as a “self-protection” mechanism in MHCC97L HCC cells, while it seems that such mechanism did not exist in MHCC97H cells; expression of RANK was maintained or even downregulated. Inhibition of NF-κB signaling was also observed in the apoptosis inducing assay. MHCC97H-IKKα-SiRNA and MHCC97H-IKKβ-SiRNA cells are more sensitive to serum-free induced apoptosis than MHCC97L-IKKα-SiRNA and MHCC97L-IKKβ-SiRNA cells because of the loss of an antiapoptotic effect executed by RANK induced NF-κB and other downstream signaling pathways. The results obtained from human HCC tissue indicated that the expression of RANK was significantly elevated in human liver cancer tissue relative to adjacent noncancerous tissue. The increased expression of RANK further implied that RANK is a critical receptor sensitizing NF-κB and other signaling molecules in tumor cells, and may also be a potential target for antitumor therapy.

When we used the kidney capsule transplantation model, the results again proved that inhibition of IKKα or IKKβ could suppress lung metastasis of the HCC. In MHCC97H cells, when IKKα was knocked down by SiRNA, expression of Maspin significantly increased. This protein was first reported in 1994 as a serpin with tumor suppressive properties, which is expressed in normal mammary epithelial cells but reduced or absent in breast carcinomas.40 Further research led to the characterization of Maspin as a class II tumor suppressor based on its ability to inhibit cell invasion, promote apoptosis and inhibit angiogenesis.41–43 In 2007, Luo et al. studied IKKα knockdown in TRAMP mice and updated Maspin as a downstream target of RANKL-RANK-IKKα signaling in prostate cancer cells. Thus, its mechanism of action in prostate cancer may provide further insights into the interpretation of increased Maspin expression in IKKα knockdown HCC cells.9 Although we selected MHCC97H cells that have high lung metastatic properties, we did not observe lung metastasis in the present subcutaneous injection model. Based on our experience in the tumorigenic study, we selected the kidney capsule transplantation model to study lung metastatogenesis, because this model could provide a better tumor microenvironment. Also, tumor microenvironment should be mentioned here to interpret the tumor-host interaction while NF-κB signaling was blocked in HCC cells. Similar to the model of “seed and soil,” tumor microenvironment contribute to many properties of tumor such as tumorigenesis, angiogenesis, proliferation and metastasis, which consist of many kinds of cells including endothelial cells, pericytes, smooth-muscle cells, fibroblasts of various phenotypes, myofibroblasts, neutrophils and other granulocytes, mast cells, T-, B- and natural killer lymphocytes, and antigen-presenting cells such as macrophages and dendritic cells.14 Moreover, based on the recently theories about genesis and development of HCC, NF-κB signaling is the predominant signaling pathway balancing proliferation and programmed cell death in many tissues including liver; complementary proliferation of transformed hepatocyte mainly regulated by NF-κB signaling induced by cytokines, such as TNF-α, IL-1 and IL-6, secreted by the cells in tumor microenvironment. Although T and B cells are absent in immunodeficiency mice that were used in our investigation, other cells like neutrophils and granulocytes still exist and secrete cytokines. The complementary proliferation and antiapoptotic effect were blocked because of the attenuation of NF-κB, therefore, cell programmed death of HCC will happened, which was contribute the anticancer effect of the gene therapy.44, 45 When we transplanted lentivirus-transduced MHCC97H cells beneath the kidney capsule, lung metastasis was easily detected using the Lightool whole body fluorescent system. Our data showed that lung metastasis was effectively suppressed in this system, suggesting that there may be another explanation for increased expression of E-cadherin in MHCC97H cells when IKKα and IKKβ were downregulated. This effect, in turn, could suppress the epithelial to mesenchymal transition, thereby inhibiting metastasis. This result is consistent with the conclusion of a previous study that expression of E-cadherin could be suppressed by NF-κB signaling.46 The higher level of metastasis suppression in MHCC97H-IKKβ-SiRNA cells also supports this contention. In this study, metastasis could only be detected using fluorescent microscopy, and it was noted that the metastasis caused by MHCC97H-IKKα-SiRNA was larger than the metastasis caused by MHCC97H-IKKβ-SiRNA cells. Previous results indicated that the high lung metastasis potential in HCC cells was because of high expression of OPG and RANKL. When SiRNA was used to knockdown the expression of IKKα and IKKβ, the expression of OPG and RANKL was decreased, and lung metastasis was inhibited.

It is commonly accepted that HCC is insensitive to radiotherapy and chemotherapy partially because of the activation of the NF-κB signaling pathway. This has provided further insights into a potential new gene therapy targeting NF-κB signaling in HCC cells. Tumor growth and metastasis could be inhibited through the inactivation of this pathway. The extent of the therapeutic effect depended on the level of inhibition of NF-κB signaling and the degree of malignancy of the HCC. We anticipate that this gene therapy will eventually be applied in a clinical setting and may overcome many problems such as poor recovery rate and chemoresistance.

Acknowledgements

This work was supported by grants from Jiangsu Province's Outstanding Medical Academic Leader program (to B.S), Jiangsu Province's Key Medical Centre, and grants from Natural Science Foundation of China (to B.S), Ministry of Health, China (to B.S) and Jiangsu Province “Six adults just” high peak.