In solid tumors, hypoxia triggers an aberrant vasculogenesis, enhances malignant character, and elevates metastatic risk. The plasma membrane organizing protein caveolin-1 (Cav1) is increased in a variety of cancers, including hepatocellular carcinoma (HCC), where it contributes to metastatic capability. However, the reason for elevation of Cav1 in tumor cells and the mechanistic basis for its contributions to metastatic risk are not fully understood. Here, we show that in HCC cells, hypoxia elevates expression of Cav1, which then acts through the calcium-binding protein S100P to promote metastasis. Hypoxic regions of HCC xenografts displayed elevated expression of Cav1. Hypoxia promoted HCC cell migration and invasion and distant pulmonary metastases, whereas Cav1 silencing abolished these effects. Gene expression profiling revealed that hypoxia-induced Cav1 functioned as a positive regulator of S100P via activation of the NF-κB pathway. S100P elevation under hypoxic conditions was abrogated by silencing of Cav1 or NF-κB function. Conversely, restoring S100P in Cav1-silenced cells rescued the migratory potential of HCC cells along with tumor formation and lung metastasis. In clinical specimens of HCC, we observed S100P overexpression to correlate with venous invasion, microsatellites, direct liver invasion, and absence of tumor encapsulation. Collectively, our findings demonstrated how hypoxia-induced expression of Cav1 in HCC cells enhances their invasive and metastatic potential. Cancer Res; 76(24); 7242–53. ©2016 AACR.

Hepatocellular carcinoma (HCC) is a typical hypervascular tumor. However, abnormal architecture of newly developed blood vessels and inefficient oxygen delivery results in the development of tumors with regions subjected to hypoxia. Cellular expansion of tumors also distances cells from the tumor vasculature and thus depriving cells of oxygen. Hypoxia is a condition where there is a reduction in the normal level of tissue oxygen tension (1). Most mammalian tissues exist at 2% to 9% O2, whereas hypoxic condition, defined as ≤2% oxygen, is characterized in many solid tumors. Increasing evidence has shown that hypoxia stimulates the malignance of cancer cells. In fact, intratumoral hypoxia is associated with enhanced proliferation, angiogenesis, metastasis, radio-, and chemoresistance, as well as an independent prognostic factor of poor clinical outcome in cancer patients (2–4). In HCC, expression of hypoxic markers predicts poor prognosis in resectable cases (5). However, the molecular basis of tumor cell adaptation to hypoxia and the resulting survival of cells within the hypoxic environment remain obscure. Caveolin-1 (Cav1), the major scaffold protein of caveolae and a component at focal adhesions, has been implicated in cellular transport, signal transductions, and human cancers. Expression of Cav1 has been found to be aberrantly altered in a broad spectrum of human cancers (6), with increasing evidence revealing upregulation of Cav1, particularly during the advanced stages of cancer metastasis (7). Cav1 overexpression is closely associated with aggressive clinical behavior and poor prognosis in various carcinomas. In the majority of HCC cases, Cav1 expression has indeed been shown to be overexpressed (8) and its overexpression is significantly correlated with metastasis and worse prognosis (9). Functionally, Cav1 has been shown to promote HCC tumourigenesis and metastasis in nude mice (10) and, correspondingly, in mouse hepatocarcinoma cell lines, strong Cav1 expression is found in those with higher invasive ability (11). Moreover, Cav1 has also been shown to enhance transformation and survival of mouse hepatoma cells (12). These studies emphasize the vital role of Cav1 in HCC metastasis. However, the involvement of Cav1 in hypoxia is yet to be clearly addressed. Downregulation of Cav1 in tumor vascular endothelial cells within the hypoxic microenvironment of solid tumors has been shown to promote angiogenesis (13). Nevertheless, Cav1 has been found to be enhanced under hypoxia in ovine fetal and neonatal lung microvascular endothelial cells (14). In human neuroblastoma cells under hypoxic condition, an enhanced Cav1 level has been shown to prevent apoptotic cell death, when mediated by inducible nitric oxide synthase (15). These findings indicate that Cav1 plays a pivotal role in regulating oxidative stress, which contributes to the hypoxic tolerance of tumor cells. However, the underlying mechanisms of Cav1 under hypoxia are complex and vary in different cellular contexts.

Although intratumoral hypoxia is originally thought to kill cancer cells, it in fact provides selective pressure for the most aggressive cancer cells to adapt and survive under hypoxic stress. Thus, the understanding of the hypoxia-induced genes that confer the aggressiveness of HCC cells is vital. Based on the tight association between Cav1 and cancer metastasis, we set out to investigate the possible role of Cav1 in HCC under hypoxic condition. Our data indicated that Cav1 expression could be induced in HCC cells and xenografts under hypoxia. We hypothesize that hypoxia-induced Cav1 contributes toward HCC cell aggressiveness in metastasis under hypoxic stress. In this study, we present for the first time that hypoxia-induced Cav1 upregulates S100P, a downstream target of Cav1, via activation of the NF-κB pathway, and this upregulation promotes HCC tumorigenesis and metastasis.

Clinical samples

Paired samples of primary HCC and the corresponding non-tumorous liver tissues were obtained at the time of surgical resection at Queen Mary Hospital, Hong Kong, and were randomly selected for the study. Tissue samples were fixed in formalin and embedded in paraffin for sectioning. Use of human samples was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB).

Cell culture, transfection, and stable cell lines

Human HCC cell lines include Huh7, Hep3B, PLC/PRF/5, and HLE were purchased from ATCC. SMMC7721 and BEL7402 were obtained from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences, People's Republic of China (PRC). Metastatic HCC cell lines MHCCLM3 and MHCC97L were gifts from Fudan University, PRC (16). MHCC97L (MHCC97L-Luc) was a gift from the Department of Surgery, The University of Hong Kong. All cell lines were obtained between 2008 and 2013 and kept at low passages for experimental use. DMEM with high glucose supplemented with 110 mg/L sodium pyruvate was used to culture MHCC97L and MHCCLM3. Low glucose DMEM was used to culture HLE. The remaining cell lines were cultured in high glucose DMEM. All medium was supplemented with 3.7 g/L sodium bicarbonate, 50 U/mL penicillin, 50 μg/mL streptomycin, and 10% heat-inactivated fetal bovine serum (FBS; Gibco). Cells were kept in 37°C humidified incubator supplied with 5% CO2. All cell lines used in this study were regularly authenticated by morphologic observation and tested for the absence of mycoplasma contamination. For hypoxic treatment, cells were kept in 37°C humidified incubator supplied with a gaseous mixture of 1% O2, 5% CO2, and balanced with N2. Transfection was performed using FuGENE6 Transfection Reagent (17) or Lipofetamine 2000 Transfection Reagent (Invitrogen). Hep3B Cav1 knockdown stable clones and MHCC97L S100P knockdown stable clones were constructed using MISSION short-hairpin (sh) RNA targeting Cav1 and S100P (Sigma-Aldrich), respectively. The rescue experiment was carried out by using a pCR-Flag-IKKβ expression construct (Addgene). Detailed procedures are described elsewhere (10).

Subcutaneous injection, orthotopic liver implantation, and hepatic artery ligation

Stable clones were analyzed for tumorigenicity by subcutaneous injection. For each stable clone, 1.5 × 106 cells were inoculated into the right flank of 4-week-old male BALB/c nude mice. The tumor weight was measured at the end of the subcutaneous injection experiment. To obtain a tumor seed for orthotopic implantation, a 100-μL cell suspension (containing 5 × 106 cells) in PBS was subcutaneously injected into the flanks of male BALB/c nude mice at the age of 6 weeks. Mice were sacrificed 2 weeks later and the tumors were cut into approximately 1 mm3 for successive orthotopic implantation into 6-week-old male BALB/c nude mice. Mice were first anesthetized and laparotomy was performed to expose the liver, before the insertion of the tumor cube into a small hole, formed by mechanical injury on the liver capsule using a needle. The tumor cube was then secured by suture. Hepatic artery ligation (18) was carried out by tying a fine thread around the main branch of the hepatic artery, so as to obstruct the blood flow into the liver; and hence mimicking hypoxia. For the control group, mice were subjected to the same orthotopic implantation but spared from hepatic artery ligation (HAL). The abdominal wall was closed by suture. For tumors that were formed from luciferase-labeled cell lines, growth of the tumor was monitored by intraperitoneal injection of D-luciferin (Xenogen; 100 mg/kg animal) followed by bioluminescence detection using IVIS 100 Imaging System (Xenogen). Six weeks after the orthotopic implantation, the mice were sacrificed and their lungs and livers were excised. The maximum diameter was taken as the tumor size for comparison among tumors. Animals (Control of Experiments) Ordinance (Hong Kong) and animal experimentation guidance from The University of Hong Kong were strictly followed for all animal work performed.

Chromatin immunoprecipitation

PLC/PRF/5 cells were seeded in 100-mm culture dish to give approximately 50% confluence, and incubated in normoxia or hypoxia for 24 hours. The chromatin immunoprecipitation (ChIP) assay was performed using the EpiQuik ChIP Kit (Epgenetek). Detailed procedures were carried out according to the manufacturer's manual. HIF-1α antibody (Novus Biologicals) was used for the immunoprecipitation step. Conditions of primer pairs were optimized for PCR and quantitative real-time RT-PCR. Genomic DNA from PLC/PRF/5 was used as a positive control for the PCR, and the normal mouse IgG provided by the CHIP manufacturer served as a negative marker.

Cell migration and invasion assay

Transwell Permeable Supports (inserts of 6.5 mm in diameter; Corning) were used for cell migration assay. BD BioCoat Matrigel invasion chamber (BD Biosciences) was used for a cell invasion assay. Cell suspension in serum-free medium was added to the upper chamber at various densities depending on the cell line. Medium supplemented with either 10% FBS or 1% FBS together with 20 ng/mL HGF (R&D Systems, Inc.) depending on the cell line was added to the lower chamber to act as chemo-attractant. Cells were then incubated in normoxia or hypoxia for 18 hours. After incubation, migrated or invaded cells were fixed with methanol for 30 minutes and stained with 1× crystal violet. Different fields of cells were randomly selected and photographed. The numbers of cells were counted. All experiments were performed in triplicates.

Western blotting

NETN lysis buffer (150 mmol/L NaCl, 5 mmol/L EDTA, 50 mmol/L Tris-Cl, pH 8.0, 0.2% NP-40, supplemented with 10 μmol/L leupeptin, 2 μg/mL aprotinin, 1 mmol/L dithiothreitol (DTT), 1 mmol/L sodium orthovanadate (Na3VO4) and 1 mmol/L PMSF) was used for protein extraction from cells. The extracted protein was subjected to sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE). Primary antibodies used were as follows: α-tubulin, β-actin (Sigma-Aldrich), S100P (Cell Signaling Technology), Cav-1 (BD Transduction Laboratories), HIF1α (Cell Signaling Technology), phospho-p65 (Cell Signaling Technology), p65 (Santa Cruz Technology), phospho-IKKβ (Santa Cruz Technology), and IKKβ (Santa Cruz Technology).

Immunohistochemistry

Organs of the mice embedded in paraffin blocks were processed to give 5-μm thick sections, and xylene was used for de-paraffinization of the sections. Rehydration was carried out by using a gradient of alcohols and distilled water. The sections were heated at 100°C for 15 minutes in antigen retrieval buffer (10 mmol/L citrate buffer, pH 6.0, 1 mmol/L EDTA, pH 8.0). After antigen retrieval, the sections were incubated in 10% hydrogen peroxidase (H2O2) in TBS for 20 minutes followed by incubation in 10% normal mouse/rabbit serum for 30 minutes. Primary antibody was then added and incubated with the sections at 4°C. After 3 successive washes with TBST, secondary antibody was added and incubated with the sections at room temperature for 30 minutes. The sections were then counterstained with hematoxylin and eosin. The DAKO EnVision system (DAKO) was used for visualization of the signal. Digital slides for analysis were created using the Aperio ScanScope CS System. Primary antibodies used were Cav-1 (D46G3; Cell Signaling Technology), HIF1α (H-206; Santa Cruz Biotechnology), S100P (abcam), phospho-p65 (Santa Cruz Biotechnology) and Pimonidazole (Hypoxyprobe).

Oligonucleotide microarray analysis

Gene expression profiles of Hep3B-Luc-sh-Cav1 (sh-8001) and Hep3B-Luc-sh-Ctl were compared using the Affymetrix human genome U133 Plus 2.0 GeneChip, covering 47,000 transcripts and variants. Experiment was carried out at the Centre for Genomic Sciences of The University of Hong Kong. Scanned output files were analyzed using the MicroArray Suite 5 (MAS5) method with Genespring GX (Affymetrix) and Ingenuity Pathway Analysis (IPA; Ingenuity Systems) software. Transcripts with more than 2-fold difference was regarded as differentially expressed. Microarray data are available publicly at http://www.ncbi.nlm.nih.gov/geo (GEO accession number, GSE75994).

Statistical analysis

One-way ANOVA performed by GraphPad Prism 5 was used for statistical analyses in various functional assays in this study. Clinicopathological analysis was analyzed with the Fisher exact test using IBM SPSS 23 for Windows. A P value less than 0.05 was considered statistically significant.

Additional experimental procedures are provided in the Supplementary Material.

Hypoxia elevates expression of Cav1 in HCC cells

To investigate if hypoxia elevates Cav1 expression in HCC cells, the expression level of Cav1 was studied and compared in HCC cell lines under both normoxia and hypoxia. Six non-metastatic HCC cell lines (Huh7, Hep3B, PLC/PRF/5, SMMC7721, BEL7402, and HLE), two metastatic HCC cell lines (MHCC97L and MHCCLM3) and an immortalized nontumorigenic liver cell line (MIHA) were subjected to 24 hours of normoxic (20% O2) or hypoxic (1% O2) treatment. Cav1 mRNA and protein were barely detected in MIHA cells under both conditions. The level of Cav1 mRNA was upregulated under hypoxia in multiple HCC cell lines except SMMC7721 and BEL7402 (Fig. 1A), while Cav1 protein was elevated in all cell lines (Fig. 1B). The upregulation of the Cav1 level in metastatic cells was revealed to be more dramatic than in nonmetastatic cells. The kinetics of the upregulation of Cav1 mRNA and protein levels under hypoxia was further examined in SMMC7721, PLC/PRF/5, and MHCC97L cells (Supplementary Fig. S1). It was noted that upregulation of Cav1 in SMMC7721 was detected at the translational level while upregulation of the Cav1 level in PLC/PRF/5 and MHCC97L cells was identified at the transcriptional level. CoCl2 was used to mimic the effect of hypoxia. Similarly, all the cell lines tested, except SMMC7721, showed an upregulation of Cav1 mRNA when CoCl2 was applied (Fig. 1C). As for the translational level, all four cell lines tested showed an increase of Cav1 protein with CoCl2 (Fig. 1D). We further analyzed whether expression of Cav1 is elevated in regions of hypoxia in tumors derived from HCC cell lines. Although HCC is a typical hypervascular tumor, abnormal architecture of newly developed blood vessels and inefficient oxygen delivery have resulted in the constitution of hypoxic regions within the tumor. Xenografts of PLC/PRF/5-derived tumors excised from nude mice were stained for Cav1 and HIF1α. A correlation between staining of Cav1 and hypoxic marker, pimonidazole in tumors was observed, suggesting that the Cav1 level is enhanced in areas of tumors subjected to hypoxic condition (Fig. 1E). Upregulation of Cav1 expression under hypoxia was HIF1α-dependent evidenced by luciferase reporter assay and chromatin immunoprecipitation assay. Our data further showed that elevation of Cav1 expression under hypoxia was HIF2-independent (Supplementary Fig. S2).

Figure 1.

Elevated expression of Cav1 in HCC cells under hypoxia. A, Upregulation of the Cav1 mRNA level in HCC cell lines under hypoxia. HCC cells were placed in hypoxic condition for 24 hours. Quantitative real-time RT-PCR was used to determine the expression of Cav1 mRNA in normoxic and hypoxic conditions. B, Hypoxia enhanced the Cav1 protein level. Western blot analysis was used to determine the expression of Cav1 protein in cells grown in normoxia and hypoxia. HIF1α/β-actin and Cav1/β-actin ratio determined by band intensity is reported for Western blot. C, CoCl2 treatment elevated expression of Cav1 mRNA. Expression level of Cav1 mRNA in cells treated with and without CoCl2 was determined by quantitative real-time RT-PCR. D, the Cav1 protein level was increased in all cell lines when CoCl2 was added. HIF1α/β-actin and Cav1/β-actin ratio determined by band intensity is reported for Western blot. E, Cav1 level was enhanced in hypoxic regions of xenografts derived from PLC cells. Xenografts developed in mice by subcutaneous injection of PLC cells were removed and subjected to H&E and IHC staining of Cav1 and pimonidazole. Representative images of IHC staining showing the hypoxic region of tumor xenograft. Magnification, ×10.

Figure 1.

Elevated expression of Cav1 in HCC cells under hypoxia. A, Upregulation of the Cav1 mRNA level in HCC cell lines under hypoxia. HCC cells were placed in hypoxic condition for 24 hours. Quantitative real-time RT-PCR was used to determine the expression of Cav1 mRNA in normoxic and hypoxic conditions. B, Hypoxia enhanced the Cav1 protein level. Western blot analysis was used to determine the expression of Cav1 protein in cells grown in normoxia and hypoxia. HIF1α/β-actin and Cav1/β-actin ratio determined by band intensity is reported for Western blot. C, CoCl2 treatment elevated expression of Cav1 mRNA. Expression level of Cav1 mRNA in cells treated with and without CoCl2 was determined by quantitative real-time RT-PCR. D, the Cav1 protein level was increased in all cell lines when CoCl2 was added. HIF1α/β-actin and Cav1/β-actin ratio determined by band intensity is reported for Western blot. E, Cav1 level was enhanced in hypoxic regions of xenografts derived from PLC cells. Xenografts developed in mice by subcutaneous injection of PLC cells were removed and subjected to H&E and IHC staining of Cav1 and pimonidazole. Representative images of IHC staining showing the hypoxic region of tumor xenograft. Magnification, ×10.

Close modal

Hypoxia-induced Cav1 promotes cell migration, invasiveness, and metastasis of HCC

Intratumoral hypoxia associates with enhanced tumor cell metastasis, thus the effect of hypoxia on HCC cells was evaluated both in vitro and in vivo. HCC cells showed significant enhancement in migration and invasiveness under hypoxia in vitro (Supplementary Fig. S3). To study the effect of hypoxia on HCC metastasis, hepatic artery ligation (18) was used to obstruct hepatic artery blood flow to promote the intratumoral hypoxic environment in the liver. Growth and metastasis of tumors, derived from luciferase-labeled metastatic MHCC97L cells, in the liver were then examined. Histological analysis revealed tumors of the HAL-treated group display a more invasive growth front than tumors of the control group (Supplementary Fig. S4). The HAL-treated group also resulted in higher incidence of distant metastases in the lungs compared with the control group. These findings prompted us to explore whether hypoxia-induced Cav1 plays a role in conferring aggressiveness to HCC cells under hypoxic stress. Hep3B and MHCC97L Cav1 knockdown stable clones were established and verified by Western blot analysis (Fig. 2A and C). In accordance with the previous findings (10), both migration and invasion of Cav1 knockdown stable clones were significantly reduced under normoxia (Fig. 2B and D). Hypoxia significantly enhanced the cell migration (Hep3B P < 0.05; MHCC97L P < 0.001) and invasion (Hep3B P < 0.001; MHCC97L P < 0.001) of the control nontarget cells when compared with cells cultured in normoxic condition. However, such enhancement in migration and invasion was not prominent when Cav1 expression was suppressed in Cav1 knockdown cells. Tumor seeds derived from MHCC97L nontarget control and Cav1 knockdown cells were implanted into liver of mice with or without receiving HAL (Fig. 2E and F). HAL increased both expressions of Cav1 and carbonic anhydrase IX (CA9; Fig. 2G). Functionally, HAL treatment enhanced the tumor growth and lung metastasis in control cells. Largest amount of metastatic tumor focus was observed in the lung tissues of animals subjected to HAL. However, this enhancement was suppressed in mice implanted with Cav1 knockdown tumor cells. These results suggest that Cav1 contributes significantly to the cell motility and metastasis of HCC cells under hypoxic stress.

Figure 2.

Hypoxia-induced Cav1 promotes HCC cell motility, tumorigenesis, and metastasis. Stable knockdown clones of Cav1 (sh-8001 and sh-8002) and nontarget control cells (sh-Ctl) were established in Hep3B (A) and MHCC97L cells (C). Expression of Cav1 in stable clones was revealed by Western blot analysis. HIF1α/α-tubulin and Cav1/α-tubulin ratio determined by band intensity is reported for Western blot. The Transwell and invasion assays were performed on Cav1 stable knockdown clones and nontarget control clone under normoxia and hypoxia. Knockdown of Cav1 suppressed hypoxia-induced cell migration (left) and invasiveness (right) in Hep3B (B) and MHCC97L (D) cells. Experiments were performed in triplicates. Results are expressed as mean ± SD of values. E, Mice implanted with tumor seed derived from Cav1 knockdown clone, sh-8001 or nontarget control; sh-Ctl were treated with or without HAL. Bioluminescence imaging of mice at the end of experiment (left). Images of the bioluminescence imaging of dissected livers are shown. Luciferase signal intensity was compared between the four experimental groups (right). F, Bioluminescence imaging of dissected lungs (left). H&E staining of lung tissues (middle). Arrows, the metastatic tumor focus found in the lung tissues. Summary of incidence of tumor formation and lung metastasis in each experimental group (right). P < 0.05 indicates statistically significant. G, HAL increased Cav1 and CA9 expressions. The hypoxic condition in the liver of mice that underwent HAL was indicated by carbonic anhydrase IX (CA9) staining (left). IHC staining revealed the expression of Cav1 in liver tumor (right).

Figure 2.

Hypoxia-induced Cav1 promotes HCC cell motility, tumorigenesis, and metastasis. Stable knockdown clones of Cav1 (sh-8001 and sh-8002) and nontarget control cells (sh-Ctl) were established in Hep3B (A) and MHCC97L cells (C). Expression of Cav1 in stable clones was revealed by Western blot analysis. HIF1α/α-tubulin and Cav1/α-tubulin ratio determined by band intensity is reported for Western blot. The Transwell and invasion assays were performed on Cav1 stable knockdown clones and nontarget control clone under normoxia and hypoxia. Knockdown of Cav1 suppressed hypoxia-induced cell migration (left) and invasiveness (right) in Hep3B (B) and MHCC97L (D) cells. Experiments were performed in triplicates. Results are expressed as mean ± SD of values. E, Mice implanted with tumor seed derived from Cav1 knockdown clone, sh-8001 or nontarget control; sh-Ctl were treated with or without HAL. Bioluminescence imaging of mice at the end of experiment (left). Images of the bioluminescence imaging of dissected livers are shown. Luciferase signal intensity was compared between the four experimental groups (right). F, Bioluminescence imaging of dissected lungs (left). H&E staining of lung tissues (middle). Arrows, the metastatic tumor focus found in the lung tissues. Summary of incidence of tumor formation and lung metastasis in each experimental group (right). P < 0.05 indicates statistically significant. G, HAL increased Cav1 and CA9 expressions. The hypoxic condition in the liver of mice that underwent HAL was indicated by carbonic anhydrase IX (CA9) staining (left). IHC staining revealed the expression of Cav1 in liver tumor (right).

Close modal

mRNA profiling identifies S100P as the downstream target of Cav1 under hypoxia

In an attempt to elucidate the molecular mechanisms by which Cav1 mediates HCC cell motility and metastasis, a genome-wide mRNA expression profiling was used to compare the gene expression profiles between Hep3B Cav1 knockdown cells and the nontarget control cells. Using a fold change of 2 as cutoff, 220 differentially expressed genes were identified, including 77 downregulated genes and 143 upregulated genes (Supplementary Table S1). Among these deregulated genes, S100P was found to be downregulated by 4-fold in the Cav1 knockdown cells. S100P has been shown to be aberrantly elevated in multiple human cancer cell lines and carcinomas (19). Moreover, the effect of S100P has also been shown to orchestrate the aggressive phenotype of cancerous cells. Downregulation of S100P was subsequently validated at both transcriptional and translational levels by real-time qPCR and Western blotting, respectively (Supplementary Fig. S5). In both Hep3B and MHCC97L nontarget control cells, S100P expression was significantly elevated in hypoxia (P < 0.001). However, in Cav1 knockdown cells, S100P expression was markedly reduced and was not elevated under hypoxia.

Restoration of S100P expression in Cav1 knockdown cells enhances HCC tumorigenesis and metastasis

To answer whether Cav1-regulated HCC cell migration and invasion is mediated by S100P, S100P was reintroduced in MHCC97L Cav1 knockdown cells (Fig. 3A). Expression of S100P in Cav1 knockdown cells significantly restored the cell migration and invasiveness (Fig. 3B). Treatment of recombinant S100P also restored the cell migration and invasiveness of Cav1 knockdown cells and such restoration was abolished when S100P-neutralizing antibody was added (Fig. 3C). To further examine the involvement of Cav1-mediated S100P downregulation in HCC tumorigenesis and metastasis, stable clones of nontarget control cells, Cav1 knockdown cells and Cav1 knockdown cells with S100P, were subcutaneously injected into nude mice. Expression of S100P in Cav1 knockdown clones (sh8001/S100P) formed bigger tumors when compared with tumors developed from Cav1 knockdown cells (sh8001; Fig. 3D; Supplementary Fig. S6). To determine whether S100P can restore metastasis in vivo, orthotopic liver implantation was performed. At the end of 7 weeks, restoration of S100P was found to enhance tumor development in liver (Fig. 3E) and incidence of distant metastasis to lungs when compared with the animal implanted with tumors derived from Cav1 knockdown cells (P < 0.05; Fig. 3F). Taken together, our findings show that S100P is a functional component downstream of Cav1 in promoting HCC aggressive phenotype. Restoration of S100P in Cav1 knockdown cells was also performed in animal model subjected to HAL (Supplementary Fig. S7). However, HAL procedure could not further enhance growth of liver tumor and distant metastasis. We suspect that tumor growth and metastases to lungs, which were reflected by the bioluminescence intensity, had already reached saturation, thus further enhancement in signal could not be revealed.

Figure 3.

Cav1 enhances HCC cell aggressiveness via the elevation of S100P. A, S100P was stably expressed in MHCC97L Cav1 stable knockdown clones. Western blot analysis revealed the expression of S100P in Cav1 stable knockdown clones (sh-8001/S100P). Cav1/α-tubulin and S100P/α-tubulin ratio determined by band intensity is reported for Western blot. B, Transwell assay (left) and invasion assay (right) were performed on nontarget control (sh-Ctl), sh-8001 Cav1 knockdown cells, and sh-8001/S100P cells. The number of cells migrated and invaded was reduced when Cav1 was knocked down. Restoration of S100P expression augmented the migration and invasion in sh-8001/S100P cells. C, Addition of recombinant S100P in sh-8001 cells resulted in increased migration (left) and invasion (right) when compared with sh-8001 cells without the addition of recombinant S100P. The enhancement in migration and invasion was abrogated when S100P-neutralizing antibody was added. Experiments were performed in triplicate. Mean ± SD of values were calculated and shown. D, Subcutaneous injection (n = 5 per group) was performed with sh-Ctl, sh8001, and sh-8001/S100P cells (left). Tumors formed were excised and weighed at the end of the experiment (middle). E, Orthotopic liver implantation of tumor seed derived from sh-Ctl, sh-8001, and sh-8001/S100P cells (n = 5 per group). Bioluminescence imaging of animals at the end of the experiment (left). Bioluminescence imaging of the dissected liver tissues (middle). Luciferase intensity was compared between different groups (right). F, Bioluminescence imaging of dissected lungs (left). Luciferase signal was plotted and compared between groups (middle). Representative images showing H&E staining of lung tissues (right). Arrows, the metastatic tumor focus found in the lung tissues. Magnification, ×10. P < 0.05 indicates statistically significant.

Figure 3.

Cav1 enhances HCC cell aggressiveness via the elevation of S100P. A, S100P was stably expressed in MHCC97L Cav1 stable knockdown clones. Western blot analysis revealed the expression of S100P in Cav1 stable knockdown clones (sh-8001/S100P). Cav1/α-tubulin and S100P/α-tubulin ratio determined by band intensity is reported for Western blot. B, Transwell assay (left) and invasion assay (right) were performed on nontarget control (sh-Ctl), sh-8001 Cav1 knockdown cells, and sh-8001/S100P cells. The number of cells migrated and invaded was reduced when Cav1 was knocked down. Restoration of S100P expression augmented the migration and invasion in sh-8001/S100P cells. C, Addition of recombinant S100P in sh-8001 cells resulted in increased migration (left) and invasion (right) when compared with sh-8001 cells without the addition of recombinant S100P. The enhancement in migration and invasion was abrogated when S100P-neutralizing antibody was added. Experiments were performed in triplicate. Mean ± SD of values were calculated and shown. D, Subcutaneous injection (n = 5 per group) was performed with sh-Ctl, sh8001, and sh-8001/S100P cells (left). Tumors formed were excised and weighed at the end of the experiment (middle). E, Orthotopic liver implantation of tumor seed derived from sh-Ctl, sh-8001, and sh-8001/S100P cells (n = 5 per group). Bioluminescence imaging of animals at the end of the experiment (left). Bioluminescence imaging of the dissected liver tissues (middle). Luciferase intensity was compared between different groups (right). F, Bioluminescence imaging of dissected lungs (left). Luciferase signal was plotted and compared between groups (middle). Representative images showing H&E staining of lung tissues (right). Arrows, the metastatic tumor focus found in the lung tissues. Magnification, ×10. P < 0.05 indicates statistically significant.

Close modal

Cav1 mediates the upregulation of S100P via the NF-κB pathway

Based on the apparent coregulation of Cav1 and S100P in HCC cell lines, we examined how Cav1 upregulates S100P. As shown in Fig. 4A, Cav1 was also able to significantly activate NF-κB cis-acting reporter (P < 0.001). Moreover, transient expression of Cav1 upregulated S100P and phospho-p65 (S276) in HEK293 cells. Addition of the NF-κB inhibitor IMD-0354 abolished the upregulation of S100P by Cav1 (Fig. 4B) and this inhibition of NF-κB was verified by the abrogation of the phospho-p65 (Ser276) level. Consistently, the expression level of phospho-p65 (Ser276) was suppressed in MHCC97L Cav1 stable knockdown cells when compared with the nontarget control cells (Fig. 4C). Our data therefore suggest that Cav1 upregulates S100P via the activation of the NF-κB pathway. It was shown that overexpression of IKKβ elevated the expression of S100P but not Cav1 in Cav1 knockdown clones (Fig. 4D). This therefore clearly demonstrates that Cav1 acts upstream of IKKβ in regulating the expression of S100P. IKKβ was transiently expressed in Cav1 knockdown cells (Fig. 4E). Functionally, expression of IKKβ significantly enhanced the cell migration (P < 0.001) and invasion (P < 0.05) in Cav1 knockdown clones (Fig. 4F).

Figure 4.

Expression of IKKβ in Cav1 knockdown cells restores S100P expression and HCC cell aggressiveness. A, Cav1 activated the NF-κB pathway. NF-κB luciferase reporter and Renilla luciferase plasmid were transfected with S100P or Cav1 in HEK293 cells. After 24 hours, luciferase activity was determined. Renilla luciferase activity was used for the normalization of transfection efficiency. Experiments were performed in triplicates. Mean ± SD of values were calculated and is shown. B, Cav1 upregulated phospho-p65 and S100P expressions. Transient transfection of Cav1 in HEK293 cells elevated expression of phospho-p65 and S100P as revealed by Western blot analysis. Such elevation was abolished in cells treated with IMD-0354. C, Activity of NF-κB and expression of S100P was suppressed in MHCC97L Cav1 stable knockdown clones. Inhibition of NF-κB activity was revealed by the reduced level of phospho-p65 in cells. D, Transient expression of IKKβ in Cav1 knockdown cells (sh-8001 and sh-8002) restored S100P expression. E, Western blot analysis revealed the restoration of S100P expression by transfecting IKKβ plasmid to Cav1 stable knockdown clones (sh-8001/IKKβ). Ratio of protein expression relative to α-tubulin or β-actin determined by band intensity is shown. F, Migration and invasion assays comparing the migratory and invasive potentials of sh-Ctl, sh-Cav1, and sh-8001/IKKβ cells. Experiments were performed in triplicates. Mean ± SD of values were calculated and is shown. P < 0.05 indicates statistically significant.

Figure 4.

Expression of IKKβ in Cav1 knockdown cells restores S100P expression and HCC cell aggressiveness. A, Cav1 activated the NF-κB pathway. NF-κB luciferase reporter and Renilla luciferase plasmid were transfected with S100P or Cav1 in HEK293 cells. After 24 hours, luciferase activity was determined. Renilla luciferase activity was used for the normalization of transfection efficiency. Experiments were performed in triplicates. Mean ± SD of values were calculated and is shown. B, Cav1 upregulated phospho-p65 and S100P expressions. Transient transfection of Cav1 in HEK293 cells elevated expression of phospho-p65 and S100P as revealed by Western blot analysis. Such elevation was abolished in cells treated with IMD-0354. C, Activity of NF-κB and expression of S100P was suppressed in MHCC97L Cav1 stable knockdown clones. Inhibition of NF-κB activity was revealed by the reduced level of phospho-p65 in cells. D, Transient expression of IKKβ in Cav1 knockdown cells (sh-8001 and sh-8002) restored S100P expression. E, Western blot analysis revealed the restoration of S100P expression by transfecting IKKβ plasmid to Cav1 stable knockdown clones (sh-8001/IKKβ). Ratio of protein expression relative to α-tubulin or β-actin determined by band intensity is shown. F, Migration and invasion assays comparing the migratory and invasive potentials of sh-Ctl, sh-Cav1, and sh-8001/IKKβ cells. Experiments were performed in triplicates. Mean ± SD of values were calculated and is shown. P < 0.05 indicates statistically significant.

Close modal

Potential therapeutic value of silencing S100P

Our data show that Cav1 upregulates S100P via the activation of the NF-κB pathway. To test whether S100P could be a potential therapeutic target for HCC, we functionally characterize metastatic MHCC97L cells after the suppression of endogenous S100P by lentiviral based knockdown. S100P expression was stably silenced by two different shRNA targeting S100P (Fig. 5A). Diminution of S100P significantly inhibited anchorage-independent growth (P < 0.001), migration (P < 0.001), and invasiveness (P < 0.001) of MHCC97L cells (Fig. 5B and C). More importantly, suppression of S100P expression largely inhibited tumorigenesis through subcutaneous injection (Fig. 5D). Smaller tumors were formed when S100P was knocked down when compared with the control cells (P < 0.001). We next tested if silencing S100P could reduce metastasis using an orthotopic liver implantation model. It was revealed that mice implanted with S100P-knockdown tumor cells had significantly smaller tumor formation in the liver (P < 0.01, Fig. 5E; Supplementary Fig. S8) and significant reduction in lung metastasis (P < 0.05, Fig. 5F) when compared with mice implanted with tumors derived from the nontarget control cells.

Figure 5.

Silencing of S100P has potential therapeutic value in inhibiting HCC tumorigenesis and metastasis. A, Western blot analysis revealed the silenced expression of S100P in MHCC97L S100P stable knockdown clones (sh-S100P#1 and sh-S100P#2). Ratio of S100P/α-tubulin determined by band intensity is shown. B, Soft agar assay was performed to examine the anchorage-independent growth of cells. C, Transwell assay (left) and invasion assay (right) were performed on nontarget control (sh-Ctl), and S100P stable knockdown cells. The number of migrated and invaded cells was reduced when S100P was knocked down. Experiments were performed in triplicates. Mean ± SD of values were calculated and shown. D, Subcutaneous injection (n = 5 per group) was performed with the nontarget control clone (sh-Ctl) and S100P stable knockdown cells (left). Tumors formed were excised and weighed at the end of the experiment (middle). Image of excised tumors is shown. E, Orthotopic liver implantation of tumor seed derived from nontarget control (sh-Ctl) and S100P knockdown cells (sh-S100P#1 and sh-S100P#2; n = 5 per group). Bioluminescence imaging of animals at the end of the experiment (left). Bioluminescence imaging of the dissected liver tissues (middle). Luciferase intensity was compared between different groups (right). Significant reduction in luciferase intensity in S100P knockdown groups when compared with the nontarget control group (P < .001) F, Bioluminescence imaging of dissected lungs (left). Luciferase signal was plotted and compared between groups (middle). Luciferase signal was significantly reduced in S100P knockdown groups. Representative images show H&E staining of lung tissues (right). Magnification, ×10. P < 0.05 indicates statistically significant.

Figure 5.

Silencing of S100P has potential therapeutic value in inhibiting HCC tumorigenesis and metastasis. A, Western blot analysis revealed the silenced expression of S100P in MHCC97L S100P stable knockdown clones (sh-S100P#1 and sh-S100P#2). Ratio of S100P/α-tubulin determined by band intensity is shown. B, Soft agar assay was performed to examine the anchorage-independent growth of cells. C, Transwell assay (left) and invasion assay (right) were performed on nontarget control (sh-Ctl), and S100P stable knockdown cells. The number of migrated and invaded cells was reduced when S100P was knocked down. Experiments were performed in triplicates. Mean ± SD of values were calculated and shown. D, Subcutaneous injection (n = 5 per group) was performed with the nontarget control clone (sh-Ctl) and S100P stable knockdown cells (left). Tumors formed were excised and weighed at the end of the experiment (middle). Image of excised tumors is shown. E, Orthotopic liver implantation of tumor seed derived from nontarget control (sh-Ctl) and S100P knockdown cells (sh-S100P#1 and sh-S100P#2; n = 5 per group). Bioluminescence imaging of animals at the end of the experiment (left). Bioluminescence imaging of the dissected liver tissues (middle). Luciferase intensity was compared between different groups (right). Significant reduction in luciferase intensity in S100P knockdown groups when compared with the nontarget control group (P < .001) F, Bioluminescence imaging of dissected lungs (left). Luciferase signal was plotted and compared between groups (middle). Luciferase signal was significantly reduced in S100P knockdown groups. Representative images show H&E staining of lung tissues (right). Magnification, ×10. P < 0.05 indicates statistically significant.

Close modal

Expressions of Cav1 and S100P are well correlated in HCC tissues

Western blot analysis revealed that S100P expression was well correlated with Cav1 expression in a panel of HCC cell lines (Fig. 6A). Despite the identification of S100P in HCC tissues, revealed by IHC elsewhere (20), its expression in paired HCC tumorous and extrahepatic metastatic tissues has not been examined. Here, S100P expression was studied in 36 sets of clinical samples, comprising of nontumorous liver tissues, primary HCCs, and extrahepatic metastatic tissues (Fig. 6B). The result showed that S100P staining was not detected in nontumorous tissues, while moderate or strong staining of S100P in cytoplasm was detected in 14 of 36 (39%) paired tumorous and extrahepatic metastatic tissues of the same patient. To determine the clinical relevance of S100P in HCC, S100P expression was revealed in 87 cases of paired tumorous and nontumorous liver tissues (Table 1). S100P was overexpressed in 34% (30/87) of the cases. Clinicopathological correlation revealed that overexpression of S100P was significantly associated with venous invasion (P = 0.003), microsatellite (P = 0.015), direct liver invasion (P = 0.045) and the absence of tumor encapsulation (P = 0.007). Phospho-p65 was overexpressed in 28.7% (25/87) of the cases and was correlated with the S100P protein expression revealed by IHC (P < 0.001, Fig. 6C).

Figure 6.

Clinical relevance of S100P in HCC. A, Expression of Cav1 and S100P in multiple HCC cell lines. Cav1 and S100P expressions were determined in a panel of HCC cells and an immortalized liver cell line, MIHA, by Western blot analysis. Ratio of Cav1/α-tubulin and S100P/α-tubulin determined by band intensity is shown. B, IHC of S100P expression in HCC. A representative case of HCC with matched nontumorous liver tissue, tumorous tissue, and distant metastatic tissue is shown. High expression of S100P was detected in nontumorous tissues and paired metastatic tissues. C, Correlation of S100P and phospho-p65 expressions in HCC. Representative images of S100P and phospho-p65 IHC staining in paired tumor and nontumor tissues. Magnification, ×20. D, Schematic diagram showing the pathway of Cav1/S100P in HCC. Under hypoxia, HIF1α induces the transcription of Cav1. Cav1 in the cytoplasm leads to the activation of the NF-κB pathway. Activated NF-κB induces the expression of S100P, leading to the promoting effect in HCC cell migration, invasiveness, and metastasis. Broken line, the tentative activation pathway.

Figure 6.

Clinical relevance of S100P in HCC. A, Expression of Cav1 and S100P in multiple HCC cell lines. Cav1 and S100P expressions were determined in a panel of HCC cells and an immortalized liver cell line, MIHA, by Western blot analysis. Ratio of Cav1/α-tubulin and S100P/α-tubulin determined by band intensity is shown. B, IHC of S100P expression in HCC. A representative case of HCC with matched nontumorous liver tissue, tumorous tissue, and distant metastatic tissue is shown. High expression of S100P was detected in nontumorous tissues and paired metastatic tissues. C, Correlation of S100P and phospho-p65 expressions in HCC. Representative images of S100P and phospho-p65 IHC staining in paired tumor and nontumor tissues. Magnification, ×20. D, Schematic diagram showing the pathway of Cav1/S100P in HCC. Under hypoxia, HIF1α induces the transcription of Cav1. Cav1 in the cytoplasm leads to the activation of the NF-κB pathway. Activated NF-κB induces the expression of S100P, leading to the promoting effect in HCC cell migration, invasiveness, and metastasis. Broken line, the tentative activation pathway.

Close modal
Table 1.

Clinicopathologic correlation of S100P expression in human HCC

No S100P overexpressionS100P overexpression
(No. of cases)(No. of cases)P
Sex Male 48 24 0.766 
 Female  
Cirrhotic liver Cirrhosis 22 0.345 
 No cirrhosis & hepatitis 35 22  
HBsAg Positive 48 24 0.756 
 Negative  
Cell Differentiation Poor 23 18 0.172 
 Differentiated 31 12  
Tumor size >5 cm 37 23 0.333 
 ≤5 cm 20  
Tumor encapsulation Absent 30 25 0.007a 
 Present 24  
Venous invasion Present 21 22 0.003a 
 Absent 35  
Microsatellite Present 22 20 0.015a 
 Absent 35 10  
Direct liver invasion Present 12 15 0.045a 
 Absent 28 12  
Tumor nodule N ≥ 2 10 0.768 
 N = 1 33 23  
No S100P overexpressionS100P overexpression
(No. of cases)(No. of cases)P
Sex Male 48 24 0.766 
 Female  
Cirrhotic liver Cirrhosis 22 0.345 
 No cirrhosis & hepatitis 35 22  
HBsAg Positive 48 24 0.756 
 Negative  
Cell Differentiation Poor 23 18 0.172 
 Differentiated 31 12  
Tumor size >5 cm 37 23 0.333 
 ≤5 cm 20  
Tumor encapsulation Absent 30 25 0.007a 
 Present 24  
Venous invasion Present 21 22 0.003a 
 Absent 35  
Microsatellite Present 22 20 0.015a 
 Absent 35 10  
Direct liver invasion Present 12 15 0.045a 
 Absent 28 12  
Tumor nodule N ≥ 2 10 0.768 
 N = 1 33 23  

NOTE: The Fisher exact test was used for the statistical analysis.

aSignificant correlation when P < 0.05.

Cav1 has been recognized as a janus-faced tumor regulator in cancers of different types and stages. Both oncogenic and tumor-suppressing roles of Cav1 have been suggested. Despite its ambiguous roles in various cancers, emerging evidence has shown a trend in the positive correlation of Cav1 expression with metastatic potential. Our previous study has shown that Cav1 acts as a tumor promoter in HCC, in which cell growth, motility, and metastasis were enhanced (10). In this study, we further showed how Cav1 plays an imperative role in HCC metastasis under hypoxia. Cav1 expression was upregulated at the transcriptional and translational levels under hypoxia in both non-metastatic and metastatic HCC cell lines. Indeed, Cav1 has been shown to be a direct transcriptional target of hypoxia-inducible factor (HIF)-1 and -2 (17). A conserved hypoxia response element was identified in the promoter of Cav1 and an HIF-dependent increase in Cav1 has been shown to promote its tumor cell proliferative, migratory, and invasive potentials (17). Despite these observations, the molecular basis underlying the functions of hypoxia-induced Cav1 remains unresolved. Using mRNA gene profiling, we identified a number of candidate genes to be differentially expressed in Cav1 knockdown cells. Among the candidate list, S100P, which has been implicated in cancer metastasis, was further analyzed. In both non-metastatic Hep3B and metastatic MHCC97L Cav1 knockdown cells, expression of S100P was verified. In concordance with Cav1, S100P was upregulated in hypoxic condition. Intriguingly, hypoxia-induced elevation of S100P was diminished when Cav1 expression was silenced, indicating that Cav1 acts upstream of S100P under hypoxia.

S100P belongs to the S100 family of calcium binding proteins, which exist as secretory or intracellular regulators. Increased level of S100P has been detected in several tumor cell lines and tissues (19). Many studies have also reported the potential role of S100P in carcinogenesis (21, 22). Regarding the effect of S100P in cancer metastasis, S100P has been shown by animal models to promote metastasis in lung cancer and colon cancers (23–25). Interaction of S100P with ezrin has been implicated in the control of tumor cell migration (26). However, very few studies have so far investigated the functions and clinical relevance of S100P in HCC. In a study reported by Kim and colleagues, S100P has been shown to be expressed at a higher level in HCC cell lines than in cell lines derived from normal hepatocytes (20). Moreover, elevated expression of S100P has been observed in HCC tissues than in normal tissue counterparts. In the present study, overexpression of S100P was detected in 34% of HCC cases, which is significantly associated with the aggressive features such as venous invasion, microsatellite formation, direct liver invasion, and absence of tumor encapsulation. This is also the first time to show overexpression of S100P in paired primary HCCs and distant metastatic tissues. These findings strongly suggest that S100P plays an imperative role in HCC. S100P gene activation is regulated by SMAD, STAT/CREB, and SP/KLF in different cancer cells (27); however, the upstream regulator of S100P in liver cancer was not studied. In this current study, we first reported that Cav1 induces expression of S100P through activation of the NF-κB pathway. NF-κB activation plays a critical role in the transcriptional response to hypoxia. In the normoxic condition, NF-κB is held inactive in the cytoplasm, and its activation occurs in response to various intra- and extracellular stimuli through the canonical and noncanonical pathway (28, 29). Hypoxia induces NF-κB activation via a canonical pathway by phosphorylation of the IKK complex. IKK subsequently mediates IκB phosphorylation and NF-κB translocation to the nucleus. Previous studies suggested that Cav1 can regulate this NF-κB activation (18, 30). Together, it is suggested that hypoxia-induced Cav1 overexpression activates NF-κB, leading to the modulation of its target molecules. We next studied the correlation between S100P and phospho-p65, the molecular effector of the NF-κB pathway. IHC revealed that expressions of S100P and phospho-p65 were significantly correlated. This suggests that the molecular pathway by which S100P expression is upregulated via the activation of NF-κB is physiologically relevant.

Most liver cancer patients are diagnosed at the late stage, and no treatment has been proven successful and thus, consequently, surgery is the main treatment for HCC. For patients with unresectable HCC, trans-arterial chemoembolization (TACE) is considered to be one of the most effective palliative measures. TACE involves the occlusion of hepatic arterial blood supply and focused administration of chemotoxins to the tumorous tissues. Despite the effectiveness of TACE, intratumoral hypoxia and metastasis are often observed after treatment. Several studies have shown that TACE inevitably induces hypoxia (31). Hypoxia has been shown to select aggressive cancer cells to adapt and survive in hypoxic condition. Thus, the understanding of the hypoxia-induced genes, which confer the aggressiveness of HCC cells, is important. Our work revealed that a high level of Cav1 was exclusively expressed in primary HCC and metastatic tissues and its level was further enhanced under hypoxic stress. Furthermore, S100P was identified to be the downstream effector of Cav1 under normoxic and hypoxic conditions. In the present study, we explored the potential therapeutic value of S100P inhibition by stable knockdown of S100P by RNA interference. Functional assays demonstrated that depletion of S100P expression in metastatic HCC cells significantly abrogated HCC cell growth, migration and invasiveness when compared with the nontarget control cells. In a nude mice model, knockdown of S100P significantly reduced tumor growth and metastasis. S100P is secretory; therefore, administrating neutralizing antibodies could also abolish its promoting effect in HCC cells. The effect of recombinant S100P in promoting HCC cell migration and invasiveness was shown to be neutralized by the addition of antibody against S100P.

In conclusion, we report a novel function of the Cav1/S100P pathway in hypoxia-induced HCC metastasis. Under hypoxic condition, HIF1 induces the expression of Cav1. Cav1 subsequently activates the NF-κB pathway. An upregulated expression of S100P was observed through the activation of NF-κB and, consequently, this increased expression contributes toward HCC tumorigenesis and metastasis (Fig. 6D). This newly identified pathway serves as a new molecular mechanism in elucidating HCC pathogenesis and thus provides numerous potential therapeutic prospects by targeting overexpressed S100P HCC cases in combination with TACE.

No potential conflicts of interest were disclosed.

Conception and design: X. Mao, E. Y, T. Tse, J.W.P. Yam

Development of methodology: X. Mao, E. Y, T. Tse, S.K. Tey, Y.S. Yeung, K. Man, J.W.P. Yam

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.Y.S. Wong, E. Y, T. Tse, F.C.F. Ko, K. Man, I.O.-L. Ng, J.W.P. Yam

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Mao, S.Y.S. Wong, F.C.F. Ko, R.C.-L. Lo, J.W.P. Yam

Writing, review, and/or revision of the manuscript: X. Mao, S.Y.S. Wong, J.W.P. Yam

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): F.C.F. Ko, J.W.P. Yam

Study supervision: J.W.P. Yam

This work was supported by the Health and Medical Research Fund (Project no. 02132846), Research Grants Council General Research Fund (Project no. 17102115), and the University of Hong Kong Seed Funding Programme for Basic Research. Faculty of Medicine Core Facility provides Xenogen imaging service.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Harris
AL
. 
Hypoxia–a key regulatory factor in tumour growth
.
Nat Rev Cancer
2002
;
2
:
38
47
.
2.
Yu
M
,
Dai
M
,
Liu
Q
,
Xiu
R
. 
Oxygen carriers and cancer chemo- and radiotherapy sensitization: bench to bedside and back
.
Cancer Treat Rev
2007
;
33
:
757
61
.
3.
Vaupel
P
,
Kelleher
DK
,
Hockel
M
. 
Oxygen status of malignant tumors: pathogenesis of hypoxia and significance for tumor therapy
.
Semin Oncol
2001
;
28
:
29
35
.
4.
Vered
M
,
Lehtonen
M
,
Hotakainen
L
,
Pirila
E
,
Teppo
S
,
Nyberg
P
, et al
Caveolin-1 accumulation in the tongue cancer tumor microenvironment is significantly associated with poor prognosis: an in-vivo and in-vitro study
.
BMC Cancer
2015
;
15
:
25
.
5.
Huang
WJ
,
Jeng
YM
,
Lai
HS
,
Fong
IU
,
Sheu
FY
,
Lai
PL
, et al
Expression of hypoxic marker carbonic anhydrase IX predicts poor prognosis in resectable hepatocellular carcinoma
.
PLoS One
2015
;
10
:
e0119181
.
6.
Goetz
JG
,
Lajoie
P
,
Wiseman
SM
,
Nabi
IR
. 
Caveolin-1 in tumor progression: the good, the bad and the ugly
.
Cancer Metastasis Rev
2008
;
27
:
715
35
.
7.
Burgermeister
E
,
Xing
X
,
Rocken
C
,
Juhasz
M
,
Chen
J
,
Hiber
M
, et al
Differential expression and function of caveolin-1 in human gastric cancer progression
.
Cancer Res
2007
;
67
:
8519
26
.
8.
Cokakli
M
,
Erdal
E
,
Nart
D
,
Yilmaz
F
,
Sagol
O
,
Kilic
M
, et al
Differential expression of Caveolin-1 in hepatocellular carcinoma: correlation with differentiation state, motility and invasion
.
BMC Cancer
2009
;
9
:
65
.
9.
Zhang
ZB
,
Cai
L
,
Zheng
SG
,
Xiong
Y
,
Dong
JH
. 
Overexpression of caveolin-1 in hepatocellular carcinoma with metastasis and worse prognosis: correlation with vascular endothelial growth factor, microvessel density and unpaired artery
.
Pathol Oncol Res
2009
;
15
:
495
502
.
10.
Tse
EY
,
Ko
FC
,
Tung
EK
,
Chan
LK
,
Lee
TK
,
Ngan
ES
, et al
Caveolin-1 overexpression is associated with hepatocellular carcinoma tumourigenesis and metastasis
.
J Pathol
2012
;
226
:
645
53
.
11.
Zhou
H
,
Jia
L
,
Wang
S
,
Wang
H
,
Chu
H
,
Hu
Y
, et al
Divergent expression and roles for caveolin-1 in mouse hepatocarcinoma cell lines with varying invasive ability
.
Biochem Biophys Res Commun
2006
;
345
:
486
94
.
12.
Wang
S
,
Jia
L
,
Zhou
H
,
Wang
X
,
Zhang
J
. 
Caveolin-1 promotes the transformation and anti-apoptotic ability of mouse hepatoma cells
.
IUBMB Life
2008
;
60
:
693
9
.
13.
Frerart
F
,
Lobysheva
I
,
Gallez
B
,
Dessy
C
,
Feron
O
. 
Vascular caveolin deficiency supports the angiogenic effects of nitrite, a major end product of nitric oxide metabolism in tumors
.
Mol Cancer Res
2009
;
7
:
1056
63
.
14.
John
TA
,
Ibe
BO
,
Usha Raj
J
. 
Oxygen alters caveolin-1 and nitric oxide synthase-3 functions in ovine fetal and neonatal lung microvascular endothelial cells
.
Am J Physiol Lung Cell Mol Physiol
2006
;
291
:
L1079
93
.
15.
Shen
J
,
Lee
W
,
Li
Y
,
Lau
CF
,
Ng
KM
,
Fung
ML
, et al
Interaction of caveolin-1, nitric oxide, and nitric oxide synthases in hypoxic human SK-N-MC neuroblastoma cells
.
J Neurochem
2008
;
107
:
478
87
.
16.
Tang
ZY
,
Ye
SL
,
Liu
YK
,
Qin
LX
,
Sun
HC
,
Ye
QH
, et al
A decade's studies on metastasis of hepatocellular carcinoma
.
J Cancer Res Clin Oncol
2004
;
130
:
187
96
.
17.
Wang
Y
,
Roche
O
,
Xu
C
,
Moriyama
EH
,
Heir
P
,
Chung
J
, et al
Hypoxia promotes ligand-independent EGF receptor signaling via hypoxia-inducible factor-mediated upregulation of caveolin-1
.
Proc Natl Acad Sci U S A
2012
;
109
:
4892
7
.
18.
Jiao
H
,
Zhang
Y
,
Yan
Z
,
Wang
ZG
,
Liu
G
,
Minshall
RD
, et al
Caveolin-1 Tyr14 phosphorylation induces interaction with TLR4 in endothelial cells and mediates MyD88-dependent signaling and sepsis-induced lung inflammation
.
J Immunol
2013
;
191
:
6191
9
.
19.
Jiang
H
,
Hu
H
,
Tong
X
,
Jiang
Q
,
Zhu
H
,
Zhang
S
. 
Calcium-binding protein S100P and cancer: mechanisms and clinical relevance
.
J Cancer Res Clin Oncol
2012
;
138
:
1
9
.
20.
Kim
JK
,
Jung
KH
,
Noh
JH
,
Eun
JW
,
Bae
HJ
,
Xie
HJ
, et al
Targeted disruption of S100P suppresses tumor cell growth by down-regulation of cyclin D1 and CDK2 in human hepatocellular carcinoma
.
Int J Oncol
2009
;
35
:
1257
64
.
21.
Gibadulinova
A
,
Tothova
V
,
Pastorek
J
,
Pastorekova
S
. 
Transcriptional regulation and functional implication of S100P in cancer
.
Amino Acids
2011
;
41
:
885
92
.
22.
Arumugam
T
,
Simeone
DM
,
Van Golen
K
,
Logsdon
CD
. 
S100P promotes pancreatic cancer growth, survival, and invasion
.
Clin Cancer Res
2005
;
11
:
5356
64
.
23.
Hsu
YL
,
Hung
JY
,
Liang
YY
,
Lin
YS
,
Tsai
MJ
,
Chou
SH
, et al
S100P interacts with integrin alpha7 and increases cancer cell migration and invasion in lung cancer
.
Oncotarget
2015
;
6
:
29585
98
.
24.
Shen
Z
,
Deng
H
,
Fang
Y
,
Zhu
X
,
Ye
GT
,
Yan
L
, et al
Identification of the interplay between SOX9 and S100P in the metastasis and invasion of colon carcinoma
.
Oncotarget
2015
;
6
:
20672
84
.
25.
Ding
Q
,
Chang
CJ
,
Xie
X
,
Xia
W
,
Yang
JY
,
Wang
SC
, et al
APOBEC3G promotes liver metastasis in an orthotopic mouse model of colorectal cancer and predicts human hepatic metastasis
.
J Clin Invest
2011
;
121
:
4526
36
.
26.
Austermann
J
,
Nazmi
AR
,
Muller-Tidow
C
,
Gerke
V
. 
Characterization of the Ca2+ -regulated ezrin-S100P interaction and its role in tumor cell migration
.
J Biol Chem
2008
;
283
:
29331
40
.
27.
Gibadulinova
A
,
Oveckova
I
,
Parkkila
S
,
Pastorekova
S
,
Pastorek
J
. 
Key promoter elements involved in transcriptional activation of the cancer-related gene coding for S100P calcium-binding protein
.
Oncol Rep
2008
;
20
:
391
6
.
28.
Culver
C
,
Sundqvist
A
,
Mudie
S
,
Melvin
A
,
Xirodimas
D
,
Rocha
S
. 
Mechanism of hypoxia-induced NF-kappaB
.
Mol Cell Biol
2010
;
30
:
4901
21
.
29.
Gilmore
TD
. 
Introduction to NF-kappaB: players, pathways, perspectives
.
Oncogene
2006
;
25
:
6680
4
.
30.
Garrean
S
,
Gao
XP
,
Brovkovych
V
,
Shimizu
J
,
Zhao
YY
,
Vogel
SM
, et al
Caveolin-1 regulates NF-kappaB activation and lung inflammatory response to sepsis induced by lipopolysaccharide
.
J Immunol
2006
;
177
:
4853
60
.
31.
Lai
JP
,
Conley
A
,
Knudsen
BS
,
Guindi
M
. 
Hypoxia after transarterial chemoembolization may trigger a progenitor cell phenotype in hepatocellular carcinoma
.
Histopathology
2015
;
67
:
442
50
.