Background
Hepatocellular carcinoma/cancer (HCC) is one of the most common malignant tumors worldwide [
1]. It most often develops in the background of underlying liver disease, such as hepatitis. Historically a challenge for Asian countries, the increasing incidence of hepatitis C has made HCC a major health problem within the United States in recent years [
2]. Early stage HCC is potentially curable with liver transplant or resection. However, most patients present with more advanced disease and for these patients' treatment options are limited. Inroads into effective therapies have been thwarted by a gap in our understanding of the molecular mechanisms involved in cancer development and progression within its complex microenvironment. Therefore, studies elucidating the mechanism and signaling pathways involved in HCC development and progression are imperative.
Previous microarray analysis from our laboratory identified autotaxin (ATX) as one a gene with enhanced mRNA expression in human hepatitis associated HCC [
3]. Reports from other labs showed that serum ATX activity and plasma lysophosphatidic acid (LPA) level are increased in various liver injuries in rats in relation to their severity [
4]. ATX was initially characterized as an autocrine motility factor from A2058 melanoma cell conditioned medium [
5]. It has been subsequently shown that ATX acts as an important mediator of tumorigenesis by stimulating angiogenesis, as well as survival, growth, migration, and invasion of tumor cells [
6‐
8]. In particular, recent studies using ATX knockout mice suggest that ATX contributes to tumor progression by stabilizing blood vessels in the vicinity of tumors [
9,
10]. Although ATX has been showed to affect adhesion through integrin-dependent focal adhesion assembly [
11,
12], the main impact of ATX on cancer biology is mostly due to its intrinsic lysophospholipase D (lyso-PLD) activity. Through the conversion of lysophosphatidylcholine (LPC) into LPA and to a less degree, sphingosylphosphorylcholine (SPC) into sphingosine-1-phosphate (S1P) [
13,
14], ATX regulates cell activation by changing signaling induced by LPC versus LPA.
LPA is an important lipid mediator that elicits a broad spectrum of biological effects by activating G protein-coupled receptors (GPCRs). The biological functions of LPA included, but not limited to cell proliferation, migration, platelet aggregation, smooth muscle contraction, and cytoskeletal reorganization. In the context of cancer, LPA could induce stress fiber formation, membrane ruffling, and lamellipodia formation [
15‐
17]. The aberrant ATX expression may lead to altered LPC/LPA balance and their receptor-mediated functions, resulting in enhanced tumor progression. Hence, the molecular events that lead to the aberrant ATX expression and the subsequent abnormal LPA production are significant for understanding the mechanisms involved in cancer progression. In this study, we examined the expression of ATX antigen in HCC tissue using immunohistochemistry. The regulatory mechanism of ATX by the key inflammatory component TNF-α/NF-κB axis was studied in human hepatoma cell lines. We also demonstrated that ATX is involved in the invasive potential of HCC cells.
Discussion
This is the first report to study the expression and the functional roles of ATX in human HCC. We showed for the first time that ATX protein was over-expressed in human HCC tissues compared with that in normal controls. Enhanced expression of ATX in HCC is significantly correlated with liver inflammation, cirrhosis, as well as risk factor such as hepatitis. ATX was also over-expressed in human hepatoma cell lines Hep3B and Huh7 cells compared to hepatoblastoma HepG2 and normal hepatocytes.
ATX-null mice embryos failed to develop into mature vessels and died at E11.5 [
9]. While ATX over-expression was presented in various cancers and promotes tumor progression by stimulating angiogenesis, tumor cell survival, growth, migration and invasion [
6‐
8]. Recently, Mills's lab demonstrated that expression of ATX or LPA receptor in mammary epithelium of transgenic mice contributes to the initiation and progression of estrogen receptor (ER)-positive, invasive, and metastatic mammary cancer [
28]. Our observations showing the overexpression of ATX in HCC tissues and cell lines, as well as its relative low levels in normal liver cell lines and tissues imply its important role in both liver physiological and pathological activities.
Being the key enzyme with lyso-PLD activity, the aberrant expression of ATX has the potential to alter the delicate balance between LPA signaling and LPC signaling in the local liver microenvironment. LPC is phospholipid with both proinflammatory activity and immunoregulatory activity by stimulating the expression of a serial of genes, including NO synthase, monocyte chemoattractant protein-1 (MCP-1), inter-cellular adhesion molecule (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1) and growth factors in endothelial cells [
29‐
33]. In addition, LPC has been reported to promote vascular smooth muscle cell proliferation, attract monocytes, inhibit endothelium dependent relaxation, reduce endothelial cell migration, and promote the development of mature dendritic cell [
34‐
37]. Moreover, LPC is also required for the cytotoxic response of human NK cells to tumor cells [
38]. On the other hand, LPA mediates a broad range of biological activities such as wound healing, vascular remodeling, and cell migration and survival. LPA and its analogs were also proposed to be critical endogenous mediators that regulate survival, motility, proliferation, and differentiation of oval cell/hepatocyte progenitors in liver regeneration [
39]. Oval cell proliferation was proposed to be associated with an increased risk for development of HCC with advancing liver disease, particularly when cirrhosis is present [
40]. Serum ATX activity and plasma LPA level were increased in chronic hepatitis C in association with liver fibrosis [
41]. Interestingly, we also found that ATX over-expression in HCC was associated with liver cirrhosis. The aberrant production of LPA may bind to its receptors and results in the altered activation of LPA signaling pathways, including, but no limited to activation of the PI3K-AKT, Ras/MEKK/MAPK, p38 MAPK, and JNK pathways. These signaling pathways have been shown to be actively involved in HCC development by controlling angiogenesis, cell motility, cell proliferation and survival [
42,
43]. Therefore, the aberrant expression of ATX along with the consequently abnormal production of LPA in the liver microenvironment may fuel the process of liver carcinogenesis.
Chronic inflammation has long been associated with the development of liver cancer. Three lines of evidence obtained from the current study support a link between ATX expression/function to inflammation in liver diseases. First, our immunohistochemistry data from human liver tissue showed the differential expression of ATX in HCC with different etiologies. Hepatitis literally means inflammation of the liver, and is the major cause of HCC. ATX expression in hepatitis-related HCC tissues is significantly elevated compared to those HCC tissues developed from non-cirrhotic "non-inflammatory" background which indeed show no signs inflammatory cell infiltration as we observed in the samples from patients with chronic active hepatitis or steatohepatitis. Secondly, the ATX expression levels correlated well to the derivative origins in HCC cell lines related to inflammation. Hep3B cells were derived from a patient with hepatitis and therefore may have unique response systems that are associated with the inflammatory background of a hepatitis infected liver [
44]. Rice lab showed that Huh7 cells had a favorable cellular environment for hepatitis C virus replication [
45]. We use it as a second cell line that may respond to inflammatory mediators thereby calling attention to unique signals associated with inflammatory associated cancers. But we need to be cautious on this data since Huh 7 cells are not well characterized although they were derived from a Japanese patient with well differentiated HCC [
46]. In contrast, HepG2 is derived from a human hepatoblastoma which almost always arise in an otherwise normal liver and is most unlikely to be associated with inflammation [
47,
48]. Hep3B and Huh7 cells, but neither HepG2 nor normal hepatocytes exhibit enhanced expression of ATX. As we are focusing on the mechanisms that potentially overlap between different etiologies of inflammatory induced HCC, whether there are any potential associations between viral antigens and the ATX/LPA pathway remain to be further studied. Finally, TNF-α, a pro-inflammatory cytokine, further promoted ATX secretion and LPA production in Hep3B and Huh7 cell lines. Moreover, the secreted enzymatically active ATX promoted Hep3B cell migration/invasion, which is dependent on extracellular LPC concentration and can be directly demonstrated by LPA's effect. Our mechanistic studies show that NF-κB activity is important for TNF-α 's activity. Our previous work showed NF-κB is constitutively activated in Hep3B, HepG2 and CL-48 cell lines, but Hep3B has the strongest basal NF-κB activity among these cell lines [
26]. The differential responsiveness of these cell lines to TNF-α in ATX stimulation suggest that these cells may also have differential signaling properties in responding to TNF-α, which remains to be further investigated.
Methods
Reagents
TNF-α, parthenolide and fatty acid-free BSA were purchased from Sigma (St. Louis, MO). LPC (1-oleoyl) was obtained from Avanti Polar Lipids, Inc. (Birmingham, AL). ATX activity assay reagents were from Echelon Biosciences, Inc. (Salt lake City, UT, USA). Purified recombinant ATX protein and rabbit polyclonal antibody against ATX and were generous gifts from Dr. Timothy Clair (National Cancer Institute, Bethesda, MD); the polyclonal antibody against ATX was prepared by immunization of rabbits with the peptide ARVRDIEHLTSLDFFRK.
Human liver tissue and cell lines
This study was approved by Indiana University Institutional Review Board. Liver tumor tissue was collected from patients undergoing resections for HCC at Indiana University Hospital. Normal tissue (n = 10) was obtained from patients undergoing non-liver disease related surgeries. The fresh tissue was formalin fixed, and paraffin embedded for immunohistochemistry. Thirty-eight HCC cases were applied for this study. Eleven of them had HCV infection, seven had HBV infection and ten had non-alcohol steatohepatitis (NASH), as confirmed by serological testing or PCR testing of benign or tumor DNA. Another ten cases of HCC samples were observed in non-cirrhotic liver and they developed in an otherwise normal liver and without identified risk factors. No pathological evidence of inflammatory infiltrates within the background liver was identified, and here they are named as normal-HCC.
CL-48, HepG2 and Hep3B cell lines were obtained from American Tissue Culture Collection and were cultured in Eagle's Minimum Essential Medium (EMEM) with 10% fetal bovine serum at 37°C, 5% CO2. Huh7 cell line was a generous gift provided by Dr. Charles M. Rice's lab (Rockefeller University, New York, NY) and was cultured in Dulbecco's Modified Eagle's Medium (DMEM) with 10% fetal bovine serum at 37°C, 5% CO2. Human normal primary hepatocytes were purchased from Lonza (Lonza, Walkersville, MD) and maintained in hepatocyte culture medium (Lonza, Walkersville, MD). Cells were serum-starved overnight and then treated with TNF-α (10 ng/ml) or parthenolide in serum free media containing 0.1% BSA. Total RNA was extracted or cell lysate was prepared after stimulation for the indicated time.
siRNA transfection
Small interfering RNA (siRNA) duplexe targeting human ATX and negative control siRNA were purchased from Ambion (Austin, TX). Cells were cultured to 60-70% confluency and then transfected with 10 nM of ATX siRNA or 30 nM negative siRNA using Transfection siPORT™ NeoFX™ kit (Ambion) according to the manufacturer's recommendations. Transfected cells were incubated at 37°C and 5% CO2 for 68-72 hours. Cells were harvested for total RNA or protein preparation and conditioned media were collected for invasion assay.
Conditioned media were prepared by incubating 70% confluent cells in 100 mm dishes for 24 hours in serum-free MEM or DMEM containing 0.1% fatty acid-free BSA. Conditioned media were harvested, clarified by centrifugation, and filtered through a 0.22 μm filter. The conditioned media were concentrated by Amicon Ultra-15 Centrifugal Filter Units before using for immunoblot. At the same time, total cell extracts were prepared from cell monolayer incubated in RIPA buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 2 mM EDTA, 1 mM sodium orthovanadate, 1% NonidetP40; 1% sodium deoxycholate; 0.1% sodium dodecylsulfate (SDS), 2 mM phenylmethylsulfonylfluoride (PMSF) and protease inhibitor cocktail. Fifteen micrograms of total cellular protein was resolved by SDS-PAGE. Blots were probed with appropriate antibodies. Anti-β-actin was used for loading control.
Quantitative real time RT-PCR (qRT-PCR)
Total RNA was isolated from cells using the RNeasy kit following the manufacturer's instructions (Qiagen, Valencia, CA). 2 μg total RNA was reverse-transcribed in a total reaction volume of 20 μl using the high capacity cDNA reverse transcriptase kit (Applied Biosystems, Foster City, CA) as described by the manufacturer. Single stranded cDNA products were then analyzed by real-time PCR using standard commercially available TaqMAN probes for ATX (Hs00196470_m1). The amount of target gene was normalized to the internal standard 18S rRNA (Hs99999901_s1) levels and reported as a relative value.
ATX/lyso-PLD activity assay
The conditioned serum-free medium from Hep3B and Huh7 cells with or without TNF-α stimulation was concentrated (40-fold) using Amicon Ultra 50,000 (Millipore). EMEM and DMEM without cells were used as control. ATX/lyso-PLD activity in concentrated conditioned media was analyzed using fluorogenic substrate FS-3 according to the manufacture protocol. Briefly, 10 ul concentrated medium was mixed with 5 uM FS-3 and assayed in 96-well plate. The change of fluorescent intensity was measured by SpectraMax Gemini EM Fluorescence Microplate Reader (Molecular Devices, Sunnyvale, CA, USA) with excitation and emission wavelengths of 485 and 528 nm, respectively.
Lipid extraction and analyses
Lipids were extracted from conditioned media and analyzed using LC-MS (API-4000, Applied Biosystems) [
49,
50]. Briefly, conditioned media were incubated with 15 μM LPC (18:1) for 3 hrs at 37°C. 1.3 mL samples were mixed with 3 mL of MeOH/chloroform (2:1) following the addition of 10 μL of 14:0 LPA (1 μM) as an internal standard and 10 μL of HCl (6N). The samples were vortexed for 1 min and incubated on ice for 10 min. Chloroform (1 mL) and PBS (1 ×) (0.5 mL) was added to separate the phases and samples were vortexed for 1 min prior to centrifugation (1,750 g for 10 minutes, at 10°C). The lower phase was transferred to a new glass tube. The upper phase was re-extracted using 2 mL chloroform and combined with the lower phase. After evaporating the solvent under nitrogen at room temperature, the dried lipids were re-suspended respectively in 100 μl of MeOH and 10 μL of sample will be used for Mass spectrometry (MS) analyses. Typical operating parameters for MS will be as follows: nebulizing gas (NEB) 15, curtain gas (CUR) 8, collision-activated dissociation (CAD) gas 35, electrospray voltage 5000 with positive-ion MRM mode, and a temperature of heater at 500°C. Precursor mode 153 will be set as the daughter ions of LPA. In MRM mode, negative monitoring ions will be at m/z 435 (the parent ion)-153 (the product ion) for 18:1 LPA. The dwell time in the MRM mode will be 75 ms. A TARGA C18 5 μM, 2.1 mm ID× 10 mm TR-0121-C185 (Higgins Analytical, Southborough, MA USA) HPLC column was used for the separation of different phospholipids and for the detection of LPAs. The mobile phase A was MeOH/water/NH4OH (90:10:0.1, v/v/v). The HPLC separations will be 12 min/sample using the following scheme: 1) 100% A for 3 min with a flow rate at 0.2 mL/min; 2) the mobile phase will be changed from 100% A to 100% B in 2 min with a flow rate from 0.2 to 0.8 mL/min; 3) a constant flow rate of 0.8 mL/min for 5 min; 4) the mobile phase will be changed from 100% B to 100% A in 1 min with a flow rate from 0.8 to 0.2 mL/min; and 5) constant flow rate of 0.2 mL/min for 1 min.
Cellular migration/invasion assays
According to previously described methods [
26], invasion assay was performed by using BD BioCoat™ Matrigel™ 24-well invasion chamber (8 μM pore size). In brief, cells were serum-starved overnight and re-suspended into serum free MEM containing 0.1% fatty acid-free BSA. 5 × 10
4 cells were added to the top insert, and 750 μl of conditioned medium with or without 1 μM LPC (18:1) was added to the bottom chamber. To determine the effect of LPA on the invasion, serum-free MEM containing 0.1% fatty acid-free BSA with or without LPA (0, 0.1, 1, 5 μM) were added to the bottom chamber. After 24 hours incubation at 37°C in a CO2 incubator, non-invaded cells were removed from the upper surface of the filter with the cotton swab; cells that migrated through the gel insert to the lower surface of the membrane were fixed with 100% methanol, stained with 1% Toluidine blue and counted using a light microscope at 50 × magnification. Each sample was tested in triplicate at least in two independent assays. Results were expressed as mean cell number per field ± SD.
Immunohistochemistry
Serial 5-micron thick sections of formalin-fixed paraffin embedded tissue were cleared with xylene and rehydrated through graded ethanol and finally immersion in distilled water. Slides were then rinsed in Tris-buffered saline (TBS). Antigen retrieval was performed by using the Dako Target Retrieval kit (Dako, Carpinteria, CA) containing a citrate buffer (pH 6.0) for 10 min at 95°C. Dako's Avidin Biotin blocking system was used for 10 min, and the tissue sections were then rinsed with TBS. The nonspecific binding sites were blocked by incubating with Dako's Protein Block for 10 min. Tissue sections were then incubated with the polyclonal rabbit antibody against ATX (7.8 μg/ml) overnight at 4°C. After washing with TBS, the secondary antibody, Dako Link (Dako LSAB2 kit) was applied for 20 min and then rinsed with TBS. Additional washing was followed by incubation with streptavidin horseradish peroxidase (Dako Label, LSAB2 kit) for 20 min. Immunoreactivity was visualized by incubation of sections with 3, 3'-diaminobenzidine in the presence of hydrogen peroxide. Sections were counterstained with light hematoxylin and mounted with a coverslip. All of the procedures were performed at room temperature except the primary antibody incubation. Microscopic fields evaluated and scored were those with the highest degree of immunoreactivity ("hot spots"). Five fields (40 × fields) per section were analyzed. An intensity score was assigned to each case on a scale from negative to high (0: no staining; +1: weak staining; +2: moderate staining; and +3: strong staining).
Statistical analysis
Data are presented as means ± SD. Analysis of the significance of differences between two groups was performed by two tailed student's t-test using Instat software (GraphPad, San Diego, CA). P-values of < 0.05 were considered statistically significant. Fisher's exact test was used for the ATX immunoreactivity analysis, and P value < 0.05 was deemed significant.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MAM supervised and coordinated the study and revised the manuscript. JMW designed the study, performed all the experiments, analyzed the data and prepared the manuscript. YX revised the manuscript and contributed to the LC-MS assay. ZZ performed the LC-MS assay. JS and HS reviewed the manuscript. RS contributed to human tissue acquisition, specimen pathology reviewing and immunohistochemistry data analysis. MY performed the statistic analysis of immunohistochemistry data. All authors read and approved the final manuscript.