Overexpression of Chemokine (C-X-C) ligand 1 (CXCL1) associated with tumor progression and poor prognosis in hepatocellular carcinoma

As previously described, 48 primary HCC specimens and normal adjacent tissues were collected from patients who underwent complete resection at the Shanghai First People's Hospital between Dec 2004 and May 2007 and were diagnosed with hepatic carcinoma (HCC) using histological diagnoses [17]. No local or systemic treatment had been received in these patients before surgical resection. Written informed consent was obtained from all patients involved in the study. This study was performed with the approval of the Medical Ethical Committee of Shanghai First People's Hospital and the ethical guidelines of the Declaration of Helsinki. All of the tissue samples were flash-frozen in liquid nitrogen immediately after collection and stored at -80°C until RNA was extracted.

The human hepatoma cell lines (SMMC-7721, Bel-7402, Hep3B and MHCC-97H) and non-malignant liver cells (LO2) were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured in High glucose Dulbecco's Modified Eagle Medium (DMEM) (HyClone, Shanghai, China) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA, USA), with 100 U/ml of penicillin G and 100 μg/ml of streptomycin at 37°C in a water-saturated atmosphere of 5% CO_2. For studies assessing the effect of NF-κB inhibition, cells were treated the NF-κB inhibitor ammonium pyrrolidinedithiocarbamate (PDTC) (Sigma-Aldrich, Poole, UK) at a final concentration of 20 μM.

As previously described [17], total RNA was isolated from HCC tissue specimens and cultured cells with TRIzol Reagent (Invitrogen) and then converted to first strand cDNA using iScript™ cDNA Synthesis Kit (Bio-Rad, CA, USA) according to the manufacturer's instructions. Real-time quantitative reverse-transcription polymerase chain reaction (qRT-PCR) was carried out using the SYBR Green PCR Master Mix (Toyobo, Osaka, Japan) according to the manufacturers' instructions by an ABI 7900HT real-time PCR system (Applied Biosystems, Foster City, CA, USA) and amplified with target gene specific primers. The primer sequences were as follows: 5′-GCCCAAACCGAAGTCATAGC-3′(sense) and 5′-GGCACAATCCAGGTGGCC-3′(antisense) for CXCL1; 5′-ACCCAGAAGACTGTGGATGG-3′(sense) and 5′-CAGTGAGCTTCCCGTTCAG-3′(antisense) for GAPDH. GAPDH was used as a loading control. The PCR reaction contained an initial denaturation at 95°C for 2 min, followed each PCR cycle by de-naturation at 95°C for 1 min, annealing and extension at 72°C for 1 min for a total of 40°Cycles. All experiments were performed in triplicate with three technical replicates.

To overexpression of CXCL1, we obtained the coding sequence (CDS) fragment of CXCL1 by PCR using the following primers: forward, 5′-CCCAAGCTTATGGCCCGCGCTGCTCTC-3′ and reverse, 5′-AAAGGATCCTCAGTTGGATTTGTCACTGTTCA-3′ (including HindIII and BamHI sites with underline, respectively). The resulting PCR product was then subcloned into pCDNA3.1 (+) vector (Invitrogen) and named it pCDNA3.1-CXCL1. The sequence of CXCL1 was confirmed by DNA sequencing (Sangon, Shanghai, China). Subconfluent cells were transfected with CXCL1 or empty vectors using X-tremeGENE (Roche, Basel, Switzerland) according to the manufacturer's protocol. The expression of CXCL1 protein was confirmed 48 hours after transfection by western blot analyses.

Cell proliferation assay was performed as recently described [17]. Briefly, differences in the proliferative capacity of 7721 and 7402 cells with or without overexpression of CXCL1 were examined by the absorbance using coulter counter (Beckman Coulter, CA, USA).

For the invasion assays, 24 hours after transfection, 3 × 10⁴ cells in 100 μl serum-free media were added to the upper chamber of an insert (8-μm pore size, millepore) coated with Matrigel (BD, San Diego, CA, USA). 600 μl media containing 10% FBS were added to the lower chamber. After incubation for 24 hours, the cells remaining on the upper membrane were removed with a cotton swab, whereas the cells that had invaded through the membrane were fixed and stained with methanol and 0.1% crystal violet. The images were acquired and counted from five randomly selected fields using an IX71 inverted microscope (Olympus, Tokyo, Japan). Experiments were independently repeated in triplicate.

As described before [17], aliquots of 40 μg of protein from each group was separated by 13% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a 0.22-μm polyvinylidene difluoride (PVDF) membrane (Millipore, Bedford, MA, USA). After blocking with 5% nonfat milk in TBS + 0.05% Tween™ 20 for 1 hour at room temperature, membranes were incubated with an Rabbit Anti-CXCL1 antibody (Proteintech, Chicago, IL, USA), Rabbit Anti-Phospho-NF-κB p65 (Ser536) antibody, Rabbit Anti-NF-κB P65 antibody (Cell signaling Technology, Beverly, MA, USA), or Anti-GAPDH (Bioworld, Nanjing, China) overnight at 4°C. Densitometric levels of CXCL1 signals were quantified and expressed as their ratio to GAPDH.

Immunohistochemical analyses were performed as described previously [17], paraffin-embedded tumor specimens were cut into 4-μm-thick sections and mounted on silane-coated slides. The slides were then treated with a graded series of ethanol solution. 3% H2O2 was applied to block endogenous peroxide activity for 15 min at room temperature. Antigen retrieval was performed at 95°C in 0.01 M sodium citrate buffer (pH 6.0) for 30 min and sections were blocked with normal goat serum. The sections were then incubated with a primary antibody against CXCL1 (Proteintech) in a humidity chamber at 4°C overnight. After washing with phosphate-buffered saline (PBS), the slides were incubated for 60 min with the DAKO Envision+/HRP system (DAKO, Carpinteria, CA, USA) and counterstained with hematoxylin. Negative control sections were incubated with normal rabbit IgG instead of a specific primary antibody.

Data are presented as the mean ± standard deviation (SD) from at least three independent experiments. Student's t test was employed to assess the differences between two groups and one-way analysis of variance (ANOVA) was used to analyze significant differences among three groups. Survival was determined using the Kaplan-Meier method, and the significance of the difference was evaluated using the log-rank test. P < 0.05 was considered to be statistically significant. All statistical analyses were performed with GraphPad Prism 5.0 software (La Jolla, CA, USA).

It has been recently reported that CXCL1 is high expressed in melanoma, breast cancer and could be associated with tumor differentiation grade and disease-specific survival of bladder tumors [12],[13],[15]. To determine CXCL1 expression in HCC tissues, we use qRT-PCR to analyze mRNA of CXCL1 in 48 HCC tissues and marched adjacent non-cancerous tissue. The level of CXCL1 mRNA was significantly increased in cancerous tissues compared with corresponding non-cancerous tissues (p < 0.01) (Figure 1A). To confirm the association between the expression of CXCL1 and HCC, four human HCC cell lines (SMMC-7721, Bel-7402, Hep3B and MHCC-97H) and a normal liver cell (LO2) were selected to detect the expression levels of CXCL1 mRNA and protein by qRT-PCR and western blot. Among the five cell lines analyzed, CXCL1 mRNA levels were found higher expression in HCC cells compared to LO2 cells (Figure 1B). Consistent with the results of mRNA, a significant increase in CXCL1 protein levels was also found in HCC cells as assessed by western blot (Figure 1C). These data indicate that abnormal CXCL1 might play important roles in HCC tumorigenesis.

Increased expression of CXCL1 in HCC tissues was further confirmed by immunochemistry (IHC) staining using an antibody to CXCL1. Representative images of positive expression of CXCL1 IHC staining were show in Figure 2. Then we determined the association of CXCL1 expression level and clinicopathological implications of 68 HCC patients in Table 1. The results demonstrated that high CXCl1 expression was significantly associated with advanced clinical stages (p = 0.034) and distant metastasis (p = 0.028). However, there was no significant correlation between CXCL1 expression and patient age, gender, tumor diameter, tumor number, AFP, and liver capsule.

We further examined whether CXCL1 expression level correlated with prognosis of HCC patients after surgery resection. Overall survival (OS) curves were plotted according to CXCL1 expression level by the Kaplan Meier analysis and log-rank test, and the results were presented in Figure 3. Remarkably, CXCL1-positive expression indicated a shorter overall survival time of patients (median OS: 39 Months) compared with CXCL1-negative expression (median OS: 64 Months). These results together suggested upregulated expression of CXCL1 in HCC was significantly correlated with patients' survival time.

To further address the biological role of CXCL1 in HCC, we investigated the effect of overexpression of CXCL1 on cell proliferation and invasion. The over-expressed cell lines were named as 7721-CXCL1 or 7402-CXCL1, while the matched control cell lines were named as 7721-vector or 7402- vector, respectively. As measured by western blotting analysis, two HCC cell lines 7721 and 7402 transfected with pCDNA3.1-CXCL1 had at least 4-fold higher levels of CXCL1 than the control (Figure 2A). MTT assay revealed that cell growth was significantly increased in CDNA3.1-CXCL1 transfected 7721 cells and 7402 cells (Figure 4B and C).Similarly, the migratory potential of 7721-CXCL1 and 7402-CXCL1 cells was significantly enhanced compared to 7721-vector or 7402- vector, respectively (p < 0.01, Figure 4D).

Cancer cell invasion is key step in metastasis that is a primary cause of cancer-related death. Since that CXCL1 were shown to be important for HCC cell proliferation and invasion, we sought to key signaling pathway responsible for CXCL1-induced cell invasion. It has been previously shown that CXCL1 increases cell migration and invasion through NF-κB/HDAC1 epigenetic regulation in prostate cancer [18], and it has also been previously established that nuclear factor-kappa B (NF-κB) positively medicated the expression of CXCL1 through binding to the CXCL1 promoter in melanoma cells [19]. Therefore, we sought to determine whether NF-κB was required for CXCL1-induced cell proliferation and invasion. Consistent with previous reports [18], we observed a marked increase in p-P65 protein levels in 7721-CXCL1 and 7402-CXCL1 cells (Figure 5A). To investigate whether NF-κB signaling pathway plays a crucial role in CXCL1-induced cell invasion in HCC cells, we blocked NF-κB pathway of 7721-CXCL1 and 7402-CXCL1 cells using NF-κB inhibitor, PDTC (Figure 5A). Our results demonstrated that the stimulatory effects of CXCL1 on cell invasion were significantly extenuated by specific inhibition of PDTC (Figure 5B), suggesting that CXCL1 activated NF-κB pathway play a critical role in CXCL1-enhanced cell invasion.