Two human HCC cell lines, SMMC7721 (a human HCC cell line with low metastatic potential, established by the Shanghai Institute of Cell Biology, Chinese Academy of Sciences, Shanghai, China) 18 and Huh7 (a well-differentiated and non-metastatic human HCC cell line, Japanese Cancer Research Resources Bank), were maintained in high-glucose Dulbecco's modified Eagle medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS), L-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. All cell lines were cultured at 37°C in a humidified incubator in 5% CO₂.

Total RNA was extracted with Trizol Reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol, and reversed transcribed with RevertAid™ first-strand cDNA synthesis kit (Fermentas, Burlington, ON). IL-17 and IL-17RA mRNA levels were determined by qPCR using SYBR Premix Ex Taq (TaKaRa, Dalian, China) and normalized with β-actin using the following primers: IL-17 forward, 5'-CGC TGA TGG GAA CGT GGA CTA C-3' and reverse, 5'-GGT GGA CAA TCG GGG TGA CA-3'. IL-17R forward, 5'-GTT CAT CAC GGG CAT CTC CAT C-3' and reverse, 5'-CAG GCA GGC CAT CGG TGT ATT-3'. β-actin forward, 5'-CAA CTG GGA CGA CAT GGA GAA AAT-3' and reverse, 5'-CCA GAG GCG TAC AGG GAT AGC AC-3'. The relative gene expression was calculated with the 2^-ΔCt method.

The effect of IL-17 on HCC cell growth was determined with the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazoliumbromide (MTT) assay. Cells were seeded into 96-well flat-bottom plates (1 × 10³/well) and cultured for 1 to 4 days in medium supplemented with recombinant human IL-17 (0, 0.1, 0.5, 1, 5, 10, 50, 100, or 500 ng/ml; 6 wells/dose), and each experiment was repeated at least three times. After IL-17 treatment, cells were incubated with MTT (20 μl/well) at 37°C for 4 h, and then 200 μl dimethylsulfoxide was added. The absorbance of individual wells was determined at 570 nm.

The wound-healing assay was used to evaluate the ability of cell migration. Cells were grown to 80-90% confluence in 24-well plates, and a wound was made by dragging a plastic pipette tip across the cell surface. The remaining cells were washed three times to remove cell debris, and then incubated at 37°C with serum-free medium. After 48 h, migrating cells at the wound front were photographed and compared. Three separate experiments were performed.

Cell invasion assays were performed using 24-well transwells (8-μm pore size; Minipore) precoated with Matrigel (BD Biosciences, San Jose, CA). In total, 1 × 10⁵ cells, which were suspended in 100 μl DMEM containing hIL-17 (50 ng/ml), 1% FBS, IL-6 mAb (10 ng/ml), and/or IL-6 (100 ng/ml), were added to the upper chamber, and 600 μl DMEM containing 10% FBS was placed in the lower chamber. After 36 h of incubation, Matrigel and the cells remaining in the upper chamber were removed using cotton swabs. Cells on the lower surface of the membrane were fixed in 4% paraformaldehyde and stained with Giemsa. Cells in five microscopic fields (at 200× magnification) were counted and photographed. All experiments were performed in triplicate.

Recombinant human IL-17 and IL-6, and an IL-6 neutralizing mAb (R&D Systems) were applied as appropriate.

For immunofluorescence staining, the monoclonal antibody (mAb) IL-17RA (R&D Systems, Minneapolis, MN) was used.

The levels of IL-6, IL-1β, TNF-α, TGF-β, IL-8, G-CSF, GM-CSF, VEGF, and MMP2 in culture supernatants were measured by ELISA, following the manufacturer's instructions (R&D Systems). Intra- and inter-assay coefficients of variation were < 5% and < 10%, respectively.

Western blotting was performed as previously described 19 . The specific primary antibodies p-JAK2 (Y1007/1008), JAK2, p-STAT3 (Y705), STAT3, p-p65 NF-κB (S536), p65 NF-κB, p-AKT (S473), AKT, p-JNK (T183/Y185), JNK, p-ERK1/2 (T202/Y204), ERK1/2, p-p38 MAPK (T180/Y182) and p38 MAPK (Cell Signaling Technology, Beverly, MA), as well as MMP2 (Abcam, Cambridge, MA) were used. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Millipore, Billerica, MA) was used as a loading control.

STAT3- and AKT-targeted siRNAs, as well as negative control mismatch sequences, were synthesized by Shanghai GeneChem Co. using the following sense and anti-sense strands: AKT sense, GUG CCA UGA UCU GUA UUU ATT and anti-sense, UAA AUA CAG AUC AUG GCA CTT; STAT3 sense, GGG ACC UGG UGU GAA UUA UTT and anti-sense, AUA AUU CAC ACC AGG UCC CTT. Transfection into SMMC7721 and Huh7 cells was performed using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions.

The pEGFP-N1-IL-17 plasmids were also constructed by Shanghai GeneChem Co. and pEGFP-N1 plasmids were used as controls. The lentiviral vector and plasmid were transfected into SMMC7721 cells as described elsewhere 20 . Stably transfected clones were validated by qRT-PCR and immunoblotting for the level of target gene expression (Additional file 1, Figure S1).

BALB/ca male nude mice (Shanghai Institute of Materia Medica, Chinese Academy of Science), 4 to 6 weeks old, were kept in laminar-flow cabinets under specific pathogen-free conditions, cared for, and handled according to the recommendations of the National Institutes of Health guidelines for care and use of laboratory animals. The experimental protocol was approved by Shanghai Medical Experimental Animal Care Committee. For the tumor challenge, 1 × 10⁶ SMMC7721-IL-17 or SMMC7721-mock tumor cells were injected subcutaneously into nude mice, and tumor growth was monitored each week 21 . There were five animals in each group. The mice were sacrificed on day 28 after tumor implantation. The levels of IL-6, IL-8, VEGF, and MMP2 in mice serum were measured by ELISA. Tumor tissues were prepared for Western blot analysis and immunostaining. Anti-mouse CD31 and Gr-1 mAbs (Abcam) were used for immunostaining.

Tumor samples were obtained from 410 patients with pathologically confirmed HCC at Liver Cancer Institute of Zhongshan Hospital, Fudan University. Serial sections from 87 patients (cohort 1) treated between December 2010 and January 2011 were used for immunohistochemistry, among which 48, 21, and 18 patients were classified as TNM stage I, II, and III, respectively 22 . For tissue microarray (TMA) construction, archived tissues were obtained from 323 consecutive HCC patients treated between January 2003 and March 2004, with a median follow-up of 60 months (range, 2.0-74.0 months; cohort 2). All patients demonstrated no distant metastasis and had not received anticancer therapy before surgery. The tumor stage was determined according to the 2010 American Joint Committee on Cancer and International Union Against Cancer tumor-node-metastasis (TNM) classification system. Tumor differentiation was graded by the Edmondson grading system. The follow-up procedure was performed as described in our previous report 23 . Overall survival (OS) or time to recurrence was defined as the intervals between the dates of surgery and death or the first recurrence, respectively. Patients without death or recurrence were censored at the last follow up. Detailed clinicopathological characteristics are given in Table S1. Ethical approval was obtained from the research ethics committee of Zhongshan Hospital and written informed consent was obtained from each patient.

To examine the relationship between the densities of intratumoral IL-17+ cells and CD66b+ neutrophils or microvessel densities (MVDs), 87 cases (cohort 1) were chosen for immunohistochemistry using serial whole tumor sections.

TMAs were constructed as described elsewhere 19 23 using tissues form cohort 2 (n = 323). Immunohistochemistry was conducted with the following primary mAbs: goat antihuman IL-17A (R&D Systems), mouse antihuman CD66b (BD Pharmingen, San Diego, CA) and CD34 (Abcam), and rabbit antihuman p-STAT3 (Y705) (Cell Signaling Technology). Blank controls were treated identically with primary antibodies omitted.

The capture of the photographs and measurement of positive staining density were performed as described previously 23 . Briefly, for IL-17, p-STAT3, and CD66b staining in consecutive sections from 87 HCC patients (cohort 1) and Gr-1 staining in xenografts, the sections were stained with indicated antibodies, and then were evaluated using light microscopy at 200× or 400× magnification by two investigators. Five representative fields of each case were captured. IL-17+ cell density or CD66b+ and Gr-1+ neutrophil density was evaluated according to the mean number of positive-staining cells in five count areas. For expression intensity of p-STAT3, the integrated absorbance and the area in a photograph were measured using Image-Pro Plus v6.0 software (Media Cybernetics, Inc.). A uniform setting of color segmentation was loaded for counting the integrated absorbance of all the pictures, and the mean p-STAT3 density was calculated as the product of the integrated absorbance/total area.

A Ki-67-specific monoclonal rat antibody (DAKO A/S) was used for detection and quantitation of tumor cell proliferation, while the In Situ Cell Death Detection kit-Peroxidase (Roche) was used for detection and quantitation of apoptosis. The Proliferation Index and Apoptosis Index were determined by calculating the number of Ki-67- or TUNEL-positive cells per total number of cells (hematoxylin-positive plus Ki-67- or TUNEL-positive cells) in 5 randomly selected fields at ×200, using Image-Pro Plus v6.0 software as previously described 19 .

For the MVD, sections were stained with CD31 or CD34 antibody and a diaminobenzidine reaction system for immunohistochemical assessment of tumor microvessels in five randomly selected fields at 200× magnification. The average number of microvessels was calculated. For the microvessel count, any brown-stained endothelial cell or endothelial cell cluster that was clearly separated from adjacent microvessels, tumor cells, and connective elements was counted as one microvessel, irrespective of the presence of a vessel lumen 4 .

IL-17 and p-STAT3 staining in TMAs were evaluated at 200× magnification using light microscopy by two investigators blinded to the clinicopathologic data of the patients. For IL-17 immunostaining, the number of positive staining cells in each 1-mm diameter cylinder were calculated manually and expressed as the mean number of the duplicates (cells/1-mm core). For the expression intensity of p-STAT3, the integrated absorbance and the area in each 1-mm diameter cylinder were measured using Image-Pro Plus v6.0 software. The mean p-STAT3 density was calculated as the product of the integrated absorbance/total area.

Statistical analysis was performed with SPSS 16.0 software (SPSS, Chicago, IL). Measurement values were expressed as means ± standard deviations. Student's t test, χ² test, and Spearman's ρ correlation were used as appropriate. The cumulative recurrence and survival rates were performed by the Kaplan-Meier method (log-rank test). Cox multivariate analysis with a stepwise method (forward, likelihood ratio) was used to determine the independent prognostic factors. Two-tailed p < 0.05 was judged to be significant.

The median value of IL-17+ cell counts was used as the cutoff to dichotomize IL-17 immunostaining. The optimal cutoff for dichotomizing p-STAT3 (Y705) expression data was determined using X-tile 3.6.1 software (Yale University of New Haven) 24 .

The expression of IL-17RA was easily detectable by qRT-PCR and immunofluorescence in Huh7 and SMMC7721 cells (Figure 1A). Because IL-17+ T cells were significantly elevated in HCC patients and correlated with poor survival 14 , we thus assumed that IL-17 could stimulate HCC cells through IL-17RA signaling pathways. Although IL-17 exerted little influence on cell proliferation (Additional file 2, Figure S2A), wound-healing assays revealed an evident increase in the wound closure rates of Huh7 and SMMC7721 cells after IL-17 (50 ng/ml) stimulation for 48 h (Figure 1B), and matrigel invasion assays showed that exogenous IL-17 (50 ng/ml for 36 h) significantly promoted the invasion of Huh7 and SMMC7721 cells (Figure 1C). These results indicated that IL-17 can promote invasion and migration of HCC in vitro.

IL-17 is known for eliciting secretion of diverse inflammatory mediators in diverse cell types, including stromal cells and tumor cells. We found that IL-17 selectively up-regulated the production of IL-6, IL-8, MMP2, and VEGF in Huh7 and SMMC7721 cells (Table 1). The most prominent secretion was IL-6, with 5.8- and 6.0- fold increases, serially followed by IL-8 (2.5- and 5.6- fold), MMP2 (1.5- and 1.6- fold) and VEGF (1.2- and 1.7- fold) in Huh7 and SMMC7721 cells respectively. By contrast, in Huh7 and SMMC7721 cells, the production of IL-1β, TNF-α, TGF-β, G-CSF and GM-CSF were not significantly affected by exogenous IL-17.

Various signaling pathways are suggested to mediate IL-17 action. Here, we scrutinized the potential signaling pathways of IL-17 action in HCC cells and found that IL-17 had no effect on p38 MAPK, ERK, JNK and p65 NF-κB activation in Huh7 and SMMC7721 cells (Additional file 2, Figure S2B). Intriguingly, phosphorylation of STAT3 and AKT were obviously increased as early as 3 h after IL-17 treatment and lasted for 24 h (Figure 2A, and Additional file 3, Figure S3A). Given that expression of IL-6, IL-8, MMP2, and VEGF increased in parallel with STAT3 and AKT activation in HCC cells treated with IL-17, STAT3, and AKT signaling may play important roles in inducing these proinvasive factors and hence tumor progression.

To explore the potential role of STAT3 and AKT in IL-17-mediated effects on HCC, STAT3 and AKT expression were reduced by small interfering RNA (siRNA).

First, in HCC cells exposed to STAT3-targeted siRNA (HCC-siRNA-STAT3), IL-17-induced AKT phosphorylation was not affected, while in HCC cells exposed to AKT-targeted siRNA (HCC-siRNA-AKT), IL-17-induced STAT3 phosphorylation was significantly reduced (Figure 2B and 2C, and Additional file 3, Figure S3B and S3C). Furthermore, IL-17-induced AKT phosphorylation was obviously increased as early as 5 minutes after IL-17 treatment, while phosphorylation of JAK2 and STAT3 were not affected during the 60 minutes treatment (Figure 2D). These results suggested that AKT phosphorylation was an earlier event as compared with STAT3 phosphorylation in HCC stimulated with IL-17.

Additionally, in HCC-siRNA-STAT3 cells, IL-17-induced expression of IL-8, MMP2, and VEGF were significantly inhibited, while IL-17-induced IL-6 upregulation was not affected (Figure 2B and 2E, and Additional file 3, Figure S3B and S3D). These results indicated that IL-8, MMP2, and VEGF were target genes instead of the upstream activators of STAT3. Furthermore, neutralizing IL-6 with a blocking mAb reduced JAK2 and STAT3 activation and inhibited upregulation of IL-8, MMP2, and VEGF by IL-17, whereas AKT activation by IL-17 was not affected (Figure 2E and 2F, and Additional file 3, Figure S3D and S3E). Thus, IL-6 may be the main upstream activator, rather than the downstream target of JAK2/STAT3, in the setting of IL-17 stimulation.

By contrast, in HCC-siRNA-AKT cells, but not HCC-siRNA-STAT3 cells, IL-17-induced IL-6 upregulation was significantly blocked (Figure 2E and 3A, and Additional file 3, Figure S3D and Additional file 4, S4A). Thus, AKT siRNA both repressed IL-6 and downregulated JAK2/STAT3 phosphorylation induced by IL-17 (Figure 2C, 3A and 3B, and Additional file 3, Figure S3C, Additional file 4, S4A and S4B), indicating that IL-6/JAK2/STAT3 may be the targets of AKT in the context of IL-17. Thus, IL-17 may upregulate IL-6 production via AKT activation. IL-6 in turn activates JAK2/STAT3 and subsequently stimulates IL-8, MMP2, and VEGF production. Supporting this hypothesis, JAK2/STAT3 phosphorylation and the expression of IL-8, MMP2, and VEGF were significantly increased after addition of IL-17 plus exogenous IL-6 in siRNA-AKT-SMMC7721 and siRNA-AKT-Huh7 cells (Figure 3A and 3C, and Additional file 4, Figure S4A and S4C).

Meanwhile, STAT3-siRNA significantly reversed tumor invasion by IL-17 stimulation in HCC (Figure 3D, and Additional file 4, Figure S4D). An IL-6 neutralizing mAb reduced STAT3 activation and also completely reversed IL-17-stimulated tumor invasion in vitro (Figure 3D, and Additional file 4, Figure S4D), while exogenous IL-6 completely recovered IL-17-induced invasion of siRNA-AKT-SMMC7721 and siRNA-AKT-Huh7 cells (Figure 3D, and Additional file 4, Figure S4D). AKT-dependent IL-6/STAT3 activation was therefore suggested to be responsible for the tumor promoting effects of IL-17 on HCC.

When tumor cells were injected into nude mice, the growth rate of SMMC7721-IL-17 tumors was drastically increased relative to SMMC7721-mock controls [mean tumor volume (mock vs IL-17), 0.589 ± 0.259 vs 4.175 ± 3.050 cm³, p = 0.031; mean tumor weight (mock vs IL-17), 0.392 ± 0.124 vs 2.140 ± 0.963 g, p = 0.004; respectively; Figure 4A]. Moreover, SMMC7721-IL-17-derived xenografts showed increased tumor cell proliferation while reduced apoptosis of HCC cells compared with the SMMC7721-mock group [proliferation index (mock vs IL-17), 71.260 ± 9.584 vs 83.820 ± 6.549, p = 0.042; apoptosis index (mock vs IL-17), 1.070 ± 0.222 vs 0.626 ± 0.320, p = 0.034; respectively; Additional file 5, Figure S5].

Increased phosphorylation of AKT was evident and the plasma concentration of IL-6 was significantly up-regulated in IL-17-transfected SMMC7721 mice compared with those in controls (Figure 4B and 4C). Overexpression of IL-17 also promoted JAK2/STAT3 activation and markedly up-regulated production of IL-8, MMP2, and VEGF (Figure 4B and 4C).

Apart from direct effects on tumor cells, MMP2 and VEGF are known for fostering angiogenesis and IL-8 is a major chemotactic factor for neutrophil recruitment. Supporting this, SMMC7721-IL-17-derived xenografts presented with more neoangiogenesis (CD31+) and Gr-1+ neutrophil infiltration than the SMMC7721-mock group (p < 0.001 for both; Figure 4D).

Because in vivo SMMC7721-IL-17 tumors were characterized by increased angiogenesis and enriched neutrophil infiltration, we further explored whether a similar phenomenon exists in human HCC tissues. By staining consecutive sections in 87 HCC patients (Figure 5A), a correlation analysis revealed that significant positive correlations were found between the density of intratumoral IL-17-producing cells and MVD (r = 0.567; p < 0.001), as well as between the densities of intratumoral CD66b+ neutrophils and IL-17+ cells (r = 0.630; p < 0.001; Figure 5B). Additionally, the p-STAT3 intensity significantly and positively correlated with levels of intratumoral IL-17+ cells, CD66b+ neutrophils, and MVD (r = 0.324 and p = 0.002; r = 0.350 and p < 0.001; r = 0.544 and p < 0.001; respectively; Additional file 6, Figure S6). Thus, IL-17+ cells may promote neoangiogenesis and neutrophil recruitment partly via STAT3 activation in human HCC.

We further investigated the prognostic value of IL-17 and its target, p-STAT3, both separately and in combination, in TMA containing 323 HCC patients (Additional file 7, Figure S7). Similarly, the density of intratumoral IL-17+ cells significantly correlated with p-STAT3 expression in this cohort (r = 0.315; p < 0.001; Figure 5C). Both IL-17^high and p-STAT3^high were significantly correlated with microvascular invasion (p = 0.016 and p = 0.013, respectively; Additional file 8, Table S1).

Among the 323 patients, the 5-year OS and cumulative recurrence rates were 50.7% and 59.7%, respectively. Univariate analysis revealed that intratumoral IL-17+ cells and p-STAT3 expression were significantly associated with recurrence and survival (Table 2). Furthermore, patients were classified into three groups: group I, high expression of both markers; group II, high expression of either markers; and group III, low expression of both markers. Significant differences in recurrence and survival were detected among the three groups (Figure 5D).

Significant clinicopathologic features identified in univariate analysis were adopted as covariates in multivariate Cox models. Multivariate analyses showed that IL-17+ cell and p-STAT3 expression independently correlated with OS and recurrence, irrespective of being used alone or in combination (Additional file 9, Table S2). However, concomitant high of IL-17 and p-STAT3 was superior to either marker alone in terms of hazard ratios and p values for both OS and recurrence (Additional file 9, Table S2). In multivariate analysis where IL-17+ cells, p-STAT3 expression and their combination were simultaneously adopted as covariates (Table 2), only their combination remained statistically significant. Specifically, compared with group III, hazard ratios of group I were 2.244 (95% CI, 1.529-3.293; p < 0.001) for death and 1.718 (95% CI, 1.193-2.473; p = 0.004) for recurrence (Table 2), respectively.