Tissues were obtained from 137 patients who underwent surgery at the Southwest Hospital (Chongqing, China) between 2005 and 2011. Patients who underwent pre-operative treatment, such as radiotherapy and/or chemotherapy, were excluded. A total of 110 cancerous and 28 lymph node metastatic samples from the patients were collected. A total of 19 chronic inflammatory bile duct samples from patients with hepatolithiasis and 8 biliary adenoma samples were also collected from Southwest Hospital. All samples were obtained with informed consent from all patients, according to the protocols approved by the Institutional Review Board of the Southwest Hospital, Third Military Medical University. The pathological reports were reviewed and the clinical features of the 110 patients with CCAs are presented in Table I. Tumor-node-metastasis stages were assigned according to the 6th Union for International Cancer Control. Overall survival was calculated from the date of surgery to the date of mortality or last contact. Recurrence-free survival was computed from the date of surgery to the date of recurrence.

Formalin-fixed, paraffin-embedded resected tissue blocks were cut into 4-mm sections and mounted on charged glass slides, deparaffinized and rehydrated in a graded series of ethanol. Endogenous peroxidase activity was blocked with a solution of 3% H₂O₂ in methanol for 30 min. Following washing in phosphate-buffered saline (PBS), antigen retrieval was performed in a citrate buffer (pH 6.0) at 120°C for 15 min. After cooling and washing 3 times with PBS (pH 7.4) for 5 min each, sections were incubated with the primary antibodies against B7-H4 (LLC.250473; dilution 1:200; Abbiotec, San Diego, CA, USA) (9), CD4 (TA802240S; dilution 1:150), CD8 (TA802079S; dilution 1:200) (both from OriGene Technologies, Inc., Rockville, MD, USA) in a humid chamber at 4°C overnight. The sections were then washed with PBS and incubated with a secondary polymeric peroxidase-labeled rabbit anti-mouse antibody (Dako, Glostrup, Denmark) for 1 h at 37°C. Subsequently, the nuclei were counterstained with hematoxylin and visualized with 3,3-diaminobenzidine tetrahydrochloride (Dako). A negative control was performed using PBS instead of the primary antibodies under the same conditions. The sections were dehydrated, cleared and mounted.

Two independent investigators (G.F. and J.P.), who were blinded to the patient clinicopathological data, analyzed the IHC images. Expression of B7-H4 was analyzed in 10 different high-power fields (HPFs). IHC for B7-H4 showed cytoplasmic and membrane staining. The intensity (I) of staining was scored as negative (0), weak (1), moderate (2), or strong (3). The proportion (P) of B7-H4-positive cells was defined as: 0, no staining; 1, <10% tumor cells with staining; 2, 10–50% tumor cells with staining; 3, 50–80% tumor cells with staining; 4, >80% tumor cells with staining. Samples with IHC scores (P × I) ≤3 were considered as negative and with scores >3 were defined as positive.

Additionally, the expression levels of CD4/CD8-positive T lymphocytes in the tumor nest and tumor stroma were separately determined according to IHC staining. The evaluation was based on a review of 10 different HPFs that showed the highest level of lymphocytic infiltrates for each case (low <10%, high ≥10%).

Two human CCA cell lines (QBC939 and RBE), were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) (10,11), and cultured in RPMI-1640 medium with 10% fetal bovine serum (FBS; GE Healthcare Life Sciences, Chalfont, UK). All cell lines were incubated at 37°C in a humidified atmosphere containing 5% CO₂.

Cells were transfected with lentiviral vectors encoding short hairpin RNA (Shanghai GenePharma Co., Ltd., Shanghai, China) targeting human B7-H4 for B7-H4 knockdown or a scrambled shRNA-NC (Shanghai GenePharma Co., Ltd.) as the control. The QBC939 and RBE cell lines were co-transfected with lentiviral vectors. The recombinant vector was named GFP&Puro-B7-H4-shRNA, according to the manufacturer's recommendations. For the QBC939 and RBE cell lines, a multiplicity of infection of 10 was used to achieve >90% transfection. The cells were cultured for 72 h following transfection. Stably transfected QBC939 and RBE cells were selected using puromycin. The efficiency of knockdown was detected using western blotting.

Seventy-two hours after co-transfection with the lentiviral vectors, cell extracts were prepared on ice using RIPA lysis buffer plus a complete protease inhibitor cocktail (both from Beyotime, China). Protein quantitation was measured by a BCA protein assay kit (Thermo Scientific, Rockford, IL, USA) according to the manufacturer's instructions. Ten micrograms of protein/lane were separated on 8% acrylamide gels by sodium dodecyl sulfate (SDS) gel electrophoresis and transferred to polyvinylidene fluoride (PVDF) membranes. After blockage of non-specific binding sites using 5% skim milk with Tris-buffered saline with Tween-20 (TBST) at 4°C overnight, the membranes were incubated for 2 h at 37°C with goat polyclonal affinity purified anti-human B7-H4 antibody (ab130151; dilution 1:1,000; Abcam, Cambridge, MA, USA). A GAPDH antibody (10494-1-AP; dilution 1:1,000; Proteintech, China) was used as an internal control. Following primary antibody incubation, the membranes were washed in TBST and incubated with horseradish peroxidase (HRP)-conjugated appropriate secondary antibodies (Dako) for 1 h at 37°C. The membranes were washed with TBST followed by visualization using enhanced chemiluminescence (Millipore, Billerica, MA, USA) according to the manufacturer's protocol.

Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Histopaque-1077 (Sigma-Aldrich, Munich, Germany) from 11 healthy donors. PBMCs were seeded into 6-well culture plates containing 2 ml RPMI-1640 medium and 10% FBS at a final concentration of 5–10×10⁶ cells/well. Following 2 h of incubation, non-adherent cells were removed by gentle washing with warm medium. The non-adherent cells (effector lymphocytes) were cryopreserved in FBS supplemented with 10% dimethyl sulfoxide. The resultant adherent cells containing dendritic cells (DCs) were cultured in medium supplemented with 500 U/ml recombinant human granulocyte-macrophage colony-stimulating factor (GM-CSF) and 1,000 U/ml recombinant human interleukin 4 (IL-4) (both from PreproTech, Inc., Rocky Hill, NJ, USA) at 37°C in 5% CO₂ (12). Every 2 days, one-half of the medium was replaced with fresh medium containing a double concentration of GM-CSF and IL-4 as indicated above. Following 5 days in culture, 10 ng/ml of recombinant human tumor necrosis factor-α (TNF-α; PreproTech, Inc.) was added to the medium to induce phenotypic and functional maturation of DCs (12). CCA cells were induced to apoptotic tumor cells (ATCs) with 100 µg/ml mitomycin for 24 h (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China), and were presented by DCs to induce specific CTLs in vitro. Subsequently, the isolated non-adherent effector lymphocytes were co-cultured with the ATC-pulsed autologous DCs in a 6-well plate in the presence of 10 ng/ml recombinant human interleukin-7 (IL-7; PreproTech, Inc.). Half the medium was replaced with complete medium supplemented with 30 IU/ml recombinant human interleukin 2 (IL-2; PreproTech, Inc.) every 3 days. Following 7 days in culture, the lymphocytes were re-stimulated with the ATC-pulsed autologous DCs in medium containing 10 ng/ml IL-7 and 20 U/ml IL-2. On day 10, following the fourth round of re-stimulation, the cells were harvested and tested using CCA-specific CTL assay (13). CD8⁺ T-mediated CTLs were purified by negative depletion using a CD8⁺ T cell isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany).

CTL activity was evaluated using the CytoTox 96^® non-radioactive cytotoxicity assay (Promega, Madison, WI, USA) based on lactate dehydrogenase (LDH) release. After washing, the target cells were counted and seeded into 96-well V-bottomed culture plates. Varying numbers of CTLs were added to a final volume of 100 µl at the effector to target (E/T) ratios of 2.5:1, 5:1, 10:1 and 20:1 and incubated for 4 h at 37°C. The supernatants were harvested and the assay plates were incubated for 30 min at room temperature, protected from light. The absorbance at 490 nm was recorded within 1 h after adding the stop solution. The corrected values were used in the following formula to compute percent cytotoxicity: Cytotoxicity = [(Experimental - Effector Spontaneous - Target Spontaneous)/(Target Maximum - Target Spontaneous)] × 100%.

Data were analyzed using SPSS software, version 19.0 (IBM SPSS, Armonk, NY, USA). Group comparisons of continuous data were made using t-test on independent means. For categorical data, Chi-square analysis or the Fisher's exact test was used. The Kaplan-Meier estimator and Cox analysis were used for overall survival and recurrence-free survival. P<0.05 was considered to indicate a statistically significant difference.

Representative IHC images of B7-H4 are presented in Fig. 1. The expression of B7-H4 protein was detected in 54/110 (49.1%) cancerous tissues, 15/28 (53.6%) lymph node metastatic tissues and 4/19 (21.1%) chronic inflammatory bile duct tissue samples (Table II). Notably, in the inflammatory bile duct tissues, B7-H4 was predominantly expressed in the infiltrating mononuclear cells rather than the epithelial cells of the bile duct, which is in line with its role during inflammatory reactions. In addition, 8 biliary adenoma samples (Fig. 1) stained negative for B7-H4. As shown in Table II, positive staining of B7-H4 was detected in cancerous and lymph node metastatic samples, which was significantly greater compared with the non-tumorous tissues. These data indicated that the high expression of B7-H4 was specific to the CCA tissues.

As B7-H4 was highly expressed in the cancer tissues of patients with CCA, its expression was investigated as to whether it correlates with clinicopathological parameters in patients with CCA. The clinicopathological features of the 110 cases of CCA were grouped by positive or negative B7-H4 expression. As shown in Table I, B7-H4 expression in CCA tissues was significantly associated with tumor status (P=0.025), lymph node metastasis (P=0.035) and UICC stage (P=0.019), however, not with gender, age, tumor location and size. The data indicate that the cases of CCA positive for B7-H4 exhibited more extensive metastatic behavior.

The results indicated that the expression of B7-H4 correlates with adverse pathological features, which are associated with patient prognosis. First, the correlation of B7-H4 expression with disease prognosis was analyzed. The results indicated that the expression of B7-H4 is associated with a poorer outcome following surgery. The median overall survival time of patients negative for B7-H4 was 19.5 months, which is longer than the 12.6 months observed in patients positive for the expression of B7-H4 (P=0.015; Table III). In addition, the median disease-free survival time of patients negative for B7-H4 expression was longer than that of patients with positive B7-H4 expression (16.7 vs. 10.9 months; P=0.046; Table III). Kaplan-Meier analysis indicated that the expression of B7-H4 is associated with reduced survival (log-rank P=0.015; Fig. 2A) and time to recurrence (log-rank P=0.046; Fig. 2B) in the 110 cases. To further validate these observations, multivariate Cox analysis was used, which showed that B7-H4 expression was an independent indicator of poorer overall survival and early recurrence [hazard ratio (HR)=1.786; 95% confidence interval (CI), 1.110–2.872; P=0.017; and HR=2.062; 95% CI, 1.160–3.665; P=0.014; Table V]. Taken together, the data suggest that aberrant expression of B7-H4 may be a risk factor for poorer prognosis in patients with CCA following surgical resection.

Previous studies have shown that B7-H4 is able to directly or indirectly modulate immune infiltrate cells, which are comprised predominantly of CD4⁺ helper T cells and CD8⁺ cytotoxic T cells (8,14–16). The presence of tumor-infiltrating lymphocytes (TILs) within the tumor stroma or nest is considered an indicator of the host immune response to the tumor (17). Therefore, whether the expression of B7-H4 correlates with CD4⁺ and CD8⁺ TILs was investigated in the 110 CCA tissues. As shown in Table IV, levels of B7-H4 expression in the tumor cells was inversely correlated with the density of CD8⁺ T cells in the tumor stroma (P=0.0004), however, was not correlated with the density of CD8⁺ T cells in the tumor nest (P= 0.776). In addition, there was no significant association between B7-H4 expression and the CD4⁺ T cells in the tumor stroma or nest (P=0.567 and P=0.822, respectively). Therefore, these data provide further evidence of the potential role of B7-H4 in the suppression of cellular immune surveillance in patients with CAA, and in particular tumor infiltrating CD8⁺ T cells.

Considering the above observations, which indicated a negative correlation between B7-H4 expression and the density of CD8⁺ T cells in the tumor stroma, the impact of B7-H4 on CD8⁺ T cells was further investigated in vitro. Western blot analysis showed that B7-H4 was expressed in QBC939 and RBE cells (Fig. 3A). Notably, the protein expression levels of B7-H4 were significantly greater in QBC939 cells compared with RBE cells (Fig. 3A). Knockdown of B7-H4 was performed in QBC939 and RBE cells using B7-H4-shRNA lentiviral transfection (Fig. 3B and C). By culturing with CD8⁺ cytotoxic T cells, the cytotoxicity of CD8⁺ T cells was markedly improved by the knockdown of B7-H4 in QBC939 and RBE cells (Fig. 3E and F). These data indicate that knockdown of B7-H4 in tumors increases CD8⁺ T cell-mediated cytotoxicity in vitro.