A total of 53 patients with HCC who underwent hepatectomy without preoperative therapy at the Nippon Medical School Hospital (Tokyo, Japan) between January 2016 and February 2018, were enrolled in this study. The tumor and adjacent normal tissues from these patients were fixed with 10% formalin at room temperature for 24 h within 2 h of resection and embedded in paraffin. Among them, 35 HCC samples were formalin-fixed within 30 min of resection and were used for immunostaining of P-STAT3. The baseline characteristics of the patients are summarized in Table I. This study was conducted according to the Declaration of Helsinki and the Japanese Society of Pathology, and was given official approval by the Ethics Committee of the Nippon Medical School Hospital (approval no. 29-03-908). Written informed consent was obtained from all patients.

Human hepatoma cell lines (Huh-7 and HuH-1) were obtained from the Japanese Collection of Research Bioresources cell bank (Osaka, Japan). A normal human hepatocyte cell line (THLE-2) was obtained from the American Type Culture Collection (Manassas, VA, USA). Huh-7 and HuH-1 cells were cultured in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Nichirei Biosciences, Inc., Tokyo, Japan) at 37°C in a humidified 5% CO₂ atmosphere.

The THLE-2 cell line was maintained in Bronchial Epithelial Cell Growth Medium (BEGM; Lonza Group Ltd., Basel, Switzerland) without gentamycin/amphotericin and epinephrine but with added 5 ng/ml epidermal growth factor (Corning, Inc., Corning, NY, USA), 70 ng/ml phosphoethanolamine (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) and 10% FBS. The THLE-2 cells were cultured in RPMI-1640 medium with 10% FBS for 7 days at 37°C in a humidified 5% CO₂ atmosphere prior to protein extraction. A JAK/STAT3 signaling inhibitor, AG490 (Merck KGaA), dissolved in dimethylsulfoxide (DMSO; Wako Pure Chemical Industries, Ltd., Osaka, Japan), was used for the suppression of STAT3 signaling at 25 or 100 µM in cell growth or western blotting, respectively.

Short-interfering RNAs (siRNAs) were purchased from Thermo Fisher Scientific, Inc. The two PDIA3 siRNAs used were designated PDIA3 si-1 and PDIA3 si-2 and known as Silencer^® Select Pre-designed siRNA cat. no. 4392420, ID s6227 (5′-GGAAUAGUCCCAUUAGCAAtt-3′) and ID s6229 (5′-GCAACUUGAGGGAUAACUAtt-3′), respectively. Silencer^® negative control #1 siRNA (cat. no. 4390844) was used as a negative control (Ctrl si). The transfection of the siRNA was performed using Lipofectamine^® RNAiMAX Reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. The optimal concentration of siRNA for use in transfections was 5 nM. The medium was refreshed 24 h after cell seeding and the cells were cultured for a further 48 and 72 h for the apoptosis and cell cycle assay and western blotting, respectively. The cell transfection efficiency was measured by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and western blotting, as described below.

A total of 2.5×10⁵ cells were seeded in 60-mm dishes and cultured for 48 h. Total RNA was extracted using the NucleoSpin RNA kit (Takara Bio Inc., Otsu, Japan), and 1 µg of total RNA was used for reverse transcription using the SuperScript VILO cDNA Synthesis kit (Thermo Fisher Scientific, Inc.) following the manufacturer's protocol. The reaction conditions were as follows: 25°C for 10 min, 42°C for 60 min and 85°C for 5 min. The qPCR was performed for PDIA3 and 18S rRNA (as an internal standard) using the StepOnePlus Real-Time PCR system with TaqMan probes and primers (18S, cat. no. Hs 03928990_g1; PDIA3, cat. no. Hs 04194196_g1) (all Thermo Fisher Scientific, Inc.). The cycling conditions were as follows: 20 sec at 95°C, followed by 40 cycles of 1 sec at 95°C and 20 sec at 60°C. The RT-qPCR results are expressed as the ratio of target mRNA to 18S rRNA and calculated using the 2^-ΔΔCq method (21). The gene expression levels were measured in triplicate.

Total protein was extracted using urea/thiourea buffer containing 7 M urea (Wako Pure Chemical Industries, Ltd.), 2 M thiourea (Nacalai Tesque, Inc., Kyoto, Japan), 3% 3-(3-[cholamidopropyl]-di methylammonio)-1-propanesulfonate (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) and 1% Triton X-100 (Sigma-Aldrich; Merck KGaA) from the cells after 96 h of siRNA transfection, as described previously (22). The protein concentration was determined using Pierce^® 660 nm Protein Assay Reagent (Thermo Fisher Scientific, Inc.). A total of 10 µg protein from each cell extract was loaded onto and separated by 5-20% SDS-PAGE (e-PAGEL^®; ATTO Corporation, Tokyo, Japan) and then transferred to a polyvinylidene difluoride membrane in a Trans-Blot^® Turbo™ Transfer Pack using a Trans-Blot Turbo transfer system (both Bio-Rad Laboratories, Inc., Hercules, CA, USA). The membranes were blocked for 1 h with 5% skimmed milk in TBS containing 0.2 M Tris-HCl, 150 mM NaCl and 0.01% Tween-20, and then incubated with specific primary antibodies overnight at 4°C. The antibodies to the following proteins were used: PDIA3 (cat. no. ab13506; 1:2,000), induced myeloid leukemia cell differentiation protein Mcl-1 (cat. no. ab32087; 1:2,000), 78 kDa glucose-regulated protein (GRP78; cat no. ab21685; 1:2,000) (all Abcam, Cambridge, UK), STAT3 (cat. no. 9139; 1:2,000), P-STAT3 (Tyr705; cat. no. 9145; 1:2,000), survivin (cat. no. 2808; 1:2,000), X-linked inhibitor of apoptosis protein (XIAP; cat. no. 14334; 1:2,000), Bcl-2-like protein 1 (Bcl-XL; cat. no. 2764; 1:2,000), cyclin D1 (cat. no. 2978; 1:1,000) (all Cell Signaling Technology, Inc., Danvers, MA, USA), B-cell lymphoma 2 (Bcl-2; cat. no. sc-492; 1:500; Santa Cruz Biotechnology, Inc., Dallas, TX, USA), cellular tumor antigen p53 (cat. no. M7001; 1:1,000; Dako; Agilent Technologies, Inc., Santa Clara, CA, USA) and β-actin (cat. no. A5316; 1:10,000; Sigma-Aldrich; Merck KGaA). Following washing in TBS with 0.01% Triton X-100 for 30 min, the membranes were incubated with a horseradish peroxidase-conjugated secondary antibody (cat. no. A106PU; 1:10,000; American Qualex Scientific Products, San Clemente, CA, USA) for 1 h at room temperature. The immunoreactive products were visualized using SuperSignal West Dura Extended Duration Substrate for each protein, and SuperSignal West Pico Chemiluminescence substrate (both Thermo Fisher Scientific, Inc.) for β-actin. The data were quantified by Quantity One 1-D analysis software version 4.6.2 (Bio-Rad Laboratories, Inc.). The experiments were performed ≥3 times.

The total protein was extracted from cells 96 h after seeding. The cells were lysed in Pierce™ IP Lysis Buffer (Thermo Fisher Scientific, Inc.) and protease inhibitor (P8340; Sigma-Aldrich; Merck KGaA) for 10 min on ice after washing in PBS. The cell lysates were centrifuged at 20,000 × g for 10 min at 4°C. The resulting supernatant was collected as a cell extract. Cell lysate (2 mg), antibody (3 µg), and protein A/G PLUS-Agarose (30 µl) were combined and incubated overnight at 4°C. The mixture was purified using Sigma Prep Spin Columns with Break-Away Tips (Sigma-Aldrich; Merck KGaA) and the proteins collected were used for immunoblotting with antibodies against PDIA3 and STAT3. Normal mouse IgG (3 µg; cat. no. SC-2025; Santa Cruz Biotechnology, Inc.) was used as a negative control for the immunoprecipitation experiments.

Cells were seeded in 96-well plates, at a density of 5,000 cells per well, and cultured at 37°C in a humidified 5% CO₂ atmosphere following siRNA transfection with PDIA3 si-1, PDIA3 si-2 or Ctrl si, as described above. After 0, 24, 48, 72 and 96 h of proliferation, cells were incubated with the WST-8 cell-counting reagent (Dojindo Molecular Technologies, Inc.) for 2 h at 37°C. The optical density of the culture solution in each well was determined at 450 nm using a microplate absorbance reader (iMark™; Bio-Rad Laboratories, Inc.). The experiments were performed ≥3 times.

Cells (5×10⁵) were seeded in a 75-ml flask and cultured in 10 ml RPMI-1640 medium. The medium was changed 24 h after PDIA3 siRNA treatment. Floating and attached cells were collected 72 h after seeding. Double staining with Annexin V and ethidium homodimer III was performed using the Apoptotic, Necrotic, and Healthy Cells Quantification kit (Biotium, Inc., Freemont, CA, USA), following the manufacturer's protocol. The apoptotic cells were counted by flow cytometry on a BD FACSCanto™ II flow cytometer and analyzed using BD FACSDiva™ software version 6.1.3 (both Becton, Dickinson and Company, Franklin Lakes, NJ, USA). The experiments were performed ≥3 times.

Cell harvesting was performed in the same manner as for the apoptosis analysis. The collected cells were washed in PBS and fixed in 100% methanol for 10 min. The cells were incubated with RNase (0.25 mg/ml) for 30 min at 37°C. Propidium iodide (50 µg/ml) was then added and incubated for 30 min at 4°C in the dark. The stained cells were analyzed by flow cytometry on a BD FACSCanto II flow cytometer and analyzed using BD FACSDiva software version 6.1.3. The experiments were performed ≥3 times.

Tissue sections (4 µm thickness) were used for hematoxylin eosin staining. Following deparaffinization, the sections were stained hematoxylin (Merck KGaA) for 30 min at room temperature, and eosin (Wako Pure Chemical Industries, Ltd.) for 20 min at room temperature, following washing in water. The sections were then washed and dehydrated. These sections were evaluated manually under a light microscope (×100 magnification; Olympus Corporation, Tokyo, Japan).

Tissue sections (4 µm thickness) were used for immunostaining. Following deparaffinization, sections were pretreated at 121°C for 15 min in 10 mM citrate buffer (pH 6.0) for PDIA3 and proliferation marker protein Ki-67 staining, and Histofine^® Antigen Activation Liquid (pH 9.0; Nichirei Biosciences, Inc.) for P-STAT3 staining. Endogenous peroxidase was blocked in 100% methanol (Wako Pure Chemical Industries, Ltd.) containing 0.3% hydrogen peroxide (Wako Pure Chemical Industries, Ltd.) for 30 min at room temperature. The sections were then incubated with antibodies for PDIA3 (1:500), P-STAT3 (1:400) and Ki-67 (MIB-1; cat. no. M7240; 1:100; Dako; Agilent Technologies, Inc.) in PBS containing 1% bovine serum albumin (Sigma-Aldrich; Merck KGaA) for 16 h at 4°C. The sections were further incubated with Histofine Simple Stain™ MAX PO (rabbit or mouse; Nichirei Biosciences, Inc.) for 30 min at room temperature, and peroxidase activity was visualized by 3,3′ diaminobenzidine. The sections were then counterstained with hematoxylin at room temperature for 1 min.

For PDIA3, the intensity and proportion of stained tumor cells were semi-quantitated. If the tumor cell staining was not apparent, the intensity and proportion were scored as 0. If it was apparent, the intensity of tumor cell staining was categorized into three grades: 1, weak; 2, moderate; and 3, strong. The proportional score of stained cells was also divided into three grades: 1, <10%; 2, 10-50%; and 3, >50%. A total score was calculated as the sum of the intensity and proportional scores. Two investigators evaluated the scores in a blind manner and the mean score was used. HCC tissue with a total score ≥4 was classified as having high expression and that with <4 was classified as having low expression. For P-STAT3, nuclear staining was considered to be a positive reaction. The P-STAT3 staining was classified as positive or negative (≥10% or <10% cells with nuclear staining, respectively). These expressions were evaluated manually under a light microscope (×100 magnification; Olympus Corporation). The Ki-67 index was calculated as the percentage of cells with positive nuclear Ki-67 immunostaining using e-Count cell counting software version 4.7 (e-Path Co, Ltd., Kanagawa, Japan) in those areas showing the highest nuclear labeling (so-called ‘hot spots’), which were determined using a light microscope (×400 magnification). The median number of tumor cells counted was 322 (range, 165-583 cells) per sample.

Cells were fixed in 4% paraformaldehyde for 20 min at room temperature. Following washing in PBS, the fixed cells were incubated in 50 mM glycine for 20 min at room temperature, and washed again. The cells were then permeabilized in 0.1% Triton X-100 for 30 min at room temperature. The cells were washed in PBS and blocking was performed using 10% goat serum (Wako Pure Chemical Industries, Ltd.) for 60 min at room temperature. The fixed cells were consecutively incubated with anti-PDIA3 mouse antibody (1:50) and anti-STAT3 rabbit antibody (cat. no. 12640S; 1:50; Cell Signaling Technology, Inc.), followed by Alexa 488-labeled anti-rabbit IgG antibody (cat. no. A11031) and Alexa 568-labeled anti-mouse IgG antibody (cat. no. A11034) (both 1:1,000; Thermo Fisher Scientific, Inc.). The slides were mounted in medium containing DAPI, and the images were captured using a Digital Eclipse C1 TE2000-E confocal microscope (×1,000 magnification; Nikon Corporation, Tokyo, Japan)

HCC tissues fixed as described above (4 µm thickness) were used for the TUNEL assay. The apoptotic cell death of HCC cells was determined by a TUNEL assay using an Apoptag^® Peroxidase In Situ Apoptosis Detection kit (Merck KGaA). The peroxidase activity was visualized by 3,3′ diaminobenzidine. The sections were then counterstained with hematoxylin at room temperature for 1 min and dehydrated. 99% xylene with Malinol (Muto Pure Chemicals Co., Ltd.) was used as a water-insoluble mounting medium. Nuclear staining was considered a positive reaction. The images were captured under a light microscope (×400 magnification; Olympus). The TUNEL index was calculated by the same method as that used for the Ki-67 index. The median number of counted tumor cells was 337 (range, 179-649 cells) per sample.

The data are expressed as the mean ± standard error of the mean or standard deviation (SD). Statistical comparisons between and among the groups were conducted using one-way analysis of variance (ANOVA) with Dunnett's post-hoc test, two-way ANOVA with Tukey's post-hoc test or a Mann-Whitney U-test. The clinicopathological parameters were analyzed using the χ² and Fisher's exact tests. P<0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using GraphPad Prism version 7.0 (GraphPad Software, Inc., La Jolla, CA, USA).

To evaluate the effects of PDIA3 expression in HCC, cell proliferation, apoptosis and clinicopathological features were investigated in 53 HCC samples. The cell proliferation was evaluated using Ki-67 immunostaining, and apoptosis was measured by a TUNEL assay. The HCC tissues with high PDIA3 expression exhibited a significantly higher Ki-67 index than those with low expression (25.8 vs. 10.1%, respectively; P<0.001; Fig. 1A and B). By contrast, HCC tissues in which PDIA3 was highly expressed revealed significantly fewer apoptotic cells than those with low expression (2.4 vs. 8.9%, respectively; P<0.001; Fig. 1A and C). However, no associations were observed between the PDIA3 expression level and any clinicopathological features (Table I). These results indicate that HCC tissues with high PDIA3 expression are associated with increased cell proliferation and decreased apoptosis.

The PDIA3 protein expression in HCC Huh-7 and HuH-1 cells was tested, and was demonstrated to be 1.7 and 2.0 times higher, respectively, than that in normal hepatocyte THLE-2 cells (P<0.001; Fig. 2A and B). This result indicates that Huh-7 and HuH-1 are appropriate cell lines for a model of HCC with increased PDIA3 levels.

To investigate the role of PDIA3 in the progression of HCC, the proliferation of HCC cells was tested following PDIA3 knockdown. First, it was ascertained that PDIA3 si-1 and si-2 successfully suppressed PDIA3 expression in Huh-7 and HuH-1 cells (Fig. 2C). The proliferative ability of Huh-7 cells treated with PDIA3 si-1 and si-2 was 50.6 and 70.0% compared with those treated with Ctrl si at 96 h, respectively. The proliferation of HuH-1 cells treated with PDIA3 si-1 and si-2 was 48.0 and 71.4% compared with the control group at 96 h, respectively (P<0.001 at 72 and 96 h; Fig. 2D). Overall, the PDIA3 knockdown significantly inhibited cell proliferation in the two HCC cell lines.

To investigate the effect of PDIA3 silencing on HCC cell proliferation, the apoptosis and cell cycle progression of HCC cells were investigated following PDIA3 knockdown. The silencing of PDIA3 induced apoptosis in the HCC cells (P<0.001 compared with Ctrl si; Fig. 2E and F). Furthermore, the cell cycle analysis demonstrated that knockdown of PDIA3 significantly increased the SubG1 population and decreased the G₀/G₁ phase cells in the two HCC cell types (P<0.001 compared with Ctrl si in the SubG1 population; Fig. 2G and H). The increased SubG1 population signified the induction of apoptosis. These findings suggest that PDIA3 knockdown inhibits cell proliferation in HCC cell lines, likely through an apoptosis-dependent mechanism.

As excessive ER stress causes apoptosis, the effect of the downregulation of PDIA3 expression on ER stress was investigated in the HCC cells. To evaluate the level of ER stress, the expression of ER stress marker GRP78 was tested following PDIA3 knockdown. No changes in GRP78 levels were observed, indicating that silencing of PDIA3 does not induce ER stress (Fig. 3).

The interaction of PDIA3 with STAT3 in HCC cells was investigated. Double immunofluorescence staining revealed the colocalization (merge; yellow) of PDIA3 (red) and STAT3 (green) (Fig. 4A). PDIA3, an ER protein, was mainly observed in the cytoplasm where it colocalized with STAT3 (Fig. 4A; top panel; merged). Notably, PDIA3 also localized to the nucleus in ~5% of Huh-7 cells (Fig. 4A; bottom panel). In these cells, nuclear PDIA3 tended to colocalize with STAT3 (Fig. 4A; bottom panel; merged). Additionally, the co-immunoprecipitation experiments demonstrated that PDIA3 was bound to STAT3 in the HCC cells (Fig. 4B). These results suggest a physiological interaction between PDIA3 and STAT3 in HCC cells. Furthermore, the western blotting revealed that inhibition of PDIA3 expression reduced P-STAT3 levels (Fig. 4C). This result indicates that PDIA3 is associated with STAT3 activation via a certain interaction in HCC cells.

The expression of downstream targets of STAT3 was investigated by immunoblotting in HCC cells following treatment with PDIA3 siRNAs (Fig. 5A and B). The knockdown of PDIA3 led to a decrease in the expression of XIAP and Bcl-XL in Huh-7 cells, and that of survivin, Mcl-1, XIAP and cyclin D1 in HuH-1 cells. The difference between the expression of Mcl-1 and cyclin D1 in the Huh-7 cells for si-1 and si-2 is possibly due to off-target effects of the siRNAs. These results indicate that PDIA3 knockdown affects apoptotic pathways by decreasing the levels of the downstream anti-apoptotic proteins of the STAT3 pathway in HCC cells.

The JAK/STAT3 signaling inhibitor AG490 suppressed the levels of P-STAT3 (Fig. 6A). Subsequently, to confirm that PDIA3 knockdown led to a decrease in cell proliferation mainly through STAT3 signaling, the cell proliferation of HCC cells treated with PDIA3 siRNAs was examined in the presence of AG490. AG490 markedly decreased cell proliferation in the two HCC cells (P<0.001 at 72 and 96 h; Ctrl si vs. Ctrl si + AG490; Fig. 6B). In the Huh-7 cells, PDIA3 si-1 or si-2 + AG490 decreased the cell proliferation only by a further 7.2 or 5.2%, respectively, compared with the effect of AG490 alone at 96 h, and only the additional effect of si-1 was statistically significant (PDIA3 si-1 + AG490 vs. Ctrl si + AG490, P<0.05; PDIA3 si-2 + AG490 vs. Ctrl si + AG490, P=0.22; Fig. 6B). In the HuH-1 cells, PDIA3 si-1 or si-2 + AG490 decreased the cell proliferation by 8.3 and 4.3% more than AG490 alone at 96 h, respectively, and only the effect of si-1 was statistically significant (P<0.001 PDIA3 si-1 + AG490 vs. Ctrl si + AG490, P<0.001; PDIA3 si-2 + AG490 vs. Ctrl si + AG490, P=0.086; Fig. 6B). Notably, these additional inhibitory effects of PDIA3 knockdown under AG490 treatment conditions were small compared with the inhibitory effect of AG490 alone (56.5 and 48.4% decrease in Huh-7 and HuH-1, respectively, compared with Ctrl si at 96 h). These results demonstrated that the inhibitory effect of PDIA3 knockdown on the proliferation of HCC cells with deactivated STAT3 was weak, suggesting that PDIA3 knockdown inhibits cell proliferation predominantly through STAT3 signaling.

Immunohistochemical staining of 35 HCC tissues revealed an association between PDIA3 and P-STAT3 levels. The HCC tissues that were positive for P-STAT3 exhibited significantly higher PDIA3 expression than those that were negative for P-STAT3 (mean PDIA3 expression score ± SD, 4.8±0.8 and 2.9±1.5, respectively; P<0.001; Fig. 7A and B). The HCC tissues that were P-STAT3 positive also demonstrated a tendency towards a higher Ki-67 index than those that were negative, although this was not statistically significant (mean Ki-67 index, 24.0 and 16.9%, respectively; P=0.069; Fig. 7C). By contrast, the P-STAT3-positive tissues demonstrated significantly fewer apoptotic cells than those that were negative (2.9 and 7.9% apoptotic cells, respectively; P<0.001; Fig. 7D). However, no association was observed between the P-STAT3 status and the clinicopathological features (Table I). Overall, PDIA3 expression was associated with a P-STAT3-positive status, which in turn was associated with decreased apoptosis in HCC tissues.