Gene expression levels of STEAP1 in non-tumor and liver cancer tissues were evaluated using gene expression profiles of GSE14520 and GSE36376 from the Gene Expression Omnibus, a public and freely available database. The GSE14520 dataset includes 488 samples of 241 non-cancerous and 247 cancerous hepatic tissues. These datasets have been widely used and well accepted in bioinformatics analysis of liver cancer. The GSE36376 dataset includes 433 samples consisting of 193 non-cancerous hepatic tissues and 240 cancerous tissues. The correlation between STEAP1 levels and clinical outcomes of patients with liver cancer was investigated using GSE14520 and the Cancer Genome Atlas Program (TCGA) (16). We used a receiver operating characteristic curve to determine the cutoff value. In total, 247 patients from the GSE14520 dataset and 360 patients from TCGA, all with liver cancer, were divided into two groups having high or low levels of STEAP1, respectively. Kaplan-Meier analyses of survival were performed based on these groups. Statistical analyses were performed using EZR software version 1.33 (17).

Gene set enrichment analysis (GSEA) was performed using the open source software, GSEA 4.0.3. Initially, we set two groups (STEAP1_high and STEAP1_low) in GSE14520-GPL3921, which includes 225 liver cancer samples in total. We conducted GSEA of the two groups using Hallmark gene sets. Gene sets showing a NOM P-val. (P-value) <0.05 and false discovery rate (FDR) Q-val. (FDR) <0.25 were considered significant. Differentially expressed genes (DEGs) between these two groups were identified using an online tool, GEO2R, with |logFC| >1.5 and an adjusted P-value <0.05.

HepG2 and Hep3B cell lines were purchased from the American Type Culture Collection; these were authenticated by short tandem repeat DNA profiling prior to all experiments. Both cell lines were cultured in DMEM containing 10% fetal bovine serum (FBS), 2 µM L-glutamine and 1% penicillin-streptomycin (the medium and all supplements from Sigma-Aldrich; Merck KGaA).

Control small-interfering RNA (siRNA; Control; #4390843; Thermo Fisher Scientific, Inc.) and two independent siRNAs targeting human STEAP1 (siSTEAP1; D-003713-01: 5′-GGAGAGAAUUUCACUAUAU-3′ and D-003713-02: 5′-UAAAGAAGAUGCCUGGAUU-3′; Dharmacon) were transfected using Lipofectamine RNAiMAX (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Cells were seeded at a density of 3×10⁵ cells/well into 6-well plates and cultured for 24 h at 37°C. Subsequently, cells were transfected with control siRNA or siRNA targeting human STEAP1, and incubated for 72 h at 37°C. Final siRNA used per well was 25 pmol. After incubation, floating cells in media were collected, adhesive cells were washed and collected, and both were immediately used for experiments.

Total RNA was extracted using TRIzol Reagent (Thermo Fisher Scientific) according to the manufacturer's protocol. Subsequently, complementary (c)DNA was synthesized from the RNA using a SuperScript VILO cDNA synthesis kit (Thermo Fisher Scientific). qPCR was performed with an Applied Biosystems 7300 Real-time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The analysis of target genes (STEAP1 and c-Myc) was conducted in quadruplicate using a POWER SYBR-Green Master Mix (Thermo Fisher Scientific, Inc.) as previously described (18). The thermal profile of the qPCR program consisted of 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 sec at 95°C and 1 min at 60°C, and a dissociation stage at the end of the run from 60°C to 95°C. Transcript levels were normalized to β-actin expression and analyzed using the 2^−ΔΔCq method. The following PCR primers were designed: 5′-CCCTTCTACTGGGCACAATACA-3′ and 5′-GCATGGCAGGAATAGTATGCTTT-3′ for STEAP1; 5′-TTTTTCGGGTAGTGGAAAACC-3′ and 5′-GCAGTAGAAATACGGCTGCAC-3′ for c-Myc; and 5′-GGCATCCTCACCCTGAAGTA-3′ and 5′-GAAGGTGTGGTGCCAGATTT-3′ for β-actin.

As previously described (19), cells were solubilized in radioimmunoprecipitation assay lysis buffer (50 mM Tris-HCl, pH 7.5, 1% NP-40, 0.5% Na-deoxycholate, 1 mM EDTA, 150 mM NaCl, 1 mM EGTA, and protease inhibitor cocktail; Sigma-Aldrich; Merck KGaA), and centrifuged at 12,000 × g for 10 min. The supernatants were collected, and protein concentrations were determined using a bicinchoninic acid Protein Assay Kit (Thermo Fisher Scientific, Inc.). Equal amounts of protein were separated on MULTIGEL II mini gels (Cosmo Bio Co., Ltd.) and transferred to polyvinylidene fluoride membranes using a QBlot Kit (ATTO, Tokyo, Japan). The blots were probed using the following primary antibodies: anti-STEAP1 (sc25514; Santa Cruz Biotechnology), anti-STEAP1 (#88677; Cell Signaling Technology), anti-cyclin D1 (#2987; Cell Signaling Technology,), anti-p27 Kip1 (#3686; Cell Signaling Technology), anti-c-Myc (OP10L; EMD Biosciences), and anti-actin-horse radish peroxidase (HRP; sc-47778; Santa Cruz Biotechnology).

Hepatocellular carcinoma cells were seeded at a density of 2×10³ cells/well into 96-well plates. Control siRNA or two independent siRNAs targeting human STEAP1 were transfected 24 h after seeding. Cell viability was assessed at 0, 24, 48 and 72 h using a WST-1 assay (Premix WST-a Cell Proliferation Assay; Takara Bio) and Infinite M1000 Pro microplate reader (Tecan Japan). A growth curve was constructed by plotting absorbance against time.

Liver cancer cells were seeded at a density of 3×10⁵ cells/well into 6-well plates and cultured for 24 h. Subsequently, cells were transfected with control siRNA or an siRNA targeting human STEAP1, and incubated for 72 h. After incubation, floating cells in media were collected and adhesive cells were washed, fixed in ethanol, and stained with propidium iodide using a cell-cycle analysis kit (FxCycle PI/RNase Staining Solution; Thermo Fisher Scientific), followed by analysis on a BD FACS II (BD Biosciences) instrument using FACSDiva (BD Biosciences) as previously described (20).

Apoptosis was evaluated using an Annexin V/7-amino-actinomycin (AAD) staining kit (BD Biosciences). Liver cancer cells were seeded at a density of 3×10⁵ cells/well into 6-well plates and cultured for 24 h. Subsequently, cells were transfected with control siRNA or an siRNA targeting human STEAP1, and incubated for 72 h. After incubation, floating cells in media were collected and adhesive cells were washed, stained with Annexin V and 7-AAD, and analyzed on a BD FACSCanto II (BD Biosciences) instrument using FACSDiva (BD Biosciences) as previously described (21).

Total RNA was reverse-transcribed using an RT² First Strand Kit (Qiagen). PCR array was performed using RT² Profiler™ PCR Array Human MYC Targets (PAHS-177Z; Qiagen) according to the manufacturer's protocol.

The significance of differences was determined by Student's t-test, Mann-Whitney U test, log-rank test or one-way ANOVA followed by Bonferroni's post-hoc test, as appropriate. Pearsons correlation was used to perform the correlation analysis. All statistical analyses were performed using EZR software version 1.33 (17). Statistical significance was defined as P<0.05.

We first investigated the expression of STEAP1 in patients with liver cancer using publicly accessible datasets (GSE14250 and GSE36376) from the Gene Expression Omnibus. In both datasets, STEAP1 is over-expressed in liver cancer tissues compared to non-cancerous hepatic tissues (Fig. 1A and B). Next, we evaluated the correlation between STEAP1 expression and survival in patients with liver cancer using GSE14520 and TCGA datasets. Patients with high STEAP1 expression presented with significantly shorter overall survival (OS) and recurrence-free survival (RFS) in GSE14520 and significantly shorter OS in TCGA (Fig. 1C-E). These data imply that STEAP1 may have oncogenic functions in liver cancer.

To evaluate the effect of STEAP1 on liver cancer, we performed STEAP1 silencing using an RNA interference method in two different liver cancer cell lines, HepG2 and Hep3B. Knockdown efficiency was examined by RT-qPCR and western blot. STEAP1 expression in these cell lines was significantly down-regulated 72 h after transfection of two independent siRNAs (Fig. 2A, B, D and E). We next evaluated the impact of STEAP1 silencing on liver cancer cell lines using WST-1 assays. STEAP1 silencing significantly reduced proliferation in both cell lines (Fig. 2C and F). Based on these data, we concluded that STEAP1 activated proliferation in liver cancer cell lines.

To evaluate the mechanism of decreasing proliferation in response to the knockdown of STEAP1, we examined the effects of STEAP1 silencing on the cell cycle in the liver cancer cell lines, HepG2 and Hep3B. STEAP1 silencing significantly induced G₁ arrest in both liver cancer cell lines (Fig. 3A and B). We also performed a flow cytometry analysis using Annexin V/7AAD staining to evaluate the rate of apoptosis. However, an increased percentage of apoptosis was not observed in STEAP1-silenced liver cancer cell lines (Fig. S1A and B). To analyze the mechanism of G₁ arrest in HCC cell lines induced by the knockdown of STEAP1, we evaluated protein levels of several cell-cycle-related proteins in liver cancer cell lines using western blot. The expression of the G₁ arrest-associated protein, cyclin D1, was decreased, whereas the expression of p27, which promotes cell-cycle arrest, was apparently increased (Fig. 3C).

To clarify the pathways related to STEAP1, we first extracted DEGs between low and high STEAP1 liver cancer samples in a publicly accessible dataset, GSE14520-GPL3921, using GEO2R. The significant DEGs with |logFC| >1.5 and adjusted P-value < 0.05 are highlighted in red and blue colors. Each gene was represented as a volcano plot (Fig. 4A) and listed in a table (Table I). Next, we conducted GSEA to explore the gene sets regulated by STEAP1 in liver cancer and found five pathways which were significantly enriched (NOM P-val <0.05 and FDR Q-val <0.25; Fig. 4B, Fig. S2, and Table SI). The genes belonging to MYC_TARGET_V2 were the most significantly enriched among these five pathways (Fig. 4C and D). Based on these findings, we hypothesized the existence of a relationship between STEAP1 and c-Myc in liver cancer. To confirm this, we evaluated their expression using the publicly accessible datasets, GSE14250, GSE36376, and TCGA. Pearson's correlation coefficient analysis revealed a significant positive relationship between STEAP1 and c-Myc in all datasets (Fig. 4E-G).

To confirm the relationship between STEAP1 and c-Myc in liver cancer, we evaluated the expression of c-Myc after STEAP1 knockdown in HepG2 and Hep3B cell lines by RT-qPCR and western blot. As we expected, downregulation of c-Myc was observed in both cell lines when transfected with siRNA targeting STEAP1 compared to non-targeting siRNA (Fig. 5A-D). Next, we conducted a PCR array to analyze components of c-Myc-related genes; most were significantly downregulated by STEAP1 silencing (Figs. 5E and S3). Taken together, our data suggest that c-Myc lies downstream of STEAP1, and that the STEAP1-c-Myc pathway promotes cell proliferation and cell-cycle progression in liver cancer.