A total of 20 matched pairs of HCC and normal adjacent liver tissues were acquired during surgery between May 2014 and April 2016 at the Shenzhen People's Hospital (Shenzhen, China). Patients met the following criteria according to our previous work: i) male/female patient ratio was 1:1, patients ranged in age from 35 to 75 years; ii) HCC patients were newly diagnosed with liver function tests of Child-Pugh grade A; iii) without history of anticancer therapy; iv) without any distant metastasis; v) without other types of malignant tumors, autoimmune, liver disease or serious heart, lung, kidney, or blood disease; and vi) seronegative for hepatitis B surface antigen and HCV. The samples were either snap-frozen or stored in liquid nitrogen for protein extraction or fixed in formalin and embedded in paraffin for immunohistochemistry analysis.

sgRNAs were designed using the Optimized CRISPR Design tool (http://crispr.mit.edu/) as follows: sgRNA719, 5′-CCTCACACGCAAATAGCTGA-3′; sgRNA720, 5′-TACTCATCCATGTGACCATG-3′; and sgRNA721, 5′-GTTATGGTTCTCACAGATGA-3′. LV-Ctrl was constructed as control lentiviruses that did not have any sgRNAs used for targeting the HIF-1α gene. The cDNAs encoding the sgRNAs for 3 gene knockout sites in the first exon of the HIF-1α gene were synthesized and purified (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA). Different programmed lentiviral plasmids based on the pLenti-CAS9-sgRNA-egfp (GeneChem Co., Ltd., Shanghai, China) were constructed as previously described (22).

The CRISPR/Cas9 lentivirus production was performed following previously described protocols (22). Briefly, 12 145-mm Petri dishes were prepared and 1×10⁷ HEK293T cells [American Type Culture Collection (ATCC) Manassas, VA, USA] were seeded in each dish with Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc.) containing 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc.). pLenti-CAS9-sgRNA-egfp (20 µg), psPAX2 (15 µg; Addgene, Inc., Cambridge, MA, USA), pMD2.G plasmid (10 µg; Addgene, Inc., Cambridge, MA, USA) and linear PEImax (Polysciences, Inc., Warrington, PA, USA) were mixed and added to the HEK293T cells in each dish. After 6 h of incubation in a CO₂ incubator at 37°C, the medium was replaced with complete medium. After 48 h, the cell supernatants were filtered through a 0.45-µm filter (EMD Millipore, Billerica, MA, USA). Subsequently, supernatants were ultracentrifuged (Beckman Coulter, Inc., Brea, CA, USA) at 20,000 × g for 2 h at 4°C, and the lentiviruses were suspended in PBS and stored at −80°C.

The titer of the concentrated lentiviruses was determined by a quantitative polymerase chain reaction (qPCR) method as described previously with 293T cells (23). Subsequently, SMMC-7721 cells (ATCC) were cultured in Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc.) with 10% FBS and 1% penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Inc.) in a CO₂ incubator at 37°C. SMMC-7721 cells were infected with the lentivirus at a multiplicity of infection (MOI) of 2.5. To check the total percentage of GFP-positive cells, single-cell suspensions were prepared in PBS plus 2% FBS from trypsinized cells at 72-h after infection. After resuspension, cells were subjected to fluorescence activated cell sorting (FACS) analysis with a BD Accuri C6 flow cytometer (BD Biosciences, San Jose, CA, USA) as described (22). Next, SMMC-7721 cell genomic DNA was extracted using a DNeasy Blood & Tissue kit (Qiagen, Hilden, Germany). A mismatch-sensitive T7E1 Assay kit (New England BioLabs, Inc., Ipswich, MA, USA) was then used to confirm CRISPR/Cas9 cleavage and targeted sequence disruption, according to the manufacturer's protocol. Subsequently, the PCR fragments containing each knockout gene locus were cloned with primers in different group using a Takara PCR amplification kit (cat. no. DR011; Takara Bio, Inc., Otsu, Japan) according to the manufacturer's protocols. The following primers were used: i) LV-H719 + 7721 cells group: 5′-TCTAATCCTTCTGTGATAAGCAG-3′ (forward) and 5′-CAAAATCAAAACATTGCGACCAC-3′ (reverse); ii) LV-H720 + 7721 cells group: 5′-ACATGAAAGCACAGAAATTGC-3′ (forward) and 5′-TGCCTTGGGTAAGTACAATAGC-3′ (reverse); and iii) LV-H721 + 7721 cells group: H721, 5′-TCTTCTTGTGCCCTTTTTAGGTG-3′ (forward) and 5′-CTTACCATTTCTGTGTGTAAGC-3′ (reverse). The cloned DNA sequences were then inserted into a plasmid using pMD19-T (Takara Bio, Inc., Otsu, Japan) for DNA sequencing (Invitrogen; Thermo Fisher Scientific, Inc.) and comparison. The SMMC-7721 cell line where HIF-1α has been successfully knockout after LV-H721 infection at MOI=2.5 was established and was named as 7721-HIF-1α-KO cells for further testing.

The total cell RNA was extracted using the TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The concentration and the quality of RNAs were measured with a NanoDrop 8000 (Thermo Fisher Scientific, Inc.). A PrimeScript^® II First Strand cDNA synthesis kit (cat. no. H6210B; Takara Bio, Inc.) was used to synthesize the cDNA. Subsequently, the gene mRNA expression levels were detected by qPCR using Thunderbird SYBR-Green qPCR Mix (cat. no. QPS-201; Toyobo Life Science, Osaka, Japan) according to the manufacturer's protocol. Primers used in this experiment were as follows: vegf, 5′-TGCTCTACCTCCACCATGCCA-3′ (forward) and 5′-GAAGATGTCCACCAGGGTCTCG-3′ (reverse); mdr1, 5′-TGATGCTGCTCAAGTTAAAGGG-3′ (forward) and 5′-TTGCCAACCATAGATGAAGGATAT-3′ (reverse); β-actin, 5′-GTCCACCGCAAATGCTTCTA-3′ (forward) and 5′-TGCTGTCACCTTCACCGTTC-3′ (reverse); and gfp, 5′-TGCTTCAGCCGCTACCC-3′ (forward) and 5′-AGTTCACCTTGATGCCGTTC-3′ (reverse).

Protein samples from patient tissues and SMMC-7721 cells were prepared using radio-immunoprecipitation assay lysis buffer (RIPA; Beyotime Institute of Biotechnology, China), and the protein concentration was detected by the BCA protein quantitation kit (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Western blot analysis was then performed based on protocols reported in a previous study (22). The following primary and secondary antibodies were used: Anti-VEGF (cat. no. 19003-1-AP; 1:2,000, ProteinTech, Inc., Chicago, IL, USA), anti-MDR1 (cat. no. PA5-28810; 1:2,000, Thermo Fisher Scientific, Inc.), anti-HIF-1α (cat. no. 20960-1-AP; 1:2,000, ProteinTech, Inc.), anti-β-actin (cat. no. MA1-140; 1:2,000, Thermo Fisher Scientific, Inc.), horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (cat. no. 7074; 1:2,000, Cell Signaling Technology, Inc., Danvers, MA, USA) and HRP-conjugated anti-mouse IgG (cat. no. 7076; 1:2,000; Cell Signaling Technology, Inc.) antibodies.

SMMC-7721 cells were cultured in 96-well (10⁴ cells per well) or 6-well (3×10⁵ cells per well) plates. A total of 150 µM cobalt chloride (CoCl₂; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), a well-known hypoxia mimetic (24,25), was used to mimic tumor hypoxic conditions. In order to examine the cell proliferation, at 8 h after the lentivirus infection, 20 µl MTT (5 mg/ml) was added to each sample and incubated for 3 h. The formazan crystals were dissolved in dimethyl sulfoxide (150 ml/well), and the absorbance of each sample at 490 nm was measured.

For cell apoptosis analysis, a PE-Annexin V Apoptosis Detection kit (cat. no. 559763; BD Biosciences) was used. Briefly, 5 µl of Annexin V-PE and 5 µl of 7-AAD were added in 100 µl cell suspension with 1× binding buffer plus 5% FBS, and incubated for 15 min in the dark. Finally, cells were washed twice and diluted with buffer and immediately subjected to FACS analysis (22). For the detection of cell cycle progression, PI-RNase solution (cat. no. 550825; BD Biosciences) was used, according to the manufacturer's protocol. Flow cytometry was then performed using a BD FACS Arial flow cytometer (BD Biosciences), and the results were analyzed with FlowJo software (Tree Star, Inc., Ashland, OR, USA).

A Transwell system with a 6.5-mm diameter and an 8.0 µm pore polycarbonate filter membrane (Corning Inc., Corning, NY, USA) was used, and Matrigel (cat. no. 356234; Corning Inc.) was used to form a thin gel layer on the wells. At 12 h after infection, the SMMC-7721 cells (1×10⁵ per well) were added to the upper chamber. In the lower chambers, 20% FBS was added as a chemoattractant. CoCl₂ (150 µM) was added to simulate hypoxia in both upper and lower chambers. After 24 h, the cells that had penetrated to the lower surface of the filter were stained with crystal violet and counted using a Nikon microscope (Nikon Corporation, Tokyo, Japan).

A scratch wound assay was also performed. Briefly, the SMMC-7721 cells were seeded into 6-well plates (3×10⁵ per well). Following infection with lentivirus after 12 h, a linear wound was created in the monolayer with a sterile 100-µl plastic pipette tip, and cells were then gently washed with PBS immediately. Cells were cultured under the CoCl₂-simulated hypoxic conditions and then observed under the microscope.

The overlapping gene mFluc2, which included the mCherry fluorophore and Fluc2 gene (the gene that expresses the firefly luciferase protein), was cloned from the pmCherry-C1 (Clontech Laboratories, Inc., Mountainview, CA, USA) and pGL4.17 plasmids (Promega Corp., Madison, WI, USA) using a Takara PCR amplification kit. The PCR cycling conditions were set as follows: 1 min at 95°C for denaturation, 45 cycles of 95°C for 30 sec, and 55°C for 30 sec, then 72°C for 1 min. Lastly 1 h s at 4°C. Next, the pWPXLd-mFluc2 plasmid was constructed, and the LV-mFluc lentivirus was packaged according to previous methods (22). The SMMC-7721 cells were infected with LV-mFluc with an MOI of 2.5, and the mCherry-positive SMMC-7721 cells were purified by flow cytometry (BD FACS Arial device) and cloned. Subsequently, a dual-luciferase reporter assay in the intended SMMC-7721-Fluc cells was conducted using the Dual-Luciferase^® Reporter Assay system according to the manufacturer's protocol (cat. no. E1910; Promega Corp. Madison, CA, USA). The primers used in this experiment are as follows: Clone for mCherry gene: 5′-GGGGATCCATGGTGAGCAAGGGCGAGGAGGATA-3′ (forward) and 5′-TCTTTATGTTTTTGGCGTCTTCCATCTTGTACAGCTCGTCCATGCCGCCG-3′ (reverse); and clone for Fluc gene, 5′-CGGCGGCATGGACGAGCTGTACAAGATGGAAGACGCCAAAAACATAAAGA-3′ (forward) and 5′-ACGGAATTCTCACTCGAGCAATTTGGACTTTCCG-3′ (reverse).

Seventy male BALB/c nu/nu mice (4–6 weeks old, 16–20 g) were purchased from the Shanghai SLAC Laboratory Animal Co., Ltd. [license no. SCXK (HU) 2007-0003; Shanghai, China]. Mice were housed in specific pathogen-free (SPF) conditions, with a 12-hours light cycle and food and water at ad libitum. In order to establish a subcutaneous HCC model, 4×10⁶ SMMC-7721 cells were injected in the right flank of the mouse subcutaneously (n=4/group). After 9 days, 5×10⁷ LV-H721, which was a lentivirus-mediated CRISPR/Cas system with small guide RNA-721, were injected into the tumor tissues. Tumor tissues injected with PBS or LV-Ctrl (a lentivirus without CRISPR/Cas system targeting HIF-1α) were used as control groups. The tumors were harvested 3 days post LV-infection for immunohistochemical examination.

The procedure for the establishment of the orthotopic liver tumor model was conducted according to a previously described method (26). Briefly, pentobarbital was dissolved in saline to obtain a 10 mg/ml solution. Mice were anesthetized with pentobarbital (50 mg/kg) through intraperitoneal injection and placed in a supine position. A small transverse incision was made below the sternum to expose the liver. A total of 5×10⁶ control lentivirus-infected SMMC-7721-Fluc cells (7721) or LV-H721-infected SMMC-7721-Fluc cells at 48 h after infection were suspended in 50 µl PBS and slowly injected into each mouse liver. At 5 days after implantation, mice were subdivided into four groups as follows: i) 7721-induced HCC + sham surgery (Ctrl; n=8); ii) 7721-induced HCC + HAL surgery (n=8); iii) control lentivirus-infected 7721-induced HCC + HAL (HAL + LV-Ctrl; n=8); and iv) LV-H721-infected 7721-induced HCC + HAL (HAL + LV-H721; n=8). HAL was performed in the mice by ligating the main branch of the hepatic artery. The tumor volume was measured by bioluminescence imaging with 200 µl D-luciferin potassium salt solutions (15 mg/ml; Sigma-Aldrich; Merck KGaA) after HAL during 2 weeks. After 2 weeks, the mice (n=5/group) were sacrificed and mouse tumor samples were collected for immunohistochemistry analysis. Furthermore, an extended observation of tumor-bearing mice with four different treatments was carried out again, and the mouse overall survivals were recorded.

Human or SMMC-7721-induced mouse tumor tissues were fixed with 4% neutral paraformaldehyde. Next, the paraffin-embedded sections (human and mouse tumor tissues) were incubated with the anti-HIF-1α antibody (cat. no. 20960-I-AP; 1:100; ProteinTech, Inc., Rosemont, IL, USA) overnight at 4°C, followed by incubation with 50–100 µl 1× the secondary antibody solution for 1-h (cat. no. KIHC-5, 1:1, ProteinTech, Inc.). The signals were detected by staining the sections with 3,3′-diaminobenzidine (DAB; ProteinTech, Inc.) and hematoxylin. Furthermore, to monitor the microvascular density (MVD), the mouse tumor sections were stained with an anti-CD31 monoclonal antibody overnight at 4°C (cat. no. 14-0319-80; 1:300, Thermo Fisher Scientific, Inc.), and the rest of the test was carried out as previously described.

The DeadEnd™ Colorimetric TUNEL System (cat. no. G7130, Promega Corp.) was used for the TUNEL staining assay. Briefly, the tumor slides were incubated with terminal deoxynucleotide transferase (TdT) and a biotinylated nucleotide mixture at 37°C for 30 min. Subsequently, the endogenous peroxidases was blocked by immersing the slides in 0.3% hydrogen peroxide in PBS for 3–5 min at room temperature. The slides were incubated with streptavidin-HRP and visualized with DAB. Negative controls were set up by substituting distilled water for TdT in the working solution. The results are presented as the ratio of the TUNEL-positive cells to the total number of cells.

The statistical analysis was performed with SPSS version 17.0 software (SPSS, Inc., Chicago, IL, USA). All the data were analyzed using one-way analysis of variance followed by least significant difference post-hoc analysis, or using an unpaired two-tailed Student's t-test (as appropriate). The survival curves were obtained by Kaplan-Meier analysis, and the prognostic significance was analyzed by the log-rank test. Statistically significant differences were denoted by P<0.05. All data were acquired from four repeated experiments.

The expression levels of HIF-1α protein were first analyzed in 20 paired HCC and normal adjacent liver tissues by western blot and immunohistochemical analyses. As shown in Fig. 1A and B, the HIF-1α protein levels in the HCC tissues were higher in comparison with those in the normal adjacent liver tissues, and the difference was found to be statistically significant (P<0.05; Fig. 1C).

SMMC-7721 cells, a human liver cancer cell line, were also used to establish an HCC model in BALB/c nu/nu mice. The results revealed that the implanted HCC cells had a significantly higher expression of HIF-1α protein as compared with the control groups (Fig. 2A). Subsequently, the SMMC-7721-induced HCC model was used in further CRISPR/Cas9-mediated HIF-1α-knockout experiments.

To identify sgRNAs specifically targeting the gene loci of HIF-1α, three sgRNAs were designed to target the first exon of HIF-1α according to guidelines of the Optimized CRISPR Design online tool (Fig. 2B). The pLenti-CAS9-sgRNA719/720/721-egfp plasmids, based on the pLenti-CAS9 vector, were established (Fig. 2B). The CRISPR/Cas9 lentiviruses with different sgRNAs were labeled as LV-H719, LV-H720 and LV-H721. Subsequently, SMMC-7721 cells were infected with the lentiviruses at an MOI of 2.5, and the gene disruption efficacies were assessed, followed by the T7E1 endonuclease assay and DNA sequencing. The results demonstrated that all three sgRNAs led to DNA frameshift mutations and had high gene disruption efficiencies (Fig. 2C). Furthermore, the results of the western blot analysis revealed that LV-H721 infection significantly reduced the expression of HIF1-α with the maximum efficiency (Fig. 2D). The percentage of total GFP-positive cells in LV-H721-infected SMMC-7721 cells was also measured to assess the lentiviral transfection efficiency, and was found to be 93.6% (Fig. 2E). LV-Ctrl (a lentivirus without CRISPR/Cas system targeting HIF-1) could also express GFP in the infected SMMC-7721 cells and had equal lentiviral transfection. It was observed that HIF-1α knockout caused by LV-H721 transfection in the cells was observed to significantly downregulate the mRNA levels of vegf and mdr1 as compared with those in the control groups (all P<0.001; Fig. 2F). As shown in Fig. 2G, the expression levels of the VEGF and MDR1 protein, which has been reported to be under the control of HIF-1α (17), were also inhibited in the present study. Thus, LV-H721 was used as a CRISPR/Cas9-mediated vector for further experiments.

The efficiency of LV-H721-mediated HIF-1α knockout was also examined in the mice. The SMMC-7721 ×enograft HCC tissues were harvested 3 days post LV infection. Immunohistochemical examination revealed that HIF-1α was highly expressed in the control and LV-Ctrl-treated HCC tissues. Furthermore, the expression of HIF-1α was significantly reduced in the LV-H721-treated tissues when compared with those in the control groups (both P<0.001; Fig. 2A). Collectively, these data indicated that the lentivirus-mediated CRISPR/Cas9 efficiently disrupted the expression levels of HIF-1α gene and its targets, including VEGF and MDR1, in liver cancer cells and xenograft tumor tissues.

First, the effect of the CRISPR/Cas9 based HIF-1α knockout on migration and invasion was determined in the SMMC-7721 cells (7721-HIF-1α-KO). The SMMC-7721 cells were infected with the lentiviruses, and CoCl₂, a well-known hypoxia mimetic, was used to simulate the hypoxic conditions in vitro. In the transwell invasion assay, the number of invasive 7721-HIF-1α-KO cells (namely the CoCl₂ + LV-H721 group) that penetrated the polycarbonate filter was significantly reduced, and this difference was statistically significant (both P<0.001; Fig. 3A). Meanwhile, 7721-HIF-1α-KO cells exhibited further reduced relative migration rates at 24 and 48 h post LV infection compared with the control groups in the scratch-wound assay, and the difference was statistically significant (all P<0.001; Fig. 3B). These results suggested that the disruption of HIF-1α with the CRISPR/Cas9 system, particularly under the hypoxic microenvironment, inhibited liver cancer cell migration and invasion.

Hypoxic conditions are reported in the majority of advanced HCC tissues, in particular in patients who undergo TAE. Therefore, the growth characteristics of the 7721-HIF-1α-KO cells under the hypoxic microenvironment were analyzed using an MTT assay. The optical density, an indicator of the number of viable cells, was monitored subsequent to infection, and a growth curve was plotted (Fig. 4A). The results demonstrated that cell proliferation was profoundly suppressed at 48 and 96 h post infection between the LV-H721-treated group and the control groups (all P<0.001).

Subsequently, the cell cycle progression under hypoxic conditions was assessed at 48 h post infection in vitro. SMMC-7721 cells infected with LV-H721 exhibited marked changes in their cell cycle progression, and there was an ~10% decrease in the number of cells in the G0/G1-phase, as well as a concomitant increase in the G2/M-phase (Fig. 4B). However, the S-phase ratio, serving as the marker of cell division, was almost consistent. These data suggested that HIF-1α knockout did not induce evident cell cycle arrest (Fig. 4B and C). Furthermore, HIF-1α knockout led to a high percentage of early cell apoptosis under hypoxic conditions, and the rates of cell apoptosis were 7.5% at 24 h and 28.6% at 48 h post infection, with a statistically significant difference observed (P<0.001; Fig. 4D). These results suggested that disruption of HIF-1α by the CRISPR/Cas9 system inhibited cell proliferation and induced cell apoptosis in the SMMC-7721 cells.

The SMMC-772-Fluc cells were established and cloned (data not shown). On days 5, 10, 15 and 20 after implantation of these cells into the mice, the tumor volume was measured by bioluminescence imaging (Fig. 5A). On day 20, the representative features of the tumors in mice with live bioluminescence imaging are shown in Fig. 5B and tumor outlines in mouse livers are shown in Fig. 5C. The results demonstrated that HAL inhibited the progression of HCC when compared with that observed in the control group (P<0.05; Fig. 5A). Notably, the combination of HAL and LV-H721 exhibited a more significant anti-tumor effect on day 20 as compared with that in the control, HAL and HAL + LV-Ctrl groups (P<0.001, P<0.05 and P<0.05, respectively; Fig. 5A). Furthermore, the overall median survival results revealed that the combination of HAL + LV-H721 significantly prolonged the survival of tumor-bearing mice from ~40 to 60 days, as compared with the other groups (P<0.05; Fig. 5D), indicating that HIF-1α disruption enhanced the efficacy of HAL in the treatment of HCC.

The expression of CD31 was used as a marker of vascular endothelial cells, or tumor microvascular density (MVD). From immunohistochemical analysis of CD31 protein expression (Fig. 6A), the CD31 was significantly higher in the mice treated with HAL and HAL+ LV-Ctrl as compared with the control group, whereas CD31 was significantly lower in the combined HAL + LV-H721 treatment group in comparison with the HAL and HAL+ LV-Ctrl groups (all P<0.05; Fig. 6B). A TUNEL assay was also performed to detect the apoptotic cells in the tumor tissues. The results demonstrated that combined treatment with HAL + LV-H721 significantly enhanced cell apoptosis compared with that observed in the control, HAL and HAL + LV-Ctrl groups (P<0.01, P<0.05 and P<0.05, respectively; Fig. 6C and D). Collectively, these results indicated that HAL treatment may promote the microvascular growth. However, HIF-1α knockout in tumor tissues significantly inhibited the tumor hypoxia-mediated increase in CD31 expression and the microvascular growth, while it promoted tumor cell apoptosis in vivo. These results further highlight the relevance of HIF-1α inhibition in combination with HAL in the treatment of advanced, unresectable HCC.