Differential gene expression analysis after DAPK1 knockout in hepatocellular carcinoma cells

Building DAPK1-knockout Cells

Cas9-DAPK1 knockout plasmids were purchased from Santa Cruz Biotechnology, Inc. PLC/PRF/5 cells were collected and seeded into a six-well plate at a density of 5 × 10⁵ cells/well. The transfection mixture was composed of 2.5 µg Cas9-DAPK1 plasmid and 5 µL Lipo3000 reagent in 125 µL of OpTI-MEM medium. The mixture was added to the wells and mixed evenly. After 24 h, the solution was changed and 2 µg /mL of purinomycin was added. Monoclonal cells were selected using the limited dilution method, and the DAPK1 gene knockout was identified using western blot. The transfection kit was purchased from Thermo Fisher Scientific.

Western blotting

Cells were seeded in 6-well plates at a density of 5 × 10⁵ cells/well. After the culture medium was discarded, the cells were washed with PBS and 100 µL buffer was added to the wells (protease inhibitor PMSF: RIPA buffer = 1: 99). The cells were scraped and collected into a sterile centrifuge tube. To lyse the cells sufficiently, the cell lysate was pipetted up and down 30 times in a centrifugal tube on ice. After incubating on ice for 30 min, the cell lysate was centrifuged at 15,294 × g for 15 min at 4 °C. The supernatant was collected and the protein concentration was measured. Then, the loading buffer was added to the supernatant and mixed. The mixture was then incubated in boiling water for 10 min. A 10% SDS-PAGE gel was used for electrophoresis and membrane transfer. The transferred membrane was blocked with 5% fat-free milk for 3–6 h (Liu et al., 2020). After rinsing the membrane with TBST (2.4228 g Tris, 8 g NaCl, 1,000 mL H₂O, 0.6 mL Tween-20) solution, the transferred membrane was incubated with primary antibodies in TTBS solution containing 5% fat-free milk while shaking overnight. The blots were washed in TBST and incubated for 1 h with a 1:2,000 dilution of secondary antibodies. Western blotting was performed using the ECL reagent.

Immunohistochemistry

The study cohort was consisted of liver samples from 90 hepatocellular carcinoma patients. The specimens were purchased from Shanghai Outdo Biotech Company. This project has been approved by the Review Committee (Ethics Application Ref: YB M-05-02). All participants obtained written consent before surgery. Liver cancer tissues and adjacent tissues were fixed in 10% formalin solution for 24 h, and the moisture in the fixed tissues was removed with graded ethanol. Xylene was used to replace the ethanol in the tissue and make the tissue transparent. The transparent tissue was immersed in melted paraffin wax for soaking and embedding. Under freezing conditions, the tissue was cut into 4 µm thick slices using a microtome. The paraffin sections were dewaxed with xylene. Paraffin sections were soaked with hydrogen peroxide (0.3%) to inhibit endogenous peroxidase activity. After washing with PBS (3.25 g Na₂HPO₄.12H₂O, 0.265 g NaH₂PO₄.2H₂O, 8.766 g NaCl, 1,000 mL H₂O), these sections were treated with 0.2%–1% Triton-100X at 37 °C for 30 min. The rabbit anti-human DAPK1 (Sigma-D1319, 1:500 dilution) antibody was kept overnight in a humidifying chamber at 4 °C. The specimens were washed with PBS and incubated at 22 ± 3 °C for 1 h with an enzyme-conjugate secondary antibody, and then washed with PBS. All the samples were then stained with both diaminobenzidine and 20% hematoxylin. Images were obtained using a microscope for histopathological analysis.

Transcriptome analysis

When the cell reached 70–80% confluency, the culture medium was discarded, and the cells were washed with PBS. TRIzol (one mL) was added to each well and incubated for 1 min, after which all TRIzol was transferred to an RNase-free tube. The samples were frozen in liquid nitrogen and stored at −80  °C. After all the samples had been collected, they were transported to Shanghai Applied Protein Technology Co. using dry ice. Samples of each cell were collected separately three times and sent to the sequencing company. Finally, Shanghai Protein Technology Co. extracted RNA from all the cell samples and sequenced them.

Screening for differential genes

Input data for the gene expression differential analysis were read and the count data was obtained from the gene expression level analysis. DESeq2 (a method for the differential analysis of count data) was used for the differential expression analysis of genes in biological duplicates. The differential analysis was done in triplicate: cell samples were collected three times, RNA was extracted for three times and gene sequencing was performed three times. The p-value obtained from the original hypothesis test was corrected during the difference analysis. The revised standard is: p-value is less than 0.05, — log2foldchange — greater than 1 as the standard significance of difference. Based on the above analysis results, the volcano map and heat map of differentially expressed genes were drawn, using the NetworkAnalyst database (https://www.networkanalyst.ca/; Xia et al., 2013; Zhou et al., 2019a).

Pathway annotation and gene ontology enrichment analyses of differential genes

Metascape (https://metascape.org/gp/index.html#/main/step1) integrates more than 40 bioinformatics databases, providing easy access to comprehensive data analysis through a simple interface for a quick one-click analysis (Zhou et al., 2019b). It not only contains a biopathway enrichment analysis, a protein interaction network structure analysis, and a rich gene annotation function, but it also presents the results in a high-quality graphical language that can be easily understood. KEGG (Kyoto Encyclopedia of Genes and Genomes, http://www.kegg.jp/) is one of the databases commonly used for pathway studies (Kanehisa, 2002; Kanehisa et al., 2017). The KEGG database contains information on metabolism, genetic information processing, environmental information processing, cellular processes, biological systems, human disease, and drug development.

Protein–protein interaction network construction and module analysis

Functional interactions between proteins were analyzed to identify the genes involved with DAPK1. The interactive gene retrieval tool used in this study is Cytoscape software. The protein-protein interaction network (PPI) was constructed using the Cytoscape software by selecting the interactions with a composite score of >0.9 from the search tool to retrieve the interacting genes/proteins (STRING;version 11.5). The Cytoscape software (version 3.7.2) is an open-source bioinformatics database for visualizing molecular interaction networks. The Cytoscape plug-in for molecular complex detection (MCODE) analyzes densely connected regions. The choice criterion for hub genes in this study were as follows: MCODE score of 5, degree cut-off of 2, node score cut-off of 0.2, max depth of 100, and k-score of 2. The KEGG and Gene Ontology (GO) analyses were performed using the Metascape software.

Selection and analysis of hub genes

The top 10 genes most closely related to DAPK1 were obtained using the MCC algorithm with the Cytoscape plug-in, cytoHubba. The protein expression profile levels (low, medium, and high) of the hub genes of both liver cancer and adjacent tissue were obtained from the Human Protein Atlas (HPA) database. The mRNA expression levels of hub genes in liver cancer and normal subjects were obtained from The Cancer Genome Atlas (TCGA) database.

MTT assay

PLC/PRF/5 (PLC) cells and DAPK1-knockout PLC/PRF/5 (KO) cells growing in the logarithmic growth phase were digested with trypsin, centrifuged, and suspended in medium. Cells were diluted to 1 × 10⁴cells/mL, and 100 µL of cell suspension was added to 96-well plates. Cells were cultured in CO₂ at 37 °C. Cells were removed every 24 h and MTT (0.5 g/mL) was added to each well and incubated for 4 h. The cells were then dissolved in DMSO. The light absorption value of each well was measured at 490 nm with a microplate reader to determine cell viability. The above experiment was repeated three times in our laboratory.

HCC growth in nude mice

A total of 20 mice were randomly divided into two groups (n = 10 per group): the PLC group and KO group. PLC or KO cells in the logarithmic growth phase were injected into the PLC group or KO group of mice, respectively. Each mouse was injected with 1 × 10⁶ cells (0.2 mL) under the skin. The mice were raised in barrier facilities with HEPA-filters and fed autoclaved laboratory rodent feed for 21 days. Tumors were resected from the mice in order to weigh and histologically treat. Animal testing in this study strictly followed the guidelines for the Care and Use of Experimental Animals. The mice used in this animal experiment were approved by the Guangdong Animal Experiment Center (approval number: 44007200082541). BALB/ C-NU/NU immunodeficient male mice aged 3–5 weeks were purchased by our laboratory. The average weight of the mice was between 10 and 12 g. The mice were placed in a room reserved for immunosuppressed mice for 2–3 days to help them acclimate to their new environment before each group was inoculated with tumor cells. Twenty-one days post-transplant, 20 mice were euthanized by cervical dislocation and the growing tumors were removed.

Statistical analyses

The SPSS software (version 20.0; SPSS Inc. Chicago, IL, USA) was used for statistical analysis of the data, and the measurement data were expressed as mean ± standard deviation. In the case of normal distribution and homogeneity of variance, t-test was used for data comparison between two groups, and analysis of variance was used for data comparison between multiple groups. For non-normal distribution or uneven variance, Mann–Whitney U test was used for comparison of two groups of data, and Kruskal–Wallis H test was used for comparison of multiple groups of data. P > 0.05, the difference was not statistically significant; P < 0.05, the difference was statistically significant. The statistical graph was made by Graphpad Prism V8.0.2.263 software.

Establishment of stable cell lines

The DAPK1 gene was knocked out from PLC/PRF/5 cells using CRISPR technology. PLC/PRF/5 cells were transfected with the Cas-9-DAPK1 plasmid and screened using a finite dilution method. As shown in Fig. 1A, a total of nine purinomycin-tolerant cell lines were screened, seven of which had the DAPK1 gene knocked out. Those cells that had the DAPK1 knocked out were then cloned and amplified for subsequent experiments.

Identification of differentially expressed genes (DEGs)

We performed immunohistochemical experiments on clinically-obtained liver cancer tissues and adjacent tissues; the results are shown in Fig. 2A. The tumor tissue and paracancer tissue of 90 patients with hepatocellular carcinoma were obtained from clinical practice, in which 70 groups had obvious expression difference and 20 groups had no obvious expression difference. The results showed that the expression of DAPK1 in adjacent tissues was lower than that in the liver cancer tissues. For samples that were biological replicates, we used DESeq2 in the NetworkAnalyst database for differential gene expression analysis. A total of 732 DEGs were identified, including 415 upregulated genes and 317 downregulated genes, as shown in the volcano map and heat map (Figs. 2B–2C).

KEGG and Gene Ontology enrichment analyses of DEGs

To analyze the features of the DEGs, we performed a GO enrichment analysis of upregulated and downregulated genes using the Metascape database. The results showed that the upregulated genes were mainly enriched in naba matrisome associated, extracellular matrix organization, platelet degranulation, naba core matrisome, and cell–cell adhesion via plasma-membrane adhesion molecules (Figs. 3A and 3C). The downregulated genes were significantly enriched in homophilic cell adhesion via plasma membrane adhesion molecules, cellular component morphogenesis, negative regulation of response to wounding, regulation of hormone levels, and morphogenesis of the epithelium (Figs. 3B and 3D). The results of the KEGG pathway analysis demonstrated that both the upregulated genes and downregulated genes were mainly enriched in biological adhesion.

PPI network construction

The PPI network of DEGs was first obtained from the STRING database and then the densest connection region (24 nodes, 276 edges) was analyzed using the Cytoscape database (Figs. 4A–4C). Functional enrichment analysis of genes in this dense region showed that these genes were mainly enriched in post-translational protein phosphorylation, hemostasis, complement and coagulation cascades, positive regulation of NABA ECM glycoprotein, and lipid localization (Table 1).

Selection of hub genes

The MCC algorithm was used to obtain the top 10 genes from the PPI network using the Cytoscape plug-in cytoHubba. The ten key genes were: KNG1, TIMP1, Alpha 2-HS glycoprotein AHSG, TF, SCG3, CP, PROC, APOA1, C3, and SERRIND1, of which KNG1, TIMP1, AHSG and C3 had the highest correlation score with DAPK1 indicating that they were closely related to DAPK1. This study focused on the relationship between these four key genes and the DAPK1 gene.

Analysis of four hub genes

To evaluate the expression of these four genes in tissues, images of human HCC tissues were obtained from The Human Protein Atlas and compared with the mRNA expression data of the four genes provided by The Cancer Genome Atlas. As shown in Figs. 5A–5D, the results of the analysis of data from these two databases were consistent. Compared with normal tissues, KNG1, TIMP1, and AHSG expression was elevated. C3 mRNA and protein expression levels were also increased, but these results were not statistically significant.

The proteins were extracted from PLC/PRF/5 cells and KO cells for WB verification, as shown in Figs. 5E. 5F shows the normalized results of the western blot analysis and the statistical analysis using mean and standard deviation methods, shown with a bar chart. The levels of C3, TIMP1, and AHSG increased in DAPK1-knockout cells, whereas the level of KNG1 decreased.

Inhibition of PLC/PRF/5 growth by DAPK1

Next, the effect of DAPK1 expression on PLC/PRF/5 cell growth in vitro was assessed. MTT assay results showed that KO cells proliferated faster than PLC cells (Fig. 6A). Furthermore, we performed in vivo tumorigenesis experiments to confirm that DAPK1 inhibits PLC/PRF/5 cell proliferation. The results showed that the tumor volume and size in mice inoculated with PLC/PRF/5 cells were significantly less than in those inoculated with KO cells (Fig. 6B). The tumor growth rate and tumor weight of mice inoculated with PLC/PRF/5 cells were significantly higher than in those inoculated with KO cells (Fig. 6C and Fig. 6D).