1 Introduction
Hepatocellular carcinoma (HCC) is a common solid tumor and the third leading cause of cancer-related death worldwide (second leading cause of cancer-related death in less developed countries) [
1]. Every year, ~500,000 new cases are diagnosed in the Asia-Pacific region, with more than 60% of the total number occurring in China alone, often due to hepatitis B virus (HBV) infection [
2]. Individuals with chronic HBV infections, especially those with chronic liver disease and cirrhosis, are at an increased risk of developing HCC [
3,
4]. Compared to HCCs associated with other risk factors, HBV-related HCCs exhibit higher rates of chromosomal alterations and p53 mutations, enhanced activation of certain signaling pathways (e.g. the WNT/β-catenin pathway) and elevated expression levels of fetal liver/hepatic progenitor genes [
5].
As highly efficacious treatment regimens for late-stage HCC patients are still lacking, early diagnosis and intervention remain the key to improve survival. Several invasive and non-invasive diagnostic biomarkers [e.g. α-fetoprotein (AFP), AFP-L3, des-γ-carboxyprothrombin (DCP)] have already been identified and evaluated in different clinical settings [
6]. However, those biomarkers display limited sensitivity and specificity, especially with respect to early HCC stages and, therefore, combinations with other newly-identified candidate biomarkers are currently being evaluated [
6]. Over the past years, technologies including whole genome sequencing (WGS), RNA-sequencing (RNA-seq) and proteomic profiling have led to a new era in biomarker development that has improved our understanding of complex interactions between proteins, genes and noncoding RNAs in hepatocarcinogenesis in different settings [
7‐
9]. Most studies did, however, not distinguish HBV-related HCCs from HCCs resulting from other factors, and the analyzed tumor tissues were rarely matched with adjacent peritumor tissues (APTs) from the same patients. Thus far, only two RNA-seq-based genome-wide transcriptome analyses have been reported, identifying HBV-related HCC biomarkers using matched patient samples. Huang et al. [
10] for the first time conducted genome-wide transcriptome analyses of 10 matched pairs of cancer and non-cancerous tissues from HCC patients. They found that the 1378 differentially expressed genes (DEGs) identified were mostly enriched in 54 bio-function terms and 41 canonical pathways, thereby providing important clues for our understanding the molecular mechanisms underlying HCC development. In a subsequent study, Miao et al. [
11] conducted comparative multi-omics profiling of a complete collection of representative HCC patient samples for HCC biomarker identification. Four DEGs (
SFN,
TTK,
BUB1,
MCM4) were found to be associated with different tumor differentiation patterns, and the dual-specificity protein kinase TTK was identified as a promising prognostic biomarker for HBV-related HCC. However, an in-depth functional characterization of the DEGs or candidate biomarkers was lacking in the above two studies.
Here, we used RNA-seq to decipher and compare whole transcriptomes of paired tumor tissues and APTs from three HBV-related HCC patients and, by doing so, generated a list of DEGs. The most highly-ranked DEGs were further verified as potential HCC signature genes using qRT-PCR in another independent cohort of 30 HCC patients. Remarkably, we found that eight out of eight top-ranked and validated DEGs (TK1, CTTN, CEP72, TRIP13, FTH1, FLAD1, CHRM2, AMBP) contained putative OCT4 binding motifs in their promoter regions, and two of them (TK1, TRIP13) were verified as direct transcriptional targets of the master pluripotency factor OCT4.
2 Materials and methods
2.1 Reagents
Monoclonal (sc-5279) and polyclonal (ab19857, ChIP grade) anti-OCT4 antibodies were purchased from Santa Cruz Biotechnology and Abcam, respectively. Monoclonal anti-TK1 (ab-76,495) and polyclonal anti-TRIP13 (ab-204,331) antibodies were purchased from Abcam. Peroxidase-conjugated anti-mouse secondary antibody (7076), peroxidase-conjugated anti-rabbit secondary antibody (7074), normal Rabbit IgG (2729) and a biotinylated protein ladder detection pack (7727) were all purchased from Cell Signaling Technology. An anti-GAPDH (AG019) antibody was purchased from the Beyotime Institute of Biotechnology.
2.2 Patients and specimens
The source of the HCC patient samples has been described before [
12]. Briefly, 33 HCC patient samples were included for RNA-seq and validation in this study; samples from three male HBV-related HCC patients (T1/P1, T2/P2, T3/P3) were used for RNA-seq. Tumor tissues and paired APTs were collected at the time of hepatic carcinectomy. The tissues were collected strictly within the boundaries of tumors, whereas the APTs were collected at least 3 cm away from the tumor margins. All collections were conducted under supervision of the same pathologist. Pathological diagnoses were conducted by two independent and expert pathologists. All tumor tissues were confirmed as primary hepatocellular carcinoma. Retrospective data including demographic, preoperative laboratory and pathologic parameters were collected from electronic medical records. The 33 HCC patient cohort encompasses 27 men and 6 women, with a mean age of 56. All 33 patients were HBV surface antigen-positive without hepatitis C virus (HCV) infection and exhibited the same underlying cirrhosis etiology. Clinical information of these patients is listed in Table
1.
Table 1
Demographic and laboratory parameters of the subjects included in the study
Demographic parameters |
Gender (M/F) | M | M | M | 24/6 |
Age (y) (x ± SD) | 52 | 39 | 57 | 51.69 ± 6.69 |
Laboratory parameters [median (range)] |
ALT (U/L) | 30 | 54 | 34 | 21.25 (12.0–170.0) |
AST (U/L) | 108 | 43 | 40 | 62.5 (20.0–355.0) |
TBiL (μmol/L) | 12.1 | 24.4 | 11.6 | 49.9 (5.0–384.0) |
AKP (U/L) | 63 | 58 | 104 | 106.25 (47.0–349.0) |
GGT (U/L) | 301 | 126 | 69 | 122.75 (14.0–591.0) |
Pathologic parameters |
AFP (ng/ml) | > 50,000 | 12,628.5 | 8214 | 2401.1 (1.5- > 50,000) |
CEA (ng/ml) | 2.3 | 2.0 | 2.8 | 1.97 ± 0.79 |
CA199 (U/ml) | 39.3 | 37.6 | 6.1 | 14.95 (2.2–212.3) |
CA125 (U/ml) | 6.1 | 7.4 | 4.7 | 121.58 (5.6–208.1) |
Ferritin (ng/ml) | 534.2 | 413.9 | 567.1 | 11,229.45 (43.5- > 40,000) |
Tumor |
Number (S/M*) | M | S | S | 14/16 |
Size (cm) | 9 | 3.5 | 8.5 | 4.65 (1.5–25.0) |
Histopathologic grading (poor/ moderate/high) | poor | moderate | moderate | 16/10/4 |
PVTT* (Yes/No) | YES | YES | YES | 4/26 |
2.3 Cell lines and culture
The human HCC lines Huh7 and Hep3B were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China) and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) (both from HycloneR GE, USA) under humidified conditions at 37 °C with 5% CO2.
2.4 RNA library preparation and sequencing
Total RNAs of HCC tumor tissues and paired APTs from three HBV-related HCC patients were isolated using Trizol reagent (Lift Technologies, USA). RNA-library preparation was conducted using an Illumina standard kit according to the manufacturer’s protocol. Briefly, poly-A containing RNAs were purified, followed by fragmentation into small pieces. Next, the RNA fragments were converted into single-strand cDNA using superscript II reverse transcriptase (Invitrogen) and random hexa-primers (IDT, Coralville, Iowa, USA), followed by second strand synthesis using DNA polymerase I (Invitrogen) and E. coli RNase H (Invitrogen). After second strand synthesis, with end repairing and A-tailing, the synthesized double-stranded cDNA fragments were subjected to purification and subsequently ligated to Illumina adapters using a Quick ligation TM kit (NEB) and DNA ligase. The resultant cDNA adapter-modified cDNA libraries were fractionated on agarose gels, after which 200-bp fragments were excised and amplified by 15 polymerase chain reaction (PCR) cycles. After purification, the quality of the cDNA libraries was checked using a Bioanalyzer 2100 (Agilent). Next, the concentrations of the cDNA libraries were measured and diluted to 10 nM in Tris-HCl buffer prior to cluster generation. Cluster formation, primer hybridization and sequencing reactions were performed sequentially according to the manufacturer’s recommended protocol. The sequencing procedure was conducted on an Illumina® Hiseq2000 apparatus.
TopHat software [
13] was used to map reads complying with quality standards to the human reference genome assembly hg19 (NCBI build: GRCh37) with default parameters. The coordinates of the mapped reads were overlaid with genomic coordinates of a human gene set defined in RefSeq (NCBI) and then counted to determine the expression level of each individual gene. The expression level for each gene was then normalized to reads per kilo base of transcript per million mapped reads (RPKM). Differential expression analysis of each gene between tumor tissues and APTs was performed using the Limma package of Bioconductor from raw read counts. The false discovery rate (FDR) of each gene was determined according to the Benjamini–Hochberg procedure and a mean log2 (fold change [RPKM of tumor/ RPKM of APTs]) was calculated across all genes. The significantly differentially expressed genes (DEGs) were selected with their FDRs < 0.05 and fold change > 2 between tumors and APTs.
2.5 Functional enrichment analysis of DEGs
Functional enrichment analysis was employed to roughly characterize the DEGs in HCC tumorigenesis. Gene ontology (GO) (web-based gene set analysis tool kit) is a standard classification system of gene function and gene products. In addition, PANTHER (
http://go.pantherdb.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (
http://www.kegg.jp/) pathway analyses were used to reveal the potential roles of DEGs in liver carcinogenesis.
2.6 RNA isolation and DEG validation
The selected DEGs were initially validated by quantitative real-time PCR (qRT-PCR) using the same RNA samples as those for RNA sequencing. Furthermore, total RNAs were extracted from tumor tissues and paired APTs of 30 independent HBV-related HCC patients using an RNeasy Mini Kit (Qiagen, the Netherlands). The concentration and quality of the RNAs were determined using Merinton SMA1000. The cDNAs were synthesized using a Prime Script
Tm RT Master Mix (TaKaRa, Japan) according to the product manual. qRT-PCR of DEGs was performed using SYBR Premix Ex TaqII (TaKaRa, Japan) and analyzed on an ABI system 7500 (Life Technologies, USA). All assays were carried out independently in triplicate. The
GAPDH or
B2M genes were used as references for quantification. Relative gene expression values, expressed as fold changes, were subsequently determined using the Delta-Delta Ct method. The primers used for qRT-PCR are listed in Table
S1.
2.7 Analyses of associations between HCC signature genes and HCC clinical outcomes
We retrieved mRNA expression levels of eight HCC signature genes and
POU5F1 from The Cancer Genome Atlas (TCGA) RNA sequence database (
https://genome-cancer.ucsc.edu/). Patients meeting the following criteria were included in this study: pathological diagnosis with HCC and the availability of overall survival (OS), relapse-free survival (RFS), progression-free survival (PFS) and clinicopathological information. After excluding incomplete clinical data obtained from TCGA dataset, 348 cases were grouped and characterized (Table
S2). Survival curves for HCC patients were generated using an online database, Kaplan-Meier plotter (
http://kmplot.com/analysis/). The above-mentioned 384 patients were divided into high-expression and low-expression groups using median values of mRNA expression. Statistical differences in survival were assessed by log-rank (Mantel-Haenszel) test, and
p < 0.05 was considered statistically significant. Hazard ratios (HR) and
p-values were calculated online.
2.8 CRISPR/Cas9 knockout assay
OCT4 knockout (OCT4-KO) Huh7 or Hep3B cells were established using a CRISPR/Cas9 lentiviral system according to a standard protocol from Feng Zhang’s laboratory. In brief, single guide RNA (sgRNA) targeting OCT4A was generated as described previously [
14] and cloned into a pLentiCRISPRv2 vector. After confirming the sgRNA efficiency, we prepared OCT4-KO and non-targeting control (NTC) lentiviruses by transfecting the lentiCRISPRv2-KO and NTC, psPAX2, pMD2.G plasmids into 293 T cells, respectively, after which the virus titers were functionally determined and used to infect Huh7 or Hep3B cells at a low multiplicity of infection (MOI; 0.3–0.5). The infected cells were selected under 0.5–1 μg/ml puromycin for 7 days, and expanded for another 7 days for Surveyor assay, qRT-PCR and further functional characterizations.
2.9 Western blotting
Protein concentrations of tumor tissues and APTs were quantified using a Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific 23,227). Next, the samples were boiled for 30 min and the supernatants were loaded for SDS-PAGE (Bio-Rad) and transferred to PVDF membranes (Bio-Rad), which were incubated sequentially with primary and secondary antibodies, and developed using ECL reagent as described previously [
15]. GAPDH was used as an internal control for sample loading.
2.10 Immunohistochemistry
Immunohistochemical (IHC) examination of formalin-fixed tumor tissue sections was conducted as previously described [
16] using anti-TK1 (diluted at 1:100), anti-TRIP13 (diluted at 1:100) or anti-OCT4 (diluted at 1:500) antibodies.
2.11 Electrophoretic mobility shift assay (EMSA)
EMSA was conducted using a LightShift chemiluminescent EMSA kit (Thermo 20,148) according to the manufacturer’s instructions. Cell nuclear extracts were prepared using NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific 78,833). Subsequently, EMSA was performed as described previously [
17]. The sequences of double-strand biotin-labeled
TK1 and
TRIP13 probes (with the putative OCT4 motifs being underlined) were as follows.
TK1 probe 1: Biotin-5′GGGACCTGGCACGCACTAGGCG
CTCTGCATGCCCACAGGAGTGCTCTAGACG 3′.
TK1 probe 2: Biotin-5′CCTGGCAGGGTCTACGGATATT
ATTAGCATAGTCAGGACTTCAATTTTCTTT 3′.
TK1 probe 3: Biotin-5′CGGGCTAACACCTTCACACTTT
ATGCAGAAAGTAACAAGGAACCATTCTGAG 3′.
TRIP13 probe: Biotin 5′GGGAATTACCTGCGTTTTCACTG
ACATGCATCTCTCTTACCAGTCTGACCCAGATGGGG 3′.
2.12 Chromatin immunoprecipitation (ChIP)
ChIP analysis was performed as described previously [
17] using an EZ-ChIP™ Chromatin Immunoprecipitation Kit (Merck Millipore 2,673,061). Briefly, 2 × 10
7 - 5 × 10
7 cells were chemically cross-linked by the addition of 1% formaldehyde solution for 10 min at room temperature. The reaction was stopped by adding glycine to a final concentration of 125 mM. Next, he cells were rinsed twice with cold PBS and harvested using a silicon scraper. The resulting cell samples were sonicated to solubilize and shear cross-linked DNA to an average size of 300–800 bp. Immunoprecipitation was carried out using 5 μg rabbit anti-OCT4 (ab 19,857) and 100 μl protein A/G agarose beads (Pierce 20,421) (with normal rabbit IgG being the negative control). The primers of
TK1 and
TRIP13 for amplifying the DNA fragments were as follows:
TK1: F 5’-CTGGCAGGGTCTACGGATAT-3′ and R 5′ -CCGTCTAGAGCACTCCTGT-3′.
TRIP13: F 5′- TCGAGGTCCCTTCTTCCCAA-3′ and R 5′- AGTAGCCCCATCTGGGTCAG-3′. ChIP-real time PCR was performed using SYBR Premix Ex TaqII (TaKaRa, Japan) and analyzed on an ABI system 7500 (Life Technologies, USA). Extracted DNA fragments and input genomic DNAs served as templates. The anti-OCT4 precipitated DNA fragments corresponding to specific genes were quantified using qPCR and expressed as fold enrichment over anti-IgG precipitated DNA fragments. The Delta-Delta Ct method was used for relative quantification.
2.13 Mouse xenograft tumor model
NOD/SCID mice (female, 3–4 weeks old) were purchased from the Shanghai Experimental Animal Centre, Chinese Academy of Science. They were kept in the central animal facility of the First Affiliated Hospital of School of Medicine, Zhejiang University and housed in laminar-flow cabinets under specific pathogen-free conditions with a 12 h light/dark cycle. All studies on mice were conducted in accordance with the National Institute Guide for the Care and Use of Laboratory Animals. The animal protocol has been approved by the Committee of the Ethics of Animal Experiments, Zhejiang University.
For subcutaneous xenografting experiments, the mice were randomly divided into a treatment group (OCT4-KO) and a control group (NTC), and OCT4-KO or NTC Huh7 cells (2 × 106) were inoculated subcutaneously into each mouse. The diameters of the tumors were measured every three days with precision calipers. The tumor mass (xenograft) volumes were calculated using the formula: volume = [(tumor length) × (tumor width)2]/2. At day 18 after inoculation, the mice were sacrificed and the tumors removed, weighed and photographed.
2.14 Statistical analysis
All continuous variables were expressed as mean ± standard deviation (S.D.) or medians, and interquartile ranges. Comparisons of continuous variables were performed using Student t test or nonparametric Mann-Whitney U test. Comparisons between paired groups were carried out using paired t test or Wilcoxon signed ranks test when necessary. Correlations between RNA-seq measures and qRT-PCR measures were determined using the Spearman rank correlation coefficient. All data analyses were performed using SPSS 20.0 (Chicago, IL).
4 Discussion
HCC represents > 90% of primary liver cancers and is a major health problem due to its poor prognosis [
8,
20]. A number of risk factors for the development of HCC has been identified, and over 50% of HCC cases are attributable to persistent HBV infections [
5]. The incidence of HCC is growing worldwide, especially in less developed regions, which suffer from more HBV infections [
21]. HBV-mediated hepatocarcinogenesis is a multistep process that involves complex interactions between viral components, host genetic and environmental factors [
5,
21]. To improve the prognosis of HBV-related HCC, it is important to identify molecular biomarkers and signature genes that act at different stages of HCC development.
RNA-seq is a recently-developed approach based on deep-sequencing technologies that can be utilized for genome-wide transcriptome profiling. It provides a more precise measurement of gene transcript levels and their isoforms than other available methods [
22]. Here, we identified 164 DEGs in HCC tumor tissues compared to their matched APTs. Confirmation of the RNA-seq data for 14 selected DEGs using qRT-PCR revealed a high correlation between the two measurements. Moreover, the eight HCC signature gene RNA-seq measurements in our study highly correlated with those reported by Huang et al. [
10], although the identified DEGs were not further filtered and ranked in the latter study. Such a high correlation indicates a high robustness and reproducibility of the experimental systems used in both studies.
Some of the eight newly-identified HCC signature genes have previously been related to the development of various human cancers including HCC. First, overexpression of
CTTN (cortactin) has been closely associated with a poor prognosis in HCC resulting from an increased cell motility and metastasis [
23,
24]. Second, upregulation of
FTH1 (ferritin heavy chain 1) in HCC cells by TNF-α has been found to attenuate starvation-induced apoptosis [
25]. Third,
TK1 (thymidine kinase-1) has been associated with the early development of breast, lung, heart, esophageal and gallbladder cancer [
26‐
29]. A clinical investigation showed that serum TK1 levels > 2.0 pmol/L may indicate an increased risk for the development of malignancies later in life [
30]. The TK1 serum levels in patients with HCC were significantly higher than those in patients with benign hepatic diseases, placing serum TK1 as a complementary biomarker for the diagnosis of HCC [
31]. Fourth, systematic analysis of data from the Gene Expression Omnibus (GEO) database revealed that
TRIP13 (thyroid hormone receptor interactor 13) was upregulated in 12 human cancers, and that a high
TRIP13 expression indicated a poor prognosis for patients with liver, breast, gastric and lung cancer [
32].
TRIP13 gene copy numbers have been found to be increased in early-stage non-small-cell-lung cancer [
33], and increased
TRIP13 transcript and protein levels have been correlated with prostate cancer progression [
34]. Overexpression of
TRIP13 in non-malignant fibroblasts resulted in malignant transformation, and high expression of
TRIP13 in head and neck cancer cells led to aggressive, treatment-resistant tumors and enhanced DNA damage repair via nonhomologous end joining [
35]. Fifth, the expression of
AMBP has been found to be down-regulated in both HCC tissues and cell lines [
36], a finding consistent with our current results. As yet, the involvement of the remaining three genes (
CEP72,
FLAD1,
CHRM2) in hepatocarcinogenesis requires further validation.
Our data indicate that OCT4 may serve as a potential common transcription factor for all the eight HCC signature genes identified in this study. OCT4, encoded by the
POU5F1 gene, is a member of the POU family of transcription factors that is abundantly expressed in pluripotent stem cells (such as embryonic stem cells, embryonal carcinoma cells, induced pluripotent stem cells) and plays an essential role in the early development of mammalian embryos. The POU domains of the OCT4 protein can independently and flexibly bind half-sites of the characteristic octamer motif (ATGCA/TAAT) through which OCT4 recognizes enhancer or promoter regions of hundreds target genes [
37]. This flexibility allows OCT4 to form heterodimers with other transcription factors such as SOX2, cooperatively activating pluripotency and self-renewal-associated genes while simultaneously repressing genes that promote differentiation [
38]. Similar to, but distinct from its well characterized roles in pluripotent stem cells, OCT4 is generally considered to promote the self-renewal, survival, metastasis and drug resistance of cancer stem cells (CSCs) [
39‐
41].
Conditions within the microenvironment of HBV-infected cells may play an important role in the programming of liver cells towards cancer stem cells (CSCs) and the divergence of multiple HBV-induced pathways towards epithelial-mesenchymal transition (EMT) or stemness [
42]. IL-6 is secreted by inflammatory and stromal cells during liver regeneration and is known to support the conversion of non-CSCs to CSCs [
43]. Chang et al. [
44] found that HBV-related HCC patients had a higher serum level of IL-6, leading to increased expression of autocrine insulin-like growth factor I (IGF-I) and IGF-I receptor (IGF-IR), which subsequently promote the expression of OCT4 and NANOG in a STAT3-dependent manner. They further found that inflammation-conditioned medium generated by lipopolysaccharide-stimulated U937 human leukemia cells significantly up-regulated the expression of OCT4/NANOG, IGF-I/IGF-IR and activated IGF-IR/AKT signaling in HBV-active (HBV + HBsAg+) HCC cells [
45]. The close association between IL-6/IGF-1R signaling and HBV-related HCC progression suggests that HBV and proinflammatory cytokines are both required for and collaboratively involved in pluripotency factor induction and CSC formation, but the exact mechanism inducing OCT4 re-expression remains to be elucidated.
HBx of the HBV viral component is a multifunctional protein that activates many viral and cellular genes, modulating multiple cellular signaling pathways and regulating host cell proliferation, apoptosis and invasion [
5]. HBx does not bind DNA directly but regulates gene expression by transactivating multiple transcription factors. So far, there is no evidence that HBx can directly interact with the OCT4 protein, but there are several reports showing that it can up-regulate OCT4 expression in HCC cells via various routes. Arzumanyan et al. [
46] reported that overexpression of HBx in HepG2 cells may be associated with enhanced expression of the pluripotency factors OCT4, NANOG and KLF4, and the stemness-associated markers EpCAM and β-catenin. Another study showed that HBV viral components including HBx may induce regulated intramembrane proteolysis (RIP) of EpCAM in HCC cells, and cleavage of the EpCAM/β-catenin complex translocated to the nucleus leading to activation of canonical Wnt signaling that is accompanied with up-regulated OCT4 expression [
47]. There is evidence that β-catenin can bind specifically and directly to the promoter of OCT4 in mouse embryonic stem cells [
48], but such a direct regulation has not been demonstrated in CSCs. Recently, HBx has been shown to stimulate AFP expression prior to up-regulation of pluripotency factors such as OCT4, leading to partial reprogramming of liver cells toward HCC progenitor/stem cells [
49]. Furthermore, HBx can activate FOXM1 expression in HCC cells via the ERK/CREB pathway [
50], and FOXM1 has been reported to directly bind to and activate the OCT4 promoter in P19 embryonal carcinoma cells [
51]. Independently, HBV PreS1 has been found to stimulate the expression of multiple pluripotency factors including OCT4 and the self-renewal of liver CSCs [
52]. However, none of the above studies has definitively identified factors/mechanisms directly linking HBx to up-regulated OCT4 expression in the context of HBV-related HCC. In a study aimed at identifying differentially-expressed transcription factors between CSCs and non-CSCs from HCC cell lines, ZIC2 was found to be highly expressed in CSCs [
16]. Additional studies demonstrated that it acts as a direct transcription factor for OCT4. It remains to be established, however, whether this also holds in the context of HBV-related HCC and whether/how HBV viral components play a role. Interestingly, the extent of increase in OCT4 mRNA levels in HCC tumor tissues was found to be significantly lower than that in OCT4 protein levels, consistent with what we found in this study.
Taken together, we identified multiple HCC signature genes that may serve as biomarkers for HBV-related HCC diagnosis and prognosis. Their common transcriptional regulation by OCT4 implicates a key regulatory role of OCT4 in the occurrence and development of HCC, and thus positions it as a potential drug target for HCC.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.