Histone acetylation and deacetylation are known epigenetic modifications, which can play a role in a gene regulation. Normally, histone acetyltransferase (HAT)-mediated histone acetylation results in a more relaxed chromatin structure that allows the transcription complexes to interact and activate the gene transcription. Inversely, deacetylation of histones is catalyzed by histone deacetylases (HDACs), and functions as a repressive mechanism of gene transcription by promoting chromatin condensation [1
]. As a consequence, aberrant activation of HDACs is considered to be involved in a process of carcinogenesis and malignant progression of tumor through silencing the tumor-suppressor genes.
HDACs are categorized into four classes; class I (HDAC1, 2, 3, and 8), class II (HDAC4, 5, 6, 7, 9 and 10), class III (Sirtuin1, 2, 3, 4, 5, 6 and 7), and class IV (HDAC11) [2
]. Classically, class I HDACs have been mainly investigated and reported to be involved in the pathogenesis and poor prognosis of patients in many types of cancers. Small molecules that block the enzymatic activity of HDACs (HDAC inhibitors), such as tricostatin A, valproic acid, romidepsin, and vorinostat, have been approved or are under investigation for the promising cancer treatment. However, these agents have possible side effects caused by inhibition of a broad range of HDACs [3
]. Moreover, which HDAC is responsible for the malignant behavior of cancer cells depends on the type of tissues. Therefore, it is important to identify which HDAC plays a pivotal role in the malignant transformation of cancer cells of the target organ.
Hepatocellular carcinoma (HCC), the most frequently diagnosed primary liver cancer, is an aggressive disease caused by chronic hepatitis B or C virus infection, excessive alcohol intake, non-alcoholic steatohepatitis, or exposure to aflatoxin B [4
]. The prognosis for patients with advanced stage HCC is poor because of its high rates of recurrence and intrahepatic metastasis. Consequently, there is an essential requirement for the development of an improved therapeutic strategy that aims to inhibit HCC progression. The malignant transformation of tumor cells is related to dedifferentiation of the cells [5
]. In addition to the primary dedifferentiation events such as increased proliferation, enhanced glycolytic activity, and a loss of cellular specific functions, subsequent genetic and epigenetic alteration often cause a loss of cell–cell adhesion, increased cell motility, and anchorage-independent cell proliferation, which are the crucial malignant phenotypes that trigger invasiveness and metastasis [6
]. In HCC, class I HDAC1 and HDAC2 were mainly reported to be associated with poor prognosis [7
], but inhibition of class II HDAC, HDAC4 and HDAC5 has also been reported to have significant anti-cancer effects [9
]. A number of HDAC inhibitors have been shown to be effective in preventing the growth of HCC as well as the other types of cancer, however, the pivotal HDAC responsible for the dedifferentiation of HCC cells remains unclear.
In this study, we analyzed the gene expression of a panel of HDACs and found that expression of HDAC9, one of the class II HDAC, is positively correlated with a dedifferentiated phenotype of HCC cells. Functional analysis using genetic and pharmacological inhibition revealed that its expression is required for the anchorage-independent cell growth in undifferentiated HCC cells. We also found a possible link between HDAC9 and aldehyde dehydrogenase (ALDH) and showed a correlation of ALDH activity to the stemness feature. Our data provide useful information for preventing malignant progression of HCC by targeting specific HDACs.
Dedifferentiation of cancer cells is a crucial cue for malignant transformation and progression. To improve the prognosis of patients, it is important to identify the factors that regulate the dedifferentiation process in order to develop target-specific drugs and gene therapies. In this study, we aimed to identify the HDAC that is responsible for the HCC cell dedifferentiation and found HDAC9 as a target molecule for inhibiting stemness features. HDAC9 is identified as a member of the class II HDACs whose N-terminal domains have an MEF-2 interacting region [12
]. HDAC9 mRNA and its alternative spliced isoform HDRP are reported to be expressed throughout various tissues while its expression is relatively low in liver [12
]. We observed a minimum expression of HDAC9 in the normal hepatic cell line HC and well-differentiated HepG2 and HuH1 hepatoma cells that retain a hepatic phenotype, such as hepatic gene expression and drug-metabolizing activities. In contrast, HDAC9 mRNA was positively detected in HLE and HLF cells, typical undifferentiated HCC cell lines that show a mesenchymal phenotype [14
]. Consistent with previous reports, class I HDACs were ubiquitously expressed among the cell lines tested, suggesting that HDAC9, but not class I HDACs, may play a specific role in undifferentiated HCC cells [17
]. Analysis of the TCGA database revealed that HDAC9 mRNA expression in human HCC patients correlated positively with the expression of marker genes of mesenchymal phenotype and stemness, while correlating negatively with those of hepatic marker genes. A recent study reports that HDAC9 expression is associated with poor prognosis, being an independent prognostic parameter in a study cohort with 37 HCC patients [18
]. These reports and our results provide a hypothesis that HDAC9 is involved in the dedifferentiation process of HCC cells.
EMT is one of the dedifferentiation processes of epithelial tumor cells toward a mesenchymal phenotype, and is a major obstacle for effective cancer treatment. TGF-β-induced EMT has been established as a model of EMT in many cancers [19
]. We observed typical morphological changes from epithelial to mesenchymal cell shape and expression change of marker genes in TGF-β-stimulated HuH1 cells. HDAC genes showed differential regulation in that only the HDAC9 gene was up-regulated, and class I HDACs were down-regulated during the EMT process. A number of studies have shown an involvement of HDAC9 in the ability of migration and invasion [20
]. However, in the case of TGF-β-induced EMT, HDAC9 suppression barely inhibited the phenotypic change of HuH1 cells, suggesting that HDAC9 does not locate in the upstream of EMT induction mediated by TGF-β-SMAD or non-SMAD pathway [19
]. HDAC9 inhibition by si-RNA and chemical inhibitor did not reduce the motility of undifferentiated HCC cells. Therefore, upregulation of HDAC9 may be one of the phenotypic changes of EMT that is upregulated in the mesenchymal type of cells. In endothelial cells, HDAC9 is induced via signal transducer and activator of transcription 3 (STAT3) signaling to proliferate by PDAC-secreted proangiogenic factors [24
]. HDAC9 is reported to be regulated post-transcriptionally by several micro RNAs in retinal, oral, breast, and gastric cancer, and as being associated with poor prognosis [21
]. Analysis of other signaling pathways that induce dedifferentiation and malignant phenotype via HDAC9 may help us to understand the essential role of HDAC9 in tumor dedifferentiation.
Anchorage-independent cell growth and self-renewal ability are the important features of cancer stem cells (CSCs) to drive tumorigenesis after metastasis. CSCs are a sub-population of tumor cells responsible for their initiation, recurrence, and metastasis. Previously, we observed a 0.1–1.0% sphere formation rate in HLE and HLF cells seeded in a low-attachment dish [16
]. In this study, HDAC9 suppression significantly reduced the growth of these spheres in both cell lines. Quantitative PCR revealed that sphere-forming cells expressed much higher HDAC9 expression than 2D-cultured population. These results suggest that HDAC9 plays a crucial role in the sphere-formation of undifferentiated HCC cells. Similar results were reported in retinoblastoma and gastric cancer, however, the molecular pathway which links HDAC9 to the stemness property remains unknown [21
Considering the question above, we searched the stemness-related genes whose expression or activity may be affected by HDAC9. Gene expression analysis of HDAC9-suppressed cells showed significant down-regulation of the ALDH1A3 gene in HLE cells. ALDH1A3 was a gene of positive correlation with HDAC9 in TCGA analysis (Table 1
). Accumulating evidence indicates that ALDH activity could be used for identifying a subpopulation of CSCs in many types of cancer including HCC [28
]. We therefore performed sphere-forming assays in the presence of the ALDH inhibitor disulfiram and confirmed that sphere-forming activity was significantly affected by the ALDH inhibitor at the low concentrations that have no inhibitory effect in 2D proliferation. This observation suggests that ALDH activity is the important regulator of stemness in undifferentiated HCC cells. It is reported that ALDH1A3 has higher catalytic efficiency than ALDH1A1 [30
]. The ALDH1A3 isotype has been reported to be responsible for ALDH activity in cardiomyocyte, breast cancer, and cholangiocarcinoma cells [31
]. These reports support the potential role of ALDH1A3 in conferring stemness properties in undifferentiated HCC cells. Interestingly, BRD4354 was able to down-regulate more stemness genes than si-HDAC9 knockdown in undifferentiated HCC cells (Figure 5
b). This result may be partly explained by the difference of target specificity of inhibition, with BRD4354 reported to inhibit the other HDACs, including class I and class II, to a lesser degree [35
Although ALDH1A1 is abundantly expressed in HLE and HLF cells, HDAC9 suppression barely affected the transcript level of ALDH1A1 in both cells. In the literature, ALDH1A1 is post-transcriptionally regulated by SIRT-mediated deacetylation at Lys353 in lung cancer cells [36
]. In addition, the HDAC9 variant that showed cytoplasmic localization due to the lack of nuclear localizing signal (NLS) sequence was identified [38
]. Class II HDACs, including HDAC5, HDAC6, and HDAC9, have been reported to catalyze the deacetylation of cytoplasmic, non-histone proteins in cancer cells [39
]. These reports allowed us to speculate that ALDH1A1 might be activated through a deacetylation by cytoplasmic HDAC9 variant. Although we only demonstrated the relationship between HDAC9 and ALDHs at the mRNA level, further study, including protein expression, acetylation status, and enzymatic activity of ALDH, will unveil the molecular mechanism of dedifferentiation regulated by HDAC9.
In summary, we found that the class II HDAC9 is a regulator of the differentiation and acquisition of stemness properties in HCC cells. TCGA analysis and the in vitro experiments suggest that the ability of anchorage-independent growth maintained by HDAC9 may be partly due to the regulation of ALDH (Figure 6
). Our findings provide useful information for drug development and gene therapies that target specific HDAC for cancer treatment.
4. Materials and Methods
4.1. Cell Culture and Reagents
Hepatoma cell lines (HepG2, HuH1, HLE and HLF) were purchased from Riken BioResource Research Center (RIKEN BRC). The human hepatic cell line Hc was kindly provided by Dr. Qin X.Y. of RIKEN, Japan. Cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) containing 10% FBS, 4.5 mg/mL glucose, and 2 mM L-glutamine in a humidified incubator at 37 °C and 5% CO2. TGF-β (Peprotech; Cranbury, NJ, USA), BRD4354 (Tocris Biosciences, Minneapolis, MN, USA), and disulfiram (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) were dissolved in a vehicle and stored in a freezer at −20 °C.
4.2. Cell Proliferation Assays
For cell proliferation assays, cells were seeded at 3 × 103 cells/well in a 96-well plate and subjected to the experiment. Cell viability was determined by the water-soluble tetrazolium (WST) assay using the Cell Counting Kit-8 (Dojin Chemical Co., Ltd.; Kumamoto, Japan) at 24, 48, and 72 h after the treatment. To determine cell viability, a 10% volume of CCK-8 reagent was added to each well, and the plate was incubated for 1 h at 37 °C. The relative amount of viable cells was calculated by measuring the absorbance at a wavelength of 450 nm.
4.3. Gene Expression Analysis
Total RNA was extracted from the cells by using the TRIZOL reagent in accordance with the manufacturer’s instructions (Invitrogen; Carlsbad, CA, USA). Half of a microgram of RNA was used to synthesize cDNA using GeneAce Reverse Transcriptase (Nippon Gene; Tokyo, Japan). A gene expression analysis was performed using PowerUp SYBR Green Master Mix (Thermo Fisher Scientific; Waltham, MA USA) and the Applied Biosystems (ABI) qPCR system (Thermo Fisher Scientific). The primers used in this study are listed in Table S1
. In public database analysis, the dataset of Liver Hepatocellular Carcinoma (Firehose Legacy) in The Cancer Genome Atlas (TCGA) database was analyzed by using the cBioPortal for Cancer Genomics (https://www.cbioportal.org/
). Samples with mRNA expression data (n
= 373) were included for the screening of co-expression genes.
4.4. Western Blot Analysis
Cell lysates were prepared from the cells subjected to the experiment by using Cell Lysis Buffer (Cell Signaling Technology; Danvers, MA, USA) and stored in a deep freezer until use. The protein concentration was determined by the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). Equal amounts of the protein sample were separated on a SDS-PAGE and immunoblotted with antibodies against human HDAC9 (Santa Cruz Biotechnology, Dallas, TX, USA), E-cadherin (Santa Cruz Biotechnology), vimentin (Santa Cruz Biotechnology), and beta-actin (Cell Signaling Technology). Horseradish peroxidase (HRP)-conjugated secondary antibodies were used to detect primary antibodies and were visualized with ECL detection reagent (GE Healthcare UK; Buckinghamshire, UK) and a light-sensitive X-ray film (Fuji film; Tokyo, Japan). The band images were analyzed by Image J software (NIH, Bethesda, MD, USA) for a quantification of protein expression. Uncropped western blot figures for Figure 1
b, Figure 2
b, and Figure 2
f are shown in Figure S4
4.5. Small-Interference RNA Experiment
A gene-knockdown experiment was performed by transfection of Stealth RNAi Pre-Designed siRNAs (Thermo Fisher Scientific) for HDAC9 mRNA (si-HDAC9) and a negative control (si-Control). Briefly, siRNA and Lipofectamine3000 (Thermo Fisher Scientific) were dissolved in Opti-MEM® I Reduced Serum Media (Thermo Fisher Scientific) separately, and then mixed to form transfection complex for 15 min. A final concentration of 10 pmol/mL of siRNA was used for the knockdown experiments.
4.6. Sphere-Forming Assay
A sphere assay was performed by culturing HLE and HLF cells in a 24-well low-attachment plate (Sumitomo Bakelite Co., Ltd.; Tokyo, Japan). Briefly, cells were seeded at 500 cells/well and treated with the reagent used in the experiments. Half of the medium was changed every three days. After 8 days of culture, formed spheres were photographed and their diameters were measured.
4.7. Statistical Analysis
Statistical comparisons for in vitro experiments were performed using Student’s t-tests. p < 0.05 was considered statistically significant.