Recent progression in the understanding of stem cell biology has greatly facilitated the identification and characterization of cancer stem cells (CSCs). Moreover, evidence has accumulated indicating that conventional cancer treatments are potentially ineffective against CSCs. Histone deacetylase inhibitors (HDACi) have multiple biologic effects consequent to alterations in the patterns of acetylation of histones and are a promising new group of anticancer agents. In this study, we investigated the effects of two HDACi, suberoylanilide hydroxamic acid (SAHA) and trichostatin A (TSA), on two CD44+ cancer stem-like cell lines from squamous cell carcinoma of the head and neck (SCCHN) cultured in serum-free medium containing epidermal growth factor and basic fibroblast growth factor. Histone deacetylase inhibitors inhibited the growth of SCCHN cell lines in a dose-dependent manner as measured by MTS assays. Moreover, HDACi induced cell cycle arrest and apoptosis in these SCCHN cell lines. Interestingly, the expression of cancer stem cell markers, CD44 and ABCG2, on SCCHN cell lines was decreased by HDACi treatment. In addition, HDACi decreased mRNA expression levels of stemness-related genes and suppressed the epithelial-mesencymal transition phenotype of CSCs. As expected, the combination of HDACi and chemotherapeutic agents, including cisplatin and docetaxel, had a synergistic effect on SCCHN cell lines. Taken together, our data indicate that HDACi not only inhibit the growth of SCCHN cell lines by inducing apoptosis and cell cycle arrest, but also alter the cancer stem cell phenotype in SCCHN, raising the possibility that HDACi may have therapeutic potential for cancer stem cells of SCCHN.
Evidence has accumulated indicating that only a minority of cancer cells with stem cell properties, cancer stem cells (CSCs), are responsible for the maintenance and growth of tumors.[1-3] These CSC subpopulations show a capacity for high tumorigenicity, self-renewal, and differentiation. To date, human CSCs have been identified, purified, and characterized in a variety of tumors.[4-8] In squamous cell carcinoma of the head and neck (SCCHN), since Prince et al. first demonstrated that the purified CD44+ population possessed the properties of CSCs, various cell surface markers, including CD133 and ALDH1, have been identified for the isolation of CSCs.[10-12] Moreover, numerous studies have demonstrated that conventional cancer treatments are potentially ineffective against CSCs, which are subsequently responsible for tumor relapse and distant metastases.[13-15] Therefore, the development of novel therapeutic strategies to overcome treatment resistance of CSCs is urgently needed.
Recent studies have demonstrated that not only genetic but also epigenetic changes, including DNA methylation and histone modifications, play an essential role in cancer development.[16, 17] Of histone modifications, acetylation and deacetylation affect the chromatin structure and gene expression, and they are regulated by the antagonistic activities of two enzymes, histone acetyltransferases (HATs) and histone deacetylases (HDACs). The aberrant expression of HDACs has been found in a variety of malignancies, which can alter gene transcription and enhance cell proliferation, indicating that HDACs are a promising target for cancer therapy. To date, several HDAC inhibitors (HDACi) have been developed and used in clinical trials for the treatment of cancers, and have been shown to induce differentiation, growth arrest, or apoptosis in tumor cells.[18, 19] In addition to these known anti-tumor effects, since HDACi are involved in the acetylation of non-histone proteins, the functions of various transcription factors, including p53, STAT3, and NFκB, are also modulated.[20-22] Thus, HDACi have multiple biologic effects, and therefore, their diverse effects have not been fully elucidated.
To explore the possibility of targeting epigenetics changes for treatment against CSCs, we investigated the effects of two HDACi, suberoylanilide hydroxamic acid (SAHA) and trichostatin A (TSA), on SCCHN cell lines. In previous studies, we and others have demonstrated that a small population possessing CSC properties was increased by culturing in serum-free medium (SFM) with epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF).[23-25] Using these culture conditions, we first examined the sensitivity of CD44+ cancer stem-like cells to HDACi. We then focused on whether the CSC properties are altered by treatment with HDACi. Interestingly, the expression of cancer stem cell markers, CD44 and ABCG2, on SCCHN cell lines was decreased by HDACi treatment. Furthermore, HDACi have demonstrated not only downregulation of the mRNA expression levels of stemness-related genes and sphere formation, but a synergistic effect with chemotherapeutic agents. Thus, our data revealed that HDACi may have therapeutic potential for CSCs in SCCHN.
Materials and Methods
Cell lines and culture conditions
Two human cultured SCCHN cell lines, HSC-2 and KUMA-1, were used in this study. HSC-2, which was established from an oral cancer, was purchased from the Japanese Collection of Research Bioresources (Osaka, Japan). KUMA-1 was established from a squamous cell carcinoma of the maxillary sinus. These cell lines were cultured in Serum-Free Expansion medium (StemCell Technologies, Vancouver, Canada) supplemented with epidermal growth factor (EGF; Calbiochem, Darmstadt, Germany) and basic fibroblast growth factor (bFGF; Calbiochem; 20 ng/mL) each, as described previously. Two HDACi, SAHA and TSA, were purchased from Wako (Osaka, Japan) and Sigma-Aldrich (St. Louis, MO, USA), respectively.
Trypsinized cells were resuspended, incubated with monoclonal antibodies for 30 min at 4°C, washed twice with PBS containing 1% FBS and 0.1% NaN3, and fixed with 1% paraformaldehyde in PBS. The antibodies used were anti-CD44- FITC (BD Pharmingen, San Diego, CA, USA) and anti-ABCG2-phycoerythrin (PE) (BD Pharmingen). Respective immunoglobulin G isotype-matched controls (BD Pharmingen) were used as negative controls.
Cell growth assays
Cell proliferation was determined using a CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA). Trypsinized cells were plated at a density of 1 × 104 cells/well in 96-well flat-bottomed plates, incubated overnight, and then exposed to various concentrations of HDACi for 48 h. Finally, 20 μL of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS solution was added to each well and the plate was incubated for 1 h at 37°C, and viable cells were quantified by measuring absorbance at 490 nm.
Apoptosis was assessed by flow cytometry as described previously. Briefly, cells were treated with HDACi for 48 h. Floating cells and adherent cells were harvested, washed, and assayed for apoptosis using a CaspACE FITC-VAD-FMK In Situ Maker (Promega). To determinate apoptotic and necrotic cells, 7-amino-actinomycin D (7-ADD) was added before analysis.
Cell cycle analysis
The cells incubated with HDACi for 48 h were trypsinized, washed, and fixed in ice-cold 70% ethanol, stained with propidium iodide RNase staining buffer (BD Pharmingen), and analyzed by flow cytometry. The population of subG1 including debris was deleted when gating singlet cells. The percentage of cells in G0/G1, S, or G2/M phase was determined using modfit lt software provided by Verity Software House (Topsham, ME, USA).
Sphere formation assay
The sphere formation assay was performed as described by Hong et al. Briefly, trypsinized cells were washed and then cells were seeded at a density of 1 × 103 cells/mL in 24-well ultra-low attachment plates in the presence or absence of HDACi. After 7 days' culture, spheres were counted using a microscope.
Total RNA was extracted from trypsinized cells using an RNeasy mini kit (Qiagen, Valencia, CA, USA). Quantitative (q)RT-PCR was performed using a Power SYBR Green RNA-to-CT 1-Step Kit on an Applied Biosystems StepOne (Applied Biosystems, Foster City, CA, USA). The melting curve was recorded at the end of every run to assess product specificity. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control gene. PCR primers used in this study were as follows: CD44 forward primer, 5′-CGGACACCATGGACAAGTTT-3′, reverse primer, 5′-GAAAGCCTTGCAGAGGTCAG-3′; ABCG2 forward primer, 5′-GCTGCAAGGAAAGATCCAAG-3′, reverse primer, 5′-CCGAAGAGCTGCTGAGAACT-3′; BMI1 forward primer, 5′-CAACTGGTTCGACCTTTGCAGATA-3′, reverse primer, 5′-GATGTGCCAATTGCTTCTAATGGA-3′; Notch forward primer, 5′-CACGCGGATTAATTTGCATCTG-3′, reverse primer, 5′-TGGGTGCACTCTTGGCATACA-3′; Nanog forward primer, 5′-TCCAACATCCTGAACCTCAGCTA-3′, reverse primer, 5′-AGGTTCCCAGTCGGGTTCAC-3′; Oct-4 forward primer, 5′-GCAATTTGCCAAGCTCCTGAA-3′, reverse primer, 5′-GCAGATGGTCGTTTGGCTGA-3′; transforming >growth factor-β (TGF-β) forward primer, 5′-AGCGACTC-GCCAGAGTGGTTA-3′, reverse primer, 5′-GCAGTGTGTTATCCCTGCTGTCA-3′, and GAPDH forward primer, 5′-GCACCGTCAAGGCTGAGAAC-3′; reverse primer, 5′-ATGGTGGTGAAGACGCCAGT-3′. The relative expression level of the target gene in HDACi-treated cells to that in control cells was determined by the method.
For immunoblot analysis of cultured cells, the cells were washed three times with PBS and lysed on ice in 1 mL lysis buffer (20 mM Tris-HCl [pH 7.6], 140 mM NaCl, 1 mM EDTA, 1% NP-40) containing aprotinin (10 μg/mL), and subjected to ultrasonic treatment. The lysates were subjected to immunoblot analysis as described by Miyashita et al. The following primary antibodies were used: CD44, ABCG2, α-tublin, E-cadherin, and β-actin (Cell Signaling Technology, Danvers, MA, USA).
Synergy between HDACi and chemotherapeutic agents, including cisplatin (kindly supplied by NIPPON KAYAKU) and Docetaxel (kindly supplied by Sanofi-Aventis) was assessed by the isobologram method, as described by Fivelman et al. Briefly, dose response assays were first carried out to obtain the 50% inhibitory concentration (IC50) of the individual drugs. For the combination assay, drug dilutions were made to allow the IC50 of the individual drugs to fall at about the fourth twofold serial dilution. Then, dilutions of each of the two drugs were prepared in fixed ratios. Tumor cells were seeded in 96-well flat-bottomed plates (BD Falcon, Franklin Lakes, NJ, USA) and treated with the combination of varying concentrations of HDACi and chemotherapeutic agent for 48 h. Two IC50s for each of the four combination curves were calculated separately, the fractional inhibitory concentration of two drugs was calculated for each point, and isobolograms were plotted.
Two-tailed Student's t-test was used for statistical analysis of data. P <0.05 was considered significant. Analyses were performed using stata 9.0 (Stata Corp., College Station, TX, USA).
HDACi inhibit the growth of SCCHN cells and induce apoptosis to SCCHN cells
Two SCCHN cell lines, HSC-2 and KUMA-1, were first cultured in serum-free medium containing EGF and bFGF. As reported previously, both cell lines expressed CD44 abundantly under this culture condition (data not shown). We then examined the effect of two HDACi, SAHA and TSA, against SCCHN cell lines. As shown in Figure 1, SAHA and TSA inhibited the growth of SCCHN cell lines in a dose-dependent manner as measured by MTS assays. The concentrations of SAHA and TSA that caused 50% inhibitory concentration (IC50) were as follows: HSC-2, 12.1 μM; KUMA-1, 7.1 μM for SAHA and HSC-2, 1.9 μM; KUMA-1, and 1.4 μM for TSA. Based on these data, subsequent experiments except the sphere formation assay were performed with 4 μM SAHA and 1 μM TSA, respectively.
To further evaluate whether HDACi could induce apoptosis in SCCHN cell lines, tumor cells were cultured with HDACi for 48 h and then analyzed for the proportion of apoptotic cells by flow cytometry. As shown in Figure 2, tumor cells treated with HDACi showed higher rates of apoptosis as compared with the control. It has been reported that CSCs derived from SCCHN could form tumor spheres, indicating that CSCs have self-renewal capacity. Therefore, we measured the sphere-forming ability of SCCHN cells by treatment with HDACi. As compared with the control, treatment with HDACi showed significantly lower numbers of tumor spheres (Fig. 3).
HDACi induced cell cycle arrest
Histone deacetylase inhibitors have been reported to induce cell cycle arrest at G1 or G2/M phase. We analyzed cell cycle distribution following HDACi treatment for 48 h by flow cytometry. Figure 4(a) shows representative diagrams of cell cycle distribution in each SCCHN cell line. In both SCCHN cell lines, the proportion of cells in G2/M phase was increased by HDACi treatment as compared with the control (Fig. 4b). At the same time, in KUMA-1, SAHA increased the proportion of cells in G0/G1 phase. Thus, we could confirm that HDACi showed inhibition of cell growth and induction of apoptosis and cell cycle arrest, as shown by numerous studies.
HDACi decreased the expression of CD44 and ABCG2 on SCCHN cells
CD44 and ABCG2 are known as cancer stem cell markers in SCCHN. We analyzed the expression of these two stem cell markers by flow cytometry, immunoblotting, and real-time quantitative RT-PCR. As indicated in Figure 5(a), HSC-2 and KUMA-1 cultured in serum-free medium containing EGF and bFGF expressed a high level of CD44 molecule, while KUMA-1 but not HSC-2 expressed ABCG2. Interestingly, treatment with each HDACi for 48 h decreased the expression level of CD44 and ABCG2 in both flow cytometric analysis and immunoblotting assay (Fig. 5a,b). Moreover, as expected, CD44 and ABCG2 mRNA expression levels were lower in HSC-2 and KUMA-1 treated with HDACi (Fig. 5c). Thus, the data in Figure 5 show that HDACi decreases the expression of cancer stem cell markers in SCCHN.
HDACi change the expression of stemness-related genes and EMT-related molecules in SCCHN cells
The key features of CSCs are activation of self-renewal and pluripotency genes; therefore, we analyzed the expression of BMI1, Notch, Nanog, and Oct-4 genes by treatment with HDACi using real-time qRT-PCR. As expected, HDACi downregulated the expression levels of four stemness-related genes in both HSC-2 and KUMA-1; however, the extent of downregulation varied (Fig. 6). On the other hand, epithelial-mesenchymal transition (EMT), which is characterized by the gain of stem cell properties, has been reported to play an important role in initiating CSCs. The hallmark of EMT is the loss of the E-cadherin, and TGF-β can induce EMT. Therefore, we determined the expression of E-cadherin using immunoblotting and TGF-β using real-time qRT-PCR. Treatment of SCCHN cell lines with HDACi showed increased expression of E-cadherin and decreased expression of TGF-β (Fig. 7), indicating that HDACi might act to suppress the EMT phenotype of CSCs.
HDACi enhance chemosensitivity of SCCHN cells
Finally, we examined whether HDACi enhanced the chemosensitivity of SCCHN cells. If HDACi induced the reduction of CSC properties, HDACi might be able to enhance the sensitivity of chemotherapeutic agents. The effects of combined of HDACi and chemotherapeutic agents were assessed by the isobologram method. As shown in Figure 8, when HDACi were combined with cisplatin and docetaxel, which are frequently used for the treatment of SCCHN clinically, synergistic effects were observed.
In previous studies, we reported that a subpopulation of CD44+ cells enriched under serum-free medium culture conditions possessed not only a marked capacity for forming tumor spheres, proliferation, migration, and invasion, but also resistance to apoptosis-inducing stimuli, including chemotherapeutic agents, as compared with that of CD44- cells.[23, 26] Based on these findings, we here investigated the possibility of HDACi for the treatment of CSCs in SCCHN, and we made the following important observations of CD44 + enriched SCCHN cell lines: (i) Two HDACi, SAHA and TSA, showed cytotoxic activity, induction of apoptosis, and cell cycle arrest as reported in other malignancies; (ii) HDACi were able to alter CSC phenotypes: decrease CD44 and ABCG2 expression, suppress sphere formation, decrease stemness gene expression, and inhibit EMT; and (iii) HDACi had synergistic effects in combination with chemotherapeutic agents.
In general, HDACi induce hyperacetylation of core histone proteins, resulting in relaxation of the chromatin structure, promoting access to transcription factors, and changing gene expression. In addition to such epigenetic effects, HDACi also induces the reversible acetylation of non-histone proteins, including transcription factors, DNA repair enzyme, and signal transduction mediators. So far, it has been demonstrated that as many as 7–10% of expressed genes are altered by treatment with HDACi using cDNA arrays.[31-33] Although the multiple mechanisms mediated by HDACi have not been completely elucidated, it has been shown that the main anti-cancer effects of HDACi are the induction of apoptosis and differentiation, and cell cycle arrest in G1 or G2/M. In our experiments, two HDACi, SAHA and TSA, showed cytotoxic activity, induction of apoptosis, and cell cycle arrest to SCCHN cell lines, confirming that these HDACi also act on SCCHN cells, as demonstrated in other tumor cells. With regard to the cell cycle arrest induced by HDACi in tumor cells, Richon et al. have demonstrated that low concentrations of HDACi predominantly induce G1 arrest, while high concentrations induce both G1 and G2/M arrest. Under our experimental conditions, HSC-2 treated with HDACi showed predominantly G2/M arrest rather than G1 arrest. On the other hand, KUMA-1 treated with TSA showed similar findings; however, treatment with SAHA induced cell cycle arrest in both G1 and G2/M. Besides the concentrations, various factors, such as exposure time, types of HDACi, and types of cell lines, affect cell cycle effects.
In addition to these anti-cancer effects, HDACi have been demonstrated to have multiple biological activities, including the induction of differentiation, inhibition of angiogenesis, and immunomodulatory activity. Interestingly, HDACi suppressed various CSC properties of CD44+-enriched SCCHN cell lines. Self-renewal is one of the well-known properties of CSCs and is regulated by interactions among a variety of stem cell regulators. Histone deacetylase inhibitors suppressed the expression of Bmi1, Notch, Nanog, and Oct-4, which are essential for stem cell self-renewal. Downregulation of each gene at the concentration used in this study was relatively modest; however, the suppression of multiple genes may lead to a strong inhibiting effect on self-renewal capacity. Indeed, tumor sphere formations reflecting self-renewal capacity were effectively inhibited by treatment with HDACi.
In this study, inhibition of CD44 and ABCG2 expression on tumor cells by treatment with HDACi are of particular importance. It has been shown that these molecules are not only cell surface markers of CSCs, but are also responsible for maintenance of the CSC phenotype. CD44 is a multifunctional transmembrane glycoprotein receptor that binds to hyaluronan. It has been shown that CD44-hyaluronan interaction induces signaling events, which promote tumor cell proliferation, migration, invasion, angiogenesis, and metastases.[35, 36] In fact, Bourguignon et al. have demonstrated that hyaluroan-CD44 interaction activates cancer stem markers, Nanog, Oct-4, and Sox2, and leads to multidrug transporter, ABCB1 (MDR1) gene expression. ABCB1 is related to the efflux of chemotherapeutic agents, including doxorubicin and paclitaxel, and its overexpression induces the acquisition of chemotherapy resistance. Meanwhile, ABCG2 is another well-known multidrug resistant protein, which effluxes not only chemotherapeutic agents, but also Hoechst 33342. In a variety of cases, the cells expressing ABCG2 appear as side population (SP) cells in a Hoechst-based flow cytometry profile, and SP cells have also been identified as stem/progenitor cells.[38, 39] The SP cells obtained from freshly isolated tumor cells and tumor cell lines enrich the cell with cancer stem cell phenotypes as compared with the corresponding non-SP cells. Thus, the inhibition of CD44 and/or ABCG2 expression on tumor cells is closely related to loss of stem cell properties and chemotherapy resistance. Contrary to our findings, To et al. have demonstrated that ABCG2 expression in some tumor cell lines was upregulated after treatment with another HDACi, depsipeptide, where ABCG2 promoter has a reproducible pattern of histone modification due to HDACi. Further studies are required to elucidate these differences.
Evidence has accumulated indicating that induction of EMT, which promotes invasion and metastases, plays an important role in the acquisition of CSC properties in a variety of cancers.[41-44] In SCCHN, Chen et al. have shown that ALDH+ putative CSCs enriched from SCCHN cell lines using an anchorage-independent culture technique increased EMT-related marker expression. In our studies, upon treatment with HDACi, E-cadherin, which acts as a master regulator of EMT, was upregulated, and the mRNA expression level of TGF-β, which is known to be an important inducer of EMT was downregulated, indicating that HDACi acts to inhibit EMT characteristics in CD44 + SCCHN cell lines. So far, SAHA has been shown to have reverted the mesenchymal phenotype by inducing E-cadherin and downregulating vimentin in SCCHN. Similarly, in renal epithelial cells, TSA has been able to prevent TGF-β1-induced ENT. Conversely, several studies have shown that in prostate cancer cells and endometrial adenocarcinoma cells, HDACi could induce EMT.[48, 49] The EMT phenomenon is regulated by a complex network consisting of a variety of EMT-related molecules, including epithelial and mesenchymal-related factors, and therefore such controversies may occur depending on the type of markers tested. Additionally, culture and treatment conditions and the type of cell line in each experiment vary, and therefore the clinical utility of HDACi should proceed carefully. Indeed, in a phase I trial of SAHA, a partial response was observed in a patient with laryngeal carcinoma; however, in a phase II trial, SAHA could not show any efficacy as defined by tumor responses.
Taken together, our studies indicated at least cell cycle arrest, downregulation of stemness-related genes, decreased CD44 and ABCG2 expression, and inhibition of EMT, and thereby HDACi would lead to decreased drug resistance. Based on such pleiotropic effects, two HDACi combined with cisplatin or docetaxel would show synergistic effects. Our data suggest the further possible use of HDACi in combination with chemotherapy, radiotherapy, and/or immunotherapy. Currently, to overcome the treatment resistance of CSCs, various approaches, such as immunotherapy, molecularly targeted drugs, and gene therapy are being developed. Better understanding of epigenetic mechanisms as well as interplay among epigenetic factors may provide new insights for developing new pharmacological strategies.
This work was supported in part by grants-in-aid (23592523 to KC, 24791820 to KS, 24791744 to MS, 25861525 to MT, and 25462626 to KT) from the Ministry of Education, Culture, Sport, Science and Technology of Japan.