Background
Glioblastoma (GBM), grade IV astrocytoma, is the most aggressive primary brain tumor in adults. Despite current advances in surgery, radiotherapy, and chemotherapy, GBM remains incurable and claims roughly 17,000 lives each year in America [
1]. Like other solid tumors, GBMs are an accumulation of heterogeneous cell populations comprised of a select few cancer stem cells (CSCs) that are able to initiate and sustain tumor growth [
2]. CSCs are multipotent, able to differentiate into multiple cell types to make up the tumor bulk [
3], and display signature characteristics of self-renewal and unlimited growth potential. Due to upregulated multi-drug transporters, altered anti-apoptotic machinery, and enhanced DNA damage response, CSCs are relatively resistant to most chemotherapy and radiotherapy [
4], therefore substantially contribute to tumor metastasis and recurrence. GBM stem-like cells (GSCs) grow in vitro as non-adherent clonal multicellular neurospheres and efficiently initiate tumor xenografts that recapitulate the genetic and histopathological features of the original neoplasm from which they were derived [
5]. Therefore, targeting GSCs or their tumor-initiating capacity will provide mechanistic insights that may more efficaciously treat this deadly cancer.
Various approaches have been tested to induce GSC differentiation or cell death to reduce their tumor-initiating potential, such as treatment with bone morphogenic protein (BMP) [
6], histone deacetylase inhibitors [
7], retinoic acid [
8], and overexpression of transcription factors [
9]. The Krüppel-like factors (KLFs) consists of 17 evolutionarily conserved zinc finger transcription factors with diverse regulatory functions [
10]. By binding to GC-GT rich regions in promoters/enhancers, KLFs regulate a variety of cellular functions such as proliferation, cell survival and differentiation [
11,
12]. It has been reported that KLF family members act as tumor suppressors and/or oncogenes under distinct cellular context [
13,
14]. Krüppel-like factor 9 (KLF9), also known as basic transcription element-binding protein 1 (BTE-B1), has been found downregulated in a number of cancers including endometrial carcinoma and colorectal cancer [
15,
16]. Our research group previously showed that expression of KLF9 in GBM was low [
9] and found it upregulated in response to diverse differentiation signals [
7,
8]. Moreover, KLF9 induces GSC differentiation and inhibits GSC self-renewal and xenograft growth in vivo [
9,
17].
DNA methylation and histone modifications are epigenetic mechanisms that contribute to the pathogenesis of cancer, including GBM [
18]. Enzymatic modifications of histone proteins have being exploited for therapeutic cancer targeting. Histone deacetylase (HDAC) inhibitors consist of a group of agents that block histone de-acetylation and neutralize positively charged lysine residues on histone tails, thereby altering chromatin structure and gene transcription [
19]. HDAC inhibitors have been reported to kill a variety of tumor cells through diverse mechanisms [
20,
21], including disruption of co-repressor complexes, induction of oxidative injury, upregulation of death receptor and ligand expression, generation of lipid second messengers, interference with chaperone protein function, modulation of NFκB activity, mitotic catastrophe, and interference with DNA repair. Thus, HDAC inhibitors reduce tumor growth mainly by inducing cell growth arrest and cell death (i.e. apoptosis and autophagy), to a less extent by modulating tumor cell migration and tumor-microenvironment interactions [
22]. Several HDAC inhibitors, such as vorinostat (SAHA) and panobinostat (LBH589), have been approved by the U.S. Food and Drug Administration (FDA) for the treatment of several malignancies. LBH589 is a non-selective histone deacetylase inhibitor that has been approved for the treatment of various cancers including multiple myeloma [
23]. In our own laboratory, we have previously shown that HDAC inhibitors are potent differentiation agents in GSCs. In the current study, GSCs were used to examine the function of KLF9 in chemotherapy sensitization to HDAC inhibitors. We found that KLF9 induction synergizes with HDAC inhibitors to induce cell death in GSCs through a mechanism that involves both apoptosis and necroptosis.
Discussion
In the present work, we found that combining two differentiation regimens, namely the transcription factor KLF9 and HDAC inhibitors, triggered synergistic cell death in GSCs via a mechanism involving both apoptosis and necroptosis (Fig.
5d). Given the resistance of GSCs to conventional radiation and chemotherapy, the tumor-cell killing effect of combined differentiation therapies will likely open a new opportunity to treat these refractory GSCs.
The tet-on system with Dox-inducible KLF9 expression was used in this study as it provided an internal controlled model for KLF9 functional analysis [
9,
17]. In other words, if we had investigated KLF9 function with stable cell lines that constitutively express KLF9, we would have compared cellular responses to HDAC inhibitors in two different cell lines, whose behavior may have differed during the long process of establishing cell lines from single clones, thus making the interpretation of the results confounded. In our study, we used several methods to quantify cell number and cell death including MTS assays, trypan blue staining, annexin-V/PI staining, and cell cycle analysis. All these assays demonstrated enhanced cell death induced by KLF9 expression + LBH589. Because of the different quantification methods, we acknowledge that there may be some variation on the percent of cell death induced by each agent alone or combined. In part, this may be because some methods measure both cell death and proliferation while other methods exclusively quantify cell death. For example, MTS assays measure both cell proliferation and cell death, whereas trypan blue and annexin-v staining measure the percentage of dead cells under treatment conditions. In addition, we used cell cycle analysis to measure percentage of cells at the sub G1/G0 phase as an indicator for cell death. We noticed that even though LBH589 treatment alone induced significant cell cycle arrest, these was no increase in the sub G1/G0 portion, confirming LBH589’s effect on cell proliferation without inducing cell death. The delayed S-phase may be due to increased stem cell differentiation under LBH589 treatment, as previously reported by our group [
7]. In our previous studies, we found that HDAC inhibitors TSA and MS-275 induced differentiation of GSCs with no effect on cell death [
7]. This is consistent with our current work that LBH589 induced cell cycle arrest in GSCs as evidenced by cell cycle analysis and increase expression in p21 and p27 on Western blot analysis.
Different types of cell death have been investigated in our study: we characterized apoptosis, autophagy, mitotic catastrophe, and necroptosis following the treatment of LBH589 in KLF9 expressing cells. We found that the enhanced cell death induced by HDAC inhibitors and forced KLF9 expression was a mixture of apoptosis and necroptosis. Several scenarios may explain the synergistic cell death in the context of dual KLF9 expression and LBH589 administration. First, compared to undifferentiated stem-like cells that are more resistant to chemotherapeutic drugs, differentiated cells induced by KLF9 expression may be more vulnerable to anti-cancer drugs, such as HDAC inhibitors. Second, our previous RNA-seq and ChIP-seq data indicated that KLF9 regulated the gene expression of both pro-apoptotic and anti-apoptotic proteins [
17]. This was confirmed in our current study that KLF9 dramatically upregulated pro-apoptotic proteins including Bak, Bik, Bax, Bid and Noxa. We also examined the expression of a panel of apoptosis regulators in KLF9 expressing cells in the presence of LBH589, and found that there was a dramatic downregulation of anti-apoptotic proteins XIAP and survivin. The synergistic effect of KLF9 expression and LBH589 elicits an enhanced cell death response in GSCs. When we examined our KLF9 ChIP-seq gene list, we found that KLF9 directly binds to the promoter regions of Bak, Bik, Bax, Bid and Noxa, [
17] but not to the promoter regions of XIAP and survivin. Therefore, we conclude that in the presence of decreased anti-apoptotic proteins, KLF9 directly upregulated pro-apoptotic gene expression to enhance cell death.
Perhaps the most interesting finding is the involvement of programmed necrosis (necroptosis) induced by KLF9 expression + LBH589 in the presence of apoptosis inhibitor z-vad. The fact that pan-caspase inhibitor z-vad did not proportionally protect GSC death induced by Dox + LBH589 suggests that [
1] Dox + LBH589 induced caspase-independent cell death, and/or [
2] Dox + LBH589 + z-vad induced a new form of cell death. The combination of z-vad with necroptosis inhibitors significantly rescued GSC death, suggesting that apoptotis and necroptosis may simultaneously occur in KLF9 expressing cells when treated with LBH589. Upon examining the expression of the necroptosis effector receptor-interacting protein (RIP) [
46], we found no significant change in RIP1/3 following drug treatment (data not shown). The exact mechanism and regulatory network by which KLF9 and HDAC inhibitors regulate the expression of necroptotic effectors and activate necroptosis in our system is currently unknown. Further analyses in the future will help to identify novel targets for anti-tumor treatments.
Moreover, the manipulation of nuclear proteins for cancer treatment is an exciting area of research. Compared with small molecular inhibitors that target receptors and/or kinases on cell membranes or in the cytoplasm that inevitably generate escape mechanisms, employing transcription factors such as KLF9 to target cancer stem cells would be beneficial because these proteins tightly control gene expression upstream of signaling transduction pathways, thereby preventing cells from developing compensatory mechanisms. With advanced gene therapy technology, our prospective in vivo testing of the synergistic cell death via KLF9 and HDAC inhibitors will help in developing new anti-tumor strategies for GBM patients.
Conclusions
Tumors are heterogeneous and comprised of a small group of tumor-initiating or cancer stem cells (CSCs). CSCs are resistant to current chemotherapy and radiotherapy, leading to metastasis and relapse of cancers, therefore significantly affect cancer therapy. In this study, we investigated the combined treatment of epigenetic modulators and forced expression of the transcription in inducing cell death in glioblastoma stem cells (GSC). We found that the combination of histone deacetylase (HDAC) inhibitors and expression of krÜppel-like factor 9 (KLF9) synergistically promote GSC death through a mechanism involving both apoptosis and necroptosis. Our findings are expected to benefit the development of effective anti-tumor strategies to treat malignant brain tumors.
Acknowledgements
We are grateful to Prof. Angelo Vescovi (San Giovanni Rotondo, Italy) for providing us HSR-GBM1A and HSR-GBM1B neurosphere cell lines isolated from patient glioma samples.