RGFP966, a Histone deacetylase 3 inhibitor, promotes glioma stem cell dif‐ ferentiation by blocking TGF-β signaling via SMAD7

Abstract

Glioma stem cells (GSC) are significant contributors to drug resistance and tumor recurrence. A genetic screen using shRNAs targeting chromatin regulators in a GSC model identified HDAC3 as a major negative regulator of GSC differentiation. Inhibition of HDAC3, either pharmacologically or through siRNA, induced GSC differentiation into astrocytes. Consequently, HDAC3 inhibition also led to a substantial reduction in the tumor-promoting and self-renewal capabilities of GSCs. These effects were strongly associated with increased acetylation of SMAD7, which protected it from ubiquitination. SMAD7 inhibits a TGF-β signaling pathway crucial for maintaining stemness. These findings suggest that HDAC3 represents a promising therapeutic target for anti-glioma therapy.

1. Introduction

Gliomas are the most frequently diagnosed primary tumors of the central nervous system. Glioblastoma multiforme (GBM), the most aggressive type of glioma, is associated with high morbidity and mortality despite advances in diagnostic methods, surgery, and therapies such as radiation and temozolomide (TMZ) chemotherapy. Stem cells play a critical role in tumor recurrence and drug resistance. Glioma stem cells (GSCs), a subpopulation of cancer cells, are responsible for tumor initiation, relapse, post-therapeutic recurrence, and resistance to chemotherapy or radiotherapy. Therefore, novel therapeutic strategies targeting the resistance mechanisms that impede effective treatment are urgently needed.

GSCs share certain characteristics with neural stem cells (NSCs) but exhibit partial differentiation while retaining high self-renewal and tumor initiation capacities. Studies have identified abnormal epigenetic changes in GSCs, including histone modification and DNA methylation. Consequently, targeting these abnormal epigenetic alterations in GSCs may represent a viable approach for treating malignant glioma.

To investigate the role of epigenetic modifications contributing to GSC survival, cultured primary GSCs from hGFAP-Cre+ p53L/L PtenL/+ mice, a model that develops spontaneous gliomas, were utilized in this study.

Chromatin modifiers regulate the expression of genes that promote cell survival and stemness. Therefore, these GSCs were transduced with a set of shRNAs capable of targeting epigenetic regulators. These experiments identified HDAC3 as a key factor in maintaining stemness, as its inhibition led to a significant decrease in the number of GSCs and promoted GSC differentiation.

Histone deacetylases (HDACs) and histone acetyltransferases (HATs) are crucial regulators in various biological processes, including the cell cycle and apoptosis, through epigenetic modulation of gene expression involved in these processes. The imbalance between HDACs and HATs is strongly associated with tumorigenesis. Overexpression or increased activity of HDACs reduces the transcription of tumor suppressor genes, such as CDKN1A and CDKN2B, and promotes the expression of oncogenes like c-MYC, RAS, and AKT. Certain non-histone proteins, such as HSP90AA1, regulated by reversible acetylation, are also positively correlated with tumorigenesis.

Abnormal expression of HDAC3 is frequently observed in various human cancers, including colorectal cancer, multiple myeloma, rhabdomyosarcoma, and malignant glioma. Inhibition of HDAC3 expression has been shown to induce differentiation in acute promyelocytic leukemia cells and cause cell cycle arrest, suggesting this deacetylase as a potential target in anti-cancer therapy. However, the effectiveness of HDAC3 inhibitors in blocking glioma stem cells remains unclear. More importantly, the mechanisms by which HDAC3 blockade inhibits GSCs are not fully understood.

In this study, we report that HDAC3 inhibition suppresses GSC proliferation and induces GSC differentiation both in vivo and in vitro. The TGF-β signaling pathway has been shown to maintain GSC stemness. Here, we demonstrate that these cellular phenotypes were strongly associated with the activation of SMAD7 by enhancing its acetylation, which protects it from ubiquitination, thereby disrupting TGF-β-dependent stemness and promoting differentiation of cells into astrocytes. In summary, targeting HDAC3 appears to be an effective strategy to inhibit glioma stem cells.

2. Materials and Methods

2.1 Cell Culture and human tissue specimens

The human GSC line TS541 was kindly provided by Dr. Cameron Brenner at the Holland and Mellinghoff laboratories (MSKCC). Murine glioma stem cells were isolated from hGFAP-Cre+ p53L/L PtenL/+ mice as previously described. The GSCs were derived from three separate transgenic mice that developed glioma. Murine GSCs were cultured in Neurobasal media, and human GSCs were maintained in NeuroCult NS-A proliferation media (human; Stem Cell Technologies) supplemented with growth factors (epidermal growth factor (EGF; 20ng/mL, Peprotech, Rocky Hill, RI, USA) and basic fibroblast growth factor (BFGF; 10ng/mL, Peprotech, Rocky Hill, RI, USA)) or differentiation medium (with 1% FBS addition but without growth factors). Frozen tumor specimens were obtained from the First Bethune Hospital of Jilin University. All tumor specimens used in this study were collected under an IRB-approved protocol with informed consent of the subjects.

2.2 Global screening with a shRNA library

A lentiviral shRNA library targeting 243 chromatin-regulating mouse genes, encompassing most ‘writers’, ‘readers’, and ‘erasers’ (3-6 shRNAs were designed per gene), was previously described. Three different hGFAP-Cre+ p53L/L PtenL/+ murine GSCs (isolated from three separate mice) were used. The percentage of GFP-positive cells in each population was determined to estimate comparable transduction efficiency. The cells were switched to differentiation culture conditions (Neurobasal media without growth factors) 48 hours post-infection. Under these conditions, the cells do not proliferate. After culturing for two weeks in the basal medium, cell growth was scored based on the size of the GFP+ colonies relative to the control shRNA (EV) using a fluorescence microscope. Selected shRNAs (scored in the range of 1-3) were subcloned into the MLP-Luc vector for transfection. A parallel control shRNA was used to eliminate gene transduction-associated artifacts.

2.3 Transfection

Lentiviral plasmids were obtained from the Cold Spring Harbor Laboratory. To generate virus particles, 293T cells were transfected with lentiviral vectors carrying shRNAs as previously described. Lentiviral particles from the culture media were concentrated after centrifugation with a Beckman LE-80K and used to infect GSCs with polybrene (1:1000, Sigma, St. Louis, MO, USA). Puromycin (1:1000, Millipore, Billerica, MA, USA) was used to select stably transfected cells. Smad7 siRNA (Genepharma, Suzhou, China) and Smad7 cDNA (MedChemExpress, Monmouth Junction, NJ, USA) were transfected using Lipofectamine reagent (Thermo Fisher Scientific, Carlsbad, CA, USA) according to the manufacturer’s recommended protocols. Before transfection, cells were cultured in 96-well or 6-well plates until they reached 80% confluency. Cells were collected 48-72 hours after transfection.

2.4 Drug treatment

GSCs were treated with different concentrations (0.05-20μM) of the HDAC3 inhibitor, RGFP966 (Cat# S7229, Sellect, Shanghai, China), prepared in DMSO. In some experiments, cells were treated with 100nM Cycloheximide (CHX), a protein synthesis inhibitor (Cat# HY-12320, MedChemExpress, Monmouth Junction, NJ, USA), 1 hour before the addition of RGFP966 (2.5μM). GSCs were treated with 0.2μM recombinant mouse TGF‑β1 (Cat# 7666-M, R&D, MI, USA) for 24 hours alone or before treatment with RGFP966 (2.5μM).

2.5 HDAC3 Inhibition Assays

HDAC3 activity was assessed using a HDAC Assays kit (Cat # 56200, Active Motif, Carlsbad, CA, USA) following the manufacturer’s protocol. GSCs (5×106) were collected after treatment with RGFP966 (0.5 to 5μM), and nuclear extracts were prepared. Nuclear extracts were incubated with the indicated substrates, and the products were measured using a fluorescence plate reader with excitation at 360 nm and emission detection at 460 nm.

2.6 Cell viability assays

To evaluate the effect of RGFP966 on GSCs, cells (1×104) were plated in 96-well plates and treated with various concentrations of RGFP966. Cell viability was determined using a Cell Counting Kit-8 (MedChemExpress, Monmouth Junction, NJ, USA) as recommended by the manufacturer.

2.7 Soft agar assay for colony formation

Soft agar assays were performed to assess the oncogenic potential of cells in vitro. The base layer of the medium contained 1% agar, and the top layer contained 0.6% agar. A total of 2.5×103 cells were seeded in the upper layer. After 15 days, colonies were stained with crystal violet and counted. In some experiments, cells were treated with RGFP966 at concentrations of 0.0, 1.0, and 2.5μM. In other experiments, cells were transfected with lentiviral vectors carrying either an empty vector (EV) or shRNAs targeting Hdac3. Each condition was tested in at least three replicates, and each experiment was repeated three times to ensure the consistency of the observations.

2.8 Self-renewal and limiting dilution assay

Self-renewal capacity was evaluated by seeding a cell suspension (1×102 cells/ml) in 6-well plates and culturing them in NSC proliferation media. The number of neurospheres formed was quantified after 10 days. Limiting dilution assays were conducted as previously described. GSCs were dissociated and plated in 96-well plates at a volume of 0.2 ml per well using Neurobasal media. Seeding densities ranged from 100 cells/well to 1 cell/well. Cultures were fed with 0.025 ml of Neurobasal media every 2 days until day 10. The percentage of wells without spheres for each cell plating density was then calculated and plotted against the number of cells per well. Regression lines were plotted, and x-intercept values were calculated, representing the number of cells required to form at least one tumor sphere in every well.

2.9 Immunostaining

Cells were fixed with 4% paraformaldehyde, permeabilized with 0.3% Triton X-100 in PBS, and then blocked with 1% BSA. Subsequently, they were incubated with the indicated primary antibodies overnight at 4 °C, followed by incubation with corresponding Alexa Fluor-conjugated secondary antibodies (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) for 30 minutes at room temperature. The cells were then mounted in a glycerinum containing DAPI. Slides were observed, and images were captured using a BX53 fluorescence microscope (Olympus, Tokyo, Japan). The primary antibodies used were against Nestin, GFAP, S-100β, NeuN, Olig2, and Smad4. Normal IgG was used as a control to ensure staining specificity.

2.10 Immunohistochemistry (IHC)

Mouse xenograft tissues treated with RGFP966 or vehicle were harvested, rinsed in PBS, and fixed in 4% paraformaldehyde overnight at 4 °C. Five-micron thick sections were mounted on poly-D-lysine-coated slides, which were then incubated with primary antibodies. A secondary antibody conjugated to HRP was applied and detected using DAB. The primary antibodies used were against HDAC3, PCNA, TGF-β RI, SOX2, GFAP, and S-100β.

2.11 Immunoprecipitation and Western blot (WB)

Cell culture samples were sonicated and suspended in ice-cold lysis buffer containing PMSF and a phosphatase inhibitor cocktail (Beyotime, Shanghai, China). After centrifugation of the lysate, the supernatant was collected, and Smad7 was immunoprecipitated using a mouse anti-Smad7 antibody. Protein-bound antibodies were pulled down using protein A or protein G agarose (Beyotime, Shanghai, China). Protein samples were then separated by 10% SDS-PAGE and transferred onto PVDF membranes. The membranes were incubated with the appropriate primary antibodies, washed, and then incubated with corresponding secondary antibodies. After washing, the blots were visualized using ECL Plus reagents. Protein expression levels were quantified after normalization with β-Actin or total protein. The primary antibodies used were against HDAC3, TGF-β RI, SOX2, Nestin, GFAP, S-100β, NeuN, Olig2, SMAD2/3, p-SMAD2/3, actin, and Smad7. For assessing acetylation or ubiquitination of Smad7, immunoprecipitated Smad7 was subjected to Western blot analysis using anti-acetyl-lysine or anti-ubiquitin antibodies, respectively.

2.12 Quantitative real-time PCR (qPCR)

qPCR was performed using SYBR Premix Ex TaqTM II (Tli RNaseH Plus) (Cat# RR820A, TaKaRa, Shiga, Japan) and a CFX96 Realtime PCR machine (Bio-Rad, USA). The threshold cycle (Ct) values for each gene were normalized to those of β-actin.

2.13 Microarray and data analysis

Total RNA was extracted from cells using TRIzol reagent (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA). cRNA was fragmented and then hybridized to Affymetrix Genome1.0 arrays (Affymetrix). The data were analyzed using the Robust Multichip Analysis (RMA) algorithm with Affymetrix default analysis settings and global scaling for normalization. Values are presented as log2 RMA signal intensity.

2.14 Animals

Five-week-old BALB/c nude male and female mice were purchased from Beijing Vital River Laboratory Animal Technology (Beijing, China). They were subcutaneously inoculated with TS541 cells (5×106 per mouse). Five days post-inoculation, the mice were randomly divided into two treatment groups (n=4/group). They were treated thrice a week with either RGFP966 (25 mg kg-1) or vehicle (5% DMSO, 45% PEG 300) via subcutaneous injection 1 cm away from the tumor, as previously described. After 24 days, the mice were euthanized due to moribundity. All animal experiments were conducted in accordance with US NIH guidelines and were approved by the institutional animal care and use committee of Jilin University.

2.15 Statistical analyses

All statistical analyses were performed using Prism GraphPad v6.00 software. Each experiment was repeated at least three times, and the data are presented as the mean ± standard deviation (SD). Comparisons between two groups were made using an unpaired Student’s t-test. For comparisons involving more than two groups, a one-way analysis of variance (ANOVA) with a Newman–Keuls multiple comparison test was used.

3. Results

3.1 HDAC3 is highly expressed in human brain gliomas.

To evaluate the role of epigenetic genes in GSC differentiation and, ultimately, tumor proliferation, three independent GSC lines were obtained from hGFAP-Cre+ p53L/L PtenL/+ mice, which exhibited high self-renewal, partial differentiation, and strong tumor initiation capacities. The expression of 243 known chromatin regulators in these cells was suppressed using a shRNA library. The aim was to identify critical epigenetic modifiers whose inactivation induces glioma cell differentiation. A screening score standard was established for candidate genes, using an empty shRNA vector and shRpa3.455 as negative and positive controls, respectively. Data from the negative control shRNA (EV) were assigned a score of ‘4’, and data from the positive control (shRpa3.455) were assigned a score of ‘1’. The goal was to identify target genes whose inactivation by cognate shRNAs scored in the range of 1-3 in at least two of the three glioma lines used (CSC2078/1534/1589). Hdac3 was one of the genes identified in this screen. The effects of HDAC3 downregulation on GSC differentiation were then tested in a cell line model, using an empty shRNA vector and shRpa3.455 as positive and negative controls. Cells transfected with shHdac3 displayed the appearance of more differentiated neuroglial-like cells. To assess the relevance of HDAC3 to primary glioma, the TCGA database was queried for HDAC3 levels in glioma patients. These analyses revealed significantly increased HDAC3 expression levels (RPKM) in GBM compared to normal tissue.

HDAC3 expression levels were also examined in different pathological grades of human glioma (n=10 per group). HDAC3 levels were higher in grade I-IV glioma than in the control. Notably, HDAC3 levels in high-grade glioma were significantly higher than those in lower grade, although no significant difference was observed between grade III and grade IV glioma. As controls, the expression of other members of the HDAC family was examined by IHC, but no significant changes were found in these samples. These findings suggest that downregulation of elevated HDAC3 levels in GSCs promotes cell differentiation and presents a therapeutic opportunity.

3.2 Reduction of HDAC3 level and/or activity leads to a suppression of GSC proliferation and self-renewal

To investigate the inhibitory effect of HDAC3 downregulation on GSC proliferation, the HDAC3-specific inhibitor RGFP966 and shRNAs targeting HDAC3 (shHDAC3) were used in human (TS541) and murine (CSC2078) GSCs. Initial HDAC assays demonstrated decreased HDAC3 activity following RGFP966 treatment. RGFP966 inhibited the proliferation and viability of both human (TS541) and murine GSCs (CSC2078) in a dose-dependent manner. To confirm the effects on cell proliferation, cells were labeled with BrdU, and BrdU incorporation was markedly reduced in the presence of RGFP966. The specificity of these effects was examined using two separate shRNAs targeting Hdac3, 1491 and 1037. Both shRNAs strongly suppressed cell viability and growth. To further demonstrate the specificity of growth suppressive effects, cells transduced with a control shRNA and shHdac3s 1491 and 1037 were treated with RGFP966. Depletion of Hdac3 caused significant cell loss compared to the control shRNA. Importantly, RGFP966 promoted cell death in control shRNA expressing cells but did not cause any further cell death beyond that observed with shHdac3 alone. A similar effect of RGFP966 on cell growth was observed in Hdac3-depleted cells, suggesting that RGFP966 effects are primarily mediated through HDAC3. Western blot analysis confirmed that the 1491 and 1037 shRNAs reduced Hdac3 expression by over 70% and 90%, respectively. The effect of RGFP966 on the ability of GSCs to form soft agar colonies (an indicator of oncogenic behavior) and neurospheres (a characteristic of self-renewing stem cells) was also assessed. RGFP966 treatment drastically reduced the formation of soft agar colonies and the number of neurospheres formed by GSCs. Together, these results demonstrate that inhibition or downregulation of HDAC3 suppresses GSC proliferation and self-renewal capacity.

3.3 Inhibition of HDAC3 activity promotes GSCs differentiation

Observation of Hdac3 knockdown significantly influencing GSC morphology suggested that Hdac3 inactivation might promote cell differentiation. To confirm this, and given the central nervous system origin of GSCs, their stem phenotype was analyzed using markers Nestin, S-100β, GFAP, Oligo2, and NeuN. Indeed, the levels of the stem cell marker Nestin were clearly decreased in RGFP966-treated and shHdac3 groups. The effect of HDAC3 inhibition on GSC differentiation by examining the expression of astrocyte markers was also investigated. Astrocyte markers GFAP and S-100β were significantly increased when HDAC3 was inhibited. Notably, GFAP levels decreased marginally at the highest dose of RGFP966 (2.5μM) compared to the low dose, whereas S-100β levels continued to increase under these conditions. This difference might be attributed to GFAP expression in all stages of astrocyte lineage, while S-100β is primarily expressed in the terminal stage, suggesting that high-dose RGFP966 could promote the differentiation of GSCs into mature astrocytes. The oligodendrocyte marker Olig2 expression was lower in the RGFP966-treated cells compared to the control. In the neural stem cell differentiation lineage, oligodendrocyte and astrocyte differentiation from progenitor cells is regulated by different stimulation factors. Thus, reduced oligodendrocyte differentiation might allow more cells to differentiate into the astrocyte lineage. The neuronal marker NeuN was elevated in the treatment group. These IHC observations were also confirmed by qPCR analyses of these markers. Thus, inhibition of HDAC3 suppresses the stemness of GSCs and promotes their differentiation into astrocytes.

3.4 HDAC3 inhibitor suppresses tumor growth and promotes differentiation in vivo

To test the in vivo relevance of the preceding observations, the effects of RGFP966 on the growth of glioma-initiating cells in BALB/c nu/nu mice were examined. RGFP966 strongly decreased tumor growth. Tumors from RGFP966-treated mice exhibited a significant reduction in PCNA, a marker of cell proliferation. These RGFP966-treated tumors also showed high GFAP and S-100β expression, consistent with the differentiation of GSCs into astrocytes. Importantly, RGFP966 at the therapeutic dose did not cause any significant toxicity, as evidenced by the absence of visible damage to the liver, spleen, kidney, lung, and cardiac tissues. Additionally, there was no significant weight loss in these mice during the entire experimental period.

3.5 Inhibition of HDAC3 blocks the TGF-β signaling in GSCs.

To elucidate the molecular mechanism underlying the action of HDAC3 on GSC differentiation, gene expression differences between RGFP966-treated and control groups were profiled and investigated using microarrays. These studies identified approximately 2900 upregulated genes and 2534 downregulated genes following RGFP966 treatment. KEGG pathway enrichment analyses indicated alterations in genes involved in cancer cell proliferation, some differentiation-related pathways, and signaling pathways regulating stem cell pluripotency in RGFP966-treated cells. Among these, the TGF-β signaling pathway was significantly downregulated. To confirm these data, self-renewal capacity was assessed using a limiting dilution assay in the presence of TGF-β1 and RGFP966 treatment. The number of cells required to generate at least one tumor sphere per well was 38.66 ± 4.69 in the control group, 23.99 ± 3.14 in the TGF-β1 group, 128.08 ± 4.42 in the RGFP966 group, and 121.93 ± 5.472 in the RGFP966+TGF-β1 group. Thus, significantly more cells were required for neurosphere formation in the RGFP966 group, and the difference between the RGFP966-treated and RGFP966/TGF-β1 groups was statistically insignificant. qPCR and Western blot analyses of critical elements of the TGF-β signaling pathway were performed. Notably, the levels of secreted TGF-β decreased mildly, and its receptor TGF-β R1 and the downstream gene Sox2 were significantly downregulated in RGFP966-treated cells. Interestingly, the transcript coding for Smad7, an inhibitor of TGF-β signaling, was strongly increased several fold in the presence of RGFP966 or following Hdac3 knockdown by shRNA. Although Smad2/3, the signal transducer for TGF-β, remained unaffected by RGFP966, its phosphorylation was strongly suppressed compared to the controls. The increased Smad7 transcription also correlated with an increase in its protein levels. Loss of TGF-β signaling was also observed in tumors treated with RGFP966 in vivo, as indicated by low TGF-β RI and Sox2 expression levels, suggesting a loss of GSC stemness in vivo. Thus, HDAC3 appears to promote GSC stemness through an upregulation of TGF-β signaling, and RGFP966 interferes with this process.

3.6 Inhibition of HDAC3 leads to a suppression of the TGF-β signaling by enhancing SMAD7 acetylation.

One mechanism by which TGF-β maintains self-renewal in tumor cells is through activation of the TGF-β RI-SOX4-SOX2 axis. Interestingly, Smad7, a negative regulator of TGF-β signaling, levels were induced by the HDAC3 inhibitor in a time-dependent manner. Smad7 is known to be an unstable protein. To determine if the increase in Smad7 levels was due to new protein synthesis or prevention of its degradation, new protein synthesis was blocked in RGFP966-treated cells using cycloheximide. Smad7 levels still increased in the presence of CHX treatment in RGFP966-treated cells, suggesting that RGFP966 augmented the stability of Smad7.

SMAD7 is downregulated through ubiquitination-mediated degradation. Therefore, immunoprecipitation experiments were conducted on cells treated with the HDAC3 inhibitor or transduced with shHdac3. Ubiquitination of Smad7 markedly decreased following the inhibition of Hdac3 with either RGFP966 or shHdac3. Conversely, acetylation of Smad7 increased under such conditions compared to the controls. Quantification of data from different experiments revealed a significant downregulation of ubiquitination and an increase in acetylation of Smad7. Phosphorylation of Smad2/3 leads to their association with the common signaling transducer Smad4, and the resulting complex translocates to the nucleus to regulate diverse biological effects. TGF-β activates this complex accumulation in the nucleus. Therefore, whether Smad4 translocation to the nucleus was inhibited by RGFP966 was investigated. Immunofluorescence staining of cells treated with TGF-β1 and/or RGFP966 was performed. As expected, TGF-β1 stimulated the nuclear translocation of Smad4. However, this translocation was blocked in the presence of RGFP966. Importantly, RGFP966 completely blocked TGF-β1-induced Smad4 translocation.

To further ascertain the role of Smad7 in promoting cell differentiation, Smad7 was overexpressed in CSC2078 cells. Ectopic Smad7 strongly reduced neurosphere formation by the GSCs. Conversely, downregulation of Smad7 levels using specific siRNAs increased neurosphere formation. More importantly, while RGFP966 was able to suppress neurosphere formation in control siRNA-transfected cells, it failed to do so in siSmad7-transfected cells. The expression of cell differentiation markers in these cells was then examined using Western blot and qPCR analyses. The levels of Nestin, p-Smad 2/3, Sox2, and TGF-β RI were decreased, while those of S-100β were increased in the Smad7 transfectants compared to control vector-transfected cells. In Smad7-depleted cells, RGFP966 failed to downregulate Nestin, p-Smad 2/3, Sox2, and TGF-β RI, and increase S-100β expression levels. These findings indicate that Smad7 blocks TGF-β-mediated cell differentiation. Taken together, these results suggest that SMAD7 is a downstream effector of growth suppression following HDAC3 inhibition.

4. Discussion

Abnormal expression of HDACs is strongly associated with the progression of various human cancers. Downregulation of HDAC3 leads to the suppression of tumorigenesis and limits the differentiation of tumor cells into relatively mature glial lineages. However, the mechanism by which the nuclear enzyme HDAC3 maintains GSC stemness has been unclear. This study demonstrates that HDAC3 exerts such actions by downmodulating critical components of the TGF-β signaling pathway. Upon blockade of HDAC3 with a specific inhibitor or a shRNA, TGF-β signaling was blunted. Importantly, the cell growth inhibitory effects of RGFP966 and its suppression of TGF-β signaling pathways were lost upon depletion of HDAC3 from the cells.

The TGF-β pathway has been implicated in glioma initiation and progression by enhancing cell propagation, tumor invasion, angiogenesis, immunodepression, and maintenance of GSC stemness. Upon HDAC3 inhibition, GSCs were able to differentiate into astrocytes. Although the downregulation of TGF-β could promote such cell differentiation, no significant changes in TGF-β levels were observed. One mechanism by which pluripotency is maintained involves the activation of TGF-β RI phosphorylation to directly induce the expression of SOX4, which promotes SOX2 expression by binding to the SOX2 enhancer. Indeed, inhibition of HDAC3 decreased p-SMAD 2/3, TGF-β RI, and SOX2 expression, providing evidence for a critical role of the TGF-β signaling pathway in maintaining GSC stemness under the control of HDAC3.

HDAC3 inhibition stabilizes the SMAD7 protein, a negative regulator of TGF-β signaling. The biological significance of SMAD7 levels in GBM was further supported by the observation that its levels negatively correlated with GBM progression. SMAD7 was previously shown to be deacetylated by HDACs, favoring its ubiquitination and degradation by SMURF1. The histone acetyltransferase p300 acetylates SMAD7, leading to its stabilization. Acetylation of SMAD7 prevents its ubiquitin-mediated degradation, possibly due to competition between acetylation and ubiquitination for the same lysine target in the SMAD7 protein.

SMAD7 is protected from ubiquitin-mediated degradation due to an increase in its acetylation. Therefore, HDAC3 inhibition potentially enhances anti-glioma therapy when combined with other agents.

The SMAD family of proteins is classified into three major subtypes based on their functions: receptor-regulated Smads (R-Smads), common partner Smads (Co-Smads), and inhibitory Smads (I-Smads). Trimers of two receptor-regulated SMADs and one co-SMAD act as transcription factors and regulate the expression of certain genes. The I-Smads disrupt TGF-β signaling through several mechanisms, such as intercepting the association of R-Smads with the receptor or Co-Smads, downregulating the receptor, and altering nuclear gene transcription. As mentioned earlier, SMAD7, an I-SMAD, negatively regulates TGF-β signaling by sequestering R-Smads. Indeed, RGFP966 blocked TGF-β1-induced SMAD4 nuclear localization. SMAD7 antagonizes R-SMAD phosphorylation, binding to R-SMADs and Co-SMADs, and binding to E3 ubiquitin ligases SMURFs, leading to the degradation of the activated type I receptor ALK5/TβRI. In actively signaling cells, TGF-β promotes the ubiquitination of SMAD7 using the ubiquitin ligase SMURF1. This study demonstrates that overexpression of SMAD7 alone negatively regulates self-renewal capacity and impedes cellular TGF-β RI, p-SMAD 2/3, and SOX2 levels while increasing astrocytic differentiation. SMAD7 levels were elevated primarily by increasing its stability in the presence of the HDAC3 inhibitor. Some studies have reported that acetylation of SMAD7 renders it resistant to the SMURF1-induced ubiquitination degradation pathway. Thus, lysine acetylation and ubiquitination of SMAD7 appear to be mutually exclusive and exert diametrically opposite effects on the stemness of the cells. In addition, these studies also indicated that part of the HDAC3 inhibition leads to an increase in the histone 3 lysine 27 acetylation mark (H3K27ac), which potentially activates SMAD7 transcription. This could lead to an additional increase in SMAD7 mRNA expression. Thus, increased mRNA expression and decreased degradation of SMAD7 could contribute to the overall inhibitory effects of RGFP966. In summary, this study reports a previously unrecognized mechanism by which HDAC3 promotes the TGF-β inhibitor pathway via SMAD7 downregulation, thereby sustaining GSC stemness.