Nrf2-miR-129-3p-mTOR Axis Controls an miRNA Regulatory Network Involved in HDACi-Induced Autophagy

Abstract

Histone deacetylase inhibitors (HDACis) are the recommended treatment for many solid tumors; however, resistance is a major clinical obstacle for their efficacy. High levels of the transcription factor nuclear factor erythroid 2 like-2 (Nrf2) in cancer cells suggest a vital role in chemoresistance, and regulation of autophagy is one mechanism by which Nrf2 mediates chemoresistance. Although the molecular mechanisms underlying this activity are unclear, understanding them may ultimately improve therapeutic outcomes following HDACi treatment. In this study, we found that HDACi treatment increased Nrf2 mRNA and protein levels and enhanced Nrf2 transcriptional activity. Conversely, Nrf2 knockdown or inhibition blocked HDACi-induced autophagy. In addition, a microRNA (miRNA) array identified upregulation of miR-129-3p in response to Nrf2 overexpression. Chromatin immunoprecipitation assays confirmed miR-129-3p to be a direct Nrf2 target. RepTar and RNAhybrid databases indicated mammalian target of rapamycin (mTOR) as a potential miR-129-3p target, which we experimentally confirmed. Finally, Nrf2 inhibition or miR-129-3p in combination with HDACis increased cell death in vitro and in vivo. Collectively, these results demonstrated that Nrf2 regulates mTOR during HDACi-induced autophagy through miRNA-129-3p and inhibition of this pathway could enhance HDACi-mediated cell death.

Histone deacetylase inhibitors (HDACis) are approved for clinical use in many cancers; however, resistance to HDACis is a major problem in clinical oncology that limits their efficacy. Sun et al. show that inhibition of the Nrf2-miR-129-3p-mTOR axis could enhance HDACi-mediated cell death.

Introduction

Histone deacetylase inhibitors (HDACis) are approved for clinical use in many cancers,1, 2 including Trichostatin A (TSA), suberoylanilide hydroxamic acid (SAHA), and romidepsin (depsipeptide).³ However, resistance to HDACis is a major problem in clinical oncology that limits their efficacy.³ Thus, a better understanding of resistance mechanisms to HDACis is urgently required.

Autophagy, an evolutionarily conserved self-digestive process that maintains cellular homeostasis in response to stress conditions,⁴ is important for cell survival during HDACi-based chemotherapy.⁵ As HDACis block histone deacetylases, they primarily regulate transcription factors (TFs),6, 7 which can cause cell differentiation, trans-differentiation, and even autophagy.8, 9 As mechanisms by which HDACis promote autophagy through TFs are unclear,¹⁰ we investigated HDACi-activated TFs capable of promoting autophagy.

Nuclear factor erythroid 2 like-2 (Nrf2) is a key TF that protects cells from oxidative and xenobiotic stresses via the regulation of a wide array of cytoprotective genes, such as heme oxygenase-1 (HO-1) and NAD(P)H dehydrogenase quinone 1 (NQO1).11, 12 Aberrant Nrf2 activation is associated with chemoresistance,12, 13 and Nrf2 knockdown resulted in increased sensitivity of cancer cells to antitumor agents.¹⁴ Furthermore, cancer patients with low Nrf2 levels showed good responses to chemotherapy.¹⁵ Nrf2-activated autophagy has been identified as an important mechanism underlying chemoresistance.16, 17 HDACis activate Nrf2, which serves as a major protective mechanism in many diseases. Yet, whether Nrf2 plays an important role in HDACi-mediated autophagy remains unclear.

MicroRNAs (miRNAs) regulate gene expression at a post-transcriptional level, and they can be targets of or contributors to epigenetic dysregulation.18, 19 Dysregulated miRNA expression affects the responses of cancer cells to chemotherapy.20, 21 Moreover, several basic research studies have suggested that the targeting of miRNAs that regulate autophagy is a promising therapeutic strategy.²² Nrf2 can bind to the promoters of miRNAs, thereby regulating fundamental biological processes such as autophagy.23, 24

In this study, we investigated the role of Nrf2 in the activation of miRNAs after HDACi treatment. Our results demonstrated that Nrf2 regulates HDACi-induced autophagy. Mechanistically, Nrf2 induces the expression of miR-129-3p, which targets mammalian target of rapamycin (mTOR), a major autophagy-associated protein. Furthermore, our data showed that inhibition of the Nrf2-miR-129-3p-mTOR axis enhanced the cytotoxic effects of HDACis in vitro and in vivo. Thus, we concluded that the Nrf2-miR-129-3p-mTOR axis controls a novel miRNA regulatory network that regulates HDACi-induced autophagy.

Results

Nrf2 Transcriptional Activity Is Implicated in HDACi-Induced Autophagy

Nrf2 was previously reported to be activated by HDACi, and Nrf2 signaling may regulate HDACi resistance by activating autophagy. However, the underlying mechanism of this chemoresistance is unclear.²⁵ Here we treated Huh7 and MGC80-3 cells with HDACis, and we investigated the effects on Nrf2 protein and mRNA levels (Figures 1A and 1B). Consistent with previous reports, HDACi treatment decreased Keap1 levels in Huh7 and MGC80-3 cell lines, resulting in Nrf2 activation (Figure 1A).¹² We found that HDACi strongly activated both Nrf2 and autophagy-related proteins. As an activator of oxidative stress, HDACi treatment caused Nrf2 to detach from Keap1. As expected, increased Nrf2 levels were blocked by actinomycin D, a commonly used inhibitor of transcription,²⁶ demonstrating that this was a direct transcriptional response (Figure 1C).

Figure 1.

Transcriptional Activity of Nrf2 Increases during Histone Deacetylase Inhibitor-Induced Autophagy

(A) Cells were treated with HDACis (Trichostatin A [TSA]: 0, 0.5, 1, and 2 μM; or suberoylanilide hydroxamic acid [SAHA]: 0, 0.5, 1, and 2 μM) for 24 h and evaluated by western blotting. The results showed that HDACi increased Nrf2 protein levels. In addition, Keap1, LC3I/II, and mTOR protein expressions were evaluated by western blot analysis. (B) HDACi upregulated Nrf2 mRNA levels by qRT-PCR. (C) Cells were pre-incubated for 3 h with or without actinomycin D (1 mg/mL), treated with TSA for 24 h, and then assayed by western blot for Nrf2 and HO-1. (D) Subcellular localization of Nrf2 was determined by cell fraction after TSA treatment in Huh7 and MGC80-3 cells. (E) After TSA (1 μM) treatment for 24 h, HO-1 mRNA levels were evaluated by qRT-PCR. (F) Huh7 cells stably expressing GFP-RFP-LC3 were transfected with siRNA-Nrf2 or pretreated with Lut (20 μM for 10 min), then exposed to TSA for 24 h. Confocal microscopy was performed to quantify Green+ and Red+ autophagosomes and Green− and Red+ autophagolysosomes. (G) Huh7 cells were transfected with siRNA-Nrf2 and subsequently exposed to 1 μM TSA for 24 h, with or without Baf-A1 pretreatment (10 nM for 10 min). Nrf2, LC3I/II, and p62 levels were analyzed by western blot. (H) Cells pretreated for 3 h in the presence or absence of 20 μM Lut and then exposed to 1 μM TSA for 24 h were analyzed by western blot analysis. Data are presented as mean ± SD; *p < 0.05 versus control; **p < 0.01 versus control.

To test whether HDACi increased the transcriptional activity of Nrf2, β-actin and histone H3 proteins were used as markers of cytosolic and nuclear fractions, respectively. The results showed that HDACi increased the transcriptional activity of Nrf2 (Figure 1D). Moreover, Nrf2 upregulated the transcription of Nrf2-responsive genes such as HO-1 (Figure 1E).

To determine whether Nrf2 is involved in HDACi-induced autophagy, autophagic flux was estimated by tandem fluorescent mRFP-GFP-LC3.27, 28 Consistent with a previous report,⁷ TSA increased the number and accumulation of punctate LC3 (Figure 1F). Furthermore, small interfering RNA (siRNA)-mediated Nrf2 knockdown (si-Nrf2) reduced the number of LC3 puncta in the presence of TSA (Figure 1F). As shown in Figure 1G, TSA led to further LC3 accumulation in the presence of bafilomycin A1 (Baf-A1). Moreover, si-Nrf2 or Nrf2 inhibition by Luteolin (Lut) reduced autophagic flux by altering LC3-II and p62 protein levels (Figure 1H).

Nrf2 Regulates miR-129 Expression

Nrf2 can bind many miRNAs to regulate their transcription.¹¹ To evaluate whether Nrf2 regulated miRNA expression after HDACi treatment, we performed an Agilent human miRNA array on Huh7 cells transfected with vector control or the pcDNA3.1-Nrf2 expression construct. The results shown in Figures 2A and 2B indicate changes in Nrf2 protein and mRNA expression in response to pcDNA3.1-Nrf2. Differentially expressed miRNAs were then identified through fold change and p values; thresholds for up- and downregulated genes were set to fold change ≥2.0 and p ≤ 0.05. The most significantly increased and decreased miRNAs in pcDNA3.1-Nrf2 cells relative to controls are shown in Figures 2C and 2D. The results identified miR-129-3p as the most highly upregulated miRNA by Nrf2.

Figure 2.

miR-129-3p Is Induced by Histone Deacetylase Inhibitors

(A and B) Huh7 cells were transfected with the pcDNA-Nrf2 plasmid, and increased Nrf2 expression was confirmed by (A) western blot and (B) qRT-PCR. (C) miRNA array of pcDNA-Nrf2 Huh7 and control cells (3 versus 3). (D) Among the 23 miRs, miR-129-3p was the most significantly upregulated. (E) Hepatocellular carcinoma (HCC, n = 366) and gastric cancer (GC, n = 413) patients were divided into low- and high-expression groups, according to the OncomiR database. Survival signature analysis was calculated with the log-rank test. (F) miR-129-3p expression was confirmed by qRT-PCR after transfecting the pcDNA-Nrf2 plasmid. (G) Huh7 cells were treated with different concentrations of HDACis (TSA: 0, 0.5, 1, and 2 μM; SAHA: 0, 0.5, 1, and 2 μM; and romidepsin: 0, 20, 40, and 80 nM), and miR-129-3p expression was assayed by qRT-PCR 24 h post-treatment. (H) Huh7 cells were transfected with si-Nrf2, treated with TSA, and then miR-129-3p was analyzed by qRT-PCR 24 h after treatment. Data are expressed as mean ± SD; *p < 0.05 versus control; **p < 0.01 versus control.

To investigate whether miR-129-3p is dysregulated in hepatocellular carcinoma (HCC) and/or gastric carcinoma (GC), we analyzed miR-129-3p levels in the OncomiR database.²⁹ We found that the high miR-129-3p expression group exhibited a significantly poorer survival rate (Figure 2E). We further confirmed that both pcDNA3.1-Nrf2 and HDACi induced miR-129-3p expression by qRT-PCR (Figures 2F and 2G). Finally, NRF2 knockdown experiments confirmed that HDACi-induced upregulation of miR-129-3p is dependent on the activation of endogenous Nrf2 (Figure 2H).

Nrf2 Binds to ARE Sites in miR-129 and HO-1 Promoters

Many studies have reported that Nrf2 not only upregulates target genes but also regulates miRNAs. Nrf2 plays a vital catalytic role in chemotherapy treatment.11, 30 From bioinformatics data, we found that Nrf2 could bind to miR-129-3p in two specific regulatory regions, termed antioxidant-responsive element 1 (ARE1) (398–409 bp) and ARE2 (992–1,003 bp) (Figure 3A). We also identified two potential ARE sites in the HO-1 promoter (Figure 3A). Chromatin immunoprecipitation (ChIP) assays confirmed binding of Nrf2 to ARE1 of miR-129, but not ARE2 (Figure 3B). As a classic Nrf2 target gene, we also confirmed increased expression of HO-1, indicating that Nrf2 bound to the HO-1 promoter.

Figure 3.

Nrf2 Binds to ARE Sites in the Promoters of miR-129 and HO-1

(A) Positions of AREs within miR-129-3p and HO-1 are indicated. (B) ChIP analysis of the miR-129-3p promoter using antibodies against Nrf2; IgG was used as a negative control. qRT-PCR was performed with primers for predicted AREs; data are presented as percent of input. (C) Specific reporter plasmids were respectively co-transfected with Nrf2-WT or Nrf2-MUT into HEK293T cells. Luciferase activity was measured after 24 h. Data are expressed as mean ± SD; *p < 0.05 versus control; **p < 0.01 versus control.

To explore whether Nrf2 influenced miR-129 and HO-1-binding sites, miR-129 (398 bp) or HO-1 (355 bp) reporter plasmids were co-transfected with the Nrf2 expression plasmid or empty vector into HEK293T cells. Luciferase activity increased after co-transfection with the Nrf2 expression plasmid compared with empty vector. Collectively, these results demonstrate that Nrf2 directly regulated transcription of the miR-129 and HO-1 cluster by binding to the respective promoters (Figure 3C).

miR-129-3p Increases Autophagy In Vitro

Previous studies have shown that miRNAs can induce autophagy through an intrinsic pathway in cancer cells.³¹ Therefore, we investigated whether miR-129-3p regulated autophagy in cancer cell lines. Western blot results showed that conversion of LC3-I to LC3-II increased in both Huh7 and MGC80-3 cells (Figure 4A). miR-129-3p induced autophagosome accumulation, as indicated by fluorescence microscopy (Figure 4B). Pretreatment with 3-methyladenine (3MA), an autophagy inhibitor, restored the ability of miR-129-3p to induce autophagic flux (Figure 4C). Additionally, when combined with rapamycin, miR-129-3p promoted the formation of many double-layered-membrane autophagosomes in Huh7 cells (Figure 4D).

Figure 4.

Effects of miR-129-3p on Autophagy

(A) Overexpression of miR-129-3p in Huh7 and MGC80-3 cells or treatment with 100 nM rapamycin (Rap) for 24 h induced LC3 conversion and p62 degradation. (B) Huh7 cells infected with mRFP-GFP-LC3 lentivirus were transfected with miR-129 for 72 h or treated with 100 nM rapamycin for 24 h. Quantification of Huh7 cells as fluorescent dots by confocal microscopy; Green+ and Red+, autophagosomes; Green− and Red+, autophagolysosomes. (C) Huh7 cells infected with miR-129 were treated with 1 μM Trichostatin A (TSA), with or without 25 mM 3-methyladenine (3MA) for 24 h. (D) Huh7 cells transfected with miR-129 were treated with DMSO or 100 nM rapamycin for 24 h and then subjected to transmission electron microscopy (TEM). Data are expressed as mean ± SD; *p < 0.05 versus control; **p < 0.01 versus control.

Nrf2 Regulates mTOR via miR-129-3p

RepTar and RNAhybrid identified mTOR as a possible target of miR-129-3p, as the 3′ UTR of human mTOR mRNA contains potential binding sequences (Figure 5A). We found that wild-type mTOR luciferase signal was decreased by miR-129-3p, whereas the mutated mTOR 3′ UTR construct did not show a significant response to miR-129-3p (Figure 5B). We next examined the effect of miR-129-3p on endogenous mTOR expression. mTOR mRNA levels were decreased in miR-129-3p-overexpressing cells compared with controls (Figure 5C). Thus, we next investigated whether miR-129-3p affected mTOR protein levels and activity. 4E-BP1 and p-S6 are mTOR substrates, and p-S6 is associated with autophagy. Consistently, mTOR and S6 signaling was decreased upon miR-129a overexpression (Figure 5C). The results shown in Figure 5D indicate that anti-miR-129-3p or NRF2 knockdown significantly restored mTOR and phosho-mTOR (p-mTOR) degradation in response to TSA. Considered together, our data suggest that Nrf2 regulates mTOR via miR-129-3p to synergistically influence autophagy.

Figure 5.

mTOR Is a Direct Target of miR-129-3p

(A) RepTar and RNAhybrid predicted miR-129-3p-binding sites in the 3′ UTR of mTOR. Wild-type and mutated 3′ UTR of mTOR were cloned into a luciferase reporter plasmid. (B) Luciferase activity of Huh7 cells co-transfected with miR-129-3p mimic or inhibitor or negative controls and plasmids containing wild-type (WT) or mutant (MT) mTOR 3′ UTR. (C) Western blotting was used to detect mTOR, p-mTOR, and p-S6 expressions after Huh7 and MGC80-3 cells were transfected with miR-129-3p mimic. (D) Cells were pretreated with anti-miR-129-3p or NRF2 knockdown, exposed to 1 μM TSA for 24 h, and then analyzed by western blot of mTOR and p-mTOR proteins. Data are expressed as mean ± SD; *p < 0.05 versus control; **p < 0.01 versus control.

Nrf2 Inhibition and Anti-miR-129-3p Increase HDACi Chemosensitivity

The results of numerous investigations suggest that HDACi-induced autophagy serves as a protective mechanism.32, 33 In our study, exposing cancer cells to Lut significantly decreased levels of Nrf2 target genes HO-1, NQO1, and p62, and it synergistically increased the cytotoxicity of TSA (Figure 6B). The p62 gene is known to be a direct transcriptional target of Nrf2, whose activation can result in p62 protein accumulation. As expected, amounts of p62 protein had increased by 6 h, but then they decreased gradually until 24 h (Figure S1). There are two possible mechanisms mediating p62 degradation: first, as previously reported, activated Nrf2 translocates into the nucleus while Keap1 detaches from Nrf2 to form a complex containing p62, which is degraded by autophagy;³⁴ second, p62 is itself a substrate of autophagy, whereby it directly conjugates to ubiquitin-like proteins, such as LC3 and Atg12-Atg5, to facilitate degradation by autophagy.³⁵ Further analyses should clarify these discrepancies.

Figure 6.

Nrf2 Inhibition or Anti-miR-129-3p Increases Chemosensitivity to Histone Deacetylase Inhibitors

(A) Cell viability was determined by Cell Counting Kit-8 (CCK-8) assay at different time points after treatment. Huh7 and MGC80-3 cells were treated with Trichostatin A (TSA, 1 μM) at different time points. Next, Huh7 and MGC80-3 cell proliferation was assessed using a CCK-8 assay. (B) Lut inhibited the Nrf2 target gene by western blot analysis, while qRT-PCR confirmed that anti-miR-129-3p inhibited miR-129-3p expression. (C) Huh7 and MGC80-3 cells were transfected with anti-miR-129-3p for 48 h or pretreated with 1 μM Lut for 20 min. Cells were then exposed to 1 μM TSA for 48 h and analyzed for apoptosis by Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide staining. Data are expressed as mean ± SD; *p < 0.05 versus control.

Using a Cell Counting Kit-8 (CCK-8) assay, we confirmed that infection of the pcDNA-Nrf2 plasmid (Nrf2) or transfection with miR-129-3p mimic (miR-129-3p) in Huh7 or MGC80-3 cells can induce TSA resistance (Figure 6A). In contrast, as shown in Figure 6C, transfection of the miR-129-3p inhibitor into Huh7 and MGC80-3 cells potentiated the cytotoxic effects of TSA. These results demonstrated that inhibiting miR-129-3p or Nrf2 increased the sensitivity of cancer cells to HDACi.

Nrf2 Inhibition Increases Chemosensitivity to HDACis In Vivo

To confirm our in vitro data in vivo, we constructed orthotopic HCC model mice using Huh7 cells transfected with control vector or miR-129-3p sponge (anti-miR). Model mice were assessed by a bioluminescence imaging system. We found that a miR-129-3p sponge significantly enhanced the chemotherapeutic effect of TSA (Figure 7A). No significant loss in average body weight occurred (Figure 7B). Tumors were further analyzed by immunoblotting. Similar to in vitro results, the Nrf2 pathway and LC3 were enhanced to an extent in TSA-treated groups while the mTOR pathway was inhibited (Figure 7C). Taken together, these data indicate that the inhibition of miR-129-3p enhanced the anticancer effects of TSA.

Figure 7.

Inhibiting miR-129-3p Sensitizes Cells to Trichostatin A Treatment In Vivo

(A) Orthotopic hepatocellular carcinoma (HCC) mice were generated with miR-129-3p sponge (anti-miR)-expressing and negative control Huh7 cells (red arrow). Tumor growth rates were monitored weekly by bioluminescence imaging. HCC model mice were treated with TSA (0.5 mg/kg, intraperitoneally) or DMSO (negative control) once every 2 days for 35 days. (B) Body weights of the HCC model mice were measured every 7 days for 35 days. (C) Levels of Nrf2, HO-1, and autophagy-related proteins in resected tumors were analyzed by western blot.

Discussion

Chemotherapeutic resistance remains a major challenge for HDACi treatment in cancer.32, 36 Activation of Nrf2, which serves as a sensor of chemical-induced oxidative stress,³⁷ plays an important role in maintaining cellular homeostasis, and it is correlated with chemoresistance in a number of human malignancies.³⁸ Autophagy plays an important role in HDACi resistance. Elevated expression of Nrf2 and other autophagy-associated genes may contribute to drug resistance by generating reductive stress.³⁹ In accordance with previous reports, we found that HDACi treatment downregulated Keap1 protein levels, consequently resulting in Nrf2 activation. Furthermore, our data indicated that Nrf2 knockdown inhibited HDACi-induced autophagy. Under oxidative stress, Nrf2 is released from Keap1, whereby it translocates to the nucleus and binds to ARE sequences to affect the expression of almost 500 genes. We confirmed that HDACi increased the transcriptional activity of Nrf2, and specifically we found that it upregulated the Nrf2-responsive gene HO-1.

miRNAs, which are approximately 22 bp long, regulate gene expression and play a significant role in drug resistance.40, 41 We employed the Agilent human miRNA array on Huh7 cells infected with pcDNA3.1-Nrf2. Among the significantly increased and decreased miRNAs in experimental cells relative to controls, miR-129-3p was the most upregulated. Using the OncomiR database, we found that miR-129-3p expression correlated with poor outcomes in HCC and GC patients.²⁹ This suggested that miR-129-3p may have important cellular functions in drug resistance. We then confirmed that HDACis upregulated miRNA-129-3p by qRT-PCR. A recent ChIP sequencing (ChIP-seq) analysis suggested that several miRNAs could be regulated by Nrf2.⁴² Our ChIP-seq analysis revealed that Nrf2 bound to the ARE1 site in the miR-129-3p regulatory region and miR-129-3p levels were upregulated by Nrf2. Furthermore, we observed that miR-129-3p increased the conversion of LC3-I to LC3-II in Huh7 and MGC80-3 cells. Moreover, p62 was downregulated by miR-129-3p. Subsequently, we confirmed that miRNA-129-3p overexpression increased autophagic activity. In accordance with previous studies, we found that Nrf2 bound to the promoter of its target gene HO-1 and this interaction may regulate autophagy.43, 44, 45

Using the RepTar and RNAhybrid databases, we identified mTOR as a possible miR-129-3p target. Specifically, there was one potential miR-129-3p seed site in the mTOR 3′ UTR. mTOR plays an essential role in autophagy, and pharmacological inhibitors of the mTOR-signaling pathway are in clinical trials as anticancer agents.46, 47 mTOR signaling is also regulated by several miRNAs.48, 49 Our data confirmed that miR-129-3p overexpression inhibited luciferase reporter activity downstream of the mTOR 3′ UTR but this effect was abolished by mutating the miR-129-3p seed site. Western blotting confirmed that miR-129-3p induced autophagy by inhibiting mTOR and its downstream effector S6, a classic mTOR substrate involved in inhibiting autophagy. Furthermore, we observed that inhibiting miRNA-129-3p expression or Nrf2 made cells more sensitive to HDACis.

To test whether anti-miR-129-3p could be used as a potential sensitizing drug for HDACis, we employed a miR-129-3p sponge (anti-miR) in Huh7 cells that were injected into livers, yielding orthotopic mouse models. The 20 orthotopically injected mice were randomly divided into four groups. Consistent with our in vitro results, the anti-miR group was more sensitive to TSA. We acknowledge the limitations of only using two cell lines for our in vitro experiments and the small number of mice investigated. In the future, we plan to conduct studies that include a variety of cancer cell lines and more mice.

In conclusion, we uncovered a novel role of Nrf2 in controlling miRNA expression in response to anticancer drugs, and we demonstrated that the Nrf2-miR-129-3p-mTOR axis controls a novel miRNA regulatory network that is active during HDACi-induced autophagy (Figure 8). Furthermore, we identified miR-129-3p as a novel regulator of autophagy. Finally, our data highlight the importance of this axis in maintaining cellular homeostasis and, thus, chemoresistance. Alterations in the Nrf2-miR-129-3p-mTOR axis may, therefore, be useful as a therapeutic target in association with other drugs.

Figure 8.

Schematic of How the Nrf2-miR-129-3p-mTOR Axis Controls a Novel miRNA Regulatory Network that Regulates Histone Deacetylase Inhibitor-Induced Autophagy

Materials and Methods

Reagents

The potent selective HDACi TSA (S1045), potent non-selective HDACi SAHA (S1047), HDACi romidepsin (S3020), Nrf2, non-selective phosphodiesterase inhibitor Lut (S2320), rapamycin (Rap, S1039), and 3MA (S2767) were purchased from Selleck (Shanghai, China). Rabbit monoclonal anti-Nrf2 (ab62352), mouse monoclonal anti-Keap1 (ab119403), mouse monoclonal anti-HO-1 (ab13248), rabbit monoclonal anti-Histone H3 (ab176842), and mouse monoclonal anti-cleaved caspase-3 (ab13585) antibodies were supplied by Abcam (Cambridge, UK). Mouse monoclonal anti-β-actin (A3853), anti-LC-3 (085M48011), and anti-actinomycin D (50-76-0) antibodies were from Sigma-Aldrich (St. Louis, MO, USA). Mouse monoclonal anti-total mTOR (4517S), rabbit polyclonal anti-p-mTOR (Ser2448, 2971S), rabbit polyclonal anti-pho-p70 S6 Kinase (p-S6K, Thr398; 9209S), and rabbit polyclonal anti-p62 (5114T) antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). Human miR-129-3p-3p mimic, miR-129-3p-3p inhibitor, miR-129-3p-3p control, and Nrf2 siRNA were obtained from RiboBio (Guangzhou, China), as were specific primers for HO-1, LC3, Nrf2, mTOR, ATG12, Beclin-1, β-actin, U6, and miR-129-3p-3p. The Nrf2-containing plasmid pCD3.1 was provided by IDOBIO (Suzhou, China). DMSO was provided by Shenggong Bioengineering (Shanghai, China). Heat-inactivated fetal bovine serum (FBS) and MEM were from HyClone (Logan, UT). Huh7 and MGC80-3 cells were purchased from Shanghai Institute of Cell Biology, Chinese Academy of Sciences.

Cell Culture and Transfection

Huh7 cells are a well-differentiated hepatocyte-derived carcinoma cell line, while MGC-803 cells are a poorly differentiated primary gastric mucinous adenocarcinoma; both cell lines were cultured under standard culture conditions (37°C, 10% FBS, in 5% CO₂) for 24 h to allow adherence of cells to the bottoms of plates. For transfection, culture medium was replaced and incubated for 4 h before miRNA vectors for miR-129-3p mimic, miR-129-3p mimic control, miR-129-3p inhibitor, miR-129-3p inhibitor control, miR-129-3p-sponge, Nrf2 overexpression plasmid, or specific siRNAs were transfected into cells with Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). After transfection, TSA and SAHA were added before PCR or western blot analyses.

ChIP Assays

Huh7 cells were transfected with Nrf2 or control plasmid for 48 h. Transfected cells were fixed with 1% formaldehyde for 10 min at room temperature. DNA was fragmented into 200- to 900-bp segments by optimized ultrasound conditions. ChIP was performed with anti-Nrf2 antibody or control immunoglobulin G (IgG) (Cell Signaling Technology). Primers that spanned ARE sites (depicted in Figure 3A) were used on immunoprecipitated chromatin to determine associations with Nrf2.

Confocal Microscopy

Coverslips coated with poly-L-lysine were placed in each well of a six-well cell culture plate before seeding cells into each well. Coverslips were washed, mounted on a glass slide, and imaged using a Fluoview FV1 200 confocal laser-scanning microscope (Olympus, Tokyo, Japan) with a 40× lens. Images were analyzed using FV10-ASW version (v.)4.1 and Cellsens v.1.14 imaging software (Olympus).

Real-time qPCR

Total RNA from cells and tissues was extracted with TRIzol (Invitrogen), and 1 μg total RNA was reverse transcribed into cDNA using a Reverse Transcription Kit (RiboBio). cDNAs were amplified with a SYBR Premix Ex Taq II (Perfect Real-Time) kit (Takara, Shiga, Japan). U6 was used as an internal control. RNA expression levels were calculated by the 2^−ΔΔCt method. PCR primers used in this study were designed against human targets and had the following sequences:

• miR-129-3p-3p, 5′-AAGCCCTTAC-CCCAAAAAGTAT-3′;

• U6 sense, 5′-GCTTCGGCAGCACATATACTAAAAT-3′;

• HO-1, forward 5′-GTGCCACCAAGTTCAAGCAG-3′ and reverse 5′-CAGCTCCTGCAACTCCTCAA-3′;

• LC3, forward 5′-GTTGGTCAAGATCATCCGGCG-3′ and reverse 5′-AAGGTTTCCTGGGAGGCGTAGA-3′;

• Nrf2, forward 5′-ATGTTGGGAGTCTGGGCAAG-3′ and reverse 5′-TGTCACCATTGCAGGTTTCGC-3′;

• mTOR, forward 5′-ATCCAGACCCTGACCCAAAC-3′ and reverse 5′-TCCACCCACTTCCTCATCTC-3′;

• ATG12, forward 5′-ATCTTTTATGATGACTGGTGC-3′ and reverse 5′-ACAAAGAAAATCAACTTGCTAC-3′; and

• Beclin-1, forward 5′-GACCGAGTGACCATTCAGGA AC-3′ and reverse 5′-GGTTCT CCATGGTGCCACCATCAG-3′.

miRNA Array

An Agilent human miRNA array (8 × 60K, Design ID 070156) was used. Total RNA was extracted from Huh7 cells, and then three groups each of experimental and control cells were cultured for 72 h. Total RNA was quantified using a NanoDrop ND-2000 (Thermo Fisher Scientific, Waltham, MA, USA), and RNA integrity was assessed using a Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). Sample labeling, microarray hybridization, and washing were performed according to the manufacturer’s standard protocols. Arrays were scanned with an Agilent G2505C Scanner.

Western Blotting

Cells and tissues were lysed in RIPA buffer (Beyotime, Nantong, China) containing 1% protease inhibitor cocktail (Sigma-Aldrich). Protein concentrations were measured using a Bicinchoninic Acid Protein Assay Kit (Thermo Fisher Scientific). Equal amounts of heat-denatured proteins were separated by SDS-PAGE and transferred onto polyvinylidene fluoride (PVDF) membranes, which were blocked for 2 h before incubation at 4°C overnight with primary antibodies (1:1,000), followed by 37°C for 2 h with an appropriate secondary antibody.

Luciferase Reporter Assay

Mutant and wild-type sequences of putative binding sites were designed and co-transfected into cells with miR-129-3p mimic, miR-129-3p inhibitor, or negative control. Fluorescence intensity 48 h after transfection was measured with a luciferase kit (Yuanpinghao Biotechnology, Beijing, China).

Flow Cytometry, CCK-8, and Transmission Electron Microscopy

All reagents and procedures were as previously described.24, 50

HCC Mouse Model Construction

Male C57BL/6J mice (18–20 g) were housed under standard conditions in a specific pathogen-free room (Zhejian University Medical College Experimental Animal Center, Hangzhou, China). HCC orthotopic models were constructed using Huh7 cells stably expressing firefly luciferase, which were transfected with control vector or miR-129-3p sponge (anti-miR). Models were established and randomly divided into four groups (n = 5 per group). Model mice were assessed with a bioluminescence imaging system after the administration of TSA or DMSO by intraperitoneal injection once every 2 days for 35 days. All animal protocols were approved by the Ethical Review Committee of Zhejian Medical University.

Statistical Analysis

Data were analyzed with SPSS v.21.0 software (IBM, Armonk, NY, USA) and GraphPad Prism 6 software (GraphPad, San Diego, CA, USA). Relative gene expression levels were analyzed by Student’s t test. Results represent mean ± SD of three independent experiments, with p < 0.05 considered statistically significant. Agilent human miRNA array images generated by the Agilent scanner and Feature Extraction software (Agilent Technologies, v.10.7.1.1) were used to acquire raw data, which were analyzed by GeneSpring GX software (Agilent Technologies, v.13.1). For these experiments, ANOVA or t tests were used to evaluate differences between groups. The p values were two-sided, with p < 0.05 considered statistically significant.

Author Contributions

Experiment Design and Performance, W.S., W.W., and X.S.; Data Analysis and Acquisition, G.X., Y.Z., Y. Yu, Y. Yi, L.L., C.H., B.H., B.Y., and C. Ye; Technical and Material Support, C. Yu, F.T., C.C., X.X., and Z.Z.; Writing and Manuscript Review, W.S., W.W., and X.S.

Conflicts of Interest

The authors declare that they have no conflicts of interest.