MIR34A regulates autophagy and apoptosis by targeting HMGB1 in the retinoblastoma cell

Abstract

MIR34A (microRNA 34a) is a tumor suppressor gene, but how it regulates chemotherapy response and resistance is not completely understood. Here, we show that the microRNA MIR34A-dependent high mobility group box 1 (HMGB1) downregulation inhibits autophagy and enhances chemotherapy-induced apoptosis in the retinoblastoma cell. HMGB1 is a multifaceted protein with a key role in autophagy, a self-degradative, homeostatic process with a context-specific role in cancer. MIR34A inhibits HMGB1 expression through a direct MIR34A-binding site within the HMGB1 3′ untranslated region. MIR34A inhibition of HMGB1 leads to a decrease in autophagy under starvation conditions or chemotherapy treatment. Inhibition of autophagy promotes oxidative injury and DNA damage and increases subsequent CASP3 activity, CASP3 cleavage, and PARP1 [poly (ADP-ribose) polymerase 1] cleavage, which are important to the apoptotic process. Finally, upregulation of MIR34A, knockdown of HMGB1, or inhibition of autophagy (e.g., knockdown of ATG5 and BECN1) restores chemosensitivity and enhances tumor cell death in the retinoblastoma cell. These data provide new insights into the mechanisms governing the regulation of HMGB1 expression by microRNA and their possible contribution to autophagy and drug resistance.

Keywords: :

• microRNA
• Hmbg1
• autophagy
• apoptosis
• chemotherapy

Introduction

microRNAs (miRNAs) are a class of post-transcriptional regulators.¹ They are short, ~22 nucleotide RNA sequences that bind to complementary sequences in the 3′ untranslated region (3′UTR) of multiple target mRNAs (mRNAs). These molecules are expressed in a tissue-specific and developmental stage-specific manner and typically reduce mRNA stability. Alterations in the expression of several miRNAs are often present in human cancers and mediate processes in tumorigenesis and cancer therapy response and resistance.^2,3

The retinoblastoma is the most common intraocular malignant tumor in infants and children.⁴ Understanding the molecular mechanism(s) underlying retinoblastoma pathogenesis is crucial for designing novel targeted therapies to avoid enucleation of the eyeball to prevent metastasis. Recent evidence has indicated that MIR34A functions as a tumor suppressor in retinoblastoma and is a potential therapeutic target.^5,6 However, the mechanisms underlying this antitumor effect are not well characterized. MIR34A belongs to one of several evolutionarily conserved families of miRNAs, namely MIR34, and was originally identified as a TP53-targeted microRNA.⁷ MIR34A is expressed in almost every tissue except lung tissue, whereas MIR34B and MIR34C are mainly expressed in lung tissue but scarcely expressed in other tissues.⁸ MIR34A has been implicated in multiple cancer types, including retinoblastoma, and its expression is a prognostic parameter.^9-13 Numerous studies demonstrate that MIR34A is critically involved in regulating diverse tumor cellular processes such as apoptosis and the cell cycle.^7,14-17 Identifying functionally important mRNA targets of MIR34A is essential to unraveling its biological function and is helpful for cancer drug development.^18-20

HMGB1 (high mobility group box 1) belongs to a family of highly conserved proteins that contain HMG box domains. HMGB1 is a ubiquitous nuclear protein that regulates and facilitates various DNA-related activities such as transcription, replication, recombination, and repair. In addition to its functions in the nucleus, HMGB1 proteins play a significant extracellular role in inflammation and immunity.^21,22 HMGB1 has been proposed as a target for cancer therapy because impaired HMGB1 is associated with the hallmarks of cancer.²³ Blocking HMGB1 release and activity reduces tumor incidence, invasion, and metastasis.²⁴ Recent studies suggest that autophagy may be important in the regulation of cancer development and progression and in determining the response of tumor cells to anticancer therapy.²⁵ HMGB1 is an important autophagic regulator.^26-28 HMGB1-mediated autophagy inhibits apoptosis and promotes drug resistance in several tumor types, including leukemia, colorectal cancer, pancreatic cancer, and osteosarcoma.^29-34 However, little is known about the mechanism of HMGB1 upregulation and its role in retinoblastoma.

In this study, we examined the impact of MIR34A on HMGB1 expression and its molecular actions in the regulation of chemosensitivity in the retinoblastoma cell. Our study suggests that MIR34A targets HMGB1 mRNA, leading to translational repression of HMGB1. MIR34A is a potent inhibitor of autophagy, which promotes chemotherapeutic agent-induced DNA damage and apoptosis in a reactive oxygen species (ROS)-dependent manner. These findings provide the first evidence that the MIR34A-HMGB1 pathway regulates autophagy and apoptosis, as well as the chemotherapy response.

Results

HMGB1 is a target of MIR34A in human retinoblastoma cells

Based on the miRNA database, HMGB1 is a predicted MIR34A target in humans (Fig. 1A). A previous study has demonstrated that MIR34A was expressed in human retinoblastoma cells.⁶ Thus, we determined the effect of MIR34A on the expression of HMGB1 in human retinoblastoma cell lines (Y79 and Weri-RB1). To this end, we used a MIR34A mimic and antagoMIR34A to modulate cellular levels of MIR34A (Fig. 1B). The mRNA levels of HMGB1 decreased following MIR34A mimic treatment, whereas antagoMIR34A increased HMGB1 expression (Fig. 1B). The 3′ UTR reporter activity of HMGB1 was further assessed by luciferase assays. MIR34A mimic inhibited HMGB1 luciferase activities, whereas antagoMIR34A increased HMGB1 luciferase activities (“HMGB1-UTRwt”) in Y79 cells (Fig. 1C). In contrast, MIR34A mimic had no effect on the expression of the luciferase reporter regulated by HMGB1-UTR with mutated MIR34A binding sites (“HMGB1-UTRmt”) (Fig. 1C). Consistently, the protein levels of HMGB1 decreased following MIR34A mimic treatment, whereas antagoMIR34A increased HMGB1 protein expression in Y79 and human HCT116 colon cancer cells (Fig. 1D). SIRT1 (sirtuin 1) is an NAD-dependent deacetylase that regulates apoptosis and autophagy during stress.^17,35 Recent data suggests that SIRT1 may function as a MIR34A-targeted gene in several cancer cells, including HCT116 cells (Fig. 1D).¹⁷ However, MIR34A did not regulate SIRT1 expression in Y79 and Weri-RB1 cells. These findings suggest that HMGB1 is the target of MIR34A, and MIR34A regulates gene expression in a tissue-specific manner.

Figure 1.MIR34A targets HMGB1. (A) Schema representing the functional interaction between MIR34A and the seed sequence (bold) in the 3′UTR of HMGB1 as predicted by miRanda-mirSVR (http://www.microrna.org/microrna/home.do). (B) Y79 and Weri-RB1 cells were transfected with antagomir or mimic of MIR34A or nontarget antagomir or mimic control (100 nM) for 72 h, and the level of MIR34A was assayed by TaqMan qRT-PCR. In parallel, the mRNA level of HMGB1 was assayed by qRT-PCR (n = 3, *P < 0.05). Data represent relative level, with control set as 1. (C) Luciferase assay of Y79 cells cotransfected with reporter constructs containing HMGB1 3′ UTRs with (HMGB1-UTRwt) or without (HMGB1-UTRmt) MIR34a-binding sites and mimic of MIR34A or scrambled control miRNA for 72 h (n = 3, *P < 0.05; ns, nonsignificant). Data represent relative level, with control set as 100%. (D) Y79, Weri-RB1 and HCT116 cells were transfected with indicated miRNAs or antagomir RNAs for 72 h, and the level of HMGB1 and SIRT1 protein expression was assayed by western blot. AU = arbitrary unit.

MIR34A induces apoptosis by modifying HMGB1 expression

Next, we investigated whether suppression of HMGB1 expression by short hairpin RNA (shRNA) has a similar function compared with treatment with MIR34A mimic. MIR34A can potentially affect several signaling pathways and is implicated in tumor malignancy. Induction of apoptosis has been identified as a major mechanism that contributes to the tumor suppressive function of MIR34A.⁹ Thus, we determined whether knockdown of HMGB1 by shRNA induces apoptosis in retinoblastoma cells. Similar to MIR34A mimic treatment, suppression of HMGB1 expression promotes apoptosis by flow cytometric analysis of the percentage of cells that are ANXA5/annexin V-positive (Fig. 2A). CASP3 (caspase 3) is a critical executioner of apoptosis, as it is either partially or totally responsible for the proteolytic cleavage of many key proteins, such as PARP1.³⁶ Knockdown of HMGB1 or MIR34A mimic increased the activity of CASP3 (Fig. 2B), cleaved-CASP3, and cleaved-PARP1 (Fig. 2C). These findings suggest that knockdown of HMGB1 or treatment with MIR34A similarly influence the regulation of apoptosis. Furthermore, overexpression of HMGB1 by transfection with full-length HMGB1 cDNA (“pUNO1-HMGB1”) inhibited MIR34A mimic-induced apoptosis (Fig. 2D), CASP3 activity (Fig. 2E), and the expression of cleaved PARP1 (Fig. 2F), suggesting that downregulation of HMGB1 is required for MIR34A mimic-induced apoptosis in retinoblastoma cells.

Figure 2.MIR34A induces apoptosis through suppression of HMGB1 expression. (A and B) Y79 cells were transfected with mimic of MIR34A, non-target mimic control, HMGB1 shRNA (HM-shRNA), or control shRNA (C-shRNA) for 0 to 72 h, and then apoptosis (A) was analyzed by measuring ANXA5-positive cells by flow cytometry and CASP3 activity (B) was measured with the CASP3 Colorimetric Assay Kit (n = 3, *P < 0.05). (C) In parallel, the indicated protein level was assayed by western blot. (D and E) Y79 cells were transfected with mimic of MIR34A, control pUNO1 cDNA, or pUNO1-HMGB1 cDNA for 72 h, and then apoptosis (D) was analyzed by measuring ANXA5-positive cells by flow cytometry and CASP3 activity (E) was measured by CASP3 Colorimetric Assay Kit (n = 3, *P < 0.05). (F) In parallel, the indicated protein level was assayed by western blot.

MIR34A inhibits autophagy by modifying HMGB1 expression

Given that HMGB1 is a target of MIR34A (Fig. 1), we investigated the role of MIR34A in autophagy. The most typical trigger of autophagy is nutrient starvation using Hank's balanced salt solution (HBSS). Starvation increased mRNA expression of HMGB1 at 1 to 2 h, and returned to the baseline levels at 6 h when MIR34A expression was increased in the retinoblastoma cell (Fig. 3A). MAP1LC3/LC3 (microtubule-associated protein 1 light chain 3) is now widely used to monitor autophagy because the amount of LC3-II correlates with the number of autophagosomes.³⁷ We found that MIR34A mimic or knockdown of HMGB1 inhibited starvation-induced MAP1LC3-II expression and the number of MAP1LC3 puncta in Y79 and HCT116 cells (Fig. 3B and C). In contrast, treatment with lysosomal protease inhibitors E64d and pepstatin A (E/P) restored MAP1LC3-II expression and MAP1LC3 puncta in MIR34A-transfection cells (Fig. 3B and C). As expected, the levels of polyubiquitin binding protein SQSTM1/p62 (sequestosome 1)³⁸ increased when starvation-induced autophagy was impaired (Fig. 3B). An assay measuring the autophagy-mediated degradation of radiolabeled long-lived proteins showed that MIR34A mimic or knockdown of HMGB1 inhibited starvation-induced protein degradation (Fig. 3D). Moreover, ultrastructural electron microscopy analysis revealed that MIR34A mimic decreased the number of autophagic vesicles after starvation (Fig. 3E). These findings suggest that MIR34A regulates autophagy flux and functions as a potent inhibitor of autophagy.

Figure 3.MIR34A inhibits autophagy through modification of HMGB1 expression. (A) Y79 cells were incubated in HBSS for different lengths of time (0 to 6 h); the relative expression levels of HMGB1 mRNA and MIR34A were assessed by qRT-PCR and TaqManqRT-PCR, respectively, with control set as 100%. (B) Y79 and HCT116 were transfected with mimic of MIR34A, mimic control, HMGB1 shRNA, or control shRNA for 72 h, and then treated with HBSS with or without E64D/ pepstatin A (E/P, 10 μg/ml) for 1 h. The levels of MAP1LC3-I/II, SQSTM1, and HMGB1 protein expression were assayed by western blot. (C–E) In parallel, the MAP1LC3 puncta formation (C, shown by red arrow), the long-lived protein degradation and ultrastructural features of autophagic vacuoles (E, shown by red arrow) was assayed as described in Materials and Methods (*P < 0.05). (F and G) Hmgb1^+/+ and hmgb1^−/− MEFs or SIRT1 knockdown Y79 cells were transfected with antagoMIR34A or nontarget antagomir for 72 h or treated with HBSS in the presence and absence of E64D/pepstatin A (E/P, 10 μg/ml) for 1 h, the levels of MAP1LC3-I/II, SQSTM1, and HMGB1 protein expression were assayed by western blot (F), and MAP1LC3 puncta formation was assayed by confocal microscopic analysis (G). *P < 0.05. (H) Y79 cells were transfected with mimic of MIR34A, control pUNO1 cDNA, or pUNO1-HMGB1 cDNA for 72 h, and then treated with HBSS in the presence and absence of E64D/pepstatin A (E/P, 10 μg/ml) for 1 h; MAP1LC3 puncta formation was assayed by confocal microscopic analysis. *P < 0.05.

Previous studies have demonstrated that deletion, depletion, or inhibition of HMGB1 in mouse embryonic fibroblasts (MEFs) or tumor cells markedly diminishes autophagy.^29-34 Interestingly, antagoMIR34A promoted HMGB1 expression (Fig. 1; Fig. 3F), MAP1LC3-II expression (Fig. 3F), and MAP1LC3 puncta (Fig. 3G) in HMGB1 wild type, but not in HMGB1 knockout MEFs in the presence and absence of E/P, suggesting that upregulation of HMGB1 is required for antagoMIR34A-induced autophagy. However, knockdown of SIRT1 by shRNA did not influence antagoMIR34A- and starvation-induced autophagy in Y79 cells (Fig. 3G), suggesting that MIR34A-mediated autophagy is SIRT1-independent in the retinoblastoma cell. Furthermore, overexpression of HMGB1 by transfection with a full-length HMGB1 cDNA (“pUNO1-HMGB1”) restored MIR34A mimic-mediated autophagy inhibition under starvation conditions (Fig. 3H), suggesting that upregulation of HMGB1 is required for autophagy under starvation conditions.

Inhibition of autophagy enhances chemotherapeutic efficacy in retinoblastoma cells

Vincristine (VCR), etoposide (ETO), and carboplatin (CBP) are agents commonly used systemically in retinoblastoma treatment. These anticancer drugs increased MAP1LC3-II and HMGB1 expression (Fig. 4A) and decreased MIR34A expression (Fig. 4A) in retinoblastoma cells. Different from ETO and CBP, inhibition of autophagy by the microtubule-destabilizing agent VCR results from an impairment of autolysosome formation because vincristine prevents SQSTM1 degradation, whereas ETO and CBP promote SQSTM1 degradation (Fig. 4A).³⁹ Similar to starvation-induced autophagy (Fig. 3), MIR34A mimic and knockdown of HMGB1 significantly decreased VCR, ETO, and CBP-induced autophagy by monitoring MAP1LC3 puncta (Fig. 4B), but enhanced chemotherapeutic efficacy in retinoblastoma cells (Fig. 4C). BECN1 and ATG5 are critical regulators of autophagy.⁴⁰ Inhibition of autophagy by knockdown of BECN1 and ATG5 (Fig. 4D and E) also increased chemotherapeutic efficacy in retinoblastoma cells (Fig. 4F). These findings suggest that enhanced autophagy increases drug resistance in retinoblastoma cells. Furthermore, overexpression of HMGB1 by transfection with a full-length HMGB1 cDNA restored MIR34A mimic-mediated autophagy inhibition (Fig. 4G) and inhibited chemotherapeutic efficacy after CBP treatment (Fig. 4H), suggesting that MIR34A inhibits autophagy and increases cell death induced by chemotherapy in an HMGB1-dependent manner.

Figure 4. Inhibition of autophagy enhances chemotherapeutic efficacy in retinoblastoma cells. (A) Y79 cells were treated with vincristine (VCR, 2.5 nM), etoposide (ETO, 1 µM), and carboplatin (CBP, 1 µM) in the presence and absence of E64D/pepstatin A (E/P, 10 μg/ml) for 48 h, and the levels of MAP1LC3-I/II and HMGB1 were assayed by western blot. In parallel, the level of MIR34A was assayed by TaqMan qRT-PCR (n = 3, *P < 0.05, VS. control group). Data represent relative level, with control set as 1. (B and C) Y79 were transfected with mimic of MIR34A, mimic control, HMGB1 shRNA, or control shRNA for 72 h, and then treated with VCR (2.5 nM), ETO (1 µM), and CBP (1 µM) in the presence and absence of E64D/pepstatin A (E/P, 10 μg/ml) for 48 h. The MAP1LC3 puncta formation (B) was assayed by confocal microscopic analysis and cell viability (C) was assayed by CCK-8 kit (*P < 0.05). (D) Western blot analysis of BECN1 and ATG5 expression in Y79 cells after transfection with indicated shRNA. (E and F) Y79 cells were transfected with BECN1 shRNA, ATG5 shRNA, or control shRNA for 72 h, and then treated with VCR (2.5 nM), ETO (1 µM), and CBP (1 µM) for 48 h. MAP1LC3 puncta formation (E) was assayed by confocal microscopic analysis and cell viability (F) was assayed by CCK-8 kit (*P < 0.05). (G) Y79 cells were transfected with mimic of MIR34A, control pUNO1 cDNA, or pUNO1-HMGB1 cDNA for 72 h, and then treated with carboplatin (CBP, 1 µM) in the presence and absence of E64D/pepstatin A (E/P, 10 μg/ml) for 48 h, and MAP1LC3 puncta formation was assayed by confocal microscopic analysis (*P < 0.05). (H) In parallel, cell viability was assayed by CCK-8 kit (n = 3, *P < 0.05).

Inhibition of autophagy enhances DNA damage in chemotherapy

The anticancer activity of most chemotherapy drugs relies on the induction of DNA damage in rapidly cycling tumor cells concomitant with the inhibition of cellular DNA repair processes.⁴¹ Exposure of Y79 cells to ETO and CBP resulted in the phosphorylation of histone H2A at Ser139 (gamma-H2AFX), a marker for DNA damage (Fig. 5A). MIR34A mimic, knockdown of HMGB1, and knockdown of ATG5 increased chemotherapy drug-induced gamma-H2AFX expression, suggesting that inhibition of autophagy enhances DNA damage in chemotherapy (Fig. 5A). Apoptosis induced by DNA damage is an important mechanism of tumor suppression, and is also significant in cancer chemotherapy. As expected, MIR34A mimic, knockdown of HMGB1, and knockdown of ATG5 increased chemotherapy-induced apoptosis (Fig. 5B and C). However, knockdown of SIRT1 cannot increase chemotherapy-induced apoptosis (Fig. 5C). Moreover, the antioxidant N-Acetyl-L-Cysteine (NAC) attenuated ETO- and CBP-induced DNA damage (Fig. 5D) and apoptosis (Fig. 5E) in ATG5 knockdown cells, suggesting that ROS-induced DNA damage is required for induction of apoptosis in autophagy-deficient cells. Mitochondrial membrane potential is an important parameter of mitochondrial function used as an indicator of cell health. There is a tight link between mitochondrial membrane potential and the rate of ROS formation. As expected, MIR34A mimic and knockdown of HMGB1 increased chemotherapy drug-induced ROS production (Fig. 5F) and decreased mitochondrial membrane potential (Fig. 5G). Taken together, these data support the notion that miRNAs regulate autophagy, thereby largely influencing mitochondrial function as well as ROS production.

Figure 5. Inhibition of autophagy enhances DNA damage in chemotherapy. (A) Indicated Y79 cells were treated with etoposide (ETO, 1 µM) and carboplatin (CBP, 1 µM) for 48 h, and the level of gamma-H2AFX was assayed by western blot. (B and C) In parallel, apoptosis was analyzed by measuring ANXA5-positive cells (n = 3, *P < 0.05). (D) ATG5 knockdown Y79 cells were treated with ETO (1 µM) and CBP (1 µM) with or without antioxidant NAC (50 mM) for 48 h, and the level of gamma-H2AFX was assayed by western blot. (E) In parallel, apoptosis was analyzed by measuring ANXA5-positive cells (n = 3, *P < 0.05). (F and G) Indicated Y79 cells were treated with ETO (1 µM) and CBP (1 µM) for 48 h. ROS production (F) was assessed by measuring the fluorescent intensity of CM-H2DCFDA in a fluorescent plate reader, and mitochondrial membrane potential depolarization (G) was measured by JC-1 fluorescence (n = 3, *P < 0.05).

Discussion

Autophagy is a catabolic process involving the degradation of a cell's own components, such as waste or excess proteins and organelles through the lysosomal machinery.⁴² It has been a fast-growing field of research in recent years, and is controlled by an ever-growing number of regulators, signalers, and miRNAs.^40,43-45 In this study, we identified that tumor suppressive miRNA, MIR34A, is a potent inhibitor of autophagy by downregulation of HMGB1, but not SIRT1 in the retinoblastoma cell. MIR34A was shown to target SIRT1 mRNA, leading to translational repression of SIRT1 in several human cancer cells. Recent evidence has demonstrated that activating SIRT1 induces autophagy, and loss of endogenous SIRT1 inhibits autophagy in several cell types. However, we found that the expression of SIRT1 in retinoblastoma cells is not required for antagoMIR34A and starvation-induced autophagy. Indeed, a recent study suggests that network organization of the human autophagy system is complex and each autophagic pathway has different components.⁴⁶

HMGB1, a conserved nuclear protein that enhances transcription, was recently discovered to be a crucial regulator of autophagy that mediates the stress response.⁴⁷ HMGB1 plays subcellular localization-dependent roles in the regulation of autophagy. Nuclear HMGB1 regulates small heat-shock protein HSPB1 expression.²⁸ Phosphorylation of HSPB1 is necessary for the regulation of the actin cytoskeleton, which affects the cellular transport required for autophagy and mitophagy in response to mitochondrial injury.²⁸ Cytosolic HMGB1 is a novel BECN1-binding protein that dissociates its inhibitory partner BCL2 to induce autophagy.²⁶ Moreover, ULK1 and TP53 positively and negatively regulate interaction between HMGB1 and BECN1 in cancer cells, respectively.³⁰ Extracellular HMGB1 binds AGER/RAGE (advanced glycosylation end product-specific receptor), but not TLR4, which inhibits MTOR and promotes the formation of BECN1-PIK3C3 core autophagic complexes in cancer cells. Induction of autophagy by both intracellular and extracellular HMGB1 is important for tumor development and a novel target for cancer therapy.^23,30,32,48 Our findings suggest that MIR34A inhibits HMGB1 expression and reduces its protein levels, thereby preventing the activation of autophagy.

Current cancer therapies, including chemotherapy, radiotherapy, and immunotherapy, are known to induce a tight interplay between autophagy and apoptosis.⁴⁹ However, autophagy plays the dual role of either promoting cell survival or promoting cell death (“autophagic cell death”) in response to these cancer treatments, depending on the cellular context and the nature of the treatment.²⁵ Indeed, accumulating evidence challenges the concept of autophagic cell death in mammalian cells.⁵⁰ Inhibition of autophagy results in a shift from an autophagic morphology to nonautophagic cell death, including apoptosis and necrosis.⁵⁰ In many cases, autophagic cell death is cell death with autophagy, but not cell death by autophagy.⁵¹ We demonstrated that MIR34A-mediated downregulation of HMGB1 and inhibition of autophagy can sensitize retinoblastoma cells to chemotherapy drug-mediated apoptosis. Of note, the molecular mechanism of drug-mediated autophagy in cancer cells is different. We showed that VCR blocks autolysosome formation, whereas ETO and CBP inhibit autophagosome formation. Therefore, mechanisms of the observed anticancer effects with MIR34A mimic and knockdown of HMGB1 could vary in different chemotherapies. These data support the notion that miRNAs regulate autophagy-related genes and autophagy regulators, thereby largely influencing the efficacy of chemotherapy.

The production of ROS by mitochondria is important because it underlies oxidative damage and cell death in many diseases.⁵² Autophagy has been implicated in the removal of damaged mitochondria (termed mitophagy) and is an important mechanism for mitochondrial quality control.⁵³ Autophagy deficiency in cells and yeast leads to mitochondrial injury and ROS production.^54,55 A previous study showed that HMGB1-mediated autophagy and mitophagy is important for mitochondrial quality control.²⁸ Knockout or knockdown of HMGB1 in fibroblasts and cancer cells significantly inhibits mitochondrial energy production and increases ROS production.²⁸ PTEN-induced putative kinase-1 (PINK1) is a kinase of the outer mitochondrial membrane and PARK2/PARKIN is a gene implicated in autosomal recessive juvenile Parkinsonism. The PINK1-PARK2 pathway has been implicated in mitophagy and mitochondrial dysfunction in humans,⁵⁶ and is required for HMGB1-mediated mitophagy and mitochondrial quality control.²⁸ We found that loss of the MIR34A-HMGB1 pathway inhibits autophagy and results in increased ROS production during chemotherapy in retinoblastoma cells. Excess ROS may induce oxidative DNA damage and chromosomal aberrations. Defects in DNA damage-induced apoptosis contribute to tumorigenesis and to the resistance of cancer cells to a variety of therapeutic agents. Thus, the level of MIR34A expression may affect the outcome of either autophagy or apoptosis.

Taken together, this work suggests that HMGB1 may be one of several distinct targets of MIR34A that contribute to its ability to regulate cell survival and death by regulation of autophagy and apoptosis (Fig. 6). MIR34A inhibits HMGB1 expression through a MIR34A-binding site within the 3′UTR of HMGB1 in the retinoblastoma cell. Accumulating data suggest that HMGB1 is an important autophagy regulator during cell stress. MIR34A inhibition of HMGB1 leads to a significant decrease in autophagy. Chemotherapy is one of the main methods of cancer treatment and is known to induce autophagy in cancer cells. Our findings further suggest that MIR34A-HMGB1 pathway-mediated autophagy increases chemotherapy-induced apoptotic resistance by modification of the ROS-mediated DNA damage response. Thus, targeting the MIR34A-HMGB1 pathway in cancer research will provide new opportunities for drug development.

Figure 6.MIR34A regulates autophagy and apoptosis by targeting HMGB1. HMGB1 is a direct transcriptional target of MIR34A. MIR34A inhibition of HMGB1 leads to a decrease in autophagy that promotes oxidative injury, DNA damage, and subsequent apoptosis during chemotherapy.

Materials and Methods

Reagents

The antibodies to CASP3 (9665), cleaved-PARP1 (5625), and SIRT1 (9475) were obtained from Cell Signaling Technology. The antibodies to HMGB1 (ab18256) and ACTB/actin (ab3280) were obtained from Abcam. The antibody to MAP1LC3 (L8918) was obtained from Sigma. The antibody to SQSTM1 (sc-28359) came from Santa Cruz Biotechnology. Hoechst 33342 (H1399) was obtained from Invitrogen. All other anticancer and chemical reagents not listed were obtained from Sigma.

Cell culture

The human retinoblastoma cell lines Y79 (HTB-18), Weri-Rb1 (HTB-169), and HCT116 (CCL-247) were obtained from American Type Culture Collection and cultured in RPMI-1640 (11875-093) or McCoy’s 5a (Life Technologies, 16600-082) plus 10% heat-inactivated fetal bovine serum and 1% penicillin-streptomycin (Invitrogen, 15140) in a humidified incubator with 5% CO₂ and 95% air. Hmgb1^+/+ and hmgb1^−/− immortalized MEFs were a kind gift from Dr Marco E Bianchi (San Raffaele Institute, Italy).

Transfection and RNAi

Short hairpin RNA (shRNA) against HMGB1 (SHCLNV-NM_002128), BECN1 (SHCLNV-NM_003766), ATG5 (SHCLNV-NM_004849), and SIRT1 (SHCLNV-NM_012238) were obtained from Sigma. The mimic and antagomir against MIR34A (C-301145-01; IH-301145-02) were obtained from Dharmacon. The pUNO1-HMGB1 plasmid (puno1-hhmgb1-s) was obtained from InvivoGen. HMGB1 3′UTR Lenti-reporter-Luciferase Vector (MT-h09960) was obtained from ABM. This HMGB1 3′UTR-containing vector (pLenti-UTR-Luc) was used as a template to create mutated MIR34A binding sites (HMGB1-UTRmt) by using the GeneArt^® Site-Directed Mutagenesis PLUS System (Invitrogen, A14604) according to the manufacturer’s instructions. These shRNAs, miRNAs, and cDNA were transfected into cells by lentiviral delivery systems (Sigma, SHP001) or Lipofectamine 2000 reagent (Invitrogen, 11668) according to the manufacturer’s instructions.⁵⁷

Cell viability assay

Cells were plated at a density of 4 × 10⁴ cells/well in 96-well plates in 100 μl medium. After treatment, cell viability was evaluated using the Cell Counting Kit-8 (Dojindo Laboratories, CK04-01) according to the manufacturer’s instructions.

Luciferase reporter assay

Cells were lysed in passive lysing buffer and then analyzed for firefly and Renilla luciferase activities using the commercial Dual-Luciferase reporter assay system (Promega, E1910) according to the manufacturer’s instructions; firefly luciferase activity was normalized to the Renilla luciferase activity.

Western blotting

Protein quantity from cell extracts was determined using the Bradford method. Protein was separated by electrophoresis in 4–12% Criterion XT Bis-Tris gels (Bio-Rad, 345-0124) and was transferred to nitrocellulose membrane, blocked with 5% milk in Tris-buffered saline with Tween 20, and probed with the indicated primary antibodies. After incubation with peroxidase-conjugated secondary antibodies, bands were visualized with enhanced chemiluminescence (Pierce, 34079). The relative band intensities were quantified using Gel-pro Analyzer^® software (Media Cybernetics).

RNA isolation and qRT-PCR

Total RNA was extracted by using Trizol (Invitrogen, 15596-026) according to the manufacturer’s instructions and reverse transcribed for quantification using the TaqMan microRNA Reverse Transcription Kit (Applied Biosystems, 4366596) according to the manufacturer’s instructions. Mature miRNAs were quantified using 2-step TaqManqRT-PCR with the TaqMan microRNA kit. miRNA expression level was normalized using RNU66rRNA as an internal control as previously described.⁵⁸ qRT-PCR for HMGB1 with specific primers (PPH00999F) from SABiosciences was performed using a RT² First Strand Kit (SABiosciences, 330401) according to the manufacturer’s instructions. Fold changes of mRNA levels in target gene relative to the POLR2A/RNA polymerase II [polymerase (RNA) II (DNA directed) polypeptide A, 220 kDa] control were calculated by relative quantification analysis.

Autophagy assays

The number of MAP1LC3 puncta was determined by quantifying the number of positively stained cells from 100 randomly chosen cells by Image-Pro Plus 5.1 software (Media Cybernetics) as previously described.²⁶ Autophagic flux assays were performed by western blotting for MAP1LC3-I/II and SQSTM1 in the presence or absence of lysosomal protease inhibitors or by long-lived protein degradation assays as previously described.^57,59 Transmission electron microscopy assessment of autophagic vacuoles was performed as previously described.²⁸

Apoptosis assays

Apoptosis in cells was assessed using the ANXA5-FITC Apoptosis Detection Kit (BD PharMingen, 556570) by flow cytometric analysis. CASP3 activity assays were performed using the CASP3 Colorimetric Assay Kit from Calbiochem (235419-1KIT) according to the manufacturer’s instructions. Cleaved PARP1 and cleaved CASP3 were measured by western blot analysis. Mitochondrial membrane potential depolarization was measured by flow cytometry using the fluorescent cationic dye JC-1, (5,5′,6,6′-tetrachloro-1,1′3,3′-tetraethylbenzamidazolocarbocyanin iodide; Molecular Probes, M34152). JC-1 dye exhibits potential-dependent accumulation in mitochondria, indicated by a fluorescence emission shift from green to red. Consequently, mitochondrial depolarization was assessed by a decrease in the red/green fluorescence intensity ratio.

Measurement of intracellular ROS

Cells were seeded in 96-well plates and cultured in the presence of a given stimuli for the indicated time. ROS were assessed with cell-permeable dye CM-H2DCFDA (Invitrogen, C6827) as described previously.²⁶ Analysis of signal intensity was performed using a fluorescent plate reader.

Statistical analysis

One-way ANOVA was used for comparison among the different groups. When the ANOVA was significant, post hoc testing of differences between groups was performed using an LSD test. A P value < 0.05 was considered significant.

Abbreviations:
AGER	=	advanced glycosylation end product-specific receptor
CBP	=	carboplatin
E/P	=	E64d and pepstatin A
ETO	=	etoposide
HBSS	=	Hank’s balanced salt solution
HMGB1	=	high mobility group box 1
H2AFX	=	H2A histone family, member X
gamma-H2AFX	=	phosphorylation of histone H2AFX at Ser139
MAP1LC3	=	microtubule-associated protein 1 light chain 3
MEFs	=	mouse embryonic fibroblasts
miRNAs	=	microRNAs
MIR34A	=	microRNA 34a
mRNAs	=	messenger RNAs
NAC	=	N-acetyl-L-cysteine
PARP1	=	poly (ADP-ribose) polymerase 1
PINK1	=	PTEN-induced putative kinase 1
ROS	=	reactive oxygen species
SQSTM1	=	sequestosome 1
shRNA	=	short hairpin RNA
UTR	=	untranslated region
VCR	=	vincristine

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Christine Heiner (Departments of Surgery and Anesthesiology, University of Pittsburgh) for her critical reading of the manuscript. This work was supported by grants from The National Natural Sciences Foundation of China (30973234 and 31171328 to LC; 81270616 to YY) and a grant from the US. National Institutes of Health (NIH) (R01CA160417 to DT).