Effects of different autophagy inhibitors on treatment with ATO/ATRA in KG-1 and HL-60 leukemia cells

Background : Few current autophagy inhibitors may have benecial effects for acute myeloid leukemia (AML) patients in clinical phases. However, there is a strong need to gure out which settings should be activated or inhibited in autophagy pathway to prevail drug resistance and improve current treatment options in leukemia. This study aimed to compare the effects of 3 well-known inhibitors of autophagy (3-MA, BafA1 and HCQ) in human leukemia KG1 and HL-60 cells exposed to ATO and/or ATRA. Methods : Cell proliferation and cytotoxicity of KG-1 and HL-60 cells were examined by MTT assay. Autophagy was studied by evaluating the development of acidic vesicular organelles, and the autophagosomes formation was investigated by acridine orange staining and transmission electron microscopy. Gene and protein expression levels of autophagy markers ( ATGs , p62/SQSTM1 and LC-3B) were also performed by qPCR and western blotting, respectively. The rate of apoptosis (annexin-V/PI staining) and cell cycle were evaluated using ow cytometry. Result: We compared the cytotoxic and apoptotic effects of arsenic trioxide (ATO) and/or all-trans retinoic acid (ATRA) in KG1 and HL-60 cells, and demonstrated that some autophagy markers can be upregulated in this context. Also, autophagy blockers HCQ and/or BafA1 could potentiate the cytotoxic effects of ATO/ATRA, which were more pronounced in KG-1 cells compared with HL-60 cell line. Conclusion: This study showed the involvement of autophagy during treatment of KG1 and HL-60 cells with ATO/ATRA.also, this study propose was to propose that combination therapy of ATO/ATRA with the autophagy inhibitor HCQ can be considered as a more effective strategy for targeting leukemic KG-1 cells.


Introduction
Macroautophagy, hereafter referring as autophagy, is a catabolic signaling pathway as a result of stress conditions or nutrient limitations, to maintain cellular metabolism and organelle integrity (1)(2)(3). Although autophagy is a homeostatic mechanism for intracellular recycling and the removal of damaged organelles or malfolded proteins. During autophagy, cargos are sequestrated into double-membrane vesicles, called autophagosomes, which then fuse with lysosomes for subsequent degradation or recycling of their contents (4, 5). Autophagy can confer stress tolerance, limit trauma, and sustain viability in harsh conditions. De ciency in autophagy has been implicated in various diseases such as aging, neurodegeneration, liver disease, and mainly in cancer (6). The function of autophagy as a prosurvival has been determined at the cellular and organismal level in different situations such as during endoplasmic reticulum stress, nutrient or growth factor deprivation, microbial infection, and the diseases characterized by the accumulation of protein aggregates. This pro-survival function is known to be adaptive; however, in the context of cancer, it might signi cantly become abnormal. Metabolic stress is a common speci cation of the tumor niche, and most of the chemotherapeutic agents promote cellular stress. Therefore, an area of intense investigation have been focused on whether autophagy-dependent survival should be blocked in these conditions to enhance tumor-cell death (7). Therefore, an area of intense investigation has been focused on whether autophagy-dependent survival should be blocked in these conditions to enhance tumor-cell death.
Acute myeloid leukemia (AML) is characterized by accumulation of abnormal hematopoietic progenitor cells and in ltration of blood, bone marrow, and other tissues (8). The inhibited cell differentiation can prevent the maturation of the progenitor cells, and subsequently prevent undergoing apoptosis (9). The hypoxic bone marrow niche in AML contributes to chemotherapy resistance, and subsequently allows persistence of minimal remaining malignant cells after chemotherapy, and nally leads to relapse. While the hypoxic and metabolically reprogaremmed niche are in the bone marrow, contact with other cellular elements of the bone marrow is anticipated to activate autophagy in leukemia cells (10).
Recent evidence revealed that impaired autophagy could alter hematopoietic stem cells (HSCs) to cells with preleukemic state phenotypes (11). In addition, myeloid leukemia have reported that leukemic cells require functional autophagy during leukemia maintenance (12,13). On the other hand, other studies have indicated that the autophagy machinery in AML might be disrupted, due to the accumulation of damaged mitochondria and increased levels of reactive oxygen species (ROS), which could potentially promote the leukemic transformation (14,15). Moreover, autophagy can be utilized as an escape mechanism by leukemic cells after treatment with chemotherapeutics such as HDAC and mTOR inhibitors (16,17).
The predominant form of retinoids, which is all-trans retinoic acid (ATRA), has been successfully used to treat acute promyelocytic leukemia (a particular AML subtype, which is known by the translocation t(15,17) encoding for the oncogene-retinoic acid receptor alpha (PML-RARA) fusion protein), which induces neutrophil differentiation of leukemic blast cells (21,22). ATRA and ATO is curative and standard regimen in the treatment of acute promyelocytic leukemia (APL). ATO and ATRA were effective in both the mouse and human model of AML cells, and their combination evidenced powerfully and precious treatment for AML subtypes in vitro. According to these data we assume that ATO, ATRA and their combination could be the possible therapeutic strategy which target different susceptibilities for other subset of AML (23).
Although different basic studies and clinical trials evaluating effects of autophagy modulators in cancers have been carried out, there is a lack of information which settings in autophagy pathway should be activated or inhibited to improve current treatment options in leukemia (24). The aim of our research is to study a possible role of autophagy and its therapeutic importance during treatment with ATO/ATRA regime in KG1 and HL-60 leukemia cell lines. We found that ATO/ATRA treatment induces autophagy in these cells and autophagy blockage at the late stage, particularly by means of hydroxyquinoline (HCQ), could enhance chemotherapy effects in KG1 cells.

Cell culture
The human acute myeloid leukemia cell lines (KG-1 and HL-60) were obtained from the National Cell Bank of Iran at Pasteur Institute of Iran. KG-1 and HL-60 were grown in suspension using RPMI 1640 medium containing 10% heat-inactivated fetal bovine serum, 100 mg/mL streptomycin, and 100 U/mL penicillin at 37 °C in a humidi ed atmosphere of 95% air and with 5% CO 2 . The passage number for all tests was between 3 and 5.

Proliferation Assay
The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to evaluate inhibition of cell proliferation potential of ATO, ATRA Rna Isolation And Real-time Pcr KG-1 and HL-60 cell lines were seeded at a density of 3 × 10^5 cell/well in 12 well culture plates, and were then incubated for 48 hours in the absence and presence of ATO, ATRA, 3-MA, HCQ, Baf-A1, and their combinations for 48 h. RNA was extracted from KG-1 and HL-60 cells using the TRI pure (Roche Applied Science, Germany), with respect to the manufacturer instructions. The RNA pellets were reconstituted in DEPC treated water. The quantity and quality of total RNA were speci ed by spectro-photometrically using Nanodrop ND-1000 (Nanodrop Technologies, Wilmington, DE) at 260 and 280 nm. First-strand complementary DNAs (cDNA) synthesis were oppositely transcribed from 1-2 µg of total RNA by the use of cDNA synthesis kit (Takara Bio Inc., Otsu, Japan) according to the manufacturer instructions. Real-time PCR was performed using a standard SYBR Green PCR protocol on Light Cycler 96 instrument (Roche). Then, the cDNA concentration was normalized in a series of PCR using HPRT and GAPDH primers. Also, the normalized cDNAs were subjected to ampli cation. The levels of HPRT mRNA expression were used to estimate the relative expression levels. The comparative mean cycle threshold value (Ct) method was used to compute relative expression values. The primers and their corresponding amplicon lengths are presented in Table 1.

Detection Of Acidic Vesicular Organelles During Autophagy Formation
The development of acidic vesicular organelles (AVOs) is one of the hallmarks of autophagy, which can be determined by acridine orange (AO) staining of the cells. KG-1 and HL-60 cells staining with AO (100 µg/ml) were accomplished based on the published procedures (32). Cells were resuspended by PBS, and then by adding AO. The uorescence of AO shifts from green to red within acidic components of the cells. Photos were obtained using a uorescence microscope (559-nm, Axioscop, Carl Zeiss, Germany).

Western Blotting Analysis
Cells were treated to ATO, ATRA, 3-MA, HCQ, Baf-A1, and their combinations for the selected concentrations and time periods. Treated cells were washed with ice-cold PBS, and lysed in ice-cold cell RIPA lysis, extraction buffer (CMG Co, Esfehan, Iran), and a cocktail of protease inhibitors (CMG Co, Esfehan, Iran). Protein concentration was quanti ed using the Bradford (1976) assay method. The Equal amounts of total extracted cellular proteins (50 mg) were used for SDS polyacrylamide gel electrophoresis and subjected to western blotting by polyvinylidene di uoride membranes (Whatman, UK). Subsequently, the transferred membranes were blocked by non-fat dry milk powder (5%) in Tris-Buffered Saline and 0.05% Tween-20 (TBST) for 90 min at room temperature, and were then incubated with the primary antibodies anti-LC3B (purchased from Abcam), and anti-GAPDH and anti-SQSTM1/P62 (purchased from Santa Cruz Biotech, USA). All the antibodies were incubated in 2% skim milk at 4 °C overnight. Membranes were washed three times in TBST (0.2%), and were then incubated for 2 h at room temperature, and anti-rabbit secondary antibody was used for LC3II (Sigma-Aldrich), and anti-mouse IgG secondary antibody for GAPDH (purchased from Santa Cruz Biotech, USA).

Transmission Electron Microscopy
For transmission electron microscopy (TEM) measurement, each sample was xed with 2% glutaraldehyde, and was then post-xed with 1% Osmium tetroxide (stock solution 1-2%) dissolved in phosphate buffer (0.1 M). After that, pure and fresh resin was used for embedding and polymerizing. Specimens were stained with 2% aqueous uranyl acetate, and lead citrate to contrast staining. Finally, all of the thin samples were observed by TEM (Philips EM208S) at an acceleration voltage of 100 kV.

Statistical analysis
Statistical analysis was performed by GraphPad prism 5 and R programming. The data were expressed as the mean ± standard deviation (SD). The difference between two mean values was analyzed using one-way ANOVA, and was considered to be statistically signi cant de ned at *P < 0.05, **P < 0.01, and ***P < 0.001.

Result
The cytotoxic effects of ATO and ATRA in KG1 and HL-60 cells we evaluated the effect of different concentrations of ATO and/or ATRA, alone or in combination with autophagy inhibitors, on proliferation and viability of KG-1 and HL-60 cells after 24 h and 48 h. The 50% inhibitory concentrations (IC50) of ATO and IC30 of ATRA were determined using MTT assay (Fig. 1). For example, IC50 for ATO were 1.61 µM and 1 µM in KG-1 and HL-60 cells, respectively, whereas IC30 for ATRA was approximately 700 nM in both cells. In addition, our result indicated that combining ATO with ATRA further reduced the cell proliferation (up to 60%) in KG-1 and HL-60 cell lines, in comparison to each monotherapy (Fig. 1). Therefore, ATO/ATRA is more potent inhibitor of the viability of leukemia cells compared to each monotherapy.

The Autophagic Effects Ato/atra In Kg-1 And Hl-60 Cells
To see the importance of autophagy in ATO/ATRA treatment, we evaluate AVO formation by the lysosomotropic and pH-sensitive AO. The results showed that the red uorescent increased in both KG-1 and HL-60 cells treated with Baf-A1 and HCQ, indicating the development of AVOs (Fig. 2a). Also, The TEM results con rmed the increased level of autophagosomes in both cells exposed to ATO/ATRA (Fig. 2c).
Then, we examined the gene and protein expression levels of key ATGs and autophagy marker protein LC3B and P62 by qPCR and western blotting (Fig. 2b). Upon induction of autophagy, LC3B-I (19 kDa) is converted to LC3B-II (17 kDa), and then conjugated with phosphatidylethanolamine (PE) within autophagic membranes. Since the rate of LC3B correlates well with the number of autophagosomes, this features of the conversion of LC3 could be used to monitor autophagic ux (33). P62, also known as sequestosome 1 (SQSTM1) in humans, directly interact with LC3 to promote the formation of autophagosome. The occurrence of autophagy was further evaluated by the mRNA expression of typical ATGs by real-time PCR. We observed the up-regulated the mRNA expression of BECN1, LC3B, P62, ATG12 and ULK in KG-1 ATO/ATRA treated cells compared with control cells (Fig. 3a). After treatment of KG-1 cells for 48 h, we observed that ATO/ATRA with a combination of HCQ, up-regulated the mRNA expression of LC3B, P62, CASP3, ATG12, and ULK; whereas they down-regulated BECN1 and ATG7 in comparison with control. After treatment of HL-60 cells for 48 h, we observed that ATO/ATRA with a combination of HCQ, up-regulated the mRNA expression of LC3B, P62, CASP3, ATG12, and ATG7; whereas they down-regulated BECN1 and ULK in comparison with control. Moreover, combination of BafA1 with ATO/ATRA up-regulated the mRNA expression of BECN1, LC3B, P62, CASP3, ATG12, and ULK; however, they down-regulated ATG7 in KG-1 cells compared to control (Fig. 3b). While 48 h treatment with ATO/ATRA increased gene expression levels LC3B, we observed no signi cant change at the protein levels of LC3B in both cells. However, the protein levels of LC3B and P62 were clearly increased after combination of ATO/ATRA with autophagy inhibitor HCQ (Fig. 2b). In addition, TEM con rmed that after treating KG-1 and HL-60 cells with selected doses of ATO/ATRA, the cells displayed the accumulation of autophagosomes after 48 h. Additionally, other morphological futures such as condensation of chromatin, a common nuclear morphology of apoptosis, was also observed in KG-1 cells (Fig. 2c)

Effects of ATO/ATRA in combination with autophagy inhibitors on proliferation and cell cycle pattern of KG-1 and HL-60
We next evaluated changes in proliferation and cell cycle pattern after treatment with ATO/ATRA either alone or in combination with the autophagy inhibitors 3-MA, HCQ and BafA1. As shown in Fig. 4 in combination with ATO/ATRA is generally followed by the cell-cycle arrest. We observed an interesting pattern for G0 and G1 in HL-60 cells. ATO/ATRA could decrease G0 and the cells enter to G1 but for KG-1 cells both the rate of G0 and G1 increased after treatment by ATO/ATRA.

Effects of ATO/ATRA in combination with autophagy inhibitors on apoptosis in KG-1 and HL-60 cells
To determine the mode of cell death induced by ATO/ATRA, we studied the occurrence of apoptosis as a typical type of programmed cell by Annexin/PI double staining and CASP3 expression. We observed a profound induction of apoptosis in both cells, and also the level of mRNA expression for CASP3 signi cantly increased in cells treated with Baf-A1/ATO/ATRA and HCQ/ATO/ATRA in both KG-1 and HL-60 cells (Fig. 3). Our data exhibited that, HCQ in combination with ATO/ATRA could signi cantly increase apoptotic (75.32 + 1.72%), when compared with control (8.64 + 0.26%) and ATO/ATRA (54.58 + 0.52%). Baf-A1 in combination with ATO/ATRA increased apoptotic (61.3 + 1.72%) when compared with control and ATO/ATRA. 3-MA in combination with ATO/ATRA slightly (60 + 1.2%) increased apoptotic, in comparison with ATO/ATRA, and no signi cant necrotic cells were in the groups exposed to ATO/ATRA (10.8% + 0.65%) alone or in association with 3-MA (7.5 + 0.94), HCQ (6.77 + 0.36), and BafA1 (5.09 + 0.52) in KG-1 cells. Thus, our data demonstrated that HCQ and BafA1 in combination with ATO/ATRA were more effective than 3-MA in KG-1 cell line (Fig. 6). These results demonstrate that the blockage of autophagy ux is more associated to apoptosis other than inhibition of autophagy induction (3-MA). In contrast with KG-1, we observed no detectable increase in apoptotic effects of ATO/ATRA in combination with autophagy inhibitors in HL-60 cells (Fig. 7).

Discussion
In this study, we showed a critical role of autophagy during treatment with the combination of ATO/ATRA. Also, we indicated that the synergistic effect of autophagy inhibitors HCQ, 3-MA, and Baf-A1 can enhance the bene cial outcomes of this therapeutic regime in leukemia cells. These inhibitors can block autophagy at different stages during the cytotoxic effects of ATO/ATRA. Therefore, it is important to investigate their unique mechanisms and molecular pathways in order to gain more insight into autophagy inhibitors while utilizing ATO/ATRA combinatorial approach.
Autophagic responses lead to the maintenance of homeostasis and adaptation to stress in normal and cancer cells (34). It has also been recognized as a common pro-survival mechanism among numerous tumor types. The association between autophagy and tumor cell survival can be explained via autophagy function in retaining cells from undergoing programmed cell death. This provides a logical rationale for autophagy inhibition promoting response to other agents (35). The results of published clinical trials exhibit evidence for the safe use of HCQ as cancer therapies (36,37). Although autophagy might be considered as functional in various cancer cells and may be needed for the response to stresses, some cancer cells may be speci cally associated with autophagy, even in the absence of added stress. This phenomenon is called autophagy-dependence, and is also crucial because it has been identi ed in some studies that only autophagy-dependent cancer cells is responding to autophagy inhibitors. Expectedly, drug synergy between pharmacological autophagy inhibitors and other anticancer compounds might exert its effects through an autophagy-dependent manner in cancer cells, and it is probable that the combination of identical agents was occasionally worked antagonistically in autophagy-independent cancer cells (35,38,39).
In hematological malignancies, autophagy acts as both tumor-suppressor and chemoresistance inducer. Nowadays, It is known that, based on both type of progenitor cells and the state of leukemia, autophagy might play contradictory roles(40). Different studies have shown that leukemic blasts and LSCs have the ability of applying autophagy to reply to the speci c energetic needs during exponential cell proliferation to neutralize chemotherapeutic stress as a survival mechanism. For instance, it has been con rmed that, in chronic myeloid leukemia (CML), patient-derived LSCs contain high levels of gene expression of basal autophagy compared to mature cells, and this targeting autophagy by pharmacological or genetic inhibition may result in reduced leukemic cell viability and improved sensitivity to common chemotherapy (41)(42)(43).
Our data indicated that the autophagic pathway involved in cell death mechanism in KG-1 during response to dual-agents treatment of ATO/ATRA. Combination of ATO/ATRA with HCQ or Baf-A1 inhibited cell proliferation, and induced apoptosis in KG-1 cells. Moreover, electron microscopy imaging con rmed rapid induction of cell death with autophagic features. The extent of cell death and phenotypes of HCQ and Baf-A1-induced autophagic markers such as an increase in microtubule-associated protein LC3-I conversion to LC3-II, BECN1, and autophagosome were considerably greater in KG-1 cells.
ATO and ATRA have major impacts on the treatment of patients with leukemia. However, AML cells can respond to ATO and ATRA in different ways ranged from activation of cell death to the initiation of survival pathways such as autophagy (44).
The relationship between apoptosis and autophagy is complicated in cancers. Apoptosis and autophagy could regulate each other, mostly during an inhibitory procedure. However, somewhen autophagy may favor the activation of apoptosis. Apoptosis may initiate with autophagy, and autophagy may end with apoptosis (45).
Different studies linked autophagy genes to programmed cell death. For instance, in human glioblastoma, blockage of autophagy genes ATG1 or ATG5 hindered the cytotoxic effect of cannabinoids, which can increase the autophagy-dependent cell death with mTORC1-dependent pathway (46).
In the central nervous system, ATG7 de ciency preserved neurons from caspase-dependent and caspaseindependent cell death after hypoxic-ischemic brain harm (47 appropriate GC-sensitive cell lines. However, they detected no effect of 3-MA in primary ALL cells, in which caspase-dependent cell death dominated. They identi ed autophagy as a primary and limiting step to steroid sensitization by obatoclax in GC-resistant cells, and emphasized on the importance of considering the cellular context during designing the strategies to target autophagy for cancer therapy (50). Kim  Folkerts et al. con rmed that AML CD34 + cells were more sensitive to the HCQ, as an autophagy inhibitor, compared to normal bone marrow CD34 + cells. In addition, inhibition of autophagy via knockdown of ATG5 or ATG7 operated apoptosis, which was occurred along with promoted expression of p53. They con rmed that targeting autophagy might provide a new therapeutic strategy for the treatment of AML(52).

Conclusion
This study shows the involvement of autophagy during treatment of KG1 and HL-60 cells with ATO/ATRA. Autophagy blockers may enhance the bene cial outcomes of this therapeutic regime in leukemia cells. Also, more studies are required to be taken for the purpose of guring out the critical contexts in which condition autophagy inhibition will be lucrative, and in which it could be detrimental.     Apoptosis assay in KG-1 after 48 h. Data are mean ± S.E of three independent experiments. Statistical signi cance were de ned at *P<0.05, **P<0.01 and ***P<0.001.