Background
Autophagy is a lysosomal-dependent process that occurs at low basal levels to support cellular homeostasis. During periods of nutrient deprivation, autophagy degrades intracellular proteins to serve as substrates for ATP generation. Autophagy also carries out housekeeping activities such as clearing the cell of damaged organelles and proteins that result from ordinary cellular metabolic activity. For example, damaged mitochondria are selectively targeted for autophagy, thus reducing the release of pro-apoptotic mediators into the cytosol and subsequent cell death [
1]. Therefore, basal levels of autophagy are necessary for cellular homeostasis.
Autophagic activity above basal levels (hereafter referred to as autophagy induction) is induced by anticancer drug treatment. While autophagy inhibition both increases anticancer drug efficacy [
2] and decreases anticancer drug efficacy [
3,
4], the majority of studies indicate that autophagy inhibition increases anticancer drug efficacy, suggesting that autophagy induction is a protective response to anticancer drug treatment. However, unrestrained drug-induced autophagy induction can lead to cell death [
5].
Osteosarcoma (OS) is the most prevalent bone tumor in children. Despite recent advances in the understanding of the molecular basis of OS and new therapeutic approaches, the mortality rate has declined only modestly. Autophagy modulation as adjuvant therapy to established anticancer therapies is currently being investigated in clinical trials, but not in OS [
6]. The use of autophagy modulation as adjuvant therapy in OS may prove beneficial. However, before considering such, the impact of anticancer drug-induced autophagy induction on cytotoxicity in OS must be better characterized.
In this study, we investigated the impact of autophagy inhibition on camptothecin (CPT)-induced cytotoxicity in OS cells. Camptothecin induces cell death by inhibiting topoisomerase I resulting in DNA single-strand breaks and subsequent cell death [
7]. Here, we show that autophagy inhibition has an opposing impact on CPT-induced cytotoxicity in two metastatic murine OS cell lines. Autophagy inhibition in K7M3 cells increased sensitivity to CPT. In contrast, autophagy inhibition in DLM8 cells decreased sensitivity to CPT. The mechanism of autophagy inhibition-mediated protection in DLM8 cells appeared to be reduced CPT-induced oxidative stress and a reduction in both mitochondrial damage and caspase activation. To our knowledge, this is the first report of an opposing effect of anticancer drug treatment on cytotoxicity in autophagy-inhibited OS cells. Furthermore, we were unable to locate any other report of autophagy inhibition decreasing anticancer drug-induced oxidative stress.
Methods
Antibodies and reagents
Camptothecin was purchased from ChemWerth (Woodbridge, CN). LC3 and ATG5 antibodies were purchased from Novus Biologicals (Littleton, CO). Cleaved PARP, total p53, phospho p53, cleaved caspase-3 and cleaved caspase-9 antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA). Pan-caspase inhibitor was purchased from Enzo Life Sciences (Farmingdale, NY). Ripa lysis buffer was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Acridine orange, 3-(4,5-dimethylthiazolyl-2)-2,5-diphenylthetrazolium bromide (MTT) reagent, Bafilomycin A1 and actin antibody were purchased from Sigma Aldrich (St. Louis, MO). Buthionine sulfoximine (BSO) was purchased from Acros Organics (Morris Plains, NJ). N-acetyl cysteine (NAC) was purchased from Calbiochem (Billerica, MA). Fetal bovine serum (FBS) was purchased from Atlanta Biologicals (Lawrenceville, GA). DMEM cell culture medium and supplements, dihydroethidium (HE), 2′,7′-dichlorofluorescein diacetate (DCFH-DA), carbonyl cyanide chlorophenylhydrozone (CCCP) and tetramethylrhodamine, ethyl ester (TMRE) were purchased from Invitrogen (Carlsbad, CA).
Cell lines and cell culture
DLM8 [
8] and K7M3 [
9] are metastatic murine OS cell lines. MC3T3 is a non-transformed murine osteoblast cell line [
10]. Cells were cultured in Dulbecco’s modified eagle medium (DMEM) containing 10% FBS supplemented with antibiotic, non-essential amino acids, glutamine, sodium pyruvate and cultured in an incubator maintained at 5% CO
2 and 37°C. Prior to experimentation, cells were karyotyped and tested for mycoplasm contamination. Cells were treated with CPT as indicated in figure legends. Treatments were based on sensitivity of each cell line to CPT. Where appropriate, cells were treated with BSO to induce oxidative stress and NAC to counter oxidative stress.
Lentiviral shRNA (Open Biosystems, Rockford, IL) targeted to autophagy-related protein-5 (ATG5) RNA was used to knockdown ATG5 protein expression. Two separate ATG5 knockdown cell lines were generated for each cell line using two different lentiviral shRNA sequences [TRCN0000099431, TRCN0000099433]. Briefly, lentivirus was produced by transfecting 293T cells with 7 ug/ml transfer plasmid [TRCN0000099431 or TRCN0000099433], 5 ug/ml psPAX2 (packaging plasmid) and 4 ug/ml pMD2. G (envelop plasmid). Forty-eight hours after 293T cell transfection, supernatant containing lentivirus was collected and immediately used for infection or stored at -80°C. For infection, 2 ml of supernatant containing lentivirus was added to each well of a 6-well plate containing 1×105 cells. Cells were incubated with lentivirus for 12 h and next transferred to a 75 mm flask. Assessment of ATG5 protein knockdown was determined when cells were approximately 70% confluent. Both ATG5 knockdown cell lines showed similar responses to CPT treatment. Control cells were infected with lentivirus containing empty shRNA vector. Cells treated with empty shRNA vector are hereafter referred to as autophagy-competent. ATG5 protein knockdown cells are hereafter referred to as autophagy-inhibited.
Cell growth was determined by seeding 4×104 cells per well in a 12-well plate followed by cell count at 48 h. Metabolic activity was assessed by MTT assay. Metabolic activity converts yellow MTT reagent to a purple formazan. Color intensity is indicative of metabolic activity. MTT reagent (1mg/ml) was added to cells (3×103 cells per well) and incubated for 1 h at 37°C followed by solubilization of formazan with DMSO followed by determination of formazan color intensity with a microplate reader set at absorbance reading 570 nm. Absorbance readings of autophagy-inhibited groups were compared to autophagy-competent groups which were normalized to one hundred percent. To determine cell viability, 4×104 cells per well were seeded in 12-well plates. Following CPT treatment, cell viability was determined by trypan blue exclusion assay using an automated cell counter (Vi-Cell, Beckman Coulter, Miami, FL). Cells restricting trypan blue entry were considered viable.
Acidic vesicular organelle (AVO) staining
Acridine orange freely diffuses the membranes of cells and organelles. Inside acidic vesicles, acridine orange is protonated and fluoresces bright red. Increased red fluorescence indicates increased acidic vesicular organelle (AVO) formation [
11]. Following CPT treatment, cell culture medium was removed from the cells and replaced with cell culture medium containing 1ug/ml acridine orange and incubated for 20 min at 37°C. Cells were then removed, washed twice and fluorescence immediately analyzed using the FL3 channel of a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA).
Western blot
Following drug treatment, supernatant and cells were collected and centrifuged at 300 g for 5 min at 4°C. The resultant pellet was lysed with RIPA lysis buffer containing protease and phosphatase inhibitor cocktail and centrifuged at 10,000 g for 10 min at 4°C. Supernatants were then collected and total protein was determined by BioRad reagent (BioRad Laboratories, Hercules, CA). Unless otherwise indicated, 30 ug of protein were resolved in SDS-polyacrylamide gels (SDS-PAGE) and transferred onto nitrocellulose membranes (BioRad Laboratories, Hercules, CA). Membranes were blocked with 5% nonfat milk then incubated with antibodies against ATG5, LC3, cleaved caspase-9, cleaved caspase-3, total p53, phospho p53 or cleaved PARP. Membranes were then washed and incubated with appropriate secondary antibody conjugated to HRP (GE Healthcare Life Sciences, Piscataway, NJ). Following secondary antibody incubation, membranes were washed and signal detected with ECL detection reagent (GE Healthcare Life Sciences, Piscataway, NJ). Beta-actin protein expression served as a protein loading control.
Oxidative stress determination
Following drug treatment, cell culture medium was removed from the cells and replaced with cell culture medium containing 5 uM dihydroethidium (HE) or 5 uM 2′,7′-dichlorofluorescein diacetate (DCFH-DA) and incubated for 20 min at 37°C to assess superoxide anion (.O2
-) and hydrogen peroxide (H2O2) levels, respectfully. Cells were then removed, washed twice and fluorescence immediately analyzed using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA). HE freely diffuses the plasma membrane and is reduced by intracellular .O2
- resulting in a red fluorescence. Intracellular DCFH-DA reacts with H2O2 to give a green fluorescence. Increased HE and DCFH-DA fluorescence indicates increased .O2
- and H2O2 presence, respectively.
Mitochondrial membrane potential (ΔΨm)
Tetramethylrhodamine ethyl ester perchlorate (TMRE) preferentially stains mitochondria producing red fluorescence and is used as an indicator of mitochondrial membrane potential (ΔΨm). Decreased TMRE fluorescence is indicative of ΔΨm depolarization and ΔΨm depolarization is associated with release of pro-apoptotic mediators [
12]. Following CPT treatment, cells were incubated with 25 ng/ml TMRE for 20 min at 37°C to assess ΔΨm. The protonophore carbonylcyanide m-chlorophenylhydrozone (CCCP) was used as a positive control for ΔΨm depolarization and to test TMRE staining efficiency.
Statistical analysis
Results are presented as means ± S.E.M. Experimental data were analyzed using 2-tailed Student t test. P values less than 0.05 were considered statistically significant.
Discussion and conclusions
The protective role of autophagy induction against anticancer therapy is supported by observations that autophagy inhibition increases anticancer drug efficacy [
2]. A literature search returned a limited number of studies reporting reduced anticancer therapy efficacy in autophagy-inhibited cells [
15]. With autophagy inhibition currently being investigated as adjuvant anticancer therapy, these limited observations are relevant. In this study, ATG5 protein expression was knocked down to inhibit autophagy. Here, we report an opposing effect of ATG5 knockdown-mediated autophagy inhibition on CPT-induced cytotoxicity within OS. Autophagy inhibition decreased sensitivity to CPT in DLM8 cells and increased sensitivity to CPT in K7M3 cells. To date, there are no reports showing an opposing impact of autophagy inhibition on anticancer therapy within OS.
Following the observation that autophagy inhibition in K7M3 cells increased sensitivity to CPT, we reasoned that autophagy plays a greater role in the overall maintenance and metabolic homeostasis in K7M3 cells and suspected that the basal level of autophagy in K7M3 cells is greater than that of DLM8 cells. Immunoblot analysis of LC3II confirmed that basal level of autophagy is higer in K7M3 cells compared to DLM8 cells and non-transformed murine MC3T3 osteoblasts (Figure
4E). This finding supports the suggestion that K7M3 cells have an increased dependence on autophagy for ordinary metabolic activities. The dependence of K7M3 on autophagy is further supported by the observation that autophagy inhibition significantly decreased both K7M3 cell metabolic activity and cell growth (Figure
4B and C). It is plausible that increased basal level of autophagy in K7M3 cells is one of several genetic influences that contribute to the cancer phenotype and decreased autophagic capability increases sensitivity to stresses such as anticancer treatment. Increased dependence on autophagy has been reported for other cancers. For example, pancreatic cancer cells [
16] and Ras oncogenic-driven cancer cells [
17] have been shown to have increased dependence on autophagy. These two studies also reported increased basal levels of autophagy.
In this study, autophagy inhibition decreased sensitivity to CPT in DLM8 cells, which contrasts the more often reported observation that autophagy inhibition increases sensitivity to anticancer drug treatment. Therefore, we were particularly interested in the response of autophagy-inhibited DLM8 cells to CPT and explored further this cell line. While it was clear that autophagy inhibition in DLM8 cells decreased CPT-induced cell death compared to autophagy-competent DLM8 cells, the mechanism was unknown. Considering that the mechanism of action for CPT is DNA damage [
18], we explored the impact of autophagy inhibition on CPT-induced DNA damage as a possible mechanism for decreased sensitivity to CPT in autophagy-inhibited DLM8 cells. DNA damage as determined by phosphorylation of p53 at Ser15 [
19] was unchanged between autophagy-competent and autophagy-inhibited DLM8 cells (Figure
4F). We also assessed the impact of autophagy inhibition on DLM8 cell growth. Autophagy inhibition did not significantly impact cell growth of DLM8 cells (Figure
1C). This is relevant because the mechanism of action for CPT is DNA damage that occurs during cell division. Had autophagy inhibition significantly reduced DLM8 cell growth, this would support the suggestion that autophagy inhibition-mediated protection is due to reduced cell division. Together, this set of data suggests that the autophagy inhibition-mediated protection observed in this study was not due to reduced DNA damage or reduced cell division.
Previous reports of CPT-induced oxidative stress [
20] led us to investigate the impact of autophagy inhibition on CPT-induced oxidative stress as a contributing factor to the observed autophagy inhibition-mediated protection. Oxidative stress, as determined by generation of
.O
2
- and H
2O
2, was higher in autophagy-competent DLM8 cells compared to autophagy-inhibited DLM8 cells following CPT treatment, indicating that autophagy inhibition decreased CPT-induced oxidative stress. Autophagy inhibition also reduced basal oxidative stress level. To our knowledge, this is the first report of autophagy inhibition-mediated reduced basal oxidative stress as well as autophagy inhibition-mediated reduced anticancer drug-induced oxidative stress.
Increased levels of CPT-induced oxidative stress coupled with increased CPT-induced cell death in autophagy-competent DLM8 cells led us to determine if autophagy-competent DLM8 cells are more sensitive to oxidative stress. The use of BSO allowed for the investigation of the impact of oxidative stress alone on cell death and autophgay induction. Autophagy-competent DLM8 cells were more sensitive than autophagy-inhibited DLM8 cells to BSO-induced cell death. In agreement with Martinez-Outschonnra et al. [
21], BSO also induced autophagy. BSO-induced cell death and BSO-induced autophagy induction were reversed by NAC pretreatment indicating a link between increased oxidative stress and both cell death and autophagy induction.
Basal levels of autophagy have been previously reported to differ among cancer cell lines [
17] and here we report varying basal levels of autophagy in two metastatic murine OS cell lines. Considering these reports, it is plausible that the threshold level of autophagy induction that causes autopahgic cell death also varies for different cancers or even different cell lines within the same type of cancer. Camptothecin-induced DNA damage and CPT-induced oxidative stress together may have caused autophagy induction in autophagy-competent DLM8 cells that exceeded the threshold level necessary to cause autophagic cell death. Considering this, autophagy inhibition would reduce or delay CPT-induced autophagic cell death, making autophagy-inhibited DLM8 cells less sensitive to CPT-induced cell death. Therefore, one explanation for decreased sensitivity of autophagy-inhibited DLM8 cells compared to autophagy-competent DLM8 cells is reduced CPT-induced autopahgic cell death. Camptothecin-induced oxidative stress was lower in autophagy-inhibited DLM8 cells compared to autophagy-competent DLM8 cells. Therefore, an alternative explanation for decreased sensitivity to CPT in autophagy-inhibited DLM8 cells is reduced oxidative stress-induced cell death unrelated to autophagic cell death. At this point in our investigation, we are unable to present data supporting an explanation for lower oxidative stress in autophagy-inhibited DLM8 cells. However, the endogenous antioxidant catalase is a reported target of selective autophagy [
22] and we suspect that autophagy inhibition may affect levels of endogenous antioxidants.
With observed CPT-induced oxidative stress in this study and reports of oxidative stress induced mitochondrial damage [
23], we investigated the impact of CPT on mitochondria. In agreement with a previous study [
24], CPT caused ΔΨm depolarization in both autophagy-competent and autophagy-inhibited DLM8 cells (Figure
6A). However, ΔΨm depolarization was greater in autophagy-competent DLM8 cells, suggesting increased mitochondrial damage. Mitochondrial membrane potential depolarization and mitochondrial damage is associated with caspase-9 activation and caspase-3 activation [
25]. Immunoblot confirmed increased caspase-9 activation and caspase-3 activation in autophagy-competent DLM8 cells compared to autophagy-inhibited DLM8 cells (Figure
6B). Thus, the observed mitochondrial damage was likely an upstream event of caspase activation and likely contributed to increased cell death in autophagy-competent cells. Conversely, caspase-3 activation was higher in autophagy-inhibited K7M3 cells compared to autophagy-competent cells (Figure
6C). This observation suggests that autophagy inhibition increased the sensitivity of K7M3 to CPT-induced apoptosis.
In conclusion, we show that autophagy inhibition can have an opposing impact on the response of OS cells following CPT treatment. Our data suggest that the protective mechanism of autophagy inhibition involves both reduced oxidative stress and mitochondrial damage. The results of this study reminds us that autophagy inhibition can decrease the efficacy of anticancer drug therapy and underscores the need to better understand and predict the response of autophagy-modulated cancer cells to anticancer drug therapy.
Competing interests
The authors declare that they have no competing interest.
Authors’ contributions
MGH conceived the study, carried out experiments, carried out data analysis and wrote the manuscript. NG assisted with project conception. JMS assisted with experiments and preparation of manuscript. ESK served as project supervisor. All authors read and approved the final manuscript.