Background
Autophagy is a normal physiological mechanism for cellular homeostasis, whereby damaged or defective cellular components, including mitochondria [
1], the endoplasmic reticulum [
2] and peroxisomes [
3], are digested within the cell by fusion with lysosomes. A basal level of autophagy occurs in all cells to ensure that only functional organelles are retained [
4]. Autophagy can also be induced by external stressors such as growth factor deprivation [
5] or upon exposure to specific chemotherapy drugs [
6‐
11]. The ultimate fate of cells under stress depends upon the net effect of apoptotic versus survival signals, often regulated by important cellular regulatory proteins such as Bcl2 and p53 [
12,
13]. Under nutrient limiting conditions, autophagy permits cells to survive by metabolizing their own organelles as a source of energy. However, this survival mechanism is considered a “double-edged sword”, since cells can also die by prolonged autophagy (also referred to as type II programmed cell death) [
14‐
16].
A number of investigations suggest that autophagy induction can promote resistance to cell death within tumor cells and has important implications for resistance to chemotherapy in cancer treatment [
17]. For instance, up-regulation of autophagy by the drug rapamycin can protect several tumor cell lines from cell death through apoptosis [
18]. In addition, the DNA damaging agents temozolomide and etoposide were found to induce an autophagy-associated increase in ATP production in multiple glioma cell lines, which protects the cells from death, possibly contributing to resistance to these drugs [
19]. Activation of autophagy was also observed when growth factors were withdrawn in apoptosis-deficient cells [
20]. It has also been suggested that autophagy induction may be associated with imatinib resistance in mouse lymphoid cells [
11]. However, a definitive role for autophagy in acquired resistance to cytotoxic chemotherapy drugs, including a temporal association between acquired drug resistance and autophagy induction, has yet to be demonstrated.
It has been previously demonstrated that most of the weakly basic chemotherapeutic drugs, such as DNA-binding anthracyclines, can accumulate in lysosomes, especially in drug resistant cells [
21‐
25]. Therefore, sequestration of chemotherapy drugs in lysosomes is widely considered to be a
bona fide mechanism of resistance to weakly basic chemotherapy drugs in cancer cells. The use of lysosomotropic agents to restore the sensitivity of drug-resistant cells to chemotherapeutic drugs has been widely investigated [
26,
27], as reviewed by Agostinelli [
28] and Kaufmann [
29]. As these agents inhibit vacuolar H
+-ATPase [
30] or change lysosomal membrane permeability [
31‐
33], they would be expected to block the accumulation of weakly basic chemotherapy drugs in lysosomes. Lysosomotropic agents such as chloroquine have recently been shown to promote the ability of the chemotherapy drug paclitaxel to kill cancer stem cells through the inhibition of autophagic survival [
34].
In this study, we investigated the role of autophagy and lysosomal drug sequestration in the acquisition of doxorubicin resistance in MCF-7 breast tumor cells. This involved the study of a panel of MCF-7 cells developed in our laboratory, whereby MCF-7 breast tumor cells were selected for survival in increasing concentrations of doxorubicin (MCF-7
DOX2 cells). Aliquots of cells were retained at each doxorubicin dose elevation. These cells do not express several drug transporters associated with doxorubicin resistance in vitro, including Abcb1, Abcc2, or Abcg2. We did, however, observe elevated expression of the Abcc1 protein at the highest doxorubicin selection dose (dose 12) [
35]. Using this panel of cell lines, we show in this study a strong temporal association between the acquisition of doxorubicin resistance and both the induction of autophagy and the sequestration of doxorubicin into lysosomes. We further provide evidence suggesting that the autophagy associated with doxorubicin resistance is distinct from starvation-induced autophagy. Blockage of this autophagy mechanism may represent a novel approach to cancer therapy, in particular for treatment of recurrent disease after prior chemotherapy administration.
Methods
This study did not require ethics approval from an ethics review committee or board because the study did not involve animals, humans, human data, or material collected from humans or animals.
Maintenance of MCF-7 cells and establishment of drug resistant variants
Human MCF-7 breast cancers cells (lot HTB-22, American Tissue Culture Collection) were grown in Dulbecco’s H21 medium (Princess Margaret Hospital, Toronto, ON) containing 10 % fetal bovine serum (FBS) (Hyclone), and incubated at 37 °C in a humidified 5 % CO
2 atmosphere. Doxorubicin-resistant MCF-7
DOX2 cells were generated in our laboratory by selecting MCF-7 cells for resistance to increasing concentrations (doses) of doxorubicin (PFS, USP, Pfizer), as described previously [
35]. The passage numbers for the doxorubicin-resistant MCF-7 cell lines at selection doses 8 through 12 are 203, 216, 220, 227 and 257, respectively. As controls, parental MCF-7 cells were identically “selected” in the absence of drug to identical passage numbers. These cell lines are referred to as “co-cultured control cell lines” for the above selection doses and help to account for any changes in drug sensitivity or other cell phenotypes associated with increased passage in culture. All of the cells used in experiments were not subcultured for more than 10 passages after thawing from frozen stocks. The parental cell line (MCF-7) has been authenticated by the American Tissue Culture Collection and all cell lines are free of mycoplasma contamination.
Measurement of cellular drug sensitivity
The sensitivity of cells to doxorubicin at various doxorubicin selection doses was measured using a clonogenic assay as described previously [
36]. The concentration of doxorubicin at which the number of colonies formed in the assay was reduced by 50 % (the IC50) was computed for both MCF-7
CC and MCF-7
DOX2 cells. The degree of drug resistance in MCF-7
DOX2 cells (the resistance factor) was then determined by dividing the IC50 value for MCF-7
DOX2 cells (at that selection dose) by the IC50 value for MCF-7
CC cells (at a similar passage number). As an example of cell nomenclature, MCF-7
DOX2–10 and MCF-7
CC10 cells represent cells selected in the presence of doxorubicin to dose level 10 and their co-cultured control cells at that selection dose, respectively.
Localization of cellular organelles and proteins by confocal microscopy
The location of mitochondria, lysosomes, and autophagosomes (and their possible co-localization with doxorubicin) in MCF-7CC10 and MCF-7DOX2–10 cells were investigated using laser scanning confocal microscopy after staining with MitoTracker® (red fluorescence) or LysoTracker® (green fluorescence), both from Molecular Probes, Thermo Fisher Scientific), or monodansyl cadeverine (MDC) from Sigma-Aldrich Chemicals, respectively. All images were obtained by confocal miscroscopy (model 510 Meta, Carl Zeiss, Toronto, ON) using lasers at specific wavelengths or under UV light. The location of doxorubicin was also visualized through confocal microscopy, since the drug naturally has red fluorescence. For the above experiments, MCF-7CC10 and MCF-7DOX2–10 cells were grown on glass coverslips for 2 days in the absence of drug in a 6 cm tissue culture plate covered with tissue culture medium. The cells were then treated with 2 μM doxorubicin [alone or together with 10 μM chloroquine (Sigma-Aldrich Chemicals)] for 8 and 48 hours for MCF-7CC10 and MCF-7DOX2–10 cells, respectively. The concentration of doxorubicin chosen for these experiments, while much higher than necessary for cytotoxicity, was the minimum concentration that permitted reliable visualization of the drug’s location in cells by its fluorescence. After staining with 50 nM LysoTracker® (using the manufacturer’s protocol) or staining with 50 μM MDC for 10 min, the coverslips were washed briefly in PBS and mounted on glass slides for examination by laser scanning confocal microscopy. All the images were taken using the same parameters for accurate comparison between treatments in one particular experiment, with the aid of an Argon laser excitation at 488 nm and emission at 560 nm (with a LP filter for doxorubicin fluorescence) and excitation at 458 nm and emission at 505–530 nm (with a BP filter for LysoTracker® fluorescence). Cells were stained with MDC after LysoTracker® staining. The filters for obtaining images by laser scanning microscope were set to Fset01 (blue), Fset17 (green) and Fset28 (red) and excited with UV light.
The locations of autophagosomes and lysosomes were also assessed by immunofluorescence using antibodies against LC3 (Cat# 2775, Cell Signaling) and LAMP1 (Cat# Sc-20011 H4A3, Santa Cruz), respectively, in an approach similar to that described by Nakagawa et al. [
37]. According to the manufacturer, the former antibody preferentially binds to LC3 conjugated to phosphatidylethanolamine (LC3-II), which is recruited to autophagosomal membranes [
38]. Hence, it is highly useful for visualizing the location of autophagosomes. Mitochondria were visualized by laser scanning confocal microscopy after staining with MitoTracker™ (Thermo Fisher). Both MCF-7
DOX2–10 and MCF-7
CC10 cells were grown on glass coverslips with or without prior treatment with 50 nM bafilomycin A1 for 24 h (to block the turnover of LC3-II, once formed). The cells on coverslips were fixed with 4 % formaldehyde in PBS. Immunocytochemical staining of cells with either anti-LC3 or anti-LAMP1 antibodies was performed as described by the manufacturer (Cell Signaling). Goat anti-rabbit IgG-TR (Cat# sc-2780, Santa Cruz) and goat anti-mouse IgG-FITC (Cat# sc-2010, Santa Cruz) secondary antibodies were used for detection of the LC3 or LAMP1 antibodies. To confirm the consistency of LC3 and LAMP1 staining, two fluorophores were switched between two secondary antibodies for the two color staining. For assessment of the location of mitochondria and autophagosomes, cells were labeled with 50 nM MitoTracker™ for 15 min prior to fixation. Immunostaining was then performed using the anti-LC3-II and goat anti-rabbit IgG-FITC antibodies. The settings for fluorescence detection were the same as described above. Images chosen were highly representative of cells views in a minimum of 5 fields (5–10 cells per field) from duplicate slides obtained in 2 to 3 independently performed experiments.
Transmission electron microscopy
For visualization of cellular ultrastructure by electron microscopy, cells were grown in 10 cm petri dishes to about 70 % confluence, after which the cells were released by trypsin treatment, washed once with PBS, and harvested by centrifugation. The cell pellet was resuspended in 10 ml of ice cold 3 % glutaraldehyde fixative in 0.1 M sodium cacodylate buffer (pH 7.2) for 35 min at 4 °C. The cells were then collected by centrifugation and resuspended in 1 ml of ice cold 0.2 M sodium cacodylate buffer (pH 7.2). The samples were then sent to the University of Western Ontario with cold packs for embedding, sectioning, and visualization by electron microscopy.
Immunoblotting experiments using whole cell lysates
Whole cell extracts of cells were prepared in modified RIPA buffer containing 1 % NP40, 0.5 % sodium deoxycholate, 1 % SDS, and 1 Complete™ protease inhibitor tablet for every 50 ml of buffer. Cultured cells were grown as a monolayer for 2 days until cell density reached 50–60 % confluence in 10 cm tissue culture plates. Twenty-four hours prior to protein extraction, the cells were treated with or without 50 nM bafilomycin A1 for 24 h under standard mammalian cell culture conditions. The culture medium was removed and the cells rinsed twice with PBS. To each plate, 0.7 ml of chilled modified RIPA buffer was added. The cells were scraped from the plates using a cell lifter, transferred to a 1.5 ml microfuge tube, and passed through a 21 gauge needle repeatedly to ensure efficient cell lysis and to shear any DNA present. The protein concentration for the whole cell extracts was determined using the BCA protein assay reagent kit (Pierce). Protein samples (30 μg) from whole cell extracts were used for SDS-PAGE analysis on 12 % or 10 % polyacrylamide gels based on the molecular weight of the target protein. Electrophoresis, protein transfer and immunoblotting were performed using standard procedures.
Measurement of the degradation of long-lived proteins (flux assay)
The degradation of long-lived proteins was measured using a modification of the standard “flux assay” [
39]. Cells were seeded in 6-well plates for 48 h in DMEM-H21 medium with 5 % FBS in a humidified incubator with 5 % CO
2. When the cell density reached about 50–60 % confluence, the medium was replaced with fresh medium supplemented with 0.2 μCi/mL [14C (U)] L-valine (MC-277, Moravek Biochemicals) and incubated for 24 h at 37 °C. Unincorporated radioisotope was then removed by three PBS washes. Cells were then incubated with 10 mM unlabeled valine (Sigma-Aldrich) for 3 h to allow for short-lived protein turnover. The medium was then replaced with fresh medium containing 10 mM unlabeled valine in the absence or presence of 10 μM chloroquine or 1 μM rapamycin (R8781, Sigma-Aldrich) in order to inhibit or activate autophagy, respectively. After a 24 or 48 h incubation period, the medium was collected from the wells. The medium with some detached cells was mixed with BSA (5 mg/ml final concentration) and 10 % trichloroacetic acid (TCA; Sigma-Aldrich), after which proteins in the medium were allowed to precipitate at 4 °C for 30 min. The precipitated proteins (along with detached cells) were harvested by centrifugation at 600 × g for 10 min, leaving behind soluble radiolabeled proteins in the supernatant. The adherent cells remaining in the tissue culture flasks were also collected by scraping in 0.5 ml of 10 % TCA, after which proteins were allowed to precipitate at 4 °C for 30 min. The cells and precipitated proteins where then harvested by centrifugation at 600 × g for 10 min, again leaving behind soluble proteins in the supernatant. Fifty μl of the supernatant from cells and 222 μl of supernatant from the medium were combined and added to scintillation vials with 5 ml of scintillation fluid. This mixture represents the acid-soluble radioactivity from degraded proteins. The TCA-precipitated protein preparations from the detached cells in the medium and adherent cells were each solubilized in 500 μl of solubilization buffer (0.1 N NaOH + 0.1 % SDS). Fifty μl from each solubilized pellet was then added to scintillation vials with 5 ml of scintillation fluid. This mixture constituted the TCA-percipitable radioactivity from both detached and adherent cells. After allowing the vials to sit overnight at room temperature, the radioactivity in the vials was quantified by liquid scintillation counting (Beckman Coulter LS6500). The total protein degradation (% proteolysis) was measured by dividing the TCA-soluble radioactivity by the radioactivity in the precipitated proteins. For all experiments, values were reported as means ± S.D. (n = 3). Statistical differences between the two groups were determined by the Student’s
t-test with Sigma plot 11.0 for Microsoft Windows. An identical experiment without isotope labeling was performed for protein extraction and immunoblot analysis of LC3-II expression after 48 h treatment.
Inhibition of ATG7 expression by siRNA interference
MCF-7CC10 and MCF-7DOX2–10 cells were grown on 10 cm plates to 30–40 % confluence in antibiotic-free DMEM medium supplemented with 10 % FBS and left to adhere overnight. The next day, the culture media was removed and the cells were washed with PBS, after which 12 ml of Opti-MEM I media (Invitrogen) were added to each plate prior to cell transfection with Lipofectamine 2000 (Invitrogen), using the manufacturer’s instructions. Briefly, an ATG7-specific siRNA oligo (Ambion Silencer® Select) was added to 1.5 ml of Opti-MEM I medium at a 20 nM final concentration in 1 well of a 6-well plate. In a separate well, 30 μl of Lipofectamine 2000 was added to 1.5 ml of Opti-MEM I medium. After 5 min incubation, the two solutions were mixed (3 ml in total). An identical procedure was performed for a Silencer® Select negative control siRNA (Ambion). The mixture was incubated for 20 min, then added to cultures of the above cell lines. The plates were incubated at 37 °C for 24 h, after which the medium was removed and replaced with antibiotic-free DMEM medium, supplemented with 10 % FBS. At 48 h post-transfection, the cells were washed, trypsinized, counted, and plated for either a clonogenic assay or a flux assay (as described above). An aliquot of cells was retained from each transfection and proteins extracted (also as described above) in order to assess the efficiency of gene knockdown using immunoblotting experiments. The siRNA sequences used in the study are as follows: ATG7-1: 5′-GGAACACUGUAUAACACCAtt-3′ and 5′-UGGUGUUAUACAGUGUUCCaa-3′. ATG7-3: 5′-GAAGCUCCCAAGGACAUU-Att-3′ and 5′-UAAUGUCCUUGGGAGCUUCat-3′.
Discussion
While previous studies have suggested a link between autophagy and chemotherapy drug resistance [
57‐
61], a temporal association between the acquisition of chemotherapy resistance and induction of autophagy has yet to be established. Moreover, it is unclear how this relates to drug uptake and drug localization in drug-resistant cells. In this study, we report for the first time that the acquisition of doxorubicin resistance can be temporally correlated with both enhanced drug sequestration into clustered perinuclear lysosomes and enhanced autophagy. The induction of autophagy upon acquisition of drug resistance is associated with increased and decreased cellular p62 and Bcl-2 levels, respectively. Inhibition of autophagy by chloroquine promotes doxorubicin-induced cell death in MCF-7
DOX2–10 cells, but not in drug-sensitive MCF-7
CC10 cells.
It has been well established that LC3 is a reliable marker of the formation of autophagosomes in mammalian cells [
62]. Its localization within cells changes from a diffuse cytosolic pattern to a punctate pattern representing its recruitment to the autophagosomal membrane during the induction of autophagy [
63]. The findings of our study are consistent with this view, since MCF-7 cells selected for survival in increasing concentrations of doxorubicin exhibited increased levels of LC3-II and this increase was temporally associated with acquisition of doxorubicin resistance. Moreover, the location of autophagosomes (LC3-II) and lysosomes (LAMP1) changed upon selection for doxorubicin resistance from a diffuse pattern throughout the cytoplasm to being clustered in the perinuclear region (Fig.
3). Similar to our observations in MCF-7
DOX2 cells, lysosomal clustering and increased cellular LC3-II levels took place during independent selection of MCF-7 cells for acquired resistance to several other chemotherapy drugs, including an analog of doxorubicin (epirubicin), and both the taxanes paclitaxel and docetaxel. These changes took place at or above selection doses where drug resistance was obtained (data not shown). Taken together, these observations suggest that increased autophagy and/or sequestration of drugs in lysosomes are highly reproducible and common mechanisms through which tumor cells acquire resistance to cytotoxic chemotherapy drugs.
Doxorubicin may have at least four possible fates upon entry into MCF-7
DOX2–10 cells. In a prior study, we have provided evidence that doxorubicin may be metabolized by cytoplasmic aldo-keto reductases (AKRs) into a considerably less toxic metabolite (13-OH doxorubicinol) in breast tumor cells [
64]. Alternatively, the drug may be sequestrated into lysosomes (either as doxorubicin or its 13-OH metabolite), due to its properties as a weak base [
21,
65]. Thirdly, before reaching the nucleus, doxorubicin may bind to mitochondrial DNA and induce oxidative damage to mitochondria (due to the drug’s ability to generate ROS). This, in turn, may result in the activation of DNA damage response/survival pathways [
66]. Finally, we have previously provided evidence that at higher selection doses, doxorubicin may simply be actively effluxed from MCF-7
DOX2–10 cells through the induced expression of drug transporters such as Abcc1 [
61]. All of these mechanisms may explain why only a small amount of doxorubicin appears to be present in the nuclei of MCF-7
DOX2–10 cells (Fig.
2).
During selection for doxorubicin resistance, it would be expected that doxorubicin would bind to mitochondrial DNA, thereby exposing the organelles to ROS produced by doxorubicin [
66]. This may result in large numbers of damaged mitochondria, which would be targeted for degradation by activation of a particular form of autophagy (namely mitophagy). This would be consistent with observations of many vesicularized mitochondria in MCF-7
DOX2–10 cells (Fig.
4). In addition, activation of autophagy has been reported to increase cellular capacity to survive stress associated with exposure to ROS [
67]. Since canonical autophagy requires the involvement of all Atg proteins [
68] and since knockdown of Atg7 did not significantly reduce doxorubicin resistance, this suggests that acquisition of doxorubicin resistance may be associated with the induction of non-canonical autophagy [
9]. The mechanism for autophagy associated with selection for doxorubicin resistance may involve selective delivery of damaged organelles into autophagosomes that then fuse with lysosomes for hydrolytic degradation [
69,
70], even under nutrient-rich conditions. This form of non-canonical autophagy is often referred to as selective autophagy.
It has been suggested that p62, as a selective cargo receptor, is involved in linking ubiquitinated protein aggregates to the autophagy machinery through LC3 [
52,
54]. In addition, p62 mediates the clustering and aggregation of dysfunctional mitochondria and binds to LC3-II to deliver aggregated mitochondria to autophagosomes [
53]. Increased p62 expression upon selection for survival in increasing concentrations of doxorubicin (beginning at selection dose 7) would help facilitate this delivery of dysfunctional mitochondria to autophagosomes. While Atg7 and Beclin1 levels remained unchanged, Bcl-2 protein levels varied throughout selection for doxorubicin resistance (Fig.
5). For example, relative to co-cultured MCF-7
CC cells, MCF-7
DOX2 cells selected to dose level 7 (6.5 nM doxorubicin) showed considerably higher expression of Bcl-2. This increase in cellular Bcl-2 levels likely enabled MCF-7
DOX2 cells to survive doxorubicin concentrations up to dose level 7, due to the ability of Bcl-2 to inhibit doxorubicin-induced apoptosis [
63,
71,
72]. However, at selection doses above 6.5 nM doxorubicin, Bcl-2 expression began to decline in a dose-dependent manner (Fig.
5). Since, Bcl-2 can negatively regulate autophagy by forming complexes with Beclin 1 [
55,
73], the loss of Bcl-2 might help promote autophagy at higher selection doses by promoting Beclin 1-dependent autophagy. There was, however, no change in the expression of Beclin 1 and Atg7 throughout selection for doxorubicin resistance, which is often seen in canonical autophagy. This suggests the activation of non-canonical or selective autophagy. There is some recent evidence that, in addition to canonical autophagy, Bcl-2 can regulate non-canonical autophagy, since knockdown of Bcl-2 activity by the Bcl-2 inhibitor Z18 induces autophagy that is unaffected by Beclin 1 and phosphatidyl inositol 3-kinase inhibition [
74]. However, overexpression of Bcl-2 in MCF-7
DOX2–10 cells did not result in autophagy inhibition (as determined by LC3-II expression levels), nor did it increase cellular sensitivity to doxorubicin (data not shown).
Our data clearly illustrates that MCF-7
DOX2–10 cells demonstrated a higher level of autophagy (as measured by LC3-II expression and electron microscopy) than equivalent co-cultured control cells. However, the rate of long lived protein hydrolysis as measured by the flux assay (a functional indicator of autophagy) was only marginally higher in MCF-7
DOX2–10 cells than in MCF-7
CC10 cells (Fig.
7). This may be because the high level of protein hydrolysis seen in canonical autophagy is used to either degrade long lived proteins for housekeeping purposes or energy production under starvation conditions. However, when cells undergo treatment with chemotherapy drugs, there is no shortage of nutrients and growth factors. Thus, organelle damage might be the main effect of drug treatment, and it may be preferable for cells in such instances to activate selective autophagy to eliminate damaged organelles rather than activation of canonical autophagy and protein hydrolysis to support cellular metabolism. After drug entry into tumor cells, mitochondria may be the first target to be affected by doxorubicin prior to its binding to nuclear DNA. Therefore, doxorubicin resistance could be partially attributed to enhanced clearance of the damaged mitochondria caused by doxorubicin via mitophagy.
Autophagy is a process that receives inputs from multiple pathways. The well documented canonical pathways regulating starvation-induced autophagy [
75‐
77] may or may not be applicable to autophagy induced by other stress inducers, such as chemotherapy agents. For example, the neurotoxin MMP+ induces autophagy in SHSY5Y human neuroblastoma cells through a pathway distinct from starvation-induced autophagy. Classic inhibitors of amino acid deprivation-associated autophagy do not inhibit the autophagic response elicited by MMP+ treatment, despite confirmation that the pathway is operative in SHSY5Y cells [
10]. Similarly, MCF-7 cells show Beclin 1-hVps34-independent autophagy or non-canonical autophagy in response to resveratrol treatment [
9,
78].
In a recent study, Sun et al. provided evidence of increased autophagy upon exposure of MCF-7 cells to epirubicin and that autophagy facilitates resistance to epirubicin [
59]. Our manuscript supports the general themes of the prior study, but differs from it in several respects. Firstly, our study demonstrates a clear dose-dependent and temporal relationship between doxorubicin selection dose and both the acquisition of doxorubicin resistance and increased autophagy, in particular at selection doses at or above 44 nM doxorubicin. We show much greater LC3-II production (autophagy) than that observed by Sun et al. when the selection dose reaches 44 nM or greater. Our study also provides evidence that autophagy induction upon selection for doxorubicin resistance appears unrelated to starvation-induced (canonical) autophagy, as siRNA-mediated downregulation of Atg7 had no effect on the sensitivity of MCF-7
DOX2–10 cells to doxorubicin and induction of the cargo protein p62 is typically associated with non-canonical or selective autophagy.
There is emerging evidence that autophagy may be highly relevant to chemotherapy drug resistance and improving the efficacy of chemotherapy treatment in cancer patients. For example, the combined inhibition of autophagy by the mTOR inhibitor temsirolimus and by the lysosomotropic agent chloroquine in a phase I study, showed the combination to be safe, with clear evidence of autophagy inhibition. 67 % of patients achieved stable disease at the maximally tolerated dose (MTD) of this regimen in patients with solid tumours. Moreover, 74 % of melanoma patients achieved stable disease at the MTD of this regimen [
79]. The combination of an mTOR and autophagy inhibitor may be important for clinical efficacy, as a study in prostate tumour xenograft models found that the combination of the mTOR inhibitor AZD5363 and chloroquine significantly reduced tumour volume, while either drug alone did not [
80]. Further evidence of the potential link between autophagy and response to chemotherapy stems from a phase II study on the efficacy of sorafenib in patients with refractory lymphoma. Patients clinically responsive to sorafenib had higher baseline levels of an autophagic biomarker and experienced a significant reduction in this biomarker during treatment [
81]. These previous investigations and our current study in drug-resistant breast tumour cells provide a compelling rationale for investigating the potential of autophagy inhibitors (possibly in combination with mTOR inhibitors) to improve clinical response to chemotherapy. This is particularly important for invasive breast cancer (not including ductal carcinoma in situ), which affects approximately 1 in 8 women in the U.S. (
http://www.breastcancer.org/symptoms/understand_bc/statistics ). According to the ClinicalTrials.gov website, two phase II clinical trials are currently recruiting patients to assess the effect of the lysosomotropic autophagy inhibitor chloroquine (alone) in patients with breast cancer or ductal carcinoma in situ prior to surgery.
Acknowledgements
This work was supported by core support to A.M.P from the Northern Cancer Foundation, Sudbury, Ontario, Canada.