Increasing intratumor C/EBP-β LIP and nitric oxide levels overcome resistance to doxorubicin in triple negative breast cancer
Triple negative breast cancer (TNBC) easily develops resistance to the first-line drug doxorubicin, because of the high levels of the drug efflux transporter P-glycoprotein (Pgp) and the activation of pro-survival pathways dependent on endoplasmic reticulum (ER). Interfering with these mechanisms may overcome the resistance to doxorubicin, a still unmet need in TNBC.
We analyzed a panel of human and murine breast cancer cells for their resistance to doxorubicin, Pgp expression, lysosome and proteasome activity, nitrite production, ER-dependent cell death and immunogenic cell death parameters. We evaluated the efficacy of genetic (C/EBP-β LIP induction) and pharmacological strategies (lysosome and proteasome inhibitors), in restoring the ER-dependent and immunogenic-dependent cell death induced by doxorubicin, in vitro and in syngeneic mice bearing chemoresistant TNBC. The results were analyzed by one-way analysis of variance test.
We found that TNBC cells characterized by high levels of Pgp and resistance to doxorubicin, had low induction of the ER-dependent pro-apoptotic factor C/EBP-β LIP upon doxorubicin treatment and high activities of lysosome and proteasome that constitutively destroyed LIP. The combination of chloroquine and bortezomib restored doxorubicin sensitivity by activating multiple and interconnected mechanisms. First, chloroquine and bortezomib prevented C/EBP-β LIP degradation and activated LIP-dependent CHOP/TRB3/caspase 3 axis in response to doxorubicin. Second, C/EBP-β LIP down-regulated Pgp and up-regulated calreticulin that triggered the dendritic cell (DC)-mediated phagocytosis of tumor cell, followed by the activation of anti-tumor CD8+T-lymphocytes upon doxorubicin treatment. Third, chloroquine and bortezomib increased the endogenous production of nitric oxide that further induced C/EBP-β LIP and inhibited Pgp activity, enhancing doxorubicin’s cytotoxicity. In orthotopic models of resistant TNBC, intratumor C/EBP-β LIP induction - achieved by a specific expression vector or by chloroquine and bortezomib - effectively reduced tumor growth and Pgp expression, increased intra-tumor apoptosis and anti-tumor immune-infiltrate, rescuing the efficacy of doxorubicin.
We suggest that preventing C/EBP-β LIP degradation by lysosome and proteasome inhibitors triggers multiple virtuous circuitries that restore ER-dependent apoptosis, down-regulate Pgp and re-activate the DC/CD8+T-lymphocytes response against TNBC. Lysosome and proteasome inhibitors associated with doxorubicin may overcome the resistance to the drug in TNBC.
KeywordsTriple negative breast cancer Doxorubicin P-glycoprotein Endoplasmic reticulum stress CAAT/enhancer binding protein (C/EBP)-β Calreticulin
One-way analysis of variance
Breast cancer resistance protein
Bovine serum albumin
CAAT/enhancer binding protein
C/EBP homologous protein
Fetal bovine serum
Green fluorescence protein
Immunogenic cell death
Multidrug resistant protein 1
Nitric oxide synthase
Quantitative Real Time-PCR
Relative luminescence units
Triple negative breast cancer
Triple negative breast cancer (TNBC) is often treated with anthracycline (e.g. doxorubicin or daunorubicin)- or taxane-based monotherapy , but the success is lower than in other breast cancer types .
Doxorubicin kills tumor cells by inducing DNA damage, increasing reactive oxygen and nitrogen species such as nitric oxide (NO), impairing mitochondrial metabolism, inducing endoplasmic reticulum (ER) stress and immunogenic cell death (ICD) [3, 4, 5]. The main mechanism of doxorubicin-induced ICD is the induction of ER stress, that triggers the translocation of calreticulin (CRT) from the ER, where it works as calcium sensor and chaperon, to the plasma-membrane. Here, CRT promotes the phagocytosis of tumor cells by dendritic cells (DC) and the activation of a durable anti-tumor response by CD8+T-lymphocytes .
Doxorubicin’s efficacy is limited by the presence of drug efflux transporters such as P-glycoprotein (Pgp) . Pgp limits the doxorubicin intracellular accumulation and the drug’s ability to elicit pleiotropic cytotoxic effects.
Pgp expression is regulated by multiple transcription factors. CAAT/enhancer binding protein (C/EBP)-β, a transcription factor with two isoforms - C/EBP-β LAP and LIP - that work as antagonists, is one of the main controller of Pgp expression in solid tumors . LAP is activated during early ER stress, induces pro-survival pathways and up-regulates Pgp; LIP is induced after prolonged ER stress, stimulates C/EBP homologous protein (CHOP)/Tribbles 3(TRB3)/caspase 3-mediated apoptosis  and down-regulates Pgp .
Besides a high expression, also a high activity of Pgp determines doxorubicin resistance. Natural and synthetic inhibitors of Pgp [10, 11], liposomal formulations , co-delivery of Pgp inhibitors plus doxorubicin , have been tested to reduce Pgp activity in vitro and in preclinical models, but until now none of these approaches were effective in patients. NO is a potent inhibitor of Pgp activity: this molecule, released by synthetic NO donors or produced by the endogenous NO synthase (NOS) enzymes, nitrates specific tyrosines that are critical for Pgp activity. Such covalent modification reduces the doxorubicin efflux through Pgp [14, 15, 16]. Curiously, doxorubicin increases the endogenous production of NO, that mediates part of the cytotoxic effects of the drug , stimulates the translocation of CRT and the ICD of tumor cell , induces ER stress [18, 19]. These events, however, occur only in doxorubicin-sensitive/Pgp-negative cells, not in doxorubicin-resistant/Pgp-positive ones [16, 17], leading to hypothesize that multiple cross-talks determine a chemo-immune-resistant phenotype. Indeed, Pgp-positive cancer cells: i) do not accumulate the intracellular amount of doxorubicin sufficient to increase NO production  and induce ICD ; ii) do not induce C/EBP-β LIP and ER stress-dependent cell death , a condition necessary for the translocation of CRT on cell surface and the subsequent ICD ; iii) are not phagocytized by DC since Pgp hampers the immune-activating functions of CRT in plasma-membrane .
Disrupting these vicious circles by decreasing Pgp expression and activity is the only way to restore the multiple cytotoxic mechanisms of doxorubicin. In this work we demonstrated that preventing C/EBP-β LIP degradation and increasing NO levels reduce at the same time expression and activity of Pgp, restore the ER stress-dependent apoptosis and the ICD induced by doxorubicin, rescuing the anthracycline’s therapeutic efficacy in Pgp-positive TNBC.
Materials and methods
Chemicals and supplies
Plastic ware were obtained from Falcon (Becton Dickinson, Franklin Lakes, NJ). Electrophoresis reagents were from Bio-Rad Laboratories (Hercules, CA). The protein content of cell lysates was assessed using the BCA kit from Sigma Chemicals Co. (St. Louis, MO). Unless specified otherwise, all reagents were purchased from Sigma Chemicals Co.
Human non-transformed breast epithelial MCF10A cells, human breast cancer MCF7, SKBR3, T47D, MDA-MB-231 cells, murine mammary cancer JC cells were purchased from ATCC (Manassas, VA). Murine mammary cancer TUBO cells were a kind gift of Prof. Federica Cavallo, Department of Molecular Biotechnology and Health Sciences, University of Torino, Italy. All human cells were authenticated by microsatellite analysis using the PowerPlex kit (Promega Corporation, Madison, WI; last authentication: January 2018). For 3D-cultures, 1 × 105 cells were seeded in 96-well plate coated with Biomimesys™ matrix (Celenys, Rouen, France). Cells were grown in DMEM/HAM F12 nutrient mixture medium (MCF710A, MCF7, SKBR3, T47D), RPMI-1640 medium (MDA-MB-231, JC), DMEM medium (TUBO) supplemented with 10% v/v fetal bovine serum (FBS) and 1% v/v penicillin-streptomycin. Cells were checked for Mycoplasma spp. contamination by PCR every three weeks; contaminated cells were discharged.
Plasma-membrane proteins were isolated using the Cell Surface Protein Isolation kit (ThermoFisher Scientific Inc., Waltham, MA) according to the manufacturer’s protocol. For whole cell lysates, cells were rinsed with lysis buffer (50 mM Tris-HCl, 1 mM EDTA, 1 mM EGTA, 150 mM NaCl, 1% v/v Triton-X100; pH 7.4), supplemented with the protease inhibitor cocktail III (Cabiochem, La Jolla, CA), sonicated and clarified at 13000×g, for 10 min at 4 °C. Protein extracts (20 μg) were subjected to SDS-PAGE and probed with the following antibodies: anti-Pgp (1:250, rabbit polyclonal, #sc-8313, Santa Cruz Biotechnology Inc., Santa Cruz, CA), anti-multidrug resistant protein 1 (MRP1; 1:500, mouse clone MRPm5, Abcam, Cambridge, UK), anti-breast cancer resistance protein (1:500, mouse clone BXP-21, Santa Cruz Biotechnology Inc.), anti-C/EBP-β (1:500, rabbit polyclonal, # sc150, Santa Cruz Biotechnology Inc.), anti-CHOP (1:500, mouse monoclonal, #ab11419, Abcam), anti-TRB3 (1:500, rabbit polyclonal, #13300–1-AP, Proteintech, Chicago, IL), anti-caspase-3 (1:1000, mouse clone C33, GeneTex, Hsinhu City, Taiwan), anti-CRT (rabbit polyclonal #PA3–900, Affinity Bioreagents, Rockford, IL), anti-NOS I (1:500, mouse clone 16, BD Biosciences, Franklin Lakes, NJ), anti-NOS II (1:1000, mouse clone 4E5, ThermoFisher Scientific Inc.), anti-NOS III (1:500, mouse clone 3, BD Biosciences), anti-pancadherin (1:500, goat clone C-19, Santa Cruz Biotechnology Inc.), anti-β-tubulin (1:1000, mouse clone D10, Santa Cruz Biotechnology Inc.), followed by the horseradish peroxidase-conjugated secondary antibodies (Bio-Rad). The membranes were washed with Tris-buffered saline (TBS)/Tween 0.01% v/v. To detect ubiquitinated C/EBP-β, 100 μg of proteins from whole cell lysates were immuno-precipitated overnight with the anti-C/EBP-β antibody, then probed with an anti-mono/poly-ubiquitin antibody (1:1000, mouse clone FK2, Axxora, Lausanne, Switzerland), using 50 μl of PureProteome Magnetic Beads (Millipore, Bedford, MA). To detect nitrated Pgp, 100 μg of proteins from plasma-membrane were immuno-precipitated overnight with anti-nitrotyrosine antibody (1:50, rabbit polyclonal, #06–284, Millipore), then probed with the anti-Pgp antibody. Proteins were detected by enhanced chemiluminescence (Bio-Rad Laboratories). Blot images were acquired with a ChemiDocTM Touch Imaging System device (Bio-Rad Laboratories). The densitometric analysis was performed with the ImageJ software (https://imagej.nih.gov/ij).
Lysosome and proteasome activities
The activity of cathepsin L, an index of lysosome activity, was measured according to . The results were expressed as nmoles/mg cellular proteins. Proteasome activity was measured with the Proteasome-Glo™ Cell-Based Assays (Promega Corporation). The results were expressed as relative luminescence units (RLU)/mg cellular proteins.
1 × 104 cells were seeded in 96-well plate and incubated as described in the experimental section for 72 h. To calculate the IC50, cells were treated with doxorubicin at scalar concentrations (from 10− 10 to 10− 3 M). Viability was measured with the ATPLite Luminescence Assay kit (PerkinElmer, Waltham, MA) as per manufacturer’s instructions. The viability in untreated cells was considered as 100%. The results were expressed as percentage of viable cells towards the untreated cells. The IC50 was calculated with the CompuSyn software (http://www.combosyn.com).
Doxorubicin accumulation and efflux
The intracellular doxorubicin content and the drug efflux were measured as detailed in . The intracellular doxorubicin concentration was expressed as nanomoles of doxorubicin/mg cellular proteins. The efflux of doxorubicin was expressed as the change in the intracellular concentration of the drug/minute (dc/dt).
Pgp ATPase activity
The Pgp ATPase activity was measured in Pgp-rich membrane vesicles as described in . The results were expressed as μmol hydrolyzed phosphate/min/mg membrane proteins.
Nitrite production and NOS activity
The production of nitrite, the stable derivative of NO, was measured spectrophotometrically by the Griess methods, as described in . Nitrite concentration was expressed as nanomoles/min/mg cellular proteins. The activity of NOS in cell lysates was measured using the Ultrasensitive Colorimetric Assay for Nitric Oxide Synthase kit (Oxford Biomedical Research, Oxford, MI), as per manufacturer’s instructions. The enzyme activity was expressed as nanomoles of nitrites/min/mg cellular proteins.
5 × 105 2D-cells were grown onto glass coverslips in 24-well plates overnight; the same number of cells was seed to produce 3D-cultures, analyzed after 1 week. Sample were fixed using 4% w/v paraformaldehyde (PFA) for 15 min at room temperature, washed with PBS, incubated for 1 h at 4 °C with an anti-Pgp antibody (1:50, mouse clone JSB-1; Abcam, diluted in 1% FBS/PBS), washed five times with PBS and incubated for 1 h at room temperature with an AlexaFluor488-conjugated secondary antibody (Abcam, diluted 1:50 in 1% FBS/PBS). Cells were incubated with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI), diluted 1:10000 in PBS for 5 min, washed four times with PBS and once with deionized water. The cover slips were mounted with the Gel Mount Aqueous Mounting and examined with a Leica DC100 fluorescence microscope (Leica Microsystems GmbH, Wetzlar, Germany). For each experimental point, a minimum of five microscopic fields were examined.
Over-expression of C/EBP-β LAP and LIP
The pcDNA4/TO expression vectors (Invitrogen Life Technologies, Milan, Italy) for LAP and LIP, produced as reported previously , were co-transduced with pcDNA6/TR vector (Invitrogen Life Technologies) in parental cells. Doxycycline-inducible (TetON) stable clones were generated by selecting cells with 2 μg/ml blasticidin S (Invitrogen Life Technologies) and 100 μg/ml zeocin (InvivoGen, San Diego, CA). LIP induction was activated by adding 1 μg/ml doxycycline in the culture medium.
Quantitative real time-PCR (qRT-PCR)
Total RNA was extracted and reverse-transcribed using the iScriptTM cDNA Synthesis Kit (Bio-Rad Laboratories). qRT-PCR was performed using the IQ™ SYBR Green Supermix (Bio-Rad Laboratories). The following PCR primer sequences were designed using the qPrimerDepot software (http://primerdepot.nci.nih.gov/): Pgp (human): 5’-TGCTGGAGCGGTTCTACG-3′, 5’-ATAGGCAATGTTCTCAGCAATG-3′; Pgp (mouse): 5’-TGCTTATGGATCCCAGAGTGAC-3′, 5’-TTGGTGAGGATCTCTCCGGCT-3’;CRT (human): 5’-TGTCAAAGATGGTGCCAGAC-3′, 5’-ACAACCCCGAGTATTCTCCC-3′; CRT (mouse): 5’-TACAAGGGCGAGTGGAAACC-3′, 5’-GCATCGGGGGAGTATTCAGG-3′; S14 (human): 5’-CGAGGCTGATGACCTGTTCT-3′, 5’-GCCCTCTCCCACTCTCTCTT-3′; S14 (mouse): 5’-TCTGGGATGAAGATTGGGCG-3′, 5’-ACCCCCTTTTCTTCGAGTGC-3′. The relative gene expression levels were calculated using the Gene Expression Quantitation software (Bio-Rad Laboratories).
Chromatin immunoprecipitation (ChIP)
The putative binding sites of C/EBP-β, containing CAAT box motif, on human and murine CRT promoter were identified using the Gene Promoter Miner software (http://gpminer.mbc.nctu.edu.tw/). The following primers were designed with the Primer3 software (http://primer3.ut.ee/): 5′-TGGGGAGGTGGAGTAGAGTG-3′; 5’-CAGGAACTGCAGGGACTGAG-3′ (site 831–843, human CRT promoter); 5’-CTCACAGGTCTCGCCTTGTC-3′; 5’-ATGCACTGTTCCGACGTTC-3′ (site 1302–1313, human CRT promoter); 5’-CCTAGCGAGCCAGAGACTC-3′; 5’-CTATTGGTCGCACTATGGGC-3′ (site 798–811, mouse CRT promoter); 5’-GCCTAACTTGCTGAGCCAAC-3′; 5’-CTACCTCTCACCCGAACCTG-3′ (site 872–883, mouse CRT promoter). To determine the binding of LAP and LIP to CRT promoter, ChIP was performed as described in .
Flow cytometry analysis
1 × 105 cells were washed with PBS, detached with Cell Dissociation Solution, wash twice with PBS, incubated for 45 min at 4 °C with the anti-CRT antibody, diluted 1:100 in 0.25% v/v bovine serum albumin (BSA)-PBS, followed by the AlexaFluor488-conjugated secondary antibody (1:50) for 30 min at 4 °C. After the fixation step in 2.5% v/v PFA for 5 min at room temperature, samples were analyzed with a Guava® EasyCyte flow cytometer (Millipore) equipped with the InCyte software (Millipore). Cells incubated with not-immune isotype antibody, followed by secondary antibody, were include as control of specificity.
Tumor cells phagocytosis and T-lymphocytes activation
DC were generated from monocytes immuno-magnetically isolated from peripheral blood of healthy donors, provided by Blood Bank of AOU Città della Salute e della Scienza, Torino, Italy as previously reported  or from the bone marrow of 6-week old female balb/C mice . The phagocytosis assay was performed as detailed in , by co-incubating DC and tumor cells at 37 °C and 4 °C for 24 h. The percentage of phagocytized cells obtained after the incubation at 4 °C was subtracted from the percentage obtained at 37 °C, and was always less than 5% (not shown). The phagocytosis rate was expressed as phagocytic index . After cell phagocytosis, DC were washed and co-cultured for 10 days with autologous T-cells, isolated by immuno-magnetic sorting with the Pan T Cell Isolation Kit (Miltenyi Biotec., Tetrow, Germany). The expression of CD107, a degranulation marker and an index of active cytotoxic CD8+T-lymphocytes, was determined by flow cytometry as previously reported , using anti-human or mouse fluorescein isothyocyanate (FITC)-conjugated-CD8 (1:10, clones BW135/80 and 53–6.7) and phycoerythrin (PE)-conjugated-CD107 (1:10, clones H4A3 and 1D4B) antibodies (Miltenyi Biotec).
1 × 105 cells were treated with 10 nM of 3 unique 27mer siRNA duplexes, targeting DDIT3/CHOP (#SR319903; Origene, Rockville, MD) or with a Trilencer-27 Universal scrambled negative control siRNA duplex (#SR30004; Origene), as per manufacturer’s instructions. The efficiency of silencing was verified by immunoblotting.
Calreticulin knock-out (KO)
JC cells were knocked-out for calreticulin using a pool of two calreticulin-targeting CRISPR/Cas9 KO-green fluorescence protein (GFP) vectors (#KN302469, Origene). Non-targeting (scrambled) CRISPR/Cas9 vector (Origene) was used as control of specificity. 1 × 105 cells were seeded in antibiotic-free medium. 1 μg of CRISPR/Cas9 plasmid was used, as per manufacturer’s instructions. Transfected cells were sorted by isolating GFP-positive cells. KO efficacy was verified by immunoblotting. Stable KO-clones were generated by culturing cells for 6 weeks in medium containing 1 μg/ml puromycin.
In vivo tumor growth
1 × 107 JC TetON LIP cells, wild-type, stably transfected with a KO-CRT vector or with a scrambled vector, were mixed with 100 μl Matrigel and orthotopically implanted in 6 week-old female immunocompetent balb/C mice (Charles River Laboratories Italia, Calco), housed (5 per cage) under 12 h light/dark cycle, with food and drinking provided ad libitum. Tumor growth was measured daily by caliper, according to the equation (LxW2)/2, where L = tumor length and W = tumor width. When tumor reached the volume of 50 mm3, mice were randomized and treated as reported in the experimental section. Tumor volumes were monitored daily. Animals were euthanized at day 21 after randomization with zolazepam (0.2 ml/kg) and xylazine (16 mg/kg). Lactate dehydrogenase, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, creatinine, creatine phosphokinase and troponin were measured on blood samples collected immediately after euthanasia, using commercially available kits from Beckman Coulter Inc. (Beckman Coulter, Miami, FL). In all studies, researchers analyzing the results were unaware of the treatments received by animals.
Tumors were resected and fixed in 4% v/v PFA, photographed and sectioned, then stained with hematoxylin/eosin or immuno-stained for Ki67 (1:50, rabbit polyclonal #AB9260, Millipore), Pgp (1:50), CHOP (1:50), cleaved(Asp175)-caspase 3 (1:200, rabbit polyclonal #9661, Cell Signaling Technology Inc., Danvers, MA), CRT (1:100), CD11c (1:50, hamster, clone HL3, BD Biosciences) to label intra-tumor DC, CD8 (1:100, rat clone YTS169.4, Abcam) to label intra-tumor cytotoxic T-lymphocytes, followed by a peroxidase-conjugated secondary antibody (1:100, Dako, Glostrup, Denmark). Sections were examined with a Leica DC100 microscope.
Tumor-draining lymph nodes were collected, homogenized for 30 s at 15 Hz using a TissueLyser II device (Qiagen, Hilden, Germany) and centrifuged at 12000×g for 5 min. The supernatant was collected to measure the amount of IFN-γ, using the Mouse IFN-γ DuoSet ELISA Kit (R&D Systems, Minneapolis, MN). The results were expressed as nmol/ml.
All data in the text and figures are provided as means±SD. The results were analysed by a one-way analysis of variance (ANOVA), using the Statistical Package for Social Science (SPSS) software (IBM, Armonk, NY). p < 0.05 was considered significant.
Pgp-positive breast cancer cells do not induce C/EBP-β LIP upon doxorubicin treatment and have high lysosome and proteasome activities
Doxorubicin resistance is associated to the lack of C/EBP LIP-β-dependent apoptosis in breast cancer cells
We thus examined if the FDA-approved lysosome inhibitor chloroquine and proteasome inhibitor bortezomib could prevent C/EBP-β LIP degradation. In preliminary dose-dependence experiments, we determined that at 1 μM chloroquine and bortezomib - used as single agents - did not reduce significantly cell viability (Additional file 2), but they decreased lysosome (Additional file 3a) and proteasome (Additional file 3b) activities, respectively, in Pgp-positive MDA-MB-231 and in JC cells. When used in combination, chloroquine and bortezomib significantly decreased the viability of these cell lines (Additional file 2).
Chloroquine and bortezomib down-regulate Pgp expression and activity by increasing C/EBP-β LIP and NO
It has been previously reported that NOS I expression is increased by the inhibition of proteasome , while NOS III activity is increased by chloroquine : indeed, the lysosome inhibitor lowers the availability of intracellular free iron ; this condition increases NOS III activity without changing its expression . In line with these findings, bortezomib, alone or in combination with chloroquine, increased NOS I expression, while chloroquine did not alter the expression of any NOS isoform (Fig. 3a). The use of the proteasome activator betulinic acid suggested that the up-regulation in NOS I induced by bortezomib was mediated by the inhibition of proteasome activity (Fig. 3a). Both chloroquine and bortezomib, alone and in particular in combination, increased the activity of NOS enzymes and the production of nitrite, the stable derivative of NO (Fig. 3b; Additional file 4). This trend can be due either to the increased expression of NOS I induced by bortezomib (Fig. 3a) or to the increased activity of NOS III induced by chloroquine .
Since NO induces ER stress , we performed the following add-back experiments in order to investigate whether chloroquine and bortezomib up-regulated C/EBP-β LIP by inhibiting lysosomal and proteasomal activity, by increasing NO production or by both mechanisms.
In a first experimental set, we co-incubated chloroquine and bortezomib with the lysosome activator torin-1 or the proteasome activator betulinic acid. As shown in Fig. 3c and Additional file 5, in both MDA-MB-231 and JC cells, torin-1 reduced the induction of C/EBP-β LIP/CHOP/TRB3/caspase 3 axis elicited by chloroquine, alone or combined with bortezomib. Since torin-1 did not affect the ubiquitination of C/EBP-β LIP, its effect was likely due to the activation of the LIP degradation via lysosome. Betulinic acid reduced the up-regulation of C/EBP-β LIP and downstream effectors induced by bortezomib, alone or associated with chloroquine. Of note, betulinic acid also reduced the levels of poly-ubiquitinated LIP, indicating that if favored the removal of ubiquitinated LIP via proteasome.
In a second experimental set, we used the NO donor sodium nitroprusside (SNP), which increased nitrite, and the NO scavenger carboxy-PTIO, which reduced the amount of nitrite in untreated, chloroquine- and bortezomib-treated cells (Additional file 4). SNP increased C/EBP-β LIP levels and CHOP/TRB3/caspase 3 axis activation, while the co-incubation with carboxy-PTIO abrogated these events (Fig. 3c; Additional file 5), suggesting the increased levels of NO may trigger the induction of LIP and ER-dependent apoptotic cascade. Of note, PTIO also reduced the increase in nitrite (Additional file 4) and in C/EBP-β LIP/CHOP/TRB3/caspase 3 axis in cells co-incubated with chloroquine and bortezomib (Fig. 3c; Additional file 5). We did not find a further enhancement of nitrite or C/EBP-β LIP/CHOP/TRB3/caspase 3 up-regulation in cells treated with SNP, chloroquine and/or bortezomib compared to cells treated with SNP alone (Fig. 3c; Additional files 4, 5), suggesting that in these experimental conditions the level of NO released by SNP was likely saturating and sufficient to reach the maximal C/EBP-β LIP induction.
Overall these results suggest that either the inhibition of lysosome and proteasome activity or the increase in endogenous NO mediate the induction of C/EBP-β LIP exerted by chloroquine and bortezomib.
To verify if the same chemosensitizing effects were maintained in 3D-cultures, a model closer to the in vivo tumor biology and characterized by higher Pgp expression and doxorubicin resistance compared to 2D-cultures , we produced 3D-cultures of T47D cells, which were Pgp-negative (Fig. 1a) and doxorubicin-sensitive (Fig. 2b) cells when grown bi-dimensionally. T47D-3D cells increased the expression of Pgp (Additional file 6a), reduced the intracellular doxorubicin retention (Additional file 6b), increased the cell viability in the presence of doxorubicin (Additional file 6c) and lost the ability to induce C/EBP-β LIP in response to the drug (Additional file 6d), behaving like the doxorubicin-resistant/Pgp-positive MDA-MB-231 cells. 3D-cultures had higher lysosome (Additional file 6e) and proteasome (Additional file 6f) activities than 2D-cultures, but they retained the sensitivity to chloroquine and bortezomib. Indeed these two agents reproduced the same effects observed in doxorubicin-resistant/Pgp-positive MDA-MB-231 cells: they increased C/EBP-β LIP/CHOP/TRB3/caspase 3 pathway (Additional file 6 g), decreased Pgp mRNA and protein (Additional file 6 g-h), increased Pgp nitration (Additional file 6 g) in consequence of the increased production of NO (Additional file 6i), restored the intracellular accumulation (Additional file 6j) and cytotoxicity of doxorubicin (Additional file 6 k) to the same levels of doxorubicin-sensitive T47D-2D cells.
C/EBP-β LIP restores doxorubicin-induced immunogenic cell death in resistant breast cancer cells
When doxycycline was added to the culture medium to induce C/EBP-β LIP (Additional file 8a-b), LIP transcriptional activity on CRT promoter (Fig. 6a-c), CRT translocation (Fig. 6d) and CRT-mediated ICD (Fig. 6e-f) were higher compared to un-induced cells. The maximal efficiency in increasing C/EBP-β LIP (Additional file 8a-b) and CRT-dependent ICD (Fig. 6a-f) was achieved in cells treated with doxycycline (that induces C/EBP-β LIP), chloroquine and bortezomib (that prevent C/EBP-β LIP degradation), and doxorubicin (that elicits ER stress up-regulating endogenous C/EBP-β LIP). These results suggest that maintaining a high level of C/EBP-β LIP, which down-regulate Pgp and up-regulate CRT, fully restores doxorubicin-dependent cell death in vitro, by triggering ER stress-mediated apoptosis and ICD.
The C/EBP-β LIP effector CHOP mediates ER stress dependent apoptosis and immunogenic cell death in response to chloroquine and bortezomib
High C/EBP-β LIP levels restore doxorubicin efficacy in drug-resistant breast cancer xenografts
Un-induced tumors treated with chloroquine and bortezomib and C/EBP-β LIP-induced tumors not treated with chloroquine and bortezomib had a comparable decrease in tumor cell proliferation, and a comparable increase in ER stress and apoptosis, as suggested by the staining for Ki67, CHOP and active caspase 3. In parallel, tumors showed decreased Pgp, increased CRT positivity, increased intratumor infiltration of DC and cytotoxic T-lymphocytes (Fig. 8c; Additional file 10a). Also the production of IFN-γ from draining lymph nodes, a marker of local immune system activation, was increased (Additional file 10b). Tumors with induced C/EBP-β LIP, treated with chloroquine and bortezomib followed by doxorubicin, displayed a further reduction in proliferation and Pgp expression, as well as a further increase in ER stress, apoptosis, CRT positivity, intratumor DC, cytotoxic T-lymphocytes (Fig. 8c, Additional file 10a), IFN-γ production (Additional file 10b).
Of note, the combined treatments did not induce signs of systemic toxicity, nor worsen the cardiac damage induced by doxorubicin, according to the animals’ hematochemical parameters (Additional file 11).
C/EBP-β LIP and calreticulin are both necessary to restore immunogenic cell death in doxorubicin-resistant breast cancers
Anthracycline-based chemotherapy is one of the first-line treatment in TNBC, but half of the patients develop resistance . Effective strategies of chemosensitization are still an unmet need.
The resistance to doxorubicin is mostly mediated by Pgp, which limits the intracellular accumulation of the drug and the possibility of exerting its pleiotropic cytotoxic mechanisms, such as increasing NO levels, inducing ER stress and ICD [5, 16, 17, 20]. Since Pgp can be inhibited at transcriptional level by C/EBP-β LIP  and at post-translational level by NO [14, 15], we set up a strategy that increases at the same time C/EBP-β LIP and NO upon treatment with doxorubicin, in order to downregulate expression and activity of Pgp and overcome the resistance to the drug in TNBC.
By screening different breast cancer cell lines, we found that TNBC human and murine doxorubicin-resistant/Pgp-positive cells did not induce C/EBP-β LIP in response to doxorubicin, differently from doxorubicin-sensitive/Pgp-negative cells. The absence of LIP, as documented in other chemoresistant tumors [5, 27], was due to its mono- and poly-ubiquitination, followed by lysosomal and proteasomal degradation. Indeed, doxorubicin-resistant/Pgp-positive TNBC MDA-MB-231 and JC cells displayed the highest activity of lysosome and proteasome, coupled with the lowest induction of C/EBP-β LIP upon doxorubicin treatment. We thus employed lysosome and proteasome inhibitors as pharmacological tools able to prevent C/EBP-β LIP degradation, either in basal conditions or upon doxorubicin treatment.
The elevated expression of the lysosomal enzyme cathepsin D has been already correlated with breast cancer progression , opening the way to the design of cathepsin D inhibitors as potential anti-tumor agents . In the present work, we used chloroquine, a FDA-approved lysosome and autophagosome inhibitor  that exerts anti-tumor effects against TNBC stem cells  and enhances the ER stress-dependent cell death by inhibiting lysosome activity [40, 41].
Also the high activity of proteasome has been correlated with poor prognosis in TNBC patients and poor response to chemotherapy in vitro . Proteasome inhibitors are under evaluation as new therapeutic options in TNBC patients . Our work may provide the rationale for the combined use of lysosome and proteasome inhibitors, in association with first-line therapy doxorubicin, against resistant TNBC as a new combination treatment that activates pleiotropic mechanisms of cell killing.
First, by increasing C/EBP-β LIP/CHOP/TRB3/caspase 3 axis, chloroquine and bortezomib induced a ER stress-dependent apoptosis with at least two mechanism. Indeed, on the one hand chloroquine and bortezomib prevented the lysosomal and proteasomal degradation of LIP, as demonstrated by the abrogation of their effects in cells co-incubated with lysosome or proteasome activators. On the other hand, chloroquine, bortezomib and doxorubicin all increase NO production [17, 29, 30]. NO in turn induces ER stress , and activates CHOP and caspase 3 . Since the NO donor SNP induced LIP as chloroquine and bortezomib did, while the NO scavenger carboxy-PTIO abrogated the effects of the lysosome and proteasome inhibitors, we propose that chloroquine and bortezomib increased C/EBP-β LIP either by preventing its degradation via lysosome and proteasome, or by eliciting a NO-dependent ER stress. The combination treatment with chloroquine, bortezomib and doxorubicin boosted the activation of ER stress-dependent cell death in TNBC cells, restoring one of the cytotoxic mechanisms of doxorubicin.
Second, by increasing C/EBP-β LIP, chloroquine and bortezomib down-regulated Pgp expression and increased CRT. The down-regulation of Pgp allowed doxorubicin to reach an intracellular concentration sufficient to activate the pro-apoptotic C/EBP-β LIP/CHOP/TRB3/caspase 3 pathway and to increase the synthesis of NO. The activation of ER stress-dependent cell death pathway and the increase in NO promote the translocation of CRT from ER to plasma-membrane. The up-regulation of CRT elicited by C/EBP-β LIP at transcriptional level further enhanced this process, contributing to restore ICD and to re-establish a second cytotoxic mechanisms of doxorubicin.
Third, chloroquine and bortezomib increased the synthesis of NO, a non-competitive inhibitor of Pgp [14, 15], that reduced the Vmax of doxorubicin efflux through the transporter. The role of NO in cancer is still controversial, since it can act as a tumorigenic or anti-cancer agent [45, 46], an activator or suppressor of the DC and T-lymphocytes , depending on the concentration and temporal production. The most studied NOS enzyme in TNBC has been NOS II, a negative prognostic factor  and an inducer of resistance to docetaxel . As demonstrated by Dávila-González and coworkers, NOS II inhibition, coupled to a constant expression of NOS III, sensitized TNBC cells to docetaxel and activated the ER stress-dependent pro-apoptotic transducers such as CHOP . This situation closely mimics the scenario of TNBC cells treated with chloroquine and bortezomib, where NOS II was undetectable, NOS III expression was constant, cells were re-sensitized to doxorubicin and activated CHOP-dependent pathways. Of note, docetaxel is a substrate of Pgp as doxorubicin : it is possible that the chemosensitization observed by Dávila-González and coworkers was in part due to the reduced efflux of docetaxel, consequent to the inhibition of Pgp elicited by NO. This hypothesis is supported by our findings in TNBC cells treated with chloroquine and bortezomib, that had increased NO synthesis, reduced Pgp activity, decrease efflux of the Pgp substrate doxorubicin.
Given the pleiotropic effects exerted by the combination of chloroquine and bortezomib, we next investigated if all the observed events were due to the C/EBP-β LIP-mediated induction of ER stress. To this aim, we transiently silenced CHOP, a downstream effector of C/EBP-β LIP  and a promoter of ER stress-dependent cell death [33, 34, 35]. The results obtained in silenced cells indicated that the activation of the caspase 3 elicited by chloroquine and bortezomib, the translocation of CRT and the consequent ICD, were dependent on ER stress. These data were in accord with several evidences demonstrating that ER stress induces apoptosis via CHOP , triggers the surface translocation of CRT and the CRT-mediated phagocytosis [3, 6]. By contrast, ER stress was not responsible for the increased production of NO, the reduction in Pgp expression and activity, the up-regulation of CRT gene. Indeed, the increase in NO was due to the higher expression of NOS I induced by bortezomib and to the higher activity of NOS III induced by chloroquine. The down-regulation of Pgp and the up-regulation of CRT were caused by the activity of C/EBP-β LIP as transcription factor. The reduced activity of Pgp was consequent to the increased production of NO that nitrated tyrosine residues critical for the pump’s catalytic activity.
We can conclude that the effects of chloroquine and bortezomib were either ER stress-dependent or independent, but they all triggered virtuous circuitries that effectively overcome the resistance to doxorubicin in Pgp-positive TNBC cells.
The effects of choloroquine and bortezomib observed in vitro were well reproduced in the preclinical model of doxorubicin-resistant/Pgp-positive JC tumors.
The use of chloroquine and bortezomib in TNBC is not new. It has been reported that chloroquine chemosensitizes TNBC-derived xenografts at 50 mg/kg , i.e. five-fold the concentration used in the present study. Bortezomib partially reduced tumor growth in TNBC patient-derived xenografts at 0.75 mg/kg , i.e. three-fold the dosage used in our in vivo protocol. In chemoresistant TNBC patients bortezomib achieved a partial response, followed by disease recurrence . The novelty of our approach is the use of lower doses of chloroquine and bortezomib in combination with doxorubicin. Such combination therapy did not elicit systemic toxicity, but it reduced tumor growth, increased intratumor ER stress, apoptosis, CRT expression, DC and CD8+T-lymphocyte infiltration. The intratumor activation of C/EBP-β LIP further enhanced the responses elicited by the chloroquine /bortezomib/doxorubicin combination, confirming that C/EBP-β LIP and its downstream effectors (such as CHOP and caspase 3) are critical factors in overcoming doxorubicin resistance.
An anti-tumor abundant immune-infiltrate, indicated by high CD8+T-lymphocytes and low T-regulatory cells, is a strong predictor of good response to chemotherapy in breast cancer patients [52, 53], suggesting that part of the anti-tumor effects of anthracyclines is mediated by the engagement of the host immune system. In agreement with this hypothesis, knocking-out CRT in tumors with induced C/EBP-β LIP completely abrogated the doxorubicin-induced ICD and reduced doxorubicin anti-tumor efficacy, indicating that both C/EBP-β LIP and CRT are necessary for a full rescue of doxorubicin efficacy in resistant tumors.
We are grateful to Mr. Costanzo Costamagna, Department of Oncology, University of Torino, for the technical assistance.
This work was supported by Italian Association for Cancer Research (IG15232 to CR; IG16985 to MM), De Benedetti-Cherasco Foundation (Torino-Weizmann Collaborative Program: Scientific Cooperation and Exchange to MR and CR). ICS is a post-doctoral research fellow supported by the “Fondazione Franco e Marilisa Caligara”, Torino, Italy. BC is a post-doc research fellow supported by American Association for Cancer Research (AACR). JK was a post-doctoral fellow of the “Fondazione Umberto Veronesi”, Milano, Italy.
The funding institutions had no role in the study design, data collection and analysis, or in writing the manuscript.
Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information files.
ICS and EG performed the in vitro and in vivo experiments and analyzed the data; AA, EM, GEFA supported the experiments in vitro; BC performed the immunological assays; MM supervised the immunological assays and revised the manuscript; MD and MR analyzed the data and revised the manuscript; CR and JK conceived and supervised the study, wrote and revised the manuscript.
Ethics approval and consent to participate
The study was approved by the Ethical Committee of the Blood Bank, AOU Città della Salute e della Scienza, Torino, Italy (#DG-767/2015). The Animal care and experimental procedures were approved by the Bio-Ethical Committee of the Italian Ministry of Health (#122/2015-PR).
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 2.Székely B, Silber AL, Pusztai L. New therapeutic strategies for triple-negative breast cancer. Oncology (Williston Park). 2017;31:130–7.Google Scholar
- 5.Riganti C, Kopecka J, Panada E, Barak S, Rubinstein M. The role of C/EBP-β LIP in multidrug resistance. J Natl Cancer Inst 2015;107(5): pii:djv046.Google Scholar
- 32.Doublier S, Belisario DC, Polimeni M, Annaratone L, Riganti C, Allia E, et al. HIF-1 activation induces doxorubicin resistance in MCF7 3-D spheroids via P-glycoprotein expression: a potential model of the chemo-resistance of invasive micropapillary carcinoma of the breast. BMC Cancer. 2012;12:4.CrossRefGoogle Scholar
- 41.Thomas S, Sharma N, Golden EB, Cho H, Agarwal P, Gaffney KJ, et al. Preferential killing of triple-negative breast cancer cells in vitro and in vivo when pharmacological aggravators of endoplasmic reticulum stress are combined with autophagy inhibitors. Cancer Lett. 2012;325:63–71.CrossRefGoogle Scholar
- 43.Meißner T, Mark A, Williams C, Berdel WE, Wiebe S, Kerkhoff A, et al. Metastatic triple-negative breast cancer patient with TP53 tumor mutation experienced 11 months progression-free survival on bortezomib monotherapy without adverse events after ending standard treatments with grade 3 adverse events. Cold Spring Harb Mol Case Stud 2017;3: pii:a001677.CrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.