Cathepsin G-Induced Cell Aggregation of Breast Cancer MCF-7 Decreases Doxorubicin Sensitivity in a Hypoxia-Inducible Factor-Independent Mechanism

Abstract

Solid tumors habitually harbor regions with insufficient oxygen away from vasculature. Hypoxia is an important factor that confers malignant phenotypes like chemoresistance to tumor cells. We have demonstrated that cathepsin G (CG) stimulates cell aggregation in breast cancer MCF-7 cells by activating insulin-like growth factor-1 signaling. We investigated whether cancer cell aggregates induced by CG acquire hypoxia-dependent chemoresistance. Pimonidazole staining and hypoxia-inducible factor (HIF)-1α and -2α expression indicated that the core of the cell aggregates was hypoxic. Electrophoretic mobility shift and reporter assays showed that the CG-induced cell aggregates displayed transcriptional activity through HIF-responsive elements. Moreover, HIF target genes PGK1 and SLC2A1 demonstrated upregulated expression in CG-induced cell aggregates, indicating that the aggregates expressed functional HIF. Doxorubicin (DXR)-induced cytotoxicity was significantly lower in the cell aggregates induced by CG compared with monolayer cells under normoxia. Unexpectedly, the upregulation of P-glycoprotein expression, which is reported to be a HIF-target gene, and decreasing intracellular accumulation of DXR was not detected in the cell aggregates as opposed to in monolayer cells under normoxia. Additionally, reduction of DXR sensitivity in the aggregates was not suppressed by treatment with the HIF inhibitor, YC-1 and HIF-1α small interfering RNA (siRNA). Therefore, we conclude that cell aggregation induced by CG decreases DXR sensitivity via a HIF-independent mechanism.

INTRODUCTION

Oxygen, an essential molecule in cell metabolism, energy synthesis, and signaling, is associated with survival and normal function.¹⁾ In solid tumors, hypoxic regions exhibit continuous proliferation of tumor cells, aberrant tumor vasculature, and specific metabolism, resulting in inadequate oxygen supply.^2,3) Hypoxia induces various biological processes, such as angiogenesis, drug resistance, and glycolysis, to adapt and survive in tumor cells; therefore, hypoxia contributes to an aggressive phenotype.^4,5) Hypoxia-inducible factors (HIFs) are key molecules in the hypoxic stress response that activate the transcription of more than hundreds of hypoxia-related genes. HIFs are heterodimeric transcription factors composed of the HIF-α subunit, HIF-1α or HIF-2α, and the HIF-1β subunit. HIF-α subunit proteins are expressed only under hypoxic conditions because HIF-α subunits are degraded by proteasome-mediated proteolysis under normoxia. Therefore, HIF functions in hypoxic conditions. HIF heterodimers bind to the hypoxia-responsive element (HRE) upstream of target genes and activate gene transcription. To improve hypoxia by building up the vascular systems in tumor tissue, HIFs promote transcription of the vascular endothelial growth factor (VEGF) gene, and then the tumor vasculature provides a route to escape to blood vessels and distant tissues, resulting in the promotion of metastasis. Furthermore, it has been reported that HIFs induce P-glycoprotein (P-gp) gene expression and confer chemoresistance in colorectal cancer cell lines.^6,7) Therefore, hypoxia is generally considered to be closely associated with poor prognosis in tumors.

The tumor microenvironment is composed of tumor cells and various non-transforming cells, such as fibroblasts and immune cells, which attract various chemokines and cytokines secreted by tumor cells. Infiltrated non-transforming cells affect tumor cell proliferation, invasion, and metastasis, thereby facilitating tumor progression.⁸⁾ Neutrophils are the most abundant immune cells that make up approximately half of leukocytes in human blood and engulf and digest by secreted proteases and reactive oxygen species to eliminate exogenous microorganisms.^9,10) Neutrophil infiltration in tumor tissues is also known as tumor-associated neutrophils (TANs), which affect tumor prognosis. TANs can polarize and play dual roles in tumor cells: N1- and N2-type TANs. N1 type TANs exert an antitumor phenotype through the secretion of cytokines and chemokines that activate immune response and inflammation, and reactive oxygen species, which are cytotoxic to tumor cells. On the other hand, N2 type TANs exhibit pro-tumoral function through secretion of matrix metalloproteinase-9 and VEGF to digest the extracellular matrix and facilitate invasion and intravasation of tumor cells. Many reports indicate that TANs predict poor overall survival in many types of cancer, although the role and importance of TANs in patients with tumors remain debatable. The number of intratumoral neutrophils in renal cell carcinoma is associated with poor prognosis.¹¹⁾ The neutrophil-to-lymphocyte ratio of the peripheral blood in patients with colorectal cancer is correlated with poor clinical outcomes.¹²⁾ Neutrophil extracellular traps (NETs), which are web-like structures composed of DNA-histone complexes, are released from activated neutrophils during microbial digestion and inflammation.¹³⁾ Recently, it was reported that NETs enhance tumor metastasis by trapping circulating tumor cells and waking dormant tumor cells, causing tumor relapse and metastasis.^14,15) In addition, neutrophil elastase (NE), a serine protease secreted by neutrophils, is strongly correlated with tumor progression.¹⁶⁾ For example, NE accelerates tumor cell proliferation in mice genetically bearing lung carcinoma and in the lung cancer cell line A549.¹⁷⁾ NE enhances tumor cell intravasation and subsequent dissemination via NE-mediated formation of dilated intratumoral vasculature.¹⁸⁾ Furthermore, NE stimulates the proliferation of human epidermal receptor 2-positive breast cancer cells.¹⁹⁾

Although NE appears to be associated with tumor progression, the role of cathepsin G (CG), another neutrophil protease, remains limited. CG is a serine protease secreted by activated neutrophils and a subset of monocytes.²⁰⁾ CG has chymotrypsin-like and trypsin-like substrate specificity, which is a preference for large hydrophobic and positive P1 residues, synthetic peptides, as substrates.^21,22) In addition to pathogen digestion, CG plays various important roles in maintaining homeostasis, such as regulating inflammation by modifying chemokines and cytokines, controlling blood pressure, and thrombogenesis.²⁰⁾ CG also increases the permeability of endothelial and epithelial cells.²³⁾ Previously, we have revealed that CG stimulates autonomous cell migration and E-cadherin-mediated cell–cell adhesion, followed by the formation of three-dimensional (3D) multicellular aggregates in human breast cancer MCF-7 cells.^24,25) Notably, the aggregation-inducing activity of CG is specific for E-cadherin-positive breast cancer cells, such as MCF-7, T47 and BT474 cells, but not E-cadherin-negative human breast cancer MDA-MB-231 cells.²⁶⁾ 3D-tumor cell aggregates are used as a model of the tumor microenvironment in vitro and mimic the characteristics of tumor tissues, such as cell–cell adhesion, deposition of extracellular matrix, and gradient distribution of oxygen and nutrients.²⁷⁾ The cell migration- and aggregation-inducing activity of CG is essential for its enzymatic activity.^28,29) We also revealed that cell aggregation induced by CG is required for the digestion of insulin-like growth factor (IGF)-binding protein (IGFBP)-2 followed by free-IGF-1 release and stimulation of IGF-1 signaling, but other serine proteases, such as NE, trypsin, and chymotrypsin, did not.^29,30) The morphological and biological characteristics of CG-induced MCF-7 cell aggregates are similar to those observed in lymphatic vessels of patients with inflammatory breast cancer, which is a highly metastatic and malignant breast cancer, resulting in poor prognosis.^31,32) Because aggregates of tumor cells in lymphatic vessels are associated with tumor embolism followed by metastasis to neighboring tissues, aggregation of breast cancer cells potentially confers malignant phenotype. Furthermore, the growth of tumor cell aggregates can become malignant due to hypoxia. Because CG forms large cell aggregates of MCF-7 cells that are tightly attached by E-cadherin-mediated cell–cell adhesion, we hypothesized that the large MCF-7 cell aggregates induced by CG acquire chemoresistance via a HIF-dependent mechanism, for example, the P-gp-mediated drug efflux system. In the present study, we investigated whether CG-induced MCF-7 cells were hypoxic and resistant to doxorubicin (DXR), a P-gp substrate, via a HIF-dependent mechanism.

MATERIALS AND METHODS

Reagents

CG, purified from human neutrophils, was purchased from BioCentrum (Krakow, Poland). Anti-HIF-1α (Cat. No. 610958), anti-HIF-2α (clone D9E3, Cat. No. 7089), and anti-proliferating cell nuclear antigen (PCNA) antibodies (Cat. No. ab29) were obtained from BD Biosciences (Franklin Lakes, NJ, U.S.A.), Cell Signaling Technology (Danvers, MA, U.S.A.), and Abcam (Cambridge, U.K.), respectively. Anti-P-gp (Cat. No. ab170904) and MRP1 (Cat. No. ab260038) antibodies were purchased from Abcam. 4′,6-Diamidino-2-phenylindole (DAPI) was purchased from Thermo Fisher Scientific (Waltham, MA, U.S.A.). DXR hydrochloride was obtained from Sigma-Aldrich (St. Louis, MO, U.S.A.). YC-1, verapamil, and tariquidar were purchased from Merck (Kenilworth, NJ, U.S.A.), FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan), and Selleck Chemicals (Houston, TX, U.S.A.), respectively.

Cells

Human breast cancer MCF-7 cells were a kind gift from Dr. Hiroshi Kosano (Teikyo University). MCF-7 cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS; MP Biomedicals, Solon, OH, U.S.A.) and 80 µg/mL kanamycin (FUJIFILM Wako Pure Chemical Corporation).

Preparation of CG-Induced MCF-7 Cell Aggregates

MCF-7 cell aggregates induced by CG were prepared using a modified method for formation of CG-induced spheroids.^27,29) Briefly, MCF-7 cells were inoculated in RPMI1640 medium supplemented with 5% FBS at a density of 6 × 10⁴ cells/cm². After 24 h, the cells were washed with serum-free RPMI1640 medium, and then treated with CG (40 nM) in RPMI1640 medium containing 1% bovine serum albumin (BSA) for 24 h.

Hypoxia Induction

The culture plates were inoculated with cells placed in a Modular Incubator Chamber (MIC-101, Billups-Rothenberg, Del Mar, CA, U.S.A.) that was infused with mixed-low oxygen gas (3% O₂, 5% CO₂, and 92% N₂) and sealed as instructed by the manufacturer. The sealed chamber was then placed in a CO₂ incubator.

Western Blotting

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequent Western blotting were performed as previously described.³⁰⁾ Nuclear extracts were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific). Nuclear extract or whole-cell lysates were prepared by solubilizing cells in radio-immunoprecipitation assay buffer (50 mM Tris–HCl (pH 7.5) containing 150 mM NaCl, 1% NP-40, and 0.1% sodium deoxycholate) supplemented with protease-inhibitor cocktails cOmplete (Roche Diagnostics, Mannheim, Germany). Samples were then mixed at a 1 : 4 ratio with 5× sample buffer (300 mM Tris–HCl (pH 6.8) containing 10% 2-mercaptoethanol, 10% SDS, 50% glycerol, and 0.05% Coomassie Brilliant Blue) and boiled for 5 min. The resulting samples (20 µg protein/lane) were subsequently separated by SDS-PAGE using pre-cast 5–20% Tris-glycine gradient gels (ATTO Corporation, Tokyo, Japan) and transferred to polyvinylidene difluoride membranes (FUJIFILM Wako Pure Chemical Corporation). Membranes were blocked by incubation with Tris-buffered saline containing 0.1% Tween 20 (TBS-T) and 5% skim milk (Megmilk Snow Bland, Tokyo, Japan) for 1 h, probed with the appropriate primary antibodies, diluted with Can Get Signal Immunoreaction Enhancer Solution (TOYOBO, Osaka, Japan) at 4 °C, washed extensively with TBS-T, and incubated with horseradish peroxidase-conjugated secondary antibodies (Cytiva, Tokyo, Japan; 1 : 5000 dilution). Finally, the membranes were washed, developed by incubation with enhanced chemiluminescence (ECL) detection reagents EzWestLumi plus (ATTO), and exposed to hyperfilm ECL (Cytiva).

Protein Assay

Protein concentrations were determined using a Pierce bicinchoninic acid protein assay kit (Thermo Fisher Scientific), and BSA was used as the standard.

Pimonidazole Treatment and Immunohistochemical Detection of Pimonidazole Adducts

Hypoxic cells were detected by pimonidazole adduct using the Hypoxyprobe™ kit (Hypoxyprobe, Burlington, MA, U.S.A.) according to the manufacturer’s instructions. Pimonidazole was added to the culture medium of CG-induced MCF-7 cell aggregates at a final concentration of 200 µM. After incubation for 2 h, the cell aggregates were embedded in iPGell (GenoStaff, Tokyo, Japan), according to the manufacturer’s instructions. The cell aggregates embedded in iPGell were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 2 h at 4 °C. After dehydration using PBS containing 20% sucrose, the blocks were rapidly frozen in liquid nitrogen and the sections with 6-µm thickness were prepared. The cells on the sections were permeabilized using 0.1% Triton-X 100 in PBS and were blocked with PBS containing 0.5% BSA and 2% normal goat serum for 1 h, and were reacted with anti-pimonidazole antibody (1 : 100 dilution) in 0.5% BSA containing PBS overnight at 4 °C. The samples were washed four times with PBS and incubated with an anti-mouse secondary antibody conjugated to Alexa Fluor 488 (1 µg/mL; Thermo Fisher Scientific). Counter-staining of nuclei was performed using DAPI. Finally, immunoreaction was observed using a confocal microscope A1Si (Nikon, Tokyo, Japan).

Electrophoretic Mobility Shift Assay (EMSA)

EMSA was performed using the LightShif Chemiluminescent EMSA Kit (Thermo Fisher Scientific), according to the manufacturer’s recommendations. Nuclear extracts were used to detect HIF-1α binding. The sequence of the 5′-Biotinylated probe containing the HIF responsive element (HRE) was below; 5′-Biotin-TCTGACGTGACCACACTCACCTC-3′, 5′-Biotin-GAGGTGAGTGTGGTCACGTACAGA-3′. Biotinylated sense and antisense oligonucleotides were annealed in an annealing buffer (10 mM Tris–HCl, 1 mM ethylenediaminetetraacetic acid (EDTA), 50 mM NaCl, pH 8.0). Nuclear extract (20 µg) was added to binding buffer in the kit containing 100 ng/µL poly (dI-dC), 5% glycerol, 5 mM MgCl₂, 0.5 mM EDTA. The mixture was then incubated for 10 min at room temperature before the addition of a 500 fmol biotinylated probe for an additional 20 min. After addition to the mixture with 5 µL of loading buffer, the DNA-protein complex was separated on a non-denaturing 4% polyacrylamide gel in 0.5× TBE buffer. The oligonucleotide probe was transferred by semidry blotting to a nylon membrane (FUJIFILM Wako Pure Chemical Corporation) and crosslinked on the membrane under a transilluminator. The membrane was incubated with blocking buffer that was contained in the EMSA kit for 15 min and reacted with streptavidin-horseradish peroxidase conjugate diluted with blocking buffer (1 : 300 dilution). After washing with the wash buffer in the kit, the membrane was developed by incubation in a chemiluminescence solution.

HIF Binding Activity Assay

HIF binding activity in cell lysates was determined using HIF-1α binding activity in an enzyme-linked immunosorbent assay (ELISA) kit (Cell Biolabs, Inc., San Diego, CA, U.S.A). Nuclear extracts (5 µg) were analyzed according to the manufacturer’s instructions.

Transfection

Transfection using Lipofectamine LTX (Thermo Fisher Scientific) supplemented with PLUS reagent (Thermo Fisher Scientific), suitable for MCF-7 cells, was performed according to the manufacturer’s instructions. Briefly, the plasmid was diluted with Opti-MEM (Thermo Fisher Scientific) to a final concentration of 5 µg/mL and was then mixed with PLUS reagent and Lipofectamine LTX. The cells and transfection mixture were added to RPMI 1640 medium containing 10% FBS without antibiotics. These cells were used in subsequent experimental procedures 24 h after transfection.

HIF Promoter Assay

The HIF reporter assay was performed using pGL4.42 (Promega, Madison, WI, U.S.A.), which contains 4-repeated HRE upstream of the firefly luciferase gene, and the control vector pGL4.27, which has a minimal promoter upstream of the firefly luciferase gene. The transfection of MCF-7 cells was performed as described previously. The reporter and internal control plasmid pRL-TK (Promega), which encodes the Renilla luciferase gene under the thymidine kinase promoter, was added at a mass ratio of 40 : 1. Firefly and Renilla luciferase activities were determined using the Dual-Luciferase Reporter Assay System (Promega), according to the manufacturer’s instructions. The transcriptional activities of the cells indicated as firefly luciferase activity was normalized to Renilla luciferase activity.

Quantitative (q)RT-PCR

Total RNA was prepared from MCF-7 cells using the RNeasy Plus Mini Kit (Qiagen, Valencia, CA, U.S.A.), according to the manufacturer’s instructions. Total RNA was reverse-transcribed using M-MLV reverse transcriptase (Thermo Fisher Scientific) and the poly(dT)₂₀ primer. Primers for qRT-PCR were designed using Primer Express Software (Life Technologies Corporation). The primer sequences were as follows: PGK1, 5′-GTCAGCCATGTGAGCACTGG-3′ and 5′-GTTGACTTAGGGGCTGTGCAC-3′; SLC2A1, 5′-AGCCAGCCAAAGTGACAAGAC-3′ and 5′-TGTGCTCCTGAGAGATCCTTAGG-3′; β-actin, 5′-TCCCTCACCCTCCCAAAAG-3′ and 5′-CATGGACGCGACCATCCT-3′. PCR was performed with SYBR Premix EX Taq II (TaKaRa Bio Inc., Shiga, Japan) using an Applied Biosystems 7500 real-time PCR system. The results were analyzed using accessory software in real-time PCR. The levels of PGK1 and SLC2A1 transcripts were normalized to those of β-actin.

Cell Viability Assay

Cell viability was determined using the CellTiter-Glo 3D viability assay reagent (Promega) or the Cell Counting Kit-8 (Dojindo, Kumamoto, Japan), according to the manufacturers’ instructions. When performing the CellTiter-Glo 3D viability assay, equal volumes of the reagent were added to the media of culturing cells, and the plates shaken gently for 5 min followed by incubation for 25 min at room temperature. The luminescence was measured using a microplate reader (ALVO X3, PerkinElmer, Inc., Waltham, MA, U.S.A). For the assay using Cell Counting Kit-8, after treatment with DXR for 24 h, the cells were incubated in a medium containing Cell Counting Kit-8 solution (1 : 10 dilution) at 37 °C for 2 h. Cell aggregates induced by CG were dispersed using 0.25% trypsin and 0.53 mM EDTA, followed by replating before the addition of Cell Counting Kit-8. The optical density at 450 nm was measured using a microplate reader.

DXR Accumulation Assay

The DXR accumulation assay was performed as described previously.^33,34) Briefly, the cells were incubated with DXR (4 µM) for 24 h, trypsinized, and washed three times with PBS. The cells were lysed in PBS containing 1% Triton X-100, followed by sonication. The amount of DXR in the lysates was quantified by measuring fluorescence intensity (Ex = 485 nm, Em = 530 nm) using a microplate reader.

RNA Interference

Silencer Select small interfering RNAs (siRNAs) (HIF-1 siRNA; IDs: s6539 and s6541, negative control siRNA; cat. Nos. 4390843 and 4390847) were obtained from Thermo Fisher Scientific. The transfection procedure using Lipofectamine RNAiMAX was performed according to manufacturer’s instructions. The cells and transfection mixture were incubated in RPMI 1640 medium containing 10% FBS without antibiotics. After 24 h of transfection, the cells were used for subsequent experiments.

RESULTS

Cells of the Central Part in CG-Induced MCF-7 Cell Aggregates Were Hypoxic

To elucidate whether the CG-induced MCF-7 cell aggregates were hypoxic, we performed morphological observations of the cell aggregates. CG-induced MCF-7 cell aggregates were of various sizes per the seeded cell density (Morimoto-Kamata, unpublished observations). As shown in Fig. 1A, confluently seeded cells formed large aggregates after stimulation with CG. The cell aggregates were of various sizes and shapes, and the large cell aggregates were more than 200 µm in diameter (Fig. 1A). It has been suggested that the center cells of its aggregates are hypoxic because cells more than 100 µm away from blood vessels are hypoxic owing to inadequate oxygen supply.³⁵⁾ To detect hypoxic cells, we performed pimonidazole labeling in CG-induced MCF-7 cells. Pimonidazole, a substituted 2-nitroimidazole, is an exogenous marker of hypoxic cells because reductively activated pimonidazole in hypoxic cells forms stable covalent adducts with thiol groups in proteins, peptides, and amino acids, and this adducted pimonidazole is detectable using antibodies against pimonidazole.³⁶⁾ Pimonidazole adducts were detected in cells at the center of the cell aggregates (Fig. 1B). Next, we examined the expression of two isoforms of the HIF-α subunit, HIF-1α and HIF-2α, in CG-induced MCF-7 cell aggregates because the HIF-α subunit is stably expressed in cells under hypoxic conditions; thus, expression of the HIF-α subunit is a marker of hypoxic cells.^4,5) Western blotting analysis demonstrated that these cell aggregates expressed both HIF-1α and -2α proteins, but the amount of HIF-1α and -2α proteins in the cell aggregates was lower than that in the monolayer cells cultured under 3% O₂ (Fig. 1C). As shown in Fig. 1B, we expected the reason for the low expression of the HIF-α subunit in the cell aggregates was that the only cells in the center of the aggregates were hypoxic, whereas almost all cells in monolayer cells under 3% O₂ became hypoxic. These results indicate that the cells in the center of the CG-induced MCF-7 cell aggregates were hypoxic.

Fig. 1. The Central Part in Cathepsin-G (CG)-Induced MCF-7 Cell Aggregates Is Hypoxic

A: Morphology of CG-induced MCF-7 cell aggregates. MCF-7 cells incubated with CG (40 nM) for 24 h. B: Localization of pimonidazole adducts in CG-induced MCF-7 cell aggregates. The cell aggregates were incubated with 200 µM pimonidazole for 2 h, and adducted pimonidazole was detected using an anti-pimonidazole antibody. Nuclei were visualized using 4′,6-diamidino-2-phenylindole (DAPI) staining. C: HIF-1α and HIF-2α protein were expressed in CG-induced MCF-7 cell aggregates. Nuclear extract (20 µg/lane) was analyzed by Western blotting. Proliferating Cell Nuclear Antigen (PCNA) was detected as a nuclear loading control.

HIF Is Functional in CG-Induced MCF-7 Cell Aggregates

Under hypoxic conditions, the HIF-α subunit dimerizes with HIF-1β, resulting in the induction of various gene expressions to adapt to hypoxic stress, for example, enzymes of the glycolytic pathway, glucose transporter, and angiogenesis. We investigated whether the expression of HIF-1α and -2α in the CG-induced MCF-7 cell aggregates was functional. HIF heterodimers exert their transcriptional activity by binding to the HRE upstream of HIF-target genes. EMSA using the HRE-containing oligonucleotide probe showed that the nuclear extract prepared from CG-induced cell aggregates contained HRE binding protein, and the elimination of the shifted band by the addition of excess unlabeled probe demonstrated that binding of the HRE binding protein to the HRE-containing oligonucleotide was specific (Fig. 2A). The shifted band of the cell aggregates induced by CG was weaker than that of the cells cultured under 3% O₂ similar to the results of Western blotting (Fig. 1B). The HIF binding assay based on ELISA showed the same results (Fig. 2B). A reporter assay using cells transfected with the HIF reporter plasmid pGL4.42, which contains 4-repeated HRE upstream of the luciferase gene, showed the transcriptional activity of CG-induced MCF-7 cell aggregates was significantly increased compared to that of the cells cultured under normoxia (Fig. 2C). Subsequently, we examined the upregulation of HIF-target genes phosphoglycerate kinase (PGK) 1 and solute carrier family 2 member 1 (SLC2A1) using qRT-PCR. The expression of PGK1 and SLC2A1 mRNA was significantly elevated in the cell aggregates (Fig. 3). These data indicate that HIFs are functional and that HIF-target genes are induced in CG-induced MCF-7 cell aggregates.

Fig. 2. CG-Induced MCF-7 Aggregates Contain Transcriptional Activity through Hypoxia-Inducible Factor (HIF) Responsive Elements

A: Electrophoretic mobility shift assay analyzed nuclear extract prepared from MCF-7 cells. Biotinylated hypoxia-responsive element (HRE)-containing probes (500 fmol) were incubated with nuclear extract (20 µg), and the complex was analyzed by electrophoretic mobility shift assay. Excess of unlabeled probe (2 pmol) was added to the mixture to compete for specific binding. B: Binding to HRE-containing oligonucleotides was detected by ELISA. Nuclear extracts (5 µg) prepared from MCF-7 cells were analyzed. C: Promoter assay using HRE-containing reporter gene. The cells were transfected with HRE containing plasmid pGL4.42 or minimal promoter containing plasmid pGL4.27. The results were normalized by Renilla luciferase activity that co-transfected pRL-TK plasmid. The results are expressed as means ± standard deviation (S.D.) (n = 3); * p < 0.05, Dunnett’s test.

Fig. 3. HIF-Target Genes Were Upregulated in CG-Induced MCF-7 Cell Aggregates

Quantification of mRNA expression was determined by quantitative RT-PCR. A: PGK1, B: SLC2A1. The results are expressed as means ± S.D. (n = 3); * p < 0.05, Dunnett’s test.

DXR Cytotoxicity Was Reduced in CG-Induced MCF-7 Cell Aggregates

HIFs induce the ABCB1 gene that encodes P-gp, and efflux various substrate drugs in cells under hypoxia, resulting in multidrug resistance.^6,7) DXR is an anthracycline-derivative anticancer agent. It is a substrate of P-gp and can be easily monitored by measuring its fluorescence.³⁷⁾ Therefore, we evaluated the chemoresistant phenotype of DXR in CG-treated MCF-7 cell aggregates. Cell viability of CG-induced cell aggregates after treatment with 40 µM DXR was significantly elevated compared with the monolayer cells under normoxia, although cell viability of the aggregates and the monolayer cells under 3% O₂ treated with less than 4 µM DXR was equal to that of the monolayer cells under normoxia (Fig. 4). We note that cell viability was measured using a CellTiter-Glo 3D cell viability assay, which lyses all the cells in the well and detects the ATP contained therein using luciferase-based luminescence technology, because of its high throughput screening (including in other anticancer drugs, Morimoto-Kamata, unpublished data), in the present experiment. However, the assay tends to reveal higher levels than morphological observations (Figs. 4A, B). The higher viability was speculated to be due to the additional detection of ATP in contracted early apoptotic cells (Fig. 4A). For analysis of cell viability in subsequent experiments, we used Cell Counting Kit-8, which is a soluble tetrazolium reduction assay and therefore reflects cell metabolism, mentioned below. The data suggested that aggregation of MCF-7 cells induced by CG decreased DXR sensitivity.

Fig. 4. CG-Induced MCF-7 Cell Aggregates Exhibited Reducing Sensitivity against Doxorubicin (DXR)

After treatment with CG (40 nM) or 3% O₂ for 24 h, MCF-7 cells were incubated with DXR for 24 h. A: Cell morphology after 24 h incubation with DXR. B: Cell viabilities after incubation with DXR were determined by CellTiter-Glo 3D assay. Cell viabilities of 40 µM DXR treated cells were analyzed using Dunnett’s test. * p < 0.05. * p < 0.05.

Decrease in DXR Sensitivity of CG-Induced MCF-7 Cell Aggregates Is Not Mediated by P-gp

Colorectal cancer and endothelial cells reportedly acquire chemoresistance via P-gp expression under hypoxic conditions in a HIF-dependent manner.^6,7) Hence, we examined whether P-gp expression and decreased sensitivity to DXR were associated with P-gp-mediated DXR efflux in CG-induced MCF-7 cell aggregates. Western blot analysis showed that P-gp was detected in human hepatocarcinoma HepG2 which constantly expressed P-gp protein, but not in the CG-induced MCF-7 cell aggregates (Fig. 5A). We hypothesized that other drug efflux transporter(s) might be associated with a decrease in DXR sensitivity in CG-induced MCF-7 cell aggregates in a HIF-dependent manner. Indeed, HIF induced by CoCl₂, a chemical inducer of HIF that mimics hypoxia, induces multidrug resistance protein MRP1 mRNA in the colon cancer cell line Lovo.³⁸⁾ Transfection of HIF-1α siRNA downregulated the expression of MDR1 and MRP1 mRNA and increased chemoresistance was reversed in the cisplatin-resistant human lung carcinoma cell line A594/CDDP.³⁹⁾ However, there was only modest expression of MRP1 protein in the 3% O₂ monolayer cells and CG-induced cell aggregates, similar to in the monolayer cells under normoxia (Fig. 5B). Therefore, we measured intracellular DXR accumulation after 24 h of incubation. Intracellular DXR accumulation was significantly increased in the monolayer cells under 3% O₂ and the CG-induced cell aggregates (Fig. 5C). We noted that intracellular DXR in HepG2 cells was significantly increased in the presence of the P-gp inhibitor verapamil and tariquidar, indicating that intracellular DXR was effluxed through P-gp in HepG2 cells (Supplementary Fig. S1). Unexpectedly, in the monolayer MCF-7 cells under 3% O₂, P-gp expression was also undetectable, MRP1 was only expressed modestly, and intracellular DXR was significantly increased (Figs. 5A, B). These results indicated that CG-induced MCF-7 cell aggregates and monolayer MCF-7 cells under 3% O₂ did not have a functional DXR efflux system, including P-gp and MRP1.

Fig. 5. CG-Induced MCF-7 Cell Aggregates and Hypoxic Monolayer MCF-7 Cells Did Not Contain DXR Efflux System

A, B: P-gp (A) and MRP1 (B) expression was detected by Western blotting. Whole-cell lysates were analyzed at 20 µg/lane. Human hepatocarcinoma HepG2 cells were used as a positive control. C: Amount of intracellularly accumulated DXR in the CG-induced MCF-7 cell aggregates. After treatment with CG (40 nM) or 3% O₂ for 24 h, the cells were incubated with DXR (4 µM) for 24 h. The amounts of DXR incorporated into the cells were quantified using the cell lysates. The results are expressed as the means ± S.D. (n = 3); * p < 0.05, Dunnett’s test.

Reduction of CG-Induced MCF-7 Cell Aggregate Sensitivity to DXR Was HIF-Independent

Finally, to examine whether the reduction of sensitivity to DXR in CG-induced MCF-7 cells depends on HIFs, we investigated the effect of the HIF inhibitor YC-1 or HIF-1α siRNA transfection on the reduction of sensitivity to DXR in the cell aggregates. YC-1 blocks HIF-1α and -2α expression at the post-transcriptional level and consequently inhibits the transcriptional activity of HIF under hypoxic conditions.⁴⁰⁾ If reduction of sensitivity to DXR in cell aggregates is HIF-dependent, treatment with YC-1 or HIF-1α siRNA transfection should re-sensitize them to DXR. We confirmed that YC-1 inhibited HIF promoter activity in a dose-dependent manner (Supplementary Fig. S2). It was estimated that YC-1 at 5 µM inhibited HIF reporter activity by approximately 63% compared to vehicle-treated cells cultured under 3% O₂. In the presence of YC-1 at 5 µM, DXR cytotoxicity was not enhanced in the cell aggregates (Figs. 6A, B). On the other hand, the viability of the monolayer cells cultured under 3% O₂ treated with DXR decreased significantly with the addition of YC-1. Because YC-1 also exhibited significant cytotoxicity to the monolayer cells cultured under 21% O₂, we compared the effect of YC-1 on DXR cytotoxicity in the monolayer cells under 21% O₂ and 3% O₂. The viability of the monolayer cells under 3% O₂ after DXR treatment was significantly lower than that under 21% O₂ in the presence of YC-1 (Figs. 6A, B). The data indicated that YC-1 was re-sensitized to DXR in the monolayer cells cultured under 3% O₂; therefore, we concluded the decrease in sensitivity to DXR in the monolayer cells under 3% O₂ was dependent on HIF. We note that hypoxia mimicry with CoCl₂ induced the expression of HIF-1α and -2α, but only slightly increased the viability of monolayer MCF-7 cells after DXR treatment (Supplementary Fig. S3). Subsequently, we investigated whether HIF-1α knockdown using siRNA restores DXR sensitivity to the cell aggregates or not, because many reports have indicated that HIF-1α plays important roles in DXR resistance.⁴¹⁾ Both independent siRNA duplexes effectively downregulated the expression of HIF-1α protein in the monolayer cells cultured under 3% O₂ and CG-induced cell aggregates (Fig. 7A). Subsequently, we determined the effect of HIF-1α knockdown on DXR sensitivity of hypoxic monolayer cells and CG-induced cell aggregates. Notably, control siRNA#2 upregulated the viability of the cells under all culture conditions when compared with that of control siRNA#1, and the results were reproducible (Fig. 7B). The siRNA were commercially available from Thermo Scientific as negative control siRNA. The effects are predicted to be off-target effects derived from control siRNA. Hence, the siRNA#1 and#2 data were integrated and are illustrated in Figs. 7C and D. HIF-1α siRNA decreased the cell viability of 3% O₂-treated monolayer cells significantly but did not affect that of CG-induced cell aggregates (Fig. 7D). The data suggest that reducing sensitivity to DXR of the monolayer cells under 3% O₂ was dependent on HIF but not CG-induced cell aggregates.

Fig. 6. Decreasing Sensitivity to DXR of Hypoxic Monolayer MCF-7 Cells Depended on HIF, but That of CG-Induced MCF-7 Cell Aggregates Did Not

A: Cell viability after incubation with DXR in the presence or absence of the HIF inhibitor YC-1. MCF-7 cells were added with YC-1 (5 µM) at 1 h before CG (40 nM) or 3% O₂ treatment. After 24 h, the cells were incubated with DXR (40 µM) for 24 h. Cell viability was determined using Cell Counting Kit-8. The results are expressed as the means ± S.D. (n = 3); ** p < 0.01; NS, not significant, Student’s t-test. B: Effect of YC-1 on the cell viability of DXR-treated MCF-7 cells. The ratio of the cell viability in the presence of YC-1 to that in the absence of YC-1 after DXR incubation described in Fig. 6A was shown. The results are expressed as means ± S.D. (n = 3); * p < 0.05, ** p < 0.01, Student’s t-test.

Fig. 7. Hypoxia-Inducible Factor 1 Alpha (HIF-1α) Knockdown Using siRNA Cancelled Decreasing DXR Sensitivity of Hypoxic Monolayer MCF-7 Cells but Did Not Cacel That of CG-Induced Cell Aggregates

A: HIF-1α knockdown using siRNA. Nuclear extract (20 µg/lane) of MCF-7 cells transfected with control or HIF-1α siRNA were analyzed by Western blotting. B: Cell viability of DXR-treated cells transfected with HIF-1α siRNA. After 24-h transfection with siRNA, cells were incubated with CG (40 nM) or 3% O₂ for 24 h and subsequently DXR (40 µM). Cell viability was determined using Cell Counting Kit-8, and that of control siRNA-transfected cells was expressed as 100%. The results are expressed as means ± S.D. (n = 4). C: Integrated cell viability of two independent siRNA-transfected MCF-7 cells after DXR incubation. Cell viability of the cells transfected with siRNA#1 and#2 was integrated, and the cell viability of vehicle-treated control siRNA was expressed as 100%. The results are expressed as means ± S.D. (n = 4); * p < 0.05, Student’s t-test. D: Effect of HIF-1α knockdown on the cell viability of DXR-treated MCF-7 cells. The ratio of the cell viability of HIF-1α-transfected cells to that of control siRNA-transfected cells after DXR incubation described in Fig. 7C is shown. The results are expressed as means ± S.D. (n = 4); * p < 0.05; NS, not significant, Dunnett’s test.

In conclusion, CG-induced MCF-7 cell aggregates exhibited decreasing sensitivity to DXR in a HIF-independent mechanism, whereas 3% O₂ conferred decreasing sensitivity to DXR in the monolayer cultured cells in a HIF-1α dependent but P-gp- and MRP1-independent manner.

DISCUSSION

In the present study, we investigated the resistance to chemotherapy of MCF-7 cells conferred by CG. We demonstrated that the center of CG-induced MCF-7 cell aggregates was hypoxic; however, decreasing sensitivity of the aggregates to DXR at 40 µM was independent of HIF, P-gp, and MRP1. CG treatment alone forms aggregates with a maximum minor axis of 300 µm, as shown in Fig. 1A. If combined with other techniques, such as rotating wall vascular cultures and spinner cultures, as opposed to incubation with CG only, it is speculated that larger cell aggregates can be created. With the formation of large aggregates, the hypoxic region becomes larger and is presumed to be more resistant to chemotherapy. However, combining CG with other culture techniques is an artificial method, and not an in vivo phenomenon; therefore, we examined chemoresistance of the cell aggregates formed using CG alone.

HIFs are master regulators responsible for hypoxic stress and upregulate the transcription of hundreds of HIF-target genes, such as erythropoietin, glycolysis enzymes, VEGF, and glucose transporter-1, for survival and adaptation under hypoxic.^4,5) Previous reports have indicated that hypoxia-induced P-gp expression is responsible for chemoresistance against antitumor agents in HMVECs, a primary endothelial cell culture isolated from the human dermis, and human clonal colon carcinoma cells T84, Caco2, and LoVo cells under hypoxic conditions.^6,7) MRP1 was also expressed in CoCl2-treated LoVo cells.³⁸⁾ In addition, Doublier et al. showed that intracellular DXR accumulation and DXR-induced cytotoxicity were suppressed in 3% O₂-treated MCF-7 cells and NE-induced MCF-7 spheroids compared with the monolayer cells under normoxia.³⁷⁾ Thus, the conclusion of Doublier et al. that NE-induced MCF-7 spheroids express HIF-1α due to hypoxia contradicts our data in this study. They demonstrated that resistance of NE-induced MCF-7 cell spheroids and 3% O₂-treated monolayer MCF-7 cells to DXR was caused by HIF-induced P-gp dependent mechanism from the following three pieces of evidence: 1) P-gp protein was expressed, 2) intracellular DXR was decreased compared to the monolayer cells under normoxia, 3) decreased cell viability after incubation with DXR was canceled by treatment with YC-1 and HIF-1α siRNA transfection. Doublier et al. concluded that these spheroids were hypoxic, and HIF-1α immunoreactivity was detected throughout spheroids, although the spheroids were less than 50 µm in diameter. It has been considered that the center of cell aggregates more than 200 µm in diameter is a hypoxic condition due to insufficient penetration of the oxygen supply.³⁷⁾ In this study, we demonstrated that the CG-induced cell aggregates were more than 200 µm in diameter, contained pimonidazole-positive cells, and expressed HIF-1α and -2α (Fig. 1); therefore, the central part of the cell aggregation was presumed to be hypoxia due to insufficient permeation of oxygen. It is unclear why there was a discrepancy between the data reported by Doublier et al. and our results. Similar to our results, Greijer et al. also demonstrated that P-gp was not detected in the monolayer MCF-7 cells cultured under 1% O₂ although HIF-1 was expressed.⁴²⁾ We speculate that HIF-1α expression in NE-induced spheroids that were small in size, as described by Doublier et al., is caused by an oxygen-independent pathway, namely, a non-canonical mechanism. The non-canonical pathway upregulates the transcription and translation of the HIF-α subunit under normoxia.⁴³⁾ The major initiator of the non-canonical pathway is the activation of receptor tyrosine kinases, such as the receptors of growth factors and cytokines. Our previous report indicated that NE stimulates transient activation of IGF-1 signaling in MCF-7 cells.³⁰⁾ In addition, Fukuda et al. showed that IGF-1 activates transcription and translation of HIF1A, regardless of the normal oxygen concentration.⁴⁴⁾ Therefore, HIF-1α expression in NE-induced spheroids, as reported by Doublier et al., may be caused by transient IGF-1R activation under normoxia. Because CG also evokes persistent activation of IGF-1 signaling,³⁰⁾ we expect that HIF expression in cell aggregates is due to insufficient oxygen supply as well as CG-evoked IGF-1 signaling. However, the center of the aggregates was suggested to be hypoxic because of the presence of pimonidazole adducts (Fig. 1C). Thus, the mechanism of DXR resistance may differ between the cell aggregates induced by CG and NE.

Hypoxia is an important factor in tumor chemoresistance. Interestingly, HIF-dependent and HIF-independent chemoresistance have been confirmed under hypoxia.⁴⁵⁾ Our study showed that CoCl₂ slightly reduced sensitivity to DXR cells in monolayer normoxic MCF-7 cells (Supplementary Fig. S3). In addition, it has been demonstrated that suppression of HIFs through gene knockdown or the use of small-molecule inhibitors did not completely abrogate resistance to cytostatic drugs, such as cisplatin, etoposide, doxorubicin, and ellipticine.⁴⁵⁾ HIF-independent chemoresistance of prostate cancer cells under hypoxic conditions may be due to increased expression of anti-apoptotic factors, including Bcl-2 family proteins and cellular inhibitors of apoptosis.⁴⁶⁾ Inhibition of the phosphoinositide 3-kinase (PI3K) pathway, nuclear factor kappa-B, and cycloxygenase-2 can also partially reverse resistance to cytostatics under hypoxia, implying a role for these pathways in hypoxia-induced drug resistance.⁴⁵⁾ Acidosis and nutrient starvation derived from the tumor microenvironment also cause chemoresistance.^47,48) The tumor microenvironment is frequently acidic because its specific metabolism shifts to the glycolytic pathway. Under hypoxia, glucose is mainly converted to lactate to produce ATP, which is known as the Warburg effect, and the increased lactate reduces the pH of the extracellular fluid. Because weakly basic compounds such as DXR are ionized under acidic conditions and cannot cross the cell membrane, it is expected that DXR does not accumulate in cells under hypoxic conditions. However, this is inconsistent with our data, that is, the intracellular accumulation of DXR was increased in the monolayer MCF-7 cells under monolayer MCF-7 cells under 3% O₂ (Fig. 5C). The reason for the increase in DXR accumulation in these cells remains unclear. One possibility is that 3% O²⁻treated monolayer cells might be able to accumulate more DXR because of acquired reduction in sensitivity to DXR. Indeed, HepG2 cells intracellularly accumulated more DXR when compared with MCF-7 cells (HepG2 cells: about 700 ng/mg protein; MCF-7 cells: 120 ng/mg protein, Fig. 5C and Supplementary Fig. S1). Another predicted mechanism of reducing sensitivity to DXR in cell aggregates is change in subcellular localization of DXR from the nucleus to the cytoplasm.⁴⁹⁾ In a DXR-sensitive mouse mammary cancer cell line EMT6-S, DXR mainly localizes in nuclei of the cells, whereas DXR is weakly detected in cytoplasm of a DXR-resistant subline, EMT6-R, with no nuclear involvement. We would like to examine subcellular localization of DXR in the cells that are in a hypoxic region in the CG-induced MCF-7 cell aggregates further.

Abnormal survival signals are also an important mechanism underlying chemoresistance to apoptosis-inducing chemotherapeutic drugs, including DXR. Constitutively activated survival signaling frequently occurs by aberrant synthesis of growth factors and/or by activating mutations in growth factor receptors, and the continuous activation of tyrosine kinase coupled with the receptor evokes downstream signaling of the PI3K/protein kinase B (Akt)/mammalian target of rapamycin (mTOR) axis in tumor cells. Akt plays a central role in the regulation of survival signaling by inhibiting apoptosis-promoting factors, such as Bad and Bax, and inducing anti-apoptotic factors, such as Bcl-2. Our previous study showed that CG stimulation persistently activates Akt via an IGF-1-dependent mechanism.²⁶⁾ In addition, we have preliminary data showing that CG induces Bcl-2 expression in MCF-7 cells (Morimoto-Kamata, R., unpublished observations). Our data indicate that DXR-induced apoptosis was suppressed by Akt-dependent survival signaling via a HIF-independent mechanism in CG-induced MCF-7 cell aggregates.

In summary, we revealed that CG-induced MCF-7 cell aggregates expressed functional HIF and decrease sensitivity to DXR via HIF- and P-gp-independent mechanisms.

Acknowledgments

This work was supported by a KAKENHI Grant (JP15K18417 to MK) from MEXT of Japan.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1) Schönenberger MJ, Kovacs WJ. Hypoxia signaling pathways: modulators of oxygen-related organelles. Front. Cell Dev. Biol., 3, 42 (2015).
• 2) Schaaf MB, Garg AD, Agostinis P. Defining the role of the tumor vasculature in antitumor immunity and immunotherapy. Cell Death Dis., 9, 115 (2018).
• 3) Martínez-Reyes I, Chandel NS. Cancer metabolism: looking forward. Nat. Rev. Cancer, 21, 669–680 (2021).
• 4) Masoud GN, Li W. HIF-1alpha pathway: role, regulation, and intervention for cancer therapy. Acta. Pharm. Sin. B., 5, 378–389 (2015).
• 5) Pezzuto A, Carico E. Role of HIF-1 in cancer progression: novel insights. A Review. Curr. Mol. Med., 18, 343–351 (2018).
• 6) Comerford KM, Wallace TJ, Karhausen J, Louis NA, Montalto MC, Colgan SP. Hypoxia-inducible factor-1-dependent regulation of the multidrug resistance (MDR1) gene. Cancer Res., 62, 3387–3394 (2002).
• 7) Chen J, Ding Z, Peng Y, Pan F, Li J, Zou L, Zhang Y, Liang H. HIF-1α inhibition reverses multidrug resistance in colon cancer cells via downregulation of MDR1/P-glycoprotein. PLOS ONE, 9, e98882 (2014).
• 8) Farc O, Cristea V. An overview of the tumor microenvironment, from cells to complex networks. Exp. Ther. Med., 21, 96 (2021).
• 9) Mantovani A, Cassatella MA, Costantini C, Jaillon S. Neutrophils in the activation and regulation of innate and adaptive immunity. Nat. Rev. Immunol., 11, 519–531 (2011).
• 10) Rosales C. Neutrophil: a cell with many roles in inflammation or several cell types? Front. Physiol., 9, 113 (2018).
• 11) Jensen HK, Donskov F, Marcussen N, Nordsmark M, Lundbeck F, von der Maase H. Presence of intratumoral neutrophils is an independent prognostic factor in localized renal cell carcinoma. J. Clin. Oncol., 27, 4709–4717 (2009).
• 12) Walsh SR, Cook EJ, Goulder F, Justin TA, Keeling NJ. Neutrophil-lymphocyte ratio as a prognostic factor in colorectal cancer. J. Surg. Oncol., 91, 181–184 (2005).
• 13) Papayannopoulos V. Neutrophil extracellular traps in immunity and disease. Nat. Rev. Immunol., 18, 134–147 (2018).
• 14) Masucci MT, Minopoli M, Del Vecchio S, Carriero MV. The emerging role of Neutrophil Extracellular Traps (NETs) in tumor progression and metastasis. Front. Immunol., 11, 1749 (2020).
• 15) Yang L, Liu Q, Zhang X, Liu X, Zhou B, Chen J, Huang D, Li J, Li H, Chen F, Liu J, Xing Y, Chen X, Su S, Song E. DNA of neutrophil extracellular traps promote cancer metastasis via CCDC25. Nature, 583, 133–138 (2020).
• 16) Lerman I, Hammes SR. Neutrophil elastase in the tumor microenvironment. Steroids, 133, 96–101 (2018).
• 17) Houghton AM, Rzymkiewicz DM, Ji H, Gregory AD, Egea EE, Metz HE, Stolz DB, Land SR, Marconcini LA, Kliment CR, Jenkins KM, Beaulieu KA, Mouded M, Frank SJ, Wong KK, Shapiro SD. Neutrophil elastase-mediated degradation of IRS-1 accelerates lung tumor growth. Nat. Med., 16, 219–223 (2010).
• 18) Deryugina E, Carre A, Ardi V, Muramatsu T, Schmidt J, Pham C, Quigley JP. Neutrophil elastase facilitates tumor cell intravasation and early metastatic events. iScience, 23, 101799 (2020).
• 19) Nawa M, Osada S, Morimitsu K, Nonaka K, Futamura M, Kawaguchi Y, Yoshida K. Growth effect of neutrophil elastase on breast cancer: favorable action of sivelestat and application to anti-HER2 therapy. Anticancer Res., 32, 13–19 (2012).
• 20) Gao S, Zhu H, Zuo X, Luo H. Cathepsin G and its role in inflammation and autoimmune diseases. Arch. Rheumatol., 33, 498–504 (2018).
• 21) Nakajima K, Powers JC, Ashe BM, Zimmerman M. Mapping the extended substrate-binding site of cathepsin G and human leukocyte elastase. Studies with peptide substrates related to the a1-protease inhibitor reactive site. J. Biol. Chem., 254, 4027–4032 (1979).
• 22) Tanaka T, Minematsu Y, Reilly CF, Travis J, Powers JC. Human leukocyte cathepsin G. Subsite mapping with 4-nitroanilides, chemical modification, and effect of possible cofactors. Biochemistry, 24, 2040–2047 (1985).
• 23) Iacoviello L, Kolpakov V, Salvatore L, Amore C, Pintucci G, de Gaetano G, Donati MB. Human endothelial cell damage by neutrophil-derived cathepsin G. Role of cytoskeleton rearrangement and matrix-bound plasminogen activator inhibitor-1. Arterioscler. Thromb. Vasc. Biol., 15, 2037–2046 (1995).
• 24) Yui S, Tomita K, Kudo T, Ando S, Yamazaki M. Induction of multicellular 3-D spheroids of MCF-7 breast carcinoma cells by neutrophil-derived cathepsin G and elastase. Cancer Sci., 96, 560–570 (2005).
• 25) Kudo T, Kigoshi H, Hagiwara T, Takino T, Yamazaki M, Yui S. Cathepsin G, a neutrophil protease, induces compact cell-cell adhesion in MCF-7 human breast cancer cells. Mediators Inflamm., 2009, 850940 (2009).
• 26) Morimoto-Kamata R, Yui S. Insulin-like growth factor-1 signaling is responsible for cathepsin G-induced aggregation of breast cancer MCF-7 cells. Cancer Sci., 108, 1574–1583 (2017).
• 27) Gunti S, Hoke ATK, Vu KP, London NR Jr. Organoid and spheroid tumor models: techniques and applications. Cancers (Basel), 13, 874 (2021).
• 28) Morimoto-Kamata R, Mizoguchi S, Ichisugi T, Yui S. Cathepsin G induces cell aggregation of human breast cancer MCF-7 cells via a 2-step mechanism: catalytic site-independent binding to the cell surface and enzymatic activity-dependent induction of the cell aggregation. Mediators Inflamm., 2012, 456462 (2012).
• 29) Yui S, Osawa Y, Ichisugi T, Morimoto-Kamata R. Neutrophil cathepsin G, but not elastase, induces aggregation of MCF-7 mammary carcinoma cells by a protease activity-dependent cell-oriented mechanism. Mediators Inflamm., 2014, 971409 (2014).
• 30) Morimoto-Kamata R, Tsuji D, Yui S. Cathepsin G-induced insulin-like growth factor (IGF) elevation in MCF-7 medium is caused by proteolysis of IGF binding protein (IGFBP)-2 but not of IGF-1. Biol. Pharm. Bull., 43, 1678–1686 (2020).
• 31) Tomlinson JS, Alpaugh ML, Barsky SH. An intact overexpressed E-cadherin/alpha, beta-catenin axis characterizes the lymphovascular emboli of inflammatory breast carcinoma. Cancer Res., 61, 5231–5241 (2001).
• 32) Alpaugh ML, Tomlinson JS, Kasraeian S, Barsky SH. Cooperative role of E- cadherin and sialyl-Lewis X/A-deficient MUC1 in the passive dissemination of tumor emboli in inflammatory breast carcinoma. Oncogene, 21, 3631–3643 (2002).
• 33) Mellor HR, Callaghan R. Accumulation and distribution of doxorubicin in tumor spheroids: the influence of acidity and expression of P-glycoprotein. Cancer Chemother. Pharmacol., 68, 1179–1190 (2011).
• 34) Kauffman MK, Kauffman ME, Zhu H, Jia Z, Li YR. Fluorescence-based assays for measuring doxorubicin in biological systems. React. Oxyg. Species (Apex), 2, 432–439 (2016).
• 35) Helmlinger G, Yuan F, Dellian M, Jain RK. Interstitial pH and pO2 gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation. Nat. Med., 3, 177–182 (1997).
• 36) Varghese AJ, Gulyas S, Mohindra JK. Hypoxia-dependent reduction of 1-(2-nitro-1-imidazolyl)-3-methoxy-2-propanol by Chinese hamster ovary cells and KHT tumor cells in vitro and in vivo. Cancer Res., 36, 3761–3765 (1976).
• 37) Doublier S, Belisario DC, Polimeni M, Annaratone L, Riganti C, Allia E, Ghigo D, Bosia A, Sapino A. 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, 12, 4 (2012).
• 38) Lv Y, Zhao S, Han J, Zheng L, Yang Z, Zhao L. Hypoxia-inducible factor-1α induces multidrug resistance protein in colon cancer. Onco. Targets Ther., 8, 1941–1948 (2015).
• 39) Min L, Chen Q, He S, Liu S, Ma Y. Hypoxia-induced increases in A549/CDDP cell drug resistance are reversed by RNA interference of HIF-1alpha expression. Mol. Med. Rep., 5, 228–232 (2012).
• 40) Yeo EJ, Chun YS, Cho YS, Kim J, Lee JC, Kim MS, Park JW. YC-1: a potential anticancer drug targeting hypoxia-inducible factor 1. J. Natl. Cancer Inst., 95, 516–525 (2003).
• 41) Warfel NA, El-Deiry WS. HIF-1 signaling in drug resistance to chemotherapy. Curr. Med. Chem., 21, 3021–3028 (2014).
• 42) Greijer AE, de Jong MC, Scheffer GL, Shvarts A, van Diest PJ, van der Wall E. Hypoxia-induced acidification causes mitoxantrone resistance not mediated by drug transporters in human breast cancer cells. Cell. Oncol., 27, 43–49 (2005).
• 43) Iommarini L, Porcelli AM, Gasparre G, Kurelac I. Non-canonical mechanisms regulating hypoxia-inducible factor 1 alpha in cancer. Front. Oncol., 7, 286 (2017).
• 44) Fukuda R, Hirota K, Fan F, Jung YD, Ellis LM, Semenza GL. Insulin-like growth factor 1 induces hypoxia-inducible factor 1-mediated vascular endothelial growth factor expression, which is dependent on MAP kinase and phosphatidylinositol 3-kinase signaling in colon cancer cells. J. Biol. Chem., 277, 38205–38211 (2002).
• 45) Doktorova H, Hrabeta J, Khalil MA, Eckschlager T. Hypoxia-induced chemoresistance in cancer cells: THE role of not only HIF-1. Biomed. Pap. Med. Fac. Univ. Palacky Olomouc Czech Repub., 159, 166–177 (2015).
• 46) Coffey RN, Morrissey C, Taylor CT, Fitzpatrick JM, Watson RW. Resistance to caspase-dependent, hypoxia-induced apoptosis is not hypoxia-inducible factor-1 alpha mediated in prostate carcinoma cells. Cancer, 103, 1363–1374 (2005).
• 47) Pillai SR, Damaghi M, Marunaka Y, Spugnini EP, Fais S, Gillies RJ. Causes, consequences, and therapy of tumors acidosis. Cancer Metastasis Rev., 38, 205–222 (2019).
• 48) Kato Y, Ozawa S, Miyamoto C, Maehata Y, Suzuki A, Maeda T, Baba Y. Acidic extracellular microenvironment and cancer. Cancer Cell Int., 13, 89 (2013).
• 49) Burrow SM, Phoenix DA, Wainwright M, Tobin MJ. Intracellular localisation studies of doxorubicin and Victoria Blue BO in EMT6-S and EMT6-R cells using confocal microscopy. Cytotechnology, 39, 15–25 (2002).