Abstract
Cisplatin (CDDP) is widely prescribed for the treatment of various cancers including bladder cancers, whereas its clinical use for breast cancer chemotherapy is restricted owing to easy acquisition of the chemoresistance. Here, we established a highly CDDP-resistant variant of human breast cancer MCF7 cells and found that procuring the resistance aberrantly elevates the expression of aldo-keto reductase (AKR) 1C3. Additionally, MCF7 cell sensitivity to CDDP was decreased and increased by overexpression and knockdown, respectively, of AKR1C3, clearly inferring that the enzyme plays a crucial role in acquiring the CDDP resistance. The CDDP-resistant cells suppressed the formation of cytotoxic reactive aldehydes by CDDP treatment, and the suppressive effects were almost completely abolished by pretreating with AKR1C3 inhibitor. The resistant cells also exhibited the elevated glutathione amount and 26S proteasomal proteolytic activities, and their CDDP sensitivity was significantly augmented by pretreatment with an inhibitor of glutathione synthesis or proteasomal proteolysis. Moreover, the combined treatment with inhibitors of AKR1C3, glutathione synthesis and/or proteasomal proteolysis potently overcame the CDDP resistance and docetaxel cross-resistance. Taken together, our results suggest that the combination of inhibitors of AKR1C3, glutathione synthesis and/or proteasomal proteolysis is effective as an adjuvant therapy to enhance CDDP sensitivity of breast cancer cells.
Abbreviations
- AKR
aldo-keto reductase
- BPS
3-bromo-5-phenylsalicylic acid
- BSO
buthionine sulfoximine
- CDDP
cis-diamminedichloroplatinum
- CDDP-R
CDDP-resistant MCF7
- DPBS
Dulbecco’s phosphate-buffered saline
- DTNB
5,5′-dithiobis(2-nitrobenzoic acid)
- DTX
docetaxel
- GCL
glutamate-cysteine ligase
- GPx
glutathione peroxidase
- GR
glutathione reductase
- GSH
reduced glutathione
- GST
glutathione S-transferase
- HNE
4-hydroxy-2-nonenal
- Keap1
Kelch-like ECH associated protein 1
- MCA
4-methylcoumaryl-7-amide
- Nrf2
nuclear factor erythroid 2-related factor 2
- PCR
polymerase-chain reaction
- PG
prostaglandin
- ROS
reactive oxygen species
- SFN
sulforaphane
- siRNA
small-interfering RNA
- TOL
tolfenamic acid
- UDCA
ursodeoxycholic acid
Introduction
A platinum-based anticancer drug cisplatin (cis-diamminedichloroplatinum, CDDP) is widely prescribed for the treatment of various types of solid cancers such as bladder, testicular, ovarian and lung cancers (1). One of the best understood anticancer mechanisms of CDDP is the covalent binding to purine bases in DNA, consequently leading to the inhibition of replication, transcription and excision repair of nucleic acids (1, 2). In addition, it is considered that the cytolethal action is in part due to the production of reactive oxygen species (ROS) and subsequent activation of apoptotic signalling pathways including alteration in mitochondrial proteins and activation of p53 and caspases (3, 4). Despite the superior efficacy of CDDP on cancer chemotherapy, its clinical use for the treatment of other cancers including breast cancer is restricted owing to low clinical response rates of its administration and high incidence of severe adverse effects, such as hepatic, renal, neurologic and auditory disorders (1). In addition to these undesired outcomes, CDDP exposure to cancer cells is known to easily develop drug resistance, the so-called platinum resistance. Therefore, it is desired that discoveries of pivotal factor(s) for ameliorating the platinum resistance lead to the exploitation of a novel strategy to cure the ineffective treatment of cancer by the drug. Previous literature indicates that the platinum resistance is ascribed to alterations in the cellular influx and efflux of the drug, promotion of the drug metabolism and DNA repair and potentiation of the cytoprotective capacity (5). To date, poly (ADP-ribose) polymerase (6), RAD6 (7), sequestosome 1 (SQSTM1/p62) (8), toll-like receptor 7 (9) and serine-arginine protein kinase 1 (10) have been proposed as the target genes for overcoming platinum resistance of breast cancer cells.
The members of the aldo-keto reductase (AKR) superfamily are in general cytosolic NAD(P)(H)-dependent oxidoreductases that metabolize a variety of carbonyl compounds including endogenous carbohydrates, steroids, prostaglandins (PGs) and xenobiotics (11). Among 15 known human members of this superfamily, 3 structurally similar members—AKR1C1, AKR1C2 and AKR1C3—are ubiquitously expressed in normal tissues at lower levels but highly up-regulated with carcinogenesis of several tissues including lung, prostate and breast cells (11). The enzymes are therefore considered as diagnostic markers relevant to the carcinogenesis and progression of these cells. In particular, AKR1C3 more potently promotes cellular proliferation, as manifested in culture cell-based experiments revealing an induction of breast cancer MCF7 cell proliferation by ectopic AKR1C3 overexpression and promotion of leukaemia (HL60 and K562) cell differentiation by knockdown of the enzyme (12). Although the detailed proliferation mechanisms by AKR1C3 remain unknown, the well-accepted hypothesis is its activity forming PGF2α isomers (PGF2α and 9α,11β-PGF2α) and steroid hormones (testosterone and 17β-estradiol), which facilitate the cellular proliferation through binding to their intracellular receptors and resultant activation of the downstream pathways dependent on mitogen-activated protein kinase and phosphoinositide 3-kinase (13). Previously, we also found that AKR1C3 efficiently reduces two isoprenyl aldehydes (farnesyl and geranylgeranyl aldehydes) to their corresponding alcohols and proposed that the enzyme is involved in protein prenylation and activation of the mitogen-activated protein kinase pathway, leading to stimulation of cell proliferation (14). Thus, AKR1C3 is suggested to facilitate cell proliferation through pathways dependent on metabolisms of the steroids, PGs, and isoprenoids. In addition, there is growing evidence indicating that AKR1C3 is a new candidate involved in chemoresistance development. For instance, cell-based investigation using cervical cancer HeLa cells and oral squamous cell carcinoma Sa3 cells revealed an enhancement of the CDDP sensitivity by incubating with AKR1C3 inhibitor (15). Chen et al. (16) also reported an up-regulation of the enzyme in ovarian cancer patients that are resistant to CDDP chemotherapy. Moreover, we previously showed an up-regulation of AKR1C3 in CDDP- (17), doxorubicin- (18) and irinotecan-resistant gastrointestinal cancer (19) and docetaxel (DTX)-resistant prostate cancer cells (20).
In this study, we have established a highly CDDP-resistant variant of human breast cancer MCF7 cells and found a significant up-regulation of AKR1C3 as a key factor for developing the chemoresistance of breast cancer cells. Because the chemoresistance development suppresses the formation of cytotoxic lipid peroxidation products such as 4-hydroxy-2-nonenal (HNE) and acrolein, we have also investigated alterations in glutathione level and activities of antioxidant enzymes and proteasome, which are involved in metabolism of the above products, in the resistant cells. Furthermore, we have evaluated the efficacy of combined treatment with inhibitors of AKR1C3, glutathione synthesis and proteasomal proteolysis for overcoming chemoresistance in the resistant cells.
Materials and Methods
Materials
CDDP and DTX were purchased from FUJIFILM Wako Pure Chemical Industries (Osaka, Japan) and Tokyo Chemical Industry (Tokyo, Japan), respectively. Sulforaphane (SFN) was obtained from LKT Laboratories (St. Paul, MN, USA); TRIzol reagent, Superscript III reverse transcriptase, oligo (dT)12–18 primer and Lipofectamine 2000 were from Invitrogen (Carlsbad, CA, USA); tolfenamic acid (TOL), glutathione ethyl ester (GSHEE), buthionine sulfoximine (BSO), Z-Leu-Leu-Leu-al (MG132) and ursodeoxycholic acid (UDCA) were from Sigma-Aldrich (St Louis, MO, USA); acetyl-Asp-Glu-Val-Asp-(4-methylcoumaryl-7-amide) (MCA), N-succinyl-Leu-Leu-Val-Tyr-MCA, t-butyloxycarbonyl-Leu-Arg-Arg-MCA and Z-Leu-Leu-Glu-MCA were from Peptide Institute (Osaka, Japan); acrolein was from FUJIFILM Wako Pure Chemical Industries; AcroleinRED was from Funakoshi (Tokyo, Japan); and SYBR Green Supermix was from Takara Bio (Shiga, Japan). Enhanced chemiluminescence substrate system, EnzyChrom Glutathione Peroxidase Assay Kit, Proteostat Aggresome Detection Kit and bicinchoninic acid protein assay reagent were obtained from GE-Healthcare (Buckinghamshire, UK), BioAssay Systems (Hayward, CA, USA), Enzo Life Sciences (Farmingdale, NY, USA) and Pierce (Rockford, IL, USA), respectively. HNE (21) and 3-bromo-5-phenylsalicylic acid (BPS) (22) were synthesized as described previously. All other chemicals were of the highest grade that could be obtained commercially.
Cell culture and transfection
MCF7 cells were obtained from American Type Culture Collection (Manassas, VA, USA) and grown in Dulbecco’s-modified Eagle medium supplemented with 10% fetal bovine serum, penicillin (100 U/ml) and streptomycin (100 μg/ml) at 37°C in a humidified incubator containing 5% CO2. To establish their variant resistant to CDDP, MCF7 cells were continuously cultured in the growth medium supplemented with the drugs, whose concentrations were increased in a stepwise manner. At incremental drug concentrations (0.5, 1, 2, 5, 10, 20 and 50 μM), the cells were passaged three times. The sensitivity of the cells to toxic compounds including CDDP was estimated by monitoring the viability, which was evaluated by a tetrazolium dye-based colorimetric sensitivity assay using 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium monosodium salt (23).
For the preparation of cells that overexpress AKR1C3, pGW1 expression vectors harbouring AKR1C3 cDNA were transfected into MCF7 cells using Lipofectamine 2000 when the cells were grown to 90% confluence in microplates or dishes (19). The empty vector was similarly transfected into the cells, which were used as control cells. After 48-h incubation, the cells were washed twice with a serum-free growth medium containing the antibiotics alone and then treated with the agents. To obtain the cells silenced for AKR1C3 or nuclear factor erythroid 2-related factor 2 (Nrf2), the duplex small-interfering RNAs (siRNA, 10 pmol) were transfected using Lipofectamine 2000 into the 50%-confluent cells. The overexpression and knockdown of the target gene were verified by western blot analysis described below.
Measurement of enzyme activities
Cells were washed twice with ice-cold Dulbecco’s phosphate-buffered saline (DPBS), suspended in 50-mM HEPES-NaOH, pH 7.4, containing 5-mM 3-[(3-cholamidopropyl) dimethylammonio]-propane sulfonic acid and 5-mM dithiothreitol, and homogenized by passing the cell suspension through a 26 gauge needle (20 strokes). After centrifuging the homogenate at 12,000 × g for 15 min, the supernatant was collected as the sample for measurement of caspase-3 activity. The activity of caspase-3 in the sample was measured using acetyl Asp-Glu-Val-Asp-MCA as the fluorogenic substrate. The proteasomal chymotrypsin-like, trypsin-like and caspase-like proteolytic activities in the sample were measured using succinyl-Leu-Leu-Val-Tyr-MCA, t-butyloxycarbonyl-Leu-Arg-Arg-MCA and Z-Leu-Leu-Glu-MCA, respectively, as the fluorogenic probes as described previously (24).
The reductase activities towards HNE, acrolein and glutathione were spectrophotometrically assayed by measuring NADPH oxidation at 37°C in 2.0-ml reaction mixtures, which consisted of 0.1-M potassium phosphate, pH 7.4, 0.1-mM NADPH, 10 μM substrate and cell sample (100 μg), except that 50-mM HEPES, pH 8.0 and 50 μM oxidized glutathione were utilized as the reaction buffer and substrate for measuring the glutathione reductase (GR) activity. The glutathione S-transferase (GST) activity was determined by the method of Habig and Jakoby (25), and the glutathione peroxidase (GPx) activity was measured using the EnzyChlom Glutathione Peroxidase Assay Kit according to the manufacturer’s instructions.
Real-time PCR analysis
Total RNA was isolated from cells using the TRIzol reagent, and single-stranded cDNA was prepared from the total RNA sample by incubation for 50 min at 42°C with Superscript III reverse transcriptase and oligo (dT)12–18 primer. The cDNAs for AKRs (AKR1C1, AKR1C2 and AKR1C3), glutamate-cysteine ligase (GCL) and β-actin were amplified from the single-stranded cDNA sample by real-time polymerase chain reaction (PCR) using a Takara Thermal Cycler Dice Real-Time PCR System with the SYBR Green Supermix and their specific primers listed in Supplementary Table S1. Specific amplification of the target product was verified by the subsequent analysis of melt curve profiles and DNA sequencing, and PCR efficiency for the primers was determined by amplifying serial dilutions of the template cDNA. The expression ratios of transcripts for the target proteins were calculated by normalizing to that for β-actin as the internal standard.
Western and dot blot analyses
Cells were washed twice with ice-cold DPBS, suspended in DPBS containing 0.5% Triton X-100 and 0.3-mM phenylmethanesulfonyl fluoride and subjected to sonication. The cell extract was isolated by centrifugation of the homogenate at 12,000 × g for 15 min. The nuclear and cytosolic fractions of the cells were prepared according to the method reported by Mohan et al. (26). Protein concentration was determined with the bicinchoninic acid protein assay reagent. For western blotting, proteins (40 μg) in the cell extracts were electrophoretically separated on a 7.5 or 12.5% sodium dodecyl sulphate-polyacrylamide gel under reducing conditions and then transferred to polyvinylidene difluoride membrane by electroblotting. To detect HNE-protein adducts, the cell extracts (100 μg) were bound to the membrane using a Sanplatec dot blot instrument (Osaka, Japan). After blocking with 0.5% bovine serum albumin, the membrane was allowed to react with primary polyclonal antibodies against AKR1C3 (19), Nrf2 (Santa Cruz Biotechnology, Dallas, TX, USA), histone H3 (Cell Signaling Technology, Danvers, MA, USA), β-actin (Cell Signaling Technology) and HNE (Alexis, San Diego, CA, USA). The immunoreactive proteins were visualized using a peroxidase-conjugated secondary antibody and the enhanced chemiluminescence substrate system. The densities of the bands and dots were estimated using a Bio-Rad GelDoc 2000 and attached program, Quantity One (Segrate, Italy).
Measurement of formation of acrolein and aggresome and glutathione level
Intracellular acrolein was detected using the fluorogenic probe, AcroleinRED (27). After washing twice with DPBS to remove detached cells, attached cells were incubated for 1 h in a fresh serum-free medium containing 20-μM AcroleinRED. The cells were sufficiently washed with DPBS, and the intensity of fluorescence derived from 4-formyl-1,2,3-triazoline, which results from the reaction of acrolein and AcroleinRED, was fluorometrically monitored at 585 nm with an excitation wavelength of 560 nm. Formation of aggresome was measured using the Proteostat Aggresome Detection Kit according to the manufacturer’s instructions and an LSM 700 confocal microscope (Carl Zeiss, Germany). The fluorescence intensity derived from Proteostat was fluorometrically measured at 600 nm with an excitation wavelength of 500 nm. Amounts of total glutathione and reduced glutathione (GSH) in surviving cells after treatment with the agent were measured according to the 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) method (28).
Statistical analysis
Data are expressed as means ± SD of at least three independent experiments unless otherwise noted. Statistical evaluation of the data was performed by using the unpaired Student’s t-test and ANOVA followed by Dunnett’s test. P < 0.05 was considered statistically significant.
Results
Up-regulation of AKR1C3 in CDDP-resistant breast cancer MCF7 cells
We established a CDDP-resistant variant of MCF7 cells by continuous incubation with stepwise incremental concentrations (from 0.5 to 50 μM) of the drug. The sensitivity assay revealed that 50% lethal concentration (147 μM) of the drug in the prepared CDDP-resistant cells is approximately 5-fold higher than that in the parental cells (28 μM) (Fig. 1a). In addition, treatment with high CDDP concentration (200 μM) of MCF7 cells considerably elevated the activity of caspase-3, a crucial mediator in apoptotic signal, whereas the treatment of the CDDP-resistant cells slightly activated the apoptotic mediator (Fig. 1b), apparently indicating their high resistance to CDDP toxicity. Therefore, the CDDP-resistant cells were designated as CDDP-R cells and subjected to monitoring alterations in expressions of AKR1C1, AKR1C2 and AKR1C3 with the development of the drug resistance.
Sensitivity to CDDP toxicity of MCF7 and CDDP-R cells. (a) Cellular viability. Parental MCF7 (|$\circ $|) and CDDP-R cells (●) were treated for 48 h with 0-, 10-, 20-, 50-, 100- or 200-μM CDDP. The viability value is normalized to that of the parental cells, which were treated with the vehicle dimethyl sulfoxide alone (shown as 0 μM). (b) Caspase-3 activity. The parental (P) and CDDP-R (R) cells were treated for 24 h with 0 or 200 μM CDDP, and the extracts (100 μg) were used for the measurement of caspase-3 activity. The viability and activity are expressed as the percentages of those in the parental cells treated with the vehicle (0 μM). **Significant difference from the vehicle-treated parental cells, P < 0.01. Significant difference from the parental cells treated with the same concentrations of CDDP, #P < 0.05 and ##P < 0.01.
Real-time PCR analysis revealed constitutive expressions of AKRs (1C1, 1C2 and 1C3) in MCF7 cells. Compared with the parental cells, CDDP-R cells showed significantly higher expression levels of mRNAs for AKR1C1 and AKR1C3, although no alteration in the expression level of that for AKR1C2 was detected (Fig. 2a). AKR1C3 was up-regulated more remarkably than AKR1C1. The aberrant up-regulation of AKR1C3 with the development of CDDP resistance in MCF7 cells was also confirmed by western blot analysis (Fig. 2b). The results clearly indicate an up-regulation of AKR1C3 by the gain of CDDP resistance in breast cancer cells.
Up-regulation of AKR1C3 by acquiring the CDDP resistance. (a) Real-time PCR analysis of gene expression of AKR1C1, AKR1C2 and AKR1C3 in parental MCF7 (P) and CDDP-R cells (R). (b) Western blot analysis of AKR1C3 in the cell extracts. Value normalized to β-actin is expressed as the percentage of that in the parental cells. (c) Effects of AKR inhibitors on cytotoxicity of CDDP. MCF7 cells were pretreated for 2 h without (|$\circ $|) or with 20 μM of BPS (●), UDCA (□) or TOL (■), before a 48-h treatment with CDDP. (d and e) Effects of AKR1C3 on cytotoxicity of CDDP. (d) Effect of AKR1C3 overexpression. AKR1C3-overpressing MCF7 cells (1C3, ●) and empty vector-transfected control cells (Vec, |$\circ $|) evidenced by western blotting (insets). (e) Effect of AKR1C3 knockdown. AKR1C3-knocked down CDDP-R cells (1C3 KD, ●) and scramble siRNA-transfected control cells (Scr, |$\circ $|) evidenced by western blotting (insets). The viability values after a 48-h treatment of CDDP were normalized to those in the respective control cells treated with the vehicle alone (0 μM). Significant difference from the parental cells, *P < 0.05 and **P < 0.01. Significant difference from the control cells treated with the same concentrations of CDDP, #P < 0.05 and ##P < 0.01. NS, no significant difference, P > 0.05.
When the parental cells were pretreated with sublethal concentrations of BPS (22), UDCA (29) and TOL (13), specific inhibitors of AKR1C1, AKR1C2 and AKR1C3, respectively, prior to 24-h treatment with CDDP, TOL most notably enhanced the sensitivity to CDDP toxicity (Fig. 2c), suggesting a major contribution of AKR1C3 to gain of CDDP resistance. To further unveil the relationship between the AKR1C3 up-regulation and CDDP resistance, we prepared an AKR1C3-overexpressing phenotype by transiently transfecting the expression vector harbouring the AKR1C3 cDNA into MCF7 cells. The AKR1C3-overexpressing cells exhibited a 3.6-fold higher level of AKR1C3 expression than the control cells introduced with the empty vector and strikingly protected from the cytolethal effect provoked by CDDP (Fig. 2d). In addition, transfection of the siRNA for AKR1C3 gene into CDDP-R cells resulted in a significant sensitization to CDDP (Fig. 2e), although the gene knockdown was insufficient (46% of the scramble siRNA-transfected control cells). Since viability curve of the AKR1C3-knocked down CDDP-R cells was almost consistent with that of the parental MCF7 cells (Fig. 1a), it is suggest that the AKR1C3 up-regulation is substantively connected to the development of the CDDP resistance.
It was reported that the expression of AKR1C3 is up-regulated through activating an antioxidant transcription factor Nrf2 (30). In order to clarify the participation of Nrf2 activation in the development of the CDDP resistance, Nrf2 levels in the nuclear and cytosolic fractions of the parental and CDDP-R cells were monitored by western blotting (Fig. 3a). Predictably, nuclear Nrf2 level of the CDDP-R cells was markedly higher than that of the parental cells, resulting in the significantly high ratio of nuclear/cytosolic Nrf2. Additionally, the expression level of AKR1C3 was significantly elevated by the 24-h pretreatment with an Nrf2 activator SFN (31) (Fig. 3b) and down-regulated by the Nrf2 knockdown (Fig. 3c). Furthermore, the CDDP toxicity was dramatically weakened and augmented by the SFN pretreatment (Fig. 3d) and Nrf2 knockdown (Fig. 3e), respectively, and viability curve of the Nrf2-silenced CDDP-R cells was similar to that of the parental cells (Fig. 1a). These data strongly suggest that AKR1C3 up-regulation with the CDDP resistance is attributed to the constitutive Nrf2 activation.
Constitutive activation of Nrf2 in the resistant cells. (a) Nrf2 translocation into the nucleus. Nuclear and cytosolic fractions of MCF7 (P) and CDDP-R cells (R) were subjected to western blot analysis using the anti-Nrf2 antibody. The band density was normalized to that of histone H3 (Histone) or β-actin, and the ratio of nuclear/cytosolic Nrf2 is expressed as the percentage of that in the parental cells in the bar graph. (b) Effect of SFN on the AKR1C3 expression. MCF7 cells were treated for 24 h with 0 (Vehicle) or 20 μM SFN (SFN). (c) Effect of Nrf2 knockdown on the AKR1C3 expression. CDDP-R cells were transfected with Nrf2 siRNA (Nrf2KD) or its scrambled siRNA (Scr). The cell extracts were applied to western blotting and value normalized to β-actin is expressed as the percentage of that in the vehicle-treated or Scr control cells. (d) Effect of SFN on the CDDP sensitivity. MCF7 cells were pretreated for 24 h without (|$\circ $|) or with 20-μM SFN (●) and then treated for 48 h with the indicated concentrations of CDDP. The viability value was expressed as the percentage of that in the control cells treated without SFN pretreatment. (e) Effect of Nrf2 knockdown on the CDDP sensitivity. Nrf2-knocked down CDDP-R cells (●) prepared in (c) were treated for 48 h with the indicated concentrations of CDDP. The viability value was expressed as the percentage of that in the control cells transfected with its scrambled siRNA (|$\circ $|). Significant difference from the parental or control cells, *P < 0.05 and **P < 0.01. Significant difference from the cells treated with the same concentrations of CDDP, #P < 0.05 and ##P < 0.01.
Association between metabolism of cytotoxic reactive aldehydes and chemoresistance development
Production of ROS is believed to be a key mechanism underlying CDDP-triggered cancer cell apoptosis. Indeed, pretreatment with an antioxidant N-acetyl-L-cysteine significantly lessened CDDP toxicity in MCF7 cells (Supplementary Fig. S1). In addition, flow-cytometric analysis using the fluorogenic ROS probe revealed a marked ROS production in MCF7 cells by CDDP treatment and the CDDP-elicited ROS production was significantly restored by acquiring the CDDP resistance (Supplementary Fig. S2), suggesting that gain of the chemoresistance elevates the antioxidant properties in the breast cancer cells. CDDP treatment of MCF7 cells notably facilitated the formation of HNE and acrolein, cytotoxic reactive aldehydes derived from lipid peroxidation (32), as apparent from the marked increases in HNE-protein adducts and intensity of AcroleinRED probe-derived fluorescence in the CDDP-treated cells (Fig. 4a and b). In CDDP-R cells overexpressing AKR1C3, the levels of the two reactive aldehydes were not affected by CDDP treatment and were highly elevated by adding the AKR1C3 inhibitor TOL. In addition, AKR1C3 knockdown in CDDP-R cells increased the CDDP-mediated formation of HNE (Fig. 4c), whose level formed is almost the same as that in the parental cells after the CDDP treatment (Fig. 4a). Since the AKR1C3 knockdown hardly affected ROS production in the resistant cells by CDDP treatment (Supplementary Fig. S3), it is suggested that AKR1C3 is involved in the detoxification of reactive aldehydes, rather than ROS. AKR1C3 is reported to reduce HNE into its less toxic alcohol 1,4-dihydroxynonene (33) and, therefore, reductase activities towards HNE and acrolein in the cells were compared (Fig. 4d and e). The HNE- and acrolein-reductase activities in CDDP-R cells were significantly higher than those in the parental cells, and similar elevations of the two activities were also observed in AKR1C3-overexpressing MCF7 cells. Furthermore, CDDP-R cells were much more resistant to the toxicities of HNE and acrolein than the parental cells (Fig. 4f). Thus, detoxification of reactive aldehydes by up-regulated AKR1C3 may play a crucial role in development of the CDDP resistance.
Reduction of HNE and acrolein by AKR1C3. (a) Formation of HNE-protein adducts by CDDP treatment. MCF7 and CDDP-R cells were pretreated for 2 h without or with 20-μM TOL and then treated for 24 h with 200 μM CDDP. The cell extracts (100 μg) were subjected to the dot blot analysis using an anti-HNE antibody (upper panel). (b) Formation of acrolein. The cells treated as described in (a) were incubated for 1 h with AcroleinRED and then the fluorescence intensity in the cell extracts (100 μg) was measured. (c) Effect of AKR1C3 knockdown on the HNE-protein adduct formation. CDDP-R cells were transfected with AKR1C3 siRNA (1C3KD) or its scrambled siRNA (Scr) and then treated for 24 h with 200-μM CDDP prior to performing the dot blot analysis. (d and e) HNE-reductase (HNE-R) and acrolein-reductase (A-R) activities in extracts (100 μg) of MCF7 (P), CDDP-R (R), empty vector-transfected MCF7 (Vector) and AKR1C3-overexpressing MCF7 cells (1C3). The activity in the cell extract is expressed as the percentage of that in the parental or vector-transfected control cells. (f) Sensitivity to HNE and acrolein. The parental (|$\circ $|) and resistant cells (●) were treated for 48 h with the indicated concentrations of HNE or acrolein. Values are expressed as the percentages of those in the cells treated with the vehicle alone. Significant difference from the parental cells, *P < 0.05 and **P < 0.01. Significant difference from the parental or control cells treated with same concentrations of the agents, #P < 0.05 and ##P < 0.01. $Significant difference from the resistant cells treated with the same concentration of CDDP, P < 0.05.
It is well accepted that one of the major metabolic pathways of HNE is conjugation with GSH catalysed by the action of GST (34). Considering that the amount of HNE formed by CDDP treatment was low in the resistant cells (Fig. 4a), it is assumed that HNE formed is subjected to the GSH conjugation as well as AKR1C3-mediated conversion into its reduced metabolite. To test this assumption, we first measured the GSH amount in the parental and CDDP-resistant cells by the DTNB method (Fig. 5a). The amounts of GSH and total glutathione in CDDP-R cells were significantly higher than those in the parental cells. Real-time PCR analysis also revealed a considerable up-regulation of GCL, a rate-limiting enzyme in de novo synthesis of GSH (35), with gain of the CDDP resistance (Fig. 5b). Moreover, cytotoxicity assay showed that the CDDP damage is augmented and reduced by pre-incubation with a GCL inhibitor (BSO) (36) and membrane-permeable analog of GSH (GSHEE), respectively (Fig. 5c and d). These results clearly suggest that the supply of GSH is accelerated through facilitating its de novo synthesis due to GCL up-regulation and attributable to gain of the CDDP resistance. The CDDP resistance had less influence on the activities of GST and GR, an enzyme that NADPH-dependently regenerates from oxidized glutathione to GSH, whereas it significantly elevated the activity of GPx (Fig. 5e). The AKR1C3 overexpression had little effect on the GSH amount and GPx activity (Fig. 5f and g), inferring that synthesis and metabolism of GSH are promoted through mechanism(s) independent of AKR1C3.
Significance of glutathione in development of the CDDP resistance. (a) Glutathione levels. Levels of total glutathione (Total) and GSH in extracts (100 μg) of MCF7 (P) and CDDP-R cells (R) were measured by the DTNB method. (b) Expression level of mRNA for GCL analysed by real-time PCR. The level in MCF7 cells is taken as 100%. (c and d) Effect of glutathione on the sensitivity to CDDP toxicity. MCF7 cells were pretreated for 2 h with 1-mM GSHEE (c) or 10 μM BSO (d) and then treated for 48 h with the indicated concentrations of CDDP. The viability values are expressed as the percentage of those in the vehicle-treated control cells (0 μM). (e) Activities of glutathione-related enzymes (GPx, GR and GST) in extracts (100 μg) of MCF7 and CDDP-R cells. (f and g) Effect of AKR1C3 on the glutathione levels and GPx activity. Levels of total glutathione and GSH (f) and GPx activity (g) in extracts (100 μg) of vector-transfected MCF7 (Vector) and AKR1C3-overexpressing MCF7 cells (1C3) were measured. Values are expressed as the percentages of those in the parental or vector-transfected control cells. Significant difference from the control cells, *P < 0.05 and **P < 0.01. #Significant difference from the cells treated with the same concentrations of CDDP, #P < 0.05 and ##P < 0.01. NS, no significant difference, P > 0.05.
Since HNE-modified proteins are known to be degraded to a considerable extent by the ubiquitin-proteasome system (37), we investigated whether proteolysis in 26S proteasome is related to the gain of CDDP resistance or not. Proteasome functional assays revealed that chymotrypsin-like, trypsin-like and caspase-like proteolytic activities of CDDP-R cells were significantly higher than those of the parental cells (Fig. 6a). The treatment of parental MCF7 cells with CDDP facilitated the formation of aggresome, insoluble aggregates composed of excessive polyubiquitinated and/or misfolded proteins (38), whereas the treatment did not influence the aggresome formation in CDDP-R cells (Fig. 6b). Additionally, pretreating with a proteasome inhibitor MG132 significantly strengthened the CDDP toxicity to MCF7 cells (Fig. 6c). The ectopic AKR1C3 overexpression did not alter the chymotrypsin-like proteolytic activity (Fig. 6d). These results suggest that the elevation of proteasome function is induced in an AKR1C3-independent manner and contributes to the CDDP resistance through degradation of the HNE-modified proteins formed by CDDP treatment.
Elevation of proteasomal proteolytic activities by the CDDP resistance. (a) Proteolytic activities of 26S proteasomes. The extracts (100 μg) of MCF7 and CDDP-R cells were used for fluorophotometric measurements of chymotrypsin-like, trypsin-like and caspase-like activities. (b) Formation of aggresome. The parental (P) and CDDP-R cells (R) were treated for 24 h with 200 μM CDDP and stained with Proteostat. The intensity of the Proteostat-derived fluorescence is expressed as the percentage of that in the parental control cells. (c) Effect of AKR1C3 on proteasomal proteolytic activity. The chymotrypsin-like activity in extracts (100 μg) of vector-transfected MCF7 (Vector) and AKR1C3-overexpressing MCF7 cells (1C3) were measured. The activities are expressed as the percentages of those in the parental or vector-transfected control cells. (d) Effect of proteasome inhibitor on the CDDP sensitivity. MCF7 cells were pretreated for 2 h with 0.2 μM MG132 (●), or the vehicle (|$\circ $|), and then treated for 48 h with the indicated concentrations of CDDP. The viability value is expressed as the percentage of that in the vehicle-treated cells (|$\circ $|, 0 μM). **Significant difference from the parental cells, P < 0.01. Significant difference from the cells treated with the same concentrations of CDDP (|$\circ $|), #P < 0.05 and ##P < 0.01. NS, no significant difference, P > 0.05.
Combined effect of inhibitors of AKR1C3, GSH synthesis and proteasomal proteolysis on CDDP resistance
Pretreatment of CDDP-R cells with any of the inhibitors (TOL, BSO or MG132) at the sublethal concentrations in part significantly elevated sensitivity to CDDP (Fig. 7a). In addition, the combined treatment of TOL, BSO and MG132 greatly sensitized to the CDDP toxicity against the resistant cells. Moreover, a similar sensitizing effect to another anti-cancer drug DTX was observed in the pretreatment of CDDP-R cells with the combination of the two inhibitors (Fig. 7b). These results clearly suggest that the combined treatment with inhibitors of AKR1C3, GSH synthesis and proteasomal proteolysis exerts a potent overcoming effect against CDDP resistance and DTX cross-resistance in the breast cancer cells.
Combined treatment with inhibitors of AKR1C3, GSH and proteasome overcomes CDDP resistance and cross-resistance in MCF7 cells. MCF7 and CDDP-R cells were pretreated for 2 h with 20-μM TOL, 10-μM BSO and/or 0.2-μM MG132 and then treated for 48 h with 200-μM CDDP (a) or 100 μM DTX (b). The viability value is expressed as the percentage to that of the parental cells treated with the vehicle alone. **Significant difference from the parental cells treated with CDDP or DTX alone, P < 0.01. Significant difference from the CDDP-R cells treated with CDDP or DTX alone, #P < 0.05 and ##P < 0.01.
Discussion
Breast cancer is one of the most prevalent types of gynecologic cancer. Similar to the chemotherapy of other malignant cancers, a platinum-based drug CDDP is also utilized as the first-line treatment of breast cancer (1). However, it is well known that effectiveness of the drug monotherapy for breast cancer is in general low (39). Previous clinical trials have therefore proposed various combination therapies, such as CDDP plus DTX (40), CDDP plus doxorubicin (41) and CDDP plus gemcitabine (42), while little is known about a potential target for adjuvant chemotherapy for alleviating the CDDP resistance. Previously, it is reported that AKR1C3 is over-expressed in invasive ductal carcinoma of the breast and might promote the proliferation through above-mentioned metabolisms of PGs and estrogens (43). Zhong et al. (44) also showed that AKR1C3 is associated with doxorubicin resistance in breast cancer via loss of PTEN. In this study, we found a marked up-regulation of AKR1C3 in CDDP-resistant variant of breast cancer MCF7 cells and accordingly propose AKR1C3 as the potential target that alleviates and overcomes development of CDDP resistance in breast cancer cells. The AKR1C3 up-regulation has been also observed in gastrointestinal cancer cells resistant to CDDP (17), doxorubicin (18) and irinotecan (19), as well as in DTX-resistant prostate cancer cells (20). A key regulator involved in cytoprotective responses against anticancer drugs is Nrf2, whose constitutive activation of Nrf2 has been detected in different types of human malignancies, potentiating an advantage for the growth and chemoresistance of cancer cells (45). In human breast cancers, overexpression and activation of Nrf2 are reported to contribute to tumour progression and chemoresistance (45, 46). As evidenced by the western blot results of Nrf2 (Fig. 3a), the CDDP resistance caused the nuclear accumulation of Nrf2, indicating its constitutive activation by the chemoresistance. Polymorphism analysis of clinical specimens found some mutations in Kelch-like ECH-associated protein 1 (Keap1), an Nrf2 inhibitor, in breast carcinoma (47). Another recent clinical investigation has also shown a positive correlation between platinum-based chemotherapy against lung cancers and formation of the somatic mutations of Nrf2, Keap1 and Cullin 3 (48). Therefore, the up-regulation of AKR1C3, the Nrf2-driven gene, may be attributed to their somatic mutations caused by continuous exposure to CDDP.
Under normal conditions, ROS act as signalling molecules in a variety of biological processes. However, excessive ROS production exceeding cellular antioxidant properties results in an enhanced formation of a highly reactive hydroxyl radical, which consequently forms cytotoxic reactive aldehydes such as HNE and acrolein by means of peroxidation of membrane lipids. HNE is well known to interact directly with various proteins and consequently to disrupt major cellular processes such as proliferation, finally leading to triggering cellular apoptosis (32, 34). Based on the results obtained in the AKR1C3-silencing studies (Fig. 4c and Supplementary Fig. S3), we here surmise that AKR1C3 up-regulated detoxifies reactive aldehydes (HNE and acrolein), rather than ROS, leading to gain of the CDDP resistance. In addition, we propose that HNE formed during the CDDP treatment is also metabolized by both GSH conjugation and proteasomal degradation of its protein adducts. Since Nrf2-target genes include GCL, GST, GR, ROS-metabolizing enzymes (GPx, thioredoxin reductase 1 and NAD(P)H: quinone oxidoreductase 1) (49), as well as some members of the AKR family (50), up-regulation of these antioxidant enzymes may also be ascribable to the constitute Nrf2 activation by gain of the CDDP resistance. Although there were no significant elevations of GST and GR activities with development of the CDDP resistance (Fig. 5e), this contradiction might imply that other transcription factors are involved in the de novo synthesis of the two glutathione-related enzymes. Our previous report showed that induced types of proteasome catalytic subunits are up-regulated through mechanisms dependent on nitric oxide and guanylate cyclase (51). Therefore, the production of reactive nitrogen species might be also promoted by resistance development. It should be noted that GPx is significantly activated by acquiring the CDDP resistance (Fig. 5e). Because GPx catalyses GSH-dependent reduction of hydrogen peroxide and fatty acid hydroperoxide into water and fatty acid alcohol, respectively (52), it is surmised that ROS production and lipid peroxidation caused during CDDP treatment are excluded in the chemoresistant cells. Eight isozymes of GPx have been identified in humans to date and are reported to differ from each other in tissue expression and substrate specificity (53). In addition, some GPx isozymes appear to be involved in cancer progression and tumorigenesis, as well as antioxidant activities. Further studies are therefore needed to identify the isozymes involved in CDDP resistance of breast cancer cells and to clarify whether and how the GPx isozymes are activated with the chemoresistance development.
Here, our study provides evidence that incubation with AKR1C3 inhibitors significantly elevates sensitivity of CDDP-R cells to CDDP. The similar overcoming effects of AKR1C3 inhibitors against drug resistance were observed in gastric and colon cancer cells as previously reported (17–19). In addition, we show that combination with more than two inhibitors of AKR1C3, GSH synthesis and proteasomal proteolysis strongly prevails the CDDP resistance. Moreover, the validity of the combination therapy was confirmed in the DTX cross-resistance of breast cancer cells. These results suggest the efficacy of the combined treatment as an adjuvant therapy to prevent the resistance to the two anti-cancer drugs in breast cancer patients. To evaluate the availability and safety of the combination therapy, further investigations using normal and other breast cancer cell lines and animal xenograft models are ongoing in our laboratories.
Supplementary Data
Supplementary Data are available at JB Online.
Author contributions
T.M., A.I., S.E. and K.I. designed the experiments. T.M., M.K, A.Y., H.T. and Y.N. conducted the research and analysed the data. All authors contributed to the preparation and discussion of the manuscript.
Acknowledgements
We are indebted to Dr. Akira Hara for his insightful comments.
Funding
This work was supported in part by grants-in-aid for scientific research (C) (17K08278 and 20K07033) from the Japan Society for the Promotion of Science.
Conflict of Interest
The authors declare that there are no conflicts of interest.
