Pancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer-related mortalities in the Western world and is responsible for more than 45,000 deaths per year in the US alone [1
]. Less than 9% of PDAC patients survive for five years or more after diagnosis [2
]. The conventional treatment approaches, such as chemotherapy, radiation therapy, surgery, and any combination of these, have had little impact on the course of this aggressive malignancy [3
]. Therefore, new therapeutic strategies based on the unique molecular biology and physiology of pancreatic cancer are needed [7
The constant need for cellular building blocks drives the overzealous metabolism in cancer cells [10
]. As a result, abundant byproducts such as reactive oxygen species (ROS) and reactive nitrogen species persistently accumulate and dysregulate cellular homeostasis, causing DNA damage and inducing senescence [11
]. To maintain optimal redox balance in the cells, efficient and counteractive antioxidant machinery is required. Glutathione (GSH), nicotinamide adenine dinucleotide phosphate (NADPH), and redox regulatory proteins such as thioredoxin reductase and thioredoxin constitute the antioxidant enzyme system and scavenge the high levels of ROS [13
Glutathione S-transferase pi-1 (GSTP1) is a principal component of the antioxidant system [14
]. It plays a cytoprotective role by catalyzing the conjugation reaction of reduced glutathione (GSH) to reactive electrophiles generated by cytochrome P450 metabolism [15
]. GSTP1 is ubiquitously expressed in mammalian tissues and is overexpressed in human tumors of diverse anatomic origin [16
], as well as in a wide variety of drug-resistant cell lines [18
]. In addition to its role in cellular detoxification and glutathionylation, GSTP1 regulates stress-induced signaling by binding to and inhibiting the phosphorylation of c-Jun N-terminal kinase (JNK) [19
]. Additionally, GSTP1 was recently shown to modulate glycolytic metabolism in breast cancer cells by enhancing glyceraldehyde-3-phosphate dehydrogenase (GAPDH) activity [20
]. These, and additional literature [15
], suggest that GSTP1 plays versatile roles in cancer cell survival, signaling mechanisms, and metabolism. With its established roles in breast [20
] and cervical cancer [24
], we postulate that overexpression of GSTP1 provides selective advantages to PDAC cells by scavenging elevated ROS and maintaining cellular homeostasis. In this present study, we provide evidence suggesting that GSTP1 contributes to pancreatic cancer cell growth and holds promise as a therapeutic target for PDAC.
Our data provide convincing evidence that GSTP1 plays a critical role in regulating PDAC cell growth, which was previously unknown. In this study, we show that the GSTP1 knockdown impairs the growth and proliferation of PDAC cells in vitro. We show, for the first time, that GSTP1 inhibition is associated with enhanced JNK activity and suppressed ERK, NF-κB, and Sp1 activity in PDAC cells. Furthermore, in an orthotopic pancreatic cancer mouse model, GSTP1 knockdown tumors showed an impressive reduction in growth compared to control tumors. Together, our results indicate that GSTP1 inhibition impairs PDAC cell growth, suggesting that GSTP1 is a viable target for PDAC therapy.
The ubiquitous expression of GSTP1 in a wide array of tissues and organisms provides evidence that GSTP1 has important cellular roles. We found GSTP1 protein expression was at least two times higher in five PDAC cell lines compared to normal pancreatic epithelial cells. Similarly, GSTP1 mRNA was reported in high levels in human PDAC tissue compared to the healthy pancreas tissue. Our analysis of The Human Protein Atlas [26
] data revealed that elevated GSTP1 expression is associated with poor survival of PDAC patients, post-diagnosis of the disease. Previous research has shown that GSTP1 is expressed at high levels in a variety of human cancers, including colon, lung, breast, and ovarian cancers [25
]. GSTP1 overexpression is associated with resistance to chemotherapeutic drugs like cisplatin, carboplatin, adriamycin, and bleomycin in ovarian and cervical cancer [33
GSTP1 is associated with a variety of cellular processes, including detoxification [15
], glutathionylation [21
], actin polymerization [22
], nitric oxide signaling [23
], kinase signaling [31
], and cellular metabolism [20
]. To investigate the role of GSTP1 in PDAC cells, we generated two GSTP1 knockdown lines for each of the three metabolically diverse PDAC cell lines (MIA PaCa-2, PANC-1, and HPAF-II). MIA PaCa-2 and PANC-1 cells are poorly differentiated mesenchymal-type PDAC cell lines compared to HPAF-II cells that belong to an epithelial subtype [27
]. GSTP1 knockdown significantly impaired the in vitro viability of all three PDAC cell lines, suggesting that this protein is vital to PDAC growth regardless of metabolic subtype. Our cell viability data are in concordance with previous reports where Louie et al. [20
] described that GSTP1 knockdown impairs the growth of triple-negative breast cancer cells. They concluded by demonstrating that GSTP1 inhibition disrupts glycolytic metabolism, resulting in reduced levels of lipids, nucleotides, and ATP. Furthermore, recently, Fujitani et al. [34
] showed that knocking down GSTP1 in cancer cells of various anatomic origins gives rise to mitochondrial stress and severely impairs cell proliferation. Interestingly, they found that pancreatic cancer cell growth was particularly sensitive to GSTP1 knockdown.
Attempts have been made to disrupt the cellular redox balance through pharmacological regulation in favor of increasing intracellular ROS and inducing apoptosis for the treatment of cancer. Arrick et al. [35
] showed that specifically inhibiting the synthesis of GSH contributed to the destruction of neoplastic cells in vitro. Inhibiting GSTP1, an integral component of the cellular antioxidant system, is one avenue to disrupt redox balance. GSTP1 protects cells from electrophiles that cause oxidative damage to DNA, proteins, and lipids by conjugating electrophiles to GSH [18
]. Here, we observed that the knockdown of GSTP1 in PDAC cells resulted in elevated ROS levels. Furthermore, the addition of GSH to GSTP1 knockdown cells enhanced cell viability and reduced the expression of stress and apoptosis-associated proteins. A previous study showed that an antioxidant (N-acetylcysteine) could reduce ROS levels in GSTP1 knockdown PDAC cells [34
]. We speculate that GSTP1 knockdown impairs the ROS scavenging function that leads to ROS accumulation in PDAC cells. Our observations complement a previous report that showed GSTP1 inhibition using siRNAs and a pharmacological inhibitor elevated ROS levels and caused DNA damage in prostate cancer cells [29
]. Moreover, GSH also restored cell viability, reduced ROS, and decreased apoptosis-associated protein expression in PDAC cells treated with a GSTP1 inhibitor [37
Elevated oxidative stress activates the JNK signaling pathway and triggers apoptosis [31
]. In a non-stressed environment, GSTP1 binds and inhibits the phosphorylation of JNK preventing the transcriptional activation of downstream cell stress pathways. However, under oxidative stress conditions, GSTP1 dimerizes into aggregates and its binding to JNK is deterred, enabling JNK activation [38
]. Previously, we showed that the interaction between JNK and GSTP1 is interrupted in PDAC cells treated with a GSTP1 inhibitor [39
]. Additionally, complementing our current results, it was shown that a JNK inhibitor could restore viability of PDAC cells treated with a GSTP1 inhibitor [39
]. As expected, GSTP1 knockdown increased the expression of phosphorylated JNK and its downstream target, c-Jun, in PDAC cells. This increase could be due to elevated levels of ROS that could activate JNK signaling and/or reduced levels of GSTP1 that would also result in enhanced JNK signaling. Our data are supported by a previous report that suggested GSTP1 knockdown elevated phosphorylated JNK expression in cervical cancer cells [24
]. The extent and duration of JNK activation can lead to ER stress, mitotic arrest, and eventually apoptosis in cancer cells [31
To elucidate additional mechanisms through which GSTP1 knockdown impedes growth and the proliferation of PDAC cells, we investigated the activation status of ERK, NF-κB, and Sp1 pathways. GSTP1 knockdown cells displayed reduced phospho-ERK and NF-κB, and reduced Sp1 protein expression. In support of this, ERK and NF-κB protein expression were reduced in cervical cancer cells upon GSTP1 inhibition [24
]. Sp transcription factors are upregulated in various cancer cells [40
] and act as negative-prognostic markers for patient survival [41
]. Our data are supported by previous reports that suggest ROS induction by chemotherapy and other anti-cancer agents lead to the downregulation of Sp proteins [42
] and the reduced phosphorylation of ERK1/2 [46
]. Similar to the previous studies [40
], we show that the restoration of Sp1 expression can be achieved by supplementing the cells with an exogenous antioxidant such as glutathione. Additionally, Sp (1, 3, or 4) knockdown induced similar cellular responses, such as enhanced cell death, and gene expression changes (increased apoptosis promoters and decreased apoptosis inhibitors) as we observed in GSTP1 knockdown PDAC cells [48
]. Furthermore, we show the reduced expression of principal cell-cycle regulators, cyclin D1 [49
] and CDK4 [50
]. Based on these results, we propose a mechanism through which GSTP1 alters MAP kinases and NF-κB signaling, averts apoptosis, and promotes cell survival and proliferation (Figure 5
D). We speculate that in the absence of GSTP1, JNK is freely phosphorylated as a result of activating the downstream cell death pathways. Moreover, elevated ROS levels reduce the expression of Sp1 that transcribes Bcl2
] and the p65 subunit of NF-κB [52
]. Reduced levels of Bcl2
and p65 in GSTP1 knockdown PDAC cells contribute to the apoptotic phenotype and decreased cell survival, respectively.
Intriguingly, we found that the orthotopic implantation of GSTP1 knockdown cells in the pancreata of athymic nude mice resulted in drastically smaller tumors compared to scrambled controls in terms of both tumor weight and volume. We also observed a lower percentage of proliferating cells and a larger population of apoptotic cells in tumors generated from GSTP1 knockdown PDAC cells. These results support our in vitro data as well as previously published literature. shRNAs targeting GSTP1 were shown to reduce breast cancer xenograft implants by more than three-fold [20
]. Similar results were observed when GSTP1 was inhibited using specific morpholinos in cervical cancer [24
Given GSTP1’s cytoprotective roles in xenobiotic detoxification, chemotherapeutics, and modulating oxidative stress, GSTP1 inhibitors emerged as promising anti-cancer compounds [55
] and have been used alone or in combination with chemotherapeutic drugs [57
]. The selective targeting of GSTP1 using 6-(7-nitro-2,1,3-benzoxadiazol-4-ylthio)hexanol (NBDHEX) has shown increased efficiency of chemotherapeutic drugs in melanoma [56
]. A potent GSTP1 inhibitor, TLK199 (Telik Inc.), has been shown to modulate cell proliferation in human myeloid leukemic cells [58
] and is under clinical trial for myelodysplastic syndrome [59
]. LAS17 was recently developed as a highly potent and selective GSTP1 inhibitor that impairs breast cancer pathogenicity [20
]. The aforementioned GSTP1 inhibitors have shown effective impairment in GSTP1 activity; however, their toxicity in normal cells is not well characterized.
Collectively, our findings illustrate the crucial role of GSTP1 in the growth of PDAC cells. The loss of GSTP1 function leads to the activation of oxidative-stress response pathways that trigger a cell death mechanism. Taken together, our data suggest that GSTP1 is a potential and promising novel therapeutic target to treat pancreatic cancer patients.
4. Materials and Methods
Puromycin was purchased from Sigma-Aldrich, St. Louis, MO, USA. A CellTiter-Glo® luminescent cell viability assay kit was purchased from Promega, Madison, WI, USA. Ki67 antibody was purchased from Vector Labs, Burlingame, CA, USA. GSTP1 antibody was obtained from Santa Cruz Biotechnology, Dallas, TX, USA. Antibodies to GAPDH, β-actin, phospho-JNK (Thr 183/Tyr 185), total JNK, p65, pERK, total ERK, cleaved caspase-3, total caspase-3, phospho-c-Jun (Ser 73), total c-Jun, and Sp1 were obtained from Cell Signaling Technology, Danvers, MA, USA. Horseradish peroxidase (HRP)-linked anti-mouse and anti-rabbit IgG secondary antibodies were obtained from Cell Signaling Technology. CF633-conjugated goat anti-mouse IgG secondary antibody was obtained from Biotium, Fremont, CA, USA. CM-H2DCFDA was purchased from Life Technologies, Carlsbad, CA, USA. Glutathione was purchased from Calbiochem, Burlington, MA, USA.
4.2. Cell Culture
Human PDAC cell lines (MIA PaCa-2, PANC-1, HPAF-II, AsPC-1, and Bx PC-3) were obtained from American Type Culture Collection, Manassas, VA, USA. hTERT-HPNE cells were obtained from Dr. Channing Der’s laboratory at UNC, Chapel Hill, NC. MIA PaCa-2 cells were cultured in DMEM (Dulbecco’s Modified Eagle Medium) high-glucose media (GE Healthcare Life Sciences, Chicago, IL, USA) containing 10% (v/v) fetal bovine serum (Atlanta Biologicals, Flowery Branch, GA, USA) and 2.5% (v/v) horse serum (Corning, Corning, NY, USA). PANC-1 cells were cultured in DMEM high-glucose media containing 10% (v/v) FBS. HPAF-II cells were cultured in Eagle’s Minimum Essential Medium (EMEM) (Corning, Corning, NY, USA) containing 10% v/v Fetal Bovine Serum (FBS). AsPC-1 were cultured in RPMI-1640 (GE Healthcare Life Sciences) containing 10% FBS (v/v). Cells were maintained at 37 °C with 5% CO2. The cell lines were subcultured by enzymatic digestion with 0.25% trypsin/1 mM EDTA solution (GE Healthcare Life Sciences, Chicago, IL, USA) when they were 80% confluent. All cell lines tested negative for Mycoplasma contamination.
4.3. Constructing Knockdown Cell Lines
We used two independent short-hairpin oligonucleotides to knock down the expression of GSTP1. Lentiviral particles containing the shRNA (Sigma, St. Louis, MO, USA, catalogue# SHCLNV-NM_000852) were used to infect the target PDAC cell lines with polybrene (Sigma-Aldrich, St. Louis, MO, USA). Transfected cells were selected over five days with 5 μg/mL puromycin. The short-hairpin sequences used to achieve the knockdown of GSTP1 expression were: shGSTP1-1, CCGGCCTCACCCTGTACCAGTCCAACTCGAGTTGGACTGGTACAGGGTGAGGTTTTG; shGSTP1-2, CCGGC GCTGACTACAACCTGCTGGACTCGAGTCCAGCAGGTTGTAGTCAGCGTTTTTG. Scrambled GSTP1 shRNA, empty vector (pLKO.1), and shRNA targeting GFP were used as controls. Knockdown was confirmed by qRT-PCR and Western blotting techniques.
4.4. Western Blotting
Cells and tissues were lysed in lysis buffer (Promega, Madison, WI, USA) containing both protease and phosphatase inhibitors. Denatured proteins were resolved on 11% SDS-polyacrylamide gel and transferred to nitrocellulose membrane (AmershamTM ProtranTM 0.2 μM, GE Healthcare Life Sciences, Chicago, IL, USA) using the wet electroblotting system (BioRad, Hercules, CA, USA). Blots were blocked using 5% BSA in Tris-buffered saline containing Tween 20 (TBS-T) solution for 1 h at room temperature, washed in TBS-T, and probed with primary antibody overnight at 4 °C. Following washes with TBS-T, the blots were incubated with HRP-linked secondary antibody at room temperature for 1 h. Blots were treated with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific, Waltham, MA, USA) and visualized using FluorChem FC2 imaging system. The expression levels were quantified using ImageJ software. The data represent average ± standard deviation for three independent experiments.
4.5. RNA Extraction and Gene Expression by qRT-PCR
Total RNA was extracted using SurePrep TrueTotal RNA purification kit (Carlsbad, CA, USA) following the manufacturer’s instructions. cDNA was synthesized using 500 ng of total RNA and the qScript cDNA synthesis kit (Quanta Biosciences, Beverly, MA, USA). Steady-state RNA levels were determined as described elsewhere [60
]. The relative change in gene expression was calculated using the 2−ΔΔCt method [61
]. HPRT, β-actin, and β-tubulin were used as internal controls. The data represent the average ± standard deviation for three independent experiments with two technical replicates each. The primer sequences of the genes analyzed are listed in Table 1
4.6. Cell Viability Assay
MIA PaCa-2, PANC-1, and HPAF-II cells (3000/well) were seeded in 96-well plates. The viability of control and GSTP1 knockdown cells after 24, 48, 72, and 96 h was evaluated by adding 100 μL of CellTiter-Glo® substrate to each well containing 100 μL of media. The plates were incubated for ten min at room temperature. The endpoint luminescence was measured using Synergy H1 Hybrid multi-mode plate reader (Winooski, VT, USA) located in the Core Biology Facility, Chemistry and Molecular Biology, North Dakota State University. The gain was maintained at 135 and the integration time of 1 s using the Gen5 v2.07 software. The data represent the average ± standard deviation of three independent experiments with eight technical replicates for each treatment.
4.7. Cell Cycle Arrest Assay
Control and GSTP1 knockdown PDAC cell lines (MIA PaCa-2, PANC-1, and HPAF-II) were seeded in 6-well plates and incubated for 24 h. Cells were synchronized overnight using serum-free medium and harvested by trypsinization, washed, and re-suspended in 70% ethanol overnight at 4 °C. Finally, 70% ethanol was removed, and cells were re-suspended in PBS containing 50 μg/mL propidium iodide (VWR Life Technologies) and 1 μg/mL RNase A (Biotium, Fremont, CA, USA). Flow cytometry was performed using BD Accuri C6 equipment to determine the cell population in each phase of the cell cycle. The data represent the average ± standard deviation of three independent experiments with three technical replicates for each treatment.
4.8. Detection of ROS Levels by the 2,7-Dichlorodihydrofluorescein Diacetate (CM-H2DCFDA) Assay
Control and GSTP1 knockdown MIA PaCa-2 and HPAF-II cells were resuspended in 20 μM CM-H2DCFDA (Life Technologies) in PBS and incubated at 37 °C for 30 min before flow cytometric analysis using a BD Accuri C6. Three technical replicates were included for each experiment, and the experiments were performed in biological triplicates for each cell line. FLOWJO software was used to create histograms. The data represent the average ± standard deviation of three independent experiments with three technical replicates for each treatment.
4.9. Orthotopic Tumor Studies
All animal experimental procedures were performed abiding by the protocol approved by North Dakota State University’s Institutional Animal Care and Use Committee (IACUC). Six- to eight-week-old female athymic nude mice (nu/nu) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). The mice were maintained in sterile conditions using individually ventilated cage (IVC) racks (Allentown and Innovive). The mice were acclimated for 1 week before tumor implantation. PDAC cells were washed twice with PBS, trypsinized, and harvested in serum-containing medium. Harvested cells were washed with serum-free medium and resuspended in PBS. Mice were anesthetized using 3% isoflurane. A small incision was made in the left abdominal flank and control or GSTP1 knockdown cells (7.5 × 105
in 25 μL) were injected into the pancreas using a 27-gauge needle. The abdomen was closed using chromic catgut and ethilon sutures by a 2-layer suture technique. Animals were monitored every day for their food and water intake and for the signs of distress and pain. The tumor volumes were estimated every ten days by abdominal ultrasounds. The mice in the HPAF-II experimental group were euthanized earlier than the previously planned endpoint, as the tumor volumes in the control group were approaching the highest acceptable values as defined in the IACUC protocol. Humane endpoints defined for removing animals from the project were: (1) if/when the tumor burden was estimated to be more than 10–15% of their body weight, if mice demonstrated significant signs of distress or pain, (2) if the tumor interfered with ambulation, if mice exhibited decreased eating or drinking, or (3) if they showed signs of infection [62
]. After 4 weeks (HPAF-II group) or 6 weeks (MIA PaCa-2 and PANC-1), animals were euthanized using a CO2
chamber (Quietek Model 1, Next Advantage, Troy, NY) that regulates the flow of CO2
in the chamber at a rate of 10–30% of the chamber volume per minute. The equipment will not exceed 30% of the chamber volume per minute. These flow rates are compliant with the AVMA regulations for euthanasia of laboratory mice. Animal death was subsequently verified by cervical dislocation. The primary tumor in the pancreata was excised and measured for weight. Each tumor was paraformaldehyde fixed and paraffin embedded for immunohistochemistry. The data represent the average ± standard deviation for the biological replicates.
Ethics approval and consent to participate: All the animal experimental procedures were approved by North Dakota State University’s Institutional Animal Care and Use Committee (protocol number: A17062). The permitted study period on the protocol was from May-2017 to April-2020. North Dakota State University maintains a registration with the United States Department of Agriculture (45-R-002) and an Animal Welfare Assurance with the National Institute of Health-Office of Laboratory Animal Welfare (D16-00156).
4.10. Murine Abdominal Ultrasound Imaging
The growth of pancreatic tumors was monitored via abdominal ultrasound imaging every ten days for all animals in the treatment groups (for the HPAF-II group, last ultrasound was performed on D27). A FUJIFILM Vevo3100 ultrasound imaging system (Toronto, ON, Canada) was used to image the pancreata. The animals were anesthetized using 3% isoflurane and were maintained at 2% isoflurane for the course of ultrasound. To support the optimal physiological conditions, mice were kept on the platform maintained at 37 °C. Intraperitoneal administration of 2 mL saline was performed to achieve a higher resolution of abdominal organs. Mice were retained in the supine position and the tumor volumes were calculated using an Mx250 transducer and Vevo Lab Software. The data represent the average ± standard deviation for the biological replicates.
Tumor tissues were collected and fixed for 24 h in formaldehyde. Paraffin-embedded 5-μm-thick sections of tumor tissues were prepared. Sections were deparaffinized with Histo-Clear and ethanol, followed by antigen retrieval in 10 mM sodium citrate buffer (0.05% Tween 20, pH 6.0) using an autoclave method. The sections were blocked for 20 min in blocking buffer (10% normal goat serum in TBS-T) and incubated with Ki67 (1:100) or cleaved caspase-3 (1:100) overnight at 4 °C. The following day, sections were incubated with CF633-conjugated goat anti-mouse secondary antibody (1:250) for 1 h at room temperature. After mounting a coverslip using Hardset Mounting Medium with 4′,6-diamidino-2-phenylindole (DAPI) (Vector Labs, Burlingame, CA, USA), slides were visualized using a Zeiss inverted Axio Observer Z1 microscope. The percentage of Ki67- or cleaved caspase-3-positive cells was measured based on the number of pink-stained cells relative to the number of blue DAPI-stained nuclei. Immunohistochemistry was performed for all the tumor samples from different treatment groups. The data represent the average ± standard deviation for the biological replicates.
4.12. Statistical Analyses
All outcome variables were analyzed using fixed-effects linear models with analysis of variance. For relative GSTP1 expression in different human PDAC cell lines, different cell lines and experimental replicate were the factors. Cell viability was analyzed separately for each PDAC cell line with knockdown line, time, and experimental replicate as the factors. The live and the dead cell population in the scrambled controls of three different cell lines were compared to the same populations in the GSTP1 knockdown cells. The G0/G1 and G2/M populations of scrambled controls were compared to the same populations in GSTP1 knockdown cells. Relative protein expressions of p-JNK, p-ERK, and p-p65 in GSTP1 knockdown cells were analyzed with protein, knockdown line, and experimental replicate as the factors. The relative expression of phosphorylated proteins was compared to total proteins. Relative cleaved caspase-3 expressions were analyzed separately for PDAC cell lines with knockdown lines and experimental replicates as the factors. Pancreas volume was analyzed separately for each PDAC cell line, with knockdown line and day as the factors. Only the results for last ultrasound are presented. The relative tumor weight was analyzed separately for each PDAC cell line using fixed-effects models with knockdown lines as the factor. Welch’s one-way analysis of variance was performed due to the observed heterogeneity of variances. Relative tumor weight, Ki-67-positive cell population, and in vivo cleaved caspase-3 cell population were analyzed separately for each PDAC cell lines using fixed-effects models with knockdown lines as the factor. The Pearson correlation test was done to analyze the association between GSTP1 expression and the survival of patients, post-diagnosis of PDAC.
For any analysis in which an interaction effect was not significant, the interaction effect was dropped from the model for the final analysis. Post-hoc pairwise comparisons using Tukey were performed following significant findings in the overall analysis of variance. All analyses were performed using the MIXED procedure in SAS software version 9.4 (SAS Institute; Cary, NC, USA).