Aspirin Regulates Gemcitabine Resistance in Pancreatic Cancer via Inhibiting the PI3K/Akt/mTOR Signaling Pathway and Reversing Epithelial-mesenchymal Transition

Background: Gemcitabine is considered a classical agent for the treatment of patients with pancreatic cancer. However, gemcitabine resistance is a common cause of treatment failure, leading to poor survival. Therapy to overcome gemcitabine resistance would benet patients with pancreatic cancer. This study investigated the impact of aspirin (ASA) to gemcitabine resistance in the biological function of pancreatic cancer cells and the potential mechanism. Methods: The MTT assay, wound healing assay and annexin V-FITC/PI test was used to determine whether ASA could inhibit gemcitabine resistance in pancreatic cancer cells. Besides, the expression of Bcl-2, Bax, E-cadherin, Vimentin, p-PI3K, p-AKT and p-mTOR was detected by the Western blot. Statistically signicant differences between groups were determined with the Student’s t-test. Results: The proliferation and migration of SW1990 and BxPC3 cells were signicantly decreased while the apoptosis rate was increased in the combination of ASA and gemcitabine group (p<0.05). The level of Bcl-2 was decreased and the level of Bax was increased signicantly in the combination of ASA and gemcitabine group (p<0.05) while the level of p-PI3K, p-AKT and p-mTOR was decreased (p<0.05). What’s more, this group signicantly reversed EMT in SW1990 and BxPC3 cells with an increase in the expression of E-cadherin and a decrease in Vimentin. Conclusion: Our research provided evidence ASA could help to inhibit gemcitabine resistance of pancreatic cancer cells, probably via inhibiting the PI3K/AKT/mTOR pathway and reversing EMT. Thus, combined use of ASA and gemcitabine is expected to be a potential therapeutic strategy for pancreatic cancer patients. of the PI3K/AKT/mTOR pathway on SW1990 and BxPC3 cells, as indicated by effectively reducing the phosphorylation of PI3K, Akt and mTOR. These data suggested that Synergistic effect of ASA plus gemcitabine therapy in pancreatic cancer may occured by reversing EMT via downregulation of the PI3K/AKT/mTOR signaling pathway. These results are similar to those seen in a previous study, in which ASA was found to inhibit the EMT and migration of oncogenic K-ras-expressing non-small cell lung carcinoma cells by down-regulating the E-cadherin repressor Slug [48]. a promising new therapeutic carcinoma


Introduction
Pancreatic cancer is the eighth leading cause of cancer deaths worldwide with early metastasis and a very poor prognosis. The overall 5-year survival rate is less than 9% [1]. At diagnosis, most patients present with local regional spread and/or distant metastasi [2]. At present, gemcitabine is recognized by many oncologists as the rst-line agent for the treatment of pancreatic cancer. However, the monotherapy with gemcitabine tends to develop acquired resistance in pancreatic cancer, which becomes a major concern in clinical practice [3][4][5] , and the underlying mechanism of chemo-resistance is poorly understood.
The PI3K/AKT/mTOR signaling pathway has been proved important in tumor development and is often activated in pancreatic cancer (PC) [6]. AKT has been proved to contribution to cancer development, which is activated by PI3K, growth factors and so on. Downstream effectors like mTOR lead to signal transduction [7]. PI3K is a downstream effector of oncogenic KRAS, which is nearly ubiquitous in pancreatic cancer [8]. A recent study which analyzed 32 cancer types in The Cancer Genome Atlas (TCGA) has identi ed KRAS exert a strong pro-tumorigenic effect through PI3K/AKT/mTOR pathway activation in pancreatic cancer [9]. A recent study shows mTOR activation is connected with gemcitabine resistance in pancreatic cancer [10].
The highly aggressive feature of pancreatic cancer may be partly attributed to the chemotherapy-resistant characteristics of pancreatic cancer cells which is related to the epithelial-mesenchymal transition (EMT) phenotype and cancer stem cells [11]. The EMT is an orchestrated series of events leading to altered cellcell and cell-extra cellular matrix interactions that transform the cancer tissue with a broblast-like morphology, and confer migratory and invasive properties on neoplastic cells [12,13]. In essence, EMT represents the transition between two fully differentiated and mature cells [14]. In tumor cells, the expression of phenotype marker proteins of epithelial cells (e.g. E-cadherin) is decreased, and the expression of phenotype marker proteins of mesenchymal cells (e.g. Vimentin and N-cadherin) is increased. Homocellular tight junction and cell polarity are also de cient [15,16]. Importantly, some research indicated the EMT may alter the sensitivity of neoplastic cells to chemotherapy agents [17,18]. Therefore, target therapy against or minimizing EMT may increase the e cacy of chemotherapy.
As we all know, several kinds of chemotherapeutic agents are now widely used in the rst line including 5FU, irinotecan, oxaliplatin and taxanes. When it comes to the reason why we speci cally choose gemcitabine for our study, here are our answer. According to National Comprehensive Cancer Network (NCCN) Guidelines Version 1.2019 Pancreatic Adenocarcinoma, gemcitabine, is contained in rst-line therapy or preferred regimens of neoadjuvant therapy (resectable/borderline resectable disease), adjuvant therapy, locally advanced disease and metastatic disease. In a word, gemcitabine is the standard rst-line chemotherapeutic drug of pancreatic cancer. In the rst line, FOLFIRINOX which including 5FU, irinotecan, oxaliplatin and taxanes is also signi cant. However, FOLFIRINOX is not recommended in poor performance status. Moreover, almost all the patients came across several side effects when treated with FOLFIRINOX, which placed restrictions on adding aspirin. As we know, adding aspirin in a therapy may increase risk of bleeding which is rather serious and could lead to death in cancer patients while anemia is common. That's part of the reasons why we didn't combine aspirin with FOLFIRINOX which including 5FU, irinotecan, oxaliplatin and taxanes. Additionally, it was reported aspirin inhibits proliferation of pancreatic cancer cells via inhibiting PI3K/Akt/mTOR signaling pathway. Moreover, aspirin also shows a therapeutic potential through targeted inhibition on tumor growth and angiogenesis [19,20]. Many studies pointed out gemcitabine activation of Akt/mTOR signaling pathway in resistant cells is a key signaling event for gemcitabine resistance, and we suppose aspirin may resensitize resistant cells to gemcitabine by inhibiting this activated pathway [21][22][23]. Considering the above two points, when it comes to combining with aspirin, gemcitabine may be more suitable for research than 5FU, irinotecan, oxaliplatin, taxanes or other chemotherapeutic agents.
ASA, a weak organic acid, show promise as cancer chemoprevention agents due to their antiin ammatory properties [24]. A pooled analysis of 25,570 patients in eight trials recently reported that daily aspirin use reduced deaths of several common cancers, including pancreatic cancer deaths, with most bene t seen after 5 years of the scheduled treatment [25], while a clinic-based case-control study showed that aspirin use is associated with lowered risk of developing pancreatic cancer [26]. Several years ago, Streicher SA, et al. conducted a study in Connecticut and found approximately 50% reduced risk of pancreatic cancer with regular use of either low-dose or regular-dose aspirin. In China, a population-based study of 761 case and 794 control subjects was performed to show ever-regular use of aspirin was associated with lowered risk of pancreatic cancer: odds ratio [OR] = 0.54; 95% CI: 0.40-0.73. However, the evidence that aspirin use may lower risk of pancreatic cancer is still full of con icts around the world and the role of ASA in the gemcitabine resistance of pancreatic cancer remain ambiguous, as well as the molecular mechanism, if ASA really bene ts.

Materials And Methods
Cell lines and cell culture Human pancreatic cancer cell lines SW1990 and BxPC3 were obtained from the Shanghai Cell Bank of the Chinese Academy of Sciences and the Changhai hospital a liated to the Second Military Medical University. Mycoplasma testing has been done for the cell lines used. Cells were cultured in 1640 medium (Gibco, Gaithersburg, MD, USA) containing 10% fetal bovine serum (FBS), and maintained in a humid atmosphere at 37°C with 5% CO 2 . The cell medium was changed every 2 to 3 days. Cell passage was performed with trypsin when the con uence of monolayer cells reached 70% to 80%.

Cell proliferation assay
Cells were cultured in 1640 medium containing 10% FBS, and maintained in a humid atmosphere at 37°C with 5% CO 2 for 72 h. Then cells were harvested and inoculated into 96-well plates with 3×10 3 cells in each well 24 h before treatment. After that, samples were divided into 4 groups. Cells were cultured in: 1)1640 medium containing 10% FBS and 2 mmol/L ASA, 2)1640 medium containing 10% FBS and 0.05 mg/L gemcitabine, 3)1640 medium containing 10% FBS, 2 mmol/L ASA and 0.05 mg/L gemcitabine, 4) negative control (1640 medium containing 10% FBS). 24 hours later, we incubated cells at 37°C with 5% CO 2 for next 0 hours, 24 hours, 48 hours and 72 hours. Then, 10% MTT (5 mg/mL) was added into each well. After adding MTT for 4 hours, total of 150 μl DMSO was then added into each well, and the samples were vibrated for 10 minutes. The absorbance (OD) was then measured with a microplate reader set at 570 nm. Each cell proliferation experiment was repeated 3 times and the readings were averaged for the statistical analysis.

Wound Healing Assay
BxPC3 and SW1990 cells were seeded in 6-well plates. When the cell con uence was about 100%, scratches were made on the monolayer cell surface with 200-μl pipette tips. After the cells were washed twice with PBS, the cells were divided into 4 groups, which was mentioned in the cell proliferation assay. Images of cells on both sides of the scratches were captured with an inverted microscope at 0 and 2 hours. Finally, images have analyzed the migration in three elds of view that were randomly selected and analyzed in an independent trial with 400X magni cation under an inverted microscope (Olympus Corp). The width (W) of the scratch measured; the percentage of wound area remaining was calculated as W 24h /W 0h × 100%. All experiments were performed in triplicate Cell apoptosis BxPC3 and SW1990 cells were seeded in 6-well plates and cultured with 1640 medium containing 10% FBS. After the cells were attached to the plate for 24 h, they were divided into 4 groups, which was mentioned in the cell proliferation assay, and cultured for 24 h. The cells were then collected, washed with phosphate buffer saline (PBS) solution 2 to 3 times. To count the number of apoptotic cells, they were stained with annexin V-uorescein isothiocyanate (FITC)/propidium iodide (PI). These cells were analyzed by using the BD FACSCalibur in each experiment. All experiments were carried out in triplicate.

Western blot Analysis
The 4 groups cells, which was mentioned in the cell proliferation assay, were harvested after cultured for 24 h and lysed with RIPA lysis buffer. Protein molecules released from the cells were separated and extracted by the SDS-PAGE (Sodium dodecyl sulfate-Polyacrylamide gel electrophoresis), and then transferred to polyvinylidene uoride (PVDF) membrane for western blotting. Membranes were blocked with 5% non-fat milk and then incubated with primary antibodies against β-actin, Bax, Bcl-2, E-cadherin, Vimentin, PI3K, phospho-PI3K, Akt, phospho-Akt, mTOR and phospho-mTOR at the dilution of 1:1000 at 4°C overnight. The membranes were then rewarmed for 30 minutes, washed with TBST (Tris Buffered saline Tween), and were incubated with the corresponding horse radish peroxidase-labeled secondary antibody at the dilution of 1:5000 at 37°C for 1 hour. Western blots were visualized using Immobilon Western Chemiluminescent HRP Substrate (Merck Millipore, Burlington, MA, USA) followed by lm exposure. β-actin was used as the internal reference. The gray level was analyzed, and the results were calculated as the gray level of the target protein divided by the gray level of the internal reference.

Statistical analysis
Statistical analysis was performed with Graphpad Prism software 7.00. Continuous data were calculated as mean ± standard deviation (SD) and statistical signi cance was de ned as p<0.05. Statistically signi cant differences among experimental groups were determined with student' t-test.

Results
1. Combination of ASA (2 mmol/L) and Gemcitabine (0.05 mg/L) Achieved More Inhibition in Cell Proliferation Compared with ASA or Gemcitabine Alone.
As shown in Figure 1, the survival rate of SW1990 has more rapid reduction after treated with combination of ASA (2 mmol/L) and gemcitabine (0.05 mg/L) than with gemcitabine or ASA alone (p<0.05). Similar results were observed with BxPC3 cells, showing the combination of ASA and gemcitabine inhibit achieved more inhibition in cell proliferation compared with ASA or gemcitabine alone. Collectively, the ASA regulates gemcitabine resistance in SW1990 and BxPC3 cells. The proportion of apoptosis of the two pancreatic cancer cell lines SW1990 and BxPC3 was signi cantly higher when treated with the combination of ASA and gemcitabine than with ASA or gemcitabine alone(p<0.05) (Fig. 3A). In control cells, the percentages of apoptosis were 4.13±0.88% and 4.79±0.65%, respectively. In cells treated with ASA alone, these values were 7.84±2.35% and 6.12±0.23%, respectively.
In cells treated with gemcitabine alone, these values were 24.68±1.53% and 7.31±0.83%, respectively. In cells treated with the combination of ASA and gemcitabine, these values were 32.13±1.95% and 11.32±0.98%, respectively. Furthermore, the protein level of Bcl-2 was signi cantly more dramatically reduced when Bax was increased in those cells treated with the combination of ASA and gemcitabine than with either ASA or gemcitabine alone (p<0.05) (Fig. 3B). Consistent with the aforementioned experimental results, the combination of ASA and gemcitabine contributes to apoptosis of pancreatic cancer cells more than ASA or gemcitabine alone. These evidences suggested that the ASA may promote antitumor effect of gemcitabine in SW1990 and BxPC3 cells. 4. The Combination of ASA (2 mmol/L) and Gemcitabine (0.05 mg/L) Increased Signi cantly the Expression of Ecadehrin and Decreased the Expression of Vimentin.
Expression of EMT biomarkers in pancreatic cancer cells in vitro was signi cantly increased for Ecadehrin but decreased for Vimentin in the combination of ASA and gemcitabine group than in ASA or gemcitabine alone group (p<0.05), indicating cell apoptosis and growth inhibition induced by the combination treatment may be closely related to EMT (Figure 4). 5. The Combination of ASA (2 mmol/L) and Gemcitabine (0.05 mg/L) Downregulated the PI3K/Akt/mTOR Signaling Pathway.
Cells presented a signi cant decrease in the expression of p-PI3K and p-AKT after treated with combination of ASA (2 mmol/L) and gemcitabine (0.05 mg/L) for 24 h, compared with those observed in control and ASA or gemcitabine alone. Concurrently, the level of p-mTOR protein of the combination treatment group was depressed, measured through western blotting. These ndings suggested that the downregulation of the PI3K/Akt/mTOR pathway may be partly involved in cell apoptosis and growth inhibition induced by the combination treatment ( Figure 5).

Discussion
ASA, a widely used non-steroidal anti-in ammatory agent, has been reported to inhibit pancreatic cancer development by inhibition of mTOR signaling pathway, supporting the results we observed [27]. Regular use of ASA has been previously observed to be related to decreasing the risk of various types of cancer, such as Hepatocellular Carcinoma and breast cancer [28,29]. Recent research found that ASA inhibited GSK-3β activation on Capan-1and PANC-1 when combining with gemcitabine [30]. Another recent study demonstrated that the combination of ASA and metformin signi cantly inhibited pancreatic cancer cell growth in vitro and in vivo by regulating the pro-and anti-apoptotic Bcl-2 gene family members [31]. There are studies showed that low-dose ASA has a moderate chemopreventive effect on adenomas in the large bowel. Daily use of ASA is associated with a signi cant reduction in the incidence of colorectal adenomas in patients with previous colorectal cancer [32,33]. Recent research has also demonstrated that ASA can reduce the risk of cancer initiation and progression and can be used to suppress several tumor properties, including tumor cell migration [34,35]. However, no detailed study has been published, to the best of our knowledge, to assess the effect of ASA on gemcitabine monotherapy-related chemotherapy in pancreatic cancer cells. In our study, which concentration of ASA could be most suitable was a issue we discussed a lot. On the basis of the study of Shengping Lin, compared with untreated cells, PANC-1 cells treated with 2 mM aspirin had a proliferation inhibition rate of about 40% at 24 h (p<0.05) [36]. We chose the 2 mmol/L concentration of ASA as a new attempt. However, compared with some reports, 2 mmol/L concentration seems to be a bit higher. We speculated it might be that the concentration of aspirin in pancreatic cancer is inconclusive and the concentration of aspirin in other tumors may be different from that in pancreatic cancer. What's more, overexpressed extracellular matrix (ECM) in pancreatic cancer limits drug penetration into the tumor and is associated with concentration of ASA in pancreatic cancer microenvironment. However, researchers have been looking for different ways. It's reported collagenase nanoparticles could enhance the penetration of drugs into pancreatic cancer. Research declared that a pretreatment based on a proteolytic-enzyme nanoparticle system disassembles the dense pancreatic cancer collagen stroma and increases drug penetration into the pancreatic cancer [37]. Moreover, to address the problem of drug penetration, some researchers demonstrated a dendrimer-camptothecin (CPT) conjugate that actively penetrates deep into pancreatic cancer through γglutamyl transpeptidase (GGT)-triggered cell endocytosis and transcytosis. This dendrimer-drug conjugate even exhibited high antitumor activity in multiple mice tumor models compared to the standard rst-line chemotherapeutic drug (gemcitabine) for advanced pancreatic cancer [38]. To sum up, increase the concentration of drugs in pancreatic cancer microenvironment requires the progress of materials science, biology, pharmacology and other elds.
The role of ASA therapy in the tumorigenesis and development of pancreatic cancer and the underlying mechanism remain unclear. The goal of this study was to validate whether ASA could help to inhibit gemcitabine resistance so that more patients are able to respond to gemcitabine. In addition, we looked for the possible molecules involved in gemcitabine resistance.
Previous research has addressed the role of EMT in tumor cell drug resistance [39]. Therefore, the mechanisms of EMT in pancreatic cancer needs research for the development of new therapeutic strategies [40]. Our study showed that ASA could increase the expression level of E-cadherin and decrease in the expression level of Vimentin in SW1990 and BxPC3, indicating ASA could inhibit gemcitabine resistance of pancreatic cancer cells through transformed EMT.
The PI3K/Akt family is among the most frequently mutated pathways in human cancer. Aberrant activation of the PI3K/Akt signaling pathway is thought to be associated with development of chemotherapeutic resistance. Multiple genetic events have been described that lead to activation of the PI3K/Akt/mTOR pathway in cancer [41]. Akt/mTOR pathways inhibitor like rapamycin and LY294002 could inhibit the EMT progression in pancreatic cancer. Study shows that ROS is concidered to regulate EMT [42]. ROS balance cellular integrity and cell death while several anticancer drugs have been shown to induce high levels of ROS and decrease of AKT/mTOR signaling pathway [43][44][45]. Our study showed that ASA could inhibit gemcitabine resistance of pancreatic cancer cells through downregulated PI3K/Akt/mTOR (e.g. p-PI3K, p-AKT and p-mTOR), which are consistent with previous reports demonstrating that gemcitabine inhibited pancreatic cancer growth and metastasis and that gemcitabine inhibited EMT with the involvement of PI3K/Akt/mTOR pathway in pancreatic cancer cells [46]. We also have found that the combination of ASA and gemcitabine exerts a remarkable inducing effect on cell proliferation and markedly promotes apoptosis in both SW1990 and BxPC3 cell lines with the involvement of PI3K/Akt/mTOR signaling pathway. Between the 2 cell lines used in our study, the difference in response rates deserves discussion. Throughout a large number of literature basis, the difference in response rates between the BxPC3 and SW1990 cell line is existent. Some researchers pointed out when treated with gemcitabine, SW1990 cell line's IC50 is 6.266mmol/L while BxPC3 cell line's is 4.01mmol/L [47]. What's clear is that, when treated with gemcitabine, the BxPC3 cell line may has stronger drug resistance in some malignant biological behaviors.
What we found is that low-dose ASA combined with gemcitabine signi cantly reduced cell proliferation, inhibited cell migration, increased the apoptosis, increased expression of E-cadherin and suppressed that of Vimentin, when compared with gemcitabine alone. When ASA plus gemcitabine therapy compared with ASA or gemcitabine alone, the result showed signi cant downregulation of the PI3K/AKT/mTOR pathway on SW1990 and BxPC3 cells, as indicated by effectively reducing the phosphorylation of PI3K, Akt and mTOR. These data suggested that Synergistic effect of ASA plus gemcitabine therapy in pancreatic cancer may occured by reversing EMT via downregulation of the PI3K/AKT/mTOR signaling pathway. These results are similar to those seen in a previous study, in which ASA was found to inhibit the EMT and migration of oncogenic K-ras-expressing non-small cell lung carcinoma cells by downregulating the E-cadherin repressor Slug [48].
In summary, our results demonstrated that ASA could help to inhibit gemcitabine resistance, which induced more inhibition in cell proliferation, reduction in migration and increase in apoptosis. Our work also revealed that ASA could regulate gemcitabine resistance by inhibiting the PI3K/Akt/mTOR signaling pathway to reverse EMT in pancreatic cancer cells, where we provided new insight into. The results provide a promising new therapeutic option for pancreatic carcinoma patients.

Declarations
Ethics approval and consent to participate Not applicable. Combination of ASA (2 mmol/L) and gemcitabine (0.05 mg/L) inhibited pancreatic cell migration. The movement ability of SW1990 and BxPC3 cells after treated with ASA and/or gemcitabine for 24 h was detected using wound healing assays. We used electronic microscope (X400) to observe. The width(W) of the scratch measured; the percentage of wound area remaining was calculated as W24h/W0h × 100%. All experiments were performed in triplicate. The results represent mean ± standard deviation, * P<0.05. Effect of the combination of ASA (2 mmol/L) and gemcitabine (0.05 mg/L) on apoptosis of pancreatic cancer cells. ASA plus gemcitabine increased pancreatic cancer cell apoptosis. Pancreatic cancer cell lines were treated with gemcitabine and/or ASA for 48 hours. (A) Annexin V-FITC/PI staining assay was used to detect cell apoptosis. The quantitative analysis of the apoptotic cell numbers is shown. (B)Western blot assay was used to evaluate the expression of apoptosis-related proteins including Bcl-2 and Bax. The data shown represent the mean ± SD from 3 independent experiments with similar results. The bars represent mean ± standard deviation, * P<0.05.

Figure 4
The combination of ASA (2 mmol/L) and gemcitabine (0.05 mg/L) increased signi cantly the expression of E-cadehrin and decreased the expression of Vimentin. SW1990 and BxPC3 cells were treated with ASA and/or gemcitabine for 24 h and the levels of EMT-related proteins(E-cadherin, Vimentin) were determined through western blotting while β-actin was included as a loading control. All experiments were performed in triplicate, * P<0.05.

Figure 5
The combination of ASA (2 mmol/L) and gemcitabine (0.05 mg/L) downregulated the PI3K/Akt/mTOR signaling pathway. SW1990 and BxPC3 cells were treated with ASA and/or gemcitabine for 24 h and the levels of PI3K, AKT, mTOR, phosphorylated PI3K, phosphorylated AKT and phosphorylated mTOR were determined through western blotting while β-actin was included as a loading control. All experiments were performed in triplicate, *P<0.05.