Introduction

The five-year survival rate for pancreatic cancer is less than 10%, making it one of the deadliest cancers1. It accounts for over 90% of pancreatic cancers, primarily pancreatic ductal adenocarcinoma2. Surgical resection, chemotherapy, and radiotherapy are the current treatments for pancreatic cancer3. Medications such as morphine are generally used to manage severe pain associated with advanced pancreatic cancer4. It appears that morphine may significantly influence the growth of tumors in cancer patients.

Morphine is used to treat severe pain, both acutely and chronically. The analgesic and sedative effects of morphine are primarily produced by activating the Mu opioid receptor5. However, morphine’s effect on cancer malignancy has been contradictory in studies. Some studies have shown that cancer cells are inhibited by morphine. For example, Zhang et al. found that using clinical concentrations of morphine (10µM in vitro experiments), tumorigenicity of hepatocellular cancer cells was inhibited6. Chen et al. demonstrated that apoptosis and blocking the cell cycle are two mechanisms by which morphine (10µM in vitro experiments) inhibits the progression of breast cancer7. On the other hand, however, morphine also promotes cancer progression. A higher level of survivin increases the aggressiveness of renal cell carcinoma caused by morphine (50µM in vitro experiments)8. Nguyen et al. found that cancer progression was stimulated and survival was impaired in transgenic mice with breast cancer by morphine (subcutaneously injected at a concentration of 1.5 mg kg−1 day−1in vivo experiments)9. Esophageal cancer cell migration and growth are stimulated by morphine (3 mg kg−1 day−1 orally) in a preclinical model10. Most scientists believe that these differences in outcomes may be due to differences in dose or route of administration11,12.

In advanced pancreatic cancer patients, morphine is commonly prescribed. However, it is unclear how morphine impacts the malignancy of pancreatic cancer and its underlying mechanisms. Consequently, it is imperative to investigate morphine’s effects on pancreatic cancer progression and its possible mechanisms of action.

Results

Morphine has minimal impact on HPDE cell growth, but it exhibits a bidirectional effect on pancreatic cancer cell growth

We selected four pancreatic cancer cell lines (BxPC-3, PANC-1, SUIT-2, and CFPAC-1) as well as the normal HPDE cell line for the CCK-8 assay. We found that different concentrations of morphine barely affected HPDE cell growth (Fig. 1a). The proliferation of pancreatic cancer cells was accelerated at a lower concentration of morphine (25 µM). However, treatment with a high concentration of morphine (100 µM) showed a significant time-dependent inhibition of proliferation (Fig. 1b–e). We selected two cell lines, BxPC-3 and PANC-1, which were the most sensitive to morphine. We chose the concentration (25 µM) at which proliferation was most pronounced and the concentration (100 µM) at which apoptosis was most pronounced for subsequent experiments. In the colony formation assay, the 25 µM group had significantly more colonies, and the 100 µM group had fewer colonies than the control (Fig. 1f,g). We also performed the EdU assay, and in both cell lines, the EdU positivity rate was significantly higher in the 25 µM group and lower in the 100 µM group than in the control group (Fig. 1h,i).

Fig. 1
figure 1

Morphine has a bidirectional effect on the growth of pancreatic cancer cells. (a) Different concentrations of morphine had little impact on HPDE cell growth. (b–e) In BxPC-3, PANC-1, SUIT-2, and CFPAC-1 cells, a morphine concentration of 25 µM significantly increased cell activity, while a concentration of 100 µM significantly reduced cell activity in a time-dependent manner. (f,g) In the colony formation assay, the 25 µM group showed a significantly higher number of colonies, whereas the 100 µM group had fewer colonies compared to the control in two cell lines. (h,i) The EdU positive rate in the 25 µM group was significantly higher, and in the 100 µM group, it was significantly lower in two cell lines compared to the control. The figure compared all 25 µM groups to group 0 and all 100 µM groups to group 0. We analyzed three independent experiments, and all data were expressed as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

Low concentrations of morphine (25 µM) promote migration and invasion of pancreatic cancer cells, while high concentrations of morphine (100 µM) inhibit migration and invasion

Additionally, we examined how morphine affected the migration and invasion of pancreatic cancer cells. In the wound healing assay, the scratch healing rate of BxPC-3 and PANC-1 cells in the 25 µM group was significantly higher, while the 100 µM group was significantly lower than the control group (Fig. 2a–d). In the transwell migration and invasion test, the number of migrated and invaded cells in the 25 µM group of the two cell lines was significantly higher, whereas the 100 µM group was significantly lower than the control group (Fig. 2e–j).

Fig. 2
figure 2

Morphine has a bidirectional effect on the invasion and migration of pancreatic cancer cells. (a–d) In the wound-healing assay, the healing rate of the scratches in the 25 µM group of both BxPC-3 and PANC-1 strains was significantly higher, while the healing rate of the scratches in the 100 µM group was noticeably lower compared to the control group. (e,f) In the transwell migration/invasion test, the number of migrated/invaded cells in the 25 µM group of both cell lines was significantly higher than that in the control group, whereas the 100 µM group showed a significant decrease compared to the control group. (g) Quantitative analysis of BxPC-3 migrating cell number in (e). (h) Quantitative analysis of BxPC-3 invasive cell number in (e). (i) Quantitative analysis of PANC-1 migrating cell number in (f). (j) Quantitative analysis of PANC-1 invasive cell number in (f). The figure compared all 25 µM groups to group 0 and all 100 µM groups to group 0. We analyzed three independent experiments, and all data were expressed as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

In vivo experiments have confirmed that morphine has a bidirectional effect on pancreatic cancer

To explore whether the bidirectional effect of morphine on pancreatic cancer progression exists in vivo, we conducted a tumor xenograft experiment. On day 31, during the sampling, we observed that the tumor volume in the low-dose morphine group (M 0.5 mg/kg) was significantly larger, while the tumor volume in the high-dose morphine group (M 5 mg/kg) was noticeably smaller compared to the control group (Fig. 3a,b). In addition, we weighed the tumors in each group and found that the tumor weight was significantly heavier in the M (0.5 mg/kg) group and obviously lighter in the M (5 mg/kg) group than in the control group (Fig. 3c). The body weights of the three groups of mice did not differ (Fig. 3d). To verify whether morphine alone affects body weight in mice, we performed the same intervention in normal mice with the above two morphine concentrations for the same period of 3 weeks. We found that there was no significant difference in body weight between the mice treated with two morphine concentrations and the control group, indicating that the two doses of morphine intervention may not have significant toxic effects on mice (Fig. 3e).

Fig. 3
figure 3

In vivo experiments confirmed that morphine exerts bidirectional effects on pancreatic cancer. a The effects of morphine (M) at doses of 0.5 mg/kg and 5 mg/kg on tumorigenesis were determined in a xenograft experiment. bIn vivo tumor growth curves were measured (n = 6). c The weights of tumor tissue were measured in the groups treated with M at doses of 0.5 mg kg−1, 5 mg kg−1, and the control group (n = 6). d There were no statistically significant differences in body weight between the groups of nude mice (morphine intervention in xenografted mice). e There were no statistically significant differences in body weight between the groups of nude mice (morphine intervention in normal mice). NS was administered based on the body weight of the control group. The number of mice used in each group was six. The data were expressed as mean ± SD, with *P < 0.05, **P < 0.01, and ***P < 0.001.

Full size image

Bidirectional effects of morphine on pancreatic cancer progression via the p38/JNK pathway

The results of the WB assay are shown in Fig. 4a. The relative expression of p-JNK/JNK was higher in the 25 µM group and lower in the 100 µM group compared to the control group (Fig. 4b). Additionally, the relative expression of p-p38/p38 was lower in the 25 µM group and higher in the 100 µM group compared to the control group (Fig. 4c). The expression statistics of Bcl-xL, Bcl-2, Bax, E-cad and Vim are shown in Fig. 4d–h. For comparison, we also treated HPDE cells with morphine at concentrations of 25 µM and 100 µM for 72 h, and measured the protein expression levels of Bcl-xL, Bcl-2, Bax, E-cad, and Vim. Our findings indicated that, relative to the control group, neither the 25 µM nor the 100 µM morphine concentration had an impact on the expression levels of these proteins, as shown in Fig. S1a, b. Furthermore, BxPC-3 cells were treated with two different concentrations of morphine for 72 h, after which the expression levels of HIF1α, p-mTOR, mTOR, p-Src, Src, and MOR (Mu opioid receptor) were assessed. Our analysis demonstrated that, relative to the control group, neither the 25 µM nor the 100 µM morphine concentration impacted the expression ratios of HIF1α/GAPDH, p-mTOR/mTOR, or p-Src/Src, as depicted in Fig. S2a–e. These findings suggest that the previously reported MOR/Src/mTOR signaling pathway13 and the HIF-1α/p38MAPK pathway14 may not significantly contribute to the effects of morphine on pancreatic cancer progression.

Fig. 4
figure 4

Bidirectional effects of morphine on pancreatic cancer progression via the p38/JNK pathway were investigated. (a) Western blot analysis was conducted to measure the levels of p-JNK, JNK, p-p38, p38, Bcl-xL, Bcl-2, Bax, E-cad, and Vim after treating BxPC-3 cells with morphine concentrations of 25 µM and 100 µM for 72 h. (b) The quantitative analysis of p-JNK/JNK in (a). (c) The quantitative analysis of p-p38/p38 in (a). (d) The quantitative analysis of Bcl-xL/GAPDH in (a). (e) The quantitative analysis of Bcl-2/GAPDH in (a). (f) The quantitative analysis of Bax/GAPDH in (a). (g) The quantitative analysis of E-cad/GAPDH in (a). (h) The quantitative analysis of Vim/GAPDH in (a). Three independent experiments were performed, and all data were expressed as mean ± SD. Statistical significance was indicated as follows: *P < 0.05, **P < 0.01, and ***P < 0.001.

Full size image

The effect of morphine on the progression of pancreatic cancer was reversed by p38 inhibitors and agonists

BxPC-3 cells were treated with the p38 inhibitor SB203580 and the p38 agonist anisomycin, respectively, and changes were detected by WB assay (Fig. 5a–c). Then we confirmed that SP600125 had a significant inhibitory effect on JNK phosphorylation by WB experiment (Fig. S3a,b). In the absence of morphine, SB203580 was found to promote the proliferation of pancreatic cancer cells, while anisomycin and the JNK inhibitor SP600125 inhibited pancreatic cancer cell proliferation when compared to controls. The combined treatment of SB203580 and SP600125 was able to counteract the proliferative effect on cancer cells that were treated with SB203580 alone (Fig. 5d–f). Anisomycin reversed the pro-growth effect of 25 µM morphine, and SB203580 reversed the inhibitory effect of 100 µM morphine on pancreatic cancer cells (Fig. 5g–i). In the absence of morphine, SB203580 was found to promote the migration and invasion of pancreatic cancer cells, while anisomycin and SP600125 inhibited the migration and invasion of pancreatic cancer cells when compared to controls. The combined treatment of SB203580 and SP600125 was able to counteract the pro-migratory and invasive effects on cancer cells that had been treated with SB203580 alone (Fig. 5j–n). Anisomycin also reversed the pro-migratory and invasive effects of 25 µM morphine on pancreatic cancer cells, while SB203580 reversed the inhibitory effects on migration and invasion caused by 100 µM morphine on pancreatic cancer cells (Fig. 5o–s). We further showed that Anisomycin could reverse the effect of 25 µM morphine concentration on the p38/JNK pathway. Moreover, SB203580 can reverse the effect of 100 µM morphine concentration on p38/JNK pathway (Fig. S3c–e).

Fig. 5
figure 5

The P38 inhibitors and agonists reversed the effects of morphine on pancreatic cancer progression. (a) BxPC-3 cells were treated with SB203580 and anisomycin, respectively, and changes were detected using a WB assay. (b) The quantitative analysis of p −JNK/JNK in (a). (c) The quantitative analysis of p −p38/p38 in (a). (d–f) Investigate the effects of SB203580, anisomycin, and SP600125, both individually and in combination, on the proliferation of pancreatic cancer cells in the absence of morphine intervention. (g–i) Anisomycin reversed the growth-promoting effects of 25 µM morphine on pancreatic cancer, while SB203580 reversed the growth-inhibiting effects of 100 µM morphine on pancreatic cancer. (j–n) Investigate the effects of SB203580, anisomycin, and SP600125, either individually or in combination, on the migration and invasion of pancreatic cancer cells in the absence of morphine intervention. (o–s) Anisomycin reversed the pro-migratory and invasive effects of 25 µM morphine on pancreatic cancer, while SB203580 reversed the inhibitory effects on migration and invasion induced by 100 µM morphine on pancreatic cancer. In Fig. 5, “SB” corresponds to SB203580, “Ani” to Anisomycin, “SP” to SP600125, and “MF” represents the concentration of morphine. We analyzed three independent experiments, and all data were expressed as mean ± SD. Statistical significance was indicated as follows: *P < 0.05, **P < 0.01, and ***P < 0.001.

Full size image

Discussion

The effect of morphine on cancer malignancy varies across different types of cancer, suggesting potential differences in the surrounding microenvironment15. Furthermore, variations in morphine concentrations, types of cancer cells16, and the use of in vivo or in vitro tests may also contribute to this inconsistency17. Based on the above information, we summarized the effects of different morphine concentrations on cancer cell growth in the cited literature in Supplementary Table 1. For these reasons, we designed our experiments with different morphine concentrations, multiple pancreatic cancer cell lines, and in vitro/in vivo experiments. Our in vitro experiments illustrated that morphine barely affects HPDE cell growth but has a bidirectional effect on pancreatic cancer cell progression. This implies that the administration of morphine in clinical settings necessitates close observation for alterations in the proliferative tendencies and metastatic behavior of pancreatic cancer. Clinicians should adopt a cautious approach when prescribing morphine for pain management in pancreatic cancer patients. It is essential to monitor patients closely for any signs of disease progression after initiating morphine therapy, especially at lower therapeutic concentrations where a potential pro-tumorigenic effect was observed. Dosing strategies should be individualized, taking into account the patient’s pain severity, overall health status, and response to treatment. Gupta et al. found that in breast cancer, morphine (0.01–100 µM) inhibited apoptosis by activating PKB/Akt pathway, and promoted cell cycle progression by increasing cyclin D1, thus promoting tumor proliferation; However, when morphine > 1000 µM was used to interfere with breast cancer cells, it inhibited the above process and induced apoptosis of cancer cells18. The drug concentration reports of morphine bidirectional effects in this study differ from ours, most likely due to different types of cancer. Bimonte et al. also mentioned that morphine may stimulate cancer cell proliferation at low concentrations but inhibit it at high concentrations19, which aligns with our conclusions.

The mechanism of action by which morphine affects the malignancy of different cancers varies. For example, in non-small cell lung cancer (NSCLC) cells, morphine (0.1 µg/µL in vitro and 1.5 mg/kg in vivo) promotes malignant behavior through MOR/Src/mTOR signaling13. Through the HIF-1alpha/p38MAPK pathway, morphine (1µM in vitro, 20 mg/kg/day and 30 mg/kg/day in vivo) suppresses tumor angiogenesis14. In the current study, we confirmed that morphine affects pancreatic cancer progression through the p38/JNK pathway, which is different from the pathway studied in other cancers. In cancer, p38 and JNK signaling are highly influenced by the cellular context and tissue of origin20. A study by Hu et al. suggested that both p38 and JNK are pro-proliferative in vascular smooth muscle cells21. However, in Zhang et al., both p38 and JNK were suggested to be pro-apoptotic in glioma cells22. Therefore, the effects of both on cell proliferation may be completely opposite in different tissues. P38 MAPK (Mitogen-Activated Protein Kinase) and JNK (c-Jun N-Terminal Kinase) constitute two pivotal signaling cascades involved in cellular responses to stress, including processes such as proliferation, differentiation, transformation, and programmed cell death (apoptosis)23. In pancreatic cancer, activation of these two pathways is strongly associated with tumor development24. In pancreatic cancer, activation of the p38 pathway has been found to induce apoptosis in cancer cells25, whereas JNK activation has been reported to promote the proliferation of pancreatic cancer cells26. We detected changes in the expression levels of MOR/Src/mTOR and HIF-1alpha/p38MAPK related pathways after the intervention of pancreatic cancer cells with morphine. We found that the expressions of Src, mTOR and HIF-1alpha were not significantly changed, while the expressions of MOR and p38MAPK were significantly changed. Further, we found that morphine affects pancreatic cancer progression through the p38/JNK pathway. The p38/JNK pathway has not been reported in studies on the effect of morphine on tumor progression. Additionally, in vitro, inhibition of JNK inhibited the progression of pancreatic cancer cells in a genetically engineered mouse model27. Zhong et al. found that high expression of p-p38 in pancreatic cancer tissues significantly increased the survival rate of pancreatic cancer patients. In vitro studies, the p38 inhibitor SB202190 promoted the proliferation of pancreatic cancer cells and the increase of p-JNK expression, while the JNK inhibitor SP600125 reversed the proliferative effect of SB202190 on pancreatic cancer. Accordingly, Zhong et al. proved that functional p38 MAPK can inhibit JNK and thereby inhibit pancreatic cancer growth28. These two studies suggest that in pancreatic cancer, JNK promotes growth, while p38 inhibits growth, and that p38 probably functions by inhibiting JNK activation. Our study also confirms that in pancreatic cancer, p38 inhibition causes phosphorylation activation of JNK; conversely, phosphorylation activation of p38 causes JNK inhibition. Therefore, it is likely that low concentrations of morphine in our study promote pancreatic cancer progression by inhibiting p38 and thereby activating the JNK pathway. Conversely, it is likely that high concentrations of morphine inhibit progression by activating p38 and thereby inhibiting the JNK pathway.

The present study further supports the notion that morphine has a bidirectional effect on the degree of cancer malignancy and, to some extent, explains the conflicting results obtained in this area of research. For example, Ustun et al. suggested that morphine stimulates angiogenesis in breast cancer (0.714 mg/kg/day)29, and Sasamura et al. suggested that morphine inhibits the growth of B16-BL6 melanoma (5 and 10 mg/kg/day)30. The present study suggests that morphine affects pancreatic cancer progression through the p38/JNK pathway, which has not been observed in other cancers. The morphine concentrations used in this study are similar to clinically relevant concentrations of morphine, which may provide valuable information for the clinical use of morphine in pancreatic cancer patients experiencing advanced cancer pain. However, this study cannot demonstrate the conclusion from a histological perspective, and the mechanism of morphine is not explored in sufficient depth. This is an area that requires further research.

Our data suggests that varying concentrations of morphine have a bidirectional effect on the progression of pancreatic cancer, and this effect may be achieved through the modulation of the p38/JNK pathway. These findings have the potential to offer clinicians a reference point for guiding the appropriate use of morphine in patients with pancreatic cancer.

Methods

Cell culture

All cell lines used in this study were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Normoductal epithelial cells of the pancreas (HPDE) and cancer cells of the pancreas (BxPC-3, PANC-1, SUIT-2, CFPAC-1) were cultured in basal medium from Gibco, which contained 10% FBS (Gibco), 1% penicillin and streptomycin mixture (Solarbio). The cells were then placed in a humidified environment at 37 °C and 5% CO2. The morphine sulfate (tablet, 30 mg) used in the experiments was purchased from Mundi (China) Pharmaceutical Co. Ltd. Morphine was dissolved in complete medium for the in vitro experiments and in normal saline (NS) for the in vivo experiments. Based on the clinically relevant drug concentration of morphine of 62.5µM31 and our preliminary experiments, the morphine concentration was set to five concentration gradients: 0 µM, 25 µM, 50 µM, 75 µM, and 100 µM for cell intervention. In our experiments, pancreatic cancer cells were pre-treated for 24 h with a concentration of 10 µM p38 inhibitor SB20358032 (Sigma-Aldrich), or alternatively, with a concentration of 5 µM p38 agonist anisomycin33 (MedChemExpress), or with a concentration of 10 µM JNK inhibitor SP60012528 (MedChemExpress).

Cell viability test

Cells were inoculated into 96-well plates at a density of 2 × 103 cells per well. After 24 h, the original medium was removed and the appropriate concentration of morphine was added to each well. The plate was then incubated for 24 h, 48 h, and 72 h. After a specific time period, we added 10% CCK-8 reagent (Meilunbio) to the supernatant. The optical density (OD) was measured at 450 nm using a BioTek Synergy H1 microplate reader after the cells had been incubated for 1 h. Each assay was repeated at least three times.

Colony formation experiment

Cells were inoculated into six-well plates at a density of 1 × 103 cells per well and incubated for 7 days. The medium was then discarded and replaced with the appropriate concentration of morphine, which was incubated for 72 h. After that, 4% paraformaldehyde was applied for 30 min, followed by a PBS wash and crystal violet staining for 15 min. ImageJ was utilized to count colonies containing more than 50 cells. Each test was repeated at least three times.

EdU test

The cells were uniformly seeded into 96-well plates at a density of 2 × 103 cells/well and incubated for 24 h to adhere to the walls. The original medium was then discarded, and the culture was continued with the appropriate concentration of morphine for 72 h. Following the manufacturer’s instructions, we measured the proliferation of each cell group using the EdU-594 Cell Proliferation Assay Kit (Beyotime). The cells were labeled with EdU to observe their proliferation, and imaging was conducted using an EVOS FL Auto microscope. Hoechst-labeled and EdU-labeled cells were counted separately using ImageJ, and the rate of EdU-positive cells was calculated. Three experiments were repeated for each condition.

Wound healing experiment

Cells were inoculated and grown to 95% confluence in six-well plates. An “injured” monolayer of cells was scratched using a 200µL pipette tip. The original medium was discarded, and incubation was continued by adding serum-free medium with the appropriate concentration of morphine for 24 h. Wound images were taken at 0 and 24 h using an Olympus camera from Tokyo, Japan. The scratch area was calculated using ImageJ, and the scratch healing rate was determined using the formula: (scratch area at 0 h - scratch area at 24 h) divided by scratch area at 0 h, multiplied by 100%. Each experiment was repeated at least three times.

Transwell migration and invasion experiment

Migration assays were conducted using transwell inserts with a 6.5-mm, 8.0-µm-pore polyester membrane (Corning). Transwell inserts for invasion were precoated with Matrigel. Cells were suspended in 200 µL medium containing 2% FBS at the appropriate morphine concentration and inoculated in the inner chamber (5 × 104 cells/chamber). Additionally, 500µL of medium containing 20% FBS was added to the outer chamber. After 72 h, we fixed the cells in 4% paraformaldehyde and stained them with crystal violet. Then, we removed the non-migrating/invading cells. Under the microscope, random fields of view were photographed. We repeated each experiment at least three times.

Tumor xenograft test

Eighteen 5-week-old female Balb/c nude mice were divided into three groups: the low-dose morphine group, the high-dose morphine group, and the control group. The number of mice used in each group was six. All mice were injected subcutaneously with 5 × 106 pancreatic cancer cells (100µL cell suspension) in the axillary region. Seven days later, the experimental group was injected intraperitoneally daily with the corresponding dose of morphine, and the control group was injected with NS for three weeks. According to the relevant literature9,12,13 and the results of our preliminary experiments, the low-dose morphine group and the high-dose morphine group were injected with 0.5 mg/kg and 5 mg/kg per day, respectively. In the case of a 60 kg adult, 30 mg and 300 mg of morphine per day are equivalent. Mice were sacrificed on day 31 by overdose of sodium pentobarbital, and tumor tissue was isolated and weighed. The calculation for tumor volume (mm3) is as follows: volume (mm3) = 1/2 × (length × width × width). To verify whether morphine alone affects body weight in mice, we performed the same intervention in normal mice with the above two morphine concentrations for the same period of 3 weeks. In 3 weeks, administer a large dose of pentobarbital sodium to euthanize the mouse. The animal experiment complied with the requirements of the Experimental Animal Ethics Committee of Guangxi Medical University (No: 202108008). The animal experiment complied with the NIH Guide for the Care and Use of Laboratory Animals. All authors complied with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

Western blotting

Cell samples were collected and washed with cold PBS. Then, the cells were lysed with RIPA lysis solution (Solarbio) supplemented with phosphatase and protease inhibitors (Solarbio). The concentration of protein was detected using the BCA test (Boster). A 10% SDS-PAGE gel was used to separate cell extracts, which were then transferred to PVDF membranes and blocked with BSA (Sangon Biotech). Antibodies were diluted with primary antibody diluent (Beyotime) and incubated for 12 h at 4 °C. They were then incubated with secondary antibodies (Invitrogen) at room temperature for 1 h. The primary antibodies employed in the study included JNK, p-JNK, p38, and p-p38 (all from CST); Bcl-xL, Bcl-2, Bax, E-cad, and Vim (all from Abcam); HIF1α (Baijia); p-mTOR, mTOR, p-Src, and Src (all from Abmart); MOR (Proteintech); and GAPDH (Abmart). Protein detection was performed using LI-COR Odyssey (LI-COR). The bands were analyzed for gray values using ImageJ. Each experiment was repeated at least three times.

Statistical analysis

Each result was calculated from the values of at least three independently repeated experiments, and all data are expressed as mean ± SD. Comparisons between two groups of samples that followed a normal distribution with Chi-squared were made using Student’s t-test; otherwise, the Mann-Whitney test or Welch’s test was used. A p-value of less than 0.05 was considered statistically significant. Statistical results were calculated and plotted using GraphPad Prism 7 software.