Modification of radiosensitivity by Curcumin in human pancreatic cancer cell lines

Pancreatic cancer is one of the most aggressive malignancies and is characterized by a low 5-year survival rate, a broad genetic diversity and a high resistance to conventional therapies. As a result, novel therapeutic agents to improve the current situation are needed urgently. Curcumin, a polyphenolic colorant derived from Curcuma longa root, showed pleiotropic influences on cellular pathways in vitro and amongst others anti-cancer properties including sensitization of tumor cells to chemo- and radiation-therapy. In this study, we evaluated the impact of Curcumin on the radiosensitivity of the established human pancreatic cancer cell lines Panc-1 and MiaPaCa-2 in vitro. In contrast to MiaPaCa-2 cells, we found a significant radiosensitization by Curcumin in the more radioresistant Panc-1 cells, possibly caused by cell cycle arrest in the most radiation-sensitive G2/M-phase at the time of irradiation. Furthermore, a significant enhancement of radiation-induced apoptosis, DNA-double-strand breaks and G2/M-arrest after curcumin treatment was observed in both cell lines. These in vitro findings suggest that especially patients with more radioresistant tumors could benefit from a radiation-concomitant, phytotherapeutic therapy with Curcumin.

mitochondrial function, cell membrane and epigenetic changes were described using Curcumin in tumor cell treatment [16][17][18][19][20] . Some in vitro studies demonstrated radiosensitization by Curcumin, e.g. in colorectal carcinoma, prostate, lung or head and neck cancer [21][22][23][24] , and it is even postulated for pancreatic cancer cells 25 . Besides the effect of Curcumin on radiation efficacy, a sensitization to chemotherapeutic drugs like Gemcitabine was shown in vitro 26 .
In this study, the effect of Curcumin on radiosensitivity of the established human pancreatic cancer cell lines Panc-1 and MiaPaCa-2, as well as the effects of a combined treatment on cell cycle distribution, apoptosis and yH2AX-phosphorylation were investigated in vitro.

Results
Panc-1 cells are more radioresistant than MiaPaCa-2 cells. To investigate the radiosensitivity of pancreatic cancer cell lines Panc-1 and MiaPaCa-2, cells were seeded for CFA and irradiated with doses of 0, 2, 4, 6 and 8 Gy. Panc-1 cells showed compared to MiaPaCa-2 cells a significantly higher survival fraction at every radiation dose (Fig. 1). D 50 (irradiation dose to reduce the survival rate to 50%) was 3.31 Gy for Panc-1 cells and 1.70 Gy for MiaPaCa-2 cells.
In contrast to Panc-1 cells, the MiaPaCa-2 cell line showed no significant radiosensitization in any treatment combination.
Curcumin treatment alone did not show apoptosis induction in both cell lines. However, treatment with a low curcumin concentration (6 µM) resulted in a significantly lower fraction of apoptotic cells (p = 0.0449) in the Panc-1 cell line (Fig. 4).
Combined treatment of 8 Gy with 12 µM Curcumin increased the radiation-induced apoptosis significantly in both cell lines (Panc-1 cells: p = 0.0048; MiaPaCa-2 cells: p = 0.0021). In MiaPaCa-2 cells, the enhancement of radiation-induced apoptosis was significant for the combination of 4 Gy and 12 µM as well (p = 0.0001). In both cell lines, a concentration of 6 µM curcumin showed no significant difference compared to the irradiated control without curcumin.
Curcumin increases radiation-induced DNA damage. Next, we evaluated yH2AX as a marker for DNA-damage, especially DNA double-strand breaks. As expected, irradiation with 4 and 8 Gy increased the yH2AX-MFI in Panc-1 and MiaPaCa-2 cells significantly (Fig. 5). 12 µM Curcumin led to significantly higher relative yH2AX-MFI in unirradiated and irradiated Panc-1 and MiaPaCa-2 cells. Consistent with the other performed experiments in this study, a concentration of 6 µM Curcumin did not significantly change the yH2AX MFI compared to the respective control group.
Curcumin enhances radiation-induced G2/M arrest. To determine if Curcumin modifies the effects of irradiation on cell cycle distribution, cell cycle analysis was performed 24 hours after irradiation. As expected, the percentage of Panc-1 and MiaPaCa-2 cells in the G2/M-phase increased significantly after exposure to 4 and 8 Gy irradiation (Fig. 6).
Treatment with 12 µM Curcumin revealed a significantly higher G2/M-fraction of Panc-1 cells in the unirradiated control (p = 0.0021) and also in combination with 4 Gy irradiation (p = 0.0034). Concurrent to the increase of the G2/M-fraction, the G0/G1-fraction was significantly reduced. Irradiation with 4 Gy in combination with 6 µM Curcumin led to a significantly higher portion of cells in the S-phase.   www.nature.com/scientificreports www.nature.com/scientificreports/ MiaPaCa-2 cells treated with 12 µM Curcumin comprised a significantly higher G2/M-fraction at every combination regime. Accordingly, the G0/G1-phase was reduced significantly by 12 µM Curcumin in unirradiated cells and in cells irradiated with 8 Gy. After a combined treatment with 4 Gy irradiation, the S-phase-fraction was significantly lowered by 12 µM Curcumin.

Curcumin significantly enhances G2/M-fraction at time of irradiation in Panc-1 cells. As a G2/M
cell cycle arrest in non-irradiated pancreatic cancer cells was observed 48 h hours after the addition of 12 µM Curcumin (Fig. 6), we figured out whether the cell cycle distribution was also affected by Curcumin at the time point of irradiation (24 h after Curcumin addition). In line with the 48 h data, 6 µM Curcumin did not show an effect on cell cycle distribution after 24 h in both cell lines (Fig. 7). Most importantly, Panc-1 cells treated with 12 µM Curcumin for 24 hours showed a significantly higher fraction of cells in the radiosensitive G2/M-phase (p = 0.0149). In contrast, cell cycle distribution of MiaPaCa-2 cells was not significantly affected by treatment with 12 µM Curcumin. This is in line with data from CFA demonstrating radiosensitization in Panc-1 but not in MiaPaCa-2 cells with 12 µM Curcumin. These data suggest that the Curcumin-induced cell cycle arrest in the G2/M-phase, which is the most radiosensitive phase, might be a possible reason for the radiosensitization by Curcumin in Panc-1 cells.  www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
Although there were great advancements in oncology achieved over the last years, prognosis and survival of patients with pancreatic cancer remain very poor. Late diagnosis, early metastasis, bad surgical accessibility and high therapy resistance mark PDAC and limit therapeutic options. Therefore, a variety of agents is evaluated in ongoing in vitro and in vivo studies to improve the efficacy of RT and to overcome high chemo-and radiation resistance of PDAC. Besides conventional and new chemotherapeutics, promising phytotherapeutics are used in pancreatic cancer research.
One potent example is Curcumin, an orange pigment derived from Curcuma longa root, which is traditionally used in Chinese medicine and showed auspicious results in in vitro studies. Besides an observed sensitization to chemotherapy, a radiosensitization of tumor cells is postulated by Curcumin treatment 5,13,27 . In contrast, anti-inflammatory and anti-fibrogenic properties of Curcumin suggest radioprotection of healthy tissues 5 .
In this study, we evaluated radiosensitization effects of Curcumin in two established human pancreatic cancer cell lines. Secondly, we investigated apoptosis induction, yH2AX as an indicator for DNA-double strand breaks and cell cycle distribution to determine the mechanisms underlying radiosensitization.
The efficacy of Curcumin treatment strongly depends on the concentration and also on the formulation used in tumor cell treatment in vitro 28 . Most in vitro studies in pancreatic cancer cells used concentrations of 5-20 µM to evaluate the impact of stand-alone Curcumin treatment on tumor cell survival and cellular pathways [29][30][31] . Therefore, in the present study Curcumin concentrations of 6, 10 and 12 µM were chosen to investigate radiosensitizing effects in the pancreatic cancer cell lines Panc-1 and MiaPaCa-2.
Both cell lines showed comparable sensitivity to Curcumin (Fig. 2) with IC 50 values of 9.5 µM for Panc-1 and 9.0 µM for MiaPaCa-2 cells. Respective other studies, which used a different method to measure cell survival, calculated slightly higher IC 50 values (e.g. 15 µM 29 or 25 µM 27 for Panc-1 cells). In line with the literature 32 . Panc-1 cells revealed higher radioresistance than MiaPaCa-2 cells (Fig. 1).  www.nature.com/scientificreports www.nature.com/scientificreports/ Most exciting in our study is the difference in radioresponse upon Curcumin treatment between the two pancreatic cancer cell lines Panc-1 and MiaPaCa-2. Whereas the more radioresistant Panc-1 cells showed a significant sensitization to irradiation in CFA, MiaPaCa-2 cells revealed no radiosensitization.
Radiosensitizing effects by Curcumin were observed in various tumor entities. For example, Javvadi et al. 33 observed significant radiosensitization by Curcumin in cervical tumor cell lines. Veeraraghavan et al. 25 first examined the combination of radiation with Curcumin in pancreatic cancer cells. They showed significantly lower cell viability, higher cell death and potentiated apoptosis induction after combination of 100 nM Curcumin with irradiation in Trypan blue exclusion assay, MTT assay and Annexin V-FITC-staining. From their findings they postulated a potential radiosensitization by Curcumin in BxPC-3, Panc-1 and MiaPaCa-2 cells, but they did not perform CFA as the gold standard to determine radiosensitization.
Curcumin has shown multiple effects on cellular pathways that are deregulated in cancer cells, as well as influences on cellular functions like DNA-damage repair, mitrochondrial function or cellular membranes in vitro 13 . Especially basal and therapy-induced overactivation of the NFkB-pathway is associated with chemo-and radio-resistance 34 . Down-regulation of NFkB by Curcumin and accompanied apoptosis induction was observed in various tumor models 35 and discussed as main mechanism of Curcumin's impact on tumor biology 28,35,36 . In vitro experiments with lung cancer cells showed down-regulation of EGFR-and NFkB-AKT-pathway leading to inhibition of proliferation, apoptosis induction and radiosensitization after Curcumin treatment 22,37 . Also in prostate cancer 23 , oesophageal cancer 38 and in head and neck squamous cell carcinoma cells 24 radiosensitization by Curcumin was observed and associated with its impact on NFkB-and EGFR-pathways. In pancreatic cancer cell lines radiation-induced NFkB activity was inhibited by Curcumin consequential leading to a significantly higher apoptosis induction 25 . Therefore, Veeraghavan et al. postulated a down-regulation of the NFkB-pathway as a cause for radiosensitization in pancreatic cancer cells.
An association between cyclooxygenase-2 (COX-2) expression and apoptosis induction by Curcumin was found for lung and pancreatic cancer cell lines, revealing increased Curcumin sensitivity of high COX-2 expressing cells compared to low-expressing ones 27,28 . In line with these data, we observed no significant apoptosis induction by a single treatment with Curcumin in the COX-2 low expressing cell lines Panc-1 and MiaPaCa-2 cells.
Interestingly, in combination with irradiation, Curcumin enhanced radiation-induced apoptosis in both cell lines. Since we observed radiosensitization in Panc-1 cells, but not in MiaPaCa-2 cells without significant difference in apoptosis induction by Curcumin treatment between the two cell lines, we concluded, that other mechanisms for radiosensitization have to be evaluated.
Since The G2/M phase is known to be the most sensitive to irradiation treatment 42 . Furthermore, DNA damage or blocked DNA replication forks lead to a G2/M arrest 43 . A single treatment with Curcumin has been shown to induce cell cycle arrest in G2/M-phase in vitro 40 . As an example, Subramaniam et al. 19 found a G2/M arrest for Panc-1 and MiaPaCa-2 cells after incubation with 30 µM Curcumin for 24 hours. Based on these data, we analyzed cell cycle distribution of curcumin-treated human pancreatic cancer cells as a potential mechanism of radiosensitization.
Comparable to Veeraraghavan et al. 25 , a pre-treatment of Panc-1 and MiaPaca-2 cells with Curcumin led to a higher amount of cells in G2/M-phase 24 hours after irradiation. But more interestingly, we found a significant increase of the G2/M-fraction at the irradiation time point only in Panc-1 but not in MiaPaCa-2 cells after treatment with Curcumin. The Curcumin-induced G2/M phase arrest in the more radioresistant Panc-1 cells might be the reason for the observed radiosensitizing effects in CFA for Panc-1 cells.
On the one hand, Curcumin sensitizes cancer cells to anti-tumor treatment in vitro; on the other hand, anti-inflammatory properties postulate lower therapy side effects under concomitant phytotherapeutical treatment. Oral intake of Curcumin showed for example, significantly reduced bowel toxicity after abdominal irradiation in rats and lower radiation-induced pneumonitis after irradiation of rat lungs 44 . Wound-healing was significantly accelerated in Curcumin pre-treated mice undergoing fractionated RT after surgery 45 . In humans, oral doses up to 12 g daily showed no toxic side effects and were well tolerated 46 . A randomized treatment of breast cancer patients medicated with 6 g Curcumin daily parallel to radiation therapy showed significant reduction of radiation dermatitis severity and moist desquamation, but no significant effects on pain, redness or attendant symptoms like nausea or fatigue 47 . CT-evaluated body consumption and weight loss were evaluated in patients with advanced pancreatic cancer receiving 8 g Curcumin per day. No significant difference compared to the control group was found 48 . Considering the metabolic rate of curcumin in human, an oral intake 6 to 8 hours before radiotherapy would be suggested as unformulated curcumin reached the maximum blood concentration at that time 49 .
However, caused by chemistry and pharmacology, Curcumin has a very low bioavailability, chemical instability and fast metabolism. Blood levels after oral intake of 8 g Curcumin daily remained very low and did not outrange a concentration of 40 ng/ml equivalent to only 0.11 µM 6 . Actually, other studies detected no Curcumin in the blood of humans after a single oral intake 50 . Compared to the effective tumor-suppressive and radiosensitizing concentrations used in vitro, blood levels in humans are more than 100 times lower. (2020) 10:3815 | https://doi.org/10.1038/s41598-020-60765-1 www.nature.com/scientificreports www.nature.com/scientificreports/ A major characteristic of PDAC is the presence of a dense desmoplastic tumor stroma with a volume up to 70% of the total tumor volume, which contains cellular, acellular and biophysical components. The dysplastic stroma serves as a physical and biological barrier for drug delivery, an unfavorable pharmacokinetic and pharmacodynamic profile resulting in a high drug resistance. Therefore, mechanisms such as vasculature permeability, blood vessel patency, drug activation/degradation enzymes or target specific biological factors are develped to improve intratumoral distribution and bioavailability aiming to overcome drug resistance 51 . Besides the stromal barrier, a highly heterogeneous and genetically diversity as well as the tumor microenvironment contribute to high therapeutic resistance in PDAC. Subtype-driven therapy decisions are not yet implemented in clinical routine and are restricted to very few molecular alterations. Although targeting of various stromal components and pathways was considered a promising therapeutic strategy, none of the efforts had yet led to efficacious and approved therapies in patients with PDAC 52 .
To address the poor bioavailability and increase blood concentrations of Curcumin for an more effective treatment with Curcumin in cancer patients, several approaches and chemical modifications like inhibition of metabolization, better absorption through covering in nanoparticles 12 or liposomes 9 , structural analoges 17 , phospholipid complexes or combinations with other dietary components are investigated in vitro and in vivo 53 . Small studies with healthy volunteers show higher blood levels of curcumin and its metabolites after oral intake of micelles or phospholipid complex formulations of curcumin. Besides the oral intake of curcumin, liposomal formulations are developed and evaluated for parenteral use. In cancer therapy especially nanoparticles are used 54 . Experiments with nanoparticles or carrier complexes in pancreatic cancer models are still ongoing 55 , while erythrocyte membrane cloaked curcumin-loaded nanoparticles showed tumor cell apoptosis as well as reduction in tumor size and tumor mass of hepatocellular tumors in mice without any observable toxicity 56 . Considering future research directions an individualization of stromal targeting and translational clinical trials comparing preclinical approaches in patient-derived xenografts with patients' response to personalize therapies are recommended 52 .

Conclusion
Treatment with Curcumin showed significant effects on survival, apoptosis, DNA-double-strand breaks and cell cycle distribution in pancreatic cancer cells in vitro. Only in the more radioresistant cell line Panc-1 a dose-dependent radiosensitization was achieved by Curcumin treatment. A Curcumin-induced cell cycle arrest in the radiosensitive G2/M phase is assumed as a possible reason for radiosensitization. In conclusion, our results warrant further studies to understand the detailed molecular mechanisms leading to radiosensitization by Curcumin and in the long term to enable an optimized and individualized treatment concept for pancreatic cancer patients.

Material and Methods
Reagents and treatment. Curcumin powder (Sigma-Aldrich, St. Louis, USA) was solved freshly in 100% DMSO to get a 50 mM stock solution and was then diluted with warm cell culture medium to concentrations of 12, 10 or 6 µM Curcumin. The highest end-concentration of DMSO was 0.02% (≙12 µM Curcumin). Cells were incubated with Curcumin 24 hours before irradiation and if not described otherwise, cell culture medium was changed by drug-free medium directly after radiation treatment.
Colony forming assay. Colony forming assay (CFA) was performed to measure radiosensitivity and sensitivity towards Curcumin. Cells were seeded in 12-well plates and 24 hours later treated with different Curcumin concentrations or drug-free cell culture medium (control group). Again 24 hours later cells were irradiated at RS225A irradiation device (Gulmay Medical Ltd/Xstrahl, Camberley, UK) with doses of 0, 2, 4, 6 or 8 Gy. After irradiation and medium change, cells grew 11 (MiaPaCa-2) or 12 (Panc-1) days and were then fixed with ice-cold methanol and stained with 0.1% crystal violet. Colonies were counted with the colony counter GelCount (Oxford Optronics). Survival curves of irradiated cells were fitted to the linear-quadratic model using GraphPad Prism (San Diego, USA). Survival curves of Curcumin-treated, non-irradiated cells were plotted by second order polynomial function.
Apoptosis. Cells were treated with 0, 6 or 12 µM Curcumin for 24 hours and then irradiated with 0, 4 or 8 Gy at RS225A irradiation device. Apoptotic cells were stained 48 hours after irradiation using CellEvent Caspase 3/7 assay in combination with SYTOX AADvanced dead cell stain (both ThermoFisher Scientific). FACS-analysis was performed using FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). Not-stained, but Curcumin-treated cells were used to exclude false-positive signals caused by autofluorescence of Curcumin.
Cell cycle. To analyze cell cycle distribution, cells were treated with Curcumin and 24 hours later irradiated with 0, 4 or 8 Gy. Cells were fixed with 70% cold ethanol and stained with propidium iodide (Invitrogen, ThermoFisher Scientific, USA) in the presence of RNAse at two different time points, directly at the time point of irradiation or 24 hours after irradiation. Results from flow cytometry (FACSCalibur flow cytometer, BD Biosciences, San Jose, CA, USA) were evaluated with ModFit-Software (Verity software house Inc). yH2AX. As a marker for DNA-double-strand-breaks, yH2AX-fluorescence intensity was measured using flow cytometry (FACSCalibur flow cytometer BD Biosciences, San Jose, CA, USA). Therefore, cell populations were