SEARCH

SEARCH BY CITATION

Keywords:

  • apoptosis;
  • gemcitabine;
  • innate resistance;
  • NME5;
  • pancreatic cancer

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The limited therapeutic effect of gemcitabine on pancreatic cancer is largely attributed to pre-existing or acquired resistance of the tumor cells. This study was aimed at screening for candidate resistance-related gene(s) and elucidating the underlying mechanisms. NME5 was found to be highly expressed in an innate gemcitabine-resistant human pancreatic cancer sample and the cell line PAXC002 derived from the sample. Downregulation of NME5 significantly reversed gemcitabine resistance in PAXC002 cells, whereas NME5 overexpression induced gemcitabine resistance in the pancreatic cancer cell line BxPC-3. NME5 attenuated the induction of apoptosis and cell cycle arrest induced by gemcitabine, probably accounting for the blunted sensitivity to gemcitabine. Furthermore, NME5 was demonstrated to play its role in a nuclear factor kappaB (NF-κB)-dependent manner. NME5 was capable of directly binding NF-κB, and possibly regulated its expression level in PAXC002 cells. Our results also suggest that NF-κB is a key executor of NME5 in regulating apoptosis and cell cycle. All of these data suggest that NME5 is a promising target for relieving innate gemcitabine resistance in pancreatic cancer cells.

Structured digital abstract


Abbreviations
FACS

fluorescence-activated cell sorting

FITC

fluorescein isothiocyanate

IC50

half-maximal inhibitory concentration

NF-κB

nuclear factor kappaB

PI

propidium iodide

shRNA

short hairpin RNA

siRNA

small interfering RNA

TCA

tumor chemosensitivity assay

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Pancreatic cancer is now the fourth leading cause of cancer death in the USA and the ninth in China, with an overall 5-year survival rate of < 5% and a median overall survival of < 6 months [1–3]. Pancreatic cancer is characterized by rapid and asymptomatic progression, and only 10–15% of the patients have a tumor localized to the pancreas at the time of diagnosis allowing a potentially curative resection [4,5].

For the chemotherapy of unresectable pancreatic cancer, gemcitabine, a cytotoxic nucleoside analog, currently remains the standard first-line chemotherapy agent available for the treatment of advanced pancreatic cancer [6]. However, gemcitabine results in a tumor response rate of only 12% [7], mainly because of inherent or acquired chemoresistance in most tumor cells [8]. Extensive research during the last decades has revealed several resistance mechanisms, including deficiencies in drug uptake, alteration of drug targets, activation of DNA repair pathways, contribution of the tumor microenvironment, and, especially, the genetic and/or epigenetic alterations involving tumor suppressor genes, proto-oncogenes, and antiapoptotic genes, which have been proposed to occur with a very high frequency in pancreatic cancer [9–13]. However, the major cause of the high level of chemoresistance seen in pancreatic cancer patients remains poorly understood. In addition, most gemcitabine-resistant models used to date could not fully reflect the characteristics of inherent resistant human pancreatic cancer, owing to their apparent phenotypic and molecular variability. In our previous work [14], we established a human pancreatic cancer cell line, PAXC002, which was shown to be more resistant to gemcitabine than several widely used pancreatic cancer cell lines. PAXC002 was derived from primary human pancreatic cancer samples without any previous chemotherapy, and preserved the pathological characteristics well, making it an ideal cell model with which to study the mechanisms underlying innate gemcitabine resistance. In this study, PAXC002 was used to screen for and to identify novel factor(s) contributing to inherent gemcitabine resistance.

NME5, a member of the mammalian gene family that encodes nucleoside diphosphate kinase-like molecules [denoted as non-metastatic 23 (nm23) genes], has been recently identified [15]. The nm23 gene family was reported to be involved in tumor metastasis suppression, and nm23-H1 was recognized as a biomarker in clinical practice to predict the prognosis of stage I non-small cell lung cancer [16,17]. The latest studies have suggested the deregulation of NME5 in urothelial carcinoma, the oral cancer cell line Tu183, and malignant breast cancer [18–20]. However, the function of NME5 in tumors remains to be elucidated.

In this study, we first performed a comprehensive evaluation of the role of NME5 in innate gemcitabine resistance in pancreatic cancer cells and the underlying mechanisms. The expression profiles of 31 candidate resistance-related genes were compared in PAXC002 and its nonresistant counterpart, and NME5 mRNA was found to be highly expressed in PAXC002 cells. A set of experiments were performed to identify the association of NME5 with inherent gemcitabine resistance. Our findings indicated that overexpression of NME5 attenuated cell apoptosis and cell cycle arrest induced by gemcitabine in a nuclear factor kappaB (NF-κB)-dependent manner, whereas NME5 knockdown significantly reversed gemcitabine resistance in PAXC002 cells, leading to the conclusion that NME5 might be an important contributor to innate resistance to gemcitabine in pancreatic cancer.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Innate gemcitabine-resistant pancreatic cancer samples/cell lines screen

Tumor xenografts generated from primary human pancreatic cancer specimens and two pancreatic cancer cell lines, MIA PaCa-2 and BxPC-3, in SCID mice were evaluated for cellular susceptibility to gemcitabine with an ex vivo tumor chemosensitivity assay (TCA). Cancer cells purified from the tumor tissues were exposed to five-fold serially diluted gemcitabine, ranging from 200 μm to 0.064 μm. A pancreatic cancer sample labeled PAX002 displayed obvious resistance to gemcitabine, with a more than 45-fold higher half-maximal inhibitory concentration (IC50) than the other samples (partial data are shown in Table 1 and Fig. 1A). For further identification of resistance, cell lines denoted PAXC002 and PAXC003 were established from PAX002 and its nonresistant counterpart PAX003, respectively, as previously described [14]. In vitro TCA was used for PAXC002, PAXC003 and pancreatic cancer cell lines including BxPC-3 and MIA PaCa-2, with treatment with five-fold serially diluted gemcitabine starting from 200 μm. PAXC002 was shown to be over 5000-fold more resistant to gemcitabine than the other cell lines (Table 2; Fig. 1B). Considering the fact that PAXC002 was derived from primary human pancreatic cancer without chemotherapy, it can be concluded that PAXC002 is an innately gemcitabine-resistant pancreatic cell line.

Table 1.   Basic information on pancreatic cancer specimens and IC50 of gemcitabine in pancreatic cancer cell lines and patient tumor xenografts (ex vivo TCA). Only the information for six pancreatic cancer specimens of 13 is shown.
ModelGenderHistological typingChemotherapy before operationIC50m)
BxPC-3FCarcinoma< 0.064
MIA Paca-2MAdenocarcinoma< 0.064
PAX001FAdenocarcinomaNo2.103
PAX002FAdenocarcinomaNo> 200
PAX003MAdenocarcinomaNo0.113
PAX005MAdenocarcinomaNo0.383
PAX011FAdenocarcinomaNo4.422
PAX016FAdenocarcinomaNo< 0.064
image

Figure 1.  A human pancreatic cancer specimen and the derived cell line showed remarkable resistance to gemcitabine. (A) Ex vivo growth inhibition assay. Tumor tissues from immunodeficient mice implanted with human pancreatic cancer specimen cell lines PAX002 and PAX003 or pancreatic cancer cell lines MIA PaCa-2 and BxPC-3 were dissociated into single cells and cultured in 96-well plates in the absence or presence of the indicated concentration of gemcitabine for 6 days. Cell viability was measured with the CellTiter Glo Viability Assay, and the inhibition rate was calculated. (B) In vitro TCA in PAXC002, PAXC003, MIA PaCa-2, and BxPC-3. Cells were cultured in a 96-well plate with or without the indicated concentrations of gemcitabine for 96 h. Cell viability was measured with the CellTiter Glo Viability Assay, and the inhibition rate was calculated.

Download figure to PowerPoint

Table 2.   IC50 of gemcitabine in pancreatic cancer cell lines and patient tumor-derived cell lines (mean ± SEM, n = 3).
Cell lineIC50m)
BxPC-30.0328 ± 0.00488
MIA Paca-20.0364 ± 0.0109
PAXC002> 200
PAXC0030.0127 ± 0.00744

Gemcitabine resistance-related gene screening

In order to explore potential gene(s) related to the resistance of PAXC002 cells to gemcitabine, quantitative real-time PCR was employed to compare the relative transcription levels of 31 candidate resistance-related genes (provided by X. Huang, Nanjing University, China) between PAXC002 and PAXC003 cells. These candidate genes were selected from over 1700 kinase-encoding genes according to their significantly altered transcription levels in three cancer cell lines, including MIA PaCa-2, with induced resistance to doxorubicin, a cytotoxic agent that is commonly used in a wide range of cancers. As shown in Fig. 2A, gene 16 (NME5) was notably highly expressed in PAXC002 cells, as indicated by an at least 15-fold increased mRNA level as compared with PAXC003 cells. Expression of NME5 protein was subsequently detected by western blot in several pancreatic cancer cell lines and primary human pancreatic cancer samples, and this also demonstrated that NME5 was particularly highly expressed in gemcitabine-resistant PAX002 and PAXC002 cells (Fig. 2B). Therefore, NME5 was postulated to be associated with the innate gemcitabine resistance of PAXC002 cells. It should be mentioned that another three genes, numbered 6, 8, and 25, also showed a > 1.5-fold expression difference between PAXC002 and PAXC003 cells. However, they appeared not to contribute to the sensitivity to gemcitabine in small interfering RNA (siRNA)-based functional target validation (Table S1). Therefore, they were excluded from our further analysis.

image

Figure 2.  A resistance-related gene screen revealed high-level expression of NME5 mRNA in the gemcitabine-resistant pancreatic cancer cell line PAXC002. (A) The relative transcription levels of 31 candidate resistance-related genes in PAXC002 and PAXC003 cells were determined by real-time PCR. Gene 16 (NME5) was shown to be exclusively highly expressed in PAXC002 cells. Each bar represents the relative expression level of NME5, with PAXC003 as 100%. *P < 0.05, ***P < 0.001. (B) NME5 protein expression levels in PAX001, PAX002, PAXC002, PAX003, PAXC003, PAX005, PAX011, PAX016, MIA PaCa-2 and BxPC-3 cells were determined by western blot. The experiment was repeated three times, and a representative blot is shown. (C) Bar diagram showing the relative expression level of NME5 in each sample, with PAXC003 as 100%. ***P < 0.001 versus PAXC003.

Download figure to PowerPoint

NME5 silencing significantly reversed gemcitabine resistance in PAXC002 cells

In order to understand the role that NME5 plays in gemcitabine resistance, we employed an siRNA approach to knock down NME5 expression in PAXC002 cells. Silencing efficacy was confirmed at the protein level by use of western blot (Fig. 3A), and the optimal conditions for transfection were determined for later use. As shown in Fig. 3B, NME5-knockdown cells became more sensitive to gemcitabine than parental cells (siControl, IC50 > 200 μm; NME5 siRNA, IC50 = 7.67 μm; more than 26-fold decrease in resistance). Specific short hairpin RNA (shRNA) targeting NME5 was used to further investigate the role of NME5 in vivo. NME5 shRNA2, the most potent of the three constructed shRNAs (Fig. 3C), was introduced into PAXC002 cells before implantation into mice. The T/C percentage (see Experimental procedures) in the NME5 shRNA group (5.00%) was significantly (P < 0.05) lower than that in the shControl group (42.13%) after 21 days of gemcitabine treatment (Fig. 3D). Taken together, the in vitro and in vivo studies indicated that NME5 interference significantly reversed the innate resistance of PAXC002 cells to gemcitabine.

image

Figure 3.  NME5 silencing sensitized PAXC002 cells to gemcitabine. (A) NME5 silencing with siRNA. Cell lysates of PAXC002 cells transfected with control siRNA (siControl) or the indicated concentration of NME5 siRNA were prepared at different time points after transfection, and subjected to western blot with antibody against NME5. NME5 expression in cells transfected with 5 pmol of NME5 siRNA was significantly lowered as compared with control. (B) In vitro efficacy study with NME5 knockdown. NME5 siRNA-transfected or siControl-transfected PAXC002 cells were cultured with gemcitabine for 96 h. Cell viability was determined with the CellTiter Glo Viability Assay, and the inhibition rate was calculated. (C) NME5 silencing with shRNA. Cell lysates of PAXC002 cells transfected with various shRNAs targeting NME5 (shNME5) or control shRNA (shControl) were subjected to western blot. shNME5-2 reduced NME5 protein expression most significantly as compared with shControl, and was chosen for later use. (D) In vivo efficacy study with NME5 knockdown. Mice were implanted with shNME5-transfected or shControl-tranfected PAXC002 cells, and treated with gemcitabine. Tumor volume was measured twice weekly, as indicated.

Download figure to PowerPoint

Overexpression of NME5 causes resistance to gemcitabine

To investigate whether NME5 overexpression induces gemcitabine resistance in nonresistant pancreatic cancer cells, NME5 was overexpressed in BxPC-3, a pancreatic cancer cell line showing no innate resistance to gemcitabine according to our study and previous reports. The cells were transfected with control or pCEP4-NEM5 plasmids. Western blot showed an approximately 4.5-fold increase in NME5 expression level in pCEP4-NEM5-transfected cells as compared with control (Fig. 4A). In vitro TCA showed that the IC50 of gemcitabine in NME5-overexpressing BxPC-3 cells (208.96 nm) was approximately seven-fold higher than that of the control (26.55 nm), as shown in Fig. 4B, indicating that NME5 overexpression induced resistance to gemcitabine in BxPC-3 cells.

image

Figure 4.  NME5 overexpression induced resistance to gemcitabine in BxPC-3. (A) BxPC-3 cells were transfected with pCEP4-NME5 or the empty vetor as control. The protein level of NME5 was detected 96 h after transfection by western blot, with antibody against NME5. (B) In vitro efficacy study with NME5 overexpression. pCEP4-transfected or pCEP4-NME5-transfected BxPC-3 cells were cultured with gemcitabine for 96 h. Cell viability was determined with the CellTiter Glo Viability Assay, and the inhibition rate was calculated.

Download figure to PowerPoint

NME5 attenuates activation of the apoptosis pathway induced by gemcitabine

Suppression of cancer cell growth can be attributed either to induction of apoptosis, or to cell cycle arrest, or to both [21]. To investigate the impact of NME5 on the ability of gemcitabine to induce apoptosis in PAXC002 cells, annexin V–fluorescein isothiocyanate (FITC) and propidium iodide (PI) labeling was used as a criterion to distinguish apoptotic cells. For siControl-transfected cells, the proportion of apoptotic cells was only increased from 1.52% to 22.82% after exposure to 40 μm gemcitabine. However, 50.44% of cells in the NME5 silencing group treated with gemcitabine had undergone apoptosis, as compared with 5.06% of cells without gemcitabine treatment (Fig. 5A,B).

image

Figure 5.  NME5 knockdown enhanced the proapoptotic effects of gemcitabine (Gem) in PAXC002 cells. (A) siNME5-transfected or siControl-transfected PAXC002 cells were cultured with or without 40 μm gemcitabine for 96 h. Cells were stained with annexin V and PI, and subjected to FACS analysis, and the percentage of cells undergoing apoptosis was measured. A representative experiment is shown, n = 3. The bottom left quadrant, negative for both annexin V and PI, shows the live cells; the bottom right quadrant, annexin V-positive and PI-negative, shows the early apoptotic cells; the top right quadrant, positive for both PI and annexin V, shows the late apoptotic or necrotic cells. (B) Bar graph showing the quantitative analysis of apoptotic cells in (A). Apoptosis (%) represented percentages of annexin V-positive and annexin V/PI-double-positive cells. ***P < 0.001. (C) Expression of Bcl-2, Bax, cytochrome c, cleaved caspase-9, total caspase-9, cleaved caspase-3 and total caspase-3 in NME5 siRNA-transfected or siControl-transfected PAXC002 cells with or without 96 h of gemcitabine treatment was detected by western blot.

Download figure to PowerPoint

To obtain further insights into the effect of NME5 knockdown on gemcitabine-induced apoptosis in PAXC002 cells, we assessed the activation of the caspase pathway. The protein levels of several key elements, including Bcl-2, Bax, cytochrome c, caspase-3, and caspase-9, were determined by western blot. The antiapoptotic Bcl-2 resides in the outer mitochondrial wall and inhibits cytochrome c release, thus preventing subsequent cleavage and activation of caspase-9 and caspase-3, which is responsible for destroying the cell from within [22,23]. A lowered ratio of endogenous levels of Bcl-2 to Bax has been shown to be related to cell apoptosis [24]. As shown in Fig. 5C, gemcitabine treatment failed to markedly activate the apoptosis pathway in the siControl-transfected group, as indicated by minor changes in the protein expression level, which was consistent with the fluorescence-activated cell sorting (FACS) results. In addition, NME5 downregulation did not alter apoptosis-related protein expression as such. However, in NME5-silenced and gemcitabine-treated cells, Bcl-2 expression was reduced to about 25%, whereas the expression levels of Bax, cytochrome c and the activated forms of caspase-9 and caspase-3 were increased by more than two-fold. All of these results suggest that the high level of NME5 in PAXC002 cells circumvented the apoptosis induced by gemcitabine, and NME5 interference made cells more prone to apoptosis.

NME5 inhibited gemcitabine-induced G1-phase arrest

Many studies have shown that the inhibition of pancreatic cancer cell growth by gemcitabine wis accompanied by cell cycle arrest in G1 phase [25–27]. Therefore, we explored the possibility that NME5 expression regulates this gemcitabine-induced cell cycle arrest. Cells were treated with control siRNA or NME5-targeting siRNA for 24 h, and subsequently exposed to 40 μm gemcitabine for 96 h. As shown in Fig. 6A,B, the cell cycle distribution of siControl-treated cells seldom changed. In contrast, NME5-silenced cells exhibited a > 10% increase in the G1-phase population after gemcitabine treatment (P < 0.05), indicating accumulation at the G1 phase of the cell cycle.

image

Figure 6.  NME5 silencing facilitated G1 arrest as well as cyclin D1 upregulation induced by gemcitabine (Gem). (A) The cell cycles of NME5 siRNA-transfected or siControl-transfected PAXC002 cells with or without gemcitabine exposure were investigated with FACS. Histograms show the DNA content distribution (PI staining intensity) of cells on the x-axis for each treatment condition. A representative experiment is shown, n = 3. (B) The percentages of cells in G1, S and G2/M phasea are shown in the bar graph. *P < 0.05 versus NME5 siRNA-transfected and vehicle-treated groups. (C) Cell lysates of NME5 siRNA-transfected or siControl-transfected PAXC002 cells treated with or without gemcitabine were subjected to western blot with antibody against cyclin D1.

Download figure to PowerPoint

As a key positive regulator of G1-phase progression, cyclin D1 actively drives transit through the G1 checkpoint. Downregulation of cyclin D1 was was shown to be associated with tumor growth suppression. Our study demonstrated that the protein level of cyclin D1 was decreased to about 31% (P < 0.05) after treatment with gemcitabine only when the expression of NME5 in PAXC002 cells was silenced (Fig. 6C). These results confirmed our assumption that NME5 attenuated the inhibitory effect of gemcitabine on cell cycle progression, probably leading to the gemcitabine resistance of PAXC002 cells.

The NF-κB signaling pathway possibly links NME5 to gemcitabine resistance

There are lines of evidence suggesting that the transcription factor NF-κB p65 subunit is closely related to gemcitabine resistance in pancreatic cancer, and plays a pivotal role in cell cycle progression and suppression of apoptosis [28]. Therefore, we investigated whether NF-κB p65 served as an important molecule linking NME5 to gemcitabine resistance in PAXC002 cells. As shown in Fig. 7A, the protein level of NF-κB p65 in PAXC002 cells was much higher than that in PAXC003, MIA PaCa-2 and BxPC-3 cells, whereas NME5 knockdown substantially lowered the expression of NF-κB p65 in PAXC002 cells (Fig. 7B), suggesting that NME5 probably regulates NF-κB p65 expression. Additionally, immunoprecipitation analysis demonstrated that NME5 was able to bind NF-κB p65 (Fig. 7C), further supporting the association between NME5 and NF-κB p65. To determine whether the effect of NME5 on innate gemcitabine resistance was dependent on NF-κB signaling, siRNA targeting NF-κB p65 was used to downregulate its expression in PAXC002 cells. As demonstrated by in vitro TCA, NF-κB p65 silencing partially restored the sensitivity to gemcitabine (Fig. 7D). NF-κB p65 knockdown also reduced the protein levels of Bcl-2 and cyclin D1 (Fig. 7E) in PAXC002 cells after treatment with gemcitabine, which was consistent with the changes caused by NME5 knockdown. Bcl-2 and cyclin D1 were both known as target genes of NF-κB p65. On the basis of these results, we may conclude that NF-κB possibly mediates the effect of NME5 on apoptosis and cell cycle in gemcitabine-resistant PAXC002 cells.

image

Figure 7.  NME5 partially depended on NF-κB to exert its effect on gemcitabine resistance. (A) The expression levels of NF-κB p65 subunit in PAXC002, PAXC003, MIA PaCa-2 and BxPC-3 cells were detected by western blot with antibody against NF-κB p65. (B) NME5 silencing decreased the expression of NF-κB in PAXC002 cells. The protein levels of NME5 in NME5 siRNA-transfected or siControl-transfected PAXC002 cells were detected by western blot. (C) Binding of NME5 to NF-κB was demonstrated by immunoprecipation assay. Lysates from NME5 siRNA-transfected or siControl-transfected PAXC002 cells were used for immunoprecipitation with mouse antibody against NME5. The immunoprecipitates were electrophoretically resolved on the polyacrylamide gel and immunoblotted with antibodies against NME5 and anti-NF-κB p65. (D) NF-κB p65 siRNA-transfected or siControl-transfected PAXC002 cells were cultured with gemcitabine for 96 h. Cell viability was determined with the CellTiter Glo Viability Assay, and the inhibition rate was calculated. (E) Ninety-six hours after NF-κB p65 siRNA or siControl transfection and gemcitabine treatment, PAXC002 cells were lysed and western blot was performed with antibodies against NF-κB p65, Bcl-2, and cyclin D1. Ab antibody; IB, immunoblotting; IP, immunoprecipitation.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Gemcitabine is considered to be the most clinically active drug for unresectable pancreatic cancer, but it is only effective in a small fraction of patients, mainly because of pre-existing or acquired chemoresistance in most of the tumor cells [29]. Current research efforts are mostly focused on the acquired resistance, and rarely on the innate resistance to the chemotherapy agent, partially because of the difficulty in obtaining primary human pancreatic cancer samples with inherent resistance. In the present study, PAXC002, a human pancreatic cancer cell line that has been well characterized in our previous work, was used to explore novel factor(s) contributing to innate gemcitabine resistance. PAXC002 was shown to be over 5000-fold more resistant to gemcitabine than its counterpart PAXC003 and several commonly used pancreatic cancer cell lines. In addition, PAXC002 was found to be over 40-fold more resistant to 5-fluorouracil, another nucleoside analog with similar anticancer mechanisms, than PAXC003 (data not shown). It should be noted that the human pancreatic cancer specimens used for gemcitabine-resistant sample screening were all derived from patients who had not received previous chemotherapy or radiation. The response of pancreatic cancer xenografts to gemcitabine was evaluated by ex vivo TCA, a method that is widely used in efficacy studies of anticancer drugs, and that has been shown to be capable of predicting outcome in patients with high accuracy [30,31]. Furthermore, these samples were not treated with any anticancer drugs prior to the establishment of cell lines, and the cell lines were maintained and propagated under normal culture conditions. Therefore, PAXC002 is a suitable cell model for research on innate resistance to gemcitabine.

Previous studies proposed the involvement of several genetic alterations in gemcitabine resistance, including p53, integrin-linked kinase, and CECACM6 [32]. However, the genetic determinants of innate gemcitabine resistance have not yet been fully elucidated. In the present study, the expression profiles of 31 candidate genes in PAXC002 and PAXC003 cells were compared by using real-time PCR based on the resistance-related gene library kindly provided by X. Huang (Nanjing University, China). Gene 16 (NME5) was found to be highly expressed in gemcitabine-resistant PAX002 and PAXC002 cells as compared with other, nonresistant, samples. NME5 is a recently identified member of the nucleoside diphosphate kinase-like molecule family (nm23 genes) [15]. The first nm23 gene was isolated on the basis of its reduced expression level in highly metastatic murine melanomas, and was proposed to be a metastatic suppressor gene [16]. Since then, several murine nm23 genes (nm23-M1 to nm23-M7) and human nm23 genes (nm23-H1 to nm23-H8) have been cloned [33–36]. Previous studies indicated that NME5 was deregulated in urothelial carcinoma, the oral cancer cell line Tu183, and malignant breast cancer [18–20], but no studies have linked NME5 to gemcitabine resistance until now. In this study, for the first time, we proposed NME5 as an important contributor to innate gemcitabine resistance in pancreatic cancer cells. Our results showed that NME5 knockdown dramatically reversed the gemcitabine resistance of PAXC002 both in vitro and in vivo. NME5 overexpression caused resistance to gemcitabine in the nonresistant pancreatic cancer cell line BxPC-3. NME5 also circumvented the induction of apoptosis and cell cycle arrest, two important pathways mediating the inhibitory effect of gemcitabine in pancreatic tumor growth. All of these data imply that NME5 plays an important role in the innate gemcitabine resistance of PAXC002 cells.

NF-κB is thought to play an antiapoptotic role in various cancer cells through its ability to induce the expression of several proteins, including the inhibitors of apoptosis family members and Bcl-2 homologs [37]. Cyclin D1, a positive regulator of G1-phase progression associated with pancreatic cancer cell growth, is also under the control of NF-κB [38]. Recent studies have shown that NF-κB overexpression has a connection with resistance in hepatocellular carcinoma through cell cycle promotion and an antiapoptotic effect [39], and that constitutive NF-κB activity confers resistance to gemcitabine [29]. In our study, NF-κB p65 was demonstrated to be a possible executor for NME5 in regulating the cell cycle and apoptosis, as indicated by the fact that NF-κB p65 knockdown partially restored cellular sensitivity to gemcitabine in PAXC002 cells and reduced Bcl-2 and cyclin D1 expression in PAXC002 cells when treated with gemcitabine, which corresponded to the downregulation of NME5. Furthermore, NME5 was capable of binding NF-κB p65 and impacting on its expression level, suggesting that NME5 probably acted upstream of NF-κB p65 to regulate cell sensitivity to gemcitabine. However, it remained to be elucidated how NME5 regulated the expression and function of NF-κB p65. In addition to NF-κB p65, we have detected several critical proteins of other signaling pathways, including PLK1, RRM1, p-AKT and β-catenin, which are all related to chemoresistance. They appeared not to be impacted by NME5, as indicated by their unchanged protein expression levels after NME5 knockdown (data not shown). However, the possibility cannot be excluded that NME5 interacted with other molecules in these pathways or other signaling pathways that were not studied in our research to induce the resistance to gemcitabine.

In summary, our data provide the first evidence that NME5 overexpression might serve as an important contributor to the innate gemcitabine resistance of pancreatic cancer, thus providing a novel potential target to relieve innate resistance and to sensitize pancreatic cancers to gemcitabine treatment. However, the conclusion was drawn from only one inherently gemcitabine-resistant sample. It will be of great importance to further explore the relationship between NME5 expression level and gemcitabine resistance in clinical trials with larger sample sizes. Furthermore, NME5 might be related not only to innate gemcitabine resistance but also to acquired resistance, which is commonly found in clinical chemotherapy practice. Additionally, as NME5 knockdown could not completely reverse gemcitabine resistance, probably because of the complex mechanisms underlying this phenomenon, it is important to develop multitarget anticancer drugs or combination therapeutic strategies directed to multiple targets for the future therapy of gemcitabine-resistant pancreatic cancers.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Patient tumors

Human pancreatic adenocarcinoma samples were obtained from Shanghai Changhai hospital, in accordance with protocols approved by the hospital’s Institutional Ethical Committee. Written informed consent was obtained from each patient. Tumor samples were taken from freshly isolated surgical resections and transported to the specific pathogen-free animal facility in Shanghai ChemPartner Co. Ltd, which is accredited by Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC), for human primary tumor xenografts and establishment of cell lines. In all cases, the diagnosis of tumor cells was confirmed by independent histopathology review.

Cell culture

The human pancreatic cancer cell lines BxPC-3 and MIA PaCa-2 were obtained from the ATCC and maintained according to the ATCC’s instructions. Two cell lines, denoted PAXC002 and PAXC003, were developed from human pancreatic tumor tissues by Shanghai ChemPartner Co. Ltd, as described previously [14]. The two cell lines were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA, USA), 10 μg·mL−1 human recombinant insulin (Invitrogen), and 1% antibiotic–antimycotic (Invitrogen).

Gene knockdown with siRNA and shRNA

siRNAs targeting genes 6, 8, 16 and 25 (corresponding to CDK10, COL6A3, NME5, and VSNL1, respectively) and control siRNAs were purchased from Santa Cruz Biotech (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Each siRNA was transfected into cells with LipofectAMINE (Invitrogen, Carlsbad, CA, USA). The silencing efficacy was confirmed by western blot at sequential time points after transfection.

For the construction of NEM5 shRNA with a lentivirus-based system, oligonucleotides corresponding to the shRNA sequence directed against NME5 were annealed and subcloned into the EcoRI and BamHI restriction sites of the pLVX-shRNA (Clontech, Palo Alto, CA, USA) vector. The pLVX-siNEM5 plasmid was cotransfected with Δ8.9 and VGVS plasmid into 293T cells. Seventy-two hours later, the supernatant was centrifuged at 4000 g for 10 min to remove debris, and aliquots of the virus solution were stored at −80 °C. A scrambled shRNA was used as control shRNA in later experiments. For lentivirus infection, 5 × 105 PAXC002 cells were seeded in six-well plates 24 h before transduction. Concentrated lentiviruses were added to the medium with 8 μg·mL−1 polybrene (Millipore, Bedford, MA, USA). Three days after infection, cells were collected for cell sorting with FACSAria (BD Biosciences, Franklin Lakes, NJ, USA). The percentage of green fluorescent protein-positive cells reached > 95% after sorting.

NME5 overexpression

A full-length human NME5 cDNA was cloned by PCR from human genomic DNA extrated from BxPC-3, with the oligonucleotide primer pair 5′-GACGAAGCTTATGGAGATATCAATGCCTC-3′ (forward) and 5′-TGCAGGATCCTTAATAAGGTTCTTCTAC-3′ (reverse). The full-length NME5 gene was sequenced, amplified, and then inserted into the BamHI and HindIII restriction sites of pCEP4 (Invitrogen). The empty vector or pCEP4-NME5 were transfected into BxPC-3 with Xfect Transfection Reagent (Clontech). The NME5 expression level was confirmed by western blot with antibody against NME5.

Development of pancreatic tumor xenografts in immunodeficient mice

All surgical procedures and care applied to the animals were in accordance with IACUC guidelines. Human patient tumor fragments of ∼ 30 mm3 were implanted into the flanks of female SCID mice (Beijing Vital River, China), or 5–10 × 106 cancer cells were subcutaneously injected into the flanks of Nu/Nu mice to establish xenograft models. The tumor length (L) and width (W) were measured with digital calipers, and the tumor volume (TV) was calculated from the following formula: TV = 1/2 × L × W2. When tumors reached 300–500 mm3, the mice were killed, and the tumors were removed under sterile conditions and then used for ex vivo TCA.

Ex vivo TCA

Tumor xenografts were cut into 3–6-mm3 fragments and dissociated into single cells. Cancer cells were isolated and purified with a Cancer Cell Isolation Kit (Panomics, Redwood City, CA, USA), as previously described [31,40]. Tumor cells were seeded into 96-well ultralow culture plates (Costar, New York, NY, USA) at 20 000 cells per well in 150 μL of culture medium. Gemcitabine (Chemiceuticals, NC, USA) was dissolved in dimethylsulfoxide (final concentration, 0.5%) and five-fold serially diluted. Various concentrations of gemcitabine (ranging from 200 μm to 0.064 μm) was added to triplicate wells at the same time as cell inoculation, with diluent-treated cells as a control. After 6 days of treatment, cell viability was measured with the CellTiter Glo Luminescent Viability Assay (Promega, Madison, WI, USA), according to the manufacturer’s instruction. Drug effects were presented as IC50 values, which was determined with xlfit software (equation 205).

In vitro TCA

Tumor cells were seeded into 96-well tissue culture plates (Costar, New York, NY, USA) (BxPC-3 at 3000 cells per well, MIA PaCa-2 at 3000 cells per well, PAXC002 at 6000 cells per well, and PAXC003 at 4000 cells per well. Plating densities were determined from cell growth curves, and the control cells were still exponentially growing at the endpoint in 150 μL of culture medium. Four-fold or five-fold serially diluted gemcitabine (original concentration, 200 μm) in dimethylsulfoxide (final concentration, 0.5%) was added to triplicate wells. After 96 h of incubation, cell viability was measured and the IC50 was determined according to the methods mentioned above.

Real-time PCR

RNA from PAXC002 and PAXC003 was extracted and reverse-transcribed to cDNA with a SuperScript III First Strand Synthesis System Kit (Invitrogen) with oligo(dT). Real-time PCR was performed with a QuantiTect SYBR Green PCR Kit (Qiagen, Hilden, Germany) with β-actin as an endogenous control. The sequence of the primers was as follows: NME5, 5′-CCCCAACTTAACAGCTTACATG-3′ (forward) and 5′-CAGCAAAGTCATTACTCCCATG-3′ (reverse); and β-actin, 5′-GATGGCCACGGCTGCTTCCAGC-3′ (forward) and 5′-GCCAGGGTACATGGTGGTGCCG-3′ (reverse). mRNA expression was normalized to β-actin and represented as relative transcription level to expression in PAXC003 cells.

In vivo efficacy study

PAXC002 cells (5 × 106) stably expressing NME5 shRNA or control shRNA were subcutaneously implanted into mice with equal volume of Matrigel (BD Biosciences). When tumors reached 150–250 mm3∼ 2 weeks later, the mice for both PAXC002-shControl and PAXC002-shNME5 groups were randomly assigned to vehicle-treated and gemcitabine-treated groups (n = 6 for each group). Mice received either 10 mL·kg−1 vehicle solution or gemcitabine (40 mg·kg−1) through intraperitoneal injection every 4 days for 3 weeks. The tumor size and body weight of mice were measured twice weekly. The tumor size was then used for calculations of T/C values. The T/C value (in percentage) was an indication of antitumor effectiveness. T is the average tumor volume in the treatment group on a specific day minus the average tumor volume in the treatment group on day 0; C is the average tumor volume in the vehicle control group on a specific day minus the average tumor volume in the vehicle control group on day 0. All animal procedures were performed according to the IACUC guidelines.

Cell cycle analysis

Cells were seeded into six-well plates at a density of 5 × 105 cells per well, and treated with 40 μm gemcitabine for 96 h. Cells were harvested after two washes with cold NaCl/Pi, and fixed in cold 75% ethanol at 4 °C overnight. Cells were then washed twice with cold NaCl/Pi, and incubated with 10 μm PI (Sigma, St Louis, MO, USA) and 0.2 mm RNase for 30 min at 37 °C in the dark. Stained cells were analyzed immediately by FACS (BD Biosciences, Franklin Lakes, NJ, USA).

Detection of apoptosis

The apoptotic rates of cells was measured with the annexin V and PI double-staining kit (Beyotime Institute of Biotechnology, Jiangsu, China), according to the manufacturer’s instructions. Briefly, cells were treated with or without gemcitabine for 96 h. Then, cells were centrifuged at 1000 g for 5 min, washed in ice-cold NaCl/Pi, and stained with annexin V and PI for 30 min at room temperature. After washing with NaCl/Pi, apoptosis was measured by flow cytometry. Total apoptosis was expressed as percentages of annexin V-positive and annexin V/PI-double-positive cells.

Western blot

Cells were collected after treatment, and washed with NaCl/Pi. The whole cell lysates and cytosolic fraction were prepared as previously described [41,42]. Protein concentration was determined with the bicinchoninic protein estimation kit (Bio-Rad, Hercules, CA, USA). Cell lysates diluted with loading buffer were heat-denatured at 100 °C for 10 min before being run on 4–12% gradient SDS/PAGE. The proteins were then transferred onto nitrocellulose membranes and incubated with primary antibodies against NME5, Bcl-2, Bax, cytochrome c, cleaved caspase-3, total caspase-3, cleaved caspase-9, total caspase-9, cyclin D1, β-tubulin (all from Cell Signaling Technology, Danvers, MA, USA), and NF-κB p65 (Santa Cruz), and this was followed by incubation with fluorescent secondary antibody. β-Tubulin was used as loading control. Protein bands were detected with an Odyssey Infrared Imaging System. Semiquantitative analysis of band intensity was performed by densitometry with odyssey software (Odyssey, USA).

Coimmunoprecipitation

Cells were grown to 70–80% confluency and harvested. Cell lysates were prepared as described previously [43], and equal amounts of protein were incubated overnight with either control IgG or antibody against NME5 in a total volume of 500 μL. The mouse IgG was used as isotype control. Protein G–Sepharose beads (Beyotime, Jiangsu, China) were added to the lysate antibody mix and incubated on a rotating platform for 2.5–3.5 h at 4 °C, and this was followed by 3–4 washes with the lysis buffer. The immunoprecipitates or total cell lysates were then immunoblotted with mouse mAb against NME5 and antibody against NF-κB p65.

Statistical analysis

Statistical analysis was carried out with graph pad prism 5 software (Graph Pad Software, San Diego, CA, USA). Data are all presented as mean ± standard error of the mean (SEM) of at least three independent experiments. Results were compared by one-way ANOVA with Dunnett’s test. A P-value of < 0.05 was considered to be statistically significant.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

This work was supported by a 2010 Yangtze River Delta region cooperation grant (No. 10495810900), Mega-projects of Science Research for the 11th Five-Year Plan (No. 2009ZX09302-002), the 111 Project (111-2-07), and central university basic research and operating expenses (20100302). We thank X. Huang for providing the candidate resistance gene library.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1. IC50 of gemcitabine in indicated cell lines before and after gene knockdown with siRNA.

Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
FEBS_8521_sm_TableS1.zip38KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.