A TrxR inhibiting gold(I) NHC complex induces apoptosis through ASK1-p38-MAPK signaling in pancreatic cancer cells

Gold(I) NHC complexes had been identified as potent anti-tumor agents with IC50 value in the low micromolar concentration range in various cancer cell lines, which represent different tumors including MCF7 (breast cancer), HT29 and HCT116 (colon cancer), and HepG2 (liver cancer)[9, 10]. Here we investigated their effects on pancreatic cancer cells, one of the most difficult to treat cancers with only 5% of 5-year survival in patients[31]. In our first analysis, we evaluated the cytotoxicity of the gold(I) metal-complexes MC2, MC3 and MC4 in four pancreatic cancer cell lines, namely Bxpc3, Miapaca2, Panc1 and ASPC1 using MTT assay. The free ligand (MC1), 1,3-diethylbenz imidazolium iodide, and the well-known chemotherapeutic drug gemcitabine (Gem) were included as references[32]. The structures were listed in Figure 1A. After 72 h treatment we observed that Gem was sufficient to block 50% of cell growth in Bxpc3 and Miapaca2 cells at 10 nM, whereas IC50 values were more than 20 μM in Panc1 and ASPC1 (Figure 1B). In MC series, we could not detect any anti-proliferative effect of the organic ligand (MC1) at 50 μM in all of the four cell lines. By contrast, MC2, possessing a chloride as leaving group, exhibited a good inhibitory effect with IC50 values of 12.05 ± 0.44, 2.44 ± 0.26, 7.01 ± 0.35 and 6.40 ± 0.32 μM in Bxpc3, Miapaca2, Panc1 and ASPC1 respectively (Figure 1B), while MC4, carrying triphenylphosphine instead of chloride, inhibited cell growth with IC50 values of 0.08 ± 0.01, 0.08 ± 0.01, 0.82 ± 0.01 and 0.77 ± 0.06 μM in the respective cell lines (Figure 1B). Interestingly, the most potent compound was MC3, substituted with a second benzimidazolylidene moiety, whose IC50 values were 0.06 ± 0.01, 0.07 ± 0.02, 0.21 ± 0.02 and 0.28 ± 0.01 μM in corresponding PDACs (Figure 1B). Additionally, we also measured the short-term effect of MC3 and MC4 and detected remarkable toxicity of MC3 and MC4 after 24 h and 48 h treatment in Panc1 and ASPC1 (Figure 1B). Of note, MC3, MC4 and Gem exhibited low toxic effects on human neonatal foreskin fibroblasts (HFF), implicating they might preferentially kill cancer cells (Figure 1B). The MTT assay is a colorimetric proliferation and cell viability assay, which is based on the measurement of the activity of mitochondrial succinate dehydrogenase[33]. Since mitochondria had been shown to be a direct target of gold(I) NHC complexes[9, 10], we also used total protein quantification with sulforhodamine B, known as SRB assay, to measure cellular growth and proliferation[34]. The results obtained were similar in both assays. Since Panc1 and ASPC1 are resistant to Gem, we focused on these two cell lines, in particular Panc1, which possesses mutations in p53, in our further experiments[35, 36].

To shed light on the mechanism of cell death we stained treated cells with FITC-conjugated annexin v, which binds to phosphatidylserine translocated to the outer cell membrane layer during apoptosis[37, 38], and propidium iodide (PI), which is only taken up after membrane damage, to distinguish necrotic cells. Panc1 cells were incubated with increasing concentrations of MC3 for 24 and 48 hrs. After treatment cells were counted and analyzed by FACS (fluorescence activated cell sorting). Treatment not only reduced cell number, but also led to a concentration-dependent increase in the population of annexin v positive cells compared to DMF treatment as mock already after 24 h (Figure 2A). The significance of apoptotic cell death is even more clearly visible after 48 h treatment. Cell numbers are further reduced and the relative amount of early apoptotic and late apoptotic cells is increased in the remaining cell population (Figure 2B). Analysis of cell cycle distribution in intact cells after 24 h treatment, detected accumulation of cells at S and G2/M phases (Figure 2C), indicating an accompanying cell cycle arrest. To investigate the ability of cells to grow into a colony after treatment, we performed a colony formation assay by seeding equal amount of cells, pre-incubated with increasing concentrations of MC3 for 24 h. After 3 weeks of continuous cultivation, cells were stained with crystal violet. Hardly any visible colony could be found, when cells were treated with 5 μM or 10 μM during pre-incubation (Figure 2D). We also performed a wound healing assay, under conditions optimized for low cell proliferation (0.5% serum) with an assay-specific optimized concentration of MC3 (1.25 μM) to minimize toxicity. After 48 h incubation in this condition, we hardly saw any cell able to enter the gap generated with a scratch in the cell layer, indicating reduced cell mobility in presence of MC3 (Figure 2E).

Since 5 μM of MC3 was sufficient to obtain a significant amount of apoptotic cells after 24 h treatment (Figure 2A), we decided to assess the formation of ROS in a time-dependent study at this concentration. We used dihydroethidum (DHE) and FACS analysis for ROS detection. ROS levels were clearly changed, showing a dramatic elevation as early as 15 min (3-fold over mock) after treatment started, with a maximum after 3 h (4.4-fold) and 3.9-fold (over mock) after 24 h (Figure 3A). The slight decrease at the later time point might be explained by an increased compensation in cancer cells attempting to overcome MC3-indcued excessive ROS for survival. We further analyzed ROS formation at various concentrations at 3 h and 24 h, both showing a concentration-dependent increase (Figure 3B) in good agreement with previous findings that MC3 acts as a potent TrxR inhibitor[9, 10]. Since increase of ROS may be initiated from dysfunctional mitochondria[39], we analyzed the mitochondrial membrane potential, with the cytofluorimetric, lipophilic cationic dye, 5,5′,6,6′-tetrachloro-1,1′,3,3′- tetraethylbenzimidazolyl-carbocyanine iodide (known as JC-1) in Panc1 upon MC3 treatment[40]. We found a constant decrease in the ratio of red to green fluorescence indicating a decline in the mitochondrial membrane potential in MC3-treated cells (Figure 3C). In addition, Bcl-xl, an anti-apoptotic protein, was concentration-dependently reduced after 24 h (Figure 3D). All above observations together provide a clear indication that MC3 is a small molecule efficiently inhibiting cell proliferation by targeting redox regulation in Panc-1 cells.

Reduced Trx can bind to ASK1 and consequently blocks ASK1 activity[22]. Therefore, we postulated that inhibition of TrxR by MC3 might keep Trx in the oxidized state, which could release free ASK1 and in turn activate the p38-MAPK pathway to promote cell apoptosis. Using immunoblotting we clearly detected concentration-dependent activation of phospho-p38-MAPK (T180/Y182) upon MC3 for 1 h (Figure 4A and Additional file1: Figure S1). We found a fast activation of p38-MAPK by MC3 (<30 min), which persisted over the test period of 24 h (Figure 4B and Additional file1: Figure S2). We also observed phosphorylation of p53 and PARP cleavage after 24 h (Figure 4B and Additional file1: Figure S2), hallmarks of cell apoptosis, in line with our above findings that MC3 induced cell cycle arrest and apoptosis. Analysis of phosphorylated p38-MAPK and HSP27 using quantitative phosphoprotein ELISA microarrays[41] confirmed time- (Figure 4C) and concentration-dependent (Figure 4D) activation of p38-MAPK signaling upon MC3[42], while phosphorylation of other signal transduction kinases like Erk1/2, Akt and GSK3ß was not significantly changed.

To further understand the profound role of the ASK1-p38-MAPK cascade in the cellular response to MC3, we isolated RNA after treatment with MC3 (5 μM) for 6 h and analyzed expression of p38-MAPK downstream genes including ATF2[43], Stat1[44] and TP53[45]. Expression of ATF2, Stat1 and TP53 was clearly induced upon MC3 treatment (Figure 4E), confirming the activation of p38-MAPK signaling by MC3 in Panc1 cells. We further tested if gold(I) NHC complexes could generally activate p38-MAPK in Panc1 and ASPC1 cells and found abundant phospho-p38-MAPK only in presence of MC3 or MC4 (Figure 4F and Additional file1: Figure S3), implicating a relationship between the anti-proliferative effect of gold(I) NHC complexes and p38-MAPK activity.

To confirm this mechanism on the molecular level, we next used a chemical p38-MAPK inhibitor (p38i, SB203580) in a co-treatment with 5 μM MC3 in Panc1 for 1 h. The results obtained confirmed that this combination could not only block MC3-induced activation of p38-MAPK-associated proteins, such as p38-MAPK and ATF2, but also activation of p53 (Figure 4G and Additional file1: Figure S4), suggesting that p38-MAPK signaling was directly involved in MC3-mediated cell death. Since MC3 is a TrxR inhibitor, we next investigated if adding a ROS scavenger could prohibit activation of the p38-MAPK cascade. Two anti-oxidants, N-actetyl-L-cysteine (NAC, 10 mM) and reduced glutathione (Glu, 5 mM), were used in co-incubation with MC3 (5 μM) for 1 h. Surprisingly, both antioxidants could not alter the level of phospho-p38-MAPK and -p53 (Figure 4G and Additional file1: Figure S4). Importantly, the higher level of ROS (Figure 4H) and the lower mitochondrial membrane potential (Figure 4I) compared to DMF control could still be detected at this early time point (1 h) of MC3 treatment in presence of NAC and Glu.

We further questioned if the p38 inhibitor (p38i) and anti-oxidant treatment are able to prevent cells from MC3-induced damage. Thus, we exposed Panc1 to MC3 in presence of ROS scavengers or p38i for 24 h and analyzed cell populations labeled with annexin v/PI as described before. The results clearly showed that ROS scavengers and p38i significantly counteracted the toxicity of MC3 (Figure 5A). Further immunoblotting data demonstrated that PARP cleavage induced by MC3 disappeared in co-treatment with p38i or ROS scavengers after 24 h (Figure 5B, Figure 5C and Additional file1: Figure S5). We also, found that ROS scavengers considerably reduced the activity of p38-MAPK (Figure 5C and Additional file1: Figure S5), as well as ROS accumulation (Figure 5D) and cell cycle arrest (Figure 5E) after 24 h, indicating a compensatory effect of ROS scavenging against MC3 in a long-term treatment. Interestingly, p38i attenuated cellular apoptosis caused by MC3, but enhanced ROS level (Figure 5D), suggesting that cells under severe oxidative stress might employ p38-MAPK to cope with excessive ROS.

Disruption of the interaction between ASK1 and Trx can activate ASK1-p38-MAPK signaling[22, 23]. To investigate whether activation of p38-MAPK was a result of releasing ASK1 from Trx caused by blocking the reductive recycling of Trx upon MC3, we used an ASK1 antibody to precipitate cellular ASK1 and examined the level of Trx binding to ASK1. After 4 h MC3 treatment the level of Trx in complex with ASK1 was clearly reduced, while ASK1 and Trx levels were not changed in the whole cell lysates (Figure 6A and Additional file1: Figure S6), demonstrating that MC3 suppressed ASK1 binding to Trx.

To further analyse the role of ASK1 in MC3-induced apoptosis, we genetically reduced the level of ASK1 using siRNA in HeLa cells and detected lack of ASK1 resulting in the reduction of p38-MAPK activity without changing total p38 levels in response to 5 μM of MC3 (Figure 6B and Additional file1: Figure S7). We used annexin v to label apoptotic cells after 24 h treatment with MC3 at 5 μM in HeLa cells transfected with increasing concentrations of ASK1 siRNA and observed a clear rescue with 20 nM of ASK1 siRNA (Figure 6C). We also transfected Panc1 with ASK1 siRNA (Figure 6D, right and Additional file1: Figure S8) and found a similar reduction of MC3 toxicity (Figure 6E). However, a higher amount of siRNA (50 nM) was required to achieve this effect (Figure 6D and Additional file1: Figure S8). Finally, we examined the influence of p38i, ROS scavengers and ASK1 siRNA on the activity of MC3 in ASPC1 cells. Comparably to the results in Panc1, viability of ASPC1 cells under MC3 treatment was markedly elevated in presence of one of the inhibitors, including ASK1 siRNA, ROS scavenger (Glu and NAC) and p38i (Figure 6F).

1,3-Diethylbenzimidazoliumiodide (MC1), chloro-(1,3-diethylbenzimidazol- 2-ylidene)gold(I) (MC2), [di-(1,3-diethylbenzylimidazol-2-ylidene)]gold(I) iodide (MC3), [triphenylphosphine-1,3-diethylbenzylimidazol-2-ylidene)]gold(I) iodide (MC4) were synthesized as described[9, 10]. Structures and purities were ascertained by ¹³C- and ¹H-NMR, and mass spectroscopy and elemental analyses. N-acetyl-L-cysteine (NAC), glutathione (Glu), gemcitabine (Gem) and p38 inhibitor (p38i, SB203580) were purchased by Sigma-Aldrich (Germany). p38 (cat: #9212), pp38 (T180/Y182, cat: #9215), PARP (cat: #9542), pp53 (S15, cat: #9284) and pATF2 (T71, cat: #5112) were obtained from cell signaling (NEB, Germany). ASK1 (cat: sc-7931) and Trx (cat: sc-20146) were from Santa Cruz (Germany).

HeLa, Miapaca, Panc1 and HFF cells were cultured in DMEM, and ASPC1 and BxPC3 in RPMI1640 containing 10% FBS and 1% Pen/Strp[32] under 5% CO₂ at 37°C in a humidified atmosphere and treated with compounds as indicated in the text. For depletion of ASK1, siRNA oligonuleotides were designed and synthesized by Riboxx (Riboxx GmbH, Radebeul, Germany), based on the gene with number NM_005923.3 and used for transfection of cells. Sequences are: ASK1a: AUUUAGAUGAAAUACAGG (guide) and CCUGUAUUUCAUCUAAAU (passenger); ASK1b: AUAUUAUCUACUCGCUG (guide) and GCAGCGAGUAGAUAAUAU (passenger); ASK1c: AUUUAAAAUACUCACU (guide) and GAGUGAGUAUUUCUAAAU (passenger); ASK1d: UUCAAUGAUUGUACAG (guide) and GCUGUACAAUCUUGAA (passenger). The two most efficient sequences were ASK1a and ASK1d. Riboxx®FECT reagent (Riboxx, Radebeul, Germany) was used to get maximal transfectional efficiency according to the manufacturer’s instructions. After 48 h transfection in 10% FBS, cells were treated with compounds as indicated in the text.

Effects on cell growth were determined using the SRB assay and MTT assay at an initial cell density of 5,000 cells/well in a 96-well plate as previously reported[9, 50].

Panc1 cells were plated in 12-well plate at a density of 200,000 cells/well and cultivated in standard condition for 24 h before cells were treated with compounds as described in the text. Cells were collected by trypsinization and centrifugation with 200 g (1500 rpm), and resuspended in FACS buffer (1% BSA in PBS) containing 5 μM dihydroethidium (DHE, Sigma-Aldrich). After 15 min incubation at room temperature in the dark cells were washed with FACS buffer, and immediately analyzed using a FACSCalibur flow-cytometer (Becton Dickinson) and CellQuest Pro (BD) analysis software (FACS analysis). Excitation and emission settings were 488 nm and 564–606 nm (FL2 filter), respectively. One representative result is presented from at least three-independent experiments showing comparable results.

Panc1 and ASPC1 cells were seeded at a density of 200,000 cells/well in a 12-well plate under standard conditions and treated with chemicals as indicated in the text. After 24 h incubation cells were trypsinized and collected by centrifugation. The annexin v/PI staining assay (BD, Germany) was used to analyze cell apoptosis according to the manufacturer’s protocol. Briefly, the cells were re-suspended in the binding buffer and stained with FITC-conjugated annexin v for 15 min, followed by PI staining for 5 min at room temperature. The cell suspension was immediately analyzed by FACS. An aliquot of the re-suspended cells was used to determine the cell number after treatment, using a hemacytometer and trypan blue. The relative amount of cells was calculated as the number of trypan blue negative cells divided by the number in the mock treatment.

Panc1 cells were seeded at a density of 200,000 cells/well in a 12-well plate under standard conditions. The cells were treated as designed, trypsinized and collected as mentioned above. The pellets were fixed in 70% Ethanol at least for 24 h, washed two times with ice-cold PBS and resuspended in 500 μL PBS. Cell suspensions were incubated with RNase A (50 μg/mL) for 30 min at 37°C and sequentially stained with PI (50 μg/mL) for 1 h and analyzed by FACS. At least two-independent experiments were performed.

Panc1 cells were cultivated under standard conditions (200,000 cells/well in a 12-well plate) and treated with various concentrations of MC3 as indicated for 24 h. Afterwards cells were trypsinized, collected, fixed in 4% paraformaldehyde for 10 min at 37°C and permeabilized in 90% methanol for 30 min on ice. Cells were blocked with 0.5% BSA solution in PBS, incubated with Alexa Flur ® 488-conjugated Bcl-xL (cat: #2767, cell signaling, NEB) for 1 h at RT, washed with PBS and analyzed by FACS. Excitation and emission settings were 488 nm and 564–606 nm (FL2 filter), respectively.

Panc1 cells (200,000 cells/well in a 12-well plate) were cultivated in DMEM with 10% FBS and treated with compounds as indicated. Cells were then stained with 2.5 μM JC-1 (5,5,6,6-tetrachloro-1,1,3,3- tetraethylbenzimidazolylcarbocyanineiodide, Sigma-Aldrich) for 30 min at 37°C, collected, and analyzed by FACS. Excitation and emission settings were 488 nm, 515–545 nm (FL1 channel) for JC monomers, and 564–606 nm (FL2 channel) for JC aggregates. Comparable results were obtained from at least in two-independent experiments.

Cell extracts were homogenized in urea-lysis buffer (1 mM EDTA, 0.5% Triton X-100, 5 mM NaF, 6 M Urea, 1 mM Na₃VO₄, 10 μg/mL Pepstatin, 100 μM PMSF and 3 μg/mL Aprotinin in PBS). Enhanced chemiluminescence (ECL) immunoblot analysis was performed. For this 40 μg of total protein was resolved on 10% SDS-PAGE gels and immunoblotted with specific antibodies. Primary antibodies were incubated at a 1:1,000 dilution in TBS (pH 7.5) with 0.1% Tween-20 and 5% BSA/nonfat milk with gentle agitation overnight at 4°C. The proper secondary antibodies were incubated in TBS (pH 7.5) with 5% BSA/nonfat milk and 0.1% Tween-20 at a 1:5,000 dilution for 1 h at room temperature. Comparable results were obtained from at least in two-independent experiments.

Immunoprecipitation was performed according to the manufacture’s protocol (Cell Signaling, Frankfurt, Germany). Briefly, cells were lysed in lysis buffern (100 mM TRIS/HCl pH 7.5, 150 mM NaCl, 1 mM MgCl2, 0.25% NP-40, 5 mM NaF, 1 mM Na₃VO₄, 10 μg/mL Pepstatin, 100 μM PMSF and 3 μg/mL Aprotinin) and incubated with Rabbit serum (DAKO) for 1 h on ice and Agarose A/G-protein bead (Santa Cruz) for 30 min. Supernatant was immunoprecipitated with ASK1 (1:100, Santa Cruz) overnight at 4°C. Immune complexes were precipitated with Agarose A/G-protein bead for 2 h at 4°C min and analyzed by immunoblotting. Clean-Blot IP Detection Reagent (Thermo) was used to reduce background caused by denatured and blotted IgGs according to manufacture’s protocol. Comparable results were obtained in at least two-independent experiments.

Quantitative reverse-transcription real-time-PCR was performed using a Light Cycler 480 (Roche, Germany) following a reference protocol from the manufacturer. Briefly, total RNA was isolated from cells using TRIzol (Qiagen, Germany). cDNA was generated by reverse-transcription using random primers and equivalent quantities of RNA. qPCR was performed using the SYBR Green PCR master mix (Roche, Germany) and the following primer pairs (MWG, Eurofins, Germany): ATF2 (5 s: CAGCGTTTTACCAACGAGGA; 3as: GAATCTTGTTGGTGTTGGGGTC); TP53 (5 s: CCTCACCATCATCACACTGGAAG; 3 s: CCTTTCTTGCGGAGATTCTCTTCC); Stat1 (5 s: GGAAAAGCAAGCGTAATCTTCAGG; 3as: GAATATTCCCCGACTGAGCC); and Actin (5 s: CTGACTACCTCATGAAGATCCTC; 3as: CATTGCCAATGGTGATGACCTG) as internal control. Results were obtained in three-independent experiments. Programs: preincubation, 1 cycle, 95°C, 5 min; quantification, 48 cycles, 95°C 10 sec, 60°C 10 sec, 72°C 20 sec; melting curve, 95°C 5 sec, 65°C 15 sec; cooling, 40°C.

Panc1 cells were plated at high density (200,000 cells/well) into 12-well plates and grown to confluence. The scratch was made by a sterile P-200 micropipette in the middle of each well. Cells were then washed three times with PBS and treated with MC3. Photographs of the same fields were taken at the beginning, after one day and at the end of the experiments (2 days).

Panc1 cells were plated into 12-well plate and grown to 70% confluence. Then cells were treated with MC3 for 24 h, harvested and re-plated into six-well plates with 2,000 cells/well. Medium were changed every 5 days. After 3 weeks, cells were stained with a staining solution containing 6% glutaraldehyde and 0.5% crystal violet for 1 h.

Phosphorylated proteins were quantified using sandwich ELISA microarrays. The microarrays are based on the ArrayStrip™ platform (Alere Technologies GmbH, Jena, Germany). A detailed description of the assay protocol and information on reagents for this assay have been previously reported[41].