Pancreatic cancer is one of the leading causes of cancer-related deaths in the western world. One-year survival rates are close to 18% with not more that 5% of patients living more than 5 years.1 Surgery still remains the best curative option especially if the tumour is detected early. However, this is not usually possible since symptoms develop gradually and detection is made at the late stages of the disease. Repeated attempts to tackle it using combinations of radiotherapy and chemotherapy have failed. 5-Fluorouracil (5-FU) has been the standard chemotherapeutic drug having response rates close to 20%, with Gemcitabine also showing promise.2, 3, 4 The aggressiveness of the disease and failure of current therapies to tackle it make it an appropriate model for the evaluation of alternative therapies.
Tetracyclines are a family of antibiotics with effectiveness against several diseases caused by microbial infection. They were discovered in 1953 and feature a 4-ring structure to which a variety of side chains are attached. Doxycycline (DC) is a member of the tetracycline family. It has good water and lipid solubility and in bacteria it is known to inhibit protein synthesis of pathogens.5 As such it has found applications in the treatment of a large number of diseases, including Lyme disease,6 Legionnaire's disease7 and malaria.8 Treatments are well tolerated by patients9 with small side-effects. In addition to its anti-microbial action, DC has also shown therapeutic potential against rheumatoid arthritis,10 abdominal aortic aneurysms,11, 12 malignant pleural effusions13 and mesenteric ischemia.14
DCs anti-tumour properties are also becoming known as in the case of prostate, breast and colon cancer.15, 16, 17, 18 Its role as a nonspecific matrix metalloproteinase and angiogenesis inhibitor is well established.19, 20, 21, 22, 23 Its pro-apoptotic effects are also known as in the case of T-lymphocytes.24
Apoptosis is a gene-directed form of cell death. It is important in a number of physiological processes like in normal development and maintenance of tissue homeostasis.25, 26 It also plays a role in tumour regulation because dysregulation of apoptosis may cause cancer cells to increase in numbers. Also chemotherapeutic agents may exert their anti-tumour action via activation of apoptosis in target cells.27, 28
In this study we demonstrate the cytotoxic effects of DC on 2 pancreatic cancer cell lines, the T3M4 and the GER cell lines. In addition we investigate the role of key molecules in apoptotic signalling pathways and report whether their regulation is necessary for DC-induced apoptosis to occur in pancreatic cancer cells.
Material and Methods
DC was purchased from SIGMA (Poole, UK) and dissolved in double-distilled water just before use. Antibodies for Bax, phospho-p38MAPK, phospho-MAPK Activated Protein Kinase 2 (MAPKAPK2), phospho-MAPK Kinase3/6 (MKK3/6) and the Caspase Cleavage Detection Kit were all purchased from Cell Signalling Technologies (Beverley, MA, USA). VDAC antibody was purchased from Abcam (Cambridge, UK). Actin antibody, Sulforhodamine B (SRB) protein dye and the caspase inhibitor Z-Val-Ala-Asp fluoromethyl ketone (Z.VAD-FMK) were purchased from Sigma. The pan-caspase inhibitor Boc-Asp(OMe)-CH2F (Boc.D-FMK) and the caspase9 inhibitor LEHD-CHO were purchased from Calbiochem (Nottingham, UK). The caspase8 inhibitor Z-I-E(OMe)-T-D(OMe)-FMK (Z.IETD-FMK) and the caspase10 inhibitor Z-A-E(OMe)-V-D(OMe)-FMK (Z.AEVD-FMK) were purchased from R&D systems (Abingdon, UK). The Chemicon colorimetric caspase3, caspase8, caspase9 and caspase10 activity kits containing caspase3, caspase8, caspase9 and caspase10 pNA-conjugated substrates and inhibitors were purchased from Chemicon International (Temecula, CA, USA). The Roche Cell Death Detection Elisa Kit was purchased from Roche (Basel, Switzerland). The Pierce Mitochondrial Isolation Kit was purchased from Perbio (Cramlington, UK).
The human pancreatic adenocarcinoma cell lines T3M4 and GER were obtained from Professor N. Lemoine (Cancer Research UK, London, UK) and Dr. A. Grant (St. George's, London, UK), respectively. Cells were maintained in a sub-confluent monolayer at 37°C in humidified atmosphere containing 5% CO2.They were propagated using RPMI-1640 medium supplemented with 10% Foetal Calf Serum (FCS), 2 mM L-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin and 25 μg/ml amphotericin B all purchased from Invitrogen (Paisley, UK). Screening for mycoplasma was carried out on a regular basis.
Cell plating and treatment
For experimental use, cells were plated at the optimal seeding density in 75 cm2 flasks or 24-well plates in medium supplemented with 2.5% FCS and were allowed to attach overnight. Then medium was changed and cells were treated with DC.
SRB assay for cell growth
The SRB assay gives an optical density output that corresponds to cell number. Removing medium from wells stopped treatments of cells in 24-well plates. To fix cells, 1 ml of 10% ice cold trichloroacetic acid in fresh medium was added to each well for 30 min. Then plates were washed with water and left to dry. Protein content was detected by addition of 200 μl of 0.4% SRB dye for 10 min to each well. Wells were washed with 1% acetic acid. SRB was solubilised with 1 ml 10 mM Trizma. Absorbance was detected using a microplate reader at 550 nm.
Cell homogenisation and western blotting
Briefly, control and treated cells were harvested from flasks, washed in PBS and lysed in SDS sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% w/v SDS, 10% glycerol, 50 mM DTT and 0.01% bromophenol blue) for 5 min on ice and sonicated for 10–15 sec. Samples were then heated for 5 min at 100°C, micro-centrifuged briefly and stored at −20°C. For the purpose of isolating mitochondrial fraction, the Pierce Mitochondrial Isolation Kit was used according to manufacturers' instructions. Cells were lysed with buffer provided in the Kit. Extract was centrifuged (700g for 5 min) and supernatant was collected and centrifuged again (12,000g for 10 min). Pellet representing the mitochondrial fraction was stored in SDS sample buffer for further examination. To check for purity of the fraction the sample was blotted for VDAC and actin. For the purpose of detecting Bcl2 family members cells were lysed in lysis buffer (50 mM Trizma pH 8.0, 150 mM NaCl, 0.1% v/v Triton X-100, 0.01 mg/ml aprotinin, 0.05 mg/ml PMSF) for 10 min on ice. Homogenates were then centrifuged at 10,000g for 10 min and supernatants were collected and stored at −70°C until further use.
Protein estimation was performed in cell extracts using the Biorad Dc protein assay (Biorad Laboratories, CA, USA). Equivalent amounts of protein were loaded on a 12 or 16% polyacrylamide-SDS gel and were electrophoresed for ∼1 hr. Proteins were then transferred from gel to membrane. After transfer was complete, primary antibodies diluted in blocking buffer (1.5% milk in PBS/T or Tris-buffered saline/Tween20 (TBS/T, pH 7.6) containing 5% BSA) were added on membranes and left incubating overnight at 4°C. After wash and secondary antibody incubation, signal from the protein of interest was visualised on X-ray films using ECL western blotting reagents (Amersham Biosciences, Little Chalfont, UK) and quantified by densitometry.
Enzyme-linked immunosorbent assay
The Roche cell death detection enzyme-linked immunosorbent assay (ELISA) kit was used to detect cleaved nucleosomes in the cytoplasm of cells according to manufacturers' instructions. Treated cells in 24-well plates were centrifuged at 200g and lysed for 30 min. Cell extract was again centrifuged at 200g for 10 min. About 20 μl from supernatant representing the cytosolic fraction were transferred from the 24-well plate into streptavidin-coated wells. Incubation with anti-histone-biotin and anti-DNA-Peroxidase (POD) antibody for 2 hr at room temperature was followed by incubation with the 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) substrate for 10 min. The reaction between POD and ABTS was photometrically determined using a microplate reader at 405 nm.
Colorimetric caspase activation assay
The assay for caspase3 activity is based on cleavage of the chromogenic substrate Acetyl-Asp-Glu-Val-Asp-p-Nitroaniline (Ac-DEVD-pNA) by caspase3, coupled with the use of the caspase3-specific inhibitor Ac-DEVD-CHO. Similarly, the assay for caspase8, caspase9 and caspase10 activities is based on cleavage of the chromogenic substrates Ac-IETD-pNA, LEHD-pNA and AEVD-pNA, respectively, coupled with the use of the caspase8 inhibitor IETD-CHO, caspase9 inhibitor LEHD-CHO and caspase10 inhibitor Z.AEVD-FMK. When cleaved, free pNA can be quantified using a microplate reader at 405 nm. Control and treated cells were harvested from flasks and lysed in cell lysis buffer (50 nmol/l, Tris-HCL, pH 7.5, 0.03% NP40, 1.0 mmol/l DTT) for 10 min. Lysates were then centrifuged at 12,000g for 12 min and protein concentration of the supernatant (cytosolic extract) was estimated using the Biorad assay. Equal amounts of protein extracts were loaded onto a 96-well microplate and incubated with reaction buffer, DTT, DMSO, the appropriate caspase substrates with or without the caspase inhibitors.
Statistical analysis of results
Statistical comparison of results using an unpaired student t test was performed. Statistical significance is indicated as ☆ p < 0.05. Standard error of the mean (SEM) is also shown.
Treatment with DC decreases pancreatic cancer cell numbers
To investigate whether DC could have a growth inhibitory effect, T3M4 and GER cells were treated with a range of concentrations of the compound for 1 and 2 days, respectively. SRB assay was used to determine cell numbers. Results show that cell lines were inhibited in a dose-dependent manner. Fifty percent reduction of T3M4 cells was evident after treatment with ∼5 and 10 μg/ml of DC for 1 day (Fig. 1a). A similar reduction in cell number was evident in GER cells after treatments with ∼10 and 20 μg/ml DC for 2 days (Fig. 1b).
DC induces caspase-dependent apoptosis in T3M4 and GER cells
To investigate whether the reduction in cell number is due to pancreatic cancer cells undergoing apoptosis, T3M4 and GER cells were treated with a range of concentrations of DC. To assess apoptosis the cell death detection ELISA assay was used. Results show a dose-dependent increase in apoptosis in both T3M4 (Fig. 2a) and GER (Fig. 2b) cells. Furthermore, to determine whether caspase-independent pathways contribute to apoptosis, T3M4 and GER cells were pretreated for 3 hr with 2 pan-caspase inhibitors, Z.VAD-FMK (50 μM) and Boc.D-FMK (25 μM) and then DC was added. Results show that apoptosis was induced in T3M4 (Fig. 3a) and GER (Fig. 3b) cells when treated with DC alone but not when co-treated with the compound together with the pan-caspase inhibitors.
Treatment of pancreatic cancer cells with DC results in activation of caspase3, caspase7, caspase8, caspase9 and caspase10
To determine which individual caspases are activated, T3M4 cells were treated with DC (10 μg/ml) for 6, 12 and 24 hr and GER cells were treated with DC (20 μg/ml) for 12, 24 and 48 hr. Individual caspase activation was assessed using either a colorimetric assay based on cleavage of pNA-conjugated substrates or immunoblotting with antibodies against activated caspases. In the case of colorimetric assays, caspase-specific inhibitors were also used to exclude any non-caspase-specific signal. Activation of caspase3 (Figs. 4a and 4h), caspase7 (Fig. 4h), caspase8 (Figs. 4b and 4e), caspase9 (Figs. 4c and 4f) and caspase10 (Figs. 4d and 4g) are evident after 12 hr in T3M4 cells (caspase7 at 24 hr only) and after 48 hr in GER cells (caspase8 at 12 hr as well (Fig. 4e)).
Death receptor pathways are likely to regulate induction of apoptosis in pancreatic cancer cells
The activation of caspase8 and caspase10 detected in previous experiments implies the activation of death receptor apoptotic pathways. To investigate whether FADD expression is increased, T3M4 and GER cells were treated with DC (10 μg/ml for 1 day and 20 μg/ml for 2 days, respectively). Using immunoblotting it was shown that DC treatments increased FADD protein levels in both cell lines (Figs. 5a and 5b). To investigate whether induction of apoptosis is directly associated with a death-receptor pathway, the caspase8-specific inhibitor Z.IETD-FMK and the caspase10-specific inhibitor Z.AEVD-FMK were used. Both pancreatic cancer cell lines were pretreated with each one of the inhibitors for 3 hr and then DC was added to the cells. Induction of apoptosis was measured using the cell death detection ELISA assay. In T3M4 cells treated for 1 day with DC (10 μg/ml) and Z.IETD-FMK (10 and 20 μM) or with DC (10 μg/ml) and Z.AEVD-FMK (20 μM) (Figs. 6a and 6b), a reduction of apoptosis at basal levels was observed. Similarly in GER cells a statistically significant reduction of apoptosis was detected after treatment with DC (20 μg/ml) and Z.IETD-FMK (10 μM) or DC (20 μg/ml) and Z.AEVD-FMK (20 μM) for 2 days relative to treatment with DC (20 μg/ml) only (Figs. 6c and 6d).
DC increases Bax protein levels in mitochondria but caspase9 does not affect induction of apoptosis
To determine whether DC affects the regulation of members of the Bcl2 family, GER cells were treated with the compound (10 and 20 μg/ml) for 2 days. Whole cell lysates and mitochondria were isolated as described in the methods section and western blotting was used to determine protein levels of Bax. Results show that DC increases Bax protein levels in whole cell lysates (Fig. 7a) and in the mitochondria (Fig. 7b) of GER cells.
This result coupled with the increase in caspase9 activity detected in previous experiments led us to investigate the effect of a caspase9-specific inhibitor in pancreatic cancer cells. Using an ELISA assay we demonstrate that induction of apoptosis in T3M4 and GER cells treated with DC (10 μg/ml for 1 day and 20 μg/ml for 2 days, respectively) is similar to cells treated with DC and the caspase9 inhibitor LEHD.CHO (20 μM) (Figs. 8a and 8b). To further understand the caspase9 activation process we investigated the residue at which caspase9 is cleaved. Using westerns we detect caspase9 cleaved on residue Asp330 in T3M4 and GER cells (Fig. 8c).
DC-induced apoptosis is associated with increased phosphorylation of members of the p38MAPK stress-activated pathway
P38MAPK is a kinase involved in mitogenic cell signalling. To evaluate its significance, T3M4 and GER cells were treated with DC (10 μg/ml for 1 day and 20 μg/ml for 2 days, respectively). Phosphorylation of p38MAPK was investigated using antibodies against the phosphorylated and total form of the kinase. In addition, phosphorylation of upstream activators, such as MKK3/6, and downstream targets, such as MAPKAPK2, of this kinase were also investigated. Results show that phospho-p38MAPK was increased after treatment with DC (20 μg/ml) in GER cells denoting activation of this kinase (Fig. 9a). Also phospho-MAPKAPK2 (Figs. 9b and 9d) and phospho-MKK3/6 (Figs. 9c and 9e) were increased.
To investigate whether activated p38MAPK is directly involved in induction of apoptosis, GER cells were pretreated with the 2 p38MAPK-specific inhibitors, SB203580 (2 μM) and SB202190 (2 μM) for 3 hr. Then DC (10 and 20 μg/ml) was added and cells were treated for 2 days. Apoptosis was detected using the ELISA assay. Results show that apoptosis due to DC was not affected by co-treatments with any of the 2 p38MAPK-inhibitors (Fig. 10).
DC has been effective against a number of tumour cell lines such as those from prostate, breast and colon cancer.15, 16, 17, 18 However, its effects on pancreatic cancer cells have not yet been investigated. The aim of this study is to establish whether DC could have an anti-tumour effect on them and to investigate its mechanism of action.
Two in vitro models of exocrine pancreatic adenocarcinoma, the cell lines T3M4 and GER were used. T3M4 cells represent tumour obtained from human metastatic lymph nodes and they are a CEA-producing cell line that contains an activated c-k-Ras oncogene.29, 30 Their doubling time is ∼24 hr. GER is another CEA-producing cell line that represents cells from human pancreatic exocrine adenocarcinoma with doubling time of about 36 hr.31, 32
In this study we show that treatment of pancreatic cancer cells with DC results in a dose-dependent reduction in cell number that is associated with increased levels of caspase-dependent apoptosis evident in T3M4 cells from day 1 and in GER cells from day 2. Importantly, the concentrations of DC used to induce apoptosis in the pancreatic cancer cells are close to the already achievable serum concentrations of the compound (∼5 μg/ml).33, 17 The fact that for potential clinical use in pancreatic cancer patients a more aggressive chemotherapeutic treatment regime may be utilised, coupled with minimal side-effects means that this antibiotic could have good therapeutic potential.
Advances in understanding the molecular pathogenesis of several diseases lead us towards an age of customised treatments for individuals, something especially true for difficult-to-treat diseases like pancreatic cancer. In this context it is important to characterize the DC-induced apoptotic process at protein level.
Using a colorimetric-based assay we demonstrate that in T3M4 cells caspase8 and caspase10 are activated after 12 hr treatment with DC and after 48 hr treatment in GER cells. Another interesting finding is the activation of caspase8 in GER cells after 12 hr of treatment with DC, the functional consequence of which is not certain. These results led us to investigate expression of FADD, which is frequently up-regulated in death receptor pro-apoptotic pathways.34 Results show that FADD was increased in both cell lines after treatment with DC. T3M4 and GER cells were also co-treated with DC and either the caspase8-specific inhibitor Z.IETD-FMK or the caspase10-specific inhibitor Z.AEVD-FMK. Then induction of apoptosis was assessed. Using this procedure we show that inhibition of caspase8 or caspase10 in T3M4 cells blocks apoptosis whereas in GER cells it significantly reduces it. It is interesting that inhibition of only one of these caspases is sufficient to abrogate apoptosis in T3M4 cells. These results imply that apoptosis induced in pancreatic cancer cells by DC is strongly associated with FADD, caspase8 and caspase10-related pathways, which is also in line with findings in other cell types such as in Jurkat T Lymphocytes.24
In addition to the colorimetric-based assay to assess caspase activation, we also used immunoblotting with antibodies against activated caspases. These antibodies do not cross-react with the respective pro-caspase zymogens or with cleaved caspases at residues other than the ones quoted. Using these methods we demonstrate that caspase3 and caspase9 are activated after 12 hr treatment in T3M4 cells and after 48 hr treatment in GER cells. Caspase3 and caspase9 activities show higher enhancement ratios relative to caspase8 and caspase10 activity especially in T3M4 cells but this is, at least partly, due to low caspase3 and caspase9 activity detected in control cells. We also investigated additional components of the mitochondrial-associated apoptotic pathway. Treatment of pancreatic cancer cells with DC increases expression of Bax in whole cell lysates and in the mitochondrial fraction of GER cells but not in that of T3M4 cells. It is an interesting observation that in GER cells complete abrogation of apoptosis is not succeeded when caspase8 or caspase10 is inhibited. Other members of the Bcl2 family such as Bak, Bclxl, Mcl1, Bim, tBid and phospho-bad failed to exhibit significant changes in their protein levels in whole cell lysates (data not shown). Furthermore, treatment of both cell lines with a caspase9 inhibitor and DC failed to show reduced apoptotic effect relative to treatments with DC alone. To explain this result, we investigated caspase9 activation using immunoblotting. Cleaved Caspase9 is detected using an antibody raised against the enzyme cleaved at residue D330 in both cell lines. This cleavage site may be a target of caspase3.35, 36, 37 The use of an antibody specific for cleaved caspase9 at residue D315, which is the one primarily targeted by apoptosome-related activity failed to show a signal in the western blot (data not shown). Hence, from these results alone we cannot conclude activation of a mitochondrial-dependent pathway.
To investigate the role of important Mitogen Activated Protein Kinases (MAPK)38 in DC treatments of T3M4 and GER cells, their phosphorylation status was assessed. No significant change was evident in p44/42MAPK and Stress Activated Protein Kinase (SAPK). However, there was an increase in the phosphorylation of p38MAPK in GER cells and in phospho-MKK3/6 and phospho-MAPKAPK2 in both T3M4 and GER cells. However, the use of 2 p38MAPK-specific inhibitors failed to reduce apoptosis in DC-treated cells, making it unlikely that this pathway is directly involved in the apoptotic process.
In conclusion this study highlights the pro-apoptotic effects of DC in pancreatic cancer cells. DC induces caspase-dependent apoptosis that is associated with activation of several individual caspases. Caspase8 and caspase10 pathways are, at least partly, responsible for induction of apoptosis in the 2 pancreatic cancer cell lines while FADD is also up-regulated in both cell lines. Together with the known properties of DC as an inhibitor of angiogenesis and MMP activity, these results strongly support the case for including DC in clinical trials of pancreatic cancer.