Background
Malignant pleural mesothelioma, developed in the pleural cavity, is often associated with asbestos exposure in the patient history [
1,
2]. The prognosis remains poor and mesothelioma is resistant to a number of chemotherapeutic and molecular-targeting agents. A novel therapeutic strategy is thereby required to improve the prognosis. Mesothelioma has a characteristic genetic change. Previous analyses showed that approximately 80 % of mesothelioma specimens had a homozygous deletion in the INK4A/ARF locus which encoded
p14
ARF
and the
16
INK4A
genes, but the
p53 genotype was infrequently mutated [
3,
4]. The genetic defect leads to inactivation of the p53 pathways and may be linked with decreased susceptibility to anti-cancer agents.
Bisphosphonates are synthetic analogues of pyrophosphates and show high binding affinity to mineralized bone matrix [
5]. Previous reports showed that bisphosphonates produced cytotoxic effects on tumors such as breast and prostate cancer [
6,
7], and these cytotoxic actions were attributable to a number of mechanisms including apoptosis induction and anti-angiogenesis [
5,
8]. Zoledronic acid (ZOL), one of the third generation of bisphosphonates, inhibits the farnesyl pyrophosphate synthetase, a key enzyme in the mevalonate pathway, and depletes isoprenoid pools, which subsequently results in decreased prenylation of small guanine-nucleotide-binding regulatory proteins (small G proteins) [
5]. Consequently, ZOL prevented growth, adhesion or spreading, and invasion of cancer cells [
5,
9]. In our previous study, we demonstrated ZOL-mediated cytotoxic effects on mesothelioma cells [
10] and showed that ZOL treatments improved cytotoxicity of adenoviruses (Ad) expressing the
p53 gene on mesothelioma [
11]. Further analyses indicated that augmentation of p53 by ZOL was essential in combinatory effects of ZOL and DNA damaging drugs which included the first-line anti-cancer agents for mesothelioma [
11].
Replicating-competent Ad is a new strategy for cancer therapy. They can spread and destroy tumors without deleterious effects in normal tissues [
12,
13]. The replicable Ad continuously release the progenies from infected tumors and consequently circumvent low transduction efficacy. This replicable propensity enhances the cytotoxicity but host immunity can be inhibitory to the viral spreading. Ad lacking the E1B-55 kDa molecules (Ad-delE1B55) are replication-competent and were originally hypothesized to target only
p53-mutated or -null tumors due to the defect in p53-inactivating E1B-55 kDa protein [
14]. Nevertheless, Ad proteins that are synthesized during the replication also regulate p53 expression in infected cells at various levels even in an epigenetic manner [
15]. Subsequent studies in fact showed that Ad-delE1B55-induced cytotoxicity was not always related to the
p53 genotype [
16]. Moreover, our previous study showed that Ad-delE1B55 produced cytotoxicity on mesothelioma cells with the wild-type
p53 gene and achieved combinatory anti-tumor agents with the first-line chemotherapeutic agents [
17].
In the present study, we examined whether ZOL and Ad-delE1B55 could produce combinatory anti-tumor effects on human mesothelioma cells carrying the wild-type p53 gene. We speculated that both agents augmented endogenous p53 levels, which resulted in augmentation of the cytotoxicity. Furthermore, we analyzed a possible mechanism of the combination and investigated involvement of apoptotic pathways and viral replication in the anti-tumor effects.
Methods
Cells
Human mesothelioma cells, MSTO-211H, NCI-H28, NCI-H226, NCI-H2452 cells, all of which were purchased from American Type Culture Collection (Manassas, VA, USA), and EHMES-10 (provided from Dr. Hironobu Hamada, Hiroshima University, Japan) [
18] and were cultured with RPMI 1640 medium with 10 % fetal calf serum. HEK 293 and A549 cells, derived from American Type Culture Collection and Dr. Katsuyuki Hamada (Ehime University), respectively, were cultured with in Dulbecco’s Modified Eagle’s Medium containing 10 % fetal calf serum. NCI-H28, NCI-H2452 and EHMES-10 cells are defective of the
p14
ARF
and
p16
INK4A
genes, and MSTO-211H and NCI-H226 cells lack the
p14 and
p16 transcription (Additional file
1: Figure S1). Sequence analyses showed that all of them possessed the wild-type
p53 gene.
Ad preparation
Replication-competent Ad-delE1B55 (Accession number for Ad; M73260), in which the 55 kDa molecule-encoding E1B region (corresponding to 2019–3509 in M73260 sequences) is deleted, and replication-incompetent Ad expressing the ß-galactosidase (NM066611) (Ad-LacZ) or the green fluorescent protein gene (U55762) (Ad-GFP) powered by the cytomegalovirus promoter (KU317610), were prepared with an Adeno-X expression system (Takara, Shiga, Japan) and HEK293 cells. The numbers of virus particles (vp) per ml was estimated with the formula [absorbance at 260 nm of purified Ad in the presence of 0.1 % sodium dodecyl sulfate].
Cell cycle analysis and Giemsa staining
Cells treated with ZOL (Novartis, Basel, Switzerland) and/or either Ad-delE1B55 or Ad-LacZ as a control were fixed in ice-cold 100 % ethanol, incubated with RNase (50 μg/ml) and stained with propidium iodide (50 μg/ml). The staining profiles were analyzed with FACSCalibur and CellQuest software (BD Biosciences, CA, USA). We set up a gated area for data collection to remove doublets signals (Additional file
2: Figure S2). For Giemsa staining, cells treated with ZOL and/or Ad-delE1B55 were treated with colcemid (10 μg/ml). They were further incubated with hypotonic buffer and stained with Giemsa solution.
Viability test in vitro
Cells (5 × 103/well) in 96-well plates were cultured with ZOL and/or Ad-delE1B55 for 5 days. Cell viability was determined with a cell-counting WST kit (Wako, Osaka, Japan) and the relative viability was calculated based on the absorbance without any treatments. Viable cell numbers were also counted with the trypan blue dye exclusion test. Combinatory effects were examined with CalcuSyn software (Biosoft, Cambridge, UK). Combination index (CI) values at respective fractions affected (Fa) points, which showed relative suppression levels of the cell viability, were calculated based on the WST assay. CI < 1, CI = 1 and CI > 1 indicate synergistic, additive and antagonistic actions, respectively.
Western blot analysis
Cells were cultured with ZOL and/or either Ad-delE1B55 or Ad-LacZ, and the cell lysate was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis. The protein was transferred to a nylon filter and was hybridized with antibodies (Ab), against p53 (Thermo Fisher Scientific, Fremont, CA, USA), phosphorylated p53 at serine (Ser) 15, Bid (detecting truncated-Bid as well), caspase-3, cleaved caspase-3, caspase-8, cleaved caspase-8, caspase-9, cleaved caspase-9, poly (ADP-ribose) polymerase (PARP), Beclin-1, Atg5, LC3A/B (Cell Signaling, Danvers, MA, USA), cyclin A, cyclin E, type 2/5 Ad E1A (Santa Cruz Biotech, Dallas, TX, USA), phosphorylated H2 histone family member (H2AX) at Ser 139 (Biolegend, San Diego, CA, USA), hexon (Abcam, Cambridge, UK) and α-Tubulin (Thermo Fisher Scientific) as a loading control. The membranes were developed with the ECL system (GE Healthcare, Buckinghamshire, UK).
Infectivity with Ad-GFP and expression of Ad receptors
Cells were infected with Ad-GFP at several vp doses for 30 min and were then washed to remove Ad. Infected cells were cultured for 48 h and then analyzed for percentages of GFP-positive cells with FACSCalibur and CellQuest software. Cells of which fluorescence was greater than the brightest 5 % of uninfected cells were judged as positively stained. For detecting Ad receptors, cells were stained with anti-coxsackie adenovirus receptors (CAR) (Upstate, Charlottesville, VA, USA), integrin αvβ3 (Chemicon, Billerica, MA, USA), integrin αvβ5 (Abcam) Ab and the fluorescence intensity was analyzed with FACSCalibur and CellQuest software.
Virus production
Cells infected with Ad-delE1B55 were treated with or without ZOL. The cell lysates were examined for the cytotoxicity with A549 cells and the virus titers were calculated with the median tissue culture infectious dose (TCID50) method.
Discussion
In this study, we demonstrated that ZOL and Ad-delE1B55 induced growth suppression and the combination of both agents produced additive or synergistic inhibitory effects on mesothelioma cells. ZOL and Ad-delE1B55 induced differential cell cycle changes and the combinatory effects were achieved through enhanced apoptosis or increased viral replication. It is the first report to our knowledge that bisphosphonates and replication-competent Ad produced combinatory effects.
We noticed that MSTO-211H cells were sensitive to ZOL with the dye exclusion test, but NCI-H226 and NCI-H28 cells required a relatively high dose to induce cell growth suppression although these 3 kinds of cells were similarly susceptible to ZOL with the WST assay. We then tested MSTO-211H cells with a low ZOL concentration and NCI-H226 and NCI-H28 cells with a high ZOL doses in other experimental conditions. Cell cycle analyses showed that Ad-delE1B55 and ZOL produced differential effects which was characterized by hyperploidy and S-phase arrest. A precise mechanism of hyperploidy and S-phase arrest remain currently unknown, but the present study showed that Ad-infected cells induced enlarged nuclei followed by pyknotic configurations, which was compatible with hyperploidy and apoptotic cell death. In addition, augmented cyclin A and decreased cyclin E expression suggested induction of S-phase arrest, and phosphorylation of H2AX indicated activation of a cellular system detecting viral DNA increase. Hyperploidy could be due to a direct or an indirect consequence of accumulated viral DNA and an activated DNA damage sensor system thereafter. Furthermore, increased S-phase populations could be resulted from impaired cell cycle progression at S- and G2-phase and from failure of cells to shift into mitosis. Previous studies in fact reported that ZOL induced S-phase arrest and that the cell cycle changes were subjected to aberrant signals induced by mutated tumor suppressor genes [
20].
The majority of human mesothelioma possesses the wild type
p53 gene but lacks the
p14
ARF
and the
16
INK4A
genes, which subsequently leads to loss of the p53 functions and activation of the pRb pathways, respectively. ZOL activated endogenous p53 downstream pathways on mesothelioma even though the cell death did not depend on the p53 pathways in our recent study [
10]. Nevertheless, up-regulated p53 levels increased sensitivity to cisplatin, one of the first-line chemotherapeutic agents for mesothelioma [
11]. Consequently, ZOL facilitated DNA damage responses through the activated p53 downstream pathways despite p53 independence of ZOL-mediated cytotoxicity. On the other hand, Ad-delE1B55 also augmented endogenous p53 levels and subsequently activated the p53 functions in
p53 wild-type mesothelioma, which was evidenced by p53 phosphorylation, pRb dephosphorylation and cleavage of caspases [
17]. The present study showed that combination of ZOL and Ad-delE1B55 increased phosphorylated p53 and cleaved caspase-3 and we therefore presume it rational to use the agent together from the viewpoint of activation of the p53 pathways. Moreover, effects of ZOL on Ad replication also needs to be clarified since viral replication itself induced cytotoxicity to the infected cells.
The ZOL inhibits prenylation of small G proteins and consequently suppresses functions of small G proteins [
5,
8‐
10]. On the other hand, mesothelioma is often defective of the Hippo pathways due to mutation of the
NF/
Merlin gene [
21,
22]. Loss of NF/Merlin functions leads to dysregulation of multiple signal pathways, and up-regulated functions of small G proteins is one of the uncontrolled regulation. Ras and RhoA family proteins, one of the major small G proteins, are in fact activated in malignant mesothelioma and subsequently mitogen-activated protein kinases are also activated in most of the mesothelioma cells [
10,
23]. ZOL thereby blocks one of the dysfunctional Hippo pathways and combinatory effects of ZOL and Ad-delE1B55 can inhibit a possible cross-talk between NF/Merlin and the p53 pathways in mesothelioma with deletion of both INK4A/ARF and NF/Merlin regions.
We examined influence of ZOL on receptor expression levels and infectivity of Ad. Expression levels of major Ad receptors, CAR and integrin αvβ3 and αvβ5, were not affected by ZOL on MSTO-211H cells, whereas those of integrin but not CAR was down-regulated in NCI-H28 cells after ZOL treatments. The differential influence on integrin expression between the tested cells were not due to ZOL concentrations since ZOL at 10 μM also decreased integrin levels on NCI-H28 cells. Interestingly, Ad infectivity in NCI-H28 cells was rather enhanced after ZOL treatments despite the down-regulated integrin expression, indicating that the integrin as well as CAR did not play a crucial role in the infection in NCI-H28 cells. ZOL treatments induced alteration of actin fiber structures followed by morphological changes since small G proteins in particular Rho family proteins played a role in the organization of actin fibers in mesothelioma [
24]. Moreover, a previous study reported that ZOL specifically suppressed integrin αvβ3 but not α5β1 expression through down-regulate focal adhesion kinase [
25]. On the other hand, suppressed αvβ5 expression, demonstrated in the present study, had not yet been reported. Down-regulated expression of integrin molecules and enhanced Ad infectivity can be in part attributable to the ZOL-mediated morphological changes, but inhibition of small G proteins also influences several intracellular signal pathways and may augment infectivity, for example, due to facilitated viral release into cytoplasm after binding to the cellular receptors. Ad endocytosis via integrin molecules in fact requires activation of the lipid kinase phosphatidylinositol-3-OH kinase, which in turn augments signaling cascades of both Ras and Rho families [
26]. In addition, RhoA, Cdc42 and Rab6 were targets for ZOL-induced actions including cell death, cell cycle arrest and modification of actin fiber structure [
24]. These reports collectively suggested that combinatory cytotoxicity of ZOL and Ad-delE1B55 was linked with suppressed small G proteins functions. We also presume that increase infectivity could directly contribute to the enhanced production of Ad progenies since ZOL-treated MSTO-211H cells did not show the increased production but ZOL treatments augmented the production in NCI-211H cells. In addition, cells infected with Ad-delE1B55 during the G1 phase of cell cycle exhibited a reduced rate of viral late protein synthesis, and produced fewer viral progenies than cells infected during the S-phase [
27]. Cell cycle arrest at S-phase that was observed in ZOL-treated NCI-H28 cells can therefore be favorable for enhanced viral progeny production.
We hypothesized a possible mechanism involved in the combinatory effects. ZOL increased an expression level of p53 in MSTO-211H cells, which was induced by Ad-delE1B55 at 48 h after the infection, and subsequently activated apoptotic signal pathways which was evidenced by up-regulated cleavages of caspase-8, −9 and −3, and PARP even without Bid truncation. Moreover, the combination markedly increased sub-G1 populations in MSTO-211H cells but ZOL did not influence progeny production of Ad-delE1B55. These data indicated that the combinatory effects in MSTO-211H cells were attributable to ZOL-mediated augmented p53 downstream pathways. In contrast, ZOL in NCI-H28 cells augmented the viral infection and the replication, which was also evidenced by elevated or prolonged expression of E1A and subsequent hexon expression. A possible mechanism of the up-regulated Ad infection can be linked with inhibited functions of small G proteins, and the augmented infection was dependent on cells. We noticed ZOL-mediated actions were different among mesothelioma [
10,
20] and the present study showed a differential pattern of cell cycle progression and varied expression levels of cell surface molecules between MSTO-211H and NCI-28 cells. These differences are probably resulted from divergent effects of small G proteins in cellular functions among the cells tested.
Ad-delE1B55 has been approved in China as an anti-cancer agent for head and neck cancer [
28] and ZOL is used in a clinical setting in many countries. We has also started a phase I clinical trial with an intrapleural injection of ZOL for mesothelioma patients, which can maintains a local concentration of ZOL to achieve cytotoxicity [
29]. The present results under these circumstances will be one of the preclinical or relevant studies for a future clinical trial with replication-competent Ad and an inhibitor for small G proteins.
Abbreviations
Ab, antibody; Ad, adenoviruses; Ad-delE1B55, Ad lacking the E1B-55 kDa molecules; CAR, coxsackie adenovirus receptor; CI, combination index; Fa, fractions affected; GFP, green fluorescent protein; H2AX, H2A histone family member X; LacZ, ß-galactosidase; PARP, poly (ADP-ribose) polymerase; small G proteins, small guanine-nucleotide-binding regulatory proteins; TCID50, median tissue culture infectious dose; vp, virus particles; ZOL, zoledronic acid