The majority of malignant mesothelioma possesses the wild-type p53 gene with a homologous deletion of the INK4A/ARF locus containing the p14ARF and the p16INK4A genes. We examined whether forced expression of p53 inhibited growth of mesothelioma cells and produced anti-tumor effects by a combination of cisplatin (CDDP) or pemetrexed (PEM), the first-line drugs for mesothelioma treatments. Transduction of mesothelioma cells with adenoviruses bearing the p53 gene (Ad-p53) induced phosphorylation of p53, upregulated Mdm2 and p21 expression levels and decreased phosphorylation of pRb. The transduction generated cleavage of caspase-8 and -3, but not caspase-9. Cell cycle analysis showed increased G0/G1- or G2/M-phase populations and subsequently sub-G1 fractions, depending on cell types and Ad-p53 doses. Transduction with Ad-p53 suppressed viability of mesothelioma cells and augmented the growth inhibition by CDDP or PEM mostly in a synergistic manner. Intrapleural injection of Ad-p53 and systemic administration of CDDP produced anti-tumor effects in an orthotopic animal model. These data collectively suggest that Ad-p53 is a possible agent for mesothelioma in combination with the first-line chemotherapeutics.
Malignant mesothelioma, most of which are linked with asbestos exposure, remains intractable despite recent treatment modalities.1 The latent period is long and no preventive procedure is currently available. The patient numbers will continuously increase in the next decades in many industrialized countries as well as in newly developing countries.2 Extrapleural pneumonectomy is a therapeutic option applicable to an early-staged case, but the recurrence is common even after the radical surgery. Mesothelioma is resistant to radiotherapy, which is mainly used for a palliative purpose. Chemotherapy is therefore a primary treatment in most of the cases and a combination of cisplatin (CDDP) and pemetrexed (PEM) is currently the first-line chemotherapy.3 Nevertheless, a mean survival rate with the CDDP plus PEM regimen is 12.1 months3 and the prognosis remains poor. A novel therapeutic strategy is thereby required to improve the prognosis and quality of the patients.
Genetic characterization of mesothelioma cells and the clinical specimens has shown that the majority is defective in the INK4A/ARF locus containing the p14ARF and the p16INK4A genes, but possesses the wild-type p53 gene.4 Deletion of p16 increases cyclin-dependent kinase 4/6 activities and subsequently phosphorylates pRb, which induces cell cycle progression. In contrast, p14 deficiency augments Mdm2 activities and consequently downregulates p53 expression, which renders mesothelioma cells resistant to apoptotic pathways. Reintroduction of the defective genes can be a possible treatment modality by activating the suppressor gene functions. Forced expression of p16 or p14 with adenovirus vectors (Ad-p16, Ad-p14) in fact inhibited proliferation of mesothelioma cells through dephosphorylation of pRb and restoration of the p53 pathways.5, 6, 7, 8 Expressed p14 however did not fully activate the p53-mediated signal transduction,9 and Ad-p14-mediated anti-tumor effects required intact p53 downstream pathways.6 Reactivation of p53 by Ad-p53 can therefore be a more direct way to reconstitute all the p53 signal pathways and can dephosphorylate pRb through p53-induced upregulation of p21. On the other hand, the Ad-p53-mediated anti-tumor effects are difficult to be achieved in wild-type p53-bearing tumors compared with those with p53 mutations.10 None of the studies in fact has demonstrated Ad-p53-mediated anti-tumor effects in mesothelioma cells, except a short report that showed the suppressed tumor growth using one mesothelioma cell line, but did not investigate the inhibitory mechanism.11 It remains unknown as to how overexpressed p53 induces growth inhibition to p14/p16-defective mesothelioma cells and whether Ad-p53 can produce anti-tumor effects in an orthotopic animal model. Moreover, combinatory effects of Ad-p53 and anticancer agents, in particular CDDP or PEM, have not been investigated with mesothelioma cells. In this study, we examined a possible Ad-p53-mediated cytotoxicity in five kinds of p14/p16-defective and p53-wild-type mesothelioma cells and demonstrated that upregulated p53 itself induced the p53 phosphorylation, activated p53 downstream pathways and produced combinatory anti-tumor effect with the anticancer agents.
Materials and methods
Cells and mice
Human mesothelioma cells, NCI-H2452, NCI-H2052, NCI-H226, NCI-H28 and MSTO-211H, were obtained from ATCC (Manassas, VA). All of them are defective of the p14 and p16 genes, but possess the wild-type p53 gene,6 which were also confirmed by our sequencing data. BALB/c nu/nu mice (6-week-old females) were purchased from Japan SLC (Hamamatsu, Japan).
Replication-incompetent type 5 Ad expressing wild-type p53 (Ad-p53) or β-galactosidase gene (Ad-LacZ), in which the cytomegalovirus promoter activated the transgene, were prepared with an Adeno-X expression vector system (Takara, Shiga, Japan) and were purified with Adeno-X virus purification kit (BD Biosciences, Palo Alto, CA). The number 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 × 1.1 × 1012).
Cell cycle analysis
Ad-infected cells were fixed in ice-cold 70% ethanol, incubated with RNase (50 μg ml−1) and stained with propidium iodide (PI) (50 μg ml−1). The staining profiles were analyzed with the FACScan and CellQuest software (BD Biosciences, San Jose, CA).
Viability test in vitro
Cells (1 × 103 per well) were seeded in 96-well plates and were infected with Ad at different doses. In a combinatory treatment, cells were treated with various concentrations of CDDP or PEM and then infected with Ad-p53 or Ad-LacZ. Cell viability was determined with a cell-counting WST kit (Wako, Osaka, Japan) on day 5 and the relative viability was calculated based on the absorbance without any treatments. IC50 (half maximal inhibitory concentration) was estimated with a program related to nonlinear least squares in FORTRAN77 Version 3.5 developed by Dr Yamaoka (Kyoto University, Kyoto, Japan; http://www.pharm.kyoto-u.ac.jp/byoyaku/Kinetics/program/manual.htm). Combinatory effects of Ad and an anticancer agent were assessed by calculating combination index (CI) values with the CalcuSyn software (Biosoft, Cambridge, UK). CI<1, CI=1 and CI>1 indicate synergistic, additive and antagonistic actions, respectively.
Western blot analysis
Cells were infected with Ad-LacZ or Ad-p53, 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 against p53 (Lab Vision, Fremont, CA), phosphorylated p53 at serine (Ser) residues 15 or at Ser 46, Fas, TRAIL receptor (DR5) (Cell Signaling, Danvers, MA), Mdm2 (Santa Cruz Biotech, Santa Cruz, CA), p21 (Santa Cruz Biotech), p27 (BD Biosciences), pRb, phosphorylated pRb at Ser 795, Bid, caspase-3, cleaved caspase-3, caspase-8, cleaved caspase-8, caspase-9, cleaved caspase-9 (Cell Signaling) or actin (Sigma-Aldrich, St Louis, MO). The membranes were developed with the ECL system (GE Healthcare, Buckinghamshire, UK).
MSTO-211H cells (1 × 106) were injected into the intrapleural cavity of BALB/c nu/nu mice. The mice were treated with intrapleural injection of Ad alone or with a combination of intraperitoneal injection of CDDP on day 3, and the tumor weights were measured on day 28. Phosphate buffered-saline and Ad-LacZ were used as a control. The animal experiments were approved by the animal experiment and welfare committee at Chiba University and were performed according to the guideline on animal experiments.
Phosphorylated p53 and activated caspases in CDDP-treated mesothelioma cells
We examined whether CDDP, a DNA-damaging agent, augmented p53 expression in p14/p16-defective and p53-wild-type MSTO-211H cells (Figure 1). The treatment upregulated p53 expression and induced the phosphorylation at Ser 15 and 46, both of which are markers for p53-mediated cell cycle arrest and apoptosis. Moreover, cleavage of caspase-8 and -9 together with the decreased pro-caspase forms, which were involved in the extrinsic death receptor- and the intrinsic mitochondria-linked apoptosis, respectively, were detected in the CDDP-treated cells. We noted that Bid expression was downregulated, but truncated Bid, which contributes to the linkage between the death receptor- and the mitochondria-mediated apoptosis, was not induced (data not shown). Caspase-3, an enzyme mapped in a common apoptotic pathway after caspase-8 and -9 activations, was also cleaved. These data collectively suggest that p14/p16-defective mesothelioma maintains p53 downstream functions and that death receptor-mediated and mitochondrial-mediated pathways were independently activated upon CDDP treatments.
Decreased viability of mesothelioma treated with Ad-p53
We examined whether transduction of p53 could suppress viability of five kinds of human mesothelioma cells with p14/p16-defective and p53-wild-type genes. Relative viabilities of mesothelioma cells were examined with different doses of Ad-p53 or Ad-LacZ as a control (Figure 2). Ad-p53 but not Ad-LacZ treatment decreased viability of all the mesothelioma cells with a dose-dependent manner, except NCI-H2052 cells. The IC50 values suggested that NCI-H28 cells were the most sensitive to Ad-p53 and the others, NCI-H226, NCI-H2452 and MSTO-211H cells, became susceptible when infected with increased virus numbers. Insensitivity of NCI-H2052 cells to Ad-p53 could be partly due to the low expression level of the major Ad receptor, coxsackie adenovirus receptor (Supplementary Figure 1). Ad infectivity to NCI-H2025 cells was in fact poor compared with that to other cells (Supplementary Figure 2), but the differential susceptibility was also attributable to other factors than the Ad infectivity, such as the vulnerability to Ad-induced cell death. In addition, greater amounts of Ad-p53 overcame the insensitivity and induced the growth suppression in NCI-H2052 cells (data not shown).
Cell cycle changes induced by Ad-p53
We examined cell cycle changes induced by Ad-p53 or Ad-LacZ at two different doses (1 × 104 and 3 × 104 vp per cell) (Table 1). Transduction of NCI-H2452 cells with Ad-p53 increased G0/G1-phase populations with reduced S-phase fractions at 24 h and the infection at 3 × 104 vp per cell increased sub-G1 populations at 48 h. The cell cycle changes, G0/G1 arrest followed by increased sub-G1 populations, were more significant in MSTO-211H cells than NCI-H2452 cells, although the IC50 value of MSTO-211H cells was greater than that of NCI-H2452 cells. Ad-p53-infected NCI-H28 cells showed increased G2/M-phase populations with decreased S-phase populations at 24 h and augmented sub-G1 populations thereafter. Sub-G1 fractions increased significantly in NCI-H28 cells compared with NCI-H2452 and MSTO-211H cells, showing their high susceptibility to Ad-p53-mediated apoptosis. Interestingly, Ad-p53-infected NCI-H28 cells did not show G0/G1 arrest in contrast to NCI-H2452 and MSTO-211H cells. Moreover, G2/M fractions increased in NCI-H28 cells with Ad-p53 treatment at 24 h and with Ad-LacZ treatment at 48 h, suggesting that increased G2/M fractions can be at least partly due to Ad infections. These cell cycle changes caused by Ad-p53 were thus subjected to cell types, and were dependent on Ad doses and treated periods as well.
Activation of p53-mediated pathways
We further examined molecular events of Ad-p53-mediated effects in MSTO-211H and NCI-H28 cells with western blot analyses (Figure 3a). Transduction with Ad-p53 increased p53 levels and the phosphorylation was readily detected at Ser 15 and then at Ser 46 residues. Ad-LacZ infection did not upregulate endogenous p53 expression or induce the phosphorylation. Expression levels of Mdm2, 90-kDa full-length molecules and 60-kDa cleaved moieties that represented a p53-bound form12 increased by transduction with Ad-p53, but not with Ad-LacZ. Anti-pRb antibody detected a change of pRb migration from the high to the low molecular weight accompanied by the Ad-p53 transduction, which corresponds to a shift from the phosphorylated to the dephosphorylated state. The decreased phosphorylation was confirmed by anti-phosphorylated pRb antibody, which detected the reduced intensity. Expression levels of p21 increased with Ad-p53 transduction in both cells, but p27 expressions altered differently. The p27 expression was upregulated in Ad-53-infected MSTO-211H cells, but the level in NCI-H28 cells remain unchanged or even decreased at 36 h after the Ad-p53 infection. These data demonstrated that Ad-p53 infection activated p53-mediated pathways, but subsequent induction of cyclin-dependent kinase inhibitors, p21 and p27, was not coordinated.
We then examined activation of caspases in Ad-p53-infected MSTO-211H and NCI-H28 cells (Figure 3b). Transduction with Ad-p53 but not Ad-LacZ induced cleavage of caspase-3 and -8, whereas Ad-p53 transduction did not stimulate caspase-9 cleavage. Bid expression levels were basically unchanged and truncated Bid was not detected in both cells (data not shown). Expressions of Bax in MSTO-211H cells remained unchanged and those in NCI-H28 cells were barely detected (data not shown). We also examined expression levels of the molecules in upstream pathways of the extrinsic apoptosis system (Figure 3c). Expression levels of Fas were augmented in both cells and those of TRAIL receptors (DR5) were also upregulated in NCI-H28 cells. FADD expressions were however almost unchanged (data not shown). These data collectively suggest that Ad-p53 transduction activated the extrinsic death receptor-mediated pathways through caspase-8 and that the intrinsic mitochondrial-mediated apoptosis was less likely to be involved.
Combination of Ad-p53 and anticancer agents
We examined possible combinatory anti-tumor effects of Ad-p53 and CDDP or PEM with the WST assay (Figure 4). Mesothelioma cells were treated with different concentrations of CDDP or PEM and then were transduced with Ad-p53. Transduction with Ad-p53 enhanced CDDP-induced anti-tumor effects and the CI values showed that the Ad-p53 and CDDP were synergistic in the inhibitory activity in most of the cells tested, except NCI-H2452 cells, which showed minimal synergism between 0.3 and 0.6 at fraction affected (Figures 4a and b). Combination of Ad-p53 and PEM also produced synergistic effects except in NCI-H28 cells, which showed slight inhibitory effects at a lower fraction affected range (Figures 4c and d). Notably, NCI-H2052 cells, which were resistant to Ad-p53-mediated growth inhibitory effects (Figure 2), came to be more sensitive to CDDP and to PEM when treated with Ad-p53 than with Ad-LacZ. We also examined cell cycle changes induced by the combination of Ad-p53 and PEM in NCI-H2052 and MSTO-211H cells (Table 2). Ad-p53 transduction alone induced minimal cell cycle changes at the virus doses used. PEM treatments increased S-phase fractions in NCI-H2052 cells and both S-phase and sub-G1 populations in MSTO-211H cells as PEM is a DNA synthesis inhibitor. Concomitant Ad-p53 treatments increased sub-G1 populations in PEM-treated cells, whereas Ad-LacZ did not influence the cell cycles. These data showed that Ad-p53 transduction enhanced cytotoxicity of the anticancer agents.
In vivo anti-tumor effects by Ad-p53 with CDDP
We investigated the anti-tumor effects of Ad-p53 in an orthotopic animal model, in which nude mice were injected with MSTO-211H cells in the pleural cavity and received intrapleural administration of Ad-p53 or Ad-LacZ as a control on day 3. Tumor weights on day 28 showed that Ad-p53 produced anti-tumor effects in a dose-dependent manner, whereas Ad-LacZ administration did not influence the weight (Figure 5a). We also examined combinatory effects with Ad-p53 and CDDP, both of which were administered on day 3 (Figure 5b). Administration of Ad-p53 and CDDP, respectively, inhibited the tumor growth and the combination decreased the tumor weight greater than Ad-p53 or CDDP treatment alone. A combinatory use of Ad-p53 and PEM was not examined as an optimal PEM dose to produce anti-tumor effects in mice was difficult to be determined.
We examined Ad-p53-mediated anti-tumor effects in five kinds of mesothelioma cells and firstly demonstrated to our knowledge combinatory effects of Ad-p53 and CDDP or PEM in p53-wild-type mesothelioma. We showed activation of p53 pathways with CDDP treatments and demonstrated that p53-mediated apoptosis was maintained in p14/p16-defective mesothelioma cells, which were evidenced by upregulated p53 expression, the phosphorylation at Ser 15 and 46 residues and cleavage of caspases. Ad-p53 likewise induced p53 expression and the phosphorylation at Ser 15 and thereafter at Ser 46, both of which are linked with cell cycle arrest and apoptosis, respectively.13, 14 It is however currently unknown about a precise mechanism of phosphorylation of p53 induced under no apparent DNA damage as Ad infection itself, as demonstrated with Ad-LacZ, did not increase endogenous p53 expression or the phosphorylation. Overexpressed p53 was not always associated with the phosphorylation,9 but increased p53 expression can augment activities of p53-phosphorylating kinases and consequently result in p53 activation.
Upon p53 induction and phosphorylation, Mdm2 and p21 expression, both of which are induced by p53-mediated transcriptional activation, were upregulated. Dephosphorylation of pRb can be induced by the upregulated p21 through inhibiting cyclin-dependent kinase activities. Mdm2 can be overexpressed in p14-defective cells because p14 suppresses the expression, but previous studies as well as this study showed no such enhancement of Mdm2 expression in p14/p16-defective mesothelioma cells.15 Full-length 90 kDa Mdm2 molecules were cleaved into 60 kDa molecules, which was induced by the binding to p53, and the cleavage level was linked with p53 induction. These data collectively demonstrated that Ad-p53 activated the downstream pathways and blocked pRb-mediated cell cycle progression, suggesting that Ad-p53 restored loss of tumor suppressor functions in p14/p16-defective mesothelioma. Expression of p27 in NCI-H28 cells, in contrast to that in MSTO-211H cells, did not coordinately increase responding to p53 activation as found in p21 expression. The dissociated expressional changes between the two cyclin-dependent kinase inhibitors might be linked with differential cell cycle alterations, increased G0/G1-phase population in MSTO-211H cells and no G0/G1 arrest in NCI-H28 cells (discussed later).
Apoptosis induction was accompanied by Ad-p53 transduction. Increased cleaved caspase-8, but not caspase-9, suggested activation of the extrinsic receptor-mediated apoptosis pathways without involvement of the mitochondria-mediated pathways. Moreover, expressions of Bax remained unchanged and truncated Bid, which is involved in a cross-talk between the extrinsic and the intrinsic apoptosis pathways, was not detected. We however detected upregulated death receptors, Fas and TRAIL receptor molecules, as reported previously.16 These biochemical data suggested that Ad-p53 activated primarily the extrinsic pathway in contrast to CDDP treatments that activated the both pathways. Shinoura et al.17 demonstrated enhanced apoptosis by a combination of Ad-p53 and forced expressions of Apaf-1 and caspase-9, suggesting less involvement of the intrinsic pathways by Ad-p53.
Cell cycle analyses showed differential effects of Ad-p53 to respective mesothelioma cells. Ad-p53 induced G0/G1 arrest in NCI-H2452 cells and a small cell population was subjected to sub-G1 fractions. The similar effects were observed in MSTO-211H cells, but increase of sub-G1 fractions was significant. In contrast, NCI-H28 cells showed increased G2/M populations at 24 h and sub-G1-phase fractions thereafter. Ad infection itself increased G2/M populations in NCI-H28 cells as found in Ad-LacZ-treated cells at 48 h. The NCI-H28 cell cycle profiles thus suggested that p53 accelerated the Ad-mediated G2/M arrest. It remains unclear as to the mechanism how Ad infection itself increased G2/M phase and whether expressed p53 differentially arrested cell cycle progression at either the G0/G1 or G2/M phase. The present biochemical data supported p53-induced cell cycle arrest at the G0/G1 phase. Phosphorylated p53 at Ser 15 followed by at Ser 46 matched to the sequential change of cell cycle arrest followed by apoptosis. Increased p21 and p27 expressions and dephosphorylated pRb in MSTO-211H cells favored G0/G1 arrest, but unchanged or decreased p27 expression in NCI-H28 cells may counteract cell cycle arrest at the G0/G1 phase. Cell viability data tested with the WST assay were not completely consistent with cell cycle analysis data in terms of Ad-p53 susceptibility. Cell cycle distributions showed that NCI-H2452 cells were resistant to Ad-p53-mediated apoptosis and MSTO-211H cells were sensitive, whereas the cell viability data indicated that NCI-H2452 cells were more susceptible than MSTO-211H cells. The differential results could be due to distinctive assays systems. Cell cycle analyses showed relative distributions of each cell cycle population, but the WST assay measured a metabolic activity of the whole cell populations. The p53 family members such as p73 and p63 may also differentially influence susceptibility to apoptosis, although expression levels of the p53 family protein were not well studied in mesothelioma cells.
Anti-tumor effects mediated by Ad-p53 need to be compared with those by Ad-p14 or Ad-p16 for the therapeutic efficacy. These Ad-mediated effects can be linked with the functions of respective genes re-expressed in mesothelioma. Ad-p14 induced upregulation of p53 expression and subsequently G0/G1 arrest and apoptosis.6 Ad-p14 transduction however did not augment CDDP-induced apoptosis or cleaved caspase-3 levels, showing that the expressed p14 did not influence DNA damage-induced p53 activations.9 Ad-p14 transduction thus did not initiate the entire p53-mediated pathways or p14 may negatively affect apoptotic pathways induced by DNA-damaging agents. In contrast, Ad-p53 transduction enhanced CDDP-induced cytotoxicity as well as PEM-mediated cell killing with a synergistic manner in most of the cells tested. Ad-p53 can thus widely activate p53 downstream pathways compared with Ad-p14.
Transduction with Ad-p16 also induced G0/G1 arrest and apoptosis in mesothelioma cells and achieved better therapeutic effects than that with Ad-p14.5, 18 Restoration of p16 expression can induce dephosphorylated pRb, but precise mechanisms of the apoptosis induction remains uncharacterized because the p16-mediated pathways are not directly linked with the p53 pathways. Moreover, Yang et al.18 showed that Ad-p16 did not increase CDDP-mediated cytotoxicity as well as Ad-p14.9 Interestingly, co-transduction with Ad-14 and Ad-p16 was less effective than Ad-p16 alone, but the underlying mechanism remains unknown.18 A combinatorial use of Ad-p53 and Ad-p16 needs to be tested to clarify whether restoration of the p53 and the pRb pathways may compete in terms of cytotoxicity in mesothelioma. We speculate at this moment that Ad-p53 would produce better anti-tumor effects than Ad-p16 as Ad-p53 can restore the p53-mediated pathways at a full scale and the pRb-mediated pathways as well.
Previous studies showed that combination of Ad-p53 and a DNA-damaging agent was greater in the cytotoxicity than Ad-p53 alone,19 although contradictory studies were also reported.20 This study firstly demonstrated in p53-wild-type mesothelioma cells that augmented p53 expressions produced synergistic anti-tumor effects with CDDP or PEM in most of the cases and even in an orthotopic animal model. Notably, NCI-H2052 cells, insensitive to Ad-p53-mediated cytotoxicity, increased the sensitivity to CDDP and to PEM with Ad-p53 transduction, which verified the benefit of the combination. In addition, the combinatory effects of Ad-p53 with PEM, an novel anticancer agent, is firstly demonstrated in this study, although a possible combination of Ad-p53 and DNA/RNA synthesis inhibitors such as 5-fluorouracil has been reported.21 Activities of enzymes involved in PEM metabolisms may play a crucial role in the anti-tumor effects, but the relationship between the p53 status and PEM sensitivity has been unclear.22, 23 An administration schedule of an anticancer agent and Ad-p53 might be influential to the cytotoxic activity, but our preliminary data did not show any significant difference in the cytotoxicity irrespective of the administration order (data not shown).
In conclusion, this study analyzed the anti-tumor effects of Ad-p53 to mesothelioma cells and demonstrated combinatory effects with CDDP or PEM, which are the first-line agents for mesothelioma. These data indicate a clinical feasibility of Ad-p53 and moreover there are several advantages of intrapleural administration of Ad-p53. Negative intrapleural pressure facilitates Ad distributions into mesothelioma that spreads along the pleura. The administration diminishes rapid Ad uptake in the liver and subsequently decreases Ad-induced acute hepatoxicity. In addition, a previous study revealed that administration of Ad into the pleural cavity did not induce any serious adverse reaction.24 We propose based on this study that intrapleural injection of Ad-p53 with systemic administration of the first-line agents can be a therapeutic strategy for intractable mesothelioma.
This work was partly supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, the Grant-in-Aid for Cancer Research from the Ministry of Health, Labor and Welfare of Japan and a Grant-in-aid from the Nichias Corporation.
About this article
Supplementary Information accompanies the paper on Cancer Gene Therapy website (http://www.nature.com/cgt)