Enhanced Antitumor Therapy by Inhibition of p21^waf1 in Human Malignant Mesothelioma Free

Malignant pleural mesothelioma (MPM) is an asbestos-related malignancy characterized by rapidly progressive and diffusely local growth, late metastases, and poor prognosis (1). Resistance of tumor cells to chemotherapies is the major clinical problem in human MPM (2, 3). Chemotherapy-induced cell cycle arrest, as an alternative to apoptosis, may be a major component of this resistance mechanism. In fact, prolonged drug-induced growth arrest can leave a residual pool of viable malignant cells from which late relapsing clones may ultimately emerge with lethal results. A therapy that could switch the chemotherapeutic response from growth arrest to apoptosis might improve the outcome of MPM treatment.

Several anticancer agents, such as inhibitors of topoisomerase I (e.g., CPT11) and topoismerase II (e.g., doxorubicin), trigger a specific type of growth arrest, namely senescence-like growth arrest, in a variety of solid tumors (4). This process engages some cell-cycle checkpoints and is accompanied by changes in gene expression and up-regulation of senescence-associated β-galactosidase (SA-β-gal; ref. 5). It is known that, such as apoptosis, senescence serves as an anticancer defense mechanism and depends, at least partially, on the p53 and Rb tumor suppressor pathways (5). However, recent findings suggest that senescent cells can also stimulate tumorigenesis in vivo (6), although the extent to which this process contributes to drug action of MPM in vivo is not known.

The p21^WAF1/CIP1 gene (here reported as p21) is transcriptionally activated by p53 and is responsible for the p53-dependent senescence pathway (6, 7). Human colon cancer cells with an intact p21-dependent checkpoint undergo a senescence-like growth arrest after DNA damage caused by ionizing radiation or chemotherapeutic drugs. In contrast, cancer cells with a defective p21 response undergo cell death that occurs during or after mitosis (8). Previous studies (9) showed that MPM expresses wild-type p21, with a frequency of >35%. In addition, unlike other malignancies, genetic alterations in p53 are rare and gene mutations in the p16^INK4a and p14^ARF cell-cycle inhibitors have been seen in up to 70% of MPM (10–13).

In a recent study, Nguyen et al. (14) showed that the loss of p21 in some MPM cells resulted in a strong enhancement of histone deacetylase inhibitor–induced apoptosis in vitro. Here, we analyzed the effect of inhibition of p21 in parallel with chemotherapy treatment on MPM experimental models both in vitro and in vivo. Our data establish that the inhibition of p21 through a RNA interference (siRNA)-based approach enhances the apoptotic potential of anticancer agents in MPM cells otherwise destined to undergo drug-induced senescence. Thus, p21 activation serves as a principal mechanism to escape apoptosis induced by chemotherapy, and inhibition of p21 could be a new adjuvant approach in selected MPM.

Cell lines, plasmids, and chemicals. Human MPM cells, NCI-H28, and NCI-H2052 cells were available in our laboratory (3) and were maintained in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, 1% l-glutamine, 1% Penicillin-Streptomycin (complete medium; all from HyClone) at 37°C, and 5% CO₂. Cells were screened periodically for Mycoplasma contamination (Celbio). HuSH p21 short hairpin RNA (shRNA; 29-mer) and control plasmids, purchased from Origene, were transfected into H28 and H2052 cells using Lipofectamine (Invitrogen) according to the manufacturer's protocols, followed by selection with 1 ng/mL puromycin (Sigma). Antibiotic-resistant pools and individual clones were isolated and maintained in the presence of puromycin. The antineoplastic agents doxorubicin (Adriamycin), etoposide, and CPT11 (irinotecan) were purchased from Sigma. Death receptor–mediated apoptosis was induced with tumor necrosis factor–related apoptosis-inducing ligand (TRAIL; Alexis Biochemicals), or cytotoxic anti-Fas CH11 antibody (Upstate Biotechnology). All concentrations used in vitro were previously determined (3, 15).

Adenovirus-mediated p21 gene transfer. The adenoviral vectors Ad-cytomegalovirus (CMV)-LacZ and Ad-CMV-p21 were previously described (16). For gene transfer experiments, MPM cells were grown in 6-well tissue culture plates to 60% confluence, washed once with PBS, and infected with adenoviral vectors at a multiplicity of infection of 10 or 30 plaque-forming units per cell as previously described (16). Thereafter, MPM cells was replaced with complete RPMI, and cells were incubated for 24 h before being used in subsequent experiments.

Immunoblot analysis. Methods for preparation of total cellular extracts in the presence of a protease inhibitors cocktail, SDS-PAGE electrophoresis of cellular proteins (50 μg), and transfer onto polyvinylidene difluoride membranes were done as previously described (3, 15). Membranes were incubated with antibodies against p53, p21, prosaposin (anti-SGP1), procaspase-3, or antiactin (all purchased from Santa Cruz); washed with Tween 20 in PBS; incubated with peroxidase-conjugated secondary antibody; and the signal was then detected with a chemiluminescence-based system (Promega). For each protein tested, Western blot analyses were repeated at least thrice.

Annexin and propidium iodide fluorescent staining. Annexin and propodium iodide fluorescent staining was carried out as described previously (17, 18). Cells positively stained with annexin and not propidium iodide were considered apoptotic, and cells negative for the two dyes were considered live cells. During flow cytometric analysis to determine propidium iodide staining, forward versus side scatter plots were also used to establish changes in cell size and granularity.

In vitro β-Gal staining. SA-β-gal activity was determined as described previously (17). Light micrographs were taken using a RTSlider Spot digital camera equipped with the Spot Advanced Software, Version 3.1 (Diagnostics Instruments) linked to a phase contrast Nikon Optiphot microscope (Nikon Corporation).

Colony formation assays. Anchorage-independent proliferation was quantified by standard soft agar assays (3). Cells (10⁴) were suspended in 1.5 mL of 0.4% (w/v) Noble agar (Sigma) and overlaid on 0.6% (w/v) agar. After 2 wk of incubation at 37°C, colonies were visualized by staining with nitrotetrazolium blue chloride (1 mg/mL), and those >0.5 μm in diameter were counted. All experiments were done in triplicate and repeated at least thrice.

Tumor formation assay in nude mice. Suspensions of 1.0 × 10⁶ H2052-derived cells (p21 shRNA–transfected cells, MM/p21-shRNA; empty vector–transfected cells, MM/C-shRNA) in PBS (50 μL) were injected s.c. into the left flanks of 5-wk-old male BALB/c nu/nu mice (weight, ∼18 g; Harlan Sprague-Dawley) at day 0. The inoculation was conducted in at least 10 mice, and tumor growth was estimated from the average volume of tumors. Tumor volume was calculated by the formula 1/2 × L² × W (L, length of the tumor; W, width of the tumor). At indicated days after inoculation, representative mice were sacrificed, and s.c. tumors were resected and fixed in 10% formaldehyde/PBS. The tumors were paraffin embedded and stained with H&E and for p21 and p53BP. Immunohistochemical staining for SA-β-gal was done as reported previously (17, 19).

I.p. MM experimental models were carried out according to our usual method (20). All the animal experiments were done in accordance with institutional guidelines of the Marche University. Mice were maintained and handled under aseptic conditions, and animals were allowed access to food and water ad libitum.

Treatment with CPT-11 in nude mice. MM/p21-shRNA and MM/C-shRNA tumors were generated both s.c. and i.p. in nude mice as above. S.c. tumors were treated 2 to 3 wk after cancer cells were injected when tumor dimensions were ∼1 cm × 1 cm. CPT11 (Sigma) was administered by tail vein injection at a dose of 33 mg/kg in 100 mL every 4 d beginning 24 h after the first viral injection. A total of five doses of CPT11 were given. PBS (100 μL) administered by tail vein injection was used as a control for drug treatment. Tumor growth was estimated from the average volume of tumors, and tumor volume was calculated as already described. At different days after inoculation, mice were sacrificed, s.c. tumors were resected, and tumor size was compared. For i.p. MM tumor models, mice received a single i.p. injection of 1.3 to 1.6 × 10⁷ MM/p21-shRNA and MM/C-shRNA cells in a volume of 100 μL PBS, and CPT11 therapy was initiated 7 d later at the dosage described above. All the animal experiments were done in accordance with institutional guidelines of the Marche University.

In vivo deoxynucleotidyl transferase–mediated dUTP nick-end labeling assay. Animals were killed by asphyxiation in a CO₂ chamber. The tumors or scar tissues were fixed in 4% paraformaldehyde at room temperature for 24 h. The specimens were cut into sections 4 mm in thickness and stained with H&E. For terminal deoxynucleotidyl-transferase–mediated dUTP nick-end labeling (TUNEL) staining, the tumors were either untreated or treated with CPT11 for different times. The tumors then were collected, embedded in ornithine carbamyl transferase compound (Invitrogen), and “snap frozen” in liquid nitrogen. Frozen sections 7 mm in thickness were cut and fixed in 4% paraformaldehyde and stained according to the manufacturer's instructions (Boehringer). This assay measures DNA strand breaks and is therefore diagnostic for cells undergoing apoptosis.

Statistical analysis. All experiments were done in duplicate or triplicate and were expressed as mean ± SE. The Bonferroni multiple comparison test and Student's t test were used for the statistical analysis of comparative data using StatView version 5.0 (NET Engineering). P values of ≤0.05 were considered significant and are indicated by asterisks in the figures.

Doxorubicin-induced p21 confers resistance to doxorubicin in human mesothelioma cells. To determine whether p21 confers resistance to genotoxic drugs, we inhibited p21 expression by shRNA in MPM cells that were then exposed to moderate doses of Doxorubicin (100 nmol/L). This widely used anticancer drug produces DNA damage and increases p21 protein expression (8). We tested four 29-mer shRNA against p21. Two such shRNAs (No.1 and No.4) decreased Doxorubicin-induced 21 protein levels, as determined by anti-p21 immunoblot analysis, when transiently introduced into H28 mesothelioma cell lines (data not shown). Stable transfection of H28 cells with constructs encoding p21 shRNA No.1 or No.4, but not the parental vector (C-shRNA), decreased steady-state levels of p21 protein and the levels of antiapoptotic protein prosaposin (SGP1), which is a p21-responsive gene (Fig. 1A). To ask whether these findings could be extended to other MPM cell lines, we repeated these experiments in H2052 cells. These cells produced p21 in response to doxorubicin, though to lesser extent than H28 cells, and down-modulation of p21 levels with shRNA dramatically inhibited p21 and SGP1 protein expression (Fig. 1A).

To analyze the effect of p21 inhibition on cellular resistance to Doxorubicin, we have measured colony formation after exposure to low doses of Doxorubicin (30-120 nmol/L). In both H28 and H2052 cells, Doxorubicin showed an IC₅₀ of 33 ± 3.0 nmol/L in the presence of p21-shRNA and 125 ± 10 nmol/L in the presence of C-shRNA. These data suggest that doxorubicin-induced p21 expression may confer drug resistance in MPM cells.

Disabling of doxorubicin-induced G₂ arrest and senescence by loss of p21. To characterize the effect of p21 inhibition in doxorubicin-treated MPM cells, we exposed H28 cells with shRNAs [C-shRNA and p21-shRNA (No. 1)] in combination with Doxorubicin and then analyzed cell cycle distribution by flow cytometry. After 1 day of treatment, H28 cells expressing C-shRNA displayed a distinct and progressive accumulation in G₂-M and reduction in S phase, a profile characteristic of G₂ arrest (Fig. 2A). This cell cycle distribution remained relatively constant from day 1 to 5 with the majority of cells (∼80-85%) in G₂ arrest. When H28 cells were treated with p21-shRNA, doxorubicin produced only minor shifts in cell cycle distribution at 1 day of treatment (Fig. 2A). Longer treatment was accompanied by a significant fraction (48%) of apparently hypodiploid nuclei (data not shown). Using Annexin V staining, we observed that cell death occurred in 30% to 35% of H28 cells treated with p21-shRNA after 5 days of treatment (Fig. 2B). In addition, p21 inhibition also increased caspase-3 cleavage (Fig. 2B, insert), a well-known inducer of apoptosis. Thus, p21-shRNA inhibits the G₂ arrest promoted by low doses of doxorubicin and sensitizes H28 cells to cell death.

Drug-induced cell cycle arrest and p21 accumulation in different cell lines have been shown to promote the senescent phenotype (8). Therefore, we investigated whether doxorubicin induces phenotypic features of cell senescence such as increased cell size and granularity and staining for the SA-β-gal marker of senescence (5, 6). Changes in cell size and granularity were investigated by flow cytometric assays through examination of the forward versus side scatter plots. The profiles show that after 24 h of treatment (at the time in which the G₂ arrest was already observed), the entire cell population was slightly larger and considerably more granular. Doxorubicin also induced SA-β-gal activity in H28 cells after 5 d of treatment (Supplementary Data S1). As observed in time course experiments, the loss-of-expression of p21 led to only 13.8% of SA-β-gal–positive cells at day 5 with respect to 83.2% of SA-β-gal–positive cells counted in C-shRNA–transfected H28 cells (Fig. 3A). To ascertain the role of p21 in modulating senescence induced by Doxorubicin, p21-shRNA–transfected cells were transduced with an adenoviral vector expressing p21 (Ad-CMV-p21) or a control vector (Ad-CMV-LacZ) 24 h before treatment with Doxorubicin. Our previous experiments using the Ad-CMV-LacZ vector showed reporter gene expression in 100% of MPM cells within 7 days after adenoviral infection at multiplicity of infections of 10 or 30 plaque-forming units per cell (16). Western blot analysis of cell lysates harvested 5 days after adenoviral infection revealed that, at comparable multiplicities of infection, Ad-CMV-p21 mediated robust p21 expression in p21-shRNA–transfected cells. In contrast, the Ad-CMV-Lac Z control vector did not rescue p21 expression in these cells (data not shown). Overexpression of p21 totally rescued the profound senescence effect of Doxorubicin in p21-shRNA–transfected H28 cells (Fig. 3B). In contrast, Ad-CMV-LacZ had no effect. To extend our result, p21-shRNA/H2052 clones and C-shRNA–transfected H2052 clones were treated with Doxorubicin, and the percentage of G₂ arrest, senescent, or apoptotic cells was determined. Confirming the results observed with H28 cells, ∼80% of C-shRNA/H2052 cells became senescent 5 days after treatment. In contrast, only 12% to 20% of p21-shRNA/H2052 cells were senescence and 18% to 25% shown G₂ arrest (Fig. 3C). Interestingly, we observed that cell death occurred only within a minority of C-shRNA/H2052 clones and that ∼33% of p21-shRNA/H2052 clones were positive to Annexin V at the 5-day time point (Fig. 3C). Thus, low doses of Doxorubicin promotes a p21-dependent senescence in human MPM cells and inhibition of p21 successfully converted a senescent, apoptotic-resistant phenotype to a more sensitive one in terms of apoptosis.

Potentiation of chemotherapy-induced cell death in vitro by p21 shRNA. We also analyzed whether p21 similarly influences the apoptotic and senescence response to other chemotherapeutic drugs and to death receptor ligands. Unlike death receptor ligands anti-CD95 (Fas) and TRAIL, low doses of etoposide and CPT11 potently increased p21 protein expression in both H28 and H2052 cells (Fig. 4A). About 80% of control cells became senescent 5 days after etoposide or CPT11 treatment. In contrast, only 20% to 35% of p21-shRNA–transfected MPM cells expressed SA-β-gal activity in the presence of drugs (Fig. 4B). Although MPM cells expressing p21-shRNA were insensitive to the apoptotic effect of TRAIL and anti-CD95, ∼35% to 45% of cells undergo cell death after treatment with chemotherapeutic drugs (Fig. 4C). These results suggest that in MPM cells, p21 expression does not interfere with the death pathway triggered by death receptors, in contrast to its strong inhibitory effect on chemotherapeutic drug–induced apoptosis.

Anticancer agents elicit tumor regression after p21 inhibition. Because of the encouraging data in vitro, we sought to determine if similar results could be attained in tumors in vivo. We generated in nude mice s.c. tumors derived from H2052 cells expressing either the control vector (H2052/C-shRNA) or the No.1 siRNA for p21 (H2052/p21-shRNA). Then, we assessed the antitumor effects after systemically administration of CPT11 in H2052/C-shRNA and H2052/p21-shRNA tumors of similar size. H2052/p21-shRNA tumors grow more slowly than H2052/C-shRNA tumors and failed to regress after therapy (Fig. 5A). However, they did not progress for extended periods. This is in marked contrast to H2052/C-shRNA tumors, which progressed rapidly (Fig. 5A and B).

We also explored the use of p21-shRNA in combination with CPT11 in i.p. implantation of tumor cells (20). In this assay, we compared the overall survival of CPT11-treated MPM tumors derived from H2052 cells (H2052/C-shRNA and H2052/p21-shRNA tumors) and H28 cells (H28/C-shRNA and H28/p21-shRNA tumors; Fig. 5C and D). Although all mice harboring MPM/p21-shRNA tumors eventually succumbed to their disease within 110 days after therapy, they lived much longer than mice with MPM/C-shRNA tumors (P < 0.0001, e.g., at 60 days: About 100% overall survival for MPM/p21-shRNA tumors versus 20% to 30% for MPM/C-shRNA group). These results show that chemotherapeutic agents were unable to elicit a growth inhibitory effect in MPM tumors unless p21 was inhibited, establishing p21 as an important mediator of MPM chemoresistance.

Inhibition of drug-mediated senescence and enhancement of apoptosis by p21 shRNA in vivo. Furthermore, we determined whether the effect of CPT11 on MPM/p21-shRNA tumors was through apoptotic or senescence mechanisms. Note that unlike H2052/p21-siRNA tumors, H2052/C-siRNA tumors displayed sustained p21 expression, even after CPT11 therapy (Fig. 6A, top). To determine cell senescence in vivo, we looked for two features evident in senescent cells, such as the expression of p53-Binding Protein (p53BP) and SA-β-gal (4–7). Staining with antibodies against p53BP revealed abundant positive cells in paraffin-embedded H2052/C-siRNA tumors, whereas H2052/p21-siRNA tumors were essentially negative (Fig. 6A, bottom). Cryosections from H2052/C-siRNA tumors stained for SA-β-gal gave an intense signal, whereas those from H2052/p21-siRNA tumors gave a weak or negative signal (Fig. 6B). We also collected several H2052/p21-siRNA and H2052/C-siRNA tumors after 5 to 45 days of CPT11 treatment and analyzed the samples by TUNEL assay, which detects DNA fragmentation and is commonly used to detect in vivo apoptosis (4). H2052/p21-shRNA tumors had widespread TUNEL-positive cells, in contrast to the staining of H2052/C-shRNA tumors (Fig. 6C), indicating that CPT11 elicited tumor cell apoptosis when p21 was inhibited. Similar results were obtained when we used H28 transfectants instead of H2052 clones (Supplementary Data S2). Therefore, our result indicated that p21 inhibition disrupted tumor cell senescence response after CPT11 treatment, leading to tumor cell death.

Chemotherapy remains the primary treatment for MPM patients. However, clinical trials have consistently shown that MPM is highly resistant to chemotherapy and that increased concentrations of cytotoxic drugs fail to improve the pharmacotherapeutic response in this malignancy (1). Thus, research efforts are aimed at determining the regulatory events involved in chemoresistance of MPM.

Resistance to apoptosis controls the ability of tumors to withstand high levels of chemotherapy (2). Moreover, cells that are resistant to apoptosis seem to have a growth advantage in tumors (21). Although there are many mechanisms by which resistance to apoptosis is achieved in MPM (3, 18), our data indicate that one protective mechanism is the expression of cyclin-dependent kinase inhibitor p21 within MPM cells in response to several chemotherapeutic drugs. In fact, inhibition of drug-induced p21 expression strongly enhances the apoptotic potential of these stimuli. Our findings probably explain the lack of efficacy of both topoisomerase I and II inhibitors in clinical trials for MPM and the reduced efficacy of genotoxic drugs in general for this type of tumor (1, 22). Moreover, we identify an important mechanism whereby MPM is induced to become resistant to cancer therapy. Consistent with our results, the use of p21 antisense approach in combination with radiation in cancer therapies has been described (23), and it is much more efficacious than a single type of treatment. Although shRNA techniques show promise and have been available for some time, their movement into the clinical arena is still under investigation. Based on our studies, p21 shRNA or gene therapy to attenuate p21 levels in MPM tumors and/or patients may have tremendous potential as therapy, likely in combination with standard DNA-damaging agents.

How does p21 expression suppress apoptosis in MPM cells in response to chemotherapy? The dominant cellular response to DNA-damaging agents in MPM cells was a prolonged G₂ growth arrest and drug-induced senescence development. The maintenance of viability during the senescence-like growth arrest after drug treatment was dependent on p21, such that apoptosis supervened as the major cellular response in MPM after p21 knockdown. Thus, maintenance of a stable p21-dependent senescence-like growth arrest contributes to the protection of MPM cells from drug-induced cell death, as has been observed with chemotherapeutics of various cytotoxic mechanism in other model systems (23, 24).

The majority of MPM cells express wild-type p53, and p21 is transcriptionally activated by p53 (8–11). In our experiments, p53 was not induced after doxorubicin treatment at the doses ranging from 3 to 150 nmol/L. Moreover, p53 knockdown with a specific shRNA did not significantly change p21 accumulation and the rate of SA-β-gal–positive cells induced by doxorubicin (data not shown). On the other hand, p21 knockdown markedly enhanced the apoptotic response of MPM cells to anticancer agents, switching the ultimate cellular fate from drug-induced senescence to apoptosis in a significant proportion of cells, as shown by apoptotic assays (see Fig. 4), reflecting an early and reversible step in the apoptotic pathway (24, 25), and by procaspase-3 cleavage assays (see Fig. 2), reflecting caspase activation in late apoptosis (18). Clearly, in p21 shRNA–transfected MPM cells, the apoptosis observed must occur via a p53-independent pathway because it was not significantly inhibited by p53 knockdown (data not shown).

p21 modulates the activity of multiple intracellular proteins, and accumulating evidence indicates that p21 is a major negative regulator of p53-dependent and p53-independent cell death (26). p21 exists in a quaternary complex with proliferating cell nuclear antigen, cyclin D, and cyclin-dependent kinases in normal cells (27). Interaction of p21 with proliferating cell nuclear antigen results in inhibition of DNA synthesis by DNA polymerase δ (28). In addition, p21 suppresses activation of procaspase 3 by masking its serine protease cleavage site (29). Although not formally shown in our experiments, it is quite likely that loss-of-expression of p21 disrupts p21/procaspase 3 complex formation and contributes to potentiation of drug-mediated cytotoxicity.

We extended our in vitro findings to the same cell lines in vivo by growing them as s.c. tumors in nude mice and measuring p21 expression, apoptosis, and senescence in vivo. Nguyen et al. (14) had contended that the improved chemocurability of MPM tumors was due to p21 inhibition that they saw in vitro, but they had never actually measured the p21 expression in tumors themselves, and there are no data that the senescence process was involved. Our study provides the first evidence that cellular senescence is induced in MPM tumors after chemotherapy in vivo. Drug-induced senescence was accompanied by substantial increases in p21, p53BP (a p53-binding protein that plays a role in cellular responses to DNA damage linked to senescence), and SA-β-gal activity. Disruption of p21 within tumor cells disabled the senescence program, prevented SA-β-gal accumulation, accelerated apoptosis, and resulted in attenuation of s.c. MPM tumor growth and in increased of overall survival in i.p. MPM tumors (see Figs. 5 and 6).

A key finding derived from this study is that a senescence program controlled by p21 may be an important determinant of treatment outcome in vivo. It is also noteworthy that although senescent cells do not divide but remain metabolically active, they can produce numerous secreted factors with diverse paracrine activities. In addition to those that inhibit tumor growth (e.g., BTG1 and BTG2, Maspin, MIC-1, and others), senescent tumor cells might also generate factors that stimulate tumor growth (e.g.,vascular endothelial growth factor, extracellular matrix component Cyr61, and prosaposin, galectin-1) and potentially contribute to metastatic cell proliferation (6, 7). Thus, successful tumor treatment based on senescence-induced strategy must take into account multiple tumor-suppressing and tumor-promoting activities in the cells. Comprehensive analysis of tumor cell senescence may offer a plausible approach to the development of novel therapeutic strategies for MPM.

In conclusion, we show that chemotherapy resistance mechanism in MPM involves the cyclin-dependent kinase inhibitor p21, and that this action is mediated at least in part by the induction of senescence. More experiments are necessary to better understand the mechanisms underlying this involvement. However, we have identified a potential therapeutic approach to circumventing such resistance. Disruption of p21 function, if achieved in tumors in vivo, using antisense technology or using small molecule inhibitors currently under development (23), might provide a means of enhancing the chemotherapeutic effect. Therefore, the combination of CPT11 and p21 shRNA merits testing in models of MPM in vivo.

No potential conflicts of interest were disclosed.

We thank Dr. Maria Cristina Albertini (Institute of Biological Chemistry, University of Urbino, Urbino, Italy) for assistance with the in vivo experiments, and Dr. Mirco Fanelli (Centre of Biotechnology, University of Urbino, Fano, Italy) for discussions.