Background
Diffuse malignant peritoneal mesothelioma (DMPM) is an uncommon though locally aggressive tumor that develops from mesothelial cells lining the peritoneal cavity [
1]. DMPM prognosis is dismal and standard therapy, including palliative surgery, systemic/intraperitoneal chemotherapy, and abdominal irradiation, showed to be ineffective, with a median survival of about 1 year [
1]. Currently, the most effective treatment is a loco-regional approach combining aggressive cytoreductive surgery (CRS) with hyperthermic intraperitoneal chemotherapy (HIPEC), which significantly extended survival in selected series of patients [
1]. However, for recurrent patients and for those who are not eligible to CRS+HIPEC, the prognosis remains severe due to the lack of alternative treatment options [
2]. Considering the rarity of the disease and the unavailability of experimental models, the biology of DMPM is still largely unknown. It is anticipated that advances in the knowledge of the mechanisms responsible for the biological aggressiveness and the relative chemoresistance of DMPM will allow the identification of relevant targets for the development of novel therapeutic strategies.
MicroRNAs (miRNAs) are single-stranded endogenous evolutionary conserved, non-coding RNA molecules acting as post-transcriptional regulators of gene expression [
3]. Deregulated miRNA expression and/or function have been observed in a variety of human solid and hematological tumors and have been causatively linked to the pathogenesis of cancer [
4]. Depending on their expression levels, cellular context, and target functions, miRNAs can act as oncogenes or tumor suppressors [
4] and may represent novel targets or tools for cancer therapy [
5].
MiR-34a is one of the most widely studied miRNAs in cancer. Its expression has been found to be decreased in a variety of human tumors [
5] due to DNA copy number loss or epigenetic silencing through aberrant CpG methylation [
6,
7]. Results of studies where the expression of miR-34a was manipulated in human tumor experimental models clearly showed that the miRNA acts as a tumor suppressor by regulating highly relevant processes such as proliferation, cell cycle, apoptosis, invasion, and metastasis [
8]. Reinforced expression of miR-34a has been found to positively modulate drug response in cancer cells [
8]. In addition, a liposomal nanoparticle-formulated synthetic miR-34 (MRX34) recently entered a phase I clinical study for patients with different tumor types [
9].
No information concerning the expression and/or the functional role of miRNAs in DMPM is currently available in the literature. Based on the knowledge that several receptor tyrosine kinases (RTKs) are validated targets of miR-34a [
8] and our previous results indicating that activation of downstream RTK signaling, in terms of phosphorylation/overexpression of extracellular signal regulated kinase ½ (ERK1/2), AKT, and mTOR, is present in a considerable fraction of DMPM clinical specimens [
10], we proposed to investigate the possible relevance of miR-34a in the disease, with the final aim to develop novel therapeutic strategies. Here, we report that miR-34a is down-regulated in DMPM clinical specimens and demonstrate that miR-34a replacement in a unique collection of in-house-developed human DMPM experimental models [
11‐
13] inhibits cell proliferation and invasion and impairs tumor growth formation in SCID mice, mainly as a consequence of c-MET and AXL inhibition. These findings identified miR-34a-AXL and -c-MET axes as promising therapeutic targets for DMPM. Moreover, we provide evidence of persistent activation of ERK1/2 and AKT as a possible cytoprotective mechanism to RTK inhibition by miR-34a.
Methods
Clinical samples
Forty-five DMPM specimens classified as epitheliod (40), sarcomatoid (1), and biphasic (4) from patients treated with CRS+HIPEC at the Fondazione IRCCS Istituto Nazionale dei Tumori, Milan (INT) from October 1997 to February 2013, and 7 normal peritoneum specimens from patients who underwent surgery for non-oncologic disease were available for miR-34a expression analysis.
This study was approved by the Institutional Review Board and Ethical Committee and each patient provided written informed consent to donate to INT the leftover tissue after diagnostic and clinical procedures.
Cell lines and culture conditions
The human mycoplasma-free DMPM cell lines MesoII, STO, MP115, MP4, and MP8 were established in our laboratory [
11‐
13]. All cells were cultured in DMEM F-12 medium (Lonza, Milano s.r.l., Treviglio, Italy) supplemented with 10% fetal bovine serum in a 37 °C humidified 5% CO
2 incubator. Cell lines were authenticated by single-tandem repeat analysis by the AmpFISTR Identifiler PCR amplification kit (Applied Biosystems, Foster City, CA, USA).
Cell transfection
Mimic pre-miR-34a precursor (miR-34a) and mimic negative control (Neg) were purchased as Pre-miR™ miRNA precursor molecules (Thermo Fisher Scientific, Monza, Italy). Knockdown of AXL and c-MET was performed using specific siRNAs (siAXL and siMET; ON-TARGET plus SMART pool) and, as a control, a siRNA with a nonsense/scrambled sequence (siNeg, ON-TARGET plus non-Targeting Pool) (Dharmacon, CO, USA) was used. Cells were transfected for 24 h with 20 pM miR-34a or Neg, or 100 nM siAXL, siMET, or siNeg, using Lipofectamine® RNAiMAX Transfection Reagent (Thermo Fisher Scientific) with Opti-MEM I (Gibco, NY, USA) according to the manufacturer’s instructions.
Quantification of miR-34a expression levels was assessed by qRT-PCR. Total RNA was isolated using the miRNeasy Mini Kit (QIAGEN, Hilden, Germany) and 1 μg of RNA was reverse transcribed by miScript II RT Kit (QIAGEN). Mature miRNA expression was assayed by miScript Primer Assays specific for miR-34a (MS00003318) and normalized on SNORD48 (MS00007511) (QIAGEN). Quantitative RT-PCR was conducted using miScript SYBR Green PCR Kit (QIAGEN). The reaction was carried out in a 96-well PCR plate at 95 °C for 15 min followed by 40 cycles of 94 °C for 15 s, 55 °C for 30 s, and 70 °C for 30 s and a dissociation step to distinguish specific from non-specific amplification products. Each sample was analyzed in triplicate.
Amplifications were run on the 7900HT Fast Real-Time PCR System (Applied Biosystem). Data were analyzed by SDS 2.2.2 software (Applied Biosystems) and reported as -ΔCt, that is the difference between the Ct of the target gene and the Ct of the housekeeping gene (where Ct is the threshold cycle), or as relative quantity (RQ) or -ΔΔCt with respect to a calibrator sample (i.e., negative control transfected cells) according to the 2−ΔΔCt method.
Cell growth assay
To assess the effect of miR-34a restoration on cell proliferation, DMPM cells were transfected with Neg or miR-34a as described above. At different intervals from transfection, cells were trypsinized and counted in a particle counter (Beckman Coulter, Cassina de’ Pecchi, Italy). Results were expressed as percent variation in the number of miR-34a-transfected cells compared with Neg-transfected cells.
Immunoblotting analyses
Cell lysates were fractionated by SDS-PAGE, transferred to nitrocellulose membranes, and probed with specific antibodies, as described in [
14]. Cells were lysed and western blot was performed using the following primary antibodies: anti-c-MET, −CDK6, −uPA, −pospho-FAK (Tyr 576/577) (Santa Cruz Biotechnology, CA, USA); anti-AXL, −phospho-AKT (Ser473), −phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204), −p44/42 MAPK (ERK1/2), −FAK, −cleaved CPP32 (Cell signaling, Beverly, USA); anti-AKT (BD Biosciences, San Jose, CA, USA); and anti-actin and -vinculin (Sigma Chemical Company, St. Louis, MO, USA).
Secondary antibodies used were conjugated to horseradish peroxidase (GE Healthcare, Little Chalfont, UK). Immunostained bands were detected by chemoluminescence method (ECL, GE Healthcare). In many experiments, membranes were stripped and reblotted with a second antibody. Moreover, membranes were cropped to allow simultaneous incubation of different primary antibodies on the same samples. For the preparation of figures, we cropped the original western blot to generate the appropriate figure panels with the relevant lanes. This cropped image was then subjected to uniform image enhancement of contrast and brightness. Molecular weights were determined using the Precision Plus Protein™ Standard (Bio-Rad, Segrate, Italy), which yields a colorimetric image only and has been removed from the chemoluminescent blot image.
Drugs
The AKT-1/2 inhibitor trifluoroacetate salt hydrate (A6730, Sigma Chemical Company) and the MEK inhibitor CI-1040 (PD184352, Selleck Chemicals, Houston, TX, USA) were dissolved and diluted in DMSO. Final concentration of DMSO in cell cultures never exceeded 0.5%. The antiproliferative activity was evaluated by cell counting at different times after exposure of miR34a-reconstituted MesoII cells to drug concentrations able to inhibit cell proliferation by 20% (IC20).
Apoptosis detection and cell cycle analysis
At different time points after transfection with miRNAs mimic or siRNAs, floating and adherent cells were harvested and processed for apoptosis evaluation by TUNEL assay according to manufacturer’s instructions (Roche, Mannheim, Germany) and for cell cycle [
15]. For cell cycle, cells were fixed in 70% ethanol 96 h after transfection, stained in phosphate-buffered saline (PBS) containing 10 μg/ml propidium iodide (PI; Sigma Chemical Company), and RNase A (66 U/ml; Sigma Chemical Company) for 18 h and analyzed by FACScan flow cytometer (Becton Dickinson, Mountain View, CA, USA).
Senescence-associated β-galactosidase staining
Cells were transfected with miR-34a or Neg for 24 h. Samples were washed in PBS 72 h after transfection and processed for senescence-associated β-galactosidase (SA-β-Gal) staining. Cells were fixed for 5 min (room temperature) in 2% formaldehyde/0.2% glutaraldehyde, washed and incubated overnight at 37 °C (no CO
2) with fresh solution as previously described [
15]. At least 300 cells were examined, and the results were expressed as percentage of SA-β-Gal positive cells over the whole population.
Transwell invasion assay
Invasion assay was performed 72 h after transfection using a 24-well Boyden chamber with 8-mm pore size filter in the inset chambers (Costar, Corning Inc., NY, USA). The Transwell membranes were previously coated with 3.47 μg Matrigel/well (BD Biosciences) and dried for 30 min. Cells were suspended in 300 μL serum-free medium and seeded into the insert chambers. After 24 h of incubation at 37 °C in 5% CO2, cells that migrated into the bottom chamber containing 1 ml of serum-free medium were fixed in 95% ethanol, stained with a solution of 0.4% sulforhodamine B in 0.1% acetic acid, counted under an inverted microscope, and then photographed.
Antibody arrays and ELISA
Cells were seeded at 2 × 104 cells/dish in complete medium and transfected with Neg or miR-34a for 24 h before serum starvation for 72 h. Conditioned media were then harvested and clarified by centrifugation at 13,000 rpm for 15 min. Cells were trypsinized, counted, and lysed for assaying protein content. Supernatant aliquots were used to assess angiogenesis-related protein content by Antibody Arrays (R&D System, SPACE Import Export, Milan, Italy) according to manufacturer’s instructions. The ELISA kit for Maspin (human Maspin “Super X” ELISA Kit, Antigenix America, Huntington Station, NY, USA) was used according to the manufacturer’s instructions for quantitative analysis.
In vivo experiments
All experimental protocols were approved by the Ethics Committee for Animal Experimentation of INT. Experiments were performed using 8-week-old female SCID mice (Charles River, Calco, Italy). Each group contained five to six mice. Cells were transfected with miR-34a or Neg for 24 h, as described above, and then inoculated subcutaneously or intraperitoneally after the analysis of the transfection efficiency by qRT-PCR.
Subcutaneous tumor models
STO, MesoII, and MP8 cells were injected subcutaneously into the right flank (1–1.2 × 10
7 cells/mouse). Inoculated animals were inspected daily to establish the time of tumor onset. Tumor growth was measured every 2 to 3 days using a Vernier caliper (Table
1). The subcutaneous tumor volume was calculated as follows: TV (mm
3) =
d
2 ×
D/2 where
d and
D are the shortest and the longest diameter, respectively. Volume inhibition percentage (TVI%) in tumors derived from miR-34a- over Neg-transfected cells was calculated as follows: TVI% = 100 − (mean miR-34a TV/mean Neg TV × 100).
Table 1
Effect of miR-34a reconstitution on DMPM cell tumorigenicity following s.c. injection in SCID mice
STO | Neg | 6/6 | 1 | 440 ± 94 | | |
miR-34a | 6/6 | 1 | 188 ± 39 | 57 (10) | 0.0003 |
MesoII | Neg | 5/5 | 7 | 201 ± 93 | | |
miR-34a | 5/5 | 12 | 9 ± 12 | 96 (12) | 0.0035 |
MP8 | Neg | 5/5 | 18 | 72 ± 35 | | |
miR-34a | 5/5 | 25 | 1 ± 2 | 98 (21) | 0.0041 |
Proteins were obtained as described previously [
16] from frozen s.c. tumors derived from two additional mice sacrified at different time points. Briefly, samples were pulverized by Mikro-Dismembrator II (B. Brown Biotech International, Melsungen, Germany) and suspended in lysis buffer supplemented with protease and phosphatase inhibitors. Proteins were processed as described [
16].
Intraperitoneal (orthotopic) tumor models
STO and MP8 cells were injected into the peritoneal cavity (10
7 and 2.5 × 10
7 cells/mouse, respectively). Animals were monitored and weighed daily and sacrificed at different times from cell injection (Table
2). A careful necropsy was performed to evaluate the take rate and spread of mesothelioma cells in the abdominal cavity.
Table 2
Effect of miR-34a reconstitution on DMPM cell tumorigenicity following i.p. injection in SCID mice
STO | Neg | 14 | 5/5 | | 10, 80, 80, 100, 110 | 76 | | |
miR-34a | | 1/5 | 0.0476 | 0, 0, 0, 0, 10 | 2 | 97 | 0.0029 |
MP8 | Neg | 31 | 5/5 | | 130, 310,250, 210,150 | 210 | | |
miR-34a | | 5/5 | | 80, 90, 70, 120, 80 | 88 | 58 | 0.0071 |
Solid masses were gently detached from organs and abdominal walls, removed, and weighed for calculating the percentage of tumor weight inhibition (TWI %) in mice inoculated with miR-34a- over Neg-transfected cells.
Statistical analyses
If not otherwise specified, in vitro data are presented as mean values ± SD from at least three independent experiments. Statistical analysis of the data was performed by two-tailed Student’s
t test. For in vivo data, two-tailed Student’s
t and Fisher’s exact test were used to compare tumor volumes/weights and tumor takes, respectively. Patient survival analysis was performed using Cox proportional regression model [
17].
p values <0.05 were considered statistically significant.
Discussion
No information is currently available on the expression and functional role of miRNAs in DMPM. Here, we demonstrated that miR-34a is down-regulated in a large series of DMPM clinical samples and in a unique panel of cell lines, established from DMPM patients in our laboratories, compared to normal peritoneum specimens. We also illustrated that miR-34a exerts oncosuppressive functions in our tumor models, consistent to what previously observed in a variety of human tumor types [
5,
39‐
41]. Indeed, miR-34a reconstitution impaired proliferation and induced an apoptotic response in DMPM cell lines, although at a variable extent and with different kinetics, mainly through the down-regulation of c-MET and AXL and the interference with the activation of downstream signaling. Interestingly, results also indicated that a transient or persistent activation of ERK1/2 and AKT can delay or prevent the cytotoxic and proapoptotic effects of miR-34a reconstitution, as observed in MesoII and MP115 cells, respectively. Noteworthy, DMPM cell feedback to AXL and c-MET down-regulation induced by miR-34a reconstitution is to directly activate ERK1/2 and AKT survival signaling cascades rather than up-regulate the expression levels of the receptors, thus ensuring a more prompt counter-response. Such findings provide the first evidence that tumor cells can exploit a well-known mechanism of resistance to RTK inhibitors—i.e., the activation of RTK downstream signaling [
23‐
26]—to counteract the antiproliferative/proapoptotic effects of miR-34a. However, such a mechanism was not found to protect DMPM cells from the anti-invasive effect of the miRNA.
Noteworthy, the cytoprotective mechanism based on ERK1/2 and AKT activation was mainly evident in the DMPM cell line MP115, derived from a biphasic subtype tumor. Such subtype is known to be more aggressive and associated with a reduced patient survival compared to the epithelioid [
42], although differences in specific relevant biological properties between the two DMPM subtypes are currently unknown. In addition, the delayed pro-apoptotic and cytotoxic effects observed in MesoII cells following miR-34a reconstitution are consistent with the finding that, unlike other epithelioid cell lines, they carry a mutant p53 [
11].
A novel mechanism of miR-34a-dependent AKT inhibition has been recently proposed by Wang et al. [
43]. In this study, miR-34a is reported to inhibit Bmi-1 by targeting c-Myc in gastric cancer cells, resulting in a PTEN-dependent reduction of phospho-AKT. The observation that in DMPM cell lines more susceptible to the cytotoxic effects of miR-34a (STO, MP4, MP8), a decrease in phospho-AKT abundance is observed early after miRNA reconstitution would suggest the possibility that the above-described mechanism is also operating in our models. However, results of phenocopy experiments showing that siRNA-mediated silencing of c-MET and AXL was able to decrease AKT activation in sensitive cells (STO) but not in those less susceptible to miR-34a (MesoII) would suggest that the main mechanism controlling AKT phosphorylation status relies on RTK activity.
Interestingly, miR-34a induced a remarkable antitumor activity in the three cell lines (STO, MesoII and MP8) able to generate tumors following xenotransplantation into immunodeficient mice. Although to a different extent, miR-34a reconstitution significantly reduced the growth of the three s.c. xenograft models. Highly relevant to the disease, miRNA ectopic expression also impaired the growth of STO and MP8 orthotopic xenografts, which properly recapitulate the dissemination pattern in the peritoneal cavity of human DMPM [
11,
12], thus representing improved models to investigate novel therapeutic approaches. Specifically, miR-34a significantly inhibited the take of STO cells, with only one mouse developing small tumor nodules in the abdominal cavity. Although the miRNA did not influence the take of MP8 cells, a significantly reduced tumor growth was observed.
Unfortunately, the inability of MP115—the only biphasic DMPM model in our panel—to grow in vivo prevented us to assess whether the in vitro cytostatic effect consequent to miR-34a reconstitution, which was paralleled by the induction of a senescence-like phenotype possibly sustained by ATK activation, may result or not in tumor growth impairment. However, the significantly reduced invasive potential induced by miR-34a in DMPM cell lines through the inhibition of FAK signaling could primarily contribute to the antitumor effect observed in the xenograft models. Moreover, the occurrence of miR-34a-induced inhibition of cell invasion in the absence of appreciable antiproliferative and proapoptotic effects that we observed in MP115 is not surprising since the same phenotype has been previously reported by Li et al. [
44] for miR-34a reconstituted HepG2 hepatocellular carcinoma cells.
Interestingly, our evidence indicating that miR-34a ectopic expression impairs the secretion of angiogenesis-related factors by MesoII cells strongly suggests that the antitumor effect observed in both s.c. and ortothopic xenograft models can also rely on miRNA-induced modification of tumor microenvironment, making it less favorable to tumor growth.
In summary, the impressive inhibitory effects induced by miR-34a on DMPM cell proliferation, invasion, and growth in immunodeficient mice suggest a possible utility of the clinically available miR-34a as novel therapeutic option for DMPM patients who are not eligible for or relapse after CRS+HIPEC. In addition, the evidence that miR-34a reconstitution positively modulates the activity of antitumor drugs in experimental models of different human tumor types [
8,
45‐
47] highlights the possibility that the miR-34a mimic could have an important role also in combined strategies for treating DMPM patients.
Conclusions
DMPM is a rapidly fatal tumor with scanty therapeutic options. Here, we demonstrated for the first time that reconstitution of miR-34a in relevant models of the disease induced a significant antitumor effect, which mainly relied on c-MET and AXL down-regulation and impairment of their downstream signaling. In vivo results were complemented by in vitro data showing significant antiproliferative, proapoptotic, and anti-invasive activities. Taken together, our results provide evidence that (i) c-MET and AXL signaling pathways are critical determinants of DMPM cell survival, growth, and invasiveness and that miR-34a reconstitution can impair all these functions and (ii) persistent activation of AKT and ERK1/2 downstream signaling pathways represents a cytoprotective mechanism against miRNA-induced proapoptotic effects, though not preventing its anti-invasive activity, which instead mainly relies on FAK inhibition.
Overall, our preclinical data form a solid foundation that could promote the clinical translation of clinically available miR-34 mimic for the treatment of a still incurable disease such as DMPM and, on the other hand, provide the first evidence of a possible cytoprotective/resistance mechanism that may arise towards miRNA-based therapeutics.
Acknowledgements
Not applicable.