MiR-29b may suppresses peritoneal metastases through inhibition of the mesothelial–mesenchymal transition (MMT) of human peritoneal mesothelial cells

Peritoneal dissemination is a major metastatic pathway for gastrointestinal and ovarian malignancies. The miR-29b family is downregulated in peritoneal fluids in patients with peritoneal metastases (PM). We examined the effect of miR-29b on mesothelial cells (MC) which play critical a role in the development of PM through mesothelial-mesenchymal transition (MMT). Human peritoneal mesothelial cells (HPMCs) were isolated from surgically resected omental tissue and MMT induced by stimulation with 10 ng/ml TGF-β1. MiR-29b mimics and negative control miR were transfected by lipofection using RNAiMAX and the effects on the MMT evaluated in vitro. HPMC produced substantial amounts of miR-29b which was markedly inhibited by TGF-β1. TGF-β1 stimulation of HPMC induced morphological changes with decreased expression of E-cadherin and calretinin, and increased expression of vimentin and fibronectin. TGF-β1 also enhanced proliferation and migration of HPMC as well as adhesion of tumor cells in a fibronectin dependent manner. However, all events were strongly abrogated by simultaneous transfection of miR-29b. MiR-29b inhibits TGF-β1 induced MMT and replacement of miR-29b in the peritoneal cavity might be effective to prevent development of PM partly through the effects on MC.

www.nature.com/scientificreports/ reported to be downregulated in various tumors as compared with normal counterparts 23 . We consistently found that the miR-29 family in exosomes derived from peritoneal fluid was markedly downregulated in patients with peritoneal metastases from gastric cancer 24 . However, the role of miR-29s in the development of PM is not fully understood. In this study, we evaluated the effects of miR-29b on the cellular function of MC to examine its possible roles in the development of PM.

Results
Expression of miR-29b in HPMC was decreased by TGF-β1. HPMC were isolated from omental tissue obtained from patients who underwent sleeve gastrectomy and cultured in DMEM supplemented with 15% FBS. Digital PCR analysis revealed that HPMC produced substantial levels of miR-29b, which were more than those from gastric cancer cells, MKN45 and NUGC as well as peripheral blood mononuclear cells (PBMC) and mesenchymal stem cells (MSC) (Fig. 1A). However, when HPMC were cultured with 10 ng/ml TGF-β1 for 24 h, the expression of miR29b was markedly decreased (Fig. 1B).

Transfection of miR-29b mimic suppresses phenotypic changes of HPMC by TGF-β1. The
HPMC were transfected with miR-29b mimic or negative control miR (NC) using lipofectamine RNAiMAX. We found 50 nM miRs are enough for the transfection to more than 90% of the HPMC and expression levels of miR-29b was significantly upregulated with digital PCR system ( Supplementary Fig. 1).
After 48 h-culture with TGF-β1, HPMCs changed their morphology from round to spindle shape (Fig. 2). Stimulation with TGF-β1 significantly downregulated the expression of E-cadherin and calretinin, while markedly enhancing vimentin expression. However, lipofection of miR-29b mimics, but not with NC significantly suppressed the morphological changes with reduced expression of vimentin and restored the expression of E-cadherin and calretinin induced by TGF-β1 (Fig. 2). The changes of E cadherin and vimentin expression were confirmed in western blotting (Supplementary Fig. 2A).

Transfection of miR-29b mimic to HPMC suppresses the expression of fibronectin (FN).
Since adhesion data suggest a contribution of fibronectin (FN) in this adhesion cascade, we next examined the effect on the expression of FN on HPMC. FN was faintly detected on the surface of resting HPMC by immunohistochemistry analysis. However, after stimulation with TGF-β1, HPMC strongly expressed FN (Fig. 5). When miR-29b mimics were transfected with TGF-β1 stimulation, the expression of FN was not increased (Fig. 5). The change of FN expression was confirmed in western blotting ( Supplementary Fig. 2B).

MiR-29b mimic partially suppresses peritoneal metastases in vivo.
Finally, we examined the effects of local transfer of miR-29b mimics on PM in vivo using syngenic mouse model. When miR-29b mimics were mixed with atelocollagen and intraperitoneally injected, metastatic nodules of YTN16P2 on mesentery did not show marked difference (data not shown). However, the volumes of metastases on omental tissue were significantly decreased as compared with negative control (miR-NC) (Fig. 7).

Discussion
MC cover the inner surface of the peritoneal cavity and secrete large amounts of lubricant that facilitates intracoelomic movement and actively contributes to fluid transport, coagulation, fibrinolysis and local immunity 25 . MC are initially considered to be a mechanical barrier to tumor metastases in the peritoneal space 26,27 . However, recent studies have suggested that MC activated by TGF-β acquire a mesenchymal-like phenotype through the MMT which mediate attachment and enhance the migration of disseminated tumor cells to the sub-peritoneal space like cancer associated fibroblasts (CAFs) [28][29][30] .
In this study, we confirmed that HPMC acquire the phenotype of CAF with enhanced proliferative and migratory properties by TGF-β1 stimulation. However, all of the observed phenomena were largely abrogated by the transfection of miR-29b. Extensive studies have already shown that the miR-29 family regulates multiple oncogenic processes, including epigenetics, proteostasis, metabolism, proliferation, apoptosis, metastasis, fibrosis, angiogenesis, and immunomodulation 15,17 . In particular, miR-29 strongly suppresses the epithelial mesenchymal transition (EMT) of tumor cells 19,31 through targeting the expression of various genes including catenin, β1-integrin and matrix metalloproteinases (MMPs) 15,17 . Given that these EMT-related genes are largely shared with MC, the results of this study indicate that TGF-β1 induces MMT mainly through the downregulation of miR-29b.
Treatment with TGF-β1 markedly upregulated FN expression on an HPMC monolayer and induced robust adhesion of gastric cancer cells which was inhibited by an antibody to β1 integrin and RGD peptide. This is consistent with the results in previous studies showing that FN-β1 integrin is an important adhesion pathway between MC and tumor cell 30,32 . However, the enhanced adhesion by TGF-β1 was also abrogated by transfection of miR-29b. The miR-29 family has been shown to suppress the synthesis of extracellular matrix proteins such as collagens, laminin and fibrillin in embryonic fibroblasts 33 , trabecular meshwork cells 34 and hepatic stellate cells 35 . Zhu et al. showed that miR-29b downregulates FN production via Hsp47 in breast cancer cells 36 . The results of the present study are consistent with those results and suggest that transduction of miR-29b in MC effectively inhibits the attachment and subsequent migration of disseminated tumor cells on MC which is the initial step in the development of PM.
In previous studies, we found that the expression of exosomal miR-29b in peritoneal fluid was markedly reduced in patients who developed peritoneal recurrence after curative resection of gastric cancer 37 as well as    www.nature.com/scientificreports/ patients with PM 38 . Generally, expression of miR-29 in tumor cells decreased mainly via hypermethylation of the CpG areas of the miR-29 promoter lesion 39,40 , and the diminished expression of miR-29 is believed to promote tumor progression 16,17 . However, since the number of disseminated tumor cells is thought to be low during the initial step of the development of PM, it is unlikely that reduced miR-29b in peritoneal fluids is dependent on tumor cells. In this study, we found that resting HPMC produce substantial amount of miR-29b which was markedly suppressed by the addition of TGF-β1. This is consistent with previous studies showing that the TGF-β-SMAD signal suppresses miR-29 expression 41,42 and considerable levels of TGF-β could be detected in malignant effusions 43,44 . These facts suggest that reduced production of miR-29b in MC downregulates the amount of miR-29b in the abdominal cavity which is a prerequisite for the development of PM. Taken together, this suggests that replacement of miR-29b in peritoneal space may be effective to prevent PM partly through its effects on MC. In fact, intraperitoneal administration of miR-29b mimics using atelocollagen partially suppressed metastases, at least, in omentum. In vivo study using appropriate delivery methods is warranted.

Methods
Reagents. The oligonucleotides of miR-29b-3p mimics and negative control miRNA were synthesized by Thermo Fisher Scientific (Waltham, MA). The sequence of the oligonucleotides used for miR-29b-3pmimics.
All Isolation and culture of human peritoneal mesothelial cells (HPMC). HPMC were isolated from 2 to 4 cm 3 samples of omentum collected from consenting patients undergoing sleeve gastrectomy. Omental samples were momentarily placed in a half TrypLE Express (Thermo Fisher Scientific, Waltham, MA) with pure PBS, and incubated in a thermostatic tank at 37 °C for 2 h. The supernatants were collected after filtration by 100 µm-pore nylon mesh, were centrifuged at 1500 rpm and 4 °C for 5 min. Explants were seeded into collagen coated 10 cm 2 tissue culture dishes (Falcon, Becton Dickinson, Oxford, UK) and cultured in DMEM with 20% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin at 37 °C in 5% CO 2 cell culture incubator (Riera, McCulloch et al. 2006). Informed consent was obtained from all study participants.
Sample collection and miRNA purification and digital polymerase reaction (PCR). Cell line miRNAs were isolated from cell pellets using the MiRNeasy kit (Qiagen, Hilden, Germany), according to the manufactures' instructions. cDNA was synthesized starting from 100 μg of extracted RNA using the TaqMan miRNA Reverse Transcription Kit and miRNA-specific stem-loop primers (Applied BioSystems, Foster City, CA) following manufacturer's instructions. Total miR-29b-3p cell levels were then quantified using the digital (dd) PCR system (Bio-Rad Laboratories, Hercules, CA). Briefly, 10 μg of synthesized cDNA were added to a 20 μl PCR reaction mixture containing 10 μl of digital PCR™ supermix (Bio-Rad Laboratories), 1 μl of TaqMan primer/probe mix (Applied BioSystems) and RNase-free H2O. Droplets were generated by loading the mixture onto a plastic cartridge with 70 μl of QX100 Droplet Generation oil (Bio-Rad Laboratories). Cartridges were then placed into the QX200 Droplet Generator (Bio-Rad Laboratories). The droplets generated from each sample were transferred to a 96-well PCR plate (Eppendorf, Hamburg, Germany), and PCR amplification was carried out on the C1000 Touch Thermal Cycler (Bio-Rad Laboratories), according to the manufacturer's protocol. The plate was then loaded on the QX200 Droplet Reader (Bio-Rad Laboratories) and read automatically. The fraction of PCR-positive droplets was quantified assuming a Poisson distribution. QuantaSoft software was used to obtain the concentration results in number of copies per microliter for each sample.
Introduction of MMT and transfection of miR-29b-3p. HPMCs were seeded in 6-well plates at 50-60% confluence and induced with 10 ng/ml of TGF-β1. Then, the cells were transfected with miR-29b-3p mimic or negative controls with lipofectamine RNAiMAX in HPMC at a final concentration of 50 nM, and cultured for 48 h at 37 °C, according to the manufacturer's instructions and then used for the following experiments.
Immunofluorescence. HPMC (5 × 10 4 ) were plated in 24-well collagen coated plates, incubated with 10 ng/ ml of TGF-β1 and then transfected miR-29b or negative controls for 48 h. Cells were washed with PBS, fixed  In vivo experiment. YTN16 established in p53 heterozygous knockout C57BL/6 mice by oral intake of N-Methyl-N-nitrosourea (MNU) was a generous gift from Dr S. Nomura (Tokyo University) 45 . YTN16P2 was a highly metastatic a subline of YTN16 by in vivo selection method. The YTN16P2 (1 × 10 6 ) were intraperitoneally (IP) transferred in C57BL/6 mice and treated with IP local injection miR-29b or negative control (miR-NC) (2 μg) were mixed with atelocollagen (80 μl) at day 3, 6, 9, 12. The mice were sacrificed at day 14 and metastatic nodules on omental tissue or mesentery were counted. All procedures were approved by the Animal Care Committee of Jichi Medical University (No 20043-01) and performed in accordance with ARRIVE guidelines.
Statistical analysis. Data were represented as mean ± standard deviation. The significance of the differences between groups was assessed with a one-way ANOVA using GraphPad Prism8. Differences were considered significant when P < 0.05.