In this study we illustrated the important role of MSC in early OCC invasion with a 3D model of metastatic nodule based on an amniochorionic membrane scaffold. We investigated the interaction between OCC and MSC and found that IL6 was determinant for OCC to migrate and infiltrate 3D structures resembling the peritoneum.
Peritoneal mesothelium is the first barrier against spreading ovarian cancer cells. OCC adhere to mesothelial cells via integrins after MMP2-mediated digestion of vitronectin and fibronectin . Previous studies showed that the mesothelium could not be detected under the proliferating OCC implants suggesting that mesothelial cells are dissociated before peritoneal involvement [17, 18]. After breaking peritoneal mesothelial layer, OCCs can invade the sub-mesothelial area, constituted by an extra-cellular matrix, and various cell types including MSCs and fibroblasts. Tumoral secreted factors such as leucine, leucine-37 (LL-37) or lysophosphatidic acid (LPA) could recruit MSCs and induce their differentiation in cancer-associated fibroblasts (CAF) through signalling pathways involving Rho kinase, ERK, PLC, and phosphoinositide-3-kinase [19, 20]. LPA present in the EOC microenvironment was also reported to induce STAT3 phosphorylation and ovarian cancer cell motility through the secretion of IL-6 and IL-8 . This is in line with the increased invasion of OCCs through the AMS after stimulation with recombinant IL6.
Several authors [22–24] showed that MSCs promoted tumor growth through increased micro-vascularization, stromal networks, and production of tumor stimulating paracrine factors. They also demonstrated that these properties were activated after Mesenchymal Stem Cell Transition to Tumor-Associated Fibroblasts, through the paracrine secretion of IL6. We observed increased production of IL6 among other cytokines in co-cultures of OCCs and MSCs indicating cross-talk between the two cell lines. Mc Lean et al. demonstrated differences between MSCs and tumor associated MSCs . They reported an enhance ovarian cancer stem cell compartment upon the interaction of cancer cells with Tumor associated mesenchymal cells compare to“wild type” MSCs . Similarly Liu et al.  reported that MSCs could support breast cancer stem cell compartment through IL6 and IL8 secretion. In concordance we found an increased number of MSCs in most OCCs nodules within the AMS suggesting an important role in the early invasion for the constitution of an inflammatory reactive stroma: “never healing wound theory” [27, 28]. As illustrated above the cross-talk between MSCs and cancer cells and the role of mesencrine factors increasing the metastatic potential have been widely illustrated. Using a unique 3D model based on an amniotic membrane scaffold we were able to replicate the early invasion steps, e.g. (i) formation of tumor aggregates, (ii) adhesion to mesothelial layer (iii) microscopic invasion. We illustrated that MSCs could also play a role in very early attachment and invasion of ovarian cancer nodules. Indeed the significant enrichment of MSCs around invasive nodules suggested their ability to provide the adequate signalling cues for attachment and invasion of ovarian cancer aggregates.
IL6 has been associated with progression in multiple cancer types, including ovarian cancer. Increased expression of IL6 and its specific receptor IL6Rα is associated with disease stage . Recent works also reported the importance of IL6 for early metastasic process in EOC. We found that IL6 inhibition limited early adhesion and infiltration of OCCs in our in-vitro 3D model. Giridhar et al. showed that IL6 regulated in-vivo adhesion of OCCs to the omentum through up-regulation of LY75 . Using mice with conditional IL6Ra deficiency, they found that host IL6 regulation was important for OCC adhesion .
In many cancer IL6 has been described as been up-regulated . IL6 signaling occurs through a hexameric complex including specific IL6-alpha receptor (IL6R; glycoprotein gp 80), and a beta-signaling receptor (gp130) . Dimerization of the receptor induces to the activation of Janus Tyrosine Kinase signaling and the translocation of a signal transducer and activator of transcription (STAT) transcription factors . IL6 mainly signals through STAT3 which translocation to the nucleus induces a complex transcriptional program resulting in Inflammation, cell survival, differentiation or prometastatic properties depending the cellular context [34, 35].
There are some limitations in our study. We used AM instead of peritoneum because it was uneasy to obtain and ethically difficult to justify regarding the quantity necessary for our study. The large resections of peritoneum mainly occur in patients undergoing debulking surgery for ovarian cancer and therefore could modify inflammation state of the tissue and bias the experiments. Our model also suffers from the fact that we couldn’t affirm by which side OCC invaded the AMS. However, our goal was to investigate the relations between OCC and MSC, and both amniotic and chorionic membrane is rich in MSC in the same proportion. Finally, we didn’t define the origin of IL6 secretion. We showed increased IL6 secretion in co-cultures of OCCs and MSCs but we did not identify if MSCs, OCCs or both secreted it. Spaeth et al. co-injected Skov3 with and without MSCs into mice and demonstrated that MSCs stimulated tumor growth through paracrine production of IL6 . They found MSCs-induced IL-6 secretion to be critical for the enhanced proliferation observed in Skov-3/MSC tumor growth assay . We showed that IL6 stimulation increased OCC mobility and invasiveness and inhibiting IL6 receptor decreased OCCs infiltration in a 3D model. Colomiere et al. reported increased epithelial to mesenchymal transition (EMT) after EGF treatment of OCCs . In their model OCCs secretion of IL6 was increased upon EGF stimulation. Several authors have demonstrated similar findings in other models [38, 39]. In our model, increased EMT and/or cancer stem cell compartment upon IL6 stimulation could explain increased invasiveness and this remains to be investigated.