Abstract
Mesenchymal stem cells (MSCs) are bone marrow derived, non-hematopoietic stem cells, which home to sites of injury and inflammation. They have also been identified at tumour sites where they can have pro- or anti-tumourigenic effects. It is unclear which specific signalling molecules and proteases secreted by the tumour cells are responsible for mobilising MSCs to the tumour sites. However, as MSCs migrate to tumour sites, they could be utilised for cancer therapy, as engineered MSCs would be able to precisely deliver gene or protein treatment to tumour sites.
Here, the role of a protease, cathepsin D, in the homing of MSCs to tumour sites is described through in vitro studies. MDA-MB-231 and HT29 cells or their conditioned media were co-cultured in Boyden chambers with MSCs with or without protease inhibitors to analyse the migration and invasion of the MSCs. The migration and invasion of MSCs was induced by the cancer cells and their 3 day conditioned media. Since increased levels of proteases play a major role in regulation of the tumours, we focussed on identifying proteases which attract MSCs to the tumour cells. Migration was decreased with protease inhibitors pepstatin A and GM6001 in both the cell lines. Pepstatin A is a potent inhibitor of cathepsin D. Furthermore, cathepsin D expression was strongly detected in tumour cells when compared to the MSCs. To confirm the effects of cathepsin D, its expression was inhibited by siRNA which also resulted in a decrease in the migration and invasion of MSCs. The signalling pathway MAPK/ERK was found to be activated in MDA-MB-231 and HT29 cells and addition of pepstatin A reversed this activation to a significant level.
Wild-type FOXL2 and tissue inhibitor metalloproteinases (TIMPs) proteins were trialled to target cancer through engineered MSCs. To target adult type ovarian tumour granulosa (KGN cells), which is characterised by a single missense point mutation of the FOXL2 gene, the wild-type FOXL2 gene was overexpressed in KGN cells and MSCs using a tdTomato expressing lentiviral vector. Cisplatin and etoposide, the first line treatment drugs, were tested on KGN and MSC cells with and without wild-type FOXL2 expressing genes. The cancer cells co-culture incubated with MSCs expressing wild-type FOXL2 showed IC50 values of almost half when compared to cells with the mutated gene. Cell cycle distribution, DNA fragmentation and mitochondrial cell death evaluation showed significant G2/M arrest, DNA cleavage and mitochondrial permeabilisation in wild type expressing KGN cells when compared to mutant gene expressing KGN control cells. Angiogenesis inhibition by MSCs, demonstrated using HUVECs in tube formation assays, was also increased by co-culture with MSCs expressing the wild-type FOXL2 gene, compared to MSC-only controls.
TIMPs are the endoproteinases that inhibit matrix metalloproteinases (MMPs), which are partly responsible for tumour metastasis. TIMP genes were successfully cloned into a tdTomato-expressing vector and overexpressed in MSCs by lentiviral infection.
These findings provide the first evidence that cathepsin D is an important protease that plays a role in attracting MSCs to tumours. We demonstrated that cancer cells secreted cathepsin D which increases migration and invasion of MSCs and that the ERK signalling pathway may be involved in this MSC migration by cancer cells. In addition, we showed that wild-type FOXL2 gene expressing KGN cells were more susceptible to the drugs cisplatin and etoposide than the cells carrying only the mutated form of the gene, suggesting that it may be clinically useful to deliver the protein using MSCs as part of the treatment regime. However more in vitro and in vivo studies are required to confirm and optimise these results. We have therefore further explored the homing of MSCs to tumour sites and taken important early steps in developing targeted therapies for cancers that could be delivered by MSCs.