Chemoresistance induced by mesenchymal stromal cells on cancer cells

Abstract

Human mesenchymal stromal cells (hMSCs) are part of bone marrow micro-environment that supports hematopoiesis. However, hMSCs also enhance tumor progression and survival when they become part of the cancer micro-environment. I aimed to investigate the interaction between hMSCs and cancer cells during chemotherapy.

Firstly, I studied the interaction between hMSCs and T-lineage acute lymphoblastic leukemia (T-ALL) cells under pegylated arginase I (BCT-100) treatment. Three T-ALL cell lines were sensitive to BCT-100 but not hMSCs. Conversely, hMSCs could partly protect all T-ALL cell lines from BCT-100 induced cell death under transwell co-culture condition. Concerning the possible mechanism, the intermediate metabolite L-ornithine could not rescue most T-ALL cells from BCT-100 treatment. But the downstream L-arginine precursor, L-citrulline could partly rescue all T-ALL cells from BCT-100 treatment. Ornithine transcarbamylase (OTC) converts L-ornithine into L-citrulline. OTC expression level in hMSCs remained relatively high during BCT-100 treatment but OTC expressions in T-ALL cell lines declined drastically. It suggested that hMSCs may protect T-ALL cells against BCT-100 treatment by having sustained OTC expression. Suppression of hMSCs by vincristine (VCR) disrupted the protective effect of hMSCs to most T-ALL cells during BCT-100 treatment. This suggests that by transiently suppressing hMSCs, we may abolish the protective effect of hMSCs to T-ALL cells during BCT-100 treatment.

Then I studied the interaction between hMSCs and neuroblastoma under cisplatin treatment. Two neuroblastoma cell lines were used for both of them are cisplatin sensitive while hMSCs are cisplatin resistant. hMSCs could partly protect neuroblastoma cells from cisplatin induced cytotoxicity. On the other hand, exogenous IL-6 but not IL-8 could also partly rescue them from cisplatin induced cytotoxicity. IL-6 activated STAT3 phosphorylation dose-dependently and enhanced expression of detoxifying enzyme (glutathione S-transferase π, GST-π) in neuroblastoma. Such effect could be counteracted by anti-IL-6R neutralizing antibody tocilizumab (TCZ). However, TCZ failed to suppress hMSCs’ protection to neuroblastoma during cisplatin treatment. This suggests involvement of multiple factors. Up-regulation of serum GST-πin some hTertMSCs/neuroblastoma co-engrafted SCID mice compared to neuroblastoma engrafted mice provided a clue that GST-π might be a possible stromal-protection factor. Caffeic acid phenethyl ester (CAPE) is a known GST inhibitor after tyrosinase activation. Neuroblastoma cells expressed tyrosinase and CAPE enhanced cisplatin cytotoxicity on them, with or without hMSCs. Paradoxically, CAPE enhanced GST-πexpression with or without cisplatin treatment in neuroblastoma suggesting possible negative feedback to GST-π inhibition. However, such additive effect of CAPE to cisplatin cytotoxicity was not observed in vivo. Further delineation of the in vivo study design may help to verify the additive effect of CAPE to cisplatin cytotoxicity in vivo.

Finally, I studied the effect of apoptotic cancer cells (AC) on the immune function of hMSCs. hMSCs could phagocytose apoptotic neuroblastoma cells with respective up-regulation of many immune-mediators including two highly-expressed cytokines IL-6 and IL-8. Up-regulation of these immune-mediators may enhance immune cells chemotaxis. Further detailed investigation on the effect of AC-engulfed hMSCs to other immune cells will help us to understand the dynamic interaction between cancer cells and stromal cells during chemotherapy.