Complex Interactions Between Osteosarcoma Cells and Fibroblasts

Abstract

This thesis investigates interactions between neoplastic cells in tumours, and the stromal-host cells with which the neoplastic cells must cooperate in order to survive. Earlier work upon which this thesis is based, established contact dependent induction of apoptosis in endothelium by SAOS-2 osteosarcoma cells (McEwen et al., 2003). This was quantitated by detecting a reduction in endothelial cell culture density (McEwen et al., 2003), but when similar experiments were performed with human gingival fibroblasts (HGF), no apoptosis was detectable despite there being an apparent reduction in HGF cell culture density (Huynh, 2007). This thesis confirms that the seeming disappearance of HGF co-cultured with SAOS-2 can be accounted for by transfer of SAOS-2 membrane marker to HGF, and explores some aspects of how the microenvironment determines specific cellular interactions between neoplastic and stromal cells. Chapter 1 summarizes background literature relevant to the work described in the remaining thesis. Chapter 2 investigates the necessary first step of cell adhesion in interaction between SAOS-2 and HGF. An adhesion assay was established in which SAOS-2 are applied to monolayers of HGF, and gently washed prior to cell counts. SAOS-2 binding to HGF was markedly increased by pre-treatment of HGF with TNF-α in a dose dependent manner, and as was seen in Chapter 3, this was found to be mediated by ICAM-1 expressed by HGF. Surprisingly, SAOS-2 pre-treated with TNF-α had reduced binding to TNF-α treated HGF, indicating that for this interaction, SAOS-2 must have a ‘permissive cellular history’ in which they have not been stimulated by TNF-α, and this was recognized as having implications regarding similar binding events in-vivo, where neoplastic cells at the invading front of tumours migrate from one micro-environment to another. Chapter 3 examines the effect of TNF- pre-treatment upon the ability of SAOS-2 to reduce the apparent culture density of HGF. Reduction in the apparent culture density of HGF was seen in broad proportion to the increased binding of SAOS-2 to HGF mediated by TNF-α. Also, work in this chapter excluded the possible contribution of autophagy to loss of HGF in co-culture with SAOS-2. Chapter 4 describes studies in which HGF and SAOS-2 were pre-labelled with a range of fluorescent markers for plasma membranes, cytoplasm and nuclear material. Co-culture of these cells was accompanied by the appearance of dual labelled cells, while quantitative analysis was performed to assess patterns of label exchange across cell types. Exchange of both plasma membrane and cytoplasmic markers was observed in these cultures, although plasma membrane marker exchange was more prevalent as judged by the proportion of dual labelled cells seen. No cells with both nuclear labels were found. The direction of label exchange was strongly influenced by the specific labels used in given experiments, so that it was not possible be certain whether bulk transfer of membrane or cytoplasm had occurred. Nonetheless, it was possible to conclude with reasonable confidence that continuity in plasma membranes and cytoplasm was established between HGF and SAOS-2, sufficient for the exchange of significant quantities of labelled proteins in the case of cytoplasm, as well as for membrane embedded alkaline phosphatase from SAOS-2 to HGF and lipophilic membrane dyes across both cell types. This ‘cellular sipping’ by SAOS-2, was suggested as a potential mechanism overcoming the effects of a progressively deranged neoplastic genome. It was also recognized that the generation of populations of cells expressing proteins of mixed origin would increase tumour diversity, with potential to affect tumour progression. In addition, quantitative morphological assessment was performed of cell circularity and cell surface area profile, as an indirect measure for phenotypic state, and the finding that dual labelled cells expressed morphological properties intermediate to HGF and SAOS-2 suggested phenotypic impact of cellular sipping. The influence of TNF-α was also studied in this chapter, which was found to increase the extent of cellular sipping. In Chapter 5, the effect of co-culture of SAOS-2 upon HGF pre-treated with TNF-α with regard to cytokine synthesis was examined in both: direct co-culture; and also in co-cultures where SAOS-2 and HGF were separated by transwell membranes. SAOS-2 made negligible contribution to levels of any of the cytokines studied (IL-6, GM-CSF, G-CSF, FGF). However, SAOS-2 did increase production by TNF-αpre-treated HGF of IL-6, GM-CSF, and G-CSF in transwell co-cultures, despite having the opposite effect in direct co-cultures where SAOS-2 were in intimate contact with HGF. In summary, results in this thesis demonstrate that SAOS-2 and HGF interact intimately, through the establishment of continuity between plasma membranes and cytoplasm, and that the resulting transfer of membrane and cytoplasmic material produces four ultimate populations of cells in co-culture being: HGF; SAOS-2; HGF with SAOS-2 membrane and cytoplasmic elements; and SAOS-2 with HGF cytoplasmic and membrane elements. Furthermore, a role for TNF-α in facilitating HGF and SAOS-2 binding via ICAM-1, as well as in mediating exchange of material between cells is demonstrated. The complexity of these interactions is underscored by the opposing effect upon SAOS-2 binding to HGF of pre-treatment of either cell type with TNF-α, as well as by the opposite cytokine synthetic profile of HGF co-cultured in direct contact with SAOS-2, as opposed to separated by a transwell membrane. These observations support the idea of there being ‘permissive microenvironments’ for a range of interactions between stromal and neoplastic cells, as well as the idea of a ‘permissive cellular history’ for some responses. The paracrine and contact dependent interactions identified support subversion of stromal cell function in support of neoplastic cells. Also, cellular sipping by neoplastic cells seems to provide a novel mechanism through which malignant neoplastic cells with irreparably damaged genomes may be able to gain survival advantage by capturing normal cytoplasmic and membrane elements from surrounding stromal cells. In addition, findings relating to permissive microenvironments, permissive cellular histories, and the generation of complex cell populations through cellular sipping, provide further explanation of architectural and cytological pleomorphism in malignant neoplasms, till now attributed primarily to the emergence of divergent sub-clones of neoplastic cells through the accumulation of genetic lesions.