Canonical vs. non-canonical E-selectin ligands on solid human tumor cells differentially determine their E-selectin binding affinity: Functional and molecular implications for a novel anti-adhesive and anti-metastatic strategy

点击翻译按钮获取中文摘要

The lectin-glycan interaction between E-selectin (on vascular endothelial cells, ECs) and carbohydrate E-selectin ligands (on circulating tumor cells, CTCs) is one important step of pulmonary metastasis formation as it mediates the adhesion of CTCs from the blood stream to vascular endothelium and hence precedes CTC extravasation. Extravasation concomitantly eludes CTCs from adverse conditions existing within the circulation. Hence, blocking adhesion displays a tempting therapeutic approach. E-selectin expression on ECs depends on a pro-inflammatory cytokine environment, which physiologically allows the extravasation of leukocytes specifically into inflamed tissues. This pro-inflammatory environment in cancer is thought to be generated through a systemic release of cytokines. The pro-inflammatory transcription factors have to be activated by the proteasome. The prototype proteasome inhibitor could block this activation is bortezomib (BZM). BZM is already in clinical use for treating multiple myeloma patients. On this background our preliminary work shows that BZM actually reduces endothelial adhesion of tumor cells in vitro and in situ and - in proof-of-principle experiments - their metastasis formation in vivo. However, the efficacy of BZM depends on the respective E-selectin ligands expressed at the tumor cell surface. We dissected that human tumors can use either canonical or non-canonical E-selectin ligands for endothelial adhesion and that this determines their E-selectin binding affinity and the potential requirement of additional CAMs for endothelial adhesion. Only tumor cells that express low affinity, non-canonical E-selectin ligands were susceptible to BZM, but not tumor cells that express high affinity, canonical E-selectin ligands. At first, it is necessary to analyze whether BZM shows divergent anti-metastatic efficacy on additional xenograft tumors in vivo. Secondly, the anti-adhesive mechanism of proteasome inhibition should be clarified in more detail since BZM was observed to down-regulate not only E-selectin, but also ICAM-1 and endothelial chemokines. Thirdly, we will determine whether the observed anti-metastatic effect of BZM is actually due to impairing CTC adhesion to pulmonary microvascular endothelium (for additional cell lines). For this purpose, we will use an established ex vivo lung perfusion model enabling us to directly observe tumor cell adhesion in the murine lung microcirculation. We will also investigate whether the presence of subcutaneous primary tumors promotes CTC adhesion at a distant site in situ. Finally, we will investigate how the endothelial adhesion of BZM-resistant, highly E-selectin-affine tumors can be targeted. Therefore, we aim to determine the functional relevance of the glycoprotein carriers and glycosyltransferases identified in BZM-resistant cell lines for their adhesion in vitro and in situ as well as metastasis formation in vivo.