Institute of Anatomy and Experimental Morphology and University Cancer Center Hamburg (UCCH), University Medical-Center Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany
Dr. Daniel Wicklein
Institute of Anatomy and Experimental Morphology and University Cancer Center Hamburg (UCCH), University Medical-Center Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany
We investigated the functional role of CEACAM1 in a spontaneous metastasis xenograft model of human melanoma in scid mice using BRAF wildtype MeWo cells with and without RNAi mediated knockdown of CEACAM1. Tumors from the xenograft model were subjected to whole genome expression analysis and metastasis was quantified histologically. Results and identified markers were verified using tissue samples of over 100 melanoma patients. Knockdown of CEACAM1 prolonged the animals’ survival by significantly reducing subcutaneous growth of MeWo tumors and spontaneous lung metastasis. Microarray analysis revealed a strong influence of CEACAM1 knockdown on the network of EMT associated genes in the xenograft tumors (e.g. downregulation of BRAF, FOSL1, NRAS and TWIST). IGFBP7 and Latexin (highest up- and downregulated expression in microarray analysis) were found to be associated with longer and shorter survival, respectively, of melanoma patients. High FOSL1 and altered TWIST1 expression were found to be correlated with shortened survival in the cohort of melanoma patients. After a stepwise selection procedure combining above markers, multivariate analysis revealed IGFBP7, Latexin and altered TWIST to be prognostic markers for death. CEACAM1 could be a target for melanoma therapy as an alternative to (or in combination with) immune checkpoint and BRAF inhibitors.
8 November 2019 in Frontiers in Molecular Biosciences
Specificity of RNAi to selected target is challenged by off-target effects, both canonical and non-canonical. Notably, more than half of all human microRNAs are co-expressed with hosting them proteincoding genes. Here we dissect regulatory subnetwork centered on IGFBP6 gene, which is associated with low proliferative state and high migratory activity of basal-like breast cancer. We inhibited expression of IGFBP6 gene in a model cell line for basal-like breast carcinoma MDA-MB-231, then traced secondary and tertiary effects of this knockdown to LAMA4, a laminin encoding gene that contributes to the phenotype of triple-negative breast cancer. LAMA4-regulating miRNA miR-4274 and its host gene SORCS2 were highlighted as intermediate regulators of the expression levels of LAMA4, which correlated in a basal-like breast carcinoma sample subset of TCGA to the levels of SORCS2 negatively. Overall, our study points that the secondary and tertiary layers of regulatory interactions are certainly underappreciated. As these types of molecular event may significantly contribute to the formation of the cell phenotypes after RNA interference based knockdowns, further studies of multilayered molecular networks affected by RNAi are warranted.
Finding additional functional targets for combination therapy could improve the outcome for melanoma patients. In a spontaneous metastasis xenograft model of human melanoma a shRNA mediated knockdown of L1CAM more than sevenfold reduced the number of lung metastases after the induction of subcutaneous tumors for two human melanoma cell lines (MeWo, MV3). Whole genome expression arrays of the initially L1CAM high MeWo subcutaneous tumors revealed unchanged or downregulated genes involved in epithelial to mesenchymal transition (EMT) except an upregulation of Jagged 1, indicating a compensatory change in Notch signaling especially as Jagged 1 expression showed an increase in MeWo L1CAM metastases and Jagged 1 was expressed in metastases of the initially L1CAM low MV3 cells as well. Expression of 17 genes showed concordant regulation for L1CAM knockdown tumors of both cell lines. The changes in gene expression indicated changes in the EMT network of the melanoma cells and an increase in p53/p21 and p38 activity contributing to the reduced metastatic potential of the L1CAM knockdowns. Taken together, these data make L1CAM a highly interesting therapeutic target to prevent further metastatic spread in melanoma patients.
Extravasation of circulating tumor cells (CTCs) is critical for metastasis and is initiated by adhesive interactions between glycoligands on CTCs and E-selectin on endothelia. Here, we show that the clinically approved proteasome inhibitor bortezomib (BZM; Velcade) counteracts the cytokine-dependent induction of E-selectin in the lung mediated by the primary tumor, thereby impairing endothelial adhesion and thus spontaneous lung metastasis in vivo. However, the efficacy of BZM crucially depends on the tumor cells' E-selectin ligands, which determine distinct adhesion patterns. The canonical ligands sialyl-Lewis A (sLeA) and sLeX mediate particularly high-affinity E-selectin binding so that the incomplete E-selectin-reducing effect of BZM is not sufficient to disrupt adhesion or metastasis. In contrast, tumor cells lacking sLeA/X nevertheless bind E-selectin, but with low affinity, so that adhesion and lung metastasis are significantly diminished. Such low-affinity E-selectin ligands apparently consist of sialylated MGAT5 products on CD44. BZM no longer has anti-metastatic activity after CD44 knockdown in sLeA/X-negative tumor cells or E-selectin knockout in mice. sLeA/X can be determined by immunohistochemistry in cancer samples, which might aid patient stratification. These data suggest that BZM might act as a drug for inhibiting extravasation and thus distant metastasis formation in malignancies expressing low-affinity E-selectin ligands.
Endothelial E- and P-selectins promote metastasis formation by interacting with sialyl-Lewis X and A (sLeX/sLeA) on circulating tumor cells. This interaction precedes extravasation and can take place under dynamic and static conditions. Metastasis formation is often studied in xenograft models. However, it is unclear whether species differences exist in the ligand specificity of human (h) vs. murine (m) selectins and whether different ligands are functional under dynamic vs. static conditions. We systematically compared the h vs. m E- and P-selectin (ESel/PSel) binding of a range of human tumor cells under dynamic vs. static conditions. The tumor cells were categorized by their sLeA/X status (sLeA+/sLeX+, sLeA−/sLeX+ and sLeA−/sLeX−). The general biological nature of the tumor–selectin interaction was analyzed by applying several tumor cell treatments (anti-sLeA/X blockade, neuraminidase, pronase and inhibition of O/N-glycosylation). We observed remarkable differences in the static vs. dynamic interaction of tumor cells with h vs. m ESel/PSel depending on their sLeA/X status. The tumor cell treatments mostly affected either static or dynamic as well as either h- or m-selectin interaction. mESel showed a higher diversity of potential ligands than hESel. Inhibition of O-GalNAc-glycosylation also affected glycosphingolipid synthesis. Summarized, different ligands on human tumor cells are functional under static vs. dynamic conditions and for the interaction with human vs. murine ESel/PSel. Non-canonical selectin ligands lacking the sLeA/X glycan epitopes exist on human tumor cells. These findings have important implications for the current development of glycomimetic, antimetastatic drugs and encourage the development of immunodeficient mice with humanized selectins.
Induced overexpression of the secretory protein YKL-40 promotes tumor growth in xenograft experiments. We investigated if targeting YKL-40 with a monoclonal antibody could inhibit tumor growth. YKL-40 expressing human melanoma cells (LOX) were injected subcutenously in Balb/c scid mice. Animals were treated with intraperitoneal injections of anti-YKL-40, isoptype control or PBS. Non-YKL-40 expressing human pancreatic carcinoma cell line PaCa 5061 served as additional control. MR imaging was used for evaluation of tumor growth. Two days after the first injections of anti-YKL-40, tumor volume had increased significantly compared with controls, whereas no effects were observed for control tumors from PaCa 5061 cells lacking YKL-40 expression. After 18 days, mean tumor size of the mice receiving repeated anti-YKL-40 injections was 1.82 g, >4 times higher than mean tumor size of the controls (0.42 g). The effect of anti-YKL-40 on the increase of tumor volume started within hours after injection and was dose dependent. Intratumoral hemorrhage was observed in the treated animals. The strong effect on tumor size indicates important roles for YKL-40 in melanoma growth and argues for a careful evaluation of antibody therapy directed against YKL-40.