br COMP leads to activation of Notch br Fig
COMP leads to activation of Notch3 117
Fig. 6. COMP expression in spontaneous breast cancer MMTV-PyMT model is required for maintenance of cancer stem cell population. (A) Mammary gland tumor volume was assessed in all mice and is presented as an illustrated scatter plot (B) Scatter plot of cancer stem IMP-1088 population percentage, displaying all mice included in the experiment. (C) Representative flow cytometry dot plot of dissociated tumor cells stained for CD24 and Thy 1 (CD90) and their corresponding antibody isotype controls. (D) Western blot analysis of tumor lysates, immunodetecting Notch3. Shown are representative samples of 2 mice per group. (E) Correlation between the presence of cancer stem cells and the intensity of Notch3 intracellular domain in western blot analysis of tumor lysates. Spearman's rank correlation coefficient was used. All experiments were repeated at least 3 times. One-way ANOVA, Bonferroni's multiple comparisons test was used (*b0.05, **b0.01, ***b0.001, ****b0.0001). NICD: Notch intracellular domain, NTM: Notch truncated protein.
“protective” effect. Strikingly, we found that COMP+/+ tumors contained 10-times more cancer stem cells than COMP−/−. Since there is no universally accept-
ed combination of stem cell markers for mouse models, we used a combination of all previously described markers (CD24, CD133, CD29, CD61,
118 COMP leads to activation of Notch3
Thy1 (CD90))  in flow cytometry analysis of excised tumors. The combination of CD24 and Thy1 (CD90) worked best to detect cancer stem cells in our mouse model, in line with previous reports [39,40]. Furthermore, a similar effect of COMP on the stem cell population was observed in vitro using human breast cancer cell lines, which formed larger tumorspheres when expressing COMP. Further-more, COMP-expressing cells had a larger propor-tion of ALDH+ cells and expression of CD133 +, both of which are well documented indicators of cancer stem cells [34,35].
To search for signaling pathways responsible for the enhanced tumor growth in the presence of COMP, we turned our attention to the Notch pathway, which is dysregulated in many solid tumors. More specifically, we assessed the potential role of Notch3, which has emerged as an important oncogenic factor in breast
cancer and where recent studies have revealed the fundamental role of Notch3 in cancer stem cells [9–12]. Indeed, Notch3 activation, as measured by the cleavage of the Notch3 receptor, was higher in the COMP+/+ background. Complementing these obser-vations, cancer stem cell frequency was reduced in vitro when Notch activation was abrogated by the Notch inhibitor DAPT, which prevents the release of the Notch intracellular domain from the plasma membrane through inhibiting its cleavage by γ-secretase . These results are consistent with our previous observation of Notch pathway upregulation detected in mRNA isolated from COMP-expressing tumors . Mechanistically, our data show that COMP interacts with both Notch3 and Jagged1 and importantly, COMP expression increases the interac-tion between Notch3 and Jagged1 (Fig. 7). Theoret-ically, the increased interaction can occur between
Fig. 7. Schematic representation of the proposed molecular mechanism. COMP acts as an adaptor protein mediating the interaction between Notch3 and Jagged1. This leads to the transcriptional activation of related genes, initiating a cancer stem cell genetic program of the cells. Notch3 also cross-talks with other molecular pathways that are stemness related (β-Catenin, AKT).
COMP leads to activation of Notch3 119
Notch3 and Jagged1 expressed in the same cell (in cis) or when expressed on juxtaposed cells (in trans). As the COMP-mediated increase in interaction is accompanied by enhanced Notch3 receptor cleavage and activation, it seems likely that COMP increases ligand-receptor interaction in trans, as this leads to Notch receptor activation, while cis-interactions are inhibitory . Furthermore, COMP-mediated activa-tion of Notch3 downstream signaling was accompa-nied by reduced Jagged1 levels, suggesting that Jagged1 along with the Notch3 extracellular domain were taken up by the signal-sending cell and degraded. Mechanistically, the increased interaction may be accomplished by polymeric COMP forming clusters of ligands and/or receptor, as ligand cluster-ing is known to be required for productive trans-activation . Alternatively, COMP may in some way facilitate the direct ligand-receptor interaction in trans, and as high-resolution crystal structure for Jagged-Notch binding has been determined , such an interaction could be structurally modelled. Interesting-ly, this mode of Notch stimulation may not be restricted only to COMP but may be achieved also by other members of the thrombospondin (TSP) family, of which COMP (TSP5) is a member. In line with this reasoning, thrombospondin 2 (TSP2) is known to facilitate the Notch3-Jagged1 interaction , where-as it serves as a negative regulator of the Notch2-Jagged1 interaction in the context of Alagille syn-drome, which is caused by Jagged1 or Notch2 mutations . The reason TSPs exert different effects on different Notch receptors remain to be established, and the role of the thrombospondin family in the regulation of Notch signaling will require further studies.