He detection threshold. Due to the selectin deficiency, the few leukemia cells escaping Rebaudioside A dormancy in the endothelial niche of the k.o. mice would again encounter great difficulties leaving the bloodstream to invade organs or establish chloromas. Future experiments will have to answer the question if (and then to what extent) a dormancy effect similar to the one described for HSCs [26] is involved in the effects of selectin deficiency in our xenograft model.E- and P-Selectin Essential in Leukemia XenograftFigure 7. Inhibition of selectin binding to human CEL and CML cells by 10457188 monoclonal antibodies and neuraminidase MK8931 treatment as determined by flow cytometry. A: Blocking of P-selectin binding by (pre)incubation with monoclonal antibodies against CD15s and CD162. FilledE- and P-Selectin Essential in Leukemia Xenograftcurves in the histograms represent incubation of the leukemia cells with isotype control, open curves represent incubation with the respective specific antibody. The only observed inhibitory effect was caused by anti-CD162 on EOL-1, but not on K562 cells. B: Blocking of E-selectin binding by (pre)incubation with monoclonal antibodies against CA19-9. Filled curves in the histograms represent incubation of the cells with isotype control, open curves represent incubation with the specific antibody. The only observed inhibitory effect by anti-CA19-9 was observed on PaCa 5061 pancreatic adenocarcinoma cells (positive control). C: Binding of selectins to the leukemia cells after neuraminidase treatment. Filled curves in the histograms: selectin binding without neuraminidase incubation (positive control); open curves, black: selectin binding to neuraminidase treated cells; open curves, grey: negative controls. Neuraminidase treatment abolished E-selectin binding to both cell lines and reduced P-selectin binding to EOL1 but not to K562 cells. All experiments were repeated twice, representative results are shown. doi:10.1371/journal.pone.0070139.gWith the exception of CD162, we were not able to identify the E- and P-selectin ligands on the surface of the leukemia cell lines used in our experiments by inhibition with monclonal antibodies specific for published selectin ligands. Only an anti-CD162 (PSGL-1) antibody inhibited P-selectin binding to EOL-1, but not to K562 cells. Surprisingly, antibodies specific for sialyl Lewis a (CA19-9) and sialyl Lewis x (CSLEX1) and the antibody HECA452 (recognizing both carbohydrate moieties) were unable to inhibit E- or P-selectin binding to EOL-1 and K562 cells (which had to be expected in the latter case as K562 cells are negative for sialyl Lewis a and x). In contrast, the antibody against sialyl Lewis a inhibited E-selectin binding to the pancreatic adenocarcinoma cell line PaCa 5061 (used as positive control), as described earlier [34]. It has been known for decades that the presence of sialyl Lewis x alone is not sufficient for a protein to function as a selectin ligand [38] and it has been shown recently that binding of HECA452 does not block simultaneous E-selectin binding to sialyl Lewis x microspheres [39]. Our results appear to verify the latter finding on a cellular level. It is unlikely, however, that the antibody (CSLEX1) against sialyl Lewis x that we used in this study is similar to HECA-452 in this respect (not able to block simultaneous selectin binding) as it has been shown to block Eand P-selectin binding on a cellular level before [40,41]. We could show that sialylated carbohydrate.He detection threshold. Due to the selectin deficiency, the few leukemia cells escaping dormancy in the endothelial niche of the k.o. mice would again encounter great difficulties leaving the bloodstream to invade organs or establish chloromas. Future experiments will have to answer the question if (and then to what extent) a dormancy effect similar to the one described for HSCs [26] is involved in the effects of selectin deficiency in our xenograft model.E- and P-Selectin Essential in Leukemia XenograftFigure 7. Inhibition of selectin binding to human CEL and CML cells by 10457188 monoclonal antibodies and neuraminidase treatment as determined by flow cytometry. A: Blocking of P-selectin binding by (pre)incubation with monoclonal antibodies against CD15s and CD162. FilledE- and P-Selectin Essential in Leukemia Xenograftcurves in the histograms represent incubation of the leukemia cells with isotype control, open curves represent incubation with the respective specific antibody. The only observed inhibitory effect was caused by anti-CD162 on EOL-1, but not on K562 cells. B: Blocking of E-selectin binding by (pre)incubation with monoclonal antibodies against CA19-9. Filled curves in the histograms represent incubation of the cells with isotype control, open curves represent incubation with the specific antibody. The only observed inhibitory effect by anti-CA19-9 was observed on PaCa 5061 pancreatic adenocarcinoma cells (positive control). C: Binding of selectins to the leukemia cells after neuraminidase treatment. Filled curves in the histograms: selectin binding without neuraminidase incubation (positive control); open curves, black: selectin binding to neuraminidase treated cells; open curves, grey: negative controls. Neuraminidase treatment abolished E-selectin binding to both cell lines and reduced P-selectin binding to EOL1 but not to K562 cells. All experiments were repeated twice, representative results are shown. doi:10.1371/journal.pone.0070139.gWith the exception of CD162, we were not able to identify the E- and P-selectin ligands on the surface of the leukemia cell lines used in our experiments by inhibition with monclonal antibodies specific for published selectin ligands. Only an anti-CD162 (PSGL-1) antibody inhibited P-selectin binding to EOL-1, but not to K562 cells. Surprisingly, antibodies specific for sialyl Lewis a (CA19-9) and sialyl Lewis x (CSLEX1) and the antibody HECA452 (recognizing both carbohydrate moieties) were unable to inhibit E- or P-selectin binding to EOL-1 and K562 cells (which had to be expected in the latter case as K562 cells are negative for sialyl Lewis a and x). In contrast, the antibody against sialyl Lewis a inhibited E-selectin binding to the pancreatic adenocarcinoma cell line PaCa 5061 (used as positive control), as described earlier [34]. It has been known for decades that the presence of sialyl Lewis x alone is not sufficient for a protein to function as a selectin ligand [38] and it has been shown recently that binding of HECA452 does not block simultaneous E-selectin binding to sialyl Lewis x microspheres [39]. Our results appear to verify the latter finding on a cellular level. It is unlikely, however, that the antibody (CSLEX1) against sialyl Lewis x that we used in this study is similar to HECA-452 in this respect (not able to block simultaneous selectin binding) as it has been shown to block Eand P-selectin binding on a cellular level before [40,41]. We could show that sialylated carbohydrate.