Y anti-CD162 (not shown). As we were unable to confirm that the described selectin Title Loaded From File ligands sialyl Lewis a and sialyl Lewis x played any role in selectin binding to the CEL or CML line used in our experiments, we treated theE- and P-Selectin Essential in Leukemia Xenografttation was: wt 48.5 days, k.o. not reached. The curves were significantly different, **: P = 0.0085 (Log-rank test). B: Human K562 CML cells (given in cells/ml) in the animals’ blood at the time of death as determined by quantitative real-time PCR (qRT-PCR). Selectin competent (wt, n = 10) compared with E- and P-selectin deficient mice (k.o., n = 10). Given in the box plot are 15481974 the median (line), highest and lowest number of K562 cells per ml of the animals’ blood (whiskers) and upper and lower quartile (box). Median cell number per ml blood was 268.8 for the wt group and 0.2 for the k.o. group. The difference between the groups was significantly different, **: P = 0.0068 (Mann Whitney test). C: Human K562 CML cells in the animals’ bone marrow at the time of death as determined by qRT-PCR (given in cells/60 ng template DNA). Given in the box plot are the median (line), highest and lowest number of K562 cells (whiskers) and upper and lower quartile (box). Median cells per 60 ng template DNA in the wt group were 243 compared with 0 cells in the k.o. group. This difference was significant, **: P = 0.0089 (Mann Whitney test). doi:10.1371/journal.pone.0070139.gcells with neuraminidase (sialidase) to investigate whether sialylated carbohydrate moieties were involved in selectin binding at all. Treatment with neuraminidase abolished E-selectin binding to both EOL-1 and K562 cells and considerably reduced Pselectin binding to EOL-1 cells. However, P-selectin binding to K562 cells was not affected by neuraminidase digestion (Figure 7C).DiscussionThis study was undertaken to analyze the functional role of Eand P-selectin in the process of leukemic dissemination in CML and CEL. Both cell lines used bound to E- and P-selectin fusion proteins: This binding was more intense in EOL-1 cells (moderate to E- and strong to P-selectin) compared to K562 (weak to E- and moderate to P-selectin) and interacted with the selectins under laminar flow conditions. The pathophysiological role of selectin binding could be verified in vivo: E- and P-selectin deficient mice showed dramatically increased survival and little organ infiltration and chloroma formation were observed compared with wt mice. Additionally, there were only few to no circulating leukemia cells in the selectin deficient animals’ blood. These observations indicate that E- and P-selectin play an important role in organ infiltration and chloroma formation by CEL and CML cells. Flow cytometric analysis in earlier Title Loaded From File xenograft experiments with EOL-1 cells showed that nearly all cells vanish from the murine bloodstream after intravenous injection and reappear in the blood about 28 days later [28]. Thus, the human leukemia cells obviously left the bloodstream to home into at least one type of survival niche in the murine organism (in the bone marrow and/or other organs). Whether this niche is similar or identical to the Leukemia Stem Cell (LSC) niche [35,36] and whether the leukemia cells can establish LSCs in the animals, is a highly pertinent question that can be studied in vivo in this model. However, it should be stated that the model applied here employed human cell lines and that cell line-specific LSC may differ from primary LSC in pa.Y anti-CD162 (not shown). As we were unable to confirm that the described selectin ligands sialyl Lewis a and sialyl Lewis x played any role in selectin binding to the CEL or CML line used in our experiments, we treated theE- and P-Selectin Essential in Leukemia Xenografttation was: wt 48.5 days, k.o. not reached. The curves were significantly different, **: P = 0.0085 (Log-rank test). B: Human K562 CML cells (given in cells/ml) in the animals’ blood at the time of death as determined by quantitative real-time PCR (qRT-PCR). Selectin competent (wt, n = 10) compared with E- and P-selectin deficient mice (k.o., n = 10). Given in the box plot are 15481974 the median (line), highest and lowest number of K562 cells per ml of the animals’ blood (whiskers) and upper and lower quartile (box). Median cell number per ml blood was 268.8 for the wt group and 0.2 for the k.o. group. The difference between the groups was significantly different, **: P = 0.0068 (Mann Whitney test). C: Human K562 CML cells in the animals’ bone marrow at the time of death as determined by qRT-PCR (given in cells/60 ng template DNA). Given in the box plot are the median (line), highest and lowest number of K562 cells (whiskers) and upper and lower quartile (box). Median cells per 60 ng template DNA in the wt group were 243 compared with 0 cells in the k.o. group. This difference was significant, **: P = 0.0089 (Mann Whitney test). doi:10.1371/journal.pone.0070139.gcells with neuraminidase (sialidase) to investigate whether sialylated carbohydrate moieties were involved in selectin binding at all. Treatment with neuraminidase abolished E-selectin binding to both EOL-1 and K562 cells and considerably reduced Pselectin binding to EOL-1 cells. However, P-selectin binding to K562 cells was not affected by neuraminidase digestion (Figure 7C).DiscussionThis study was undertaken to analyze the functional role of Eand P-selectin in the process of leukemic dissemination in CML and CEL. Both cell lines used bound to E- and P-selectin fusion proteins: This binding was more intense in EOL-1 cells (moderate to E- and strong to P-selectin) compared to K562 (weak to E- and moderate to P-selectin) and interacted with the selectins under laminar flow conditions. The pathophysiological role of selectin binding could be verified in vivo: E- and P-selectin deficient mice showed dramatically increased survival and little organ infiltration and chloroma formation were observed compared with wt mice. Additionally, there were only few to no circulating leukemia cells in the selectin deficient animals’ blood. These observations indicate that E- and P-selectin play an important role in organ infiltration and chloroma formation by CEL and CML cells. Flow cytometric analysis in earlier xenograft experiments with EOL-1 cells showed that nearly all cells vanish from the murine bloodstream after intravenous injection and reappear in the blood about 28 days later [28]. Thus, the human leukemia cells obviously left the bloodstream to home into at least one type of survival niche in the murine organism (in the bone marrow and/or other organs). Whether this niche is similar or identical to the Leukemia Stem Cell (LSC) niche [35,36] and whether the leukemia cells can establish LSCs in the animals, is a highly pertinent question that can be studied in vivo in this model. However, it should be stated that the model applied here employed human cell lines and that cell line-specific LSC may differ from primary LSC in pa.
