6B). In CXCR3− NK cells, CD27+ NK cells displayed slightly stronger IFN-γ production than CD27− NK cells, whereas in CXCR3+ NK cells no difference was detected between CD27− and CD27+ NK cells. CD27−/dim/bright NK cells appeared in the CXCR3+ subset after stimulation of

the NK cells, which downregulated CD27 expression (see also Fig. 3C). Induction of IFN-γ was also detected upon contact with YAC-1 cells as assessed by the CD107a assay (data not shown). In general, CXCR3 expression correlated PD0325901 research buy positively with IFN-γ, TNF-α, and MIP-1α production. We did not detect any cytokine production in unstimulated NK cells (data not shown). In humans, CD56dim and CD56bright NK cells represent functionally distinct subsets 9, 12, 13. In contrast, mouse NK cells express neither CD56 nor a correlate, which limits investigations of extrapolations of murine data to the human system. Thus, the definition and characterization of NK-cell subsets in mice is a major topic of current NK-cell research. Recently, markers such as CD94 or CD27 were proposed as potential markers for murine NK-cell subsets corresponding to the human CD56dim and CD56bright

paradigm 23, 32. Based on microarray gene analyses, we previously demonstrated the almost exclusive coexpression of CXCR3 (CD183) on human CD56bright NK cells, and we suggest this molecule to allow comparisons between human and mouse NK-cell subsets 15, 29. check details In this study, CXCR3 expression, and particularly coexpression of CD27 on murine NK cells, was analyzed in order to determine the optimal marker constellation to define a murine NK-cell subset. The percentages of NK-cell subsets in humans and mice vary considerably among the compartments. For instance, in humans 90% of circulating and 85% of splenic NK cells are CD56dimCD16bright, whereas in LN up to 90% of NK cells display a CD56brightCD16−/dim phenotype 18, 33. In mice, we also detected higher percentages of CXCR3+ and CD27+ NK cells in LN and Interleukin-3 receptor other compartments such as BM, uterus and liver. Only lung-derived NK cells presented a very low CXCR3 but high CD27 expression. In healthy humans, the majority

of lung NK cells displays a CD56dim phenotype 34. However, the similar expression patterns of CXCR3 and CD27 suggest a coexpression of both markers. In fact, CXCR3 was exclusively expressed on CD27bright NK cells, although this could not be shown for human NK cells 26. In recent publications, mouse NK-cell subsets were defined as CD27+(high) and CD27−(low)23. According to our data regarding CXCR3 and CD27 expression, murine NK-cell subsets can be more precisely differentiated into CD27−CXCR3−, CD27dimCXCR3−, CD27brightCXCR3− and CD27brightCXCR3+ NK cells. Regarding the phenotype, the CXCR3+CD27bright NK-cell subset contained a greater proportion of CD69+, CD94+, CD62L−, CD16−/dim, CD11b− and Ly49s− NK cells as compared with CXCR3−CD27bright NK cells.