To improve our ability to identify hematopoietic stem cells (HSCs) and their localization in vivo, we compared the gene expression profiles of highly purified HSCs and non-self-renewing multipotent hematopoietic progenitors (MPPs). Cell surface receptors of the SLAM family, including CD150, CD244, and CD48, were differentially expressed among functionally distinct progenitors. HSCs were highly purified as CD150(+)CD244(-)CD48(-) cells while MPPs were CD244(+)CD150(-)CD48(-) and most restricted progenitors were CD48(+)CD244(+)CD150(-). The primitiveness of hematopoietic progenitors could thus be predicted based on the combination of SLAM family members they expressed. This is the first family of receptors whose combinatorial expression precisely distinguishes stem and progenitor cells. The ability to purify HSCs based on a simple combination of SLAM receptors allowed us to identify HSCs in tissue sections. Many HSCs were associated with sinusoidal endothelium in spleen and bone marrow, though some HSCs were associated with endosteum. HSCs thus occupy multiple niches, including sinusoidal endothelium in diverse tissues.

In addition to triggering the activation of B- or T-cell antigen receptors, the binding of a ligand to its receptor at the cell surface can sometimes determine the physiological outcome of interactions between antigen-presenting cells, T and B lymphocytes. The protein SLAM (also known as CDw150), which is present on the surface of B and T cells, forms such a receptor-ligand pair as it is a self-ligand. We now show that a T-cell-specific, SLAM-associated protein (SAP), which contains an SH2 domain and a short tall, acts as an inhibitor by blocking recruitment of the SH2-domain-containing signal-transduction molecule SHP-2 to a docking site in the SLAM cytoplasmic region. The gene encoding SAP maps to the same area of the X chromosome as the locus for X-linked lymphoproliferative disease (XLP) and we found mutations in the SAP gene in three XLP patients. Absence of the inhibitor SAP in XLP patients affects T/B-cell interactions induced by SLAM, leading to an inability to control B-cell proliferation caused by Epstein-Barr virus infections.

SummaryThe presence of natural killer (NK) cells contributes to early defense against murine cytomegalovirus (MCMV) infection. Although NK cells can mediate in vivo protection against MCMV, the mechanism by which they do so has not been defined. The studies presented here evaluate cytokine production by NK cells activated during MCMV infection and the role of NK cellproduced cytokines in early in vivo antiviral defenses. Experiments with normal C57BL/6, T cell-deficient C57BL/6 nude, and severe combined immunodeficient mice lacking T and B cells demonstrated that both interferon ~/(IFN-~/) and tumor necrosis factor (TNF) production were induced at early times after infection with MCMV. Conditioned media samples prepared with cells from these mice, on day 2 after infection, produced 11-43 pg/million cells oflFN-~ and 12-19 pg/million cells of TNF as evaluated by specific protein enzyme-linked immunosorbent assays. Studies in the NK-and T cell-deficient mouse line, E26, in mice that had been depleted in vivo of NK cells by treatment with antibodies eliminating NK cells, anti-asialo ganglio-N-tetraosylceramide or anti-NKl.1, and with populations of cells that had been depleted of NK cells by complement treatment with the anti-NK cell antibody, SW3A4, demonstrated that NK cells were solely responsible for the IFN-~/ but were not required for TNF production. The in vivo absence of NK cells was accompanied by increased viral hepatitis and viral replication in both immunocompetent and immunodeficient mice, as well as decreased survival time ofimmunodeficient mice. In vivo treatments with antibodies neutralizing IFN-~/ demonstrated that this factor contributed to the NK cell-mediated antiviral defense and reduced the measured parameters of viral defense to levels indistinguishable from those observed in NK cell-deficient mice. These effects appeared to be independent of cytolytic activity, as NK cells isolated from anti-IFN-~/-treated mice mediated killing at levels comparable to those observed in control-treated mice. The consequences ofinterleukin 12 (IL-12) administration, a known potent inducer of IFN-~/production by NK cells, were evaluated in MCMV-infected mice. Low IL-12 doses, i.e., 1 ng/d, increased NK cell cytotoxicity and IFN-~/production up to twofold and resulted in improved antiviral status; virus-induced hepatitis was decreased as much as fivefold, and viral burdens were decreased to levels below detection. The beneficial effects oflL-12 treatment were prevented by depleting either NK cells or IFN-~/. As infection of severe combined immunodeficiency mice was exacerbated by anti-IFN-~/administration and benefited from IL-12, the antiviral effects of both administered IL-12 and NK cell-produced IFN-'y in vivo could be independent of T and B cells. The results elucidate a cytokine-mediated mechanism for antiviral function of NK cells in vivo. Furthermore, they demonstrate the therapeutic efficacy of IL-12 administration in the context of an acute viral infection.M urine cytomegalovirus (MCMV) I is a herpesvi...