The Pleiotropic Functions of Interleukin 15: Not So Interleukin 2–like After All

Since its discovery in 1994, IL-15 has been studied in the context of its structural cousin, IL-2 (1). Early biochemical and in vitro lymphocyte studies revealed many similarities between IL-15 and IL-2: (a) IL-2 and IL-15 constitute the only two members of a family of cytokines containing a four a helix bundle; (b) the heterotrimeric receptors for IL-2 and IL-15 share both the IL-2/15R b and IL-2R common g (IL-2R g or g c) chains; (c) the shared IL-2/ 15R b and g c chains are responsible for downstream cytoplasmic signals; and (d) the IL-2R a and IL-15R a chains, which confer specificity to IL-2R and IL-15R for their respective ligands, are closely related proteins (2). However, the highly restricted expression of IL-2 and IL-2R a to activated T cells suggests that IL-2R signals perform unique functions in adaptive T cells, whereas the pleiotropic expression of IL-15 and IL-15R a hints at a much broader role for IL-15R in multiple cell types. Indeed, distinct functions for IL-2 and IL-15 have been suggested from in vivo studies. IFN regulatory factor (IRF-1)–deficient mice lack NK cells due to an inability of their bone marrow stroma to elaborate IL-15, demonstrating a unique role for IL-15 in supporting NK cell development (3). Comparative studies of IL-2R a –deficient and IL-2/15R b –deficient mice (where only the latter should lack IL-15R signals) suggest that IL-15 might also be important for the differentiation of NK T cells and certain subsets of intraepithelial lymphocytes (IELs) (4, 5). Finally, exogenous IL-15, but not IL-2, selectively induces the proliferation of memory phenotype CD8 1 T cells in normal mice (6). Thus, emerging from the shadow of IL-2, IL-15 may serve a wide range of immune functions. The unique roles for IL-15 in vivo have become clearer with the initial reports of IL-15–deficient (IL-15 2 / 2 ) mice by Kennedy et al. in this issue (7) and IL-15R a 2 / 2 mice (8). While IL-2 2 / 2 and IL-2R a 2 / 2 mice spontaneously accumulate activated T and B cells and die prematurely from autoimmune disease, IL-15 2 / 2 and IL-15R a 2 / 2 mice are generally healthy, lymphopenic, and specifically lack NK cells, NK T cells, intestinally derived subsets of IELs, and activated CD8 1 T cells (7–10). The loss of these cells demonstrates that IL-15 signals via IL-15R a are critical for lymphoid development and/or maintenance, particularly for innate immune cells. By showing that 7 d of IL-15 treatment salvages these cells, Kennedy et al. suggest that the peripheral maintenance of these cells may be a prominent aspect of IL-15 function (7). This observation is consistent with prior work indicating that IL-15 supports peripheral memory phenotype CD8 1 T cells (6, 8). It also suggests that CD8 1 T cells may partially resemble innate immune cells. Furthermore, IL-15 also induces the activation and cytotoxicity of mature NK cells (11). In addition, IL-15 clearly regulates innate immune cell differentiation. Bone marrow stroma elaborates IL-15 to support NK cell differentiation (3, 12), and IL-15 skews the differentiation of thymic progenitors towards NK T cells (13; Fig. 1). Thus, IL-15 is critical for mediating the development, homeostasis, and activation of innate immune cells. The finding that IL-15 2 / 2 and IL-15R a 2 / 2 mice are lymphopenic suggests that IL-15 may also support adaptive CD4 1 and CD8 1 T cells (7, 8). Whether IL-15 regulates adaptive lymphocyte differentiation remains to be elucidated, but the reduced thymic cellularity of IL-15R a 2 / 2