1.  Unlike T cells, immature and mature NK cells require very high concentrations of soluble IL2 for growth in vitro.

2.  Immature and mature NK cells also require very high concentrations of IL15 for growth, making it unlikely that soluble IL15 alone is capable of driving the expansion of NK cell progenitors in vivo.

3.  Surprisingly, T cells also require very high concentrations of soluble IL15 for growth in vitro, with the exception of the CTLL2 line which responds efficiently to low doses of human and murine IL15.

4.  As judged by the kinetics and pattern of acquisition of key surface receptors, the differentiation of immature cells into mature NK cells occurs in an identical manner regardless of whether it is driven by IL15 or IL2.

5.  The requirement for very high concentrations of IL2 and IL15 might suggest that NK cells respond to these cytokines only through low affinity receptors, and may not express high affinity receptors for IL2 or IL15.

6.  In the absence of suitable mAbs it is not possible to determine whether NK cells express IL15Ra chains at the surface, but by RTPCR analysis they clearly express mRNA for IL15Ra.

7.  However, the IL15Ra transcripts are subject to extensive alternate splicing, and we identified transcripts that potentially encode forms of the IL15Ra chain that could either disrupt high affinity signalling complexes at the cell surface or compete for signalling components in the cytoplasm.

8.  Contrary to expectation and widely held views, murine NK cells were found to express mRNA for IL2Ra chains.

9.  In addition, established NK cell lines express easily detectable IL2Ra chains on the cell surface.

10.  Freshly explanted NK cells grown in IL15 also express detectable surface IL2Ra chains, but IL2 induces down regulation of IL2Ra chains, and NK cells grown in IL2 therefore have undetectable surface expression.

11.  The IL2Ra chains on NK cells appear to be associated with functional high affinity receptors which, as shown by blocking with anti-IL2Ra mAbs, are essential for the proliferation of NK cells at least when these are grown in low concentrations of IL2.

12.  The requirement for very high, presumably non-physiological, concentrations of IL2 and IL15 suggests that these cytokines are either not directly involved in driving the proliferation of NK cell progenitors, or their effectiveness is enhanced by association with other factors.

13.  In support of the latter hypothesis we found that low doses of IL21 could markedly enhance the responsiveness of immature and mature NK cells to limiting concentrations of IL2 and IL15.  By contrast, high doses of IL21 inhibited NK cell growth.

14.  The recent demonstration by Kasaian et al. that mice lacking the only known receptor for IL21 have normal numbers of NK cells indicates that IL21 is not essential for the development of NK cells in vivo.  The present findings suggest, however, that IL21 may have an important role in controlling NK cell expansion during immune responses, high concentrations of IL21 at sites of infection or inflammation inhibiting in situ NK cell proliferation and the low concentrations of IL21 that reach sites of NK cell development enhancing the production of NK cells.