Together, these results suggest the existence of a strong positive-feedback loop, using IL-15 as a common trophic signal, in

early GC development. Once IL-15 signalling is induced, proliferation of GC-B cells and FDCs is augmented, and the amount of IL-15 per se will be dramatically amplified by reciprocal signalling between the cells. Given the urgency of generation and production of protective high-affinity antibodies in case of infection, this sharing of common pro-proliferative cytokines, by both functional CP-690550 in vitro GC-B cells and microenvironmental stromal cells, FDCs, may be advantageous for the timely development of the GC reaction. Moreover, proliferation of FDCs is thereby coupled to antigen-specific proliferation of GC-B cells, augmenting the selective generation of GC-B cells with high-affinity B-cell receptors for antigen. Interleukin-15 does not have a significant effect on the apoptosis of FDC in our in vitro culture model (Fig. 3c) in contrast to previous reports on the anti-apoptotic effects of IL-15 in various cells.44,56,57 The reported anti-apoptotic effects were measured in the presence of strong apoptotic signals, including stimulation of other surface molecules by anti-Fas, TNF-α, anti-CD3 and IgM, or use of toxic chemicals. In contrast, we examined the effect of IL-15 in the absence of apoptotic inducers, which may be more relevant to the early GC reaction in vivo. We attempted

to induce apoptosis ICG-001 mouse of FDCs using anti-Fas antibody or TNF-α to investigate an anti-apoptotic function of IL-15 on FDCs; however, apoptosis was not detected in freshly isolated FDCs (C-S. Park, unpublished data). Therefore,

although an anti-apoptotic effect of IL-15 on FDCs undergoing apoptosis during the GC response54 cannot be excluded, the major role of IL-15 in the developing GC is to enhance proliferation of both FDCs and GC-B cells. Another important question regarding the function of IL-15 on FDCs is whether IL-15 is involved in FDC differentiation. One of the major obstacles in FDC research has been the lack of a reliable, functional, experimental system. For instance, it is difficult to distinguish between any changes in FDCs from those of other cellular components of the GC reaction, using a genetically modified many mouse model. Immunohistochemical analysis has limitations because such analysis cannot be used to measure functional changes. In vitro culture experiments are a plausible alternative. However, the culture experiments also have limitations, including the possible loss of functional competency during in vitro culture. The FDCs needs various factors from GC-B cells to develop and to maintain their function. To compensate for these problems, we designed a culture protocol to mimic in vivo functional FDCs by co-culturing primary human FDCs with GC-B cells. Hence, signals from GC-B cells essential for FDC function16,58 are provided in our experimental model. The TNF-α control set is included for two purposes.