As a consequence, the level of LTα is not sufficient in CXCR5-deficient mice for the development of follicular structures. In these animals, the T-cell zone is surrounded by a narrow ring of BP3hi biglycanhi stromal cells (Fig. 4C) in which B cells are embedded (data not shown) 27. As in wild-type
animals, the spatially differential expression of the chemokines CXCL13 and CCL21 (Fig. 4G, left and center panel) controls the localization of B and T cells. BP3hi stromal cells expressed in addition to Cxcl13, Enpp2, but the expression levels were lower than in mature FDC (Fig. 4G, right panel). No expression was detectable for the genes Serpina1, Cilp, Postn, Lbp3, Lrat, Coch and 9130213B05Rik (Table 1), supporting that in CXCR5-deficient mice the level of LTα expression is not sufficient for full development of mature FDC. In LTα-deficient mice, the lymphoid compartments LBH589 order are partially established (Fig. 4D and H) 28. The network of reticular cells was visualized with biglycan and Vcam-1-specific Ab (Fig. 4D). Again the organization of a B- and a T-cell area is supported by spatially differential expression of Cxcl13 and Ccl21 (Fig. 4H). In situ hybridization showed
that the expression level of Cxcl13 is even lower than in BP3hi reticular cells of SCID mice (Fig. 4F and H). Expression of the genes Enpp2, Serpina1, Cilp, Postn, Lbp3, Lrat 9130213B05Rik Seliciclib nmr and Coch was not detectable (Table 1). These data show that the newly defined Cyclin-dependent kinase 3 set of FDC specific genes allows us to follow modifications in the gene expression profile leading to the differentiation of mature FDC. FDC have an essential role for B-cell homeostasis and during the GC reaction they support the activation and differentiation of B cells into memory and plasma cells 1–3, 5. As the isolation of intact FDC to homogeneity is not yet technically possible 6, 7, 11, rather little is currently known about their function and origin. In contrast to other approaches analyzing gene expression in FDC, we used laser capture micro-dissection (LCM),
an isolation technique that does not affect the transcriptional in vivo situation. The problematic issue of co-isolation of additional cell types was overcome by using an in silico subtraction approach. The number of follicular T cells and tingible body macrophages, which localize in the FDC network, was shown to be too low to have a significant impact on the gene expression profile of FDC as demonstrated by barely detectable signals of major transcriptional products of these cell types. Subsequently, it was sufficient to subtract genes expressed in co-isolated B cells to determine the transcriptome of FDC. Nevertheless, a dilution of FDC-expressed RNA by that of co-isolated B cells was seen, when the gene expression profile of FDC and BP3hi stromal cells isolated from the SCID mouse was compared (Fig. 3).