With the next set of experiments we addressed the question whether surface IgE-positive B cells can be detected in IgE knock-in mice. First, we stimulated total spleen cells for 5 days with LPS and IL-4. We used IgE knock-in mice on the CD23−/− background in order to avoid passive binding of soluble Akt inhibitor IgE to the low

affinity IgE receptor (CD23) on B cells [23]. Surface IgE and IgG1 were detected by flow cytometry. LPS alone neither induced significant IgE nor IgG1 expression (0.4–1.5%) (Fig. 2A and Supporting Information Fig. 1). In B cells from WT mice LPS+IL-4 induces IgG1 (23%), but only very little IgE (1.5%). In contrast, both cells isolated from either heterozygous or homozygous IgE knock-in mice express comparably high amounts of IgE (ca. 15%) on the cell surface. However, the find more small fraction of positively stained cells might be due to a cross-reactivity or background staining of

the detection antibodies (see also Fig. 2E). WT mice express 23% and heterozygous IgE knock-in mice 10% IgG1 and, as predicted, no IgG1 was found in IgEki/ki mice. These results suggest that in vitro the chimeric membrane IgE molecule can be transported to the surface with a slightly lower efficiency than natural IgG1. To confirm these results, we performed a RT-PCR analysis of the membrane forms of IgE, IgG1, and the chimeric membrane IgG1-IgE form (Fig. 2B). The results of LPS+IL-4 stimulated cultures are in line with the protein expression data (Fig. 2A); however, LPS alone induces mRNA transcripts with little IgG1 or chimeric IgE being expressed on the surface of the cells (Fig. 2B). Second, we analyzed B cells from bone marrow, lymph nodes (data not shown), and spleens of heterozygous IgE knock-in mice and their WT littermates. We could find a normal B-cell subset distribution in vivo (data not shown). However, we could not detect membrane IgE-positive B cells (Fig. 2C) in the

spleen. The absence of CD23 demonstrates that the increase in IgE expression is not a result of an increase in membrane IgE expressing B cells in unchallenged, naïve mice (Fig. 2C) [23]. Additionally, immunization and boost with the T-dependent antigen unless trinitro-phenyl-chicken ovalbumin (TNP-OVA) and the subsequent immunohistochemical analysis of splenic B-cell follicles shows only very rare IgE-positive cells located at the edge of the B-cell follicle in IgE knock-in mice of the CD23−/− background (Fig. 2D). Surface IgE and IgG1 expression in vivo were then analyzed after infection with the helminth Nippostrongyus brasiliensis (Nb), which leads to pronounced Th-2 responses [29]. Mesenteric lymph nodes of IgEki/ki, IgEki/wt, and WT mice were taken at day 14 after infection, at the peak of the germinal center response. IgEki/ki mice, as expected, showed no staining for IgG1, whereas IgEki/wt had intermediate expression of surface IgG1 when compared to WT.