Regulation of IgE production requires oligomerization of CD23

Here we describe the production of a rabbit polyclonal Ab (RAS1) raised against the stalk of murine CD23. RAS1 inhibits release of CD23 from the surface of both M12 and B cells resulting in an increase of CD23 on the cell surface. Despite this increase, these cells are unable to bind IgE as determined by FACS. CD23 has previously been shown to bind IgE with both a high (4-10 x 10(7) M(-1)) and low (4-10 x 10(6) M(-1)) affinity. Closer examination by direct binding of (125)I-IgE revealed that RAS1 blocks high affinity binding while having no effect on low affinity binding. These data support the model proposing that oligomers of CD23 mediate high affinity IgE binding. These experiments suggest that RAS1 binding to cell surface CD23 results in a shift from oligomers to monomers, which, according to the model, only bind IgE with low affinity. These experiments also suggest that high affinity binding of IgE is required for IgE regulation by CD23 and is demonstrated by the fact that treatment of Ag/Alum-immunized mice treated with RAS1 results in a significant increase in IgE production similar to the levels seen in CD23-deficient mice. These mice also had significantly decreased levels of serum soluble CD23 and Ag-specific IgG1. RAS1 had no effect on IgE or Ag-specific IgG1 production in CD23-deficient mice.

Metalloprotease inhibitor-mediated inhibition of mouse immunoglobulin production

High levels of membrane CD23 have been shown to decrease immunoglobulin E (IgE). CD23 is a very labile molecule and is cleaved from the cell surface by an unknown metalloprotease. Two metalloprotease inhibitors, compound A (N-[4-hydoxyamino-2-(R)-isobutyl-3-(S)propargylthiomethylsuccinyl]-(S)-phenylalnine-N'-methyl-amide) and compound B (N-[3-(S)-hydroxy-4-hydroxyamino-2-(R)-(2-naphthylmethyl) succinyl]-(S)-tert-leucinamide), were chosen for their ability to inhibit human CD23 cleavage and selectively inhibit IgE production. The ability of these inhibitors to block cleavage of murine CD23 and immunoglobulin production in an in vitro system was examined. The inhibitors blocked sCD23 release from B cells. The inhibitors also decreased IgE production by B cells; however, 20-30 times more inhibitor was needed to give a similar amount of inhibition as compared with sCD23 release. The effects on immunoglobulin production did not require the presence of CD23 in that these inhibitors also blocked in vitro immunoglobulin production when B cells from CD23-/- mice were used. The inhibitors decreased production of all other immunoglobulin isotypes examined and reduced the number of IgE antibody-forming cells (AFC) while having no effect on cell proliferation or viability. The level of Iepsilon transcripts in cells treated with compounds A and B were not different as compared with control cells. These results suggest that while these inhibitors effectively inhibit IgE production in a CD23-specific manner in the human, these compounds, in the mouse, inhibit immunoglobulin production by an unknown mechanism that is unrelated to CD23.

B cell activation and Ig, especially IgE, production is inhibited by high CD23 levels in vivo and in vitro

The capacity of CD23 to regulate IgE production was evaluated in both an in vitro and an in vivo system. The decreased IgE response seen in CD23 transgenic mice was confirmed and observed to occur at all antigen doses used. In addition, purified B cells from the Tg animals in general exhibited lower IgE production when stimulated with CD40L and IL-4. To examine this down-regulating activity of CD23 an in vitro model system was developed. CHO cells were transfected with CD23, ICAM-1, or both CD23 and ICAM-1. ICAM-1 was chosen to enhance B cell-B cell interaction. Purified resting B cells were placed into culture with the mitomycin C-treated transfected or control CHO cells and activated with IL-4, IL-5, and CD40L-CHO. A dose-dependent decrease in IgE production was observed with increasing cell numbers of the CHO transfectants that expressed CD23. The effect lasted up to Day 3 of culture. B cell proliferation was also inhibited in a dose-dependent fashion by increasing numbers of CD23-expressing cells suggesting a potential effect of CD23 on B cell apoptosis. In contrast, ICAM-1-transfected or CHO control cells had minimal effects on either Ig production or B cell proliferation. While IgE production was inhibited up to 95% by high numbers of CD23-transfected CHO cells, some inhibition of IgG and IgM production was also seen. Finally, the mechanism of CD23-mediated inhibition of IgE production was compared with the inhibition in IgE production seen when B cell were coactivated with multivalent anti-IgD in conjunction with CD40L plus optimal IL-4. To this end we used RT-PCR to compare the relative levels of epsilon-germline transcripts in control cultures and cultures coactivated by anti-IgD, CD40L, and IL-5 or activated in the presence of high levels of CD23-expressing cells. CD22 was used as an internal standard since levels change little with B cell activation. Coactivation strongly inhibited epsilon-germline transcript levels but the presence of CD23-expressing cells did not. Thus, coactivation potentially operates prior to isotype switching, while high CD23 coculture blocks either recombination or more likely B cell differentiation to high Ig producers stage. Our data support the hypothesis that IL-4 induces both IgE and a controlling agent for IgE, namely, CD23.

Induction of B cell apoptosis by co-cross-linking CD23 and sIg involves aberrant regulation of c-myc and is inhibited by bcl-2

A novel system to study the effects of co-cross-linking CD23/FceRII and sIg on murine B lymphocytes utilizes a highly multivalent form of anti-Ig prepared by covalently linking anti-Ig antibodies to a DNP-dextran backbone. CD23-sIg co-cross-linking is accomplished by the addition of DNP-specific monoclonal IgE. Previous studies demonstrated that co-cross-linking CD23 and sIg significantly inhibited mouse B cell proliferation, especially at high doses of the multivalent anti-Ig. Interestingly, examination of early activation signals reveals no difference in B cells subjected to co-cross-linking conditions as compared to B cells activated with anti-Ig alone. Total cellular protein tyrosine phosphorylation levels are unchanged by co-cross-linking. Analysis of B cell mRNA reveals that co-cross-linking the receptors does not alter the expression levels of ornithine decarboxylase 8 h after stimulation as compared to the controls. In contrast, levels of the proto-oncogene c-myc were significantly elevated 1 h after inducing B cell activation under co-cross-linking conditions. However, it remains unclear whether this aberrant c-myc regulation plays any role in inducing apoptosis. In addition, on day 3 after stimulation, the co-cross-linking of CD23 and sIg resulted in the formation of apoptotic B cells, determined by both photomicroscopy of the B cell cultures and FACS analysis of B cell nuclei. B cells obtained from bcl-2 transgenic mice proliferated as well as controls, and failed to undergo apoptosis when CD23 and sIg were co-cross-linked on their surface. These studies indicate that co-cross-linking of CD23 with B cell sIg inhibits B cell proliferation by a mechanism that is distinct from that seen by co-cross-linking of the Fc gamma RII and sIg. In addition, these results suggest a means by which antigen-specific IgE can down-regulate additional B cell activation and IgE synthesis.