Discussion
Targeting IgE with therapeutic anti-IgE antibodies such as omalizumab is
a validated strategy to treat IgE-mediated diseases. Naturally occurring
anti-IgE IgG antibodies have been identified in patients with various
diseases, including asthma and atopic dermatitis, but also healthy
individuals. Our recent work demonstrated that a single immunization
with IgE-allergen complexes elicits an anti-IgE IgG response without
adding adjuvants. The resulting anti-IgE IgGs bind glycosylated IgE
significantly better than deglycosylated IgE. Here we show that booster
immunizations can even enhance the anti-IgE IgG response suggesting that
repeated exposure to IgE-allergen IC might lead to a beneficial anti-IgE
IgG response. Glycosylation of IgE plays a vital role in this process.
The use of glycosylated IgE leads to an increased anti-IgE IgG and even
an increased anti-Fel d 1 IgG response compared to deglycosylated IgE.
The adjuvant effect of IgE for antigens has been previously
demonstrated, but here we show that glycans are critical for this
effect. The reduced immunogenicity of deglycosylated IgE might be
counterintuitive. Removal of the glycans makes the IgE less soluble and
could result in the formation of aggregates. It is known that aggregates
can be important for antigenicity. Hence, it is reasonable to assume
that glycan-recognizing receptors are involved in the process since the
interaction between IgE and CD23 is glycan independent and CD23KO mice
also show an anti-IgE response. The fact that IgE is the most
glycosylated immunoglobulin further strengthens this assumption. About
12% of its molecular weight is carbohydrate. Moreover, glycosylation of
proteins influences their serum kinetic. A possible candidate could be
the lectin galectin-9, which is known to bind IgE. Even though it has an
anti-allergic effect by preventing IgE-antigen complex formation, it is
not clear whether it can bind IgE-antigen complexes or not. However, it
is known to induce DC maturation and promote T cell proliferation and
Th1 cytokine production. This would explain the higher anti-IgE IgG
response in mice immunized with glycosylated IgE-ICs. Other receptors in
mice that can bind IgE are FcγRIIIA and FcγRIV. Specifically for FcγRIV,
one study has shown that this receptor binds IgE when IgE is present as
a complex but not monomeric IgE. Moreover, the engagement of FcγRIV led
to the antigen presentation to T cells However, whether glycosylation of
IgE plays a role in this process is unclear. Interestingly, two amino
acids (K117 and E132) essential for FcεRI binding to IgE are conserved
in FcγRIV. Since IgE glycosylation is required for binding to FcεRI, it
may also be necessary for binding to FcγRIV.
Further differences can be observed in the ability of anti-IgE and
anti-IgE(PNG) IgG antibodies to clear IgE from serum and prevent binding
to effector cells. The results show that anti-IgE IgG antibodies are
significantly better at controlling the binding of IgE to FcεRI of
basophils than anti-IgE(PNG) antibodies. This indicates that IgE
glycosylation affects the quantity of anti-IgE antibodies generated and
their quality. An explanation for this could be the difference in
affinity since it is higher for the anti-IgE IgG antibodies than
anti-IgE(PNG) antibodies. It could be that glycans are essential for
interacting with anti-IgE IgG to IgE. Thus, in the case of missing
glycans, the anti-IgE IgG binding site could be altered, resulting in
decreased affinity. Another explanation could be that deglycosylation
leads to a conformational change. This change would result in different
conformational epitopes between glycosylated and deglycosylated IgE.
Passive immunization with anti-IgE IgG and anti-IgE(PNG) IgG allowed us
to see if the difference in serum clearance is also reflected in the
ability to protect mice from anaphylaxis upon allergen challenges. Our
results show that both anti-IgE IgG antibodies protect BALB/c WT mice
compared with controls. However, protection was always better in the
anti-IgE IgG group than in the anti-IgE(PNG) group. These results are
consistent with the clearance experiment results, where significantly
more IgE was found on the basophils of the anti-IgE(PNG) IgG group than
in those of the anti-IgE-IgG group. To better understand the protection
mechanism, allergen challenges were also performed in FcγRIIBKO and
CD23KO mice. Previous studies have shown that co-aggregation of FcγRIIb
with FcεRI inhibits IgE-mediated degranulation of effector cells.
Moreover, it has been observed that allergen-specific desensitization
(SIT) induced allergen-specific IgGs, which act via FcγRIIb. However,
since the FcγRIIbKO mice were as protected as the WT mice, we speculate
that the protection is mainly mediated by the neutralization of IgE via
CD23. In both B cells and monocyte-derived DCs, CD23 was shown to bind
and internalize IgE-ICs better than IgE alone and could therefore play a
role in the clearance of the complexes. Indeed, CD23KO mice were not as
well protected as wild-type Balb/c mice, suggesting that CD23 is
important for the clearance of IgE-IgG complexes. Indeed, our serum
kinetic results in CD23KO confirmed that most complexes are eliminated
via CD23. Not enough elimination of complexes leads to the accumulation
of IgE-IgG complexes and then to their dissociation, during which the
released IgE can bind preferentially to FcεRI because of its higher
affinity for FcεRI than for the IgG antibodies (see Graphical Abstract).
This results in the sensitization of effector cells, which can
degranulate upon encountering the allergen. Additionally, while IgE-ICs
do not trigger empty FcεRI-mediated degranulation, IgE pre-sensitized
basophils/mast cells can be degranulated by IgE-ICs. A lack of clearance
could thus favor degranulation of IgE pre-sensitized effector cells.
In conclusion, we have demonstrated the importance of IgE glycosylation
for the anti-IgE-IgG response and a possible pathway to eliminate
IgG-IgE complexes via CD23 in a mouse model. Further research is needed
to investigate the effects of IgE glycosylation on anti-IgE IgG
antibodies and serum IgE levels in humans. In particular, the extent to
which different IgE glycosylation patterns affect anti-IgE IgG
antibodies should also be investigated. More insights into the role of
IgE glycosylation for anti-IgE IgG antibodies could help to find even
more efficient strategies against type I hypersensitivity diseases.
References and Notes: