Discussion
Our study was the largest number of East Asian samples to analyze IgE
GWAS to date. By imputing the HLA allele, we obtained a higher
resolution of the association between HLA and serum IgE level.
Furthermore, we explore the pleiotropy relationship between IgE and
atopic diseases. Finally, we construct the PRS of IgE and found the
positive relationship between IgE PRS and asthma and atopic dermatitis.
The most significant SNP rs147642819 on chr6 has been reported to be
associate with several blood cell including white blood cell count,
neutrophil count, eosinophil counts26, and type 1
diabete27. Besides, this SNP is an eQTL forHLA-DQB1 in several tissues (including venous blood, thyroid
gland, and esophagus squamous epithelium), and locates within active
enhancer region in CD14 positive monocyte with the RegulomeDB score of
1f28. This evidence suggested that the variant may be
biological relevance and participated in gene expression via
transcriptional regulation.
Previous family study points out the linkage of 2q33 with total serum
IgE in asthma families29. This cytoband harbors two
candidate genes - CTLA4 and CD28, and both of them are
involved in regulation of T-cell activation. Howard et al.demonstrates that the polymorphisms of CTLA4 associate with the
total serum IgE levels, but no association is observed of the variants
from CD28 30. Our study firstly demonstrated an
intron variant of CD28 - rs1181388 in cytoband 2q33.2 is
significantly associated with IgE. Interestingly, this SNP is in DHS
promoter of CTLA431 , and is an eQTL ofCTLA4 in blood as well. This variant hasn’t been reported as
genome wide significant signals with other traits but being nominal
significance (p=2.5×10-4) in age hay fever, rhinitis
or eczema diagnosed in UKBB32. Besides, the SNP is in
highly LD (R2=0.92) with the asthma lead SNP -
rs55730955, which is previous reported in the Japanese
population22. Further studies are still needed to
deeper elucidate the relationship and impact of variants withinCD28 and CTLA4 to IgE level.
Multiple signals were identified in the MHC regions after conditional
analysis, suggesting the importance and high heterogeneity of HLAgenes. Several studies mentioned the association between HLA genes and
serum IgE, but the identified gene and HLA allele vary across
different populations. In Latino and Hispanic populations, they reportHLA-DQB1 *03:02 as the most significant signal for
IgE8. Our study identified HLA-DQB*03:03 as the
most significant star allele in the Taiwanese Han population, whileHLA-DQB1 *03:02 did not show significant association. The
frequency of HLA-DQB1 *03:02 is higher in Hispanic (0.194)
compared with HLA-DQB*03:02 (0.02), while the frequency ofHLA-DQB*03:03 (0.14) is higher in our cohort compared withHLA-DQB1 *03:02 (0.08). The difference distribution of HLA allele
frequencies across ancestries emphasizes the value of using diverse
ethnic groups to identify the shared and unique population-specific
association signals.
For atopic dermatitis, the global genetic correlation did not show
significance, but the local genetic correlation revealed borderline
significant signals. Our study revealed that IgE and atopic dermatitis
shared genetic components in 5q31.1. Previous study has demonstrated the
association between 5q31.1 and IgE33, and the linkage
association of 5q31-33 with AD has been reported as
well34. The local genetic correlation enabled us
having a deeper understanding of the shared genetic components of both
traits.
The IgE PRS was significantly correlated with the serum IgE level and
the risk of asthma and atopic dermatitis in testing group. Furthermore,
the association between IgE PRS and asthma was confirmed in a Japanese
population. This leaded to the high possibility of the future
application of IgE PRS. However, the AD did not show significance signal
in Japanese population. This may be due to relatively smaller sample
size in Japanese cohort (case = 4,288; control = 124,753) compared with
our testing group (case = 17,327; control = 247,433). Currently in
clinical practice, patients undergo testing for total IgE levels only
when allergic symptoms appear35. Therefore, our study
aims to identify individuals with high-risk to allergic disease before
their IgE levels rise even in newborn babies. With the decreasing cost
and increasing accessibility of DNA microarrays, genotyping screening
may become easier to implement in hospitals. By applying IgE PRS, each
subject can obtain their risk of allergic diseases before symptoms
manifest. This information can be particularly beneficial for high-risk
patients, as it can aid in preventing antigen exposure and reducing the
risk of allergies through early prevention measures.
In summary, our study identified 8 variants associated with IgE in the
Taiwanese Han population, successfully replicated six of them.
Additionally, the IgE PRS showcased its potential for future clinical
applications. These findings contribute to a deeper understanding of the
genetic architecture of serum IgE and its relationship with atopic
diseases.