A genome-wide association study identifies polymorphisms in the HLA-DR region associated with non-response to hepatitis B vaccination in Chinese Han populations
Human Molecular Genetics
A genome-wide association study identifies polymorphisms in the HLA-DR region associated with non-response to hepatitis B vaccination in Chinese Han populations
Liping Pan 2
Li Zhang 0
Wei Zhang 5
Xiaopan Wu 2
Yuanfeng Li 4
Bingyu Yan 0
Xilin Zhu 2
Chao Yang 2
Jianfeng Xu 3
Gangqiao Zhou 4
Aiqiang Xu 0
Hui Li 1
Ying Liu 2
0 Shandong Center for Disease Control and Prevention , Jinan , China
1 Department of Epidemiology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, School of Basic Medicine, Peking Union Medical College , Beijing , China
2 National Laboratory of Medical Molecular Biology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences
3 Center for Cancer Genomics, Wake Forest School of Medicine , Winston-Salem, NC , USA
4 State Key Laboratory of Proteomics, Beijing Proteome Research Center, Beijing Institute of Radiation Medicine , Beijing , China
5 Beijing Center for Disease Control and Prevention , Beijing , China
Vaccination against hepatitis B virus is an effective and routine practice that can prevent infection. However, 510% of healthy adults fail to produce protective levels of antibody against the hepatitis B vaccination. It has been reported that host genetic variants might affect the immune response to hepatitis B vaccination. Here, we reported a genome-wide association study in a Chinese Han population consisting of 108 primary high-responders and 77 booster non-responders to hepatitis B vaccination using the Illumina HumanOmniExpress Beadchip. We identified 21 SNPs at 6p21.32 were significantly associated with non-response to booster hepatitis B vaccination (P-value <1 3 1026). The most significant SNP in the region was rs477515, located ∼12 kb upstream of the HLADRB1 gene. Its P-value (4.81 3 1028) exceeded the Bonferroni-corrected genome-wide significance threshold. Four tagging SNPs (rs477515, rs28366298, rs3763316 and rs13204672) that capture genetic information of these 21 SNPs were validated in three additional Chinese Han populations, consisting of 1336 primary high-responders and 420 primary non-responders. The four SNPs continued to show significant associations with non-response to hepatitis B vaccination (P-combined 5 3.98 3 10213 - 1.42 3 1028). Further analysis showed that the rs477515 was independently associated with non-response to hepatitis B vaccination with correction for other three SNPs in our GWAS and the known hepatitis B vaccine immunity associated SNP in previous GWAS. Our findings suggest thatthe rs477515 was an independentmarker associated withnon-response tohepatitis B vaccination and HLA-DR region might be a critical susceptibility locus of hepatitis B vaccine-induced immunity.
Infection with hepatitis B virus (HBV) is a public health problem
that seriously threatens human life. Approximately more than
400 million people worldwide are chronically infected with
HBV and more than half a million people die each year from
HBV-related liver disease (
). HBV infection shows a marked
regional diversity and is prevalent in the Asia and Sub-Saharan
Africa (3). In China, 93 million people are carriers of HBV,
and 300 000 deaths occur each year as a result of HBV infection
). Recombinant vaccines have been proven to be effective in
preventing HBV infection and have been used widely for .30
years. However, 5 – 10% of healthy adults fail to produce
protective levels of antibody (non-responders) after standard three
doses of hepatitis B vaccination schedule (
). Furthermore, a
subset of primary non-responders could not elicit protective
antibodies yet, even though they are administered booster
vaccination. These individuals remain at high risk for HBV infection.
Several epidemiological factors such as older age, male
gender, higher body mass index (BMI) and a history of smoking
are associated with a decreased antibody response to hepatitis
B vaccination (
6 – 9
). Different injection methods could also
affect the efficacy of hepatitis B vaccine (10). It has been
reported that parts of primary non-responders might develop to
responders after booster vaccination (another one dose or
revaccinated with another standard three doses of hepatitis B vaccine)
). In addition, a twin study indicated that host genetic
background account for 77% of all factors influencing immune
response to hepatitis B vaccination (13). Although
candidategene-based case – control studies have implicated that genetic
variants in interleukins (IL)1b, IL4, IL10, IL12B, CD3Z, and
several HLA loci were associated with variable immune response
to hepatitis B vaccination (
14 – 20
), none of these associations
Genome-wide association study (GWAS) has been developed
to systematically investigate the associations between
polymorphisms and polygenic inheritance disorders (
a GWAS reported that genetic variants in HLA-DR, HLA-DP
and HLA-III loci were strongly associated with postvaccination
antibody titers of hepatitis B vaccination in an Indonesian
). In this study, they grouped the subjects into three
groups on the basis of their antibody titers and focused on
investigating the association between polymorphisms and the
postvaccination antibody titers, rather than the non-responses. In
fact, World Health Organization reported that persons with
antibody ≥10 mIU/ml could prevent HBV infection, namely, the
non-responders whose antibody titers ,10 mIU/ml are
susceptible and at a high risk of infection with HBV (
). Therefore, it is
clinically meaningful to identify genetic variants contributing to
non-response to hepatitis B vaccination. Until now, only one
candidate-gene-based case – control study has paid attention to
non-responders to hepatitis B vaccination in Chinese Han
population, but the sample size and the number of candidate SNPs
were limited (24 non-responders and 46 responders, 51 SNPs)
). Therefore, we performed a multistage GWAS consisting
of 498 non-responders and 1449 high-responders in four
Chinese Han populations to identify the susceptible loci of
nonresponse to hepatitis B vaccination. Our findings might
contribute to a better understanding of the genetic variants that were
associated with the non-response to hepatitis B vaccination,
and facilitating to investigate detailed mechanisms of
nonresponse to hepatitis B vaccination. Furthermore, the results
might be helpful to identify specific genes as targets in the
development of novel and effective vaccines.
In the GWAS stage, we genotyped 731 442 SNPs across the
genome in 113 primary high-responders (cases) and 78 booster
non-responders (controls). After quality control, a total of
588 026 SNPs in 108 cases and 77 controls were remained
for further analysis. Principal component analysis (PCA)
showed that cases and controls had similar distributions of
the top three eigenvectors (Supplementary material, Fig. S1).
Consistent with this result, we found no evidence of population
stratification, with a genomic inflation factor (l) of 1.03
(Supplementary material, Fig. S2). We also performed a logistic
regression model to adjust for the top five principal components
(generated in EIGENTSTRAT) and derived a similar genomic
inflation factor (l ¼ 1.03). In addition, the direct comparisons
showed that the association results of top 1000 SNPs from the
analysis with or without adjustment for top five principal
components were generally similar. All these results indicated that the
population stratification in the remaining samples had a minimal
impact on the association results.
The SNPs showing strongest association were on
chromosome 6p21.32 (Fig. 1; Supplementary materials, Table S1).
The lead SNP in this region was rs477515 (P ¼ 4.81 × 1028,
OR ¼ 3.59), which exceeded the Bonferroni-corrected
genomewide significance threshold (0.05/588 026 ¼ 8.50 × 1028). A
detailed examination of this region revealed that a total of 21
SNPs were significantly associated with non-response to the
vaccination (P-values ,1 × 1026, Table 1 and Fig. 2). The lead
SNP (rs477515) locates 12 kb upstream of the HLA-DRB1
gene, and these 21 SNPs overlapped with four known genes;
C6orf10, BTNL2, HLA-DRA and HLA-DRB1.
Linkage disequilibrium (LD) analysis identified four tagging
SNPs (rs477515, rs28366298, rs3763316 and rs13204672) that
capture genetic information of these 21 SNPs (Supplementary
material, Fig. S3). So, we only genotyped these four SNPs in
the confirmation Ia (consisting of 263 primary non-responders
and 825 primary high-responders) and confirmation Ib (111
primary non-responders and 297 primary high-responders)
cohorts from Shandong. We found that all four SNPs were
significantly associated with non-response to primary hepatitis B
vaccination in the two confirmation cohorts (P-values ranged
from 0.026 to 5.05 × 1028 in these confirmation stages,
Table 2). To further validate these results, we analyzed these
four SNPs in an additional Chinese Han population consisting
of 46 primary non-responders and 214 primary high-responders
from Beijing. Associations with response to the vaccination
were also confirmed for all four SNPs in this cohort (P-values
ranged from 0.048 to 1.58 × 1024, Table 2), even after adjusting
for the age, BMI and smoking (P-values ranged from 0.031 to
2.25 × 1024). When compared the genotypes of these four
SNPs in GWAS stage and subsequent confirmation stage with
the 11 populations from HapMap, we found that the allele
frequencies of these four SNPs were similar to those of Asian
By pooling the three confirmation case – control studies, the
associations of these four SNPs with non-response to primary
hepatitis B vaccination reached genome-wide significance
(P-values ¼ 3.98 × 10213 – 1.42 × 1028). The combined
analysis of the GWAS and confirmation stage data revealed more
prominent associations (P-values ¼ 2.63 × 10219 – 3.75 ×
10213; Table 2).
We also evaluated the genetic effects of previously reported
loci, HLA-DR (rs3135363), HLA-DP (rs9277535) and HLA-III
(rs9267665) on 6p21.3, in our GWAS and confirmation sets
(Supplementary materials, Table S2). The rs3135363 was
replicated consistently in both GWAS (P-value ¼ 0.0007) and
confirmation Ia set (P-values ¼ 0.0002), but it showed a weaker
association than the identified four SNPs in our GWAS.
Neither rs9277535 nor rs9267665 showed significant
associations with non-response to hepatitis B vaccination in both
GWAS and confirmation sets (P-values . 0.05). Because the
known rs3135363 was replicated in our GWAS and
confirmation sets, and it was located near the identified SNPs in our
study, we next investigated whether the identified four SNPs
might simply be from the tracking of rs3135363. LD between
rs3135363 and each of the identified four SNPs (rs477515,
rs28366298, rs3763316 and rs13204672) were performed
firstly. The results indicated that no tight LDs were observed
in the analysis (r2 ≤ 0.120, Supplementary Materials,
Table S3). Furthermore, the identified four SNPs were strongly
associated with non-response to hepatitis B vaccination even
after stratification with rs3135363 in GWAS and
confirmation sets, using a forward stepwise conditional logistic
regression analysis (Supplementary material, Table S4).
These results together implicated the independent effects of
the identified four SNPs on immune response to hepatitis B
To further test whether these four SNPs were independently
associated with response to the hepatitis B vaccination, we fit a
logistic regression model using a forward selection procedure,
starting with the most significant SNP rs477515 in all study
subjects (GWAS and three confirmation stages). Only rs477515
showed independent association with non-response to hepatitis
B vaccination, while none of the other three SNPs was
significant, suggesting the observed associations of these four SNPs
(and perhaps 21 SNPs) at 6p21.32 with response to the
vaccination are dependent and likely due to a common variant near or
at rs477515 (P-value ¼ 1.64 × 10217, OR ¼ 2.00, 95% CI ¼
1.71 – 2.35).
The HLA-DRB1 region was also genotyped using direct
sequencing method among the 185 samples in the GWAS stage.
In line with previous studies, we also found significant
The SNPs that were not detected in previous published GWAS were emphasized in bold.
‘_’, the minor allele. MAF, minor allele frequency; HR, high responder to primary vaccination (anti-HBs ≥ 1000 mIU/ml after three doses); NR, non-responder to
booster vaccination (anti-HBs , 10 mIU/ml after six doses).
association of HLA-DRB1∗0701 allele with non-response to
hepatitis B vaccination (P-value ¼ 1.50e206, OR ¼ 3.69;
Supplementary Material, Table S5). Further LD analysis indicated
that there was a strong LD between identified rs13204672 and
HLA-DRB1∗0701 (r2 ¼ 0.982), while other three SNPs
showed weaker LD with HLA-DRB1∗0701 (r2 ≤ 0.548).
To further increase genome coverage, we carried out an
imputation analysis to infer the genotypes of additional common
SNPs using the genotype data from 1000 Genomes Project
Han Chinese in Beijing (CHB) and Han Chinese South (CHS)
(March 2012 release) as a reference. For all over 38 million
autosome imputed SNPs, 5.7 million imputed SNPs were retained
for the association analysis after rigorous quality control. The
SNPs showing strongest association were also located on
chromosome 6p21.32 (Supplementary materials, Fig. S4,
Table S6). The top associated SNPs with P-values of ,1 ×
1026 were located near the identified four SNPs in our GWAS
(, 340 kb). No novel region associated with non-response to
hepatitis B vaccination was found in the imputation.
Of the total 588 026 SNPs, 288 743 SNPs were mapped to 17
492 genes within 5 kb upstream and downstream, which
were assigned to 104 pathways. Pathway enrichment analysis
using the improved gene-set-enrichment analysis approach
(i-GSEA4GWAS) have identified that 11 pathways were
significantly enriched with association signals with FDR , 0.05
(Supplementary material, Table S7). The top one significant
pathway is antigen processing and presentation, which is
consisting of several genes in HLA region and is directly involved
in immune response. The rs477515 nearby gene, HLA-DRB1,
was also included in this significantly enriched gene sets in
the pathway analysis. We also performed a pathway analysis
with omitting the SNPs in HLA region, and found that no
significant pathway was enriched after omitting the HLA
Hepatitis B vaccine-induced immunity is a complex process that
is controlled by numerous factors. In addition to environmental
factors and host-related physical factors, genetic variations
also play an important role in regulating the immune response
to hepatitis B vaccination (
). Previous candidate-gene
studies have suggested that polymorphisms in the HLA-DR
region were associated with immune response to hepatitis B
vaccination (15), and HLA-DRB1∗0701 allele was consistently
validated as a risk factor for immune response to hepatitis B
). Recently, a GWAS focused on
postvaccination antibody titer has reported that rs3135363 in the HLA-DR
region was associated with differed postvaccination antibody
titer (22). In our study, a GWAS and subsequent confirmation
focused on non-response to hepatitis B vaccination has also
detected that polymorphisms in the HLA-DR region did show
strong associations with non-response to hepatitis B vaccination.
Therefore, previous studies and our study have together
implicated that the HLA-DR region is very likely the key loci
associated with immune response to hepatitis B vaccination.
Hepatitis B vaccine is an exogenous antigen. It cannot be
presented to the CD4+ T lymphocyte unless it is degraded into
peptides and combined with HLA class II molecules (
HLA-DR, a subtype of the HLA class II molecule, is consisted
of two subunits HLA-DRA and HLA-DRB (
genes are highly polymorphic, especially the HLA-DRB genes.
It has been reported that allelic differences of HLA-DR genes
would affect the T-cell recognition of MHC-peptides complexes
in hepatitis B vaccine-triggered immunity, and then affect the
outcomes of immune responses (
). Furthermore, HLA-DR
genes are in high LD with numerous genes involved in
immunity. Although polymorphisms that located in non-coding
sequences of HLA-DR genes looks like no direct effect, it
HR, high-responder to primary vaccination (anti-HBs ≥ 1000 mIU/ml after three doses); NR, non-responder to booster vaccination (anti-HBs , 10 mIU/ml after six
doses); NR, non-responder to primary vaccination (anti-HBs , 10 mIU/ml after three doses); ∗the combined analysis of GWAS and confirmation stages.
cannot rule out that they might link to other causal loci affecting
the quality of immune response. Here, we detected that the
rs477515, which located in the upstream of HLA-DRB1 gene,
was shown to be an independent and the strongest signal
associated with non-response to hepatitis B vaccination. We
suspected that it might be a genetic marker in HLA-DR region and
in high LD with other function-related SNPs that affect the
T-cell recognition of MHC – peptide complexes on immune
cells, and then result in weak or no immune response to hepatitis
Previous twin study has reported that non-HLA account for
60% of genetic factors that influence the immune response to
hepatitis B vaccination (
). Furthermore, several case –
control studies have detected polymorphisms of genes involved
in immune response were associated with hepatitis B-induced
). However, the two GWA studies have
verified that the strongest association signals were located
within HLA region, and other SNPs located in non-HLA region
were shown to be weaker association signals. Immune response
elicited by a vaccine is a cumulative result of interactions driven
by genes in the immune response network, where HLA and other
immunity-relevant genes are both indispensable (29). In the
present study, genome-wide complex trait analysis (GCTA)
showed that heritability of response to hepatitis B vaccination
due to the whole-genome SNPs was estimated to be 56.7%,
while 5.62% for the associated 21 SNPs within HLA region.
This result implicated that the polymorphisms of HLA alone
do not explain all variations in the immune response to vaccines,
and the polymorphic contributions appear to be the norm (
It is conceivable that the effect of HLA polymorphisms might
be principal, and the effect of polymorphisms of non-HLA
genes themselves might be limited, or linked with HLA alleles
). Thus, our results did not deny the effect of non-HLA
genes, but implicated that antigen presentation and recognition
of T-cell receptor repertoire relying on the HLA molecules
might play a critical role in immune response triggered by
hepatitis B vaccine.
To the best of our knowledge, this study is the first GWA
study focusing on non-response to hepatitis B vaccination in
Chinese Han population. In the present study, we replicated
the previously reported rs3135363 within HLA-DR region in
our GWAS and confirmation Ia sets (
). However, we found
that rs477515 within the HLA-DR region was the strongest
association signal and was independently associated with
nonresponse to hepatitis B vaccination even after adjustment of
rs3135363 within the HLA-DR region, which implicated the
independent effects of rs477515 on immune response to hepatitis
B vaccination. The different associated SNPs in the same
HLA-DR loci that were detected in the two GWAS might be
due to the different ethnic populations in the two GWAS.
Furthermore, in previous GWAS, they focused on postvaccination
antibody titer and assigned the subjects to three subgroups
on the basis of antibody titers, rather than analyzing as a
binary trait. In contrast, we used the international standard that
the subjects with antibody ,10 mIU/ml was defined as
nonresponders and the subjects with antibody ≥1000 mIU/ml was
defined as high-responders (
). In addition, the
nonresponders in our GWAS stage were administered two rounds
of standard three dose of hepatitis B vaccine, which represented
a more precise and true non-response status. Since booster
nonresponders are the persons who have no protective antibody titers
after revaccination and remain high risk of infection with HBV,
studies focused on these subjects will be helpful and meaningful
to clarify the genuine polymorphisms associated with
nonresponse to hepatitis B vaccination and find the really useful
loci to improve the efficacy of vaccine. The previously reported
SNP rs9277535 within HLA-DP region and rs9267665 within
HLA-III region failed to show significant association with
nonresponse to hepatitis B vaccination in both our GWAS and
following confirmation stage. Together the GWAS and
confirmation sets, our study had .85% statistical power to detect
these two SNPs with an OR .2.0 at a significance level of
0.05; hence, the sample size might not be the critical reason for
failed replication, while these different results might owing to
ancestry difference or different grouping patterns in the two
We also investigated the known HBV infection associated
SNPs (rs3077, rs9277535, rs9277542, rs2856718, rs7453920,
rs652888 and rs1419881 within HLA region) (
33 – 36
hepatitis B-related hepatocellular carcinoma (HCC)-associated SNPs
(rs17401966 within 1p36.22, rs7574865 within STAT4,
rs455804 within 21q21.3, and rs9275319, rs9272105 within
HLA region) (
37 – 39
) with non-response to hepatitis B
vaccination in our GWAS discovery analysis. Due to the lack of
rs9277542, rs17401966 and rs9275319 in our GWAS data,
their linkage SNPs (rs9277533 instead of rs9277542, r2 ¼ 1.0;
rs3748578 instead of rs17401966, r2 ¼ 0.847; and rs9275371
instead of rs9275319, r2 ¼ 0.847.) were investigated in the
GWAS analysis. LD analysis identified that there is no evidence
of strong LD between rs477515 and each of the abovementioned
SNPs (r2 ≤ 0.27). We found the evidence of the association at
rs3077 (P-value ¼ 9.8 × 1023, OR ¼ 0.56), rs3748578
(P-value ¼ 8 × 1023, OR ¼ 1.84) and rs9272105 (P-value ¼
3.6 × 1024, OR ¼ 0.45) with non-response to hepatitis B
vaccination. The rs3077 showed both protective effects for response
to hepatitis B vaccination and for HBV infection, suggesting
the similar genetic background in these two phenotypes. These
results were consistent with previous report that the vaccine
response may provide a useful experimental model of a natural
infection for genetic study of infection, and the vaccine is least
effective in those who need it most (22). However, the minor
allele of the shard rs3748578 and rs9272105 between response
to hepatitis B vaccination and HCC showed protective effects
on response to hepatitis B vaccination, while as a risk effect on
development of HCC. Further studies will be needed to
demonstrate whether the opposite associations of response to
hepatitis B and HCC at these two SNPs are due to different causal
In previous GWAS and candidate studies, a total of 103 SNPs
have been reported to be significantly associated with immune
response to hepatitis B vaccination. In our study, we also
evaluated the genetic effects of these SNPs in our GWAS stage
(Supplementary materials, Table S8). Of these 103 SNPs, the
genotype data of 96 SNPs could be got from our GWAS array
or the imputation analysis. Among the 96 SNPs, 41 SNPs
showed significant association with non-response to hepatitis
B vaccination (P-values , 0.05), while there is no association
between other 55 SNPs and non-response to hepatitis B
vaccination. Given the frequencies of occurrence of the risk factors of
these 103 SNPs ranged from 1.4 to 54.2%, the present GWA
study demonstrated a statistical power to detect allelic
association of 11.5 – 63.5% with an OR of .2.0 at a significance
level of 0.05. Due to the actual proportion of booster
nonresponders is limited in a general population (
sample size in our GWAS was indeed limited, and further
decreased the statistical power. However, almost half of
previously reported SNPs were validated in our GWAS, and the
rs3135363 within HLA-DR region that reported in previous
GWAS was also replicated, as well as the consistent validation
that HLA-DRB1∗0701 allele was a risk factor on immune
). All these results together confirmed our
GWAS results. Furthermore, we also detected that rs477515
within HLA-DR region achieved a genome-wide significance
(P-gwas ¼ 4.81 × 1028) in GWAS stage, and was validated as
an independent association signal in the following confirmation
studies. These results might implicate that the rigorous
enrollment of boost non-responders may have helpful effect of
enriching for ‘non-response alleles’ in the GWAS. However, since
power is still an important factor in association study (
increasing sample size in order to improve the study power is a
necessary option in future studies. Due to the limited sample size in
our GWAS stage, stringent statistical significance level was
applied in our GWAS stage in order to control for false-positive
associations. However, such approach may simultaneously omit
the true associations that show modest effects, which might be
the reason why some of previously reported associations were
not replicated in our study. Furthermore, due to the small
impact of the associated 21 SNPs on heritability (5.62%) of
nonresponse to hepatitis B vaccination, further studies focusing on
whole-genome SNP results deserve to be performed. In the
present study, the confirmation studies were not strict
replications for the GWAS, because primary non-responders were
included in the confirmation stage and the subjects in
confirmation Ia and Ib sets, as well as the subjects in GWAS stage,
were coming from the same population. However, comparing
the frequencies of risk allele of each associated SNPs
(Table 2), we found that the order of frequency of risk allele
was booster non-responders . primary non-responders .
primary high-responders. These results were plausible and
wellfounded, because following parts of the primary non-responders
convert to high or normal responders after booster vaccination
), the frequency of risk allele in booster non-response group
must be increased. Once they were deviated from this order,
the associations might be suspicious. In the present study, all
the four SNPs were demonstrated to associate with booster
non-response as well as primary non-response to hepatitis B
vaccination and were in accordance with this order concurrently,
suggesting that the four SNPs were the true association signals
and implicating that it was plausible in some extent to use
primary non-responders to confirm the associations obtained in
using booster non-responders. In our GWAS, although there
are two kinds of doses (10/20 mg) in the first round of standard
vaccination, the primary non-responders were all revaccinated
with another round of standard vaccination; therefore the
different kinds of doses were unremarkable. Since the sample size in
our GWAS was not large, it is necessary to replicate this
association in a larger cohort. Furthermore, given that the genetic
structures are different among different geographical-origin
populations, further validations should be performed in other
In conclusion, we have detected that a polymorphism
(rs477515) within the HLA-DR region was strongly associated
with non-response to hepatitis B vaccination, and this
association was independent from rs3135363 that was reported
previously. Considering the importance of HLA region in immune
response to hepatitis B vaccination, our results implicate that
antigen presentation and recognition of T-cell receptor
repertoire on HLA-DR molecules might be critical in total process
of immune response triggered by hepatitis B vaccination.
Further in-depth analysis focusing on HLA-DR region might
be helpful to understand the mechanisms of immune response
to hepatitis B vaccination and detect novel loci for developing
more effective hepatitis B vaccine.
MATERIALS AND METHODS
Participants in the GWAS stage, together with confirmation Ia
and Ib stages were recruited from a large hepatitis B vaccination
campaign in Shandong province in 2009. Participants in the
confirmation II stage were recruited from among healthy volunteers
in the community in Beijing in 2007. Written informed consent
and completed questionnaires including demographic
information, smoking history, vaccination history, chronic disease and
immunosuppressive disease/medications were obtained from
all participants. All individuals were tested for five markers of
hepatitis B using an Abbott i2000 detection kit (Abbott
Laboratory, Chicago, IL, USA). Individuals who were negative for
the five markers of hepatitis B were tested further for HBV
DNA and for anti-HCV and anti-HIV. Participants were
excluded if: (i) they were positive for HBV DNA, hepatitis B
surface antigen (HBsAg), hepatitis B e antigen (HBeAg),
anti-HBs, anti-HBc, anti-HBe, anti-HCV and/or anti-HIV; (ii)
they had been vaccinated previously with any hepatitis B
vaccine; (iii) they had a chronic disease, such as diabetes,
cancer or cardio-cerebrovascular disease, or were undergoing
renal dialysis; (iv) they had any immunosuppressive disease or
were taking any immunosuppressive medication; (v) they were
not of Han ethnicity; (vi) they were ,18 years.
A total of 3985 qualifying individuals from Shandong
province were administered three doses of 10 or 20 mg recombinant
hepatitis B vaccines according to standard 0-1-6 schedule (20 mg
recombinant hepatitis B vaccine, GlaxoSmithKline Investment
Co., Ltd, UK; North China Pharmaceutical Co., Ltd, Beijing.
10 mg recombinant hepatitis B vaccine, Dalian Hissen
Biopharmaceutical Co., Ltd., Dalian). Levels of anti-HBs were
measured at 1 month after the final dose using the Abbott
i2000 detection kit. Four hundred and twenty-two primary
nonresponders (anti-HBs , 10 mIU/ml after three doses
immunization) were revaccinated with another standard three doses of
hepatitis B vaccines (20 mg recombinant hepatitis B vaccine,
GlaxoSmithKline Investment Co., Ltd, UK). Levels of
antiHBs were measured once again at 1 month after the final dose
using the Abbott i2000 detection kit. After completing the
revaccination schedule, 78 vaccines with an antibody titer ,10 mIU/
ml were booster non-responders and included in case group in
the GWAS stage. One hundred and thirteen primary high
responders (anti-HBs ≥ 1000 mIU/ml after three dose immunization)
were randomly selected as control group to match for vaccine
doses, age and gender ratio against case group. The remaining
374 primary non-responders, together with 1122 age and
gender matched primary high-responders were assigned to
confirmation Ia and Ib cohorts (confirmation Ia: 20 mg group, 263
primary non-responders and 825 primary high-responders;
confirmation Ib: 10 mg group, 111 primary non-responders and 297
A total of 599 qualifying individuals from Beijing were
administered three doses of 10 mg recombinant hepatitis B vaccine
according to standard 0-1-6 schedule (North China
Pharmaceutical Co., Ltd, Beijing, China). Levels of anti-HBs were measured
at 1 month after the final dose using the Abbott i2000 detection
kit. Forty-six primary non-responders and 214 primary
highresponders were assigned to confirmation II stage.
The demographic characteristics and clinical features of
subjects in the GWAS stage and the confirmation stages are
summarized in Supplementary Materials, Table S9.
The study was performed in accordance with the guidelines of
the Helsinki Declaration and was approved by the Ethics
Committee of the Institute of Basic Medical Sciences, Chinese
Academy of Medical Science.
SNP genotyping and quality control
Genomic DNA was extracted from peripheral blood using the
phenol – chloroform method. In the GWAS stage, 191 samples
(78 cases and 113 controls) were genotyped for 731 442 SNPs
using Illumina HumanOmniExpress BeadChip (Illumina, CA,
USA). The SNP with call rate ,95% (n ¼ 13 567), minor
allele frequency , 0.01 in both control and case group (n ¼
123 224) and/or P values of Hardy – Weinberg equilibrium
(HWE) test , 1 × 1025 in control group (n ¼ 8,454) were
excluded. The 191 samples were subsequently assessed for
population stratification using a PCA, and six genetic outlier
samples (one case and five controls) were excluded.
After quality control, 588 026 SNPs in 185 samples (77 cases
and 108 controls) were remained for the subsequent association
analyses. The association analysis was performed with and
without adjusting for the top five principal components
(generated in EIGENTSTRAT). A direct comparison was also
performed between the association results of the top 1000 SNPs
from the genome-wide analyses with PCA adjustment and the
one without adjustment (
The top 21 SNPs (P-values , 1 × 1026) were considered to be
further analyzed. LD analysis based on tagger pairwise method
using Haploview software showed that four SNPs (rs477515,
rs28366298, rs3763316 and rs13204672) were in high LD with
other SNPs (r2 . 0.8), respectively. Hence, the four SNPs
were further genotyped in the subsequent confirmation studies.
The previously reported three SNPs (rs3135363, rs9277535
and rs9267665) were also genotyped in our GWAS and
confirmation sets. All of these seven SNPs were genotyped using the
TaqMan-MGB (Genecore Biotech Co., Ltd, Shanghai, China)
or TaqMan-BHQ (Sangon Biotech Co., Ltd, Shanghai, China)
probe-based real-time polymerase chain reaction (PCR). The
primer and probe sequences that were used to genotype each
SNP are shown in Supplementary Material, Table S10.
Amplification and detection were conducted using a Bio-Rad iQ5
Multicolor Real-Time PCR Detection System (Bio-Rad, Hercules,
To confirm the genotyping results in the GWAS stage, 20% of
the samples from GWAS stage were selected randomly and were
replicated with the TaqMan-MGB or TaqMan-BHQ probes that
were used in the confirmation stages, and we obtained 100%
HLA-DRB1 region was genotyped using the direct
sequencebased genotyping method (PCR-SBT), among the 108
highresponders and 77 booster non-responders in the GWAS stage.
To generate additional genotypes, we performed imputation
using SHAPEIT [Shape-IT: new rapid and accurate algorithm
for haplotype inference] and IMPUTE2 [a flexible and accurate
genotype imputation method for the next generation of
genomewide association studies] using the genotype data from CHB and
CHS samples of the 1000 Genomes Project as reference. For all
datasets, cases and controls were imputed together. We
performed quality control for all 38 043 533 imputed SNPs. In
detail, we removed SNPs with low imputation quality (info
score ,0.8 for IMPUTE2) as suggested. We also removed
SNPs which showed MAF , 0.01 or showed call rate of
,95% or significant deviation from HWE in controls (the
same as for the genotyped SNPs). Many of the imputed SNPs
were removed from further analyses because of low call rate
and deviation from HWE, probably owing to low LD between
genotyped and imputed SNPs. Association test of non-response
to hepatitis B vaccination risk with each of these imputed SNPs
was carried out using PLINK software (version 1.07).
A x 2 goodness-of-fit test was used to examine whether genotype
distributions of each SNP were conformed toHWE in both case
group and control group. A PCA was performed to check for
signs of population stratification in GWAS stage using the
EIGENSTRAT software. The statistical significance of the
association with each SNP was assessed with and without correction
for the top five principal components (generated in
EIGENTSTRAT), using a logistic regression model. These calculations
were conducted with PLINK software (version 1.07). Regional
association plot was performed by the online software
). Pathway analysis was performed to explore the
most biologically relevant pathways impacted by a list of input
SNPs in the GWAS stage, using the improved gene-set-enrichment
analysis approach (i-GSEA4GWAS). GCTA (GCTA software,
version 1.04) was used to estimate the phenotypic variance
explained by the 21 statistically significant SNPs or
wholegenome-wide SNPs in the GWAS stage with correction for the
top five principal components (generated in EIGENTSTRAT)
). LD test was conducted using Haploview software
(version 4.2). The allele frequencies for each SNP were
compared between case and control groups using the x 2-test in the
confirmation stages, and the calculation was conducted using
the online software SHEsis (
). Logistic regression analysis
was also used to adjust for confounding factors and the previous
reported top SNP rs3135363 in the HLA-DR region (
independent test for the four SNPs was also performed by a
logistic regression model using forward selection procedure. These
analyses were performed using SPSS software (version 11.0).
Supplementary Material is available at HMG online.
We thank all the subjects who participated in this study.
Conflict of Interest statement. None declared.
This work was supported by grants from the Major Project of
National Science and Technology (No. 2008zx10002-001), the
National Basic Research Program of China (973 Program) (No.
2012CB519005) and the Shandong Medical Health Science
and Technology Development Programs (No. 2009QZ017).
1. Lai , C.L. and Yuen , M.F. ( 2008 ) Chronic hepatitis B - new goals, new treatment . N. Engl. J. Med ., 359 , 2488 - 2491 .
2. Shepard , C.W. , Simard , E.P. , Finelli , L. , Fiore , A.E. and Bell , B.P. ( 2006 ) Hepatitis B virus infection: epidemiology and vaccination . Epidemiol. Rev. , 28 , 112 - 125 .
3. Custer , B. , Sullivan , S.D. , Hazlet , T.K. , Iloeje , U. , Veenstra , D.L. and Kowdley , K.V. ( 2004 ) Global epidemiology of hepatitis B virus . J. Clin. Gastroenterol ., 38 , S158 - S168 .
4. Lu , F.M. and Zhuang , H. ( 2009 ) Prevention of hepatitis B in China: achievements and challenges . Chin. Med. J. (Engl.) , 122 , 2925 - 2927 .
5. Zuckerman , J.N. ( 1996 ) Nonresponse to hepatitis B vaccines and the kinetics of anti-HBs production . J. Med . Virol., 50 , 283 - 288 .
6. Boland , G.J., van Bommel, T. , Rulos-van den Berg , A., van den Berg, J.P., van Loon , A.M. and van Hattum , J. ( 2003 ) The efficacy of a two-dose hepatitis B vaccination scheme . Adv. Exp. Med . Biol., 531 , 185 - 190 .
7. Hollinger , F.B. ( 1989 ) Factors influencing the immune response to hepatitis B vaccine, booster dose guidelines, and vaccine protocol recommendations . Am. J. Med ., 87 , 36S - 40S .
8. Weber , D.J. , Rutala , W.A. , Samsa , G.P. , Santimaw , J.E. and Lemon , S.M. ( 1985 ) Obesity as a predictor of poor antibody response to hepatitis B plasma vaccine . JAMA , 254 , 3187 - 3189 .
9. Wood , R.C. , MacDonald , K.L. , White , K.E. , Hedberg , C.W. , Hanson , M. and Osterholm , M.T. ( 1993 ) Risk factors for lack of detectable antibody following hepatitis B vaccination of Minnesota health care workers . JAMA , 270 , 2935 - 2939 .
10. Barraclough , K.A. , Wiggins , K.J. , Hawley , C.M., van Eps , C.L. , Mudge , D.W. , Johnson , D.W. , Whitby , M. , Carpenter , S. and Playford , E.G. ( 2009 ) Intradermal versus intramuscular hepatitis B vaccination in hemodialysis patients: a prospective open-label randomized controlled trial in nonresponders to primary vaccination . Am. J. Kidney Dis ., 54 , 95 - 103 .
11. Jafarzadeh , A. , Zarei , S. and Shokri , F. ( 2008 ) Low dose revaccination induces robust protective anti-HBs antibody response in the majority of healthy non-responder neonates . Vaccine , 26 , 269 - 276 .
12. Zhuang , G.H. , Yan , H. , Wang , X.L. , Hwang , L.Y. , Wu , Q. , Wang , L.R. and Gao , H.Y. ( 2006 ) Hepatitis B revaccination in healthy non-responder Chinese children: five-year follow-up of immune response and immunologic memory . Vaccine , 24 , 2186 - 2192 .
13. Newport , M.J. , Goetghebuer , T. , Weiss , H.A. , Whittle , H. , Siegrist , C.A. and Marchant , A. ( 2004 ) Genetic regulation of immune responses to vaccines in early life . Genes Immun., 5 , 122 - 129 .
14. Chen , J. , Liang , Z. , Lu , F. , Fang , X. , Liu , S. , Zeng , Y. , Zhu , F. , Chen , X. , Shen , T. , Li , J. et al. ( 2010 ) Toll-like receptors and cytokines/cytokine receptors polymorphisms associate with non-response to hepatitis B vaccine . Vaccine , 29 , 706 - 711 .
15. Davila , S. , Froeling , F.E. , Tan , A. , Bonnard , C. , Boland , G.J. , Snippe , H. , Hibberd , M.L. and Seielstad , M. ( 2010 ) New genetic associations detected in a host response study to hepatitis B vaccine . Genes Immun. , 11 , 232 - 238 .
16. Godkin , A. , Davenport , M. and Hill , A.V. ( 2005 ) Molecular analysis of HLA class II associations with hepatitis B virus clearance and vaccine nonresponsiveness . Hepatology , 41 , 1383 - 1390 .
17. Pan , L. , Zhang , W. , Liang , Z. , Wu , X. , Zhu , X. , Li , J. , Li , T. , Wang , L. , Li , H. and Liu , Y. ( 2012 ) Association between polymorphisms of the cytokine and cytokine receptor genes and immune response to hepatitis B vaccination in a Chinese Han population . J. Med . Virol., 84 , 26 - 33 .
18. Wang , C. , Tang , J. , Song , W. , Lobashevsky , E. , Wilson, C.M. and Kaslow , R.A. ( 2004 ) HLA and cytokine gene polymorphisms are independently associated with responses to hepatitis B vaccination . Hepatology , 39 , 978 - 988 .
19. Yucesoy , B. , Sleijffers , A. , Kashon , M. , Garssen , J., de Gruijl , F.R. , Boland , G. J. , van Hattum, J. , Simeonova , P.P. , Luster , M.I. and van Loveren , H. ( 2002 ) IL-1beta gene polymorphisms influence hepatitis B vaccination . Vaccine , 20 , 3193 - 3196 .
20. Pan , L.P. , Zhang, W. , Zhang, L. , Wu , X.P. , Zhu , X.L. , Yan , B.Y. , Li , J.Y. , Xu , A.Q. , Liu , Y. and Li , H. ( 2012 ) CD3Z genetic polymorphism in immune response to hepatitis B vaccination in two independent Chinese populations . PLoS ONE , 7 , e35303 .
21. Zondervan , K.T. and Cardon , L.R. ( 2007 ) Designing candidate gene and genome-wide case-control association studies . Nat . Protoc., 2 , 2492 - 2501 .
22. Png , E. , Thalamuthu , A. , Ong , R.T., Snippe , H. , Boland , G.J. and Seielstad , M. ( 2011 ) A genome-wide association study of hepatitis B vaccine response in an Indonesian population reveals multiple independent risk variants in the HLA region . Hum. Mol. Genet ., 20 , 3893 - 3898 .
23. WHO. ( 1992 ) Informal Consultation on Quadrivalent Diphtheria- TetanusPertussis-Hepatitis B Vaccine: Final Report . World Health Organization, Geneva.
24. Hohler , T. , Reuss , E. , Evers , N. , Dietrich , E. , Rittner , C. , Freitag , C.M. , Vollmar , J. , Schneider , P.M. and Fimmers , R. ( 2002 ) Differential genetic determination of immune responsiveness to hepatitis B surface antigen and to hepatitis A virus: a vaccination study in twins . Lancet , 360 , 991 - 995 .
25. Li , Y. , Ni , R. , Song , W. , Shao , W. , Shrestha , S. , Ahmad , S. , Cunningham , C.K. , Flynn , P.M. , Kapogiannis , B.G. , Wilson, C.M. et al. ( 2009 ) Clear and independent associations of several HLA-DRB1 alleles with differential antibody responses to hepatitis B vaccination in youth . Hum. Genet ., 126 , 685 - 696 .
26. Kimman , T.G. , Vandebriel , R.J. and Hoebee , B. ( 2007 ) Genetic variation in the response to vaccination . Commun. Genet ., 10 , 201 - 217 .
27. Milich , D.R. ( 1988 ) T- and B-cell recognition of hepatitis B viral antigens . Immunol. Today , 9 , 380 - 386 .
28. Kruger , A. , Adams , P. , Hammer , J. , Bocher , W.O. , Schneider , P.M. , Rittner , C. and Hoehler , T. ( 2005 ) Hepatitis B surface antigen presentation and HLA-DRB1 ∗ -lessons from twins and peptide binding studies . Clin. Exp. Immunol. , 140 , 325 - 332 .
29. Ovsyannikova , I.G. and Poland , G.A. ( 2011 ) Vaccinomics: current findings, challenges and novel approaches for vaccine development . AAPS J ., 13 , 438 - 444 .
30. Poland , G.A. , Ovsyannikova , I.G. , Jacobson , R.M. and Smith , D.I. ( 2007 ) Heterogeneity in vaccine immune response: the role of immunogenetics and the emerging field of vaccinomics . Clin. Pharmacol . Ther., 82 , 653 - 664 .
31. Yucesoy , B. , Johnson , V.J., Fluharty , K. , Kashon , M.L. , Slaven , J.E. , Wilson, N.W. , Weissman , D.N. , Biagini , R.E. , Germolec , D.R. and Luster , M.I. ( 2009 ) Influence of cytokine gene variations on immunization to childhood vaccines . Vaccine , 27 , 6991 - 6997 .
32. Zuckerman , J.N. ( 2006 ) Protective efficacy, immunotherapeutic potential, and safety of hepatitis B vaccines . J. Med . Virol., 78 , 169 - 177 .
33. Kamatani , Y. , Wattanapokayakit , S. , Ochi , H. , Kawaguchi , T. , Takahashi , A. , Hosono , N. , Kubo , M. , Tsunoda , T. , Kamatani , N. , Kumada , H. et al. ( 2009 ) A genome-wide association study identifies variants in the HLA-DP locus associated with chronic hepatitis B in Asians . Nat. Genet ., 41 , 591 - 595 .
34. Mbarek , H. , Ochi , H. , Urabe , Y. , Kumar , V. , Kubo , M. , Hosono , N. , Takahashi , A. , Kamatani , Y. , Miki , D. , Abe , H. et al. ( 2011 ) A genome-wide association study of chronic hepatitis B identified novel risk locus in a Japanese population . Hum. Mol. Genet ., 20 , 3884 - 3892 .
35. Nishida , N. , Sawai , H. , Matsuura , K. , Sugiyama , M. , Ahn , S.H. , Park , J.Y. , Hige , S. , Kang , J.H. , Suzuki , K. , Kurosaki , M. et al. ( 2012 ) Genome-wide association study confirming association of HLA-DP with protection against chronic hepatitis B and viral clearance in Japanese and Korean . PLoS ONE , 7 , e39175 .
36. Kim , Y.-J., Kim , H.Y. , Lee , J.-H. , Yu , S.J. , Yoon , J.-H. , Lee , H.-S., Kim , C.Y. , Cheong , J.Y. , Cho , S.W. , Park , N.H. et al. ( 2013 ) Genome-wide association study identified new variants associated with the risk of chronic hepatitis B . Hum . Mol. Genet ., 22 , 4233 - 4238 .
37. Jiang , D.K. , Sun , J. , Cao , G. , Liu, Y. , Lin , D. , Gao , Y.Z. , Ren , W.H. , Long , X.D. , Zhang, H., Ma , X.P. et al. ( 2013 ) Genetic variants in STAT4 and HLA-DQ genes confer risk of hepatitis B virus-related hepatocellular carcinoma . Nat. Genet ., 45 , 72 - 75 .
38. Li , S. , Qian , J. , Yang , Y. , Zhao , W. , Dai , J. , Bei , J.X. , Foo , J.N. , McLaren , P.J. , Li , Z. , Yang , J. et al. ( 2012 ) GWAS identifies novel susceptibility loci on 6p21.32 and 21q21.3 for hepatocellular carcinoma in chronic hepatitis B virus carriers . PLoS Genet ., 8 , e1002791 .
39. Zhang , H., Zhai , Y. , Hu , Z. , Wu , C. , Qian , J. , Jia , W. , Ma , F. , Huang , W. , Yu , L. , Yue , W. et al. ( 2010 ) Genome-wide association study identifies 1p36.22 as a new susceptibility locus for hepatocellular carcinoma in chronic hepatitis B virus carriers . Nat. Genet ., 42 , 755 - 758 .
40. Clemens , R. , Sanger , R. , Kruppenbacher , J. , Hobel , W. , Stanbury , W. , Bock , H.L. and Jilg , W. ( 1997 ) Booster immunization of low- and non-responders after a standard three dose hepatitis B vaccine schedule - results of a post-marketing surveillance . Vaccine , 15 , 349 - 352 .
41. Shokri , F. and Amani , A. ( 1997 ) High rate of seroconversion following administration of a single supplementary dose of recombinant hepatitis B vaccine in Iranian healthy nonresponder neonates . Med . Microbiol. Immunol., 185 , 231 - 235 .
42. Zhang , F.R. , Huang , W. , Chen , S.M. , Sun , L.D. , Liu, H. , Li , Y. , Cui , Y. , Yan , X.X. , Yang , H.T. , Yang , R.D. et al. ( 2009 ) Genomewide association study of leprosy . N. Engl. J. Med ., 361 , 2609 - 2618 .
43. Pruim , R.J. , Welch , R.P. , Sanna , S. , Teslovich , T.M. , Chines , P.S. , Gliedt , T.P. , Boehnke , M. , Abecasis , G.R. and Willer , C.J. ( 2010 ) LocusZoom: regional visualization of genome-wide association scan results . Bioinformatics , 26 , 2336 - 2337 .
44. Yang , J. , Lee , S.H. , Goddard , M.E. and Visscher , P.M. ( 2011 ) GCTA: a tool for genome-wide complex trait analysis . Am. J. Hum. Genet ., 88 , 76 - 82 .
45. Shi , Y.Y. and He , L. ( 2005 ) SHEsis, a powerful software platform for analyses of linkage disequilibrium, haplotype construction, and genetic association at polymorphism loci . Cell Res ., 15 , 97 - 98 .