Article Text
PDF
Abstract
Objective To investigate the molecular basis of occult hepatitis B virus (HBV) infection (OBI) in Asian blood donors.
Design OBI donors from Hong Kong, Malaysia, Singapore, Taiwan and Thailand were tested for HBV serological markers, and strains were molecularly characterised.
Results Among 138 confirmed OBI carriers (median age 47 years), HBV genotypes B and C were dominant (60% and 34%, respectively) in agreement with the genotype distribution in chronically infected donors in the region. Viral load ranged between unquantifiable and 3670 IU/ml (median 11 IU/ml). Eleven per cent of OBIs showed an unusual anti-HBs-only serological profile without evidence of past vaccination for most of these individuals. Occult HBV strains showed a higher genetic diversity than strains from matched hepatitis B surface antigen (HBsAg)+ donors, irrespective of genotype. No unique genetic signature or evidence of reduced replication competence was found. Mutations in the vicinity of the pre-S2/S splice donor site were common in OBIB (44%) and OBIC (36%) strains. S regions from four OBI cases were transfected in HuH7 cells. Results showed limited HBsAg secretion and suggested that mutations disrupting the splice donor site structure may affect pre-S2/S mRNA splicing.
Conclusions There is indirect evidence that incomplete immune control is involved in the occurrence of OBI in Asian blood donors infected with genotypes B and C as observed in Europe with genotype A2 but to a lower extent than with genotype D. A post-transcriptional mechanism may play a role in HBsAg expression in some OBIs irrespective of HBV genotype.
- Hepatitis B virus
- occult HBV infection
- blood donors
- RNA splicing
- diagnostic virology
- hepatitis C
- hepatitis B
- molecular epidemiology
- chronic viral hepatitis
- hepatitis
- hepatitis D
- hepatitis E
Statistics from Altmetric.com
- Hepatitis B virus
- occult HBV infection
- blood donors
- RNA splicing
- diagnostic virology
- hepatitis C
- hepatitis B
- molecular epidemiology
- chronic viral hepatitis
- hepatitis
- hepatitis D
- hepatitis E
Significance of this study
What is already known about this subject?
-
Occult hepatitis B virus (HBV) infection/carriage (OBI) in asymptomatic blood donors is a new facet of HBV infection natural history observed worldwide, with prevalence varying according to HBV level of endemicity and HBV genotype.
-
Mechanisms leading to OBI include imperfect host's immune response (in Europe) and viral factors (in South Africa).
-
There is emerging evidence of the potential clinical relevance of OBI, since it has been involved in HBV transmission through liver transplant and anecdotally by blood transfusion, in HBV re-activation in immunodeficient patients, and as a risk factor for hepatocellular carcinoma.
What are the new findings?
-
A higher genetic diversity was confirmed in occult HBV genotype B and genotype C strains from South East Asian blood donors compared with strains from hepatitis B surface antigen (HBsAg)+ chronic carriers, but no unique genetic signature for OBI was found.
-
There is indirect evidence that incomplete immune control or post-transcriptional mechanisms involving S gene splicing are involved in the occurrence of OBI in Asian blood donors infected with genotypes B and C as observed in Europe with genotype A2 but to a lower extent than with genotype D, rather than reduced replication competence associated with specific genetic defects as reported in South African OBIA1 carriers.
-
Look back investigations conducted on a relatively large number of OBI cases (n=51) support previous observations from a single case that very low HBV DNA levels, either consistently or intermittently detected, is a common characteristic of OBI blood donors infected with HBV genotype B or genotype C, raising issues regarding strategies for blood testing.
-
Several OBI samples contain anti-HBs as the only antibody marker in non-vaccinated individuals.
Significance of this study
How might it impact on clinical practice in the foreseeable future?
-
These findings emphasise the critical importance of test sensitivity to identify OBIs that are potentially infectious by transfusion and the need to repeat testing at intervals in a given individual.
-
Anti-HBc screening does not reliably detect OBI.
-
HBV genotype seems to influence levels of viral markers as well as frequency of OBI.
-
Determination of mechanisms leading to reduced viral replication and HBsAg production may lead to new approaches to antiviral therapy.
Introduction
Occult hepatitis B virus (HBV) infection/carriage (OBI) is characterised by the presence of very low levels of HBV DNA in plasma and/or in liver, with undetectable hepatitis B surface antigen (HBsAg) using the most sensitive commercial assays, with or without antibodies to hepatitis B core antigen (anti-HBc) or hepatitis B surface antigen (anti-HBs), outside the pre-seroconversion window period.1 OBI prevalence varies between different geographic areas and populations according to HBV endemicity and HBV genotype.2–4 In South East Asia, where genotypes B and C are prevalent, up to 90% of the population has evidence of past exposure to HBV.5–7 The prevalence of OBI previously reported among blood donors in Asia is 1:570–1:7517 in China,4 ,8 1:3248 in Hong Kong,9 1:894–1:1029 in Taiwan7 ,10 and 1:3832 in Thailand.11
HBV DNA load quantification, full-length viral genome obtained in plasma, documented transfusion transmission, and re-activation of OBI indicate a sustained but low viral replication in most OBI carriers.1 However, the virological features and the mechanisms leading to OBI remain unknown. OBI may be related to: (1) mutations in the HBV S proteins affecting in vitro antigen detection8 ,12 ,13; (2) mutations in S promoters that may decrease the circulating HBsAg level14–16; (3) imperfect control of HBV replication by the host humoral and cellular immune response3 ,17; (4) various mutations in genomic regulatory elements that may negatively affect viral replication2; and (5) mutations affecting post-transcriptional mechanisms regulating HBsAg expression and secretion.18 In a genotype A-infected patient with OBI, the lack of HBsAg expression was apparently caused by a single mutation (G458A) at a 5′ splice site of the pre-S2/S mRNA.18 This mutation interfered with pre-S2/S mRNA splicing and reduced the accumulation of functional unspliced pre-S2/S mRNA. However, splicing would yield non-functional S mRNA, eventually coding for truncated or hybrid HB proteins. How splicing is essential for the optimal production of functional unspliced transcript and HBsAg expression is still unclear, but it may involve co- or post-transcriptional mechanisms including processing, nuclear export, and/or stability of the unspliced mRNA.
This study reports on the clinical and molecular characteristics of OBI with genotype B and genotype C in blood donors from South East Asia.
Materials and methods
Collaborative centres and sample identification
Blood centres participating in the study were from Hong Kong Red Cross Blood Transfusion Centre, Hong Kong, PRC, the National Blood Centre, Thai Red Cross Society, Bangkok, Thailand, Taiwan Blood Services Foundation, Taipei, Taiwan, the Centre for Transfusion Medicine, Health Science Authority, Singapore, and the National Blood Centre, Kuala Lumpur, Malaysia. The study was approved by the Hong Kong Red Cross Blood Transfusion Centre Internal Review Board.
Samples were from HBsAg−/HBV DNA+ donors. Donations were tested for HBV DNA individually with the PROCLEIX TIGRIS/PROCLEIX ULTRIO assay (Chiron/Gen-Probe, Emeryville/San Diego, California, USA) or in minipools of six with the Roche cobas s 201/cobas TaqScreen multiplex (Roche Diagnostics, Mannheim, Germany). The reported 95% HBV DNA detection limits of the two assays were 10 and 3 IU/ml, respectively.7 ,11
Blood donation archive samples were retrieved for 51 blood donors with OBI in Hong Kong and were included in the study.
HBV serological testing
Each collaborating centre provided results of serological HBV markers. HBsAg was tested with PRISM (Abbott Laboratories, Abbott Park, Illinois, USA). In OBI-affected donors, anti-HBc IgM/IgG and anti-HBs IgG were tested with Architect (Abbott Diagnostics, Sligo, Ireland), and alanine aminotransferase (ALT) with the ALT Roche/Hitachi cobas C System. The healthy upper limit for serum ALT was considered to be 35 IU/l in Asian populations.19
HBV DNA analyses
Each initially reactive sample was further tested for HBV DNA by real-time quantitative PCR (QPCR) and nested PCR of different regions of the HBV genome from DNA purified either directly from 0.5 ml plasma or after ultracentrifugation of 2–10 ml plasma as previously described.3 The limit of detection and the limit of quantification of the QPCR assay were 5 and 10 IU/ml, respectively. Amplified products were sequenced, and sequences were compared with sequences obtained from 124 and 93 HBsAg+ blood donors from the same populations infected with HBV genotype B (Hong Kong, 59; Malaysia, 35; Taiwan, 22; Thailand, 8) and genotype C (Hong Kong, 14; Malaysia, 21; Taiwan, 2; Thailand, 56), respectively.
RNA structure prediction
RNA secondary structures were predicted using the program MFOLD (version 3.0) at the server (http://mfold.bioinfo.rpi.edu/cgi-bin/rna-form1.cgi).
In vitro HBsAg expression
A subgenomic fragment (positions 156–1803) containing the S coding region and the HBV post-transcriptional regulatory element (PRE) was PCR-amplified by using SPL3 (5′-GCGCGCGCTAGCACCATGGGGARCAYCRYATCRGGA-3′; nucleotides 513–1534) and SPL2 (5′-GCCTTTGCAAGCTTCASACCAATTTATGCCTAC-3′; nucleotides 3161–3143) primers, and cloned into pcDNA3.1+ vector (Invitrogen, Carlsbad, California, USA) under the control of the human cytomegalovirus (CMV) immediate-early promoter. HBsAg constructs were transfected into HuH7 cells.20 Transfection efficiency was evaluated by co-transfection of a reporter plasmid expressing LacZ and determination of β-galactosidase enzymatic activity in culture. To test for intracellular HBsAg, cytosolic extract was obtained by washing transfected cells three times in 1×phosphate-buffered saline (PBS) followed by centrifugation at 13 000 rpm for 15 min in a lysis buffer containing 1×PBS, 1% Triton X-100, 10 U/ml DNase I and cOmplete Proteinase Inhibitor Cocktail Tablets with EDTA (Roche). HBsAg was tested in culture supernatants and cytosolic extracts by using Murex HBsAg v3 immunoassay (Abbott).
Detection of spliced RNA
Total RNA was purified from transfected cells by using RNeasy Mini kit (QIAGEN, Hilden, Germany), amplified by reverse transcriptase (RT)-PCR using primers SPL4 (5′-ACCACAGAGTCTAGACTYGTGGT-3′; nucleotides 1597–1619) and SPL5 (5′-GGTCGGAACRRCAGRCGRAGAAG-3′; nucleotides 2878–2852), and sequenced.
Statistical analysis
Categorical variables were compared using the Fisher exact test and, for continuous variables, the non-parametric Mann–Whitney test. The significance of HBsAg production in the cell transfection experiment was assessed by using a one-way analysis of variance with Tukey's multiple comparison (post hoc) test. p<0.05 was considered significant.
Results
Sample classification
Initial screening of blood donations identified 177 HBV DNA+/HBsAg− donors (92 from Hong Kong, 10 from Malaysia, 15 from Singapore, 25 from Taiwan and 35 from Thailand). Further PCR testing of each reactive sample and of archive samples from 51 repeat donors from Hong Kong collected 3–15 months before the index donation confirmed the presence of HBV DNA in 167 (94%) samples (table 1). Seventy-six (45.5%) DNA+ samples were reactive for anti-HBc only, and 47 (28%) carried both anti-HBc and anti-HBs. Fifteen samples (10.9%) were reactive for anti-HBs only yielding a total of 138 samples that were classified as OBI (table 1). Owing to incomplete antibody or follow-up data, 29 samples that were confirmed HBV DNA+ remained unclassified, as window period infection could not be excluded. Ten unconfirmed HBV DNA samples (eight containing anti-HBc and/or anti-HBs) were considered unclassifiable. The prevalence of confirmed occult HBV carriage was 1:4255, 1:13 602, 1:12 807 and 1:22 575 in Hong Kong, Malaysia, Thailand and Singapore, respectively. The prevalence of OBI was 1:640 in anti-HBc+ Taiwanese donors.
Molecular and serological confirmation of hepatitis B virus (HBV) infection in index DNA-positive blood donor samples
Characterisation of OBI samples
Among the 138 OBI samples, 60 BCP/PC (43%), 98 pre-S/S (71%) and 49 whole HBV genomes (35.5%) were PCR-amplified and sequenced. Phylogenetic analysis of the pre-S/S sequences or the BCP/PC sequences in the absence of pre-S/S showed that three sequences clustered with genotype A2, 68 with genotype B, 39 with genotype C, and four with genotype D. Genotype B was dominant (60%) in all countries except Thailand (table 2). In this study, donors carrying OBI were significantly older (median 47 years; range 17–64) than 330 HBsAg+ donors (median 29 years; range 18–61) irrespective of geographic origin (p=0.0001) and of the overall donor median age in the five countries involved (39, 27, 30, 33 and 28 years in Malaysia, Hong Kong, Singapore, Taiwan and Thailand, respectively). The majority of HBV-infected donors were male (73% and 59% of OBI and HBsAg carriers, respectively). Among the donor populations studied, the proportion of male donors varied between 48% and 59.6% (data not shown). Six OBI donors had elevated ALT level (≥35 IU/l; range 41–78 IU/l). The median HBV DNA load of the 138 OBI donors was 11 IU/ml ranging between <10 and 3670 IU/ml, when samples giving no QPCR signal but confirmed DNA positive as containing <10 IU/ml were included (table 2). The OBI median viral load was significantly lower than in HBsAg+ controls (944 IU/ml; range <10−2–109 IU/ml; p=0.0001). No significant difference in viral load distribution was observed according to age, genotype or serological profile (table 3), except for the 15 samples carrying anti-HBs only (median 49 IU/ml; range <10–3670; p=0.011). These 15 OBI donors were significantly younger (p<0.0001) than those with other serological profiles (table 3).
Characterisation of 138 occult hepatitis B virus (HBV) infections from South East Asia
Host and viral markers in 135 OBI donors stratified according to serological status*
A look back study of HBV DNA load was conducted in previous donations from 51 OBI donors (two to five donations per donor over a ≤20-month period). In 23 OBI donors (45%), viral load was quantified in all donations and fluctuated between <10 and 219 IU/ml, whereas, in 25 donors (49%), HBV DNA was not detected in some of the archived samples but was successfully quantified in others (<10–99 IU/ml). Limited sample volume preventing viral particle concentration, HBV DNA remained consistently undetectable in three anti-HBc+ donors (6%), two of them carrying anti-HBs (59 and 430 IU/l) classification in either group could be done. Anti-HBs was present in 72% of OBI donors with randomly detectable viral DNA.
Pre-S/S region analysis
The pre-S1, pre-S2 and/or S amino acid sequences of 55 OBI genotype B (OBIB) and 34 OBI genotype C (OBIC) strains were analysed. OBIB and OBIC sequences were compared with their respective counterparts obtained concomitantly from 124 and 93 HBsAg+ blood donors infected with HBV genotype B and genotype C, respectively. No specific mutations were more commonly present in OBI than in HBsAg+ strains. The average amino acid diversity over the long surface protein was similar in OBIB and OBIC strains, but significantly higher in OBI than in HBsAg+ strains irrespective of HBV genotype (table 4). In both OBIB and OBIC groups, the amino acid diversity was significantly higher in immunoreactive regions, including the major hydrophilic region (MHR) and 12 putative CD8+ cytotoxic T lymphocyte (CTL) epitopes, than in the non-immunoreactive regions (table 4 and figure 1). Substitutions were also significantly more common in the MHR than CTL epitopes (10.1% vs 7.2% for OBIB strains (p<0.0001) and 11.6% vs 7.1% for OBIC strains (p<0.0001)). Eight OBIB (six anti-HBs+/two anti-HBs−) and 10 OBIc (five anti-HBs+/five anti-HBs−) MHR sequences were identical with the corresponding consensus sequence derived from the alignment of HBsAg+ strain sequences. Multiple amino acid substitutions were observed in OBIB and OBIc immunoreactive domains, including several not found in HBsAg+ controls, although many were only present in one OBI strain. Several substitutions occurred at sites within the MHR that have been previously identified as affecting antigenicity, immune escape, or infectivity as indicated in figure 1A. Among these critical sites were four cysteines and four sG145R substitutions. One OBIB strain had sC121Y substitution, one had dual substitutions sC137W + sG145R, seven had sG145A/R, and one had an additional threonine residue inserted at position s115. Substitutions sC124S and sG145A/D/K were present in one and three OBIC strains, respectively. Pre-S2 deletions of amino acids 6–15 and 13–22 were observed in one and two OBIB strains, respectively. One OBIC strain had amino acids 39–55 and 1–5 deleted in pre-S2 and S, respectively.
Average intra-group amino acid diversity (range) of occult HBV genotype B and C
Amino acid variability in major hydrophilic region (MHR) (A) and pre-S/S cytotoxic T lymphocyte (CTL) epitopes (B) of occult hepatitis B virus (HBV) infection (OBI) and hepatitis B surface antigen (HBsAg)+ strains. Black bars and white bars characterise OBI and HBsAg+ strains, respectively. Values are given as percentage of strains with substitution at a given position compared with a genotype-specific consensus sequence. Residues previously identified as affecting antigenicity and infectivity are indicated by ↑ and *, respectively.
Analysis of the pol, pre-core/core and X regions
Both OBIB and OBIC Pol sequences had an overall amino acid diversity similar to the sequences of HBsAg+ controls that distributed across the whole protein (table 4). One OBIB sequence had a deletion of one nucleotide (nucleotide position 930) in the pol gene which introduced a frameshift in the open reading frame resulting in a stop codon (amino acid position 625).
Similar frequency of mutations preventing HBeAg synthesis was observed in OBI cases and controls. Mutations abolishing the pre-core start codon (A1814C ×3 and A1814G ×1) and insertion of a nucleotide A at position 1839 (two cases) creating a stop signal terminating the HBeAg were found only in OBIB strains.
Amino acid variability within the Core protein was significantly higher in 33 OBIB sequences than in controls, but this was the reverse between OBIC and control sequences (table 4). Within the nucleic acid-binding domain, mutation R151G/Q was detected in seven OBIB sequences, the double mutations R149Q and R152G in one OBIB sequence, and R150S and R152G in one OBIC sequence. The phosphorylation state of the Core protein is important for HBV replication. Serine substitutions were observed in three OBIB sequences (S155T, S168F and S176T) and in one OBIC sequence (S178C + S181C), but in none of the control sequences. No critical mutation was observed in the core sequences of seven anti-HBs-only OBIs.
The average amino acid variability of the HBx protein was significantly higher in OBIB than in controls, but it was the reverse for genotype C strains (table 4). Mutations A44V/V125I and A47T were commonly found in OBIB and OBIC sequences, respectively. HBx protein was truncated by a stop codon at position 101 in one OBIB sequence.
Analysis of regulatory elements
No major difference regarding frequency and nature of nucleotide substitutions within the pre-core/pre-genomic, S1, S2 and X promoters, the enhancers I and II, and the cis-acting negative regulatory element was observed in OBI (33 OBIB and 16 OBIC) and control sequences.
Recent data from Hass and co-workers suggested that pre-S2/S mRNA splicing might be essential for HBsAg expression.18 Pre-S2/S mRNA splicing is controlled by a 5′ splice donor site (nucleotides 426–464) and the PRE that contains the 3′ splice acceptor site. While the PRE was mainly conserved in both OBIs and controls, 25/55 (45%) OBIB and 14/33 (42%) OBIC sequences presented mutations in the vicinity of the 5′ splice donor site, compared with 5/47 (11%) and 5/48 (10%) of genotype B and C controls, respectively (p<0.0001). The predicted RNA secondary structure in this region for each variant was compared with the structure calculated for wild-type sequences. Genotypes B and C wild-type sequences were predicted to fold similarly in a stem–loop secondary structure. Among OBI variants, 11/25 (44%) OBIB and 5/14 (36%) OBIC splice donor mutations disrupted this predicted stem–loop structure (figure 2). Viral load was not significantly different in the 16 OBI donors carrying critical mutations compared with other OBI donors. None of the few substitutions found in HBsAg+ control variants was predictive of affecting the predicted stem–loop structure.
Nucleotide mutations at the S 5′ donor splice site of occult hepatitis B virus (HBV) infection (OBI) genotype B and C strain. Within each genotype, sequences from OBI strains were aligned, and a consensus sequence was derived. OBI sequences differing from the consensus are presented. The corresponding secondary structures were predicted using MFOLD. Alteration of the RNA secondary structure is reported on the right. Nucleotide substitutions affecting the predicted RNA secondary structure at the splice site are in bold. An arrow indicates the splice site (position 458).
To test whether mutations adjacent to the 5′ splice donor site influenced splicing and HBsAg production, cloned S gene-coding sequences including PRE isolated from two OBI strains with predicted disrupted stem–loop structure (OBIB HK8663 and OBIC HK8442), and from two OBI strains without predicted disrupted stem–loop structure (OBIB HK3110 and OBIC TW6083) were used in transient transfection experiments in HuH7 cells (figure 3A). Wild-type sequences of two genotype B (M88 and M90) and two genotype C (M86 and M95) HBsAg+ strains were used as controls. Both unspliced and spliced forms of S mRNA differing in length were detected in all transfected cells 3 days after transfection (figure 3B) and were confirmed by sequencing. No PCR amplification was obtained when the RT step was omitted. Sequences showed identical splicing positions from nucleotides 458 to 1305 and/or 458 to 1385 in all HBV S RNA tested except M90. M90 RNA was alternatively spliced from positions 459 to 1309 and 459 to 1361. The higher yield of unspliced RNA observed with HK8663 and HK8442 OBI strains suggested a reduced efficiency in splicing compared with OBI strains and controls without putative structural mutations (figure 3B). HBsAg was released into the cell culture supernatant (figure 3C). However, the relative amount of HBsAg secreted from cells transfected with OBIs HK8663, HK3110 and TW6083 genes was significantly lower than that from both genotypes B and C control transfected cells (p<0.001). In contrast, HBsAg was detected at similar level in the supernatant of cells transfected with HK8442 and controls. A similar difference in intracellular HBsAg levels was also observed between cells transfected with OBI and control sequences except for the HK3110 construct (data not shown). HK3110-transfected cells showed an intracellular level of HBsAg slightly lower than controls, but was apparently defective in HBsAg secretion (figure 3C).
S mRNA splicing and hepatitis B surface antigen (HBsAg) production in HuH7 cells transfected with HBs constructs from genotype B and C occult hepatitis B virus (HBV) infection (OBI) and control HBV strains. (A) Predicted RNA structure of the 5′ splice site (position 458) region of wild-type HBVB (M88 and M90) and HBVC (M86 and M95) controls and OBIB (HK3110 and HK8663) and OBIC (HK8442 and TW6083) using MFOLD. Mutations are circled, and an arrow indicates the splice site. (B) Total cellular RNA was extracted from HuH7 cells transfected with wild-type controls and OBI S constructs. For each sample, 500 ng DNAse-treated RNA was tested for the presence of spliced and/or unspliced S mRNA by reverse transcriptase PCR. Amplified products were analysed by electrophoresis in a 1.2% agarose gel and SYBR green staining (Safe View; NBS Biologicals Ltd, Huntingdon, UK). (C) HBsAg levels in transfected cell supernatants were measured by enzyme-linked immunoassay. The relative level of HBsAg was expressed as the log value of the absorbance measured for each sample divided by the reaction cut-off. The means and standard deviations of six independent cultures are shown.
Discussion
The majority (94%) of the HBV DNA+/HBsAg− blood donor samples collected from five different Asian countries were confirmed as HBV-infected, and 78% of these samples were identified as OBIs (table 1). Lack of confirmation and discrepancies between the different HBV DNA assays used may be related to (a) limited sample volume preventing viral particle concentration, (b) viral load below the confirmatory assay detection limit, (c) stochastic sampling variation, or (d) nucleic acid testing (NAT) screening false positive. Twenty-nine confirmed samples remained unclassified, as they could not be differentiated between window period and seronegative primary or transient OBI because of the lack of follow-up/archive samples or complete serological data.21 Among these 29 unclassified cases, eight were from Hong Kong, four from Malaysia, four from Singapore, and 13 from Thailand (table 1). A percentage of window period infection of 2.5 and 21 was observed in Hong Kong (unpublished data) and was previously reported in Thailand.22 Therefore it may be estimated that at least 18 of the 21 HBV-seronegative/DNA-positive donors from Hong Kong and Thailand were seronegative OBIs. No window period samples were reported in Malaysia and Singapore (unpublished data).
Compared with African and European OBI blood donors, Asian OBIB and OBIC carriers shared a male dominance, normal ALT levels, low viral load, and older age than HBsAg+ donors (table 2).2 ,3 Asian OBIB and OBIC carriers were older (47 years) than South African OBIA1 donors (31 years), but younger than European OBIA2 or OBID donors (51 years). These results differed from a recent study reporting that the average age of OBI blood donors from southeast China (Fujian Province) was 30 years.8 Yuan and co-authors found that the prevalence of genotype C was significantly higher in OBIs than in HBsAg+ asymptomatic carriers or patients with chronic hepatitis, whereas genotype B has been reported to be dominant in the Fujian Province.8 In the present study, the overall distribution of OBIs between genotype B and C was 60% and 34%, respectively (table 2), and reflected the general HBV genotype distribution observed in matched HBsAg+ blood donors—genotype B being dominant in Hong Kong, Malaysia, Singapore and Taiwan, and genotype C in Thailand. The discrepancies between the two studies may be related to the difference in the number of samples analysed (138 vs 30 cases).
Irrespective of genotype, 45% of OBI donors were positive for anti-HBs (table 3) as reported in OBIs in Europe and South Africa.2 ,3 A higher frequency of anti-HBs+ OBIs was observed in Shenzen, China.4 The concomitant presence of HBV DNA and anti-HBs as the only markers of HBV infection in 11% of the OBI donors was intriguing. This phenomenon was not restricted to a particular HBV genotype (table 3). Similar cases were reported in one (∼1%) genotype A2-infected OBI donor from Poland, in four (7%) South African donors probably infected with genotype A1, and in two vaccinated Chinese donors probably infected with genotype C.2–4 No vaccination was documented in 9/12 anti-HBs-only OBI donors for whom the information was available. Acute-phase vaccine breakthrough may be suspected in three immunised donors, but, in the absence of follow-up, this point could not be clarified. No difference in age and viral load between vaccinated and unvaccinated donors was observed. No particular genetic feature in the BCP/PC and core sequences could be associated with a lack of Core antigen production or altered antigenicity. This unusual serological profile was significantly associated with younger age of carriers and higher HBV DNA load (table 3), suggesting a shorter infection history compared with anti-HBc+ OBIs. Further studies are needed to document the prevalence among blood donors of anti-HBs-only/HBV DNA+ OBI carriers and their potential infectivity, and to characterise the virological and immunological mechanisms responsible for this unusual HBV infection profile.
A common feature among OBI carriers is the extremely low viral load irrespective of HBV genotypes and epidemiological settings.1–4 Look back analysis showed that viraemia can be either consistently or intermittently detected, raising issues regarding blood testing strategies and supporting previous observations from an individual case of OBI transfusion transmission.23–26 Low viraemia may be related to imperfect control of HBV replication by the host humoral and cellular immune defences, as suggested by the presence of anti-HBs in 72% of OBI donors with randomly detectable viral DNA. In addition, compared with HBsAg+ strains, a significantly higher frequency of mutated humoral and cellular epitopes in the large surface protein of OBI genotypes B or C strains was observed (figure 1). A similar observation has been reported in OBI donors infected with HBV genotype D and A2 and in OBI donors from southeast China infected with genotypes B and C, but not in donors from Shenzen, China.3 ,4 ,8 MHR diversity in OBIB and OBIC strains was lower than in OBID strains, similar to that in OBIA2 strains, and higher than in OBIA1 strains.2 ,3 Immune pressure may play a role in the occurrence of OBI carriage, but with differences according to HBV genotype.
As OBI is defined by a lack of detectable HBsAg in plasma and low viral load, it is intuitively attractive to investigate mutations in the S gene and key regulatory regions for plausible mechanisms that could impair HBsAg detection, S gene expression, and viral replication. In many studies, mutations associated with OBI have often been identified without a robust comparison with relevant control cases. Consequently, it was not possible to exclude natural polymorphism and/or differences related to the geographical origin of the patients, clinical status of the HBV-infected individuals, or tissue source of the virus.27 In order to overcome these limitations, we compared OBI sequences with genotype-matched sequences obtained from HBsAg+ asymptomatic, apparently healthy blood donors identified during the same HBV-screening process (table 4). Detailed analysis of HBV structural proteins and key regulatory regions showed multiple mutations in OBIB and OBIc strains, but the overall nature and locations of these mutations seemed to reflect the polymorphic pattern observed in HBsAg+ strains. No hotspots of mutations shared by most of the OBI strains and absent in HBsAg+ strains were identified. The majority of the mutations observed were only present in unique OBI strains. However, it is possible that the accumulation of mutations across the whole HBV genome or unrecognised combinations of mutations may play a role in the genesis of OBI carriage. Approximately 52% of HBsAg+ donors were <30 years old and likely to be in the immunotolerance phase of chronic HBV infection, accounting for the relatively low sequence diversity. HBe antigenaemia has been reported in 60% and 80% of individuals <30 years old infected with genotype B and genotype C, respectively.28 It can also be argued that the high level of viral replication observed in these individuals in the present study (median viral load 2.4×104 IU/ml) may potentially favour the accumulation of mutations. However, as HBV infection mainly occurs early in life in the South East Asian population, older OBI carriers may be more likely to have been infected for decades and the high genetic diversity observed in OBIs may reflect an accumulation of mutations over a longer period of time than in younger HBsAg+ carriers.
A variety of mutations have previously been identified in the HBV S proteins, which affects HBsAg detection assays, immune response recognition, HBV infectivity, cell tropism and virion morphogenesis.12 ,29 Only a few mutations (sC121Y, sC124S, sC137W and sG145A/D/K/R) were found in a minority of OBIB-C strains, irrespective of genotype. Multiple mutations were also present in OBIB and OBIc immunoreactive domains, including several not found in non-OBI controls, but they were only present in one OBI strain (table 4). Similar observations have been reported in OBIA2 and OBID strains.3 Within regulatory regions, OBIB and OBIc strains showed a similar overall genetic polymorphism to OBIA2 and OBID strains, with no strong evidence of a putative negative effect on viral replication, as opposed to what has been reported in OBIA1 strains.2 ,3
A new type of mutant-induced downregulation of HBsAg expression by a post-transcriptional mechanism, not yet fully elucidated, that involves surface RNA splicing has been described.18 Pre-S2/S splicing occurs during replication of full-length HBV genomes, and a G458A mutation has been described that inhibited pre-S2/S splicing, resulting in a marked decrease in the unspliced pre-S2/S transcript and HBsAg.18 ,30 Compared with HBsAg+ strains, OBIB and OBIC strains showed a significantly higher frequency of nucleotide mutations in the vicinity of the pre-S2/S splice donor site. In 44% and 36% of these OBIB and OBIC variants, respectively, these mutations were predicted to disrupt the stem–loop structure of the 5′ donor splice site and consequently to interfere with pre-S2/S mRNA splicing (figure 2). There is increasing evidence that RNA secondary structure at the 5′ splice site can modulate the splicing efficiency by affecting the binding of RNA splicing factors or the recognition of conserved splice site consensus sequences.31 Thus, it was hypothesised that these mutations may alter the interaction of RNA with RNA-binding proteins constituting the spliceosome and modulate post-transcriptional RNA processing or nuclear export via the PRE.18 ,30 Retrospective analysis of OBIA1, OBIA2 and OBID sequences indicated that disruptive mutations in the vicinity of the pre-S2/S 5′ splice donor site were significantly more common in OBI than in HBsAg+ controls (data not shown). In the present study, transient transfection of HuH7 cells with the S-coding region from a limited number of OBI strains and HBsAg+ controls suggested that splicing positions seemed to be mainly conserved across strains irrespective of genotype, but splicing variants may arise as observed in the M90 control strain. The presence of 5′ donor site disruptive mutations in two OBI sequences seemed to be associated with decreased splicing efficiency, but it did not correlate with the level of HBsAg secretion (figure 3). The lack of correlation between the apparent yield of amplified spliced RNAs (figure 3B) and HBsAg production data (figure 3C) may be related to more efficient amplification of the shorter spliced RNA than unspliced RNA. The four OBI strains investigated here appeared to be competent in HBsAg production. However, the amount of HBsAg in culture supernatant was significantly lower in cells transfected with three of the S sequences from OBI strains than in HBsAg+ wild-type strains. Low HBsAg secretion was independent of the level of transcription, as both OBI and control sequences were under the control of the same CMV promoter. These data suggest that alteration of a post-transcriptional mechanism may be involved in the lack of HBsAg detection characterising occult HBV carriage. It is still unclear whether mutations potentially affecting the structure of the S donor splicing site directly negatively influence HBsAg production, and further investigations on full pre-S2/S transcripts from a larger number of OBI strains are needed.
In conclusion, occult HBV genotype B and genotype C strains from South East Asian blood donors showed a higher genetic diversity than strains from HBsAg+ individuals, but no unique genetic signature for occult HBV or evidence of reduced replication competence was found. There is indirect evidence that incomplete immune control is involved in the occurrence of OBI in Asian blood donors infected with genotypes B and C as observed in Europe with genotype A2 but to a lower extent than with genotype D, rather than reduced viral load associated with specific genetic defects as reported in South African OBIA1 carriers. Preliminary data suggest that a post-transcriptional mechanism may play a role in HBsAg expression, and consequently viral replication, in OBI irrespective of HBV genotype.
Acknowledgments
The authors are grateful to the staff of all blood centres participating in the study, who collected, tested and stored the blood samples.
References
Footnotes
-
Funding This study was supported by grants from the International Society of Blood Transfusion Foundation, the National Health Service Blood and Transplant, England, and Novartis.
-
Competing interests None.
-
Ethics approval Ethics approval was supplied by Hong Kong Red Cross Blood Centre Internal Review Board.
-
Provenance and peer review Not commissioned; externally peer reviewed.