Potential conflict of interest: Nothing to report.
An acute hepatitis B virus (HBV) infection was diagnosed in a regular apheresis (plasma/platelet) donor by the hepatitis B surface antigen (HBsAg) assay and minipool nucleic acid amplification technology (NAT). The acute infection was confirmed by detection of anti-HBc (IgM) and anti-HBs 2 weeks later. The donor showed no clinical symptoms and had normal alanine aminotransferase levels. He had a history of weekly apheresis plasma or platelet donations. Archived material from the donor and the respective recipients was investigated by sensitive HBV NATs as part of a look-back procedure. HBV DNA was detectable in previous donations as well as in two recipients transfused with platelet concentrates. The rare HBV genotype G was identified in all HBV-DNA-positive samples. Strong evidence of genotype G monoinfection was obtained by clonal sequencing, HBV genotype line probe assay, genotype-specific NATs, and restriction pattern analysis. In contrast to previously described genotype G infections, which all appeared as coinfections with genotype A, neither the hepatitis B e antigen (HBeAg) nor anti-HBe was detectable in any of the samples. This shows that HBeAg is dispensable for viral replication. The delay in detecting HBsAg in both the donor and recipient samples may be explained by either decreased genotype G–specific synthesis of incomplete viral forms in early HBV infection or the lower sensitivity to genotype G of the current HBsAg assays. In conclusion, this reported case of an HBV infection was caused exclusively by genotype G. (HEPATOLOGY 2006;44:99–107.)
Hepatitis B virus (HBV) infection is a serious global health problem, with 2 billion people infected worldwide and 350 million suffering from chronic HBV infection.1 HBV is a prototype member of the Hepadnaviridae family and consists of a circular, partially double-stranded DNA genome of approximately 3.2 kb. The genome carries four open-reading frames (ORFs). The presurface/surface (pre-S/S) region encodes three forms of the viral hepatitis B surface antigen (HBsAg). The precore/core (pre-C/C) region encodes the viral e antigen (HBeAg) and the core protein (HBcAg). The P coding region is specific for the viral polymerase, and the X-ORF encodes the X protein.
HBV is currently classified into eight genotypes (A-H) by sequence divergence with the entire genome exceeding 8%.2–7 Only a few infections with the rare HBV genotype G have been previously described. All were characterized by simultaneous coinfection with genotype A. The typical features of the genotype G genome are a 36-nucleotide insertion at the 5′ end of the C gene and translational stop codons in the pre-C region (codons 2 and 28), preventing HBeAg synthesis. Such sequences had already been described in a U.S. patient coinfected with HIV in 1990 and in an isolate (B1-89) from a French patient in 1991.8, 9 However, these isolates were not described as a new genotype. Genotype G was detected in 11 of 82 (13%) HBV carriers from Georgia (in the U.S.) and in 2 of 39 (5%) HBV-infected individuals in France.5 This genotype was also reported in a single patient from Mexico,6 from Germany,10 and, recently, from Japan.11 The Japanese and German patients were also coinfected with HIV type 1. The patient in Japan belonged to a risk group presumed to have acquired HBV and HIV infections while traveling in the United States. All these HBV patients suffered from chronic infection.
HBV transmission by transfusion became rare after the introduction of the HBsAg test into blood screening in the early 1970s. Routine HBV nucleic acid amplification technology (NAT) on minipools was introduced on a voluntary basis at some blood transfusion services in order to further improve blood safety by reducing the length of the diagnostic window.12 However, both screening tests are insufficient for diagnosing all chronic HBV infections because viral load may be below the NAT detection limit and HBsAg may become undetectable some months after the infection.13
This report describes a recent HBV genotype G infection in a regular blood/plasma donor and the transmission of the virus to two platelet recipients. The look-back procedure identified archived positive donor samples so that the early HBV infection period could be analyzed. Furthermore, the respective recipient materials were available, and more recent blood samples could be taken for further investigation. We were able to confirm the chain of infection, to characterize the implicated viral genotype, and to exclude typical infections by mixed HBV genotypes (G and other) using molecular approaches.
HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; NAT, nucleic acid amplification technology; PCR, polymerase chain reaction.
Patients and Methods
As the donor (male, 60 years old) sample from May 28, 2003 (index donation), was positive for HBsAg and HBV DNA, a look-back procedure was initiated as required by the German guideline.14 Archived samples obtained up to 6 months before the last negative donation were identified and repeatedly tested for several HBV markers. Recipients of blood components from the donor in question were traced and tested for HBV markers. The samples available for retrospective investigations are listed in Tables 1 and 2.
Table 1. Donor History: HBV NAT and Serological Testing of Plasma or Platelet Donations and of Plasma Samples Drawn for Follow-up Investigations
The PRISM HBsAg assay (Abbott GmbH & Co. KG, Wiesbaden, Germany) was used for routine HBsAg screening. The cutoff was determined to be 9 pg/mL using the PEI HBsAg standard preparation, subtype ad (1 PEI unit corresponds to 1 ng of immunologically active HBsAg). To detect HBeAg and antibodies to hepatitis core, surface, and e antigens (anti-HBc, anti-HBs, and anti-HBe), the AxSYM HBe 2.0, AxSYM Core, AxSYM Core-M; PRISM HBc, PRISM HBcore; AxSYM AUSAB, and AxSYM Anti-HBe 2.0 tests (all Abbott GmbH & Co. KG, Wiesbaden, Germany) were used. All serological tests were performed according to the manufacturers' instructions. To compare HBsAg results, different commercial assays—Immulite 2000 HBsAg (DPC, Los Angeles, CA), bio-ELISA HBsAg color (Biokit, S.A., Barcelona, Spain), Murex HBsAg version 3 (Abbott/Murex Biotech Ltd., Datford, UK), and IMx HBsAg (Abbott GmbH & Co. KG, Wiesbaden, Germany)—were used.
Screening for HBV DNA was performed at the blood screening laboratory with C-gene-specific polymerase chain reaction (PCR; nucleotide positions [nt] 1906–2086, referencing a unique EcoRI site15) in a minipool format (n = 96 donations) and chemiluminescence analysis. The 95% detection limit of the NAT assay was determined to be about 720 copies/mL, calculated for the single donation entering the minipool. In the two confirmatory laboratories, nested PCR or the TaqMan assay was used for amplification of the S region (nt 192–704 and 339–430) and the pre-C/C region (nt 1861–2315), as previously published.13, 16 HBV DNA was quantified using the COBAS TaqMan HBV Test in combination with the High Pure System Nucleic Acid Test (Roche Diagnostics GmbH, Mannheim, Germany).
For the genotypes A-F– and genotype G–specific PCR approaches, primer pairs were used spanning two regions (nt 978–1650 and 2676–2912). The well-characterized international reference plasma preparation of HBV genotype A, subtype adw2, was used for retrospective NAT analysis (nucleotide sequence data available in GenBank under accession number AJ012207).16
Restriction Enzyme Analysis.
Amplicons from the nested PCR of the S region (nt 192–704) were digested with HinfI for 1 hour at 37°C following 20 minutes at 80°C for enzyme inactivation. Fragments were separated on an agarose gel and visualized by ethidium bromide staining on a UV illuminator.
Sequencing and Phylogenetic Analysis.
HBV S- and C-gene-specific primers were used for sequencing plus and minus strands. Amplification products were sequenced directly (TaqMan assay) and after cloning (nested PCR) into the pCR2.1 vector (Invitrogen BV, Leek, The Netherlands) using a BigDye Terminator Cycle Sequencing Ready Reaction Kit on an ABI PRISM DNA 310 sequencer (Perkin Elmer Applied Biosystems, Darmstadt, Germany). Four clones per fragment were sequenced in both directions. S gene sequences (nt 282–409) of prototype genomes of genotypes A–G deposited in GenBank (identified by accession number) were aligned. Phylogenetic analysis was performed as published.13 Twenty additional clones with a core insert (491 bp) and 10 additional clones with a surface insert (513 bp) derived from the amplified HBV DNA of infected donor sample 05/14 were sequenced using the respective PCR sense primers.
Sequencing of Full-Length HBV Genome.
Full-length HBV DNA derived from the donor sample (May 14) was generated by the sequencing of PCR amplicons from 8 overlapping fragments spanning the whole viral genome and cloned into the pCR2.1 vector. Four clones per fragment were sequenced in both directions. A full-length sequence was compared to published sequences of genotypes A and G, GenBank accession numbers AJ012207 and AF160501, respectively. The new complete sequence is accessible in GenBank under accession number DQ207798.
HBV Genotyping Assay.
The INNO-LiPA HBV Genotyping assay (Innogenetics N.V., Ghent, Belgium) is a line probe assay designed to differentiate among HBV genotypes A–G by hybridizing type-specific sequences in P gene domains B–C.17 The new genotype H can also be identified.18 The assay was performed according to the manufacturer's instructions.
HBV Markers of Samples Involved in the Transmission Episode.
In May 2003 a 60-year-old male apheresis donor tested positive for HBsAg as well as for HBV DNA by minipool PCR in routine screening. Look-back procedures using NAT on archived samples identified six previous donations of his that were HBV DNA positive (three plasmapheresis units, three double-thrombapheresis units; Table 1). All three plasma units were still under inventory hold and could be discarded before further processing. Four recipients had already been transfused with six platelet components derived from three double-pheresis units (Table 2). Two weeks after the initial positive test results, the donor tested positive for immunoglobulin M anti-HBc and anti-HBs. One week later the HBsAg and HBV PCR test results were negative once again. The infection route of the donor was unclear because operations, blood transfusions, travel to high-risk countries, and changing sexual contacts all were denied.
Recipient 1 was given platelets of both bags from a single donation (Table 2). Seven weeks after transfusion he was negative for HBsAg and HBV DNA, but positive for anti-HBc and anti-HBs. These data indicate an old, resolved HBV infection rather than a recent one. Recipient 2 received platelet units from two different donations of this infected donor and tested negative for HBsAg and HBV DNA. He was determined to have an anti-HBs titer of 164 IU/L. Past vaccination against HBV seems likely.
Recipient 3 had been transfused with a platelet concentrate and tested HBV DNA positive 1 month later (52 copies/mL). The viral DNA concentration increased to more than 9 million copies/mL 4.5 months after transfusion. This recipient tested positive for HBsAg and remained negative for anti-HBc and anti-HBs during this period.
Recipient 4 tested positive for HBV DNA (69 copies/mL) 3 weeks after platelet transfusion. During follow-up, HBV DNA concentration increased and anti-HBc antibodies became detectable 2.5 months after transfusion, whereas tests for HBsAg and anti-HBs remained negative. Recipients 3 and 4 had elevated ALT levels. They died of heart failure. None of the recipients showed clinical symptoms of acute hepatitis.
HBV Sequences and Phylogenetic Relatedness Derived from Samples Involved in the Transmission Episode.
Two laboratories performed PCR with primer pairs mainly overlapping for the amplified S and C regions. HBV DNA–containing plasma samples (from the donor and recipients 3 and 4), which tested positive with the S and C regions as PCR targets, were used for sequencing analysis in order to analyze the putative chain of infection. All examined plasma samples showed 100% sequence identity for both HBV regions, providing strong evidence of transmission by the same viral strain. In contrast to the positive run control (genotype A, subype adw 2; GenBank accession number AJ012207), the sequences from donor and recipient samples were characterized by an insertion of 36 bp after the fifth nucleotide in the C gene and by a translational stop at codon position 28 in the pre-C region (TGG to TAG). A phylogenetic comparison with 70 other HBV genomes representing all known genotypes was performed. Homology percentages and phylogenetic distances were calculated for S-gene nucleotide positions 282–409. This analysis demonstrated clustering of the sequences from the donor and the recipients in a separate branch representing the HBV genotype G sequences (Fig. 1).
Full-Length HBV Sequence of Genotype G.
The complete HBV genome derived from the donor was found to consist of 3,245 bp. All characteristic differences with other genotypes were confirmed: translational stop codons at codon positions 2 and 28 of the pre-C region, an insert of 12 amino acids in the amino-terminal part and deletion of two amino acids in the carboxy-terminal part of HBcAg, and deletion of three nucleotides in the pre-S1 region. In addition, a further deletion of three nucleotides (TTT) in the overlapping pre-S2/P ORFs resulting in an amino acid deletion of phenylalanine in both proteins was found when compared to the published full-length sequence of genotype G (AF160501). No frameshift was observed in either ORF because of the presence of four consecutive T nucleotides. The full-length sequence had 87.7% homology with the genotype A sequence (AJ012207). It was nearly identical (99.5% homology) to the first published genotype G reference sequence.5
Molecular Studies to Identify Mixed Genotype Infection.
The results from cloning and sequencing of the amplified HBV DNA, derived from samples of the donor (including additional sequencing of 20 clones with a core insert and 10 clones with a surface insert) and the two recipients, clearly hinted that no sequences of additional HBV genotypes had been identified. Moreover, no sequence showed a recombination of genotype G sequences with sequences from other genotypes. Nevertheless, other methods for direct molecular typing of the HBV genome were performed to challenge these findings.
The INNO-LiPA HBV Genotyping line probe assay allows discrimination among HBV genotypes A–G. Moreover, LiPA has been shown to identify mixed-genotype infections, which may remain undetected by amplicon sequencing.17, 18 As shown in Fig. 2, the samples of the donor and the two recipients (R3 and R4) were positive for genotype G–specific line 16 and for genotype A–specific line 4, which was expected to cross-hybridize with the genotype G sequences. By contrast, a mixture of genotypes A and G would be reactive for genotype G (line 16) and at least two of three lines indicative of genotype A (lines 3–5), as explained in the manufacturer's instructions. Because there were not even any faint bands of other genotypes visible, there was no indication of an additional HBV genotype in the donor or recipients samples.
Genotypes A-F– and G–specific PCR was performed with primer pairs for two regions (nt 978–1650 and 2676–2912). HBV DNA extracted from the donor and recipient samples was amplified exclusively with the genotype G–specific primer pairs, whereas, as the control, the genotype A reference sample was amplified with the genotypes A–F primer pairs (Fig. 3). Thus, the results did not give any indication of coinfection of the donor and recipient samples with HBV genotypes other than G.
PCR amplification of the pre-S or S region followed by analysis of restriction fragment length polymorphism is another means to distinguish HBV genotypes. Figure 4 shows the results of restriction enzyme analysis with HinfI of the 513-bp amplicon derived from the S gene. It shows a restriction pattern for the reference sample of genotype A (fragment sizes of 7, 19, 34, 175, and 278 bp), which is clearly different from that of the samples of the donor and the recipients (fragment sizes of 7, 19, 34, 48, 175, and 230 bp). The occurrence of an additional restriction site for HinfI in the genotype G sequence of this part of the S gene compared to in the respective genotype A sequence resulted in two fragments 48 and 230 bp in size.
Furthermore, the negative test results for HBeAg and anti-HBe in all relevant donor and recipient samples were consistent with a genotype G–only infection. This is explained by the absence of an intact pre-C-encoding sequence in genotype G. By contrast, all published genotype G infections appeared as coinfections with genotype A and were associated with positive HBeAg or anti-HBe results, which again are only explained by the input of genotype A.
This reported case of putative HBV infections was caused exclusively by genotype G. All cases described so far have been coinfections with genotype A.5, 10, 19 Two of four recipients transfused with HBV-contaminated platelets were infected with the same rare genotype G. Neither recipient showed any clinical symptoms of acute hepatitis, but both had elevated ALT levels, the cause of which was difficult to determine as both patients died of heart failure 4–6 months after infection. The other two recipients of the platelet concentrates did not develop an HBV infection, possibly because of their HBV-specific immune status: there was evidence of prior vaccination in one recipient and of resolved HBV infection in the other. The lack of genotype G infection after putative HBV vaccination strengthens the concept of cross-genotype protection through genotype A–derived vaccines.
The identity of the HBV sequences in the donor and the two recipient samples has been demonstrated. These sequences were identical to the first published sequence of genotype G5 and clustered on a branch separate from the other 6 genotypes in phylogenetic analysis of the S-gene amplicons. The full-length sequence derived from the donor sample consisting of 3,245 bp showed all characteristic genotype G–specific features: 2 stop codons in the pre-C region, insertion of 36 nucleotides at the 5′ end of the C gene, and deletion of 1 codon in the pre-S region. The HBV genome of the donor had 87.7% homology with the genotype A sequence (AJ012207) and was nearly identical (99.5% homology) to the first published genotype G sequence (AF160501). We identified a new deletion of an amino acid (F) in the pre-S2/P region with as-yet unclear potential relevance for the biology of the virus.
In all previously described cases, HBeAg was detected in the patient samples, although the genotype G genome contains two stop codons in the pre-C region (codons 2 and 28). This finding, puzzling at first glance, is only explained by coinfection with genotype A, missing the stop codons and competent for HBeAg production.19 Alternative explanations are that the 12 extra amino acids at the N-terminus of the genotype G core protein created a novel signaling sequence directing a fraction of the core protein into the secretory pathway10 or that nonparticulate core protein-sharing epitopes with HBeAg were detected.20 However, in our case HBeAg was consistently negative for all tested donor and recipient samples. Neither was any anti-HBe detectable. Thus, these results strongly indicate the dispensability of HBeAg for viral replication. It was assumed that HBeAg is necessary for the induction of a persistent infection.21 The seroconversion of the donor to anti-HBc and anti-HBs accompanied by the nonreactivity of the HBV DNA shown by NAT most probably reflects recovery from the asymptomatic acute infection. Follow-up investigations of the donor will take place as he gives his consent. For the two recipients follow-up investigations are not possible because of their deaths from heart failure.
The serological data are consistent with the sequence analysis of multiple clones derived from amplified sequences of the donor and recipient samples, which exclusively detected genotype G sequences. Moreover, the results of the line probe INNO-LiPA HBV Genotyping assay did not show any indication of an HBV mixed-genotype infection in any of the samples investigated. Independent investigations demonstrated a high correlation between the LiPA assay and other genotyping methods combined with the potential to detect very sensitive mixed-genotype infections.17 Data from the genotype-specific PCR approaches and from the restriction pattern analysis support the absence of further HBV strains in the donor and recipient samples. In all genotype G cases reported so far, the patients were chronically HBV infected. Here we describe the early period of HBV infection in a donor patient. Although the details of the infection are unknown, HBV genotype G seems to be replication competent without the need for helper viruses. The donor and recipients samples tested negative for antibodies against the hepatitis delta virus (anti-HDV, data not shown). Though very unlikely, HDV coinfection cannot be definitively ruled out because anti-HDV might not yet be detectable in the early infection period.
Transfection studies with cloned HBV genomes are currently ongoing to clarify the relative replication competence of different HBV genotypes. Nevertheless, evolutionary relatedness of the new genotype G to genotype A seems likely, as was suggested by the 6-year follow-up study of a French chronic HBV carrier.9 At the beginning of the study major form B1-83 represented genotype A, whereas 6 years later major form B1-89 was detected as a viral mutant representing a genotype G sequence. This mutant may thus be interpreted as an outcome of selection that might be more capable of escaping the host immune response.22 For the time being, the geographic origin of pure genotype G and its distribution are unclear. Furthermore, nothing is known about a possible epidemiological link between the reported U.S. and European cases. However, no genotype G cases have been found so far in Japan, in a geographic area with a higher HBV prevalence23 except for one case imported from the United States. Interestingly, some HIV patients, preferentially by risk group, are coinfected with HBV of rare genotypes.11
The differences at the nucleic acid level of the S gene of our genotype G with other genotypes reflect only a few unique changes at the amino acid level (A/S45V, Q51L, T63I). The genotype G sequences revealed subtype adw2. No typical HBsAg escape variant is obvious from the sequence analysis. However, because exchanged amino acid positions are N-terminal to the antigen loop, steric alterations cannot be ruled out and may explain reduced sensitivity of the HBsAg tests. Several specimens from the early-infection phase would have been expected to give reactive HBsAg results given the corresponding viral DNA concentrations. This potential lack of sensitivity was also confirmed by five additional HBsAg assays, where again only the plasma sample with the highest DNA (> 9 × 106 copies/mL; recipient 3) tested positive for HBsAg (Table 2). We analyzed parallel titration series of this plasma specimen and of the HBV DNA International Standard for NAT assays (97/746, genotype A; NIBSC, Hertfordshire, UK) with the PRISM HBsAg assay. This study revealed a viral DNA concentration corresponding to the HBsAg cutoff of about 600 copies/mL for the WHO standard and of about 24,000 copies/mL for the high-titer recipient sample (data not shown). Normally, an HBsAg excess ratio of more than 1,000:1 between empty HBsAg particles and virions has been described.24 In our case the ratio was reduced by a factor of more than 10. This considerable difference might reflect either different genotype-specific sensitivities of the HBsAg assays or genotype-specific expression rates of the HBsAg -containing particles. The latter might again result in different ratios of noninfectious viral forms (empty spheres and filamentous HBsAg particles) to viral DNA containing infectious virions. The observed significant decrease in envelope protein secretion of the French B1-89 mutant form was explained by a decrease in the envelope transcripts. The putative down-regulation of these transcripts might be a result of mutations in transcriptional control elements.22 Because of the increased number of amino acid changes in the pre-S region of genotype G compared to in the amino acid sequences of the other genotypes, an influence on surface antigen synthesis and secretion cannot be ruled out. Further investigations using in vitro systems will provide more information.
The observed miscorrelation between detectable HBsAg and HBV DNA levels might lead to a prolonged diagnostic window period and complicate HBV detection during donor screening in transfusion medicine. Therefore, all available donor samples were tested in triplicate with the COBAS AmpliScreen HBV Test (Roche Diagnostics GmbH; data not shown). The earliest detection of HBV DNA was found in the donor sample from April 10, 2003 (two of three replicates were reactive). In this follow-up with frequent specimens (<1 week between donations), an estimated reduction of the diagnostic window of 7 weeks would be the result of sensitive single-donation HBV NAT. However, minipool PCR of the screening laboratory detected the recent HBV infection quite late, reflecting 8,210 copies of HBV DNA/mL in the index donation. The validated 95% detection limit of this screening NAT assay was about 720 copies/mL. Therefore, even a donation with an HBV DNA concentration of 855 copies/mL would have been expected to give a positive NAT result. Suboptimal primer design (with the C region as target) resulted in a 3′ end mismatch as the probable reason for the lower NAT sensitivity of this genotype G sample. By contrast, another in-house minipool TaqMan assay with primers targeted to the S region detected this virus strain in 3 of 4 runs at an HBV DNA concentration of 42 copies/mL (data not shown). Although NAT testing may reduce the length of the diagnostic window significantly, it should be taken into account that apheresis donations usually are available after much shorter intervals than are whole blood donations. In our case platelets or plasma were donated weekly. This high donation frequency increases the potential risk of obtaining several HBV-contaminated window period donations in the absence of highly sensitive NAT systems. The introduction of sensitive NAT systems or larger donation intervals might be considered suitable measures.
In conclusion, this report shows that HBV infection may occur exclusively with genotype G and that this infection may be transmitted by transfusion of blood components. No indication of coinfection with a second HBV genotype could be found using a variety of sensitive molecular approaches. The course of infection demonstrated the replication competence of the virus and provided evidence that HBeAg is dispensable for viral replication. Further studies of the obvious relative lack of detection of genotype G HBsAg by screening assays are ongoing.
We thank Christine Pfannkuch for excellent technical assistance and Ina Plumbaum for critically reading the manuscript.