LETTER

The hepatitis B virus (HBV) infection rate has been effectively reduced by universal vaccination. Concentrations of vaccine-induced neutralizing antibodies (anti-HBs) above 10 mIU/ml are generally regarded as sufficient for protection against HBV infection (4). The detection of the hepatitis B surface antigen (HBsAg) by enzyme immunoassay (EIA) utilizes either polyclonal and/or monoclonal anti-HBs and targets the major "a" determinant of HBsAg, although other targets may also be available for detection. Such assays may be unable to detect HBsAg "a" determinant mutants (2, 5, 6).

To evaluate HBV persistence in Singaporeans who were negative for HBsAg by EIA following hepatitis B (HB) vaccination, we tested for the presence of HBV DNA by PCR using primers specific to the immunogenic major hydrophilic region (covering amino acid residues 110 to 160) of HBsAg. Sixty-three adults and 15 vaccinated children (aged below 15 years) from different ethnic groups, who had tested negative for HBsAg by four independent commercial immunobased diagnostic kits, were selected (Table 1). They were all positive for total antibodies to HBV core antigen (anti-HBc) (Corzyme; Abbott Laboratories, North Chicago, Ill.). Some also displayed levels of anti-HBs (Ausab; Abbott Laboratories) that were much higher than 10 mIU/ml. Results indicated that HBV DNA was amplified in 3 of the 15 children (20%) (Table 1, patients 1 to 3) and 8 of the 63 adults (13%) (Table 1, patients 4 to 11). This detection rate was comparable to that reported previously (1, 8). Direct sequencing of the amplified DNA fragments revealed mutations at various positions of the major hydrophilic region in these 11 patient samples (Table 1). These include the most common vaccine escape mutation, G145R (2), in eight cases. There were also four cases with the G130D mutation associated with lamivudine therapy (7). In addition, the T131N mutation was identified in four cases.

To investigate the underlying mechanism of the identified HBsAg mutants that escape detection by current immunobased diagnostic kits, recombinant HBV genomes carrying individual HBsAg mutations were generated and analyzed in mammalian HepG2 cells after transfection. To this end, a replicative form of the wild-type HBV genome with a redundant 300-bp regulatory region (nucleotides 1630 to 1930) (9) was cloned into the mammalian expression vector pcDNA3.1. The mutations G130D, T131N, M133T, and G145R and a double mutation, G130D G145R, were generated by site-directed mutagenesis using the wild-type HBV as the template. Cell culture supernatant containing secreted viral particles (either the wild type or the HBsAg mutant) was collected 3 days after transfection and assayed for their recognition by the four diagnostic kits described above. These kits were based on either polyclonal or monoclonal anti-HBs. The same amount of HBV DNA for each construct, measured by liquid hybridization assays (Genostics; Abbott Laboratories), was used for the assays. Results of the HBsAg detection assays showed that none of the mutants was detected (data not shown).

Our findings indicate that HBsAg mutants can escape detection by current methods and persist in HBV-infected individuals after the loss of HBsAg and seroconversion to anti-HBs. These mutants are not neutralized by concentrations of anti-HBs that exceed the generally believed protective levels (10 mIU/ml). The detection of HBsAg mutants in HBsAg-negative vaccinated Singapore children suggests a need for closer monitoring since they may spread undetected to the general population through horizontal transmission (3) or through blood supplies and cause liver diseases.