Analysis of false positive HIV-1 serologic testing in Kenya

Reactivity of genotype-specific recombinant proteins of human erythrovirus B19 with plasmas from areas where genotype 1 or 3 is endemic

Human erythrovirus (parvovirus) B19 (B19) is a common human pathogen. The recent discovery of three genotypes, 1 to 3, raised issues related to the ability of genotype-specific antigens to cross-react with antibodies elicited by other genotypes. This study assessed antibody capture and immunoglobulin G (IgG) cross-reactivity between genotype 1 and genotype 3 recombinant antigens and analyzed the potential gain of adding VP1 protein to VP2 capsid antigen. Plasma samples from genotype 1- or genotype 3-infected populations were blindly tested with blindly prepared reagents. The IgG reactivity was assessed with baculovirus-expressed VP2 or VP1 and VP2 recombinant genotype 1 or genotype 3 proteins in a standardized enzyme immunoassay (EIA). A high degree of agreement (>95%) between EIA results was observed, with Spearman correlation coefficient and kappa reliability coefficient results of >/=0.95 for samples from the United Kingdom and >/=0.77 for samples from Ghanaian children, respectively. Most discrepant results were related to equivocal reactivity. The addition of VP1 to VP2 capsids did not significantly impact antibody detection. These data suggest that the currently available genotype-1-based IgG EIA is suitable to detect antibody to B19 genotype 3 in Ghanaian children. However, samples from the Ghanaian adult population exhibited atypical results in the assay, possibly due to the high levels of nonspecific IgG antibodies found in adults living in this region. Within these limitations, the study demonstrates that genotype 1 and genotype 3 antigens are equally effective in detecting either antibody species.

Identification of regions of HIV-1 p24 reactive with sera which give "indeterminate" results in electrophoretic immunoblots with the help of long synthetic peptides

We analyzed nine sera from persons unlikely to be HIV infected which had an IgG reactivity directed against HIV-1 p24, and in two cases also to its precursor p55, but to no other HIV proteins, nor to proteins of the H9 host cell, in electrophoretic immunoblots (EIB). These sera are also referred to as having an indeterminate HIV EIB pattern or as HIV antibody false positive sera. Seven of nine sera reacted with longer (61-77 amino acids) and none with shorter (17-25 amino acids) p24-derived peptides in enzyme immunoassays (EIAs). This is compatible with a conformational (discontinuous) nature of the epitopes involved in many false positive HIV-1 p24 antibody reactions. Four sera reacted with an N-terminal, one with an internal, and two with a C-terminal fragment. Each of the seven sera thus only reacted with one of the long p24 peptides. The specificity and singularity of the reaction was further demonstrated by competition and/or absorption experiments with synthetic peptides. In contrast, 18 of 20 confirmed HIV-1+ sera with p24 reactivity in EIB reacted with at least one and often several of the longer peptides, most frequently the C-terminal one. Thus, the distribution of peptide reactivity of true HIV-1 antibody-positive sera was different from that of the falsely reactive sera. According to two of several explanations, these antibodies may have arisen because of (1) molecular mimicry by chance or by functional selection, (2) immunization by activation, noninfectious exposure, or infection involving non-HIV endogenous or exogenous retroviral antigens. The latter gains some support from our finding of antibody reactions with capsid proteins of the simian viruses, simian sarcoma-associated virus (SSAV), and Mason-Pfizer monkey retrovirus in some of the p24 +/- p55 reactive sera.