The potential for oral polio vaccines to initiate the AIDS pandemic has been investigated previously2,3. Many species of African non-human primates are naturally infected with simian immunodeficiency viruses (SIV) and the common chimpanzee (Pan troglodytes) harbours SIVCPZ, the closest relative to modern strains of the human immunodefiency virus HIV-1 (ref. 4). The use of chimpanzee cells to prepare CHAT vaccine may therefore have resulted in the inadvertent transmission of SIVCPZ to humans.

To test this, we analysed two stocks of OPV CHAT. The first, labelled 19 CHAT 10A-11, was received by NIBSC in 1981 from the Karolinska Institute, Stockholm, and represents an original vial sent from the Wistar Institute, Philadelphia in 1958. The second, CHAT 6039 (an internal NIBSC number) made at the Institute of Immunology, Zagreb, was received by NIBSC in 1987. The presence of CHAT virus was confirmed by molecular assay and both samples yielded viable poliovirus in culture as evidence of quality of long-term storage (J. Martin, personal communication).

We tested extracted nucleic acid for HIV-1 RNA using polymerase chain reaction with reverse transcription (RT-PCR) assays based on gag (Amplicor HIV-1 Monitor assay, version 1.5) and combinations of long terminal repeat (LTR) sequences5, which are capable of detecting HIV-1 group M (subtypes A–H)6, group N and O viruses and SIVCPZ. Additional combinations of LTR sequences used in nested PCR reactions were 507–529 and 524–543 (forward orientation), 622–641 and 628–648 (reverse orientation); the numbering is based on HIV-1 HXB2. HIV-1 primers in pol and env (gp41; ref. 7) and for HIV-2/SIVSM (ref. 5) were also included. Sensitivity was shown to be less than 400 RNA equivalents per ml for each primer set. Using assays that detect this wide range of genetic variants of HIV-1, including SIVCPZ and the earliest known sequence of HIV-1 present in the Belgian Congo in 1959 (ref. 7), we were unable to find any evidence of HIV/SIV in either CHAT stock.

Substrate cells were identified by targeting the D-loop control region of mitochondrial DNA using PCR. Chimpanzee-specific primers failed to amplify a product from either sample. Furthermore, these primers were capable of detecting limiting amounts of chimpanzee template in the presence of 105 macaque-cell equivalents, making it unlikely that chimpanzee cells were used to propagate poliovirus in the two prior passages of vaccine material. However, generic D-loop primers capable of amplifying DNA from baboon, African green monkey (vervet), chimpanzee, Chinese and Indian rhesus or cynomolgus macaques successfully yielded product from both batches. Sequencing and phylogenetic analyses indicate that the cells used to prepare 19 CHAT 10A-11 were obtained from rhesus macaques, and those for CHAT 6039 were from cynomolgus macaques (Fig. 1).

Figure 1: Nucleic-acid sequences were aligned using ClustalW (version 1.5) and evaluated using the Phylip Software package.
figure 1

Evolutionary distances were estimated using DNADIST (default: Kimura 2-parameter method, transition:transversion ratio 2.0) and phylogenetic relationships estimated by the neighbour-joining method using NEIGHBOR. Reproducibility of the branching pattern was assessed with SEQBOOT (boot strap, 100 replicates).

Failure to detect HIV/SIV sequences or chimpanzee cellular components in two OPV CHAT stocks, together with the positive identification of macaque mitochondrial sequences, provides no support for the hypothesis that these materials were responsible for the entry of HIV into humans and the source of AIDS.