River of tears

In June 1981, two unusual events occurred in very different parts of the world. In Los Angeles, five gay men fell sick with rare symptoms suggestive of immunological problems, which prompted two local doctors to write a paper for the Morbidity and Mortality Weekly Report. Meanwhile, in Mugana, northern Tanzania, a German missionary doctor saw five women from the Ugandan border region, all suffering from untreatable anaerobic ulcers of the groin and anus. Some new pathogen was abroad, and these doctors were among the first to recognise that fact.

Before long, the beast was named. By 1983, the condition of immune collapse had swept through gay communities in the western world, and was given the acronym Aids. Meanwhile, in Africa, the condition spread almost as rapidly among men, women and children. Tanzanian villagers christened the newly-arrived disease "Juliana" and Ugandans named it "Slim". By 1985, it was realised that the US and African outbreaks were caused by the same pathogen. In both cases, it seems, this pathogen, HIV, had been planted in fertile seed-beds (the gay bath-houses of Los Angeles; the smuggling zone around Lake Victoria) and had prospered. But this begged a further question: where had the new virus sprung from?

The classic view about emerging viruses is that there is an unprecedented interaction between humans and their environment (traveller visits bat cave, logging company opens up virgin rain forest), which triggers a zoonosis - a human disease caused by crossover of an animal pathogen. This scenario seems to work well for killer diseases like Ebola, Marburg and Lassa Fever, and has considerably boosted funds to American virology labs. But does it apply equally well to HIV?

The disciples of "natural transfer", such as the microbiologist Beatrice Hahn, from the University of Alabama, and geneticist Paul Sharp, from the University of Nottingham, believe it does. They propose that a hunter or bushmeat- seller who had skinned or butchered a chimpanzee somewhere in west-central Africa became infected with a simian immunodeficiency virus (SIV), and that this gave birth to the HIV variant (HIV-1 Group M) which has now infected more than 50 million persons worldwide, causing 16 million deaths from Aids.

Two other outbreaks of HIV-1 infection (Groups O and N) have been recognised in the 1990s, and these, they say, arose in the same way. When asked why there have apparently been no Aids outbreaks before the late 20th century, they reply that perhaps there have been, but that it took road-building, urbanisation, and new sexual freedoms following independence, to allow the viruses to break free from sequestered African environments and begin their awful spread.

This theory of west-central African origin was presented in February 1999, with Hahn adding that a reduction in the chimpanzee bushmeat trade would lessen the chances of future Aids outbreaks. She further claimed (misleadingly) that the local subspecies of chimp was in danger of extinction, which might prevent the development of an effective vaccine. It was a sexy package and it grabbed headlines.

However, there were some problems with the theory. First, in the slave trade up to the 1860s, more than 10m Africans from the western seaboard were transported to the Americas. Their viruses and retroviruses came too, but apparently no SIVs or HIVs. Furthermore, there were huge internal population movements in Africa - caused by colonial labour recruitment for plantations, mines and the construction of roads and railways - in the early 20th century, while millions of Africans fought in the two world wars. Yet there was not one visible outbreak of Aids.

Second, although the two minor HIV-1 variants, Group O and Group N, have a clear epidemiological hearth in west-central Africa, the pandemic variant, Group M, does not. There is no evidence of Group M being present there before 1981.

There is, meanwhile, another HIV, HIV-2, which appears to have descended from the SIV of the sooty mangabey, and which has a hearth in West Africa, especially the tiny state of Guinea-Bissau.

The American virologist, Preston Marx, carried out extensive research on sooty mangabey SIV and HIV-2 infections in West Africa and he reached a startling conclusion: that sooty mangabey SIV acquired by hunters or bushmeat sellers resulted in non-pathogenic infections that were hardly, if at all, transmissible; "dead-end infections", in short. Natural transfer, he decided, was not by itself enough to cause Aids.

Marx felt that an extra factor was needed to turn such viruses into epidemic forms, and he proposed needles - either those reused in poorly- equipped "country clinics" or those inadequately sterilised during mass vaccination programmes. What Marx was proposing was a natural transfer from sooty mangabeys, followed by an iatrogenic (physician-caused) amplification of the virus in humans.

So, could the needles hypothesis apply in west-central Africa also, and explain the outbreaks of HIV-1-related Aids?

The natural transfer school thought it could. In February this year, one of Hahn's colleagues, Bette Korber, addressed a press conference about her research on the dating of the pandemic. She had used a "super- computer" to compare many different Group M viral sequences, and proposed that the date of the "last common ancestor" was 1930, plus or minus 20 years. Most reporters interpreted this as a chimp hunter getting infected in west-central Africa in 1930, and Korber did not disavow them.

When other scientists looked for mass vaccination campaigns in the area, they came up with inoculations against smallpox, yaws, sleeping sickness and yellow fever, starting as long ago as the late 19th century and continuing right through to the present day.

I believe, however, that this actually weighs against the natural transfer theory, for despite claiming mechanisms for crossover and amplification, Hahn and Korber still cannot provide evidence of Group M infection in west-central Africa before 1981. However, there is considerable evidence further east, in the Democratic Republic of Congo (DRC), Rwanda and Burundi, where M-infected blood is seen from the late 1950s onwards, and Group M-related Aids from the early 1970s.

Hahn and Sharp argue that a chimp hunter from west-central Africa could have travelled here soon after getting infected, and sparked the Group M pandemic. The problem is there are approximately 12 subtypes of Group M, which appear to have emerged in the 1950s. Just how did the itinerant hunter start all these?

The answer, I believe, is that he did not. I propose a different origin for HIV-1 Group M, pertaining to Chat, an experimental oral polio vaccine (OPV) which, I believe, was prepared in the kidney cells of chimpanzees, nearly 400 of which were sacrificed between 1956 and 1958 at Lindi camp, a vaccine research facility in the then Belgian Congo (now the DRC).

The primate cells used to produce Chat were never documented at the time and have never since been revealed. But between 1957 and 1960, Chat vaccine was fed to over a million Africans in the Congo, Rwanda and Burundi - and 64% of all the earliest (retrospectively recognised) Aids cases in Africa, through 1980, come from the same places where Chat was fed. So do 82% of the earliest HIV-1 Group M infections, again through 1980. I have documented 28 Chat campaigns in these former Belgian colonies, and the introduction of different chimp SIV variants in different vaccination venues may provide the explanation for the very different Group M subtypes.

It is only 200 miles from the Ruzizi valley, where the biggest mass trial of Chat was staged in 1958, to northern Tanzania, where the Tanzanian army camped for three months in 1978 before driving Idi Amin from Uganda, and where the first cluster of African Aids cases was seen in 1980. And the only other place where experimental variants of Chat were tested was on babies born at Clinton, a women's prison in New Jersey, which lies just 50 miles from where the first plausible HIV-1 infection in the US was identified.

My own position has changed somewhat since my book, The River, was published. I now believe that both methods of transfer (natural and iatrogenic, cut hunter and contaminated vaccine) may work. Scientists are now finding "sputtering forms" of HIV-1 in the Cameroon, variants which are not related to any of the known subtypes and which look suspiciously like those dead-end HIV-2 infections. Increasingly it seems that natural transfers of SIVs from sooty mangabeys and chimps may occasionally take in humans, but that they neither spread easily, nor readily cause Aids.

But for an Aids epidemic or pandemic to occur, I believe what is needed is a mass exposure, as through a vaccination campaign. Perhaps SIVs stand a better chance of infecting humans when prepared in cell culture, and are then introduced by an "unnatural" route. Or perhaps it is simply a question of numbers - and only 1 in 10,000 (or 100,000) human exposures to SIV results in a productive infection.

My bet is that even those smaller outbreaks of HIV-1 Group O and HIV-2 may have been sparked by experimental vaccine trials. In the late 50s', French polio vaccines, prepared in cells from African primates, were tested in the French colonies of west Africa and west-central Africa. The preferred primates were Guinea baboons which, records reveal, were frequently gang-caged with other monkeys, including sooty mangabeys.

However, scores of chimps were also present in the manufacturing labs for vaccine safety testing, and it may be their kidneys (undamaged by this testing) were later incorporated into the French vaccines - just as I believe happened with the Lindi chimps and Chat vaccine.