Four years ago I wrote The River, a book in which I argued for a new theory of how the Aids pandemic began.The book proved very controversial, and provoked what I would consider a defensive response from many in the scientific community, who damned the theory on insubstantial grounds. I am returning to this subject now because there is new evidence, both historical and scientific, to demonstrate that the theory was buried prematurely.
After 27 million deaths and the infection of more than 66 million people with HIV, there are now strong indications that human hands – in particular, those of the doctor and the scientist – started the Aids pandemic. This is not the theory of origin favoured by most in the medical establishment: the familiar ‘cut hunter’ or natural transfer theory proposes that a single hunter or bushmeat seller became infected with simian immunodeficiency virus (SIV) while skinning or butchering a chimp, and that the pandemic started from that one infection.
The theory of origin that I supported in The River is the OPV (oral polio vaccine) theory, and it requires a little background. In the 1950s, OPVs were prepared in primate cells, as most still are today. As a result, each OPV contained not only weakened poliovirus, but also whichever monkey viruses happened to be present in the cell substrate. One such virus was SV-40 (the 40th simian virus to be discovered), which was found in 1960 to cause tumours in hamsters. By then, tens of millions of people around the world had been given SV-40-contaminated polio vaccines, and over the next two years the producers switched from using Asian monkeys, which are susceptible to infection with SV-40, to African monkeys, which are not. Forty years on, it is recognised that exposure to SV-40 leads to a slightly heightened risk of contracting certain cancers such as mesothelioma.
But the OPV theory relates to a different polio vaccine. It proposes that an experimental polio vaccine called CHAT, developed at the Wistar Institute in Philadelphia, initiated the Aids pandemic by introducing simian immunodeficiency virus (SIV) from the common chimpanzee into some of the million Africans who were given the vaccine between 1957 and 1960. Chimpanzee SIV is now widely recognised as the direct ancestor of the strain of HIV (HIV-1 Group M) that has caused approximately 99 per cent of infections to date. In Africa, CHAT vaccine was administered only in Belgian-ruled territories: the Belgian Congo (now the Democratic Republic of Congo) and the former UN trusteeship of Ruanda-Urundi (now Rwanda and Burundi). These are also the countries that represent the epicentre of Group M-related Aids. The Laboratoire Médical de Stanleyville (LMS), which tested CHAT vaccine for safety and co-ordinated the early African vaccinations, was situated just a few miles from a chimpanzee colony, Lindi camp, which operated between 1956 and 1960. During those years, more than five hundred chimps and pygmy chimps (bonobos) were sacrificed there, mostly in the course of the polio research.
It has sometimes been claimed that people do not become HIV-infected by the oral route. This is demonstrably wrong: gay men have become infected after having only oral sex, and there are several recorded instances in which babies of HIV-negative mothers have been infected through being breast-fed by HIV-positive wet nurses. All mucosal cells, including those of the mouth and throat, represent potential portals of entry for HIVs and SIVs, and the likelihood of infection increases when there are oral lesions, such as those caused by teething or mouth ulcers. The dendritic cells around the tonsils are also significant target cells for HIV, and much of the CHAT administered in Africa, in contrast to OPVs given elsewhere, was delivered by a squirt with a syringe to the back of the throat.
By September 2000, in the run-up to a Royal Society conference on the origin of HIV and Aids, half a dozen samples of CHAT vaccine selected from the freezers at the Wistar Institute had been independently tested in three separate labs. No trace of HIV, SIV or chimpanzee DNA was found. Rather surprisingly, given that none of the samples selected had been prepared specifically for the African trials, the results were widely taken by medical journals, and by the press at large, as concrete proof against the OPV theory. In April 2001, brief reports by the labs that had tested the CHAT samples were published in Nature and Science, and both journals made a big splash of the findings in editorials and commentaries. Robin Weiss’s piece in Nature was headed ‘Polio vaccines exonerated’, as if the safety of all such vaccines were in question, and ended with the claim that ‘some beautiful facts have destroyed an ugly theory.’
Weiss wrongly stated that the samples had ‘included the OPV batch (designated CHAT 13) used in Leopoldville’, confusing the terms ‘pool’ and ‘batch’. To explain the importance of his error requires a brief digression into how polio vaccines are made. Passaging (growing) a virus such as poliovirus in different substrates or tissue cultures (sheets of animal cells grown in vitro, in the laboratory) alters the pathogenicity of the virus. What polio vaccine-makers are looking for is a weakened, or attenuated version that will produce protective antibodies in humans, but not the disease itself. A ‘pool’ of polio vaccine describes material produced at a specific level of attenuation – ‘Pool 9’, for instance, might indicate a poliovirus that has been attenuated by being passaged 28 times through chick embryo tissue culture, and eight times through monkey kidney tissue culture. (In most American and European virology labs in the 1950s, this meant kidney cells from macaques, a species of monkey found in India and the Philippines; it was the accepted final substrate for most polio vaccines.) By contrast, a specific ‘batch’ of Pool 9 would describe a quantity of vaccine that had been prepared from that pool on a specific date in a specific lab: in other words, in a single production run. In practice, batches are prepared by taking a small quantity of the vaccine pool (or a batch made from that pool), and amplifying it by a final passage in monkey kidney tissue culture. The different numbered pools of CHAT vaccine had all been made at the Wistar Institute, or later at pharmaceutical houses, mainly in the USA and Belgium. The crucial factor, however, is where and how the individual batches were made. Some of these batches, I have now found, were prepared in labs far from the US and Belgium, and in substrates derived from different species of primate.
Just a few days before the articles appeared in Nature and Science, some colleagues and I had made a remarkable discovery in Kisangani (formerly Stanleyville). We tracked down and talked to three former lab technicians who had worked at LMS in the late 1950s. Just five minutes into his interview, one of the virology technicians, Jacques Kanyama, revealed that batches of CHAT vaccine had been prepared locally in Stanleyville. This was the detail I had missed in The River. Until now, all the Belgian and American witnesses who had worked at LMS, or who had been involved with CHAT, had insisted that the vaccine had not, indeed could not, have been made locally. They didn’t have the equipment, they said; they couldn’t possibly have produced a vaccine at a primitive lab like that. But Kanyama explained that he had started working in Paul Osterrieth’s virology department at LMS on 12 February 1958, and that Osterrieth had already been making polio vaccine before his arrival. He described how Osterrieth would bring the vaccine from his sterile room, after which the assistants would divide it into phials. Each time a new order came in, Osterrieth made fresh polio vaccine. Sometimes the technicians also helped with immunisations, and Kanyama recalled one particular episode when they took the vaccine across the river to Lukusa military camp to vaccinate the soldiers and their families, giving each person a few drops by mouth. This not only established a chain of custody, but also identified the vaccine as CHAT, the only vaccine then given by mouth in the Congo.
It was now essential to discover which primate tissue cultures the Stanleyville doctors had used to prepare the vaccine. Philippe Elebe worked from April 1956 as a technician in the microbiology department (where Osterrieth had worked until the virology section opened in 1957). He told us that they had indeed been producing tissue culture, and that he had been in charge of culture media – the balanced salt solutions, sera and antibiotics that are used to keep cultures alive and biologically ‘sterile’ (free from known pathogens). We asked which primate had been used to make the tissue cultures, and Elebe’s reply was prompt: ‘Surtout les chimpanzés.’
After returning from Africa, I did further research. It transpired that there were no rules in the 1950s about which primate cells to use for growing polio vaccines: any species could be used provided it made good cultures. Chimp kidneys made ‘very good’ cultures, with ‘very good’ sensitivity to poliovirus. And it was routine practice for oral polio vaccine to be amplified locally, in labs around the world. Some virologists, including Albert Sabin, whose sugar lump OPV was adopted for use the world over, acknowledged this fact in their publications. Others, including makers of CHAT vaccine, did not. However, the papers written about vaccinations in Poland in 1959-60 by collaborators of the Wistar Institute reveal that the titre (concentration) of the CHAT vaccine had risen tenfold between the time of its arrival at the virology department of the State Institute of Hygiene in Warsaw and the moment, a few weeks later, when the vaccine was diluted and distributed to smaller labs around the country. Because vaccine titre decreases with time and temperature change, the only possible explanation was that the titre had been boosted by further passage in primate cells in Warsaw prior to dilution. The primates used by the Polish scientists were the standard Asian macaques, which are not naturally infected with SIV. This indicates why, of the seven million children given CHAT in Poland, and the million elsewhere in Europe, none became HIV-infected as a result.
For trials in Africa (then at least two days from the US by plane), the standard practice was to fly a small bottle of the vaccine, packed in an ice-box, to the destination lab, where it was amplified in whichever primate cells were locally available. This meant that less vaccine had to be transported, and that vaccine quantity and titre could be boosted on arrival. The higher the titre, the more it could be diluted and the more people could be vaccinated. By the second half of the 1950s, virologists in South Africa were using African green monkey cells to amplify the Sabin vaccines, while their colleagues in French West Africa and French Equatorial Africa were apparently using cells from baboons (and perhaps other species, too) to amplify the Pasteur Institute vaccines of Pierre Lepine. In Stanleyville, they had the Lindi chimpanzees.
This new information does not prove that CHAT vaccine started Aids, but it does significantly strengthen the OPV hypothesis. Over the last three years, various ‘disproofs’ have been put forward, but none stands up to scrutiny. One claim involved an attempt to calculate whether a potential contaminating virus such as chimp SIV could have survived the vaccine-making process, and concluded that the chances were trillions to one against. This is wrong. The tissue cultures used in Stanleyville, at least until mid-1958, were clearly of a primitive type that would have provided ideal substrates not just for attenuated polioviruses, but also for SIVs. At least some of these cultures also employed chimpanzee sera to nourish the cells, which means there was substantial potential for recombination (the exchange of fragments of DNA) between different SIV strains in vitro.
It has also been claimed, for instance by a group of researchers led by Beatrice Hahn, Bette Korber and Paul Sharp, that Aids could not have started in the 1950s, because their calculations reveal that the first HIV infection (the so-called ‘Eve virus’) must have existed in the 1940s or 1930s. However, in Science last July, Jon Cohen, the magazine’s leading Aids correspondent, wrote about a ‘beautiful study’ of rampant recombination in the spleen of an HIV-infected patient, observing that the research raised ‘serious questions about phylogeny trees that attempt to date the origin of HIV, all of which discard suspected recombinants to make the data interpretable’. Cohen was only stating what other scientists had been murmuring for some time: that the phylogenetic model used by Hahn, Sharp and Korber to date the epidemic appears inherently flawed, in that it allows for evolution through mutation, but not through recombination, which is now revealed as a major factor in the development of immunodeficiency viruses. Another geneticist, Mikkel Schierup, has pointed out that all it would have required to generate the diversity of HIV-1 variants seen in the world today would be two different chimp SIVs which somehow recombined in humans. Such an event could have happened in vivo (for instance in human vaccinees) or in vitro (for instance in polio vaccines cultured in chimp cells and later fed to humans).
The Hahn group has also insisted that the closest ancestor of HIV-1 is the SIV found in chimpanzees from west central Africa, and that the chimp SIVs from the Democratic Republic of Congo and central Africa are more distantly related. I believe this is a dangerous claim, and not only because recent mitochondrial DNA analysis suggests that the chimps from these two regions should be redesignated as a single subspecies. Even now, there are only four available SIV sequences from the west central African chimps, and just two from the central African chimps. The latter are slightly less closely related to HIV-1, but crucially they differ massively from each other, which emphasises the urgent need for extensive SIV testing right across the chimpanzee range. The five hundred common chimps and eighty bonobos that were housed at Lindi were collected from a vast area – 100,000 square miles of the Congolese rain forest – and we know that at least one chimp (which survived at Lindi for more than two years) came from the Mbandaka area, so may well have been one of Hahn’s west central African chimps. Because cages and play-cages were shared, just one such SIV-infected chimp might have caused widespread SIV infection throughout the colony. So even if Hahn is right, it doesn’t disprove the OPV theory.
Perhaps the most important area of this debate, however, relates to the early epidemiology of Aids. We know that CHAT vaccine was administered in at least 27 different places, all in the Democratic Republic of Congo, Rwanda and Burundi. I have found that 68 per cent of all the earliest Aids cases in Africa (and therefore, with minor exceptions, the world), and 76 per cent of all the earliest HIV infections in the continent, come from the very same towns and villages where CHAT was administered between 1957 and 1960.Recently, a software programme has been devised to analyse five competing theories for the emergence of Aids in Africa – and found that only the OPV scenario achieved a good fit.
The most interesting argument against OPV has been put forward by two American scientists, Preston Marx and Ernest Drucker. They take the cut hunter/natural transfer scenario (which, theoretically, might have occurred at any point during the last few million years, since chimps and humans became separate species), and provide it with a plausible timeframe. They do this by proposing an amplification factor: the arrival in Africa of disposable needles, which were nonetheless reused. Needle deliveries to Africa experienced an exponential rise in the 1950s. This is a theory not of iatrogenic origin, but of iatrogenic spread. It is the only version of the ‘cut hunter’ theory that seems at all credible; otherwise, supporters of the theory have to rely on urbanisation as the factor that allowed a primate virus that had recently passed to humans suddenly to explode. In fact, this explosion would have to have happened not just once, but four times almost simultaneously, since in addition to the pandemic strain, three other HIVs have emerged and become established in Africa in the last fifty to seventy years, infecting between a few hundred thousand and a few dozen persons in each case. These minor outbreaks fit rather well with the OPV theory, for it is known that experimental French-made polio vaccines were tested in the 1950s in both French Equatorial Africa and French West Africa – the epicentres of the three minor HIV outbreaks. Among the primates the French were using for their research were chimps and sooty mangabeys, the SIVs of which are the closest relatives to the HIV strains associated with the outbreaks.
The Marx/Drucker theory has been embraced by the scientific establishment, even though the medical profession plays a key role in it. This may be because the theory is non-specific, so that no individual physicians or institutions can be held responsible. And there is an unspoken subtext: that the physicians in question were probably not Western ones. However, the reuse of unsterilised needles did not happen only in the Belgian Congo, so this theory doesn’t explain the strong correlation between early HIV/Aids and the 27 known CHAT vaccination sites.
In recent weeks there has been a new spin on the needles debate, provided by David Gisselquist. He heads a group of American academics who conducted a survey of various epidemiological studies of HIV, which they combined with the proposal that levels of sexual activity in Africa can be roughly equated with those in America and Europe. They came up with a remarkable claim, which elicited much press coverage, that in marked contrast to previous analyses, 60 per cent of all HIV spread in Africa was caused by unsterilised needles and unscreened transfusions, and only some 30 per cent by sex. (The usual ball-park figures are 10 per cent for parenteral – or blood-borne – transmission and 80 per cent for sexual.) ‘For the last 15 years,’ Gisselquist commented to Reuters, ‘the Aids establishment somehow got on to the notion that we need to scare people about sex to prevent HIV transmission.’
The Gisselquist group provides a long list of surveys they have examined, and impressive pages of mathematical formulae, but their research smacks of a lack of experience of the African epidemic on the ground. They highlight early HIV surveys conducted between 1984 and 1988, but the experience of most people who worked on Aids in Africa during that period is very different. In Uganda, which was probably the first country in the world to experience a visible community-wide Aids epidemic, almost every survey from 1985 onwards reveals intermediate levels of HIV-1 infection for 0-to-4-year-olds (presumably largely caused by perinatal spread), which plummet to virtually zero for 5-to-14-year-olds, then rise steeply for the ages of 15 to 45 for women, and 20 to 55 for men, before tailing off to zero for older individuals. Yet all age groups and both sexes would have experienced comparable exposure to unsterile injections and unsafe transfusions. The ‘Christmas tree’ pattern, which was recognised right across the continent in the 1980s, is strongly suggestive of a pathogen that is spread largely by the sexual route, with only minor roles played by perinatal and parenteral spread.
Of course, the more publicity that is given to claims that dirty needles are responsible for most of the HIV spread in Africa, the easier it becomes to promote the idea that dirty needles might also have initiated the epidemic. And there is clearly some linkage between the two theories, since one of Gisselquist’s co-authors is Ernest Drucker. My own belief (based on many years’ study of African epidemiological surveys) is that parenteral spread may have increased in importance since the start of the epidemic, and that nowadays it might cause between 10 and 20 per cent of new infections. Most of those who have studied the African epidemic at close range believe that Gisselquist’s estimates are speculative, and that some of his public utterances are simply irresponsible, given how many people are eager to hear, dying to hear, that unprotected sex is not so dangerous after all.
The arguments, denials and protestations will continue for some time, but I believe that over the next few years it will gradually come to be realised that the Aids pandemic was sparked by large-scale field trials of an experimental polio vaccine – trials that employed African ‘volunteers’ as guinea pigs.
An extended version of this article, entitled ‘Dephlogistication, Imperial Display, Apes, Angels and the Return of Mr Emile Zola’, is available online at www.uow.edu.au/arts/sts/bmartin/dissent/documents/AIDS/.
Send Letters To:
London Review of Books,
28 Little Russell Street
London, WC1A 2HN
Please include name, address, and a telephone number.