"A Dog's Breakfast": Michael Worobey's New Paper Seeks to "Exonerate" Patient Zero
A new paper from the team of the University of Arizona molecular biologist, Michael Worobey, has just been published in Nature, to the usual fanfare of publicity. It seeks to demonstrate that the Canadian air steward, Gaetan Dugas, was not responsible for seeding the AIDS epidemic in the United States.
To those who have spent any time examining the history of AIDS, this is hardly new news. However, it is a plausible peg on which to hang an article in an attempt to render it newsworthy.
Even if Randy Shilts, in his wonderful 1987 panorama of the early AIDS epidemic in America, "And The Band Played On", coined the term "Patient Zero", he himself never proposed that Dugas seeded the epidemic in the US. (However, what he failed to report in his book was that Patient Zero was a misnomer. It should have been "Patient O", with "O" standing for "Out-of-California".) My book The River, published in 1999, went into rather more detail, and made it even clearer that Dugas had not been demonstrated to be the source of the US AIDS epidemic. [See The River, pages 63-69.]
Nature's accompanying news piece is headlined "HIV's Patient Zero Exonerated". A similar theme is adopted in an article by Michaeleen Doucleff, put out by NPR Now, entitled "Researchers Clear 'Patient Zero' from AIDS Origin Story". This portrays Dugas as "charismatic, kind and energetic", and goes on to assert that "Dugas wasn't a villain. Not at all." The article ends with Worobey saying that "there's something nice about going back and correcting the record....[Dugas] has been blamed for things that no-one should be blamed for."
The reality, however, is rather different. After being advised by CDC researchers that he was almost certainly carrying an infectious pathogen which had caused him to contract AIDS, Gaetan Dugas refused to stop attending the baths, and thus may well have been responsible for passing on his HIV infection to others. He was hardly a hero, therefore.
Where the latest study is genuinely interesting is in its analysis of the early history of AIDS in the United States. Worobey's molecular analysis of several early samples of HIV from the US strongly suggests that the first infectees were gay men in New York, followed by others from San Francisco. The problems start when Dr Worobey attempts to date the different stages of the epidemic. This is a man who when I knew him (in 2000 to 2002) frequently stated that attempts to date the AIDS epidemic through molecular analysis represented "a dog's breakfast", meaning "a complete mess". Nowadays, however, Dr Worobey seeks fame and fortune through his painstaking research into that same dog's breakfast.
The abstract of the Nature article claims that pandemic HIV must have entered New York City in about 1970, though the article itself specifies 1971. Worobey also proposes that HIV migrated to San Francisco from New York in about 1976. The article includes many pages of data detailing Worobey's analytic methods, but fails to reveal that they are all based on a "molecular dating hypothesis" which is itself based on two highly questionable assumptions: that HIV evolves predominantly through mutation, and that the pandemic began with a single human getting infected with the causative virus (rather than a group of people getting infected from a common source, for instance a contaminated vaccine).
In fact, over 90% of HIV's evolution takes place though recombination - a process whereby two viruses meet in a cell and exchange large pieces of genetic material. (This has been coined "viral sex", and is the equivalent of a baby being created with genetic elements from both parents.) The concept of molecular dating for HIV, the most recombinogenic virus known, is therefore controversial. It seems to me (and to many others, including several molecular biologists) that all Worobey and his fellows are doing with their molecular dating of the AIDS pandemic is to justify their existing findings through a series of assumptions and checks which are self-confirming, and which can therefore produce only one result.
Worobey seeks to clean up the HIV samples employed in his analysis by various methods, and claims that he ends up with a "recombination-free" dataset, but these adjustments make almost no difference to his dates, suggesting that he has very likely failed to identify all the recombinant samples. (When recombination occurs early in the history of a virus, it is almost impossible to detect.)
Ignoring recombination allows one to come up with a putative date of onset that is pushed further back in time than is warranted. Thus Worobey proposes an arrival of the virus in New York City in 1971, and in the Doucleff article he comments: "The virus got to New York City pretty darn early...It was really under the radar for a decade or so." In his paper, Worobey mentions an early AIDS case from New York City which dates from January 1975, adding that this was the date of onset for the man's Kaposi's sarcoma (KS). However, KS alone is not sufficient for a diagnosis of AIDS, and this man might well have been infected with the KS virus in or before 1975, and then with HIV at a later date.
My own researches reveal that the earliest date for a proven HIV-positive sample from NYC relates to 1977, and the earliest likely retrospective date for a New Yorker contracting AIDS relates to 1979. Four putative AIDS cases from 1978 were apparently cases of KS, and thus not definable as AIDS. As far as I am aware, no earlier samples of HIV or AIDS have been reported from the US. [See The River, pages 55-73. This is part of the early section of the book comprising Chapters 3 to 6, in which I attempt to summarise the early history of HIV/AIDS.]
The earliest evidence of HIV from San Francisco relates to 1977, and the first AIDS cases in that city were reported in 1981.
Interestingly, all the HIV-positive samples from 1977 (these being six from New York and one from San Francisco) were obtained not from gay men, but from HIV-positive children born to drug-injecting mothers. This suggests that all seven children were infected perinatally, and that the mothers were HIV-infected by that year. This in turn suggests that there may have been early crossover in the US between the first two risk groups to be recognised: gay men and intravenous drug users.
Especially in view of the rapid rate at which HIV could have spread through the drug-injecting and gay club scenes of the 1970s, I find it hard to believe that the virus could have been present in NYC for eight years (1971-79) before anyone fell sick with AIDS. And if HIV really did arrive in the US in 1971, it is remarkable that no evidence has yet been found in a blood sample taken before 1977.
My own gut feeling, for what it is worth, is that HIV may have arrived in New York City in around 1973 to 1975. By contrast, an arrival in San Francisco in around 1976 seems reasonable to me.
One final thought. Worobey's analysis suggests the same order of spread that Shilts proposed in 1987, and that I proposed in 1999: that HIV spread from the Democratic Republic of Congo (the former Belgian Congo) to Haiti (which had supplied technocrats to fill the jobs vacated by Belgians when the DRC achieved Independence in 1960), and then later from Haiti to the US. Looking backwards in time, Worobey's graphs and figures date the arrival of HIV in Haiti in around 1967, and seem to place the earliest node on the HIV tree (which he identifies as the divergence between two subtypes of HIV-1 called B and D) in around 1955.
This, of course, is almost exactly the time when the first trials of an oral polio vaccine called CHAT which, unusually, had been locally prepared in chimpanzee cells, took place in the Belgian Congo. (The common chimpanzee is host to a version of simian immunodeficiency virus, SIV, that is the direct ancestor of pandemic HIV-1 in humans.)
However, Worobey has previously proposed that HIV-1 began 500 miles north of the Belgian Congo, when a Cameroonian became infected with chimpanzee SIV in around 1908. (That date of origin is entirely hypothetical, being based on his molecular dating theories, but is nowadays increasingly spoken of as if it were a proven fact.) More recently, a Belgian paper has proposed that HIV-1 remained quiescent for many decades until the late 1950s or early 1960s, when it emerged from Leopoldville (now Kinshasa), the capital of the DRC, and began spreading round the world at an exponential rate.
The earliest proven samples of HIV remain two samples from Leopoldville dating from 1959 and 1960, just after the CHAT trials in the Congo. The next earliest samples date from the 1970s, but come from the same country. Although the 1959 sample was discovered in 1985, no earlier sample of HIV has been discovered since.
Given these uncomfortable details, it seems that what Worobey and the Belgians are trying to do is to come up with an explanation for the emergence of HIV that differs from the obvious one: that early HIV was seeded in Homo sapiens by a version of CHAT vaccine which, uniquely in the history of the world, was prepared in chimpanzee cells.
Ed Hooper. 28th October, 2016