Insight Editions · No. 252
Virus Hunt
The Search for the Origin of HIV
Dorothy H. Crawford
How thirty years of detective work traced AIDS to a chimpanzee virus that crossed into a person in Cameroon around 1900, and why that crossing became a pandemic.
A study guide to Virus Hunt by Dorothy H. Crawford. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
The Book on One Page
Dorothy Crawford tells the detective story, not the tragedy: how international teams spent thirty years working out that AIDS began with a chimpanzee virus crossing into a person in southeast Cameroon around 1900, and why that one crossing, out of at least twelve, became a pandemic.
The book runs in two directions at once. Forward, it follows the virus from a forest in Cameroon down the Sangha and Congo rivers to Leopoldville, out to Haiti about 1966, to the US about 1969, and into the Los Angeles clinics of 1981. Backward, it follows the scientists: from the isolation of HIV in 1983 to the sooty mangabey origin of HIV-2, the P. T. Troglodytes origin of HIV-1, and the dating of the founding infection with two fossil viruses recovered from a 1959 blood sample and a 1960 pathology block.
Crawford's central claim is that the pandemic was made by circumstances, not by a superior virus. Cross-species transmission turns out to be common: HIV-2 crossed from mangabeys at least eight times, HIV-1 from chimpanzees and gorillas four times, and the twelve known crossings are probably the survivors of many more that infected one person and died out. Groups N and P between them have infected fewer than twenty people. What distinguished group M was that it reached a colonial boom town at the moment its demography was most favourable: migrant men outnumbering women more than 4 to 1, widespread untreated genital ulcer disease, a lower rate of male circumcision than today, and mass injection campaigns sharing unsterile needles.
The method is as much the subject as the answer. Evolutionary trees and the molecular clock, calibrated on real outbreaks and corrected for hypervariable genes, recombination and within-host diversity, turned fragments of RNA into dates and places. Along the way the field had to discard its own favourite stories: the Manchester sailor, destroyed by PCR contamination; the oral polio vaccine theory, refuted by vaccine assays, chimpanzee geography and the revised dating; and, after Crawford wrote, the "patient zero" myth, cleared by Worobey in 2016.
- The proof that HIV causes AIDS rests on four natural experiments, not on Koch's postulates: mother-to-child transmission, the haemophiliac outbreak, the Florida dentist cluster, and three laboratory accidents with a pure cloned virus.
- The reservoirs are precise. HIV-2 from sooty mangabeys in West Africa; HIV-1 groups M and N from chimpanzee communities in southeast and south central Cameroon; groups O and P from western lowland gorillas, which caught the virus from chimpanzees 100-200 years ago.
- SIVcpz is itself a hybrid, assembled in one chimpanzee from SIVgsn and SIVrcm, two viruses of monkeys that chimpanzees hunt and eat.
- SIVcpz is not harmless to chimpanzees. At Gombe, infected animals died at 10-16 times the rate of uninfected ones, and the prevalence that dooms a community is only 3.4%.
- The molecular difference that may matter is tetherin: human tetherin lacks four amino acids that Nef needs, and only group M's Vpu evolved to block it.
| Ch | Title | One-line payload |
|---|---|---|
| Intro | A New Disease | 1981-87: panic, scapegoating, the virus found in under three years |
| 1 | The Puzzle of HIV-1 | HIV causes AIDS; the toolkit; simian AIDS solved as a rehearsal |
| 2 | Tracing HIV to its Roots | Africa's epidemic found; HIV-2 solved in eight jumps; ZR59 survives |
| 3 | The Primate Connection | Virus tree matched to chimpanzee subspecies: P. T. Troglodytes |
| 4 | From Rainforest to Research Laboratory | Genomes from faeces pin group M to southeast Cameroon |
| 5 | Timing the Jump | Relaxed clock plus DRC60 gives 1884-1924; Haiti before the US |
| 6 | The First Steps into Humans | Cut-hunter theory stands; the polio vaccine theory falls |
| 7 | The Epic Journey Begins | Rivers, syringes and Leopoldville's demography make a pandemic |
| 8 | Adapting to Humans | Restriction factors versus accessory genes; the tetherin switch |
| 9 | The Challenge of Pandemics | HIV against plague, flu and syphilis; why a vaccine is so hard |
Crawford is careful about what remains unknown. Nobody will ever see the founding event; the needle theory can no longer be proved or disproved; the origin of group O was unresolved at the time of writing, and she leaves three competing phylogenies on the page rather than choosing. Her conclusion is that the pandemic's cause is historical as much as biological: a virus that had been crossing into people for centuries finally met a society that could carry it.
A New Disease
AIDS arrived in 1981 without warning, and within three years its cause, HIV, had been isolated. This book is not about the pandemic's suffering but about the thirty-year scientific hunt that worked out when, where, how and why HIV first infected a human, roughly a century before anyone noticed. Crawford sets the scene with the panic, scapegoating and rivalry of the early 1980s.
Why a book about origins
Dorothy Crawford, a virologist who watched the story "from the sidelines", calls it a detective story and a tale of scientific endeavour. The question of where HIV came from was hotly debated for thirty years after the virus was found in 1983, and cracking it needed laboratory virologists, epidemiologists, animal ecologists and evolutionary biologists working together. Progress often stalled until a new technology arrived, and the hunt turned up many unknown viruses on the way, including twelve "new" HIVs now spreading among us.
The journey runs from the United States, where AIDS was first seen, to Africa and its origins; back in time from the Pasteur Institute to the rainforests of west central Africa in search of HIV's closest living relatives; and then forward again, following the virus from rural Africa to colonial cities, the Caribbean, and finally the US, where the discovery of AIDS closes the circle. Crawford is candid that holes remain and later finds may change the picture, and she closes her preface with Francis Bacon: begin with doubts and you may end in certainties.
1981: the shock
In the West, doctors were congratulating themselves on getting to grips with infectious disease when a new, fatal one appeared. Early in 1981 physicians in Los Angeles reported an immune deficiency so severe that patients died of overwhelming infections; San Francisco and New York followed, then London and other European capitals. Pneumocystis pneumonia and an unusually aggressive tumour, Kaposi's sarcoma, topped the list of killers.
- The "gay plague". The disease seemed to strike only sexually active gay men; the press coined the label in 1982 even after the official name AIDS (acquired immunodeficiency syndrome) was adopted. Sufferers had often had very many partners, some claiming over a thousand, which raised two rival suspicions: "poppers" (amyl nitrite) or a new sexually transmitted microbe.
- "Patient zero". A Canadian airline steward who reckoned on about 2,500 sexual encounters in ten years developed swollen glands in 1979, Kaposi's sarcoma in 1980, and died in 1984. Tracers linked him to at least forty early cases in ten cities on two continents, which looked like strong evidence of a sexually transmitted agent.
Crawford repeats the 2013 consensus that this man was a likely spreader. In 2016 Michael Worobey and colleagues (Nature) sequenced HIV from 1970s blood samples and from the steward himself (Gaëtan Dugas). His virus was typical of strains already circulating in the US; HIV had reached New York around 1970-71, via the Caribbean, a decade before AIDS was named. The "0" in the CDC cluster study originally meant "Outside California", later misread as zero. He was neither the first case nor the source.
FROM RISK GROUPS TO EVERYONE
Within six months of the first reports injecting drug users were added to the risk list; six months later came haemophiliacs, who depend on infusions of clotting factor. In July 1982 Haitians living in the US were proposed as a source, completing the cruel shorthand of the "4-H Club": homosexuals, haemophiliacs, heroin addicts and Haitians.
The comfort of thinking AIDS belonged to "them" ended between December 1982 and January 1983: a multiply transfused child developed AIDS, mother-to-child transmission was suspected, and female partners of bisexual men fell ill. In 1984 two task forces sent to central Africa found hospitals full of AIDS patients of both sexes, a massive heterosexual epidemic.
The press stoked a parallel epidemic of "fear of AIDS". Headlines such as "Exterminate gays" fed stigma and scapegoating: children excluded from school, adults losing jobs and homes, families abused. Gay rights groups and AIDS charities pushed back with education, and the death of Rock Hudson in 1985 helped by "giving AIDS a face".
The virus and the virus war
In May 1983 the Pasteur Institute team under Luc Montagnier and Françoise Barré-Sinoussi isolated a new virus, LAV (lymphadenopathy-associated virus), from an AIDS patient's lymph gland. In April 1984 Robert Gallo's group at the US National Cancer Institute announced "the virus that causes AIDS", HTLV-III. The "great Franco-American virus war" over priority was officially settled in March 1987 when Presidents Reagan and Chirac declared co-discovery, but it rumbled on until the 2008 Nobel Prize went to the French pair alone; most scientists, Crawford says, felt right had prevailed.
The cause of a brand-new disease was identified in under three years, its genome sequenced and a blood test in use by March 1985, so carriers could be found and the blood supply made safe. The priority fight should not obscure how fast the science moved.
Crawford was then in an immunology unit at University College Hospital, London, whose collaborative group identified CD4, the receptor HIV uses to enter T lymphocytes. She assumed, naively, that 1981 marked the epidemic's start. Evolutionary biologists knew better and, from 1985, began asking where HIV came from, how it first infected humans, and why it spread so widely. At the time of writing the virus had infected 60-80 million people and killed at least 25 million.
UNAIDS's 2026 fact sheet (data for 2025): about 92.0 million people infected since the epidemic began and 44.2 million AIDS deaths; 41.0 million living with HIV; 1.2 million new infections (down from 3.5 million in 1994); 570,000 deaths (peak 2.1 million in 2004); 32.1 million on treatment.
- 1981 LA and New York report gay men with severe immunodeficiency and Kaposi's sarcoma; first UK cases; 108 US cases.
- 1982 Cluster suggests an infectious agent; haemophiliacs and Haitians affected; name AIDS adopted; 593 cases, 243 deaths.
- 1983 LAV isolated (May); WHO counts cases on five continents; 1,972 cases, 759 deaths.
- 1984 HTLV-III announced; CD4 identified as receptor; African heterosexual epidemic reported; 6,993 cases.
- 1985 First antibody test (March); first International AIDS Conference, Atlanta; UK screens all blood; 10,000 cases.
- 1986 "HIV" becomes the official name; first zidovudine trial; 28,098 cases.
- 1987 Zidovudine (AZT) approved; Reagan-Chirac accord; 40,051 cases, 23,165 deaths.
The Puzzle of HIV-1
Before hunting HIV's origin, Crawford settles that HIV causes AIDS, taking Peter Duesberg's denialism apart against Koch's postulates and four tragic "natural experiments". She then lays out the tools of the hunt: retrovirus biology, mutation rates, the molecular clock and evolutionary trees. A dress rehearsal follows: the solved mystery of simian AIDS in US primate centres, traced to a sooty mangabey virus passed to macaques during kuru experiments.
Proving the culprit
The 2008 Nobel Prize to Montagnier and Barré-Sinoussi came twenty-five years after the discovery, and Crawford thinks the committee over-cautious, but in 1983 causation was genuinely open. Everyone carries persistent viruses that reactivate when immunity fails, so a virus isolated from an AIDS patient might be a passenger rather than the cause. Scientists waited for formal proof.
THE DENIALIST CASE
Peter Duesberg, a Berkeley retrovirologist, led the opposition from the start, culminating in Inventing the AIDS Virus (1996). His claims shifted over time:
- Drugs, not a virus. Poppers and injected drugs poisoned gay men's and addicts' immune systems.
- Protein overload. For haemophiliacs and transfusion recipients he blamed foreign protein in blood, noting that half of all transfusion recipients die within a year for unrelated reasons.
- African AIDS is old disease. "A battery of previously known" illnesses (malnutrition, gastroenteritis, pneumonia, TB), with HIV an ancient, harmless infection passed from mother to child.
- A harmless passenger. HIV and AIDS coexist in the same risk groups, but the virus is innocent; he reportedly offered to inject himself with HIV.
- The drugs cause AIDS. Later he added antiretroviral therapy to his list of immune poisons.
Thabo Mbeki (president 1999-2008), advised by Duesberg, denied that HIV-caused AIDS was spreading and withheld antiretrovirals. A 2008 Harvard study estimated that in 2000-05 this caused 330,000 unnecessary deaths, 35,000 preventable infant infections and 3.8 million lost person-years. Duesberg's 2009 rebuttal paper was retracted by its publisher because of its public-health implications.
KOCH'S POSTULATES APPLIED
Robert Koch, who isolated the first human pathogenic bacterium (Bacillus anthracis, 1877), set out his postulates in the 1890s. Duesberg was right at first that HIV did not meet them; healthy scepticism is fine while evidence accumulates. But Koch wrote before viruses were known, admitted the rules would not fit every microbe, and scientists now accept indirect evidence such as antibodies, as with the SARS coronavirus in 2003, where only patients carried antibodies. PCR, invented by Kary Mullis at Cetus (Nobel 1993), then allowed tiny amounts of viral genetic sequence to be amplified and detected without culturing the virus.
| Koch's postulate | Difficulty for HIV | How it was met |
|---|---|---|
| 1. Microbe in every case | Virus hard to culture; antibodies thought "past infection" | Antibody and PCR tests positive in virtually all AIDS |
| 1b. Accounts for pathology | HIV kills indirectly; symptoms vary | CD4 T cell loss; drugs halt it, resistance brings it back |
| 2. Not found in other diseases | Persistent viruses are everywhere | Carriage alone proves nothing, as with herpesviruses |
| 3. Pure culture reproduces disease | Unethical in humans; no animal model | Four "natural experiments" |
On postulate one, Duesberg argued that most PCR-positive people did not have AIDS. The answer was time: US cohorts of gay men with identical risk factors except HIV status showed that only the HIV-positive developed AIDS, after a lag of six to fifteen years.
The key pathology is the steady loss of CD4 T cells. HIV infects and destroys an estimated 1-2 billion CD4 cells a day, producing up to 100 billion new viruses. When counts fall below 200 per microlitre, opportunistic microbes invade. The 1982 CDC definition was a loose list of conditions (pneumocystis pneumonia, thrush, shingles) in people without other causes of immune failure; in 1993 it was revised to include CD4 < 200/µL in an HIV-positive person. Drugs lower virus load and stop the decline; AIDS returns only when resistance emerges or treatment stops, a clean demonstration of cause.
Rare "HIV-negative AIDS" cases briefly boosted Duesberg. Immunologists had seen such patients before AIDS: late-onset congenital immune disorders, outside all risk groups and not rising in number.
FOUR NATURAL EXPERIMENTS
- Mother to child. Without drugs about a quarter of babies of infected mothers acquire HIV; mother and baby viruses share ~97% sequence identity, the same as two samples from one person. Only infected babies develop AIDS.
- Haemophiliacs. UK factor VIII imported from the US carried HIV from 1979 until imports stopped in 1986: 1,227 of 6,278 UK haemophiliacs were infected; in 1984-91 their death rate was ten times that of uninfected haemophiliacs, 85% from AIDS.
- The Florida dentist. Kimberly Bergalis, with no risk factors, said in 1990 that her dentist David Acer infected her. Of 1,100+ patients tested, 10 were positive; five (plus one "indeterminate") had viruses so close to Acer's that they shared one source.
- Laboratory accidents (1994). Three workers infected with a pure laboratory clone; all had low CD4 counts, one developed pneumocystis pneumonia 68 months later, and the same clone was recovered from their blood.
By the early 1990s all but a few accepted the proof. The remainder are denialists, who reject objective reality to sustain "a flawed, hurtful, and ultimately dangerous belief system".
What kind of virus
HIV works by stealth: no outward signs for about ten years while it silently dismantles the defences built to fight it, and its unknowing carriers spread it. That latency is the key to HIV's success: people living with the virus, unaware of it, carry on their lives and pass it on.
A SHORT HISTORY OF RETROVIRUSES
Retroviruses are an ancient family infecting many vertebrates, harmless in natural hosts but able to cause cancer, anaemia or immune deficiency when they cross species.
- 1908 Ellermann and Bang (Denmark) transmit chicken leukaemia with a cell-free filtrate.
- 1911 Peyton Rous (US) transfers a solid chicken tumour the same way; his Nobel comes only in 1966.
- 1930s John Bittner finds a "milk agent" (mouse mammary tumour virus).
- 1936, 1951 Ludwig Gross reports mouse leukaemia virus.
- 1970 Temin and Baltimore independently discover reverse transcriptase (Nobel 1975).
- 1981 Robert Gallo isolates HTLV-I, cause of adult T cell leukaemia; HTLV-II follows, no known disease.
- 1983 HIV isolated, using Gallo's culture and enzyme-detection methods.
Gallo's achievement rested on two techniques: growing malignant blood cells with growth factors and detecting reverse transcriptase, an enzyme only growing retroviruses make. Those methods set the stage for HIV.
THE LIFE CYCLE
Like flu and measles, retroviruses carry RNA genomes, but each particle carries two RNA copies plus reverse transcriptase, which writes RNA back into DNA. Its discovery seemed to break the central dogma (DNA → RNA → protein) and was met with disbelief before the evidence won.
Once integrated, the provirus is shielded from immune attack and copied into every daughter cell. If it reaches germ cells, it is inherited indefinitely. The human genome bears this out: about 96,000 retrovirus-like elements, roughly 8% of our DNA, probably fossils of ancient infections. Crawford speculates that some caused plagues like AIDS, and that a fossilized HIV might one day sit in the human genome.
Mutation and the molecular clock
DNA uses A, C, T and G (A pairs with T, G with C); RNA swaps thymine for uracil (U). Each three-letter group codes one of twenty amino acids. A human cell copies 3 × 10⁹ base pairs at each division, proof-reads them, and passes on roughly one error per 200 million nucleotides per year.
RNA viruses have no proofreading: about one mutation per thousand nucleotides per year, roughly a million times the human rate, with generations of hours and thousands of offspring per cycle. Most mutations are harmful or neutral; the few that raise replication rate, immune evasion or transmissibility increase viral fitness and take over. Rapid mutation is one of the keys to HIV's success.
- Premise. Mutations occur randomly but at a roughly constant rate over long periods, like radioactive decay.
- Similarity = kinship. The more alike two sequences, the more closely related the organisms.
- Difference = time. The amount of difference estimates time to the most recent common ancestor.
- Calibration. The clock is set using species whose split is dated by fossils.
- The human test. Fossils suggested a human-chimp split ~15 million years ago; the clock gave 6-7 million, a shock then, now confirmed by new fossils and the ~1% genome difference.
In an evolutionary (phylogenetic) tree, each node marks a common ancestor and branch lengths estimate time since divergence. Viruses leave no fossils, so the clock is the only way into their history, and their speed means it works best for the recent rather than the ancient past, partly because viruses swap genome segments, which blurs deep history.
Crawford illustrates the ticking with a short stretch of a flu virus gene sampled from 1933 to 1985: the 19-letter sequence picks up a changed letter every few years, and the changes accumulate rather than revert, which is what lets a clock be read from them.
Retrovirus families and the SIVs
Genome sequencing, beginning with the first complete human virus genome in 1985, replaced classification by shape, size and genome type. Retroviruses were split into genera alpha to epsilon, all sharing three major genes: gag (core proteins), pol (enzymes) and env (envelope proteins).
HIV belongs to the lentiviruses ("slow viruses": disease, if any, comes after a long lag). In 1983 its closest known relative was visna virus of sheep. Once simian immunodeficiency viruses (SIVs) were found from 1985, they proved far closer; about forty are now known, named by host: SIVagm from African green monkeys, SIVcpz from chimpanzees.
Rehearsal: simian AIDS in American primate centres
Soon after 1981 the New England Regional Primate Research Center in Boston noticed captive macaques, an Asian genus, dying of an AIDS-like disease: swollen glands, diarrhoea, wasting, lymphomas, pneumocystis, TB, thrush and herpesvirus infections, always fatal. In 1980 a third of its Taiwanese rock macaques died; stored samples showed simian AIDS as early as 1970, in animals bought from the California Primate Center, where four outbreaks had occurred in 1969-81.
In 1985 Boston isolated SIVmac from four sick rhesus macaques; injected into healthy macaques it caused simian AIDS within weeks, satisfying Koch outright. But the macaques had been shipped among centres and housed in outdoor corrals of fifty or more animals of mixed species, so macaques might not be the natural host. A second virus, SIVstm, appeared in stump-tailed macaques in California and at Yerkes in Atlanta.
About forty SIVs have been found in wild African primates and none in wild Asian primates. Macaques do not carry SIV in nature, so they must have caught it in captivity from an African species.
TRACING THE SOURCE
- 1986. SIVmac resembles a virus in healthy captive sooty mangabeys (Cercocebus atys) from West Africa.
- 1989. SIVsmm is isolated from captive and wild mangabeys, proving them the natural host. One gene differs by 12% between SIVsmm and SIVmac, no more than between two wild mangabey isolates: a split in the late 1960s or early 1970s.
- Imports. Wild mangabeys were imported for fifty years until 1968, to four centres: California, New Iberia, Yerkes and Tulane.
- The phylogeny. 84 SIVsmm genomes showed at least nine imported mangabeys were already infected and the viruses spread within colonies; just two jumped to macaques, both in California, becoming the ancestors of SIVmac and SIVstm. In macaques the virus spread by sex, mother to child and fights, and by the 1980s several subtypes were co-circulating.
Two transfers in fifty years of mixed housing suggests that ordinary contact between caged animals rarely transfers the virus. The real route was experimental. In the 1960s Carleton Gajdusek was trying to reproduce kuru, the neurodegenerative disease of the Fore people of Papua New Guinea (spread by eating the dead, later shown to be a prion, now extinct with the custom), by injecting blood and tissue into primates, brain and bloodstream, then passing "infected" material on to fresh animals. Over 1,500 animals were used. The second transfer came in the 1980s at Tulane, where leprosy material from naturally infected mangabeys (one of the few animals carrying Mycobacterium leprae) was injected into rhesus macaques. One mangabey given contaminated material became superinfected and within a month carried a recombinant virus mixing both strains, a sign of how labile these viruses are.
SIVmac is the only SIV that causes disease in macaques, and it does so in weeks, odd for a "slow virus". Direct transfer from mangabeys does not reliably cause simian AIDS; some macaques clear it. The suspicion is that injection into brain and blood plus serial passage selected an unusually virulent strain.
The simian AIDS investigation is the method of this whole book in miniature: stored samples, sequence comparison, an evolutionary tree, and archival records of who shipped which animal where. It produced an answer no clinical observation could: a man-made cross-species jump hidden for a decade, identified only because people kept old specimens and paperwork. The same combination, applied to HIV-1, is a far harder problem, as the next chapter shows.
Crawford notes the outcome's two faces. Present-day animal-research law makes a repeat unlikely; meanwhile SIVmac in macaques remains the best model for HIV pathogenesis and for testing antiretrovirals and candidate vaccines. And the detective work revealed viral diversity as great as that of pandemic HIV-1.
Tracing HIV to its Roots
Two stories run side by side here. One is the discovery, from 1979 onwards, that a huge heterosexual epidemic was already under way in central Africa. The other is the hunt for HIV's geography: the easy solution of HIV-2 (eight separate jumps from sooty mangabeys), the collapse of the famous Manchester sailor case under PCR contamination, and the survival of one 1959 Kinshasa blood sample, ZR59, which put the epicentre of the pandemic in west central Africa.
Africa, before anyone was looking
Peter Piot, a Belgian clinical microbiologist at the Institute of Tropical Medicine, Antwerp, began seeing "weird" infections in 1979, all in patients with African links, mostly from Zaire (the former Belgian Congo, now the DRC). A Greek fisherman from Lake Tanganyika died of a disseminated atypical mycobacterial infection, a normally harmless environmental microbe destroying his organs. Stored blood tested five years later showed HIV.
Piot did not connect these cases with the June 1981 American reports, because the pictures differed: US reports described gay men with Pneumocystis pneumonia and Kaposi's sarcoma; Piot was seeing heterosexual men and women equally, with neither. The link came late in 1981, when gay men at his STD clinic, most of them recent visitors to the US, began turning up with classic AIDS.
By 1982 Piot reckoned about 100 African cases had been treated in Belgium. Since only a wealthy minority could seek care abroad, this was the tip of the iceberg. Funding was hard to find, and the team of European and American specialists reached Kinshasa only in 1983.
Piot had worked in Zaire before, on one of the first recorded Ebola outbreaks at Yambuku in 1976, and was still unprepared. At Mama Yemo, 2,000 beds and among the best-equipped hospitals in sub-Saharan Africa, wards were full of dying young men and women with weight loss, persistent diarrhoea, TB and Kaposi's sarcoma. Dr Kapita, head of internal medicine and no infectious-disease specialist, was the only doctor who had noticed, and met the team with a pile of files, most of the patients already dead. Simple blood tests in a lab set up at the University Hospital of Kinshasa found the same CD4 T cell deficiency reported in America: the same disease. Piot calls it his "aha moment".
Piot joined Jonathan Mann of the CDC in 1984 to found Project SIDA in Zaire, the first and largest AIDS research project in Africa; in 1995 he became the first head of UNAIDS, serving until 2008. The team's report in The Lancet in 1984, published beside a parallel report from Kigali, Rwanda, showed HIV well established in central Africa, in both sexes, spread heterosexually, with an annual AIDS incidence higher than San Francisco or New York. That finished the idea that AIDS was a disease of sexual or ethnic minorities. By 1985 studies in both cities showed very high infection levels in female sex workers, identifying them as a route into the wider population.
SLIM DISEASE
Publicity made the African findings politically explosive. Some governments ignored the threat; most looked for someone to blame; several read the reports as the West trying to blame Africa for its own crisis. Uganda was the exception, rising to the challenge, and gave African AIDS its name: slim disease, for the chronic wasting that distinguished it from the Western picture. It appeared in the rural district of Rakai on Lake Victoria in 1982 and was widespread by the time it was described in 1985.
Months 1-6: malaise and intermittent fevers, self-treated with aspirin, chloroquine or chloramphenicol; appetite fails. Months 6-12: intermittent diarrhoea, gradual weight loss, pallor; most patients turn to traditional healers, since many attribute the illness to witchcraft. After a year: an itchy maculopapular rash over the body, skin scarred with hyperpigmentation, a cough; by the time hospital is reached, little can be done, and death follows.
Risk factors were multiple partners and prior STDs, affecting both sexes equally: clearly sexual, clearly HIV. Doctors suspected the virus reached Rakai with Tanzanian traders or soldiers and spread through female sex workers, a pattern that would repeat across Africa.
- 42% of consenting patients admitted to fifteen Ugandan hospitals in one week of 1987 tested HIV positive (early tests exaggerated African results, but the figure shocked the world).
- Over 18% of people in Kampala were positive: typically young, travelling businessmen and clerks with multiple partners and past STDs, with no evident homosexuality or injecting drug use.
- 37 of 47 African countries surveyed had HIV by 1987; worst hit were Zaire, Rwanda, Burundi, Uganda, Zambia and Tanzania. Spread followed highways and trade routes.
- Nairobi: sex workers 4% positive in 1981, 61% by 1985. Men with genital ulcers: 0 of 100+ in 1981, 15% by 1985. Pregnant women, standing for the general population: 2% by 1985.
HIV-2: the easy case
West Africa was oddly quiet. Senegal, Nigeria, Ghana, Mali, Sierra Leone, The Gambia, Guinea-Bissau and Ivory Coast reported few cases by the mid-1980s. In 1985, blood from Senegalese sex workers found 7% "HIV positive", all healthy, with antibodies that bound SIVmac proteins more strongly than HIV's. A different virus was circulating.
Montagnier's group isolated it in 1986 from two patients: a 29-year-old man from Guinea-Bissau treated in Lisbon, whose sample a visiting Portuguese scientist had brought to Paris, and a 32-year-old man from the Cape Verde Islands treated at the Claude Bernard Hospital in Paris since 1983. Both had AIDS symptoms and both had repeatedly tested HIV antibody negative.
| Feature | HIV-1 | HIV-2 |
|---|---|---|
| Closest relative | SIVcpz (chimpanzee) | SIVsmm (sooty mangabey) |
| Target and routes | CD4 T cells; sex, blood, mother to child | Same |
| Virus load in carriers | Higher | Lower |
| Transmission rate | Higher | Lower |
| Lag to AIDS | ~10 years | Longer; many never progress |
| Spread | Pandemic | Local to West Africa |
The lower virulence explains the geography: HIV-2 stayed in West Africa because it transmits poorly, not because it arrived late. Named LAV-2 at first, it became HIV-2, and the original virus HIV-1.
EIGHT JUMPS FROM THE SAME MONKEY
HIV-2's resemblance to SIVmac redirected the hunt towards African primates, and the two investigations converged: SIVsmm from sooty mangabeys turned out to be the ancestor of both. HIV-2 and SIVmac are genetically closer to each other than either is to HIV-1. SIVsmm and HIV-2 proteins are 62-87% identical, implying a common ancestor perhaps 50-60 years before.
Which way did the virus jump? Early sequences showed more variation among HIV-2 genomes than among SIVsmm, hinting that it had evolved longer in humans. More samples equalised the diversity, and three facts settled it for the monkey:
- Prevalence. About 50% of wild sooty mangabeys carry SIVsmm; the highest national HIV-2 rate in humans is 8-10%.
- Pathology. SIVsmm is non-pathogenic in mangabeys, the mark of long association; HIV-2 causes AIDS.
- Local matching. SIVsmm isolates from pet mangabeys in Liberia and Sierra Leone are genetically divergent from each other but closest to the HIV-2 strains of people living in the same area.
The route is blood contact: bites or cuts while hunting and butchering bushmeat, or bites from orphaned animals kept as pets.
Each HIV-2 group A to H is closer to a particular SIVsmm strain than to the other HIV-2 groups. They cannot descend from one another or from a single jump, so each represents a separate transmission event from mangabey to human. Only A and B ever spread; the rest were each found in a single person. The lesson for HIV-1 is blunt: cross-species jumps are common; epidemics are rare.
Dating 33 group A and B genomes with known sampling dates gave a most recent common ancestor around 1940 for group A and 1945 for group B, the latest possible jump dates; the common ancestor of A and B within mangabeys dated to the late nineteenth century, the earliest. A study of free-ranging mangabeys in the Taï Forest, Ivory Coast, found viruses clustering with five of the eight groups, including A and B, suggesting that area as the site of transfer, though it leaves unexplained how A and B reached Guinea-Bissau.
WHY GUINEA-BISSAU, WHY 1955
Guinea-Bissau has the highest HIV-2 prevalence; the epicentre is the rural northeast around Canchungo, where a 1992 survey found 8% HIV-2 against 0.1% HIV-1. In the capital, Bissau, the age pattern is inverted compared with HIV-1: infection peaks at about 15% in people aged 50-69, where HIV-1 stands at 0.5%.
Formerly Portuguese Guinea, the country won independence in 1974 after an eleven-year war. The 50-69 cohort surveyed in 2000 were in their twenties and thirties during the fighting. Sequence analysis fits: HIV-2 grew slowly through the 1940s and early 1950s, then switched to epidemic mode around 1955 and grew exponentially. Between 1955 and 1970 its growth rate in Guinea-Bissau exceeded that of HIV-1 in west central Africa.
- Soldiers and movement. An influx of Portuguese troops, migrant labour and commercial sex multiplied sexual contacts; being in the army was itself a risk factor for men.
- Mass injection campaigns. Army doctors ran large vaccination and treatment programmes for TB and sleeping sickness, with reuse of unsterilised needles then normal practice.
- Transfusion. Blood transfusion spread to treat war casualties and civilians. A 57-year-old Portuguese woman, whose family lived in Guinea-Bissau 1960-74, was found HIV-2 positive in 1991; her only risk was a transfusion in 1967 after a miscarriage. Her husband and children, including one born later, were uninfected. The blood supply was contaminated by 1967.
- The export. Portugal was the first country outside West Africa to report HIV-2 AIDS; about 12% of Portuguese AIDS is HIV-2, mostly in people from Guinea-Bissau, visitors, workers and armed-forces veterans.
HIV-1: wrong turns
By the 1980s many suspected west central Africa with import to the US in the 1970s, but there was no evidence and no consensus. Fringe claims filled the gap: an angry god, outer space, germ warfare by the US or by the Nazis. A serious minority argued the virus was long endemic in the US or Europe, unnoticed because it could not spread in a less promiscuous society. One version derived HIV-1 from visna virus, its closest known relative before the SIVs: isolated from sheep in 1949, causing encephalitis and pneumonia in Icelandic sheep since the 1930s. On that theory the trigger was social: the sexual revolution of the 1960s and gay liberation of the 1970s, with syphilis and hepatitis B both tripling in gay men during that decade.
- The literature trawl. Search medical journals before 1981 for fatal immunodeficiency or Kaposi's sarcoma meeting the AIDS definition.
- The freezer hunt. Find and test old stored blood collections.
- The sequence collection. Gather HIV-1 genomes worldwide and map their diversity. Only the three combined gave the answer.
Dozens of plausible retrospective cases emerged. A 28-year-old man in Memphis diagnosed with viral pneumonia in February 1952 died that December with cytomegalovirus and Pneumocystis. European cases from the 1970s were mostly gay men who had visited the US or Haiti, suggesting arrival through international sexual networks, but several were Europeans who had lived in central Africa: a 47-year-old Danish surgeon who worked in northern Zaire 1972-75 and Kinshasa until 1977, ill with diarrhoea from 1976, dead of Pneumocystis pneumonia in December 1977. In August 1977 a 34-year-old woman from Kinshasa brought her three-month-old daughter to hospital in Louvain; two of her children had already died at six months, a third had thrush since birth, and she herself fell ill, returning to Kinshasa to die in February 1978. None of these cases left testable stored samples.
Dr Kapita remembered his first Kinshasa AIDS case in 1978: a Haitian nurse whose Congolese husband stayed healthy. He suspected that professionally trained Haitians who came in the 1960s and 1970s had brought the virus, an idea the later chapters invert.
Africa's background of fatal infection in the young hid early cases, but Kaposi's sarcoma was conspicuous there. Before the pandemic it was common in Zaire, Rwanda, Burundi, Cameroon, Tanzania and Uganda, indolent, in adult men, with survival of ten years or more; an aggressive childhood form existed. In the West it was rare, mostly in elderly men of Mediterranean or Jewish descent. US incidence rose in the late 1970s as the tumour turned aggressive. African incidence rose from about 1980, and in 1983 Anne Bailey, a surgeon in Lusaka, Zambia, identified an atypical aggressive form. In her series from Zambia and Uganda, 96% of aggressive cases were HIV-1 positive against 17% of classic indolent cases. Her 1985 report also found 20% of healthy Ugandan controls positive against 2% of Zambian controls: HIV was long established in Uganda and newly arrived in Zambia.
TWO SAILORS
- The Manchester sailor: 25-year-old Royal Navy national serviceman, ill from December 1958, dead September 1959 of Pneumocystis pneumonia and cytomegalovirus; claimed in 1990 as the world's first AIDS case.
- The Norwegian sailor: Royal Norwegian Navy 1961-65, sailed to Nigerian ports in 1961-62 and on to Asia, Europe, Canada, the Caribbean and East Africa; two STDs while serving; ill from 1966, dementia and lung disease, died April 1976.
- His wife: ill from 1967 aged 24 with thrush and recurrent infections, then wasting and encephalitis; died of acute leukaemia, December 1976.
- Their youngest daughter, born 1967: recurrent severe infections from age two, died of chickenpox in 1976. All three had low lymphocyte function by 1971; all three tested positive on retesting in 1988.
The Manchester case looked ideal. His foreign travel amounted to Gibraltar and perhaps a day in Tangiers; TB treatment failed; he wasted away with herpes ulcers and died of pneumonia, with organs fixed in formalin and stored. In the late 1980s PCR on formalin-fixed tissue became feasible; Manchester virologists ran blinded tests with tissue from a 1959 road-accident victim as negative control and an HIV-1 cell line as positive. Kidney, bone marrow, spleen and pharynx came back positive, reported in 1990.
Sequences amplified at the Aaron Diamond AIDS Research Center and placed in the Los Alamos tree fell among the most recent subtype B viruses circulating in the US and Europe, impossible after thirty years of fast evolution. Tissue typing then showed the samples came from two different people. David Ho and Tuofu Zhu reported in Nature in 1995 that the result raised "the spectre of specimen contamination". A laboratory HIV-1 strain had contaminated the samples. Crawford's moral: a cautionary tale for would-be PCR technologists.
The Norwegian family survived scrutiny but changed the question. In 1997, partial genomes from the sailor and his daughter (none could be amplified from his wife) showed infection with HIV-1 group O, not the pandemic strain, presumably picked up in West African ports in 1961-62. Group O's divergence also explains why the first antibody tests on the family in 1986 were negative.
Diversity, groups and the oldest genome
Global sequencing from the mid-1980s, pooled at the Los Alamos HIV database, revealed extraordinary diversity: no two isolates identical, all clearly from one ancestor. Then came the exceptions.
| Group | Found | Where | Share of HIV-1 |
|---|---|---|---|
| M (major) | 1983 | Worldwide | Virtually all |
| O (outlier) | 1990, Piot's lab, Cameroonian patient | Cameroon and neighbours | 1-5% locally |
| N (non-M, non-O) | 1998, Cameroonian patient | West/central Africa | Extremely rare |
Group O proteins are only about 50% identical to pandemic HIV-1. Because of that diversity and their interspersed positions among the SIVs, the three groups must have been introduced into humans separately.
Group M's global spread plus its mutation rate produced nine subtypes (A-D, F-H, J, K) differing by 25-35% in envelope protein sequence, diverging almost simultaneously in the "starburst", plus sub-subtypes and recombinants, formed when one person carries two viruses. Recombination is common: subtypes E and I do not exist because the original isolates proved to be recombinants (E was A plus E; I was mainly A, G, H and K). Recombinants are about 18% of all group M infections and rising.
The geography tells the story: subtype C alone accounts for almost half of all infections, but virtually every subtype and many recombinants circulate in west central Africa, the signature of a long residence.
Blood collections were scarce. Nothing positive was found in the US, Haiti or Europe before the 1970s, and DRC freezers failed during civil unrest. Of 659 samples from Yambuku taken during the 1976 Ebola outbreak, just five (0.8%) were HIV-1 positive, and three of those villagers had died by 1985-86; rates there were unchanged ten years later, so the rural epidemic had not yet taken off while Kinshasa reached 5% by 1985. A collection of over 1,000 African samples (1959-82), stored in a US freezer, included 818 from 1959. Early tests called 92% positive, a known artefact; retesting the strongest by two further methods left one unequivocal positive, L70, from an adult Bantu man in Leopoldville in 1959. In 1998, from a minute quantity after forty years, Zhu and Ho amplified short sequences: ZR59, still the oldest HIV-1 genome in existence.
ZR59 is not a recombinant, which makes it easy to place. It shows features of both subtype B and D, which are closer to each other than to other subtypes, and it sits near their common ancestor. That single position rules out subtype B circulating in Europe in the 1930s, and with it Europe as the birthplace of the pandemic.
ZR59, plus the extraordinary diversity and many recombinant forms of group M in Kinshasa, where all known subtypes co-circulate with A predominating, point to one conclusion: the virus has been in that city longer than anywhere else on Earth. West central Africa was the epicentre, so that is where the search for HIV-1's animal source went.
The Primate Connection
Finding HIV-2's parent took a few years; finding HIV-1's took twenty. This chapter follows the SIV survey from African green monkeys through Martine Peeters's three infected chimpanzees and a dead US research chimp called Marilyn, to the moment when Beatrice Hahn and Paul Sharp matched virus phylogeny to chimpanzee subspecies and fingered Pan troglodytes troglodytes. The last twist: SIVcpz is itself a recombinant, assembled inside a chimpanzee from two monkey viruses.
What the SIV survey showed
By the mid-1990s the epicentre was fixed in west central Africa: DRC and Cameroon. The virus had crossed into humans at least three times, yet its animal reservoir was unknown, and finding it mattered for preventing the next jump.
Two strands of evidence pointed at a local SIV. First, as isolates accumulated (two species in 1985, twenty by the late 1990s, forty by 2010), every one was a lentivirus with HIV-1-like genome structure, all genetically related to HIV-1, so all from a shared ancestor. Second, the primates of the DRC, Gabon, Republic of Congo and Cameroon forests, chimpanzees, bonobos, gorillas, baboons, drills, mandrills and many small monkeys, are hunted for food with young kept as pets, which supplies the blood contact route.
The practical obstacle was sampling. Capturing and bleeding wild primates at scale is impossible, so the first surveys used zoo and primate-centre animals and found little: by 1989 just two additions, SIVagm from African green monkeys and SIVmnd from the mandrill. Both sat roughly equidistant from the HIV-1 and HIV-2 clusters, and both lacked vpu, a tiny accessory gene then found only in HIV-1. The presence or absence of vpu became the single most useful marker in the hunt for HIV-1's ancestor.
CO-EVOLUTION OR JUMPING: THE GREEN MONKEY TEST
Co-evolution One ancestral lentivirus infected the common ancestor of African Old World monkeys millions of years ago and diverged as the hosts speciated. Prediction: virus relatedness mirrors host relatedness.
Cross-species spread A lentivirus of one species jumped repeatedly to others. Prediction: virus relatedness mirrors overlapping ranges rather than host family trees.
African green monkeys make a good test case: among Africa's most common primates, living in territorial groups of up to eighty, in four species with non-overlapping ranges, vervet, grivet, sabaeus and tantalus. Well over half of wild adults carry a lentivirus harmlessly.
- The co-evolution evidence. SIVagm isolates cluster by monkey species, even when the monkeys live hundreds of miles apart. High prevalence plus harmlessness points to long association, from the green monkey common ancestor several million years ago.
- The Caribbean anomaly. Green monkeys taken across the Atlantic as pets during the 17th and 18th century slave trade still live wild on some islands. Over a hundred tested showed no SIV, which would mean the virus entered African populations only in the last 200 years. The counterargument: only young animals were shipped, too young to catch a mainly sexually transmitted virus.
- Jumping does happen. A captive-born white-crowned mangabey carried a vervet-type virus, probably from fights. More tellingly, two free-ranging yellow baboons in Mikumi National Park, Tanzania, living alongside vervets, had SIVagm antibodies, and one carried a vervet-like virus. Baboons there eat small monkeys.
- Ancient recombination. The SIVagm of West African sabaeus monkeys is a recombinant: half its genome clusters with other SIVagm, half with the mangabey/HIV-2 group. A mangabey virus jumped into a sabaeus monkey and recombined with the resident virus, long enough ago that both parents were predecessors of the modern SIVs.
Crawford's verdict is honest: with both co-evolution and repeated cross-transmission plus recombination on the record, which predominates is still unresolved.
Peeters finds chimpanzees with HIV-like virus
Martine Peeters, a Belgian scientist and colleague of Piot in Antwerp, moved in 1985 to the Centre International de Recherches Médicales at Franceville, Gabon, with Eric Delaporte, to work on STDs. Finding little HIV in the Gabonese population, she turned her antibody screen on non-human primates: hundreds of samples, mostly captive animals at the centre plus a few household pets, including fifty wild-born chimpanzees (Pan troglodytes) caught in Gabon's rainforest.
One of fifty chimpanzees reacted with HIV-1 proteins: a healthy 4-year-old female, caught at six months after hunters killed her mother. While Peeters worked on isolating her virus, a second young chimpanzee arrived needing treatment, its mother also shot and the infant wounded; it died a week later. It too was HIV-1 antibody positive. The viruses became SIVcpz-Gab-1 and SIVcpz-Gab-2.
Peeters's 1989 paper reported a lentivirus related to but not identical with HIV-1: its Env protein, normally the most variable part of the genome, was about 65% identical to HIV-1's, making it HIV-1's closest known relative. Decisively, it carried the vpu gene.
Back in Antwerp from 1989, Peeters screened captive primates in Belgium and found one more: Noah, a 5-year-old male born in the DRC rainforest, caught for the pet trade, smuggled into Belgium at 2 or 3 years old and seized by Brussels customs. His virus, SIVcpz-ant, also had vpu, but its sequence was twice as far from the two Gabon viruses as they were from each other, a distance typical of different primate species.
Most experts now expected SIVcpz to be HIV-1's ancestor, but two facts argued against chimpanzees being the natural reservoir: the diversity of the three isolates and the very low infection rate, only three carriers among over 100 wild-born chimpanzees tested. Either infection varied by region, with undiscovered high-prevalence pockets, or both chimpanzees and humans had caught the virus from some other species.
The animal you find the virus in need not be its home. In 2003, SARS coronavirus was isolated from several species in Guangdong wet markets, most often the Himalayan masked palm civet, the likely source of human infection. The true wild reservoir turned out to be horseshoe bats. A market intermediary can look exactly like a reservoir.
Marilyn, and a phone call
Beatrice Hahn had written her doctoral thesis in Germany on retrovirus evolution, trained as a molecular biologist in Robert Gallo's laboratory cloning HTLV-I and -II, and took up genetic analysis of HIV when it was discovered in 1983, work that stood apart from the ownership dispute. At the University of Alabama at Birmingham she ran her own group, isolating HIV-2 strains from Ivory Coast.
The break came by telephone. Larry Arthur, a former colleague at the National Cancer Institute, had screened a colony of over ninety chimpanzees at a New Mexico primate centre used for hepatitis research, looking for an HIV animal model. One sample reacted strongly with HIV proteins in all five tests.
- Born in African rainforest, country unknown; her records burned with the building that stored them.
- 1963: caught aged about 4 and imported to the US; used in US armed forces space research at Holloman Air Force Base.
- 1966-69: hepatitis virus experiments at the primate centre, receiving infusions of human blood products.
- 14 pregnancies, 6 live young, none infected. Died a week after delivering stillborn twins, of uterine infection and pneumonia, with enlarged glands and spleen but no typical early AIDS changes.
- 1994: Arthur offers Hahn the frozen lymph gland and spleen he had kept for over ten years and was about to discard, warning that the samples might carry hepatitis viruses, HIV "or potentially something worse".
Hahn's group amplified a complete lentivirus genome, different enough from HIV-1 to rule out contamination from Marilyn's human blood infusions. The virus, SIVcpz-us, carried vpu and was about 90% identical to the two Gabon viruses, while SIVcpz-ant was only 77% identical to the other three. The tree now held two distinct clusters: three viruses with HIV-1, and Noah's outlier alone.
Sharp's chimpanzee papers
Paul Sharp, an evolutionary biologist trained in genetics at Edinburgh, had analysed Hahn's HIV and SIV sequences for twenty years from Nottingham, Dublin and, from 2007, Edinburgh again. He was first to spot recombination among HIV subtypes. In conversation he pointed Hahn to two papers on chimpanzee evolution from 1994 and 1997.
Almost nothing was known about the evolutionary history of Pan until the 1990s. Chimpanzees and bonobos (Pan paniscus) are separated by the Congo River, bonobos living south of it; both are poor or reluctant swimmers, so the river's formation 1-2 million years ago probably split them.
Three subspecies were recognised in the 1990s: western P.t. Verus, central P.t. Troglodytes, eastern P.t. Schweinfurthii. They cannot be told apart by coat or bone structure, and with no fossils available their definition rested entirely on geography, with the Niger and Ubangi rivers assumed to be barriers.
HAIR FROM ABANDONED NESTS
Genetic study became possible in 1994 when a UC San Diego team worked at Gombe, Tanzania's smallest national park on Lake Tanganyika, where Jane Goodall had studied chimpanzees for over fifty years by strictly non-invasive methods. Adults build a fresh treetop nest of twigs and leaves each night and sleep alone except for mothers with infants, so the researchers collected hairs from abandoned nests for DNA: 43 animals of the Kasakela community, all individually recognisable after study since the 1960s, plus 24 animals from twenty other sites across the range.
The survey mostly confirmed the morphological scheme but showed P.t. Verus far more distant from the other two, isolated for about 1.6 million years, possibly a species in its own right. A 1997 study of Nigerian chimpanzees, again from nest hairs, found two distinct West African groups differing more from each other than troglodytes and schweinfurthii do: P.t. Verus in Senegal and Ghana, and a Nigerian group, now P.t. Ellioti. Chimpanzees on both banks of the Niger proved to be one genetic group, so the real divide was the Sanaga River in Cameroon.
Matching virus to subspecies
Hahn needed the subspecies of all four infected chimpanzees. She had Marilyn; Peeters, by then in Montpellier, still had frozen samples from Noah, Gab 1 and Gab 2, and sent them. Crawford notes the collegiate exchange of material between the two groups as a reason for their success.
They used mitochondrial DNA, which comes from bacteria that invaded primitive cells, sits outside the chromosomes and mutates faster than chromosomal DNA, making it ideal for distinguishing individuals and subspecies.
| Chimpanzee | Origin | Subspecies | Virus |
|---|---|---|---|
| Marilyn | Africa, unknown country | P.t. Troglodytes | SIVcpz-us |
| Gab 1 | Gabon | P.t. Troglodytes | SIVcpz-Gab-1 |
| Gab 2 | Gabon | P.t. Troglodytes | SIVcpz-Gab-2 |
| Noah | DRC | P.t. Schweinfurthii | SIVcpz-ant (divergent) |
The virus split mirrored the host split exactly: SIVcpz had co-evolved with its host subspecies, just as SIVagm had with green monkey species. The result was published in 1999.
By 1999 HIV-1 groups M, N and O were known, and in the tree they are interspersed with SIVcpz isolates rather than being each other's closest relatives, so they cannot descend one from another after a single transfer: each jumped from chimpanzee to human independently. All three are closer to viruses of P.t. Troglodytes than to Noah's P.t. Schweinfurthii virus, and the troglodytes range in south Cameroon, Gabon and Republic of Congo matches where groups N and O are found.
The earlier low prevalence figure was an artefact. Wild P.t. Troglodytes infection varies by locality and reaches 50% in some groups; the other subspecies, P.t. Verus and P.t. Ellioti, are free of SIVcpz, and most of the roughly 1,000 captive chimpanzees tested earlier had been P.t. Verus.
HIV-1's ancestor was itself a hybrid
A lingering anomaly: SIVs are overwhelmingly monkey viruses, so why did an ape carry one? Two of the newly isolated viruses answered it, both spotted by Sharp's group: SIVrcm from the red-capped mangabey and SIVgsn from the greater spot-nosed monkey.
- Split resemblance. The pol gene in the first half of SIVgsn is closest to pol in SIVcpz; the env gene at the other end of SIVrcm resembles env in SIVcpz.
- The right reading. Rather than both monkey viruses being recombinants of SIVcpz with something unknown, SIVcpz is the recombinant: roughly half SIVgsn genome, half SIVrcm genome. The vpu gene of SIVcpz and HIV-1 came from the SIVgsn side and is found only in that small group of viruses.
- How it happened. Both monkeys' ranges overlap P.t. Troglodytes, and chimpanzees prey on small monkeys. One or both viruses probably circulated in chimpanzees before meeting in one animal, where they recombined. Both troglodytes and schweinfurthii carry the recombinant, so the event predates their split about 100,000 years ago, or else spread between them later.
Crawford closes with the caveat the authors themselves made: with only a handful of chimpanzees and viruses studied, a monkey reservoir could still be hiding in the rainforest. The answer required sampling wild chimpanzees, which, as the next chapter shows, was far easier to propose than to do.
This is a reconstruction built from borrowed freezer scraps, nest hair, mitochondrial DNA and colonial-era paperwork, with the decisive step being a cross-domain conversation: a virologist's puzzle solved by two papers on chimpanzee population genetics. The marker that mattered, vpu, was a detail nobody had collected for that purpose. Keeping specimens and reading outside your field are the two cheap habits that made a twenty-year problem solvable at all.
From Rainforest to Research Laboratory
To prove that wild chimpanzees carry HIV-1's ancestor, Hahn's group had to read virus genomes out of faeces picked off the forest floor. The payoff was exact: group M came from chimpanzees in the extreme southeast of Cameroon, group N from one community at EK in the south centre. The same method then found SIVgor in wild gorillas, a fourth human group (P) in a Paris patient, and the unwelcome news that SIVcpz is not harmless to chimpanzees.
Why seven captive animals were not enough
Hahn, Sharp and Peeters had their candidate, SIVcpz of P. T. Troglodytes, but only three captive-chimpanzee genomes supported it. A 2000 report showed why that was shaky: three more antibody-positive wild-born Cameroonian chimpanzees, Cam3, Cam4 and Cam5. Cam3 and Cam4 were caught in 1992, first kept as pets, then housed together at a wildlife rescue centre; Cam5 was taken as an infant in 1998 to the Yaoundé zoo. Cam4 belonged to the Nigerian subspecies P.t. Ellioti, yet its virus was ~96% identical to Cam3's, which is P.t. Troglodytes.
Cam3 almost certainly arrived infected and passed the virus to Cam4 in captivity: the first demonstration of SIVcpz spreading between captive animals. Data from captive primates cannot be trusted to describe what viruses do in the wild.
All four of these viruses, plus SIVcpz-us from Marilyn (caught 1963, a thirty-year gap), clustered with HIV-1 group N, the very rare Cameroonian group. That raised the question the next studies had to answer: why did every chimpanzee virus resemble group N rather than the pandemic group M? With a seventh infected captive animal, Cam13, the sample was still too small to rule out a third, unknown host that had infected both chimpanzees and humans.
The problems with wild chimpanzees
- Communities run 5 to 150 animals, in remote forest, constantly moving, a new nest every night.
- Mature males stand about 1.7 m and weigh up to 70 kg: not animals to approach casually.
- A century ago a few million chimpanzees; now perhaps 150,000, classified endangered.
- Gombe: about 150 chimpanzees when Goodall began in 1960, about 100 now, in a 52 km² forest island surrounded by cleared land.
- Bushmeat is a $2 billion a year international trade removing about 1 million tonnes of meat from African forests annually and cutting chimpanzee numbers 5-7% a year.
- A female produces one offspring every 5-6 years, so losses outrun replacement. Some conservationists predict extinction in the wild by 2040.
Habitat loss pushes chimpanzees into closer contact with people, and they catch what we catch: flu, polio, and Ebola from fruit bats. After the human outbreak at Yambuku in 1976, Ebola struck chimpanzees in Taï National Park, Ivory Coast, in 1994, killing a quarter of a 43-strong community and infecting the pathologist who autopsied one. A 2004 report found that every human Ebola outbreak in Gabon and DRC over the previous three years began with someone, usually a hunter, handling the carcass of a chimpanzee, gorilla or antelope.
Hunting pressure has three sources: food, long sustainable; defence of homes and crops as farms encroach; and the commercial bushmeat trade, the serious threat. The Jane Goodall Institute, the WWF, governments and local villagers work on it, with mixed prospects.
Research ethics closed the obvious door. Blood requires tranquillizer darts, which can harm the animals. Goodall's fifty years of non-invasive work at Gombe changed how chimpanzees are regarded: invasive research on great apes is now banned in most countries because it induces anxiety and depression like that of traumatized humans, and the world's captive population has shrunk as research animals move to sanctuaries.
Reading a virus out of faeces
Hair roots, the method used for subspecies work, carry host DNA but too little blood for viral sequences. That left urine or faeces.
| Sample type | Antibody detection | Viral RNA by PCR | Collectable in wild forest |
|---|---|---|---|
| Urine | 100% of infected animals | None detected | Very hard |
| Faeces | 65% | Two-thirds positive | Yes, from the forest floor |
Urine won the first round: stored samples from primatologists at Kibale were all negative, but the second batch, from Gombe, gave the first proof that antibodies to HIV-1 proteins can be found in the urine of infected wild chimpanzees. For genome sequences, though, faeces were the only option.
- Antibody detection and PCR amplification from spiked chimpanzee faeces: two years of development.
- Sex, species and subspecies identification from the same sample.
- Genetic fingerprinting to name the individual animal that produced a given dropping.
- Validation against faeces from captive and habituated animals of known SIV and HIV status.
- Preservation: viral RNA is chopped up by enzymes within hours, and neither refrigeration nor freezing is reliably available in the rainforest. A commercial RNA stabilising fluid held up even at rainforest temperatures.
Peeters, by then at the IRD and University of Montpellier, was working in Gabon and Cameroon and organised collection. Ten field sites were chosen, nine south of the Sanaga River (P.t. Troglodytes) and one north (P.t. Ellioti). Local trackers gathered fresh samples from the forest floor, dropping each into preserving fluid for shipment to France and the US.
- 513 well enough preserved for RNA extraction after the journey.
- 67 turned out to be gorilla, not chimpanzee.
- 446 chimpanzee samples: 423 from P.t. Troglodytes, 23 from P.t. Ellioti, all of the latter from the site north of the Sanaga, exactly as the subspecies map predicts.
- 34 samples reacted with HIV-1 proteins, all from P.t. Troglodytes; every P.t. Ellioti sample was negative.
- The 34 came from 16 individual chimpanzees (7 male, 9 female), each yielding viral RNA.
- Infected animals were in 5 of the 10 communities; within an infected community prevalence ran 4% to 35%.
THE 2006 RESULT
Sharp's tree put all the new sequences with the captive P.t. Troglodytes viruses and the Cam isolates, in a cluster that included HIV-1 groups M and N but excluded group O and the P.t. Schweinfurthii virus: neither of those came from any Cameroon site sampled.
The sequences showed evolutionary-geographic clustering: viruses from the same community are most alike, neighbouring communities next, and rivers mark the breaks. That let Sharp place the human viruses on the map.
The 2006 paper concluded that no significantly closer SIVcpz strains to groups M and N were likely to exist, pointing to southeastern and south central Cameroon as the two sources. That statement has held: no SIVcpz closer to HIV-1 group M or N has been found since. A later survey at fifteen Cameroon sites confirmed that the group M ancestor is unique to chimpanzees in the extreme southeast, and the group N ancestor only at EK: chimpanzees just across the Dja River carry a quite different strain.
Gorillas enter the story
The team had already been collecting gorilla faeces almost by accident. Gorillas are the largest living apes, sharing a common ancestor with us about 9 million years ago, ground-dwelling, strictly herbivorous, living in troops of up to thirty, and endangered, with Ebola outbreaks wiping out whole troops.
| Species | Subspecies | Range |
|---|---|---|
| Western gorilla (G. Gorilla) | Cross River (G. G. Diehli) | Nigeria, western Cameroon |
| Western lowland (G. G. Gorilla) | S Cameroon, Gabon, Eq. Guinea, Congo | |
| Eastern gorilla (G. Beringei) | Mountain (G. B. Beringei) | North and east of the Congo River |
| Grauer's (G. B. Graueri) | Eastern DRC | |
| Bwindi (possible third) | Uganda |
Of 213 wild gorilla faecal samples, six were positive, yielding three distinct virus sequences from three western lowland gorillas. The new virus, SIVgor, formed its own lineage inside the SIV/HIV-1 cluster and is the closest known relative of HIV-1 group O. (All 55 P.t. Ellioti samples were negative, as expected.)
WHERE SIVgor CAME FROM
SIVgor is closest to SIVcpz, and like every HIV-1 group it resembles the P.t. Troglodytes virus more than the schweinfurthii one, so it came from chimpanzees, recently. Two scenarios were possible: one jump followed by endemic spread among gorillas, or repeated occasional jumps. Three facts favour the first: the gorillas sampled lived nearly 400 km apart, their viruses cluster together, and where chimpanzee and gorilla territories overlap the two species' viruses are distinct enough to rule out transfer within the carriers' lifetimes. The jump is dated 100-200 years ago, possibly earlier once more sequences exist.
In the laboratory SIVgor behaves like HIV-1 and SIVcpz: it infects and replicates in chimpanzee and human CD4 T cells via the CD4 receptor, so presumably spreads by blood and sex. How a herbivore first caught it is unclear. Gorillas have acquired other chimpanzee viruses, including hepatitis B, which travels the same routes; the two apes feed in the same trees when fruit is good, so a fight drawing blood is possible, as is indirect contact with urine, faeces or saliva. Once inside a troop, spread is easy: a silverback mates with all adult females, mothers can infect infants, young males leave and found new troops, and animals of both sexes transfer between troops with overlapping territories.
- Scenario 1. SIVcpz jumped independently from chimpanzees to humans (group O) and to gorillas (SIVgor). Confirmed if separate SIVcpz isolates ancestral to each are found.
- Scenario 2. Chimpanzee to gorilla, then gorilla to human. Confirmed if new SIVgor isolates ancestral to group O (or to the whole O/SIVgor cluster) turn up.
- Scenario 3. Chimpanzee to human, then human to gorilla, unlikely since wild gorillas rarely meet human blood. Confirmed if new HIV-1 isolates ancestral to SIVgor appear. The test sites are where chimpanzees and gorillas coexist in Gabon, DRC or Equatorial Guinea, outside the heavily sampled parts of Cameroon.
The most recent survey in the book tested western lowland and Grauer's gorillas and found SIVgor only in the western species in Cameroon, at two of thirteen field sites, an overall prevalence of 1.6% against 5.9% for SIVcpz in chimpanzees in the same study.
Crawford leaves the three scenarios open. In 2015 D'arc, Ayouba, Peeters, Sharp, Hahn and colleagues (PNAS) reported SIVgor strains from western lowland gorillas in southwestern Cameroon that fall ancestral to HIV-1 group O, which supports scenario 2: chimpanzee to gorilla, then gorilla to human, for both group O and probably group P. Gorillas, not chimpanzees, are the proximate source of two of the four HIV-1 groups.
Group P: a fourth jump, found in Paris
In 2004 a 62-year-old woman from Yaoundé moved to Paris and was diagnosed with HIV-1 after a year of weight loss and recurrent fevers, without full AIDS. Her unusual antibody pattern suggested group O, which causes about 6% of Cameroonian infections, but no group O sequences were present. A near-complete genome placed her virus beside SIVgor, closer to SIVgor and SIVcpz than to any HIV-1 group: a separate transmission, named HIV-1 group P, the fourth HIV-1 and the twelfth SIV known to have crossed into humans.
A screen of 1,736 HIV-positive Cameroonian samples found exactly one more: a 54-year-old man at the Jamot Hospital, Yaoundé, whose virus was 87% identical to the first. Both are closer to SIVgor than group O viruses are.
Is SIVcpz harmless to chimpanzees?
The assumption was yes, on the usual reasoning: SIVs are ancient parasites co-evolved with their hosts, infected captive and wild chimpanzees showed no AIDS-like illness, and experimental HIV-1 infections of chimpanzees rarely caused disease. Crawford lists the reasons to doubt it:
- Thin evidence. Only two SIV infections, in sooty mangabeys and African green monkeys, had been studied closely enough to be called harmless. Of seven naturally infected captive chimpanzees, several died as infants, and only Noah was followed in the laboratory; he was alive and well 25 years after capture.
- The pattern looks recent. SIVcpz prevalence is lower and patchier than SIVsmm or SIVagm; SIVs are primarily monkey viruses; and SIVcpz is a recombinant assembled from two monkey viruses a chimpanzee ate.
Hahn, Sharp and colleagues tested most members of the Kasekela (~65 animals) and Mitumba (~25) communities annually by urine or faeces. Prevalence ran 9-18%. Nine animals were infected at first test; eight more acquired infection during the study. Two were infants of infected mothers, with mother and infant viruses nearly identical. Four Kasekela animals (two male, two female) were infected within a twenty-month window with near-identical viruses, pointing to sexual transmission. SIVcpz travels among chimpanzees exactly as HIV does among people.
- 10 to 16 times higher death rate in infected animals over nine years.
- Three times lower birth rate in infected females, with significantly higher infant mortality (echoing Marilyn: six live young from fourteen pregnancies).
- One infected female, weak and lethargic before death, had multiple abdominal abscesses and tissues severely depleted of T lymphocytes, as in human HIV.
- Lymph glands of infected animals held significantly fewer CD4 T cells than uninfected ones.
SIVcpz is not a harmless passenger in chimpanzees: it causes progressive immune deficiency, an AIDS-like death, and damages fertility and infant survival.
THE MATHEMATICAL MODEL
To ask whether the virus is driving the wider decline, the team compared the two habituated communities with Kalande, a non-habituated community of about sixteen in the south of the park, shrinking since 1999, blamed on poaching and food shortage.
| Community | Annual growth, 2002-09 | SIVcpz prevalence |
|---|---|---|
| Kasekela | +2.4% | 12% |
| Mitumba | +1.9% | 13% |
| Kalande | -7% | 46% |
Feeding population size, habitat loss, mating, migration, transmission and deaths into a model, with faecal genetics supplying individual migrations, paternities and who gave the virus to whom, they found the prevalence that tips a community into negative growth and eventual extinction is only 3.4%. All three Gombe communities were above it.
The model shows a fine balance between extinction of the virus and extinction of the community, decided by population structure. Immigration of breeding females from neighbouring communities rescues a declining group, and the chimpanzees win even when some immigrants are infected, because the gain in breeders outweighs the virus. Kalande lost that protection: the death of four males in 2002 prompted females and their offspring to leave, with no comparable immigration back. Habitat loss was not a significant factor in its decline.
Phylogeny traced virtually all Gombe virus strains back to Kalande, so migrating females carried the infection north into Kasekela and Mitumba. If the lag before symptoms matches HIV-1's, around nine years as estimated in rural Uganda, those communities' recent infection plus new female immigrants explains why they are still growing. Whether continued habitat destruction cuts off migration, and with it the rescue mechanism, is the open question.
The patchiness fits the model: screening in national parks found infected chimpanzees only in DRC, with animals in Uganda, Tanzania and Rwanda negative, and all screened bonobo communities negative too.
Gombe's chimpanzees are P.t. Schweinfurthii, not the P.t. Troglodytes that carry HIV-1's ancestor, so the disease findings may not transfer. One case suggests they do: CAM 155, a male P.t. Troglodytes caught in southern Cameroon in 2003 at 18 months and kept in a sanctuary, tested positive soon after capture and over nine years had recurrent infections, weight loss, a falling CD4 count and high viral load: an AIDS-like illness. Field studies in genuinely wild animals are still needed.
Timing the Jump
Dating HIV-1 looked simple and was not. The molecular clock gave estimates ranging from 1951 to 25 million years ago before anyone understood how badly hypervariable genes, recombination and within-host virus populations distort it. The relaxed clock, validated on the Thai epidemic and a thirteen-year Swedish transmission chain, plus a second fossil virus, DRC60, pushed the common ancestor of HIV-1 group M back to 1884-1924. Worobey then used five forgotten Miami blood samples to prove the virus reached Haiti before the US.
Why the date matters
Epidemiology had taken the story as far as central Africa; molecular work had narrowed it to Kinshasa, with ZR59 (1959) as the oldest genome. The next question, when the virus entered humans, belonged to evolutionary biologists, and it was not academic: knowing HIV-1's past trajectory would help predict its future one and inform vaccine strategy.
Viruses leave no fossils, but genomes like ZR59 are called fossil viruses because they serve the same function as rock. HIV-1's RNA mutates about a million times faster than the DNA of other species, so even a short residence in humans leaves a vast accumulated record.
Why the clock misfired
- 1988: one group dates the HIV-1/HIV-2 common ancestor to 1951; another argues all lentiviruses co-evolved with hosts whose ancestor lived 25 million years ago.
- 1988: Sharp's group, using 15 sequences from Africa, the Americas and Europe, dates group M's ancestor to about 1960, which matches ZR59 and holds for a decade.
- 1998: sequences from ZR59 place it near but not at the B/D ancestor, implying group M's ancestor fell between 1940 and the early 1950s; a second 1998 paper says 1942.
- 2000-01: three independent estimates agree on the 1930s.
- 2008: with two fossil viruses, 1884-1924, no later than the 1930s.
Early estimates were thin because few sequences existed and most had similar isolation dates, leaving little divergence to measure. The deeper problem was the assumption of a constant mutation rate.
- The rate varies along the genome. About 7 mutations per 1,000 nucleotides per year in the most variable part of env, about 3 per 1,000 in conserved parts of pol. env codes the protein that binds CD4, so antibodies drive relentless selection for escape mutants in a cat-and-mouse game; pol codes reverse transcriptase, where most changes kill the virus. The receptor-binding region of env differs by up to 40% between subtypes and up to 15% between viruses in one person.
- Repeat mutations hide. A nucleotide in a hypervariable region can mutate more than once; ignore that and the rate is underestimated, biasing deep dates towards the present.
- Recombination. Sharp spotted it in 1988 in an African isolate whose env clustered with African viruses while gag and pol clustered with non-African ones. It was ignored because dual infection was thought rare, until 1995, when his group found at least 10 recombinants among 100 sequences. Recombinants in Kinshasa and other cities proved dual infection common enough to change the virus far faster than point mutation alone.
- Dating the sample is not dating the infection. Primary HIV infection causes no distinctive illness, so isolates are dated by year of isolation, after the virus may have evolved in that host for years; or the sampled virus may have sat archived as an integrated provirus, not evolving at all.
- The rate changes with disease stage. It is low in the primary and terminal phases, when immune pressure is weak, and high through the long lag period. Viruses from rapid progressors accumulate fewer mutations than those from slow progressors, which again points to immune pressure as the engine.
THE BIOKO ISLAND CALIBRATION
Mike Worobey (University of Arizona) and colleagues at Tulane used an unusual natural experiment. Bioko Island lies 32 km off the west African coast and was cut off when sea levels rose about 12,000 years ago. SIVs from Bioko bushmeat each had a mainland counterpart in the same species, so the separation date could calibrate the clock.
| Method | Estimated age of all SIVs' common ancestor |
|---|---|
| Small RNA fragments | about 49,129 years |
| Amino acid sequences of viral proteins | about 76,800 years |
Both show SIVs are ancient, and the discrepancy makes the methodological point: the faster a sequence evolves, the more its clock reading is biased towards the present. Molecular clocks are reliable over hundreds of years; over thousands to millions, repeat mutations compress the apparent age. For old dates, amino acid sequences beat nucleotide sequences, because not every nucleotide change alters an amino acid.
VALIDATING THE RELAXED CLOCK
Computer models that compensate for these variables became the relaxed molecular clock, and each revision pushed group M's ancestor further back. Each was tested on real outbreaks, calibrating on ZR59 with its fixed 1959 date.
Thailand's population was 54 million; HIV-1 infections went from near zero in 1980 to 300,000 by 1990 and 800,000 by 1994. A 1985 screen of 600 high-risk people found three positives, and thousands of workers tested in 1986 for Middle East jobs were all negative, yet the virus was already there. It entered the resident gay community of Chiang Mai and Chiang Rai around 1984 through sex tourism, and the government underplayed the crisis to protect tourism. It hit Thailand's estimated half million female sex workers, then their clients, then everyone. The new variant, first called subtype E, proved to be a recombinant and was renamed CRF01. A separate introduction in 1988 carried subtype B through Bangkok's 100,000 injecting drug users. The models dated CRF01's common ancestor to 1986, close enough to the epidemiological date of 1984 to be encouraging.
In 1980 a Swedish man acquired HIV in Haiti. Of eight later partners, six women were infected; two of them infected male partners and two passed the virus to a child. Thirteen years separated the first and last transmission, which the authors note is equivalent to 13 million years of evolution in higher organisms. Interviews fixed each transmission to within months, and stored blood gave sequences that had diverged for up to 25 years. Genetic distance tracked time intervals exactly: the relaxed clock fits HIV-1. One detail matters: even at the branch point there was about 2% distance in the env hypervariable region, because each person carries a diverse virus population, so what is transmitted to two recipients is not identical to begin with.
DRC60: hunting a second fossil
With a validated clock the date landed in the 1930s, but only one pre-1976 fossil, ZR59, existed for calibration. Worobey, a Canadian evolutionary biologist who took up HIV at Oxford as a Rhodes Scholar in the late 1990s, went looking for another.
The obvious place was DRC, where others had failed, not because samples were never stored but because two civil wars in the late 1990s brought power cuts that destroyed every freezer collection. The surviving Congolese samples, ZR59 included, had been stored abroad.
Blood smears hold perhaps ten microlitres of blood dried on a slide at room temperature for decades. Refined PCR recovered human DNA from them but no HIV. Pathology blocks were richer: tissue fixed in formalin, embedded in wax, sliced and stained for diagnosis, then stored indefinitely at room temperature. Formalin degrades DNA, but modern protocols can recover sequences, and RNA is easier to reach than DNA. Crawford, a pathologist's daughter, vouches that pathologists never throw blocks away; she inherited a set from her father.
Jean-Jacques Muyembe of the University of Kinshasa supplied blocks from the late 1950s and early 1960s. One of hundreds gave an HIV-1 RNA signal: a lymph gland taken in 1960 from an adult woman in Leopoldville. Worobey sent material to an independent laboratory, which extracted the same ancient sequences, ruling out contamination. He named it DRC60.
DRC60 sits near the ancestral branch of subtype A; ZR59 near the root of B and D. They differ by about 11%, far more than is seen within a subtype. Two samples taken a year apart in the same city already belonged to separate, long-diverged subtypes, each representing decades of independent evolution in people. Calibrating on both, Worobey put group M's most recent common ancestor between 1884 and 1924, certainly no later than the 1930s.
Worobey noted in Nature how closely that window matches the founding of the first cities of west central Africa: Kinshasa 1881, Brazzaville 1883, Yaoundé 1889, Bangui 1899. Both fossils cluster with other DRC isolates but sit nearer the roots, while elsewhere subtypes cluster geographically (B in the US and Europe, C in southern Africa and India). That pattern is the signature of founder events: single viruses seeding each region.
Crawford's window of 1884-1924 was narrowed in 2014 by Faria, Rambaut, Suchard, Pybus and colleagues (Science), who used far more central African sequences to place group M's ancestor in Kinshasa around 1920 (roughly 1909-1930), and traced spread from there along the Belgian Congo rail and river network to Brazzaville, Lubumbashi and Mbuji-Mayi by the end of the 1930s and 1940s. Group M and the non-pandemic group O behaved similarly until about 1960, when group M entered an epidemiological transition and grew faster than the population. A second update: in 2019-20 Gryseels, Worobey, Vandamme and colleagues recovered a near-full-length genome, DRC66, from a 1966 Kinshasa tissue block, confirming and tightening the dating rather than overturning it.
Haiti before the US
In the early 1980s Haitians made up about 5% of early US AIDS cases and were briefly listed by the CDC as a high-risk group, giving the "4-H club". The label was withdrawn; the stigma stayed.
Haiti occupies the western third of Hispaniola, with the Dominican Republic on the rest. In the 1970s it was the poorest country in the western hemisphere: six million people, widespread malnutrition, crowding, no clean water, little medical care, TB, malaria, typhoid, diarrhoeal disease and STDs, and a life expectancy of 47 years. Those who could emigrate went to Miami, 600 miles away.
At Jackson Memorial Hospital, 17 men and 3 women aged 22 to 43, all arrived in the US after 1975, presented with full AIDS: Kaposi's sarcoma, Pneumocystis pneumonia, cryptococcal meningitis, thrush, TB, cytomegalovirus, cerebral toxoplasmosis and more. All denied homosexual contact and illicit drugs; none had prior immune deficiency; all had low CD4 counts; half died within the two-year study. Blood from some was frozen, stored, and forgotten for over twenty years.
The report set off speculation about voodoo blood-drinking, blood-letting, promiscuity, drug abuse, near-starvation diets, over-the-counter antibiotics and a recent swine flu epidemic in Haitian pigs. Haiti's tourist industry collapsed overnight. A Haitian and Cornell team then showed in the New England Journal of Medicine that AIDS patients in Port-au-Prince had the same risk behaviours as American patients: young men, homosexual or bisexual, recipients of blood products, clients of sex workers, or with past STDs. The finger-pointing continued anyway.
The direction of travel was genuinely open: the first authenticated cases in both countries fall in 1978-79. Haiti had become a destination for gay sex tourism from the US since the 1970s, which Haitian doctors argued had seeded the virus into Haiti. Early evolutionary work hinted the other way, with Haitian subtype B viruses dating slightly earlier than American ones, but there were too few to be sure.
In 2008 Worobey retrieved six of the forgotten Miami blood samples and recovered HIV-1 sequences from five: the oldest Haitian fossil viruses in existence, which he called "the next best thing to time travelling". Placed in a tree alongside isolates from the Americas, the Caribbean, Europe, Africa, Asia and Australia, they settled the question at once. Haitian isolates were more diverse than all other subtype B viruses in the world combined, and sat basal to all non-African subtype B, the same pattern group M shows in DRC. Hahn called it "a very nice piece of evolutionary sleuthing".
| Founder event | Dated range | Best estimate |
|---|---|---|
| Africa to Haiti (all non-African subtype B descends from it) | 1962-1970 | 1966 |
| Haiti to the US | 1966-1972 | 1969 |
| Haiti to Trinidad and Tobago | around 1973 | single event |
If HIV-1 reached the US in 1969, it circulated there for nine years before the first retrospectively positive blood sample and twelve years before anyone recognised AIDS. By 1978 about 5% of sexually active gay men in San Francisco and New York were infected, several thousand people. Crawford's explanation for the silence: slow early spread, then acceleration once the virus reached the highest-risk networks, and a ten-year lag before the symptoms that would make doctors notice.
The Trinidad and Tobago finding overturned the assumption that its heterosexual epidemic began in the late 1970s or early 1980s through contact with American gay men.
A founder event seeds a population from one or very few individuals, producing a bottleneck: a sudden loss of genetic diversity, as when antibiotics wipe out all but a few resistant bacteria. Several HIV-1 viruses reached the US from Haiti and at least one reached Brazil, but all except one US and one Trinidadian lineage were dead ends. Global spread has been punctuated by such bottlenecks, which is why the subtypes exist. Searches for evidence that the successful lineages were fitter have found none, leaving the question of why this virus travelled so well to later chapters.
Worobey's team sequenced full genomes from 1970s US serum samples in 2016 (Nature), dating the arrival of subtype B in New York around 1970-71 and in San Francisco about 1976, consistent with the 1969 estimate here. The same work cleared Gaëtan Dugas, the "patient zero" of the Introduction, whose virus was typical of strains already circulating.
Vital First Steps
How did the virus actually get into a person? The front runner is the cut-hunter theory: blood contact while killing and butchering a chimpanzee. The serious rival was the oral polio vaccine (OPV) theory, that contaminated CHAT vaccine sprayed into a million African mouths in 1957-60 seeded the pandemic. Crawford reconstructs the whole dispute, the Rolling Stone lawsuit, Hooper's The River, the Royal Society meeting of 2000 and Bill Hamilton's death, and explains the four findings that finished the OPV theory.
What has to be explained
By 2000 the consensus was that the ancestors of HIV-1 groups M, N and O came from P. T. Troglodytes chimpanzees, which live only in west central Africa, so the transfer happened there. The exact event will never be observed, but it was not unique: each group jumped independently, and HIV-2 crossed from sooty mangabeys at least eight times. Twelve known crossings are probably the survivors of many more that infected someone and then died out unrecorded.
Inside their natural hosts, SIVs circulate in blood and concentrate in blood-rich organs: gut, lymph glands, spleen, brain. They appear at much lower levels in semen, breast milk and possibly saliva. Since mucosal routes transmit far less efficiently than blood, cross-species transfer is assumed to have come from blood contact.
Hunting is the human activity that guarantees regular contact with animal blood, and hominids have hunted in Africa for millions of years; modern humans appeared there around 100,000 years ago. Bushmeat remains a lifeline for thousands of villagers in the chimpanzee range. Crawford's colleagues who have joined hunts testify to how bloody it is to kill, skin and butcher an animal as large and aggressive as a chimpanzee: bites and scratches from trapped animals, cuts from spears and knives are routine. Nobody doubts the exposure is sufficient. What nobody has is proof that this is what happened around 1900.
Without direct proof, the question becomes which hypothesis is most probable, and which can be tested.
THE MALARIA EXPERIMENTS
A 1990 letter to Nature suggested a different spark. In 1922 two European doctors in Freetown, Sierra Leone, injected themselves with fresh blood from a malaria-infected chimpanzee. The medical literature records 34 people injected with blood from chimpanzees, mangabeys and other primates, for two reasons: to learn whether primate malaria was the same parasite as the human disease, and to find a better treatment for syphilis, then incurable and sometimes treated by inducing a malarial fever. Such work continued into the 1960s, mostly in the US, and since both chimpanzees and mangabeys were used it offered a single explanation for HIV-1 and HIV-2.
An American doctor who had run such experiments on prisoners in the Atlanta penitentiary in 1962 replied that the local ethics board never permitted direct injection of primate blood: transfer was attempted through parasite-carrying mosquitoes. No stored samples survived, and he dismissed the proposal on the grounds that AIDS cannot be transmitted by mosquitoes or other arthropods.
The polio vaccine theory
- Tom Curtis: investigative journalist; published the theory in Rolling Stone, 1992, after being alerted by AIDS treatment activist Blaine Elswood.
- Hilary Koprowski: director of the Wistar Institute, Philadelphia; developed CHAT, the first oral polio vaccine used in mass trials; sued Curtis for defamation.
- Edward Hooper: British journalist; interviewed over 600 people, published The River (1999), 1,000+ pages, nearly 2,500 footnotes.
- W. D. (Bill) Hamilton: Royal Society Professor of Evolutionary Biology, Oxford; not a believer but insisted the theory be examined; wrote The River's foreword and pushed for the Royal Society meeting.
- Paul Sharp, Kevin De Cock, Robin Weiss, Andrew Rambaut, Mike Worobey: the scientists whose evidence closed the case.
WHY THE THEORY WAS PLAUSIBLE
Vaccines come inactivated (virus killed with formaldehyde) or attenuated (weakened live virus). The Salk inactivated vaccine (IPV) reached the West in 1955 and cut paralytic polio sharply, but it required injection and boosters, so it was judged unsuitable for the developing world. The race was on for an oral vaccine.
Polio virus had to be grown in living cells. Human cells were avoided in the 1950s over fears of cancer-causing genes, so monkey kidney cells were used, mostly from Asian rhesus and cynomolgus monkeys. Monkey viruses were barely understood. In the 1930s Albert Sabin, who eventually produced the licensed oral vaccine, isolated Herpes B virus from the brain of a colleague who was bitten by a laboratory monkey and died; the virus causes cold-sore lesions in monkeys and is often lethal in humans. A test existed, so the donor animals were Herpes B negative. The open question was what else was in the culture.
Simian virus 40, the fortieth virus isolated from monkeys, was found in rhesus monkeys in the 1960s. Retrospective testing found infectious SV40 in up to 30% of batches of both inactivated and oral polio vaccine: the formaldehyde that killed polio virus did not kill it. Between 1955 and 1963 roughly 90 million Americans, and countless more worldwide, received it. Recipients made antibodies, so their cells were infected, and SV40 caused tumours in experimental animals. No firm evidence of harm in people has emerged, though reports surface periodically, so the case is not closed.
SV40 proved that unknown monkey viruses could ride along. Nobody in the 1950s knew SIVs existed, and a latent virus can infect cells for life with no outward sign. After the SV40 scare and an Indian ban on rhesus monkey exports, producers switched to African green monkey kidneys, and Curtis named SIVagm as HIV's possible ancestor, which later proved to be far too distant a relative.
- CHAT was first tested in a small number of institutionalised children and infants of women prisoners in the US, then in several European countries.
- February to April 1957: 3,000 to 10,000 people a day queued in the Ruzizi Valley (modern DRC, Rwanda, Burundi) after village chiefs directed them to rallying points. 215,504 people vaccinated there.
- The 1958 British Medical Journal report: 244,596 inhabitants of the Belgian Congo and Ruanda-Urundi given the vaccine by oral spray.
- Further campaigns as a polio epidemic swept the region, including Leopoldville; by the end, about one million people, mostly young children.
Curtis's argument was geographical and chronological: the vaccination zone overlapped the areas with the highest HIV-1 levels and earliest AIDS cases twenty years later, and the campaign dates fitted the then-current estimate of 1960 for group M's common ancestor. Koprowski sued; the case settled out of court with a clarification note, damages of $1, and legal fees of half a million dollars.
Hooper took up the case, documented unpublished laboratory protocols and unconventional practices, and, once SIVagm proved too distant, proposed that the vaccine virus had been grown in chimpanzee kidney cells, making SIVcpz the contaminant. By the end he believed that scientists, editors and doctors had colluded to prevent exposure of a medical accident. Hamilton's foreword to The River called the closing of ranks a block on discussing "an accident that is bidding to prove itself more expensive in lives" than the atrocities of Hitler, Stalin and Pol Pot.
Leave it alone Those involved in production had acted in good faith to save lives. Others argued that nothing could reverse the past, and that the theory, if aired, would destroy public faith in vaccination.
Investigate it Most thought the facts mattered precisely so that a repeat could be prevented, and that the only way to answer the charge was to test the vaccine and the chimpanzees.
The Wistar Institute expert committee, set up at Curtis's own suggestion, reported in 1992 that the probability of the epidemic starting with the 1957 trials was extremely low, since "almost every step in this hypothetical mode of transmission is problematic". But nobody could state categorically that early CHAT batches were clean, and Crawford lists the chain the theory needed, every link of which was possible:
The committee cited two counterpoints. The Manchester sailor had AIDS-like illness in 1958, before the trials, though by 1995 that case was shown to be unrelated to HIV. And ZR59, taken in Leopoldville in 1959, could in principle have come from a vaccinee, but the evolutionary distance between any known SIV and HIV-1 could not be crossed in two years.
Hamilton's last journey
Hamilton was struck by the geographic coincidence and, without believing the theory, thought it deserved a hearing. He persuaded the Royal Society to hold a discussion meeting and set two lines of investigation: find out whether chimpanzee kidneys were ever used to make CHAT and whether those animals carried an HIV-1-like SIV, and test stored CHAT batches for SIVs and for host-cell DNA to identify the species used.
Koprowski maintained that only rhesus kidneys from India and the Philippines were used, but could produce few records; kidneys were sometimes removed before shipment, so laboratory workers did not always know the species. Hooper proposed that when Asian supply was threatened, African primates, possibly chimpanzees, were used instead.
A research station in the rainforest near Stanleyville (now Kisangani) in the northeast Belgian Congo, opened June 1956 specifically for polio vaccine testing ahead of the mass trials. It housed chimpanzees and bonobos caught by local hunters in the Parisi Forest. Animals were given intraspinal injections of vaccine to prove it harmless, or oral vaccine then challenged with wild polio virus to prove it protective. By February 1958, 416 chimpanzees had passed through Lindi; by the end of the trials most had been killed and about fifty remained. Hooper's question was what happened to the organs of the dead animals. Records were missing or incomplete, and he became convinced chimpanzee kidneys were sometimes used in the final stages of CHAT preparation.
Hamilton visited Kisangani in 1999 and got only pet chimpanzee samples, none of which survived the journey. In early 2000 he returned with Mike Worobey, then at Oxford and sceptical of the theory but keen to get wild DRC chimpanzee viruses for evolutionary work. They collected faeces and urine from wild P. T. Schweinfurthii that later yielded SIVs.
- Two days into the forest Worobey impales his hand on a spiny branch, leaving an inch-long splinter in his thumb; with no antibiotics the wound blackens and he becomes seriously ill, but walks out carrying the first samples. Médecins Sans Frontières removes the splinter and treats him.
- Back in Kisangani, Hamilton wakes with the fever, sweats and chills of malaria; they make a dash for the first flight out.
- Kampala: antimalarial treatment; then Nairobi and London, with the parasite apparently cleared by the time they land.
- The next day Hamilton collapses. He dies three weeks later, aged 63, in March 2000.
Crawford knew Hamilton's work through Dawkins's The Selfish Gene. The obituaries describe "a gentle giant", "a solitary scholar", "a one-off", with "a radical originality" and "a disdain for authority".
The Royal Society meeting, September 2000
The meeting Hamilton had instigated went ahead after a postponement caused by rows over the speaker list, with virologists, vaccine experts, epidemiologists, molecular and evolutionary biologists and the theory's proponents. Worobey recalls "a uniquely adversarial tone".
- The date had moved. By 2000 the consensus for group M's most recent common ancestor was 1915-1941, best estimate 1931, decades too early for a 1957 vaccine. Hooper's counter, without evidence, was that the subtypes had diversified inside chimpanzees and were each delivered to humans by the vaccine, which would also explain the starburst.
- Wrong chimpanzees. Camp Lindi lies in the range of P. T. Schweinfurthii, not the P. T. Troglodytes that carry the ancestral virus. Vaccine producers denied chimpanzee kidneys were ever used, and Lindi had no tissue culture facility capable of growing polio virus.
- The vaccine tested clean. Stored CHAT batches, tested at the Wistar Institute and independent laboratories, were negative for primate lentivirus sequences; where host DNA was detectable it came from Asian macaques, not chimpanzees.
- The geographic overlap is not causal. Kevin De Cock (CDC, Nairobi) called the association "ecologic": proof would need HIV incidence to be higher in vaccinees than non-vaccinees, and no records of the 1950s vaccinees exist. Hooper's 29 early AIDS cases from the literature cannot measure incidence in a country the size of western Europe that had about 400 hospitals and 1,200 doctors by the mid-1980s, where only academically inclined physicians ever wrote cases up. Many of the 29 were in towns along the Congo River, a 2,700 km thoroughfare whose social mix would spread HIV anyway.
Robin Weiss (UCL) summed up: the theory had once been plausible, but the weight of evidence now said it was not how HIV-1 entered humans, while admitting such an accident could happen, and concluding that exactly how, when and where the first person was infected is "likely, however, to remain a matter of conjecture".
Hooper and a small group continue to argue the case online, alleging a conspiracy. De Cock answered that with Peter Medawar: the intensity of conviction that a hypothesis is true has no bearing on whether it is true.
Four later findings
| Year | Finding | Effect on the OPV theory |
|---|---|---|
| 2001 | Actual CHAT pool 13 used in the DRC campaign located and tested: Old World monkey DNA, no SIV, no chimpanzee DNA | Removes the physical evidence |
| 2001 | Rambaut (Oxford) shows the starburst formed after the jump to humans | Removes the theory's best explanatory claim |
| 2004 | Worobey compares SIVs from the Parisi Forest near Camp Lindi with Gombe P. T. Schweinfurthii viruses: they cluster together, distinct from HIV-1 and from P. T. Troglodytes viruses | Lindi chimpanzees could not be the source |
| 2006-08 | Hahn, Sharp, Peeters locate the group M ancestor in southeast Cameroon; Worobey's DRC60 pushes the jump to before the 1930s | Wrong place and wrong time |
Rambaut's argument deserves spelling out. Among several hundred HIV-1 group M sequences from DRC, the recognised subtypes are lost in a background of diversity, unlike everywhere else, where a few subtypes dominate a region. The starburst was created in the 1940s or 1950s by individual viruses leaving Kinshasa to found epidemics elsewhere, each becoming a subtype through a founder effect, exactly as the single lineage that went to Haiti around 1966 and the US about three years later became subtype B. In Kinshasa itself there is no starburst: by the late 1950s, when CHAT was made, several hundred strains already existed in DRC, and the theory would require every one of them to have come through the vaccine. His conclusion in Nature: the claim that the OPV theory is the only hypothesis that explains the starburst is incorrect.
She backs the cut-hunter theory along with every scientist she knows, but prefers the name "natural transmission" theory, since the jump could equally have happened at any stage of bushmeat preparation or through a bite from a pet chimpanzee. Even so, three questions remain: why did HIV/AIDS surface in Leopoldville when the source chimpanzees live in Cameroon; why, if hunting is ancient, did the HIVs appear only in the late 19th and 20th centuries, and then cross at least twelve times; and why did group M alone go global while N, O and P stayed in Cameroon?
Crawford's position has held. Later work, including the 2014 dating of group M to Kinshasa around 1920 and surveys showing 1-2% of bushmeat hunters in Cameroon carrying simian foamy virus or antibodies to monkey retroviruses, supports the picture of repeated, ongoing cross-species transmission at the hunting interface. What none of it supplies is the one thing Crawford says is missing: direct evidence of the founding event itself, which almost certainly left no record at all.
The Epic Journey Begins
Twelve HIVs crossed into humans in a single century, which means the question is not how a virus gets from a chimpanzee into a hunter, but why it suddenly stopped dying out. Crawford assembles the answer from bushmeat surveys, Marx's unsterile-needle theory, the hepatitis C disaster in Egypt, the Sangha and Congo rivers, and above all the demography of colonial Leopoldville: a boom town with four men to every woman, widespread genital ulcer disease and a window of opportunity that opened around 1919 and closed by the 1950s.
Twelve jumps, one century
The chimpanzees carrying ancestral group M live in southeast Cameroon, but the eight HIV-2 groups came from sooty mangabeys over 2,000 km away in West Africa. Any theory of emergence has to explain twelve crossings in the 20th century, not one.
Everything found so far is native to west or west central Africa, which raises the question of what was special about that region at that time. The usual candidates are the social changes of colonialism and the First World War: guns for hunting large game, logging concessions opening inaccessible forest, and forced labour that pulled men from their villages so they hunted bushmeat instead of farming. None can be proved.
HOW OFTEN DO MONKEY VIRUSES REACH PEOPLE NOW?
Simian foamy virus is the test case: named for the frothy appearance of infected cells, it spreads between juvenile animals in saliva, harms neither its natural hosts nor humans, is ubiquitous in captive African and Asian primates, and is known to cross into people, so it works as a marker for the potential of the less transmissible SIVs.
- 4 of 200+ North American animal handlers tested carried simian foamy virus; all four had been bitten or injured by a captive primate or a contaminated instrument. Sequences identified the sources as one African green monkey and three baboons.
- 10 of over 1,000 people from nine Cameroonian villages who handled fresh primate blood were infected; three sequenced viruses came from a gorilla, a mandrill and a De Brazza's guenon, all local species.
- Over 500 bushmeat carcasses from Cameroon markets tested by Peeters: SIVs in the majority of primate species, prevalence 5-40% per species.
- 1 of over 6,500 HIV-positive samples from Cameroon and Gabon reacted more strongly with mandrill SIV proteins than with HIV; the donor was a healthy 65-year-old man.
- 17% of 76 samples from people who hunt, butcher or keep primates had SIV-reactive antibodies, one strongly matching the eastern black and white colobus virus.
- In both of the last two studies no viral sequences could be amplified, so neither infection is confirmed.
Primate viruses reach African hunters regularly; what almost never happens is that one of them goes anywhere afterwards.
Dead ends and chains
- No infection at all. The virus fails to establish.
- Transient infection. The host's immune response eliminates it.
- Persistent infection, no onward spread. A dead end: the virus dies with the host.
- Persistent infection with transmission. A small focus of infected people, which still dies out unless, on average, each infected person infects more than one other.
Flu and measles travel on air currents; the HIVs could only go where their hosts carried them, by blood or sex. A cut hunter in a traditional village in rural Cameroon had very limited opportunity to pass anything on, so almost every jump died out unrecorded. Crawford notes that groups N and P are in exactly that category today, despite having been found in people.
WHAT ABOUT THE SLAVE TRADE?
From the 15th century Europeans traded gold, ivory and slaves along the west African coast; Columbus's 1492 voyage turned slaving into a 300-year business feeding the sugar plantations. By 1867 about 11 million people had been shipped to the Americas, with countless millions dying before embarkation.
| Microbe | How it crossed the Atlantic |
|---|---|
| Malaria, river blindness, schistosomiasis, elephantiasis | Carried in enslaved people; established in the Americas |
| Yellow fever virus | Serial infection aboard ship, spread by mosquitoes breeding in drinking-water casks |
| Hepatitis B | Persistent infection with a long silent period: travels inside the host |
| HTLV-I | Like HIV, persists in CD4 T cells, spreads by blood and sex, causes leukaemia decades later or nothing at all; probably jumped from simians about 20,000 years ago; slaves carried it to the Caribbean, where the focus remains |
About 800,000 slaves were shipped from the area now covering Cameroon, Gabon and DRC, the range of the ancestral-HIV chimpanzees, and the real figure may be far higher. If an HIV precursor had existed then, it could have crossed and thrived. There is no evidence that it did, which supports the conclusion that SIV/HIV infection in the local human population was rare or absent before the early 20th century. Something changed around 1900.
Could the virus adapt on its own?
A virus entering a new host usually goes through rapid adaptation, improving its fitness by mutation and natural selection. HIV mutates fast, so adaptation can be fast, but since nobody knows how many mutations turned SIVcpz-Ptt into HIV-1, the time required is unknown.
Researchers built SHIV, an SIVmac genome with its env gene replaced by HIV-1's, so that cell binding, infection and immune escape all behaved like HIV-1. SHIV causes no symptoms in pig-tailed macaques, and the first pair stayed healthy. Infected cells were passed to a second pair, then a third, through five pairs. In groups one and two the virus grew poorly in CD4 T cells and caused no disease. In groups three, four and five it grew well: high virus loads, low CD4 counts, and three of six animals with clear AIDS within a year. Serial passage selected a fast-growing, CD4-destroying strain.
MARX'S OBJECTION AND THE NEEDLE THEORY
Preston Marx (Tulane) and colleagues argue that a poorly replicating SIV would be cleared by the human immune response before it could adapt. Their computer model estimated the time needed for SIVcpz to become HIV-1 as 65 days if 100 mutations are required and 80 days if 200 are, while the human immune response peaks in about ten days. Their conclusion: the emergence of even one epidemic HIV strain after a single exposure was very unlikely, and four or more such transitions (groups M and O, HIV-2 A and B) in a brief period is "vanishingly small".
Oral polio vaccine theory (chapter 6) SIVcpz contaminated the vaccine itself and was delivered to humans in the oral spray. Rejected on evidence.
Marx's serial passage theory Accepts the cut-hunter jump as the origin. Argues the virus could not persist in that first host long enough to adapt and spread sexually, so it had to be picked up from the bloodstream during the acute, high-viraemia phase and passed on by unsterile injection equipment, staying ahead of the immune response.
Needles unquestionably transmit HIV efficiently, as the speed of spread among injecting drug users in the US and Europe in the early 1980s showed. The question is whether the history supports it.
- 1848 Hypodermic syringe invented: handmade glass and metal, costly, sterilised in boiling water and reused.
- 1920s Global production rises from 100,000 to over 2 million a year.
- 1917-19 A sleeping-sickness campaign in the Central African Republic screens about 90,000 people and treats over 5,000 cases reportedly using six syringes.
- Late 1940s Penicillin, injectable only, drives demand.
- Early 1950s UN yaws eradication campaign treats around 35 million children.
- 1950 7.5 million syringes a year; disposables arrive and the price falls to US 1.5 cents.
- 1960 Global production reaches one billion.
From the early 20th century, Belgian, French, British, German, Portuguese and Spanish authorities treated hundreds of thousands of west central Africans with injectable drugs for malaria, syphilis, yaws, tuberculosis and leprosy.
The campaign against schistosomiasis, a sometimes fatal blood-fluke disease spread by water snails in the Nile's irrigation channels, ran from 1920 to 1980. Treatment was tartar emetic, given as 12 to 16 intravenous injections; at its height, 1966-69, about three million injections a year went into 250,000 patients. A 1964 WHO report describes a doctor completing 504 injections between 9.20 and 10.10 a.m., under five seconds each, with 20 to 30 syringes in rotation, each washed and boiled for a minute or two. The consequence surfaced in the 1990s: about one in five Egyptians carries hepatitis C. Rates are 28% in the Nile Delta farming communities against 6-8% in Cairo and Alexandria, where schistosomiasis was rare, and reach 55% in the oldest Delta age group, the people who were young adults at the campaign's peak. One or two minutes of boiling does not reliably inactivate HIV either.
Direct data on HIV are impossible to collect, since anyone infected during those campaigns died of AIDS long ago. Two studies used long-survivable blood-borne viruses as proxies, both in rural areas near the range of the SIVcpz-Ptt chimpanzees:
- Southern Cameroon, hepatitis C. Prevalence second only to Egypt's: 56% of participants positive, risk highest in older people and those who had received intravenous treatment for malaria. Molecular clock analysis showed exponential growth coinciding with the treatment programme in the first half of the century.
- Central African Republic, HTLV-I. Villages near the southeast Cameroon border where sleeping sickness was rife 1936-50. Infection was linked to sleeping-sickness treatment or prevention. Colonial records predicted that about 60% of participants aged 65 or over should have been treated; only 11% reported it. The authors suggest the gap exists because many of the treated acquired HIV-1 along with the drugs and died early of AIDS.
Both studies show blood-borne viruses were spread by unsterile needles in the right place at the right time. Neither proves that this step was necessary for HIV-1's adaptation, and Crawford accepts it can no longer be proved or disproved.
From the Sangha to the Pool
Group M's ancestor dates to 1884-1924, so the person who carried it to Kinshasa was infected in that window, almost certainly not the original hunter. Sharp, Hahn and Worobey all stress the coincidence between the virus's emergence and the rise of large colonial cities.
- 1884 Cameroon comes under German control under Bismarck; road and rail built, and the Atlantic port of Douala developed to exploit ivory, rubber, timber, coffee and cocoa.
- 1914 Britain invades from Nigeria, France from French Equatorial Africa; the Germans retreat.
- After the war The country is divided, France taking the larger share including southeast Cameroon, where group M was born.
- 1916 Yaoundé becomes capital for its central position; Douala stays the commercial centre but remains small and attracts few migrants.
In the remote, forested southeast, the main thoroughfare was the Sangha River, a tributary of the Congo. The Congo basin covers over four million km²; the river runs 4,700 km from the East African Rift to the Gulf of Guinea. The Stanley Falls make the river impassable to boats between the coast and the interior, but above them lies the Malebo Pool, with Kinshasa on the south bank and Brazzaville on the north, and from there river traffic penetrates forest that is otherwise impassable.
Whether the virus made the journey in one leap inside a traveller or in short hops between fishing communities and river traders, it reached Leopoldville in the early 1900s.
A SHORT HISTORY OF A BRUTAL COLONY
- 15th century Portuguese reach the Congo estuary and trade amicably with the Kongo people, who give the country its name.
- 16th century European slave traders plunder the area.
- 1871 Henry Morton Stanley, backed by the New York Herald, finds David Livingstone near Lake Tanganyika; together they rule the lake out as the Nile's source.
- 1874-77 Stanley follows the Lualaba and then the Congo for 999 days with 356 men; over two-thirds desert or die, and the survivors reach Boma close to death.
- 1885-1908 King Leopold II holds the Congo Free State as personal property, extracting rubber, ivory and wood with forced labour; the population falls sharply through killing, disease and flight. Joseph Conrad, sent out to replace a murdered steamer captain, writes Heart of Darkness (1902) from what he saw.
- 1908 International outrage dissolves the Free State; the colony passes to Belgium. Conditions improve, but racism and extraction continue. By 1960 reportedly only 17 of 14 million Congolese had a university education, none of them doctors.
- 1881 Stanley founds Leopoldville at a fishing village on the Pool. Goods are head-carried 250 km around the rapids to Matadi until the railway opens in 1898; the town becomes the commercial hub and, in 1920, the capital.
Why Leopoldville rather than Brazzaville is partly chance; Brazzaville is the nearer bank for someone coming from Cameroon, and modern Brazzaville isolates are just as diverse as Kinshasa's, though no fossil viruses have been found there. Constant traffic across the Pool would have moved the virus back and forth anyway. But Brazzaville was the administrative capital of a resource-poor country, while Leopoldville was the capital of one rich in diamonds, copper, cobalt, tin, zinc, manganese and uranium, with rubber and coffee grown at scale: a boom town.
- 1900: about 10,000 people, twice Brazzaville's.
- 1920: 16,000. 1929: 47,000, swollen by migrant workers coerced into industry and building.
- A dip with the 1930s world recession, then 200,000 by 1950 and about 400,000 by the 1960s.
Why the city made the pandemic
The decisive factor, in Crawford's account, was Belgian labour policy: men were conscripted into the workforce while their wives and families were discouraged from leaving the villages. In the late 1920s men outnumbered women in Leopoldville by over 4 to 1.
A gender ratio of that kind produces commercial sex and extremely high levels of sexually transmitted disease, and the same conditions that let syphilis and chancroid thrive let a new, silent STD thrive too. Genital ulcer disease multiplies HIV transmission: early African studies found ulcers in female sex workers raised male-to-female transmission 10-50 times per act and female-to-male 50-300 times. Experts concluded that over half of new African infections were assisted by genital ulcer disease.
Male circumcision cuts the risk sharply. Among men who caught genital ulcer disease from sex workers, HIV transmission per act was 43% in uncircumcised men against 4% in circumcised men.
A University of Leuven group reconstructed the conditions from colonial health reports and medical journals for Leopoldville, Brazzaville and Douala:
| Factor | Early 20th century Leopoldville |
|---|---|
| Ulcer-causing infections | Syphilis (infectious ulcers about 5 months) and chancroid (about 10 weeks), both probably introduced by Europeans in the late 1800s |
| Sex work | A 1928 survey estimated nearly half of the city's 6,000 women were "mainly living on prostitution", a colonial category covering full-time sex workers and femmes libres with a few regular partners |
| Ulcer prevalence | Early 1930s surveys: 5% of those women had active genital ulcer disease; syphilis prevalence 10% |
| Circumcision | About 70-80% of men in 1910, rising to over 95% by 1960 |
| Trend | Treatment campaigns and screening cut STDs from the mid-1930s until independence |
Their simulation of HIV-1 group M's emergence concluded that the virus could not have survived in a pre-colonial village, for want of commercial sex and STDs, and that its best chance of sustaining a chain of infection in Leopoldville fell between 1919 and 1929, precisely the period predicted for its arrival. The window then closed as health care improved: a virus arriving in 1950 would probably have died out.
Either the virus survived under its own steam in a city whose demography happened to suit it, or it needed an initial boost from contaminated needles to adapt. Crawford's view: we will probably never know, and the outcome is the same. A chain of infection held long enough to reach Leopoldville, where social change turned it into an epidemic.
Chaos, Haiti, and the closing of the circle
- 1960 Patrice Lumumba, a former post office clerk, becomes prime minister of the new DRC with almost no preparation for the handover.
- Within months The army mutinies, riots spread, most remaining Belgians flee, the government collapses. Joseph Mobutu takes charge and rules corruptly for thirty years.
- January 1961 Lumumba, imprisoned, escaped and recaptured, is executed six months after independence. Civil war follows; thousands of refugees reach Kinshasa, unemployment hits 25%, poverty is rife, brothels flourish. Crawford places the start of epidemic growth here.
- Mid-1960s With almost no Congolese teachers or doctors, the UN and WHO recruit abroad; several thousand Haitian teachers, doctors, nurses and technicians work in DRC.
- About 1966 One strain, in one person, most likely a returning Haitian professional, seeds Haiti.
- About 1969 A single strain moves on to the US, becoming subtype B.
Having worked through local sexual networks, the virus followed trade routes across central Africa, then left the continent. That closes the circle begun with the Los Angeles reports of 1981.
Crawford's reconstruction is impressionistic where the 2014 Faria analysis is quantitative. That study placed the ancestor in Kinshasa about 1920 and showed group M radiating along the Belgian Congo's rail and river network, reaching Brazzaville, Lubumbashi and Mbuji-Mayi by the 1930s and 1940s, carried by the over one million people a year moving through Kinshasa's rail hub by the late 1940s. It also found that group M and the non-pandemic group O grew at similar rates until about 1960, after which group M alone accelerated, which puts more weight on the post-independence upheaval that Crawford describes and less on any intrinsic superiority of group M.
Adapting to Humans
Why did group M infect tens of millions while O, N and P stalled at 25,000 and fewer than 20 people? The molecular answer is a contest of host restriction factors against accessory genes. Crawford walks through the Gag-30 mutation, the Red Queen arms race, Nef and T cell activation, TRIM5α, and the finding that matters most: pandemic HIV-1 alone learned to disable human tetherin by switching the job from Nef to Vpu.
What has to be explained
Two SIVs out of dozens have colonised humans, despite constant contact with the rest, and the four HIV-1 groups behave completely differently in the same species.
| Virus | People infected (2013) | Death rate versus uninfected |
|---|---|---|
| HIV-1 group M | 60-80 million | Up to 60 times |
| HIV-1 group O | about 25,000 | Pathogenic, local to Cameroon and Gabon |
| HIV-1 groups N and P | fewer than 20 | Barely established |
| HIV-2 | West Africa only | 2-5 times |
| SIVcpz in chimpanzees | - | 10-16 times |
| SIVsmm, SIVagm in natural hosts | - | No effect on life expectancy |
The viruses newest to their hosts are the most damaging; the ancient ones, SIVsmm and SIVagm, have co-evolved to the point of harmlessness.
DOES THE VIRUS GET MILDER WITH TIME?
Each transmission is a severe bottleneck: often a single virus particle establishes the new infection, which is what produced the founder effect and the subtypes, now 10-35% apart. Whether any of those differences represent adaptation to humans is unclear, and there is no firm evidence that group M's virulence has changed in thirty years. No cohort large enough was set up in the 1980s, and such a study is impossible now because HAART interrupts the natural course.
The expectation that HIV will mellow rests mostly on myxomatosis: mild in Brazilian rabbits, it killed 99.8% of infected British and Australian rabbits on release, only about a quarter seven years later, and eventually most survived. Crawford warns against generalising.
Towards mildness A less virulent strain keeps its host alive through the long lag period, giving more chances to transmit.
Towards virulence A strain producing a higher viral load in primary infection spreads better per unit time, even if the host dies sooner.
Subtype C is the test case: it accounts for over 50% of infections worldwide and has displaced resident subtypes in places, which some read as greater transmissibility and others as the luck of being seeded into the huge populations of South Africa and Asia. Laboratory work is contradictory: some finds subtype C replicates less efficiently (less virulent, equally transmissible), some finds it produces higher viral loads (more virulent).
One mutation: Gag-30
In 2007 Sharp and colleagues compared the reconstructed ancestors of HIV-1 groups M, N and O with twelve SIVcpz-Ptt genomes, hunting for changes that differ between chimpanzee and human viruses but are conserved within each species. They found exactly one, at site Gag-30.
SIVcpz codes methionine (Met), a neutral sulphur-containing amino acid; all three HIV-1 ancestors code arginine (Arg), a basic one. A non-conservative replacement like this changes the protein's overall charge, and so potentially its shape and function. The familiar parallel: sickle cell anaemia arises from one such swap at position six of β-globin, bending red cells into sickles. All three HIV-1 groups made the same switch independently, which rules out chance and implies host-specific adaptation.
The confirmation came from two chimpanzees experimentally infected with HIV-1 in the 1990s during the search for an AIDS model. Both recovered viruses had reverted from Arg back to Met, resembling SIVcpz. In culture, the Met viruses grew much better in chimpanzee cells than human cells; mutating Gag-30 back to Arg reversed that. Why is unknown: the region codes the matrix protein that coats the virus inside its capsid and interacts with several host proteins during assembly of new particles, so disrupting one of those interactions after a host jump could plausibly change fitness.
A loose end: subtype C alone among group M subtypes carries Met at Gag-30, so its founder virus reverted. Whether that bears on subtype C's disputed virulence is unknown.
The Red Queen's arms race
The human genome's 8% retroviral remnants, revealed when sequencing finished in 2003, outweigh all protein-coding genes put together: evidence that our ancestors endured retroviral infections and evolved countermeasures. Evolutionary biologists borrow Lewis Carroll's Red Queen: both sides must keep running to stay in the same place.
Molecules produced by cells to block virus entry or replication, acting long before antibodies and T cells can be mustered. They attack every stage of the life cycle, from entry to the release of new particles, by interlocking mechanisms. One expert likens them to HAART, which hits the same life cycle at several points. The APOBEC family ("apolipoprotein B editing complex catalytic subunits") are packaged into new virus particles as they assemble, travel to the next cell, and there both suppress reverse transcription and mutate the viral genome so it cannot code working proteins.
To see how deep the arms race runs, researchers tracked 140 genes implicated in HIV-1 infection plus 100 random controls across five primates (human, chimpanzee, orang-utan, rhesus monkey, common marmoset) diverging for about 40 million years. Thirteen genes, several of them known restriction factors, showed unexpectedly high mutation rates in the human lineage: positive selection driven by a succession of lethal retroviruses. We inherited partial protection from what killed our ancestors. Most incoming SIVs are stopped by that inheritance; the dangerous ones are those that have learned to sidestep it.
THE VIRAL ANSWER: ACCESSORY GENES
Beyond gag, pol and env, all HIVs and most primate lentiviruses carry three or four accessory genes: vif (virion infectivity factor), vpr (viral protein R), vpu (viral protein U) and nef (negative factor). They are dispensable in the laboratory but decide the outcome of real infection and transmission, chiefly by neutralising host defences. Their conservation across millions of years marks them as survival kit, and they are especially important during a host jump, when the new cell is hostile.
Nef, T cell activation and why some infections are harmless
In 1995 researchers described seven HIV-positive people infected through blood or blood products from one donor. All eight, donor included, were long-term non-progressors: well 10 to 14 years on, with low or undetectable viral loads and normal CD4 counts. The only thing they shared was the virus, and the virus had no nef gene and no other significant abnormality. Similar nef deletions have turned up in other long-term survivors.
The mechanism of AIDS is chronic CD4 T cell loss. After primary infection, large numbers of activated CD4 T cells appear, some infected, some part of the response to the virus, and all activated T cells die quickly, with estimates reaching a billion a day. The cycle continues through the silent phase; production compensates until it cannot, and then counts fall and AIDS follows, on average about ten years after infection. The best clinical predictors of progression are viral load and the proportion of activated CD4 T cells.
The surprise is what non-pathogenic infections look like. SIVsmm and SIVagm also produce high virus levels and high T cell activation at first, and the viral load stays high. What falls is the number of activated CD4 T cells.
A non-pathogenic infection is not one with little virus. It is one in which activation drops after primary infection, so that T cell loss balances T cell renewal. The host keeps its immune system; the virus keeps replicating and spreading. Nef is the protein that strikes the bargain, reducing expression of host proteins involved in T cell activation.
Nef's function is conserved across primate lentivirus evolution, which is why those infections are survivable. The Nef of SIVcpz and every HIV-1 group has lost it, and actually enhances T cell activation, which is why they are pathogenic. The loss happened somewhere in SIVcpz's complicated ancestry, before the jump to humans.
Recall the recombination: SIVgsn (greater spot-nosed monkey) and SIVrcm (red-capped mangabey) met in one P. T. Troglodytes chimpanzee, probably before the troglodytes/ellioti split, so 50,000 to 1 million years ago, and the hybrid spread through the chimpanzee lineages before crossing to humans about 100 years ago. Nef's function must have been given up along that path: SIVcpz's vpu, from the SIVgsn side, now partly does Nef's old job by destroying CD4 molecules in infected cells.
Harmless viruses are not harmless by nature: SIVsmm causes simian AIDS in rhesus macaques (as SIVmac) and SIVagm in pig-tailed macaques, and HIV-2, straight from SIVsmm, is pathogenic in people, if less so than HIV-1. Nef must have other interactions with host proteins, and unidentified restriction factors plus differences in the immune response also decide the outcome.
TRIM5α: the gate that lets apes be infected by monkey viruses
TRIM5α (tripartite motif family) was found while researchers were trying to work out why rhesus macaques resist HIV-1, a resistance that ruined them as an AIDS model. The protein binds incoming virus capsids and destroys them, aborting infection before the genome can integrate.
Primate TRIM5α proteins are among those evolving rapidly under ancient viral pressure, and have become species-matched: by swapping capsid genes between lentiviruses experimentally, researchers showed that TRIM5α of higher primates, chimpanzees and humans included, inhibits their own lentiviruses but not those of most monkey species.
Because chimpanzee TRIM5α does not block incoming monkey lentiviruses, the unlucky P. T. Troglodytes that acquired both SIVgsn and SIVrcm could be colonised by both at once. The defence that protects a species against its own viruses is precisely the one that leaves it open to another species'.
TRIM5α has clinical weight too. A 2009 study of people at high risk in Durban, South Africa found lower TRIM5α expression in those who became infected, with the low levels present before infection and unchanged by it, so not virus-induced. High expression appears to reduce susceptibility, which may help explain why HIV is a poor transmitter compared with other sexually transmitted viruses: hepatitis B is about 300 times more infectious.
Tetherin: the step that made the pandemic
Tetherin sits in the cell membrane and holds newly formed virus particles there, stopping them from spreading, a defence also seen against Ebola. It was found by researchers asking why Vpu was needed for virus release from some cell types but not others: they fused the two cell types, found the hybrid always required Vpu, and deduced that a blocking factor was present in one and had to be overcome.
Every SIV has evolved some way of defeating tetherin, which marks it as a serious hurdle for a colonising virus. Different viruses use different tools: some Vpu, some Nef, some even Env. HIV-2 group A uses Env although its ancestor SIVsmm uses Nef; whether groups B to H ever solved the problem is unknown, and Crawford suggests the answer might explain why A and B spread and the rest did not.
Human tetherin differs from the chimpanzee protein by a small deletion of four amino acids, and they sit in the tail, exactly where Nef binds. That makes human tetherin insensitive to SIVcpz Nef. In the laboratory, putting the four residues back restores Nef's activity; an incoming virus had no such option. SIVgor's Nef works in gorillas, which is why the chimpanzee-to-gorilla jump was easier than the chimpanzee-to-human one.
Group M solved the problem by reverting to Vpu, the tool its distant ancestor SIVgsn used. In the laboratory, group M Vpu is a potent inhibitor of human tetherin; group O and P Vpu have no such activity at all; group N Vpu has a little, apparently gained at the cost of its other job, destroying CD4. This is the first adaptation found in HIV-1 that lets it replicate efficiently in human cells, and since it works best for group M, it may be why group M alone spread. The review accompanying the papers was titled "A Tail of Tetherin: How Pandemic HIV-1 Conquered the World".
The barrier between species is not one wall but a layered set of defences, each specific to the species that evolved it, and a virus crosses only when it happens to carry a tool that fits the new lock. That is why the same hunting exposure produces twelve crossings and one pandemic: the rare event is not transmission, it is a chance molecular match plus the social conditions to exploit it. Any assessment of a new zoonosis has to ask both questions, and the molecular one cannot be answered by counting cases.
The tetherin story is Crawford's strongest candidate for group M's success, and it still stands as the clearest molecular difference between the groups. The later 2014 Faria analysis complicates the causal claim: group M and group O grew at similar rates until about 1960, with group M pulling away only afterwards, which fits an explanation based on transport networks and post-independence upheaval at least as well as one based on Vpu. The honest position is that the molecular advantage is real but was probably permissive rather than decisive.
The Challenge of Pandemics
Crawford closes by putting HIV beside plague, smallpox, measles, Spanish flu, SARS and syphilis, and asks what the molecular history is worth. Her answer: the clock that dated HIV has dated our oldest killers too; most new human diseases are zoonoses; HIV's career was built on chance at every step; and its genetic diversity, up to 30% within one subtype against 4% for measles, is why a vaccine remains the hardest problem in the field.
The clock applied to older plagues
The same sequencing and tree-building that solved HIV has rewritten the origins of our classic killers, which are, like HIV, mostly zoonoses: microbes of animals that occasionally find a way into people.
| Disease | True origin revealed | Transmission oddity |
|---|---|---|
| Bubonic plague (Yersinia pestis) | A rodent microbe that became a blood parasite within the last 20,000 years | Never learned to pass between humans: every case is a rat flea bite, even in a pandemic |
| Smallpox (variola major) | Not from cowpox, as long assumed: closest relatives are rodent and camel poxviruses, all from one ancestor; probably entered humans in Asia about 6,000 years ago | Entirely human, airborne, highly efficient |
| Measles | Closest relative rinderpest of cattle, then canine distemper; measles and rinderpest diverged about 2,000 years ago | Airborne, needs large populations |
| HTLV-I | Simian ancestor in central Africa, several jumps tens of thousands of years ago | Persistent and silent; reached Japan by perhaps 300 BC, the Caribbean via the slave trade |
Viruses are inert particles with no say in where they go; spread is chance, and fast evolution lets them exploit whatever chance offers. HIV-1 group M's history is a chain of accidents: a hunting injury, possibly a contaminated needle, a river journey to Leopoldville, a city with four men to every woman, and then a passport. By the early 1960s that chain had produced only a few thousand infections, and that was enough.
WHEN SOCIETIES CHANGE, MICROBES MOVE
- The farming revolution, about 11,000 years ago, is the largest upheaval of all: domestication put hundreds of animal microbes within reach of a new host. Flu, naturally a virus of wild birds, could use domestic waterfowl as a stepping stone and pigs as the adapter before reaching people.
- The European arrival in the Americas gave Old World microbes a new continent; smallpox and measles destroyed whole Native American populations and did much of the work of the Spanish conquest of the Incas and Aztecs.
- Colonial Africa did the same for HIV, as chapter 7 showed.
A new H1N1 strain emerged as the Great War ended, circled the world in six months, infected 500 million people, a third of humanity, and killed about 50 million, far more than the roughly 10 million killed in five years of combat. Troop encampments bred it and troop movements carried it. To find out why it was so lethal, researchers reconstructed its genome from the lungs of a victim buried in Alaskan permafrost since 1918: every gene came from an avian flu virus that had never infected pigs or people, and the specific mutations conferring human transmissibility and virulence could be identified. Those same markers are now monitored in H5N1, which is highly virulent, has jumped to humans repeatedly, and so far cannot spread between them. Most flu experts expect that to change.
Why HIV went unnoticed
Plague, smallpox, measles and flu announce themselves: short incubation, devastating illness, huge death tolls, obviously infectious even to people who knew nothing of microbes. HIV did the opposite, causing no symptoms for years, so it spread widely before anyone noticed.
There is a precedent in the deepest human past. Primate-to-human microbe transfer in sub-Saharan Africa predates farming by millennia, but among small, mobile hunter-gatherer bands an acute infection burns through everyone and then dies out. Only microbes that establish persistent infection, letting the host live long enough to pass them to the next generation, could survive. HTLV-I is the surviving example, and HIV is built on the same strategy.
- UK to Australia: about a year in the 18th century, 100 days in the 19th, 50 days with steamships in the 20th, under 24 hours by the time of SARS.
- Aircraft now move over a billion people a year between more than 200 countries.
- One sneeze by an infected doctor in a crowded Hong Kong hotel lift seeded SARS to 27 countries.
SARS was containable because it made people obviously and severely ill and did not travel far through the air. HIV's "patient zero" spread virus across two continents for years while apparently in good health.
The slow pandemic
Compared with flu, smallpox, plague and SARS, HIV spreads slowly and infects inefficiently, largely because of its sexual route, which is unusual for a pandemic microbe. The only other sexually transmitted pandemic is syphilis.
The disease appeared in Spain as Columbus and his crew returned, so it was assumed to have come from the Americas; its origin is still debated. Recognised in 1494, it spread through Europe, Asia and North Africa, helped by war: Charles VIII of France marched on Naples, the disease tore through troops, mercenaries and camp followers, the king caught it and ordered a retreat, and the disbanding army carried it home. Rates of 10-20% persisted in Paris, London, Berlin and New York until penicillin.
"The Italians called it the French disease, the French the disease of Naples, the Poles the German disease." On naming syphilis, 1490s
The blame reflex of the 1490s matches that of the 1980s exactly: in the Middle East it was the European pustule, in India the Franks, in China the ulcer of Canton, in Japan Tang sore. In Britain it was the great pox, to distinguish it from smallpox, and contemporary descriptions suggest its symptoms were far more severe then than now.
| HIV-1 | Syphilis | |
|---|---|---|
| Primary infection | Mild, easily missed | Mild, easily missed |
| Persistence | Lifelong | Lifelong |
| Progression untreated | Virtually all develop AIDS | About a quarter develop tertiary disease |
| Timing | Kills young adults | Usually silent until mid or late life |
| Social impact | Devastating globally | Limited, chiefly through a few infected rulers, possibly Ivan the Terrible and Henry VIII |
The vaccine problem
Vaccines ended smallpox and are closing on measles; where none exists, as for syphilis, drugs cure the individual but do not stop transmission. Thirty years after discovery, HIV had HAART, which prevents progression, and no vaccine.
Within a single HIV-1 subtype, amino acid variation reaches 30%. For measles, where one vaccine works everywhere, it is about 4%. With subtypes and recombinants evolving continuously, no single vaccine is likely to protect against all HIV-1 variants.
- Consensus sequences. Identify sequences common to clusters of subtypes and recombinants and build a consensus amino acid sequence to induce immunity across the cluster.
- Ancestral sequences. Use a sequence from a virus ancestral to a cluster, in the hope of broad coverage.
- The flu model. Prepare a combination vaccine of several circulating strains, varied by region, as is done annually for flu.
- Fitness mapping. Continue the hunt for mutations tied to viral fitness, as was done by reconstructing the 1918 flu, to inform both vaccine and drug design.
Past pandemics often ran for over a century before dying out naturally. Crawford expected the HIV pandemic to continue into the 22nd century, and closed with Bill Clinton at the 18th International AIDS Conference, Vienna, 2010, paraphrasing Churchill: "This is not the end. It's not even the beginning of the end. It is only the end of the beginning." With over two million new infections a year at the time of writing, she concluded that the people alive today will not see the pandemic's final shape, but that understanding where, how, when and why this virus emerged is what makes preventing the next one possible.
Crawford's closing predictions have aged in two directions at once. Better than expected. New infections fell from over 2 million to about 1.2 million a year, deaths from a 2004 peak of 2.1 million to about 570,000, and 32.1 million people are on treatment. Treatment as prevention was proved (undetectable = untransmittable), and prevention became long-acting: cabotegravir injections every two months, then lenacapavir, approved by the FDA in June 2025 and in the EU that August, which is given twice a year and showed near-complete protection in trials; WHO recommended it in July 2025. Seven or more people have been cured or put into long remission by CCR5-delta32 stem cell transplants (Berlin, London, Düsseldorf, New York, City of Hope, the "second Berlin patient" with a heterozygous donor, and others since), though transplantation is not a scalable treatment. Worse than expected. No vaccine exists. The Imbokodo (2021) and Mosaico (2023) efficacy trials both failed, and the field has shifted to germline-targeting immunogens and broadly neutralising antibodies, still years from licensure. Funding collapsed: international funding fell 18% in 2025, the sharpest drop in nearly two decades, as PEPFAR and other US programmes were cut, and UNAIDS warned that sustained loss of that support could add millions of infections and deaths by 2030. The Churchill line still fits, with the stress now on the first clause.
Crawford's book is a case study in reconstructing an unwitnessed event from indirect evidence: mutation rates as a clock, sequence trees as a map, archival specimens as fossils, and colonial paperwork as context, each unreliable alone and decisive in combination. The pattern repeats whenever a new pathogen appears, as it did with SARS-CoV-2 in 2020, when the same phylogenetic toolkit was turned on a live outbreak in weeks rather than decades. The practical lessons are unglamorous: keep the specimens, date them properly, publish sequences to a shared database, and assume the first recognised case is not the first case.
Master Takeaways
The twenty-five things worth carrying away from Virus Hunt: the findings, the methods that produced them, the errors the field had to retract, and the places where thirteen years of later work have changed the picture.
- Species jumps are common; pandemics are rare. At least twelve simian immunodeficiency viruses have entered humans, yet only HIV-1 group M ever went global.
- The pandemic was a social event as much as a biological one. Colonial Leopoldville supplied migrant men outnumbering women over 4 to 1, commercial sex and untreated genital ulcer disease.
- HIV-1 group M was born in southeast Cameroon, in the chimpanzee communities at field sites MB and LB; group N came from one community, EK, in the Dja Reserve.
- HIV-2's parent is SIVsmm of the sooty mangabey, and its eight human groups sit interspersed among mangabey viruses in the tree, which is what proves each one is a separate transmission.
- SIVcpz is itself a recombinant, assembled inside one chimpanzee from SIVgsn (which supplied pol and vpu) and SIVrcm (which supplied env), both viruses of monkeys that chimpanzees hunt and eat.
- Groups O and P came via gorillas. SIVgor crossed from chimpanzees to western lowland gorillas perhaps 100-200 years ago, and work published after this book confirmed gorillas as the proximate source of group O.
- Two fossil viruses carry the dating. ZR59, from a 1959 Kinshasa blood sample, and DRC60, from a 1960 lymph gland in a wax block, differ by about 11%, so the subtypes had already separated by 1960 and the founding infection long preceded it: 1884-1924.
- The molecular clock had to be rebuilt before it could be trusted. Mutation rates vary along the genome (about 7 per 1,000 nucleotides per year in variable env, 3 in conserved pol), nucleotides mutate more than once, viruses recombine, sampling dates are not infection dates, and the rate changes with disease stage.
- Fast-evolving sequences bias deep dates towards the present. The Bioko Island calibration gave 49,129 years from RNA fragments and 76,800 years from amino acid sequences for the same ancestor.
- A relaxed clock was validated against reality on the Thai epidemic (model 1986 against an epidemiological 1984) and a thirteen-year Swedish transmission chain dated to within months.
- HIV reached Haiti before the US. Haitian subtype B isolates are more diverse than all other non-African subtype B put together: Africa to Haiti about 1966, Haiti to the US about 1969, Haiti to Trinidad about 1973, each by a single founder virus.
- The virus circulated in America for a decade unnoticed, reaching about 5% of sexually active gay men in San Francisco and New York by 1978.
- Cross-species transfer happens through blood, not sex. SIVs are concentrated in blood and blood-rich organs, and hunting, butchering and pet bites supply the exposure. The modern rate is measurable: 10 of 1,000 Cameroonian hunters carried simian foamy virus from local species.
- Captive animals corrupt the data. Cam3 infected Cam4 in a rescue centre; SIVmac and SIVstm were created by kuru and leprosy experiments that injected mangabey material into macaques. Nearly every early inference from captive primates needed correction.
- Sampling method decided what could be learned. Reading virus genomes from faeces collected off the forest floor, with RNA preservative, subspecies identification and individual genetic fingerprinting, converted an unanswerable question into a map reference.
- SIVcpz harms chimpanzees. At Gombe, infected animals died at 10-16 times the rate of uninfected ones, with a third the birth rate, and the modelled prevalence that tips a community into extinction is only 3.4%; immigrating females are what rescues a community.
- The oral polio vaccine theory failed on four independent grounds: the date moved to before the 1930s; Camp Lindi lay in the range of the wrong chimpanzee subspecies; stored CHAT pool 13 contained macaque DNA and no SIV; and the starburst of subtypes formed after the jump to humans.
- The "patient zero" story was wrong twice over. The CDC cluster's "O" meant "Outside California", and sequencing in 2016 showed Gaëtan Dugas carried a strain typical of viruses already circulating.
- The Manchester sailor was laboratory contamination. The 1990 claim to the world's first AIDS case collapsed when the sequence proved modern and the tissues came from two different people.
- Pathogenicity tracks novelty. Untreated HIV-1 raises mortality up to 60 times, SIVcpz 10-16 times, HIV-2 2-5 times, and SIVsmm and SIVagm not at all in their ancient hosts.
- A harmless lentivirus is not one with little virus. SIVsmm and SIVagm keep high viral loads but shut down T cell activation after primary infection, so loss balances renewal. Nef does that in old infections and fails to in HIV-1 and SIVcpz, where it increases activation.
- Four missing amino acids may explain the pandemic. Human tetherin lacks the tail residues Nef binds, so SIVcpz Nef cannot disable it. Group M's Vpu re-evolved to do the job; group O and P Vpu cannot, and group N's barely can.
- One conserved adaptation marks every human lineage: the Met to Arg switch at Gag-30, made independently by the ancestors of groups M, N and O, and reversed by the ancestors of subtype C and by HIV-1 passaged back through chimpanzees.
- Our genome carries the scars of this war. 8% of human DNA is retroviral remnant, and thirteen genes including known restriction factors show positive selection across 40 million years of primate evolution.
- The pandemic is not over, and the trend has reversed direction. New infections fell to about 1.2 million a year and deaths to about 570,000, with 32.1 million on treatment and twice-yearly lenacapavir approved in 2025, but international funding fell 18% in 2025 and no vaccine exists.
Frameworks and mental models
| Name | What it says | Use it when |
|---|---|---|
| Molecular clock | Genetic distance is proportional to divergence time, once corrected for rate variation and repeat mutation | Dating an event nobody recorded, from surviving copies of a fast-changing thing |
| Relaxed clock | Allow the rate to vary by site, gene and stage, then validate against outbreaks with known dates | Any inference whose model assumptions are stronger than the data |
| Founder event and bottleneck | A population seeded by one or a few individuals loses diversity; descendants carry the founder's quirks | Explaining why regional variants differ without invoking selection |
| Dead-end infection | Transmission without an onward chain extinguishes the pathogen regardless of severity | Judging whether a new zoonosis is a threat or a curiosity |
| Koch's postulates, revised | Indirect markers (antibodies, PCR) replace culture; natural experiments replace inoculation | Establishing causation when the clean experiment is impossible or unethical |
| Red Queen arms race | Host and pathogen must keep evolving to maintain the same balance | Reading defences and countermeasures as a matched pair, not as fixed traits |
| Restriction factor versus accessory gene | Species-specific cell defences (TRIM5α, APOBEC, tetherin) against viral genes (vif, vpr, vpu, nef) evolved to disable them | Asking why a pathogen crosses into one species and not another |
| Evolutionary-geographic clustering | Viruses from the same community are most alike; rivers and distance mark the breaks | Locating a source population from sequences alone |
| Ecologic association | Overlap in place and time may be real without being causal | Resisting a compelling map that has no denominator behind it |
Where it is strong: Crawford is scrupulous about uncertainty, shows failed hypotheses in full, names the people and their disagreements, and never pretends the founding event was observed. Where it has dated: the group O origin, left open with three competing trees in chapter 4, was resolved in favour of a gorilla source in 2015; the dating window of 1884-1924 was narrowed in 2014 to about 1920 in Kinshasa, with the rail and river network doing the work she attributes more loosely to river traffic; the "patient zero" account in the Introduction was overturned in 2016; and the treatment and prevention picture of chapter 9 has been transformed by U=U, long-acting PrEP and the transplant cures. Where it is one-sided: the tetherin explanation for group M's success is presented as close to settled, and the later finding that groups M and O grew at similar rates until about 1960 suggests the molecular advantage was permissive rather than decisive. The Marx needle hypothesis is given a fair hearing but, as Crawford concedes, cannot now be tested either way. Where the frame is narrow by design: this is a book about scientists and viruses. The political economy of the epidemic, treatment access, activism and the consequences for the countries worst affected are deliberately out of scope, and readers should not mistake the omission for a judgement about importance.
Who's who
| Name | Role | Why they matter |
|---|---|---|
| Dorothy H. Crawford | Virologist, Edinburgh; the author | Watched from an immunology unit that helped identify the CD4 receptor |
| Luc Montagnier, Françoise Barré-Sinoussi | Pasteur Institute | Isolated LAV in 1983 and HIV-2 in 1986; Nobel 2008 |
| Robert Gallo | US National Cancer Institute | Found HTLV-I and II and the methods that made HIV's isolation possible |
| Peter Duesberg | Berkeley retrovirologist | The denialist case, and the South African policy disaster that followed |
| Peter Piot | Antwerp, then UNAIDS | Found the African epidemic in 1983; founded Project SIDA |
| Beatrice Hahn | University of Alabama, then Pennsylvania | Sequenced Marilyn's virus; led the wild chimpanzee programme |
| Paul Sharp | Nottingham, then Edinburgh | Built the trees; found recombination, Gag-30 and the SIVcpz hybrid origin |
| Martine Peeters | Antwerp, Gabon, Montpellier | Found the first infected chimpanzees; ran the Cameroon field collections |
| Mike Worobey | Oxford, then Arizona | Recovered DRC60 and the Haitian fossils; refuted the polio vaccine theory |
| Marilyn and Noah | Two captive chimpanzees | Their stored tissue defined the two SIVcpz lineages |
| Hilary Koprowski | Wistar Institute | Made CHAT; sued over the polio vaccine theory |
| Tom Curtis, Edward Hooper | Journalists | Raised and then championed the OPV theory |
| W. D. Hamilton | Oxford evolutionary biologist | Forced the theory to be tested; died of malaria after the 2000 expedition |
| Preston Marx | Tulane | Argues unsterile needles were needed for the virus to adapt |
| Kevin De Cock, Robin Weiss, Andrew Rambaut | CDC, UCL, Oxford | Supplied the epidemiological, summary and starburst arguments that closed the debate |
| Kimberly Bergalis, David Acer | Patient and dentist, Florida | The transmission cluster that met Koch's third postulate |
| Jane Goodall | Gombe | Fifty years of non-invasive study that made the chimpanzee work possible |
Connect it
- Provenance beats recollection. The whole reconstruction rests on specimens and records nobody expected to need: a wax block from 1960, blood frozen and forgotten in Miami, colonial health reports. Systems that keep immutable, dated records with their context let you answer questions you had not thought to ask.
- Validate the model on cases with known answers. The clock was only believed once it reproduced the Thai epidemic and a transmission chain dated to the month. Backtesting against ground truth, not internal plausibility, is what moved the field.
- An ecologic association is not a signal. Hooper's map overlapped perfectly and meant nothing without a denominator. The same failure mode appears whenever a dashboard shows two things rising in the same region.
- Watch for the contamination that confirms you. The Manchester result was positive because the laboratory was full of the thing it was testing for. Any pipeline that processes positives and unknowns together will eventually produce a spectacular false positive.
- The rare event is usually the combination, not the step. Exposure happened constantly; what was rare was exposure plus a molecular match plus a population that could sustain a chain. Risk models that score single steps in isolation will systematically misprice the tail.
About this book
Virus Hunt: The Search for the Origin of HIV by Dorothy H. Crawford, 2013. This is a chapter-by-chapter study of the book, with summaries, maps, tables and diagrams; short quotations keep their page references. It is shared free for personal study, is not for sale, and is not endorsed by the author or publisher. Please read and buy the original.
Insight Editions · rajeshrmahapatra.com/insight-editions