Out of Africa, there is always something new, wrote historian Pliny the
Elder more than 2,000 years ago. He was quite right. As early as 1984, just three years after the first description of the new disease, it was suspected that HIV, its recently discovered aetiological agent (then known as human T-cell lymphotropic virus (HTLV)-III in the US, LAV (lymphadenopathy-associated virus) in Europe), originated in central Africa.
This was mainly because the first studies in Africa, conducted in Zaire and Rwanda, showed that AIDS was common in Kinshasa and Kigali, where nearly 90% of sex workers were
infected. These field studies were prompted by the observation that of the first few hundred cases of AIDS diagnosed in Europe, about half occurred among patients coming from central Africa, mostly from Zaire.
Over the following years, the epidemiology of HIV-1 infection in Kinshasa would be described in great detail by a group of American, Belgian and Congolese researchers known as Projet Sida, based at
Hôpital Mama Yemo (Mama Yemo was dictator
Mobutu’s mother, a former sex worker, and she suffered the same fate as the Belgian colonists after her son was overthrown: this institution is now called Hôpital Général de Kinshasa). Projet Sida came to an abrupt end in 1991, when the whole of Kinshasa was looted by the city’s poor people
. During the same period and until the 1994 genocide, similar epidemiological work was conducted in Kigali, 1,500 kilometres east of Kinshasa
.
1
–
3
In retrospect, this early vision of central Africa as the source of HIV-1 was rather naive. Researchers assumed that since this was at the time the region with the highest prevalence (i.e. the proportion of the population that is infected) among groups representative of the general adult population, the virus must have originated there. There were at least two problems with this assumption.
First, there was an obvious bias, as little information on HIV prevalence was available from other parts of the continent, especially East and
southern Africa. Belgian researchers, the most prominent being Peter
Piot, from the Institute of Tropical
Medicine in Antwerp (who would later become the founding executive director of UNAIDS, the UN programme specifically dedicated to the control of HIV/AIDS), had naturally initiated HIV research in the former Belgian colonies where their institutions had maintained networks and contacts over the preceding decades. Much to their credit, Zaire and Rwanda were open to AIDS research from the start, but this was not the case in other countries such as
Burundi and some English-speaking countries of East Africa, where there was a strong temptation to keep AIDS under wraps: if we ignore it, perhaps it will go away
.
Second, the relationship between HIV prevalence and duration of the epidemic is not straightforward: it all depends on the annual
incidence (the proportion of previously uninfected individuals who acquire HIV each year). We now know that in Kinshasa the HIV incidence among the general adult population was probably never higher than 1% per year. However, in some countries of southern Africa, annual incidence reached the extraordinary level of 5% in the 1990s (one seronegative adult out of twenty got newly infected with HIV each year). A prevalence of 10% could reflect an annual incidence of 1% continuing for more than ten years, or an incidence of 5% over just a couple of years. However, even if these assumptions about a central African origin of HIV/AIDS were naive, eventually they proved to be correct, showing that in the scientific domain intuition can sometimes be trusted
!
Additional support for a central African origin of HIV-1 came from the testing of archival samples of blood. In the mid- and late 1980s, to understand the dynamics of HIV in the recent past, researchers tried to locate collections of sera obtained earlier for other purposes and kept frozen. Scientists tend to clean out their freezers once in a while to make room for new samples, or their samples are destroyed when they retire or move on to other positions. However, sometimes samples are forgotten for a long time or deliberately conserved. In Kinshasa, among mothers attending a well-baby clinic in the
Lemba district, HIV-1 prevalence was 0.25% in 1970 (n=805) and 3.0% in 1980 (n=498)
.
In the remote Catholic mission of Yambuku and surrounding communities of the
Equateur province of Zaire, 0.8% of 659 samples collected in 1976
during an investigation of an epidemic of Ebola fever were found to be HIV-1 seropositive when tested ten years later. This proved that the virus had existed in this part of the world for some time, but not necessarily that it originated there; testing of archived samples of serum from American gay
men who participated in epidemiological studies of hepatitis B also retrospectively documented cases of HIV-1 in the late 1970s, and even earlier for drug
addicts.
4
–
8
The Yambuku epidemic of Ebola fever which had prompted the collection of these samples had largely been ‘
iatrogenic’ (healthcare related). In this small rural hospital, syringes and needles were scarce and constantly re-used, fuelling transmission of the blood-borne Ebola virus between patients attending the hospital for other reasons (malaria, gonorrhoea, etc.). The nuns issued only five syringes to the nurses each morning, which were then used and re-used on the 300 patients attending each day. Three-fourths of the first 100 cases of Ebola in Yambuku were infected through injections received at the hospital. The epidemic came to an end after the hospital was closed following the death of several nurses and nuns, infected by their patients. Clearly, noble intentions for providing health care to the underprivileged could have disastrous consequences when the risk of transmission of blood-borne viruses was not appreciated. This unfortunate situation was not new or specific to the Yambuku hospital, and had already had infinitely more dramatic consequences, although this was not known at the time, than these few hundred deaths from Ebola fever. It was decided to call this new disease after a nearby river rather than after the Yambuku mission, to avoid further stigmatisation after all it had gone through. The contrast between the two diseases is an excellent illustration of the genius of HIV. People infected with the Ebola virus quickly fall ill and die. This causes a spectacular epidemic, which triggers a massive (and always successful) reaction to control it. People infected with HIV, on the other hand, can live and quietly pass on the virus for ten years or more, and it will take even longer before physicians can recognise the emergence of this new disease, because symptomatic cases are not clustered within a short period of time
.
9
,
10
Elsewhere in Africa, no trace of HIV was found before the 1980s, which increasingly pointed to a central African origin of this ‘new’ virus. In West Africa, out of more than 6,000 samples obtained in
Nigeria,
Liberia,
Ivory Coast,
Togo,
Senegal,
Sierra Leone,
Mali,
Niger and
Ghana in the 1960s and 1970s, not a single case of HIV-1 infection
was found. A few cases of
HIV-2 infection were documented, however. Among 789 samples obtained in Senegal in 1981, one was positive but it is unclear whether this corresponded to HIV-2 or HIV-1.
11
–
16
Meanwhile in East and southern Africa, in samples obtained from low-risk groups between 1959 and 1981, HIV was not found in Mozambique,
Zimbabwe,
Zambia,
Uganda,
Tanzania and northern
Kenya, nor in mine workers in South Africa (who originated from
Mozambique,
Malawi,
Lesotho,
Botswana,
Angola,
Swaziland and
South Africa itself). The earliest evidence of HIV in East Africa comes from Nairobi in 1980–1 where 1% of patients with
STDs and 5% of sex workers were HIV-1-infected. Just three years later, 82% of
Nairobi sex workers were HIV-1-infected. This exponential transmission among
prostitutes is central to the story and will be examined later.
14
,
17
–
19
Documentation of early cases of full-blown AIDS was also achieved retrospectively. First, let me say that no conclusions can be drawn from isolated cases of apparently immunocompetent patients found to have had, many years ago, a diagnosis of a condition now frequently associated with AIDS such as
Pneumocystis
pneumonia, if this is not substantiated by a specimen positive for HIV in the patient or his/her spouse. This is because there are rare non-infectious diseases of the immune system which lead to very low counts of CD4 lymphocytes (the cells which are the main target for destruction by HIV), and subsequently to any of a long list of opportunistic infections. Short of an archived specimen positive for HIV, the clustering of cases, geographically or temporally, or within a couple, is more suggestive of AIDS but never conclusive.
20
–
21
Valuable journalistic information about some documented early cases can be found in
The river
as well as in
And the band played on
.
The most interesting is that of a Norwegian family (father, mother and nine-year-old daughter), who all tragically died in 1976 from AIDS caused by
HIV-1 group O, and whose sera were found to be HIV-positive when tested twelve years later. The child was born in 1967, which implies that the mother was already HIV-infected by then. The father had been a sailor, visiting a number of ports in Africa in the early 1960s, where he developed STDs, presumably after contacts with
prostitutes. He probably acquired HIV-1 group O in
Nigeria or Cameroon
, where his boat stopped for a few days in 1961–2.
A Danish
surgeon died of
AIDS in 1977, after working at the Abumonbazi rural hospital in Zaire in 1972–5 and in Kinshasa in 1975–7, following an earlier stint in
the same country around 1964. An eight-year-old Zairean child, infected perinatally in 1974–5, died in
Sweden in 1982, and AIDS was serologically proven later on. A very unfortunate
Canadian pilot involved in a plane crash in 1976 in northern Zaire, where he had surgery and received a blood transfusion, died of AIDS in 1980; his serum was later found to be HIV-1-positive (transfusion-acquired HIV infection progresses rapidly to AIDS, because of the huge quantity of viruses present in the blood bag). Former physicians at the university hospital in Kinshasa reported seven cases of AIDS diagnosed retrospectively, five of them confirmed serologically, which had been acquired sexually in the DRC (or
Burundi in one case) in the late 1960s or the 1970s, mostly among Belgian nationals. Then in 1979, cases of AIDS started trickling down among the small proportion of Zaireans rich enough to seek health care in
Belgium.
22
–
30
We do not know whether other researchers tested ancient samples from other parts of Africa without reporting their findings. Studies with negative results tend not to be published in scientific journals, a phenomenon known as ‘publication bias’. Thus although sketchy, testing of archival samples suggested that HIV-1 was present in the 1960s and 1970s, albeit at a low prevalence, in several locations in central Africa but not in West or East Africa.
The next step came from the documentation of the earliest case of HIV-1 infection in a sample obtained in the Belgian Congo around 1959, during the course of a study on genetic diseases of red blood cells. Of 672 samples, collected in Léopoldville and other locations, and miraculously kept (probably forgotten) in a freezer, one was found twenty-six years later to contain antibodies against HIV-1. Apparently, the HIV-1-positive specimen came from a male adult recruited in Léopoldville. HIV genetic material was amplified from this sample, and analyses confirmed that this was indeed the oldest HIV-1 isolate ever documented. It was named ZR59.
31
–
33
It took more than twenty years for another ancient specimen containing HIV-1 to be located. Finding old tissue blocks collected between 1958 and 1960 and kept at the pathology department of the University of
Kinshasa, scientists discovered HIV-1 sequences in a lymph node biopsy obtained in 1960 from an adult woman. It was given the name DRC60. Twenty-six other specimens (lymph nodes, livers and placentas) did not contain HIV. DRC60 and ZR59 differed by about 12%. It was calculated that DRC60 and ZR59 shared a common ancestor
around 1921, as we will discuss later. Although the exact time of its introduction into human populations remains debated, there is no doubt that HIV-1 was present in Léopoldville by 1959–60
.
34
Now we will examine how the genetic diversity of HIV-1 in different parts of the world helped scientists trace back the origins of the virus. But first, we need to review quickly what ‘sequencing’ is all about. Sequencing is the identification in their proper order of the series of ‘nucleotides’ that constitute a gene. There are four types of nucleotides: adenine (A), thymine (T), guanine (G) and cytosine (C). The genome of any living organism is a long list of these four letters. When scientists compare viruses, the similarity between sequences is called ‘homology’, and non-similarity ‘divergence’. If 90% of the nucleotide sequences between two isolates are the same, they have 90% homology or 10% divergence. This degree of divergence is used to decide whether two isolates constitute subtypes of one viral species, or two distinct species. For instance, sequences of HIV-1 and
HIV-2 differ by more than 50%.
Based on such analyses, HIV-1 is now divided into four ‘groups’: group M (main), which is responsible for the current pandemic and causes more than 99% of all HIV-1 infections in the world, group O (outlier), group N (non-M non-O) and group P, which did not spread outside central Africa, for reasons still unclear.
HIV-1 group M is further subdivided into nine ‘subtypes’: A, B, C, D, F, G, H, J and K (the alphabet is not respected because subtypes E and I were found not to be real subtypes and have been renamed). HIV-1 often makes mistakes when replicating, a phenomenon compounded by the high level of viral production throughout the long natural history of the infection. Up to ten billion copies of the virus are produced every day, and the potential for errors in replication is commensurate. Over time, the accumulation of these errors leads to viral diversity. When around 20% of the nucleotide sequences of the initial virus have undergone replication errors, the result will be a new subtype, as defined arbitrarily by scientific consensus
.
35
–
36
High-risk individuals (especially in Africa) can get infected with a first subtype, and later with a second subtype, which can recombine into ‘circulating recombinant forms’ (CRF): part of their genome is derived from the first subtype, part from the second. Recombinants can be
transmitted forward. Forty-eight recombinants have now been recognised. Their names correspond to the two subtypes involved in the recombination, for example CRF02_AG is a recombination of subtypes A and G. Some recombinants have generated their own epidemic in specific countries or regions.
37
There is no clear-cut difference between subtypes with regard to their propensity to cause AIDS, with one exception: individuals infected with subtype D die faster than others. Do subtypes differ in their transmissibility? Shedding of HIV-1 in the genital tract of women infected with subtype C is higher than for other subtypes, which would imply more effective female-to-male sexual transmission. With subtype C, the high degree of ‘
viraemia’ (the quantity of virus in the blood) that characterises acute infections may be worse than with other subtypes, increasing its infectiousness. Subtype C spread like wildfire in southern Africa, even if other subtypes had been introduced at the same time. These findings do suggest that subtype C is transmitted more efficiently than the others, which might explain its current worldwide preponderance.
38
–
42
Some subtypes are associated in specific locations with particular modes of transmission. This represents a founder
effect within specific risk groups: a subtype originally introduced in a group of intravenous
drug users will continue to be transmitted preferentially to other addicts, another subtype originally introduced in sexual networks of homosexuals will be transmitted preferentially to other gay men, and so on. A good example of this is
South Africa, where subtype B is found in 96% of white homosexuals (probably after it was imported from the US in the 1970s–80s), while subtype C accounts for 81% of infections of black heterosexuals. There is limited mixing between these two groups: few homosexual
Afrikaners have sex with
heterosexual Zulus
! To date, there is no evidence that some subtypes are intrinsically better transmitted by one route than another. As the relative contribution of some modes of transmission varies over time, depending on the effectiveness of control efforts targeting a specific risk group, the distribution of subtypes within a given country can also change.
43
HIV-1 evolves at a rate about one million times faster than that of animal desoxyribonucleic acid (DNA). This means that, in just over a decade, HIV-1 will change as much as all the genetic changes and the ensuing diversity that accrued among the common predecessors of
Homo sapiens
, chimpanzees and gorillas over ten million years. The longer HIV-1 has been present somewhere, the more opportunities it
will have had to undergo a series of mutations which will eventually allow it to evolve into different subtypes, and the more likely it is that recombinants will be created. Conversely, if we were to examine all of the HIV-1 isolates in a city or country in which the very first case was introduced only a year earlier, for example in a population of drug addicts, we would find little genetic variation and the viruses of all the individuals infected after this first case would still belong to the same original subtype, the founder. There would not have been enough replication errors to result in new subtypes.
Because HIV evolves in only one direction, from a single model of virus to an increasingly complex differentiation into numerous subtypes and recombinants, we can reconstruct the sequence of its progress in a particular region or country by examining the local distribution of subtypes. Starting in the early 1990s, as new tools made it easier to examine
nucleotide sequences from a large number of HIV-1 isolates obtained in various locations, an additional and most convincing argument emerged which supported a central African origin: the extreme genetic diversity of HIV-1 isolates from this part of the world.
Worldwide, subtype C accounts for about 50% of all HIV-1 infections, followed by subtypes B and A (10–12% each), G (6%), CRF02_AG (5%), CRF01_AE (5%) and D (2.5%), while subtypes F, H, J and K have undergone limited transmission (each fewer than 1% of cases). However, the contribution of each subtype varies dramatically from region to region.
44
–
45
In North America and Western Europe, respectively 98% and 88% of HIV-1 infections correspond to subtype B, which is clearly the subtype that was originally introduced into these two continents, the
founder strain. Subtypes other than B are usually found in
migrants, who acquired HIV-1 in their countries of origin. In contrast, in Eastern Europe and central Asia, subtype A accounts for 79% of HIV-1 infections: clearly, this epidemic had a different origin than that of Western Europe, and it spread initially through needles rather than gay sex.
36
,
44
–
46
In Latin America and the
Caribbean, subtype B accounts for respectively 75–80% and 95% of strains. Cuba stands out as the country with not only the lowest HIV prevalence in the Americas but also the highest diversity: about half of Cuban isolates are either non-B subtypes or recombinants.
This reflects the acquisition of multiple subtypes of HIV-1 (or recombinants) by some of the
internationalistas
, the soldiers
that
Castro sent to fight alongside the leftist Movimento Popular de Libertação de Angola during the civil war in Angola, and very limited opportunities for transmission upon their return to the island. The whole Cuban population was screened for HIV in 1986–9; seropositives were quarantined for years in AIDS sanatoria and brainwashed with preventive messages (Cuba was indeed the only country that tried to control HIV like an infectious disease, rather than making it a human rights issue). At the peak of their intervention in 1986, 35,000 Cuban troops were stationed in Angola, which became one of the most corrupt and capitalist regimes in Africa, while smaller numbers of Cuban soldiers were stationed in sixteen other African countries. Recent studies documented a high diversity in HIV-1 isolates in Angola, where all non-B subtypes found in Cuba are present. This illustrates how political and military events, even ideologies, had a measurable impact on the transmission of HIV
.
47
–
52
In southern Africa, subtype C corresponds to 92–8% of HIV-1 infections. This implies that the virus was introduced relatively recently into this region now so afflicted by AIDS, a finding corroborated by epidemiological investigations. Subtype C accounts for 99% of infections in
Ethiopia and also predominates in
Zambia, while subtype A accounts for 70% of infections in
Kenya. In
Tanzania, subtypes A, C and D are the major players. In
Uganda, which borders not only Tanzania and Kenya but also the DRC, there is more diversity, with a high prevalence of A, D and recombinants, and lower frequencies of C, B and G. In West Africa, all the way from
Nigeria to
Senegal, CRF02_AG predominates, implying that this part of the continent was infected only after subtypes A and G had had the opportunity to recombine.
36
,
45
,
46
Countries of central Africa (the two Congos,
Cameroon, Gabon, the Central African Republic and Equatorial Guinea) display by far the widest diversity in HIV-1 subtypes. All subtypes of
HIV-1 group M and many recombinants have been found in this region, where there is also more genetic diversity within each subtype.
Map 2
, reproduced from a 2003 review paper, illustrates the extreme genetic variation of HIV-1 in central Africa compared to other parts of the continent. Additional studies published since might indicate small changes in the distribution of this or that subtype, without modifying the general pattern. The conclusion is crystal clear: HIV-1 must have originated in central Africa, where it has had more time to diversify genetically.
45