911 911.920

Preston A. Marx1,2*, Phillip G. Alcabes3 and Ernest Drucker4 1Aaron Diamond AIDS Research Center,The Rockefeller University, New York, NY 10016, USA 2Tulane Regional Primate Research Center and School of Public Health andTropical Medicine, Tulane University Health Sciences Center, Covington, LA 70433, USA 3Hunter Col ege School of Health Sciences, City University of New York, NewYork, NY 10010, USA 4Monte¢ore Medical Center, Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, Bronx, NY 10467, USA There is compelling evidence that both human immunode¢ciency virus (HIV)types emerged from two dissimilar simian immunode¢ciency viruses (SIVs)in separate geographical regions of Africa. Each of the two HIVs has its own simian progenitor and speci¢c genetic precursor, and all of the primates that carry these SIVs have been in close contact with humans for thousands of years without the emergence of epidemic HIV. To date no plausible mechanism has been identi¢ed to account for the sudden emergence in the mid-20th century of these epidemic HIVs.
In this study we examine the conditions needed for SIV to complete the genetic transition from indivi- dual human SIV infections to epidemic HIV in humans. The genetic distance from SIV to HIV and the mutational activity needed to achieve this degree of adaptation to human hosts is placed within a mathe- matical model to estimate the probabilities of SIV completing this transition within a single SIV-infected human host. We found that the emergence of even one epidemic HIV strain, following a single human exposure to SIV, was very unlikely. And the probability of four or more such transitions (i.e. HIV-1 groups M, O and HIV-2 subtypes A and B)occurring in a brief period is vanishingly small. We conclude that SIV cannot become a zoonosis, but requires adaptive mutations to become HIV. Some modern event must have aided in the transition of SIV to HIV.
Our research indicates that serial passage of partially adapted SIV between humans could produce the series of cumulative mutations su¤cient for the emergence of epidemic HIV strains. We examined the rapid growth of unsterile injections in Africa beginning in the 1950s as a biologically plausible event capable of greatly increasing serial human passage of SIV and generating HIV by a series of multiple genetic transitions. We conclude that increased unsterile injecting in Africa during the period 1950^1970 provided the agent for SIV human infections to emerge as epidemic HIV in the modern era.
Keywords: acquired immune de¢ciency syndrome; simian immunode¢ciency virus; human immunode¢ciency virus; Africa; unsterile injections; zoonosis evidence that HIV ever existed in Africa or reached Europe and the New World prior to the 20th-century Multiple lines of evidence indicate that all known human epidemic. How, therefore, may we understand the nearly immunode¢ciency viruses (HIVs)derive from a group of simultaneous emergence of two genetically distinct types simian immunode¢ciency viruses (SIVs)that pre-date of epidemic HIV from ancient and separate simian the emergence of the acquired immune de¢ciency sources occurring in two di¡erent regions of Africa? syndrome (AIDS)epidemic by hundreds of thousands, Two paths are possible for the emergence of epidemic perhaps even millions of years (Sharp et al. 1994). Yet, HIV strains from ancestral SIV. The ¢rst is that HIVs despite centuries of opportunity to emerge (e.g. the evolved in Africa from relatively ancient, multiple and African slave trade that went on for over 300 years and separate cross-species transmissions of SIV to humans. In dislocated millions (Hochschild 1998)), there is no this model, HIV remained sequestered in remote areas of Africa, without spread to the general population within Africa or to other continents, until the population growth Author and address for correspondence: AIDS Research Center, Tulane Regional Primate Research Center, 18703 Three Rivers Road, and dislocations of the 20th century enabled HIV to Covington, LA 70433, USA (pmarx@adarc.org).
break out as a human epidemic. The second possibility is Phil. Trans. R. Soc. Lond. B (2001) 356, 911^920 912 P. A. Marx and others Serial passage of SIVand the origin of HIV that new ecological factors arose in the 20th century that Of the SIV lineages, four are divergent from HIV facilitated the transformation of SIV to epidemic HIV while the remaining two, from chimpanzees (SIVcpz) types. However, no models have yet been developed to and sooty mangabeys (SIVsm), are closely related to exclude either scenario or identify a plausible biological HIV-1 and -2, respectively (Peeters et al. 1989; Hirsch et al.
mechanism and speci¢c ecological factors, new to the 1989; Gao et al. 1999; Chen et al. 1996). Since HIV-1 and mid-20th century, that could be responsible for multiple -2 are genetically closer to naturally occurring SIVcpz transitions of SIV to distinct HIVs.
and SIVsm lineages, respectively (Peeters et al. 1989; Here we examine the plausibility of a hypothesis that Hirsch et al. 1989; Gao et al. 1999; Chen et al. 1996), than serial transmission of SIV between humans was the bio- they are to each other, types 1 and 2 therefore must have logical mechanism that permitted the accumulation of adaptive mutations of SIV, which led to the emergence of HIV-1's closest known relatives SIVcpzGab1, SIVcpzGab2 epidemic HIV strains. In this model we make a speci¢c and SIVcpzUS, occur in P. t. troglodytes, a chimpanzee distinction between epidemic and non-epidemic strains of subspecies whose natural range (Kingdon 1997)coincides HIV. The hypothesis was assessed in a model that calcu- with the occurrence of all three HIV-1 groups (M, N, O) lates the probabilities for individual non-epidemic SIV found thus far (Gao et al. 1999). In contrast, SIVcpzAnt human infections completing the genetic transition to from the P. t. schweinfurthii of eastern Africa has no known epidemic HIV. Using in vivo mutation rates and virus HIV counterpart (Gao et al. 1999)(¢gure 1a). A similar burst size data for lentiviruses, we measured the accumu- phylogenetic relationship was established for sooty lation of adaptive mutations per SIV genome per person.
mangabeys, C. t. atys (Hirsch et al. 1989; Georges-Courbot The model shows that even a single epidemic strain of et al. 1998; Chen et al. 1996), in that HIV-2 clusters with HIV is unlikely to have arisen spontaneously from direct SIV from only the sooty mangabey subspecies. The sister SIV infections of individuals in contact with SIV-infected taxon of sooty mangabeys, C. t. torquatus the red-capped mangabeys or chimpanzees. Moreover, the probability of mangabey (Georges-Courbot et al. 1998), harbours an several spontaneous transitions, which appear to have SIVsm-related virus (Georges-Courbot et al. 1998; F.
occurred within a decade in Africa, is vanishingly small.
Gao, personal communication)that has not been found in While many animal viruses can and have made the humans (F. Simon, personal communication)(¢gure 1b).
genetic transition to human pathogens (e.g. smallpox), the rapid transformation of four or ¢ve ancient SIVs into the modern HIVs in di¡erent parts of Africa suggests a single modern event that occurred in multiple locations.
To establish a speci¢c and biologically plausible agent The genetic lineages of the HIV-1 and -2 are shown in that could adequately amplify the serial transmission of ¢gure 1. A striking feature of HIV trees is the dis- SIV, we focused on the exponential growth of unsterile cordance in prevalence of HIV genetic variants. Of six injections that was associated with the introduction of HIV-2 subtypes, only A and B are epidemic (Chen et al.
injectable medications in these regions of Africa in the 1997; Gao et al. 1994). The rest, subtypes C^G (Yama- guchi et al. 2000), are non-epidemic HIV-2 strains that are weakly pathogenic, replicate poorly in infected humans and are found only within the range of the sooty mangabey (Chen et al. 1997; Gao et al. 1994). In conserved SIVoccurs as ¢ve or six di¡erent genetic lineages distrib- genes, the least divergent SIVsm is ca. 7.0% divergent uted throughout sub-Saharan Africa (Fukasawa et al. 1988; from non-epidemic HIV-2 strains and 12.1% from Peeters et al. 1989; Hirsch et al. 1989; Tsujimoto et al. 1988; epidemic strains (Chen et al. 1997). SIVsmH4 is ca. 25% Emau et al. 1991; Georges-Courbot et al. 1998). Recombi- distant from epidemic HIV-2 (Hirsch et al. 1989; Chen et nation (Robertson et al. 1995), phylogenetic data (Fuka- al. 1997). Non-epidemic subtypes are very rare, found sawa et al. 1988; Peeters et al. 1989; Hirsch et al. 1989; only in individuals who either live in or emigrated from Tsujimoto et al. 1988; Emau et al. 1991; Georges-Courbot et western Africa (Chen et al. 1997; Gao et al. 1994). For al. 1998)and the continent-wide distribution of naturally example, HIV-2 subtype F was found in only one person infected simian hosts (Georges-Courbot et al. 1998; among 9306 individuals living in the same area of rural Muller et al. 1993; Gao et al. 1999)provide strong evidence Sierra Leone (Chen et al. 1997)where sooty mangabeys that SIVs in mangabeys, African green monkeys and harbour natural SIVsm infections in this same region.
chimpanzees are ancient, up to one million years old.
SIVcpz strains from the P. t. troglodytes subspecies show a Other SIVs have spread across Africa through more similar relationship with epidemic and non-epidemic recent cross-species transmissions (Jin et al. 1994). SIV is HIV-1. The HIV N group, which is also rare in humans therefore ancient in Africa and all six SIV lineages pre- (Simon et al. 1998)(¢gure 1a), has infected humans for at date the AIDS epidemic by many thousands of years.
least a decade (F. Simon, personal communication).
Although the ancestral origin of both HIV types is well Group M occurs as the major epidemic group in the documented (Hirsch et al. 1989; Gao et al. 1999; Chen et world, whereas groups O and N are relatively rare and al. 1996), the mechanism for the transition of these SIVs are found within the range of the Central African to epidemic HIV strains is still unexplained. Most impor- chimpanzee (Gao et al. 1999). We used the prevalence of tantly, the simian hosts of the HIV ancestor lineages, Pan rare HIV-1 and -2 strains as the basis to test a mechanism troglodytes troglodytes, and Cercocebus torquatus atys, are not that produces epidemic HIV subtypes from poorly found in the same areas of Africa, but exist in natural, non- adapted, non-epidemic HIV subtypes, and employed a contiguous ranges over 1500 km apart (Kingdon 1997).
broad range of mutations, from 20 to 2000 nucleotides, to Serial passage of SIVand the origin of HIV P. A. Marx and others 913 test the genetic transition of SIV to epidemic HIV strains. Therefore, it is not necessary to know the precise number of mutations required for SIV to epidemic HIV genetic transitions, because the model allows for very few genetic changes, as well as large numbers of mutations.
A key issue in understanding the origin of HIV is whether or not SIV has the potential to become a zoonosis by making the transition from SIV to HIV in a single human exposed to simian blood. Because of our data showing multiple examples of dead-end SIV infec- tions (¢gure 1), we assume that adaptive mutations are necessary for SIV to become epidemic. If this were true, then SIV is not a zoonosis. Normally, human host defences would suppress poorly adapted SIV strains within a few weeks of infection. If the requisite genetic events needed for the transition to HIV would have taken a longer time to accrue than is normally available in a human SIV infection, then the single-infection zoonotic Therefore, we sought to establish a probability distribu- tion for the timing of the SIV to HIV transition following a single human infection. In that way, the plausibility of the scenario by which SIV to HIV transition might occur in a single infection could be examined. Very low prob- abilities would cast doubt on the underlying theory (i.e.
single infection)and lead to a search for new theories that would allow for serial passage of SIV in transition to We used a stochastic approach with continuously valued generation times, similar to the treatment of the coalescent theory approach taken by Rodrigo & Felsen- stein (1999). Our goal was to generate a probability distri- bution for the elapsed time to produce HIV, given di¡erent values of the proportion of the SIV genome that must change in order to produce HIV.
Speci¢cally, we assumed a human is infected with wild- type, replication-competent SIV. We then used binomial theory to estimate the probability of one or more progeny SIV virions in each generation carrying a genetic change Figure 1. Phylogenetic analysis was performed on a 453 bp in the direction of HIV. We used the error rate of the fragment of the gag gene. Nucleotidic sequences were reverse transcriptase as the underlying change probability aligned using the CLUSTAL W v. 1.7 program (Higgins et al.
1996), the ¢nal alignment being adjusted by eye. Genetic and conditioned on the length of the SIV genome. We distances between pairs of DNA sequences were calculated allowed the number of virions produced per generation to using Kimura's two parameter model. Phylogenetic analysis vary (table 1). Treating this probability as a hazard rate of sequences consisted of minimum evolution estimated by allowed us to estimate the time (in number of genera- the neighbour-joining method of Saitou & Nei (1987) tions)to the next HIV-productive genetic change. Finally, implemented in the CLUSTAL program, without taking gaps we cumulated the estimated times over the total number into account. The reproducibility of the branching order was of required changes to reach the total time for SIV to estimated by applying a bootstrap procedure to 100 replicates of the original data set. Phylogenetic relationships between In order to capture the full sampling variability in the SIV and HIV. (a)SIVcpz strains found in Central African process, we used resampling with a range of possible chimpanzees (GAB1 and US)are closest to HIV-1 groups M, values of generation size in calculating per-change N and O (Gao et al. 1999). SIVcpzAnt, from the East African chimpanzee, is more distant. (b)Distinct SIV strains occur in hazards. Using the exponential distribution to determine two mangabey subspecies, and only one is closely related to HIV-2. SIVrcm, found in the red-capped mangabey of equatorial Africa, is more distant from HIV-2 when compared SIVcpz and HIV-1, are each con¢ned to one subspecies of with SIVsm, which is found in mangabeys from western mangabey and chimpanzee, respectively. Although both Africa (Georges-Courbot et al. 1998). Close ancestral subspecies gave rise to HIV-1 and -2 at about the same time, relationships between SIVsm and HIV-2, and between they are found in di¡erent parts of Africa.
914 P. A. Marx and others Serial passage of SIVand the origin of HIV Table 1. Estimation of the number of acute infection days the earliest known HIV infection in Central Africa by (period of relatively high virus load) required for m mutations Zhu et al. (1998)place the ¢rst known case no later than 1959. For HIV-1 group M to emerge during this time- (CI ˆ 90% con¢dence interval, expectation is median to m mutations (Rodrigo & Felsenstein 1999).) period, the event(s)associated with an increased risk of serial passage of human SIV infections in the area of Central Africa must have occurred prior to that date. But not by many years, given HIV-1's clinical latency and time of progression to AIDS, i.e. less than a decade, or else we would have had earlier evidence of it. For western Africa and the emergence of HIV-2 the timing is perhaps We therefore searched for a biologically plausible event that occurred in this region of Central Africa within the decade before 1959 and continued to operate through time from one change to the next allowed for continuous the 1960s elsewhere in Africa. A massive increase in the generation times and avoided heavier parameterizations number of unsterile injections in sub-Saharan Africa in which might induce bias. Finally, we made assumptions this period quali¢es on all pointsöit is a speci¢c and that would be most favourable to a rapid SIV to HIV parsimonious explanation for signi¢cantly increased transition. The most signi¢cant of these was our assump- serial transfer of SIV, in both Central and western Africa tion that SIV plasma virus loads approached human during the 1950s and 1960s, respectively. We propose that HIV-1 loads during the acute infection period, before this event greatly increased the probability of serial trans- immune responses inhibited SIV replication. In this way, mission of partially adapted SIV during acute, but we would be assured that a ¢nding of a low probability of normally time-limited, SIV infections in humans.
the single-infection theory was not due to bias.
To document this event and address the regional and The number of mutations required for SIV to HIV temporal speci¢city of HIV's emergence, we assembled transition is unknown, so we took a broad approach. In and reviewed the available literature on the changes in this mutation-only illustration, there is almost no the supply and demand for injection equipment world- (55%)transition-probability density at 435 days post- wide during the 20th century. We also examined changes infection, when the number of mutations required is in the uses, costs, demand for, and availability of syringes greater than 60. At m460 mutations, the probability of and their e¡ects on injecting practices in sub-Saharan successful transition in 435 days remains small ((50%).
The probability density for duration of transition at m ˆ100 mutations is centred at 65 days and at m ˆ 200 mutations is centred at 80 days. The immune responses would greatly suppress an un¢t virus, making the SIV to The history of the hypodermic syringe is punctuated by HIV transition highly unlikely during the chronic important manufacturing and drug developments that dramatically a¡ected their availability, price, demand and use worldwide over the past century (table 2).
Following their invention in 1848 and until the end of World War I (WWI), sterile syringes were considered Serial transfer of SIV between humans would have a precision medical instruments and individually hand- markedly di¡erent outcome. In that case, m mutations made from glass and metal. The cost was high, about $50 would be achieved by allowing them to accumulate in an per unit in 1900 (adjusted to current dollars), and even by incremental and additive fashion in several persons. We 1920, after the considerable increases associated with conclude that a single, initial human SIV infection would WWI, production was still very limited (The Echo not accumulate su¤cient mutations before host immune 1991a)öonly about 100 000 syringes per year worldwide.
responses largely suppressed growth of the poorly adapted However, beginning in the period between the World viruses. These ¢ndings explain the rare occurrences of Wars, syringe manufacture was increasingly mechanized, non-epidemic subtypes of HIV-2 and suggest a similar using interchangeable components and mass production phenomenon for HIV-1 group N (Simon et al. 1998), methods. Global production reached two million per year which has existed in humans for over a decade, but has (by 1930)and eight million per year by 1952. Throughout not emerged as a signi¢cant epidemic strain thus far (F.
this period, the unit price declined steadily, by 80% Simon, personal communication). Additional serial between 1920 and 1950. At the same time the number passage would be required before these viruses could and signi¢cance of clinical applications grew, e.g. for the increased injection of insulin in the USA and Europe in the 1930s and 1940s (The Echo 1991a). But it was penicillin that drove the greatest increase in demand for injection equipment worldwide, and it did so in the 1950s.
While a precise determination of the date for the emer- While penicillin was ¢rst manufactured during WWII, gence of the HIV-1 M group is not yet possible (Goudsmit it did not become generally available (especially outside & Lukashov 1999; Korber et al. 1999), sequence data from the USA and Europe)until the early 1950s. By that time Serial passage of SIVand the origin of HIV P. A. Marx and others 915 Table 2. History of injectable medications and needle reuse in Africa: 1900^1998 injection equipment production 1 dose; 1943, in USA and Europe; injecting campaigns Central Africa campaign; 2.5Â106 account for over 3Â107 injections yr71 with injections in region adequate sterilization campaign focused on increase in unsterile persons. 1960^1964, b Including injections into the urethra.
d First documentation of transmission of VD and hepatitis from unsterile injections of penicillin.
916 P. A. Marx and others Serial passage of SIVand the origin of HIV These `single-use' syringes were never intended for sterilization or reuse. The material from which they are fabricated, polypropylene plastic, does not maintain integrity of the seals and deforms at autoclave tempera- tures and they cannot be e¡ectively sterilized unless disas- sembled, heated at temperatures above 80 8C, and reassembled under sterile conditions (The Echo 1991c).
Nonetheless, in much of the developing world, espe- cially in sub-Saharan Africa, these new syringes £ooded the market in the 1950s adding signi¢cantly to the pool of injecting equipment already in circulation. And as more injectable medications became available (especially the antibiotics)multiple reuse quickly became common prac- 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 tice, often without even attempts at sterilization (Wyatt 1984; UNICEF 1987; Van der Geest 1982; Whyte & Van der Geest 1994; CIBA 1977)(table 2).
Figure 2. The growth in global production of injecting equipment in relationship to unit price from 1898 to 1998.
Cost was adjusted to unit prices of syringes in US dollars for Although the use of injectable medications in Africa the year 1998. The almost total replacement of reusable glass and elsewhere was common by the late 19th century (The syringes occurred in the period from 1950 to 1960 and was Echo 1991a), and always included some unsterile use associated with a 100-fold growth of production from seven (Sheehan 1944; Nickum 1933), the period after WWII saw million to a thousand million. The decrease in unit price was $3.75 to $0.18 in the decade. A comparable increase in a dramatic surge in the frequency of medical injections production and decline in unit prices for penicillin occurred worldwideöand especially of injecting in the developing in the same period, which was the initial decade of mass worldömost of which were performed under unsterile availability following World War II (WWII). WWI, World conditions (Wyatt 1984; WHO 1997; Mollaret & Reilly 1947)(table 2).The consequences of this huge increase in unsterile injecting (Parascandola 1980; Reeler 1990)were the mass production of antibiotics was substantially soon evident in the worldwide rise of injection-related lowering prices for these drugs (table 2)(Mahoney et al.
transmission of infectious diseasesöincluding hepatitis 1943; Brandt 1987)and increasing their global availability (Sheehan 1944), malaria (Nickum 1933), syphilis (Parascandola 1980). As the e¤cacy of these new drugs (Mollaret & Reilly 1947)and possibly poliovirus (Wyatt became apparent, popular demand increased worldwide 1984). Sub-Saharan Africa, then entering the last decades (Wyatt 1984; Reeler 1990; UNICEF 1987). But, even with of the European colonial period (Wyatt 1984; Reeler declines in the cost of antibiotics, the injection equipment 1990; Whyte & Van der Geest 1994), was particularly needed to administer them was relatively expensive and the safe reuse of glass syringes was still dependent on a In the 75 years prior to WWII, a network of colonial costly infrastructure to ensure sterilization (Van der and missionary clinics was the principal base of Western Geest 1982)(¢gure 2). And, even at the production rate medical practice in this region of Africa (Wyatt 1984; of 150 000 reusable syringes a week (in 1952), it was Reeler 1990; Whyte & Van der Geest 1994). Speci¢c prac- impossible to meet the growing demand (The Echo tices varied, depending on the medical traditions of French, British or Belgian colonial powers (UNICEF By the 1950s manufacturing injection equipment was 1987), but most administered injectable drugs. This was increasingly a centralized industry. While even at the done on site and under medical supervision, while closely time prior to WWII there were scores of companies controlling access to the relatively costly drugs and making injecting equipment, by the 1950s Becton- injecting equipment (UNICEF 1987), using sterile Dickinson were acquiring other smaller manufacturers injecting procedures and with access to sterilization and (by 1960)were making over 50% of all the injecting equipment on hand. There is little evidence of injection- equipment manufactured in the world.
related transmission of disease in this pre-WWII colonial This increased demand was anticipated by the industry and led to a series of important changes in the conception However, in the period following WWII, with indepen- of injecting equipment, i.e. the development of inexpen- dence movements sweeping across Africa, Europe's sive, disposable `single-use' syringes. And with it a mass control of civic a¡airs in the region began to weaken manufacturing technology for plastic injection equip- (Alland 1970), including its controls on medical practice ment, dramatically lowering prices, and massively (Whyte 1982). And despite substantial new investments by increasing availability worldwide (The Echo 1991a^c) some colonial powers (Britain and France)in educational (¢gure 2). During the decade 1950^1960 sterilizable glass and administrative preparation for independence, the and metal units were largely replaced with single-use shrinking colonial medical care system was not quickly plastic syringes. This change represented a 100-fold replaced by the newly independent, but impoverished, increase in global production of syringes, up to one billion African states (UNICEF 1987; CIBA 1977; Alland 1970; units per year in 1960 and was coupled with a 56-fold decline in priceödown to $0.18 per unit when adjusted Soon the old medical care system was supplemented by a growing number of indigenous practitioners with Serial passage of SIVand the origin of HIV P. A. Marx and others 917 varying degrees of training and minimal controls, e.g. in tions used to create and validate the model may have `country clinics' using Western practices, injecting equip- adversely a¡ected its accuracy. For example, the model ment and medicines. Medications and injecting equip- ignores natural selection, which is certain to play a role.
ment (most previously used)were easily and often The model may overestimate the age of HIV-1 by diverted or salvaged from the former system. These missing mutations that were under negative selection.
formed the basis for an `informal' parallel system of Moreover, the model relies on steady-state infections.
injecting with little or no awareness of the need or the Mutations during serial passage to induce epidemic capability for sterilization procedures (UNICEF 1987; strains would be strongly driven by natural selection.
Van der Geest 1982; Whyte & Van der Geest 1994; CIBA Finally the model is only for HIV-1 group M, and does not deal with the emergence of other HIV-1 and -2 viruses. The theory does not take into account that (d) Reuse of single use syringes in mass public shorter times to transition could be achieved by serial In the 1950s, the ¢rst mass campaigns using injectable The number of mutations needed for an SIV to HIV antibiotics took place in India and Africa. In Central transition is unknown. However, recent ¢ndings in Africa (table 2)between 1952 and 1959, there were 35 animal models strongly support the serial passage million injections under UNICEF's yaws eradication mechanism for HIVs' emergence and give an estimate of campaign (UNICEF 1987). In this programme, and in the number of mutations for one gene, env, to adapt to a many subsequent antibiotic and anti-malaria treatment new host. Several in vivo passages of SHIV clone HxB2 campaigns in sub-Saharan Africa, the mass adminis- were needed to achieve necessary adaptive mutations.
tration of injectable medications by poorly trained local SHIV clone HxB2, although containing non-env genes aids using unsterile practices became the norm (Wyatt adapted for macaques, replicated to low levels in vivo 1984; Reeler 1990; UNICEF 1987; Van der Geest 1982; (Cayabyab et al. 1999). Replication of the parental virus Whyte & Van der Geest 1994; CIBA 1977; Alland 1970; was low throughout the acute infection and was undetect- able at about 15 weeks. Low levels of replication by poorly Further, throughout this period and in the following adapted viruses, and strong suppression after an acute decades, there was a sharp growth in the use of injections replication phase are key concepts in the serial passage throughout Africa (Wyatt 1984; Reeler 1990; UNICEF mechanism for HIV emergence from SIV. Pathogenicity 1987; Birungi et al. 1994; Gumodoka et al. 1996). They and replication increased 1000-fold after serial passage in were expected at every medical visit and for the treat- three macaques and animals developed AIDS. Cayabyab ment of any condition (Alland 1970). While earlier docu- et al. (1999)showed that only 12 amino-acid changes in mentation is sparse, by the mid-1960s, several studies env were required for enhanced pathogenicity. Serial establish that more than 75% of households in sub- passage, therefore, enhanced the in vivo replicative Saharan Africa had received an injection within the capacity and persistence of SHIV clone HxB2 in vivo. The previous two-week period (Birungi et al. 1994). Ethno- relevance of these ¢ndings is that three serial passages graphic and public health surveys conducted in several were required to e¡ect even a small number of adaptive parts of Africa (and India)in the 1960s found very high changes in the env gene of this poorly adapted SIV^HIV levels of injectingöin one study in Uganda 80% of hybrid virus. Poorly adapted SIV infections would face households owned their own syringe. Under these condi- tions, new to Africa at the time, the probabilities of serial Alternative events in post-colonial Africa, such as passage of any infectious agent multiplied rapidly and population growth, changing sexual practices, migration, presented a plausible mechanism in the right place at the social upheaval, increased hunting and deforestation have right time to facilitate the mutation of SIV to HIV in this also been considered as primary causes for the emergence region of Africa in the latter half of the 20th century.
of epidemic HIV (Pela & Platt 1989). Yet, despite centu- ries of forced migration and social upheaval among African peoples (Hochschild 1998; Hunt et al. 1997), epidemics of HIV did not emerge. The slave trade is Any theory of the origin of HIV must explain how ancient in Africa, and even ante-dates the massive distinct strains of SIV, which were native to dissimilar European and Arabic slave trade that took hold in the Central and West African simian species, and to which 17th century (Hochschild 1998; Hunt et al. 1997). Ulti- humans were routinely exposed for thousands of years mately, over 30 million rural people were displaced over (through bites, scratches and butchering of monkeys for a period of 400 years. Further, displacement did not stop food), could evolve into all known variants of epidemic with the end of the American Civil War in 1865, but HIV-1 and -2 in such a relatively brief evolutionary time- continued into the early 20th century in the form of frame. And, of equal importance, why they did not This slave trade originated in the same areas of Africa There is a challenge to the timing of HIV-1's emer- where epidemic HIV-1 and HIV-2 emerged in the 20th gence only in the 1950s or slightly before. Korber et al.
century. Had HIV existed, even in small numbers of (2000)proposes a model using a molecular clock to place people, it is likely that HIV would have spread as a the date of the emergence of HIV-1 group M in the 1930s sexually transmitted virus. In contrast to HIV, human to the early 1940s. The establishment of an epidemic HIV T-cell leukaemia virus appears to have spread during the long before the large increase in needle reuse would be period of slave trade (Gallo et al. 1983), showing that evidence against the hypothesis. However, the assump- retroviruses were disseminated during those voyages.
918 P. A. Marx and others Serial passage of SIVand the origin of HIV Deforestation and hunting practices may have promoted by adding diversity to an already diverse group of HIV's emergence by increasing contact between humans epidemic strains of HIV. Our ¢ndings also indicate that and SIV-infected simian species. However, these social future studies should be directed toward tracking SIV-like factors could have increased opportunities for newly infections in areas where needle reuse is widespread.
emerged epidemic strains to spread, but were not ¢rst Most signi¢cantly, the use of unsterile injection prac- causes since no mechanism for SIV adaptation was tices is still widespread and is increasing each year.
Global production of injection equipment is currently at Traditional practices, such as tattooing and clitori- 40 billion units per year (up from one billion in 1960), ectomy, existed in Africa long before the advent or rise of more injectable drugs are now available, and the intra- unsterile injecting and have been discussed as being venous use of illicit drugs (WHO 1997)is burgeoning related to the emergence of HIV. However, these pro- worldwideöresponsible for igniting explosive AIDS cedures were limited in number (occurring once a year in epidemic outbreaks in many regions of the developing many areas), have been declining in post-colonial Africa, world (Birungi et al. 1994; Gumodoka et al. 1996). And and have involved use of fresh razor blades in modern while there are a few needle exchange programmes for times (Pela & Platt 1989; Hunt et al. 1997). The limited heroin injectors in Asia, and there has been some attempt and declining prevalence of ritual cutting must be to introduce safer injecting equipment in immunization contrasted with our striking ¢nding that 80% of African programmes, WHO immunization guidelines (as recently households had experienced needle use in a two-week as 1998)still proposed up to 200 reuses of (resterilized) period by the 1960s (Birungi et al. 1994).
plastic syringes, despite the common lack of adequate Finally, in recent years, SIV contamination of oral facilities for such sterilization throughout the developing polio vaccine (OPV)and subsequent OPV mucosal expo- world (WHO 1997). If we are to avoid the emergence of sure has been invoked as a ¢rst cause of HIV's emergence new human pathogens, it is critical that we take measures (Gallo et al. 1983). This mechanism is not plausible to curtail unsterile injecting worldwide and limit the because kidneys from Asian macaques and African green serial passage of pathogens that it makes possible.
monkeys were used, and neither species naturally Evidence for proof or disproof of the model will require harbours SIVs closely related to HIV type 1 or 2 (Hirsch tracking the emergence of new HIV strains. This will et al. 1989). Therefore, contamination of OPV with both require studying regions where needle reuse and exposure mangabey and chimpanzee kidneys must be invoked.
to SIVsm or SIVcpz are common. Tracking and charac- However, these species were not routinely available for terization of SIV infections in these populations could vaccine production. Furthermore, contamination with at lead to proof of the cut-hunter hypothesis (that direct least four, and as many as nine SIVs (Chen et al. 1996; exposure to SIV through contact with SIV-infected blood Simon et al. 1998)(consider six HIV-2 subtypes and three is su¤cient to launch an epidemic strain)or the needle HIV-1 groups)from separate chimpanzee and mangabey sources would be required. Lastly, if OPV theories are correct, then it must be explained why HIV did not emerge earlier from mucosal exposures to blood and The research of P.A.M. was supported by a grant from the US tissue of hunted chimpanzees and mangabeys.
National Institutes of Health, no. AI44596. We thank Phillip Ye Unsterile needle reuse was the only mechanism for programming assistance, Lisa Chakrabarti for help with the common enough to cause serial passage of SIV. The inva- phylogenetic trees, and Dr Sol Levinson for his advice and data sive procedures in needle reuse are primarily percutaneous on the history of injecting equipment. We thank Theresa Secrist needle sticks. The risk of percutaneous needle trans- mission of HIV is 0.25%, or more (Goudsmit & Luka- shov 1999; McCray 1986). In Ippolito et al. (1993), HIV was transmitted to a student nurse from a seronegative, acutely infected patient. This is the precise mechanism implicated in this study for SIV serial transmission.
Alland, A. 1970 Adaptation in cultural evolution: an approach to Therefore, percutaneous needle reuse would be capable of medical anthropology. NewYork: Columbia University.
serially passaging SIV, if the injection event occurred Birungi, H., Asiimwe, D. & Whyte, S. R. 1994 Injection use and practices in Uganda: WHO action program on essential drugs.
during the acute phase of the SIV human infection, when Geneva, Switzerland: World Health Organization.
virus load would be relatively high. Thus, serial passage Brandt, A. 1987 No magic bullet. Oxford University Press.
of SIV during acute infections is also rare and plausible Cayabyab, M., Karlsson, G. B., Etemad-Moghadam, B. A., only with the advent of large increases in needle reuse.
Hofmann, W., Steenbeke, T., Halloran, M., Fanton, J. W., Axthelm, M. D., Letvin, N. L. & Sodroski, J. G. 1999 Changes in human immunode¢ciency virus type 1 envelope Beyond proposing an explanation of the modern glycoproteins responsible for the pathogenicity of a multiply origins of HIV, this study calls attention to a mechanism passaged simian-human immunode¢ciency virus (SHIV- for continued acceleration of the adaptation of other animal viruses to a human host. It also establishes a Chen, Z., Telfer, P., Reed, P., Gettie, A., Zhang, L. Q., Ho, critical need for alternative drug delivery systems and D. D. & Marx, P. A. 1996 Genetic characterization of a new west African simian immunode¢ciency virus SIVsm: better control of needle use, e.g. through development and geographic clustering of household-derived SIV strains with use of single-use or auto-destruct injecting equipment.
human immunode¢ciency virus type 2 subtypes and genetically Otherwise, this model suggests, new HIV strains will diverse viruses from a single feral sooty mangabey troop. J.Virol.
continue to emerge and may thwart vaccine programmes Serial passage of SIVand the origin of HIV P. A. Marx and others 919 Chen, Z. (and 10 others)1997 HIV-2 seroprevalence and char- McCray, E. 1986 Cooperative Needlestick Surveillance Group.
acterization of a new HIV-2 genetic subtype (F)within the Occupational risk of the acquired immunode¢ciency syndrome natural range of SIV infected sooty mangabeys. J. Virol. 71, among health care workers. New Engl. J. Med. 314, 1127^1132.
Mahoney, J. F., Arnold, R. C. & Harris, A. 1943 Penicillin CIBA 1977 Health and disease in tribal societies, CIBA Foundation treatment of early syphilis: a preliminary report. Am. J. Pub.
Symposium no. 49. Oxford, UK: Elsevier.
The Echo 1991aVol. 11, no. 1, Spring. Franklin Lakes, NJ: Becton- Mollaret, P. & Reilly, J. 1947 Danger of interhuman contamina- tion in serial mycotherapy: contribution to diseases caused by The Echo 1991b Vol. 11, no. 2, September. Franklin Lakes, NJ: syringe. Bill. Mem. Hosp. Paris 63, 80^82.
Muller, M. C., Saksena, N. K., Nerrienet, E., Chappey, C., The Echo 1991c Vol. 11, no. 3, December. Franklin Lakes, NJ: Herve, V. M., Durand, J. P., Legal-Campodonico, P., Lang, M. C., Digoutte, J. P. & Georges, A. J. 1993 Simian immuno- Emau, P., McClure, H. M., Isahakia, M., Else, J. G. & Fultz, P.
de¢ciency viruses from central and western Africa: evidence N. 1991 Isolation from African Sykes' monkeys (Cercopithecus for a new species-speci¢c lentivirus in tantalus monkeys. J.
mitis)of a lentivirus related to human and simian immunode- ¢ciency viruses. J.Virol. 65, 2135^2140.
Nickum, O. C. 1933 Malaria in Nebraska from contaminated Fukasawa, M., Miura, T., Hasegawa, A., Morikawa, S., hypodermic syringes. J. Am. Med. Soc. 100, 1401^1402.
Tsujimoto, H., Miki, K., Kitamura, T. & Hayami, M. 1988 Parascandola, J. 1980 The history of antibiotics. University of Sequence of simian immunode¢ciency virus from African green monkeys, a new member of the HIV/SIV group. Nature Peeters, M., Honore, C., Huet, T., Bedjabaga, L., Ossari, S., Bussi, P., Cooper, R. W. & Delaporte, E. 1989 Isolation and Gallo, R., Sliski, A. & Wong-Staal, F. 1983 Origin of human T- partial characterization of an HIV-related virus occurring cell leukaemia-lymphoma virus.The Lancet 2, 962^963.
naturally in chimpanzees in Gabon. AIDS 3, 625^630.
Gao, F. (and 10 others)1994 Genetic diversity of human Pela, A. V. & Platt, J. J. 1989 AIDS in Africa: emerging trends.
immunode¢ciency virus type 2: evidence for distinct sequence subtypes with di¡erences in virus biology. J. Virol. 68, 7433^ Reeler, A. V. 1990 Injections: a fatal attraction. Soc. Sci. Med. 31, Gao, F. (and 11 others)1999 Origin of HIV-1 in the chimpanzee Robertson, D., Hahn, B. H. & Sharp, P. M. 1995 Pan troglodytes troglodytes. Nature 397, 436^441.
Recombination in AIDS viruses. J. Mol. Evol. 40, 249^259.
Georges-Courbot, M. C. (and 11 others)1998 Natural infection Rodrigo, A. G. & Felsenstein, J. 1999 Coalescent approaches to of a household pet red-capped mangabey (Cercocebus torquatus HIV population genetics. In The evolution of HIV (ed. K. A.
torquatus)with a new simian immunode¢ciency virus. J. Virol.
Crandall), pp. 236^241. Baltimore, MD: Johns Hopkins Goudsmit, J. & Lukashov, V. V. 1999 Dating the origin of HIV-1 Saitou, N. & Nei, M. 1987 The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol.
Gumodoka, B., Vos, J., Berge, Z. A., Van Asten, H. A., Dlomans, W. M. & Borgdor¡, M. W. 1996 Injection practices Sharp, P. M., Robertson, D. L., Gao, F. & Hahn, B. 1994 in Mwanza region, Tanzania: prescriptions, patient demand Origins and diversity of human immunode¢ciency viruses.
and sterility.Trop. Med. Int. Hlth 1, 874^880.
Higgins, D. G., Thompson, J. D. & Gibson, T. J. 1996 Using Sheehan, H. L. 1944 Epidemiology of infective hepatitis: role of CLUSTAL for multiple sequence alignments. Methods unsterilized syringes.The Lancet 2, 8^11.
Simon, F., Mauclere, P., Roques, P., Loussert-Ajaka, I., Muller- Hirsch, V. M., Olmsted, R. A., Murphy-Corb, M., Purcell, R.
Trutwin, M. C., Saragosti, S., Georges-Courbot, M. C., H. & Johnson, P. R. 1989 An African primate lentivirus Barre-Sinoussi, F. & Bran-Vezinet, F. 1998 Identi¢cation of a (SIVsm)closely related to HIV-2. Nature 339, 389^392.
new human immunode¢ciency virus type 1 distinct from Hochschild, A. 1998 King Leopold's ghost. New York: Houghton group M and group O. Nature Med. 4, 1032^1037.
Tsujimoto, H., Cooper, R. W., Kodama, T., Fukasawa, M., Hunt, N. R., Liu, T. P. & Quataert, J. (eds)1997 Gendered colon- Miura, T., Ohta, Y., Ishikawa, K., Nakai, M., Frost, E. & ialism in African history. Oxford, UK: Blackwell.
Roelants, G. E. 1988 Isolation and characterization of simian Ippolito, G., Puro, V. & DeCarli, G. 1993 The Italian study immunode¢ciency virus from mandrills in Africa and its rela- group on occupational risk of HIV infection. The risk of occu- tionship to other human and simian immunode¢ciency pational human immunode¢ciency virus infection in health care workers: Italian multicenter study. Arch. Intern. Med. 153, UNICEF 1987 UNICEF in Africa south of the Sahara: a historical perspective. UNICEF History Series, monograph VI. New York, Jin, M. J., Rogers, J., Phillips-Conroy, J. E., Allan, J. S., Desrosiers, R. C., Shaw, G. M., Sharp, P. M. & Hahn, B. H.
Van der Geest, S. 1982 The illegal distribution of Western medi- 1994 Infection of a yellow baboon with simian immuno- cines in developing countries: pharmacists, drug peddlers, de¢ciency virus from African green monkeys: evidence for injection doctors and others. A bibliographic exploration.
cross-species transmission in the wild. J. Virol. 68, 8454^ WHO 1997 Product information sheets: global program for vaccine and Kingdon, J. 1997 The Kingdon ¢eld guide to African mammals (ed.
immunization. expanded program on immunization. Geneva, C. B. Robbins). San Diego, CA: Academic Press.
Switzerland: World Health Organization/UNICEF.
Korber, B. T., Sharp, P. M. & Ho, D. D. 1999 Dating the origin Whyte, S. R. 1982 Penicillin, battery acid and sacri¢ce: cures of HIV-1 subtypes (reply). Nature 22, 326.
and causes in Nyole medicine. Soc. Sci. Med. 16, 2055^2064.
Korber, B., Muldoon, M., Theiler, J., Gao, F., Gupta, R., Whyte, S. R. & Van der Geest, S. 1994 Injections: issues and Lapedes, A., Hahu, B. H., Wolinsky, S. & Bhattacharya, T.
methods for anthropological research. In Medicines: meanings 2000 Timing the ancestor of the HIV-1 pandemic strains.
and context (ed. N. L. Etkin & M. L. Tan). Quezon City, Philippines: Health Action Information Network.
920 P. A. Marx and others Serial passage of SIVand the origin of HIV Wyatt, H. V. 1984 The popularity of injections in the Third analysis of full-length genomic sequence. AIDS Res. Hum.
World: origins and consequences for poliomyelitis. Soc. Sci.
Zhu,T., Korber, B.T., Nahmias, A. J., Hooper, E., Sharp, P. M. & Yamaguchi, J., Devore, S. G. & Brennan, C. A. 2000 Ho, D. D. 1998 An African HIV-1 sequence from 1959 and Identi¢cation of a new HIV-2 subtype based on phylogenetic implications for the origin of the epidemic. Nature 391, 594^597.

Source: http://www.uow.biz/arts/sts/bmartin/dissent/documents/AIDS/rs/papers/Marx.pdf