A Brief Chronicle of Retrovirology

Retroviruses as Natural Pathogens of Vertebrates

Retroviruses were discovered at the turn of the century in two investigations devoted to neoplastic diseases in chickens. In 1908, the Danish physician-veterinarian team of Vilhelm Ellermann and Oluf Bang showed that chicken leukosis, a form of leukemia and of lymphoma, was caused by a virus. In 1911, Peyton Rous at the Rockefeller Institute in New York reported the cell-free transmission of a sarcoma in chickens (Ellermann and Bang 1908; Rous 1911). The agents discovered by Ellermann and Bang are now known under the collective term avian leukosis virus or ALV. The virus isolated by Rous bears the name of its discoverer: Rous sarcoma virus. Together, these viruses constitute the avian C-type virus genus, often referred to as avian sarcoma/ leukosis viruses (ASLV).

In due time, the discovery of virus-induced tumors was extended to mammalian species. In 1936, John Bittner established that mammary carcinoma in mice was caused by a milk-transmitted, filterable agent, and Ludwik Gross in 1957 reported on successful efforts to select a potent leukemia virus in mice by a combination of inbreeding and inoculation at an early age (Bittner 1936; Gross 1957). During the next two decades, many such viruses that cause neoplastic disease in mice, cats, cattle, and monkeys were identified (see Table 2). The number of virus-induced tumors in fowl also greatly increased—by the early 1930s, about 20 viral isolates that cause histologically distinct tumors had been reported (Claude and Murphy 1933). The list of new avian Retroviruses is still growing (Payne 1992).

Table 2

Principal Retroviruses and Their Origins.

Several of the viruses isolated during this period became important model systems, actively studied at the cellular and molecular levels to this day. The Friend and Rauscher murine leukemia viruses provided models for the study of erythropoiesis (Friend 1957; Rauscher 1962). The rodent sarcoma viruses of Kirsten, Harvey, and Moloney; the feline sarcoma virus of McDonough; the avian erythroblastosis virus of Engelbreth-Holm and Rothe Meyer; and the avian leukemia virus MC29, to name a few examples, yielded oncogenes that have critical roles in cellular signal transduction and are also important in the genesis of human tumors of nonviral origin (see Table 2) (Engelbreth-Holm and Rothe Meyer 1932; Ivanov et al. 1962; Rauscher 1962; Harvey 1964; Moloney 1966; Kirsten and Mayer 1967; McDonough et al. 1971). Studies in animals also produced evidence for the occurrence of endogenous retroviruses which initially revealed themselves by their oncogenicity for uninfected animals. Otto Mühlbock found in 1955 that MMTV-induced mammary carcinoma in mice could be inherited as well as virally transmitted, and in 1958/1959, transmissible leukemogenic agents were obtained from X-ray-induced murine leukemias (Mühlbock 1955; Lieberman and Kaplan 1959; Latarjet and Duplan 1962).

The first description of what later turned out to be a lentiviral disease, equine infectious anemia, goes back to before the turn of the century (Vallée and Carré 1904), although recognition of the viral origin of this disease came later. Visna, a neurological disease in sheep caused by a lentivirus, was described in 1957 and gave rise to the concept of slow (Latin: lentus, slow) viral infections (Sigurdsson 1954). Several lentiviruses that can induce immunodeficiency in various species of mammals, including monkeys and cats, have been isolated in recent years and serve as models for HIV infection (Letvin et al. 1985; Pedersen et al. 1987).

The history of the spumavirus discovery is unusual in that the isolates did not come from sick animals but from cell cultures prepared from healthy monkeys. Spumaviruses (Latin: spuma, foam) induce a characteristic foamy appearance in the cytoplasm of cultured cells; the first isolate, simian foamy virus, was made from macaque cultures in 1954 (Enders and Peebles 1954; Rustigian et al. 1955).

The search for and isolation of animal retroviruses continue to this day. As recent examples show, spontaneous tumors in animals form a rich and far from exhausted resource for naturalists to cast their nets. The reward can be significant for our knowledge of viral replication and of genes that have important roles in cell growth and differentiation (Maki et al. 1987; Mayer et al. 1988; Souyri et al. 1990; Kawai et al. 1992).

Quantitation of Infectivity and Oncogenicity

In the first five decades after the discovery of Retroviruses, infectivity and oncogenicity were commonly titrated in the animal host, a costly, time-consuming, and inaccurate procedure. A first step toward direct quantitation was accomplished with the inoculation of RSV onto the chorioallantoic membrane of the chick embryo (Keogh 1938). Individual small tumors appeared and their numbers could be approximately correlated with viral concentrations, but great variability in sensitivity existed between embryos. It was not until the 1950s that cell culture became an effective and precise tool in animal virology, allowing the kind of quantitative analysis that had been practiced successfully with bacteriophage. In 1958, Howard Temin and Harry Rubin applied the powers of cell culture to retrovirology by showing not only that RSV caused oncogenic transformation of chick embryo fibroblasts in culture, but also that single viral particles could induce discrete foci of transformed cells (Temin and Rubin 1958). The numbers of these microtumors in culture are directly proportional to the viral inoculum: A sensitive and quantitative assay for infectivity and tumorigenicity had been created.

The focus assay facilitated important advances in the study of the viral life cycle and of viral oncogenes and led to the discovery of the defectiveness in replication of acutely transforming oncoviruses and their dependence on a companion nontransforming helper virus. The analysis of defectiveness in turn uncovered helper virus functions that reside in normal cells and produced early evidence for the occurrence of Endogenous retroviral information in the genomes of normal cells. The focus assay also allowed the isolation of conditional and nonconditional viral mutants affecting replicative and oncogenic characteristics. These mutants, in turn, prepared for the discovery and isolation of oncogenes. Finally, the focus assay was instrumental in demonstrating recombination between related Retroviruses. Similar quantal assays, based on oncogenic transformation in cell culture, were developed for all retroviruses that transduce an oncogene (Baluda and Goetz 1961; Hartley and Rowe 1966; Rosenberg et al. 1975). Even some of the retroviruses that fail to transform cells in culture can be assayed by their focal effects on a cell monolayer, such as formation of fused giant cells, or syncytia, or plaques due to killing of infected cells (Klement et al. 1969; Graf 1972).

Defectiveness in Replication

Under ideal conditions, a focus of transformed cells in culture originates from the infection of a single cell by a single viral particle. In such single infections, almost all acutely transforming oncogenic retroviruses fail to produce infectious progeny, although they convert a normal cell into a cancer cell. This defectiveness in replication was detected first with the Bryan Hightiter strain of RSV (Hanafusa et al. 1963; Temin 1963b). It results from the substitution of essential virus-coding sequences by a cell-derived oncogene in the viral genome. Superinfection of the transformed nonproducer cells with a related replication-competent but nontransforming retrovirus leads to the synthesis of infectious oncogenic viral progeny. The replicative functions missing in the acutely oncogenic virus are complemented at the phenotypic level by the helper virus. The interaction between helper and defective virus can be viewed as an extreme case of phenotypic mixing. The oncogenic progeny virus remains genetically defective, and all viral proteins for which genetic information is lacking are provided in trans by the helper (Hanafusa 1965). The model provided by the defective oncogene-containing viruses provided the biological precedent for the development of retroviral vectors (Mann et al. 1983).

Endogenous Retroviruses

Some nonproducer cells infected and transformed by the Bryan Hightiter strain of RSV can spontaneously begin to release infectious virus without the addition of an exogenous helper. An analysis of this viral production showed that certain types of normal cells can supply helper functions endogenously, indicating not only the presence, but also the expression of retroviral genetic information (Dougherty et al. 1967; Vogt 1967; Weiss 1967; Hanafusa et al. 1970; Weiss and Payne 1971). Endogenous retroviral genomes have indeed been found in all vertebrates where they have been seriously looked for, including humans (Aaronson and Stephenson 1976; Hughes et al. 1981; Repaske et al. 1985; Larsson et al. 1989; Löwer et al. 1993). They occur in expressed and in silent forms, as complete or as partially deleted defective viral genomes. Endogenous proviruses are present in somatic and germ cells alike as evidenced by genetically defined chicken and mouse lines in which inheritance of specific endogenous proviruses has been followed (Payne and Chubb 1968; Robinson 1978; Pincus 1980). They are transmitted from parent to offspring in the form of single dominant Mendelian loci. Complete retroviral genomes Endogenous to normal chicken or mouse cells can be induced by irradiation of the cell or exposure to demethylating agents; these genomes then direct the synthesis of infectious virus (Rowe et al. 1971; Weiss et al. 1971). Production and release of endogenous virus can also occur spontaneously (Vogt and Friis 1971; Levy 1978; Pincus 1980). All endogenous retroviral genomes belong to the simple category. So far, no endogenous lenti-, spuma-, or HTLV-like viruses have been identified.

The Virion

The adaptation of modern cell culture techniques to retrovirology was also accompanied by a steady improvement in the methods of viral growth and virion purification. Analysis of viral proteins from infected cells and from mature viral particles and information generated by the study of viral mutants defined three principal groups of virion proteins (Fleissner 1971; Oroszlan et al. 1971; Schäfer et al. 1972; August et al. 1974). Located at the virion surface are glycosylated envelope (Env) proteins. The structural proteins of the matrix and nucleocapsid core are nonglycosylated and are referred to as Gag proteins. The acronym Gag is derived from group-specific antigen, in reference to the cross-reactive immunological tests in which a single antiserum was able to detect related retroviruses infecting the same host species. Associated with nucleocapsid and RNA are the polymerase (Pol) proteins that are responsible for reverse transcription and integration.

An important advance in the understanding of viral gene expression and assembly of progeny virus came with the discovery that the primary translational Products of retroviruses are three polyproteins translated from polycistronic messages corresponding to Gag, Pro, Pol, and Env sequences (von der Helm 1977; Eisenman and Vogt 1978). These are cleaved proteolytically to generate the functional virion components. The processed Gag proteins are referred to as MA (matrix), CA (capsid), and NC (nucleocapsid). The viral protease, PR, cleaves Gag, Pol, and sometimes Env precursors. It is encoded by the pro gene and maps between gag and pol. The Pol cleavage products are RT (reverse transcriptase) and IN (integrase). The Env precursor is processed by a cellular enzyme into SU (surface) and TM (transmembrane) proteins that are linked by disulfide bonds or noncovalent interactions (Leis et al. 1988; Hunter and Swanstrom 1990; Oroszlan and Luftig 1990). Translation of the polycistronic messages of retroviruses is subject to controls that include suppression of termination and ribosomal frameshifting at the intersections between the gag, pro, and pol genes (Jacks 1990; and Chapter 7).

Virion RNA sediments as a complex of 60–70 Svedberg (S) units (corresponding to about 20–30 kb of RNA) in the ultracentrifuge; it can be dissociated by heat into 30 S (7–12 kb) components (Robinson et al. 1965; Duesberg 1968). This observation immediately raised the question of whether the smaller components are functionally and genetically identical or whether they are different, performing distinct, nonoverlapping tasks in the viral growth cycle. The answer came from a determination of the genetic complexity of retroviruses by RNA fingerprinting. The results favored the lower 30 S limit for the whole genome, corresponding to 7–12 kb in size, and provided evidence for diploidy of Retroviruses (Beemon et al. 1974; Billeter et al. 1974). RNA fingerprinting of viral genomes with specific deletions and of recombinants defined the basic gene order as 5′-gag-pol-env-3′ (Joho et al. 1975; Coffin and Billeter 1976; Wang et al. 1976a,b,c).

Viral Mutants

The focus assay also provided the technical basis for the first genetic experiments with retroviruses. Much of the early work was done with nondefective RSV, which is both strongly transforming-competent and replication-competent. The ability to clone virus by picking single foci of transformed cells and to work with the progeny of single viral particles was a prerequisite for the isolation and characterization of conditional and nonconditional viral mutants (Temin 1961; Toyoshima and Vogt 1969; Martin 1970; Hanafusa et al. 1972; Wyke 1973). RSV mutants defined viral gene functions, provided markers for construction of the genetic map, and was used to reveal the cellular origin of retroviral oncogenes. The mutants could be grouped into two basic categories, one—in the oncogene—affecting only virus-induced transformation and oncogenesis and the other—in gag, pol, or env—interfering primarily with viral growth. The characteristics of these mutants allowed the important conclusion that transformation and replication were separately coded functions and provided the first genetic definition of retroviral oncogenes.

Mutants of RSV that were temperature-sensitive for focus formation in cultures of chicken embryo fibroblasts demonstrated unequivocally that both the induction and maintenance of the transformed phenotype depended on the activity of a viral gene (Toyoshima and Vogt 1969; Martin 1970; Kawai and Hanafusa 1971). Nonconditional transformation-defective mutants of RSV were found to have suffered a deletion of the src oncogene. These mutants were still able to produce infectious nontransforming progeny, marking the oncogene as dispensable for viral growth and survival (Duesberg and Vogt 1970; Vogt 1971a). The congenic pair of RSV and its derivative src deletion mutant later provided the raw material for the synthesis of a specific src cDNA probe, the tool that traced the origin of the src gene to the genome of the normal cell.

Genetic Recombination

Biological cloning in cell culture and viral host range markers defined through the use of genetically resistant cells played a decisive part in the discovery of stable genetic exchanges between related retroviruses (Vogt 1971b; Kawai and Hanafusa 1972). The nature of these genetic interactions was not immediately clear. When the initial experiments were done, the complexity of the genome had not been resolved, and there was still a possibility that it consisted of independently replicating segments that could reassort at the observed high frequencies. Only later did molecular analysis of recombinant RNA reveal true crossing over as the mechanism of retroviral recombination, favored by the diploidy of the nonsegmented retroviral genome (Beemon et al. 1974; Joho et al. 1975).

Provirus and Reverse Transcriptase

The molecular tools available throughout the 1960s were primitive by today's standards. Many conclusions concerning the replication of animal viruses were based on the effects of metabolic inhibitors. For retroviruses, the use of such inhibitors created a paradox: The replication of these RNA viruses was widely assumed to follow the model of other single-stranded RNA viruses such as poliovirus, where replication involves RNA-dependent RNA synthesis carried out by a virus-coded enzyme that creates partially double-stranded RNA replicative forms. Yet no double-stranded viral RNA species could be detected in retrovirus-infected cells. Moreover, retroviral replication was found to be sensitive to inhibitors of DNA synthesis and DNA-directed RNA synthesis (Temin 1963a; Bader 1965). In 1964, Howard Temin provided a coherent explanation for these puzzling data by proposing the provirus hypothesis, which postulates the generation of a DNA copy of the viral genome and its subsequent integration into the cell genome (Temin 1964). This radical idea gained general acceptance only with the discovery of reverse transcriptase in retroviral virions (Baltimore 1970; Temin and Mizutani 1970). Further evidence in support of the provirus hypothesis came from DNA transfection experiments. Total DNA extracted from RSV-transformed cells and introduced into uninfected recipients induced oncogenic transformation and, under appropriate conditions, synthesis of complete RSV (Hill and Hillova 1972).

The discovery of reverse transcriptase was a watershed event. The specific enzymatic and mechanistic problems posed by reverse transcription of the RNA viral genome could now be intensely studied. The primer for initiation of DNA synthesis was identified as a species of cellular tRNA (Harada et al. 1975; Peters and Dahlberg 1979). Synthesis of the viral DNA was shown to start near the 5′end of the virion RNA template (Taylor and Ilmensee 1975). To continue reverse transcription, the polymerase-nascent DNA complex must jump to the 3′end of virion RNA, a feat made possible by the RNase H activity of reverse transcriptase (Moelling et al. 1971) and by the existence of short direct repeats at the 5′and 3′ends of the genome (Coffin and Haseltine 1977; Stoll et al. 1977). The initial product of reverse transcription is an RNA-DNA duplex; RNase H digests the RNA template, allowing newly synthesized DNA to base pair with the repeat at the 3′end of the genome. A second similar template jump was found to take place later in the process of reverse transcription (Gilboa et al. 1979) to produce a complete proviral copy with two LTRs. Integration was found to take place into a very large number of sites in the host genome, without circular permutation of the proviral sequence (Hughes et al. 1978; Steffen and Weinberg 1978). Structure and function of proviral RNA transcripts were determined as well as their mechanism of synthesis, the essential features being the synthesis of full-length and spliced RNA copies with characteristics of host mRNAs (Hayward 1977; Weiss et al. 1977). The former serve as progeny genomes and as mRNA for Gag, Pro, and Pol proteins; the latter function as messages for Env proteins and, in the case of the complex retroviruses, for regulatory proteins. Cloning and sequencing of DNA in the early 1980s greatly expanded the technical possibilities of the field, and the first complete sequence of a retrovirus, Moloney murine leukemia virus, was published in 1981 (Shinnick et al. 1981).

The Cellular Origin of Oncogenes

Early studies on reverse transcriptase had shown that cDNA copies of the viral genome could be synthesized in preparations of purified virions. Although the DNA transcripts were much smaller than the virion RNA template, they were representative of all sequences in the genome. Being able to generate short cDNA copies that contained sequences from the entire genome allowed the construction of probes for specific genes by the use of defined deletion mutants as selection agents. A probe specific for the src oncogene was obtained in this way (Stehelin et al. 1976a) and was found to hybridize to sequences in normal cellular DNA (Stehelin et al. 1976b). This surprising result showed that src was not originally a retroviral gene, but a gene of cellular origin carried in the viral genome and transduced by the virus from cell to cell. In short order, all retroviral oncogenes were found to be recent acquisitions from the cell. Many were eventually identified as cellular genes with normal functions in mitogenic signal transduction: coding for peptide growth factors, growth factor receptors, protein kinases, G proteins, and transcription factors (Bishop 1983; Varmus 1984; Weinberg 1989; Cooper 1990).

The discoveries at the nucleic acid level were complemented by important advances at the protein level. The product of the src oncogene was identified as a 60-kD phosphoprotein by immunoprecipitation using sera from rabbits bearing RSV-induced tumors, and translation in vitro from src mRNA (Brugge and Erickson 1977; Beemon and Hunter 1978; Purchio et al. 1978). The src gene product was soon found to be a protein kinase, specifically one that targets tyrosine (Hunter and Sefton 1980). Tyrosine kinases are now known to have a central role in cellular signal tranduction (Chapter 10).

The discovery of the cellular origin of retroviral oncogenes also proved to be key to unlocking the origins of human cancer in the absence of viruses. The question was: If normal cells contain Protooncogenes, can they be made oncogenic by mutation alone, without being acquired by a virus? The first suggestion that this might be the case came from the analysis of tumors induced by ALV, a virus that causes lymphoma in a high percentage of infected chickens but does not contain an oncogene. Tumors induced by this virus were found to carry a provirus inserted within a common region of the protooncogene c-myc, previously identified as the cellular counterpart of the oncogene of several avian retroviruses (Hayward et al. 1981; Payne et al. 1982). The insertions were quickly shown to lead to deregulated expression of c-myc, demonstrating that oncogenes could be activated by juxtaposition of protooncogene and regulatory sequences without the requirement of incorporation into a virus. This precedent led other investigators to look among chromosomal rearrangements in human tumors for evidence of DNA rearrangements involving c-myc, and several examples were rapidly uncovered (Dalla-Favera et al. 1982; Marcu et al. 1983). Indeed, chromosomal rearrangements that place Protooncogenes under strong cellular transcriptional control signals or create oncogenic fusion proteins are a common feature of many naturally occurring malignancies (Gale and Canaani 1984; Groffen et al. 1984).

The second landmark discovery in this area was made with an extension of the technique which had demonstrated that the DNA of RSV is biologically active: Transfection of normal cells with DNA extracted from cells infected with RSV produced foci (Hill and Hillova 1972; Shih et al. 1979; Krontiris and Cooper 1981). Now the sources of DNA were nonviral tumors, including human cancers and tumor cell lines. In recipient mouse cells, such DNA induced neoplastic transformation. The effect was traced to a single, dominant gene that in serial DNA transfections retained the ability to confer the malignant phenotype upon the recipient cell. In several types of human tumors, this gene was identified as the mutated cellular form of the retroviral oncogene ras, previously studied in a murine sarcoma virus (Der et al. 1982; Parada et al. 1982; Santos et al. 1982).

With these discoveries, oncogenes appeared on the scene as candidate causative factors of human cancer. Transfection studies characterized oncogenes as dominant effectors of neoplastic transformation. Further analysis of individual oncogenes and their transforming potential revealed cooperative effects between oncogenes, such as myc and ras, introduced in the same cell (Land et al. 1983). The often inefficient transformation of normal cells freshly cultured from an animal organism and the enhanced oncogenicity of certain combinations of oncogenes probably reflect the multistep nature of carcinogenesis. For the induction of full oncogenic transformation, a change in a single gene is generally insufficient; multiple genetic events that promote growth-promoting genetic events are required for the development of a malignant tumor. Several human tumors show a high incidence of genetic change in specific oncogenes, and in some cancers, such as Burkitt's lymphoma or chronic myelogenous leukemia, a specific change involving an oncogene occurs in virtually all cases (Haluska et al. 1987). Such consistent deregulation of growth-promoting genes very likely has an etiological role in carcinogenesis.

The study of oncogenes has blossomed into a major field of biology and has moved far from its retroviral roots. The production of transgenic animals and the germ line inactivation of genes by homologous recombination added new dimensions to oncogene research (Palmiter and Brinster 1986; Sinn et al. 1987; Mansour et al. 1988). Investigations with these techniques firmly established the causative role of mutated and ectopically expressed cellular oncogenes in cancer and uncovered often unexpected functions of these genes in embryonal development and differentiation.

Human Infections: The Challenge to Retrovirology

The isolation of retroviruses from tumors in higher mammals in the 1960s intensified the search for human cancer viruses. These efforts were supported by the Special Virus Cancer Program of the U.S. National Cancer Institute, initiated in 1971. This program greatly enhanced basic knowledge of animal tumor viruses at the organismic, cellular, and molecular levels but was not immediately successful in one of its primary missions, namely, finding an oncogenic human retrovirus. Although several retroviral isolates from human material were described, closer scrutiny relegated all of them to the category of contaminants from animal sources (Weiss 1982). By the end of the 1970s, Retroviruses had been searched for without success in most types of human tumors. Despite the universal occurrence of these viruses in a number of mammalian species, including some primates, it seemed questionable whether a human retrovirus existed at all. However, persistence in the face of declining odds was eventually rewarded.

The initial clue came in 1977 from clinical and epidemiological studies that revealed an unusually high frequency and a suspicious clustering of adult T-cell leukemia in areas of the Kyushu and Shikoku islands of Japan (Takatsuki et al. 1977; Uchiyama et al. 1977). These epidemiological patterns were suggestive of a transmissible leukemogenic agent. The isolation and identification of that agent, however, depended on a technical breakthrough, the long-term cultivation in vitro of human T lymphocytes (Morgan et al. 1976; Ruscetti et al. 1977). Success in the culture of T cells depended on the discovery of the T-cell growth factor, now known as interleukin-2 (IL-2). Cultures of leukemic T cells then were the source of the first human oncogenic retrovirus, human T-cell leukemia virus 1 (HTLV-1) (Poiesz et al. 1980; Yoshida et al. 1982). HTLV-1 is the causative agent of the aggressive T-cell leukemia of southern Japan and parts of the Caribbean.

HTLV-1 soon became the prototype of the complex retroviruses. Cloning and sequencing of its viral genome showed that HTLV-1 lacked a cell-derived oncogene, yet it was more complex than other oncogenic retroviruses (see Seiki et al. 1983). Coding information for new types of retroviral proteins, then referred to as X proteins and now termed Tax and Rex, was discovered. Subsequent analysis revealed several nonvirion regulatory proteins in the HTLV-1 genome which thus became the first example of a complex retrovirus (see Fig. 1).

HTLV-1 infection is strongly linked to adult T-cell leukemia epidemiologically. A similar link of HTLV-1 was discovered to a disease of the central nervous system, tropical spastic paraparesis (TSP) (Gessain et al. 1985; Osame et al. 1986). Adult T-cell leukemia is seen in less than 1% of individuals who show immunological evidence of past infection with HTLV-1; the percentage for TSP is even smaller. HTLV-1 infection therefore appears to be necessary but not sufficient to cause hematopoietic or neurological disorders. In adult T-cell leukemia, the HTLV-1 provirus was found to be integrated at the same chromosomal site in all leukemic lymphocytes from a given patient, but at different sites in cells from different patients (Wong-Staal et al. 1983; Yoshida et al. 1983). This integration pattern demonstrated both the clonal origin of the disease and absence of insertional activation of a cellular oncogene. The two well-known mechanisms of retroviral Oncogenesis, transduction and cis-activation of an oncogene, therefore did not apply to HTLV-1. A plausible scenario, in accord with available data, is that Leukemogenesis is caused by expression of a viral regulatory protein, probably Tax. Tax can act as a transcriptional regulator and has been shown to affect expression levels of several host genes (Sodroski et al. 1984; Feuer and Chen 1992). However, the low incidence and long latency of this disease implies collaboration of other factors or genetic changes.

About the time of the isolation of HTLV-1, an epidemic of AIDS arose in several developed countries. Again, the ability to culture human T cells played a critical part in isolating a candidate retrovirus variously referred to by its discoverers as lymphadenopathy-associated virus (LAV), HTLV-3, or AIDS-related virus (ARV) (Barré-Sinoussi et al. 1983; Gallo et al. 1984; Levy et al. 1984b), but now known as human immunodeficiency virus (HIV). Cloning and sequencing of the HIV genome showed that this virus was of the complex variety, capable of coding for several regulatory proteins in addition to the standard set of Retroviral virion proteins. HIV was rapidly and firmly linked to the causation of AIDS: Epidemiological data showed a strong correlation with the disease (Gallo 1988, 1990). Sexual intercourse was found to be the major mode of transmission, and promiscuity defined major risk groups. Blood-borne transmission was also recognized as an important factor in the spread of HIV and responsible for the high rates of infection among intravenous drug users and hemophiliacs. Blood-borne infections provided dramatic proof of the etiological role of HIV in AIDS: Transfusions given in the course of surgery at the beginning of the AIDS epidemic but before contamination with HIV could be tested led directly to the transmission of the disease, as did administration of HIV-tainted coagulation factors to hemophiliacs. Such accidental infections had invariably tragic consequences. Screening for HIV virtually eliminated such iatrogenic infections. Key elements of AIDS pathogenicity were recognized early: drastic depletion of CD4-positive lymphocytes, and the consequent heightened susceptibility to opportunistic microbial and parasitic infections, severe neurological complications, and high incidences of lymphoma and Kaposi's sarcoma (Gottlieb et al. 1981; Safai and Koziner 1985; Navia et al. 1986). At this time, neither protective vaccine nor curative therapy is available for HIV infection. The immense human suffering caused by the continued spread of HIV constitutes the greatest challenge to retrovirology.