|Home | About | Journals | Submit | Contact Us | Français|
Foamy viruses (FV) are complex retroviruses which are widespread in many species. Despite being discovered over 40 years ago, FV are among the least well characterized retroviruses. The replication of these viruses is different in many interesting respects from that of all other retroviruses. Infection of natural hosts by FV leads to a lifelong persistent infection, without any evidence of pathology. A large number of studies have looked at the prevalence of primate foamy viruses in the human population. Many of these studies have suggested that FV infections are prevalent in some human populations and are associated with specific diseases. More recent data, using more rigorous criteria for the presence of viruses, have not confirmed these studies. Thus, while FV are ubiquitous in all nonhuman primates, they are only acquired as rare zoonotic infections in humans. In this communication, we briefly discuss the current status of FV research and review the history of FV epidemiology, as well as the lack of pathogenicity in natural, experimental, and zoonotic infections.
Foamy viruses (FV), or Spumaviruses, comprise one of the seven genera of Retroviridae. They have a distinct morphology and bud primarily from the endoplasmic reticulum rather than the plasma membrane (Fig. (Fig.1).1). The particles are characterized by an immature-looking core with an electron-lucent center and very distinct glycoprotein spikes on the surface. Since their discovery in a human-derived cell culture in 1971, FV have been implicated in a variety of human pathologies. In this review we examine FV epidemiology and the putative relationship between FV infection and human disease.
FV are considered complex retroviruses because they encode viral proteins which are not incorporated into viral particles. The prototype FV genome, SFVcpz(hu), which has been most commonly designated human FV in the literature, is depicted in Fig. Fig.2.2. Throughout this review, the nomenclature presented in Table Table11 is used. Such nomenclature has been proposed because it clearly indicates the origin of each FV isolate. The FV genome encodes the canonical retroviral gag, pol, and env genes, as well as at least two additional genes termed tas (bel-1) and bet. Although a number of putative functions have been proposed for the SFVcpz(hu) bet gene, its role in vivo is still unknown; it is dispensable for replication in tissue culture. In contrast, the tas gene, for transactivator of spumavirus, is required for viral replication 68. The tas gene encodes a protein which functions to transactivate the long terminal repeat (LTR) promoter, which in the absence of Tas is transcriptionally silent 77, 87. In addition to the LTR, FV contain a second promoter, termed the internal promoter (IP) because it is located within the env gene 23, 65, 67, 75. The IP drives expression of the tas and bet genes but, unlike the LTR, has significant basal transcription activity 66. In a manner similar to the LTR, the IP is also transactivated by Tas protein. Studies in infected fibroblasts indicate that early in infection, the IP drives expression of tas and bet mRNAs 66. Once these mRNAs are translated and Tas protein is synthesized, transcription begins from the LTR, leading to the accumulation of gag, pol, and env mRNAs and ultimately to new viral particles.
FV structural genes have features distinct to this genus. The Gag protein is not efficiently cleaved into the mature viral proteins seen in most other retroviruses, leading to the immature morphology. The Pol precursor protein is only partially cleaved; the integrase domain is removed by viral protease, but the protease and reverse transcriptase domains apparently are found in the same molecule. The Env protein is cleaved into surface and transmembrane domains, as in other retroviruses. However, the Env protein contains an endoplasmic reticulum retention signal which contributes to its site of budding 39.
FV replication differs from that of conventional retroviruses like type C oncoviruses and lentiviruses in many, sometimes remarkable, ways. Interestingly, in some respects FV replication more closely resembles that of the other family of reverse transcriptase-encoding viruses, the Hepadnaviridae (reviewed in reference 63). An outline of the FV life cycle is given in Fig. Fig.3.3. Some of the unique aspects of FV replication are denoted with red text and arrows, while those shared by FV and conventional retroviruses such as human immunodeficiency virus (HIV) are denoted in black. Reverse transcription occurs at a late step in the replication cycle, resulting in infectious SFVcpz(hu) particles containing DNA rather than RNA 79, 130. Efforts to strictly classify FV as either DNA or RNA viruses are problematic. Although the infectious genome of SFVcpz(hu) is DNA, viral RNA is specifically packaged into virions (Fig (Fig3).3). Furthermore, only SFVcpz(hu) isolates have been shown to contain an infectious DNA genome. During viral maturation, cleavage of the Gag protein by viral protease is incomplete, and mature particles do not contain matrix, capsid, and nucleocapsid proteins but instead comprise two large Gag proteins differing by ca. 4 kDa at the carboxy terminus. Particles bud from cells primarily through the endoplasmic reticulum 37, 38, and in the absence of Env glycoproteins, particles are not released from cells 14, 31. The cellular receptor for FV infection is not known but must be ubiquitous, since FV have wide tropism (see below).
In vitro, FV can infect most cell types from a wide variety of species 47, 76. However, the result of infection can be markedly different depending on the cell type. Infection of many cell types leads to the outcome resulting in the characteristic foamy cell appearance which gives these viruses their name. Infected cells undergo rapid syncytium formation, cytoplasmic vacuolization, and cell death. Fibroblast cell lines such as human embryonic lung cells, baby hamster kidney cells, and Mus dunni tail fibroblasts are particularly sensitive to the cytopathic effects (CPE) of FV infection.
In contrast, it is possible to chronically infect a variety of cells lines in culture. Transformed cell lines of myeloid, erythroid, and lymphoid origin have all been persistently infected 76, 78, 129. However, primary human lymphoid cells were found to be sensitive to the FV CPE, while in contrast, monocytes were infected with FV but showed no CPE 78. Chronic infections in vitro result in continual virus production in the absence of notable cell death but, compared to cytopathic infections, produce significantly lower viral titers 129. The factors, either viral or host, which mediate a lytic versus chronic infection in vitro are not well understood. A number of reports indicate that increased expression of Bet may enable chronic infection 93, 95, 127 as well as block superinfection 16. However, the role of Bet in establishment of a chronic infection is not completely clear, as Bet− viruses can also establish chronic infections in vivo 21.
The first description of an FV was published in 1954, when it was found as a contaminant in primary monkey kidney cultures 27. The first isolate of this “foamy viral agent” was obtained in 1955 91. Shortly thereafter, FV were isolated from a wide variety of New and Old World monkeys, cats, cows, and prosimians (Table (Table11).
It was not until 1971 that a putative human FV was isolated. In this case, a viral agent with FV-like characteristics was isolated from lymphoblastoid cells released from a human nasopharyngeal carcinoma (NPC) isolated from a Kenyan patient 3. This isolate was termed human FV because of its origin and became the prototypic laboratory strain when it was subsequently sequenced and infectious molecular clones were derived 68, 85, 86. Shortly after the isolation of SFVcpz(hu), the original authors concluded that SFVcpz(hu) was a distinct type of FV most closely related to Simian FV (SFV) types 6 and 7, both isolated from chimpanzees 28. In a contradictory report, Brown et al. concluded that SFVcpz(hu) was not a distinct type of FV but rather a variant strain of chimpanzee FV 20. The origin of SFVcpz(hu) was debated until 1994, when SFVcpz was cloned and sequenced 46. The sequence showed that there is 86 to 95% amino acid identity between SFVcpz and SFVcpz(hu) 46, indicating that SFVcpz(hu) is likely a variant of SFVcpz and not a unique isolate. Further phylogenetic analysis of the pol regions of these genomes indicate 89 to 92% nucleotide homology and 95 to 97% amino acid homology between SFVcpz(hu) and the various SFVcpz strains 101. Recently, sequence comparisons between the original SFVcpz(hu) isolate and SFV from four distinct subspecies of chimpanzee demonstrate that SFVcpz(hu) is most closely related to FV from Pan troglodytes schweinfurthii, whose natural habitat includes Kenya (W. Heneine, W. M. Switzer, G. Shanmugam, P. Sandstrom, J. Ely, M. Peeters, L. Chapman, and T. M. Folks, submitted for publication). Taken together, these results clearly indicate that SFVcpz(hu) is a variant FV of chimpanzee origin. Since the individual from whom SFVcpz(hu) was isolated lived in proximity to P. troglodytes schweinfurthii, it is possible that the virus was acquired as a zoonotic infection. However, it is not possible to rule out laboratory contamination as the source of SFVcpz(hu).
The prevalence of FV infection in naturally infected animals is generally high and varies widely depending on the species and environmental conditions (Table (Table1).1). Some species, such as rhesus macaques 58, African green monkeys 102, chimpanzees 52, and cats 124, harbor closely related yet serologically distinct subtypes of FV. Seroprevalence is generally higher in animals housed in captivity, reaching 100% compared to animals studied in the wild. For example, a 1995 study showed that 93% of African green monkeys housed in captivity were infected with FV 103, while only 36% of wild animals tested positive in a separate study 109. The prevalence data presented in Table Table11 indicate that FV efficiently infect not only nonhuman primates but cats and cows as well. Unconfirmed FV infections have also been reported for hamsters 53, sheep 32, and sea lions 59.
The precise mechanisms of FV transmission are not well understood. However, the data collected thus far indicate that a saliva-based means of transmission, such as licking or biting, is most probable. The earliest controlled study of FV transmission was conducted in cattle, in which bovine FV (BFV) infection is endemic 55. This study showed that newborn calves had passive immunity, which was lost by 3 to 5 months of age. All calves were BFV negative at birth, but once placed in contact with infected adults, approximately 35% became infected within 10 weeks and 85% by 3 years. A similar study was performed with captive baboons 18. As in cattle, infants were initially FV negative and had passive maternal antibody. By 15 months of age, 1 of 10 juveniles tested positive for FV. When adults were analyzed, 100% tested positive. Additionally, SFV antibodies can be found in juvenile animals prior to sexual maturity, indicating a nonsexual route of transmission (J. Allan, personal communication).
Recently, a comprehensive study of feline FV (FFV) infection in age and sex-matched domestic and feral cats was performed 125. In domestic cats, the study found a gradual increase in FFV infection over time, but no differences between males (56%) and females (57%). Feline immunodeficiency virus (FIV) infection, however, was significantly higher in male cats (16%) and highest in nondesexed males (35%), consistent with aggressive behavior, such as biting, as a mode of transmission. When feral cats were analyzed, FFV infection was lower in males (23%) than females (52%), while FIV infection was higher in males (14%) than in females (3%). FFV transmission does not correlate well with aggressive or sexual behavior but does correlate well with nonaggressive social behavior, such as grooming.
As early as 1955, Rustigian et al. infected monkeys, rabbits, chicks, and mice with “foamy agents” from rhesus macaque kidney cultures 91. Disease was demonstrated in newborn mice, but because the authors were attempting to adapt Dengue virus to monkey cells, it is likely that this result was due to Dengue virus infection and not FV.
In 1974, Johnston infected rabbits with SFVmac via intradermal inoculation 57. The only pathology noted by the author was inflammation and necrosis at the site of injection. Very low titer virus could be recovered from the dermal lesions as well as from tissue kidney explant cultures up to 1 year postinfection. A second study of SFVmac infection of rabbits was performed by Swack and Hsuing in 1975 112. This report demonstrated that rabbits could be inoculated by either intraperitoneal or intranasal inoculation. Virus could be recovered at all time points up to 264 days postinfection by cocultivation from spleen, liver, kidney, lung, salivary gland, brain, and blood cells, but never from serum. Tissue distribution of FV in the experimentally infected rabbits mirrored that of naturally infected macaques and baboons (see below). Neutralizing antibody titers were low at all time points but increased over time. No pathology was noted in any of the FV-infected animals. Another study performed in 1979 investigated whether SFVcpz induces immunosuppression in rabbits 49. Again, no pathology was noted in any animals, but a transient immunosuppressive effect in the absence of interferon production was observed in infected rabbits. Another study of FV-induced immunosuppression in rabbits and mice was published in 1993 96. Similar to the previous study but using in vitro proliferation assays, the authors observed a transient immunosuppressive effect only in FV-infected animals. No animals showed any sign of disease, and FV was only recovered after coculture on permissive cells.
Recently, a long-term study of SFVcpz(hu)-infected rabbits was performed 94. In this 5-year study, all animals developed robust antibody responses against all major FV proteins. FV was recovered from peripheral blood leukocytes (PBL) by coculture after 2 weeks, and all major organs were positive for FV by nested PCR. FV was never detected in freshly isolated PBL by indirect immunofluorescence (IFA) or by electron microscopy, despite the fact that the same cells yielded FV when cocultured on permissive cells. As in other studies, there was no gross pathology or histopathological lesions in any of the infected animals.
There are two published studies using mice as a model system for FV infection. The first employed SFVcpz (type 6) and Swiss-Webster mice 19. As in previous studies with rabbits, all animals were asymptomatic throughout the course of the experiment. Virus was recovered from spleen and kidney by coculture, but histological examination and indirect immunofluorescence of the same tissues were negative. A more recent study utilized a variety of inbred mouse species and the prototype SFVcpz(hu) 98. A differential humoral immune response was observed between CBA/Ca and C57BL/6 mice, wherein both developed antibody against Gag protein but only C57BL/6 mice produced antibody against the Bet protein. The authors note that FV was detected in a wide variety of organs by nested PCR and that the frequency of detection increased over time in CBA/Ca mice but decreased over time in C57BL/6 animals. Unlike the previous study using Swiss-Webster mice, recovery of FV from infected mice by coculture was very inefficient. Like the previous study, no pathology was observed throughout the experiment.
In an effort to understand the pathogenic potential of FV, two transgenic strains of mice were generated 17. The first expressed all SFVcpz(hu) genes from the LTR but was replication defective due to a frameshift in the integrase domain of pol. A second strain contained deletions in gag, pol, and env, expressing only the accessory genes tas and bet from the IP. In both cases, clinical symptoms, which eventually proved fatal, included ataxia, spastic tetraparesis, and blindness. The rate of disease progression, however, was faster in animals expressing all FV genes. In affected animals, transgene expression was limited to the forebrain and cerebellum. Pathology in all affected animals was limited to tissues of the central nervous system (CNS) and striated muscle. At the onset of symptoms there was progressive degeneration of primarily the CA3 region of the hippocampus and the telencephalic cortex. Some animals showed degeneration of striated muscle as well. Interestingly, transgene expression was observed in only a fraction of the cells of a given lineage. This is unusual for transgenes and indicates that cell type-specific factors are likely required for FV expression. The gene products which were responsible for the pathology were not identified, but it was likely to be tas and/or bet because these are the only intact open reading frames (ORFs) in both viral constructs.
A possible explanation for the observation that mice expressing FV structural genes progressed more rapidly than those which only expressed tas and bet was provided by analysis of FV protein expression in affected animals 8. It was noted that in both transgenic strains Tas expression was observed, but only in cells expressing FV structural genes was there syncytium formation in astrocytes. The fusogenic property is likely due to expression of FV Env. Env expression may accelerate disease progression because of its fusogenic property per se or because by fusing Tas-expressing cells with neighboring cells, Tas can induce FV expression in those neighboring cells which otherwise would have not expressed any FV genes.
Further work was performed investigating transgene expression during development 6. These authors determined that during embyrogenesis the transgene was broadly expressed, albeit at low levels, in tissues of ectodermal and mesodermal origin, but this expression waned as animals neared gestation. Perinatal expression of the transgene was restricted to the CNS, and by 3 weeks of age, transgene expression was undetectable. In young adult mice, transgene expression reappeared in the CNS and striated muscle and reached levels much higher than observed during embyrogenesis. Interestingly, the authors noted that transgene expression in adult mice was invariably high, but the cellular distribution was variable even among littermates. These results are most likely explained by the positive-feedback loop created whereby the transactivator, Tas, activates its own promoter as well as the LTR 67, 87. Hence, if a basal level of tas expression is present, the result is an unregulated positive feedback loop.
Transgenic mice containing the tas and bet ORFs under control of the LTR were generated in an attempt to determine which gene(s) was responsible for the neurotoxicity observed in the previous studies 116. None of these animals showed disease progression despite the presence of mRNA corresponding to both tas and bet forms in a wide variety of tissues, including the CNS. Interestingly, only Tas protein was detected in the brains of these animals, while affected animals from the previous studies expressed not only Tas, but also high levels of Bet. These data indicate that Bet expression may be required for disease progression in transgenic mice.
Taken together, the studies in transgenic animals have clearly demonstrated the pathogenic potential of FV gene expression when it is targeted to the brain and striated muscle. They have also shown that expression of structural genes enhances pathogenicity and that Bet may be responsible for the pathology. However, whether FV can productively infect the tissues in which pathology is seen in transgenic animals during a natural infection remains unclear. The results of the majority of these studies are reviewed in detail elsewhere 5, 7.
Since the discovery of FV, and especially since the isolation of SFVcpz(hu), researchers have searched for diseases associated with FV infection in humans. A wide variety of diseases have been tenuously associated with FV infection of humans (Table (Table2).2). Most of these studies have relied upon only one assay to demonstrate FV infection, leading to the possibility of spurious positive results. In most cases, more comprehensive studies using multiple assays to confirm virus presence have failed to verify the initial disease association.
A transmissible agent was associated with thyroditis de Quervain in 1974 by coculturing of patient samples with susceptible cell lines 106, 107. The agent was later studied by electron microscopy and tentatively classified as a paramyxovirus 105. Subsequently, the agent was reanalyzed by immunofluorescence and electron microscopy and reclassified as an FV 119. A more comprehensive study using IFA, radioimmunoprecipitation (RIPA), Western blot (WB), and enzyme-linked immunoabsorbent assay (ELISA) on samples from 19 patients demonstrated no association between FV infection and thyroiditis de Quervain 26. As part of a larger study investigating the prevalence of FV infection in humans, 59 patients with thyroid disorders, including 28 with Quervain's thyroiditis, were analyzed by IFA, RIPA, WB, and PCR 103. Again, there was no prevalence of FV infection in thyroiditis de Quervain patients. The origin of the FV-like agent in the original publications 106, 107 remains unclear, but because the more comprehensive study was unable to detect FV infection in Quervain patients, it is highly unlikely that FV infection is a causative agent of this condition.
A putative association of FV infection with Graves' disease (GD) was first reported by Wick et al. in 1992 122, 123. Using antibodies against FV Gag protein, the authors detected a signal by IFA in tissue sections from GD patients. A second study of French GD patients revealed the presence of FV DNA by PCR in 19 of 29 patients 61. DNA samples from these patients as well as samples from 41 German GD patients were subsequently analyzed by Southern blot and PCR 104. This study confirmed the presence of FV DNA in the French cohort but did not detect FV DNA in any of the 41 German patients. This leaves open the possibility of artifactual contamination of the French samples. Another study analyzed 28 more GD patients, this time using IFA, RIPA, and WB and did not detect FV markers in any samples 43. In a separate study, 5 of 99 GD patient samples were positive for FV sequences when analyzed by PCR 128. However, 6 of 109 control patients also tested positive using this assay, indicating no statistical association between the presence of FV sequences and GD. As part of a study investigating the presence of FV antibody in African patients, serum from 45 patients was analyzed by ELISA and WB 71. None of the GD patients had FV-reactive antibodies using these assays. Recently, nested PCR was used to identify FV gag, env, and LTR sequences in normal and Korean GD patients. FV sequences were detected in 13 of 24 patients, and in 7 patients all three regions were amplified 62. However, 9 of 23 apparently normal controls were also positive for at least one locus under the same conditions. Although intriguing, as no other methods were used to confirm FV infection, these studies provide no evidence for a causative role for FV in GD. However, the data do represent the possibility that FV-like sequences may be present in some geographically distinct populations.
The long search for an etiologic agent in multiple sclerosis (MS) and the findings of neurodegenerative disease in transgenic mice expressing FV genes led to an idea that FV infection may be associated with MS. Westarp et al. used an ELISA-based technique to analyze the sera of more than 85 MS patients for the presence of FV-specific antibodies 120, 121. The authors found that sera from 33 of 85 patients cross-reacted with FV antigen and concluded that there may be a link between FV infection and MS pathology. In four subsequent studies, 153 patient samples were analyzed using a variety of techniques, including coculture, PCR, ELISA, WB, and IFA 70, 71, 90, 111. Only in cases which relied exclusively on the ELISA technique was there any indication of FV infection in MS patients. However, except for the original studies by Westarp et al., these studies all reached the conclusion that there is no association between FV infection and MS.
The cross-reactivity of a patient serum sample with FV antigen by WB led to the speculation that FV infection may be involved in myasthenia gravis 92. A more recent study of four myasthenia gravis patients showed the presence of FV DNA in patient thymuses by PCR and Southern blot as well as low levels of SFVcpz(hu)-specific neutralizing antibody in all patients 64. The authors of this study were unable to recover infectious FV from these patients, and their Southern blot data indicate that the FV sequences detected may be from incomplete proviruses. For these reasons, the authors suggest the need for further study to confirm or refute the role of FV as an etiological agent in myasthenia gravis. No further confirmatory studies have been published to date.
FV has also been investigated as an etiologic agent in a variety of other diseases. Two studies in 1992 searched for FV-specific antibodies in patients with chronic fatigue syndrome 40, 71. In both studies the authors were unable to find FV seroreactivity in 71 samples tested. A number of unconfirmed reports, largely based on single assays, have implicated FV infection in familial Mediterranean fever 113, sensorineural hearing loss 84, and dialysis encephalopathy 22.
As a whole these studies clearly demonstrate that the link between FV infection and human disease is unfounded. The only exception is myasthenia gravis, but the inability to isolate FV from myasthenia gravis patients confounds the issue. The ability to amplify FV-like sequences from human DNA by PCR is used by many as evidence for FV infection, but may be explained by the sequence homology of FV with other retroviruses. Phylogenetic analyses indicate that FV are distantly related to other complex retroviruses, like human T-cell lymphotrophic virus (HTLV) and HIV, but most closely related to human endogenous retrovirus type L, which may be present at up to 200 copies per cell in the human genome 24. Furthermore, the probability of FV as an etiologic agent in human disease is further reduced because controlled studies have indicated there is no evidence of FV infection in the human population (see below).
In conjunction with the search for human diseases associated with FV infection, a number of studies attempted to determine whether any specific human populations are naturally infected with SFVcpz(hu). The study which first hinted that the so-called human FV isolate was not of human origin also demonstrated by serology that FV infection was not prevalent in humans 20. More than 250 patient samples were assayed, including 50 from individuals with NPC and Burkitt's lymphoma. None of these were found to contain neutralizing antibodies specific for SFVcpz(hu). The same group also tested 14 animal caretakers and 24 laboratory personnel, all of whom had regular contact with nonhuman primates, and found that none carried anti-FV antibodies as measured using IFA and neutralizing antibody (NAB) tests. Three seroepidemiological studies by Achong and Epstein 1,2 and by Muller et al. 80 sought to confirm or refute the negative data obtained by Brown et al. 20 and by Nemo et al. 82. In these studies, samples were analyzed from over 800 individuals from various regions of Africa, as well as Singapore and Britain. Using IFA analysis of FV-infected cells, it was found that while healthy East African patients have a low seroprevalence (~4%), East Africans with tumors, especially NPC, were >20% seropositive using this assay. No samples obtained from individuals outside the regions of East Africa studied, including Tunisia, were seropositive. At about the same time, a study of over 1,700 persons from various Pacific island territories was performed using IFA and NAB methods 69. Overall, 6.9% of the population showed FV-specific antibodies by IFA, with up to 15% of some individual populations testing positive. The presence of NAB correlated well with the IFA data. In another study, ELISA as well as WB techniques were used to analyze 3,020 serum samples from African and European patients 71. Overall, 3.2% were seropositive using these techniques. Only 1.6% of European samples were positive, while 6.3% of samples from East Africa were positive. A seroprevalence of 16.6% in patients with NPC was in accordance with previous studies.
Two comprehensive studies were performed in 1995 and 1996 which cast doubt on the conclusion of previous studies which had suggested that FV infections are prevalent in some human populations. The first used WB, RIPA, IFA, and PCR to screen over 2,500 serum samples from Germany and east and west Africa 103. Sera from patients with a variety of diseases and conditions, including various cancers, thyroid disorders, AIDS, malaria, sudden deafness, graft-related immunosuppression, and other autoimmune disorders, were included in the study. This study differed from previous studies in that samples from FV-infected animals as well as accidentally infected and zoonotically infected humans were used to establish the baseline criteria for an authentic FV infection. Patient samples had to test positive in all four assays to be considered FV infected. Control experiments showed that 80% of rhesus monkeys, 93% of African green monkeys, 75% of chimpanzees, and two humans with documented infection by FV were positive for FV in all serological assays. When PCR was performed on 32 human and primate subjects, 14 samples tested positive. In these 14 cases, all serological results were also positive, while the 18 cases that were negative by PCR were also negative by serology. Tests of unknown sera indicated that 99.7% of the samples were negative in at least one of the serological assays. It is important to note that a significant proportion of the test sera were positive by one or two of the serological assays, but only 7 of 2,688 samples were positive for all three, and in all these cases the WB results were questionable. However, the rate of false-positive results seen in this study when using only a single serological assay could readily explain the apparent positive results observed in previous studies which relied upon only one or two serological assays. A second comprehensive study using sample from diverse geographic regions was performed using ELISA as a preliminary screen 9. ELISA-positive samples were then tested by focal immunoassay, WB, NAB, and, when possible, PCR. The initial screen by ELISA produced results similar to previous studies, in that 17% of Pacific Islanders and 34% of samples from Malawi in central Africa were positive. However, when subjected to further analyses, none of these samples were positive using the confirmation techniques. Taken together, these well-controlled studies indicate that FV infections are not prevalent in the human population.
In an effort to understand the risk of zoonotic transmission of FV in a natural setting, Goepfert et al. examined the prevalence of FV infection in West African hunters 36. This cohort is presumably at high risk for zoonotic transmission of FV because of the close contact with body fluid and tissue from a wide variety of nonhuman primates. Using a recombinant RIPA and PCR techniques, none of the 17 hunters examined tested positive for FV infection.
Another group at high risk for zoonotic transmission of SFV is frequent handlers of nonhuman primates, such as veterinarians and zookeepers. A voluntary study sponsored by the Centers for Disease Control tested animal handlers for simian immunodeficiency virus, simian T-cell lymphotrophic virus, and simian type D retrovirus as well as SFV infection. The results of this study show that 3 of the 231 participants tested positive for FV infection 13. A virus characteristic of FV was isolated from one of the three individuals by coculture of peripheral blood mononuclear cells (PBMC) on permissive cells. Virus from this caretaker was subsequently determined to be of SFVagm origin 100. PBL from this individual were sorted by fluorescence-activated cell sorting (FACS) and then tested for FV DNA by nested PCR. This analysis determined that CD8+ T cells were the major cellular reservoir infected with FV 117. A follow-up study identified a fourth individual infected with FV and confirmed the three previous zoonotic infections 44. In this study, sequence analysis of PCR products indicated that one individual was infected with SFVagm, while the remaining three were infected with SFVbab. In all cases, the positive individuals have histories of severe bite wounds or, in one case, two instances of instrument puncture wounds. Archived serum samples from one worker showed the presence of anti-FV antibodies dating to 1978, while those from another dated to 1988. In no cases have associated clinical symptoms been reported. Sexual transmission was not observed in any of the cases, nor was there transmission to immediate family members.
It should be noted that there are no published studies on the presence of other primate FV in the general population, nor are there studies on the transmission of nonprimate viruses such as FFV to humans.
Transmission of SFVbab to humans was studied in two patients who had received baboon liver transplants 11, 12. The two patients had AIDS and end-stage hepatic disease associated with hepatitis B virus infection. Both patients received baboon livers, and both died shortly afterwards from complications unrelated to FV infection 108. Using PCR, SFVbab sequences were detected in various organs from both patients. In addition, baboon endogenous virus was also detected in a wide variety of organs at distal locations from the site of transplant. The presence of baboon mitochondrial DNA in the same locations suggests that the viral sequences were present in donor cells, possibly baboon leukocytes, which had migrated to distant sites in the recipients. However, infection of the human cells was not specifically analyzed. No SFVbab virus could be isolated from either patient.
Persistence in the absence of disease is the defining characteristic of natural, accidental, and experimental FV infection. Despite the presence of a robust antibody response against FV, infectious virus from infected animals can be easily recovered by coculturing throat swabbings, PBMC, or infected tissue with susceptible cell lines 50, 55, 56, 60. The recovery of virus from cross-species and experimental infections is more difficult and can only be achieved by coculture of PBMC or infected tissues, but interestingly not from saliva. However, the recovery of infectious SFVagm from an accidentally infected human after 20 years, despite high antibody titers, underscores the ability of FV to persist for long periods of time even in cross-species transmissions 44, 100. An SFVagm virus isolated from the infected person was found to lack a functional bet gene, which confounds the evidence for the role of the bet gene in in vivo persistence 21.
It is important to note the incredible genetic stability of FV compared to other retroviruses. For example, in one case of SFVcpz infection of a veterinarian who had been severely bitten by a chimpanzee, the viral sequences in the human and chimp could be compared 15 years after the zoonosis. Very few nucleotide differences could be detected in the integrase domain of pol by PCR and sequencing (Heneine et al., submitted). A study of the genetic stability of African green monkeys infected with FV also found remarkably high homology rates 102. In one animal, sequence analysis from samples taken over a 13-year period indicated homology rates of 99.5 to 100%. These data clearly reflect the very low replication rate of FV in vivo, similar to the case of HTLV.
The tropism of FV in vivo appears to be very broad. Using coculture techniques as well as PCR, FV can be detected in most tissues in infected animals 30, 50. In contrast, however, the direct culture of FV from sera and the detection of replicating FV in infected tissues using IFA and electron microscopy have been consistently negative 19, 49, 50, 112. In a recent study of naturally infected African green monkeys, nested PCR identified proviral FV DNA in all organs tested 30. To detect sites of active viral replication, reverse transcriptase-PCR analysis was performed on RNA from the same organs. Only in the oral mucosa of a single animal was a positive reverse transcriptase-PCR result observed. This result was confirmed by in situ hybridization using an env-specific probe. This location is consistent with our limited knowledge of FV transmission, which appears to be through saliva. The ability of the authors to find integrated FV DNA by PCR in every organ tested highlights the central paradox of FV biology: the processes of FV infection, dissemination, both intra- and interhost, and persistence are very efficient despite the apparent absence of viral replication in infected tissues. It is important to note that detection of FV DNA in a particular tissue does not indicate that the tissue is infected per se. Infection of a circulating population of cells could account for such PCR results. Nevertheless, the mechanism(s) which underlies the process of virus dissemination in vivo is poorly understood, and this underscores the importance of determining the actual cell type(s) in which FV can persist and replicate. Investigations of both the host immunological and viral factors involved in FV persistence also need to be undertaken.
A number of studies have sought to identify which subset(s) of peripheral blood cells is persistently infected in FV-infected hosts. PBL from naturally infected African green monkeys, chimpanzees, and two accidentally infected humans were analyzed for the presence of FV DNA. One human was accidentally infected with the prototype lab strain, SFVcpz(hu), while the second was infected zoonotically with SFVagm. PBL from these individuals were separated into PBMC and polymorphonuclear leukocytes (PMNL), then positively sorted by FACS into CD8+, CD4+, and CD14+ (monocyte) populations. Human samples were also sorted for CD19+ (B cells) cells, while for African green monkeys and chimpanzee, CD4− CD8− fractions, mostly B cells, were collected. CD8+ cells from all subjects were positive for FV DNA. All other human cell types were negative for FV DNA. However in monkey samples, 4 of 7 CD4+ samples, 10 of 13 CD4− CD8−samples, 3 of 11 CD14+ samples, and 4 of 11 PMNL samples tested positive.
A second study attempting to identify the cellular tropism of FV was performed on PBMC from a human accidentally infected with SFVagm 21. Virus isolated from this individual was determined to have a defective bel-2 ORF and thus a presumably defective bet gene. Antibody-conjugated magnetic beads were used to positively sort PBMC into CD4, CD8, CD14, and CD19 populations. Nested PCR on DNA from these populations indicated that CD14+ and CDI9+ populations were infected, while CD4+ and CD8+ cells were not infected. These results directly contradict the results obtained by Von Laer et al. 117. In both studies, humans infected with SFVagm were analyzed, and thus one possible explanation for these contradictory results is that an intact bel-2 ORF somehow alters tropism in vivo. Clearly, further investigation of the in vivo cellular targets of FV is needed.
The characteristics of FV infection, such as lack of pathogenicity, broad host range, and wide tissue tropism, persistence in the presence of neutralizing antibody, and the large genome size, make FV attractive vectors for gene therapy. There is evidence that FV vectors can infect CD34+ hematopoietic cells more efficiently than murine leukemia virus based vectors 48. The apparent lack of pathogenicity provides the option of using replication-competent vectors in which the bet gene can be replaced with a promoter and gene of interest 99. Viruses which lack the transactivator tas can be constructed by replacing the U3 region of the LTR with constitutive promoters such as that from cytomegalovirus 97, 115. However, several problems need to be overcome. Currently, there are no highly efficient packaging cell lines for FV, and the apparent inability to pseudotype FV particle with other envelopes is problematic 83, 126. In addition, the vector systems which have been reported have relatively low titers 115, 126. Furthermore, several reports have described multiple cis-acting sequences in the FV genome that are required for vector transfer 29, 42. The first packaging signal is located at the 5′ end of the RNA, and a second is located near the 3′ end of the pol gene. A better understanding of these cis-acting elements is required for optimal vector and packaging cell line construction. A further unknown is whether viral vectors will continue to be expressed at long times after administration.
To date, except for the FV, all retroviruses have been shown to cause disease in some host, either naturally or accidentally infected. Simple retroviruses such as murine leukemia virus and avian leukosis virus induce tumors and/or immunodeficiencies with long latency periods. Complex retroviruses like bovine leukemia virus, HTLV, and HIV can also be highly pathogenic. Most simian immunodeficiency viruses are nonpathogenic in their natural hosts but can cause rapid, fatal disease when cross-species infection occurs. There is mounting evidence that both HIV-1 and HIV-2 have been introduced into the human population via zoonotic transmission on multiple occasions 33, 34, 81. Human FV has been described as “a virus in search of a disease” 118. However, as mentioned, many viruses persist in their natural hosts in the absence of disease only to reveal their true pathogenic potential when they cross species barriers. Despite evidence to the contrary regarding FV zoonosis, great caution should taken when purposefully introducing viruses or vectors into new host species. In our opinion, even greater caution should be taken in performing xenotransplantations using organs from foreign species which are known to harbor endogenous or naturally occurring viruses. Use of primate material for production of biologicals, such as vaccines, could also inadvertently introduce FV into the human population. Our knowledge to date indicates that FV are truly nonpathogenic, even in other host species; however, very few transmission to humans have been documented. Given enough opportunities to cross species barriers, a hitherto unforeseen FV pathogenicity may reveal itself.