FV isolation is relatively easy on primary fibroblasts from throat swabs, although any other tissue as source of virus can be used.⁹⁴,¹⁵⁴,¹⁶⁶,¹⁸⁶ Virus isolation from peripheral blood lymphocytes is greatly enhanced by the addition of anti–γ-interferon (γ-IFN) antibodies.⁵⁶ A virus isolate displaying the typical giant cells’ CPE (Fig. 52.1A) can be confirmed by the detection of reverse transcriptase activity, by immunofluorescence assay (IFA) demonstrating a predominantly nuclear antigen (Fig. 52.1B), or by nucleic acid detection methods.⁹⁴,¹⁵⁴,¹⁶⁶,¹⁸⁶,²²²

The diagnosis of an FV infection can be made by demonstrating antibodies against the main structural proteins (typically the Gag doublet, see later discussion) in serologic assays,⁷¹ such as immunoblots or radioimmunoprecipitation assays (Fig. 52.2). The choice of the right antigen in these assays is important, because the reaction is, at least in part, virus type specific.⁹⁶,¹⁰²,¹¹² To verify an FV infection, the serologic analysis should be combined with nucleic acid detection methods (typically polymerase chain reaction [PCR]). For this, the amplification of a conserved (∼420 bp) fragment from the integrase (IN) domain of the pol gene has proven to be extremely useful.²²⁰

Various vertebrate species are naturally infected with FVs. Table 52.1 and Figure 52.3A give an overview of some isolates.
In addition to FVs of Great Apes,⁹⁴,⁹⁵,¹⁵⁴,¹⁸⁶ Old and New World simians,⁹⁴,¹⁵⁴,¹⁸⁶ and prosimians,⁹⁹,¹⁵⁴,¹⁸⁶ FV infections appear to occur worldwide in bovines and other Artiodactyla,⁵,⁵⁹,¹⁴⁸,^149a,¹⁵⁴,¹⁸⁶ equines, ¹¹⁹,¹⁵⁴,¹⁸⁶,²⁴² and felidae.¹⁵⁴,¹⁸⁶,¹⁹⁶ Whether sea mammals are natural hosts has not been intensively investigated.¹¹⁰,¹⁵⁴,¹⁸⁶ Probably all monkey species harbor an FV.⁹⁴,¹⁰⁰,¹³⁵,¹⁵⁴,¹⁸⁶ Prevalence in the natural host in the wild may be as high as 100% and is usually more than 30%.⁹⁴,¹³⁵,¹⁵⁴,¹⁸⁶ It is therefore not unlikely that FVs are, in terms of prevalence, the most successful of all retroviruses. Although host restriction factors show some species restriction (see later discussion), FVs have been reported to cross the species barrier between monkeys and apes in captivity or in the wild.⁹⁴,¹²⁴,¹³⁵ In their hosts, FVs cause lifelong persistent infections of a benign nature, often in the presence of neutralizing antibodies.⁹⁴,¹⁵⁴,¹⁸⁶ Laboratory animals, such as mice and rabbits, have also been infected in the absence of overt disease.²⁶,⁹³,⁹⁴,²⁰⁸,²¹⁴

Humans are not a natural host of FVs.¹⁸⁶ Indeed, the best-studied FV was once believed to be of human origin.¹,¹⁴³ It has now been designated as the prototype foamy virus (PFV).¹⁸⁷ Initial reports on naturally occurring human infections¹⁴⁴,¹⁴⁷ were not confirmed by the large-scale screening of more than 5,000 samples using sophisticated methods such as a combination of antibody detection and PCR.²,²²² Furthermore, postulated associations of FVs with human disease states, mainly autoimmune disorders and neurologic diseases of unknown origin could not be validated.⁴⁰,⁸⁶,²⁰¹,²²³ Because of the close nucleotide sequence homology to FVs from the Pan troglodytes schweinfurthii chimpanzee sub-species,⁸⁸,¹³⁵ the single human isolate from a Kenyan patient is now believed to have resulted from a trans-species transmission of a virus from these chimpanzees that were once more prevalent in East Africa.⁵¹ Primate FVs are currently not circulating in the human population. However, humans are susceptible to zoonotic transmissions of nonhuman primate (NHP) FVs.²¹⁸,²³² Altogether, around 100 human infections with NHP FVs have been confirmed worldwide. In a survey, the Centers for Disease Control and Prevention identified approximately 2% seropositives and virus DNA positives among several hundred samples from occupationally to NHP-exposed persons and could isolate virus in several instances.²²,⁸⁷ Some of these

infections date back decades and were the result of severe monkey bites.²²,⁸⁷ FV infection has also been identified in African bushmeat hunters.³⁰ Other persons at risk are those living in close proximity to quasi-wild NHPs (e.g., at Asian temple sites) or individuals possessing an NHP pet.¹⁰¹ A very special case may be human recipients of NHP xenotransplants.⁴ Human infections are lifelong; however, none induced any disease and remained unrecognized prior to the investigation. The viral load in buffy-coat cells of infected humans revealed a very low number of FV DNA copies.³⁰ Moreover, in contrast to lentiviruses, in vivo adaptation to what can be called a human FV has not occurred. Even after decades of infection, FVs in humans have remained relatively unchanged at the nucleotide level, and the transmitting donor species can be readily identified²²,⁸⁷ (i.e., a baboon FV will remain a baboon FV even after decades in a human host) (Fig. 52.3B). Although the sample size is relatively small, there is no indication of human-to-human transmission even between close contacts.²²,^64a Because FV can be transmitted by transfusion,²⁵,¹¹¹ infected persons have been advised not to donate blood to avoid new human retrovirus infections.⁸⁵

For reasons that are not fully understood, men appear to be a “dead-end” host for primate FVs. With respect to human infections by non-NHP FVs, it is worth mentioning that the antibody screening of more than 200 veterinarians at risk of acquiring a feline foamy virus (FFV) infection did not reveal a single positive case.²⁸ Based on this result, it appears unlikely that bovine or equine foamy virus (BFV and EFV, respectively) can infect humans. However, this has not been thoroughly investigated.

Exogenous FVs are very ancient retroviruses. Exogenous FVs are very ancient retroviruses. There exist three examples of endogenous FVs, which suggest that exogenous FVs were around before that times: 1) Katzourakis et al.¹⁰⁸ reported the detection of endogenous FVs (SloEFV) in the South American sloths genome (Choloepus hoffmanni), 2) Han and Worobey^71a described endogenous FVs (PSFVaye) in the Madagascan aye-aye, a primitive lemur species (Daubentonia madagascarensis), and more surprinsingly, 3) in the latimera (Coelacanth) genome^71b (CoeEFV), suggesting that exogenous FVs existed more than 400 million years ago in species outside the mammals. Moreover, once an exogenous FV has adapted to its host, it mutates only slightly faster than the host mitochondrial DNA.²³⁴ The substitution rate of simian foamy viruses (SFVs) has been estimated to be around 1.7 × 10⁻⁸ per site and year.²³⁴ This makes FVs the most genetically stable of viruses, with an RNA phase in replication. For instance, the FV mutation rate is approximately ten times lower than primate T-cell lymphotropic virus type 1 (PTLV-1), a virus that replicates primarily by a proviral expansion mechanism (i.e., through DNA) (see Chapter 48).

FVs combine two favorable features: They are of extraordinary genetic stability and, if not acquired by trans-species transmissions, always point to the host species from which they were derived, and they are shed in feces. Owing to these features, NHP phylogeographic and conservational issues can be addressed easily without animal disturbance.¹³⁵,²²⁰

Curiously, the fidelity of the reverse transcriptase enzyme (RT) does not reflect this enormous genetic stability. If analyzed in vitro, the PFV RT was found to be of exceptional high processivity but of low fidelity.²³,²⁴,¹⁹⁷ The mutation rate of PFV RT (approximately 1.7 × 10⁻⁴ per site and replication round) is similar to that of human immunodeficiency virus (HIV).²⁴ Most mutations found were small deletions and insertions.²⁴ If analyzed in cell culture, such deletions and insertions were not found. However, a point mutation error rate of 1.1 × 10⁻⁵ per site and round of replication remained.⁶²

Thus, the genetic stability of FVs at the molecular level is not currently understood, and the involvement of as yet unidentified specific cellular factor(s) may play a role in this process.

The host cell range for FVs is quite broad and includes species-independent primary cells or cell lines of fibroblastoid, epithelial, and lymphoblastoid origin, such as various B and T lymphocytes, and cells of erythroid and of myeloid lineages.⁹⁴,¹⁵⁴,¹⁵⁹,¹⁸⁶,²⁶⁸ Upon replication in adherent cell cultures, FVs induce massive multinucleated giant cell CPE (Fig. 52.1A), and apoptosis is thought to be the ultimate cause of cell death.¹⁵⁸ Vacuolization of cells is often only observed using primary isolates.

The paucity of cell lines resistant to FV infections has hindered the identification of the cellular receptor(s) required for entry by classical approaches. It is now appreciated that all FVs use the same cellular receptor, including those present on bird, reptile, and fish cells.¹²,⁸⁹,²³⁰ Only two cell lines—Pac-2 zebrafish embryonic fibroblasts and the G1E-ER4 human erythroid precursors—have been reported to be refractory to infection.²³⁰ Thus, the means to screen complementary DNA (cDNA) libraries for FV receptor–related genes has now been established.

Whereas the characteristic CPE develops in adherent cells, this hallmark of FVs is often absent in cells of lymphoblastoid origin, in which FVs appear to become latent and intermittently reactivatable. Latently infected cells do not undergo syncytium formation and death but proliferate with normal kinetics and produce low amounts of virus.²⁶⁸ As judged from Southern blots, the viral DNA copy number oscillated in infected lymphocytes (unpublished observation). Interestingly, chemical treatment of lymphocytes (e.g., with phorbol esters) may induce the latent virus and cause cell death owing to activated viral replication.¹⁵⁷,²⁶⁸ Although this is reminiscent of the lymphotropic herpesviruses, in FVs the molecular basis for virus reactivation has not been investigated nor have sites within the cellular or FV DNA genome responsive to the drug-mediated reactivation been mapped. Whether the methylation of FV DNA that has been observed in a cell culture model²¹⁹ contributes to in vivo latency remains unresolved, because there is no evidence of transcriptional down-regulation of FV vectors by methylation following their introduction in vivo.⁹⁰,¹⁶⁹,²²⁵

Recently, Liu et al¹³⁵ used methods similar to those employed to demonstrate that human immunodeficiency virus type 1 (HIV-1) was derived from simian immunodeficiency virus from chimpanzees (SIVcpz; i.e., they collected and analyzed fecal samples from wild chimpanzees; see Chapters 49 and 50). They found that simian foamy virus from chimpanzees (SFVcpz) is widely distributed among wild chimpanzees with a
phylogeographic distribution and is transmitted horizontally, because babies younger than 2 years were free from FVs and infection rates increase with age. Moreover, they determined that superinfection by SFVs from lower primates and frequent recombination events occur. The most interesting finding by Liu et al has been the detection of viral RNA but not viral DNA in the fecal samples.¹³⁵ This finding directly relates to the FV replication pathway (see later discussion). However, Liu et al have not investigated whether the RNA-containing virus transmits the infection, and it is not known in which cell type these viruses were produced. It is possible that the DNA content in the fecal samples may have been too low to be detected even with very sensitive methods.

The tissue distribution of SFV and sites of in vivo replication have been investigated using another approach. As expected from the broad host cell range of FVs seen in vitro, viral DNA was detected at a frequency of one genome copy per 10² to 10³ cells in every organ examined.⁵⁴ Because animals were perfused prior to the analysis of nucleic acids, infected lymphocytes were not detected, although these cells were the probable vehicles of virus dissemination in vivo.⁵⁴,⁵⁵ As judged from in situ hybridization experiments, viral RNA, indicative of active virus replication, was confined to superficial cells of the oral mucosa.⁵⁴,¹⁶⁵ Thus, it appears that only cells, which are destined to be shed, are productively infected and undergo lytic replication in vivo. Differential expression of yet undisclosed host factors restricting viral replication in other tissues is a likely explanation for this observation.

It is generally assumed that FVs are transmitted among NHPs through saliva via social contacts, including aggressive activities such as biting among young animals.³¹,¹³⁵ These contacts and lactation are also suspected to be the main transmission route of the FFV,²⁵⁹ whereas BFV probably is mainly transmitted via milk from infected cows to offspring.¹⁹⁸

Early studies suggested that drug-induced immunosuppression of infected African green monkeys did not result in SFV-related symptoms yet enhanced the frequency of virus isolation (D. Neumann-Haefelin, Freiburg, Germany, personal communication). It was subsequently found that in dually SFV- and simian immunodeficiency virus (SIV)-infected and severely immunosuppressed macaques, the predominant site of FV replication changed from the oral mucosa to the small intestine.¹⁶⁴ However, SFV-related diseases did not occur. This was also observed in cats dually infected with feline immunodeficiency virus (FIV) and FFV.⁸,²⁷² A case of human co-infections by HIV-1 and SFV from mandrills has also been reported, without clinical consequences that could be attributed to the SFV infection.²³³

By ultrastructural analysis, FVs appear as immature-looking core particles surrounded by a lipid bilayer with embedded prominent Env proteins⁹⁴,¹⁵⁴,¹⁸⁶,²⁵⁶ (Fig. 52.4). In negative-staining electron microscopy (EM), the virion has a diameter
of approximately 110 nm and a core of approximately 60 nm.⁹⁴,¹⁵⁴,¹⁸⁶,²⁵⁶ The core has an immature morphology owing to the very limited cleavage of the Gag precursor protein by the viral protease (PR) (see later discussion). The cores of infectious PFV virions are made up of the Gag precursor pr71^Gag and its larger processing product p68^Gag at a ratio of 1:1 up to 1:4.³³ This Gag doublet is seen with all FVs,⁷¹ although the nonprimate FV capsid proteins are considerably smaller than their primate relatives.⁹²,¹¹⁹,¹⁹⁹,²⁴² The core is probably oriented radially with the Gag N-terminus pointing outward in the direction of the Env proteins²⁵⁶ (Fig. 52.4C). The prominent surface spikes average 15 nm in length and are organized as trimers in ring-like structures.²⁵⁶ A feature that PFV shares with HBV is the formation of subviral particles (SVPs) consisting only of membranous Env-containing vesicles and devoid of cores.²²⁴,²²⁸

It appears that FV particles contain fewer Pol molecules than orthoretroviruses, which is consistent with the high processivity activity of its RT.²³,¹⁹⁷ The interpretation of older studies demonstrating equal amounts of Pol in foamy and orthoretroviruses³³ have been complicated by the detection of significant amounts of extraparticular Pol protein present in FV particle preparations of different origin as reported by Swiersy et al.²³¹ This study also revealed that in sharp contrast to orthoretroviruses, FV replication tolerates a great imbalance in the relative ratios of Gag and Pol molecules in virus producing cells.²³¹ This is owing to the unusual Pol encapsidation strategy of FVs (see later discussion), in which viral RNA is the limiting factor at conditions of high cellular levels of Pol.

The physical stability of FVs has not been directly compared with that of orthoretroviruses. However, owing to their immature core and a particular Env topology (see later discussion), FVs are probably quite stable. This is illustrated by the fact that vector particles can be concentrated more than 100-fold (i.e., by ultracentrifugation) without great loss of infectivity.⁶⁵,⁸⁹

Viral nucleic acid is present within the core of extracellular FV particles, and it appears to be a mixture of RNA and DNA. RNA dominates at a ratio of approximately 1:1 to 7:1, depending on the genomic region analyzed (long terminal repeat [LTR] vs. gag).²⁰³,²⁶⁹,²⁷⁰ Consistent with this, both forms of nucleic acid were detected in plasma and saliva of an experimentally infected cynomolgus macaque by PCR and RT-PCR.²⁵ Full-length DNA was shown to be present in roughly 5% to 20% of FV virions, and functional studies using the RT-inhibiting drug AZT indicated DNA to be the relevant genome for infection, at least at the high multiplicity of infection (MOI) studied.¹⁵⁰,¹⁶⁰ Because DNA was found in PFV and FFV virions, the idea that FVs are facultative DNA viruses may be generalized.²⁰³

However, it has not been investigated in much detail to what extent reverse transcription is taking place late in the replication cycle. For instance, it is likely that there are genomic regions, such as the gap in the plus strand cDNA (see later discussion), that still remain single stranded. The finding of more LTR than gag region reverse transcripts in virions is a clear indication of this.²⁰³,²⁶⁹ In addition, because more RNA than DNA is found in virions, it would be interesting to know whether purely RNA-containing viruses exist or whether most viruses contain both forms of nucleic acids.

It is with respect to reverse transcription taking place at a late step in viral replication that FVs diverge at most from orthoretroviruses, which display a RNA to DNA ratio of approximately 10⁵:1 in extracellular virions.²⁰³ The most compelling evidence for this are experiments in which the infectivity of virion-extracted FV DNA has been demonstrated.¹⁵⁰,²⁰³,²⁶⁹ Thus, FVs functionally bridge the replication pathways of orthoretroviruses and hepadnaviruses (Fig. 52.5). It is because of this analogy to the hepadnaviral replication cycle (see Chapter 68) that the full-length FV RNA has been termed (pre-) genomic.

In addition to reverse transcription in late phases of FV particle morphogenesis, it has been shown that, similar to orthoretroviruses, reverse transcription also occurs during the early phases of FV replication upon target cell entry.⁴²,²⁷⁰ Given
the aforementioned ratios of RNA to DNA in extracellular viruses, reverse transcription during the early phase of FV replication may only become relevant at a very low MOI, when the amount of virion DNA may be too low to sustain a productive infection.²⁷⁰ Furthermore, the discrepant results reported about the importance of the differentially timed reverse transcription events for FV infectivity may reflect the inherent differences in the cell types used for virus production.⁴²,¹⁵⁰,¹⁶⁰,²⁰³,²⁶⁹,²⁷⁰ In essence, the generally accepted view that virions contain either an RNA or DNA genome, but not both, may not apply to FVs.

The schematic representation of the genome of PFV is shown in Figure 52.6. All FV genomes share common genome structures.¹⁹¹ Between the LTRs, the canonical gag, pol, and env genes are found; downstream of env, the accessory open reading frames (ORFs) are found. Proviruses are 12 to 13 kb—long in comparison to those of other retroviruses. The large size of FVs genomes is partially attributed to the extraordinary long U3 regions of the LTR, which can be explained in part by the overlapping ORF-2 (see Fig. 52.6). However, the accessory ORF-2 reaches only to some extent (approximately 300 bp, in the case of PFV) into the more than 1.4 kb long U3 region, leaving several hundred bases without known function. In the latter, very few enhancer elements, such as those for AP-1 and Ets-1,¹⁵²,²¹³ and the short sequence motifs responsive to the viral transactivator (see later discussion), are present.