STLV-1 co-infection is correlated with an increased SFV proviral load in the peripheral blood of SFV/STLV-1 naturally infected non-human primates

Simian T-Leukemia Virus type 1 and Simian Foamy Virus infect non-human primates. While STLV-1, as HTLV-1, causes Adult T-cell Leukemia/lymphoma, SFV infection is asymptomatic. Both retroviruses can be transmitted from NHPs to humans through bites that allow contact between infected saliva and recipient blood. Because both viruses infect CD4+ T-cells, they might interfere with each other replication, and this might impact viral transmission. Impact of STLV-1 co-infection on SFV replication was analyzed in 18 SFV-positive/STLV-1-negative and 18 naturally SFV/STLV-1 co-infected Papio anubis. Even if 9 animals were found STLV-1-positive in saliva, STLV-1 PVL was much higher in the blood. SFV proviruses were detected in the saliva of all animals. Interestingly, SFV proviral load was much higher in the blood of STLV-1/SFV co-infected animals, compared to STLV-1-negative animals. Given that soluble Tax protein can enter uninfected cells, we tested its effect on foamy virus promoter and we show that Tax protein can transactivate the foamy LTR. This demonstrates that true STLV-1 co-infection or Tax only has an impact on SFV replication and may influence the ability of the virus to be zoonotically transmitted as well as its ability to promote hematological abnormalities.

Foamy viruses are also members of the Retroviridae family. Even if more than 100 cases of Human Foamy Virus (HFV) infection of zoonotic origin have been reported to date, the number of HFV-infected carriers is unknown (for reviews see [17,18]). HFV inter-human transmission does not seem to occur and the consequences of this infection in vivo remain unknown [17,19]. Thus, this virus is considered as non-pathogenic in humans, although a recent report demonstrated hematological abnormalities [20]. HFV simian counterpart, SFV, is highly prevalent in NHPs [21]. It has co-evolved with Old World primates for 30 to 100 million years [22], is well adapted to these hosts and is considered non-pathogenic [23]. SFV is readily transmitted among individuals via the saliva, and some reports suggest that in NHPs, SFV active replication is mainly restricted to epithelial cells of the oral mucosa [24]. However, viral DNA is also detected in CD4+ and CD8+ T-cells and in B cells in vivo [25], and it was previously suggested that PBMCs represent a site of foamy virus viral latency [26].
Although HIV-1/HFV co-infection has been reported to occur in vivo [32], pathogenesisrelated data have only been obtained in NHPs. Indeed, co-infected animals have a decreased CD4+ T-cell population and an accelerated death rate [38]. Consistently, persistent infection with foamy virus increases HIV-1 cell-to-cell transmission in vitro [39].
These results imply that retroviral co-infection can have a strong effect in vivo, even if a given cell might not simultaneously be infected with both retroviruses.
We have access to a colony of baboons (Papio anubis) through the CNRS primate center of Rousset-sur-Arc, France. A series of individuals is naturally co-infected with STLV-1 and SFV, while others are only infected with SFV. These animals are followed up on a yearly basis and blood samples are taken routinely. We previously showed that the ratio between CD4+ and CD8+ T-cells remains constant between STLV-1-positive and STLV-1-negative animals and that STLV-1 infection increases the proportion of effector CD8+ T-cells [40]. Given that STLV-1 and SFV may compete for CD4+ T-cells infection in vivo, we investigated here, in a series of 36 animals, whether STLV-1 co-infection impacted or not SFV proviral load in the blood and in the saliva. This could then result in an increased risk of zoonotic transmission in case of severe bites [41]. As all animals in the cohort were infected with SFV, we could not decipher whether SFV infection impacts STLV-1 PVL. Of note, STLV-1 provirus has never been searched in saliva, while studies on the presence of HTLV-1 provirus in this fluid, which contains lymphocytes, are rare [42,43], and there is no evidence that HTLV-1 could be transmitted through this fluid. We first demonstrate that STLV-1 provirus can be found in the saliva in half of the infected animals that were tested, which is not negligible. In addition, STLV-1/SFV co-infection does not affect SFV PVL in the saliva but leads to a very significant increase in SFV proviral load in the blood. In addition, we show that PTLV-1 Tax protein can activate foamy LTR. These results demonstrate that STLV-1 co-infection has some beneficial effect on SFV replication in vivo. This may influence the ability of the virus to be zoonotically transmitted.

Ethics statement
The use of animals was approved by ethic committee (APAFIS#4227-201604130940121) from French Minister of education and research. Animals are housed at the primate center of the CNRS (UPS 846) in Rousset-sur-Arc and cared for in compliance with French regulations. All individuals live in social groups and have free access to outdoor areas connected to indoor areas. Enclosures are enriched by wooden platforms and vertical structures. Baboons are fed four times a day with monkey pellets, seed mixture, fresh vegetables and fruits. Water is available ad libitum. The experimental procedure complied with the current French laws and the European directive 86/609/CEE. Blood and saliva samples were obtained after anesthesia. Animals were cared for in compliance with French regulations. Animals were anesthetized by ketamine and Medetomidin to allow blood (at the saphenous vein) and saliva collection.

Animals
Thirty-six SFV naturally infected baboons (Papio anubis) were included in this study. Eighteen animals were STLV-1 uninfected and eighteen were STLV-1 naturally infected as determined by immunofluorescence with HTLV-1 infected MT-2 cells confirmed by western-blot (MP diagnostics HTLV-1/II 2.4 for some of them). All animals but one were females.

Blood samples
Whole blood was obtained by venipuncture from the internal saphenous vein and transferred into Vacutainer tubes containing EDTA. Peripheral blood mononuclear cells (PBMCs) were separated on Ficoll gradient. Briefly, blood was diluted 2-fold in PBS-2mM EDTA before loading on Ficoll-Paque PLUS (GE Healthcare). After centrifugation (30 min 400g at room temperature), PBMCs were collected, washed in PBS-2mM EDTA and frozen in 90% SVF, 10% DMSO. High-molecular-weight DNA was then extracted from frozen PBMC samples using the NucleoSpin Tissue kit (Macherey Nagel) with slight modifications, as follows. Cells were defrosted in complete medium, counted and aliquoted (5 million by tube). Lysis was performed for 1 hour at 56˚C and elution was performed with 90 μl of elution buffer supplied in the kit. DNA concentration was then measured (NanoDrop ND-1000 Spectrophomometer, Thermo Scientific).

Saliva samples
Saliva and buccal cells were obtained by harvesting 1-2 ml of PBS deposited in the mouth of the animals at the jwol level without scrapping cheeks. After centrifugation for 10 min at 400g, high-molecular-weight DNA was extracted using the NucleoSpin Tissue kit (Macherey Nagel). DNA concentration was measured (NanoDrop ND-1000 Spectrophomometer, Thermo Scientific).

SFV real-time PCR
Real-time PCR (qPCR) was performed with 20-40 ng of genomic DNA for β-actin amplification and with 100-120 ng of genomic DNA for LTR amplification, using the FastStart Universal SYBR Green Master (Roche), 0,2μM of each primer (PBF3 and 3R see below) using the SYBR Green Master mix (Roche) on a StepOnePlus system (Applied Biosystem) in a total volume of 50μl. After denaturation at 95˚C for 10 min, the reaction mixtures containing DNA were cycled 50 times at 95˚C for 15 sec, 60˚C for 30 sec. The melt curve was then performed as follow: 15 sec at 95˚C, 1 min at 60˚C followed by a gradual increase until 95˚C with acquisition signal every 0.7˚C. As a positive control that was used for all experiments, a fragment of P. anubis SFV 5'LTR was amplified using the same primers i.e. PBF3 (5' AAAGATATTGTATATTA GCCTTGCT) [44] and 3R (5' CACGTTGGGCGCCAATTG) [45]. Amplification of human βactin (Act-F 5 0 TGAGCTGCGTGTGGCTCC and Act-R 5 0 GGCATGGGGGAGGGCATACC) was done on each experimental sample. A series of diluted amplified products was obtained and used for normalization (i.e. SFV copy numbers for 100.000 cells). Sensitivity of the protocol was determined to be 10 copies/for 100ng of genomic DNA.

STLV-1 real-time PCR
Real time PCR was performed with 20-40 ng of genomic DNA for β-actin amplification and with 100-120 ng of genomic DNA for tax amplification, using the FastStart Universal SYBR Green Master (Roche) containing ROX (Carboxy-X-Rhodamine) as an internal reference, in 50 μl of final volume. Samples were incubated for 10 min at 95˚C; then 50 cycles were performed using a StepOnePlus system (Applied Biosystem) (10 sec at 95˚C and 30 sec at 60˚C), and then the melt curve was performed between 60˚C and 95˚C. Tax and β-actin primer efficacy was determined on a series of specific amplicon dilutions. Specific Tax primers Tax-F 5 0 -GTTGTATGAGTGATTGGCGGGGTAA and Tax-R 5 0 -TGTTTGGAGACTGTGTACAAG GCG were used. Amplification of human β-actin was done as described above and allowed to determine STLV-1 copy numbers for 100.000 cells. Sensitivity of the protocol was determined to be 10 copies/sample. Transfection and luciferase assay 10 6 Jurkat cells were transfected with 100 ng of foamy LTR-luc construct or of HTLV-1 LTRluc construct in the presence of increasing amount of Tas [46,47] or of Tax [48] constructs (0-800ng) using the Superfect reagent (Qiagen). Transfections were carried out in the presence of a renilla luciferase vector (phRG-TK) in order to normalize for transfection efficiency. Thirty-six hours after transfection, cells were lysed and luciferase activity was measured using the Dual-luciferase assay system (Promega).

Statistical analyses
Graphs and statistical analyses were performed with the GraphPad Prism software. Comparison of PVL and of age between the different groups was performed using non-parametric unpaired T test (Mann-Whitney test). Comparison of the PVL in the saliva and in PBMCs was performed using non-parametric paired T test (Wilcoxon test, Ã Ã Ã = p-value < 0.001; Ã Ã Ã Ã = pvalue < 0.0001 and ns = not significant). Comparison of frequencies was performed by a χ 2 test.

Results
SFV and HTLV-1 proviruses, i.e. viral DNA integrated into the host genome, have been found in T-cells and saliva in vivo [25,43]. Free SFV viral particles, i.e. viral RNA in cell-free samples, can also be found in saliva, while HTLV-1 particles are extremely rare and poorly infectious, as a result of which cell-cell contact is required for an efficient viral transmission [49]. We wanted to assess the effect of STLV-1 co-infection on SFV proviral load (PVL) in saliva and blood, and therefore the putative inter-individual viral transmission through infected cells. We therefore took advantage of a cohort of Papio anubis, and analyzed eighteen animals naturally coinfected with SFV and STLV-1, and the same number of animals that were only SFV-positive. The animals from both groups had a similar age (S1 Fig).
Because STLV-1 provirus has never been searched in saliva, we first assessed STLV-1 PVL in both the blood and saliva for each STLV-1-infected animal, by qPCR (Fig 1). Samples were classified as negative, positive but non-quantifiable (i.e. a quantifiable signal at least twice among 2 to 3 qPCR experiments, each performed in duplicate) but STLV-1 tax or SFV LTR copy number beyond the amplicon standard range in addition to a positive signal visible on a gel loaded with the PCR products, S2 Fig) or positive and quantified (positive qPCR and tax copy number within the amplicon standard range). Interestingly, saliva samples analyzed by qPCR were positive for STLV-1 proviral DNA in 9/18 animals, including 7 positive but nonquantifiable and 2 positive and quantified samples ( Fig 1A). As a control, qPCR analyses performed on the blood showed positive and quantified results in all animals. The frequencies of negative, positive but non-quantifiable and positive and quantified samples among blood or saliva samples were significantly different (Fig 1A, p<0.0001). We then compared quantified STLV-1 PVL in the blood and in the saliva ( Fig 1B). As expected, quantified STLV-1 PVL were significantly higher in the blood compared to the saliva (Fig 1B, median value for PBMCs = 363.7 and median value for saliva = 266.7, p<0.001). There was no significant correlation between the PVL in the blood and in the saliva since a high PVL in the blood was not associated with a high PVL in the saliva. This result suggests that HTLV-1 PVL in the blood cannot be used as a prediction factor for the presence of the virus in cells present in the saliva. Taken together, these results indicate that the STLV-1 provirus can be found in the saliva in half of the infected animals, which is not negligible.
To decipher whether STLV-1 infection impacts SFV replication, we then performed SFV proviral load analysis by qPCR in the blood and saliva of animals that were or not co-infected with STLV-1 (Fig 2). As for STLV-1, samples were classified as negative, positive but nonquantifiable (positive qPCR but LTR copy number beyond the amplicon standard range) or positive and quantified (positive qPCR and LTR copy number within the amplicon standard range). First, we did not observe any difference in the frequencies of negative, positive but non-quantifiable and positive and quantified samples among STLV-1-negative or STLV-1-positive samples in the saliva (Fig 2A, right columns). The quantified SFV PVL values in the Regarding blood samples, even if the frequency of positive samples was similar between STLV-1-negative and positive groups (13 vs. 12+4 animals), the frequencies of negative, positive but non-quantifiable and positive and quantified samples were significantly different (Fig  2A, left columns, p<0.0001). Indeed, there was no animal where the SFV PVL could be quantified among STLV-1-negative animals, while 75% (i.e. 12 out of 18) of samples were quantified among SFV/STLV-1 co-infected animals. Thus, the SFV PVL in the blood is much higher in STLV-1 co-infected animals, indicating that STLV-1 co-infection may impact SFV replication. These results also suggest that exposure to blood from co-infected animals should be linked to a higher risk of SFV transmission.
Finally, we compared STLV-1 and SFV PVL in blood and saliva of co-infected animals ( Fig  3). The frequencies of negative, positive but unquantifiable and positive and quantified blood samples were significantly different between STLV-1 and SFV (Fig 3A, p = 0.025). STLV-1 PVL were also significantly higher than quantified SFV PVL in PMBCs (Fig 3B, median value for SFV = 50,21 and median value for STLV-1 = 363,7, p<0.001). Similar analyses were performed with saliva samples. The frequencies of distribution in 3 categories (negative, unquantifiable-and quantifiable-positive animals) were significantly different between two viruses (Fig 3C, p = 0,0014). However, contrary to results obtained in the blood, the quantified PVL values were not significantly different between the two viruses ( Fig 3D, median value for SFV = 100,82 and median value for STLV-1 = 266).
Since it was shown that Tax protein is also present under a soluble cell-free form [50,51], it is therefore possible that such version of the viral protein is present in vivo in naturally infected animals where it could enter foamy infected cells. We then tested whether Tax can activate transcription from the foamy promoter in Jurkat T-cells (Fig 4). Indeed, and although less https://doi.org/10.1371/journal.pntd.0006812.g004 efficient than Tas, Tax can activate transcription from foamy-LTR-luc construct in a dosedependent manner. This observation might explain the observed increase SFV PVL in coinfected animals.

Discussion
In this report, we first tried to detect STLV-1 provirus in saliva samples obtained without scratching cheeks. Our results clearly demonstrate that STLV-1 DNA is detected in the saliva in 50% of all infected and asymptomatic animals tested. Of note, STLV-1 PVL in saliva is much lower than in PBMCs. Such a study has not been frequently performed with HTLV-1, and, to our knowledge, never with STLV-1. Some results have suggested that, when proviral HTLV-1 is detected in the saliva as well as in epithelial cells and lymphocytes present in saliva, neutralizing antibodies are also present and prevent HTLV-1 transmission through saliva [42,43,52,53]. However, given the fact that it is now well established that STLV-1 can be transmitted to humans through sever bites [8,41], it would be interesting to determine if there is a minimal STLV-1 PVL value in NHP saliva required for putative STLV-1 transmission through bite. Unfortunately, while blood samples from people who had been bitten have been obtained, neither the NHP source nor its saliva were available.
Given the high prevalence of SFV among NHPs, humans are bitten by NHP that are probably all SFV-positive, and, in some cases STLV-1-positive. Our results clearly demonstrate that co-infection has a positive impact on SFV proviral load in PBMCs but not in the saliva. Thus, the risk of becoming infected by SFV after being bitten by an SFV/STLV-1 co-infected animal, compared to a SFV-only infected animal, should not be different. However, people exposed to the blood of SFV/STLV-1 co-infected animals during bush meat consumption [54] likely have a higher risk of becoming SFV-infected.
Previous reports have shown that HTLV-1 infected cells produce soluble Tax protein that can act as a viral transactivator on cellular genes [50,51]. The observation of a positive effect of STLV-1 co-infection on SFV PVL suggested either that cells are truly co-infected with both viruses, or, alternatively, that soluble Tax transactivator enters cells already infected with SFV and acts on its promoter. Our results demonstrate that indeed Tax is able to activate transcription from the foamy LTR. Although this does not exclude the fact that true co-infection exists in animals, it is sufficient to suggest that Tax expression and capture leads to an increase in SFV viral replication and thus an increased SFV PVL. It remains to determine whether soluble Tax is secreted from STLV-1-infected cells in vivo, whether it escapes anti-Tax antibodies present in the plasma, and if it can activate SFV LTR in SFV-infected cells. Why this positive impact of STLV-1 co-infection is not visible in the saliva remains to be determined. Foamy infection has not been so far associated with a disease either in non-human primates or in humans. However, it is worth noting that foamy proviral load is usually low in vivo, and this might explain the lack of symptoms. The major difference between SFV proviral load in coinfected animals vs. mono-infected ones suggests that it would be of interest to pay more attention to these animals. As an example, one should perform autopsy and compare results with those of mono-infected animals in order to detect any change.
Of note, we did not have any SFV-negative and STLV-1-positive animal in our simian cohort. This prevented any study aimed at determining whether SFV infection is linked to an increased STLV-1 PVL compared to animals that would be STLV-1-mono-infected, although this would be of interest to gain a better insight into the mutual interactions between both retroviruses. Given that humans co-infected with both viruses or only with HTLV-1 are present in Africa, it would then be interesting to determine whether they are at higher risks for developing an HTLV-1 associated disease compared to HTLV-1 infected individuals. However, performing epidemiological studies will be difficult, due to the fact that these individuals are so far in very limited number.
In conclusion, using in vivo samples, our results demonstrate for the first time the presence of STLV-1 and SFV retroviruses in the blood and saliva of infected asymptomatic baboons. It would be of interest to perform similar studies among animals living in the wild, although this will be technically very difficult. Our results also strongly suggest that STLV-1 co-infection as a positive effect on SFV replication in the blood, thus suggesting that STLV-1 co-infection might allow a higher rate of inter-individual and possibly inter-species transmission of foamy virus after exposure to infected blood.