Putative progressive and abortive feline leukemia virus infection outcomes in captive jaguarundis (Puma yagouaroundi)

Feline leukemia virus (FeLV) is an exogenous gammaretrovirus of domestic cats (Felis catus) and some wild felids. The outcomes of FeLV infection in domestic cats vary according to host susceptibility, virus strain, and infectious challenge dose. Jaguarundis (Puma yagouaroundi) are small wild felids from South and Central America. We previously reported on FeLV infections in jaguarundis. We hypothesized here that the outcomes of FeLV infection in P. yagouaroundi mimic those observed in domestic cats. The aim of this study was to investigate the population of jaguarundis at Fundação Parque Zoológico de São Paulo for natural FeLV infection and resulting outcomes. We investigated the jaguarundis using serological and molecular methods and monitored them for FeLV-related diseases for 5 years. We retrieved relevant biological and clinical information for the entire population of 23 jaguarundis held at zoo. Post-mortem findings from necropsies were recorded and histopathological and immunohistopathological analyses were performed. Sequencing and phylogenetic analyses were performed for FeLV-positive samples. For sample prevalence, 95% confidence intervals (CI) were calculated. Fisher’s exact test was used to compare frequencies between infected and uninfected animals. P-values <0.05 were considered significant. In total, we detected evidence of FeLV exposure in four out of 23 animals (17%; 95% CI 5–39%). No endogenous FeLV (enFeLV) sequences were detected. An intestinal B-cell lymphoma in one jaguarundi was not associated with FeLV. Two jaguarundis presented FeLV test results consistent with an abortive FeLV infection with seroconversion, and two other jaguarundis had results consistent with a progressive infection and potentially FeLV-associated clinical disorders and post-mortem changes. Phylogenetic analysis of env revealed the presence of FeLV-A, a common origin of the virus in both animals (100% identity) and the closest similarity to FeLV-FAIDS and FeLV-3281 (98.4% identity), originally isolated from cats in the USA. We found evidence of progressive and abortive FeLV infection outcomes in jaguarundis, and domestic cats were probably the source of infection in these jaguarundis.


Background
Jaguarundis (Puma yagouaroundi) are small diurnal felids that have several unpatterned color morphsbrownish-black, gray and reddish-yellow furand are protected across most of their range. The species occurs at low densities and has a decreasing population in the wild, despite being widely distributed throughout South and Central America and occupying a broad range of habitats [1]. Mainly due to displacement from nature and, to a lesser extent, to captive breeding, jaguarundis are commonly found in zoos and similar captive settings in Brazil [2].
The feline leukemia virus (FeLV) is an exogenous, oncogenic, immunosuppressive gammaretrovirus that can establish persistent infections in domestic cats (Felis catus) [3]. The prevalence of FeLV ranges from 1% to 8% in healthy cats almost everywhere in the world and is usually higher when sick cats are included [4]. FeLV is naturally transmitted by oronasal exposure to viruscontaining secretions, mainly saliva, but also in feces and urine [5][6][7]. The effects of FeLV are cytoproliferative diseases including lymphomas and myeloproliferative disorders; degenerative illnesses, such as anemia and leukopenia; and immunosuppressive diseases associated with opportunistic infections [3,8].
FeLV infections tend to be rare or absent in many nondomestic felid species [9][10][11][12][13][14], except for the European wildcat (Felis silvestris silvestris), a species very closely related to the domestic cat, in which FeLV appears to be endemic [15][16][17]. However, documentation of FeLV is becoming more common in wild felid species less closely related to the genus Felis, which highlights the omnipresent threat that FeLV represents to the conservation of wild felids worldwide. FeLV has been shown to represent a major threat to the survival of critically endangered populations of Iberian lynxes (Lynx pardinus) in Europe [18,19] and to Florida panthers (Puma concolor coryi) in North America [20,21]. In Brazil, antibodies against FeLV have been detected in two freeranging pumas (Puma concolor) and two jaguarundis; FeLV DNA was detected in the following captive felids: an ocelot (Leopardus pardalis), an oncilla (Leopardus tigrinus) and two jaguarundis [22][23][24].
For domestic cats, the detection of the structural viral protein FeLV p27 in serum or plasma is used as a marker of infection and, in most cases, as a parameter for viremia. The outcomes of FeLV infection vary according to infectious challenge dose, route of challenge, and possibly host susceptibility and virus strain. The classification of FeLV outcomes in the domestic cat has been refined using sensitive molecular assays that detect and quantify proviral FeLV DNA and viral FeLV RNA, in addition to traditional serological and virological methods [25][26][27][28][29][30]. In this regard, the main host response categories of FeLV infection in domestic cats have been redefined as abortive, regressive, and progressive infection. In short, those cats that abort infection do not show any evidence of virus infection, except for seroconversion. Cats with regressive infection overcome viremia after an undetectable or transient initial phase by means of efficient cellular and humoral immune responses. These cats seroconvert, permanently harbor low to moderate FeLV proviral loads integrated in mononuclear cells, i.e., lymphocytes, and may or may not clear their plasma viral RNA loads [25,27,31]. Cats with progressive infection are constantly viremic, have low or no antibodies to FeLV, show elevated FeLV proviral and viral loads in peripheral blood cells and plasma and may develop FeLV-associated disease [25,26,28,29].
In previous surveillance work, we detected FeLV infection in two captive-born jaguarundis (#1 and #4) as well as previous exposure to FeLV in two other captive-born jaguarundis (#2 and #22) among a population of 23 jaguarundis held at Fundação Parque Zoológico de São Paulo (FPZSP), Brazil [23]. Jaguarundis #1 and #4 tested positive for FeLV p27 antigen by sandwich ELISA. FeLV proviral DNA was detected in blood from both animals by quantitative real-time polymerase chain reaction (qPCR). The sampling for this previous study occurred between 2003 and 2004 (Table 1).
By analogy, we hypothesized here that FeLV infection in P. yagouaroundi mimics the main outcomes observed for the domestic cat. Thus, the aim of this study was to perform additional serological and molecular tests and monitor the population of jaguarundis at FPZSP for FeLV infection and development of FeLV-related diseases for 5 years (2003)(2004)(2005)(2006)(2007).

Animals and sample collections
We retrieved relevant biological and clinical information for the entire population of 23 jaguarundis held at FPZSP, from birth or admission of the animals into the zoo until 2007. Most animals were mature adults (n = 16), five were immature and two were geriatric; kittens were absent. Three jaguarundis were wild born, while 20 were captive born at FPZSP. Six females (two of wild origin) and three zoo-born males were reproductively active. Eleven animals from the population were full siblings, and 16 were half siblings. The females (n = 9) averaged 4.28 ± 0.57 kg in body weight, and the males (n = 14) averaged 5.08 ± 0.49 kg in body weight. Detailed biological data such as age, sex, weight, origin (captive or wild born) and the parental history for the 23 jaguarundis are presented in Table 2.
For hematological analysis, several laboratories were used over time, and registries of results from the animals were not completed and standardized. All jaguarundis  Notes: FeLV-infected (#1, #4) and FeLV seropositive (#2, #22) jaguarundis are identified in bold. a For the wild-born jaguarundis (#3, #10 and #18), the dates refer to entrance in the zoo as their dates of birth are unknown. b Immature adults (n = 5) were defined as jaguarundis in the age range of 6 months ≤ age < 24 months for males and 6 months ≤ age < 36 months for females; mature adults (n = 16) were defined as those in the age range of 24 months ≤ age < 10 years (120 months) for males and 36 months ≤ age < 10 years for females; geriatric jaguarundis (n = 2) were those aged ≥10 years for both sexes. c For two captive-born jaguarundis (#8, #9), the identities of the father and the mother, respectively, were not available were vaccinated regularly against feline herpesvirus 1 (FHV-1), feline calicivirus (FCV), feline parvovirus (FPV) and rabies lyssavirus. We monitored the potential development of FeLV-related diseases through clinical data provided by the veterinarian staff at the zoo and registered at the zoo archives. Serological and molecular FeLV tests were performed, in addition to those previously published [23], using samples that had been previously obtained or additional samples collected throughout the observation period. FeLV tests previously performed included p27 FeLV antigen detection by a sandwich ELISA and the commercial immunoassay Snap™ Combo FeLV Antigen/FIV Antibody Test Kit (IDEXX Laboratories Inc., Westbrook, ME, USA), detection of antibodies against the recombinant non-glycosylated form of the FeLV gp70 surface glycoprotein (p45) by indirect ELISA, detection of antibodies against FeLV p12, p15, p27, p58, and gp70 antigens by Western blot and detection of FeLV proviral DNA by real-time PCR. We investigated the presence of antibodies to FeLV transmembrane protein p15E in serum samples from six jaguarundis (#1, #3, #4, #5, #14 and #17) using the previously described assay [32,33].
Six jaguarundis (#1, #2, #4, #5, #10 and #14) died between 2005 and 2008 and underwent necropsy at the zoo; for these jaguarundis, relevant clinical and post-mortem findings from necropsies were recorded, and several tissues were collected for analyses. Fragments of tissues from these animals were fixed in 10% formalin, embedded in paraffin wax and sectioned. The sections were stained with hematoxylin-eosin (HE) on glass slides for histopathological analyses. Tissue specimens from animals #1 and #5 (bone marrow, mesenteric lymph node and spleen) were also stored at ≤ −70°C. A saliva specimen from animal #1 was collected at first sampling using a sterile cotton swab immersed in virus transport media, which consisted of phosphate-buffered saline (PBS) balanced salt solution supplemented with 0.5% bovine albumin (BSA), antimicrobial agents (200 U/mL penicillin G, 200 U/mL streptomycin, 25 μg/mL fungizone and 6 μg/mL gentamycin) and was kept at ≤ −70°C. No additional material could be obtained from animal #2.

Immunohistology
To clarify the histogenesis of an intra-abdominal mass of tissue that developed in jaguarundi #5, we performed immunohistologic assays by the streptavidin-biotinperoxidase complex technique with a commercial immunoperoxidase kit (LSAB kit, Dako, Glostrup, Denmark) according to the manufacturer's protocol. We used silanized microscope slides to produce histologic slides, which were processed by usual deparaffination techniques in xylol and hydrated in alcohols (absolute, 95%, 70%, and distilled water). A microwave antigen retrieval technique using EDTA buffer, pH 9.0, was applied [34]. The sections were incubated overnight at 4°C with primary antibodies against the T-cell marker CD3 (rabbit polyclonal antibodies A0452; Dako; diluted 1 in 500), the B-cell marker CD79 (mouse monoclonal antibodies; clone HM57; M7051; Dako; diluted 1 in 500), and vimentin (mouse monoclonal antibodies; M0725; Dako; diluted 1 in 100) as well as antibodies against proliferating cell nuclear antigen antibodies (PCNA) (mouse monoclonal antibodies; M0879; Dako; diluted 1 in 800). For all reagents, negative controls were generated by substituting the primary antibody with a class-matched immunoglobulin. These procedures were conducted using the logistics of the Department of Pathology, School of Veterinary Medicine and Animal Sciences, University of São Paulo (USP), São Paulo, SP, Brazil.
We investigated the presence of FeLV antigens in several paraffin-embedded tissues from the deceased jaguarundis #1, #4, #5 and #14. This investigation was carried out using the methods and logistics of the Institute of Veterinary Pathology, University of Giessen, Giessen, Germany [35].

Extraction of nucleic acids
Total nucleic acid (TNA) was extracted from 100 μL of EDTA-anticoagulated whole blood or buffy coat or from 200 μL of EDTA-anticoagulated plasma using the MagNA Pure LC Total Nucleic Acid Isolation Kit (Roche Diagnostics, Rotkreuz, Switzerland). TNA was eluted into 100 μL of elution buffer and stored at −70°C until PCR testing was performed. RNA was extracted from the saliva, plasma samples collected from jaguarundi #1 and from serum samples from jaguarundis #1 and #4 using the QIAmp® Viral RNA Mini Kit (Qiagen). gDNA and RNA were extracted from tissues (bone marrow, mesenteric lymph node and spleen) from animals #1 and #5 upon necropsy, as previously described [36,37]. During all extractions, negative controls consisting of 100 μL of PBS were concurrently prepared with each batch of samples to monitor for cross-contamination.

Molecular tests for FeLV
FeLV viral RNA loads and proviral loads were investigated using real-time RT-PCR and PCR and primers targeting the U3 region of exogenous FeLV, as previously described [30]. For tissue samples, the total copy numbers detected per reaction were normalized to the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene as described [38].
FeLV subgroups A, B and C were investigated in the bone marrow, mesenteric lymph node, and spleen of the jaguarundi #1 and in the serum and plasma of jaguarundi #4 by conventional PCR using the FeLV-A specific primers RB59 and RB17; the FeLV-B specific primers RB53 and RB17 and the FeLV-C specific primers RB58 and RB47 as described [39,40].

Feline immunodeficiency virus (FIV) serology and molecular assays
Serum from jaguarundi #5 was tested for the presence of antibodies against feline immunodeficiency virus (FIV) using a commercial immunoassay Snap™ Combo FeLV Antigen/FIV Antibody Test Kit (IDEXX Laboratories) and by Western blotting, as described [41]. FIV real-time and conventional RT-PCR were performed from whole blood from jaguarundi #5, as previously described [42,43].

Sequencing of FeLV env from jaguarundis #1 and #4
FeLV env sequences were analyzed for the FeLV-positive jaguarundis #1 and #4 from TNA of previously collected buffy coats. For comparison, a second sequence obtained from bone marrow of animal #1 was analyzed 1.4 years later, at the time of euthanasia. For the analysis of the full-length FeLV env sequences, genomic DNA from these samples was amplified using PCR with the primers 5847F (5′ ACATATCGTCCTCCTGACCAC 3′) and 8197R (5′ GAAGGTCGAACCCTGGTCAACT 3′) [43], yielding an approximately 2´370 bp product. To ensure high-fidelity amplification, the PCR was performed using Phusion polymerase and HF buffer (Finnzyme, Ipswich, UK). PCR products were sequenced by Microsynth (Balgach, Switzerland) after purification using the GenElute PCR Clean-Up Kit (Sigma, Fluka GmbH, Buchs, Switzerland). The jaguarundi FeLV-A env sequence was submitted to GenBank [KR349469].

Phylogenetic analysis
Phylogenetic analyses were conducted using MEGA version 6 [45]. The FeLV env sequences were aligned using CLUSTAL W [46]. Bootstrap support (1000 replicates) was calculated by the neighbor-joining (NJ) [47] and maximum parsimony (MP) [48] methods, and results >70% were considered significant [49]. The MP tree was obtained using the subtree-pruning-regrafting (SPR) algorithm [48] with search level 1 in which the initial trees were obtained by the random addition of sequences (10 replicates). All positions containing gaps and missing data were eliminated from the dataset (complete deletion option).

Statistics
For sample prevalence, 95% confidence intervals (CI) were calculated. Fisher's exact test was used to compare frequencies between infected and uninfected animals. Pvalues <0.05 were considered significant.

Clinical data of all 23 jaguarundis
The population of jaguarundis kept at FPZSP from 2003 to 2007 consisted of 23 captive jaguarundis. The animals presented good general condition at sampling. Twenty animals presented discrete to moderate oral disorders, including one or more of the following conditions: gingivitis, presence of tartar, periodontitis, tooth loss and/or fractured teeth. No oral disorders were registered in three immature adults, #19, #21 and #23. At least once during the experiment, 15 animals presented episodes of diarrhea, 11 jaguarundis demonstrated transitory weight loss and three suffered from dermatological disorders, such as alopecia; for all these clinical conditions, the underlying causes of the problems were unknown. Despite regular deworming, intestinal parasites were detected in 18 animals during the study period. In hematological analysis, most animals (n = 16) showed some degree of leukopenia (marked: n = 1; mild: n = 8) and leukocytosis (marked: n = 2; mild: n = 5). Jaguarundi #1, which was FeLV positive, and jaguarundi #2, which was FeLV seropositive, showed mild transient leukocytosis. Several (n = 12) animals showed mildly increased packed cell volume (PCV) and two showed markedly increased PCV in at least one sampling. Jaguarundis #1 and #2 showed mildly increased PCV, and jaguarundi #22, which was FeLV seropositive, showed a markedly increased PCV. The increased PCV could be due to dehydration as the animals had been fasted before immobilization.

Deceased animals
Six jaguarundis (#1, #2, #4, #5, #10 and #14) in the investigated population died from 2005 to 2008. Jaguarundi #1 was euthanized 1.4 years after being diagnosed with FeLV infection [23]. This jaguarundi was euthanized because the zoo could not keep it segregated under good ethological conditions, although the animal was in good clinical condition. Jaguarundi #5 developed a large intra-abdominal mass of tissue about 1 year after testing negative for FeLV. After clinical and ultrasonographic examinations, the animal underwent laparotomy for excision of the mass and died 3 days after surgery due to clinical deterioration. Animals #2, #4, #10 and #14 died naturally during the period of study. For the six deceased animals, clinical conditions and main post-mortem findings are presented in Table 3.

Indirect ELISA for p15E
Antibodies against p15E antigen were not detected to a level that is considered positive for privately owned domestic cats and the results from all the jaguarundis tested (#1, #3, #4, #5, #14 and #17) were negative (Table 1).

FeLV-negative B-cell lymphoma in jaguarundi #5
The excised mass of tissue from jaguarundi #5 weighed approximately 0.25 kg and affected all layers of the wall of the jejunum. In addition, several small foci were present across the mesentery. There was multifocal necrosis as well as intense reactional fibrosis and proliferation of round cells arranged in dense blocks limited by delicate conjunctive septa. The cells were neoplastic and malignant and presented anisokaryosis, anisocytosis, and a high mitotic index. The tumor was found to be an intestinal B-cell lymphoma, as the cells stained positive for a B-cell marker (CD79) but not a T-cell marker (CD3) by immunohistology. No FeLV RNA or provirus was detected in the tissues collected upon necropsy, including tumoral tissues, bone marrow, mesenteric lymph nodes and spleen, and no FeLV antigens were detected in paraffin-embedded tissues by immunohistological analysis (Table 1). Because B-cell lymphomas can also be associated with FIV infection, jaguarundi #5 was tested for FIV by serology and PCR. The animal tested negative for antibodies to FIV in the commercial assay. In the FIV Western blot, one band against p24 was detectable.

Molecular detection of FeLV-A in two jaguarundis (#1 and #4)
FeLV RNA and proviral DNA were detected in two out of the 23 animals, namely, jaguarundis #1 and # 4. From jaguarundi #1, a saliva sample was available; it was FeLV RNA positive, indicating shedding of FeLV at the time of sampling. FeLV was also demonstrated in the plasma sample collected at a second sampling from jaguarundi #4 upon necropsy 1.4 years later, indicating persistent antigenemia in this animal. For jaguarundis #1 and # 4, respectively, the absolute copy numbers of proviral DNA were 4.26 × 10 7 and 3.49 × 10 7 copies of DNA/mL of buffy coat; the absolute copy numbers of viral RNA were 2.63 × 10 8 and 2.97 × 10 9 copies of RNA/mL of serum. The results from all FeLV tests are presented in Tables 1  and 4. Furthermore, FeLV could be demonstrated in the tissue samples collected at necropsy from jaguarundi #1, and FeLV antigens were confirmed by immunohistological analysis of paraffin-embedded tissue samples from deceased jaguarundis #1 and #4. For both jaguarundis, #1 and #4, FeLV-A was the only FeLV subtype identified; no FeLV-B or FeLV-C was detected.

No detection of enFeLV
No enFeLV was detected in the blood or buffy coat samples from any of the 23 jaguarundis.  (Fig. 1). When the env surface unit from the FeLV detected in jaguarundis was compared with those found in Iberian lynxes [EU293175 to EU293194], a sequence identity of approximately 97% was obtained (data not shown).

Discussion
We detected evidence of FeLV exposure in four (#1, #2, #4 and #22) out of 23 jaguarundis in the FPZSP (17%; 95% CI 5-39%). The remaining 19 jaguarundis were negative for FeLV in all serological and molecular tests performed. The population of jaguarundis presented clinical disorders that were common at FPZSP, as in any other captive setting in Brazil [2,12], and could be associated with a multitude of causes. Although the FeLV-positive jaguarundis also presented some conditions that could be related to FeLV [8], our data were insufficient to prove a causal association. Two captive-born male jaguarundis, the geriatric #1 and the mature adult #4, presented serological and molecular FeLV test results similar to the progressive FeLV infection outcome in domestic cats [25]. Jaguarundis #1 and #4 were both antigenemic. Antigenemia is usually used as a measure for viremia in domestic cats and is consistently found in domestic cats with a progressive FeLV infection outcome [26,28]. Additionally, blood, buffy coat, serum and tissue samples from both jaguarundis were positive for FeLV proviral DNA and viral RNA with high proviral and viral loads, respectively (Table 4). While the presence of proviral DNA attests to the integration of provirus into blood cells, the detection of viral RNA usually indicates replicating virus. Once the infection outcome has been established, high proviral and FeLV RNA loads are characteristically found in cats with progressive infection, while low loads are detected in cats with regressive infection that are provirus positive but not viremic at the time [25,29,50]. FeLV proviral and viral RNA loads were measured in the two FeLV-positive jaguarundis #1 and #4. Comparison of the FeLV tissue loads of jaguarundi #1 revealed loads similar to those reported earlier for domestic cats with progressive infection using identical methods. Viral RNA loads in serum from the two FeLV-positive jaguarundis were comparable to those described earlier in persistently antigenemic cats [30]. The viral RNA load in the clinically healthy jaguarundi #1 was slightly lower than that in jaguarundi #4, which showed weight loss, vomiting, anorexia and diarrhea before death. This is also similar to previous findings from a study in cats, comparing viremic healthy and viremic ill cats [51]. Moreover, consistent with findings in domestic cats with a progressive FeLV infection, no antibodies to FeLV antigens were detected in jaguarundis #1 and #4.
FeLV RNA and FeLV provirus were also detected in the saliva of jaguarundi #1. This finding is indicative of virus shedding; thus, jaguarundi #1 was a potential source of FeLV infection to other felids. Jaguarundi #1 was euthanized in good clinical condition as a biosafety measure. However, upon necropsy, histopathologic evaluation revealed splenic lymphoid depletion and muscular atrophy of the hind limbs; these findings may be associated with immunosuppression and neurological impairment in this animal. Jaguarundi #4 died presenting weight loss and gastrointestinal disorders including vomiting, anorexia and diarrhea. Histopathology revealed splenic alterations and evidence of enteritis (Table 3). Although these conditions are commonly associated with FeLV in domestic cats, it was not possible to ascertain whether they were caused by FeLV infection in this case.
Two captive-born jaguarundis, #2 and #22, presented test results similar to those reported for domestic cats with abortive FeLV infection and seroconversion as the only marker of FeLV exposure [28]. The female jaguarundi #2 had a history of direct contact with the two jaguarundis with progressive FeLV infection, #1 and #4: she was a sibling of jaguarundi #1 and had mated with the male #4.  Relationships at the protein level. The most parsimonious tree, with length = 637, is shown. The consistency index is (0.750000), the retention index is (0.747664), and the composite index is 0.620901 (0.560748) for all sites and parsimony-informative sites (in parentheses). There were a total of 605 positions in the final dataset. The percentages of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with the length being relative to the number of changes over the entire sequence. The MP tree was obtained using the subtree-pruning-regrafting (SPR) algorithm with search level 1 in which the initial trees were obtained by the random addition of sequences (10 replicates). All positions containing gaps and missing data were eliminated. Analyses were conducted in MEGA 6 [44][45][46][47]63] This would explain how she was exposed to FeLV, either at the same time as her sibling #1 or during mating with #4. Subsequently, jaguarundi #2 gave birth to jaguarundi #23, who did not show any signs of FeLV exposure (Tables 1, 2, 4). This may well be the case, since the mother, #2, had developed an antibody response and probably had never shed FeLV (abortive infection). It is unknown for how long and at what virus challenge conditions jaguarundi #2 was exposed, as animals were naturally and not experimentally infected in the present study. Overall, we speculate that jaguarundi #2 developed an abortive FeLV infection for similar (and possibly several) reasons that may drive FeLV abortive infections in some domestic cats. The parents of male jaguarundi #22 were the FeLV-negative male jaguarundi #5 and the female #7. Thus, it is unknown at what time point jaguarundi #22 was exposed to FeLV. Overall, we provide evidence of progressive and abortive FeLV infection in P. yagouaroundi. Similar results have been found in Florida panthers (Puma concolor coryi), which belong to the felid lineage 6 (genus Puma), the same phylogenetic lineage as the jaguarundis [52]. In Florida panthers, infection outcomes resembled those of domestic cats with progressive, regressive and abortive infection (previously persistent, regressive, and latent infection) [20,21]. We speculate that FeLV infection in jaguarundis is similarly unpredictable, and it is influenced for the same diverse aspects cited above for domestic cats.
Complimentary to the fact that FeLV causes various tumors in domestic cats [53], the literature has reported a FeLV-associated multicentric T-cell lymphoma in a captive non-domestic cheetah (Acinonyx jubatus) [54] and non-FeLV-associated T-and B-cell lymphomas in geriatric African lions (Panthera leo) [55]. These data motivated the search for possible FeLV involvement in the appearance of the neoplastic intra-abdominal mass in jaguarundi #5. Although jaguarundi #5 had not shown any evidence of FeLV infection intra vitam (Table 1), we further investigated whether FeLV antigens or proviruses were present in intestinal lymphoid cells and whether these cells produced FeLV viral RNA locally without it entering the bloodstreamas would be expected in a case of sequestered FeLV infection [4]. However, no FeLV antigens were detected by immunohistochemistry and no FeLV provirus or viral RNA were detected by molecular assays. Infection with FIV, which is a lentivirus associated with lymphomagenesis in domestic cats [56], was not detected by the commercial immunoassay, and only antibodies specific to FIV p24 capsid proteins were detected by Western blot, which may be indicative of an early or very late FIV infection or may have resulted from unspecific cross-reactivity. The FIV RT-PCR assays performed from whole blood were negative; however, this may be due to low viral loads or lack of specificity of the assays due to sequence diversity of different FIV isolates. In conclusion, jaguarundi #5 had developed a non FeLV-associated intestinal B-cell lymphoma; involvement of FIV in the development of the neoplasia cannot completely be ruled out.
Notably, a spleen sample from jaguarundi #5 had been previously analyzed for the presence of the FeLV receptor fTHTR1, which permits the virus to enter the cell. The fTHTR1 complementary DNA (cDNA) from that animal had shown 99% nucleotide and amino acid identity to the fTHTR1 cDNA sequences of the domestic cat and other wild felid species, such as the lynx (Lynx pardinus), African lion (P. leo bleyenberghi), Asiatic lion (P. leo persica), and European wildcat (F. silvestris silvestris) [57]. In addition, fTHTR1 was quantified by real-time PCR in a few tissues from two jaguarundis where tissue was available (the viremic jaguarundi #1 and jaguarundi #5, which developed a tumor, but was not FeLV-infected).These limited results showed no significant difference between the fTHTR1 tissue loads in cat and jaguarundi tissue (results not shown). However, it also needs to be mentioned that the real-time PCR assay was designed for the domestic cat THTR1 and not the jaguarundi THTR1. Two-point mutations can be found in the region of the assay: one-point mutation in the middle of the forward primer and one in the probe. Thus, although similar fTHTR1 expression levels were found in jaguarundi tissue compared with cat tissue, we cannot exclude differences in the efficiency of the real-time PCR assay. These findings support the perspective that the first phase of the FeLV virus cycleviral entrymight be similar among domestic cats and other wild felid species.
We did not detect enFeLV in the jaguarundis by qPCR. This is consistent with earlier data [58] in which enFeLV was not detected in jaguarundis. Accordingly, FeLV-B, which has greater pathogenicity than FeLV-A in domestic cats and arises by recombination of exogenous FeLV-A with enFeLV sequences [59], was not found in the two FeLV-positive jaguarundis #1 and #4. Moreover, the highly virulent FeLV-C was not detected. In FeLVinfected Iberian lynxes, enFeLV sequences were also not detected, but their FeLV-A variants were shown to be highly virulent, suggesting that the mechanisms inducing disease in these wild felids might be distinct from those in domestic cats [19].
The env sequence obtained from the FeLV infections of jaguarundis #1 and #4 [KR349469] clustered with the highest identity with the FeLV-FAIDS and FeLV-3281 strains, which represent members of a highly conserved group of horizontally transmitted, minimally pathogenic FeLV-A present in all naturally occurring infections in domestic cats. Both FeLV strains were originally isolated from cats in the United States of America (USA) [60,61]. Interestingly, the surface unit of the env sequence of FeLV detected in the jaguarundis also showed 97% identity to those of the env gene of FeLV isolated from the critically endangered Iberian lynxes [EU293175 to EU293194] in Spain, whose small wild population suffered from an outbreak of FeLV with several deaths [18].
Both PCR-positive jaguarundis, #1 and #4, were infected with a FeLV with an identical env sequence, indicating a common origin of the virus. It is unknown at what time point each of the jaguarundis became infected. The reproductive success achieved for jaguarundis in captivity is probably a result of the breeding program conceived for small felids at FPZSP [62] and a number of measures for improving the well-being of the animals, including the location of the jaguarundi enclosures in an isolated and forested area of the zoo away from the public. Nonetheless, domestic cats commonly are seen invading the zoo, especially the more isolated areas, and they were probably the source of infection of FeLV for the jaguarundis. Domestic cats are the main reservoir for FeLV worldwide and greatly outnumber FeLV-infected wild felids. Given this, it is prudent to prevent the presence of domestic cats in nature reserves, zoos and other settings with captive wild felids and to reinforce the importance of control measures for FeLV in domestic cats, such as testing and vaccination.
A severe FeLV outbreak occurred in a previously naïve population of Florida panthers (Puma concolor coryi) in North America from 2002 to 2005 in which five FeLV antigen-positive panthers died [20,21]. During a six-month period in 2007, six provirus-positive antigenemic Iberian lynxes (Lynx pardinus) died in Spain [18,19]. In both situations, FeLV vaccination programs were initiated. We recommend safe recombinant subunits or inactivated FeLV vaccines for captive jaguarundis, especially considering that invading domestic cats, potential sources of infection, are a frequent problem faced by zoos. In addition, we recommend the inclusion of FeLV testing in breeding programs for jaguarundis.

Conclusions
Our findings support our hypothesis that FeLV infection in P. yagouaroundi mimics the outcomes observed for the domestic cat. In addition, we found phylogenetic evidence that domestic cats may have been the source of the FeLV-A infection of the jaguarundis. Despite the minimal pathogenicity of FeLV-A for domestic cats, this virus may represent a threat to the jaguarundi population, and regular FeLV testing and vaccination are encouraged.