International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Contents lists available at ScienceDirect International Journal for Parasitology: Parasites and Wildlife journal homepage: www.elsevier.com/locate/ijppaw Equal contributions of feline immunodeficiency virus and coinfections to morbidity in African lions Heather Broughton a, *, Danny Govender b, Emmanuel Serrano c, Purvance Shikwambana b, Anna Jolles d a Department of Biology, Oregon State University Cascades, 1500 SW Chandler Ave, Bend, OR, 97702, USA b Scientific Services, South African National Park Services (SANPARKs), Kruger National Park, P.O. Box 176, Skukuza, Mpumalanga, 1350, South Africa c Wildlife Ecology and Health Group (WE&H), and Servei d’Ecopatologia de Fauna Salvatje (SEFaS), Universitat Autónoma de Barcelona (UAB), Facultat Veterinària, Placa Civica, Bellaterra, 08193, Spain d Department of Biomedical Sciences, Oregon State University, 1500 SW Jefferson St., Corvallis, OR, 97331, USA A R T I C L E I N F O A B S T R A C T Keywords: Feline immunodeficiency virus (FIV) is a pathogenic lentivirus related to human and simian immunodeficiency Panthera leo viruses that has been associated with AIDS-like pathologies in domestic and wild cats, as well as in hyenas. Concomitant infections Despite known pathologies, progressive immunosuppression and ill health effects driven by these lentiviruses in Wildlife diseases association with other secondary infections remain understudied in free-ranging species. Here, the role of coinfections by gastrointestinal parasites and tick-borne hemoparasites for FIV disease progression was explored in 195 free-ranging African lions (Panthera leo) living in Kruger National Park (KNP), South Africa. Using sta­ tistical methodology, we evaluated the effects of FIV on a range of health indicators to explore how direct and indirect effects of FIV and associated coinfections align to determine lion health outcomes. Findings show direct negative effects of FIV on host immunity and nutritional status, and exacerbation of aggressive behaviors, conditions which may increase exposure/susceptibility to other secondary infections. When taken together, the contribution of coinfecting parasites to morbidity in lions is of similar magnitude as direct effects of FIV infection alone, suggesting that the particular coinfection assemblage may play a role in mediating disease progression within natural lion populations. 1. Introduction found throughout Africa; North, Central, and South America; and Asia, as well as in domestic cats and hyenas (Troyer et al., 2005; VandeWoude Feline immunodeficiency virus (FIV) is a single-stranded RNA and Apetrei, 2006; Antunes et al., 2008). While characterization of FIV lentivirus that is closely related to well-known pathogens such as human infection in most wild felids is still on-going, evidence from natural and immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) experimental infection in domestic cats supports altered immune cell (Pedersen et al., 1987; Brown et al., 1994; Biek et al., 2003; Kanzaki and levels in some infected hosts, which mimics AIDS in humans (Hof­ Looney, 2004; Hartmann, 2011). While the pathogenicity of these vi­ mann-Lehmann et al., 1997; Elder et al., 1998; Hartmann, 2011). Recent ruses varies widely by strain, subtype, geographic location, and degree studies on FIV-ple (the lion-specific strain of the virus) in African lions of host-pathogen coevolution, the general hallmark of infection is a (Panthera leo) of Botswana suggest similar health effects including wide protracted decrease in CD4+ T helper cell levels with resultant immu­ scale immunosuppression, increased liver pathologies, and declining nosuppression, decreased resistance to parasitic exposures, progressive clinical health (Kennedy-Stoskopf et al., 2003; Roelke et al., 2006, weight loss, and increased mortality (VandeWoude and Apetrei, 2006; 2009). Despite these advances, potential health interactions between Pecon-Slattery et al., 2008; Hartmann, 2011, 2012; Troyer et al., 2011). FIV and other parasitic coinfections in this and other species with FIV is primarily transmitted via saliva to blood contact during limited access to disease control remain difficult to characterize. aggressive encounters and has been documented in 19 wild felid species As an immunosuppressive pathogen, FIV may cause significant * Corresponding author. Oregon State University Cascades, 1500 SW Chandler Ave, Bend, OR, 97702, USA. E-mail addresses: broughth@oregonstate.edu, dr2broughton@gmail.com (H. Broughton), danny.govender@sanparks.org (D. Govender), emmanuel.serrano@uab. cat (E. Serrano), purvance.shikwambana@sanparks.org (P. Shikwambana), anna.jolles@oregonstate.edu (A. Jolles). https://doi.org/10.1016/j.ijppaw.2021.07.003 Received 22 February 2021; Received in revised form 12 July 2021; Accepted 12 July 2021 Available online 19 August 2021 2213-2244/Published by Elsevier Ltd on behalf of Australian Society for Parasitology. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 morbidity in host species by lowering the threshold for invasion and heparin, Cyto-Chex® BCT, erythrocyte sedimentation rate (ESR) ®, and persistence of secondary infections (Bendinelli et al., 1995; Dean et al., whole blood (red-topped) tubes. Feces were collected directly from the 1998; Hartmann, 2012). Similar to HIV in humans, destruction of im­ rectum. A full physical examination was performed and the following mune cells and immune dysregulation may lead to increased coin­ demographics recorded: approximate age, sex, and pride membership. fections and inflammation; disease related nutritional deficiency may Age was estimated using established methods for lions (Smuts et al., limit the host’s ability to compensate by increasing investment in im­ 1978), and sex determined via visualization of genitalia. Pride mem­ munity; and organ pathology may break down mechanical barriers to bership was provided through known pride ranges and long-term parasitic invasion (Bendinelli et al., 1995; Dean et al., 1998; Hartmann, behavioral monitoring data provided by the parent study. 2012). In the face of these changes, alterations to sex steroids and glial cell populations of the brain may serve to upregulate aggressive be­ 2.3. Health, immunity, and behavior haviors, further increasing transmission potential (Fromont et al., 2003; Hawley et al., 2011; Mccaig et al., 2011; Tejerizo et al., 2012). Histor­ In order to evaluate health and immune status in FIV-positive and ically, the role of immunosuppression-driven secondary infections in negative individuals, this study used a combination of physical exam disease progression of FIV and other pathogenic lentiviruses has been parameters, blood biochemistry and endocrinology, and functional and difficult to assess, as management of immunodeficiency virus–associ­ quantitative immune assays. ated disease in model species such as humans and domestic animals typically includes preventing or treating coinfections (Bentwich et al., 2.3.1. Physical examination 2000). However, in natural animal populations not housed in the sterile Each lion was evaluated for the presence of lesions associated with indoor environment of a lab, the chronic nature of lentiviral infection is FIV or regionally endemic coinfections, which included gingivitis, oral likely to increase coinfection by common parasites due to the increased papillomas, ocular lesions, lymph node enlargement, dehydration, and window of exposure throughout an individual’s lifetime (Bentwich hygromas (Roelke et al., 2009). Body condition was assessed by a score et al., 1999; Bentwich et al., 2000; Ivan et al., 2013). Thus, under­ (BCS) ranked from one (poor) to five (ideal) using fat coverage and standing how parasite coinfection assemblage contributes to disease palpation of bony prominences over the ribs, sacrum, hips, and stomach progression is necessary to understand the impact of FIV in natural host (Ferreira and Funston, 2010). Oral papillomas were identified via populations, as these associations are likely to play a central role in visualization of large raised plaques on the gingiva, tongue, or oral mediating health outcomes of FIV disease. mucosa, while gingivitis was identified as inflammation or ulceration of To investigate the role of coinfections by gastrointestinal parasites the gingiva seen along the gum line and ranked on a scale of zero (no and tick-borne hemoparasites in FIV disease progression, we conducted inflammation) to four (marked inflammation and ulceration; Roelke a structured survey of 195 free-ranging lions in Kruger National Park, et al., 2009). Scores of two or above merited inclusion in the South Africa. Specifically, we asked: (1) what are the consequences of gingivitis-positive category. Ocular lesions were identified using light­ FIV for lion health in terms of immunity, maintenance of body condition, –facilitated examination of the cornea and anterior chamber of the eye aggressive behaviors, biochemical homeostasis, and clinical signs of for signs of corneal ulceration or infiltrate/increased vascularization disease; (2) to what extent does FIV infection alter the richness (total that may signify uveitis. A gross examination of each lion was used to number) of coinfecting parasite species; and (3) what role do coin­ determine the presence of dehydration (seen as sunken eyes and fections play in causing morbidity in FIV disease? Key findings suggest decreased skin turgor over the ribs), hygromas (soft fluctuant swellings that FIV plays a crucial role in structuring host health and susceptibility over the elbow), and peripheral lymph node enlargement (palpable through direct pathologies, but that the contributions of coinfecting enlargement of one or more lymph nodes). For further evaluation, de­ parasites may be as important or even more critical when predicting gree of systemic inflammation was determined from the number of disease progression and morbidity related to FIV. enlarged lymph nodes ranked on a scale from negative six (all nodes reduced) to six (all nodes enlarged). This score was calculated by adding 2. Materials and methods the number of enlarged lymph nodes and subtracting the number of reduced nodes detected in the submandibular, prescapular, axillary, 2.1. Ethics statement inguinal, and popliteal regions, as well as a separate “other” category that included lymph nodes that were only rarely palpable when Data and samples were collected between March 2010 and enlarged, such as the lateral retropharyngeal and accessory axillary. September 2013 from 195 free-ranging lions as part of a larger study on lion health, behavior, and demographics (PIs Funston and Ferreira). 2.3.2. Blood biochemistry and endocrinology Lion captures were conducted by South African National Parks veteri­ Blood was collected via venipuncture of the jugular or femoral veins nary staff under AUCC permit #FERSM5-767. Chemical immobilization and transported to the laboratory at ambient temperature. Serum was protocols used for sample collection were in accordance with SAN­ separated from whole blood within 8 h of collection and stored at − 60 ◦ C PARKS Animal Use and Care Committee (AUCC) approved guidelines, as until analysis. Liver (alanine aminotransferase [ALT], alkaline phos­ validated by previous clinical studies in this species (Jacquier et al., phatase [ALP]) and kidney (blood urea nitrogen [BUN], creatinine 2006; Wenger et al., 2010). [CRE]) biochemical markers, blood glucose (GLU), and total protein (TP) concentrations were assessed using the commercially available 2.2. Sample collection Prep II Profile (Abaxis Inc.; PN 500–7124) run on an Abaxis VS2 Vet Scan machine (Abaxis Inc., Union City, California, USA). When possible, all pride members over six months of age were Leptin and ghrelin were evaluated as a measure of body condition immobilized with a cocktail of either butorphanol (0.31 ± 0.034 mg/ (Kotler et al., 1984; Yamamoto et al., 1989; Freeman et al., 2004). Leptin kg), midazolam (0.21 ± 0.024 mg/kg), and medetomidine (0.052 ± is secreted at times of high nutrient availability due to increased body 0.006 mg/kg); or tiletamine-zolazepam (1.8 ± 0.5 mg/kg), and mede­ fat, while ghrelin is secreted in response to food scarcity and resulting tomidine (0.07 ± 0.01 mg/kg; Kyron Laboratories (Pty) Ltd, South Af­ stomach emptiness (Shibata et al., 2003; Martin et al., 2010; Depauw rica) delivered using a CO2 pressured dart rifle (Dan-Inject International, et al., 2012). Serum for the ghrelin assay was collected from whole blood Skukuza, South Africa; Jacquier et al., 2006; Wenger et al., 2010). After tubes, which were transported on ice until separation. Once separated, sample collection, animals received atipamezole (0.3 ± 0.1 mg/kg; HCl was added to acidify the sample to prevent protein breakdown and Orion Pharma, Finland) intramuscularly for reversal. During immobili­ all samples were stored at − 60 ◦ C until use. Plasma for leptin was zation, blood was collected via venipuncture into EDTA, sodium collected from heparin tubes and was also transported on ice until 84 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 storage at − 60 ◦ C. Both hormones were measured at Oregon State Michigan, USA; PN 582701) from serum samples stored at − 60 ◦ C. University using commercially available RIA kits following manufac­ turer instructions (Millipore Inc., Billerica, Massachusetts, USA; PN 2.4. Coinfections XL-85K and GHRA-88HK). The Leptin Multispecies RIA kit and Human Acylated Ghrelin RIA kit were used due to strong cross-reactivity with 2.4.1. Viral diagnostics the corresponding feline hormones (Martin et al., 2010; Depauw et al., Diagnostic virology was run on serum samples using in-house ELISAs 2012). validated for use in lions by van Vuuren et al. at the Department for Veterinary Tropical Disease, Onderstepoort Veterinary Institute in 2.3.3. Hematology and immunity Onderstepoort, South Africa (van Vureen et al., 2003). Exposure to Immune status of lions was examined using a combination of quan­ canine distemper, feline parvovirus, feline enteric coronavirus, and fe­ titative measures including clinical hematology, flow cytometry, line calicivirus was measured qualitatively using antibody-based ELI­ erythrocyte sedimentation rate, and bactericidal killing capacity. White SAs. Feline immunodeficiency virus was also detected using blood cell counts were obtained by electrical impedance counting via an antibody-based methods, but since FIV is a life-long infection, pres­ automated hematology analyzer (Vet ABC, Scil Animal Care Company, ence of circulating antibodies was presumed indicative of active infec­ Gurnee, Illinois, USA) from blood collected into EDTA tubes. Differential tion (Hartmann, 2011). Specificity of the FIV diagnostic is 100%, while counts of lymphocytes and granulocytes were calculated via blood sensitivity is lower at 78.6% (van Vureen et al., 2003). smears prepared using Hemacolor Diff-Quick Staining Solution (VWR International, Radnor, Pennsylvania, USA). Packed cell volume and total 2.4.2. Gastrointestinal parasitology solids were measured from heparinized blood spun at 15000XG in a Fecal parasites were detected through the use of standardized fecal hematocrit centrifuge (Gemmy Industrial Corporation, Taipei, Taiwan; sugar flotation and centrifugation techniques (Foreyt, 2001). An PN HKT-400) using microcapillary tubes and a refractometer (Optika, approximate number of worm eggs for each species of parasite was Ponteranica, Italy; PN HR-160). identified visually through microscopy and quantified as the number of Immunolabeling and flow cytometry were used to count populations eggs counted per gram of feces. Parasite families identified were solely of lymphocytes including CD4+ T-helper cells, CD8+ cytotoxic T cells, those detected using flotation methods and included nematodes, ces­ and CD21+ B cells, which are known targets of FIV infection (Roelke todes, and protozoa (for a complete list of those identified reference et al., 2006). Cells were stored as whole blood in Cyto-Chex® BCT cell Table 1). preservation tubes (VWR International, Radnor, Pennsylvania, USA; PN 87005-196) for no longer than 2 weeks at 4 ◦ C. For quantification, cells 2.4.3. Blood parasitology were labeled with feline-specific antibodies against CD4, CD8, or CD21 Hemoparasite DNA to be used for infection diagnostics was extracted cell surface markers and allowed to interact for 15 min. Following in­ from stored EDTA whole blood using a Qiagen DNeasy kit (Qiagen, cubation, red cells were lysed to prevent sample contamination using Hilden, Germany; PN 69506) and run using a polymerase chain reaction lysis buffer (500 mls of deionized water with 4.13g NH4Cl, 0.5g KHCO3 (PCR) based reverse line blot. This method was chosen due to its ability [Sigma Aldrich, St. Louis, Missouri, USA; PN A0171, 60339, respec­ to detect multiple parasites at once, as well as its capability for catch-all tively], and 0.185g Na4EDTA [Fisher Scientific, Waltham, Maryland, detection of unidentified species. The PCR membranes used for this USA; PN S657]). Antibodies were then conjugated to horse anti-mouse study contained existing markers for parasites previously identified in IgG-FITC over a second 15-min incubation period (Vector Labs, Burlin­ cheetahs and lions and included Theileria, Hepatozoon, Babesia, and game, California, USA; PN FI-2000). Control tubes with unlabeled cells Ehrlichia species (for a complete list reference Table 1). All PCR di­ were also provided for proper gating of cell populations. Felid-specific agnostics were run by M. Oosthuizen (Molecular Diagnostic Services, antibodies and protocols for their use in lions were provided by Peter Department of Veterinary Tropical Diseases, Onderstepoort Veterinary Moore at UC Davis (Leukocyte Antigen Biology Laboratory, University Institute, Pretoria, South Africa) according to established protocols in of California, Davis, CA, USA) and final counts for each subset of lions (Bosman et al., 2010; Kelly et al., 2014). lymphoid cells were performed by faculty at Onderstepoort Veterinary Institute’s Flow Cytometry Laboratory in Pretoria, South Africa. Erythrocyte sedimentation rate (ESR), which has been used as a gauge of systemic inflammation in late stage AIDS patients, was assessed following protocols validated in humans (Ndakotsu et al., 2008). Sam­ Table 1 ples were collected into citrate ESR tubes and processed immediately Coinfecting parasite species identified in African lions from Kruger National upon return to the lab (Greiner Bio-One, Lasec Laboratory and Scientific, Park, South Africa. Cape Town, South Africa; Vacuette ESR Systems). ESR was measured at COINFECTIONS 60 and 120 min post initiation of the assay. A bactericidal assay, which Hemoparasites Gastrointestinal Viruses measures the ability of whole blood (cellular and protein constituents) parasites or plasma (protein constituents) to kill bacteria, was also conducted as a Ehrlichia/Anaplasma Ascarid spp. Canine distemper virus measure of innate immune function using protocols established for other spp. wildlife species (Tieleman et al., 2005; Matson et al., 2006; Beechler Babesia felis Tapeworm spp. Feline calicivirus et al., 2012). Babesia leo Hookworm spp. Feline enteric coronavirus Babesia microti Whipworm spp. Feline parvovirus Babesia lengua Coccidia spp. Feline immunodeficiency 2.3.4. Behavioral metrics virus Aggressive encounters may speed the transmission of both FIV and Babesia rossi Toxoplasma gondii other parasites (Fromont et al., 2003; Hawley et al., 2011; Mccaig et al., Babesia vogeli 2011). AIDS in other species is often associated with an increase in sex Babesia canis Hepatozoon spp. Parasite abundance Parasite richness steroids that promote these behaviors (Tejerizo et al., 2012). As such, Theileria annae Worm eggs per gram of Number of parasite species wounds and testosterone levels were quantified as proxies for aggression feces per host in lions. Wounds from bites and scratches were identified visually and Theileria bicornis noted when severe enough to penetrate the skin or if there was evidence of suppurative inflammation. Testosterone was measured using a a Viral measures are based on seroprevalence (exposure) and do not denote commercially available EIA kit (Cayman Chemical, Ann Arbor, active infection. 85 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 2.5. Data analysis In the case of flow cytometric cell counts, sample size was prohibi­ tively small for FIV-negative individuals (n = 4). Thus, relationships 2.5.1. Statistical analysis between FIV status and CD4+, CD8+, and CD21+ cell counts were Direct associations between FIV and nutritional status, biochemical examined using t-tests (R Foundation for Statistical Computing, 2016) and clinical health, inflammation and immunity, aggression, and coin­ independent of other fixed effects. To account for confounding variables, fection were explored using a combination of linear mixed (LMM) and cell counts were also modeled against demographic parameters for age generalized linear mixed (GLMM) models with varying error and func­ and sex using generalized linear models with a Quasipoisson distribu­ tion link structures as dictated by the data. First, to examine the direct tion to account for overdispersion as indicated by the dispersiontest associations between FIV infection and nutritional status, models with function in the AER package (Kleiber and Zeileis, 2008; R Foundation for response variables for nutritional metrics (leptin, ghrelin, BUN, and Statistical Computing, 2016). Wound status, body condition score, and hydration status), were compared against fixed effects for FIV status, pride were not included as fixed effects in the full model to avoid further age, sex, and body condition, as well as two-way interactions between data partitioning. For quality control, animals with CD4+, CD8+, or age and FIV, sex and FIV, and a random effect accounting for pride CD21+ cell counts significantly greater than the gated T/B cell catch-all membership. Inclusion of this last variable was necessitated by the fact measure were also dropped from the final analysis. Model selection was that animals from the same pride are not fully independent (Bolker et al., conducted via manual backwards selection using AIC values and 2009) due to commonalities in disease exposures, prey availability, and p-values via the drop1 function (Kleiber and Zeileis, 2008). relatedness. Similar models were used to explore associations between FIV status and clinical or biochemical health, and included response 2.5.2. Path analysis variables for physical exam parameters (body condition score, gingivitis, Partial Least Squares Path Models (PLS-PM) are very useful in oral papillomas, ocular lesions, lymph node enlargement, dehydration, establishing causal relationships in observational studies on natural and hygromas) or biochemical measures (ALT, ALP, BUN, CRE, GLU, processes (Serrano et al., 2014), including the impact of diseases (Fan TP), paired against the previously mentioned fixed and random effects. et al., 2016). Two types of relationships are described in PLS-PM. The Models to test whether or not aggression increases in FIV positive lions first one involves relationships between the latent variables (e.g. a also followed this format and included the same independent variables combination of physiological parameters defining an abstract concept compared against the number of bites, scratches, suppurative wounds, such as “health status”), while the other considers the links between the and circulating levels of testosterone detected via physical examination latent variables (LVs) and its own block of manifest variables (MVs). or EIA methods. An increase in any of these variables was seen as a proxy These MVs are variables directly measured in the field or in the labo­ for increased aggression. Finally, evaluation of coinfection prevalence in ratory (e.g. body condition or leptin concentration). The PLS-PM FIV-positive and FIV-negative animals was evaluated individually for approach does not require strong assumptions with respect to the dis­ each parasite of interest (Table 1), as well as in combination for parasite tributions of the MVs, the sample size, or the measurement scale (Wold, richness (the number of parasite species per host) and abundance (the 1985). number of eggs/oocysts per gram of feces) compared against all inde­ For our purposes, we have defined five blocks of LVs in our path pendent variables of interest. model: FIV infectious status (0 for FIV-free and 1 for FIV-infected lions), While models for nutrition, aggression, coinfection, and health status immune response (IMM), hemoparasite infections (Hemoparasites), followed similar formats, models for immune parameters (hematology, helminth infections (Helminths), and morbidity (Morbidity). Because of packed cell volume, total solids, ESR, and bactericidal ability) required its importance for our research purposes, FIV infectious status was the inclusion of an additional parameter to account for wound-derived considered a latent variable with a single manifest variable (see exam­ systemic inflammation, which correlates significantly with circulating ples in Aleuy et al., 2020; Serrano et al., 2014). Descriptive statistics for numbers of immune cells. Thus, models exploring the direct effects of the set of manifest variables defining each block are shown in Table 2. FIV on immune dysregulation examined singular associations between Additional manifest variables describing immune response or health each immune parameter of interest paired individually against inde­ status that were initially included in our path model, but were finally pendent variables for age, sex, FIV status, body condition, suppurative excluded because of their low contribution to their own LVs, are also wounds, two-way interactions between FIV and age/sex, and a random shown in Table 2. Other viral (e.g., FCV, FPV, CDV) infections, as well as effect for pride status. Within models for bactericidal ability, this last Toxoplasma gondii, detected in our lions were initially considered for variable (pride status) was replaced by a random effect for “batch”, being included in independent LVs, but finally excluded for not which accounted for variability in bacterial plating between days (1–2 contributing to the path model due to low co-occurrence and overall prides were captured per day). prevalence in the population. Noting that old lions were more prone to All model assumptions were evaluated independently after exploring be infected with FIV (mean age of FIV-infected lions = 5.5 yrs whereas normality (via the Shapiro-Wilk’s Test), residual patterns, and data age of FIV-free lions = 3.9 yrs; t = - 4.5; df = 131.8; p < 0.01) the age of overdispersion (R Foundation for Statistical Computing, 2016; Kleiber animals was also included as a LV. and Zeileis, 2008), and models fit using either LMMs or GLMMs using When specifying the conceptual model, it was assumed that FIV and the nlme and lme4 packages, respectively (Bates et al., 2014; Pinheiro IMM should have both a direct and indirect impact (through Helminths et al., 2015). Model selection was based on stepwise comparison of AIC and Hemoparasites) on Morbidity. The vector of outer weights (Wj) values provided by the stepAIC function in the MASS package for LMMs associated at each block of MVs was obtained as the vector of the (Venables and Ripley, 2002), or the dredge function in the MuMIn regression coefficients in the multiple regression of the inner estimate of package for GLMMs (Bartoń, 2017). GLMMs were fit with a gaussian the LVs on their associated MVs (Tenenhaus et al., 2005). Each outer distribution and log link for continuous data; and a binomial distribution weight can be considered as a proxy for the importance of each MV in and logit link for binary data (Bartoń, 2017). Data was backtransformed the construction of the LV. The path coefficients (β) were obtained by for presentation. Final LMMs were modeled using REML (Pinheiro et al., least squares regression of connected LV scores, and they were inter­ 2015). As overdispersion was a common problem in count data, and preted as standard regression coefficients (Vinzi et al., 2010). Finally, means of Poisson modeled variables were not prohibitively small (>5), once the full model including all possible relationships between LVs discrete count variables were modeled using penalized quasi-likelihood (Table 2) was fitted, a model simplification was performed by removing GLMMs in the MASS package (Venables and Ripley, 2002) and back­ those MVs uncorrelated with their own LVs and later those relationships wards selection via Wald T statistics provided by the aod package between LVs with the lowest R2 values. The fit of the final model was (Lesnoff and Lancelot, 2012). Samples with zero values were removed to measured by the goodness-of-fit index (Tenenhaus et al., 2004). Model avoid zero inflation. parameters and fit indices were validated by bootstrapping. All the 86 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Table 2 Description of latent and manifest variables used to fit a partial least squares path model in order to examine the relationships between FIV infection, immune status, coinfection with helminths or hemoparasites, and morbidity in 195 lions from Kruger National Park, South Africa. Latent Variable Manifest Variable (Units) Descriptive statistics according to FIV infection status Reference Range Free Infected Feline Immunodeficiency Virus (FIV) 0 = FIV-free 27.20% 72.80% – 1 = FIV-infected n = 53 n = 142 Age (Age) Years 3.9 5.5 – Immune response (IMM) Leukocytes (cells/ul) 17305.2 ± 1505.9 (5000–35050) 14200.8 ± 547.9 (5350–29000) 7139–31045 Neutrophils (cells/ul) 10414.3 ± 970.4 (386.6–22256.3) 8654.7 ± 466.3 (583.3–28040.1) 3893–18545 Lymphocytes (cells/ul) 2731.1 ± 308.4 (31.4–5390) 1726 ± 109.1 (22.7–4011) 761–7220 Monocytes (cells/ul) 928.7 ± 139.4 (20.9–2430) 636 ± 62.5 (0–2804) 153–3606 Eosinophils (cells/ul) 1709.4 ± 299.2 (73.2–7495) 1049.6 ± 98.4 (0–3985) 217–4015 Basophils (cells/ul)* 37.8 ± 22.5 (0–350.5) 30.8 ± 9.49 (0–318) 0–1 Packed cell volume (%)* 38 ± 0.9 (27.3–45) 38.2 ± 0.5 (20–51) 33.7–46.7 Total solids (g/dl)* 8.4 ± 0.2 (8.0–9.5) 8.4 ± 0.1 (6.9–10.4) 7.5–9.1 CD4 (cells/10000) 513.3 ± 17.9 (375–658) 412.5 ± 16.4 (77–978) 242–745 CD21 (cells/10000) 434.7 ± 16.6 (223–542) 339.29 ± 16.7 (16–855) 119–1874 CD8 (cells/10000) 277.7 ± 13.7 (100–401) 255.6 ± 10.6 (20–709) 51–1140 ESR120 (units at 120 min) 49.3 ± 3.1 (8–104) 55.5 ± 2.3 (1–126) 7.8–107.6 BKaWB (BKA) 35.7 ± 3.25 (2.5–59.2) 22.8 ± 1.5 (0–57.6) – BKaPlasma (BKA) 15.2 ± 2.4 (0–53.9) 20.6 ± 1.2 (0–65.2) – Hemoparasite infections (Hemoparasites) Richness 2.4 ± 0.18 (0–4) 3.1 ± 0.10 (0–5) – (No. species/lion host) Helminth infections (Helminths) Burden (eggs/gram of feces) 4 ± 1.81 (0–81) 25 ± 6.6 (0–700) – Richness (No. taxa/lion host) 0.34 ± 0.08 (0–2) 0.83 ± 0.08 (0–3) – Health impairment (HealthIMP) Emaciation (- log BCS) − 1.14 ± 0.003 (− 1.2 to − 1.08) − 1.13 ± 0.002 (− 1.2 to − 1.04) – ALT (U/L) 23.3 ± 1.5 (4–54) 39.2 ± 1.9 (4–133) 6.6–91.4 ALP (U/L) 30.2 ± 3.6 (4–66) 32.1 ± 3.5 (4–129) 6.8–86.0 GLU (mg/dl) 92.3 ± 5.9 (29–272) 96.2 ± 2.9 (9.8–244) 42.6–151.0 BUN (md/dl) 48.1 ± 2.6 (19–96) 44.9 ± 1.42 (21–113) 8.3–78.5 TP (g/dl) 10.1 ± 0.2 (7.5–14.9) 10.1 ± 0.11 (7.2–15.1) – CRE (mg/dl) 1.5 ± 0.1 (0.1–3.2) 1.8 ± 0.14 (0.1–12.5) 0.8–2.6 GHR (pg/ml) 21.3 ± 7.3 (10.5–35.2) 40.8 ± 37.3 (9.5–178.6) 11.4–37.5 Lept (ng/ml) 3.1 ± 0.6 (0.07–27.1) 2.2 ± 0.3 (0.15–20.3) 0.3–4.8 Aggression* Testosterone (ng/ml)* 12.1 ± 5.0 (0.6–91.8) 17.8 ± 3.9 (3.8–171.9) 0.06–10.5 a Values in the table describe the sample mean ± standard error, range minimum, and range maximum for each manifest variable included in the full path model. Asterisked parameters were measured for the purposes of this study, but were not included as manifest variables in the path analysis. Manifest variables marked in bold represent those that made a statistically significant contribution towards predicting the outcome of their respective latent variable. Only those values were retained for the final path analysis. Reference intervals for wild lions are provided for comparison and were generated using clinically healthy animals from the same sample population (Broughton et al., 2017). variables have been standardized (mean = 0, var = 1) for our PLS-PM positive hosts (Fig. 2a; Table 2 and S1Table). purposes. The final sample size in our PLS-PM was 106 lions once excluding individuals with missing information in our manifest vari­ 3.2. FIV infected lions show biochemical alterations and clinical signs ables. This statistical procedure was performed using the package suggestive of feline AIDS ‘plspm’ version 0.4.736 (Sanchez et al., 2017) of the statistical software R version 4.1.0. (R Foundation for Statistical Computing, 2021). FIV infection was associated with an increase in liver enzymes (ALP and ALT; n = 193; β = 1.89 and 1.65; p-values < 0.01 and 0.01, 3. Results respectively; Table 2 and S1 Table); an increased prevalence of ocular lesions (n = 184; β = 10.19; p-value < 0.01); and increases in gingivitis FIV infection was common in KNP lions, affecting 142 of the total with age (n = 181; β = 1.63; p-value = 0.04). Contrastingly, FIV infection 195 sampled (72.8%; 95% CI = 66.5–79.1%). Prevalence of FIV infec­ was associated with a lower prevalence of oral papillomas (n = 180; β = tion was similar among male (75.6%; 95% CI = 65.8–85.3%) and female 0.50; p-value = 0.04; Fig. 2a; S1 Table). (70.2%; 95% CI = 62.1–78.3%) lions and between park regions, but increased with host age (n = 195; β = 1.29; p-value < 0.01; for a com­ 3.3. FIV alters inflammation and immunity plete demographic breakdown reference Fig. 1). Overall, results of this study showed negative effects of FIV infection on host nutrient balance, FIV-positive lions showed an increase in erythrocyte sedimentation clinical health, and immunocompetence, as well as alterations to rate (n = 100; β = 3.10; p-value = 0.02); and higher total solids (n = 94; endocrine and behavioral parameters that may reflect increased β = 1.04; p-value = 0.05), suggesting elevated protein levels. Infected aggression in FIV-positive hosts. animals also showed a reduction in total lymphocyte count (n = 90; β = 0.58; p-value = 0.04; Fig. 2a–b; Table 2 and S2 Table). While no dif­ 3.1. FIV infected lions are malnourished ference was detected in flow cytometric counts for CD4+, CD8+, and CD21+ cells based solely on FIV status (n = 57; t = 1.69, 0.52, and Lions infected with FIV showed higher levels of serum ghrelin (n = − 0.94; p = 0.17, 0.63, and 0.41, respectively; Table 2 and S3 Table), 194; β = 3.40; p-value < 0.01), indicating low frequency of feeding; CD4+ and CD8+ helper cell populations decreased with age (n = 57; β = lower levels of leptin (n = 194; β = 0.23; p-value < 0.01), indicating 0.94 and 0.94; p-values < 0.01 and 0.02, respectively; Fig. 2c), sug­ decreased fat reserves; and reduced blood urea nitrogen (n = 193; β = gesting progressive age-related impairment of adaptive immunity, 0.88; p-value = 0.04), reflecting reduced protein intake. Together, these including humoral and cell-mediated responses (S3 Table). Monocytes findings suggest impaired resource acquisition or retention in FIV- and neutrophils also showed progressive declines with age, with a more 87 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Fig. 1. FIV infection is common in lions of Kruger National Park. Within the sample population, 72.8% (142/195) of lions were infected with FIV. Prevalence of FIV infection was similar between males and females (75.6% versus 70.2%; n = 74 and 121, respectively), but increased with host age (a). For ease of visualization, age has been broken up by life stage into cubs (0–2yrs), subadults (2.1–4yrs), young adults (4.1–6yrs), prime adults (6.1–8yrs), and seniors (>8yrs) based on previous age classifications (Schaller, 1976). Regional prevalence of FIV was highest in the central region and lowest in the north (b). The map to the right shows locations where lion prides were sampled. drastic decrease in FIV-infected lions, supporting a reduction in innate strikingly common in FIV-positive lions (Fig. 3a). In addition to T. gondii immunity (n = 90; β = 0.85 and 0.94; p-values < 0.01 and 0.03, and whipworms, infections with tapeworms (Taenia and Echinococcus respectively; Table 2 and S2 Table). spp.; n = 114; β = 2.87; p-value = 0.03), hookworms (Ancylostoma spp.; n = 114; β = 8.51; p-value = 0.01), and coccidia (Isospora spp.; n = 114; 3.4. FIV may increase aggressive behaviors or delay tissue healing β = 221.02; p-value = 0.02) were increased among FIV-positive lions (S5 Table). Total gastrointestinal parasite burden reflected this trend and FIV-positive lions exhibited a higher risk for severe wounds when was increased in FIV-positive hosts (n = 104; β = 4.73; p-value = 0.03), compared to FIV-negative individuals (n = 193; β = 3.76; p-value = largely driven by a marginal increase in the intestinal burden of hook­ 0.04), suggesting increased fighting, more severe outcomes of fights, or worm species (n = 48; β = 4.67; p-value = 0.09; S5 Table). Finally, reduced ability to heal wounds. Male FIV-infected lions had higher overall richness of gastrointestinal parasites was significantly higher in testosterone levels (n = 73; β = 3.36; p-value = 0.03) than their FIV- FIV-positive hosts (n = 114; β = 1.61; p-value = 0.02; Fig. 3b). negative counterparts, again pointing towards a potential for The prevalence of Babesia microti, a tick-transmitted hemoparasite, increased aggression; however, FIV-positive females had lower testos­ was also higher in FIV-positive lions (n = 190; β = 2.19; p-value = 0.05; terone levels than healthy females (n = 119; β = 0.44; p-value < 0.01; Fig. 3a), as was hemoparasite richness (n = 190; β = 1.35; p-value = Fig. 2a; Table 2 and S4 Table). 0.04; Fig. 3c). Interestingly, despite these trends, overall parasite rich­ ness, including both gastrointestinal and hemoparasites, was not sta­ 3.5. FIV increases the richness and abundance of coinfecting parasites tistically different in FIV-positive and FIV-negative lions (n = 111; β = 1.17; p-value = 0.12; S5 Table), suggesting a trade-off in infection This study measured exposure to 21 different parasite taxa known to prevalence between groups. infect lions. Overall, infections were common and diverse among KNP lions, with individuals typically hosting from one to twelve of the taxa 3.6. FIV affects lion health directly and indirectly via immunity and we identified (median = 7; IQR = 3; Fig. 3a). Several of the taxa were coinfections detected only in FIV-infected lions: feline enteric coronavirus (4% prevalence), whipworms (Trichuris spp.; 9% prevalence), Toxoplasma In our path analysis, we asked whether progressive wasting and gondii (12% prevalence), Theileria annae (0.07% prevalence), and Thei­ organ damage associated with FIV are mediated directly due to FIV leria bicornis (4% prevalence). Gastrointestinal parasite infections were infection, or indirectly via increased coinfections. Our models reveal 88 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Fig. 2. FIV has broad effects on lion health and is associated with progressive immune impairment. Box (2a) above provides a complete list of the health metrics measured for the purposes of this study. Arrows to the right of each variable summarize the directionality of statistically significant changes with FIV infection. Descriptive statistics and reference values can be found in Table 2. Complete model output can be found in supplementary tables S1–S4. For ease of visualization, each parameter has been broken into categories of clinical relevance. To the right, age-related changes in lymphocyte profiles are shown for total lymphocyte counts in FIV-positive versus FIV-negative lions (b); as well as specific lymphocyte subsets in FIV-positive lions (c). Due to small sample size for lymphocyte subsets, subadults, and cubs have been included together under the ‘juvenile’ category. three major findings. First, FIV, hemoparasites, and gastrointestinal and biochemical changes commonly associated with feline AIDS parasites have direct negative effects of similar magnitude on lion health (FAIDS) (Roelke et al., 2009; Hartmann, 2011, 2012). Importantly, (Fig. 4 and Tables 3a and 3b), but they explain different aspects of findings suggest that these morbidities may be driven in similar parts by morbidity. All three contribute to emaciation based on body condition FIV itself, and by gastrointestinal and hemoparasitic coinfections, which score, but FIV and hemoparasites are associated with liver pathology appear to be facilitated by FIV-related immune dysfunction. Overall, (ALT) while GI parasites are correlated with the most variation in leptin these findings suggest that coinfections may play a strong role in shaping and ghrelin levels (Fig. 4, Table 3a). Second, FIV-related immune FIV-mediated disease progression in lion hosts. impairment itself contributes strongly to ill health in lions in addition to Using quantitative metrics, alterations in immune parameters were the infections that we measured (Fig. 4). In fact, the effects of immunity common in FIV-positive hosts. Total lymphocyte counts were lower; on health parameters are equal to, or stronger than, direct effects of FIV CD4+ and CD8+ T cell populations decreased with age in infected ani­ and coinfections (Tables 3a and 3b). Negative health trends are associ­ mals, as did counts of innate effector cells including monocytes and ated primarily with low B cell populations and inflammation (ESR; neutrophils; and inflammation, as measured by erythrocyte sedimenta­ Tables 3a and 3b). Finally, FIV shows strong indirect associations with tion rate (ESR) and total solids, was elevated. Observed changes to im­ hemoparasitic and gastrointestinal coinfections through its effects on mune parameters were consistent with trends previously identified in host immunity (Fig. 4 and Tables 3a and 3b). GI parasite coinfections domestic cats infected with FIV-fca, which includes broader cell tropism show strong associations with low B cell populations and increased than HIV and SIV, a difference that leads to infection and destruction of inflammation (ESR); whereas hemoparasites were associated with low B lymphocytes, T lymphocytes, regulatory T cells, and macrophages general populations of lymphocytes, as well as poor bactericidal activity (Ackley et al., 1990; Brown et al., 1991; Yamamoto et al., 2007, 2010; of whole blood (Tables 3a and 3b). Reggeti et al., 2008; Elder et al., 2008; Murphy et al., 2014). Immune dysfunction may be further supported in FIV-positive lions by elevations 4. Discussion in ESR and total solids. In humans and domestic cats, these elevations have been shown to correlate well with a shift towards general immune As indicated by our results, FIVple-infection in lions with viral sub­ overstimulation – a compensatory response to immune impairment due types A, D, or E (as identified via PCR in a partner study; Kerr et al., to CD4+ T-helper cell depletion (Bendinelli et al., 1995; Ndakotsu et al., 2018) is characterized by immune impairment and dysfunction associ­ 2008; Roelke et al., 2009; Hartmann, 2012). ated with nutrient imbalance and progressive wasting, as well as clinical Coinfections were overwhelmingly common in the lion population. 89 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Fig. 3. FIV significantly increases the prevalence of select gastrointestinal and hemoparasitic coinfections, as well as overall parasite richness for both groups. Graph (a) shows the prevalence of coinfecting parasites isolated in lions from this study. Sample size for the parasite groups included is as follows: n = 114 for gastro­ intestinal parasites; n = 190 for hemoparasites; and n = 195 for viral parasites. Coinfections are broken down by FIV status (positive versus negative). The two additional graphs illustrate the relationship between FIV status and gastrointestinal parasite richness (b) and hemoparasite richness (c). FIV-positive lions regularly exhibited increases in the richness and samples despite a lack of affinity for the PCR markers. Of the species abundance of co-infecting parasites, which in-turn may contribute to FIV detected via PCR, hemoparasite coinfections increased as immune cell disease progression. Positive associations with FIV were particularly populations and bactericidal capacity of blood decreased, likely striking for the gastrointestinal parasites we investigated, with preva­ reflecting the role of adaptive and innate immune defenses in limiting lence of hookworms, tapeworms, and coccidia increased, and whip­ the invasion, growth, and persistence of these intracellular parasites worms and Toxoplasma gondii exclusively found in FIV-positive lions. (Olivier et al., 2003; Levinson, 2012). Hemoparasites may thus benefit Overall intestinal parasite burdens were increased with FIV infection directly from a decrease in humoral and cell-mediated immunity and were driven largely by increased egg shedding by hookworm species brought on by clinical AIDS (Sher et al., 1992; Dean et al., 1998; Olivier based on parasite eggs/gram detected in feces. While mechanisms et al., 2003). explainging these findings have not been fully identified, observed Another hallmark of AIDs in humans and FAIDS in cats is progressive changes may be due to impaired gastrointestinal immunity, as similar wasting and malnourishment. Here, these findings were evident through lentiviruses (HIV and FIV-fca) have a strong affinity for CD4+ T cells high levels of ghrelin, a gastric hormone released in response to nutrient exhibiting either CCR5 (HIV) or CXCR4 (FIV-fca) chemokine coreceptors deprivation and stomach shrinkage; and low levels of leptin, which is commonly found in gut associated lymphoid tissue (GALT) (Reggeti produced by adipose cells in respone to high nutrient balance (Appleton et al., 2008; Hartmann, 2011). As such, the correlations between et al., 2000; Shibata et al., 2003; Ida et al., 2007; Klok et al., 2007; FIV-infection and gastrointestinal parasite prevalence may reflect a Martin et al., 2010; Depauw et al., 2012; Jensen et al., 2015; Witzel breakdown in gut immunity as observed with HIV in humans and FIV in et al., 2015). Further support for malnourishment was provided by domestic cats (Willett et al., 2006; Walson et al., 2009; McSorley and consistently low levels of blood urea nitrogen (BUN) in FIV-positive Maizels, 2012; Murphy et al., 2014; Costiniuk and Angel, 2012). hosts, even in the face of dehydration. While BUN is more frequently Alongside gastrointestinal parasites, the richness of hemoparasite used to evaluate renal health, low values of BUN may reflect decreased infections was increased in FIV-positive lions. This alteration was most protein balance and subsequent muscle wasting due to insufficient apparent in the prevalence of Babesia microti, which was significantly alimentary intake or increased losses through the kidneys and intestines higher in FIV-infected hosts. B. microti infection occurred against a (Thrall, 2004; Stockham and Scott, 2008). Of the parasites diagnosed, GI backdrop of other Babesia, Ehrlichia, Theileria, and Hepatozoon species parasites contributed most profoundly to these changes and were more that were also prevalent in combination, including one unknown Hep­ tightly associated with low levels of leptin and high levels of ghrelin atozoon species that was visually identified in most of the blood smear than FIV itself or hemoparasites. Given that reduced immunity strongly 90 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Fig. 4. FIV has strong direct and indirect effects on overall health. The final path model (a) shows only statistically significant relationships between manifest variables (rectangles) and latent variables (circles) for FIV infection, immune response (IMM), co- infections with hemoparasites (Hemopar­ asites), co-infections with gastrointestinal parasites (Helminths), and morbidity (Morbidity). Note that the parameter βx,z between each variable of interest represents the path coefficient obtained from least squares regression examining the relation­ ship between one latent variable and the next (for example, FIV to IMM). Blue arrows represent positive relationships, whereas red arrows denote negative relationships. Final sample size was 106 lions. and positively affects the prevalence, richness, and abundance of GI et al., 2009). We worked with live animals only, and so were not in a parasites, these findings suggest that lions with FAIDS may maintain position to evaluate liver histology; however, the tight association of FIV condition while they are able to mount compensatory immune re­ and GI parasite infections in our study is consistent with this explana­ sponses, but that metabolic homeostasis may decline as B and T cells tion. Contrastingly, increased richness and prevalence of hemoparasites wane. contributed much more strongly to elevations in ALT alongside clinical In AIDS patients, systemic inflammation and antigen deposition wasting despite normal nutrient balance (as measured through leptin along filtration barriers due to primary and secondary infections often and ghrelin), suggesting a dominant role for hemoparasites in driving leads to extensive immunopathology in the form of glomeruloneph­ liver pathology in the form of hepatocellular injury (Reyers et al., 1998; ropathies and enteropathies (Yamamoto et al., 1989; Freeman et al., Solano-Gallego et al., 2016). In addition, while some changes in 2004; Roelke et al., 2009). While markers for these pathologies were biochemistry and endocrine markers were subclinical, in that they did included in our biochemistry panels, BUN levels in FIV-positive lions not fall outside reference intervals established for healthy wild lions, typically remained within the normal range (Broughton et al., 2017), as clinical signs of FIV infection including gingivitis and ocular lesions did creatinine, suggesting that they do not suffer from overt renal pa­ were present in FIV-positive lions, especially as they aged. Within the thology. Liver enzymes were increased with FIV-infection relative to context of environmental variation to which wild animals are exposed, uninfected lions, but again were within the normal range previously particularly variation in food availability and parasitic exposures, the established for this species (Broughton et al., 2017). Recent studies on observed subtle, subclinical changes in organ function and minor clin­ the pathological mechanisms of FIV in lions of Botswana found similar ical manifestations may thus be relevant to animal health (Munson et al., trends in liver enzymes, which were attributed to parasite visceral larval 2008). migrans as diagnosed via gross assessment and histopathology (Roelke While this was one of the most comprehensive predator disease 91 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Table 3 3a and 3b. Correlations and % contributions between latent and manifest variables describing the final partial least squares path model examining the relationships between FIV infection, immune status, coinfection with helminths or hemoparasites, and morbidity in 195 lions from Kruger National Park, South Africa. Box 3a: Final cross correlations between significant manifest variables (MVs) and their respective latent variables (LVs) Variables FIV Immunity (IMM) Hemoparasites Helminths Health Impairment (HealthIMP) Age 0.25 − 0.28 0.34 − 0.11 0.12 Immunity (IMM) − 0.46 1 − 0.38 − 0.39 − 0.43 CD4+ − 0.38 0.62 − 0.48 − 0.05 − 0.11 CD21+ − 0.29 0.88 − 0.25 − 0.46 − 0.45 BKAWB − 0.39 0.57 − 0.39 0.01 − 0.16 ESR120 0.28 − 0.56 0.03 0.41 0.23 Hemoparasites 0.22 − 0.39 1 − 0.01 0.29 Helminths 0.2 − 0.4 − 0.01 1 0.28 Burden 0.19 − 0.29 0.03 0.86 0.29 Richness 0.19 − 0.41 0.02 0.95 0.24 Health Impairment 0.36 − 0.43 0.29 0.28 1 Emaciation 0.15 − 0.15 0.18 0.11 0.44 ALT 0.34 − 0.36 0.29 0.15 0.86 Ghrelin (GHR) 0.08 − 0.19 − 0.01 0.28 0.4 Leptin (Lep) − 0.04 0.14 0.01 − 0.15 − 0.23 Box 3b: Contribution (%) of each latent variable (LV) to global observed variability in health impairment Latent Variables β Correlation Contribution to global R2 (%) FIV 0.2 0.36 27 Immunity − 0.24 − 0.424 38 Helminths 0.16 0.28 17 Hemoparasites 0.17 0.29 18 *Relationships between latent variables are denoted by bold face font. studies to date, capture logistics within a free-ranging predator popu­ 5. Conclusions lation did impose several limitations. First, we were unable to investi­ gate the role of directly transmitted viruses in FIV disease progression With a sample population of 195 wild lions out of 1700, which because our diagnostics for viruses were limited to detection of anti­ equates to 11.5% of the total population living in Kruger National Park, bodies at a single time point. As such, we cannot be sure whether viral as well as diagnostic and health data on 22 different parasites, ours is exposures were current or occurred at some point in the past and one of the largest and most comprehensive infectious disease studies of whether they occurred prior to, or after FIV infection. Directly trans­ wild predators. Findings from this study significantly extend current mitted viruses would be of particular interest here due to the potential knowledge of FIV’s immune effects in lions, as well as its potential role feedback between increased aggressive behaviors in FIV positive hosts for structuring parasite communities via these immune alterations. Both and enhanced transmission – both of FIV itself, and of other directly FIV’s direct mechanisms, and indirect pathology mediated by the higher transmitted infections. Indeed, we found evidence of increased aggres­ richness and abundance of other parasites in FIV-infected individuals, sion or delayed wound healing particularly in FIV-infected male lions, may in turn contribute to this pathogen serving as a key mediator of host which showed a higher prevalence of combative wounds and elevations pathology through organ injury, malnutrition, and further immune in circulating levels of testosterone. These results are consistent with dysfunction. Most importantly, this is the first study to show strong findings in domestic cats with FIV and humans with HIV, which show supportive evidence that secondary infections may play an essential part upregulation of sex steroid levels and an increased propensity to engage of FIV disease progression in lions, rather than just a symptom, a finding in “risky” behaviors (increased fighting in cats and sexual encounters in that may also have clinical relevancy for other lentiviral impacted ani­ humans; Fromont et al., 1997; Podell et al., 1999; Lloyd-smith et al., mal and human populations with high exposure to neglected parasites. 2005). Based on lion behavioral ecology, similar findings in this species would be likely to increase host contact rates, with the possibility of Declaration of competing interest increasing transmission of FIV and other directly transmitted infections (Fromont et al., 1997, 2003). In addition, this study’s finding of strong The authors declare that they have no known competing financial direct effects of FIV-driven immunosuppression on host health (inde­ interests or personal relationships that could have appeared to influence pendent of coinfections) could also be related to the diagnostic limita­ the work reported in this paper. tions of this study, which failed to detect relevant parasites. Future studies would benefit from the use of NextGen Sequencing and longi­ Acknowledgements tudinal data collection in order to assign causal pathways with more certainty as they pertain to coinfections. Lion captures and sample collection for this study would not have Finally, while we attempted to avoid sampling bias by catching been possible without the supporting infrastructure of a larger study on entire prides and conducting captures across all areas of KNP, our lion behavior, demographics, and disease dynamics (PIs Dr. Paul Fun­ sampling methods were not truly random. Lions that were too weak to ston and Dr. Sam Ferreira), as well as the continued efforts and expertise respond, or to compete with other lions at the carcass, were unlikely to of staff and researchers from SANParks Veterinary Wildlife and Scien­ be sampled. We may thus have missed the most emaciated individuals, tific Services. Specifically, we would like to thank Drs. Sam Ferreira, as well nomadic sub adults, as these animals often do not associate with Peter Buss, Michelle Miller, and Markus Hofmeyr, as well as technician a pride (Schaller, 1976). On the other hand, lions that had recently eaten Jennifer Joubert-Hofmeyr, for their continued diligence and help initi­ might not respond to a bait carcass, and as such we might also have ating and orchestrating this project. In addition, collection of informa­ missed especially well-fed individuals with higher fitness. tion regarding pride dynamics and demographics would not have been possible without the dedicated efforts of the lion behavioral monitoring team comprised of Darius Scholtz, Nkabeng Maruping, Trevor Valoyi, 92 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Travis Smit, and Katlego Makutulela. Coupled with this information, virus as a tool to develop intervention strategies against human immunodeficiency virus type 1. AIDS Res. Hum. Retrovir. 14 (9), 797–801. laboratory efforts for this study were widely supported by the partici­ Elder, J., Sundstrom, M., de Rozieres, S., de Parseval, A., Grant, C., Lin, Y., 2008. pation of several graduate and undergraduate students including Tanya Molecular mechanisms of FIV infection. Vet. Immunol. Immunopathol. 123, 3–13. Kerr, Zaya McSky, Emily Magnuson, Kadie Anderson, Elena Alonso, and Fan, Y., Chen, J., Shirkey, G., John, R., Wu, S., Park, H., Shao, C., 2016. Applications of Ignatius Viljoen, as well as Dr. Patrick Chappell at Oregon State Uni­ structural equation modeling (SEM) in ecological studies: an updated review. Ecol. Process 5 (19). https://doi.org/10.1186/s13717-016-0063-3. versity. This work was supported by Oregon State University, SAN­ Ferreira, S., Funston, P., 2010. Estimating lion population variables: prey and disease PARKs Scientific Services, Ajubatus, and Morris Animal Foundation effects in Kruger National Park, South Africa. Wildl. Res. 37, 194–206. [Grant #D13ZO-420], as well as the Spanish Ministerio de Economia y Foreyt, W., 2001. Veterinary Parasitology: Reference Manual, fifth ed. Iowa State University Press, Ames, Iowa. Competitividad (MINECO) through a Ramon y Cajal agreement [RYC- Freeman, L., Mansfield, K., Goldin, B., Woods, M., Gualtieri, L., Li, W., Bussell, S., 2016-21120] and private donors through Morris Animal Foundation’s Lackner, A., Gorbach, S., 2004. Body-composition changes in the simian research sponsor program. Sponsors of this research did not participate immunodeficiency virus-infected juvenile rhesus macaque. J. Infect. Dis. 189 (11), 2010–2015. in the design, data collection, analysis, interpretation, writing, or deci­ Fromont, E., Courchamp, F., Artois, M., 1997. Infection strategies of retroviruses and sion for submission of this work. social grouping of domestic cats. Can. J. Zool. 75, 1994–2002. Fromont, E., Pontier, D., Langlais, M., 2003. Disease propagation in connected host populations with density-dependent dynamics: the case of the Feline Leukemia Appendix A. Supplementary data Virus. J. Theor. Biol. 223 (4), 465–475. Hartmann, K., 2011. Clinical aspects of feline immunodeficiency and feline leukemia Supplementary data to this article can be found online at https://doi. virus infection. Vet. Immunol. Immunopathol. 143 (3–4), 190–201. Hartmann, K., 2012. Clinical aspects of feline retroviruses: a review. Viruses 4, org/10.1016/j.ijppaw.2021.07.003. 2684–2710. Hawley, D., Etienne, R., Ezenwa, V., Jolles, A., 2011. Does animal behavior underlie References covariation between hosts’ exposure to infectious agents and susceptibility to infection? Implications for disease dynamics. iCB 528–539, 2001. Hofmann-Lehmann, R., Holznagel, E., Ossent, P., Lutz, H., 1997. Parameters of disease Aleuy, O.A., Serrano, E., Ruckstuhl, K.E., Hoberg, E.P., Kutz, S., 2020. Parasite intensity progression in long-term experimental feline retrovirus (feline immunodeficiency drives fetal development and sex allocation in a wild ungulate. Sci. Rep. 10, 15626. virus and feline leukemia virus) infections: hematology, clinical chemistry, and Ackley, C., Yamamoto, J., Levy, N., Pedersen, N., Cooper, M., 1990. Immunologic lymphocyte subsets. Clin. Diagn. Lab. Immunol. 4 (1), 33–42. abnormalities in pathogen-free cats experimentally infected with feline Ida, T., Miyazato, M., Naganobu, K., Nakahara, K., Sato, M., Lin, X., Kaiya, H., Doi, K., immunodeficiency virus. J. Virol. 64 (11), 5652–5655. Noda, S., Kubo, A., Murakami, N., Kangawa, K., 2007. Purification and Antunes, A., Troyer, J., Roelke, M., Pecon-Slattery, J., Packer, C., Winterbach, C., characterization of feline ghrelin and its possible role. Domest. Anim. Endocrinol. 32 Winterbach, H., Hemson, G., Frank, L., Stander, P., Siefert, L., Driciru, M., (2), 93–105. Funston, P., Johnson, W., 2008. The evolutionary dynamics of the lion Panthera leo Ivan, E., Crowther, N., Mutimura, E., Osuwat, L., Janssen, S., Grobusch, M., 2013. revealed by host and viral population genomics. PLoS Genet. 4 (11), e1000251. Helminthic infections rates and malaria in HIV-infected pregnant women on anti- Appleton, D., Rand, J., Sunvold, G., 2000. Plasma leptin concentrations in cats: reference retroviral therapy in Rwanda. PLoS Neglected Trop. Dis. 7 (8), e2380. range, effect of weight gain and relationship with adiposity as measured by dual Jacquier, M., Aarhaug, P., Arnemo, J., Bauer, H., Enriquez, B., 2006. Reversible energy X-ray absorptiometry. J. Feline Med. Surg. 2, 191–199. immobilization of free-ranging african lions (Panthera leo) with medetomidine- Bartoń, K., 2017 [cited 2016 Jan 10]. MuMIn: Multi-model inference. Version 1.40.0 tiletamine-zolazepam and atipamezole. J. Wildl. Dis. Wildl. Dis. Assoc. 42 (2), [Internet]. Available at: http://CRANR-project.org/package=MuMIn. 432–436. Bates, D., Maechler, M., Bolker, B., Walker, S., Christensen, R., 2014. cited 2015 Aug 15]. Jensen, K., Forcada, Y., Church, D., Niessen, S., 2015. Evaluation and diagnostic lme4: Linear mixed-effects models using Eigen and S4. 1.1-7 [Internet]. Available potential of serum ghrelin in feline hypersomatotropism and diabetes mellitus. from: https://CRAN.R-project.org/package=lme4. J. Vet. Intern. Med. 29, 14–20. Beechler, B., Broughton, H., Bell, A., Ezenwa, V., Jolles, A., 2012. Innate immunity in Kanzaki, L., Looney, D., 2004. Feline immunodeficiency virus: a concise review. Front. free-ranging african buffalo (syncerus caffer): associations with parasite infection Biosci. 9, 370–377. and white blood cell counts. Physiol. Biochem. Zool. 8500 (3), 255–264. Kelly, P., Marabini, L., Dutlow, K., Zhang, J., Loftis, A., Wang, C., 2014. Molecular Bendinelli, M., Pistello, M., Lombardi, S., Poli, A., Garzelli, C., Matteucci, D., Ceccherini- detection of tick-borne pathogens in captive wild felids, Zimbabwe. Parasites Vectors Nelli, L., Malvaldi, G., Tozzini, F., 1995. Feline immunodeficiency virus: an 7 (1), 514. interesting model for AIDS studies and an important cat pathogen. Clin. Microbiol. Kennedy-Stoskopf, S., Levine, J., Tompkins, W., Loomis, M., Gebhard, D., Bull, M., 2003. Rev. 8 (1), 87–112. Evaluation of T lymphocytes in captive African lions (Panthera leo) infected with Bentwich, Z., Kalinkovich, A., Weisman, Z., Borkow, G., Beyers, N., Beyers, A., 1999. Can feline immunodeficiency virus. Am. J. Vet. Res. 64 (10), 1293–1300. eradication of helminthic infections change the face of AIDS and tuberculosis? Kerr, T., Matthee, S., Govender, D., Tromp, G., Engelbrecht, S., Matthee, C., 2018. Immunol. Today 20 (11), 485–487. Viruses as indicators of contemporary host dispersal and phylogeography: an Bentwich, Z., Maartens, G., Torten, D., Lal, A., Lal, R., 2000. Concurrent infections and example of feline immunodeficiency virus (FIVple) in free-ranging African lion HIV pathogenesis. AIDS 14, 2071-2018. (Panthera leo). J. Evol. Biol. 21, 1529–1543. Biek, R., Rodrigo, A., Holley, D., Drummond, A., Anderson, C., Ross, H., Poss, M., 2003. Kleiber, C., Zeileis, A., 2008 [cited 2016 Jan 10]. Applied Econometrics with R Epidemiology, genetic diversity, and evolution of endemic feline immunodeficiency [Internet]. Available from: http://eeecon.uibk.ac.at/~zeileis/. virus in a population of wild cougars. J. Virol. 77 (17), 9578–9589. Klok, M., Jakobsdottir, S., Drent, M., 2007. The role of leptin and ghrelin in the Bolker, B., Brooks, M., Clark, C., Geange, S., Poulsen, J., Stevens, M., White, J., 2009. regulation of food intake and body weight in humans: a review. Obes. Rev. 8 (1), Generalized linear mixed models: a practical guide for ecology and evolution. Trends 21–34. Ecol. Evol. 24 (3), 127–175. Kotler, D., Gaetz, H., Lange, M., Klein, E., Holt, P., 1984. Enteropathy associated with the Bosman, A., Oosthuizen, M., Peirce, M., Venter, E., Penzhorn, B., 2010. Babesia lengau acquired immunodeficiency syndrome. Ann. Intern. Med. 101 (4), 421–428. sp. nov., a novel Babesia species in Cheetah (Acinonyx jubatus, Schreber, 1775) Lesnoff, M., Lancelot, R., 2012 [cited 2015 Aug 21]. aod: Analysis of Overdispersed Data populations in South Africa. J. Clin. Microbiol. 48 (8), 2703–2708. [Internet]. Available from: http://cran.rproject. org/package=aod. Broughton, H., Govender, D., Shikwambana, P., Chappell, P., Jolles, A., 2017. Bridging Levinson, W., 2012. Review of Medical Microbiology and Immunology. McGraw Hill, gaps between zoo and wildlife medicine: establishing reference intervals for free- New York. ranging african lions (Panthera leo). J. Zoo Wildl. Med. 48 (2), 298–311. Lloyd-smith, J., Schreiber, S., Kopp, P., Getz, M., 2005. Superspreading and the effect of Brown, P., Hopper, C., Harbour, D., 1991. Pathological features of lymphoid tissues in individual variation on disease emergence. Nature 438, 355–359. cats with natural feline immunodeficiency virus infection. J. Comp. Pathol. 104 (4), Martin, L., Siliart, B., Lutz, T., Biourge, V., Nguyen, P., Dumon, H., 2010. Postprandial 345–355. response of plasma insulin, amylin and acylated ghrelin to various test meals in lean Brown, E., Yuhki, N., Packer, C., O ’Brien, S., 1994. A lion lentivirus related to feline and obese cats. Br. J. Nutr. 103, 1610–1619. immunodeficiency virus: epidemiologic and phylogenetic aspects. J. Virol. 68, Matson, K., Tieleman, B., Klasing, K., 2006. Capture stress and the bactericidal 5953–5968. competence of blood and plasma in five species of tropical birds. Physiol. Biochem. Costiniuk, C., Angel, J., 2012. Human immunodeficiency virus and the gastrointestinal Zool. 79 (3), 556–564. immune system: does highly active antiretroviral therapy restore gut immunity. Mccaig, C., Begon, M., Norman, R., Shankland, C., 2011. A symbolic investigation of Mucosal Immunol. 5 (6), 596–604. superspreaders. Bull. Math. Biol. 73, 777–794. Dean, G., Bernales, J., Pedersen, N., 1998. Effect of feline immunodeficiency virus on McSorley, H., Maizels, R., 2012. Helminth infections and host immune regulation. Clin. cytokine response to Listeria monocytogenes in vivo. Vet. Immunol. Immunopathol. Microbiol. Rev. 25 (4), 585–608. 65 (2–4), 125–138. Munson, L., Terio, K., Kock, R., Mlengeya, T., Roelke, M., Dubovi, E., Summers, B., Depauw, S., Hesta, M., Whitehouse-Tedd, K., Stagegaard, J., Buyse, J., Janssens, G., Sinclair, A., Packer, C., 2008. Climate extremes promote fatal co-infections during 2012. Blood values of adult captive cheetahs (Acinonyx jubatus) fed either canine distemper epidemics in African lions. PloS One 3 (6), e2545. supplemented beef or whole rabbit carcasses. Zoo Biol. 31, 629–641. Murphy, B., Hillman, C., McDonnel, S., 2014. Peripheral immunophenotype and viral Elder, J., Dean, G., Hoover, E., Hoxie, J., Malim, M., Mathes, L., Neil, J., North, T., promoter variants during the asymptomatic phase of feline immunodeficiency virus Sparger, E., Tompkins, M., Tompkins, W., Yamamoto, J., Yuhki, N., Pedersen, N., infection. Virus Res. 179, 34–43. Miller, R., 1998. Workshop summary: lessons from the cat: feline immunodeficiency 93 H. Broughton et al. International Journal for Parasitology: Parasites and Wildlife 16 (2021) 83–94 Ndakotsu, M., Salawu, L., Durosinmi, M., 2008. Relation between erythrocyte Tejerizo, G., Doménech, A., Illera, J., Silván, G., Gómez-Lucía, E., 2012. Altered plasma sedimentation rate, clinical and immune status in HIV-infected patients. Niger. J. concentrations of sex hormones in cats infected by feline immunodeficiency virus or Med. 18 (2), 208–210. feline leukemia virus. Domest. Anim. Endocrinol. 42 (2), 113–120. Olivier, M., Badaró, R., Medrano, F., Moreno, J., 2003. The pathogenesis of Leishmania/ Tenenhaus, M., Amato, S., Vinzi, E., 2004. A global goodness-of-fit index for PLS HIV co-infection: cellular and immunological mechanisms. Ann. Trop. Med. structural equation modelling. XLII SIS Sci Meet. Parasitol. 97 (1), 79–98. Tenenhaus, M., Esposito-Vinzi, V., Chatelin, Y., Lauro, C., 2005. PLS path modeling. Pecon-Slattery, J., Troyer, J., Johnson, W., O’Brien, S., 2008. Evolution of feline Comput. Stat. Data Anal. 48 (1), 159–205. immunodeficiency virus in Felidae: implications for human health and wildlife Thrall, M., 2004. Veterinary Hematology and Clinical Chemistry. Lippincott, Williams, ecology. Vet. Immunol. Immunopathol. 123, 32–44. and Wilkins, Philadelphia, Pennsylvania. Pedersen, N., Ho, E., Brown, M., Yamamoto, J., 1987. Isolation of a T-lymphotropic virus Tieleman, I., Williams, J., Ricklefs, R., Klasing, K., 2005. Constitutive innate immunity is from domestic cats with an immunodeficiency-like syndrome. Science 235 (4790), a component of the pace-of-life syndrome in tropical birds. Proc. R. Soc. B Biol. Sci. 790–793. 272 (1573), 1715–1720. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., 2015. Team TRC [cited 2015 Aug 21]. Troyer, J., Pecon-Slattery, J., Roelke, M., Johnson, W., VandeWoude, S., Vazquez- nlme: Linear and Nonlinear Mixed Effects Models [Internet]. Available from: http:// Salat, N., Brown, M., Frank, L., Woodroffe, R., Winterback, C., Winterbach, H., CRAN.R-project.org/package=nlme. Hemson, G., Bush, M., Alexander, K., Revilla, A., O’Brien, S., 2005. Seroprevalence Podell, M., Maruyama, K., Smith, M., Hayes, K., Buck, W., Ruehlmann, D., Mathes, L., and genomic divergence of circulating strains of feline immunodeficiency virus 1999. Frontal lobe neuronal injury correlates to altered function in FIV-infected cats. among felidae and hyaenidae species. J. Virol. 79, 8282–8294. J. Acquir. Immune Defic. Syndr. 22, 10–18. Troyer, J., Roelke, M., Jespersen, J., Baggett, N., Buckley-Beason, V., MacNulty, D., Reggeti, F., Ackerley, C., Bienzle, D., 2008. CD134 and CXCR4 expression corresponds to Craft, M., Packer, C., Pecon-Slattery, J., O’Brien, S., 2011. FIV diversity: FIVPle feline immunodeficiency virus infection of lymphocytes, macrophages and dendritic subtype composition may influence disease outcome in African lions. Vet. Immunol. cells. J. Gen. Virol. 89, 277–287. Immunopathol. 143, 338–346. R Foundation for Statistical Computing, 2021. R: A Language and Environment for VandeWoude, S., Apetrei, C., 2006. Going wild: lessons from naturally occurring T- Statistical Computing 4.1.0. lymphotropic lentiviruses. Clin. Microbiol. Rev. 19, 728–762. Reyers, F., Leisewitz, A., Lobetti, R., Milner, R., Jacobson, L., Van Zyl, M., 1998. Canine van Vureen, M., Stylianides, E., Kania, S., Hardy, W., 2003. Evaluation of an indirect babesiosis in South Africa: more than one disease. Does this serve as a model for enzyme-linked immunosorbent assay for the detection of feline lentivirus-reactive falciparum malaria? Ann. Trop. Med. Parasitol. 92 (4), 503–511. antibodies in wild felids, employing a puma lentivirus-derived synthetic peptide Roelke, M., Pecon-Slattery, J., Taylor, S., Citino, S., Brown, E., Packer, C., antige. Onderstepoort J. Vet. Res. 70 (1), 1–6. Vandewoude, S., O’Brien, S., 2006. T-lymphocyte profiles in FIV-infected wild lions Venables, W., Ripley, B., 2002. Modern Applied Statistics with S, fourth ed. Springer, and pumas reveal CD4 depletion. J. Wildl. Dis. 42, 234–248. New York. Roelke, M., Brown, M., Troyer, J., Winterbach, H., Winterbach, C., Hemson, G., Vinzi, V., Trinchera, L., Amato, S., 2010. PLS path modeling: from foundations to recent Smith, D., Johnson, R., Pecon-Slattery, J., Roca, A., Alexander, K., Klein, L., developments and open issues for model assessment and improvement. In: Vinzi, V., Martelli, P., Krishnasamy, K., O’Brien, S., 2009. Pathological manifestations of feline Henseler, J., Chin, W., Wang, H. (Eds.), Handbook of Partial Least Squares. Springer, immunodeficiency virus (FIV) infection in wild African lions. Virology 390, 1–12. London and New York. Sanchez, G., Trinchera, L., Russolillo, G., 2017 [cited 2017 May 20]. plspm: Tools for Walson, J., Herrin, B., John-Stewart, G., 2009. Deworming helminth co-infected Partial Least Squares Path Modeling (PLS-PM) [Internet]. Available from: https:// individuals for delaying HIV disease progression. Cochrane Database Syst. Rev., cran.r-project.org/package=plspm. CD006419, 2009. Serrano, E., López-Soria, S., Trinchera, L., Segalés, J., 2014. The use of null models and Wenger, S., Buss, P., Joubert, J., Steenkamp, J., Shikwambana, P., Hatt, J., 2010. partial least squares approach path modelling (PLS-PM) for investigating risk factors Evaluation of butorphanol, medetomidine and midazolam as a reversible narcotic influencing post-weaning mortality in indoor pig farms. Epidemiol. Infect. 142 (3), combination in free-ranging African lions (Panthera leo). Vet. Anaesth. Analg. 37 530–539. (6), 491–500. Schaller, G., 1976. The Serengeti Lion: A Study of Predator-Prey Relations. Univ Chicago, Willett, B., McMonagle, E., Ridha, S., Hosie, M., 2006. Differential utilization of CD134 Chicago, Illinois. as a functional receptor by diverse strains of feline immunodeficiency virus. J. Virol. Sher, A., Gazzinelli, R., Oswald, I., Clerici, M., Kullberg, M., Pearce, E., Berzofsky, J., 80 (7), 3386–3394. Mosmann, T., James, S., Morse, H., 1992. Role of T-cell derived cytokines in the Witzel, A., Kirk, C., Kania, S., Bartges, J., Boston, R., Moyers, T., Byrd, H., Lauten, S., downregulation of immune responses in parasitic and retroviral infection. Immunol. 2015. Relationship of adiponectin and its multimers to metabolic indices in cats Rev. 127 (1), 183–204. during weight change. Domest. Anim. Endocrinol. 53, 70–77. Shibata, H., Sasaki, N., Honjoh, T., Ohishi, I., Takiguchi, M., Ishioka, K., Ahmed, M., Wold, H., 1985. Partial least squares. In: Kotz, S., Johnson, N. (Eds.), Encyclopedia of Soliman, M., Kimura, K., Saito, M., 2003. Feline leptin: immunogenic and biological Statistical Sciences. Wiley, New York. activities of the recombinant protein, and its measurement by ELISA. J. Vet. Med. Yamamoto, J., Hansen, H., Ho, E., Morishita, T., Okuda, T., Sawa, T., Nakamura, R., Sci. 65 (11), 1207–1211. Pedersen, N., 1989. Epidemiologic and clinical aspects of feline immunodeficiency Smuts, G., Anderson, J., Austin, J., 1978. Age determination of the African lion (Panthera virus infection in cats from the continental United States and Canada and possible leo). J. Zool. 185, 115–146. mode of transmission. J. Am. Vet. Med. Assoc. 194, 213e220. Solano-Gallego, L., Sainz, Á., Roura, X., Estrada-Peña, A., Miró, G., 2016. A review of Yamamoto, J., Pu, R., Sato, E., Hohdatsu, T., 2007. Feline immunodeficiency virus canine babesiosis: the European perspective. Parasites Vectors 9, 336. pathogenesis and development of a dual-subtype feline- immunodeficiency-virus Stockham, S., Scott, M., 2008. Fundamentals of Veterinary Clinical Pathology. Blackwell vaccine. AIDS 21, 547–563. Publishing, Iowa. Yamamoto, J., Sanou, M., Abbott, J., Coleman, J., 2010. Feline immunodeficiency virus model for designing HIV/AIDS vaccines. Curr. HIV Res. 8 (1), 14–25. 94