key: cord-102226-aioqogaw
authors: Chiu, Elliott S.; VandeWoude, Sue
title: Presence of endogenous viral elements negatively correlates with FeLV susceptibility in puma and domestic cat cells
date: 2020-06-24
journal: bioRxiv
DOI: 10.1101/2020.06.23.168351
sha: 
doc_id: 102226
cord_uid: aioqogaw

While feline leukemia virus (FeLV) has been shown to infect felid species other than the endemic domestic cat host, differences in FeLV susceptibility among species has not been evaluated. Previous reports have noted a negative correlation between enFeLV copy number and exogenous FeLV infection outcomes in domestic cats. Since felids outside the genus Felis do not harbor enFeLV genomes, we hypothesized absence of enFeLV results in more severe disease consequences in felid species lacking these genomic elements. We infected primary fibroblasts isolated from domestic cats (Felis catus) and pumas (Puma concolor) with FeLV and quantitated proviral and viral antigen loads. Domestic cat enFeLV env and LTR copy numbers were determined for each individual and compared to FeLV viral outcomes. FeLV proviral and antigen levels were also measured in 6 naturally infected domestic cats and 11 naturally infected Florida panthers (P. concolor coryi). We demonstrated that puma fibroblasts are more permissive to FeLV than domestic cat cells, and domestic cat FeLV restriction was highly related to enFeLV LTR copy number. Terminal tissues from FeLV-infected Florida panthers and domestic cats had similar exFeLV proviral copy numbers, but Florida panther tissues have higher FeLV antigen loads. Our work indicates enFeLV LTR elements negatively regulate exogenous FeLV replication. Further, Puma concolor lacking enFeLV are more permissive to FeLV infection than domestic cats, suggesting endogenization can play a beneficial role in mitigating exogenous retroviral infections. Conversely, presence of endogenous retroelements may relate to new host susceptibility during viral spillover events. Importance Feline leukemia virus (FeLV) can infect a variety of felid species. Only the primary domestic cat host and related small cat species harbor a related endogenous virus in their genomes. Previous studies noted a negative association between the endogenous virus copy number and exogenous virus infection in domestic cats. This report shows that puma cells, which lack endogenous FeLV, produce more virus more rapidly than domestic cat fibroblasts following cell culture challenge. We document a strong association between domestic cat cell susceptibility and FeLV long terminal repeat (LTR) copy number, similar to observations in natural FeLV infections. Viral replication does not, however, correlate with FeLV env copy number, suggesting this effect is specific to FeLV LTR elements. This discovery indicates a protective capacity of the endogenous virus against the exogenous form, either via direct interference or indirectly via gene regulation, and may suggest evolutionary outcomes of retroviral endogenization.

to FeLV viral outcomes. FeLV proviral and antigen levels were also measured in 6 naturally infected 23 domestic cats and 11 naturally infected Florida panthers (P. concolor coryi). We demonstrated that 24 puma fibroblasts are more permissive to FeLV than domestic cat cells, and domestic cat FeLV 25 restriction was highly related to enFeLV LTR copy number. Terminal tissues from FeLV-infected 26

Florida panthers and domestic cats had similar exFeLV proviral copy numbers, but Florida panther 27 tissues have higher FeLV antigen loads. Our work indicates enFeLV LTR elements negatively 28 regulate exogenous FeLV replication. Further, Puma concolor lacking enFeLV are more permissive to 29

FeLV infection than domestic cats, suggesting endogenization can play a beneficial role in mitigating 30 exogenous retroviral infections. Conversely, presence of endogenous retroelements may relate to 31 new host susceptibility during viral spillover events. 32 33 Importance 34

Feline leukemia virus (FeLV) can infect a variety of felid species. Only the primary domestic cat host 35 and related small cat species harbor a related endogenous virus in their genomes. Previous studies 36 noted a negative association between the endogenous virus copy number and exogenous virus 37 infection in domestic cats. This report shows that puma cells, which lack endogenous FeLV, produce 38 more virus more rapidly than domestic cat fibroblasts following cell culture challenge. We document a 39

The vast majority of vertebrate genomes, including up to 8% of the human genome, harbor 48 fossils of ancient viral infections made up predominantly of retroviral genetic material (1-3). During 49 infection, the retroviral RNA genome is reverse-transcribed to form double-stranded DNA, which is in 50 turn integrated into the host's genome (4). While most of these infections target somatic cells, these 51 viruses are capable of infecting and integrating into germ cells (5). The consequences of viral 52 integration into the germline is profound, and ultimately the virus is vertically transmitted as 53 permanent genetic elements inherited in a Mendelian fashion (6). Fixation of the retroviral content in 54 host genomes is a process termed endogenization and leads to new host genetic elements called 55 endogenous retroviruses (ERVs) (7). ERVs in their early stages are believed to undergo massive 56 changes during host cell transcription, and the foreign, potentially deleterious genetic material 57 accumulates mutations and deletions that often render the newly endogenized virus defunct (8). 58

While typically unable to produce infectious virions, many ERVs are still capable of undergoing 59 transcription and may even produce functional viral proteins. Certain ERVs are known to function in 60 important physiologic, cellular, or biological processes, including placentation, oncogenesis, immune 61 modulation, and infectious disease progression (9, 10). 62

Endogenous feline leukemia virus (enFeLV) is an example of an ERV which has a horizontally 63 transmitted retroviral counterpart (feline leukemia virus, FeLV). Only members of the Felis genus 64 harbor enFeLV as endogenization is believed to have originated after the Felis genus split off from other members of the Felidae family (11, 12) . FeLV can infect felid species that harbor enFeLV (i.e., 66 domestic cats) as well as species that lack enFeLV (i.e., puma). FeLV epizootics have been 67 documented in multiple non-Felis species populations including the North American puma (Puma 68 concolor) (13) (14) (15) (16) (17) (18) (19) . 69

FeLV represents an endogenous-exogenous retroviral system that has perhaps been best 70 studied with regard to disease biology and outcome during naturally occurring infections in an 71 outbred, highly dispersed mammalian host. Thus, evaluation of this system provides opportunities to 72 better understand ERV-exogenous viral interactions that are highly relevant to virus and host 73 evolution and ecology. FeLV epizootics in wild felids are characterized by serious disease of epidemic 74

proportions, (20, 21), whereas FeLV infection in adult cats frequently results in regressive and 75 abortive infections (22). It has been hypothesized that enFeLV may be associated with differences in 76 infection outcome. We previously demonstrated enFeLV long terminal repeat (LTR) copy number was 77 associated with better infection outcomes during a natural FeLV outbreak in a multi-cat household 78 (23). Here, we evaluate FeLV infection of puma (P. concolor) and domestic cat cells in vitro and in 79 situ to examine the susceptibility of endemic and novel hosts to FeLV infection with respect to 80 enFeLV to provide further evaluation of this relationship. 81 82

Primary fibroblasts were successfully propagated from ear punches from three free-ranging 84 puma (two kittens of unknown sex and one adult male) and 1 abdominal skin incision from a mature 85 adult female captive puma. Primary fibroblasts were cultured from 7 domestic cats abdominal full skin 86 biopsies from necropsied cats at Colorado State University (6 male, 1 female). 87

Exogenous FeLV proviral load was significantly and substantially greater in puma fibroblasts 89 Domestic cat and puma viable cell counts were equivalent on days 5 or 10, averaging 00 1.28x10 5 cells per 2 cm 2 for domestic cat cells and 1.92x10 5 cells per 2 cm 2 for puma cells 01 (Supplemental figure 1A) . CrFKs are smaller than domestic cat and puma primary fibroblasts and 02 therefore cell density was higher in these cultures (3.54 x10 5 cells per 2 cm 2 at day 5 and 4.13 x10 5 03 cells per 2 cm 2 at day 10). Percent dead cells (as measured by trypan blue exclusion) on days 5 and 04 10 ranged between 3.94±0.70% in for domestic cat cells and 4.73%±1.45% for puma cells regardless 05 of infection status (Supplemental figure 1B ). There was a consistent trend for lower percent mortality 06 at day 5 when cells first reached confluency. CrFKs experienced greater cell mortality over the course 07 of the infection regardless of infection status (2.59% at day 5 and 14.26% at day 10). 08

As anticipated, domestic cat cells harbored more enFeLV LTRs than enFeLV env genes 10 ( Figure 2A ). Normalized LTR sequence copy numbers in domestic cat cells ranged from 32 and 74 11 copies per cell with an average of 57 copies per cell. Copy numbers for enFeLV env were significantly 12 lower, ranging from 9 to 13 copies per cell with an average of 11 copies. 13

Variation in domestic cat fibroblasts enFeLV-LTR copy number correlated to FeLV antigen 14 loads (day 7, Pearson's correlation coefficient=-0.894; p<0.05; Figure 2B ), whereas variation in 15 enFeLV-env did not (day 7, Pearson's correlation coefficient=-0.107, p=0.840; Figure 2C ). enFeLV-16 exFeLV correlations were calculated at day 7 since this is the timepoint that cells reached complete confluency and the rate at which antigen was produced waned after this timepoint. Linear regression 18 analysis of FeLV proviral load against antigen load showed that only 44% of the variation in antigen 19 production could be explained by proviral load (R 2 = 0.438; Figure 3 ). 20 21

FeLV loads in bone marrow, spleen, thymus and peripheral lymph nodes were assessed in 6 23 experimentally infected cats and 11 naturally infected pumas (not all tissues were available for each 24 animal, see Table 1 ). Both tissue proviral load and tissue antigen load failed the Kolmogorov-Smirnov 25 test for log-normality, indicating a non-normal distribution. FeLV proviral load was greater in domestic 26 cats compared to panther by a median difference of 1.07 by Mann-Whitney test of log-transformed 27 copy numbers per cell (U=122, p=0.0001; Figure 4A ). Despite lower mean proviral load in Florida 28 panther tissues, p27 capsid antigen in tissues tended to be higher (median value 3.41 vs. 2.87, 29

Mann-Whitney test, U=250, p=0.127; Figure 4B ). There was no difference in antigen or proviral load 30 among tissues, with the exception of bone marrow antigen. 31 32

Alterations in infectivity and virulence has been noted following cross species viral 34 transmission (24). In some cases, disease spillovers into novel species of the same family can result 35 in dead-end hosts for the virus (Infectious hematopoietic necrosis virus (25); Feline immunodeficiency 36 virus strain lru (26)). In other cases, disease spillover may result in active infections that maintain 37 persistent transmission (Mycoplasmosis; (27); Feline foamy virus (28)). Further still, some cases 38 result in adaptation of the virus in novel hosts leading to increased morbidity and mortality (HIV (29); 39

Covid-19 (30)). Outcomes of diseases associated with spillover are dependent on the given specifics 40 of host, environment and agent interactions (31). Multiple FeLV spillover events have been 41 documented in free-ranging pumas with resultant significant morbidity (13, 32). The apparent 42 virulence of FeLV in pumas and other nondomestic felids has led to speculation that this virus may have enhanced virulence in novel hosts (13, 14) . This study was thus undertaken to evaluate the 44 hypothesis that FeLV infection of nondomestic felids might be more competent in virus replication 45 than the domestic cat reservoir host, examined specifically in the context of the presence of enFeLV. 46

Experimental infections were conducted in vitro to establish differences in FeLV replication in puma 48 and domestic cat cells in the absence of immunological and physiological parameters. In fibroblast 49

cultures, FeLV infection resulted in higher proviral load in puma cells than in domestic cat cells 50 ( Figure 1A ). Additionally, increased viral antigen was documented in infected puma cells, which is 51 suggestive of increased viral production ( Figure 1B) . Therefore, at a cellular level, puma cells appear 52 to be more competent at supporting FeLV infection and replication than fibroblasts of the primary host 53 domestic cats. This is further supported by the fact that proviral load could only explain 43% of the 54 variation in viral antigen production, indicating that proviral integration events alone are not a 55 surrogate for virus replication. Other factors may be influencing the increased viral production in 56 puma cells or restriction of viral replication in domestic cats 57 enFeLV-LTR copy number is associated with resistance to FeLV infection and antigen 58 production. Endogenous elements constitute a sizable component of an animal's genome (2) and 59 solo LTRs vastly outnumber full endogenous genomes/pseudogenomes (7). This occurs because two 60 flanking LTRs allow for the intervening genes to be removed by homologous recombination, leaving 61 behind just one copy of LTR it its place (33). As such, env copy number serves as a proxy for full 62 enFeLV genomes, since loss of one gene would not occur frequently. In this sampling of domestic cat 63 fibroblast enFeLV components, solo LTRs range from 32-74 copies per cell, while env ranges from 9-64 13 copies per cell, similar to previous observations ( Figure 2A ). This is consistent with previously 65 reported measures of full-length enFeLV range of 6-26 copies per cell (23, 34, 35), versus 19-58 66 copies of enFeLV LTR per cell (23). LTR copy number variation, but not full enFeLV genomes, 67 correlated with FeLV replication as evidenced by directly proportional antigen load; enFeLV-env 68 showed no correlation to either FeLV replication ( Figure 2B -C). While FeLV replication may intuitively seem to be correlated to proviral integration number, only 44% of FeLV antigen variation is explained 70 by proviral load (Figure 3 ). Linear regression analysis therefore suggests that factors other than 71 proviral integration number contributes to host susceptibility. This also corroborates observations 72 made in vivo in a domestic cat colony naturally infected with FeLV (23) Alternatively, it is possible that the enFeLV genetic elements may directly interfere with 96 exFeLV infection by encoding for small interfering RNAs or PIWI interacting RNAs that activate host 97 DICER complexes to specifically target FeLV transcripts (38, 39). Our results are more suggestive of 98 direct interference mechanisms due to the linearity of the FeLV restriction afforded by enFeLV-LTRs. 99

It is unlikely that all LTR integration sites are influencing the transcription of host genetic factors and 00

would have an effect in a dose-dependent manner. 01

FeLV reaches high viral load in lymphoid tissues during natural infections. Naturally infected 02 pumas with FeLV had lower proviral load than domestic cats, with the exception of two bone marrow 03 samples, that achieved 2x10 7 proviruses per 1x10 6 cells ( Figure 4A ). Interestingly, a much wider 04 range of proviral copy numbers were noted in pumas, and viral antigen loads in pumas equaled or 05 exceeded that of domestic cats ( Figure 4B ). Field collections were performed opportunistically on 06

Florida panthers when animals were either found deceased or hit by vehicle, often hours to days after 07 death occurred. In contrast, FeLV positive shelter cats were euthanized prior to death, and tissues 08 were collected rapidly following death. The timeliness of collection likely impacted quality of the 09 sample prior to DNA extraction for puma samples. While normalization to feline CCR5 helped to 10 address these issues for proviral copy number calculations, it is possible that viral antigen loads 11 measured in puma tissues underestimate actual values. 12

Biological aspects of FeLV transmission that differ between domestic cats and pumas may 13 impact subsequent infection kinetics. The initial spillover events of FeLV to pumas has been 14 associated with predation of domestic cats (13, 40), whereas FeLV in domestic cats is believed to be 15 transmitted in households through social interactions such as grooming, or via antagonstic 16 interactions (11). Domestic cats interact socially, so behaviors like grooming may sustain infection in 17 animals in close contact and infections may result from repeated exposures. Unlike domestic cats, 18 pumas are much more solitary and interactions between pumas outside of mother-offspring groups 19

are primarily believed to be antagonistic (41). 20

Gammaretroviruses require the dissolution of the nucleus during mitosis in order to integrate into cells 22

(42), and therefore dividing cells are more susceptible to FeLV infection and replication. We 23 measured cell count as a proxy for rate of cell division, and percent cell mortality as a measure of 24 viability. Neither measure differed between domestic cat and puma fibroblasts, though CrFK cells 25 displayed greater cell count and cell mortality at confluency (Supplemental figure 1) . Immortalized cell 26 lines have accumulated multiple changes that fundamentally alter their morphology and physiologic 27 behaviors, including decreases in contact inhibition (43). FeLV proviral integration and antigen 28 production in CrFK infections had far less within-and between-run variation than primary fibroblasts, 29 likely attributable to the clonal nature of CrFK versus wild type derived primary tissue cultures. 30

Observation of cell culture parameters did not suggest differences in growth characteristics that 31 explain the variant FeLV susceptibly of puma and domestic cat fibroblasts. 32

In this report, we present information that demonstrates that enFeLV-LTR confers protection 33 against exFeLV infection in vitro through the limitation of FeLV replication. The exact mechanism by 34 which these constituents act has yet to be determined, but leaves room for further investigation in the 35

FeLV system as well as other endogenous-exogenous retroviral dyads. We hypothesize that FeLV 36 restriction may manifest as direct interference through RNA silencing mechanisms, or by indirect 37 enFeLV-LTR-mediated promotion of host anti-viral genes. FeLV provides an opportunity to directly 38 interrogate the mechanisms that govern related exogenous-endogenous retroviral interactions in an 39 outbred and diverse population. penicillin/streptomycin/fungizone). One puma culture was infected with feline foamy virus and was 49 treated prophylactically with the anti-retroviral drug, AZT (100 ug/ml; Sigma) per manufacturer's 50 direction until in cultures where FFV CPE were detected. Cells were passaged two times for 51 approximately 10 days in media without AZT washout prior to infection. Primary cultures were 52 expanded for at most four passages before being frozen in (20% DMSO, 10% FBS, 70% serum free 53 DMEM) using a freezing container (Nalgene) and stored at -80˚C. 54

Bone marrow, thymus, spleen, and lymph node from naturally FeLV-infected domestic cats 55 were washed with sterile PBS, were given fresh media, and were incubated with 5% CO 2 at 37˚C for 73 ten days. Titration was repeated three times. FeLV antigen ELISA detection, described below, was 74 used to detect viral capsid antigen p27 in the supernatant. The quantity of virus necessary to infect 75 50% of tissue cultures (TCID 50) was calculated by previously published methods (45). 76 enFeLV and exFeLV quantification by Real Time qPCR 77 LTR and env enFeLV copy number was quantified in domestic cat cells. env was used as a 78 proxy for full-length endogenous FeLV, and LTR copy number detected both full length enFeLV as 79 well as solo LTRs. Exogenous FeLV proviral DNA was measured by a third qPCR protocol targeting 80 exFeLV specific LTRs, which vary from enFeLV (22). enFeLV-env, enFELV-LTR and exFeLV-LTR 81 primers and probes were previously designed and reactions were performed as described (46) 

Biorad CFX96 thermocycler. In order to determine enFeLV and exFeLV proviral load, quantified FeLV 83 was normalized against feline CCR5 (C-C chemokine receptor type 5; (47)) recognizing both 84 domestic cat and puma CCR5 sequences. We used the delta cT method accounting for two CCR5 85 genes per cell (48). Custom DNA oligos were synthetically constructed with target regions of enFeLV 86 LTR and env, exFeLV, and CCR5 on one DNA construct for quantification (gBlock, IDT; 87

Primer and probe sequences and qPCR thermocycling conditions are reported in Table 2 ( (Bio-Rad), water, and DNA template. FeLV and CCR5 reactions were run simultaneously on the 97 same plate on a Bio-rad CFX96 at 95˚C for 3 minutes, followed by 40 cycles of 95˚C for 5 seconds 98 and 60˚C for 15 seconds. The limit of detection for this assay is ≥10 copies per reaction. Standards 99 for this assay were created as custom synthetic oligos (gBlocks, iDT) containing a relevant fragment 00 of the exogenous FeLV and CCR5 genes (Supplementary Figure 2) . Standard dilution and controls 01 were run in duplicate and samples were run in triplicate. 02

Primary puma and domestic cat fibroblast cultures passaged fewer than five times were 04 cultured in 20% FBS-supplemented DMEM high glucose media. Cells were plated at a density of 05 50,000 cells per 2 cm 2 in a 24-well plate and infected with a multiplicity of infection (MOI) of 0.01 06

FeLV-61E in triplicate and cultured with 1.2mL media. 120 μL of supernatant was collected and 07 stored at 80˚C at days 0, 1, 3, 5, 7, and 10 for detection of p27 ELISA. At days 5 and 10, cells were 08 harvested to determine cellular viability based on cell number and percent mortality by counting cells 09 stained with trypan blue (Gibco) on a hemocytometer. One domestic cat cell culture, one puma cell 10 culture, and one CrFK cell culture infection were terminated at day 7 due to equipment failure. One 11 puma primary cell culture triplicate infection was repeated twice. 12

FeLV capsid antigen p27 was measured by sandwich ELISA. Costar Immulon 2HB plates were 14 coated with 600ug CM1 capture antibody (Custom Monoclonal, Inc., US) in 100uL 0.1M Carbonate 15 buffer (7.5 g/L Sodium Bicarbonate, 2.0 g/L Sodium Carbonate, pH ~9.5) overnight at 4˚C. Plates 16 were blocked with 200uL 2% BSA in TEN buffer for two hours. One hundred μL of samples buffered 17 with 50 μL ELISA diluent were incubated for two hours on a plate shaker. Six hundred micrograms of 18 biotinylated secondary antibody (CM2-B; Custom Monoclonal, Inc., US) was incubated in each well, 19 followed by 1:4000 dilution of HRP-conjugated streptavidin (ThermoFisherScientific, MA). Each step 20 following sample incubation was followed with 5x wash with TEN buffer (0.05M TRIS Base, 0.001M 21 EDTA, 0.15M NaCl, pH 7.2-7.4) with 0.1% Tween. All incubations were performed at room 22 temperature. p27 antigen was detected indirectly following the addition of 3, 3', 5, 5' tetramethyl 23 benzidine (TMB) substrate and peroxidase (Biolegend, San Diego, CA) at room temperature for 7.5 min before adding 2.5 N H 2 SO 4 was quantified by Bioanalyzer at 450nm. Semi-purified FeLV p27 25 diluted in appropriate media (DMEM or RPMI) was used as a standard curve. Cutoff values for 26 negative samples were three times the standard error over the average OD measured for control 27 media samples. One puma, one domestic cat, and one control cat experiment were ended at day 7. Cutoff values 50 were established by 3x standard error above average value for negative control wells (red line). 51 52

variation between enFeLV-LTRs and env. In individual cats, LTR copy number was greater than env 54 copy number. enFeLV-LTR ranged between 9-13 copies per cell. enFeLV-env was more variable 55 ranging between 32-74 copies per cell. B) Domestic cat enFeLV-LTR copy number was 56 negatively correlated with FeLV antigen production (Pearson's coefficient= -0.8943; p= 0.0066). 

6:e1001191. 92 2. Consortium IHGS. 2001. Initial sequencing and analysis of the human genome

Endogenous retroviruses and the human germline. Current Opinion in 95

Evolution of retroviruses: Fossils in our DNA

Endogenous retroviruses of non-avian/mammalian 99 vertebrates illuminate diversity and deep history of retroviruses

Transmission, evolution, and 01 endogenization: lessons learned from recent retroviral invasions. Microbiology and Molecular 02

Retrotransposons, endogenous retroviruses, and the evolution of 04 retroelements

Degradation and remobilization of endogenous retroviruses by recombination during the 09 earliest stages of a germ-line invasion

Endogenous retroviral syncytin: compilation of experimental research on syncytin and its possible 12 role in normal and disturbed human placentogenesis

Expression and regulation of human endogenous retrovirus W elements

Feline leukaemia virus: half a century since its discovery

Evolutionary dynamics 18 of endogenous feline leukemia virus proliferation among species of the domestic cat lineage

Multiple 21 Introductions of Domestic Cat Feline Leukemia Virus in Endangered Florida Panthers

Feline 25 leukemia virus and other pathogens as important threats to the survival of the critically 26 endangered Iberian lynx (Lynx pardinus)

Prevalence of infectious diseases in feral cats in Northern Florida

Prevalence of FIV and FeLV infections in cats in Istanbul

Prevalence of feline leukaemia virus and antibodies to feline immunodeficiency 33 virus and feline coronavirus in stray cats sent to an RSPCA hospital

36 Prevalence of feline immunodeficiency virus and other retroviral infections in sick cats in Italy

Lymph node Spleen Thymus