HIGH PREVALENCE OF CYTAUXZOON FELIS IN BOBCATS (LYNX RUFUS) ACROSS OKLAHOMA AND OCCURRENCE IN WEST TEXAS, USA

Cytauxzoon felis is an emerging tick-borne apicomplexan parasite that causes cytauxzoonosis in domestic and wild felids. Severe disease may occur, characterized by depression, fever, lethargy, inappetence, dehydration, icterus, and possibly death (Wagner 1976; Kier-Schroeder 1979; Meinkoth and Kocan 2005). Cytauxzoonosis was first reported in four domestic cats (Felis catus) from rural southwestern Missouri, US, in 1976 (Wagner 1976). Bobcats (Lynx rufus) have since been recognized as the wild vertebrate reservoir of C. felis (Kier et al. 1982a, b). The prevalence of C. felis in wild bobcats is often high in areas where known tick vectors, Amblyomma americanum (lone star tick) and Dermacentor variabilis (American dog tick), are common (Supplementary Material Table S1). Cases of cytauxzoonosis in domestic cats have been reported from the southeastern, south-central, and mid-Atlantic US (Meinkoth and Kocan 2005; Sherrill and Cohn 2015; Reichard et al. 2021).

The complete life cycle of C. felis is not known; however, schizogony of C. felis occurs in feline macrophages and is the most pathogenic stage for felids (Kier-Schroeder 1979; Kier et al. 1987). Ticks ingest piroplasms of C. felis along with a blood meal, and sexual reproduction of the protozoan occurs in the tick definitive host. Though both domestic and wild felids can be persistently infected and serve as reservoirs for C. felis, initial studies by Kier et al. (1982a, b) identified the bobcat (L. rufus) as the natural wild animal reservoir in North America. When C. felis was experimentally infected into 30 different species of domestic farm, laboratory, and wildlife animals, only bobcats developed clinical signs of cytauxzoonosis. Additionally, only the blood samples collected from those infected bobcats could infect naive domestic cats following subinoculation (Kier et al. 1982b). Subsequent studies demonstrated that bobcats remain, for the most part, subclinically infected during the schizogenous phase of C. felis and gradually become chronically parasitemic (Glenn et al. 1982, 1983; Blouin et al. 1984, 1987). However, fatal cases of cytauxzoonosis in bobcats have been recorded in experimental studies (Kier et al. 1982a) as well as from natural C. felis infections (Nietfeld and Pollock 2002). Initial experimental trials demonstrated that the American dog tick (D. variabilis) can transmit C. felis from bobcats to domestic cats (Blouin et al. 1984). Subsequent studies with domestic cats demonstrated that the lone star tick, A. americanum, was a more competent vector for C. felis compared to D. variabilis. (Reichard et al. 2009, 2010; Allen et al. 2019). Nonetheless, both D. variabilis and A. americanum are considered vectors of C. felis.

Bobcats infected with C. felis have been documented in 16 US states: Arkansas, California, Colorado, Florida, Georgia, Illinois, Kansas, Kentucky, Missouri, North Carolina, North Dakota, Pennsylvania, Ohio, Oklahoma, South Carolina, and West Virginia (Supplementary Material Table S1). Previous surveys to ascertain the prevalence of C. felis in bobcats from Oklahoma found that 31–50% of the wild felids were infected (Supplementary Material Table S1; Glenn et al. 1982, 1983; Blouin et al. 1984, Kocan et al. 1985, Blouin et al. 1987). These studies probably underestimated the prevalence of C. felis because infection was based on identification of piroplasms in stained blood smears, a technique that is less sensitive than current molecular detection methods. Also, examination of blood smears for piroplasms does not allow for differentiation of morphologically similar species. For example, Shock et al. (2013) reported a female bobcat from Thomas County, Georgia, that was infected with a Babesia sp. However, other than this report of Babesia sp. in a single bobcat from Georgia, the only piroplasm of bobcats known in North America is C. felis. In another study, Shock et al. (2011) used PCR with piroplasm-specific primers and reported that 65% of 20 bobcats from 2 counties in north-central Oklahoma were infected with C. felis. Much more sensitive PCR assays are now available (Kao et al. 2021).

Bobcats are widely distributed throughout Oklahoma, and their population is considered to be increasing (Roberts and Crimmins 2010). Large portions of Oklahoma contain overlapping distributions of both tick vectors, lone star ticks and the American dog ticks (Barrett et al. 2015; Mitcham et al. 2017), where bobcats are located. Thus, the purpose of our study was to determine the regional prevalence and distribution of C. felis in bobcats using a highly sensitive droplet digital PCR (ddPCR) assay.

Bobcat tongue samples were collected by Oklahoma Department of Wildlife Conservation designated officials from carcasses of legally hunted or trapped bobcats during winter 2018–19 and winter 2019–20 from 53 counties in Oklahoma and three counties in Texas (Table 1). Tongues were removed, placed in individually labeled zip-close bags, and frozen at –20 C until they were processed for C. felis DNA extraction. Information on the age, sex, and detailed location of collection, other than the county, was not available for the bobcats tested.

Total DNA was extracted from frozen bobcat tongue samples (∼0.025 g) using a commercially available kit according to the manufacturer's instructions (Quick-DNA™ Miniprep plus Kit, Zymo Research, Irvine, California, USA). Following DNA extraction, concentration was measured using a Thermo Scientific Spectrophotometer NanoDrop 2000 (Thermo Fisher Scientific, Waltham, Massachusetts, USA), and the eluted solution was stored at –80 C for subsequent PCR reactions. Before the ddPCR assay, DNA concentration of each sample was adjusted to 55 ng/mL.

Droplet digital PCR was performed using the QX200™ ddPCR System (Bio-Rad Laboratories, Inc., Hercules, California, USA) as described by Kao et al. (2021). Briefly, 13.3 µL of ddPCR master mix (prepared by mixing 12 µlL super mix for probes [no dUTP] (Bio-Rad Laboratories), 1.2 µL of primer-probe mix (forward: 5′-CTA-CACTCTTTACACGTTTGTG-3′, reverse: 5′-CGAAATGCCAGTATACTCCT-3′, probe: 5′-TGAGTTTGCAAGGGCCATTATAACACC-3′) and 1.2 µL of DNase-free water) was mixed with a 8.8 µL aliquot of the DNA samples to generate the reaction mixes. From each of the prepared mixes, 20 µL were transferred to respective wells of DG8™ Cartridges (Bio-Rad Laboratories followed by 70 µL of Droplet Generation Oil for Probes (no dUTP) (Bio-Rad Laboratories). The DG8 cartridges were inserted into the QX200™ Droplet Generator (Bio-Rad Laboratories) to generate 40 µL droplet suspensions, which were then transferred to 96 Deep Well Reaction Modules (Bio-Rad Laboratories). The 96 well-plate was sealed, and PCR was carried out with a C1000 Touch™ Thermal Cycler (Bio-Rad Laboratories) with temperatures and times as described previously (Kao et al. 2021). After thermocycling, modules were transferred to a QX200 Droplet Reader (Bio-Rad Laboratories) for absolute quantification of the target DNA. All samples were run in duplicate, with a positive control from an experimentally infected domestic cat (Kao et al. 2021), a negative control from a specific-pathogen-free domestic cat, and a no-template control. Results were analyzed using QuantaSoft software version 1.7 (Bio-Rad Laboratories). Thresholds were set between the values of negative control (<1500) and the positive control (ranging from 2000 to 3000) as described previously (Kao et al. 2021). Samples were determined as positive when at least one droplet was classified as positive in each duplicate. The absolute copy numbers per reaction was calculated by the software based on the fraction of positive to negative droplets, using the Poisson law of small numbers (Hindson et al. 2011).

Prevalence and 95% confidence intervals (CI) were calculated using GraphPad Prism 8.0 software (La Jolla, California, USA). Prevalence of C. felis infection was determined according to Bush et al. (1997) and calculated for each county sampled. Because age, sex, and precise locations where bobcats were collected, other than county, were not available, data from individual counties were combined according to geographic regions (Table 2) of Oklahoma (Travel Oklahoma, June 2021). Differences in the prevalence of C. felis in bobcats were calculated for each geographic region compared to the rest of Oklahoma using chi-square tests and odds ratios (Sokal and Rohlf 1995). P values ≤ 0.05 were considered significant. Samples collected from the three counties of Texas were analyzed separately.

A total of 360 bobcat samples from 53/77 (68.8%) counties in Oklahoma (Table 1) were tested for infection with C. felis. The overall prevalence of C. felis in bobcats from Oklahoma was 80.0% (95% CI, 75.6–83.8). Cytauxzoon felis was not detected in 10 bobcats sampled from Ellis County or in one bobcat sampled from Texas County, both in western Oklahoma (Table 1 and Fig. 1). Thirteen bobcats were tested for infection from three counties in Texas including three from Hemphill County, nine from Lipscome County, and one from Wheeler County. When combined, the occurrence of C. felis in bobcats sampled from the three counties in Texas was 30.8% (95% CI, 12.4–58.0). The occurrence in Hemphill County was 33.3% (95% CI, 5.63–79.8), Lipscome County was 22.2% (95% CI, 5.3–55.7), and Wheeler County was 100.0% (95% CI, 16.8–100.0).

The overall prevalences of C. felis in bobcats from counties in southeast, northeast, central, and south-central regions of Oklahoma were >90% (Fig. 1 and Table 2). Of the bobcats sampled from central and south-central Oklahoma counties, all the samples were infected with C. felis, indicating a 100% prevalence. Prevalence estimates of C. felis in bobcats from southwest and northwest Oklahoma ranged from 55.6% to 67.9%. The odds ratio of a bobcat being infected with C. felis was 25.693 in central counties in Oklahoma followed by 12.422 in south-central counties, 7.646 in northeast counties, 3.824 in southeast counties, 0.358 in northwest counties, and 0.217 in southwest counties.

Cytauxzoonosis is the most severe tick-borne disease of cats in North America, and bobcats serve as wild vertebrate reservoirs for C. felis infection in domestic cats. Compared to the 65% prevalence of C. felis in bobcats in Oklahoma reported in 2011 (Shock et al. 2011), we report a 15% increase in the statewide C. felis prevalence in bobcats, to 80.0% (95% CI, 75.6–83.8). Bobcats from central, south-central, northeast, and southeast counties had the highest odds of being infected with C. felis; those from northwest and southwest counties had lower odds (Table 2).

A high prevalence of C. felis in bobcats from Oklahoma was expected and is probably due to several reasons including sampling strategy, high abundance of A. americanum, D. variabilis, and bobcats, and high sensitivity of detection using ddPCR. Previous estimates of the prevalence of C. felis in bobcats from Oklahoma were limited to sampling wild felids from one or only a few counties (Glenn et al. 1982, 1983; Blouin et al. 1984, 1987; Kocan et al. 1985; Shock et al. 2011). The 360 bobcats that we tested originated from 68.8% of the counties in Oklahoma and encompassed all major geographic areas of the state.

Both A. americanum (Reichard et al. 2009, 2010) and D. variabilis (Blouin et al. 1984, 1987) have been shown to transmit C. felis. Historically, A. americanum was limited to the southeastern and south-central US states (Bishopp and Trembley 1945), but the tick is now found widely dispersed across the southern and eastern US (Springer et al. 2014; Monzón et al. 2016; Saleh et al. 2021). Amblyomma americanum is widely distributed in Oklahoma (Barrett et al. 2015) and is less abundant in western regions that become arid (Supplementary Material Fig. S2). Dermacentor variabilis is widespread across North America and has expanded into parts of the western US (Minigan et al. 2018; Duncan et al. 2021; Saleh et al. 2021). Dermacentor variabilis is also widespread across Oklahoma (Supplementary Material Fig. S3) and is considered an abundant tick in this region (Mitcham et al. 2017). Both of these tick species have been identified outside their established distribution areas (Raghavan et al. 2019; Boorgula et al. 2020). As large portions of Oklahoma contain populations of both A. americanum (Supplementary Material Fig. S2; Barrett et al. 2015) and D. variabilis (Supplementary material Fig. S3; Mitcham et al. 2017), it is possible that bobcats (and domestic cats) in Oklahoma are at greater risk of becoming infected due to tick abundance and activity (Fig. 1 and Supplementary Material Figs. S2 and S3). Other distinct geographic areas with high abundances of bobcats and ticks would also be expected to have a high prevalence of C. felis in bobcats, as has been demonstrated in Missouri (Shock et al. 2011) and Illinois (Zieman et al. 2017).

Bobcats are also widely distributed across North America, and estimates of their populations suggest that the number of bobcats have increased over time (Woolf and Hubert 1998; Roberts and Crimmins 2010). Bobcats range through a variety of habitats such as forest, savanna, shrubland, grassland, and desert (Kelly et al. 2016); studies have demonstrated that bobcat densities have a positive relationship with forest cover and high prey densities, and a negative relationship with extensively modified landscape (Donovan et al. 2011; Reding et al. 2013; Broman et al. 2014; Thornton and Pekins 2015). Bobcats are considered widespread in Oklahoma, and samples included in the current study were collected from a large proportion of the counties (53/77) in Oklahoma. Estimates of bobcat populations in Oklahoma indicate that their numbers have increased since 1981 (Roberts and Crimmins 2010); the abundance of both bobcats and competent tick vectors across Oklahoma probably contributed to the high prevalence of C. felis that we detected. Within Oklahoma, several geographic regions were noted with a higher infection prevalence (central, south-central, northeast, and southeast), whereas in other areas (northwest, southwest), bobcats were less likely to be infected with C. felis. The 12 level III ecoregions of Oklahoma comprise vast plains, elevated karst plateaus, hills, and folded low mountains; the fauna including mammals and insects are greatly influenced by the east-west zonation of vegetation and climate (Woods et al. 2005). Eastern Oklahoma has the most forest cover, changing into more stunted, open grasslands in the west regions (Woods et al. 2005). Additionally, the climate of Oklahoma ranges from humid subtropical to semiarid moving from east to west (Oklahoma Climatological Survey 2022). These climatological differences result in a higher tick prevalence in the eastern regions compared to western regions because the subtropic climates favor tick survival (Estrada-Peña et al. 2012). As such, the differences in the prevalence of C. felis among regions within the state Oklahoma are probably due to ecological differences (such as ecosystem, landscape, and population) among the areas influencing bobcat abundance and activity as well as tick abundance and activity.

Previous studies estimating the prevalence of C. felis in bobcats from Oklahoma used either blood smear analysis (Glenn et al. 1982, 1983; Blouin et al. 1984, 1987; Kocan et al. 1985) or conventional PCR (Shock et al. 2011). In our current study, we used a ddPCR assay that has a high sensitivity and able to detect low levels of C. felis DNA in tissues. The ddPCR technique partitions the sample into approximately 20,000 water-in-oil droplets that either contain C. felis DNA or not (Pinheiro et al. 2012). After the PCR reaction, droplets are counted to detect the droplets containing target gene fragments, and absolute quantification is done using Poisson statistics (Hindson et al. 2011). The methodology employed by ddPCR permits high sensitivity and specificity, detecting as little as 0.00232 ng of C. felis DNA/reaction (Kao et al. 2021). This minute level of detection permitted application of these methods to detect C. felis piroplasms in red blood cells remaining in tongue tissues. Detecting or estimating the prevalence of C. felis in blood, spleen, or similar tissues would have been preferred as these tissues contain more red blood cells, but such samples were not available from the bobcats sampled in this study.

Although it is common knowledge that C. felis occurs in Texas, published cases on the occurrence, prevalence, and distribution are limited and outdated (Bendele et al. 1976). We found C. felis in 4/13 (31%) bobcats collected in Hemphill, Lipscome, and Wheeler counties in the Texas panhandle. There are no previous published reports of C. felis in bobcats from Texas (Supplementary Material Table S1). Future studies should focus on sampling more bobcats and other wild felids from diverse areas to determine where C. felis occurs in Texas wild felids.

Although the present study provided a comprehensive evaluation on the occurrence of C. felis in bobcats within Oklahoma, there were some limitations. Of the 53 counties sampled, for 22 counties (41.5%) there were fewer than five samples, which may have limited the detection of C. felis in low-prevalence areas. Similarly, 24 counties were not sampled, leaving the distribution of C. felis in bobcats from Oklahoma incomplete. Our study was also restricted by not knowing the exact locations where bobcats were collected, nor were data on sex or age of sampled bobcats available for analyses. Such data on location, sex, and age of bobcats would have permitted more in-depth analyses and would undoubtedly enhance our understanding of C. felis occurrence, distribution, and ecology within Oklahoma.

We demonstrated that C. felis infections in bobcats were highest in areas where A. americanum and D. variabilis populations and activity are well established and considered common. Although it is known that both A. americanum and D. variabilis transmit C. felis, much is still unknown regarding transmission dynamics. Continued surveillance of bobcats from enzootic areas for C. felis should be used to investigate the variables affecting transmission dynamics among the wild felids and to domestic cats.

We would like to thank personnel at the Oklahoma Department of Wildlife Conservation for collection of bobcat samples. We also thank the hunters and trappers who provided tongue samples. The research was supported through internal funds of MVR and CAM from the College of Veterinary Medicine, Oklahoma State University.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-22-00152.