HEPATOZOON RUFI N. SP. (APICOMPLEXA: HEPATOZOIDAE) OF BOBCATS (LYNX RUFUS) FROM MISSISSIPPI

During surveys for Toxoplasma gondii and Sarcocystis infections in bobcats, muscle samples of 25 bobcats were obtained from Mississippi in 2017 as reported previously (Dubey et al., 2023, 2024). Bobcats were trapped legally in February of 2017. The sampling was from remote counties (Claiborne, Warren, and Jefferson) of Mississippi and included bobcats of both sexes. Samples of the tongue, heart, and limb muscle were collected, put in Ziploc bags, and transported to the Animal Parasitic Diseases Laboratory (APDL), U.S. Department of Agriculture (USDA), Beltsville, Maryland for testing. Up to 4 days elapsed between collection and transport of samples to APDL. This survey, intended to also examine the prevalence of T. gondii, was suspended in 2018 when the USDA redirected its food safety research efforts; the remaining frozen samples were then incinerated and discarded, but paraffin-embedded muscle samples were retained.

From each bobcat, we prepared one-half of the cross-section from ventricular myocardium, 1 cross-section of the tongue at midpoint, and a 2 × 2 cm section of leg muscle. These samples were fixed in 10% buffered formalin, embedded in paraffin, sectioned at 5 µm, stained with hematoxylin and eosin (HE), periodic acid Schiff (PAS) counter-stained with hematoxylin or Gomori’s methenamine silver (GMS), and examined microscopically for parasites as described (Dubey et al., 2024). PAS is used to stain polysaccharides, and GMS is used to stain the cyst wall. Numerous histological sections were made of the hearts of the 3 most infected bobcats (nos. 4, 16, and 17), and each section was examined at ×1,000 magnification for stages of Hepatozoon spp. Parasites detected in histological sections were photographed with an Olympus AX-70 microscope and DP-73 digital camera, and measurements were made digitally with Olympus Imaging Software cellSens Standard 1.18 (Olympus Optical Ltd., Tokyo, Japan).

Genomic DNA was extracted from paraffin sections (5 µm thick) from the myocardium of 2 bobcats (nos. 4 and 17) with the Qiagen (Hilden, Germany) DNeasy® Blood and Tissue Kit following the manufacturer’s instructions. These specimens were selected because they had several Hepatozoon meronts. Amplification of the 18S rRNA gene was performed for molecular identification with primers retrieved from the literature (PCR1F: 5′-CCAGCAGCCGCGGTAATTC-3′, HepaR: 5′-GACAGTTAAATACAAATGCC-3′; Baneth et al., 2013; Díaz-Regañón et al., 2017; van As et al., 2020). The 13.5 µl PCR mix consisted of a 2 μl DNA template, 6.25 μl of Platinum Hot Start PCR Master Mix (Invitrogen, Thermo Fisher Scientific, Waltham, Massachusetts), 0.5 μl of each 10 pmol/μl primer (Integrated DNA Technology, Coralville, Iowa), and 4.25 μl of molecular grade water. After initial denaturation at 94 C for 3 min; 45 cycles were performed consisting of denaturation at 94 C for 30 sec, annealing at 60 C for 30 sec, and elongation at 68 C for 20 min; terminal elongation was accomplished by incubating the products at 68 C for 5 min. Initially, no amplification could be achieved using the above PCR conditions; hence, samples were re-amplified under the same conditions with PCR-amplified products as templates. The final PCR products were analyzed on a 2% agarose gel running at 100 V for 45 min, and the size of the amplicons was estimated by comparison with thae 100 bp DNA ladder (Promega Corporation, Madison, Wisconsin). PCR products were cleaned with the ExoSAP method (Bell, 2008). Sanger sequencing was performed at the DNA sequencing unit of Psomagen (Rockville, Maryland) on an ABI 3500x1 Genetic Analyzer (Applied Biosystems, Thermo Fisher Scientific, Waltham, Massachusetts). The resulting sequences were visualized, assembled, and edited using Geneious Prime 2023.0.4 (http://www.geneious.com). The sequencing of the 18S rRNA gene resulted in a 467 bp sequence with a guanine-cytosine content of 31.9%. The edited and assembled sequence from 1 bobcat (no. 17) was deposited in GenBank (accession no. PP492441).

We tried additional PCRs (touchdown and gradient) but achieved an amplicon of only 467 bp (primer pairs H14HepaF and HepR) after many sets of amplifications and re-amplifications.

In addition, we designed new primers for 28S, ITS1, and CO1 genes using reference Hepatozoon sequences from the NCBI but failed to achieve amplification. Likely, the paraffin blocks contain low-quantity or degraded DNA. Our primers also might not perfectly target the particular species of Hepatozoon present in this sample. For reference and primer design, only a few representative Hepatozoon spp. are available at the NCBI.

To reconstruct the phylogenetic position of the sequences, we used the Basic Local Alignment Search Tool (BLAST) searches against the NCBI nr database. The search identified more than 99% identity with Hepatozoon sp. KEA-2009a, Hepatozoon felis, Hepatozoon luiperdjie, and Hepatozoon sp. NCM11 and 95% with Hepatozoon silvestris (Table III). We inspected sequencing traces using Geneious Prime to identify any ambiguous base calls or alignment gaps. We also used the web server Guidance 2 (Sela et al., 2015) to align and remove ambiguously aligned positions. Specifically, the sequences were aligned with the MAFFT algorithm under the options Max-Iterate: 1.000 and Pairwise Alignment Method: –localpair. The final dataset included 12 taxa and 1,839 positions. Positions with a score below 0.93 were removed as were positions with more than 25% missing data. Phylogenetic relationships were reconstructed under the maximum likelihood (ML) criterion. ML analyses were performed with the program IQ-TREE version 1.6.12 (Nguyen et al., 2015). The analyses were run with the options –m MFP –b 1,000 –nt 5 (i.e., ModelFinder + tree reconstruction + 1,000 non-parametric bootstrap replicates + tree topology test). The model selected based on the lowest BIC value was the HKY+G model with a score of 2452.556.

In 2014, 35 bobcats from Mississippi were surveyed for Sarcocystis and Toxoplasma (Verma et al., 2015, 2017). For histological examination, samples of tongues were fixed in 10% formalin, and paraffin-embedded sections were examined microscopically. Frozen samples had been discarded. These sections from the tongue were examined recently for Hepatozoon infections.

Developmental stages of Hepatozoon sp. were detected in the myocardium, tongue, and limb muscle of 11 (nos. 2, 4, 6, 9, 13, and 14–19) of 25 bobcats (myocardium of all 11, tongues of 4, and limb muscle of 5); meronts were seen in all 3 tissue types from 4 cats (nos. 4, 14, 15, and 17) (Table I). Concurrent Sarcocystis sarcocysts were detected in 8 of the 11 Hepatozoon-infected cats (Table I). More hepatozoans were found in the myocardium than in the tongue or limb muscle (Figs. 1–6). The distribution of Hepatozoon stages was erratic; several parasites were present in some fields (Fig. 1), whereas they appeared sparsely distributed in others. Hepatozoon spp. stages were difficult to locate in sections stained with HE because they were not distinctively different from host tissue and because of lack of inflammatory host response. PAS staining helped to locate meronts, particularly immature stages, even under low magnification (Fig. 1A, B). The host cell or cells parasitized were not identified but were presumed to be a myocyte or a mononuclear cell.

Three types of meronts were recognized (Figs. 2–6), which we here designate as types 1, 2, and 3 (Table II). Type 1 meronts were found in the myocardium of 2 bobcats (nos. 4 and 16). These meronts were so small as to be overlooked during the initial examination. Most of the stages found were in a small focus (<0.5 mm) in the myocardium of bobcat no. 16 (Fig. 2A), confined to 2 (5 µm) sections; they were not found in 12 other nearby sections. The following description of type 1 meronts is based on samples from bobcat no. 16.

Type 1 meronts were enclosed in a vacuole. The meront maximum dimension was 15 µm, including the vacuole, and contained 2 or 4 merozoites, often in pairs (Fig. 2). Merozoites were robust, with rounded ends and were 9.5–11 × 3.0–4.5 µm; the nucleus was 2.5–3.5 µm wide. The merozoite cytoplasm was eosinophilic and appeared foamy (Fig. 2D). The basophilic nucleus was located centrally and appeared lobed in some merozoites. A residual body or multinucleated stages were not recognized. Some paired merozoites appeared to be connected (Fig. 2D). Type 1 meronts were PAS negative (Fig. 2G–J). Two type 2 meronts (described later) were present in another area of the myocardium of bobcat no. 4. One presumed type 1 meront was also found in the myocardium of bobcat no. 4 (Fig. 4H).

Type 2 meronts were found in all 11 infected bobcats (Figs. 3, 4). They were round, oval, or pear shaped. The earliest stage or stages were uninucleate zoites that varied in morphology (Figs. 1B, 3A, 3B). Most zoites were surrounded by a vacuole. The nucleus in these zoites was vesicular, mostly located centrally, and occupied nearly half of the zoite width. These zoites were PAS negative (Fig. 3B, C). Figure 3A shows a 12 × 5.5 µm zoite probably in a phagocytic cell. Figure 3B shows a 12.7 × 4.0 µm zoite in a vacuole; the nucleus is around 2 µm wide. Figure 3C shows a robust zoite (10.2 × 5.5 µm) that may be the youngest stage of type 2 meront. Figure 3D shows a 17 × 8.5 µm oblong meront with a 6 × 3 µm nucleus. A later stage was a 14.8 × 8.8 µm meront with 1 prominent nucleus and surrounded by granular PAS-positive cytoplasm (Fig. 3E). The first indication of nuclear division is depicted in Figure 3F; the meront has 2 nuclei, one of which is dividing further. In subsequent stages observed, the nuclei were located mostly at the periphery of the meront. Figure 3H shows a 13.2 µm long meront with 7 peripherally arranged rectangular nuclei in focus; the meront is enclosed in a 2.8 µm wide capsule. In more advanced meronts containing 10 or more nuclei, the nuclei became denser (Figs. 3K, 3L, 4A). Merozoites budded at the periphery of a residual body/mass (Fig. 4D, E). Figure 4E shows a mature meront with a condensed, discrete residual body and peripherally located merozoites. The number of merozoites or nuclei was not determined, but up to 44 nuclei could be counted (Fig. 4B). Type 2 meronts, inclusive of their capsule, were up to 39 µm long and up to 24 µm wide. Ten longitudinally cut merozoites were 7.5–9.0 µm long and around 2 µm wide (Figs. 4F, 4G, 6B). The merozoite nucleus was rectangular and occupied the entire width of the merozoite. In some meronts, mature merozoites were located within a capsule, but a meront membrane could not be identified around merozoites (Fig. 4G). Mature meronts without a residual body were PAS negative; merozoites were eosinophilic with a basal nucleus. Figure 4G shows a PAS-positive meront next to a mature meront with PAS-negative merozoites. In sections stained with GMS, the merozoite nuclei stained black and were enclosed in a silver-negative thin membrane; the capsule was prominent and unstained (Fig. 4C). The capsule size differed with the development of the type 2 meronts; it was 0.5–6.0 µm wide (Fig. 4).

A third type of meront was discovered (type 3) while scanning slides stained with PAS. These meronts appeared to be present intracellularly between myocytes and were up to 37 µm long and contained up to 42 nuclei. These meronts were PAS positive, nuclei/merozoites (Fig. 2) filled the meront, and there was no residual body. The merozoites were short, stubby (around 6 × 2 µm), with a central nucleus. A capsule could not be detected around meronts (Fig. 5).

Ruptured or rupturing meronts were seen rarely (Fig. 6), and merozoites were seen among mononuclear cells (Fig. 6B). Whether merozoites belonged to different types or generations was not determined.

Individual intracellular zoites were seen in the blood (Fig. 6E, F); the parasitized host cells could not be identified but appeared to be leukocytes. Whether the zoites were merozoites or gamonts could not be ascertained.

Amplicons were derived from H. rufi n. sp. from the paraffin-embedded muscle section of bobcats from Mississippi. Although we generated sequences from 2 bobcats (nos. 4 and 17), sequences from only 1 bobcat (No. 17) were used in the phylogenetic analysis because the 2 sequences shared 100% identity. The final assembled sequence had more than 99% homogeneity to Hepatozoon spp. reported from South Africa, the United States, Japan, and Hungary and 95% for identity with Hepatozoon spp. from Bosnia and Herzegovina and the Czech Republic, indicating the vast geographical range of these highly enigmatic parasites (Table III). The initial few bases at the beginning and a few bases at the end of the initial alignment were trimmed to create the final alignment to focus on regions of interest. The final aligned file had the following pattern of differences from the sequence under study: it differed from Hepatozoon sp. KEA-2009a by 2 gaps at positions 503 and 980 and single nucleotide polymorphisms (SNPs) at positions 499 and 592; from H. felis by 1 gap at position 981 and SNPs at positions 652, 681, and 991; from H. silvestris by 12 gaps at positions 682–687, 707, 744, 745, 746, 747, and 981 and SNPs at positions 498, 541, 542, 733, 736, 770, 773, 776, and 867; from H. luiperdjie by 1 gap at position 981 and SNPs at positions 552, 652, and 990; and from Hepatozoon sp. NCM11 by 1 gap at position 981 and SNPs at positions 552, 652, and 990 (Table III; Suppl. Data, File S1).

The phylogenetic tree was reconstructed including the present H. rufi n. sp. (accession no. PP492441) with the closely related species retrieved from the BLAST database. Adelina grylli (accession no. DQ096836) was used as an outgroup (Fig. 7). The ML phylogenetic tree based on the partial 18S rRNA gene formed 2 clusters; 1 cluster includes the sequence under study and Hepatozoon sp. KEA-2009a (accession no. FJ895406) obtained from Sylvilagus floridanus from the United States. Strong bootstrap support (89) exists for this pair, which is then mostly closely related to 6 other isolates from various felids (Hepatozoon sp. isolate NCM11 (acc. no. MK621318) from South Africa, H. luiperdjie (acc. no. MN793004) from South Africa and 4 isolates of H. felis (acc. nos. AB771545-Japan, OM422756-Hungary, AB983400-Japan and LC179796-Japan) with high consensus support. Those might be subdivided, but the evidence supporting each division is uncertain. As a group, these appear distinct from H. silvestris (acc. no. KX757032) from Bosnia and Herzegovina and Hepatozoon americanum (acc. no. AF176836). The evidence for the monophyly of these two is weak (39%) and long branches separate them. The distinction between H. canis and the remaining exemplars appears stronger, supported by 80% of bootstrap replicates. Thus, the phylogenetic evidence suggests that parasites infecting feline intermediate hosts share a distinct common ancestry from those infecting canid intermediate hosts; the only exception to this view is the equivocal placement of C. familiaris.

Immature and mature meronts were detected in tongues of 3 (nos. 10, 12, and 30 – details of cats were stated in Verma et al., 2017) of 35 bobcats. Structurally, the meronts were like those in bobcats sampled in 2017 (Fig. 1C).

Hepatozoon rufi n. sp.
(Figs. 1–7)

Three types of meronts in muscular tissues. Type 1 meronts, PAS-negative, up to 15 µm in maximum dimension, contain 2–4 robust 9.5–11 × 3.0–4.5 µm merozoites, division into 2, probably by endodyogeny. Type 2 meronts enclosed in a 0.5–6.0 µm wide capsule, contain up to 44 merozoites/nuclei, and measure up to 39 µm long and up to 24 µm wide including capsule. Immature meronts, PAS-positive, contained peripherally located nuclei. Merozoites formed at the periphery around a prominent PAS-positive residual body. Type 2 merozoites were 8.2–9.0 × 1.9–2.6 µm, PAS-negative, nucleus occupying the entire width of the merozoite. Type 3 meronts, PAS-positive, not enclosed in a capsule, contained up to 42 nuclei, and no detectable residual body. Gamonts unknown.

Bobcat (Lynx rufus).

Heart, tongue, limb muscle.

Unknown.

Mississippi, USA.

The specimens were deposited in United States National Parasite Collection in the Division of Invertebrate Zoology and National Museum of Natural History, Smithsonian Institution, Museum Support Center, MRC 534, 4210 Silver Hill Road, Suitland, Maryland 20746, USA, under numbers USNM (Table IV). DNA sequences were deposited in GenBank with accession number PP492441.

urn:lsid:zoobank.org:act:0355BB6B-31C8-4549-9F27-6FBCE7CC94EA.

Species named after the species of the bobcat, Lynx rufus.

This is the first morphological description of the development of meronts of Hepatozoon in the bobcat and given a new name, Hepatozoon rufi. The new species is distinguished from the 2 known species of Hepatozoon of felids (Table II). It more closely resembles H. felis than H. silvestris. However, merozoites of H. felis are arranged haphazardly without a residual body whereas in H. rufi merozoites are arranged peripherally around a prominent residual body. Additionally, Type I meronts were not found in H. felis. Hepatozoon rufi is morphologically distinct from H. silvestris based on the size and structure of merozoites and the capsule (Table II). Hepatozoon silvestris type 1 merozoites (macromerozoites) are much smaller (around half the size) than merozoites of H. rufi (Table II). Additionally, the capsule surrounding H. silvestris meronts is much thinner than the one surrounding H. rufi (Table II). Though H. rufi resembles some features of H. felis and H. silvestris, morphological and genetic evidence differentiate H. rufi as a distinct species derived from ancestors that also gave rise to congeners which likewise infect feline intermediate hosts.

The present study described the development of Hepatozoon meronts in the muscular tissues of bobcats for the first time. Of the 3 muscular organs examined, the myocardium was parasitized more than the skeletal muscles of the tongue or limb. In the present study, the host cell parasitized was not determined but appeared to be a monocyte or a myocyte. Klopfer et al. (1973), who first described Hepatozoon spp. in the heart of domestic cats, stated that meronts were intravascular. However, according to Baneth et al. (2013), H. felis meronts in domestic cats occurred in monocytes or myocytes. Our findings are in accord with Baneth et al. (2013), but ultrastructural studies are needed for confirmation. The host cell parasitized with H. silvestris was not stated.

In the present study, 3 types of meronts were recognized. Whether there is more than one replicative cycle of merogony of Hepatozoon in wild felids could not be determined, and will be difficult to determine without experimental infections, likely not possible in bobcats. To our knowledge, more than 2 types of meronts have been reported first time in Hepatozoon spp. for any Hepatozoon species in felids.

In the present study in bobcats, type 1 meronts were considered H. rufi solely based on structure; detailed observations were limited to parasites in bobcat #16. Currently, there are no specific immunohistochemical tests for H. rufi, and no frozen tissue is available for the extraction of good-quality DNA. Two type 2 meronts were also present in the same bobcat.

The site of development of Hepatozoon spp. probably varies with the species of the host and the parasite. Our review of available literature suggests that Hepatozoon spp. in felids parasitize muscles. Hepatozoon felis, H. silvestris, and H. rufi predominantly parasitize the myocardium, although tissue predilections differ (Table II). For example, the tongue was not parasitized by H. felis and H. silvestris whereas H. rufi was found in tongues (in addition to myocardium and limb muscle). Lane and Kocan (1983) may not have examined muscle tissues for infection.

Pathogenicity of Hepatozoon spp. varies with the parasite and the host. In the present study, only a few localized Hepatozoon-associated inflammatory lesions were found in bobcats. Nothing was known of the clinical status of bobcats because they had been hunted. Among domestic cats, most cases of clinical hepatozoonosis were in cats with immunosuppressive conditions (Baneth et al., 1998, 2013; Díaz-Regañón et al., 2017; Kegler et al., 2018; Basso et al., 2019; Schäfer et al., 2022; Simonato et al., 2022). However, fatal H. silvestris-associated myocarditis has been reported in a domestic cat in Switzerland; that case lacked evidence for immunosuppression and had been immunized against Feline Immunodeficiency Virus (FIV) and Feline Leukemia Virus (each of which can cause immunosuppression) (Kegler et al., 2018). It suffered severe myocarditis in both auricles and ventricles; its numerous H. silvestris meronts were confined to the heart; the pancreas, kidneys, stomach, intestines, brain, and bone marrow were not affected (Kegler et al., 2018). Myocarditis was also found in 3 of 4 European wild cats infected with H. silvestris (Hodžić et al., 2017). Here, we examined only the heart, tongue, and limb muscles.

This research was supported in part by an appointment of Aditya Gupta and Larissa Araujo to the Agricultural Research Service (ARS) Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the U.S. Department of Energy (DOE) and the U.S. Department of Agriculture (USDA). ORISE is managed by Oak Ridge Associated Universities (ORAU) under DOE contract number DE-SC 0014664. We thank Dr. Matthew J. Lovallo, Game Mammals Section, Bureau of Wildlife Management, Pennsylvania Game Commission, Harrisburg, Pennsylvania for samples of bobcats. We are indebted to Dr. Gad Baneth for advice. We also thank Dr. Adnan Hodžić for his helpful suggestions.