Toxoplasma gondii Infection in the Eurasian Beaver (Castor fiber) in Switzerland: Seroprevalence, Genetic Characterization, and Clinicopathologic Relevance

Beavers (Rodentia, Castoridae) are the second largest living rodents, after the capybaras (Hydrochoerus spp.). The family Castoridae today consists of a single genus, Castor, which includes two species: the Eurasian beaver (Castor fiber), also called European beaver, and the Canadian beaver (Castor canadensis). In recent centuries, the Eurasian beaver was almost completely eradicated from vast areas in Europe, except for a few thousand specimens. Almost all of today’s occurrences in Europe are due to reintroduction projects in the 20th century (Angst and Fasel 2021; Zahner et al. 2021). In Switzerland, the Eurasian beaver had completely disappeared and was reintroduced in 1956. The Swiss population amounts to approximately 4,900 individuals (latest estimate 2022; Angst et al. 2023). Since the releases, two separate populations have formed in the catchment areas of the Rhone and Rhine rivers. Today, the beaver mainly inhabits water bodies in the Swiss Plateau, as well as the Alpine Rhine valley (Angst and Fasel 2021). As the largest rodent in Europe, the beaver is herbivorous and feeds mainly on herbaceous plants in summer and bark and buds of woody plants in winter (Angst and Fasel 2021).

Toxoplasma gondii (Apicomplexa, Sarcocystidae) is a protozoan parasite that occurs worldwide. Intermediate hosts are practically all warm-blooded animals, including a wide variety of mammals and birds (Dubey 2022). Definitive hosts are members of the family Felidae, which in Switzerland are represented by the domestic cat (Felis catus), the European wildcat (Felis silvestris), and the Eurasian lynx (Lynx lynx). Toxoplasma gondii causes one of the most common zoonotic diseases worldwide—toxoplasmosis in humans—and it is estimated that one third of the human population is infected (Halonen and Weiss 2013). Toxoplasma gondii exhibits a great genetic variability, which may be associated with different virulence phenotypes. Initially, three clonal archetypes, I, II, and III, were recognized, and these seem to prevail in the US and Europe, but in recent years, so-called nonclonal and variant or recombinant genotypes have been found (Grigg et al. 2001; Deiró et al. 2021). Those nonclonal genotypes are sometimes associated with more severe clinical disease in humans and animals (de-la-Torre et al. 2013; Hamilton et al. 2019; Pardini et al. 2019; Casartelli-Alves et al. 2021).

Rodents are among the most important intermediate hosts for T. gondii, as they are frequent prey of cats, which in turn shed infectious oocysts with their feces into the environment. Additionally, T. gondii is thought to influence rodent behavior (e.g., reduced fear of cats), which contributes to cycle maintenance (Webster 2001; Pardo Gil et al. 2023). Response to infection depends on different factors related to the host (e.g., species, age, immune status, coinfections with other agents) and the parasite (e.g., infection dose, parasitic stage, genetic background; Hassan et al. 2019; Stelzer et al. 2019; Dubey 2022). For example, strains from the Type I lineage are considered lethal for laboratory mice, whereas for rats they do not seem to be pathogenic. Type II and III lineages are commonly less virulent to mice, whereas some new nonclonal genotypes in Brazil caused lethal infections (Pena et al. 2008).

Little is known about T. gondii infection and toxoplasmosis in beavers. Of the three published studies investigating seroprevalence, two studies in North America on Canadian beavers reported seroprevalences ranging from 7% to 10% (Smith and Frenkel 1995; Jordan et al. 2005), and a study from the Netherlands on Eurasian beavers documented a seroprevalence of 6% (Maas et al. 2022). There are also very few studies on the clinical relevance of T. gondii infection in beavers. In one Canadian beaver, systemic toxoplasmosis was found to be the cause of death (Forzán and Frasca 2004). One study from Germany isolated T. gondii from tissues of two beavers, and in one of these animals, histopathologic changes mainly characterized by encephalitis associated to the presence of tachyzoites were found and were consistent with the death of the animal (Herrmann et al. 2013). In recent years, T. gondii infections have also been detected in Swiss beavers, and encephalitis was observed in some animals (Zürcher-Giovannini 2018), raising the question of the clinical relevance of infection in this animal species and of the role of beavers in T. gondii epidemiology.

Until recently, unlike Cryptosporidium and Giardia spp., T. gondii was not considered to be a significant waterborne pathogen, but several outbreaks in humans have led to a growing concern (Bowie et al. 1997; Keenihan 2002; Minuzzi et al. 2021). Additionally, in southern sea otters (Enhydra lutris nereis) in California, US, mass mortalities have been associated with T. gondii–contaminated freshwater runoff into coastal waters (Miller et al. 2002). Other rodent species with a semiaquatic lifestyle similar to beavers are the nutria (Myocastor coypus) and the muskrat (Ondatra zibethicus). Both have been used as sentinel species in serologic studies for T. gondii. Factors such as proximity to cities, agricultural land, or the amount of municipal waste in waterbodies have been assumed to play an important role for infection (Hejlícek et al. 1997; Ahlers et al. 2015, 2020; Zanzani et al. 2016). Given that beavers are mainly distributed along the water bodies of the Swiss Plateau, they are often located close to settlements and agricultural land and thus in potentially oocyst-contaminated areas.

The aims of our study were to investigate the epidemiologic role of the Eurasian beaver in the life cycle of T. gondii as an intermediate host and the clinical relevance of infection. For this, we combined several approaches: a seroprevalence study, assessment of risk factors for seropositivity, and parasite detection by molecular and histologic methods, as well as genetic characterization of the detected parasites.

Switzerland is a small country (41,285 km²) in central Europe and is home to numerous inland waters and watercourses that flow into various seas. The Rhine and the larger Rhine tributaries (Aare, Reuss, Limmat) in the east, northeast, and Swiss Plateau, as well as the Rhone in the canton of Valais, are of particular importance for the distribution area of the beaver. We obtained samples from beaver carcasses submitted by game wardens within the frames of the general surveillance program of wildlife health in Switzerland (Ryser-Degiorgis and Segner 2015) and of a beaver health surveillance project. Surveillance was intensified from 2002 and a detailed necropsy and sampling protocol was established at the Institute for Fish and Wildlife Health (FIWI) in collaboration with the beaver consultancy office. We included 247 dead beavers found 2002–2022, from which we obtained serosanguinous fluid samples from the heart and thoracic cavity. We centrifuged these samples to obtain sera, which were then stored at −20 C until further analysis. During necropsies, we collected brain samples (n=41) whenever possible. A set of metadata was compiled including sex, age, weight, body condition, coordinates of the finding site, and proximity to urban areas. We classified beavers into two age categories based on body size, weight, and body condition (Dežkin and Safonov 2004): juveniles (0–12 mo) and adults ≥1 yr old (cutoff at 16.95 kg in good body condition). The weight was taken in kilograms with a commercial scale, except for four animals. To investigate whether beavers were found in the immediate vicinity of humans (within/outside urban areas), we compared the coordinates of the finding sites with a map provided by the mapping platform of the Swiss Confederation (Swiss Confederation 2022) depicting the urban areas of Switzerland, generated with information from the Federal Statistics Office. Because of missing coordinates, the location of two animals could be assigned on only a cantonal level.

We drew the maps with the free software QGIS3 (QGIS.org 2017). Sample composition is provided in Supplementary Material (Supplementary Table 1).

We tested all 247 sera with a commercially available multispecies ELISA kit (ID Screen Toxoplasmosis Indirect, ID vet, Grabels, France), designed to detect specific antibodies against T. gondii in serum, plasma, or meat juice from different animal species, including dogs, cats, goats, sheep, cattle, and pigs (TOXO-MS). ELISA plates were precoated with cell culture–derived T. gondii tachyzoite antigen (TgSAG1/P30). The multispecies conjugate to be used was supplied in the kit. According to the manufacturer’s instructions, a sample to positive ratio (S/P ratio) based on the optical density of the samples was calculated for each serum. Samples with an S/P% ≤40% were considered negative, 40%<S/P%<50% inconclusive, and S/P% ≥50% positive. We included internal serum controls from a pig experimentally infected with T. gondii oocysts and from a seronegative control pig (Basso et al. 2017) in each plate.

We extracted DNA from fresh brain tissue samples as follows: 500 mg of tissue was homogenized with 900 µL Buffer ATL and 100 µL proteinase K and incubated overnight at 56 C. We then used 200 μL of the homogenate to complete further DNA extraction steps as indicated by the manufacturer (DNeasy Blood and Tissue Kit, Qiagen, Hilden, Germany).

We performed a Taq-Man–based quantitative PCR (qPCR) in a CFX96 qPCR instrument (Bio-Rad Laboratories AG, Cressier, Switzerland) to detect and quantify T. gondii DNA from a subset of fresh brain samples (n=41). We used the protocol from previous research (Scherrer et al. 2023) to target the 529-bp repetitive genomic sequence of T. gondii with a specific Taq-Man–based real-time qPCR and analyzed the PCR results with the CFX manager software version 1.6 (Bio-Rad).

Since 2012 the FIWI has systematically examined brain macroscopically and collected samples for histologic examination and archiving at −20 C. Histologic specimens were fixed in 4% buffered formalin, embedded in paraffin, stained with H&E, and further examined by light microscopy. For this study we included and screened only histologic specimens from which real-time PCR results from brain tissue were available (40/41; in one animal no histologic examination was performed).

In the H&E-stained histologic specimens, we looked for hallmarks including encephalitis, meningitis or meningoencephalitis, gliosis, tachyzoite aggregations, and tissue cysts containing bradyzoites.

We performed strain typing with the widely known PCR restriction fragment length polymorphism (RFLP) method using a multilocus nested PCR (Mn-PCR) amplification of 10 known genetic markers of T. gondii (SAG1, SAG2 [5′-SAG2, 3′-SAG2 and alt. SAG2 amplified separately], SAG3, BTUB, GRA6, c22-8, c29-2, L358, PK1, and Apico; Su and Dubey 2020). We adapted the protocol proposed by Su and Dubey (2020) by using a Multiplex PCR Kit (Qiagen) including a HotStarTaq DNA polymerase. The PCR products were purified and sequenced in both directions with the internal primers used in the previous Mn-PCR protocol (Mycrosynth, Balgach, Switzerland). To create alignments for analysis, we imported the resulting sequences into the Geneious Prime software version 2022.2.1 (Biomatters Inc. 2022). To obtain genotype specific fragment patterns we used an online program for in silico digestion of the sequences with specific enzymes (NEBcutter V2.0; Vincze et al. 2003; Castro et al. 2020). We assigned RFLP genotype numbers according to the ToxoDB database (Harb and Roos 2020).

Toxoplasma gondii population structure was evaluated through DNA sequence–based analyses. To generate an unrooted phylogenetic tree, all SAG3 consensus sequences and reference sequences were aligned using MEGA X software (version 11.0.13; Kumar et al. 2018). We inferred an evolutionary history using the neighbor-joining method.

Apparent seroprevalence was estimated based on the results obtained by ELISA and calculated with a free online sample size tool (Kohn and Senyak 2021). Parasite detection was considered positive when T. gondii DNA could be amplified from thawed tissues.

For statistical analyses we used the R program (R Core Team 2022). Level of significance was set at 0.05. To identify potential risk factors for seropositivity we first performed univariate comparisons between serology results (positive vs. negative) and explanatory variables such as sex (male, female), age (juvenile, adult), urban areas (within, outside) and weight (in kilograms). When explanatory variables consisted of two categories (i.e., sex, age, urban areas) we applied Pearson’s chi-square test. For continuous nonnormally distributed data (weight) we used Spearman’s rank correlation. In a next step, we investigated the effects of sex, weight, and occurrence in an urban area on the serology results by fitting a generalized linear model. We did not include age itself in the model because it was partially determined by the animal weight and because the factor age might act as a confounder rather than a causal factor (Opsteegh et al. 2012). To select the best model, Akaike’s information criterion (AIC) was used considering that models with a ΔAIC of 0–2 provide similar support (Anderson et al. 2001). For a subset of samples, we performed only univariate comparison between serology and PCR results (Fisher’s exact test) and PCR results and occurrence of encephalitis in histology (Pearson’s χ² test).

Overall, 113/247 (45.8%; 95% CI, 39.5–52.0%) sera tested positive for specific anti–T. gondii IgG antibodies according to the manufacturer’s thresholds of the ELISA kit. None of the samples had an inconclusive result. The Switzerland-wide distribution of seropositive and seronegative individuals along the Swiss water bodies is shown in Figure 1. Apparent prevalences according to the different explanatory variables are shown in Table 1. Univariate comparison showed a highly significant difference (P<0.001) with the factor weight and a significant difference regarding factors age (P=0.005) and proximity to urban areas (P=0.007). Sex did not show a significant effect (P=0.621). In the best-fitting generalized linear model, “weight” and “urban area” were retained as significant (P<0.05) predictors of seropositivity (Table 2). Heavier animals and animals found within an urban area were more likely to be seropositive. When looking at a subset of animals with PCR results available, univariate comparison revealed a strong association (P<0.001) between seropositivity and PCR positive results for T. gondii DNA in brain tissue (Table 3). The geographical distribution of beavers for which both serology and PCR results were available is shown in Figure 2.

For brain samples, 23/41 (56%; 95% CI, 41–71%) tested positive for T. gondii DNA by real-time PCR. We found inflammatory changes (i.e., encephalitis, meningitis, choroid plexitis, or a combination of these) in 26/40 histologic specimens and observed parasitic stages (tachyzoite aggregations or tissue cysts) in 6/26 samples (examples in Fig. 3).

Real-time PCR of the corresponding fresh brain samples was positive for all six specimens in which parasites were detected and for an additional 11 samples without visible parasite stages. Univariate comparison showed a mild correlation (P<0.05) between positive PCR results and histopathologic findings of encephalitis (Table 3).

We successfully amplified and sequenced all 10 markers (including the three separately amplified SAG2 regions) from 22/23 brain samples that were positive by real-time PCR. Final allele patterns for all the samples are shown in Table 4. Two samples corresponded to the clonal Type II (ToxoDB 1), 19 samples to a lineage II variant strain (ToxoDB 3), and one sample revealed a novel genotype (data submitted to ToxoDB for inclusion) with alleles corresponding to allele Type II for nine markers and allele Type I for two markers (alt. SAG2, Apico). Sequencing revealed single-nucleotide polymorphisms (SNPs) in 21 analyzed sequences (whole amplified sequences including primers were considered to define the SNP positions). Concerning marker SAG3, 17 of 21 sequences showed 100% identity, with isolates having an SNP at position 187 with a change from G to T compared with the clonal Type II reference strain (ME49). Another SAG3 marker sample featured an SNP at location 153 with a change from G to C that did not show 100% homology with any sequence deposited on GenBank. A sequence from the marker alt. SAG2 had an SNP at bp 118. The affected site is a known interface for the enzyme TaqI in the RFLP assay. The reference sequences, both ME49 and VEG, have a TCGA nucleotide sequence at this location, which in our sample had a TCAA sequence. According to the RFLP cut pattern, this sequence is attributed to the Type I allele, but phylogenetically the sequence was more closely related to the Type II reference sequence ME49 (Fig. 4). In two PK1 sequences there was an SNP (bp 816) with a nucleotide change from A to G compared with the reference (ME49). Respective sequences are provided with GenBank accession numbers in Table 4.

With an estimated apparent prevalence of 45.8%, beavers in Switzerland show a higher prevalence than reported, for example, in beavers in the Netherlands (6%; Maas et al. 2022) or in Canadian beavers in Missouri-Kansas and Massachusetts (7% and 10% respectively; Smith and Frenkel 1995; Jordan et al. 2005), suggesting a higher exposure of beavers to T. gondii in Switzerland. In some of the mentioned studies, a different detection method was used (modified agglutination test); therefore, the results cannot be directly compared. Although the commercial ELISA used in this study has not been validated for beaver, the clearly distinguishable values of positive and negative results (Fig. 5) are a good indication that the method works for the detection of antibodies in beaver. Furthermore, there was a strong correlation between seropositive animals and positive PCR results from the brain when available. This is interesting because it is not always possible to detect the parasites in the tissues of seropositive animals (Glor et al. 2013).

Our finding of positive association between seropositivity and weight has also been seen in a survey of rats from France, and suggests postnatal transmission (Ayral et al. 2015). This is also reflected by the fact that adult animals were more likely to be seropositive, which is explained by an increasing cumulative likelihood of exposure during an animal’s lifetime and long-lasting antibody persistence (Zarnke et al. 2001; Roelke et al. 2008; Opsteegh et al. 2012; Basso et al. 2020, 2022).

A major issue with T. gondii in (semi)aquatic ecosystems is the route of transmission. Because beavers are herbivorous, we suspect that they become infected directly through water or plants contaminated with oocysts. Lindsay et al. (2003) demonstrated that oocysts remained viable for several months in seawater and were able to infect marine mammals (van de Velde et al. 2016; Dubey et al. 2020). In some cases, in both humans and animals, toxoplasmosis outbreaks have been linked to oocyst-contaminated fresh water (e.g., creeks and streams), or water reservoirs, which also confirms viability in freshwater systems (Bowie et al. 1997; Aramini et al. 1999; Keenihan 2002). Often these systems were equipped with inadequate filtration systems, but various methods of water treatment sometimes fail to eliminate oocysts (Dubey 2022). Beavers in our study found within an urban area had a higher probability of being seropositive; such associations between urban areas and higher T. gondii seroprevalence have also been found in otters (E. lutris nereis), muskrats, and nutrias (Miller et al. 2002; Ahlers et al. 2015; Zanzani et al. 2016). This may be explained by the higher cat density in regions highly populated by humans and thus more oocysts in the environment. Modified drainage systems used in agriculture and urbanized areas have been associated with facilitating transport of T. gondii into freshwater ecosystems (Ahlers et al. 2015). In comparison, in a study in Minnesota, far away from most human influences, no antibodies against T. gondii were detected in various populations of muskrats (Ahlers et al. 2020). To accurately estimate the risk of oocyst ingestion, additional research on the environmental contamination burden is necessary. It would be interesting to attempt to detect T. gondii DNA directly from the water in affected areas. The lack of standardized assays for detection of T. gondii oocysts in water bodies may have contributed to an underestimation of the role that oocysts play in the epidemiology of T. gondii (Shapiro et al. 2019).

In most of the T. gondii–positive brains analyzed in our study, encephalitis was mild and of chronic nature. This suggests that infected beavers tend to develop subclinical and chronic disease courses. The cases with acute encephalitis were usually beavers with systemic toxoplasmosis, that is, with involvement of other organs such as the lungs, heart, or liver; with inflammatory and necrotic changes; and with presence of tachyzoites associated to these lesions. Such changes were seen in the histology of a case from 2022 (ID W22_6329; Fig. 3). The pathologic changes are not new to Swiss beavers, as they have been reported associated with T. gondii infection in a previous study on beaver health (Zürcher-Giovannini 2018). Documented cases of T. gondii infection in beavers exist at the FIWI only from 2012 onwards; this probably does not indicate that toxoplasmosis is an emerging disease in the area, but rather is explained by the fact that previously beaver carcasses, especially skulls, were spared for later preparation and brain tissue was not routinely examined. Deaths due to clinical toxoplasmosis have also been reported in cases from Germany (free-ranging beaver; Herrmann et al. 2013) and the US (captive C. canadensis; Forzán and Frasca 2004).

It is difficult to assess to what extent the observed histopathologic changes affected the life of beavers. In our study population the most common cause of death was trauma (Ryser-Degiorgis and Zürcher, pers. comm.). It may be that encephalitis affects the behavior or responsiveness of the animal; for example, sea otters were observed to be significantly more likely to be attacked and killed by sharks if they had concurrent moderate to severe encephalitis caused by T. gondii (Kreuder et al. 2003). Loss of aversion towards predators has been described in small rodents such as mice and rats (Vyas et al. 2007) but not yet in beavers. When considering that this change in behavior helps the parasite to be ingested by the final host, the beaver seems likely to be of lesser importance in this role: Domestic cats, wild cats, and Eurasian lynx do not prey on beaver, although in rare cases they might ingest beaver meat through scavenging. Scavenging of dead beavers may play a role in the maintenance of the life cycle among intermediate hosts, (e.g., wild canids, mustelids, scavenging birds). Histopathologic examination is not a sensitive method to detect T. gondii because parasitic stages in tissues are generally scarce, especially during the chronic phase of infection, and their distribution is not homogeneous. In some of our cases we were able to clearly distinguish between acute infections (tachyzoite aggregates) and chronic infections (tissue cysts). Conversely, in many other samples we were not able to see the parasites in tissue sections, only the inflammatory reaction to their presence. In such cases the diagnostic value of PCR was shown.

We detected most frequently ToxoDB genotype 3 (in 19/22 positive samples). Within the different marker sequences, we detected SNPs that further distinguish the type. We found most frequently the mutation in marker SAG3. In Switzerland, this mutation has also been detected in a lynx (OQ230332; Scherrer et al., 2023) and different voles (OQ108542–108544; Pardo Gil et al. 2023). ToxoDB 3 is generally considered to be weakly virulent in mice; however, fatalities caused by this genotype have been reported in other species (Spycher et al. 2011; Jokelainen 2012). In our study, 14/19 beavers with this genotype had encephalitis, and in five of them toxoplasmosis was assumed to be the cause of death, showing that clinical toxoplasmosis can be an important disease in beavers.

In Spain, a flock of sheep suffered an abortion storm due to Toxoplasma from this genotype exhibiting the same mutation (Fernández-Escobar et al. 2020). In contrast to that Spanish study, we did not see a clear differentiation of ToxoDB 3 from ToxoDB 1 (ME49) due to this mutation.

Two animals infected with T. gondii of a ToxoDB 3 genotype exhibiting the mentioned mutation in the SAG3 sequence had an additional mutation in the PK1marker sequence. These two beavers were found in close proximity to each other 6 mo apart, suggesting a local variation of this genotype. The 22 genotyped samples date from 2012 to 2022, showing that genotypes (in our case mainly ToxoDB genotype 3) may persist in a region for several years, sometimes passing on these distinct mutations.

One animal showed a mutation in the alt. SAG2 marker, which resulted in an allele pattern not yet described (ID W21_1294, OQ402909). Although a Type I allele was present, which in principle is considered more virulent (Behnke et al. 2011), pathology did not differ compared with animals infected with other genotypes. However, a more detailed investigation of pathogenicity in this case could not be performed because DNA was isolated from frozen material and a bioassay was not possible.

Our findings confirm that toxoplasmosis is a relevant disease in beavers and that they are able to act as an intermediate host for T. gondii, in which role they may contribute to further parasite spread in semiaquatic habitats. The high seroprevalence indicates frequent contact with the parasite, and most beavers appear to develop subclinical to chronic disease courses, though acute and fatal outcomes also occur. Transmission through water contaminated with oocysts seems likely, which opens the floor for further research.

We would like to thank the Swiss cantonal hunting authorities, game wardens and hunters for submitting beaver carcasses for necropsy, all FIWI collaborators who compiled data and collected samples during necropsy and the diagnostic staff of the Institute of Parasitology for their excellent technical assistance in the laboratory, and especially Gaston More for his help with genotyping.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-23-00077.