CANINE DISTEMPER VIRUS ECOLOGY: INSIGHTS FROM A LONGITUDINAL SEROLOGIC STUDY IN WILD RACCOONS (PROCYON LOTOR)

Raccoons (Procyon lotor), a ubiquitous species in North America, are reservoirs for several pathogens that are of zoonotic and conservation importance. In addition to rabies virus, canine distemper virus (CDV) and parvovirus circulate in raccoon populations (Mitchell et al. 1999; Junge et al. 2007; Rainwater et al. 2017). Increasing awareness of CDV in diverse hosts, and of changing CDV dynamics (Beineke et al. 2015; Riley and Wilkes 2015; Viana et al. 2015), have led to renewed interest in the ecology of CDV in wild animal populations. Many wildlife CDV studies have focused on documenting prevalence across different populations or geographies (Hoff et al. 1974; Roscoe 1993; Mitchell et al. 1999), understanding pathogenesis and epidemiologic characteristics (Pfeffermann et al. 2018; Needle et al. 2019), and characterizing genetic variability (Anis et al. 2018b; Giacinti et al. 2022). Few longitudinal serologic studies have been conducted in wildlife. Data from such studies may contribute valuable insights into intrapopulation and intraindividual pathogen dynamics. They also may establish sequence of events and follow change over time in individuals, to identify periods of increased pathogen exposure risk and clinical and physiologic changes before and after exposure.

Canine distemper virus is a morbillivirus closely related to human measles virus (MV): both have similar structure, function, and pathogenesis (da Fontoura Budaszewski and von Messling 2016). Consequently, new findings from MV research may yield important hypotheses for studies concerning the impact of CDV in wildlife. Researchers have used longitudinal studies of human hosts infected with MV to assess the occurrence of a phenomenon known as immune amnesia (de Vries et al. 2012; de Vries and de Swart 2014; Mina et al. 2015, 2019). Researchers detected changes in the qualitative composition of immune cell populations before and after recovery from MV infections, specifically a targeted decrease in immune memory cells (de Vries et al. 2012; de Vries and de Swart 2014; Mina et al. 2015, 2019). This may lead to reduced immunity, including decreases in antibodies against previously encountered pathogens. Similar mechanisms have been postulated for other morbilliviruses (Pfeffermann et al. 2018), but have not been explored for CDV in wildlife.

We used longitudinal data to 1) report seasonal CDV seroprevalence from wild raccoons captured on more than one occasion over a period of 2 yr; 2) describe how CDV titers change over time within individual raccoons; and 3) explore whether exposure to CDV in raccoons is associated with a reduction in preexisting parvovirus antibodies, using paired serum samples from individuals before and after CDV exposure.

We used serum samples collected from raccoons captured on more than one occasion during a live-trapping study carried out from May to November 2011–2013. Sympatric species captured, but not included, in our study were striped skunks (Mephitis mephitis) and opossums (Didelphis virginiana). Trapping and processing procedures were approved by the Animal Care Committee, University of Guelph, Guelph, Ontario Canada (University of Guelph Animal Use Protocol no. 1482), following guidelines of the Canadian Committee on Animal Care, and have been described previously (Bondo et al. 2016; Schulte-Hostedde et al. 2018). In brief, trapping was carried out at 10 locations (Fig. 1) once every 5 wk over the study period. At each location, 20–40 Tomahawk live traps (Tomahawk Live Trap Co., Tomahawk, Wisconsin, USA) were set three to four nights per week. Captured individuals were anesthetized using an intramuscular injection of 0.025 mg/kg dexmedetomidine hydrochloride (Dexdomitor 0.5 mg/mL, Pfizer Animal Health, Kirkland, Quebec, Canada) and 5 mg/kg ketamine hydrochloride (Vetalar 100 mg/mL, Bioniche Animal Health, Belleville, Ontario, Canada), and a passive integrated transponder tag was injected subcutaneously for subsequent identification. Sex and age were recorded for each animal at the time of capture. Age was recorded as adult or juvenile (birth year) based on animal size and tooth wear and staining (Grau et al. 1970). Captured individuals were assessed for any abnormal behavioral or clinical signs in the trap before anesthesia and also during handling. Behavioral and clinical observations were recorded for each individual. Blood (<5 mL) was collected from the jugular vein into a vacutainer with no additive (Becton Dickinson, Mississauga, Ontario, Canada). After sample collection, raccoons were given 0.25 mg/kg atipamezole (Antisedan 5 mg/mL, Pfizer Animal Health) intramuscularly to reverse the dexmedetomidine and then placed back into traps to recover before being released at their capture location. Individual raccoons were sampled once per trapping session (i.e., 5-wk period). If an individual animal was recaptured in a subsequent trapping session, sampling was repeated. Individual animals were not captured at more than one location.

In the field, samples were stored in a cooler with freezer packs. After transport to the laboratory, blood samples were spun down to extract serum within 12 h. Serum samples were stored frozen at –80 C until laboratory analysis.

Virus neutralization (VN) for the detection of antibodies against CDV, carried out at the Animal Health Laboratory (AHL), University of Guelph, was performed on all available serum samples collected from raccoons captured on more than one occasion. Methods used for VN are provided in the Supplementary Material. Based on the results of CDV VN, the following two subsets of individuals were identified for inclusion in the CDV and canine parvovirus (CPV) study (i.e., objective 3).

Individual raccoons with a change in CDV titer from nondetectable to seropositive during the study period: A nondetectable CDV titer was defined as a titer <2, a detectable CDV titer was defined as ≥2, and a positive CDV titer was defined as >8 (Junge et al. 2007). For ease of reference, a change in CDV titer from nondetectable at one time point to seropositive at a subsequent capture was referred to as CDV exposure. It is acknowledged that individuals with a detectable titer may have been reexposed to CDV during the study period; however, these individuals were not considered for the purposes of analysis related to objective 3. Individuals were excluded if a positive CDV titer preceded the nondetectable titer or if there was insufficient serum remaining at one or both timepoints. The serum sample selected for CPV serology was obtained at the first capture timepoint where the individual had a positive CDV titer (i.e., timepoint 2) and from the most recent preceding capture for which sufficient serum was available (i.e., timepoint 1).

Individual raccoons that had nondetectable CDV titers at all captures: These were divided into approximate time interval and location cohorts. We aimed to randomly select individuals from each cohort, but were limited to individuals that had sufficient serum remaining from two capture timepoints. Serum samples from timepoints 1 and 2 from these individuals were submitted to the AHL for hemagglutination inhibition (HI) testing to detect antibodies against CPV-2. Methods used for HI are provided in the Supplementary Material.

For VN and HI tests, samples from the same individual were run concurrently using the same cell lines, and the results of both tests were interpreted and recorded by the same technician. Parvovirus was selected as an example pathogen and as a preliminary step to explore the association between exposure to CDV and change in preexisting antibody titers. Most raccoons were expected to have been exposed and therefore to have preexisting antibodies against CPV (Rainwater et al. 2017).

Month of capture was divided into two categories: May–July and August–November to correspond with the rearing and predenningdispersal periods, respectively, described by Rosatte et al. (2010). STATA 15.0 (StataCorp LLC, College Station, Texas, USA) and a significance level of 5% (α=0.05) were used for all analyses. Additional model building and model diagnostic methods are provided in the Supplementary Material.

Univariable mixed logistic regression was used to explore the association between proportion of raccoon samples seropositive for CDV and season modeled as a dichotomous variable (i.e., May–July and August–November) as the primary independent variable of interest. The following independent variables were also explored for association with the outcome: year (2011, 2012, or 2013), age (adult or juvenile), and sex (male or female). Nested random intercepts were included to account for autocorrelation among samples from the same location and individual at different timepoints. Variables were considered for multivariable analysis by using a liberal significance level (α=0.2).

Mixed multivariable logistic regression models were fitted including all variables that were significant on univariable analysis; season was included in all models, regardless of significance, because it was the primary variable of interest. Nested random intercepts were included as described earlier. Interactions were evaluated between all statistically significant variables for which covariate patterns contained a minimum of 10 samples. The final model included variables that were either statistically significant, part of a significant interaction, or acted as explanatory antecedents or distorter variables (i.e., nonintervening variables whose removal changed the coefficient of a significant variable by >30%).

Two different statistical models were used to explore the association between change in CPV titer and exposure to CDV. The first model considered change in CPV titer as the number of dilutions between timepoints 1 and 2 modeled as a continuous variable (e.g., –1,0,1). The second model considered change in CPV titer between timepoints 1 and 2 as a dichotomous outcome (titer decrease versus titer increase or no change). Univariable models (linear and exact logistic, respectively) were fitted to initially explore the association between the change in CPV titer for each individual between timepoints 1 and 2 and the following independent variables: CDV exposure between timepoints 1 and 2 modeled (yes or no) as the primary independent variable of interest; time interval between timepoints 1 and 2 (i.e., <100 d, ≥100 but <365 d, or ≥365 d); parvovirus titer at timepoint 1 (i.e., <1,024 or ≥1,024); age at timepoint 1 (adult or juvenile); and sex (male or female). Trapping location was not evaluated in either of the above-mentioned models due to limited sample size. In both cases, variables were retained for multivariable analysis by using a liberal significance level (i.e., α=0.2).

Multivariable models (linear and exact logistic, respectively) were fitted including all variables that were significant on univariable analysis except for CDV exposure, which was included regardless of significance as the primary variable of interest. The final model included variables that were either statistically significant or acted as explanatory antecedents or distorter variables. Because of the limited sample size, interactions were not evaluated.

Overall, 620 serum samples were collected from 235 raccoons trapped on two or more occasions between May 2011 and November 2013 (Table 1 and Fig. 1). These included 140 individuals trapped on two occasions, 68 on three occasions, and 27 on more than three occasions. Overall, 154 (67%) were adults, 67 (29%) were juveniles, and 14 (6%) were captured initially as juveniles and later as adults (Table 1).

Detectable CDV antibodies were found in 222 (36%) of 620 samples (95% confidence interval [CI], 32–40), and of these, 156 were classified as seropositive. Of the 222 samples with detectable CDV antibodies, 49 were obtained from juveniles. Overall, 119 (51%) of 235 (95% CI, 44–57) raccoons had detectable antibodies to CDV on one or more occasions across all years, and of these, 84 were seropositive on at least one occasion. Clinical comments, largely consistent with trauma-related incidents (e.g., missing tail), were noted for 13 (15%) of 84 seropositive individuals, including 9 of 84 at the time they were seropositive. Additional descriptive data, including stratification by age, are available in Table 1.

The highest proportion of detectable titers occurred between May and July and in 2013 (Table 1). The proportion of samples with detectable CDV titers ranged from 22% in 2012 to 63% in 2013 (11 and 52% were seropositive, respectively; Table 1). Seasonal proportions ranged from 44% in May–July to 27% in August–November (34 and 15% were seropositive, respectively; Table 1). Based on univariable analysis, the following variables had statistically significant associations with CDV seropositivity: year, season, and age (see Supplementary Material Table S1). The final multivariable model included year, season, age, and the interaction between season and age. Based on this model, juvenile raccoons were less likely to be seropositive for CDV if sampled in August–November than if sampled in May–July (Tables 2 and 3). There were no significant differences in CDV seropositivity between seasons in adult raccoons (Table 3). Juveniles sampled between August and November were less likely to be seropositive for CDV than adult raccoons sampled between August and November (Tables 2 and 3). There were no significant differences in CDV seropositivity between adults and juveniles sampled in May–July (Table 3). Based on the variance components, capture location, animal, and sample levels accounted for 4.3, 81.3, and 14.3% of the variance in seropositivity, respectively (Table 2).

Canine distemper virus titers at each capture from all individuals with at least one detectable titer are available in Supplementary Material Table S2. Detectable titers ranged from 2–768 in adults (median=32) and 2–192 in juveniles (median=8). The yearly median titer (range, 2–768) was highest (32) in 2013.

Of 22 instances of a change in CDV titer from nondetectable to seropositive, all but 2 occurred in adults, mainly over the non-sampling winter months between 2012 and 2013 (Table 4). Of eight individuals with evidence of CDV exposure at a 2013 capture and that were caught in a subsequent month, five showed increased titers; a single juvenile became seropositive between July and August 2012 (Table 4).

Of the individuals seropositive at one capture, 21 then had a nondetectable titer on a subsequent capture (Table 5). Of these, 16 occurred in juveniles between July or August and September within the same year (Table 5). Another juvenile recorded as seropositive in 2011 (titer 32) was subsequently found to have a nondetectable titer as an adult in 2012 and recorded as seropositive (titer 64) again in 2013 (Table 5). Three adults showed a change in CDV serostatus from positive to nondetectable (Table 5).

Paired samples from 17 individuals (11 female, 6 male; 1 juvenile, 15 adult, and 1 juvenile at first capture and adult at subsequent capture [J–A]) with evidence of CDV exposure were tested for antibodies to CPV. Similarly, paired samples from 23 individuals (13 female, 10 male; 4 juvenile, 18 adult, and 1 J–A) that had nondetectable CDV titers at all captures were tested.

We detected CPV antibodies in 71 (89%) of 80 (95% CI, 80–95) samples. Parvovirus titers ranged from 16 to 2,048. Six individuals (nine samples) did not have detectable CPV titers at one or both timepoints and were excluded from the remaining analyses; all six individuals had been exposed to CDV. Of the 34 remaining individuals, 11 (32%; 95% CI, 17–51) had a decreased CPV titer between timepoints 1 and 2 of one (n=7), two (n=3), or three (n=1) dilutions. Of these individuals with a decreased CPV titer, 8 (73%) of 11 (95% CI, 39–94) individuals had been exposed to CDV (see Supplementary Material Table S3). All nine individuals with an increased CPV titer between timepoints 1 and 2 had negative CDV titers at all captures (see Supplementary Material Table S3). Parvovirus titer increases were of one (n=7) or two (n=2) dilutions. The remaining individuals had no change in CPV titer between timepoints (see Supplementary Material Table S3).

Based on univariable analysis, the following variables had a statistically significant association with CPV titer changes between timepoints 1 and 2 in the linear and exact logistic regression models: CDV exposure and the CPV titer at timepoint 1 (see Supplementary Material Tables S4 and S5). Based on the final multivariable linear regression model, raccoons exposed to CDV between timepoints 1 and 2 had a statistically significant decrease in CPV titer between these timepoints, compared with raccoons with no evidence of CDV exposure (Table 6). Raccoons with a higher CPV titer at timepoint 1 had a statistically significant decrease in CPV titer at timepoint 2 (Table 6). Juvenile raccoons had a statistically significant decrease in CPV titer between timepoints 1 and 2 compared with adults (Table 6).

Based on the final multivariable exact logistic regression model, raccoons exposed to CDV between timepoints 1 and 2 were more likely to have a decrease in CPV titer between these timepoints, compared with raccoons with no evidence of CDV exposure (Table 7). Raccoons with a higher CPV titer at timepoint 1 were also more likely to have a decreased titer at timepoint 2 (Table 7).

Canine distemper is a major cause of mortality for raccoons in Ontario (Giacinti et al. 2021); nevertheless, our findings from raccoons captured on more than one occasion corroborate reports that many wild raccoons survive for weeks to years after exposure (Junge et al. 2007). We were unable to address whether survival after CDV exposure is primarily associated with the development of subclinical versus clinical infection. Typically, respiratory and neurologic or behavioral signs accompany clinical CDV infections in raccoons (Pfefferman et al. 2018). In our study, most CDV seropositive individuals were assessed as behaviorally and clinically unremarkable, with multiple individuals recorded as both seropositive and clinically unremarkable at repeated captures over time periods spanning several months to 2 yr. Similarly, in a smaller study, nine CDV seropositive individuals were reported to be clinically healthy at a subsequent capture (Junge et al. 2007). It is important to remember that loss to follow-up is common in longitudinal studies in wildlife due to the dynamic nature of wildlife populations (e.g., animals die, move, or develop trap avoidance). The individuals in our study represent those that were recaptured and may not be representative of CDV dynamics in all seropositive raccoons.

Based on paired titers from individuals exposed to CDV during the study period, and within the limitations of the field season, our results indicate that the majority of CDV exposure occurred sometime between November and April. This time period approximately corresponds to the winter breeding period of raccoons (Rosatte et al. 2010). Similarly, an analysis of the Canadian Wildlife Health Cooperative (CWHC) raccoon carcass surveillance data found an increased odds of CDV infection in the winter breeding season compared with the rearing season in adult raccoons (Giacinti et al. 2021). Based on raccoon carcasses collected in New Jersey, Roscoe (1993) found CDV exposure to be highest during the winter. These findings make biological sense based on the degree of intraspecific contact between raccoons during this period of communal denning and the increase in susceptible juveniles after loss of maternal antibodies (Rosatte 2000; Rosatte et al. 2006).

We found that juvenile raccoon samples were significantly more likely to be seronegative from August to November compared with samples from May to July. Indeed, several seropositive juvenile individuals trapped from May to July had nondetectable titers on subsequent captures in August and September. Raccoons typically have young between April and May (Rosatte et al. 2006); therefore, these data are consistent with a loss of maternal antibodies, which is expected to occur by 20 wk of age (Paré et al. 1999).

Adult raccoon CDV titers also fluctuated over time. Notably, and inconsistent with the commonly accepted notion that immunity against CDV is lifelong (Appel and Summers 1995; Martella et al. 2006; Greene 2013), three adult raccoons that were CDV seropositive on one capture were found to have nondetectable titers from 1 mo to 1 yr later. These animals were reported to be clinically and behaviorally unremarkable, with little to no weight fluctuations. A similar phenomenon was reported by Junge et al. (2007), who found that 33.3% of a small sample of subadult or older raccoons that were CDV seropositive at timepoint 1 were seronegative at timepoint 2. Acknowledging that the three individuals reported in our study represent a small percentage of the individuals sampled, these findings nevertheless warrant attention. Most studies of the duration of CDV immunity concerned domestic dogs; the present results provide raccoon-specific insights and highlight the importance of studying CDV dynamics across hosts to understand differences between host-specific immunologic and physiologic responses to exposure. Although reexposure to CDV might explain an increase in titer, we were unable to track such occurrences.

An early and substantial humoral response is known to be a critical component of host defenses against CDV, and the neutralizing antibody titer has been described as the “best indicator of protection against infection” (Greene 2013). However, cell-mediated immunity may be important for initial protection in individuals with low CDV titers at the time of challenge (Sadler et al. 2016; Ortiz et al. 2018). The titer suggested to be protective varies within and between species (Perrone et al. 2010; Gray et al. 2012). Schultz (2006) posited that the presence of a detectable antibody titer is more important than its magnitude for determining immune response after core vaccination (including against CDV) in adult domestic dogs. Host immune response to CDV may also differ based on the CDV variant to which individuals were exposed (Anis et al. 2018a). In Ontario, there are multiple cocirculating CDV variants in raccoons (Giacinti et al. 2022). Because our study detected antibodies, not virus, we were unable to determine CDV infection status or identify the CDV variant to which individuals were exposed.

Although this was a preliminary investigation, it is noteworthy that CDV exposure was associated with a decrease in CPV titer. This result aligns with recent MV literature that has provided mechanistic insights into the longer term immune suppression observed in humans after MV infection (de Vries et al. 2012; de Vries and de Swart 2014; Mina et al. 2019). Mina et al. (2019), studying unvaccinated children, found a reduction in their antibody repertoire of 11–73% after natural infection with MV. The substantial loss of preexisting immune memory cells, termed immunologic amnesia, may cause a renewed susceptibility to previously encountered pathogens and has been implicated in childhood morbidity and mortality years after initial MV infection (de Vries and de Swart 2014; Mina et al. 2019). Recovery of immune memory populations may require restimulation through subsequent antigen exposure (de Vries and de Swart 2014; Mina et al. 2019). Interestingly, a similar loss of antibody repertoire was not found in MV vaccinated individuals (Mina et al. 2019). The results of our limited study are not sufficient to confirm the presence of CDV-induced immune amnesia; thus, an alternate hypothesis (e.g., immune modulation for other host and/or environmental reasons) also must be considered for the observed association between CDV exposure and change in CPV titer. However, the results of the present study do raise interesting questions and highlight the need for further research into this phenomenon. If present, it would be important to understand the impacts of a reduced population immunity secondary to CDV exposure in raccoons and other wildlife particularly in the event of increased or changing viral activity, and any potential management implications for pathogen control in raccoon populations. This is particularly relevant for rabies virus where, in Ontario, there are active control operations centered around building humoral immunity against rabies virus through oral vaccine baiting in wildlife vector species (e.g., raccoons; Ministry of Northern Development, Mines, Natural Resources and Forestry 2022). Although there are limitations associated with the current preliminary analysis, namely related to the use of observational data, we attempted to control for potential confounding factors (e.g., time interval) where possible. We were unable to confirm the occurrence of reexposure to CPV during the study, but we would only expect this to bias our results away from the null if the risk of reexposure was different in CDV-exposed versus unexposed individuals. Although we attempted to maximize the reliability of test results as described in the Methods, test error must be considered. At worst, we expect any misclassification bias to be nondifferential and only result in underestimation of the true effect.

Further research is needed to investigate the possible occurrence of CDV-induced immune amnesia in wild raccoon populations, considering antibodies to a broader range of pathogens and, where possible, using data collected in a more controlled setting. Future research would benefit from an assessment of the clinical relevance of any observed titer decreases. In particular, future studies should consider whether and to what extent CDV-induced immune amnesia might impact rabies-specific humoral immunity generated through oral baiting programs and the potential value of co administering CDV vaccines as a mitigation strategy.

J.A.G. was funded by the Natural Sciences and Engineering Research Council, Ontario Veterinary College and the University of Guelph. Additional project funding was received from Pet Trust at the Ontario Veterinary College. Erin Harkness, Samantha Allen, CWHC staff, pathologists, summer students, and Animal Health Laboratory virology staff collected and/or assisted with the processing of samples for this study.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-22-00052.