4. Discussion
This study provides strong evidences for the exposure of Iberian hare populations to MYXV or other virus antigenically similar to MYXV more than two decades before the occurrence of the first reported outbreak, and thus new insights into the emergence of a new natural recombinant, and highly pathogenic, ha-MYXV strain in Iberian hares in mid-2018 (Bocanegra et al., 2019; Dalton et al., 2019; Pinto et al., 2019; Carvalho et al., 2020).
To assess the earlier circulation of MYXV or other MYXV-like viruses and contact with Iberian hare populations we used and validated a commercial indirect ELISA originally developed to detect rabbit antibodies that recognise MYXV antigens to detect antibodies in hares, using protein G as conjugate. Whenever possible samples were investigated for MYXV-DNA. Although rabbits and hares belong to the same taxonomic family Leporidae, the relative affinity of protein G towards Iberian hare’s immunoglobulins (Ig) has not been estimated and compared to the affinity towards rabbit Ig. In fact, there is a significant difference in the G protein affinity towards Ig from different species (Biolabs, 2020). Furthermore, the iELISA includes rabbit positive and negative control sera. Since a dissimilar affinity to the Iberian hare Ig could affect the performance and outcome of the iELISA, the adequacy of the test for hare sera was evaluated and the cut-off for positivity was firstly estimated.
It was then necessary to ensure that the hares testing positive had been truly exposed to the myxoma virus or an antigenically similar virus. We thus aimed for a cut-off maximising the positive predictive value of the serological test (iELISA). The qPCR results for ha-MYXV in healthy hares and myxoma-affected hares collected after the 2018 emergence, were used to assign the “infected/non-infected” reference categories. While molecular tests detect viral DNA, which is assumed to mean infection, the serological tests reveal the presence of specific antibodies, which last longer than viremia, as assessed in European rabbits (Kerr, 2012). The mismatch between the timeframes of viral positivity (earlier and shorter) and antibody positivity (later and long-lasting) may therefore have introduced some noise in our data.
The selected cut-off was further supported by finite mixture models, which do not assume a reference test. Under the assumption of normality of the log transformed IR10 of the seronegative subset of our sample, 99.9% of the estimated seronegative population would be correctly assigned by the iELISA, using the selected cut-off. The iELISA with a cut-off RI10=6.1 yielded a good performance, as shown by an estimated positive predictive value of 96.4% (CI95% 82.6-98.0%) and a specificity of 99.1% (CI95% 95.2-100%). The sensitivity of the test was moderate (52.9%, CI95%38.5-67.1%), which is due to most (55.8%) of the qPCR positive Iberian hares not having detectable antibodies recognising MYXV antigens. Therefore, the overall seroprevalence is probably slightly underestimated. Hares infected with MYXV but without detectable antibodies may represent the initial stage of infection or a hyperacute myxomatosis presentation (Carvalho et al., 2020; García-Bocanegra et al., 2019). Either way, the iELISA allows a high degree of certainty in establishing that a hare with detectable antibodies had contacted and become infected by virus identical or antigenically similar to MYXY.
Given the large scale of the current outbreak of myxomatosis in Iberian hares, it is not surprising that the seroprevalence in randomly collected samples was higher in 2018-2019, when compared to 1994-1999 (Fig. 2). In fact, the GLM identified 2019 as the year having significantly higher seroprevalence (21.5%, CI95%13.3-33.0%) (Fig. 2). Interestingly, the year 1999 showed non-significant statistically elevated values of seroprevalence (10.0%, CI95% 4.4-21.4%), not associated with any known reported case of mortality. This lack of reporting could be due to deficient sanitary surveillance in Iberian hare during the 1990’s (Leighton et al., 1995). Sanitary surveillance in Europe, and particularly in the Iberian Peninsula, has been improving in the last decades, despite still far from an optimized system.
Despite being recognised that some hare species other than Lepus granatensis are susceptible to myxomatosis (Magallon et al., 1953; Jacotot et al., 1954; Barlow et al., 2014), the scarcity of scientific evidence regarding the putative infection of Iberian hare by MYXV give room to different explanations. Iberian hares have for long been considered refractory to infection by the MYXV strains, contrasting with European rabbit where a severe disease usually develops (Fig. 5A) (Wibbelt and Frolich, 2005). Others have suggested that the Iberian hare could be naturally resistant to the infection (Fig. 5C) (Fenner and Fantini, 1999). However, no serological evidences were provided to support or exclude any of these theories.
Our results corroborate the serology data published very recently (García-Bocanegra et al., 2020) and report for the first time the presence of antibodies recognising MYXV antigens in Iberian hare populations since at least 1996, supporting the theory of enduring circulation of MYXV, or a virus antigenically similar to MYXV, in this host species, possibly in unapparent forms of disease that justify the lack of published reports. In infected domestic rabbits, IgM can be detected 5–6 days post-infection and during the following 30 to 40 days, whereas IgG peaks at day 20 to 30 and can be detected up to 2 years (OIE Terrestrial Manual, 2018). This hypothesis is further reinforced by the detection of myxoma virus-specific DNA (gene M005) that does not differentiate from ha-MYXV.
The relatively high seropositivity in 1999 suggests that, in this year, the sampled hare populations had eventually more contact with a virus antigenically identical or similar to MYXV, which could have been a virulent MYXV strain grade I or II. In certain conditions, such as immunosuppression induced by pathogens, as the recently described gammaherpesvirus LeHV-5 (Abade dos Santos et al., 2020), hares may eventually become more susceptible to rabbit MYXV (Fig. 5B). Another possibility is that ha-MYXV was already circulating unnoticed in Iberian hares (Fig. 5E). Also, the recombination event that generated ha-MYXV may have involved an intermediate host (Fig. 5C), similarly to many emerging viruses (e.g. SARS-CoV1, MERS-CoV, SRAS-CoV-2). Rather than a recent species jump from rabbit to hare, our results support the possibility of a genetic change in a MYXV strain that was already circulating unnoticed in the hare populations (Fig. 5E). An increased virulence of this strain would lead to the recent outbreaks, as previously suggested (Dalton et al., 2019; Pinto et al., 2019). The possibility that ha-MYXV originated from a distinct hare MYXV that was already circulating unnoticed in wild hare populations for decades under subclinical forms of disease, would explain myxomatosis never being reported before in hares (Fig. 5F).
The higher seroprevalence in our sample during the spring/summer suggests seasonal transmission like that of MYXV in European rabbits, possibly also mediated by arthropod vectors (García-Bocanegra et al., 2010). The seroprevalence found in the Iberian hare is much lower than the seroprevalence reported in European rabbits (Ferreira et al., 2009; Boag et al., 2013), which can be justified by the high fatality rate of ha-MYXV infection in hares (Bocanegra et al., 2019). Also, the Iberian hare is a more solitary species than the gregarious European rabbit, attaining lower densities (Gortazar et al., 2007), which could hinder MYXV transmission (Villafuerte et al., 2017b). Also, the reproductive characteristics of hares, with less and smaller litters compared to the European rabbit (Alves et al., 2002; Gonçalves et al., 2002) introduce fewer susceptible juveniles to the population, which may lead to greater containment of outbreaks due to the existence of fewer susceptible animals (García-Bocanegra et al., 2010).
The case fatality rate of the ha-MYXV infection in Iberian hares in Spain has been estimated at 69.2% (García-Bocanegra et al., 2019). We detected a significant proportion of the hares as qPCR-negative but antibody-positive. Given the lack of skin lesions, liver and lungs were also used to investigate MYXV-DNA in apparently healthy hare tissues. However, since low titres are generally observed in these organs (liver, spleen lungs), even during active MYXV infection (Kerr et al., 2015), one cannot exclude that some of the MYXV qPCR negative results may be false negatives at initial stages of infection, before clinical signs become apparent.
The detection of viral MYXV-DNA in four serum samples from 1997 collected in Alcochete, Portugal, strongly corroborates the conclusions based in serological data.
In conclusion, despite the origin of the ha-MYXV still being unclear, our study shows the long-term circulation of virus antigenically similar to MYXV in Iberian hare populations and support the hypothesis of a recent increase in virulence of MYXV strain or antigenically similar strains already circulating in hares, that could have caused the ongoing outbreak. Furthermore, we describe and evaluate the performance of a commercial serological ELISA to detect MYXV antibodies in hare sera, which proved reliable and may contribute to extend the knowledge on the epidemiology of myxomatosis in hares. Our results also highlight the need to continue and improve sanitary surveillance of this neglected game species.