The results of this work demonstrated that the red-legged partridge is susceptible to WNV disease, and provides further evidence that some European wild bird species can be clinically affected by WNV. This concept is supported by recent reports describing field WNV cases in different wild bird species in Europe. While previous work reported three types of European eagles (golden eagle -Aquila chrysaetos-, Spanish imperial eagle -Aquila adalberti- and Bonnelli's eagle -Hieraaetus fasciatus-) clinically affected by WNV in Spain [6, 7], the disease was also described in goshawks (Accipiter gentilis), sparrow hawks (Accipiter nisus) and geese (Anser anser) in Hungary , sparrow hawks in Austria , and house sparrows (Passer domesticus) and magpies (Pica pica) in the Camargue region of France . Just prior to the arrival of WNV to North America in 1999, migrating storks and domestic geese were adversely affected by WNV infection in Israel . Subsequent experimental studies on domestic geese conducted in Israel confirmed their susceptibility to WNV disease . Nevertheless, to our knowledge, the present study is the first reporting an experimental infection with WNV in a European wild bird showing symptoms of disease.
In addition, our results demonstrate experimentally that at least some Euro-Mediterranean WNV strains are pathogenic in an indigenous wild bird species. Previous results showed that one of the strains used in this experiment (Morocco/2003) was as pathogenic for mice as the NY99 prototype strain . Thus, the apparently limited wild bird morbidity caused by WNV in Europe, as compared to the high virulence for WNV in wild birds in North America, demands an explanation accounting for the fact, reflected in this work, that at least some Euro-Mediterranean WNV strains are highly pathogenic for certain indigenous wild birds such as the red-legged partridge. In the European context the apparently mild consequences of WNV disease for wild birds would likely arise from the difficulties inherent to the surveillance of mortality events in the affected wild bird species rather than from low pathogenicity of local WNV strains, or from low susceptibility of indigenous wild birds.
In this work, contact partridges did not show disease symptoms, and remained negative both for virus detection in blood and/or tissues, as well as for WNV-specific antibodies. Thus, direct WNV transmission did not occur in the experimental conditions used in this assay, despite close interaction between inoculated and non-inoculated partridges. While contact transmission has been demonstrated for certain avian species , evidence of this is lacking for other wild bird species. For instance, in an exhaustive study evaluating WNV transmission dynamics in 25 different avian species, contact (cage-mate) transmission was observed in only four species (Blue Jays, Black-billed Magpies, American Crows and Ring-billed Gulls). Interestingly, contact transmission was not observed in the two species of Galliformes evaluated (Japanese Quail and Northern Bobwhite) . In a different study addressing long-term immunity in House Sparrows as important amplifying hosts for WNV, contact transmission was not observed in this species . Viral shedding by the oral and faecal routes occurred in all the inoculated partridges monitored, so the absence of contact transmission in this model needs further investigation, in particular if this finding is attributable to the avian model, or to the viral isolates used.
New animal models of WNV infection are of interest not only for virulence studies but also for antiviral and vaccine development . This work demonstrates that the red-legged partridge constitutes a suitable avian model of WNV infection, since this bird species is not only susceptible to infection and disease but is also easy to obtain, can be raised in captivity, and is relatively easy to handle. In addition, our data demonstrate that the red-legged partridge could act as amplifying host for WNV, as shown by the maximum viremia levels attained during the experiment (approximately 107.0 pfu/mL blood), clearly above the established threshold level of infectious viremia, estimated at 105.0 pfu/mL of blood .
The red-legged partridge is indigenous to the Mediterranean region, where it constitutes a key prey species for many predators. Those predators feeding on infected partridges or carrion eaters feeding on their carcasses could be at risk of infection since non-vectored transmission of WNV through ingested infected food has been documented [10, 31, 32]. In Spain, some endangered species such as the Golden eagle, the Bonnelli's eagle and the Spanish Imperial eagle prey regularly on red-legged partridges and have been shown to be susceptible to WNV-associated disease [6, 7]. Thus the occurrence of WNV outbreaks in red-legged partridges could potentially constitute a risk for the conservation of these endangered species.
The red-legged partridge also represents an important small game species being intensively raised in open air facilities and released into the wild for restocking and hunting purposes. The species also exists in the South of the UK and the Canary Islands, where there is concern about the introduction and spread of WNV. A closely related partridge species (Chukar partridge) is also susceptible to WNV-associated disease in North America, where farms raising these wild birds have suffered outbreaks of the disease . Besides their economic importance, partridge farms could play a relevant role for public health since in case of a hypothetical WNV outbreak, and provided the presence of suitable mosquito species with a broad host spectrum, farmed red-legged partridges could potentially act as the reservoir for accidental transmission of WNV into humans and horses. Suitable mosquito vectors as Culex pipiens are abundant in the red-legged partridges' natural habitat as well as in the farms (unpublished observation). Thus, in surveillance programs, WNV disease should be considered as a potential hazard for farms producing red-legged partridges. This is of particular importance bearing in mind that eventually these captive-raised partridges are to be released into the wild.
In the experimental work described here, the viral strains (Morocco/2003, and Spain/2007) used to inoculate differed in 13 out of 3433 amino acid residues (99.62% amino acid identity), one of them being NS3249. This site appears to play a role in the genetic modulation of pathogenicity for wild birds, as substitution of threonine (T) by proline (P) in a moderately pathogenic WNV strain (Ken98) resulted in an increase in virulence for American crows, whereas, the opposite change (P249T) resulted in a decrease in virulence for the highly pathogenic NY99 strain in the same hosts . This site is occupied by T in Morocco/2003, and by P in Spain/2007. We showed recently in a mouse model that Morocco/2003 is more pathogenic than Spain/2007 . Here we confirm and extend this observation to a wild bird species. This suggests that a proline residue in position 249 of the NS3 protein is not sufficient to enhance virulence for any given WNV strain. Possibly, other residues that differ between both strains are also playing a role in the observed pathogenicity, and could have produced apparent discrepancies with previous studies using infectious clones in American crows [22, 33]. The NS3249-P genotype has arisen recently in the Western Mediterranean region. This WNV genotype has been involved in recent outbreaks in northeastern Italy , in which WNV strains have been similar to those used in the present study . An assessment of the pathogenicity of these recent Italian isolates would provide further insight into the epidemiology of the NS3249 genotype in the Western Mediterranean area.
The pathogenicity of the Spanish WNV isolate was less than that of another Western Mediterranean strain, i.e., Morocco/2003, not only in a mouse model, but also in the red-legged partridge. This finding could explain why this disease has remained essentially silent in Spain to date, and provide a possible mechanism accounting for the occurrence of sporadic WNV outbreaks interspersed by silent WNV circulation in some regions such as the Western Mediterranean. Small genetic changes in the virus genome (T249P at NS3, or others), could substantially reduce the pathogenicity of circulating viruses, and vice-versa. However, other factors not linked to the virus, but to host susceptibility and vectors, could also contribute to this epidemiological pattern, the relative importance of which merits further investigations.