Serological Survey of West Nile Virus in Domestic Animals from Northwest Senegal

West Nile fever is caused by the West Nile virus (WNV), a mosquito-borne Flavivirus. Birds are the natural reservoir of the virus, which is maintained in nature in a mosquito–bird–mosquito transmission cycle. WNV has been detected in many regions worldwide, including North America, Europe, Africa, Near East, Asia, and Australia (Ozdenerol et al. 2013). WNV has been shown to cause meningoencephalitis in humans and horses.

In Africa, infection with WNV is frequent but almost always asymptomatic in humans and equids. Clinical cases in humans and horses have been reported in Morocco, Algeria, Tunisia, Egypt, Sudan, DR Congo, and South Africa. In Senegal, several studies have reported the serological prevalence of WNV in horses (Cabre et al. 2006, Chevalier et al. 2006, Chevalier et al. 2010) and dogs (Davoust et al. 2014). The aim of this study was to identify whether any other domestic animal living in the same enzootic locality may be the sentinel of WNV circulation.

In northwest Senegal, blood samples were collected from 283 adult domestic animals (136 sheep, 64 horses, 29 donkeys, 29 goats, 14 cattle, and 11 dogs), in three localities near Keur Momar Sarr (latitude: 15.916667; longitude: −15.966667; altitude: 8 m.): Gankette Balla (15.980710, −15.928506), Loboudou (15.953050, −15.919963), and Ndimb (16.049155, −16.002941). Keur Momar Sarr is a village of 1,800 inhabitants in the Louga region, located on Lake Guiers, one of the largest freshwater (partially brackish water) lakes in West Africa. The survey was conducted in May 2014 at the end of the dry season. We sampled all the animals belonging to volunteer owners for the epidemiological study.

Sera, after centrifugation, were separated, frozen at −20°C, and sent to the French National Reference Centre for Arbovirus (Marseille, France). Each sample was systematically tested for WNV immunoglobulin G (IgG) using an in-house enzyme-linked immunosorbent assay (ELISA) with inactivated WNV as an antigen (Cabre et al. 2006). The negative antigen was prepared from a supernatant of noninfected cells. Specific binding was demonstrated by using a peroxidase-labeled goat antihorse IgG conjugate (Jackson) at a dilution of 1:10,000 for horse samples, peroxidase-labeled goat antidonkey IgG conjugate (Thermofisher) at a dilution of 1:10,000 for donkey samples, peroxidase-labeled rabbit antisheep IgG conjugate (Jackson) at a dilution of 1:4,000 for sheep samples, peroxidase-labeled rabbit antigoat IgG conjugate (BETHYL) at a dilution of 1:4,000 for goat samples, peroxidase-labeled sheep antibovine IgG conjugate (BETHYL) at a dilution of 1:20,000 for bovine samples, and a peroxidase-labeled sheep antidog IgG conjugate (BETHYL) at a dilution of 1:2,000 for dog samples. Serum samples were considered positive if the optical density at 450 nm was >3-fold the mean of the negative antigen (Cabre et al. 2006). Owing to antigenic cross-reactivity among Flaviviruses, all positive samples were confirmed as true positives by a Western Blot test (Cabre et al. 2006, Mansfield et al. 2011).

The p values and 95% confidence interval were calculated separately for each variable using the Epi Info Software (v5.01; CDC Atlanta). The chi-square test was used to evaluate associations (α = 5%). The differences were considered statistically significant when p≤0.05. The Fisher exact test was used to test for significance when the chi-square test was not appropriate.

Table 1 shows the characteristics of the populations of the study and summarizes the results of the serology. The prevalence of WNV infection among donkeys, horses, dogs, goats, cattle, and sheep was 86.2% (25/29), 68.7% (44/64), 27.3% (3/11), 6.9% (2/29), 0%, and 0%, respectively. Statistically, a very significant difference was established between equids (69/93–74.2%) and small ruminants (2/165–1.2%) (p≤0.00001) and between equids and cattle (p≤0.00001). Also, there is a significant difference between horses and dogs (p≤0.01) and between small ruminants and dogs (p≤0.01). However, no statistically significant difference was observed between horses and donkeys (p≥0.05), dogs and cattle (p≥0.05), and between small ruminants and cattle (p≥0.01). No difference was observed between the three villages (p≥0.05).

OPEN IN VIEWER

Table 1.

Results of Serology (ELISA) for West Nile Virus in Domestic Animals (N = 283) of Three Localities Near Keur Momar Sarr (Northwest Senegal–May 2014)

Localities	Horses	Donkeys	Dogs	Cattle	Sheep	Goats
Gankette Balla	8/12	5/6	1/3	0/4	0/51	0/1
Loboudou	18/30	2/3	1/6	0/7	0/12	2/16
Ndimb	18/22	18/20	1/2	0/3	0/73	0/12
Total	44/64	25/29	3/11	0/14	0/136	2/29
	68.7%	86.2%	27.3%	0%	0%	6.9%
	[CI: 55.9–79.8]	[CI: 68.3–96.1]	[CI: 6–61]	[CI: 0–23.2]	[CI: 0–2.7]	[CI: 0.8–22.8]

ELISA, enzyme-linked immunosorbent assay; CI, confidence interval.

For studies on the equids, the observed seroprevalence (74.2%) was comparable to those reported in Dakar (92% of 25 horses), the Senegal River basin (85% of 367 horses), and the Ferlo area (78.3%) (Cabre et al. 2006, Chevalier et al. 2006, Chevalier et al. 2010). For dogs, our results (27.3%) are intermediate between those obtained in Dakar (18.7% of 16 dogs), Sine Saloum (3% −6/33), and Casamance (3.7% −3/81) (Davoust et al. 2014). The absence of seropositive cattle is comparable to the very low (0.11% −2/2695 cattle) prevalence observed in Croatia (Barbić et al. 2012). In horses, persistence of IgG several years after infection has already been described (Petersen et al. 2013). This ELISA test is particularly valuable for serological surveillance as it is very sensitive and WNV IgG lasts for at least 15 months after infection.

Our results demonstrate the circulation of WNV on peridomestic animals in the villages we studied. These people (Fulani and Wolof) have an extensive livestock in the Sahel region of this dry savannah. Equids and dogs are known sentinels for the risk of human infection (Cabre et al. 2006, Davoust et al. 2014). The Lake Guiers area is an ornithological site where potentially WNV reservoir birds are abundant (Lanius senator, Anthus trivialis, Hippolais opaca, Jynx torquilla, and Cercotrichas galactotes). Several WNV strains have been isolated from birds and mosquitoes in northwest Senegal, the Ferlo Valley, and the Senegal River basin (Chevalier et al. 2009, Diallo et al. 2011, Fall et al. 2014). It has been shown that the environmental factors affecting the distribution of the vector (presence of Culex spp.) are more important in the transmission of WNV than the distribution of birds (Ozdenerol et al. 2013). Serological prevalence in equids is lower in seaside areas and close to brackish waters because of the limitation factors for Culex spp. development (Chevalier et al. 2010).

All three villages are situated on the edge of the big Lake Guiers. Quite often, before 1985, the salt water from the delta of the Senegal River freely entered the lake and the salinity rose up to 1375 mg/L. However, since the construction of two dams (Diama and Manantali), the salinity has decreased significantly and does not exceed 200 mg/L. Increased salinity is one of the natural factors limiting the development of Culex spp. larvae, especially C. quinquefasciatus, one of the main vectors of WNV in Africa (Patrick et al. 2001). The desalinization of Lake Guiers could, probably, also favor the development of WNV vectors, including C. quinquefasciatus and C. neavi (Fall et al. 2011). These mosquitoes feed preferably on horses and birds. Feeding patterns of these mosquito species may help to understand the viral transmission.

This survey confirms that equids and dogs could be the best sentinel animals for surveillance of WNV; other animals (cattle and small ruminants) do not develop the WNV-directed antibodies in enzootic regions. The ruminants do not play a role in viral transmission. In addition to improving our knowledge of the bioecological characteristics that influence transmission, it would be worthwhile to assess the clinical impact of WNV on human and equine populations. The lineage of the WNV circulating near Lake Guiers remains to be characterized.