Detection of anti-ebolavirus antibodies in Ghanaian pigs

Abstract

Some filoviruses such as ebolaviruses and marburgviruses, cause hemorrhagic fever in humans and nonhuman primates. Pigs are suggested to play a potential role in the filovirus ecology. We investigated the seroprevalence of filovirus infection in pigs in Ghana. Using a viral glycoprotein (GP)-based enzyme-linked immunosorbent assay, we detected filovirus-specific immunoglobulin G antibodies in 5 of 139 samples. These positive sera showed specificities to four different filovirus species. Particularly, two of the positive sera reacted to GPs of two African ebolaviruses (i.e., Ebola virus and Taï Forest virus) in Western blotting. Our results suggest that these Ghanaian pigs were exposed to multiple filoviruses and emphasize the importance of continuous monitoring of filovirus infection in pig populations in West African countries.

Ebolaviruses and Marburgviruses belonging to the family Filoviridae cause severe hemorrhagic fever in humans and nonhuman primates. The genus Ebolavirus consists of six species represented by Ebola virus (EBOV), Sudan virus (SUDV), Taï Forest virus (TAFV), Bundibugyo virus (BDBV), Reston virus (RESTV), and Bombali virus. The genus Marburgvirus consists of a single species including two viruses: Marburg virus (MARV) and Ravn virus [11]. The genus Cuevavirus is another member of the Filoviridae family [17] and consists of a single species represented by Lloviu virus (LLOV) discovered in insectivorous bats (Miniopterus schreibersii) in Europe.

Outbreaks of filovirus infection in humans, i.e., Ebola virus disease (EVD) and Marburg virus disease, frequently occur in Central Africa. EVD was also reported in West Africa. The only EVD case caused by TAFV was reported in 1994 in Côte d’Ivoire [13]. The 2014–2016 EBOV outbreak in West African countries (e.g., Guinea, Liberia, and Sierra Leone) became the largest and most complex outbreak [3]. The epidemic origins of these two West African ebolaviruses are unknown. Although some bat species are suspected as natural hosts, the ecology of ebolaviruses is largely unknown [1]. Several studies suggest that pigs may act as an intermediate or amplifying host in the ebolavirus ecology: RESTV was isolated from pigs in the Philippines and subsequently discovered in China [4, 20] and pigs have been experimentally shown to be susceptible to EBOV and RESTV infection [10, 14]. Furthermore, it was shown that one percent (4/400) of pigs in Sierra Leone were EBOV-seropositive [7], and two of these serum samples were cross-reactive to other ebolaviruses (RESTV or SUDV).

Here, we investigated the seroprevalence of filovirus infection of pigs in Ghana. Although no human case of EVD or Marburg virus disease has been reported in Ghana, there may be a risk due to wild animal and human cross-border movements from the adjacent affected country (i.e., Côte d’Ivoire). This study was conducted under the agreement between the University of Ghana and Gifu University, which was reviewed by the Gifu University Animal Care Committee on behalf of the two institutions (Approval No. 17070) and was approved by the College of Basic and Applied Sciences, University of Ghana locally as well. We screened 139 serum samples collected from pigs by veterinarians for the purpose of clinical testing (Ashanti Black and Large White) in three regions (Greater Accra, Upper West, and Upper East) in Ghana in 2014 and 2016 for immunoglobulin G (IgG) antibodies against seven filoviruses in the genera Ebolavirus, Marburgvirus, and Cuevavirus using a viral glycoprotein (GP)-based enzyme-linked immunosorbent assay (ELISA) [5, 6, 16, 18, 19] (Supplementary Table 1, Supplementary Information). We analyzed optical density (OD) values obtained from the 139 samples for seven filovirus GPs concurrently by Smirnov-Grubbs rejection test (n=973, T=4.24; P<0.01) to detect outliers that had significantly higher OD values than the others [19]. ELISA-positive samples were subsequently analyzed by Western blotting using virus-like particles as antigens (Supplementary Information).

Five of 139 serum samples had OD values above the cut-off value and were considered reactive to one of the filovirus GPs: Sample #126 for TAFV, #46 for EBOV, #125 for LLOV, #118 for RESTV and #131 for TAFV (Fig. 1A and 1B, Table 1). None of these samples showed significant cross-reactivity to GPs of multiple filovirus species. The ELISA-positive sera were further analyzed by Western blotting [19]. Samples #126 and #46 reacted to TAFV and EBOV GPs, respectively (Fig. 1C). However, specific GP bands were not visible with the other ELISA-positive sera. Although the #131 serum reacted to approximately 90–104-kD bands corresponding to the ebolavirus nucleoprotein (NP), these were likely non-specific reactions since similar bands were also seen with negative pig (#146) and specific pathogen-free pig serum samples (Fig. 1C). This serum also reacted to another size of bands, including slightly visible approximately 40-kD band corresponding to the matrix protein (VP40) (Supplementary Fig. 1). Similar finding has been shown in another study in which swine serum samples collected in Sierra Leone were evaluated by NP-based ELISA and Western blotting (i.e., ELISA-positive serum reacted to only VP40 in Western blotting) [7]. However, since seroepidemiologic study is limited and nothing is known about antibody response in field pig populations, we were not able to reach a conclusion about this reaction. Since all of 3 samples (#125, #118 and #131) were relatively low OD values as compared to those of samples #126 and #46, specific GP bands might be visible using lower diluted serum (e.g., 1:50).

Fig. 1.

Filovirus species-specific immunoglobulin G (IgG) antibodies detected in glycoprotein (GP)-based enzyme-linked immunosorbent assay (ELISA) (A and B); reactivity of ELISA-positive sera in Western blotting (C). (A) Optical density (OD) values for all of the tested filovirus species (139 sera for seven GPs) with 1–99 percentile whiskers and for each filovirus species. *Significantly higher OD values by the Smirnov-Grubbs rejection test (P<0.01). (B) Filovirus species-specificity of IgG antibodies in nine sera (the top nine OD values including *five significantly positive samples). (C) Sera were tested for reactivity against each filovirus antigen showing the highest OD value. NC: negative control antigens (Supplementary Information). Mouse monoclonal antibodies ZGP42/3.7 [21] and LGP14-2 [15] were used as positive controls for Ebolavirus and Cuevavirus, respectively. Specific pathogen free (SPF) miniature pig and #146 serum samples were used as negative controls.

Table 1. Optical density (OD) values of the five significantly positive pig serum samples.

Sample ID	Year	Location	The highest reactivity	OD value
				EBOV	SUDV	TAFV	BUDV	RESTV	MARV	LLOV
#126	2016	Upper West	TAFV	0.1935	0.2297	3.024	0.182	0.2386	0.1871	0.1436
#46	2014	Upper East	EBOV	1.7548	0.1367	0.0967	0.1241	0.1617	0.1505	0.2359
#125	2016	Upper West	LLOV	0.2045	0.2343	0.2017	0.2028	0.2307	0.214	0.7653
#118	2016	Greater Accra	RESTV	0.1532	0.1536	−0.0348	0.4868	0.7373	0.0324	−0.0532
#131	2016	Upper West	TAFV	0.195	0.2108	0.7136	0.1156	0.2376	0.1535	0.0399

Bold font indicates OD values showing significantly higher values (i.e., outliers). EBOV, Ebola virus; SUDV, Sudan virus; TAFV, Taï Forest virus; BDBV, Bundibugyo virus; RESTV, Reston virus; MARV, Marburg virus; LLOV, Lloviu virus.

Surveillance of filovirus infections of animals in Africa has been conducted for bats, gorillas, chimpanzees, duikers, dogs, and pigs [1, 7], but it mostly focused on EBOV and/or RESTV. Herein, we examined antibodies against seven filoviruses using the GP-based ELISA and detected IgG antibodies specific to EBOV, TAFV, RESTV, and LLOV. Two of these serum samples showing high OD values to TAFV and EBOV reacted to the respective filovirus GPs (i.e., EBOV and TAFV) in Western blotting. It is particularly noteworthy that TAFV-specific IgG was detected in two pigs since only one human case has been reported (through contact with a dead chimpanzee) in Côte d’Ivoire [13], and epidemiological information such the distribution of TAFV is quite limited. Although a Colobus monkey was a suspected source of the TAFV infection [12], it has not been fully proven. It is conceivable that shared feeding spots among animals, including pigs, bats and non-human primates may support the transmission of ebolaviruses [2, 12]. Since Côte d’Ivoire borders Ghana and both countries are thought to share the ecosystem in which provides the domestic pigs-wildlife interface, TAFV might be maintained in some animals in these countries.

Hayman et al. detected EBOV and RESTV antibodies in fruit bat species including the straw-colored fruit bat (Eidolon helvum) in Ghana, suggesting the introduction of both ebolavirus species into this migratory bat population [8]. Although RESTV has been believed to be of Asian origin, another study also demonstrated RESTV-specific IgG in E. helvum migrating from central Africa to Zambia, suggesting the existence of RESTV or RESTV-like virus broadly in Africa [19]. Interestingly, herein we detected IgG reacting to RESTV GP in a pig. Although potential sources of the virus introduced into Ghana’s pig population are unclear, it might be possible that RESTV or a RESTV-like virus is already widely distributed and locally maintained in West Africa including nonendemic countries. This hypothesis is supported by the detection of RESTV-specific IgG in pigs in Sierra Leone [7].

It is also notable that LLOV-specific IgG was detected in the present study, although it was detected so far in only ELISA. LLOV was detected in M. schreibersii in Europe [9, 17]. However, M. schreibersii may not be the natural host of LLOV since this species presents high mortality by LLOV infection [17]. Although it remains unclear whether LLOV causes clinical diseases in other animal species including humans, our data suggest that pigs may be susceptible to LLOV and may play a role in LLOV transmission to other animal species.

Taken together, our results suggest that multiple filovirus species might have been transmitted to pig populations not only in the countries previously affected by EVD but also in distant nonendemic countries such as Ghana.

CONFLICT OF INTEREST

No conflicts of interest declared.