Serological evidence for Borna disease virus infection in humans, wild rodents and other vertebrates in Finland

1. Introduction 

Borna disease virus (BDV), the causative agent of Borna disease, is a single-stranded RNA virus belonging to the family Bornaviridae, order Mononegavirales (de la Torre et al., 2002). In Central Europe, Borna disease has been endemic for centuries and natural BDV infections have been traditionally recognised in horses and sheep (von Sind 1767, 1781; Durrwald and Ludwig, 1997). More recently, signs of BDV infection have been found in many vertebrate species elsewhere in Europe and around the world (Ikuta et al., 2002) and natural infections have been described also in cats (Berg, 1999, Kamhieh and Flower, 2006), dogs (Weissenbock et al., 1998, Okamoto et al., 2002), shrews (Hilbe et al., 2006) and other vertebrate species (Ikuta et al., 2002). Experimental infections have been carried out in many rodent species such as rats (Hirano et al., 1983, Narayan et al., 1983), mice (Kao et al., 1984, Rubin et al., 1993, Hallensleben et al., 1998), hamsters, gerbils and guinea pigs, but also in rabbits, chickens, tree shrews and rhesus monkeys (Pletnikov et al., 2002). In animals, the infection may induce a usually fatal chronic progressive meningoencephalomyelitis (Ikuta et al., 2002). Various combinations of neurological symptoms are seen, including alterations in behaviour (Rott and Becht, 1995). Mild or asymptomatic infections are also common (Kao et al., 1993, Berg et al., 1999, Vahlenkamp et al., 2002).

In humans, BDV-specific RNA and proteins, and antibodies against BDV have been reported (e.g. Rott et al., 1985, Chalmers et al., 2005). The prevalence has most often been higher in neuropsychiatric patients, but the results have varied considerably between different studies (Planz et al., 2002).

The mode of BDV transmission is unknown (Durrwald et al., 2006). A common reservoir has been suspected (Staeheli et al., 2000, Berg et al., 2001, Durrwald et al., 2006). Vertical transmission has been reported in horses (Hagiwara et al., 2000) and laboratory rodents (Okamoto et al., 2003). Several studies might indicate indirectly a role of wild rodents as a reservoir: (1) horizontal transmission between laboratory rodents (Sauder and Staeheli, 2003), (2) geographic clustering of strains (Kolodziejek et al., 2005), (3) temporal clustering of cases (Rott and Becht, 1995, Durrwald and Ludwig, 1997, Vahlenkamp et al., 2002) and (4) the biggest risk for BDV infection in free-roaming tomcats in the countryside (Berg et al., 1998). Although insectivorous shrews [Soricomorpha; not rodents] were recently suggested as a reservoir (Hilbe et al., 2006), BDV infection has not been previously detected in wild rodents (Rodentia) (Tsujimura et al., 1999, Hagiwara et al., 2001, Hilbe et al., 2006).

BDV infections do not always induce detectable levels of antibodies (Grabner and Fischer, 1991, Katz et al., 1998, Vahlenkamp et al., 2002). Some authors have reported that BDV antibodies are found in circulating immune complexes (Bode et al., 2001), although there has been discussion about the reproducibility of these results (Wolff et al., 2006, Flower and Ludwig, 2006). Concerns about the specificity of serological methods have been raised (Sauder et al., 2002). To improve the specificity of the serological detection of BDV infection, confirmatory peptide array has been introduced (Billich et al., 2002).

Aim of this study was to find out whether BDV infections occur in Finland. We aimed at screening several species, including candidate reservoir species and humans, for BDV antibodies, and at verifying potential positive results with independent methods.

2. Methods 

Please see http://www.hi.helsinki.fi/zoonoosivirukset/bornavirukset/Final-Matmeth for details.

We screened for BDV antibodies 2464 samples from horses, cats, dogs, sheep, cattle, large predators (mainly lynx), grouse, small mammals [mainly rodents, especially bank voles, Myodes (Clethrionomys) glareolus], and 499 human sera (Table 1) by IFA (Fig. 1). Subsets of neurologically or behaviourally symptomatic horse, cat and dog patients were included in the study, and the human samples included veterinarians and patients suspected for hantavirus infection. Antibody titres repeatedly ≥15 were considered positive.

Table 1. BDV-antibodies by BDV/He-80 C6 cell IFA

	Panel	BDV-antibodies
		n (sera)	Positive	Prevalence (%)
	Horse	490	12	2.4
	Neurological/behavioural patients	89	3	3.4
	Symptomless	387	7	1.8
	From same farm with seropositive horse	14	2	14.3
	Cat	283	2	0.7
	Neurological/behavioural patients	30	1	3.3
	Symptomless	253	1	0.4
	Dog	92	1	1.1
	Neurological patients	89	0	0
	From same farm with seropositive horse	3	1	33
	Sheep	22	0	0
	Cattle	52	0	0
	Predators, large	122	0	0
	Bird, grouse	292a	0	0
	Small mammals/wild rodents	1111	3	0.3
	Near seropositive horses	393b	2	0.5
	Elsewhere	718	1	0.1
	Human	499	3	0.6
	Veterinarians	138	1	0.7
	Suspected for hantavirus infection	361	2	0.6

a For birds, dried and later dissolved whole blood was used. b For wild rodents, either serum, blood or pulmonary cavity extract was used.

View Large Image Download to PowerPoint

Fig. 1. Detection of BDV antibodies by immunofluorescence assay based on BDV-infected, acetone-fixed rat astroglial C6-cells (ratio of infected to uninfected control cells 1:3), with the following samples: (a) confirmed Borna-diseased horse from Germany (serum, diluted 1:40). (b) Root vole R1 (lung extract, diluted 1:160). (c) Bank vole R2 (lung extract, diluted 1:10). (d) Bank vole R3 (whole blood, diluted 1:40).

To confirm the specificity of the antibody findings in the screening by BDV-IFA (Table 2), and to characterize them in more detail, especially the wild rodent and human samples, we established the following tests: (1) measurement of the avidity of the BDV-IgG-antibodies (Billich et al., 2002). (2) Detection of antibodies by immunoblotting and IFA using recombinant BDV-N and -P proteins expressed either as GST-fusion proteins in E. coli or using baculovirus expression in insect cells. (The reactivity of control sera with insect cells expressing BDV-P (a) and in negative control cells (b) is shown in Fig. 2.) (3) Epitope mapping using a peptide array with 15- and 20-mers covering the coding regions of BDV-N and -P (Billich et al., 2002). A sample was considered reactive towards an epitope if the signal was clearly stronger than the background reaction of secondary antibodies, and ≥3 consecutive overlapping peptides were reactive (Fig. 3). The analysed IFA-negative sera did not react with the peptides.

Table 2. Further analysis of BDV-IgG-findings

	Sample	Symptoms	Average			Recombinant proteins
			IFA titre	Aviditya	Peptide array	N	P
	Horse
	Ho1	Ataxia, paresis	320	High	1–2 epitopes	Positive	Negative
	Ho2	Fever, ataxia, apathy	80	High	Multiple epitopes	ND	ND
	Ho3	Ataxia, decreased proprioception, hypersensitivity	15	ND	ND	Positive	Negative
	Ho4	None	20	High	ND	Positive	Positive
	Ho5	Ataxia	40	Low	ND	ND	ND
	Ho6	None	40	Low	ND	Negative	Positive
	Ho7	None	320	High	Multiple epitopes	ND	ND
	Ho8	None	160	High	ND	Negative	Positive
	Ho9	None	80	Low	ND	Negative	Positive
	Ho10	None	160	High	ND	ND	ND
	Ho11	None	40	Low	ND	Negative	Positive
	Ho12	None	160	High	ND	Positive	Negative
	Cat
	C1	Aggressivity	20	Low	ND	Negative	Positive
	C2	None	80	High	ND	Negative	Positive
	Dog
	D1	None	15	ND	ND	Positive	Positive
	Rodent
	R1	?	320	High	Multiple epitopes	Negative	Positive
	R2	?	15	ND	ND	ND	ND
	R3	None	80	Low	ND	Positive	Negative
	Human
	Hu1	?	20	High	Multiple epitopes	Negative	Positive
	Hu2	?	20	Low	ND	Negative	Positive
	Hu3	?	40	Low	ND	ND	ND

ND, not determined.

a Avidity was interpreted high, if AIX (=titre with urea wash/titre without urea wash×100%)>25% (Billich et al., 2002).

View Large Image Download to PowerPoint

Fig. 2. Confirmation of BDV antibody findings by immunofluorescence assay based on acetone-fixed Sf9 insect cells expressing BDV-P protein. Antigens in panels (a and c): 10% of cells infected with recombinant baculovirus expressing BDV-P; panels (b and d): non-infected control Sf9 cells. (a and b) Polyclonal anti-P rabbit serum reacts with cells expressing BDV-P protein but not with control cells. (c and d) Antibodies of the bank vole R1 react with cells expressing BDV-P protein but not with control cells.

View Large Image Download to PowerPoint

Fig. 3. Confirmation of BDV antibody findings by peptide arrays consisting of the peptides of BDV-N and -P proteins (amino acids 19–626 of the whole BDV sequence). Detection by electrochemiluminescence (a–d) or alkaline phosphatase (e and f). (a) Antibodies of the horse Ho7 detect multiple BDV-N and -P protein epitopes (serum, 1:750). (b) Control reaction with anti-horse IgG secondary antibody. (c) Antibodies of the vole R1 detect multiple BDV-N and -P protein epitopes (lung extract, dilution 1:100). (d) Antibodies of IFA-negative rodent pool do not detect BDV-N and -P protein epitopes (lung extract, dilution 1:100). (e) Antibodies of the veterinarian Hu1 detect multiple BDV-N protein epitopes (serum, 1:20). (f) Antibodies of the veterinarian Hu1 detect multiple BDV-P protein epitopes (serum, 1:20).

3. Results and discussion 

BDV antibodies were detected in horses, cats and dogs (Table 1). The prevalence of BDV-seropositive animals in the whole horse material was 2.4% (12/490) and in cat material 0.7% (2/283). No significant correlation to the symptom status could be seen. Positive antibody reactions were observed in two asymptomatic horses, a healthy dog and two rodents (see below) near/from two farms housing a symptomatic horse that scored seropositive in our BDV antibody assay. The prevalence in contact horses (14.3%; 2/14) was significantly higher than in other symptomless horses (1.8%; 7/387) (χ2=9.587; p<0.005). A typical immunofluorescence image seen in seropositive horses is shown in Fig. 1a. IFA antibody titres of the seropositive individuals are given in Table 2.

All tested IFA-positive horse and cat samples reacted with one or both recombinant antigens, and 7/11 horses and 1/2 cats had high avidity (index>25%) antibodies (Table 2). Furthermore, all three tested horses recognised several epitopes in BDV-N and -P in peptide array assays (Fig. 3a) differing from background reactions (Fig. 3b).

BDV antibodies were detected in 3/1111 small mammals (Table 1); all of them voles (Rodentia, Arvicolinae). One root/tundra vole (Microtus oeconomus; code R1) had a titre of 320 in BDV-IFA (Fig. 1b, Table 2) and was one of the 291 rodents caught in Northern Finland from the field, where feed was cut for a symptomatic seropositive horse (Ho1 and his stable mate with similar encephalitic symptoms). Two other seropositive wild rodents (R2 and R3; Fig. 1c and d) were bank voles. R2 was caught near seropositive symptomatic horse Ho2 in Western Finland. R3 was an individually marked wild bank vole with a clear positive BDV-IgG-IFA result obtained from all four fresh blood samples collected during a catch-mark-and-recapture study in Central Finland within 4 months. Most negative samples were from larger, geographically scattered monitoring of rodents. The prevalences were higher from neighbourhood of positive horses (0.5%) than elsewhere (0.1%). However, more studies are needed to elucidate possible causal relationships of these geographical findings.

The rodent R1 had high-avidity IgG-antibodies (Table 2), was reactive with recombinant BDV-P by IFA (Fig. 2c and d) and immunoblotting (data not shown), and bound to multiple epitopes of BDV-N and -P in the peptide array (Fig. 3c). In contrast, antibodies of BDV-IgG IFA-negative rodents did not bind specifically to the membrane-bound peptides (Fig. 3d; the secondary antibody bound to three peptides. Such a binding was not observed with other secondary antibodies or IFA-negative sera). Antibodies from the rodent R3 reacted with recombinant N protein confirming the seropositivity (data not shown). Due to small sample volume, confirmatory testing of the positive rodent R2 could not be performed.

BDV antibodies were detected in 3/499 humans (Table 1): 1 veterinarian (Hu1) attending a serological survey in a veterinary meeting and 2 patients suspected to have a Puumala hantavirus infection (Hu2 and Hu3). Due to ethical reasons, no other samples or clinical data concerning neuropsychiatric symptoms were available from these individuals. The patient Hu2 was also positive for Puumala hantavirus IgG-antibodies, demonstrating previous exposure to rodent excreta and rodent-borne pathogens.

The sample from BDV-IgG IFA-positive human (Hu1, veterinarian) had a high avidity of BDV-IgG-antibodies, reacted with P protein and detected multiple linear BDV-N and -P epitopes in the peptide array (Table 2, Fig. 3e and f). The other two IFA-positive humans had a low IgG avidity, but the positivity of the serum from Hu2 could be verified by its reaction with recombinant P protein.

Altogether, by N/P protein/peptide assays, we studied and confirmed the BDV-IgG-positivity of 10 horses, 2 cats, 1 dog, 2 rodents and 2 humans. In conclusion, we have shown serologically that BDV or a closely related virus exist in Finland. Furthermore, it seems very likely that a Finnish veterinarian was infected with BDV as his/her serum tested positive by several distinct methods. In addition, we present serological evidence, for the first time, for BDV infection of wild rodents.