Vaccine-induced Rabies in a Red Fox (Vulpes vulpes): Isolation of Vaccine Virus in Brain Tissue and Salivary Glands

Oral immunization of wildlife via dissemination of vaccine baits in the environment in the late 1980s allowed several European countries to become rabies-free (Cliquet and Aubert 2004). Most vaccines used are modified live-virus vaccines descending from the Evelyn Rokitnicki Abelseth (ERA) virus strain derived from the original Street Alabama Dufferin (SAD) virus isolated from the salivary glands of a rabid dog (Canis lupus familiaris). Since 1992, three commercial oral vaccines have been used in Europe. The SAD strain has been passaged in mouse brain cell cultures and then adapted to baby hamster kidney (BHK) cells by various passages to obtain the SAD Bern vaccine. The SAD B19 strain is derived from SAD Bern by additional passages on cloned BHK cells. The Street Alabama Gif (SAG2) strain was selected from SAD Bern by two successive mutations affecting the triplet coding for amino acid 333 of the glycoprotein.

Since 1995 large-scale oral rabies vaccination (ORV) programs have been carried out in Slovenia. Each ORV campaign (using only SAD B19 vaccines) was evaluated through rabies incidence, bait uptake (indicated by a tetracycline biomarker), and rabies immunization rate on foxes collected in vaccinated areas (Hostnik et al. 2011; WHO 2013).

Some of the modified live-rabies virus vaccines currently used in wildlife may have a residual pathogenicity related to the viral strain's attenuation level (WHO 2013). Therefore, any rabies virus isolated from animals in vaccination areas should be characterized using monoclonal antibodies or molecular techniques to detect any vaccine-induced rabies (WHO 2013).

We analyzed the genome of a vaccine virus strain isolated in Slovenia (KC522613) in spring 2012 in a red fox (Vulpes vulpes) exhibiting clinical signs and identified positive for rabies. The nucleoprotein, phosphoprotein, matrix protein, and glycoprotein genes were analyzed and compared to other referenced lyssavirus species in GenBank.

During rabies surveillance, a young male fox was found in the village of Most na Soči (Primorska region) about 10 km from the Italian border. The animal was wandering around the village at midday on 31 May 2012. The fox exhibited clinical signs suggestive of rabies infection, including unusual behavior, aggressiveness, and uncoordinated movements. The animal was shot and submitted for rabies diagnosis. The fox was found in an area that had been treated in May 2012 with SAD B19 Fuchsoral vaccine (batch 7880811, titer 10^6.2 PFFU/mL, EF 206709).

Two fox samples (brain and salivary glands) were tested using standard rabies diagnostic techniques (fluorescent antibody test [FAT] and Rabies Tissue Culture Infection Test (RTCIT) (Meslin et al., 1996) as well as with real time reverse-transcriptase PCR (RT-qPCR) (Hayman et al. 2011). Fox age was determined by histologic dental examination, and a tooth sample was tested for tetracycline marker (Johnston et al. 1987).

The partial lyssavirus genome (N-G fragment, 4.5 kb) was amplified with one long distance PCR. Briefly, reverse transcription was performed with 2 µg of total RNA extracted from the original brain homogenate and 20 pmol of random primer using Superscript III (Invitrogen, Saint Aubin, France). The 4.5 kb PCR products were amplified as previously described (Marston et al. 2007).

The PCR products were purified, cloned into a pJet vector (ThermoFischer Scientific, Courtaboeuf, France), and bidirectionally sequenced. A dataset of 31 sequences was constituted for the analysis of the 4.5 kb Lyssavirus genome including the fox isolate (KC522613), four classical field lyssavirus species, 25 referenced vaccine strains, and the European Bat Lyssavirus type-1 acting as an outgroup. Phylogenetic analysis was performed with the neighbor-joining method with MEGA software (Tamura et al. 2011. The bootstrap probabilities of each node were calculated using 1,000 replicates to assess the robustness of the method.

The tooth sample showed a line of tetracycline, indicating that the animal had consumed a bait; the fox's age was <1 yr. Brain and salivary glands were positive for rabies by FAT, RTCIT, and RT-qPCR. By using RTCIT, one passage was sufficient to detect infectious virus in both brain and salivary glands.

Phylogenetic analysis of the 4,351-base pair fragment amplified from the isolate KC522613 showed that it clustered with the group forming the SAD B19 vaccines with a bootstrap value of 100% (Fig. 1). Nucleotide identity between the partial N-gene sequence of the brain and salivary glands was 100%.

The pairwise comparison of the four rabies genes between the isolate KC522613 and rabies vaccines confirmed the virus as a rabies vaccine strain (Table 1). The sequence identities ranged from 99.9 to 100% among the N, P, M, and G genes from the fox isolate and SAD B19 vaccines (Table 1). The nucleotide mutations between the isolate and vaccine SAD B19-10th (EU877071) were shown on position 1335 of the N gene (mutation G→A) and on position 3114 of the M gene noncoding region (mutation A→G).

The isolate KC522613 was characterized in codon 333 (G gene) by the presence of amino acid arginine. The sequence was clearly distinct from SAG2 (EF206719), characterized by two mutations in the codon 333 yielding glutamine (Gln) at this position instead of the arginine (Arg). The nucleotide variation on position 1335 of the KC522613 N gene resulted in an amino acid exchange from Arg to Gln (amino acid position 422).

Although SAD B19 and SAD Bern are attenuated rabies virus strains, few vaccine-induced rabies cases have been documented in target and nontarget species. Very few studies investigated the presence of rabies virus in salivary glands of such cases, and if done, results were negative (as described below). Experimental studies on SAD B19 vaccine safety were undertaken by the manufacturer in 16 animal species (Vos et al. 1999). No vaccine virus was detected in the saliva of vaccinated animals except in foxes, in which the virus was detected in the tonsils and parts of the mucosa up to 4 days postinoculation. Brain and salivary glands, on the other hand, were negative. Despite attenuation, SAD Bern and SAD B19 are pathogenic for adult mice and other rodent species (Apodemus sp., Microtus sp., Mus musculus, Microtus epiroticus, Apodemus sylvaticus), whatever the route of inoculation (cerebral, muscular, or oral) (Artois et al. 1992; Vos et al. 1999). A laboratory study (Rupprecht et al. 1990) demonstrated that the SAD B19 vaccine caused rabies in skunks (Mephitis mephitis) via the oral route. Several vaccine-derived rabies cases in wild and domestic carnivores have been reported in Switzerland following field distribution of SAD Bern vaccines (Wandeler et al. 1988) and in baboons (Papio ursinus; Bingham et al. 1992) following oral vaccination. More recently, Muller et al. (2009) reported six vaccine-induced rabies cases in foxes caused by SAD B19 and SADP5/88 (adaptation of SAD Bern to the BHK-21 clone BSR) in vaccinated areas in Germany and Austria, respectively. Three of six brain tissue samples were negative using FAT but positive with RTCIT and RT-PCR. Salivary glands were not investigated. In Canada, nine vaccine-associated cases in four red foxes, two raccoons (Procyon lotor), two striped skunks, and one bovine calf (Bos primigenius) were identified in areas vaccinated with the ERA vaccine (Fehlner-Gardiner et al. 2008). Salivary glands from one animal were tested, but no rabies virus was detected.

This is the first report demonstrating that a vaccine-associated virus might be detected in the salivary gland tissues of a fox naturally infected with the SAD B19 strain contained in oral vaccines. Rabies virus was detected in the fox's brain and salivary glands on the first cell passage, suggesting that rabies virus was above the threshold of detection. The fox was found in an area where 81.5% and 56% of sampled foxes were positive for bait uptake and rabies antibodies, respectively. The fox's age was <1 yr, similar to previous findings reporting vaccine-induced cases mainly in juveniles (Fehlner-Gardiner et al. 2008). The fox was discovered 3 wk after vaccine distribution. As a tooth sample was positive for tetracycline, it is likely that this case resulted from infection following bait consumption during the spring campaign, and that incubation was no longer than 3 wk. The fox had no detectable rabies antibodies (data not shown), confirming that infected animals do not systematically produce antibodies in the final stage of the disease. This case confirms that in rare circumstances the vaccine virus strain SAD B19 can enter neurons, propagate within the central nervous system, and cause rabies, resulting in death. The SAD B19 virus might be released into the salivary glands after propagation in the central nervous system. Because the infected fox showed aggressiveness during the daytime and could therefore bite a human being or another animal, and because vaccine rabies virus was identified in the salivary glands, this infected fox could transmit the infection to other mammals. We cannot rule out the possibility that the infection of this young fox could be the result of a lateral transmission from another SAD B19-infected animal. Further investigations are needed to establish if a fox experimentally vaccinated with SAD B19 can transmit the virus through saliva to healthy animals.

Even though case reports of the residual pathogenicity of SAD B19 remain limited, at a time when environmental issues are being raised, even a single case of vaccine-induced rabies in Europe in the 21st century is reason for debate, there being safer and more effective modern alternatives.

We thank Jedrt Maurer Wernig, Head of the Animal Health Department, Veterinary Administration of the Republic of Slovenia, for constant support and Mélanie Biarnais from Anses Nancy Laboratory for technical assistance.