Mystery of fatal ‘staggering disease’ unravelled: novel rustrela virus causes severe meningoencephalomyelitis in domestic cats

Inflammatory central nervous system diseases in mammals often have unknown causes, particularly within the broad group of non-suppurative, lymphohistiocytic encephalitides where conventional diagnostics frequently fail to identify an agent. Since the 1970s, a distinct clinicopathological syndrome in cats termed ‘staggering disease’—characterized by hind limb ataxia, increased muscle tone, and a consistent histopathology of non-suppurative meningoencephalomyelitis—has been reported mainly in Sweden and later near Vienna, Austria. For decades, Borna disease virus 1 (BoDV-1) was suspected, but robust, reproducible evidence in affected cats from Sweden and Austria was lacking and reported sequences conflicted with BoDV-1’s known phylogeographic patterns. Recent advances in clinical metagenomics have uncovered novel encephalitic viruses, including rustrela virus (RusV), a rubella-related matonavirus associated with encephalitis in diverse mammals and linked to Apodemus spp. rodents. This study investigates whether RusV is the causative agent of feline ‘staggering disease’.

Early reports from the 1970s–1990s described a consistent feline syndrome with non-suppurative meningoencephalomyelitis and characteristic clinical signs, suggesting a viral etiology. BoDV-1 was long considered a candidate due to experimental infections and sporadic natural infections in domestic mammals within limited endemic regions in Central Europe. However, unequivocal detection of BoDV-1 in ‘staggering disease’ cats from Sweden was not achieved and alleged sequences were likely artifacts, contradicting established regional clustering of BoDV-1. Concurrently, metagenomic approaches have identified unexpected encephalitic pathogens in animals and humans, including novel bornaviruses and other RNA viruses. RusV was recently discovered in encephalitic zoo mammals in northern Germany with putative reservoir in yellow-necked field mice (Apodemus flavicollis), expanding the known diversity and host range of Matonaviridae. These findings motivated examination of RusV in long-standing feline ‘staggering disease’ hotspots.

Study design and samples: Brain and/or spinal cord from 29 cats (Sweden n=15; Austria n=9; Germany n=5) fulfilling inclusion criteria for ‘staggering disease’ (non-suppurative lymphohistiocytic meningoencephalomyelitis/encephalitis predominating in grey matter, plus compatible clinical signs) from 1991–1993 (Austria) and 2017–2022 (Sweden and Germany) were analyzed. Controls included 21 cats without encephalitis (Sweden, Austria, Germany) and 8 German cats with other encephalitides or causes (e.g., FIP, vasculitis, immune-mediated limbic encephalitis). Clinical metadata (age, sex, outdoor access, disease duration, seasonality, signs) were collated. Additionally, 116 archived rodent brains from near Grimsö, Örebro County, Sweden (Apodemus sylvaticus n=106; A. flavicollis n=10) collected 1995–2019 were screened.

BoDV-1 testing: Bornavirus RNA detection used panBorna v7.2 and BoDV-1 Mix-1 RT-qPCR assays; BoDV-1 nucleoprotein detection used IHC (monoclonal Bo18). No bornavirus RNA or antigen was found in any of the 29 ‘staggering disease’ cats.

Metagenomic sequencing: Selected cat samples underwent unbiased metagenomic HTS (Ion Torrent S5XL and Illumina NovaSeq 6000) with target capture enrichment using panRubi myBaits panels v2 (based on northeastern Germany RusV) and updated v3 (adding Swedish and Austrian RusV). Some samples had pan-mammal rRNA depletion prior to library prep. Reads were quality trimmed (Trim Galore/454 software), GC-filtered (PRINSEQ-lite), assembled (SPAdes), and mapped to RusV references in Geneious to generate consensus genomes with subsequent annotation.

RusV-specific detection: Initial RusV RT-qPCR (RusV Mix-1) targeted German sequences. A new broad-range RT-qPCR (panRusV-2) targeting a highly conserved 5′ genomic region was designed from German and Swedish sequences. For Sanger sequencing, primers targeting a conserved 449 nt fragment at the 5′ end of p150 (yielding 409 nt consensus) were developed.

In situ hybridization (ISH): RNAscope ISH employed a custom probe against the conserved 5′ genomic region (based on Swedish consensus). Positive and negative control probes (Felis catus PPIB; bacterial DapB) and tissue controls were used. Signals were semi-quantitatively scored (0–3).

Immunohistochemistry (IHC): A mouse monoclonal antibody (2H11B1) against RusV capsid (aa 128–308) was generated using recombinant protein expressed in Expi293 cells with Strep-tag purification. Brain sections were processed with citrate antigen retrieval, primary antibody incubation, polymer detection and DAB visualization; controls included RusV-positive capybara brain, reagent controls, and irrelevant antibody controls. Signals were scored (0–3).

Histopathology: Formalin-fixed paraffin-embedded sections of brain and spinal cord were stained with H&E; selected cases also had Luxol Fast Blue-Cresyl Echt Violet to emphasize grey/white matter. Inflammation severity was graded. Rodent brains were also examined.

Rodent screening: PanRusV RT-qPCR screened 116 rodents; subset underwent ISH. Species IDs were confirmed via cytochrome b sequencing.

Phylogenetics and mapping: Amino acid and nucleotide phylogenies (IQ-TREE2 with appropriate models; 100,000 ultrafast bootstraps) were constructed for complete and partial sequences. A sliding window analysis (JC69) assessed genomic variability. Geographic origins were mapped for sequence clades and cases.

• BoDV-1 not detected: Bornavirus RNA and BoDV-1 antigen were absent in all 29 cats fulfilling ‘staggering disease’ criteria.
• Strong RusV association: RusV infection was confirmed in 27/29 cats (93.1%) by at least two independent methods (RT-qPCR, ISH, IHC, and/or HTS). Method-specific detection among ‘staggering disease’ cats: HTS 15/17; RT-qPCR 26/29 (Sweden 15/15; Austria 8/9; Germany 3/5); ISH 24/29; IHC 27/29. No RusV detected in controls: 0/21 cats without encephalitis and 0/8 cats with other encephalitides across all methods applied.
• Genomics and phylogeny: Complete or near-complete RusV genomes were obtained for cats from Sweden (3), Austria (2), and Germany (1). Swedish and Austrian sequences formed distinct clades more closely related to each other than to northeastern German sequences. Nucleotide identity ranged down to ~75–77% between regional lineages; GER_04 shared ≥92.1% identity with previously published German RusV. A conserved 5′ genomic region (approx. nt 1–300) enabled broad-range assay design; higher variability was observed in the intergenic region and within p150 (nt ~2100–2600).
• Tissue tropism: ISH and IHC localized RusV RNA and capsid protein predominantly to neuronal cytoplasm, including Purkinje cells, dentate gyrus granule cells, cortical pyramidal neurons, brainstem and spinal ventral horn neurons; signals sometimes extended beyond overt inflammatory areas.
• Clinicopathology: Affected cats were all adults (median age 3.2 years; range 1.5–12.3); 21/27 (77.8%) were males; all reported had outdoor access. Onset was more frequent in winter–spring (Dec–May: 18 cases) vs summer–fall (7 cases). Clinical signs included progressive gait abnormalities, ataxia, hind-limb weakness, proprioceptive deficits, behavioral changes, hyperaesthesia, cranial nerve signs, and occasional seizures. Histology showed polio-predominant, angiocentric lymphocytic/lymphohistiocytic meningoencephalomyelitis, often most pronounced in brainstem, cerebral cortex, and spinal cord grey matter.
• Reservoir screening: RusV RNA was detected in 8/106 (7.5%) wood mice (Apodemus sylvaticus) from Örebro County, Sweden (Cq 20–35), across years 1996, 1997, 2005, and 2011; 0/10 yellow-necked field mice from the same area were positive. ISH confirmed RusV RNA in 4/4 RT-qPCR-positive wood mice tested; no encephalitic lesions were seen in rodents.
• Spatial patterns: Three geographic RusV clades corresponded to Sweden, Austria, and northeastern Germany; Swedish sequences had three subclades including one shared with wood mice. The cat GER_01 sequence from Hannover clustered with the Austrian clade despite no travel to Austria; the cat had been imported from China a year before disease onset.
• Diagnostic advancement: Broadly reactive RT-qPCR, ISH probes, and a monoclonal antibody detected diverse RusV lineages despite substantial sequence divergence.

The findings refute BoDV-1 as the cause of feline ‘staggering disease’ and strongly implicate rustrela virus as the etiologic agent. The near-universal detection of RusV RNA and/or antigen in cats meeting ‘staggering disease’ criteria, contrasted with complete absence in control cats, supports causality. Clinicopathological features—neuronal tropism, grey matter–predominant inflammatory lesions, and distribution in cerebellar and brainstem structures—align with RusV-associated encephalitis described in other mammals, reinforcing a cohesive disease entity. Significant genomic diversity among regional RusV lineages complicated assay design, but identification of a conserved 5′ genomic region enabled sensitive, broad-range molecular and ISH diagnostics, and a monoclonal anti-capsid antibody provided robust antigen detection across clades. The detection of RusV in Swedish wood mice suggests Apodemus spp. as probable reservoirs, though species differences across regions (A. flavicollis in Germany versus A. sylvaticus in Sweden) imply local adaptation or sampling composition effects. The geographically clustered phylogenies and sporadic occurrence in cats with outdoor access suggest local maintenance within small mammal reservoirs and spillover into cats as dead-end hosts. Seasonality trends (more winter–spring onsets) may relate to reservoir population dynamics and seasonal rodent–human/animal contact patterns. While experimental infection and virus isolation are pending, the strength and consistency of associations, supported by multiple orthogonal detection methods and controls, justify considering RusV as the causative agent and integrating RusV testing into diagnostic criteria. Broader host range and distribution raise concern that RusV may underlie additional mammalian neuropathologies, potentially including human disease, warranting further surveillance and research.

This study provides compelling evidence that rustrela virus is the long-sought cause of feline ‘staggering disease’, overturning long-standing assumptions about BoDV-1 involvement. The authors detected RusV RNA and antigen in the vast majority of affected cats across Sweden, Austria, and Germany, identified phylogenetically distinct regional clades, and discovered RusV in Swedish wood mice as a likely reservoir. They developed and validated broadly reactive molecular, ISH, and IHC tools that can facilitate diagnosis and surveillance. Future work should focus on virus isolation, experimental infection models to fully satisfy causality criteria, expanded epidemiologic studies across regions and host species, development of serological assays, investigation of tissue tropism and shedding, transmission pathways, reservoir ecology, and assessment of zoonotic potential.

• No RusV isolate was available; experimental infection to fulfill Koch’s postulates was not performed.
• Some analyses relied on archived FFPE tissues, which may reduce nucleic acid quality and lead to occasional false negatives or fragmented sequences.
• Whole-genome sequencing was challenging for FFPE material, limiting completeness of some sequences.
• Transmission routes, shedding, tissue distribution outside the CNS, and incubation period remain unknown.
• Rodent surveillance was geographically restricted and species composition was uneven (many more wood mice than yellow-necked field mice), potentially biasing reservoir inference.
• Cross-sectional design precludes temporal inference of infection relative to disease onset.
• Considerable RusV genetic diversity may still challenge universal assay performance despite targeting conserved regions.