Novel Pestivirus Species in Pigs, Austria, 2015

A novel pestivirus species was discovered in a piglet-producing farm in Austria during virologic examinations of congenital tremor cases. The emergence of this novel pestivirus species, provisionally termed Linda virus, in domestic pigs may have implications for classical swine fever virus surveillance and porcine health management.


The Study
A small-scale piglet-producing farm in Styria, in southeastern Austria, reported major piglet losses from CT in January 2015. Animals from this farm were examined and samples were taken by authorized veterinarians. The CTaffected piglets showed a severe lateral shaking and were often incapable of sucking milk, which led to elevated preweaning death rates (Video, https://wwwnc.cdc.gov/EID/ article/23/7/17-0163-V1.htm). The prevalence of CT varied from 20% to 100% within the affected litters. Litters affected by CT were not used for the production of replacement gilts. The outbreak of CT abruptly stopped in July 2015, when all sows had produced 1 CT-affected litter. Only 22.4 piglets per sow were weaned in that year, compared with an average of 25.8 piglets/sow/year before CT symptoms occurred. Pathological examinations were performed as described previously (8), confirming the presence of typical CT A-II lesions ( Figure 1, panel A).
We tested various samples from the farm using an APPV-specific TaqMan probe-based reverse transcription PCR (RT-PCR) (8) but obtained negative results. However, we obtained an amplicon of appropriate length from CT-piglet serum samples by using a novel panpestivirus RT-PCR (PPF 5′-GTKATHCAATACCCTGARGC-3′ and PPR 5′-GGRTTCCAGGARTACATCA-3′), which enables detection of CSFV, BVDV-1, BVDV-2, BDV, BV, and APPV. Surprisingly, sequencing of this RT-PCR product yielded an unknown sequence with a noninterrupted open reading frame. An initial BLAST search (https://blast.ncbi. nlm.nih.gov/Blast.cgi) resulted in "no significant similarity found," but the translated sequence of 270 aa aligned well with different pestiviruses. The viral RNA was detected in all samples (serum, tonsils, lung, liver, spleen, and central nervous system material) of CT-affected piglets and their littermates from the farm.
After inoculating SK-6 cells with serum samples of affected piglets, we detected the amplification of this unknown pestivirus using RT-PCR. We provisionally termed the agent "Linda" (lateral-shaking inducing neurodegenerative agent) to avoid confusion with other pestiviruses. We determined the full genome of Linda virus (LV) using the standard primer walking RT-PCR approach together with rapid amplification of cRNA ends PCR to identify the ultimate 5′-and 3′termini using the cultured virus. We determined the length of the LV genome (GenBank accession no. KY436034) to be 12,614 nt, with 381 nt 5′-nontranslated region, an open reading frame of 11,772 nt, and 461 nt of 3′-nontranslated region.
We performed a phylogenetic analysis of LV using CLC Workbench 7.6 (CLCBIO, Aarhus, Denmark), which demonstrated LV's divergence from other pestiviruses ( Figure 1,  LV, approved pestiviruses, and APPV to be only 60%, but we found an identity of 68% between LV and BV. Comparison of the pestiviral polyprotein sequences yielded an amino acid identity of 69% between LV and BV and of <54% with all other pestiviruses ( Figure 1, panel C). All known cleavage sites of the NS3 protease were present in LV (10). We mapped the NS4A-NS4B cleavage, which takes place at an L/A site in classical pestiviruses and at an L/S site in BV, to the L 2354 /S 2355 motive in LV; the intramolecular NS3 cleavage site (L 1834 /A 1835 ) was conserved in LV (11). Each hypothetical mature protein of LV matched best with the mature proteins of BV (Table), even if the identities were relatively low in the structural protein region.
We saw a strong reactivity against LV infected cells when using a BVDV E2-specific antibody termed 6A5 (12). This antibody detected the E1-E2 heterodimer of LV (75 kDa) as well as the E2 monomer (50 kDa) in Western blot analysis, although the reactivity against BVDV-1 E2 was much stronger (Figure 2, panel A). With the help of this antibody, we were able to study the in vitro growth of LV. E2 antigen signals of LV were detectable in the cytoplasm of SK-6 and MDBK cells after infection, but the antigen-positive foci were 10-fold larger in SK-6 than in MDBK cells (Figure 2, panels B,C). Already in the primary passage on porcine cells, considerably high infectious titers of LV were measured in the supernatant (>10 7 50% tissue culture infectious dose/ mL). Using monoclonal antibody 6A5 in immunohistochemical analysis, we detected pestiviral antigen in CT-affected piglets in regions with histological lesions. Specific stains in neurons of the nucleus of the trigeminal nerve (Figure 2, panels D,E), in glial cells in the cerebellar and the cerebral white matter, and in tubular epithelium of the kidneys proved that a pestivirus crossed the blood-brain barrier. As in the case of other pestiviruses, the mechanisms inducing hypomyelination during fetal development remain unclear (8,13).

Conclusions
A previously unknown pestiviral agent was found in CTaffected piglets on a farm in Austria. We observed a severe hypomyelination in the entire white matter of the spinal cord and detected pestivirus antigen in the brain of the CTaffected piglets, suggesting a causal relationship between infection and lesions. Analyses of the assembled genome allowed an unambiguous assignment of LV within the genus Pestivirus with regard to the presence of pestivirusspecific genes (N pro and E rns ) and sequence homology to other pestiviruses (14). In contrast to APPV, which hardly infects cultured cells at all, LV could be easily propagated on porcine cell lines without the need for adaptation, similar to what has been reported for BV. We suggest that LV likely shares a common ancestor with BV. After its description ≈10 years ago, BV was intensively sought worldwide, but the virus has never been detected outside Australia. The broad cell culture tropism of BV led to the hypothesis that the virus recently jumped from another species to porcine hosts (15). The discovery of a related pestivirus with substantial sequence divergence in swine on a different continent suggests that both viruses probably have a porcine origin. The identification of the cross-reacting E2-specific monoclonal antibody 6A5 indicates that a cross-reactivity to related pestiviral proteins exists, which might interfere with the serologic testing for CSFV. Further work is needed to investigate the prevalence and the epidemiology of LV in Europe and to assess its virulence in controlled animal experiments.