Introduction
In June 2003, an outbreak of sudden deaths occurred in 2–3-week-old sucking piglets on a large corporate farm complex in southern New South Wales, Australia. This event was soon followed by an elevation in the proportion of stillborn fetuses, increased preweaning losses and to a lesser extent increased mummification rates. Although some aspects of the clinical presentation were consistent with known infectious diseases such as encephalomyocarditis virus infection, the scale of the outbreak, the rapid spread and some aspects of the clinical presentation suggested that this was likely to be a novel infectious disease.
Disease outbreak
While the first clinical cases were observed in a single production module on the index farm, after 4–6 weeks, the disease spread to other modules on the same farm and eventually spread to a second farm approximately 60 km away. This farm had received some pigs from the first farm and was also a large complex with five production modules. Clinical cases were reported from these farms for approximately 13 months. During this time it was estimated that approximately 50,000 piglets were lost due to this disease.
Clinical signs
Initially, there was an increase in sudden deaths in unweaned piglets. Soon after, there was also a dramatic increase in the incidence of stillbirths and a smaller rise in mummified piglets but few abortions were observed. Placentas appeared normal and dams of the affected piglets also appeared clinically normal with normal feed intake and body temperature. However, just prior to the start of the outbreak, there was an indication that some sows showed vague signs of inappetence or appeared unsettled.
Piglets born alive in affected litters showed greatly reduced viability, with a preweaning mortality of up to 35%. Piglets born alive from affected litters but dying within the first 3 weeks of life were usually of normal size and weight and rarely showed premonitory signs before death. Some affected piglets developed cyanosis of the snout and ears, vocalization and hyperventilation before death. During the early stages of the outbreak, trucking piglets after weaning at 3 weeks of age sometimes resulted in sudden deaths. In-contact older weaners and growers were unaffected. Growth rates and mortality data for grower and finisher pigs that were companion pigs to the affected litters were all similar to those of the previous year and those of unaffected farms under similar ownership and management (McOrist et al., Reference McOrist, Thornton, Peake, Walker, Robson, Finlaison, Kirkland, Reece, Ross, Walker, Hyatt and Morrissy2004).
Pathology
While many affected stillborn piglets showed few gross abnormalities, a proportion showed distinctive gross changes of anasarca. Gross findings at autopsy included moderate subcutaneous edema, an excess of pericardial, pleural and peritoneal fluid, bilateral cardiac dilation with ecchymotic hemorrhages in the myocardium and well-developed diffuse venous congestion in the liver. Some hearts showed irregular areas of pallor and there were sometimes fibrin tags on thoracic and abdominal viscera.
Histopathological findings in affected piglets consistently included acute to subacute multi-focal non-suppurative myocarditis with myonecrosis. A minority of cases in older unweaned piglets showed early myocardial fibrosis. Occasionally affected piglets showed non-suppurative interstitial pneumonitis, encephalitis, hepatitis and lymphadenitis (McOrist et al., Reference McOrist, Thornton, Peake, Walker, Robson, Finlaison, Kirkland, Reece, Ross, Walker, Hyatt and Morrissy2004).
There were elevated IgG levels in the body fluids of about half of the stillborn piglets, consistent with an in utero infection. Extensive serological and virological investigations failed to identify a known etiological agent.
Identification of Bungowannah virus
After virus isolation and PCR investigations were unsuccessful, direct fetal inoculation was undertaken with tissue from an infected piglet. Fetuses were collected 17–21 days after inoculation. Nucleic acid was purified from serum from several littermates of fetuses that were presumptively infected and was subsequently subjected to sequence-independent single-primer amplification (SISPA) followed by nucleic acid sequencing. PCR primers matching terminal sequences of nucleic acid fragments detected by SISPA were used to amplify intervening RNA segments. The resulting products were cloned, sequenced and the results compared with data in Genbank. A small number of sequences consistent with a pestivirus were obtained (Kirkland et al., Reference Kirkland, Frost, Finlaison, King, Ridpath and Gu2007). These sequence data had about 60% homology with classical swine fever virus (CSFV). These results caused some consternation because, apart from a rarely encountered BVDV infection, there were no known pestiviruses circulating in pigs in Australia. The last detection of CSFV had occurred more than 40 years ago.
Virus characterization
PCR with ‘primer walking’ was later used to determine the entire viral genome. The characteristics of the genome of this new pestivirus were subsequently compared with published data for representative viruses from other pestivirus species. It was shown that the virus had all of the sequences to encode the recognized structural and non-structural proteins of a pestivirus. However, there was considerable variability at both the nucleotide and amino acid level, sufficient to suggest that this virus represents a novel pestivirus species.
Virus isolation was attempted using PK-15 cell cultures and viral replication was detected initially by PCR. This isolate was named Bungowannah virus after the location where the first cases were identified. Later, after optimizing cell fixation, viral antigens were detected in cell culture by immunoperoxidase (IPX) staining using convalescent pig serum. Antigenic studies, involving both peroxidase linked assays and virus neutralization, were undertaken with a collection of polyclonal antisera raised against a range of representative pestiviruses. The results showed little evidence of antigenic relatedness to recognized pestiviruses. A wide range of pan-pestivirus reactive monoclonal antibodies were also used in IPX assays but no reactivity was detected.
More recently, monoclonal antibodies specific to Bungowannah virus have been developed and a range of in vitro cell culture-based studies have been undertaken (Richter et al., Reference Richter, Reimann, Wegelt, Kirkland and Beer2011; Kirkland et al., Reference Kirkland, Frost, King, Finlaison, Hornitzky, Richter, Reimann, Dauber, Schirrmeier, Beer and Ridpath2014). These include an assessment of the potential host range and have shown that Bungowannah virus will grow in cells extending from humans to bats as well as farm animals.
Molecular studies involving the exchange of genes between Bungowannah virus and BVDV show that genes encoding both key structural and non-structural proteins remain fully functional when inserted into a BVDV backbone (Richter et al., Reference Richter, Reimann, Wegelt, Kirkland and Beer2011, Reference Richter, Konig, Reimann and Beer2014). However, the wide in vitro host range of Bungowannah virus was not retained when a chimeric BVDV expressed the envelope proteins of Bungowannah virus, leading to the conclusion that both the envelope proteins and replication machinery of Bungowannah virus are responsible for its capacity to infect a wide range of cell cultures (Richter et al., Reference Richter, Reimann, Wegelt, Kirkland and Beer2011).
Evidence for Bungowannah virus causing disease in pigs
Soon after the identification of this novel pestivirus, several different approaches were adopted to accumulate evidence that this virus alone was the cause of this disease outbreak. Firstly, a PCR directed at sequences unique to Bungowannah virus was used to test tissues of naturally infected fetuses. Bungowannah-specific nucleic acid sequences were amplified from a number of piglets (Kirkland et al., Reference Kirkland, Frost, Finlaison, King, Ridpath and Gu2007; Finlaison et al., Reference Finlaison, King, Frost and Kirkland2009). Utilizing a peroxidase-linked assay, extensive cross-sectional serological studies were also undertaken on the affected farms. These studies showed a very high prevalence of infection in breeding sows in production modules where the disease had occurred and, at the time of initial sampling, most of the sows on the second farm complex were seronegative and disease free. Seroconversions in apparently unaffected groups of sows were soon followed by fetal losses until the virus had spread through all of the production modules on both of the large farm complexes (Finlaison et al., Reference Finlaison, King, Frost and Kirkland2009). Sows on other farms under the same ownership remained seronegative and disease free. Bungowannah-specific antibodies were also identified in fetal fluids of stillborn piglets that had elevated IgG levels. Finally, Bungowannah viral RNA was localized in the myocardial lesions of affected fetuses by using in situ hybridization (Finlaison et al., Reference Finlaison, King, Frost and Kirkland2009). Collectively, these data show widespread circulation of this pestivirus on the affected farms, with both positive serology in sows producing affected litters and both antibodies and viral RNA in affected piglets.
Experimental infections
Pregnant sows
While a myocardopathy was considered to be the primary lesion affecting fetuses, there was a range of clinical signs and gross pathology observed when piglets were born. It was thought that these variations were likely to be the outcome of fetal infection at different stages of gestation. Consequently, experimental infections were established in sows at different stages of gestation (Finlaison et al., Reference Finlaison, Cook, Srivastava, Frost, King and Kirkland2010). Sows were infected intranasally at 35, 55, 75 and 90 days of pregnancy and were allowed to proceed to term. Piglets were collected and sampled prior to suckling. All sows became infected, 87% (20/23) of litters were infected and, within an affected litter, all but one of the fetuses were infected. There was a range of clinical presentations depending on the stage at which infection had occurred but all were consistent with field observations. Following infection at day 35, there was a high incidence of stillbirths and mummified piglets. The reduction in viable piglets was much less at days 55 and 75 and only slightly higher than the uninfected control litters. Cumulative pre-weaning losses of 70 and 29% were observed following infection very early (day 35) and late in gestation (approximately day 90), respectively. Most piglets infected at day 35 were shown to be persistently infected, stunted and had poor viability while chronic infections were observed after fetal infection at day 55. Chronically infected piglets showed growth retardation and were viremic for up to 7 months, after which time they were able to clear the virus and showed improved growth. Acute myocarditis was associated with infection in late gestation (day 90).
Weaners and growing pigs
Experimentally infected weaner age pigs underwent an acute infection that is characteristic of many pestiviruses – they became viremic from 3 days post-infection and also commenced excretion of virus (Finlaison et al., Reference Finlaison, King, Gabor and Kirkland2012). Moderate quantities of virus were detected in oropharyngeal fluid. Seroconversion was detected from 10 days post-inoculation and both viremia and shedding decreased rapidly after seroconversion. There were no apparent clinical signs and hematological parameters mostly fell within normal ranges. When naïve weaners were housed with acutely infected pigs there was no evidence of transmission. In contrast, in-contact pigs rapidly became infected when co-housed with either persistently or chronically infected pigs.
Sheep and cattle
In limited studies, non-pregnant sheep and cattle could become infected after experimental intranasal challenge but there was no evidence of disease. There was limited virus excretion with RNA detected in nasal secretions for up to 14 days after infection (Read et al., Reference Read, Finlaison, Frost, King, Gu and Kirkland2011). A calf was also readily infected when housed with a persistently infected piglet. In an experiment, to determine whether Bungowannah virus could induce persistent infection in cattle, seven cows were infected intranasally between 53 and 65 days of pregnancy. About half of the cows showed a very brief viremia of 1–3 days duration but all cows shed virus in nasal secretions and seroconverted. Cows were induced to calve at approximately 275 days (one at day 255) after insemination and all produced live calves. Three calves (born at days 274 and 276) had high neutralizing antibody titers in pre-colostral sera and relatively high levels of viral RNA were detected in skin biopsy samples by Bungowannah virus-specific qRT–PCR. One of these calves had a very low-level viremia and had moderately high levels of viral RNA in testicular tissue when euthanized at 4 weeks of age. Virus was not detected in any other tissues of this calf and in no tissues of any of the other calves.
Disease control and eradication
After the initial outbreak subsided in late 2004, Bungowannah virus remained endemic on both of the large farm complexes where there had been infected pigs. As the virus was not known to be present on any other farm in Australia, these two farms remained in quarantine. Subsequently, this novel pestivirus has been eradicated from one farm following a program of depopulation and disinfection. However, similar attempts on the second larger farm have been unsuccessful. After the virus was cleared from several production units, there was a re-introduction, thought to be by aerosols generated during decontamination of an infected module 0.5 to 1.5 km away from the Bungowannah virus-free units. Introduction of this virus into a large susceptible population was associated with severe reproductive losses until widespread immunity in breeding animals was re-established. In some instances a CSFV-like enteric disease was also observed in growing pigs although the role of Bungowannah virus in this clinical manifestation remains unclear. It is possible that these clinical signs may have been due to a combination of pestivirus-induced immunosuppression and bacterial infections.
Monitoring of boar semen
One of the farms on which Bungowannah virus was detected produces replacement boars for the company artificial breeding program. As semen is supplied to farms on which Bungowannah virus has not been detected, potential donor boars undergo a period of quarantine and Bungowannah virus screening to ensure that they are not acutely infected and possibly shedding virus in semen. Only young boars that have high neutralizing antibody titers are selected and semen samples are collected for testing by qRT–PCR. Bungowannah viral RNA has been detected in the semen of up to 20% of these boars at initial screening. Virus has not been isolated from any of these PCR-positive samples, but few have been examined, nor has the semen been used for artificial insemination to determine whether it would be infective for female pigs. However, the RNA load in some of these samples is sufficiently high that transmission could be expected.
Conclusions
Bungowannah virus has been shown to have all of the characteristics of a pestivirus and in some respects is similar to ‘low virulence’ strains of classical swine fever virus, the only other pestivirus to commonly infect pigs. Although it does not cause obvious disease in pigs after birth, it is a significant cause of disease following in utero infection. Similar to other pestiviruses, some of the clinical effects that result from in utero infection are apparent for many months after pigs have been born. Its genetic and antigenic characteristics, while being consistent with a pestivirus, also suggest that it represents a novel species within the Pestivirus genus. Although Bungowannah virus spread rapidly and throughout the two affected farms on which the myocarditis outbreak occurred, the virus did not spread to any other farms. This is believed to be largely due to the ‘farrow to finishing’ nature of the enterprise involved, the fact that the second farm was also the site of the company abattoir and that the affected farms were placed in quarantine. Although there has been considerable speculation about the origin of this virus, the likelihood that it was introduced to these farms as a contaminant in a vaccine remains a strong possibility (Finlaison, Reference Finlaison2010).
Acknowledgments
We are indebted to Australian Pork Limited, the Australian Biosecurity Co-operative Research Centre and NSW Department of Primary Industries for funding of this research. The staff of the Virology Laboratory at EMAI made significant supporting contributions to these studies. Katherine King, Jing Zhang and Glenda Macnamara deserve special recognition for their efforts. We also thank Drs Julia Ridpath and Martin Beer for productive and enjoyable collaboration after the discovery of this novel virus.
