Survey for pestiviruses in backyard pigs in southern Brazil

Abstract

The Pestivirus genus comprises species that affect animal health and productivity worldwide. Members of the Suidae family are hosts for classical swine fever virus (CSFV), an important pathogen tracked by the World Organization for Animal Health (OIE). However, swine are also susceptible to other pestivirus species that can result in disease or compromise CSFV detection. We searched for pestivirus infection in swine sera collected from 320 backyard pig herds in southern Brazil. We used reverse-transcription PCR primers for Bungowannah virus; atypical porcine pestivirus (APPV); and a panpestivirus pair that detects bovine viral diarrhea virus (BVDV)-1, -2, and HoBi-like pestivirus (HoBiPeV), border disease virus (BDV), and CSFV. Two samples were positive using the panpestivirus primer pair and were classified as BVDV-1d and -2a, respectively. Serum samples were tested for virus neutralization against BVDV-1a, -1b, and -2 strains, resulting in 28 (4.4%) positive samples. Of those, 16 samples had the highest titers against BVDV-1a (2), BVDV-1b (5), and BVDV-2 (9). Our results indicate that Bungowannah virus, APPV, CSFV, BDV, and HoBiPeV have not been circulating in these specific backyard swine populations. However, ruminant pestiviruses were detected and must be considered in future pestivirus control programs conducted in Brazil.

The Pestivirus genus (family Flaviviridae) was previously composed of 4 recognized species named Bovine viral diarrhea virus 1 (BVDV-1), BVDV-2, Border disease virus (BDV), and Classical swine fever virus (CSFV). A new taxonomy and species nomenclature for the genus was adopted in 2018. Now, BVDV-1 has been named Pestivirus A, BVDV-2 is Pestivirus B, CSFV is Pestivirus C, and BDV is Pestivirus D. As previously described, 7 putative pestivirus species—from Pestivirus E to Pestivirus K—were recognized as official members of the genus, including Bungowannah virus and atypical porcine pestivirus (APPV), which are known to infect and cause disease in pigs.¹⁵

Pestiviruses have positive single-stranded RNA genomes of ~12.3 kb, with a single open reading frame (ORF) translated into 12 viral polypeptides flanked by untranslated regions (UTRs) at the 5’- and 3’-ends. Conserved genomic regions, especially 5’-UTR, have been used for genotyping pestivirus species and variants through phylogenetic analysis, which has resulted in at least 21 subtypes for BVDV-1 (1a–1u) and 3 for BVDV-2 (2a–2c).³

The pestivirus species were initially named according to their host, with pigs as hosts for CSFV, sheep for BDV, and cattle for BVDV-1 and -2, although alternative hosts have been described over time. CSFV is considered one of the most important swine viral pathogens, which has led to the official eradication of this disease in many countries. BVDV has spread worldwide and is considered an important pathogen for the cattle population, having economic impacts, especially as a result of reproductive losses and the consequent birth of persistently infected calves. Given this fact, some countries have included BVDV in official eradication programs in cattle.

Some pestiviruses, such as BVDV, can infect unusual hosts sporadically; BVDV has been found in several members of the order Artiodactyla, such as sheep, goats, deer, bison, swine, and wild boars.¹⁵ BVDV infection in swine can be asymptomatic, which allows propagation of the virus in the herd. In some cases, BVDV infection can cause signs similar to those of CSFV infection; antigenic and structural similarities among pestiviruses can also lead to misleading detection of CSFV.¹⁴

Other pestiviruses, such as Bungowannah virus and APPV, are apparently restricted to infecting only Suidae members.^7,8 Bungowannah virus was described and detected in only Australia, causing myocarditis in piglets⁸; APPV was recently associated with congenital tremors in piglets worldwide, including in Brazil.¹¹

To assess information about several pestivirus species infections in pigs, we tested sera from backyard pig herds for pestivirus genomes and BVDV antibodies. These samples composed the target population for the CSFV Surveillance Program given that backyard pig farms typically lack biosecurity measures. Samples were analyzed by reverse-transcription PCR (RT-PCR) for CSFV, APPV, Bungowannah virus, and ruminant pestiviruses (BVDV-1, BVDV-2, HoBiPeV, and BDV). Additionally, we used virus neutralization (VN) to quantify the amount of antibody reaction to such infections.

Brazil is considered largely as a CSFV-free zone by the World Organization for Animal Health (OIE), resulting from established eradication and control programs. Despite the large volume of pork meat production in southern Brazil, more than half of the swine farms in Rio Grande do Sul State are backyard herds, characterized as having a small number of animals that are slaughtered and consumed mostly on the farm. This type of herd is characterized by lack of biosecurity measures, and hence these animals were an important target population for the CSFV Surveillance Program. The 741 swine sera from backyard pigs used in our study were collected in 2014 during CSFV surveillance tests in the state of Rio Grande do Sul, southern Brazil. The Official Veterinary Office (Secretaria Estadual de Agricultura, Pecuária e Irrigação, SEAPI-RS) randomly collected blood samples from 320 backyard farms selected from a source population while setting predefined risk criteria for CSFV, such as proximity to waste dumps and the practice of feeding food waste to pigs, to increase the effectiveness of the surveillance system. The samples were collected in 202 counties and comprised sera from male and female animals 6–72 mo old. All samples tested negative in CSFV serology assays performed at the official laboratory in Brazil.

Total RNA was isolated (TRIzol LS reagent; Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. Complementary DNA was synthesized (GoScript reverse transcriptase; Promega, Madison, WI), and PCR was performed (GoTaq; Promega). For ruminant pestivirus and CSFV detection, PCR was carried out using primers named PanPesti F and PanPesti R, which amplified a 118-bp fragment of the conserved 5’-UTR genomic region of ruminant pestivirus species.¹⁸ Additionally, the samples were assayed for Bungowannah virus and APPV using the primer pairs UTR_Left and UTR_Right¹ and Pesti-11453-F and PestiV NS5-R,² respectively. PCR products were subjected to electrophoresis in 2% agarose gels and visualized under ultraviolet illumination.

Samples that were found to be positive in the ruminant pestivirus RT-PCR assay were further tested with the primers 324 and 326,¹⁷ resulting in a longer amplification product of 288 bp that was sequenced and used in phylogenetic analysis. Amplification products were purified (PureLink PCR purification kit; Thermo Fisher Scientific). Both DNA strands were sequenced (ABI PRISM 3100 genetic analyzer; BigDye Terminator v.3.1 cycle sequencing kit; Thermo Fisher Scientific).

The sequences were assembled (Geneious v.9.1.5; Biomatters Limited, Auckland, New Zealand) and analyzed (nucleotide BLAST; https://blast.ncbi.nlm.nih.gov/Blast.cgi). For phylogenetic analysis, 27 reference sequences were retrieved from GenBank (https://www.ncbi.nlm.nih.gov/genbank/), and multiple alignments were performed (ClustalW; available in MEGA v.6.06, https://www.megasoftware.net/). A phylogenetic tree based on the 5’-UTR nucleotide sequences was constructed with MEGA 6 using the neighbor-joining method and the Kimura-2 substitution model.

The serum samples were also tested for the presence of antibodies by VN, using strains of the pestivirus species that were detected by RT-PCR and sequencing. Pestivirus strains used in the comparative VN assay included cytopathic (cp) BVDV-1a Oregon C24V, BVDV-1b Osloss, and BVDV-2a VS-253. BVDV strains were propagated and titered in Madin–Darby bovine kidney (MDBK) cells. MDBK cells used during the study were grown in Dulbecco modified eagle medium (DMEM; Thermo Fisher Scientific) supplemented with antibiotics at a final concentration of 100 units/mL penicillin, 100 µg/mL streptomycin, and 5% equine serum. MDBK cells were previously found to be free of pestivirus RNA by RT-PCR.

Serum aliquots were placed in a water bath at 60°C for 60 min for complement system deactivation. A first screening by VN was performed against the reference strains (BVDV-1a, -1b, and -2), as described in the OIE Manual of Standards for Diagnostic Tests and Vaccines.²⁰ The samples were initially tested in triplicate at a dilution of 1:8 given the rare infection of swine by ruminant pestiviruses and the low titers produced against these infections. Seropositive samples at a 1:8 dilution were retested using 2-fold serial dilutions in DMEM from 1:8 to 1:1,024; 100 tissue culture infectious doses (TCID₅₀) of virus were added to each well. Plates containing serum dilutions and virus were incubated for 100 min at 37°C in 5% CO₂. Following incubation, a 50 μL/well cell suspension containing 10⁵ cells/mL was added. The plates were incubated for another 96 h at 37°C in 5% CO₂. Sera of animals that previously tested negative by VN for BVDV-1 and -2 were used as negative controls. Samples that neutralized the virus at a dilution of 1:1,024 were further tested in dilutions from 1:512 to 1:32,768. Plates were observed using an optical microscope, searching for characteristic cytopathic effects.

In the survey of pestivirus genomes through RT-PCR, we tested all 741 samples. No sample was found to be positive for Bungowannah virus or APPV using specific RT-PCRs. Two serum samples (0.3%) were positive in the ruminant pestivirus RT-PCR with either the PanPesti¹⁸ and 324/32634¹⁷ primers. These samples were named SUI 1 and SUI 2, and the 5’-UTR partial sequences obtained from DNA sequencing of the amplification products using the 324/326 primers were deposited in GenBank as accessions MK334041 and MK334042, respectively.

In a nucleotide BLAST search, the SUI 1 nucleotide sequence was found to have the highest similarity (~98%) to isolates LV/C3P/13 (KP715116) and LF80/11 (JX122862), both BVDV-1d samples from Brazil. In the phylogenetic tree produced based on 5’-UTR nucleotide sequences, the strain SUI 1 and those cited above with the highest nucleotide identity were clustered with the reference strain (BJ1308) for BVDV-1d, along with other Brazilian strains that were classified as the same subtype. These nucleotide sequences were grouped into a BVDV-1d strain branch within the phylogenetic tree with a bootstrap value of 100% (Fig. 1).

Figure 1.

Phylogenetic tree based on the 5’-UTR nucleotide sequence. Neighbor-joining linear tree based on the partial 5’-UTR nucleotide sequence of representative strains for some BVDV-1 and BVDV-2 subtypes. For phylogenetic tree construction, representative sequences were used from GenBank. The samples from our study, SUI 1 and SUI 2, are marked with a black dot. Kimura-2 model with gamma-distributed rate variation and 1,000-bootstrap repetitions were used as parameters to build the tree.

For the SUI 2 sample, the nucleotide sequence was found to be 94% identical to that of BVDV-2 strains from Argentina (isolates 106 and 76/08; GenBank accessions JX848364 and MF120586, respectively). Strain SUI 2 and other strains previously described as BVDV-2a as well as the reference strain for the subtype known as New York93 were grouped to form a branch composed of BVDV-2a strains with a bootstrap of 89% (Fig. 1).

The 741 swine sera tested by VN resulted in 102 (13.7%) samples with toxic effects on the cell culture that were excluded from the VN results. From the 639 nontoxic serum samples, antibodies against the ruminant pestiviruses were detected in 27 samples (4.2%) from 18 households (5.6%).

Despite great genetic and antigenic diversity within the pestiviruses, there is serologic cross-reaction over an extended range between species and even subtypes. To accurately detect antibody titers, OIE establishes a difference in titer by >4-fold to define a strain for which the antibodies have a specific neutralization reaction greater than the cross-neutralization reaction.²⁰ Comparative neutralization of the 27 VN-positive samples from among the 639 tested samples with the BVDV-1a, -1b, and -2b strains revealed the following: 3 of 27 (11.1%) had significantly higher titers against BVDV-1a than against the other subtypes, 9 of 27 (33.3%) had higher titers against BVDV-1b, and 9 of 27 (33.3%) had higher titers against BVDV-2 (Fig. 2). Six samples had a higher titer against 1 of the strains than against the other 2 strains (but <4-fold increase). Of these 6 samples, 2 had higher titers against BVDV-1a (7.4%), 4 against BVDV-1b (14.8%), and 1 against BVDV-2 (3.7%). One sample had the same titer for BVDV-1a and BVDV-2 (3.7%; Supplementary Table 1).

Figure 2.

Virus neutralization (VN) results. Antibody titers were determined in swine serum samples tested in the VN assay against BVDV-1a, -1b, and -2 strains. During cross-neutralization between 2 or 3 reference strains, highest titers (>4-fold) are considered as a significant neutralization reaction against the particular strain (outer circles); for cross-neutralization results with highest titer against 1 strain (but <4-fold), the strain for which the neutralization reaction occurred could not be determined (inner circles).

Studies involving the detection of pestiviruses other than CSFV in pigs are scarce. We tested 741 backyard swine sera for pestiviruses using RT-PCR and sequencing, followed by VN. We found no evidence of Bungowannah virus and APPV genomes. Apparently, Bungowannah virus infection is restricted to Australia.⁸ Bungowannah virus was not observed in one study in the United States,¹ in accord with our results. We also searched for APPV without success, although it has been detected previously in southern Brazil.¹¹ We speculate that the age of the sampled animals contributed to the negative results, given that APPV is associated with congenital tremors in newborn piglets.⁷

We found 2 (0.3%) pig samples to be positive for a ruminant pestivirus genome. The occurrence of positive samples in our study was lower than in previous studies given differences in sampling. Our study was conducted by random sampling representative of the backyard swine population in the State, in which clinical appearance data were not collected, whereas most surveys for these pestiviruses in swine were performed in clinically affected animals, which may have overestimated the data. A study with 511 samples from sick pigs in China detected 26.8% BVDV RT-PCR positivity,⁴ suggesting its widespread circulation in the swine herds in that country. Although previously detected in a wild boar from southern Brazil,¹⁹ a search for the BVDV genome in domestic pigs through RT-PCR has not been conducted in South America, which suggests a lack of studies on pestiviruses other than CSFV in Suidae hosts in this region.

In phylogenetic analysis, the samples that we sequenced were classified as BVDV-1d and BVDV-2a. BVDV-1d is frequently reported in cattle in Brazil,¹⁸ which could raise suspicions about the cattle origin of the SUI 1 strain given that backyard pigs usually come in contact or are raised with other animal species. We detected one BVDV-2a sample, although 2a strains are rarely reported in Brazil,⁵ but BVDV-2b is frequently found in cattle in southern Brazil.¹⁸ BVDV-2a was reported in cattle from Argentina,¹² and the sequence that we obtained had a high degree of identity with those samples.

The BVDV seroprevalence observed herein was 4.2% at the individual level and 5.6% at the herd level, similar to that in other studies performed in commercial farm pigs in Brazil, ranging from 2.4–4.7% at the animal level.⁶ In Europe, commercial farm pigs had a lower BVDV seroprevalence at the animal level using VN, especially in Norway (2.2%) and The Netherlands (2.5%).¹⁰ Higher BVDV seroprevalences in pigs were found in Asian countries, such as in South Korea,⁹ with a seroprevalence of 5.3%. Regarding Chinese swine herds,¹⁶ the epidemic status of BVDV infection was a serious issue, linking low seroprevalence rates to high biosecurity in farms and abolished mixed farming practices.

BVDV-infected cattle are speculated to be the main source of BVDV infection in swine given increased seroprevalence in pigs when cattle are reared on the same farm¹⁰; close contact with cattle farms or cattle in the same facilities were identified as risk factors for BVDV infection in swine. Rio Grande do Sul, southern Brazil, contains >13 million cattle, according to the State Veterinary Office (SEAPA-RS), and serologic studies performed on cattle from this region found BVDV seroprevalences of 56–68%.¹³ Backyard or noncommercial farms are frequently found to have ≥2 livestock species reared in the same property, which was true in the farms in our study as 96.6% of the farms were identified as raising pigs and cattle. Among the 18 swine herds with BVDV-positive serologic results, in 10 of them, the pigs and cattle had nose-to-nose contact, whereas only one farm did not rear any ruminant species. Although our aim was strictly to determine the frequency of pestivirus genomes and antibodies, sampling was performed according to predefined risk criteria for CSFV infection. Hence, it is not possible to define the risk factors related to our study.

Epidemiologic investigation of pestiviruses in heterologous host species is important to understand their impact on health, potential virus reservoirs in nature, and laboratory testing. Thus, our results reinforce the notion that Bungowannah virus, APPV, CSFV, BDV, and HoBiPeV have not been circulating in this Brazilian backyard swine population. However, there is a need to generate epidemiologic studies for pestiviruses, especially BVDV, in heterologous host species for future control programs.