One of the biggest challenges of this study was the use of archival swine tissue, stored for up to 50 years, in the molecular diagnosis and characterization of PCV2. Many clinical specimens are kept in collections world-wide, and the genetic information that this type of material may provide is extremely important for research related to human and animal health [21]. Tissue treatment with 10% formaldehyde often leads to DNA fragmentation, resulting in false negative results by degradation of target nucleic acid, and/or presence of substances that could inhibit the PCR reaction [22, 23]. For this reason, the choice of the target fragment was important in DNA amplification of ancient FFPE tissues.
The earliest detection of PCV2 from archived tissues/cell lines is from Germany in 1962, but the fragment (only 66 bp) was unfortunately outside of the target region for genotyping. In this study, PCV2 was detected in 25 samples, including 1 sample from 1967 and several samples from the late 60s and early 70s, which were the oldest FFPE tissue fragments available in the PESAGRO-RIO collection. RJ018/1967 is the second oldest positive sample in the world, and the oldest in the Americas. These results corroborate a retrospective study suggesting that PCV2 has possibly caused isolated episodes of systemic disease since 1962, long before the major epidemic from the 1990s [24].
In 2008, a genotype classification based on the nucleotide diversity in the ORF2 was accepted, and so far, 8 PCV2 genotypes have been described (a-h) based on analysis of the genomic region encoding the ORF2 capsid protein (nt 1486 − 1473) [25, 26, 27]. Among these, PCV2a, PCV2b, and PCV2d (earlier known as mutant PCV2b) are the major genotypes, with marked differences in virulence among them [2, 28]. To date, there is no PCV2 evolutionary study in Brazil, just the analysis of the genotype of small groups of animals. Of the genotypes circulating in Brazil, Vidigal et al. [5] found in 9 samples the co-circulation of PCV2a (5 samples) and PCV2b (4 samples). Cruz et al. in 2016 [6], studying 31 Brazilian swine herds, detected the genotype PCV2a in 73.3% of the plasma samples, followed by PCV2b (26.7%). To the best of our knowledge, this retrospective study is the first description of PCV2d in Brazil, and its predominance comparing to 2a and 2b.
PCV2 vaccine development has been in a state of continuous updating because of the high mutation rate of the PCV2 genome, and the emergence of new PCV2 subtypes due to selective pressures promoted also by vaccination-induced immunity escape [29]. PCV2 vaccination started in 2004 in Europe and in 2006 in the Americas [2]. Based on the criteria previously described [27] and phylogenetic analyses, the PCV2d genotype was the only genotype detected in this study, which included samples collected prior to the introduction of vaccines in Brazil (2007), demonstrating the circulation of PCV2d genotype in Brazilian herds many years before the introduction of vaccination. Similar results were obtained by Tsai et al. [30] who detected PCV2a, PCV2b, and even PCV2d genotypes in Taiwan in 2000, 2001, and 2002, also before the introduction of PCV2 vaccines in this country.
Although the DNA fragmentation of ancient FFPE has led to a limited fragment, the mutation rate found in the samples of this study (1.54 × 10− 3) are very close to those obtained by other authors based on complete capsid sequences (1.2 × 10− 3) [2, 3, 29]. Our study demonstrated the circulation of this genotype since 1967 (the oldest sequence detected in Brazil), and our results suggest that it still represents the main circulating genotype even 50 years later.
Bayesian inference has become the standard for PCV2 phylogenetic analyses [31]. In this study, the topology of the phylogeographic tree was also recovered by the network analysis. The high haplotypic diversity of PCV2 observed in this study is related to its high mutational rate, as described by several authors [29, 32, 33], even considering that the sequences obtained were partial. The large number of median vectors in the PCV2a and PCV2c clusters can be justified since PCV2a is the first reported genotype [2] and because PCV2c was first identified from pig sera collected in 1980, 1987, and 1990 in Denmark, and not long reported since then [34]. The larger number of median vectors (15 vectors) in clusters with genotypes 2b, 2d, 2f, and 2g was an interesting finding, and suggests that these genotypes share a common origin. This hypothesis is supported by the low intergroup genetic distances among these genotypes (Supplementary File 2), suggesting once again a close evolutionary relationship.
As demonstrated by the phylogeographic tree, PCV2b and PCV2d share a common ancestor, with the emergence of PCV2d possibly predating PCV2b. This is unlike what was previously assumed [3, 35]. In our analysis, the origin of PCV2d dated back to the 60s. Furthermore, some of our samples clustered on a distinct branch of the Bayesian tree (pp = 1), preceding the origin of the "modern" 2d. We are aware that this surprising topology may be a methodological bias due to the short length of our fragment and the lack of old reference samples available in the databases. However, it is worth noting that several of the samples described in this study are the oldest sequences to date, and the discovery of PCV2 circulation in the 1960s and 1970s could completely change what is known about the evolutionary history of this virus. Therefore, further studies using the whole PCV2 genome, and a more comprehensive database are needed to confirm or refute this hypothesis.
In conclusion, the technique of DNA extraction in ancient tissue made it possible to obtain PCV2 genetic material from FFPE samples up to 50 years old, expanding the information about the evolution of PCV2 in Brazil and the positive PCV2 sample from 1967, which represented the second oldest description in the world. All PCV2 samples from this study were genotyped as PCV2d, even those obtained before the outbreak of circoviruses and introduction of PCV2 vaccine in Brazilian herds. Phylogenetic analysis and haplotype network confirmed the high diversity of PCV2, as well as its high mutation rate. This information may guide further studies to clarify the routes of entry and the spread of PCV2 in Brazil and contributes to increased knowledge about the evolution of the virus.