Despite a rapid increase in the number of sublineage 8.7 virus infections in Asian countries over recent years, very little was known about the patterns of virus emergence and spread. Relying on a national long-term PRRSV surveillance project, we collected over 6,000 suspected positive samples and obtained 242 newly ORF5 sequences and 42 complete genomes belonging to sublineage 8.7 in approximately two decades and integrated them with public genomic data to form the largest collection of available PRRSV sublineage 8.7 sequences to answer the question of how sublineage 8.7 emerged, evolved, transmitted, and recombined (intra- and interlineage) in the nearly two decades 3,4,41,47. More pragmatically, we confirmed the high weight of rural swine activities and provincial distance contributing to the sublineage 8.7 spatial spread. Note that several HP-PRRSV MLVs were hastily approved for use on a remarkably emergency basis in China at that time whereas few studies focused on its potential impact on the field spanning a long temporal vaccination, we further sequenced all HP-PRRSV MLVs. As such, we found strong leaky evidence of HP-PRRSV MLVs based on a multivariate perspective, which may have restored its virulence in clinical farming.
In our nextstrain analysis of the total sublineage 8.7 clusters, although several offsets were detected in the USA and Russia, nearly the whole phylogeny trunk was located in Asia, suggesting sporadic transmission events from China to other countries and without any outbreak events identified in other regions. In addition, the classical sublineage 8.7 cluster and transition cluster have exhibited, with absolute resolution, longer branch length, indicative of comprehensive genetic divergence than the subsequently emerging HP-PRRSV cluster (Fig. 2). Nonetheless, it also explained that a huge mutation spectrum presented prior to the emergence of sublineage 8.7 HP-PRRSV, termed by “transition lineage”, suggesting the virus under greater host innate immune pressure and adaptative evolution during the early invasion period. This observation coincides with a study suggesting that the ongoing convergence of SARS-CoV-2 lineages includes multiple mutations that can enhance the persistence of diverse virus lineages during host immune recognition 48. In the dispersal history, the nextstrain result formulated a major hypothesis on how PRRSV sublineage 8.7 may be maintained in strict transmission foci. The dissemination pattern of sublineage 8.7 constitutes a connected network of Asian regions; that is, South China serves as a principal province of PRRSV maintaining and spreading not only for the Chinese population but also for the other neighboring Asian regions such as Vietnam and Thailand. As such, Thailand and Vietnam could act as second hubs of spreading sources and diffusing viral sources to neighboring countries (Laos and Cambodia). We deployed a subsampling approach in which the sampling counts of each specific region were strictly limited up to 60 to re-estimate sublineage 8.7 phylogeographic analysis to test this hypothesis, which supported the major results, indicative of the likely accurate estimation of regional spread in Southeast Asia (Supplementary Fig. 5). As well, the epicenter role of Guangdong was also corroborated by our GLM analyses in China. In our ensuing Bayesian discrete phylogeographic result, we accurately estimated the early transmission link from Guangdong to nearby provinces (Guangxi) and central China, such as Henan and Hubei, with strong Markov jump support. Similarly, He et al also has proved the epicenter role of Guangdong with another important porcine virus i.e., PEDV using Bayesian discrete analysis, with less weight compared with that in PRRSV 17. This study also successfully linked the swine industry trade and pork consumption with PEDV spread in China in their GLM extensions. In our GLM model, we found strong support for provincial distance as well as demographic factors such as population amount at origin, pork sale at rural, and per capita disposable income at destination to PRRSV spread in China. We estimated this variation may be attributable to the host infectivity heterogeneity (PEDV: piglets; PRRSV: boar and pregnant sow) and different transmission capabilities between PRRSV and PEDV.
Recombination occurred with ubiquity as a result of virulence enhancement, host shifting, and adaptability strengthening. Likewise, PRRSV recombination is also significant and pervasive in that it largely enhances genetic diversities and reduces the cross-protection of vaccines. In this study, we systematically analyzed the intra- and interlineage recombination of PRRSV lineage 8 with its temporal dynamics and found a principal recombination wave spanning 2014 to 2016. As we know, frequent homogeneous RNA viral recombination is the major result of random template conversion during replication and is thought to be deployed by the “copy-choice” mechanism of RdRp. In such a long-term evolution, any probability distributions may turn to be tendency events if such stochastic recombination could be conducive to viral survival. Although high-level recombination existed among intra- and interlineage, we found that interlineage recombination was more targeted to structural protein regions (GP3-GP5), whereas intralineage recombination was more concentrated on non-structural protein regions (ORF1a), with specifically the case for breakpoint at nsp2-nsp5, which mainly involved into antagonizing with host innate immune systems such as deubiquitin, IFN antagonist and membrane modification 1. Besides, such a great difference among the number of inter- and intralineage recombinations may be involved with the flush vaccination of lineage 8 MLVs. Until now, lineage 8 possessed the largest amounts of approved MLV vaccines of PRRSV in China. Since all PRRSV MLV could continue to replicate in the host, the “copy-choice” characteristic of RNA polymerase offers a possibility to recombine with field strains in host. On the other hand, it should be noted that China currently possesses only L5 lineage vaccines, derived from the VR2332 lineage, as well as L8 lineage vaccine strains. The extensive use of L8 lineage vaccines significantly outweighs that of L5 lineage vaccines, thereby elevating the probability of genetic recombination occurrences. Hence, lineage 8 MLV vaccines may gain more possibilities to recombine with field strains. However, it is under an intricate field that we may not interpret, with unilateralism, the erratic phenomenon just in terms of one aspect.
Our study provided the first exploration of quantifying how the MLVs are likely to affect immunized herds under the field. By multiple independent phylogenetic reconstructions and recombination elimination, we have identified four MLV groups characterized with evolutionarily homogeneity. We inspected the temporal signal of potential descendants within each group. Each strain in the JXA1-R, TJM-F92, and HuN4-F112 groups coincidentally supported our scenario whose prerequisite was that the time of vaccines approved is prior to the prevalence of associated field isolates. However, the temporal signal and the haplotype analysis of GDr180 cluster was on the contrary, showing temporal irrelevance between GDr180 and field isolates. We hypothesized that it is partly due to that GDr180 is the latest approved MLV vaccines (2015), as such, it has been vaccinated with relatively low frequency in the field. It is less impacted then in the clinic and would still continue to be monitored. In the other three HP-PRRSV MLV vaccines, JXA1-R, the most frequently administered HP-PRRSV vaccine and mandatory immunized before 2017 in China, were also the vaccines that pose relevance with the most numerous field strains. It, in turn, reflected the impact that MLV vaccines brought to the field, and vice versa. Given that the JXA1-R-homogeneity strain, KU842720/Hanvet1/Vietnam, was detected under the context of without approved JXA1-R vaccine importing in Vietnam, the consistently spatial spreading scale with our estimation of sublineage 8.7 national transmission further highlighted the significance of continuous monitoring and restrict quarantine underlying cross-regional livestock trading. Although multiple approaches such as infectious clones and challenge experiments have been attempted previously, the common characteristic of these results related to reversion sites was merely appropriate for specific cases whereas we are still in the paucity of amino acid markers from MLV supported by comprehensive clinical whole genome data. Our results firstly showed several common amino acid substitution positions spanning whole-genome scale, which may be associated with HP-PRRSV MLV reversion markers albeit specific molecular markers varied with different vaccine clusters. These results should be invaluable hints and be helpful when it comes to potential vaccine reversion cases and potential vaccine escape mutants and other potentially problematic variants.
In summary, we constituted the largest dataset to reconstruct sublineage 8.7 spatial dynamics, its associated ecological, demographic, and swine-farming practices implication, and potential leaky evidence of HP-PRRSV MLVs. Given that the phylogeographic studies in temporospatial dynamics of porcine-associated infectious diseases increase exponentially, further studies that integrate the spread patterns of different pathogens and investigate how and why they may vary against different objects such as PEDV and PRRSV are invaluable for effective targeted control of swine pathogens. Importantly, as PRRSV and SARS-CoV-2 are two important members of Nidovirales order, the long-term clinical data of PRRSV MLVs immunization can be a good reference for SARS-CoV-2 vaccine safety evaluation and future RNA virus MLV development.