Porcine reproductive and respiratory syndrome virus (PRRSV) is a positive-stranded enveloped RNA virus which belongs to the genus Arterivirus, family Arteriviridae and order Nidovirales (Lunney et al., 2016). The genome size of PRRSV is about 15 kb and is organized with replicase genes located at the 5′ end followed by the genes encoding structural proteins toward the 3′ end (Snijder and Meulenberg, 1998; Dokland, 2010). The genome of PRRSV contains over 10 open reading frames (ORFs). ORF1a and ORF1b account for over two thirds of the viral genome and encode non-structural proteins that are necessary for viral replication (Lunney et al., 2016), while ORFs 2-7 encode structural proteins (Lunney et al., 2016). There are two well recognized PRRSV genotypes: Type 1, or European-like (prototype Lelystad) and Type 2, or North American-like (prototype VR-2332) (Wensvoort et al., 1991; Mardassi et al., 1994). Recently, PRRSV Type 1 and Type 2 were classified into two species in the genus Porartevirus: PRRSV-1 and PRRSV-2, respectively, in the new taxonomy (Adams et al., 2016; Kuhn et al., 2016).

PRRSV-1 and PRRSV-2 strains share approximately 60% nucleotide sequence identity and exhibit serotype differences (van Woensel et al., 1998; Forsberg, 2005). However, overall disease phenotype, gross clinical signs, genomic organization and temporal emergence are all similar between the two species (Kappes and Faaberg, 2015). Unlike other members of Arterivirus, which have relatively broad tropisms for cells of various origins (Zhang and Yoo, 2015), PRRSV infection is highly restricted to cells of the monocyte-macrophage lineage, such as porcine alveolar macrophages (PAMs) (Albina et al., 1998; Morgan et al., 2014), macrophages from the spleen, tonsils, lymph nodes, liver, Peyer’s patches and thymus, as well as peritoneal macrophages from blood and progenitor cells in bone marrow (Sur et al., 1996; Duan et al., 1997a,b; Wang et al., 2016). Moreover, bone marrow-derived dendritic cells and macrophages are susceptible to PRRSV infection in vitro as well (Chang et al., 2008; Chaudhuri et al., 2016). Generally, only PAMs in lung are considered to be the primary target of PRRSV in vivo (Albina et al., 1998; Morgan et al., 2014).

Numerous studies have demonstrated that PRRSV infection is mediated by various cellular receptors or factors (Shi et al., 2015) such as heparin sulfate (HS) (Delputte et al., 2002), vimentin (Kim et al., 2006), CD151 (Wu et al., 2014), porcine CD163 (CD163) (Guo et al., 2014), sialoadhesin (CD169) (Delputte et al., 2007), DC-SIGN (CD209) (Huang et al., 2009; Pineyro et al., 2016), and MYH9 (Gao et al., 2016). A list of receptors utilized by PRRSV was summarized as Table 1. However, only CD163 is indispensable for PRRSV infection both in vitro and in vivo (Burkard et al., 2017). In addition to PAMs, immortalized cell lines commonly used for in vitro PRRSV propagation are sub-clones derived from the African green monkey kidney cell line MA104, such as MARC-145, CRL11171 and CRL2621a. While MARC-145 cells are predominantly used in academic laboratories (Benfield et al., 1992; Meng et al., 1996). Moreover, several cell lines from various species after introduction of CD163 cDNA, such as PK-15, CRL2843, HEK293T and BHK21, have been shown to support PRRSV replication as well (Calvert et al., 2007; Delrue et al., 2010; Wang et al., 2013d).

Current control of PRRS is inadequate despite substantial efforts have been dedicated to minimize the impact of this disease. Since the first report of PRRSV in the United States in 1987, PRRSV remains one of the major challenges for the swine industry globally and continuously evolves to cause new outbreaks (Tian et al., 2007; Rowland et al., 2012). Although the first vaccine, Ingelvac PRRS^® MLV, has been commercially available and widely used for more than two decades, the prevalence of PRRSV infection in swine herds is still high and vaccination has demonstrated only limited control of PRRS (Butler et al., 2014). Indeed, this dilemma of PRRSV vaccine development has been somewhat surprising, since a number of vaccines against equine arteritis virus (EAV, another member of the genus Arterivirus) are available and are highly effective (Balasuriya and MacLachlan, 2004).

Porcine reproductive and respiratory syndrome virus can replicate in numerous cell types from a variety of species as long as the cells express functional porcine CD163 (Van Breedam et al., 2010). However, swine (except wild boar) is the only natural host for PRRSV and there is no evidence demonstrating cross-species infection. This situation is reminiscent of smallpox infection in humans or rinderpest infection in cattle, which have been successfully controlled through vaccination. Therefore, in an analogous manner, it is hoped that soon large-scale immunization with highly effective vaccines will finally eliminate PRRSV from swine herds. In fact, a highly effective vaccine of another swine RNA virus, such as C strain MLV for classic swine fever virus (CSFV), is characterized by its genetic stability and safety to pigs of all ages, as well as its ability to induce sterile immunity and provide rapid, long-lasting and complete protection against CSFV of various genotypes (Luo et al., 2014).

In 2007, the Colloquium on Prospects for Development of an Effective PRRSV Virus Vaccine was held at the University of Illinois, College of Veterinary Medicine, United States to discuss the state of current knowledge about PRRS vaccination (Rock, 2007). All attendees, including experts in PRRS, virology, immunology and vaccinology, as well as clinical veterinarians, academics and vaccine industry scientists set new standards for the next generation of PRRSV vaccines. These standards include rapid induction of immunity, protection against most currently prevalent PRRSV strains, no adverse outcomes to swine health and ability to differentiate vaccinated from infected animals (Rock, 2007). However, although almost 10 years have passed since that meeting, no new vaccine candidates are commercially available that meet all of the above criteria.

In recent years, multiple techniques have been used to develop new PRRSV vaccines and some novel discoveries have been made with the potential to transform current PRRSV vaccinology. In this review, we summarize both the progress and challenges faced by current PRRSV vaccine researchers and provide new insights to guide future efforts.

Although the origin of PRRSV is still a myth, initial outbreak of PRRS was reported nearly simultaneous both in North America (1987) and Western Europe (1990) (Goyal, 1993). Since then, this disease has rapidly emerged across the rest of the world (Lunney et al., 2016). However, since the identification and characterization of prototype PRRSVs for both PRRSV-1 and PRRSV-2, new variants of PRRSVs have constantly evolved and appeared in outbreaks with increasingly divergent and virulent phenotypes (Kappes and Faaberg, 2015).

Notably, retrospective studies have suggested that PRRSV infection existed for years in swine herds globally before the official recognition of PRRS by practitioners of swine industry. In a serologic study conducted in Ontario, Canada, screening was performed to detect PRRSV antibody-positive samples from pig sera collected between 1978 and 1982, prior to the 1987 outbreak. Among those samples, the earliest positive sample for PRRSV antibodies was identified as early as 1979, with increasing PRRSV-positive frequencies detected in later samples (Carman et al., 1995). In Iowa, United States, similar studies detected PRRSV infection in swine herds during or shortly prior to 1985 (Zimmerman et al., 1997). Moreover, similar retrospective screenings in Europe and Asia have also confirmed the presence of PRRSV in local swine herds prior to the first known outbreak and PRRSV antibody-positive ratio in tested samples increased rapidly in a very short time frame before the emergence of major clinical disease outbreaks (Zimmerman, 2003). In other words, these retrospective investigations suggested that non-pathogenic ancestral strains for both PRRSV genotypes had been circulating in swine herds in the pre-PRRS era for years before becoming pathogenic and appearing in outbreaks since 1987.

In addition to non-pathogenic PRRSV strains circulating before the first outbreak, VR2385, a virulent PRRSV-2 strain identified in the mid-1990s, was isolated from PRRSV-infected herds soon after the identification of the PRRSV-2 prototype strain (ATCC VR2332) and diverged from VR2332 about 8% in nucleotide identity (Meng et al., 1996). Lately in 1998, another atypical PRRSV strain emerged and caused high fetal mortality and abortion in vaccinated herds in the United States (Mengeling et al., 1998). Subsequently, since 2001 many virulent isolates belonging to the same group of viruses (characterized by restriction fragment length polymorphism type 1-8-4) have been identified, leading to the discovery of the highly virulent MN184 strain, which is quite distinct (>14.5% nucleotide difference) from other genotype 2 strains (Han et al., 2006). In 2006, the key event reforming the concept of PRRSV pathogenesis was the emergence of a highly pathogenic PRRSV strain (now recognized as HP-PRRSV) with a unique molecular marker (deletion of 30 amino acids in nsp2) and high mortality rate (20–100%) in sows in South China and North Vietnam (Tian et al., 2007) and later to Southeast Asia and India (An et al., 2011; Rajkhowa et al., 2015). Additionally, since the identification of genotype 2 strain NADC30 in the United States in 2008 (Brockmeier et al., 2012), NADC30-like strains (with mortality rates of 30–50%) were soon isolated across China as well (Zhou et al., 2015; Li et al., 2016). Challenge experiments with NADC30-like virus suggested that almost all commercial vaccines licensed in China confer limited protection (Bai et al., 2016).

Although less attention has been paid to PRRSV-1 as it was previously thought to be less divergent than PRRSV-2, accumulating evidence suggests that its genetic diversity is actually similar to that of PRRSV-2. Moreover, PRRSV-1 has evolved in the same direction as PRRSV-2 to become more pathogenic (Forsberg et al., 2002; Morgan et al., 2013). In 2010, the hallmark of PRRSV-1 evolution was identification of the Lena strain (isolated in 2007), a highly pathogenic strain which shares 87% nucleotide sequence identity with the PRRSV-1 prototype Lelystad virus. This is a similar degree of identity as observed between PRRSV-2 strains MN184 and VR2332 (Karniychuk et al., 2010). In spite of its similarity to the PRRSV-1 prototype virus, Lena infection cannot be fully prevented by attenuated European subtype 1 vaccine (Porcilis^® PRRS, Merck) (Trus et al., 2014). Notably, spontaneous deletion of nsp2 has also been observed in PRRSV-1 Lena strains as well (Karniychuk et al., 2010; Trus et al., 2014).

The high degree of observed variability suggests that both PRRSV-1 and PRRSV-2 are constantly evolving to adapt to existing immunity and re-emerging as new variants to cause new outbreaks continuously (Morgan et al., 2013). Although the molecular mechanism of PRRSV evolution is still not fully understood, both epidemiological and molecular evolution data point to a time of divergence between PRRSV-1 and PRRSV-2 of approximately 1980. If this estimate is correct, PRRSV has evolved at the highest evolutionary rate (on the order of 10^-2/site/year) of all known RNA viruses (with rates ranging from 10^-3 to 10^-5/site/year) (Hanada et al., 2005). Taking all information together, including outbreaks of PRRSV in vaccinated herds, epidemiological monitoring data and molecular evolutionary analysis, it appears that PRRSV is constantly evolving to cause new outbreaks and is becoming more virulent with ability to evade vaccine-induced immunity. Therefore, creating an effective vaccine to target constantly evolving PRRSV is a top priority for controlling PRRS.

Although antibodies against PRRSV were initially considered as an ineffective component of the PRRSV-protective immune response or even deleterious due to the ADE concerns (Yoon, 1995), NAs have been considered effective against PRRSV (Lopez and Osorio, 2004; Charerntantanakul, 2012). This is consistent with the observation that inactivated PRRSV vaccines that failed to induce NA were not protective, while MLV which induced NA did confer protection to vaccinated animals against homologous PRRSV challenge (Lopez and Osorio, 2004; Charerntantanakul, 2012). Moreover, PRRSV-specific antibody kinetics suggested that the onset of NA after experimental infection correlates with clearance of the virus from the circulation and tissues (Labarque et al., 2000; Robinson et al., 2015).

In addition to the fact that genetic and antigenic variability among PRRSV isolates has hampered development of effective prevention or control strategies based on antibody-mediated virus neutralization (Kim et al., 2013), the major neutralizing targets among PRRSV antigens are still controversial. However, it is known that immune responses to PRRSV isolates are strain-specific (Mengeling et al., 2003). Soon after characterization of the PRRSV genome and identification of encoded ORFs, it was postulated that GP5, the major glycosylated envelope protein encoded by PRRSV-ORF5, acts as a major inducer of NA, as had been earlier observed for related viruses lactate dehydrogenase-elevating virus (LDV) and EAV (Zhang et al., 1998). Indeed, early reports investigating PRRSV-GP5-specific monoclonal antibodies (mAbs) identified specific epitopes of GP5 that correlated with viral neutralization (Pirzadeh and Dea, 1997; Zhang et al., 1998; Weiland et al., 1999). Based on mAb screening, a liner neutralizing epitope (designated the B epitope) and a non-neutralizing epitope (designated the A epitope) of VR-2332 GP5 protein were identified (Ostrowski et al., 2002; Plagemann et al., 2002). The core sequence of the B epitope was mapped to aa37-45 of GP5 and the antibody recognizing the B epitope was consistent with neutralizing activity of sera from VR-2332-infected pigs (Plagemann et al., 2002). Meanwhile, the hypothesis that GP5 is the major neutralization target of PRRSV was further supported by identification of neutralization-resistant mutants containing amino acid substitutions in GP5 or chimeric virus containing regions of ORF5 that had been swapped among virus strains that were susceptible or resistant to NAs (Kim and Yoon, 2008; Fan et al., 2015). Moreover, recent studies also indicated that GP5 regions that are highly variable among PRRSV strains, such as aa32-34, aa38-39 and aa57-59 regions within the N-terminal ectodomain of GP5, significantly influenced the susceptibility of the mutant viruses to NA (Kim et al., 2013). Thus, numerous lines of evidence support G5 as a major target of NAs.

In another line of research, M protein encoded by PRRSV-ORF6, an unglycosylated membrane protein of 18–19 kDa, had been shown to be important in virus assembly and budding (Conzelmann et al., 1993). M protein associates with GP5 through formation of heterodimers via a disulfide bond between the N-terminal ectodomains of both GP5 and M (Mardassi et al., 1995, 1996; Wieringa et al., 2004). Because a mAb against M protein is able to neutralize PRRSV infection (Yang et al., 2000), M protein has also been studied as a vaccine candidate. Indeed, immunization of pigs with both PRRSV-GP5 and M protein expressed by M. bovis BCG strain confers a certain degree of protection that correlates with appearance of NAs (Bastos et al., 2004). Notably, a recent study demonstrated that variation of a single amino acid (Tyr-10) in M protein confers virus resistance to pig serum with broad NA activity (Trible et al., 2015). Together, all of these data imply that the M protein of PRRSV also plays a role in viral neutralization in addition to the role played by GP5.

Conversely, several mAbs reacting with antigenic regions corresponding to the putative ”major neutralizing epitope” for PRRSV have been demonstrated to possess little neutralizing activity (Van Breedam et al., 2011). Moreover, one report demonstrated that PRRSV M-GP5 ectodomain-specific antibodies from PRRSV-neutralizing serum bound to virus but did not neutralize it (Li and Murtaugh, 2012). Therefore, it appears that antibody binding to ectodomain alone is not sufficient to ensure complete neutralization of PRRSV.

For European prototype strain Lelystad, a pepscan did not identify any virus-neutralizing epitopes in E, GP5 or M, while GP2, GP3 and GP4 were shown to possess neutralizing epitopes. In fact, GP3 appears to be the major target of NA from sera of Lelystad-infected pigs (Vanhee et al., 2011). It is important to point out that GP2, GP3 and GP4 are able to form a multi-protein complex that plays an important role in viral infectivity and receptor binding (Lee et al., 2004; Wissink et al., 2005; Das et al., 2011). Furthermore, GP4 of both the North American prototype strain VR-2332 and European prototype strain Lelystad contains a viral-neutralizing epitope and might be a driving force in PRRSV evolution (Weiland et al., 1999; Costers et al., 2010; Vanhee et al., 2010). A list of the function of all PRRSV proteins and their role in viral infection and neutralization was summarized in Table 3.

To date, the mechanism of antibody-mediated PRRSV neutralization is still unclear, due to conflicting data from various research studies (see Table 1). One plausible explanation for the discrepancies lies in the variability among PRRSV isolates used for antibody production; among the numerous studies, antigenic determinants and biological properties may differ radically (Trible et al., 2015). Supporting evidence for the heterologous nature of PRRSV neutralization stems from observations that the B epitope of GP5, a major linear neutralizing epitope for VR-2332 (Ostrowski et al., 2002; Plagemann et al., 2002), is no longer sufficient for neutralization of HP-PRRSV-HuN4 after mutation of a single amino acid of G5 (aa39) that is not predicted to affect antibody recognition (Leng C.L. et al., 2012). Moreover, in a study to systematically investigate neutralization susceptibility among PRRSV isolates, the presence of previously identified neutralizing epitopes among GP3, GP4 and GP5 did not correlate with the neutralization phenotype of the corresponding PRRSV isolates (Martinez-Lobo et al., 2011). A recent study focusing on the cross-reactivity of immune responses to PRRSV suggested that CMI and total antibody responses against PRRSV are broadly cross-reactive among PRRSV-2 isolates (Correas et al., 2017). However, NA titers are specific for the challenge isolate and homologous T cell responses show a positive association with homologous NA titers (Correas et al., 2017).

Our understanding of PRRSV envelope antigens and epitopes related to viral neutralization are inconclusive. Previous reports have demonstrated that cellular receptors for PRRSV interact with various PRRSV envelope proteins (Lunney et al., 2016). However, no crystal structure information is available that describes most PRRSV envelope proteins, which makes it difficult to understand the structural basis of the virus-receptor interaction or antibody-mediated neutralization. Furthermore, most structural data or models used for visualizing PRRSV envelope proteins (such as GP5) still rely on bioinformatic analysis or comparison with related virus counterparts (e.g., EAV) to predict putative domains or structures. Consequently, the predicted ectodomain or transmembrane regions that are used to evaluate neutralization sites among various studies are not in agreement (Thaa et al., 2013; Do et al., 2016). Moreover, most studies have used artificially synthesized peptides or bacterially expressed viral proteins to mimic authentic PRRSV envelope proteins for identification of PRRSV-neutralizing epitopes. However, these tools apparently do not mimic authentic proteins in terms of membrane association or proper protein folding in PRRSV virions. Of special note, conserved conformational epitopes or epitopes requiring post-translational modification, such as glycosylation, may exist among heterogeneous PRRSV isolates and are not reliably reconstructed using many conventional research models.

The highly heterogeneous nature of PRRSV and its CD163 dependency for host cell tropism is analogous to human immunodeficiency virus-1 (HIV-1) and its receptor CD4. Therefore, it is a reasonable to expect that antibodies recognizing certain conserved epitopes may play an essential role in broad PRRSV neutralizing effects. In HIV-1 research, HIV-1-infected donor serum samples exhibiting the ability to neutralize diverse primary strains of HIV-1 were documented as early as in 1993, about 10 years after discovery of HIV (Mascola and Haynes, 2013). Since then, only a handful of neutralizing human monoclonal antibodies (hmAbs) has been reported (Burton et al., 1994; Trkola et al., 1995). Analysis of epitopes recognized by these hmAbs revealed several unique regions of HIV-1 gp120 (the major envelope protein of HIV-1) that correlate with broad neutralization, such as the CD4-binding site of gp120 (hmAb b12) and surface glycans on the outer domain of gp120 (hmAb 2G12) (Mascola and Haynes, 2013). To date, analysis of available HIV-1 broad NAs suggests that the majority of them recognize peptidoglycan moieties located in the V1, V2 and V3 regions of gp120 or glycans located in the outer domain, and mimic the viral CD4 receptor-binding region via their complementarity-determining regions rather than a linear epitope (Mascola and Haynes, 2013; Eroshkin et al., 2014). In PRRSV, little is known with regard to how glycosylation sites located in GP5 or other envelope proteins, such as GP2, GP3 and GP4, are associated with antibody recognition sites. This lack of information is partly due to the lack of structural information regarding receptor interaction regions. Furthermore, no report yet exists that describes any PRRSV antibody from swine or mice that recognizes peptidoglycan on the virion envelope. In spite of this lack of information, most studies still suggest that glycosylation sites of PRRSV envelope proteins (especially GP5) play a role in PRRSV escape, blocking, or minimization of virus-neutralizing antibody responses rather than directly functioning as potential NA targets (Jiang et al., 2007).

Meanwhile, it is highly interesting that hosts infected with HIV develop antibodies against major envelope protein surface glycans (Calarese et al., 2003). Notably, crystal structures of the antibody-antigen complex revealed that the hmAb 2G12 binds a cluster of surface glycans and was shown to adopt an unusual domain swap configuration in which the two heavy chains interact to form a large monovalent binding surface (Calarese et al., 2003). Indeed, no antibody with a similar configuration had been described before that report. Furthermore, it was unclear how such an antibody was induced in the HIV-infected individual. Since no applicable technique currently exists to generate mAb in pigs (as was done for human mAbs), it is still unknown if antibodies recognizing glycan or peptidoglycan exist in PRRSV-infected pigs. However, it is reasonable to speculate that infected pigs produce PRRSV-specific broad NAs that recognize conserved epitopes (Doria-Rose et al., 2009; Simek et al., 2009). In fact, pig sera from swine herds exposed to circulating field PRRSV strains contained PRRSV-specific broad NAs against both PRRSV genotypes (Robinson et al., 2015).

In recent years, a combination of FACS and single cell isolation techniques makes it possible to isolate a single Ig-secreting B cell from peripheral blood mononuclear cells (PBMC) and obtain the cDNA sequence encoding paired light chain and heavy chain via single cell DNA sequencing (Tiller et al., 2008; Scheid et al., 2009a,b; Pan, 2014). These techniques has been successfully used for HIV-1 broad NA screening and production of recombinant antibodies in mammalian cells (Tiller et al., 2008; Scheid et al., 2009a,b; Pan, 2014). Considering the fact that PBMCs from PRRSV-infected swine are easy to isolate, similar techniques should be applied to PRRSV research for screening of Ig-secreting B cells with broad neutralizing activity from PRRSV-infected swine. Ultimately, if conserved epitopes for broad NAs can be characterized, it will impact the development of vaccines conferring broad protection against heterogeneous PRRSV strains.

There is no agreement regarding a predominant PRRSV neutralizing target, as current data suggests PRRSV neutralization appears to be strain-specific. Therefore, other approaches have been used to broaden cross-protection of single strain-based MLVs. As the PRRSV genome is single-stranded and consists of positive-sense RNA, reverse genetics techniques have been used for construction of infectious cDNA clones. One strategy has been to generate chimeric virus by swapping gene segments encoding structural proteins from heterologous PRRSV strains. In one study, chimeric viruses containing shuffled GP3 genes from six different PRRSV strains were generated (Zhou et al., 2012). However, only one chimeric virus was able to induce significantly higher levels of cross-NAs in pigs against heterologous PRRSV strain FL-12, suggesting that shuffling of areas of a single structural protein may not be enough to induce cross-NAs against heterologous PRRSV. To support this explanation, in a modified study conducted by the same group, shuffling of both the GP4 and M genes from different parental viruses to create chimeric viruses did indeed broaden induction of heterologous cross-NAs (Zhou et al., 2013). Thus, this approach appears to be more promising than shuffling of a single ORF for inducing broad protection. Moreover, a recent study demonstrated that by shuffling even a greater number of structural genes (ORFs 3-6) from six heterologous PRRSV strains into a PRRSV-VR2385-based backbone, rescued chimeric viruses exhibiting improved cross-protective efficacy against multiple heterologous strains (Tian et al., 2017).

In addition to gene shuffling to generate chimeric virus strains, phylogenetic analysis offers another tool to understand genetic diversity of PRRSV and seek a common antigen-coding sequence among heterologous PRRSV strains. Using this approach, broadly protective candidate strains were generated to counter the extraordinary genetic diversity of PRRSV. After performing phylogenetic analysis with alignment of 59 non-redundant, full-genome sequences of type 2 PRRSV “centralized” sequences were identified that were of equal genetic distance to all 59 wild-type PRRSV strains (Vu et al., 2015). In this way, centralized sequences of these PRRSV isolates were computationally designed and synthesized to generate a novel infectious clone designated PRRSV-Con. Rescued PRRSV-Con virus from the infectious clone was viable and replicated effectively in MARC-145 cells; moreover, pigs infected with this virus exhibited expanded levels of heterologous protection (Vu et al., 2015). However, based on challenge data, PRRSV-Con was only able to confer slightly improved protection against the virulent MN184 strain. This protection was not enough to prevent disease and PRRSV-Con was itself virulent in host animals (Vu et al., 2015). Therefore, sequence optimization and attenuation of PRRSV-Con is further required to gain maximum heterologous protection and reduce host virulence [for detail of PRRSV phylogenetic analysis, please see following reviews (Murtaugh et al., 2010; Shi et al., 2010; Stadejek et al., 2013)]. In addition, a single strain containing “centralized” sequences may confer better protection, but cannot provide complete protection to numerous heterologous PRRSV strains. In summary, although PRRSV-Con is still far from being an effective vaccine candidate, the artificial design approach to generate novel PRRSV strains based on centralized sequences holds great promise as a new for future vaccine development.

Three decades have passed since the emergence of PRRS in 1987. Unfortunately, even with sustained efforts to understand PRRSV pathogenesis and vaccinology, an effective vaccine to prevent PRRSV has yet to be successfully developed. Concerns about safety of MLVs as vaccines have persisted in addition to concerns regarding the inability of MLVs to prevent new outbreaks. DNA vaccines, subunit vaccines or virus-vectored vaccines have all been tested, but their potential value as replacements for PRRSV-MLVs currently in use remains uncertain. Unfortunately, 10 years after the Colloquium on Prospects for Development of an Effective PRRS Virus Vaccine in 2007, a successful vaccine meeting all criteria set in that meeting is still not available. Therefore, single strain-based MLVs are still the only choice we have to control evolving PRRSV at the current time.

Fortunately, ongoing research has provided clues to guide creation of strategies to meet the 2007 goals mentioned above. A2MC2-P90, an attenuated strain developed via passaging in cell culture, shows promise because it shares a unique IFN-inducing phenotype with its parental strain. It also has a unique deletion in the nsp2 region that could be helpful for use as a negative DIVA vaccine marker. In addition, it is avirulent in swine and does not undergo virus shedding. Thus, A2MC2-P90 could serve as an ideal backbone for vaccine development. Alternatively, a novel hybrid PRRSV strain could be developed using DNA shuffling or artificial synthesis of PRRSV structural protein gene segments based on phylogenetic analysis (the method used for PRRSV-Con). As a further consideration, cross-protection capability of this novel hybrid against heterogeneous PRRSV could be included in the design if conserved epitopes of broad NA could be identified and characterized. Hopefully, a PRRSV vaccine designed using a combination of advanced techniques will be available that exhibits superior efficacy and safety to current vaccines.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.