Recombinant Porcine Reproductive and Respiratory Syndrome Virus Expressing Membrane-Bound Interleukin-15 as an Immunomodulatory Adjuvant Enhances NK and γδ T Cell Responses and Confers Heterologous Protection

ABSTRACT

Cytokines are often used as adjuvants to improve vaccine immunogenicity, since they are important in initiating and shaping the immune response. The available commercial modified live-attenuated vaccines (MLVs) against porcine reproductive and respiratory syndrome virus (PRRSV) are unable to mount sufficient heterologous protection, as they typically induce weak innate and inadequate T cell responses. In this study, we investigated the immunogenicity and vaccine efficacy of recombinant PRRSV MLVs incorporated with the porcine cytokine interleukin-15 (IL-15) or IL-18 gene fused to a glycosylphosphatidylinositol (GPI) modification signal that can anchor the cytokines to the cell membrane. We demonstrated that both cytokines were successfully expressed on the cell membrane of porcine alveolar macrophages after infection with recombinant MLVs. Pigs vaccinated with recombinant MLVs or the parental Suvaxyn MLV had significantly reduced lung lesions and viral RNA loads in the lungs after heterologous challenge with the PRRSV NADC20 strain. The recombinant MLVs SUV-IL-15 and SUV-IL-18 recovered the inhibition of the NK cell response seen with Suvaxyn MLV. The recombinant MLV SUV-IL-15 significantly increased the numbers of gamma interferon (IFN-γ)-producing cells in circulation at 49 days postvaccination (dpv), especially for IFN-γ-producing CD4⁻ CD8⁺ T cells and γδ T cells, compared to the Suvaxyn MLV and SUV-IL-18. Additionally, MLV SUV-IL-15-vaccinated pigs also had elevated levels of γδ T cell responses observed at 7 dpv, 49 dpv, and 7 days postchallenge. These data demonstrate that the recombinant MLV expressing membrane-bound IL-15 enhances NK and T cell immune responses after vaccination and confers improved heterologous protection, although this was not statistically significant compared to the parental MLV.

IMPORTANCE Porcine reproductive and respiratory syndrome (PRRS) has arguably been the most economically important global swine disease, causing immense economic losses worldwide. The available commercial modified live-attenuated vaccines (MLVs) against PRRS virus (PRRSV) are generally effective against only homologous or closely related virus strains but are ineffective against heterologous strains, partially due to the insufficient immune response induced by the vaccine virus. To improve the immunogenicity of MLVs, in this study, we present a novel approach of using porcine IL-15 or IL-18 as an adjuvant by directly incorporating its encoding gene into a PRRSV MLV and expressing it as an adjuvant. Importantly, we directed the expression of the incorporated cytokines to the cell membrane surface by fusing the genes with a membrane-targeting signal from CD59. The recombinant MLV virus expressing the membrane-bound IL-15 cytokine greatly enhanced NK cell and γδ T cell responses and also conferred improved protection against heterologous challenge with the PRRSV NADC20 strain.

INTRODUCTION

Porcine reproductive and respiratory syndrome (PRRS) is arguably the most economically important global swine disease, causing immense economic losses worldwide, with an estimated annual loss of $664 million in the United States alone (1, 2). The causative agent, PRRS virus (PRRSV), is an enveloped RNA virus that belongs to the family Arteriviridae in the order Nidovirales (3). The genome of PRRSV is a single-stranded positive-sense RNA molecule of approximately 15 kb, consisting of at least 10 open reading frames (ORFs): ORF1a, ORF1b, ORF2a, ORF2b, ORF3 to ORF5, ORF5a, ORF6, and ORF7 (4, 5). As is the case for all arteriviruses, the structural proteins of PRRSV are expressed from a set of 3′-coterminal subgenomic mRNAs (sgmRNAs) using the discontinuous mRNA transcription mechanism (6, 7).

Sequence analyses revealed that PRRSV can be divided into two genotypes, European type 1 and North American type 2 (8). There is also high genetic diversity within each genotype, which is often caused by mutations and recombination among strains (9). Type 2 PRRSV was systematically classified into 9 genetically distinct lineages based on the ORF5 gene sequences of 8,624 PRRSV strains (10). The Suvaxyn PRRSV modified live-attenuated vaccine (MLV) used in this study is derived from PRRSV isolate ISU-55, which was isolated in the early 1990s and belongs to genetic lineage 5 of type 2 PRRSV (8, 11, 12). A heterologous PRRSV strain, NADC20, used as the challenge virus in this study, belongs to genetic lineage 9 and shares approximately 87% amino acid sequence identity in ORF5 with the PRRSV Suvaxyn MLV.

Current commercially available PRRSV vaccines include both MLVs and inactivated vaccines with limited efficacy (13–15). MLVs are generally effective against homologous or closely related strains but are largely ineffective against heterologous strains (16). The ineffectiveness of the commercial vaccines is due mainly to the significant antigenic variations among circulating viruses and is also due to a compromised immune response induced by PRRSV upon exposure or vaccination. Since innate cytokines or costimulatory molecules are critically important in activating antigen-presenting cells (APCs) and shaping adaptive immunity, the use of these molecules as vaccine adjuvants has been explored in numerous studies (17), but none were tested for their adjuvant effects on PRRSV MLVs. Interleukin-15 (IL-15), which has been shown to promote the development and function of cytotoxic T cells and NK cells (18), is thus a good candidate to augment the immune response of PRRSV MLVs. Additionally, IL-18, a gamma interferon (IFN-γ)-inducing factor similar to IL-12, has also been reported to effectively enhance Th1 immunity and NK cell function (19, 20). Furthermore, the coding regions of bioactive IL-15 and IL-18 are both <500 bp, thus making them suitable for insertion into the PRRSV genome for more-stable expression without affecting the viability of recombinant viruses.

As a glycosylphosphatidylinositol (GPI)-anchored protein, porcine CD59 is constitutively expressed on leukocytes and is usually concentrated in the lipid raft, which is thought to function as a platform for many cell-to-cell contact events (21). In order to decrease the adverse systemic effects of soluble cytokines in circulation, we incorporated the gene sequence of porcine IL-15 (pIL-15) or IL-18 (pIL-18) fused with the GPI-anchoring signal from porcine CD59 into the PRRSV Suvaxyn MLV infectious clone. These recombinant PRRSV MLVs were rescued and used to infect porcine alveolar macrophages (PAMs), and the expression of pIL-15 and pIL-18 on the cell surface was characterized. Nursery pigs were vaccinated with recombinant PRRSV MLVs expressing cytokines, followed by challenge with a heterologous PRRSV strain, NADC20. The adjuvant effects of the membrane-bound cytokines on anti-PRRSV protective immunity, including NK cell and T cell responses, were investigated.

RESULTS

The GPI modification signal from pCD59 and the transmembrane region from influenza virus HA successfully target cytokine expression to the cell plasma membrane surface.

We generated expression plasmids in which the coding regions of mature pIL-15 or pIL-18 were individually fused with a short N-terminal signal sequence from hemagglutinin (HA) of influenza virus (influenza virus A/WSN/33; GenBank accession no. J02176.1) and a C-terminal GPI modification signal from pCD59. Additionally, we also tested the influenza virus HA transmembrane (HATM) region as an alternative to the GPI modification signal for targeting cytokine expression to the cell surface, since the HATM region, containing the transmembrane domain, the cytoplasmic tail, and a portion of the short stalk region, was shown previously to direct fusion proteins to the lipid raft of the cell surface (22–24). The resulting plasmids were designated pIL-15-CD59GPI, pIL-15-HATM, pIL-18-CD59GPI, and pIL-18-HATM, respectively. Their plasma membrane-targeting abilities in cells transfected with each of these plasmids were characterized. As expected, substantial cell surface expression of IL-15 (Fig. 1A and B) or IL-18 (Fig. 1C and D) was detected in cells transfected with plasmids containing cytokines fused to either the CD59 GPI signal or the HATM region, while only minimal surface expression of cytokines was found in cells transfected with the control plasmid pIL-15-TAA or pIL-18-TAA with only an N-terminal signal sequence preceding the cytokine genes. These results suggest that both the GPI-anchoring signal from pCD59 and the transmembrane region from HA can serve as appropriate plasma membrane-targeting sequences for the expression of selected cytokines.

FIG 1.

Characterization of the cell plasma membrane-targeting abilities of the GPI modification signal from pCD59 and the HA transmembrane (HATM) region. Porcine kidney PK-15 cells were transfected with individual expression plasmids, fused or not fused with the respective membrane-targeting sequence. Cell surface cytokines were stained with anti-IL-15- or anti-IL-18-specific antibody and analyzed by flow cytometry. (A and C) Representative histograms showing the frequencies of cells expressing membrane-bound cytokines in individual transfections. (B and D) Above-mentioned frequencies on average for each group. Asterisks indicate a statistically significant difference between the designated groups (**, P < 0.01; ***, P < 0.001).

Successful rescue of vaccine virus from a DNA-launched infectious clone of PRRSV Suvaxyn MLV.

The complete genomic sequence of the PRRSV Suvaxyn MLV was determined by using primers based on the reported partial sequence of parental strain ISU-55 of the Suvaxyn MLV and the conserved sequences of type 2 PRRSV strains. 5′ rapid amplification of cDNA ends (RACE) and 3′ RACE were used to determine the extreme 5′- and 3′-end sequences of the PRRSV Suvaxyn MLV genome. A DNA-launched infectious clone of PRRSV Suvaxyn MLV was successfully generated according to similar strategies reported previously (25). The PRRSV Suvaxyn MLV genome fused with ribozyme fragments was successfully assembled in a modified vector, pIRES-EGFP2, downstream of a cytomegalovirus (CMV) promoter, which is used for in vitro transcription (Fig. 2A). The resulting full-length clone of the PRRSV Suvaxyn MLV was designated pIR-SUV. The viable infectious MLV virus was successfully rescued from this DNA-launched infectious clone in BHK-21 cells and MARC-145 cells, as previously described (25) (data not shown). The sequence of the rescued MLV virus was confirmed by reverse transcription-PCR (RT-PCR) amplification and sequencing analysis of the supernatant of MARC-145 cells infected with the rescued Suvaxyn MLV virus.

FIG 2.

Construction of DNA-launched infectious clones of the PRRSV Suvaxyn MLV and recombinant PRRSV MLVs with foreign genes incorporated in the vicinity of the ORF1b/2 junction region. (A) Construction of a DNA-launched full-length clone of the PRRSV Suvaxyn MLV. The 5 genomic fragments, overlapping each other, all contain unique restriction enzyme sites for full-length viral genome assembly and are assembled into a modified pIRES-EGFP2 vector downstream of a CMV promoter. The resulting infectious clone of the PRRSV Suvaxyn MLV is designated pIR-SUV. (B) pIR-SUV contains a nonoverlapping region of one single nucleotide between ORF1b and ORF2a. pIR-SUV-2RE serves as a PRRSV MLV expression cassette for foreign gene insertion. The two introduced unique restriction enzyme sites, SbfI and PacI, are indicated by boldface type and underlining with italic type, followed by the introduced synthetic body TRS with flanking sequences. The conserved junction motif for the synthetic body TRS is boxed. The asterisks beneath the underlined sequences indicate stop codons of the corresponding ORFs. (C) Hairpin structure of the synthesized body TRS and its flanking region. (D) Nucleotide sequences of GFP-inserted recombinant PRRSV and cytokine-incorporated recombinant PRRSV MLVs in the vicinity of the ORF1b/2 junction region. The recombinant PRRSV MLV SUV-IL-15-CD59 contains a GPI modification signal at the C terminus, while SUV-IL-15-HATM contains the HATM region. The recombinant PRRSV MLVs SUV-IL-18-CD59 and SUV-IL-18-HATM have an additional N-terminal signal peptide sequence from HA. SUV-IL-15-TAA and SUV-IL-18-TAA contain only the cytokine-coding regions.

Successful construction and rescue of viable recombinant PRRSV MLVs expressing pIL-15 or pIL-18.

To determine whether porcine cytokines could be incorporated into the PRRSV genome, thus being expressed by the virus in the membrane-bound form, we first generated a new DNA-launched PRRSV MLV infectious clone that allows for foreign gene insertion and expression. Two unique restriction enzyme sites, SbfI and PacI, were introduced into the nonoverlapping region between ORF1b and ORF2a of the infectious clone pIR-SUV (Fig. 2B). Therefore, any foreign gene sequence can be inserted into this modified PRRSV MLV infectious clone between the SbfI and PacI sites and subsequently transcribed into sgmRNAs guided by body transcription-regulating sequence 2 (TRS2) that is embedded in ORF1b and originally responsible for ORF2a transcription. In order to compensate for ORF2a transcription, a synthetic body TRS with a flanking sequence of a total of 40 bp (Fig. 2C) was inserted immediately after the PacI site and upstream of the start codon of ORF2a. The resulting PRRSV MLV infectious clone was designated pIR-SUV-2RE (Fig. 2B). Viable infectious viruses were successfully rescued from pIR-SUV-2RE (Fig. 3A). Subsequently, we demonstrated that pIR-SUV-2RE can also be used as a PRRSV expression cassette for foreign gene insertion, as a green fluorescent protein (GFP) sequence was cloned into the vector (Fig. 2D), and a GFP-expressing PRRSV was successfully rescued (Fig. 3A).

FIG 3.

Rescue and characterization of recombinant PRRSV MLVs expressing membrane-bound cytokines. All the recombinant PRRSVs were rescued as live infectious viruses, as confirmed by an IFA with a PRRSV-specific monoclonal antibody. (A, top) MARC-145 cells were infected with passage 1 virus rescued from pIR-SUV-2RE or passage 1 virus rescued from pIR-SUV-GFP and stained with an anti-PRRSV N monoclonal antibody (SDOW17). (Bottom) GFP fluorescence in fresh cells after infection with the SUV-2RE or SUV-GFP virus. (B). Similarly, recombinant MLVs incorporated with the IL-15 or IL-18 gene were rescued and confirmed by an IFA. (C to E) PAMs were infected with the indicated viruses, followed by surface staining with anti-IL-15 or anti-IL-18 antibodies and intracellular staining of anti-PRRSV N protein at 20 hpi. (C) Representative FACS data showing that the PRRSV N-positive cells were further analyzed for the surface expression of cytokines. In the noninfected control group, frequencies of cytokine-expressing cells in the PRRSV N-negative population were calculated. FSC-A, forward scatter area. (D and E) The frequencies of cells expressing surface IL-15 (D) and IL-18 (E) in individual infections were compared with those in noninfected cells. Asterisks indicate statistical significance between the designated groups, as determined by one-way ANOVA (*, P < 0.05; ****, P < 0.0001; ns, not significant).

The nucleotide sequence of mature pIL-15 or pIL-18 with a short N-terminal signal sequence from HA as well as the above-mentioned C-terminal pCD59-GPI signal or C-terminal HATM region, respectively, was incorporated into pIR-SUV-2RE (Fig. 2D). Recombinant MLV viruses, designated SUV-IL-15-CD59, SUV-IL-18-CD59, SUV-IL-15-HATM, and SUV-IL-18-HATM, were successfully rescued (Fig. 3B). Two additional control recombinant MLV viruses, SUV-IL-15TAA and SUV-IL-18TAA, without a membrane-targeting signal (Fig. 2D), were also rescued similarly (Fig. 3B). The rescued viruses were sequenced and confirmed to be genetically stable for at least 8 serial passages in cell culture. The SUV-IL-15-CD59 and SUV-IL-18-CD59 viruses replicated at lower levels in vitro (0.5 to 1 log units lower at passages 4 to 6) than the parental Suvaxyn virus (data not shown).

Recombinant PRRSV MLVs successfully express membrane-bound pIL-15 or pIL-18 on the plasma membrane of infected porcine alveolar macrophages.

In order to determine if the inserted cytokine can be expressed by the recombinant MLVs on the plasma membrane of infected cells, PAM cells were infected with each of the recombinant PRRSV MLVs at a multiplicity of infection (MOI) of 0.1. At 20 h postinfection (hpi), surface cytokine expression as well as the intracellular viral N protein were detected by flow cytometry, as described previously (26). Infected PAMs can be divided into PRRSV N-positive or -negative cells, and the frequencies of cells expressing membrane-bound cytokines in the PRRSV N-positive population were compared (Fig. 3C). As expected, compared to noninfected cells, there was a significant increase in the plasma membrane surface expression level of IL-15 (Fig. 3D) or IL-18 (Fig. 3E) on PAM cells infected by recombinant SUV-IL-15-CD59 or SUV-IL-18-CD59. In contrast, only a few cells expressing membrane-bound cytokines were detected among cells infected with the Suvaxyn SUV-IL-15TAA and SUV-IL-18TAA viruses, which contain cytokine genes but without the membrane-targeting signal. Interestingly, the HATM region appeared to be less efficient in terms of tethering cytokines onto the plasma membrane, as shown by the results for cells infected with the recombinant SUV-IL-15HATM and SUV-IL-18HATM viruses. Given the fact that PAMs, as the natural host cells for PRRSV, rapidly die or undergo apoptosis within 24 to 30 h postinfection in vitro, in this study, we chose the GPI modification signal from pCD59, with a higher potency for expressing membrane-bound cytokines, in subsequent investigations of the recombinant PRRSV MLVs.

Recombinant PRRSV MLVs SUV-IL-15 and SUV-IL-18, but not the parental PRRSV Suvaxyn MLV, activate NK cells after vaccination.

To evaluate the immunomodulatory effects of the membrane-bound cytokines expressed by the recombinant PRRSV MLVs as adjuvants, we conducted a vaccination-challenge study with 3-week-old piglets. Groups of pigs (Table 1) were vaccinated with the parental PRRSV Suvaxyn MLV, the recombinant MLV SUV-IL-15-CD59, the recombinant MLV SUV-IL-18-CD59, and Dulbecco's modified Eagle's medium (DMEM) via a combination of intranasal and intramuscular routes. At 49 days postvaccination (dpv), the animals were challenged with a high dose of a heterologous PRRSV (strain NADC20) to determine the protective efficacy of the recombinant PRRSV MLVs.

TABLE 1.

Experiment design for the vaccination-challenge study in pigs to evaluate the immunogenicity and efficacy of the recombinant PRRSV MLVs expressing membrane-bound cytokines^a

Group	No. of pigs	Vaccination at 0 dpv (5.0 × 104 TCID50/pig)	Challenge at 49 dpv (5.0 × 105 TCID50/pig)	No. of pigs at necropsy at 14 dpc
1	8	PRRSV Suvaxyn MLV	PRRSV NADC20	8
2	8	SUV-IL-18-CD59b	PRRSV NADC20	8
3	8	SUV-IL-15-CD59b	PRRSV NADC20	7b
4	8	DMEM control	PRRSV NADC20	8
5	8	DMEM control	DMEM	8

^aPigs were vaccinated with the respective recombinant PRRSV MLVs by a combination of intranasal (half) and intramuscular (half) inoculation routes. The pigs were subsequently challenged with a heterologous PRRSV strain, NADC20, by intramuscular inoculation. PRRSV Suvaxyn MLV is the commercial vaccine virus rescued from a DNA-launched PRRSV infectious clone, pIR-Suvaxyn. SUV-IL-15-CD59 and SUV-IL-18-CD59 are recombinant PRRSV MLVs expressing plasma membrane-bound IL-15 and IL-18, respectively. TCID₅₀, 50% tissue culture infective dose.

^bOne pig died from intestinal hemorrhage unrelated to this study.

We found that porcine NK cells (CD3⁻ CD16⁺) in the circulation can be further categorized into two subgroups, CD8⁻ NK and CD8⁺ NK cells, according to CD8 expression (Fig. 4A), although the functional difference between these two subgroups is not well known (20, 27). At 7 dpv, the frequencies of NK cells were decreased in pigs vaccinated with the PRRSV Suvaxyn MLV compared to those in nonvaccinated pigs, and these decreases were observed mainly in the subpopulation of CD8⁺ NK cells (Fig. 4B). However, we did not find such a reduction of NK cells in recombinant PRRSV MLV SUV-IL-15- or SUV-IL-18-vaccinated pigs.

FIG 4.

NK cell responses after vaccination at 7 dpv and 49 dpv. PBMCs collected at 7 dpv and 49 dpv were restimulated with PRRSV NADC20. Cells were surface stained with anti-pig CD107a, CD3, CD8, and CD16 antibodies and intracellularly stained with IFN-γ antibody. (A) Gating strategies for NK cells. Porcine NK cells were pregated from PBMCs as CD3⁻ CD16⁺ cells based on lymphocyte morphology. They were further divided into CD8⁺ and CD8⁻ subpopulations. CD107a and IFN-γ expression levels were then evaluated in NK cells. Shown here are representative graphs. FSC-A, forward scatter area; SSC-A, side scatter area. (B) NK cells in circulation at 7 dpv. (C) Total IFN-γ-producing NK cells and IFN-γ-producing CD8⁺ NK or CD8⁻ NK cells among CD3⁻ PBMCs at 7 dpv. (D) IFN-γ-producing NK cells at 49 dpv. Data on the CD107a expression levels of NK cells are not shown. Asterisks indicate statistical significance between the designated groups, as determined by one-way ANOVA (*, P < 0.05; **, P < 0.01).

Additionally, at 7 dpv, there was an increase in the frequency of IFN-γ-producing CD8⁻ NK cells in the recombinant SUV-IL-15-vaccinated group compared to that in the parental Suvaxyn MLV group (Fig. 4C). Later, at 49 dpv, Suvaxyn MLV-vaccinated but not recombinant MLV SUV-IL-15- or SUV IL-18-vaccinated pigs had significantly lower numbers of total IFN-γ-producing NK cells and IFN-γ⁺ CD8⁺ NK cells than did the nonvaccinated pigs (Fig. 4D). These data suggest that the Suvaxyn MLV inhibited the proliferation and IFN-γ production of NK cells, whereas the recombinant MLVs SUV-IL-15 and SUV-IL-18 greatly ameliorated this inhibition.

Recombinant MLV SUV-IL-15 facilitates long-term γδ T cell responses and greatly increases the number of IFN-γ-producing T cells during late stages of vaccination.

To evaluate T cell responses after vaccination, we first analyzed the surface expression of CD107a as an indicator of lymphocyte degranulation. At 7 dpv, the total numbers of CD107a-expressing T cells were comparable among all groups (data not shown). However, the Suvaxyn MLV-vaccinated and recombinant MLV SUV-IL-15-vaccinated pigs had substantially increased numbers of CD107a⁺ γδ T cells compared to those in nonvaccinated pigs (Fig. 5B). We also found increased total numbers of γδ T cells in these 2 vaccinated groups (Fig. 5C), which may partially explain the enhanced CD107a expression from γδ T cells at 7 dpv. Similarly, at 49 dpv, total numbers of CD107a-expressing T cells were comparable among different groups (Fig. 5D), but only the recombinant MLV SUV-IL-15-vaccinated group had increased numbers of both total T cells (Fig. 5C) and CD107a-expressing γδ T cells (Fig. 5E) compared to those in the Suvaxyn MLV-vaccinated group.

FIG 5.

T cell responses after vaccination at both 7 dpv and 49 dpv. PBMCs were collected from each pig at 7 dpv and 49 dpv. (A) Gating strategy for T cells. PBMCs were pregated as CD3⁺ T cells and further divided into CD4⁺, CD8⁺, CD4⁺ CD8⁺ (double-positive), and γδ T cell subpopulations. CD107a and IFN-γ expression levels were evaluated in these T cell populations. Only representative graphs are shown. FSC-A, forward scatter area; SSC-A, side scatter area. (B) Different subpopulations of T cells expressing surface CD107a at 7 dpv. (C) Frequencies of γδ T cells among PBMCs at 7 dpv and 49 dpv. (D) Total CD3⁺ T cells expressing CD107a at 49 dpv. (E) Different T cell subpopulations expressing CD107a at 49 dpv. (F) Total T cells producing IFN-γ at 49 dpv. (G) Different T cell subpopulations producing IFN-γ at 49 dpv. Asterisks indicate statistical significance between the designated groups, as determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

We also evaluated IFN-γ production in T cells after vaccination. At 7 dpv, the numbers of IFN-γ-producing T cells were comparable among different groups (data not shown). In contrast, at 49 dpv, the group vaccinated with the recombinant MLV SUV-IL-15, but not with the Suvaxyn MLV or the recombinant MLV SUV-IL-18, had a significant increase in the total number of IFN-γ-producing T cells compared to other groups (Fig. 5F). Moreover, there were substantially increased numbers of IFN-γ⁺ CD8⁺ T cells and IFN-γ⁺ γδ T cells among T cells in recombinant MLV SUV-IL-15-vaccinated pigs compared to those in the Suvaxyn MLV-vaccinated pigs (Fig. 5G). These results demonstrate that the recombinant MLV SUV-IL-15, but not the recombinant MLV SUV-IL-18, greatly enhanced both the proliferation and anti-PRRSV functions of γδ T cells, as indicated by the increased levels of expression of the degranulation marker CD107a and intracellular IFN-γ production. More importantly, the recombinant MLV expressing IL-15 was able to activate IFN-γ-producing CD8⁺ T cells at a late stage of vaccination.

Recombinant MLV SUV-IL-15 enhances CD107a surface expression in γδ T cells after heterologous PRRSV challenge.

To investigate the protective immune response against heterologous challenge with PRRSV strain NADC20, we analyzed both PRRSV-specific T cells and NK cells at 7 days postchallenge (dpc). Different from the time point of 49 dpv, the numbers of IFN-γ⁺ T cells in the circulation were comparable among the groups, whereas the numbers of CD107a-expressing T cells at 7 dpc in recombinant MLV SUV-IL-15-vaccinated pigs increased significantly compared to those in Suvaxyn MLV-vaccinated pigs or to those in the unvaccinated and nonchallenged DMEM/DMEM control pigs (Fig. 6A and B). More specifically, the numbers of both activated γδ T cells (CD107a⁺ or IFN-γ⁺) and total γδ T cells in the recombinant MLV SUV-IL-15 vaccination group were elevated compared to those in the Suvaxyn MLV vaccination group (Fig. 6C and D). Interestingly, a similarly increased CD107a⁺ or IFN-γ⁺ γδ T cell response in the DMEM group was also observed. Unlike the MLV SUV-IL-15 group, however, the total number of γδ T cells in the DMEM group was not increased compared to that in the Suvaxyn MLV group or to that in the DMEM/DMEM group. There was no significant difference in either CD107a expression or IFN-γ production from T cells in the recombinant MLV SUV-IL-18-vaccinated group compared to the Suvaxyn MLV-vaccinated group.

FIG 6.

Anti-PRRSV immune responses after heterologous PRRSV NADC20 challenge. PBMCs collected at 7 dpc were evaluated for CD107a-expressing and IFN-γ-producing lymphocytes. (A and B) Total CD3⁺ T cells expressing CD107a (A) or IFN-γ (B) among PBMCs. (C) γδ T cells expressing CD107a or IFN-γ among PBMCs. (D) Frequencies of γδ T cells among PBMCs. (E) Frequencies of NK cells among PBMCs. (F) Frequencies of total NK cells expressing CD107a or IFN-γ. Asterisks indicate statistical significance, as determined by one-way ANOVA (*, P < 0.05; **, P < 0.01).

Regarding the NK cell response, the nonvaccinated challenged-only pigs had the fewest NK cells in circulation among the groups (Fig. 6E). The numbers of CD107a-expressing NK cells in this challenged-only group were also lower than those in the other groups (except for the recombinant MLV SUV-IL-18 group), even though the numbers of IFN-γ-producing NK cells were comparable among different groups (Fig. 6F). These results suggested that PRRSV NADC20, similar to PRRSV Suvaxyn MLV, likely inhibited the proliferation of NK cells and impaired their degranulating activity.

Furthermore, the total circulating cells, regardless of cell type, were also evaluated for the surface expression of CD107a and IFN-γ before and after PRRSV NADC20 challenge (Fig. 7A and B). At 49 dpv, recombinant MLV SUV-IL-15 significantly increased the number of IFN-γ-producing cells in the peripheral blood compared to those in the parental Suvaxyn MLV-vaccinated group as well as the nonvaccinated group, suggesting an overall enhanced immunogenicity by the recombinant MLV expressing IL-15 after vaccination. However, after heterologous challenge with PRRSV NADC20, there was no significant difference in numbers of IFN-γ-producing cells among the groups, while numbers of CD107a-expressing cells were increased in both recombinant MLV SUV-IL-15-vaccinated pigs and the unvaccinated challenged-only pigs.

FIG 7.

CD107a-expressing or IFN-γ-producing cells regardless of cell type in the peripheral blood at 7 dpv, 49 dpv, and 7 dpc. PBMCs were collected at the indicated time points, and all the cells were evaluated for CD107a expression (A) and IFN-γ production (B). Asterisks indicate statistical significance, as determined by two-way ANOVA (*, P < 0.05).

Pigs vaccinated with recombinant MLV SUV-IL-15 have significantly reduced numbers of lung lesions and viral RNA loads in lungs, similar to those of Suvaxyn MLV-vaccinated pigs after heterologous challenge with PRRSV.

In order to measure the replication level of the challenge virus NADC20, viral RNA loads in the sera and lung tissues were compared among different groups. Pigs vaccinated with the recombinant MLVs SUV-IL-15 and SUV-IL-18 had serum viral RNA titers similar to those of pigs vaccinated with the parental Suvaxyn MLV against heterologous virus challenge at both 7 dpc and 14 dpc (Fig. 8A and B). In the lung tissues at 14 dpc, there were lower viral RNA loads in the recombinant MLV SUV-IL-15-vaccinated group than in the parental Suvaxyn MLV- and MLV SUV-IL-18-vaccinated groups (Fig. 9A). All three vaccinated groups had significantly lower viral RNA copy numbers than those in the unvaccinated challenged-only group, indicating that the parental Suvaxyn MLV effectively reduced lung viral RNA loads against heterologous PRRSV challenge. The microscopic lung lesion scores for the recombinant MLV SUV-IL-15-vaccinated group were lower than those for the Suvaxyn MLV-vaccinated, recombinant MLV SUV-IL-18-vaccinated, or unvaccinated challenged-only group (Fig. 9B). There was no significant difference in the gross lung lesions among three vaccinated groups (data not shown). The results suggested that the recombinant MLV SUV-IL-15 exhibited enhanced protection, although not statistically significant, against heterologous challenge, as evidenced by the low lung viral RNA loads and low lung lesion scores compared to the parental Suvaxyn MLV. It is important to point out that the parental Suvaxyn MLV is a commercial vaccine; thus, any incremental improvement of the efficacy of an existing vaccine is an achievement, as significant improvement of the efficacy of a commercial vaccine is usually very challenging.

FIG 8.

Serum viral RNA loads after challenge with the PRRSV NADC20 strain at 7 dpc (A) and 14 dpc (B). Viral RNAs were extracted from serum samples at 7 and 14 dpc for the quantification of PRRSV RNA copy numbers in sera by qRT-PCR. The detection limit is calculated as 2 log₁₀ copies for each real-time PCR. Samples below the detection limit were considered negative and labeled as 2 log₁₀ copies. Asterisks indicate statistical significance between the designated groups, as determined by one-way ANOVA (**, P < 0.01; ***, P < 0.001; ****, P < 0.0001).

FIG 9.

Viral RNA loads, microscopic lesions, and cytokine expression in lung tissues after heterologous challenge with PRRSV NADC20 at 14 dpc. (A) Viral RNA loads in lung homogenates were determined by qRT-PCR. The detection limit is calculated as 2 log₁₀ copies for each real-time PCR. Samples below the detection limit were considered negative and labeled as 2 log₁₀ copies. (B) Microscopic lung lesion scores at 14 dpc. (C and D) IL-15 (C) or IL-18 (D) expression levels in lung tissue homogenates at 14 dpc were quantified by qRT-PCR and normalized to the value for the housekeeping gene YWHAZ, shown as fold changes over the values for the DMEM (unvaccinated challenged-only) group. Asterisks indicate statistical significance between the designated groups, as determined by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

Recombinant MLVs SUV-IL-15 and SUV-IL-18 express IL-15 or IL-18 in vivo.

As expected, the IL-15 expression level in the lung tissues of recombinant MLV SUV-IL-15-vaccinated pigs was significantly higher than those in any of the other groups at 14 dpc (Fig. 9C). Similarly, the IL-18 expression level in lung tissues was also significantly higher in the recombinant MLV SUV-IL-18-vaccinated group than in the other groups (Fig. 9D). These data indicate that the recombinant MLVs carrying the cytokine genes successfully expressed IL-15 or IL-18 in pigs.

DISCUSSION

PRRSV infection usually induces insufficient activation of the immune response, including a weak cell-mediated immune response and a strikingly delayed neutralizing antibody response (28). This is partially due to virus inhibition of type I interferons (29–31), tumor necrosis factor alpha (TNF-α) (32), or other innate cytokines and costimulatory molecules (26, 33, 34) that play an important role in priming APCs and in bridging adaptive immunity (28, 32, 35, 36). This immune suppression by PRRSV has remained a major challenge for developing effective MLVs against PRRSV (14, 37). Therefore, novel strategies that can improve the efficacies of PRRSV MLVs are needed for the effective control of PRRSV.

Previous studies evaluated the use of cytokines as adjuvants to improve the immunogenicity and efficacy of the vaccines, either by codelivering the cytokines along with vaccines in their bioactive form or, alternatively, by fusing their DNA sequences with certain vaccines (38–40). In this study, for the first time, we investigated the immunogenicity and the protective efficacy of recombinant PRRSV MLVs expressing membrane-bound porcine cytokines as adjuvants. We present a novel approach of using porcine IL-15 and IL-18 as adjuvants by incorporating their genes into recombinant PRRSV MLVs, in attempts to induce cytotoxic NK cells during the innate immune response as well as to enhance the adaptive T cell response. In the meantime, we also attempted to direct cytokine expression to the plasma membrane of infected cells by the recombinant PRRSV MLV, aiming to reduce cytokine diffusion into the circulation, thus avoiding excessive systemic inflammation. Furthermore, these membrane-bound cytokines are more likely to stimulate immune cells and facilitate cytotoxic lysis due to their proximity to viral antigen presentation on the same cell.

By using a DNA-launched infectious clone of the commercial PRRSV Suvaxyn MLV as the backbone, we successfully generated a panel of recombinant Suvaxyn MLVs expressing cytokines. We further demonstrated that the recombinant MLVs can readily infect MARC-145 cells and PAMs in vitro and express the cytokines of interest during virus infection. Both the GPI signal and HATM region can direct cytokine protein expression to the cell plasma membrane surface by expression plasmids. However, the GPI modification signal from CD59, but not the HATM region, can tether the cytokines onto the plasma membrane much more efficiently when they are expressed by the recombinant PRRSV MLVs in PAMs. We speculate that since the sequence from CD59 is of porcine origin, it has already been adapted for expression in PAMs and/or can be expressed by the recombinant PRRSV with a higher efficiency than the HATM region. When the recombinant MLVs were inoculated into pigs, the MLVs incorporated with the cytokine gene fused to the GPI signal expressed IL-15 and IL-18 in vivo, as demonstrated by the elevated levels of IL-18 or IL-15 expression in the lung tissues, as measured by RT-quantitative PCR (qRT-PCR) for gene transcription.

Among the circulating lymphocytes, γδ T cells are a prominent subset in pigs, especially in young pigs (27). Unfortunately, how γδ T cells respond to PRRSV infection has rarely been studied. Consistent with data from previous studies, we found that γδ T cells are the most abundant porcine T cell subpopulation by making up more than 40% of the T cells across the entire time period of the animal study (data not shown). We also found that γδ T cells were much more responsive to PRRSV infection than CD8⁺, CD4⁺ CD8⁺ (double-positive [DP]), and CD4⁺ T cells. This is reflected by the substantially expanded and activated γδ T cells at an early stage of vaccination with the Suvaxyn MLV or with the recombinant MLV SUV-IL-15. This is further demonstrated by the significantly increased number of CD107a⁺ γδ T cells in the unvaccinated challenged-only group 7 days after PRRSV NADC20 challenge.

It is also worth mentioning that CD8⁺ T cells and DP T cells are two primary T cell populations that are capable of expressing surface CD107a as an indicator of lymphocyte degranulation/cytotoxicity, followed by the γδ T cell population, in response to PRRSV infection (Fig. 5A and D and 6C). In contrast to γδ T cells, together they compose less than 30% of the total T cells (data not shown). CD4⁺ T cells, as expected, were the least capable of expressing CD107a. These observations suggested that CD8⁺ T, DP T, and γδ T cells are likely major players in lymphocyte degranulation against PRRSV infection. Whether γδ T cells exert this cytotoxicity toward PRRSV-infected cells in a major histocompatibility complex (MHC)-unrestricted manner or not is unclear. Meanwhile, NK cells were also capable of expressing surface CD107a at a considerable level compared to T cells, suggesting their important role in degranulation and cytotoxicity as well.

Early after vaccination with the parental Suvaxyn MLV, we observed significant decreases in the numbers of total NK cells and IFN-γ⁺ NK cells in pigs at 7 dpv. Similarly, numbers of total NK cells and CD107a⁺ NK cells were significantly reduced in the unvaccinated challenged-only pigs after PRRSV NADC20 challenge. NK cell-mediated cytotoxicity was suppressed in PRRSV-infected PAMs, as reported previously (41), and the results of our study reinforce the fact that PRRSV greatly impairs NK cell activation. However, the PRRSV MLVs SUV-IL-15 and SUV-IL-18 greatly restored the proliferation of NK cells and their functions, suggesting that IL-15 and IL-18 expressed by recombinant MLVs as adjuvants can enhance the anti-PRRSV innate immune response.

At a late stage of vaccination, the total number of IFN-γ-producing T cells and, moreover, the total number of cells in the circulation producing IFN-γ were significantly increased only in the MLV SUV-IL-15 group compared to any of the other groups. This suggests that the MLV SUV-IL-15 is able to elicit a long-term protective immune response with enhanced IFN-γ production. When we analyzed the different T cell subsets at 49 dpv, we noticed that the CD8⁺ and γδ T cell responses in MLV SUV-IL-15-vaccinated pigs were significantly more robust than those in Suvaxyn MLV-vaccinated pigs but were not statistically different compared to those in unvaccinated animals. This observation further implies that the parental Suvaxyn MLV may have inhibitory effects, especially on inducing long-term CD8⁺ and γδ T cell responses, whereas the recombinant MLV SUV-IL-15 was able to overcome and/or offset this suppression by enhancing IFN-γ production not only from these two T cell subpopulations but also overall in the circulation. Again, these results from vaccination with the Suvaxyn MLV substantiate the accumulating evidence for the compromised host immune response induced by PRRSV.

After heterologous challenge with PRRSV NADC20, we found increases in the numbers of both total T cells and total cells among peripheral blood mononuclear cells (PBMCs) expressing surface CD107a only in the MLV SUV-IL-15 group, compared to those in the Suvaxyn MLV or DMEM/DMEM control group. To a large extent, such an increase seems to be attributed to the activated γδ T cells. Interestingly, we also found an increased γδ T cell response in the unvaccinated challenged-only group at a level similar to that in the MLV SUV-IL-15-vaccinated group. However, we noticed that the number of total γδ T cells was not increased in this group but was substantially elevated in the MLV SUV-IL-15 group compared to the Suvaxyn MLV group or the uninfected DMEM/DMEM control. Based on these observations, we postulate that the γδ T cells typically responded, in a yet-to-be-defined manner, to the high level of PRRSV replication in pigs of the challenged-only group at 7 dpc. In fact, it has been demonstrated that porcine γδ T cells may directly respond to pathogen-associated molecular patterns (PAMPs) in the absence of APCs by recognition via Toll-like receptors (TLRs) (42, 43) and can exert antiviral functions quickly after viral infection by producing IFN-γ and perforin and upregulating MHC class II (44, 45). On the other hand, it is important to note that the MLV SUV-IL-15 enhanced both the proliferation and the antiviral function of γδ T cells at 7 dpv, 49 dpv, and 7 dpc, especially compared to the uninfected (unvaccinated and unchallenged) DMEM/DMEM control group. Therefore, whether the γδ T cell response that we observed in the unvaccinated challenged-only group at 7 dpc was sustainable and/or protective against NADC20 challenge is debatable, as γδ T cells were not profoundly proliferating in this group compared to the uninfected DMEM/DMEM control group. Clearly, further in-depth study is warranted to investigate the role of γδ T cells in anti-PRRSV immunity.

Taken together, the recombinant MLV expressing membrane-bound IL-15, but not the parental Suvaxyn MLV or the recombinant MLV expressing IL-18, is able to enhance the immunogenicity of the MLV mainly by recovering the early response of NK cells, increasing the numbers of IFN-γ-producing T cells in the long term, and facilitating a quick and robust γδ T cell response against PRRSV NADC20 challenge.

At necropsy at 14 dpc, we found that the commercial Suvaxyn MLV significantly reduced viral RNA loads in the serum and lung tissues against heterologous PRRSV NADC20. The recombinant MLV SUV-IL-15 group had lower viral RNA loads in lung tissues and lower microscopic lung lesion scores than those of the Suvaxyn MLV group and the recombinant MLV SUV-IL-18 group. We believed that this incremental enhancement of vaccine protection is attributed largely to the above-mentioned improved immune response in recombinant MLV SUV-IL-15-vaccinated pigs. These data demonstrate that the recombinant MLV expressing membrane-bound IL-15 can not only augment immunogenicity after vaccination but also confer improved protection against the heterologous PRRSV NADC20 strain. Additionally, it appears that PRRSV is generally less efficient in stimulating IFN-γ production from T cells, since the numbers of IFN-γ⁺ T cells were increased only in the recombinant MLV SUV-IL-15-vaccinated pigs at as late as 49 dpv. Therefore, it will be interesting in the future to evaluate the use of other cytokines or chemokines as potential adjuvants expressed by recombinant PRRSV MLVs.

In summary, in this study, we present a novel strategy of utilizing the discontinuous transcription mechanism of PRRSV to express membrane-bound cytokines as vaccine adjuvants upon infection with recombinant PRRSV MLVs. This is achieved by directly incorporating the cytokine genes and fusing them to a selected plasma membrane-targeting signal into a PRRSV MLV genome. Our results demonstrate that the cytokines of interest are successfully expressed on the cell plasma membrane surface by the recombinant PRRSV MLVs in vitro and in vivo. Therefore, this would potentially eliminate the need for administering soluble adjuvants along with the vaccines. The cytokine-incorporated MLVs could serve as potential vaccines with an improved immune response and cross-protection against heterologous PRRSVs.

MATERIALS AND METHODS

Cells and viruses.

BHK-21 and MARC-145 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and antibiotics. PBMCs were isolated from heparinized blood of PRRSV-negative pigs and cryopreserved for later use. Upon thawing, PBMCs were cultured with RPMI 1640 medium supplemented with 10% FBS, 25 mM HEPES, and antibiotics. The PRRSV Suvaxyn MLV used in this study (kindly provided by Zoetis Inc., Kalamazoo, MI) is a vaccine virus that was attenuated by serial passages in vitro of wild-type PRRSV strain ISU-55. PRRSV ISU-55 was isolated from lungs of a diseased pig during a PRRS outbreak in Iowa in the early 1990s (46). PRRSV strain NADC20, of genetic lineage 9 (47), was kindly provided by Kelly Lager of the USDA National Animal Disease Center, Ames, IA.

Construction of expression vectors for membrane-bound IL-15 and IL-18.

Porcine IL-15 (486 bp) and mature IL-18 (471 bp) were commercially synthesized by Integrated DNA Technologies Inc. (Coralville, IA) as gBlock fragments, designated pIL-15 and HAsp-pIL-18, respectively, both of which contain two flanking restriction enzyme sites (NheI and XhoI) at each of the two ends as well as a C-terminal Flag tag. Additionally, HAsp-pIL-18 has an engineered N-terminal signal peptide of 51 bp from HA of influenza virus A/WSN/33 (GenBank accession no. J02176.1), whereas pIL-15 retains its own signal peptide. Furthermore, the GPI modification signal from pCD59 (114 bp) and the HA transmembrane region (210 bp) were also commercially synthesized as pCD59-GPI or HATM. All the sequences for gBlock fragments and primers used in this study will be provided upon request.

To construct plasmids for expressing cytokines with the GPI signal, the cytokine gBlock fragment (IL-15 or IL-18) was digested with NheI and XhoI, and the GPI modification signal from pCD59 or the HATM region was amplified from gBlock with specific primers (primers pCD59_XhoI-F and pCD59_ NotI-R and primers HA_XhoI-F and HA_NotI-R, respectively) containing XhoI and NotI restriction sites and subsequently digested with the corresponding enzymes. These two flanking fragments were then included in a three-fragment ligation with NheI- and NotI-digested vector pIHA (26). The resulting constructs, pIL-15-CD59GPI, pIL-18-CD59GPI, pIL-15-HATM, and pIL-18-HATM, contained either a GPI modification signal or a HA transmembrane region at the C terminus. To construct plasmids for expressing cytokines without a GPI signal as controls, two primer sets with an engineered stop codon (primer set IL-15TAA-F and IL-15TAA-R and primer set HAsp-IL-18-TAA-F and HAsp-IL-18-TAA-R) were used to amplify the cytokine gBlock fragment and subsequently cloned into NheI- and NotI-digested vector pIHA. The resulting plasmids were designated pIL-15TAA and pIL-18TAA. Each of the cytokine coding regions was fused to a Flag tag prior to the membrane-targeting signal (or prior to the stop codon in pIL-15TAA and pIL-18TAA). The specificity of the rabbit antibody against porcine IL-18 or porcine IL-15 was confirmed by simultaneous staining of the Flag-tagged cytokines with a rat anti-Flag antibody (data not shown).

Determination of the full-length genomic sequence of the PRRSV Suvaxyn MLV and sequence analyses.

The sequences of the ORF2 to ORF7 genes, but not ORF1, of the wild-type PRRSV ISU-55 isolate were reported previously (8, 11). To determine the complete genomic sequence of the PRRSV Suvaxyn MLV, total RNAs were isolated from the Suvaxyn MLV virus by using Tri reagent (MRC, Cambridge, UK). Reverse transcription and cDNA synthesis were performed at 42°C for 60 min in a 20-μl reaction mixture containing 100 U of Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA), 10 mM deoxyribonucleoside triphosphate, 100 mM dithiothreitol (DTT), 1 U of RNasin (Promega, Madison, WI), and 0.5 μg of oligo(dT) primers (Invitrogen, Carlsbad, CA). A total of 10 overlapping PCR fragments covering the entire genome of the Suvaxyn MLV were amplified and subsequently cloned into a pCR-2.1 vector (Invitrogen). Three individual clones of each fragment were sequenced. The consensus sequences were assembled and used for sequence analysis utilizing Lasergene software (version 10; DNAStar Inc., Madison, WI). The sequence of the extreme 5′ end of the viral genome was determined by using the GeneRACER kit (Invitrogen, Carlsbad, CA) with two reverse primers, SV5RACE1 and SV5RACE2. The sequence of the extreme 3′ end of the viral genome was determined by nested PCR using the same forward primer, SV10F, and two reverse primers, SV3RACE adaptorT and SV3RACE primer.

Construction of a DNA-launched infectious clone of the PRRSV Suvaxyn MLV.

After determining the complete genomic sequences at the extreme 5′ end and 3′ end of the viral genome, a total of 5 overlapping fragments (AX, XP, PN, NE, and EX) with unique restriction enzyme sites were amplified from the viral cDNA of the PRRSV Suvaxyn MLV. Twenty poly(A) nucleotides were introduced immediately downstream of the 3′ end of the viral genome. Nested PCR was used to introduce the ribozyme elements to the 5′ and 3′ ends of the viral genome for the construction of the DNA-launched infectious clone, as previously described (47). Two restriction sites, AscI and XbaI, were engineered upstream and downstream of the viral genome for genome assembly. A total of 6 individual clones for each fragment were selected for sequencing, and the clone containing the consensus sequence was used for the assembly of the full-length clone. Fragment ACYC-XA from plasmid pACYC-177 was amplified by using primers pACYC-Xbalf (5′-AAACCCGATATCAAACCCTCTAGAGCCCTTCGCCCTTCCGGCTGGCT-3′) and pACYC-AscIr (5′-GGTTTCATATGGGGTTTGTTTAAACGGGTTTGGCGCGCCGGATCCTCCGGCGTT-3′), incorporating the unique restriction sites AscI, XhoI, PmeI, NdeI, EcoRV, and XbaI, which were then used to ligate the fragment upstream of the viral genome. Among the five fragments, each of the PN, NE, and EX fragments was used to sequentially replace the fragments with the same restriction enzyme sites on the modified expression vector pIRES-EGFP2. A three-fragment ligation was used to assemble fragments AX and XP into the vector using the restriction sites AscI and PmeI, resulting in the assembly of a full-length genome of the PRRSV Suvaxyn MLV, which was then used to replace the PRRSV VR2385 genome in the pIR-VR2385-CA DNA-launched infectious clone (47) with the restriction sites AscI and XbaI. The final full-length DNA-launched infectious clone of the PRRSV Suvaxyn MLV was designated pIR-SUV.

Construction of cytokine gene-incorporated recombinant PRRSV MLV infectious clones.

In order to generate the PRRSV MLV expression cassette, an upstream fragment of 3,906 bp was amplified from the original pIR-SUV infectious clone backbone with forward primer Suv-NS_F1 containing the NotI restriction site and reverse primer Suv-NS_R1 containing the SbfI and PacI restriction enzyme sites. Nested PCR was used to amplify the downstream fragment of 576 bp with two forward primers (Suv-NS F2-1 and Suv-NS F2-2) containing the PacI restriction enzyme site and a synthetic TRS of 40 nucleotides (nt) (IDT, Coralville, IA) and a reverse primer (Suv-NS_R2) containing the ScaI site. These two flanking fragments were then included in a three-fragment ligation with the NotI- and ScaI-digested pIR-SUV infectious clone to generate a new PRRSV MLV infectious clone, designated pIR-SUV-2RE, which contains two unique enzyme sites, SbfI and PacI, and a synthetic TRS that allows for foreign gene insertion and expression.

To insert pIL-15 and pIL-18 together with the GPI signal in pIR-SUV-2RE, a primer set comprising IL-18_SbfI-F1, IL-18-R1, IL-18-F2, and CD59_PacI-R and a primer set comprising IL-15_SbfI-F1, IL-15-R1, IL-15-F2, and CD59_PacI-R were used for overlapping PCR to amplify the cytokine fused with GPI from expression plasmids pIL-15-GPI and pIL-18-GPI. The amplified regions and pIR-SUV-2RE were separately digested with SbfI and PacI and subsequently ligated together to generate recombinant PRRSV MLV SUV-IL-15 and SUV-IL-18 infectious clones. The control recombinant MLVs SUV-GFP, SUV-IL-15TAA, and SUV-IL-18TAA were generated similarly with primer sets GFP-F and GFP-R, IL-15_SbfI-F1 and IL15_PacI-R, and IL18mtr_SbfI-F and IL18_PacI-R.

Rescue of recombinant PRRSV MLVs and indirect immunofluorescence assay.

BHK-21 cells were transfected with the respective DNA-launched PRRSV recombinant MLV infectious clone plasmid DNAs. The supernatant of transfected cells at 26 h posttransfection (hpt) was used to infect MARC-145 cells, as described previously (26). Transfected or infected cells were examined by an indirect immunofluorescence assay (IFA) using PRRSV anti-N monoclonal antibody (MAb) SDOW17, as described previously (48). To further verify the stability and authenticity of the introduced mutations, viral RNA from the fourth and fifth passages of the recombinant PRRSV MLVs in MARC-145 cells was extracted from the supernatants by using the ZR viral RNA kit (Zymo Research) and subjected to RT-PCR amplification (Superscript III one-step RT-PCR system) and DNA sequencing analyses of the inserted genes between ORF1b and ORF2a.

Antibodies.

The following antibodies against porcine antigens were used for multiparameter fluorescence-activated cell sorter (FACS) analysis. Anti-CD8α IgG2a (clone 76-2-11), anti-CD8α IgG1 (clone PT36A), anti-CD4 IgG2b (clone 74-12-4), anti-CD3 IgG2b (clone 8E6-2b3c), and anti-γδ T cell IgG1 (clone PGBL31A) were purchased from Washington State University (Pullman, WA). Phycoerythrin (PE)-conjugated anti-human/porcine CD107a and mouse anti-porcine CD3ε-biotin (clone PPT3) were purchased from Southern Biotech (Birmingham, AL). Other primary antibodies included mouse anti-pig CD16 IgG1 (clone G7; Bio-Rad, Hercules, CA) and peridinin chlorophyll protein (PerCP)-Cy5.5 mouse anti-pig IFN-γ (clone P2G10; BD eBioscience, San Jose, CA). The secondary antibodies used for FACS analysis, fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG2b, PE-Cy7 anti-mouse IgG1, allophycocyanin-conjugated streptavidin, and allophycocyanin-conjugated anti-mouse IgG2a, were purchased from BioLegend (San Diego, CA). Rabbit polyclonal antibodies against porcine IL-15 or pIL-18 were purchased from MyBioSource (San Diego, CA). The SDOW17 antibody against the PRRSV N protein used for IFAs was purchased from Research Technology Innovation Inc. (Brookings, SD).

Flow cytometry.

Cells were dispensed into 96-well U-bottom plates (BD Falcon, Bedford, MA) and stimulated with PRRSV strain NADC20 at a MOI of 1 in a final volume of 150 μl. As described previously (26, 49), GolgiPlug and GolgiStop (Thermo Scientific, Waltham, MA) were added at a final dilution of 1:1,000 at 12 hpi. PE-CD107a was also added at 5 μl per well and incubated for four additional hours, according to the manufacturer's instructions. Cells were washed and suspended in FACS buffer (phosphate-buffered saline [PBS] containing 2% FBS at 4°C). They were divided into different panels and incubated with anti-porcine CD3, CD4, and CD8 antibodies of different isotypes as panel 1 or separately with anti-porcine CD3 and anti-γδ antibodies as panel 2. This was followed by incubation with isotype-matched PE-Cy7-, FITC-, and allophycocyanin-conjugated secondary antibodies. As panel 3, NK cells were incubated with anti-porcine CD3, CD8, and CD16 antibodies of different isotypes, followed by isotype-matched PE-Cy7-, FITC-, and APC-conjugated secondary antibodies. After staining of these surface markers, cells were fixed with 3.2% paraformaldehyde for 10 min at room temperature, followed by permeabilization in 0.2% saponin for 10 min at room temperature. Cells were incubated with PerCP-Cy5.5 mouse anti-porcine IFN-γ for 15 min and subsequently washed with FACS buffer with 0.2% saponin. The fluorescence of each panel was assessed with a BD FACSAria II instrument.

Vaccination and challenge study in pigs.

The animal study was approved by the Virginia Tech Institutional Animal Care and Use Committee (IACUC) (approval no. 16229). A total of 40 pigs were divided into 5 groups of 8 pigs each and vaccinated with recombinant PRRSV MLVs or the cell culture medium DMEM (Table 1). At 49 dpv, all the pigs were challenged with either a heterologous PRRSV strain, NADC20, or cell culture medium as a control. PRRSV NADC20 was inoculated into pigs via the intramuscular route, as this has been widely used in studies reported previously (47, 48, 50). At 14 dpc, all pigs were necropsied. At necropsy, the gross lung pathological lesions were recorded in a blind fashion by a board-certified veterinary pathologist (T.L.). Samples of lung tissues were also collected during necropsy for histological examination in a blind fashion by a board-certified veterinary pathologist (T.L.) and quantification of viral RNA loads. Weekly serum samples and PBMCs were also collected from each pig for a total of 9 weeks.

Quantitation of viral RNA loads in sera and lung tissues.

Viral RNAs were extracted from serum samples at 7 and 14 dpc by using the ZR viral RNA kit (Zymo Research, Irvine, CA), and the total RNAs from the lung tissues were extracted by using Tri reagent (MRC, Cambridge, UK), both according to the manufacturers' protocols. The quantification of PRRSV RNA copy numbers in sera and in lung tissues was performed by qRT-PCR using a SYBR green one-step qRT-PCR kit (Bioline), as described previously (48, 50). The RNA standard used for qRT-PCR was derived from the in vitro transcription of a PRRSV full-length cDNA clone, pACYC-VR2385, by using the mMessage mMachine T7 kit (Ambion). Each qRT-PCR was performed in triplicate.

Gross pathological and histological evaluations of lung lesions.

All pigs were humanely euthanized by an intravenous overdose of pentobarbital (Fatal-Plus; Vortech Pharmaceutical Ltd., Dearborn, MI). At necropsy, the lungs were scored for gross pathology in a blind manner, as described previously (12). Five sections of lung tissues were collected from each pig, fixed in 10% neutral buffered formalin, and processed for routine histopathological evaluation. The criteria for evaluating gross pathology and histopathology have been well established and were described previously (12).

Statistical analyses.

One-way analysis of variance (ANOVA) was used to evaluate the data for statistical differences. The data were analyzed by using GraphPad Prism (version 6.0).