Introduction

Porcine reproductive and respiratory syndrome (PRRS), characterized by reproductive failure in pregnant sows and respiratory distress in pigs of all ages, continues to be one of the most economically devastating diseases in the pig farming industry worldwide since it emerged in North America in 1987 [16]. PRRS virus (PRRSV), the causative agent [22] of the disease, is an enveloped virus containing a 15.4-kb positive-strand RNA genome [23] and is a member of the family Arteriviridae within the order Nidovirales [3]. The viral genome contains the following nine open reading frames (ORFs): ORF1a and ORF1b encode viral replicase polyproteins, while ORF2a, ORF2b, and ORFs 3–7 encode the viral structural proteins GP2, E, GP3, GP4, GP5, M, and N, respectively [23, 33, 38]. There are two PRRSV genotypes: North American and European (NA-type and EU-type, respectively) [27].

In April 2006, a highly pathogenic PRRSV (HP-PRRSV) emerged in China and spread quickly throughout most of its provinces. Infection with this virus led to high fever, high morbidity (50–100 %), and high mortality (20–100 %) in pigs of all ages and caused more than two million deaths [34, 35]. HP-PRRSV also affected pigs in Vietnam, Thailand, and some other Southeast Asian countries [1], causing great concern to the global swine industry. The molecular mechanism underlying the significantly enhanced pathogenicity of HP-PRRSV compared with the classical PRRSV isolated in China before 2006 remains elusive.

Macrophages, as important components of the innate immune system, play key roles in clearing microbial pathogens and maintaining host homeostasis through their scavenger-like abilities. During a viral infection, specialized proteins expressed by host cells recognize viral nucleic acids or proteins, thereby triggering signaling pathways within the infected cell. PRRSV has a strict tropism for cells of the monocyte/macrophage lineage. Lung-resident porcine alveolar macrophages (PAMs) are the primary target cells for PRRSV infection in vivo [32], although macrophages isolated from the spleen, tonsils, lymph nodes, and thymus can also support productive PRRSV replication [9, 36]. Because of the importance of PAM-PRRSV interactions during PRRSV infections, investigating PAM responses may help to unravel how HP-PRRSV causes increased pathogenicity of the virus in pigs. In particular, we hypothesize that investigating which PAM genes are differentially expressed before and during infection with HP-PRRSV might help us to learn more about the molecular basis of the disease. Genini et al. previously evaluated the gene expression profile of innate immune genes from PAMs infected in vitro with the European PRRSV Lelystad strain using a porcine Chip microarray and showed that a few innate immune genes were upregulated, including the antiviral cytokine type I IFN-β gene [10]. Because the biological effects of the European strains are distinct from those of the North American isolates, in a separate study, the transcriptomic changes in PAMs infected with the North American PRRSV VR2332 strain were also evaluated using serial analysis of gene expression methodology, and very little or no change was observed in the innate immune genes at any time point following infection [25]. However, no study has included a comprehensive analysis of the innate immune changes induced specifically within PAM cells upon infection with the highly pathogenic strain.

In this study, we constructed two cDNA libraries and used the suppression subtractive hybridization (SSH) method to investigate differentially expressed genes in HP-PRRSV-infected or uninfected PAM cells and identified some differentially expressed genes in vitro and in vivo.

Materials and methods

Ethics statement

Animal experiments were approved by the Harbin Veterinary Research Institute of the Chinese Academy of Agricultural Sciences. All procedures were conducted in accordance with animal ethics guidelines and approved protocols. The Animal Ethics Committee approval number is Heilongjiang-SYXK 2006-032.

Cells and virus

PAMs were obtained under sterile conditions from 6-week-old PRRS antibody- and virus-negative Large White/Landrace cross piglets as described previously [12, 37]. The cells were cultivated in RPMI 1640 medium supplemented with 10 % fetal bovine serum (Gibco BRL, Gaithersburg MD, USA) and 100 IU of penicillin and 100 mg of streptomycin per mL in a humidified 5 % CO2 atmosphere at 37 °C. The HP-PRRSV strain HuN4 (GenBank no. EF635006), isolated in 2006 in China, was maintained in our laboratory and was confirmed to be highly pathogenic to piglets [35].

Construction of the SSH cDNA library

To determine the appropriate time point for extraction of cellular RNA, we investigated the replication kinetics of HP-PRRSV in PAM cells using PRRSV-specific quantitative real-time RT-PCR. The primers and probe (forward, 5′-AAATACAACGGCAAGCAGC-3′; reverse, 5′-AGACACAATTGCCGCTCACTA-3′; probe, FAM-CAAAACCAGTCCAGAGGCAAGGG-BHQ1) correspond to the conserved region of the PRRSV ORF7 gene. PAM cells were freshly obtained from a donor pig, seeded onto a six-well plate, and allowed to adsorb to the plate at 37 °C for 8 hours. The cells were then infected with the HuN4 virus at a titer of 1000 copies/well. As an uninfected control, cells were mock treated with sterile PBS and processed exactly like the infected cells. Cells were harvested for RNA extraction every 12 hours from 0 to 96 hours postinfection (hpi) from both the uninfected control and the PRRSV-infected cultures. Real-time RT-PCR was performed using an Mx3000P Real-Time PCR System (Genetimes Technology Inc.). The total volume of each reaction was 25 μL, and the concentration of each primer was 100 nM. A hot-start enzyme was used with the following PCR conditions: 50 °C for 2 min, 95 °C for 10 min, and 40 cycles of 94 °C for 30 s, 53 °C for 1 min, and 72 °C for 30 s.

Total RNA from HP-PRRSV-HuN4-infected or control PAM cells at 40 hpi (just before the peak of viral replication at 48 hpi) were extracted using an RNeasy Plus Mini Kit (QIAGEN, Valencia, CA). cDNA libraries were constructed by SSH using a PCR-Select™ cDNA Subtraction Kit (Clontech, USA) following the manufacturer’s protocol. Briefly, mRNA (2 μg) from control or HP-PRRSV-infected cells was reverse transcribed into cDNA. For the forward subtraction, cDNA from the control cells was used as the driver, and cDNA from the HP-PRRSV-infected cells was used as the tester for identifying upregulated gene expression. For the reverse subtraction, cDNA from the control or HP-PRRSV-infected cells was applied as the tester and the driver, respectively, to detect downregulated gene expression. All procedures, including cDNA synthesis, Rsa Ι digestion, adaptor ligation, first and second hybridizations, and first and second PCR amplifications, were performed according to manufacturer’s instructions. The PCR products obtained from both the forward and reverse subtractions were cloned separately into a pGEM-T vector (Promega) and then used to transform E. coli DH5α cells to establish different cDNA libraries. The transformed bacteria were plated onto Luria-Bertani (LB) agar plates containing ampicillin, isopropyl-β-D-thiogalactopyranoside and X-gal. After overnight incubation at 37 °C, recombinant white colonies were randomly selected and cultured in LB broth containing ampicillin, followed by PCR identification.

Screening cDNA libraries by Southern blot analysis

Approximately 600 individual recombinant clones were randomly chosen, and each individual clone was PCR amplified using the same nested primers, followed by analysis on a 2 % agarose gel. PCR products over 100 bp were selected, and 200 individual clones from each library were analyzed by Southern blotting. Each sample from the forward- or reverse-subtracted cDNA libraries was spotted onto two nylon membranes, one for each membrane, in the same position. The avian reticuloendotheliosis virus (REV) gp90 gene acted as a non-porcine-derived control. For denaturing purposes, each kit-purified (Watson, Shanghai, China) PCR product was mixed with the same volume of 0.6 N NaOH, denatured for 10 min at 100 °C, and then chilled on ice. Each denatured product was individually blotted onto a Nytran® nylon membrane (Schleicher Schuell, Keene, NH) in duplicate and fixed. DIG-labeled cDNA probes from either the forward or reverse subtractions were prepared separately using a DIG Random Labeling and Detection Kit II (Jishide, Wuhan, China) according to the manufacturer’s protocol. The duplicate membranes were then hybridized with either of the two probes, and the gray density on the hybrid membranes was measured. Densitometry scanning with a Disk Scanning Biological Fluorescent Microscope (Olympus IX2-DSU) followed by analysis with Image-Pro Plus 5.0 software (Media Cybernetics, USA) was used to quantify the density value of each sample dot. Samples hybridized to the tester probe exhibiting values of two-fold or greater over the control as determined by densitometry were considered positive and were selected for further analysis.

DNA sequencing and gene annotation

Positive samples from the Southern blot analysis were sequenced using an ABI PRISM 3370 DNA Sequencer at Invitrogen Inc. (Shanghai, China). The sequences generated were subjected to Basic Local Alignment Search Tool (BLAST) analysis using various online databases to identify the clones, as described previously [11]. These databases included the National Center for Biotechnology Information nonredundant (NR) database (NR GenBank, EMBL, DDBJ, and PDB) and the EST database (NR GenBank, EMBL, and DDBJ EST divisions). Clone redundancy was determined by sequence alignment using Lasergene software (DNASTAR, Madison, WI). Classification of gene function was done using the DAVID bioinformatics resource [8, 13, 14], and gene interactions were analyzed using Cytoscape software [6].

IL-8 quantification by ELISA

For quantitative determination of IL-8, the cell culture medium from HP-PRRSV-HuN4-infected or noninfected PAM cells was collected at 6, 24, 48, and 72 hpi. IL-8 was detected using a commercial swine IL-8 ELISA kit (Invitrogen) according to the manufacturer’s instructions. The absorbance was read at 450 nm. The process was repeated three times. Data were analyzed using unpaired Student’s t-tests and expressed as the mean ± SD (pg/mL).

Identification of Mx1 and HSP70 by western blot

Western blot analysis of Mx1 or HSP70 protein was carried out using PRRSV-infected PAM cells. The cells were harvested using lysis buffer containing 150 mM NaCl, 1 mM CaCl2, 25 mM Tris-HCl (pH 7.4), and 1 % Triton-X-100 supplemented with a protease inhibitor cocktail (Sigma). Total cell extracts (50 μg per lane) were separated on a 12 % SDS polyacrylamide gel and transferred onto a PVDF membrane. The Mx1 protein was visualized with a polyclonal goat anti-Mx1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at a 1:1000 dilution, and the HSP70 protein was visualized with a mouse anti-HSP70 monoclonal antibody (Abcam) at a 1:100 dilution. Antibody–antigen complexes were detected with horseradish peroxidase-labeled rabbit anti-goat IgG (Sigma) at a 1:8000 dilution. β-actin was detected with an anti-β-actin antibody (Sigma) at a 1:5000 dilution, and the bound antibodies were detected with horseradish-peroxidase-labeled goat anti-mouse IgG (Sigma) at a 1:5000 dilution. The membranes were incubated with ECL Western Blotting Substrate (Pierce, Rockford, IL, USA) and analyzed using the ImageQuant 350 imaging system (GE Healthcare).

mRNA analysis of differentially expressed genes in infected piglets

Fifteen 6-week-old PRRS-antibody-negative and PRRSV-negative pigs were randomly divided into two groups. Pigs in group 1 (n = 6) were inoculated intramuscularly (1 mL) and intranasally (2 mL) with the PRRSV HuN4 strain (3 × 104.0 TCID50 per pig). Uninfected control pigs in group 2 (n = 9) were inoculated by the same routes with the same volumes of Dulbecco’s minimal essential medium (Gibco BRL, Gaithersburg MD, USA). Animals were kept in Biosafety Level 2 conditions and housed separately. Clinical symptoms of infection, including reduced appetite, coughing, and respiratory distress, were recorded daily, as were rectal temperatures. Serum samples were collected at 0, 3, 5, 7, 11, and 14 days postinfection (dpi) and were used to detect viral titers by real-time RT-PCR as described above.

To obtain PAM cells from infected piglets, the piglets were euthanized at different dpi. In the infected group (n = 6), three piglets were euthanized at 7 dpi, and the remaining three at 14 dpi. In the uninfected group (n = 9), the piglets were euthanized at 0 dpi, 7 dpi and 14 dpi, respectively, using three piglets for each time point. The right lung of each pig was used to prepare PAM cells to detect up- and downregulated genes. The left lungs were harvested, fixed in 10 % neutral-buffered formalin, and routinely processed for histological examination.

The mRNA expression levels of differentially expressed genes were analyzed by real-time RT-PCR. PAM RNA was extracted and reverse transcribed into cDNA using oligo(dT) primers and a commercial Primescript RT reagent kit (TaKaRa, Dalian, China). RNA expression was quantified using a Perfect Real Time quantitative RT-PCR kit (Takara) according to the manufacturer’s instructions. For each gene detected, double standard curves were established: one curve was made using the glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) housekeeping gene (used to normalize the amount of cDNA present in each PCR experiment), while the other curve was made using gene-specific primers (Table 1). Real-time PCR was performed using an Mx3000P Real-Time PCR System (Genetimes Technology Inc.) as follows: 50 °C for 2 min (initial step), 95 °C for 5 min, then 40 cycles at 94 °C for 30 s, 53 °C for 1 min, and 72 °C for 30 s. Relative changes in gene-specific mRNA expression levels compared with the uninfected control and normalized against GAPDH were quantified by the 2−ΔΔCt method [19] using the instrument settings.

Table 1 Primers used for real time RT-PCR detection of differentially expressed genes
Full size table

Results

HP-PRRSV replication kinetics in PAM cells in vitro

The replication kinetics of HP-PRRSV in PAM cells were evaluated to identify the appropriate time point to test for changes in gene expression. Replication kinetics analysis showed that the viral copy number gradually increased from 0 to 48 hpi, reached its peak at 48 hpi, and then decreased over the period of 48 to 96 hpi (data not shown). Virally induced cytopathic effects were obvious in cultured PAM cells at 48 hpi, and these effects continuously increased in severity until 60 hpi, at which point the infected cells began detaching from the cell culture plates with increased frequency. Based on these results, we selected the 40 hpi time point for analysis to enable total RNA to be extracted from living PAM cells for SSH library construction just before the peak of viral infection at 48 hpi.

cDNA library construction and screening

During SSH library construction, each step was validated by PCR or agarose gel electrophoresis. Southern blotting was used to evaluate the 200 samples obtained from each of the forward- and reverse-subtracted libraries (Fig. 1); the positive and negative controls produced the expected results. From the 400 samples, 358 positive clones exhibiting a twofold or greater gray density value in comparison to the control were identified with ~90 % subtraction efficiency (358/400); 183 and 148 clones were derived from the forward and reverse subtraction library, respectively, and all positive clones were sequenced.

Fig. 1
figure 1

Southern blot analysis of differentially expressed genes (partial genes). The denatured PCR products of differentially expressed genes were dotted in duplicate onto two pieces of nylon membrane at the same position. The membranes were analyzed using DIG-labeled cDNA probes from either the (a) forward or the reverse (a′) subtraction libraries. The unlabeled subtraction libraries that served as positive controls were dotted onto the four corners of the membrane (◆, or ◊, according to which DIG-labeled probe was used). The gp90 gene from the avian reticuloendotheliosis virus (REV) was used as a non-pig-derived negative control (△)

Full size image

Gene annotation and assignment to categories of similar gene function

To identify the up- or down-regulated genes that were isolated by SSH screening, the 358 sequences were imported into GenBank to search for homology to known genes by BLAST analysis. Of the 183 clones from the forward subtraction library, 151 were identified by their homologous sequence, 15 matched EST sequences from unknown genes, and 17 did not match any homologous sequence. Of the 148 clones from the reverse subtraction library, 101 were identified by their homologous sequence, 39 matched EST sequences from unknown genes, and 12 did not match any homologous sequence.

The frequencies of the differentially expressed genes in the forward- and reverse-subtraction libraries were evaluated (Table 2) and were classified into eight categories by gene function: metabolic, immune-related, cell signaling, transcription/translation, cytoskeleton regulation, cell adhesion and migration, ribosomal proteins, apoptosis/cell proliferation/cell differentiation regulators, and other genes (Table 2). The proportion of genes in each category was also considered, and the immune, metabolic, and cell signaling gene categories together represented approximately 70 % of the differentially expressed genes within these top three gene families (Fig. 2a). After PRRSV HuN4 infection, 40 % of immune-related genes were upregulated in PAM cells; additionally, infection also induced expression of cell signaling, metabolic, and adhesion/migration genes (Fig. 2b). In contrast, PRRSV HuN4 infection downregulated genes related to ribosomal proteins, cytoskeleton regulation, and apoptosis/cell proliferation/cell differentiation (Fig. 2c).

Table 2 Genes that were differentially expressed between uninfected PAM cells and those infected with HP-PRRSV strain HuN4. The genes were functionally classified according to DAVID bioinformatics resource [8, 13, 14]
Full size table
Fig. 2
figure 2

Functional classification of differentially expressed genes. Genes were classified into eight groups based on their functional similarity, including metabolic, immune-related, cell signaling, transcription/translation, cytoskeleton regulation, cell adhesion and migration, ribosomal proteins, apoptosis/cell proliferation/cell differentiation regulation, and other genes. (a) Overall changes in the differentially expressed genes. Among the eight categories, the top three changes in gene expression were in the metabolic, immune-related, and cell signaling categories. (b) Upregulated genes. Among genes with upregulated expression, immune-related genes were the most common. Only five categories exhibited upregulated gene expression, excluding those whose function is unknown. (c) Downregulated genes. Among genes with downregulated expression, ribosomal, cytoskeleton, and apoptosis/cell differentiation genes were inhibited in PAM cells after infection with highly pathogenic PRRSV in vitro

Full size image

IL-8 production is increased in PAM cells upon HP-PRRSV HuN4 infection

Because macrophages are capable of producing IL-8, using ELISA, we investigated whether PAMs were able to upregulate IL-8 production in response to HP-PRRSV HuN4 infection. Indeed, the IL-8 concentration was significantly higher than that of the uninfected control in PAM cells after infection (p < 0.01), reaching 1113 pg/mL at 48 hpi and 1437 pg/mL at 72 hpi (Fig. 3).

Fig. 3
figure 3

IL-8 protein expression. PAM cells were infected with the highly pathogenic PRRSV HuN4 strain. Cell supernatants were collected and assayed using a commercial ELISA IL-8 kit. Data are expressed as the mean ± SD (n = 3). **, significant increase of IL-8 in HuN4-infected PAM cells compared with the uninfected control (p < 0.01)

Full size image

Expression of innate immune Mx1 and HSP70 antiviral proteins is increased in PRRSV-infected PAM cells

Expression of antiviral proteins known to be induced within macrophages under viral infection conditions was also analyzed by western blotting. mAbs targeting Mx1 or HSP70 were used to analyze protein expression over time in uninfected control cells or PRRSV-HuN4-infected PAM cells. Mx1 (72 kDa) was detected in uninfected PAM control cells and infected PAM cells, but its expression level increased upon PRRSV HuN4 infection (Fig. 4). Similarly, the HSP70 protein (70 kDa) expression level also increased after PRRSV infection. Although the expression levels of Mx1 and HSP70 showed a tendency to decrease at the late stage of infection, the expression level in the infected cells was higher than that in the uninfected control cells.

Fig. 4
figure 4

Western blot analysis of Mx1 and HSP70 protein expression. PAM cells infected with 100 TCID50 of PRRSV HuN4 were collected at 6, 12, 24, 48, and 72 hpi to analyze the changes in Mx1 and HSP70 protein expression under infection conditions. Compared with the uninfected control, Mx1 and HSP70 expression increased in the PRRSV-infected PAM cells. The molecular weights of Mx1, HSP70, and the β-actin loading control were 72 kDa, 70 kDa, and 42 kDa, respectively. hpi, hours postinfection

Full size image

mRNA expression level changes of differentially expressed genes in PAM cells from HP-PRRSV infected piglets

To determine if similar patterns of gene expression were detectable in lung-resident PAM cells after in vivo infection with HP-PRRSV, we inoculated piglets with the PRRSV HuN4 strain and evaluated signs of infection in the piglets as well as gene expression levels in PAM cells harvested directly from the infected piglets at various time points after infection. Although all piglets had normal body temperatures before the experimental infections were administered, all infected piglets exhibited fever (40.5–41.6 °C) beginning at 2 dpi (Fig. 5a) and showed signs of anorexia, coughing, skin cyanopathy, shivering or lameness. One infected piglet died at 11 dpi, and the remaining piglets were euthanized at 14 dpi. Using quantitative RT-PCR analysis of the serum samples that were harvested, viremia was found to start at 3 dpi (Fig. 5b). Compared with the lungs from uninfected piglets (Fig. 5c), severe proliferative interstitial pneumonia and massive lymphomononuclear infiltration was evident in the lungs of infected piglets (Fig. 5d). These results showed that piglets infected with the PRRSV HuN4 strain experienced the effects of severe illness.

Fig. 5
figure 5

Clinical and pathologic features of HP-PRRSV-infected and uninfected control piglets. (a) Mean rectal temperature (°C) of the piglets. (b) Mean PRRSV titers (expressed as viral RNA copies per mL) in serum estimated by real-time RT-PCR analysis. (c–d) Histopathological examination of lung tissue from an uninfected control (c) and an HP-PRRSV-infected pig (d) at 14 dpi. Lungs from HP-PRRSV-infected piglets displayed signs of robust interstitial pneumonia with thickening of the alveolar septa accompanied by extensive infiltration of lymphomononuclear cells. dpi, days postinfection

Full size image

From the differentially expressed genes that were identified in the in vitro infection experiment, we selected 21 representatives to test in the PAM cells harvested from the in vivo infections, including 14 upregulated and seven downregulated genes. mRNA expression of the 21 genes was analyzed by real-time PCR. Among the selected 14 upregulated genes, 13 were exposed at higher levels in PAM cells from the infected piglets than from the uninfected mock control at both 7 and 14 dpi; only ferritin heavy polypeptide 1 (FTH1) was reduced, which contrasts with the in vitro result. Quantitatively, the differentially expressed genes exhibited changes in the 1- to 5-fold range in the infected pigs compared with the uninfected controls (Fig. 6).

Fig. 6
figure 6

mRNA expression levels of differentially expressed genes in vivo determined by quantitative real-time RT-PCR at 7 and 14 dpi. Data represent the average values from three animals. Up- or downregulated genes were detected together with the internal reference gene, GAPDH, which was used to normalize the amount of cDNA in each PCR experiment. Gene expression at 0 dpi was defined to be 1, and gene expression at 7 or 14 dpi was calculated relative to gene expression at 0 dpi. Thirteen out of 14 genes were confirmed to be upregulated, and 6 out of 7 genes were confirmed to be downregulated in PAM cells from infected piglets. Genes shown in red were upregulated, and those shown in green were downregulated in the infected piglets as compared with the uninfected controls. dpi, days postinfection

Full size image

Among the seven downregulated genes, six were expressed at lower levels in PAMs from the infected piglets than in PAMs from uninfected piglets at 0 dpi; only the polyglutamine binding protein variant 4 (PQBP1) was higher than at 0 dpi. Specifically, Toll-like receptors, IL-16, TGF-beta type 1 receptor gene, agouti signaling protein gene, TCR-α/β chain gene, CC chemokine receptor gene, PQBP1, IL-8, epidermal growth factor gene, Ak3L1, MHC-I SLA genes, Mx1, and HSP70 were upregulated in the infected piglets compared with the uninfected control group (p < 0.01). Genes encoding the hepatoma-derived growth factor, skeletal muscle ryanodine receptor, apaf1-like mRNA, ferritin heavy polypeptide 1 (FTH1), fibrinogen alpha chain preproprotein (FGA) gene, mandelonitrile lyase, and MHC-II (SLA-DRB1) were downregulated (p < 0.01).

Discussion

In this study, we investigated changes in gene expression in PAMs in response to infection with HP-PRRSV to better understand the possible mechanism(s) underlying its enhanced infectivity and pathogenicity. Here, the expression profiles of genes in HP-PRRSV-HuN4-infected PAM cells were compared with their uninfected counterparts using the SSH method, and 358 clones were identified and verified by Southern blotting. Classification of these genes according to their biological functions revealed upregulated genes related to immune function (40 %) and cell signaling (24 %) and downregulated genes related to ribosomal proteins, cytoskeleton regulation, and apoptosis/cell differentiation in the HP-PRRSV-infected PAM cells.

Among the 21 differentially expressed genes in PAM cells harvested from piglets after in vivo HP-PRRSV infection that were selected to be analyzed by real-time RT-PCR, 19/21 (90.5 %) exhibited the same trends observed by Southern blot analysis of PAM cells infected in vitro. Specifically, 13 of the 14 genes that were upregulated in vitro were confirmed to increase, and 6 of the 7 downregulated genes were confirmed to decrease in PAM cells after in vivo HP-PRRSV infection. The finding that two genes did not follow the same trend observed in the Southern blot analysis might result from differences between the in vitro and in vivo infections, such as the influence of other cell types or the time point after PRRSV infection that was chosen for testing.

During the process of verifying protein expression of the differentially expressed genes, we confirmed that the representative proteins Mx1, HSP70 (by western blot), and IL-8 (by ELISA), were upregulated. Protein expression for most of the genes identified was not evaluated, because antibodies specifically recognizing these porcine proteins were not available. Mx1 protein, induced by antiviral type I IFNs, is responsible for mediating a specific antiviral state against influenza virus infection in murine [30] and porcine [29] hosts; it accumulates in the nucleus and inhibits influenza virus replication at the level of primary transcription [5]. In pigs, Mx1 expression was also found to be upregulated after PRRSV infection in a peptide-conjugated morpholino oligomer inhibition assay [28] as well as in foot-and-mouth disease virus (FMDV)-infected nasal mucosal epithelial cells in vitro and in vivo. Interestingly, FMDV infection also induced IL-8 protein expression [7]. IL-8 and IFN-γ are linked to PRRS virus clearance [21]. IL-8, an important leukocyte chemokine, is produced by monocytes and macrophages and primarily functions in chemotaxis, inflammation, and neutrophil granulocyte activation. In the present study, expression of the inflammatory cytokine IL-16 and CC chemokine receptor genes (CCR3, CCR2, CCR5, and CCRL2) was also induced after PRRSV infection.

The real-time PCR results indicate that PRRSV infection triggers TCR-α/β, MHC class I, IL-8, IL-16, and Toll-like receptor expression and turns off MHC class II expression. MHC molecules are critical for antigen processing, management, and presentation, and TCR-α/β recognizes MHC molecules by their CDR1 and CDR2 domains. The data obtained in our study are consistent with the different functions of MHC class I and II molecules, because class II molecules mainly present exogenous antigen to CD4 + T cells, while class I molecules present endogenous antigens to CD8+ cytotoxic T lymphocytes, and these cells are known to be required for viral clearance during viral infections.

Xiao et al. evaluated the lung response to either HP-PRRSV or classical North American PRRSV infection using whole lung tissue and found that an anti-apoptotic and anti-inflammatory state was induced to subvert the host innate immune response [39, 40]. Another recent study revealed that HP-PRRSV infection induced gene expression related to cytoskeleton and exocytosis organization, protein degradation and folding, intracellular calcium regulation, and zinc homeostasis in PAMs harvested from infected piglets [41]. However, no comprehensive study has been performed to investigate the innate immune changes induced specifically within PAMs upon infection with the highly pathogenic strain. SSH is a powerful technique for comparison of two mRNA populations and for obtaining genes exclusively expressed by one population. SSH has overcome the technical limitations of traditional subtraction methods requiring several rounds of hybridization and that are not well suited to identifying rare mRNAs. Furthermore, SSH has become widely used because of its simple operation, high efficiency, and low false-positive rate [2, 18, 24, 26].

The highly pathogenic 1918 Spanish influenza A virus has been suggested to have triggered aberrantly high and sustained expression of many of the above-mentioned innate immune response genes in humans, including proinflammatory cytokines and chemokines, which could have contributed to the virulence of this pathogenic virus [4, 15, 17, 20, 31]. Compared with the highly pathogenic 1918 influenza A virus, the clinical and pathological features of HP-PRRSV are quite similar, including the high viral titers caused by rapid viral replication, severe lung tissue damage (bleeding, immune cell infiltration), and high morbidity and mortality. Therefore, the data presented in the present study suggest that the highly pathogenic PRRSV variant may, to some degree, cause excessive stimulation of the innate immune response.

In summary, the results from this study reveal that the differential gene expression we observed between HP-PRRSV-infected PAM cells and uninfected PAM cells is mainly related to differences in cell signaling and innate immunity. The results also indicate that HP-PRRSV infection excessively stimulates innate immune response genes, including proinflammatory cytokines and chemokines. Further studies will be necessary to decipher the roles of the differentially expressed genes in HP-PRRSV infection that were identified in this study.