Actinobacillus pleuropneumoniae induces SJPL cell cycle arrest in G2/M-phase and inhibits porcine reproductive and respiratory syndrome virus replication

Porcine reproductive and respiratory syndrome virus (PRRSV) is one of the most important pathogens in the swine industry and causes important economic losses. No effective antiviral drugs against it are commercially available. We recently reported that the culture supernatant of Actinobacillus pleuropneumoniae, the porcine pleuropneumonia causative agent, has an antiviral activity in vitro against PRRSV in SJPL cells. Objectives of this study were (i) to identify the mechanism behind the antiviral activity displayed by A. pleuropneumoniae and (ii) to characterize the active molecules present in the bacterial culture supernatant. Antibody microarray analysis was used in order to point out cellular pathways modulated by the A. pleuropneumoniae supernatant. Subsequent, flow cytometry analysis and cell cycle inhibitors were used to confirm antibody microarray data and to link them to the antiviral activity of the A. pleuropneumoniae supernatant. Finally, A. pleuropneumoniae supernatant characterization was partially achieved using mass spectrometry. Using antibody microarray, we observed modulations in G2/M-phase cell cycle regulation pathway when SJPL cells were treated with A. pleuropneumoniae culture supernatant. These modulations were confirmed by a cell cycle arrest at the G2/M-phase when cells were treated with the A. pleuropneumoniae culture supernatant. Furthermore, two G2/M-phase cell cycle inhibitors demonstrated the ability to inhibit PRRSV infection, indicating a potential key role for PRRSV infection. Finally, mass spectrometry lead to identify two molecules (m/z 515.2 and m/z 663.6) present only in the culture supernatant. We demonstrated for the first time that A. pleuropneumoniae is able to disrupt SJPL cell cycle resulting in inhibitory activity against PRRSV. Furthermore, two putative molecules were identified from the culture supernatant. This study highlighted the cell cycle importance for PRRSV and will allow the development of new prophylactic or therapeutic approaches against PRRSV.


Background
Coinfections are likely more frequent in farm than reported and it is well established that a primary infection with a pathogen, viral or bacterial, may enhance the infectious potential of a secondary pathogen [1][2][3][4]. The porcine respiratory disease complex (PRDC) is a multifactorial disease affecting fattening pigs that is caused by coinfections with viral and/or bacterial pathogens including the porcine reproductive and respiratory syndrome virus (PRRSV), Actinobacillus pleuropneumoniae and others [2,5,6]. PRDC is the most common disease in swine industry resulting in significant economic losses and is characterized by several symptoms including respiratory distress, fever, lethargy, stunted growth and death [2,5,6].
Coinfections are often studied by observing clinical symptoms in model animals; however, the basic mechanisms involved in these pathogen-pathogen interactions are often overlooked. In vitro investigations can provide insights for understanding coinfections. Our laboratory recently developed a model to study co-infections by A. pleuropneumoniae and PRRSV using SJPL cells [7].
Actinobacillus pleuropneumoniae (App) is the causative agent of porcine pleuropneumonia, an important disease in swine industry. The disease is well controlled in USA and Canada but still a significant problem in Latin America and some Asian and European countries [17]. Actinobacillus pleuropneumoniae is a Gram-negative rodshaped bacteria and member of the Pasteurellaceae family. This bacterium is known to possess many virulence factors including lipopolysaccharides, capsular polysaccharides, outer membrane proteins involved in the acquisition of essential nutrients, surface molecules involved in adherence to the respiratory tract and Apx toxins [18]. For a recent review about virulence factors of A. pleuropneumoniae see Chiers and collaborators [18].
We recently reported that A. pleuropneumoniae culture supernatant has an antiviral activity against PRRSV in vitro in SJPL infected cells and in porcine alveolar macrophages [7]. This antiviral activity is not induced by A. pleuropneumoniae lipopolysaccharides or by peptidoglycan fragments (i.e. NOD1 and NOD2 ligands) [7].
The identity of the molecules responsible for the antiviral activity are unknown and their identification could provide the basis for the development of new therapeutic drugs, including prophylactic drugs with suitable biopharmaceutical properties against PRRSV infection. It is of note that experiments performed with culture supernatant of Haemophilus parasuis (strain Nagasaki), a close relative of A. pleuropneumoniae, did not show any antiviral activity against PRRSV (J. Labrie and M. Jacques, unpublished data).
We hypothesize that the culture supernatant of A. pleuropneumoniae induces a specific SJPL cell response which has an antiviral activity against PRRSV. The first objective of the present study was to identify the mechanism behind the antiviral activity displayed by A. pleuropneumoniae. The second objective was to identify the molecules present in the A. pleuropneumoniae culture supernatant which are responsible for the antiviral activity against PRRSV. Therefore, we first used an antibody microarray to identify cell pathways modulated by the A. pleuropneumoniae culture supernatant, observed modulations in cell cycle regulation pathways and confirm these modulations by cell cycle analysis using flow cytometry. We also demonstrated the ability of two known cell cycle inhibitors to inhibit PRRSV. Finally, mass spectrometry was used to detect and identify two molecules present only in the culture supernatant of A. pleuropneumoniae.

Protein profiling of SJPL cells
Lévesque and collaborators recently demonstrated that both AppΔapxICΔapxIIC culture supernatant and its ≤ 3 kDa ultrafiltrate have an antiviral activity in vitro against PRRSV [7]. Therefore, protein profiling of SJPL cells infected or not with PRRSV (MOI 0.5) and/or treated or not with the AppΔapxICΔapxIIC culture supernatant was performed using Kinex KAM-850 antibody microarray. Eight hundred and fifty four cell signaling proteins were targeted, using 337 phosphosite-specific antibodies and 517 pan-specific antibodies. Pan-specific antibodies targeted both phosphorylated and unphosphorylated proteins forms. Proteins were classified into nine groups according to their cellular functions: (1) transcription and translation factors; (2) proteins implicated in signal transduction pathway; (3) proteins implicated in host-pathogen interaction or in immune response; (4) proteins implicated in stress response; (5) proteins implicated in cell spreading, cell migration, cell survival, cell growth, cell cycle and cell proliferation; (6) apoptosis signaling pathway; (7) cytoskeleton-associated proteins; (8) proteins implicated in cell biosynthesis and metabolism; (9) proteins implicated in other biological process or unclassifiable proteins. Data are expressed in percent of change from control (mock-infected and untreated SJPL cells; % CFC) and are presented in Fig. 1 and Additional file 1: Table S1. The AppΔapxICΔapxIIC culture supernatant upregulates inhibitor of nuclear factor kappa-B kinase subunit beta (IKKβ), receptor-interacting serine/threonine-protein kinase 2 (Rip2) and caspase-9 (CASP9). In cells treated with AppΔapxICΔapxIIC culture supernatant, cell cycle regulation proteins (M-phase inducer phosphatase 3 inactive form (CDC25c (ser216)) and cyclin-dependent kinase 1/2 active forms (CDK1/2 (tyr161))) and caspase-3 (CASP3) were modulated by the culture supernatant regardless of whether they were infected with PRRSV or not. Also, PRRSV seems to upregulate mitogen-activated protein kinase (MAPK)/mitogen activated protein kinase ERK pathway and NF-κB pathway. Indeed, serine/threonine-protein kinase A-Raf (Raf-A), ERK1/2, interleukin-1 receptorassociated kinase 4 (IRAK4) and cyclic AMP-responsive element-binding protein 1 (CREB1) were differentially modulated in presence or absence of culture supernatant. Thus, PRRSV upregulates MAPK/ERK and NF-κB pathways. The culture supernatant seems to decrease the PRRSV induced upregulation for MAPK/ERK and NF-κB pathways. Overall, the antibodies microarray data suggests that A. pleuropneumoniae culture supernatant modulates SJPL cell cycle.

Effect of the AppΔapxICΔapxIIC culture supernatant on cell cycle
The microarray antibodies data suggesting that AppΔapxICΔapxIIC culture supernatant modulates SJPL cell cycle, we decided to analyze SJPL and MARC-145 cell cycle after an 18 h treatment with the AppΔapxICΔapxIIC culture supernatant, using flow cytometry. SJPL cells incubated in DMEM were used as normal cell cycle and we obtained a mean 50.6 % of SJPL cells in Gap-1 phase (G1-phase); 35.1 % of SJPL cells in DNA-synthesis phase (S-phase); and 14.3 % of SJPL cells in Gap-2 and mitotic phase (G2/M-phase) ( Fig. 2a and b). We then used as controls, two molecules known to modulate the cell cycle at different phases. Nocodazole interferes with the polymerization of microtubules, blocking treated cells in G2/M-phase and inducing cell death by apoptosis [19]. Aphidicolin is a DNA-polymerase inhibitor and induces cell cycle arrest in S-phase [20]. When SJPL cells were treated with nocodazole or aphidicolin, we observed Fig. 1 AppΔapxICΔapxIIC culture supernatant upregulated cellular proteins when SJPL cells were infected or not with PRRSV. Proteins represented were selected using a Z-score difference (Z-score SJPL -Z-score SJPL treated/infected cells) threshold of plus 75 % (corresponding in log2 to 1.68 fold change versus the control). CDC25c: ser216 phosphorylated form; CDK1/2: tyr161 phosphorylated form; Tau: ser516 and ser519 phosphorylated forms; Rb: ser807 and ser811 phosphorylated forms differences between treated and untreated cells (Fig. 2a). With nocodazole treatment, SJPL cell number in G2/Mphase increased (in mean) to 41.5% and an additional apoptotic peak before the G1-phase peak was seen as expected (Fig. 2a). With aphidicolin treatment, the proportion of SJPL cells in S-phase increased (in mean) to 51.9 % (Fig. 2a). Following validation with cell cycle controls,  33.9 %, respectively ( Fig. 2a and b), indicating that A. pleuropneumoniae culture supernatant had an effect on SJPL cell cycle.
Lévesque and collaborators reported that the antiviral effect of AppΔapxICΔapxIIC culture supernatant against PRRSV was not observed in MARC-145 infected cells [7]. To provide supplemental insight of the impact of In contrast to SJPL cells, when MARC-145 cells were treated with the culture supernatant, no differences were observed between treated and control cells ( Fig. 2a and b).

Effect of the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate on SJPL cell cycle
Knowing that AppΔapxICΔapxIIC culture supernatant modulates SJPL cells cycle, we treated cells with the ≤ 3 kDa DMEM ultrafiltrate (as control) or the ≤ 3 kDa App-ΔapxICΔapxIIC culture supernatant ultrafiltrate to know if the ≤ 3 kDa fraction also modulates the SJPL cell cycle. The ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate significantly (n = 4; P ≤ 0.01) increased the number of cells in G2/M-phase from 11.8 % to 22.1 % (in mean) ( Fig. 3a and b (left side)). Thus, the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate which possesses the antiviral activity against PRRSV also modulates SJPL cell cycle. No modulation of the cell cycle was also observed in MARC-145 cells treated with the AppΔapxICΔapxIIC culture supernatant ( Fig. 3a and b right side). This suggests a potential link between

Effect of DIM and SBE-13 on SJPL cell cycle
We have previously shown, using protein profiling, that AppΔapxICΔapxIIC culture supernatant upregulates CDC25c (ser216) and CDK1/2 (tyr161), both of which are implicated in the G2/M cell cycle regulation pathway. Two specific cell cycle inhibitors were used to verify whether cell cycle modulation induced by the culture supernatant is due to CDC25c and/or CDK1/2. We first used DIM which is a specific activator of cell cycle inhibitor serine/threonine-protein kinase (Chk2), Chk2 is an inhibitor of CDC25 and leads to a cell cycle arrest in G2/M-phase [21]. Then, we used SBE-13 which is a selective inhibitor of cell cycle activator polo-kinase isoforms: PLK1, PLK2, PLK3 [22]. These proteins are implicated in the G2/M-phase transition pathway; PLK1 is a CDC25c activator and CDC25c promotes the G2/ M-phase transition using cyclin-B/CDK1 complex. Both DIM and SBE-13 treatments induced a SJPL cell cycle modulation (Fig. 4). G1-phase of SJPL cells was not modulated by DIM or SBE-13 ( Fig. 4a and b). The number of cells in S-phase was significantly (n ≥ 6; P ≤ 0.001) decreased (in mean) from 35.1 % to 17.7 % and to 3.6 % by DIM and SBE-13 treatments, respectively ( Fig. 4a and c) and proportions of cells in G2/M-phase were significantly (n ≥ 6; P ≤ 0.001) increased (in mean) from 14.3 % to 37.0 % and to 51.8 % by DIM and SBE-13 treatments, respectively as previously observed with A. pleuropneumoniae culture supernatant ( Fig. 4a and d).

Evaluation of SJPL cell proliferation
In order to confirm the cell cycle arrest we performed cell counts as an indicator of cell proliferation. Cells were counted before and after treatments and ratios between the second and the first counts were determined. These ratios were then compared to untreated cells ratio (DMEM) or to the ≤ 3 kDa DMEM ultrafiltrate ratio (Fig. 5a). We have previously shown that nocodazole and aphidicolin blocks SJPL cell cycle (Fig. 2a, upper panel) and decreases SJPL proliferation (Fig. 5a). Furthermore, cells treated with the AppΔapxICΔapxIIC culture supernatant, DIM or SBE-13 showed significantly (n = 8; P ≤ 0.05) less proliferation than cells treated with DMEM. Same effect was observed between cells treated with the ≤ 3 kDa DMEM ultrafiltrate and cells treated with the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate; from a fold change of 2.4 to 1.3 ( n = 8; P ≤ 0.01). These data indicated that the AppΔapxICΔapxIIC culture supernatant or the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate significantly decreased SJPL cells proliferation.

Detection of PRRSV in SJPL infected and treated cells
Knowing that the culture supernatant disrupts SJPL cell cycle and decreases SJPL cell proliferation we were interested to see whether both cell cycle inhibitors, DIM and SBE-13, can also block PRRSV infection in SJPL cells. When SJPL cells were infected with PRRSV and treated with the AppΔapxICΔapxIIC culture supernatant, an antiviral activity was observed (Fig. 6) as previously reported by Lévesque and collaborators [7]. When SJPL infected cells were treated with DIM, no PRRSV was detected; in contrast to DIM solvent control (Fig. 6).
Similar observations can be done for cells treated with SBE-13 and SBE-13 solvent control (Fig. 6). Absence of PRRSV in SJPL infected cells treated with DIM and SBE-13 suggests an antiviral effect of these two cell cycle inhibitors against PRRSV.

Mass spectrometry analysis
In order to identify the actives molecules present in the AppΔapxICΔapxIIC culture supernatant, mass spectrometry was used. Interestingly, mass spectrometry and collision-induced dissociation can provide comprehensive spectra allowing structural information to be derived with assistance of bioinformatics. Thus, mass spectrometry was used to analyze the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate to detect, identify and characterize active molecules. The ≤ 3 kDa DMEM and ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrates were fractionated using a simple salting-out liquid-liquid extraction in order to remove undesired inorganic molecules. Supernatant were then analyzed in full scan positive ion mode by LC-MS/MS and results are presented with a total ion current chromatogram (TIC) (Fig. 7a). Both ≤ 3 kDa DMEM ultrafiltrate and ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate were then compared and two unique peaks were observed in the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate (Fig. 7a, black arrows). Full scan spectra and extracted ion  Fig. 7b) and m/z of 663.6 ( Fig. 7c). Product ion spectra for these two peaks were acquired. At m/z 515.2, four ions were observed at m/z 160.6; 239.6; 328.6; 347.6, and six ions were observed for m/z 663.6 at m/z 495.0; 496.11; 551.0; 552.0; 606.9; and 607.9. However, using in silico fragmentation predictor and database search (e.g. METLIN) we did not succeed to unequivocally identify these two molecules present in A. pleuropneumoniae culture supernatant.
Separation of the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate using thin-layer chromatography In order to improve fractionation of the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate and test molecules activities, we separated the suspected active ingredients using thin-layer chromatography and compare results with standards run on the same chromatogram. Migrations were carried out using a chloroform/methanol mixture and were then read using UV-A. We observed a distinctive spot (retention factor (Rf) = 0.52) present in the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate (Fig. 8). Interestingly, this distinctive spot was also present in the ≤ 3 kDa culture supernatant ultrafiltrate of two reference strains of A. pleuropneumoniae: L20 (serotype 5b) and S4074 (serotype 1) (Fig. 8).

Discussion
Lévesque and collaborators recently reported an in vitro antiviral activity of the A. pleuropneumoniae culture supernatant against PRRSV in SJPL infected cells and in porcine alveolar macrophages [7]. The objective of the study presented here was to identify the mechanism behind the antiviral activity displayed by A. pleuropneumoniae and to identify the active molecules present in A. pleuropneumoniae culture supernatant.
In SJPL infected cells, we demonstrated that PRRSV strongly upregulates mitogen-activated protein kinase (MAPK)/ERK pathway using ERK1/2, RAF-A and CREB. Our results are in agreement with Lee and Lee which demonstrated an activation of ERK1/2 during PRRSV infection [23] and with Han and collaborators which highlighted that PRRSV induces degradation of CREB-binding protein (CBP) [24]. In contrast, treatment of cells with A. pleuropneumoniae culture supernatant decreased the upregulation of CREB1 and ERK1/2 induced by PRRSV demonstrating that the A. pleuropneumoniae culture supernatant decrease MAPK/ERK and NF-κB pathways upregulation induced by PRRSV. This observation is interesting as it can partly Peptidoglycan fragments can stimulate the nucleotidebinding oligomerization domain (NOD) pathway [25,26]. Following activation, NOD1 or NOD2 recruits RIP2 and nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor alpha (IκBα), leading to NF-κB activation [27]. Protein profiling indicated that only RIP2, a specific NF-κB activator, was upregulated by the culture supernatant in absence of PRRSV. The absence of NF-κB activation by A. pleuropneumoniae culture supernatant alone is certainly puzzling, but Lévesque and collaborators demonstrated that NOD pathway induction using peptidoglycan fragments did not lead to an antiviral effect [7]. Our results confirm their observation.
Auger and collaborators demonstrated a differential expression of NF-κB pathway in SJPL cells after 3 h of infection with A. pleuropneumoniae serotype 1 (S4074) [28]. In SJPL cells treated with A. pleuropneumoniae culture supernatant, we only observed a weak upregulation of IKKβ. Auger and collaborators implicated the A. pleuropneumoniae lipopolysaccharides in NF-κB pathway activation [28]. Absence of NF-κB activation by the culture supernatant cannot be explained because A. pleuropneumoniae lipopolysaccharides were not removed from the culture supernatant. Conversely, PRRSV seems to be able to activate the NF-κB pathway in SJPL cells, as was already observed in MARC-145 and alveolar macrophages [29].
Apoptosis is an important cell process which can be regulated by caspases proteins [30]. We observed that A. pleuropneumoniae culture supernatant weakly regulates CASP3 and CASP9. Apx toxins from A. pleuropneumoniae are known to induce apoptosis through activation of c-Jun N-terminal kinase (JNK)/MAPK pathway and CASP3 [31,32], as using an Apx mutant prevent this induction of JNK/MAPK pathway. Our results suggest that A. pleuropneumoniae culture supernatant do not induce apoptosis. We did not observe any peak corresponding to apoptosis during cell cycle analysis (Figs. 2a and 3a). These results were in agreement with Auger and collaborator, who carried out Western blot to detect CASP3 activation and concluded that A. pleuropneumoniae was not inducing apoptosis in SJPL cells [28].
Cell proliferation is a tightly controlled process. Many cell cycle checkpoints can be modulated to provoke cell cycle arrest. The G2/M-phase checkpoint is activated after DNA damages and provides opportunity to stop and repair the damaged cells. To the best of our knowledge, the observed cell cycle arrest induced by A. pleuropneumoniae has never been reported before. However, Aggregatibacter actinomycetemcomitans (formerly Actinobacillus actinomycetemcomitans) was able to stop growth and induce cell cycle arrest in G2/M-phase of HeLa cells, by production of the cytolethal distending toxin (CDT) [33,34]. With regards to PRRSV, Sun and collaborators demonstrated the ability of its non structural protein 11 to control the cell cycle and to induce a cell cycle arrest in S-phase [35]. We demonstrated that A. pleuropneumoniae culture supernatant is able to induce a G2/M-phase cell cycle arrest in SJPL cells, resulting in decreased cell proliferation. Using CDC25c, CDK1/2 and PLK inhibitors, we confirmed the protein profiling results. DIM and SBE-13 are known to induce a cell cycle arrest in G2/M-phase and both decreased proliferation of SJPL cells. The cell cycle arrest induced by A. pleuropneumoniae can be due to a CDC25c or CDK1/2 inhibition. We will further investigate this pathway using Western-blot analysis of regulation proteins.
We also studied the effect of the culture supernatant on the MARC-145 cell line. Lévesque and collaborators have shown that the culture supernatant of A. pleuropneumoniae has no antiviral effect in MARC-145 cells [7]. In agreement with this finding, our cell cycle observations demonstrated a differentiated effect of the culture supernatant on SJPL and on MARC-145 cells. In SJPL cells, the culture supernatant induced a cell cycle arrest in G2/Mphase, whereas the same culture supernatant has no antiviral and no cell cycle effects on MARC-145 cells. Similar effects were observed for cell proliferation and for the cell cycle arrest induce by DIM and SBE-13. These differences between SJPL and MARC-145 cells support our results indicating that the culture supernatant inhibited PRRSV in SJPL cells by inducing a G2/M-phase cell cycle arrest.
Using tandem mass spectrometry we have isolated and identified two unique peaks (m/z 515 and m/z 663) from the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate. We were able to separate and purify the ≤ 3 kDa AppΔapxICΔapxIIC culture supernatant ultrafiltrate using TLC plates, and more interestingly, to recover the two peaks after TLC spots extraction. Unfortunately, mass spectrometry fingerprinting did not lead to identification of these molecules corresponding to these two peaks. In order to obtain more information and to definitely identify active molecules, we have to refine extraction protocols and to obtain higher purity products in amounts compatible with 1 H and 13 C nuclei nuclear magnetic resonance (NMR).

Conclusions
We have demonstrated for the first time that A. pleuropneumoniae disrupts the SJPL cell cycle and prevents We have also demonstrated the importance of cell cycle in PRRSV infection. Furthermore, two putative molecules were identified from the culture supernatant. It seems likely that these molecules are responsible for the antiviral activity of A. pleuropneumoniae against PRRSV. We have to further investigate these molecules in the hope that it will lead to the development of a new antiviral drug against PRRSV.

Bacterial and viral strains
Actinobacillus pleuropneumoniae non-hemolytic and non-cytotoxic MBHPP147 was kindly provided by Ruud P.A.M. Segers (MSD Animal Health, Boxmeer, The Netherlands). This strain is a mutant of the serotype 1 reference strain S4074 producing non-active ApxI and ApxII toxins (AppΔapxICΔapxIIC) [41]. Bacteria was grown at 37°C in brain heart infusion broth (BHI; Oxoid Ltd., Basingstoke, Hampshire, England) or in BHI agar The PRRSV North American reference strain IAF-Klop was used in this study. This strain is a genotype II strain [42,43].

Bacterial culture supernatant
Actinobacillus pleuropneumoniae culture supernatant was prepared as recently described by Lévesque and collaborators [7]. Briefly, an overnight culture of AppΔapxICΔapxIIC was diluted into fresh BHI broth and grown to an OD 600 of 0.6. This culture was then diluted into DMEM medium, supplemented with 10 % FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 1 % NEAA, to 1 × 10 6 CFU/mL and incubated overnight at 37°C. Bacterial cells were centrifuged at 4,000 g for 20 min and the supernatant was harvested and passed through a 0.22 μm filter (Sarstedt Inc., Newton, NC, USA). Filter-sterilized supernatant was ultrafiltrated through a 3 kDa membrane (Merck Millipore Ltd., Carrigtohill, Cork, Ireland). Resulting culture supernatant ultrafiltrate and the filter-sterilized supernatant were stored at −20°C up to 6 months. Supplemented DMEM medium was also ultrafiltrated through a 3 kDa membrane and used as a control.

Protein profiling of SJPL cells
Sample preparation was performed as described by Auger and collaborators [28]. Briefly, a T25 flask (Sarstedt Inc.) of confluent SJPL cells (0.5 × 10 6 cells) were infected or not with IAF-Klop PRRSV strain for 4 h (multiplicity of infection (MOI): 0.5) and then treated with AppΔapxICΔapxIIC culture supernatant or left untreated for 18 h. These conditions are known to block PRRSV replication [7]. Cells were washed twice with Dulbecco's phosphate-buffered saline (DPBS; Life Technologies) and 500 μL of lysis buffer was added (20 mM MOPS (pH 7.0; Sigma-Aldrich), 2 mM EGTA (Sigma-Aldrich), 5 mM EDTA (Fisher Scientific, Fair Lawn, NJ, USA), 30 mM sodium fluoride (Sigma-Aldrich), 60 mM β-glycerophosphate (pH 7.2; Sigma-Aldrich), 20 mM sodium pyrophosphate (Sigma-Aldrich), 1 mM sodium orthovanadate (Sigma-Aldrich), 1 % (v/v) Triton X-100 (Sigma-Aldrich) and one protease inhibitor cocktail tablet Complete Mini EDTA-free (Roche Diagnostics GmbH, Mannheim, Germany), pH 7.2). Cells were removed using a cell scraper (Nalge Nunc International, Rochester, NY, USA) and transferred into a microcentrifuge tube on ice. Sonication was performed at 180 Joules using an ultrasonic processor (Cole-Parmer, Vernin Hills, IL, USA) in order to lyse the cells. Samples were then ultracentrifuged at 90,000 g for 30 min in a Sorvall RC M100 ultracentrifuge (Beckman Coulter, Mississauga, ON, Canada). Protein concentration was determined on the supernatant using Pierce BCA protein assay kit (Thermo Fisher Scientific Inc., Rockford, IL, USA) and was adjusted to 2 mg/mL. Samples were then analyzed using Kinex KAM-850 antibody microarray (Kinexus Bioinformatics Corporation, Vancouver, BC, Canada). Kinex KAM-850 antibody microarray includes 854 antibodies which target different cell signaling pathways and included 517 pan-specific antibodies and 337 phosphosite-specific antibodies. These include 466 antibodies against protein kinase, 44 antibodies against protein phosphatase, 46 antibodies against stress response proteins, 75 antibodies against protein implicated in transcription and 223 others. Data were expressed in a chemiluminescence signal ratio versus the untreated cells. The percentage change from control (% CFC) was calculated for each protein by dividing the treated condition Z-ratio by the control Z-scores X 100 as described previously [44].

Cell cycle analysis
Cell cycle analysis was done using a modified protocol from Sazer and Sherwood [45]. Briefly, 0.5 × 10 6 SJPL cells were cultured overnight in T75 flask (Sarstedt Inc.) in DMEM. Cells were treated for 18 h at 37°C in a 5 % CO 2 atmosphere with either DMEM, the ≤ 3 kDa DMEM ultrafiltrate, AppΔapxICΔapxIIC undiluted culture supernatant, the ≤ 3 kDa culture supernatant ultrafiltrate, 0.4 μM nocodazole (Sigma-Aldrich), 7.4 μM aphidicolin (Sigma-Aldrich), 100 μM 3,3′-diindolylmethane (DIM; R&D Systems Inc., Minneapolis, MN, USA) or 50 μM SBE-13 hydrochloride (SBE-13; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA). Cells were washed twice with DPBS and then trypsinized at 37°C for 10 min. Trypsin (Life Technologies) was inhibited using complete medium and cells were centrifuged at 380 g for 15 min at 4°C. Cells were washed one more time with DPBS and then fixed and permeabilized using 70 % (v/v) cold ethanol for 2 h at room temperature followed by an overnight incubation at 4°C. Cells were washed twice with DPBS then treated for 2 h at 37°C with 2 mg of ribonuclease-A (RNase-A; Sigma-Aldrich) in 900 μL of DPBS and cells were finally stained with 100 μL of propidium iodide (PI) 0.2 mg/mL (Sigma-Aldrich). Stained cells were analyzed on BD FACS Calibur (Becton Dickinson, Mississauga, ON, Canada), measuring fluorescence emission at 585 nm after excitation at 488 nm. For each analysis, 10,000 events were evaluated and DNA content was determined by ModFit LT v3.0 (Verity software house Inc., Topsham, ME, USA), which provided histograms to evaluate cell cycle distribution.

Cell counting
To compare cell proliferation and cell viability during treatments, we used a Countess automated cell counter (Life Technologies). After the trypsinization step in the cell cycle analysis protocol, cells were washed twice in DPBS and centrifuged at 380 g for 15 min at 4°C. Cells were resuspended into 1 mL DPBS and 100 μL of this cell suspension was added to 100 μL of trypan blue (Life Technologies) and 10 μL of this dilution was added to a Countess cell counting chamber slide (Life Technologies) to perform cells count and viability measurement. Data were automatically expressed as number of viable cells per mL.

Immunofluorescence assay
To detect PRRSV infected cells, we used a modified immunofluorescence assay (IF) protocol from Provost and collaborators which detects PRRSV antigens [39].

Sample preparation for biochemical analysis
In order to eliminate salts and remove inorganic contaminants, a salting-out liquid-liquid extraction was performed. Briefly, 2 mL of ultrafiltrate samples (≤3 kDa) were mixed vigorously for 1 min with 2 g of NaCl (Sigma-Aldrich) and 2 mL of acetonitrile (Fisher Scientific), organic layer was kept after decantation for analysis.

Thin-layer chromatography
Thin-layer chromatography (TLC) plates were used to analyze samples and preparative TLC plates were used to isolate molecules of interest. For samples analysis, 40 μL of the ≤ 3 kDa ultrafiltrate were applied on a 5 cm × 10 cm (length × width) silica gel plate (Sigma-Aldrich). For preparative TLC plates, 1 mL of the ≤ 3 kDa ultrafiltrates were applied on a 20 cm × 20 cm (length × width) silica gel plate (Sigma-Aldrich). Samples were separated with a mobile phase composed of 75 % (v/v) chloroform (BDH Inc., Toronto, Ontario, Canada) and 25 % (v/v) methanol (Fisher Scientific) in a saturated-closed glass chamber. Migration was performed at ambient temperature (21°C ± 3°C). After migration, plates were airdried for 5 min and then observed under a ultra-violet A (UV-A) light (Ultra-Lum Inc., Claremont, CA). Areas which correspond to spots of interest were scraped from the silica gel plate and transferred into a microcentrifuge tube. The extracted silica was weighed and molecules were extracted with acetonitrile at a ratio of 1:3 (w/v) for 2 h at room temperature. Samples were then centrifuged at 18,000 g for 5 min and the supernatant was kept for liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis.

Instrumentation and bioanalytical methods
The LC-MS/MS system included a Thermo Surveyor autosampler, a Thermo Surveyor MS pump and a Thermo LCQ Advantage Ion Trap Mass Spectrometer (Thermo Scientific, San Jose, CA, USA). Data was acquired and analyzed with Xcalibur 1.4 (Thermo Scientific). Samples were analyzed by LC-MS/MS in full scan positive ion mode. Briefly, 2 μL of sample was injected and the separation was achieved with a gradient mobile phase along with a microbore column Thermo Biobasic C8 (100 × 1 mm) (Thermo Scientific), with a particle size of 5 μm. The initial mobile phase condition consisted of acetonitrile and water (both fortified with 0.4 % (v/v) of formic acid) at a ratio of 5:95. From 0 to 1 min, the ratio was maintained at 5:95. From 1 to 21 min, a linear gradient was applied up to a ratio of 90:10 and maintained for 5 min. The mobile phase composition ratio was then reverted to the initial conditions and the column was allowed to re-equilibrate for 14 min for a total run time of 40 min. The flow rate was fixed at 75 μL/min. Mass spectrometer was coupled with the LC system using a pneumatically assisted electrospray ion source (ESI). Sheath nitrogen gas was set to 10 units and ESI electrode was set to 4000 V in positive mode. Capillary temperature was set at 300°C and capillary voltage to 34 V. Instrument was operating in full scan MS mode (m/z 200-1000) and MS/MS data was collected on the most abundant peaks in a data dependent acquisition (DDA) mode using a peak intensity threshold of 1 × 10 5 counts per second (cps). The ≤ 3 kDa ultrafiltrate of DMEM was used as a negative control and spectra were compared with the ≤ 3 kDa ultrafiltrate of the A. pleuropneumoniae culture supernatant or the TLC extracted molecules. This study was supported by discovery grants from the Natural Sciences and Engineering Research Council (NSERC) to MJ (003428) and to CAG (312113) and by a team grant from the "Fonds de recherche du Québec -Nature et technologies" (FRQNT) to CAG and MJ. JAFB received a scholarship from the "Centre de recherche en infectiologie porcine et avicole" (CRIPA), a FRQNT strategic cluster.