The response of porcine monocyte derived macrophages and dendritic cells to Salmonella Typhimurium and lipopolysaccharide

Background Following infection and initial multiplication in the gut lumen, Salmonella Typhimurium crosses the intestinal epithelial barrier and comes into contact with cells of the host immune system. Mononuclear phagocytes which comprise macrophages and dendritic cells (DC) are of key importance for the outcome of Salmonella infection. Although macrophages and DC may differentiate from a common precursor, their capacities to process and present antigen differ significantly. In this study, we therefore compared the response of porcine macrophages and DC differentiated from peripheral blood monocytes to S. Typhimurium and one of the most potent bacterial pathogen associated molecular patterns, bacterial lipopolysaccharide. To avoid any bias, the expression was determined by protein LC-MS/MS and verified at the level of transcription by quantitative RT-PCR. Results Within 4 days of culture, peripheral blood monocytes differentiated into two populations with distinct morphology and expression of MHC II. Mass spectrometry identified 446 proteins in macrophages and 672 in DC. Out of these, 433 proteins were inducible in macrophages either after infection with S. Typhimurium or LPS exposure and 144 proteins were inducible in DC. The expression of the 46 most inducible proteins was verified at the level of transcription and the differential expression was confirmed in 22 of them. Out of these, 16 genes were induced in both cell types, 3 genes (VCAM1, HMOX1 and Serglycin) were significantly induced in macrophages only and OLDLR1 and CDC42 were induced exclusively in DC. Thirteen out of 22 up-regulated genes contained the NF-kappaB binding site in their promoters and could be considered as either part of the NF-kappaB feedback loop (IkappaBalpha and ISG15) or as NF-kappaB targets (IL1beta, IL1alpha, AMCF2, IL8, SOD2, CD14, CD48, OPN, OLDLR1, HMOX1 and VCAM1). Conclusions The difference in the response of monocyte derived macrophages and DC was quantitative rather than qualitative. Despite the similarity of the responses, compared to DC, the macrophages responded in a more pro-inflammatory fashion. Electronic supplementary material The online version of this article (doi:10.1186/s12917-014-0244-1) contains supplementary material, which is available to authorized users.


Background
Pigs are one of the most important sources of animal proteins for the human population. However, due to intensive rearing, pigs are also highly susceptible to various pathogens including those with zoonotic potential. Salmonella enterica serovar Typhimurium is an example of a zoonotic agent for which pigs represent a reservoir for the human population. In fact, pigs represent the second most common source of Salmonella for humans after poultry and are the most important source if only serovar S. Typhimurium is considered.
Salmonella in pigs is transmitted by the oral-fecal route. Infections of pigs with S. Typhimurium are often asymptomatic [1] although mild diarrhea may be recorded at the early stage of infection [2,3]. After initial multiplication in the gut lumen, S. Typhimurium invades the intestinal epithelial cells and comes into contact with the host's immune system [4]. The innate immune responses of different parts of porcine intestinal tract have been repeatedly described [5][6][7]. Of the different leukocyte subpopulations, mononuclear phagocyte cells are of key importance for the outcome of Salmonella infection which comprise of macrophages (MΦ) and dendritic cells (DC). When these cells come into contact with bacterial pathogens, they recognize the pathogens through the presence of pathogen associated molecular patterns (PAMP) present in prokaryotic pathogens but absent from eukaryotic host cells. This leads to the modification of gene expression in MΦ and DC, and secretion of signaling molecules to coordinate responses of other cells of the host immune system. In addition, both MΦ and DC are able to take up, process and present antigens to lymphocytes, thereby inducing the development of an adaptive immune response [8,9].
Although only particular subsets of MΦ and DC can be differentiated from a monocyte precursor, their capacity to take up, process and present antigen differ significantly. There are several papers to date describing the expression profiles of selected genes in porcine monocytes, dendritic cells or macrophages in response to external stimuli [10][11][12][13][14][15]. However, these studies compared the responses either monocytes and monocyte derived dendritic cells (MoDC) [13], two differently generated macrophages [15] or two different DC populations [11,12]. The response of MΦ and DC derived from monocytes has never been compared. Moreover, all the studies either determined the expression of preselected genes such as TLRs, MHC-II molecules, chemokines and cytokines by quantitative RT-PCR or used the Affymetrix microarray, so the measurements were limited to the level of transcription. Since the general understanding of antigen presentation and associated processes, especially in a porcine model, is far from being completely understood, in this study, we therefore differentiated porcine monocyte-derived macrophages (MoMΦ) and MoDC, and compared their response to S. Typhimurium and one of the most potent bacterial PAMP, bacterial LPS. To avoid any bias, expression was determined by protein LC-MS/MS and verified at the level of transcription by QRT-PCR. Using such an approach, we concluded that the difference in the response of MoMΦ and MoDC was quantitative rather than qualitative, i.e. MoDC responded less extensively than MoMΦ to LPS or S. Typhimurium stimulation.

Cell differentiation
Depending on culture conditions, peripheral blood monocytes differentiated into two populations with distinct morphology within 4 days of culture. Adhered monocytes differentiated into MoMΦ of spherical shape with characteristic short hairy protrusions on their surface. On the other hand, monocytes treated with IL4 and GM-CSF differentiated into MoDC characteristic by elongated cells with numerous dendrites typical of dendritic cells ( Figure 1).
Flow cytometric analysis further confirmed the difference between the two cell populations. The most remarkable difference was a more than 40 times higher expression of MHC-II molecules on the surface of MoDC compared to MoMΦ. The expression of CD14 and CD11a, when compared to MoMΦ, was numerically but not significantly higher in MoDC ( Figure 2). On the other hand, expression of CD172α, CD16, CD163, CD45, TLR-2 and TLR-4 did not differ between both cell types (data not shown). The expression of surface molecules on MoDC and MoMΦ was also determined after S. Typhimurium infection. In response to infection, the amount of CD14 increased in both cell types whereas the expression of the remaining cell surface molecules did not change ( Figure 2).

Protein mass spectrometry
The previous experiment showed that CD14 increased both in MoMΦ and MoDC after S. Typhimurium infection. In the next experiment, we therefore determined what other proteins might have increased in expression in both MoMΦ and MoDC. In addition to S. Typhimurium infection, protein expression was also determined in the LPS exposed cell cultures.
Mass spectrometry detected at least once 2191 proteins in MoMΦ and 2204 proteins in MoDC [see Additional file 1]. However, when we applied the filter criteria described below, the number of proteins decreased to 446 proteins which were repeatedly detected in MoMΦ and 672 in MoDC ( Figure 3A). Out of the filtered proteins, 400 were detected in both MoMΦ and MoDC, 46 proteins were identified only in MoMΦ and 272 proteins were identified only in MoDC. The comparison of the MoMΦ and MoDC response to LPS and S. Typhimurium showed that MoMΦ were more responsive as the number of proteins induced more than 2 fold was higher in MoMΦ than in MoDC ( Figure 3B). The MoMΦ also showed a higher mean fold induction than the MoDC when comparing the proteins induced in both cell types [see Additional file 1].

QRT-PCR
Since mass spectrometry was performed in cells originating from only two donor pigs (Table 1), proteomic expression data were therefore confirmed at the transcriptional level by quantitative real-time PCR to which additional samples from 3 different pigs were included. Genes selected for QRT-PCR verification included those in which   Table 2).
Out of these, 16 genes were induced both in MoMΦ and MoDC, 3 genes (VCAM1, HMOX1 and Serglycin) were significantly induced in MoMΦ only and OLDLR1 and CDC42 were significantly induced exclusively in MoDC. Out of the 16 genes up-regulated in both cell types, 5 the most up-regulated ones encoded cytokines (IL1β, AMCF2 (the porcine homolog of CXCL 5/6), CXCL2, IL8 and IL1α). In agreement with the proteomic data, the fold inductions determined by QRT-PCR were usually higher in MoMΦ than in MoDCs, despite the fact that the differences between MoMΦ and MoDC responsiveness reached statistical significance only in 2 genes after exposure to LPS and 7 genes after infection with S. Typhimurium. This also means that the responsiveness of MoMΦ and MoDCs to LPS was similar but DCs, unlike MΦ, did not extensively respond to S. Typhimurium infection.

Discussion
In this study we were interested in the interactions of porcine MoMΦ and MoDCs with LPS and S. Typhimurium. Both cell morphology and flow cytometry indicated that the culture conditions led to the differentiation of distinct cell types. Expression of surface molecules observed on MoDC was in agreement with the previous observations [10,16]. The up-regulation of CD14 in porcine alveolar macrophages following stimulation with LPS or Salmonella infection was also reported earlier [10,17].
CXCL2, IL1α, MARCKSL1, SOD2, PPA1 and ISG15 were also found among the top 10% of the most inducible genes of bone marrow derived porcine macrophages treated with LPS [15].
The differences between MoMΦ and MoDC responses were quite low. This could be caused by the fact that we selected the most inducible and highly abundant proteins, which had similar expression profiles following exposure to LPS or S. Typhimurium, and minority proteins responsible for the specificity of MoMΦ and MoDC remained unrecognized. The exceptions were represented by MARCKS-related gene/protein (MARCKSL1). This gene was expressed and induced in MoMΦ after exposure to both LPS and S. Typhimurium at a significantly higher rate than in MoDC. Finally, its induction in LPS stimulated MoMΦ was more than 40 fold indicating that this was a highly inducible gene.
MARCKSL1 (synonyms MLP, MacMARCKS, F52) is expressed mainly in the brain, reproductive tissues, and macrophages [29,30] and belongs to a family of the unstructured proteins that mediate cross-talk in cell-to-cell signaling. MARCKSL1 is involved in the regulation of cell migration, adhesion and phagocytosis, as well as in neurosecretion and brain development [31][32][33]. Transcription of MARCKSL1 is also strongly increased upon stimulation with bacterial LPS [34][35][36][37]. Mancek-Keber et al. [38] showed that the MARCKSL1 protein binds LPS with an affinity sufficient enough to sequester a large fraction of the internalized LPS or intracellular Gram-negative bacteria within the endosome/phagosome.

Conclusions
In this study we have shown that S. Typhimurium and LPS in particular represent one of the most potent activators of signaling pathways of MoMΦ and MoDC leading to a similar response in both cell types. Despite the similarity of the responses, compared to MoDC, the MoMΦ were more pro-inflammatory. This difference was less obvious in response to LPS and more obvious in response to S. Typhimurium. Our results showed that increased pro-inflammatory signaling by MoMΦ compared to MoDC might be, at least in some subsets of mononuclear phagocytes, associated with differential MARCKSL1 protein expression though the exact role of these proteins in differential responsiveness of MoMΦ and MoDC to S. Typhimurium infection remains to be determined.

Differentiation of monocyte-derived macrophages and dendritic cells
Blood was collected from healthy 8-10 month old conventional pigs kept at the Veterinary Research Institute for the educational purposes. The animal care protocol for this experiment followed the Czech guidelines for animal experimentation and was approved by the Branch Commission for Animal Welfare of the Ministry of Agriculture of the Czech Republic (permission No MZe 921). Peripheral Blood Mononuclear cells (PBMC) were isolated by gradient centrifugation using Histopaque-1077 (Sigma-Aldrich). Monocytes were sorted from PBMC using indirect magnetic labelling based on the expression of CD14. In the first step, anti-CD14 monoclonal antibody (MIL-2, IgG2b, AbD Serotec) was used and consequently, the CD14-positive cells were captured by goat anti-mouse IgG MicroBeads (Miltenyi Biotec) and sorted using a QuadroMACS separator (Miltenyi Biotec) according to manufacturer's recommendations. The purity of the sorted cells was checked using a flow cytometer (LSR Fortessa, Becton Dickinson) with more than 90% of the cells being CD14 positive. Approx. 5 × 10 5 CD-14 positive monocytes resuspended in 1 ml of media were seeded per well on a 24-well microplate.
Half of the isolated monocytes were differentiated using protocol described previously [39] and the resulting cells were considered as MoDC throughout the study. RPMI-1640 medium (PAA) supplemented with antibiotics penicillin/streptomycin (Sigma-Aldrich), 10% heatinactivated fetal calf serum (PAA), recombinant porcine IL4 (50 ng/ml, R&D Systems) and GM-CSF (20 ng/ml, R&D Systems) was used. The second half of the cells were left to adhere to plastic microplates in DMEM (Gibco Invitrogen) supplemented with penicillin/streptomycin (Sigma-Aldrich) and 10% heat-inactivated and filtered porcine serum (PAA) as described previously [40]. These cells were considered as MoMΦ for the purposes of this study. Both MoMΦ and MoDC were differentiated at 37°C in a 5% CO 2 incubator for 4 days. MoMΦ and MoDC were differentiated from peripheral blood monocytes from 8 different pigs, although not all of the cultures were used for all of the analyses (Table 1).

Fluorescence microscopy and flow cytometry
Differentiation of MoMΦ and MoDC was confirmed by flow cytometry and light microscopy. For light microscopy, the cells were grown on 13 mm glass slides as described above. After a 4-day culture, the cells were fixed with 4% paraformaldehyde and labelled with DAPI as a nuclear stain and Alexa Fluor (AF) 488 conjugated phalloidin (Invitrogen) to visualize the actin cytoskeleton. Microscopy was performed using epifluorescence inverted microscope Olympus IX81 equipped with PlanSAPO 40× (NA 0.95) objective using differential interference contrast (DIC), and fluorescence mode to detect DAPI and AF-488 fluorescence.

Bacteria and culture conditions
Salmonella enterica serovar Typhimurium 16E5 of porcine origin belonging to phage-type DT104 [41] was used in this study. Bacteria were grown statically in LB broth at 37°C for 18 hours. This culture was diluted 800 × in LB broth and incubated for an additional 6 hours at 37°C to obtain the bacteria in the late logarithmic growth-phase of highly invasive phenotype. Prior to infection of MoMΦ and MoDC, the bacteria were collected by centrifugation and re-suspended in PBS to OD 600 = 0.3.

Experimental infection
Prior to infection, the medium was replaced by serum free DMEM without antibiotics. MoMΦ and MoDC were infected with S. Typhimurium at a multiplicity of infection equal to 1 for 1 h. Free bacteria were then washed away and gentamicin was added to fresh medium (100 μg/ml) to kill any remaining extracellular bacteria. One hour later, the medium was replaced with fresh medium containing 15 μg/ml gentamicin to prevent multiplication of extracellular bacteria that were eventually released during culture from dead cells. LPS from S. Typhimurium (Sigma-Aldrich) at a concentration of 1 μg/ml was used as another stimulus. Negative controls included an assay performed without any contact with S. Typhimurium or LPS. Eighteen hours after infection or LPS stimulation, the extracellular transport of proteins was blocked by the addition of 10 μg/ml of Brefeldin A (Sigma-Aldrich). Six hours later, the cells were washed with PBS and lysed by TRI reagent (Sigma-Aldrich) for RNA and protein purification.

Mass spectrometry and proteome analysis
Proteins from cell cultures were lysed with TRI Reagent and precipitated from the phenolic phase with acetone according to the manufacturer's recommendations. Dried protein pellets were dissolved in 300 μl of 8 M urea and processed by modified FASP method [42] using a Vivacon 500 device with MWCO of 10 kDa (Sartorius Stedim Biotech). Dissolved proteins were washed twice with 8 M urea and reduced by 10 m M DTT. After reduction proteins were incubated with 50 mM iodoacetamide and washed twice with 25 mM triethylammonium bicarbonate buffer. Trypsin added to sample in 1:50 ratio (w/w) was used as digestive enzyme.
Tryptic peptides were labeled using the stable isotope dimethyl labeling method as described elsewhere [43]. Three combinations of formaldehyde and cyanoborohydride isotopomers were used for labelling; proteins from non-stimulated cells were labeled with a combination of CH 2 O and NaBH 3 CN (light tag), Salmonella-infected cells with CD 2 O and NaBH 3 CN (medium tag) and LPSstimulated cells with 13 CD 2 O and NaBD 3 CN (heavy tag). Prior to LC-MS/MS analysis, the labeled samples were mixed at a 1:1:1 ratio and analyzed in 3 independent LC-MS/MS runs using the Dionex UltiMate 3000 RSLC nano system connected to a Obitrap Velos Pro mass spectrometer (Thermo Scientific). Such setup enabled semi quantitative analysis of protein abundance. The injected sample was desalted and pre-concentrated during the first 3 min of LC-MS/MS run on the Acclaim PepMap C18 trapping column (2 cm × 75 μm, 3 μm particles) at a flow rate of 5 μl/min using the loading mobile phase consisting of 0.1% formic acid in 98/2 H 2 O/ACN (vol/vol). Chromatographic separation was performed on an EASY-Spray C18 separation column (15 cm × 75 μm, 3 μm particles) at a flow rate of 400 nl/min using the acetonitrile gradient. High resolution (30,000 FWHM at 400 m/z) MS spectra were acquired for the 390-1700 m/z interval. Low resolution MS/MS spectra were acquired in Linear Ion Trap in a data-dependent mannerthe top 10 precursors (in terms of abundance) were fragmented using CID fragmentation at a normalized collision energy of 35.
Data were analyzed using the Proteome Discoverer (v.1.4). MS/MS spectra identification was performed by searching SEQUEST against the Sus scrofa database (Uniprot, on 4th of September, 2013) and precursor and fragment mass tolerance for searches were 10 ppm and 0.6 Da respectively. Only peptides with a false discovery rate of <1% were included in quantification. Only proteins which quantification was based on 12 or more particular peptide quantifications were considered as reliable. The peptide quantifications were based on ratios of peptide peak areas in stimulated and nonstimulated cells.
RNA purification, reverse transcription and quantitative real-time PCR Total RNA was isolated from the water phase of TRIreagent using RNeasy kit (Qiagen) and immediately reverse transcribed using M-MLV reverse transcriptase (Invitrogen)