Interaction of Salmonella Typhimurium with Dendritic Cells Derived from Pluripotent Embryonic Stem Cells

Using an in vitro differentiation protocol we isolated cells with the properties of dendritic cells (DCs) from immunologically refractive pluripotent murine embryonic stem cells (ESCs). These ES-derived dendritic cells (ESDCs) expressed cytokines and were able to present antigen to a T cell line. Infection of ESDCs with Salmonella Typhimurium stimulated the expression of immune cell markers and thousands of murine genes, many associated with the immune response. Consequently, this system provides a novel in vitro model, amenable to genetic modification, for monitoring host/pathogen interactions.


Introduction
Dendritic cells (DCs) play a central role in the immune response, acting as sentinel cells for the detection and presentation of antigens to other immune cells. They are involved in the recognition and processing of antigens from potentially pathogenic microbes and are important for priming and stimulating immune cells involved in mediating protection against such agents. Conversely, invading microbes have developed mechanisms for subverting the activity of DCs. Some pathogens can interfere with the antigen presentation process or even stimulate innate death responses in DCs [1,2]. Others, such as Salmonella enterica, actually exploit activated DCs as portals of entry into their hosts, using them to traverse the epithelial barriers of the intestine [3,4].
Salmonella are potentially pathogenic for different vertebrate hosts, causing a variety of disease syndromes ranging from localized gastroenteritis to systemic typhoid [5]. One of the key pathogenic traits associated with the Salmonella is their ability to actively invade and persist within eukaryotic immune cells, including DCs [6,7,8]. These bacteria harbor specific sets of genes that facilitate the invasion and persistence processes and moderate vesicular trafficking within the targeted cells [9,10]. There is evidence that they can also moderate the immune response by interfering with the cellular responses leading to both innate and acquired immunity [11,12,13].
Both animal studies, such as those based on the mouse, and in vitro models using cultured cells have proven valuable for studying host-pathogen interactions, including those involving Salmonella. Genetically manipulated mice and eukaryotic cells have added an extra dimension to the versatility of these approaches [14]. Cell based studies have normally relied on populations of primary cells grown directly from the host or on transformed and established differentiated cells adapted for tissue culture in the laboratory. Another potential approach is to exploit pluripotent embryonic stem cells (ESCs) as the host, as these can potentially be differentiated along different lineage pathways or genetically manipulated to create targeted and defined mutant lines [15]. Early studies on pluripotent ESCs from mice have shown them to lack immunological competence. For example, murine ESCs exposed to different pathogenic bacteria, including Salmonella enterica serovar Typhimurium (S. Typhimurium), have no obvious innate or other immune signature even though they are readily invaded by these bacteria [16]. Protocols have been developed for reproducibly differentiating pluripotent ESCs into immune cells, including those of the DC lineage [17]. These protocols enable the potential to study the immune response stimulated by pathogens in carefully controlled cellular environments, including the application of genetic modification. Here we report methods for the exploitation of ESC derived DCs for the investigation of the immune response to Salmonella. These approaches are potentially applicable to a range of pathogens and their products.

Bone Marrow DCs (BMDCs)
Mice were sacrificed by neck dislocation. Bone marrow from the tibia and femur of 5-10 week old 129/Sv mice were flushed using a syringe and needle containing RPMI-1640, 16 penicillin/ streptomycin and 2% FBS. The cells were allowed to settle for 5 min at room temperature, the supernatant was transferred in a fresh tube and bone marrow collected by centrifugation. The cells were suspended and cultured in Iscoves modified Dulbecco's medium (IMDM) with 10% FBS, 2 mM L-Glutamine, 1 mM Sodium Pyruvate, 16 MEM-NEAA, 2-b-Mercaptoethanol and penicillin/streptomycin. The medium was supplemented with differentiation factors GM-CSF at 500 WHOSU/ml and IL-4 at 150 WHOSU/ml. The bone marrow derived DCs were harvested at 6 and 9 days for analysis. The procedure was ethically approved by the Wellcome Trust Sanger Institute's Ethical Review Committee and all animal work was carried out in accordance with UK Home office regulations under licence number 80/2099.

Flow Cytometric Analysis of Surface Cell Markers
The mouse ESC line AB2.2 and the embryonic stem cell derived DCs were washed with PBS, incubated with cell dissociation buffer (Sigma), centrifuged and fixed in 1% paraformaldehyde in PBS for 20 minutes at 4uC, washed with PBS and incubated with specific antibodies including CD44 PE-Cy5conjugated (BD Biosciences); Integrin a-6 (CD49f, R&D Systems) and CD11b APC-conjugated (BD Biosciences); Octamer-Binding Protein 3/4 (OCT3/4, R&D Systems), Stage-Specific Mouse Embryonic Antigen 1 (SSEA-1), CD11c and CD80 PE-conjugated; CD86, IA/IE and H-2K FITC-conjugated (all from BD Biosciences). Cellular membranes were permeabilized with saponin buffer for intracellular staining when needed. Similar isotype matched antibodies conjugated to the same fluorophore were employed as negative controls (BD Biosciences). Surface marker expression analysis was performed using a FACS Aria I (BD Biosciences).

Bacterial Growth and ES Cell Infection
S. Typhimurium SL1344(p1C/1), which expresses GFP under the control of the promoter ssaG [22], were used to infect cells. In order to carry out in vitro infection, S. Typhimurium SL1344(p1C/ 1) were grown from a single colony in Luria Bertani broth with 50 mg/ml Ampicillin selection at 37uC over night in static conditions until OD 600 , 0.6-0.7. The bacterial suspension was diluted with culture medium to a desired multiplicity of infection (MOI) of 100 and added to the culture cells. The in vitro cells were incubated with bacteria for 30 minutes at 37uC and after 2 washes with PBS they were incubated with culture media containing Gentamicin at 50 mg/ml for the indicated time.

Confocal Microscopy
To track bacterial intra-cellular location, ESC derived DCs were grown on glass coverslips treated with poly-L-lysine and infected with S. Typhimurium SL1344(p1C/1). At the indicated time, the cells were washed and fixed with 1% paraformaldehyde buffer. They were then permeabilized with saponin buffer and stained with primary antibodies against Early Endosome Marker 1 (EEA-1) (Rabbit polyclonal primary antibody, AbCam), Lysosome Associated Membrane Protein 1 (LAMP-1) (Rat monoclonal primary antibody, AbCam) or LAMP-2 (Rat monoclonal primary antibody, AbCam) respectively following the manufacturer's recommended dilutions, incubated for 30-40 minutes, prior to being washed. Finally they were incubated in the dark with secondary goat anti-rabbit or goat anti-rat antibodies conjugated to APC-Cy7 (Santa Cruz Biotechnology) according to the manufacturer's instructions. After washing, the coverslips were mounted on glass slides with ProLong Gold containing DAPI (Invitrogen).

Electron Microscopy
Cells were fixed in culture medium containing 2.5% glutaraldehyde and 4% paraformaldehyde at 37uC for 5 minutes and then transferred to ice for 25 minutes. Cells were then scraped from the dish with a freshly cut Teflon strip and transferred to an Eppendorf centrifuge in fresh cold fix, pelleted and fixed for a further 30 minutes. The pellet was carefully rinsed in 0.1 M cacodylate buffer and post-fixed for one hour at room temperature in 1% buffered osmium tetroxide followed by a brief rinse and then mordanted for one hour in 1% buffered tannic acid, briefly rinsed in 1% aqueous sodium sulphate and dehydrated in an

Cell Maturation and Cytokine Measurement
Bone marrow and ESC derived DCs were incubated overnight with the activating agents S. Typhimurium LPS (Sigma) at 10 mg/ ml or Tumour Necrosis Factor a (TNFa) (R&D Systems) at 5000 U/ml and cytokine measurements were carried out on the culture supernatants. These cells were also used for confocal imaging. Undifferentiated ESCs and ESDCs were harvested and incubated overnight then exposed to S. Typhimurium for 2 and 4 hours at a multiplicity of infection or MOI 100 and their ability to secrete cytokines into the supernatant was tested. Cytokine concentrations were measured using Mouse Cytometric Bead Array (CBA) kit from BD Biosciences that detects and measures the concentration of the cytokines Interleukin 6 (IL-6), IL-10, IL-

12, TNFa, Monocyte Chemotactic Protein-1 (MCP-1) and
Interferon c (IFNc) captured onto beads. The experiment was performed following the manufacturer's instruction. The cytokine bound beads were washed once in PBS and analyzed with a FACS Aria I. The data obtained was processed with FCAP Array software. T cell activation was measured by the production of IL-2 upon antigen presentation using mouse CBA IL-2 Flex Kit (BD Biosciences).

Antigen Presentation Assay
Allogeneic T cell hybridoma MF2.2D9 (specific for I-A b+ OVA258-276 peptide Major Histocompatibility Complex (MHC) class II restricted peptide, gift from Dr B. Chain, University College London) were seeded with DCs at a 5 to 1 ratio in 96 well flat bottom plates (in triplicate). Whole ovalbumin protein was added at 10 mg/ml. As a positive control MF2.2D9 T cells were supplemented with 5 mg/ml of Concanavalin A (ConA). All cells were incubated for 24 hours at 37uC, 5% CO 2 and the supernatants were obtained without disturbing the cells and stored at 220uC. Levels of IL-2 expressed by activated T cells were measured by using a mouse CBA IL-2 Flex Kit assay (BD Biosciences).

Cell Sorting and Total RNA Extraction
Infected and uninfected ESC derived DCs were sorted by FACS (Aria cell sorter I, BD Biosciences). Single cell suspensions were obtained using cell dissociation buffer and the cells were maintained in PBS during sorting. The cells were sorted into sterile 15 ml Falcon tubes containing 2 ml of RNAlaterH solution (Ambion). Total RNA was extracted using a RNeasy Mini kit (QIAGEN). Additionally, an uninfected sample of cells was also sorted to obtain total RNA from uninfected cells. The total RNA quantity was measured by Spectrophotometer NanoDrop-1000 v 3.1.0 (Thermo Scientific) and the RNA quality was analyzed using a 2100 Bioanalyzer (Agilent) following the manufacturer's instructions. The differentiation protocol was performed at least three times in order to carry out infection and cell sorting for total RNA extraction.
indicates , 20% of the cells were positive;+indicates that at least 30% of the cells were positive;++indicates $60% of the cells were positive;+++indicates that $80% of the cells were positive for the marker tested. doi:10.1371/journal.pone.0052232.t002

Microarray Hybridization and Data Analysis
Microarray chip hybridization was carried out following the manufacturer's instructions. Briefly, Biotin-labeled cRNA was synthesized from 300 ng total RNA with the TotalPrep RNA Amplification kit (Ambion, Foster City, CA, USA). The hybridization mix containing 1500 ng of labeled cRNA was prepared according to the Illumina BeadStation 5006 System Manual (Illumina, San Diego, CA, USA). The hybridization of Illumina MouseWG-6 v1.1 Expression BeadChip array was performed overnight at 58uC on the BeadChip Hyb Wheel. The BeadChips were washed with E1BC solution for 15 minutes, and then washed in 100% Ethanol for 10 minutes and after a final wash in E1BC for 2 minutes they were treated with Block Buffer E1 on a rocker mixer for 10 minutes. The staining reaction was performed using 2 ml of buffer E1 with 1 mg/ml of Cy3-streptavidin followed by a 5 minutes' wash with buffer E1BC. The BeadChips were then dried by centrifugation and scanned using the Illumina BeadStation 500 platform operated by Illumina BeadArray Reader software. For each time point, three independent sample replicates were hybridized to three different chips.
Microarray data were analyzed using Bioconductor [23] to perform a pair-comparison analysis between mRNA expressed by uninfected cells versus infected cells at 2 hours and 4 hours post infection. The data obtained from Bioconductor analysis were further analyzed for cellular Pathways enrichment using InnateBD database [24]. The corresponding human orthologs were determined and the results were enriched using the 'over-representation analysis' option which allows selection of genes on the basis of their fold change (+/21.5) and p-value (0.05). The default settings for the analysis algorithm (Hypergeometric) and the multiple testing correction method (The Benjamin & Hochberg correction for the False Discovery Rate (FDR)) were used. Only pathways that are significantly up-or down-regulated with p-value ,0.05 are reported here.

Semi-quantitative Real Time-PCR
The cDNA was synthesized using QuantiTect Rev. Transcription kit (Qiagen) from 30 ng of total RNA following the manufacturer's instructions. Primers were designed using primer3 website [http://frodo.wi.mit.edu/primer3]. Primers are shown in Table 1. The RT-PCR reactions were performed using SYBR green dye from SensiMix Plus SYBR kit (Quantace) on a Mx3000P RT-PCR machine (Stratagene).
Each sample was run in triplicate for each biological replicate. To test the primers for specificity and quality, efficiency curves were performed and dissociation products were obtained. The Ct (threshold cycle) values were analyzed for relative quantification using ribosomal protein S18 transcript as housekeeping control gene. The results are reported as Ln2 ' 2(DDCt).

Statistics
Colony formation units (CFUs) were reported as mean of three counts +/21 standard deviation (SD). Cytokine concentrations were analyzed using one-way ANOVA test. Semi quantitative real time RT-PR data were analyzed using RESTß software [25].

Differentiation and Characterization of DCs from ESCs
In a previous study we characterized the interactions between undifferentiated pluripotent mouse ESCs and a number of pathogenic bacteria, including S. Typhimurium and Shigella flexneri [16]. Interestingly, although S. Typhimurium was able to invade and persist within such ESCs there was virtually no evidence for the activation of genes associated with innate or adaptive immune responses. This observation encouraged us to investigate if ESCs could be differentiated into a cell type that can mount a pronounced immune response following exposure to this pathogen. DCs play a key role in the pathogenesis of salmonellosis and the subsequent host immune response. Thus, we explored these interactions following differentiation of ES cells along the DC lineage [26].
Undifferentiated mouse ES cells derived from the line AB2.2 were initially characterized for surface marker expression by flow cytometry. Such cells expressed markers characteristic of pluripotent and self-renewing ESCs including Oct3/4, Integrin a-6 and SSEA-1 whilst being negative for CD44, a marker used to confirm the undifferentiated status of pluripotent and self renewing cells (Fig. 1A). These cells were also analyzed for the expression of other markers normally associated with populations of leukocytes including CD11b, CD11c, CD45, CD80 (Fig. 1B) CD4, CD8, CD86, H-2K and IA-IE (Table 2). They were found to be generally negative for all these markers, although low levels of CD40 and CD54 were detected on some cells in the population. Next, to drive differentiation down the DC lineage, single ESCs were seeded under conditions lacking LIF for at least 10 days, the time necessary for the development of embryoid bodies. Subsequently these embryoid bodies were incubated with GM-CSF and IL-3 as described previously to drive them along the DC lineage. After about 4 days of incubation in the conditioned media, cells resembling DCs started to differentiate from the edges of the adherent embryoid bodies. Under the light microscope the ESC derived DCs were visible as bright cells growing in small clusters budding all around the culture dish. They were recognizable by characteristic comma-like protrusions. Resting dendritic-like cells were visible around both beating and non-beating embryoid bodies. Observations made using confocal and electron microscopy revealed striking morphology differences between undifferentiated ESCs and the corresponding DCs. Undifferentiated ESCs grew attached to a layer of gelatin, close to each other and often one on the top of the other. Once differentiated into DCs, they were visible via the microscope either as single cells or organized in small clusters. ES derived DCs had long and thin protrusions branching from the cell body, sometimes as dendritic extensions and other times as obvious lamellipodia ( Fig. 2A).
There was a striking difference in the organelles present in undifferentiated ESCs compared to the ES derived DCs when they were observed using electron microscopy ( Fig. 3A and B respectively). The much larger ESDC (Fig. 3B) had a different nuclear/cytoplasm volume ratio and exhibited relatively increased numbers of cellular organelles including mitochondria and Golgi. In addition, features resembling MHC class II associated multilaminar compartmental structures were observed in ESDCs but not in undifferentiated ESCs. Similar compartments were described previously in the DC line CB1 [12]. These data provide a visual confirmation that ES cells exhibit a different overall morphology after differentiation.
Populations of ESC derived DCs were harvested by vigorous pipetting and the detached cells were characterized for the expression of selected surface markers by flow cytometry. The ESC derived DCs had enhanced expression of multiple surface markers including CD11b, CD11c, CD45, CD80 (Fig. 1C), CD4, CD40, CD54, CD80, H-2K and selected Toll like receptors (TLRs) ( Table 2) compared to undifferentiated ES cells (Fig. 1B and Table 2). Interestingly, these cells did not express detectably . Gentamicin invasion assay. The ability of S. Typhimurium SL1344(p1C/1) to invade ESDCs was measured by the gentamicin assay described in Methods. ESDCs were exposed to either wild type S. Typhimurium or a SipB mutant for 30 minutes and then incubated with media containing 50 mg/ml of Gentamicin for 2 or 4 hours and then invading bacteria were enumerated. doi:10.1371/journal.pone.0052232.g004 higher levels of some markers typical of activated classical DCs such as IA-IE and CD86.
Next, harvested ESDCs cells were incubated with either Tumor Necrosis Factor a (TNFa) or lipopolysaccharide (LPS) [27]. After overnight incubation with LPS, the ESDCs exhibited increased expression levels of some markers including H-2K CD11c, CD11b, CD4, CD40, CD54, TLR4 and MHC Class II ( Table 2, Fig. 2B), compared to untreated ESDCs. Following treatment with TNFa, these cells exhibited increased expression of similar markers ( Table 2). For comparative purposes flow cytometry was   performed on murine BMDCs prepared as described in Methods. These BMDCs expressed all the markers used (Table 2).

Antigen Presentation to T cells
A key characteristic of DCs is their ability to process and present specific antigens to naïve T cells using MHC Class II. Consequently, ESDCs were tested for their ability to process and present antigens by co-incubating them with the allogeneic naïve T cells line MF2.2D9 (I-A b+ OVA258-276 peptide-Class II specific T cell hybridoma) with or without whole ovalbumin. The cell line MF2.2D9 can respond to exposure to ovalbumin antigen presented on MHC class II by producing IL-2. The concentration of IL-2 in the culture supernatants was thus assessed after 24 hours of incubation with ESDCs and ovalbumin. BMDCs were used as positive controls. The level of IL-2 produced by the T cells mixed with either the ES cell derived dendritic cells or BMDCs was quantified by bead assay (Table 3, representative of independent experiments). Prolonged time incubation to 48 hours did not significantly increase the concentration of IL-2 (data not show). These experiments confirmed the ability of ESC derived DCs to specifically activate antigen specific T cells line.

Interaction of ESDCs with S. Typhimurium SL1344(p1C/1)
The ability of S. Typhimurium SL1344(p1C/1) to invade ESC derived DCs was investigated using a combination of microbiological counts, FACS and microscopy. Microbiological count of S. Typhimurium SL1344(p1C/1) was carried out using an optimized multiplicity of infection (MOI) of 100 bacteria per cell in each well. S. Typhimurium(p1C/1) sipB, a non-invasive mutant, was used as a control. Fig. 4 shows the colony forming units (CFUs) recovered at 2 and 4 hours post infection in this in vitro assay, following Gentamicin selection for intracellular bacteria. The number of   Metabolism of amino acids and derivatives; 4 Respiratory electron transport, APT synthesis by chemiosmotic coupling, and heat production by uncoupling proteins; 5 Integration of energy metabolism; 6 Mitochondrial fatty acid beta-oxidation. doi:10.1371/journal.pone.0052232.t006 bacteria recovered increased between 2 hours and 4 hours post infection but was lower after 24 hours of incubation (data not shown), due to increasing cell death of ES cell derived DCs [28]. FACS was carried out using S. Typhimurium SL1344(p1C/1), which harbors an in vivo activated promoter that drives the expression of Green Fluorescent Protein (GFP) once inside the host cell (Fig. 5). These data showed that more than 30% of ESDCs consistently harbored intracellular bacteria at 2 hours post-infection and between 35 and 42% at 4 hours post-infection. ESDCs were tightly adherent to the plastic after exposure to S. Typhimurium SL1344(p1C/1). As early as 30 minutes after infection S. Typhimurium SL1344(p1C/1) were found co-localized with the EEA1 (Fig. 6.A) as has been previously described in BMDCs [29]. In addition at 3 hours post-infection the bacteria were often found co-localized with the late lysosome marker LAMP1 (Fig. 6.B).

ESDCs Express Cytokines before and during Salmonella Infection
One important characteristic of DCs is the ability to produce cytokines and present antigens to other immune cells. For this reason we monitored the ability of cultured ESDCs to produce a number of cytokines including TNFa, IFNc, IL-6, IL-12 subunit 70 (IL-12p70), MCP-1 and IL-10 in response to exposure to S. Typhimurium SL1344(p1C/1). Once differentiated into ESDCs the supernatants were sampled at 24 hours after harvesting and at 30 minutes, 2 hours and 4 hours after S. Typhimurium SL1344(p1C/1) infection. Low levels of MCP-1 and TNFa were already present in the culture supernatant of uninfected ES cell derived DCs (Table 4). After 30 minutes of S. Typhimurium SL1344(p1C/1) infection the supernatants from ESC derived DCs harbored IL-6, MCP-1 and TNFa. Statistical analysis showed a significant difference in the ESDCs expression at all infection time points for MCP-1 and at 4 hours infection of IL-6 and TNFa compared to undifferentiated ESCs. At 2 and 4 hours post infection, detectable levels of IL-10 were also present but this was not statistically significant. However, little or no IFNc or IL-12p70 was detectable at any time during the S. Typhimurium SL1344(p1C/1) infection of the ESC derived DCs (Table 4). It is striking to note that no detectable levels of cytokines were produced by undifferentiated ESC during the time points here reported (Table 4 and reference 17).
The Transcriptome of Murine ESDCs Infected by S. Typhimurium SL1344(p1C/1) Once we had established that ESDCs have the ability to present antigen, their RNA expression profile was investigated during S. Typhimurium SL1344(p1C/1) infection (Table 5 and 6). Salmonella infected cells were isolated from uninfected ESDC by fluorescent activated cell sorting into RNAlater at 2 hours and 4 hours postinfection by exploiting the GFP tag associated with S. Typhimurium SL1344(p1C/1). The mRNA was extracted and cDNA was synthesized and hybridized on Illumina arrays carrying 35,000 mouse probes and the data were analyzed by pair-comparison using Bioconductor. This experimental analysis highlighted the ability of the differentiated cells to mount a strong response to bacterial invasion with 3615 genes significantly differentially expressed with p,0.01 at 2 hours post-infection. Of these genes, 325 had a .2 fold difference in expression levels and confidence p value ,0.0001, compared to uninfected cells. One hundred genes showed a .4 fold differential expression with p-value ,0.01 (one of which was down regulated). At 4 hours post-infection more than 4000 genes were significantly differentially expressed with p,0.01.
Of these genes, 525 had an expression difference .2 fold compared to uninfected cells with a confidence p value of ,0.001. One-hundred and 75 genes showed a .4 fold difference in expression with p-value ,0.01. The Bioconductor analysis results were confirmed by semi-quantitative Real Time-polymerase chain reaction (PCR) performed on selected genes including those encoding IL-6, H2-Dma and Cxcr4 and these data were in agreement with the microarray analysis (Fig. 7). These included Oasl1 (2'-5' oligoadenylate synthetase-like 1) a gene involved in the innate immune response, Tyki (cytidine monophosphate (UMP-CMP) kinase 2, mitochondrial) an LPS-inducible member of the thymidylate kinase family and IL-6, which is produced by activated DC in vivo [30] and plays a role in DC maturation [31].
The lists of differentially expressed genes obtained by Bioconductor analysis were further analyzed for pathway enrichment analysis using a specialist software package known as InnateDB designed for the focused assessment of immune-associated pathways [24]. The analysis was performed on corresponding human orthologue genes with the fold expression and the p-value reported by Bioconductor analysis at 2 and 4 hours post-infection. This analysis revealed that at 2 hours post-infection, pathways that were significantly differentially expressed (data not shown) included those termed as 'cytokine-cytokine receptor interaction', 'MAPK signaling pathway', 'Jak-STAT signaling pathway', 'IL-12 mediated signaling events' and 'hematopoietic cell lineage'. These pathways have been previously associated with the early response to pathogens [32,33]. The InnateDB analysis did not reveal any statistically significant down-regulated pathways at 2 hours post infection. Pathway analysis of the genome expression profile at 4 hours compared to 2 hours post-infection reported 'arachidonic acid metabolism' as being down regulated. However, the InnateDB analysis of expression profile at 4 hours post-infection comparing S. Typhimurium SL1344(p1C/1) versus uninfected ESC derived DCs revealed a number of noteworthy pathways with 18 pathways up-regulated (Table 5) and 40 pathways downregulated ( Table 6). The analysis reported several pathways that are part of the same Network, for example 'Carbohydrate Metabolism', 'DNA-replication' and 'Metabolism of amino acids and derivatives'. Up regulated pathways included those termed 'cytokine-cytokine receptor interaction', 'IL-4', 'IL-23' and 'IL-12 mediated signaling events'. The activation of these particular pathways confirms the immune signature associated with these differentiated ESCs. In particular the identification of the IL-12 pathway suggests that this cytokine although not detected by quantification methods is nevertheless expressed. It is possible that this protein is expressed in low concentration not detectable by our assays. IL-23 is part of the IL-12 family and shares the subunit p40 with IL12 itself [34]. These two cytokines can have a proinflammatory effect connecting innate and adaptive immune responses via the activation and differentiation of T cells [35,36]. Patients with inherited deficiency of these cytokines are more susceptible to salmonellosis [37]. IL-4 pathway, which is used for in vitro differentiation into DCs was also up-regulated during Salmonella infection [38,39]. Other interesting pathways include the 'Vpu mediated degradation of CD4'. Vpu is an HIV protein that, on association with CD4, induces self degradation by the proteasome. It is possible that Salmonella induce CD4 degradation in a similar way as suggested by the observation of the lack of CD4 positive DCs at 5 days post-Salmonella infection in B cell follicle borders in the spleen [40]. Among the downregulated pathway at 4 hours post-infection reported by the InnateDB analysis, many can be connected to a few network ( Table 6) including 'pyruvate metabolism and TCA cycle', 'Energy metabolism', 'Fatty acid b-oxidation'. These pathways are all involved in the mitochondrial metabolism and Salmonella is known to target this organelle with specific effector proteins [41]. The variety of Interaction Nodes reported to be up-regulated in our analysis highlight the deep remodeling process that the differentiated cells undergo during bacterial invasion in contrast to undifferentiated ESCs.

Discussion
We showed previously that undifferentiated ESCs can be infected with bacterial pathogens including S. Typhimurium and S. flexneri [16]. However, such infected cells exhibited no obvious immune signature and consequently they are not ideal to explore immune responsiveness to invading bacteria. DCs are specialized sentinel cells of the immune system, able to detect the presence of pathogens. Therefore, here ESCs were differentiated along this lineage and the impact of exposure to the invasive S. Typhimurium SL1344(p1C/1), which is known to target and exploit such cells, was investigated. We showed through a combination of flow cytometric analysis, microscopy and transcriptome profile that the ESDCs respond robustly to exposure to S. Typhimurium SL1344 and associated stimuli such as LPS or TNFa. They were able to express immune surface markers, secrete cytokines and even present MHC class II antigen to activate T cells hybridoma cell line in an in vitro assay. These ESDCs were CD8 2 and expressed markers such as CD4, CD11b, and CD11c. Based on these characteristics [42], in vitro differentiated ESDCs resemble in vivo DCs present in the spleen and in the intestinal Peyer's patches. Evidence from transcriptomic and cytokine analysis indicates that an array of different immune pathways was activated by exposure to pathogenic insult by ESDCs. Interestingly, we were unable to quantify IL-12 protein even though evidence from the transcriptome analysis suggested the gene is activated during infection. It is possible that deriving from embryonic cells, ESDCs may be more 'primitive' or less differentiated than similar cells developing in vivo during hematopoiesis. Nevertheless the potential of this model is evident and in the future it may be possible to optimize protocols to obtained specific subsets of specialized DC from murine and perhaps human induced pluripotent stem cells (iPSC) opening the opportunity to investigate the interaction of pathogenic bacteria with human DCs [43]. These approaches may also find utility in the exploitation of mutant mouse and human ESC lines [44,45,46]. In conclusion, we believe that this stem cell based approach has a great potential in the future studies of hostpathogen interactions.