Molecular Consequences of Proprotein Convertase 1/3 (PC1/3) Inhibition in Macrophages for Application to Cancer Immunotherapy: A Proteomic Study*

Macrophages provide the first line of host immune defense. Their activation triggers the secretion of pro-inflammatory cytokines and chemokines recruiting other immune cells. In cancer, macrophages present an M2 anti-inflammatory phenotype promoting tumor growth. In this way, strategies need to be develop to reactivate macrophages. Previously thought to be expressed only in cells with a neural/neuroendocrine phenotype, the proprotein convertase 1/3 has been shown to also be expressed in macrophages and regulated as a function of the Toll-like receptor immune response. Here, we investigated the intracellular impact of the down-regulation of the proprotein convertase 1/3 in NR8383 macrophages and confirmed the results on macrophages from PC1/3 deficient mice. A complete proteomic study of secretomes and intracellular proteins was undertaken and revealed that inhibition of proprotein convertase 1/3 orient macrophages toward an M1 activated phenotype. This phenotype is characterized by filopodial extensions, Toll-like receptor 4 MyD88-dependent signaling, calcium entry augmentation and the secretion of pro-inflammatory factors. In response to endotoxin/lipopolysaccharide, these intracellular modifications increased, and the secreted factors attracted naïve T helper lymphocytes to promote the cytotoxic response. Importantly, the application of these factors onto breast and ovarian cancer cells resulted in a decrease viability or resistance. Under inhibitory conditions using interleukin 10, PC1/3-knockdown macrophages continued to secrete inflammatory factors. These data indicate that targeted inhibition of proprotein convertase 1/3 could represent a novel type of immune therapy to reactivate intra-tumoral macrophages.

phenotype is characterized by filopodial extensions, Tolllike receptor 4 MyD88-dependent signaling, calcium entry augmentation and the secretion of pro-inflammatory factors. In response to endotoxin/lipopolysaccharide, these intracellular modifications increased, and the secreted factors attracted naïve T helper lymphocytes to promote the cytotoxic response. Importantly, the application of these factors onto breast and ovarian cancer cells resulted in a decrease viability or resistance. Under inhibitory conditions using interleukin 10, PC1/3-knockdown macrophages continued to secrete inflammatory factors. These data indicate that targeted inhibition of proprotein convertase 1/3 could represent a novel type of immune therapy to reactivate intra-tumoral macrophages. Molecular & Cellular Proteomics 14 Innate immunity is the first line of immune defense and is common to all metazoans (1,2). In this immune system, macrophages play a crucial role in the maintenance of tissue homeostasis. These cells are involved in almost every disease through their immunological and wound-healing functions (1,2). During a pathogenic infection, trauma or neurodegeneration, macrophages are recruited and activated contributing to the phagocytosis of pathogens and the secretion of cytokines and chemokines activating other immune cells. Macrophages can develop into classically pro-inflammatory (M1) or alternatively (M2) activated macrophages. M1 macrophages are characterized by the secretion of pro-inflammatory cytokines whereas M2 macrophages secrete anti-inflammatory cytokines (3). Stimulation of macrophages with LPS activates TLR4 signaling leading to the nucleus translocation of NF-B or IRF3 which activate genes encoding proteins involved in innate immune response (4). Many of these proteins are secreted (cytokines, chemokines…) to attract and activate other immune cells like T lymphocytes. In tumors, macrophages are oriented toward the M2 phenotype and promote cancer growth by suppressing immune cells function (5). Current research in the therapeutic field focus on ways to reactivate macrophages.
Surprisingly, we have shown that during immune responses, macrophages secrete typical neuroendocrine molecules (6 -8), such as neuropeptides (9) or the proprotein convertases (PC) 1 PC2 and PC1/3 and that PC1/3 is an important regulator of innate immune responses (10 -12). Proprotein convertases cleave precursor proteins which can lead to the activation, inactivation or functional changes. PC2 and PC1/3 operate within the regulated secretory pathway. Their expression is not restricted to neuroendocrine tissues, they are also expressed in macrophages and lymphocytes (12). In a previous study from our group, PC1/3 knockout (KO) in mice challenged with LPS caused innate immune defects and uncontrolled cytokine secretion (10). Th1 pathway is enhanced in PC1/3 KO mice. Following LPS treatment, PC1/3 colocalized with TLR4 in the endosomal compartment (11). We concluded that PC1/3 contributes to the regulation of TLR4 signaling and the resulting cytokine secretion.
The NR8383 rat pulmonary macrophage cell line was previously shown as a good model to study the role of PC1/3 in the macrophage innate immune response (13). In the present study, we developed a PC1/3-knockdown (K D ) NR8383 cell line using lentiviral-delivered shRNAs. Our aim is to understand the cellular impact of PC1/3 inhibition in macrophages and the consequences on their activation. Proteomic analysis of secreted proteins allowed us to identify pro-inflammatory cytokines and alarmins already at 24h of LPS stimulation in PC1/3-K D secretomes which was confirmed by cytokines array. Proteomic studies of PC1/3-K D NR8383 cellular extracts revealed an important perturbation in the intracellular trafficking machinery through the disorganization of cytoskeletal protein expression. These results were confirmed on macrophages from PC1/3 KO mice. Cytokines secretion and cytoskeleton reorganization can be linked to intracellular calcium increase in PC1/3-K D cells. Moreover, we showed that MyD88-dependant TLR4 signaling was sustained when PC1/3 is down-regulated. We describe here that inhibition of PC1/3 induced classically activated phenotype (M1) in macrophages. The chemotactic and anti-tumor properties of the PC1/3-K D macrophage secretome promoted the cytotoxic immune response and inhibited cancer cell viability. The down-regulation of PC1/3 could be used in cancer immunotherapy to reactivate macrophages.
Confocal Microscopy-The NR8383 cells were grown in culture flasks and treated or not with LPS (InvivoGen, Toulouse, France) at a concentration of 200 ng/ml before being subjected to immunofluorescence studies. For actin immunostaining, the cells were fixed with 4% paraformaldehyde (PFA) for 10 min, washed with PBS, permeabilized with 0.2% Triton X-100 for 10 min at room temperature, blocked with 1% BSA, 1% OVA and 1% normal donkey serum for 1 h and stained with phalloidin labeled with rhodamine (1/100, Santa Cruz Biotechnology, Heidelberg, Germany) at 4°C for 30 min. After washing with PBS, the nuclei were stained with Hoechst 33342 (1/10000), and the cells were visualized by confocal microscopy. Fluorescence analysis was conducted using a Zeiss LSM 510 confocal microscope (488 nm excitation for Alexa 488 and 543 nm for Alexa 546) connected to a Zeiss Axiovert 200 M with a 63X1.4 numerical aperture oil immersion objective. Both channels were excited, collected separately and then merged to examine the colocalization. The image acquisition characteristics (pinhole aperture, laser intensity, scan speed) were the same throughout the experiments to ensure comparability of the results.
Identification of Cytokines and Chemokines Using Rat Cytokine Antibody Arrays-NR8383-K D and -NT cells were plated on sterile six-well plates until confluence was attained. The cells were starved overnight with Ham's F12K medium supplemented with 2% FBS and stimulated for 24 h with 20 ng/ml IL-10 (Peprotech) in serum-free medium or left untreated. The medium was then replaced, and the cells were stimulated for 24 h with 200 ng/ml LPS or left untreated. The cell supernatants were collected, centrifuged at 500 ϫ g, passed through a 0.22-m filter to remove cells and immediately frozen in liquid nitrogen.
The Rat Cytokine Array Panel A from R&D system was used to probe the cytokines in the secretome of stimulated and nonstimulated NR8383 cells by following the procedures recommended by the manufacturer. The membranes were quantified by densitometry using ImageJ software. Statistical analyses were performed using the paired t test. Error bars represent the S.E.
Total and Nuclear Protein Extracts-NR8383-K D and -NT cells were plated on sterile six-well plates until confluent. For LPS stimulation, the cells were starved overnight with Ham's F12K medium supplemented with 2% FBS. The cells were stimulated with 200 ng/ml LPS in serum-free medium or left untreated. At 1 h, 3 h, 6 h (for Western blot analysis), and 24 h (for FASP), the cells were collected, washed once with ice-cold PBS and then lysed with RIPA buffer for total protein extraction (150 mM NaCl, 50 mM Tris, 5 mM EGTA, 2 mM EDTA, 100 mM NaF, 10 mM sodium pyrophosphate, 1% Nonidet P-40, 1 mM PMSF, 1X protease inhibitors). Cell debris was removed by centrifugation (20000 ϫ g, 10 min, 4°C); the supernatants were collected. For the nuclear extracts, NE-PER Nuclear and Cytoplasmic Extraction Reagents were used (Thermo Scientific) according to the manufacturer's instructions. The supernatants were collected, and the protein concentrations were measured using the Bio-Rad Protein Assay.
Filter-aided Sample Preparation (FASP)-The total protein extract (0.1 mg) was used for FASP analysis as described previously (14). We performed FASP using Microcon devices YM-10 (Millipore) before adding trypsin (Promega) for protein digestion (40 g/ml in 0.05 M NH 4 HCO 3 ). The samples were incubated overnight at 37°C. The digests were collected by centrifugation, and the filter device was rinsed with 50 l of NaCl 0.5 M. Next, 5%TFA was added to the digests, and the peptides were desalted with a Millipore ZipTip device before LC-MS/MS analysis.
Secretome Preparation and Protein Digestion-NR8383-K D and -NT cells were plated in sterile 24-well plates until confluent. For LPS stimulation, the cells were starved overnight with Ham's F12K medium supplemented with 2% FBS. The cells were stimulated with 200 ng/ml LPS in serum-free medium or left untreated. At 1, 16, 24, 48, and 72 h, the cell supernatants were collected, centrifuged at 500 ϫ g, passed through a 0.22-m filter to remove the cells and immediately frozen in liquid nitrogen. The experiments were performed in biological triplicates.
Four hundred microliters of the secretome was collected for each condition. The volume was reduced to 100 l in a SpeedVac. Secretome digestion was performed as previously described (15). In brief, the cell supernatants were denatured with 2 M urea in 10 mM HEPES, pH 8.0 by sonication on ice. The proteins were reduced with 10 mM DTT for 40 min followed by alkylation with 55 mM iodoacetamide for 40 min in the dark. The iodoacetamide was quenched with 100 mM thiourea. The proteins were digested with 1 g LysC/Trypsin mixture (Promega) overnight at 37°C. The digestion was stopped with 0.5% TFA. The peptides were desalted with a Millipore ZipTip device in a final volume of 20 l of 80% ACN elution solution. The solution was then dried using the SpeedVac. Dried samples were solubilized in water/0.1% formic acid before LC MS/MS analysis.
LC MS/MS Analysis-The samples were separated by online reversed-phase chromatography using a Thermo Scientific Proxeon Easy-nLC system equipped with a Proxeon trap column (100 m ID ϫ 2 cm, Thermo Scientific) and a C18 packed-tip column (75 m ID ϫ 10 cm, Thermo Scientific). The peptides were separated using an increasing amount of acetonitrile (5-35% over 100 min) at a flow rate of 300 nl/min. The LC eluent was electrosprayed directly from the analytical column, and a voltage of 1.7 kV was applied via the liquid junction of the nanospray source. The chromatography system was coupled to a Thermo Scientific Q Exactive mass spectrometer that was programmed to acquire in a data-dependent Top 10 method. The survey scans were acquired at a resolution of 70 000 at m/z 400.
Data Analyses-All MS data were processed with MaxQuant (25) using the Andromeda (26) search engine. The proteins were identified by searching MS and MS/MS data against the Decoy version of the complete proteome for Rattus norvegicus in the UniProt database (UniProt Consortium. Reorganizing the protein space at the Universal Protein Resource (UniProt. Nucleic Acids Res. 2012, 40 (Database issue), D71-5.) (Release June 2014, 33675 entries) combined with 262 commonly detected contaminants. Trypsin specificity was used for digestion mode, with N-terminal acetylation and methionine oxidation selected as variable, and carbamidomethylation of cysteines set as a fixed modification. We allowed up to two missed cleavages. For the MS spectra, an initial mass accuracy of 6 ppm was selected, and the MS/MS tolerance was set to 20 ppm for the HCD data. For identifi-cation, the FDR at the peptide spectrum matches (PSM) and protein level was set to 0.01. Relative, label-free quantification of the proteins was conducted using the MaxLFQ algorithm (27) integrated into Max-Quant with default parameters. The data sets and Perseus result files used for analysis were deposited at the ProteomeXchange Consortium (28) (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository (29) with the data set identifier PXD001984 for cellular extracts (For reviewer access only Username: reviewer62003@ebi.ac.uk; Password: zokVruKN) and PXD001986 for the secretome analyses (For reviewer access only Username: reviewer19925@ebi.ac.uk; Password: PUUJeVfO). Analysis of the identified proteins was performed using Perseus software (http:// www.perseus-framework.org/) (version 1.5.0.31). The file containing the information from the identification and hits from the reverse database were used, and proteins with modified peptides and potential contaminants were removed. The LFQ intensity was logarithmized (log2(x)). Categorical annotation of the rows was used to define the different group depending on the following: (1) the cell line (NT or K D ); (2) the treatment (Ctrl/LPS); and 3) the kinetics of the secretomes (1 h, 16 h, 24 h, 48 h, or 72 h). Multiple-sample tests were performed using ANOVA with a FDR of 5% and preservation of the group randomization. To evaluate the enrichment of the categorical annotations (Gene Ontology terms and KEGG pathway), Fisher's exact test was performed taking in account the results of the ANOVA for each group. Normalization was achieved using a Z-score with matrix access by rows. Only proteins that were significant by ANOVA were used for the statistical analysis. Hierarchical clustering was first performed using the Euclidean parameter for the distance calculation, and the average option for linkage in the rows and columns of the trees was used with a maximum of 300 clusters. To quantify fold changes in proteins across the samples, we used MaxLFQ. To visualize these fold changes in the context of individual protein abundances in the proteome, we projected them onto the summed peptide intensities normalized by the number of theoretically observable peptides. Specifically, to compare the relative protein abundances between and within samples, the protein length normalized to the log 2 protein intensities (termed the "iBAQ" value in MaxQuant) was included in the MaxLFQ differences. Functional annotation and characterization of the identified proteins were performed using PANTHER software (version 9.0, http://www.pantherdb.org) and STRING (version 9.1, http://stringdb.org). The GeneMANIA Cytoscape plugin (30) was used to generate co-expression networks from cell extracts proteomics data. A "basal" network composed of 95,886 recognized interactions was generated from the data obtained by the analysis of K D or NT cells under basal conditions. A supervised clustering was then performed to identify the top 100 molecules that coregulated with Anxa6 in this "basal network." The list of 100 genes which encoded molecules were identified was then assessed for gene set enrichment using EnrichR (31) and the GO classification. Following the same approach, two "LPS-stimulated" networks were then generated from the data obtained by the analysis of: (1) unstimulated or LPS-stimulated K D cells (89,017 recognized interactions) and (2) unstimulated or LPS-stimulated NT cells (96 563 recognized interactions). A supervised clustering was then performed to identify the top 100 molecules that coregulated with NFKB1 in each of these "LPS-stimulated" networks. Both lists of 100 genes that encoded molecules were identified and then assessed for gene set enrichment using EnrichR and the GO classification. Finally, subnetworks of genes presenting significant enrichments for specific GO terms were selected and visualized on Cytoscape. For presentation purposes, nodes were assigned equal weights and subnetworks were slightly distorted to avoid nodes superimposition.
Systems Biology Analysis-The altered pathways relevant to overexpressed proteins in K D PC1/3 cells were analyzed using Pathway Studio software v.10 (Ariadne Genomics, Rockville, MD). This soft-ware helps to interpret biological meaning based on gene (protein) expression, to build and analyze pathways, and to identify relationships among genes, proteins, cell processes, and diseases. This software contains a built-in resource named ResNet, which is a database of molecular interactions based on natural language processing of scientific abstracts in PubMed. Using ResNet, a researcher can analyze the gene product/protein list and build a pathway using well-known interactions that are discussed in the existing literature. The program searches the current pathway database and ResNet for interactions with the selected entities and then adds them to the pathway. After the new pathway was constructed, we were able to obtain more detailed information regarding the putative pathways that were altered in response to LPS treatment.
Data Analysis of Mouse Peritoneal Macrophages-An analysis of previously publish data was done to obtain information concerning modifications of proteins induced in peritoneal macrophages from wild type (WT) or PC1/3 KO (KO) mice challenged by LPS (8 h) or not (10). Briefly, peritoneal macrophage from wild type (WT) and PC1/3 knocked out (KO) mice were collected as described previously (10). Cytosolic and membrane fractions were then prepared according to (16) and processed for Gel-LC-MSMS as described in 10. The resulting tryptic peptides were purified and identified by reversed-phase chromatography coupled to an LTQ-Orbitrap Velos (Thermo Scientific). For identification, raw files were filtered for high quality spectra and converted to mgf files using scaffold software (4.4.5) (17) and analyzed using ProteinPilot software 4.5 (18). For each samples all gel band of a particular condition were combined for protein identification. FDR was calculated using PSPEP files (Sciex) and protein alignment was done with the protein alignment template 2.0 (Sciex) using a local protein FDR of 95% (19) over a combined list of all identified proteins. Only proteins with a score of over 6, which represents the proteins identified with two or more unique peptides and at least a fold change of 4 (converted log2 value) were kept for analysis. The relative protein expression was calculated based on the protein score, which was shown to be an adequate relative indicator of the relative differential expression (20). Ratio between WT and KO was calculated using the sum of corresponding fractions. Gene ontology analysis was performed using Blast2go (21).The network analysis was performed as follows: The gene names of identified proteins were used as input to retrieve a network from STRING (22), and this network was then loaded into Cytoscape 3.2 (23,24). with relative expression data using Id mapper. The Reactome FI pluggin was used to select a subnetwork of gene ontology terms and NCI database-associated specific proteins Calcium Imaging-Cells were plated onto glass coverslips and loaded with 4 M Fura-2 AM at room temperature for 45 min in growth medium. Recordings were performed in HBSS containing the following: 140 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 0.3 mM Na 2 HPO 3 , 0.4 mM KH 2 PO 4 , 4 mM NaHCO 3 , 5 mM glucose, and 10 mM HEPES adjusted to pH 7.4 with NaOH. The cells were then washed three times in HBSS. The fluorescence intensity of Fura-2 in each cell was monitored and recorded at 340 and 380 nm. To represent the variation in the intracellular free calcium concentration, the fluorescence intensity ratio represented by F340/F380 was used as an indicator of the changes in cytosolic Ca 2ϩ concentrations.
Isolation of Human Immune Cell Subsets for the Migration Assay-Human blood samples were collected from healthy adult donors with informed consent obtained in accordance with Institutional Review Board approval from the Institut de Biologie de Lille. Mononuclear cells (PBMCs) were isolated from peripheral blood samples by density gradient centrifugation using Ficoll. Human NK, CD8ϩ and CD4 ϩ CD25 Ϫ conventional T cells were purified from PBMCs by magnetic separation using NK and T cell isolation kits as described by the manufacturer (Miltenyi Biotech, Bergisch Gladbach, Germany). The purity of the isolated NK, CD8 ϩ , and CD4 ϩ T cells was Ͼ95%. NK cells were activated using recombinant IL-2 (200 UI/ml, BD Pharmingen, San Jose, CA) and IL-15 (200 UI/ml, BD Pharmingen). CD8 ϩ and CD4 ϩ T cells were activated with plate-bound anti-CD3 (1.5 g/ml, Clinisciences, Nanterre, France) antibody and incubated at 37°C for 2 h prior to culturing. Soluble mouse anti-human CD28 antibody (100 ng/ml, Clinisciences) was added to the cells at the time of culture.
Migration Assay-The different NK and T cell populations were harvested and suspended at a concentration of 10 6 cells/ml in RPMI 1640. The chemotaxis protocol was performed as previously described (32) using a 48-well microchemotaxis Boyden chamber with 5-m pore polycarbonate filters. The cells were incubated for 2 h 30 min at 37°C in 5% CO 2 with NR8383 secretomes (NT and PC1/3-K D nonstimulated or stimulated with 200 ng/ml LPS for 24 h). Each condition was performed in triplicate. Cells that migrated through the filter were counted in the inferior well. The results are expressed as the number of activated cells that were attracted by each secretome compared with the nonactivated ones.
Cell Viability Measured by the CellTiter-Glo Assay-SKBR3 and SKOV3 cells were seeded into 96-well white plates (6500 and 3000 cells per well, respectively) with NR8383 secretomes obtained after 24 h of LPS stimulation or no stimulation. The assay was conducted for 24, 48, 72, or 96 h. For the 96 hϩ medium, conditioned secretomes were removed at 72 h and replaced with fresh ones. CellTiter-Glo reagent (Promega) was added to the wells and incubated at room temperature for 10 min protected from light. The luminescence was recorded using a Berthold luminometer Centro LB960. The results are expressed as relative light units (RLUs).

PC1/3-knockdown NR8383 Cells Express an Inflammatory
Profile-A stable PC1/3-K D cell line has been developed by lentiviral delivery of shRNAs (11). To better understand the interaction and regulation of TLR4 by PC1/3, we used these PC1/3-K D NR8383 cells in the present study. A proteomics analysis was performed to identify secreted proteins subsequent to PC1/3 silencing in NR8383 cells, before and after LPS challenge. Shotgun proteomics of secreted proteins (from 1 h to 72 h after stimulation) was performed for nontarget (NT) control cells and K D cells (Fig. 1). More than 1400 proteins were identified in all analyses of the NR8383 secretomes (supplemental Data S1).
Comparison of the identified proteins between resting K D or NT cells and cells challenged with LPS allowed the identification of 28 specific proteins that were directly related to PC1/3 (Fig. 1A). These proteins were only secreted by K D cells, both resting and those stimulated with LPS. The proteins are involved in different functions, e.g. responses to stimuli, RNA processing, endocytosis, regulation of transcription, catabolism, protein binding, regulation of axogenesis and immune responses (Table I).
This study was then further coupled to a kinetic study at 1 h, 16 h, 24 h, 48 h, and 72 h after LPS challenge. Proteins with an abundance that was significantly different among the conditions were determined according to the MaxQuant and Perseus software. As a criterion of significance, we applied an ANOVA significance threshold of p Ͻ 0.05, and heat maps were created. A total of 125 proteins in the cell culture super-natants of the macrophages stimulated with LPS versus nonstimulated macrophages were considered reliable based on the statistical analysis (Fig. 1B, supplemental Data S2).
Two major branches of the heat map separate the 1-h series from the other time points. In the second branch, two branches distribute the control versus the stimulated cells. Each experiment was performed 3 times, and each sample corresponded to the statistical data obtained per condition. Five specific clusters of over-expressed proteins were retrieved. Cluster 1 is specific to the control (16 h-72 h). Clusters 2 and 5 are specific to LPS stimulation (16 h-72 h). Cluster 3 represents proteins found after 1 h in both control and LPS-stimulated cells. Cluster 4 is specific to later time points (48 h and 72 h). For each cluster, proteins with specific functions were characterized. For example, in cluster 2, proteins implicated in macrophage activation and the immune response were specific to LPS stimulation. In cluster 1, the majority of the proteins were implicated in protein synthesis associated with early stimulation. The complete list of proteins in these clusters is provided in Table II.
Among the proteins identified in clusters 4 and 5 (Table II) were the alarmins (e.g. GRP78, HSP84, HSP86, HSP73, calreticulin, capthepsin B, nucleolin, and granulins) (33). Alarmins are produced by immune cells through the endoplasmic reticulum (ER)-Golgi secretion pathway and are involved in trig-gering the adaptive immune response. Analyses of the other clusters revealed the release of chemokines (CXCL10 and CCL3), cytokines (MIF), interferon-inducible protein (IFI30) and growth factor (GDF15). Both chemokines and alarmins have known functions in immune responses. These proteins recruit and activate receptor-expressing cells of the innate immune system, including dendritic cells and CD4ϩ cells, and they also directly or indirectly promote adaptive immune responses (34).
To analyze the chemokine/cytokine pattern between NT and K D PC1/3 NR8383 cells in detail, time course analyses of chemokine and cytokine protein expression were performed without quantification ( Fig. 2A).
Both nonstimulated and challenged K D cells secreted CCL3, CCL6, CCL9, CXCL10, CXCL2, MIF, and IL-1RA. Challenged K D cells also secreted CCL7, CCL2, and CXCL3. Comparisons with NT cells revealed that most chemokines were released by both cell types. However, differences in the timing and nature of the secreted chemokines were evident. CXCL3 was only secreted by K D cells, whereas the inhibitory chemokine TGF-␤ was secreted by NT cells. MIF protein was not secreted by NT cells in response to LPS stimulation. To validate these results and to quantify the chemokine and cytokine levels released by NR8383 cells, cytokine arrays were performed after 24 h of stimulation (Fig. 2B). As shown, NR8383 cells secreted several chemokines, including CCL5, CCL3, CXCL9, CCL20, CXCL10, CXCL1, and CXCL2. We established that without LPS challenge, NR8383 K D cells produced significantly more CCL5, CXCL1, CXCL2, CXCL10, IL-6, and TNF-␣ than NT cells (Fig. 2B). Moreover, following LPS challenge, K D cells additionally produced CXCL9 and CCL20 and concomitantly released the previous cytokines and chemokines CXCL1, CXCL2, CXCL10, IL-1␣, and IL-1␤. These chemokine and cytokine profiles are characteristic of secretion through the unconventional secretory pathway, which involves synthesis in the cytoplasm and release without passing through the ER and Golgi complex (35). PC1/3-K D NR8383 Cells and Modulation of the Intracellular Trafficking Machinery-Members of the IL-1 cytokine family, particularly IL-1␣ and IL-1␤ are key inflammatory cytokines that are released through the unconventional secretory pathway. Unconventional vesicular or organellar pathways have been discussed. To determine whether the secretory pathway is affected by PC1/3 K D , proteomic studies of the cellular contents were undertaken in NT and K D cells under either resting conditions or at 24 h post-LPS stimulation (Fig. 3, supplemental Data S3). For the ANOVA test, the samples were classified between NT and K D . Two clear clusters were highlighted. Cluster number 1 represents over-expressed proteins in K D cells, whereas cluster number 2 groups over-expressed proteins in NT cells (Fig. 3A, Table III and supplemental Data S4). The proteins in each cluster were then analyzed using PANTHER software (http://www.pantherdb. org) to determine the biological functions based on the protein classes. Using this analysis, we demonstrated that over-expressed proteins in K D cells were implicated in cell adhesion, extracellular matrix or cytoskeleton, whereas in NT cells, overexpressed proteins were involved in nucleic acid binding or oxidoreduction functions (Fig. 3B). For a more detailed analysis, we performed a systems biology analysis for network identification of the over-expressed proteins in PC1/3-K D cells (Fig. 3C). Differential pathways were generated using the "direct interaction" algorithm to map the relationships of the identified proteins (supplemental Data S5). We found that among the 85 altered proteins, 32 proteins had direct regulatory relationships, including binding, post-translational modifications and transcriptional regulation. Different biological processes are represented by the over-expressed proteins in PC1/3 K D cells (Fig. 3C). For instance, these proteins are involved in actin organization, inflammatory responses, cytoskeletal assembly, T cell activation and calcium channels. Several of the identified proteins were clustered under certain functional classes, such as cellular remodeling (Cluster 1, supplemental Data S5), immune activation and calcium export (Cluster 2, supplemental Data S5), which is consistent with the Protein binding Reticulon-4 Regulation of axonogenesis results obtained using PANTHER software analysis (Fig. 3B). Further demonstrating the impact of PC1/3 K D on the cytoskeleton organization, we found that molecules co-up-regulated with Anxa6 in K D cells were highly significantly related with the GO terms "actin binding" or "extracellular vesicular exosome" (Fig. 4A). Next, we focused on proteins that are important for the cytoskeleton (Table III). Among these cytoskeletal proteins, STRING analysis revealed that most of them are known to be involved in the ARP2/3 complex and the WAVE complex (WASF2, ACTR3, ARPC1B) (supplemental Data S6). Wiskott-Aldrich syndrome protein (WASP) is known to lead to upstream signals resulting in activation of the ARP2/3 complex, which causes a burst of actin polymerization and the formation of a lamellipodium structure. ARP2/3-complexmediated actin polymerization is crucial for the reorganiza-tion of the actin cytoskeleton at the cell cortex during processes such as cell movement and vesicular trafficking (36). Moreover, macrophage activation promotes changes in macrophage cell elasticity that depend on actin polymerization. Cytoskeletal rearrangements are observed between activated and nonactivated macrophages: activated macrophages are elongated, whereas nonactivated cells are more circular (37).
To validate our hypothesis regarding cytoskeletal rearrangements, we incubated NR8383 macrophages with phalloidin to stain the actin filaments. We clearly demonstrated that resting PC1/3-K D cells were more elongated and expressed a high level of directional actin-filled structures compared with NT cells (Fig. 4Ba and b). In this context, PC1/3-K D cells were polarized, with long actin filopodia that emanated from one side of the cell and actin-associated membrane ruffling on the other side (Fig. 4Bb). Thus, PC1/3-K D cells expressed an activated phenotype, and, in response to LPS challenge, the number of filopodia increased corresponding to a higher level of activation in these cells ( Fig. 4Bc and d). Filopodia in LPS-stimulated NT cells increased as compared with control cells demonstrating an activation state under LPS. These results are consistent with the data obtained using proteomics for cytoskeletal reorganization (Figs. 3 and 4A). It has been shown that the cytoskeleton regulates cell polarity, migration and cytokine secretion via vesicular trafficking. Based on the proteomic data, proteins specific to Golgi vesicle transport were clearly under-expressed in K D cells (Table IV).
Among them, we found the following proteins in the same cluster: STX7, USO1, NSF, and COPG1. These proteins are required for transport from the endoplasmic reticulum to the Golgi stack (38) and catalyze the fusion of transport vesicles within the Golgi cisternae. By contrast, endosome-specific proteins (early, late, recycling) APPL1, VAC14, EHD4, VPS4B, ANXA6, and RAP2B and GGA2 proteins involved in protein trafficking between the trans-Golgi network (TGN) and endosomes were over-expressed in K D cells (Table IV). These

FIG. 2. Chemokines and cytokines secreted by NT and K D cells over the time that were challenged or not with LPS.
A, Mass spectrometry analyses allowed for the identification of chemokines and cytokines over time. For identification, FDR was set at 0.01. Blue indicates proteins that were absent, whereas red shows proteins that were present in the secretome. B, Rat cytokine array assay. Cells were untreated (control) or treated with LPS for 24 h. The NT cell secretomes are shown in blue, and the K D cell secretomes are in orange. The bar diagrams represent the ratio of the spot mean pixel densities/reference point pixel densities. Significant differences were analyzed using Student's t test. *p Ͻ 0.05. proteins are required for the regulation of cell proliferation in response to extracellular signals from an early endosomal compartment and also play a role in the biogenesis of endo-some carrier vesicle (ECV)/multivesicular body (MVB) transport intermediates from early endosomes (38). Moreover, these proteins are involved in the late steps of the endosomal FIG. 3. Whole-cell extract analysis strategy. A, NR8383 cells were stimulated with LPS (200 ng/ml) or not (control) and lysed before FASP and LC-MS/MS analysis. MaxQuant and Perseus software were used for the statistical analysis, and a heat map was generated to show proteins that were significantly different between NT and PC1/3-K D NR8383 macrophages in the cell extracts. Two clusters are highlighted. B, The proteins in each cluster were analyzed using Panther software. The biological functions associated with these proteins are shown. The functions framed in red correspond to those of proteins that were over-expressed in K D cells, whereas the functions framed in black correspond to those of proteins that were over-expressed in NT cells. C, Global pathway analysis of the over-expressed proteins identified in PC1/3-K D cells. Over-expressed proteins in PC1/3-K D cells were involved in global altered molecular pathways. The different colors reflect the degree of expression. Proteins in blue are over-expressed in PC1/3-K D cells, whereas those in orange show no expression differences between NT and K D cells. MVB pathway. MVBs contain intraluminal vesicles that are generated by the invagination and scission from the limiting membrane of the endosomes and are mostly delivered to lysosomes, enabling the degradation of membrane proteins. This phenomenon indicates that PC1/3 K D remodels the endosomal compartment, which is consistent with our data obtained using macrophages isolated from PC1/3-KO mice. Ontogenic enrichment of proteins issued from our previous publish proteomic analyses of peritoneal macrophages (10) from wild type (WT) or PC1/3 KO (KO) mice challenged by LPS (8 h) or not, confirms the over-expression of cytoskeleton proteins as well as GTPase activity (5 Go terms) (Fig. 5A). STRING analysis of proteins with GTPase activity has been performed and the network subsequently analyzed using the visualization tool Cytoscape followed by the Integrated Reactome FI analysis tool which selects subnets with specific ontogenetic functions. The biological interaction network GTPase proteins included 202 proteins. Among the groups of proteins with specific ontogenic functions two groups are overexpressed in KO samples i.e. receptor activity and vesicles trafficking. We focused our attention on the last one (Fig.  5B). The network consists mainly of small G proteins and Rab proteins responsible for their regulation. The center of the network, the protein Agfg1 (Arf-GAP domain and FG repeatscontaining protein (1), has an important role in endocytosis (39) At the periphery of this protein, several Rab show a similar expression profile which is over-expressed in unstimulated KO sub-sample and expressed when stimulated with LPS. These Rab (Rab8, Rab11, Rab14 and Rab27) all have a common function that is exocytosis and secretion (40). This disruption of molecular pathways secretion and the increase of membrane targeted molecules involved in vesicle-associated functions in PC1/3 KO mice (Fig. 5C) correlate with dysregulated cytokine secretion as well as intracellular disruption as we found in NR8383 PC1/3 K D cells. Moreover, these results in both NR8383 PC1/3 K D cells and PC1/3 KO macrophages are also consistent with the data obtained from the kinetic study of secretion showing a lot of proteins involved in exosomes.
The large amount of inflammatory cytokines produced by the activated PC1/3-K D cells is consistent with MVB discharge. Interestingly, the cytoskeleton, MVB discharge (38,41) and Ca 2ϩ signaling are linked. Indeed, increasing intracellular Ca 2ϩ levels stimulate exosome secretion (41). Moreover, the disruption of actin filaments inhibits Ca 2ϩ signaling (42). It has been shown that T cells deficient in WAVE2 and WASP present impaired Ca 2ϩ mobilization (43). Gelsolin is another protein that promotes the assembly of actin filaments and has Ca 2ϩ binding sites. A low concentration of Ca 2ϩ inhibits actin binding, whereas a high Ca 2ϩ concentration exposes actin binding sites (44). We have shown that in PC1/3-K D cells, gelsolin and certain proteins implicated in the WAVE2 and WASP complexes are over-expressed (Table III). Moreover, cytokine secretion is dependent on an increase in intracellular Ca 2ϩ (45). The inhibition of store-operated channels (SOCs) leads to a decrease in TNF-␣ and IL-6 secretion (46). To determine whether calcium homeostasis was impacted in NR8383 cells, we performed calcium imaging experiments (Fig. 6).
For that purpose, we used the calcium probe fura-2AM to evaluate the cytosolic calcium concentration ((Ca 2ϩ )) in sterile conditions or in response to LPS exposure. We clearly demonstrated that LPS (n ϭ 3) rapidly induced a pronounced elevation of (Ca 2ϩ )c in NR8383-K D cells compared with NT cells (Fig. 6A, 6B). Fig. 6A shows the Ca 2ϩ signals from individual cells, which are quantified in Fig. 6B. We next investigated whether other features important for calcium homeostasis were affected in NR8383-K D cells. Fig. 6C shows the quantitative results obtained from original traces of similar Ca 2ϩ imaging experiments. Resting (Ca 2ϩ )c levels were significantly increased in NR8383-K D cells. To more thoroughly investigate how the (Ca 2ϩ )c in NR8383-K D cells was affected, we examined two key processes that determine the basal (Ca 2ϩ )c in cells. First, we measured store-operated calcium entry (SOCE) mediated by SOCs. SOCs are located in the plasma membrane and are activated by depletion of the internal Ca 2ϩ store in response to stimulation of the surface receptor-coupled signaling pathway (47). Ca 2ϩ entry mediated by SOCE directly influences (Ca 2ϩ )c. New intracellular Ca 2ϩ imaging experiments were conducted using the SERCA pump inhibitor thapsigargin (TG) as a store-depleting agent (48). As expected, the addition of 2 mM extracellular calcium to NR8383 cells that had been pre-incubated for 10 min with TG (1 M) in Ca 2ϩ -free medium to achieve complete ER Ca 2ϩ store depletion resulted in marked and sustained elevation of the (Ca 2ϩ )c because of the activation of SOCE (Fig. 6D). The peak (Ca 2ϩ )c elevation in relation to SOCE increased by ϳ70% in K D NR8383 cells (Fig. 6E). Moreover, an increasing (Ca 2ϩ ) in the bathing solution from 0 mM to 2 mM produced more significant elevations of the basal (Ca 2ϩ )c in K D NR8383 compared with NT cells (Figs. 6F,  6G). Notably, we observed higher resting (Ca 2ϩ )c levels in NR8383-K D cells compared with NT cells at normal 2 mM external Ca 2ϩ concentrations prior to the application of TG (Fig. 6F). This finding suggests that the PC1/3 K D also promoted enhanced basal Ca 2ϩ influx and not only SOCE. The application of TG in the presence of 2 mM extracellular calcium clearly showed that SOCE was not sustained in control cells compared with NR8383-K D cells. Taken together, these results demonstrate that the basal (Ca 2ϩ )c increased in K D NR8383-K D cells in relation to SOCE and the increase in the constitutive Ca 2ϩ influx, which is consistent with the results obtained for the cytoskeletal rearrangements and cytokine release in these cells.
PC1/3-K D NR8383 Cells and the TLR4 Intracellular Signaling Pathway-As shown previously, in response to LPS treatment, PC1/3 and TLR4 colocalize in the endosomal compartment (11). In PC1/3-K D cells, this compartment is disorganized (10,11) (Table IV), and uncontrolled cytokine secretion is observed following LPS stimulation (10). Altogether, these results indicate that PC1/3 or its products may contribute to the regulation of TLR4 signaling. Proteomic data for the cellular extracts allowed us to focus on proteins implicated in immune system responses, particularly TLR signaling (Table V, supplemental Data S3). LPS stimulation of TLR4 results in the activation of MyD88dependent signaling and subsequently NF-B signaling activation. TLR4 can also mediate the activation of MyD88-independent signaling after its internalization leading to the activation of interferon response factor (IRF) 3 (4). It appears that proteins involved in TLR4 signaling were modulated in NR8383-K D cells. Most importantly, the relative abundance of IRF3 decreased, whereas that of NF-B1 (also known as p105) increased (Table V). This finding demonstrated that NF-B was solicited rather than IRF3 and, thus, suggested that the TLR4 MyD88-dependent pathway was activated in K D cells. Detailed analyses of the results reinforced this hypothesis (supplemental Data S3). Indeed, we observed an over-expression of interleukin-1 receptor accessory protein (Il1Rap), a known active factor in TLR4 signaling that functions by mediating IL-1-dependent activation of NF-B in the MyD88-dependent pathway. Moreover, the cytoplasmic inhibitor of TLR4 signaling, NLRX1, was under-expressed in both resting and challenged PC1/3-K D cells (49). Finally, we observed that molecules co-expressed with NF-B1 in LPSstimulated K D cells comprised Stat1 and Stat2, two major pro-inflammatory transcription factors that were not coexpressed with NF-B1 in LPS-stimulated NT cells (Fig. 7).
To determine whether NF-B showed greater activation in PC1/3-K D cells, we conducted Western blot studies of the TLR4 signaling pathway by studying IB-␣ phosphorylation and IRF3 nuclear translocation (Fig. 8).
The fold change represents the ratio of the band intensity between LPS-stimulated and nonstimulated samples at each time point. The kinetics of the phosphorylation of IB-␣ after LPS challenge from 1 h to 6 h clearly showed that the fold change in phosphorylated IB-␣ was higher at 3 h in K D cells compared with NT cells (Fig. 8Aa, aЈ and 8Ba). At 1 h, the phosphorylation of IB-␣ was high in both NT and K D cells. This phosphorylation remained high in K D cells at 3 h, whereas it decreased in NT cells. In K D cells, a decrease in IB-␣ FIG. 5. PC1/3 KO (KO) mice challenged by LPS exhibit disruption of molecular pathways secretion correlates with dysregulated cytokine secretion. A, Ontogenic enrichment of proteins issued from proteomic analyses of peritoneal macrophages from wild type (WT) or PC1/3 KO (KO) mice challenged by LPS (8 h) B, Cytoscape followed by the Integrated Reactome FI analysis tools focus on vesicles trafficking has been realized. The network consists mainly of small G proteins and Rab proteins responsible for their regulation. Color code corresponds to : green over-expressed, red under-expressed of the selected proteins (KO versus WT) and including treatment (center of the circle untreated, outline treated with LPS). C, Macrophages from PC1 KO mice exhibit increased membrane targeted molecules involved in vesicle-associated functions. The number of membrane-targeted molecules, here defined as molecules with a membrane/cytoplasm ratio Ͼ2, was higher in macrophages derived from KO mice as compared with macrophages from WT mice. Genes coding for KO-specific membrane-targeted molecules were significantly enriched in genes annotated with GO terms describing vesicle-associated functions or cellular compartments. These included the GO term "extracellular vesicular exosome" (p ϭ 1.75E-11) in unstimulated macrophages and the GO term "post-Golgi vesicle-mediated transport" (p ϭ 0.009) in LPS-stimulated macrophages.
phosphorylation was observed at 6 h. These features demonstrate that in K D cells, IB-␣ phosphorylation was observed for a longer period, leading to NF-B activation. The TLR4 MyD88-dependent pathway was involved at later stages compared with NT cells. We also focused on the TLR4 MyD88independent pathway by examining the nuclear translocation of IRF3. After 3 h and 6 h of LPS stimulation, a duplication of the band corresponding to the various phosphorylated states of IRF3 was observed (Fig. 8AbЈ). However, no significant difference in IRF3 nuclear translocation was observed between NT and K D cells over time (Fig. 8Bb), suggesting that this pathway was not impacted by PC1/3 K D (Fig. 8Ab, bЈ and  8Bc). These data confirmed that NF-B was mainly activated in PC1/3-K D cells, as expected based on the proteomic analysis of the NR8383 cellular extracts. This phenomenon is also consistent with our results showing the remodeling of the endosomal compartment (10), (11) (Table IV) and perturbations in calcium homeostasis (Fig. 6).

FIG. 7. NFKB1 co-expression networks in PC1/3 K D versus NT cells.
Lists of 100 genes which encoded molecules are the most tightly coregulated with NFKB1 in K D or NT cells under LPS stimulation were established and assessed for gene set enrichment. Shown are subnetworks of genes annotated by the GO term "cellular response to cytokine stimuli" (adjusted p value for enrichment significance ϭ 9.48e-6 in NT cells and 3.4e-7 in K D cells). Note that Arg1, a prototypic M2 molecule, is specific to the NT subnetwork whereas Stat1 and Stat2, major pro-inflammatory transcription factors, are co-expressed with NFKB1 in K D but not NT cells. Chemoattraction and Antitumor Activities of PC1/3-K D Macrophage Secretomes-PC1/3 down-regulation in NR8383 macrophages strongly affected morphology and intracellular signaling pathways, leading to the secretion of large amounts of alarmins and pro-inflammatory chemokines and cytokines. These changes clearly suggest that PC1/3 K D polarizes NR8383 toward an M1-like profile. M1 phenotype is characterized by high antigen presentation, high production of nitric oxide and high production of pro-inflammatory cytokines. They are known to have a killing function. In contract, M2 macrophages have repairing functions and produce high concentration of anti-inflammatory cytokines like IL-10.
CXCL10 (also known as IFN-␥-inducible protein of 10 kDa or IP-10) and CXCL9 (also known as monokine induced by IFN-␥ or MIG) mainly function in T cell chemoattraction. These two chemokines share a common receptor: CXCR3. CXCR3 appears to be preferentially expressed on T lymphocytes (50). CCL20 is another chemokine that has been implicated in T cell chemoattraction, and the expression of its receptor, CCR6, has been detected in T and B lymphocytes (51). To investigate whether PC1/3-K D NR8383 secretomes have chemotactic properties, assays were conducted in naïve (NA) or activated (A) T lymphocytes (CD4ϩ, CD8ϩ) and natural killer (NK) cells (Fig. 9A). Chemotaxis was performed in Boyden chambers using NR8383 secretomes as chemoattractant agents. The secretomes were obtained from FIG. 9. Chemoattraction and antitumor properties of NR8383 secretomes. A, NR8383 PC1/3-K D secretome enhanced CD4ϩ NA cell chemotaxis. Human primary leukocytes (CD4ϩ (a), natural killer (NK) cells (b) and CD8ϩ (c) cells that were activated (A) or nonactivated (NA)) were incubated in a Boyden chamber with NR8383 secretomes obtained at 24 h (control nonstimulated or stimulated with LPS) or F12 medium in the lower compartment. SDF1a and IL15 were used as controls for migration. Migration was conducted for 150 min. The results are presented as the number of cells that migrated. The experience was done in triplicates. B, C, The cell viability of SKBR3 cells (B) and SKOV3 cells (C) was determined using the CellTiter-Glo assay. The cells were incubated with NR8383 secretomes obtained at 24 h after no stimulation (a) or LPS stimulation (b). The assays were conducted for 24 h, 48 h, 72 h and 96 h. At 72 h, conditioned medium was replaced completely with fresh medium (96 hϩ medium). The results were representative of three independent experiments. Significant differences were identified using Student's t test. *p Ͻ 0.05. nonstimulated or LPS-stimulated NT or PC1/3-K D macrophages after 24 h.
There was a significant attraction of CD4ϩ NA cells in the presence of the stimulated NR8383 PC1/3-K D secretome (K D LPS). We noticed a 6-fold increase in the attraction of CD4ϩ NA between NT LPS and K D LPS secretomes (Fig. 9Aa). No significant differences in chemoattraction were observed between NK and CD8ϩ cells (Figs. 9Ab and c). The data indicated that immune factors produced by K D cells following LPS challenge attracted naïve T helper lymphocytes (Th0; Fig.  9Aa). Moreover, the nature of the chemokines and cytokines produced by the PC1/3-K D cells could polarized these cells from a Th0 to a Th1 profile (TNF-␣, IL-1␣ and IL-1␤).
Next, we assessed the cytotoxic activities of NR8383 secretomes in the SKBR3 breast cancer cell line (Fig. 9B). The relative luminescence is proportional to cell viability and represents a measure of the level of ATP. Treatment with the supernatants of nonstimulated macrophages (control) did not affect the number of living cells (Fig. 9Ba). Treatment with the supernatants of LPS-stimulated macrophages led to a decrease in the number of living cells at 72 h and 96 h (Fig. 9Bb). We then conducted these tests in an ovarian cancer cell line (SKOV3; Fig. 9C). The results confirmed that the secreted factors from PC1/3-K D macrophages exerted antitumor activities, but in different ways, depending on the cell line considered. In SKOV3 cells, the effects were registered at 96 h after refreshing the conditioned medium with activated secreted factors from LPS-challenged PC1/3-K D macrophages (Fig.  9Cb). Under these conditions, NT and K D cells in sterile conditions affected the viability of the SKOV3 tumor cells. A 1.7-fold decrease in luminescence intensity was observed when the cells were exposed to nonstimulated NT and K D cell secretomes (Fig. 9Ca). In LPS-challenged conditions, this decrease was 2.2 fold for the K D secretome and 1.4-fold for the NT secretome.
These results suggested that these two cell lines did not have the same sensitivity toward factors secreted by PC1/ 3-K D macrophages. SKBR3 cells were more sensitive to the direct actions of the secreted factors. SKOV3 cells were sensitized by these factors, and renewal of the conditioned medium with fresh medium containing the same secreted proteins affected their proliferation. This difference in response to treatment between SKBR3 and SKOV3 cells was noted re-cently (52). One major difference between these two cell lines is the production of inhibitory cytokines, such as IL-10: SKBR3 cells produce IL-10, whereas SKOV3 cells do not. Nevertheless, in both cases, the factors secreted by PC1/ 3-K D NR8383 cells were active toward the tumor cells. In those conditions, we assessed whether PC1/3-K D cells were susceptible to IL-10 in terms of their secretion of pro-inflammatory cytokines. As observed in Fig. 10, in response to Il-10 treatment, NT and K D cells secreted decreased amounts of the pro-inflammatory chemokines CXCL1, CXCL2, TNF-␣, and IL1-10. However, K D cells were more resistant to this inhibition and could be more resistant to the tumor inhibitory medium. After IL-10 inhibition, cells were then challenged with LPS. In that case chemokine secretion was restored in both NT and K D cells. Of note, a significant increase in chemokine release was again observed in K D cells, especially for CXCL1, CXCL2, CXCL10, IL1-␣, and IL-6. These results clearly demonstrate that PC1/3 K D cells still conserve their M1-like phenotype under inhibitory conditions and that the immune profile is pro-inflammatory. DISCUSSION The data from our current work demonstrate that PC1/3 is a key enzyme involved in the regulation of cytokine secretion and, consequently, is important for the regulation of macrophage activation, as depicted in Fig. 11. When PC1/3 was inhibited, immune factors such as proinflammatory chemokines as well as alarmins (GRP78, HSP84, HSP86, HSP73, calreticulin, capthepsin B, nucleolin, granulins) were secreted in the absence of challenge. These alarmins may act as autocrine and paracrine factors and activate an immune response through TLR4 (53). Immune factors, such as IL-6 and TNF-␣, are known to be produced by immune cells through the canonical ER-Golgi secretion pathway. However, in NR8383-K D macrophages, numerous unconventional vesicular or organellar pathways have been found. Indeed, proteomic analyses of cell extracts have revealed that the canonical secretion pathway is deregulated and the noncanonical secretion machinery, with an accumulation of endosomes and MVBs, is reinforced. A large number of studies have shown that many proteins known to regulate endocytosis also participate in nuclear signaling. Among them, an adaptor protein containing a pleckstrin homology domain, a phosphotyrosine binding domain and leucine zipper motif 1 (APPL1) has a role in the positive regulation of NF-B signaling. The overexpression of APPL1 triggers p65 translocation to the nucleus in the absence of stimulation (54). Annexin A6 has been shown to increase NF-B activity in response to activation signals (55). These two proteins are over-expressed in PC1/3-K D macrophages and may have an impact on TLR4 signaling (Table IV). Co-expression networks analysis showed that two major pro-inflammatory transcription factors, Stat1 and Stat2, are co-expressed with NF-B1 in PC1/3-K D cells whereas Arg1, a prototypic M2 molecule, is specific to the NT subnetwork (Fig. 7). Proteomics and Western blot analysis revealed that in PC1/3-K D cells, the MyD88dependent pathway was sustained. As a result, a greater activation of NF-B was observed. This transcription factor is known to regulate the expression of the pore-forming Ca 2ϩ channel unit, Orai1, and its activator, STIM1, to control Ca 2ϩ entry and affect cellular functions (56). The induction of NF-B by ER stress, i.e. Ca 2ϩ efflux from the ER, has been also observed (57), and this induction is inhibited by Ca 2ϩ chelators (58). Moreover, the treatment of cells with thapsigargin, which causes an efflux of Ca 2ϩ from the ER, leads to NF-B translocation (58). In the present study, we showed that, in PC1/3-K D cells, both basal Ca 2ϩ and constitutive Ca 2ϩ influx increased. This result is consistent with our findings about NF-B activation (Figs. 7 and 8). Calcium is also known to interact with cytoskeletal proteins. In PC1/3-K D cells, cytoskeletal proteins were over-expressed, leading to the formation of a large number of filopodia (Figs. 3 and 4). The cell elasticity determines macrophage functions and is known to regulate phagocytosis or LPS responsiveness. In a positive feedback loop, the cytoskeleton can modulate the intracellular trafficking machinery through microtubules and calcium mobilization (59,60). This event may amplify such a phenomenon and explain the spontaneous release of pro-inflammatory cytokines. In response to LPS challenge, amplified effects were registered in these PC1/3-K D cells, provoking increased Ca 2ϩ entry and thus impacting the morphology of the cells by producing a greater number of filopodia. As a result, these cells secreted more chemokines and cytokines to recruit naïve T helper lymphocytes (CD4ϩ) and to orient the immune response from Th0 to Th1. Altogether, these findings show that PC1/3-K D macrophages exhibit an M1-like phenotype.
It is known that, during the tumor development process, M1-polarized macrophages switch to an M2-like phenotype that is characterized by IL-12 low IL-10 high and lose their tumoricidal activities (61). According to our results, PC1/3 inhibition could be a relevant strategy to reverse the macrophage phenotype from an M2-like to an M1-like phenotype. As a first step, we tested whether the inhibitory cytokine IL-10 could FIG. 10. Effect of PC1/3 down-regulation on cytokine secretion under inhibitory conditions. The rat cytokine array assay was performed in NR8383 secretomes (NT/K D ). The cells were untreated (control) or treated with IL-10 for 24 h without stimulation (IL-10) or with LPS for 24 h (IL-10ϩLPS). Blue shows NT cell secretomes, and orange indicates K D cell secretomes. The bar diagrams represent the ratio of the spot mean pixel densities/reference point pixel densities. Significant differences were analyzed using Student's t test. *p Ͻ 0.05. affect the secretion properties of PC1/3-K D macrophages. Despite IL-10 treatment, these cells still spontaneously released pro-inflammatory cytokines and oriented the immune response toward a cytotoxic one. Thus, these cells clearly remained highly active. Moreover, in response to LPS challenge, their level of reactivity was restored to the same level of significance as that in cells challenged with LPS without IL-10. We also confirmed the antitumoral properties of their secretomes toward two different cell lines, i.e. one that is known to produce large amounts of IL-10 (SKBR3), whereas the other does not (SKOV3). Taken together, we demonstrated that PC1/3 K D affected cell viability and resistance of cancer cells by the more abundant release of antitumoricidal factors such as TNF-␣ (62). PC1/3 is thus a promising target to reactivate dormant immune cells in tumors and in immunotherapeutic strategies.