Article pubs.acs.org/ac
Combined Proteome and Eicosanoid Profiling Approach for Revealing Implications of Human Fibroblasts in Chronic Inflammation Ammar Tahir, Andrea Bileck, Besnik Muqaku, Laura Niederstaetter, Dominique Kreutz, Rupert L. Mayer, Denise Wolrab, Samuel M. Meier, Astrid Slany, and Christopher Gerner* Department of Analytical Chemistry, Faculty of Chemistry, University of Vienna, 1090 Vienna, Austria S Supporting Information *
ABSTRACT: During inflammation, proteins and lipids act in a concerted fashion, calling for combined analyses. Fibroblasts are powerful mediators of chronic inflammation. However, little is known about eicosanoid formation by human fibroblasts. The aim of this study was to analyze the formation of the most relevant inflammation mediators including proteins and lipids in human fibroblasts upon inflammatory stimulation and subsequent treatment with dexamethasone, a powerful antiphlogistic drug. Label-free quantification was applied for proteome profiling, while an in-house established datadependent analysis method based on high-resolution mass spectrometry was applied for eicosadomics. Furthermore, a set of 188 metabolites was determined by targeted analysis. The secretion of 40 proteins including cytokines, proteases, and other inflammation agonists as well as 14 proinflammatory and nine anti-inflammatory eicosanoids was found significantly induced, while several acylcarnithins and sphingomyelins were found significantly downregulated upon inflammatory stimulation. Treatment with dexamethasone downregulated most cytokines and proteases, abrogated the formation of pro- but also anti-inflammatory eicosanoids, and restored normal levels of acylcarnithins but not of sphingomyelins. In addition, the chemokines CXCL1, CXCL5, CXCL6, and complement C3, known to contribute to chronic inflammation, were not counter-regulated by dexamethasone. Similar findings were obtained with human mesenchymal stem cells, and results were confirmed by targeted analysis with multiple reaction monitoring. Comparative proteome profiling regarding other cells demonstrated cell-type-specific synthesis of, among others, eicosanoid-forming enzymes as well as relevant transcription factors, allowing us to better understand cell-type-specific regulation of inflammation mediators and shedding new light on the role of fibroblasts in chronic inflammation.
T
to sustained treatments with these drugs and still represents a great challenge in clinical practice.10,11 Fibroblasts are important players of inflammation, displaying strong chemokine secretion activities.12,13 Moreover, fibroblasts have been identified as important players in chronic inflammation over the past few years.14,15 One reason therefore may be that, after they have been activated upon inflammation, fibroblasts hardly undergo cell death as typically observed in case of leukocytes16 so that inflammation-associated activities may be maintained for long time periods in the body. Furthermore, these cells may not be affected by classical antiphlogistic drugs in a same way as leukocytes. For an improved understanding of the molecular processes occurring in these cells, a reliable determination of inflammation mediators, which are eicosanoids and cytokines, is mandatory.
hroughout decades of intense research, eicosanoids have been proven to play important roles in physiological and pathophysiological processes related to inflammation, such as diabetes, atherosclerosis, and cancers.1−3 Neutrophils, macrophages, and platelets are known as major producers of eicosanoids that regulate important processes such as antibody and cytokine release, as well as cell proliferation, differentiation, and migration. On the basis of the proinflammatory effects of eicosanoids,4,5 most anti-inflammatory drugs are designed to target enzymes responsible for the synthesis of eicosanoids, either by inhibiting the enzymatic functions or by repressing the transcription of the genes encoding for the enzymes.6,7 However, these drugs may also interrupt the formation of anti-inflammatory eicosanoids,8,9 thus potentially interfering with the resolution of inflammation. This is a scenario with some potential relevance in case of chronic inflammation and related diseases. Actually, while classical antiphlogistic drugs have been successfully implemented in the treatment of acute inflammation, chronic inflammation responds in an unsatisfying fashion © 2017 American Chemical Society
Received: November 10, 2016 Accepted: January 13, 2017 Published: January 13, 2017 1945
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry
acid (FA; Sigma-Aldrich) (ACN/MeOH/FA, 49:49:2). Solvents were evaporated by vacuum centrifugation, and lipids were dissolved in 200 μL of reconstitution buffer (ACN/H2O/ FA, 30:70:0.02). LC−MS Method for Eicosanoids. Samples were pipetted into 200 μL glass inserts. All eluents such as mobile phase A (H2O/FA, 100:0.02) and mobile phase B (ACN/MeOH/FA, 90:10:0.02) were degassed prior to their usage. Using an Infinity 1290 ultrahigh-performance liquid chromatograph (UHPLC) (Agilent Technologies Austria GmbH, Vienna, Austria), a 40 min gradient flow method was applied using a Kinetex 2.1 mm × 15 cm, 2.6 μm, C18, 100 Å reversed-phase column (Phenomenex); the flow rate was set to 250 μL/min, and a gradient was applied (0−2 min, 5% mobile phase B; 2− 28 min, 5−85% mobile phase B; 28−35 min, 95% mobile phase B; re-equilibration with 5% mobile phase B). All samples (20 μL injected) were analyzed in technical duplicates. Mass spectrometric detection was performed with a Q Exactive orbitrap mass spectrometer (Thermo Fisher Scientific) using the HESI source to achieve negative ion mode ionization. MS scans were performed with an m/z range from 250 to 750 and a resolution of 35 000 (at m/z = 300). A lock mass was set; drift was ±5 ppm over all experiments. MS/MS scans of the six most abundant ions were achieved through higher-energy collisional dissociation (HCD) fragmentation at 30% normalized collision energy and analyzed in the orbitrap at a resolution of 17 500 (at m/z = 300). Data have been converted into mzXML to be readable independent of vendor-specific software and made freely available to download following the link: https://anchem. univie.ac.at/ueber-uns/internes/geschuetzt/ download9156546/. Eicosanoid Data Interpretation. Raw files generated by the Q Exactive orbitrap were analyzed and assessed manually using Thermo Xcalibur 2.2 Sp1.48 (Qual browser). Candidate ions were extracted and further evaluated using an in-house developed data processing software (programmed by AT). Libraries from Lipid Maps depository were used and implemented as references.24,25 Extraction and Digestion of Proteins from Secretomes, Cytoplasm, and Nuclear Extracts. Secreted proteins were prepared by overnight precipitation with ethanol at −20 °C of three cell supernatants each of control, activated, and dexamethasone-treated NHDF. After centrifugation, proteins were dissolved in sample buffer (7.5 M urea, 1.5 M thiourea, 4% CHAPS, 0.05% SDS, 100 mM dithiothreitol). In case of dexamethasone-treated NHDF, two biological replicates were further processed in order to obtain cytoplasmic and nuclear proteins, proceeding as previously described.17 In short, cells were lysed in isotonic lysis buffer supplemented with protease inhibitors by applying mechanical shear stress. Cytoplasmic and nuclear proteins were extracted and dissolved in sample buffer. Protein concentrations were determined by means of a Bradford assay (Bio-Rad Laboratories, Germany). Thereafter, in-solution digestion of proteins was performed with trypsin (Roche Diagnostics, Germany). LC−MS Method for Proteins. Samples were solubilized in 5 μL of 30% FA containing 10 fmol each of four synthetic standard peptides and diluted with 40 μL of mobile phase A (98% H2O, 2% ACN, and 0.1% FA). An amount of 10 μL was injected into the nano-HPLC system (Dionex Ultimate 3000) loading peptides on a 2 cm × 75 μm C18 Pepmap100 precolumn (Thermo Fisher Scientific) at a flow rate of 10 μL/ min using mobile phase A. Peptides were then eluted to a 50
Referring to a previous study about proteome alterations induced in inflammatory stimulated primary human dermal fibroblasts,12 we here analyzed these cells with respect to their responsiveness to one of the most important anti-inflammatory drugs, dexamethasone. Thereby, focus was put on eicosanoids and chemokines, as well as other lipids, proteins, and metabolites, applying mass spectrometry (MS)-based shotgun and targeted analyses. On the basis of our practical experience with in-depth proteome profiling of inflammatory processes using a Q Exactive orbitrap,12,17,18 we developed a data-dependent profiling strategy for eicosanoids similar to shotgun analysis of peptides. Only a few research groups have yet employed orbitrap mass spectrometry for eicosanoid profiling,19 and to the best of our knowledge, this is the first analytical study screening for eicosanoids released by human fibroblasts. The present study demonstrates that the combined application of eicosadomics, proteomics, and metabolomics uncovers previously unrecognized features of fibroblasts, highlighting their relevance for chronic inflammation and potentially offering new molecular targets for improved therapy.
■
MATERIAL AND METHODS Cell Culture. Normal human dermal fibroblasts (NHDF), kindly provided by Verena Paulitschke, were cultured as previously described.12 Experiments were performed with cells at same passages (20−22) for three biological replicates each of control and treated cells, using 1.5 × 106 cells per 25 cm2 culture flask. Control cells were incubated for 24 h at 37 °C and 5% CO2. In parallel, inflammatory stimulated NHDF were obtained by treating cells with 10 ng/mL of IL-1β (SigmaAldrich, Vienna, Austria) for 24 h, based on experience from previous studies.20−22 One aliquot of stimulated cells was additionally treated with 100 ng/mL of dexamethasone (SigmaAldrich) 1 h after stimulation and cultured like the other for additional 23 h. Thereafter, all cells were cultured for additional 6 h in 3 mL of serum-free medium to obtain the fraction of secreted proteins.23 Similar experiments were performed with NHDF derived from another donor at passage 6−7 (Lonza, France, Amboise) and human mesenchymal stem cells (hMSC, passage 4−5; Lonza). hMSC were cultured in mesenchymal stem cell growth medium (Lonza) supplemented with the associated Bulletkit and 100U/mL penicillin/streptomycin (ATCC/LGC Standards, London, U.K.). All further processing steps and shotgun proteomics were done in the same way as for NHDF. Extraction of Eicosanoids. Supernatants from cell culture (3 mL each) were collected each in an individual 15 mL falcon tube. Concentrations of 100 nM of each of the internal standards (PGE2-d4, 15S-HETE-d8, and PGF2a-d4; from Cayman Europe, Tallinn, Estonia) were added. After protein precipitation, ethanol (EtOH) was evaporated in a SpeedVac at 35 °C for 25 min until the original sample volumebefore protein precipitationwas restored. Samples were then diluted 1:3 with MS grade water and extracted using 30 mg/mL StrataX solid-phase extraction (SPE) columns (Phenomenex, Torrance, CA, U.S.A.). Columns were conditioned and equilibrated with HiPerSolv grade methanol (MeOH; VWR International, Vienna, Austria) and liquid chromatography− mass spectrometry (LC−MS) grade water (Sigma-Aldrich), respectively. After sample loading, the columns were washed with 15% MeOH, and eicosanoids were eluted with 1 mL of acetonitrile (ACN, VWR International), MeOH, and formic 1946
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry
Figure 1. Protein regulation in secretomes of NHDF upon inflammatory activation and treatment with dexamethasone. (A) Differences of LFQ values (ln2) of proteins with corresponding p-values (log10) represented as volcano plots: act vs con, activated vs control cells; act-dex vs act, activated and dexamethasone-treated vs activated cells; area above the black lines, significantly, at least 2-fold regulated proteins with a global FDR < 0.05; yellow, unregulated proteins; red, inflammation-induced proteins responsive to dexamethasone; green, proteins downregulated upon inflammatory stimulation and responsive to dexamethasone; blue, inflammation-induced proteins not responsive to dexamethasone. (B) Profile plots showing LFQ values (ln2) obtained for the proteins highlighted in panel A from six LC−MS/MS measurements.
cm × 75 μm Pepmap100 analytical column (Thermo Fisher Scientific) with a flow rate of 300 nL/min, using a gradient from 8% to 40% mobile phase B (80% ACN, 20% H2O, 0.1% FA) over 95 min for secreted proteins, and over 235 min for cytoplasmic and nuclear proteins. The nano-HPLC system was coupled to a Q Exactive orbitrap with a nanospray ion source (Thermo Fisher Scientific). MS scans were performed in the range from m/z 400 to 1400 at a resolution of 70 000 (at m/z = 200), MS/MS scans at a resolution of 17500 (at m/z = 200), using a top 8 method for secreted proteins and a top 12 method for cytoplasmic and nuclear proteins and applying HCD fragmentation at 30% normalized collision energy. Protein Data Interpretation. Identification of proteins and label-free quantification (LFQ) were performed using the MaxQuant 1.5.2.8 software including the Andromeda search engine and the Perseus statistical analysis package version 1.5.2.3,26,27 searching against the UniProt database for human proteins (version 102014). A minimum of two peptide identifications, at least one of them unique, was required for positive protein identification. For peptides and proteins, a false discovery rate (FDR) of less than 0.01 was applied. Protein regulation was determined by comparing the LFQ values for each individual protein in the different samples using Perseus, normalizing to the same initial protein amount of 20 μg. Changes in protein abundance values (Supporting Information Tables S1−S3) were determined by a two-sided t test, considering proteins as significantly regulated when the
abundance difference was at least 2-fold with p < 0.05. Proteins described in the manuscript had further a global FDR < 0.05 determined by a permutation-based method.28,29 The proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE30 partner repository with the identifier PXD003963, 3965, 3966, 3967, and 3971. Using Perseus, a principle component analysis (PCA) was performed. Additionally, euclidean distance and average linkage clustering was used for hierarchical clustering of z-scored LFQ values. Enrichment analysis was based on gene ontology terms for biological processes. Furthermore, for selected proteins, heat maps representing LFQ values determined in different cells were generated using an R script.31 To this end, raw data of LFQ values determined in NHDF were used and compared to similar data previously obtained from peripheral blood mononuclear cells (PBMCs).12,17 Targeted Proteomics. Targeted analysis using multiple reaction monitoring (MRM) was performed for selected proteins secreted by NHDF, measuring five biological replicates per cell state, each as technical triplicate. An Agilent 6490 triple quadrupole mass spectrometer coupled to a nano-chip-LC Agilent Infinity series HPLC1290 system was used. Method development for MRM was based on shotgun data, as recently described.21 Optimized transitions are listed in Supporting Information Table S4. After injection of 1 μL of sample to the nano-LC system, peptides were separated by applying a 20 min gradient from 8% to 30% solvent B (99.8% ACN; 0.2% FA). 1947
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry
Figure 2. PCA and cluster analysis demonstrate that dexamethasone does not completely restore the state of control cells. (A) PCA of shotgun proteomics data. (B) Clusters of protein groups differentially regulated in untreated and treated NHDF and main biological processes including Benjamini−Hochberg-corrected FDR values.
pass quality control, resulting in 178 metabolites being finally determined. The excluded metabolites were Asp, Gly, Ac-Orn, ADMA, SDMA, PEA, DOPA, histamine, serotonin, and putrescine. Quality controls were analyzed every 20th sample. Amino acids and biogenic amines were quantified by constructing five-point calibration curves and showed correlation coefficients between 0.9764 (Arg) and 0.9997 (Thr).
Data evaluation was done using Skyline software (version 3.1)32 and was based on manual peak selection and exportation of the total peak area for each peptide. Statistical analysis of the data was performed with Excel, normalizing data to the peak areas of four standard peptides, the percentage of sample volume used for digestion, and the number of cells used per sample.21 Targeted Metabolomics. Targeted metabolomics was performed using AbsoluteIDQ p180 kits (Biocrates Life Sciences AG, Innsbruck, Austria). The kit allows the identification and (semi)quantitation of metabolites including acylcarnitines, amino acids, and biogenic amines, the sum of hexoses, sphingolipids, and glycerophospholipids by LC and flow injection analysis (FIA) MRM. Samples were analyzed on an AB SCIEX Trap 4000 mass spectrometer operated with Analyst 1.6.2 (AB SCIEX), using an Agilent 1200 RR HPLC system with a chromatographic column from Biocrates. Cell lysates of control, inflammatory-activated, and dexamethasonetreated inflammatory-activated NHDF, additional blanks, calibration standards, and quality controls were prepared according to the user manual. All amino acids and biogenic amines were derivatized with phenylisothiocyanate. The experiments were validated with the supplied software (MetIDQ, version 5-4-8-DB100-Boron-2607, Biocrates Life Sciences, and Innsbruck, Austria), and 10 metabolites did not
■
RESULTS Regulation of Proteins in Inflammatory-Activated Fibroblasts Treated with Dexamethasone. Inflammatory stimulation of human dermal fibroblasts (NHDF) was achieved by treating the cells for 24 h with the canonical inflammation mediator IL-1β, as applied in previous studies.20−22 A previously described data-dependent shotgun proteomics method12 was applied to cell supernatants of treated and untreated NHDF. A total of 40 secreted proteins were found to be significantly upregulated by the IL-1β treatment (FDR < 0.05), including many inflammation mediators and promoters such as interleukins IL6 and IL8, chemokines CCL2, CXCL1, CXCL2, CXCL3, CXCL5, CXCL6, matrix metalloproteinases MMP1, MMP2, MMP3, and others (Supporting Information Table S5). In the same experimental setup, the effects of the anti-inflammatory drug dexamethasone were assessed. Dexa1948
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry
Figure 3. Targeted analyses of selected proteins and metabolites demonstrate that dexamethasone does not completely restore the state of control cells. (A) MRM analysis of selected proteins isolated from five biological replicates of NHDF, each analyzed in technical triplicates. Con, controls; IL-1β, inflammatory stimulated; Dex, stimulated and dexamethasone-treated cells; TAN, normalized total area under the curve. (B) Targeted metabolomics analysis of acylcarnitins, sphingolipids, and glycerophospholipids.
of z-scored LFQ values (Figure 2B), using again Perseus, was performed. Subjecting the proteins of each of the four main clusters to the DAVID functional annotation tool,33,34 overrepresented biological processes as indicated in Figure 2B became apparent. To verify the results of the quantitative analysis, we applied targeted proteomics analysis using MRM, currently considered as the most accurate independent validation technique available for multiplex data analysis (Figure 3A).35 Inflammation-Related Metabolic Changes. An accompanying metabolomics study was performed using a targeted approach on cell lysates with a commercial kit. Out of 178 metabolites determined in a valid fashion, only six were significantly regulated upon inflammatory activation (Figure 3B). While the downregulation of acylcarnithins was fully compensated by dexamethasone, the downregulation of several phosphatidylcholins and sphingomyelins was not. Thus, several
methasone was applied 1 h after inflammatory stimulation of the cells in order to mimic a probable in vivo situation where an anti-inflammatory drug is applied when inflammatory signaling has already commenced. Dexamethasone proved to be effective by abrogating or significantly downregulating the secretion of many effector molecules such as IL6 and MMP1 (Figure 1, parts A and B). However, several important inflammation mediators such as CXCL1, CXCL6, and complement factor C3 were not downregulated under the given conditions (Figure 1, parts A and B; Supporting Information Table S5). These findings were verified using NHDF of another donor, and similar results were found in case of hMSC which are precursor cells of fibroblasts (Supporting Information Figure S1). Using Perseus, data were further subjected to a PCA, demonstrating that dexamethasone treatment introduced a distinct cell state rather then bringing inflammatory stimulated cells back to normal (Figure 2A). Furthermore, a hierarchical cluster analysis 1949
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry
Figure 4. Levels of eicosanoids demonstrating that dexamethasone successfully downregulated all IL-1β-induced eicosanoids in NHDF. (A) Eicosanoids upregulated in NHDF upon stimulation with IL-1β. (B) Eicosanoids downregulated in NHDF upon stimulation with IL-1β. (C) Eicosanoids not regulated in NHDF upon stimulation with IL-1β. (D) Eicosanoids newly induced in NHDF upon stimulation with IL-1β.
Figure 5. Comprehensive eicosanoid pathway analysis in NHDF. The metabolic pathways covered by the present analysis methods, indicating successfully identified molecules in black and not identified ones in gray.
dexamethasone. Inflammatory stimulated fibroblasts display a complex eicosanoid profile
but not all inflammation-induced changes in the metabolism of fibroblasts were brought back to normal by the action of 1950
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry
Figure 6. Comparison of proteins levels in NHDF and PBMCs, represented as heat maps, demonstrating cell-type- and cell-state-specific protein regulation. (A) Levels of selected enzymes necessary for lipid synthesis. (B) Levels of selected proinflammatory effector proteins and transcription factors.
enzymes involved in eicosanoid formation was found to differ (Figure 6A). This applied to basal levels as well as to the consequences of inflammatory activation (Figure 6A). For example, arachidonate 12-lipoxygenase and arachidonate 5lipoxygenase-activating protein were highly abundant in PBMCs and rather low in NHDF, whereas prostaglandin reductase 1 and prostaglandin G/H synthase 1 showed higher levels in NHDF. Furthermore, several enzymes involved in eicosanoid synthesis also displayed specific proteome patterns. Actually, prostaglandin G/H synthase 2 (COX-2) was strongly upregulated in inflammatory-activated NHDF and PBMCs. Enzymes such as prostaglandin E synthase and fatty acid desaturase 1 were identified in NHDF and undetectable in PBMCs, and strongly upregulated upon inflammatory stimulation of NHDF, thus displaying both cell-type- and cell-statespecific proteome patterns. The previous study had shown that dexamethasone successfully downregulates all proinflammatory mediators in PBMCs.17 The known capability of dexamethasone to suppress the formation of IL-1β37,38 was verified in NHDF as well (Figure 6B). Also other inflammation regulators such as interleukin-1 receptor antagonist protein and pentraxin-related protein PTX3 were found regulated similarly in PBMCs and NHDF. This raised the question whether the sustained secretion of some proinflammatory mediators in fibroblasts (Figures 1 and 3A) could be the consequence of altered transcription factor activities. Altogether, 449 proteins involved in transcriptional regulation according to gene ontology were identified in nuclear extracts of fibroblasts. Several of those were strongly regulated upon inflammatory stimulation such as nuclear factor κB p100 and p105, AP-1, and jun-D (Figure 6B). The regulation of transcription factors was found to strongly differ between PBMCs and fibroblasts. To give an example, junD was found upregulated in fibroblasts but downregulated in PMBCs (Figure 6B). Using the oPOSSUM software (version 3.0),39 NFκB and AP1 were identified as most important transcription factors with respect to overrepresented binding sites in the promoter regions of proteins upregulated in
Besides cytokines and chemokines, eicosanoids are the most powerful regulators of inflammation.36 We have developed a data-dependent eicosanoid shotgun analysis approach. Analytical parameters such as extraction efficiencies, calibration functions, and LC retention time stability for the evaluation of this method are presented in Supporting Information Figure S2. By applying this method to the secretomes derived from treated and untreated NHDF, 46 molecules with exact masses related to eicosanoids were determined. By matching the corresponding MS2 ion fragments to references from Lipid Maps,24,25 25 distinct eicosanoids were identified in addition to the four deuterated internal standards (Supporting Information Figure S3 and Tables S6 and S7). From those, six eicosanoids were rather unaffected by the inflammatory stimulation (Figure 4C), while another nine were significantly upregulated (Figure 4A), three eicosanoid precursor molecules were downregulated (Figure 4B), and seven eicosanoids were only detectable in the stimulated cells (Figure 4D). As demonstrated in Figure 4, dexamethasone successfully counter-regulated all inflammationinduced eicosanoids. As a data-dependent shotgun approach was applied, we were also able to observe 18 yet unidentified molecules, six of which were newly induced upon inflammatory stimulation and downregulated by dexamethasone (Supporting Information Table S6). The identified molecules were classified into pro- and anti-inflammatory mediators and ordered according to their biochemical generation pathway (Figure 5). As observed in case of proteins, fibroblasts also proved to be potent contributors of powerful and highly relevant lipid inflammation mediators. Effects in Fibroblasts in Comparison to Those Observed Previously in Peripheral Blood Mononuclear Cells. Published proteome profiling data of untreated, inflammatory-activated, and inflammatory-activated plus dexamethasone-treated peripheral blood mononuclear cells (PBMCs) supported a data comparison across these different cells. While proteins involved in household metabolic processes such as β-oxidation were identified in both kinds of cells at rather similar abundance levels, relative abundance levels of 1951
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry inflammatory-activated NHDF. While NFκB p100 was downregulated by dexamethasone in NHDF and PBMCs, AP1 was not downregulated in NHDF, unlike PBMCs. Thus, AP1 might represent a candidate regulatory transcription factor contributing to the sustained formation of proinflammatory mediators such as CXCL6 and C3.
■
HETE and 15-HETE, and to be inhibited by treatment with dexamethasone.45 In summary, dexamethasone successfully abrogated the formation of several proinflammatory eicosanoids in inflammatory-activated NHDF. However, fibroblasts were observed to produce also anti-inflammatory eicosanoids, which were also downregulated by dexamethasone. In consequence, application of dexamethasone may result in a lack of eicosanoids involved in the resolution of inflammation11 and may thus contribute to the persistence of inflammation. Interestingly, the synthesis of enzymes contributing to the production of eicosanoids was found to be regulated in a cell-type-specific way in NHDF and PBMCs. These data show that the different kinds of cells promote the synthesis of eicosanoids in a different fashion, suggesting that their contribution to the inflammatory process and response to antiphlogistic drugs may also be different. Together with these investigations in terms of eicosanoids, the present secretome and metabolomics analyses gave rise to a data set of previously unmatched comprehensiveness combining three independent methodologies. This allowed us to describe previously unrecognized alterations induced in fibroblasts upon inflammation and treatment with a prototypic anti-inflammatory drug. Actually, it is usually expected that the suppression of eicosanoid formation representing the most important anti-inflammatory effect of an antiphlogistic drug would subsequently shut off the secretion of proinflammatory cytokines and chemokines in an indirect fashion.7 This was, however, only observed for PBMCs.17 In fibroblasts, several inflammatory mediators such as CXCL1, CXCL6, and complement C3, described to play important roles in the establishment of chronic inflammation,16,53 were not counterregulated by dexamethasone. These observations were similar when using NHDF from another donor or human mesenchymal stem cells (hMSC), close relatives of fibroblasts. Even though hMSC responded to inflammatory stimulation somewhat differently in comparison to NHDF, IL6 and CXCL8 (IL8) were downregulated likewise by dexamethasone in contrast to C3, CXCL1, CXCL5, and CXCL6. Furthermore, the bottom-up shotgun proteomics data were independently reproduced by targeted analyses. Since dexamethasone acts as transcriptional repressor downregulating the expression of NFkB54 and corresponding target genes such as phospholipase A2, COX-2, IL6, or IL8, related effects were also expected in fibroblasts. Indeed, dexamethasone downregulated important inflammation players such as MMP1, MMP3, CXCL2, PAI-2. The biological process most significantly represented by proteins induced by IL-1β and repressed by dexamethasone was “wound healing”, whereas proteins not affected by dexamethasone were most closely related to “inflammatory response” (Figure 2B). This observation pointed to a different transcriptional regulation of these molecules in fibroblasts. Actually, the levels of classical proinflammatory transcription factors such as jun-D or AP1 were found to be regulated in a cell-type-specific way in NHDF and PBMCs. Regarding the downregulation of sphingomyelins in IL-1βtreated fibroblasts, observations were again pointing to fibroblasts as contributors to chronic inflammation. These metabolites were not counter-regulated when the cells were treated with dexamethasone. Actually, enhanced levels of sphingomyelinase, responsible for degrading sphingomyelins, have been associated with chronic inflammation.55 The present data thus suggest that fibroblasts may support chronic
DISCUSSION
Fibroblasts have only recently been characterized as relevant targets for the treatment of chronic inflammation.15,40 The aim of this study was to analyze the regulation of important players in these cells upon inflammatory stimulation and treatment with the antiphlogistic drug dexamethasone. The analysis of eicosanoids posed a technical challenge due to low concentrations within a complex biological matrix. This is the reason why, so far, mainly targeted mass spectrometry methods have been established for the analysis of eicosanoids.41−43 However, besides technical issues such as the risk of false assignments of signals resulting from unintended in-source fragmentation of labile ions, targeted analyses have limitations with regard to the detection of unexpected molecules. Here we have successfully established a working data-dependent shotgun analysis method for eicosanoids, which allowed us to identify not only wellknown inflammation-regulated eicosanoids but also unexpected ones. While several of the eicosanoids such as 9-HODE and 13HODE as well as TXB2, 11-HETE, and PGE2 have already been described in connection with inflammatory activities of fibroblasts,44−46 most of the eicosanoids identified in the present study were not yet described to occur in these cells. This actually applies to all newly induced eicosanoids presently analyzed (Figure 4D). From those, 8-iso-PGA1 and 11b-PGE2 belong to the isoprostane family and are assigned to COX activity or free radical pathways, thus potentially representing biomarkers for oxidative stress.47,48 Eicosanoids found not or only slightly regulated by inflammation, such as EPET family members or 8-HETE, whose biological functions are not yet fully established, may represent relevant household eicosanoids contributed by fibroblasts. Furthermore, consecutive chains of precursor molecules and their catabolic products were identified. Eicosanoids may undergo rapid catabolism, either enzymatically such as in case of PG 15-OH49 where the end products are the inactive 13,14-dihydro metabolites, or nonenzymatically via oxidation. In our cell model, we observed two degradation events. Thromboxane 2 is a stable but biologically inactive product of the nonenzymatic hydrolysis of the active short-lived compound thromboxane A1. This molecule was found 8 times upregulated upon stimulation with IL-1ß. 6-keto-PGF1a is the result of nonenzymatic hydrolysis of the major inflammation mediator hormone prostacyclin PGI2 which has been described previously to be released from inflammatory-activated fibroblasts.50 This metabolite is physiologically active and is involved in regulating many biological processes related to inflammation.51,52 Furthermore, we found that tetranor-12R-HETE was synthesized in the activated NHDF. This molecule is the result of ß-oxidation of the active compound 12R-HETE, which is produced by 12-LOXR. On the other hand we have identified 12S-HHTrE, which is a metabolite of the thromboxane synthase. Both compounds, which were newly produced and upregulated are not yet functionally characterized. The precursor, 12-HETE, is known to be released from inflammatory fibroblasts, together with 111952
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry Author Contributions
inflammation, even in a state where the synthesis of eicosanoids was successfully abrogated by a potent anti-inflammatory drug. This finding may have an important impact for an improved understanding of chronic inflammation. It is clinically wellknown that withdrawal of dexamethasone may cause worsening of inflammatory symptoms, called rebound effect. While dexamethasone-induced upregulation of cytokine receptors was considered responsible for the rebound effect,56 here we offer an additional, fully compatible but previously unrecognized, mechanism. Leukocytes typically die after inflammation̈ cells are constantly induced cell activation. New and naive formed in the bone marrow. Fibroblasts are long-lived cells which do not die due to exhaustion.16 Considering a potential in vivo scenario where leukocytes are effectively controlled by dexamethasone, explaining the known clinical success of this drug, the continuous release of inflammatory chemokines by fibroblasts could keep the blood levels of these molecules relatively high. Upon attenuation of the immediate effect of dexamethasone, new blood leukocytes could easily become inflammatory stimulated by both high cytokine and cytokine receptor levels, thus also contributing to the feared immune rebound. The present observation may actually direct future investigations in order to improve clinical effectiveness against chronic inflammation and the immune rebound. A complete knock-down of eicosanoids would maybe not help out of this situation; a targeted suppression of cytokine release may be considerable. When investigating transcription factors responsible for the induction of the observed cytokine secretion with the software oPOSSUM,39 AP1 was listed on top as observed previously.12 As AP1 was not counter-regulated by dexamethasone, it may represent an important new target molecule. In this way, the multidimensional molecular profiling data revealed novel functional aspects and may trigger future research on therapeutic strategies targeting those proinflammatory mediators whose secretion was not abrogated during application of classical drugs.
■
C.G. planned the study, A.T. established the method for the analysis of eicosanoids, A.T. and L.N. made the eicosadomics experiments, A.B. and D.K. did the cell culture experiments, A.B. and R.L.M. did the proteome profiling experiments, B.M. developed and performed the targeted proteomics methods, S.M.M. and D.W. did the targeted metabolomics experiments, and A.T., A.S., and C.G. did the data interpretation and wrote the main manuscript text. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank Johanna M. Mader and Peter Frühauf for helpful discussion and practical support.
■
(1) Kortz, L.; Dorow, J.; Becker, S.; Thiery, J.; Ceglarek, U. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2013, 927, 209−213. (2) Greene, E. R.; Huang, S.; Serhan, C. N.; Panigrahy, D. Prostaglandins Other Lipid Mediators 2011, 96, 27−36. (3) Harizi, H.; Corcuff, J.-B.; Gualde, N. Trends Mol. Med. 2008, 14, 461−469. (4) Tam, V. C. Semin. Immunol. 2013, 25, 240−248. (5) Balderramas, H. A.; Penitenti, M.; Rodrigues, D. R.; Bachiega, T. F.; Fernandes, R. K.; Ikoma, M. R. V.; Dias-Melicio, L. A.; Oliveira, S. L.; Soares, Â . M. V. C. Cytokine 2014, 67, 36−43. (6) Huwiler, A.; Pfeilschifter, J. Pharmacol. Ther. 2009, 124, 96−112. (7) Yazid, S.; Norling, L. V.; Flower, R. J. Prostaglandins Other Lipid Mediators 2012, 98, 94−100. (8) Gilroy, D. W. Int. J. Biochem. Cell Biol. 2010, 42, 524−528. (9) Serhan, C. N.; Yang, R.; Martinod, K.; Kasuga, K.; Pillai, P. S.; Porter, T. F.; Oh, S. F.; Spite, M. J. Exp. Med. 2009, 206, 15−23. (10) Czapski, G. A.; Czubowicz, K.; Strosznajder, J. B.; Strosznajder, R. P. Neurochem. Res. 2016, 41, 243−257. (11) Dennis, E. A.; Norris, P. C. Nat. Rev. Immunol. 2015, 15, 511− 523. (12) Slany, A.; Bileck, A.; Kreutz, D.; Mayer, R. L.; Muqaku, B.; Gerner, C. Mol. Cell. Proteomics 2016, 15, 1982−1997. (13) Smith, R. S.; Smith, T. J.; Blieden, T. M.; Phipps, R. P. Am. J. Pathol. 1997, 151, 317−322. (14) Ortega-Gomez, A.; Perretti, M.; Soehnlein, O. EMBO Mol. Med. 2013, 5, 661−674. (15) Flavell, S. J.; Hou, T. Z.; Lax, S.; Filer, A. D.; Salmon, M.; Buckley, C. D. Br. J. Pharmacol. 2008, 153, S241−S246. (16) Naylor, A. J.; Filer, A.; Buckley, C. D. Clin. Exp. Immunol. 2013, 171, 30−35. (17) Bileck, A.; Kreutz, D.; Muqaku, B.; Slany, A.; Gerner, C. J. Proteome Res. 2014, 13, 5989−6000. (18) Meier, S. M.; Muqaku, B.; Ullmann, R.; Bileck, A.; Kreutz, D.; Mader, J. C.; Knasmuller, S.; Gerner, C. PLoS One 2015, 10, e0140367. (19) Masoodi, M.; Eiden, M.; Koulman, A.; Spaner, D.; Volmer, D. A. Anal. Chem. 2010, 82, 8176−8185. (20) Groessl, M.; Slany, A.; Bileck, A.; Gloessmann, K.; Kreutz, D.; Jaeger, W.; Pfeiler, G.; Gerner, C. J. Proteome Res. 2014, 13, 4773− 4782. (21) Muqaku, B.; Slany, A.; Bileck, A.; Kreutz, D.; Gerner, C. Anal. Bioanal. Chem. 2015, 407, 6525−6536. (22) Slany, A.; Meshcheryakova, A.; Beer, A.; Ankersmit, H. J.; Paulitschke, V.; Gerner, C. Clin Proteomics 2014, 11, 41. (23) Gundacker, N. C.; Haudek, V. J.; Wimmer, H.; Slany, A.; Griss, J.; Bochkov, V.; Zielinski, C.; Wagner, O.; Stockl, J.; Gerner, C. J. Proteome Res. 2009, 8, 2799−2811. (24) Fahy, E.; Sud, M.; Cotter, D.; Subramaniam, S. Nucleic Acids Res. 2007, 35, W606−612.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b04433. Proteome analysis data from an independent donor and hMSC, eicosanoid shotgun analysis method evaluation data, MS2 spectra of identified eicosanoids, list of secreted, cytoplasmic, and nuclear proteins, list of transitions, regulation of secreted proteins determined by MRM, chromatographic features with exact masses related to eicosanoids, and raw data for calculating eicosanoid regulation values (PDF)
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: +43-1-427752302. ORCID
Samuel M. Meier: 0000-0002-8930-4574 Astrid Slany: 0000-0002-2217-5800 Christopher Gerner: 0000-0003-4964-0642 1953
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
Article
Analytical Chemistry (25) Sud, M.; Fahy, E.; Cotter, D.; Brown, A.; Dennis, E. A.; Glass, C. K.; Merrill, A. H., Jr.; Murphy, R. C.; Raetz, C. R.; Russell, D. W.; Subramaniam, S. Nucleic Acids Res. 2007, 35, D527−532. (26) Cox, J.; Mann, M. Nat. Biotechnol. 2008, 26, 1367−1372. (27) Cox, J.; Mann, M. BMC Bioinf. 2012, 13, S12. (28) Cox, J.; Hein, M. Y.; Luber, C. A.; Paron, I.; Nagaraj, N.; Mann, M. Mol. Cell. Proteomics 2014, 13, 2513−2526. (29) Tusher, V. G.; Tibshirani, R.; Chu, G. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 5116−5121. (30) Vizcaino, J. A.; Csordas, A.; del-Toro, N.; Dianes, J. A.; Griss, J.; Lavidas, I.; Mayer, G.; Perez-Riverol, Y.; Reisinger, F.; Ternent, T.; Xu, Q. W.; Wang, R.; Hermjakob, H. Nucleic Acids Res. 2016, 44, D447− 456. (31) R Development Core Team. R Foundation for Statistical Computing; R Foundation: Vienna, Austria, 2010. (32) MacLean, B.; Tomazela, D. M.; Shulman, N.; Chambers, M.; Finney, G. L.; Frewen, B.; Kern, R.; Tabb, D. L.; Liebler, D. C.; MacCoss, M. J. Bioinformatics 2010, 26, 966−968. (33) Huang, D. W.; Sherman, B. T.; Lempicki, R. A. Nat. Protoc. 2008, 4, 44−57. (34) Huang, D. W.; Sherman, B. T.; Lempicki, R. A. Nucleic Acids Res. 2009, 37, 1−13. (35) Aebersold, R.; Burlingame, A. L.; Bradshaw, R. A. Mol. Cell. Proteomics 2013, 12, 2381−2382. (36) Wymann, M. P.; Schneiter, R. Nat. Rev. Mol. Cell Biol. 2008, 9, 162−176. (37) Bleeker, M. W.; Netea, M. G.; Kullberg, B. J.; Van der VenJongekrijg, J.; Van der Meer, J. W. Immunology 1997, 91, 548−552. (38) van Furth, A. M.; Seijmonsbergen, E. M.; Langermans, J. A.; van der Meide, P. H.; van Furth, R. Clin. Diagn. Lab. Immunol. 1995, 2, 689−692. (39) Ho Sui, S. J.; Mortimer, J. R.; Arenillas, D. J.; Brumm, J.; Walsh, C. J.; Kennedy, B. P.; Wasserman, W. W. Nucleic Acids Res. 2005, 33, 3154−3164. (40) Buckley, C. D. Immunol. Lett. 2011, 138, 12−14. (41) Song, J.; Liu, X.; Wu, J.; Meehan, M. J.; Blevitt, J. M.; Dorrestein, P. C.; Milla, M. E. Anal. Biochem. 2013, 433, 181−188. (42) Sterz, K.; Scherer, G.; Ecker, J. J. Lipid Res. 2012, 53, 1026− 1036. (43) Kortz, L.; Dorow, J.; Ceglarek, U. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2014, 964, 1−11. (44) Godessart, N.; Camacho, M.; Lopez-Belmonte, J.; Anton, R.; Garcia, M.; de Moragas, J. M.; Vila, L. J. Invest. Dermatol. 1996, 107, 726−732. (45) Godessart, N.; Vila, L.; Puig, L.; de Moragas, J. M. J. Invest. Dermatol. 1994, 102, 98−104. (46) Islam, A. B.; Dave, M.; Amin, S.; Jensen, R. V.; Amin, A. R. Genomics, Proteomics Bioinf. 2016, 14, 81−93. (47) Brose, S. A.; Thuen, B. T.; Golovko, M. Y. J. Lipid Res. 2011, 52, 850−859. (48) Morrow, J. D. Curr. Pharm. Des. 2006, 12, 895−902. (49) Tai, H. H.; Ensor, C. M.; Tong, M.; Zhou, H.; Yan, F. Prostaglandins Other Lipid Mediators 2002, 68−69, 483−493. (50) Myers, S. I.; Evans, C. T.; Bartula, L.; Riva, A.; Kalley-Taylor, B. Mol. Cell. Endocrinol. 1995, 115, 29−39. (51) Terada, K.; Hayashi, G.; Hokama, Y. Gynecol. Oncol. 1990, 36, 212−214. (52) Dorris, S. L.; Peebles, R. S., Jr. Mediators Inflammation 2012, 926968, 18. (53) Fujita, T.; Hemmi, S.; Kajiwara, M.; Yabuki, M.; Fuke, Y.; Satomura, A.; Soma, M. Diabetes/Metab. Res. Rev. 2013, 29, 220−226. (54) Liberman, A. C.; Druker, J.; Perone, M. J.; Arzt, E. Cytokine Growth Factor Rev. 2007, 18, 45−56. (55) Dinkla, S.; Wessels, K.; Verdurmen, W. P.; Tomelleri, C.; Cluitmans, J. C.; Fransen, J.; Fuchs, B.; Schiller, J.; Joosten, I.; Brock, R.; Bosman, G. J. Cell Death Dis. 2012, 3, e410. (56) Almawi, W. Y.; Hess, D. A.; Assi, J. W.; Chudzik, D. M.; Rieder, M. J. Cell Transplant. 1999, 8, 637−647.
1954
DOI: 10.1021/acs.analchem.6b04433 Anal. Chem. 2017, 89, 1945−1954
