trans, trans-2,4-decadienal, a lipid peroxidation product, induces inflammatory responses via Hsp90- or 14-3-3ζ-dependent mechanisms

Abstract

Peroxidation of polyunsaturated fatty acids leads to the formation of a large array of lipid-derived electrophiles (LDEs), many of which are important signaling molecules involved in the pathogenesis of human diseases. Previous research has shown that one of such LDEs, trans, trans-2,4-decadienal (tt-DDE), increases inflammation, however, the underlying mechanisms are not well understood. Here we used click chemistry-based proteomics to identify the cellular targets which are required for the pro-inflammatory effects of tt-DDE. We found that treatment with tt-DDE increased cytokine production, JNK phosphorylation, and activation of NF-κB signaling in macrophage cells, and increased severity of dextran sulfate sodium (DSS)-induced colonic inflammation in mice, demonstrating its pro-inflammatory effects in vitro and in vivo. Using click chemistry-based proteomics, we found that tt-DDE directly interacts with Hsp90 and 14-3-3ζ, which are two important proteins involved in inflammation and tumorigenesis. Furthermore, siRNA knockdown of Hsp90 or 14-3-3ζ abolished the pro-inflammatory effects of tt-DDE in macrophage cells. Together, our results support that tt-DDE increases inflammatory responses via Hsp90- and 14-3-3ζ-dependent mechanisms.

Graphical Abstract

1. Introduction

Redox stress is a common feature of many human illnesses, including inflammatory diseases, cardiovascular diseases, and cancer [1]. Enhanced formation of reactive oxygen species (ROS) during redox stress mediates attack of membrane-incorporated polyunsaturated fatty acids (PUFAs), leading to the formation of a large array of lipid-derived electrophiles (LDEs), including 4-hydroxynonenal (4-HNE), malondialdehyde (MDA), and trans, trans-2,4-decadienal (tt-DDE) [2]. Previous studies have shown that the concentrations of these LDEs are increased in mice and human patients with inflammatory diseases [3–5], and many of these compounds can induce inflammatory responses in vitro and in vivo [6, 7]. These results support potential roles of these LDE compounds in the pathogenesis of human diseases.

One of such LDEs is tt-DDE, which is produced from lipid peroxidation of ω−6 PUFAs such as linoleic acid (LA, 18:2) and arachidonic acid (ARA, 20:4) [8, 9]. tt-DDE is widespread in nature and contributes to activated defense of microalgaes [10]. In addition, it is frequently detected in food products in particular fried food, and this compound is also used as a food additive [11, 12]. Feron et al. reported that tt-DDE was detected in 80 food products, with a concentration as high as 500 ppm [13]. Treatment with tt-DDE in human bronchial epithelial cells enhanced cell proliferation, increased redox stress, and induced production of pro-inflammatory cytokines, illustrating its pro-inflammatory effects in vitro [14]. Wang et al. reported treatment with tt-DDE induced lung inflammation in CD-1 mice, suggesting its pro-inflammatory effect in vivo [15]. Together, these results demonstrate the pro-inflammatory effects of tt-DDE in vitro and in vivo. However, the mechanisms by which tt-DDE increases inflammatory responses are not well understood.

To explore the mechanism of the pro-inflammatory effect of tt-DDE, we employed a combination of the “click” probe of tt-DDE and mass spectrometry (MS) proteomics. Alkyne and azide [3+2] cycloaddition “click” chemistry-based proteomics is a powerful tool for identification of cellular targets of LDEs and similar protein-reactive compounds. This strategy involves several steps: (1) to design and synthesize a “click” probe of the LDE, by functionalizing it with an alkyne- or azide- moiety, (2) to treat the cells or animals with the “click” probe, during this period, the protein targets of the LDE are labelled with the alkyne- or azide- moiety, and (3) the labelled protein targets are then enriched by click chemistry and affinity purification, and the identities of the enriched protein targets could be elucidated by LC-MS analysis[16, 17]. In a previous study, we designed and synthesized trans, trans-2,4-decadien-9-ynal (DDY), as a “click” probe mimicking tt-DDE [18]. In this study, we further evaluated the pro-inflammatory effects of tt-DDE in vitro and in vivo, and used DDY to perform click chemistry-based proteomics to identify the potential cellular targets that are required for the pro-inflammatory effects of tt-DDE.

2. Materials and Methods

2.1. DSS-induced colonic inflammation model in C57BL/6 mice

All animal experiments were conducted in accordance with the protocols approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Massachusetts-Amherst. C57BL/6 male mice (age = 6 weeks) were purchased from Charles River (Wilmington, MA) and maintained in a standard animal facility and fed with standard mouse chow (Prolab Isopro RMH 3000 #5P76, LabDiet, St. Louis, MO) at the University of Massachusetts-Amherst. After one week of acclimation, the mice were stimulated with DSS (2% w/v, molecular weight in the range of 36–50 kDa, MP Biomedicals, Santa Ana, CA) in drinking water, as well as intraperitoneal injection with tt-DDE (dose = 2 mg/kg/day, ACROS Organics, Morris, NJ) or vehicle DMSO (volume = 50 μL per mouse, Thermo Fisher Scientific, Waltham, MA). After 7 days of treatment, the mice were sacrificed to collect tissues for analysis, as we described previously [19].

2.2. H&E staining

The formalin-fixed distal colon tissue was embedded in paraffin (Thermo Fisher Scientific), and then sliced to 5-μm sections, dewaxed in serial xylene treatments (Thermo Fisher Scientific), rehydrated through graded ethanol solutions (Pharmco-Aaper, Brookfield, CT), then stained with hematoxylin and eosin (Sigma-Aldrich, St. Louis, MO), and examined under a light microscope. The pathological scores were evaluated by a blinded observer according to the following principles: crypt architecture, degree of inflammatory cells infiltration, muscle thickening, goblet cell depletion and crypt abscess. The pathological score is the sum of each individual score.

2.3. qRT-PCR analysis of gene expression

The colon tissues from the same location of the distal colon were grounded after frozen by liquid nitrogen. Total RNA was isolated from the colon tissues using TRIzol reagent (Ambion, Waltham, MA). The concentration and quality of the extracted RNA was measured using a NanoDrop Spectrophotometer (Thermo Fisher Scientific), then the RNA was reverse transcribed into cDNA using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Waltham, MA) according to manufacturer’s instructions. qRT-PCR was performed using a DNA Engine Opticon system (Bio-Rad Laboratories, Hercules, CA) with Maxima SYBR-green Master Mix (Thermo Fisher Scientific). The sequences of mouse-specific primers (Thermo Fisher Scientific) were: Il-1β (forward) 5’-GCAACTGTTCCTGAACTCAACT-3’ and (reverse) 5’-ATCTTTTGGGGTCCGTCAACT-3’,Tnf-α (forward) 5’-CCCTCACACTCAGATCATCTTCT-3’ and (reverse) 5’- GCTACGACGTGGGCTACAG-3’, Cox-2 (forward) 5’-TTCAACACACTCTATCACTGGC-3’ and (reverse) 5’- AGAAGCGTTTGCGGTACTCAT-3’, Il-6 (forward) 5’-AGGTCGGTGTGAACGGATTTG-3’ and (reverse) 5’-TGTAGACCATGTAGTTGAGGTCA-3’, Ifn-γ (forward) 5’-ATGAACGCTACACACTGCATC-3’ and (reverse) 5’- CCATCCTTTTGCCAGTTCCTC-3’, Tlr-4 (forward) 5’- ATGGCATGGCTTACACCACC-3’ and (reverse) 5’- GAGGCCAATTTTGTCTCCACA-3’. The results of target genes were normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) and expressed to vehicle treated mice using the 2 ^−ΔΔCt method. The primer sequence for Gapdh was (forward) 5’- AGGTCGGTGTGAACGGATTTG-3’ and (reverse) 5’- TGTAGACCATGTAGTTGAGGTCA-3’.

2.4. Cell culture

Mouse macrophage RAW 264.7 cells (ATCC, Manassas, VA) were cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, Gaithersburg, MD) fortified with 10% fetal bovine serum (Gibco), 100 unit/ml penicillin, and 100 mg/ml streptomycin (Gibco) at 37 °C under an atmosphere with 5% CO₂, and treated with tt-DDE, DDY or vehicle (DMSO), then the medium or cell lysates were collected for analysis.

2.5. ELISA analysis of cytokines

The concentration of cytokine IFN-γ in the cell culture medium was measured using a CBA Mouse Inflammation Kit (BD Biosciences, Franklin Lakes, NJ) according to the manufacturer’s instructions.

2.6. Immunoblotting

Proteins were extracted with RIPA lysis buffer (Boston BioProducts, Ashland, MA) containing 50 mM Tris-HCl, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS and 5 mM EDTA with a protease inhibitor cocktail (Boston BioProducts). Protein concentrations were determined using BCA protein assay kit (Thermo Fisher Scientific). The samples with equal amount of protein (20 μg) were resolved on sodium dodecyl sulphatepolyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto an Odyssey nitrocellulose membrane (LI-COR Biosciences, Lincoln, NE). The membrane was blocked in 5% bovine serum albumin (BSA, Thermo Fisher Scientific) buffer at room temperature for 1 h, then incubated with primary antibodies against phospho-SAPK/JNK (Thr183/Tyr185), SAPK/JNK, IκBα antibody (Cell Signaling Technology, Danvers, MA) and β-actin (Sigma-Aldrich) in 5% BSA solution at 4 °C overnight. The membrane was then probed with LI-COR IRDye 800CW goat anti-rabbit and IRDye 680RD goat anti-mouse secondary antibodies and was scanned using the Odyssey imaging system (LI-COR Biosciences). Quantification of immunoblotting was performed using Image Studio Lite Software (LI-COR Biosciences). β-actin was used as the loading control. Data are normalized against those of the corresponding β-actin.

2.7. Click chemistry-based in-gel SDS-PAGE fluorescence imaging

RAW 264.7 cells were treated with vehicle (DMSO), tt-DDE, or DDY for 1–4 h, then harvested with RIPA buffer (Boston BioProducts) containing 1% protease inhibitor cocktail (Boston BioProducts). 50 μL of the cell lysate was incubated with 1 mM CuSO₄ (Thermal Fisher Scientific), 1 mM ascorbic acid (Sigma-Aldrich), 2 mM tris-hydroxypropyltriazolylmethylamine (THPTA) ligand (Click Chemistry Tools, Scottsdale, AZ), and 50 μM CruzFluor sm™ 8 azide (Santa Cruz Biotechnology, Dallas, TX) at room temperature for 1 hour in dark, as described ¹⁷. After the reaction, the solution was mixed with SDS-PAGE loading buffer (Amresco, Solon, OH), resolved by 8% SDS-PAGE, and the gels were directly scanned with excitation wavelength 774 nm and emission wavelength 789 nm using an Odyssey imaging system (LI-COR Biosciences). Then the gels were stained with Coomassie blue (Thermo Fisher Scientific) to quantify total proteins in the gel.

2.8. Click chemistry reaction, biotinylation and streptavidin affinity purification

The following biotinylation protocol was modified from the protocol used in Szychowski et al. [20]. RAW 264.7 cells (a kind gift from Dr. Karen Liby) received one of the following five treatments: 1) vehicle (DMSO), 2) 1 mM tt-DDE, 3) 25 μM DDY (DDY1), 4) 20 μM DDY (DDY2), or 5) 1 mM tt-DDE + 20 μM DDY (tt-DDE +DDY). Following treatment, cells were incubated for 2 h, followed by lyzing the cells with RIPA buffer containing 1% protease inhibitor cocktail. The cell lysate (100 μL, protein concentration = 0.5 mg/mL) was added to phosphate buffer saline (PBS, 760 μL, pH 7.8), followed by addition of SDS (100 μL, 10% w/v stock solution). A solution of PEG3 azide (10 μL, 5 mM, Lumiprobe, Hunt Valley, MD) and a solution of THPTA ligand (10 μL, 50 mM, Lumiprobe) dissolved in DMSO were added to the mixture. Next, a solution of CuSO4 (10 μL, 25 mM, Sigma-Aldrich) in deionized distilled water and an aqueous ascorbic acid solution (10 μL, 1M, pH 7.4, Sigma-Aldrich) were added and stirred vigorously at 30 °C for 1 h. The reaction mixture was then transferred dropwise to a centrifugation tube containing 5 mL acetone and stored at −20 °C for 2 h followed by centrifugation at 2000 g for 10 min at 4 °C. The supernatant was removed, and the pellet was resuspended via sonication in PBS (100 μL, pH 7.8, containing 3% SDS). The suspension was then centrifuged at 2000 g for 10 min at 4 °C. The supernatant was collected, and the labeled proteins in the supernatant were purified using Pierce™ High Capacity Streptavidin Agarose (Thermo Fisher Scientific, catalog number 20361), according to manufacturer’s instructions. Upon completion of the affinity purification, 1 mL of ice-cold 50 mM ammonium bicarbonate (ABC) was added to the resin and stored at 4 °C for trypsin digestion.

2.9. Trypsin Digest and Proteomics Analysis

The trypsin digestion was performed by the proteomics core facility at Michigan State University. Bead-bound proteins were digested by washing 4 times using 50 mM BAC buffer. Trypsin, in the same buffer, was then added to the beads at 5 ng/μL so that the beads were just submerged in digestion buffer and allowed to incubate at 37 °C for 6 h. The solution was acidified to 5% formic acid and centrifuged at 14,000 g. Peptide supernatant was removed and concentrated by solid phase extraction using C18 OMIX tips (Varian, Milpitas, CA) according to manufacturer’s instructions. Purified peptides eluates were dried by vacuum centrifugation and re-suspended in 2% acetonitrile/0.1% TFA to 20 μL.

2.10. LC-MS/MS and data analysis

For LC-MS/MS analysis, 8 μL samples were injected via Thermo EASYnLC 1000 onto a Thermo Acclaim PepMap RSLC 0.075mm x 500mm C18 column and eluted over 35 min with a gradient of 5% B to 25% B gradient in 24 min, at which point, B was ramped to and remained at 90% for another 10 min. The flow rate was held constant at 0.3 μL/min (Buffer A: 0.01% formic acid in water, Buffer B: 0.1% formic acid in acetonitrile). An integrated column heater (PRSOV1) was used to hold the column temperature at 50 °C.

A FlexSpray spray ion source was used to spray the eluted peptides into the mass spectrometer (Thermo Fisher Q-Exactive). An Orbi trap (70,000 resolution, m/z = 200) was used to identify the top ten ions in every survey scan, which were then subjected to automatic higher energy collision induced dissociation (HCD). Fragment spectra were acquired at a resolution of 17,500, and Mascot Distiller (v2.7.0, Matrix science, Boston, MA) converted MS/MS spectra data to peak lists. These lists were then run against all available mouse protein sequences from Uniprot (www.uniprot.org) and appended with common laboratory contaminants (from the cRAP project, www.thegpm.org) via the Mascot searching algorithm (v2.6) on an in-house server. Further analysis of the Mascot output was then performed using Scaffold (v4.8.7), which allowed for the validation of protein identification probabilities. An FDR of 1% in Scaffold was used as a confidence filter for considering true assignments. Other parameters were: MS/MS tolerance of 0.02 Da, peptide tolerance of +/−1ppm, allowance of 1 missed tryptic site, variable modification of oxidation of methionine, and a limited peptide charge state at +2/+3. Maxquant was then used to conduct label free quantification (LFQ) match-between-runs comparisons (parameters included in supplemental information) to determine potential protein leads.

From the Maxquant search (protein threshold at FDR 1.0%, minimum peptide count of 2, and:

1. The normalized value of quantitative total ion count (TIC), total spectrum count (TSC), and percent total spectrum (PTS) of the DDY probe samples is greater than the control sample (DDY1 > the vehicle control and DDY2 > the tt-DDE control sample);

2. TIC, TSC, and PTS values of DDY2 is greater than that of tt-DDE + DDY sample and;

3. High identification probability (>95%) as assigned by Scaffold.

DDY1 was not compared to the competition sample because the two had different concentrations of DDY (25 μM in DDY1 compared to 20 μM in tt-DDE + DDY). Hence, proteins considered were DDY1 > Veh control, DDY2 > tt-DDE control, DDY2 > tt-DDE+DDY for TIC, TSC, PTS, and protein probability > 95%.

2.11. Small Interfering RNA (siRNA) transfection

siRNA duplexes of siHsp90b1 (Gene ID: 22027) and si14-3-3ζ (Gene ID: 22631) oligos were purchased from Silencer™ Select siRNAs (Thermo Fisher Scientific). The siRNA oligos were transfected into cells by using Lipofectamine RNAiMAX transfection reagent (Thermo Fisher Scientific) according to manufacturer’s instructions. The Silencer™ Select Negative Control was used as negative control. The sequences of mouse-specific primers (Thermo Fisher Scientific) to detect the knockdown effect of tt-DDE target genes were: Hsp90b1: (forward) 5’-TCGTCAGAGCTGATGATGAAGT-3’ and (reverse) 5’- GCGTTTAACCCATCCAACTGAAT-3’, 14-3-3ζ: (forward) 5’-GAAAAGTTCTTGATCCCCAATGC-3’ and (reverse) 5’-TGTGACTGGTCCACAATTCCTT-3’.

2.12. Cell viability assay

RAW 264.7 cells were seeded in 96-well plates at a cell density of 5,000 cells per well and were treated with tt-DDE or DMSO vehicle (0.1%) for 24 h, cell viability was measured using 3 −(4,5 −dimethylthiazol −2 −yl) −2,5-diphenyltetrazolium bromide (MTT, Sigma Aldrich, St. Louis, MO).

2.13. Data analysis

All data are expressed as the mean ± standard error of the mean (SEM). For the comparison between two groups (e.g. vehicle versus tt-DDE), Shapiro-Wilk test was used to verify the normality of data; when data were normally distributed, statistical significance was determined using two-side t-test; otherwise, significance was determined by Mann-Whitney U test. Analysis of three groups (e.g. effect of vehicle, tt-DDE, and DDY on expression of IκBα) was performed by one-way ANOVA. Analysis of four groups (e.g. roles of JNK signaling in the pro-inflammatory effect of tt-DDE) was performed by two-way ANOVA, followed by Tukey-Kramer’s method. The statistical analyses were performed using SAS statistical software, and P < 0.05 were considered statistically significant.

3. Results

3.1. Effects of tt-DDE on DSS-induced colonic inflammation in mice

Our recent study showed that treatment with 4-HNE, a tt-DDE-similar LDE compound, exacerbated DSS-induced colonic inflammation in mice [19]. Here we further studied the effect of tt-DDE on inflammation using this model (see scheme of animal experiment in Fig. 1A). We Compared with vehicle control, treatment with tt-DDE (dose = 2 mg/kg/day, determined from our previous study of 4-HNE in the DSS model [19]) reduced colon length (P < 0.05, Fig. 1B), increased expression of pro-inflammatory genes (Il-1β, Ifn-γ, Tlr-4 and Cox-2) in the colon (P < 0.05, Fig. 1C), and exaggerated tissue damage in the colon (P < 0.001, Fig. 1D). Together, these results showed that systemic treatment with tt-DDE increased DSS-induced colonic inflammation, demonstrating its pro-inflammatory effects in vivo.

Fig. 1. Administration of tt-DDE increased DSS-induced colitis in mice.

(A) Scheme of animal experiment. (B) Colon length. (C) RT-PCR analysis of gene expressions in colon. (D) H&E staining of colon (magnification ×300, Scale bars: 50 μm). The data are mean±S.E.M., n=5–6 mice per group.

3.2. Effects of tt-DDE on inflammatory responses in RAW 264.7 cells

To determine the effect of tt-DDE on inflammation in vitro, we tested its action on cytokines production in mouse macrophage RAW 264.7 cells. Treatment with 1 μM tt-DDE for 24 h increased gene expression of pro-inflammatory cytokines (Il-1β, Il-6 and Ifn-γ) (P < 0.05, Fig. 2A) and elevated medium concentration of IFN-γ (P < 0.05, Fig. 2B) in RAW 264.7 cells, while treatment of tt-DDE had little effect on the cell viability (Fig S2), demonstrating the pro-inflammatory effect of tt-DDE in vitro.

Fig. 2. tt-DDE induced inflammation in RAW 264.7 cells.

(A) tt-DDE (1 μM) increased gene expression of pro-inflammatory cytokines (n=3–4 per group). (B) tt-DDE (1 μM) increased concentration of proinflammatory cytokine IFN-γ in cell culture medium (n=4–5 per group). (C) tt-DDE increased JNK phosphorylation and enhanced IκBα degradation (n=3 per group). (D) Chemical inhibition of JNK signaling attenuated pro-inflammatory effects of tt-DDE (n=5–6 per group). The data are mean±S.E.M.

Previous studies showed that NF-κB and JNK signaling pathways play critical roles in inflammation [21, 22]. We found that treatment with tt-DDE induced a rapid phosphorylation of JNK in RAW 264.7 cells: after 15 min treatment, tt-DDE increased expression of phosphorylated JNK in RAW 264.7 cells (Fig. 2C). Such activity was also observed after 30 min treatment with tt-DDE, though the effect was weaker; and this result is in agreement with a previous study about the effect of 4-HNE on JNK phosphorylation [23]. In addition, treatment with tt-DDE induced degradation of IκBα in RAW 264.7 cells, suggesting that tt-DDE activates NF-κB signaling (Fig. 2C), however, this effect required longer time compared with that of JNK phosphorylation.

To test the contributions of the JNK pathway in the pro-inflammatory effects of tt-DDE, we used a chemical inhibitor approach and found that chemical inhibition of JNK signaling abolished the pro-inflammatory effects of tt-DDE (Fig. 2D). Without co-administration of a JNK inhibitor, treatment with tt-DDE increased gene expression of Il-1β, Il-6 and Ifn-γ in RAW 264.7 cells; however, with co-administration of a JNK inhibitor SP600125 (concentration = 100 nM), the pro-inflammatory effects of tt-DDE were abolished. Two-way ANOVA analysis validated that there was a statistically significant interaction (P < 0.05) between JNK signaling pathway and tt-DDE on inflammatory responses in vitro (Fig. 2D). Treatment with the JNK inhibitor alone slightly increased cytokine expression (Fig. 2D), and this is in agreement with previous studies [24, 25]. Together, these results support that the JNK signaling pathway contributes to the pro-inflammatory effect of tt-DDE.

3.3. Effects of DDY, a click chemistry probe mimicking tt-DDE, on inflammation and protein labeling in RAW 264.7 cells

To better understand the mechanisms by which tt-DDE increases inflammation, we used a click chemistry-based proteomics approach to identify its potential cellular targets. We have designed and synthesized a click chemistry probe of tt-DDE termed DDY (see chemical structure in Fig. 3A) and we evaluated the biological activities of this probe. Similar to tt-DDE, treatment with DDY (concentration = 1 μM) increased expression of pro-inflammatory cytokines (Il-1 β, Il-6 and Ifn-γ) and enhanced degradation of IκBα in RAW 264.7 cells (Fig. 3B–C), suggesting these two compounds have similar pro-inflammatory effects. In addition, click chemistry-based fluorescence SDS-PAGE imaging showed that treatment with DDY in RAW 264.7 cells labeled cellular proteins in a dose- and time-dependent manner (Fig. 4). Together, these results support the use of DDY to explore the cellular targets of tt-DDE.

Fig. 3. DDY has the same biological function with tt-DDE.

(A) Chemical structure of tt-DDE and DDY. (B) Both tt-DDE and DDY increased gene expression of pro-inflammatory cytokines in RAW 264.7 cells (n=4–6 per group). (C) Both tt-DDE and DDY increased IκBα degradation in RAW 264.7 cells (n=4 per group). The data are mean±S.E.M.

Fig. 4. DDY labelled cellular proteins in RAW 264.7 cells.

(A) Scheme of the click chemistry-based in-gel fluorescence SDS-PAGE imaging. (B) DDY labeled cellular proteins in a dose-dependent manner in RAW 264.7 cells (2 h treatment). Left panel: fluorescence imaging. Right panel: Coomassie blue staining. (C) DDY labeled cellular proteins in a time-dependent manner in RAW 264.7 cells. Left panel: fluorescence imaging. Right panel: Coomassie blue staining.

3.4. Identification of potential cellular targets of tt-DDE using click chemistry-based proteomics

To identify the potential protein target of tt-DDE, we employed click-based proteomics, using DDY as our click probe of tt-DDE (Please see Fig. S1 and Section 2.8 for scheme of the proteomic analysis). Based on our proteomic analysis (see proteomics result in supplemental information Table S1), 209 proteins/clustered were identified, 36 proteins (Table S1) were determined to be in greater abundance in DDY probe samples than in the relative control and competition sample(s), as measured by TIC, PTS, and TSC. Included were proteins whose primarily roles were metabolism (aldose reductase, fructose-bisphosphate aldolase, glyceraldehyde-3-phosphate dehydrogenase, l-lactate dehydrogenase, malate dehydrogenase (mitochondrial), phosphoglycerate kinase, pyruvate kinase, triosephosphate isomerase), translation and mRNA processing (40S ribosomal proteins S25, S26, S3a, S8, and S9, 60S ribosomal proteins L13, L18, and L7a (MCG11348), elongation factors 1-γ, and 2, receptor of activated protein C kinase 1, and ribosomal proteins L4, S14, L13a (MCG23455), and S16 (MCG123443), heterogeneous nuclear ribonucleoprotein K, and poly(rC)-binding protein 1), structural (type II cytoskeletal 2 oral keratin, Tpm3, tubulin β−5 chain, and uncharacterized protein (Actb)), nucleotide synthesis (inosine-5’-monophosphate dehydrogenase 2 and multifunctional protein ADE2), nuclear transport (GTP-binding nuclear protein Ran), mitochondrial transport (MCG10343), stress response/ regulatory (heat shock protein 90 and 14-3-3 ζ), and others (carbonic anhydrase 2, MCG23377, nucleophosmin, and uncharacterized protein (Cct8)). Unfortunately, we were unable to directly identify any peptides modified by DDY by the MS/MS data, which may be due to their low relative abundances. In addition, we have to point out we used a relatively high dose of DDY (concentration = 20–25 μM) for the proteomics experiment, therefore some of the identified protein targets could be not specific. And while it is possible that there are multiple mechanisms of action by tt-DDE/ DDY which include the interactions/ modifications of many proteins included amongst those identified, we opted to further investigate those proteins which are known to play critical roles in stress response signaling in inflammation. HSP90 and 14-3-3ζ stood out as likely contributors [26–28]. HSP90 was found to have been present at ~2.32 average fold difference in the DDY-treated samples relative to vehicle-treated controls, comprising an average of 0.46% of the total spectrum count compared to 0.12% for the controls. The TIC for 14-3-3ζ in the DDY-treated samples had an average fold difference of 7.13 over that of the vehicle-treated controls, and accounted for 0.19% of the total spectra as compared to 0.025% in the controls.

3.5. Roles of Hsp90 and 14-3-3ζ in the pro-inflammatory effects of tt-DDE

To validate the identified targets, we used siRNA knockdown to determine the contributions of the identified cellular proteins in the pro-inflammatory effects of tt-DDE. Transfection of siRNA targeting Hsp90 reduced expression of the target gene (P < 0.001, Fig.5A). With transfection of negative control siRNA, treatment with tt-DDE increased gene expression of pro-inflammatory cytokines (Il-1 β, Il-6 and Ifn-γ) in RAW 264.7 cells; however, with the transfection of siRNA to knockdown Hsp90, the pro-inflammatory effects of tt-DDE were abolished (Fig. 5B). Two-way ANOVA analysis validated that there was a statistically significant interaction (P < 0.05) between Hsp90 with tt-DDE on inflammation in vitro (Fig. 5B), supporting that Hsp90 contribute to the pro-inflammatory effects of tt-DDE. A similar result was also obtained for 14-3-3ζ (Fig. 5C–D), supporting that a potential role of 14-3-3ζ in the pro-inflammatory effects of tt-DDE. Treatment with tt-DDE had little effect on gene expression of Hsp90 and 14-3-3ζ in cells (Fig. S3).

Fig. 5. siRNA knockdown of Hsp90 or 14–3-3ζ abolished the pro-inflammatory effects of tt-DDE.

(A) siRNA knockdown of Hsp90 reduced gene expression of Hsp90 in RAW 264.7 cells (n=4 per group). (B) siRNA knockdown of Hsp90 abolished the pro-inflammatory effects of tt-DDE (n=5–6 per group). (C) siRNA knockdown of 14–3-3ζ reduced gene expression of 14–3-3ζ in RAW 264.7 cells (n=4 per group). (D) siRNA knockdown of 14–3-3ζ abolished the pro-inflammatory effects of tt-DDE (n=5–6 per group). The results are mean±S.E.M., NC: negative control (the cells were transfected with control siRNA).

4. Discussion

Previous studies have shown that tt-DDE, a lipid peroxidation-derived LDE compound, induces inflammation in vitro and in vivo [14, 15], however, the underlying mechanisms are not well understood. Using a click chemistry-based proteomics approach, the central finding of our research is that tt-DDE increases inflammation via Hsp90- and 14-3-3ζ-dependent mechanisms. This finding will enhance our understanding for the action mechanism of tt-DDE, as well as other lipid peroxidation-derived LDEs.

In agreement with previous studies [14, 15], our results further support that tt-DDE has potent pro-inflammatory effects in vitro and in vivo. We found that treatment with low-dose tt-DDE increased cytokines production, induced JNK phosphorylation, and enhanced activation of NF-κB signaling in RAW 264.7 macrophage cells, demonstrating its pro-inflammatory actions in vitro. JNK signaling plays a critical role in mediating the pro-inflammatory of tt-DDE, since tt-DDE induced a rapid phosphorylation of JNK, and chemical inhibition of the JNK signaling abolished the pro-inflammatory effects of tt-DDE. We also found that treatment with tt-DDE via intraperitoneal injection exaggerated DSS-induced colonic inflammation in C57BL/6 mice, further validating its pro-inflammatory actions in vivo. We did not study the effect of tt-DDE on inflammatory responses in naive mice, while a previous study by Wang et.al demonstrated that treatment with tt-DDE in CD-1 mice induced lung inflammation, suggesting that tt-DDE can cause inflammation in naive animals [15]. In the animal experiment, we did not use oral treatment to administer tt-DDE, since it is highly reactive toward proteins and therefore could be degraded in the gastrointestinal tract [29]. Our recent study showed that dietary administration of frying oil, which contains high concentrations of lipid oxidation-derived LDE compounds, exaggerates colonic inflammation and colon cancer in mice [30]. Based on our finding, it is feasible that the presence of tt-DDE, as well as other LDE compounds, could contribute to this effect. Together, our results further validate that tt-DDE has potent pro-inflammatory effects in vitro and in vivo. Besides tt-DDE, our recent research also showed that systematic treatment with other lipid peroxidation-derived LDEs, such as 4-HNE, also increased inflammation in mouse models [19]. These results highlight the biological significance of these LDEs in human health.

The molecular mechanisms by which tt-DDE increases inflammation were not well understood. Here, using DDY as a “click” chemistry probe of tt-DDE, we find that Hsp90 and 14-3-3ζ play critical roles in the pro-inflammatory actions of tt-DDE. Compared with tt-DDE, DDY showed more potent pro-inflammatory effects in RAW 264.7 cells. There could be multiple mechanisms why DDY has more potent biological actions than tt-DDE. Compared with tt-DDE, DDY has increased polarity, which could make it easier to pass cell membrane to enter intracellular space, and this could in part contribute to its more potent actions. The click chemistry-based proteomics showed that Hsp90 and 14-3-3ζ are two binding proteins of DDY. Furthermore, siRNA knockdown of either gene abolished the pro-inflammatory effects of tt-DDE in macrophage cells. Previous research has shown that Hsp90 and 14-3-3ζ play critical roles in inflammation. Hsp90 is overexpressed in tumor tissues and is implicated in tumorigenesis and cancer resistance [31, 32]. Acting as a chaperone, Hsp90 can stabilize many oncogenes and contribute to activation of NF-κB signaling [33]. HSP90 inhibitors are currently being evaluated as anti-cancer and anti-inflammatory therapeutics [34, 35]. 14-3-3 is a family of highly conserved proteins that are ubiquitously expressed in various mammalian tissues [36]. There are seven isoforms (β, γ, ε, η, σ, τ, and ζ) of 14-3-3 encoded in humans [37]. 14-3-3ζ is overexpressed in LPS-stimulated macrophage cells [38], as well as in hepatocellular carcinoma cells and infiltrating T lymphocytes [39]. And 14-3-3ζ is being evaluated as a potential prognostic marker and therapeutic target for cancer [40]. Together, these results support that Hsp90 and 14-3-3ζ contribute to the pro-inflammatory effects of tt-DDY. Our results are largely in agreement with previous studies, which showed that Hsp90 is a binding protein of 4-HNE, a lipid peroxidation-derived LDE with a similar α, β-unsaturated carbonyl moiety [41]. Further studies are needed to characterize the functional roles of Hsp90 and 14-3-3ζ in the pro-inflammatory effects of tt-DDE. Notably, it remains to determine whether tt-DDE directly modifies Hsp90 and 14-3-3ζ to modulate their activities or properties, resulting in enhanced inflammation, or whether Hsp90 and 14-3-3ζ are involved in down-stream signaling for the pro-inflammatory effects of tt-DDE.

In summary, here our results showed that tt-DDE potently increases inflammation in macrophage cells and in mouse models. Click chemistry-based proteomics supports that Hsp90 and 14-3-3ζ play critical roles in mediating the pro-inflammatory effects of tt-DDE. These results could broadly enhance our understanding for the biological effects and mechanisms of lipid peroxidation-derived LDEs.

Table 1. Top protein candidates.

The Protein candidates are listed in alphabetical order and color coded according to their primary functions.

Identified Protein (35)	Gene
14-3-3 protein zeta/delta	Ywhaz
40S ribosomal protein S25	Rps25
40S ribosomal protein S26	Rps26
40S ribosomal protein S3a	Rps3a1
40S ribosomal protein S8	Rps8
Aldose reductase	Akr1b1
Carbonic anhydrase 2	Ca2
Elongation factor 1-gamma	Eef1g
Elongation factor 2	Eef2
Fructose-bisphosphate aldolase	Aldoa
Glyceraldehyde-3-phosphate dehydrogenase	Gapdh
GTP-binding nuclear protein Ran	Ran
Heat shock protein HSP 90-beta	Hsp90ab1
Heterogeneous nuclear ribonucleoprotein K	Hnrnpk
Inosine-5’-monophosphate dehydrogenase 2 (Fragment)	Impdh2
Keratin, type II cytoskeletal 2 oral	Krt76
L-lactate dehydrogenase	Ldha
Malate dehydrogenase, mitochondrial	Mdh2
MCG10343, isoform CRA_b	Slc25a3
MCG11348	Rpl7a
MCG123443	Rps16
MCG23377, isoform CRA_b	Gm8797
MCG23455, isoform CRA_e	Rpl13a
Multifunctional protein ADE2 (Fragment)	Paics
Nucleophosmin	Npm1
Phosphoglycerate kinase 1	Pgk1
Poly(rC)-binding protein 1	Pcbp1
Pyruvate kinase PKM	Pkm
Receptor of activated protein C kinase 1	Rack1
Ribosomal protein L4	Rpl4
Ribosomal protein S14	Rps14
Tpm3 protein	Tpm3
Triosephosphate isomerase	Tpi1
Tubulin beta-5 chain	Tubb5
Uncharacterized protein	Actb
Uncharacterized protein	Cct8
Protein Group	# of proteins
Translation/ mRNA processing	13
Metabolism	9
Structural	4
Regulatory/ Signaling	2
Nucleotide Synthesis	2
Nuclear/ Mitochondrial Transport	2
Other	2