A FABP4-PPARγ signaling axis regulates human monocyte responses to electrophilic fatty acid nitroalkenes

Abstract

Nitro-fatty acids (NO₂-FA) are electrophilic lipid mediators derived from unsaturated fatty acid nitration. These species are produced endogenously by metabolic and inflammatory reactions and mediate anti-oxidative and anti-inflammatory responses. NO₂-FA have been postulated as partial agonists of the Peroxisome Proliferator-Activated Receptor gamma (PPARγ), which is predominantly expressed in adipocytes and myeloid cells. Herein, we explored molecular and cellular events associated with PPARγ activation by NO₂-FA in monocytes and macrophages. NO₂-FA induced the expression of two PPARγ reporter genes, Fatty Acid Binding Protein 4 (FABP4) and the scavenger receptor CD36, at early stages of monocyte differentiation into macrophages. These responses were inhibited by the specific PPARγ inhibitor GW9662. Attenuated NO₂-FA effects on PPARγ signaling were observed once cells were differentiated into macrophages, with a significant but lower FABP4 upregulation, and no induction of CD36. Using in vitro and in silico approaches, we demonstrated that NO₂-FA bind to FABP4. Furthermore, the inhibition of monocyte FA binding by FABP4 diminished NO₂-FA-induced upregulation of reporter genes that are transcriptionally regulated by PPARγ, Keap1/Nrf2 and HSF1, indicating that FABP4 inhibition mitigates NO₂-FA signaling actions. Overall, our results affirm that NO₂-FA activate PPARγ in monocytes and upregulate FABP4 expression, thus promoting a positive amplification loop for the downstream signaling actions of this mediator.

Graphical abstract

Highlights

• •Nitro-Fatty Acids (NO₂-FA) behave as partial PPARγ ligands in human monocytes and macrophages.
• •PPARγ activation induces Fatty Acid Binding Protein 4 (FABP4) expression in human monocytes and macrophages.
• •FABP4 binds NO₂-FA in vitro, suggesting its involvement in NO₂-FA intracellular transport.
• •FABP4 participates in NO₂-FAcell signaling actions in human monocytes.

1. Introduction

Nitrated derivatives of unsaturated fatty acids (NO₂-FA) mediate pleiotropic cell signaling actions [1,2]. NO₂-FA are endogenously generated upon nitrogen dioxide (·NO₂) addition to double bonds [3]. Among unsaturated fatty acids, those containing conjugated diene moieties are preferential substrates for nitration, because of the greater reactivity of the external flanking carbons of the conjugated diene moiety [4]. Conjugated linoleic acid (CLA), the most abundant dietary conjugated fatty acid, has been identified as a major endogenous substrate for nitration, generating nitro conjugated linoleic acid (NO₂-CLA) [5]. In healthy humans, CLA nitration occurs during digestion of lipid-containing foods, leading to nanomolar concentrations of unesterified NO₂-CLA in plasma [6]. Current data indicate that NO₂-CLA levels rise in metabolically-stressed and/or inflamed tissues because of downstream nitro-oxidative reactions stemming from nitric oxide synthase induction in inflammatory cells and tissues [2,7,8]. Despite the much lower reactivity of monounsaturated fatty acids, low levels of nitrated oleic acid have been reported in human plasma [9].

NO₂-FA modulate cell metabolic and inflammatory responses [5,[10], [11], [12], [13], [14], [15], [16], [17]] by multiple mechanisms, predominantly those involving post-translational modifications (PTM) of target proteins that regulate intracellular signaling responses and gene expression. These PTMs are a consequence of the soft electrophilic nature of NO₂-FA, which promotes reversible Michael addition reactions with soft nucleophilic groups such as His or Cys in proteins, leading to alterations in protein cell distribution and/or function [18]. The nuclear receptor PPARγ (Peroxisome Proliferator-Activated Receptor γ) is one target of NO₂-FA. PPARγ is a multi-domain protein with high affinity for lipophilic ligands. Agonists stabilize the active state of PPARγ, regulating the expression of genes involved in lipid metabolism or inflammation [19,20]. NO₂-FA bind to the PPARγ ligand-binding domain (LBD) and covalently react with a redox-sensitive Cys-285 in the LBD, partially activating the transcriptional function of the receptor [[21], [22], [23], [24], [25]]. This contrasts with the PPARγ activation profile induced by stronger ligands or full agonists such as the family of synthetic thiazolidinedione activators [26]. Besides PPARγ, Keap1 (Kelch-like ECH-associated protein 1) was described as a NO₂-FA target with implications for cytoprotective responses [[27], [28], [29], [30]]. NO₂-FA react with key cysteine residues in Keap1, resulting in reduced proteasomal degradation of the transcription factor Nrf2 (Nuclear factor-erythroid 2-related factor 2), and activation of Nrf2-dependent transcription of anti-oxidant and detoxifying genes, such as heme oxygenase-1 (HMOX1) and glutamate cysteine ligase modulatory gene (GCLM). Finally, NO₂-FA adduction to members of the heat shock protein family (HSP) that sense electrophiles, leads to activation of HSF1-regulated gene expression [30] and downstream anti-inflammatory effects in myeloid cells [31,32]. This NO₂-FA signaling mechanism in endothelial cells stimulates HSPA1A expression [28]. Beyond the pleiotropic nature of NO₂-FA, their protein targets will depend on several factors including cell redox and activation status as well as the intracellular half-life and stability of NO₂-FA.

The effect of NO₂-FA on PPARγ activation has primarily been studied in a metabolic context using fibroblasts, adipocytes, mammary epithelial (MCF7), or kidney cell lines (CV-1) [[33], [34], [35]]. Reporter assays have also been used but do not reliably reproduce physiological PPARγ expression levels and interactions with co-regulators (co-activators and co-repressors) that modulate its transcriptional activity [[21], [22], [23]]. In aggregate, NO₂-FA activation of PPARγ and subsequent modulation of cell functions is still poorly understood, particularly in the context of immunological responses. Monocytes and macrophages are innate cell populations of foremost importance in mediating integrated inflammatory responses, eliminating pathogens and contributing to tissue homeostasis. The recruitment of monocytes and their subsequent differentiation into macrophages gains relevance during inflammation to reinforce the immune response. Previously, PPARγ-independent effects on monocyte and macrophage inflammatory responses to NO₂-FA have been reported [10,36]. In this work, we examined NO₂-FA activation of PPARγ in both human monocytes undergoing differentiation into macrophages (termed differentiating monocytes) and macrophages.

We report herein that low μM levels of NO₂-FA activated PPARγ in differentiating monocytes and to a lesser extent in already-differentiated macrophages. The most robust PPARγ-regulated gene expression response in these cells was the upregulation of FABP4, a member of the fatty acid-binding protein family (FABP) that transports fatty acids between cell compartments [37]. NO₂-FA have limited solubility in aqueous milieu, thus this FABP4 upregulation and transport capability induced a significant impact on NO₂-FA trafficking to nuclear and cytoplasmic targets including PPARγ, in turn regulating downstream cell signaling responses. These responses were abrogated by FABP4 inhibitors in differentiating monocytes, affirming that FABP4 plays a crucial role in the transduction of NO₂-FA by (at least) PPARγ, Keap1/Nrf2 and HSP-regulated signaling networks.

2. Materials and methods

2.1. Chemical reagents

Reagents of analytical grade were purchased from Sigma (St. Louis, MO, USA) unless otherwise stated. Octadec-9-enoic acid (oleic acid, OA), octadec-9,11-dienoic acid (conjugated linoleic acid, CLA) and 5,8,11,14-eicosatetraenoic acid (arachidonic acid, AA) were obtained from Nu-Check Prep, Inc (Elysian, MN, USA). 9- and 12-nitro-octadec-9,11-dienoic acid (9-NO₂-CLA and 12-NO₂-CLA), 9- and 10-nitro-octadec-9-enoic acid (9-NO₂-OA and 10-NO₂-OA) and 10-nitro-octadecanoic acid (NO₂-SA) were synthesized as described previously [5,38,39]. The terms NO₂-CLA and nitro oleic acid (NO₂-OA) refer to the mixture of the corresponding above-mentioned positional isomers. AA nitration was carried out as previously described [11] to obtain a mixture of positional isomers referred to as nitroarachidonic acid (NO₂-AA). Rosiglitazone (Rosi), GW9662, and HTS01037 (HTS) were obtained from Cayman Chem (USA) while BMS 309403 (BMS) was acquired from ApexBio (USA).

2.2. Recombinant mouse FABP4 and rabbit anti-mouse FABP4 polyclonal antibodies

Recombinant mouse FABP4 (rFABP4) was expressed and purified following conventional protocols as previously described (Supplementary Fig. 1) [40]. Polyclonal antiserum against rFABP4 was raised in rabbits following standard protocols. All procedures were carried out in accordance with the ethical guidelines of the Honorary Commission of Animal Experimentation (CHEA) from UdelaR. Briefly, a New Zealand rabbit was injected subcutaneously with 500 μg of purified rFABP4 in an emulsion made of water in oil, prepared with Incomplete Freund Adjuvant. A second dose (booster) was similarly performed when the serum antibody titer anti-rFABP4 significantly dropped (about 16-times lower than the maximum reached). Bleeding was done at day 47 post-booster. The polyclonal antisera obtained showed a titer of 1/48.000 by ELISA and showed to recognize by Western blot a 14 kDa band present in a THP-1 cell extract, which was compatible with FABP4. The fraction of polyclonal anti-rFABP4 antibodies was purified by immunoaffinity using rFABP4 conjugated to NHS-Sepharose and 0.1 M glycine pH 2.0 for elution. In parallel, we obtained the polyclonal antisera fraction non-specifically bound to Sepharose as a control.

2.3. Cell culture for NO₂-FA stimulation

Human pre-monocytic THP-1 cells (ATCC, USA), a suitable model of primary human monocytes and macrophages [41], was used. Cells were cultured in RPMI medium (RPMI 1640 supplemented with 10 mM HEPES, 1.5 g/L sodium bicarbonate, 1 mM sodium pyruvate, 2 mM glutamine, 100 U/mL penicillin, 0.1 mg/mL streptomycin and 250 ng/mL amphotericin B) plus 10% fetal bovine serum (FBS, GIBCO). Cells were maintained at 37 °C in a humidified atmosphere of 5% CO₂ in air, and subcultured every 3–4 days to maintain cell density between 0.2 and 1.0 × 10⁶ cells/mL. Cells were plated at 5 × 10⁵ cells/well in RPMI medium supplemented with 5% FBS and macrophage differentiation was induced with 50 ng/mL phorbol esters (PMA, SIGMA). NO₂-FA, their corresponding fatty acid precursors (FA) and Rosi (a well-known PPARγ ligand) were tested at a final concentration of 1 μM, and the vehicle (DMSO) was used as a negative control. For assessing effects on differentiating monocytes, tested compounds were added together with PMA. For assessing effects on completely differentiated macrophages, macrophages were obtained after 72 h differentiation with PMA, cultured in RPMI medium without PMA for 48 h, and subsequently stimulated with NO₂-FA, Rosi or controls. In all assays, stimulation was stopped after 6 h or 16 h stimulation for mRNA levels and protein synthesis analysis, respectively. In order to evaluate the contribution of PPARγ activation to NO₂-FA-induced effects, similar experiments were done in the presence of a PPARγ inhibitor, GW9662. Briefly, cells were treated with PMA for 3.5 h, and then for 30 min with GW9662 (1 or 2.5 μM). Cells were then stimulated with NO₂-FA or controls and mRNA and protein levels examined at 6 h or 16 h post-stimulation, respectively. The role of FABP4 on NO₂-FA cell signaling actions was evaluated using the FABP4 inhibitors BMS [42] and HTS [43] in the presence of 5% delipidated FBS (PAN Biotech, Germany). Cells were pretreated with PMA together with BMS (25 μM) or HTS (15 μM) for 2 h and then stimulated with NO₂-FA or controls for 6 h mRNA expression was then examined as described below. In the case of HTS, an additional pre-incubation (24 h) was carried as previously described [44].

2.4. Cell viability

Cell viability was analyzed using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide] assay. Briefly, cells were cultured in the presence 200 μg/ml of MTT in PBS with 0.1% glucose for 3 h at 37 °C, 5% CO₂. Afterward, cells were washed with PBS, formazan was dissolved in DMSO and the absorbance at 560 nm was measured. In the case of differentiating monocytes, it was also analyzed on the basis of exclusion of the fixable viability dye (LIVE/DEAD, Invitrogen) and measured by flow cytometry.

2.5. Gene expression analysis

After stimulation, total cellular mRNA was isolated from cells using the TRIzol Reagent (Ambion, Life Technologies, USA) accordingly to the manufacturer's recommendations. Traces of DNA were removed by treatment with DNAse-I following the manufacturer's instructions (Fermentas, Thermo Fisher Scientific, USA). The synthesis of the first strand complementary DNA (cDNA) was carried out using 1 μg of total mRNA, M-MLV reverse transcriptase (200 Units, Life Technologies, USA), 0,5 mM dNTPs, 5 μM random primers, Ribolock (40 Units, Fermentas) at 25 °C for 10 min, followed by 37 °C for 50 min and 70 °C for 15 min. Template DNA was amplified by quantitative real-time PCR using Quantitect SYBR Green PCR kit (Qiagen, Hilden-Germany) in a Rotor-Gene Q real-time PCR cycler (Qiagen), according to the following protocol: 15 min at 95 °C, followed by 40 cycles at 94 °C for 15 s, at 50 °C for 30 s and at 72 °C for 30 s.

The primers, designed using Primer Express software (Applied Biosystems, USA), are listed in Table 1 and were used at a final concentration of 0.9 μM. The expression of genes of interest was normalized using GAPDH as housekeeping gene, except for assays using BMS where 18S was used. The relative mRNA amount in each sample was estimated using the 2^−ΔΔCt method [45], where ΔCt = Ct gene of interest-Ct housekeeping, and expressed as relative mRNA levels in the test group compared to the control group.

Table 1.

Primers for real-time PCR.

Gene	Forward primer	Reverse primer
18S	GTAACCCGTTGAACCCCATT	CCATCCAATCGGTAGTAGCG
ABCA-1	TTTCCAGGCCAGTACGGAAT	TCGCCAAACCAGTAGGACTT
ABCG-1	CTGACATTTCCCCTGGAGATG	TCCAGTACACGATGCTGCAGTA
CD36	GTGGCAGCTGCATCCCATAT	TCTGACTTGGAACATAGAAGATTTTGA
CPT1A	GGTGGTGGGCGTGATGA	CAGTTGGCCGTTTCCAGAGT
FABP4	GCCAGGAATTTGACGAAGTCAC	TTCTGCACATGTACCAGGACAC
FABP5	CCCTGGGAGAGAAGTTTGAAGA	AATGCACCATCTGTAAAGTTGCA
GAPDH	ACCCACTCCTCCACCTTTG	CTCTTGTGCTCTTGCTGGG
GCLM	AGACGGGGAACCTGCTGAA	TCATGAAGCTCCTCGCTGTC
HO1	AAGACTGCGTTCCTGCTCAA	GGGGGCAGAATCTTGCACTT
HSP70	CCACCAAGCAGACGCAGAT	GCCCTCGTACACCTGGATCA
Il1B	TTGGTGATGTCTGGTCCATATGA	GGACATGGAGAACACCACTTGTT
MCP1	GCTCAGCCAGATGCAATCAA	GCCTCTGCACTGAGATCTTCCT
PPARG	CAACAGACAAATCACCATTCGTTAT	GGATGGCCACCTCTTTGCT

2.6. FABP4 detection by immunofluorescence

THP-1 cells (5 × 10⁵ cells) were seeded on coverslips and stimulated with PMA (50 ng/mL) plus 1 μM concentration of NO₂-FA, FA, Rosi or DMSO (vehicle) in RPMI supplemented with 5% fetal delipidated FBS, for 10 h at 37 °C and in a humidified atmosphere of 5% CO₂. Cells were subsequently fixed with 4% paraformaldehyde in PBS, and coverslips were incubated overnight at 4 °C with the affinity-purified rabbit anti-rFABP4 antibody (1:100) or the corresponding control antibody. After washing 3 times with PBS, cells were stained using goat anti-rabbit IgG (H+L) conjugated with Alexa Fluor TM 488 (1:1600, Invitrogen) for 1.5 h, and then with DAPI (300 μM, Calbiochem-Novabiochem, USA) for 30 min. Cells were observed under epifluorescence microscopy (Nikon E800, Japan). To measure FABP4 nuclear translocation, cells were similarly stimulated for 10 h with PMA plus 1 μM NO₂-FA in order to reach detectable FABP4 levels, Leptomycin B (10 ng/ml) was added for 30 min and a short re-stimulation (30 or 60 min) with NO₂-FA, FA or Rosi (between 1 and 10 μM) was performed. Cells were stained as described above, but also using DAPI and rhodamine-phalloidin (0.3 μg/ml, Santa Cruz Biotechnology, USA) for nuclear and actin staining, respectively. Images were obtained with an LSM 800-AiryScan confocal microscope (Zeiss, Germany) and processed using FIJI/ImageJ [46]. Quantification was performed using at least 30 cells per condition in three independent experiments. The nucleus and the whole cell were delimitated according to DAPI and rhodamine staining, respectively, and integrated density (selected area*mean grey value, ID) was measured for each cell compartment. FABP4 distribution was analyzed by determining the nuclear/cytoplasm ratio, calculated as the Nuclear ID/(Total ID – Nuclear ID) (N/T-N).

2.7. FABP4 detection in cell extracts by Western blot

THP-1 cell extracts were obtained after 16 h of NO₂-FA stimulation. Briefly, monocytes were washed two times with RPMI without FBS at 37 °C, lysed using hypotonic buffer (Hepes 10 mM, EDTA 0.5 mM, KCl 10 mM, DTT 1 mM and protease inhibitor cocktail SIGMA, pH 7.5) and sonicated (three rounds of 1 min pulses at 15% and 30% of total potency) using an Omni-Ruptor 4000 (OMNI International Inc.). Samples were analyzed by SDS-PAGE at reducing conditions using polyacrylamide 12.5% gels and following conventional protocols. The proteins were transferred to a PVDF membrane (Millipore) and, after blocking with PBS 0.5% bovine serum albumin (BSA), probed with rabbit anti-rFABP4 antisera (1:20000 in PBS containing 0.05% BSA and 0.05% Tween 20) or rabbit anti-human tubulin IgG (1:1000, Cell Signaling, USA) followed by peroxidase-conjugated anti-rabbit IgG (1:2000, Calbiochem). Detection was carried out using SuperSignal West Pico Chemiluminescent Substrate (Thermo Fischer Scientific) in an imaging system (G:BOX, Syngene, India).

2.8. CD36 detection by flow cytometry

Stimulated THP-1 monocytes (1 × 10⁶ cells/well) were washed twice with cold PBS and detached by treatment with PBS containing 1 mM EDTA, 0.1% NaN₃, 0.1% glucose (PBS_FC buffer) for 10 min at 4 °C. After blocking with PBS_FC supplemented with 0.1% SFB, cells were incubated with mouse IgM anti-human CD36 (BD Biosciences, USA) or with an IgM isotype control (BD Biosciences) following manufacturer's recommendations. Then, cells were washed 3 times and development was carried out by incubation with phycoerythrin-conjugated anti-mouse IgM (BD Biosciences) according to the manufacturer's protocol. Cells were recorded on a FACSCalibur equipment (BD Biosciences) and analyzed with FlowJo (version 7.6, USA, www.flowjo.com).

2.9. FABP4 binding assays

The fluorimetric binding assay was performed as previously described [47]. rFABP4 (3 μM) was incubated with the hydrophobic fluorescence probe 8-anilinonaphthalene-1-sulfonic acid (ANS, 10 μM, Molecular Probes) for 5 min in 1 mL final volume of phosphate buffer (5 mM K₂HPO₄, 5 mM KH₂PO₄, 150 mM KCl, pH 7,4). Using a Fluorolog-3 (Horiba-Yvon) spectrofluorometer, the probe was excited at 400 nm (excitation slit 1), and emission spectra was registered between 420 and 600 nm (emission slit 2) at 25 °C, using FluorEssence™ software. Then, small amounts of the NO₂-FA were added progressively and the emission spectrum was registered 2 min after each addition to reach FABP4 saturation. NO₂-CLA, NO₂-OA, and NO₂-AA were tested and compared with their FA precursors (CLA, OA and AA, respectively). Each experiment was done at least in triplicates. Fluorescence data were fitted to a hyperbolic decay using SigmaPlot (version 11.0, Systat Software, Inc., USA, www.systatsoftware.com). The apparent dissociation constant (K_dapp) was calculated using the equation EC_50ligand/[ANS] = K_d,ligand/K_d,ANS as previously described [48]. To obtain K_d,ANS, FABP4 (2 μM) was titrated with the fluorescent probe until reaching saturation. Curves were fitted using SigmaPlot to a one-site ligand binding saturation model, and K_d,ANS was calculated as the mean of four independent experiments.

2.10. In silico prediction and analysis of FA and NO₂-FA binding to murine FABP4 in 1:1 complexes

The crystallographic structure of the available murine mFABP4:oleate complex in the open portal (inactive for translocation) conformation (PDB 1LID) [49] was taken as starting point for constructing the corresponding mFABP4:FA or mFABP4:NO₂-FA 1:1 complexes in solution and running 1.2 μs molecular dynamics (MD) NPT simulations at 310 K and 1 atm using the AMBER 16 suite [50] (see details on the simulations and trajectory analysis in the supplementary information). The study included OA and conjugated CLA, and the corresponding regioisomeric nitroalkene derivatives (9-/10-NO₂-OA and 9-/12-NO₂-CLA) as ligands, whose anionic structure in solution was determined by Density Functional Theory (DFT) quantum mechanics (QM) calculations [51] in a IEF-PCM continuous solvent [52] using Gaussian09 [53] (details on QM modeling in the supplementary information). Representative structures for each of the six complexes under study were obtained from MD by hierarchical-agglomerative clustering analysis and the presence of active/inactive for translocation conformations was monitored inspecting the distance between Phe57 (βC-βD loop) and Thr29 (αII helix) along the trajectories with cpptraj, taking values smaller than 8.5 Å as indicative of a closed conformation. In order to analyze the trends on binding affinity along the series of FA/NO₂-FA looking for evidence on eventual regioisomeric modulation, MM/PB(GB)SA binding free energies [54] were also obtained within a single-trajectory approach, extracting 201 snapshots from the 1.2 μs MD simulations separated every 500 ns (the first 200 ns were discarded) with the mmpbsa.py module [55] of AmberTools17 [50].

2.11. Data analysis

All experiments were performed at least three independent times (n = 3) in duplicate. Statistical analysis was carried out with GraphPad Prism (version 5, GraphPad Software, USA, www.graphpad.com) using two-way analysis of variance (ANOVA) for multiple comparisons, with Tukey post-test, or one-way ANOVA with Dunn post-test as indicated. For immunofluorescence studies, when comparing two groups, an unpaired t-test was used. Differences were considered significant when p ≤ 0.05.

3. Results

3.1. NO₂-FA activate PPARγ in differentiating monocytes and to a lesser extent in macrophages

For studying the ability of NO₂-FA to activate PPARγ in differentiating monocytes and macrophages, cell response to the full PPARγ agonist Rosiglitazone (Rosi) was firstly examined by quantitative analysis of mRNA expression of a series of potential PPARγ reporter genes (Supplementary Fig. 2). FABP4 was the most upregulated gene in both cell types, exhibiting the highest increment at 6 h post-stimulation (about 20-fold and 3-fold, Supplementary Figs. 2e and 2f, respectively). FABP5 and CD36 expression were moderately upregulated by Rosi in differentiating monocytes and macrophages, with increases slightly higher in monocytes than macrophages. Regarding ABCA1, ABCG1, and CPT1A, Rosi induced dissimilar and modest responses in monocytes and macrophages. Based on these results, FABP4, FABP5, and CD36 were chosen as target genes for studying the effect of NO₂-FA on PPARγ activation.

Studies were carried out using low micromolar concentrations of NO₂-OA, NO₂-CLA, and NO₂-AA (up to 1 μM), which are likely close to physiological levels in an inflammatory milieu [17] and did not cause alterations in cell viability (Supplementary Fig. 3). We found that all NO₂-FA, but not their corresponding FA precursors, were able to induce statistically significant increases in FABP4 mRNA levels in differentiating monocytes (Fig. 1a), showing the relevance of nitration for PPARγ activation. In contrast, FABP5 mRNA levels were unchanged after NO₂-FA stimulation (Fig. 1b) and only NO₂-CLA caused a modest increment in CD36 mRNA levels, suggesting that NO₂-CLA was the most potent PPARγ activator among tested NO₂-FA (Fig. 1c). In macrophages, NO₂-FA stimulation led to increases in FABP4 mRNA levels (Fig. 1e), but the magnitude of these increases was much lower than that observed in monocytes (1.6-fold vs. 7.5-fold for NO₂-CLA, respectively). Furthermore, no changes in FABP5 and CD36 expression were observed in macrophages after NO₂-FA stimulation while, as expected, Rosi induced significant rises in these PPARγ target genes (Fig. 1f and g). Moreover, we measured PPARG mRNA levels in both cell types to determine whether Rosi and NO₂-FA effects could induce an increase in PPARG expression. Results showed that exposure to Rosi and NO₂-FA did not significantly affect PPARG expression (Fig. 1d and h). In aggregate, these results indicate that NO₂-FA could act as partial PPARγ agonists in human differentiating monocytes and, to a lesser extent, in human macrophages, enhancing the transcription of genes associated with lipid metabolism such as CD36 and FABP4.

Fig. 1.

NO₂-FA stimulated FABP4 and CD36 mRNA expression in differentiating monocytes, and to a lesser extent in macrophages.

Differentiating THP-1 monocytes (a, b, c, d) and THP-1 macrophages (e, f, g, h), obtained by treatment with PMA, were stimulated with 1 μM concentration of Rosi, NO₂-FA or their corresponding FA precursors. In parallel, cells were incubated with DMSO (vehicle) as a control. After 6 h stimulation, cells were homogenized to obtain total mRNA, and FABP4, FABP5, CD36 and PPARγ mRNA levels were determined by real-time-qPCR and normalized to the housekeeping gene GAPDH. Results are represented as relative mRNA levels (fold increase) referred to Medium group, and correspond to the mean of at least three independent experiments ± SD. (*) Indicates statistically significant differences compared with DMSO (Two-way ANOVA, Tukey's Multiple Comparison test * p˂0.05, **p˂0.01, ***p˂0.001, ****p˂0.0001).

Next, we analyzed whether upregulation of FABP4 and CD36 transcription by NO₂-FA-stimulation led to an increment in protein levels. To that end, FABP4 levels were determined by immunofluorescence, using affinity-purified anti-rFABP4 antibodies or its corresponding control. In agreement with the effects observed at the mRNA level, stimulation of differentiating monocytes with Rosi or NO₂-FA, but not with native FA, led to significant increases in FABP4 protein expression (Fig. 2a and b). No reactivity was observed using control antibodies (Supplementary Fig. 4). FABP5 might cause cross-reactivity in this assay because of its similar amino acid sequence and 3D-structure to FABP4 (Supplementary Fig. 5). However, the fluorescence response was due to upregulation of FABP4 expression since FABP5 mRNA levels were unchanged after NO₂-FA stimulation of differentiating monocytes (Fig. 1b). A contrasting scenario was observed in macrophages, in which NO₂-FA treatment did not lead to increases in FABP4 protein levels (Fig. 2c). Moreover, Rosi was unable to increase FABP4 protein expression, despite an ability to enhance FABP4 transcription (Fig. 1e), suggesting that FABP4 synthesis might have reached a plateau under these conditions. On the other hand, CD36 protein levels were determined by flow cytometry. Among the tested NO₂-FA, only NO₂-CLA caused a weak, but statistically significant rise in CD36 protein expression in differentiating monocytes (Fig. 2d), which correlates with its ability to upregulate CD36 transcription.

Fig. 2.

NO₂-FA induced an increase in FABP4 and CD36 protein expression in differentiating monocytes, but not in macrophages.

Differentiating THP-1 monocytes (a, b, d) and THP-1 macrophages (c), obtained by treatment with PMA, were stimulated with 1 μM concentration of Rosi, NO₂-FA or their corresponding FA precursors, or the concentration indicated in the figure. In parallel, cells were incubated with DMSO (vehicle) as a control. After 12 h stimulation, cells were washed and stained using affinity-purified rabbit anti-rFABP4 antibodies and Alexa488 conjugated anti-rabbit IgG for development. Nuclei were stained using DAPI. Cells were observed and counted using epifluorescence microscopy. Images were analyzed using Image J software. In (a), panels show representative images of differentiating monocytes following incubation with the indicated stimuli or control. Graphs in (b) and (c) correspond to the quantification of FABP4 protein levels in differentiating monocytes and macrophages, respectively. Graphs indicate the Integrated Density (ID, in arbitrary units) for at least 70 cells per treatment ± SD. (*) and (#) Indicate statistically significant differences compared with DMSO or their non-nitrated precursor, respectively (One-way ANOVA, Dunn's Multiple Comparison test */#p˂0.05, ***/###p˂0.001). In (d) treatments were carried out for 16 h, then cells were harvested, and CD36 protein levels were determined by flow cytometry using a mouse IgM anti-human CD36 antibody and a phycoerythrin-conjugated anti-murine IgM for development. Graph represents protein levels normalized to Medium group, and are the mean of three independent experiments ± SD. (*) Indicates statistically significant differences compared with DMSO (Two-way ANOVA, Tukey's Multiple Comparison test * p˂0.05, **p˂0.01, ***p˂0.001, ****p˂0.0001).

Considering the robustness of NO₂-FA transcriptional responses and their impact on protein expression, studies were then focused on differentiating monocytes. The role of PPARγ on NO₂-FA induction of FABP4 and CD36 expression was evaluated using the PPARγ specific inhibitor GW9662. GW9662 treatment inhibited FABP4 and CD36 mRNA induction triggered by Rosi and NO₂-FA in differentiating monocytes (Fig. 3a and b, respectively). Similarly, analysis of FABP4 levels in cell extracts by Western blot (Fig. 3c) showed that GW9662 caused a reduction in FABP4 protein levels induced by Rosi (80% inhibition) or NO₂-FA (72 to 89% inhibition, Fig. 3d). Altogether, these results indicate that NO₂-FA activated PPARγ in differentiating monocytes. Moreover, NO₂-FA effects on FABP4 and CD36 expression were much lower than those induced by Rosi, supporting the partial PPARγ agonist actions of nitroalkenes, in agreement with previous studies using transfected cell lines [25].

Fig. 3.

Upregulation of FABP4 and CD36 expression by NO₂-FA was abolished bya PPARγ specific inhibitor.

Differentiating monocytes were pretreated with the PPARγ specific inhibitor GW9662 (1 μM) for 30 min and afterward, Rosi, NO₂-FA and their controls (DMSO and non-nitrated FA) were added at 1 μM to the culture. After 6 h post-stimulation, cells were homogenized to preserve total mRNA, and FABP4 and CD36 mRNA were measured by real-time qPCR and normalized to the housekeeping gene GAPDH. The expression of FABP4(a) and CD36(b) is represented as relative mRNA levels (fold increase) referred to Medium group (not shown), and correspond to the mean of three independent experiments ± SD. (*) and (#) Indicate statistically significant differences with DMSO or GW9662 treatment, respectively (Two-way ANOVA, Tukey's Multiple Comparison test */#p˂0.05, **/##p˂0.01, ***/###p˂0.001, ****/####p˂0.001). After 16 h stimulation, GW9662 effects on FABP4 upregulation by NO₂-FA were examined at the protein level by Western Blot using α-tubulin as loading control; a representative Western blot result is shown in (c), where two separate membranes from the same experiment were placed next to each other. FABP4 protein levels were quantified by densitometry of Western blot bands from three independent experiments and normalized to Rosi. The relative induction of FABP4 protein levels is shown in (d).

3.2. FA nitration did not abrogate FABP4 binding

The induction of FABP4 synthesis by NO₂-FA might influence their own intracellular transport and signaling because FABP4 is the main FABP isoform involved in intracellular FA transport in monocytes and macrophages. Nonetheless, the role of NO₂-FA binding to FABPs is unknown, thus we hypothesized that the presence of a nitro group on the fatty acyl chain might modify FA interactions with FABP4. Therefore, we studied NO₂-FA binding to FABP4 in vitro using a competitive fluorescent binding assay based on the displacement of a fluorescent hydrophobic probe (ANS), previously bound to FABP4. For this assay, we used delipidated murine rFABP4 as a model, recognizing that FABP4 is highly conserved among mammals, with murine and human FABP4 sharing a high level of primary and secondary structure identity (both higher than 91%, Supplementary Fig. 6). The addition of increasing amounts of NO₂-FA to rFABP4 bound to ANS (at saturation conditions induced by a protein to probe molar ratio of 1:3) caused a progressive decrease in fluorescence, indicating the displacement of ANS from rFABP4 (Fig. 4). Displacement curves were adjusted to a hyperbolic decay, suggesting a 1:1 FABP4:NO₂-FA stoichiometry ratio in agreement with previous reports [56,57], and allowed K_dapp calculation. Differences in binding affinity among NO₂-FA were noted. NO₂-OA behaved like a high-affinity ligand, whereas NO₂-CLA and NO₂-AA acted as weak affinity ligands since they displaced about 84%, 58%, and 46% of ANS at protein saturation condition, respectively [48]. This correlated with the fact that NO₂-OA showed the lowest K_dapp (Table 2, indicated with *). On the other hand, according to K_dapp, nitration of OA, CLA and AA caused a reduction in their rFABP4 affinity (Table 2, indicated with #), related to variations in binding free-energies of 0.27, 1.376 and 1.47 kcal mol⁻¹ for each mixture of NO₂-OA, NO₂-CLA and NO₂-AA regioisomers, respectively. Despite such variations, acyl chain nitration did not abrogate FA binding to FABP4, supporting the idea that FABP4 could also act as a NO₂-FA intracellular carrier.

Fig. 4.

ANS displacementfrom rFABP4 byNO₂-FA.

rFABP4 was loaded with ANS, and FA or NO₂-FA were added progressively, registering the emission spectrum of ANS after each addition. The increment in FA or NO₂-FA concentration induced a decrease in ANS fluorescence intensity. Each point corresponds to the integral of the ANS emission spectra at the indicated ligand concentration, and the line represents the adjustment to a hyperbolic decay. ANS displacements induced by FA (open circles) or NO₂-FA (filled circles) are shown for OA/NO₂-OA (a), CLA/NO₂-CLA (b), AA/NO₂-AA (c). Curves are representative of three independent experiments.

Table 2.

K_dapp and binding free energies at 298 K for murine FABP4:FA and FABP4:NO₂-FA (1:1 complexes).

Ligand	In vitro (rFABP4)		In silico murine FABP4 ΔGbind (kcal mol−1)
	Kdapp (μM)a,b	ΔGbindc (kcal mol−1)	MMGBSA	MMPBSA
OA	0.59 ± 0.20	−8.49 ± 0.34	−41 ± 5	−41 ± 9
9-NO2-OA	0.93 ± 0.16##e	−8.22 ± 0.17	−48 ± 5	−49 ± 8
10-NO2-OA			−49 ± 5	−49 ± 9
CLA	0.28 ± 0.03	−8.90 ± 0.11	−42 ± 6	−44 ± 7
9-NO2-CLA	3.03 ± 0.01∗∗∗∗/####e	−7.524 ± 0.003	−49 ± 5	−49 ± 7
12-NO2-CLA			−45 ± 5	−46 ± 8
AA	0.25 ± 0.02	−9.00 ± 0.08	n/ad	n/ad
NO2-AA	3.02 ± 0.08∗∗∗∗/####	−7.53 ± 0.03	n/ad	n/ad

^aValues correspond to the mean of three independent experiments ± SD, determined by ANS displacement experiments using each NO₂-FA regioisomeric mixture or the corresponding FA precursor.

^bThe (^#) and (*) superscripts indicate significant differences with respect to the corresponding FA precursor or NO₂-OA, respectively (One-way ANOVA, Tukey multiple comparison test */#p˂0.05, **/##p˂0.01, ***/###p˂0.001).

^cΔG_bind = -RTlnK_d-app at T = 298 K.

^dn/a: data not available, AA/NO₂-AA complexes were not simulated.

^eCorresponding to the mixture of positional isomers.

The in silico results show that FABP4-bound FA/NO₂-FA remained thermodynamically stable throughout the 1.2 μs simulation periods (Table 2, right) as previously reported for other 1:1 complexes established between murine FABP4 and different ligands [49,[58], [59], [60]]. The corresponding binding free-energies extracted from MM/GB(PB)SA calculations also enabled the evaluation of ligand affinities between regioisomers of NO₂-OA and NO₂-CLA (not assessed in vitro), without significant changes between isomers. These results provide further support that nitration did not compromise FA binding to FABP4.

3.3. FABP4 nuclear import occurs independently of NO₂-FA stimulation in differentiating monocytes

NO₂-FA promotion of FABP4 translocation into the nucleus was examined in differentiating monocytes, since this translocation has been correlated with PPARγ activation by synthetic agonists [61,62]. FABP4 translocation has been previously analyzed using artificial systems, where FABP4 was overexpressed in non-myeloid cell lines [63] or expressed as a fusion protein with GFP, which likely interferes with the ability of FABP4 to freely diffuse across the nuclear membrane [59,61,62]. To avoid this interference, studies were performed by confocal immunofluorescence microscopy using immunoaffinity-purified anti-rFABP4 antibodies for detection. FABP4 localization was first examined in differentiating monocytes stimulated with PMA plus NO₂-FA for 10 h in order to achieve a detectable FABP4 signal (FABP4 levels were very low in the absence of a PPARγ agonist, Fig. 2). FABP4 was equally distributed throughout the cell, showing a nucleus to cytoplasm ratio (N/T-N) of around 1 (Fig. 5a). Then, the effect of re-stimulation with Rosi, NO₂-FA or FAs on FABP4 localization was assessed after 30 or 60 min, as previously described for other ligands [62]. There were no significant changes in FABP4 nuclear transport in differentiating monocytes treated with a range of NO₂-FA and Rosi concentrations (1–10 μM) (Fig. 5b–e). In these assay conditions, total FABP4 levels in treated and control cells were comparable, indicating that re-stimulation did not induce FABP4 synthesis, which might have contributed to the cytoplasmic signal (Fig. 5f).

Fig. 5.

NO₂-FA did not increase FABP4 nuclear localization in differentiating monocytes.

(a-f) Differentiating THP-1 monocytes obtained by treatment with PMA were stimulated with 1 μM concentration of NO₂-FA for 10 h and then treated with LMB (10 ng/ml) and re-stimulated with Rosi, NO₂-FA or their corresponding FA precursors in a range of concentrations (1–10 μM). In parallel, cells were re-stimulated with DMSO (vehicle) as a control. After 30 min of re-stimulation, cells were washed and stained using affinity purified rabbit anti-rFABP4 or control antibodies, and Alexa488 conjugated anti-rabbit IgG for development. Nuclei were stained using DAPI. Cells were observed and counted using confocal microscopy. Images were analyzed using Image J software. Magnification (63x) of representative FABP4 distribution in cells treated with DMSO (a), CLA (b), NO₂-CLA (c), or Rosi (d). Quantification of the intracellular distribution calculated as the ratio Nuclear ID/(Total ID-Nuclear ID) (N/T-N) (e) and total integrated density (f) after 30 min of re-stimulation. Plots showed box and whisker with bars from minimum to maximum value corresponding to the quantification of at least 30 cells per condition. Results are representative from three independent experiments. (g) and (h) Superimposed representative structures for the most populated conformation corresponding to a closed portal (NLS exposed), for each FA or NO₂-FA, 1:1 complexes with murine FABP4 (backbone in grey) from MD simulations. Phe57 and key residues interacting with polar heads in FA (carboxylate) and NO₂-FA (carboxylate and NO₂) are evidenced in stick representation. (g) OA series: OA, 9-NO₂-OA and 10-NO₂-OA ligands in cyan, violet and red, respectively. (h) CLA series: CLA, 9 NO₂-CLA and 12-NO₂-CLA ligands in cyan, violet and red, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

In addition to FABP4 diffusion into the nucleus as a small protein, nuclear translocation could also depend on the presence of an atypical, and undetectable from inspection of the linear sequence, nuclear localization signal (NLS) in the three-dimensional structure of FABP4. This is constituted by three basic residues (Lys21, Arg30, and Lys31) [62]. The NLS is exposed in FABP4 by conformational changes induced upon binding of an activating ligand, thus promoting a transition from an inactive (open) to an active (closed) portal conformation. This is characterized by the Phe57 pointing inwards and pushing outwards the αI-loop-αII cap [53]. Because NO₂-FA did not promote FABP4 nuclear translocation in our cell model, we explored through molecular dynamics simulations whether NO₂-FA and their corresponding native FA precursors would facilitate NLS exposition. In agreement with previous observations [60], the dynamics of the apo-FABP4 showed the presence of both open and closed portal conformations, with the majority of the population in the latter (76%, Table 3). Similarly, all FABP4:FA and FABP4:NO₂-FA complexes, showed both open and closed conformations of FABP4, but there were differences in the percentage of structures exhibiting the closed conformation within the spectrum of conformations adopted by the protein. CLA and OA binding to FABP4 promoted a redistribution of the native population, favoring the closed portal conformation that exposes the NLS region (Fig. 5g and h and Table 3). In contrast, NO₂-FA binding showed dissimilar effects (Table 3). Binding of 10-NO₂-OA, but not of 9-NO₂-OA, enhanced the predominance of the Phe57 closed portal conformation to a similar extent as OA. On the other hand, 9- and 12-NO₂-CLA binding to FABP4 did not increase the predominance of the closed conformation as FA and 10-NO₂-OA did. Overall, these results indicated that NO₂-FA are not better inducers of an active/closed FABP4 conformation than their corresponding FA precursors.

Table 3.

Relative weight of closed portal conformations (expressed in percentage) from FABP4:FA and FABP4:NO₂-FA complexes 1.2 μs molecular dynamics simulations.

Ligand	FABP4:ligand complexes (1:1) closed portal conformation (%)
OA	87
9-NO2-OA	71
10-NO2-OA	87
CLA	91
9-NO2-CLA	71
12-NO2-CLA	66
No ligand	76a

^aPreviously reported as 62% by Li et al. [60] using a different Amber2012 force field (ff03) for protein simulation.

3.4. FABP4 transduces NO₂-FA signaling

Evaluation of FABP4 involvement in cell signaling by NO₂-FA in differentiating monocytes was performed using the FABP4 inhibitors BMS and HTS, which have higher affinities for FABP4 than NO₂-FA [42,43]. Monocytes were differentiated in the presence of FABP4 inhibitors and afterward, cells were stimulated with NO₂-FA or native FA as control, and FABP4 and CD36 expression followed to indicate PPARγ activation. In these assays, inhibitors did not affect basal levels of FABP4 (Supplementary Fig. 7). BMS treatment caused a reduction in the upregulation of FABP4 expression induced by NO₂-FA and abrogated the CD36 expression induced by NO₂-CLA (Fig. 6a and b). These results were similarly reproduced by HTS, which strongly inhibited FABP4 and CD36 expression by NO₂-FA (Fig. 6c and d). Then, we examined the effect of FABP4 inhibitors on the NO₂-FA-dependent induction of HMOX1, GCLM, and HSP70 expression. We found that BMS treatment elicited different effects on NO₂-OA and NO₂-CLA signaling. NO₂-OA induced significant increases in both HMOX1 and GCLM expression that were partially or completely inhibited by BMS, respectively (Fig. 6e and f). In contrast, NO₂-CLA caused a modest upregulation of HMOX1 expression that was not inhibited by BMS treatment (Fig. 6e). Analysis of NO₂-FA activation of the HSF1 signaling showed that NO₂-OA, but not NO₂-CLA, upregulated HSP70 expression, which was significantly reduced by BMS (Fig. 6g). In contrast, when the effect of HTS was assessed on Keap1/Nrf2 and HSF1 reporter gene expression, a significant increase in mRNA levels of target genes was observed, which impeded a direct comparison of gene expression between control and treated cells. Despite this effect, normalization of gene expression by their corresponding basal levels revealed that HTS and BMS induced comparable inhibitory effects on NO₂-FA-induced activation of Keap1/Nrf2 (Supplementary Fig. 8).

Fig. 6.

Blocking FABP4 binding activity inhibits NO₂-FA activation of PPARγ, Keap1/Nrf2 and HSF1 in differentiating monocytes.

(a-g) THP-1 cells were differentiated with PMA in the presence of the FABP4 inhibitor BMS (25 μM)/HTS (15 μM) or DMSO (control) for 2 h. Then, NO₂-FA (1 μM) or DMSO as control were added for 6 h. To examine NO₂-FA signaling actions, total mRNA was purified and FABP4(a and c), CD36(b and d),HMOX1(e), GCLM(f), or HSP70(g) expression, by real-time qPCR. Results are represented as relative mRNA levels (fold increase) referred to DMSO group and correspond to the mean of four independent experiments ± SD. (*) Indicates statistically significant differences with the corresponding control (DMSO or DMSO+BMS/DMSO+HTS, Two-way ANOVA, Tukey's Multiple Comparison test * p˂0.05, **p˂0.01, ***p˂0.001, ****p˂0.001). (#) Indicates statistically significant differences with the same treatment without inhibitor (Two-way ANOVA, Tukey's Multiple Comparison test #p˂0.05, ##p˂0.01, ###p˂0.001, ####p˂0.001).

3.5. FABP4 regulates NO₂-FA modulation of cytokine expression

Cytokine expression is a feature of the inflammatory profile of monocytes and macrophages, which is known to be modulated by NO₂-FA. Since PPARγ modulates inflammatory responses, we explored whether activation of the PPARγ/FABP4 axis might influence cytokine expression by differentiating monocytes. To that end, we determined MCP1 and IL1B expression in the presence or absence of GW9662 or BMS. NO₂-OA inhibited MCP1 and IL1B expression (Fig. 7a and d), while NO₂-CLA had a marginal effect on MCP1. Moreover, considering that GW9662 did not modify NO₂-OA effects, activation of PPARγ is likely not involved in NO₂-OA modulation of cytokine expression. Although PPARγ does not mitigate these anti-inflammatory effects, FABP4 does, since NO₂-OA effects on MCP1 and IL1B were inhibited by FABP4 blockade (around a 87% and 44%, respectively).

Fig. 7.

FABP4 modulates NO₂-FA inhibition of cytokine expression.

Differentiating THP-1 monocytes obtained by treatment with PMA were stimulated with 1 μM concentration of Rosi, NO₂-FA or their corresponding FA precursors (a–f). In parallel, cells were incubated with DMSO (vehicle) as a control. After 6 h stimulation, cells were homogenized to obtain total mRNA and MCP1 and IL1β mRNA levels were determined by real-time-qPCR and normalized to the housekeeping genes GAPDH or 18S. For BMS experiments, THP-1 cells were differentiated with PMA in the presence of the FABP4 inhibitor BMS (25 μM) for 2 h previous to NO₂-FA treatment (b) and (e). For GW9662 experiments, differentiating monocytes were pre-treated with GW9662 (1 μM) for 30 min previous to NO₂-FA treatment (c) and (f). The expression of MCP1(a–c) and IL1β(d–f) is represented as relative mRNA levels (fold increase) referred to DMSO, and correspond to the mean of three independent experiments ± SD. (*) Indicates statistically significant differences with DMSO/DMSO+BMS/DMSO+GW9662 treatment (Two-way ANOVA, Tukey's Multiple Comparison test *p˂0.05, **p˂0.01, ***p˂0.001, ****p˂0.001).

4. Discussion

The pleiotropic anti-oxidant and anti-inflammatory actions of NO₂-FA have led to the present evaluation of these species as new drug candidates designed to alleviate disorders in which physiological regulatory circuits intended to control inflammation fail [30,64]. In vivo studies in various pre-clinical inflammatory models support this possibility [65]. Despite the considerable past and current understanding of NO₂-FA actions in modulating inflammatory responses, these mostly focused on those regulated by NF-κB and Nrf2 [[27], [28], [29],66]. At present there is limited insight into the ability of NO₂-FA to activate PPARγ in macrophages and differentiating monocytes [10,36].

We found that NO₂-FA, but not native FA at similarly low concentrations, induced PPARγ activation in differentiating monocytes and macrophages, reinforcing the unique impact of the electrophilic character imparted by unsaturated FA nitration. While non-physiological concentrations of native FA activate PPARγ (around 100 μM in vitro) [[67], [68]], several characteristics differentiate the activity exerted by NO₂-FA from native FA. NO₂-FA, are orders of magnitude more potent activators than native FA, but are present at orders of magnitude lower concentrations, making comparisons challenging. Importantly, by reacting covalently, NO₂-FA have much longer occupancy times that are determined by the rate of the elimination reaction or protein turnover as opposed to the pharmacokinetic properties of native FAs. In addition, while NO₂-FA can compete with FA for ligand binding domain association, FAs will not be able to compete with adducted NO₂-FA, thus intracellular levels of NO₂-FA become less relevant for initiating and sustaining an activated state of the receptor.

Plasma levels of NO₂-FA are in the very low nM range [6,9], with net intracellular concentrations difficult to estimate as NO₂-FA are found esterified in complex lipids and covalently bound to proteins, small thiols, and other nucleophilic intracellular molecules [69]. Thus, the free acid fraction of NO₂-FA, while in equilibrium with these other components, is viewed to be a minor percentage [70]. Regarding intracellular concentrations of free FAs, limited data has been reported for monocytes and macrophages, with one report indicating at least 1 μmol/10⁶ THP-1 cells [71]. In any case, in our assays 10⁶ THP-1 cells were treated with 2 nmoles of NO₂-FA, a level orders of magnitude lower than native FA levels, affirming that NO₂-FA exert effects even when native FA are in excess. Extents of PPARγ activation by NO₂-FA were lower in macrophages, yet there were no changes in PPARG expression levels found between differentiating monocytes and macrophages. The transcriptomic profile of human monocytes undergoes alterations during differentiation towards macrophages, which include the induction of transcripts involved in palmitate and oleate synthesis, as well as in fatty acid desaturation and elongation, leading to differences in the lipidomic profile [72]. In this scenario, the lower ability of NO₂-FA to activate PPARγ in macrophages may be associated with the presence of higher-affinity and/or higher concentrations of competing endogenous ligands in this cell type, which might render NO₂-FA less effective in PPARγ activation. In line with this hypothesis, basal PPARγ activation, measured in terms of FABP4 mRNA levels, seems to be much greater in macrophages than in monocytes (around 16-fold, Supplementary Fig. 9).

NO₂-FA acted as partial PPARγ agonists in monocyte and macrophages, a result consistent with previous studies using biochemical approaches and transfected cell lines. In all conditions and for all reporter genes measured, the full PPARγ agonist Rosi caused a stronger activation than NO₂-OA and NO₂-CLA. This behavior is likely associated with less displacement of co-repressors and lower recruitment of co-activators in comparison with Rosi, as previously reported for NO₂-OA [25]. Considering the potential use of NO₂-OA as an anti-inflammatory drug, this feature is not a disadvantage, as partial agonists of PPARγ that discriminate between the beneficial actions and the adverse effects associated with full PPARγ activation may be of clinical benefit. Thiazolidinedione-related increases in cardiovascular events led to both market withdrawal and a restricted use in several countries [34]. In this regard, understanding the differences observed in the potency and transcriptional output of distinct NO₂-FA, can help explain the molecular events leading to the beneficial and/or undesirable effects triggered by PPARγ activation. Moreover,particular NO₂-FA regiosiomers may undergo different interactions with this nuclear receptor [22,23]).

Among NO₂-FA, NO₂-CLA induced a greater magnitude of PPARγ activation, as shown by its unique upregulation of CD36 expression, a response inhibited by GW9662. This expands our knowledge of the biological actions of NO₂-CLA in monocytes and macrophages and presumably other cells [5,17,73,74]. Because CLA behaves as a preferential FA target for nitration at physiological conditions [5], PPARγ activation by NO₂-CLA could be biologically more relevant than that mediated by other NO₂-FA. However, whether NO₂-CLA plays a role as an endogenous PPARγ agonist and physiological regulator is still controversial since NO₂-CLA levels in inflamed fluids or specific anatomic compartments of tissues from patients with acute or chronic inflammatory diseases are unknown. In a mouse peritonitis model, NO₂-CLA levels were ~20 nM in peritoneal lavage, indicating ~μM concentrations in undiluted native peritoneal fluid, as there is a limited recovery of this chemically-reactive lipid from the peritoneal cavity because of alternative reactions [17].

FABP4, the most highly upregulated PPARγ-reporter gene in response to NO₂-FA treatment, is a lipid carrier which binds and delivers ligands to PPARγ [[61], [62], [63],75]. FABP4 binds a series of NO₂-FA with an affinity similar or lower than that corresponding to native FA precursors, but in the intracellular milieu, this binding would depend on the concentration of FABP4, NO₂-FA as well as other competing FA ligands. On the other hand, the NO₂-FA binding affinity for FABP4 did not correlate with an ability to activate PPARγ, since NO₂-OA showed the highest FABP4 binding affinity (Table 2), but not the strongest PPARγ activation activity among NO₂-FA. Upon ligand binding, FABP4 undergoes structural changes, adopting the closed/active conformation that promotes nuclear translocation and, in turn, PPARγ activation. Thus, we compared the degree of FABP4 nuclear translocation induced by NO₂-FA and corresponding native FA precursors. In our experimental conditions, FABP4 nuclear translocation was similar in differentiating monocytes stimulated with NO₂-FA or FA precursor. Furthermore, molecular simulations showed that the closed/active conformation of FABP4 is present in the dynamics of both FABP4:FA and FABP4:NO₂-FA complexes as in the apo-FABP4. This agrees with the precept that in all assay conditions at least half of total FABP4 was present in the nucleus (Fig. 5). Taking into account the unique nature of NO₂-FA-FABP4 interactions, NO₂-FA seem not to play a major role in stabilizing the active FABP4 conformation. However, NO₂-FA-mediated FABP4 translocation might be difficult to detect, considering that, as discussed above, in the cellular milieu NO₂-FA must compete with relatively higher concentrations of endogenous FA, and NO₂-FA-FABP4 complexes are not more active in terms of nuclear translocation. Importantly, the lack of correlation between the binding affinity of NO₂-FA for FABP4 and the strength of the PPARγ activation response does not exclude the involvement of FABP4 in the mechanism of PPARγ activation by NO₂-FA. FABP4 levels might be high enough to carry high -and low-affinity ligands, even those in low concentration such as NO₂-FA. Indeed, two different FABP4 inhibitors inhibited NO₂-FA-mediated PPARγ activation. In aggregate, this data reveals that PPARγ activation by NO₂-FA induces FABP4 upregulation and thus instigates a positive amplification loop of this signaling pathway (Fig. 8).

Fig. 8.

Involvement oftheFABP4/PPARγsignalingaxis indifferentiating monocyteNO₂-FA signaling.

NO₂-FA behave as potential ligands of FABP4. NO₂-FA binding to FABP4 may protect them from irrelevant adduction reactions favoring their interaction with factors involved in intracellular pathways associated with anti-oxidative and anti-inflammatory responses (Keap1/Nrf2, HSPs/HSF1, NF-κB). In addition, FABP4 is involved in the activation of PPARγ triggered by NO₂-FA, which likely includes the trafficking and delivery of these lipids to PPARγ. As a result of PPARγ activation, FABP4 expression is upregulated, which would operate as a positive regulation mechanism of the FABP4/PPARγ axis.

Any differences in PPARγ activation by different NO₂-FA and FA could not be attributed to differences in NO₂-FA-FABP4 interactions, as nitration did not enhance FA binding to FABP4 or consequent FABP4 nuclear translocation. Therefore, it is more likely that nitration could modify the set of interactions that determine how the lipid acyl chain is accommodated in the ligand-binding domain (LBD) of PPARγ, thus influencing receptor conformation and activity. The chemical features of the ligand define the PPARγ-LBD structural changes that stabilize a particular receptor conformation and determine the position adopted by helix 12 and this domain's binding of co-activators and release of co-repressors [76]. In the case of NO₂-FA, the crystal structure of PPARγ complexed with two nitrolinoleic acid (NO₂-LA) regioisomers revealed that this nitroalkene establishes specific interactions with amino acids of the PPARγ binding pocket, which are not shared by the Rosi interaction network at the LBD, thus conferring partial agonism of transcriptional responses [24]. These unique events include interactions between the nitro group and Arg288 or Glu343, as well as interactions between Phe287 and the NO₂-LA backbone, required for stable binding of NO₂-LA. Furthermore, NO₂-FA interactions with PPARγ-LBD are further stabilized by the formation of a covalent bond with the redox-sensitive Cys285 in the LBD of PPARγ [25]. Consequently, differences in PPARγ activation by NO₂-CLA, NO₂-OA and NO₂-AA may be related to the set of non-covalent and covalent interactions that each nitroalkene establishes with the PPARγ-LBD, which in turn determine the complex stability and the degree of conformational changes induced by the occupancy of the ligand on the binding pocket and the downstream recruitment/release of coactivators and corepressors.

The potential of NO₂-FA to bind FABP4 and to upregulate FABP4 expression might have additional implications for tissue NO₂-FA responses. To explore this, we took advantage of the ability of NO₂-FA to induce HMOX1/GCLM and HSP70 expression via Keap1/Nrf2 and HSF1, respectively. In contrast with a modest relative activation of PPARγ, NO₂-OA was more active than NO₂-CLA. Using breast cancer cells, NO₂-OA elicited greater Nrf2 and lower PPARγ transcriptional activity responses than NO₂-LA [77]. Several factors might be involved in this differential potency for activating these signaling pathways. As noted for PPARγ, differences in the set of non-covalent/covalent bonds formed between NO₂-FA and the target protein could affect the strength of the signaling response. We speculate that for stress-induced transcriptional responses, differences related to NO₂-FA Michael addition kinetics and equilibrium constants with thiol targets may be more important, since this will affect thiol residency time, net populations of thiol targets and downstream propagation of the signal. These differences could also affect “signaling-productive” versus alternative NO₂-FA reactions, as this would impact the formation of adducts with thiol-bearing pools of molecules, including glutathione and proteins not involved in catalysis or the regulation of gene expression. In this context, the greater activity showed by NO₂-OA compared to NO₂-CLA might be explained by the fact that in our assays, at the same molar concentration, NO₂-OA could react more rapidly with nucleophiles than NO₂-CLA [78]. When analyzing the involvement of FABP4 in NO₂-FA-induced signaling responses, FABP4 inhibitors in turn limited the HMOX1 upregulation induced by NO₂-OA, but not by NO₂-CLA. We envision that FABP4 could also limit NO₂-FA reaction with non-productive cellular nucleophiles such as glutathione, that in turn promotes cellular export of NO₂-FA-glutathione adducts [79]. Thus, the net effect of FABP4 sequestration would be to promote NO₂-FA reactions with more productive targets rather than inactivation and export.

Expanding the relevance of FABP4 in regulating NO₂-FA signaling, FABP4 inhibition attenuated NO₂-FA inhibition of pro-inflammatory cytokine expression in monocytes and macrophages. There was not a role for PPARγ in mediating these anti-inflammatory responses, suggesting that NF-κB transrepression by NO₂-FA-mediated PPARγ activation is less impactful than the inhibition of pro-inflammatory gene expression via the direct adduction of NF-κB p65 by NO₂-FA [10].

5. Conclusions

This work reveals that the lipid binding protein FABP4 is upregulated in monocytes in response to NO₂-FA mediated PPARγ activation. It also shows that FABP4 regulates NO₂-FA signaling through a chaperone function that probably improves the cellular transport of NO₂-FA and the downstream amplification of adaptive signaling responses. Upon induction of FABP4 expression by inflammatory stimuli and NO₂-FA-FABP4 binding, the expression of Nrf2 and HSF-1 regulated tissue-protective genes is enhanced and expression of pro-inflammatory cytokines is inhibited. These responses to NO₂-FA, all associated with inflammatory resolution and tissue repair [[80], [81], [82], [83]] are facilitated by the chaperone activity of FABP4.

Funding

This work was supported by Comisión Sectorial de Investigación Científica (CSIC, Uruguay, grants Programa Iniciación a la Investigación C272 and Programa Grupos 536) and Agencia Nacional de Investigación e Innovación (ANII, Uruguay, fellowships to MLB) and National Institutes of Health (NIH, USA, grants R01GM125944 and R01DK112854 to FJS and grants P01HL103455 and R01HL132550 to BAF. JB, ELC, HR, and AMF are active members of the Sistema Nacional de Investigadores (SNI-ANII, Uruguay) and Programa de Desarrollo de las Ciencias Básicas (PEDECIBA Áreas Química y Biología, Uruguay), whose continuous support is gratefully acknowledged.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: BAF and FJS acknowledge an interest in Complexa, Inc. and Creegh Pharmaceuticals.

Appendix A. Supplementary data

The following is the Supplementary data to this article: