Nitrite and nitrate-dependent generation of anti-inflammatory fatty acid nitroalkenes

Abstract

A gap in our understanding of the beneficial systemic responses to dietary constituents nitrate (NO₃⁻), nitrite (NO₂⁻) and conjugated linoleic acid (cLA) is the identification of the downstream metabolites that mediate their actions. To examine these reactions in a clinical context, investigational drug preparations of ¹⁵N-labeled NO₃⁻ and NO₂⁻ were orally administered to healthy humans with and without cLA. Mass spectrometry analysis of plasma and urine indicated that the nitrating species nitrogen dioxide was formed and reacted with the olefinic carbons of unsaturated fatty acids to yield the electrophilic fatty acid, nitro-cLA (NO₂-cLA). These species mediate the post-translational modification (PTM) of proteins via reversible Michael addition with nucleophilic amino acids. The PTM of critical target proteins by electrophilic lipids has been described as a sensing mechanism that regulates adaptive cellular responses, but little is known about the endogenous generation of fatty acid nitroalkenes and their metabolites. We report that healthy humans consuming ¹⁵N-labeled NO₃⁻ or NO₂⁻, with and without cLA supplementation, produce ¹⁵NO₂-cLA and corresponding metabolites that are detected in plasma and urine. These data support that the dietary constituents NO₃⁻, NO₂⁻ and cLA promote the further generation of secondary electrophilic lipid products that are absorbed into the circulation at concentrations sufficient to exert systemic effects before being catabolized or excreted.

Graphical abstract

Introduction

Fatty acid nitroalkene derivatives (NO₂-FA) are detected in species as diverse as plants, insects and humans [1–3], but little is known of how NO₂-FA are endogenously generated. NO₂-FA display signaling actions that expands both the pharmacokinetics and scope of nitric oxide (^·NO) signaling events to include actions beyond the activation of guanylate cyclase-dependent cGMP production [4–7]. Thus, it is important to better understand both the mechanisms of formation and actions of these lipid signaling mediators. In addition to previously described biomolecule nitration mechanisms [8, 9], metal center-independent reactions of ^·NO and nitrite (NO₂⁻) at neutral pH yield both symmetric and asymmetric dinitrogen trioxide (N₂O₃), that in turn mediates the concerted S-nitrosation of thiols and nitration of unsaturated fatty acid conjugated dienes [10]. We now report that the reactions of these oxides of nitrogen support substantial NO₂-FA generation in humans, the levels of which can be altered by dietary supplementation.

The unique electrophilic character of fatty acid nitroalkenes promotes the reversible Michael addition of susceptible cysteines in a limited number of transcriptional regulatory factors and enzymes [11]. This post-translational modification of functionally-significant hyperreactive thiols induces pleiotropic responses that include the cGMP-independent activation of transcriptional programs regulated by nuclear factor (erythroid-derived-2)-like 2 (Nrf2) and nuclear lipid receptors, as well as inhibition of pro-inflammatory signaling by nuclear factor kappa-β (NF-kβ) [12–15]. Herein, the mass spectrometry analysis of healthy human urine and plasma at different times after oral consumption of ¹⁵N-labeled stable isotopes of NO₂⁻ and NO₃⁻ and conjugated linoleic acid (cLA) reveals substantial NO₂-FA generation at concentrations capable of mediating metabolic and inflammatory signaling responses.

Materials and Methods

Study Population

This study was performed with University of Pittsburgh IRB (PRO11120134) and FDA IND (#115926) approval. Informed consent was obtained from all subjects before participation. Subjects who had a normal blood pressure (BP) defined as systolic BP≤130 and diastolic BP≤85 mm Hg were recruited from university/community advertisements. Subjects meeting any of the following were excluded: 1) Positive urine pregnancy test; 2) Recent addition/change in birth control dosing (pills, shot or intrauterine device); 3) Concurrent use of medications affecting glucose or lipid metabolism; 4) Current use of BP medications regardless of BP control; 5) Current use of PD5 inhibitors or organic nitrates; 6) Not stable on treatments for the prior three months or not planning to remain on current dose of medications for birth control, etc.; 7) Chronic mental health or medical conditions including diabetes, obesity syndromes, liver or kidney disease; 8) Smoker. Healthy volunteers fasted overnight (10–12 h) and were randomized to receive either Na¹⁵NO₂ (20 mg) or Na¹⁵NO₃ (1 gm). Following ~7 day washout period, volunteers were crossed over to receive the other nitrogen oxide (NOx) (Trial I). In Trial II, cLA (3 gm) was given concurrently with both NOx (Fig. 1). Blood was drawn at baseline, 0.5, 1, 2, 3, 6, and 24 h after oral administration of NOx ± cLA. Urine samples were collected at baseline, 6, and 24 h for Trial II only. Vital signs including blood and mean arterial pressure, respiratory rate, heart rate, and methemoglobin % (noninvasively by co-oximetry) were monitored as a safety precaution and remained within normal ranges for the duration of the trials.

Figure 1.

Trial I and II study design. Trial I: volunteers were randomized to receive either ¹⁵NO₂⁻ (20 mg) or ¹⁵NO₃⁻ (1 gm) and blood samples were collected at t = 0, 0.5, 1, 2, 3, 6, and 24 h. After a 7 day washout period, volunteers returned to receive the other oxide of nitrogen. In Trial II, the same ¹⁵NO₂⁻ or ¹⁵NO₃⁻ administration and blood collection protocol was implemented, along with cLA (3 gm) supplementation and urine collection at t = 0, 6, and 24 h.

Materials

Total Lean™ cLA capsules (1 gm) were purchased from General Nutrition Company (Pittsburgh, PA). The safflower oil-derived capsules are a racemic mixture of (9Z,11E)-cLA and (10E,12Z)-cLA. 2,3-diaminonaphthalene (DAN) and 4-phenyl-1,2,4-triazoline-3,5-dione (PTAD) were purchased from Sigma-Aldrich (St. Louis, MO). (9Z,11E)-cLA was purchased from Nu-Chek Prep (Elysian, MN) and (11Z,13E)-cLA acid standard was purchased from Matreya (Pleasant Gap, PA). [¹³C₁₈]-NO₂-OA internal standard and NO₂-cLA standard were synthesized as described previously [3, 16]. MS grade solvents for liquid chromatography electrospray ionization tandem mass spectrometry (LC-MS/MS) were purchased from Burdick and Jackson (Muskegon, MI). Solvents for extractions and solid phase extraction (SPE) columns (C18 reversed phase; 500 mg, 6 mL capacity) were purchased from Thermo Scientific (Pittsburgh, PA).

¹⁵NO₃⁻ and ¹⁵NO₂⁻ Capsule Formulation

Stable isotope ¹⁵N-labeled sodium nitrite (10 mg) and sodium nitrate (500 mg) capsules were prepared by the NIH Pharmaceutical Development Section. The ¹⁵N isotopes were purchased from Cambridge Isotope Laboratories (Andover, MA). Each isotope was milled separately to provide uniformity in particle size with the needed excipients to ensure uniform powder mixture content and controlled disintegration and dissolution characteristics.

¹⁵Nitrite was especially difficult because of the variation in particle size and difficulty in milling. A mini blender was used initially followed by passage through a #40 mesh screen. After this, microcrystalline cellulose, pregelatinized starch, sodium starch glycolate, crospovidone and colloidal silicon dioxide were added sequentially. The mixture was milled on a Waring blender at low speed for 2 min before passing the through the #40 mesh screen 4 times. This produced an approximate particle size of 425 micrometers. ¹⁵Nitrate was mixed with all the same excipients used for the nitrites, but it did not require the same degree of mechanical milling and were passed through a #60 mesh screen. This produced an approximate particle size of 250 micrometers. Uniformity of capsule strength was affected by use of a mini cap-100 capsule filler making full fill capsules of each salt form. The quality and stability of the capsules were measured by stability indicating HPLC assays. All operations were performed in a GMP pharmaceutical manufacturing facility with chemical analysis preformed on site.

Lipid extraction

Plasma samples collected at each time point were extracted in triplicate using a modified Bligh-Dyer with interim vortexing. Samples were centrifuged at 2800 × g for 5 min. The top layer (aqueous) was transferred to a clean vial, [¹³C₁₈]-NO₂-OA internal standard (10 µL of 1 µM) was added, samples (5.5 mL) were vortexed, and placed on ice for 15 min. HgCl₂ (10 mM) was added and the samples were incubated for 30 min at 37°C to release adducted nitro-fatty acids. A 10% w/v sulfanilamide solution (100 µL) was then added to inhibit artifactual nitration. The aqueous layer was extracted with 1 M formic acid:isopropanol:hexanes (2:20:30, vol/vol/vol) followed by an additional 1 mL of hexanes, vortexing, and centrifugation. The upper organic phase was dried and reconstituted in 100 µL methanol before MS analysis. This extraction procedure allowed for measurement of NO₂-cLA that was non-covalently bound to albumin and other proteins as well as the NO₂-cLA that was covalently adducted to small thiols and proteins. These species were found in the protein precipitate interface and aqueous layer that form during the extraction procedure. The extraction of NO₂-cLA and its metabolites, including Cys-NO₂-cLA, from urine was performed and normalized to creatinine (Cre) as previously described [3, 17].

LC-MS/MS

A CTC PAL autosampler (Leap Technologies, Carrboro, NC) and a Shimadzu LC-20AD pump (Columbia, MD) coupled to an AB Sciex (Framingham, MA) 5000 triple quadrupole mass spectrometer was used for the quantification of modified fatty acids. Samples (10 µL) were separated on a reversed phase HPLC column (Luna C18(2), 5µ, 2 × 100 mm, Phenomenex, Torrence, CA) at a 0.75 mL/min flow rate using a gradient solvent system consisting of water containing 0.1% acetic acid (solvent A) and acetonitrile containing 0.1% acetic acid (solvent B). Samples were applied to the column at 40% B (0.5 min) and eluted with a linear increase in solvent B (40–100%) over 10 min. This was followed by a wash using 100% B for 2 min. The gradient then returned to starting conditions at 40% B for 3 min. MS analyses used electrospray ionization in the negative ion mode with the collision gas set at 5 units, curtain gas 50 units, ion source gas #1 55 units and #2 50 units, ion spray voltage −4500 V, and temperature 650°C. The declustering potential was −60, entrance potential −10, collision energy −35, and the collision exit potential −3. MRM was used for sample analysis and quantification of lipids showing loss of a nitro group (m/z 46) upon collision-induced dissociation. A standard curve using synthetic NO₂-cLA and [¹³C₁₈]-NO₂-OA internal standard allowed for the quantification of endogenous products.

Nitrite determination

Total plasma and urine NO₂⁻ was measured using the triiodide method as previously described [18]. To determine the contribution of ¹⁴NO₂⁻ and ¹⁵NO₂⁻ to these measurements, plasma and urine was reacted with diaminonaphthalene (DAN) and the products, ¹⁴N-napthotriazole (¹⁴NT) and ¹⁵N-napthotriazole (¹⁵NT) were measured using LC-MS/MS. A calibration curve using a Na¹⁴NO₂ standard undergoing the same sample preparation was used for quantification. Additionally, the “bleed over” of ¹⁴N into the ¹⁵N transition has been accounted for in the calculations of the ¹⁵N contribution to total nitrite. For samples, 100 µL of plasma or urine was deproteinized with 1 vol of cold acetonitrile and centrifuged (5000×g, 5 min). Supernatants were adjusted to pH 6 with 5 µL of 0.62 M HCl, then 10 µL of 0.5 mg ml⁻¹ DAN in 0.62 M HCl was added, samples incubated at RT for 15 min, and reactions were neutralized with 10 µL of 1.4 M NaOH. The samples (10 µL injection) containing ¹⁴NT and ¹⁵NT were separated on a reversed phase HPLC column (Phenomenex Gemini C18, 3 µ, 2 × 20 mm) at a 0.65 mL min⁻¹ flow rate. Samples were applied to the column at 5% B (0.3 min) and eluted with a linear increase in solvent B (5–100%) over 3.5 min. This was followed by a wash using 100% B for 1 min. The gradient was returned to starting conditions with 5% B for 0.5 min. MS analyses were conducted using electrospray ionization in positive ion mode with the collision gas set at 5 units, curtain gas 40 units, ion source gas #1 at 55 units and 2 at 50 units, ion spray voltage −5500 V, and temperature at 650°C. The declustering potential was set to 70, entrance potential 5, collision energy 35, and the collision exit potential 5. MRM transitions for ¹⁴NT (170→115) and ¹⁵NT (171→115) were used for analyses.

cLA determination

cLA detection and characterization was performed by LC-MS/MS in negative ion mode after derivatization with 10 mM 4-phenyl-1,2,4 triazoline-3,5 dione (PTAD) to form the respective Diels-Alder adducts. As an internal standard, 1 µM of “non-natural” (11Z,13E)-cLA was added. The following transitions were used for detecting derivatized cLA positional isomers: (10E,12Z)-cLA (454→182, 454→205, 454→220, 454→238) and (9Z,11E)-cLA (454→168, 454→191, 454→206, 454→224). Quantification was performed using standard curves of PTAD-cLA (0–10 µM) and (11Z,13E)-cLA (400 nM).

Statistical analysis

Plasma and urine extractions were performed on samples collected from 5 subjects, with analyses done in duplicate or triplicate for each subject per time point. Data is represented as mean ± SEM for plasma cLA and overall ¹⁴NO₂-cLA levels after ¹⁵NO₂⁻ and ¹⁵NO₃⁻ supplementation. The ¹⁴N/¹⁵N contributions to total nitrite were determined as percentages of the total nitrite concentration for each time point in plasma and urine. Box plots represent the concentration of ¹⁴NO₂-cLA and ¹⁵NO₂-cLA in plasma and urine over time; with median and range reported. Plasma levels are reported as concentration and urine levels are reported as pmol mg⁻¹ creatinine.

Results

NO₂-cLA formation upon oral ingestion of NO₃⁻ and NO₂⁻

In Trial I, healthy volunteers (3 males, 2 females, mean age 29 ± 7 yr, other clinical indices in Supplementary Table 1) fasted overnight prior to supplement randomization with either oral ¹⁵NO₂⁻ (20 mg) or ¹⁵NO₃⁻ (1 gm). Blood was drawn at baseline, 0.5, 1, 2, 3, 6, and 24 h post-supplementation (Fig. 1). After a 7 day washout, the same subjects received the opposite NOx and blood was drawn as before. Plasma lipids were extracted and both ¹⁴NO₂-cLA (endogenous) and ¹⁵NO₂-cLA (formed from ¹⁵NOx) were measured and quantified by LC-MS/MS (Fig. 2A). Endogenous ¹⁴NO₂-cLA was present in the plasma of all volunteers and did not vary significantly over the 24 h sampling period after both ¹⁵NO₂⁻ and ¹⁵NO₃⁻ supplementation (¹⁵NO₂⁻: 1.0 ± 0.3 nM and ¹⁵NO₃⁻: 2.1 ± 0.7 nM, Fig. 2B). ¹⁵NO₂-cLA was not detectable over the sampling period after ¹⁵NO₂⁻ supplementation and was only detectable 24 h after ¹⁵NO₃⁻ consumption (0.14 nM median, range = 0 to 25.3 nM, Fig. 2C).

Figure 2.

Dietary supplementation of nitrite, nitrate and conjugated linoleic acid supports NO₂-cLA formation detected in plasma. Oral ¹⁵NO₂⁻ (20 mg) or ¹⁵NO₃⁻ (1 gm) was ingested ± cLA (3 gm). (A) LC-MS/MS chromatograms of plasma lipid extract show endogenous ¹⁴NO₂-cLA (MRM, 326/46) present at each time point and ¹⁵NO₂-cLA (MRM, 327/47) formed following ¹⁵NOx supplementation for ¹⁵NO₂⁻ at 2 h and ¹⁵NO₃⁻ at 24 h. (B) ¹⁴NO₂-cLA remains at basal levels over 24 h after ¹⁵NO₂⁻ (1.0 ± 0.3 nM) or ¹⁵NO₃⁻ supplementation (2.1 ± 0.7 nM). (C) Only ¹⁵NO₃⁻ administration led to detectable ¹⁵NO₂-cLA at 24 h (range = 0 to 25.3 nM, 0.14 nM median). (D) ¹⁴NO₂-cLA remained at basal levels after cLA supplementation (+¹⁵NO₂⁻, 2.6 ± 0.8 nM) and (+¹⁵NO₃⁻, 3.1 ± 0.9 nM). (E) ¹⁵NO₂⁻ + cLA resulted in early and significant increases from 1–6 h in ¹⁵NO₂-cLA (1 h range = 0 to 25.7 nM, 2.5 nM median and 6 h range = 0 to 11.5 nM, 3.3 nM median). Following ¹⁵NO₃⁻ + cLA, ¹⁵NO₂-cLA was detectable at 2 h and highest at 24 h (range = 0 to 9.9 nM, 1.6 nM median). n=5 volunteers.

NO₂-cLA generation is promoted by cLA consumption

The same 5 healthy volunteers returned 3 months later for Trial II, identical in design with respect to oral ¹⁵NOx administration and the timing of blood sampling. The protocol also included a) the additional consumption of cLA (3 gm) in concert with each ¹⁵NOx and b) the collection of urine at baseline, 6 and 24 h after ¹⁵NOx + cLA administration. Similar concentrations of endogenous ¹⁴NO₂-cLA were detectable over the 24 h sampling period and not affected by ¹⁵NOx + cLA supplementation (¹⁵NO₂⁻ + cLA : 2.6 ± 0.8 nM and ¹⁵NO₃⁻ + cLA L 3.1 ± 0.9 nM, Fig. 2D). The modest decrease over time in ¹⁴NO₂-cLA may have been due to volunteers fasting prior to ¹⁵NOx + cLA supplementation during the first 6 h of blood collection. Furthermore, volunteers were instructed to consume an evening meal and beverage low in NOx and cLA before overnight fasting prior to 24 h sampling. ¹⁵NO₂-cLA generation became detectable at 1 h (2.5 nM median, range = 0 to 25.7 nM) after ¹⁵NO₂⁻ + cLA supplementation and remained elevated until 6 h (3.3 nM median, range = 0 to 11.5 nM). In contrast, ¹⁵NO₃⁻ + cLA mediated ¹⁵NO₂-cLA generation became detectable at 2 h and was the greatest at 24 h (1.6 nM median, range = 0 to 9.9 nM), compared to baseline (Fig. 2E).

Plasma levels of free cLA and NO₂⁻

The consumption of cLA in Trial II supported significant increases in plasma ¹⁵NO₂-cLA levels from 1 to 6 h following ¹⁵NO₂⁻ + cLA administration (Fig. 2E). These changes corresponded with the timing of peak plasma concentrations of both (9Z,11E)-cLA and (10E,12Z)-cLA (Fig. 3A). The (9Z,11E)-cLA and (10E,12Z)-cLA isomers were quantified by LC-MS using a PTAD derivatization and the non-naturally occurring 11,13-cLA isomer as an internal standard (Fig. 3B). The (9Z,11E)-cLA isomer remained elevated above baseline between 0.5 and 6 h before decreasing back to baseline while the (10E,12Z)-cLA isomer, derived from the cLA supplement, was only detected from 2 to 6 h (Fig. 3A). The concentration of both isomers peaked at 3 h after ¹⁵NO₂⁻ + cLA administration with (9Z,11E)-cLA ranging from 530 to 1350 nM with an average of 858 ± 138 nM and (10E,12Z)-cLA ranging from 271 to 710 nM, with an average of 430.3 ± 77.8 nM (Fig. 3A). The levels of ¹⁵NO₂-cLA after ¹⁵NO₃⁻ + cLA administration were lower and detectable at later time points compared to those after ¹⁵NO₂⁻ + cLA supplementation. The lower levels of cLA at later time points may have contributed to the decreased ¹⁵NO₃⁻-derived ¹⁵NO₂-cLA formation (Fig. 2D and E).

Figure 3.

Plasma levels of free cLA and NO₂⁻. (A) Free (10E,12Z)-cLA and (9Z,11E)-cLA levels measured from plasma over time in Trial II. The mean represents a combination of values from ¹⁵NO₃⁻ and ¹⁵NO₂⁻ supplementation. (9Z,11E)-cLA is found endogenously whereas (10E, 12Z)-cLA is solely derived from the supplement. Both isomers of free cLA peaked at 3 h with (9Z,11E)-cLA levels ranging from 530 to 1350 nM with a mean of 858 ± 138 nM. (10E,12Z)-cLA ranging from 270 to 710 nM with a mean of 430 ± 77.8 nM. (B) Representative chromatograms of the PTAD-derivatized (9Z,11E)-cLA, (10E,12Z)-cLA, and (11Z,13E)-cLA internal standard in plasma. (C) The ¹⁴N and ¹⁵N contribution to total plasma NO₂⁻ concentration after ¹⁵NO₂⁻ or ¹⁵NO₃⁻ supplementation in Trial I based on the mean NO₂⁻ value (Fig S1A). (D) The ¹⁴N and ¹⁵N contribution to total plasma NO₂⁻ concentration after ¹⁵NO₂⁻ + cLA or ¹⁵NO₃⁻ + cLA supplementation in Trial II based on the mean NO₂⁻ value (Fig S1B). LC-MS/MS was used to differentiate between the contributions of ¹⁴NOx and ¹⁵NOx species to overall NO₂⁻ levels and the mean concentration for each time point was used for both (B) and (C). The mean NO₂⁻ concentration ± SEM can be found in Supplementary Fig. S1 for n=5 volunteers.

Plasma NO₂⁻ levels after ¹⁵NOx supplementation were determined for both trials using both chemiluminescence and tri-iodide-based reactions (Supplementary Fig S1A and B) [18, 19]. To distinguish between the contributions of endogenous and exogenously-administered NOx species to total NO₂⁻, 2,3-diaminonaphthalene (DAN) was reacted with NO₂⁻ and the reaction products ¹⁴N- and ¹⁵N-napthotriazole (NT) were differentiated using LC-MS/MS (Fig. 3C and D) [20]. The mean NO₂⁻ concentration for each time point (Supplementary Fig S1A and B) was used to calculate relative ¹⁴N and ¹⁵N contributions to total NO₂⁻ concentration. Plasma NO₂⁻ concentrations were greatest at 0.5 h post-¹⁵NO₂⁻ supplementation in both trials; however, total NO₂⁻ concentrations were greatest at 0.5 h in Trial I, where there was a larger contribution of ¹⁵NO₂⁻ to total NO₂⁻ (Fig. 3C). The peak plasma ¹⁵NO₂⁻ concentration at 0.5 h preceded the highest ¹⁵NO₂⁻ + cLA-dependent ¹⁵NO₂-cLA level, seen at 2 h (Fig. 2E). The NO₂⁻ concentrations after ¹⁵NO₃⁻ supplementation in Trials I and II peaked between 2 and 3 h with a slightly longer lag time when cLA was added (Fig. 3C and D). As expected, ¹⁵NO₃⁻ did not contribute as much to the net NO₂⁻ concentration as direct administration of ¹⁵NO₂⁻ (Fig. 3C and D). Notably, dietary supplementation with ¹⁵NO₃⁻ and ¹⁵NO₂⁻ caused an increase of all detectable endogenous NOx species, thereby increasing ¹⁴NO₂⁻ levels that paralleled concentrations of ¹⁵NOx species. This is most evident with ¹⁵NO₃⁻ supplementation in both Trial I and II (Fig. 3C ad D).

Urinary NO₂-cLA and NO₂⁻ concentrations

Urine was collected in Trial II at baseline, 6 and 24 h following ¹⁵NO₂⁻ or ¹⁵NO₃⁻ + cLA supplementation. Endogenous urinary mean ¹⁴NO₂-cLA concentrations modestly decreased after ¹⁵NO₂⁻ + cLA or ¹⁵NO₃⁻ + cLA supplementation over the 24 h period (Fig. 4A). The concentration of urinary ¹⁵NO₂-cLA was greatest 6 h after ¹⁵NO₂⁻ + cLA consumption (10.7 pmol/mg creatinine (Cre) median, range = 0 to 55.0 pmol/mg Cre) and 24 h after ¹⁵NO₃⁻ + cLA consumption (12.5 pmol/mg Cre median, range = 3.8 to 44.3 pmol/mg Cre, Fig. 4B). Urinary NO₂⁻ was also measured via triiodide chemiluminescence (Supplementary Fig S1C) [18, 19] and the mean NO₂⁻ concentration for each time point was used to calculate relative ¹⁴N and ¹⁵N contributions to total NO₂⁻ concentration (Fig. 4C and Supplementary Fig. S1C). This paralleled ¹⁵NO₂-cLA levels, peaking at 6 h with an equimolar contribution from ¹⁴NOx and ¹⁵NOx species (Fig. 4C). Total urinary NO₂⁻ concentrations decreased after ¹⁵NO₃⁻ + cLA administration over the 24 h time period, with little contribution from ¹⁵NO₃⁻ (Fig. 4C). Urinary β-oxidation products and cysteine-adducts (Supplementary Fig. S2 and S3) of ¹⁴NO₂-cLA and ¹⁵NO₂-cLA were also measured, confirming the electrophilic reactivity of the parent molecules and metabolites [3]. Cysteine-NO₂-cLA adducts were detected in urine, with the ion intensity of Cys-¹⁵NO₂-cLA metabolites being ~10-fold greater than Cys-¹⁴NO₂-cLA metabolites at 24 h (Supplementary Fig. S3).

Figure 4.

Urinary NO₂-cLA and NO₂⁻ concentrations. Urine was collected at 0, 6, and 24 h following oral ¹⁵NO₂⁻+ cLA or ¹⁵NO₃⁻ + cLA administration. (A) ¹⁴NO₂-cLA levels in urine are highest at baseline and decrease over time. (B) Urinary ¹⁵NO₂-cLA was greatest 6 h after ¹⁵NO₂⁻ + cLA consumption (range = 0 to 55.0 pmol/mg Cre, 10.7 pmol/mg Cre median) and 24 h after ¹⁵NO₃⁻ + cLA consumption (range = 3.8 to 44.3 pmol/mg Cre, 12.5 pmol/mg Cre median). (C) The ¹⁴N and ¹⁵N contribution to total urinary NO₂⁻ concentration after ¹⁵NO₂⁻ + cLA or ¹⁵NO₃⁻ + cLA supplementation in Trial II based on the mean NO₂⁻ value (Supplementary Fig. S1C). n=5 volunteers.

Discussion

Intrinsic to the metabolism of NO₃⁻ and NO₂⁻ is the formation of the nitrogen dioxide radical (^·NO₂). This species is formed by metalloprotein-catalyzed oxidation of NO₂⁻, the autoxidation of ^·NO, the protonation of NO₂⁻ to N₂O₃ followed by its disproportionation, decomposition of nitrous acid, and the homolytic scission of two key products of ^·NO and superoxide (O₂ ^·−) reaction, peroxynitrous acid and nitrosoperoxocarbonate [21–23]. Unsaturated fatty acids having conjugated double bonds, such as cLA, are highly reactive with ^·NO₂, a reaction that occurs much more readily than tyrosine nitration [24]. ^·NO₂-induced nitration is also several orders of magnitude greater for the conjugated diene system of cLA compared to the bis-allylic diene system found in linoleic and oleic acids due to resonance stabilization of the radical product formed during ^·NO₂ radical addition [25]. This makes cLA the preferable substrate for nitration despite the fact that conjugated dienes are present in ~100-fold lower concentrations than bis-allylic unsaturated fatty acids in plasma [26]. Dietary sources of cLA include dairy products and meat and cLA is also a product of enterosalivary bacterial Δ⁹-desaturase activity [27]. The predominant endogenous cLA isomer is 9-cis, 11-trans-cLA, whereas plant-derived oils contain a racemic mixture of 9-cis, 11-trans and 10-trans, 12-cis-cLA [28]. The consumption of cLA is linked with beneficial clinical responses, including the reduction of obesity-related body mass index, improved insulin sensitivity and attenuation of airway hyperactivity, all of which we propose are due in part to the formation and signaling actions of NO₂-cLA [29–32].

NO₂-cLA is found basally in the plasma and urine of healthy humans and its concentrations increase in organelles and tissues undergoing metabolic and inflammatory stress [3, 16] (Fig. 5). Electrophilic nitroalkenes such as NO₂-cLA induce kinetically rapid posttranslational modification of nucleophilic cysteines of key transcriptional regulatory proteins and inhibit the activity of functionally-significant electrophile-sensitive enzymes [11, 33, 34]. Previously reported targets of nitroalkylation include the Cys285 of PPARγ (peroxisome proliferator activator receptor γ) [35], Cys273 and Cys288 of the Nrf2 regulatory protein Keap1 [Kelch-like ECH-associated protein 1 (Keap1)/regulator of nuclear factor (erythroid-derived-2)-like 2 (Nrf2)] [12], heat shock proteins[12], and Cys38 of the p65 subunit of nuclear factor κB (NF-κB) [36] (Fig. 5). Consequently, synthetic homologs of endogenous nitroalkenes induce human coronary artery endothelium to alter the expression of over 400 genes related to metabolism, tissue repair and inflammatory responses [12]. Nitroalkylation also directly inhibits the catalytic activity of enzymes including xanthine oxidoreductase [33], cyclooxygenase [37] and soluble epoxide hydrolase (Cys521) [34], reactions that result in significant beneficial physiological responses. These alkylation reactions are reversible, with the exception of xanthine oxidoreductase, and can be modulated by competing tissue nucleophiles such as cysteine, glutathione (GSH) and hydrogen sulfide [11, 38]. Thus, lipid nitroalkene derivatives represent salutary signaling molecules that both modulate inflammation and serve as sensitive biomarkers for tissue nitrative and nitrosative reactions, much like 8-iso-prostaglandin F_2α is a marker for oxidative reactions.

Figure 5.

Formation and signaling of NO₂-cLA. Nitrate (NO₃⁻) and nitrite (NO₂⁻) are dietary sources of nitrogen dioxide (•NO₂). Nitrate is reduced to nitrite by entero-salivary bacteria. Nitrite, combined with the low pH of the stomach favors •NO₂ formation via nitrous acid (HNO₂) generation. Various oxides of nitrogen can form from the decomposition of HNO₂ in the gut, including ^•NO₂. In inflammation, ^•NO₂ can arise from the protonation of NO₂⁻ to nitrous acid (HNO₂) or NO₂⁻ oxidation by heme peroxidases. Another significant mechanism of ^•NO₂ formation involves peroxynitrite (ONOO⁻), peroxynitrous acid (ONOOH), which are formed through a reaction of ^•NO and superoxide (O₂^•−). These species can readily diffuse through the membrane to mediate unsaturated fatty acid nitration and oxidation via homolysis of ONOOH to ^•NO₂ and ^•OH. Peroxynitrite also reacts with CO₂ to form nitrosoperoxocarbonate (ONOOCO₂) and like HNO₂, this compound can undergo homolytic scission to form ^•NO₂. Nitrogen dioxide reacts with the π electrons of alkenes via an addition reaction and a reaction with a second ^•NO₂ results in the reformation of the double bond. Conjugated diene containing PUFAs, such as cLA, are especially susceptible to nitration, as opposed to methylene-interrupted species. The endogenous production and exogenous administration of electrophilic fatty acids targets multiple redox-sensing transcriptional regulators. It has been previously demonstrated that nitroalkenes (a) putatively bind HSP70, releasing HSF-1 and thus driving HSF-1-dependent gene transcription; (b) covalently adduct Keap1, causing dissociation from and translocation of Nrf2 to induce ARE gene transcription, and (c) modify the p65 subunit of NF-κB, sustaining inhibition by IκB and blocking p50/p65-dependent gene transcription. In the nucleus, (d) nitroalkenes covalently bind and act as partial PPARγ agonists, stimulating gene transcription.

It was hypothesized that dietary NOx would be limiting in NO₂-cLA formation; however, in the absence of cLA supplementation in combination with overnight fasting, plasma ¹⁵NO₂-cLA concentrations were barely detectable and only became significantly elevated 24 h after ¹⁵NO₃⁻ consumption (Fig. 2C). With cLA supplementation, greater concentrations of ¹⁵NO₂-cLA were detected earlier after both ¹⁵NO₂⁻ and ¹⁵NO₃⁻ supplementation (Fig. 2E). Plasma ¹⁵NO₂-cLA concentrations in the ¹⁵NO₂⁻-supplemented cohort paralleled plasma cLA concentrations (Fig. 2E and 3A). Maximum levels of plasma ¹⁵NO₂⁻ after dietary intake occurred after 30 min (Fig. 3C), thus preceding ¹⁵NO₂-cLA formation. The plasma half-life of NO₂⁻ is 5–6 h, also tracking with the decrease of ¹⁵NO₂-cLA in plasma between 6 and 24 h [39, 40]. The ¹⁴NO₂⁻ levels were increased in plasma after ¹⁵NOx supplementation (Fig. 3C and D). It is possible that the 1 gm of orally-consumed ¹⁵NO₃⁻ exchanged with and increased concentrations of endogenous ¹⁴NOx counterions available for forming nitrating species.

The detection and quantification of NO₂-cLA and other electrophilic fatty acids is challenged by their facile reaction with biological nucleophiles (e.g., cysteine, GSH and protein thiols and histidine). Also, electrophilic fatty acids become esterified in complex lipids [2], are enzymatically reduced to non-electrophilic nitroalkane derivatives by prostaglandin reductase 1 [41], can be further oxidized and readily undergo mitochondrial β-oxidation [11, 42]. Evidence for these reactions occurring in humans was reflected by the detection of urinary ¹⁵NO₂-cLA β-oxidation products and corresponding cysteine-NO₂-cLA and cysteine-NO₂-cLA β-oxidation products detected after ¹⁵NOx + cLA supplementation (Supplementary Fig S2 and S3). From a pharmacological perspective, some of these transformations retain intrinsic signaling competency via preservation of their electrophilic moiety and others do not. The relative signaling activities of electrophilic β-oxidation products, the extents of NO₂-fatty acid esterification into glycerolipids and the half-lives of nitro-fatty acids esterified into various complex lipid and tissue compartments remain to be fully characterized.

In summary, the formation of NO₂-cLA and its metabolites after dietary nitrogen oxide and cLA consumption reveals that diet, digestion and metabolism yield a spectrum of absorbable electrophilic metabolites that are detectable in both plasma and urine. Model systems show that these products mediate distinctive downstream post-translational protein modifications, altered gene expression and signaling responses at nM concentrations. Phase 1 safety studies using FDA-approved nitro-oleic acid reveal a wide therapeutic window for NO₂-fatty acids (clinicaltrials.gov ID: NCT02127190). Significant metabolic and anti-inflammatory responses are induced by exogenously-administered fatty acid nitroalkenes in murine models after administration of doses giving 5–10 nM plasma concentrations [43, 44]. Herein, human plasma NO₂-cLA concentrations after ¹⁵NOx + cLA supplementation increased from an average basal level of 3 nM to 10 nM. These data support the concept that the consumption of dietary constituents high in NO₃⁻ and cLA, such as leafy vegetables, dairy products and meat, can mediate physiological responses via the generation of NO₂-cLA.

Highlights.

• ¹⁵N-labeled NO₃⁻/NO₂⁻ supplementation results in ¹⁵NO₂-cLA formation

• NO₂-cLA formation is dependent on levels of free cLA

• NO₂-cLA concentration can be modulated by dietary components

• Detectable NO₂-cLA concentrations are sufficient to exert systemic effects

Abbreviations

PTAD

4-phenyl-1,2,4 triazoline-3,5 dione

[¹³C₁₈]-NO₂-OA

[¹³C₁₈]-nitro-oleic acid

¹⁴NO₂-cLA

¹⁴nitro-conjugated linoleic acid

¹⁵NO₂-cLA

¹⁵nitro-conjugated linoleic acid

cLA

conjugated linoleic acid

cGMP

cyclic GMP

DAN

diaminonaphthalene

GSH

glutathione

HPLC

high performance liquid chromatography

KEAP1

Kelch-like ECH-associated protein 1

LC-MS/MS

liquid chromatography tandem mass spectrometry

NT

N-napthotriazole

NO₂-FA

nitro-fatty acid

Nrf2

Nuclear factor (erythroid-derived 2)-like 2

NF-κB

nuclear factor kappa-light-chain-enhancer of activated B cells

PPARγ

Peroxisome proliferator-activated receptor gamma

PTM

post-translational modification

RT

retention time

SPE

solid phase extraction