Stable isotope-assisted LC-MS/MS monitoring of glyceryl trinitrate bioactivation in a cell culture model of nitrate tolerance

Abstract

The nitric oxide (NO) metabolites nitrite (NO₂⁻) and nitrate (NO₃⁻) can be quantified as an endpoint of endothelial function. We developed a LC-MS/MS method of measuring nitrite and nitrate isotopologues, which has a lower limit of quantification (LLOQ) of 1 nM. This method allows for isotopic labeling to differentiate newly formed nitrite and nitrate from nanomolar to micromolar background levels of nitrite and nitrate in biological matrices. This method utilizes 2,3-diaminonaphthalene (DAN) derivatization, which reacts with nitrite under acidic conditions to produce 2,3-naphthotriazole (NAT). NAT was chromatographically separated on a Shimadzu LC System with an Agilent Extend-C₁₈ 5 μm 2.1 × 150 mm column and detected using a multiple reaction monitoring (MRM) method on an ABSciex 3200 QTRAP mass spectrometer operated in positive mode. Mass spectrometry allows for the quantification of ¹⁴N-NAT (m/z 170.1) and ¹⁵N-NAT (m/z 171.1). Both nitrite and nitrate demonstrated a linear detector response (1 nM – 10 μM, 1 nM – 100 nM, respectively), and were unaffected by common interferences (Dulbecco's Modified Eagle Medium (DMEM), fetal bovine serum (FBS), phenol red, and NADPH). This method requires minimal sample preparation, making it ideal for most biological applications. We applied this method to develop a cell culture model to study the development of nitrate tolerance in human endothelial cells (EA.hy926).

1. Introduction

Nitric oxide (NO) is a signaling molecule and free radical that plays a diverse role in cardiovascular, immune, and nervous systems [1][2]. NO is produced from L-arginine by NO synthase (NOS) in mammals [1], yet it can also be produced via the reduction of nitrite (NO₂⁻) and nitrate (NO₃⁻) by xanthine oxidase (XO) [3; 4; 5; 6; 7]. Nitrite and nitrate are present in the environment and diet. NO is released by cells in the picomolar to nanomolar range [8], and has an estimated in vivo half-life of 3-4 seconds in human blood [9]. It is therefore very difficult to accurately measure, so its nitrite and nitrate metabolites are often measured as surrogates [10]. Measurement of nitrite and nitrate is the most suitable and practical method to assess NO synthesis in vivo [11]. Given that nitrite and nitrate can be reduced to NO by XO, nitrite and nitrate can be viewed as bioavailable storage pools for NO [12].

Nitrate tolerance

Glyceryl trinitrate (GTN), more commonly known as nitroglycerin, is used for the treatment of angina pectoris, coronary artery disease, acute myocardial infarction, hypertension, and congestive heart failure. GTN is a prodrug that must be bioactivated in endothelial cells to produce NO (Figure 1), which activates soluble guanylyl cyclase (sGC) to release the second messenger cyclic guanosine monophosphate (cGMP), stimulating the relaxation of smooth muscle cells [13; 14]. Continuous treatment with GTN results in nitrate tolerance after only a few weeks, which limits the long-term use and benefits of treatment [15]. Currently, nitrate tolerance is managed by implementing nitrate-free intervals, but this strategy could become obsolete when co-therapies that effectively prevent nitrate tolerance become available. The cause of nitrate tolerance and the mechanism of GTN bioactivation remain relatively unknown. In the present study we measure nitrite and nitrate as endpoints of GTN bioactivation in endothelial cells. Our intent is to determine whether cell culture and the measurement of nitrite and nitrate isotopologues can be used as an appropriate model to determine the mechanism of GTN bioactivation and the cause of nitrate tolerance.

Figure 1.

Glyceryl Trinitrate (GTN) is denitrified by xanthine oxidase (XO) in endothelial cells to produce 1,2-glyceryl dinitrate (1,2-GDN) and NO₂⁻, which is reduced into NO and subsequently non-enzymatically oxidized to NO₂⁻ and NO₃⁻.

Quantification of nitrite and nitrate

There are over 200 analytical methods for the measurement of nitrite and nitrate, yet a small portion of these can be applied to human biological fluids [11]. Current methods that are used for nitrite and nitrate analysis include fluorescence, chemiluminescence, high performance liquid chromatography (HPLC), capillary electrophoresis, colorimetric and ultraviolet (UV) spectrophotometric methods, and gas chromatography-mass spectrometry (GC-MS). The current methods have been extensively reviewed [8; 11; 16; 17; 18]. Mass spectrometry is the only analytical technique that is able to measure isotopologues of nitrite and nitrate, and is an indispensable analytical tool for reliable quantitative analysis of nitrite and nitrate [8]. In the present study, we modify a method of measuring nitrite with LC-MS/MS [19; 20], and expand upon it to include quantification of nitrate. We utilized the well-established 2,3-diaminonaphthalene (DAN) derivatization, which reacts with nitrite under acidic conditions to produce the highly fluorescent 2,3-naphthotriazole (NAT)[21](Figure 2). Measuring NAT with LC-MS/MS, instead of by fluorescence, allows for the use of isotope labeling which reduces the impact of high background levels of nitrite and nitrate in biological matrices.

Figure 2.

NO₂⁻ reacts with 2,3-diaminonaphthalene (DAN) under acidic conditions to produce 2,3-naphthotriazole (NAT).

This LC-MS/MS method is a significant contribution to the field because it is a sensitive and reproducible method for quantifying nitrite and nitrate isotopologues as metabolic products of GTN. This method is used to investigate the mechanism of nitrate tolerance in human endothelial cells, which is made possible with the use of ¹⁵N₃-GTN.

2. Materials and Methods

2.1 Chemicals

2,3-diaminonaphthalene (DAN) was procured from TCI America (Portland, OR). ¹⁵N_3-glyceryl trinitate (GTN) in acetonitrile (¹⁵N, 98%+), sodium ¹⁵N-nitrite (¹⁵N, 98%+), and sodium ¹⁵N-nitrate (¹⁵N, 98%+) were from Cambridge Isotope Laboratories (Tewksbery, MA). GTN, NADPH, sodium hydroxide, and nitrate reductase were from Sigma Aldrich (St. Louis, MO). Dulbecco's Modified Eagle Medium (DMEM), Hank's Buffered Sodium Salt (HBSS), trypsin-EDTA, fetal bovine serum (FBS), penicillin, and streptomycin were procured from Invitrogen (Carlsbad, CA). Ammonium bicarbonate, LCMS-grade water and methanol were purchased from J.T Baker (Center Valley, PA).

2.2 Endothelial cell culture and treatments with glyceryl trinitrate (GTN)

EA.hy926 human hybrid endothelial cells (American Type Culture Collection; Manassas, VA) were maintained in culture in T75 flasks with DMEM medium containing 4.5 g/L glucose and 84 mg/L L-arginine, and supplemented with 10% FBS and 1% penicillin-streptomycin (PS) in a humidified incubator at 37°C with 5% CO₂. Endothelial cells were chosen because they have been previously shown to produce NO from nitrite, which caused changes in markers of intracellular nitrosation reactions [22]. Endothelial cells express xanthine oxidase, mitochondrial aldehyde dehydrogenase, and nitric oxide synthase, among others.

EA.hy926 cells were plated at a density of 5.0 × 10⁵ cells per well in 6-well plates with 2 mL medium following a trypsinization protocol. Cells were returned to the incubator for 24 hours before treatment. Nitrate tolerance was induced in cells using a modified method from Hinz and colleagues [23]. In short, cells were given an initial 20 μM GTN challenge in phenol red-free DMEM with 1% FBS, and incubated for 5 hours. Medium was aspirated, washed once with 2 mL HBSS, then treated for 60 minutes at four concentration levels of ¹⁵N-GTN (1 μM – 20 μM) in 1 mL phenol red-free DMEM with 1% FBS. Results were compared to control cells, which were treated with 5 μL of acetonitrile as a vehicle control. The entire final volume (1 mL) was transferred to Eppendorf tubes and immediately frozen at −80°C until analysis.

2.3 Enzymatic reduction of nitrate into nitrite

Nitrate was enzymatically converted to nitrite by incubating with nitrate reductase [24; 25]. The enzymatic method of reducing nitrate to nitrite is preferred to cadmium reduction because it is non-toxic, and offers high specificity, simplicity, rapidity, and complete conversion of nitrate into nitrite [24; 25]. One hundred μL of the cell medium, or ¹⁵N-nitrate standard, was incubated with 10 μL of 1 U/mL nitrate reductase and 10 μL of 120 μM NADPH and incubated at room temperature for 60 minutes. There was complete conversion of nitrate into nitrite, as evidenced by conversion of known amounts of ¹⁵N standards of sodium nitrite and sodium nitrate. After incubation, samples were immediately derivatized for nitrite analysis.

2.4 2,3-Diaminonaphthalene (DAN) derivatization

DAN reacts with nitrite under acidic conditions to form NAT [21], which is detectable with mass spectrometry. For mass spectrometric analysis of NAT, the protocol developed by Misko et al. [26] was adapted. No filtration is necessary because it was determined that our culture media is not a source of interference for the measurement of nitrite and nitrate. A 100 μL sample aliquot (culture medium, nitrite standards, or nitrate-derived standards) was incubated at 24°C for 60 minutes with 10 μL of 316 μM DAN in 50% EtOH with 0.62 M HCl. The reaction was quenched with 5 μL of 2.8 M NaOH. In alkaline conditions, NAT is stable for at least 24 hours [25]. The samples were centrifuged at 16,000 x g for 1 minute, then transferred to glass LCMS vials (Microsolv, Eatontown, NJ) and analyzed for ¹⁴N-NAT and ¹⁵N-NAT.

2.5 Detection of 2,3-naphthotriazole (NAT) with LC-MS/MS

Mass spectrometric detection of ¹⁴N-NAT and ¹⁵N-NAT was performed on a Shimadzu Prominence LC system coupled to an AB Sciex 3200 QTRAP operated in positive ion mode. An Agilent Extend-C₁₈ 5 μm 2.1 × 150 mm column was maintained at 40°C with a flow rate of 0.2 mL/min (0-13 minutes) and 0.4 mL/min (13-25 minutes). Mobile phase A was 5 mM aqueous NH₄HCO₃, mobile phase B was 100% methanol, and the needle wash solvent was 50% methanol. The total run time was 25 minutes, column eluent was directed to MS between 8-12 minutes only, and injections were 15 μL. The solvent gradient profile started at 5% B for 4 minutes, changed linearly to 50% B over 1 minute, was held at 50% B for 8 minutes, changed linearly to 95% over 3 minutes, was held at 95% B for 5 minutes, changed linearly to 5% B over 3 minutes, and remained at 5% B to equilibrate the column. Retention time was monitored, and no significant shifts were observed throughout sample sets. A multiple reaction monitoring (MRM) method monitored the following transitions over 100 ms for the analyte indicated: m/z 171.1 → 115.1 for ¹⁵N-NAT, m/z 159.1 → 159.1for DAN, and m/z 170.1 → 115.1 for ¹⁴N-NAT. The collision voltages were 33 V for all except DAN, which was 5 V. Declustering, entrance, and collision cell exit potentials were 51, 11.5 and 4 V, respectively. The IonSpray voltage was 5500 V and the ion source temperature was 550 °C.

¹⁴N-NAT and ¹⁵N-NAT concentrations were quantified by integrating the peaks in Analyst 1.5.1 to obtain peak area. Peak area was converted to concentration (nM) by using individual seven-point concentration curves of the analytes. Nitrate concentration was obtained after reduction with nitrate reductase.

2.6 Method validation and quality control

Seven-point concentration curves (1 nM – 10 μM) for sodium ¹⁵N-nitrite and sodium ¹⁵N-nitrate were made in combination and individually in milli-Q water, 200 mM HEPES, and DMEM with 1% FBS. These were used to calculate LLOQ, linear range, percent recovery in DMEM, and percent conversion of nitrate reduction to nitrite. Background levels of nitrate and nitrite were measured in water, HEPES, and DMEM. Three concentrations (10 nM, 100 nM, and 1000 nM) of nitrite and nitrate isotopologues were analyzed ten times in water and DMEM to determine precision and accuracy. For every day of analysis of cell culture or enzyme samples, a seven-point concentration curve of nitrite in water was run to assess instrument performance. If a drop in sensitivity was observed, the instrument was cleaned and calibrated before analysis of samples. All samples were randomized to prevent batch effect, yet nitrite and nitrate samples were paired and run in sequence.

2.7 Statistical analysis

Results were analyzed using a one-way ANOVA, followed by Tukey post hoc analysis with a P-value of < 0.05 indicating significance. All statistical analyses were performed using GraphPad Prism 4.

3. Results

3.1 Quantification of nitrite and nitrate

The derivatization product of NO₂⁻ is 2,3-naphthotriazole (NAT), which is detected by LC-MS/MS. ¹⁴N-NAT has a parent mass of 169.1 Da, which has an [M+H]⁺ mass of 170.1 Da. The major product ions of ¹⁴N-NAT are C₁₀H₈N⁺ (m/z 142.1) and C₉H₇⁺ (m/z 115.1). The proposed fragmentation pattern of ¹⁴N-NAT is in Figure 3. The transition of m/z 170.1→115.1 was chosen for quantitation because of the high abundance and specificity. ¹⁵N-NAT has a m/z of 171.1 and a quantitation transition of m/z 171.1→115.1.

Figure 3.

Proposed fragmentation pattern of 2,3-naphthotriazole (NAT). The molecular cation of ¹⁴N-NAT (m/z 170.1) is fragmented into the qualifier ion C₁₀H₈N⁺ (m/z 142.1), and undergoes additional fragmentation into C₉H₇⁺ (m/z 115.1) which is used for quantification.

¹⁴N-NAT and ¹⁵N-NAT co-elute at 9.60 minutes. There is a significant peak at 10.04 minutes that was identified as a degradation product of ¹⁴N-NAT and ¹⁵N-NAT. ¹⁴N-NAT and ¹⁵N-NAT degrade into their N-oxide metabolites (m/z 186 and 187)(Supplemental Figure 1), which give rise to protonated ¹⁴N-NAT and ¹⁵N-NAT species upon in-source loss of •OH (m/z 169 and 170). This “degradation peak” is larger in cell culture media samples, and increases over time if the samples remain at room temperature for several days. This demonstrates the necessity of measuring all derivatized samples within 24 hours of preparation.

3.2 Method Validation

The LC-MS/MS method of measuring nitrite and nitrate was analyzed for sensitivity, reproducibility, accuracy, precision, interferences, percent recovery, and carryover. Standard curves of nitrate and nitrite were used to determine LLOQ and the linear range. The LLOQ was defined as the lowest standard on the calibration curve at which the analyte peak was at least 5 times the response relative to blank response, with a precision of ≤ 20% RSD and an accuracy of 80-120%. ¹⁵N-Nitrite and ¹⁵N-nitrate had a LLOQ of 1 nM in both water and in DMEM (n=10). This value is within the same range as previous studies that have measured ¹⁵N-nitrite with LC-MS/MS, which reported LLOQ of 4 nM [19] and 5 nM [20]. We postulate that our quantitation limit is lower because different instrumentation and column, as well as our choice in solvent system. NAT is more stable under alkaline conditions, so using a solvent system with 0.1% formic acid [19; 20] may lead to degradation of NAT and a reduction in sensitivity.

¹⁵N-Nitrite had a linear range from 1 nM to 10 μM (correlation coefficient > 0.99, P<0.01), whereas ¹⁵N-nitrate was linear from 1 nM – 100 nM (correlation coefficient > 0.99, P<0.05) (n=5). Though nitrate is detectable at very low concentrations, it demonstrated a low precision, high level of variability, and low recovery at concentrations greater than 100 nM (correlation coefficient = 0.72, P>0.05). This is most likely due to the saturation of nitrate reductase. For these reasons, this method is only suitable to measure nitrate at nanomolar levels without further method development. However, increasing the amount of enzyme was not desirable due to potential interferences with mass spectrometry.

Precision and accuracy were analyzed by spiking known amounts (10 nM, 100 nM, 1000 nM) of ¹⁵NO₂⁻, NO₂⁻, ¹⁵NO₃⁻, and NO₃⁻ into water and DMEM (phenol red-free, with 1% FBS and PS), n=10. Individual calibration curves (seven point, 1 nM – 10 μM) were used for quantitation. Accuracy is determined as the percentage deviation of calculated concentration from the nominal concentrations, and was calculated with the equation: $\frac{calculatedconcentration-nominalconcentratin}{nominalconcentration}\ast 100\%$. Precision is determined as the relative standard deviation (RSD) of the 10 measurements, and is calculated as $\frac{standarddeviation}{mean}\ast 100\%$. Accuracy and precision is presented in Table 1. Satisfactory accuracies were obtained for ¹⁴N-nitrite and ¹⁵N-nitrite in both DMEM and water, with biases less than 7.5% and 5.6%, respectively. Accuracy for ¹⁴N-nitrate and ¹⁵N-nitrate was satisfactory up to 100 nM (less than 7.3% and 4.7%, respectively) but was not acceptable for 1000 nM (up to 20.4% and 17.9%, respectively). There was good precision for ¹⁴N-nitrite (4.76 to 10.9%) and ¹⁵N-nitrite (2.41 to 9.17%) in water and DMEM. Precision for ¹⁴N-nitrate (6.32 to 9.43%) and ¹⁵N-nitrate (5.16 to 8.51%) was high at 100 nM and lower, but was unacceptably low for the 1000 nM concentration (up to 19.2 and 15.6%, respectively). The results demonstrate that using ¹⁵N labeled analytes improves accuracy and precision compared to the ¹⁴N analytes. Nitrate has very low recovery above 100 nM, and should not be quantified at high nanomolar to micromolar concentrations without further method validation. This method is precise and accurate for the measurement of ¹⁵N-nitrite and ¹⁵N-nitrate in both water and DMEM.

Table 1.

Three concentrations (10 nM, 100 nM, 1000 nM) of ¹⁵NO₂^-, NO₂^-, ¹⁵NO₃^-, and NO₃^- were spiked into water (n=10) and DMEM (n=10), and the concentration was calculated by comparing to a seven-point concentration curve. Accuracy is determined as the percentage deviation of calculated concentration from the nominal concentrations. Precision is determined as the relative standard deviation (RSD) of the 10 measurements. Blank values for ¹⁴N-nitrite and ¹⁴N-nitrate in DMEM and water are provided.

Analyte	Isotope	Matrix	Concentration (nM)	Measured Concentration (nM)	Accuracy (%)	Precision (RSD, %)
Nitrite	14N	Water	Blank	91	N/A	8.71
			10	9.22	−7.80	9.43
			100	96.3	−3.70	5.38
			1000	926	−7.40	6.38
		DMEM	Blank	339	N/A	10.3
			10	9.25	−7.50	10.9
			100	101	1.00	4.76
			1000	1060	6.00	9.06
	15N	Water	10	9.49	−5.10	9.17
			100	94.7	−5.30	2.41
			1000	964	−3.60	5.39
		DMEM	10	10.1	1.00	4.42
			100	96	−4.00	3.07
			1000	944	−5.60	6.02
Nitrate	14N	Water	Blank	141	N/A	12.3
			10	9.43	−5.70	6.32
			100	92.7	−7.30	10.4
			1000	841	−15.9	17.6
		DMEM	Blank	467	N/A	13.2
			10	9.31	−6.90	8.59
			100	94.7	−5.30	9.43
			1000	796	−20.4	19.2
	15N	Water	10	9.61	−3.90	7.26
			100	95.3	−4.70	8.51
			1000	884	−11.6	15.6
		DMEM	10	10.1	1.00	5.16
			100	98.9	−1.10	6.48
			1000	821	−17.9	12.4

Neither nitrite nor nitrate was found to interfere with each other. Percent recovery of nitrite and nitrate in the presence of the other analyte was 97% ± 4 and 98% ± 5, respectively. This demonstrates that we can measure nitrate and nitrite in a single sample without compromising the quantitation. Phenol-red free DMEM culture media with 1% FBS and PS was found to not be a source of interference and caused minimal matrix effects. There was high recovery of spiked ¹⁵N-nitrite and ¹⁵N-nitrate within their linear range, n=10. The percent recovery of nitrite and nitrate in DMEM for all concentration points was 91% ± 3 and 99% ±11, respectively. This demonstrates that we can analyze nitrite and nitrate in cell culture medium with minimal sample preparation.

Both NADPH and nitrate reductase were added individually to ¹⁵N-nitrite standards (5 points, 1 nM- 1 μM) to determine whether they would interfere with mass spectrometric detection of nitrite. NADPH and nitrate reductase did not affect either derivatization of nitrite, or mass spectrometric quantification of NAT. Samples incubated with 10 μL of 120 μM NADPH had a percent recovery of 100% ± 5, and samples with 10 μL of 1 U/mL nitrate reductase had a percent recovery of 99% ± 3. This indicates that at the concentrations tested, the presence of NADPH and nitrate reductase will not interfere with the recovery of nitrite from nitrate. In both water and DMEM, there was greater than 99% conversion of nitrate into nitrite within the linear range.

We quantified the concentration of nitrite in milli-Q water, 200 mM HEPES, phenol red-free DMEM with 1% FBS, and phenol red-free DMEM with 1% FBS that was used to culture 5 × 10⁶ EA.hy926 cells in 2 mL medium for 24 hours at 37°C. The results are reported in Table 2. Background levels of nitrate exceeded our linear range of measurement, so these values are not reported. These high levels of nitrite and nitrate in both water and media demonstrate the necessity of using isotope labeling in cell culture experiments.

Table 2.

Nitrite was measured in 1) reverse osmosis water filtered with Millipore water filter, 2) 200 mM HEPES made from Milli-Q water, 3) phenol red-free Dulbecco's modified Eagle's medium (DMEM) with 1% FBS, and 4) phenol red-free DMEM with 1% FBS that was used to culture 5 × 10⁶ EA.hy926 cells in 2 mL medium for 24 hours at 37°C. Nitrate was not measured because all values were above the linear range.

	Nitrite (nM)
Millipore water	91 ± 13
200 mM HEPES2	103 ± 15
DMEM – blank3	339 ± 28
DMEM - EA.hy9264	461 ± 61

Value are mean ± standard deviation., n=6

3.1 GTN dose-response: nitrate tolerant and control

Nitrate tolerance was induced in EA.hy926 cells with a 5-hour pre-treatment with GTN. The endothelial cells bioactivated GTN to produce nitrite, whereas nitrate was not a significant product. Representative chromatograms are in Figure 4. A linear dose-response for ¹⁵N-nitrite and ¹⁵N-nitrate production was measured. GTN pre-treatment caused decreased production of ¹⁵N-nitrite compared to the vehicle control, at all 4 concentrations tested (1, 5, 10, 20 μM) (P= 0.006, 0.014, 0.047, 0.029) (Figure 5a). However, ¹⁵N-nitrate production was not altered by pre-treatment with GTN (P>0.05 for all concentration points) (Figure 5b). The concentration of nitrate was near the LLOQ, indicating that nitrate is not a significant metabolic product of GTN. There was no difference in ¹⁴N-nitrite and ¹⁴N-nitrate concentration between nitrate tolerant cells and control cells, demonstrating that there was no competition between GTN and ¹⁵N₃-GTN that would invalidate results (Figure 5c and 5d). ¹⁴N-nitrite levels were similar for all concentration points, ranging from 411 to 487 nM, whereas ¹⁴N-nitrate was 150 to 184 nM. The lower levels of ¹⁴N-nitrate can be explained by the observation that these values were above our detection limit for nitrate, and thus suffer from incomplete conversion. This demonstrates that EA.hy926 cells can become nitrate tolerant, and may be a model for measuring the development of nitrate tolerance. This is a significant finding because it suggests that nitrate tolerance develops at the level of GTN bioactivation. The results indicate that the enzymes that are responsible for GTN metabolism may be inactivated or damaged by continuous GTN treatment.

Figure 4.

Representative extracted ion-chromatograms (XIC) of 2,3-naphthotriazole (NAT). ¹⁵N-NAT (m/z 171.1→115.1) XIC is presented in a) ¹⁵NO₂⁻ standard in DMEM, b) 20 μM ¹⁵N₃-GTN (vehicle control), and c) 20 μM ¹⁵N₃-GTN (pre-treatment with GTN). These data demonstrate that pre-treatment with GTN resulted in reduced metabolism of ¹⁵N₃-GTN into ¹⁵NO₂⁻. ¹⁴N-NAT (m/z 170.1→115.1) is presented in d) ¹⁴N-nitrite standard in DMEM, e) 20 μM ¹⁵N₃-GTN (Vehicle Control), and f) 20 μM ¹⁵N₃-GTN (pre-treatment with GTN). These data demonstrate that pre-treatment with GTN did not cause a change in ¹⁴NO₂⁻ concentration.

Figure 5.

¹⁵NO₂⁻ and ¹⁵NO₃⁻ production from ¹⁵N₃-GTN treatment in EA.hy926 endothelial cells. Cells were pre-treated with 20 μM glyceryl trinitrate (GTN), washed, and then treated with ¹⁵N₃-GTN. Acetonitrile (5 μL) was used as a vehicle control. a) ¹⁵NO₂⁻ levels decreased in response to GTN pre-treatment, n=3 (P<0.05 at all concentration points). A dose-response curve with ¹⁵N₃-GTN demonstrated linear production of ¹⁵NO₂⁻ in nanomolar levels. b) GTN pre-treatment had no significant effect on ¹⁵NO₃⁻ levels (P>0.5 at all concentration points). A dose-response curve with ¹⁵N₃-GTN demonstrated linear production of ¹⁵NO₃⁻ in low nanomolar levels, indicating very low conversion of nitrite into nitrate. The reduction of nitrite to nitrate may occur nonenzymatically. c) ¹⁴NO₂⁻ concentration and d) ¹⁴NO₃⁻ concentration was not significantly affected (P>0.05 at all concentration points). This demonstrates that the pre-treatment was completely washed away and did not compete with ¹⁵N-GTN. The high level of ¹⁴NO₂⁻ and ¹⁴NO₃⁻ is a result of background levels of nitrite and nitrate, potentially from NO synthesis from L-arginine.

These results are supported by the MTT cell viability assay, which determined that 24-hour treatment with a wide range of GTN treatments (0.1 μM – 50 μM) does not cause significant levels of cell death (Supplemental Information Figure 2). In addition, the ATP Glo Assay to measure cell proliferation showed no significant differences between controls and multiple concentrations of GTN (1, 5, 10, 20 μM) (Supplemental Information Figure 3). Both assays demonstrated minor, yet insignificant, increases in cell viability/proliferation with GTN treatment. Therefore, the reduction in nitrite and nitrate production in nitrate tolerant cells is not caused by a reduced amount of viable cells.

4. Discussion

The results demonstrate that our method of measuring nitrite and nitrate is highly sensitive, can differentiate between isotopologues of nitrite and nitrate, is accurate and precise, is free of common interferences, and requires minimal sample preparation. HPLC methods that use fluorescence detection have detection limits in low nanomolar levels, however they are not able to distinguish between newly formed and pre-existing (i.e., background) nitrite and nitrate [25] [27],[28]. The ability of these methods to measure minute changes in nitrite and nitrate is questionable, particularly in light of the fact that we measured high nanomolar background levels of nitrite in cell culture media. Small changes in NO, and its metabolites, can have large physiological effects. We demonstrated that LC-MS/MS can measure changes in ¹⁵N-nitrite and ¹⁵N-nitrate concentrations in low nanomolar concentrations, which would undoubtedly be obscured without isotope labeling. Measuring NAT with LC-MS/MS, instead of by fluorescence, has the additional benefit of increased stability, because fluorescence quenching by biological components is not a concern. Though the merit of measuring nitrite and nitrate as surrogates for NO synthesis has been recently questioned [29], it remains the most suitable and practical method to assess NO production from GTN.

Despite the numerous benefits of measuring nitrite and nitrate with LC-MS/MS, there is one limitation that should be considered. This method requires for separate analysis of nitrite and nitrate, which increases sample preparation and analysis time. For many applications, this may not be a concern. We have demonstrated that nitrite is a better marker of GTN metabolism than nitrate, thus it may suffice to measure nitrite only. Human and mammalian in vivo studies indicate that circulating nitrite, rather than nitrate, reflect endothelial-dependent NO synthesis [30]. We acknowledge that methods that allow for simultaneous quantification of nitrite and nitrate may be preferable for human biological fluids. The GCMS method developed by Tsikas and coworkers, which uses pentafluorobenzyl derivatization, allows for the measurement of isotopes and has the benefit of being able to measure nitrite and nitrate simultaneously [31]. However, this GCMS method requires extensive sample preparation. The sensitivity of our LC-MS/MS method, as well as the ability to use isotope labeling, is necessary for mechanistic cell cultures studies.

One source of concern with this method could be that ¹⁵N-NAT and ¹⁴N-NAT only differ by a single mass unit. In theory, this could lead to issues with quantitation and interference with natural isotope abundance. We determined that ¹⁵N-NAT is not detectable in cell culture samples with no treatment, indicating that the natural abundance of ¹⁵N (0.36%) will not affect quantification. In addition, natural abundance of ¹³C could not be mistaken for ¹⁵N labeling because of the fragmentation pattern of NAT. There is a significant mass fragment peak at m/z 116.0, which results from the natural abundance of ¹³C₁ labeled NAT (9.7%). Since we measure the m/z 171.1→115.1 transition for ¹⁵N-NAT, ¹³C labeling is excluded from detection.

The cell culture results demonstrate that endothelial cells are capable of metabolizing GTN into nitrite. Pre-treatment with GTN resulted in nitrate tolerance, as evidenced by reduced ¹⁵N-nitrite production. However, only low nanomolar levels of ¹⁵N-nitrate were measured, suggesting that it is not a significant metabolic product of GTN. These results indicate that this may be a good model to determine the mechanism of GTN bioactivation, as well as the cause of nitrate tolerance. Measuring nitrite as an endpoint of GTN metabolism is a valuable tool to determine the cause of nitrate tolerance. We hypothesize that inactivation of GTN-metabolizing enzymes, such as xanthine oxidase (XO), is the cause of nitrate tolerance.

5. Conclusion

LC-MS/MS is a sensitive and reproducible method of quantifying isotopologues of nitrite and nitrate, which will bring new insight into the phenomena of nitrate tolerance. We demonstrated that nitrate tolerance is associated with a significant decrease in nitrite production in human endothelial cells, which indicates inactivation of GTN-metabolizing enzymes.