Method to Simultaneously Determine the Sphingosine 1-Phosphate

Disruption of the Sj?gren-Larsson Syndrome Gene Aldh3a2 in Mice Increases Keratinocyte Growth and Retards Skin Barrier Recovery. Tatsuro Naganuma ...
0 downloads 0 Views 827KB Size
Article pubs.acs.org/ac

Method to Simultaneously Determine the Sphingosine 1‑Phosphate Breakdown Product (2E)‑Hexadecenal and Its Fatty Acid Derivatives Using Isotope-Dilution HPLC−Electrospray Ionization−Quadrupole/ Time-of-Flight Mass Spectrometry Corinna Neuber,† Fabian Schumacher,†,‡ Erich Gulbins,‡ and Burkhard Kleuser*,† †

Department of Nutritional Toxicology, Institute of Nutritional Science, University of Potsdam, Arthur-Scheunert-Allee 114-116, 14558 Nuthetal, Germany ‡ Department of Molecular Biology, University of Duisburg-Essen, Hufelandstraße 55, 45147 Essen, Germany ABSTRACT: Sphingosine 1-phosphate (S1P), a bioactive lipid involved in various physiological processes, can be irreversibly degraded by the membrane-bound S1P lyase (S1PL) yielding (2E)-hexadecenal and phosphoethanolamine. It is discussed that (2E)-hexadecenal is further oxidized to (2E)-hexadecenoic acid by the long-chain fatty aldehyde dehydrogenase ALDH3A2 (also known as FALDH) prior to activation via coupling to coenzyme A (CoA). Inhibition or defects in these enzymes, S1PL or FALDH, result in severe immunological disorders or the Sjögren-Larsson syndrome, respectively. Hence, it is of enormous importance to simultaneously determine the S1P breakdown product (2E)hexadecenal and its fatty acid metabolites in biological samples. However, no method is available so far. Here, we present a sensitive and selective isotope-dilution high performance liquid chromatography−electrospray ionization−quadrupole/time-offlight mass spectrometry method for simultaneous quantification of (2E)-hexadecenal and its fatty acid metabolites following derivatization with 2-diphenylacetyl-1,3-indandione-1-hydrazone and 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide. Optimized conditions for sample derivatization, chromatographic separation, and MS/MS detection are presented as well as an extensive method validation. Finally, our method was successfully applied to biological samples. We found that (2E)-hexadecenal is almost quantitatively oxidized to (2E)-hexadecenoic acid, that is further activated as verified by cotreatment of HepG2 cell lysates with (2E)-hexadecenal and the acyl-CoA synthetase inhibitor triacsin C. Moreover, incubations of cell lysates with deuterated (2E)-hexadecenal revealed that no hexadecanoic acid is formed from the aldehyde. Thus, our method provides new insights into the sphingolipid metabolism and will be useful to investigate diseases known for abnormalities in long-chain fatty acid metabolism, e.g., the Sjögren-Larsson syndrome, in more detail. egress from lymph nodes.9,10 This offers new opportunities to investigate S1PL as a novel immunosuppressant drug target. Interestingly, the fate of the S1P degradation products (2E)hexadecenal and phosphoethanolamine is not fully examined yet. Due to the α,β-unsaturated aldehyde function, (2E)hexadecenal is a reactive compound and may be subjected to various enzymatic or spontaneous intra- or extra-cellular reactions. For instance, it was reported that (2E)-hexadecenal forms adducts with the DNA nucleoside 2′-deoxyguanosine in vitro via Schiff base formation.11 Furthermore, it has been discussed that an enzymatic conversion of (2E)-hexadecenal to (2E)-hexadecenoic acid and/or hexadecanoic (palmitic) acid occurs, which makes a re-entering of the aldehyde into the sphingolipid metabolism plausible. One enzyme capable to

S

phingosine 1-phosphate (S1P), a bioactive lipid mediator, is involved in many signaling pathways such as cell growth, differentiation, migration, and apoptosis.1−5 S1P, which is initially formed from sphingosine by sphingosine kinases, can be degraded via at least two mechanisms.6 First, S1P may be able to re-enter the sphingolipid synthesis pathway after dephosphorylation to sphingosine by specific intracellular S1P phosphatases. On the other hand, it can be irreversibly cleaved into (2E)-hexadecenal and phosphoethanolamine by the membrane-bound S1P lyase (S1PL). Thus, S1PL plays a central role in the regulation of the biological activity of S1P. Changes in S1PL activity were correlated with developmental defects or with resistance to specific anticancer therapies indicating the importance of this enzyme in cellular physiology.7−9 Moreover, inhibition of S1PL or knockout of the S1PL gene in mice and cells in culture led not only to elevated S1P levels but also to a blockage of the lymphocyte © 2014 American Chemical Society

Received: May 6, 2014 Accepted: August 19, 2014 Published: August 19, 2014 9065

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

Using these chemicals, we were able to develop a sensitive quantification method for fatty aldehydes and acids based on isotope-dilution HPLC-ESI-QTOF, which then was successfully applied to biological samples.

catalyze this reaction is the long-chain fatty aldehyde dehydrogenase ALDH3A2 (also known as FALDH), which in general oxidizes long-chain fatty aldehydes to the corresponding fatty acids. Mutations in the ALDH3A2 gene can be found in patients suffering from the Sjögren-Larsson syndrome. This inherited disorder is characterized by ichthyosis, spasticity, and mental retardation.12 Due to the fundamental role of both enzymes, S1PL and FALDH, in cellular processes, it is of great interest to quantify the S1P breakdown product (2E)-hexadecenal and its putative metabolites (2E)-hexadecenoic acid and hexadecanoic acid. Up to now, there is no method available suitable for a simultaneous quantification of (2E)-hexadecenal and its fatty acid metabolites using mass spectrometric techniques. Here, we report on the development of a sensitive and specific high performance liquid chromatography−electrospray ionization−quadrupole/time-offlight mass spectrometry (HPLC-ESI-QTOF) method using characteristic mass transitions and deuterated internal standards for unambiguous and simultaneous quantification of (2E)hexadecenal and the putative metabolites (2E)-hexadecenoic acid and hexadecanoic acid. The chemical structures of all analytes and internal standards monitored in the present study are depicted in Scheme 1. Due to their relatively high



EXPERIMENTAL SECTION Materials. (2E)-Hexadecenal and (2E)-hexadecenal (d5) were purchased from Avanti Polar Lipids (Alabaster, USA). DAIH, EDC hydrochloride, (2E)-hexadecenoic acid, hexadecanoic acid, hexadecanoic acid (d5), and heptadecanoic acid were obtained from Sigma-Aldrich (Taufkirchen, Germany). Hexadecanal, pentadecanal, and (9Z)-hexadecenoic acid were ordered from ABCR (Karlsruhe, Germany). Fatty acid free bovine serum albumin (BSA) was ordered from Carl Roth (Karlsruhe, Germany). Fetal bovine serum (FBS) was obtained from Moregate Biotech (München, Germany). RPMI-1640 medium, DMEM medium, and penicillin/streptomycin were purchased from Biochrom AG (Berlin, Germany). Triacsin C was obtained from Santa Cruz Biotechnology (Heidelberg, Germany). All chemicals and solvents used were of LC-MS grade. Cell Culture. HepG2 cells or human dermal fibroblasts were grown in 15 cm dishes using RPMI-1640 medium or DMEM medium, respectively, each supplemented with 10% FBS, 100 U mL−1 penicillin, and 100 μg mL−1 streptomycin. Cells that grew at 37 °C in an atmosphere of 5% CO2 and 95% air-humidity were used between passage 2 and 12 (HepG2 cells) or 2 and 6 (fibroblasts). Cells in exponential phase were harvested through trypsination. Cell pellets containing 2 × 106 cells were stored at −80 °C until usage. Cell Lysis and Treatment of Cell Lysates. Cell pellets were resuspended in 500 μL of 100 mM potassium phosphate buffer (pH 7.4) and sonicated three times for 20 s on ice. The whole cell homogenate was warmed for 5 min at 37 °C prior to the addition of (2E)-hexadecenal or (2E)-hexadecenal (d5) with final concentrations of 8 μM. After incubation at 37 °C, 10 μL of the cell lysates was taken at different time points (5, 10, 15, 30, and 60 min) and immediately extracted as described below. In further experiments, cell lysates, which were incubated with (2E)-hexadecenal (d5), were coincubated with 20 μM triacsin C. Extraction of Fatty Aldehydes and Acids from Cell Homogenates. For extraction of fatty aldehydes and acids, a modified extraction procedure published by Püttmann et al.14 was used. Before use, all glassware was rinsed with ethanol and air-dried to avoid contamination of samples with exogenous sources of fatty acids. Ten μL of cell homogenate was placed in a 1.5 mL glass vial on ice and filled up with water to 100 μL. For analysis of (2E)-hexadecenal and the corresponding acids, 5 μL of a mixture containing 20 μM of both (2E)-hexadecenal (d5) and hexadecanoic acid (d5) in ethanol was added as internal standard. For cell culture experiments addressing the metabolic oxidation of (2E)-hexadecenal (d5), 5 μL of a mixture containing 20 μM pentadecanal and heptadecanoic acid in ethanol was taken as internal standard. Both chemicals are frequently used as internal standards for quantification of longchain fatty aldehydes and acids in biological samples, since they do not occur physiologically. Extraction started by adding 500 μL of ice-cold Dole solution consisting of isopropanol/nheptane/2 M phosphoric acid (40:10:1, v/v/v). After vortexing for 30 s, 200 μL of n-heptane and 300 μL of water were added, and the tubes were again thoroughly vortexed. After a 10 min incubation step on ice, the samples were centrifuged at 400g for

Scheme 1. Chemical Structures of All Analytes Investigated and Internal Standards Used in the Present Studya

a a, hexadecanal; b, (2E)-hexadecenal; c, hexadecanoic acid; d, (2E)hexadecenoic acid; e, (9Z)-hexadecenoic acid; f, (2E)-hexadecenoic acid (d5); g, (2E)-hexadecenal (d5); h, hexadecanoic acid (d5); i, pentadecanal; j, heptadecanoic acid.

hydrophobicity, long-chain fatty aldehydes and acids are poorly ionizable using electrospray ionization and, therefore, require a derivatization step prior to ESI-MS detection. We developed a protocol including a combination of the derivatization agent 2diphenylacetyl-1,3-indandione-1-hydrazone (DAIH), known for good reactivity toward fatty aldehydes,13 and the catalyst 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide (EDC), necessary for an additional sufficient derivatization of fatty acids. 9066

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

5 min. 200 μL of the upper organic phase was transferred into a 400 μL glass inlet and evaporated to dryness. The residue was subjected to derivatization. Derivatization Procedure. Derivatization using a saturated solution of DAIH was optimized by sequentially varying the reaction temperature (60 and 80 °C), reaction duration (5, 10, 20, 30, and 60 min), and EDC concentration (12.5, 25.0, 37.5, 50.0, 62.5, 75.0, and 87.5 mM). In our final protocol, fatty aldehydes and acids were dissolved in 25 μL of ethanol. The sample solution was mixed with 50 μL of 1.69 mM DAIH in acetonitrile, 50 μL of 250 mM EDC·HCl in ethanolic 3% pyridine, and 75 μL of water. This mixture was allowed to react for 15 min at 80 °C in a water bath and was then immediately cooled down to 4 °C. Ten μL of the derivatized sample was directly injected into the HPLC-ESI-QTOF system to determine the investigated fatty aldehydes and acids. Detection of Derivatized Long-Chain Fatty Aldehydes and Acids by HPLC-ESI-QTOF. Analysis of the aldehyde and fatty acid derivatives was conducted with an Agilent 1200 liquid chromatography system coupled to an Agilent 6530 quadrupole/time-of-flight mass spectrometer (both from Waldbronn, Germany). Chromatographic separation was performed on a ZORBAX Eclipse XDB-C18 column (4.6 × 50 mm, 1.8 μm). The injection volume was 10 μL per sample. At a flow rate of 0.9 mL min−1, the following elution gradient was applied: the initial solvent composition of 85% methanol containing 0.01% ammonium hydroxide (eluent B) and 15% water (eluent A) was changed to 100% eluent B within 3 min. Then, the system was run isocratically for 9 min before the gradient was changed to the starting conditions. The total run time of one analysis was 18 min. For mass spectrometric detection, the derivatized fatty aldehydes and fatty acids were ionized in an electrospray source operating in the negative ion mode. The following settings of the ion source and the MS/MS detector were used: sheath gas temperature = 400 °C, sheath gas flow = 9 L min−1, nebulizer pressure = 22 psig, drying gas temperature = 350 °C, drying gas flow = 9 L min−1, capillary voltage = 4000 V, fragmentor voltage = 200 V, and nozzle voltage = 250 V. Table 1 presents the precursor ions of the investigated DAIH derivatives as well as the corresponding fragment ions generated under optimized collision energies. Method Validation and Quantification of Analytes. The following method validation parameters were determined: linearity of detection, limit of detection (LOD), limit of quantification (LOQ), recovery, precision, and accuracy (intraand interday variability). All samples had to be derivatized prior to LC-MS detection as described in a previous paragraph. Linearity of detection of (2E)-hexadecenal and (2E)hexadecenoic acid was determined using ethanolic standard solutions. The remaining validation parameters were determined after extraction of spiked samples containing PBS and 4.5% BSA (fatty acid free) as described for cell homogenates in a previous section. To all samples of the method validation and cell culture assays, constant amounts of (2E)-hexadecenal (d5) and hexadecanoic acid (d5) were added as internal standards (1 pmol each on column). To samples of the metabolism study of (2E)-hexadecenal (d5) in lysates of HepG2 cells, pentadecanal and heptadecanoic acid were added as internal standards instead (1 pmol each on column). Additionally, for all samples (method validation and cell culture assays), an external calibration was recorded using nine ethanolic standards of (2E)-hexadecenal and (2E)-hexadecenoic acid in the range of 10 fmol to 25 pmol. Quantification was carried out by dividing

Table 1. Precursor Ions, Corresponding Fragment Ions, and Optimized Collision Energies for the Selected Reaction Monitoring (SRM) of DAIH Derivatives of Fatty Aldehydes and Acids Investigated DAIH derivative of

precursor ion [M − H]−

fragment ion [M − H]−

collision energy [V]

pentadecanal (2E)-hexadecenala hexadecanal (2E)-hexadecenal (d5)b (2E)-hexadecenoic acidc (9Z)-hexadecenoic acid hexadecanoic acid (2E)-hexadecenoic acid (d5)d,e hexadecanoic acid (d5)d heptadecanoic acid

561.3487 573.3487 575.3643 578.3800 589.3436 589.3436 591.3592 594.3750 596.3906 605.3749

167.0860 249.2336 381.2911 254.2650 395.2704 395.2704 397.2861 400.3042 402.3174 411.3013

35 24 35 24 30 30 30 30 30 30

a

(2E)-hexadecenal (d5) used for internal calibration. bPentadecanal used for internal calibration. cHexadecanoic acid (d5) used for internal calibration. dHeptadecanoic acid used for internal calibration. eNot available as purified standard; detection after formation in situ.

the ratio of the peak areas of analyte and corresponding internal standard by the slope of the regression line of the external calibration. LOD and LOQ were estimated using the signal-tonoise (S/N) ratio plotted against the applied amount of the analyte. Thereby, LOD was defined as analyte amount that produces a signal with an S/N ratio of 3. LOQ was defined as three times the LOD. S/N ratios were calculated using the Agilent MassHunter Workstation Software Qualitative Analysis (Version B.06.00). The following calculation parameters were set: signal definition: height; noise definition: root-meansquare; automatic noise detection (maximum length: 1 min and minimum length: 0.1 min). Recoveries were determined using the absolute signal responses of the internal standards in samples of the external calibration compared with those in spiked PBS and BSA-containing samples or cell lysates after lipid extraction. Intra- and interday precision and accuracy were evaluated by assaying six consecutive replicates of a low, lowmedium, medium, and high quality control sample (0.5, 1, 5, 10 pmol), representing the concentration range of the investigated analytes in biological samples. Interday precision and accuracy were determined on three different days.



RESULTS Derivatization of (2E)-Hexadecenal and (2E)-Hexadecenoic Acid Using DAIH and EDC. Simultaneous determination of (2E)-hexadecenal and its fatty acid metabolites gives important information about the metabolism of the bioactive sphingolipid S1P and, moreover, allows an assessment of the activity of the critical enzymes in S1P degradation, S1PL and FALDH. The former enzyme is essential for the regulation of various immunological functions. Moreover, mutations in the gene encoding FALDH have been observed in patients suffering from the Sjögren-Larsson syndrome. Hence, it is important to determine the S1P breakdown product (2E)hexadecenal and its putative fatty acid metabolite (2E)hexadecenoic acid. Unfortunately, no method for simultaneous quantification of these compounds is available so far. The present study introduces a combination of the derivatization agent DAIH and coupling agent EDC capable for simultaneous detection of (2E)-hexadecenal and (2E)-hexadecenoic acid. Consequently, we optimized the derivatization parameters to

9067

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

Scheme 2. Chemical Structures of (2E)-Hexadecenal (in the Following Depicted as (2E)-16:1-CHO), (2E)-Hexadecenoic Acid (in the Following Depicted as (2E)-16:1-COOH), and the Corresponding Reaction Products Following Derivatization with DAIH in the Presence of EDC

Figure 1. Impact of incubation time (A) and applied final amount of the catalyst EDC (B) on the simultaneous derivatization of (2E)-16:1-CHO and (2E)-16:1-COOH (5 pmol each) with 0.42 mM DAIH. Both experiments were performed at 80 °C. For optimization of the incubation time, 62.5 mM EDC was used. To determine the optimal EDC concentration, samples were allowed to react for 15 min. The arrows indicate the optimized values of both investigated parameters used in all further experiments. Data of two parallel derivatization reactions are presented.

detect long-chain fatty aldehydes and acids within a single HPLC-ESI-QTOF run. Formation and chemical structures of DAIH derivatives of (2E)-hexadecenal and (2E)-hexadecenoic acid are depicted in Scheme 2. Derivatization Conditions. In order to optimize the derivatization temperature, we allowed the investigated fatty aldehydes and acids to react with DAIH (0.42 mM) in the presence of EDC (62.5 mM) at 60 or 80 °C for different durations. These experiments revealed maximum yields of azines (fatty aldehyde-DAIH derivatives) and hydrazides (fatty acid-DAIH derivatives) between 30 and 60 min, respectively, at 60 °C (not shown). Derivatization at 80 °C provided comparable maximum product yields of (2E)-hexadecenal and (2E)-hexadecenoic acid derivatives after 10 and 20 min (Figure 1A), respectively, without formation of any byproducts as verified by untargeted HPLC-ESI-QTOF. To minimize the time consumption of our experiments, we used a derivatization time of 15 min at 80 °C for further investigations. Second, we evaluated the impact of the applied amount of the catalyst EDC on the formation of azine and hydrazide derivatives using a saturated solution of DAIH and constant derivatization temperature (80 °C) and duration (15 min). As shown in Figure 1B, maximum yields for both derivatives were obtained using a final EDC concentration of 62.5 mM, which therefore was selected for all further experiments. Detection of Derivatized Long-Chain Fatty Aldehydes and Acids Using HPLC-ESI-QTOF. Chromatographic separation of the investigated, derivatized long-chain fatty aldehydes

and acids was conducted by reversed-phase rapid resolution liquid chromatography. To avoid confounding of signals and ion suppression during MS/MS detection, it was important to separate (9Z)-hexadecenoic acid, an abundant product of the mitochondrial fatty acid β-oxidation, from the much less abundant S1P breakdown product (2E)-hexadecenoic acid, as both isomers possess the same precursor and product ions. Figure 2A exemplifies the separation of derivatized (2E)hexadecenal and two isomers (9Z and 2E) of hexadecenoic acid. Both isomers were derivatized with comparable efficiency under the reported optimized conditions. Separation of both isomers was sufficient, as experiments with HepG2 cell lysates incubated with unlabeled or deuterated (2E)-hexadecenal revealed almost equal results in respect to conversion of the aldehyde to the corresponding fatty acid (compare Figures 5 and 6). In order to quantify the analytes investigated in cell lysates incubated with (2E)-hexadecenal, we used the deuterated internal standards (2E)-hexadecenal (d5) and hexadecanoic acid (d5) for internal calibration and unambiguous identification (Figure 2B). When the oxidative metabolism of (2E)-hexadecenal (d5) was monitored, pentadecanal and heptadecanoic acid were used as internal standards (Figure 2C). Using the LC conditions presented in a previous section, the fatty aldehyde and acid derivatives were baseline separated within 9 min. Hereby, DAIH derivatives of (2E)-hexadecenal and (2E)-hexadecenal (d5) coeluted at 8.3 min. As fatty acids are more polar than their corresponding aldehydes, hydrazide derivatives of hexadecenoic acid eluted at 1.4 min (9Z-isomer) 9068

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

Figure 2. Merged SRM chromatograms of DAIH derivatives of (A) two isomers (9Z and 2E) of hexadecenoic acid and (2E)-hexadecenal, (B) hexadecanoic acid (d5) and (2E)-hexadecenal (d5), and (C) heptadecanoic acid and pentadecanal. All analytes and internal standards were analyzed in the same run with each peak set at 100%. Both, (2E)-hexadecenal and (2E)-hexadecenal (d5), possessed a second minor peak, probably due to isomerization during the derivatization process. The main peaks, indicated with asterisks, were used for quantification.

Figure 3. Product ion mass spectra and fragmentation patterns of derivatized (A) (2E)-16:1-CHO and (B) (2E)-16:1-COOH using negative electrospray ionization and collision energies as given in Table 1.

aldehydes and acids, we recorded product ion mass spectra to identify characteristic fragment ions of each analyte. After derivatization, all analytes were readily ionized within the ESI source to give [M − H]− precursor ions. That was not the case when derivatization was omitted (data not shown). Using the

and 1.9 min (2E-isomer). The latter coeluted with its internal standard hexadecanoic acid (d5). DAIH derivatives of the internal standards heptadecanoic acid and pentadecanal eluted at 2.2 and 7.7 min. In order to develop selective and sensitive SRM methods for the detection of the investigated fatty 9069

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

hexadecenal and (2E)-hexadecenoic acid. After lipid extraction and derivatization, the accuracy was calculated as the percentage of the deviation between the nominal and the determined concentrations. Precision was expressed in terms of the relative standard deviation (RSD) within a single assay (intraday) and between different assays (interday). As shown in Table 2, the coefficients of variation were less than 20%, and the precision and accuracy were in the range of 3.3−11.3% and 90−114%, respectively, which is in accordance with FDA guidance.15 Recovery of (2E)-hexadecenal and (2E)-hexadecenoic acid from samples containing 4.5% BSA as well as cell lysates was found to be in the range of 80−95%. The derivatized samples were stable for up to 3 days when stored at 4 °C. Degradation of (2E)-Hexadecenal, Formation of (2E)Hexadecenoic Acid and Further Activation via AcylCoA-Synthetase in Cell Lysates. After optimization of the derivatization parameters, the method was applied to HepG2 cells and human dermal fibroblasts in order to study the metabolism of the S1P degradation product (2E)-hexadecenal. Therefore, cell lysates were incubated with 8 μM (2E)hexadecenal for up to 60 min at 37 °C. As depicted in Figure 5, detected amounts of (2E)-hexadecenal were reduced to about half within 5 min in both types of cell lysates. However, after 30 min, the amount of the aldehyde was close to the LOQ in lysed HepG2 cells, whereas still 7% of the aldehyde was detectable after 60 min in lysed fibroblasts. Simultaneously, (2E)hexadecenoic acid was formed, suggesting the involvement of certain oxidoreductases such as FALDH in the degradation of (2E)-hexadecenal. In lysates of HepG2 cells, (2E)-hexadecenoic acid reached a maximum following a saturation curve at 60 min amounting to approximately 40% of the initial aldehyde amount. As expected, the aldehyde-to-acid conversion was less pronounced in human dermal fibroblasts. Here, after 60 min of incubation, a maximum of 12% of the applied amount of (2E)-hexadecenal was recovered as (2E)-hexadecenoic. As verified by incubating HepG2 cell lysates with (2E)hexadecenal (d5) (pentadecanal and heptadecanoic acid were used as internal standards in this case), we clearly showed the formation of only (2E)-hexadecenoic acid (d5) representing approximately 40% conversion of the added (2E)-hexadecenal (d5) after 60 min. Neither deuterated (9Z)-hexadecenoic acid nor hexadecanal or hexadecanoic acid were formed within a 60 min incubation at 37 °C. This is in contrast to the general assumption that (2E)-hexadecenal re-enters the sphingolipid metabolism after conversion to hexadecanoic acid. Fatty acids

collision energies given in Table 1, all analytes were fragmented into characteristic product ions. Figure 3 presents the product ion mass spectra of derivatized (2E)-hexadecenal and (2E)hexadecenoic acid using optimized collision energies. The spectral data (putative underlying fragmentation reactions are given in the insets of Figure 3) revealed a rather different fragmentation behavior of derivatized (2E)-hexadecenal compared to (2E)-hexadecenoic acid. Method Validation. As depicted in Figure 4, the detection of (2E)-hexadecenal and (2E)-hexadecenoic acid in ethanolic

Figure 4. Linear detection of derivatized (2E)-16:1-CHO and (2E)16:1-COOH in a concentration range of 10 fmol to 25 pmol per injection of 10 μL sample volume. Peak areas were normalized to those of the corresponding derivatized internal standards, (2E)-16:1CHO (d5) or 16:0-COOH (d5), respectively. The inset shows an enlargement of the low concentration range (10 fmol to 2.5 pmol per injection). Data of two injections from the same derivatized sample are presented.

solutions was linear over the whole tested concentration range (1−2500 fmol/μL) with correlation coefficients of r2 ≥ 0.998. Both, LOD and LOQ were estimated using spiked samples containing PBS and 4.5% fatty acid-free BSA that were subjected to lipid extraction before derivatization. The LOD, defined as the amount of analyte that causes a peak with an S/ N ratio of 3, was found to be 33 fmol for (2E)-hexadecenal and 69 fmol for (2E)-hexadecenoic acid. The LOQ, defined as three times the LOD, consequently amounted to 99 fmol for (2E)hexadecenal and 207 fmol for (2E)-hexadecenoic acid. Intraand interday precision and accuracy were evaluated by analyzing six consecutive replicates of PBS and BSA-containing samples spiked with 0.5, 1, 5, or 10 pmol of both (2E)-

Table 2. Method Validation Parameters Determined for the Detection of Derivatized (2E)-Hexadecenal and (2E)-Hexadecenoic Acid (2E)-hexadecenal nominal (pmol) intraday (n = 6)

interday (n = 6)

a

0.5 1 5 10 0.5 1 5 10

(2E)-hexadecenoic acid

found (pmol)a

precision (RSD, %)

accuracy (%)

nominal (pmol)

± ± ± ± ± ± ± ±

8.8 5.8 5.4 3.3 7.1 5.6 4.0 5.5

114 104 100 102 112 107 101 97

0.5 1 5 10 0.5 1 5 10

0.57 1.04 5.02 10.19 0.56 1.07 5.03 9.74

0.05 0.06 0.27 0.34 0.04 0.06 0.20 0.54

found (pmol)a

precision (RSD, %)

accuracy (%)

± ± ± ± ± ± ± ±

11.3 8.9 5.0 7.3 11.3 8.8 5.9 5.8

106 90 91 102 106 91 95 101

0.53 0.90 4.57 10.21 0.53 0.91 4.76 10.08

0.06 0.08 0.23 0.75 0.06 0.08 0.28 0.58

Data are means ± SD of six replicates. 9070

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

Figure 5. Time course of (2E)-16:1-CHO added to lysates of (A) HepG2 cells or (B) human dermal fibroblasts and formation of (2E)-16:1-COOH after incubation at 37 °C. 100% correspond to 4 pmol of extracted and derivatized (2E)-16:1-CHO injected into the LC-MS system. (2E)-16:1CHO and (2E)-16:1-COOH were quantified in relation to their internal standards (2E)-16:1-CHO (d5) and 16:0-COOH (d5), respectively. Data of two separate extractions (and subsequent derivatizations) from the same cell lysate are presented.

proposed the following metabolic pathway for (2E)-hexadecenal: first, it is oxidized by the fatty aldehyde dehydrogenase ALDH3A2 (also known as FALDH) to (2E)-hexadecenoic acid followed by conversion to (2E)-hexadecenoyl-CoA catalyzed by acyl-CoA-synthetases. Then, an up to now unidentified reductase saturates the double bond of (2E)-hexadecenoylCoA yielding hexadecanoic acid (palmitoyl)-CoA, that can reenter the sphingolipid synthesis pathway via serine palmitoyl transferase.18 Hence, the conversion of (2E)-hexadecenal into a fatty acid is believed to be the link between S1P degradation and reformation. To investigate this issue in a quantitative manner, a method for simultaneous determination of (2E)hexadecenal and the conceivable unsaturated and saturated fatty acid metabolites (2E)-hexadecenoic acid and hexadecanoic acid was required. We have recently published a highly sensitive method to determine (2E)-hexadecenal in cells and plasma after derivatization with DAIH using HPLC-ESI-QTOF.13 This hydrazone rapidly reacts with short- and long-chain aldehydes to form stable azines that are suitable for UV−vis, fluorescence, or MS detection. Among the popular hydrazone agents for derivatization of aldehydes, 2,4-dinitrophenylhydrazine (DNPH) is the most common one.19 By means of EDC as coupling agent, DNPH also forms adducts with fatty acids.20 However, to our best knowledge, only one study is published showing the simultaneous detection of both compound classes, fatty aldehydes and acids, after derivatization using directinfusion ESI-MS/MS.21 However, neither (2E)-hexadecenal nor (2E)-hexadecenoic acid were examined. While the mentioned study focused on the identification of an extensive set of carbonylated compounds, an absolute quantification using stable-isotopic labeled standards was not addressed. In the present study, we recommend an optimized derivatization protocol using DAIH and EDC that allows sensitive mass spectrometric quantification of both fatty aldehydes and acids after negative ionization using deuterated standards for structure confirmation and internal calibration. Our optimized conditions comprise an incubation time of 15 min and a derivatization temperature of 80 °C. This rather high temperature may raise the question about stability and oxidation status of the educts as well as the derivatization products. Regarding this, we scanned (2E)-hexadecenal and (2E)-hexadecenoic acid standards heated to 80 °C for byproducts using accurate mass QTOF. We did not find any degradation products and, therefore, have no hints for heat-

can be further activated via coupling to coenzyme A (CoA), representing the initial step for further metabolic processes such as β-oxidation within the mitochondria. In order to investigate to which extent (2E)-hexadecenoic acid (d5) follows a metabolic route depending on CoA-activation, HepG2 cell lysates were coincubated with deuterated (2E)-hexadecenal and triacsin C, a strong inhibitor of the long-chain fatty acid acylCoA-synthetase. As can be seen in Figure 6, triacsin C

Figure 6. Time course of (2E)-16:1-CHO (d5) added to HepG2 cell lysates and formation of (2E)-16:1-COOH (d5) and 16:0-COOH (d5) after incubation at 37 °C. 100% corresponds to 4 pmol of extracted and derivatized (2E)-16:1-CHO (d5) injected into the LC-MS system. (2E)-16:1-CHO (d5) and (2E)-16:1-COOH (d5) were quantified in relation to their internal standards pentadecanal and heptadecanoic acid, respectively. 16:0-COOH (d5) was not detectable in any taken sample. In a second set of samples, triacsin C, an inhibitor of acyl-CoA synthetases, was added to the cell lysates to block the further activation of fatty acids formed from (2E)-16:1-CHO (d5). Data of two separate extractions (and subsequent derivatizations) from the same cell lysate are presented.

treatment significantly elevated levels of (2E)-hexadecenoic acid (d5) detected. After 60 min of incubation, ca. 80% of deuterated (2E)-hexadecenal was converted into (2E)hexadecenoic acid (d5).



DISCUSSION S1P is a biologically active sphingolipid that modulates an enormous number of physiological processes.1,2,4−6,16 The activity of S1P is strongly regulated by the enzyme S1P lyase, which irreversibly cleaves it into phosphoethanolamine and (2E)-hexadecenal.17 Up to now, it is not fully understood how the aldehyde is further metabolized. Nakahara et al. recently 9071

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

triacsin C to inhibit the further activation of (2E)-hexadecenoic acid (d5), and indeed, inhibition of the acyl-CoA synthetase increased the levels of (2E)-hexadecenoic acid (d5) up to 80% (after 60 min) of the applied aldehyde amount. Thus, the majority of (2E)-hexadecenal in HepG2 cells is oxidized to (2E)-hexadecenoic acid that is further activated to (2E)hexadecenoyl-CoA by acyl-CoA synthetases. Statements about the fate of the remaining 20% of (2E)-hexadecenal are speculative. As (2E)-hexadecenal contains a reactive α,βunsaturated aldehyde function, it is conceivable that it may react with various cellular nucleophiles under Schiff base formation or Michael addition. Indeed, in vitro DNA adduct formation of (2E)-hexadecenal was described by Upadhyaya et al.11 Accordingly, we could mass spectrometrically detect both Schiff base and Michael adducts of glutathione (γ-L-glutamyl-Lcysteinylglycine) with (2E)-hexadecenal after incubation of both substances under buffered conditions in vitro (C. Neuber, F. Schumacher, B. Kleuser, unpublished results to be described elsewhere). Moreover, the 9Z-isomer of hexadecenoic acid and hexadecanoic acid were detectable in lysates of HepG2 cells and human fibroblasts incubated with (2E)-hexadecenal. However, incubation of HepG2 cell lysates with deuterated (2E)hexadecenal allowed us to exclude the formation of these two metabolites from (2E)-hexadecenal, because neither (9Z)hexadecenoic acid (d5) nor hexadecanoic acid (d5) was detectable. Ohkuni et al. cotreated HeLa cells with radiolabeled sphingosine or dehydrosphingosine and triacsin C and analyzed fatty acids after methyl esterification via TLC.22 In contrast to our findings, they detected methyl esters of (9Z)-hexadecenoic acid and hexadecanoic acid for both tritiated compounds. Additionally, they found labeled (2E)-hexadecenoic acid derived from [11,12-3H]sphingosine. Using compound specific MS/MS transitions and deuterated internal standards for structure identification and unambiguous quantification, as in the present study, it is almost impossible to confound signals with the wrong analytes. Due to poorer resolution and selectivity, this may not always be the case with TLC. Moreover, a different cell line used in the mentioned study may be a reason for the varying outcomes. Taken together, we detected only one fatty acid metabolite derived from (2E)hexadecenal in the cell lysates investigated, namely, (2E)hexadecenoic acid. We suggest that the formation of (9Z)hexadecenoic acid and hexadecanoic acid detected by Ohkuni et al. follows other metabolic pathways that do not include (2E)-hexadecenal. To clarify this issue, further studies should be envisaged. For that, again, our developed mass spectrometric approach will be a useful tool.

mediated autoxidation of (2E)-hexadecenal or (2E)-hexadecenoic acid in our assay. However, incubation times longer than 20 min (at 80 °C) led to substantially decreased yields of (2E)hexadecenal-DAIH-derivatives. It is likely that the derivatization products undergo transformation for instance by oxidation using prolonged incubation times at high temperature and are therefore not detectable using the reported targeted mass spectrometric approach. The detection of derivatized (2E)hexadecenal and (2E)-hexadecenoic acid was linear from 10 fmol to 25 pmol per injection. Thus, our new method provides a sensitivity for (2E)-hexadecenal-DAIH-adducts comparable to the method it is based on.13 Moreover, our method shows very good precision and robustness and was successfully applied to biological samples. We used lysates of HepG2 cells and human dermal fibroblasts for incubation with unlabeled and deuterated (2E)-hexadecenal at 37 °C. At defined time points, we took samples and determined the (2E)-hexadecenal and (2E)hexadecenoic acid contents. Up to now, only two studies were published, showing the conversion of (2E)-hexadecenal into the corresponding unsaturated fatty acid.18,22 The authors of the most recent study used thin-layer chromatography (TLC) for detection of (2E)-hexadecenoic acid in HeLa cells treated with the acyl-CoA synthetase inhibitor triacsin C.22 However, only qualitative data were presented, and (2E)hexadecenal was not included in the analyses. Therefore, no conclusions about the conversion rate of (2E)-hexadecenal to (2E)-hexadecenoic acid could be drawn. Using our newly developed HPLC-ESI-QTOF method, we were able to simultaneously analyze (2E)-hexadecenal and (2E)-hexadecenoic acid in lysates of HepG2 cells and human fibroblasts using deuterated internal standards for absolute quantification and, therefore, provide new insights into the metabolism of the S1P degradation product (2E)-hexadecenal. Already after 15 min of incubation, (2E)-hexadecenal was nearly completely degraded in HepG2 cell lysates. On the other hand, about 7% of the applied amount of (2E)-hexadecenal remained detectable in lysed fibroblasts after 60 min of incubation. In lysed HepG2 cells, the level of (2E)-hexadecenoic acid increased and reached a plateau at approximately 40% of the added amount of aldehyde after 60 min of incubation. The time course of the fatty acid formation indicates an involvement of enzymatic processes, which is in agreement with the metabolism pathway proposed by Nakahara et al. that includes conversion of (2E)hexadecenal to (2E)-hexadecenoic acid catalyzed by FALDH.18 Defects in this enzyme are associated with the Sjögren-Larsson syndrome. FALDH activity (using octadecanal as substrate) was reported to be substantially higher in preparations of human liver23 compared to cultured human fibroblasts.24 Nevertheless, our method allowed us to monitor the formation of (2E)-hexadecenoic acid from (2E)-hexadecenal also in lysed human dermal fibroblasts, cells that are of particular interest in the research of the molecular pathogenesis of the SjögrenLarsson syndrome.25,26 Hence, our developed HPLC-ESIQTOF method could be useful to analyze samples from affected patients in order to study metabolic processes of this autosomal, recessive, neurocutaneous disease in more detail. Since we have recovered not more than 40% of the applied amount of (2E)-hexadecenal as corresponding unsaturated fatty acid in our time-course assay using HepG2 cell lysates, we presumed a further metabolic activation of (2E)-hexadecenoic acid as proposed by the group of Kihara.18 Consequently, we have cotreated cell lysates with the deuterated aldehyde and



CONCLUSION We have developed a sensitive and specific isotope-dilution HPLC-ESI-QTOF method for simultaneous determination of (2E)-hexadecenal and its fatty acid metabolites following derivatization with DAIH and EDC. Studies with cell lysates treated with unlabeled or deuterated (2E)-hexadecenal revealed a major metabolic route including an almost exclusive oxidation of (2E)-hexadecenal to (2E)-hexadecenoic acid with a subsequent activation of the latter by acyl-CoA synthetases. Our method will be instrumental to investigate the physiological function of the critical enzymes in S1P degradation, S1PL and FALDH, in a quantitative manner. Likewise, the presented analytical approach will be useful to study diseases known for abnormalities in long-chain fatty acid metabolism, e.g., the Sjögren-Larsson syndrome, and intox9072

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073

Analytical Chemistry

Article

(25) Keller, M. A.; Watschinger, K.; Lange, K.; Golderer, G.; WernerFelmayer, G.; Hermetter, A.; Wanders, R. J.; Werner, E. R. J. Lipid Res. 2012, 53, 1410−1416. (26) Rizzo, W. B.; Heinz, E.; Simon, M.; Craft, D. A. Biochim. Biophys. Acta 2000, 1535, 1−9.

ications with certain fusarium toxins (e.g., fumonisin B1) known to impair the sphingolipid metabolism, in more detail.



AUTHOR INFORMATION

Corresponding Author

*Phone: +49 33200-885301. Fax: +49 33200-885541. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This study was financially supported by the Deutsche Forschungsgemeinschaft (KL988/7-1).



REFERENCES

(1) Fyrst, H.; Saba, J. D. Nat. Chem. Biol. 2010, 6, 489−497. (2) Gangoiti, P.; Camacho, L.; Arana, L.; Ouro, A.; Granado, M. H.; Brizuela, L.; Casas, J.; Fabriás, G.; Abad, J. L.; Delgado, A.; GómezMuñoz, A. Prog. Lipid Res. 2010, 49, 316−334. (3) Ogretmen, B.; Hannun, Y. A. Nat. Rev. Cancer 2004, 4, 604−616. (4) Spiegel, S.; Milstien, S. Nat. Rev. Mol. Cell Biol. 2003, 4, 397−407. (5) Tabasinezhad, M.; Samadi, N.; Ghanbari, P.; Mohseni, M.; Saei, A. A.; Sharifi, S.; Saeedi, N.; Pourhassan, A. J. Cancer Res. Ther. 2013, 9, 556−563. (6) Gault, C. R.; Obeid, L. M.; Hannun, Y. A. Adv. Exp. Med. Biol. 2010, 688, 1−23. (7) Colié, S.; Van Veldhoven, P. P.; Kedjouar, B.; Bedia, C.; Albinet, V.; Sorli, S. C.; Garcia, V.; Djavaheri-Mergny, M.; Bauvy, C.; Codogno, P.; Levade, T.; Andrieu-Abadie, N. Cancer Res. 2009, 69, 9346−9353. (8) Newbigging, S.; Zhang, M.; Saba, J. D. Gene Expression Patterns 2013, 13, 21−29. (9) Vogel, P.; Donoviel, M. S.; Read, R.; Hansen, G. M.; Hazlewood, J.; Anderson, S. J.; Sun, W.; Swaffield, J.; Oravecz, T. PLoS One 2009, 4, No. e4112. (10) Schwab, S. R.; Pereira, J. P.; Matloubian, M.; Xu, Y.; Huang, Y.; Cyster, J. G. Science 2005, 309, 1735−1739. (11) Upadhyaya, P.; Kumar, A.; Byun, H. S.; Bittman, R.; Saba, J. D.; Hecht, S. S. Biochem. Biophys. Res. Commun. 2012, 424, 18−21. (12) De Laurenzi, V.; Rogers, G. R.; Hamrock, D. J.; Marekov, L. N.; Steinert, P. M.; Compton, J. G.; Markova, N.; Rizzo, W. B. Nat. Genet. 1996, 12, 52−57. (13) Lüth, A.; Neuber, C.; Kleuser, B. Anal. Chim. Acta 2012, 722, 70−79. (14) Püttmann, M.; Krug, H.; von Ochsenstein, E.; Kattermann, R. Clin. Chem. 1993, 39, 825−832. (15) USFDA. Guidance for Industry, Bioanalytical Method Validation; US Department of Health and Human Services, FDA, Center for Drug Evaluation and Research and Center for Veterinary Medicine: Rockville, MD, 2001. (16) Fayyaz, S.; Henkel, J.; Japtok, L.; Krämer, S.; Damm, G.; Seehofer, D.; Püschel, G. P.; Kleuser, B. Diabetologia 2014, 57, 373− 382. (17) Van Veldhoven, P. P. Methods Enzymol. 2000, 311, 244−254. (18) Nakahara, K.; Ohkuni, A.; Kitamura, T.; Abe, K.; Naganuma, T.; Ohno, Y.; Zoeller, R. A.; Kihara, A. Mol. Cell 2012, 46, 461−471. (19) Dalle-Donne, I.; Carini, M.; Orioli, M.; Vistoli, G.; Regazzoni, L.; Colombo, G.; Rossi, R.; Milzani, A.; Aldini, G. Free Radical Biol. Med. 2009, 46, 1411−1419. (20) Horikawa, R.; Tanimura, T. Anal. Lett. 1982, 15, 1629−1642. (21) Milic, I.; Hoffmann, R.; Fedorova, M. Anal. Chem. 2013, 85, 156−162. (22) Ohkuni, A.; Ohno, Y.; Kihara, A. Biochem. Biophys. Res. Commun. 2013, 442, 195−201. (23) Kelson, T. L.; Secor McVoy, J. R.; Rizzo, W. B. Biochim. Biophys. Acta 1997, 1335, 99−110. (24) Rizzo, W. B.; Craft, D. A. J. Clin. Invest. 1991, 88, 1643−1648. 9073

dx.doi.org/10.1021/ac501677y | Anal. Chem. 2014, 86, 9065−9073