Anal. Chem. 2005, 77, 3383-3389
Analysis of Derivatized Biogenic Aldehydes by LC Tandem Mass Spectrometry Taufika Islam Williams,† Mark A. Lovell,†,‡ and Bert C. Lynn*,†,‡
Department of Chemistry, University of Kentucky, Lexington, Kentucky 40506, and Sanders-Brown Center on Aging, University of Kentucky, Lexington, Kentucky 40356
Lipid peroxidation has been linked to the etiology of several diseases, including Alzheimer’s disease (AD). End products of this phenomenon include low molecular weight, water-soluble aldehydes, compounds that covalently modify proteins and nucleic acids, thereby altering function. Aliphatic aldehydes (C3-C10) are generated during lipid peroxidation, along with r,β-unsaturated aldehydes, including acrolein and 4-hydroxynonenal (HNE). The Hantzsch reaction was used to produce heterocyclic aldehyde derivatives that can be conveniently analyzed with mass spectrometry. Liquid chromatographic analyses revealed increasing retention times from derivatized methanal to octanal. HNE derivatives were observed to elute between heptanal and octanal derivatives, while the acrolein derivatives had a retention time similar to the propanal derivative. Smaller aliphatic aldehyde derivatives fragmented in a similar manner to produce a base peak of m/z 273, while the larger derivatives yielded m/z 274 as the base peak. Acrolein and HNE derivatives fragmented in a slightly different manner compared to their aliphatic counterparts. Calibration plots of aliphatic and unsaturated aldehydes were linear (r2 g 0.99) in the concentration range explored (∼5-1500 pg on column). The LC-MS/MS methodology developed here will be used in subsequent studies to determine aldehyde concentrations for comparing age-matched controls to AD tissues from human subjects. Lipid peroxidation (LPO), a process that generates low molecular weight, water-soluble aldehydes, is commonly observed in several different types of neurodegenerative conditions, such as stroke and Alzheimer’s disease (AD).1 Compared to other organ systems, the brain contains a high lipid content and relatively poor defenses against LPO.2 Neuronal membranes readily succumb to oxidative damage because the polyunsaturated fatty acid (PUFA) constituent of their phospholipids can easily react with free radicals from cell metabolism.3,5 Ordinarily, the brain is protected from such damage by a closely controlled balance between * Corresponding author. Phone: 859-257-2300 ext. 287. Fax: 859-257-2489. E-mail:
[email protected]. † Department of Chemistry. ‡ Sanders-Brown Center on Aging. (1) Markesbery, W. R.; Lovell, M. A. Neurobiol. Aging 1998, 19, 33-36. (2) Markesbery, W. R. Free Radic. Biol. Med. 1997, 23, 134-147. (3) O′Brioen-Coker, I. C.; Mallet, G. P. A. I. Rapid Commun. Mass Spectrom. 2001, 15, 920-928. 10.1021/ac048265+ CCC: $30.25 Published on Web 04/19/2005
© 2005 American Chemical Society
prooxidant and antioxidant mechanisms. Oxidative stress results from an imbalance in free radical production and their removal by antioxidants.3 Recently, there has been increased interest in biogenic aldehydes as potential biomarkers for oxidative stress and associated ailments. Saturated aliphatic aldehydes (C3-C10) are generated in LPO, along with several R,β-unsaturated aldehydes such as acrolein and 4-hydroxynonenal (HNE, a product of ω-6 PUFA peroxidation).4 Aliphatic aldehydes have no evident toxicity, but acrolein (CH2d CHCHO) and HNE (C5H11sCH(OH)CHdCHCHO) are quite neurotoxic and may be associated with AD development. 4-Hydroxyalkenals are strong electrophiles that react in tissues by alkylating nucleophiles, such as sulfhydryl, amino, and imidazole groups.6 Acrolein is produced in vivo during metal-catalyzed oxidation of PUFAs, including arachidonic acid.7 It is the strongest electrophile of the unsaturated biogenic aldehydes and exhibits the most reactivity toward nucleophilic species.8 Liquid chromatography (LC) separations, coupled to a variety of detection methods, have been employed in the analysis of derivatized biogenic aldehydes.9-17 In cases where analysis was carried out by gas chromatography (GC), the detection method of choice has been mass spectrometry (MS).13,18-20 For instance, (4) Lovell, M. A.; Markesbery, W. R. In Methods in Biological Oxidative Stress; Hensley, K., Floyd, R. S., Eds.; Humana Press: Totowa, NJ, 2003; pp 1721. (5) Porter, N. A.; Lehman, L. S.; Weber, B. A.; Smith, K. J. J. Am. Chem. Soc., 1981, 103, 6447-6455. (6) Spies-Martin, D.; Sommerburg, O.; Langhans, C.-D.; Leichsenring, M. J. Chromatogr., B 2002, 774, 231-239. (7) Lovell, M. A.; Xie, C.; Markesbery, W. R. Neurobiol. Aging 2001, 22, 187194. (8) Lovell, M. A.; Xie, C.; Markesbery, W. R. Free Radic. Biol. Med. 2000, 29, 714-720. (9) Cordis, G. A.; Bagchi, D.; Maulik, N.; Das, D. K. J. Chromatogr., A 1994, 661, 181-191. (10) Esterbauer, H.; Schaur, R. J.; Zollner, H. Free Radic. Biol. Med. 1991, 11, 81-128. (11) Draper, H. H.; Csallany, A. S.; Hadley, M. Free Radic. Biol. Med. 2000, 29, 1071-1077. (12) Bu ¨ ldt, A.; Karst, U. Anal. Chem. 1997, 69, 3617-3622. (13) Thomas, M. J.; Robison, T. W.; Samuel, M.; Forman, H. J. Free Radic. Biol. Med. 1995, 18, 553-557. (14) Yoshinko, K.; Matzuura, T.; Sano, M.; Saito, S.; Tomita, T. Chem. Pharm. Bull. 1986, 34, 1694-1700. (15) Holley, A. E.; Walker, M. K.; Cheeseman, K. H.; Slater, T. E. Free Radic. Biol. Med. 1993, 15, 281-289. (16) Bailey, A. L.; Wortley, G.; Southon, S. Free Radic. Biol. Med. 1997, 23, 1078-1085. (17) Matsuoka, M.; Imado, N.; Maki, T. Chromatographia 1996, 43, 501-506. (18) Luo, X. P.; Yazdanpanah, M.; Bhooi, N.; Lehotay, D. C. Anal. Biochem. 1995, 228, 294-298.
Analytical Chemistry, Vol. 77, No. 10, May 15, 2005 3383
Zhang and co-workers30 have recently developed a simple, rapid, sensitive, solvent-free GC/MS method with solid-phase microextraction and on-fiber derivatization for the analysis of aldehyde levels in the blood of lung cancer patients. The blood samples were shown to contain significantly elevated levels of hexanal and heptanal compared to controls, suggesting the potential use of these aldehydes as biomarkers for this disease. There are some noteworthy reports that describe the use of LC in association with mass spectrometry in aldehyde analysis of tissue samples.3,21-24 Several derivatization techniques have been explored, including the Hantzsch reaction.25 In this reaction, an aldehyde, two β-dicarbonyl compounds, and ammonia combine to form a heterocyclic system. The resulting compound, known as a dimedone derivative, has fluorescing properties.25 Several β-dicarbonyl compounds have been studied.26-28 O’Brioen-Coker and Mallet3 developed a LC-MS procedure for analyzing hexanal and larger aldehydes in blood plasma. Aldehyde derivatization using the Hantzsch reaction preceded analysis by a LC-Quattro instrument with electrospray. Quantification was achieved by multiple reaction monitoring, with benzaldehyde as an internal standard. This is a nonphysiological aldehyde of little structural similarity with the target compounds. The approach has the potential drawback that benzaldehyde reactivity in the derivatization may not be identical to that of the target aldehydes. However, experimental data did not suffer and indicated the applicability of LC-MS in plasma aldehyde quantification. Smaller (C1-C5) and unsaturated aldehydes were not studied. Zurek and Karst25 described an LC-MS method for the determination of aliphatic aldehydes following derivatization with acetylacetone through the Hantzsch reaction. The authors employed atmospheric pressure chemical ionization and ESI in the positive ion mode to ionize the dihydropyridine and decahydroacridine derivatives, protonated at their basic secondary amine groups. The oxidation products of the formaldehyde derivatives were identified as side products. The authors concluded that the combination of LC and mass spectrometry offered considerably high selectivity in comparison to the conventional detection approaches. Here, we extend and apply the protocols of Karst et al. and Mallet et al. in developing methodology for analyzing small aliphatic and unsaturated aldehydes of biological importance, especially in AD etiology. In contrast to other biogenic aldehyde studies, 5,5′-dimethyl-1,3-cyclohexanedione (dimethyl CHD) was our derivative of choice due to improved chromatographic quality. (19) Bruenner, B. A.; Jones, A. D.; German, J. B. Anal. Biochem. 1996, 241, 212-219. (20) Spiteller, G.; Kern, W.; Spiteller, P. J. Chromatogr.. A. 1999, 843, 29-98. (21) Sim, A. S.; Salonikas, C.; Naidoo, D.; Wilcken, D. E. L. J. Chromatogr.. B 2003, 785, 337-344. (22) Deighton, N.; Magill, W. L.; Bremner, D. H.; Benson, E. E. Free Radic. Res. 1997, 27, 255-265. (23) Ravandi, A.; Kuksis, A.; Myher, J. J.; Marai, L. J. Biochem. Biophys. Methods 1995, 30, 271-285. (24) Gioacchini, A. M.; Calonghi, N.; Boga, C.; Cappadone, C.; Masotti, L.; Roda, A.; Traldi, P. Rapid Commun. Mass Spectrom. 1999, 13, 1573-1579. (25) Zurek, G.; Karst, U. J. Chromatogr., A 1999, 864, 191-197. (26) Nash, T.; Biochemistry 1953, 55, 416-421. (27) Sawicki, E.; Carnes, R. A. Mikrochim. Acta 1968, 148-159. (28) Compton, B. J.; Purdy, W. C. Can. J. Chem. 1980, 58, 2207-2211. (29) Deng, C.; Zhang, X. Rapid Commun. Mass Spectrom. 2004, 18, 1715-1720. (30) Wenner, B. R.; Lovell, M. A.; Lynn, B. C. J. Proteome Res. 2004, 3, 97-103.
3384
Analytical Chemistry, Vol. 77, No. 10, May 15, 2005
A range of aliphatic aldehydes were studied (C1-C6), along with acrolein and HNE. Furthermore, two internal standards were employed (heptanal and octanal) to accurately quantify low- and high-abundance aldehydes. The influence of several instrument parameters on the quality of mass spectral data is explored, and a description of how their values were optimized is provided. We detail the development of sensitive, selective, and dependable methods to identify and quantify biogenic aldehydes in tissues by LC-MS/MS. EXPERIMENTAL SECTION Chemicals. All aliphatic aldehydes except methanal (Sigma, St. Louis, MO) and ethanal (EM Industries, Inc., Hawthorne NY) were obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). HNE was obtained from Cayman Chemical Co. (Ann Arbor, MI). Acrolein, 1,3-cyclohexanedione (CHD), 5-methyl-1, 3-cyclohexanedione (methyl CHD), and 5,5-dimethyl-1,3-cyclohexanedione (dimethyl CHD) came from Aldrich (St. Louis, MO). LC grade water, acetonitrile (ACN), methanol, glacial acetic acid, PrepSep C18 solid-phase extraction (SPE) cartridges, and a PrepSep 12-port vacuum manifold were obtained from Fisher Scientific (Pittsburgh, PA). Ammonium acetate came from Sigma. Solutions for aldehyde quantification were prepared with Wiretrol disposable micropipets from VWR Scientific Products (St. Paul, MN). Brain Tissue. Specimens of superior and middle temporal gyrus (SMTG) from AD patients and age-matched control subjects were obtained from short postmortem interval autopsies, immediately frozen in liquid nitrogen, and subsequently stored at -80°C until used for analysis. Derivatization of Aldehydes. Derivatizing Agent. Aldehydes were derivatized by heating them with appropriate amounts of a stock dimethyl CHD derivatizing agent solution, as described below. This solution was prepared by combining 10 g of ammonium acetate, 10 mL of glacial acetic acid, and 0.25 g of dimethyl CHD and diluting to 100 mL with water.4 CHD and methyl CHD derivatizing solutions were prepared in a similar fashion. Standard Aldehyde Derivatives for Quantitative Analysis. For the aliphatic aldehydes (C1-C6), a standardized aldehyde solution mixture was prepared by adding precalculated, weighed amounts of stock aliphatic aldehydes (C1-C6) to a class A volumetric flask and diluting with methanol. A separate solution of heptanal internal standard was prepared in a similar manner. A set of standardized solutions with increasing concentration of aldehydes (∼1 × 10-9-∼3 × 10-7g/mL) was prepared in volumetric glassware by adding suitable amounts of the aliphatic aldehyde mixture, an excess of derivatizing agent, a fixed amount of internal standard and diluting to the mark with methanol. The aldehyde solutions, upon derivatization, were calculated to provide 1-300 pg/µL. The standardized solutions were incubated in a 60 °C water bath for 1 h and subsequently desalted with C18 SPE cartridges. These derivatized aldehydes were then eluted with 2 mL of methanol (Figure 1). For the unsaturated aldehydes (acrolein and HNE), standardized solutions of acrolein and HNE were prepared in a similar manner as described above for the aliphatic species. The internal standard employed here was octanal (Figure 1).
Figure 1. Cyclization of two β-dicarbonyl compounds and an aldehyde in the presence of ammonia.25
Brain Tissue. For aldehyde analysis, 100 mg of tissue sample was homogenized with 5 mL of HEPES buffer (pH 7.4) with 137 mM NaCl, 4.6 mM KCl, 1.1 mM KH2PO4, 0.6 mM MgSO4, pepstatin (0.7 µg/mL), leupeptin (0.5 µg/mL), aprotinin (0.5 µg/mL), and phenylmethylsulfonyl fluoride (40 µg/mL). The 500-µL aliquots of homogenate were added to glass test tubes, followed by appropriate volumes of standardized internal standard solution (prepared in methanol). Heptanal was chosen as the internal standard for the aliphatic and octanal for the unsaturated aldehydes. Octanal internal standard was added at a 10-fold lower concentration compared to heptanal so that solvent concentration of the final derivatized brain sample by 10 enabled accurate quantification of low-abundance unsaturated aldehydes. A review of the literature on aldehyde quantification in the brain indicates that this is adequate for acrolein and HNE analysis.1,4,7 Samples were vortexed for 30 s for aldehyde extraction and centrifuged (850g for 10 min). A 500-µL aliquot of supernatant was then mixed with 1 mL of dimethyl CHD derivatizing reagent and heated in a 60 °C water bath for 1 h. After cooling to room temperature, the reaction mixture was added to preconditioned C18 SPE columns, which were then washed with water to remove excess ammonium acetate. Derivatized aldehydes were eluted with methanol.4 Instrumentation. A Hewlett-Packard LC system (1100 series) with a laboratory-constructed C18 reversed-phase capillary LC column30 was employed in the separation of aldehyde derivatives for quantitative analysis. Gradient elution using a mobile-phase mixture of water and ACN, each containing 0.1% formic acid, was used in all experiments. For the first 2 min of a given run, the mobile-phase composition was kept at 5% ACN. In the next 8 min, the proportion of ACN was linearly increased to 20%. Another linear increase in ACN occurred for 20 min more to reach a concentration of 70%. In the following 20 min, the ACN was linearly changed to 90%. During the last 5 min of the run, ACN content rose, once again linearly, to a maximum concentration of 95%. The column was restored to starting mobile-phase conditions prior to beginning the next run. LC solvent flow rate was maintained at 4 µL/min throughout the course of each chromatographic run using a laboratoryconstructed 50 to 1 splitter between the pump and injector. Manual injections were performed with a six-port Rheodyne injection valve (model 7725i) equipped with a 5-µL sample loop. A Finnigan LCQ Classic quadrupole ion trap mass spectrometer was interfaced to the LC system for reversed-phase LC-MS/ MS analyses. Samples with derivatized aldehydes were introduced into the LCQ by LC separation or direct infusion.
Figure 2. Ion intensity vs % NCE for (a) dimethyl CHD heptanal and (b) dimethyl CHD acrolein.
Tuning LCQ Parameters. Dimethyl CHD heptanal was infused into the LCQ mass spectrometer, and instrument parameters, such as lens offsets, capillary voltages, etc., were tuned for this compound to improve the quality of LC-MS/MS data obtained for aldehyde derivatives. Electrospray source parameters were tuned to the following values: 4 kV applied voltage, 20 arbitrary units of sheath gas flow, 45 V inlet capillary voltage, and 175 °C inlet capillary temperature. RESULTS AND DISCUSSION Initial experiments were performed to determine retention times for the aldehydes derivatives. In data-dependent scanning, the instrument acquires a full scan mass spectrum, evaluates the relative intensities of ions in a given mass range, and then isolates the ion with the highest intensity for fragmentation and analysis (MS/MS mode). Masses utilized as precursors for MS/MS experiments were placed on an exclusion list and were not sampled again for a preselected time window. CHD, methyl CHD, Analytical Chemistry, Vol. 77, No. 10, May 15, 2005
3385
and dimethyl CHD derivatives of standard aldehyde solutions, and several brain tissue samples with derivatized aldehydes (SMTG), were analyzed in this manner. Apart from its simplicity and compatibility with LC-MS/MS analyses, the Hantzsch reaction offers near 100% efficiencies, which is well suited for these studies.16 After determining the retention times for each derivatized aldehyde peak, method files for segmented analysis were constructed to perform MS/MS analysis on the aldehyde derivatives. In segmented analysis, acquisition time is divided into segments for exclusive MS/MS analysis of particular analytes. This offers a dramatically improved signal-to-noise ratio, and the isolation of the desired species in this manner for exclusive MS/MS analysis represents a form of ion purification. As described in the Experimental Section and in the paragraphs that follow, several measures were taken to improve data for derivatized aldehydes from such experiments. Parameters for resonant CID, such as percent normalized collision energy (% NCE) and CID time were explored for spectral data improvement. The collision energy needed to achieve optimum fragmentation has a mass dependency, which is compensated by the normalized collision energy used with the LCQ. CID time is the time during which the resonant rf signal is applied to the end cap electrodes for ion excitation. Infusion studies measuring ion intensity as a function of NCE and CID time were performed to determine suitable conditions for segmented analysis for aldehyde derivatives. Aldehyde derivative solutions (nanograms per microliter) were introduced into the LCQ at a rate of 3 µL/min. NCE and CID time were separately ramped while acquiring data on precursor and product ion intensities. For the ion intensity versus NCE plot of dimethyl CHD heptanal (Figure 2a), conversion of precursor to products at higher collision energies was ∼80%. A similar situation was observed when varying ion intensity with CID time for this species. At 38% NCE and 30-ms CID time, most of the precursor ion current was converted into products, with a little precursor species remaining intact. Indeed, all the aliphatic aldehydes provided similar responses. In the case of ion intensity versus NCE for dimethyl CHD acrolein, ∼50% of the precursor ion current was converted into products (Figure 2b), while in the case of the HNE derivative, only a third of the precursor ion current resulted in product formation (data not shown). It is possible that some major products are below the low-mass cutoff of the ion trap and are therefore not detected. All aldehyde derivatives required similar NCE (38%) and CID times (30 ms). Using optimized parameters for aldehyde derivatives in the MS methods resulted in >95% reduction in precursor ion counts and subsequent conversion to products. Data from Segmented Analyses. CHD, methyl CHD, and dimethyl CHD aldehydes were studied using segmented analysis. The dimethyl CHD-derivatized aldehydes offered the best data in terms of chromatographic peak quality. A possible explanation for this observation is that the greater hydrophobic nature of the dimethyl CHD derivatives promoted increased chromatographic focusing at the head of the column, thereby reducing peak width and improving the overall separation process. As such, this derivative was used exclusively for all further experiments. 3386 Analytical Chemistry, Vol. 77, No. 10, May 15, 2005
Figure 3. Segmented chromatograms of (a) dimethyl CHD aliphatic and (b) dimethyl CHD unsaturated aldehydes.
Chromatographic Data. The aliphatic aldehyde derivatives displayed increasing retention times with increasing mass of aldehyde derivative (Figure 3a) in good agreement with the rise in hydrophobicity associated with greater mass. Reconstructed ion chromatograms for dimethyl CHD methanal to dimethyl CHD heptanal formed a stair-step of peaks. For the unsaturated aldehydes, dimethyl CHD acrolein had a shorter retention time than dimethyl CHD HNE, as was expected due to the increased mass of the latter (Figure 3b). Indeed, the acrolein and propanal derivatives showed similar retention times, in agreement with their similar masses. Keeping with this hydrophobicity trend, the internal standard for unsaturated aldehydes, dimethyl CHD
Table 1. Dimethyl CHD Aldehyde Derivatives neutral loss from fragmentation
Figure 4. MS/MS spectra of (a) dimethyl CHD butanal and (b) dimethyl CHD unsaturated aldehydes.
octanal, had the longest retention time of all. The R group (Figure 1) of dimethyl CHD HNE has a hydrophilic OH group at the γ position from the dimedone ring, which reduces its hydrophobicity compared to aliphatic derivatives of similar size. So, this species has a shorter retention time than dimethyl CHD heptanal, which has a 6-carbon chain and no hydroxyl moiety. Mass Spectrometric Data. The aliphatic aldehyde derivatives fragmented in a characteristic pattern to produce a base peak of either m/z 273 or 274, depending on the size of the R group derived from the precursor aldehyde. The MS/MS spectrum of dimethyl CHD butanal is shown in Figure 4a. For smaller aldehyde derivatives (C1-C4), the appearance of m/z 273 was consistent with a simple C-C bond dissociation between the dimedone group and the R group, followed by neutral loss of a saturated R radical. For larger aldehyde derivatives (C5-C8), a hydrogen rearrangement followed by cleavage of the (R-H) group became more feasible. For these derivatives, the C-H bond dissociation energies of 2° carbon atoms in the R group were sufficiently low enough for a base peak of m/z 274 to be generated as a result of
aldehyde
MW
parent mol ion of deriv, [M + H]+
methanal ethanal propanal butanal pentanal hexanal heptanal octanal acrolein HNE
30 44 58 72 86 100 114 128 56 156
274 288 302 316 330 344 358 372 300 400
daughter ion from fragment, m/z
fragment
MW
273 273 273 273 274 274 274 274 272 382, 339
CH3 C2H5 C3H7 C4H8 C5H10 C6H12 C7H14 C2H4 H2O, C3H7
15 29 43 56 70 84 98 28 18, 43
the loss of an unsaturated (R-H) species.31 For all the derivatives, the major bond cleavage was between the R group and the heterocyclic ring, with the latter retaining the charge at the nitrogen atom. Dimethyl CHD acrolein fragmented to produce a base peak at m/z 272, which appeared to be the loss of ethylene from the precursor species. In the case of dimethyl CHD HNE, the base peak was at m/z 339 but a significant peak was also observed at m/z 382 (Figure 4b). It is probable that m/z 382 was produced from the loss of water with a subsequent loss of 43 amu (equivalent in mass to a propyl radical), resulting in the formation of the m/z 339 species. Mass spectrometric data for all aldehyde derivatives are summarized in Table 1. Construction of Calibration Plots. Prior to proceeding with quantitative analysis, standard dimethyl CHD heptanal solutions of increasing concentrations were analyzed by segmented LC-MS/MS to determine the linearity of a plot of area counts versus concentration. Each solution was analyzed in triplicate to obtain standard deviation data. The plot was linear and so the heptanal derivative was a good choice for an internal standard in these experiments. Additionally, neither heptanal nor octanal is present in brain tissues to any significant concentration, minimizing concerns regarding contributions of these aldehydes to internal standard peak areas.4 Due to similarity in structure, the octanal derivative produced a similar response. For each of the standardized solutions of increasing concentration prepared from the aliphatic and unsaturated aldehyde derivatives, segmented analysis was performed in triplicate to obtain standard deviation data. The area counts of aldehyde derivative X was divided by the area counts of internal standard to obtain the relative response factor (RRF) of aldehyde derivative X with respect to its internal standard. Calibration plots were constructed by plotting the RRF of each aldehyde derivative versus the number of picograms on column delivered for each LC-MS/MS experiment. The aliphatic aldehyde derivatives produced linear calibration plots with r2 g 0.99 values and relative standard deviations (RSDs) of ∼e5%. The RRF values for these derivatives were comparable, with a slight trend toward lower RRFs with increasing aldehyde derivative mass, in agreement with their similarity in structure. The ethanal derivative showed a somewhat higher response factor (31) McLafferty, F. W.; Turee`ek, F. Interpretation of Mass Spectra, 4th ed.; University Science Books: Mill Valley, CA, 1993.
Analytical Chemistry, Vol. 77, No. 10, May 15, 2005
3387
Table 2. Calibration Curve Data for Dimethyl CHD Aldehydes aldehydes
r2
methanal
0.997
ethanal
0.999
propanal
0.995
butanal
0.996
pentanal
0.996
hexanal
0.996
acrolein
0.999
HNE
0.996
a
pg on col ald deriv
RRFa
RSDb, %
11.3 113 226 564 1128 1804 8.5 67 167 334 501 835 1169 7.1 57 143 286 429 716 1002 5.9 59 118 294 588 940 1175 6.1 61 122 306 612 979 1224 6.5 65 129 323 646 1033 1292 6.5 52 130 259 389 648 907 6.4 51 128 256 384 639 895
0.65 1.40 1.85 2.94 5.03 7.41 0.31 0.76 1.24 2.03 2.87 4.48 6.23 0.28 0.47 0.92 1.36 1.74 2.63 3.50 0.05 0.12 0.25 0.91 1.80 2.89 3.95 0.06 0.13 0.21 0.94 1.73 3.13 3.72 0.02 0.12 0.18 0.76 1.67 2.98 3.82 0.19 0.29 0.61 1.06 1.46 2.22 3.10 0.01 0.03 0.05 0.08 0.09 0.16 0.22
3.72 3.10 3.06 2.67 2.68 1.67 5.27 4.74 3.70 1.85 2.51 2.85 2.87 5.31 5.93 4.35 2.87 2.70 2.95 2.90 4.71 3.36 1.00 3.41 2.88 2.13 2.01 5.82 5.35 3.68 3.11 2.66 1.72 2.94 4.77 4.83 4.56 3.33 0.94 2.94 1.97 5.29 5.92 4.87 4.80 3.74 4.84 2.33 5.79 5.39 4.56 3.56 3.42 3.90 3.46%
Relative response factor. b Relative standard deviation.
compared to the others. The greatest RRFs would probably have been expected of the methanal derivative, which has the smallest mass. It should be noted, however, that methanal, which is a 37% stock solution, can exist in the form of other species, such as paraformaldehyde, which may not derivatize efficiently through the Hantzsch reaction or perhaps form different derivatives. This was evidenced by the observation of several “extraneous” peaks at earlier retention times in the reconstructed ion chromatogram for dimethyl CHD methanal. As such, it may be that a somewhat 3388 Analytical Chemistry, Vol. 77, No. 10, May 15, 2005
Table 3. Comparison of Aldehyde Concentrations in the Brains (SMTG) of AD vs Control Subjects
aldehyde
subject with AD (nmol/mg of protein) mean ( SEM
age-matched control (nmol/mg of protein) mean ( SEM
methanal ethanal propanal butanal pentanal hexanal acrolein HNE
133 ( 20 173 ( 21 84.9 ( 14.5 6.29 ( 1.09 9.99 ( 1.56 56.9 ( 5.7 1.45 ( 0.21 0.550 ( 0.106
290 ( 39 175 ( 27 91.0 ( 14.7 8.42 ( 0.93 11.4 ( 2.0 13.8 ( 1.7 0.475 ( 0.077 0.209 ( 0.039
lower proportion of the methanal (compared to the calculated value) is actually partaking in the derivatization process to form the desired heterocyclic derivative. In other words, if the stock methanal solution was indeed 37% “pure methanal”, then this aldehyde derivative would probably have produced the highest RRF values compared to other aliphatic aldehyde derivatives. Calibration plots for unsaturated aldehyde derivatives were linear and with r2 and RSD values comparable to those of their aliphatic counterparts. The HNE derivative had a much lower RRF than the acrolein derivative, in agreement with infusion studies data (not shown). It should be noted that the area counts versus % NCE infusion experiment for dimethyl CHD HNE generated product ion curves at much lower intensity values (about a third of the initial ion intensity of the precursor species at low % NCE) compared to the acrolein derivative. Data from calibration plots for all the aldehyde derivatives are given in Table 2. These calibration plots can be employed in the analysis of real tissue that has been subjected to oxidative stress for aldehyde content determination. As an example of how this methodology may be applied toward this purpose, brain samples (SMTG) from an AD subject and a suitable age-matched control were analyzed in triplicate as described above. Data from these findings are summarized in Table 3. Aldehyde content per milligram of protein from the brain tissue was reported as the mean value ( standard error of the mean (SEM). The AD sample showed similar levels of most aliphatic aldehydes and higher levels of unsaturated aldehydes, compared to the control. It is worth pointing out that AD development has been related to an increase in the production of unsaturated aldehydes, in agreement with these findings. The results demonstrate that the protocol outlined here can be effectively applied in aldehyde quantification in real tissue samples. CONCLUSIONS These studies aimed to develop sensitive, selective, and reliable methods to analyze aldehydes in tissues using LC-MS/MS. It was shown that aldehydes typically found in brain tissues plagued by oxidative stress can be derivatized through the Hantzsch reaction to form cyclohexanedione derivatives, which can easily be quantified by MS. The dimethyl CHD derivatives offered better peak quality due to their increased hydrophobicity, which resulted in greater chromatographic focusing of the sample at the head of the column. Analyte derivatization coupled with LC-MS/MS provided increased sensitivity and selectivity over contemporary UV methods for the analysis of low-abundance biogenic aldehydes. Unsaturated aldehydes occur in the brain at much lower concen-
trations than their aliphatic counterparts, and LC-MS/MS proved to be an ideal method for analyzing these two classes of compounds using two internal standards differing in concentration by an order of magnitude. The developed method was applied to the analysis of samples of human brain (SMTG) from an AD patient and an age-matched control subject. In comparison to the control sample, the tissue obtained from the AD subject displayed similar concentrations of most aliphatic aldehydes; however, significantly elevated levels of acrolein and HNE were detected. These results confirm that the protocol described here can be reliably applied in the quantification of aldehydes in brain tissue. In the future, concentration variations of aliphatic and unsaturated aldehydes will be investigated by exploring a larger data set of AD and control
subjects, as well as by investigating several different regions of the brain, based on the level of AD pathology (study in progress). ACKNOWLEDGMENT These authors thank the University of Kentucky Mass Spectrometry Facility (www.rgs.uky.edu/ukmsf) and the SandersBrown Center on Aging for laboratory resources. This work was supported by the National Institute of Health Grants 5P50-AG05144 and 1P01-AG05119 and a grant from the Abercrombie Foundation.
Received for review November 23, 2004. Accepted March 18, 2005. AC048265+
Analytical Chemistry, Vol. 77, No. 10, May 15, 2005
3389
