Imaging of Lipids in Atheroma by Desorption Electrospray Ionization

Oct 5, 2009 - lipid composition can vary among atheroma, even within a single ... Finally, cho- lesterol was imaged in the atheroma by doping the char...
0 downloads 0 Views 4MB Size
Download PDF
Anal. Chem. 2009, 81, 8702–8707

Imaging of Lipids in Atheroma by Desorption Electrospray Ionization Mass Spectrometry Nicholas E. Manicke,†,‡ Marcela Nefliu,†,| Chunping Wu,†,‡ John W. Woods,§ Vladimir Reiser,§ Ronald C. Hendrickson,§ and R. Graham Cooks*,†,‡ Department of Chemistry, Purdue University, West Lafayette, Indiana 47907, Bindley Biosciences Center, West Lafayette, Indiana 47907, and Merck, Rahway, New Jersey 07065 One of the hallmarks of atherosclerosis is the accumulation of lipoproteins within the wall of blood vessels. The lipid composition can vary among atheroma, even within a single individual, and is also dynamic, changing as the lesion progresses. One desirable characteristic of atheroma is their stability, as the rupture of unstable plaques can interfere with normal blood flow to the brain or heart, leading to stroke or heart attack. Desorption electrospray ionization mass spectrometry (DESI-MS) was used in this study for the profiling and imaging of arterial plaques. DESI-MS is an ambient ionization method in which a charged, nebulized solvent spray is directed a surface. In the positive and negative ion modes, sodium and chloride adducts, respectively, of diacyl glycerophosphocholines (GPChos), sphingomyelins (SMs), and hydrolyzed GPChos were detected. Also, cholesteryl esters were detected via adduct formation with ammonium cations. Finally, cholesterol was imaged in the atheroma by doping the charge labeling reagent betaine aldehyde directly into the DESI solvent spray, leading to in situ chemical derivatization of the otherwise nonionic cholesterol. DESI imaging experiments, in which the spatial distribution of the various chemical species is determined by scanning the DESI probe across an entire sample surface, revealed that there are lipid rich regions within the arterial walls, and the lipid rich regions seem to have one of two different lipid profiles. These lipid rich regions likely correspond to the areas of the tissue where lipoprotein particles have accumulated. It is also possible that the different lipid distributions may correlate with the stability or vulnerability of that particular region of the plaque. Atherosclerosis is the major source of mortality in the developed world.1 The major clinical events (such as myocardial infarction and stroke) resulting from atherosclerosis are precipitated by plaque rupture. The propensity of atherosclerotic plaques to rupture may be influenced by their lipid content and the distribution of these lipids within the plaque.2,3 Differences in both the content of specific lipids and the percentage of three important * To whom correspondence should be addressed. E-mail: [email protected]. † Purdue University. ‡ Bindley Biosciences Center. | Current address: Merck, West Point, Pennsylvania 19486. § Merck. (1) Breslow, J. L. Nature Medicine 1997, 3, 600–601.

8702

Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

components present in the plaque (cholesterol, cholesteryl esters, and phospholipids) have been detected among the different types of plaque, although the specific values have been found to differ inter- and intraplaque for different stages of progression.4 For example, free cholesterol concentrations are higher at the center of disrupted plaques compared to intact plaques, whereas at the edge of disrupted plaques, the free-to-esterified cholesterol ratio was lower because of the accumulation of esterified cholesterol.5 These data suggest that temporal changes in the chemical composition of plaques during disease progression may be used to discriminate stable plaques from vulnerable ones by identifying morphologic and molecular markers before clinical syndromes develop. Selected oxidants, oxidation products, and enzymes shown to be involved in the development of atherosclerosis have already been examined as potential biomarkers.6 A number of invasive and noninvasive imaging methods are currently used to study atherosclerosis. Most of the standard techniques (including X-ray angiography, angioscopy, and intravascular ultrasound) identify luminal diameter, abnormal narrowing of blood vessels, wall thickness, and plaque volume but are ineffective in identifying the high-risk plaques that are vulnerable to rupture and thrombosis. In vivo, high-resolution, multicontrast magnetic resonance imaging (MRI) holds the best promise of noninvasively imaging high-risk plaques.7 Fourier transform infrared (FTIR) imaging8 as well as direct microscopy following tissue staining with antibody and fluorescent probes9,10 have been used for lipid localization in atherosclerotic lesions. Mass spectrometry imaging techniques, including secondary ion mass spectrometry (SIMS)11-14 and matrix assisted laser (2) Libby, P.; Ross, R. Cytokines and Growth Regulatory Molecules; LippincottRaven: Philadelphia, 1996. (3) Ross, R. N. Engl. J. Med. 1999, 340, 115–126. (4) Marinello, E.; Setacci, C.; Giubbolini, M.; Cinci, G.; Frosi, B.; Porcelli, B.; Terzuoli, L. Life Sci. 2003, 72, 2689–2694. (5) Felton, C. V.; Crook, D.; Davies, M. J.; Oliver, M. F. Arterioscler. Thromb. Vasc. Biol. 1997, 17, 1337–1345. (6) Tsimikas, S. In Oxidative Biomarkers in the Diagnosis and Prognosis of Cardiovascular Disease, Symposium on Emerging Evidence for Pleiotropic Effects of Statins, New York, NY, May 06, 2005; Excerpta Medica Inc: New York, 2005; pp 9P-17P. (7) Fayad, Z. A.; Fuster, V. Circ. Res. 2001, 89, 305–316. (8) Colley, C. S.; Kazarian, S. G.; Weinberg, P. D.; Lever, M. J. Biopolymers 2004, 74, 328–335. (9) Bugelski, P. J.; Maleeff, B. E.; Klinkner, A. M.; Louden, C. S.; Hart, T. K. Microsc. Microanal. 2000, 6, 532–541. (10) Bobryshev, Y. V.; Lord, R. S. A. Atherosclerosis 1999, 142, 121–131. (11) Todd, P. J.; McMahon, J. M.; Short, R. T.; McCandlish, C. A. Anal. Chem. 1997, 69, A529–A535. 10.1021/ac901739s CCC: $40.75  2009 American Chemical Society Published on Web 10/05/2009

desorption/ionization(MALDI),15,16 can provide chemical images of tissue with great sensitivity, specificity (through added MS/ MS and high resolution), and spatial resolution. They are invasive techniques that cannot be applied in clinical settings for in vivo testing but have value for fundamental research. More recently, attention has been given to imaging by DESI-MS, in which a pneumatically assisted stream of charged microdroplets is directed at a surface in the ambient environment.17,18 A thin liquid film collects on the surface, and the impact of incoming droplets on this thin film causes the ejection of small secondary droplets containing the dissolved analyte.19,20 Reports in the literature include examples of imaging of lipids, drugs, and drug metabolites from thin tissue sections.21-24 More recently, the sensitivity and selectivity of DESI toward cholesterol and other biological alcohols have been improved using a selective charge-labeling reagent,25 an approach related to that employed previously for the derivatization of diacylglycerols.26 In this experiment, called reactive DESI, the reagent betaine aldehyde was doped into the DESI spray and used to derivatize cholesterol in situ without additional preparation steps. In this study, we use DESI-MS for chemical profiling and imaging of plaque tissue in ambient conditions using both normal DESI conditions and reactive DESI for detection of cholesterol. METHODS DESI-MS. The DESI source utilized in these studies was described in detail elsewhere.27 Briefly, a fused silica capillary with a 50 µm inner diameter (i.d.) and 150 µm outer diameter (o.d.) (Polymicro Technologies, AZ) was used for delivering the spray solvent, and an annular outer capillary (250 µm i.d., 350 µm o.d.) for delivering the nebulizing gas.28 The sample position was manipulated using two 200 steps/revolution stepping motors. (12) Sjovall, P.; Lausmaa, J.; Johansson, B. Anal. Chem. 2004, 76, 4271–4278. (13) Malmberg, P.; Borner, K.; Chen, Y.; Friberg, P.; Hagenhoff, B.; Mansson, J. E.; Nygren, H. Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 2007, 1771, 185–195. (14) Mas, S.; Touboul, D.; Brunelle, A.; Arangocillo, A.; Laprevote, O.; Barderas, M. G.; Egido, J.; Vivanco, F. Mol. Cell. Proteomics 2006, 5, 1169. (15) Todd, P. J.; Schaaff, T. G.; Chaurand, P.; Caprioli, R. M. J. Mass Spectrom. 2001, 36, 355–369. (16) Puolitaival, S. M.; Burnum, K. E.; Cornett, D. S.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 2008, 19, 882–886. (17) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471–473. (18) Cooks, R. G.; Ouyang, Z.; Takats, Z.; Wiseman, J. M. Science 2006, 311, 1566–1570. (19) Takats, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 1261– 1275. (20) Costa, A. B.; Cooks, R. G. Chem. Commun. 2007, 3915–3917. (21) Wiseman, J. M.; Ifa, D. R.; Song, Q. Y.; Cooks, R. G. Angew. Chem., Int. Ed. 2006, 45, 7188–7192. (22) Wiseman, J. M.; Ifa, D. R.; Zhu, Y. X.; Kissinger, C. B.; Manicke, N. E.; Kissinger, P. T.; Cooks, R. G. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 18120–18125. (23) Dill, A. L.; Ifa, D. R.; Manicke, N. E.; Ouyang, Z.; Cooks, R. G. J. Chromatogr., B, 2009, 877, 2883-2889. (24) Dill, A. L.; Ifa, D.; Manicke, N.; Costa, A.; Ramos-Vara, J.; Knapp, D.; Cooks, R. Anal. Chem., in press. (25) Wu, C. P.; Ifa, D. R.; Manicke, N. E.; Cooks, R. G. Anal. Chem. 2009, 81, 7618-7624. (26) Li, Y. L.; Su, X.; Stahl, P. D.; Gross, M. L. Anal. Chem. 2007, 79, 1569– 1574. (27) Manicke, N. E.; Kistler, T.; Ifa, D. R.; Cooks, R. G.; Ouyang, Z. J. Am. Soc. Mass Spectrom. 2009, 20, 321–325. (28) Takats, Z.; Wiseman, J. M.; Cooks, R. G. J. Mass Spectrom. 2005, 40, 1261– 1275.

Normal DESI imaging experiments were performed in the positive ion mode with a spray solvent of 1:1 (v:v) methanol:water with 40 ppm formic acid. Reactive DESI experiments were performed in the positive ion mode using a spray solvent of 8:3:1 (v:v:v) acetonitrile:water:dimethylformamide with 66 ppm betaine aldehyde. This solvent was used because betaine aldehyde reacts with methanol. In both cases, the pixel size was 200 × 200 µm. All experiments were performed using a Thermo-Scientific LTQ (San Jose, CA) over a m/z range of 150-1200. Reagents and Materials. N-palmitoyl-D-erythro-sphingosylphosphorylcholine (16:0 SM, C39H80N2O6P, MW 703.6) (1), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (16:0-18:1 GPCho, C42H83NO8P, MW 760.1) (2), and 1-palmitoyl-2-hydroxysn-glycero-3-phosphocholine (14:0 Lyso GPCho, C24H51NO7P, MW 496.6) (3) were purchased from Avanti Polar Lipids (Alabaster, AL). HPLC grade methanol, acetonitrile, chlororform, sodium acetate and ammonium chloride were purchased from Mallinckrodt Baker Inc. (Phillipsburg, NJ), butylated hydroxytoluene and LDL isolated from human plasma from Sigma-Aldrich (St. Louis, MO), and water (18.2 MΩ-cm) was produced using a PureLab ultra system by Elga LabWater (High Wycombe, UK). The reference compounds were dissolved in MeOH/chloroform 3:1 (v/v) and used at concentrations of 1 mM for DESI experiments. Sample Preparation. Human atherosclerotic plaque tissue was removed by carotid endarterectomy, quickly washed in icecold PBS to remove blood contamination, and frozen in liquid nitrogen within 5 min after the procedure completion followed by storage at -80 °C. Prior to sectioning, the plaque was divided into four portions and embedded in 2% aqueous carboxymethyl cellulose. The blocks were frozen, and 10 µm thick sections were cut using a cryotome and thaw mounted on glass slides. The glass slides holding the tissue sections were stored at -80 °C and dried under vacuum for about 3 h prior to analysis. RESULTS AND DISCUSSION Analysis of Phospholipids. Figure 1A shows a representative negative ion DESI mass spectrum from direct analysis of a plaque tissue section. Signals due to intact phospholipids appear in the range m/z 700-850 while some hydrolyzed phospholipids were identified at m/z 500-600. Table 1 and Supporting Information (SI) Table S2 summarize the main compounds identified in the negative ion mode using DESI-MS. The nomenclature used to explain structural assignments include R for the alkyl residue of Lyso-GPCho and the sn-2 position of SM and R1 and R2 for the alkyl residues in the sn-1 and sn-2 positions, respectively, for diacyl-GPCho.29 Chemical assignments were performed by tandem mass spectrometry using DESI-MS directly from the tissue and by comparing the product ion mass spectra with those obtained from standard compounds. Tandem mass spectrometry data for one of the phospholipids in the negative ion mode are shown in SI Figure S1. The intact phospholipids were identified as chloride adducts of GPCho or SM. The peaks are chloride adducts due, presumably, to the presence of this electrolyte in the extracellular matrix. The presence of these two (29) Harrison, K. A.; Davies, S. S.; Marathe, G. K.; McIntyre, T.; Prescott, S.; Reddy, K. M.; Falck, J. R.; Murphy, R. C. J. Mass Spectrom. 2000, 35, 224– 236.

Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

8703

Figure 1. (A) Negative-ion DESI mass spectrum, acquired from a 10 µm slice of artery plaque tissue, using methanol/water 3:1 (v/v), (B) positive ion mass spectrum using 3:1 methanol/water, (C) positive ion mass spectrum acquired using using methanol/water 3:1 (v/v) with 0.01% (w/v) ammonium chloride. Table 1. Intact Phosphocholines and Sphingomyelins Detected in the Artery Plaque Tissue by DESI-MS name

m/z of [M+Cl]-

m/z of [M+Na]+

14:0 SM 15:0 SM 16:1 SM 16:0 SM 18:0 SM 32:0 GPCho 33:1 GPCho 34:2 GPCho 34:1 GPCho 36:4 GPCho 36:3 GPCho 36:2 GPCho 22:0 SM 24:2 SM 24:1 SM

709 723 735 737 765 768

697 711 723 725 753 756 768 780 782 804 806 808 809 833 835

792 794 816 818 820 821 845 847

polar lipid classes is expected because of the presence in the plaque of low density lipoprotein (LDL), which has a monolayer consisting of cholesterol and phospholipids, mainly GPCho and SM.30 Additionally, cells and cellular debris are present in atherosclerotic lesions,31 and GPCho and SM are major constitu(30) Kinnunen, P. K. J.; Holopainen, J. M. Trends Cardiovasc. Med. 2002, 12, 37–42. (31) Katz, S. S.; Small, D. M. J. Biol. Chem. 1980, 255, 9753–9759.

8704

Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

ents of biological membranes.32 Some peak overlap occurs, and this may hinder compound identification through tandem MS, especially for low intensity peaks. For example, the peak at m/z 794.5 assigned to [34:1 GPCho + Cl]- overlaps with the 37Cl isotope of [34:2 GPCho + Cl]- with nominal m/z of 792.5. Several hydrolyzed phosphatidylcholines were also identified in the DESI mass spectrum of the plaque tissue (SI Table S1). Among these, Lyso-GPCho 16:0 and Lyso-GPCho 18:0 dominate (see peaks m/z 530 and m/z 558 in Figure 1A). Structural identification of these lipids was performed by tandem MS (SI Figure S2), and a previous study analyzing GPCho standards in the positive and negative ion modes indicates that these lipids do not arise during the spray process or because of in-source fragmentation.33 Lyso-GPCho lipids are ubiquitous species generated following phospholipase A2 hydrolysis of phosphatidylcholine.34 Many other reactions also generate phosphatidylcholine species with a free hydroxyl group in the sn-2 position. LysoGPCho augments inflammation through effects on adhesion molecules and growth factors, monocytes, and macrophages.34-36 Lyso-GPChos are known to be significantly more abundant in the oxidized LDL found in plaques and are believed to play an important role in atherosclerosis because of their proinflammatory properties.37,38 Inflammation is common to all stages of atherosclerosis; immune cells are abundant during the early stages of the disease, and various inflammatory mediators are known to accelerate the process in animal models. A number of studies have attributed some of the proatherothrombotic properties of LDL to Lyso-GPChos, such as induction of adhesion molecules in endothelial cells,39 augmenting foam cell formation by causing an increase in LDL scavenger receptors on macrophages,40 stimulation of growth factor and cytokine release by macrophages,41 mediating apoptosis,42 and a number of others.38 Data obtained from the direct analysis of plaque tissue in the positive ion mode yielded complementary information (Table 1 and SI Table S1). Figure 1B shows a positive ion DESI mass spectrum obtained from the direct analysis of the plaque tissue using MeOH/water 3:1 (v/v). A representative product ion spectrum is shown in SI Figure S3. Due to the presence of endogenous sodium, the phospholipids ionize by adduct formation with sodium ions even though no salt is added to the DESI spray. The lipid profile is very similar to that obtained in the negative (32) Vance, D. E.; Vance, J. Biochemistry of Lipids, Lipoproteins, and Membranes; Elsevier: New York, 1996. (33) Manicke, N. E.; Wiseman, J. M.; Ifa, D. R.; Cooks, R. G. J. Am. Soc. Mass Spectrom. 2008, 19, 531–543. (34) Balsinde, J.; Winstead, M. V.; Dennis, E. A. Elsevier Science Bv: Madrid, Spain, May 20-22 2002; pp 2-6. (35) Yan, J. J.; Jung, J. S.; Lee, J. E.; Lee, J.; Huh, S. O.; Kim, H. S.; Jung, K. C.; Cho, J. Y.; Nam, J. S.; Suh, H. W.; Kim, Y. H.; Song, D. K. Nat. Med. 2004, 10, 161–167. (36) Kume, N.; Gimbrone, M. A. J. Clin. Invest. 1994, 93, 907–911. (37) Chen, L.; Liang, B. H.; Froese, D. E.; Liu, S. Y.; Wong, J. T.; Tran, K.; Hatch, G. M.; Mymin, D.; Kroeger, E. A.; Man, R. Y. K.; Choy, P. C. J. Lipid Res. 1997, 38, 546–553. (38) Matsumoto, T.; Kobayashi, T.; Kamata, K. Curr, Med, Chem, 2007, 14, 3209–3220. (39) Rong, J. X.; Berman, J. W.; Taubman, M. B.; Fisher, E. A. Arterioscler, Thromb. Vasc. Biol. 2002, 22, 1617–1623. (40) Kita, T.; Kume, N.; Yokode, M.; Ishii, K.; Arai, H.; Horiuchi, H.; Moriwaki, H.; Minami, M.; Kataoka, H.; Wakatsuki, Y. New York Academy of Sciences: Barcelona, Spain, May 18-20 1999; pp 95-102. (41) Liu-Wu, Y.; Hurt-Camejo, E.; Wiklund, O. Atherosclerosis 1998, 137, 351– 357. (42) Matsubara, M.; Hasegawa, K. Atherosclerosis 2005, 178, 57–66.

Figure 2. Selected ion images for the m/z 725, 782, and 518 from sample p600 1-4 (columns 1-4).

ion mode, although the relative intensity of various species is slightly different and the positive ion spectrum contains some additional low abundance peaks (Figure 1B). Analysis of Cholesterol and Cholesteryl Esters. The addition of reagent chemicals or salts to the DESI spray can lead to the detection of new chemical species, generated either through reaction or adduct formation (SI Table S2). Because detection of cholesteryl esters as ammonium adducts by ESI has been previously reported,43 ammonium chloride 0.01% (w/v) was added to the spray solvent to facilitate the detection of cholesteryl esters. In addition to sodiated and protonated phospholipid ions, several new peaks of low intensity were observed at m/z 640, 666, and 690 (Figure 1C). Their identification as ammonium adducts of cholesteryl esters (CE) was confirmed through CID (SI Figure S4). The ability to identify this type of compound is significant, as CE comprises a major fraction of the lipid-rich plaque and its abundance has been associated with plaque rupture.5 To improve the selectivity and sensitivity for imaging of cholesterol and other biologically complex alcohols, reactive DESI was implemented using precharged betaine aldehyde (BA). In this experiment, an in situ chemical reaction is performed at the site being sprayed to yield a more easily ionized form of the analyte(s) of interest. In the case of BA, the chemical selectively reacts with alcohols to generate a salt, the ready ionization of which improves analytical performance.25 BA tags the hemiacetal reaction product with a positive charge for higher sensitivity in positive mode detection (SI Scheme 1). For a target molecule of molecular weight M, the detected product hemiacetal gives a signal at m/z (43) Liebisch, G.; Binder, M.; Schifferer, R.; Langmann, T.; Schulz, B.; Schmitz, G. Biochim. Biophys. Acta, Mol. Cell Biol. Lipids 2006, 1761, 121–128.

(M+102), where 102 corresponds to the molecular weight of BA. This charge labeling method greatly improves the sensitivity; detection limits of ∼1 ppm (one ng absolute amount) were achieved for neat solutions of cholesterol,25 which is several orders of magnitude better than obtainable for conventional DESI. A representative mass spectrum when using the reagent betaine aldehyde is shown in SI Figure S6. An intense peak at m/z 488 is seen representing [cholesterol + BA]+. A much smaller peak at m/z 504 is also present. This is likely an oxidation product of cholesterol, [cholesterol + O + BA]+. However, this peak has also been observed during analysis of pure cholesterol, so the peak may be an artifact arising from oxidation reactions during ionization.44 Detection and identification of oxidized sterols is complicated because even the most abundant oxidized species in the atheroma are 100 times less abundant than cholesterol. The concentration of cholesterol in atheroma is high, with a concentration of over 200 µg/mg tissue. The most abundant oxidized sterols are 2 µg/mg or less.45 The importance of oxidized cholesterol in atherosclerosis is recognized,46 however the sensitivity or selectivity of the method needs improvement to detect these molecules. DESI Imaging. DESI-MS imaging was performed on sections taken from four different portions of a human atherosclerotic plaque tissue sample (designated P1 through P4). Imaging was performed in the positive ion mode with both normal DESI and reactive DESI using betaine aldehyde. (44) Benassi, M.; Wu, C. P.; Nefliu, M.; Ifa, D. R.; Volny, M.; Cooks, R. G. Int. J. Mass Spectrom. 2009, 280, 235–240. (45) Carpenter, K. L. H.; Taylor, S. E.; Ballantine, J. A.; Fussell, B.; Halliwell, B.; Mitchinson, M. J. Biochim. Biophys. Acta 1993, 1167, 121–130. (46) Schroepfer, G. J. Physiol. Rev. 2000, 80, 361–554.

Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

8705

Figure 3. Averaged normal DESI mass spectra from three different regions within sample P3 showing intact phospholipids and sphingolipids. Spectrum A is from a region of relatively low lipid signal. Spectra B and C are representative of the two different lipid distributions seen within the lipid rich regions. The image shown in the upper left is a total ion image, reflecting the ion intensity over the entire mass range for each pixel.

The ions detected by normal DESI imaging include the intact polar lipids and hydrolyzed lipids in Table 1 and SI Table S1, along with the osmolyte glycerophosphorylcholine at m/z 280. Selected ion images obtained from each sample using normal DESI together with a corresponding H&E stained section are shown in Figure 2, with additional selected ion images shown in SI Figures S7-S10. Each selected ion image within a given sample set is plotted on the same color scale to allow comparison of ion intensity between images. The DESI imaging data reveal that each sample set contains lipid rich regions in which the signal from the lipid species is significantly higher. While all of the lipid species detected seem to be more abundant in the lipid rich regions, the lipid composition is not constant within these regions. Figure 3 shows average mass spectra from three different regions of the tissue from sample P3. The areas of the tissue that do not show abundant lipid signal (Figure 3A) show a fairly constant lipid composition typified by abundant signal from 16:0 sphingomyelin at m/z 725. Some of the lipid rich regions have a very similar lipid composition, except that the overall lipid abundance is significantly higher (Figure 3B). Other lipid rich regions, on the other hand, have a significantly different lipid distribution, in which 34:2 GPCho, 34:1 GPCho, and 36:2 GPCho (among others) are significantly more abundant relative to 16:0 SM (Figure 3C). The different lipid regions can be visualized using the ratio of 725 to 782 and vice versa, as shown in SI Figure S11. The selected ion 8706

Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

image of 725 is overlaid with a 2D plot of 782:725 (SI Figure S11A) and 725:782 (SI Figure S11B). Using reactive DESI, an ion corresponding to cholesterol + betaine aldehyde (m/z 488) is clearly detected from the tissue in the positive ion mode. Selected ion images for the ion at m/z 488 are shown in SI Figure S12. As with the other species detected by normal DESI in the positive and negative ions modes, the distribution of cholesterol is detected more abundantly in certain lipid rich regions of the tissue. The fact that there are lipid rich regions is consistent with what is known about plaque formation. An early step in plaque formation is the accumulation of lipoprotein particles and aggregates in the tunica intima. Monocytes then adhere to the endothelium at the site of lipoprotein accumulation and penetrate into the intima where they differentiate into macrophages and take up the lipoproteins via endocytosis. As these cells die, they add their lipid filled contents to the growing necrotic core of the intima.47 The lipid rich regions detected by DESI could be the sites within the plaque in which more lipoprotein and macrophages have accumulated. In support of this hypothesis, cholesterol, GPChos, and SMs were detected from the plaque tissue by DESI with higher signal coming from certain lipid rich regions. Higher levels of all of these molecules are expected within the plaque based on previous studies showing that lesions at all stages are higher in (47) Ross, R. Nature 1993, 362, 801–809.

these lipid types compared to healthy artery.45,48,49 The higher levels of these lipids are primarily due to the accumulation of LDL, which contains a lipid core comprised mostly of cholesteryl esters surrounded by a membrane consisting of cholesterol, GPCho and SM.30 Analysis of LDL isolated from human serum showed a similar chemical profile (SI Figure S13). The detection of these two polar lipids, cholesteryl esters, and cholesterol are therefore consistent with the hypothesis that the lipid rich regions detected by DESI correspond to areas of more abundant LDL accumulation. As depicted in Figure 3, two grossly different polar lipid distributions were observed in the tissue. The profiles seen in Figures 3A and C, which were detected from the areas of the tissue with low lipid signal and from some of the lipid rich regions, show relative SM/GPCho amounts of 64% SM and 36% GPCho, while the profile in Figure 3B has relative SM and GPCho amounts of 43% and 57%, respectively. The percentages are based on the relative intensities detected for the species in Table 1. The ratio of SM to GPCho is an important variable in plaque formation, as the SM/GPCho ratio varies in different lipoprotein subclasses and the ratio is higher in plaque tissue than in healthy artery.49,50 The increased sphingomyelin in atherosclerotic lesions plays a part in the aggregation of LDL particles in lesions.51 Sphingomyelinase enzymes released by macrophages promote aggregation by converting SM to the more nonpolar ceramide.49 As a cautionary note, it cannot be concluded that the relative amounts of SM and GPCho detected by DESI reflect the relative amounts in the tissue. However, a previous imaging study in rat brain suggests that the spatial intensity distributions are in qualitative agreement with relative concentration differences.21 DESI-MS may therefore be (48) McCandless, E. L.; Zilversmit, D. B. Arch. Biochem. Biophys. 1956, 62, 402–410. (49) Schissel, S. L.; TweedieHardman, J.; Rapp, J. H.; Graham, G.; Williams, K. J.; Tabas, I. J. Clin. Invest. 1996, 98, 1455–1464. (50) Subbaiah, P. V.; Davidson, M. H.; Ritter, M. C.; Buchanan, W.; Bagdade, J. D. Atherosclerosis 1989, 79, 157–166. (51) Guyton, J. R.; Klemp, K. F. Arterioscler. Thromb. Vasc. Biol. 1996, 16, 4– 11.

a useful as a tool for determining the relative amounts and composition of GPCho and SM in artery tissue, which could in turn give insights into the subclasses of lipoprotein at various sites within the tissue. CONCLUSION DESI-MS was used for chemical profiling and imaging of arterial plaques. In the positive and negative ion modes, sodium and chloride adducts, respectively, of diacyl glycerophosphocholines (GPChos), sphingomyelins (SMs), and hydrolyzed GPChos were detected. Further, addition of ammonium to the DESI spray was used to detect cholesteryl esters and reactive DESI using betaine aldehyde was used to image cholesterol. DESI imaging revealed that there are lipid rich regions within the arterial walls, which likely correspond to the areas of the tissue where lipoprotein particles have accumulated. Additionally, the lipid rich regions seem to have one of two different lipid profiles. The lipid composition can vary among plaques and also changes during plaque progression. Characterization of the lipid composition of plaques will therefore likely lead to a better understanding of plaque vulnerability and atherosclerosis. ACKNOWLEDGMENT This work was supported by the National Institutes of Health grant # 1R21EB009459-01 and in part by a fellowship from Merck Research Laboratories (NM) and ACS-DAC Fellowship (MN). We would like to thank Sam Wright for his support for this project, and Manny and the ev3 team for providing support during methods development and human plaque collection efforts. SUPPORTING INFORMATION AVAILABLE Scheme 1, Tables S1 and S2, Figures S1-S13. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review August 3, 2009. Accepted September 21, 2009. AC901739S

Analytical Chemistry, Vol. 81, No. 21, November 1, 2009

8707