Lipid cartography of atherosclerotic plaque by cluster-TOF-SIMS imaging

Abstract

Several frozen vessels bearing atherosclerotic lesion were analysed by cluster TOF-SIMS (time-of-flight secondary ion mass spectrometry ) to map their lipid (fatty acids, cholesterol, vitamin E, phosphatidic acids, phosphatidylinositols and triglycerides) content at a micrometric resolution.

---

Atherosclerosis is traditionally considered as a lipid-mediated pathology since the XIX-century when Virchow's lipid hypothesis was postulated.¹ The lipid role is not only limited to the genesis of this pathology but the lipid contents of the plaque also seem to be key players on the progression of the lesion.²

Lipid accumulation within the vessels measured by histochemical methods can be correlated with histological classification, particularly on unstable lesions characterized by the presence of a large lipid-rich necrotic core.³ It has been shown that aortic plaques, with a lipid core covering more than 40% of the total plaque surface, are at the highest risk of rupture and this has been reinforced by the fact that lipid reducing treatments, such as statins, or a combination of dietetic intervention plus exercise are able to reduce the risk of acute coronary events.⁴

Qualitative analyses of the lipid content on the plaque have been based on lipid solvent extraction from artery homogenates followed by mass spectrometry experiments.⁵ These techniques have proved to be efficient in the characterization of the most abundant lipids on the atheroma core. Despite the sensitivity of this approach, the spatial information on individual molecular species is lost during the sample extraction. New methods for assessing the vascular morphology as intravascular ultrasound imaging (IVUS),⁶ or more recently intravascular magnetic resonance (IVMI),⁷ provide in vivo information on the size and physical characteristics of the lipid core, but there is still the need for analytical tools that could provide qualitative and spatial information on the molecular composition of the plaque.

TOF-SIMS using cluster liquid metal ion guns (gold or bismuth clusters) can be an interesting approach for this purpose, without the need of specific sample preparation and a very good yield for hydrophobic molecules.⁸ High-resolution images of specific metabolite distributions on tissues can thus be obtained with a routine spatial resolution of 1–2 µm, together with a mass resolution of several thousands (FWHM). Moreover, the data acquisition time is relatively short (less than 1 h) and the technique is easy to integrate among other imaging techniques. Then cluster-TOF-SIMS imaging seems to be very well adapted to study a lot of pathologies, and particularly those in which lipids play a fundamental or major role.

Samples from four male patients (carotid stenosis >70%) undergoing carotid endarterectomy at the Fundación Jiménez Díaz were used in the study. Informed consent and the clinical characteristics were obtained before enrolment. For mass spectrometry analyses, 15 µm-thickness samples were sectioned in a cryostat, at a constant temperature of –25 °C. The tissue sections were deposited onto a stainless steel plate and stored again at –80 °C. After drying under a pressure of a few hPa during 15 min, they were directly analyzed in the TOF-SIMS IV mass spectrometer fitted with a bismuth cluster ion source located at the ICSN. For histochemical staining, we used consecutive sections to those used in TOF-SIMS imaging. 4 µm thickness sections of the frozen samples were thaw-mounted onto aminopropyltriethoxysilane (APES)-coated glass slides. Masson trichrome stains were performed to evaluate vascular lesion.

Surface rastering from atherosclerotic lesion thin-sections generates numerous secondary ion signals in the m/z range 1 to 1000. All the recorded peaks correspond to singly charged ions in both positive and negative ion modes. A representative negative ion spectrum obtained from atherosclerotic plaque surfaces is shown in Fig. 1. Below m/z 200 most peaks are attributed to inorganic ions and organic fragment ions arising from collisions between primary ions and surface molecules or from in-source dissociation. On the mid-range, between m/z 200 and 500, several peaks have been assigned to non-esterified fatty acids: palmitic (C16:0, m/z 255), stearic (C18:0, m/z 283), palmitoleic (C16:1, m/z 253), oleic (C18:1, m/z 281), hexadecadienoic (C16:2, m/z 251), linoleic (C18:2, m/z 279), gamma-linolenic (C18:3, m/z 277) or arachidonic (C20:4, m/z 303) acids. In addition, well-defined peaks attributed to cholesterol and vitamin E are found at m/z 385 and m/z 429, respectively. Finally, phosphatidylinositol fragments (m/z 223 and 241), sphingomyeline fragments (m/z 616 and 642), phosphatidic acids (m/z 650–750) and triglycerids (m/z 829–882) can be also observed on the average mass spectrum recorded from an atherosclerotic-bearing vessel (Fig. 1). The mass assignments are based on accurate mass measurements and biological significance. Additional confirmation was also obtained using post-source decay-like (PSD-like) experiments already described in ref. 9.

Fig. 1 TOF-SIMS spectrum recorded from an atherosclerotic-bearing vessel (Fig. 3) in the negative ion mode.

Lipid images are presented in Fig. 2, together with a view of an adjacent slice stained with Masson's trichrome. It corresponds to a whole transversal section of the endartery specimen. Fatty acids are more abundant on the boundary between neointimal and medial layer on a heterogeneous and discontinuous fashion, generating fatty-acid-rich islets. The most intense signals are found for palmitic and stearic acids, which are known to be the most abundant fatty acids on human tissues. Oleic acid and other unsaturated fatty acid are located on the same positions. Cholesterol ion distribution differs from those observed from fatty acids. Other molecules of interest were displayed, such as vitamin E, and are colocalized with cholesterol. Phosphatidic acid and sphingomyeline fragments show a wide distribution on the tissue surface, which is partially overlapping with cholesterol-rich areas.

Fig. 2 TOF-SIMS images in a non-calcified endartery sample. (A) Masson's stained slice consecutive to the one used for TOF-SIMS analysis, showing major morphological features: L (arterial lumen), Neo (atherosclerotic intima), M (medial fragments), IL (internal elastic lamina), FC (fibrous cap); (B) Eight of the most intense peaks found in the same sample are represented as bidimensional ion density maps. The field of view is 8.4 × 8.4 mm² (256 × 256 pixels). To increase the contrast, the images are compressed to 128 × 128 pixels (final spatial resolution 65.6 µm). The primary ion dose density is about 5 × 10⁹ ions cm^–2. The name of the compounds and the m/z value of the peak centroid, the maximal number of counts in a pixel (mc) and the total number of counts (tc) are written below each image. The color scales correspond to the interval [0, mc].

High resolution images of the vulnerable sites, which are areas of fissuration and rupture, help us to define functional regions on apparently homogeneous tissue. In Fig. 3, the previously described negative ions observed in the thrombosis boundary are shown together with the simultaneously recorded video image. In this area, fatty acids with sixteen carbons are widely distributed. C18:0 fatty acid is abundant in the luminal site and the thrombus, C18:1 and C18:2 fatty acids are abundant in the luminal site but less in the thrombi compared to C18:0, while C18:3 is only present on the outer areas of the neointimal tissue. Cholesterol is found in the inner border, suggesting contact with the luminal flow, while vitamin E is located in the surrounding area, including the thrombus itself. Phosphatidic acids and sphingomyeline fragments are colocalized at the inner border, whereas phosphoinositol fragments are located in the outer parts of the neointimal tissue.

Fig. 3 (A) Optical image of a plaque vulnerable site. The 500 × 500 µm² area ion image is recorded inside the green square. (B) Ion density maps of seventeen ions. The field of view is 500 × 500 µm² (256 × 256 pixels). To increase the contrast, the images are compressed to 128 × 128 pixels (final spatial resolution 3.9 µm). The primary ion dose density is about 1 × 10¹² ions cm^–2. The name of the compounds and the m/z value of the peak centroid, the maximal number of counts in a pixel (mc) and the total number of counts (tc) are written below each image. The color scales correspond to the interval [0, mc]. PI; phosphatidylinositol; SM: sphingomyeline.

An interesting finding from two of the analysed samples is the presence of intense signals of several non-esterified fatty acids (NEFA). Several experiments were carried out to test if the presence of these compounds is an analytical artefact, due to in-source fragmentation of complex lipids, such as triglycerids. It was shown that lipid fragmentation was only a minor contributor to the NEFA's presence.⁹ Further, it can be seen in Fig. 3 that NEFA signals are only partially colocalized with those of triglycerides. Rationalization of the presence of free fatty acids in these areas require further investigation with a larger number of patients.

The present work was initially intended as a proof-of-concept study, using just four atherosclerotic plaques, in order to test the usefulness of TOF-SIMS for metabolite imaging of human vascular samples. In our opinion, cluster TOF-SIMS has proven to be a valuable technique for atherosclerosis pathological studies, being capable in only one single experiment to display spatial information of multiple low molecular weight molecules directly from biological tissues, as no other technique is able to accomplish it at the present time.

With cluster TOF-SIMS imaging experiments, the secondary ions generated from the sample show a bias towards hydrophobic molecules. Hence, several lipids might be overrepresented in the spectra while other very abundant metabolites, such as sugars, are completely absent. Nevertheless, this drawback of the technique becomes useful on cardiovascular research, as lipids play an important role on the genesis and development of this pathology. Therefore, lipid imaging on human samples can be accomplished with high sensitivity together with the high resolution inherent to this ionization source, achieving precise molecular maps of the lipid distribution across the tissue. Furthermore, as the primary ion current has proven to be very stable along the data acquisition time and as the secondary ion emission yields can be assumed to be constant over the whole sample surface, the intensities in the density maps can therefore be correlated to relative concentrations, enabling relative quantification over the tissue surface.

This paper presents for the first time, as far as we know, the use of this technique on human pathology, adding new data on lipid metabolism on plaque. Thus, from the data we obtain in our samples we highlight the presence of high amounts of endogenous non-esterified fatty acids in some samples, as well as the presence of vitamin E (a putative redox state marker) at the innermost areas of the intima bordering extra-cellular cholesterol accumulation. In order to assess the physiological relevance, new experiments are being carried out to confirm their role on the plaque stability.

TOF-SIMS Imaging is still far from having widespread use in clinical pathology, but its use in experimental pathology has proven valuable, particularly in lipid mediated diseases.

Acknowledgements

This work was supported by grants from the Fundación Mutua Madrileña and Pfizer-Spain. S.M.’s research grant was funded by the Ministerio de Educación y Ciencia (SAF 2004/06109). D.T. is indebted to the Institut de Chimie des Substances Naturelles (CNRS) for a PhD research fellowship.

Notes and references

1. V. Fuster, L. Badimon, J. J. Badimon and J. H. Chesebro, N. Engl. J. Med., 1992, 326, 242–50; V. Fuster, L. Badimon, J. J. Badimon and J. H. Chesebro, N. Engl. J. Med., 1992, 326, 310–318.
2. V. Fuster, J. Badimon, J. H. Chesebro and J. T. Fallon, Haemostasis, 1996, 26(suppl 4), 269–284; M. Navab, A. M. Fogelman, J. S. Frank, L. L. Demer, P. A. Edwards, A. D. Watson and A. J. Lusis, Circulation, 1995, 91, 2488–2496.
3. H. C. Stary, A. B. Chandler, R. E. Dinsmore, V. Fuster, S. Glagov, W. Insull, Jr., M. E. Rosenfel, C. J. Schwartz, W. D. Wagner and R. E. Wissler, Arterioscler. Thromb. Vasc. Biol., 1995, 15(9), 1512–31; P. K. Shah, J. Am. Coll. Cardiol., 2003, 41(4 Suppl S), 15S–22S.
4. M. J. Davies, P. D. Richardson, N. Woolf, D. R. Katz and J. Mann, Br. Heart J., 1993, 69, 377–381; R. Corti, Z. A. Fayad, V. Fuster, S. G. Worthley, G. Helft, J. Chesebro, M. Mercuri and J. J. Badimon, Circulation, 2001, 104(3), 249–52; R. P. Wildman, L. L. Schott, S. Brockwell, L. H. Kuller and K. Sutton-Tyrrell, J. Am. Coll. Cardiol., 2004, 44(3), 579–85.
5. Z. Mallat, T. Nakamura, J. Ohan, G. Leseche, A. Tedgui, J. Maclouf and R. C. Murphy, J. Clin. Invest., 1999, 103(3), 421–7; H. G. Rose and M. Oklander, J. Lipid Res., 1965, 6, 428–431.
6. D. M. Small, Arteriosclerosis, 1988, 8, 103–129.
7. A. Blank, G. Alexandrowicz, L. Muchnik, G. Tidhar, J. Schneiderman, R. Virmani and E. Golan, Magn. Reson. Med., 2005, 54(1), 105–12.
8. A. Brunelle, D. Touboul and O. Laprévote, J. Mass Spectrom., 2005, 40(8), 985–99; D. Touboul, F. Kollmer, E. Niehuis, A. Brunelle and O. Laprévote, J. Am. Soc. Mass Spectrom., 2005, 16(10), 1608–18; D. Touboul, A. Brunelle, F. Halgand, S. De La Porte and O. Laprévote, J. Lipid Res., 2005, 46, 1388–1395.
9. D. Touboul, A. Brunelle and O. Laprévote, Rapid Commun. Mass Spectrom., 2006, 20(4), 703–9.

---

Footnote

† These two authors equally contributed to this work.

---