Spatially Resolved Metabolites in Stable and Unstable Human Atherosclerotic Plaques Identified by Mass Spectrometry Imaging

Abstract

Background –

Impairments in carbohydrate, lipid, and amino acid metabolism drive features of plaque instability. However, where these impairments occur within the atheroma remains largely unknown. Therefore, we sought to characterize the spatial distribution of metabolites within stable and unstable atherosclerosis in both the fibrous cap and necrotic core.

Methods –

Atherosclerotic tissue specimens from nine unmatched individuals were scored based on the Stary classification scale and subdivided into stable and unstable atheromas. After performing mass spectrometry imaging (MSI) on these samples, we identified over 850 metabolite-related peaks. Using MetaboScape, METASPACE, and HMDB, we confidently annotated 170 of these metabolites and found over 60 of these were different between stable and unstable atheromas. We then integrated these results with an RNA-sequencing dataset comparing stable and unstable human atherosclerosis.

Results –

Upon integrating our MSI results with the RNAseq dataset, we discovered that pathways related to lipid metabolism and long-chain fatty acids were enriched in stable plaques, whereas reactive oxygen species, aromatic amino acid, and tryptophan metabolism were increased in unstable plaques. In addition, acylcarnitines and acylglycines were increased in stable plaques whereas tryptophan metabolites were enriched in unstable plaques. Evaluating spatial differences in stable plaques revealed lactic acid in the necrotic core, whereas pyruvic acid was elevated in the fibrous cap. In unstable plaques, 5-hydroxyindoleacetic acid was enriched in the fibrous cap.

Conclusions –

Our work herein represents the first step to defining an atlas of metabolic pathways involved in plaque destabilization in human atherosclerosis. We anticipate this will be a valuable resource and open new avenues of research in cardiovascular disease.

Graphical Abstract

INTRODUCTION

Despite the considerable advancements in diagnostic techniques and therapeutic strategies, atherosclerotic cardiovascular disease (ASCVD) remains the leading cause of morbidity and mortality worldwide¹. Acute thrombogenic events that lead to myocardial infarction and stroke are frequently caused by the rupture of unstable plaques, characterized by the presence of a large necrotic core covered by a thin fibrous cap^2,3. In contrast, stable plaques show thicker, collagen-rich fibrous caps with smaller necrotic cores. Understanding these features of plaque stability has provided a framework for much of the work regarding the complexity of mechanisms that eventually lead to ASCVD-related events.

There is a growing interest in elucidating the mechanisms that govern plaque instability, which has so far highlighted dysregulations in lipid, carbohydrate, and amino acid metabolism^4–9. Impaired metabolism plays a critical role in all stages of atherosclerosis, from its inception as a benign fatty streak to the precipitation of acute clinical events. However, where these impairments occur in the atherosclerotic plaque has yet to be revealed. This is critically important because lipid deposition, cell composition, and extracellular matrix remodeling are heterogeneous throughout human atheromas, particularly in regions of the fibrous cap and near the necrotic core. Furthermore, lesional cells are constantly bombarded with various intrinsic and extrinsic metabolic insults that influence cell phenotype^4,5. While our understandings of metabolism in atherosclerosis are being elucidated, technical limitations have prevented investigators from determining where these dysregulations in metabolism occur within the plaque.

Mass spectrometry imaging (MSI) is a powerful technology for visualizing the spatial localization and relative abundance of hundreds to thousands of analytes in thin tissue sections without a priori knowledge of analytes present in the samples¹⁰. Imaging is performed using matrix-assisted laser desorption/ionization (MALDI) to interrogate small areas, or pixels, from a tissue surface after adding a chemical matrix to the section, which serves to extract, co-crystallize, and ionize molecules from the sample¹¹. With the capability to image hundreds-to-thousands of molecules without labeling, the information gained from mass spectrometry and the visualization of its spatial distributions makes this platform uniquely suited to characterize the spatial distribution of metabolites within an atherosclerotic plaque. In the present study, we performed MALDI-MSI of late-stage, stable and unstable human atherosclerotic plaques to visualize the distribution of metabolites in the fibrous cap and necrotic core.

MATERIALS AND METHODS

The data not provided in this manuscript that support the findings of this study are available from the corresponding authors upon reasonable request. An expanded materials and methods section can be found in the online supplemental file.

Human atheroma specimen collection and scoring

Human tissue was deemed non-human research by the local IRB due to the exclusive use of post-mortem samples. Human tissue was excised postmortem during autopsy and fixed in 4% formaldehyde, embedded in paraffin, then cut into 5 μm sections. Based on stains described in Figure 1, sections were scored into type I (lesions with macrophage foam cells), type II (fatty streaks with intracellular lipid accumulation), type III (increased presence of extracellular lipid pools), type IV (presence of necrotic core and fibrous cap), or type V (similar to type IV except with the presence of spotty calcific nodules or calcific sheets)¹².

Figure 1. Mass Spectrometry Imaging of Human Atheromas.

(A) We obtained 141 human atheromas and identified nine type V human atherosclerotic plaques suitable for MSI. (B) Plaque staging and scoring for plaque vulnerability were aided by H&E, Picrosirius red, Alizarin red, and von Kossa staining. Representative images are shown for stable and unstable atheromas. Regions of the fibrous cap are identified in the cyan dashed lines and the necrotic core regions are identified in the blue dashed lines. (C) MSI was performed after staging and scoring for plaque instability. Matrix was applied to tissue sections using a Robotic Reagent Sprayer. 10 mg/mL α-cyano-4-hydroxycinnamic acid (CHCA) in 50% acetonitrile, 0.1% trifluoroacetic acid was used for positive mode, whereas 10 mg/mL 1,5-diaminonaphthalene in 50% acetonitrile was applied for negative ion mode. Images were acquired using Bruker FlexImaging 5.1 on a Bruker timsTOF fleX mass spectrometer in both positive and negative ion mode at a spatial resolution of 30 μm over the m/z range 50–1000. Example images for homo-L-arginine and 2-nonenal are shown. (D) A principal component analysis was performed on the intensities of metabolite-related peaks. The first two dimensions, PC1 and PC2, show distinct clusters of stable (St) and unstable (Unst) plaque specimens. (E) Data were displayed as a Volcano plot and significance was determined as -Log10 P values above 1.95 (determined by Benjamini-Hochberg test) with a Log2 fold change ± 1.5. Peaks that are significantly elevated in the whole vessel of unstable plaques are displayed as blue circles, whereas peaks significantly elevated in the whole vessel of stable plaques are displayed as green circles. (F) Identification of metabolites was carried out by searching for deposited metabolites in MetaboScape. Mass-to-charge values were also compared to metabolite image identifications in METASPACE followed by formula searching using the Human Metabolome Database then queried for the designation of being detected and quantified. Data from annotated metabolites are displayed as a Volcano plot and significance was determined as -Log10 P values above 1.95 (determined by Benjamini-Hochberg test) with a log2 fold change ± 1.5. Metabolites that are significantly elevated in the whole vessel of unstable plaques are displayed as blue circles, whereas metabolites significantly elevated in the whole vessel of stable plaques are displayed as green circles. (G) Joint pathway analysis was performed with MetaboAnalyst 5.0 on top differentially expressed genes (P < 0.05) and significantly altered metabolites (P < 0.05). To integrate the transcriptomic and metabolic data, genes and metabolites of interest were pooled together in the enrichment and topological analyses to search for significant metabolic pathways in the KEGG database. Metabolic pathways are ranked by an impact factor indicating their topological impact (degree centrality) on the pathway.

Mass Spectrometry Imaging

Detailed methods can be found in the online supplemental file. Briefly, cross-sections were deparaffinized and applied with matrix used an HTX M5 Robotic Reagent Sprayer. Images were acquired using Bruker FlexImaging 5.1 on a Bruker timsTOF fleX mass spectrometer. After data acquisition the matrix was solubilized and the sections were stained with hematoxylin and eosin. Images were captured and merged with for co-registration, which was carried out using the GNU image program. Metabolite identification was performed using MetaboScape 2021b, the Human Metabolome Database, and METASPACE.

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 9.0. Student’s t-test was conducted on the identified peaks then adjusted using a Benjamini-Hochberg correction test. Both adjusted and nominal P values are shown in the volcano plots as horizontal dashed and dotted lines, respectively.

RESULTS

To select stable and unstable human atherosclerosis samples for mass spectrometry imaging (MSI) analysis, we utilized a biobank of atherosclerotic tissue specimens taken from the right coronary artery, left coronary artery, left anterior descending artery, and the carotid sinuses of 141 individuals. These tissues were formalin-fixed, paraffin-embedded, sectioned and scored based on the Stary Classification scale¹², then subdivided into stable and unstable atheromas. These samples were then processed for MSI and integrated with a publicly available RNAseq dataset comparing stable and unstable human atherosclerosis¹³. A general overview of the experimental design is described in Figure 1A. To score the atherosclerotic plaques, H&E, picrosirius red (identifies collagen), and Alizarin red and von Kossa (identifies extracellular calcium deposition) staining were performed (Figure 1B). Based on the features from these stains, we identified 41 late-stage (type V) human atheromas. Because MSI demands high-quality tissue sections, we next removed samples that contained histological artifacts, such as wrinkles, tears, folds, or incomplete sectioning (n = 32). Based on fibrous cap thickness, we obtained five stable (fibrous cap > 200 microns¹⁴) and four unstable (fibrous cap < 200 microns¹⁴) atheromas (Supplemental Figure 1 and Supplemental Table 1).

Cross-sections of these stable and unstable plaques were then deparaffinized, and matrix was applied using an automated reagent sprayer. Samples were then analyzed in positive and negative ion modes, and images were acquired using Bruker FlexImaging 5.1 on a Bruker timsTOF fleX mass spectrometer with a total of 1200 laser shots per pixel. Images of all samples were combined and viewed in SCiLS Lab Pro 2021b after root mean squared normalization. MSI examples of our collected human atherosclerosis specimens are shown in Figure 1C and Supplemental Figure 1. The MALDI-MSI platform was able to identify 855 peaks with assigned chemical formulas. Principal component analysis (PCA) was carried out by centering the peak intensities by mean subtraction and scaled by dividing them by standard deviations. Resultant data were used based on R functions and visualized with ggplot2. Importantly, PC analysis supported our assignments of the atheromas into stable and unstable atheromas, showing clear separation between the two groups (Figure 1D). After Benjamini-Hochberg adjustments, we found that 51 peaks were significantly elevated in the whole vessel of the stable plaques, and 128 peaks were significantly elevated in the entire vessel of unstable plaques (Figure 1E). Metabolite identification was performed by loading the ROIs into MetaboScape 2021b (Bruker) and searching against software-deposited metabolite databases. Mass-to-charge values were also compared to metabolite image identifications in METASPACE (https://metaspace2020.eu/) followed by formula searching using the Human Metabolome Database (HMDB, https://hmdb.ca/) then queried for the designation of being previously detected and quantified. We then selected twelve metabolites within different classes of molecules (e.g., nucleotides and amino acids) and compared these to commercially purchased standards. Each standard was spotted to a MALDI target, and an MS/MS spectrum was collected using collision-induced dissociation (CID). The same parent m/z values observed in the standards were targeted in tissue sections and subjected to MS/MS. The fragments observed in the MS/MS spectra from tissues were compared to those obtained from the standards (Supplemental Figures 2–5, green arrows in each set of spectra indicate fragment peaks common between the standards and tissue). Unsurprisingly, we found that more than one compound was fragmented simultaneously from tissue sections due to multiple peaks that were present within a ±1 Da isolation window, resulting in chimeric spectra with additional fragment peaks that do not entirely match the MS/MS spectra obtained from the standards. This filtering system resulted in 170 annotated metabolites (Figure 1F).

We then coupled our annotated metabolites with RNAseq of human stable and unstable atheromas¹³, which revealed metabolic pathways primarily related to amino acid metabolism, including arginine, histidine, alanine, proline, glycine, and tryptophan. We also uncovered alterations in glycerophospholipid metabolism, unsaturated fatty acid biosynthesis, pyrimidine metabolism, and pyruvate metabolism (Figure 1G). When evaluating RNAseq analysis of stable and unstable plaques alone, differences in pathways related to the metabolism of collagen, arginine, nitric oxide, N-acetylglucosamine, reactive oxygen species, aromatic amino acids, long-chain fatty acids, and lipids were uncovered (Figure 2A). Twenty-nine of the individually annotated metabolites that were significantly different between stable and unstable atheromas could be assigned to at least one atherosclerosis-relevant pathway (Figure 2A). The pathway with the highest number of assigned metabolites was lipid metabolism. Whereas acylcarnitines (tiglylcarnitine, butenylcarnitine, butyrylcarnitine, and 3-hydroxybutyrylcarnitine) and acylglycines (hexanoylglycine and capryloylglycine), both known to be involved in fatty acid metabolism/oxidation, were enriched in stable plaques, saturated fatty acids (palmitic acid and stearic acid) and lipid peroxidation metabolites (4-hydroxynonenal and 2-nonenal) were increased in unstable plaques. Because xylene can remove neutral lipids, the process of deparaffinizing FFPE tissue sections results in the removal of cholesterol. However, many lipid species, including saturated fatty acids, are solvent-resistant¹⁵. To determine whether xylene could remove these lipid species, we isolated fresh-frozen brain tissue slices, known to be rich in fatty acids, and performed MSI before and after xylene washing. We found that stearic acid, palmitic acid, and DHA were largely unaffected by xylene and retained their spatial properties (Supplemental Figure 6). In addition to fatty acids, we discovered that aromatic amino acids and tryptophan metabolites (phenylacetic acid, phenylacetaldehyde, N-acetyltryptophan, N-acetylserotonin, indoleacetaldehyde, and 5-hydroxyindoleacetic acid) were predominantly found in unstable plaques. Interestingly, most metabolites could not be assigned to atherosclerosis-relevant pathways identified through RNAseq. This suggests that there are other metabolic pathways relevant to atherosclerosis that have yet to be explored and highlights the potential for new avenues of research linking dysregulations in metabolism and plaque instability.

Figure 2. Integration with RNAseq of Stable and Unstable Atheromas and Spatial Resolution of Metabolites in the Fibrous Cap and Necrotic Core.

(A) Differential gene expression analysis was performed on an NCBI curated RNAseq dataset (GSE120521) comparing unstable and stable human atherosclerotic plaques. The impact of differentially expressed genes (P < 0.05) on biological processes was determined by gene set enrichment analysis using the Gene Ontology database. Metabolic pathways that are significantly enriched (P < 0.05) are shown below the annotated metabolites. (B) Images of lactic acid and 5-HIAA are shown as examples of metabolites enriched in the necrotic core and fibrous cap, respectively. (C-F) Metabolites that are significantly elevated in the fibrous cap (C) and necrotic core (D) of unstable plaques are displayed as blue circles, whereas metabolites significantly elevated in stable plaques are displayed as green circles. Data are displayed as a Volcano plot and significance was determined as -Log10 P values above 2.25 and 2.15, respectively, as determined by the Benjamini-Hochberg test (horizontal dashed line; nominal P value is a dotted horizontal line at 1.3), with a Log2 fold change ± 1.5 (vertical dashed lines). In stable (E) and unstable plaques (F), data are displayed as a Volcano plot and significance was determined as -Log10 P values above 3.33 and 3.36, respectively, as determined by the Benjamini-Hochberg test (horizontal dashed line; nominal P value is a dotted horizontal line at 1.3), with a Log2 fold change ± 1.5 (vertical dashed lines).

We next assessed differences in the fibrous cap and the necrotic core, comparing stable and unstable atheromas. An example of two metabolites and their colocalization is shown in Figure 2B. In the fibrous cap, we identified thirteen metabolites significantly elevated in unstable plaques, and four metabolites significantly elevated in stable plaques (Figure 2C). In the necrotic core microenvironment, we discovered eleven metabolites significantly elevated in the unstable atheromas and five significantly elevated in the stable atheromas (Figure 2D). Interestingly, homo-L-arginine, a substrate for nitric oxide synthase and an emerging protective cardiovascular risk factor^16–18, was enriched in the fibrous cap and necrotic core of stable atheromas (Figure 2C and D). Whereas the lipid peroxidation metabolite 2-nonenal was elevated in both the fibrous cap and necrotic core of unstable atheromas (Figure 2C and D), pantothenic acid (vitamin B5), which contributes to fatty acid synthesis and reduces oxidative stress^19,20, was enriched only in the fibrous cap (Figure 2C). We next assessed metabolite differences in the fibrous cap versus the necrotic core. While there were no metabolites that did not pass the statistical threshold set by the FDR calculation, nominal P values indicated a trend of increased pyruvic acid in the fibrous cap and elevated lactic acid in the necrotic core of stable plaques (Figure 2E). In unstable plaques, the tryptophan metabolite 5-hydroxyindoleacetic acid, linked to metabolic syndrome and systemic inflammation²¹, met the nominal P value threshold, suggesting a potential increase in the fibrous cap (Figure 2F). Altogether, using MALDI-MSI coupled with transcriptomics, we identified metabolic differences in the whole vessel, fibrous cap, and necrotic core between stable and unstable atheromas.

DISCUSSION

Elegant omics studies, including RNAseq, scRNAseq, and bulk metabolomics, have revealed differences in gene expression, cell composition, and metabolite profiles in human atherosclerosis^22,23. However, these studies do not inform us wherein the plaques these differences occur. Here, we aimed to spatially resolve metabolites within the necrotic cores and fibrous caps of stable and unstable human atheromas in an unbiased manner. Integration of MSI with multi-omics approaches have yielded powerful results that more fully reveal tissue function during pathophysiologies^24,25. Here, we annotated 170 metabolites out of 855 m/z peaks and integrated these results with a publicly available RNAseq dataset that determined a majority of annotated metabolites belonged to metabolic pathways that are not currently associated with atherosclerosis. We also found a variety of classes of metabolites in the fibrous cap and necrotic cores of stable and unstable atheromas.

Our dataset revealed that acylcarnitines and acylglycines were enriched in stable plaques. While our data might suggest that enhancing acylcarnitine levels would be atheroprotective, L-carnitine supplements, which enhance acylcarnitine levels²⁶, worsen stenosis in individuals with metabolic syndrome²⁷. Furthermore, links between elevated acylcarnitines and insulin resistance have been described^28,29. Elevations in acylcarnitines are also associated with an increased risk of acute cardiovascular events^30–32. In contrast to acylcarnitines, increased circulating glycine levels are linked to protections in clinical and experimental atherosclerosis^9,33,34, and enhanced glycine levels increase circulating levels of acylglycines³⁵. These fatty acid-conjugated amino acids, particularly acylcarnitines, are critical for transporting fatty acids across the inner mitochondrial membrane³⁶, and the presence of these fatty acid-conjugated amino acids in stable plaques is in line with enhanced fatty acid oxidation, an atheroprotective metabolic pathway^4,5.

In contrast to acylcarnitines and acylglycines, saturated fatty acids (palmitic acid and stearic acid) were found to be increased in unstable plaques. Data from the Rotterdam study revealed that intake of palmitic acid was strongly associated with coronary artery disease³⁷. Consistently, feeding mice a palmitic acid-enriched diet modulates vascular smooth muscle cell (vSMC) phenotype³⁸, drives inflammasome activation³⁹, and enhances reactive oxygen species⁴⁰, cellular processes that drive vascular calcification and plaque instability^41,42. In contrast to palmitic acid, the role of stearic acid in atherosclerosis remains inconclusive. For instance, stearic acid not only lowers LDL-cholesterol but also lowers HDL-cholesterol⁴³. Also, dietary stearic acid does not increase atherosclerosis risk and is linked with reductions in systolic blood pressure and improvements in cardiac function^44–46. However, stearic acid accumulation in macrophages induces endoplasmic reticulum stress-mediated apoptosis⁴⁷, a process known to drive necrotic core formation and plaque instability. While stearic acid alone may not impact the incidence of atherosclerosis, the action of both stearic acid and palmitic acid may drive vSMC phenotypic modulation, enhance proinflammatory responses in macrophages, and increase apoptosis.

In addition to saturated fatty acids, aromatic amino acids and tryptophan metabolites were also enriched in unstable plaques. Interestingly, phenylacetic acid induces reactive oxygen species in vSMCs and TNFα production in endothelial cells^48,49. The fatty acid-conjugated amino acid N-acetyltryptophan is modified from tryptophan by the gut microbiota that is then absorbed by the intestine⁵⁰. Supporting our observation that N-acetyltryptophan was present in unstable atheromas, enhanced levels of N-acetyltryptophan are associated with an increased risk of coronary heart disease⁵¹, which may be attributed to the effect of N-acetyltryptophan on hyperlipidemia⁵². While not passing the adjusted P value, the serotonin derivative 5-hydroxyindoleacetic acid (5-HIAA) did pass the nominal P value in the fibrous cap of unstable plaques, and interestingly, 5-HIAA is associated with metabolic syndrome and C-reactive protein (CRP), a biomarker of inflammation²¹. 5-HIAA can be oxidized by myeloperoxidase to form reactive quinones, and these 5-HIAA reactive quinones are found in advanced human atherosclerotic lesions⁵³. However, a causative role for 5-HIAA in atherosclerosis has yet to be explored.

Despite not finding any metabolites passing the False Discovery Rate (FDR) correction threshold for Figures 2E and 2F, our study interestingly revealed that pyruvic acid and lactic acid were distinctly located in the fibrous cap and necrotic core of stable plaques, respectively – findings that did meet the nominal P value. During acute and chronic inflammation, a shift in immune cell metabolism towards glycolysis that leads to extracellular lactate accumulation occurs⁵⁴, and elevations in circulating lactate are associated with hypertension⁵⁵, type 2 diabetes⁵⁶, and carotid atherosclerosis⁵⁷. Lactate being enriched in the necrotic core support this concept that macrophages localized in these regions are proinflammatory. However, instead of promoting inflammatory responses that drive atherosclerosis progression, lactate plays a crucial role in resolution by driving IL-10 production in phagocytes engulfing apoptotic cells⁵⁸, inducing histone lactylation⁵⁹, and suppressing proinflammatory responses by inhibiting YAP and NFκB activation in macrophages⁶⁰. In contrast to lactate, we found pyruvate might be enriched in the fibrous caps of stable plaques. Notably, pyruvate seems to promote features that drive fibrous cap formation. For instance, deleting pyruvate kinase M2, which substantially reduces the formation of pyruvate, in vSMCs suppresses phenotypic modulation and attenuates neointimal hyperplasia⁶¹. Interestingly, our analyses did not yield any metabolites that surpassed the adjusted P value threshold when we conducted a comparison between the fibrous cap and the necrotic core in both stable and unstable plaque formations (Figures 2E and 2F). This absence of discernable difference points towards a potential similarity in the dysregulation of metabolic pathways within the fibrous cap and necrotic cores of these plaques. These findings thus highlight a key concept – that there exists a highly coordinated metabolic crosstalk between lesional cells.

Our MSI data presented here validate known metabolites associated with atherosclerosis progression while also pointing to classes of metabolites that may play significant roles in plaque instability, highlighting the potential for new avenues of research. However, several technical and sample limitations exist that reveal new opportunities to extend our findings. First, we were unable to confidently annotate the majority of metabolite-related peaks that were identified, highlighting the need to improve the currently existing platforms for high-confidence peak annotation. Thus, future technical innovations that permit accurate annotation may reveal more metabolites that play critical roles in plaque instability. Also, metabolite concentrations and their spatial distributions in post-mortem specimens used in our study may be different compared to live tissues. Therefore, future studies that examine samples collected from live donors (e.g., carotid endarterectomies) will enhance our understanding of the metabolic differences between post-mortem and live specimens. In addition, while the number of specimens used here is similar to those of other studies utilizing MSI technology^62–64 future work that includes more samples will extend our findings. However, this may be challenging due to the time- and cost-prohibitive nature of currently existing MSI platforms, which may be alleviated as technology advances. Ultimately, due to the inability to procure complete clinical data from the deceased subjects, there is a clear need for follow-up research. Future studies that involve a larger sample size will also address how factors such as age, sex, ethnicity, and the involvement of specific vascular beds (e.g., coronaries vs. carotids) might influence the metabolite profile and spatial distribution within stable and unstable atherosclerotic plaques. This enhanced focus will ensure a more comprehensive understanding of these influential factors. Nonetheless, our work presented here serves as a first step to defining an atlas of metabolic pathways involved in plaque destabilization in human atherosclerosis. We anticipate that this work will be a valuable resource for the scientific community and will ultimately open new avenues of research in cardiovascular disease.

HIGHLIGHTS.

• MALDI-MSI revealed differences in metabolic pathways between stable and unstable atherosclerotic plaques.

• Reactive oxygen species, aromatic amino acid, and tryptophan metabolism were found to be increased in unstable plaques while lipid metabolism and long-chain fatty acids were enriched in stable plaques.

• This work represents one of the first steps toward creating an atlas of metabolic pathways involved in plaque destabilization, potentially creating new avenues of research for cardiovascular disease.

Nonstandard Abbreviations and Acronyms

5-HIAA

5-hydroxyindoleacetic acid

13-HODE

13-hydroxyoctadecadienoic acid

MALDI

Matrix-assisted laser desorption/ionization

MSI

Mass spectroscopy imaging

RNAseq

RNA sequencing