MALDI and DESI: Complementary Lipid Imaging on a Single Mass Spectrometry System

MALDI and DESI: Complementary Lipid Imaging on a Single Mass Spectrometry System — Summary

Significance of the topic

The spatial distribution and composition of lipids in atherosclerotic plaques are critical determinants of plaque stability and the risk of rupture that can lead to myocardial infarction, stroke or local tissue infarction. Mass spectrometry imaging (MSI) enables label-free, spatially resolved molecular profiling directly from tissue sections with minimal sample preparation compared with bulk LC‑MS workflows. Combining orthogonal MSI ionization methods—MALDI (matrix-assisted laser desorption/ionization) and DESI (desorption electrospray ionization)—can broaden lipid coverage, reveal complementary molecular information, and support more robust biological interpretations of heterogeneous tissue such as carotid atherosclerotic plaques.

Objectives and study overview

This application study compared MALDI and DESI MSI performed on a single SYNAPT XS mass spectrometer to assess: (1) the breadth of putatively identified lipid species detected by each ionization technique, (2) differences in spectral patterns, sensitivity and image quality, and (3) which lipid classes are preferentially ionized by each method. Fourteen human carotid plaque specimens were analyzed by both DESI and MALDI on consecutive tissue sections to allow direct comparison.

Methodology

Sample preparation and acquisition

• Human carotid plaques (n = 14) were snap‑frozen, sectioned at 10 µm after embedding in 10% porcine gelatin, and imaged as consecutive sections for DESI and MALDI to permit paired comparisons.
• MALDI matrix: 2,5‑dihydroxybenzoic acid (DHB), 10 mg/mL in 70% acetonitrile; applied with a SunCollect sprayer, 25 layers, 1.5 mm line spacing, N2 nebulizing gas at 1.7 bar and 20 µL/min flow.
• DESI sections were analyzed without additional sample preparation.

Instrument and acquisition parameters

• Mass spectrometer: SYNAPT XS, positive ion mode, acquisition m/z 100–1200, analyzer set to sensitivity mode.
• DESI source: solvent 98% methanol / 2% water, flow 2 µL/min, capillary voltage 0.6–0.8 kV, cone 30 V, source temp 150 °C, heated transfer line 250 °C, nitrogen 10–15 psi, pixel size 100 µm, equivalent scan speed 2 scans/s.
• MALDI source: laser 1 kHz, attenuation 260, 500 shots/pixel, cooling gas 50 mL/min, pixel size 100 µm, equivalent scan speed 2 scans/s.

Data processing and identification strategy

• Raw data inspected in MassLynx and visualized in HDI (v1.9.1). A top‑2000 peak target list was compiled across acquisitions.
• Representative files were processed in the Analyte Browser MicroApp (v2.2.2) and searched against LIPID MAPS with a mass tolerance of ±5 ppm.
• Putative identifications considered common adducts [M+H]+, [M+Na]+ and [M+K]+; duplicate molecular formulas were removed (to account for multiple adducts and identical formula matches).
• Note: Analyses were performed in MS1 only; phosphatidylcholine (PC) and phosphatidylethanolamine (PE) annotations in positive mode may be ambiguous (PC favored), as no diagnostic MS/MS headgroup data were acquired.

Used instrumentation

• SYNAPT XS Mass Spectrometer (Waters)
• DESI XS source with high‑performance sprayer and heated transfer line
• SunCollect sprayer (SunChrom) for MALDI matrix deposition
• Software: MassLynx, HDI (v1.9.1), Analyte Browser MicroApp (v2.2.2)

Main results and discussion

Global detection and image quality

• Both MALDI and DESI produced clear lipid images at 100 µm pixel size with comparable distributions for shared analytes (example: m/z 725.5560, putatively SM 34:3, was robustly detected by both methods).
• DESI images often exhibited slightly sharper feature definition and somewhat higher overall signal intensity in examined ROIs, potentially due to matrix application effects in MALDI that can broaden spatial signal.

Spectral differences and fragmentation patterns

• MALDI spectra showed notable signal at m/z 184.074 consistent with phosphocholine headgroup fragments, suggesting partial in‑source fragmentation for some PCs under the applied MALDI conditions.
• DESI spectra revealed a strong m/z 301.14 signal, attributed to a solvent/background contaminant present off‑tissue in these acquisitions.

Coverage statistics and complementarity

• After duplicate removal, MALDI yielded 465 unique putative lipid annotations and DESI yielded 428; the overlap between methods was modest (34 lipids detected by both), indicating significant complementarity.
• Unique annotations: 431 lipids observed only in MALDI and 394 observed only in DESI (numbers derived from unique counts minus shared identifications).

Class‑level selectivity

• Across combined data, 26 lipid sub‑classes were observed (classes defined when combined identifications exceeded 15 species). Examples include ceramides, triacylglycerols, oxidized glycerophospholipids, phosphocholines/phosphoethanolamines, glycerophosphoinositols, fatty esters and steroid conjugates.
• Applying an ≥80% class identification cutoff revealed eight classes with pronounced differences between ionization modes:
• Classes favoring MALDI: eicosanoids; glycerophosphoglycerides/glycans; glycerophosphoinositols; oxidized glycerophospholipids; assorted phospholipids and steroid conjugates.
• Classes favoring DESI: fatty esters and triacylglycerols.

Benefits and practical applications

• Either MALDI or DESI alone provides broad lipid coverage in plaque tissue, but using both sequentially substantially increases the number of putative lipid detections and class coverage.
• DESI is non‑destructive and requires no matrix application, enabling prior DESI imaging followed by MALDI imaging or histological staining on the same section to maximize information from limited material.
• Selection of ionization approach can be tailored to analytical goals: MALDI for classes prone to fragmentation or requiring matrix‑assisted extraction/crystallization; DESI for neutral lipids and triacylglycerols.
• Workflow integration with HDI and Analyte Browser MicroApp enables rapid, high‑throughput putative annotation against lipid databases for exploratory studies.

Limitations

• Identifications are putative (MS1 only, ±5 ppm) and can be confounded by isomeric/isobaric species and multiple adduct forms; orthogonal MS/MS or ion mobility separations are needed for confident structural assignments.
• Matrix application can affect apparent spatial resolution in MALDI; DESI is susceptible to solvent contaminants and surface‑dependent ionization variability.

Future trends and potential applications

• Routine incorporation of MS/MS imaging and ion mobility separation (e.g., IMS) to resolve isomers and provide confident structural assignments for lipids in situ.
• Higher spatial resolution MSI (smaller pixel sizes) while maintaining sensitivity to visualize micro‑domains within plaques and map cell‑specific lipid signatures.
• Standardized, multi‑modal workflows: non‑destructive DESI first, then MALDI and downstream histology on the same section to integrate molecular and morphological data.
• Improved automated annotation pipelines, curated spectral libraries for tissue MSI, and AI‑assisted annotation to reduce false positives and accelerate discovery.
• Quantitative MSI or hybrid LC‑MS validation of key findings to translate putative spatial markers into validated biomarkers for plaque stability and clinical research.

Conclusion

The study demonstrates that MALDI and DESI MSI on a single SYNAPT XS system are complementary approaches for lipid profiling in heterogeneous human atherosclerotic plaques. Each ionization method preferentially ionizes particular lipid classes and yields many unique putative annotations; combining both methods increases molecular coverage and supports more comprehensive spatial lipidomics. A practical workflow is to perform non‑destructive DESI imaging first, followed by MALDI imaging and orthogonal analyses as needed, while employing MS/MS or ion mobility for definitive identification in targeted follow‑up studies.

References

1. Slijkhuis N, Towers M, Claude E, van Soest G. MALDI versus DESI Mass Spectrometry Imaging of lipids in atherosclerotic plaque. Rapid Commun Mass Spectrom. 2025 Jan 15;39(1):e9927. doi: 10.1002/rcm.9927. PMID: 39435741.