MALDI Mass Spectrometry Imaging of Extracellular Matrix Proteins

Abstract

The identity, quantity, and spatial distribution of extracellular matrix (ECM) proteins in tissues defines the function of cells within. Dysregulation of ECM proteins can be coincident with or even drive various pathological conditions. Common techniques that are used to spatially detect and identify ECM proteins include magnetic resonance imaging (MRI), immunohistochemistry (IHC), second harmonic generation (SHG) imaging, and scanning electron microscopy (SEM). However, these techniques typically limit detection to only a few proteins simultaneously. Matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) has emerged as a tool capable of detecting various biomolecules such as lipids, proteins, peptides, and glycans in a spatially-defined manner due to its high molecular specificity. However, it is rarely applied to ECM proteins because of their highly cross-linked nature, relatively low abundance, and prevalence of post-translational modifications. This perspective discusses the current state and future of MALDI-MSI of ECM proteins, details the technical hurdles limiting the adoption of MALDI-MSI for ECM imaging, and summarizes the potential opportunities that MALDI-MSI has for spatially resolving ECM proteins in healthy and diseased tissues.

1. The extracellular matrix

1.1. Composition and structure of the extracellular matrix

The extracellular matrix (ECM) is the non-cellular component of tissues, and it is a complex network of highly cross-linked proteins and sugars that play both structural and biochemical roles in tissue development and function.^1,2 In sum, the ECM is comprised of approximately 300 core proteins, with collagens accounting for 30% of the total ECM protein content by mass.^3,4 Other key components include fibronectin, elastin, and laminins, which are interconnected via proteins such as nidogen and heparan sulfate proteoglycans.⁵ Tissues have unique ECM compositions, and the cells residing in each tissue produce tissue-specific ECM components.^5–7 Given the vast number of different ECM proteins, we cannot discuss each of them in a concise perspective piece. Instead, we have briefly summarized some key features of select ECM proteins that are the most abundant in human tissues and then describe how MALDI-MSI may be used to better spatially identify them (Table 1).

Table 1.

Select, abundant ECM proteins, their function, structure properties, and distribution in and across tissues.

ECM protein	Function	Structural properties	Tissue distribution
Collagens	Structural support and impart tensile strength to tissues.2	Form fibrous, fibrillar networks, or are fibril-associated.8 High proline, hydroxyproline, and glycine content stabilize the triple helix structure of collagens,8 undergo lysine hydroxylation that strengthens collagen structure.9,10	Types I, III, VI, VIII co-distributed throughout tissues and type II & IX are abundant in connective tissues2. Other collagens are less abundant.
Elastin	Tissue elasticity and resiliency.6	60 kDa molecular weight,11 forms elastic fibers of different shapes based on tissue needs, can form continuous sheets in pulmonary vessels.12	Abundant in large arteries that experience pulsatile flow and tissues that undergo stretch and relaxation (e.g. lung, skin, bladder, etc.).6
Proteoglycans	Assemble ECM; aggrecan provides elasticity; decorin regulates collagen assembly.6,13	Core protein with glycosaminoglycan (GAG) chains, examples include versican, aggrecan, decorin, and lumican.6	All tissues.6,13
Laminins	Primary component of the basement membrane. Developmental and organogenesis, organ homeostasis.14	~850 kDa molecular weight,15 assembles into networks with collagen IV, nidogen, and others.14	All tissues.14
Fibronectins	Regulate adhesion, migration, differentiation.6,16 Promote wound healing, and re-epithelization at injury sites.17,18	230–270 kDa dimers,16 polymerize into networks, multiple binding sites (collagens, heparins, integrins, etc.).	Most abundant in connective tissues and in wound repair.19

Given the diverse and complex nature of the ECM, understanding its composition and function in various tissues is critical. There has been increasing interest to study the specific roles of the ECM proteins in various diseases such as cardiovascular and pulmonary diseases,^20,21 as well as its role in cancer invasion, embryonic development, liver regeneration, and wound healing.^7,22–24 Research has focused on how changes in ECM composition, structure, and stiffness contribute to disease, with ECM molecules influencing cancer progression and offering potential biomarkers for diagnosis and prognosis.²⁵ Spatial information about ECM proteins is essential for elucidating disease mechanisms, as it can reveal whether elevated ECM expression correlates with localized changes in cell behavior or therapeutic responses.

While immunohistochemistry (IHC) has been a key method in biomarker discovery, its limitations, such as antibody availability and spectral overlap, restricts its multiplexing ability.^26–28 Further, although IHC can be used to confirm existence of a protein within a tissue, it requires prior knowledge of the target. Because it is an antibody-based technique, immunohistochemistry often cannot differentiate ECM proteins with varying post-translational modifications (e.g., hydroxylation, oxidation, glycosylation).^29–31 These modifications are crucial for protein function and may offer valuable insights into disease mechanisms, opening the door to new therapeutic strategies. In addition, the antibody-based methods may not reveal changes in protein modifications that occur due to biological or pathogenic processes. While antibodies that are specific for certain post-translational modifications (PTMs) are available, they usually are not specific for ECM proteins that contain these modifications, thereby minimizing their utility for ECM studies. In contrast, MALDI-MSI allows for discovery-driven, target-agnostic, spatial mapping of biomolecules with virtually unlimited multiplexing.²⁷ Combining MALDI-MSI with IHC could enable more targeted analysis, and integrating MALDI-MSI into clinical practices could enhance diagnostics and biomarker discovery.^26,27,32

1.2. Cell-ECM crosstalk and dynamics

Cells adhere to the ECM via the heterodimeric transmembrane receptors integrins and syndecans, which both drive activation of intracellular signaling, transcription, cell phenotype, differentiation, and even translation of new ECM proteins.^2,6,33–36 The ECM is dynamically and continuously remodeled by fibroblasts (and less frequently by endothelial cells, osteoblasts, lymphocytes, neutrophils, and more). Fibroblasts secrete matrix metalloproteases (MMPs) and cathepsins that degrade ECM proteins.³⁷ In healthy tissues, this dynamic reciprocity between ECM deposition and subsequent degradation by proteases maintains tissue homeostasis,³⁸ but in diseased tissues, dysregulation in protease expression and activity leads to scarring, blistering, and more. Techniques such as MALDI-MSI can help to visualize these ECM and enzyme dynamics during both tissue homeostasis and disease.

Given that tissues have unique ECM compositions, and that these compositions are dynamic in nature, there is increasing interest in studying the specific roles of ECM proteins in various diseases.^7,20–24 For example, in a study on triple negative breast cancer (TNBC) associated with obesity,³⁹ proteomic analyses revealed distinct differences in the protein composition of lean vs. obese and tumor tissues. Collagen VI, collagen XII, fibronectin, laminin V, elastin, and vimentin were all upregulated in the ECM of tumor and obese mammary glands. Furthermore, invasion and migration assays indicated that collagen VI increased the migration of MCF-10As (a breast cancer cell line) more than other ECM proteins via the epidermal growth factor receptor (EGFR) and mitogen-activated protein kinase (MAPK) signaling pathways.³⁹ While these studies revealed new roles of collagens in breast cancer, adding spatially resolved information about ECM proteins would provide more localized information, which could help monitor changes in tumor progression throughout treatment.

Collagen type V α2 (COL5A2) in colorectal cancer is a target for attenuating tumor cell growth and progression in mammary tumors.³⁵ Overexpression of COL5A2 in different cell types, including a colon cancer cell line, SW60, caused cells to proliferate more and led to the activation of the PI3K/mTOR signaling pathway.⁴⁰ While these studies highlight COL5A2’s role in cancer progression, spatial information on its distribution within the tumor microenvironment, such as that potentially gained via MALDI-MSI, could determine if cell growth and/or drug sensitization is localized to areas of high COL5A2. Collagen V has been associated more broadly with tumor progression and malignancy as well.^22,40,41 Laminin and fibronectin have elevated expression in tumors such as breast, lung, and brain, thereby promoting growth and invasion of cancer cells.^42,43 Collagen VIα3, fibrillin-1, fibronectin, fibrinogen (α,β,γ), and periostin have increased expression in pancreatic ductal adenocarcinoma.⁴⁴ These studies and many more demonstrate that dysregulation of ECM proteins is causative or at least coincident with tumor growth. Since the spatial distribution of ECM proteins is crucial for defining the function and morphology of the cells and tissues,^45–47 understanding how these distributions change during disease progression, such as may be captured with MALDI-MSI, is essential for deciphering the precise disease mechanisms and identifying potential therapeutic targets.

Most studies of ECM proteins have relied on proteomics, which is target-agnostic but lacks spatial information, or immunohistochemistry, which can identify proteins spatially but requires a prior knowledge of the targets and has limited multiplexing capability.^48,49 Monitoring multiple ECM proteins at once, without the need for a target-specific antibody, would provide the most comprehensive snapshot of how ECM shapes tissue homeostasis and disease progression. Below are some of the common ways to image and visualize ECM proteins with a focus on MALDI-MSI as an emerging tool for measuring ECM protein distributions in tissues.

2. Common methods to image and quantify the ECM

Common methods to image protein distributions in tissues are magnetic resonance imaging (MRI), IHC/fluorescence microscopy, second harmonic generation (SHG) imaging, scanning electron microscopy (SEM), and two-photon excitation microscopy (Table 2).^46,50,51 In IHC, an antibody is used that binds to the protein of interest with high affinity. These antibody-antigen complexes are then imaged, typically with fluorescence microscopy using fluorophore-conjugated antibodies.²⁸ Given the limited multiplexing of traditional IHC, methods like mass cytometry have been developed, which allows up to 40 proteins to be imaged simultaneously.^52,53 Mass cytometry uses an inductively coupled plasma time-of-flight mass spectrometry (ICP-TOF-MS) readout to measure metal signals from metal-tagged antibodies. This approach has recently been used to study the relationship between cellular niches and the surrounding ECM in allergic airway inflammation in lungs.⁵⁴ While mass cytometry enables an increased number of proteins to be imaged simultaneously, it is still a targeted approach that relies on antibodies and pre-existing knowledge of the target.

Table 2.

Advantages and limitations of commonly used techniques to identify and visualize the ECM proteins.

Imaging technique	Advantages	Limitations	Commonly imaged ECM and non-ECM components
Second harmonic generation imaging (SHG)	Used on live, unstained, unfixed tissues, high resolution (400–1000 nm).55	Limited to a small number of noncentosymmetric proteins (collagens 1–3, myosin),49 limited to thin samples.55	Fibrillar collagen,57,55 myosin, polysaccharides48,66
Scanning electron microscopy (SEM)	High resolution imaging (below 10 nm for high-contrast materials),58 target agnostic.59	Destructive, structure/morphology only.66	Direct references to collagen28, but can be used to see all proteins (cannot distinguish them).
Two-photon excitation microscopy	Deep tissue imaging compared to confocal microscopy.62	High photobleaching, limited wavelengths/spectra.49,62	Elastin,60 auto-fluorescent molecules, such as NAD(P)H,68,69 lipids, and collagens 1–3.
Magnetic Resonance Imaging (MRI)	Non-invasive51, deep tissue penetration.	Lack of standards for data acquisition and analysis.51	Collagen,61 elastin/tropoelastin, ECM fibrosis.70
Immuno-histochemistry (IHC)	Protein specific, easy to use, determines localization of the proteins, can be used to study cell-ECM interactions.28	Requires prior knowledge of sample to target proteins of interest, tissue fixation and processing artifacts.51	Cytoskeletal proteins,71 cellular proteins,58 various ECM proteins including collagen,49,72,73 elastin, fibronectin.74
Mass cytometry	Higher multiplexing capability than IHC, protein specific.52,54	Requires prior knowledge of sample to target proteins of interest,52 isotopically pure lanthanide metals are needed, signal intensity can be varied over time.75	Hyaluronan, type 1 and type 3 collagen,54 cell surface proteins, cell signaling proteins.76

SHG microscopy can identify collagen fibrils in cells and tissues via the optical polarization caused by collagen’s helical structure.^55,56 It is commonly used to study changes in fibrillar collagen structure in various pathological conditions, such as lymph node remodeling, where alterations to the fibrillar collagen structure are evident in metastatic lesions.⁵⁷ SHG enables imaging live and unfixed tissues, without any staining, at resolutions ranging from 400–1000 nm but can only provide morphological information.⁵⁵ SEM provides morphological information, with high resolution information at length scales as low as 10 nm, but only for high-contrast materials.^58,59 Two-photon excitation microscopy and MRI have been used primarily to image elastin and collagen, respectively, both on living tissues.^60,61 Advantages of these techniques are that both are non-invasive,⁵¹ and two-photon excitation microscopy enables deeper tissue imaging than traditional confocal microscopy.⁶² Although they have distinct advantages, these techniques all suffer from limited multiplexing. MS-based proteomics methods work without antibodies, without the limitation of spectral overlap, can reveal the presence of PTMs, and also provide relative abundances of proteins.^63–65 However, MS-based proteomics does not provide spatial information.²⁹ Mass spectrometry imaging is an emerging tool that combines spatial information from imaging with proteomics capabilities.

3. Mass spectrometry imaging

The acquisition of spatially-resolved molecular-level information directly from biological specimens using MS is possible by coupling ionization techniques like MALDI, desorption electrospray ionization (DESI), secondary ion mass spectrometry (SIMS), laser-ablation inductively coupled plasma mass spectrometry imaging (LA-ICP-MSI), matrix-assisted laser desorption electrospray ionization (MALDESI),⁷⁷ and laser ablation electrospray ionization (LAESI).^78–81 MS can be used to detect a wide variety of biomolecules such as peptides, proteins, lipids, glycans, and metabolites as well as trace elements.^82,83 In this perspective, we focus on MALDI-MSI as an emerging tool to study the distribution of ECM proteins.⁸⁴

3.1. MALDI-MSI to spatially locate biomolecules

MALDI-MSI has been used to image the biodistributions of a wide range of molecules in a variety of ex vivo tissues.^78,85,86 It also has the potential to provide spatial information about ECM proteins. This technique utilizes an organic matrix that co-crystallizes with the analytes of interest⁸⁷ and absorbs energy from the laser to enable desorption and ionization of the analytes present in the sample. The m/z ratio of the analytes is often measured on a time-of-flight (TOF) mass analyzer, but other mass analyzers can be used as well, such as orbitraps and quadrupole-TOFs. MALDI-MS can detect many biomolecules including proteins, peptides, glycans, lipids, and metabolites.⁸⁸ When MALDI is used to image the distribution of molecules in tissues, sample preparation is critically important. Most analyses are performed on fresh frozen tissue sections that have a thickness of ~10–20 μm.⁸⁹ The tissue section is placed on a conductive glass slide, and the organic matrix is applied onto the tissue to facilitate ionization. A laser is rastered across the surface of the matrix-coated tissue section, generating thousands of spectra that can be reconstructed into images that indicate the locations of the measured ions, thereby revealing molecular distributions across the tissue (Figure 1).

Figure 1.

Biomolecular imaging of tissues using MALDI-MSI. The schematic shows the spatial distributions of the detected ions in the tissue sample.

Among the various analytes imaged with MALDI-MSI, lipids have been most extensively studied.⁹⁰ Lipids require minimal prior tissue treatment to be detected because of their abundance and ease with which they are ionized by MALDI. Changes in lipid distribution have been studied in various healthy and pathological conditions, such as cancer, and in tissues such as brain, liver, intestines, and embryos.^84,91,92 Choosing the right matrix is important to enable effective analyte ionization and maximize analytical sensitivity.⁹³ 2,5-dihydroxybenzoic acid (DHB) is one of the mostly commonly used matrices for lipid detection in the positive mode of MALDI-MSI.⁹⁴

N-glycans are another commonly measured biomolecule in MALDI-MSI.^95,96 Over 50% of human proteins are glycosylated, making it the most common PTM, and ~80% of cell surface and ECM proteins are glycoproteins.⁹⁶ Sample preparation and pre-treatment, such as using PNGaseF to deglycosylate proteins, are usually necessary to enable efficient detection and imaging of N-glycans in tissues.⁹⁷ For N-glycan imaging, DHB and α-cyano-4-hydroxycinnamic acid (CHCA) matrices are most often used.⁹⁶

Protein analysis using MALDI-MSI is more challenging. Intact proteins are not detected as sensitively as smaller molecules due to poor matrix-protein co-crystallization and ion suppression by more abundant low molecular weight compounds like lipids. For intact proteins, a matrix such as sinapinic acid or even a mixture of two matrices, such as CHCA and DHB,⁹⁸ are used. Proteins are more commonly detected indirectly via their peptide fragments that are produced after enzymatic digestion with proteases such as trypsin. The detected peptides are then matched to the protein from which they arise using protein databases.⁸⁶ Proteolysis and peptide-based detection increase the number of different proteins that can be detected due to better matrix-peptide co-crystallization and less ion suppression.⁹⁹ Often, tissue washing steps using solvents such as ethanol, Carnoy’s fluid, and water are required to remove highly abundant lipids and salts before enzymatic digestion to further avoid the suppression of peptide ion signals (Figure 2).^82,100 These washing steps must be optimized to avoid any delocalization of the analytes of interest. The most commonly used matrices for peptides are CHCA and DHB. Correctly identifying peptides, and the proteins from which they come, is often achieved from protein databases that are created from liquid chromatography and tandem MS (LC-MS/MS) datasets that are acquired on adjacent tissue sections.⁸⁶ Detecting and matching two or more unique peptides per protein helps increase confidence in protein identification by MALDI-MSI.

Figure 2.

Typical workflow for protein detection using MALDI-MS imaging. (a) The fresh frozen tissues are first sliced, (b) placed on a conductive slide, (c) washed to remove interfering compounds like lipids, (d) deposited with trypsin for proteolytic digestion and (e) coated with matrix. (f) The tissues are then imaged using MALDI-MS. Created using BioRender.com

MALDI-MSI works best on fresh-frozen tissue sections, yet the majority of clinical biopsies are fixed and stored as formalin-fixed paraffin-embedded (FFPE) tissues, which pose a unique challenge.⁸⁹ Formalin-fixing results in protein cross-linking via primary amines, which stabilizes them for histological analyses¹⁰¹ but makes proteolytic digestion with trypsin and subsequent imaging by MALDI-MSI prohibitively difficult. Recent work has shown that antigen retrieval with citraconic anhydride and deparaffinization with xylenes can improve detection of both lipids and peptides in such tissues, thereby expanding the capability of MALDI-MSI as a tool for analyzing clinical specimens;^89,102–105 however, peptide signals are still higher from frozen tissue sections.

3.2. Advantages of MALDI-MSI

MALDI-MSI has a variety of advantages compared to other techniques used for imaging biological and clinical specimens. MS is a universal method that relies on an inherent property of all analytes (i.e. mass) and usually does not require analyte derivatization for its detection and imaging.^82,83,85 MALDI-MSI can detect hundreds of molecules in a single experiment.^49,51,85 In addition, unlike most other imaging techniques, MALDI-MSI is an untargeted approach and does not require any prior knowledge of the tissue composition, allowing it to be a valuable discovery tool.⁸³ MALDI-MSI can be used to generate molecular snapshots of tissues in a multiplexed manner at resolutions down to ~10 μm.¹⁰⁰ Recent studies have coupled MALDI-MSI with IHC, maintaining the targeted nature of IHC while leveraging the inherent multiplexing and imaging capabilities of MALDI-MSI.^27,32

MALDI-MSI is often complemented by staining techniques such as hematoxylin and eosin (H&E) or antibody staining to enhance anatomical interpretation of tissue sections.¹⁰⁶ H&E staining can reveal any abnormal morphologies, providing spatial context for the molecular information from MALDI-MSI. For example, a study by Seeley and Caprioli used MALDI-MSI with histological stains to explore breast cancer progression, identifying distinct proteomic profiles by MALDI-MSI in different tissue regions (ductal carcinoma, invasive carcinoma, and stroma) that were evident from staining. Specific proteins like histone H2A were visible in ductal carcinoma, and thymosin β4 was present in the stroma.¹⁰⁷ The technique has been applied broadly to study changes in biomolecule distributions in tissues such as brain, breast, lung, and liver, under various pathological conditions.^84,107–111 These studies demonstrate the capability of MALDI-MSI to map proteins within tissues and disease-related regions, as well as to identify potential markers associated with disease mechanisms and/or tissue remodeling.

4. MALDI-MSI as an emerging technique to study ECM proteins

It is important to map the spatial arrangement of ECM proteins in tissues to better understand how changes in their distributions affect the morphology, biochemistry, and function of tissues and cells in healthy and diseased states. ECM proteins are highly cross-linked and large (>100 kDa), making them more difficult to detect and image by MALDI-MS(I) than intracellular proteins.^29,112–114 Some ECM proteins also form higher-order molecular structures, such as collagen triple helices, disulfide-bonded fibronectin dimers, and covalently cross-linked elastin fibers, which further complicate their detection. These chemical properties make ECM proteins less soluble than intracellular proteins, posing detection challenges as they are difficult to digest and liberate for MALDI-MS analysis. Moreover, because many ECM proteins are glycosylated,^29,112,114 they are often difficult to digest because their cleavage sites are inaccessible, complicating their analyses by conventional MALDI-MSI approaches that rely on peptide fragment detection.

Because of these technical hurdles, relatively few studies have used MALDI-MSI to image ECM protein distributions. An early study used decellularization to remove more easily detected soluble proteins present in cells of a whole tissue.¹¹⁴ This study took inspiration from tissue engineering, where decellularization is commonly used to isolate ECM components while preserving their native architecture.^115,116 In this study, the researchers used SDS to decellularize whole tissues, followed by tissue sectioning, proteolytic digestion with trypsin, and analysis via MALDI-MSI.¹¹⁸ Decellularization led to a 50% increase in detection of peptides from ECM proteins.¹¹⁴ Our own group has expanded on this technique, demonstrating that decellularization of liver tissue sections (rather than whole tissues) using 0.1% SDS allowed detection of threefold more ECM proteins than possible without decellularization, including collagens II and IV, emilin-2, fibrinogen α, hemicentin-1, tenascin-C, thrombospondin-1, and laminin α2 and α3.¹¹⁷

Another way to improve ECM protein detection is to utilize more ECM targeted enzymes. Lysine residues in collagen and elastin are often hydroxylated, limiting proteolysis by trypsin.¹⁰⁵ For this reason, collagenase III can sometimes be a better choice for digesting ECM proteins.^105,118 Angel et al. demonstrated that collagenase III digestion enhances detection of collagen peptides compared to untreated tissues (Figure 3), enabling the mapping of many more ECM proteins.¹⁰⁵ In a related study, this approach was applied to low-grade lung adenocarcinoma, enabling detection of 40 proteins, 67% of which were ECM proteins, including collagen, emilin-1, tenascin, fibronectin, and vimentin.¹¹¹ Notably, MALDI-MSI could track hydroxylation of proline and lysine residues in collagen after comparison to parallel LC-MS/MS experiments, thereby helping identify ECM proteins with PTMs. The findings suggested that collagen peptides with hydroxylated prolines were found more extensively in the early grade lung adenocarcinoma as compared to healthy lungs.¹¹¹ These studies are excellent examples of using ECM-specific enzymes such as collagenase III to digest and detect collagen, elastin, and other ECM proteins, as well as identify PTMs using MALDI-MSI in tissues at different stages of disease progression. Despite their importance, PTMs on ECM proteins remain underexplored by MALDI-MSI. To our knowledge, hydroxylation of ECM proteins has been the primary PTM detected on the peptide digestion products of ECM proteins. Increasing the types of PTMs that can be detected on ECM proteins represents an important area for future MSI studies.

Figure 3.

MALDI-MS images of liver tissues with and without collagenase III digestion. (a) Composite mass spectra across the entire liver tissue section with and without collagenase III digestion. (b) Representative MALDI-MS images with and without collagenase III treatment.¹⁰⁵ Figure adapted with permission from P. M. Angel, S. Comte-Walters, L. E. Ball, K. Talbot, A. Mehta, K. G. M. Brockbank and R. R. Drake, J. Proteome Res., 2018, 17, 635–646. 2018 American Chemical Society

Additionally, many ECM proteins are glycosylated, which complicates proteolytic digestion and increases spectral congestion due to the molecular heterogeneity associated with glycans.^95,96 Another tactic to increase ECM protein signals in MALDI-MSI is to perform serial enzymatic digestions.¹¹² Clift et al. applied chondroitinase for glycosaminoglycan (GAG) digestion, PNGaseF for deglycosylation, and elastase for elastin digestion before finally digesting the remaining ECM proteins with collagenase type III.¹¹² This series of enzymatic digests improved the access of collagenase III to the ECM proteins in the tissues and improved the resulting mass spectra. In a direct comparison, tissues treated with a series of enzymes followed by collagenase III showed a 3.1% increase in signal for proteins such as collagen VIα3, fibrinogen, decorin, elastin, and periostin, and an 82% increase in the ECM peptides detected compared to treatment with collagenase III alone.

Our group has also demonstrated the value of serial digestions in combination with decellularization of tissue sections. Treating decellularized tissue sections with PNGaseF for deglycosylation and trypsin for proteolytic digestion improved MALDI-MSI of ECM proteins.¹¹⁷ We observed a threefold increase in the number of ECM proteins identified after decellularization and deglycosylation compared to decellularized tissues that were not deglycosylated. Ion signals for glycosylated ECM proteins like tenascin-C were particularly enhanced after deglycosylation, while collagen I peptide signals were less improved upon deglycosylation, consistent with collagen’s lower glycosylation levels (Figure 4).¹¹⁷ Overall, these studies suggest that sample preparation should be tailored to the target protein of interest. They also indicate the promise of MALDI-MSI for detecting a wide variety of ECM proteins, including glycosylated proteins.

Figure 4.

Representative MALDI-MS images showing the spatial distribution of peptide fragments from ECM proteins in liver tissue. Shown are peptides from collagen I (α1) (left) and tenascin-C (right) in samples with and without decellularization (DC and NDC, respectively). Tissue sections were digested with either trypsin alone or a combination of trypsin and PNGase F after lipid removal.¹¹⁷ Figure adapted with permission from A. Rampal, I. F. de la Fuente, N. K. Vu, J. Doungchawee, U. Ranjan, S. R. Peyton and R. W. Vachet, Anal. Chem., 2025, 97, 886–893. 2025 American Chemical Society.

While in this perspective we have primarily focused on ECM protein imaging, proteoglycans and their GAG chains are also important for ECM structure and function. As mentioned earlier, a common way to detect N-glycans is to digest the ECM proteins using PNGaseF and separately detect the resulting glycans. In addition, the detection of GAGs can further be facilitated with enzymes such as chondroitinase.^95–97 A recent study by Devlin et al. further improved the separation and spatial analysis of GAGs, like chondroitin sulfate, dermatan sulfate, and hyaluronic acid, using trapped ion mobility spectrometry (TIMS).¹¹⁹ This approach represents an important advance for expanding MSI detection of proteoglycans, thereby offering a more comprehensive view of the ECM.

5. Limitations and future perspectives on MALDI-MSI of ECM proteins

In sum, MALDI-MSI is a powerful approach to detect and locate proteins in tissues without the need for prior knowledge of the targets. There are at least three roadblocks that continue to limit progress in MALDI-MSI of ECM proteins: low abundance, low detection sensitivity, and low digestion efficiency. Decellularization, deglycosylation, and serial targeted enzyme application, are all promising methods to overcome at least two of these issues, and we expect increasing adoption for MALDI-MSI for ECM proteins.

On-tissue chemical derivatization (OTCD) approaches could help overcome the limited detection sensitivity for some ECM proteins by specifically tagging proteins or their peptide digestion products to enhance their ionization efficiency.¹²⁰ Recently, OTCD approaches have improved the detection of specific amino acids, N-glycans, hormones, fatty acids, and many other analytes during MALDI-MSI.^120–122 New OTCD reagents might selectively enhance ECM protein/peptide signals for some ECM proteins. For example, structural support proteins like collagen and elastin are crosslinked due to the presence of allysine residues that are formed upon oxidation of lysine residues by lysyl oxidase (LOX).^123,124 Using a hydrazine-based reagent to label these residues could potentially improve detection of collagen and elastin by MALDI-MSI since allysine are specific to these ECM proteins. If wisely designed, such hydrazine-based reagents could also facilitate the identification of the proteolytic peptides from these ECM proteins if they were isotopically encoded.

Another avenue of research would be to improve the detection of ECM proteins with PTMs still attached. At least one study has revealed that hydroxylated prolines are found more extensively in collagen from early stages of lung cancer as compared to healthy lung samples.¹¹¹ PTMs of various types are associated with the progression and onsets of other cancers,^125,126 and the ability to find the unique spatial distributions of proteins with specific PTMs would represent a significant advance in clinical diagnoses, especially since antibody-based approaches often fail to detect such modifications. Other MS-based imaging techniques such as desorption electrospray ionization (DESI),^127,128 which has proven effective for the detection of proteins and post-translationally modified proteins, could also be applied to this detection problem, if appropriate methods to overcome the highly crosslinked nature of ECM proteins are developed.

6. Summary

As our understanding of the ECM’s role in health and disease continues to expand, so does the need for tools that can capture its molecular complexity in situ and in a spatially-resolved manner. MALDI-MSI holds significant promise for overcoming current limitations in ECM imaging by enabling multiplexed, label-free detection of a broader range of ECM proteins. Continued innovation in sample preparation, enzymatic digestion, and data analysis will be key to unlocking the full potential of MALDI-MSI in this space, ultimately offering new insights into ECM biology and its contributions to disease.