Direct Molecular Tissue Analysis by MALDI Imaging Mass Spectrometry in the Field of Gastrointestinal Disease

Abstract

Mass spectrometry has become among the most important analytical approaches in life sciences in the last few years because it allows the unlabeled and multiplexed analysis of a broad variety of molecules ranging from small molecules to proteins and their modifications in biological samples. In tissues, all these molecule classes may vary according to the cell type and the disease state of the organism. Because tissues are complex systems of molecular and morphologic heterogeneity, a differentiated tissue analysis can only be performed by morphology- or microscopy-based techniques. Matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry (“MALDI imaging”) combines mass spectrometry with microscopy and has emerged as a promising technique for combined morphologic and molecular tissue analyses. It enables a spatially resolved and unlabeled imaging of different molecules in their histologic context and the allocation of molecular profiles to specific cell types like tumor, preneoplastic, or inflammatory cells. MALDI imaging has proven to provide novel and clinically relevant information to a variety of different biomedical questions, with focus in oncology and inflammatory diseases, including gastrointestinal disorders.1Schwamborn K. Caprioli R.M. Molecular imaging by mass spectrometry–looking beyond classical histology.Nat Rev Cancer. 2010; 10: 639-646Crossref PubMed Scopus (274) Google Scholar, 2McDonnell L.A. Corthals G.L. Willems S.M. et al.Peptide and protein imaging mass spectrometry in cancer research.J Proteomics. 2010; 73: 1921-1944Crossref PubMed Scopus (128) Google Scholar In these studies, molecular signatures could be correlated with disease phenotypes or compiled clinical information, for example, histologic tumor type, disease stage, patient outcome, or therapy response. Of increasing interest is the ability of MALDI imaging to localize drugs and their metabolites in tissues, which is valuable for drug development and efficacy studies in animal models and even in individual patients. Because of its practical simplicity and ability to gain reliable information, even from the smallest tissue amounts, which may also originate from endoscopic biopsy sections, MALDI imaging might have the potential to complement histopathologic evaluation for assisting in diagnostics, risk assessment, or response prediction to therapy (Figure 1). Mass spectrometry (MS) relies on separating ionized molecules on the basis of their mass-to-charge ratio (m/z) to determine the chemical composition of samples. This results in a mass spectrum, which plots the measured ion intensity versus its m/z value. Soft ionization techniques, like the MALDI, enable the ionization of a broad variety of molecules ranging from small molecules to molecules of high molecular weight, such as peptides and proteins, and hence the investigation of the molecular complexity and dynamic nature of proteomes, including posttranslational modifications. MALDI requires the sample to be mixed with a laser-energy–absorbing matrix, where each type of matrix favors the ionization of different molecule classes. The 3 most commonly employed matrices are sinapinic acid for proteins, α-cyano-4-hydroxycinnamic acid for peptides and small molecules, and 2,5-dihydroxybenzoic acid for peptides, lipids, and small molecules. The analyte–matrix co-crystals are then irradiated by a laser beam, which results in desorption and mostly single-charged ionization of the analytes. In MALDI protein analyses, the resulting ions are usually separated according to their time of flight (TOF) through a vacuum drift tube, which corresponds with their m/z value. Molecules of lower molecular weight, such as lipids, drugs, and metabolites, are analyzed in reflectron mode on a tandem (TOF/TOF) instrument. This mode allows for greater resolution and mass accuracy, as well as for fragmentation of an ion of interest, which may provide structural information for small molecules or lipids, or may lead to protein identities through database lookups of their fragmentation fingerprints. There exist many other ion source–mass analyzer combinations that differ in their ionization products, their mass and spatial resolution, and sensitivity. For a detailed technological overview please refer to the review of Pol et al.3Pol J. Strohalm M. Havlicek V. et al.Molecular mass spectrometry imaging in biomedical and life science research.Histochem Cell Biol. 2010; 134: 423-443Crossref PubMed Scopus (67) Google Scholar MALDI imaging extends the MALDI approach to intact tissue sections. The principle and workflow is described in detail throughout this article and schematically in Figure 2. A tissue sample is cut in 10- to 15-μm slices and placed onto a conductive glass side. Importantly, the use of a conductive glass slide enables optical microscopic image acquisition and MALDI imaging mass spectrometry to be performed on the very same tissue section and therefore allows the direct correlation of the tissue image and mass spectrometric data. The tissue section may be either from fresh frozen or archived material like formalin-fixed, paraffin-embedded tissues. However, the analysis of formalin-fixed, paraffin-embedded tissues by MALDI imaging requires preceding antigen retrieval and tryptic digestion steps. Next, the section has to be coated with the matrix. This is a crucial step because it determines the maximum spatial resolution, sensitivity, and reproducibility of a MALDI imaging experiment. The matrix can be applied through spotting, sublimation, or spraying. Although spotting offers the highest sensitivity, spraying is used for image acquisitions at high spatial resolutions (up to 5 μm versus 150 μm for spotting). Sublimation results in the highest crystal density and thus offers the highest spatial resolution. Because a solvent is missing, it is better suited for the measurements of lipids than for proteins. Although any approach may be done manually, a robotic application of the matrix is recommended to ensure reproducibility and comparability between measurements. A histology-directed deposition of matrix droplets to extract molecular data from specific morphologic features—not for creating images—is called MALDI profiling. The coated glass slide is introduced into the mass spectrometer and mass spectra are acquired by laser irradiation of the matrix across the entire tissue section in an ordered raster process. Each individual measurement spot has an associated mass spectrum that will later constitute a pixel in the resulting MALDI image. The pixel size is technically limited by both the laser focus diameter and the average matrix droplet size. The spatial resolution typically ranges from 5 to 200 μm. Another limitation of MALDI imaging is that proteins >30 kDa are rarely detected by standard sample preparation. Because the tissue section remains intact during the MALDI measurement process, it can be counterstained afterward, as in conventional hematoxylin and eosin staining, and correlated with the mass spectrometric data. MALDI images are created by displaying the detected intensities of single mass signals as a function of location. This allows study of the distribution of molecules in tissues together within the underlying histology. Conversely, spectra derived from histology-defined regions of interest (eg, tumor lesions) can be used for the discovery of molecular signatures or single m/z signals that are linked to specific disease phenotypes by using various statistical methods. A comprehensive description of the MALDI imaging workflow, including sample preparation, measurement, and data interpretation can be found in the review from Gustafsson et al.4Gustafsson J.O. Oehler M.K. Ruszkiewicz A. et al.MALDI imaging mass spectrometry (MALDI-IMS)-application of spatial proteomics for ovarian cancer classification and diagnosis.Int J Mol Sci. 2011; 12: 773-794Crossref PubMed Scopus (80) Google Scholar Knowledge of the identity behind mass signals of interest is not always given. It depends on the molecule type, weight, and experimental conditions. Although peptides and lipids can usually be identified on-tissue through fragmentation, proteins need individually adapted, external identification experiments. These usually start with the isolation of the protein of interest from the tissue through fractionation using gel electrophoresis or liquid chromatography. Depending on its weight, the protein is then identified with (bottom-up) or without (top-down) previous digestion through fragmentation in the mass spectrometer. Validation is an important step to confirm the true importance of the identified molecules. One common way is to perform targeted in situ experiments, like immunohistochemistry for proteins, on larger, independent sample collections. MALDI imaging fills the gap of existing targeted and untargeted protein analysis methods. As a mass spectrometric approach, it offers the untargeted measurement of different molecule classes and events (like protein modifications) that are hardly accessible by other targeted in situ methods such as immunohistochemistry. One drawback is that it lacks the same high optical resolution. Compared with other state-of-the-art discovery techniques in proteomics (like fractionation coupled to MS), MALDI imaging suffers from sensitivity. However, by conserving the natural context of the sample, substantial specificity is added to the obtained results. Promising results of studies in diseases of the human lower and upper gastrointestinal system are presented (Supplemental Table 1). In a Barrett's adenocarcinoma study, MALDI imaging was used to elucidate proteomic changes in the progression from Barrett's esophagus to Barrett's adenocarcinoma and its potential to metastasize.5Elsner M. Rauser S. Maier S. et al.MALDI imaging mass spectrometry reveals COX7A2, TAGLN2 and S100-A10 as novel prognostic markers in barrett's adenocarcinoma.J Proteomics. 2012 Feb 17; (Epub ahead of Print)Crossref PubMed Scopus (86) Google Scholar The specific analysis of histologically defined precursor and invasive carcinoma lesions in human tissues revealed proteomic signatures for both tumor development and metastasis. Three of the identified proteins—COX7A2, TAGLN2, and S100-A10—could be successfully validated on an independent patient cohort (n = 102), and were also found to be novel prognostic indicators in Barrett's adenocarcinoma. That such information can be obtained by MALDI imaging, even from the smallest amounts of tissue samples, like endoscopic biopsies, has been shown in a study of gastric cancer. Here, the potential use of MALDI imaging as a diagnostic tool to identify early-stage tumors has been investigated.6Kim H.K. Reyzer M.L. Choi I.J. et al.Gastric cancer-specific protein profile identified using endoscopic biopsy samples via MALDI mass spectrometry.J Proteome Res. 2010; 9: 4123-4130Crossref PubMed Scopus (71) Google Scholar By histology-directed profiling of 63 gastric cancer and 43 healthy endoscopic biopsies, the authors identified profiles for separating tumor from healthy tissue, and for distinguishing pathologic stage Ia (pT1N0M0) from more advanced stages. This could be clinically important, because stage Ia lesions are potential candidates for endoscopic treatment. For patients with more advanced-stage disease, clinically relevant information is related to improving risk stratification. This has been addressed by our group, where 63 intestinal-type primary resected gastric cancer tissues were submitted to MALDI imaging analysis.7Balluff B. Rauser S. Meding S. et al.MALDI imaging identifies prognostic seven-protein signature of novel tissue markers in intestinal-type gastric cancer.Am J Pathol. 2011; 179: 2720-2729Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar A tumor-specific, 7-protein signature was found to be associated with an unfavorable overall survival, independent of major clinical covariates (Supplementary Figure 1). One of the proteins was identified via extraction, liquid fractionation and top-down proteomics analysis as CRIP1. CRIP1, previously unknown in gastric cancer, was confirmed to be an independent prognostic factor by immunohistochemistry on an independent patient cohort (n = 118). In terms of hepatologic questions, imaging mass spectrometry approaches have been applied to study autoimmune diseases and hepatocellular carcinomas (HCC). Le Faouder et al8Le Faouder J. Laouirem S. Chapelle M. et al.Imaging mass spectrometry provides fingerprints for distinguishing hepatocellular carcinoma from cirrhosis.J Proteome Res. 2011; 10: 3755-3765Crossref PubMed Scopus (43) Google Scholar showed in a recent publication that MALDI imaging could be a useful tool for distinguishing well-differentiated human HCC from cirrhotic macronodules, which is of high relevance in clinical practice. By comparing the in situ proteomes of HCC and cirrhotic peritumoral tissue (n = 30), they found a set of 13 proteins that could accurately define cancer from adjacent cirrhotic tissue. In tissue sections of a validation cohort, this set of proteins was able to classify both HCC and peritumoral cirrhotic areas in concordance with the underlying histology. Autoimmune liver diseases are each characterized by the destruction of a specific liver cell type. MALDI imaging was used for the histology-directed analysis of involved cell types in 37 liver biopsies to investigate proteomic differences between major autoimmune liver diseases: Primary biliary cirrhosis, primary sclerosing cholangitis, and autoimmune hepatitis.9Bowlus C.L. Seeley E.H. Roder J. et al.In situ mass spectrometry of autoimmune liver diseases.Cell Mol Immunol. 2011; 8: 237-242Crossref PubMed Scopus (11) Google Scholar Ten proteins were found differentially expressed between autoimmune hepatitis and primary sclerosing cholangitis, among them 2 belonging to neutrophil defensins. This is concordant with the primary sclerosing cholangitis diagnostic test by perinuclear staining of antineutrophil cytoplasmatic antibodies. Thus, the remaining proteins may represent also novel autoantigens or effector molecules. In inflammatory bowel diseases, the distinction between ulcerative colitis and Crohn's colitis can be challenging, but is of utmost importance for therapy decision making. M'Koma et al10M'Koma A.E. Seeley E.H. Washington M.K. et al.Proteomic profiling of mucosal and submucosal colonic tissues yields protein signatures that differentiate the inflammatory colitides.Inflamm Bowel Dis. 2011; 17: 875-883Crossref PubMed Scopus (48) Google Scholar used MALDI imaging to characterize the protein microenvironments of colonic mucosa and submucosa to determine if molecular differences between Crohn's colitis (n = 24) and ulcerative colitis (n = 27) exist. These authors found significant discriminatory proteins in both inflamed and uninflamed colonic submucosa, resulting in the ability to distinguish between the 2 inflammatory entities with 80% accuracy. In colon cancer, Meding et al11Meding S. Nitsche U. Balluff B. et al.Tumor classification of six common cancer types based on proteomic profiling by MALDI imaging.J Proteome Res. 2012; 11: 1996-2003Crossref PubMed Scopus (107) Google Scholar addressed 2 important clinical challenges: the classification of cancers of unknown primary and the identification of markers for regional lymph node metastasis. In their first study, protein profiles of 6 different primary adenocarcinoma types (colon, stomach, breast, esophagus, liver, and thyroid cancers) were used for the determination of the origin of distant metastasis samples, here, colon cancer liver metastases.11Meding S. Nitsche U. Balluff B. et al.Tumor classification of six common cancer types based on proteomic profiling by MALDI imaging.J Proteome Res. 2012; 11: 1996-2003Crossref PubMed Scopus (107) Google Scholar The results indicate that even closely related entities, such as the primary tumor of colon cancer, its liver metastasis, and hepatocellular cancer, could be distinguished accurately (accuracies >80%). In their second study, colon tumors with (n = 33) and without lymph node metastasis (n = 21) were compared by MALDI imaging analyses.12Meding S. Balluff B. Elsner M. et al.Tissue based proteomics reveals FXYD3, S100A11 and GSTM3 as Novel markers for regional lymph node metastasis in colon cancer.J Pathol. 2012; https://doi.org/10.1002/path.4021Crossref PubMed Scopus (92) Google Scholar Two candidate proteins, FXYD3 and S100-A11, were validated by immunohistochemistry and correlated significantly with the regional lymph node metastasis in the independent patient cohort (n = 168). These studies provide evidence that MALDI imaging, a novel tissue-based approach, may provide important information for clinically relevant questions. However, it is also gaining special interest for the imaging of small molecules, such as drugs, in biological tissues. Imaging modalities are important for characterizing pharmaceutical entities during drug discovery and development in targeted organs. Several studies have demonstrated the usefulness of MALDI imaging for assessing pharmacokinetic properties and the distribution of drugs and their metabolites in tissues or even whole body sections.1Schwamborn K. Caprioli R.M. Molecular imaging by mass spectrometry–looking beyond classical histology.Nat Rev Cancer. 2010; 10: 639-646Crossref PubMed Scopus (274) Google Scholar Figure 3 gives an example of the spatial distribution of lapatinib metabolites measured by MALDI imaging in a dog liver tissue section. Castellino et al13Castellino S. Groseclose M.R. Wagner D. MALDI imaging mass spectrometry: bridging biology and chemistry in drug development.Bioanalysis. 2011; 3: 2427-2441Crossref PubMed Scopus (180) Google Scholar conducted this experiment with high spatial resolution (50 μm) to clearly localize metabolites of lapatinib and endogenous molecules in small liver subcompartments (Figure 3A). In addition, they used a mass spectrometer with high mass resolution to resolve drug-related signals with very proximate m/z signals (Figure 3B). Detecting the distribution of a drug in its target site may be crucial for determining its efficacy; recent evidence points to the role of poor drug delivery as a key resistance mechanism in widely therapy-resistant tumors, such as in pancreatic adenocarcinoma.14Olive K.P. Jacobetz M.A. Davidson C.J. et al.Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer.Science. 2009; 324: 1457-1461Crossref PubMed Scopus (2409) Google Scholar This may therefore not only lead to optimization of the delivery and efficacy in general, but may also identify individual patients with low drug levels in the target tissue compartments, who may show a poorer response to treatment and may be candidates for alternative treatment protocols. Furthermore, the assessment of the proteome by MALDI imaging in combination with drug distributions (“pharmacoproteomics”), clinical endpoints, and more traditional assays, such as for apoptosis or cell proliferation, should allow for a more complete understanding of the biochemical effects of therapeutics. In gastroenterology, MALDI imaging might constitute a novel tool for tissue-based research, for example, for the analysis of animal models, drug candidates, or endoscopic biopsies from patients. Some of the already conducted studies have been presented in this article. Although these are still low in sample numbers, they already point out that novel and clinical useful information can be obtained by MALDI imaging. High-throughput techniques are prone to producing false results, which may be the result of experimental design, effects during sample preparation and measurement, or statistical overfitting. Although MALDI imaging adds substantial specificity to its results by incorporating the histologic context, a validation of the results on independent sample sets is essential to indicate their suitability for further studies and potential clinical application. However, a bottleneck for validation is the protein identification in tissue, which still has to be overcome. Another important step toward interlaboratory comparability and therefore clinical applicability is the quantification of the spectral data of MALDI imaging. This has been recognized and is currently under investigation by the MALDI imaging community. In general, MALDI imaging mass spectrometry is now receiving increasing interest, which is driving the further development of this technology to overcome some of its limitations, for example, in terms of spatial and mass resolution, sensitivity, or protein identification.15Balluff B. Schone C. Hofler H. et al.MALDI imaging mass spectrometry for direct tissue analysis: technological advancements and recent applications.Histochem Cell Biol. 2011; 136: 227-244Crossref PubMed Scopus (102) Google Scholar Nevertheless, MALDI analyses are quick, inexpensive, and even applicable on the smallest amounts of tissue. Practically, one can envision the translation of MALDI imaging into clinical practice. Especially in gastroenterology, which is significantly based on bioptic diagnostics, MALDI imaging might complement histopathologic evaluation with additional molecular information, to aid in diagnostics, risk assessment, and response prediction to therapy, thereby facilitating a step toward more personalized medicine. However, future preclinical studies, prospective clinical trials, and technological developments will have to prove these expectations. Supplemental Table 1Proteomics Studies in Gastrointestinal Diseases Using MALDI ImagingDiseasePublicationProteinMass [Da]Clinical purposeValidationBarrett's cancerElsner et al5Elsner M. Rauser S. Maier S. et al.MALDI imaging mass spectrometry reveals COX7A2, TAGLN2 and S100-A10 as novel prognostic markers in barrett's adenocarcinoma.J Proteomics. 2012 Feb 17; (Epub ahead of Print)Crossref PubMed Scopus (86) Google Scholar61-protein signature—Carcinogenesis—COX7A26720Carcinogenesis/prognosisIHCS100-A1011185Carcinogenesis/prognosisIHC28-protein signature—Marker for regional lymph node metastasis—TAGLN222262Marker for regional lymph node metastasis/prognosisIHCStomach cancerKim et al6Kim H.K. Reyzer M.L. Choi I.J. et al.Gastric cancer-specific protein profile identified using endoscopic biopsy samples via MALDI mass spectrometry.J Proteome Res. 2010; 9: 4123-4130Crossref PubMed Scopus (71) Google Scholar73-protein signature—Tumor detection—DEFA13439Tumor detection—DEFA23368Tumor detection—S100-A810840Tumor detection—S100-A913158/12694Tumor detection—17-protein signature—Early vs advanced stage—Balluff et al7Balluff B. Rauser S. Meding S. et al.MALDI imaging identifies prognostic seven-protein signature of novel tissue markers in intestinal-type gastric cancer.Am J Pathol. 2011; 179: 2720-2729Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar7-protein signature—Prognosis in intestinal type gastric cancer—DEFA13445Prognosis in intestinal type gastric cancerIHCCRIP18406Prognosis in intestinal type gastric cancerIHCS100-A610098Prognosis in intestinal type gastric cancerIHCLiver cancerLe Faouder et al8Le Faouder J. Laouirem S. Chapelle M. et al.Imaging mass spectrometry provides fingerprints for distinguishing hepatocellular carcinoma from cirrhosis.J Proteome Res. 2011; 10: 3755-3765Crossref PubMed Scopus (43) Google Scholar13-protein signature—Tumor marker—Ubiquitin8565Tumor markerIHC, PCRLiver autoimmune diseasesBowlus et al9Bowlus C.L. Seeley E.H. Roder J. et al.In situ mass spectrometry of autoimmune liver diseases.Cell Mol Immunol. 2011; 8: 237-242Crossref PubMed Scopus (11) Google Scholar10-protein signature—Distinction of autoimmune hepatitis and primary sclerosing cholangitis—Pancreatic cancerDjidja et al16Djidja M.C. Claude E. Snel M.F. et al.MALDI-ion mobility separation-mass spectrometry imaging of glucose-regulated protein 78 kDa (Grp78) in human formalin-fixed, paraffin-embedded pancreatic adenocarcinoma tissue sections.J Proteome Res. 2009; 8: 4876-4884Crossref PubMed Scopus (106) Google ScholarGrp7872288Tumor markerIHCColon colitidesM'Koma et al10M'Koma A.E. Seeley E.H. Washington M.K. et al.Proteomic profiling of mucosal and submucosal colonic tissues yields protein signatures that differentiate the inflammatory colitides.Inflamm Bowel Dis. 2011; 17: 875-883Crossref PubMed Scopus (48) Google Scholar5-protein signature—Distinction of ulcerative colitis and Crohn's colitis—Colon cancerMeding et al11Meding S. Nitsche U. Balluff B. et al.Tumor classification of six common cancer types based on proteomic profiling by MALDI imaging.J Proteome Res. 2012; 11: 1996-2003Crossref PubMed Scopus (107) Google Scholar50–118 protein signature—Classification of cancer of unknown primary (liver metastasis from primary colon cancer)—Meding et al12Meding S. Balluff B. Elsner M. et al.Tissue based proteomics reveals FXYD3, S100A11 and GSTM3 as Novel markers for regional lymph node metastasis in colon cancer.J Pathol. 2012; https://doi.org/10.1002/path.4021Crossref PubMed Scopus (92) Google ScholarFXYD39264Marker for regional lymph node metastasisIHCS100-A1111646Marker for regional lymph node metastasisIHCDa, Dalton; IHC, immunohistochemistry; MALDI, matrix-assisted laser desorption/ionization; PCR, polymerase chain reaction. Open table in a new tab Da, Dalton; IHC, immunohistochemistry; MALDI, matrix-assisted laser desorption/ionization; PCR, polymerase chain reaction.