596202

review-article2015

JHCXXX10.1369/0022155415596202Yalcin and de la MonteMALDI-IMS Tissue Lipids

Review Journal of Histochemistry & Cytochemistry 2015, Vol. 63(10) 762­–771 © The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1369/0022155415596202 jhc.sagepub.com

Review of Matrix-Assisted Laser Desorption Ionization-Imaging Mass Spectrometry for Lipid Biochemical Histopathology Emine B. Yalcin and Suzanne M. de la Monte

Liver Research Center, Division of Gastroenterology and Department of Medicine (EBY, SMD); Departments of Neurology, Neurosurgery, and Pathology (SMD); and Rhode Island Hospital and the Alpert Medical School of Brown University, Providence, RI (EBY, SMD)

Summary Matrix-Assisted Laser Desorption Ionization-Imaging Mass Spectrometry (MALDI-IMS) is a rapidly evolving method used for the in situ visualization and localization of molecules such as drugs, lipids, peptides, and proteins in tissue sections. Therefore, molecules such as lipids, for which antibodies and other convenient detection reagents do not exist, can be detected, quantified, and correlated with histopathology and disease mechanisms. Furthermore, MALDI-IMS has the potential to enhance our understanding of disease pathogenesis through the use of “biochemical histopathology”. Herein, we review the underlying concepts, basic methods, and practical applications of MALDI-IMS, including post-processing steps such as data analysis and identification of molecules. The potential utility of MALDI-IMS as a companion diagnostic aid for lipid-related pathological states is discussed. (J Histochem Cytochem 63:762–771, 2015) Keywords Matrix-assisted laser desorption ionization, histology, imaging mass spectrometry, lipids

Overview Matrix-assisted laser desorption ionization-imaging mass spectrometry (MALDI-IMS) is a high-throughput analytical tool for investigating the distributions of proteins, peptides, lipids, and small molecules in situ, particularly, in tissue sections (Aichler and Walch 2015; Cornett et al. 2007; Eriksson et al. 2013; Franck et al. 2009; Walch et al. 2008). Mass spectrometry identifies molecules by measuring the mass to charge ratios (m/z) of their ionized forms. Mass spectrometry entails three major steps: 1) conversion of molecules into ions by the ionizing source, e.g. laser; 2) separation of ions according to their m/z ratios in a mass analyzer; and 3) detection of separated ions according to their relative abundance using an ion detector (Di Girolamo et al. 2013). IMS detects ion species generated from the surfaces of solid samples, and thus provides information on the in situ localization and identity of endogenous (lipid, protein, peptide) (Caprioli et al. 1997; Groseclose et al. 2007; Murphy et al. 2011; Taban et al. 2007) and exogenous

(xenobiotic, metabolite) (Cornett et al. 2008; KhatibShahidi et al. 2006) molecules. Therefore, MALDI-IMS could be regarded as a ‘biochemical histology’ tool.

The Need for MALDI-IMS as a Biochemical Histology Diagnostic Aid Despite major advances in molecular biology, cell biology, and chemistry, the vast majority of tissue diagnoses are made using routine histopathology, in which 3–10-µm-thick, formalin-fixed, paraffin-embedded (FFPE) tissue sections are stained with hematoxylin and eosin (H&E) and examined by a pathologist using light microscopy. This technology has not Received for publication March 24, 2015; accepted June 18, 2015. Corresponding Author: Suzanne M. de la Monte, Pierre Galletti Research Building, Rhode Island Hospital, 55 Claverick Street, Room 419, Providence, RI 02903, USA. E-mail: [email protected]

763

MALDI-IMS Tissue Lipids been replaced because it is well established, relatively rapid, reproducible across institutions and generations, inexpensive, and predictive of disease stage, severity, and prognosis. However, advances in therapeutic strategies are forcing us to gain a better understanding of disease mechanisms and appreciate the basis for heterogeneity, as responses to treatment can vary with disease subtypes. In many respects, these problems are being addressed through the use of immunophenotyping by immunohistochemical staining or fluorescence-activated cell sorting (FACS), gene expression analysis by polymerase chain reaction (PCR) or in situ hybridization studies, and gene mutation or polymorphism detection by next-generation sequencing and fluorescence in situ hybridization (FISH). Unfortunately, many diseases, particularly those that are metabolic or degenerative rather than neoplastic in nature, do not readily render themselves more diagnosable by any of these approaches. One of the main reasons is that their underlying bases are often anchored in chronically altered states that lead to pathogenic shifts in intracellular signaling, organelle dysfunction, stress, and metabolic derangements. Furthermore, different pathogenic agents can produce similar or even indistinguishable histopathological abnormalities, limiting the utility of routine histology as a stand-alone diagnostic tool. For example, alcoholic liver disease, non-alcoholic fatty liver disease, and hepatitis C virus infection can all cause steatohepatitis that progresses to cirrhosis and liver failure (Neuman et al. 2014). Chronic alcohol abuse, heavy tobacco smoking and Alzheimer’s disease are all associated with atrophy of cerebral white matter (de la Monte 1988; Durazzo et al. 2014; Wang et al. 2009). Inflammatory diseases of the gastrointestinal tract have a wide range of etiologies (Escobedo et al. 2014; Hajela et al. 2015). Therefore, additional approaches are needed to more effectively diagnose and guide therapy on the basis of disease pathogenesis. Since mass spectrometry enables the detection and quantification of abnormally expressed or modified (e.g. phosphorylated, glycosylated, or nitrosated) lipids, proteins and nucleic acids, without the need for specific antibodies, this approach holds tremendous promise as a complementary diagnostic aid, particularly for non-neoplastic disease. Among all of the biochemical and molecular abnormalities that occur with disease, lipids are the most challenging to study because of the lack of suitable reagents, including antibodies, for analysis; yet, their roles in membrane structure and function, metabolism, and intracellular signaling are vast. Lipids are the principle component of the myelin sheath that insulates nerve fibers in the central and peripheral nervous systems. Myelin degenerates and fails to recover in inflammatory diseases, such as multiple sclerosis, and in metabolic/degenerative diseases due to alcohol abuse or Alzheimer’s disease; yet, we have no way to adequately study these problems in humans. Lipids accumulate in hepatocytes following a wide range of insults; but only in a

subset of individuals does the lipid accumulation (steatosis) lead to progressive disease that culminates in liver failure. Currently, we do not understand how hepatocyte lipid profiles change in relation to proneness for disease progression. These problems could be addressed using mass spectrometry and, specifically, MALDI-IMS because the abnormalities could be linked to specific cellular and structural pathological changes. Thus far, a number of ionization methods, mass analyzers, and ion detectors have been developed for IMS. Below, we discuss the most commonly used instruments and analytical approaches for lipid studies.

How MALDI Works MALDI is a soft ionization technique in which sample molecules or analytes are co-crystallized with the matrix to form sample–matrix crystals. Pulsed UV laser targeting of the crystals results in vaporization of matrix that carries the analyte with it. The matrix acts as a proton donor or acceptor, ionizing the analyte (Di Girolamo et al. 2013; Lewis et al. 2000). Proton transfer between the matrix and the analyte renders neutral analyte molecules [M] protonated ([M+H]+) or deprotonated ([M-H]-), and resulting ions are detectable in positive or negative ion modes, respectively. Analytes can also form adducts with salts ([M+Na]+) via cation transfer. Laser irradiation is influenced by wavelength, pulse duration, and fluence (Dreisewerd 2003; Knochenmuss and Zenobi 2003). To achieve the best mass spectrometric performance, MALDI typically employs a nitrogen UV laser (337 nm) or frequency tripled (355 nm) or quadrupled (266 nm) Q-switched neodymium-doped:yttrium aluminum garnet (Nd:YAG) lasers. Pulse durations are usually short, between 0.5 and 10 ns, as longer pulses may cause destructive thermal excitation of analytes prior to their release into the gas phase via desorption/sublimation. Laser fluence is the amount of laser energy pulsed to a focused area (spot). Standard fluence values used for MALDI are between 30 and 600 mJ/cm-2. If the laser fluence is too low, no ions are produced, and if it is too far above the threshold, the generation of ion currents may exceed the capacity of the mass analyzer and detector. In most cases, the parameters for laser fluence needed for sample analysis can be adjusted prior to image execution based on published literature. MALDI, including with IMS, is usually coupled with time-of-flight (TOF) mass analysis and detection by ion-toelectron conversion. In TOF, ions accelerated by an electrostatic field are allowed to drift in a neutral field. Due to acceleration, ions with equal charge have the same kinetic energy, but the differences in ion velocities distinguish ions based on mass. Smaller particles move faster through the flight tube and reach the detector in shorter periods of time than heavy particles. The time required to reach the detector determines the ion’s m/z ratio (Wiley and McLaren 1955). A mass spectrum is generated at each position irradiated by

764

Yalcin and de la Monte Yalcin and de la Monte

Figure 1.  MALDI imaging workflow diagram. Fresh-frozen tissue samples are cryosectioned at a 10-µm thickness and thaw-mounted on conductive MALDI slides. Matrix is applied onto the tissue section by sublimation and a single mass spectrum is acquired from every pixel in the image. For molecular identification, selected precursor ion is fragmented by laser energy generating product ions collected in MS/MS spectrum, which are searched against the LIPID MAPS database. Additional data analysis steps include visualization of the distribution of single molecules across the tissue section, Principal Component Analysis (PCA), and hierarchical clustering.

the laser. The tissue distribution and relative intensity of individual ions can be visualized using a color scale corresponding to the relative intensity of each ion at every pixel. (Fig. 1). Most important, the positions of individual or clusters of molecules visualized on the image can be co-localized with histological structures and histopathology. In a typical mass spectrum, the x-axis represents m/z values, whereas the y-axis reflects total ion counts. In IMS, the m/z ion values are rasterized across tissue sections such that molecular information is collected from regularly spaced pixel arrays (Seeley and Caprioli 2008). The distance between pixels governs spatial resolution of the image. A highly resolved image can be acquired by adding grid points within the area but the additional information can decrease overall signal intensity and the number of peaks detected while increasing acquisition time and data file size. Protocol optimization considers these variables.

Overview of Sample Preparation Sample preparation and storage are critical for ensuring that the integrity and spatial distribution of molecules in tissues are preserved, and that delocalization and degradation of

analytes are minimized or avoided. Earlier protocols focused on imaging proteins and peptides, whereas later applications have attended to the detection of lipids. In deciding upon an experimental strategy, the sample source and nature of the analytes to be detected should be major considerations, and protocols should be adapted accordingly. Typically, freshfrozen tissues are cryosectioned and mounted onto glass slides coated with a conductive MALDI target. The sections must then be washed with an aqueous buffer, and then matrix applied to the tissue section. The samples can then be imaged and analyzed for presence, abundance and distributions of molecules in the tissue. Since the literature is replete with methods for analyzing proteins and peptides (Groseclose et al. 2007; Meding and Walch 2013), and our major research areas concern lipid abnormalities in steatohepatitis and myelin degeneration of different etiologies, in this review, we provide general guidelines as well as specific recommendations for imaging lipids using MALDI-IMS.

Sample Collection As is the case for all biological tissue-based studies, rapid and efficient collection and storage of samples are critical.

765

MALDI-IMS Tissue Lipids Delays, such as those associated with postmortem human tissue harvesting, particularly from subjects with prolonged antemortem agonal periods, can cause artefacts and reduce sensitivity and specificity of results. Therefore, under all circumstances, validation and reproducibility studies, including the use of appropriate controls, are needed. Rapid freezing of tissue helps prevent molecular degradation due to endogenous enzymatic activity, oxidation, and posttranslational modifications. It is best to snap-freeze small pieces (generally up to 1 cm in diameter and 3–4 mm in thickness) of well-oriented (gently flattened and free of folds) tissue samples by submerging them in crushed (pulverized) dry ice or liquid nitrogen. Dry ice freezing can be preferable to liquid nitrogen due to reduced cracking. Frozen tissue should be wrapped, e.g., in aluminum foil, or inserted into a pre-labeled plastic, disposable histology cassette prior to storage in an air-tight container at -80°C. The use of standard histology cassettes enables convenient, clear labeling, rapid sample retrieval, and specimen mounting onto cryostat cutting blocks. Besides fresh-frozen tissue, which is the “gold standard”, MALDI-IMS can be performed on tissues stabilized by formalin fixation, heat, or microwave irradiation (Carter et al. 2011; Goodwin et al. 2012; Sugiura et al. 2014). Furthermore, although investigators have also established MALDI-IMS protocols for analyzing formalin-fixed, paraffin-embedded (FFPE) tissues (Groseclose et al. 2008), those approaches are not suitable for lipid studies because tissue lipids are extracted during the processing steps required for dehydration before paraffin embedding and re-hydration before staining and imaging. On the other hand, as the technology advances, we might anticipate that the barriers to accessing the large repertoire of archived FFPE human and experimental animal model tissues that span decades of collection from countries around the world, will be resolved, enabling us to better understand disease pathogenesis.

Sectioning The ultra-sensitivity of MALDI-IMS demands avoidance of contaminants. Therefore, cryostats used for sectioning must be thoroughly cleaned before and after use. In some instances, it may be appropriate to have a cryostat dedicated for MALDI-IMS, particularly when there are multiple heavy users of the instrument. Embedding material, particularly Optimal Cutting Temperature (OCT) compound, must not contaminate the tissue sections mounted on slides because OCT can suppress ionization and analyte detection. In general, just a very small amount of OCT should be applied to the center of the frozen tissue block for attachment to the chuck. The OCT compound/medium must not extend beyond the tissue perimeter. Alternatively, glue can be used to affix tissue to the chuck. Sectioning very tiny fragile specimens can be challenging. However, the process

can be facilitated by embedding the tissue in 2% or 5% carboxy methylcellulose, 1% agarose, or 5% to 10% gelatin, and attaching the specimen to the chuck using either OCT or glue. Optimum section thickness is between 10 and 20 µm, made at cabin temperatures between -10°C to -20°C. Flat tissue sections free of folds and cracks are thawmounted onto room temperature glass slides coated with indium-tin-oxide. Serial sections should be mounted on separate glass slides for repeated imaging or histological staining. However, different samples, e.g. diseased versus control, can be mounted onto the same slide to control for subsequent specimen handling and imaging. After sectioning, the slides can be stored in -80°C for approximately 1 year. Prior to sample processing, slides from the freezer should be brought up to room temperature in a desiccator under vacuum.

Aqueous Wash The tissue sections must be washed in an aqueous buffer to remove native salts and minerals. This straight-forward step vastly improves ion signal intensities while retaining their spatial resolution. A number of buffers and washing protocols have been tested and optimized (Angel et al. 2012). For example, Angel et al. tested ammonium bicarbonate (pH 8.4), ammonium acetate (pH 6.7), ammonium formate (pH 6.4), and formic acid (pH 2.5) at 50 mM concentrations for 5, 10, 15, and 30 sec. A 15-sec ammonium formate wash was found to be optimum for maximizing signal intensity (Angel et al. 2012). For our research, we further optimized the protocol for studying liver and brain by washing the fresh-frozen tissue sections (4 times, 5 sec each) in 50 mM ammonium formate (pH 6.4). These washes were performed by gently pipetting 100–200 µl of buffer directly onto the tissue, waiting 5 sec, pouring off the liquid, then repeating the cycle. This approach is preferable to submerging the slides in buffer because it reduces the risk of losing specimens due to dislodgment from the slides. As an additional precaution, tissue section adherence can be increased by pre-coating the slides with poly-l-lysine (Zavalin et al. 2012). Ammonium formate buffer washes increased lipid ion signal intensities by at least three-fold when imaged in the negative ion mode (Fig. 2). After the buffer washes, the slides were dried under vacuum for 30 min to enhance tissue adhesion.

Matrix Application Matrix is then applied to fresh-frozen ammonium formatewashed, vacuum-dried tissue sections. Matrix is composed of small, acidic, and aromatic compounds that can absorb laser energy at the same wavelength as the irradiating laser. Sinapinic acid is commonly used to analyze high molecular weight proteins, whereas α-cyano-4-hydroxycinnamic acid

766

Yalcin and de la Monte Yalcin and de la Monte

Figure 2.  Effect of aqueous washing on lipid ion intensity in histological sections of (A) human livers and (B) rat brains. Two adjacent tissue sections were thaw-mounted on the same MALDI slide. One section was not washed, and the other was washed by pipetting ~100–200 µl of 50 mM ammonium formate buffer four times for 5 sec. The sections were dried in a desiccator under vacuum for ~30 min before matrix application to increase the adhesion of tissue section to the slide during MS acquisition. The distribution of lipid ions in tissue sections and the average MS spectra detected in washed versus un-washed sections show that aqueous washing increases lipid ion intensities by at least 3-fold when imaged in the negative ion mode.

is preferred for analyzing low molecular weight peptides. DHB (2,5-dehydroxybenzoic acid) is commonly used to image lipids because of the resulting high spatial resolution

and sensitivity. The quality of matrix application can significantly impact mass and spatial resolution, sensitivity, and reproducibility of the MALDI-IMS results (Murphy

MALDI-IMS Tissue Lipids et al. 2011). Matrix application approaches can be manual or automated. Manual application techniques, including the use of airbrushes or thin-layer chromatography (TLC) sprayers (Wang et al. 2008), can be problematic due to difficulty controlling the thickness and distribution of the matrix, and achieving consistent results from slide to slide. Automated matrix sprayers are more suitable for achieving matrix applications that are reproducible and afford high resolution, excellent spectral quality imaging (Yamada et al. 2011). For lipid studies, although both manual and automated matrix application methods are widely used, an alternative approach is to employ sublimation. Sublimation is a dry method that does not require a matrix solvent, which can cause lipid analytes to become delocalized. Sublimation is also relatively inexpensive and can be completed within a short period of time (less than 10 min). Most important, sublimation produces even layers of small crystals across targets (tissue sections), substantially increasing the resolution of ions detected. Using a commercial sublimation apparatus, matrix deposition can be rendered highly reproducible by standardizing the time, temperature, and pressure settings (Hankin et al. 2007). The commercial sublimation apparatus (Chemglass Life Sciences, Vineland, NJ) consists of upper and lower flasks coupled to a rough pump and digital thermocouple vacuum gauge controller (Fig. 3A). A cold trap is placed between the sublimation flask and pump to avoid contaminating the pump with matrix crystals. DHB crystals (300 ± 5 mg) should be evenly dispersed on the bottom of lower flask. The MALDI target, i.e., the glass slide with mounted tissue sections, is attached with conductive copper tape to the bottom surface of the upper flask facing the matrix crystals (Fig. 3B). After positioning the target and matrix, the upper and lower flasks are sealed tightly, then the upper flask is filled with wet ice to chill the glass slide, and vacuum is applied using the pump. Once the pressure reaches 0.05 Torr, the sublimation chamber is placed onto the sand bath heated over a hot plate. With heat and vacuum applied, matrix molecules transition from the solid phase to gas phase, and subsequently condensate onto the surface of the cool slide. Within approximately 7 min, with the temperature at 120°C, the MALDI slide should be coated with 200 ± 13 mg/cm2 of DHB matrix as a thin uniform layer (Fig. 3C), which is needed to obtain high-quality images.

MALDI-IMS Execution MALDI targets must have an electrically conductive surface. To achieve this, the tissue mounting surfaces of the glass slides are coated with indium tin oxide (ITO), which is transparent and conductive. Pre-coated slides are commercially available. The matrix is uniformly applied to the entire conductive surface of the slide, covering the tissue and the bare ITO-coated glass. The slide edges must be wiped free of matrix to prevent electrical conductivity noise.

767

Figure 3.  Matrix Coating by Sublimation. (A) Photograph of sublimation system used to apply matrix on tissue sections for MALDI-IMS. The system is composed of a sublimation chamber coupled to a rough pump and digital thermocouple vacuum gauge controller. A cold trap is placed between the sublimation flask and pump to avoid contamination of the pump with matrix crystals. (B) An enlarged view of the bottom of the sublimation chamber. Matrix is placed on the bottom of the lower sublimation flask. The MALDI slide with mounted tissue sections is attached to the bottom surface of the upper sublimation flask facing the matrix crystals. With pressure and heat, matrix molecules undergo phase transition from solid to gas, subsequently undergoing condensation on the surface of the cool slide. (C) Pre- and post-sublimation images are shown. Sublimation produces a thin, uniform layer of small crystals across the tissue, which is needed to yield high-quality images. The right panel shows a higher magnification image of the fine, uniformly distributed crystals. Scale (left, middle) 6.25 mm; (right) 300 µm.

768 Two additional steps are then needed before the tissues can be imaged. The first is to spot a calibration standard onto the MALDI target for instrument calibration. The calibration standard consists of a mixture of compounds; e.g., peptides or lipids that have m/z values spanning the range of the analytes of interest. The standards are mixed with the matrix solution that is applied to the tissue sections; e.g., DHB in 50% acetonitrile, 0.1% TFA (1:1). The calibration standard (1 µl) is applied directly to the MALDI target and air-dried. The second pre-imaging step is to mark the target with three teaching points that will be used to align the slide and tissue image with the actual target position in the instrument. The matrix-coated conductive glass slide with mounted tissue sections, calibration spot, and teaching points is placed onto the MALDI slide adapter and inserted into mass spectrometer. The optical image generated at a minimum resolution of 1200 dpi is loaded into the imaging software to generate auto-execution sequences and set the teaching points for individual IMS experiments. The tissue sections are encircled to facilitate localizing their boundaries while being imaged. A raster size is determined based upon the resolution required to address specific biological questions. For high-resolution imaging, the laser ablation targets can be spaced at 10–50 µm intervals. For exploratory imaging, larger raster sizes (200–300 µm) would suffice to select more focused regions of interest for high-resolution imaging (Andersson et al. 2010). However, compromises with respect to resolution must be considered because, although high resolution images provide more spatial information, they can substantially increase acquisition time, datasets, and length of instrument use, and decrease sensitivity and the number of ions produced/detected (Norris and Caprioli 2013). Besides raster size, other factors affecting image detection or resolution include laser beam diameter and instrument polarity. Laser beam diameter is adjustable on just a limited number of instruments. The Bruker Smartbeam® has a 5-step adjustable focus that is sized as ultra-large, large, medium, small, and minimum. Estimating the importance of the effective beam diameter for specific data acquisition can be achieved by comparing the tissue ion spectra acquired from adjacent raster points using each of the beam diameters (Setou et al. 2010). With regard to polarity, images are usually acquired in both the positive and negative ion modes for untargeted analysis. Alternatively, the instrument can be set to detect ions in either the positive or negative ion mode if the ionization tendencies of the analytes of interest are known.

MALDI-IMS Data Acquisition and Analysis IMS data, consisting of m/z values of ions, signal intensity, and x- and y-coordinates, are used to visualize ion distributions and intensities throughout the tissue section.

Yalcin and de la Monte Yalcin and de la Monte Ion intensity within a specified area or region can also be represented as an averaged spectrum. Distributions of single or multiple m/z values can be visualized within specified regions of interest (ROIs), and signal intensities can be compared among different ROIs. IMS results should be validated by imaging identical samples (serial sections) on separate days, and under the same matrix application and MS acquisition conditions to assess intra-sample variability. In addition, IMS should be performed on numerous different subjects within each group to assess inter-sample variability (Fig. 4). Because of the large amount of complex data acquired by IMS, it is often necessary to apply algorithms that characterize sample groups. Principal component analysis (PCA) is one of the commonly used methods of analyzing complex MALDI-IMS data sets. PCA is a statistical differentiation method that extracts variances within data sets. PCA reduces the dimensionality of the mass spectra data to either a 2D or 3D coordinate system, where each spectrum is represented by a point or symbol. Spectra with similar characteristics form clusters, rendering inter-sample and inter-group differences readily visible on PCA plots (Cho et al. 2012). Hierarchical clustering is another statistical method that provides grouping of the spectra into different classes. Spectra are clustered by similarity in a dendrogram, and specific nodes can be visualized with different colors on the MALDI image. This enables classification of different tissues such as healthy versus diseased state (Deininger et al. 2008). A unique feature of IMS is its ability to correlate biochemical data with histopathology. This is best accomplished by pre- or post-staining of histological sections used for IMS, and staining of immediately adjacent histological sections made at the same time as those used for IMS. Staining prior to IMS provides unambiguous correlation between histology and IMS, but staining can compromise the quality of MS spectra (Walch et al. 2008). Although staining after IMS acquisition also enables clear correlation between spectra and histology, the tissue integrity can be lost due to the prolonged acquisition periods (laser irradiation) needed to generate high-resolution images. Either way, we have found it prudent to stain the immediately adjacent section because the histological details are crisp and the regional co-localization of analytes remains fairly accurate. We recommend staining the tissue sections after IMS analysis, as well as the subsequent serial section (stored at -80°C). With guidance from a pathologist, standardized ROIs that reflect representative structural features within the sections can be co-registered and digitally marked on the scanned images for further IMS analysis (Fig. 1). This approach generates more refined data from uniform areas across multiple sections/cases, and it reduces the data volume by focusing on subsets to be analyzed.

769

MALDI-IMS Tissue Lipids

Figure 4.  Method Reproducibility. Serial sections from human liver were thaw-mounted onto separate MALDI slides and imaged on different days. The nearly exact overlays of MS spectra from each experiment demonstrate that the method is highly reproducible.

Identification of Molecules Following MALDI-IMS, it is desirable to identify specific ions that distinguish different regions of the tissue or different experimental or human study groups. The identification of molecules can be accomplished by tandem mass spectrometry analysis (MS/MS). Using MS/MS, parent ions (analytes to be identified) are fragmented by laser energy, generating product ions that are collected in MS/MS spectra. The characteristic patterns of product ion fragmentation enable their identification via comparison with database libraries. For example, complete assignment of glycerophospholipids requires product ions that are indicative of the head group (such as ethanolamine, serine, choline or inositol) plus two fatty acyl chains. Fatty acyl chains are fragmented by the loss of carboxylic acid or corresponding ketene group. The parent ion and all of the fragments are used to search a lipid database, such as LIPID MAPS. The structure/identity of the lipid is assigned based upon fragment matches in the library. Despite the potentially broad application of “biochemical histopathology”, such as its diagnostic applicability in distinguishing different causes and disease stages of steatohepatitis, many MALDI-IMS routines lack the ability to distinguish isobaric and isomeric compounds because of omission of pre-separation steps used for liquid

chromatography coupled with MS/MS (LC-MS/MS). Isobaric and isomeric molecules have the same masses, but isobaric molecules have the same total number of protons and neutrons whereas isomeric compounds have identical numbers of protons and neutrons. Thus, whereas the chemical composition and m/z is the same between isobaric and isomeric proteins, the orientations of their atoms differ due to alterations in double bond positions, backbone substitution of fatty acyls, and stereochemistry. To circumvent this problem, MALDI-IMS can be coupled with high-resolution mass analyzers such as orbitrap (Landgraf et al. 2009), fourier transform ion cyclotron resonance (FT-ICR) (Vidova et al. 2010), or ion mobility mass spectrometry. For example, ion mobility mass spectrometry has been used to detect glycerophosphoethanolamines (PEs), glycerophosphoserines (PSs), glycerophosphoglycerols (PGs), glycerolphosphoinositols (PIs), glycerophosphates (PAs), sulfatides (STs), and gangliosides directly in mouse brain tissue sections, as well as in brain tissue extracts (Jackson et al. 2014).

Conclusions MALDI-IMS is a powerful method that enables investigators to visualize, localize and quantify molecules in situ in

770 tissue sections. This emerging field of biochemical histopathology should be embraced by pathologists, as it offers the opportunity to expand diagnostic assessments and better distinguish underlying causes of disease that share similar histopathological features. Integrating biochemical histopathology with the already established molecular and structural pathological approaches could help refine diagnostic criteria, disease stage, and analysis of complex patterns of tissue injury. Acknowledgments The authors acknowledge the valuable contributions of Dr. Dale Shannon Cornett, Bruker Daltonics, Inc. Billerica, MA for his ongoing support in developing the MALDI-IMS applications in our laboratory.

Competing Interests The authors declared no potential competing interests with respect to the research, authorship, and/or publication of this article.

Author Contributions EBY had a major role in establishing the MALDI-IMS methods for lipid analysis in the laboratory, prepared the initial draft of the manuscript, planned the manuscript, and took the lead role in generating all post-initial drafts of the manuscript, including its final version. SMD conceived of the ‘biochemical histopathology’ concept and the eventual utility of MALDI-IMS to enhance diagnostics of metabolic and degenerative diseases, and provided critical guidance to EBY. Both authors read and approved of the manuscript prior to its submission.

Funding The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: All funding for this research was provided by grants from the National Institutes of Health. Supported by AA-11431 and AA-12908 from the National Institutes of Health The research was supported by grants AA-11431 and AA-12908 from the National Institutes on Alcohol Abuse and Alcoholism at the National Institutes of Health.

References Aichler M, Walch A (2015). MALDI Imaging mass spectrometry: current frontiers and perspectives in pathology research and practice. Lab Invest 95:422-431. Andersson M, Andren P, Caprioli RM (2010). MALDI Imaging and Profiling Mass Spectrometry in Neuroproteomics Neuroproteomics. Boca Raton FL, Llc. Angel PM, Spraggins JM, Baldwin HS, Caprioli R (2012). Enhanced sensitivity for high spatial resolution lipid analysis by negative ion mode matrix assisted laser desorption ionization imaging mass spectrometry. Anal Chem 84:1557-1564. Caprioli RM, Farmer TB, Gile J (1997). Molecular imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Anal Chem 69:4751-4760.

Yalcin and de la Monte Yalcin and de la Monte Carter CL, McLeod CW, Bunch J (2011). Imaging of phospholipids in formalin fixed rat brain sections by matrix assisted laser desorption/ionization mass spectrometry. J Am Soc Mass Spectrom 22:1991-1998. Cho Y-T, Chiang Y-Y, Shiea J, Hou M-F (2012). Combining MALDI-TOF and molecular imaging with principal component analysis for biomarker discovery and clinical diagnosis of cancer. Genomic Medicine, Biomarkers, and Health Sciences 4:3-6. Cornett DS, Frappier SL, Caprioli RM (2008). MALDI-FTICR imaging mass spectrometry of drugs and metabolites in tissue. Anal Chem 80:5648-5653. Cornett DS, Reyzer ML, Chaurand P, Caprioli RM (2007). MALDI imaging mass spectrometry: molecular snapshots of biochemical systems. Nat Methods 4:828-833. de la Monte SM (1988). Disproportionate atrophy of cerebral white matter in chronic alcoholics. Arch Neurol 45:990-992. Deininger SO, Ebert MP, Futterer A, Gerhard M, Rocken C (2008). MALDI imaging combined with hierarchical clustering as a new tool for the interpretation of complex human cancers. J Proteome Res 7:5230-5236. Di Girolamo F, Lante I, Muraca M, Putignani L (2013). The Role of Mass Spectrometry in the “Omics” Era. Curr Org Chem 17:2891-2905. Dreisewerd K (2003). The desorption process in MALDI. Chem Rev 103:395-426. Durazzo TC, Mattsson N, Weiner MW, Alzheimer’s Disease Neuroimaging I (2014). Smoking and increased Alzheimer’s disease risk: a review of potential mechanisms. Alzheimers Dement 10:S122-145. Eriksson C, Masaki N, Yao I, Hayasaka T, Setou M (2013). MALDI Imaging Mass Spectrometry-A Mini Review of Methods and Recent Developments. Mass Spectrom (Tokyo) 2:S0022. Escobedo G, Lopez-Ortiz E, Torres-Castro I (2014). Gut microbiota as a key player in triggering obesity, systemic inflammation and insulin resistance. Rev Invest Clin 66:450-459. Franck J, Arafah K, Elayed M, Bonnel D, Vergara D, Jacquet A, Vinatier D, Wisztorski M, Day R, Fournier I, Salzet M (2009). MALDI imaging mass spectrometry: state of the art technology in clinical proteomics. Mol Cell Proteomics 8:2023-2033. Goodwin RJ, Nilsson A, Borg D, Langridge-Smith PR, Harrison DJ, Mackay CL, Iverson SL, Andren PE (2012). Conductive carbon tape used for support and mounting of both whole animal and fragile heat-treated tissue sections for MALDI MS imaging and quantitation. J Proteomics 75:4912-4920. Groseclose MR, Andersson M, Hardesty WM, Caprioli RM (2007). Identification of proteins directly from tissue: in situ tryptic digestions coupled with imaging mass spectrometry. J Mass Spectrom 42:254-262. Groseclose MR, Massion PP, Chaurand P, Caprioli RM (2008). High-throughput proteomic analysis of formalin-fixed paraffin-embedded tissue microarrays using MALDI imaging mass spectrometry. Proteomics 8:3715-3724. Hajela N, Ramakrishna BS, Nair GB, Abraham P, Gopalan S, Ganguly NK (2015). Gut microbiome, gut function, and probiotics: Implications for health. Indian J Gastroenterol Hankin JA, Barkley RM, Murphy RC (2007). Sublimation as a method of matrix application for mass spectrometric imaging. J Am Soc Mass Spectrom 18:1646-1652.

MALDI-IMS Tissue Lipids Jackson SN, Barbacci D, Egan T, Lewis EK, Schultz JA, Woods AS (2014). MALDI-Ion Mobility Mass Spectrometry of Lipids in Negative Ion Mode. Anal Methods 6:5001-5007. Khatib-Shahidi S, Andersson M, Herman JL, Gillespie TA, Caprioli RM (2006). Direct molecular analysis of whole-body animal tissue sections by imaging MALDI mass spectrometry. Anal Chem 78:6448-6456. Knochenmuss R, Zenobi R (2003). MALDI ionization: the role of in-plume processes. Chem Rev 103:441-452. Landgraf RR, Prieto Conaway MC, Garrett TJ, Stacpoole PW, Yost RA (2009). Imaging of lipids in spinal cord using intermediate pressure matrix-assisted laser desorption-linear ion trap/Orbitrap MS. Anal Chem 81:8488-8495. Lewis JK, Wei J, Siuzdak G (2000). Matrix-assisted Laser Desorption/Ionization Mass Spectrometry in Peptide and Protein Analysis. In Encyclopedia of Analytical Chemistry. John Wiley & Sons Ltd, 5880-5894 Meding S, Walch A (2013). MALDI imaging mass spectrometry for direct tissue analysis. Methods Mol Biol 931:537-546. Murphy RC, Hankin JA, Barkley RM, Zemski Berry KA (2011). MALDI imaging of lipids after matrix sublimation/deposition. Biochim Biophys Acta 1811:970-975. Neuman MG, French SW, French BA, Seitz HK, Cohen LB, Mueller S, Osna NA, Kharbanda KK, Seth D, Bautista A, Thompson KJ, McKillop IH, Kirpich IA, McClain CJ, Bataller R, Nanau RM, Voiculescu M, Opris M, Shen H, Tillman B, Li J, Liu H, Thomes PG, Ganesan M, Malnick S (2014). Alcoholic and non-alcoholic steatohepatitis. Exp Mol Pathol 97:492-510. Norris JL, Caprioli RM (2013). Analysis of tissue specimens by matrix-assisted laser desorption/ionization imaging mass spectrometry in biological and clinical research. Chem Rev 113:2309-2342. Seeley EH, Caprioli RM (2008). Molecular imaging of proteins in tissues by mass spectrometry. Proc Natl Acad Sci U S A 105:18126-18131. Setou M, Shrivas K, Sroyraya M, Yang H, Sugiura Y, Moribe J, Kondo A, Tsutsumi K, Kimura Y, Kurabe N, Hayasaka T,

771 Goto-Inoue N, Zaima N, Ikegami K, Sobhon P, Konishi Y (2010). Developments and applications of mass microscopy. Med Mol Morphol 43:1-5. Sugiura Y, Honda K, Kajimura M, Suematsu M (2014). Visualization and quantification of cerebral metabolic fluxes of glucose in awake mice. Proteomics 14:829-838. Taban IM, Altelaar AF, van der Burgt YE, McDonnell LA, Heeren RM, Fuchser J, Baykut G (2007). Imaging of peptides in the rat brain using MALDI-FTICR mass spectrometry. J Am Soc Mass Spectrom 18:145-151. Vidova V, Pol J, Volny M, Novak P, Havlicek V, Wiedmer SK, Holopainen JM (2010). Visualizing spatial lipid distribution in porcine lens by MALDI imaging high-resolution mass spectrometry. J Lipid Res 51:2295-2302. Walch A, Rauser S, Deininger SO, Hofler H (2008). MALDI imaging mass spectrometry for direct tissue analysis: a new frontier for molecular histology. Histochem Cell Biol 130:421-434. Wang HY, Jackson SN, Post J, Woods AS (2008). A Minimalist Approach to MALDI Imaging of Glycerophospholipids and Sphingolipids in Rat Brain Sections. Int J Mass Spectrom 278:143-149. Wang JJ, Durazzo TC, Gazdzinski S, Yeh PH, Mon A, Meyerhoff DJ (2009). MRSI and DTI: a multimodal approach for improved detection of white matter abnormalities in alcohol and nicotine dependence. NMR Biomed 22:516-522. Wiley WC, McLaren IH (1955). Time-of-Flight Mass Spectrometer with Imrpoved Resolution. The Review of Scientific Instruments 26:1150-1157. Yamada Y, Hidefumi K, Shion H, Oshikata M, Haramaki Y (2011). Distribution of chloroquine in ocular tissue of pigmented rat using matrix-assisted laser desorption/ionization imaging quadrupole time-of-flight tandem mass spectrometry. Rapid Commun Mass Spectrom 25:1600-1608. Zavalin A, Todd EM, Rawhouser PD, Yang J, Norris JL, Caprioli RM (2012). Direct imaging of single cells and tissue at sub-cellular spatial resolution using transmission geometry MALDI MS. J Mass Spectrom 47:i.