The aim of this protocol is to provide detailed guidance on the proper sample preparation for lipid and metabolite analysis in small tissues, such as the Drosophila brain, using matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging.
Lipid profiling, or lipidomics, is a well-established technique used to study the entire lipid content of a cell or tissue. Information acquired from lipidomics is valuable in studying the pathways involved in development, disease, and cellular metabolism. Many tools and instrumentations have aided lipidomics projects, most notably various combinations of mass spectrometry and liquid chromatography techniques. Matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI MSI) has recently emerged as a powerful imaging technique that complements conventional approaches. This novel technique provides unique information on the spatial distribution of lipids within tissue compartments, which was previously unattainable without the use of excessive modifications. The sample preparation of the MALDI MSI approach is critical and, therefore, is the focus of this paper. This paper presents a rapid lipid analysis of a large number of Drosophila brains embedded in optimal cutting temperature compound (OCT) to provide a detailed protocol for the preparation of small tissues for lipid analysis or metabolite and small molecule analysis through MALDI MSI.
Lipids are involved in a wide range of biological processes and can be broadly classified into five categories based on their structural diversity: fatty acids, triacylglycerols (TAGs), phospholipids, sterol lipids, and sphingolipids1. The fundamental functions of lipids are to provide energy sources for biological processes (i.e., TAGs) and form cellular membranes (i.e., phospholipids and cholesterol). However, additional roles of lipids have been noted in development and diseases, and have been extensively studied in the biomedical field. For instance, reports have shown that fatty acids of different lengths may have unique therapeutic roles. Short fatty acid chains can be involved in defense mechanisms against autoimmune diseases, medium-length fatty acid chains produce metabolites that can mitigate seizures, and long fatty acid chains generate metabolites that can be used to treat metabolic disorders2. In the nervous system, glia-derived cholesterol and phospholipids have been shown to be vital for synaptogenesis3,4. Other types of lipids have shown promise in medical applications, including sphingolipids utilized in drug delivery systems and saccharolipids used to support the immune system5,6. The numerous roles and potential therapeutic applications of lipids in the biomedical field have made lipidomics—the study of the pathways and interactions of cellular lipids—a critical and increasingly important field.
Lipidomics makes use of analytical chemistry to study the lipidome on a large scale. The main experimental methods utilized in lipidomics are based on mass spectrometry (MS) coupled with various chromatography and ion-mobility techniques7,8. The use of MS in the area is advantageous due to its high specificity and sensitivity, speed of acquisition, and unique capabilities to (1) detect lipids and lipid metabolites occurring even at low and transient levels, (2) detect hundreds of different lipid compounds in a single experiment, (3) identify previously unknown lipids, and (4) distinguish between lipid isomers. Among the developments in MS, including desorption electrospray ionization (DESI), MALDI, and secondary ion mass spectrometry (SIMS), MALDI MSI has emerged as a powerful imaging technique that complements conventional MS-based approaches by providing unique information on the spatial distribution of lipids within tissue compartments9,10.
The typical workflow of lipidomics consists of sample preparation, data acquisition using mass-spectrometry technology, and data analysis11. The study of lipids and metabolites in samples has led to the emergence of techniques to understand the physiological and pathological conditions of metabolic processes in organisms. While understanding biological interactions is important, the sensitivity of lipids and metabolites makes them difficult to image and identify without dyes or other modification. Changes in metabolite levels or distribution may lead to phenotypic changes. One tool used for metabolomic profiling is MALDI MSI, a label-free, in situ imaging technique capable of detecting hundreds of molecules simultaneously. MALDI imaging allows for the visualization of metabolites and lipids in samples while preserving their integrity and spatial distribution. Previous technology for lipid profiling involved the use of radioactive chemicals to individually map lipids, while MALDI imaging forgoes this and allows for the detection of a range of lipids simultaneously.
Lipid metabolism and homeostasis play important functions in cell physiology, such as the maintenance and development of the nervous system. One essential aspect of nervous system lipid metabolism is the lipid shuttling between neurons and glial cells, which is mediated by molecular carrier lipoproteins, including very-low-density lipoprotein (VLDL), low-density lipoproteins (LDL), and high-density lipoproteins (HDL)12. Lipoproteins contain apolipoproteins (Apo), such as ApoB and ApoD, which function as structural blocks of lipid cargo and as ligands for lipoprotein receptors. The neuron-glia crosstalk of lipids involves multiple players such as glia-derived ApoD, ApoE, and ApoJ, and their neuronal LDL receptors (LDLRs)13,14. In Drosophila, apolipophorin, a member of the ApoB family, is a major hemolymph lipid carrier15. Apolipophorin has two closely related lipophorin receptors (LpRs), LpR1 and LpR2, which are homologs of mammalian LDLR15,16. In previous studies, the astrocyte-secreted lipocalin Glial Lazarillo (GLaz), a Drosophila homolog of human ApoD, and its neuronal receptor LpR1 were discovered to cooperatively mediate neuron-glia lipid shuttling, thus regulating dendrite morphogenesis17. Therefore, it was speculated that the loss of LpR1 would cause a decrease in overall lipid content in the Drosophila brain. MALDI MSI would be a suitable tool for profiling the lipid contents in small tissues of LpR1−/− mutant and wild-type Drosophila brains, as demonstrated in this study.
Despite the growing popularity of MALDI MSI, the instrument's high cost and experimental complexity often impede its implementation in individual laboratories. Thus, most MALDI MSI studies are conducted using shared core facilities. As with other applications of MALDI MSI, a careful sample preparation process for lipidomics is critical to achieve reliable results. However, because sample slide preparation is typically performed in individual research laboratories, there is a possibility of variation in MALDI MSI acquisition. To combat this, this paper aims to provide a detailed protocol for the sample preparation of small biological samples prior to MALDI MSI measurement using lipid analysis of a large group of adult Drosophila brains in positive ion mode as an example11,17. However, some phospholipid classes and the majority of small metabolites are favorably detected by MALDI imaging in negative ion mode, which was described previously11. Therefore, with these two example studies, we hope to provide detailed sample preparation protocols of various combinations: free-standing large tissue versus embedded small tissue, thaw-mounting versus warm-slide mounting, and positive ion mode versus negative ion mode.
1. Fly head embedding
NOTE: The whole procedure takes ~45-60 min.
2. Cryosectioning the tissue
NOTE: When handling the indium tin oxide (ITO) slides, wear gloves at all times to avoid tissue contamination. Wearing a mask is also recommended to avoid breathing directly onto the slide.
The loss of the neuronal receptor LpR1 of the astrocyte-secreted lipocalin Glial Lazarillo (GLaz), a Drosophila homolog of human ApoD, was hypothesized to be able to cause a decrease in overall lipid content in the Drosophila brain. To test this, MALDI MSI was used to profile the lipids in LpR1−/− mutant and wild-type Drosophila brains, which is elaborated on below.
The experiment was performed according to the workflow shown in Figure 1. Adult fly brains were harvested as described above. They were embedded in OCT, and dry ice was used to freeze the OCT block. The OCT tissue block was cryosectioned at 12 µm thickness and at −18 °C for both the specimen head and the chamber. ITO slides with cryosections mounted on them were desiccated in a vacuum for 30 min at room temperature. Next, matrix deposition was carried out using the automatic HTX M5 matrix sprayer and a 40 mg/mL solution of DHB in methanol/water (70/30, v/v).
The MALDI time of flight (TOF) MS Autoflex instrument in positive ion mode acquired a mass spectrum within the mass ranges from 50-1,000 m/z. The spectrum of both the OCT and brain tissue from the same experiment were aligned in SCiLS to evaluate the overlapping peaks and the ion suppression interference of the OCT (Supplemental Figure S1).
The results in Figure 2 show output images from MALDI MSI data analysis software of selected m/z spectra where the selected values showed a significant difference in lipid distribution between LpR1 knockout and wild-type Drosophila. The scans revealed a general reduction in lipid contents in the LpR1−/− mutant. Each spectrum was analyzed manually, and analyte identification was achieved by cross-referencing METLIN databases and previously published studies3,4. Triacylglycerol (TAG) and phosphatidylglycerol (PG) were highly expressed in control brains compared to the LpR1−/− mutant (showing up to a 10-fold change in TAG). Additionally, diglycerol (DAG), phosphatidylcholine (PC), and phosphatidyl-ethanolamine (PE) were also minimally expressed in the LpR1−/− mutant genotype.
Figure 1: Workflow of MALDI-TOF mass spectrometry imaging. (A) Drosophila brains are aligned in OCT placed on top of dry ice. (B–D) The tissue block is cryosectioned into 12 µm thick sections at −15 °C. The section is flattened using a precooled brush and thaw-melted onto the ITO-coated side of a glass slide kept at room temperature. (E) The slide is marked with fiducial markers, coated with matrix, and placed into the MALDI instrument. (F) The MALDI image of a Drosophila brain is obtained at 40 µm resolution using DHB as matrix. The regions of eyes (red, m/z 370.2), head tissue (green, m/z 184.1), and brain (blue, m/z 788.6) are shown in the overlayed multi-channel image. (G) H&E staining is performed on the same tissue section after MALDI MSI. (H) The workflow of sample preparation, storage, and transportation. Scale bars = 500 µm (F,G). Abbreviations: MALDI-TOF = matrix-assisted laser desorption/ionization-time of flight; OCT = optimal cutting temperature compound; ITO = indium tin oxide; DHB = 2,5,-dihydroxybenzoic acid; H&E = hematoxylin and eosin; MSI = mass spectrometry imaging. Please click here to view a larger version of this figure.
Figure 2: Representative images from MALDI MS analysis revealing a general decrease in lipid contents in the LpR1−/− mutant brain. Representative H&E-stained adult fly brain sections are shown (left). The lipid species, their respective m/z values, and the scales of the heatmap are as indicated. At the bottom, the average fold of reduction in the LpR1−/− mutant as compared to the controls from at least four biological replicates are shown as numbers next to the arrows. Scale bar = 1 mm. This figure has been modified from Yin et al.17. Please click here to view a larger version of this figure.
Supplemental Figure S1: Mass spectra of OCT and Drosophila brain tissue. The cryosections of Drosophila brain OCT block from Figure 1 were covered evenly with DHB matrix before MALDI imaging. The mass spectrum of the selected blank OCT region and brain tissue region from both control and mutant flies were shown in the range of m/z 1-1,000 and m/z 520-900 (lipid-enriched), respectively. The interference of OCT is associated with both ion suppression phenomena and overlapping signal issues. Abbreviations: OCT = optimal cutting temperature compound; DHB = 2,5-dihydroxybenzoic acid; MALDI = matrix-assisted laser desorption/ionization. Please click here to download this File.
As demonstrated in the study on the variations in lipid composition in mutant and wild-type Drosophila brains, MALDI MSI can be a valuable label-free imaging technique for in situ analysis of molecular distribution patterns within organs of small insects. Indeed, because lipids are distributed in both the brain tissue and fat bodies of Drosophila heads, conventional lipidomics approaches based on liquid-chromatography and mass-spectrometry (LC-MS) can detect only combined signals from both regions when whole-head extraction is used. Broadly, differential lipidomics analysis of subregions within organs of such a small insect like Drosophila is a very laborious and challenging task for LC-MS approaches18. It would require dissection of those regions, which imposes great challenges. In contrast, MALDI MSI provides adequate resolution to distinguish different anatomic structures, even within the small Drosophila brain. The benefit of a minimum amount of tissue requirement compared to conventional LC-MS could also be applied to precious samples such as clinical biopsies. Furthermore, the sample slides prepared for MALDI MSI could also be examined by complementary techniques such as TOF-SIMS and immunohistochemistry, which would enable the integration of the data obtained from cross-disciplinary approaches19,20.
One of the most important steps in MALDI MSI is the sample preparation—the part of the experiment that accounts for most of the differences in the results from metabolomics studies carried out in different labs21. The main goal here is to provide a comprehensive yet practical sample preparation protocol for MALDI MSI analysis of lipids and metabolites in organisms as small as Drosophila insects in the hope of aiding researchers with the implementation of MALDI MSI in their studies from basic biology to translational science.
In addition to the precautions to minimize changes in the molecular profiles (both abundance and spatial distribution) as previously discussed11, several other practices need to be adapted when handling small tissues. First, the time between biological tissue harvesting and freezing in OCT should be minimized to reduce the likelihood of postmortem ischemia21. Second, one should ensure that a flat stage of OCT is prepared before placing the tissue into the block, which is crucial for having sections from comparable planes across different tissues during sectioning. Third, the cutting of small-scale biological tissue or precious samples needs adequate practice before the experiment. While the OCT aids in cutting even sections, complications may arise in terms of tissue curling or flaking. To combat curling, one should allow the section to rest on the stage for a few seconds before removing the roll plate. With two brushes, one can be used to hold the top of the section still, while the other can be used to unfurl the section. Care should be taken to minimize contact with the tissue by primarily working with the edges of the block sections. Fourth, although the thaw-mounting method could be used for fewer fly brains (<10 brains), it would be impossible to transfer a section containing a large number of fly brains (>30 brains) to a cold ITO slide without losing a remarkable number of brains, which would easily fall out of the block section once lifted. Therefore, for rapid analysis of a large number of fly brains, warm-slide mounting needs to be used. Notably, the temperature difference between the OCT section and the slide will cause the section to cling and adhere to the slide very quickly. Hence, one should not press the warm slide against the section and cryostat stage; instead, one must gently and quickly approach the sections to allow the section to attach to the ITO slide without touching the cryostat stage. We did not observe any noticeable trace of the sections left behind on the cryostat stage compared to the thaw-mounting method. Fifth, a uniform and fine deposition of MALDI matrix is crucial to achieve strong and accurate spatial information during acquisition. It is recommended to use a blank slide to test the matrix deposition for proper coverage before proceeding to the slide containing the sample.
With its growing popularity and advancements, MALDI MSI is expected to reach a broader base of users and become a standard tool for molecular mass measurements of analytes such as amino acids, metabolites, lipids, peptides and proteins, and other small molecules directly extracted from biological tissues10. However, it has its own limitations and challenges. When preparing fragile and small-scale samples for MALDI MSI, embedding agents such as OCT are necessary for sectioning. However, OCT contains a combination of polymers (poly(vinyl alcohol), polyethylene glycerol, and benzalkonium chloride), which creates a polymeric background signal that may interfere with the analyte signal22. Alternative embedding compounds such as gelatin and carboxymethyl cellulose (CMC) have been reported to demonstrate less ion suppression effects. Gelatin shows less intense signals that are more scattered in the low-mass range of 100-600 m/z20,23,24. CMC has also been used in MALDI MSI of Drosophila to visualize metabolites such as GABA, though it requires immersing the heads in 70% ethanol prior to embedding for optimal adhesion25.
Despite OCT's tendency for signal suppression, this protocol proceeds with its use due to the superior ability of OCT to provide reliable sectioning of a large number of samples while preserving brain morphology. Researchers have found that, while gelatin and CMC section well at high thicknesses such as 20 µm, only OCT is able to produce consecutive 12 µm sections reliably26. CMC and gelatin lack the structural integrity and tight hold of tissue that OCT provides, which sets the limit of the number of small tissues such as fly brain to be accommodated in one block26. It has been reported that MS signals from OCT-embedded fly sections are comparable to those embedded in gelatin and CMC26. Our unpublished data also indicated that warm-mounted, OCT-embedded fly brain sections generated robust lipid signals comparable to thaw-mounted gelatin-embedded tissue. Overall, due the superior ability of OCT to preserve the integrity of tissue morphology and enable precise sample sectioning compared to other embedding compounds, its usage with samples of high fragility, such as arrays of Drosophila brains, remains an option. In that scenario, only the relative quantification of comparing samples from the same experiment, but not absolute quantification, should be made. With respect to our reported study of the Drosophila model in lipid analysis, it was concluded that the benefits of OCT outweigh its limitations. In addition, trial experiments must be performed to test whether the MS signals of the targeted analytes would be masked by the OCT signals.
Furthermore, several practices need to be adapted to minimize the effects of ion suppression. First, one should process the samples of different groups at the same time to ensure the same extent of ion suppression, if there is any. Second, in the selection of the region of interest for MALDI scanning, one should avoid the OCT region and outline only the tissue region. Lastly, a region of blank OCT should be selected to be scanned as the negative control to exclude the signals from OCT during data analysis.
The field of lipidomics emerged in the early 2000s and has rapidly advanced in recent years, largely due to developments in mass spectrometry, including MALDI MSI27. These advancing techniques have aided in driving the field toward biological and biomedical applications. For example, lipidomics can be employed in neurological studies as changes in levels of brain lipid trafficking and composition can be used as biomarkers of disease. Additionally, lipidomics has led to the identification of new signaling molecules, the development of drug efficacy tests, the discovery of mechanisms underlying pathophysiological conditions, and more28. Despite the remarkable achievements made by the field in the past few years, there are still areas that demand growth. For instance, the ability for a lipid to be accurately quantified using current technology remains under debate29. In addition, the complete mapping of the cellular lipidome has much progress to make. It is expected that these areas, among others in the field, will experience significant growth in the coming years.
The authors have nothing to disclose.
Yuki X. Chen, Kelly Veerasammy, and Mayan Hein are supported by the Sloan Foundation CUNY Summer Research Program (CSURP). Jun Yin is supported by the intramural research program of the National Institutes of Health Project Number 1ZIANS003137. Support for this project was provided by a PSC-CUNY Award to Ye He and Rinat Abzalimov, jointly funded by The Professional Staff Congress and The City University of New York.
2,5-Dihydroxybenzoic acid (DHB) | Millipore Sigma Aldrich | 85707-1G-F | |
Andwin Scientific CRYOMOLD 15X15X5 | Fisher Scientific | NC9464347 | |
Andwin Scientific Tissue-Tek CRYO-OCT Compound | Fisher Scientific | 14-373-65 | |
Artist brush MSC #5 1/8 X 9/16 TRIM RED SABLE | Fisher Scientific | 50-111-2302 | |
autoflex speed MALDI-TOF MS system | Bruker Daltonics Inc | MALDI-TOF MS instrument | |
BD Syringe with Luer-Lok Tips | Fisher Scientific | 14-823-16E | |
BD Vacutainer General Use Syringe Needles | Fisher Scientific | 23-021-020 | |
Bruker Daltonics GLASS SLIDES MALDI IMAGNG | Fisher Scientific | NC0380464 | |
Drierite, with indicator, 8 mesh, ACROS Organics | AC219095000 | ||
Epson Perfection V600 Photo Scanner | Amazon | Perfection V600 | |
Fisherbrand 5-Place Slide Mailer | Fisher Scientific | HS15986 | |
Fisherbrand Digital Auto-Range Multimeter | Fisher Scientific | 01-241-1 | |
FlexImaging v3.0 | Bruker Daltonics Inc | Bruker MS imaging analysis software | |
HPLC Grade Methanol | Fisher Scientific | MMX04751 | |
HPLC Grade Water | Fisher Scientific | W5-1 | |
HTX M5 Sprayer | HTX Technologies, LLC | Automatic heated matrix sprayer | |
Kimberly-Clark Professional Kimtech Science Kimwipes Delicate Task Wipers | Fisher Scientific | 06-666A | |
MSC Ziploc Freezer Bag | Fisher Scientific | 50-111-3769 | |
SCiLS Lab (2015b) | SCiLS Lab | Advanced MALDI MSI data analysis software | |
Thermo Scientific CryoStar NX50 Cryostat | Fisher Thermo Scientific | 95-713-0 | |
Thermo Scientific Nalgene Transparent Polycarbonate Classic Design Desiccator | Fisher Scientific | 08-642-7 |