Sequential Detection of Biomolecules in Formalin-fixed, Paraffin-embedded Samples with Mass Spectrometry Imaging

Summary

Here, we present a Mass Spectrometry Imaging protocol for sequential metabolite, N-linked glycan, and tryptic peptide detection in formalin-fixed, paraffin-embedded tissue samples.

Abstract

Tissues are complex cellular environments, made up of a vast array of cell types and biomolecules all interacting with each other to carry out the functions of the organ. Traditionally, techniques for the analysis of biomolecules such as metabolites, glycans, and proteins involved the homogenization of tissue, destroying all spatial information. These traditional methods inhibited the complete understanding of complex intra- and extracellular molecular interactions. Mass Spectrometry Imaging (MSI), on the other hand, not only preserves the spatial information of biomolecules in tissue but also allows for multiple classes of analytes to be detected from the same tissue section through sequential analysis at a near-single-cell resolution. This enables us to derive a more complete picture of molecular interactions across different classes of biomolecules. The protocol presented here outlines the steps for performing mass spectrometry imaging of metabolites, N-linked glycans, and tryptic peptides sequentially from the same tissue section at a 20 µm resolution. Conscientious consideration of the order in which the classes of analytes are imaged, along with careful handling of the sections to ensure integrity, allows for multiple high-quality images to be collected from the same section. These data can subsequently be integrated with other spatial omics data (transcriptomics, immunohistochemistry, etc.) collected from serial sections, where the same cell can be analyzed in these adjacent sections.

Introduction

Mass Spectrometry Imaging (MSI) is a powerful technique that allows for the detection and visualization of hundreds to thousands of biomolecules from thin tissue sections, without the need for a priori knowledge of the exact molecules present in the tissue, making it an excellent tool for biomarker discovery¹. While several types of MSI are used by different labs, including Desorption ElectroSpray Ionization (DESI)^2,3, Infrared Matrix Assisted Laser Desorption ElectroSpray Ionization (IR-MALDESI)⁴, and Secondary Ion Mass Spectrometry (SIMS)⁵, Matrix Assisted Laser Desorption/Ionization (MALDI) remains the most commonly used and most versatile. Most classes of biomolecules can be detected by MSI by tailoring the sample preparation, including washing/pretreatment of the tissue, enzymatic digestion, and the matrix/solvent used^6,7. MALDI MSI has been used for the detection of metabolites, lipids, glycans, proteins, and proteolytic peptides.

MSI studies have been used in a variety of clinical and preclinical studies to better understand disease mechanisms^8,9, improve cancer grading and staging^10,11, predict treatment outcome¹², improve diagnosis¹³, and determine molecular tumor margins¹⁴, among others. Most of these studies have focused on the detection of only a single class of analytes. However, it is often of interest to analyze more than one class of biomolecules from the same sample. Traditionally, that has been performed with serial sections of tissues. But more recently, examples have been shown of sequentially analyzing multiple analyte classes from the same tissue section. For example, Yagnik et al. have demonstrated lipid imaging of frozen tissue followed by MALDI Immunohistochemistry from the same section for cell type classification¹⁵. Clift et al. have shown sequential application of PNGaseF, Collagenase, and Trypsin to formalin-fixed, paraffin-embedded (FFPE) tissue sections for imaging of N-linked glycans, extracellular matrix proteins, and general proteins, respectively¹⁶. Escobar et al. demonstrated the sequential analysis of N-linked glycans, O-GlcNAc, and tryptic peptides from the same section of frozen tissue¹⁷. These types of experiments are accomplished through careful experimental planning to analyze the most easily lost analytes first, followed by those that are more stable in the tissue. In many cases, the washes that are used to remove unwanted classes of molecules help enhance other classes of molecules⁶, a property that can be taken advantage of, where each wash that removes the previously applied matrix also helps to enhance the signal of the next class of analytes to be imaged.

Recently, we published a workflow for the sequential analysis of metabolites, N-linked glycans, and tryptic peptides from the same section of FFPE ovarian cancer tissue¹⁸. Here, we present the detailed workflow for the analysis of these three classes of biomolecules from the same tissue section; an overview of the protocol is detailed in Figure 1. To our knowledge, this is first protocol describing in detail this workflow for sequential imaging from FFPE tissue. Briefly, tissue sections are dewaxed with xylene before coating with 1,5-diaminonaphthalene matrix (10 mg/mL in 50% acetonitrile) for negative ion mode metabolite imaging. The matrix is then removed, the sections rehydrated, and antigen-retrieved, before being sprayed with PNGase F (0.1 µg/µL in ammonium bicarbonate) to release N-linked glycans in situ. After incubation, the sections are coated with α-cyano-4-hydroxycinnamic acid (CHCA) (10 mg/mL in 70% acetonitrile, 0.1% trifluoroacetic acid) matrix and imaged in positive ion mode. Matrix is then removed again, rehydration and antigen retrieval repeated, and the sections are sprayed with trypsin (0.075 µg/µL in ammonium bicarbonate) and incubated before being again coated with CHCA (10 mg/mL in 70% acetonitrile, 0.1% trifluoroacetic acid) matrix and imaged in positive ion mode. All solutions should be made fresh, immediately before use, and if at all possible, the experiments should be carried out over 3 consecutive days. If delays are encountered, sections should be stored in a -80 °C freezer under desiccation.

Protocol

This protocol uses formalin-fixed paraffin-embedded (FFPE) tissues collected from previously untreated patients undergoing primary cytoreductive surgery for high-grade serous ovarian carcinoma. All clinical data were obtained from the ovarian cancer repository of the Department of Gynecologic Oncology and Reproductive Medicine under protocols approved by the University of Texas MD Anderson's Institutional Review Board. Written informed consent from the patients was obtained by front desk personnel, and the studies were conducted in accordance with recognized ethical guidelines.

1. Metabolite imaging sample preparation

NOTE: This protocol assumes that sections of formalin-fixed, paraffin-embedded tissue (4 µm thickness) have already been mounted on glass microscope slides, as clinical samples must generally be sectioned within a clinical pathology core by histotechnicians. As the work presented here is performed on the referenced TOF instrument, the orthogonal TOF measurement alleviates the need for conductive slides traditionally used in MSI experiments. The use of standard microscope slides is also more conducive to standard workflows in clinical pathology labs, as well as enabling the use of archival tissue already on slides.

1. Make fiducial marks at the four corners of the slide using a diamond scribe.
2. Deparaffinize sections by submersion in histological grade xylene for 3 min.
   NOTE: Xylene is noxious and should be used in a chemical fume hood.
3. Submerge sections in fresh xylene for 3 min.
4. Allow the slide to dry at room temperature.
5. Place the slide into a Slide Adapter Target.
6. Acquire an optical image of the slide using a flatbed scanner at 4,800 dpi resolution.
7. Make a MALDI matrix by dissolving 15 mg of 1,5-diaminonaphthalene (DAN) in 1.5 mL of 50% ACN. Sonicate until completely dissolved, ~5 min. Filter the matrix solution before loading into the sample loop.
   NOTE: DAN is a suspected carcinogen and should be used with proper personal protective equipment. Matrix spraying should be performed in a chemical fume hood.
8. Spray the slide with DAN solution using a Robotic Reagent Sprayer with the following parameters: Track Speed - 1,200 mm/s; Track Spacing - 3 mm; Nozzle Temperature - 60 °C; Nozzle Height - 40 mm; Flow Rate - 100 µL/min; Passes - 4; N ₂ Pressure - 10 psi; Pattern - CC
9. Apply 0.5 µL of red phosphorus suspended in methanol to a known location on the slide.

2. Metabolite imaging data collection

1. Place the prepared slide(s) into a Slide Adapter II and load into a timsTOF fleX MALDI QTOF mass spectrometer.
2. Load the appropriate method for negative ion mode data collection in timsControl with tims OFF. If the instrument was not already in negative ion mode, allow at least 25 min after switching polarity for equilibration before calibration and data collection with the following instrument parameters for metabolites: m/z range - 50 - 600; Laser Scan Range - 16.0 µm in both X and Y; Resulting Field Size - 20.0 µm in both X and Y; Laser Shots - 150; Laser Fluence - ~30%; Funnel 1 RF - 150.0 Vpp; Funnel 2 RF - 200.0 Vpp; Multipole RF - 200.0 Vpp; Collision Energy - 10.0 eV; Collision RF - 500.0 Vpp; Transfer Time - 55.0 µs; PrePulse Storage - 5.0 µs (Supplemental Figure S1).
3. Perform a Target Profile (click Generate Now button) and Focus Adjustment (click Adjust Now button) on the Sample Carrier Tab. Adjust laser position in X and Y, if needed on the laser tab to ensure that the laser is exactly aligned with the crosshairs.
4. Collect a spectrum of red phosphorus by summing at least five sets of laser shots from the spot on the slide.
5. Perform a calibration by clicking the Calibrate button on the Calibration Tab. Remove any peaks with very low intensity or high ppm Error.
   NOTE: Even phosphorus clusters (P2 and P4) generally give a very weak signal. The calibration score should be 100% before clicking accept. If not, clear the summed spectrum and acquire a new one.
6. Launch the flexImaging software. The image acquisition GUI will start automatically.
7. Select Set up a new imaging run on the Choose Imaging Run window.
8. Choose Classic workflow on the Select Workflow window.
9. On the Data Storage Window, provide an Imaging Run Name and a Result Directory for the data.
   NOTE: Spaces should not be used in file names.
10. On the Acquisition Settings window, choose User-defined measurement regions and type 20 into the Raster Width box. Click Browse to select the timsControl method previously created.
11. On the Sample Image window, click Browse to select the optical image acquired using the flatbed scanner.
12. On the Teach Sample window, perform a three-point alignment between the optical image and the slide; then, move the sample carrier to the first etched fiducial on the slide and align the crosshairs in timsControl with a distinct feature in the etching. Zoom in on the picture (100% Zoom recommended), set the Sample Image to Set Teach Points, and click on the same location in the optical picture. Repeat this process 2x more for a total of three teach points. Click Next to complete the teaching.
13. Click OK to exit the New Imaging Run Wizard.
14. Check the accuracy of the teaching by clicking on the Position Sample Carrier button in the software. Zoom in and click on a distinct feature of the fourth (unused) fiducial point in the picture; the sample carrier will move to this position within timsControl. Confirm that the crosshairs line up with the selected feature. If they do not line up, clear the teaching in flexImaging and repeat the teaching process to improve accuracy.
15. In the software, select the Add Polygon Measurement Region tool and outline the tissue(s) to be imaged. Each click of the mouse will create an anchor point. Press the backspace button on the keyboard to remove points, one-by-one. Once the region is drawn, double-click the mouse to close the polygon.
16. Check the signal by firing the laser at a few locations within the tissue. Click on a location in the picture using the Position Sample Carrier button, as in Step 2.14, for checking the teaching. The spectrum will be dominated by matrix peaks around m/z 157 and 313; zoom in on the m/z region below 154 and 250-300 to show many small molecule metabolite and fatty acid peaks, respectively. Adjust laser fluence as needed for optimal signal.
17. Save the timsControl method.
18. Start the imaging acquisition by clicking the Start AutoXecute Run button in the software
    NOTE: Imaging in negative ion mode with DAN matrix results in the detection of numerous metabolites from the TCA cycle, amino acids, and fatty acids, among others.

3. Sample preparation for N-linked glycan imaging

1. After removing the sample from the mass spectrometer, remove the DAN matrix by submerging the slide in 100% ethanol (200 proof) for 1 min.
2. Rehydrate the slide in Coplin jars: Fresh 100% ethanol for 1 min; 95% ethanol for 1 min; 70% ethanol for 1 min; Water for 3 min; Fresh water for 3 min.
3. Allow the slide to dry for ~15 min.
4. Perform antigen retrieval by submerging the slide in 100 mM Tris buffer (made by dissolving 6.06 g of Tris Base in 500 mL of water and adjust pH as needed), pH 9, in a Coplin jar with a lid placed loosely on top that is placed into a Decloaking Chamber with 500 mL of water in the chamber. Heat to 95 °C for 20 min.
5. Remove the Coplin Jar from the Decloaking Chamber and place it on the benchtop to cool for 20 min.
6. Buffer-exchange the slide to water over five changes. For the first three, pour out half of the solution and replace it with water. For the final two, pour out all of the solution and replace it with water.
7. Allow the slide to completely dry.
8. Create a humidity chamber for incubation of the slide after PNGaseF application. Cut a circle of an absorbent wipe to cover the bottom of a 100 mm diameter Petri dish. Roll two laboratory wipes to create pillows at either side of the Petri dish (Figure 2). Add 8 mL of water to saturate these wipes and put the lid on the Petri dish and place it in an oven at 37 °C for at least 30 min to preheat.
   NOTE: The inside of the dish should appear steamy.
9. Prepare PNGaseF by reconstituting 100 µg of lyophilized enzyme in 200 µL of 100 mM ammonium bicarbonate, pH 7.8, and vortex to mix. Transfer the solution to a 1.5 mL microcentrifuge tube and dilute to 1 mL with water.
10. Spray the slide with PNGaseF using the Robotic Reagent Sprayer with the following parameters. Use the syringe pump for low volume/low flow rate applications: Track Speed - 1200 mm/s; Track Spacing - 3 mm; Nozzle Temperature - 45 °C; Nozzle Height - 40 mm; Flow Rate - 25 µL/min; Passes - 15; N ₂ Pressure - 10 psi; Pattern - CC
11. Place the slide in the preheated Petri dish, tissue side up, with one end on each of the laboratory wipe pillows. Replace the slide and seal the Petri dish with adhesive sealing film. Place in the 37 °C oven and place a reptile heating pad set to 40 °C on top of the Petri dish.
    NOTE: This keeps the lid the warmest part of the dish and prevents condensation from forming on the lid and dropping onto the tissue, which will cause artifacts in the images.
12. Incubate the slide for 2 h to allow digestion to proceed.
13. Remove the slide from the Petri dish and allow to dry for ~5 min.
14. Make the MALDI matrix by dissolving 15 mg of α-cyano-4-hydroxycinnamic acid (CHCA) in 1.5 mL of 70% ACN, 0.1% TFA, 10 mM ammonium phosphate.
    NOTE: TFA is caustic and should be used in a hood with proper personal protective equipment.
15. Sonicate until completely dissolved, ~5 min.
16. Filter the matrix solution before loading into the sample loop.
17. Spray the slide with CHCA using the Robotic Reagent Sprayer with the following parameters: Track Speed - 1,200 mm/s; Track Spacing - 3 mm; Nozzle Temperature - 75 °C; Nozzle Height - 40 mm; Flow Rate - 120 µL/min; Passes - 3; N ₂ Pressure - 10 psi; Pattern - HH
18. Apply 0.5 µL of red phosphorus suspended in methanol to a known location on the slide.

4. N -Linked glycan imaging data collection

1. Place the prepared slide(s) into the Slide Adapter target and load into the MALDI QTOF mass spectrometer.
2. Load the appropriate method for positive ion mode data collection in timsControl. If the instrument was not already in positive ion mode, allow at least 25 min after switching polarity for equilibration before calibration and data collection with the following instrument parameters for glycans: m/z range - 700-3500; Laser Scan Range - 16.0 µm in both X and Y; Resulting Field Size - 20.0 µm in both X and Y; Laser Shots - 200; Laser Fluence - ~25%; Funnel 1 RF - 450.0 Vpp; Funnel 2 RF - 500.0 Vpp; Multipole RF - 500.0 Vpp; Collision Energy - 10.0 eV; Collision RF - 2,700.0 Vpp; Transfer Time - 140.0 µs; PrePulse Storage - 14.0 µs (Supplemental Figure S2).
3. Perform a Target Profile (click Generate Now button) and Focus Adjustment (click Adjust Now button) on the Sample Carrier Tab. Adjust the laser position in X and Y, if needed, on the laser tab to ensure the laser is aligned with the crosshairs .
4. Collect a spectrum of red phosphorus by summing at least five sets of laser shots from the spot on the slide.
5. Perform a calibration by clicking the Calibrate button on the Calibration Tab. Remove any peaks with very low intensity or high ppm Error.
   NOTE: The calibration score should be 100% before clicking accept. If not, clear the summed spectrum and acquire a new one.
6. Open the .mis file previously used for metabolite imaging in the software.
7. Save the file with a new name indicating that it is glycan data.
8. Under Edit | Imaging Run Properties, change the acquisition method to the desired glycan method.
9. Clear the previous teaching using Edit | Clear Teach Points.
   NOTE: The slide will not be in the exact same position after removing it from and reloading it in the slide adapter.
10. Perform new teaching using Edit | Set Teach Points; set 3 points as was done in the original Imaging Wizard (Step 2.12).
11. Check the accuracy of the teaching by clicking on the Position Sample Carrier button in the software. Zoom in and click on a distinct feature of the fourth (unused) fiducial point in the picture. The sample carrier will move to this position within timsControl; confirm that the crosshairs line up with the selected feature.
    NOTE: If the crosshairs do not line up, clear the teaching and repeat the teaching process to improve accuracy.
12. Save the file.
13. Check the signal by firing the laser at a few locations within the tissue. Click on a location in the image using the Position Sample Carrier button as in Step 4.11 to check the teaching. Confirm that the spectrum has abundant glycan peaks with dominant peaks typically at m/z 1,663 and 1,809. Adjust laser fluence as needed for optimal signal.
14. Save the timsControl method.
15. Start the imaging acquisition by clicking the Start AutoXecute Run button.

5. Sample preparation for tryptic peptide imaging

1. Remove the CHCA matrix by submerging the slide in 100% ethanol (200 proof) for ~1 min
2. Rehydrate the slide in Coplin jars as follows: Fresh 100% ethanol for 1 min; 95% ethanol for 1 min; 70% ethanol for 1 min; Water for 3 min; Fresh water for 3 min.
3. Allow the slide to dry for ~15 min.
4. Perform antigen retrieval by submerging the slide in 100 mM Tris buffer, pH 9, in a Coplin jar with a lid placed loosely on top that is placed into a Decloaking Chamber with 500 mL of water in the chamber. Heat to 95 °C for 20 min.
   NOTE: This second antigen retrieval is necessary to denature the proteins for more efficient access of the trypsin to cleavage sites.
5. Remove the Coplin Jar from the Decloaking Chamber and place it on the benchtop to cool for 20 min.
6. Buffer-exchange the slide to water over five changes. For the first three, pour out half of the solution and replace it with water. For the final two, pour out all of the solution and replace it with water.
7. Allow the slide to completely dry.
8. Prepare trypsin by reconstituting 100 µg of lyophilized enzyme in 200 µL of 100 mM acetic acid and pipet up and down to mix. Aliquot into appropriate sizes for experiment (typically 40 µL) and store at -20 °C.
9. Prepare a 40 µL aliquot of trypsin by adding 350 µL of 100 mM ammonium bicarbonate, pH 7.8, and 39 µL of ACN. Mix well.
   NOTE: Acetonitrile (10% ACN) increases the activity of trypsin compared to aqueous solution and reduces the surface tension of the solution for spraying¹⁹.
10. Spray the slide with Trypsin using a Robotic Reagent Sprayer with the following parameters. Use the syringe pump for low volume/low flow rate applications: Track Speed - 750 mm/s; Track Spacing - 3 mm; Nozzle Temperature - 30 °C; Nozzle Height - 40 mm; Flow Rate - 10 µL/min; Passes - 12; N ₂ Pressure - 10 psi; Pattern - HH
11. Place a square of absorbent wipe in the bottom of a 100 mm diameter Petri dish; ensure that the corners just touch the edge of the dish, and distribute 1 mL of water over the wipe.
12. Place the slide on top of the wipe, tissue side up, place the lid on the Petri dish, and seal the Petri dish with adhesive sealing film.
13. Incubate the slide for 4 h at 37 °C to allow digestion to proceed.
14. Remove the slide from the Petri dish and allow it to dry for ~5 min.
15. Make MALDI matrix by dissolving 15 mg of α-cyano-4-hydroxycinnamic acid (CHCA) in 1.5 mL of 70% ACN, 0.1% TFA, 10 mM ammonium phosphate.
16. Sonicate until completely dissolved, ~5 min.
17. Filter the matrix solution before loading into the sample loop.
18. Spray the slide with CHCA using a Robotic Reagent Sprayer with the following parameters: Track Speed - 1200 mm/s; Track Spacing - 3 mm; Nozzle Temperature - 75 °C; Nozzle Height - 40 mm; Flow Rate - 120 µL/min; Passes - 4; N ₂ Pressure - 10 psi; Pattern - HH
19. Apply 0.5 µL of red phosphorus suspended in methanol to a known location on the slide.

6. Tryptic peptide imaging data collection

1. Place the prepared slide(s) into a Slide Adapter and load into a timsTOF MALDI QTOF mass spectrometer.
2. Load the appropriate method for positive ion mode data collection in timsControl. If the instrument was not already in positive ion mode, allow at least 25 min after switching polarity for equilibration before calibration and data collection using the following instrument parameters for peptides: m/z range - 600-4500; Laser Scan Range - 16.0 µm in both X and Y; Resulting Field Size - 20.0 µm in both X and Y; Laser Shots - 200; Laser Fluence - ~25%; Funnel 1 RF - 450.0 Vpp; Funnel 2 RF - 500.0 Vpp; Multipole RF - 600.0 Vpp; Collision Energy - 10.0 eV; Collision RF - 3400.0 Vpp; Transfer Time - 180.0 µs; PrePulse Storage - 18.0 µs (Supplemental Figure S3).
3. Perform a Target Profile (click Generate Now button) and Focus Adjustment (click Adjust Now button) on the Sample Carrier Tab. Adjust laser position in X and Y, if needed, on the laser tab to ensure that the laser is aligned with the crosshairs.
4. Collect a spectrum of red phosphorus by summing at least five sets of laser shots from the spot on the slide.
5. Perform a calibration by clicking the Calibrate button on the Calibration Tab. Remove any peaks with very low intensity or high ppm Error. Confirm that the calibration score is 100% before clicking accept. If not, clear the summed spectrum and acquire a new one.
6. Open the .mis file previously used for glycan imaging.
7. Save the file with a new name indicating that it is peptide data.
8. Under Edit | Imaging Run Properties, change the acquisition method to the desired peptide method.
9. Clear the previous teaching using Edit | Clear Teach Points.
   NOTE: The slide will not be in the exact same position after removing it from and reloading it in the slide adapter.
10. Perform new teaching using Edit | Set Teach Points; set three points as was done in the original Imaging Wizard (Step 2.12).
11. Check the accuracy of the teaching by clicking on the Position Sample Carrier button. Zoom in and click on a distinct feature of the fourth (unused) fiducial point in the picture. The sample carrier will move to this position within timsControl; confirm that the crosshairs line up with the selected feature.
    NOTE: If the crosshairs do not line up, clear the teaching and repeat the teaching process to improve accuracy.
12. Save the file.
13. Check signal by firing the laser at a few locations within the tissue. Click on a location in the image using the Position Sample Carrier button as in Step 6.11 for checking the teaching. Confirm that the spectrum has abundant peptide peaks with dominant peaks typically at m/z 944, 1,105, and 1,198. Adjust laser fluence as needed for optimal signal.
14. Save the timsControl method.
15. Start the imaging acquisition by clicking the Start AutoXecute Run button.

7. Histological staining

NOTE: There are several methods for hematoxylin and eosin available. Presented here is a modified Carazzi Method frequently used in this lab.

1. Remove the CHCA matrix by submerging the slide in 100% ethanol for 1 min.
2. Stain the slide as follows: 95% ethanol for 30 s; 70% ethanol for 30 s; Water for 30 s; Hematoxylin for 2 min; Water for 30 s; 70% ethanol for 30 s; 95% ethanol for 30 s; Eosin for 1 min; 95% ethanol for 30 s; 100% ethanol for 30 s; Xylene for 2.5 min; Coverslip with mounting medium.
3. Allow the slide to dry for >1 h and remove any bubbles by gently pressing on the coverslip.
4. Digitize with a digital slide scanner.

8. Data visualization

NOTE: There is extensive analysis that can be done with SCiLS Lab that is beyond the scope of this protocol. Here, we will just describe basic file creation, data visualization, and searching of peaks against databases for putative identification.

1. Launch SCiLS Lab 2025b and choose New.
2. Select the datafile(s) to be loaded into the software.
   NOTE: With SCiLS Lab Pro, multiple datafiles that were collected with the same parameters can be loaded together for normalization to each other and simultaneously visualized.
3. Arrange the samples in the preferred layout. Move the samples by left-clicking and dragging. Rotate the samples by right-clicking and dragging. Select Next.
4. In the Mass Axis Settings window, make any necessary adjustments. The default values generally work well for timsTOF data. Click Next.
5. On the Feature Finding window, set appropriate values for automated peak picking during import. Click Next.
   NOTE: This does not work well for metabolite data from FFPE tissue and should be skipped.
6. Import settings will be visualized in the Import Summary window. Click Import.
7. Provide a file name and a location to save the SCiLS file.
   NOTE: Depending on the size of the file, it may take a few min to several hours to import.
8. When completed, a summary window will be visible. Click Open Imported Dataset to interact with the SCiLS file.
9. In the top right of the software, click the dropdown beside Normalization and select Root Mean Square.
   NOTE: This is the most appropriate method for timsTOF data given the sparsity of the imported centroid data and will provide the best visualization.
10. Under File | File Properties, set the interval width to an appropriate value, typically 10 - 15 ppm for timsTOF data. Confirm the value is appropriate by evaluating if it spans the width of a peak in the average spectrum.
11. Zoom in on the average spectrum at the bottom of the software, either by clicking and dragging or rolling the mouse wheel forward.
12. Select a peak to visualize by positioning the crosshairs at the apex and left-clicking the mouse wheel.
13. Add ions of interest to the active Ion Images list by pressing control + space bar.
14. Save the curated feature list by clicking the Save as Feature List button at the bottom of the ion images window on the right side of the software.
15. Open a saved feature list by clicking on File | Feature Table and search using MetaboScape for putative metabolite or glycan IDs, if MetaboScape is installed and configured on the same network. Alternatively, convert the SCiLS files to .imzML files for import into METASPACE (https://metaspace2020.org/) for putative identification.
    NOTE: A full description of the capabilities of SCiLS Lab can be found in the User Manual under the Help menu option.

Results

The completion of this protocol should result in three robust MSI datasets from the same section of tissue. In the visualization of the full spectrum of the metabolite data, the spectrum will be heavily dominated by matrix peaks at m/z 157 and 315. This is normal for FFPE tissue. Many metabolite and fatty acid signals will be observed by zooming in on the m/z ranges <155 and between 250 and 300. Figure 3 shows examples of the full spectrum and zoomed ranges highlighting the complex metabolite spectra obtained from the tissue. Typically, between 700 and 900 metabolite peaks are observed from FFPE tissue sections in negative mode with this sample preparation.

PNGaseF digestion should result in the detection of ~2-300 glycan peaks in the tissue. Peaks originating from glycans in the m/z range 1,000-2,000, should have decimal masses of 0.3-0.7 due to their atomic composition. This decimal mass helps to distinguish glycans from matrix clusters that may be observed with this sample preparation. Figure 4A shows examples of robust glycan signal with abundant peaks, especially m/z 1,127.55, 1,325.58, 1,663.58, and 1,809.64, while Figure 4B shows a spectrum with poor glycan signal. This spectrum is dominated by matrix cluster peaks with decimal masses of 0.0-0.2, likely the result of poor efficiency of the PNGaseF digestion on tissue, possibly a result of incomplete washing of the tissue or the presence of small molecule inhibitors in the tissue.

On tissue, tryptic digestion should result in the detection of 800-1,000 peptide peaks in the average mass spectrum. Figure 5A displays a robust average peptide spectrum from a successful sequential imaging experiment. Hundreds of peptides are detected across the analyzed m/z range with high intensity. In contrast, Figure 5B shows a relatively unsuccessful tryptic digest average spectrum. In this case, fewer peptides are detected in the higher m/z range. This is most likely attributed to errors in spraying, in active enzyme, or incomplete washing of the tissue to remove interfering molecules.

Since all three datasets were collected from the same piece of tissue, they can be easily co-registered using the SCiLS Ion Image Mapper to visualize co- and differential localization of molecules belonging to different classes. Figure 6 shows an example of a glycan (green), a metabolite (taurine, red), and a tryptic peptide (blue), both separately (left) and combined (right). This type of data co-registration enables a deeper understanding of the spatial biology occurring in complex tissue samples.

Figure 1: General outline of FFPE SeqMSI protocol.
After tissue slides are mounted, they are (A) marked for fiducials and deparaffinized to be prepared for optical image scanning. (B) Subsequently, they are sprayed with DAN for metabolite imaging, which is (C) removed before proceeding with PNGaseF and digestion incubation. (D) Slides are then prepared for Glycan imaging in which they are sprayed with CHCA matrix and imaged; (E) the CHCA matrix is subsequently removed before spraying with Trypsin and going through a second digestion incubation. (F) Lastly, slides are prepared for peptide imaging in which they are sprayed with CHCA and imaged. (G) Finally, the matrix is removed and slides are stained with H&E for histology and colocalization analysis. Abbreviations: FFPE = formalin-fixed, paraffin-embedded; SeqMSI = sequential mass spectrometry imaging; MALDI = matrix-assisted laser desorption/ionization; DAN = 1,5-diaminonaphthalene; CHCA = α-cyano-4-hydroxycinnamic acid; H&E = hematoxylin and eosin. Please click here to view a larger version of this figure.

Figure 2: Incubation chamber for PNGaseF digestion.
A piece of absorbent wipe is cut to fit the bottom of a 100 mm diameter Petri dish. Laboratory wipes are folded and rolled to form pillows that the slide can be elevated on. Please click here to view a larger version of this figure.

Figure 3: Example metabolite average mass spectrum.
(A) The full spectrum is dominated by matrix peaks around m/z 157 and clusters between m/z 300 and 350. (B) Zoomed spectrum from m/z 80 to 154 showing the wealth of metabolite peaks detected in this range. (C) Zoomed spectrum from m/z 250 to 300, a range where several free fatty acids are detected in tissue. Please click here to view a larger version of this figure.

Figure 4: Example glycan average mass spectra.
(A) An example spectrum with abundant glycan signals. These are evidenced by the decimal masses between 0.5 and 0.7. This is an example of a good PNGase F digest spectrum. (B) An example spectrum of poor glycan signal. This spectrum is dominated by peaks originating from the MALDI matrix, evidenced by decimal masses between 0.0 and 0.1. This is an example of an unsuccessful PNGase F digest spectrum. Please click here to view a larger version of this figure.

Figure 5: Example of peptide average mass spectra.
(A) An example of a spectrum with abundant peptide signal. The spectrum is robust with high-intensity peaks. (B) An example of a spectrum with poor peptide signal. The overall spectral intensity is considerably lower than A with less total peaks. Please click here to view a larger version of this figure.

Figure 6: Example MSI data from two ovarian cancer samples. (A) A metabolite (taurine) image shown in red; (B) An N-linked glycan shown in green. The cartoon indicates the composition of the sugars in the glycan, but the linkages are not known. (C) A tryptic peptide from Histone H4 shown in blue. (D) Composite image of all three ion images with insets of the H&E images. The dashed white line divides the tumor (above) and stroma (below). Abbreviations: MSI = mass spectrometry imaging; H&E = hematoxylin and eosin. Please click here to view a larger version of this figure.

Supplemental Figure S1: Instrument parameters for metabolite data acquisition. (A) m/z range, number of laser shots, and laser fluence. (B) Laser scan range and adjustment. (C) Instrument tuning parameters. Please click here to download this File.

Supplemental Figure S2: Instrument parameters for glycan data acquisition. (A) m/z range, number of laser shots, and laser fluence. (B) Laser scan range and adjustment. (C) Instrument tuning parameters. Please click here to download this File.

Supplemental Figure S3: Instrument parameters for peptide data acquisition. (A) m/z range, number of laser shots, and laser fluence. (B) Laser scan range and adjustment. (C) Instrument tuning parameters. Please click here to download this File.

Discussion

The collection of sequential MSI data from a single tissue section requires careful attention to detail. There are a few steps that are absolutely crucial to achieving high-quality data. First, care should be taken if more than one slide is prepared at the same time. When placing the slides into, or moving between Coplin jars, be careful not to place two slides into the same position in the jar. This will result in inadequate solvent exposure of the tissue surface, or one slide can scrape the tissue off the other slide. Second, it is critical that the instrument be allowed to equilibrate for a minimum of 25 min when changing polarity for analysis. Over this time, the measured m/z values of molecules will drift slightly. This can be observed when spraying tune mix over this equilibration time, as the masses of the peaks will change slightly. If an image is started before the system is equilibrated, there will be mass drift of the detected peaks in the first part of the image, resulting in incorrect calibration if the calibration was performed during this settling period.

In addition, keeping the top of the Petri dish warm during the PNGaseF incubation is incredibly important. During this 2 h incubation in high humidity, it is possible for water to condense on the lid of the dish and rain down onto the tissue on the slide. If this happens, there will be noticeable areas of the tissue where the matrix will not crystallize, and no glycan signal will be observed. Images from samples with this artifact will have a pockmarked appearance. Furthermore, matrices should be filtered before loading into the sample loop of the sprayer. This helps to prevent clogging of the system. A disposable syringe filter works well for this purpose. Finally, it should be noted that when initially removing the wax from the section, the tissue will become nearly transparent after submersion in xylene. Do not be alarmed by this. It will "reappear" after it dries.

It is important to consider that the methods described herein may not be optimal for all tissue types. They have been shown to work well for most cancers, including breast, liver, and ovarian^16,17,18, but some tissues, such as the brain with much higher lipid content, may not work well with these methods. It may be necessary to perform additional washing steps, such as a Carnoy's fluid wash⁷, if lipids still remain in the tissue after fixation. Longer incubation times for enzymes may be needed if incomplete digestion is observed. In some cases, additional passes of matrix spray may be necessary to achieve complete matrix coverage of the sample. Or fewer passes of matrix may be needed if the imaging experiments are to be carried out on a mass spectrometer with an atmospheric pressure MALDI source.

Traditionally, only a single MSI analysis has been carried out on a tissue section, or serial sections are used for different analyte types. However, this can lead to challenges when tissue is limited or when the objective is to correlate results from multiple spatial omics studies, where the goal is to analyze the same cells across multiple platforms^18,20. The ability to collect multiple MSI datasets from the same tissue section helps to reduce the amount of material needed for spatial omics studies, as well as enable more facile registration of datasets; the MSI datasets are inherently registered to each other since they are collected from the same tissue section. The protocol presented here is not limited to only cancer specimens. We, and we hope others as well, are applying these workflows to a variety of tissue types and diseases. This type of workflow is greatly advancing our ability to perform meaningful multi-omics analysis for a deeper understanding of disease biology.

As with any method, there are limitations that must be considered. The extensive sample preparation of sequential imaging of the same tissue section may result in tissue loss from the slide during extended steps of solvent exposure, especially during repeat antigen retrieval of the section. If areas (or all) of the tissue are lost during the process, the datasets may not be able to be successfully co-registered to each other. Despite these challenges, we believe that this workflow will be invaluable for future clinical studies where tissue available for research is limited and/or there is a need to collect multiple datasets from the exact same cells. We also envision that future technology advancements will enable more datasets to be collected from the same tissue section, further enhancing our ability to obtain and interpret multi-omics data from cancer and other diseased tissue.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1,5-diaminonaphthalene	Fisher Scientific	D010125G	MALDI Matrix for metabolite imaging
Acetonitrile	Fisher Scientific	A955-4	LC-MS grade solvent used for matrix preparation
α-cyano-4-hydroxycinnamic acid	Sigma-Aldrich	70990-1G-F	MALDI matrix for glycan and peptide imaging
Ammonium Bicarbonate	Fisher Scientific	A643-500	Buffer for enzymes
Ammonium Phosphate	Sigma-Aldrich	467782-50G	Additive to reduce matrix clusters during imaging
Decloaking Chamber NxGen	Biocare Medical	N/A	Used for antigen retrieval of tissue
Ethanol	Fisher Scientific	04-355-223	LC-MS grade solvent used for matrix preparation and staining
M5 Robotic Reagent Sprayer	HTX Imaging	N/A	Used for application of enzymes and matrices to tissue
Methanol	Fisher Scientific	A456-4	LC-MS grade solvent for making red phosphorus suspension
MS Grade Trypsin	Fisher Scientific	PI90058	Enzyme for protein digestion
MTP Slide Adapter II	Bruker Daltonics	8235380	Adapter to insert micrscope slides into mass spectrometer
NanoZoomer SQ Digital Slide Scanner	Hamamatsu Corp	N/A	Used for generating digital microscopy images of stained tissue
Perfection V600 Flatbed Scanner	Epson	N/A	Used for generating optical image of the slide for MSI data collection
Petri-seal	Fisher Scientific	50-212-518	For sealing petri dish during enzymatic digestion
PNGaseF	Bulldog Bio	NZPP550LY	Enzyme for cleavage of N-linked glycans from proteins
Red Phosphorus	Sigma-Aldrich	04004-250G	MALDI calibrant for both positive and negative ion mode
SCiLS Lab (2025b)	Bruker Daltonics	N/A	Software for MSI data visualization
timsTOF fleX QTOF Mass Spectrometer	Bruker Daltonics	N/A	Used for mass spectrometry data collection
Trifluoracetic acid	Fisher Scientific	85183	Matrix additive to decrease pH for positive ion mode imaging
Tris Base	Fisher Scientific	BP152-500	Buffer for antigen retrieval
Water	Fisher Scientific	W64	LC-MS grade solvent used for matrices, enzymes, and staining
WypAll X60	Fisher Scientific	19-413-113	Absorbent wipe for humidified enzyme incubation
Xylene	Fisher Scientific	X3P-1GAL	Clearing agent for staining