This Tutorial focuses on the use of secondary ion mass spectrometry for the analysis of cellular and tissue samples. The Tutorial aims to cover the considerations in sample preparation analytical set up and some specific aspects of data interpretation associated with such analysis.

Secondary ion mass spectrometry (SIMS), while often viewed by the mainstream mass spectrometry community as an inorganic analysis tool due to its unique ability to provide detailed compositional information from hard or composite materials, also offers benefits for the analysis of biological samples such as cell and tissue specimen.1 Compared with conventional optical microscopy-based imaging methods, the technique can operate “label free” and hence can be utilized as a discovery technique. Further, compared with other imaging mass spectrometry approaches, SIMS provides routine high spatial resolution imaging with minimal alteration of the sample chemistry (e.g., no additional matrix is required) and with the unique feature of high surface sensitivity. This Tutorial will focus predominantly on molecular mass spectrometry using SIMS, as opposed to targeted elemental analysis. Hence, we are considering what has historically been referred to as static SIMS and, due to the almost ubiquitous use of the time-of-flight analyzer for such analysis since the 1980s, sometimes known synonymously as ToF-SIMS. However, in the modern era with beams that can provide molecular information beyond the static limit, and the availability of new mass analyzers, things are less clear cut and it is perhaps better to use the general label of molecular SIMS.

The desorption process via focused ion beam bombardment (sputtering) enables the acquisition of mass spectra from specific coordinates or areas on a biological specimen. Hence, spatial chemical information can be delivered through the creation of chemical maps for different ions of interest. Notably, in the last 15 years, SIMS has become a versatile tool to detect surface molecules up to approximately 1000 Da.2 

In the SIMS experiment, the nature of the primary ion beam can have a profound effect on the analytical capabilities influencing spatial resolution, depth resolution, and the type/relative abundance of secondary ions that are generated. There is a general trend that the ability to detect intact molecules comes with a degradation in achievable spot size as illustrated in Fig. 1. Atomic ion beams typically provide the best opportunities for very high resolution imaging (>100 nm spot sizes) and are the default for inorganic analysis or the detection of elements or small fragments but with limited production of intact molecular type secondary ions. Au3 and similarly Bi3 produce nonlinear yield enhancements of higher mass signals versus atomic ion beams while maintaining a relatively easy to focus ion source. C60 increases high mass performance further while gas cluster ion beams (GCIBs), such as (Ar)n+, and recently (H2O)n+ offer further enhancements.3–10 However, the increase in intact molecular ion signal comes with the penalty of reduced spatial resolution. For depth profile or 3D imaging experiments, conventional ToF-SIMS instruments are normally operated in a dual beam mode where one beam is used for analysis (typically Bi3+) while a different beam is used to sputter the material between analysis cycles (for organic samples this is now typically a GCIB). Mass spectrometers where the mass analyzer is decoupled from secondary ion formation (e.g., The Ionoptika J105 or the Iontof Hybrid SIMS) offer the opportunity to perform single beam depth profiling/3D imaging.11,12

FIG. 1.

Illustration of the tendency toward larger ion beams for the detection of intact molecular signals from biological compounds while sacrificing lateral resolution (not to scale). Arrows indicate improvements (i.e. reduced beam diameter).

FIG. 1.

Illustration of the tendency toward larger ion beams for the detection of intact molecular signals from biological compounds while sacrificing lateral resolution (not to scale). Arrows indicate improvements (i.e. reduced beam diameter).

Close modal

While the availability of different primary ion beams will as depend on the specific SIMS instruments, the optimal pixel size may depend on factors beyond the ability to focus the primary ion beam. Smaller pixels may mean more pixels in an image and hence extended analysis time, and ultimately the amount of signal per pixel (or voxel for 3D images) of the target analyte(s) may be the deciding factor for useful resolution.

Cellular analysis is often made by bulk methods which reveal average features of a given cell type. On the other hand, single-cell analysis can capture individual cell differences within a cell population. Especially for early diagnosis for heterogeneous diseases like cancer where discriminating these particular changes can be essential.13 The capacity of a cell to generate a scale of responses due to environmental signals could result in heterogeneous biochemical profiles in a given cell population.14,15

Revealing the localized function of molecules on the cell plasma membrane and inside the cell depends strongly on the methodology of the analysis. For this purpose, fluorescent tags have been used widely. However, it can be challenging to obtain sufficient labels across the wide range of available cellular components while also accounting for potential alterations caused by the insertion of tags into the cell structure. At this point, imaging mass spectrometry provides label-free imaging of cellular components to elucidate their natural behaviors. Still, in both SIMS and matrix assisted laser desorption ionization (MALDI) MS analyses, it is crucial to implement the correct sample preparation techniques to maintain the structural integrity of cells and prevent molecular migration prior to analysis.16 To date, for cellular analysis, monoatomic (e.g., Bi+ and Au+), cluster and polyatomic (C60+, Bi3+ and Aun+) primary ion sources, and gas cluster beams [Arn+, (CO2)n+, and (H2O)n+] have all been used to create chemical images of cells.17,18 While monoatomic sources can target a smaller area providing higher resolution and so have been a favorite for subcellular imaging, polyatomic sources can yield more secondary ions due to softer ionization ability.17,18 For example, Sparvero et. al. recently revealed ferroptotic death signals by imaging peroxidized PE species in H9c2 cells using GCIB.19 

The size of single cells can vary from few micrometers (red blood cells, 8 μm) to about 1 mm (Xenopus laevis oocyte).20,21 Imaging of smaller cells is typically the most challenging, emphasizing the need to optimize resolution versus signal intensity. Additionally, the cell passage number should be considered along with whether there will be single-cell analysis or multiple cellular analyses where average signals coming from all cells in a region of interest/analysis area. Depending on the ion beam or ion dose (the number of ions per unit area, typically stated as ions/cm2, and also referred to as dose density and fluence), signals may not be enough to detect molecular information (e.g., less ionized lipid species) from each cell. In those cases, to increase sensitivity, summing up individual cell signals can enrich the signal intensity and enable the researcher to discriminate the signals from noise. Ideally, this should only be considered when the cell population show similar signals as cell heterogeneity or subpopulations inside the given cell type is neglected. The use of a low (typically < 1012 ions/cm2) primary ion dose (static SIMS) enables analysis of intact lipids from the very surface of the sample.22 However, sometimes the chemistry of interest lies below the surface of the cell.

Since analysis with SIMS requires a vacuum environment, sample preparation is a critical step prior to analysis. Receiving the true spatial chemical information depends on the retained structure of the sample. To meet this requirement, there have been several techniques used to study the cell specimen in a high vacuum environment based on approaches for more routine cellular microscopy including electron microscopy. Besides chemical fixation methods, several cryogenic preparation techniques have been used to minimize structural deformation between sampling time to analysis.2 

Many methods have been used up to now are transferred from electron microscopy (EM) sample preparation techniques, either by applying modification to the existing method or by combining several methods to implement them to the SIMS imaging. The overall workflow is illustrated in Fig. 2.

FIG. 2.

Workflow for cellular analysis with SIMS. PDL/PLL refer to poly-D-lysine and poly-L-lysine, respectively.

FIG. 2.

Workflow for cellular analysis with SIMS. PDL/PLL refer to poly-D-lysine and poly-L-lysine, respectively.

Close modal

1. Washing

Typically, the first step of preparation for cells is the removal of exogenous salts and other chemicals used in the culturing process. Phosphate buffered saline is often used to remove cell media components. However, nonvolatile buffers create accumulation of salts on the surface creating a barrier on top of the cells and modifying the ionization probability of molecules. Hence, a gentle washing procedure is essential to remove the salts and possible contaminants prior to freezing is needed. Depending on the durability of the cell type and the fixation method, the washing solution and washing time can differ. Since chemical fixation protects cells to some degree against osmotic pressure, various washing solutions can be applied to those samples such as de-ionized water, ammonium acetate, etc.23,24 Specific cells like oocytes are relatively durable to osmotic pressure changes and de-ionized water has also been used to clear undesirable molecules from the surface.25 Alternatively, Malm et al. used ammonium formate, a volatile salt, as a washing solution on the surface for more delicate cell types prior to the freezing process. They observed an effective removal of salts coming from the buffer solutions while remaining the sample quality.26 Ammonium acetate has also been used to clean the residues from the surface.27,28 0.15 M ammonium formate is the most common washing solution for mammalian cell samples.

2. Chemical fixation

Fixation can be performed either chemically or physically or using a combination of both. Commonly used chemical fixation methods are glutaraldehyde, osmium tetraoxide (OsO4), uranyl acetate, etc.24,26,29 A combination of glutaraldehyde and uranyl acetate fixation has been recently used to image organelle level fatty acid distribution without disrupting ionization. However, the method lacks information from intact phospholipids, sphingolipids, and cholesterol.24 Following chemical fixation, the challenge of introducing a “wet” sample into the vacuum remains. Drying techniques also have been used such as alcohol substitution and/or freeze drying.26 

3. Cryogenic fixation

Cryofixation has generally been claimed as advantageous to maintain a more intact membrane, a better distributed diffusible ion profile, and higher signals from secondary ions compared to chemical fixation.26 Cryofixation has been generally made by plunge freezing/dipping the cell coated substrates into cooled liquid solutions such as liquid ethane, propane, or isopentane.26,28 The use of liquid propane results in rapid freezing without the formation of ice crystals by avoiding the Leidenfrost effect associated with liquid nitrogen freezing.

Frozen samples are either freeze dried by gradually warming the sample under vacuum or analyzed frozen hydrated. Schaepe and co-workers evaluated different sample fixation approaches for lipid detection from human mesenchymal stem cells and found glutaraldehyde and/or paraformaldehyde fixation and subsequent plunge freezing (followed by freeze drying) provided the most reproducible results for lipid detection.31 

While specialized freeze-drying instruments, developed for SEM sample preparation can be purchased, this is typically performed in home-built systems, repurposed vacuum desiccators or in the preparation/entry chamber of the SIMS instruments.

4. Cryogenic analysis

If suitably equipped, the SIMS analysis can be performed at cryogenic temperatures and the need to remove water from the sample is negated. A challenge arises during transfer of the sample to the SIMS instrument where sample frosting must be prevented. Transferring the sample under liquid nitrogen into a purged glove box surrounding the sample insertion port is the best approach. Cells may still be partially buried beneath ice and this can be removed by careful warming of the sample within the SIMS instrument (sublimation of water in a SIMS instrument occurs around 165 K). Alternatively, the ice can be etched off using an ion beam—a GCIB being the recommended beam in order to preserve the underlying cell chemistry. A third approach is to use in vacuum freeze fracture to present a fresh cellular surface for analysis.

5. Freeze fracture

The freeze fracture approach has been utilized for EM and then adapted for dynamic SIMS by Chandra and Morrison and for molecular SIMS by Roddy and co-workers.30–32 Another variation of this method was implemented by Kulp et al. using sandwich freeze fracture on cancer cells followed by freeze drying to use the samples in ambient temperature.33 Since the late 1990s, Winograd and Ewing have published a series of papers using freeze fracture and frozen hydrated analysis conditions to preserve cellular integrity including studies often incorporating additional techniques such as fluorescent labeling of the membranes in order to identify different fractured states of the cells.34 During the sample handling development of the J105, the MousetrapTM in vacuo freeze fracture device was developed and this led to the construction of a similar type of device for use on Iontof instruments called the “Bugtrap.”35,36

While cryofixation followed by frozen hydrated analysis is considered the gold standard for preserving sample integrity in both two and three dimensions, it also provides additional advantages with respect to ionization as samples in ice matrices have been shown to yield higher molecular signal levels.39 Tian et al. recently compared chemical fixation (fixed with 2% glutaraldehyde for 15 min, then air-dried) and frozen hydrated analysis on single Hela cells and observed lipid and metabolite migration regardless of the beam type.37 However, the caveat is that reproducibility in frozen hydrated analysis (and additionally freeze fracture) is more difficult to achieve. Frozen analysis is also typically more labor intensive.

6. Sample modification

While preserving sample integrity is important, some adulteration of the sample can be advantageous for specific applications. Any sample modification must be matched to the other sample preparation steps and analysis conditions. Different surface modification or coating techniques on cellular samples have been utilized by research groups in order to increase ionization efficiency and reduce signal suppression and matrix effects (Table I).16,38 Several metals such as silver and gold have been used to enhance ionization in SIMS analysis. For example, Malmberg and Sjövall introduced a silver imprinting system that greatly increased cholesterol detection by the formation of silver adduct ions.42,43 MALDI matrices have also been applied for SIMS analysis resulting in elevated molecular ion yields. Graphene quantum dots have been used on cancer cells and enhanced lipid signals significantly.38 It was proposed that the graphene quantum dots’ capability of conducting the electrons facilitates energy transmission from primary ion beam to target molecules, and improves ionization.

TABLE I.

Examples of different SIMS cell imaging analysis from the literature including various cell types, sample preparation, beam choices, ion doses, and pixel sizes.

Cell typeSubstrate/substrate treatment/washing bufferCell fixation and additional modificationSurface modification/isotope labelingBeamIon dose/target currentPixel sizeReference
Breast cancer cells MCF-7 Silicon wafers/amiodarone/ammonium formate Cryofixation: freeze-dry Plunge freezing with liquid nitrogen, dried under vacuum freeze drier (−40 C°, 6h) Graphene quantum dots /− 30 kV Bi3+ Depth profiling: 10 kV Ar2000+ GCIB 0.2 pA pulsed current ∼200 nm 38  
Oligodendrocytes Silicon shards/poly(L-lysine) and collagen/DI water Cryofixation: plunge freeze with liquid ethane (stored under liquid nitrogen) /freeze etching −/− 15 kV In+ and 40 kV C60+ 1 nA and 20 pA 1 μ2  
Hela cells Si wafer/ammonia acetate Cryofixation: plunge freezing with liq. ethane, carried with N2 gas. *Precooled instrument) In vivo isotope incorporation [15N]Ser/[13C370 kV (CO2)n+ (n > 10 000) GCIB 1.7 × 1013 ions/cm2 1 μ28  
Hela cells Si wafer/glutaraldehyde fix; plunge freeze/ammonium formate wash Chemical fixation ; plunge freeze/air dry; LN2 −/− 70 kV (CO2)11.5k+ and (H2O)28k+ GCIBs 5.63 × 1012 ions/cm2 1 μ37  
Breast cancer cells MCF-7 Si wafer/ammonium formate wash Cryofixation: frozen in liquid nitrogen (Stored in – 80 C°) Freeze dried inside ToF-SIMS instrument −/− 40 kV (CO2)7k+, H2O)18k+, (CO2)6k+ GCIBs 1.02 × 1013 ions/cm2 4.7 μ18  
H9c2 cells (2000 cells per ml) Si wafer Ammonium formate Cryofixation: plunge frozen with liquid ethane (liquid N2 cooled) Transferred to precooled (LN2) sample holder −/− 70 kV (H2O)30k+ GCIB 1.1 × 1012 ions/cm2 3.1 μ19  
PC12 cells Si shards (freeze fracture device 37 C°)/washed with HEPES Freeze fracture device fast frozen in liquid propane at −185 C°) Mounted under liquid nitrogen —/Cells incubated with deuterated phospholipids 25 kV Bi3+ 4 × 1012 ions/cm2 4 μ39  
PC12 -cells Poly-D-lysine coated silicon wafer Fast frozen with precooled propane and freeze dried —/— 25 kV Bi3+ 0.3 pA pulsed current > 1 × 1013 ions/com2 Not stated 40  
Chromaffin cells Poly-D-lysine coated ITO glass Frozen with liquid nitrogen cooled 2-methylbutane and freeze dried under vacuum —/— 40 kV (CO2)6k+ GCIB 2.56 × 1012 ions/cm2 0.39 and 2 μm per pixel 41  
Cell typeSubstrate/substrate treatment/washing bufferCell fixation and additional modificationSurface modification/isotope labelingBeamIon dose/target currentPixel sizeReference
Breast cancer cells MCF-7 Silicon wafers/amiodarone/ammonium formate Cryofixation: freeze-dry Plunge freezing with liquid nitrogen, dried under vacuum freeze drier (−40 C°, 6h) Graphene quantum dots /− 30 kV Bi3+ Depth profiling: 10 kV Ar2000+ GCIB 0.2 pA pulsed current ∼200 nm 38  
Oligodendrocytes Silicon shards/poly(L-lysine) and collagen/DI water Cryofixation: plunge freeze with liquid ethane (stored under liquid nitrogen) /freeze etching −/− 15 kV In+ and 40 kV C60+ 1 nA and 20 pA 1 μ2  
Hela cells Si wafer/ammonia acetate Cryofixation: plunge freezing with liq. ethane, carried with N2 gas. *Precooled instrument) In vivo isotope incorporation [15N]Ser/[13C370 kV (CO2)n+ (n > 10 000) GCIB 1.7 × 1013 ions/cm2 1 μ28  
Hela cells Si wafer/glutaraldehyde fix; plunge freeze/ammonium formate wash Chemical fixation ; plunge freeze/air dry; LN2 −/− 70 kV (CO2)11.5k+ and (H2O)28k+ GCIBs 5.63 × 1012 ions/cm2 1 μ37  
Breast cancer cells MCF-7 Si wafer/ammonium formate wash Cryofixation: frozen in liquid nitrogen (Stored in – 80 C°) Freeze dried inside ToF-SIMS instrument −/− 40 kV (CO2)7k+, H2O)18k+, (CO2)6k+ GCIBs 1.02 × 1013 ions/cm2 4.7 μ18  
H9c2 cells (2000 cells per ml) Si wafer Ammonium formate Cryofixation: plunge frozen with liquid ethane (liquid N2 cooled) Transferred to precooled (LN2) sample holder −/− 70 kV (H2O)30k+ GCIB 1.1 × 1012 ions/cm2 3.1 μ19  
PC12 cells Si shards (freeze fracture device 37 C°)/washed with HEPES Freeze fracture device fast frozen in liquid propane at −185 C°) Mounted under liquid nitrogen —/Cells incubated with deuterated phospholipids 25 kV Bi3+ 4 × 1012 ions/cm2 4 μ39  
PC12 -cells Poly-D-lysine coated silicon wafer Fast frozen with precooled propane and freeze dried —/— 25 kV Bi3+ 0.3 pA pulsed current > 1 × 1013 ions/com2 Not stated 40  
Chromaffin cells Poly-D-lysine coated ITO glass Frozen with liquid nitrogen cooled 2-methylbutane and freeze dried under vacuum —/— 40 kV (CO2)6k+ GCIB 2.56 × 1012 ions/cm2 0.39 and 2 μm per pixel 41  

Sample modification can also be employed to help target specific chemistry or biological processes. Isotopic labeling has been used for several cell imaging studies. Vihdi et al. used isotopically labeled nutrients in the cell media to explore the purine synthesis biochemical pathway.28 Lanekoff et al. used isotopically labeled phospholipids fed to PC12 cells and were able to measure the incorporation of lipids by their deuterated fragment ions.39 Tyler et al. have demonstrated the ability to identify 15N treated cells in mixed cell populations.44 

The sputtering process in SIMS offers unique opportunities for 3D cellular imaging. Every analysis removes some materials from the sample and continued analysis gradually erodes the sample completely. The greatest challenge for a long time was the maintenance of molecular signals during the 3D analysis as subsurface damage (in the form of molecular fragmentation) accumulation occurs under bombardment with atomic and small cluster ion beams. The advent of C60, and particularly GCIBs, has led to a significant (sometimes almost complete) removal of subsurface damage.

Instruments with decoupled sputtering and mass analysis (e.g., the J105 or Hybrid SIMS systems including Orbitrap-SIMS instruments) can be operated in single beam depth profiling mode where one ion beam is used to image the sample many times, at relatively high dose, to create a continuous stack of images that can be reconstructed into a 3D data set. The pixels now become voxels. C60 or GCIB's are typically used.

Conventional ToF-SIMS instruments typically employ a dual beam approach where one beam is pulsed for analysis and another operated continuously for etching the sample. This decreases the time required to “dig” through a cell. The ratio of the analysis beam dose to the etching beam dose needs to be considered for successful 3D molecular imaging.45 Either the analysis beam dose is kept low enough during the entire experiment that no significant damage becomes evident or subsurface damage from the analysis beam needs to be etched away by the GCIB. A 20 kV Au+ ion will cause damage down to approximately 25 nm below the surface of an organic sample.46 

While the dual beam approach allows combinations of different ion beams to be used and can reduce 3D analysis time (high current broadly focused etching beams can be used to increase primary ion flux), the single beam approach offers the highest potential sensitivity as all the sputtered material is fed to the mass spectrometer. In the dual beam approach, the material removed by the continuous etching beam is discarded.

The 3D image data are recorded as a stack of images that can be viewed as a “cube” with the initial image as the upper face of the cube. However, unless the sample was flat before the analysis, this will not represent a lifelike view of the sample. Data reshaping to match the initial topography of the sample can be employed. For cellular analysis, this is typically done by completely eroding the cell and identifying the point when the substrate was reached for each pixel. A threshold of the indium signal (for ITO coated slides), for example, can be used or multivariate analysis such as PCA can be employed to find this point [these approaches were used in the examples in Figs. 3(a)3(f)].35,47,48 The data are then reshaped to make the substrate flat. The caveat here is the assumption of an equal sputter rate through all parts of the cell. AFM measurements prior to 3D imaging have illustrated that this is a reasonable assumption with polyatomic ion beam etching as shown demonstrated by Robinson et al. [Figs. 3(e) and 3(f)].47 However, the presence of regions in the sample with grossly different sputter rates presents a challenge as can be seen by the red signal in Fig. 3(d), originating from accumulated TiO2 nanoparticles, seemingly penetrating the silicon substrate while the organic signals are correctly reconstructed.49 Additionally, in some cases, image shift correction may need to be employed to compensate for the angle of incidence of the erosion beam and changing sample height.

FIG. 3.

Examples of cellular analysis using SIMS. Freeze fractured frozen hydrated Hela cells (red m/z 136, adenine, green m/z 184 PC-lipid head group) (a). Reproduced with permission from Lanekoff et al., Anal. Chem. 25, 925 (2011).35 Copyright 2011, John Wiley & Sons, Ltd. Single-cell organism tetrahymena pyriformis cryogenically preserved for analysis with internal structure preserved (b) vs a dried sample cell where the contents (green and red signals) have leaked out (c). 3D reconstruction of tetrahymena showing incorrect apparent penetration of TiO2 NPs into the silicon substrate due to the disparity in sputter rate (d). Reproduced with permission from Angerer and Fletcher, Surf. Interface Anal. 46, 198 (2014).49 Copyright 2014, John Wiley & Sons Ltd. Comparison of reconstructed SIMS data with AFM measurement of cell morphology showing that reconstruction is reasonably accurate in the absence of inorganic inclusions (e) and (f). Reproduced with permission from Robinson et al., Anal. Chem. 84, 4880 (2012).47 Copyright 2012, American Chemical Society. GCIB images of HT22 cells highlighting the move to intact 3D molecular imaging. Total ion image of the first layer showing the outline of the cells (white arrow pointing to a cell). (g) Total ion image of the first layer showing the outline of the cells (white arrow pointing to a cell). (h) and (i) m/z 1404.0 peak identified as CL(68:2) and (c) the m/z 885.5 peak identified as PI (38:4) respectively. (j)–(l) Panels show overlay images at different depths in to the cells. Green, PI (38:4); blue, deoxyribose phosphate (m/z 257.0); magenta, CL(68:2). Images were captured at 1 μm per pixel. Reproduced with permission from Tian et al., Angew. Chem. 131, 3188 (2019).50 Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA.

FIG. 3.

Examples of cellular analysis using SIMS. Freeze fractured frozen hydrated Hela cells (red m/z 136, adenine, green m/z 184 PC-lipid head group) (a). Reproduced with permission from Lanekoff et al., Anal. Chem. 25, 925 (2011).35 Copyright 2011, John Wiley & Sons, Ltd. Single-cell organism tetrahymena pyriformis cryogenically preserved for analysis with internal structure preserved (b) vs a dried sample cell where the contents (green and red signals) have leaked out (c). 3D reconstruction of tetrahymena showing incorrect apparent penetration of TiO2 NPs into the silicon substrate due to the disparity in sputter rate (d). Reproduced with permission from Angerer and Fletcher, Surf. Interface Anal. 46, 198 (2014).49 Copyright 2014, John Wiley & Sons Ltd. Comparison of reconstructed SIMS data with AFM measurement of cell morphology showing that reconstruction is reasonably accurate in the absence of inorganic inclusions (e) and (f). Reproduced with permission from Robinson et al., Anal. Chem. 84, 4880 (2012).47 Copyright 2012, American Chemical Society. GCIB images of HT22 cells highlighting the move to intact 3D molecular imaging. Total ion image of the first layer showing the outline of the cells (white arrow pointing to a cell). (g) Total ion image of the first layer showing the outline of the cells (white arrow pointing to a cell). (h) and (i) m/z 1404.0 peak identified as CL(68:2) and (c) the m/z 885.5 peak identified as PI (38:4) respectively. (j)–(l) Panels show overlay images at different depths in to the cells. Green, PI (38:4); blue, deoxyribose phosphate (m/z 257.0); magenta, CL(68:2). Images were captured at 1 μm per pixel. Reproduced with permission from Tian et al., Angew. Chem. 131, 3188 (2019).50 Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA.

Close modal

In order to increase visualization capacity or to validate/correlate the SIMS experiments, complementary studies can be utilized. Different microscopic techniques such as atomic force microscopy (AFM), optical microscopy, fluorescence microscopy, scanning electron microscopy (SEM), confocal microscopy, confocal Raman microscopy have been employed for several reasons such as topography analysis, localization validation, complementing signals from lipids with other molecular classes including carbohydrates, proteins, etc.21,51–53 Lanni et al. used confocal Raman microscopy besides detecting lipids with SIMS (C60) to analyze further molecular classes such as proteins carbohydrates and quinolone signaling molecules in pseudomonas derived biofilms.51 Fletcher et al. used fluorescence microscopy to identify transfected subpopulations of cells on conductively coated finder grid marked microscope slides prior to GCIB-SIMS analysis.57 

Apart from strategies to combine microscopic techniques, there are studies that have correlated SIMS with electrochemistry; however, so far these have used cells cultured in parallel for correlated analysis and not the same individual cells. Ranjbari et al. combined ToF-SIMS and amperometry to study the effects of modafinil on PC12 cells; the results suggested that the changes in cell membrane lipid profile affected dynamics of exocytosis.40 Using chromaffin cells, ToF-SIMS imaging and amperometry were also recently correlated to show dose dependent effects on membrane lipid changes and neuroexocytotic events.41 

Tissue imaging using SIMS is typically performed on thin slices of tissue. These tissue sections are normally in the range of 5–20 μm thick. Thinner samples often present less of a challenge in terms of charge compensation, but the ability to cut the tissue, this thin can be quite sample dependent. Factors affecting the minimum thickness that can be achieved include sample size, water content, and operator skill/experience. While pathologists often work with formalin fixed, paraffin embedded tissue blocks these samples are not recommended for SIMS analysis and instead fresh frozen tissue is best. A schematic of the tissue analysis procedure is shown in Fig. 4.

FIG. 4.

Workflow for tissue analysis by SIMS.

FIG. 4.

Workflow for tissue analysis by SIMS.

Close modal

Samples are typically sliced using a cryo-microtome at approximately −20 °C although this can be optimized for different samples. For best results, the cryo-microtome should be purged using a suitable inert gas (Ar and N2) to avoid water condensation on the tissue sample of substrate.

While any substrate can be used, a conducting substrate is preferred to minimize charging effects. Steel or silicon wafer substrates have been used in many studies; however, the most common substrate for tissue imaging is ITO coated glass. This glass can be supplied as “full size” microscope slides or in smaller pieces and offers the advantage of being optically transparent facilitating the incorporation of optical microscopy-based steps into the imaging workflow. For example, tissue that has been subjected to SIMS imaging can be removed from the instrument and stained for conventional histology.54 

The tissue slices are most commonly “thaw mounted” onto the substrate by gently warming the reverse of the substrate using ones finger. This step is quite user sensitive and can have adverse effects on the analysis if parts of the tissue are not well adhered to the substrate (localized sample charging, for example). An alternative method for sample mounting is the use of adhesive tape; this has been shown to minimize some tissue shrinkage and chemical migration if/when samples are dried prior to SIMS analysis at room temperature.55 

A final consideration during the sectioning process is the possibility of smearing/sample delocalization of material during the cutting process. Obviously, a clean microtome blade should be used but care must be taken if particularly mobile chemicals are present either in the sample of in any embedding medium used to either support the tissue (larger tissue samples may not require embedding) or to adhere the tissue sample to the holder used in the microtome. “Optimal cutting temperature” (OCT) embedding compound is often used by bio/medical researchers but can smear and spread over tissue sections, covering material and also causing ion suppression.56 Gelatin or carboxymethyl cellulose are generally preferred for embedding samples prior to MS imaging analysis. In the event of minor smearing on the surface of the tissue, GCIB sputtering has sometimes been used to clean the sample prior to data analysis.57 

Unfortunately, in many cases, it is not the SIMS expert who is involved in all of the above steps especially for clinical samples. The whole workflow of tissue sample preparation should be checked and partners educated to the requirements for SIMS analysis—keep the samples as clean, cold, and dry as much as possible. Common practices in pathology workflows such as breathing heavily on the substrate to facilitate thaw mounting/tissue adhesion are not recommended for SIMS studies!

As with cellular imaging, the gold standard is analysis of the frozen tissue sample. This approach minimizes morphological changes in the tissue that can occur during drying and vacuum exposure; it reduces the chance of chemical relocation and may increase secondary ion signal as result of the ice matrix.

The benefits of cryogenic analysis have been reported multiple times and the most common example is that of rodent brain imaging where cholesterol migration in the sample has been a persistent hindrance for SIMS analysis. Sjövall and co-workers first showed that this effect could be reduced by performing analysis at −130 °C even though the sample had been dried prior to introduction to the SIMS instrument.58 Angerer et al. showed that higher mass signals such as gangliosides, detected using GCIB analysis, were much higher on frozen hydrated samples analyzed at 130 K versus dried samples at room temperature.59  Figure 5 shows a comparison of SIMS images of the hippocampus region in sectioned rat brain. The RGB overlays feature a phosphatidylinositol (PI) lipid in blue [PI(38:4) at m/z 885.5] and two gangliosides (GM) in red [GM1(36:1) at m/z 1544.9] and green [GM1(38:1) at m/z 1572.9]. The ganglioside signal is greatest in the frozen sample while being partially recovered when the freeze-dried sample is exposed to ammonia vapor. Figure 5(d) shows an overlay of spectra from mouse heart obtained frozen hydrated and freeze dried. While cholesterol signal increases in the dry sample, the other intact lipid signals are decreased upon drying.

FIG. 5.

ToF-SIMS images comparing (a) frozen hydrated (FH), (b) freeze dried (FD), and c) ammonia treatment (NH3) tissue analysis. RGB overlay generated using m/z 885.5 PI (38:4) [M-H], m/z 1544.9 GM1(36:1) [M-H], and m/z 1572.9 GM1(38:1) [M-H]. Scale bar = 1 mm. Reproduced with permission from Angerer et al., Biointerphases 11, 02A319 (2016).59 Copyright 2016, American Vacuum Society. (d) Comparison of positive ion SIMS spectra from frozen hydrated and freeze-dried mouse heart. Reproduced with permission from Sämfors et al., Int. J. Mass. Spectrom. 437, 77 (2019).60 Copyright 2019, Elsevier. All data were acquired using a 40 kV GCIB producing clusters using 8% CO2 in argon gas with a nominal cluster size of 4000.

FIG. 5.

ToF-SIMS images comparing (a) frozen hydrated (FH), (b) freeze dried (FD), and c) ammonia treatment (NH3) tissue analysis. RGB overlay generated using m/z 885.5 PI (38:4) [M-H], m/z 1544.9 GM1(36:1) [M-H], and m/z 1572.9 GM1(38:1) [M-H]. Scale bar = 1 mm. Reproduced with permission from Angerer et al., Biointerphases 11, 02A319 (2016).59 Copyright 2016, American Vacuum Society. (d) Comparison of positive ion SIMS spectra from frozen hydrated and freeze-dried mouse heart. Reproduced with permission from Sämfors et al., Int. J. Mass. Spectrom. 437, 77 (2019).60 Copyright 2019, Elsevier. All data were acquired using a 40 kV GCIB producing clusters using 8% CO2 in argon gas with a nominal cluster size of 4000.

Close modal

While frozen hydrated analysis may be considered the best approach overall, it is not always an option as not all instruments are equipped with cryogenic sample handling and those that are do not always have the same capabilities. Some SIMS instruments include purgeable glove boxes to facilitate transfer of frozen samples while minimizing water condensation and most recently instruments have been adapted to interface with cryo-transfer devices from manufacturers of common cryo-microtome and microscopy equipment.12 

Cryogenic analysis is normally more complicated than room temperature analysis, is generally more labor intensive, and in some cases, is not the best option depending on how, and in what state, the sample is being delivered to the SIMS laboratory. It is much easier to reduce water contamination, for example, if the sample is sliced and mounted in very close proximity to the SIMS instrument.

When samples cannot, for any reason, be analyzed frozen, then potential artifacts from the drying procedure must always be considered when interpreting the data.

Preparation of tissue samples for room temperature analysis starts with two steps: warming and drying. These two steps are sometimes combined, for example by, freeze drying. If the tissue sections are frozen, they should not be exposed to air/water vapor while warming to room temperature. This can be achieved by warming in and inert atmosphere, warming under vacuum (crude freeze drying) of warming with a stream of inert gas. Tissue that is already at room temperature is typically vacuum desiccated for an hour prior to SIMS analysis, based on the series of tissue imaging studies by Brunelle and co-workers.61–64 

One advantage of room temperature analysis/dry sample analysis is that once a sample is at room temperature, the possibility for additional modification to enhance secondary ion signal becomes available. The obvious approach to consider is that of matrix addition similar to MALDI analysis and this has been demonstrated to be beneficial particularly by Wu and Odom65 and subsequently Heeren and co-workers.66,67 For SIMS applications, the matrix should be thinner than in typical MALDI-MS experiments and care should be taken to minimize matrix crystal size that would negate some of the benefits of SIMS. Sublimation of the matrix generally produces the smallest “crystals” of matrix by not actually forming crystals. Most recently, matrix deposition within the SIMS instrument has been demonstrated. Here, intact matrix molecules are sputtered from a target using a GCIB and these ejected molecules are transferred to the sample.

Sample treatment may need to be tailored to the specific application, acid matrices, for example, will favor protonation of positive ions at the detriment of negative ion formation.

Other examples of sample pretreatment have utilized reactive vapors, for instance, trifluoroacetic acid (TFA) or ammonia vapor.59,68 Both of these have been demonstrated in rodent brain sections. Despite being an acid, TFA enhance signals in both positive and negative polarity much of which was attributed to the removal of accumulations of cholesterol on the sample surface that may physically mask chemistry especially in the white matter regions of brain samples. Ammonia vapor exposure has been used to enhance plant metabolite signals and has been shown to increase signals in SIMS analysis of mouse brain in negative ion modes.69 Specifically, NH3 exposure recovered gangliosides signals that were lost following freeze drying of tissue albeit not to the levels observed in frozen hydrated tissue. A note of caution regarding treatment with reactive species such as TFA is that such exposure may ultimately degrade the tissue. The optimum time for TFA vapor exposure was reported to be 30 min following which intact molecular signals started to diminish in the SIMS spectrum.

While most tissue treatment approaches have aimed at increasing as many molecular signals as possible, there are some approaches where specific classes of molecules or even individual proteins have been targeted. Kaya et al. demonstrated the in tissue chemical derivatization of catecholamines (e.g., adrenaline, nor-adrenaline) by reaction with a boronic acid reagent.70 Also, in recent years, there has been considerable development in the use of antibodies with tags that are amenable to MS detection for targeted protein imaging. While the majority of this work has been on specifically developed ToF-SIMS instruments, this approach, sometimes referred to as multiplexed ion beam imaging (MIBI),71 has shown to also be applicable to more general use SIMS instruments and was recently demonstrated on a J105 SIMS instrument by Tian et al.72 

Increasingly, tissue analysis with SIMS has utilized GCIBs and this has led to a convergence with MALDI imaging for lipid analysis. For example, MALDI imaging has been used to show regio-specific localization of gangliosides at plaque sites in Alzheimer's disease model mouse brain (Table II).73 Similar localization can now be observed through SIMS imaging as shown in Fig. 6(c).

FIG. 6.

Tissue imaging examples. Pancreatic B-cell islet tumor in mouse tissue (25 kV Bi3+), scale bar = 100 μm (a). Reproduced with permission from Bluestein et al., Biointerphases 13, 06D402 (2018).78 Copyright 2018, American Vacuum Society. Sequential imaging of lipids [70 kV (H2O)30 k+—GCIB, frozen hydrated] and then metal tagged proteins (40 kV C60+, room temperature) in human breast cancer tissue (b). Reproduced with from Tian et al., Anal. Chem. 93, 8143 (2021).72 Copyright 2021, Author(s), licensed under Creative Commons Agreement. RGB overlays of lipid species in the hippocampus of an Alzheimer's mouse model different ganglioside (GM) lipids accumulate in different regions of the tissue [40 kV (CO2)6 k+—GCIB] unpublished images from University of Gothenburg (Ibrahim Kaya and John Fletcher) (c).

FIG. 6.

Tissue imaging examples. Pancreatic B-cell islet tumor in mouse tissue (25 kV Bi3+), scale bar = 100 μm (a). Reproduced with permission from Bluestein et al., Biointerphases 13, 06D402 (2018).78 Copyright 2018, American Vacuum Society. Sequential imaging of lipids [70 kV (H2O)30 k+—GCIB, frozen hydrated] and then metal tagged proteins (40 kV C60+, room temperature) in human breast cancer tissue (b). Reproduced with from Tian et al., Anal. Chem. 93, 8143 (2021).72 Copyright 2021, Author(s), licensed under Creative Commons Agreement. RGB overlays of lipid species in the hippocampus of an Alzheimer's mouse model different ganglioside (GM) lipids accumulate in different regions of the tissue [40 kV (CO2)6 k+—GCIB] unpublished images from University of Gothenburg (Ibrahim Kaya and John Fletcher) (c).

Close modal
TABLE II.

Small selection of tissue imaging examples highlighting some of the diversity in ion beams, sample preparation methods, analysis conditions, sample types, and spatial resolution.

Tissue typeSubstrate/substrate treatment/washing bufferFixation and additional modificationSurface modification/isotope labelingBeamIon dose/target currentPixel sizeReference
Mouse kidney ITO coated glass Vacuum desiccated from frozen +/− 25 kV Bi3+ 2 × 1012 ions/cm2 2 μ63  
Lung (human) ITO coated glass Vacuum desiccated − 25 kV 
Bi3+ 1 × 1012 ions/cm2 1 μ74  
Alzheimer model mouse brain Glass Freeze dried (some analysis at −80 °C) Dried and fixed Aβ-labeled with liposomes − 25 kV Bi3+ >1 × 1012 ions/cm2 0.4–2 μ75  
Breast cancer (human) ITO coated glass Freeze dried +/− 40 kV Ar/CO2—GCIB 3.2 × 1011–1.3 × 1013 ions/cm2 3.9−10 μ76  
Skin cancer (BCC, human) ITO coated glass Freeze dried +/− CO2—GCIB 3 × 1011–1.3 × 1013 ions/cm2 4–30 μ54 and 81  
Porcine adrenal gland Si wafer On tissue derivatisation of catecholamines 40 kV CO2—GCIB 1 × 1011–2.8 × 1012 ions/cm2 3–30 μ70  
Infarcted mouse heart ITO coated glass Frozen hydrated and freeze-dried analysis +/− 40 kV Ar/CO2—GCIB 5 × 1012 ions/cm2 20 μ60  
Drosophila melanogaster ITO coated glass Frozen hydrated +/− 40 kV Ar/CO2 GCIB (and C60 comparison) ∼5 × 1012 ions/cm2 3–6 μ77  
Pancreatic β cell tumors in mouse model Si wafer Room temperature +/− 25 kV Bi3+ 5 × 1011 ions/cm2 0.8 μ78  
Planarian flatworm Phagocata gracilis ITO coated glass Warmed under a stream of N2 prior to analysis −/− 40 kV Ar/CO2)4k+—GCIB 5 × 1012 ions/cm2 1 × 1013 ions/cm2 12 μm 3 μ79  
Mouse brain Steel plate Freeze dried (3h, −80 °C in vacuum) −/− ToF-SIMS: 20 kV Ar3000+ GCIB Orbitrap -SIMS: 20 kV A2000+ GCIB ToF-SIMS: 2.5 × 1011 ions/cm2 Orbitrap-SIMS: 3.37 × 1013 ions/cm2 30 μ80  
Tissue typeSubstrate/substrate treatment/washing bufferFixation and additional modificationSurface modification/isotope labelingBeamIon dose/target currentPixel sizeReference
Mouse kidney ITO coated glass Vacuum desiccated from frozen +/− 25 kV Bi3+ 2 × 1012 ions/cm2 2 μ63  
Lung (human) ITO coated glass Vacuum desiccated − 25 kV 
Bi3+ 1 × 1012 ions/cm2 1 μ74  
Alzheimer model mouse brain Glass Freeze dried (some analysis at −80 °C) Dried and fixed Aβ-labeled with liposomes − 25 kV Bi3+ >1 × 1012 ions/cm2 0.4–2 μ75  
Breast cancer (human) ITO coated glass Freeze dried +/− 40 kV Ar/CO2—GCIB 3.2 × 1011–1.3 × 1013 ions/cm2 3.9−10 μ76  
Skin cancer (BCC, human) ITO coated glass Freeze dried +/− CO2—GCIB 3 × 1011–1.3 × 1013 ions/cm2 4–30 μ54 and 81  
Porcine adrenal gland Si wafer On tissue derivatisation of catecholamines 40 kV CO2—GCIB 1 × 1011–2.8 × 1012 ions/cm2 3–30 μ70  
Infarcted mouse heart ITO coated glass Frozen hydrated and freeze-dried analysis +/− 40 kV Ar/CO2—GCIB 5 × 1012 ions/cm2 20 μ60  
Drosophila melanogaster ITO coated glass Frozen hydrated +/− 40 kV Ar/CO2 GCIB (and C60 comparison) ∼5 × 1012 ions/cm2 3–6 μ77  
Pancreatic β cell tumors in mouse model Si wafer Room temperature +/− 25 kV Bi3+ 5 × 1011 ions/cm2 0.8 μ78  
Planarian flatworm Phagocata gracilis ITO coated glass Warmed under a stream of N2 prior to analysis −/− 40 kV Ar/CO2)4k+—GCIB 5 × 1012 ions/cm2 1 × 1013 ions/cm2 12 μm 3 μ79  
Mouse brain Steel plate Freeze dried (3h, −80 °C in vacuum) −/− ToF-SIMS: 20 kV Ar3000+ GCIB Orbitrap -SIMS: 20 kV A2000+ GCIB ToF-SIMS: 2.5 × 1011 ions/cm2 Orbitrap-SIMS: 3.37 × 1013 ions/cm2 30 μ80  

The complexity of cell and tissue data can make interpretation a challenge, even for the experienced SIMS user. In this final section, the most common approaches will be described. Different SIMS instrument can produce quite different looking data. The available ion beam will change the character of the mass spectrum with atomic beams producing lower mass species and more fragment ions with reduced signal from molecular ions. Mass analyzers have become more varied as has their performance. ToF analyzers are still the most common but these may be combined with bunchers as in the J105 instrument11,81 or have an option for a second mass spectrometer such as in the 3D OrbiSIMS hybrid instrument. Ionoptika, Iontof, and Physical Electronics all have options for tandem mass spectrometry on their instruments.12,82,83

The rich complexity of SIMS data means that multivariate analysis methods have long been employed for reducing the dimensionality of the data. Principal components analysis (PCA) has been used for feature identification, background/substrate removal, spectral classification, etc. A tutorial including details and guidance for MVA of SIMS data has recently been published and so such approaches will not be described in detail here.84 Furthermore, the computational tools for data analysis are evolving at a significant rate with developments in supervised classification methods often incorporating machine learning. Recent examples include the use of artificial neural networks in an interlaboratory study by Aoyagi and co-workers (including different ion beams and mass analyzer types) and study of multilayered binary mixtures of organic electronic material related to predict composition of mixtures where matrix effects influence the results and also provide indication of which peaks are most useful for quantitative assessment.85 Forbes et al. have recently demonstrated the use of non-negative matrix factorization (NMF) with k-means (NMFk) for image analysis of heterogeneous mixtures of pharmaceuticals.86 The journey from MVA toward machine learning approaches is described in a recent review by Gardner et al.87 

Following MVA, the challenge of interpreting the data still remains and there are particular challenges associated with peak identification for biological samples while on the other hand, due to the wide use of other MS methods in the biological research fields, there are several non-SIMS specific tools/databases that can help.

While many laboratories have their own standards and libraries available, there are many available databases that can be searched to identify peaks in SIMS spectra. Typically, the first requirement for peak matching is the m/z value of the (pseudo) molecular ion, e.g., [M ± H]± often [M + Na/K]+. A mass accuracy or mass tolerance window is typically required before searching. Hence, good mass calibration across the mass range of interest, typically performed on the same type of ions as the analyte is the starting point. An assessment of the mass accuracy of the measurement should be made, perhaps using an additional peak from a known molecule within the sample. For instruments where the mass spectrometry is decoupled from the sputtering event, i.e., if using an Orbitrap or J105 Instrument, the mass accuracy is likely to be much better than on a directly coupled ToF-MS system (as the secondary ion energy will vary based on topography and possible charging effects) but this may be sample dependent.

Specific isotope patterns can help increase confidence in peak assignment. Most databases do not include this information in peak searching but the metaSPACE web platform (https://metaspace2020.eu/) uses colocalization of isotopes as a contributor to confidence of assignment.88 

Fragmentation during secondary ion generation can be both a useful tool and a complication. MS databases do not identify common SIMS fragments and care must be taken to avoid incorrectly assigning a fragment ion as a molecular ion—not all fragments are low mass! Common issues from lipid fragmentation in SIMS are the loss of trimethylamine (59 Da) from Na/K adduct ions of phosphatidylcholines as these are identified as phosphatidic acid (PA) lipids in computationally generated databases such as “lipid maps” (www.lipidmaps.org) despite PAs ionizing negatively. Similarly, phosphatidylserine (PS) lipids can fragment via a neutral loss of 87 Da, again producing a potential database match with PA lipids, this time in the correct polarity as PS lipids are detected in negative ion mode.89 Other common fragmentation overlaps from lipids include diacylglyceride (DAG) [M + H – H2O]+ ions from fragmentation of triacylglycerides (TAG), loss of the head group (183 Da) from [M + Na]+ ions of phosphatidylcholine lipids, ceramide fragments from sphingomyelins and lyso-lipids. There are no doubt many more but these are the commonly miss assigned or ambiguous species when using database matching in lipid maps.

Conversely, it can be argued that the presence of fragments within the SIMS spectra may be considered a “built in MSMS” and colocalization of known fragments with a putatively assigned molecular ion precursor can increase confidence in assignment with the caveat that in a complex biological many common fragments may originate from different starting molecules. There is generally good agreement between fragment ions generated in SIMS and those obtained through tandem MS measurements and so comparison with complementary data (e.g., from electrospray-MSMS data) can improve confidence in peak assignment.90 While these data can be generated in house, given access to suitable instrumentation, there are also online databases such as Mass Bank (https://massbank.eu/MassBank/) that provide MSMS data for a wide range of compounds relevant to cell and tissue samples analyzed using SIMS. It should be noted that some low mass ions are more SIMS specific and are not produced by gas phase CID in the energy ranges typically employed (10–100 eV).

Cluster analysis is a quantitative method used to identify clusters of related samples by calculating high similarity coefficients between each pair of samples and subsequently grouping them accordingly.

In order to understand a biological question from different types of experiments, a data integration approach might be useful to compare the outcomes for the same conditions such as drug doses and disease conditions.

If there will be an integrated approach such as multiomics, trend comparison upon drug treatment to combine different method outcomes, first data should be normalized to make them comparable. To normalize data either an excel sheet can be used or popular tools such as MetaboAnalyst or custom packages written in R/Python programming languages. Lotfollahi et al. recently suggested a multimodal data integration approach, Multigrate, aiming to create joint representations of multiomic single-cell data sets.91 After the normalization step, coclustering/ multiclustering of multimodal data can be made by either using popular tools such as MetaboAnalyst or custom packages. Commonly used clustering algorithms are k-means clustering and hierarchical clustering.92,93 One example was used to cluster electrochemical amperometric data with ToF-SIMS lipidomic data to visualize (heat map) the relationship between exocytotic events with lipid alteration.41 

After finding altered lipids under a certain biological condition such as disease, drug effects, etc., it is important to understand the link between those lipid groups within biological pathways. Up to now, genomic and proteomic disciplines use pathway enrichment tools widely to elucidate the meaning of the altered gene/protein signatures.94 As an emerging field of the omics area, lipidomics has recently been producing large amounts of data. Hence, scientists have started to use pathway enrichment tools such as LIPEA, Metaboanalyst, BioPAN, IMPaLA, and ConsensusPathDB. These tools are connected with databases that save annotations and interaction data from previously collected and curated experimental entry sets or other literature mining databases such as LipidPedia.95 When the altered lipid query list (with assigned reference IDs) is entered based on the biological condition, the database will suggest which pathways might be enriched in the specific condition of interest such as a disease and drug. Hence, it enables researchers to see the biological relevance of the differed lipid groups. There are several examples where lipid pathway enrichment tools have been used with the data outcome of secondary ion mass spectrometry. The LIPEA web tool has been used to investigate enrichment of the pathways with altered membrane lipids analyzed by GCIB-SIMS upon lactacystin treatment of the cells and found glycerophospholipid metabolism, autophagy, GPI-anchor biosynthesis, and ferroptosis pathways were related significantly.41 Ren et al. used the pathway analysis tool of Metaboanalyst platform to understand the single-cell lipidomic profile of cardiomyocytes in the condition of heart failure analyzed by ToF-SIMS imaging. They found that signal transduction, lipid, and lipoprotein metabolism were enriched with the altered lipids.96 

The analysis of cells and tissue using secondary ion mass spectrometry is still a rapidly evolving area. The diversity of different sample types and the different biological questions that need to be answered mean that there are a wide range of different approaches that are used. Development of new ion beams, new sample treatment strategies and new mass spectrometry increase the amount of information that can be extracted from a sample. While preparation and analysis conditions can vary, it is apparent that, for untargeted analysis, cryogenic preservation combined with frozen hydrated analysis provides the best preservation of the sample along with potential signal benefits associated with the presence of an ice matrix.

John Fletcher gratefully acknowledges financial support from the Swedish Research Council (VR, Ref. No. 2022-04498).

The authors have no conflict of interest.

Ethical approval is not required.

Inci Barut: Writing – original draft (equal); Writing – review & editing (equal). Fletcher: Conceptualization (equal); Writing – original draft (equal); Writing – review & editing (equal).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

References

1.
J. C.
Vickerman
, in
Molecular Surface Mass Spectrometry by SIMS, Surface Analysis—The Principal Techniques
, edited by
J. C. V. I. S.
Gilmore
(
Wiley
,
New York
,
2009
), Vol. 2, pp.
113
203
.
2.
P. D.
Piehowski
,
M. E.
Kurczy
,
D.
Willingham
,
S.
Parry
,
M. L.
Heien
,
N.
Winograd
, and
A. G.
Ewing
,
Langmuir
24
,
7906
(
2008
).
3.
N.
Davies
,
D. E.
Weibel
,
P.
Blenkinsopp
,
N.
Lockyer
,
R.
Hill
, and
J. C.
Vickerman
,
Appl. Surf. Sci.
203
,
223
(
2003
).
4.
D.
Touboul
,
F.
Kollmer
,
E.
Niehuis
,
A.
Brunelle
, and
O.
Laprevote
,
J. Am. Soc. Mass Spectrom.
16
,
1608
(
2005
).
5.
D.
Weibel
,
S.
Wong
,
N.
Lockyer
,
P.
Blenkinsopp
,
R.
Hill
, and
J. C.
Vickerman
,
Anal. Chem.
75
,
1754
(
2003
).
6.
E. A.
Jones
,
J. S.
Fletcher
,
C. E.
Thompson
,
D. A.
Jackson
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Appl. Surf. Sci.
252
,
6844
(
2006
).
7.
J. S.
Fletcher
,
S.
Rabbani
,
A. M.
Barber
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Surf. Interface Anal.
45
,
273
(
2013
).
8.
T. B.
Angerer
,
P.
Blenkinsopp
, and
J. S.
Fletcher
,
Int. J. Mass Spectrom.
377
,
591
(
2015
).
9.
S.
Sheraz
,
A.
Barber
,
J. S.
Fletcher
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Anal. Chem.
85
,
5654
(
2013
).
10.
K.
Dimovska Nilsson
,
A.
Karagianni
,
I.
Kaya
,
M.
Henricsson
, and
J. S.
Fletcher
,
Anal. Bioanal. Chem.
413
,
4181
(
2021
).
11.
J. S.
Fletcher
,
S.
Rabbani
,
A.
Henderson
,
P.
Blenkinsopp
,
S. P.
Thompson
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Anal. Chem.
80
,
9058
(
2008
).
12.
M. K.
Passarelli
et al,
Nat. Methods
14
,
1175
(
2017
).
13.
I.
Dagogo-Jack
and
A. T.
Shaw
,
Nat. Rev. Clin. Oncol.
15
,
81
(
2018
).
14.
S. J.
Altschuler
and
L. F.
Wu
,
Cell
141
,
559
(
2010
).
15.
M. J.
Taylor
,
J. K.
Lukowski
, and
C. R.
Anderton
,
J. Am. Soc. Mass Spectrom.
32
,
872
(
2021
).
16.
A. F.
Altelaar
,
I.
Klinkert
,
K.
Jalink
,
R. P.
de Lange
,
R. A.
Adan
,
R. M.
Heeren
, and
S. R.
Piersma
,
Anal. Chem.
78
,
734
(
2006
).
17.
A.
Bodzon-Kulakowska
and
P.
Suder
,
Mass Spectrom. Rev.
35
,
147
(
2016
).
18.
H.
Tian
,
L. J.
Sparvero
,
P.
Blenkinsopp
,
A. A.
Amoscato
,
S. C.
Watkins
,
H.
Bayir
,
V. E.
Kagan
, and
N.
Winograd
,
Angew. Chem. Int. Ed. Engl.
58
,
3156
(
2019
).
19.
L. J.
Sparvero
et al,
Angew. Chem. Int. Ed. Engl.
60
,
11784
(
2021
).
20.
N.
Yeow
,
R. F.
Tabor
, and
G.
Garnier
,
Adv. Colloid Interface Sci.
249
,
149
(
2017
).
21.
A.
Gulin
,
V.
Nadtochenko
,
A.
Astafiev
,
V.
Pogorelova
,
S.
Rtimi
, and
A.
Pogorelov
,
Analyst
141
,
4121
(
2016
).
22.
S.
Samfors
and
J. S.
Fletcher
,
Annu. Rev. Anal. Chem.
13
,
249
(
2020
).
23.
D.
Breitenstein
,
C. E.
Rommel
,
J.
Stolwijk
,
J.
Wegener
, and
B.
Hagenhoff
,
Appl. Surf. Sci.
255
,
1249
(
2008
).
24.
C.
Zhang
,
M.
Horikawa
,
T.
Kahyo
,
T.
Matsudaira
,
T.
Tanaka
,
L.
Xu
,
S.
Takei
, and
M.
Setou
,
Microscopy
71
,
324
(
2022
).
25.
J. S.
Fletcher
,
N. P.
Lockyer
,
S.
Vaidyanathan
, and
J. C.
Vickerman
,
Anal. Chem.
79
,
2199
(
2007
).
26.
J.
Malm
,
D.
Giannaras
,
M. O.
Riehle
,
N.
Gadegaard
, and
P.
Sjovall
,
Anal. Chem.
81
,
7197
(
2009
).
27.
E. S. F.
Berman
,
S. L.
Fortson
,
K. D.
Checchi
,
L.
Wu
,
J. S.
Felton
,
K. J. J.
Wu
, and
K. S.
Kulp
,
J. Am. Soc. Mass Spectrom.
19
,
1230
(
2008
).
28.
V.
Pareek
,
H.
Tian
,
N.
Winograd
, and
S. J.
Benkovic
,
Science
368
,
283
(
2020
).
29.
K.
Schaepe
,
J.
Kokesch-Himmelreich
,
M.
Rohnke
,
A.-S.
Wagner
,
T.
Schaaf
,
S.
Wenisch
, and
J.
Janek
,
Biointerphases
10
,
019016
(
2015
).
30.
W. A.
Ausserer
,
S.
Chandra
, and
G. H.
Morrison
,
J. Microscopy
154
,
39
(
1989
).
31.
32.
T. P.
Roddy
,
D. M.
Cannon
, Jr.
,
S. G.
Ostrowski
,
N.
Winograd
, and
A. G.
Ewing
,
Anal. Chem.
74
,
4020
(
2002
).
33.
K. S.
Kulp
,
E. S.
Berman
,
M. G.
Knize
,
D. L.
Shattuck
,
E. J.
Nelson
,
L.
Wu
,
J. L.
Montgomery
,
J. S.
Felton
, and
K. J.
Wu
,
Anal. Chem.
78
,
3651
(
2006
).
34.
S. G.
Ostrowski
,
M. E.
Kurczy
,
T. P.
Roddy
,
N.
Winograd
, and
A. G.
Ewing
,
Anal. Chem.
79
,
3554
(
2007
).
35.
J. S.
Fletcher
,
S.
Rabbani
,
A.
Henderson
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Rapid Commun. Mass Spectrom.
25
,
925
(
2011
).
36.
I.
Lanekoff
,
M. E.
Kurczy
,
R.
Hill
,
J. S.
Fletcher
,
J. C.
Vickerman
,
N.
Winograd
,
P.
Sjovall
, and
A. G.
Ewing
,
Anal. Chem.
82
,
6652
(
2010
).
37.
H.
Tian
,
S.
Sheraz Nee Rabbani
,
J. C.
Vickerman
, and
N.
Winograd
,
Anal. Chem.
93
,
7808
(
2021
).
38.
H. W.
Li
,
X.
Hua
, and
Y. T.
Long
,
Anal. Bioanal. Chem.
411
,
4025
(
2019
).
39.
P.
Sjovall
,
J.
Lausmaa
,
H.
Nygren
,
L.
Carlsson
, and
P.
Malmberg
,
Anal. Chem.
75
,
3429
(
2003
).
40.
H.
Nygren
,
C.
Eriksson
,
P.
Malmberg
,
H.
Sahlin
,
L.
Carlsson
,
J.
Lausmaa
, and
P.
Sjovall
,
Colloid Surf. B
30
,
87
(
2003
).
41.
I.
Lanekoff
,
P.
Sjovall
, and
A. G.
Ewing
,
Anal. Chem.
83
,
5337
(
2011
).
42.
B. J.
Tyler
,
M. M.
Takeno
, and
K. D.
Hauch
,
Surf. Interface Analy.
43
,
336
(
2011
).
43.
T. B.
Angerer
and
J. S.
Fletcher
,
Surf. Interface Anal.
46
,
198
(
2014
).
44.
M. A.
Robinson
,
D. J.
Graham
, and
D. G.
Castner
,
Anal. Chem.
84
,
4880
(
2012
).
45.
H.
Tian
,
L. J.
Sparvero
,
P.
Blenkinsopp
,
A. A.
Amoscato
,
S. C.
Watkins
,
H.
Bayır
,
V. E.
Kagan
, and
N.
Winograd
,
Angew. Chem.
131
,
3188
(
2019
).
46.
S.
Muramoto
,
J.
Brison
, and
D.
Castner
,
Surf. Interface Anal.
43
,
58
(
2011
).
47.
J. S.
Fletcher
,
A.
Henderson
,
G. X.
Biddulph
,
S.
Vaidyanathan
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Appl. Surf. Sci.
255
,
1264
(
2008
).
48.
D.
Breitenstein
,
C. E.
Rommel
,
R.
Mollers
,
J.
Wegener
, and
B.
Hagenhoff
,
Angew. Chem. Int. Ed.
46
,
5332
(
2007
).
49.
E. J.
Lanni
,
R. N.
Masyuko
,
C. M.
Driscoll
,
S. J.
Dunham
,
J. D.
Shrout
,
P. W.
Bohn
, and
J. V.
Sweedler
,
Anal. Chem.
86
,
10885
(
2014
).
50.
C. F.
Shao
,
Y.
Zhao
,
K.
Wu
,
F. F.
Jia
,
Q.
Luo
,
Z.
Liu
, and
F. Y.
Wang
,
Chin. J. Anal. Chem.
46
,
1005
(
2018
).
51.
J. S.
Fletcher
,
S.
Sämfors
,
J.
Vallin
,
A.
Svanström
, and
J.
Grantham
,
Anal. Bioanal. Chem.
413
,
445
(
2021
).
52.
E.
Ranjbari
,
M. H.
Philipsen
,
Z.
Wang
, and
A. G.
Ewing
,
QRB Discov.
2
,
e6
(
2021
).
53.
I.
Barut
,
X.
He
,
E.
Sener
,
S.
Samfors
,
A. G.
Ewing
, and
J. S.
Fletcher
,
Angew. Chem. Int. Ed. Engl.
62
,
e202217993
(
2023
).
54.
K. D.
Nilsson
,
N.
Neittaanmäki
,
O.
Zaar
,
T. B.
Angerer
,
J.
Paoli
, and
J. S.
Fletcher
,
Biointerphases
15
,
041012
(
2020
).
55.
M. U. T.
Le
,
J. G.
Son
,
H. K.
Shon
,
J. H.
Park
,
S. B.
Lee
, and
T. G.
Lee
,
Biointerphases
13
,
03b414
(
2018
).
56.
J. X. M.
Truong
,
X.
Spotbeen
,
J.
White
,
J. V.
Swinnen
,
L. M.
Butler
,
M. F.
Snel
, and
P. J.
Trim
,
Anal. Bioanal. Chem.
413
,
2695
(
2021
).
57.
M.
Munem
,
O.
Zaar
,
K. D.
Nilsson
,
N.
Neittaanmaki
,
J.
Paoli
, and
J. S.
Fletcher
,
Biointerphases
13
,
03B402
(
2018
).
58.
P.
Sjövall
,
B.
Johansson
, and
J.
Lausmaa
,
Appl. Surf. Sci.
252
,
6966
(
2006
).
59.
T. B.
Angerer
,
A. S.
Mohammadi
, and
J. S.
Fletcher
,
Biointerphases
11
,
02A319
(
2016
).
60.
S.
Sämfors
,
M.
Ståhlman
,
M.
Klevstig
,
J.
Borén
, and
J. S.
Fletcher
,
Int. J. Mass Spectrom.
437, 77 (
2019
).
61.
S.
Mas
,
D.
Touboul
,
A.
Brunelle
,
P.
Aragoncillo
,
J.
Egido
,
O.
Laprevote
, and
F.
Vivanco
,
Analyst
132
,
24
(
2007
).
62.
D.
Debois
,
M. P.
Bralet
,
F.
Le Naour
,
A.
Brunelle
, and
O.
Laprevote
,
Anal. Chem.
81
,
2823
(
2009
).
63.
I.
Abbas
,
M.
Noun
,
D.
Touboul
,
D.
Sahali
,
A.
Brunelle
, and
M.
Ollero
,
Int. J. Molecular Sci.
20
,
1623
(
2019
).
64.
M.
Brulet
,
A.
Seyer
,
A.
Edelman
,
A.
Brunelle
,
J.
Fritsch
,
M.
Ollero
, and
O.
Laprevote
,
J. Lipid Res.
51
,
3034
(
2010
).
65.
K. J.
Wu
and
R. W.
Odom
,
Anal. Chem.
68
,
873
(
1996
).
66.
K.
Krijnen
,
J. D.
Keelor
,
S.
Böhm
,
S. R.
Ellis
,
C.
Köster
,
J.
Höhndorf
,
R. M. A.
Heeren
, and
I. G. M.
Anthony
,
J. Am. Soc. Mass Spectrom.
34
,
720
(
2023
).
67.
L. A.
McDonnell
,
S. R.
Piersma
,
A. F. M.
Altelaar
,
T. H.
Mize
,
S. L.
Luxembourg
,
P. D. E. M.
Verhaert
,
J.
van Minnen
, and
R. M. A.
Heeren
,
J. Mass Spectrometry
40
,
160
(
2005
).
68.
T. B.
Angerer
,
M. D.
Pour
,
P.
Malmberg
, and
J. S.
Fletcher
,
Anal. Chem.
87
,
4305
(
2015
).
69.
S.
Vaidyanathan
,
M.
Salim
,
C.
Hurley
, and
J.
Pugh
,
Surf. Interface Anal.
45
,
290
(
2013
).
70.
I.
Kaya
,
S. M.
Brulls
,
J.
Dunevall
,
E.
Jennische
,
S.
Lange
,
J.
Martensson
,
A. G.
Ewing
,
P.
Malmberg
, and
J. S.
Fletcher
,
Anal. Chem.
90
,
13580
(
2018
).
72.
H.
Tian
,
L. J.
Sparvero
,
T. S.
Anthonymuthu
,
W. Y.
Sun
,
A. A.
Amoscato
,
R. R.
He
,
H.
Bayir
,
V. E.
Kagan
, and
N.
Winograd
,
Anal. Chem.
93
,
8143
(
2021
).
73.
I.
Kaya
,
E.
Jennische
,
J.
Dunevall
,
S.
Lange
,
A. G.
Ewing
,
P.
Malmberg
,
A. T.
Baykal
, and
J. S.
Fletcher
,
ACS Chem. Neurosci.
11
,
14
(
2020
).
74.
B. M.
Bluestein
,
F.
Morrish
,
D. J.
Graham
,
L.
Huang
,
D.
Hockenbery
, and
L. J.
Gamble
,
Biointerphases
13
,
06D402
(
2018
).
75.
76.
L.
Carlred
et al,
J. Am. Chem. Soc.
136
,
9973
(
2014
).
77.
T. B.
Angerer
,
Y.
Magnusson
,
G.
Landberg
, and
J. S.
Fletcher
,
Anal. Chem.
88
,
11946
(
2016
).
78.
N. T. N.
Phan
,
J. S.
Fletcher
, and
A. G.
Ewing
,
Anal. Chem.
87
,
4063
(
2015
).
79.
T. B.
Angerer
,
N.
Chakravarty
,
M. J.
Taylor
,
C. D.
Nicora
,
D. J.
Graham
,
C. R.
Anderton
,
E. H.
Chudler
, and
L. J.
Gamble
,
Biochim. Biophys. Acta Mol. Cell Biol. Lip.
1864
,
733
(
2019
).
80.
H. K.
Shon
,
J. G.
Son
,
S. Y.
Lee
,
J. H.
Moon
,
G. S.
Lee
,
K.-S.
Kim
, and
T. G.
Lee
,
Biointerphases
18
, 031005 (
2023
).
81.
R.
Hill
,
P.
Blenkinsopp
,
S.
Thompson
,
J.
Vickerman
, and
J. S.
Fletcher
,
Surf. Interface Anal.
43
,
506
(
2011
).
82.
N. T. N.
Phan
,
M.
Munem
,
A. G.
Ewing
, and
J. S.
Fletcher
,
Anal. Bioanal. Chem.
409
,
3923
(
2017
).
83.
N.
Ogrinc Potocnik
,
G. L.
Fisher
,
A.
Prop
, and
R. M. A.
Heeren
,
Anal. Chem.
89
,
8223
(
2017
).
84.
D. J.
Graham
and
L. J.
Gamble
,
Biointerphases
18
, 031201 (
2023
).
86.
T. P.
Forbes
,
J. G.
Gillen
,
A. J.
Souna
, and
J.
Lawrence
,
Anal. Chem.
94
,
16443
(
2022
).
87.
W.
Gardner
,
D. A.
Winkler
,
B. W.
Muir
, and
P. J.
Pigram
,
Biointerphases
17
,
020802
(
2022
).
88.
89.
F. F.
Hsu
and
J.
Turk
,
J. Chromatogr. B
877
,
2673
(
2009
).
90.
J. S.
Fletcher
,
H. L.
Kotze
,
E. G.
Armitage
,
N. P.
Lockyer
, and
J. C.
Vickerman
,
Metabolomics
9
,
535
(
2013
).
91.
M.
Lotfollahi
,
A.
Litinetskaya
, and
F. J.
Theis
, bioRxiv, 2022.2003.2016.484643 (2022).
92.
K. C.
O’Neill
,
P.
Hinners
, and
Y. J.
Lee
,
Anal. Methods
12
,
792
(
2020
).
94.
C.
Monti
,
M.
Zilocchi
,
I.
Colugnat
, and
T.
Alberio
,
J. Proteomics
198
,
36
(
2019
).
95.
S. M.
Lam
,
Z.
Wang
,
B.
Li
, and
G.
Shui
,
Anal. Chim. Acta
1147
,
199
(
2021
).