5.3 Desorption/Ionization Techniques used for MSI
5.3.3 MSI Methods Using Electrospray Ionization
Desorption electrospray ionization (DESI) is an ambient MSI method that involves the exposure of the sample (e.g. tissue section) to charged microdroplets generated by ESI [84]. These charged, high‐velocity microdroplets impinge the sample and extract the analytes [85]. The droplets scatter off the surface toward the inlet capillary of the mass spectrometer. Ionization occurs via ESI‐like processes as the charged microdroplets undergo desolvation. Thus, DESI analysis does not require external matrices and allows samples to be analyzed under ambient conditions. Spray sol- vents consisting of MeOH:H2O or ACN:H2O mixtures are most commonly used for
5 Mass Spectrometry Imaging of Lipids 126
DESI and are suitable for positive‐ and negative‐ion mode analysis. However, these can disturb the underlying tissue structure, prohibiting informative histological staining after DESI. Use of 1:1 N,N‐dimethylformamide:EtOH spray solvents has been shown to preserve the tissue structure and generate rich lipid spectra [86]. The spot size of the electrospray largely determines the spatial resolution of DESI‐MSI on the sample surface with pixels as low as 20 μm being achieved for tissue imaging [87].
However, pixel sizes of 50–150 μm are more common.
Single cell
100%
(a) (b) (c)
(d) (e) (f)
0%
100%
0%
(g) 1330
Single pixel
(h)
1470 m/z
1330 m/z 1470
Figure 5.2 GCIB-SIMS-MSI of single Ht22 cells at 1 μm pixel size using a 70 keV (CO2)10 000+ primary ion beam. Before analysis, the cells were treated with (1-ethyl-3-(3- dimethylaminopropyl)carbodiimide hydrochloride) and phospholipase C. (a) Total-ion current image from the first analysis layer showing the cell locations (scale bar = 50 μm). (b and c) SIMS-MSI data form the second analysis layer hosing distribution of (b) [CL(68:2)−H]− and (c) [PI(38:4)−H]−. (D–F) shows distributions of [PI(38:4)−H]− (green), [CL(68:2)−H]− (magenta), and deoxyribosediphosphate (m/z 257.0, blue) from the () first (0–200 nm), (e) second (200–400 nm), and (f) third (400–600 nm) analysis layers (depth indicated from the initial sample surface).
(g) Mass spectrum showing the cardiolipin mass range from the cell marked with an arrow in (a).
(h) Representative single-pixel spectrum from the same indicated cell. Source: Tian et al. [70].
DESI‐MSI is well suited for imaging a wide range of lipid species from tissues, including many phospholipids and sphingolipids [88, 89]. For example, one study demonstrated that DESI detects many phospholipid classes from tissue surfaces that are also detected by liquid chromatography mass spectrometry (LC‐MS) [90].
Another study used DESI‐MSI to map lipid changes throughout rat sciatic nerve tissue with lipid identifications supported by LC‐MS/MS data [91]. Many lipid spe- cies, including phospholipids (PLs), TAGs, diacylglycerols (DAGs), glycosphingolip- ids (GSLs), sulfatides, and ceramide‐1‐phosphates, were detected and imaged with distinct lipid profiles observed across fibers, connective tissue, and adipose tissue.
DESI also generated higher signals for FFAs than MALDI, which has been explored in a variety of studies. As one example, Sato et al. investigated the distribution of polyunsaturated fatty acids (PUFAs) such as FA(22:5) and FA(22:6) along with PUFA metabolites including 12‐hydroxyeicosapentaenoic acid (12‐HEPE), 15‐
hydroxydocosahexaenoic acid (15‐HDoHE), Protectin D1, and protectin D1 and leu- kotriene B5 (LTB5) within plaques derived from apolipoprotein E‐deficient mice supplemented with either FA(22:5) or FA(22:6) diets [92].
DESI‐MS has been applied to study lipid distributions in many tissue types, includ- ing cancerous tissues such as brain [93–95], breast [96], ovarian [97–99], and colorec- tal [100, 101] where it can also classify the nature of the diseased tissue. For example, DESI‐MSI can differentiate ovarian high‐grade carcinomas and borderline ovarian tumors [97] in addition to breast cancer subtypes [96]. DESI‐MSI in three dimensions has been applied to study cancer tissue such as glioblastoma [102] and colorectal can- cers [103]. Henderson et al. combined both positive‐ and negative‐ion mode DESI‐MSI to reveal 3D lipid distributions throughout xenograft glioblastoma tumors from mice [102], revealing distinct lipid within hypoxic and normoxic regions with acylcar- nitines localizing to the hypoxic regions and PI(38:4) higher in viable tissue.
5.3.3.2 Laser Ablation Electrospray Ionization and IR-Matrix-Assisted Laser Desorption-Electrospray Ionization
Laser ablation electrospray ionization (LAESI) [104] and infrared‐matrix‐assisted laser desorption‐electrospray ionization (IR‐MALDESI) [105] are two similar ambient‐based approaches using an infrared laser to desorb analytes from the sample surface into a plume of charged droplets generated by an ESI emitter. Desorption is initiated by absorption of the 2.94 μm laser pulse by water, while ionization occurs via ESI‐like processes for analytes captured by the charged droplets. For tissue imaging experi- ments, a thin film of amorphous ice is formed on the surface by cooling the sample and acts as the energy‐absorbing matrix. The spatial resolution is determined by the size of the laser spot on the sample and is typically 100–200 μm. However, recent implementa- tions have reported spatial resolutions for MSI experiments of 50 μm [106].
LAESI/IR‐MALDIESI has been applied to the imaging of lipids and metabolites from a variety of samples, including plant material [107, 108], ovarian tissue [109], zebrafish [110], dog liver [111], mouse brain [112], and mouse skin tissue [113]. The latter study demonstrated the 3D analysis of lipids in skin tissue by utilizing the ability to sequentially probe deeper into the sample with each successive laser pulse. Monitoring signals as a function of laser shots thus provides insights into the distribution of analytes as a function of depth. A depth resolution of 7 μm was achieved along with a lateral
5 Mass Spectrometry Imaging of Lipids 128
resolution of 50 μm and allowed 3D imaging of whole skin without the need for section- ing of skin samples. Numerous lipid species, including PLs, sphingolipids, TAGs, DAGs, and cholesterol, were detected. As a selected example, sphingosine‐1‐phosphate was found with higher signal intensity in the dermis compared to the epidermis and hypo- dermis, while cholesterol was higher in the epidermis and TAGs in the hypoder- mis [113]. In addition, unsaturated lipids that may be poorly ionized using typical ESI spay solutions can be enhanced by using dopants such as silver [114].
5.3.3.3 Nanospray Desorption Electrospray Ionization
Nanospray desorption electrospray ionization (nano‐DESI) is another ambient MSI method for imaging lipids from tissue sections [115]. It involves the formation of a liquid microjunction between two capillaries on the sample surface. The solvent flow forming the microjunction extracts analytes from the sample and is aspirated into a nano‐ESI capillary for ionization and transfer into the inlet capillary. As per other ESI‐
based methods, nano‐DESI is well suited for detecting many lipid classes well ionized by ESI and spatial resolutions as low as ~10 μm reported. For example, Yin et al. dem- onstrated the localization of lipid signals, including a variety of phosphatidylcholine (PC) and oxidized PC species to individual islets at a spatial resolution of 11 μm [116].
A comparison of lipid coverage from mouse lung tissue compared to that obtained following lipid extraction and LC‐MS/MS revealed that nano‐DESI could detect 265 lipid signals across the positive‐ and negative‐ion modes from 20 lipid subclasses using 9 : 1 MeOH:H2O (v/v) as the solvent [117]. This corresponds to roughly half of the sig- nals detected following Folch extraction and LC‐MS/MS and confirms the suitability of nano‐DESI for many lipid species. TAGs were detected significantly better using LC‐MS/MS compared to nano‐DESI. TAG detection using nano‐DESI‐MSI can be improved by using less polar solvents such as 5 : 3.5 : 1.5 (v/v/v) MeOH:CAN:toluene [118].
The addition of dopants such as silver to the solvent can also aid in detecting additional lipid classes. For example, the use of silver dopant promoted the formation of [M+Ag]+ ions of prostaglandins and has been applied for imaging prostaglandins such as PGE2, PGF2α, dimethyl‐PGE2, PGA1, and PGK2 within the mouse uterine tissue from four day pregnant mice. The sensitivity of [M+Ag]+ ions was ~30‐fold higher than that obtained when using the [M−H]− ions. Prostaglandins were primarily localized to the luminal epithelium and glandular epithelium regions of the tissue.