Mass spectrometry imaging and its place in nephrology

The last two decades have seen enormous technical improvements in molecular imaging techniques in terms of sensitivity, specificity and spatial resolution. We now have at our disposal a plethora of different methods to display structural, functional and molecular changes in tissues, including optical imaging, computed tomography, magnetic resonance imaging, positron emission tomography, single-photon emission computed tomography and ultrasound [1].

In this article, we will discuss a method probably less known by physicians, yet holding tremendous potential for future research: mass spectrometry imaging (MSI). MSI allows the untargeted measurement of molecules such as drugs, metabolites, lipids, peptides and proteins directly on tissue at cellular resolution without the need for labelling. This means that no a priori knowledge of the analytes of interest is necessary; on the contrary, modern instruments allow data acquisition of thousands of molecules in parallel. The analyst only decides after data acquisition which compounds to focus on, or, with the help of modern bioinformatics, employs algorithms to identify molecules and regions of interest based on their molecular fingerprint.

Herein, we will focus on molecular MSI; alternative approaches that allow the imaging of specific elements, e.g. platinum accumulation in specific renal regions upon cisplatin chemotherapy, are beyond the scope of this article [2]. For a typical MSI experiment, tissue preparation on a microscope slide is comparable to histology (sample thickness ∼20 µm). Next, the sample is either coated with an organic acid in the case of matrix-assisted laser desorption ionization (MALDI) MSI or can be measured directly in the case of desorption electrospray ionization (DESI) MSI. The sample is then scanned either with a laser beam (MALDI) or a spray of organic solvent (DESI), allowing the generation of a mass spectrum at each point of the raster (=pixel). Each peak recorded by the mass spectrometer corresponds to a molecular mass that can be directly linked to a chemical formula and therefore to a specific compound. After data acquisition, molecule-specific images of the entire tissue can be generated for each peak in these mass spectra, which in turn represent individual biomolecules.

An example of MSI performed on rat kidney tissue is shown in Figure 1. Thousands of individual mass spectra are recorded at a spatial resolution of 15 µm. Afterwards, single molecules as here illustrated for different regiospecific phospholipids can be extracted from the dataset, and visualized.

FIGURE 1:

MALDI MSI of a rat kidney. (A) Optical microscopy image of the H&E stained tissue. (B) Overlaid and individual images for selected molecules from individual mass spectra. (C) Average mass spectrum for the entire kidney. Signals for molecules selected for visualization in panel (B) are marked (*). Reprinted with permission from reference number [3]. Copyright 2019 American Chemical Society.

Open in new tabDownload slide

A very common clinical application of MSI is to study cancer biology and progression, such as the molecular microheterogeneity of renal cell carcinoma. By using a panel consisting of 20 peptide markers, Morgan et al. [4] were able to define the margins separating healthy and tumour regions. The approach was applied to a cohort of 70 patients, resulting in predictive accuracies of >90%. Additionally, MSI might improve the definition of the tumour boundary: it has been shown that tumour markers often progress beyond the histologically defined tumour margin, indicating that biochemical changes occur before phenotypical changes are observed [5].

Another very active field of MSI research is diabetic nephropathy (DN). It has been shown that several classes of phospholipids and glycolipids are differentially expressed in the glomeruli and tubules of diabetic compared with non-diabetic mice. Employing inhibitors of lipid oxidation pathways leads to an increase in DN-associated lipid levels while improving renal pathology without affecting hyperglycaemia [6]. More recently, Smith et al. [7] investigated proteomic markers in human kidney biopsies to distinguish DN from hypertensive nephrosclerosis. Proteomic signatures associated with the progression of DN were identified, with two proteins (PGRMC1 and CO3) strongly correlated with disease progression.

MSI can also actively support pre-clinical drug development. For example, Bruinen et al. [8] studied the crystal-like structures formed in the cortex in rabbits upon application of a c-Met tyrosine kinase inhibitor as a potential anti-cancer drug. Based on histological examinations, the crystals lead to tubular degeneration, thus causing renal toxicity. Using MSI, it was demonstrated that the crystals were predominantly composed of drug metabolites (de-methylation and oxidation).

The same laboratory also published a study on ischaemic injury in renal tissue using a porcine model, in which paired kidneys received warm (severe) or cold (minor) ischaemia through perfusion [9]. In contrast to two blinded pathologists, MSI was capable of a detailed discrimination of severe and mild ischaemia by differential expression of characteristic lipid-degradation products (lysocardiolipins, lysophosphatidylcholines and lysophosphatidylinositol) throughout the tissue.

These examples represent only a small excerpt of the vast number of MSI studies that have been performed in nephrology. Nonetheless, they convincingly demonstrate the power and future potential that untargeted molecular imaging has to offer. In addition to basic and pre-clinical research, current applications also explore the use of MSI intraoperative tissue analysis. It has been shown that combined MSI and hematoxylin and eosin (H&E) images can be obtained within 5 min and, by use of machine learning algorithms, provide a surgeon with clinically actionable information based on molecular tissue signatures [10]. With the continuous improvement of MSI instrumentation, we are convinced that our understanding of renal pathologies on the molecular level enabled by MSI will improve as well.

CONFLICT OF INTEREST STATEMENT

None declared.

REFERENCES