Liver

Primary human hepatocytes are a commonly used model for investigating drug metabolism, as this system is often reflective of in vivo hepatic metabolism (Ponsoda et al., 2001; McGinnity et al., 2004). However, hepatocytes only make up approximately 80% of cells in the liver. Other cell types, including hepatic stellate cells, endothelial cells, cholangiocytes, and Kupffer cells, may also influence drug metabolism (MacParland et al., 2018; Aizarani et al., 2019). Importantly, these cells make up the complex framework of the liver and express a degree of intrapopulation heterogeneity themselves, which enables the multifunctionality of the organ (MacParland et al., 2018; Marcu et al., 2018; Payen et al., 2021).

Although a significant proportion of the body’s drug metabolism can be attributed to the liver, several other organs, including the kidneys, intestines, lungs, and brain, express metabolizing enzymes, such as the cytochromes P450 (P450s), UDP-glucuronosyltransferases (UGTs), sulfotransferases (SULTs), N-acetyltransferases (NATs), and others (Table 1). Notably, extrahepatic drug metabolizing enzymes can impact drug bioavailability and pharmacokinetics in both an organ and cell-type-specific manner (Paine et al., 1996; Zhang et al., 2007; Fonsi, 2014; Shibata and Chiba, 2015). For example, Shibata and Chiba observed significantly higher predicted bioavailability based on hepatocyte metabolism than calculated bioavailability in the plasma of rats, dogs, monkeys, and humans (Shibata and Chiba, 2015).

Kidney

Drug metabolism in the kidneys has been extensively studied and has been shown to contribute considerably to the biotransformation of a variety of different drug classes, including immunosuppressants, such as cyclosporin A, sedatives, such as propofol, and antiretrovirals, such as zidovudine (antiretroviral) (Dai et al., 2004; Hiraoka et al., 2005; Knights et al., 2016). Using scRNA-seq, ten distinct cell types were classified in adult human kidneys. Of these, glomerular parietal epithelial cells and proximal tubule cells accounted for 80% of the cells present (Liao et al., 2020). Primary human proximal tubule cells have been noted to express UGT1A1, UGT1A6, and UGT2B7, as well as CYP1B1, CYP3A4/5, and CYP4A11; however, the only enzymatic activity that was directly observed was that of CYP3A4/5 (Lash et al., 2008). UGTs 1A8/9 and 2B7 have been observed in primary renal epithelial and cortical renal epithelial cells, in addition to proximal tubule cells, but at significantly lower levels than in renal tissue as a whole (Van der Hauwaert et al., 2014). In immortalized cell lines, such as HK2 (proximal tubular cell line) and human embryonic kidney 293 cells, UGT transcript expression has not been detected (Van der Hauwaert et al., 2014). Differences between cellular and microsomal/tissue expression data suggest that there could be a range of cell types contributing to renal metabolism. These cells may be overlooked or underappreciated in bulk analyses; however, single-cell technologies have the potential to fill this gap when studying renal metabolism, as well as drug metabolism in other regions of the body.

Intestine

Drug-metabolizing enzymes are also expressed throughout the intestines, and have been found to contribute to the overall clearance of drugs, such as cyclosporin, midazolam, and docetaxel (Kolars et al., 1991; Thummel et al., 1996; van Herwaarden et al., 2007). RNA sequencing of human intestinal epithelial cells revealed the presence of several phase 1 and phase 2 drug metabolizing enzymes, with notable differences in transcript expression between the duodenum, ileum, colon, and rectum. In the ileum, the most abundant P450 transcripts observed were CYP3A4, CYP2C18, CYP3A5, and CYP2J2 (Takayama et al., 2021).

Enterocytes, the major subtype of intestinal epithelial cells, from the duodenum exhibit nonspecific UGT and SULT activity as well as CYP3A4, CYP2C9, CYP2C19, and CYP2J2 activity (Ho et al., 2017). Of the four intestinal regions examined, these enzymes were more highly expressed in the ileum and duodenum than in the colon or rectum. This region-specific expression also included UGTs and SULTs, where UGT8, UGT1A1, UGT2B7, UGT2A3, and SULT2A1, SULT1B1, SULT1A2, and SULT1A1 were the top four enzymes of each class identified in the ileum (Takayama et al., 2021). An orthogonal study employing immunoquantification in tissue from the ileum and duodenum identified many of the same SULTs, but at slightly different proportions, confirming that whole tissue analyses do not necessarily reflect the protein abundance in specific cell populations (Riches et al., 2009).

Intestinal cellular models, such as Caco-2 cells, have shown low transcript expression of CYP3A4, which can be artificially induced, but limited endogenous expression of drug metabolizing enzymes make it a poor model for exploring intestinal drug metabolism (Sergent et al., 2009; Li, 2020). Intestinal metabolism has also been explored using precision-cut intestinal slices as a more comprehensive approach than classic in vitro models, such as cultured cells or microsomes (de Graaf et al., 2010). In precision-cut intestinal slices taken from both the ileum and the colon, CYP3A and SULT1 metabolite formation was higher in the ileum, while CYP2A and UGT1A metabolite formation was higher in the colon, indicating regional variability (Iswandana et al., 2018). However, this method does not allow for the discernment between different cell types. Understanding cell-type-specific metabolism within different intestinal regions is especially relevant for regimens that are delivered rectally or drugs that specifically target the intestines.

Lung

Regional differences in metabolizing enzymes have also been noted in the lung, another extrahepatic organ in which local metabolism may have significant effects on overall drug biotransformation, particularly for medications administered via inhalation. scRNA-seq has characterized 58 different cell types in the human lung, encompassing epithelial, endothelial, stromal, and immune cell populations (Travaglini et al., 2020).

Parenchymal cells, part of the alveolar epithelial population, have been found to express the following P450 enzymes: CYP1A1, CYP1A2, CYP2A6, CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP2J2, CYP3A4 and CYP3A5, all at a fraction of the expression levels found in the hepatocytes. However, phase 2 metabolizing enzymes SULT1A1, SULT1A3/4, SULT1C2, SULT2B1/2, and UGT2A1 were expressed more highly in the parenchymal cells. Several other UGTs and SULT1A2 were also observed, but with lower relative expression to hepatocytes (Somers et al., 2007).

Further, primary human bronchial epithelial cells have been noted to express up to 27 different P450 isoforms, including CYP1A1, CYP2C9, CYP2J2, and CYP3A5, which were also observed in alveolar epithelial cells (Courcot et al., 2012). Courcot et al. also presented phase 2 metabolizing enzyme mRNA expression of UGTs 1A1 and 1A6, NAT5, and SULTs 1A3/4, 1B1, 1E1, 2B1, and 4A1 (Courcot et al., 2012). Of particular interest, transcripts of P450s 1A1 and 1B1 have been observed at higher concentrations in the lungs of smokers and ex-smokers as opposed to non-smokers, indicating that environmental and/or health-related events could impact drug metabolism in the lung (Kim et al., 2004).

Brain

The brain is perhaps the most pharmacologically elusive and heterogenous of the organs discussed here. When considering the presence of drug metabolizing enzymes, there are not only several brain regions, but a multitude of cell types that must be taken into account. In the cortex alone, there have been 180 perceptible areas (Glasser et al., 2016) and 75 cell types identified (Hodge et al., 2019). Between the frontal cortex, striatum, hippocampus, thalamus/hypothalamus, basal ganglia, substantia nigra, amygdala, pons, and cerebellum, the frontal cortex and cerebellum are the most expressive of P450 enzymes, with 11 and 10 P450s identified, respectively, while the thalamus and hypothalamus have only two P450 enzymes expressed: CYP2D6 and CYP3A4 (McFadyen et al., 1998; Toselli et al., 2015; Kuban and Daniel, 2021). Although CYP2B6 mRNA has been observed in five of the different brain regions explored (frontal cortex, striatum, basal ganglia, pons, and cerebellum), a study examining the microsomal metabolism of various CYP2B6 substrates — nicotine, ketamine, and methadone — found no evidence of enzyme activity (Bloom et al., 2019).

Importantly, CYP2E1 plays a role in the metabolism of ethanol in the brain (Carpenter et al., 1996; Upadhya et al., 2000; Zimatkin et al., 2006; Zhong et al., 2012; Toselli et al., 2015) and has also been linked to Parkinson’s disease (Shahabi et al., 2008; Kaut et al., 2012). CYP2E1 mRNA is found primarily in neurons, but has also been observed in Purkinje cells of the cerebellum (Upadhya et al., 2000). In a study of the blood-brain barrier (BBB), CYP1A1, CYP1B1, CYP2A6, CYP2B6, CYP2C9, CYP2E1, CYP2J2, CYP3A4, and CYP4A11 mRNA were observed in primary human brain endothelial cells, each at a significantly higher level than commercially available immortalized human brain microvascular cerebral endothelial cells and human-derived umbilical vein endothelial cells (Ghosh et al., 2010). Further study of human brain microvessels, however, revealed that CYP1B1 and CYP2U1 were the only two P450s able to be quantified at the protein level (Shawahna et al., 2011).

Regarding phase 2 metabolizing enzymes, UGTs and NATs were not observed in the microvasculature of the BBB, but SULT1A1 was expressed (Shawahna et al., 2011). UGT mRNA expressed in brain tissues includes UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A9, UGT2A3, UGT2B4, UGT2B7, UGT2B10, UGT2B11, UGT2B15, UGT2B17, and UGT2B28 (Court et al., 2012; Jones and Lazarus, 2014). Both UGT1A6 and UGT2B7 mRNA have been observed specifically in the cerebellum (King et al., 1999). UGT2B7, the enzyme responsible for the biotransformation of morphine (Coffman et al., 1997), showed activity in some, but not all, human brain microsomes tested in a study by Wahlstrom et al. (Wahlstrom et al., 1988). Lastly, the protein of SULTs 1A1 and 1A3 have been identified in neurons and glial cells from the hippocampus, cerebellum, and frontal, occipital, and temporal lobes (Salman et al., 2009). Overall, these data indicate the presence of drug metabolizing enzymes in specific regions of the brain, which is of particular relevance to local metabolism of drugs that cross the blood-brain barrier.

Moving Forward

While the liver is the primary site of metabolism of most drugs, extrahepatic organs, such as the kidneys and the intestines, have been shown to contribute to the overall clearance of certain drugs. Additionally, extrahepatic metabolism has the potential to impact pharmacology when target tissues exhibit local metabolism. Further, with the complexity of cellular heterogeneity in pathophysiological states, such as tumorigenesis and the sensitivity of cell state to the microenvironment, investigation of cell-type-specific drug metabolism could reveal unique insight into cellular phenotype in these contexts (Wilting and Dannenberg, 2012; Kirschner et al., 2020; Guan et al., 2021; Xu et al., 2022). We envision the utilization of single-cell technologies in drug metabolism research will aid in more accurate targeting and dosing for a wide variety of treatments.

Current literature primarily uses scRNA-seq to address differences in the expression of drug metabolizing enzyme transcripts between different cell types, but this strategy is most powerful when employed in conjunction with other single-cell technologies. Few studies have sought to uncover intercellular heterogeneity of drug metabolizing enzymes in extrahepatic metabolizing organs, and the recent development of a variety of single-cell analysis tools will be indispensable in remedying this gap in knowledge.

Current and Emerging Single-Cell Methods

While some progress has been made using scRNA-seq to study drug-metabolizing enzyme transcript expression, there is still little known about drug metabolism and action in a single cell. The following methods can be employed to ascertain cell-type heterogeneity in metabolic activity and cellular response to aid in pharmacological research (Table 2).

RNA Transcripts

Transcriptomics at the single-cell level has been used as a means to determine therapeutic targets and to identify genes related to drug tolerance. Drop-seq is a technique used to analyze mRNA of individual cells and has been applied to the investigation of genes associated with drug-tolerance (Macosko et al., 2015; Aissa et al., 2021). Using this method, Aissa et al. employed PC9 cells to define subpopulations of drug-tolerant cells after treatment with erlotinib, a tyrosine kinase inhibitor, and characterized previously unidentified drug-tolerant markers. In determining drug-tolerant markers in subpopulations, Aissa et al. identified drugs that could downregulate such tolerance markers, sensitizing drug-tolerant cells to erlotinib treatment, a feat not possible using classic bulk cell RNA sequencing technologies. Further, their data determined that higher transcript expression of tolerance markers correlated with worse patient outcomes (Aissa et al., 2021). Interestingly, drug tolerance was shown to be continual such that cells can transition between tolerance states, which is an important finding relevant to improving therapeutics. Overall, transcriptomics can be used to characterize subpopulation tolerances that can better inform personalized treatments, as well as identify new targets for drug discovery.

In recent years, bioinformaticians have generated various scRNA-seq atlases and databases, as the curation of datasets facilitates the continual analysis of published data. Several scRNA-seq databases have been established, such as the EMBL-EBI Single Cell Expression Atlas, which as of March 2022 includes 304 studies, encompassing 20 species and over 8.5 million cells (Moreno et al., 2022). Other initiatives, such as Gene Expression Omnibus (Barrett et al., 2009) and CellDepot (Lin et al., 2022), provide raw sequencing and visually processed data from single-cell sequencing data, respectively. In utilizing these databases, pharmacologists can uncover the transcript expression of drug-metabolizing enzymes in specific cell populations and identify potential therapeutic targets.

Proteins

While measurements of transcript abundance are valuable and have the potential to be integrated with other -omics data sets, single-cell protein abundance and localization data can further improve our understanding of complex biologic systems and lend insight into drug disposition. Of note, flow cytometry can be used to characterize heterogeneity in cell surface markers and intracellular proteins in many different biologic systems. However, this technique is limited in multiplexing capabilities and does not allow for characterization of protein localization within cells and intact tissues (Tracey et al., 2021).

Mass cytometry (also referred to as CyTOF) is a complementary method to flow cytometry that can produce comprehensive data sets. Like flow cytometry, CyTOF uses antibodies to cell surface markers and intracellular proteins to enable detection. However, while flow cytometry multiplexing is limited by spectral overlap of fluorophores, CyTOF uses antibodies conjugated to heavy metals and inductively coupled mass spectrometry to measure the abundance of these metals in each cell. This technique allows for routine 40-plex experiments and theoretically enables 100-plex studies.

Similar to scRNA-seq and flow cytometry, CyTOF has been used to study the immune tumor microenvironment (Norton et al., 2019; Zhang et al., 2019; Zhu et al., 2021). Zhu et al. were able to use such data to link AKT activity to chemotherapy resistance in breast cancer cells and showed that AKT inhibition increased cell death (Zhu et al., 2021). CyTOF provides a unique profile of a tumor and has been used to develop a computational framework to identify the optimal combination therapy for a patient (Anchang et al., 2018). Additionally, Zhang and colleagues incorporated CyTOF data into a multi-omics analysis including RNA-seq and metabolomics data to study the heterogeneity of hepatocellular carcinoma (Zhang et al., 2019).

The same antibodies that enable high multiplexing in CyTOF can also be used to build upon classic immunohistochemistry in imaging mass cytometry and multiplexed ion beam imaging (MIBI) experiments. Both methods involve labeling a tissue section with metal-conjugated antibodies and detecting these metals by mass spectrometry. To transfer the metals from the tissue to the mass spectrometer, imaging mass cytometry and MIBI use laser ablation and a beam of primary ions, respectively. Both techniques can achieve spatial resolution at or below 1 µm, rendering them capable of imaging at the single-cell level. Employing MIBI technology, Keren et al. characterized breast cancer tumor immune microenvironments using 36 markers simultaneously at 260 nm spatial resolution (Keren et al., 2019).

An emerging multiplexed protein imaging technique is co-detection by indexing (CODEX). CODEX utilizes oligonucleotide-conjugated antibodies and involves iterative in situ polymerization to selectively incorporate dNTP-tethered fluorophores on specific target barcodes. Following imaging, fluorophores are removed and additional rounds of in situ polymerization are completed, theoretically allowing for limitless protein detection. Goltsev and colleagues used CODEX to characterize the architecture of normal and diseased mouse spleens using 28 antibodies on the same tissue section (Goltsev et al., 2018).

Typically, these single-cell methods are used to compare the presence of surface markers and occasionally intracellular proteins. However, these methods can be used to inform knowledge of drug metabolism and pharmacokinetics by applying them to the study of drug transporters and intracellular drug metabolizing enzymes.

MS-based proteomics removes the need for cell surface markers to characterize proteins of a single cell. Combined with other biochemical methods, mass spectrometry can deliver data on regulatory correlation, signaling pathways, metabolic processes, and posttranslational modifications, all of which are factors that can drive cellular pathobiology and drug response (Olsen and Mann, 2013; Carvalho and Matthiesen, 2016; Marx, 2019; Ctortecka and Mechtler, 2021; Meissner et al., 2022).

The standard bottom-up proteomics workflow includes cell lysis, protein extraction and solubilization, reduction of disulfide bonds and alkylation of cysteines, digestion of proteins to peptides, sample desalting, clean-up, chromatographic separation, MS acquisition, and analysis (Aebersold and Mann, 2003; Bantscheff et al., 2007; Gundry et al., 2009). Currently, single cells are isolated by fluorescence-activated cell sorting or microscopy-based sorting techniques into multi-well plates, where lysis, digestion, and multiplex labeling, if employed, are performed. To date, multiple methods for single-cell proteomics processing have been described, including minimal proteomic sample preparation (Budnik et al., 2018), nanodroplet processing in one pot for trace samples (nanoPOTS) (Zhu et al., 2018), and Single Cell ProtEomics by Mass Spectrometry using the cellenONE automated single cell seeding platform (Vallone et al Rinke et al., 2014; Hartlmayr et al., 2021).

The nanoPOTS platform (Zhu et al., 2018) consists of microfabricated nanowell array chips and a nanoliter robotic pipetting system (Fig. 1A). The nanoPOTS platform demonstrated enhanced efficiency and sample recovery by downscaling processing volumes to < 200 µl to minimize surface losses (Zhu et al., 2018). Recent trials of the nanoPOTS platform have identified and quantified over 1,000 proteins from single mammalian cells (Kelly et al., 2019).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Three sample preparation strategies for single-cell proteomic analysis. (A) The nanoPOTS platform uses a nanoliter robotic pipetting to dispense cells and reagents onto microfabricated nanoarray chips (Zhu et al., 2018). (B) The ScoPE2 method uses TMT isobaric tags to label peptides of single cells and a 200-cell channel to boost protein sequence coverage (Petelski et al., 2021). (C) Whole-cell patch-clamp electrophysiology involves sample injection into a patch pipette, bottom-up proteomics on collected protein, and identification and quantification using a capillary electrophoresis electrospray ionization platform followed by high-resolution mass spectrometry (Choi et al., 2021).

Modern proteomics often relies on the use of chemical “barcodes” to facilitate sample multiplexing, allowing up to an 18-fold increase in sample throughput. Specifically, isobaric tandem mass tags (TMTs) have an identical total mass while heavy isotopes are distributed differently across the tag (Budnik et al., 2018; Myers et al., 2019). Thus, the fragmentation of precursor ions generates TMT reporter ions with isotopic loading unique to each TMT tag, and quantification of TMT-labeled peptides relies on reporter ions whose levels reflect both peptide abundances and noise contributions. Multiplexed peptides have an additive signal when combined, which enhances low-level samples, such as single cells.

Single Cell ProtEomics by Mass Spectrometry (SCoPE2) further exploits the additive nature of multiplexed peptides through the use of a single labeled channel that contains a significantly higher peptide concentration (Fig. 1B). This “carrier” or “boost” channel provides sufficient signal to trigger fragmentation enabling peptide sequencing, while the isobaric tags for each single cell are sufficient for obtaining relative peptide quantification. Recent work by Petelski et al. described the preparation of single mammalian macrophages, where all stages of preparation were carried out in a total volume of 1 µl (Petelski et al., 2021). Minimal sample loss and efficient sample preparation enabled over 1,000 proteins to be identified per cell.

Our group recently optimized the multiplexed single-cell proteomics methods from SCoPE2 and analyzed human-derived single-cell samples using trapped ion mobility time-of-flight mass spectrometry (Orsburn et al., 2022, preprint). When employing a peptide carrier channel to boost protein sequence coverage, we obtained over 40,000 tandem mass spectra and identified over 1,000 proteins in 30-minute LC gradients, providing unprecedented sequence coverage of each identified protein. We described the first observations of multiple classes of protein posttranslational modifications (PTMs) using liquid chromatography-mass spectrometry (LC-MS) (TIMSTOF Flex) in single human cells. PTMs play an essential role in various cellular processes that modulate the folding, stability, interaction, and functional activity of proteins (Su et al., 2017). Understanding the impact of PTMs on proteins, especially near active sites and the interfaces of protein-protein interaction, is critical for ongoing drug development research and can be leveraged to improve the efficacy of future drug design (Meng et al., 2021). Precisely targeting PTMs or protein isoforms can advance personalized treatment and precision medicine (Meng et al., 2021).

While these methods rely on traditional LC-MS, other groups have explored alternative techniques to further intensify peptide signal. Choi et al. described a custom-built capillary electrophoresis ionization platform using high-resolution mass spectrometry (Choi et al., 2021) (Fig. 1C).

This platform combines patch-clamp electrophysiology with subcellular CE-electrospray ionization-high-resolution mass spectrometry to quantify proteins in targeted neurons in brain tissues. Analysis of 1 picogram of protein digest quantified 157 proteins among the somas of three different dopaminergic neurons from a mouse (Choi et al., 2021). Additionally, Webber et al. optimized a label-free single-cell proteomics method which mitigates the slow MS2 sampling rate (Webber et al., 2022). A two-column LC system provides two parallel subsystems to multiplex sample loading, desalting, analysis, and column regeneration, which doubles the throughput as an average 621, 774, 952, and 1622 protein groups were identified with total analysis times of 7, 10, 15, and 30 minutes, respectively (Webber et al., 2022).

Small Molecule Measurements

An advancement in single-cell analysis that directly impacts pharmacology is the quantification of intracellular drug concentrations. This approach provides data regarding drug delivery, influx, and efflux at the intended target cell (i.e., CD4+ T-cells in the treatment and prevention of HIV) and can offer a real-time readout of pharmacokinetic parameters. Drug concentrations in single cells can be quantified under ambient conditions using single-probe mass spectrometry. In this technique, a probe containing dual capillaries allows for the delivery of sampling solvent and internal standard to a cell and the withdrawal of cellular extracts by capillary action. The extracts flow through a nano-electrospray ionization emitter, and ions are detected by mass spectrometry (Pan et al., 2019). Recent developments have coupled the single-probe system with an integrated cell manipulation platform, allowing for this technique to be used on cells in suspension, thereby increasing the applicability. Using this method, intracellular drug levels determined in single bladder cancer cells from patients treated with gemcitabine revealed significantly different intracellular concentrations of the drug between two patients, even after multiple rounds of treatment (Bensen et al., 2021), unveiling a need for precision medicine.

Single-probe single-cell mass spectrometry can also be used to understand the metabolic changes in cells upon treatment with drugs. These data can provide a key understanding of the drugs’ mechanisms of action and how cells can become resistant to therapeutics. This technique has been used to understand the synergism of metformin and irinotecan in an induced-irinotecan-resistant colorectal cancer cell line, where it was shown that the combination therapy had the largest synergistic effect on the reduced abundance of ceramide phosphoethanolamine, likely due to a reduction in enzymatic activity of fatty acid synthase, an enzyme overexpressed in cancer cells (Chen et al., 2022). Using this technique to identify changes in metabolite levels in single-cells upon treatment with a drug can assist in understanding the pharmacodynamics and eventual development of new combination therapies.

Another technique in development to quantify drugs in single cells is capillary electrophoresis coupled with ultrasensitive laser-induced fluorescence detection. In this method, the native fluorescence of drug molecules is exploited to quantitate single-cell drug levels. Importantly, this method can be used in tandem with flow cytometry to correlate intracellular drug levels to the abundance of cell-surface transporters, such as p-glycoprotein (Deng et al., 2011). Studies using this technique identified cell-to-cell variability in drug uptake and how dosing schedules and p-glycoprotein can impact drug uptake heterogeneity, which are factors that can have clinical implications (Deng et al., 2014).

Imaging

Finally, to decipher the molecular diversity of single cells, powerful imaging techniques, such as mass spectrometry imaging (MSI), can be used to map the localizations of a range of biomolecules including small metabolites, lipids, peptides, proteins, drugs, and drug metabolites. One of the major advantages of MSI over other imaging techniques is that it can be used to visualize the spatial distribution of hundreds of biomolecules simultaneously without labeling (Buchberger et al., 2018). The main ionization techniques used for MSI are matrix-assisted laser desorption/ionization (MALDI), desorption electrospray ionization, laser ablation electrospray ionization, and secondary ion mass spectrometry (SIMS) (Buchberger et al., 2018). Of these ionization techniques, MALDI MSI is the most frequently employed for simultaneous visualization of the spatial distribution of drugs, their metabolites, and endogenous lipid molecules at tissue level (Prideaux et al., 2010; Seneviratne et al., 2018; Tang et al., 2019; Seneviratne et al., 2020). Due to recent improvements in MALDI MS sources, this approach can also be used for single cell imaging. Several research groups have demonstrated the use of MALDI MSI to image analytes at 5 µm spatial resolution (Zavalin et al., 2013; Korte et al., 2015). However, the observed ion yields were low, and the investigators had to partially overlay individual pixels to remedy this problem (Korte et al., 2015). Technological and methodological advancements in MALDI MSI have enabled high spatial resolution; however, sensitivity is a major challenge. To overcome this, laser desorption/ionization is employed, as it avoids spectral interference due to the presence of a chemical matrix (Stopka et al., 2016). Currently, SIMS is the predominant modality for single cell MSI, and it uses a primary ion beam instead of a laser for sample ablation and ionization. Because of this, it can achieve greater spatial resolution and sensitivity than either MALDI or laser desorption/ionization. Interestingly, Malherbe et al. demonstrated that a radio frequency plasma oxygen primary ion source coupled to NanoSIMS can image endogenous cations (e.g., calcium) at a 37 nm spatial resolution (Malherbe et al., 2016). While SIMS is typically coupled to time-of-flight mass spectrometers, this ionization method can also be coupled to an orbitrap mass spectrometer to achieve higher mass resolution and improve identification of biologic molecules (Passarelli et al., 2017).

Single-cell-level imaging of endogenous metabolites can improve our understanding of the natural heterogeneity of tissues and cell types as well as enhance research in disease biology and drug discovery. These same tools can be used to study patterns of drug absorption, distribution, and metabolism. In addition to lipids and neurotransmitters, Passarelli et al. were able to detect and image the distribution of amiadarone, an antiarrhythmic drug, using OrbiSIMS (Passarelli et al., 2017). SIMS can achieve spatial resolution closer to 200 nm when imaging heavy metals and has been leveraged to image the distribution of platinum-containing drugs (Lee et al., 2017; Proetto et al., 2018).

Recently, there has been a trend toward the combined use of multiple imaging modalities (Veličković et al., 2018; Tuck et al., 2022). For example, imaging mass cytometry can obtain snapshots of tissue and cell microenvironments, and when coupled with MSI, may facilitate the correlation of drugs, drug metabolites, and endogenous molecules with cellular markers. Further, coupling microscopy and mass spectrometry can help unravel biologic processes.

Van Acker et al. demonstrated a direct comparison between fluorescence confocal microscopy and laser ablation-inductively coupled plasma-MSI techniques through the use of hybrid tracers targeting CXCR4 and EGFR (Van Acker et al., 2019). Additionally, Schoeberl et al. developed a high-throughput laser ablation-inductively coupled plasma MS workflow for single-cell analysis following cytospin preparation of cells (Schoeberl et al., 2021). Using this technique, the authors investigated cisplatin, an anticancer drug, uptake in macrophage subtypes (Schoeberl et al., 2021). The above innovative multimodal imaging strategies can be employed broadly to high-resolution imaging of drugs/metabolites and single-cell analysis in precision medicine.

In addition to these original investigations, the results of single cell studies are often available for meta-analysis through public data repositories. Today, nearly all mass spectrometry journals require that the original data and processed results are made directly accessible through the ProteomeXchange Consortium of global public repositories (Vizcaino et al., 2014). Recent efforts by the European Bioinformatics Institute repository PRIDE have enabled the public deposition of MALDI imaging data when converted to a universal XML format (Rompp et al., 2015).

With the ongoing evolution of bioinformatics tools and computational techniques, the exploration of publicly deposited single cell data will likely lead to further discoveries in addition to those made in the original studies.