Approaching complexity: systems biology and ms-based techniques to address immune signaling

Abstract

Introduction:

Studying immune signaling has been critical for our understanding of immunology, pathogenesis, cancer, and homeostasis. To enhance the breadth of the analysis, high throughput methods have been developed to survey multiple areas simultaneously, including transcriptomics, reporter assays, and ELISAs. While these techniques have been extremely informative, mass-spectrometry-based technologies have been gaining momentum and starting to be widely used in the studies of immune signaling and systems immunology.

Areas covered:

We present established proteomic methods that have been used to address immune signaling and discuss the new mass-spectrometry- based techniques of interest to the expanding field of systems immunology. Established and new proteomic methods and their applications discussed here include post-translational modification analysis, protein quantification, secretome analysis, and interactomics. In addition, we present developments in small molecule and metabolite analysis, mass spectrometry imaging, and single cell analysis. Finally, we discuss the role of multi-omic integration in aiding leading edge investigation.

Expert opinion:

In science, available techniques enhance the breadth and depth of the studies. By incorporating proteomic techniques and their innovative use, it will be possible to expand the current studies and to address novel questions at the forefront of scientific discovery.

1. Introduction

Don’t miss the forest for the trees

– English proverb of unknown origin, prior to 1546.

For decades immunology discoveries have been driven by the available assays, with the remarkable contribution of flow cytometry, often focusing on a specific point or a single variable. This reductive approach has been beneficial and led to seminal discoveries that shape our understanding of every aspect of immunology but also led to a focus on specific ‘trees’ that limit our understanding of the entire ‘forest’.

Systems biology addresses this limitation, allowing to ‘visualize the forest’. By combining high-throughput experimental techniques with bioinformatic skills, systems biology examines whole cells or the interactions of multiple cells with a holistic approach rather than the traditional reductive approach. The growing view of systems questions requiring systems answers has crafted a cooperative space for a multi-omic approach to immunology and cell signaling. While the reductive methods have been extremely informative, immune signaling presents challenges for these methods because of the complex interactions between cells, and the complex series of pathogen-elicited signals that propagate through cytosol via multilayered signaling networks involving proteins, protein complexes, modifications and small molecules.

Some of the methods in systems biology experiments are reporter assays, mRNA screens, and ELISA-based assays. These assays all measure the downstream effects following the introduction of a variable such as a mutation or knockdown of a specific gene, stimulation, or introduction of an inhibitor. While focusing on the downstream effects can improve our understanding of the upstream signaling involved in an immune response when used in conjunction with a properly designed experiment, these experiments can be difficult to interpret due to the complexity of the pathways involved or if the specific proteins involved in the pathway are unknown or novel. The complexities of a whole biological system, like those of model organisms, make interpretation of these downstream effects difficult to interpret due to the overwhelming number of variables. This leads to increasing numbers of assumptions required for valid experimentation, but with the addition of each assumption we stray further from biological relevance.

2. Common non-proteomic tools

Complementary macromolecule interactions have been the basis of cellular and molecular experimentation in the reductive approach that has dominated science. Interaction specificity lends confidence in results, a common aspect of the three non-proteomic methods we outline here, but that same dependence on interactors limits their assay range and resulting biological relevance; as is the case in reporter based assays and ELISAs, while transcriptomics lend context to biological mechanics, but focusing on the RNA, are not applicable for the analysis of the protein-protein interactions which drive cellular actions. Prior to presenting the proteomic methods, we believe it is important to briefly introduce these methods for proper context. We acknowledge the importance and contributions of flow cytometry to immunology, but the description of all the aspects of this extremely wide field is beyond the scope of this review and has been a topic of many excellent recent overview articles, for example [1].

2.1. Reporter-based assays

Reporter assays constitute the in vitro stimulation of a promoter and reporter gene pair to measure a response [2]. These assays can be extremely useful and are a robust method for high throughput screens of different drugs or compounds by the direct measurement of expressed gene products or the indirect changes in signal molecules regulated by genes. This is accomplished by transfecting a desired cell line to express a transcription factor activated promoter and a common reporter gene. The cells can then be stimulated to drive promoter activity. Ligand or agonist induction of the transcription factors leading to expression of the reporter gene. In the case of a luciferase reporter gene, luciferin, a luciferase substrate, can be added and the resulting luminescence measured as a corollary of stimulation. The unattractive variables of reporter assays tend to revolve around the limits in experimental design: reporter assays may take vast resources and long time periods to design and to target assays to transcription factors. Common reporter assays exploit the activity of reporter genes producing β-galactosidase, green fluorescent protein (GFP), and luciferase [2].

2.2. Transcriptomics

Transcriptomics as a field covers a wide range of techniques all meant to investigate the RNA transcripts in a cell. Methods like microarrays and RNAseq can provide insight into the set of RNA molecules transcribed and present in a single cell or across biological sample. This provides hints into the genome expression in different types of cells. Microarrays function by the matching of an array of predefined complementary probes to transcripts to quantify their abundance. Using sample probes from the target organism, a transcriptional profile can be made by arraying RNA probes together. RNAseq works by creating a cDNA library, from the RNA sample, then addition of sequence adaptors followed by high throughput sequencing of the sample [3]. These techniques are effective, quick, and inform the mechanics of drug treatments or disease states on the genome. Major drawbacks of transcriptomic approaches are the loose association between mRNA levels and actual protein expression, as well as the lack of information on proteoforms or other regulatory mechanisms not encoded in the transcriptome. Microarrays and RNAseq are attractive to researchers due to their relatively low cost and high throughput abilities and many labs will successfully continue to incorporate transcriptomics into their research [4].

2.3. ELISA

Enzyme-Linked Immunosorbent Assays (ELISA) are used to determine the concentration of an antigen or antibody based on its binding to a plate bound antibody or antigen respectively, and the resulting change in color of a substrate interacting with enzymes linked to the target agent. The selectivity and confidence in antibody-based assays make ELISAs a standard for many common tests, from measuring cytokine concentration to many serological tests for disease states. ELISAs can accurately measure the amount of proteins produced by the cells in response to stimuli and can be used with secreted proteins [5]. There are certain downsides of ELISA assays: rigidity of antibodies confers recognition limitations; the presence of a known protein of interest is required and an antibody against said target has to be available to the user beforehand. These drawbacks make the prerequisite immunological knowledge of the sample a high bar for any discovery based investigations [6].

2.4. Systems biology studies

Two studies utilizing reporter assays to analyze the innate immune and inflammatory signaling response were published by Tay et al. [7], and Sung et al. [8]. Both used real time fluorescent measurements in conjunction with transcriptomics to analyze the response of single cells to stimuli.

Tay et al. examined the response to tumor necrosis factor (TNF)-α at the single-cell level by first utilizing a fluorescent reporter chimera of NF-κB to measure the response to increasing concentrations of TNF-α over a 7-hour time course. By examining the cells at the single cell level, they found that the fraction of cells stimulated by the TNF-α was proportional to the concentration added. In addition, the long time course of live measurements showed the NF-κB oscillating between the nucleus and cytoplasm with the rate of oscillation being TNF-α dose-dependent. To analyze the effects of the stimulation, transcriptomics was employed to measure the mRNA produced during the response. Combining these data with mathematical modeling, allowed to generate a model for the individual cell responses to TNF-α stimulation that faithfully reproduced the individual cellular responses to stimulation [7].

Sung et al. examined the macrophage response to lipopolysaccharide (LPS) utilizing a dual reporter system measuring the responses of RelA and TNF-α promoters. This system used fluorescent reporters downstream of each promoter to measure the responses over a 24-hour time course at the single cell level. The single cell measurements allowed to show that in contrast to previous studies that analyzed the response of nonimmune cells, which showed the portion of cells responding to LPS is proportional to the amount of LPS, all of the macrophages in the sample responded to the LPS addition but the strength of their response was proportional to the dose used. Using transcriptomics, they identified the genes induced following lower and higher doses LPS stimulation and found that genes linked to a positive feedback loop were specifically stimulated by the higher doses. Analysis of this this subset of genes, led to the identification of the transcription factor Ikaros (Ikzf1) as the regulator of the positive feedback loop [8].

Both studies make effective use of reporter genes to characterize the responses of single cells from populations followed by downstream analysis of the cellular responses with transcriptomics. However, each study is limited in a couple of potentially key areas. First, while the production of mRNA is correlated to the amount of the encoded protein, the exact quantity of the protein is dependent upon many variables beyond the amount of mRNA. Second, while the time courses can be extended to hours following the stimulation, the transcription (mRNA readout) or secretion of newly transcribed proteins are slow compared to the modification of existing proteins. Lastly, while the phenotypes following deletion of a protein from a cell can indicate the protein’s function, the exact role of the protein (i. E. its binding partners, substrates, or regulators) can only be inferred through previous knowledge or comparison with similar proteins or systems when using such broad tools.

If the only tool you have is a hammer, everything looks like a nail

– adapted from Abraham Maslow, 1966

3. Introduction to the screwdriver (Proteomics)

The theme of a toolbox is quite fitting for researchers approaching both immunology and systems biology. A problem may require many tools to fix, and often we bring only a metaphorical ‘hammer’ to fix these problems. Maslow’s Hammer (the law of the instrument), is a concept of a bias toward a familiar tool; if all you have is a hammer, you treat everything as a nail. It is the hope that a holistic approach to immunology, incorporating many of the methods outlined below will help illuminate the emergent properties of immune cell signaling. While all three of methods described above have been and will continue to be critically important tools for biology, they may not be the only tool for every experiment.

Proteomics is the large-scale study of proteins, or the study of proteomes, the complete sets of proteins expressed by the organism, cell, or any studied system. The role of proteomics in the realm of immune cell signaling invites some review and highlighting. Genomic and transcriptomic approaches have been very important to understanding the reasons behind several immune mechanisms. A more proximal descriptor of the mechanisms at play in the immune system may require input from a proteomics viewpoint and integration of these viewpoints to offer the full explanation. Studying the complexities of the human immune system require a close analysis of the proteins responsible for the effector mechanisms and signaling pathways controlling our response to immune challenges. Proteomics offers new abilities that can enhance previously performed experiments or address previously impossible experiments.

Although proteomics was long thought to be expensive and laborious, the increased number of reliable reagents and protocols published have improved both the quality of sample preparation, and decreased its cost. Another common complaint about proteomics is the analysis of the data is too complex for the researchers outside the field. LC-MS/MS spectra contain a wealth of information, but they are still difficult to read directly. Bioinformatic tools help process these files into easy to analyze tables with protein identifications, quantifications, and modifications. In addition, a plethora of newly developed software tools (whose detailed description is beyond the scope of this review) is now available to assist researchers in turning these protein catalogs into comparisons and to map the results to pathways. These improvements in both ease and cost of using proteomics tools have made proteomic analysis much more approachable than ever before. The most comprehensive studies are performed with the broadest range of tools, so incorporating proteomic methods and their possible uses for the analysis of immune signaling is vital.

Proteomics offers more and more broad strategies for approaching the study of proteins (Figure 1). Bottom-up proteomics is by far the most common approach and entails the digestion of protein mixtures with proteolytic enzymes, peptide separation via chromatography, and the creation of a mass spectra in tandem mass spectrometry through peptide fragmentation [9]. This approach is carried out on protein mixtures, and is the most common variety of bottom-up proteomic analysis, known as shotgun proteomics [10]. Fast and robust, shotgun proteomics does have drawbacks including the loss of some PTMs and structure. In contrast, top-down proteomics is the analysis of intact proteins and protein complexes [11]. Analysis of difficult post-translational modifications and structural artifacts of proteins is possible using top-down proteomics, but the increased complexity of the resulting signals and difficulty in pre-MS protein separation of complex mixtures for analysis are still barriers. Lastly, middle-down proteomics bridge the two approaches by using incomplete or perturbed proteolytic digestion to prepare mixtures of larger peptide fragments than in bottom-up approaches [9]. When properly utilized, proteomic methods can add important insights to larger studies and open new routes to understand the complex nature of immune signaling.

Figure 1.

Graphical abstract: a summary of approaches to immune signaling utilizing MS based techniques.

When compared to the non-MS methods described earlier, proteomic analysis has many benefits that can either enhance or replace these methods. While transcriptomics records changes in the mRNA levels, proteomics can directly measure the amount of protein (quantification) and record changes in the protein beyond its sequence, such as post-translational modifications (PTM analysis). While reporter assays can measure the changes in activation of specific pathways or transcription factors with extremely high sensitivity and ELISAs can measure changes in protein levels or specific protein modifications, both assays are dependent on having specific targets, with reporter assays using the introduction of a reporter gene downstream of a specific promoter sequence and ELISA assays using antibodies linked to an enzyme. Unlike reporter assays, proteomic analysis requires no targeting of promoter sequences or transcription factors. In addition, while ELISAs are dependent on having an antibody toward the protein or small molecule being measured, proteomics can measure not only the amount of a specific target without using an antibody but also simultaneously measure the amount of other proteins or small molecules in the sample. This direct analysis of the proteins involved in immune signaling cascades for their amounts, modifications, secretion, interactomes, etc. without prior identification and targeting is a strength of proteomics.

4. Proteomic analysis of immune signaling

While the identification of proteins contained in a sample, such as a gel slice, may be the first thing associated with the phrase ‘mass spec analysis,’ improvements in the resolving power and sensitivity of mass spectrometers have expanded the usefulness of MS. The high resolution and sensitivity of modern LC-MS systems enabled the development of new techniques and experimental methodologies that address questions crucial for understanding human health and disease.

4.1. Post-translational modifications

One of the methods for controlling protein activity, localization, or interaction is post-translational modification (PTM) of the protein. While there are more than 200 PTM types, the two most common regulatory PTMs in the signaling networks are phosphorylation and ubiquitination [12,13].

4.1.1. Post-translational modifications – phosphorylation

One of the most abundant and dynamic PTMs, phosphorylation can rapidly induce changes in protein localization, activation, stabilization, or interaction. Covalently bound to serine, threonine, or tyrosine residues by kinases, the phosphate can also be readily removed by phosphatases, making phosphorylation a strong but also reversible signal for the protein. Biochemically, phosphate groups have a strong negative charge that can be used to purify phosphorylated peptides from mixtures by their binding to positively charged metal ions, such as TiO₂ or Fe-NTA.

Given phosphorylation’s common use by cells in signaling cascades, it is not surprising that numerous phosphoproteomic screen studies have been performed throughout the years. In the area of innate immunity, Weintz et al. used metabolic labeling (see Section 4.2 on quantification) to identify nearly seven thousand phosphosites in LPS-stimulated macrophages at three time-points [14]. Sjoelund et al. used a phosphopeptide enrichment to identify the phosphorylation dynamics in at five time-points within the first 30 minutes in macrophages following stimulation with three agonists of toll-like receptors of the innate immune system, refining the temporal resolution of the analysis. By comparing the results of the three stimuli, the researchers found that while some pathways were conserved amongst all three stimulations there were several pathways that were unique to each stimulation. In addition, by combining phosphopeptide enrichment with a time course, the researchers found that the phosphorylation levels changed over time as the cells responded to the stimuli [15].

Another immune cell system extensively studied using phosphoproteomics is T cell signaling. A study by Tan et al. describes measurements of thousands of phosphorylation events during the time course following stimulation and identification of several distinct activation patterns during the time course. By using systems biology approaches to analyze the phosphorylated proteins along with their activation profiles, the authors generated a map of the signaling cascades regulating protein translation following activation. Furthermore, they found a strong correlation between T-cell activation was dependent upon the mitochondria with the inhibition of mitochondria-dependent oxidative phosphorylation preventing the conversion of naïve T cells into Th1 cells [16].

4.1.2. Post-translational modifications – ubiquitination

Like phosphorylation, ubiquitination is a common PTM found following immune stimulation; but unlike phosphorylation with its multiple receptor residues, ubiquitin molecules are almost exclusively bound to lysine residues in the target protein. Additionally, while phosphorylations consist of a single phosphoryl group, ubiquitin PTMs can exist as single ubiquitin modification (mono), two ubiquitin molecules bound to each other (di), or long chains of ubiquitin (poly). To further complicate the analysis of ubiquitin and its effect on proteins, polyubiquitin chains can be either linear or branched and can be linked to each other through the ubiquitin’s N-terminal methionine (M1) or any of the seven lysine residues in the ubiquitin molecule. Each of these binding types can in turn lead to different downstream results [17].

Fortunately, because ubiquitin is itself a protein and thus is digested by the same enzymes used to prepare large protein molecules for analysis on the LC-MS, the digestion of ubiquitin by trypsin leaves a distinct diglycine remnant on the lysine residue where the ubiquitin was bound. This motif (K-ε-G-G) can be bound to an antibody for enrichment prior to analysis and the addition of the diglycine remnant is readily visible in LC-MS analysis. Using this enrichment technique, Dybas et al. examined the proteins ubiquitinated following T cell stimulation and found over 1000 proteins with the majority of the ubiquitinated proteins being affected by the state of the T cells (resting v. restimulated). In addition, the use of the tandem ubiquitin binding entity (TUBE) enrichment system allowed to analyze the types of ubiquitin chains bound to the proteins. By using this technique, they demonstrated that following T cell reactivation over 200 proteins are ubiquitinated and nearly two-thirds of those proteins are being modified by non-degradative ubiquitin [18].

Another method for studying protein ubiquitination is the use of molecularly tagged ubiquitin to allow for the enrichment of the ubiquitinated proteins. In a study by Lectez et al., the authors biotinylated the ubiquitin molecules of a mouse in vivo and then used the strong binding of biotin to streptavidin to purify the ubiquitinated proteins from the livers of the mice. With nearly 400 ubiquitinated proteins identified, they were able to link ubiquitination with proteins found in all portions of the cell and involved in a wide variety of functions [19].

4.2. Quantification

Qualitative techniques have focused on protein identification, providing long catalogs of proteins found various types of samples. Quantitative approaches retain this power but allow to calculate protein abundance using various technical approaches. While qualitative proteomics offers valuable information based on protein identification in samples, the biological relevance of qualitative proteomic data can be limited. Reviewing several excellent workflows for quantitative proteomics, Sukumaran et al. highlight the power of quantitative approaches. Even though the MS-based approaches have long been resource prohibitive, the study designs reviewed by Sukumaran et al., cover the emergence of quantitative proteomics as an extremely important tool for dissecting and understanding the dynamics of immune cell signaling [20].

4.2.1. Label-free quantification

Label-free quantification (LFQ) is the most straightforward protein quantification method. This method consists of sequential MS analysis of separate samples and comparison of the resulting peak intensities for peptide IDs or spectral counting for proteins across several samples to determine relative abundances. One of the most widely used quantitative proteomics techniques, LFQ is a strong and accepted technique for performing relative quantification. One example of this strategies application is in a recent study by Montoya et al., where biological samples from patients with ulcerative leishmaniasis were analyzed by LC-MS/MS and differentially expressed proteins between treated and untreated patients were identified by comparison of relative quantification [21]. This discovery approach was used to identify a set of potential biomarkers to identify successful leishmaniasis treatment.

4.2.2. Chemical labeling

Chemical labeling by isobaric tagging of peptides consist tagging digested peptides before sample mixture and subsequent MS analysis. Peptides are labeled with tags consisting of varying combinations of ¹³C and ¹⁵N substitutions. The isotopic molecules consist of an amine reactive group, a variable mass reporter encompassing the variable isotopes, and a mass normalizer. Isobaric tags can be spiked into samples, which can be subsequently combined to shorten MS run time and allow for sample differentiation of the protein identifications and quantification. Relative quantification can be achieved through the comparison of the spectra intensities between the labeled samples. In an analysis by Wu et al., this quantitative approach was used to investigate the mechanism of infection of Bombyx mori with Bombyx mori nucleopolyhedrovirus (BmNPV). Heat shock protein 90 was known to be an important chaperone protein to this viral replication, and inhibition of Hsp90 lead to decreased viral replication. Drug treatments to inhibit Hsp90 were compared to control groups using TMT labeled samples, allowing for the identification of differentially expressed proteins between treatment groups, based on quantitative data. These data built a systems level overview of changing protein levels which suggested the changing Hsp90 levels influenced wide changes in other chaperone proteins effecting cellular processes ranging from innate immune activation, to transcription factor binding, which may be the mechanism for inhibition of viral replication [22]. This study highlights the rich array of systems data available in a quantitative proteomic investigation. Dimethyl chemical tagging is an affordable alternative to isobaric tagging. Developed by Hsu et al. as stable isotope dimethyl labeling, this method tags methyl groups on primary amines through reductive amination [23]. The resulting mass shifts on tagged peptides allows for relative quantification. This method uses combinations of hydrogen, deuterium, and ¹³C to provide a light, heavy and medium channel for quantification. Various stable isotope dimethyl labeling techniques for peptide tagging were developed by Boersema and others, in 2009, for quantitative proteomics [24].

4.2.3. Metabolic labeling

The chemical labeling techniques outlined above occur at the peptide level, after protein digestion and sample preparation for MS analysis. In contrast, metabolic labeling techniques rely on the incorporation of stable isotopes into proteins at the cellular or organismal level to allow for relative quantification between sample groups. The stable isotope labeling by amino acids in cell culture (SILAC) technique currently allows up to three multiplex channels based on isotope combinations of lysine and arginine. In vivo cell culture in the absence of arginine and lysine but supplemented with these amino acids in stable isotope-labed versions will eventually result in the full, or nearly full, incorporation of the stable isotope labeled amino acids into proteins [25]. Full incorporation of these labels in different biological samples allow for the proteomic quantification in experimental design based on the stable isotope’s mass shift in peptides analyzed by MS. The quick growth of cells in culture allows for the labeling of immortalized lines after several passages, but this stable isotope labeling approach has been applied at the organismal level as well. Though lifespans and slow label incorporation times can be expensive and time intensive, stable isotope labeling in mammals (SILAM) was developed to enable proteomic profiling and quantification of rodent model proteins [26]. This facilitates the proteomic study of disease states with greater biological relevance, achieved by the incorporation of ¹⁵N into rodent proteins.

4.2.4. Absolute quantification

The quantitative strategies described above are all relative. They allow for the identification of differentially expressed protein abundances between disease states, treatments, or across a temporal profile. On the other hand, absolute quantification is a proteomic strategy dependent on the spiking of a pre-quantified peptide standard into a biological sample allowing users to determine the concentration of analyte per sample volume, and resulting in the absolute quantification of the number of molecules of interest in the sample or in a cell. Synthetic peptides are designed which elute with the target peptide of interest to be quantified. These peptides must be proteotypic and quantypic. These synthetic peptide’s concentration is obtained through amino acid analysis. This absolute abundance per volume provides users with a predefined, quantified peptide standard. The spiking of LC-MS samples with synthesized peptides, used as internal standards, and the comparison of the native peptide peak with intensity with that of the internal standard peptide allows the target peptide and protein (based on several target peptides per protein) concentration to be calculated [27]. These approaches can yield useful biological information, the absolute quantification of proteins can be crucial for defining signaling mechanics within cells and defining the pathophysiology of diseases.

4.2.5. Hyperplexing

The quantification strategies outlined above are very versatile, but experimental design can be somewhat limited depending on your choice of labeling; 3 channels for SILAC, up to 16 channels for isotopic post-processing labeling, 3 channels for dimethyl labeling. This limitation can be overcome by multiplexing different labeling strategies together. Hyperplexing is an exciting frontier of protein quantification which is the facilitating of simultaneous quantification of a greater number of sample groups than can be achieved with specific labeling techniques alone, by utilizing multiple labeling techniques at once. In a study conducted by Welle et al., this hyperplexing strategy was applied by combining SILAC labeled cells and subsequent isobaric tagging to simultaneously measure protein turnover kinetics in human fibroblast cells. This approach allowed for the measurement of peptide turnover in quiescent cells in an efficient and cheap experimental design, on a larger scale, when compared against more traditional proteomic profiling methodologies [28].

4.2.6. Examples of quantitative analysis

Two studies that utilized protein quantification to measure the downstream effects of T cell activation are the studies published by Howden et al., and Marchingo et al. Both studies used the quantification of the cellular proteins in T cells following activation to characterize the processes of T cell expansion and differentiation, two process critical for the adaptive immune response.

In the study by Howden et al., the proteomes of naïve CD4⁺ and CD8⁺ T cells were examined before and after antigen activation. Nearly all (93% and 91%) proteins were affected by the activation with the majority (54% and 56%) of proteins increasing in abundance. By examining the specific proteins with changing abundances, the authors explained the previously reported differences in nutrient uptake by CD4⁺ effector cells versus CD8⁺ effector cells: CD8⁺ have an increased abundance of glucose and amino acid transporters in response to stimuli. In addition, mRNA translation regulators were accumulated following activation, including eukaryotic initiation factor 4F (eIF4F) complexes and eukaryotic initiation factor 2 (eIF2) complexes. Increase in these complexes was related to the increased protein synthesis of activated T cells. Interestingly, the study identified also several environmental-sensing molecules with increased abundances following activation. Combining these data with previous reports on protein function allowed to identify the nutrient-sensing kinase mTORC1 as a key regulator of T cell differentiation by affecting glucose transport, glycolysis, fatty acid metabolism, translation, and cell adhesion [29].

Another study using quantitative analysis to investigate the signaling pathways involved in T cell activation was published by Marchingo et al. This work focused specifically on the roles of the transcription factor Myc. Using the histone proteins as a ‘proteomic ruler’, the authors estimated the protein mass and copy number per cell, which showed that the increase in mass of T cells following activation is dependent on Myc. An examination of the proteins with increased expression showed that Myc expression affects the expression of lactate transporters, which control the rate of glycolytic flux and glycolysis. Furthermore, the quantification data showed that following activation T cells produced an increased amount of amino acid transporters and this increased activation leads to a positive feedforward loop that further increases in amino acids [30].

4.3. Extracellular signaling and secretome

While the intracellular signaling pathways are a common topic of research in most systems, immune signaling is also dependent on intercellular signaling via cell-cell direct contact and the secretion of free molecules and vesicles. Using proteomics (and its subfield, secretomics, examining the secreted proteins rather than internal proteins), the whole secretome of immune cells could be examined following stimulation and novel secreted proteins could be analyzed when dozens of secreted proteins could be identified simultaneously [31], and the contents of extracellular vesicles released during inflammation could be determined [32].

In an examination of the innate immune response, Koppenol-Raab et al. quantified hundreds of proteins secreted in response to TLR stimulation. In addition, analysis of the specific proteins found and their quantity following stimulation with three TLR agonists showed that the secretomes of macrophages stimulated with the different stimuli had portion (between 47 and 70%) of conserved secretome but a significant portion of each secretome was unique to the type of stimuli. In addition, the researchers compared the proteome and secretome results to the transcriptome and showed that the transcriptome results were lacking between 20 and 40% of the proteins identified by LC-MS [31].

While secreted proteins are often common signaling molecules, recently the roles of extracellular vesicles in signaling surrounding cells have been recognized as a promising research area. Mesenchymal stromal cells (MSC) are multipotent cells that form the connective tissue of the body and release a variety of signaling molecules in response to infections. To examine the proteins and other molecules contained in the small vesicles secreted by MSCs, Adamo et al. isolated the vesicles from the media by centrifugation and then analyzed the proteins contained in them by shotgun proteomics. By comparing the proteins found in samples from resting MSCs to primed MSCs, they were able to identify dozens of proteins with differential secretion following priming with a strong relationship between these proteins and the regulation of the actin cytoskeleton and to conclude that given MSC roles in wound healing and migration toward the site of injury or infection, modulation of the cytoskeleton is a process critical for proper MSC function [32].

4.4. Interactomics

During the cellular response to stimulation, there are two fast mechanisms that define the response of the cells: post-translational modifications and protein-protein interactions (PPI). The PPI, or interactome, has been extensively studied using proteomics methods and a wide variety of techniques have been employed to identify and analyze protein binding partners (reviewed in [33]). To briefly summarize, interactomic experiments commonly use one of three general protocols: affinity purification – mass spectrometry (AP-MS), proximity-dependent labeling – mass spectrometry (PDL-MS), and implied interactomics.

AP-MS is the classical approach to interactomics. Using a purification method such as an antibody toward the target protein or the chimeric addition of an ectopic tag to the target protein, the target protein along with all the associated proteins are extracted for identification and analysis. Using this technique, Lum et al. analyzed the interactome of the DNA-sensor protein cyclic GMP-AMP synthetase (cGAS) following HSV-1 infection. By comparing the cGAS-bound proteins in mock and HSV-1 infected cells, the researchers identified proteins involved in apoptosis, cytoskeleton maintenance, mitochondrion maintenance, and signal transduction bind to cGAS in response to HSV-1 infection. Furthermore, by comparing the cGAS-interactomes recovered following infections with mutant HSV-1 virions, the researchers identified a crosstalk pathway between the RNA and DNA sensing pathways [34].

In addition to analyzing the results using data dependent acquisition (DDA), several labs have used data independent acquisition (DIA) to improve the sensitivity of their AP-MS experiments. Briefly, DDA is a process where the mass spectrometer analyzes and identifies the most abundant ions, and then further analysis, identification, and quantification are performed on only the most abundant ions present in the mixture collected from the LC spray, often this is noted as TopX where the X is how many ion signals are fragmented and analyzed. The limitation of this setup is that if an ion is not in the TopX portion then it is not further analyzed. In DIA all ions within a specified window are fragmented, analyzed, and quantified. While this would be a possible improvement to an analysis, there are a few critical issues to remember including that DIA experiments are not compatible with all MS instruments and that the data produced can be highly complex with multiple overlapping ion signals [35]. Despite these issues, several studies have combined DIA with AP-MS to improve their analysis.

One study where the combination of AP-MS with DIA techniques to produce detailed interactomes with temporal and cell specific resolution is the study by Caron et al. that analyzed the interactions of growth-factor-receptor-bound protein 2 (GRB2), an essential adaptor protein involved in cell signaling. In this study, a mouse model with a chimeric GRB2 construct was used to allow for affinity purification of GRB2 from developing T cells isolated from the thymus and from mature T cells isolated from the periphery. By comparing the GRB2 interactomes produced by each condition, the authors identified 53 proteins with high-confidence interactions along with 8 proteins that had significantly different temporal patterns when comparing the developing to mature T cells [36].

Voisinne et al. combined mouse genetics, AP-MS and sophisticated computational approaches to monitor the dynamics of the complexes formed by ubiquitin E3 ligases CBL and CBLB and defined the signalosomes associated with these two hub proteins. They were able to explore the time-dependent correlations of protein associations with these baits and gain insights into how CBL and CBLB regulate protein ubiquitination following TCR stimulation [37].

Another study by the same group combined the specificity of AP-MS with the flexibility of DIA is the second study by Voisinne et al. Here, the interactomes of 15 proteins downstream of the T cell antigen receptor (TCR) in primary CD4⁺ T cells isolated from mice. In addition to analyze the interactomes in static cells, the researchers stimulated the T cells in vitro by adding purified antibodies to the isolated cells. The interactomes of all 15 proteins were combined into a network of 277 unique proteins and 366 high-confidence interactions. Furthermore, the analysis of these interactions it was possible to generate a complex and informative model for the interactions downstream of the TCR response [38].

While AP-MS uses the purification of the target protein-containing complexes as the method for isolating the binding proteins from the background, PDL-MS uses the addition of a ligase to target protein that will label all proteins in close proximity to the target. A commonly used enzyme is the bacteria-derived biotin ligase, BirA or BioID [39]. BioID, or the enhanced versions, TurboID and MiniTurboID [40], is fused to the target protein and labels all lysine residues in close proximity to the target with biotin. The labeling with biotin allows for stringent lysis of the cells followed by purification of the labeled proteins by the strong affinity of biotin to streptavidin. In immune signaling, PDL-MS was used by Rider et al. to analyze the Epstein-Barr Virus (EBV) protein latent membrane protein 1 (LMP1). This study showed that LMP1-interacting proteins includes proteins involved in many cellular pathways/processes including the MAPK cascade, protein ubiquitination, and protein transport. In addition, the authors characterized the role of LMP1 in the production of exosomal vesicles that pass proteins and signal molecules from cell to cell [41].

While the interactomic methods used in the above described studies examine the interactomes of specific proteins, implied interactomics is a high throughput way to analyze the interactomes of multiple proteins simultaneously by combining the interactomes characterized by AP-MS or PDL-MS. One common method is to separate the protein complexes from cells using chromatography or gel electrophoresis based upon the complexes’ size, pH, isoelectric point, charge, or a combination of factors. Following the separation, the resulting complexes are degraded, digested, and analyzed by LC-MS/MS. The identification of proteins within the same fraction implies a possible interaction between the proteins (reviewed in [42]). Another method proposed to do high throughput analysis of the whole cell interactome is to use the principle that proteins bound in a complex then denatured by heating will aggregate together [43,44]. While both methods can indicate a possible interaction and be used to characterize a known interaction, it is important to remember that multiple protein complexes can simultaneously coelute.

New truths become evident when new tools become available

– Rosalyn Sussman Yalow, Winner of 1977 Nobel Prize in Physiology or Medicine

5. New directions of mass spectrometry

While these methods have been informative and critical for our understanding of immune signaling, there are several advancements in MS analysis that may provide new avenues for research as they become more developed. Four techniques and areas where MS analysis could provide additional information includes small molecule and metabolite analysis (metabolomics), mass spectrometry imaging (MSI), single cell analysis, and bioinformatics.

5.1. Small molecules and metabolites

While the analysis of proteins has dominated the use of the MS in biology, the high sensitivity of MS systems means that small peptides and metabolites can also be detected and quantified. Varying levels of these molecules can be indictive of changes to the cell, including mutations of the cell, modifications of cellular pathways, or response to stimuli. In addition to small molecules produced by the normal cellular processes, another small molecule target of analysis specific for immune signaling is the peptides bound to the major histocompatibility complex (MHC). The selection and presentation of these peptides is critical for the adaptive immune response and the inhibition of autoimmune diseases. While this critical aspect of immunology has been long analyzed, the analysis of these peptides has been difficult for researchers due to unknown variables that define the selection and production of the peptides from the foreign antigen, the similar sizes of the peptides presented on MHC, and the low abundance of specific peptides presented during the immune response [45].

While the generation of the peptides has long been hypothesized to be related to the ubiquitin-proteasome system, the specific mechanisms that select antigenic proteins for degradation and presentation were unclear. One possible pathway, proposed by Trentini et al., is the ribosome-associated quality control (RQC) machinery which degrades defective ribosomal products (DRiPs). To test their hypothesis, the researchers knocked out proteins required for RQC and analyzed the peptides presented on the MHC-I by MS analysis. Overall, wildtype or complemented mutants presented fewer peptides from fewer proteins than the mutant cell line. In addition, analysis of the types of peptides presented found that transmembrane proteins are preferentially degraded and presented by this mechanism [46].

Following the generation of the peptides, the next step is the loading of the peptide into the MHC-I complex for presentation on the plasma membrane. Prior to the presentation, the MHC-I: peptide complexes are edited to control the presentation of peptides and tapasin was originally identified as the sole peptide editor. Recently, a second protein was identified as a peptide editor, the tapasin-related protein TAPBPR. In the study by Ilca et al., MS-analysis of the peptides presented following mutation of TAPBPR to analyze the mechanisms involved in peptide dissociation from MHC-I was used and a loop of amino acids containing a leucine in TAPBPR was found to act as a level to lift peptides out of the MHC-I groove. In addition, soluble TAPBPR in the media can promote the ejection of peptides presented by cultured cells, allowing for the MHC-I complexes to then bind synthesized peptides. This process could enhance the effectiveness of immunotherapies in targeting tumors [47].

5.2. Mass spectrometry imaging

The intact, three-dimensional investigation of biologically relevant structures is a near ideal goal for many physiological questions. Throughout history the increasing power to visualize has crept from digital radiography, X-ray, computed tomography, nuclear imaging and so on. All these methods with their varying strengths and weaknesses. With increasing computational power, and mass spectrometry’s ability to record hundreds or thousands of proteomic, metabolomic, or lipidomic datum in every sample the multi-omic, three-dimensional composition and distribution from biological samples can be reconstructed with greater accuracy than ever before. The technique of mass spectrometry imaging (MSI) is dependent on selectively ionizing the surface of a biological surface that has had minimal sample prep and directly injecting ionized analytes into a mass spectrometer for analysis and replicating this process throughout a predetermined area on the surface of the biological sample. This method allows researchers to render an incredibly rich image of the analyte distribution across a surface. This rastering across sample surfaces, and repetition of this process through multiple slices of the same biological sample allow 3-D rendering of molecular distribution within tissues. MSI files from a singular plane of a biological sample can contain hundreds if not thousands of sample positions which form a rich map of the biochemical topography across the sample.

Matrix Assisted Laser Desorption Ionization (MALDI) MSI is a method of mass spectrometry imaging in which an applied organic matrix softens the ionization process to decrease fragmentation and provide a uniform layer from which to sample, which helps to normalize analyte variability entering the mass spectrometer when sampling nonuniform surfaces. After the application of a matrix across the sample surface, a laser ablation focused to a precise sample area desorbs surface analytes and ionizes sample into the MS inlet. The sample is then resolved, and mass spectra are acquired. While the rich data characteristics of a single MS sample run makes MSI an incredibly attractive tool for visualizing biological distribution of analytes, but an issue with any imaging tool is resolution. In work by Kompauer et al., the lateral resolution of an atmospheric pressure MALDI-MSI setup reached 1.4 μm. This was done using a numerical aperture on a single focusing objective, allowing for higher contrast and resolution than other available MALDI-MSI set ups. With coupling to an Orbitrap MS the ability to differentiate lipid, metabolite and peptide sample topography was successfuly demonstrated [48].

Secondary Ion Mass Spectrometry (SIMS) MSI is the bombardment of a sample surface from a primary ion source and the MS analysis of resulting secondary ions ejected from the sample surface that are directed into the MS inlet. Passarelli et al. created a 3D OrbiSIMS with subcellular resolution for lipids and metabolites. SIMS coupled to a tandem Orbitrap analyzer and TOF analyzer allowed for the resolution of the metabolic profiles of the nucleus within an interneuron [49].

Desorption Electrospray Ionization Mass Spectrometry (DESI) MSI operates based on analyte desorption after electrospray ionization focus on the sample surface. In a study by Garza et al., a DESI- field asymmetric waveform ion mobility (FAIMS)-MS tool was used to directly image top down proteins from tissue samples. FAIMS separates ions in their gas phase by their movement in an electric field. It is a useful technique of top-down proteomics to further separate intact structures, downstream of the usual liquid chromatography but in this case, DESI set up. As all analytes in the sample region are transmitted to the MS inlet this coupling reduces signal noise. This group used DESI-MS for top down protein ID of common and biologically relevant proteins like hemoglobin and S100 proteins. Simple sample preparation and top down nature of this approach makes it attractive for proteomic profiling of tissues, but the overall protein coverage of less abundant proteins is apparent in comparison to MALDI-MSI [50].

5.2.1. Immune system mass spectrometry imaging

The above mentioned MSI techniques are examples of exciting advances in the application of mass spectrometry being thought up over the last decade. Biological relevance is best preserved in many forms of imaging. This is why advances in technologies ranging from x-ray, MRI, PET and CT have had such high applications in clinical settings. As systems approaches become more common in study designs, the popularity of MSI is sure to grow. MSI offers the ability to build richer in situ imaging data without the need of arduous sample preparation, labeling, antibodies etc.

Application of MS imaging to the structural investigation of the immune system is an exciting frontier. Holzlechner and Strasser et al. illustrate the application of MALDI-MSI to investigate the proteomes of tissue resident immune cells, their distributions, and polarized states. They clearly demonstrate that MALDI-MSI on tissue resident cells produced matching proteomes to intact cell mass spectrometry prepared for traditional LC-MS, lending validity to the data produced by MSI as biologically relevant, and reproducible. An important step in setting MSI as a confident and rapid approach to proteomic investigation of immune cells [51].

When investigating molecules in situ, the question of molecular origin becomes apparent, especially when investigating the interface between host and microbe. Geier et al. address this problem by investigating the metabolomic phenotypes of host and microbe cells in situ. The effective rastering and high resolution of the current MALDI techniques, as previously discussed by Kompauer et al. [52] meant that there was minimal sample damage after MALDI-MSI. This translates to sample integrity for further downstream experimentation. Fluorescent in situ hybridization (FISH) is a method of fluorescently labeling complementary nucleic acid sequences on cytochromes. Geier et al. developed a pipeline of high resolution MALDI-MSI and subsequent FISH microscopy on the same biological sample. This image overlay allows for the metabolomic topography to be visualized and host or microbe metabolites to be identified at a resolution of 3 μm when optimized [53]. This pipeline would have many applications in infectious disease and immunology research beyond the exemplary host-microbe symbiosis described in their study.

5.3. Single cell analysis

While the analysis of organs, tissues, or dishes of cultured cells has informed our understanding of all biology, many processes are dependent on the responses of single cells which can then be propagated by either direct or indirect cell to cell signaling. Because these responses often begin in a single cell, analysis of individual cells could offer a glimpse into the initial stages of any response. With the improvements in flow cytometry and nanoflow machinery, the selection and isolation of single cells from a mixed pool has become routine, but the analysis of the cell’s proteome has been only hypothesized for many years.

One of the earliest methods that uses mass spectrometry to analyze single cells is mass cytometry, also known as cytometry by time-of-flight or CyTOF. There are many excellent reviews of CyTOF that explain the mechanisms, benefits, and concerns involved in these experiments (reviewed in [54–56]). Briefly, specific targets are labeled with heavy metal tags then the single cells isolated and atomized using argon plasma. Then using the sensitivity and accuracy of an elemental time-of-flight mass spectrometer to detect the heavy metal tags, CyTOF can quantify up to 40 possible elements at a single cell level. While the measurement of 40 individual elements is an improvement over the number of elements that can be simultaneously measured with fluorophores (a process limited to less than 20 simultaneous measurements), the method of does have drawbacks. First, CyTOF uses atomization of the cells with argon plasma to free the metal markers for detection and cannot be used to collect live cells and the sensitivity for detection of the mass reporters prevents the detection of extremely rare molecules.

To address the limitations of CyTOF, researchers have begun to develop single cell whole proteome analysis methods. Two recent methods that analyze single cells for protein quantification are SCoPE-MS and nanoPOTS. In both cases, the researchers use chromatography to isolate individual cells then lyse and digest the individual cells separately. The peptides are then labeled using post-digestion isobaric tandem mass tagging to identify the peptides as being from a specific cell. Finally, the peptides are pooled to create a mixture with enough signal for analysis by the LC-MS/MS system.

In the method developed by by Budnik et al., the researchers used a method called Single Cell ProtEomics by MS (SCoPE-MS) to isolate the cells. Initially, manual picking of cells under the microscope was used to isolate single cells and in the improved method [57], it was changed to completely automated single cell picking based on acoustic single cell tracking and dispensing, followed by digestion and labeling. Then, the labeled single cell derived peptides are mixed with peptides derived from a pool of 200 similar cells. The goal of adding the pooled carrier cells was to enhance the signals of the cells peptides over the noise of the LC-MS/MS system. Once the system could identify the presence of a peptide, it would analyze the specific ions in the mixture and thus measure the amount of the labeled peptide in the mixture compared to the carrier peptides. By this method, the researchers successfully analyzed the protein content of mouse embryonic stem (ES) cells during differentiation and found that cells could differentiate cells of differing fates by their protein contents [58].

Another notable method for single cell proteomic analysis, is the use of nanodroplet processing in one-pot for trace detection (nanoPOTS) described by Dou et al. Unlike the SCoPE-MS method which places the cells onto a glass slide, in nanoPOTS the single cells are isolated using microscopy-assisted robotic platform into nanowells on chips to prevent evaporation of the media. The cells were lysed, and the proteins digested followed by labeling. Finally, the peptides from each cell were mixed along with a ‘boosting’ channel containing a mass of peptides derived from a different cell line, like the carrier peptides in SCoPE-MS, prior to analysis by LC-MS/MS. By this method, the researchers were able to achieve reproducible results across multiple tuns and differentiate between mouse cell lines of differing origins [59].

5.4. Integrative multi-omics

While the techniques described above are extremely informative and useful, a potentially powerful improvement is called multi-omics. Based upon the integration of multiple -omic techniques, multi-omics could shed light on the difficult to assay regions of the cell. By combining proteomic analysis of the proteins in the cell, transcriptomics of the mRNAs produced following a stimulation, and phosphoproteomics of the proteins modified in the cell following a stimulation, researchers can map the full pathways stimulated during their experiment.

Two elegant reports used multi-omics to analyze the toll-like receptor (TLR) pathways. In the report by Chevrier et al., transcriptomics, proteomics, and phosphoproteomics were combined to analyze the responses of bone marrow-derived dendritic cells (BMDCs) to stimulations with several TLR agonists. In their transcriptomics data, 280 genes were found to change expression following stimulation. The 280 potential targets were then reduced to a subset of 23 with strong differential expression and their previously unknown relationship with TLR pathway. A further analysis of these candidate proteins identified one, Polo-like kinase 2 (Plk2), as a novel non-regulator of the TLR pathway. Interestingly, while the knockdown or knockout of Plk2 had no effect on the TLR pathway alone, when both Plk2 and Plk4, a related kinase that had similar differential expression, were inhibited, the antiviral responses of the BMDCs was significantly decreased. Furthermore, treatment with a pan-specific Plk inhibitor lead to modified production of antiviral RNAs, secretion of antiviral proteins, and phosphorylation of 413 distinct proteins following stimulation with LPS [60].

In the follow-up study, Mertins et al. used multiple omics techniques to analyze the responses of dendritic cells (DCs) to LPS stimulation. Here, SILAC-based relative quantification was used to identify proteins with differential expression upon stimulation. Next, the phosphoproteome was analyzed during a time course from 15 to 360 minutes post stimulation to characterize the activation patterns of known TLR components. These patterns were used to define the early events of the TLR response and, when combined with previously reported results, highlighted 169 candidate genes. Knockdowns of these candidate genes by shRNA lead to 27 potential regulators of the TLR response as measure by transcriptome analysis. To further characterize the roles of these potential regulators, interactomics was advantageously employed to identify the binding partners of several of the candidate proteins. This process characterized the adaptor related protein complex 1 associated regulatory protein (AP1AR) and its binding partner, phosphatidylinositol binding clathrin assembly protein (PICALM), as being likely regulators of pro-inflammatory TLR4 signaling. Furthermore, by combining their results with previously reported results in public repositories the authors created a network of the signaling-to-transcription paths in DCs that respond to LPS stimulation [61].

These studies highlight the power of multi-omics and systems biology. Only by combining an understanding of transcriptomics, quantitative proteomics, phosphoproteomics, and interactomics using bioinformatics, a clear understanding of the pathways activated following LPS stimulation could be reached. In addition, the analysis of their results led to new avenues of analysis and branches of the pathways that could have gone unexamined if not for their reproducible result in multiple omics assays. These studies and many more currently in the literature show the concept of multi-omics is a very well-developed and readily applicable idea, clearly to be explored as one of the avenues whichh will gain more and more interest in the future (Figure 2).

Figure 2.

A network of techniques employed for multi-omic investigations in the immune cell signaling studies reviewed in this article, with MS-based methods merging with other accepted stand-alone techniques.

6. Expert opinion

We have sought to give researchers a broad and thorough survey of proteomic strategies that can be implemented when investigating the immune system. In order to approach such a complex field as immunobiology, a systems approach is required. This entails the use of traditional methods, but also expand the tools to the mass spectrometry-based approaches.

To date, several current MS techniques including PTM analysis with special attention to two common PTMs, phosphorylation and ubiquitination, have resulted in may studies elucidating the roles of these PTMs in the various cell signaling landscapes. Next, the field of quantitative proteomics has provided new tools and paved the way to investigate disease states and immune signaling with greater biological context. Then, the analysis of whole secretomes offers a more complete glimpse at cellular expression and communication and the extracellular signaling molecules that are valuable and may be lost in other types of experimental design. Lastly, reconstructing the functional networks within cells using emerging techniques to investigate protein-protein interactions, or interactomics, can help inform our assumptions of cellular and molecular functionality.

Beyond the current techniques previously used with immune signaling, advances in MS technology have led to new methods that will help address experimental challenges within systems immunology. With progresses in MS detection limits and sensitivity, studies of small molecules and metabolites, products of cellular processes often left out of context in proteomic investigation, began to complement the studies of signaling networks. Using new methods of sampling, the varied techniques of MS-imaging provide even greater biological context as investigation of proteins, lipids, and metabolites are possible in situ. In addition, 3D-reconstruction of structures and microenvironments can inform immune research in powerful ways. Continued improvement in data analysis and greater integration of bioinformatic tools, whose deeper analysis is beyond the scope of this review, into the standard workflows of labs has led to an improved understanding of biological processes and mechanisms.

Recent years were also marked by the advances in single cell proteomics that will allow to probe the vast heterogeneity of cell states and cellular responses dependent on cell-to-cell signaling with higher resolution and depth than before. The continuity of cell states visible in the populations and acknowledged, for example, by Mosser and Edwards who describe the spectrum of macrophage states between the dichotomy of M1 and M2 [62], is possible to visualize using flow cytometry, but single cell proteomics will undoubtedly inform the field as to the proportion of single cells and their changes and differences in response to pathogens and tissue damage.

Another area we visualize as expanding with the new quantitative mass spectrometry tools is the research on a large number of subjects in human studies, which has been lacking and necessary to systematically address the multiple genetic, dietary and lifestyle factors in combination with the microbiome, immunity and molecular tissue biomarkers as well as the influence of germline and somatic genetic variations on cellular signaling and immune function. Proteomics will add a new dimension to the research performed according to the Molecular Pathological Epidemiology (MPE) paradigm [63,64] whose contributions to epidemiology and precision medicine are becoming evident.

Proteomics is especially amenable to complement other orthogonal methods in immunology, such as quantitative microscopy, other -omics, optogenetics and genetic modifications, animal studies, and flow cytometry. The combinatorial approach and conceptually original, carefully designed multi-disciplinary studies involving teams of scientists with diverse expertise will generate answers to crucial questions unattainable by individual techniques.

Article highlights.

• Combining proteomic techniques with systems biology approaches offers new tools for the analysis of immune signaling.

• Current proteomic techniques used to address immune signaling:

  • Post-translational modification screens (phosphorylation and ubiquitination), Protein quantification, secretome analysis, and interactomics.
• Future mass spectrometry–based techniques to use in immune signaling:

  • Small molecule and metabolite analysis, mass spectrometry imaging, single cell analysis.

• Expanding our use of these established and developing techniques will lead to an improved understanding of immune signaling.