How can platelet proteomics best be used to interrogate disease?

Abstract Various modifications of proteins and the resulting proteoforms of a protein can associate with many diseases and are also significantly involved in the rapid regulation of hemostasis and thrombosis. For example, the release of prostacyclin from the intact endothelium and the consequent following phosphorylation of VASP in platelets is a post-translational regulation to keep them in a quiescent state. In Alzheimer’s disease, proteoforms arise from the altered cleavage of the amyloid precursor protein, which finally causes amyloid plaques in the brain. This changed processing of the amyloid precursor protein can also be detected in platelets, making them an attractive source of biomarkers for this neurodegenerative disease. Age-related or prothrombotic disorders can have multiple origins, including genomic, transcriptional, and translational factors, which together can be mapped at the proteome level. Hence, recording these dynamic protein changes under physiological and pathophysiological conditions is paramount in platelet proteomics. To effectively study diseases through platelet proteomics, it is crucial to consider platelets’ primary regulatory mechanism and thoroughly evaluate the disparities between the two leading proteomics technologies, top-down and bottom-up approaches. This commentary provides insights into the differences between these two technologies, which are particularly noticeable in detecting the different proteoforms of a protein. Plain Language Summary What is the context? The repertoire of all proteins in a biological sample is the proteome. Proteomics refers to different biochemical technologies that detect and quantify the proteins in a biological sample, such as platelets. If proteome analyses are carried out on a representative number of samples from a specific patient group and compared to a matched control group, disease-dependent changes in proteins can be found that indicate unknown causes of the disease or also be used as biomarkers for diagnosis and prognosis. It is also essential to consider that the proteins in biological samples can occur in various variations, the proteoforms. These proteoforms of a protein can arise, for example, through genetically-based variations or regulatory post-translational protein modifications. What’s new? There are two fundamentally different methods in proteomics technology: top-down and bottom-up. For bottom-up proteomics, the proteins must be digested into peptides for technical reasons, whereas top-down proteomics analyzes intact proteins. These different sample processing steps significantly impact the resulting data set. This is particularly crucial for platelet proteomics studies, as primary hemostasis is mainly carried out through post-translational modifications of proteins, resulting in various proteoforms that regulate platelet reactivity and thrombus formation. What’s the impact? Bottom-up proteomics can quickly and automatically identify an extensive repertoire of proteins from a platelet sample. This is much more cumbersome with top-down proteomics. In contrast, here, various intact proteoforms of intact proteins can be unbiasedly detected and directly quantified, which is particularly important for examining the global proteome of platelets in clinical samples. This qualitative and quantitative relative assignment to the respective proteoforms of a protein is not possible in bottom-up proteomics.


Post-translational modifications are essential to the main functions of platelets
Platelets are responsible for primary hemostasis, a vital process that stops bleeding when blood vessels are injured.To fulfill this life-protecting function in time, this physiological reaction must be initiated immediately.Accordingly, these primary hemostasis steps must be very rapidly executed by the platelets and consist of adhesion, activation, platelet shape change, degranulation, and aggregation.Post-translational modifications are mainly responsible for the quick implementation of these signaling cascades.Thus, one of the first steps in platelet activation, such as binding of vWF to GPIb-V-IX, triggers a phosphorylation signaling cascade via the adapter molecule 14-3-3ζ and the Src kinase, which mediates downstream activation of the major effector enzyme PLCγ.These signaling events in activated platelets lead, among others, to the binding of cytoskeletal protein Talin-1 to GPIIb/IIIa, triggering a conformational change to its active form, which recognizes proteins containing an Arg-Gly-Asp motif, such as fibrinogen.The activation mechanism of the GPIIb/IIIa integrin complex involves calpain-mediated cleavage of Talin-1. 1,2More generally considered, proteolytic cleavage, a post-translational modification, is a significant step in regulating hemostasis to activate coagulation factors and fibrinolysis.
On the other hand, the intact endothelium secretes nitric oxide (NO) and prostacyclin (PGI 2 ), which immediately prevents platelet activation and keeps platelets quiescent in blood circulation.Thus, PGI 2 and NO elevate intracellular levels of cAMP and cGMP in platelets by activating the kinases PKA and PKG, whose phosphorylation of various cytoskeletal proteins such as VASP 3 and RAP1B 4 directly reduces the platelet reactivity potential.Concurrent with these highly specialized traits of rapid response in hemostasis, however, platelets lack a nucleus and perform virtually no timeconsuming de novo protein synthesis.For this reason, proteome analysis has emerged as the preferred tool to unbiasedly study the molecular composition of these highly specialized cell particles in physiological and pathophysiological states.

Technical differences in top-down and bottom-up proteomics
However, two fundamentally different technologies are available to perform these analyses, top-down and bottom-up proteomics.The main difference between these technologies is that for bottom-up proteomics, also called shotgun analysis, the proteins of a biological sample have to be digested into peptides to enable the analysis process.This strange step of preparing biological samples using enzymatic digestion is necessary because an adequate liquid chromatographic separation and subsequent mass spectrometric measurement of complex protein mixtures are only possible with their peptides than with the corresponding intact proteins.On the other hand, top-down proteomic methods involve the analysis of intact proteins from a biological sample.The term "top-down proteomics" was first coined in the early 2000s to define mass spectrometry-based method developments that should also enable the analysis of intact proteins.6][7] Different staining methods directly visualize the two-dimensional gel electrophoresis separated protein profiles in a nice analogue image of a proteomic spot map.In this process step the intact proteins can be quantified and biological differences can be identified.It is important to note that only after the intact proteins have been fractionated, and their identity is still unknown, are they digested into peptides for identification by MS.
In contrast, only a digital visual of the much more complex peptide mixture from a digested biological sample is possible in shotgun proteomics.However, since the two-dimensional gel electrophoresis top-down proteomics method is not as easy to automate as bottom-up proteomics technologies and is significantly more labor-intensive, the bottom-up variant has increasingly become established in proteomics. 8

Top-down and bottom-up proteomics yields fundamentally different data sets
In bottom-up proteomics, the peptide masses of the digested proteins must then be reassigned to their canonical protein sequence for protein identification and quantification to be performed.However, for a biological sample such as platelets, this bottom-up analysis process loses most information about the proteoform composition and the post-translational modification profiles of the proteins, which are also essential for the rapid regulation of primary hemostasis.
Although the canonical protein repertoire of a platelet sample can be quickly identified and quantified using shotgun analysis, detecting unbiasedly hemostasis-regulating PTM-modified proteoforms of the respective proteins is limited.Thus, it can be assumed that the data sets obtained from a shotgun analysis of platelets should primarily provide information about their preceding transcription and translation phase from the megakaryocytes.In addition, the bottom-up analysis may also detect post-translational regulated decreases and increases in proteins, which are lost from the platelets due to degranulation or microvesicle shedding or are uptaken from the plasma.

Most proteins can have numerous proteoforms
Although PTMs have been known for decades, the enormous complexity and importance of these regulatory protein changes have only gained more attention in proteome research in recent years.To give this extensive variety of protein variants of a protein a uniform name, the term proteoform was defined in 2013. 9Thus, the term proteoform includes variants of a protein based on variations or mutations in its gene or alternative splicing of its RNA.From highly sensitive shotgun analysis data it is currently estimated that platelets contain around 5000 different canonical proteins. 10Currently, there are approximately 400 recognized PTMs alongside other genetic and transcriptional variations. 11Thus, up to millions of different regulatory proteoforms can be present only in platelets.
However, since the top-down method, two-dimensional gel electrophoresis, separates intact proteins based on their isoelectric point (pI) and molecular weight (MW), the proteins and their proteoforms can be detected as exchanged amino acids or PTMs alter the electrophoretic properties of proteins.The 2D protein spot map of the global platelet proteome, in particular, clearly shows that many proteins that are functionally important for hemostasis, such as fibrinogen, thrombospondin, VASP or coagulation factor XIIIA, are present in numerous of their proteoforms.However, it should also be noted that the corresponding PTMs of differentially regulated proteoforms have received too little attention in bottom-up proteomics and are also rarely studied more deeply in two-dimensional gel electrophoresis-based global platelet proteomics and other proteomics studies.One reason is that the PTMs responsible for a given pI and MW of a proteoform are also often very difficult to identify from a 2D protein spot.When conducting a proteome investigation of different biological sample groups, top-down and bottom-up approaches typically aim to identify as many significantly altered "protein events" as possible and then analyze them using bioinformatic tools to group them into pathways or interactomes.However, it must be noted that in most cases, not all corresponding proteoforms of these protein signals were found.Moreover, all these other proteoforms do not necessarily have to be quantitatively changed in the same direction, and their abundance profile should therefore be examined more closely.A method for the strongly recommended detection of the intact proteoform profiles of specific proteins in a biological sample, such as platelets, would be the 2D Western blot analysis as long as antibodies against these proteins are available.After these rather theoretical considerations on the application of proteomics in platelet research, some practical examples from the literature should help to show what information can be found in the platelet proteome with these different proteomic technologies.

Proteoforms of Coagulation factor XIIIA (F13A1) are changed in platelets of lung cancer and COVID-19 patients
In an unbiased clinical study of the global platelet proteome using the top-down proteomics method two-dimensional gel electrophoresis, we identified an increased 55 kDa cleavage product of F13A1, the fibrin-stabilizing coagulation factor, in platelets from patients with lung cancer, a disease associated with a high risk of thrombosis.A significant correlation of this F13A1 cleavage product with plasma concentrations of the plasmin-alpha2-antiplasmin complex and D-dimer in this lung cancer study indicates a previously unknown inactivation of F13A1 by the fibrinolytic system in human. 12In a subsequent proteomics study of COVID-19 patients, we also found an increased amount of this 55 kDa cleavage product of F13A1 compared to healthy controls in the platelets.In parallel, we found decreased F13A1 levels in the corresponding plasma samples of this study cohort, which were even more reduced in the non-surviving COVID-19 patients. 135][16][17] However, this proteolytic proteoform of F13A1 would be undetectable using bottom-up proteomics.Similarly, another two-dimensional gel electrophoresis-based top-down proteomics study showed that the integrin αIIbβ3-activating protein Kindlin-3 is progressively cleaved in human platelets during severe myocardial infarction and may thereby affect the conformational change of this fibrinogen receptor for thrombus formation. 18

Genetic and splicing-based alterations of proteoforms in platelets from Alzheimer's patients
In the platelet proteome of Alzheimer's patients, we also found more changes based on the proteoforms of proteins with twodimensional gel electrophoresis, such as apolipoprotein E4, derived from the well-known single nucleotide polymorphism (SNP) of this neurodegenerative disease or just a single increased out of seven different TPM1 proteoforms. 19Only increased protein amounts of MAOB in the blood platelets of Alzheimer's patients are also due to increased amounts of RNA from it. 20he amyloid precursor protein was not detectable with twodimensional gel electrophoresis method setting used in this study, 19 since it has a molecular weight of 140 kDa and is difficult to separate as a membrane protein with this top-down method.This protein may be better detectable using bottom-up proteomics, but not its cleavage products.

Mass spectrometry-based top-down proteomics analysis of platelets
Only one MS-based top-down proteomics analysis of platelets is currently published.This work characterized the composition of the intact proteoforms of platelets and many other blood components such as plasma, monocytes, neutrophils, erythrocytes and various subgroups of lymphocytes. 21The fractionation of the intact proteins and their proteoforms for MS-based top-down proteomics is mainly performed with gel-eluted liquid fraction entrapment electrophoresis (GELFrEE). 22But as mentioned before, this MS-coupled fractionation and identification method does not yet work as well as two-dimensional gel electrophoresis.Only proteins/proteoforms with an MW smaller than 30 kDa can be identified.Therefore, 93% of all identified proteoforms had MW < 20 kDA using this top-down MS proteomics method.Thus, this top-down proteomics analysis identified 344 distinct proteins and 1595 unique proteoforms from platelets. 21n contrast, a well-equipped and well-run bottom-up standard shotgun analysis is much more sensitive to identifying as many different proteins in the global platelet proteome, but only based on their canonical amino acid sequences.As mentioned above, 4000 unique proteins have been identified in the platelet proteome, and their respective amounts (copy numbers) have been estimated in platelets. 10

Affinity enrichment of modified peptides to improve detection of certain PTMs with bottom-up proteomics
For better detection of PTMs with bottom-up analyses, affinity enrichment of the respective PTM is recommended, for example, the enrichment of phosphopeptides from a platelet sample.Likewise the phosphoproteome of the GPVI/ITAM pathway, 23,24 cAMP 25 and cGMP system 26 are extensively studied as well as from patients and with the Scott syndrome 27 and in obesity. 28hese approaches are well suited to look deeper and more precisely into the respective pathways of platelet reactivity regulation.For example, analysis of phosphopeptide profiles in obese subjects suggested a possible alteration in signaling pathways mediated by platelet Src family kinases.To elucidate this further mechanistically, it could be shown that the collagen receptor GPVI is increasingly present on the surface of the platelets of these patients and their platelets have an increased adhesion to collagen. 28owever, with these highly sensitive proteomic application techniques, it is essential to note that the stoichiometric relationship to its global proteome is lost due to specific affinity enrichment of a particular PTM, as is the case with phosphopeptide isolation from a platelet protein sample digest.For example, we had also studied cAMP/PKA-dependent phosphorylation in platelets; however, we analyzed the global platelet proteome by 2D electrophoresis, 29 in contrast to the enriched phosphopeptide proteome as before. 25However, we could not identify nearly as many phosphorylation targets with two-dimensional gel electrophoresis (15 proteins) as the shotgun platelet proteomics studies (360 proteins) using phosphopeptide enrichment.On the other hand, we found that some PKA phosphorylation targets, such as RAP1B and LASP1, are significantly more phosphorylated in the presence of prostacyclin than the known platelet inactivation marker VASP, which would provide an attractive new diagnostic target for more robust quantification of platelet reactivity state. 29Also, the complexity of changes in the proteoforms of a protein relative to one another could only be observed in two-dimensional gel electrophoresis, as the unphosphorylated proteoforms, such as VASP and LASP, decreased in contrast DOI: https://doi.org/10.1080/09537104.2023.2220046 Considerations for Platelet Proteomics in Diseases 3 to the increase in the phosphorylated ones.These different phosphorylation profiles are undetectable in a shotgun analysis of the enriched phosphoproteome from a similar platelet inactivation experiment. 25

Conclusion
In conclusion, platelet proteomics is a helpful tool for studying disease-related changes globally and unbiasedly.However, greater awareness should be raised that quantitative stoichiometric relations and changes in a protein's most diverse proteoform profiles are hardly detectable in bottom-up proteomics.Thus, information, for example, in proteolytic hemostatic regulations, can be lost in such platelet proteome analyses.On the other hand, affinity enrichment, in combination with bottom-up proteomics, can characterize a selected group of specific proteoform peptides much more comprehensively and sensitively.Therefore, it would be advantageous to study platelet-dependent diseases in a combination of top-down and bottom-up proteomics to gain a comprehensive insight, but also with the possibility to capture the regulatory diversity of proteoforms.