How can we use proteomics to learn more about platelets?

Abstract Proteomics tools provide a powerful means to identify, detect, and quantify protein-related details in studies of platelet phenotype and function. Here, we consider how historical and recent advances in proteomics approaches have informed our understanding of platelet biology, and, how proteomics tools can be used going forward to advance studies of platelets. It is now apparent that the platelet proteome is comprised of thousands of different proteins, where specific changes in platelet protein systems can accompany alterations in platelet function in health and disease. Going forward, many challenges remain in how to best carry out, validate and interpret platelet proteomics experiments. Future studies of platelet protein post-translational modifications such as glycosylation, or studies that take advantage of single cell proteomics and top-down proteomics methods all represent areas of interest to profiling and more richly understanding platelets in human wellness and disease. Plain Language Summary What is the context? Platelets are well known as cellular mediators of hemostasis and drivers of thrombosis and inflammation. Thousands of different proteins come together to make up the molecular structure of platelets and support their functions in health and disease. What is new? “Proteomics” refers to laboratory approaches that study many proteins simultaneously to provide details on proteomes and the biology of a given system of interest. Many different high-throughput biochemical methods can be considered as proteomics, including mass spectrometry analysis. What is the impact? Proteomics has already identified more than 5,000 individual proteins in human platelets and provided important insights into how platelets may serve as biomarkers or therapeutic targets in human health and disease. In coming years, advances in proteomics technologies will likely help to characterize additional proteins and molecular systems in platelets to shed light on platelet-mediated mechanisms of physiology and pathology that have been largely unapproachable with other methods.


Introduction: how have platelets already been studied with proteomics?
By definition, proteomics tools systematically identify, characterize, and quantify protein features, or "proteomes" of a given biological sample of interest. 1Since the advent of proteomics technologies in the 1980s, there has been significant, continual progress in identifying and characterizing the protein composition of platelets. 2,3Using highthroughput methods such as biochemical labeling, 2D gel electrophoresis, affinity array capture, and mass spectrometry, at least 5,000 unique proteins and proteoforms (i.e., all of the different molecular variants of a protein product of a single gene, including changes due to genetics, alternative splicing of RNA transcripts, truncations, and post-translational modifications) 4 have been observed in platelets. 5n the past decade, advances in quantitative mass spectrometry and computational biology tools have allowed for even greater sensitivity in proteomic analysis, where approximately 3,000 to 5,000 unique proteins have been identified in human platelets.Notably, in 2012, Burkhart et al provided the first quantitative mass spectrometry analysis of human platelet proteomes, which identified and rank-ordered the relative abundances of over 3,000 different proteins expressed in human platelets from multiple healthy subjects. 6To date, this and other studies of platelet proteomes report that the majority of proteins expressed by platelets have low coefficients of variation between individuals regardless of age, sex, or pathology. 7,8For instance, by protein copy number and total mass, most of the protein content of platelets is made up of cytoskeletal and structural proteins, as well as metabolic, secretory, signaling, and other proteins essential to platelet function.][11][12] In addition to identifying and cataloging proteins in platelets, mass spectrometry and other proteomics tools have also allowed for the characterization and quantitation of protein posttranslational modifications (PTMs), such as reversible phosphorylation, which transduces signaling events to platelet function.Initial computational estimates noted that as many as 28 000 possible phosphorylation sites may be found on tyrosine, serine and threonine residues of the thousands of different proteins that are expressed by platelets. 13At present, results from mass spectrometry experiments show that at least 3,000 phosphorylation sites on over 1,000 different proteins are significantly regulated in platelet activation programs, where at least 300 signaling events can be causally placed into established signaling pathways. 14hese and other protein modifications, indicative of platelet activation state, are expected to offer a more mechanistic approach to assaying platelet phenotype relative to protein content on its own.For instance, phosphorylation of the protein kinase A substrate Vasodilator-Stimulated Phosphoprotein (VASP) can serve as a marker of the activation state of circulating platelets. 15][18] In recent years, advances in mass spectrometry, computational and other proteomics technologies have continued to expand our understanding of the molecular basis of platelet phenotype and function. 9For instance, in 2018, the Maguire group cataloged more than 700 unique proteins that are consistently released from platelets ex vivo following stimulation with thrombin. 19uantitative proteomics efforts have also detailed how the protein content of platelets changes over their 5-10 days in circulation as platelets lose hemostatic activity. 20Proteomics studies of platelets from different species with adaptations of interest to human health and disease have also shed light on mechanisms of platelet function.Notably, recent proteomics studies of platelets from hibernating 13-lined ground squirrels and brown bears have pointed to new molecular targets in platelets that may protect against platelet cold storage lesions or deep vein thrombosis. 21,22 major goal of platelet proteomics studies has been, and remains, to understand how platelets change in wellness and disease, where platelets are attractive candidates as biomarkers or therapeutic targets.Accordingly, many studies have used proteomics methods to profile platelet proteomes in different inflammatory, thrombotic and other pathological states, where platelets are suspected participants in disease or poor clinical outcomes.Studies of platelet proteomes in conditions such as heart failure, 23 cancer, 24 myocardial infarction, 25 and COVID-19, 26 all note proteomic changes in platelets compared to specific controls (usually platelets isolated from self-reported healthy, age and sex matched subjects).Interestingly, these studies suggest a general, physical interaction between platelets and leukocytes in disease -especially neutrophils -as the alarmin protein calprotectin, or, S100A8/A9, becomes associated with platelets in a number of acute and chronic pathologies.It is increasingly apparent that inflammatory markers such as S100A8/A9 find their way from leukocytes to platelets in disease in a manner that can alter platelet function. 27Nonetheless, much work remains in determining whether and how such changes in platelet proteomes may reflect a specific disease state or therapeutic target.Moreover, it remains unclear whether such findings are a common hallmark of disease, especially as interlaboratory and preanalytical variables can dampen the reproducibility and impact of proteomics studies. 28

How should platelets be prepared for proteomics experiments?
Although mass spectrometry and other proteomics tools offer great potential for scientific and clinical research on platelets, consensus and discussion on how platelets should be studied with proteomics remains limited.Currently, most platelet proteomics studies use similar protocols to isolate and prepare washed platelets from anticoagulated blood and platelet-rich plasma (PRP) using modified HEPES Tyrode's buffer, as well as comparable centrifugation times and forces. 29However, as more research teams outside of the hematology community become interested in platelets, and, as more advanced proteomics methodologies become increasingly complex, it is likely that some commonalities will increasingly diverge.][32] While methods for washing platelets have been generally consistent among groups studying platelet proteomes, the handling of blood samples before platelet preparation can vary significantly and may have a significant impact on proteomics results.4][35] As blood samples await processing, proteins from platelets and other blood cells are released into the plasma, altering the proteome composition of the plasma in concert with changes in platelet protein content.It is expected that preanalytical variables will also impact platelet proteomes in a similar way, but the effects of many sample processing variables on platelets have yet to be thoroughly examined.
To this end, our group recently used a quantitative mass spectrometry approach to investigate how sample processing time and blood anticoagulant affect the platelet proteome. 36We found that delays in sample processing were associated with a loss of secretory proteins from platelet proteomes, in line with studies of preanalytical variables in plasma proteomics.We also found that anticoagulant choice can impact the platelet proteome, with citrate anticoagulants having a lesser effect on secretion, complement activation, and platelet: leukocyte interactions compared to EDTA or heparin.

How should platelet proteomics experiments be designed and executed?
In addition to accounting for preanalytical variables and methodological factors, it is essential to consider several prescribed best practices when conducting platelet proteomics studies. 28For example, in order to avoid batch effects from sample processing, it is recommended that a single team at a specific site obtain, prepare, and store platelet samples before analysis.However, this scenario may not be feasible in practice, particularly for samples obtained from specialized collaborators or multiple clinics.In general, if there is any uncertainty regarding the quality or source of the samples prepared for a study, we would recommend avoiding further analysis and moving on to a different sample set or project.Mass spectrometry analyses require significant effort and typically cost tens of thousands of dollars to produce a reliable dataset.Once the dataset is generated, it may take several months to examine the data for basic quality controls and to identify potential hypotheses.After becoming familiar and confident with a given dataset, another one to three years may be required to follow up on any hypotheses and validate results prior to reporting them in a manuscript or grant proposal.Therefore, initiating a project using samples with any level of doubt about their quality or source is unlikely to be worthwhile.
Several other general and specific questions will arise regarding specific platelet proteomics experiments.For example, how many samples and replicates should be included?What positive and negative controls should be used?Sample number is a challenging subject, and increasing the number of samples has both positive and negative effects on data quality.For instance, when designing a quantitative, ultrahigh resolution 18-plex tandem mass tag (TMT) experiment, it may be tempting to simply compare six groups with three replicates each.However, statistical power may be much better if only two groups are compared using nine vs. nine samples.Accordingly, researchers should consider whether they want conclusive data to publish a paper, or, if they want to screen samples in a shotgun, fishing expedition style experiment to identify targets to follow up.Given the cost of and time commitment of proteomics studies, many researchers may first intend to only run screening experiments, then later decide to try publish their results, only to find themselves trapped with a potentially revolutionary dataset that is not appropriately structured for publication.
Ideally, before conducting a platelet proteomics experiment, researchers should have a specific plan in place for what they will ultimately do with the resulting dataset.It is important to have a clear, well-controlled, research question or goal in mind for proteomics experiments and not conduct experiments merely for the sake of using advanced technologies.Researchers planning to try out proteomics experiments should carefully consider the bestand worst-case scenarios of what their results may or may not provide.On a practical level, it is also important to note that proteomics projects will likely require several months or years of collaborative, open, team-style science before their presentation or publication to ensure that their efforts will not eventually be abandoned and can have the support of all players necessary for completion.

How should platelet proteomics data be validated?
A general goal of many platelet proteomics experiments is to discover molecular features, targets and processes in platelets, and, to generate novel hypotheses around mechanisms of platelet phenotype and function.The subsequent follow-up, hypothesis testing, verification and validation of proteomics results can be a contentious issue, particularly when interdisciplinary teams of platelet biologists and omics scientists come together for big projects.It is important to recognize these and other potential cultural and philosophical differences between collaborating research groups to avoid internal conflict and ensure that results are ultimately shared in a meaningful way with the wider scientific community.For instance, proteomics scientists with expertise in mass spectrometry are often well-versed in the problems associated with using antibodies and other commercial probes to study proteins, as many such reagents have fueled a large part of the current reproducibility crisis. 37Moreover, many proteomics scientists are not experienced in carrying out and interpreting black box assays from commercial sources.This can result in a disconnect with platelet biologists, who have traditionally relied on many such tools to build their field over the past several decades.In contrast, mass spectrometrists may ultimately present platelet biologists with proteomics data that points to obscure and, perhaps, seemingly nonsensical results, leaving platelet biologists uncertain as to how to interpret the data and move forward.
Many validation-related steps should occur well before any samples are analyzed with mass spectrometry, such as assessing platelet count as well as leukocyte and erythrocyte content.One potential option for validation is to examine samples and datasets from as many perspectives as possible.For example, if a quantified proteome for a platelet sample of interest contains abundant levels of proteins that have never before been found in platelets, further investigation may be necessary.Many different groups have independently quantified and rank-ordered proteins in platelets from hundreds of different subjects globally, and there is an emerging consensus around which proteins are highly expressed in platelets and which are found in lower quantities.Comparing the rank of proteins in a sample to known values can serve as a good internal quality control.In cases where a traditional assay such as a Western blot or ELISA is needed for validation, but no available antibody reagent exists, targeted assays using peptide and protein standards can be developed to quantify proteins or PTMs in a sample as a follow-up to shotgun experiments.While there is no straightforward answer for how to best validate proteomics results, continued collaboration and compromise between proteomics scientists and platelet biologists will be essential to move the field forward.

What are the next exciting things that can be looked at in platelets with proteomics?
9][40] As proteomics technologies continue to advance, platelets will continue to be a valuable sample for discovery, including in areas beyond PTMs such as phosphorylation, acetylation, methylation, and especially, glycosylation. 41 42,43In recent years, studies have revealed extensive glycosylation of platelet proteins, particularly in platelets that are prepared and stored for transfusion, raising questions about how glycosylation may impact platelet phenotype and function. 44,45Investigating the role of glycosylation in platelet biology represents an important area of future research that is likely to yield significant mechanistic as well as clinically translational insights.
Another emerging area of interest to platelet profiling is single cell proteomics, which involves analyzing the proteomic details of individual cells, one at a time, using mass spectrometry tools. 46hile there are many practical barriers to carrying out, analyzing, and understanding single cell proteomics experiments, recent single cell proteomics studies of macrophages provide a working model for how such studies are possible and valuable for platelets. 47Moreover, recent cell sorting and flow cytometry analyses of protein heterogeneity in single platelets provides additional rationale for further considering single cell proteomics studies of platelets. 48n addition to more traditional "bottom-up" style mass spectrometry analyses of proteomes, where lysates are digested into specific peptides and then analyzed, "top-down" proteomics methods analyze proteins before they are digested, allowing for the study of intact proteoforms. 49,50A recent top-down analysis, "The Blood Proteoform Atlas," recently identified more than 30 000 unique proteoforms in 21 blood cell types, including platelets, where more than 700 proteoforms were found to be unique to platelets. 51While the physiological relevance and translational impact of these specific results remain to be actualized, they demonstrate the feasibility of top-down studies in platelets, and point to exciting opportunities for discovery and advancement in the field.

Conclusion
Proteomics tools have already helped to define the platelet proteome and support systems biology studies of platelets in health and disease.The total protein content of platelets shows mostly minor or specific variations between individual human subjects, regardless of age, sex or health, and small changes in platelet protein content likely accompany or support maladaptive platelet function.While it is clear that changes in platelet proteomes can reflect changes in platelet function, much work remains before specific changes in platelet protein systems can be considered as mechanistic targets or biomarkers in platelet associated disease states.As proteomics tools become more advanced, it is important to work together as a community of proteomics scientists and platelet biologists to design, carry out, validate and interpret studies of platelet proteomes.There is no doubt that future advances in proteomics methods will enrich our understanding of platelet function, where platelets will also serve exemplary roles in proteomic, cellular and systems biology studies of wellness and disease.

Disclosure statement
No potential conflict of interest was reported by the author(s).