Editor's note:

Plasma proteins are important "players" in a wide range of biological processes, taking part in growth, repair, signal transduction, transport and defence against infection. At the same time, plasma proteins are differentially regulated as disease develops and progresses. A systematic and precise understanding of the properties of plasma proteins in normal physiological and pathological states will therefore effectively advance the discovery of disease molecular markers and drug targets.

The following article comes from the Human Proteome Organization, by HUPO

"Proteins in plasma have been studied since before we knew genes existed."

—— Leigh Anderson & Norman  Anderson

The human plasma proteome plays a pivotal role in research on human biology and pathophysiology, and holds the promise of creating a disruptive revolution in precise disease diagnosis and treatment monitoring. 

On the one hand, plasma proteins are important "players" in a wide range of biological processes, participating in key life activities such as growth, repair, signal transduction, transport and defence against infection. On the other hand, plasma proteins are differentially regulated as disease develops and progresses. A systematic and precise understanding of the properties of plasma proteins in normal physiological and pathological states will therefore effectively advance the discovery of disease molecular markers and drug targets.

Dilemma | "Science fiction"

The analysis of the human plasma proteome is, however, extremely challenging, and that challenge comes not only from the plasma proteins themselves but also from the limitations of analytical techniques and platforms.

First, plasma is the most complex and most diverse sample in body fluid analysis, because (1) the difference in concentration between the highest- and lowest-abundance proteins in plasma spans more than 10 orders of magnitude. High-abundance proteins such as albumin and immunoglobulins account for 80% of total plasma protein, while the remaining proteins exist at very low concentrations — yet they are associated with the evolution of many diseases and are also potential biomarkers. Removing the interference of high-abundance proteins is therefore critical in plasma proteome analysis. Second, (2) plasma contains a large number of proteoforms, such as endogenous short peptides and proteins with different post-translational modifications (PTMs), adding another layer of complexity and analytical difficulty. In addition, (3) individual differences arising from age, sex and genetic factors introduce a higher level of variability, making the analytical workload and research difficulty harder still to imagine.

The technical bottleneck is another key reason why plasma protein research has been so hard to advance. Conventional proteomic techniques such as two-dimensional gels, limited in detection range, cannot deliver unbiased, high-depth analysis of the plasma proteome. In other words, in the era before mass spectrometry was widely applied, researchers had no way to "see" more plasma proteins accurately, which made the discovery of plasma protein biomarkers something of "science fiction" for a time.

Beyond the problem of "seeing", there was also a gulf between the number of plasma proteins discovered and their application in clinical diagnosis. A 2002 article pointed out that at that time only a few hundred plasma proteins could be detected, and the number that could be applied to clinical diagnosis and treatment monitoring was a mere few dozen — very few indeed.

▲ Between 1994 and 2002, the number of FDA-approved plasma protein markers showed a downward trend (source: The human plasma proteome: history, character, and diagnostic prospects)

Pathfinders | The founding of the HPPP

To address these problems, Sam Hanash, the first president of the Human Proteome Organization (HUPO), launched a key pilot programme in 2002 — the Human Plasma Proteome Project (HPPP). The project was led by Gilbert Omenn, one of HUPO's founders, with the participation of 35 proteomics laboratories from 13 countries.

▲ Gilbert Omenn (left) and Sam Hanash (right)

The HPPP is one of the "pioneers" of human proteome research, both because it was one of HUPO's early flagship projects and because of the irreplaceable research value of plasma proteins. First, (1) the plasma proteome is the most complex category of the human proteome, encompassing the sub-proteomes of different tissues; (2) plasma is the principal clinical sample. Compared with other body fluids (such as cerebrospinal fluid, gastric juice, bile and synovial fluid), plasma is easier to obtain and to standardise; (3) plasma proteins are of great significance for disease diagnosis and efficacy monitoring. Because plasma proteins travel with the blood circulation throughout the body, they can reflect a wide range of disease problems, and plasma has become a potentially rich source of protein biomarkers for clinical diagnostic monitoring and for judging drug response. A number of plasma and serum proteins have already been identified as potential biomarkers of disease, including cardiovascular disease, autoimmune disease, infectious disease and neurological disease; (4) plasma proteins correlate poorly with their corresponding mRNA expression, so studying them at the protein level serves as a complement. 

At the same time, given the complexity of plasma proteins noted above and the diversity of separation, analysis techniques and platforms, the HPPP defined its initial research goals: (1) to compare the advantages and limitations of proteomic analysis platforms, such as sensitivity and analytical depth; (2) to analyse human plasma/serum reference specimens; and (3) to build a knowledge base of the human plasma proteome.

Progress | The human plasma proteome under mass spectrometry

In fact, long before the HPPP was launched, some scientists had already begun to gnaw on this "hard bone" of the plasma proteome.

In the 1970s, with the introduction of high-resolution two-dimensional electrophoresis at the time, the detection of plasma proteins became feasible. In 1977, two researchers — Leigh Anderson (founder of the Plasma Proteome Institute) and Norman Anderson (Senior Scientific Advisor of the Plasma Proteome Institute) — successfully identified 49 plasma proteins.

▲ Leigh Anderson and Norman Anderson successfully identified 49 plasma proteins using two-dimensional gel electrophoresis

In 2002, Leigh and Norman pressed on, compiling 289 plasma proteins from the literature — the first large-scale plasma protein compilation. In the same year, Joshua Adkins (Deputy Director of the Biological Sciences Division, Pacific Northwest National Laboratory) used on-line reversed-phase microcapillary liquid chromatography coupled with ion trap mass spectrometry to analyse proteins in serum and successfully identified 490 plasma proteins.

▲ Leigh Anderson and Norman Anderson compiled 289 proteins (left); Joshua Adkins identified 490 plasma proteins using on-line reversed-phase microcapillary liquid chromatography coupled with ion trap mass spectrometry (right)

With the advance and widespread application of mass spectrometry, plasma proteome data emerged rapidly, but the comparability between data sets was a major problem. In 2003, the HPPP obtained a batch of samples from the world's major human populations — Caucasian-American, African-American and Asian-American — and prepared them as pooled plasma and serum.

All 35 laboratories collaborating with the HPPP received these samples, but the technology platforms and the data extraction and analysis methods they used differed, leading to large discrepancies in the results. The project report stated that 9,504 non-redundant proteins were detected in total, and that every laboratory could detect at least one peptide from these 9,504 proteins. Of these, 3,020 proteins could be detected with at least two peptides, and 1,500 proteins with three peptides. Because the data were poorly comparable, the researchers re-analysed the report in 2006 and concluded that only 889 proteins had high-reliability evidence of their existence.

Gradually, mass spectrometry developed towards the twin goals of "high" sensitivity and "high" accuracy, and the scale of quantitative protein identification rose with it. Compared with early plasma proteome databases, today's MS2 spectra are of high resolution, greatly improving the accuracy of sample analysis and greatly reducing the false positive rate. Researchers can therefore "see" "deeper" and "more" — observing more peptides and raising sequence coverage.

In 2005, Eric Deutsch (Chief Scientist of the Moritz Laboratory) and eight laboratories applied liquid chromatography–tandem mass spectrometry to the study of plasma/serum and mapped the experimental results onto the Ensembl human genome database (Ensembl Genomes). The data showed that 6,929 peptides were mapped to 960 proteins. In 2011, Terry Farrah's research team (formerly of the Institute for Systems Biology) identified 1,929 high-reliability proteins (FDR of 1%); in 2013, 3,553 non-redundant proteins were successfully identified.

In 2015, Hasmik Keshishian's team (Senior Scientist at the Proteomics Research Center of the Broad Institute of MIT and Harvard) used plasma from four patients with hypertrophic cardiomyopathy (HCM) who had undergone myocardial infarction surgery as samples, applying iTRAQ technology and a high-resolution Q Exactive Plus Orbitrap instrument. They successfully quantified at least 3,400 proteins in each sample, with 5,300 proteins identified in some samples (FDR<0.02%). In 2016, researchers identified 3,509 plasma proteins meeting the standard of the HPP Guidelines for Mass Spectrometry Data Interpretation v2.1.

On the other hand, the PeptideAtlas database (which stores and records peptide and protein mass spectrometry data from different species), as part of the HPPP, has continuously collected mass spectrometry data from laboratories worldwide since 2004. Through analysis and processing of the raw data, a standardised, unified peptide compilation is produced and made public to researchers around the world. Moreover, the Human Plasma PeptideAtlas has, since 2005 and in step with the growth and change of plasma proteome data, published updated high-stringency snapshots of plasma proteins. The latest Human Plasma PeptideAtlas data (2021-07) show that researchers identified 4,395 canonical proteins from 240 mass-spectrometry-based experiments.

▲ The latest Human Plasma PeptideAtlas data (2021-07) show that 4,395 canonical proteins have been identified

In 2021, Eric Deutsch and other scientists published the article "Advances and Utility of the Human Plasma Proteome" on the progress and application areas of the human plasma proteome. The article points out that plasma proteomics research and its related applications will inevitably play an ever greater role in precision medicine, and has already brought more research ideas to three hot research areas: COVID-19 infection, ageing and hemostasis.

▲ In 2021, Eric Deutsch and other scientists published "Advances and Utility of the Human Plasma Proteome" on the progress and application areas of the human plasma proteome

Awaiting the metamorphosis | The future of plasma protein research

Research on the plasma proteome will undoubtedly bring disruptive progress to the life sciences. Yet it is not hard to see that over the past 20 years the speed and number of human plasma protein identifications have remained greatly constrained.

Although today's proteomics technologies and analytical methods have made considerable progress, with automated high-throughput separation, identification technologies and platforms appearing one after another, they still cannot match the high complexity of plasma proteins. As a result, the data measured by scientists lack sufficient precision to reveal deep-level information, and are even harder to apply to clinical diagnosis and treatment monitoring. 

In truth, this kind of research predicament is the normal state of science. Research has never been a myth; it is a protracted war of attrition. It is water held at a constant 99 °C: let a disruptive technology or method appear and it shoots up to boiling; then it falls back to 99 °C and waits for the next boil, over and over. Metamorphosis takes time, and the more tortuous and lengthy the process, the more beautiful and striking the result.

In the same way, although the progress of the HPPP has not been ideal, we can still optimistically predict that in the near future it will help researchers and clinicians draw a precise and comprehensive picture of disease treatment and monitoring.