Could serum proteins act as holistic biomarkers for personalised medicine? Whilst personalised medicine is a highly promising new treatment, in many cases its full potential has not yet been reached. One main restriction is that novel biomarkers (used to predict treatment efficacy) can only be validated in clinical settings. This makes personalising treatments risky for patients as well as expensive and time consuming. This highlights the need to identify biomarkers with the highest diagnostic and predictive accuracy. To combat this, a proposed strategy is to use a broad-based measure, which holistically captures the molecular state of individuals: serum proteins.
Protein profiling technology has evolved in recent decades. We are now capable of measuring 1000s of proteins simultaneously in large population studies. A recent paper, published in Cell, has proposed that serum proteins could provide an all-encompassing molecular measure, which could have profound impacts on therapeutic drug development as well as early disease diagnoses and interventions.
Vast networks of serum proteins are associated with disease:
The cardiovascular system regulates the flow of blood to and from all tissues in the body. This maintains a global homeostasis, controlling our physiological temperature, oxygen transportation and the transfer of nutrients and waste. This highly sophisticated coordination across various tissues of the body is facilitated by blood.
Separate work published in Science suggests that serum proteins exist in networks. These groups of proteins share similar functions but originate from different tissues and facilitate homeostasis through intricate cross-tissue regulation. For example, insulin and glucagon are known regulators of blood glucose homeostasis. However, there are also other secreted proteins produced in a range of tissues, which also aid this homeostasis. Thus, a complex web of interactions forms, mediated by serum proteins travelling in the blood. Importantly, the disruption of this careful balance can lead to the development of diabetes, obesity and other diseases.
Proteins in circulation have been closely linked to a wide range of diseases, but can also be used to detect diseases which may arise after blood has been sampled. Analysing serum protein networks may therefore provide a robust reflection of the existing and future disease state of an individual.
Complex cross-tissue regulation of serum protein networks:
The interaction of serum proteins across various bodily tissues is highly complex. Protein quantitative trait loci (pQTLs) are naturally occurring DNA sequence variants, which are linked to protein expression. These pQTLs can be located proximal to (cis) or distant from (trans) the gene encoding a regulated protein. Recent studies have discovered many cis-trans pairs of serum proteins.
For example, a serum protein in the liver is controlled by a cis genetic variant, close to its coding region. However, the same variant also controls the levels of other serum proteins in the brain, spleen and liver, and the coding regions for each of these proteins exist on different chromosomes. This reveals a deep and complex network of serum proteins, which spans across all tissues of the body and is under genetic control. This offers strong support for the role of these proteins in the development of complex diseases.
Serum proteins: the link between genetics and disease
Over time, large-scale genome-wide association studies (GWASs) have recorded over 70,000 DNA variants related to disease phenotypes. These GWASs have also demonstrated that complex disease traits are linked to a multitude of risk loci, each producing individual small effect sizes.
When GWAS risk loci were linked to serum proteins, many of the identified loci were found to regulate several proteins in cis or trans, many of which were also involved in the same protein networks. This indicates that disease-associated genetic variants may impact key protein networks in the body, causing the breakdown of essential bodily processes. This paper proposes that many of the genetic signals for common diseases, are cis and/or trans acting pQTL pairs, which connect different tissues in the body and produce diverse disease-related effects via the serum protein network.
A serum protein-based biomarker platform:
Previous studies have demonstrated that for individual protein biomarkers used to diagnose disease, ~75% differ significantly throughout the population. These also depend on genetic and clinical factors which are often not linked to the disease of interest. Thus, their generalisability for clinical application is reduced.
In comparison, protein clusters have been demonstrated to provide more robust biomarkers for disease diagnosis. There is also evidence that individuals can be grouped based on the particular serum protein networks which they possess.
This paper suggests that the production of a large, proteomic screening
platform may allow the categorisation of individuals by their molecular state in a holistic, disease-relevant manner. The routine and holistic assessment of individual disease states could provide thorough clinical evaluations and potentially allow for early diagnoses. A serum protein-based platform could identify many different diseases at once, providing depth that is missing from current screening practises.
Many aspects of the current healthcare system rest on treating patients once clinical symptoms of disease have already developed. Whilst the clinical trials for this technology would be expensive, and involve the testing of many otherwise healthy individuals, it may be highly beneficial to shift research paradigms towards disease prevention and early intervention.
Final Remarks:
Whilst much remains to be uncovered and investigated, it seems that serum proteins exist in coregulatory networks, under genetic control, bridging all bodily tissues and correlating with all common diseases. Thus, they provide the opportunity for detailed measurements of the molecular state of individuals, and across time.
Producing a platform capable of measuring large numbers of proteins precisely and cost effectively will not be easy. However, with technological advancements and investments, it could become feasible for these measurements to be routinely collected, improving treatments, drug development and early interventions.
