Genetic Maps of Blood Proteins Open New Paths for Cancer Drug Repurposing

A huge international team has just completed the largest study ever on how our genes control the levels of proteins circulating in our blood. The work, led by researchers at Queen Mary University of London and the Berlin Institute of Health, brings together data from more than 78,000 people across 38 cohorts worldwide and has been published in the journal Cell. For cancer patients and researchers, this kind of study could quietly reshape how we find drug targets, repurpose existing medicines and monitor disease over time.

Most people know that DNA carries our genetic information, but it is easy to forget that its main job is to provide instructions for building proteins. Proteins do most of the real work in the body: they form structures, carry signals, control metabolism and steer the immune system. When disease develops, the underlying problem is often reflected in disturbed protein levels or in proteins that no longer function properly, sometimes long before symptoms appear.

Over the last two decades, genome‑wide association studies have linked thousands of genetic variants to different diseases, including many types of cancer. However, these studies often stop at “this region of DNA matters”, without telling us clearly which protein to target or how it affects disease. By measuring thousands of proteins in blood and linking these levels to genetic variation, this new study builds a crucial bridge between DNA, the blood proteome and clinical disease. That bridge is exactly where most drugs act, so understanding it in detail is highly relevant for future therapies.

The team carried out a large “proteogenomic” analysis, combining genetic data with blood protein measurements from many existing cohorts and analysing them together as one very large dataset. They harmonised the different protein measurement technologies, used large‑scale statistical models, and then applied machine‑learning methods to see how genetic variants change protein levels and how those protein changes relate to disease risk. This allowed them to map thousands of protein quantitative trait loci, or pQTLs, which are places in the genome where genetic differences lead to measurable differences in blood protein levels. They then linked these pQTLs to known genetic risk for a range of diseases, effectively building a molecular “disease map” in which diseases are connected by shared gene–protein patterns rather than by organ or ICD code.

To show how this can translate into therapy, the researchers highlight TYK2, a kinase involved in immune regulation that is already targeted by drugs in conditions like psoriasis. By combining genetic signals, protein data and disease information, they found strong evidence that inhibiting TYK2 may also be helpful in rheumatoid arthritis, suggesting a clear drug‑repurposing opportunity. The key idea is simple but powerful: if inherited differences that reduce or increase a protein are clearly linked to lower or higher disease risk, then drugs that push the protein in the same direction as the protective genetic pattern are more likely to be effective and safe. The same logic can be applied to cancer, where many targets have been selected with far less genetic and proteomic evidence.

For oncology, this type of map could be particularly valuable. Cancer is not one disease but many, driven by complex combinations of inherited and acquired changes. Blood proteins sit in an interesting position because they can provide a readout of tumour biology indirectly, for example through inflammatory cytokines, growth factors or angiogenesis markers, and they also capture the host response, such as systemic inflammation or immune activation. At the same time, they can be measured repeatedly from a simple blood draw, which is ideal for monitoring.

Using this proteogenomic framework, researchers can strengthen target validation in cancer. If a certain protein is elevated in a tumour, this alone does not prove that blocking it will help the patient. But if genetic variants that raise that protein are also associated with higher cancer risk or worse outcomes, this makes it a far more convincing target. If the genetic pattern points in the opposite direction, targeting that protein might be risky. The large catalogue of genetic–protein links from this study can therefore be mined to select or exclude potential oncology targets more rationally.

The molecular “diseasome” created in this work may also expose shared pathways between cancer and non‑cancer diseases. Many cancers share inflammatory, metabolic or vascular signalling with autoimmune and cardiometabolic conditions. If a protein sits at the centre of such a shared module, and there is already a licensed drug acting on that protein for a non‑cancer indication, that agent becomes a logical candidate for repurposing in oncology, either as an anti‑tumour therapy, an immune modulator or a supportive‑care drug. This is especially interesting in settings like metastatic disease, where combination approaches and host‑directed strategies are increasingly important.

Traditionally, many oncology drug targets have been discovered in tumour cell lines, animal models or small patient series. These approaches remain essential, but they often lack strong support from human genetics. The new proteogenomic map shows that it is now realistic to start from human data at scale, identify proteins that lie on causal paths from genotype to disease, and then use functional work to validate and refine those targets. For cancer drug development, this can mean prioritising targets where genetic, protein and clinical evidence align, de‑prioritising weak candidates earlier, and systematically scanning existing drug catalogues for repurposing opportunities.

Source.