The Hundred-Person Platelet Releasate Project (100PREPs): A Barcode for Human Health and Disease

Background:

Upon activation platelets release a host of soluble and vesicular signals, collectively termed the 'platelet releasate' (PR). The contents of this PR play a significant role in haemostasis, wound healing, tissue regeneration, inflammation, and pathologic sequelae. We recently established that the proteome of the platelet releasate was highly reproducible across 32 healthy adult donors, with low variation in secretion levels for individual PR proteins. As there is accumulating evidence that circulating platelets may sense or be 'educated' by their environment, we hypothesised that the PR may store a relevant and bespoke collection of molecules during pregnancy and in platelet-related disease.

Methods:

PR from 18 healthy pregnant women, 22 pregnant women diagnosed with early-onset preeclampsia (EOP) (onset <34 gestational weeks), 15 people with relapsing-remitting multiple sclerosis and 15 people with primary-progressive multiple sclerosis were obtained with informed consent. After trypsin/lys-C double digest, each platelet releasate was individually subjected to label-free quantitative proteomic analysis. A machine learning approach based on differential peptides was used as a training set and then to predict outcome.

Results:

69 PR proteins were found to be differentially released from platelets in healthy pregnancy in comparison to age-matched non-pregnant female controls, including several pregnancy-specific proteins. A further 36 PR proteins were statistically modified in EOP, a disease associated with life-threatening consequences for both mother and baby, and another 26 proteins were altered in multiple sclerosis. Based on a training set of differential PR peptides, we developed several diagnostic algorithms, including a proof-of-concept pregnancy algorithm that could predict outcome with 100% accuracy. Strikingly, our proprietary algorithm based on several platelet releasate peptides could predict the severity of maternal outcome in preeclampsia. Moreover, again based on specific platelet-derived peptides, we could differentiate between clinical subtypes of multiple sclerosis.

Conclusion:

We found that the PR proteome significantly changes in physiologic and pathologic conditions. We conclude that the PR is a barcode for the health status of an individual at a given time and provides an easily accessible discrete set of peptides and proteins that can differentiate disease.