Shopping Cart
Professional article · Analytics & standardisation
PRP beyond platelet count: what the proteome and metabolome reveal
PRP is not a uniformly composed preparation. Platelet count, leukocyte content, plasma matrix and processing explain only part of its biological variability. Proteomics and metabolomics reveal further differences: recent studies show formulation-specific protein and metabolic signatures as well as initial biomarkers that correlate with clinical response. What research has not yet established is a routine-ready rule of the type “this molecular profile = this PRP protocol for this patient”.
716
Proteins in the TRAP-stimulated platelet releasate; 225 reproducibly detected in all donors studied
~600
Proteins across plasma, PRP and PPP — with more than 50% overlap between fractions
5–10 Mrd.
Total platelet-dose range associated with the largest functional effect in a 2026 knee-osteoarthritis meta-analysis; not a universal target
30
Differentially expressed proteins between responders and non-responders in a PRP cohort
Why growth-factor counting reaches its limits
The conventional explanatory model for Platelet-Rich Plasma (PRP) names four to six growth factors — PDGF, TGF-β, VEGF, EGF, IGF-1 and bFGF — and derives biological effects from them. Analytically, this model is highly simplified. Mass-spectrometric studies of the platelet releasate identified as early as 2009 716 proteins, 225 of which were detectable in all three donors. A later label-free analysis of thrombin-induced releasates found 318 proteins, nine of which were differentially released — including tetranectin (CLEC3B), protein disulfide-isomerase A3, factor V and fibronectin; notably, all nine were assigned to the GO category “extracellular vesicles”.
The direct comparison of fractions yields the more uncomfortable figure: almost 600 proteins were detected across plasma, leukocyte-poor PRP and PPP, with more than 50% overlap between formulations. A substantial portion of what is measured in a PRP preparation is therefore plasma proteome rather than platelet-derived secretory product.
Analytical consequence
The therapeutically differentiating part of the proteome lies in the low-abundance range and is masked by albumin and immunoglobulins. Without depletion or enrichment, analysis mainly measures what is present in every plasma sample anyway. In addition, platelets become activated ex vivo during sample processing and themselves release proteins and vesicles — a known source of bias in plasma proteomics.
Fig. 1 — The three analytical levels are not interchangeable. Growth-factor measurement by ELISA covers only part of level 1; it cannot support conclusions about redox state, energy metabolism or vesicle transfer.
Comparison of analytical depth
Each square represents one of the approximately 716 proteins identified in the stimulated platelet releasate. Select an analytical method.
5proteins detected
0,7 %coverage of the identified releasate
711not detected
Schematic representation. The reference is the extent of the releasate proteome reported in the literature; multiplex panels also measure analytes not contained in this dataset. Greater coverage does not automatically mean greater clinical relevance.
What metabolomics shows about preparation systems
An exploratory PRP metabolomics study published in 2025 used untargeted Q-TOF LC-MS to compare ten PRP types: two formulations from six donors and eight preparations from the blood of a single donor. At least within the single-donor part, this design makes it possible to examine differences in processing without interindividual blood variability.
In this study, the formulations differed in pathways related to inflammation, redox homeostasis, steroid metabolism, energy production and platelet activation. The authors regarded the preparation method as a major determinant of biochemical composition. Several preparations from the same donor also separated from one another in principal component analysis. This demonstrates molecular differences between preparations, but not clinical superiority of any particular formulation.
Chemically activated PRP (A-PRP)
BilirubinArachidonic acidCoproporphyrinogenCreatine
The authors interpret this pattern as high redox-buffering capacity combined with early inflammatory activation and pronounced energy availability. Arachidonic acid is a substrate for eicosanoid synthesis (prostaglandins, thromboxanes, leukotrienes); coproporphyrinogen III is an intermediate in haem biosynthesis. The functional interpretation discussed in the publication is exploratory and is not a clinically validated indication assignment.
Leukocyte-rich PRP (L-PRP)
L-TyrosinD-glucuronic acidNicotinamidribosid
L-tyrosine, a catecholamine precursor, was markedly elevated, as were D-glucuronic acid — a building block of hyaluronic-acid biosynthesis — and nicotinamide riboside as an NAD⁺ precursor. The authors classify the profile as immune-activating and remodelling-oriented. A selection rule for specific treatment cases derived from this has not yet been clinically validated.
Leukocyte-poor PRP (P-PRP)
low L-tyrosinelow D-glucuronic acidbalanced profile
Compared with L-PRP, P-PRP showed lower levels of several metabolites associated with inflammation and remodelling. The authors classify this as a less inflammatory profile. Whether a reliable clinical selection rule for intra-articular or other applications can be derived from this remains open.
PRP with added hyaluronic acid (TFF-PRP)
CarnosinL-ValinDocosahexaenoic acidTestosteronglucuronid
Metabolites associated with antioxidant capacity, tissue remodelling and anabolic signalling were elevated. The publication discusses these relationships mechanistically; the metabolomics analysis did not demonstrate a clinical advantage of this formulation.
The key methodological point
The enriched metabolic pathways are not based on the presence or absence of individual metabolites. The substances concerned are present in all formulations — it is the quantitative differences in abundance, that separate the groups. Personalisation here therefore means controlling proportions, not adding ingredients.
Compare formulation profiles
Ordinal classification of the four formulations studied along the axes that separated in the enrichment analysis.
No measured values. The levels reflect how the study authors describe each formulation on the basis of discriminating metabolites and enriched metabolic pathways. They are useful for understanding differences, not for selecting a preparation for a specific treatment case.
The same applies to Platelet-Rich Fibrin (PRF), which should not be equated with PRP from a processing perspective. A Mainz research group compared cell content, immune potential, proteome and secretome of PRF versus PRP using nano-LC-MS and, with comparable platelet counts, found differences in protein composition and abundance between PRF, PRP and plasma-free fractions. Formulation-specific proteomic signatures are also seen within the PRF family. These data describe biological differences; they do not demonstrate general superiority of any PRF variant.
The vesicle fraction: why growth factors alone do not explain what happens
Protein cargo
Label-free proteomics of sorted PRP vesicles showed that 32% of the identified proteins were assigned to the category “defence and immunity”, with highly significant enrichment for vesicle-mediated transport and wound healing.
Nucleic acids
PRP EVs transport platelet-derived miRNAs such as miR-24, miR-223 and miR-126 into target cells. Altered circRNA profiles after thrombin activation have also been described.
Organelles
Platelets release functional mitochondria as vesicle cargo that can influence the metabolism of recipient cells — a mechanism not captured by growth-factor measurements.
Greater purification is not monotonically better
For those pursuing EV-based approaches, a finding from platelet-lysate research is relevant: EVs obtained by tangential-flow filtration promoted skin-organoid formation and inhibited T-cell proliferation more strongly than EVs purified by column chromatography — and recombining the purified vesicles with the soluble fraction restored the effect. The reason is a protein corona which, unlike in synthetic nanoparticles, improves function. More intensive purification may therefore remove precisely what contributes to the effect.
Variability: processing and donor influence the molecular profile
The most practically consequential insight from the omics data concerns the relative influence of different factors. Preparation technique generates substantial metabolic diversity even without interindividual variation. Donor factors are therefore not irrelevant, but they are less predictable than the literature often suggests — and findings are in some cases clearly contradictory.
| Cohort | Finding on donor influences | Direction |
|---|---|---|
| Weibrich et al. n = 115 |
PDGF-AB 117 ± 63, TGF-β1 169 ± 84, IGF-I 84 ± 23 ng/ml; growth-factor content correlated only weakly with platelet count (r ≈ 0.35); no effect of age or sex except for IGF-I | no clear donor effect |
| Japanese cohort n = 39, LP-PRP |
Age negatively correlated with PDGF-BB (r = −0.32) and IGF-1 (r = −0.39); PRP platelet count positively correlated with TGF-β1 (r = 0.75), EGF (0.71), HGF (0.48), PDGF-BB (0.39) | age and count relevant |
| US cohort n = 102 |
All seven growth factors studied were higher in women and in individuals aged ≤25 years; four of seven differed significantly by sex and five of seven by age | age and sex relevant |
| Platelet proteome by age |
DIA mass spectrometry of washed platelets: high similarity between young and middle-aged donors, significant proteomic differences between older and young donors | cell state changes with age |
| Platelet proteome in diabetes |
Even freshly prepared platelets from people with diabetes show changes in multiple immunological and haemostatic mediators | comorbidity relevant |
A comparatively well-quantifiable parameter is the delivered platelet dose. A systematic review and meta-analysis on knee osteoarthritis published in 2026 classified PRP by the total deliverable platelet count into 1–5, 5–10 and more than 10 billion. Compared with hyaluronic acid, the group with 5–10 billion platelets was associated with the largest functional effect; more than 10 billion was not automatically superior. This is indication-specific evidence and not a universal therapeutic threshold or target for PRP. For the general clinical context of PRP, see also the article on mechanism, evidence and limitations of PRP therapy .
Fig. 2 — Schematic representation based on published system comparisons. The range of deliverable absolute dose results from platelet concentration × injected volume × number of injections and is often not reported in studies.
Calculate platelet dose
The deliverable absolute dose is limited by the processed blood volume, not by the enrichment factor. This module illustrates the relationship.
Amount of blood from which the concentrate is obtained
Reference range approx. 150–400 ×10³/µl
Proportion of platelets recovered in the concentrate
Determines concentration, not absolute dose
05 billion · DEPA A at >515 billion
6,00billion platelets per injection
1500×10³/µl in the concentrate
6,0×enrichment versus whole blood
ADEPA dose category
Calculation model, not a product or treatment claim: dose = platelet concentration × processed blood volume × recovery. The actual recovery of a system depends on the device, protocol and operator and must be determined in each case. DEPA categories: A >5 billion, B 3–5 billion, C 1–3 billion, D <1 billion. The DEPA dose category is a classification, not an efficacy or quality score.
Responder signatures: early clues to possible biomarkers
At this point, description begins to move towards prediction. A research group compared PRP samples from patients with knee osteoarthritis according to clinical response and found 30 differentially expressed proteins, which separated responders from non-responders. Cathepsin G and the calprotectin subunit S100A8 were upregulated in responders, while no substantially upregulated protein was found in non-responders.
| Protein | Function | Correlation with Δ KOOS pain |
|---|---|---|
| S100A8 | Calprotectin subunit, alarmin | r = 0,65 (positiv) |
| Cathepsin G | neutrophil serine protease | r = 0,51 (positiv) |
| Glutathionperoxidase 1 | antioxidant enzyme | r = −0,55 (negativ) |
| Integrin-linked Kinase | adhesion and signalling mediator | r = −0,50 (negativ) |
The result is counterintuitive and therefore particularly interesting: neutrophil granule proteins and an alarmin as positive markers do not fit well with the blanket doctrine that “leukocyte-poor is fundamentally better”. Complementary functional assays show that in-vitro responses to PRP in a macrophage–fibroblast co-culture system can be linked to composition and clinical course. Algorithmic analysis has also entered the field: a model based on 102 patients used pre-intervention clinical and biochemical features to predict six-month response.
Limitation that must be kept in mind
The cohorts are small (15 responders versus 11 non-responders) and have not been prospectively externally validated. It also remains unclear whether the markers are causal or merely reflect an inflammatory osteoarthritis phenotype that is inherently responsive to anti-inflammatory interventions. In that case they could still be useful as a stratification tool — but not as an argument for optimising PRP composition.
What actually stands in the way of personalisation
Analytical coverage
Untargeted metabolomics does not provide a complete picture: on average only 3,000–5,000 features are detected per injection, of which 700–1,000 can be annotated and reliably compared. The same study acknowledges that bioactivity is not driven by metabolites alone but depends substantially on proteins, cytokines and vesicles — which were not measured there.
Time factor
No mass-spectrometric method that is clinically practical today provides a result before a point-of-care injection. Realistically, omics can serve as a characterisation and stratification tool: system qualification, cohort formation and panel development. The likely route into practice is through reduced, validated multiplex panels derived from discovery data.
Reporting quality
Without uniform reporting standards, omics data cannot be pooled. PAW, PLRA, DEPA and MARSPILL are established as classification systems, while MIBO is a reporting standard with 23 mandatory items in 12 categories, developed through Delphi consensus among 24 international experts.
Implementation in the literature
Voluntary adherence is weak: implementation of MIBO showed no improvement when comparing publications before and after the guideline. In systematic reviews, the final platelet concentration is missing from the large majority of studies — precisely the parameter with the strongest dose–response evidence.
What the classifications cover
- PAW / DEPA — cell composition, platelet dose, recovery, purity, activation.
- MARSPILL — extends this to the preparation workflow and procedural variables, including image guidance.
- PLRA — platelets, leukocytes, erythrocytes and activation as a minimum dataset.
- MIBO — not a classification, but a reporting standard: completeness and methodological transparency.
Assessment of the evidence
Supported
Preparation methods can reproducibly generate different protein and metabolite signatures. Different PRP formulations should therefore not be treated analytically as compositionally identical. For knee osteoarthritis, there is evidence that the delivered platelet dose may also be clinically relevant.
Plausible, not validated
It is plausible, but not validated, that an individually suitable formulation could in future be derived from a proteomic or metabolomic signature. The available responder signatures come from small cohorts without prospective external validation.
Unresolved
Whether the identified markers are causal or merely reflect the patient state. Also unresolved are transferability across indications, long-term stability of the signatures and practical point-of-care analytics.
What this means for documentation
Regardless of when omics-assisted personalisation becomes ready for routine practice, the current evidence already suggests requirements for documentation. These include not only cell counts, but also tube system, centrifugation, final volume and activation status. Clean process documentation makes studies and internal treatment courses more comparable.
- Document absolute dose rather than enrichment factor. “5-fold concentrated” is not meaningful without a baseline value and volume. Platelet concentration × volume × number of injections gives the comparable quantity.
- Record the baseline platelet count from the blood count. Without it, system recovery cannot be determined — and recovery differentiates systems more than catalogue wording does.
- Decide leukocyte content according to indication, not as a matter of principle. The evidence does not support a blanket rule in either direction.
- Report centrifugation parameters in RCF, not only rpm. Without rotor radius, an rpm value cannot be transferred between devices. The technical basis is explained in the article RCF, RPM, rotor and time in PRP centrifugation .
- Record activation status and additives. The metabolomics data show that activation method and tube material measurably alter the profile.
- Observe and document the processing time window. Metabolic profiles shift over storage time; freshly prepared material is the referenced state.
↗PRP tubes
Tube systems for professional PRP preparation with information on material, additives and protocol.
↗PRP centrifuges
Centrifuges for clinical practice and laboratory use with technical information on rotor, RCF and compatibility.
↗Tube and centrifuge as a system
Technical guide to compatibility, RCF, RPM, rotor radius and reproducible PRP protocols.