Recently, thrombotic diseases have surged to unprecedented levels, causing a serious threat to human health. In the present study, we utilized enzymatic hydrolysis technology to transform hemp seed processing by‐products, such as hemp seed cake, into peptides with antithrombotic activity. Our findings confirmed hemp seed peptides as potent antithrombotic agents that promote angiogenesis and inhibit intracellular apoptosis.

Antithrombotic activity evaluation techniques can be categorized into in vitro assays, in vivo models, and emerging in silico methods. Among in vitro techniques, thrombin inhibition assays serve as a fundamental tool for preliminary antithrombotic assessment due to their strong controllability and low cost. Nonetheless, these assays are inherently limited by their controlled experimental conditions, which fail to fully recapitulate the physiological complexity of human tissues. In vivo studies are particularly valuable for evaluating therapeutic interventions, as they provide critical insights into the physiological responses and efficacy of peptides within living organisms. While conventional peptide discovery remains limited by complexity, high costs, and labor intensity, animal models offer an indispensable and reliable platform to understand the underlying pathogenesis of thrombosis and assess the efficacy of novel antithrombotic peptides in preclinical studies (He, Du, et al. 2024). Currently, with the rapid advancement of bioinformatics, in silico analysis represents a cost‐effective and efficient approach for target identification of antithrombotic peptides. Some databases, such as ToxinPred, AllerCatPro, and Expasy, have been used as bioinformatics tools to provide structure and physical–chemical properties information about peptides. In addition, as a high‐throughput bioinformatics tool, molecular docking can not only predict the strength of protein‐peptide interactions, but also reveal possible binding mechanisms between ligands and receptors (Li et al. 2025). For instance, Ouyang et al. (2024) successfully utilized molecular docking to screen potential anticoagulant peptides in protein sequences from animals inhabiting deep‐sea hydrothermal vents and cold seeps. Similarly, Yang et al. (2025) identified six anticoagulant peptides from Larimichthys crocea protein using molecular docking in combination with in vitro evaluation. Nevertheless, molecular docking typically treats receptors as rigid, ignoring that both ligands and receptors can undergo conformational changes to achieve optimal binding (Yu et al. 2023). Therefore, the results of molecular docking may occasionally lack accuracy and require further validation. It is noteworthy that the main location of action for polypeptides with antithrombotic activity is still in the blood. Hence, the majority of these bioactive peptides require blood entry to function (Liu et al. 2023). An antithrombotic peptide may demonstrate promising bioactivity in vitro, but this does not guarantee similar effects will be observed in vivo due to the complexity and variability of biological systems (Toldra et al. 2020). Moreover, while these studies demonstrated strong in vitro thrombin‐inhibitory activity, they are limited by the lack of in vivo validation of peptide biological effects. Furthermore, industrial‐scale bioactive peptide preparation often faces challenges in scaling up and costly purification, which hinders commercialization (Cheng et al. 2024). In the present study, an integrated in vitro to in vivo approach provides a more comprehensive evaluation of enzymatic hydrolysate bioactivity.

The < 3 kDa fraction demonstrated remarkable thrombin inhibition (99.38% at 0.01 g/mL) and a low IC₅₀ (4.55 mg/mL), outperforming rapeseed peptides of IC₅₀ = 30 mg/mL (Zhang et al. 2008), scorpion‐derived peptides of IC₅₀ = 23.65 mg/mL (Zhang et al. 2023) and egg peptides of IC₅₀ > 50 mg/mL (Yang et al. 2007). This superiority is likely attributed to the unique peptide profile of hemp hydrolysates, where approximately 67% of identified sequences (0.5–1.5 kDa) fall within the optimal molecular weight range for antithrombotic bioactive peptide (Cheng et al. 2019). The complex composition of crude enzymatic hydrolysates may exert synergistic antithrombotic effects through multiple‐targeted mechanisms, such as inhibiting platelet aggregation, promoting fibrinolysis, and enhancing production of endothelial nitric oxide. These characteristics not only enhance therapeutic potential but also provide obvious advantages for reducing the difficulty of separation in industrial‐scale production. Moreover, using DOCK 6, we screened 10 representative hemp seed peptides, though their practical role and functional mechanisms require full elucidation. Future studies on their physicochemical and structural properties, along with the establishment of structure–activity relationships, will facilitate the rational design of novel targeted antithrombotic peptides.

The other primary purpose of this experiment was to investigate the antithrombotic activity of hemp seed peptide in vivo and its potential mechanism. Researchers have developed animal models like rats and mice, which are often costly, time‐consuming, and yield low success rates. Zebrafish share 87% genomic homology with humans. Hence, zebrafish have been well‐established as a vertebrate model for hematopoiesis and hemostasis research, widely used in high‐throughput drug screening (Sato et al. 2013). Furthermore, research has identified several key genes in zebrafish that significantly influence human thrombosis, demonstrating a notable degree of homology. Thrombosis models are commonly induced via microinjection or direct immersion using phenylhydrazine (PHZ), ferric chloride (FeCl₃), arachidonic acid (AA), or ponatinib. PHZ, a strong oxidant, primarily induces thrombosis by causing oxidative damage to red blood cells, making this model particularly suitable for studying thrombosis associated with hemolytic anemia (Sato et al. 2013). Exogenous administration of AA triggers thrombosis via platelet aggregation in zebrafish (Luo et al. 2025), indicating its appropriateness for evaluating antiplatelet agents. Thrombosis involves tightly regulated processes including vasoconstriction, platelet activation, and blood coagulation, with endothelial cells playing a crucial role in the vascular response to injury leading to thrombus formation (Bochenek and Schafer 2019). Although FeCl₃ induces thrombosis through direct chemical oxidation and damage to the vascular endothelium, this model is predominantly applied in murine studies (Li et al. 2016) and has not been widely adopted for zebrafish. In contrast, ponatinib, a tyrosine kinase inhibitor initially identified for its thrombotic side effects in leukemia treatment, causes endothelial dysfunction in zebrafish. It inhibits vascular endothelial proliferation, migration, and angiogenesis by blocking vascular endothelial growth factor receptor (VEGFR) signaling, ultimately leading to thrombosis (Hamadi et al. 2019). The ponatinib‐induced zebrafish thrombosis model is now utilized for in vivo thrombosis research and rapid screening/evaluation of antithrombotic drugs. For instance, He, Wu, et al. (2024) demonstrated that 1.5 μg/mL ponatinib treatment reduces blood flow in zebrafish larvae, inducing thrombosis, whereas Ye et al. (2020) employed this model to identify three dammarane‐type triterpenoids from Rhus verniciflua root with antithrombotic activity. An additional advantage of the ponatinib model is its pathological relevance, especially for mimicking thrombosis complications in leukemia patients undergoing therapy. We induced thrombosis in a zebrafish model using ponatinib, and our findings were consistent with previous results (Zhang et al. 2022). Ponatinib significantly reduced cardiac staining intensity by 43%, blood flow velocity by 177%, and vessel diameter by 21%. Nevertheless, the hemp seed peptide improved vascular endothelial injury, increasing the staining intensity of erythrocytes in the cardiac, enhancing the blood flow velocity, and expanding the diameter of intersegmental blood vessels. Notably, at 125 μg/mL, hemp peptides restored vascular function to near‐normal levels without inducing toxicity, suggesting favorable therapeutic prospects compared to heparin's bleeding risks (Li et al. 2020). In our study, the thrombosis model of zebrafish was used for the first time to verify the antithrombotic activity of hemp seed peptide, and the results showed that hemp seed peptide had significant antithrombotic activity. However, zebrafish models cannot fully replicate the complexity and variability of thrombosis in human patients (Griffin et al. 2024). Therefore, the specific clinical antithrombotic effects of hemp seed peptides require validation across multiple thrombosis models in the future. Data covering diverse thrombotic mechanisms would enhance the accuracy and robustness of our experimental findings.

Mechanistically, the observed antithrombotic effects align with the dual modulation of vegfr1 and caspase‐3. Dose‐dependent improvements in hemodynamic parameters align with vegfr1‐mediated endothelial proliferation, analogous to Rubia cordifolia's pro‐angiogenic effects via VEGF‐A/kdrl upregulation (Chen et al. 2018). Vegfr1 activation promotes endothelial cell proliferation and vascular repair (Griffin et al. 2024), while caspase‐3 suppression mitigates apoptosis‐driven platelet aggregation (Raghunath et al. 2022). This dual‐action mechanism contrasts with single‐target approaches, such as direct thrombin inhibitors, potentially offering broader clinical applicability (Schwienhorst 2006). Intriguingly, lower peptide concentrations (31.25 μg/mL) maximally suppressed caspase‐3, possibly due to enhanced cellular uptake efficiency, whereas higher doses (125 μg/mL) optimized hemodynamic outcomes. This dose‐dependent divergence underscores the need for pharmacokinetic studies to balance bioavailability and bioactivity.

As is well known, thrombin drives platelet aggregation by activating platelets via PARs and GPIbα, enabling fibrinogen binding (De Candia 2012). It further promotes coagulation by activating Factor XIII (FXIIIa), which cross‐links fibrin to form a stable clot resistant to plasmin degradation (Souri et al. 2015; Zabczyk et al. 2021). Additionally, thrombin activates PAR1 on endothelial cells, reducing NO production and thereby facilitating platelet aggregation and thrombus formation (Li et al. 2022). Therefore, anticoagulant peptides may simultaneously possess platelet aggregation inhibitory activity, fibrinolytic properties, and the ability to promote NO secretion in endothelial cells. Among the 10 candidate peptides analyzed, two peptides—GARFDER and RASQEQLKA—demonstrate significant sequence similarity to known platelet aggregation inhibitors, suggesting potential shared inhibitory mechanisms. GARFDER contains a characteristic “FD” motif and exhibits a high density of acidic residues, closely resembling the oat‐derived antiplatelet peptide GEEFDAFTPK (Yu et al. 2016). Similarly, RASQEQLKA shares a conserved “QL” receptor‐binding motif and mixed charge profile with two established inhibitors, QLAQIPR and the scolopendra‐derived SQL peptide (Kong et al. 2013). These structural similarities indicate that GARFDER and RASQEQLKA may possess dual antithrombotic activities, combining thrombin inhibition with platelet aggregation suppression. Additional analysis identified two peptides—GRHSSENYR and RADVIVVPAG—that exhibit structural features associated with fibrinolytic activity. GRHSSENYR displays a hydrophobic‐polar “HSS” motif resembling the “TVAA”/“LLSY” sequences found in fibrinolytic peptides from Cordyceps militaris culture supernatant, along with a C‐terminal tyrosine residue that may enhance fibrin binding (Liu et al. 2015). RADVIVVPAG features a hydrophobic “IVV” core similar to the “VLL” region in known fibrinolytic peptides. These structural characteristics suggest potential roles in promoting fibrinolysis. Future research should focus on characterizing specific effects of these peptides on platelet aggregation pathways (particularly through PAR and GPIbα interactions), their fibrinolytic potential (via FXIIIa modulation), and their influence on endothelial nitric oxide production (through eNOS regulation). Such investigations would provide crucial insights into their therapeutic potential for antithrombotic applications, particularly in conditions requiring multi‐target intervention strategies. The combination of thrombin inhibition, antiplatelet activity, and fibrinolytic enhancement could offer novel approaches for developing more effective and safer antithrombotic therapies.