The present study is a continuation of our previous work wherein we showed the antithrombotic activity of Potentilla erecta rhizome extract in normoglycemic rats and mice (Marcinczyk et al., 2017). The genus Potentilla L (Rosaceae family) consists of approximately 700 species of annual and biennial plants and small shrubs. Extract from the underground parts of Potentilla erecta (L.) Raeusch (tormentil extract [TE]) has been used in traditional ethnomedicine for the treatment of diarrhea and mild inflammation of the oral mucosa because of its antioxidant and anti-inflammatory activity (Tomczyk and Latté, 2009; Melzig and Böttger, 2020; Augustynowicz et al., 2021). The pharmacological properties of tormentil have been discussed in several reviews and it’s monographs are included in the Russian Pharmacopoeia (14th edition), State Pharmacopoeia of the Republic of Belarus, ESCOP, in the British Pharmacopoeia, or the European Pharmacopoeia (10th edition) (Shikov et al., 2021). These actions are induced by ellagitannins and procyanidins, which are rarely found together in a plant material, although the presence of ellagitannins and procyanidins in TE was confirmed by us and other researchers (Geiger et al., 1994; Mari et al., 2013; Marcinczyk et al., 2017). Ellagitannins and procyanidins are characterized by poor bioavailability after oral intake. However, their gut microbiota-synthesized metabolites have good bioavailability and are responsible for in vivo activity of TE. Urolithins are metabolites of ellagitannins (Piwowarski et al., 2017), while phenolic acids as well as dimeric and trimeric procyanidins are metabolites of high-molecular-weight procyanidins (Sano et al., 2003; Sánchez-Patán et al., 2012). It has been shown that procyanidins and ellagitannins exert multidirectional effects on hemostasis; however, most of these studies have been conducted in vitro, which makes it difficult to predict the activity of this class of compounds after oral intake. The effect of ellagitannins on platelet activity, coagulation, and fibrinolysis was mainly studied in vitro by using plasma (Dong et al., 1998), washed platelets (Shen et al., 2003), and chromogenic substrates (Zhang et al., 2004; Yamamoto et al., 2011). The effects of ellagitannin-rich products other than TE on hemostasis has also been studied in humans after oral intake. The consumption of pomegranate juice has been shown to decrease collagen-induced platelet aggregation (Haber et al., 2011) while blended frozen red raspberries were found to improve flow-mediated vasodilation of the brachial artery (Istas et al., 2018). The effects of procyanidins on hemostasis reported thus far include mainly increase in nitric oxide (NO) and prostacyclin (PGI₂) production (in vitro and in vivo) (Schramm et al., 2001; Byun, 2012; Novakovic et al., 2017) and decrease in platelet activity (in vitro and in vivo) (Murphy et al., 2003; Bouaziz et al., 2007; Morel et al., 2014). Furthermore, by using stasis-induced model of venous thrombosis in rats (Jiang et al., 2007) and laser-induced model of arterial thrombosis in mice, the antithrombotic effect of procyanidins after oral intake was also demonstrated (Sano et al., 2005).

In our previous study, we showed for the first time the antithrombotic activity of ellagitannin- and procyanidin-rich TE in an animal model. In our study, TE inhibited arterial thrombosis in normoglycemic rats through a mechanism dependent on platelet inhibition. The underlying mechanism of this antiplatelet effect was based on the inhibition of thromboxane (TXA₂) production in platelets, which was comparable to the effect of acetylsalicylic acid. We also showed that TE slightly enhanced fibrin formation and attenuated fibrinolysis, but these effects did not abolish the antithrombotic effect of TE (Marcinczyk et al., 2017).

Considering the multidirectional activity of TE in normoglycemic rats, the main goal of the present study was to assess whether TE exerts the antithrombotic effect in streptozotocin (STZ)-induced diabetes, i.e., in conditions with intrinsically increased fibrin formation and impaired fibrinolysis. Hyperglycemia in type 1 diabetes is a risk factor for cardiovascular diseases and thromboembolic complications (Lee et al., 2019). Enhanced coagulation, impaired fibrinolysis, and increased platelet activity expressed as an enhanced production of TXA₂ and P-selectin secretion are often observed in patients with diabetes (Wisinski and Kimple, 2016; Sobczak and Stewart, 2019). Diabetes also impairs endothelial functions and vessel wall contractility and leads to a reduction in blood flow (Lunder et al., 2018), which may affect the dynamics of thrombus formation. Furthermore, vascular complications of diabetes, such as macro- and microangiopathy, contribute to hemostasis activation (Tarnow et al., 2009; Zahran et al., 2018). Because of multidirectional hemostasis dysfunctions, the evaluation of the activity of TE, which acts on multiple targets, in diabetes seems to be justified.

In our present study, we used a model of STZ-induced diabetes (STZ-diabetes), which clinically reflects type 1 diabetes (Furman, 2015), and conducted experiments in two models of thrombosis (electrically induced and laser-induced thrombosis) that differ in the mechanism of thrombus formation. Electrically induced arterial thrombosis allows to assess thrombus weight and blood flow disturbances due to vessel-occluding thrombus formation and growth after extensive endothelium damage. Dynamics and the extent of the thrombotic process assessed in this model involve both hemodynamic conditions and hemostatic activity. Thrombus formed in this model is mainly composed of platelets and fibrin, with a minor amount of red and white blood cells (Zakrzeska et al., 2015). The model of laser-induced thrombosis combined with confocal imaging was used for the intravital observation of the thrombotic process at the site of mesenteric vein injury, where thrombus formation occurs on the exposed subendothelial matrix. Because the thrombus is composed of platelets, this model allows to assess the thrombus area and platelet activation (Marcinczyk et al., 2020). Furthermore, by using a wide range of advanced techniques, we determined the effect of TE on the components of hemostasis: platelets, coagulation system, fibrinolysis, and endothelial-dependent vascular functions. This enabled more detailed and multidimensional study of the effect of TE on the thrombotic process and hemostasis in STZ-diabetes.

In the present study, we observed the multidirectional effects of TE on hemostasis in STZ-diabetic rats and mice, which were not associated with the blood glucose level. An increase in the dynamics and extent of thrombus formation in STZ-diabetes was demonstrated in the models of electrically induced and laser-induced thrombosis. Thrombosis was enhanced due to increased platelet (Figures 4–6) and coagulation activity (Figures 7, 8), impaired fibrinolysis (Figures 9C,D), endothelial dysfunction (Figure 1B), reduced blood flow (Figure 2B), and inflammatory state (Figures 1C,D). In vivo experiments were performed in both rats and mice to make the overall effect of TE species-independent. We observed that TE in diabetic rats, unlike in normoglycemic rats (Marcinczyk et al., 2017), does not exert antithrombotic effect in the rat carotid artery. Moreover, it intensified electrically induced thrombotic process at 200 mg/kg dose, which was manifested as increased thrombus weight (Figure 2D). Considering the slight increase in fibrin formation in normoglycemia, we presumed that the increase in thrombus weight in STZ-diabetes could be due to enhanced coagulation. This hypothesis was confirmed in the experiment where the most pronounced fibrin net density was observed at TE dose of 200 mg/kg in clots formed in both PPP and PRP. TE at all the tested doses increased fibrin net density in PRP, while it increased the fibrin net density in PPP only at 100 and 200 mg/kg doses. This observation indicated an increase in platelet activity at 400 mg/kg dose as compared to that noted at 100 mg/kg dose and confirmed the key role of platelets in the process of fibrin formation. Increased fibrin net density was the reason for further studies on the effect of TE on coagulation. Because the TF/VIIa complex triggers coagulation in vivo (Smith, 2009), the next stage of the study was to determine the expression of TF in a mouse mesenteric artery. TE at all the tested doses decreased the expression of TF, and this effect was the strongest at 200 mg/kg dose (Figure 8). Because TE at 200 mg/kg dose caused the most dense fibrin net formation, the contribution of TF in the enhancement of coagulation was ruled out. As platelets play an essential role in the process of fibrin formation, the next stage of the study was to evaluate the influence of TE on platelet procoagulant activity. For this purpose, we used a flow chamber combined with a confocal microscopic imaging system. This model enables real-time monitoring of the process of thrombus formation on collagen fibers (ex vivo). Platelets in thrombus can be divided into two subpopulations: procoagulant and aggregating platelets. During sustained and potent activation, platelets change their discoidal shape to irregular shape, and the platelet cell membrane undergoes irreversible reorganization, which causes exposure of PS from the inner to the outer leaflet of the platelet plasma membrane. Because PS catalyzes the activation of coagulation factors, platelets with exposed PS are called procoagulant platelets. Aggregating platelets do not show exposure of PS (PS-negative platelets). However, because they are close to activating factors, they form pseudopods and undergo secretion (Heemskerk et al., 2013). As PS catalyzes coagulation reactions, the ratio of the area of PS-positive platelets to the area of PS-negative platelets is termed as PI and reflects the extent of platelet procoagulant response. PI was reduced at all tested doses of TE; the strongest antiprocoagulant effect was observed at 100 mg/kg dose, while the weakest reduction in PI was observed at 400 mg/kg dose (Figure 4). Therefore, it can be assumed that the weakened antiprocoagulant effect contributed to the increased fibrin net density in clots formed in PRP at 400 mg/kg dose compared to that in the Diabetes group. However, the decreased fibrin net density at 400 mg/kg dose as compared to that at 200 mg/kg dose with and without platelets (PRP and PPP) indicates an unknown plasma-related mechanism of coagulation activation. Further experiments on the influence of TE on platelet activity were performed in a model of laser-induced thrombosis. In this model, because of the small area of the exposed subendothelial matrix, potent and sustained platelet activation does not occur, which results in thrombus composed of platelets that do not undergo irreversible activation (aggregating, PS-negative platelets) and can easily detach from the site of injury. Therefore, this model is suitable to investigate platelet activity in the aggregation state. TE only at the lowest dose (100 mg/kg) decreased the thrombus area in the model of laser-induced thrombosis (Figure 5A); this finding is consistent with the antiplatelet activity observed in the flow chamber model. The previously observed antithrombotic effect in normoglycemic rats and mice was noted only at the highest dose (400 mg/kg) and corresponded to a potent antiplatelet activity (Marcinczyk et al., 2017). Next, the assessment of the activity of aggregating platelets in thrombus was measured by the PECAM-1/thrombus ratio, which indicates the proportion of platelet-endothelial cell adhesion molecule 1 (PECAM-1) in thrombus. As PECAM-1 is considered as an antithrombotic molecule, the higher the PECAM-1/thrombus ratio, the less activated are the platelets in thrombus (Marcinczyk et al., 2020). TE increased the PECAM-1/thrombus ratio with the most pronounced effect at 100 mg/kg dose (Figure 5B). The antiplatelet effect of TE at 100 mg/kg dose was significant enough to translate into the antithrombotic effect. These observations indicate that TE at 100 mg/kg dose most effectively reduced platelet activity in two modes of activation: reversible, which was expressed as increased PECAM-1/thrombus ratio, and irreversible, which was expressed as decreased PI. The sections of arterial thrombi showed an increase in entrapped erythrocytes at 400 mg/kg dose (Figure 3). This indicates hemolysis, which leads to the enhanced incorporation of erythrocytes into thrombus (Helms et al., 2013). Furthermore, the pale red color of plasma of the 400 mg/kg group (data not shown) indicated the occurrence of hemolysis. During hemolysis, erythrocytes release platelet agonists, (e.g., ADP), which also contributes to enhanced platelet activity (Noh et al., 2010) and could partially explain the alleviation of the antiplatelet effect of TE at 400 mg/kg dose. The next stage of the study was the assessment of P-selectin secretion at the site of laser-induced mesenteric vein injury under intravital conditions. P-selectin, a marker of platelet secretion, is responsible for thrombus progression and its stability, which allows direct interactions of platelets with endothelial cells and leukocytes (Prakash et al., 2017). Inhibition of P-selectin secretion was observed only at 100 and 400 mg/kg doses; this finding may suggest the decreased stability of thrombus (Figure 6). The curve of thrombus formation kinetics in the 100 mg/kg- and 400 mg/kg-treated groups is characterized with a high peak and the slope of the curve, which correspond to the addition of the thrombotic material shortly after vessel injury and subsequent elution of the thrombus. The curve of thrombus kinetics in the 200 mg/kg-treated group shows two peaks that correspond to two massive additions of the thrombotic material; this finding suggests increased thrombus stability as compared to that noted in the 100 and 400 mg/kg groups (Figure 5C). We observed that TE at 200 and 400 mg/kg doses increased the IBF in the rat carotid artery and prolonged TTO (Figures 2B,C); this finding indicated the contribution of the vessel wall in changes of thrombus formation kinetics. Furthermore, TE at 400 mg/kg dose also prolonged BT (Figure 1A), which reflects the dependence of primary hemostasis on the blood vessel response (vasoconstriction/vasodilatation) and on the activity of platelets that form a platelet plug at the site of injury (Mattix and Singh, 1999). Because of the lack of the antiplatelet effect of TE at 400 mg/kg dose in this model, the prolongation of BT indicated an improvement in vascular function. Furthermore, H and E staining revealed that TE increased the thickness of the middle layer of the artery wall (Figure 3), which may be an indirect effect associated with an increased blood flow (Basu et al., 2010). Some of the compounds of TE and their metabolites have potential to improve vascular functions, which may be reflected as increased blood flow observed in Figure 2B. Metabolites generated from ellagic acid and ellagitannins by the intestinal microbiota such as urolithin A and B have been shown to enhance eNOS expression (Han et al., 2016; Spigoni et al., 2016), and the activity of procyanidin B leads to vasorelaxation of the human mammary artery (Novakovic et al., 2017). This indicated that the observed improvement in blood flow might be partially due to an endothelial-dependent mechanism. Therefore, in the next stage of the study, the concentrations of NO₂ ⁻ and NO₃ ⁻, the stable metabolites of NO – the main vasoactive molecule released from the endothelium (Mitchell et al., 2008) – were measured. TE at 200 mg/kg dose increased the concentrations of NO₂ ⁻ and NO₃ ⁻, which may contribute to IBF improvement and TTO prolongation (Figures 2B,C, respectively). Previous studies on endothelial NO release have shown that the increase in eNOS activity is linked to the decrease in TF expression (Solovey et al., 2010). Therefore, the largest decrease in TF fluorescence at 200 mg/kg dose may be due to the increased production of NO at this dose. However, the reason why NO₂ ⁻ and NO₃ ⁻ concentrations increased only at 200 mg/kg dose and not at 400 mg/kg dose remains unclear.

The next component of the prothrombotic effect of TE at 200 mg/kg dose was the attenuation of fibrinolysis, as the fibrinolytic activity of plasma measured as ECLT was most strongly inhibited at this dose (Figure 9A). However, TE at 200 mg/kg dose increased the concentration of the active form of t-PA (Figure 9C), which might be due to increase NO release (Giannarelli et al., 2007). Furthermore, compared to 200 mg/kg dose, TE at 400 mg/kg dose decreased the concentration of active form of t-PA, which may result from decreased concentration of bradykinin – one of the activators of t-PA release (Marcinczyk et al., 2017). The antifibrinolytic effect of TE could also be, to some extent, explained by an increase in the concentration of the active form of PAI-1 (Figure 9D). This may be due to the increased activation of platelets, which release PAI-1 during the activation process (Morrow et al., 2020). Furthermore, the tendency of increasing concentration of active form of PAI-1 may be due to the decreased concentration of t-PA that binds to PAI-1, thus diminishing the measurable pool of free PAI-1 (Chandler et al., 1995).

Considering the anti-inflammatory activity of TE (Tunón et al., 1995; Wölfle et al., 2017) and the prothrombotic effect of inflammation (Esmon, 2003), the concentration of TNF-α, a commonly accepted inflammation marker, was measured. TE did not change the level of TNF-α in rat plasma (Figure 1D). However, no effect on TNF-α does not imply a lack of anti-inflammatory properties of TE in diabetes. The less number of leukocytes in thrombi from the 400 mg/kg-treated group than that in thrombi from the Diabetes group may indicate the anti-inflammatory activity of TE (Figure 3). We have also shown that TE at 400 mg/kg dose lowered the concentration of 6-keto PGF_1α, thus indicating the inhibition of PGI₂ production by endothelial cells in the presence of high dose of TE (Figure 1C). The effect of decreased PGI₂ production may be due to the known inhibitory effect of TE or its agrimoniin-rich fraction on COX activity (Tunón et al., 1995; Hoffmann et al., 2016). Furthermore, the results of our study are consistent with those which show that the anti-inflammatory effect of TE is dependent on the inhibition of arachidonic acid (AA) metabolism. However, in those studies, the anti-inflammatory effect was demonstrated in a human keratinocyte cell line (HaCaT, in vitro) and in human skin (in vivo, applied topically) (Hoffmann et al., 2016). Thus, our study for the first time showed that after oral intake, the anti-inflammatory effect of TE could be dependent on the inhibition of AA metabolism.

In our present study, the bioavailability and plasma concentrations of specific compounds originating from TE and their metabolites were not assessed; thus, we cannot confirm which compounds or metabolites are primarily responsible for the changes observed in hemostasis. Furthermore, it is possible that the inflammatory state changed the bioavailability of specific TE components. Urolithins, which are the gut microbiota-derived metabolites of ellagitannins, undergo glucuronidation rapidly after absorption. Glucuronides of urolithins are inactive, but under inflammatory conditions, they are deconjugated by ß-glucuronidase to free urolithins (Piwowarski et al., 2017). Therefore, it can be presumed that the amount of free urolithins after the administration of equivalent doses of TE is different in normoglycemia and STZ-diabetes. Furthermore, the differences in gut microbiota in normoglycemia and diabetes (Ma et al., 2020) could lead to the formation of different metabolites of TE. In our present study, we showed for the first time that the activity of TE is dependent on the pathological condition. It is also possible that TE activity in STZ-diabetes was associated with the production of inflammatory-related TE metabolites. To summarize, we have shown that TE exerts multidirectional effects on the activity of platelets, coagulation, fibrinolysis, and endothelial-dependent vascular functions in STZ-diabetic rats and mice. These multidirectional effects of TE on hemostasis translated into a model-dependent effect on the thrombotic process. Furthermore, the contrasting effects of TE in normoglycemia (Marcinczyk et al., 2017) and diabetes indicates that the mechanism of action of TE is related to the pathological-dependent baseline hemostatic activity.

Since the TE has been used to treat diarrhea (Tomczyk and Latté, 2009), our study may have clinical relevance in hemostasis regulation. However, considering the activation of coagulation and the inhibition of fibrinolysis by TE, the beneficial impact of ellagitannins (Haber et al., 2011; Istas et al., 2018) on hemostasis after short-term use is not so obvious. Our study’s TE’s doses were selected based on studies with rats and mice (Shushunov et al., 2009). Pharmacokinetics of TE in humans have not been investigated yet, but TE’s doses are lower (approximately 2.5–40 mg/kg) (Subbotina et al., 2003; Huber et al., 2007). Due to the fast metabolism rate, mice and rats often require higher doses of the drug (e.g., acetylsalicylic acid) (Wientjes and Levy, 1988; Nagelschmitz et al., 2014). Nevertheless, we cannot state whether the same pharmacological effects can be expected after using the same doses of TE in humans as in rodents. Raising these issues, further studies are needed to assess the potential clinical significance of our study particularly in patients with increased risk of thromboembolic events.