Circulating platelet levels decrease about 10-fold in multiple hibernating species during torpor, with a return to normal levels as the animals rewarm during arousals (Figure 4). These dynamics could decrease the risk of a large intravascular clot forming during torpor, while restoring normal clotting activity upon arousal. The thrombocytopenia is temperature dependent and can be induced in hamsters, rats, and mice with forced chilling (de Vrij et al., 2014). Cold acclimation of rats and hamsters also reduced circulating platelet levels even though the animal’s body temperatures did not decrease (Deveci et al., 2001). The rapid return of platelets in circulation after arousal, in the absence of large numbers of circulating immature platelets, suggests that hibernator platelets are stored during the cold and released into circulation during rewarming. Removal of the spleens of Syrian hamsters and 13-lined ground squirrels before or during torpor did not affect the platelet storage and release indicating that spleen is not required for storage as originally proposed (Reddick et al., 1973; Cooper et al., 2012; Cooper et al., 2016; de Vrij et al., 2021). Rather, platelet storage during torpor occurs in the sinusoids of the liver, consistent with this organ being responsible for storing platelets (Cooper et al., 2017; de Vrij et al., 2021). The mechanism of hepatic storage of platelets during torpor is currently unknown, potentially platelet margination may underlie the reversible storage and release in liver (de Vrij et al., 2021). Although platelets cluster together in liver sinusoids during torpor, they do not form irreversible thrombi. Several adaptations may prevent platelet activation during their slumber in cold liver sinusoids. Platelet GpIbα receptors can bind to VWF, but the plasma concentration of VWF decreases during torpor in bears, ground squirrels and hamsters, with a selective loss of the thrombogenic high molecular weight multimers (Cooper et al., 2016; Friedrich et al., 2017). Additionally, platelets from torpid ground squirrels bind less VWF and collagen, consistent with decreased GpIbα activity on their surface. Upon arousal, VWF levels return to normal rapidly, possibly by release of stores in platelets or endothelial cells (Cooper et al., 2016). The combination of decreased platelet and VWF levels along with decreased binding of platelets to VWF would all help to suppress activation of primary hemostasis during torpor-induced stasis.

Shape change is another prominent feature of cooled and hibernating platelets. Under normal conditions, platelets have a circumferential microtubule ring that keeps them in a disc conformation. In non-hibernating mammals, chilling platelets leads to depolymerization of their microtubules and the platelets become spherical and do not repolymerize upon warming (White and Krivit, 1967; Winokur and Hartwig, 1995). In contrast, ground squirrels and Syrian hamster platelet microtubules form rods when chilled and can reconstitute the circumferential ring upon rewarming (Cooper et al., 2017; de Vrij et al., 2021). This allows ground squirrel platelets to retain their overall shape after chilling. It is unclear if this rod shape has any physiological or adaptive significance, but it may protect platelets from damage caused by cold storage.

Cold storage of human and mouse platelets damages platelets in several ways, collectively termed cold storage lesions. Sialic acid residues on surface glycoproteins are cleaved leading to the rapid clearance of chilled platelets from circulation by the liver (Hoffmeister et al., 2003; Hoffmeister, 2011). Cold storage also induces intrinsic mitochondrial activated apoptosis in human and mouse platelets (Van Der Wal et al., 2010; Lebois and Josefsson, 2016; Stolla et al., 2020). In contrast, ground squirrel platelets stored in vitro in the cold are not cleared rapidly upon re-transfusion (Cooper et al., 2012). Ground squirrel platelets stored in the cold show decreased desialylation of surface glycoproteins and decreased phagocytosis by hepatocytes. They also appear to be resistant to induction of apoptosis as measured by caspase activation and surface phosphatidylserine on platelet membranes (Splinter et al., 2023). These adaptations of hibernator platelets to cold storage are consistent with their storage and release after weeks of cold storage during torpor and likely add to the rapid restoration of a hibernator’s ability to form clots upon arousal.

Platelets are anucleate cells and their protein content is determined primarily by progenitor megakaryocytes in the bone marrow. Surprisingly, a bone marrow transcriptome of samples collected in the fall, winter, and summer from 13-lined ground squirrels revealed no significant differences in any platelet protein transcripts, including P-selectin, membrane receptors, and integrins (Cooper et al., 2016). The platelet proteome collected in the same species and at the same times of the year, also displayed no differences in integrins and signaling proteins, but did reveal significant seasonal differences in other proteins (Cooper et al., 2021). In the summer, platelets have increased heat shock proteins which could help with proper protein folding at warm temperatures. In the fall, as animals are going through short bouts of shallow torpor, pro-inflammatory and clotting proteins become less abundant in platelets. During torpor, more plasma-derived proteins such as albumin and lipoproteins are present in platelets which are known to inhibit apoptosis and thrombosis (Cooper et al., 2021). A very recent platelet proteome study on brown bears showed a 55-fold decrease in the heat shock protein HSP47 (SerpinH1) during hibernation (Thienel et al., 2023). This decrease correlated with a decline in DVT formation during prolonged immobilization, and was also demonstrated in humans, pigs, and knock-out mice. Interestingly, and in sharp contrast, a previous study of 13-lined ground squirrel platelet proteome documented unchanged HSP47 abundance across various phases of hibernation (Cooper et al., 2021). This marked difference in Hsp47 regulation could be due to bears being shallow hibernators and also of greater body mass compared to ground squirrel, so the immobilization puts bears at greater risk of DVT. Moreover, torpid bears have a modest drop in circulating platelets compared to deep hibernators (Figure 4), possibly warranting additional protection against DVT. It would be interesting to look more closely at the inhibition of HSP47 in deep hibernators other than 13 lined ground squirrel to appreciate its role in hibernation. In addition to the decrease in circulating platelet numbers, these proteomic changes could prevent unwanted activation of platelets during stasis or during their storage in liver. While there were no significant differences in the amounts of surface receptors or proteins in signaling pathways at the proteome level, it is possible that the activities of these pathways are altered. Alternatively, some of the proteins may be sequestered in the cytoplasm or secretory granules which could be resolved by immunohistochemical analysis. Studying the phosphoproteome and metabolome of platelets in torpor would help to resolve which pathways are affected.

In recent years, cultured cells from hibernators have demonstrated to hold specific adaptations that confer protection from cooling in a cell-autonomous way (Talaei et al., 2011; Ou et al., 2018; Hendriks et al., 2020). Mechanistically, these studies in hamster cell lines and ground squirrel pluripotent stem cells derived neurons indicate that adaptations in mitochondrial function, allowing cooled cells to maintain ATP production and lower reactive oxygen species production, constitute a key element of cell protection. Given that platelets contain mitochondria, we re-analyzed aforementioned ground squirrel platelet proteome data (Cooper et al., 2021) by separating out mitochondrial proteins. Analysis of the platelet proteins involved in metabolism revealed an increase in 79 proteins with clusters in lipoprotein metabolism, cholesterol esterification, and lipase activity during the winter. With respect to mitochondrial proteins, 24 were increased during the winter relative to summer and fall, with clusters in fatty acid beta-oxidation, the electron transport chain, and protein targeting to the mitochondria (Figure 5). These proteome changes correspond well with changes in most other tissues of hibernators, principally denoting the shift to lipid metabolism during hibernation as animals rely on fat stores for energy. As platelets appear to be no exception, it is conceivable that they hold similar cell autonomous protection from cooling damage as found in hamster and ground squirrel smooth muscle cells, kidney cells and neurons. Interestingly, ground squirrel neurons maintain their tubulin network during cooling by limitation of the production of reactive oxygen species (Ou et al., 2018). Such mechanism may also support tubulin stability and possibly underlie the difference in cooling induced platelet shape change between hibernators (reversible rod formation) and non-hibernators (irreversible sphere formation). Given this clearcut difference in response to cooling and the ease of obtaining primary cells, platelets constitute an important platform for the future study of cell-autonomous adaptations in hibernators.

Secondary hemostasis occurs simultaneously with primary hemostasis forming a fibrin network to trap red and white blood cells and further strengthen the hemostatic platelet plug (Figure 6). Secondary hemostasis occurs via a plasma coagulation cascade divided into the intrinsic and extrinsic pathways. Most clotting factors are serine proteases that are always present in plasma as inactive zymogens (identified by Roman numerals II, VII, IX, X, XI, and XII) and are activated by proteolysis (addition of an “a”). Factors V and VIII are non-enzymatic cofactors also activated by proteolysis. The extrinsic pathway starts with exposure and binding of tissue factor (TF) to plasma coagulation factor VII, which forms a TF/VIIa complex. The TF/VIIa complex proteolytically activates factors IX and X, commencing the common pathway, creating a prothrombinase complex with Va that converts prothrombin (factor II) into thrombin (IIa) (Versteeg et al., 2013). Thrombin slowly accumulates during the amplification phase, activating platelets and platelet derived factor V, amplifying the prothrombinase activity. In a positive feedback loop, thrombin activates factor XI and VIII, generating more factor Xa. Hemophilia A and B are defined by deficiencies in factors VIII and IX, respectively, which is associated with spontaneous and prolonged bleeding. The extrinsic pathway can be assayed in vitro by measuring the prothrombin time (PT). The intrinsic pathway of coagulation can be triggered independently by collagen, polyphosphates secreted by platelets, neutrophil extracellular traps (NETs), and artificial material such as glass, leading to activation of factors XII, XI, kallikrein and the subsequent downstream coagulation factors (Muller et al., 2009; Versteeg et al., 2013). The intrinsic pathway can be assessed in vitro by measuring the activated partial thromboplastin time (APTT).

Secondary hemostasis is reduced during torpor in ground squirrels, bears, hedgehogs, and hamsters. In general, hibernating animals in torpor reduce the level of coagulation factors VIII, IX and XI resembling mild hemophilia A, B and C respectively (Denyes and Carter, 1961; Biorck et al., 1962; Lechler and Penick, 1963; Cooper et al., 2016; Welinder et al., 2016). The decreases in factor VIII, IX and XI during torpor in examined hibernating species are on average 78%, 61% and 51%, respectively, compared to euthermic non-hibernating values (Figure 7). Interestingly, the extrinsic pathway of the coagulation cascade remains fairly unaltered during hibernation. A logical explanation could be that thrombosis due to stasis of blood should be prevented, but blood clotting due to tissue damage by an intruder or predator (activating the extrinsic pathway) may still be vital. An alternative explanation is that during torpor, damage to blood vessels is unlikely, so suppressing the extrinsic pathway is not necessary and protection against stasis-induced clots is more important. Remarkably, one clotting factor consistently showed an increase in protein levels during torpor: prothrombin. Possibly, this is caused by a decrease in its normal baseline formation resulting from inhibition of spontaneous, low-level activation of the clotting cascade, thus diminishing the conversion of prothrombin into thrombin (Lechler and Penick, 1963; Welinder et al., 2016). Prothrombin mRNA levels were slightly decreased in the livers of hibernating 13-lined ground squirrels, consistent with this explanation (Gillen et al., 2021). To assess the overall level of activation of clotting, levels of the irreversible complex between thrombin and its inhibitor antithrombin may be used. Thrombin-antithrombin complexes were decreased during hibernation in 13-lined ground squirrels, consistent with suppression of secondary hemostasis (Bonis et al., 2019). An alternative to reducing clotting factors would be to increase anticoagulants. Yet, the anticoagulant proteins antithrombin and protein C are not reduced in torpid hamsters and ground squirrels (Bonis et al., 2019). However, in hibernating bears antithrombin levels are reduced, but it is not known if this is due to decreased production of antithrombin or increased consumption (Welinder et al., 2016). Collectively, torpor features the suppression of both the intrinsic and extrinsic arms of the coagulation cascade leading to prolonged APTT times and a mild extension of PT times (Biorck et al., 1962; Lechler and Penick, 1963; Pivorun and Sinnamon, 1981; De Wit et al., 1985; Iles et al., 2017). Diverse hibernators including shallow hibernators (bears), facultative hibernators (hamsters), and deep seasonal hibernators (ground squirrels and bats) use similar adaptations. These common strategies reduce both primary and secondary hemostasis leading to prolonged APTT (Figure 7).

The decreased activities of clotting factors in the blood correlate well with the changes in APTT and PT, however, the differences in protein levels could be due to decreased synthesis or increased consumption during torpor. Many of the blood proteins involved in secondary hemostasis are produced in liver, and a few are released from platelets and endothelial cells. In addition to traditional activity assays, transcriptome, proteome, and metabolomes are being analyzed in liver and plasma to identify changes in hemostatic regulators. The livers of torpid 13-lined ground squirrels revealed decreases in prothrombin, factor V, factor IX, tissue factor, and heparin cofactor II mRNA while VWF mRNA was decreased in bone marrow (Cooper et al., 2016; Gillen et al., 2021). There was also an increase in α2 macroglobulin mRNA in liver during hibernation, a serine protease inhibitor that can inhibit a broad range of clotting factors and its protein level has been previously shown to increase during torpor (Srere et al., 1995), potentially contributing to the suppressed state of hemostasis in torpor. Although these findings are in line with a suppression of hemostasis in torpor, some other studies demonstrate changes that are less unidirectional or involve both procoagulant and anticoagulant factors. For example, clotting factor mRNA expression was unexpectedly decreased in the brains of hibernating Himalayan marmots, but not in their livers (Bai et al., 2019). Another marmot transcriptome study compared two species, but did not look into seasonal variation in expression (Liu et al., 2016). While transcriptomes are useful in measuring new synthesis of proteins, activity in plasma is ultimately estimated by measuring levels of circulating protein which can be detected using proteomics. The bear plasma proteome revealed decreases in clotting factors similar to that reported for individual factor assays with winter increases in α2 macroglobulin, potentially inhibiting thrombin, whereas thrombin itself and fibrinogen were also increased. In contrast, factors VIII and IX and VWF decreased in winter, in line with a suppression of hemostasis, but also the anticoagulant factors antithrombin and protein C decreased (Welinder et al., 2016). An arctic ground squirrel liver proteome also revealed increased α2 macroglobulin but a decrease in antithrombin during torpor (Shao et al., 2010). Finally, a liver proteome of 13-lined ground squirrels showed no seasonal differential expression of proteins involved in hemostasis (Rose et al., 2011). Although all analyzed hibernating mammals demonstrated some suppression within the hemostatic system, there remains some variation between species regarding which factors are involved and whether some procoagulant factors are increased. For the most part proteomic studies seem to align better with assays of individual proteins than do transcriptomic studies. This suggests that protein levels may be regulated more by consumption than transcriptionally. Further analysis of the plasma and liver proteomes may reveal other adaptations that regulate hemostasis.

Proteins released from endothelial cells after clot formation initiate fibrinolysis. The cross-linked fibrin network is enzymatically degraded by plasmin, which is formed from plasminogen by tissue plasminogen activator (t-PA). t-PA is slowly released by damaged endothelium enabling a gradual degradation of fibrin after the bleeding has stopped and tissue regeneration has started (Gonias, 2021). A marker of fibrinolysis activation is the level of circulating complexes between tPA and its inhibitor plasminogen activator inhibitor 1 (PAI-1). In ground squirrel torpor, PAI-1 levels are decreased along with tPA-PAI-1 complexes consistent with a hyperfibrinolytic state (Bonis et al., 2019). In ground squirrels, plasminogen levels drop two-fold while in bears and hamsters an increase is seen during torpor. Upon fibrinolysis, fibrin in a clot is cleaved into fibrin degradation products, of which D-dimers can be detected in plasma and is commonly used in the diagnosis of venous or arterial thrombosis. In bears, hamsters, and ground squirrels no increases in D-dimers are seen during hibernation, consistent with low levels of clot formation and fibrinolysis (Iles et al., 2017; Bonis et al., 2019). The lack of D-dimers is consistent with suppression of secondary hemostasis and absence of actual clot fibrinolysis during torpor, while the hyperfibrinolytic state reflected by increased tPA and PAI-1 may be an added layer of protection in the event that a clot does form (Figure 7).