Hypercoagulability in patients with primary biliary cirrhosis has been attributed to elevated fibrinogen and hyperactivity of platelets, which is not observed in non-cholestatic liver diseases, such as chronic hepatitis C, and alcoholic cirrhosis. These changes are believed to be the result of a marked systemic inflammatory activity (Ben-Ari et al., 1997; Papadopoulos et al., 2007). The underlying mechanisms for altered platelet function in cholestasis are unclear. Several mechanisms have been suggested (Bowen et al., 1988; Pihusch et al., 2002; Atucha et al., 2007; Witters et al., 2010; Tripodi et al., 2011), including an intrinsic change in platelet function, e.g., increased aggregation resulting from altered intracellular calcium (Ca²⁺) homeostasis or decreased aggregation due to a storage pool deficiency, the effect of a plasmatic factor such as bile acids or bilirubin or because of the effect of release of ADP-degrading enzymes in the circulation. Regarding platelet activation, the necessary increase in cytoplasmic Ca²⁺ levels, both by release from internal stores and entry of extracellular Ca²⁺, has been reported to be defective both in experimental models of liver cirrhosis and patients (Bandi et al., 1997; Atucha et al., 2007; Annie-Jeyachristy et al., 2008). In previous studies from our laboratory, a greater amount of stored Ca²⁺ together with an increased activity of SERCA, were observed in platelets obtained from a rat model of biliary cirrhosis (Atucha et al., 2007).

The role of bile acids in mediating these alterations in Ca²⁺ signaling and platelet function is not known. Deoxycholic acid (DCA) has been observed to cause profound Ca²⁺ release from intracellular stores and entry from the extracellular medium of intact or permeabilized fibroblast cells (Lau et al., 2005). In diets containing cholesterol and cholic acid (CA), increased thrombin and adenosine diphosphate (ADP)-induced platelet aggregation and an acceleration of the platelet-rich plasma clotting time were found (McGregor et al., 1980). Other authors, however, observed that sodium salts of some bile acids, such as cholic acid, showed a clear inhibitory effect on platelet aggregation by ADP, while the sodium salt of chenodeoxycholic acid (CDA) was an inducer of platelet aggregation (Baele et al., 1980). In other experiments, deoxycholic acid inhibited platelet aggregation and ADP secretion in platelets activated with either collagen or thrombin (Tan et al., 2012).

Thus, in the present study we aimed at analyzing the roles of bile acids such as cholic acid and deoxycolic acid on the altered platelet function observed in a model of cholestasis, the bile duct-ligated rat.

The BDL rats used in the present study had the typical features of this model, jaundice, enlarged liver and spleen, and mesenteric edema. Ascites was present in variable amounts only in the animals included in this group. Sham-operated animals did not have any of these alterations.

Baseline P-selectin expression was significantly greater in the BDL group as compared to the controls and the BDL-ascitic rats, and CA or DCA did not affect these values in any group (Table 1). Platelet aggregation, which we found significantly elevated in the BDL non-ascitic rats in a previous paper (Romecín et al., 2017) was not modified by pretreatment with those bile acids (data not shown). Basal platelets Ca²⁺ levels was significantly elevated by incubation with both CA and DCA (Table 2) in platelets of the control animals groups, while in platelets of the BDL rats only DCA elevated basal Ca²⁺. In platelets of BDL rats with ascites, there was no significant effect of bile acids (BA).

In relation with thrombin responses, incubation with both BA reduced the Ca²⁺ release after stimulation with thrombin (0.3 U/ml) in the absence of extracellular Ca²⁺ (Figure 1A, top panels), but only CA in the BDL group showed a significant difference as compared to its basal (Figure 1B, top panels). In the presence of extracellular Ca²⁺ (Figure 1A, bottom panels), thrombin responses were much greater in all three groups and pretreatment with CA but not with DCA, reduced significantly those Ca²⁺ responses in all three experimental groups (Figure 1B, bottom panels).

The capacitative Ca²⁺ response is shown in Figure 2. Pretreatment with thapsigargin to deplete Ca²⁺ from the intracellular stores slightly elevated Ca²⁺ levels, being significantly greater in the platelets of the BDL rats. However, regarding the effect of BA, only DCA significantly elevated the thapsigargin effect in platelets of the control rats (Figure 2B, top panels). But, the addition of Ca²⁺ in those platelets pretreated with thapsigargin to evoke the capacitative Ca²⁺ entry induced a significantly lower entry of Ca²⁺ in platelets pretreated with both CA and DCA (Figures 2A,B, bottom panels) in all three experimental groups.

To estimate the amount of Ca²⁺ accumulated into intracellular stores, we used the simultaneous addition of thapsigargin and ionomycin. As shown in Figures 3A,B, total Ca²⁺ stored was significantly decreased by pretreatment with both CA and DCA, although these changes were significantly different in the control rats only with CA and in the BDL platelets with DCA.

In previous studies (Atucha et al., 2007; Romecín et al., 2017), we showed in a rat model of liver cirrhosis by bile duct ligation, several platelet alterations compatible with the existence of a hyperaggregatory state. These alterations are related to a defective platelet Ca²⁺ handling, specifically to an enhanced intracellular Ca²⁺ release evoked by thrombin and an increased amount of Ca²⁺ stored in the intracellular organelles and are present before the appearance of ascites. In the present study we confirm these alterations and extend them to show their lack of relationship to bile acids. Our results indicate that, if any, bile acids, would tend to reduce Ca²⁺ movements across platelet membranes which would reduce this hyperaggregatory state characterisitic of the non-ascitic phase of the disease. A similar result would also occur in the ascitic phase, characterized in this case by a lower platelet aggregatory function.

This hyperaggregatory state is related firstly to a higher expression of P-selectin in both groups of cirrhotic rats, indicative of an activated state, which is typical of several types of human cirrhosis (Panasiuk et al., 2001; Vardareli et al., 2007; Xianghong et al., 2013). These data are similar to those found by Watanabe et al. (1998), in cirrhotic patients, who explained it by a change in the fatty acid composition of the platelet membrane. Although bile acids are well known detergents that exhibit membrane-damaging properties against a variety of cells, including blood platelets (Shiao et al., 1993), this does not seem the case, since in our data the expression of P-selectin was unchanged by pretreatment with CA and CDA.

Regarding thrombin-induced intracellular Ca²⁺ changes, both bile acids clearly decreased both Ca²⁺ release from internal stores (that observed in the absence of Ca²⁺) and Ca²⁺ entry from the extracellular space, although only CA, but not DCA, showed a significantly lower response in all three groups, control and BDL-ligated rats. This result agrees with previous data (Bowen et al., 1988; Shiao et al., 1993) but disagrees with those of McGregor (McGregor et al., 1980) showing that cholic acid increased thrombin and ADP-induced platelet aggregation and accelerated plasma clotting time. A similar result was obtained in fibroblasts, where DCA directly induced both Ca²⁺ release from internal stores and persistent entry at the plasma membrane (Lau et al., 2005).

Our results are in keeping with more recent data indicating that bile acids also act as regulatory molecules (Hylemon et al., 2009). Bile acids have been discovered to activate specific nuclear receptors such as farnesoid X receptor (FXR) and others, as well as cell signaling pathways in cells of the liver and gastrointestinal tract. A recent study by Moraes et al. (2016) indicates that FXR ligands inhibited the activation of platelets in response to stimulation of collagen or thrombin receptors, resulting in diminished intracellular Ca²⁺ mobilization, secretion, fibrinogen binding, and aggregation. However, bile acids also activate a number of cellular Ca²⁺ related channel receptors (Dopico et al., 2002; Lee et al., 2012; Rainer et al., 2013). Since Ca²⁺ regulation is critical for multiple platelet functions (Sage, 1997; Rosado and Sage, 2002; Dolan and Diamond, 2014), we suggest that bile acids, acting through platelet receptors, induce inhibition of platelets by reducing Ca²⁺ signaling. Clearly, more experiments would be necessary in order to prove this hypothesis.

A similar effect of bile acids was found when we studied the capacitative Ca²⁺ entry, again suggesting an interference with the channels involved in those responses. Interestingly, bile acids did not reduce the intracellular Ca²⁺ elevation induced by thapsigargin, an inhibitor of SERCA Ca²⁺ ATP-ase, that although modest compared to the capacitative entry, was significantly elevated with DCA in the platelets of the control animals.

Finally, total amount of Ca²⁺ in internal stores, as estimated with the use of thapsigargin and ionomycin, was again reduced by the bile acids, suggesting again that bile acids interfere with Ca²⁺ signaling receptors both at internal and cellular membranes. Clearly, further studies will be necessary to analyze the platelets receptors involved in these alterations.

Patients with liver cirrhosis and experimental models show, in the compensated phase prior to the development of ascites, platelet alterations compatible with a hyperaggregatory state. This is related to enhanced Ca²⁺ release evoked by agonists in activated platelets, secondary to an increased amount of Ca²⁺ stored in the intracellular organelles. These alterations are not related to bile acids, since they reduce the amount of internal Ca²⁺ and the entry from the extracellular space, both necessary to induce platelet aggregation.

PR was the main performer of most of the experiments, EN is our lab technician and without her nothing can be done, MO and DI played a key role with the flow cytometry technique, JG was the main writer of the manuscript and NA supervised all the experimental protocols and procedure laboratories and directed the research.