Background
Hepatopulmonary syndrome (HPS) and portopulmonary hypertension (POPH) are common complications of liver diseases with high morbidity and mortality around the world. Pulmonary artery pressure decreases in HPS as a result of pulmonary vasodilation and vascularization. In addition, the alveolar-arterial O
2 difference increases in the patients with HPS due to ventilation-perfusion inequality, intrapulmonary shunt and diffusion limitation [
1,
2]. On the other hand, pulmonary artery pressure increases in POPH as a consequence of pulmonary vasoconstriction and vascular remodeling [
3‐
5]. Furthermore, the impairment of gas exchange may occur in POPH patients [
6,
7].
The mechanisms of POPH and HPS have not been fully understood yet. It has been proposed that some vasoconstrictive and proliferative substances may escape the liver through porto-systemic shunt, and influence directly the pulmonary smooth muscle cells and capillary endothelium. Also, a prolonged increase in cardiac output may harm the pulmonary capillary endothelium. All aforementioned may pave the way for vascular remodeling and induction of POPH [
3,
8,
9]. In addition, the inflammatory cytokines and carbon monoxide may involve in the development of HPS [
9,
10]. It has been reported that endothelin-1 increases the expression of endothelial NO synthase and NO production through the activation of endothelin B receptors in HPS [
11,
12]. By contrast, endothelin-1 has a vasoconstrictive effect through endothelin A and B receptors in the smooth muscle cells of pulmonary vessels [
13]. We previously indicated that ET-B receptors on the pulmonary smooth muscle cells play roles in the regulation of pulmonary vascular tone in animal model of liver cirrhosis [
14]. Human studies have shown that reactive oxygen species (ROS) increase in POPH, but not in HPS [
6,
15], whereas, ROS are reportedly increased in animal models of HPS [
7,
16‐
18]. Furthermore, high level of ROS may decrease the bioavailability of NO and lead to the constriction of pulmonary vessels [
19]. Taken together, all these suggested mechanisms cannot be linked directly to POPH because of lacking an accepted animal model for POPH, and some limitations in human studies [
20‐
22].
It has been reported that the prevalence of pulmonary artery hypertension in women is higher than men [
23]. Some scientists link this difference to sexual hormones like estradiol, whereas, others are against this conclusion [
24‐
26]. Also, there are inconsistent studies about the effect of the estradiol receptors of ERα and ERβ in pulmonary hypertension [
27]. Besides, plasma concentration of estradiol increases in male and female animals with liver dysfunctions. This may be due to the production of estradiol in the stomach or the lack of its metabolism in the injured liver [
28,
29]. However, little attention has been paid to the association of pulmonary hemodynamic with serum estradiol in liver dysfunctions.
There are a few controversial investigations concerning the relationship between the severity of liver disease and pulmonary dysfunctions [
30‐
32]. However, due to the complexity in human studies, the classification of liver damage in patients may not be as accurate as the animal studies. HPS can be developed by a common bile duct ligation or drug administration in animals [
7,
18]. Nevertheless, a reliable method has not been introduced for the induction of POPH thus far. Although, POPH can be linked to portal hypertension, the increase in portal pressure by partial mechanical obstruction of the portal vein in animals does not lead to pulmonary hypertension [
33]. In addition, the increase of pulmonary artery pressure following the injection of sephadex microsphere into the portal vein [
22], creation of porto-systemic shunt [
21] and intraperitoneal administration of carbon tetrachloride [
20] could not entirely create a condition like POPH in human.
Unlike the systemic circulation, alveolar hypoxia, constricts the vessels in the affected region of the lung, thereby moves the blood from the area with low oxygen pressure to the well- ventilated ones. This physiologic phenomenon is called hypoxic pulmonary vasoconstriction (HPV). The limited studies have reported that HPV decreases or disappears in patients with liver disease and animal model of cirrhosis [
34,
35]. Nevertheless, the question remains whether the sensitivity of pulmonary vessels to alveolar hypoxia is influenced by the kind or severity of liver disorders.
With the above background, in this study, we established three graded models of liver dysfunctions including mild (PPVL), moderate (CBDL), and for the first time, severe (CBDL+ PPVL), based on liver histology and blood-borne variables. We assessed pulmonary and systemic hemodynamic, and the sensitivity of pulmonary vessels to hypoxia during repeated ventilation with hyperoxic and hypoxic gases. This study was performed in female rats, because the relationship between the estradiol level and the prevalence of pulmonary hypertension with liver dysfunction has not been reported in female animals yet.
Discussion
In this study, we developed three graded levels of liver damage based on the liver histology and blood-borne variables and compared RVSP during ventilation with hyperoxic and hypoxic gases. Although the plasma MDA in the CBDL group tended to be lower than that in the CBDL+PPVL group, it was not statistically significant. Furthermore, the enhancement of serum estradiol in the CBDL+PPVL group was less pronounced than the CBDL group. Also, there was a considerable decrease in plasma platelet level in the CBDL+PPVL group. Besides, the impairment of gas exchange through the blood gas barrier occurred only in the CBDL+PPVL group. The results of WBC suggest a substantial inflammatory reaction in the CBDL+PPVL group. Together, we considered the CBDL+PPVL, CBDL, and PPVL groups as models of severe, moderate and mild liver dysfunction, respectively. Though the previous studies have shown the response of pulmonary vessels to hypoxia in liver dysfunction model, we investigated the sensitivity of pulmonary circulation based on the severity of liver damage, and observed the recovery of repeated HPV only in the PPVL group with mild liver dysfunction. We also proposed, for the first time, ligation of both CBDL and PPVL as a possible animal model for induction of POPH.
Although plasma MDA and NO metabolites increased identically in the CBDL and CBDL+ PPVL groups, data of liver histological scores, liver enzymes, low platelet level, and high WBC are indicating a remarkable liver injury and inflammatory reactions in the CBDL+ PPVL group. Therefore, this group was considered as a model for severe liver damage. Furthermore, plasma MDA and NO metabolites, liver enzymes, WBC and liver histology in the CBDL group were higher than those in the PPVL group, and lower than the ones in the CBDL+PPVL group. Consequently, the CBDL group was considered as a model for moderate liver damage. Also, all the above-mentioned variables did not differ between the PPVL and Sham groups which suggests portal vein ligation per se could not lead to a significant liver dysfunction. Therefore, the model of PPVL was considered as a model for mild liver damage. The survival rate of CBDL+PPVL group was almost similar to that in the CBDL group during 28 days of experiments. However, the mortality rate of animals in the CBDL+PPVL groups increases to 50% after 40 days. Therefore, it cannot be recommended for investigating in longer times if untreated.
In order to exclude the effects of different concentrations of estradiol during the estrus cycle, all female animals were entered the study during diestrus phase. High serum estradiol in the CBDL group can be linked to the overproduction of estradiol or the lack of its metabolism in the stomach [
28,
29]. On the other hand, estradiol in the CBDL+PPVL group did not increase significantly. This may explain partly the high level of RVSP in this group, as estradiol has preventive effects in pulmonary hypertension induced by monocrtotaline or hypoxia [
24,
26,
50,
51]. Therefore, the effects of pulmonary vasoconstrictors may not be counterbalanced substantially by a low estradiol level in the CBDL+PPVL group. As a result, the pulmonary artery hypertension may occur.
High level of RVSPs during ventilation with the first hyperoxic gas in the CBDL+PPVL and CBDL groups were not comparable with other reports. There are a few studies reporting the pulmonary artery pressure or vascular resistance in CBDL model. In one study, pulmonary artery pressure has been measured by inserting a catheter into the pulmonary artery through the umbilical vessel at 2 days before the hemodynamic study in Wistar rats [
35]. This may affect the normal conditions of pulmonary hemodynamic. In another study, pulmonary vascular resistance was measured only 2 weeks after induction of CBDL, which may not be enough long for the induction of HPS [
22]. However, we measured RVSP directly, by inserting a catheter into the right ventricle, with caution, 28 days after induction of liver dysfunction which can be more accurate compared with other investigations. Since RVSP increased markedly in the CBDL+PPVL group, therefore, it can be suggested that ligations of both portal vein and common bile duct in animals may induce a POPH model in rat. However, we did not measure cardiac output and vascular resistance in this study because of some limitations. Therefore, the increase in RVSP might be caused by volume overload, hyperdynamic circulation, vascular remodeling, or combination of them which must be specified in the further studies.
Little change occurred in RVSP during ventilation of animals with the first and second hypoxic gas in the CBDL+PPVL and CBDL groups. A few data have indicated disruption of pulmonary vascular responses to alveolar hypoxia in cirrhotic patients and conscious animals with cirrhosis [
34,
35]. In addition, we evaluated the sensitivity of pulmonary vessels to hypoxia by repeating the hypoxia maneuvers. On the other hand, RVSP increased in the Sham and PPVL groups during ventilation with the first hypoxic gas which were amplified during the second hypoxia maneuver. These data may suggest that elimination of hypoxia response depends on the severity of liver dysfunction. The confirmation is that the hypoxia response in the PPVL group with low liver injury was retained in the second hypoxia maneuver. The increase in RVDP in the CBDl+PPVL group can be related to the damage to the heart and decreased the ability of the heart to tolerate the stress induced by hypoxia. This may also explain the relationship between the severe liver damage with detrimental effects in other organs such as the heart. RVSP decreased in the Sham and PPVL groups during ventilation with hyperoxic gas. However, RVSP increased markedly in the CBDL+PPVL group with no change in the CBDL group during ventilation with hyperoxic gas, which may be related to reducing the bioavailability of NO by oxygen [
52].
mBP in the CBDL+PPVL and CBDL groups was lower than that in the Sham group. Since plasma overload and high cardiac output have been reported in liver cirrhosis [
53], the low mBP at above mentioned groups can be related to the pronounced effect of peripheral vasodilation relative to the volume overload. The results of mBP in the CBDL groups are consistent with the results of Moezi et al. and Nunes et al. that indicated decreased mBP after 4 and 6 weeks of CBDL in male rats, respectively [
35,
54]. Besides, in this study, mBP in the PPVL group was similar to the Sham group due to a mild liver damage. Ventilation of animals with the first and second hypoxic gas decreased mBP roughly in all groups due to the reduction in peripheral vascular resistance [
55] which is consistent with the results of Edmunds, et al. that indicated ventilation of animals with 10% oxygen decreases sharply the arterial pressure in male Wistar rats [
46]. Also, other studies have indicated the arterial pressure falls during acute hypoxia exposure in animals [
56,
57]. Even continuous exposure to hypoxia in conscious animals may decrease both heart rate and systolic blood pressure [
58]. However, in our study, the reduced mBP by hypoxic gas was much pronounced as compared with our previous study and some other works in male rats [
38,
43]. Sex differences in blood pressure response to hypoxia may explain partly this different response [
45]. Also, it is important to mention that in our study, animals were ventilated with hyperoxic gas before each hypoxic maneuver. Ventilation with heperoxic gas constricts the systemic vessels, thereby increases the blood pressure, similar to our study [
59]. Therefore, ventilation with hypoxic gas may exacerbate the systemic vasodilatory response leading to a sharp drop in systemic arterial blood pressure. Acute hypoxia in human rises, decreases or doesn’t change the systemic arterial pressure depends on the interaction between sympathetic outflow of chemoreceptors and peripheral vascular resistance [
60,
61]. However, the effect of peripheral vasodilation may be principal in our experiments compared with the effect of the sympathetic activity. On the other hand, the intermittent or chronic exposure to hypoxic gas increases blood pressure linked to enhancing the sympathetic activity which is different relative to the acute hypoxic condition in our study [
45,
62‐
64]. It should be noted that alterations of mBP in the CBDL and CBDL+PPVL groups were small during ventilation with hypoxic gas which may be linked to the increase of heart rate and volume overload in cirrhotic animals. Hyperoxic gas recovered the arterial pressure because of direct effect of oxygen on increasing the vascular resistance and reducing the bioavailability of NO. [
52,
59]
The heart rate decreased a little (data not shown) at the beginning of hypoxic maneuver linked to a reduction in depolarization rate of cardiac pacemaker cells [
65]. It was followed by a tachycardia induced by both sympathetic activity and vagal withdrawal subsequent to the chemo-reflex activity [
57]. Also, the heart rate decreased transiently during switching from hypoxia to hyperoxia because of increasing the vagal activity [
66].
There was not a difference between the values of PaO
2 in the CBDL and Sham groups, while PaCO
2 and pH in the CBDL group were less than those in the Sham group. This could be caused by cirrhosis-induced hyperventilation [
67]. On the other hand, low PaO
2 in the CBDL+ PPVL group may be linked to the dominant effect of diffusion impairment compared with the hyperventilation [
67]. Furthermore, a low PaO
2/FIO
2 ratio in the CBDL+PPVL group is verifying the injury of the blood gas barrier in the severe liver dysfunction. In both HPS and POPH, low PaO
2 and low saturation of hemoglobin with oxygen have been reported [
6,
7,
68‐
70]. On the other hand, low PaO
2 may lead to pulmonary vasoconstriction and explain partly increased RVSP in the CBDL+PPVL group (Table
2).
Table 2
The comparison of the experimental variables in the CBDL+PPVL versus CBDL models
Liver damage ↑↑ | Liver dysfunction | Liver damage ↑↑ | Liver dysfunction |
Liver enzymes ↑↑ | Liver dysfunction | Liver enzymes ↑↑ | Liver dysfunction |
WBC ↑↑ | Systemic Inflammation | WBC ↑ | Systemic inflammation |
PLT ↓↓ | Vascular thrombosis | PLT ↔ | No change |
Basal RVSP ↑↑ | | Basal RVSP ↑ | |
HPV | No response | HPV | No response |
ROS ↑ | Systemic Inflammation | ROS ↑ | Systemic Inflammation |
PaO2 ↓ | Impairment of gas exchange | PaO2 ↓ | Little effect |
PaO2/FIO2 ↓ | Impairment of gas exchange | PaO2/FIO2 ↓ | Little effect |
Estradiol ↑ | | Estradiol ↑↑ | |
The plasma concentration of MDA in the CBDL group was more than those in the sham and PPVL groups. Other investigators also have expressed that oxidants increase and antioxidants decrease in liver cirrhosis [
6,
71]. ROS may be involved in HPV [
72]. Therefore, the disruption of HPV in the CBDL groups may imply that the pulmonary vasculatures were already maximally stimulated by the observed oxidative stress before hypoxia maneuver. High level of ROS may also lead to vascular remodeling and promotes the pulmonary hypertension [
72] which needs to be investigated in the long term (Table
2). High levels of NO metabolites and MDA in both CBDL groups, suggests the productions of large amounts of NO and ROS. The combination of NO and ROS may produce the peroxynitrite, a potent vasoconstrictor oxidant which increases the pulmonary vascular resistance and pulmonary artery pressure [
73]. In addition, it can be speculated that a part of the inhibitory response to hypoxia is related to the increase in NO production. There are inconsistent results regarding the NO production in the patients with liver diseases or animal models of cirrhosis. For instance, a few studies have shown that NO production increases in human cirrhosis [
74,
75], and NO plays a role in the regulation of pulmonary vascular tone in the animal model of cirrhosis [
35]. In contrary, NO synthase inhibitor protein increases, thereby, NO production decreases in the cirrhotic patients [
6,
76]. However, we measured the plasma NO metabolites which could release from different sources in the body tissues. Furthermore, high level of NO can be partly linked to high estradiol level in the CBDL groups [
77].
The platelet level and RVSP in the CBDL+PPVL group was lower than those in the other groups. It has been indicated that platelet level is linked to pulmonary hypertension and the rate of survival [
78,
79]. On the other hand, the level of metaloproteinase decreases in the patients with liver cirrhosis. This enzyme regulates the Von Willberand factor size and platelet adhesive activity. Then, the reduction of the enzymes may lead to platelet deposition in the afferent pulmonary vessels [
80]. Furthermore, thrombocytopenia in cirrhosis may be caused by the reduction of hematopoietic growth factor thrombopoietin activity in the liver or platelet sequestration in the spleen [
81]. All above possibilities may increase the chance for thrombosis formation and pulmonary hypertension [
82]. Therefore, it may be assumed that at least a part of increased RVSP is caused by low platelet level in the CBDL+PPVL group (Table
2).
Together, POPH occurs following liver cirrhosis with portal hypertension, and is associated with a mild hypoxemia, thrombocytopenia and inflammation [
1,
6,
81,
83]. In this study, RVSP increased substantially in the CBDL+PPVL group. It was also associated with liver damage and portal hypertension induced by partial portal vein ligation. Furthermore, there was a mild hypoxemia in this group. Also, the platelet level decreased only in this group which could increase the prevalence of pulmonary hypertension following thrombotic abnormality. Therefore, the model of CBDL+PPVL can be suggested as a reliable animal model for POPH.