Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Portal hypertension, the most important complication in patients with cirrhosis of the liver, is a serious and life-threatening disease for which there are few therapeutic options. Because angiogenesis is a pathological hallmark of portal hypertension, the goal of this study was to determine the effects of sorafenib—a potent inhibitor of proangiogenic vascular endothelial growth factor receptor 2 (VEGFR-2), platelet-derived growth factor receptor β (PDGFR-β), and Raf kinases—on splanchnic, intrahepatic, systemic, and portosystemic collateral circulations in two different experimental models of portal hypertension: rats with prehepatic portal hypertension induced by partial portal vein ligation and rats with intrahepatic portal hypertension and secondary biliary cirrhosis induced by bile duct ligation. Such a comprehensive approach is necessary for any translational research directed toward defining the efficacy and potential clinical application of new therapeutic agents. Sorafenib administered orally once a day for 2 weeks in experimental models of portal hypertension and cirrhosis effectively inhibited VEGF, PDGF, and Raf signaling pathways, and produced several protective effects by inducing an approximately 80% decrease in splanchnic neovascularization and a marked attenuation of hyperdynamic splanchnic and systemic circulations, as well as a significant 18% decrease in the extent of portosystemic collaterals. In cirrhotic rats, sorafenib treatment also resulted in a 25% reduction in portal pressure, as well as a remarkable improvement in liver damage and intrahepatic fibrosis, inflammation, and angiogenesis. Notably, beneficial effects of sorafenib against tissue damage and inflammation were also observed in splanchnic organs. Conclusion: Taking into account the limitations of translating animal study results into humans, we believe that our findings will stimulate consideration of sorafenib as an effective therapeutic agent in patients suffering from advanced portal hypertension. (HEPATOLOGY 2009.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Portal hypertension is the most important complication that develops in patients with cirrhosis of the liver and remains a leading cause of morbidity and mortality worldwide.1 The portal hypertensive syndrome is characterized by a pathological increase in portal pressure and by the development of hyperdynamic splanchnic circulation, with an increase in blood flow in splanchnic organs draining into the portal vein and a subsequent elevation in portal venous inflow. Such increased portal venous inflow is a significant factor maintaining and worsening portal pressure elevation and determining the formation of ascites. Another characteristic feature of portal hypertension is the formation of an extensive network of portosystemic collateral vessels; these include gastroesophageal varices, which are prone to rupture that can cause massive, life-threatening gastroesophageal bleeding. In addition, collateral vessels result in shunting of portal blood into the systemic circulation and play an important role in other major complications of chronic liver disease, including portosystemic encephalopathy and sepsis.1
Unfortunately, the management of patients with portal hypertension still is a relevant clinical problem, and current therapy with nonselective β-adrenergic blockers has significant limitations due to adverse events and unpredictable response.2 Therefore, it is clear that new treatment strategies are needed to improve the prognosis of patients with advanced portal hypertension. Understanding the vascular and molecular alterations occurring in portal hypertension could contribute to this aim.
Recent studies from our laboratory have highlighted that angiogenesis, the growth of new blood vessels from a preexisting vascular bed,3 is a pathological hallmark of portal hypertension. Thus, increased splanchnic neovascularization regulated through the coordinated action of vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF) has been found in experimental models of portal hypertension and has been shown to be a crucial mechanism by which portal hypertension, hyperdynamic splanchnic circulation, and portosystemic collateralization are initiated and stabilized.4-9 Importantly, these prior studies have made the control of new blood vessel formation a promising therapeutic target for patients suffering from portal hypertension and chronic liver disease.
Until now, however, investigation of antiangiogenesis agents focusing simultaneously on several circulatory disturbances associated with the pathophysiology of portal hypertension (disturbances in splanchnic, intrahepatic, systemic, and portosystemic collateral circulations) has not been conducted. Such a comprehensive and systematic approach is required for any careful translational research directed toward defining the effectiveness of new agents with potential therapeutic application in human patients. For example, a treatment could have clinical benefit on the intrahepatic circulation but be detrimental to the splanchnic circulation, and vice versa. In the present study, we determined the effects of sorafenib, an orally active multikinase inhibitor, on the splanchnic, intrahepatic, and systemic circulations and on portosystemic collateral vessels in two different experimental models of portal hypertension: rats with prehepatic portal hypertension induced by partial portal vein ligation, and rats with intrahepatic portal hypertension and secondary biliary cirrhosis induced by common bile duct ligation. Sorafenib potently blocks the tyrosine kinases of VEGF receptor 2 (VEGFR-2) and PDGF receptor β (PDGFR-β), as well as the Raf serine/threonine kinases along the Raf/mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK) pathway.10 Integration of the signals decoded by these different transduction pathways orchestrates the cellular events acting at different stages during angiogenesis.3 Sorafenib is farthest along in clinical development and has been approved in several countries worldwide for treatment of renal cell carcinoma and hepatocellular carcinoma.11, 12 The latter is a common complication of advanced cirrhosis, which means that sorafenib is currently being used in patients with portal hypertension. Hence, evaluation of the effects of sorafenib in animal models of portal hypertension seems timely and pertinent.
Sorafenib was purchased from Bayer Pharmaceuticals (West Haven, CT). Antibodies against CD31, VEGF, VEGFR-2, PDGF, PDGFR-β, lipopolysaccharide-binding protein (LBP), interleukin-1β (IL-1β), tumor necrosis factor α (TNF-α), and glyceraldehyde 3-phosphate dehydrogenase were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti–endothelial nitric oxide synthase (eNOS) antibody was obtained from BD Biosciences (Franklin Lakes, NJ). Antibodies against heme oxygenase-1 (HO-1) and peroxidase-conjugated secondary antibodies were obtained from Stressgen (Sidney, Canada). Antibodies against ERK1/2, phospho-ERK1/2, phospho-eNOS, and inducible NOS (iNOS) were obtained from Cell Signaling Technology (Danvers, MA). Antibody against von Willebrand factor (vWF) was obtained from Dako (Carpinteria, CA), and antibody against CD43 was obtained from Serotec (Raleigh, NC). 51Cr-labeled microspheres were obtained from Perkin-Elmer (Boston, MA). All other reagents were obtained from Sigma (St. Louis, MO).
Studies were performed in male Sprague-Dawley rats (300 to 350 g). All procedures were approved by the Laboratory Animal Care and Use Committee of the University of Barcelona and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication 86-23, revised 1985).
Prehepatic portal hypertension was induced by partial portal vein ligation (PPVL) as described.13 Briefly, under anesthesia with a combination of ketamine (100 mg/kg) (Merial, Lyon, France) plus midazolam (5 mg/kg) (Baxter, Deerfield, IL), a calibrated constriction of the portal vein was performed using a single ligature of 3-0 silk tied around the portal vein and a 20-gauge blunt-tipped needle. The needle was then removed, leaving a calibrated constriction of the portal vein.
Secondary biliary cirrhosis with intrahepatic portal hypertension was induced by common bile duct ligation (CBDL) as described.14 While each animal was under anesthesia, the common bile duct was occluded by double ligature with 5-0 silk thread. The bile duct was then resected between the two ligatures. Two days after surgery, the presence of bilirubin in the urine turned it a dark brown color, indicating successful ligation. In sham-operated (SHAM) animals, portal vein and common bile duct were similarly manipulated but not ligated.
Sorafenib (PPVL, n = 7; CBDL, n = 12; SHAM, n = 8) or vehicle (0.9% sodium chloride; PPVL, n = 8; CBDL, n = 8; SHAM, n = 7) was administered orally by gavage, once a day, for 2 weeks at dose levels of 2 mg/kg/day in PPVL and SHAM rats and 1 mg/kg/day in CBDL animals. Treatments began when portal hypertension was already developed (1 week after PPVL or 2 weeks after CBDL).5, 15 This is closer to the clinical situation of the portal hypertensive patient, who is usually diagnosed, and eventually treated, when portal hypertension is already quite advanced and has caused clinical manifestations. Doses and schedule of administration of sorafenib were selected on the basis of earlier work showing that these were efficacious in inhibiting angiogenesis, that sorafenib has a long half-life (24 to 48 hours), and that steady-state concentrations are reached after 7 days.10
Under anesthesia, PE-50 catheters were introduced into the femoral artery and portal vein to measure mean arterial pressure (MAP) (mm Hg), heart rate (beats/minute), and portal pressure (PP) (mm Hg). A perivascular ultrasonic flowprobe (Transonic Systems, NY) was placed around the superior mesenteric artery to measure superior mesenteric artery blood flow (SMABF) (mL · minute−1 · 100 g−1). Superior mesenteric artery resistance (SMAR) (mm Hg · mL−1 · minute−1 · 100 g−1) was calculated as (MAP − PP)/SMABF.8 Cardiac output (mL/minute) was measured using the thermodilution technique. Briefly, a thermocouple probe was placed in the aortic arch via the carotid artery and connected to a Cardiac Output System (ADInstruments, Colorado Springs, CO). The thermal indicator (0.1 mL of room temperature 0.9% saline) was injected into the right atrium through a PE-50 catheter placed in the jugular vein.
Determination of the Extent of Portosystemic Collateral Vessels.
The extent of portosystemic collateralization was quantified by injection of 51Cr-labeled microspheres (diameter, 15 ± 3 μm; specific activity, 52.68 mCi/g) into the spleen8 as collateralization (%) = [lungs radioactivity/(lungs radioactivity + liver radioactivity)] × 100.
Western Blot Analysis.
Lysate proteins were separated via sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and western blotting was performed using the corresponding primary antibodies and a chemiluminescence detection system. Loading accuracy was evaluated via membrane rehybridization with antibodies against β-actin or glyceraldehyde 3-phosphate dehydrogenase.8
Tissues (mesentery, jejunum, or liver) were fixed in 10% formalin, embedded in paraffin, sectioned, and stained with hematoxylin-eosin (H&E).8 Computer-based quantitating analysis of angiogenesis was performed in mesenteric sections with the assistance of Axiovision software (Zeiss, Germany) by scanning of the entire section and counting areas occupied by vascular structures. We obtained a value from each tissue section (one section per rat) and then averaged the results from individual sections in each group (six to eight animals per group). Group values reflect the average readings from all sections in the group.8
For semiquantitative analysis of hepatic fibrosis, liver sections were stained with 0.1% Sirius red,16 photographed, and analyzed using a microscope equipped with a digital camera. Six fields from each slide were randomly selected, and the red-stained area per total area was measured using AxioVision software. Values are expressed as the mean of 18 to 24 fields taken from three to four animals per group.
Immunostaining of paraffin-embedded liver sections was performed with anti-CD43 and anti-vWF antibodies diluted 1:500 or, as a negative control, with phosphate-buffered saline. Bound antibodies were visualized using diaminobenzidine as chromogen, and slides were then counterstained with hematoxylin. The number of CD43- and vWF-positive cells and vessels was quantified using AxioVision software.
Values are expressed as the mean ± standard error of the mean and compared via an unpaired Student t test and analysis of variance. Statistical significance was established at P < 0.05.
In all the animals studied, the 2-week course of treatment with sorafenib was well tolerated without any sign of toxicity, adverse effects (reduced body weight gain, diarrhea, or hemorrhage), or drug-induced mortality.
Effects on Splanchnic Neovascularization.
Consistent with our recent observations,4-9 overexpression of the proangiogenic factors VEGF and PDGF, the endothelial cell markers VEGFR-2 and CD31, and the perivascular cell markers PDGFR-β and α-smooth muscle actin (α-SMA) was observed in mesenteries from rats with prehepatic portal hypertension induced by PPVL when compared with SHAM control animals (P < 0.05) (Fig. 1A). Of note, these findings were also reproduced in the CBDL model of intrahepatic portal hypertension and cirrhosis (Fig. 2A). Interestingly, in both experimental models, therapy with sorafenib was highly effective in reducing VEGF and PDGF expressions, which translated into substantial inhibition of splanchnic neovascularization (down-regulation of CD31 and VEGFR-2 expressions) and perivascular cell coverage of neovessels (down-regulation of α-SMA and PDGFR-β expressions), compared with the corresponding vehicle-treated groups (Figs. 1A, 2A). In agreement with these western blot results, computer-based quantitative analysis of neovascularization in tissue sections also demonstrated that the vascular area was markedly reduced by approximately 80% in mesenteries from drug-treated relative to vehicle-treated PPVL and CBDL rats (Figs. 1B, 2B). These anti-angiogenic effects of sorafenib were not observed in SHAM rats (Figs. 1A, 2A).
The anti-angiogenic effects of sorafenib can be mediated not only by inhibition of receptor tyrosine kinases such as VEGFR-2 and PDGFR-β, but also by blockade of the ubiquitous Raf/MEK/ERK signaling pathway, which plays a major role in the regulation of endothelial cell proliferation and neoangiogenesis.10 Therefore, we determined the effectiveness of sorafenib in inhibiting the Raf/MEK/ERK signaling cascade in portal hypertensive rats by immunoblotting with antibodies specific for phosphorylated ERK-1 and ERK-2 kinases and total ERK1/2. Indeed, sorafenib markedly inhibited signaling through the Raf kinase pathway, as evidenced by a decrease in ERK1/2 phosphorylation in the intestines of sorafenib-treated PPVL and CBDL rats, compared with those receiving vehicle (Figs. 1C, 2C).
Effects on Portosystemic Collateral Circulation.
An extremely high degree of portosystemic collateral vessels developed 2 weeks after PPVL (98.69 ± 0.87%) and 4 weeks after CBDL (98.87 ± 0.37%) in animals receiving vehicle. Sorafenib significantly reduced this portosystemic collateralization both in PPVL rats (18% reduction; P < 0.05) (Fig. 1D) and CBDL rats (18% decrease; P < 0.05) (Fig. 2D) compared with their corresponding vehicle-treated groups.
Both vehicle-treated PPVL and CBDL rats exhibited portal hypertension, increased SMABF, and decreased SMAR and MAP compared with vehicle-treated SHAM animals (Fig. 3A), thus demonstrating the hyperkinetic splanchnic circulation characteristic of portal hypertension.1 Values of heart rate were similar among groups (Fig. 3A).
In PPVL rats, sorafenib administered as a daily oral treatment for 2 weeks markedly attenuated the hyperdynamic splanchnic circulation, as indicated by a 17% decrease in SMABF (P = 0.012) and a 30% increase in SMAR (P = 0.004) compared with vehicle-treated PPVL rats (Fig. 3A). Despite the decrease in splanchnic blood flow, PP was not reduced by sorafenib in PPVL animals. This was most likely due to the reduction of portosystemic collateralization observed in PPVL rats following administration of sorafenib (Fig. 1D). In addition, sorafenib did not modify MAP (P = 0.2) or heart rate (P = 0.8) in PPVL rats compared with vehicle (Fig. 3A), but it caused a marked and significant 24% decrease in cardiac output (143.4 ± 5.9 mL/minute in vehicle-treated PPVL rats and 108.7 ± 7.9 mL/minute in sorafenib-treated PPVL rats [P = 0.040]).
In CBDL rats, the hyperdynamic splanchnic circulation was also reduced after sorafenib treatment with a 28% decrease in SMABF (P = 0.0002) and a 39% increase in SMAR (P = 0.022) (Fig. 3A). Unlike what was observed in the prehepatic portal hypertensive model, the attenuation of splanchnic hyperemia translated in cirrhotic rats into a marked (25%) decrease in PP (P = 0.0002) (Fig. 3A), despite the degree of portosystemic collateral vessels being similarly reduced in both experimental models (Fig. 2D). These findings suggest that the intrahepatic vascular resistance of CBDL rats could have been reduced by sorafenib treatment (see Effects on Intrahepatic Fibrosis section and Discussion below). Sorafenib did not modify MAP in CBDL rats (P = 0.7) but induced a significant 11% reduction in heart rate (P = 0.02) (Fig. 3A) and a marked 31% decrease in cardiac output that did not reach statistical significance (156.6 ± 14.8 mL/minute in vehicle-treated CBDL rats; 107.7 ± 18.5 mL/minute in sorafenib-treated CBDL rats [P = 0.16]). In SHAM rats, sorafenib had no significant hemodynamic effects (Fig. 3A).
The sorafenib-induced attenuation of hyperdynamic splanchnic circulation observed in PPVL and CBDL rats was most likely due to inhibition of splanchnic neovascularization rather than inhibition of vasodilatation, because no effects were seen after sorafenib on the expression of eNOS and HO-1, which catalyze the production of the vasodilators nitric oxide and carbon monoxide, respectively (Fig. 3B).17, 18
Effects on Intrahepatic Fibrosis.
To evaluate the possibility that sorafenib decreased hepatic vascular resistance, we determined the intrahepatic effects of sorafenib in CBDL rats through a series of complementary experimental approaches. CBDL resulted in distortion of the normal liver architecture, with a marked proliferation of bile ductules and an extensive deposition of fibrillar collagen in portal tracts and central veins of the liver, as identified by staining of liver sections with H&E and Sirius red (Fig. 4A). Impressively, these histological and structural changes were substantially decreased in response to administration of sorafenib, with a 49% reduction in fibrosis (Sirius red staining area) in the liver from sorafenib-treated CBDL rats compared with vehicle-treated CBDL animals (Fig. 4A,B).
Because activated hepatic stellate cells are a major contributor to hepatic fibrogenesis,16, 19 we measured the protein expression of α-SMA, which is expressed by hepatic stellate cells when they gain a myofibroblast-like phenotype in response to liver injury. α-SMA expression was remarkably high in livers from vehicle-treated CBDL rats compared with SHAM animals, consistent with stellate cell activation in fibrotic livers (Fig. 4C). Interestingly, sorafenib caused a profound decrease in α-SMA expression in CBDL rats (Fig. 4C), suggesting that this treatment inhibited activation of stellate cells. These results were further supported by our findings of an increased expression of PDGF, which is the most potent mitogen for hepatic stellate cells,16, 19 and its receptor PDGFR-β in livers from vehicle-treated CBDL rats (Fig. 4C), together with a marked reduction of these overexpressions by sorafenib (P < 0.05) (Fig. 4C).
Effects on Intrahepatic Inflammation.
Current evidence suggests that the process of hepatic fibrosis is driven primarily by the development of inflammation in response to liver injury.16 Therefore, we determined whether the sorafenib-stimulated attenuation of fibrosis was also associated with a reduction in inflammatory activity. Indeed, using histological analyses (H&E) as well as immunohistochemistry with staining for CD43, we found an increased inflammatory cell infiltrate in livers from CBDL rats, compared with SHAM control animals, and attenuation of this inflammatory infiltration (24% decrease) after sorafenib treatment (Fig. 5A,B).
These immunohistochemical studies were supported and confirmed via western blot analyses showing an increased protein expression of the proinflammatory mediators TNF-α, LBP, and iNOS in livers from vehicle-treated CBDL rats compared with SHAM animals (Fig. 5C), and a marked reduction of this overexpression by sorafenib treatment over a period of 2 weeks (Fig. 5C).
Effects on Intrahepatic Neovascularization.
Because liver fibrogenesis and inflammation are frequently associated with hepatic angiogenesis,20 we also conducted studies to determine hepatic neovascularization and the expression of angiogenic factors in livers of sorafenib-treated and vehicle-treated CBDL rats and in corresponding SHAM-operated control animals. Western blot analysis revealed a marked 200% increase in VEGFR-2 expression in vehicle-treated CBDL rats compared with SHAM rats (P < 0.05) (Fig 6A). VEGFR-2 is predominantly expressed on endothelial cells, and its expression is up-regulated during angiogenesis.3 Therefore, VEGFR-2 overexpression most likely reflects an increase in intrahepatic neovascularization in livers from CBDL rats. Notably, sorafenib decreased the expression of VEGFR-2 in CBDL rats by 45% (P < 0.05) (Fig. 6A). These results fit well with the immunohistochemical studies of vWF protein expression, a marker of endothelial cells,3 which confirmed an increased number of vWF-positive vessels in livers from vehicle-treated CBDL rats compared with SHAM animals (322% increase), and a marked 59% reduction of this neovascularization following sorafenib treatment (Fig. 6B). Intriguingly, increased intrahepatic angiogenesis in CBDL rats was not associated with VEGF overexpression. Rather, the expression of VEGF was lower in CBDL rats than in SHAM animals, and it was not modified by sorafenib (Fig. 6A).
Effects on Intrahepatic eNOS and HO-1 Expression.
To determine whether the decrease of intrahepatic vascular resistance in livers from CBDL rats could also be assigned to sorafenib-induced vasodilatation, we measured the protein expression of the vasodilatory enzymatic systems eNOS and HO-1. As expected,17, 18 a robust decrease in phosphorylated eNOS expression, as well as overexpression of HO-1, was observed in livers from vehicle-treated CBDL rats compared with SHAM animals (Fig. 6C,D). We found no detectable changes in the expression of phospho-eNOS, total eNOS, or HO-1 in CBDL rats in response to sorafenib (Fig. 6C,D), indicating that the sorafenib-mediated reduction in intrahepatic vascular resistance is unlikely to be due to nitric oxide and/or carbon monoxide overproduction.
Effects on Intestinal Inflammation.
Next, we wanted to determine whether sorafenib had anti-inflammatory effects, not only in the liver, but also in splanchnic organs. To this end, histological studies using H&E were performed on portions of the jejunum from all experimental groups. Small intestinal mucosa of vehicle-treated PPVL rats had decreased intestinal villus height and abundant inflammatory infiltration compared with SHAM animals (Fig. 7A). These histological changes were also present and even more prominent in CBDL animals (Fig. 7A). Interestingly, daily treatment with sorafenib produced a marked attenuation of the intestinal inflammation observed in both PPVL and CBDL rats (Fig. 7A). In addition, the expression of the proinflammatory proteins TNF-α and iNOS was increased in mesenteries from PPVL and CBDL rats receiving vehicle, and markedly reduced by sorafenib (Fig. 7B). Sorafenib also decreased the IL-1β overexpression observed in PPVL rats (Fig. 7B). Interestingly, the expression of LBP was also elevated in PPVL and CBDL rats, as compared with SHAM rats, and sorafenib attenuated this overexpression (Fig. 7B).
Remarkable and very promising amelioration of portal hypertension has been observed in animal models with several antiangiogenic strategies.4-9, 21-23 Our present study determined, in a comprehensive and systematic manner, whether the multiple kinase inhibitor sorafenib causes beneficial effects on the splanchnic, intrahepatic, and systemic circulations and on portosystemic collateral vessels in two different experimental models of portal hypertension: prehepatic portal hypertension induced by PPVL and intrahepatic portal hypertension with secondary biliary cirrhosis induced by CBDL. This approach, which focused simultaneously on several pathophysiological processes associated with portal hypertension, is necessary for any careful translational research directed toward defining the efficacy and potential clinical application of new agents. Importantly, the use of sorafenib as an orally active multikinase inhibitor has an additional advantage in that its efficacy, safety, and clinical benefit in the treatment of several human malignancies has been documented,11, 12 including renal cell carcinoma and hepatocellular carcinoma. The latter is a common complication of advanced cirrhosis, which means that sorafenib is currently being used in patients with portal hypertension.
Our study demonstrated that sorafenib has several protective effects on splanchnic, intrahepatic, systemic, and portosystemic collateral circulations in PPVL and CBDL rats. First, our results indicate that a 2-week course of treatment with sorafenib administered orally to PPVL and CBDL rats is efficacious in interfering with the VEGF and PDGF signal transduction pathways, as well as with the Raf/MEK/ERK signaling cascade, all of which are crucial to the tightly regulated multistep angiogenic process contributing to the formation, stabilization, and maturation of newly formed vessels.3, 10 Accordingly with the ability of sorafenib to target several proangiogenic signaling pathways, it also showed substantial angioinhibitory efficacy, inducing an approximately 80% decrease in splanchnic neovascularization in both PPVL and CBDL animals. It is important to note that these effects were not observed in SHAM animals, indicating that sorafenib acts only against migrating and proliferating capillary endothelial cells at sites of angiogenesis, but it does not affect the normal nonproliferating vessels.
The prolonged suppression of pathologic neovascularity after sorafenib also translated into a marked attenuation of the hyperdynamic splanchnic and systemic circulations, and decreased the extent of portosystemic collateral vessels as well. In the PPVL model, the decrease in splanchnic blood flow caused by sorafenib did not translate into a significant reduction in PP. That could be due to the significant decrease observed in the extent of portosystemic collaterals after sorafenib treatment, which increases portocollateral resistance and, therefore, increases the vascular resistance of the portal venous system. However, in the CBDL model, the sorafenib-induced attenuation of splanchnic hyperemia (28% reduction in SMABF and 39% increase in SMAR) resulted in a 25% decrease in PP, despite a similar decrease in portosystemic collateralization as in the prehepatic portal hypertensive model, indicating that the vascular resistance of the cirrhotic liver was lowered by sorafenib.
Indeed, we found a remarkable improvement in liver damage and intrahepatic fibrosis and inflammation in CBDL rats following sorafenib treatment, which is likely to be the basis for the diminished intrahepatic resistance to portal blood flow. Histopathological analyses using H&E and Sirius red stainings revealed a robust attenuation in bile duct proliferation, and reduction of fibrillar collagen accumulation and fibrosis in livers from sorafenib-treated CBDL rats compared with the vehicle-treated group. Sorafenib also induced a marked decrease in intrahepatic inflammation, as demonstrated in histological sections by immunostaining for CD43, which is typically expressed by infiltrating inflammatory cells,16 as well as in protein extracts analyzed by western blotting for the proinflammatory mediators TNF-α, iNOS, and LBP.17, 24, 25 The mechanisms by which sorafenib attenuates fibrosis and inflammation are most likely mediated through inhibition of hepatic stellate cell activation, evidenced by strong reduction of α-SMA and PDGFR-β expressions, which are markers of activated stellate cells.16, 19 Of note, sorafenib also markedly decreased PDGF expression, which is the most potent proliferative factor for hepatic stellate cells.16, 19, 26 It is important to note that activation of perisinusoidal hepatic stellate cells is a critical step in hepatic fibrogenesis.16, 19 Consistent with our data, antifibrotic effectiveness has been also observed with some VEGF and PDGF inhibitors in different animal models.21-23
In the present study, we also found a higher degree of neovascularization in livers from CBDL rats than in SHAM animals, as evidenced by elevated VEGFR-2 protein expression and vWF-positive vessels. The occurrence of hepatic angiogenesis has been frequently described in chronic liver disease,20 where it appears to play a critical pathogenic role, either as part of the formation of liver fibrosis and inflammation or as a compensatory response to decreased blood supply in the cirrhotic liver. We observed a marked reduction of this neovascularization following sorafenib treatment. Intriguingly, the increased intrahepatic angiogenesis observed in CBDL rats was not associated with VEGF overexpression. Instead, the expression of VEGF was even lower in CBDL rats than in SHAM animals, and it was not modified by sorafenib. The reason why VEGF expression is decreased in livers of CBDL rats despite having increased intrahepatic neovascularization is at present unknown. Notably, reduced VEGF expression associated with VEGFR-2 overexpression and increased vascular density has also been found recently in cirrhotic livers.27, 28 In addition, studies in CBDL rats have shown a transient up-regulated expression of VEGF in the liver, almost limited to the first days after the BDL procedure.29 It can be proposed, but not yet proven, that VEGF production increases in the early stages of cirrhosis to induce endothelial cell proliferation and formation of new vessels, whereas in advanced stages, stabilization and maturation of newly formed vessels is mainly mediated by other growth factors.30, 31 In any case, angiogenesis is a multifactorial, extremely complex process involving a huge number of molecules and cells regulated at different levels by various factors.3 Hence, it is not surprising that different angiogenic factors may be observed depending on the fibrosis stage, grade of inflammation, or maturation of vasculature.
The protective effects of sorafenib against tissue damage and inflammation were observed not only in livers of cirrhotic rats, but also in splanchnic organs from both PPVL and CBDL animals. Indeed, we found increased inflammation in the intestines of PPVL and CBDL rats, as well as elevated mesenteric expression of the proinflammatory mediators TNF-α, IL-1β, and iNOS, compared with the SHAM group.17, 24 The LPS-binding protein LBP, which enhances cellular responses to LPS,25 was also overexpressed in PPVL and CBDL animals. These data support that small intestinal bacterial overgrowth and translocation of bacteria across the intestinal wall play an important role in the intestinal inflammation associated with portal hypertension.6, 7, 32 Interestingly, sorafenib markedly reduced the splanchnic inflammatory response in both portal hypertensive experimental models. In addition, the finding that iNOS protein expression was increased in mesenteries from PPVL and CBDL rats could suggest that, although the main enzymatic source of the splanchnic NO overproduction that contributes to hyperdynamic splanchnic circulation in portal hypertension is eNOS,13, 17 a small part of NO derived from iNOS could also participate in the portal hypertensive splanchnic hyperemia.
As noted earlier, sorafenib caused a marked 80% decrease in splanchnic neovascularization, but only a moderate (18%) reduction in the extent of portosystemic collaterals. It is likely that a greater decrease in collateralization may take longer than the 2-week course of treatment evaluated in this study. It should be noted in this regard that patients with cirrhosis who undergo orthotopic liver transplantation maintain an increased portocollateral blood flow for months.33 On the other hand, it is also possible that there are intrinsic differences in the sensitivity of vessels to the lack of growth factors or differences in the extent of vessel maturation.3, 34 In this regard, it is important to point out that collaterals are frequently large vessels, suggesting a high degree of vessel maturation.35 Therefore, the integrity of mature collateral vessels may be less influenced by antiangiogenic therapy than that of the newly formed splanchnic vasculature. The marked decrease in PP and splanchnic hyperemia in response to sorafenib is a very relevant effect, because even if collaterals are still present, their potential risk to result in clinical complications would be substantially reduced by this treatment.35-37
In conclusion, our current study provides in vivo evidence that sorafenib has several beneficial effects in experimental models of portal hypertension by inducing an approximately 80% decrease in splanchnic neovascularization, which translates into a marked attenuation of the hyperdynamic splanchnic and systemic circulations, as well as a significant 18% decrease in the extent of portosystemic collateral vessels. In cirrhotic rats, treatment with the multitargeted tyrosine kinase inhibitor also resulted in a 25% reduction in portal pressure and a remarkable improvement in liver damage and intrahepatic fibrosis, inflammation, and angiogenesis. Notably, the beneficial effects of sorafenib against tissue damage and inflammation were observed not only in livers of cirrhotic rats, but also in splanchnic organs from portal hypertensive and cirrhotic animals, as described above. Keeping in mind the limitations of translating results in animal models into clinical practice, we believe that our findings will be stimulating for consideration of this therapeutic approach in patients suffering from advanced portal hypertension, and we predict accelerated progress in this field of research.
We thank Hector Garcia for technical assistance and Raul Mendez for helpful discussions and critical reading of the manuscript.