LIVER INJURY/REGENERATION

Antiangiogenic Treatment with Sunitinib AmelioratesInflammatory Infiltrate, Fibrosis, and Portal Pressure

in Cirrhotic RatsSonia Tugues,1 Guillermo Fernandez-Varo,1 Javier Munoz-Luque,1 Josefa Ros,1 Vicente Arroyo,2 Juan Rodes,2

Scott L. Friedman,3 Peter Carmeliet,4,5 Wladimiro Jimenez,1 and Manuel Morales-Ruiz1

Liver cirrhosis is a very complex disease in which several pathological processes such asinflammation, fibrosis, and pathological angiogenesis are closely integrated. We hypothe-sized that treatment with pharmacological agents with multiple mechanisms of action willproduce superior results to those achieved by only targeting individual mechanisms. Thisstudy thus evaluates the therapeutic use of the multitargeted receptor tyrosine kinase inhib-itor Sunitinib (SU11248). The in vitro effects of SU11248 were evaluated in the humanhepatic stellate cell line LX-2 by measuring cell viability. The in vivo effects of SU11248treatment were monitored in the livers of cirrhotic rats by measuring angiogenesis, inflam-matory infiltrate, fibrosis, �-smooth muscle actin (�-SMA) accumulation, differential geneexpression by microarrays, and portal pressure. Cirrhosis progression was associated with asignificant enhancement of vascular density and expression of vascular endothelial growthfactor-A, angiopoietin-1, angiopoietin-2, and placental growth factor in cirrhotic livers. Thenewly formed hepatic vasculature expressed vascular cellular adhesion molecule 1 and in-tercellular adhesion molecule 1. Interestingly, the expression of these adhesion moleculeswas adjacent to areas of local inflammatory infiltration. SU11248 treatment resulted in asignificant decrease in hepatic vascular density, inflammatory infiltrate, �-SMA abundance,LX-2 viability, collagen expression, and portal pressure. Conclusion: These results suggestthat multitargeted therapies against angiogenesis, inflammation, and fibrosis merit consid-eration in the treatment of cirrhosis. (HEPATOLOGY 2007;46:1919-1926.)

According to reports of the World Health Organi-zation, cirrhosis is one of the leading causes ofmortality worldwide and a shortened life expect-

ancy despite recent therapeutic advances. Cirrhosis ischaracterized by the presence of a hepatic inflammatoryinfiltrate that occurs early in the history of the disease,prior to the onset of significant clinical manifestations; itbecomes chronic during the evolution of the illness, par-ticularly in chronic viral hepatitis.1 This dynamic inflam-matory state is believed to contribute to the progression ofliver fibrogenesis and cirrhosis.2

Angiogenesis is of major importance during adult tis-sue repair3 and is also a hallmark of inflammatory pro-cesses where both phenomena are closely integrated. Forinstance, many inflammatory mediators have direct an-giogenic activities, and may also indirectly stimulate othercells to produce angiogenic factors such as vascular endo-thelial growth factor (VEGF)-A.4 Angiogenesis, in turn,contributes to the perpetuation and the amplification ofthe inflammatory state due to the expression of adhesionmolecules and chemokines in the neovasculature, whichpromote the recruitment of inflammatory cells. The in-

Abbreviations: ang-1, angiopoietin-1; ang-2, angiopoietin-2; FITC, fluoresceinisothiocyanate; HSC, hepatic stellate cell; MAPK, mitogen-activated protein kinase;PDGFR, platelet-derived growth factor receptor; �-SMA, �-smooth muscle actin;SU11248, Sunitinib; VEGF, vascular endothelial growth factor; VEGFR, vascularendothelial growth factor receptor.

From the 1Department of Biochemistry and Molecular Genetics, University ofBarcelona, Barcelona, Spain; 2Liver Unit-Institut de Malalties Digestives, HospitalClınic, IDIBAPS, University of Barcelona, Ciberehd, Spain; the 3Division of LiverDiseases, Department of Medicine, Mount Sinai School of Medicine, New York,NY; and the 4Department for Transgene Technology and Gene Therapy and 5TheCenter for Transgene Technology and Gene Therapy, Leuven, Belgium.

Received March 8, 2007; accepted July 9, 2007.Supported by grants from the Ministerio de Educacion y Ciencia Plan Nacional

de I�D�I (SAF2007-63069 to M.M.-R., SAF2006-07053 to W. J.), the Na-tional Institutes of Health (DK56621 to S. L. F.), and Ministerio de Sanidad yConsumo (PI041198 to M.M.-R.; and CC02/03 and PI041198 to M.M.-R.).

Address reprint requests to: Manuel Morales-Ruiz, Ph.D., Department of Bio-chemistry and Molecular Genetics, Hospital Clınic, Villarroel 170, Barcelona08036, Spain. E-mail: morales@clinic.ub.es; fax: (34) 93-2275697.

Copyright © 2007 by the American Association for the Study of Liver Diseases.Published online in Wiley InterScience (www.interscience.wiley.com).DOI 10.1002/hep.21921Potential conflict of interest: Nothing to report.Supplementary material for this article can be found on the HEPATOLOGY website

(http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html).

1919

flammatory response can also be accentuated by new ves-sels acting as transporters of nutrients to the site ofinflammation and tissue remodeling.5 In this context,there is considerable evidence suggesting that during earlyinflammatory states, angiogenesis may contribute to thetransition from acute to chronic inflammation.

The existence of important angiogenic processes in cir-rhotic liver is widely accepted. For instance, in cirrhoticlivers, scar tissue is surrounded by a dense vasculature.6

Additionally, neovascularization is significantly increasedduring the development of liver fibrosis in both humanand animal studies.7 In light of these observations, andconsidering that chronic inflammation and angiogenesisare associated in many disorders, this link could also beexploited therapeutically in cirrhosis to modulate chronicinflammation.

An interesting therapeutic candidate to be used as aninhibitor of angiogenesis in cirrhosis is Sunitinib(SU11248). This indolinone molecule was designed tohave a broad selectivity for the split kinase family of re-ceptor tyrosine kinases, including KIT and FLT3,8 andhas efficacy as a potent antitumor and antiangiogenicagent in clinical trials for treatment of cancer.9,10 Its anti-angiogenic efficacy is attributable to inhibition of vascularendothelial growth factor receptor (VEGFR) and platelet-derived growth factor receptor (PDGFR), both of whichare essential for angiogenesis development.11 Apart fromangiogenesis, the inhibition of PDGFR has an additionaltherapeutic capability, because PDGF is among the mostpotent mitogenic and chemotactic agents for hepatic stel-late cells (HSCs).12 These cells, which rapidly induce thereceptor � subunit (PDGFR-�) when they are activat-ed,13 play a key role in the development and progressionof hepatic fibrosis.

We hypothesized that SU11248 treatment may havemultiple mechanisms of action against cirrhosis progres-sion. Therefore, we tested its efficacy in culture and in ananimal model of cirrhosis.

Materials and Methods

Animals and Treatment. The study was performedin male adult Wistar rats according to the criteria of theInvestigation and Ethics Committees of the HospitalClınic. Cirrhosis was induced by inhalation of CCl4 asdescribed previously.14 For the in vivo study, 10 cirrhoticrats were treated daily with 40 mg/kg of SU11248 (Pfizer,Inc., NY) via oral administration for 1 week. In parallel,another group of 10 cirrhotic rats received a daily intra-gastric administration of citrate-buffered solution (pH3.5) as a control.

Fluorescein Isothiocyanate–Dextran Angiography,Histology, Immunohistochemistry, and Immunofluo-rescence. See online expanded Experimental Procedures.

Reverse-Transcription Polymerase Chain Reactionand GeneChip Hybridization. Total RNA was ex-tracted from the frozen liver with Trizol reagent (LifeTechnologies, Rockville, MD). One microgram of totalRNA was reverse-transcribed using the First StrandcDNA Synthesis Kit (Roche, Mannheim, Germany).Then, complementary DNA samples were amplified for30-35 cycles (94°C for 30 seconds, 55-60°C for 30 sec-onds, 72°C for 1 minute). Specific primers used for com-plementary DNA amplification were: placental growthfactor, 5�-GTCGCTGTAGTGGCTGCTGTGGTG-3�(forward primer) and 5�-TGTGGGGTTTTGCTTTG-CTTCCTC-3� (reverse primer); and HPRT, 5�-GGG-GCTATAAGTTCTTTGCTGAC-3� (forward primer)and 5�-CATTTTGGGGCTGTACTGCTTGAC-3� (re-verse primer).

RNA from livers of 4 nontreated cirrhotic rats and 4SU11248-treated cirrhotic rats was hybridized to 8 high-density oligonucleotide microarrays (Affymetrix, SantaClara, CA) as described previously.14 Microarray resultshave been deposited in the Gene Expression OmnibusMIAME-compliant database (http://www.ncbi.nlm.nih.gov/geo/; accession number: GSE6929).

Cell Culture and Western Blotting. Human umbil-ical vein endothelial cells were isolated and cultured asdescribed previously.15 The LX-2 cell line, which is de-rived from normal primary human HSCs that have beenimmortalized by selection in low serum, were maintainedas described previously. These cells expressed markers ofactivated HSCs and presented a similar phenotype to thatof activated HSCs in vivo.16

Serum-starved cells were incubated with or without 1�M SU11248, 100 ng/mL human VEGF-A, 200 ng/mLhuman angiopoietin-1, and 20 ng/mL PDGF-BB (R&DSystems, Minneapolis, MN). Then, cell lysates were pre-pared in lysis buffer and Western blotting was performedusing anti-Erk-1/2 mitogen-activated protein ki-nase (MAPK), anti–phospho-Erk-1/2 (Thr202/Tyr204)MAPK, anti-Akt, and anti–phospho-Akt (Ser473)–spe-cific antibodies (Cell Signaling Technology, Beverly,MA) as described previously.17

Cell Viability Assay. For the quantification of cell via-bility, we performed the MTT assay using the 3-(4,5-dim-ethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide reagent(Sigma Chemical, St. Louis, MO). (See also online expandedExperimental Procedures available at http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html.)

Hemodynamic Measurements. Animals were anes-thetized with inactine (0.5 mL/kg) and prepared for mea-

1920 TUGUES ET AL. HEPATOLOGY, December 2007

surement of hemodynamic parameters as describedpreviously.17 Mean arterial pressure and portal pressurewere continuously recorded in a multichannel system(MX4P and MT4; Lectromed, Ltd., Jersey, Channel Is-lands, UK).

Statistical Analysis. Statistical differences were ana-lyzed via Student t test. A P value of less than 0.05 wasconsidered significant. Gene expression analysis was per-formed using libraries from the Bioconductor project(http://www.bioconductor.org) rooted in the computingenvironment R (see also online expanded ExperimentalProcedures).

Results

Vascular Proliferation During Cirrhosis Progres-sion. Vascular proliferation was investigated via immu-nolabeling using the anti-vWF antibody and byfluorescein angiography. In cirrhotic livers, vWF-labeledvessels were nearly 10 times more prevalent than in con-trol rats (Fig. 1A, panels b and a, respectively; Fig. 1C).Similarly, fluorescein isothiocyanate (FITC)–dextran an-giography revealed significant growth of a disorganizedvascular network in cirrhotic livers that was associated

with the absence of normal lobular drainage through thecentral vein (Fig. 1A, panels d and c, respectively). Vas-cular growth was much more pronounced as cirrhosisprogressed (data not shown). Next, vascular changes werealso analyzed in other splanchnic beds, including the mes-enteric region (Fig. 1B, panels e and f) and the tissueadjacent to the hepatic hilum (Fig. 1B, panels g and h).vWF immunolabeling showed an increase in the numberof mesenteric vessels in cirrhotic rats (Fig. 1B, panels e andf; Fig.1D). These results were corroborated by FITC–dextran angiographies (Fig. 1B, panels g and h).

Induction of Angiogenic Factors in Cirrhotic Liv-ers. Liver tissue was processed from control and cirrhoticrats at 8-12 weeks (early cirrhosis) and 17-20 weeks (ad-vanced cirrhosis), respectively, after the beginning ofCCl4 treatment. All cirrhotic animals included in the sec-ond group had well-developed cirrhosis and ascites. Therewas a marked increase in the abundance of VEGF-A, an-giopoietin-1 (ang-1), angiopoietin-2 (ang-2), and placen-tal growth factor in cirrhotic livers compared withcontrols. The expression of these growth factors washigher in cirrhotic rats with ascites than in those withoutascites (Fig. 2A).

Fig. 1. Increased hepatic and splanchnic neovascularization in cirrhotic rats. (A) Hepatic vasculature was analyzed by immunostaining of paraffinliver sections with anti-vWF antibodies (panels A and B; original magnification �100) and by FITC–dextran angiography (panels C and D). Cirrhoticlivers (panels B and D) exhibited an increased number of microvessels compared with control livers (panels A and C). The histogram in (C) showsthe number of vWF-stained vessels quantified in control and cirrhotic livers (n � 10 for each group). Data are expressed as the mean � standarderror of the mean. *P � 0.001 compared with control rats. (B) Splanchnic vasculature was visualized with anti-vWF antibodies (panels e and f;original magnification �100) and by performing FITC–dextran angiographies (panels g and h). Cirrhotic rats (CH), (panels F and H) had a significantvascular proliferation in both the mesenteric area (panel F) and in the proximity of the hepatic hilum area (panel h) compared with control rats (CT),(panels E and G). (D) Histogram shows the number of vWF-stained vessels quantified in the mesenteric tissue of control and cirrhotic rats (n � 10for each group). Data are expressed as the mean � standard error of the mean. *P � 0.001 compared with control rats.

HEPATOLOGY, Vol. 46, No. 6, 2007 TUGUES ET AL. 1921

Because the expression of VEGF-A and ang-2 is acti-vated by hypoxia, we next investigated whether the distri-bution of these proteins may be attributable to hypoxicconditions. Immunohistochemistry for pimonidazoleprotein adducts has been used as a tissue marker of hyp-oxia.18 As shown in Figure 2A, pimonidazole staining wasnot detected in control livers. In contrast, the distributionof pimonidazole adducts displayed a similar pattern tothat of VEGF-A and ang-2 expression in cirrhotic livers.

Cellular Infiltration and Correlation with the Ex-pression of VCAM-1 and ICAM-1 in the Vasculatureof Cirrhotic Livers. As shown in Fig. 3, CD11b-positiveand CD3-positive cells were barely detected in livers ofcontrol rats. In contrast, a significant increase in CD11band CD3 immunoreactivity was detected in cirrhotic liv-ers. This increase in inflammatory infiltrate density wasmainly localized in portal tracts and fibrous septa close tothe newly formed vasculature. Next, the expression of

VCAM-1 and ICAM-1 cell adhesion molecules in thesenewly formed blood vessels was examined with VCAM-1/ICAM-1 and vWF immunofluorescent staining (Fig.4). VCAM-1 was weakly expressed in the sinusoidal lin-ing cells of control livers and did not colocalize with bloodvessels. In contrast, VCAM-1 immunostaining in cir-rhotic livers was stronger and colocalized well with theneovasculature (yellow color in the merge panels). Similarto VCAM-1, a weak constitutive ICAM-1 expression waspresent in the hepatic sinusoid of control animals. In con-trast, portal tracts and sinusoids of cirrhotic livers dis-played a strong ICAM-1 immunoreactivity thatcolocalized with vWF immunostaining (yellow color inthe merge panels).

SU11248 Inhibits VEGF-A and Ang-1–DependentActivation of Erk-1/2 In Vitro and Angiogenesis InVivo. The activity of SU11248 against VEGFR and Tie2receptors was evaluated in vitro. Stimulation of humanumbilical vein endothelial cells with VEGF-A and ang-1induced time-dependent activation of Erk-1/2 MAPK.This effect was inhibited by preincubation of human um-

Fig. 2. Hepatic expression of angiogenic growth factors and colocal-ization with hypoxic tissue during cirrhosis progression. (A) Immunostain-ing profile is shown for VEGF-A, ang-1, ang-2, and pimonidazole (pim)adducts in livers of control rats (n � 10), cirrhotic rats with early cirrhosis(early CH, n � 10), and cirrhotic rats with advanced cirrhosis (advancedCH, n � 10). In all panels, the brown color indicates positive immuno-staining. (Original magnification �100.) (B) Levels of messenger RNAencoding placental growth factor were evaluated via reverse-transcriptionpolymerase chain reaction using total liver RNA isolated from control rats,cirrhotic rats with early cirrhosis, and cirrhotic rats with advanced cirrho-sis. Amplification of HPRT was used as a housekeeping control.

Fig. 3. Immunofluorescent localization of CD11b-positive, CD3-posi-tive, and vWF-positive cells in cirrhotic livers. CD11b (red), CD3 (red),and vWF (green) immunofluorescent staining in control (CT) and cirrhoticlivers (CH) was performed using specific antibodies. In cirrhotic livers,staining with CD11b and CD3 revealed a significant increase in areas ofinflammatory infiltrate that were adjacent to the neovasculature withoutsignificant colocalization with the vWF vascular marker. (Original magni-fication �100.)

1922 TUGUES ET AL. HEPATOLOGY, December 2007

bilical vein endothelial cells with 1 �M SU11248 (Fig.5A).

Next, the antiangiogenic activity of SU11248 was eval-uated in vivo via immunostaining with an anti-vWF an-tibody and via FITC–dextran angiography. As shown inFig. 5B, cirrhotic rats treated with SU11248 (panel b)exhibited a significant reduction (P � 0.001) in hepaticvascularization compared with cirrhotic rats treated withvehicle (panel a) (5.1 � 0.2 vessels/field versus 16.4 � 0.4vessels/field, respectively). This decrease in hepatic vascu-lar density was also observed on FITC–dextran angiogra-phy (panels c and d). Similarly, the number of secondaryvessel branches quantified in the small intestine was sig-nificantly lower (P � 0.001) in SU11248-treated cir-rhotic animals (panel f) than in cirrhotic rats treated withvehicle (panel e) (2.8 � 0.2 vessels/field versus 20.2 � 1.7vessels/field, respectively). No tertiary vessel brancheswere detected in the bowel of SU11248-treated animalscompared with the significant number of tertiary

branches detected in vehicle-treated cirrhotic animals.Furthermore, SU11248 treatment was also associatedwith a significant decrease in vessel density in the tissueadjacent to the portal vein (data not shown). These resultsdemonstrated that the treatment of cirrhotic rats withSU11248 effectively inhibits angiogenesis in all of thetissues examined.

SU11248 Treatment Decreases Inflammatory Infil-trate in Cirrhotic Livers. In agreement with Fig. 3,portal tracts of cirrhotic livers displayed a strong immu-noreactivity for anti-CD11b and anti-CD3 antibodies(Fig.6, panels a and c, respectively) that was adjacent to

Fig. 4. Immunofluorescent localization of VCAM-1–positive, ICAM-1–positive, and vWF-positive cells in cirrhotic livers. VCAM-1, ICAM-1, andvWF immunofluorescent staining in control (CT) and cirrhotic livers (CH)was performed using specific antibodies. In cirrhotic livers, the extent ofVCAM-1 and ICAM-1 staining was greater than in control animals andwas mainly expressed in the newly formed blood vessels as visualized bycolocalization with the vWF marker in the merge panels (yellow). FITC-conjugated secondary antibody was used for vWF detection—except forthe ICAM-1 and vWF combination in cirrhosis, where Cy3-conjugatedsecondary antibody was used for vWF inmunofluorescent detection in-stead. (Original magnification �100.)

Fig. 5. In vitro and in vivo efficacy of SU11248 treatment. (A) Humanumbilical vein endothelial cells were preincubated with or withoutSU11248 (1 �M) and then stimulated with VEGF-A or ang-1 for differenttime points. Lysates (40 �g of proteins) were analyzed via Westernblotting (wb) with the specific antibodies anti–phospho-Erk-1/2-Thr202/Tyr204 or anti–Erk-1/2, (n � 4). (B) Hepatic vasculature of cirrhotic ratswas analyzed via immunostaining with anti-vWF antibodies (panels a andb; original magnification �100) and FITC–dextran angiography (panels cand d). As detected using both techniques, SU11248 treatment [panelsb and d (n � 10)] was associated with a significant decrease in thenumber of hepatic microvessels compared with nontreated cirrhotic rats[panels a and c (n � 10)]. This inhibitory effect of SU11248 was alsoevident in the small intestines of cirrhotic rats treated chronically withSU11248 (panel f), compared with cirrhotic animals that receivedvehicle (panel e).

HEPATOLOGY, Vol. 46, No. 6, 2007 TUGUES ET AL. 1923

vWF-positive areas (Fig. 6, panels e and g). After inhibi-tion of angiogenesis by SU11248, a significant reductionin macrophages and lymphocyte infiltrate was detected incirrhotic livers (Fig. 6, panels b and d and histograms i andj, respectively) which correlated well with the diminutionof the vasculature detected in portal areas (Fig. 6, panels fand h). Similar results were obtained by quantifyingCD68-positive cells as a measure of macrophage cell in-filtration in cirrhotic livers (Supplementary Fig. 1).

Effect of SU11248 Treatment on HSC Activation,Fibrosis, and Portal Pressure. Immunostaining dem-onstrated that many �-SMA–positive cells accumulatedalong the fibrous septa (Fig. 7A, panel a). Treatment withSU11248 drastically reduced the number of positive cellsin cirrhotic livers as shown via immunohistochemistryand computer-assisted quantification (Fig. 7A, panel b;Fig. 7B). Next, liver fibrosis was assessed using Masson’strichrome stain. As expected, strong collagen staining wasseen around the portal tracks and fibrotic septa in livers ofcirrhotic rats (Fig. 7A, panel c). Importantly, cirrhotic ratstreated with SU11248 showed a significant 30% decreasein hepatic collagen accumulation (Fig. 7A, panel d; Fig.7C).

Next, the transcriptional modifications induced bySU11248 in cirrhotic livers were investigated. Gene ex-pression was analyzed on high-density oligonucleotidemicroarrays. In agreement with Fig. 7A-C, SU11248treatment induced a significant down-regulation of genesmainly related to HSC activation and extracellular matrixformation such as �-SMA, Col4�1, Col1�2, Col5�2,and Col1�1 (Supplementary Table 1).

The hemodynamic effect of SU11248 was evaluatedthrough the measurement of mean arterial pressure and

portal pressure in cirrhotic rats. As expected, cirrhotic ratshad hypotension, a characteristic feature of cirrhosis (107mm Hg). However, treatment with SU11248 did notmodify mean arterial pressure (data not shown). In con-trast, the SU11248-treated group exhibited a significant40% decrease in portal pressure compared with non-treated cirrhotic rats (Fig. 7D).

To test the inhibitory effect of SU11248 treatment onPDGFR activation, we treated serum-starved LX-2 cellswith PDGF-BB (20 ng/mL) for 0 to 30 minutes in thepresence or absence of 1 �M SU11248. PDGFR-� activ-ity was assessed by measuring activation-specific phos-phorylation of Erk-1/2 MAPK and Akt viaimmunoblotting. As seen in Fig. 7E, PDGF-BB treat-ment markedly increased Erk-1/2 MAPK and Aktphosphorylation, which were detectable at the earliesttime point examined (5 minutes). SU11248 inhibitedErk-1/2 MAPK and Akt activation at all time pointsexamined.

We next determined the effect of SU11248 on PDGF-stimulated LX-2 viability. Cells were supplemented withPDGF-BB (20 ng/mL) in the presence or absence ofSU11248 for 48 hours. Cell viability was then determinedby measuring MTT reduction. As shown in Fig. 7F, in-cubation with 20 ng/mL PDGF-BB resulted in a signifi-cant increase in LX-2 viability compared with the controlcondition. In contrast, LX-2 cell viability was reducedsignificantly after 48 hours of incubation with SU11248.These results suggest that the inhibitory effect ofSU11248 on PDGFR activity may contribute to the an-tifibrotic effect observed in cirrhotic livers after SU11248treatment.

Fig. 6. SU11248 treatment decreased in-flammatory infiltrate in cirrhotic livers. (A,B)CD11b, (C,D) CD3, and (E-H) vWF immunoflu-orescent staining in (A,C,E,G) nontreated cir-rhotic animals (n � 10) and (B,D,F,H) cirrhoticrats treated with SU11248 (n � 10) wasperformed using specific antibodies. (Originalmagnification �100.) (I,J) The inflammatoryinfiltrate of macrophages (CD11b-positivecells) and lymphocytes (CD3-positive cells)was quantitatively measured as described inExperimental Procedures. Each column repre-sents the mean � standard error of the mean.*P � 0.01 compared with cirrhotic rats nottreated with SU11248.

1924 TUGUES ET AL. HEPATOLOGY, December 2007

DiscussionThis study revealed that cirrhosis is accompanied not

only by splanchnic vessel growth but also by extensive

hepatic angiogenesis. Additionally, newly formed hepaticvasculature expressed VCAM-1 and ICAM-1. The natureof the inflammatory response is regulated by the induc-tion of adhesion molecules in endothelial cells, which is anecessary event in the recruitment of inflammatory cellsto sites of inflammation. Among these adhesion mole-cules, VCAM-1 and ICAM-1 appear to play a major rolein the adhesion and transmigration of macrophages andlymphocytes across endothelial cells in a variety of acuteand chronic inflammatory diseases.19,20 Double immuno-histochemical staining demonstrated that VCAM-1 andICAM-1 expression in the hepatic vasculature was adja-cent to areas of local inflammatory infiltrate. Interest-ingly, hepatic vascular density as well as VCAM-1 andICAM-1 endothelial expression were higher as cirrhosisprogressed. In this scenario, we hypothesized that liverendothelial cells are activated to form new vessels and toexpress adhesion molecules. Circulating cells are then at-tracted to the activated vasculature, where they adhere toand migrate into the liver parenchyma.

VEGF-A and the angiopoietin families of growth fac-tors are essential for vascular growth.21,22 We found asignificant correlation between placental growth factor,VEGF-A, ang-1, and ang-2 expression and vascular den-sity during cirrhosis progression. Several studies haveshown that these angiogenic factors are up-regulated bycytokines and chemokines that are released by leukocytesand damaged hepatocytes.23 These observations suggestthat initially, proinflammatory cytokines may contributeto the up-regulation of these growth factors. Because tis-sue hypoxia often occurs in inflamed tissue after extensivefibrosis,24 the colocalization of VEGF-A and ang-2 withhypoxic areas was assessed. Our findings showed a clearcolocalization between VEGF-A/ang-2 overexpressionand hypoxia, in agreement with earlier studies.25

Linkage of angiogenesis to inflammatory infiltrate incirrhotic livers suggests that angiogenesis inhibitors mayinterfere with progression of the disease. In fact, studies inexperimental models of cirrhosis have shown that angio-genic inhibitors such as TNP-470 and neutralizingmonoclonal antibodies anti-VEGFR1 or anti-VEGFR2can effectively decrease liver fibrosis.26,27 Consistent withthese reports, we have also demonstrated that cirrhoticrats treated with SU11248 showed a significant decreasein hepatic fibrosis. However, our study extends these find-ings in several significant ways. First, SU11248 treatmentdecreases inflammatory infiltrate in cirrhotic livers (Fig.6). This beneficial effect of SU11248 is likely due to thedecrease in the number of hepatic vessels expressingVCAM-1 and ICAM-1. Second, SU11248 treatment sig-nificantly decreased portal pressure in cirrhotic rats (Fig.7D), which may be ascribed to several mechanisms. First,

Fig. 7. Cirrhotic rats treated with SU11248 showed a significantdecrease in HSC activation, extracellular matrix accumulation, and portalpressure. (A) Liver sections from cirrhotic rats were immunostained for�-SMA. Abundant �-SMA staining (brown) indicated the presence ofactivated HSCs in cirrhosis [panel a (n � 10)]. In SU11248-treatedcirrhotic rats, the number of �-SMA positive cells was reduced signifi-cantly [panel b (n � 10)]. (Original magnification �40.) In addition,liver sections from cirrhotic rats were stained with Masson’s trichromestain. Abundant collagen deposition (blue) indicates the presence of liverfibrosis [panel c (n � 10)]. In SU11248-treated cirrhotic rats, collagendeposition was significantly reduced [panel d (n � 10)]. (Originalmagnification �40.) (B,C) �-SMA and fibrotic positive areas werequantitatively measured as described in Experimental Procedures. (D)Portal pressure was measured in nontreated and SU11248-treatedcirrhotic rats as described. SU11248 significantly decreased portalpressure in cirrhotic rats (n � 10). *P � 0.01 compared with cirrhoticrats not treated with SU11248. (E) LX-2 cells were preincubated with orwithout SU11248 (1 �M) and then stimulated with 20 ng/mL PDGF-BBfor different time points. Lysates (40 �g of proteins) were analyzed viaWestern blotting with antibodies anti–phospho-Erk-1/2-Thr202/Tyr204, an-ti–Erk-1/2, anti–phospho-Akt-Ser473, or anti-Akt (n � 3). (F) LX-2 cellviability was determined by measuring MTT reduction. Cells were incu-bated for 48 hours with PDGF-BB in the absence or presence of 1 �MSU11248. Each column represents the mean � standard error of themean. *P � 0.01 compared with vehicle. #P � 0.01 compared withPDGF-BB treatment (n � 3).

HEPATOLOGY, Vol. 46, No. 6, 2007 TUGUES ET AL. 1925

considerable evidence has demonstrated that HSCs arethe primary source of extracellular matrix accumulationin livers, and PDGFR signaling is known to have a majorrole in this process. Therefore, it is likely that SU11248decreased �-SMA and extracellular matrix accumulationin cirrhotic livers through the inhibition of the PDGFsignaling pathway in HSCs. This hypothesis is supportedby the results obtained with the LX-2 cell line where wedemonstrated that the PDGF-BB–dependent increase incell viability was reduced significantly after SU11248treatment (Fig. 7F).

The increase in portal blood flow is another importantcomponent contributing to portal hypertension. There issignificant evidence suggesting that the increase in portalblood flow is due not only to splanchnic vasodilation butalso to an enlargement of the splanchnic vascular treecaused by angiogenesis.7,28,29 Therefore, we anticipatedthat the significant inhibition of angiogenesis promotedby SU11248 in this anatomic region might translate intoa significant decrease in portal blood flow.

In conclusion, this study demonstrates that therapieswith several molecular targets against angiogenesis, in-flammation, and fibrosis might be beneficial in the treat-ment of cirrhosis. The advantage of using multitargetedinhibitors such as SU11248 is further supported by theobservation that SU5416, which is similar to SU11248but exclusively targets VEGFRs, failed to modify portalpressure in the experimental model of partial portal veinligation.30 This discrepancy may be explained by the lackof specificity of SU5416 for targeting PDGFR comparedwith the multitargeting capability of SU11248.

Acknowledgment: SU11248 was generously suppliedby Pfizer Inc.

References1. Bissell DM. Hepatic fibrosis as wound repair: a progress report. J Gastro-

enterol 1998;33:295-302.2. Friedman SL. Molecular regulation of hepatic fibrosis, an integrated cel-

lular response to tissue injury. J Biol Chem 2000;275:2247-2250.3. Carmeliet P. Angiogenesis in life, disease and medicine. Nature 2005;438:

932-936.4. Cohen T, Nahari D, Cerem LW, Neufeld G, Levi BZ. Interleukin 6

induces the expression of vascular endothelial growth factor. J Biol Chem1996;271:736-741.

5. Jackson JR, Seed MP, Kircher CH, Willoughby DA, Winkler JD. Thecodependence of angiogenesis and chronic inflammation. FASEB J 1997;11:457-465.

6. Rappaport AM, MacPhee PJ, Fisher MM, Phillips MJ. The scarring of theliver acini (cirrhosis). Tridimensional and microcirculatory considerations.Virchows Arch A Pathol Anat Histopathol 1983;402:107-137.

7. Morales-Ruiz M, Jimenez W. Neovascularization, angiogenesis, and vas-cular remodeling in portal hypertension. In: Sanyal AJ, Shah VH, eds.Portal Hypertension: Pathobiology, Evaluation, and Treatment. Totowa,NJ: Humana Press, 2005:99-112.

8. Smith JK, Mamoon NM, Duhe RJ. Emerging roles of targeted smallmolecule protein-tyrosine kinase inhibitors in cancer therapy. Oncol Res2004;14:175-225.

9. Abrams TJ, Lee LB, Murray LJ, Pryer NK, Cherrington JM. SU11248 inhib-its KIT and platelet-derived growth factor receptor beta in preclinical modelsof human small cell lung cancer. Mol Cancer Ther 2003;2:471-478.

10. Deeks ED, Keating GM. Sunitinib. Drugs 2006;66:2255-2266.11. Armulik A, Abramsson A, Betsholtz C. Endothelial/pericyte interactions.

Circ Res 2005;97:512-523.12. Pinzani M, Gesualdo L, Sabbah GM, Abboud HE. Effects of platelet-

derived growth factor and other polypeptide mitogens on DNA synthesisand growth of cultured rat liver fat-storing cells. J Clin Invest 1989;84:1786-1793.

13. Wong L, Yamasaki G, Johnson RJ, Friedman SL. Induction of beta-plate-let-derived growth factor receptor in rat hepatic lipocytes during cellularactivation in vivo and in culture. J Clin Invest 1994;94:1563-1569.

14. Tugues S, Morales-Ruiz M, Fernandez-Varo G, Ros J, Arteta D, Munoz-Luque J, et al. Microarray analysis of endothelial differentially expressedgenes in liver of cirrhotic rats. Gastroenterology 2005;129:1686-1695.

15. Ros J, Leivas A, Jimenez W, Morales M, Bosch-Marce M, Arroyo V, et al.Effect of bacterial lipopolysaccharide on endothelin-1 production in hu-man vascular endothelial cells. J Hepatol 1997;26:81-87.

16. Xu L, Hui AY, Albanis E, Arthur MJ, O’Byrne SM, Blaner WS, et al.Human hepatic stellate cell lines, LX-1 and LX-2: new tools for analysis ofhepatic fibrosis. Gut 2005;54:142-151.

17. Morales-Ruiz M, Cejudo-Martin P, Fernandez-Varo G, Tugues S, Ros J,Angeli P, et al. Transduction of the liver with activated Akt normalizesportal pressure in cirrhotic rats. Gastroenterology 2003;125:522-531.

18. Arteel GE, Thurman RG, Yates JM, Raleigh JA. Evidence that hypoxiamarkers detect oxygen gradients in liver: pimonidazole and retrogradeperfusion of rat liver. Br J Cancer 1995;72:889-895.

19. Cook-Mills JM, Deem TL. Active participation of endothelial cells ininflammation. J Leukoc Biol 2005;77:487-495.

20. Kukielka GL, Hawkins HK, Michael L, Manning AM, Youker K, Lane C,et al. Regulation of intercellular adhesion molecule-1 (ICAM-1) in isch-emic and reperfused canine myocardium. J Clin Invest 1993;92:1504-1516.

21. Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors.Nat Med 2003;9:669-676.

22. Jain RK. Molecular regulation of vessel maturation. Nat Med 2003;9:685-693.

23. Osawa Y, Nagaki M, Banno Y, Brenner DA, Asano T, Nozawa Y, et al.Tumor necrosis factor alpha-induced interleukin-8 production via NF-kappaB and phosphatidylinositol 3-kinase/Akt pathways inhibits cell apo-ptosis in human hepatocytes. Infect Immun 2002;70:6294-6301.

24. Norman JT, Fine LG. Intrarenal oxygenation in chronic renal failure. ClinExp Pharmacol Physiol 2006;33:989-996.

25. Pugh CW, Ratcliffe PJ. Regulation of angiogenesis by hypoxia: role of theHIF system. Nat Med 2003;9:677-684.

26. Wang YQ, Ikeda K, Ikebe T, Hirakawa K, Sowa M, Nakatani K, et al.Inhibition of hepatic stellate cell proliferation and activation by the semi-synthetic analogue of fumagillin TNP-470 in rats. HEPATOLOGY 2000;32:980-989.

27. Yoshiji H, Kuriyama S, Yoshii J, Ikenaka Y, Noguchi R, Hicklin DJ, et al.Vascular endothelial growth factor and receptor interaction is a prerequi-site for murine hepatic fibrogenesis. Gut 2003;52:1347-1354.

28. Fernandez M, Vizzutti F, Garcia-Pagan JC, Rodes J, Bosch J. Anti-VEGFreceptor-2 monoclonal antibody prevents portal-systemic collateral vessel for-mation in portal hypertensive mice. Gastroenterology 2004;126:886-894.

29. Morales-Ruiz M, Tugues S, Cejudo-Martin P, Ros J, Melgar-Lesmes P,Fernandez-Llama P, et al. Ascites from cirrhotic patients induces angiogen-esis through the phosphoinositide 3-kinase/Akt signaling pathway. J Hepa-tol 2005;43:85-91.

30. Fernandez M, Mejias M, Angermayr B, Garcia-Pagan JC, Rodes J, BoschJ. Inhibition of VEGF receptor-2 decreases the development of hyperdy-namic splanchnic circulation and portal-systemic collateral vessels in portalhypertensive rats. J Hepatol 2005;43:98-103.

1926 TUGUES ET AL. HEPATOLOGY, December 2007