For consideration of publication in Am J Physiol February 7, 2006

Transcription

1 Page 1 of 45 Articles in PresS. Am J Physiol Gastrointest Liver Physiol (March 30, 2006). doi: /ajpgi Gaudio et al. 1 For consideration of publication in Am J Physiol February 7, 2006 ADMINISTRATION OF r-vegf-a PREVENTS HEPATIC ARTERY LIGATION INDUCED BILE DUCT DAMAGE IN BILE DUCT LIGATED RATS Eugenio Gaudio, M. D. 7# Barbara Barbaro, Ph. D. 2, 6, 7# Domenico Alvaro, M. D. 6 Shannon Glaser, M. S. 3 Heather Francis, B. S. 3 Antonio Franchitto, B. S. 7 Paolo Onori, M. D. 8 Yoshiyuki Ueno, M. D., Ph. D. 5 Marco Marzioni, M. D. 4 Giammarco Fava, M. D. 4 Julie Venter, B. S. 2 Ramona Reichenbach, B. S. 2 Ryun Summers, B. S. 2 Gianfranco Alpini, Ph. D. 1, 2, 4 From 1 Central Texas Veterans Health Care System, 2 Department of Medicine and 4 Systems Biology & Translation Medicine, 3 Division of Research and Education, Scott & White Hospital and The Texas A&M University System Health Science Center, College of Medicine, Temple, TX 76504, 5 Div. Gastroenterol, Tohoku University School of Med, Aobaku, Sendai, Japan, and 6 Div. Gastroenterol and 7 Anatomy, University "La Sapienza", Rome, Italy and 8 Dept. Experimental Medicine, University of L Aquila, Italy. This work was supported by a grant award to Dr. Alpini from Scott & White Hospital and The Texas A&M University System, by a Grant Award from Scott & White Hospital to Shannon Glaser, and by a VA Merit Award, a VA Research Scholar Award and the NIH grants DK58411 and DK to Dr. Alpini and by a Grant from MIUR (COFIN 2003, # _002 and COFIN 2005 # _002) Dr. Domenico Alvaro and by a grant from MIUR (PRIN 2003 and ex 60%) to Prof Eugenio Gaudio and from MIUR Biomedicina, Cluster04, progetto n.5 to Prof. E. Gaudio / P. Onori and by a grant from Health and Labor Sciences Research Grants for the Research on Measures for Intractable Diseases (from the Ministry of Health, Labor and Welfare of Japan) and from Grant-in Aid for Scientific Research C ( ) from JSPS to Dr. Ueno. # Dr. Gaudio and Dr. Barbaro equally contributed to this study. Short title: VEGF regulation of cholangiocyte functions Key words: camp; ductal secretion; intrahepatic biliary epithelium; mitosis; microcirculation; secretin. Copyright 2006 by the American Physiological Society.

2 Page 2 of 45 Gaudio et al. 2 Abbreviations used: BDI = bile duct incannulation; BDL = bile duct ligation; camp = adenosine 3', 5 -monophosphate; PCNA = proliferating cellular nuclear antigen; r-vegf- A = recombinant-vegf-a; VEGF = vascular endothelial growth factor. Address correspondence to: Gianfranco Alpini, Ph. D. VA Research Scholar Award Recipient Professor, Medicine and Systems Biology & Translation Medicine Dr. Nicholas C. Hightower Centennial Chair of Gastroenterology Central Texas Veterans Health Care System The Texas A & M University System Health Science Center College of Medicine Medical Research Building 702 SW H.K. Dodgen Loop, Temple, TX, Phone: or Fax: or galpini@tamu.edu or galpini@medicine.tamhsc.edu

3 Page 3 of 45 Gaudio et al. 3 ABSTRACT The hepatic artery, through the peribiliary vascular plexus, nourishes the intrahepatic biliary tree. During obstructive cholestasis, the nutritional demands of intrahepatic bile ducts are increased as a consequence of enhanced proliferation; in fact, the peribiliary plexus displays adaptive expansion. The effects of hepatic artery ligation (HAL) on cholangiocyte functions during cholestasis are unknown although ischemic lesions of the biliary tree complicate the course of transplanted livers and are encountered in cholangiopathies. We evaluated the effects of HAL on cholangiocyte functions in experimental cholestasis induced by bile duct ligation (BDL). By using BDL and BDL + HAL rats or BDL + HAL rats treated with r-vegf-a for 1 week, we evaluated liver morphology, the degree of portal inflammation and peri-ductular fibrosis, microcirculation, cholangiocyte apoptosis, proliferation, and secretion. Microcirculation was evaluated by scanning electron microscopy vascular corrosion cast technique. HAL induced in BDL rats: (i) the disappearance of the peribiliary plexus; (ii) increased apoptosis and impaired cholangiocyte proliferation and secretin-stimulated ductal secretion; and (iii) decreased cholangiocyte VEGF secretion. The effects of HAL on peribiliary plexus and cholangiocyte functions were prevented by r-vegf-a that, by maintaining the integrity of the peribiliary plexus and cholangiocyte proliferation, prevents damage of bile ducts following ischemic injury. INTRODUCTION Cholangiocytes, the epithelial cells lining the intrahepatic biliary epithelium (6), modify bile, originally secreted at the bile canaliculus (43), by a series of absorptive and

4 Page 4 of 45 Gaudio et al. 4 secretory events regulated by a number of factors including gastrointestinal hormones/peptides, bile salts and nerve receptor agonists (4-7, 11, 31, 38). The gastrointestinal hormone secretin increases ductal secretion by interaction with specific receptors (expressed only by cholangiocytes) (10), an interaction that induces an increase in intracellular adenosine 3',5'-monophosphate (camp) levels (4, 7, 25, 26, 31, 38). Increased intracellular camp levels induces activation of the CFTR Cl - channels (9, 31) and Cl - /HCO 3 - exchanger activity (7, 11, 31), a series of events that leads to secretinstimulated bile and bicarbonate secretion (5, 26). In normal liver, cholangiocytes have low basal DNA synthesis (4, 51). However, cholangiocytes proliferate in a number of experimental models of cholestasis including bile duct ligation (BDL) (4, 5, 12, 25, 37, 38). Cholangiocyte proliferation in the course of cholangiopathies compensates for the loss of injured ducts and, in fact, proliferating cholangiocytes display enhanced basal and secretin-stimulated ductal secretory activities (4-6, 26, 31, 38). A number of studies in rats (3-6, 8-10, 25, 36-40) and humans (45) have shown that changes in cholangiocyte proliferation are associated with parallel modifications in secretin receptor gene expression, secretin-stimulated camp levels and secretin-induced bile and bicarbonate secretion. For example, we have shown in rats that in pathological conditions associated with increased cholangiocyte proliferation (e.g., following BDL or partial hepatectomy) (4, 5, 25, 38), there is enhanced secretin receptor gene expression and augmented basal and secretin-stimulated camp levels and bile and bicarbonate secretion. On the other hand, reduced cholangiocyte proliferation or enhanced cholangiocyte loss (e.g., following vagotomy, acute administration of CCl 4 or depletion of endogenous bile acid pool) (3, 36, 39) is coupled with decreased basal and

5 Page 5 of 45 Gaudio et al. 5 secretin-stimulated camp levels and bile and bicarbonate secretion. In humans, an impaired response to secretin was observed in cholestatic conditions (45). Cholangiocyte proliferation is regulated by neuropeptides, hormones and growth factors including vascular endothelial growth factor (VEGF) (6, 21, 37). Rat cholangiocytes express the protein for VEGF-A and secrete VEGF, and express the VEGF receptor subtypes, VEGFR-2 and VEGFR-3 but not VEGFR-1 (21). VEGF secretion is enhanced in proliferating cholangiocytes from BDL rats, where it stimulates, by autocrine mechanisms, cholangiocyte proliferation (21). The peribiliary plexus (PBP) stems from the hepatic artery, nourishes the biliary tree and sustains a contercurrent of substances reabsorbed from bile toward hepatocytes (23). A true microvascular plexus vascularizes larger ducts, whereas around the smaller ducts the plexus gets progressively simpler (up to a single capillary) and thinner (23). In normal rats, where cholangiocytes are in a quiescent status (39), ligation of the main hepatic artery by its own is not sufficient to induce bile duct damage, suggesting that accessory arteries, collateral vessels or anastomosis between PBP and portal system may overcome the interruption of arterial flow in the main hepatic artery (16, 17). Consistently, interruption of blood flow to intrahepatic bile ducts (by short-term ligation of the hepatic artery of normal guinea pigs) does not alter cholangiocyte secretion (50). The extrahepatic and intrahepatic PBP in the normal liver has been the subject of a number of studies using three-dimensional observations (22-24). Changes in intrahepatic bile duct mass are always associated with changes of the PBP architecture (22, 23). After BDL, the increase in intrahepatic bile duct mass is followed by a parallel growth of the PBP (23), which is fundamental in sustaining the enhanced nutritional and functional demands

6 Page 6 of 45 Gaudio et al. 6 of proliferating cholangiocytes (4-6). Nevertheless, the proliferation of the PBP occurs only after the hyperplasia of the intrahepatic biliary epithelium (23). This finding suggests a cross-talk mechanism between cholangiocytes and endothelial cells, an interaction that mediates the adaptive changes of these cells during liver damage. However, limited information exists on the role of blood supply through the hepatic artery in pathological conditions characterized by cholangiocyte proliferation/loss (14, 48). This concept has clinical implications since ischemic bile duct lesions are considered possible causes of cholestatic disorders, in particular after liver transplantation, hepatic surgery and intra-arterial chemotherapy (13, 15). MATERIAL AND METHODS Materials Reagents were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise indicated. The substrate for -glutamyltranspeptidase ( -GT), N ( -L-glutamyl)-4- methoxy-2-naphthylamide was purchased from Polysciences (Warrington, PA). Porcine secretin was purchased from Peninsula (Belmont, CA). The antibodies against proliferating cellular nuclear antigen (PCNA), VEGF-A and the VEGF receptor subtypes VEGFR-1, VEGFR-2 and VEGFR-3 were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). The recombinant mouse VEGF (r-vegf-a) was purchased from Leinco Technologies Inc. (St. Louis, MO).

7 Page 7 of 45 Gaudio et al. 7 Animal Models Male Fischer 344 rats (150 to 175 gm) were purchased from Charles River (Wilmington, MA). The animals were kept in a temperature-controlled environment (22 C) with a 12- hour light-dark cycle and fed ad libitum rat chow. The studies were performed in: (i) normal rats and normal rats + HAL (for evaluation of cholangiocyte apoptosis and proliferation in liver sections); (ii) 1 week BDL (for isolation of cells) (4, 5) or bile duct incannulated (BDI, for bile collection) (5) rats; and (iii) rats that (immediately after BDL or BDI + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% bovine serum albumin (BSA) or r-vegf-a (2.5 nmol/kg/hour with 0.2% BSA) for 1 week. The dose (nm range) of r-vegf-a administered to BDL + HAL rats was chosen according to the concentration (nm range) of VEGF found in the serum of rats and human in other studies (12, 52). The group of BDL + HAL rats was studied since we observed impaired secretion of VEGF in BDL cholangiocytes after HAL. Since we observed impaired secretion of VEGF in BDL cholangiocytes after HAL, the group of BDL + HAL + VEGF rats was chosen to evaluate whether the effects of HAL observed in BDL rats are prevented by chronic administration of r-vegf-a. BDL and BDI were performed as described (5). HAL was performed as described (29). Before each procedure, animals were anesthetized with sodium pentobarbital (50 mg/kg body weight, IP). Study protocols were performed in compliance with the institution guidelines.

8 Page 8 of 45 Gaudio et al. 8 Isolation of Hepatocytes and Cholangiocytes Hepatocytes were isolated as described (2). Cholangiocytes (97-100% pure by -GT histochemistry) (46) were purified by immunoaffinity separation (4, 28). Cell number and viability (greater than 97%) was assessed by trypan blue exclusion. Body Weight, Liver Morphology, Necrosis, Inflammation and Peri-ductular Fibrosis We evaluated the effect of BDL, HAL and HAL + r-vegf-a administration to BDL + HAL rats on body weight, liver morphology, necrosis, inflammation and peri-ductular fibrosis. We evaluated, by H&E staining of paraffin-embedded liver sections (4 µm thick, 6 slides evaluated for each group), the degree of portal inflammation (18), necrosis and lobular morphology (disarrangement of hepatocytes). At least 10 different portal areas were evaluated. Following H&E staining, liver sections were examined in a coded fashion with an Olympus BX-40 (Tokyo, Japan) microscope equipped with a camera. We evaluated, by Masson s trichrome staining of paraffin-embedded liver sections (4 µm thick, 6 slides evaluated for each group), the degree of fibrosis around proliferating ducts from BDL, BDL+HAL and BDL+HAL+r-VEGF-A treated rats. At least 6 different microscopic fields (10X and 20X) for each slide were analysed in a coded fashion with an Olympus BX-51 microscope (Tokyo, Japan) equipped with a Videocam (Spot Insight, Diagnostic Instrument, Inc. Sterling Heights, MI) and processed with an Image Analysis System (IAS - Delta Sistemi, Roma- Italy). Peri-ductular fibrosis was measured as the volume fraction of the entire liver tissue specimen (% volume fraction of the green

9 Page 9 of 45 Gaudio et al. 9 stained collagen fibres less volume fraction occupied respectively by portal triads and by the parenchyma). Evaluation of Liver Microcirculation Following anesthesia, the abdomen of the selected animal was opened and a cannula (Inpharven diameter 1.4 mm, Inphardial, Italy) was inserted into the aorta and fixed with 2 silk ties. Before flushing the vascular bed with heparinized saline solution (23), the thorax was opened and the right atrium incised to allow the efflux of the perfusate. When the outflow fluid appeared clear of blood, Mercox CL2R resin, diluted with methyl methacrylate monomer (4:1) (34) and mixed with a standard amount of its catalyzer (up to 2 ml catalyzer per 20 ml of base compound), was injected at room temperature. A constant pressure control was maintained (by a CONEL electronic manometer, Rome, Italy) through the lateral port of the cannula s injection valve until resin polymerization was visible. The animals were left at room temperature for 24 hours and after the polymerization of the resin, the livers were removed and macerated in 20% NaOH solution at room temperature. After rinsing in distilled water, the liver casts were placed in 5% trichloroacetic acid solution to free the cast from tissue remnants. The casts were isolated, frozen in distilled water and frozen-dried. They were then glued onto stubs by means of Silver Dag and gold coated in an Sl50 sputterer (Edwards, London, UK). The prepared casts were examined with a Hitachi S4000 Field Emission scanning electron microscope (Hitachi Ltd., Tokyo, Japan) operating at 5-8 kv (23). Evaluation of Cholangiocyte VEGF Protein Expression and Secretion

10 Page 10 of 45 Gaudio et al. 10 Protein expression for VEGF-A, VEGFR-1, VEGFR-2 and VEGFR-3 was evaluated by immunohistochemistry in liver sections (5 µm thick, 6 slides evaluated per group) mounted on glass slides coated with 0.1% poly-l-lysine. Following staining, sections were analyzed in a coded fashion with an Olympus BX-51 microscope (Tokyo, Japan) equipped with a Videocam (Spot Insight, Diagnostic Instrument, Inc. Sterling Heights, MI) and processed with an Image Analysis System (IAS - Delta Sistemi, Roma- Italy). The intensity and distribution of immunostaining were assessed in a coded fashion. Following isolation, hepatocytes or cholangiocytes (1x10 6 ) were incubated at 37 o C for zero or six hours, centrifuged at 1,500 rpm for 10 minutes, the supernatant (100 µl) transferred to a tube and analyzed for VEGF concentration by ELISA (Peninsula Laboratories, Inc, San Carlos, CA). VEGF secretion (ng/1x10 6 cells) was calculated as the difference between the amount of VEGF detected at six hours and the amount detected at time zero. Cholangiocyte Apoptosis and Proliferation We evaluated cholangiocyte and hepatocyte apoptosis by TUNEL analysis (39) in liver sections (3 slides evaluated for each group, 5 µm thick) from the selected group of animals. Following counterstaining with Hematoxylin solution, liver sections were examined by light microscopy with an Olympus BX-40 microscope (Tokyo, Japan) equipped with a camera. Approximately 100 cells per slide were counted in a coded fashion in ten non-overlapping fields. Cholangiocyte and hepatocyte proliferation was evaluated by measuring the number of PCNA- and hepatocyte-positive cholangiocytes. Cholangiocyte growth was also

11 Page 11 of 45 Gaudio et al. 11 evaluated by measuring the % of CK-19- and -GT-positive ducts in liver sections (3 slides evaluated for each group of animals, 5 µm thick) (39). Sections were counterstained with hematoxylin and examined in a random, blinded fashion with an Olympus BX 51 light microscope (Tokyo, Japan). Data were expressed as number of: (i) PCNA-positive cholangiocytes or hepatocytes per each 100 cells measured; and (ii) % of C 19-positive ducts area evaluated in ten different fields (10X or 20X) of slide taken from each 6 blocks randomly taken from medial lobe. Histochemistry for -GT-positive ducts (46) was performed in frozen sections (5 µm thick, 3 slides evaluated for each group). Measurement of Basal and Secretin-Stimulated Ductal Secretion At the functional level, cholangiocyte proliferation was evaluated by measurement of basal and secretin-stimulated camp levels (32, 39) in purified cholangiocytes and bile and bicarbonate secretion (5) in bile fistula rats, two functional indices of cholangiocyte proliferation) (4, 5, 20, 25, 26, 36-39). For the measurement of camp levels, purified cholangiocytes were incubated for 1 hour at 37 C (32) and incubated for 5 minutes at room temperature (32) with 0.2% BSA (basal) or 100 nm secretin with 0.2% BSA. Intracellular camp levels were measured by commercially available RIA kits (Amersham Life Science) (32). Following anesthesia, rats were surgically prepared for bile collection as described by us (5). One jugular vein was incannulated with a plastic cannula to infuse either Krebs Ringer Henseleit (KRH) or secretin (100 nm) dissolved in KRH. Biliary bicarbonate concentration (measured as total CO 2 ) was determined by an ABL 520 Blood Gas System (Radiometer Medical A/S, Copenhagen, Denmark).

12 Page 12 of 45 Gaudio et al. 12 Statistical Analysis All data are expressed as mean ± SEM. Differences between groups were analyzed by the Student unpaired t test when two groups were analyzed and analysis of variance (ANOVA) when more than two groups were analyzed, followed by an appropriate post hoc test. RESULTS Effects of HAL and r-vegf-a on Body Weight, Liver Morphology, Necrosis and Inflammation and Peri-ductular Fibrosis No changes in body weight were observed among BDL (161.7 ± 4.0 gm), BDL + HAL (172.3 ± 7.0 gm) and BDL + HAL + r-vegf-a (165.6 ± 5.6 gm) rats. There were no differences in the amount of lobular damage among BDL rats and rats that (immediately after BDL + HAL) were treated with 0.2% BSA or r-vegf-a for 1 week (see representative Figure 1 a and Table 1). The degree of necrosis, portal inflammation and the % volume fraction of peri-ductular fibrosis observed in BDL rats were reduced in BDL+HAL compared to the value of BDL rats (Figure 1 a-b and Table 1). Following the administration of r-vegf-a to BDL + HAL rats, the degree of necrosis, portal inflammation and the % volume fraction were similar than that of the BDL rat (see representative Figure 1 a-b and Table 1). Effect of HAL and r-vegf-a on the Peribiliary Plexus and VEGF Expression The peribiliary plexus, observed in BDL rats (23), was not demonstrated in BDL + HAL rats by scanning electron microscopy corrosion cast technique. We did not observe PBP

13 Page 13 of 45 Gaudio et al. 13 in BDL+HAL rats because the PBP is nourished by the hepatic artery. In normal rats, PBP was observed more easily in large portal tracts (not shown) (23). In the small portal tracts, the PBP was characterized by single layer of capillaries or even by a single capillary; in BDL+ HAL we do not observed PBP. Administration of r-vegf-a to BDL+HAL rats prevented the HAL-induced microvascular modification (absence of PBP) (Figure 2). The microvascular pattern observed after administration of r-vegf-a to BDL + HAL rats demonstrated the presence of a PBP (Figure 2) with similar characteristics previously described in BDL rats (23). In vascular corrosion casts, we observe only vascular tree because the tissue is completely digested. Therefore in Figure 2, we did not observe bile duct (completely digested) but only the absence of peribiliary plexus. Immunohistochemistry in BDL liver sections shows that bile ducts express VEGF-A, VEGFR-2 and VEGFR-3 (Figure 3 a-b and Table 2). HAL induced a decrease in the number of cholangiocytes positive for VEGF-A and VEGFR-2 and VEGFR-3 receptors compared to liver sections from 1 week BDL rats (Figure 3 a-b and Table 2). Following the administration of r-vegf-a to BDL + HAL rats, the expression of VEGF-A and VEGFR-2 and VEGFR-3 was similar or higher than that of the BDL rat (Figure 3 a-b and Table 2). VEGFR-1 was not expressed by cholangiocytes (Table 2). VEGF-A was predominantly expressed by hepatocytes of centrolobular zone (Table 3). Hepatocyte VEGF-A protein expression did not increase significantly in BDL liver sections compared to normal sections, decreased after HAL and returned to values similar to those of BDL rats (Table 3). Normal rat hepatocytes and cholangiocytes secrete VEGF (Figure 4). Following BDL, there was an increase in VEGF secretion in cholangiocytes (Figure 4, top). VEGF

14 Page 14 of 45 Gaudio et al. 14 secretion significantly decreased in BDL hepatocytes compared to normal hepatocytes (Figure 4, bottom). In cholangiocytes and hepatocytes from BDL+HAL rats, there was decreased VEGF secretion (Figure 4) compared with cholangiocytes and hepatocytes from BDL rats (Figure 4 b). Administration of r-vegf-a prevented the decrease of cholangiocyte VEGF secretion induced by HAL in cholangiocytes and hepatocytes (Figure 4). Cholangiocyte Apoptosis and Proliferation Parallel to previous studies (41), TUNEL analysis showed a few apoptotic bodies in the liver sections of normal (results not shown) and BDL rats (Figure 5). The number of cholangiocytes undergoing apoptosis increased in liver sections from BDL + HAL rats compared to BDL rats (Figure 5). HAL had no effect on cholangiocyte apoptosis of normal rats (not shown). Administration of r-vegf-a prevented the increase in cholangiocyte apoptosis induced by HAL in BDL rats (Figure 5). The number of apoptotic hepatocytes was not changed in the different groups of animals (BDL = 2.87 ± 0.13; BDL + HAL = 3.47 ± 0.16; BDL + HAL + r-vegf-a = 3.39 ± 0.16). Following HAL, the number of PCNA-positive cholangiocytes and the % of CK-19 and -GT-positive ducts decreased compared to liver sections from BDL rats (Table 4 and representative Figure 6 showing the % of CK-19-positive ducts in liver sections). Administration of r-vegf-a prevented the inhibitory effect of HAL on the number of PCNA-positive cholangiocytes and the % of CK-19 and -GT-positive ducts, values that were similar to that of BDL rats (Table 4 and Figure 6). HAL did not alter the number of PCNA-positive cholangiocytes and the % of CK-19 and -GT-positive ducts in normal liver sections (not shown). The proliferation of hepatocytes (PCNA positive) was similar

15 Page 15 of 45 Gaudio et al. 15 in the different groups of animals (BDL = 3.44 ± 0.19; BDL + HAL = 3.04 ± 0.16; BDL +HAL + r-vegf-a = 3.68 ± Basal and Secretin-stimulated camp Levels and Ductal Secretion In agreement with previous studies (25), secretin increased intracellular camp levels of cholangiocytes from BDL rats (Figure 7). HAL significantly reduced basal cholangiocyte camp levels and inhibited secretin-stimulated camp levels of cholangiocytes compared to cholangiocytes from BDL rats (Figure 7). Administration of r-vegf-a prevented the inhibition of basal and secretin-stimulated camp levels induced by HAL (Figure 7). Secretin stimulated bile flow, bicarbonate concentration and secretion of BDI rats (Table 5). In HAL + BDL rats, secretin did not increase bile flow and bicarbonate concentration and secretion of BDI rats (Table 5). Administration of r-vegf-a prevented the inhibition of secretin-stimulated bile and bicarbonate concentration and secretion, which were similar to that of BDL rats (Table 5). DISCUSSION The study demonstrates that in BDL rats, HAL: (i) induced the absence of the peribiliary plexus; (ii) decreased the immunolocalization of VEGFR-2 and VEGFR-3 receptors and VEGF-A in liver sections, and VEGF secretion in purified cholangiocytes and hepatocytes; (iii) induced loss of intrahepatic bile ducts in BDL but not normal rats, caused by both increased apoptosis and decreased cholangiocyte proliferation; no changes in hepatocyte apoptosis and proliferation were observed in BDL rats and rats that

16 Page 16 of 45 Gaudio et al. 16 (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a for 1 week; and (iv) decreased basal and secretin-induced intracellular camp synthesis and impaired basal and secretin-stimulated bile flow and bicarbonate secretion. The adverse effects of HAL on peribiliary plexus, VEGF expression and secretion (in cholangiocytes and hepatocytes), and on cholangiocyte proliferative and secretory activities in BDL rats were all prevented by chronic administration of r-vegf-a to BDL + HAL rats. After BDL, the intrahepatic biliary epithelium undergoes cholangiocyte proliferation (4-6, 25), which leads to bile duct mass expansion, which is followed by an adaptive proliferation of the PBP (23). Proliferating bile ducts are characterized by enhanced cholangiocyte secretory and proliferative activities (5). Therefore, the adaptive proliferation of the PBP (and its circulating factors including VEGF) is fundamental to sustain the enhanced functional and nutritional demands of the proliferating biliary tree. Since proliferation of the PBP follows in order of time the proliferation of bile ducts (23), it is reasonable to suppose that proliferating cholangiocytes modulate the adaptive response of the vascular bed. Consistently, proliferating cholangiocytes express VEGF-A and secrete VEGF (21), which modulates cholangiocyte proliferation by autocrine mechanisms. The fact that: (i) hepatocytes express and secrete VEGF (35, 49): and (ii) hepatocyte VEGF secretion is decreased by HAL and maintained by administration of r- VEGF-A raises the possibility that cholangiocyte growth may be also regulated by a paracrine mechanism. However, we have demonstrated that: (i) hepatocyte VEGF secretion decreases following BDL and it is much lower than cholangiocyte VEGF secretion; and (ii) the proliferation/apoptosis of hepatocytes is similar in BDL rats and in

17 Page 17 of 45 Gaudio et al. 17 rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with BSA or r-vegf-a for 1 week. Taken together, we propose that regulation of BDL cholangiocyte apoptosis/proliferation occurs mainly by an autocrine mechanism (by changes in cholangiocyte VEGF secretion), although a paracrine mechanism (by hepatocyte or vascular VEGF secretion) cannot be ruled out completely by this study. We submitted BDL rats to hepatic artery ligation and evaluated the pathophysiology of the intrahepatic bile duct system at the morphological and functional levels in comparison with BDL and BDL + HAL rats. Interestingly, VEGF-A expression and secretion in cholangiocytes was decreased by HAL. On the basis of this finding, we evaluated whether the effects of HAL on BDL rats are prevented by chronic in vivo administration of r-vegf-a. The dose of VEGF used in our studies is similar to the concentration of VEGF found in the serum of rats and human (12, 52). The decreased VEGF expression and secretion by BDL cholangiocytes was an unexpected finding since, in different tissues, VEGF is usually induced by ischemia (42). However, VEGF is expressed at low levels by quiescent non-proliferating cholangiocytes of normal rat liver (21) but markedly expressed and secreted by cholangiocytes proliferating following BDL (21). This suggests that interruption of hepatic artery blood supply to proliferating bile ducts, characterized by the vanishing of the peribiliary plexus, as demonstrated by vascular corrosion cast, compromises VEGF protein synthesis in cholangiocytes, which is associated with impaired proliferation and with activation of apoptosis. However, the fact that the reduced VEGF expression and secretion by BDLcholangiocytes plays a role in HAL-induced impairment of proliferative and secretory

18 Page 18 of 45 Gaudio et al. 18 activities of BDL rats was suggested by the findings that all the effects of HAL on cholangiocyte function of BDL rats were prevented by chronic in vivo r-vegf-a administration. Furthermore, the administered dose of r-vegf-a induced serum levels of VEGF, which was similar to control BDL rats, both being higher with respects to BDL rats submitted to HAL. Interestingly, we found that in BDL rats HAL induced a decrease in the degree of necrosis, portal inflammation (presumably due to decreased cholangiocyte cytokine secretion) (19) and peri-ductular fibrosis (may be due to decreased cholangiocyte secretion of specific growth factors including platelet-derived growth factor) (27). This finding should be interpreted as a consequence of the reduced cholangiocyte proliferation following HAL, since portal inflammation and peri-ductular fibrosis in the BDL model are triggered by cholangiocyte proliferation via secretion of a number of cytokines/chemokines able to activate stellate cells (27, 44). HAL induced no effect in the liver of normal rats confirming findings from several previous studies (16, 17, 30). This has been attributed to accessory arteries or to anastomosis between the PBP and the portal system, which, if necessary, may overcome interruption of the blood supply through the main hepatic artery (16, 17, 30). Evidently, when the intrahepatic bile duct mass is expanded as occurs after BDL (5), blood supply through the hepatic artery becomes fundamental in sustaining the enhanced nutritional and functional demands of proliferating intrahepatic biliary epithelium (23). Regarding the deleterious effects of HAL on BDL cholangiocyte function, VEGF seems to play a role although we cannot exclude that decreased VEGF expression and secretion is a consequence of impaired proliferation caused by decreased blood supply. However,

19 Page 19 of 45 Gaudio et al. 19 since VEGF is a player in the complex loop of agents sustaining cholangiocyte proliferation after BDL (6, 37), the impaired synthesis and release of VEGF by cholangiocytes after HAL certainly has a role in compromising cell proliferation. In support of this, we have shown (21) that VEGF stimulates cholangiocyte proliferation. Furthermore, we found that HAL-induced impairment of proliferation is associated with decreased basal and secretin induced bile flow and bicarbonate secretion, all these effects being prevented by r-vegf-a administration. This further supports the role of VEGF in mediating the effects of HAL on the functions of the intrahepatic biliary epithelium. In cholangiocytes from HAL + BDL rats, we also observed a decreased level of basal and secretin-induced camp levels that were normalized by r-vegf-a in vivo administration. The second messenger, camp, plays an important role in modulating cholangiocyte growth (20, 25, 36). Furthermore, camp-related intracellular pathways are activated/deactivated by different agents involved in the regulation of cholangiocyte proliferation including the cholinergic system, gastrin, somatostatin and bile salts including taurocholate and ursodeoxycholate (1, 4, 25, 36). For example, intracellular camp levels are elevated in BDL cholangiocytes compared to normal cholangiocytes (4, 7). Chronic in vivo administration of forskolin, a camp activator (33), increases cholangiocyte proliferation and secretin-stimulated ductal secretion in normal rats (20). Thus, a complex loop of neuropeptides, hormones and growth factors, which potentiate each other, sustains cholangiocyte proliferation after BDL (6) and in this loop VEGF is an important player. Evidently, fall of VEGF after HAL compromises the global function of the proliferative machinery, including the function of agents acting by camprelated pathways. Thus, the decrease of camp levels found in HAL+BDL rat

20 Page 20 of 45 Gaudio et al. 20 cholangiocytes is consistent with impaired proliferation and increased apoptosis. While increased basal and secretin-stimulated cholangiocyte secretory activity is coupled with enhanced ductal hyperplasia (4, 5, 8, 10, 38), impaired ductal secretion is associated with conditions causing reduction of cholangiocyte proliferation (1, 25, 36, 39). In agreement with this study, HAL-induced impairment of proliferation is associated with decreased basal and secretin-induced ductal secretion, all these effects being prevented by VEGF. After liver transplantation, ischemic lesions of bile ducts may occur presumably by surgical lesions of hepatic artery (13, 15, 47). In cholangiopathies, mainly primary sclerosing cholangitis, lesion of hepatic artery and its branches play a causal role in bile duct damage and in the related ductal cholestasis (13). Our study gives the pathophysiological basis for these cholestatic conditions since reduction of the blood supply through the hepatic artery impairs the proliferative and the repairing capacities of damaged ducts. The beneficial effects of VEGF may provide the experimental background for using this growth factor in the management of cholangiopathies. REFERENCES 1. Alpini G, Baiocchi L, Glaser S, Ueno Y, Marzioni M, Francis H, Phinizy JL, Angelico M, and LeSage G. Ursodeoxycholate and tauroursodeoxycholate inhibit cholangiocyte growth and secretion of BDL rats through activation of PKC alpha. Hepatology 35: , Alpini G, Garrick RA, Jones MJ, Nunes R, and Tavoloni N. Water and nonelectrolyte permeability of isolated rat hepatocytes. Am J Physiol Cell Physiol 251: C872-C882, 1986.

21 Page 21 of 45 Gaudio et al Alpini G, Glaser S, Alvaro D, Ueno Y, Marzioni M, Francis H, Baiocchi L, Stati T, Barbaro B, Phinizy JL, Mauldin J, and LeSage G. Bile acid depletion and repletion regulate cholangiocyte growth and secretion by a phosphatidylinositol 3-kinasedependent pathway in rats. Gastroenterology 123: , Alpini G, Glaser S, Ueno Y, Pham L, Podila PV, Caligiuri A, LeSage G, and LaRusso NF. Heterogeneity of the proliferative capacity of rat cholangiocytes after bile duct ligation. Am J Physiol Gastrointest Liver Physiol 274: G767-G775, Alpini G, Lenzi R, Sarkozi L, and Tavoloni N. Biliary physiology in rats with bile ductular cell hyperplasia. Evidence for a secretory function of proliferated bile ductules. J Clin Invest 81: , Alpini G, Prall RT, and LaRusso NF. The pathobiology of biliary epithelia. In: The Liver; Biology & Pathobiology, 4 th Ed, edited by Arias IM, Boyer JL, Chisari FV, Fausto N, Jakoby W, Schachter D, and Shafritz DA. Philadelphia, PA: Lippincott Williams & Wilkins, 2001, p Alpini G, Roberts S, Kuntz SM, Ueno Y, Gubba S, Podila PV, LeSage G, and LaRusso NF. Morphological, molecular, and functional heterogeneity of cholangiocytes from normal rat liver. Gastroenterology 110: , Alpini G, Ueno Y, Glaser S, Marzioni M, Phinizy JL, Francis H, and LeSage G. Bile acid feeding increased proliferative activity and apical bile acid transporter expression in both small and large rat cholangiocytes. Hepatology 34: , Alpini G, Ulrich C, Roberts S, Phillips JO, Ueno Y, Podila PV, Colegio O, LeSage G, Miller LJ, and LaRusso NF. Molecular and functional heterogeneity of

22 Page 22 of 45 Gaudio et al. 22 cholangiocytes from rat liver after bile duct ligation. Am J Physiol Gastrointest Liver Physiol 272: G , Alpini G, Ulrich CD, 2nd, Phillips JO, Pham LD, Miller LJ, and LaRusso NF. Upregulation of secretin receptor gene expression in rat cholangiocytes after bile duct ligation. Am J Physiol Gastrointest Liver Physiol 266: G922-G928, Alvaro D, Cho WKC, Mennone A, and Boyer JL. Effect of secretin on intracellular ph regulation in isolated rat bile duct epithelial cells. J Clin Invest 92: , Assy N, Paizi M, Gaitini D, Baruch Y, and Spira G. Clinical implication of VEGF serum levels in cirrhotic patients with or without portal hypertension. World J Gastroenterol 5: , Batts KP. Ischemic cholangitis. Mayo Clin Proc 73: , Beaussier M, Wendum D, Fouassier L, Rey C, Barbu V, Lasnier E, Lienhart A, Scoazec JY, Rosmorduc O, and Housset C. Adaptative bile duct proliferative response in experimental bile duct ischemia. J Hepatol 42: , Ben-Ari Z, Pappo O, and Mor E. Intrahepatic cholestasis after liver transplantation. Liver Transpl 9: , Burczynski FJ, Luxon BA, and Weisiger RA. Intrahepatic blood flow distribution in the perfused rat liver: effect of hepatic artery perfusion. Am J Physiol Gastrointest Liver Physiol 271: G561-G567, Castaing D, Houssin D, and Bismuth H. Anatomy of the liver and portal system. In: Hepatic and portal surgery in the rat, edited by Castaing D, Houssin D, Bismuth H. Paris, France: Masson, 1980, p

23 Page 23 of 45 Gaudio et al Davis B and Madri J. Type I and type III procollagen peptides during hepatic fibrinogenesis. Am J Pathol 126: , Demetris AJ. Participation of cytokines and growth factors in biliary cell proliferation and mito-inhibition during ductular reaction. In: The Pathophysiology of Biliary Epithelia, edited by Alpini G, Alvaro D, Marzioni M, LeSage G, LaRusso N. Georgetown, TX: Landes Bioscience, 2004, p Francis H, Glaser S, Ueno Y, LeSage G, Marucci L, Benedetti A, Taffetani S, Marzioni M, Alvaro D, Venter J, Reichenbach R, Fava G, Phinizy JL, and Alpini G. camp stimulates the secretory and proliferative capacity of the rat intrahepatic biliary epithelium through changes in the PKA/Src/MEK/ERK1/2 pathway. J Hepatol 41: , Gaudio E, Barbaro B, Alvaro D, Glaser S, Francis H, Ueno Y, Meininger C, Franchitto A, Onori P, Marzioni M, Taffetani S, Fava G, Stoica G, Venter J, Reichenbach R, De Morrow S, Summers R, and Alpini G. Vascular endothelial growth factor stimulates rat cholangiocyte proliferation via an autocrine mechanism. Gastroenterology: In press, Gaudio E, Onori P, and Pannarale L. Microcirculation of the extrahepatic biliary tree: a scanning electron microscopy study of corrosion casts. J Anat 182: 37-44, Gaudio E, Onori P, Pannarale L, and Alvaro D. Hepatic microcirculation and peribiliary plexus in experimental biliary cirrhosis: a morphological study. Gastroenterology 111: , 1996.

24 Page 24 of 45 Gaudio et al Gaudio E, Pannarale L, Ripani M, Onori P, and Riggio O. The hepatic microcirculation in experimental cirrhosis. A Scanning Electron Microscopy study of microcorrosion casts. Scanning Microscopy 5: , Glaser S, Benedetti A, Marucci L, Alvaro D, Baiocchi L, Kanno N, Caligiuri A, Phinizy JL, Chowdhury U, Papa E, LeSage G, and Alpini G. Gastrin inhibits cholangiocyte growth in bile duct-ligated rats by interaction with cholecystokinin- B/Gastrin receptors via D-myo-inositol 1,4,5-triphosphate-, Ca( 2+ )-, and protein kinase C alpha-dependent mechanisms. Hepatology 32: 17-25, Glaser S, Rodgers R, Phinizy JL, Robertson WE, Lasater J, Caligiuri A, Tretjak Z, LeSage G, and Alpini G. Gastrin inhibits secretin-induced ductal secretion by interaction with specific receptors on rat cholangiocytes. Am J Physiol Gastrointest Liver Physiol 273: G1061-G1070, Grappone C, Pinzani M, Parola M, Pellegrini G, Caligiuri A, DeFranco R, Marra F, Herbst H, Alpini G, and Milani S. Expression of platelet-derived growth factor in newly formed cholangiocytes during experimental biliary fibrosis in rats. J Hepatol 31: , Ishii M, Vroman B, and LaRusso NF. Isolation and morphological characterization of bile duct epithelial cells from normal rat liver. Gastroenterology 97: , Ishikawa M, Fjii M, Iuchi M, Miyauchi T, and Tashiro S. Effect of intrahepatic omental implantation on angiogenesis in rat liver with hepatic artery ligation. Clin Exp Med 1: 27 33, 2001.

25 Page 25 of 45 Gaudio et al Itai Y and Matsui O. Blood flow and liver imaging. Radiology 202: , Kanno N, LeSage G, Glaser S, and Alpini G. Regulation of cholangiocyte bicarbonate secretion. Am J Physiol Gastrointest Liver Physiol 281: G , Kato A, Gores GJ, and LaRusso NF. Secretin stimulates exocytosis in isolated bile duct epithelial cells by a Cyclic AMP-mediated mechanism. J Biol Chem 267: , Komalavilas P and Lincoln TM. Phosphorylation of the inositol 1,4,5- trisphosphate receptor. Cyclic GMP-dependent protein kinase mediates camp and cgmp dependent phosphorylation in the intact rat aorta. J Biol Chem 271: , Lametschwandtner A, Lametschwandtner U, and Weiger T. Scanning electron microscopy of vascular corrosion casts. Technique and applications: updated review. Scanning Mic 4: , LeCouter J, Moritz DR, Li B, Phillips GL, Liang XH, Gerber H-P, Hillan KJ, and Ferrara N. Angiogenesis-independent endothelial protection of liver: role of VEGFR-1. Science 299: , LeSage G, Alvaro D, Benedetti A, Glaser S, Marucci L, Baiocchi L, Eisel W, Caligiuri A, Phinizy JL, Rodgers R, Francis H, and Alpini G. Cholinergic system modulates growth, apoptosis, and secretion of cholangiocytes from bile duct-ligated rats. Gastroenterology 117: , LeSage G, Glaser S, and Alpini G. Regulation of cholangiocyte proliferation. Liver 21: 73-80, 2001.

26 Page 26 of 45 Gaudio et al LeSage G, Glaser S, Gubba S, Robertson WE, Phinizy JL, Lasater J, Rodgers R, and Alpini G. Regrowth of the rat biliary tree after 70% partial hepatectomy is coupled to increased secretin-induced ductal secretion. Gastroenterology 111: , LeSage G, Glaser S, Marucci L, Benedetti A, Phinizy JL, Rodgers R, Caligiuri A, Papa E, Tretjak Z, Jezequel AM, Holcomb LA, and Alpini G. Acute carbon tetrachloride feeding induces damage of large but not small cholangiocytes from BDL rat liver. Am J Physiol Gastrointest Liver Physiol 276: G1289-G1301, LeSage G, Glaser S, Ueno Y, Alvaro D, Baiocchi L, Kanno N, Phinizy JL, Francis H, and Alpini G. Regression of cholangiocyte proliferation after cessation of ANIT feeding is coupled with increased apoptosis. Am J Physiol Gastrointest Liver Physiol 281: G182-G190, Marzioni M, LeSage G, Glaser S, Patel T, Marienfeld C, Ueno Y, Francis H, Alvaro D, Phinizy JL, Tadlock L, Benedetti A, Marucci L, Baiocchi L, and Alpini G. Taurocholate prevents the loss of intrahepatic bile ducts due to vagotomy in bile duct ligated rats. Am J Physiol Gastrointest Liver Physiol 284: G837-G852, Nagy JA, Dvorak AM, and Dvorak HF. VEGF-A(164/165) and PlGF: roles in angiogenesis and arteriogenesis. Trends Cardiovasc Med 13: , Nathanson MH and Boyer JL. Mechanisms and regulation of bile secretion. Hepatology 14: , Pinzani M. Cholestasis and fibrogenesis. In: The Pathophysiology of Biliary Epithelia, edited by Alpini G, Alvaro D, Marzioni M, LeSage G, LaRusso N. Georgetown, TX: Landes Bioscience, 2004, p

27 Page 27 of 45 Gaudio et al Prieto J, Garcia N, Marti-Climent JM, Penuelas I, Richter JA, and Medina JF. Assessment of biliary bicarbonate secretion in humans by positron emission tomography. Gastroenterology 117: , Rutenburg AM, Kim H, Fischbein JW, Hanker JS, Wasserkrug HL, and Seligman AM. Histochemical and ultrastructural demonstration of -glutamyl transpeptidase activity. J Histochem Cytochem 17: , Siegel EG, Schmidt WE, and Folsch UR. Severe ischemic-type biliary strictures due to hepatic artery occlusion seven years after liver transplantation--a rare cause of late cholestatic graft failure. Z Gastroenterol 36: , Soares AF, Castro e Silva Junior O, Ceneviva R, Roselino JE, and Zucoloto S. Biochemical and morphological changes in the liver after hepatic artery ligation in the presence or absence of extrahepatic cholestasis. Int J Exp Pathol 74: , Taniguchi E, Sakisaka S, Mattsuo K, Tanikawa K, and Sata M. Expression and role of vascular endothelial growth factor in liver regeneration after partial hepatectomy in rats. J Histochem Cytochem 49: , Tavoloni N and Schaffner F. The intrahepatic biliary epithelium in the guinea pig: is hepatic artery blood flow essential in maintaining its function and structure? Hepatology 55: , Yang L, Faris RA, and Hixson DC. Long-term culture and characteristics of normal rat liver bile duct epithelial cells. Gastroenterology 104: , Yin R, Feng J, and Yao Z. Dynamic changes of serum vascular endothelial growth factor levels in a rat myocardial infarction model. Chin Med Sci J 15: , 2000.

28 Page 28 of 45 Gaudio et al. 28 LEGENDS Figure 1 [a] Light microscopy of liver sections (stained with H&E) from 1 week BDL rats, and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% BSA for 1 week. There were no significant differences in the amount of lobular damage in liver sections from 1 week BDL rats, and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a for 1 week (for quantitative data see Table 1). In BDL + HAL rats, the degree of necrosis and portal inflammation decreases compared to BDL rats (for quantitative data see Table 1). Following the administration of r-vegf-a to BDL + HAL rats, the degree of necrosis and portal inflammation was similar than that of the BDL rat (for quantitative data see Table 1). Orig. magn., x80. [b] Light microscopy of liver sections (Masson s stain) from 1 week BDL rats, and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% BSA for 1 week. The % volume fraction of peri-ductular fibrosis observed in BDL rats was reduced in BDL+HAL rats compared to the value of BDL rats. Following the administration of r- VEGF-A to BDL + HAL rats, the % volume fraction was similar than that of the BDL rat (for quantitative data see Table 1). Orig. magn., x20. Figure 2 Scanning electron microscopy vascular corrosion cast from rats (that immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% BSA for 1 week. Observe in BDL+HAL rats

29 Page 29 of 45 Gaudio et al. 29 the presence of sinusoidal network (S) and portal vein (P) and the absence of peribiliary plexus (orig. magn. 50X); differently, in BDL + HAL+ r-vegf-a rats the peribiliary plexus (PBP) that run together with portal tract was observed (orig. magn. 100X). Figure 3 Immunohistochemistry for [a] VEGF-A and [b] VEGFR-2 and VEGFR-3 in liver sections from 1 week BDL rats and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% BSA for 1 week. [a] Bile ducts from BDL rats express VEGF-A. The administration of r-vegf-a to BDL + HAL rats prevented the decrease in cholangiocyte VEGF-A expression observed in BDL + HAL rats. Orig. magn. 40X. [b] Following the administration of r-vegf-a to BDL+HAL rats, the expression of VEGFR-2 and VEGFR-3 was similar to that of the BDL rat alone. VEGFR-1 was not expressed by cholangiocytes. Orig. magn. 40X. For statistical evaluation of the number of VEGF-A-, VEGFR-2 and VEGFR-3-positive cholangiocytes see Table 2. Figure 4 VEGF secretion in primary cultures (6 hours) of [top] cholangiocytes or [bottom] hepatocytes from normal and 1 week BDL rats, and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a (2.5 nmol/kg/hour with 0.2% BSA) for 1 week. Primary cultures (i.e. 6 hours) of normal rat hepatocytes and cholangiocytes secrete VEGF. Following BDL, there was an increase in VEGF secretion in primary cultures of cholangiocytes. VEGF secretion significantly decreased in BDL hepatocytes compared to normal hepatocytes. In both cholangiocytes and hepatocytes from BDL+HAL rats, there was a marked

30 Page 30 of 45 Gaudio et al. 30 decrease of VEGF secretion compared with cholangiocytes and hepatocytes from control 1-week BDL rats. Administration of r-vegf-a prevented the decrease of cholangiocyte VEGF secretion induced by HAL in cholangiocytes and hepatocytes. Data are mean ± SE of 7 (cholangiocytes) and 4 (hepatocytes) experiments. [top] *p < 0.05 vs. VEGF secretion of cholangiocytes from BDL rats. #p < 0.05 vs. VEGF secretion of normal cholangiocytes. [bottom] *p < 0.05 vs. VEGF secretion of hepatocytes from BDL rats. #p < 0.05 vs. VEGF secretion of normal hepatocytes. Figure 5 Measurement of cholangiocyte apoptosis in liver sections by TUNEL analysis from 1 week BDL, and rats that [immediately after BDL or BDI + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA, or r-vegf-a (2.5 nmol/kg/hour with 0.2% BSA) for 1 week. TUNEL analysis showed only a few apoptotic bodies in the liver sections of 1 week BDL rats. The number of cholangiocytes undergoing apoptosis increased in liver sections from BDL + HAL compared to 1 week BDL rats. Chronic administration of r-vegf-a prevented the increase in cholangiocyte apoptosis (by TUNEL analysis) induced by HAL. Data are mean ± SE of 8 values. *p < 0.05 vs. cholangiocyte apoptosis of all the other groups. Figure 6 Measurement of cholangiocyte proliferation by quantitative measurement of the % of CK-19-positive ducts in liver sections from 1 week BDL rats and rats that [immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a (2.5 nmol/kg/hour with 0.2% BSA) for 1 week. Following HAL, the % of CK-19- positive ducts decreased in liver sections compared to liver

31 Page 31 of 45 Gaudio et al. 31 sections from 1 week BDL rats. Chronic administration of r-vegf-a prevented the inhibitory effect of HAL on the % of CK-19-positive ducts, a value that was not statistically different from that of 1 week BDL rats. For statistical evaluation of the % of CK-19-positive ducts see Table 4. Figure 7 Measurement of basal and secretin-stimulated camp levels in purified cholangiocytes from 1 week BDL, and rats that [immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA, or r-vegf-a (2.5 nmol/kg/hour with 0.2% BSA) for 1 week. Secretin increased intracellular camp levels of cholangiocytes from 1-week BDL rats. HAL significantly reduced basal cholangiocyte camp levels and inhibited secretin-stimulated camp levels of cholangiocytes compared to cholangiocytes from 1 week BDL rats. Chronic administration of r-vegf-a prevented the inhibition of basal and secretin-stimulated camp levels induced by HAL. Data are mean ± SE of 7 experiments. *p < 0.05 vs. its corresponding basal value. #p < 0.05 vs. basal camp levels of cholangiocytes from BDL rats and BDL + HAL rats treated with r-vegf-a for 1 week.

32 Page 32 of 45 Table 1 H&E staining of liver sections from 1 week BDL rats and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% BSA for 1 week. Treatment Portal Inflammation Necrosis Lobular Damage Peri-ductular Fibrosis (% volume fraction) BDL 1 week 1.7 ± ± ± ± 0.8 BDL + HAL + 0.2% BSA 1 week 1.2 ± 0.1 * 0.7 ± 0.1 * 1.2 ± ± 0.7 * BDL + HAL + r-vegf-a 1 week 1.4 ± ± ± ± 1.0 BDL = bile duct ligated; HAL = hepatic artery ligation. There were no significant differences in the extent of lobular damage in liver sections (4 µm thick) from the different groups of animals. In BDL + HAL rats, the degree of necrosis, portal inflammation and peri-ductular fibrosis decreased compared to BDL rats. Following the administration of r-vegf-a to BDL + HAL rats, the degree of necrosis, portal inflammation and peri-ductular fibrosis were similar than that of the BDL rat. Data are mean ± SEM of 17 values from the evaluation of 3 slides for each group. * p<0.05 vs. all the other groups.

33 Page 33 of 45 Table 2 Immunohistochemical evaluation of cholangiocyte VEGF-A, VEGFR-1, VEGFR-2 and VEGFR-3 expression in liver sections from 1 week BDL rats, and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% BSA for 1 week. Treatment VEGF-A expression (% positive cholangiocytes) VEGFR-1 expression (% positive cholangiocytes) VEGFR-2 expression (% positive cholangiocytes) VEGFR-3 expression (% positive cholangiocytes) BDL 1 week (n = 6) 78.3 ± 3.9 Negative 46.2 ± ± 2.9 BDL + HAL + 0.2% BSA 1 week (n = 6) 23.1 ± 2.7* Negative 25.6 ± 1.8* 10.5 ± 2.1* BDL + HAL + r-vegf-a 1 week (n = 6) 74.9 ± 2.9 # Negative 48.1 ± ± 3.2 # BDL = bile duct ligation; HAL = hepatic artery ligation. The immunohistochemical cholangiocyte expression of VEGF-A, VEGFR-2 and VEGFR-3 decreased in liver sections from BDL + HAL rats compared to BDL rats and BDL + HAL rats treated with r-vegf-a. Chronic administration of r-vegf-a prevented HAL-induced loss of cholangiocyte expression of VEGF-A, VEGFR-2 and VEGFR-3, values that were similar or higher ( # p<0.05) to that of liver sections from 1 week BDL rats. Cholangiocytes did not express VEGFR-1. *p<0.05 vs. all the other groups. # p<0.05 vs. corresponding value of 1 week BDL rat liver sections.

34 Page 34 of 45 Table 3 Measurement of hepatocyte VEGF-A protein expression in liver sections from normal and 1 week BDL rats and rats that (immediately after BDL + HAL) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a with 0.2% for 1 week. Treatments VEGF-A (% positive hepatocytes) Normal (n = 6) 20.3 ± 1.7 BDL 1 week (n = 6) 24.5 ± 2.1 BDL + HAL + 0.2% BSA 1 week (n = 6) 13.2 ± 1.6* BDL + HAL + r-vegf-a 1 week (n = 6) 28.5 ± 2.5 BSA = bovine serum albumin; HAL = hepatic artery ligation. Hepatocyte VEGF-A protein expression did not increase significantly in BDL liver sections compared to normal sections, decreased after HAL and returned to values similar to those of BDL rats in BDL + HAL rats treated with r-vegf-a. *p<0.05 vs. corresponding value of liver sections from 1 week BDL rats and BDL + HAL rats.

35 Page 35 of 45 Table 4 Measurement of cholangiocyte proliferation by quantitative measurement of the number of PCNA-positive cholangiocytes and the % of CK-19- and -GT-positive ducts in liver sections from 1 week BDL, and rats that (immediately after BDL + HAL) were treated with 0.2% BSA or r- VEGF-A with 0.2% BSA for 1 week. Treatment PCNA-positive cholangiocytes (per portal tract) (n = 12) % CK-19-positive ducts (per portal tract) (n = 10) % -GT-positive ducts (per portal tract) (n = 14) BDL 1 week 16.4 ± ± ± 0.5 BDL + HAL + 0.2% BSA 1 week 4.6 ± 05* 3.7 ± 0.4* 1.1 ± 0.1* BDL + HAL + r-vegf-a 1 week 11.7 ± ± ± 0.5 Following HAL, the number of PCNA-positive cholangiocytes and the % of CK-19 and -GT-positive ducts decreased in liver sections compared to liver sections from BDL rats and BDL + HAL rats treated with r-vegf-a. Chronic administration of r-vegf-a prevented the inhibitory effect of HAL on the number of the number of PCNA-positive cholangiocytes and the % of CK-19 and -GT-positive ducts, values that were not statistically different from those of 1 week BDL rats. Data are mean ± SE of 8 values. *p < 0.05 vs. all other groups.

36 Page 36 of 45 Table 5 Measurement of basal and secretin-stimulated bile flow, bicarbonate concentration and secretion in 1-week BDI rats and rats that (immediately after BDI + hepatic artery ligation) were treated by IP implanted Alzet osmotic minipumps with 0.2% BSA or r-vegf-a for 1 week. Bile Flow Bicarbonate Concentration Bicarbonate Secretion Treatment Basal (µl / min / Kg BW) Secretin (µl / min / Kg BW) Basal (meq / Liter) Secretin (meq / Liter) Basal (µeq / min / Kg BW) Secretin (µeq / min / Kg BW) BDI (n = 8) 98.6 ± ± 6.5 # 38.1 ± ± 3.4 # 3.7 ± ± 0.7 # BDI + HAL + 0.2% BSA (n = 8) 96.6 ± ± 6.7 ns 39.5 ± ± 1.5 ns 3.8 ± ± 0.4 ns BDI + HAL + r-vegf-a (n = 6) 82.8 ± ± 7.5 # ± ± 4.1 # 3.6 ± ± 0.3 # BDI bile duct incannulation; HAL = hepatic artery ligation. When steady spontaneous bile flow was reached (60-70 minutes from the infusion of Krebs Ringer Henseleit (KRH), rats were infused for 30 minutes with secretin followed by a final infusion of KRH for 30 minutes. After the rats were surgically prepared for bile flow experiments, bile was collected every 10 minutes in pre-weighed tubes and used for determining bicarbonate concentration. Data are mean ± SEM. # p < 0.05 vs. their corresponding basal values of bile flow, bicarbonate concentration or bicarbonate secretion. ns vs. corresponding basal value of bile flow, bicarbonate concentration or bicarbonate secretion of BDI rats. Differences between groups were analyzed by the Student unpaired t test when two groups were analyzed and analysis of variance (ANOVA) when more than two groups were analyzed.

37 Page 37 of 45 Figure 1 a BDL BDL + HAL BDL + HAL + r-vegf-a

38 Page 38 of 45 Figure 1 b BDL BDL+HAL BDL+HAL+r-VEGF-A

39 Page 39 of 45 Figure 2 BDL + HAL BDL + HAL + r-vegf-a

40 Page 40 of 45 Figure 3 a BDL BDL+HAL BDL+HAL+r-VEGF-A

41 Page 41 of 45

42 140 Figure 4 Page 42 of 45 Cholangiocyte VEGF secretion (ng / 1 x 10 6 cells) # * * 0 Normal BDL BDL + HAL BDL + HAL + r-vegf-a Normal BDL BDL + HAL BDL + HAL + r-vegf-a Hepatocyte VEGF secretion (ng / 1 x 10 6 cells) #

43 Page 43 of 45

44 Page 44 of 45