- Email: [email protected]
Effect of fibrin glue occlusion of the hepatobiliary tract on thioacetamide-induced liver failure
The American Journal of Surgery 182 (2001) 58 – 63
Effect of fibrin glue occlusion of the hepatobiliary tract on thioacetamide-induced liver failure Thomas C. Schmandra, M.D., Holger Bauer, M.D., Henrik Petrowsky, M.D., Gu¨nther Herrmann, M.D., Albrecht Encke, M.D., Ernst Hanisch, M.D.* Department of Surgery, Johann Wolfgang Goethe-University Hospital, Theodor Stern Kai 7, 60590 Frankfurt am Main, Germany Manuscript received September 5, 2000; revised manuscript February 19, 2001
Abstract Background: Expression and activation of hepatocyte growth factor (HGF) is stimulated by a complex system of interacting proteins, with thrombin playing an initial role in this process. The impact of temporary occlusion of the hepatobiliary tract with fibrin glue (major component thrombin) on the HGF system in acute and chronic liver damage in a rat model was investigated. Methods: Chronic liver damage was induced in 40 rats by daily intraperitoneal application of thioacetamide (100 mg/kg) for 14 days. After 7 days half of them received an injection of 0.2 mL fibrin glue into the hepatobiliary system. Daily intraperitoneal administration of thioacetamide continued for 7 consecutive days. The rats were then sacrificed for blood and tissue analysis. Acute liver failure was induced in 12 rats by intraperitoneal administration of a lethal dose of thioacetamide (500 mg/kg per day for 3 days) after an injection with 0.2 mL fibrin glue into their hepatobiliary tract. Survival rates and histological outcome were investigated and compared with control animals. Results: Fibrin glue occluded rats showed significantly lower liver enzyme activities and serum levels of bilirubin, creatinine and urea nitrogen. Immunohistochemistry revealed a significant increase in c-met-, HGF␣- and especially HGF-positive cells. Rats subjected to a lethal dose of thioacetamide survived when fibrin glue was applied 24 hours prior to the toxic challenge. These animals showed normal liver structure and no clinical abnormalities. Conclusion: Fibrin glue occlusion of the hepatobiliary tract induces therapeutic and prophylactic effects on chronic and acute liver failure by stimulating the HGF system. Therefore, fibrin glue occlusion might be useful in treating toxic liver failure. © 2001 Excerpta Medica, Inc. All rights reserved. Keywords: Fibrin glue occlusion; Hepatocyte growth factor; Liver regeneration; Liver failure
Hepatocyte growth factor (HGF), a multifunctional cytokine originally identified as a potent mitogen agent for rat hepatocytes [1–3], is the most potent stimulus for hepatocyte growth and DNA synthesis. This process is essential for hepatocyte proliferation during liver regeneration. Following the onset of liver disease or after liver resection, a marked increase in circulating HGF and an upregulation of HGF mRNA expression in the liver occurs [4 –9]. HGF strongly stimulates DNA synthesis of hepatocytes in the liver damaged by surgery or toxic agents, and accelerates liver regeneration, whether applied both systemically or locally [10 –13]. The response to HGF is mediated by the
* Corresponding author. Tel.: ⫹49-69-6301-5142; fax: ⫹49-69-63017452. E-mail address: [email protected]
high-affinity signalling receptor for HGF, which is the protein product of the proto-oncogene c-met [14]. HGF is secreted as an inactive single-chain precursor and remains in this form until it is converted into an active heterodimeric molecule in response to hepatic injury [15]. Activation is exclusively initialized in the injured liver by limited proteolysis at a single site of the HGF precursor. The protease responsible for HGF conversion has been identified as HGF activator (HGFA) [16,17]. HGFA itself is first synthesized as an inactive precursor and then activated by proteolytic processing in response to tissue injury. Thrombin seems to play an initial role in this process, catalyzing the conversion of the precursor to active HGFA [18]. Assuming that activated HGF is the essential factor in regulating liver regeneration after surgical resection or functional damage caused by hepatic disease or toxic influence, activating the HGFA zymogen by thrombin might be of
0002-9610/01/$ – see front matter © 2001 Excerpta Medica, Inc. All rights reserved. PII: S 0 0 0 2 - 9 6 1 0 ( 0 1 ) 0 0 6 5 9 - 6
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
59
Materials and methods
these animals laparotomy was performed 24 hours prior to the first injection of thioacetamide. During laparotomy one half received an injection of 0.2 mL fibrin glue into the hepatobiliary system (n ⫽ 12, group C), the other half underwent mere liver inspection (n ⫽ 12, group D). In the control group (n ⫽ 10) no laparotomy was performed prior to induction of acute liver failure. After the first application of thioacetamide, survival rates, histological outcome, and blood status were recorded for 10 days.
Animal treatments
Histological procedures
All animals (male Sprague-Dawley rats, weight 280 ⫾ 20 g) were maintained under standardized conditions with free access to commercial rodent diet and tap water ad libitum throughout the experiments. The animals received humane care in compliance with the institution’s guidelines and were subjected to a 12-hour day-night cycle. In experiment 1, after induction of general anesthesia (nembutal, 40 mg/kg, intraperitoneal) and midline incision, the common bile duct was microsurgically isolated and incised (1 cm below the lowest tributary) in 15 rats. The bile duct was cannulated with a silicone tube (inner diameter 0.25 mm) and 0.2 mL fibrin glue (32.5 to 57.5 mg fibrinogen, 500 IU aprotinin, 300 IU thrombin, 40 mmol/L calcium chloride in 0.5 mL; Behring, Marburg, Germany) was injected into the proximal biliary tract. The silicone tube was removed and the bile duct incision closed (polyethylene 8-0). All animals survived surgery. To evaluate the reabsorption rate of fibrin glue applied into the hepatobiliary tract, the animals were sacrificed after 6, 12, 18, 24, and 36 hours for liver tissue analysis by fibrin staining. In livers explanted after 24 and 36 hours (n ⫽ 6) immunohistochemical detection of c-met, HGF␣ and HGF was performed. For experiment 2, chronic liver damage was induced by daily intraperitoneal application of a sublethal dose of thioacetamide (100 mg/kg) for 14 consecutive days in 40 rats. After 7 days laparotomy was performed in all animals after induction of general anesthesia (nembutal, 40 mg/kg, intraperitoneal). In group A (n ⫽ 20) fibrin glue (0.2 mL) was injected into the hepatobiliary tract as described above. Animals in group B (n ⫽ 20) underwent mere liver inspection during laparotomy. Following surgery, daily intraperitoneal administration of thioacetamide (100 mg/kg) continued in all animals for 7 consecutive days. On day 15 all animals were sacrificed for tissue and blood analysis. The findings pertaining to groups A and B were compared with those in control animals (group L [n ⫽ 10] to control laparotomy: no thioacetamide application/laparotomy; and group V [n ⫽ 10] to control the vehicle of administering thioacetamide: intraperitoneal administration of sodium chloride 0.9%/no thioacetamide/no laparotomy). For experiment 3, acute liver failure was induced by daily intraperitoneal application of thioacetamide (500 mg/ kg) for 3 consecutive days (lethal dose) in 34 rats. In 24 of
For immunohistochemical detection of activated HGF (consisting of an ␣- and -chain) and the c-met receptor, freshly frozen liver tissue was serially sliced (5 m, 25 slices per animal) in a cryostat, immersed in ice-cold acetone for 10 minutes and dried at ⫺75°C. The slices were then incubated with primary antibodies against recombinant rat HGF␣ (1:200), HGF (1:100), and c-met-peptide (1: 100; Santa Cruz Biotechnology, Santa Cruz, New Mexico). Following administration of secondary antibodies (biotinylated immunoglobulin G; Dianova, Hamburg, Germany), the slices were incubated with alkaline phosphatase-conjugated streptoavidine. The slices were then immersed in a solution with 5% new fuchsin in 0.2 mol/L Tris-HCl (pH 8.3 to 8.5). Nuclei counterstaining was performed with hematoxylin solution. A possible endogenous biotin immunoreactivity was excluded by using a biotin blocking kit before incubating the slides with primary antibodies. Immunohistochemical staining was evaluated with a staining score ranging from 0 (no staining) to 3 (strongly positive reaction). Both primary antibodies and animal test groups were not known to the evaluators. In experiments 1 and 3 excised liver tissue was fixed in 10% neutralized formaldehyde, embedded in paraffin, and stained with hematoxylin and eosin.
potential benefit in treating subjects with liver failure. Therefore, we applied fibrin glue (major component thrombin) into the hepatobiliary tract to investigate the impact of a temporary fibrin glue occlusion on HGF, the proto-oncogene c-met, and hepatic function in induced chronic and acute liver failure in a rat model.
Biochemical measurements The liver-specific cytosolic enzyme activities of aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyltranspeptidase (␥GT), and alkaline phosphatase (AP); and the serum level of bilirubin, creatinine, and urea nitrogen were determined using an automated analyzer system (Hitachi Co., Tokyo, Japan). Statistical analysis Student’s t test (biochemical measurements), the MantelHaenszel test, and the Mantel-Haenszel-Zimmermann test (quantification of immunohistochemical staining) [19,20] were used to determine statistical differences between the groups. Values were considered significantly different if P was less than 0.05.
60
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
Table 1 Serum activity of liver enzymes and serum levels of bilirubin, creatinine, and urea nitrogen in induced chronic liver damage in group A and group B
Group A: thioacetamide ⫹ fibrin glue Group B: thioacetamide
AST (U/L)
ALT (U/L)
␥GT (U/L)
AP (U/L)
Bilirubin (mg/dL)
Creatinine (mg/dL)
Urea (mg/dL)
32.5 ⫾ 3.6* 40.1 ⫾ 6.8
17.4 ⫾ 5.8* 33 ⫾ 9
3 ⫾ 0.9 3.6 ⫾ 0.9
237 ⫾ 42* 335 ⫾ 49
0.41 ⫾ 0.2* 0.73 ⫾ 0.17
0.46 ⫾ 0.08* 0.63 ⫾ 0.1
37.4 ⫾ 4.8* 50.3 ⫾ 9.6
Group A was given daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days and fibrin glue occlusion of the hepatobiliary tract on day 7; n ⫽ 20 and group B was given daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days; n ⫽ 20. Values are means ⫾ SD; P ⬍0.05 was deemed significant. * P ⬍0.05. AST ⫽ aspartate aminotransferase; ALT ⫽ alanine aminotransferase; ␥GT ⫽ gamma-glutamyetranspeptidase; AP ⫽ alkaline phosphatase.
Results Reabsorption of fibrin glue applied to the hepatobiliary system (experiment 1) Application of 0.2 mL fibrin glue into the hepatobiliary tract led to temporary occlusion. The plug was for the most part reabsorbed within 24 hours. After 36 hours no fibrin glue could be detected in the hepatobiliary system. In comparison with normal liver tissue, immunohistochemistry in livers explanted 24 or 36 hours after fibrin glue occlusion revealed an increased number of cells positive for c-met, HGF␣, and HGF. Effect of fibrin glue occlusion on chronic liver damage (experiment 2) In group A (fibrin glue occluded), all animals survived thioacetamide administration, laparotomy, and hepatobiliaric application of fibrin glue. After surgery the animals recovered well and their body weight increased to 294 ⫾ 27.6 g (⫹5.8%). Serum enzyme activities of AST, ALT, AP, and the serum levels of bilirubin, creatinine, and urea
nitrogen were significantly lower than in group B (Table 1) but showed no significant differences compared with control animals. Standard histopathology in group A showed abundant inflammatory cells infiltrating the portal areas. Bile ducts and ductules were not extended. Compared with group B and the control groups L and V, immunohistochemistry revealed a significantly higher number of HGF␣-positive (score: 2.73 ⫾ 0.44) and especially HGF-positive cells (score: 1.27 ⫾ 1.06) (Figs. 1, 2). HGF-positive cells were predominantly found around the portal triads, to a lesser degree around the central veins, and diffusely in the liver. C-met–positive cells were abundant with a diffuse distribution pattern throughout the liver. Immunoreactivity of c-met–positive cells was significantly increased (staining score: 2.73 ⫾ 0.44; Fig. 2). In group B (not fibrin glue occluded), animals appeared less vital and agile, compared with animals of group A (fibrin glue occlusion) and the control groups. Mean body weight decreased to 261 ⫾ 17.6 g (⫺6.8%) and all animals developed moderate to severe ascites. Three animals (15%) died before day 15 (2 on day 12, 1 on day 13). Immunohistochemistry showed a great number of HGF␣-positive cells (score: 1.67 ⫾ 1.18), particularly around the liver
Fig. 1. Immunostaining of HGF-positive cells by primary antibody in group A (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days and fibrin glue occlusion of the hepatobiliary tract on day 7; n ⫽ 20) and group B (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days; n ⫽ 20). The sections were immersed in a solution with 5% new fuchsin in 0.2 mol/L Tris-HCl. Counterstaining of the nuclei was performed with hematoxylin solution (immunostained cells: red; nuclei: blue).
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
61
statistically different to the control groups (score: 0). Cmet–positive cells were numerous (score: 2.25 ⫾ 0.59) and found diffusely in the liver. Staining was predominantly localized on the plasma membrane. Compared with the control groups L and V immunoreactivity of c-met–positive cells was not significantly different (Fig. 2). Effect of fibrin glue occlusion on acute liver failure (experiment 3)
Fig. 2. Immunostaining of HGF␣-, HGF-, and c-met–positive cells by primary antibodies in group A (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days and fibrin glue occlusion of the hepatobiliary tract on day 7; n ⫽ 20), group B (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days; n ⫽ 20), and control groups L (laparotomy, no thioacetamide, no fibrin glue; n ⫽ 10) and V (intraperitoneal administration of sodium chloride 0.9%, no laparotomy, no thioacetamide, no fibrin glue; n ⫽ 10). Immunoreactivity is quantified using a special scoring system ranging from 0 (no staining) to 3 (strong positive reaction). The Mantel-Haenszel and the Mantel-HaenszelZimmermann tests were used to determine statistical significance. Values are means ⫾ SD; P ⬍0.05 was deemed significant. agroup A versus group B and control groups; P ⬍0.05. bgroup B versus control groups; P ⬍0.05.
portal triads. Most of these cells had large nuclei and could be histologically identified as nonparenchymal cells. Around the central veins, HGF staining increased diffusely, but the number of stained cells was fewer than those surrounding the portal triads. In comparison with the control groups L (score: 0.67 ⫾ 0.47) and V (score: 0.71 ⫾ 0.59) the animals revealed a significantly higher number of HGF␣-positive cells. Incubation with primary antibodies against the HGF-chain revealed no, or weak, immunohistochemical staining (score: 0.09 ⫾ 0.18), which was not
In group C (fibrin glue occluded prior to thioacetamide application), all animals survived the application of a lethal dose of thioacetamide. During the experiment they appeared to be clinically normal with no differences to the control animals. Blood analysis, which was performed on day 10 of the experiment, revealed no statistical differences to normal animals. Histological analysis showed a normal liver structure and almost no effects of the lethal application of thioacetamide. In group D (not fibrin glue occluded prior to thioacetamide application), all animals died within 48 hours after the application of a lethal dose of thioacetamide. Histological analysis of liver tissue revealed fulminant liver necrosis, predominantly localized around the central veins. No differences from the control animals, which also died within 48 hours, were observed (Fig. 3; Table 2).
Comments HGF plays a significant role in liver regeneration, stimulating DNA synthesis of hepatocytes in toxically or mechanically injured liver tissue or during liver infection. The serum concentration of HGF rises dramatically after liver
Fig. 3. Histology of the liver after induction of acute liver failure in group C (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; fibrin glue occlusion of the hepatobiliary tract 24 hours prior to the first thioacetamide administration; n ⫽ 12) and group D (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; liver inspection 24 hours prior to the first thioacetamide administration; n ⫽ 12). Liver tissue is fixed in 10% neutralized formaldehyde, embedded in paraffin, and stained with hematoxylin and eosin (nuclei: blue; tissue necrosis: red).
62
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
Table 2 Mortality 2, 5, and 10 days after induction of acute liver failure 2 Days Group C: laparotomy ⫹ fibrin glue ⫹ thioacetamide Group D: laparotomy ⫹ thioacetamide Control: thioacetamide
0/12 (0%)
5 Days
10 Days
0/12 (0%)
0/12 (0%)
12/12 (100%) 10/10 (100%)
Group C (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; fibrin glue occlusion of the hepatobiliary tract 24 hours prior to the first thioacetamide administration; n ⫽ 12); group D (liver inspection 24 hours prior to the first thioacetamide application; n ⫽ 12); and control group (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; n ⫽ 10).
surgery [6] and remains elevated for the duration of the regenerative process in the liver. Similar increases in HGF are found when hepatic regenerative activity is induced by drugs, toxic substances, or peroxisome proliferators [21– 23]. The HGF response correlates with the extent of hepatic injury. In patients with acute or chronic liver disease the serum concentration of HGF is elevated. The rise correlates with the severity of the disease and— especially in fulminant liver failure—the HGF level is inversely correlated with the prognosis of the patient [4,7]. However, no detailed studies of the nature of the elevated HGF in serum have been performed. Therefore, it is possible that the circulating HGF may be in the inactive form. In animal models infusion of HGF into the portal vein or the systemic circulation has stimulated regenerative response in the injured liver [13,24]. In our study thioacetamide was applied intraperitoneal to induce chronic and acute liver injury. The chronic application of thioacetamide, which results in chronic liver damage (experiment 2), caused a significant increase in HGF␣positive cells in the liver. Almost no HGF-positive cells were found in these animals, inferring that HGF is in the inactive, single-chain form, and a conversion to the active heterodimeric molecule is needed for an effective response to hepatic injury. Fibrin glue occlusion of the hepatobiliary system induced an effective regenerative response to the thioacetamide-induced chronic liver damage. The animals treated with fibrin glue showed superior liver function with lower liver enzyme activities and serum bilirubin (despite temporary occlusion of the bile tract), creatinine and urea nitrogen levels. No catabolic metabolism was observed, as was apparent in nonoccluded animals. Immunohistochemistry revealed a significant increase in HGF␣- and especially HGF-positive cells (13-fold higher) representing the activated form of HGF. The protease HGFA is responsible for the conversion of HGF into the active form. HGFA itself is synthesized as an inactive molecule, which is activated by proteolytic processing. Thrombin is suggested to be responsible for this local effect, as it has been proved to play the initial role in activating the
HGFA zymogen [18]. Therefore, the protective effect of fibrin glue may be due to activation of the HGFA/HGF process. Since permanent bile duct ligation is a potent stimulus for hepatocellular proliferation [25], the temporary occlusion of the hepatobiliary tract after fibrin glue application may also influence HGF expression. In our experiments fibrin glue was reabsorbed within 36 hours after occlusion and liver tissue analysis showed that bile ducts and ductules were not extended. Serum levels of bilirubin and parameters indicating cholestasis were significantly lower in the fibrin glue occluded animals. This makes occlusion itself unlikely to be the main cause for HGF activation. The response to activated HGF is mediated by the high-affinity signalling receptor for HGF, the protein product of the proto-oncogene c-met [14]. Immunoreactivity of c-met–positive cells was significantly higher in fibrin glue occluded animals, indicating an upregulation of the HGF receptor expression. Upregulation of the HGF receptor is inducible and HGF itself is known to induce expression of its own receptor [26]. Whether the increase in c-met–positive cells is caused by stimulation of the HGF activation process itself, or fibrin glue stimulates the c-met oncogene is not clear. In our study c-met–positive cells were uniformly distributed throughout the liver, whereas HGF-positive cells were predominantly found in the periportal region and—to lesser degree— around the central veins. Possibly, HGF, attached to high affinity sites in the periportal region, is not bound to the HGF receptor. Investigators have partly characterized the mechanisms whereby HGF binds to periportal high affinity sites, but the precise process is not completely clear [27]. Aside from fibrin glue application reducing liver damage, prophylactic fibrin glue occlusion significantly increased the survival rate in acute liver failure. After fibrin glue occlusion the animals withstood a lethal challenge of thioacetamide and showed negligible clinical or histological pathology. This underlines the importance of activated HGF for adequate liver regeneration, since survival after fulminant hepatic failure is determined by the extent of liver injury and the ability of hepatocyte regeneration. In conclusion, local application of fibrin glue into the hepatobiliary tract might have positive effects, useful in treating toxic liver damage or fulminant hepatic failure. References [1] Nakamura T, Nawa K, Ichihara A. Partial purification and charcterization of hepatocyte growth factor from serum of hepatectomized rats. Biochem Biophys Res Commun 1984;122:1450 –9. [2] Russell WE, McGowan JA, Bucher NLR. Partial characterization of hepatocyte growth factor from rat platelets. J Cell Physiol 1984;119: 183–92. [3] Thaler J, Michalopoulos GK. Hepatopoitin A: a partial characterization and trypsin activation of a hepatocyte growth factor. Cancer Res 1985;45:2545–9. [4] Tsobouchi H, Niitani Y, Hirono S, et al. Levels of the human hepatocyte growth factor in serum of patients with various liver diseases
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
determined by an enzyme-linked immunosorbent assay. Hepatology 1991;13:1–5. Shiota G, Okano J, Kawasaki H, et al. Serum hepatocyte growth factor levels in liver diseases: clinical implications. Hepatology 1995; 21:106 –12. Lindroos PM, Zarnegar R, Michalopoulos GK. Hepatocyte growth factor (hepatopoitin A) rapidly increases in plasma before DNA synthesis and liver regeneration stimulation by partial hepatectomy and carbon tetrachloride administration. Hepatology 1991;13:743–9. Tomiya T, Nagoshi S, Fujiwara K. Significance of serum human hepatocyte growth factor levels in patients with hepatic failure. Hepatology 1992;15:1– 4. Kinoshita T, Tashiro K, Nakamura T. Marked increase of HGF mRNA in non-parenchymal liver cells of rats treated with hepatotoxines. Biochem Biophys Res Commun 1989;165:1229 –34. Zarnegar R, DeFrances MC, Kost DP, et al. Expression of hepatocyte growth factor mRNA in regenerative rat liver after partial hepatectomy. Biochem Biophys Res Commun 1991;177:559 – 65. Ishiki Y, Ohnishi H, Muto Y, et al. Direct evidence that hepatocyte growth factor is a hepatotrophic factor for liver regeneration and for potent anti-hepatitis action in vivo. Hepatology 1992;16:1227–35. Fujiwara K, Nagoshi S, Ohno A, et al. Stimulation of liver growth by exogenous human hepatocyte growth factor in normal and partially hepatectomized rats. Hepatology 1993;18:1443–9. Ishii T, Sato M, Sudo K, et al. Hepatocyte growth factor stimulates liver regeneration and elevates blood protein level in normal and partially hepatectomized rats. J Biochem 1995;117:1105–12. Kobayashi Y, Hamanoue M, Ueno S, et al. Induction of hepatocyte growth by intraportal infusion of HGF into beagle dogs. Biochem Biophys Res Commun 1996;220:7–12. Naldini L, Vigna E, Narsimhan RP, et al. Hepatocyte growth factor (HGF) stimulates the tyrosine kinase activity of the receptor encoded by the proto-oncogene c-MET. Oncogene 1991;6:501– 4. Miyazawa K, Shimomoura T, Naka D, Kitamura N. Proteolytic activation of hepatocyte growth factor in response to tissue injury. J Biol Chem 1994;269:8966 –70.
63
[16] Miyazawa K, Shimomura T, Kitamura A, et al. Molecular cloning and sequence analysis of the cDNA for a human serine protease responsible for activation of hepatocyte growth factor. J Biol Chem 1993;268:10024 – 8. [17] Miyazawa K, Shimomura T, Kitamura N. Activation of hepatocyte growth factor in the injured tissues is mediated by hepatocyte growth factor activator. J Biol Chem 1996;271:3615–18. [18] Shimomura T, Kondo J, Ochiai M, et al. Activation of the zymogen of hepatocyte growth factor activator by thrombin. J Biol Chem 1993;268:22927–32. [19] Mantel N. Chi2-tests with one degree of freedom: extensions of the Mantel-Haenszel procedure. JASA 1963;58:690 –700. [20] Zimmermann H. Erweiterung des Mantel-Haenszel-Testes auf Kl⬎2 gruppen. Biomed J 1984;26:223– 4. [21] Lindroos P, Tsai WH, Zarnegar R, Michalopoulos GK. Plasma levels of HGF in rats treated with tumor promoters. Carcinogenesis 1992; 13:139 – 41. [22] Matsumoto K, Nakamura T. Hepatocyte growth factor: molecular structure and implications for a central role in liver regeneration. J Gastroenterol Hepatol 1991;6:509 –19. [23] Matsumoto K, Nakamura T. Hepatocyte growth factor: molecular structure, roles in liver regeneration, and other biological functions. Crit Rev Oncogen 1992;3:27–54. [24] Fausto N, Laird AD, Webber EM. Liver regeneration. 2. Role of growth factors and cytokines in hepatic regeneration. FASEB J 1995; 9:1527–36. [25] Polimeno L, Azzarone A, Zeng XH, et al. Cell proliferation and oncogene expression after bile duct ligation in the rat: evidence of a specific growth effect on bile duct cells. Hepatology 1995;21:1070 – 8. [26] Boccacio C, Gaudino G, Gambarotta G, et al. Hepatocyte growth factor receptor expression is inducible and is part of the delayed-early response to HGF. J Biol Chem 1994;269:12846 –51. [27] Kobayashi T, Honke K, Miyazaki T, et al. Hepatocyte growth factor specifically binds to sulfoglycolipids. J Biol Chem 1994;269:9817– 21.
Effect of fibrin glue occlusion of the hepatobiliary tract on thioacetamide-induced liver failure Thomas C. Schmandra, M.D., Holger Bauer, M.D., Henrik Petrowsky, M.D., Gu¨nther Herrmann, M.D., Albrecht Encke, M.D., Ernst Hanisch, M.D.* Department of Surgery, Johann Wolfgang Goethe-University Hospital, Theodor Stern Kai 7, 60590 Frankfurt am Main, Germany Manuscript received September 5, 2000; revised manuscript February 19, 2001
Abstract Background: Expression and activation of hepatocyte growth factor (HGF) is stimulated by a complex system of interacting proteins, with thrombin playing an initial role in this process. The impact of temporary occlusion of the hepatobiliary tract with fibrin glue (major component thrombin) on the HGF system in acute and chronic liver damage in a rat model was investigated. Methods: Chronic liver damage was induced in 40 rats by daily intraperitoneal application of thioacetamide (100 mg/kg) for 14 days. After 7 days half of them received an injection of 0.2 mL fibrin glue into the hepatobiliary system. Daily intraperitoneal administration of thioacetamide continued for 7 consecutive days. The rats were then sacrificed for blood and tissue analysis. Acute liver failure was induced in 12 rats by intraperitoneal administration of a lethal dose of thioacetamide (500 mg/kg per day for 3 days) after an injection with 0.2 mL fibrin glue into their hepatobiliary tract. Survival rates and histological outcome were investigated and compared with control animals. Results: Fibrin glue occluded rats showed significantly lower liver enzyme activities and serum levels of bilirubin, creatinine and urea nitrogen. Immunohistochemistry revealed a significant increase in c-met-, HGF␣- and especially HGF-positive cells. Rats subjected to a lethal dose of thioacetamide survived when fibrin glue was applied 24 hours prior to the toxic challenge. These animals showed normal liver structure and no clinical abnormalities. Conclusion: Fibrin glue occlusion of the hepatobiliary tract induces therapeutic and prophylactic effects on chronic and acute liver failure by stimulating the HGF system. Therefore, fibrin glue occlusion might be useful in treating toxic liver failure. © 2001 Excerpta Medica, Inc. All rights reserved. Keywords: Fibrin glue occlusion; Hepatocyte growth factor; Liver regeneration; Liver failure
Hepatocyte growth factor (HGF), a multifunctional cytokine originally identified as a potent mitogen agent for rat hepatocytes [1–3], is the most potent stimulus for hepatocyte growth and DNA synthesis. This process is essential for hepatocyte proliferation during liver regeneration. Following the onset of liver disease or after liver resection, a marked increase in circulating HGF and an upregulation of HGF mRNA expression in the liver occurs [4 –9]. HGF strongly stimulates DNA synthesis of hepatocytes in the liver damaged by surgery or toxic agents, and accelerates liver regeneration, whether applied both systemically or locally [10 –13]. The response to HGF is mediated by the
* Corresponding author. Tel.: ⫹49-69-6301-5142; fax: ⫹49-69-63017452. E-mail address: [email protected]
high-affinity signalling receptor for HGF, which is the protein product of the proto-oncogene c-met [14]. HGF is secreted as an inactive single-chain precursor and remains in this form until it is converted into an active heterodimeric molecule in response to hepatic injury [15]. Activation is exclusively initialized in the injured liver by limited proteolysis at a single site of the HGF precursor. The protease responsible for HGF conversion has been identified as HGF activator (HGFA) [16,17]. HGFA itself is first synthesized as an inactive precursor and then activated by proteolytic processing in response to tissue injury. Thrombin seems to play an initial role in this process, catalyzing the conversion of the precursor to active HGFA [18]. Assuming that activated HGF is the essential factor in regulating liver regeneration after surgical resection or functional damage caused by hepatic disease or toxic influence, activating the HGFA zymogen by thrombin might be of
0002-9610/01/$ – see front matter © 2001 Excerpta Medica, Inc. All rights reserved. PII: S 0 0 0 2 - 9 6 1 0 ( 0 1 ) 0 0 6 5 9 - 6
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
59
Materials and methods
these animals laparotomy was performed 24 hours prior to the first injection of thioacetamide. During laparotomy one half received an injection of 0.2 mL fibrin glue into the hepatobiliary system (n ⫽ 12, group C), the other half underwent mere liver inspection (n ⫽ 12, group D). In the control group (n ⫽ 10) no laparotomy was performed prior to induction of acute liver failure. After the first application of thioacetamide, survival rates, histological outcome, and blood status were recorded for 10 days.
Animal treatments
Histological procedures
All animals (male Sprague-Dawley rats, weight 280 ⫾ 20 g) were maintained under standardized conditions with free access to commercial rodent diet and tap water ad libitum throughout the experiments. The animals received humane care in compliance with the institution’s guidelines and were subjected to a 12-hour day-night cycle. In experiment 1, after induction of general anesthesia (nembutal, 40 mg/kg, intraperitoneal) and midline incision, the common bile duct was microsurgically isolated and incised (1 cm below the lowest tributary) in 15 rats. The bile duct was cannulated with a silicone tube (inner diameter 0.25 mm) and 0.2 mL fibrin glue (32.5 to 57.5 mg fibrinogen, 500 IU aprotinin, 300 IU thrombin, 40 mmol/L calcium chloride in 0.5 mL; Behring, Marburg, Germany) was injected into the proximal biliary tract. The silicone tube was removed and the bile duct incision closed (polyethylene 8-0). All animals survived surgery. To evaluate the reabsorption rate of fibrin glue applied into the hepatobiliary tract, the animals were sacrificed after 6, 12, 18, 24, and 36 hours for liver tissue analysis by fibrin staining. In livers explanted after 24 and 36 hours (n ⫽ 6) immunohistochemical detection of c-met, HGF␣ and HGF was performed. For experiment 2, chronic liver damage was induced by daily intraperitoneal application of a sublethal dose of thioacetamide (100 mg/kg) for 14 consecutive days in 40 rats. After 7 days laparotomy was performed in all animals after induction of general anesthesia (nembutal, 40 mg/kg, intraperitoneal). In group A (n ⫽ 20) fibrin glue (0.2 mL) was injected into the hepatobiliary tract as described above. Animals in group B (n ⫽ 20) underwent mere liver inspection during laparotomy. Following surgery, daily intraperitoneal administration of thioacetamide (100 mg/kg) continued in all animals for 7 consecutive days. On day 15 all animals were sacrificed for tissue and blood analysis. The findings pertaining to groups A and B were compared with those in control animals (group L [n ⫽ 10] to control laparotomy: no thioacetamide application/laparotomy; and group V [n ⫽ 10] to control the vehicle of administering thioacetamide: intraperitoneal administration of sodium chloride 0.9%/no thioacetamide/no laparotomy). For experiment 3, acute liver failure was induced by daily intraperitoneal application of thioacetamide (500 mg/ kg) for 3 consecutive days (lethal dose) in 34 rats. In 24 of
For immunohistochemical detection of activated HGF (consisting of an ␣- and -chain) and the c-met receptor, freshly frozen liver tissue was serially sliced (5 m, 25 slices per animal) in a cryostat, immersed in ice-cold acetone for 10 minutes and dried at ⫺75°C. The slices were then incubated with primary antibodies against recombinant rat HGF␣ (1:200), HGF (1:100), and c-met-peptide (1: 100; Santa Cruz Biotechnology, Santa Cruz, New Mexico). Following administration of secondary antibodies (biotinylated immunoglobulin G; Dianova, Hamburg, Germany), the slices were incubated with alkaline phosphatase-conjugated streptoavidine. The slices were then immersed in a solution with 5% new fuchsin in 0.2 mol/L Tris-HCl (pH 8.3 to 8.5). Nuclei counterstaining was performed with hematoxylin solution. A possible endogenous biotin immunoreactivity was excluded by using a biotin blocking kit before incubating the slides with primary antibodies. Immunohistochemical staining was evaluated with a staining score ranging from 0 (no staining) to 3 (strongly positive reaction). Both primary antibodies and animal test groups were not known to the evaluators. In experiments 1 and 3 excised liver tissue was fixed in 10% neutralized formaldehyde, embedded in paraffin, and stained with hematoxylin and eosin.
potential benefit in treating subjects with liver failure. Therefore, we applied fibrin glue (major component thrombin) into the hepatobiliary tract to investigate the impact of a temporary fibrin glue occlusion on HGF, the proto-oncogene c-met, and hepatic function in induced chronic and acute liver failure in a rat model.
Biochemical measurements The liver-specific cytosolic enzyme activities of aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyltranspeptidase (␥GT), and alkaline phosphatase (AP); and the serum level of bilirubin, creatinine, and urea nitrogen were determined using an automated analyzer system (Hitachi Co., Tokyo, Japan). Statistical analysis Student’s t test (biochemical measurements), the MantelHaenszel test, and the Mantel-Haenszel-Zimmermann test (quantification of immunohistochemical staining) [19,20] were used to determine statistical differences between the groups. Values were considered significantly different if P was less than 0.05.
60
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
Table 1 Serum activity of liver enzymes and serum levels of bilirubin, creatinine, and urea nitrogen in induced chronic liver damage in group A and group B
Group A: thioacetamide ⫹ fibrin glue Group B: thioacetamide
AST (U/L)
ALT (U/L)
␥GT (U/L)
AP (U/L)
Bilirubin (mg/dL)
Creatinine (mg/dL)
Urea (mg/dL)
32.5 ⫾ 3.6* 40.1 ⫾ 6.8
17.4 ⫾ 5.8* 33 ⫾ 9
3 ⫾ 0.9 3.6 ⫾ 0.9
237 ⫾ 42* 335 ⫾ 49
0.41 ⫾ 0.2* 0.73 ⫾ 0.17
0.46 ⫾ 0.08* 0.63 ⫾ 0.1
37.4 ⫾ 4.8* 50.3 ⫾ 9.6
Group A was given daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days and fibrin glue occlusion of the hepatobiliary tract on day 7; n ⫽ 20 and group B was given daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days; n ⫽ 20. Values are means ⫾ SD; P ⬍0.05 was deemed significant. * P ⬍0.05. AST ⫽ aspartate aminotransferase; ALT ⫽ alanine aminotransferase; ␥GT ⫽ gamma-glutamyetranspeptidase; AP ⫽ alkaline phosphatase.
Results Reabsorption of fibrin glue applied to the hepatobiliary system (experiment 1) Application of 0.2 mL fibrin glue into the hepatobiliary tract led to temporary occlusion. The plug was for the most part reabsorbed within 24 hours. After 36 hours no fibrin glue could be detected in the hepatobiliary system. In comparison with normal liver tissue, immunohistochemistry in livers explanted 24 or 36 hours after fibrin glue occlusion revealed an increased number of cells positive for c-met, HGF␣, and HGF. Effect of fibrin glue occlusion on chronic liver damage (experiment 2) In group A (fibrin glue occluded), all animals survived thioacetamide administration, laparotomy, and hepatobiliaric application of fibrin glue. After surgery the animals recovered well and their body weight increased to 294 ⫾ 27.6 g (⫹5.8%). Serum enzyme activities of AST, ALT, AP, and the serum levels of bilirubin, creatinine, and urea
nitrogen were significantly lower than in group B (Table 1) but showed no significant differences compared with control animals. Standard histopathology in group A showed abundant inflammatory cells infiltrating the portal areas. Bile ducts and ductules were not extended. Compared with group B and the control groups L and V, immunohistochemistry revealed a significantly higher number of HGF␣-positive (score: 2.73 ⫾ 0.44) and especially HGF-positive cells (score: 1.27 ⫾ 1.06) (Figs. 1, 2). HGF-positive cells were predominantly found around the portal triads, to a lesser degree around the central veins, and diffusely in the liver. C-met–positive cells were abundant with a diffuse distribution pattern throughout the liver. Immunoreactivity of c-met–positive cells was significantly increased (staining score: 2.73 ⫾ 0.44; Fig. 2). In group B (not fibrin glue occluded), animals appeared less vital and agile, compared with animals of group A (fibrin glue occlusion) and the control groups. Mean body weight decreased to 261 ⫾ 17.6 g (⫺6.8%) and all animals developed moderate to severe ascites. Three animals (15%) died before day 15 (2 on day 12, 1 on day 13). Immunohistochemistry showed a great number of HGF␣-positive cells (score: 1.67 ⫾ 1.18), particularly around the liver
Fig. 1. Immunostaining of HGF-positive cells by primary antibody in group A (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days and fibrin glue occlusion of the hepatobiliary tract on day 7; n ⫽ 20) and group B (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days; n ⫽ 20). The sections were immersed in a solution with 5% new fuchsin in 0.2 mol/L Tris-HCl. Counterstaining of the nuclei was performed with hematoxylin solution (immunostained cells: red; nuclei: blue).
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
61
statistically different to the control groups (score: 0). Cmet–positive cells were numerous (score: 2.25 ⫾ 0.59) and found diffusely in the liver. Staining was predominantly localized on the plasma membrane. Compared with the control groups L and V immunoreactivity of c-met–positive cells was not significantly different (Fig. 2). Effect of fibrin glue occlusion on acute liver failure (experiment 3)
Fig. 2. Immunostaining of HGF␣-, HGF-, and c-met–positive cells by primary antibodies in group A (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days and fibrin glue occlusion of the hepatobiliary tract on day 7; n ⫽ 20), group B (daily intraperitoneal administration of 100 mg/kg thioacetamide on 14 consecutive days; n ⫽ 20), and control groups L (laparotomy, no thioacetamide, no fibrin glue; n ⫽ 10) and V (intraperitoneal administration of sodium chloride 0.9%, no laparotomy, no thioacetamide, no fibrin glue; n ⫽ 10). Immunoreactivity is quantified using a special scoring system ranging from 0 (no staining) to 3 (strong positive reaction). The Mantel-Haenszel and the Mantel-HaenszelZimmermann tests were used to determine statistical significance. Values are means ⫾ SD; P ⬍0.05 was deemed significant. agroup A versus group B and control groups; P ⬍0.05. bgroup B versus control groups; P ⬍0.05.
portal triads. Most of these cells had large nuclei and could be histologically identified as nonparenchymal cells. Around the central veins, HGF staining increased diffusely, but the number of stained cells was fewer than those surrounding the portal triads. In comparison with the control groups L (score: 0.67 ⫾ 0.47) and V (score: 0.71 ⫾ 0.59) the animals revealed a significantly higher number of HGF␣-positive cells. Incubation with primary antibodies against the HGF-chain revealed no, or weak, immunohistochemical staining (score: 0.09 ⫾ 0.18), which was not
In group C (fibrin glue occluded prior to thioacetamide application), all animals survived the application of a lethal dose of thioacetamide. During the experiment they appeared to be clinically normal with no differences to the control animals. Blood analysis, which was performed on day 10 of the experiment, revealed no statistical differences to normal animals. Histological analysis showed a normal liver structure and almost no effects of the lethal application of thioacetamide. In group D (not fibrin glue occluded prior to thioacetamide application), all animals died within 48 hours after the application of a lethal dose of thioacetamide. Histological analysis of liver tissue revealed fulminant liver necrosis, predominantly localized around the central veins. No differences from the control animals, which also died within 48 hours, were observed (Fig. 3; Table 2).
Comments HGF plays a significant role in liver regeneration, stimulating DNA synthesis of hepatocytes in toxically or mechanically injured liver tissue or during liver infection. The serum concentration of HGF rises dramatically after liver
Fig. 3. Histology of the liver after induction of acute liver failure in group C (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; fibrin glue occlusion of the hepatobiliary tract 24 hours prior to the first thioacetamide administration; n ⫽ 12) and group D (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; liver inspection 24 hours prior to the first thioacetamide administration; n ⫽ 12). Liver tissue is fixed in 10% neutralized formaldehyde, embedded in paraffin, and stained with hematoxylin and eosin (nuclei: blue; tissue necrosis: red).
62
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
Table 2 Mortality 2, 5, and 10 days after induction of acute liver failure 2 Days Group C: laparotomy ⫹ fibrin glue ⫹ thioacetamide Group D: laparotomy ⫹ thioacetamide Control: thioacetamide
0/12 (0%)
5 Days
10 Days
0/12 (0%)
0/12 (0%)
12/12 (100%) 10/10 (100%)
Group C (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; fibrin glue occlusion of the hepatobiliary tract 24 hours prior to the first thioacetamide administration; n ⫽ 12); group D (liver inspection 24 hours prior to the first thioacetamide application; n ⫽ 12); and control group (daily intraperitoneal administration of 500 mg/kg thioacetamide on 3 consecutive days; n ⫽ 10).
surgery [6] and remains elevated for the duration of the regenerative process in the liver. Similar increases in HGF are found when hepatic regenerative activity is induced by drugs, toxic substances, or peroxisome proliferators [21– 23]. The HGF response correlates with the extent of hepatic injury. In patients with acute or chronic liver disease the serum concentration of HGF is elevated. The rise correlates with the severity of the disease and— especially in fulminant liver failure—the HGF level is inversely correlated with the prognosis of the patient [4,7]. However, no detailed studies of the nature of the elevated HGF in serum have been performed. Therefore, it is possible that the circulating HGF may be in the inactive form. In animal models infusion of HGF into the portal vein or the systemic circulation has stimulated regenerative response in the injured liver [13,24]. In our study thioacetamide was applied intraperitoneal to induce chronic and acute liver injury. The chronic application of thioacetamide, which results in chronic liver damage (experiment 2), caused a significant increase in HGF␣positive cells in the liver. Almost no HGF-positive cells were found in these animals, inferring that HGF is in the inactive, single-chain form, and a conversion to the active heterodimeric molecule is needed for an effective response to hepatic injury. Fibrin glue occlusion of the hepatobiliary system induced an effective regenerative response to the thioacetamide-induced chronic liver damage. The animals treated with fibrin glue showed superior liver function with lower liver enzyme activities and serum bilirubin (despite temporary occlusion of the bile tract), creatinine and urea nitrogen levels. No catabolic metabolism was observed, as was apparent in nonoccluded animals. Immunohistochemistry revealed a significant increase in HGF␣- and especially HGF-positive cells (13-fold higher) representing the activated form of HGF. The protease HGFA is responsible for the conversion of HGF into the active form. HGFA itself is synthesized as an inactive molecule, which is activated by proteolytic processing. Thrombin is suggested to be responsible for this local effect, as it has been proved to play the initial role in activating the
HGFA zymogen [18]. Therefore, the protective effect of fibrin glue may be due to activation of the HGFA/HGF process. Since permanent bile duct ligation is a potent stimulus for hepatocellular proliferation [25], the temporary occlusion of the hepatobiliary tract after fibrin glue application may also influence HGF expression. In our experiments fibrin glue was reabsorbed within 36 hours after occlusion and liver tissue analysis showed that bile ducts and ductules were not extended. Serum levels of bilirubin and parameters indicating cholestasis were significantly lower in the fibrin glue occluded animals. This makes occlusion itself unlikely to be the main cause for HGF activation. The response to activated HGF is mediated by the high-affinity signalling receptor for HGF, the protein product of the proto-oncogene c-met [14]. Immunoreactivity of c-met–positive cells was significantly higher in fibrin glue occluded animals, indicating an upregulation of the HGF receptor expression. Upregulation of the HGF receptor is inducible and HGF itself is known to induce expression of its own receptor [26]. Whether the increase in c-met–positive cells is caused by stimulation of the HGF activation process itself, or fibrin glue stimulates the c-met oncogene is not clear. In our study c-met–positive cells were uniformly distributed throughout the liver, whereas HGF-positive cells were predominantly found in the periportal region and—to lesser degree— around the central veins. Possibly, HGF, attached to high affinity sites in the periportal region, is not bound to the HGF receptor. Investigators have partly characterized the mechanisms whereby HGF binds to periportal high affinity sites, but the precise process is not completely clear [27]. Aside from fibrin glue application reducing liver damage, prophylactic fibrin glue occlusion significantly increased the survival rate in acute liver failure. After fibrin glue occlusion the animals withstood a lethal challenge of thioacetamide and showed negligible clinical or histological pathology. This underlines the importance of activated HGF for adequate liver regeneration, since survival after fulminant hepatic failure is determined by the extent of liver injury and the ability of hepatocyte regeneration. In conclusion, local application of fibrin glue into the hepatobiliary tract might have positive effects, useful in treating toxic liver damage or fulminant hepatic failure. References [1] Nakamura T, Nawa K, Ichihara A. Partial purification and charcterization of hepatocyte growth factor from serum of hepatectomized rats. Biochem Biophys Res Commun 1984;122:1450 –9. [2] Russell WE, McGowan JA, Bucher NLR. Partial characterization of hepatocyte growth factor from rat platelets. J Cell Physiol 1984;119: 183–92. [3] Thaler J, Michalopoulos GK. Hepatopoitin A: a partial characterization and trypsin activation of a hepatocyte growth factor. Cancer Res 1985;45:2545–9. [4] Tsobouchi H, Niitani Y, Hirono S, et al. Levels of the human hepatocyte growth factor in serum of patients with various liver diseases
T.C. Schmandra et al. / The American Journal of Surgery 182 (2001) 58 – 63
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
determined by an enzyme-linked immunosorbent assay. Hepatology 1991;13:1–5. Shiota G, Okano J, Kawasaki H, et al. Serum hepatocyte growth factor levels in liver diseases: clinical implications. Hepatology 1995; 21:106 –12. Lindroos PM, Zarnegar R, Michalopoulos GK. Hepatocyte growth factor (hepatopoitin A) rapidly increases in plasma before DNA synthesis and liver regeneration stimulation by partial hepatectomy and carbon tetrachloride administration. Hepatology 1991;13:743–9. Tomiya T, Nagoshi S, Fujiwara K. Significance of serum human hepatocyte growth factor levels in patients with hepatic failure. Hepatology 1992;15:1– 4. Kinoshita T, Tashiro K, Nakamura T. Marked increase of HGF mRNA in non-parenchymal liver cells of rats treated with hepatotoxines. Biochem Biophys Res Commun 1989;165:1229 –34. Zarnegar R, DeFrances MC, Kost DP, et al. Expression of hepatocyte growth factor mRNA in regenerative rat liver after partial hepatectomy. Biochem Biophys Res Commun 1991;177:559 – 65. Ishiki Y, Ohnishi H, Muto Y, et al. Direct evidence that hepatocyte growth factor is a hepatotrophic factor for liver regeneration and for potent anti-hepatitis action in vivo. Hepatology 1992;16:1227–35. Fujiwara K, Nagoshi S, Ohno A, et al. Stimulation of liver growth by exogenous human hepatocyte growth factor in normal and partially hepatectomized rats. Hepatology 1993;18:1443–9. Ishii T, Sato M, Sudo K, et al. Hepatocyte growth factor stimulates liver regeneration and elevates blood protein level in normal and partially hepatectomized rats. J Biochem 1995;117:1105–12. Kobayashi Y, Hamanoue M, Ueno S, et al. Induction of hepatocyte growth by intraportal infusion of HGF into beagle dogs. Biochem Biophys Res Commun 1996;220:7–12. Naldini L, Vigna E, Narsimhan RP, et al. Hepatocyte growth factor (HGF) stimulates the tyrosine kinase activity of the receptor encoded by the proto-oncogene c-MET. Oncogene 1991;6:501– 4. Miyazawa K, Shimomoura T, Naka D, Kitamura N. Proteolytic activation of hepatocyte growth factor in response to tissue injury. J Biol Chem 1994;269:8966 –70.
63
[16] Miyazawa K, Shimomura T, Kitamura A, et al. Molecular cloning and sequence analysis of the cDNA for a human serine protease responsible for activation of hepatocyte growth factor. J Biol Chem 1993;268:10024 – 8. [17] Miyazawa K, Shimomura T, Kitamura N. Activation of hepatocyte growth factor in the injured tissues is mediated by hepatocyte growth factor activator. J Biol Chem 1996;271:3615–18. [18] Shimomura T, Kondo J, Ochiai M, et al. Activation of the zymogen of hepatocyte growth factor activator by thrombin. J Biol Chem 1993;268:22927–32. [19] Mantel N. Chi2-tests with one degree of freedom: extensions of the Mantel-Haenszel procedure. JASA 1963;58:690 –700. [20] Zimmermann H. Erweiterung des Mantel-Haenszel-Testes auf Kl⬎2 gruppen. Biomed J 1984;26:223– 4. [21] Lindroos P, Tsai WH, Zarnegar R, Michalopoulos GK. Plasma levels of HGF in rats treated with tumor promoters. Carcinogenesis 1992; 13:139 – 41. [22] Matsumoto K, Nakamura T. Hepatocyte growth factor: molecular structure and implications for a central role in liver regeneration. J Gastroenterol Hepatol 1991;6:509 –19. [23] Matsumoto K, Nakamura T. Hepatocyte growth factor: molecular structure, roles in liver regeneration, and other biological functions. Crit Rev Oncogen 1992;3:27–54. [24] Fausto N, Laird AD, Webber EM. Liver regeneration. 2. Role of growth factors and cytokines in hepatic regeneration. FASEB J 1995; 9:1527–36. [25] Polimeno L, Azzarone A, Zeng XH, et al. Cell proliferation and oncogene expression after bile duct ligation in the rat: evidence of a specific growth effect on bile duct cells. Hepatology 1995;21:1070 – 8. [26] Boccacio C, Gaudino G, Gambarotta G, et al. Hepatocyte growth factor receptor expression is inducible and is part of the delayed-early response to HGF. J Biol Chem 1994;269:12846 –51. [27] Kobayashi T, Honke K, Miyazaki T, et al. Hepatocyte growth factor specifically binds to sulfoglycolipids. J Biol Chem 1994;269:9817– 21.