Increased Serum Levels and Sinusoidal Expression of Thrombomodulin in Acute Liver Damage

Increased Serum Levels and Sinusoidal Expression of Thrombomodulin in Acute Liver Damage

Thrombosis Research 93 (1999) 113–120 REGULAR ARTICLE Increased Serum Levels and Sinusoidal Expression of Thrombomodulin in Acute Liver Damage Masao...

799KB Sizes 0 Downloads 24 Views

Thrombosis Research 93 (1999) 113–120

REGULAR ARTICLE

Increased Serum Levels and Sinusoidal Expression of Thrombomodulin in Acute Liver Damage Masao Takatori1, S. Iwabuchi1, S. Ro1, M. Murayama1, S. Maeyama2, T. Uchikoshi2, M. Nakano3 and H. Ishii4 1 Second Department of Internal Medicine, St. Marianna University School of Medicine, Sugao, Miyamae-ku, Kawasaki, 2-16-1; 2Second Department of Pathology, St. Marianna University School of Medicine, Sugao, Miyamae-ku, Kawasaki, 2-16-1; 3Niigata Research Laboratory, Mitsubishi Gas Chemical Company, Inc., 182 Shinwari, Tayuhama, Niigata 950-31; 4Department of Public Health, Showa College of Pharmaceutical Sciences, 3-3165 Higashi Tamagawa Gakuen, Machida, Tokyo 194, Japan. (Received 1 April 1998 by Editor A. Takada; revised/accepted 10 September 1998)

Abstract Thrombomodulin (TM) is a surface glycoprotein of endothelial cells involved in both anticoagulation and antifibrinolysis. In this study, we assessed the clinical significance of TM in acute liver damage by using a rat model induced by intraperitoneal injection of D-galactosamine (Gal-N). Serum TM levels were measured with enzyme immunoassay utilizing rabbit anti-rat TM antibody. Simultaneously, immunohistochemical examination was performed using the same antibody. Serum TM levels increased significantly after the injection of Gal-N compared with preinjection levels, peaking from 48 to 72 hours after injection and normalizing by 168 hours. Changes in parenchymal damage were synchronized with changes of TM, and changes of TM levels mirrored changes of liver weight. In immunohistochemical examination, TM immunoreactivity was observed only on the endothelial surfaces of both the artery and portal vein within Abbreviations: TM, thrombomodulin; EIA, enzyme immunoassay; Gal-N, D-galactosamine; DIC, disseminated intravascular coagulation; DM, diabetes mellitus; ALT, alanine aminotransferase; T. Bil, total bilirubin; Cr, creatinine; AI, after injection; BI, before injection; MMP, matrix metallo protease. Corresponding author: M. Takatori, Second Department of Internal Medicine, St. Marianna University School of Medicine, Sugao, Miyamae-ku, Kawasaki, 2-16-1, Japan. Tel: 181 (44) 977 8111; Fax: 181 (44) 976 7093.

Glisson’s sheath in controls. After injection of GalN, TM immunoreactivity was gradually intensified, especially around the necrotic area and the central veins. These findings disappeared with improvement of parenchymal damage. Both the increase of serum TM levels and intensified TM immunoreactivity in the liver were synchronized with acute liver parenchymal damage induced by Gal-N. These findings on TM are related to endothelial damage with parenchymal necrosis and liver regeneration interacting with both homeostasis of microcirculation and healing of parenchymal damage.  1999 Elsevier Science Ltd. Key Words: Thrombomodulin; Rat; EIA; Acute liver damage; D-Galactosamine

T

hrombomodulin (TM) is a surface glycoprotein of endothelial cells that participates in anticoagulation on the endothelial surface by inactivating thrombin and activating Protein C [1,2]. Simultaneously, TM inhibits fibrolysis by activating proplasma carboxypeptidase B [3,4] and degradating single-chain urokinase-type plasminogen activator in the presence of thrombin [5]. Ishii et al. [6,7] reported that TM antigen could be detected in both human urine and blood and called it soluble TM. The levels of soluble TM detected in blood are maintained by the turnover of endothelial cells [8], and they are increased in patients

0049-3848/99 $–see front matter  1999 Elsevier Science Ltd. Printed in the USA. All rights reserved. PII S0049-3848(98)00167-4

114

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

with disseminated intravascular coagulation (DIC) [9–11], collagen disease [12,13], and diabetes mellitus (DM) [14,15]. Therefore, the levels of serum soluble TM have clinical significance as a marker of endothelial cell damage [16,17]. Although we have reported increases of serum TM levels in various types of liver damage [18], the mechanism of increase is still obscure. Recently, an enzyme immunoassay (EIA) using rabbit anti-rat TM polyclonal antibody was established [19]. In this study, we made a rat model of acute liver damage by intraperitoneal injection of D-galactosamine (Gal-N) and measured serum TM levels serially using EIA. Additionally, we examined immunohistochemical distribution of TM in rat liver using the same antibody. The aim of this study is to investigate the clinical significance of TM in acute liver damage.

1. Materials and Methods 1.1. Animals Male Sprague-Dawley (SD) rats weighing approximately 200 g were used. They were fed a commercial pellet diet and water ad libitum and kept in a room maintained at 22628C under normal laboratory lighting conditions. One group of six rats injected with saline served as a control. Five groups of 10 rats were administered Gal-N after adjusted to pH 7.4 by an intraperitoneal injection of 200 mg/100 g body weight. Serum samples were obtained from the six groups before injection and 24, 48, 72, 96, and 168 hours after injection from the inferior vena cava under ether anesthesia. The same rat group was used for immunohistochemical examination of the liver. The study protocol conformed to the National Research Council criteria for humane care.

1.2. Measurement of Serum TM Levels Rabbit anti-rat TM antibodies were prepared as previously reported [19]. In short, male rabbits were immunized with 100 mg of purified rat TM in complete Freund’s adjuvant injected to multiple intradermal sites. Anti-rat TM antibodies were isolated from antisera by using Affi-Prep Protein A chromatography. Anti-rat TM antibodies showed approximately

10000-fold specificity to immobilized rat lung TM when compared with preimmune IgG. The cofactor activity of TM (2.4 ng) was inhibited by 50% with 216 ng of anti-TM antibody, although the activity of TM (2.4 ng) was not inhibited at all with 4.8 mg of preimmune IgG. The sandwich enzyme immunoassay (EIA) for rat TM was developed using the anti-rat TM antibodies [19]. Each well of a microtiter plate was coated with 200 ml of 10 mg/ml rabbit anti-rat TM antibodies dissolved in 50 mM carbonate buffer, pH 9.5, overnight at 48C. The plate was washed six times with PBS containing 0.05% Tween 20 and left to stand with 300 ml of PBS containing 0.2% BSA and 0.05% Tween 20 for 1 hour at room temperature. After discarding the buffer, 200 ml of rat serum specimen diluted 200 times with PBS containing 0.2% BSA and 0.05% Tween 20 was placed in each well. The plate was incubated overnight at 48C. The plate was washed six times with PBS containing 0.05% Tween 20. Two hundred ml of HRPlabeled Fab9 of rabbit anti-rat TM (Fab9 conc.51.0 mg/ml) was placed in each well and left to stand for 3 hours at 258C. After washing the plate six times with PBS containing 0.05% Tween 20, 200 ml of substrate solution containing 0.015% H2O2 and 1.0 mg/ml o-phenylendiamine in 0.1 M phosphate-citrate buffer (pH 5.9) was placed in each well, and the plate was incubated for 30 minutes at 258C. Color development was terminated with 50 ml of 4N H2SO4. Absorbance at 490 nm was measured using an Immunoreader NJ-2000 (Japan Intermed Co., Tokyo, Japan). The immunoaffinity purified rat lung TM was used as a standard. The standard curve was linear below 8 ng/ml. The minimum concentration detected was 0.05 ng/ ml. The coefficients of variation in intra-assay and interassay were both below 4.0%. This EIA can be assumed to hold a high species specificity to rat TM, as the TM from human, rabbit, murine, and canine was not detected in the EIA for rat TM. Serum ALT, T. Bil, and Cr were measured using commercial kits (ALT, Wako Junyaku Co.; T. Bil, Nihon Shoji Co.; Cr, Yatoron Co., Japan)

1.3. Immunohistochemical Examination For immunohistochemical examination of the liver, the liver was soaked in formaldehyde immediately after removal and fixed in paraffin. After deparaffin-

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

ization with alcohol, immunohistochemical staining was conducted by LSAB method with a LSAB kit (DAKO Inc., USA) using rabbit anti-rat TM polyclonal antibody as the primary antibody. The reaction product was developed by incubation in a solution of diaminobenzidine tetrahydrochloride (containing 30% H2O2, pH 7.6). Serially cut sections were also stained by hematoxylin-eosin for studying liver histology.

1.4. Statistical Analysis The means between two groups (e.g., before injection of Gal-N and 24 hours after injection) were compared using the unpaired Student’s t-test. All tests of significance were two-tailed, with a p value of less than 0.05 considered to indicate significance.

2. Results 2.1. Serial Change of Serum TM, ALT, and T. Bil levels after Gal-N Injection After intraperitoneal injection of Gal-N, serum TM levels increased, peaking from 48 to 72 hours after

Fig. 1. Serial changes of serum TM, ALT, and T. Bil levels after Gal-N injection. Dashed lines indicate the mean6SD.

115

injection (AI) and then normalizing by 168 hours AI (see Figure 1). Serum TM levels were 592.86 35.6 ng/ml (n510) before injection (BI), 900.86 108.7 ng/ml (n59) at 24 hours AI, 1475.16190.6 ng/ml (n59) at 48 hours AI, 1490.46105.4 ng/ml (n55) at 72 hours AI, 1015.26390.3 ng/ml (n59) at 96 hours AI, and 623.6635.2 ng/ml (n54) at 168 hours AI. Serum TM levels at all measurement points after injection except 168 hours were significantly higher than the BI level (BI vs. 24, 48, 72, and 96 hours AI: p,0.01). Serum ALT levels of Gal-N after injection increased up to a peak at 48 hours AI and were normalized at 168 hours AI. Serum ALT levels were 52.7620.1 IU/l (n56) BI, 2602.962738.1 IU/l (n57) at 24 hours AI, 8704.764914.0 IU/l (n59) at 48 hours AI, 7062.361129.6 IU/l (n53) at 72 hours AI, 403.16371.4 IU/l (n57) at 96 hours AI, and 48.0616.1 IU/l (n54) at 168 hours AI. Serum ALT levels at all measurement points after injection except 168 hours, were significantly higher than the BI level (BI vs. 24 and 96 hours AI: p,0.05; BI vs. 48 and 72 hours AI: p,0.01). Serum T. Bil levels of Gal-N after injection increased up to a peak at 72 hours AI and were

116

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

Fig. 2. Serial changes of serum creatinine and TM levels after Gal-N injection.

normalized by 168 hours AI. Serum T. Bil levels were 0.1560.12 mg/dl (n56) BI, 0.2460.18 mg/dl (n510) at 24 hours AI, 3.2461.37 mg/dl (n511) at 48 hours AI, 4.8761.96 mg/dl (n53) at 72 hours AI, 1.1361.61 mg/dl (n57) at 96 hours AI, and 0.1560.10 mg/dl (n54) at 168 hours AI. Serum T. Bil levels 48 and 72 hours AI were significantly higher than the BI level (BI vs. 48 and 72 hours AI: p,0.01).

2.2. Serial Change of Serum Creatinine and TM Levels after Gal-N Injection Serum creatinine levels increased gradually AI, peaking at 48 hours AI (see Figure 2). However, a significant difference from the BI level was observed only at 48 hours AI. (BI vs. 48 hours AI: p,0.05). Serum creatinine levels were 0.3860.14 mg/dl (n56) BI, 0.5060.08 mg/dl (n57) at 24 hours AI, 0.5560.15 mg/dl (n511) at 48 hours AI, 0.3560.21 mg/dl (n52) at 72 hours AI, 0.6060.31 mg/dl (n57) at 96 hours AI, and 0.4760.05 mg/dl (n53) at 168 hours AI.

2.3. Serial Change of Serum TM Levels and Liver Weight after Gal-N Injection Liver weights decreased after injection of Gal-N, reaching their lowest levels at 72 hours AI and then recovering gradually with a return to BI weights at 168 hours AI (see Figure 3). Serial changes of serum TM levels mirrored the changes in liver weight. Liver weights were 9.8560.20 g (n53) BI, 8.4160.47 g (n59) at 24 hours AI, 6.6060.37 g (n59) at 48 hours AI, 4.9260.67 g (n55) at 72 hours AI, 6.9662.80 g (n59) at 96 hours AI, and 9.4562.02 g (n57) at 168 hours AI. Liver weights at 24, 48, and 72 hours AI were significantly lower than the BI level (BI vs. 24, 48, and 72 hours AI: p,0.01).

2.4. Immunohistostaining of TM 2.4.1. HE Staining 24 Hours after Gal-N Injection As shown in Figure 4A, scattered hepatocyte necrosis in parenchyma and mild infiltration of in-

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

117

Fig. 3. Serial changes of serum TM levels and liver weight after Gal-N injection.

flammatory cells in Glisson’s sheath were identified.

2.4.2. TM Staining 24 Hours after Gal-N Injection With the control groups, immunopositive TM staining was observed only on the endothelial surface of both artery and portal vein within Glisson’s sheath. As shown in Figure 4B, 24 hours after Gal-N injection, a mild increase in TM immunoreactivity was identified on the surface of the sinusoid surrounding the parenchymal necrotic area. 2.4.3. HE Staining 48 Hours after Gal-N Injection As shown in Figure 4C, spread of the parenchymal necrotic area and increased infiltration of inflammatory cells in Glisson’s sheath compared with the levels 24 hours after injection were identified. 2.4.4. TM Staining 48 Hours after Gal-N Injection As shown in Figure 4D, spreading and increases of TM immunoreactivity were identified on the surface of the sinusoid surrounding the parenchymal necrotic area. TM staining was also identified on the surface of the sinusoid around the central vein. These findings on TM immunoreactivity peaked from 48 to 72 hours after injection of Gal-N

and dissapeared by 168 hours AI. Control staining with normal rabbit sera was negative.

3. Discussion The degree of acute parenchymal liver damage was estimated by the change in serum ALT and T. Bil levels. We found that the serum TM levels in this model increased in parallel with acute parenchymal liver damage induced by Gal-N and that course of these parameters peaked at 48 to 72 hours of injection. In liver TM immunohistochemical staining, TM immunoreactivity increased on the sinusoidal endothelial cells, especially around the parenchymal necrotic area and central veins, and these findings were also synchronized with the course of liver damage and the change of serum TM levels. The liver is an organ constituting with the portal vein, artery, and hepatic vein vessel systems and containing many vascular endothelial cells. Therefore, the synchronicity between the increase of serum TM levels in Gal-N liver damage and the increase of immunohistostaining of TM in the necrotic area suggest that the release of TM from the damaged endothelial cells in those areas might contribute to the increase of serum TM levels [16].

118

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

Fig. 4. (A) HE staining, 24 hours after Gal-N injection (3200). (B) TM staining, 24 hours after Gal-N injection (3200). (C) HE staining, 48 hours after Gal-N injection (3200). (D) TM staining, 48 hours after Gal-N injection (3200).

Severe hepatocellular necrosis with massive hemorrhage in parenchyma might induce impaired intrahepatic microcirculation. It is well established that TM has anticoagulation properties in which TM forms a 1:1 complex with thrombin and serves as a cofactor for thrombin-catalyzed activation of protein C [1,2]. However, it was recently reported that TM also plays a role in antifibrinolysis by accelerating thrombin-induced TAFI (thrombinactivatable fibrinolysis inhibitor) or plasma procarboxypeptidase B activation. Activated TAFI inhibits plasminogen activation by removing C-terminal lysines from partially degrated fibrin [3,4]. Small amounts of TM may predominantly act as a cofactor for antifibrinolysis rather than for anticoagulant [20]. Additionally, TM accelerates the inactivation of single-chain urokinase-type plasminogen activator by thrombin [5]. Thus, TM contributes to inhibi-

tion of plasmin generation. Consequently, inhibition of plasmin-induced matrix metallo protease (MMP) activation might suppress the degradation of extracellular matrix protein, thus indicating a relation between TM and the healing of damaged tissue. TM shows antifibrinolysis properties with the presence of thrombin that could be useful to prevent rebleeding in the process of healing damaged vascular tissue. Serial change of serum TM levels and TM immunoreactivity in liver tissue mirrored the change in liver weight. In rat hepatectomy models, serum TM levels increased immediately after hepatectomy and peaked at 48 and 72 hours, which corresponded to the timing of maximum liver regeneration [21]. Additionally, TM immunoreactivity occured on the surface of sinusouidal endothelium and also exhibited a synchronization with the course of serum

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

TM levels similar to the synchronizations seen in Gal-N models [21]. These serial changes of serum TM levels after hepatectomy and Gal-N–induced liver damage are similar to those of hepatocyte growth factor in the course of liver regeneration [22]. It is reported that the epidermal growth factor–like domain of recombinant human TM can exhibit mitogenic activity for fibroblast cell line [23]. These findings indicate the close relationship between increase of serum TM levels, immunoreactivity on sinusoid, and liver regeneration. Since TM is metabolized mainly by the kidney, impaired renal function increases serum TM levels [16]. In this study, serum creatinine levels increased gradually in the acute phase of liver damage. However, significant differences from the preinjection levels of serum creatinine were only observed at 48 hours after injection. Within the dosage of GalN administered in this study, little effect on renal function was observed. It was reported previously that serum TM levels in patients with various diseases increase to approximately five or six times the levels of normal controls [8–10,12–14]. However, in this report, serum TM levels increased to a maximum of only approximately three times the control level. The reduced influence of Gal-N on renal function in this model might contribute to this difference in the increase of serum TM levels. The serum TM levels of controls in this rat model were approximately 20 or 40 times higher than the levels in humans, and those of rabbit were reported to be approximately 1600 ng/ml [11]. These findings indicate that the specificity of species might contribute to these variations of serum TM levels between humans and other species.

References 1. Esmon CT. The roles of protein C and thrombomodulin in the regulation of blood coagulation. J Biol Chem 1989;264:4743–6. 2. Dittman WA, Majerus PW. Structure and function of thrombomodulin: A natural anticoagulant. Blood 1990;75:329–36. 3. Redritz A, Tan AK, Eaton DL, Plow EF. Plasma carboxypeptidases as regulators of the plasminogen system. J Clin Invest 1995;96: 2534–8. 4. Bajzar L, Morser J, Nesheim M. TAFI, or

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

119

plasma procarboxypeptidase B, couples the coagulation and fibrinolytic cascades through the thrombin-thrombomodulin complex. J Biol Chem 1996;271:16603–8. de Munk GAW, Groeneverd E, Ridjken DC. Acceleration of the thrombin inactivation of single chain urokinase-type plasminogen activator (Pro-urokinase) by thrombomodulin. J Clin Invest 1991;88:1680–4. Ishii H, Majerus PW. Thrombomodulin is present in human plasma and urine. J Clin Invest 1985;76:2178–81. Ishii H, Nakano M, Tsubouchi J, Ishikawa T, Uchiyama H, Hiraishi S, Tahara C, Miyajima Y, Kazama M. Establishment of enzyme immunoassay of human thrombomudulin in plasma and urine using monoclonal antibodies. Thromb Haemost 1990;63:157–62. Kazama M. Soluble Thrombomodulin: A specific parameter of endothelial injury. Jpn J Clin Hematol 1990;32:103–7. Asakura H, Jokaji H, Saito M, Uotani C, Kumabashiri I, Morishita E, Yamazaki M, Matsuda T. Plasma levels of soluble thrombomodulin increase in cases of disseminated intravascular coagulation with organ failure. Am J Hematol 1991;38:281–7. Amano K, Tateyama M, Inaba H, Fukutake K, Fujimaki M. Fluctuations in plasma levels of thrombomodulin in patients with DIC. Thromb Haemost 1992;68:404–6. Uchiyama H, Ohtani H, Hiraishi S, Horie S, Ishii H, Kazama M. Changes in plasma thrombomodulin antigen in rabbit developing endotoxin-induced desseminated intravascular coagulation and the effect of heparin. Thromb Res 1992;65:593–604. Kazama M. Clinical evaluation of hemostatic molecular markers. Acta Haematol 1988;51: 1387. Takaya M, Ichikawa N, Kobayashi N, Kawada T, Shimizu H, Uchiyama M, Moriushi J, Watanabe K, Arimori S. Serum thrombomodulin and anticardiolipin antibodies in patients with systemic lupus erythematosus. Clin Exp Rheumatol 1991;9:395–499. Iwashima Y, Sato T, Watanabe K, Ooshima E, Hiraishi S, Ishii H, Kazama M, Makino I. Elevation of plasma thrombomodulin level in

120

15.

16.

17.

18.

M. Takatori et al./Thrombosis Research 93 (1999) 113–120

diabetic patients with early diabetic neuropathy. Diabetes 1990;39:983–8. Shimano H, Takahashi K, Kawakami M, Gotoda T, Harada K, Shimada M, Yazaki Y, Yamada N. Elevated serum and urinary thrombomodulin levels in patients with non-insulindependent diabetes mellitus. Clinica Chimica Acta 1994;225:89–96. Ishii H, Ushiyama H, Kazama M. Soluble thrombomodulin antigen in conditioned medium is increased by damage of endothelial cells. Thromb Haemostas 1991;65:618–23. Takahashi H, Ito S, Hamano M, Wada K, Niwano H, Seki Y, Shibata A. Circulating thrombomodulin as a novel endothelial cell marker: Comparison of its behavior with von Willebrand factor and tissue-type plasminogen activator. Am J Hematol 1992;41:32–9. Iwabuchi S, Yoshida Y, Kamogawa A. Plasma thrombomodulin in liver diseases. Gastroenterol Jpn 1991;25:661.

19. Daimon T, Kazama M, Miyajima Y, Nakano M. Immunocytochemical localization of thrombomodulin in the aqueous humor pallage of the rat eye. Histochem Cell Biol 1997;108:121–31. 20. Hosaka Y, Takahashi Y, Ishii H. Thrombomodulin in human plasma contributes to inhibit fibrinolysis through accleration of thrombindependent activation of plasma procarboxypeptidase B. Thromb Haemost. In Press 1998. 21. Takahashi Y. Significance of thrombomodulin in liver diseases. Jpn J Clin Pathol 1995;43 (Suppl):80. 22. Shiota G, Okano J, Kawasaki H, Kawamoto T, Nakamura T. Serum hepatocyte growth factor levels in liver diseases: Clinical implication. Hepatology 1995;21:106–12. 23. Hamada H, Ishii H, Sakyo K, Horie S, Nishiki K, Kazama M. The epidermal growth factorlike domain of recombinant human thrombomodulin ehxibits mitogenic activity for Swiss 3T3 cells. Blood 1995;86:225–33.