and others with azan. Hydroxyproline determinations. The remainder of the liver was homogenized in methyl alcohol in a Potter-Elvehjem glass homogenizer ...
Br. J. Exp. Path. (I989) 70, I43-I52
cx1-Antitrypsin and hepatic fibrosis Tsuneo Ozeki, Kenichi Imanishi,* Hitoshi Ueda,t Takehiko Uchiyama,* Keisuke Funakoshi,4 Ikuo Suzuki,* Kazuo Ohuchi,§ Mikio Kan,T and Takeshi SatohT
The 3rd Department of Internal Medicine, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan; * The Department of Bacteriology, Tokyo Women's Medical College; t The Department of Anatomy, University of Occupational and Environmental Health; t The Department of Pathology, Kyushu Dental College, Kitakyushu; § Department of Biochemistry, Faculty of Pharmaceutical Sciences, Tohoku University, Sendai, Japan and T Department of Cell Biology, Research Institute for Tuberculosis and Cancer, Tohoku University Received 17 November I987 Accepted for publication 5 October I988
Summary. The administration of ax-antitrypsin, both sialo and asialo types, to rats with chronic liver injury accelerated hepatic fibrosis. More fibrosis was seen histologically and the amount of hydroxyproline in the liver increased. Moreover, the growth of HEL cells (human embryonal lung fibroblast) in culture was promoted by x1-antitrypsin though the action of the asialo type was very weak by contrast with the sialo type.
a1-antitrypsin (oal-AT) is a human serum protein contained in the ax1-fraction. a1-AT is a glycoprotein (MW 54 000 Da), its sugar content is about I2.4% and it contains galactose, mannose and N-acetylglucosamine as well as sialic acid (Clamp 1975). xiAT is synthesized primarily in the liver (Asofsky & Thorbecke I96I; Prunier et al. I964). The serum content of this protein increases in acute inflammation and it is considered, therefore, to be one of the acute phase reactants (Courtoy et al. I98I). It is known to inhibit the protease activities of trypsin, chymotrypsin, pancreatic elastase, plasmin, thrombin, kallikrein and bacterial collagenase (Schultze et al. I962). It is thought therefore, that a-1AT may protect tissues from degradation by proteases during inflammation. However, the function of o1AT in inflammatory conditions has not yet been elucidated completely. Evidence is
reported here which suggests that ax1-AT accelerates hepatic fibrosis in liver injury.
Materials and methods ax1-AT was purchased from Sigma Chemical Company (USA). Asialo-x1-AT was made as follows: I5 ml of sialidase (Sigma Chemical, USA) solution (5 unit/I 5 ml in O.I M sodium bicarbonate pH 8.3, containing 0. 5 M, NaCl), and i 5 ml of activated CNBr-sepharose CL6B (Pharmacia, Sweden), washed by the same coupling buffer, were slowly and automatically stirred at 40C for 24 h. The sialidase-binding gel was poured into a glass filter and the liquid phase removed. To block the rest of the CNBr functional group, the antibody-binding gel was suspended in 30 ml of I M ethanolamine at 4"C overnight. In addition, the gel was washed on a glass filter with i 1 of o.i M acetate buffer, pH 4.0,
Correspondence: Dr Tsuneo Ozeki, The 3rd Department of Internal Medicine, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan.
I43
Tsuneo Ozeki et al. I44 containing 0.5 M NaCl and i I of O.I M this dry defatted liver powder by the method of Prockop (Prockop et al. I 96 7), i.e. i mg of sodium bicarbonate, pH 8.o, respectively. Finally, the gel was washed with i 1 of o. I M the powder was dissolved in i ml of water acetate buffer, pH 4.0. Subsequently, I 5 ml and I.0 ml of conc. HCI was added. The of the gel was suspended in 15 ml of 0.1 M solution was heated at ioo0C overnight. acetate buffer, pH 4.8. Ten milligrams of a1- Subsequently, Prockop's experimental proAT was added to the suspended solution and cedure was followed. the mixture was slowly stirred at 3 70C for 1 2 h. After enzyme digestion, the mixture was Electron microscopy centrifuged at 3000 r.p.m. at 40C and the supernatant dialysed against physiological Portions of livers were excised for electron saline for I2 h at 40C. After exchanging the microscopic examination. Samples were physiological saline, one further dialysis was fixed in a mixture of 2.5% glutaraldehyde carried out and the inner solution from the and 2% p-formaldehyde in 0.2 M cacodylate Visking tube was used for the animal experi- buffer (TAAB laboratories) at pH 7.4 for 2 h at 4TC, and then postfixed in cacodylatements. buffered 2% osmium tetroxide for 2 h at 40C. They were embedded in Epon 812 (TAAB Animal experiments laboratories) and ultrathin sections, cut with Male albino rats of the Wistar strain weigh- a diamond knife on a Porter-Blum MT-I ing approximately I 50 g were used. Animals ultramicrotome, were stained with uranyl acetate and Reynold's lead citrate (TAAB were divided into two sets of four groups, each consisting of seven rats. Groups 3 and 4 laboratories), and examined in a JEM-ioo of set A were treated with sialo-a1-AT, and CX electron microscope. Groups 3 and 4 of set B with asialo-a1-AT. Group I received no treatment. Rats in the Cell culture other groups were injected intramuscularly Human embryonal diploid lung fibroblasts with carbon tetrachloride (2 ml/kg body weight) twice a week for 8 weeks to induce (HEL) were kindly provided by the Departchronic hepatic injury. They were then ment of Cell Biology Research Institute for injected daily with i ml of physiological Tuberculosis and Cancer, Tohoku Universaline i.p. for 5 days (Group 2) or with o.i% sity, Sendai, Japan. The culture medium, from a flask in which HEL cells had been (Group 3) or 0.2% (w/v) (Group 4) ci1-AT (i ml/I50 g body weight), daily for 5 days. planted and stored, was removed by pipette Animals were killed by cutting the bilateral and the flask was then washed twice with jugular veins 24 h after the final injections. phosphate buffered saline (PBS) containing The liver was removed and portions saved for 0.02% EDTA, added to the flask in a 3 70C histological and electron microscopical water-bath for 30 s. A 0.3% trypsin solution was added to the cells remaining in the flask examination. For histology, some sections and the cells removed by repeated pipetting were stained with haematoxylin and eosin, of the trypsin solution. The cell-containing and others with azan. solution was collected and centrifuged at I000 r.p.m. for 3 min at room temperature. Hydroxyproline determinations The supernatant was decanted with a pipette The remainder of the liver was homogenized and the remaining cells suspended in 5 ml PBS. Cells were counted in a Fuchs-Rosenin methyl alcohol in a Potter-Elvehjem glass homogenizer, defatted with ether thal chamber and the cell number adjusted (methanol/ether, I: i, v/v) and dried in to 4.6 x 104 cell/ml with PBS containing EDTA. Subsequently, 1.5 ml of MEM medium vacuo. Hydroxyproline was determined in
Antitrypsin and hepatic fibrosis 145 (containing io% calf serum) was added to a there was a dose-related effect and animals Petri dish (35 mm in diameter) and 3 x IO4 which received the higher level of the asialo HEL cells were planted in the dish. The dish compound (Group 4) had nodular livers was incubated in an atmosphere of 5% CO2 with more marked changes than those in and 95% air at 3 70C for 6 h. The culture Group 3. medium was discarded and the cells washed The percentage ratios of liver weight to twice with PBS containing EDTA. The culbody weight were lower, and of spleen ture medium was then exchanged for RITC weight to body weight were higher, after 80-7 (no plasma protein) (Yamane et al. both the sialo and asialo types of a,1-AT than I98I). Subsequently, various concentra- in untreated animals or in saline controls tions ofoc1-AT (sialo type: o, I, 3, 10, 30, I00 treated with carbon tetrachloride (Tables i mg/I, asialo type: o, 10, 30 mg/l) were added and 2). to the culture medium. For each concentraThe carbon tetrachloride-treated, saline tion, three samples were used per group. The control animals in Group 2 had inflammacell cultures were incubated in 5% CO2 and tion and fibrosis around both the central 95% air at 370C for 5 days. At the end of this veins and portal tracts (Figs Ia and b). time cells were again removed by trypsinizaAdministration of sialo-a1-AT (set A anition and repeated pipetting. The cell suspension was centrifuged at I000 r.p.m. for 3 min at room temperature. The supernatant was decanted and the remaining cells suspended in 5 ml PBS. Cells were counted in a FuchsRosenthal chamber. All results are expressed as the mean ISD and statistical significance was
Table i. The effects of sialo-ae1-antitrypsin on the ratio of liver weight and spleen weight to body weight
assayed by Student's t test.
Results
Effect of a,-AT on chronic liver injury produced by carbon tetrachloride Histology. The livers from animals in Group 2, which had received carbon tetrachloride followed only by saline, were swollen and their surfaces were granular. When sialo-o1AT was given (set A, Groups 3 and 4), instead of saline, the liver surfaces were more uneven than those from Group 2 and there was a dose-related effect such that Group 4 animals which had been exposed to the higher dose of sialo-ax1-AT, had nodular livers with more obvious and marked pathological changes than those in Group 3. When asialo-oxI-AT was used (set B, Groups 3 and 4) instead of saline, the liver surfaces again were more uneven than those in the Group 2 saline controls but the degree of change was less severe than seen in the sialo-a1-AT-treated animals in set A. Again
Group Ai untreated A2 CC14+saline A3 CC14+sialo-ax-AT
A4 CC14+sialo-a1-AT
Percentage liver weight/ body weight
Percentage spleen weight/ body weight
5.9+0.5a
0.29+0.04c
5.4+0.5
0.3I+0.03
47 o.6b 4.5 ±0.7
0.35 +0.04d 0.37+0.04
a vs b, Po.oo5. c vs d, 0.025 > P>O.0I.
0.27±0.02c 0.28 + 0.03 0.32+0.04d 0.35 + 0.04
I46
Tsuneo Ozeki et al.
(a)
(b)
Fig. i. Section from carbon tetrachloride-treated, saline control animal (Group 2). Inflammation and slight fibrosis can be seen around the central zones and portal tracts. (a) H and E; (b) azan staining, x Ioo.
Antitrypsin and hepatic fibrosis (a)
a
I47 4.
W
(b)
A
-
~
Fig. 2. Section from a Set A Group 3 rat after treatment with sialo-cl -At. Necrosis and degeneration of hepatocytes and moderate fibrosis (portal-to-portal bridging) is illustrated. (a) H and E; (b) azan staining, x 100.
148
Tsuneo Ozeki et al.
Fig. 3. Section from a Set A Group 4 animal after sialo-al-AT. Advanced fibrosis and the presence of numerous nodules can be seen. H and E staining, x IOO.
mals) increased the inflammation and fibroand produced both central-to-portal and portal-to-portal bridging fibrosis (Figs 2a and b). The damage was more severe after the higher dose of sialo-ac1-AT (Fig. 3) than after the lower dose (Figs 2a and b). Asialo-a1-AT also increased the extent of hepatic inflammation and fibrosis in set B animals by comparison with saline controls but the effect was less marked (Fig. 4) than seen in set A animals after the sialo type (Fig. 3).
sis
Sub-cellular changes. Examination of the liver tissues by electron microscopy revealed cystic dilatation of the endoplasmic reticulum in the hepatocytes of animals treated with either sialo or asialo-a1-AT. In addition, the mitochondria were irregularly shaped and had fewer than normal cristae. Figure 5 illustrates typical changes seen in a set B animal treated with asialo-ac1-AT.
Hydroxyproline analysis. Both sialo and asialo-a1-AT increased the hydroxyproline content of the livers but the asialo type was less effective in this respect than the sialo type (Table 3). Hydroxyproline is present in high concentration in collagen and so these findings parallel the histological changes.
Fibroblast cell culture. There was a linear relationship between the planted cell number and incubation time for cell numbers between 3 x I04 and 4 x 104. For these experiments the number of cells planted and cultured was 3 x 104. Fibroblast (HEL) cell growth increased in proportion to the concentration of the sialo type of a1-AT in the culture medium (Table 4) but the effect of the asialo type was relatively weak (Table 5). Even with 30 mg/l of asialo-a1-AT the HEL cell growth was only slightly increased over that of controls without any added alAT.
Antitrypsin and hepatic fibrosis O-
(a)
(b)
149
1
v.
Fig. 4 Section from a Set B GrouP 4 animal after asialo-ci -AT. Advanced fibrosis is present in this section but the fibrotic bundles are thin. (a) H and E; (b) azan staining. x ioo.
Tsuneo Ozeki et al.
I50
p::e.
Fig. 5. Electron micrograph of the liver from a Set B rat treated with asialo-a1-AT. There is cystic dilatation of the endoplasmic reticulum in hepatocytes and the mitochondria are irregularly shaped and have fewer cristae than normal. L, Lipid droplet. M, mitochondria; E, cystic dilatation of ER; arrow, lysosome. x
1
5 000.
3. The effect of sialo and asialo types of a,antitrypsin on the amount of hydroxyproline in carbon tetrachloride-injured livers
Table
Experimental
group
A I. Untreated A2. CC14+ saline A3. CC14+ sialo-a,-AT A4. CC14+sialo-cx-AT B3. CC14+asialo-cx,-AT B4. CCI4 + asialo-x1-AT
Hydroxyproline (pg/mg dry liver)
Concentration of sialo-a1-AT
(pg/ml)
No. of cells x
I04
0
5.92 ± 0.83a
I
8.88±O.I33b
0.42 + o.o8a I .99 + 0.32b 2.98 + 0.3 8c 3.95 +o.4 Id
3
9.96±o.ig
10
II.17±0.80c
2.50+0.36e 2.98 + 0.3 3f
30 100
I4.28 I.29 150.3 3.90
a vs b, PP. b vs c, o.oi > P>o.oos.
Antitrypsin and hepatic fibrosis Table 5. Effect of aI-antitrypsin (asialo-type) on the growth of HEL cells in vitro Concentration of
asialo-cel-AT (Ug/ml)
No. of cells x I04
Io
5.92±0.I6a 6.45 ±0.36b
30
6.56±o.36c
0
a vs b, o.i0 > P>o.o5. a vs c, 0.05 > P>0.025.
Discussion The data presented here show that administration of sialo-a1-AT to rats with chronic carbon tetrachloride-induced injury, accelerates liver fibrosis in vivo. Asialo-a1-AT also causes hepatic fibrosis though the reaction is less severe than after the sialo type. Both types of a1-AT also directly damage hepatocytes and cause both marked degenerative subcellular changes and cell necrosis. It is noteworthy that more severe fibrosis and associated cell damage is seen after either type of axl-AT treatment than in saline controls, even in animals already in the acute stage of carbon tetrachloride-induced liver inflammation. Sialo-a1-AT also stimulates the growth of fibroblasts (HEL cells) in vitro; in this particular study the doubling time was relatively slow and cell growth in the absence of plasma protein was poor but nevertheless, the stimulatory effect of sialo-al-AT was quite clear. By contrast, in the same experiments, asialo-a1-AT had very little accelerating effect on HEL cell growth. It is logical that, in vivo, direct stimulation of fibroblast growth by sialo-a1-AT could contribute to the observed increase in hepatic fibrosis even in acutely inflamed livers, but it is more difficult to explain the increased fibrosis induced by the asialo type. One explanation could be that, in vivo, there may be some conversion in the liver of the relatively
I 5I inactive asialo type to the more active sialo type of a,1-AT. A similar situation has been reported following clinical observations in patients with emphysema and liver cirrhosis in whom a,-AT deficiency has been reported (Laurell & Eriksson I963). It is known that human serum a1-AT shows genetic polymorphism and the recognized determinants are designated PiMM and PiZZ; the PiZZ type, which currently is thought to be the asialo type, is found in patients with liver cirrhosis (Fagerhol I968). Laurell and Eriksson (1963 ) reported the PiZZ type of a1-AT to be a very insoluble protein and considerable amounts are thought to be deposited in the liver in cases of a1-AT deficiency. This suggests that in those patients where there is a deficiency in total a1-AT activity, the relatively insoluble asialo or partially desialo types of a1-AT are synthesized in the liver and, while some will be released into the circulation from necrotic cells, a proportion will be deposited subcellularly within the endoplasmic reticulum. Such deposits would directly account for the liver (hepatocyte) damage observed and can be assumed to be indirectly responsible for the fibrosis. Another component of the a1-fraction of human serum protein is orosomucoid which it is known can promote fibroblast cell growth (Ozeki I986). It might be suggested, therefore, that it is orosomucoid rather than a1-AT which is responsible for the fibrosis seen in chronic hepatitis. However, this is unlikely in view of both the relative amounts of orosomucoid (less than i o%) and of a,1-AT (90%) present in the a1-serum protein fraction, and the stronger fibroblast-promoting activity of a1-AT by comparison with orosomucoid (Ozeki I986). We suggest that liver cirrhosis may be symptomatic of an overall deficiency of serum a1-AT accompanied by deposition of abnormal forms of the protein within the liver.
References ASOFSKY R. & THORBECKE G.J. (I96I) Sites of formation of immune globulins and of a compo-
I S2
Tsuneo Ozeki et al.
nent of C'3. II Production of immnoelectrophoretically identified serum proteins by human and monkey tissues in vitro. J. Exp. Med., I I4, 471-
483.
CLAMP J.R. (I975). In The Plasma Proteins. Volume 2. Second edition. Ed. F. W. Putnam. New York: Academic. pp. I 63-2 I 1. COURTOY P.J., LoMBART C., FELDMANN G., MoGuILEUSKY N. & ROGIER E. (I 98 I) Synchronous
increase of four acute phase proteins synthesized by the same hepatocytes during the inflammatory reaction: a combined biochemical and morphologic kinetics study in the rat. Lab. Invest. 44, I05-I I 5. FAGERHOL M.K. (I968) The Pi-system: genetic variants of serum o1-antitrypsin. Ser. Hematol. I, I53.
LAURELL C.B. & ERIKSSON S. (I963) The electrophoretic a,-globulin pattern of serum in a-
antitrypsin deficiency. Scand. J. Clin. lab. Invest., 15, I32.
OzEKI T. (1986) Orosomucoid as the accelerator of hepatic fibrosis. Br. J. exp. Path. 67, 731-736. PROCKOP D.J., UDENFRIENDS S., KARI I.K. & Ossi L.P. (I967) Modifications of a specific assay for hydroxyproline in urine. Anal Biochem. I9, 249-255.
PRUNIER J.H., BEARN A.G. & CLEVE H. (I964) Site of formation of the group specific component and certain other serum proteins. Proc. Soc. Exper. Biol. Med. 115, 1005. SCHULTZE H.E., HEIDE K. & HAUPT, H. (I962) 0c1Antitrypsin aus Humanserum. Klin. Wshr. 40, 427-429. YAMANE I., KAN, N., HOSHI H. & MINAMOTO Y. (I98I) Primary culture of human diploid cell and its long-term transfer in a serum-free medium. Exp. Cell. Res. 134, 470-474.