- Email: [email protected]
Higher levels of aminoterminal type III procollagen peptide, and laminin in alcoholic in nonalcoholic cirrhosis of equal severity
Jolarnalof Hepatoalogy,1992 14: 71-77 fj?J1992 Elsevier Science PublishersB.V. All lights reserved.0168-8278/92/$03.50
71
HEPAT 00904
E. Eotterd,
A.
‘Department offclinic~dPharmacoiogy, Universityof GoettingeTand ‘Department of Clinical Chemistry and Central Laboratory, Universiryof Marburg, Federal Republic of Germany {Received 7 September 1390)
In vitro models have shown t at metabolites of ethanol (acetaldehyde and lactate) stimulate collagen synthesis, thereby,
they may be important as fibrogenic mediators. The relevance of these findings for fibrogenesis in the human liver in vivo, however, has not as yet been demonstrated. Serum markers for collagen (PIIINP, using radioimmunoassays employing polyclonal antibodies and Fab-fragments (PIIINP-Fab), respectively) and basement membrane (laminin) metabolism were therefore investigated in 25 alcoholic cirrhotics (Pugh-Score: 6.7 -+ 1.9 SD.) and in 19 comparable nonalcoholic cirrhf*tics (Pugh-Score: 6.3 zb 1.5, n.s.) with only slight evidence for inflammation: COT 28 & 22 vs. 24 + 21 Wl; GPT 24 + 23 vs. 31 + 28 U/l; y-globulins 24 + 8 ‘4s. 22 +_5%, respectively {all n.s.). Severity of tbe disease was assessed by quantitative liver function tests. Levels of PIIINP, INP-Fab and laminin measured by RIA were 21 C 19ptgn, 90 f 42&l and 2.5 f: 0.8 U/ml in alcoholic cirrhosis and 10 + 6&l, 61 f 10&l and 1.9 -C0.4 U/ml in nonalcoholic cirrhosis, respectively (all p K 0.01). Differences on PIIINP and PIIINP-Fab remained significant even after accurate matching for galactose elimination capacity, aminopyrine breath test and hepatic sorbitol clearance. Laminin levels were higher in alcoholic cirrhosis only after matching for the hepatic sorbitol clearance (p < 0.01). The higher levels of serum markers for collagen and basement membrane metabolism in alcoholic vs. nonalcoholic patients with cirrhosis at equal severity of the disease and with only minimal signs of inflammation may be the clinical reflection of a specific fibrogenic effect of ethanol metabolites. suggestirng that
Ethanof-induced liver injury may trigger a fibrotic response in which perivenulat fibrosis is associated with proliferation of myofibrobiasrs, the mosi common type of mesenchymal cell around the perivenular areas in both the normal and the diseased liver (1). Metabolites of ethanol, such as acetaldehyde and lactate, have been shown to stimulate collagen synthesis in *znious celi cuiture systems (2-6). Increases of collagen synthesis and of messenger RNA have been described in human fibroblasts and in the liver of baboons given acetaldehyde or ethanol, respectively (7,8). Voss et al. (9) also observed cells similar to myofibroblasts in vitro in the outgrowth of explants of fi-
brotic livers. Sush cells synthesized collagen types I, III, IV, and V. Taken together, these investigations suggest a specific fibrogemc effect of ethanol or its metabolites. The relevance of these in vitro findings for in vivo fibrogenesis in the human liver has not as yet been demonstrated satisfactorily. In various liver diseases the serum markers for collagen and basement membrane metabolism, such as serum concentrations of the aminoterminal type III procollagen pe@e (PIIINP and PIIWB>-Fab, using polycPonal iantibodies and Fab-fragments, respectively), and of laminin have been demonstrated to exhibit significant correlations with activity of the fibrogenic pro-
* Present address: University of Witten-Herdecke, Bcckweg 4, D-5604 Herdecke, F.R.G. Correspondence andpresent address: Dr. E. Letterer, M.D., Department of Medicine H,University of Erlangen-Niimberg, 12, D-5520 Erlangeo, F.R.G.
Krankenhausstrasse
E. LOITERER
72
cess (10-26). Their plasma levels, however, may not depend only on collagen formation! but also on the disposition of these markers by hepatic or renal elimination ‘(27,28). In order to investigate the role of ethanol in fibrogenesis in intact man, the serum markers in patients with alcoholic and with nonalcoholic cirrhosis who are well defined in regard to both inflammatoy activity and severity of organ dysfunction must be compared. This study is an attempt to achieve such a goal.
Material and Methods Patients
This study is based on a group of 25 patients with alcoholic cirrhosis and another group of 19 patients with nonalcoholic cirrhosis (Table 1). This second group consisted of 12 patients with hepatitis B virus (HBV)- and seven patients with hepatitis NANB virus (HNANB)-related liver disease. All patients were clinically stable cirrhotics. There had been no ascites or upper GI-hemorrhage for 41 months (median, range: 8-79 months) before the study and patients had received biopsies within the last 4 years. The diagnosis of alcoholic cirrhosis was established by the following criteria:
et al.
- alcohol consumption of more than 60 g per day for females and more than 80 g per day for males for at least 10 years, and abstinence for at least 6 months before the study, as documented by negative alcohol plasma concentrations at each visit; - routine laboratory tests compatible with alcoholic cirrhosis; - sonography with typical alterations of the liver and lack of abnormalities of extrahepatic bile ducts; - liver biopsy showing cirrhosis and lesions compatible with chronic alcohol consumption in 18 patients (29) and oesophageal varices or ascites in the others. The diagnosis of nonalcoholic cirrhosis, i.e., viral-associated cirrhosis, was established by the following criteria: - history of a transfusion or another pathway for infection with hepatitis B- or NANB-virus; - elevated serum aspartate aminotransferase (AST, GOT), alanine aminotransferase (ALT, GPT) for more than 6 months; - for hepatitis B: IgG-Anti-HBs, IgG-Anti-HBc and lack of HBs-antigen, IgM-Anti-HBs and IgM-Anti-HBc; - for hepatitis NANB: negative serology of hepatitis A and B, Anti-HC was not sought; - alterations associated with cirrhosis and no abnormalities of extrahepatic bile ducts by ultrasonography; - liver histology showing lesions associated with cirrhosis and/or chronic inflammation in 14 patients (29).
TABLE 1 Characterization of the patients with liver cirrhosis (median, range)a Alcoholic cirrhosis Number of cases Sex (m/f)b Age (years)
Nonalcoholic cirrhosis
25 2213
19 1118
(::-67)
(G-74)
(Z-115) 18
(i-96) 14
(56-12)
(5611)
Body weight (kg) Cirrhosis on liver biopsy Pugh-Score Oesophageal varices (stage I-IV) History of oesophageal variceal hemorrhageC Splenomegaly’ History of ascitesC Treatment of ascites at the time of study
p
co.05 n.s. ns ns ns
Design of the study
All patients were investigated as out-patients in the liver clinic of the Division of Gastroenterology and Endocrinology, University of Goettingen. At the initial visit clinical and laboratory evaluation was conducted before further tests. Treatment modalities were adapted to the individual needs of each case. Patients with alcoholic cirrhosis were encouraged to stop drinking. Investigations (i) Specimen collection of blood for determination of
20
7
co.01
connective tissue parameters in serum. After allowing the
9 16 13
3 7 2
n.s. co.05 co.01
11
2
co.05
blood to clot for 1 h at room temperature serum was prepared by centrifugation (15 mitt, 5000 x g) and stored at -70 “C for a maximum of 1 month before the assays were performed.
a Some of the patients were treated with following drugs: acetylcysteine, atenolol, captopril, colestyramine, digitoxin, digoxin, famotidine, ferrum, glibenclamide, lactulose, nifedipine, ranitidine, sucralfate, theophylline, verapamil, vitamine K. Furosemide and spironolactone were used for treatment of ascites. b When calculated by the stastistical tests there is not any significant difference between male and female patients between both groups. ’ Oesophagtal varices (stage I-IV) and oesophageai variceal hemorrhage were verified by gastroscopy, ascites and splenomegaly were detected clinically and sonographically.
(ii) Determination of N-terminal propeptide of type III procollagen
(PZZINP). Method A: The propeptide was measured in various dilutions of serum using a radioimmunoassay (Behring-Hoechst, Frankfur&iMain, F.R.G.) based on intact polyclonal antibodies recognizing mainly the higher molecular mass (intact) propeptide Co1 1-3 (10,30). The intraday coefficient of variation (CV) l’vas 4.9% (at 12,@1; n = 14), the interday CV was 7.2% (:at 5
-l-YPEIii’ PROCOLLAGEN
PEPTIDE
AND LAMININ
IN CI
&I; n = 12). A reference (normal) range of PIIIN rum 0:2-12&l was accepted (18). Method B: This method based on a recently described radioimmunoassay employing the Fab-fragment of the antibody to PIIINP rather than intact antibody (31). The Fab-fragment has equal affinity for Co1 1-3 and Co1 1 and, hence, measures intact propeptide and its globular fragment with parallel inhibition profiles. The intraday CV was 6.9% (at 26 pg/l; R = 8), the interday CV was 5.2% (at 20 @gl; n = 7). The reference range of PIIINPFab was 15-63bg/l(31). (ii!) Detemindon oflaminin. Concentrations of the glycoprotein were determined with a competitive radioimmunoassay (Behring-Hoechst, FrankfurtMain, F.R.G.). The analytical del-ails have been reported previously (32,33). The antibody is directed against antigenic determinants of the pep&r-resistant fragment Pl of laminin (32). The CV of int,-aday and interday reproducibility were 3.2% and 4.9% (at a concentration of 2.0 U/ml), respectively. The concen:!ration of laminin is expressed in arbitrary U/ml since an International standard of laminin is not available: 1 U is defined as the mean quantity of Iaminin per ml serum in a group of obviously healthy volunteers, and corresponds to about 0.23 pg (32). The reference range’of serum lamini:? in our laboratory is 1.1-1.9 u/ml (34). (iv) Quantitative fiverjimction test. The galactose elimination capacity (GEC) was ca’culated from venous blood samples taken between 20-60 min at 5 min intervals after i.v.-injection of 0.5 g/kg galactose as described by Tygstrup (35). A factor of 5 min was used for correction of uneven distribution of galactose iu the body (28). For the aminopyrine breath test (ABT) samples were examined I;0 min after iv.-injection of 1 mg/l.5&i dimethylamine]14C]aminopyrine according to Miotti et al. (36). The hepatic sorbitol clearance (SCI) which represents a measure of hepatic parenchymal plasma flow was determined as described by Zeeh et al. (37). All blood samples were taken under fasting conditions. Routine biochemical and hematological examinations were carried out with automated techniques. Alcohol plasma levels were determined enzymatically and were below the limit of detection in all patients at the time of the investigation. Written informed consent was obtained from each participant in accordance with the declaration of Helsinki, revised in Tokyo in 1975. Statistical analysis The Wilcoxon test for unpaired data and the Chi-square test were used for statistical comparisons. Both tests were corrected for multiple testing by the Friedman rank test.
73 earman rank test was used for calculation of correlations. The level of significance wasp < 8.05. esults Initially the group of 25 patients with alcoholic cirrhosis was similar to the group of 19 patients with viral-associated cirrhosis in age (52 + 8 VS. 52 + 12 years, n.s,), body weight (76 + 14 vs. 73 f 14 kg, n.s.), Pugh-Score (6.7 + 1.9 vs. 6.1 + 1.7) and drug intake (Table 1). Tbe frequency of oesophageal varices, ascites, and splenomegaly were significantly higher in the alcoholic group (Table 1). The routine laboratory tests are summarizad in Table 2. In the group with alcoholic cirrhosis levels of alkaline phosphatase, y-GT and serum bilirubin were signiflcantly higher than in the patients with nonalcoholic cirrhosis, whereas pseudocholinesterase activity and prothrombin time were significantly lower. No significant differences were found in regard to GOT, GPT, albumin, yand serum-creatinine. The globulins, thrombocytes values of sodium, potassium, urea, cholesterol, triglyTABLE
2
Routine laboratory tests in patients with alcoholic and nonalcoholic cirrhosis (median, range) Tests (reference range)
Alcoholic cirrhosis
Nonalcoholic cirrhosis
GOT (<17 U/l)
&3)
(8-99)
GPT (~23 U/l)
&J6) (C96)
y-GT ( ~28
16 (8-247)
(E351)
A,lkaline phosphatase (so- 200 un)
184 (70-585)
117 (61-362)
co.05
Serum-bilirubin (t1.2mgtdl)
(EL.!)
eo.05
(::63-27.2)
y-Globulins (lP-20%)
(Z-37)
(E-30)
Pseudocholinesterase (2300-7400 U/l)
3100 (7OO-BOo)
4800 (1400-9600)
Albumin (3.5-5.0g/dl)
g-4.6)(2.9-5.2)
ns.
(Lm)
co.01 (Z-100)
Thrombocytes (150-300.103Ij.q
116 (33-291)
141 (40-312)
n.5.
Creatinine (0.6-l .3 mg/dl)
1.0 (0.6-1.3)
n.s.
(E-1.3)
18
un)
Prothrombin (70- 100%)
p ns.
n.5.
n.5.
4.0
time
co.01
E. LOTIERFX
74 ceride and haennotologic parameters were within the normal range. me median ABT was 0.22 %dose.kg per mmolC@ (0.01-1.13) in patients with alcoholic cirrhosis and 0.55 %dose+kgper mmolC0, (0.13-1.18;~ C 0.01) in patients with nonalcoholic cirrhosis. In contrast, for the GEC no significant difference between the groups was observed (3.7 mg/min per kg (1.5-5.2) vs. 4.1 mg/min per kg (2.1-6.83, respectively). Levels of the SC1 were significantly lower in the alcoholic group than in the nonalcoholic group (5.8 ml/min per kg (2.4-11.6) vs. 7.3 ml/min per kg (3.4-10.6), p c 0.05). The renal sorbitol clearance, thought to express glomerular filtration rate (38) was 1.3 ml/min per kg (0.4-2.1) and 1.4 mYmin per kg (0.6-2.0, n.s.), respectively. In patients with alcoholic cirrhosis serum concentrations of PIIINP, PIIINP-Fab and laminin were 19 ,@l (5-SO), 79 pg/l (36-193) and 2.6 U/ml (1.4-4.1) respectively, i.e., significantly higher than the values of patients with viral-associated cirrhosis (8 pg/l, 4-25, 59 @I, 49-83, and 1.9 U/ml, 1.1-2.7, respectively, all p < 0.005). The median ratio of PIIINP-Fab/PIIINP levels was 5.7 (1.9-10.6) in alcoholic and 7.8 (3.8-14.8) in nonalcoholic cirrhotics (p c 0.01). In order to compare serum levels of PIIINP, PIIINPFab and laminin of both groups at equal liver function a matching procedure was applied to the data as shown in Table 3. After matching for the Pugh-Score, significantly elevated concentrations of all three parameters were found in alcoholic patienis compared to nonalcoholic patients. After matching for the GEC and the ABT, only
80
1
TABLE
et al.
3
Matching of patients with liver cirrhosis for Pugh-Score, GEC, ABT, and SC1 (median, range)a Alcoholic cirrhosis
Nonalcoholic cirrhosis
P
a) Matching for Pugh-Score (n = 13) PIIINP @il) 15 (5-46) PIIINP-Fab (J@) 79 (36-115) Laminin (U/ml) 2.5 (1.4-3.8) GOT (U/l) 20 (8-113) 16 (g-116) GPT (U/l) y-Globulins (%) 22 (10-31)
19 (4-24) 56 (49-80) 1.7 (1.2-2.4) 12 (8-99) 20 (5-96) 21 (15-30)
co.05 co.05 co.05 n-s. ns. ns.
b) Matching for GEC PIIINP (&I) PIIINP-Fab @g/l) Laminin (U/ml) GOT (U/1) GPT (U/l) y-Globulins (%)
(n = 14) 16 (S-24) 74 (39-115) 2.4 (1.4-3.8) 20 (8-37) 17 (8-48) 23 (10-34)
9 (4-24) 60 (49-80) 2.0 (1.2-2.6) 26 (8-99) 28 (5-96) 24 (14-30)
co.05 co.05 n.s. ns. ns.
c) Matching for ABT PIIINP f&g/l) PIIINP-Fab (qil) Laminin (U/ml) GOT (U/I) GPT (U/l) y-Globulins (%)
(n = 12) 18 (5-26) 82 (36-147) 2.5 (1.4-2.9) 21 (10-41) 19 (13-53) 22 (10-35)
9 61 2.0 21 22 23
co.05 co.05 n.s. n.s. n.s. n.5.
d) Matching for SC1 (n = 13) PIIINP &g/l) 21 (5-33) PIIINP-Fab (.@l) 88 (36-147) Laminin (U/ml) 2.6(1.4-3.8) GOT (U/l) 20 (10-113) GPT (U/l) 16 (7-116) v-Globulins (%) 23 (15-35)
(4-24) (49-83) (1.2-2.7) (10-99) (6-96) (15-30)
10 (4-t4j 59 (49-80) 1.9(1.2-2.6) 16 (8-99) 20 (5-96) 22 (18-28)
-
l-l.S.
co.02 co.01 co.01
n.s. ns. n.s.
a Matching for Pugh-Score (+O), GEC (+O.l mg/min per kg), ABT (+0.05 %dose.kg per mmol CO,), and SCI (20.2 ml/mm per kg),
a *
I 60
R,=-O.S6 p (
R,=-3.46
0.01
a
p c 0.002
7 2 A 0 4o z = 0 20
0
--0
O+l-------
01
2
4
G E C [mg/min/kg]
6
Cl
2
4
G E C [mg/min/ka]
6
2
4
6
G E C [mg/min/kg]
Fig. 1. Spearman rank correlations (R,) between the galactose elimtnation capacity (GEC) and PIIINP, PIIINP-Fab, and laminin in patients with alcoholic cirrhosis (n = 25, and in patients with nonalcoholic cirrhosis (n = 19, A). Horizontal lines are the upper limits of the corresponding normal ranges.
TYPE III PROCOLLAGEN
PEPTIDE
AND LAMININ
IN CIRRHOSIS
75
80-s R,s-0.56
a,=-0.45 P
c
p c O.OOI
0.002
Oi
0.0
1 .o
0.5
0.0
0.5
ABT[E]
1 .o
-0.0
1 .o
0.5
ABT[E]
ABT[E]
Fig. 2. Spearman rank correlations (R3 between the aminopyrine breath test (ABT) and PIIINP, PIIINP-Fab, and laminin in patients with alcoholic cirrhosis (n = 25, ) and in patients with nonalcoholic cirrhosis (n = 19, A j. Horizontal Lnes are the upper limits of the corresponding normal ranges.
80
0
4 S C
8
I [ml/min/kg]
12
0
4 S C
8
I [ml./min/kg]
12
0
4 S C
a
12
I [ml/min/kg]
Fig. 3. Spearman ra& correlations (R,) between the hepatlc sorbitol clearance (SCl) and PIIINP, PIIINP-Fab. and laminin in pat:ents with alcoholic cirrhosis (n = 25. apd in patients with nortalcoho!ic cirrhosis (n = 19, A). Horizontal lines are the upper limits of the corresponding normal ranges.
PIIINP- and PIIINP-Fab levels were significantly higher in the alcoholic groilp than in the no,lalcoholic group, whereas for laminin no significant differences were detected. In contrast, after matching for SC1 all three parameters showed significant differences between both groups. As shown in Figs. l-3 significant, negative correlations were found between serum concentrations of PIIINP, PIIINP-Fab, and laminin and the GEC (all p < O.Ol), the AI3T (all p : 0.002), and the SC1 (all p < 0.05). PIIINP was positively correlated with PIIINP-Fab (R, = 0.58) and laminin (R,= 0.63), and PIIINP-Fab with laminin (R, = 0.63; all p < 0.0001). In agreement with other investi-
gations (10,12,20,23,39) we found in the group with nonalcoholic cirrhosis PIIINP positively or negatively correlated to GOT (R, = 0.66), y-GT (R, = 0.68), serum bilirubin (R!; = 0.67), y-globulins (R, = 0.73). albumin (R, = -0.73), pseudocholinesterase (R, = -OH), prothrombin time (R, = -0.75), and thrombocyte count (R, = -0.64, all p < O.Ol), whereas laminin was related only to GOT, alkaline phosphatase, and albumin (all R,values were > f 0.64, all p < 0.01). In the group with alcoholic cirrhosis PIIINP, PIIINP-Fab and laminin concentrations showed a significant negative relationship only to serum albumin respectively (all R, values and pseudocclolinesterase, were 2 -0.52, allp < 0.01).
E. LOTTERER
76 Discussion The data of this study seem to show that PIIINP, PIIINP-Fab, and laminin are higher in alcoholic cirrhosis than in nonalcoholic cirrhosis, even when data are controlled for the severity of the disease and for the degree of inflammation. Consequently, these results support the hypothesis that in alcoholics acetaldehyde and other ethanol metabolites exert fibrogenic activity over and above the fibrogenesis which results from hepatocyte injury. me results of our study extend concepts derived from animal and tissue culture experiments to explain the development of alcoholic cirrhosis in man (2-8). Before accepting these conclusions, however, the methodological limitations have to b: analysed in detail. A relation between the investigated connective tissue markers and fibrogenesis in the liver is well documented in literature (lo-26,39). Serum levels always depend on at least two factors, however, the rate of formation and the rate of elimination of the compottnd under consideration. Consequently, when comparing serum markers of two groups of patients it is important to control for alterations in the function of organs which may be involved in the disposition of serum aminoterminal type III procollagen peptide and laminin (40). In our study both groups of cirrhotics had equal serum creatinine concentrations and renal sorbitol clearances and both groups were matched for GEC, ABT, and SC1 (Tables 2 and 3). Consequently, in this study it was possible to investigate patients with alcoholic and nonalcoholic cirrhosis. Since these patients had comparable renal and hepatic functions, it may therefore be reasonable to assume that processes responsible for the disposition of the serum markers were adequately controlled. On this basis it was concluded that the higher serum markers in alcoholics are likely to be due to an increased rate of formation (21,23,24,41). It has also been well documented that the degree of hepatocellular damage and inflammation within the liver correlates with elevation in the investigated serum markers of fibrogenesis (10,12,17,19-21,23,39). This was confirmed in our patients with nonalcoholic cirrhosis since the levels of PIIINP were correlated with GOT and y-globulins, and the levels of laminin with GOT. It seemed important therefore to compare patients with similar signs of inflammation. There were no differences in inflammation values in the total groups or after matching (Table 3). The data, therefore, support the hypothesis that intrahepatic cell damage and inflammation was slight in our patients and comparable among both groups. Consequently, the observed differences between alcoholic and nonalcoholic cirrhotics in PIIINP, PIIINP-Fab, and laminin cannot be
et al.
explained by differences in hepatocellular damage or inflammation. In this study the conclusion tl=i t increased serum levels of connective tissue markers and presumably increased fibrogenesis in alcoholic cirrhosis may be related to the toxic effects of acetaldehyde an other ethanol metabolites, is based on an exclusion of possible alternative explanations. Although these exclusions can never be exhaustive, the data appear to be fully compatible with in vitro experiments which have shown that acetaldehyde leads to fibrogenesis with increases in collagen synthesis and messenger RNA in various cell culture systems (2-9). The concentrations of acetaldehyde used in these studies was in accordance with concentrations found in the liver during alcohol abuse (2-8). Together with this experimental evidence our findings support the idea that acetaldehyde may also be a specific factor for collagen synthesis in the liver and thereby may contribute to the pathogenesis of alcoholic cirrhosis in man. Al?hough it might have been interesting to investigate hepatic fibrosis more directly, i.e., the degree of hepatic fibrosis evaluated accurately by liver biopsy, microscopic examinations could only provide a static view. Information concerning the dynamics of hepatic fibrogenesis might be better obtained by direct determination of changes in amounts and the molecular composition of various types of connective tissue. These studies, however, are difficult to execute. An interesting detail of our investigation remains difficult to explain, i.e., why the matching for GEC and ABT eliminated differences in laminin levels, whereas matching for SCl made the difference between alcoholic and nonalcoholic cirrhosis highly significant (Table 3). Taken at face value this may imply that the severity of the disease, as reflected by hepatocyte mass (GEC and ABT), is not a pertinent determinant of laminin. In contrast, limitations of the vascular bed to equal functional liver plasma flow values results in higher laminin concentrations in alcoholics. This could suggest that more basement membrane is formed in alcoholic cirrhosis which may be due to a higher degree of capillarization of sinusoids (24). Although speculative, these changes could again be related to acetaldehyde or other ethanol metabolites.
Acknowledgements The authors thank the nurses Mrs. A. Marx and Mrs. M. Suhr-Gottschaldt for their care for the patients and Mrs. C. Guthardt and Mrs. H. Steinmetz for their technical assistance.
TYPE III PROCOLLAGEN
PE
DE AND LAMININ
IN CIR
77
I. Nakano M, Worner TM, Lieber CS. Perivenuiar fibrosis in al& hofic injUrY: ultrastructure and histologic progression. Gastroenterology 1982; 83: 777-85. 2 Savolainen ER, Leo MA, Timpl R, Lieber CS. Acetaldehyde and lactate stimulate collagen synthesis of cultured baboon liver myofibroblasts. Gastroenterology 1984; 87: 777-87. 3 Holt K, Bennett M, Chojkier M. Acetaldehyde stimulates collagen and noncollagen protein production by human fibroblasts. Hepatology 1984; 4: 843-8. 4 Moshage H, Casini A, Lieber CS. Acetaldehyde selectively increases collagen-synthesis in cultured fat-storing cells but not in hepatocytes. Hepatology 1953; 10: 630. 5 Tsutsumi M, Takade A, Takas? S, Siiiimanaka K, Enyama EC.Effects of ethanol and its metabolites on collagen-synthesis by CL& tured rat-liver cells. J Gastroenterol Hepatol1989; 4: 229-40. 6 Kamegaya K, Qkazaki I, Maruyama K, et al. Ethanol, acetaldehyde and lactic-acid stimulate type-III collagen-synthesis of cultured chang liver-cells. Hepatology 1988; 8: 1435. 7 Brenner DA, Chojkier M. Acetaldehyde increases collagen gene transcription in cultured human fibroblasts. J Biol Chem 1987; 262: 17690-5. 8 tirn Ua,, Leo MA., Giambrone MA, Lieber CS. Increased type I procollagen mRNA !eve!s and in :r+_:o Fro:& qj&esis & &e baboon model of chronic alcoholic Liver disease. Gastroenterology 1985; 89: 1123-31. 9 Voss B, Rauterberg .I, Pott G, et al. Nonparenchymal cells cultivated from explants of fibrotic liver resemble endothelial and smooth muscle cells from blood vessel walls. Hepatology 19B2; 2: 19-28. 10 Rohde H, Vargas L, Hahn EG, Kalbfleisch H, Bruguera M, Timpl R. Radioimmunoassay for type III procollagen peptide and its application to human liver disease. Eur J Clin Invest 1979; 9: 451-9. 11 Raedsch R, Stiebl A, Waldherr R, et al. Procollagen type IIIpeptide serum concentrations in chronic persistent and chronic active hepatitis and in. irrhosis of the liver and their diagnostic value. Z Gastroenterolqie 1982; 20: 738-43. 12 Frei A, Zimmermarln PI Weigand K. The N-terminal propeptide of collagen type III iE sr rum refIects activity and degree of fibrosis in patients with CL~ ~tnit liver disease. Hepatology 1984; 5: 830-4. 13 Weigand K, Zaugg P-Y, ,:rei A, Zimmermann A. Lone-term follow-up of serum N-teti al propeptide of collagen type III levels in patients with chronic li ver disease. Hepatology 1984; 4: 835-8. 14 Savolainen E-R, Goldbtq B, Leo M.4, Velez M, Lieber CS. Diagnostic value of semi procoiiagen peptide measurement in alcoholic liver disease. Alcoholism 1984; 8: 384-9. 15 Tanaka Y, Minato 9, Hasumura Y, Takeuchi J. Evaluation of hepatic fibrosis by serum proline and amino-terminal type III procollagen peptide !evels in alcoholic patients. Dig Dis Sci 1985: 31: 712-7. 16 Schuppan D, Dumont JM, Kim KY, Hennings G, Hahn EG. Serum concentration of the aminoterminal procollagen type III peptide in the r& reflects eaily formation of connective tissue in experimental livc,r cirrhosis. J Hepatoll986; 3: 27-37. 17 McCullough AJ, Stassen WN, Wiesner RH, Czaja AJ. Serial determinations of the amino-terminal peptide of type III procollagen in severe chronic active h.:patitis. .I Lab Clin Med 1987; 109: 55-61. 18 Niemelij 0, Ristcli L, Sotaniemi EA, Risteii J. Aminoterminal propeptide of type III procollagen in serum in alcoholic iiver disease. Gastroenterology 1983; 85: 254-9. 19 Torres-Salinas M, Par& A, Caballeria .I, et al. Serum procollagen type III peptide as a marker of hepatic fibrogenesis in akoholic%epatitis. Gastroenterology 1986; 90: 1241-6. 20 cban.g -I-I’, l_imHC, Lee SD, et al. Clinical significance of serum
type-III procollagen aminopropeptide in hepatitis B virus-related liver diseases. Stand J Gastroenterol1989; 24: 533-8. Annoni G, Colombo M, Cantaluppi MC, Kbiat B, Lampertico P, Rojkind M. Serum type III procollagen peptide and laminin (Lam-PI) detect alcoholic hepatitis in cl.:rJnic alcohol abusers. Hepatology 1989; 9: 693-7. van Zanten RAA, van Leeuwen RE, Wilson JHP. Serum proc& lagen III N-terminal peptide and laminin PI fragment concentrations in alcoholic disease and primary biliary cirrhosis. Clin Cbm Acta 1988; 177: 141-6. McCullough AJ, Stassen WN, Wiesner RI-I, Czaja AJ. Serum type III procollagen peptide concentrations in severe chronic active hepatitis: relationship to cirrhosis and disease activity. Nepatology 1987; 7: 49254. :Niemefl 0, Risteli L, Sotaniemi EA, Risteli J. Type IV collagen and laminin-related antigens in human serum in alcoholic liver disease. Eur J Clin Invest 1985; 15: 132-7. Sato S, Nouchi T, Womer TM. Lieber CS. Liver fibrosis in altoholics. Detection by Fab radioimmunoassay of semm procoUagen III peptides. J Am Med Assoc 1986; 256: 1471-3. ‘Kropf J, Gressner AM, Negwer A. Efficacy of serum laminin measurement for diagnosis of fibrotic liver diseases. Clin Chem 1988; 34: 2026-30. Hahn EG. Blood analysis for liver fibs&s. .I >!*r\atol 1984; 1: 67-73. Bircher J. Quantitative assessment of deranged hepatic function: a missed opportunity? Sem Liver Dis 1983; 3: 275-84. Scheuer PJ. I iver Biopsy Interpretation. 3rd Edn, London: BailIi&e Tindall, 1481. Niemalti 0. Radioimmunoassays for type III procollagen aminoterminal peptides in humans. Clin Chem 1985; 31: 1301-4. Rohde H, Langer I, Krieg T, Timpl R. Serum and urine ana&sis of the aminoterminal procollagen peptide type III by radioimmunoassay with antibody Fab fragment. Collagen Rel Res 1983; 3: 371-9. Brocks DG, Strecker H, Neubauer HP, Tiipl R. Radioimmunoassay of laminin in serum and its application to cancer patients. Clin Chem 1986; 32: 787-91. Pick-Kober KH, Negwer A, Gressner AM. Determination of laminin in serum by a competitilve radialmmunoassay directed against the pepsin-resistant Fragment P 1. Clin Chem Clin Biothem 1985; 23: 572-3. Gressner AM, Tittor W. Serum laminin - its concentration increases with portal hypertension in cirrhotic liver disease. Klin Wochenschr 1986; 64: 1240-8. Tygsmp N. Determination of the hepatic elimination capacity (Lm) of galactose by si@Je injection. Stand J CIio Lab Invest 1966; 18 (Suppl.): 118-25. Miotti T, Bin-her .I, Preisig R. The 30-minute aminOpyrine breath test: optimization of sampling times after intravenous adrr;inisuahon of’4C-Aminopyrine. Digestion 1988; 39: 241-50. aeb J, Lange H, Bosch J, et al. Steady-state extrarenal sorbitol clearance as a measure of hepatic plasma flow. Gastroenterology 1988; 95: 749-59. Smith WW, Finkelstein N, Smith HW. Renal excretion of hexitols (sorbitol, marmitol, and dulcitol) and their derivates (sorbitan, isomannide, and sorbide) and of endogenous creatinine-like chromogen in dog and man. J Biol Chem 1940; 135: 231-50. Grabrielli GB, Faccioli G, Casaril M, et al. Procollagen III peptide and fibronectin in alcohol-related chronic Liver disease: correlations with morphological features and bicchemical tests. Clin Chim Acta 1989; 179: 315-U. Biagini G, Ballardini G. Liver fibrosis and extracellular matrix. J Hepatoll989; 8: 115-24. Gressner AM, Kropf J. Estimation of the production rates of serum aminoterminal propeptide of type III procollagen and laminin in human fibrotic liver. J Clin Chem Clin Biochem 1987; 25: 553-9.
eferemes 21
22
23
24
25
26
27 28 29 30 31
32
33
34
35
36
37
38
39
40 41
71
HEPAT 00904
E. Eotterd,
A.
‘Department offclinic~dPharmacoiogy, Universityof GoettingeTand ‘Department of Clinical Chemistry and Central Laboratory, Universiryof Marburg, Federal Republic of Germany {Received 7 September 1390)
In vitro models have shown t at metabolites of ethanol (acetaldehyde and lactate) stimulate collagen synthesis, thereby,
they may be important as fibrogenic mediators. The relevance of these findings for fibrogenesis in the human liver in vivo, however, has not as yet been demonstrated. Serum markers for collagen (PIIINP, using radioimmunoassays employing polyclonal antibodies and Fab-fragments (PIIINP-Fab), respectively) and basement membrane (laminin) metabolism were therefore investigated in 25 alcoholic cirrhotics (Pugh-Score: 6.7 -+ 1.9 SD.) and in 19 comparable nonalcoholic cirrhf*tics (Pugh-Score: 6.3 zb 1.5, n.s.) with only slight evidence for inflammation: COT 28 & 22 vs. 24 + 21 Wl; GPT 24 + 23 vs. 31 + 28 U/l; y-globulins 24 + 8 ‘4s. 22 +_5%, respectively {all n.s.). Severity of tbe disease was assessed by quantitative liver function tests. Levels of PIIINP, INP-Fab and laminin measured by RIA were 21 C 19ptgn, 90 f 42&l and 2.5 f: 0.8 U/ml in alcoholic cirrhosis and 10 + 6&l, 61 f 10&l and 1.9 -C0.4 U/ml in nonalcoholic cirrhosis, respectively (all p K 0.01). Differences on PIIINP and PIIINP-Fab remained significant even after accurate matching for galactose elimination capacity, aminopyrine breath test and hepatic sorbitol clearance. Laminin levels were higher in alcoholic cirrhosis only after matching for the hepatic sorbitol clearance (p < 0.01). The higher levels of serum markers for collagen and basement membrane metabolism in alcoholic vs. nonalcoholic patients with cirrhosis at equal severity of the disease and with only minimal signs of inflammation may be the clinical reflection of a specific fibrogenic effect of ethanol metabolites. suggestirng that
Ethanof-induced liver injury may trigger a fibrotic response in which perivenulat fibrosis is associated with proliferation of myofibrobiasrs, the mosi common type of mesenchymal cell around the perivenular areas in both the normal and the diseased liver (1). Metabolites of ethanol, such as acetaldehyde and lactate, have been shown to stimulate collagen synthesis in *znious celi cuiture systems (2-6). Increases of collagen synthesis and of messenger RNA have been described in human fibroblasts and in the liver of baboons given acetaldehyde or ethanol, respectively (7,8). Voss et al. (9) also observed cells similar to myofibroblasts in vitro in the outgrowth of explants of fi-
brotic livers. Sush cells synthesized collagen types I, III, IV, and V. Taken together, these investigations suggest a specific fibrogemc effect of ethanol or its metabolites. The relevance of these in vitro findings for in vivo fibrogenesis in the human liver has not as yet been demonstrated satisfactorily. In various liver diseases the serum markers for collagen and basement membrane metabolism, such as serum concentrations of the aminoterminal type III procollagen pe@e (PIIINP and PIIWB>-Fab, using polycPonal iantibodies and Fab-fragments, respectively), and of laminin have been demonstrated to exhibit significant correlations with activity of the fibrogenic pro-
* Present address: University of Witten-Herdecke, Bcckweg 4, D-5604 Herdecke, F.R.G. Correspondence andpresent address: Dr. E. Letterer, M.D., Department of Medicine H,University of Erlangen-Niimberg, 12, D-5520 Erlangeo, F.R.G.
Krankenhausstrasse
E. LOITERER
72
cess (10-26). Their plasma levels, however, may not depend only on collagen formation! but also on the disposition of these markers by hepatic or renal elimination ‘(27,28). In order to investigate the role of ethanol in fibrogenesis in intact man, the serum markers in patients with alcoholic and with nonalcoholic cirrhosis who are well defined in regard to both inflammatoy activity and severity of organ dysfunction must be compared. This study is an attempt to achieve such a goal.
Material and Methods Patients
This study is based on a group of 25 patients with alcoholic cirrhosis and another group of 19 patients with nonalcoholic cirrhosis (Table 1). This second group consisted of 12 patients with hepatitis B virus (HBV)- and seven patients with hepatitis NANB virus (HNANB)-related liver disease. All patients were clinically stable cirrhotics. There had been no ascites or upper GI-hemorrhage for 41 months (median, range: 8-79 months) before the study and patients had received biopsies within the last 4 years. The diagnosis of alcoholic cirrhosis was established by the following criteria:
et al.
- alcohol consumption of more than 60 g per day for females and more than 80 g per day for males for at least 10 years, and abstinence for at least 6 months before the study, as documented by negative alcohol plasma concentrations at each visit; - routine laboratory tests compatible with alcoholic cirrhosis; - sonography with typical alterations of the liver and lack of abnormalities of extrahepatic bile ducts; - liver biopsy showing cirrhosis and lesions compatible with chronic alcohol consumption in 18 patients (29) and oesophageal varices or ascites in the others. The diagnosis of nonalcoholic cirrhosis, i.e., viral-associated cirrhosis, was established by the following criteria: - history of a transfusion or another pathway for infection with hepatitis B- or NANB-virus; - elevated serum aspartate aminotransferase (AST, GOT), alanine aminotransferase (ALT, GPT) for more than 6 months; - for hepatitis B: IgG-Anti-HBs, IgG-Anti-HBc and lack of HBs-antigen, IgM-Anti-HBs and IgM-Anti-HBc; - for hepatitis NANB: negative serology of hepatitis A and B, Anti-HC was not sought; - alterations associated with cirrhosis and no abnormalities of extrahepatic bile ducts by ultrasonography; - liver histology showing lesions associated with cirrhosis and/or chronic inflammation in 14 patients (29).
TABLE 1 Characterization of the patients with liver cirrhosis (median, range)a Alcoholic cirrhosis Number of cases Sex (m/f)b Age (years)
Nonalcoholic cirrhosis
25 2213
19 1118
(::-67)
(G-74)
(Z-115) 18
(i-96) 14
(56-12)
(5611)
Body weight (kg) Cirrhosis on liver biopsy Pugh-Score Oesophageal varices (stage I-IV) History of oesophageal variceal hemorrhageC Splenomegaly’ History of ascitesC Treatment of ascites at the time of study
p
co.05 n.s. ns ns ns
Design of the study
All patients were investigated as out-patients in the liver clinic of the Division of Gastroenterology and Endocrinology, University of Goettingen. At the initial visit clinical and laboratory evaluation was conducted before further tests. Treatment modalities were adapted to the individual needs of each case. Patients with alcoholic cirrhosis were encouraged to stop drinking. Investigations (i) Specimen collection of blood for determination of
20
7
co.01
connective tissue parameters in serum. After allowing the
9 16 13
3 7 2
n.s. co.05 co.01
11
2
co.05
blood to clot for 1 h at room temperature serum was prepared by centrifugation (15 mitt, 5000 x g) and stored at -70 “C for a maximum of 1 month before the assays were performed.
a Some of the patients were treated with following drugs: acetylcysteine, atenolol, captopril, colestyramine, digitoxin, digoxin, famotidine, ferrum, glibenclamide, lactulose, nifedipine, ranitidine, sucralfate, theophylline, verapamil, vitamine K. Furosemide and spironolactone were used for treatment of ascites. b When calculated by the stastistical tests there is not any significant difference between male and female patients between both groups. ’ Oesophagtal varices (stage I-IV) and oesophageai variceal hemorrhage were verified by gastroscopy, ascites and splenomegaly were detected clinically and sonographically.
(ii) Determination of N-terminal propeptide of type III procollagen
(PZZINP). Method A: The propeptide was measured in various dilutions of serum using a radioimmunoassay (Behring-Hoechst, Frankfur&iMain, F.R.G.) based on intact polyclonal antibodies recognizing mainly the higher molecular mass (intact) propeptide Co1 1-3 (10,30). The intraday coefficient of variation (CV) l’vas 4.9% (at 12,@1; n = 14), the interday CV was 7.2% (:at 5
-l-YPEIii’ PROCOLLAGEN
PEPTIDE
AND LAMININ
IN CI
&I; n = 12). A reference (normal) range of PIIIN rum 0:2-12&l was accepted (18). Method B: This method based on a recently described radioimmunoassay employing the Fab-fragment of the antibody to PIIINP rather than intact antibody (31). The Fab-fragment has equal affinity for Co1 1-3 and Co1 1 and, hence, measures intact propeptide and its globular fragment with parallel inhibition profiles. The intraday CV was 6.9% (at 26 pg/l; R = 8), the interday CV was 5.2% (at 20 @gl; n = 7). The reference range of PIIINPFab was 15-63bg/l(31). (ii!) Detemindon oflaminin. Concentrations of the glycoprotein were determined with a competitive radioimmunoassay (Behring-Hoechst, FrankfurtMain, F.R.G.). The analytical del-ails have been reported previously (32,33). The antibody is directed against antigenic determinants of the pep&r-resistant fragment Pl of laminin (32). The CV of int,-aday and interday reproducibility were 3.2% and 4.9% (at a concentration of 2.0 U/ml), respectively. The concen:!ration of laminin is expressed in arbitrary U/ml since an International standard of laminin is not available: 1 U is defined as the mean quantity of Iaminin per ml serum in a group of obviously healthy volunteers, and corresponds to about 0.23 pg (32). The reference range’of serum lamini:? in our laboratory is 1.1-1.9 u/ml (34). (iv) Quantitative fiverjimction test. The galactose elimination capacity (GEC) was ca’culated from venous blood samples taken between 20-60 min at 5 min intervals after i.v.-injection of 0.5 g/kg galactose as described by Tygstrup (35). A factor of 5 min was used for correction of uneven distribution of galactose iu the body (28). For the aminopyrine breath test (ABT) samples were examined I;0 min after iv.-injection of 1 mg/l.5&i dimethylamine]14C]aminopyrine according to Miotti et al. (36). The hepatic sorbitol clearance (SCI) which represents a measure of hepatic parenchymal plasma flow was determined as described by Zeeh et al. (37). All blood samples were taken under fasting conditions. Routine biochemical and hematological examinations were carried out with automated techniques. Alcohol plasma levels were determined enzymatically and were below the limit of detection in all patients at the time of the investigation. Written informed consent was obtained from each participant in accordance with the declaration of Helsinki, revised in Tokyo in 1975. Statistical analysis The Wilcoxon test for unpaired data and the Chi-square test were used for statistical comparisons. Both tests were corrected for multiple testing by the Friedman rank test.
73 earman rank test was used for calculation of correlations. The level of significance wasp < 8.05. esults Initially the group of 25 patients with alcoholic cirrhosis was similar to the group of 19 patients with viral-associated cirrhosis in age (52 + 8 VS. 52 + 12 years, n.s,), body weight (76 + 14 vs. 73 f 14 kg, n.s.), Pugh-Score (6.7 + 1.9 vs. 6.1 + 1.7) and drug intake (Table 1). Tbe frequency of oesophageal varices, ascites, and splenomegaly were significantly higher in the alcoholic group (Table 1). The routine laboratory tests are summarizad in Table 2. In the group with alcoholic cirrhosis levels of alkaline phosphatase, y-GT and serum bilirubin were signiflcantly higher than in the patients with nonalcoholic cirrhosis, whereas pseudocholinesterase activity and prothrombin time were significantly lower. No significant differences were found in regard to GOT, GPT, albumin, yand serum-creatinine. The globulins, thrombocytes values of sodium, potassium, urea, cholesterol, triglyTABLE
2
Routine laboratory tests in patients with alcoholic and nonalcoholic cirrhosis (median, range) Tests (reference range)
Alcoholic cirrhosis
Nonalcoholic cirrhosis
GOT (<17 U/l)
&3)
(8-99)
GPT (~23 U/l)
&J6) (C96)
y-GT ( ~28
16 (8-247)
(E351)
A,lkaline phosphatase (so- 200 un)
184 (70-585)
117 (61-362)
co.05
Serum-bilirubin (t1.2mgtdl)
(EL.!)
eo.05
(::63-27.2)
y-Globulins (lP-20%)
(Z-37)
(E-30)
Pseudocholinesterase (2300-7400 U/l)
3100 (7OO-BOo)
4800 (1400-9600)
Albumin (3.5-5.0g/dl)
g-4.6)(2.9-5.2)
ns.
(Lm)
co.01 (Z-100)
Thrombocytes (150-300.103Ij.q
116 (33-291)
141 (40-312)
n.5.
Creatinine (0.6-l .3 mg/dl)
1.0 (0.6-1.3)
n.s.
(E-1.3)
18
un)
Prothrombin (70- 100%)
p ns.
n.5.
n.5.
4.0
time
co.01
E. LOTIERFX
74 ceride and haennotologic parameters were within the normal range. me median ABT was 0.22 %dose.kg per mmolC@ (0.01-1.13) in patients with alcoholic cirrhosis and 0.55 %dose+kgper mmolC0, (0.13-1.18;~ C 0.01) in patients with nonalcoholic cirrhosis. In contrast, for the GEC no significant difference between the groups was observed (3.7 mg/min per kg (1.5-5.2) vs. 4.1 mg/min per kg (2.1-6.83, respectively). Levels of the SC1 were significantly lower in the alcoholic group than in the nonalcoholic group (5.8 ml/min per kg (2.4-11.6) vs. 7.3 ml/min per kg (3.4-10.6), p c 0.05). The renal sorbitol clearance, thought to express glomerular filtration rate (38) was 1.3 ml/min per kg (0.4-2.1) and 1.4 mYmin per kg (0.6-2.0, n.s.), respectively. In patients with alcoholic cirrhosis serum concentrations of PIIINP, PIIINP-Fab and laminin were 19 ,@l (5-SO), 79 pg/l (36-193) and 2.6 U/ml (1.4-4.1) respectively, i.e., significantly higher than the values of patients with viral-associated cirrhosis (8 pg/l, 4-25, 59 @I, 49-83, and 1.9 U/ml, 1.1-2.7, respectively, all p < 0.005). The median ratio of PIIINP-Fab/PIIINP levels was 5.7 (1.9-10.6) in alcoholic and 7.8 (3.8-14.8) in nonalcoholic cirrhotics (p c 0.01). In order to compare serum levels of PIIINP, PIIINPFab and laminin of both groups at equal liver function a matching procedure was applied to the data as shown in Table 3. After matching for the Pugh-Score, significantly elevated concentrations of all three parameters were found in alcoholic patienis compared to nonalcoholic patients. After matching for the GEC and the ABT, only
80
1
TABLE
et al.
3
Matching of patients with liver cirrhosis for Pugh-Score, GEC, ABT, and SC1 (median, range)a Alcoholic cirrhosis
Nonalcoholic cirrhosis
P
a) Matching for Pugh-Score (n = 13) PIIINP @il) 15 (5-46) PIIINP-Fab (J@) 79 (36-115) Laminin (U/ml) 2.5 (1.4-3.8) GOT (U/l) 20 (8-113) 16 (g-116) GPT (U/l) y-Globulins (%) 22 (10-31)
19 (4-24) 56 (49-80) 1.7 (1.2-2.4) 12 (8-99) 20 (5-96) 21 (15-30)
co.05 co.05 co.05 n-s. ns. ns.
b) Matching for GEC PIIINP (&I) PIIINP-Fab @g/l) Laminin (U/ml) GOT (U/1) GPT (U/l) y-Globulins (%)
(n = 14) 16 (S-24) 74 (39-115) 2.4 (1.4-3.8) 20 (8-37) 17 (8-48) 23 (10-34)
9 (4-24) 60 (49-80) 2.0 (1.2-2.6) 26 (8-99) 28 (5-96) 24 (14-30)
co.05 co.05 n.s. ns. ns.
c) Matching for ABT PIIINP f&g/l) PIIINP-Fab (qil) Laminin (U/ml) GOT (U/I) GPT (U/l) y-Globulins (%)
(n = 12) 18 (5-26) 82 (36-147) 2.5 (1.4-2.9) 21 (10-41) 19 (13-53) 22 (10-35)
9 61 2.0 21 22 23
co.05 co.05 n.s. n.s. n.s. n.5.
d) Matching for SC1 (n = 13) PIIINP &g/l) 21 (5-33) PIIINP-Fab (.@l) 88 (36-147) Laminin (U/ml) 2.6(1.4-3.8) GOT (U/l) 20 (10-113) GPT (U/l) 16 (7-116) v-Globulins (%) 23 (15-35)
(4-24) (49-83) (1.2-2.7) (10-99) (6-96) (15-30)
10 (4-t4j 59 (49-80) 1.9(1.2-2.6) 16 (8-99) 20 (5-96) 22 (18-28)
-
l-l.S.
co.02 co.01 co.01
n.s. ns. n.s.
a Matching for Pugh-Score (+O), GEC (+O.l mg/min per kg), ABT (+0.05 %dose.kg per mmol CO,), and SCI (20.2 ml/mm per kg),
a *
I 60
R,=-O.S6 p (
R,=-3.46
0.01
a
p c 0.002
7 2 A 0 4o z = 0 20
0
--0
O+l-------
01
2
4
G E C [mg/min/kg]
6
Cl
2
4
G E C [mg/min/ka]
6
2
4
6
G E C [mg/min/kg]
Fig. 1. Spearman rank correlations (R,) between the galactose elimtnation capacity (GEC) and PIIINP, PIIINP-Fab, and laminin in patients with alcoholic cirrhosis (n = 25, and in patients with nonalcoholic cirrhosis (n = 19, A). Horizontal lines are the upper limits of the corresponding normal ranges.
TYPE III PROCOLLAGEN
PEPTIDE
AND LAMININ
IN CIRRHOSIS
75
80-s R,s-0.56
a,=-0.45 P
c
p c O.OOI
0.002
Oi
0.0
1 .o
0.5
0.0
0.5
ABT[E]
1 .o
-0.0
1 .o
0.5
ABT[E]
ABT[E]
Fig. 2. Spearman rank correlations (R3 between the aminopyrine breath test (ABT) and PIIINP, PIIINP-Fab, and laminin in patients with alcoholic cirrhosis (n = 25, ) and in patients with nonalcoholic cirrhosis (n = 19, A j. Horizontal Lnes are the upper limits of the corresponding normal ranges.
80
0
4 S C
8
I [ml/min/kg]
12
0
4 S C
8
I [ml./min/kg]
12
0
4 S C
a
12
I [ml/min/kg]
Fig. 3. Spearman ra& correlations (R,) between the hepatlc sorbitol clearance (SCl) and PIIINP, PIIINP-Fab. and laminin in pat:ents with alcoholic cirrhosis (n = 25. apd in patients with nortalcoho!ic cirrhosis (n = 19, A). Horizontal lines are the upper limits of the corresponding normal ranges.
PIIINP- and PIIINP-Fab levels were significantly higher in the alcoholic groilp than in the no,lalcoholic group, whereas for laminin no significant differences were detected. In contrast, after matching for SC1 all three parameters showed significant differences between both groups. As shown in Figs. l-3 significant, negative correlations were found between serum concentrations of PIIINP, PIIINP-Fab, and laminin and the GEC (all p < O.Ol), the AI3T (all p : 0.002), and the SC1 (all p < 0.05). PIIINP was positively correlated with PIIINP-Fab (R, = 0.58) and laminin (R,= 0.63), and PIIINP-Fab with laminin (R, = 0.63; all p < 0.0001). In agreement with other investi-
gations (10,12,20,23,39) we found in the group with nonalcoholic cirrhosis PIIINP positively or negatively correlated to GOT (R, = 0.66), y-GT (R, = 0.68), serum bilirubin (R!; = 0.67), y-globulins (R, = 0.73). albumin (R, = -0.73), pseudocholinesterase (R, = -OH), prothrombin time (R, = -0.75), and thrombocyte count (R, = -0.64, all p < O.Ol), whereas laminin was related only to GOT, alkaline phosphatase, and albumin (all R,values were > f 0.64, all p < 0.01). In the group with alcoholic cirrhosis PIIINP, PIIINP-Fab and laminin concentrations showed a significant negative relationship only to serum albumin respectively (all R, values and pseudocclolinesterase, were 2 -0.52, allp < 0.01).
E. LOTTERER
76 Discussion The data of this study seem to show that PIIINP, PIIINP-Fab, and laminin are higher in alcoholic cirrhosis than in nonalcoholic cirrhosis, even when data are controlled for the severity of the disease and for the degree of inflammation. Consequently, these results support the hypothesis that in alcoholics acetaldehyde and other ethanol metabolites exert fibrogenic activity over and above the fibrogenesis which results from hepatocyte injury. me results of our study extend concepts derived from animal and tissue culture experiments to explain the development of alcoholic cirrhosis in man (2-8). Before accepting these conclusions, however, the methodological limitations have to b: analysed in detail. A relation between the investigated connective tissue markers and fibrogenesis in the liver is well documented in literature (lo-26,39). Serum levels always depend on at least two factors, however, the rate of formation and the rate of elimination of the compottnd under consideration. Consequently, when comparing serum markers of two groups of patients it is important to control for alterations in the function of organs which may be involved in the disposition of serum aminoterminal type III procollagen peptide and laminin (40). In our study both groups of cirrhotics had equal serum creatinine concentrations and renal sorbitol clearances and both groups were matched for GEC, ABT, and SC1 (Tables 2 and 3). Consequently, in this study it was possible to investigate patients with alcoholic and nonalcoholic cirrhosis. Since these patients had comparable renal and hepatic functions, it may therefore be reasonable to assume that processes responsible for the disposition of the serum markers were adequately controlled. On this basis it was concluded that the higher serum markers in alcoholics are likely to be due to an increased rate of formation (21,23,24,41). It has also been well documented that the degree of hepatocellular damage and inflammation within the liver correlates with elevation in the investigated serum markers of fibrogenesis (10,12,17,19-21,23,39). This was confirmed in our patients with nonalcoholic cirrhosis since the levels of PIIINP were correlated with GOT and y-globulins, and the levels of laminin with GOT. It seemed important therefore to compare patients with similar signs of inflammation. There were no differences in inflammation values in the total groups or after matching (Table 3). The data, therefore, support the hypothesis that intrahepatic cell damage and inflammation was slight in our patients and comparable among both groups. Consequently, the observed differences between alcoholic and nonalcoholic cirrhotics in PIIINP, PIIINP-Fab, and laminin cannot be
et al.
explained by differences in hepatocellular damage or inflammation. In this study the conclusion tl=i t increased serum levels of connective tissue markers and presumably increased fibrogenesis in alcoholic cirrhosis may be related to the toxic effects of acetaldehyde an other ethanol metabolites, is based on an exclusion of possible alternative explanations. Although these exclusions can never be exhaustive, the data appear to be fully compatible with in vitro experiments which have shown that acetaldehyde leads to fibrogenesis with increases in collagen synthesis and messenger RNA in various cell culture systems (2-9). The concentrations of acetaldehyde used in these studies was in accordance with concentrations found in the liver during alcohol abuse (2-8). Together with this experimental evidence our findings support the idea that acetaldehyde may also be a specific factor for collagen synthesis in the liver and thereby may contribute to the pathogenesis of alcoholic cirrhosis in man. Al?hough it might have been interesting to investigate hepatic fibrosis more directly, i.e., the degree of hepatic fibrosis evaluated accurately by liver biopsy, microscopic examinations could only provide a static view. Information concerning the dynamics of hepatic fibrogenesis might be better obtained by direct determination of changes in amounts and the molecular composition of various types of connective tissue. These studies, however, are difficult to execute. An interesting detail of our investigation remains difficult to explain, i.e., why the matching for GEC and ABT eliminated differences in laminin levels, whereas matching for SCl made the difference between alcoholic and nonalcoholic cirrhosis highly significant (Table 3). Taken at face value this may imply that the severity of the disease, as reflected by hepatocyte mass (GEC and ABT), is not a pertinent determinant of laminin. In contrast, limitations of the vascular bed to equal functional liver plasma flow values results in higher laminin concentrations in alcoholics. This could suggest that more basement membrane is formed in alcoholic cirrhosis which may be due to a higher degree of capillarization of sinusoids (24). Although speculative, these changes could again be related to acetaldehyde or other ethanol metabolites.
Acknowledgements The authors thank the nurses Mrs. A. Marx and Mrs. M. Suhr-Gottschaldt for their care for the patients and Mrs. C. Guthardt and Mrs. H. Steinmetz for their technical assistance.
TYPE III PROCOLLAGEN
PE
DE AND LAMININ
IN CIR
77
I. Nakano M, Worner TM, Lieber CS. Perivenuiar fibrosis in al& hofic injUrY: ultrastructure and histologic progression. Gastroenterology 1982; 83: 777-85. 2 Savolainen ER, Leo MA, Timpl R, Lieber CS. Acetaldehyde and lactate stimulate collagen synthesis of cultured baboon liver myofibroblasts. Gastroenterology 1984; 87: 777-87. 3 Holt K, Bennett M, Chojkier M. Acetaldehyde stimulates collagen and noncollagen protein production by human fibroblasts. Hepatology 1984; 4: 843-8. 4 Moshage H, Casini A, Lieber CS. Acetaldehyde selectively increases collagen-synthesis in cultured fat-storing cells but not in hepatocytes. Hepatology 1953; 10: 630. 5 Tsutsumi M, Takade A, Takas? S, Siiiimanaka K, Enyama EC.Effects of ethanol and its metabolites on collagen-synthesis by CL& tured rat-liver cells. J Gastroenterol Hepatol1989; 4: 229-40. 6 Kamegaya K, Qkazaki I, Maruyama K, et al. Ethanol, acetaldehyde and lactic-acid stimulate type-III collagen-synthesis of cultured chang liver-cells. Hepatology 1988; 8: 1435. 7 Brenner DA, Chojkier M. Acetaldehyde increases collagen gene transcription in cultured human fibroblasts. J Biol Chem 1987; 262: 17690-5. 8 tirn Ua,, Leo MA., Giambrone MA, Lieber CS. Increased type I procollagen mRNA !eve!s and in :r+_:o Fro:& qj&esis & &e baboon model of chronic alcoholic Liver disease. Gastroenterology 1985; 89: 1123-31. 9 Voss B, Rauterberg .I, Pott G, et al. Nonparenchymal cells cultivated from explants of fibrotic liver resemble endothelial and smooth muscle cells from blood vessel walls. Hepatology 19B2; 2: 19-28. 10 Rohde H, Vargas L, Hahn EG, Kalbfleisch H, Bruguera M, Timpl R. Radioimmunoassay for type III procollagen peptide and its application to human liver disease. Eur J Clin Invest 1979; 9: 451-9. 11 Raedsch R, Stiebl A, Waldherr R, et al. Procollagen type IIIpeptide serum concentrations in chronic persistent and chronic active hepatitis and in. irrhosis of the liver and their diagnostic value. Z Gastroenterolqie 1982; 20: 738-43. 12 Frei A, Zimmermarln PI Weigand K. The N-terminal propeptide of collagen type III iE sr rum refIects activity and degree of fibrosis in patients with CL~ ~tnit liver disease. Hepatology 1984; 5: 830-4. 13 Weigand K, Zaugg P-Y, ,:rei A, Zimmermann A. Lone-term follow-up of serum N-teti al propeptide of collagen type III levels in patients with chronic li ver disease. Hepatology 1984; 4: 835-8. 14 Savolainen E-R, Goldbtq B, Leo M.4, Velez M, Lieber CS. Diagnostic value of semi procoiiagen peptide measurement in alcoholic liver disease. Alcoholism 1984; 8: 384-9. 15 Tanaka Y, Minato 9, Hasumura Y, Takeuchi J. Evaluation of hepatic fibrosis by serum proline and amino-terminal type III procollagen peptide !evels in alcoholic patients. Dig Dis Sci 1985: 31: 712-7. 16 Schuppan D, Dumont JM, Kim KY, Hennings G, Hahn EG. Serum concentration of the aminoterminal procollagen type III peptide in the r& reflects eaily formation of connective tissue in experimental livc,r cirrhosis. J Hepatoll986; 3: 27-37. 17 McCullough AJ, Stassen WN, Wiesner RH, Czaja AJ. Serial determinations of the amino-terminal peptide of type III procollagen in severe chronic active h.:patitis. .I Lab Clin Med 1987; 109: 55-61. 18 Niemelij 0, Ristcli L, Sotaniemi EA, Risteii J. Aminoterminal propeptide of type III procollagen in serum in alcoholic iiver disease. Gastroenterology 1983; 85: 254-9. 19 Torres-Salinas M, Par& A, Caballeria .I, et al. Serum procollagen type III peptide as a marker of hepatic fibrogenesis in akoholic%epatitis. Gastroenterology 1986; 90: 1241-6. 20 cban.g -I-I’, l_imHC, Lee SD, et al. Clinical significance of serum
type-III procollagen aminopropeptide in hepatitis B virus-related liver diseases. Stand J Gastroenterol1989; 24: 533-8. Annoni G, Colombo M, Cantaluppi MC, Kbiat B, Lampertico P, Rojkind M. Serum type III procollagen peptide and laminin (Lam-PI) detect alcoholic hepatitis in cl.:rJnic alcohol abusers. Hepatology 1989; 9: 693-7. van Zanten RAA, van Leeuwen RE, Wilson JHP. Serum proc& lagen III N-terminal peptide and laminin PI fragment concentrations in alcoholic disease and primary biliary cirrhosis. Clin Cbm Acta 1988; 177: 141-6. McCullough AJ, Stassen WN, Wiesner RI-I, Czaja AJ. Serum type III procollagen peptide concentrations in severe chronic active hepatitis: relationship to cirrhosis and disease activity. Nepatology 1987; 7: 49254. :Niemefl 0, Risteli L, Sotaniemi EA, Risteli J. Type IV collagen and laminin-related antigens in human serum in alcoholic liver disease. Eur J Clin Invest 1985; 15: 132-7. Sato S, Nouchi T, Womer TM. Lieber CS. Liver fibrosis in altoholics. Detection by Fab radioimmunoassay of semm procoUagen III peptides. J Am Med Assoc 1986; 256: 1471-3. ‘Kropf J, Gressner AM, Negwer A. Efficacy of serum laminin measurement for diagnosis of fibrotic liver diseases. Clin Chem 1988; 34: 2026-30. Hahn EG. Blood analysis for liver fibs&s. .I >!*r\atol 1984; 1: 67-73. Bircher J. Quantitative assessment of deranged hepatic function: a missed opportunity? Sem Liver Dis 1983; 3: 275-84. Scheuer PJ. I iver Biopsy Interpretation. 3rd Edn, London: BailIi&e Tindall, 1481. Niemalti 0. Radioimmunoassays for type III procollagen aminoterminal peptides in humans. Clin Chem 1985; 31: 1301-4. Rohde H, Langer I, Krieg T, Timpl R. Serum and urine ana&sis of the aminoterminal procollagen peptide type III by radioimmunoassay with antibody Fab fragment. Collagen Rel Res 1983; 3: 371-9. Brocks DG, Strecker H, Neubauer HP, Tiipl R. Radioimmunoassay of laminin in serum and its application to cancer patients. Clin Chem 1986; 32: 787-91. Pick-Kober KH, Negwer A, Gressner AM. Determination of laminin in serum by a competitilve radialmmunoassay directed against the pepsin-resistant Fragment P 1. Clin Chem Clin Biothem 1985; 23: 572-3. Gressner AM, Tittor W. Serum laminin - its concentration increases with portal hypertension in cirrhotic liver disease. Klin Wochenschr 1986; 64: 1240-8. Tygsmp N. Determination of the hepatic elimination capacity (Lm) of galactose by si@Je injection. Stand J CIio Lab Invest 1966; 18 (Suppl.): 118-25. Miotti T, Bin-her .I, Preisig R. The 30-minute aminOpyrine breath test: optimization of sampling times after intravenous adrr;inisuahon of’4C-Aminopyrine. Digestion 1988; 39: 241-50. aeb J, Lange H, Bosch J, et al. Steady-state extrarenal sorbitol clearance as a measure of hepatic plasma flow. Gastroenterology 1988; 95: 749-59. Smith WW, Finkelstein N, Smith HW. Renal excretion of hexitols (sorbitol, marmitol, and dulcitol) and their derivates (sorbitan, isomannide, and sorbide) and of endogenous creatinine-like chromogen in dog and man. J Biol Chem 1940; 135: 231-50. Grabrielli GB, Faccioli G, Casaril M, et al. Procollagen III peptide and fibronectin in alcohol-related chronic Liver disease: correlations with morphological features and bicchemical tests. Clin Chim Acta 1989; 179: 315-U. Biagini G, Ballardini G. Liver fibrosis and extracellular matrix. J Hepatoll989; 8: 115-24. Gressner AM, Kropf J. Estimation of the production rates of serum aminoterminal propeptide of type III procollagen and laminin in human fibrotic liver. J Clin Chem Clin Biochem 1987; 25: 553-9.
eferemes 21
22
23
24
25
26
27 28 29 30 31
32
33
34
35
36
37
38
39
40 41