- Email: [email protected]
Modifications of Concanavalin A patterns of α1-acid glycoprotein and α2-HS glycoprotein in alcoholic liver disease
Chnrca Chimlca Acfa, 176 (1988) 49-57
49
Elsevier
CCA 04225
Modifications of Concanavalin A patterns of al-acid glycoprotein and (r2-HS glycoprotein in alcoholic liver disease Maryvonne Jezequel a, Nathalie S. Seta a, Michkle M. Corbic b, Jeanne M. Feger a and Genevikve M. Durand a ULaboratowe de Blochlmre A, Hsplral Blchar 7.5018 Paris and Laboratorre de Blochrmre, CNRS UA 622, UER des Scrences Pharmaceutiques. 92290 Chatenav-Malabry. and h INSERM lJ 24. H;pprtal Beaujon, 92 Chchv (France) (Received
19 November
1987: reviston received
Key words: Concanavalin
A; al-Acid
15 March
1988; accepted
glycoprotein: a2-HS Liver drsease; Alcohol
after revision
Glycoprotein;
26 April 1988)
Microheterogeneity;
In order to test whether abnormalities in hepatocytes affect the glycoprotein carbohydrate moiety, crossed immunoaffinoelectrophoresis (CIAE) with Concanavalin A (Con A) was used to study serum alpha l-acid glycoprotein (al-AGP) and alpha 2-HS glycoprotein (a2-HS) obtained from alcoholic patients with biopsy-proven liver disease. Cirrhotic patients, placed in groups Cl, C2 or C3, according to Pugh’s classification, were compared to healthy donors (N) and to steatosic non-cirrhotic patients (S). Con A CIAE patterns revealed in group N three subpopulations for a2-HS and four for d-AGP. Two main results emerged from this study: (1) in the alcoholic groups, the proportions of Con A-unreactive subpopulations of both glycoproteins increased. Moreover, group N could be separated from group S and group S from all the cirrhotic groups. (2) There was a good correlation between the relative amounts in Con A-unreactive subpopulations of d-AGP and a2-HS. The increases observed in Con A-unreactive subpopulations are probably a general phenomenom related to alterations in glycosylation processing during liver cell damage.
Correspondence Pans, France.
to: N. Seta, Laboratoire
0009-8981/88/$03.50
de Biochimie
0 1988 Elsevrer Science Publishers
A. HBpital
Bichat.
B.V. (Btomedical
46. rue H. Huchard,
Division)
75018
50
Introduction A number of chemical abnormalities related to excessive alcohol consumption and alcoholic liver injury have been detected. Particularly, changes in the concentration of plasma proteins have often been reported [l--3]. Lately, however, much interest has been shown in plasma glycoprotein microheterogeneities. The presence of glycoprotein ~croheterog~neities can be demonstrated directly in the serum by lectin-crossed immunoaffinoelectrophoresis (CIAE) in which the proteins are separated by their electrophoretic mobility and their retardation due to the affinity of their carbohydrate structures for lectin. Several authors studying serum a,-acid glycoprotein ( (Y,-AGP) and other glycoproteins in acute inflammatory disease have observed pattern modifications using CIAE in the presence of Concanavalin A (Con A-CIAE) with an increase in Con A-reactive forms [4,5]. These reports raise the question of a same regulatory system of glycoprotein glycan moiety synthesis in infla~ation. We have recently shown that (r,-AGP in alcohol-induced cirrhotic patients, exhibits an increase in Con A-unreactive microheterogenous forms, when separated by Con A-CIAE. This alteration has been shown to be related to liver damage [6]. The purpose of this work was: 1) to confirm and extend our previous results using larger groups and to compare them with those of non-cirrhotic alcoholic patients. 2) To study simultaneously by Con A-CIAE, the two glycoproteins a,-AGP and (u,-HS glycoprotein (LY*-HS), which are respectively positive and negative acute phase proteins [7,8]. Materials and methods Patients
Eighty-two alcoholic patients with biopsy-proven liver disease were studied (58 men, 24 women; age 24-72 years, mean 46 years). Histological findings included steatosis without cirrhosis (20 patients, group S) and cirrhosis with or without steatosis, fibrosis and alcoholic hepatitis (62 patients, group C). Cirrhotic patients were placed in group C,, C, or C,, according to Pugh’s classification [9], which is based on the grading (score l-3) of five clinical and biological complications of the cirrhotic process (encephalopathy, ascitis, bilirubinemia, albuminemia, prothrombin time). Details of the 62 cirrhotic patients are given in Table 1. Alcoholic patients were compared with 22 healthy donors (group N) (19 men, 3 non-pregnant women with no oestroprogestative treatment; age 21-39 years, mean 31 years) with alcohol consumption of between 0 and 15 g/day and with no known liver disease. Liver biopsies were not performed in the control subjects for ethical reasons. At the time of liver biopsy, blood specimens were drawn from the patients for the measurement of serum ALAT, ASAT, bilirubin, alkaline phosphatase, gammaglutamyltransferase ( y-GT), albumin, immunoglobulin A (&A). prothrombin time, proaccelerin, haemoglobin and mean corpuscular volume (MCV). Serum samples were collected at the same time and stored at - 20 o C.
51 TABLE
I
Conventional
liver function
Groups
N
n
22
Pugh’s score ASAT [5-25 UI/l) ALAT [5-25 UI,‘l] YGT [ -c30 UI/l] Alkaline phosphatase
test values in healthy S
and patients
14
+ 4
93
f
69 a
11
f
63
+
37 a
22
+11
200
73
rt23
92
with alcoholic
C2
Cl 20
5
donors
16 5-6 101 *107 66
=
+ 56 a
+247
366
+532
a
85
f
26
153
+ 42
&I.5
89
rt 12
70
f
15%
15
+
15
f
1.3
13
f
1.9
90
rt 4.5
f
5.7 a
99
f
7.5 =
liver disease f3
21 7-9 83 f
57 d
35
17 h,c
*
25 10-15 100 *
58 a
45
+ 27
401
&348 a
175
+136
251
+ 182 a,h
214
+177
67
rf: 19 a.’
36
4
10
&
2.2 a.b.c
i
8.4 ’
a.b
[ < 150VI/I] Proaccelenn [75-lOO%] Hemoglobin 112-17 g%] MCV [85-95 pm31
2.3f ;g-3.0
1.5
0.9
loo
2.3*
0.7
7.2+
11.75
5.0 a,h
103
i
5.8*
2.5 a.b 12.4 a 1.7
103
9.5 f
16 ‘.‘.’
3.5 a.b
g/l]
Results are expressed as mean+sD. Numbers in brackets N = healthy donors; group S = steatosis without cirrhosis; according to Pugh’s score. a vsN(p<0.002). h vss(p<0.001). c vs Cl (p i 0.001).
are the normal laboratory range. groups C = cirrhosis at different
Group stages.
Reagents Rabbit antisera to human oil-AGP and a2-HS and NOR Part&en oil-AGP and M Partigen cuZHS plates were purchased or gifts from Behringwerke, Marburg, FRG. Concanavalin A was from 1’Industrie Biologique Franqaise, Villeneuve-laGarenne, France. Agarose type II and a-D-methylglucopyranoside were from Sigma, USA. Polyethylene glycol (PEG 6000) and Coomassie blue G 250 were from Fluka and Merck (FRG) respectively. Crossed imm~noaffi~ity electrophoresis Crossed immunoaffinity electrophoresis (CIAE) was carried out essentially as described by Bog Hansen [lo]. The buffer (pH 8.7) contained 72 mmol/l Tris, 24 mmol/l Veronal, 0.4 mmol/l calcium lactate, 0.2 mmol/l sodium azide. Con A was dissolved at a final concentration of 5% in a solution containing 1 mmol/l MgCl,, M&l, and CaCl,. This solution was added to the first dimension gel (140 ,ul/cm2 for ail-AGP, 280 pi/cm’ for a2-HS). Specific antiserum was incorporated into the second dimension gel (7.5 &‘cm* for al-AGP, 15 $/cm2 for CYZ-HS)together with ~-D-methyl~ucop~anoside (4%) and PEG 6000 (2%). The gels were 1.5 mm-deep 1% agarose.
52
Sample Cal-AGP and (Y~-HS contents were determined by radial immunodiffusion and their concentrations were adjusted to 0.30 g/l + 0.05 so that the CIAE patterns would be comparable. Sample volumes of 4 ~1 ((wl-AGP) or 10 ~1 (a2-HS) of various dilutions were deposited on the gel. The first dimension was run at 9 V/cm for 1.5 h (al-AGP) or 2.5 h (a2-HS) and the second at 2.5 V/cm for 18 h for both glycoproteins. After drying and staining, peak areas corresponding to the various subpopulations were determined. Designating peak 1 as the most anodic, results were expressed as the surface ratio of the main peak (peak 1 for al-AGP and peak 3 for a2-HS) to the minor ones, ie
R&LAW = (I)/(2
+ 3 + 4) and &us
= (3)/(2
+ 1)
Statistical methods Calculations were made by analysis of variance for multiple comparison between groups. When a significant difference was found (0.05) comparison was performed with a modified t test and the Bonferroni method was used to determine the level of significance. Results Results of conventional liver function tests for healthy donors and alcoholic patient groups are reported in Table I. Serum ASAT, ALAT and the MCV differed significantly between groups N and S (sensitivity 70%, specificity 95%) and IgA between groups S and Cl (sensitivity 88%, specificity 82%).
TABLE
II
Comparative study variants of al-acid alcoholic patients
of plasma content and surface ratios of Con glycoprotein (al-AGP) and a2-HS glycoprotein
A-crossed ((~2-HS)
immunoelectrophoresis in healthy donors and
Groups
N
S
Cl
c2
n Pugh’s score (rl-AGP (g/l) Rh,.,,, = peaks l/(2 + 3 + 4) a2-HS (g/l)
22
20
16 5-6 0.52 * 0.28 b 1.53+0.53 a
21 25 7-9 10-15 0.83 f 0.57 0.50+ 1.75 f 0.87 a.b,c 1.88+
R aNIs=
0.72 f 0.12 0.72 f 0.07
0.91 f 0.24 1.07 f 0.47
0.66 + 0.08 3.04kO.42
0.63 + 0.13 2.37kO.55 =
0.59 f 0.16 1.75 *0.51 a.b
c3
0.56 * 0.12 1.68 + 0.62 a,h
0.21 0.77 h
0.34+0.13 a.h,‘ 1.78 + 0.51 a.b
peaks 3/(2 + 1) Results are expressed as mean f SD. Group N = healthy donors; group groups C = cirrhosis at different stages, according to Pugh’s score. a vs N (p i 0.002). b vss(p~0.005). c “S Cl (p i 0.005).
S = steatosis
without
cirrhosis;
53
1B
Fig. 1. Con A-crossed
immunoaffinoelectrophoresis patterns of serum d-acid a2-HS glycoprotein (B) from a healthy donor.
glycoprotein
(A) and
Total serum glycoprotein levels are presented in Table II. Total serum al-AGP concentrations were not found to be significantly different between groups, except for group Cl compared to the steatosis group. Total serum a2-HS concentrations did not differ significantly, except for group C3 which had significantly reduced levels. Typical Con A-CIAE patterns of al-AGP and a2-HS are presented in Figs. 1 and 2, respectively. Fig. 1A shows an al-AGP pattern with four peaks for a healthy donor: 1, Con A-unreactive; 2, weakly reactive; 3 and 4, the most reactive with Con A. Fig. 1B shows the corresponding a2-HS pattern: peak 1, Con A-unreactive and peaks 2 and 3, Con A-reactive. As can be seen in Figs 2A and B, showing al-AGP and a2-HS Con A-patterns from the same cirrhotic patient with a dramatic increase in Con A-unreactive peak 1 for both glycoproteins, the patterns are different but the direction of the shift is the same. These pattern modifications were quantified for all the groups through the corresponding ratio values (Table II). These ratios, R&t_,,, and Ra2_HS, gave a good representation of the observed pathologic-related variations: R&I_AGPvalues rose progressively from normal subjects (0.72 + 0.07) to cirrhotic group C3 (1.88 k
54
1
2A
-,- - ._
,,_
-7---
,
I
2B
Fig. 2. Con A-crossed
immunoaffmoelectrophoresls patterns of serum q-acid a,-HS glycoprotein (B) from a clrrhotlc patlent.
glycoprotein
(A) and
400.
RalAGP
. 200.
I. .
- . 2.00
4.00
RoZ-HS
Fig. 3. Correlation
between
ratios Rb,_ac,d glywprote,n (R&, = peaks l/2 + 3 + 4) and a2-HS glycoprotein (R& = peaks 3/2 + 1) (‘r’ = 0.70, n = 104).
0.77). This marked increase allowed group N to be differentiated from Group Cl (sensitivity 1008, specificity 82%) and group S from group C2 (sensitivity 47%, specificity 958). RazsHS values decreased progressively from normal subjects (3.04 + 0.42) to cirrhotic group C3 (1.75 rt 0.51). This decrease allowed group N to be separated from group S (sensitivity 65% specificity 82%) and from all the cirrhotic groups (~nsiti~ty 94%, specificity 82%); furth~~ore, group S could be separated from all the cirrhotic groups (se~siti~ty 52W, specificity 90%). By associating the two ratios R&i,, and RaZ_uS,each group coutd be separated from the others with good sensitivity and over 95% specificity. Moreover a significant correlation was found between R,,_Hs and R&,oe, as represented in Fig. 3 (correlation ‘r ’ = 0.70, n = 104). Finally, no correlation was found between the two ratios R&l_AGPand RaZ_HS,and the corresponding glycoprotein plasma levels. A Con A-CIAE analysis of LU,-AGPwas carried out using the serum of a steatosic patient during an alcoholic withdrawal period of 15 days: a histologically proven improvement in steatosis was observed and the &-AGP pattern was modified, with a decrease in the Con A-unreactive peak (before: R',,_,,, = 1.14; after: RL,_,o, = 0.68). Discussion The biochemical heterogeneity of serum a,-AGP and a,-HS in normal subjects and alcoholic patients was studied by CIAE using free Con A in the first dimension gel. Gly~proteins are separated during the two-step ~lectrophoresis, and their electrophoretic mobility is affected by their affinity for the lectin used. Various authors have previously described a relationship between alcoholic liver damage and a decreased electrophoretic mobility of transfer+ [ll], tw,-AGP [12] and C,-inactivator [13], due to a loss of sialic acid residues. The undersialylation was not found to correlate with the corresponding glycoprotein plasma levels. In this work, with the addition of Con A, a second type of ~crohcterogeneity was demonstrate. Con A is a lectin which binds specifically to mannosyl residues proportionally to their accessibility. In liver damage, the affinity of (Y,-AGP and at,-HS appears to be modified towards an increase in unreactive forms. A similar observation for LU,-AGP has been related to an increase in tri- and tetra-antennary glycan structure [ 141, The results presented here confirm our previous work [6], despite the fact that the patients’ classification was different. Pugh’s classification, compared to that of Child and Turcotte, includes a further biological criterion and adds flexibility to the method of grading the severity of the liver disease. The present study, based on a larger alcoholic population and a more discriminating classification, allowed normal subjects to be separated from steatosic patients and the latter group from cirrhotic patients. The modifications in Con A patterns reflected closely the hepatocellular state. As shown in the iterative study of a steatosic patient’s serum, following 1.5 days’ alcoholic withdrawal and a subsequent histological improvement in his liver condition, Con A pattern modifications appear to be reversible, and can change rapidly. These rapid changes in human
56
serum are consistent with our results on the glycan microheterogeneities of serum crl-AGP in galactosamine-treated rats [15]. In the present study, the Con A pattern shift was identical for Cal-AGP and a2-HS. A similar type of modification has been observed with transferrin, in the same pathology [16]. On the contrary, in acute inflammatory disorders, the proportion of Con A-separated components of different proteins (oil-AGP, al-antichymot~psin) is altered towards a higher level of Con A-reactive components [4,5,17]. One can therefore postulate that different types or levels of dysfunction might affect a common glycosylation mechanism. This may explain the identical Con A-CIAE pattern shift of some, or all, serum glycoproteins in a given pathological state. Moreover, the lack of correlation between glycoprotein plasma levels and the two types of described microheterogeneity (sialylation and Con A reactivity) on the glycan moiety indicates that synthesis and glycosylation of glycoproteins are not modulated by the same regulatory mechanisms. The fact that qualitative alterations of serum glycoproteins may be as significant as quantitative changes, is a new concept. Further studies should take the importance of such modification into account. Acknowledgements
We wish to thank Dr. P.W. Becker (Ber~ngwerke AG) for the gift of cx2-HS antisera and immunoplates and David Young for revising the English form of the manuscript. This work was supported by grants from the Institut National Scientifique et de la Recherche Mtdicale (contrat libre externe No 857-007) and the Institut de Recherches Scientifiques Economiques et Sociales sur les Boissons. References 1 Naipas B, Botgne JM. Zafrani ES, Z~~e~ann R, et al. perturbations de 10 proteines pl~matlques au cours des hepatopathies alcooliques. Gastroenterol Clin Biol 1980;4:646-654. 2 Murray-Lyon IM, William R. Quantitative immunoelectrophoresis of plasma protems m acute vtral hepatitis. extrahepatic biliary obstruction, primary biliary cirrhosis and idiopathic haemochromatosis. Clin Chim Acta 1974;51:303-308. 3 Teppo AM, Maury CPJ. Serum albumm transfernn and immunoglobulins m fatty hver, alcoholic cirrhosis and primary biliary cirrhosis. Clin Chim Acta 1983;129:279-286. 4 Raynes J. Variations in the relative proportions of microheterogenous forms of plasma glycoproteins in pregnancy and disease. Biomedicine 1982;36:77-86. 5 Nicollet I, Lebreton J-P, Fontaine M, Hiron M. Evidence for al-acid ~ycoprotein populations of drfferent pl values after Concanavalin A affinity chromatography. Bmchim Biophys Acta 1981;668:235-245. 6 Serbource-Goguel Seta N, Durand G, Corbic M. Agneray J, FCger J. Alterattons m relative proportions of nucroheterogenous forms of human al-Acid glycoprotem in hver disease. J Hepatol 1986;2:245-252. 7 Whither J,d Dieppe PA. Acute phase proteins. In: Clinics in immunology and allergy, Vol 5, no 3, October 1985;~~ 425-446. 8 Lebreton JP. Joisel F, Raoult JP, Lannuzel B, et al. Serum concentration of human a2-HS glycoprotein during the ln~~ato~ process. J Clm Invest 1979;64:1118-1121.
9 Pugh RNH, Murray-Lyon IM, Dawson JL. Pietrom MC et al. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646-649. 10 Bog-Hansen TC. Affmity electrophoresis of glycoproteins. In: Scouten WH, ed. Sohd phase biochemistry, analytical and synthettc aspects. New York: J Wiley and Sons. 1983;223-251. 11 Stibler H, Borg S. Allgulander C. Abnormal heterogeneity of serum transfernn. A new diagnostic marker of alcoholism? Acta Psych Stand 1981;62 (suppl 286):189-194. 12 Serbource-Goguel N. Corbtc M, Erlinger S, Durand G, Agneray J and Ftger J. Measurement of serum al-acid glycoprotein and cYl-antitrypsm desialylatton m liver disease. Hepatology 1983;3:356-359. 13 Serbource-Goguel Seta N, Bordas M, Davy J. and Durand G. Evaluation of the degree of destalylation of serum Cl-inactivator and haemopexm. Clin Chim Acta 1984;143:235-241. 14 Biou D. Konan D, FCger J, Agneray J et al. Alterations in the carbohydrate moiety of al-acid glycoprotein punfied from human cirrhottc ascetic fluid. Bmchim Biophys Acta 1987;913:308-312. 15 Monnet D, Durand D, Biou D, Ftger J et al. o-galactosamme-Induced liver injury: a rat model to study the heterogeneity of the oligosaccharide chains of al-acid glycoprotein. J Clin Chem Clin Biochem 1985;23:249-253. 16 Spik G. Debruyne V and Montreuil J. Alterations of the carbohydrate structure of human serotransferrin. In: Popper H, Reutter W, Gudat F and Klittgen E (eds). Liver dtseases. structural carbohydrates in the liver. Lancaster: MTP Press, 1983;447-483. 17 KOJ A, Dubin A, Kasperczyk H, Bereta J, et al. Changes in blood level and affinity to Concanavalin A of rat plasma glycoproteins during acute inflammatron and hepatoma growth. Biochem J 1982;206:545-553.
49
Elsevier
CCA 04225
Modifications of Concanavalin A patterns of al-acid glycoprotein and (r2-HS glycoprotein in alcoholic liver disease Maryvonne Jezequel a, Nathalie S. Seta a, Michkle M. Corbic b, Jeanne M. Feger a and Genevikve M. Durand a ULaboratowe de Blochlmre A, Hsplral Blchar 7.5018 Paris and Laboratorre de Blochrmre, CNRS UA 622, UER des Scrences Pharmaceutiques. 92290 Chatenav-Malabry. and h INSERM lJ 24. H;pprtal Beaujon, 92 Chchv (France) (Received
19 November
1987: reviston received
Key words: Concanavalin
A; al-Acid
15 March
1988; accepted
glycoprotein: a2-HS Liver drsease; Alcohol
after revision
Glycoprotein;
26 April 1988)
Microheterogeneity;
In order to test whether abnormalities in hepatocytes affect the glycoprotein carbohydrate moiety, crossed immunoaffinoelectrophoresis (CIAE) with Concanavalin A (Con A) was used to study serum alpha l-acid glycoprotein (al-AGP) and alpha 2-HS glycoprotein (a2-HS) obtained from alcoholic patients with biopsy-proven liver disease. Cirrhotic patients, placed in groups Cl, C2 or C3, according to Pugh’s classification, were compared to healthy donors (N) and to steatosic non-cirrhotic patients (S). Con A CIAE patterns revealed in group N three subpopulations for a2-HS and four for d-AGP. Two main results emerged from this study: (1) in the alcoholic groups, the proportions of Con A-unreactive subpopulations of both glycoproteins increased. Moreover, group N could be separated from group S and group S from all the cirrhotic groups. (2) There was a good correlation between the relative amounts in Con A-unreactive subpopulations of d-AGP and a2-HS. The increases observed in Con A-unreactive subpopulations are probably a general phenomenom related to alterations in glycosylation processing during liver cell damage.
Correspondence Pans, France.
to: N. Seta, Laboratoire
0009-8981/88/$03.50
de Biochimie
0 1988 Elsevrer Science Publishers
A. HBpital
Bichat.
B.V. (Btomedical
46. rue H. Huchard,
Division)
75018
50
Introduction A number of chemical abnormalities related to excessive alcohol consumption and alcoholic liver injury have been detected. Particularly, changes in the concentration of plasma proteins have often been reported [l--3]. Lately, however, much interest has been shown in plasma glycoprotein microheterogeneities. The presence of glycoprotein ~croheterog~neities can be demonstrated directly in the serum by lectin-crossed immunoaffinoelectrophoresis (CIAE) in which the proteins are separated by their electrophoretic mobility and their retardation due to the affinity of their carbohydrate structures for lectin. Several authors studying serum a,-acid glycoprotein ( (Y,-AGP) and other glycoproteins in acute inflammatory disease have observed pattern modifications using CIAE in the presence of Concanavalin A (Con A-CIAE) with an increase in Con A-reactive forms [4,5]. These reports raise the question of a same regulatory system of glycoprotein glycan moiety synthesis in infla~ation. We have recently shown that (r,-AGP in alcohol-induced cirrhotic patients, exhibits an increase in Con A-unreactive microheterogenous forms, when separated by Con A-CIAE. This alteration has been shown to be related to liver damage [6]. The purpose of this work was: 1) to confirm and extend our previous results using larger groups and to compare them with those of non-cirrhotic alcoholic patients. 2) To study simultaneously by Con A-CIAE, the two glycoproteins a,-AGP and (u,-HS glycoprotein (LY*-HS), which are respectively positive and negative acute phase proteins [7,8]. Materials and methods Patients
Eighty-two alcoholic patients with biopsy-proven liver disease were studied (58 men, 24 women; age 24-72 years, mean 46 years). Histological findings included steatosis without cirrhosis (20 patients, group S) and cirrhosis with or without steatosis, fibrosis and alcoholic hepatitis (62 patients, group C). Cirrhotic patients were placed in group C,, C, or C,, according to Pugh’s classification [9], which is based on the grading (score l-3) of five clinical and biological complications of the cirrhotic process (encephalopathy, ascitis, bilirubinemia, albuminemia, prothrombin time). Details of the 62 cirrhotic patients are given in Table 1. Alcoholic patients were compared with 22 healthy donors (group N) (19 men, 3 non-pregnant women with no oestroprogestative treatment; age 21-39 years, mean 31 years) with alcohol consumption of between 0 and 15 g/day and with no known liver disease. Liver biopsies were not performed in the control subjects for ethical reasons. At the time of liver biopsy, blood specimens were drawn from the patients for the measurement of serum ALAT, ASAT, bilirubin, alkaline phosphatase, gammaglutamyltransferase ( y-GT), albumin, immunoglobulin A (&A). prothrombin time, proaccelerin, haemoglobin and mean corpuscular volume (MCV). Serum samples were collected at the same time and stored at - 20 o C.
51 TABLE
I
Conventional
liver function
Groups
N
n
22
Pugh’s score ASAT [5-25 UI/l) ALAT [5-25 UI,‘l] YGT [ -c30 UI/l] Alkaline phosphatase
test values in healthy S
and patients
14
+ 4
93
f
69 a
11
f
63
+
37 a
22
+11
200
73
rt23
92
with alcoholic
C2
Cl 20
5
donors
16 5-6 101 *107 66
=
+ 56 a
+247
366
+532
a
85
f
26
153
+ 42
&I.5
89
rt 12
70
f
15%
15
+
15
f
1.3
13
f
1.9
90
rt 4.5
f
5.7 a
99
f
7.5 =
liver disease f3
21 7-9 83 f
57 d
35
17 h,c
*
25 10-15 100 *
58 a
45
+ 27
401
&348 a
175
+136
251
+ 182 a,h
214
+177
67
rf: 19 a.’
36
4
10
&
2.2 a.b.c
i
8.4 ’
a.b
[ < 150VI/I] Proaccelenn [75-lOO%] Hemoglobin 112-17 g%] MCV [85-95 pm31
2.3f ;g-3.0
1.5
0.9
loo
2.3*
0.7
7.2+
11.75
5.0 a,h
103
i
5.8*
2.5 a.b 12.4 a 1.7
103
9.5 f
16 ‘.‘.’
3.5 a.b
g/l]
Results are expressed as mean+sD. Numbers in brackets N = healthy donors; group S = steatosis without cirrhosis; according to Pugh’s score. a vsN(p<0.002). h vss(p<0.001). c vs Cl (p i 0.001).
are the normal laboratory range. groups C = cirrhosis at different
Group stages.
Reagents Rabbit antisera to human oil-AGP and a2-HS and NOR Part&en oil-AGP and M Partigen cuZHS plates were purchased or gifts from Behringwerke, Marburg, FRG. Concanavalin A was from 1’Industrie Biologique Franqaise, Villeneuve-laGarenne, France. Agarose type II and a-D-methylglucopyranoside were from Sigma, USA. Polyethylene glycol (PEG 6000) and Coomassie blue G 250 were from Fluka and Merck (FRG) respectively. Crossed imm~noaffi~ity electrophoresis Crossed immunoaffinity electrophoresis (CIAE) was carried out essentially as described by Bog Hansen [lo]. The buffer (pH 8.7) contained 72 mmol/l Tris, 24 mmol/l Veronal, 0.4 mmol/l calcium lactate, 0.2 mmol/l sodium azide. Con A was dissolved at a final concentration of 5% in a solution containing 1 mmol/l MgCl,, M&l, and CaCl,. This solution was added to the first dimension gel (140 ,ul/cm2 for ail-AGP, 280 pi/cm’ for a2-HS). Specific antiserum was incorporated into the second dimension gel (7.5 &‘cm* for al-AGP, 15 $/cm2 for CYZ-HS)together with ~-D-methyl~ucop~anoside (4%) and PEG 6000 (2%). The gels were 1.5 mm-deep 1% agarose.
52
Sample Cal-AGP and (Y~-HS contents were determined by radial immunodiffusion and their concentrations were adjusted to 0.30 g/l + 0.05 so that the CIAE patterns would be comparable. Sample volumes of 4 ~1 ((wl-AGP) or 10 ~1 (a2-HS) of various dilutions were deposited on the gel. The first dimension was run at 9 V/cm for 1.5 h (al-AGP) or 2.5 h (a2-HS) and the second at 2.5 V/cm for 18 h for both glycoproteins. After drying and staining, peak areas corresponding to the various subpopulations were determined. Designating peak 1 as the most anodic, results were expressed as the surface ratio of the main peak (peak 1 for al-AGP and peak 3 for a2-HS) to the minor ones, ie
R&LAW = (I)/(2
+ 3 + 4) and &us
= (3)/(2
+ 1)
Statistical methods Calculations were made by analysis of variance for multiple comparison between groups. When a significant difference was found (0.05) comparison was performed with a modified t test and the Bonferroni method was used to determine the level of significance. Results Results of conventional liver function tests for healthy donors and alcoholic patient groups are reported in Table I. Serum ASAT, ALAT and the MCV differed significantly between groups N and S (sensitivity 70%, specificity 95%) and IgA between groups S and Cl (sensitivity 88%, specificity 82%).
TABLE
II
Comparative study variants of al-acid alcoholic patients
of plasma content and surface ratios of Con glycoprotein (al-AGP) and a2-HS glycoprotein
A-crossed ((~2-HS)
immunoelectrophoresis in healthy donors and
Groups
N
S
Cl
c2
n Pugh’s score (rl-AGP (g/l) Rh,.,,, = peaks l/(2 + 3 + 4) a2-HS (g/l)
22
20
16 5-6 0.52 * 0.28 b 1.53+0.53 a
21 25 7-9 10-15 0.83 f 0.57 0.50+ 1.75 f 0.87 a.b,c 1.88+
R aNIs=
0.72 f 0.12 0.72 f 0.07
0.91 f 0.24 1.07 f 0.47
0.66 + 0.08 3.04kO.42
0.63 + 0.13 2.37kO.55 =
0.59 f 0.16 1.75 *0.51 a.b
c3
0.56 * 0.12 1.68 + 0.62 a,h
0.21 0.77 h
0.34+0.13 a.h,‘ 1.78 + 0.51 a.b
peaks 3/(2 + 1) Results are expressed as mean f SD. Group N = healthy donors; group groups C = cirrhosis at different stages, according to Pugh’s score. a vs N (p i 0.002). b vss(p~0.005). c “S Cl (p i 0.005).
S = steatosis
without
cirrhosis;
53
1B
Fig. 1. Con A-crossed
immunoaffinoelectrophoresis patterns of serum d-acid a2-HS glycoprotein (B) from a healthy donor.
glycoprotein
(A) and
Total serum glycoprotein levels are presented in Table II. Total serum al-AGP concentrations were not found to be significantly different between groups, except for group Cl compared to the steatosis group. Total serum a2-HS concentrations did not differ significantly, except for group C3 which had significantly reduced levels. Typical Con A-CIAE patterns of al-AGP and a2-HS are presented in Figs. 1 and 2, respectively. Fig. 1A shows an al-AGP pattern with four peaks for a healthy donor: 1, Con A-unreactive; 2, weakly reactive; 3 and 4, the most reactive with Con A. Fig. 1B shows the corresponding a2-HS pattern: peak 1, Con A-unreactive and peaks 2 and 3, Con A-reactive. As can be seen in Figs 2A and B, showing al-AGP and a2-HS Con A-patterns from the same cirrhotic patient with a dramatic increase in Con A-unreactive peak 1 for both glycoproteins, the patterns are different but the direction of the shift is the same. These pattern modifications were quantified for all the groups through the corresponding ratio values (Table II). These ratios, R&t_,,, and Ra2_HS, gave a good representation of the observed pathologic-related variations: R&I_AGPvalues rose progressively from normal subjects (0.72 + 0.07) to cirrhotic group C3 (1.88 k
54
1
2A
-,- - ._
,,_
-7---
,
I
2B
Fig. 2. Con A-crossed
immunoaffmoelectrophoresls patterns of serum q-acid a,-HS glycoprotein (B) from a clrrhotlc patlent.
glycoprotein
(A) and
400.
RalAGP
. 200.
I. .
- . 2.00
4.00
RoZ-HS
Fig. 3. Correlation
between
ratios Rb,_ac,d glywprote,n (R&, = peaks l/2 + 3 + 4) and a2-HS glycoprotein (R& = peaks 3/2 + 1) (‘r’ = 0.70, n = 104).
0.77). This marked increase allowed group N to be differentiated from Group Cl (sensitivity 1008, specificity 82%) and group S from group C2 (sensitivity 47%, specificity 958). RazsHS values decreased progressively from normal subjects (3.04 + 0.42) to cirrhotic group C3 (1.75 rt 0.51). This decrease allowed group N to be separated from group S (sensitivity 65% specificity 82%) and from all the cirrhotic groups (~nsiti~ty 94%, specificity 82%); furth~~ore, group S could be separated from all the cirrhotic groups (se~siti~ty 52W, specificity 90%). By associating the two ratios R&i,, and RaZ_uS,each group coutd be separated from the others with good sensitivity and over 95% specificity. Moreover a significant correlation was found between R,,_Hs and R&,oe, as represented in Fig. 3 (correlation ‘r ’ = 0.70, n = 104). Finally, no correlation was found between the two ratios R&l_AGPand RaZ_HS,and the corresponding glycoprotein plasma levels. A Con A-CIAE analysis of LU,-AGPwas carried out using the serum of a steatosic patient during an alcoholic withdrawal period of 15 days: a histologically proven improvement in steatosis was observed and the &-AGP pattern was modified, with a decrease in the Con A-unreactive peak (before: R',,_,,, = 1.14; after: RL,_,o, = 0.68). Discussion The biochemical heterogeneity of serum a,-AGP and a,-HS in normal subjects and alcoholic patients was studied by CIAE using free Con A in the first dimension gel. Gly~proteins are separated during the two-step ~lectrophoresis, and their electrophoretic mobility is affected by their affinity for the lectin used. Various authors have previously described a relationship between alcoholic liver damage and a decreased electrophoretic mobility of transfer+ [ll], tw,-AGP [12] and C,-inactivator [13], due to a loss of sialic acid residues. The undersialylation was not found to correlate with the corresponding glycoprotein plasma levels. In this work, with the addition of Con A, a second type of ~crohcterogeneity was demonstrate. Con A is a lectin which binds specifically to mannosyl residues proportionally to their accessibility. In liver damage, the affinity of (Y,-AGP and at,-HS appears to be modified towards an increase in unreactive forms. A similar observation for LU,-AGP has been related to an increase in tri- and tetra-antennary glycan structure [ 141, The results presented here confirm our previous work [6], despite the fact that the patients’ classification was different. Pugh’s classification, compared to that of Child and Turcotte, includes a further biological criterion and adds flexibility to the method of grading the severity of the liver disease. The present study, based on a larger alcoholic population and a more discriminating classification, allowed normal subjects to be separated from steatosic patients and the latter group from cirrhotic patients. The modifications in Con A patterns reflected closely the hepatocellular state. As shown in the iterative study of a steatosic patient’s serum, following 1.5 days’ alcoholic withdrawal and a subsequent histological improvement in his liver condition, Con A pattern modifications appear to be reversible, and can change rapidly. These rapid changes in human
56
serum are consistent with our results on the glycan microheterogeneities of serum crl-AGP in galactosamine-treated rats [15]. In the present study, the Con A pattern shift was identical for Cal-AGP and a2-HS. A similar type of modification has been observed with transferrin, in the same pathology [16]. On the contrary, in acute inflammatory disorders, the proportion of Con A-separated components of different proteins (oil-AGP, al-antichymot~psin) is altered towards a higher level of Con A-reactive components [4,5,17]. One can therefore postulate that different types or levels of dysfunction might affect a common glycosylation mechanism. This may explain the identical Con A-CIAE pattern shift of some, or all, serum glycoproteins in a given pathological state. Moreover, the lack of correlation between glycoprotein plasma levels and the two types of described microheterogeneity (sialylation and Con A reactivity) on the glycan moiety indicates that synthesis and glycosylation of glycoproteins are not modulated by the same regulatory mechanisms. The fact that qualitative alterations of serum glycoproteins may be as significant as quantitative changes, is a new concept. Further studies should take the importance of such modification into account. Acknowledgements
We wish to thank Dr. P.W. Becker (Ber~ngwerke AG) for the gift of cx2-HS antisera and immunoplates and David Young for revising the English form of the manuscript. This work was supported by grants from the Institut National Scientifique et de la Recherche Mtdicale (contrat libre externe No 857-007) and the Institut de Recherches Scientifiques Economiques et Sociales sur les Boissons. References 1 Naipas B, Botgne JM. Zafrani ES, Z~~e~ann R, et al. perturbations de 10 proteines pl~matlques au cours des hepatopathies alcooliques. Gastroenterol Clin Biol 1980;4:646-654. 2 Murray-Lyon IM, William R. Quantitative immunoelectrophoresis of plasma protems m acute vtral hepatitis. extrahepatic biliary obstruction, primary biliary cirrhosis and idiopathic haemochromatosis. Clin Chim Acta 1974;51:303-308. 3 Teppo AM, Maury CPJ. Serum albumm transfernn and immunoglobulins m fatty hver, alcoholic cirrhosis and primary biliary cirrhosis. Clin Chim Acta 1983;129:279-286. 4 Raynes J. Variations in the relative proportions of microheterogenous forms of plasma glycoproteins in pregnancy and disease. Biomedicine 1982;36:77-86. 5 Nicollet I, Lebreton J-P, Fontaine M, Hiron M. Evidence for al-acid ~ycoprotein populations of drfferent pl values after Concanavalin A affinity chromatography. Bmchim Biophys Acta 1981;668:235-245. 6 Serbource-Goguel Seta N, Durand G, Corbic M. Agneray J, FCger J. Alterattons m relative proportions of nucroheterogenous forms of human al-Acid glycoprotem in hver disease. J Hepatol 1986;2:245-252. 7 Whither J,d Dieppe PA. Acute phase proteins. In: Clinics in immunology and allergy, Vol 5, no 3, October 1985;~~ 425-446. 8 Lebreton JP. Joisel F, Raoult JP, Lannuzel B, et al. Serum concentration of human a2-HS glycoprotein during the ln~~ato~ process. J Clm Invest 1979;64:1118-1121.
9 Pugh RNH, Murray-Lyon IM, Dawson JL. Pietrom MC et al. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646-649. 10 Bog-Hansen TC. Affmity electrophoresis of glycoproteins. In: Scouten WH, ed. Sohd phase biochemistry, analytical and synthettc aspects. New York: J Wiley and Sons. 1983;223-251. 11 Stibler H, Borg S. Allgulander C. Abnormal heterogeneity of serum transfernn. A new diagnostic marker of alcoholism? Acta Psych Stand 1981;62 (suppl 286):189-194. 12 Serbource-Goguel N. Corbtc M, Erlinger S, Durand G, Agneray J and Ftger J. Measurement of serum al-acid glycoprotein and cYl-antitrypsm desialylatton m liver disease. Hepatology 1983;3:356-359. 13 Serbource-Goguel Seta N, Bordas M, Davy J. and Durand G. Evaluation of the degree of destalylation of serum Cl-inactivator and haemopexm. Clin Chim Acta 1984;143:235-241. 14 Biou D. Konan D, FCger J, Agneray J et al. Alterations in the carbohydrate moiety of al-acid glycoprotein punfied from human cirrhottc ascetic fluid. Bmchim Biophys Acta 1987;913:308-312. 15 Monnet D, Durand D, Biou D, Ftger J et al. o-galactosamme-Induced liver injury: a rat model to study the heterogeneity of the oligosaccharide chains of al-acid glycoprotein. J Clin Chem Clin Biochem 1985;23:249-253. 16 Spik G. Debruyne V and Montreuil J. Alterations of the carbohydrate structure of human serotransferrin. In: Popper H, Reutter W, Gudat F and Klittgen E (eds). Liver dtseases. structural carbohydrates in the liver. Lancaster: MTP Press, 1983;447-483. 17 KOJ A, Dubin A, Kasperczyk H, Bereta J, et al. Changes in blood level and affinity to Concanavalin A of rat plasma glycoproteins during acute inflammatron and hepatoma growth. Biochem J 1982;206:545-553.