Close topographical relationship in alpha foetoprotein (AFP) between a lens culinaris binding glycan and the epitope recognized by AFP-reactive monoclonal antibody, 18H4.

Monoclonal antibodies 18H4 and 19F12 against alpha-foetoprotein (AFP) were examined by enzyme immunoassay for binding to two forms of AFP that were separated on the basis of the reactivity with lentil lectin (LCA). LCA-binding and LCA non-binding AFP, coated on a solid phase, reacted with 18H4 but reactivity with the LCA-binding species was inhibited by 60% following pretreatment of the AFP with LCA. The lectin was a very poor inhibitor of binding of 18H4 to the AFP which did not interact with LCA. In an alternative binding assay, a polyclonal anti-AFP coated solid phase was reacted with beta-galactosidase-labelled 18H4. Pre-treatment with LCA of the LCA-reactive AFP gave 56% inhibition of binding of conjugated 18H4 while little inhibition was achieved with the LCA-nonreactive AFP component. These findings show that the epitope recognised by 18H4 is distinct from the glycan sequence that is reactive with LCA. However, the LCA-binding oligosaccharides occur in close proximity to the 18H4 epitope in the native AFP.


Alpha-foetoprotein (AFP) is one of the major plasma
Materials and methods glycoproteins in the early foetal stage (Rouslahti & Seppala, 1979), and a useful tumour marker for some malignant AFP neoplasms such as hepatocellular carcinomas, terato-AFP was isolated from serum of a patient with hepatocarcinomas and germ cell tumours (Abelev, 1968; cellular carcinoma by affinity chromatography and DEAEal., 1968;Smith, 1970;Kurman et al., 1977). Serial Sephadex chromatography (Pharmacia Fine Chemicals measurement of the serum concentration of AFP may be Sephadex c atography phrmaia Fine Ce als, important in the follow-up of patients with chronic liver Uppsala Sweden) as described previously (Aoyagi et al.c diseases (Nishi & Hirai, 1973;Okuda et al., 1980). The 1977). The LCA-reactive and nonreactive molecular species serum concentration of AFP, however, also increases in of AFP were obtained by affinity chromatography with chronic liver diseases such as hepatic cirrhosis and chronic lC p e P ch a) g hepatitis (Rouslahti et al., 1974;Lehmann, 1976; Alpert & al., 1985). Feller, 1978). Therefore, it would be desirable if the molecular species of AFP occurring in hepatocellular Chemicals carcinomas can be distinguished from those in non-Salt-free lyophilized powder of LCA (L-5880) was purchased neoplastic liver diseases.
from Sigma Chemical Co., St Louis, MO., USA. Our previous study has shown that the chemical structures Horseradish peroxidase-labelled conjugate of goat anti-mouse of foetal and hepatocellular carcinoma-derived AFPs are IgG was obtained from Jackson Immuno Research almost identical except for a difference in carbohydrate Laboratories Inc. (Avondale, PA, USA). fl-D-galactoside composition (Aoyagi et al., , 1978(Aoyagi et al., , 1979(Aoyagi et al., , 1982. Several galactohydrolase was from Boeringer Mannheim Bioinvestigators have reported heterogeneous reactivity of AFP chemica (W. Germany). with lectins (Smith et al., 1977;Bayard & Kerckaert, 1977;Miyazaki et al., 1981; Monoclonal anti-AFP antibodies Breborowicz et al., 1981;Taketa et al., 1983) (Aoyagi et al., 1984(Aoyagi et al., , 1985a(Aoyagi et al., , 1986. human foetal cord serum. These monoclonal antibodies were Ths eto ha indiate an. inres of th LC-reactive........... purified from ascites fluids by affinity chromatography with species of AFP in patients ithrea to eLu Arcacinm Protein A-Sepharose 4B (Pharmacia Fine Chemicals). 1 8H4 specieto al-} 1984)-patientsylatin ohepato lular charinomas t was of the IgGi subclass, and 19F12 was of the IgG2b. The (Aoyagi et al., 1984). Fucosylation of the sugar chain is the isolated immunoglobulins were pure as revelaed by sodium molecular basis for this variation of AFP (Aoyagi et al., dodecyl sulphate polyacrylamide gel electrophoresis andim-1985a,b, 1986). However, the method with crossed immunoaffino-electropnoresis requires a Skilfl techninque. aKeta munoelectrophoresis. et al. (1985) have recently reported an antibody-affinity blotting method that is able to distinguish LCA-reactive and Polyclonal anti-AFP antibody nonreactive species of AFP at very low concentration of AFP. Polyclonal anti-AFP antibody (PcAb) was obtained by im-Although this method is beneficial for the discrimination of munization of rabbits with pure AFP, and purified using AFP species, it also requires skilful technique.
AFP coupled-agarose column as previously described The present work was initiated to develop a simpler . method to distinguish the LCA-reactive species of AFP from the LCA-nonreactive species by enzyme immunoassay using AEP-coated polystyrene beads monoclonal antibody.
test. Data are represented as the mean + s.d.

PcAb-coated polystyrene beads Results
Polystyrene beads were coated with 15 pg ml1 of purified PcAb in buffer A at 4°C for one week. Figure  Inhibition with LCA of the binding of 18H4 to the solid assay, the following experimental procedures were phase AFPs was measured with the monoclonal antibody performed. AFP-coated beads were preincubated at 4°C for at concentrations of 450, 750, 1125 and 1500ngml-1 3 h with 100 pg ml-I of LCA in buffer A in order to allow ( Figure 2a). These concentrations were in the range over this lectin to bind to a carbohydrate chain of the LCAwhich the antibody was bound in a dose-dependent manner reactive species of AFP. After washing with buffer A on to the solid phase (Figure la). filter paper, pretreated beads were incubated with various The results indicated that pretreatment with LCA of the amounts of monoclonal antibody at 4°C for 2 h. The solid solid phase LCA-reactive species of AFP resulted in phase polystyrene beads for control experiments were not approximately 60% inhibition of the binding of 18H4 to the pretreated with LCA. The beads were then incubated at 4°C solid phase at maximum. With 19F12 only 27% inhibition for 1 h with horseradish peroxidase-labelled goat anti-mouse was attained at maximum. Thus, 19F12 may serve as a IgG anti-serum. The enzyme activity associated with the reference monoclonal antibody. At all 4 concentrations solid phase was determined by measuring the absorbance at examined, the % inhibition with LCA for the binding of 492 nm after incubation with H202 and o-phenylenediamine 18H4 to the solid phase LCA-reactive AFP species was at room temperature for 30min.
higher than that in the system with 19F112, and the In control experiments, the AFP-coated beads were differences between 18H4 and 19F12 were statistically allowed to react directly with monoclonal antibody at significant (P<0.02-P<0.001) ( Figure 2a). varying concentration which was determined using a value of When the LCA-nonreactive AFP species was used as a E28°n-=15, and the enzyme activity associated with the solid phase, pretreatment with LCA had little effect on the beads determined. Concentrations of monoclonal antibodies binding of 18H4 to the solid phase (Figure 2b), and the bound to the solid phase in the control experiments were difference in the inhibitory effects of this treatment was determined in quintuplicate and the results are presented as significant (P<0.05-P<0.001) between the systems with the mean of triplicate values with the elimination of the two different LCA molecular species of AFP (Figure 2b). maximum and minimum values. Concentrations of mono-Pretreatment of LCA did not affect the binding of 18H4 clonal antibodies bound to the solid phase in the inhibition to the solid phase coated with the LCA-nonreactive species assay with LCA were measured in septuplicate and the of AFP (Figure 2c), indicating that the binding of 18H4 to a results expressed as the mean of quintuplicate values with the solid phase was inhibited with LCA only when the solid elimination of the maximum and minimum values. phase was coated with the LCA-reactive species of AFP. Secondly, as a model system for the analysis of serum Figure 3a shows a standard curve in the control enzyme samples of AFP species, the following procedures were immunoassay system with the LCA-reactive species of AFP carried out: PcAb-coated beads were incubated with 200trapped by PcAb-coated solid phase at indicated 2,000ngml-1 of either LCA-reactive or -nonreactive species concentrations and with fl-D-galactosidase-labelled McAb of AFP in buffer A at 4°C for 2 h in order to allow each 18H4. This standard curve without the LCA incubation of AFP species to bind to solid phase PcAb. After washing the AFP trapped by solid phase PcAb showed a good dosewith buffer A, these beads were incubated at 4°C for 3 h dependent relationship from 200 to 2,000ngml-' of LCAwith 100 jg ml-I of LCA in buffer A in order to allow this lectin reactive species of AFP. A similar dose-dependent standard to bind to a carbohydrate chain of the LCA-reactive species curve was also obtained in the system with conjugate 18H4 of AFP trapped by solid phase PcAb. The solid phase and the LCA-nonreactive species of AFP trapped by solid polystyrene beads for control experiments were not treated phase PcAb (Figure 3b). with LCA. The beads were then incubated at 37°C for 2 h Inhibition with LCA of the binding of conjugate 18H4 to with JJ-D-galactosidase-labelled McAbl8H4. Tne enzyme the LCA-reactive species of AFP trapped by solid phase activity associated with the solid phase was determined by PcAb was measured at AFP concentrations of 200, 400, 600, measuring the absorbance at 405 nm after incubation with 800 and 1,000 ng ml-1 (Figure 4a). These AFP concentrations p-nitrophenyl-,B-D-galactopyranoside (Nakarai Chemicals, were in the range where a good dose-dependent relation-Ltd., Kyoto, Japan) at 37°C for 2 h.
ship was observed (Figure 3a). The results indicated that In control experiments, the PcAb-coated beads were the LCA treatment for the LCA-reactive species of AFP allowed to react with AFP at varying concentration which trapped by solid phase PcAb resulted in -~56% inhibition was determined using a value of E28P0m= 5.3, and the of binding of LCA-reactive species of AFP to conjugate enzyme activity associated with the beads determined.

Concentration of McAb (ng ml-')
Figure 2 (a) The effects of LCA pretreatment on the binding of monoclonal antibodies to the solid phase LCA-reactive species of AFP. LCA-reactive AFP coated polystyrene beads were pretreated with LCA (100 gml-1) at 4°C for 3h and then allowed to react with either 18H4 (0) or 19F12 (0) at concentrations of 450, 750, 1,125 and l,SOOngml-1. The bound antibody was quantitated as described in the legend of Figurel(a) and % inhibition was calculated by comparison with the data from control experiments in which the LCA pretreatment was omitted. Values represent the mean + s.d. (vertical bars, n = 5). The differences in the mean % inhibition were statistically significant between the experiments with 18H4 and 19F12 at concentrations of 450 ng ml 1 (P < 0.02), 750 ng ml -(P < 0.02), 1,125 ng ml -1 (P < 0.01) and 1,500 ng ml -1 (P < 0.001).   10~00 Concentration of AFP (ng ml-) Figure 4 (a) Effects of LCA treatment on the binding of ,B-D-galactosidase-labelled 18H4 to the LCA-reactive species of AFP trapped by PcAb-coated solid phase. PcAb-coated polystyrene beads were incubated with LCA-reactive species of AFP at indicated concentrations at 37°C for 2 h. After washing, the beads were treated (0) with LCA (100 ,gmP 1) at 4°C for 3 h or untreated (0), and were then allowed to react with conjugate 18H4. The conjugate 18H4 bound to the LCA-reactive species of AFP was quantitated as described in the legend of Figure 3 PcAb. PcAb-coated polystyrene beads which were incubated with either LCA-reactive (0) or LCA-nonreactive (0) species of AFP were treated with LCA. The beads were then allowed to react with conjugate 18H4, and conjugate 18H4 bound to each AFP species was quantitated as described in the legends of Figures 3(a) and 4(a). Inhibition % values shown are the mean + s.d.
When the LCA-nonreactive species of AFP was examined, treatment with LCA had little effect on the binding of conjugate 18H4 to the AFP trapped by solid phase PcAb (Figure 4b), and the difference in the inhibitory effect of this treatment was statistically significant (P <0.05-P <0.001) for the systems with two different LCA molecular species of AFP (Figure 4b).
Discussion Application of the monoclonal antibody technique to the quantitative measurement of serum markers in patients with several diseases has been described (Wands et al., 1982;Hedin et al., 1983). Bellet et al. (1984) recently reported the specific radioimmunoassay method for hepatocellular carcinoma with use of monoclonal anti-AFP antibodies which were supposed to recognize certain unique epitopes of AFP. They have claimed that the method is more useful than conventional radioimmunoassays for the detection and monitoring of AFP-producing tumours in high risk populations, and for distinguishing hepatocellular carcinomas from nonmalignant liver diseases or healthy subjects.
We have recently found that the degree of fucosylation of AFP is a good marker for distinguishing hepatocellular carcinomas from nonmalignant liver diseases (Aoyagi et al., 1984(Aoyagi et al., , 1 985a, 1986. The fucosylated and non-fucosylated molecular species of AFP can be measured by crossed immuno-affinoelectrophoresis with LCA, by taking advantage of the reactivity of the fucosylated species with this lectin. Similar methods have been widely used for diagnosis of neural tube defects (Smith et al., 1979;Toftager-Larsen et al., 1983) and liver diseases (Miyazaki et al., 1981;Breborowicz et al., 1981) with lectins such as Con A and LCA. Recently a new attempt by immunoblotting technique for the discrimination of AFP species has been reported by Taketa et al. (1985).
However, the method of crossed immuno-affinoelectrophoresis and antibody-affinity blotting has certain limitations as described above. Therefore, we aimed to provide a more convenient method which distinguishes between the fucosylated and non-fucosylated molecular species of AFP.
In the present study, we prepared monoclonal antibody 18H4 for such a purpose. The binding of 18H4 to the solid phase coated with the LCA-reactive species of AFP was inhibited by pretreating the solid phase with LCA ( Figure 2b). Such inhibition was not observed when the solid phase was coated with the LCA-nonreactive species of AFP (Figure 2b). In similar experiments with monoclonal antibody 19F12, LCA only slightly inhibited the binding of this antibody to the solid phase coated with the LCAreactive species of AFP (Figure 2a).
As described above, McAbl8H4 was selected in the system with AFP-coated solid phase. Thereafter we developed a model system for the analysis of serum samples of AFP species. PcAb-coated polystyrene beads were prepared, and incubated with LCA-reactive or nonreactive (Figures 4a, b) species of AFP at various concentrations from 200 to 2,000 ng ml-1. These AFP species trapped by solid phase PcAb were then treated with LCA, and incubated with /3-Dgalactosidase-labelled 188H4. The bound conjugate of 18H4 to AFP species was quantified, and % inhibition was calculated by comparison with the data from control experiments without LCA treatment. The binding of conjugate 18H4 to the LCA-reactive species of AFP trapped by solid phase PcAb was inhibited by treating with LCA (Figures 4a, b). Such inhibition was not obtained when the LCA-nonreactive species of AFP was examined in the system with PcAb-coated polystyrene beads and with LCA treatment (Figure 4b).
Inhibition of the binding of 1 8H4 to the LCA-reactive species of AFP can be explained by assuming that this monoclonal antibody recognizes an epitope of AFP which is closely located to the attachment site of a carbohydrate chain. (Pre)treatment with LCA of the LCA-reactive species of AFP would result in the binding of this lectin to its fucosylated sugar chain, and prevent 18H4 from binding to the AFP-molecular species by steric hindrance. The reason why complete inhibition was not achieved by the (pre)treatment with LCA may be that 1 8H4 does not recognize a carbohydrate chain itself.