(CANCER RESEARCH 49. 361-366. January 15. 1989]
Detection of Membrane-bound a-Fetoprotein in Human Hepatoma Cell Lines by
Monoclonal Antibody 19F12
Saiko Hosokawa,' Minoru Muramatsu, and Kazuhiro Nagaike
Biosciences Laboratory, Research Center, Mitsubishi Kasei Corporation, 1000 Kamoshida-cho, Midori-ku, Yokohama, Kanagawa 227, Japan
ABSTRACT
Monoclonal antibodies against human a-fetoprotein (AFP) were ob
tained by the hybridoma technique and studied with regard to their
reactivities with the human hepatoma cell lines PLC/PRF/5 and KN, and
a spontaneously immortalized cell line derived from fetal liver, NuE, all
of which synthesize AFP. One of the monoclonal antibodies, 19F12
(IgG2b) became bound to free AFP which was used as the immunogen
with an affinity constant of 3.4 x IO8 M"'. This value was not much
higher than those of two other antibodies, 19B1 (IgGl) and 9D12
(IgG2b). However, only antibody 191 12 showed definite reactivity with
AFP-producing cells in analysis using flow cytometry. Immunofluorescence microscopy showed that antibody 19F12 detected AFP over the
surface of NuF and PLC/PRF/5 cells with a uniform distribution, whereas
definite reactivities of antibodies 19B1 and 9D12 to these cells were not
detected. These antibodies did not show the specific binding to a nonAFP-producing human lung cancer cell line, PC-9, or to human peripheral
blood lymphocytes. The binding ability of 19F12 to hepatoma cells was
shown in both viable and fixed cells. Addition of free AFP inhibited the
binding of antibody I'll 12 to PLC/PRF/5 cells in a concentrationdependent manner. The specific reactivity of 11112 to human AFP was
also confirmed by immunostaining of a tissue section of human cancer
proved to be AFP positive with AFP-specific antisera. In two-dimensional
polyacrylamide gel electrophoresis of the antigen (from membrane frac
tion of PLC/PRF/5 cells)-antibody (19F12) complex, spots derived from
the antibody and a spot (pi 4.7, U, 65,000) corresponding in pi and
molecular weight to AFP were detected. Western blot analysis showed
that material in the membrane fraction of PLC/PRF/5 cells recognized
by antibody 19F12 has the same molecular weight as human AFP derived
from placenta. In a study of reactivities to PLC/PRF/5 cells treated with
various enzymes, the reactivity of this antibody decreased when cells
were treated with protease and trypsin and increased when lipase was
used. The binding of 19F12 to AFP was not inhibited by concanavalin A.
The antibody 19112 appeared to recognize an epitope that is considered
to be part of the peptide area of AFP. These results indicate that the
reactivity, the amount of bound antibodies, and the distribution of mono
clonal antibodies on antigen-producing cells vary, respectively, even
though these antibodies were produced using the same antigen as an
immunogen. Monoclonal antibody 191 12 binds to the epitope of AFP
present on the membrane surface of hepatoma cells with a wide and
uniform distribution. This antibody may therefore be a suitable one for
immunotoxin therapy or imaging of AFP-producing cancer cells.
INTRODUCTION
The preparation of monoclonal antibodies against tumorassociated antigens using the hybridoma technique (1) has
stimulated much interest with regard to their use in tumor
detection by radiolabeling and the targeting of cytotoxic agents
to tumor tissues. Success in these procedures depends on the
specificity of antibody to target tumors.
AFP2 is a fetal serum glycoprotein produced by the liver and
yolk sac. Liver cell carcinomas and teratomas containing yolk
Received 12/4/87; revised 4/29/88, 9/29/88; accepted 10/11/88.
The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' To whom requests for reprints should be addressed.
2The abbreviations used are: AFP, a-fetoprotein; SDS, sodium dodecyl sulfate;
PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline: Con
A, concanavalin A.
sac tissue often produce large amounts of AFP and this is
helpful in the diagnosis of these tumors and as a marker during
therapy (2, 3).
AFP is recognized to be a secretory-type protein (4). Never
theless, it has been reported that horse serum or purified
antibodies to rat AFP show a cytotoxic effect on AFP-producing
cells both in vitro (5, 6) and in vivo (7, 8). Furthermore it has
been found to be a suitable target for immunolocalization of
tumors (9-12).
Tsukada et al. (13) showed that polyclonal and monoclonal
antibodies to AFP conjugated to daunomycin show specific
cytotoxicity and growth inhibition to AFP-producing hepatoma
cells in vitro and in vivo.
However, Tsukazaki et al. (14) reported that ricin toxin A
chain-conjugated monoclonal antibody to AFP showed no spe
cific cytotoxicity to AFP-producing hepatoma cells. They con
sidered that the discrepancy between the former results and
theirs may have been due to the difference in cytotoxic drugs
used as conjugates and that ricin toxin A chain was not able to
exert its cytotoxic activity because AFP is a secreted protein.
We also reported that a monoclonal antibody against human
AFP, 19F12, conjugated with liposome containing Adriamycin,
showed antitumor effects both in vivo and in vitro. This effect
was specific to AFP-producing human hepatoma cells and it
was not shown when normal mouse IgG (IgG2b fraction) was
conjugated to liposome (15).
As a continuation of this line of study, we investigated the
reactivities of anti-AFP monoclonal antibodies to human hep
atoma and other cell lines in order to determine whether the
success or failure of toxin-targeting therapy depends on the
nature of the toxin, the antigen selected, or the reactivity of the
antibody. The results reported here suggested that the reactivity
of the antibody may be the major determining factor.
MATERIALS
AND METHODS
Cells and Culture. The hepatoma cell lines PLC/PRF/5 (16) and
KN, and a spontaneously immortalized cell line derived from a fetal
liver cell, NuE, were obtained from Dr. N. Ishida (Tohoku University,
Sendai, Japan). A lung cancer cell line, PC-9 (17), was obtained from
IBL (Gunma, Japan). These cells and hybridomas were grown in
Dulbecco's modified Eagle's medium:RPMI 1640 (1:1) with 10% fetal
calf serum. In the immunocytochemical studies, short-term (12 h)
serum-free cultured cells were used.
Production of Anti-Human AFP Monoclonal Antibodies. BALB/c
mice were immunized with purified human AFP derived from placenta
(purity more than 99% detected by SDS-PAGE) (a gift from Morinaga
Biochemical Research Center, Tokyo, Japan). Spleen cells from im
munized mice were fused with P3U1 myeloma cells. Antigen-specific
antibody-producing hybridomas were screened by enzyme-linked immunosorbent assay, and four monoclonal antibodies were chosen for
studies. Antibodies were purified from culture supernatant by affinity
chromatography on protein A-Sepharose CL-4B (Pharmacia). Sub
classes of antibodies were determined by diffusion assay using antimouse IgG subclasses (ICN immunoBiologicals, Lisle, IL).
Detection of Reactivities of Monoclonal Antibodies to Various Cell
Lines Using Flow Cytometry. Cultured cells were dispersed by incuba
tion in 0.02% EDTA-containing PBS. These cells ( 106cells) and normal
361
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
DETECTION OF MEMBRANE-BOUND
human peripheral blood lymphocytes were washed with PBS and in
cubated with 1 ml of 10 Mg/ml of purified anti-AFP antibody. The
concentration of antibody was determined from titration study. Reac
tivity of antibody 19F12 to hepatoma cells saturated on this level. Then
cells were stained with fluorescein isothiocyanate-labeled goat antimouse antibody (Cappel Laboratories, Cochranville, PA), diluted 1:20.
Binding of the antibody to the cells was examined using a FACS 440
flow cytometer (Becton/Dickinson).
Immunofluorescence Study by Fluorescence Microscopy and Inuminostaining of Tissue Sections. PLC/PRF/5 and NuE cells were cultured
on slide glasses and fixed with 4% (w/v) paraformaldehyde in PBS for
20 min at room temperature. Formalin-fixed tissue sections of human
hepatoma proved to be AFP positive were deparaffinized. These cells
and sections were incubated with 5% bovine serum albumin in PBS for
l h at 37"C to minimize any nonspecific binding. They were then
incubated with 10 ^g/ml of monoclonal anti-AFP antibodies, normal
mouse IgG (IgGl or IgG2b fraction) that were purified from normal
mouse serum by affinity chromatography ( 18), or polyclonal rabbit antiAFP antibody for 2 h at 37°C.Fluorescein isothiocyanate-labeled goat
anti-mouse antibody (Cappel) or anti-rabbit antibody (Cappel) were
used for detection of the bound antibodies.
Binding Activities of Antibodies to AFP. The equilibrium constants
of antibodies to human AFP were determined by solid-phase immunoassay using the method previously described (13, 19). However, we
used microassay plates (Falcon No. 3912) instead of paper discs, and
peroxidase-conjugated anti-mouse IgG (Cappel) was used for measuring
the antibodies bound to AFP.
Biosynthetic Labeling of Cells. Hybridoma cells (5 x IO6 cells)
secreting I9F12 antibody were grown in 5 ml of leucine-free RPMI
1640 (GIBCO) supplemented with 5 MCi/ml of [14C]leucine for 60 h.
Radionuclide-labeled antibodies were purified by affinity chromatog
raphy on protein A-Sepharose CL-4B from culture supernatant. Puri
fied antibody was deionized using a PD-10 column equilibrated with
PBS.
Competition Assay. 14C-Labeled 19F12 antibody (3000 cpm) and
various concentrations (0-30 ^g) of pure AFP (immunogen) were added
to IO5 PLC/PRF/5 cells. The mixture was incubated for 12 h at 4°C
and centrifuged for 10 min at 3000 rpm. Radioactivities in the super
natant and in the cell pellet were counted.
Two-Dimensional PAGE of Antibody 19F12-Antigen(from PLC/PRF/
5 Cells) Complex. PLC/PRF/5 cells (5 x 10" cells) were homogenized
in a Polytron in 5 ml of water for 5 min at 4°Cand centrifuged at
AFP
peroxidase-conjugated goat antibodies to mouse immunoglobulins
(Cappel). The activity of peroxidase revealing the amount of antibody
19F12 bound to cells was detected by colorimetrie assay (reaction
mixture: 1 mM 2,4-dichlorophenol, 0.2 mM 4-aminoantipyrine, and 1
mM H2C>2dissolved in 50 mM phosphate buffer, pH 7.0). Color devel
opment was quantitated by measuring the absorbance at 500 nm.
RESULTS
Reactivities of Anti-Human AFP Antibodies to Various AFPproducing Human Cell Lines. Four monoclonal antibodies to
human AFP were selected, and their reactivities with AFPproducing and non-AFP-producing cell lines were tested using
flow cytometry (Fig. 1). Antibody 19F12 (IgG2b) showed defi
nite binding activity to all AFP-producing cell lines (PLC/
PRF/5, KN, and NuE). However, the reactivities to each of
these cell lines were different. Binding of this antibody to NuE
cells and PLC/PRF/5 cells was shown in almost all of the cells
involved, but a difference was observed in the amount of bound
antibodies. It seemed that an almost equal amount of antibody
per cell became bound to NuE cells, while PLC/PRF/5 cells
could be divided into two or three types according to the amount
of bound antibody per cell. In contrast with these reactivities,
antibody 19F12 did not show binding to a minor proportion of
KN cells. The reactivities of antibodies 9D12 (IgG2b) and 19B1
(IgGl) with these cell lines were also detected in all of the cell
lines tested, but the amount of reactivity was sparse for both
(data for 19B1 are not shown because they were almost identical
to those for 9D12). Antibody 13B1 (IgGl) had no reactivity to
NuE cells at all, while binding could be detected only in a small
proportion of KN cells and a minor proportion of PLC/PRF/
5 cells. None of these antibodies showed specific binding activ
ity with the non-AFP-producing cancer cell line, PC-9 (Fig.
ID), or peripheral blood lymphocytes (data not shown).
Next, in order to investigate the reason for the difference in
reactivity of antibody 19F12 with these cell lines and those of
antibodies 9D12 and 19B1, the binding profiles of these anti-
100,000 x g for 60 min. The precipitate was used as membrane fraction.
The precipitate was solubili/ed by incubation in 100 ¿il
of 20 mM TrisHC1 buffer (pH 8.0) containing 1% Triton X-100 for 30 min. This was
then diluted with 9 volumes of PBS and mixed with 2 Mgof 19F12
antibody for l h at 37°C.Antigen-antibody complex was purified by
affinity chromatography on protein A-Sepharose CL-4B. SDS twodimensional PAGE of the complex was performed using the method
of O'Farrell (20). After electrophoresis, the gel was stained with silver.
Immunoblot Analysis of Antigen in Hepatoma Cells. The membrane
fraction of PLC/PRF/5 cells was solubilized with SDS at a final
concentration of 10% (w/v). SDS-PAGE of the samples (150 pg) was
performed and proteins in the gel were electrophoretically transferred
to a nitrocellulose filter. The filter was incubated with 5% bovine serum
albumin in PBS at 37°Cfor 2 h and washed with PBS. Then the filter
0)
-0
was incubated with 10 ng/ml of monoclonal antibody 19F12, washed,
and incubated with peroxidase-conjugated anti-mouse IgG at 37°Cfor
I h. Detection of bound antibody was performed by incubation of the
filter with 5 mM 4-chloro-l-naphthol in 20% methanol/PBS solution
containing 0.01% H2O2.
Reactivity of 19F12 Antibody with PLC/PRF/5 Cells Treated with
Various Enzymes. PLC/PRF/5 cells (IO7cells) were treated with 0.25%
(w/v) trypsin (/V-tosyl-L-phenylalanine chloromethyl ketone-treated;
Sigma type X1I1), chymotrypsin (sodium p-tosyl-L-lysine chloromethyl
ketone-treated; Sigma type VII), protease (Sigma type XI), or lipase
(Sigma type VII) for 30 min at 37°Cand washed 3 times with PBS.
The enzyme-treated cells were incubated with 1 ml of 10 Mg/m' of
19F12 antibody for 1 h at 37°C.The mixture was centrifuged for 10
min at 1000 rpm, washed 3 times with PBS, and then incubated with
a, b, c,d
ioo
irji
io2 10°
io1
IO2
Fluorescence
Intensity
Fig. 1. Study of binding activities of monoclonal antibodies to AFP-producing
cells. The cells were incubated with each antibody at 4 ( for 3 h, washed, and
stained with fluorescein isothiocyanate-labeled goat anti-mouse antibody. Stained
cells were analyzed by flow cytometry using a FACS 440. NuE cells (A), KN cells
(B), PLC/PRF/5 cells (C), or PC-9 cells (D) were stained with antibody I9F12
(Curve a), antibody 9D12 (Curve b), antibody 13B1 (Curve c), or PBS (Curve d)
and fluorescein isothiocyanate-labeled anti-mouse antibody.
362
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
DETECTION OF MEMBRANE-BOUND
bodies to hepatoma cells were initially investigated by fluores
cence microscopy.
Immunofluorescence Study Using Fluorescence Microscopy.
Fig. 2, A and E, shows that antibody 19F12 became bound to
the surface of PLC/PRF/5 and NuE cells with a uniform
distribution. Cells were also stained with antibodies 19B1 and
9D12, but the staining showed a speckled distribution in each
case (Fig. 2, B, C, F, and G). This difference in reactivity was
thus likely to be reflected in the results obtained from flow
cytometry analysis. Similar staining with speckled distribution
was shown with normal mouse IgGl and IgG2b (Fig. 2, D and
H) (data for normal mouse IgGl were not shown). Therefore,
it is difficult to state that the bindings of antibodies 9D12 and
19B1 to hepatoma cells were specific. However, with regard to
antibody 19F12, definite reactivity to hepatoma cells was shown
in the immunofluorescence study (Figs. 1 and 2).
The binding activities of these antibodies to hepatoma cells
were then compared.
Binding Activities of Monoclonal Antibodies to Human AFP.
The association constants of these antibodies were determined
by solid-phase immunoassay (Table 1). The binding activity to
AFP of antibody 19F12 was no higher than those of the other
antibodies, 9D12 and 19B1. This finding indicated that the
reactivity of antibody 19F12 with hepatoma cells was not due
to the strength of its affinity to AFP or to the subclass of
antibody, but to its character.
These results suggested that the reactivities of the monoclonal
antibodies with antigen-producing cells varied remarkably ac
cording to differences in the epitopes recognized by the anti
bodies produced using the same antigen. Subsequent studies
were done to investigate the reactivity of antibody 19F12 with
hepatoma cells and the epitope recognized by this antibody.
Competition between Binding of Antibody 19F12 to PLC/PRF/
5 Cells and That to Pure AFP. Competition assay was performed
in order to confirm whether antibody 19F12 definitely recog
nized AFP present on the surface of hepatoma cells. Fig. 3
shows that the amount of antibody binding to PLC/PRF/5
cells was decreased and that the radioactivity in the supernatant
was increased following addition of pure AFP. The decrease
and increase were dependent on the amount of pure AFP added,
indicating that the binding of this antibody to PLC/PRF/5 cells
was competed for by pure AFP.
AFP
Immunostaining of Tissue Sections. Antibody 19F12 showed
positive staining of tissue section of human hepatoma (Fig. 4,
A, D) proved to be AFP positive with AFP-specific antisera
(Fig. 4, B, E). The area stained by antibody 19F12 was the
same as that by polyclonal AFP-specific antisera. This result
also confirmed that antibody 19F12 is specific for human AFP.
Two-Dimensional PAGE of Affinity Purified Antibody 19112Antigen (from Membrane Fraction of PLC/PRF/5 Cells) Com
plex. For determination of the target material of 19F12 anti
body on the hepatoma cell surface, two-dimensional PAGE of
antibody-antigen complex was performed. Antibody 19F12 and
a Triton X-100 extract of the membrane fraction of PLC/PRF/
5 cells were mixed, and this mixture was chromatographed on
protein A-Sepharose CL-4B as described in "Materials and
Methods." The results of two-dimensional PAGE of this anti
body-antigen complex are shown in Fig. 5. Besides the spots
derived from antibody (H and L chain), a spot with a pi of 4.7
and a molecular weight of 65,000 was detected. This spot
corresponded in both pi and molecular weight to AFP. The
same spot pattern was obtained when culture supernatant of
PLC/PRF/5 cells was used as the antigen source.
Western Blot Analysis of Antigen Recognized by 19F12 Anti
body. Western blot analysis of membrane fraction prepared
from PLC/PRF/5 cells with 19F12 antibody revealed an immunoprecipitate band (Fig. 6Ä).The molecular weight of this
protein was to be approximately 65,000 and this was identical
with AFP (Fig. 6, A and B).
Reactivities of 19F12 Antibody with PLC/PRF/5 Cells Treated
with Various Enzymes. In order to obtain information concern
ing the epitope recognized by antibody 19F12, the following
experiments were done. PLC/PRF/5 cells were treated with
trypsin, chymotrypsin, protease, or lipase and the reactivities
of the antibody with these treated cells were measured (Table
2). Antibody reactivities with cells treated with protease and
Table 1 Comparison of the binding activity of monoclonal antibodies to AFP
The association constants of antibodies to AFP were determined by solidphase immunoassay.
Antibody
19F12(IgG2b)
9D12(IgG2b)
19B 1 (IgG 1)
Association con
stant (M~')
3.4 X 10s
7.7 x 10'
6.7 x 10'
Fig. 2. Immunofluorescence
staining of
AFP-producing cells with monoclonal anti
bodies 19F12. 9D12, and 19B1. Cells fixed
with paraformaldehyde were stained with
monoclonal antibodies and FITC-labeled goat
anti-mouse antibody. A-D, PLC/PRF/5 cells;
E-H, NuE cells. A and E, stained with antibody
19F12; B and F, stained with antibody 9D12;
C and G, stained with antibody 19B1; D and
//, stained with normal mouse IgG2b and fluorescein isothiocyanate-labeled anti-mouse an
tibody.
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
DETECTION OF MEMBRANE-BOUND
trypsin were decreased compared with that for untreated control
cells, but reactivity for cells treated with lipase was increased.
The binding of the antibody to AFP was not inhibited by Con
A. Con A solution (50 ^1 at 0-500 Mg/ml) was added to each
well of an AFP-coated microassay plate and then reacted with
50 M' of 10 Mg/ml antibody 19F12. The amounts of antibody
which bound to AFP treated or untreated with various concenterations of Con A were equal (data not shown).
From these results, it seems that the epitope recognized by
antibody 19F12 is part of the peptide area of AFP. Also, the
increase in the binding activity for cells treated with lipase
suggests the presence of AFP buried under the cell membrane.
2000
E
O.
I
>>
•¿â€¢-1000
o
T3
O
10
Concentration
20
of AFP - jjg
Fig. 3. Competition between reactivity of 19F12 antibody to PLC/PRF/5
cells and that to pure AFP. '4C-Labeled 19F12 and various concentrations of
pure AFP were added to 10*cells. After incubation for 12 h at 4°C,radioactivities
of precipitate (•)and supernatant (O) were counted.
AFP
DISCUSSION
AFP is recognized to be a secreted protein (4, 21). However,
the presence of AFP at the cell surface of AFP-producing rat
hepatoma cells has been demonstrated previously using immunofluorescence staining (6, 22), and antibodies to AFP conju
gated with antitumor drugs or drug-entrapped liposome-conjugated antibodies to AFP show specific cytotoxicity (13, 15).
The presence of membrane-bound AFP was confirmed in the
present study. Monoclonal antibody 19F12 to AFP became
bound to both AFP-producing hepatoma cells and immortal
ized fetal liver cells (Figs. 1 and 2), pure AFP competed the
binding of this antibody to hepatoma cells (Fig. 3), and it was
demonstrated that the material recognized by antibody 19F12
in the membrane fraction of PLC/PRF/5 hepatoma cells could
be extracted by treatment with Triton X-100 and that this was
indistinguishable from AFP on the basis of molecular weight
and pi (Figs. 5 and 6). Furthermore, we have examined the
AFP production of the cells used for the study of the antibody
binding to AFP both in vitro and in vivo (data not shown). The
amounts of AFP in sera from tumor-bearing nude mice, extracts
of membrane fractions, and cytoplasmic fractions of cultured
cells and those secreted into culture supernatants were meas
ured. It was found that PLC/PRF/5 cells produced AFP in
large amounts, being detected not only as secreted AFP but
also as a membrane-associated fraction. On the other hand, in
KN and NuE cells, AFP was detected mainly in the membrane
fractions, hardly any secreted AFP being apparent. Although it
is recognized that AFP is generally destined for secretion, a
type of cell from which it is hard for AFP to be secreted may
exist. It seems that AFP produced by these cells stays in the
membrane, and the patterns obtained using flow cytometry
reflect this difference in the nature of these cells (Fig. 1, A-C).
Monoclonal antibodies to AFP have various reactivities with
antigen-producing cells. Antibody 19F12 in particular binds to
Fig. 4. Immunostaining of human hepa
toma tissues. Deparaffinized human hepatoma
tissue section fixed in formalin was stained
with anti-AFP monoclonal antibody 19F12 (A
and I>) or anti-AFP rabbit antisera (It and / I.
Tissue section was not stained with only sec
ond antibody (fluorescein isothiocyanate-labeled anti-mouse antibody) (C and F).
364
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
DETECTION OF MEMBRANE-BOUND
MW 4
Fig. 5. Two-dimensional PAGE of anti
body 19F12-antigen (from PLC/PRF/5 cells)
complex. Triton X-100 extracts from the
membrane fraction of PLC/PRF/5 cells (5 x
10' cells) were diluted with 9 volumes of PBS
and mixed with 2 jig of antibody 19F12. An
tigen-antibody complex was purified on pro
tein A-Sepharose CL-4B and electrophoresed
two-dimensionally. Arrow, spot that corre
sponds to AFP in both pi and molecular weight
(MW). K, thousands.
AFP
ISOELECTRIC
PH
7
65K -
such cells with a uniform distribution (Fig. 2, A and E). The
other antibodies, 9D12 and 19B1, bind to these cells with a
speckled distrtibution (Fig. 2, A, C, F, and G). It seems that
94these varying reactivities are due to differences in the epitopes
recognized by these antibodies. Tsukazaki et al. (14) showed
67that ricin toxin A chain-conjugated anti-AFP monoclonal an
tibody has no specific cytotoxicity to AFP-producing hepatoma
43cells, while the same toxin-conjugated monoclonal antibody to
placenta! alkaline phosphatase has specific cytotoxicity to this
antigen-producing cancer cell. This difference in the binding
30profile of these antibodies is similar to that observed in the
distribution on the cell surface of target cells in the present
immunofluorescence study. The former antibody binds to target
20cells in a speckled distribution, similar to the results obtained
using antibodies 9D12 and 19B1, the latter binding to target
cells with a uniform distribution. According to our unpublished
results, ricin toxin A chain-conjugated antibody 19F12 is 10
times more cytotoxic to AFP-producing human hepatoma cells
Fig. 6. Western blot. SDS-PAGE of the membrane fraction of PLC/PRF/5
cells (150 pg) (B and O and the same sample (150 #ig) plus human AFP from
(PLC/PRF/5 cells) than ricin toxin A chain or this antibody
placenta (1 ^g) (A) was performed, and proteins in the gel were electrophoretically
only. We therefore consider that successful cell killing by drug
transferred to a nitrocellulose filter (A and B). The filter was blotted with 19F12
or toxin conjugated to antibody depends on the amount of
antibody (10 ng/ml) immunochemically. Total proteins of the membrane fraction
of PLC/PRF/5 cells were stained with Coomassie blue (C). One protein band
antibody bound to cells. As is evident from analysis using flow
was stained by 19F12 antibody (B) and the molecular weight (MW) of this was
cytometry,
the amount of antibody 19F12 binding per single
identical to that of human AFP (A and B). k, thousands.
cell is high. Bovine AFP was present in the culture medium,
but it has been demonstrated that these monoclonal antibodies
were not reactive with bovine AFP by sandwich immunoassay
Table 2 Reactivities ofl9F12 antibody to PLC/PRF/5 cells treated with various
and consequently it seems that antibody 19F12 recognizes
enzymes
human AFP derived from cells rather than being absorbed from
PLC/PRF/5 cells treated with various enzymes were incubated with antibody
19F12 and then with horseradish peroxidase-conjugated goat anti-mouse anti
the medium. Furthermore, the specific binding of antibody
body. The amount of antibody bound to cells is shown as bound horseradish
19F12 to human AFP was confirmed by positive immunostainperoxidase activity.
ing to tissue sections of human hepatomas proved to be AFP
of bound
positive with AFP-specific antiserum (Fig. 4). Therefore, anti
Treatment of
peroxidase
binding
cellsUntreated
(¿MO™)0.329
(%)100
body 19F12 is considered to be a good one for the targeting of
±0.051°
drugs or toxin to antigen-producing cells and also for the
Trypsin
0.092 ±0.051
27.9
detection of tumors. Antibody 19F12-conjugated liposome con
Protease
0.1 11 ±0.038
33.7
taining Adriamycin has also been shown to have specific cyto
Chymotrypsin
0.224 ±0.069
68.1
toxicity for antigen-producing tumors both in vivo and in vitro
LipaseActivity
0.638 ±0.077Antibody
197
°Mean ±SD.
(15), but we and another collaborator (K. Nakamura, Keio
365
MW
-k
A
B
C
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
DETECTION OF MEMBRANE-BOUND
University, Japan) have obtained satisfactory imaging data
using I25l-labeled 19F12 and tumor (NuE cells)-bearing nude
mice.3
As for the epitope recognized by antibody 19F12, this may
be part of the peptide exposed above the cell surface, as sug
gested from the results shown in Table 2 and the Con A
inhibition assay. AFP derived from human hepatoma cells has
one Con A-specific asparagine-linked sugar chain and its sugar
content is about 4% (23, 24). Therefore it seems unlikely that
this epitope exists on the sugar chain.
ACKNOWLEDGMENTS
We wish to thank Mihoko Nakajima for her expert technical assis
tance.
REFERENCES
1. Köhler,G., and Milstein, C. Continuous cultures of fused cells secreting
antibody of predefined specificity. Nature (Lond.), 256:495-497, 1975.
2. Alpert, E. Human a-fetoprotein (AFP): developmental biology and clinical
significance. In: Progress in Liver Diseases, pp. 337-349. New York: Gruñe
and Stratton, 1976.
3. Ruoslahti. E., I mila. M., and Engvall, E. Radioimmunoassay of o-fetoprotein with polyclonal and monoclonal antibodies. Methods Enzymol., 84: 319, 1982.
4. Leffert, H. L., and Sell, S. a-Fetoprotein biosynthesis during the growth cycle
of differentiated fetal rat hepatocytes in primary monolayer culture. J. Cell
Biol., 61:823-829, 1974.
5. Mizejewski, G. J., Young, S. R., and Allen, R. P. a-Fetoprotein: effect of
heterologous antiserum on hepatoma cells in vitro. 3. Nati. Cancer Inst., 54:
1361-1367, 1975.
6. Tsukada, Y., Mikuni. M., Watabe, H., Nishi, S., and Mirai, H. Effect of antitt-fetoprotein serum on some cultured tumor cells. Int. J. Cancer, 13: 187195, 1974.
7. Mizejewski, G. J., and Allen, R. P. Immunotherapeutic suppression in
transplantable solid tumors. Nature (Lond.), 250: 50-52, 1974.
8. Wepsic, H. T., Tsukada, Y., Takeichi, N., Nishi, S., and Hirai, H. Effect of
horse antibody to rat a-fetoprotein upon the growth of AH-66 in Donryu
rats. Int. J. Cancer. 25: 655-661, 1980.
9. Kim, E. E., DeLand. F. H., Nelson, M. O., Benne«,S., Simmons, G., Alpert,
E., and Goldenberg, D. M. Radioimmunodetection of cancer with radiola
beled antibodies to a-fetoprotein. Cancer Res., 40: 3008-3012, 1980.
' Unpublished data.
AFP
10. Koji, T., Ishii, N., Munehisa, T., Kusumoto, Y., Nakamura, S., Tamenishi,
A., Hará,A., Kobayashi, K., Tsukada, Y., Nishi, S., and Hirai, H. Localiza
tion of radioiodinated antibody to a-fetoprotein in hepatoma transplanted in
rats and a case report of a-fetoprotein antibody treatment of a hepatoma
patient. Cancer Res., 40: 3013-3015, 1980.
11. Kono, K., Nakata, K., Muro, T., Furukawa, R., Ogino, T., Kusumoto, Y.,
Ishii, N., Munehisa, T., Koji, T., Nagataki, S., Tsukada, Y., and Hirai, H.
Localizing properties of antibody to a-fetoprotein (AFP) to AFP-producing
tumor cells. Oncodev. Biol. Med., 4: c87-c94, 1983.
12. Sakahara, H., Endo, K., Nakashima, T., Koizumi, M., Ohta, H., Torizuka,
K.,Okada,K., Yoshida,<>., and Nishi,S. Radioimmunoimagingusing I (ah'h
fragment of monoclonal antibodies against human a-fetoprotein. Kakuigaku
(Jpn. J. NucÃ-.Med.), 22: 377-382, 1985.
13. Tsukada, Y., Bischof, W. K.-D., Hibi, N., Hirai, H., Hurwitz, E., and Sela,
M. Effect of a conjugate of daunomycin and antibodies to rat a-fetoprotein
on the growth of a-fetoprotein-producing tumor cells. Proc. Nati. Acad. Sci.
USA, 79:621-625, 1982.
14. Tsukazaki, K., Hayman, E. G., and Ruoslahti, E. Effects of ricin A chain
conjugates of monoclonal antibodies to human a-fetoprotein and placenta!
alkaline phosphatase on antigen-producing tumor cells in culture. Cancer
Res., 45:1834-1838, 1985.
15. Konno, H., Suzuki, H., Tadakuma, T., Kumai, K., Yasuda, T., Kubota, T.,
Ohta, S., Nagaike, K., Hosokawa, S., Ishibiki, K., Abe, O., and Saito, K.
Antitumor effect of Adriamycin entrapped in liposomes conjugated with antihuman a-fetoprotein monoclonal antibody. Cancer Res., 47: 4471-4477,
1987.
16. Alexander, J. J., Bey, E. M., Geddes, E. W., and Lecatsas, G. Establishment
of a continuously growing cell line from primary carcinoma of the liver.
South Afr. Med. J., 50: 2124-2128, 1976.
17. Kinjo, M., Oka, K., Naito, S., Kohga, S., Tanaka, K., Oboshi, S., Hayata,
Y., and Yasumoto, K. Thromboplastic and fibrinolytic activities of cultured
human cancer cell lines. Br. J. Cancer, 39: 15-23, 1979.
18. Ey, P. L., Prowse, S. J., and Jenkin, C. R. Isolation of pure IgGl, IgG2a and
IgG2b immunoglobulins from mouse serum using Protein A-Sepharose.
Immunochemistry, 15: 429-436, 1978.
19. Tsukada, Y., Hurwitz, E., Kashi, R., Sela, M., Hibi, N., Hará,A., and Hirai,
H. Chemotherapy by intravenous administration of conjugates of daunomy
cin with monoclonal and conventional anti-rat a-fetoprotein antibodies. Proc.
Nati. Acad. Sci. USA, 79:7896-7899, 1982.
20. O'Farrell, P. H. High resolution two-dimensional electrophoresis of proteins.
J. Biol. Chem., 250: 4007-4021, 1975.
21. Law, S. W., and Dugaiczyk, A. Homology between the primary structure of
a-fetoprotein, deduced from a complete cDNA sequence, and serum albumin.
Nature (Lond.), 297: 201-205, 1981.
22. Tsukada, Y., and Hirai, H. a-Fetoprotein and albumin synthesis during the
cell cycle. Ann. NY Acad. Sci., 259: 37-44, 1975.
23. Morinaga, T., Sakai, M., Wegmann, T. G., and Tamaoki, T. Primary
structures of human a-fetoprotein and its mRNA. Proc. Nati. Acad. Sci.
USA, 80:4604-4608, 1983.
24. Yoshima, H., Mizuochi, T., Ishii, M., and Kobata, A. Structure of the
asparagine-linked sugar chains of a-fetoprotein purified from human ascites
fluid. Cancer Res., 40:4276-4281, 1980.
366
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
Detection of Membrane-bound α-Fetoprotein in Human
Hepatoma Cell Lines by Monoclonal Antibody 19F12
Saiko Hosokawa, Minoru Muramatsu and Kazuhiro Nagaike
Cancer Res 1989;49:361-366.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/49/2/361
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.