Available via license: CC BY 4.0
Content may be subject to copyright.
THE
JOURNAL
OF
BIOLOGICAL
CHEMISTRY
Q
1985
by
The
American
Society
of
Biological
Chemists,
Inc.
Vol.
260,
No.
22,
Issue
of
October
5,
pp.
12035-12041,1985
Printed
in
U.S.A.
Purified Immunotoxins That Are Reactive with Human Lymphoid
Cells
MONOCLONAL ANTIBODIES CONJUGATED TO THE RIBOSOME-INACTIVATING PROTEINS GELONIN
AND THE POKEWEED ANTIVIRAL PROTEINS*
(Received for publication, February 25, 1985)
John
M.
Lambert, Peter
D.
Senter, Annie Yau-Young$, Walter A. Blattler, and
Victor
S.
Goldmacher
From the Dana-Farbet Cancer Institute, Division
of
Tumor
Immumlogy,
Boston, Massachusetts
02115
Seven different monoclonal antibodies of the IgG
class that are reactive with four different antigens on
human lymphoid cells were utilized to form immuno-
toxins with the ribosome-inactivating proteins gelonin
and the three known pokeweed antiviral proteins.
Thirteen different immunotoxin combinations were
prepared. The ribosome-inactivating proteins were
modified with 2-iminothiolane. The sulfhydryl groups
so
introduced were reacted with maleimido groups or
with dithiopyridyl groups that had been introduced
into the antibodies. The toxin-antibody conjugates
so
formed were purified by affinity chromatography on
protein A-Sepharose CL-4B, ion exchange chromatog-
raphy, and by
gel
filtration and were characterized by
polyacrylamide-dodecyl sulfate gel electrophoresis.
The purified immunotoxins were free of nonconjugated
monomeric proteins and aggregates of very high mo-
lecular weight. All the immunotoxins~showed the
spe-
cific
binding of the component antibody as measured
by indirect immunofluorescence binding assays. The
activities of the ribosome-inactivating proteins were
unaffected by conjugation where the cross-link to the
antibody contained a disulfide bond and when assayed
after reductive cleavage of the linker. Disulfide-linked
immunotoxins with six of the antibodies were highly
cytotoxic for the target cells. However, immunotoxins
containing an anti-B1 antibody showed no cytotoxic-
ity.
The possible use of antibodies to target pharmacologic
agents, such
as
toxins, was first proposed by Ehrlich
(1).
Research to exploit this idea has developed rapidly in the last
decade, owing much to the ability to produce pure highly
specific monoclonal antibodies using the hybridoma technol-
ogy
(2).
Recently, monoclonal antibodies have been developed
that recognize tumor-associated antigens
(3,
4,
11, 12), and it
is the hope that such antibodies can be exploited to deliver
toxic agents to particular types of tumor cells in order to kill
them selectively. The ribosome-inactivating proteins
(5,
6)
seem to be ideal toxic agents for this purpose. Most effort has
been directed toward using ricin (extracted from castor beans,
Ricinus communis)
which consists of two nonidentical sub-
*
This work was supported by a grant from ImmunoGen Inc. 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.
$Present address: Liposome Technology Inc., 1050 Hamilton
Court, Menlo Park,
CA
94025.
units (A- and B-chains) that are joined
by
a disulfide bond
(6).
The B-chain has the property of binding to cell-surface
carbohydrates and promotes the uptake
of
the A-chain into
cells
(6).
Entry of the A-chain into the cytoplasm of a cell
then results in the death of the cell by catalytic inactivation
of its ribosomes. Immunotoxins have been made by conjugat-
ing intact ricin to antibodies (7-9). Such immunotoxins ex-
hibit
specific
toxicity only in the presence of lactose which at
high concentration competes with the cell surface carbohy-
drates for the ricin B-chain binding site(s).
In uiuo,
these
immunotoxins are expected to be nonspecifically toxic, as is
ricin itself, and are, therefore, unlikely to be of therapeutic
value, although they may have limited use in the
in
uitro
treatment of bone marrow for transplantation
(9,lO).
There is a class
of
ribosome-inactivating proteins that have
properties and characteristics similar to those of ricin A-chain
alone
(5).
Gelonin
(20)
and the three known pokeweed anti-
viral proteins
(21)
are examples of such proteins. They are
basic proteins, of
M,
about 30,000
(5).
These proteins have
several advantages over ricin A-chain in the preparation of
immunotoxins: they are extremely stable proteins, they do
not bind to cells and
so
are nontoxic to intact cells (except at
very high concentrations), and they are safe to purify and
manipulate in the laboratory without the extreme precautions
necessary
for
work with ricin
(5).
These proteins are good
candidates for the preparation of immunotoxins since, at least
in principle, such immunotoxins will only bind to the cells
selected by the antibody. Immunotoxins have been made using
gelonin and PAP,l and in general they showed specific cyto-
toxicity similar
to
immunotoxins prepared with ricin A-chain
There
is
an important caveat in the interpretation of the
experiments that have been reported using immunotoxins
made with ricin A-chain, gelonin,
or
PAP.
There is not a
single example of an immunotoxin that was completely puri-
fied from nonconjugated antibody. This is an important ob-
stacle to the proper interpretation of these reports and ham-
pers understanding of the mechanisms involved in the cyto-
toxicity of such immunotoxins. We describe here the prepa-
ration of highly purified immunotoxin conjugates using seven
different monoclonal antibodies and using the ribosome-in-
The abbreviations used are:
PAP,
pokeweed antiviral protein;
PAP 11, pokeweed antiviral protein type
II;
PAP-S,
pokeweed anti-
viral protein from seeds; SPDP, n-succinimidyl 3-(2-pyridyldi-
thi0)propionate; SMCC, succinimidyl
4-(N-maleimidomethyl)cyclo-
hexane-1-carboxylate; bis-tris,
2-[bis(2-hydroxyethyl)amino-2-(hy-
droxymethyl)-propane-1,3-diol;
HEPES,
4-(2-hydroxyethyl)-l-piper-
azineethanesulfonic acid; CALLA, common acute lymphoblastic leu-
kemia antigen; RPMI, Roswell Park Memorial Institute.
(13-19,22-25).
12035
12036
Purified Immunotoxins
activating proteins gelonin, PAP, PAP
11,
and PAP-S. The
principle of the purification methods should be generally
applicable. The immunotoxins reported here were prepared
both with a cleavable linker containing a disulfide bond and
with a noncleavable linker, in order to evaluate the impor-
tance of release of the toxic agent from the antibody. All the
immunotoxins were analyzed for their ribosome-inactivating
capacity, their ability to bind to cell-surface antigens, and
their
in
vitro
cytotoxic potency toward lymphoid cell lines.
EXPERIMENTAL PROCEDURES
AND
RESULTS*
DISCUSSION
We have developed methods for the preparation of purified
immunotoxin conjugates in high yield, utilizing the techniques
of
affinity chromatography, gel filtration, and ion exchange
chromatography with buffers of carefully defined composi-
tion. Thirteen different immunotoxin conjugates were made
using seven different monoclonal antibodies and four different
ribosome-inactivating proteins. The immunotoxin prepara-
tions contain no nonconjugated antibody, no nonconjugated
ribosome-inactivating proteins, and no aggregates of very high
molecular weight. These purified and well defined immuno-
toxins made it possible for the first time to perform quanti-
tative binding studies and to perform cytotoxicity tests with-
out fear of the effect
of
competition by nonconjugated anti-
body on the experimental result, for example, by blocking
antigens
or
saturating the internalization pathways. Also, the
purity of the immunotoxin preparations permitted a careful
comparison to be made of the biological activities of the
component proteins with their nonconjugated counterparts.
Gelonin and the pokeweed antiviral proteins were not af-
fected by modification with 2-iminothiolane in their ability
to inhibit protein synthesis. The ability of the modified toxins
to inactivate ribosomes in a cell-free system of protein syn-
thesis was indistinguishable from that of the native proteins.
This is in contrast with the results of earlier work using
N-
succinimidyl3-(2-pyridyldithio)propionate
to modify gelonin
(22); the gelonin was inactivated by about 90%.
It
is possible
that 2-iminothiolane and N-succinimidyl 3-(2-pyridyldi-
thio)propionate each react preferentially with different amino
groups of the gelonin molecule. Another factor is the preser-
vation of the positive charge at amino groups upon reaction
with 2-iminothiolane that may account
for
this difference.
Our results suggest that 2-iminothiolane is the reagent of
choice for modifying ribosome-inactivating proteins, in order
to introduce sulfhydryl groups while preserving their toxic
activity.
When the modified toxins were covalently linked to an
antibody, the ribosome-inactivating activity of the toxins was
reduced by about
70%.
When the covalent link between the
toxin and the antibody included a disulfide bond, then the
full ribosome-inactivating activity as measured in a cell-free
system could be restored by reductive cleavage of the linker.
This may account, in part, for the observation that the im-
munotoxins J5-gelonin and J5-PAP-S exhibited greater cy-
totoxicity (about 103-fold) on cultured CALLA-positive cell
Portions of this paper (including “Experimental Procedures,”
“Results,” Figs.
1-5,
and Table
1)
are presented in miniprint at the
end of this paper. Miniprint is easily read with the aid of a standard
magnifying glass. Full size photocopies are available from the Journal
of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814.
Request Document
No.
85M-546, cite the authors, and include a
check
or
money order for $9.20 per set
of
photocopies.
Full
size
photocopies are also included
in
the microfilm edition of the Journal
that
is
available from Waverly Press.
lines when the linker included a disulfide bond than when the
linker was noncleavable. There
is
one claim in the literature
that immunotoxins containing pokeweed antiviral proteins
are more cytotoxic when linked by a noncleavable linker than
when linked by a disulfide linker (25). However, we could not
confirm this result, and our observations with purified im-
munotoxins were consistent with previous findings with im-
munotoxins containing ricin A-chain made with cleavable and
noncleavable linkers (17).
A
cleavable linkage between the
toxin and the antibody presumably allows the toxin to escape
more easily into the cytoplasm from the membrane-bound
antibody/antigen complex
(48).
Binding studies showed that the immunotoxins made with
all seven antibodies showed specific binding to antigen-bear-
ing cells. Cytotoxicity experiments
in
vitro
showed that im-
munotoxins made with six of the seven antibodies showed
considerable enhancement
of
the toxic effect of the ribosome-
inactivating proteins, with ID5,, concentrations that were
about 104-fold lower than that of the native ribosome-inacti-
vating protein. The enhanced toxicity was completely specific
for cells bearing the cognate antigen. Immunotoxins made
with gelonin
or
with the pokeweed antiviral proteins gave
qualitatively similar results. However, none of the immuno-
toxins made with anti-El showed any more toxicity than that
exhibited by the native ribosome-inactivating protein, even
at concentrations that were near the saturation of antibody/
antigen binding. This result was obtained on several E1
antigen-positive cell lines, including some that were also
positive for CALLA and
for
Ia antigens, and which were
sensitive to immunotoxins made with the antibodies 55 or
I-
2,
respectively.
Receptor-mediated endocytosis of an antigen/immunotoxin
complex may be essential for cytotoxicity. This has been
suggested on the basis of experiments with lysosomotropic
agents which raise the pH of intracellular acidic vesicles (51)
and which increase the potency of immunotoxins (52, 53).
There is some evidence to suggest that the
T11
surface antigen
(29) and the CALLA
(49)
can be internalized, and perhaps
these antigens utilize the mechanisms involving coated pits,
coated vesicles, and endosomes
(48).
Our preliminary results
suggest that the Ia antigens can also be internalized carrying
the 1-2 antibody. However, the current evidence suggests that
the B1 antigen (50) remains firmly on the cell surface and
shows no tendency to be internalized? Thus, the
I31
antigen
may be excluded from the coated pits involved in receptor-
mediated endocytosis
(48).
This suggests
a
possible explana-
tion for the lack of cytotoxicity of immunotoxins made with
anti-B1; that is, that the B1 antigen does not transport the
immunotoxin complex inside the cell. Further evidence will
be required to assess the relationship between endocytosis
and the cytotoxicity
of
the immunotoxins made with ribo-
some-inactivating proteins. However, it is clear from our
present results that the properties of the target antigen require
careful consideration when designing immunotoxin conju-
gates.
The availability of pure immunotoxin conjugates will allow
us to examine in greater detail than hitherto the effect of
various parameters on the cytotoxicity exhibited by the im-
munotoxins. For example, the antibody-binding affinity of
the immunotoxins, the number of cell-surface antigens capa-
ble of binding the immunotoxins, and the properties of differ-
ent antigens once bound by an immunotoxin such as in
receptor-mediated endocytosis may all influence the degree of
cytotoxicity of the immunotoxins. We are now employing
L.
M.
Nadler, personal communication.
Purified
Im
several different
in
vitro
cytotoxicity assays, including those
that directly measure cell survival, to investigate these param-
eters and to study potentiators of cytotoxicity such as ade-
novirus
(54)
and lysosomotropic agents
(53).
These experi-
ments may help to determine how to improve the efficacy of
the immunotoxins. Also, little is known about the biological
properties of immunotoxins
in
uiuo.
Purified conjugates using
different anti-Tll antibodies, three of which are described
here, may be useful biological reagents for this purpose since
the antibodies react with a surface antigen found on T cells
of various species of monkey that is analogous to the human
T11 surface antigen
(34).
Acknowledgments-We
wish to thank Christina Doyle, Susan Bro-
deur, Nancy Tinnel, and Jean Anderson for skilled technical work
and Diana Sam for excellent typing.
REFERENCES
1.
Ehrlich,
P.
(1956)
in
Collected Papers
of
Paul Ehrlich
(Himmel-
2.
Kohler, G., and Milstein, C.
(1975)
Nature
256, 495-496
weit,
F.,
ed) Vol.
2,
pp.
442-447,
Pergamon Press, London
3.
Ritz, J., Pesando,
J.
M., Notis-McConarty, J., Lazarus, H., and
4.
Seon, B. K., Negoro,
S.,
and Barcos, M.
P.
(1983)
Proc. Natl.
5.
Barbieri,
L.,
and Stirpe,
F.
(1982)
Cancer
Surveys
1,489-520
6.
Olsnes,
S.,
and Pihl, A.
(1982)
in
Molecular Action
of
Toxins and
Viruses
(Cohen, P., and Van Heyningen,
S.,
eds) pp.
51-105,
Elsevier, Amsterdam
7.
Youle, R.
J.,
and Neville, D. M., Jr.
(1980)
Proc. Natl. Acad. Sci.
Schlossman,
S.
F.
(1980)
Nature
283, 583-585
Acad. Sci.
U.
S.
A.
80,845-849
U.
S.
A.
77,5483-5486
8.
Thoroe.
P.
E.,
Mason. D. W., Brown. A. N.
F.,
Simmonds,
S.
J.,
his,'
W.
C.
J., Cumber,
A.
J., and Forrester,
J.
A.
(1982)
Nature
297,594-596
9.
Vallera, D. A., Ash, R. C., Zanjani, E. D., Kersey,
J.
H. LeBien,
T.
W., Beverley,
P.
C.
L.,
Neville,
D.
M., and Youle, R.
J.
(1983)
Science
222, 512-515
10.
Quinones, R.
R.,
Youle, R.
J.,
Kersey,
J.
H.
Zanjani, E.
D.,
Azemove,
S.
M., Soderling, C. C. B., LeBien,
T.
W., Beverley,
P.
C. L., Neville, D. M., Jr., and Vallera,
D.
A.
(1984)
J.
Zmmunol.
132,678-683
11.
Woodbury, R. G., Brown,
J.
P., Yeh, M.-Y., Hellstrom,
I.,
and
Hellstrom, K. E.
(1980)
Proc. Natl. Acad. Sci.
U.
S.
A.
77,
12.
Herlyn, M., Steplewski,
Z.,
Herlyn, D., and Koprowski, H.
(1979)
Proc. Natl.
Acad.
Sci.
U.
S.
A.
76, 1438-1442
13.
Krolick,
K.
A., Villemez,
C.,
Isakson, P., Uhr,
J.
W.,
and Vitetta,
'E.
S.
(1980)
Proc.
Natl. Acad. Sci.
U.
S.
A.
77, 5419-5423
14.
Youle, R. J., and Neville,
D.
M., Jr.
(1982)
J.
Biol. Chem.
257,
15.
Raso, V., Ritz, J., Basala, M., and Schlossman,
S.
F.
(1982)
16.
Thorpe,
P.
E., and Ross, W. C.
J.
(1982)
Zmmunol.
Rev.
62,119-
17.
Masuho, Y., Kishida,
K.,
Saito, M., Umemoto, N., and Hara,
T.
18.
Vitetta,
E.
S.,
Krolick, K. A., Miyama-Inaba, M., Cushley, W.,
19.
Vitetta, E.
S.,
Cushley, W., and Uhr,
J.
W.
(1983)
Proc. Natl.
20.
Stirpe, F., Olsnes,
S.,
and Pihl, A.
(1980)
J.
Biol. Chem.
255,
21.
Irvin,
J.
D.
(1983)
Pharmacol. Ther.
21,371-387
2183-2186
1598-1601
Cancer
Res.
42,457-464
158
(1982)
J.
Biochem.
(Tokyo)
91, 1583-1591
and Uhr,
J.
W.
(1983)
Science
219,644-650
Acad. Sci.
U.
S.
A.
80,
6332-6335
6947-6953
:munotorins
12037
22.
Thorpe,
P.
E., Brown, A.
N.
F., Ross, W. C. J., Cumber, A. J.,
Detre,
S.
I.,
Edwards, D. C., Davies, A.
J.
S.,
and Stirpe,
F.
(1981)
Eur.
J.
Biochem.
116,447-454
23.
Wiels, J., Junqua,
S.,
Dujardin, P., LePecq, J.-B., and Tursz,
T.
(1984)
Cancer
Res.
44,129-133
24.
Colombatti, M., Nabholz, M., Gross,
O.,
and Bron, C.
(1983)
J.
Zmmunol.
131,3091-3095
25. Ramakrishnan,
S.,
and Houston, L. L.
(1984)
Cancer
Res.
44,
26.
Hercend,
T.,
Nadler, L. M., Pesando,
J.
M., Reinherz, E.
L.,
Schlossman,
S.
F., and Ritz,
J.
(1981)
Cell.
Zmmunol.
64, 192-
199
27.
Stashenko, P., Nadler, L. M., Hardy, R., and Schlossman,
S.
F.
(1980)
J.
Zmmunol.
125, 1678-1685
28.
Nadler,
L.
M., Stashenko, P., Hardy, R., Pesando,
J.
M., Yunis,
E.
J.,
and Schlossman,
S.
F.
(1981)
Hum. Zmmunol.
1,77-90
29.
Meuer,
S.
C., Hussey, R. E., Fabbi, M., Fox, D., Acuto,
O.,
Fitzgerald, K. A., Hodgdon,
J.
C., Protentis,
J.
P., Schlossman,
S.
F.,
and Reinherz, E. L.
(1984)
Cell
36,897-906
201-208
30.
Goding,
J.
W.
(1976)
J.
Immunol. Methods
13,215-226
31.
Cowan, N. T.,, Secher,
D.
S.,
Cotton, R. G. H., and Milstein, C.
32.
Barbieri,
L.,
Aron, G. M., Irvin,
J.
D., and Stirpe, F.
(1982)
33.
Deleted in proof
34.
Letvin, N. L., King, N. W., Reinherz, E.
L.,
Hunt, R. D.,
347
H., and Schlossman,
S.
F.
(1983)
Eur.
J.
Zmmunol.
13,
345-
35.
Benjamin, D., Magrath,
I.
T.,
Maguire, R., Janus, C., Todd,
H.
D., and Parsons, R. G.
(1982)
J.
Zmmunol.
129,1336-1342
36.
Falk, L., Wright,
J.,
Wolfe, L., and Deinhardt, F.
(1974).
Int.
J.
Cancer
14,244-251
37.
'Hurwitz,
R.,
Hozier,
J.,
LeBien,
T.,
Gajl-Peczalska,
K.,
Kuboni-
shi, I., and Kersey,
J.
(1979)
Znt.
J.
Cancer
23, 174-180
38.
Lazarus, H., Barell,
E.
F., Oppenheim,
S.,
and Krishnan, A.
(1974)
Zn
Vitro
9, 303-310
39.
Collins.
S.
J..
Gallo, R. C., and Gallagher,
R.
E.
(1977)
Nature
(1973)
FEBS Lett.
30,343-346
Biochem.
J.
203, 55-59
270,'347-349
.
"
40.
Carlsson, J., Drevin, H., and Axen,
R.
(1978)
Biochem.
J.
173,
723-737
41.
Lambert,
J.
M.,
Jue, R., and Traut, R.
R.
(1978)
Biochemistry
42.
Ellman, G. L.
(1959)
Arch. Biochem. Biophys.
82,
70-77
43.
Houston, L.
L.,
Ramakrishnan,
S.,
and Hermodson, M. A.
(1983)
44.
Laemmli, U. K.
(1970)
Nature
227,680-685
45.
Ritz,
J.,
Pesando,
J.
M.; Notis-McConarty, J., and Schlossman,
46.
Pelham, H.
R.
B., and Jackson, R.
J.
(1976)
Eur.
J.
Bwchem.
67,
47.
Yoshitake,
S.,
Yamada, Y., Ishikawa, E., and Masseyeff, R.
(1979)
48.
de Duve, C.
(1983)
Eur.
J.
Biochem.
137, 391-397
49.
Pesando,
J.
M., Ritz, J., Lazarus, H., Tomaselli, K. J., and
50.
Oettgen,
H.
C., Bayard,
P.
J.,
van Ewijk, W., Nadler, L. M., and
51.
Ohkuma,
S.,
and Poole, B.
(1978)
Proc. Natl. Acad. Sci.
U.
S.
A.
52. Myers, C. D., Thorpe,
P.
E., Ross, W. C.
J.,
Cumber, A.
J.,
Katz,
F.
E.,
Tax, W., and Greaves, M. F.
(1984)
Blood
63,1178-1185
53.
Casellas, P., Bourrie, B.
J.
P., Gros, P., and Jansen, F. K.
(1984)
J.
Bwl. Chem.
259,9359-9364
54.
Fitzgerald,
D.
J.
P.,
Trowbridge,
1.
S.,
Pastan, I., and Willingham,
M.
C.
(1983)
Proc. Natl. Acad. Sci.
U.
S.
A.
80,4134-4138
17,5406-5416
J.
Biol.
Chem.
258,960-9604
S.
F.
(1980)
J.
Immunol.
125,1506-1514
247-256
Eur.
J.
Biochem.
101,395-399
Schlossman,
S.
F.
(1981)
J.
Immunol.
126, 540-544
Terhorst,
C.
P.
(1983)
Hybridoma
2,17-28
75,3327-3331
Continued on next page.
12038
Purified
Immunotoxins
SUPPLEHeNT*RY
KATERW
TO
PURlFlED ~OTOXlNS
TbXT
*Be
REACTIVE
WITH
RlBOSOHe-INACTIVATING PROTEINS GELONIN
AM)
THE
POyEueED
ANTIVIRAL
PROTEINS
John
K.
Lambert, Peter
0.
Seater,
Annie
Yau-¶aung,
oonoelooal antibodies
were
kindly provided by-Drs.
J.
Bit=
(J5,
J30),
L.K. Nidler (anti-al,
1-21
and
E.L. Reinhers (anti-T11 antibodies) of the Dana-Farber Cancer Institute, Division
of
Tumor
Ilmunology.
Purification of lloooclooal Antibodies
-All
steps
wexe
done
at 40
C.
All
of
the
antibodies.
except
for anti-Tllu
(eo
lgcl),
were
purified
fra
ascites
fluid by affinity
chr-tography
on
protein
A-Sepharose
CL-48
"sing
the nethod described previovely
(30).
Antibodies honnd by the
protein
A
were
elvred wich 0.1K
acetic
acid containing
NaC1
(0.15
M).
Antibody-containing fractions
were
bdiatedly neutralized by the
addition
of
one-tenth
volume
Of
1.0
M
NaBco3,
and
were
then
didyzed against
10
mH
NaPi buffer,
pU
6.0. containire zlvcine
(50
mK,
and
NaN?
(0.4
oll).
for further wrification bv
".
.
.
ion-exchange
chromatography
on
columns
of
C~~ellvlose
(yhauan,
cK.52)
equilibrated
io
the
sane
buffer. The
columns
(30
ml
bed volume far
120
mg
of
protein)
were
developed with
gradients of UaC1
in
the
same
buffer; C-100
OH
for 55,
-200
aM
for
1-2.
anti-81,
anti31118
and
anri-Tll1C.
and
C-3W
OH
for
530.
The
purified antibodies
yere
finally
dialyzed
against
10
oll
KPi
buffer,
pU
7.2.
containing NaCl
(145
.M)
and
stored at
-700
C.
.
..
passed through protein A-Sepharose
CL-48
to
remove
rraeee
of
murine
inmvnoglobvlios
that
do
Anti-TllU does
not
bind to
protein
A.
Aecites fluid
containing
this antibody
was
hind
to
protein
A.
The eluate
was
then
fractionated
by adding
(NE$)2So4
to
50%
containing
NaCl (145
a),
and then dialyzed
into
the
pa
6.0 buffer for chromatography
an
a
saturation.
The
precipitated protein
vas
disolved
in
10
ol(
KPi
buffer.
pH 7.2,
column Of Cm-cellvlose
aa
described *bow:
the
column
ras
developed with
a
gradient Of
0-300
OH
NaC1. Fractions containing anti-T1Iu
were
pooled.
caneeotra~ed
and submitted
to
gel filtration
on
a
col-
(99
cm
I
2.6
cm)
of Sephacryl
S-300
equilibraced
in
10
mM
KPi,
pH
7.2,
containing
NeCl
(145
a).
All
the
antibodies
were
judged
pure
by pdyacrylamidJdodecy1 sulfate
gel
electrophoresis and hy isoelectrofoc*ssing
(pH
range
3.5
co
9.5)
om
polyacrylamide
gels.
The is0electzofacussing gels showed that zhe purified
monoel~nal
antibodies
gave
the
typical
Pattern
of
3
to 5 closely wmed bands that ha8
been
deeeribed
previoYs1y
(31). The
averape
final yield of each antibod?,
per
m1
of ascites fluid.
Yere
3.0-mg
for 55, 0.06
mg
for
530;
mg
for
anti-Tll1c.
2.4
mg
for aoti-Bl,
1.6
mg
for
1-2.
3.2
mg
for anLi-TllU,
3.0
mg
for
anti-Tl11B
and
3.2
Source
and
Purification Of Ribosome-Inactivating Profeine
-
Seeds frm Gelonium multiflocum
were
from United
Chemlcal
and Allied Products. 10 Clive
Bow.
Calcutta-1, India,
and
were
obtained through the
mer
Corporation. North
Bergeo,
NJ. Gelonin
was
purified
by
the method
previously described
(20).
Seeds and leases of PhycoLaeea
anericana
(pokeweed)
were
collected
in
Brookiioe,
MA.
The
pokeweed antiviral
proteine,
PAP
and PAP
11,
-re
pvrified
from
leavea
as
described previovsly
(21),
and PAP-S
was
purified from the seeds by the
method described ~reviovsly
(32).
Dana-Parber
Cancer Institute,
and
one
m&y lymphoblasroid
eell
line obtained frao
Dr.
N.
Cell
Lines
-
Seventeen
different human lymphoblasroid
cell
lines
available
at
the
L. Lezsio,
N.
E.
Regional Primate Center, Sonthboro,
M.
were used
in
the
course
of this
work.
Representative
experiments described
io
the
text
employed
the
nine
eell
lines
described below. CALLA-positive human
cell
lines inclvded Dial-6
(37),
BJAB
(35), Namlva
(ATCC
CBL
1432).
and Raji
(35).
Nal-6. Namalva and
aaji
were
also meed
as
positive cell
lines
in
teste
with antibody 1-2,
and
Ualn;6
-6
used
as
a
positive
line with the antibody
530.
BJAB,
Raji and
Rams
(35)
were
usad
as
positive cell
lines
in
tests
with anti-81. The
anti-Tl1 antibodies. vhieh
were
produced against
the
Ill
antigen found
on
h-n
T
cells,
also
react
with the homologova molecule
on
T
cella frao
other
primates
(34):
the
experiments
with
the
anti-TU
antibodies described
in
this work utilized the
cell
line 1022
derived from lymphocytes from
a
cotcoo-topped
ma-set
(36). The human eell liuee XOLT-4
with
various
antibodies.
Cells
were
grom
et
37O
c
in
a
hvnidified amsphere containing
(ATCC
CRL
1582).
USB-2
(38)
and
HL-60 (39)
were
uSed
as
negative
cell
lines
in
experiments
5%
CO2
in
FmlI-1640 lmdivn (Gibco Laboraforie~,
Grand
Island.
NY)
suppleneoted with
hear-inactivated (560
C
for
30
min)
fecal
calf
#e-
(10%) (Flow Laboratories, KcLeao,
pghl).
Cells
were
maintained
io
asynchronous exponential growth by dilution
twice
per
week
VA), L-glutamine
(2
a).
pyruvate
(1
a),
penicillin
(50
unirslml) and streptomycin
(50
to
about 1-2
x
lo5
cellslml.
mTU0DS
Kodificarioo of 55 with SPDP
-
Portions of
a
sfmk
solution
(10
nH)
of SPDP
in
ethanol
Yere
added
to
J5
(1
dml)
in
100
mM
NaP, buffer.
PU
7.0.
containinl
EDTA
(0.5
a),
and the
mixture
YBS
incvbared
at
300
c
for
50
an.
aeaetio&
yere
terPinaeed.by
gel
filtration
or
by dialysis
as
described above for SKCC wdifications. The level of incorporation of
dithiopyridyl
grovpa
MS
lineear vith
respect
to
the
initial
concentration
of
reagent
(about
measured
as
described previously
(40).
Dithiopyridyl groups
were
etable for many days at
pH
1
group iotrcdvced per molecule of antibody at
20
p
reagent
under
these
conditions)
and
was
7.0
(40).
..
Modification of Gelouin with 2-Iminothiolane
-
Gelanin
(3-4
mglml)
in
10
Rei
buffer.
pH
7.2.
containing
NaCl (145
m)
vas
diluted
to
2
nglml
with dietilled
water.
0.5
K
triethaMlaminelHC1 buffer,
pH
8.0,
and
0.1
M
EDTA
80
that the final
caneentration
of
triethanolamine end EDTA
=re
60
OH
and
la,
respectively.
The
solufloo
vas
then
degassed
and held under
argon
at
00
C.
Stock
eolutions
of Z-inioothiolaoe.EC1
(0.5
K)
were
nM)
at
00
C
for
90
min
under
argon.
The
weme
reagent
=a
reoved
by
gel
filtration
at
prepared
as
described previously
(41).
The gelonin
was
them
treated with 2-iminothiolane (1
bo
C
on
columns of
Sephadea
025 (fine) equilibrated with
5
ml4
bis-trislaeetare buffer.
pH
5.8,
containing
NaCl (50
aM)
and
EDTA
(1
OH).
sulfhydryl groups iorrodued
into
the
protein
Were
quantified BpeetrophofomeLriCelly by zhe method of
Ellman
(42):
the
eondifiona
described here
resvlt
in the
addition
of 0.6
fo
0.7
sulfhydryl
groups
per
gelonin molecule.
a).
The
pH
of the mixture
was
raised to
7.0
by the addition of 0.5
I
triethaoolaminelHC1
buffer.
pH
8.0,
and
the mixture
was
then held
under
argon
et
4O
C
for
20
h. Finally,
iadaaeetamide
(2
IPM)
*a8
added to block any
rmaining
free sulfhydryl groups
and
inevhatioo
continued for
a0
additional hour at
250
C.
Purification of J5-Gelouin Conjugates
-
Nom-conjugated gelonin
was
renorred from the
conjugation
mixture
by
passage
of the
s~1utIoo
through
a
col- of
protein
A-Sepharose
CL-YB
(15
01
column for
1W
mg
antibody)
as
described
above.
The bound
protein
was
eluted with
0.1H
acetic
acid containing
NaC1
(0.15 K) and 0.1
"01
of 1.0
I
EPi
buffer.
pH
7.5,
vas
added
to
each fraction immediately
after
collection.
The procein
was
dialyzed
against
5
mK
NaPi
buffer.
pH 6.5, containing NaCl
(35
mK)
and
NaN3
(0.4
m)
and
YBS
then
applied
to
a
column
of
Cwcellulose (Whatman, CI-52;
30
ml
col- for 100
mg
antibody) vhieh had been
eqvilibrated wizh the
same
buffer. Non-conjugated
55
dces
nof
bind to Cm"cellu1ose under
these
precise
conditions of
ionic
strength and
pH,
and
was
reoved
from the column by
eluted
in
a
-11
~olume with 100
mn
NaPi buffer.
pH
6.5,
containing
Nacl
(1.0
K). The
washing with buffer.
The
gelonlo-containing
eonjvgare
was
band by the eolm
and
was
conjugated
protein,
now
free of non-eoojugated
55
and
gelonin,
MS
submitted
to
gel
filtration
on
a
colvoo
of Sephacryl
s-300
(99
em
x
2.6
cm)
eqvilibrated with
10
mM
KPi
buffer,
pH
7.0,
containing
NaCl
(145
mn)
io
order
to
remove
aggregates of high molecular
weight. The
J5-gelonio
conjugate
was
finally sterilized by passage of the
solution
throvgh
a
0.22
p
filtration
membrane
(Killex+V, Millipore CorporPLion, Bedford,
"4).
Purification of Conjqates with
Other
Antibodies. and Conj
ace6
with Polrereed Antiviral
Proteins
-
The procedure
for
the conjugation of
gelonin
wig
anti-81, and the purification
of
the
conjugate,
were
exactly
as
described for
55.
The methods for
making
and purifying
eooiwates of
eelonin
with
the
antibodies
1-2.
530.
anti-Tlhn
and anti-TlhI- differed
_"
from the example of
J5
only
in
the
solutions
used for the
separation
of noo-conjugared
antibody from conjugate
using
Cm-cellulose.
All
far buffera contained NaPi
(5
mM)
and
NaNg
(0.4
mM)
and
were
adjusted
fo
pH
6.5,
except
for the buffer for 530 which
We
pH
7.0.
The
eoncencration of UaCl
-8
40
m
in
the solvrions
for
1-2 and
530.
34
mn
in
the
wt
bind
to
protein
A.
Following
conjugation
of this antibody
to
gelomin
as
described for
solvtion for anti-T1llC, and
25
mM
in
the
solution for aoti-Tl11B. Anti-Tllu does
J5.
the
reaction
mixture
was
eooeentrafed
by
ultrafiltration,
and
the
excess
non-conjugated
gelonin and
aggregates
of high
molecular
veight
were
sepazated
by
gel
filtration
on
a
column
equilibrated with 10
mH
ai
buffer, pH
7.0,
cooraining
NaCl
(145
a).
Nowconjugated
(99
em
x
2.5
cm,
for
a
12
ml
sample containing 100
mg
of antibody) of Sephacryl
S-300
J5
except
that
the
buffer
-6
5
on
NaPi buffer. pU 6.5.
cooraining
NaCl
(21.5
OH)
and
anti-Tllu
MS
separated frm the conjugate
by
Crcelldaee fractionation
as
described far
NaNg
(0.4
m).
The pvrified conjugate
was
elvted from the column
a9
described
above
ad
dialyzed into 10
mK
KPi buffer.
pH
7.0,
containing
NaCl
(145
d).
PAP, PAP
11
and
PAP-S
were
all
conjugated with
55
and
with anti-B1.
PAP-s
-8
also
canjugaced with anti-TllU
and anti-Tlllg.
All
these
conjugates
were
prepared
and
purified
using
methods identical
to
those
for
the
corresponding gelonin conjugates.
Meaavr-nt of
Protein
Concentration
-
Protein
cowentratioos
of solurlans of purified
proteios
Mre
determined from their A280.
essdng
El:.
Values
of
14.0
for IgG,
6.7
far
gelonin
(ZO),
and
8.3.
8.9 and 8.9 for PAP,
PAP
I1
(21). and PAP-S (43), respectively.
The
protein
concentrations
Of purified conjugates
were
also
e~~hafed
fern A280
values:
for
example. antibody-gelonin cmnjugaces of
molar
ratio
13
ez
122
were
cltlculaced
to
have
Efn
Valves
of
12.8
and
12.0.
respeoriuely. The
actual
ratio
may
be esLinaeed from
polyacrylamideldodecyl rvlfate
gels.
Polyaerylardde gel electrmhoresis
-
Cross-linking reactions
and
conjvgate purification
were
analyzed by polyaerykatde/dodccyl sulfate
gel
electrophoresis
in
gel
slabs
(14.5
-
I
90
mn
x
0.75
.sn)
Cast with acrylamide gradients (5-102
(-1")
for
gele
run
under
no!Yredueiug
conditioos)
prepared
as
described previously
(44).
Sample buffers for gels
Iyn
under
mn-reducing conditions contained 10
aglml
iodoacetamtde
(41).
Gel*
for samples under
reducing
conditions
rere
cast with 12.5%
(wlv)
Icrylamide.
or
with 1C-2OX
(riv)
aerylamide
gradients.
some
gel8
were
dried onto dialysis nabrams for
sfaming
vith
a
Quick-Scan gel
wanner
from uelena Laboratories. Iswleccrofocvssing
gels
were
No
on
a
mltipbor
electrophoresis apparatus from
LRB using
PAG plates
(W)
containing
aapholines of
a
pH
,.
L
_"
range
from
3.5
to 9.5.
Ant
e=Bindi
Activif of Antibodies
or
COD
ate*
-
The binding activity of antibodies
or
conkates wa?measuredyby
indirect
iolunoflu%ceoce (45).
Cells
(1
x
lo6)
were
incubated
at
0'
C
for
30
mi0
with
serial
dilutions of antibody
or
eonjugace
in
100 pl of
Eagle's Kinhum Essential Kedium for
suspension
cultures
(Gibe0
Laboratories) supplemented
with
2.5% (71")
pooled
human
serm
of &type
and
1x
(TI%?)
1
H
WPES buffer,
pU
7.2.
containing
NaC1 (0.9%
wh).
The
cells
were
then washed three
times
with iee-eold medium
before they
were
stained with fluoreecei.rlabelled goat
anti-use
IgG
antibody for
30
mio
at
Oo
C,
using
100
pl
of
a
1:25 dilvtion of the stock
SolnLion
(Keloy
Laboratories)
with
medium. The
cells
yere
again
washed three
rioes
with ice-cold medium. The flvoreseent
antibody-coated cells
were
finally analyzed
on
an
EPICS
IV
fluorescence-activated cell
sorter
(Covlrer
Electronics, Hialeah,
a).
The binding activity Of conjugates that contained gelonin
MS
a180
determined by
antiaervl
before
staining
with fluorescein-labelled
goat
anti-rabbit
Ig
antibody,
wing
incubating
cells. which had been treated vith conjugate. with
B
rabbit anti-gelonin
procedures
similar
to
those
described
above.
The
rabbit
anti-gelonio
antisem
was
kindly
gelonin
ia
complete Freund's adjuvanc
3
times
per
week
for
2
weeks
wing
1
mg
of gelonin
prepared by
Dr.
Jerome
Ritz.
uew
zealand
vhite rabbits
were
injected
(B.c.
and
i.m.)
with
each time.
The
rabbits
were
bled follaving
1
or
2
booster injections of gelonin (1
mg)
in
incomplete adjuvant.
pH
7.4,
containing
NaCl
(20
mH)
and
bovine
*e-
albumin
(0.2
Iglml),
yere
added
fo
the
samples
of
gelonin
Qr
conjugates. dilvted
to
0.02
pglml
of
gelonin with
10
mH
KPf
buffer,
vere
started by
the
addition
of 16
p
of
a
mixture
comainlog
aalrs
and
bvffer eoekfail (New
reticulocyte lysate (10
p)
in
0.5
ml
Eppendorf
centrifuge tubes
at
O'
C.
The
reactions
England Nuclear),
a
mixture
of
19
amino
acids
as
described
preriavsly
(46),
creatine
phosphate (0.15
pmol),
creatine
phosphokinase
(2.5
pg).
&A
(80
ng).
and
13Ul-lemire
(16
pCi)
diluted
to
a
specific radioactivity of
57
mCilpol. After rapid
mixing,
the cvbes
yere
incubated at
300
c.
sanples
(3
pl)
were
taken
at
different
times
and
the
incorporation of
[%l-leueine
into
prorein
vas
quenched
by
dilutioa.
into
dietilled
Yarer
(0.4
ml).
Radiolabelled
protein
was
quantified
a8
described
previously
(46).
Purified
Immunotoxins
12039
RESULTS
Preparation
and
Pvrification of Conjugates
Between
Antibodies
and
Pokeweed Antiviral
~
Proteins
-
PAP, PAP
I1
and
PAP-S
were
each
conjugated
to
55
using
SPDP
to
modify the
J5,
thus farninp.
cooiuaares
that
conrained
a
disulfide bond.
We
fod that ezacclv
the
8-
conditions
iould-be
wed
for
the
preparation
and
purification of each of the cbnjugates with
the three different pokeweed
antiviral
proteins
aa
described for the corresponding gelonin
conjugate. PAP-S
ha8
also
been
conjugated
to
aoti-Bl, nnti-Tllm
and
anti-Tlllg
using
the
disulfide-linker,
and
has
been
conjugated to J5
using
the linker containing
a
ehioeeher
bond.
Again,
their
preparation
and
purification followed exactly
that
of the corresponding
gelonin
conjugate.
Elboame-Inaetiwting Activity of the Conjugarea
-
Figure
3
(a)
Shows
Chat
20
pg
of
gelonin
completely inhibited
protein
synthesis
io
the
rabbit reticulocyte lysate system. Assays of
gelonin
and of gelonin modified with 2-Mnothiolane (1.4 sulfhydryl gmoupslmol) after prior
reduction with dirhioerythritol
(20
m,
300
C
for
30
nin),
shod exactly the
-
gelonin
by rasetion with
2-imioothiolsne,
without
impairing
its
cappaciry
to
inhibit protein
inhibition
(Fig.
3
(a)).
Up
to
4 sulfhydryl groups
per
molecule
could
be
introduced
into
the
rate
of inhibition of protein
synthesis
was
slwer
than that of native gelonin and,
as
egothesis (data
not
shown).
When
gelonio
ns
linked
to
J5
wing
the non-cleavable linker,
Assays done with fnrther dilution- of Dntivc gelonin
swested
that
the activity of
the
expected.
this could not
be
affected
by
prior
reduction with dithioerythritol
(Fig.
3
(a)).
gelonio
linked
LO
55
in
this
ray
was
aht
20-30%
of that of native
gelonin.
The
gelonio
within
a
J5-gelonin conjugate formed with
the
disulfide-linker
was
ala0
lese
active
t-a
native
gelonin when assayed
withrmt
prior reduction
(Pig.
3(b)).
Eosevel,
prrincubetion
of
the
conjugate dth dithioerythritol released fully
active
gelonin
that
me
indistinguishable
froc
nafise
gelonin
in
its
ability
fo
inhibit
protein
synthesis
in
fbse
assays
(Fig.
3
(b)).
Figure
3
(c)
shorr
tkt
ideacical
results
were
ebcained
with
the
anti-Bl-gelonin
canjugarea fomd with
the
disulfidrlinker. These
experiment*
with J5
and
anti-81
are
Aollogous
results
were
obtained with all
the
conjugates
using
the pokered antiviral
repreaeotatiye of the
results
obtained with
the
other
antibodies uaed
in
tUis
work.
proteins:
that
is,
the inhibitory activity Of these
toxins
when
conjugated
LO
ao
antibody
ns
reduced
to
abovt
20-301
of
that
Of the
native
proteim,
and
reductio0
af the conjugates
eantaioiug
il
disulfide linker resulted
in
release of Proteins
that
were
fully fuuctiozd
(for example,
Figure
3
(d)).
12040
Purified Immunotoxins
567
0
0.
m
I
0
-
x
s'
93
30
60
80
100
120
(b)
FRACTION
(3
mL)
FRACTION NUMBER
75
80
84
88
92
94
96
98
100102
104
600
-
(
a
200
22
I
I90
160
IO0
50
400:
200
fb'^
:
.
kp-n-@
rr)
I
0
-
X
L
s
93
53
100
50
30
IO
30 50
Purified Immunotoxins
TABLE
1
TBB
CTTOTOXICITT
OF
GELONIA
AND
ITS
DISULFIDE
LINBED
IBlONIOMXIN
CONJUGATES
TOWABDS
IPLIPBOBLASTOID
CELL
LINES*
12041
Cell line: Nanalra Baln-6
%JAB
Raji
1022
MOLT-4
&L-60
Gelonin
JSIgelanin
1-2Ige1onin
J301gelonia
anri-T1llAlgelonin
anri-TlllE/gelooin
anri-T1llC/gelonin
aori-Bl/gelooin
700
400
,100
0.4(+)
0.02(+)
0.08(+)
0.02(+)
0.001(+)
0.03(+)
0.04(+)
,100
(-)
,100
(-)
.IO0
(e)
'loo(+)
1000 >1w
200
300
0.2(+)
>lw(-)
.loo(-)
'loo(-)
8.0(+)
0.3(+)
2.0(+)
,loo(+) >loo(-)
40
lO"0
10-6
*
Cytotoxicity
MS
measured
after
3
days exposure of the cells
to
the gelonin
or
fnmunotaxio,
measuring
[%]-thymidine
incorporation
into
DNA during
a
2
h pulse
as
a"
index of cytotoxicity.
The
ID50
values
(OM)
are
the
mean
of
at
least
three independent
experiments,
such
as
those
shown
in
Figure
7,
where
each
point
ia
the
mean
of triplicate
deierminatiom.
The ID50 of
J5-gelonin
on
Nal-6 cells
was
0.02
2
0.01
nM ("-8).
The
presence
Or
absence of the relevant antigen
00
the
cell
surface,
to
vhich the
particular
antibody would bind,
is
indicated by
(+)
or
(-),
respectively.
1
P
i_
-10
10-6
ANTIBODY CONCENTRATION
(MI
Figure
4.
Capariaon of
the
binding of
itmuootoxios
with
that
of
native
antibody
oeasvred by
indirecr
itmunofluoresceace. The
relative
fluozescence
in
arbitrary
units
vas
cell
sorter,
and
was
plotted
against
antibody concentration of
the
Mfive
antibody
(W,
or
taken
as
the
mid-point Of flvorescence
histograms
determined
"sing
a
flYo~'escenceacfiva~ed
Panel (b),
J5
and 55-gelonin
containing
the
chioefher
link. All of
the
remaining
panels
(c
of
che
imunomxin
(0).
Panel
(a),
J5
and
J5-gelonin
containing
the
disvlfide
link.
(d).
55
and
J5-gelonin
eoacsioing
exactly
2
molecules of gelooin per
35
molecule. Panel
LO
h)
were
of conjugates
cooraining
disulfide
links.
Panel
(c),
55
and
55-PAP-S. Panel
(e),
anti-81
and
anti-El-geloaio. Panel (f). anti-TllM
and
anti-Tll~-gelonin. Pam1
(8),
aori-Tlllg and anri-Tlllg-gelonin. Panel
(h),
anci-Tl11C
ad
aori-IlllC-gelonin.
10-
1
