(CANCER RESEARCH 50, 3249-3256. June I, 1990]
Antitumor Activity of L6-Ricin Immunotoxin against the H2981-T3 Lung
Adenocarcinoma Cell Line in Vitro and in Vivo1
Heinz Schmidberger,2 Laurel King, Larry C. Lasky, and Daniel A. Vallera3
Department of Therapeutic Radiology, Section on Cancer Immunology [H. S., D. A. V.], Department of Obstetrics and Cynecology [L. K.J, and Department of Laboratory
Medicine/Pathology [L. C. L., D. A. V.], University of Minnesota Hospitals and Clinics, Minneapolis, Minnesota 54455, and Memorial Blood Center of Minneapolis,
Minneapolis, Minnesota 55404 ¡L.C. /../
ABSTRACT
The monoclonal antibody L6 recognizes a determinant that is ex
pressed on lung, breast, colon, and ovarian carcinomas and is present
only at trace levels in normal tissues. 16 was covalently linked to intact
ricin by a thioether bond to produce an immunotoxin (IT). Gel analysis
revealed that this IT was heterogeneous, but mostly one monoclonal
antibody molecule linked to one ricin molecule. The L6-ricin IT selec
tively bound and was selectively toxic to L6-positive H2981-T3 adenocarcinoma cells in protein synthesis inhibition assays in which lactose
was added to block the native ricin binding site. Clonogenic studies
showed that 1 Mg/ml L6-ricin could inhibit about 99.99% of H2981-T3
growth in a limiting dilution assay, even in the presence of a 20-fold
excess of human bone marrow cells. Treatment of bone marrow cells with
the same dose of I d-riciii resulted in the growth of ample numbers of
bone marrow progenitor cells (colony-forming units-mixed, colony-form
ing units granulocyte/macrophage, and blast-forming units erythroid)
after 14 days.
We also evaluated the antitumor effect of 1.6-rinn administered intratumorally with lactose against established H2981-T3 tumors in a nude
mouse model. Thirty % of the tumor-bearing animals responded com
pletely to single-dose treatment, while 60% gave partial responses. The
in vivo effects were not absolutely specific, since irrelevant anti-CD5 IT
also induced tumor regression in this model (10% responded completely,
while 30% gave partial responses). However, irrelevant IT gave higher
systemic toxicity (50% mortality) than L6-ricin (23% mortality). The
nonspecific activity of IT was possibly due to Fc binding, which was
demonstrated in vitro, or due to ricin B-chain binding. Ricin alone was
too toxic for sustained tumor protection. Unconjugated L6 had no antitumor effect. The data suggest that I 6-ricin may be useful for in vitro
purging of autologous bone marrow from patients with solid tumors and
marrow involvement and for in vivo regional therapy of Lo-positive
carcinomas.
INTRODUCTION
IT4 are antibodies covalently linked to bacterial or plant
toxins (reviewed in Ref. l). The antibody moiety specifically
binds the IT to cells expressing the appropriate antigen, while
the toxin catalytically inhibits protein synthesis and causes cell
death. Intact ricin is particularly effective and has been conju
gated to a variety of MoAb for preclinical studies (2-5). Clinical
trials have evaluated intact ricin IT for donor T-cell depletion
prior to allogeneic bone marrow transplantation (6) and for the
purging of leukemia cells from autologous bone marrow grafts
(7). We (8) and others (9) reported the therapeutic efficacy of
ricin IT on human lymphoma xenografts in nude mice. The
efficacy of this treatment can be increased by combining it with
chemotherapy (10) or using radioimmunotherapy (11, 12).
The ricin molecule consists of two A/r 32,000 subunits joined
by a disulfide bond. The A-chain subunit is an enzyme which
inactivates the 28S component of the 60S subunit of ribosomes
(13). The B-chain subunit contains a binding site that binds to
terminal nonreducing galactose residues on the surface of all
eukaryotic cells. Studies have shown that B-chain facilitates Achain activity (14). B-chain also slows IT trafficking to the
lysosomes, likely protecting the A-chain from lysosomal deg
radation (15). Native ricin binding can be prevented by coincubating IT with lactose, which blocks the galactose binding
site of the B-chain, rendering the IT more immunologically
specific (16).
The MoAb L6 recognizes a carbohydrate antigen that is
strongly expressed in lung, breast, colon, and ovarian carcino
mas (17, 18), while this antigen is minimally expressed on
normal human cells. In this study, we have investigated the
therapeutic potential of L6-ricin IT in vitro against a human
adenocarcinoma, H2981-T3. This is in view of the fact that
major centers are currently evaluating the efficacy of strategies
designed to eliminate residual tumor cells from autologous bone
marrow grafts in patients with solid tumors (19-22). Reagents
destroying tumor cells and not bone marrow progenitor cells
could be used to purge marrow in vitro, permitting the clinical
use of more aggressive therapy by radiation and/or anticancer
drugs.
Intact ricin IT may also be useful for the regional treatment
of cancers (3, 23), such as ovarian carcinomas which often occur
in areas amenable to intracavitary treatment (24). Therefore,
we investigated the activity of L6-ricin against H2981-T3 car
cinomas in vivo, using a nude mouse model.
MATERIALS
AND METHODS
Antibodies. L6 is an IgG2a mouse MoAb recognizing a carbohydrate
determinant expressed on most human carcinomas. Normal adult tis
sues express no more than trace amounts of this antigen (17). The
hybridoma resulted from a fusion of BALB/c splenocytes (immunized
against human non-small cell carcinoma) and NS-1 mouse myeloma
cells. T101 is an IgG2a mouse MoAb that recognizes CDS, a M, 67,000
glycoprotein on normal and malignant T-lymphocytes (25). UCHT1 is
an IgGl MoAb which recognizes the CD3 (p20) component of the Tcell receptor (26). Human 7-globulin was obtained from Sigma Chem
icals (St. Louis, MO).
Immunotoxin Preparation. L6 or anti-CD5 were linked to intact ricin
via a covalent thioether linkage, which has been previously described
(2). Briefly, ricin was reacted with m-maleimidobenzoyl-A/-hydroxysuccinimide ester at a 3:1 molar ratio to attach maleimide linkage residues
to the lysines of ricin. Monoclonal antibodies were prepared for crosslinking by reduction of disulfide bonds with 91 ITIMdithiothreitol for
30 min. Derivatized ricin was then mixed with the sulfhydryl-containing
MoAb, resulting in the formation of a nonreducible thioether linkage
3249
Received 8/29/89; revised 2/14/90.
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.
1This work was supported in part by USPHS Grants PO1-CA-21737, RO1CA-31618, and RO1-CA-36725 awarded by the National Cancer Institute.
DHHS; American Cancer Society Research Grant 1M-502; and the Minnesota
Medical Foundation. Center for Experimental Transplantation and Cancer Re
search publication 58. This paper is no. 4 in a series on immunotoxin therapy in
nude mice.
1Supported in part by a fellowship grant from the Rotary' Foundation.
3 To whom requests for reprints should be addressed, at Box 367 UMHC,
Harvard Street at East River Road, University of Minnesota. Minneapolis.
Minnesota 55455.
4 The abbreviations used are: IT, immunotoxin(s); MoAb, monoclonal anti
body; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis;
PBS, phosphate-buffered saline; PCS, fetal calf serum; FACS. fluorescenceactivated cell sorting; MFI, mean fluorescence intensity; TCA, trichloroacetic
acid; CFU-GM, colony-forming unit granulocyte/macrophage;
BFU-E, burstforming unit erythroid; PBMC, peripheral blood mononuclear cells; CPU-mix,
colony-forming unit mixed.
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-R1CIN ANTITUMOR
between the two species. In a typical conjugation, 5.0 mg of purified
MoAb were reduced and reacted with 12.4 mg of m-maleimidobenzoyl/V-hydroxysuccinimide ester-derivatized ricin to give a molar ratio of
ricin to MoAb of 6:1. IT was purified by high performance liquid
chromatography on a 21.5- x 600-mm SW size exclusion column (Toya
Soda, Japan and Beckman Instruments) to remove unreacted ricin.
Unreacted MoAb was removed by passage over a Sepharose 4B affinity
column, which binds the galactose binding site of ricin B-chain. Purified
IT was eluted from the column with 50 HIM lactose. The yield of
purified IT was 20-40% of the concentration of MoAb used.
Gel Analysis. Analysis was performed on a 10% PAGE plate (Daiichi
Pure Chemicals, Tokyo, Japan), using a Mini Protean II slab cell (BioRad, Richmond, CA). Five Mgof IT were suspended in sample buffer
(40 HIMTris buffer, pH 6.8, 1.5% SDS, 7.5% glycerol, 0.005% bromophenol blue), boiled for 5 min, and added to the plate. Electrophoresis was performed using a constant current (26 mA), for a running
time of 100 min, in electrode buffer (0.025 M Tris buffer, 0.192 M
glycine, 0.1% SDS). The plates were stained with Coomassie blue,
dried, and scanned using a Du-62 spectrophotometer with gel-scanning
accessory (Beckman Instruments, Fullerton, CA). The photometry was
performed at 560-nm wavelength. Quantitation of the bands was carried
out by densitometric analysis using the Gel Scan Soft Pack Module
software package (Beckman Instruments, Fullerton, CA).
Target Cell Line and Tumor. The cell line H2981 was established at
Oncogen Corporation (Seattle, WA) from a primary human adenocarcinoma of the lung (17) and was further propagated at the University
of Minnesota under the designation H2981-T3. The cells grew as
adherent monolayers in RPMI 1640 medium supplemented with 10%
heat-inactivated FCS, 1% glutamine, and 1% penicillin/streptomycin,
at 37°Cin a 5% CO2/95% air atmosphere. Once a week, cells were
removed from monolayers with a PBS solution containing 0.05%
trypsin and 0.02% EDTA and were passaged. Before resuspended cells
were used for in vitro assays, they were incubated at 37°Cfor l h to
allow reexpression of receptors that may have been damaged by trypsinization (FACS studies of the kinetics of reexpression of the L6
antigen showed that reexpression was complete by this time). Prior to
Scatchard analysis, adherent cells were scraped from the flasks rather
than trypsinized. The human T-cell leukemia line CEM was maintained
at 37°Cin 5% CO2/95% air and passaged every 3 days in RPMI 1640
culture medium supplemented with 10% heat-inactivated FCS, 1%
glutamine, and 1% penicillin/streptomycin.
Immunofluorescence of Cell Lines. The binding of unconjugated
MoAb and IT to H2981-T3 and CEM cells was determined by indirect
immunofluorescence. In brief, IO6 cells in 0.1 ml were incubated with
various concentrations of MoAb or IT for 30 min at 4°C,in the presence
of 0.015% NaNj. After incubation, cells were washed and incubated
with 50 M' of a 1:20 dilution of fluorescein isothiocyanate-conjugated
goat anti-mouse 7-globulin G (Meloy Laboratories, Springfield, VA).
Cells were washed 3 times, fixed in 400 ¿il1% paraformaldehyde, and
analyzed using a FACS IV (Becton Dickinson, Mountain View, CA).
To test for the presence of the L6-defmed antigen in tumors not
cured by treatment with 1.6 ricin, relapsing tumor that had initially
responded to L6-ricin treatment (95% tumor size reduction) was re
moved from the animal and placed in a spinner flask containing 7 ml
of trypsin-EDTA. After 30 min, a 0.5-ml aliquot was removed, pipetted
into medium containing 20% FCS, spun, and washed again. This was
repeated every 10 min for 40 min. Cells were pooled, incubated for 1
h, and then tested by FACS analysis for the presence of L6 antigen.
Modulation Assay. Modulation was assessed as previously described
(27). One hundred-^! aliquots of 10' cells were incubated with various
amounts of MoAb for 10 min at 4°Cto bind antigen sites. Subsequently,
cultures were transferred to a 37°Cwater bath for the appropriate
incubation period with the respective MoAb. To halt any MoAbinduced modulation, the test samples were diluted with ice-cold PBS,
5% FCS, containing 0.015% NaN3. Cells were washed and suspended
for 30 min at 4°Cin 100 n%/m\ L6 or anti-CD5 MoAb, respectively,
to bind any remaining or reexpressing receptor sites.
As a positive control, one sample was incubated in 100 ng/m\ L6 or
anti-CD5. As a negative control, one sample was incubated with 100
ng/ml IgG2a mouse myeloma protein (Cooper Biomedicai, Malvern,
PA). Both samples were incubated for 30 min at 4°C.For fluorescence
ACTIVITY
staining, all samples were washed twice and incubated for 30 min at
4°Cwith 50 n\ of fluorescein isothiocyanate-goat anti-mouse 7-globulin
G diluted 1:20. Cells were then washed 3 times, diluted in 400 /il 1%
paraformaldehyde, and analyzed by FACS. Nonviable cells were gated
out of the analysis. For each specimen, MFI was calculated over 256
channels and plotted in a linear fashion. Modulation (M) was quantitated by the following equation:
M = 1-
MFI of the test sample - MFI of negative control
MFI of positive control —MFI of negative control
x 100%
Protein Synthesis Inhibition Assay. Cells were incubated overnight in
flat-bottomed 96-well plates (Costar, Cambridge, MA) in RPMI 1640
supplemented with 10% heat-inactivated FCS (5 x \04 cells/200 ^1), to
allow them to adhere to plastic. The medium was then replaced by 100
¡¡I
of the same incubation medium containing the appropriate amount
of IT and 200 mM lactose. After a 2-h incubation, quadruplicate wells
were washed twice and incubated for 24 h in leucine-free RPMI 1640
containing 20% dialyzed FCS and 1% penicillin/streptomycin. The
cells were then pulsed for 4 h at 37°Cwith 1 ¿iCi['Hjleucine (Amersham, Arlington Heights, IL) per well. The monolayers were washed
twice in serum-free medium, resuspended with trypsin-EDTA (5 min,
37°C),and harvested onto glass filter paper using an automated har
vester (Bramici. Gaithersburg, MD). Filter paper discs were dried, and
pH]leucine incorporated into the cells was determined by using stand
ard scintillation counting techniques. The results were expressed as
percentages of the lactose control.
Radiolabeling. L6 MoAb and ricin were labeled with iodine-125 using
an iodine-monochloride micro-method which has been previously de
scribed (28). Briefly, 200-^g aliquots of MoAb or ricin were labeled
with 3 mCi of radionuclide and 12 equivalents of ICI. Free iodide ions
were removed by passage over a Dowex 1-X4 resin column. The
contents of the column were collected into vials containing 5% human
serum albumin. When ricin was radiolabelcd, the column was washed
with 1 ml of 100 mM lactose to block ricin binding. A 20% loss of
protein content during the labeling procedure was assumed, based on
previous studies (11), and all concentrations were adjusted using this
value.
To determine the amount of label incorporated into protein, 10%
TCA was added to the radiolabeled reagents, after which the tubes were
mixed and incubated for 30 min at room temperature. Precipitated
protein was pelleted. The pellet and supernatant were counted in a welltype gamma counter (Gamma 5500; Beckman Instruments, Fullerton,
CA). The percentage of TCA-precipitable counts was determined using
the following equation:
% TCA-precipitable counts
cpm pellet
x 100%
cpm pellet + cpm supernatant
Scatchard Analysis. Binding of radiolabeled MoAb and ricin was
assayed using a modification of previously described techniques (15,
29-32). Target cells (106/tube) were washed and resuspended in 100 p\
RPMI, 5% FCS, containing increasing amounts of radiolabeled MoAb
or ricin. Binding of the labeled reagents was lowered 100-fold by adding
unlabeled protein. Thus, it was possible to incubate with saturating
protein concentrations without surpasssing the counting limits of the
gamma counter used. The protein concentrations used for binding
studies ranged from 3.9 ng/ml to 390 jig/ml for L6 MoAb and from
2.25 to 180 Mg/ml for ricin. The binding at each protein concentration
was determined in duplicate samples. Nonspecific binding of the radi
olabeled ligand was determined for each protein concentration by
adding excessive amounts (20 times the saturating concentration) of
unlabeled ligand to the suspension prior to incubation with radioligand.
(The percentage of bound activity under blocked conditions was sub
tracted from bound activity under nonblocked conditions. The resulting
activity was taken as the corrected, specific, bound activity. The cell
suspensions were vortexed and incubated for l h at 4°C.Preliminary
experiments indicated that binding of both radiolabeled L6 and ricin
3250
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-RICIN ANTITUMOR ACTIVITY
was complete by 1 h. After incubation, 90 ß\
of the cell suspension were
layered onto a mixture of 84% silicone oil/16% mineral oil (200 ^1) m
0.5-ml polypropylene microcentrifuge tubes (Bio-Rad) and centrifuged
through the oil at 10,000 x g. Pellets and supernatants were counted
in a gamma counter.
Scatchard analysis was performed using H2981-T3 cells, to deter
mine the association constant of the equilibrium binding of labeled L6,
the number of L6 binding sites per cell, and the number of ricin binding
sites per cell. The data were plotted as the number of molecules bound
to each cell at a given dilution versus the ratio of bound MoAb to free
MoAb divided by the concentration of cells in cells/liter. The amount
of free MoAb was corrected for the immunoreactive fraction, and the
corrected binding data for specific binding were considered for the
Scatchard representation. K, values were derived from the slopes of the
linear regression lines and expressed as liters/mol (M~'). The binding
sites were expressed as «/cell.
Clonogenic Assay. Neoplastic target cells were assayed for clonogenic
growth by limiting dilution, as previously described (33).
Measurement of Normal Bone Marrow Progenitor Cells in Mixed
Colony Assays. Human bone marrow, obtained from normal donors
following informed consent, was cultured for progenitor cells using a
modification of a previously described technique (34). Briefly, cells
suspended in culture medium at concentrations ranging from 0.5 x IO5
to 2 x 10s were plated in triplicate in 15-mm wells (Nunc, Roskilde
Denmark). Culture medium consisted of Iscove's modified Dulbecco's
medium with 30% PCS, 5 x IO'2 M 2-mercaptoethanol 0.01 g/ml
methylcellulose, 100 pM recombinant granulocyte/macrophage colonystimulating factor, 5 ng/ml recombinant interleukin 3, and 1 unit/ml
recombinant erythropoietin. The recombinant cytokines were kindly
provided by Genetics Institute (Cambridge, MA). After 14 days in
culture at 37°Cin 4% CO2/96% air, colonies were visually scored using
a stereo dissecting microscope. Mixed colonies containing erythroid
and nonerythroid cells were scored as CFU-mix. Pure erythroid colo
nies were scored as BFU-E, and nonerythroid colonies were scored as
CFU-GM.
Animals. BALB/c (nu/nu) male mice, 4 weeks old, were obtained
from Charles River (Wilmington, MA) or Simonson Laboratories
(Gilroy, CA). The animals were kept under sterile conditions in cages
with filter bonnets. They were maintained on Purina Lab Chow mouse
diet and given water, supplemented with hypochlorite, ad libitum.
Establishment of H2981-T3 Tumors in Nude Mice. H2981-T3 cells
(5 x IO6 in 0.1 ml RPM1) were injected s.c. into the flanks of nude
mice. Palpable tumors were detected after 2-3 weeks. Tumor establish
ment in the animals was 80-90%. Tumor size was evaluated by meas
uring three perpendicular diameters of each tumor. Multiplication of
these diameters yielded an index for tumor size, given in cm3. Statistical
comparisons of tumor sizes between various treatment groups were
performed by Student's / test at different time points after treatment.
Tumors were read blindly.
Tumor Treatment with Immunotoxins. Tumors with a size index of
between 0.3 and 0.6 cm3 were treated by intratumoral injection of 5 or
10 /ug of L6-ricin, control anti-CD5-ricin, or control anti-CD3-ricin.
The intratumoral injections were preceded by 0.5-ml i.v. injection of
750 IHMlactose solution per mouse. The intratumoral injections con
tained 300 IHM lactose in PBS and 5 or 10 Mg/0.1 ml IT. Control
animals which only received lactose were included and usually had
tumors of 3-6 cm3 by day 30. Mice treated with unconjugated ricin did
not receive a lactose blockade prior to treatment. Responses were
considered complete if there was no detectable tumor on day 30.
Responses were defined as partial if tumors were <50% of the size of
the untreated control tumor, a definition established by the World
Health Organization.
RESULTS
Gel and HPLC Analysis of Immunotoxins. Two different
batches of L6-ricin and two different batches of anti-CD5-ricin
were tested. High performance liquid chromatography profiles
were similar if not identical for all (not shown). Data from
SDS-PAGE under nonreducing conditions and gel analysis by
scanning are shown in Fig. 1. Gel profiles of L6-ricin (Fig. 1,
«•I
Fig. 1. Gel analysis of IT. Analysis was performed on a 10% PAGE plate
under nonreducing conditions. Five ¿ig
of IT were added to the plate. The plates
were stained with Coomassie blue, dried, and scanned. Lanes I and 2, different
batches of L6-ricin IT (final material was isolated by affinity chromatography
after lactose elution from Sepharose 4B). Lanes 3 and 4, different batches of antiCD5-ricin IT. Lane S, molecular weight standards. F, reduced fibronectin (M,
220,000); A, antibody (M, 150,000); K, two species of ricin (M, 61,400 and
63.400).
lanes I and 2) and anti-CD5-ricin (Fig. 1, lanes 3 and 4) revealed
four major IT bands exceeding M, 150,000. The protein bands
corresponded to IT species of molecular weights 170,000,
190,000, 220,000, and 260,000. Scanning showed that the
largest band (on a percentage basis) for all IT was at M, 220,000
(43.3-51.6%). These bands contained over twice the amount of
any other IT band and likely represent one ricin molecule linked
to one antibody molecule. Only a small amount of M, 260,000
species was present for all IT (0.1-6.0%). This band likely
represents two ricin molecules linked to one antibody molecule.
There were no major differences in IT bands (bands of M,
> 150,000) obtained with either L6-ricin or anti-CD5-ricin. The
M, 170,000 and 190,000 species likely represent antibody frag
ments linked to ricin and represent <9.4% and <16.3% of the
total protein, respectively. We did not attempt to remove these
species. Gel scanning revealed 0.3-7.8% contaminating free
MoAb and 2.3-8.9% free ricin. L6-ricin showed a band at M,
125,000 that was not present for anti-CD5-ricin. The band
could represent one light chain reannealed to two reannealed
heavy chains, while the band at M, 100,000 for all IT could
represent reannealed heavy chain. The batch of anti-CD3-ricin
that was used for in vivo studies showed less than 5% free ricin
(not shown).
Affinity and Binding Sites of I25l-Ricin on H2981-T3 Cells.
Since the native galactose binding site of ricin is present on the
surface of eukaryotic cells, a difference in the number of galac
tose binding sites on H2981-T3 and the control CEM line
might influence our results. Thus, we studied the binding of
ricin to H2981-T3 and CEM cells. Table 1 shows the Scatchard
calculations of the binding data. Ricin bound to H2981-T3 cells
with an association constant, A"a,of 1.8 x IO6 M"1 (average of
two experiments) and each tumor cell had an average of 16 x
10" receptors. CEM cells had a similar affinity of 3.7 x IO6 M"'
and 9 x IO6 ricin binding sites/cell (average of two experi
ments). The data indicate that the binding of ricin to H2981T3 and CEM cells is comparable. Affinity of L6 MoAb and L6
3251
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-RICIN ANTITIIMOR
ACTIVITY
Table 1 Affinity of ricin and number of ricin receptors on H2981-T3 and CEM
cells as measured hy Scatchard analysis
One million cells were incubated with increasing concentrations of '"l-labcled
ricin for 60 min at 4°C.Cell-associated radioactivity was separated from unbound
activity by centrifugation through an oil layer, as described in "Materials and
Methods." Bound and free radioactivity (cpm) were analyzed by the Scatchard
studied CDS modulation of a CDS-positive neoplastic cell line,
CEM. Modulation of CDS on PBMC was 6-15-fold higher
than the modulation of the defined L6 antigen on H2981-T3
cells (49.8 and 55.2% at 1.0 and 10.0 Mg/ml antibody). Modu
lation of CDS on CEM cells was 2-fold higher (14.2 and 16.7%
method. The analysis was performed without lactose.
at 1.0 and 10.0 ng/m\ antibody). A prolonged incubation time
receptor
(M'1)1.5
with MoAb L6 (up to 16 h) did not increase the modulation of
Experiment1
lineH2981-T3
sites/cell21
the L6-defined antigen, whereas the modulation of CDS on
x 10'
x 10'
10'5.1
x
10'10
11 x
H2981-T3CEM 2.0
234Cell
PBMC increased more than 30% (data not shown). We con
clude that the L6-defined antigen does not modulate well.
x 10'
10'8
x
2.2 x 10'Ricin
x 10*
In Vitro Cytotoxic Activity of L6-Ricin Immunotoxins. To
CEMK.
evaluate the cytotoxic potential of L6-ricin IT, we incubated
adherent H2981-T3 cells with various concentrations of the IT
Table 2 Binding of MoAb L6 and L6-ricin IT to H298/-T3 cells as measured by
for 2 h at 37°Cin a lactose-containing medium (Table 4). At
flow cytometry
concentrations of 0.01, 0.1, 0.3, 1.0, and 5.0 Mg/ml, L6-ricin
Indirect ¡mmunofluorescence was performed with one million Lo-positive
H2981-T3 cells or CDS-positive CEM cells. Cells were washed, treated with
IT reduced protein synthesis to 66.2, 50.5, 42.1, 23.5, and
MoAb or IT at 4"C for 30 min, washed again, and then stained with fluorescein
7.9%, respectively. The IC50 value (i.e., the concentration of
isothiocyanate-goat anti-mouse -y-globulin G. Binding was measured by flow
L6-ricin IT inhibiting protein synthesis to 50% of control) was
cytometry using a FACS IV and expressed as percentage of positive cells. For
each cell line, background binding was determined by incubating cells with mouse
0.1 Mg/m'- When L6-ricin was incubated with a 200-fold excess
myeloma protein followed by fluorescein isothiocyanate-goat anti-mouse 7-globof unconjugated MoAb L6, inhibition of protein synthesis was
ulinG.
blocked 35-fold (IC50 = 3.5 Mg/ml). Blocking with IgG de
(%)H2981-T353.570.795.15.081.281.33.36.16.65.04.634.4CEM13.45.60.66.2ND°0.797.097.291.595.497.689.0
cells
creased the inhibition of protein synthesis 7-fold (IC50 = 0.7
ITL6L6-ricinAnti-CD5Anti-CD5-ricinConcentration
MoAb or
(Mg/ml)11050110501105011050Positive
Mg/ml). Blocking was inefficient at 5 ng/ml, which was probably
due to the inefficiency of the lactose blockade (36).
When treated with 0.01, 0.1, and 1.0 ng/ml levels of a control
IT, anti-CDS-ricin, the protein synthesis of H2981-T3 cells as
compared to controls, was 70.7, 73.0, and 52.7%, respectively.
Thus, the IC50 value for anti-CDS-ricin was 10-fold higher (1
Mg/ml) than the IC50value for L6-ricin. The effect of anti-CDSricin could be blocked almost completely by addition of 1 mg/
ml human IgG, suggesting that ricin IT can enter and kill cells
in more ways than binding of specific antibody or B-chain.
Effect of L6-Ricin on the Clonogenic Growth of H2981-T3
Cells. To determine the potency of L6-ricin, IO6 adherent
" ND, not determined.
H2981-T3 cells were treated with 1 ng/m\ L6-ricin for 24 h at
37°Cand then plated in a limiting dilution assay following
receptor number on H2981-T3 cells were also determined.
Scatchard analysis revealed a ATaof 1.6 x IO8 M~' for the L6
antigen-MoAb interaction and 170,000 L6 receptors/cell. All
Scatchard plots were curvilinear. The r value for the regression
analysis was 0.97.
Binding of MoAb or Immunotoxin to Target Cells. The binding
of unconjugated MoAb or IT to H2981-T3 (L6 positive, CDS
negative) and to CEM (L6 negative, CDS positive) cells was
determined by indirect immunofluorescence at various protein
concentrations. A representative experiment is shown in Table
2. The cell suspensions contained 200 niM lactose to block the
binding of ricin B-chain. At 1, 10, and 50 Mg/ml, L6 MoAb
bound to 53.5, 70.7, and 95.1% of the H2981-T3 cells and L6ricin bound to 5.0, 81.2, and 81.3% at the same concentrations.
This indicates that conjugation to ricin had minimal effect on
MoAb binding. Binding to the control cell line CEM was
minimal for both L6 MoAb and L6 IT. Conversely, anti-CDS
and anti-CDS-ricin IT bound to CEM cells, while binding to
H2981-T3 cells only occurred at high IT concentrations which
might overcome the lactose blockade.
Antigenic Modulation of L6 on H2981-T3 Cells. To determine
the degree of Lo-induced modulation of the receptor for MoAb
L6 on H2981-T3 cells, these cells were incubated with 0.1, 1.0,
or 10.0 Mg/ml MoAb L6 for 2 h at 37°C.We measured 7.3,
resuspension of trypsinized cells. Comparison of the colony
counts for L6-ricin-treated cells versus nontreated controls gave
an estimate of 4.1-log killing of clonogenic cells on day 7. In a
second experiment, we observed 3.5-log kill. In an experiment
in which H2981-T3 cells were admixed with a 20-fold higher
number of human bone marrow mononuclear cells, we meas
ured about 4-log kill of H2981-T3.
Effect of L6-Ricin on in Vitro Cultured Normal Bone Marrow
Progenitor Cells. To determine the toxicity of L6-ricin to human
stem cells, bone marrow was preincubated with L6-ricin,
Table 3 Modulation of the antigen defined hy MoAb L6 on H298I-T3 cells
Cells were incubated with 0.1, 1, or 10 ><g/ml MoAb for 30 min at 4'C and
then for 2 h at 37°C.Cells were washed, incubated with the same MoAb again
for 30 min at 4'C to bind any reexpressed antigen, washed again, incubated with
fluorescein ¡sothiocyanate-goat anti-mouse ->-globulin G, and analyzed by FACS.
Percentage of modulation was calculated as described in "Materials and Meth
ods." Modulation of L6 antigen was measured on H2981-T3 cells, and modulation
of CD5 antigen was measured on PBMC or CEM cells. Two experiments were
performed.
6.5, and 5.1% modulation, respectively (Table 3). As a positive
control, we measured the modulation of CDS on PBMC after
incubation with the anti-CD5 antibody. Anti-CD5 was chosen
as a control because a detailed study of the modulation of CDS
was previously published (27, 35). As a further control, we
Concentration
((jg/ml)L6
MoAb
lineH2981-T3
0.1
1.0
10.0Anti-CD5
0.1
1.0
10.0Anti-CDS
ND/16.7"
ND, not determined.
1.0
10.0Cell
(%) (expt.
1/expt.
2)8.3/7.3
H2981-T3
H2981-T3PBMC
8.8/6.5
3.1/5.147.7/56.4
PBMC
PBMCCEM
55.0/49.8
49.0/55.2ND"/14.2
CEMModulation
3252
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-RIC1N ANTITUMOR
ACTIVITY
Table 4 Effect of L6-ridn and anti-CD5-ricin on protein synthesis as measured by tritiated leucine incorporation
In a representative experiment. H298I-T3 cells were treated with various amounts of L6-ricin or anti-CD5-ricin for 2 h. For blocking. 1 mg/ml unconjugated L6
MoAb or human IgG were added to the preincubation mixture. Cells were washed, incubated for 24 h, pulsed with tritiated leucine. and then harvested. The cpm/
well were determined (from quadruplicate samples) and data are expressed as a percentage of control protein synthesis, as described in "Materials and Methods." The
cpm/well for untreated control cultures were 21,034 ±4,805 (mean ±l SD). SD for the data shown in this table did not exceed 23% of the mean value.
(%L6-ricin/IgG
synthesis
Concentration
of IT(Mg/ml)0.01
50.5
42.1
23.5
7.90.1L6-ricin/L6
0.1
0.31
5IC50'Lo-ricin66.2
block90.8
block88.6
103.8
92.5
81.2
42.83.5Protein
80.3
90.4
37.2
6.40.7of
control)Anti-CD5-ricin70.7
block94.0
73.0
ND°
52.7
8.41.0Anti-CD5-ricin/IgG
94.6
92.6
92.6
86.8NA*
" ND, not determined.
* NA, not available.
c IC50, 50% inhibitory concentration.
washed, and then cultured for the growth of CPU-mix, BFUE, and CFU-GM after 14 days. In Table 5, experiment 1, Loricin at 1 Mg/ml did not inhibit CPU-mix or BFU-E; CFU-GM
were inhibited by 64%. In a second experiment with cells from
a different donor, CPU-mix was inhibited by 40%, BFU-E was
inhibited by 24%, and CFU-GM was inhibited by 31%.
In Vivo Treatment of Established H2981-T3 Tumors with
Intratumoral Ricin. Before evaluating the antitumor efficacy of
ricin IT in vivo, the effect of free ricin on tumor growth was
examined. Nude mice were s.c. inoculated with 5 x IO6H2981-
Control
(4/4)
0.2ug
(3/3)
u
in
o
T3 cells. Three weeks later, they had tumors with a size index
of 0.3-0.6 cm3, which were treated with 0.2, 0.4, 0.6, or 1.0 ßg
ricin in 100 ^1 PBS. Control animals received 100 /ul lactose.
Intratumoral injection of ricin had a dose-dependent inhibitory
effect on tumor growth in vivo. As shown in Fig. 2, tumors
injected with 0.2, 0.4, or 0.6 jug of ricin grew more slowly than
the controls. The difference in tumor size indices between
animals treated with 0.4 /ug ricin and controls was significant,
as shown by Student's / test on day 30 after treatment (P <
0.021). However, at a dose of 0.6 pg, two of the four treated
animals died within 6 days due to ricin toxicity. These animals
showed a characteristic wasting syndrome with a purpuric rash
distributed over the whole body, which appeared 1-2 days after
treatment. At a dose of 1 ¿¿g
ricin, all of the six treated animals
died within 5 days. Animals which survived treatment with ricin
for longer than 6 days usually recovered from the toxic effects
within 1 or 2 weeks.
Antitumor Activity of in Vivo Treatment of Established
H2981-T3 Tumors with Intratumoral L6-Ricin and Anti-CD5Ricin plus Lactose. To determine antitumor efficacy, nude mice
bearing H2981-T3 tumors with a size index of 0.3-0.6 cm3
were randomly allocated to different treatment groups, after
which 5 or 10 /¿gof L6-ricin IT in 100 n\ of 300 HIM lactose
Table 5 Effect ofL6-ricin on normal bone marrow progenitor cells
Human bone marrow was aspirated, enriched for bone marrow light density
cells, pretreated with L6-ricin plus lactose for 2 h, washed, and cultured for
progenitor cell growth. CPU-mix, CFU-GM, and BFU-E colonies were micro
scopically counted from the same plated bone marrow samples after 14 days in
culture. Results are expressed as mean ±SE from triplicate samples.
Colonies/106 cells plated
TreatmentExpl.
1
Lactose control
L6-ricin. 500 ng/ml
ng/mlExpt.
L6-ricin. 1000
±20
33 ±17
101
22 ±
80633
±
±67
1731289
489 ±
±17
267 ±37
±47711
144
2
Lactose control
11 ±40
±140
±40577
ND°
L6-ricin. 500 ng/ml
800 ±200
±93
L6-ricin.
1000
ng/mlCFU-mix22
1000
+
20CFU-GM400
160"
67 ±33BFU-E489
489 ±
ND, not determined.
(4/4>
(2/4)
Days Atter Treatment
Fig. 2. Growth of s.c. H2981-T3 tumors following intratumoral administra
tion of ricin in nude mice. Groups of 3-6 animals with palpable tumors of a size
index between 0.3 and 0.5 cm' were given injections of various doses of ricin.
Tumor growth was monitored twice weekly over 4 weeks after treatment. The
data points represent means and SE for each treatment group: lactose (B); ricin,
0.2 (jg (•);ricin, 0.4 ^g (O); ricin. 0.6 /jg (•).The data in parentheses are the
number of survivors after 30 days over the total number of treated animals in
each group. Of the 6 animals treated with 1.0 >igricin, none survived longer than
6 days.
were injected into each tumor. Five min prior to the application
of IT, the animals received 500 n\ of 750 HIM lactose i.v. The
same dose of lactose alone was given to the control mice. Fig.
3A shows a representative experiment. Both of the two groups
treated with 5 or 10 ng of L6-ricin showed a marked inhibition
of tumor growth, in comparison to the lactose controls. This
effect was dose dependent, since 10 ug of L6-ricin caused more
tumor inhibition than 5 fig. The difference in tumor size be
tween either IT group and lactose controls was significantly
different in a Student's t test on day 30 (10 MgL6-ricin versus
controls, P < 0.001; 5 /¿gL6-ricin versus controls, P < 0.001).
Of the six animals treated with 5 ug L6-ricin, reexamination
on day 45 showed stabilization in one case, relapse in three
cases, and two cures. Of the four survivors treated with 10 ug
L6-ricin, two cures and two relapses were apparent upon reex
amination on day 45. The administration of 50 ug unmodified
MoAb L6 had no effect on established tumors.
To evaluate the specificity of L6-ricin treatment, we deter
mined the effect of anti-CD5-ricin IT on H2981-T3 tumors,
since anti-CD5-ricin did not bind to H2981-T3 cells according
to FACS analysis. Fig. 35 shows the growth curves of tumors
treated with 5 or 10 ug of anti-CD5-ricin, in comparison to
lactose controls. Anti-CD5-ricin had a dose-dependent inhibi-
3253
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-RICIN ANTITUMOR
Control
(4/4)
S
en
o
B. Control IT
Control
Bug
10ug
(4/4)
(3/4)
(3/7)
Days After Treatment
Fig. 3. Growth of H2981-T3 tumors following intratumoral administration of
L6-ricin (A) and anti-CDS-ricin (A) in nude mice. Groups of 4-7 animals bearing
H2981-T3 tumors between 0.3 and 0.5 cm3 were given injections of 0 (0), 5 (•),
or 10 Mg(O) IT in 300 HIMlactose. Animals received 500 M' of 750 m\i lactose
i.v. 5 min prior to the application of IT. Tumor size is plotted against the day of
measurement: the data points represent the means and SE in each treatment
group. The data in the parentheses are the number of survivors after 30 days over
the total number of treated animals in each group.
tory effect on the growth of H2981-T3 tumors, as did L6-ricin.
Comparison of the anti-CD5-ricin group to lactose controls on
day 30 after treatment showed a significant difference in tumor
size (10 ng anti-CD5-ricin versus controls, P < 0.03; 5 /¿gantiCD5-ricin versus controls, P < 0.017).
Our pooled data from three experiments are shown in Table
6. At a dose of 5 fig, L6-ricin induced comparatively more
ACTIVITY
complete regressions (2 of the 9 survivors) than anti-CD5-ricin
(0 of 8) or anti-CD3-ricin (0 of 8). However, if overall responses
are considered (complete responses and partial responses com
bined), antitumor effects were similar for all three IT.
The data in Table 6 also address toxicity. At a dose of 10 /¿g,
animals which received L6-ricin had a better survival (10 of 13)
than the anti-CD5-ricin (5 of 10) or the anti-CD3-ricin (1 of
12) controls. At a dose of 5 ¿ig,survival was similar for all IT.
Transient wasting was not observed in any 5 ¿/g-treatedmice
but was observed in some 10 ¿¿g-treated
mice. Wasting was seen
more often in animals treated with anti-CD5-ricin than with
L6-ricin.
In Vivo Treatment of Established H2981-T3 Tumors with
Intratumoral L6-Ricin and Anti-CD5-Ricin without Lactose.
Treatment of H2981-T3 tumors with 10 ^g of L6-ricin without
lactose resulted in pronounced antitumor effects in 4 of 4 mice.
L6-ricin induced complete tumor regression in 3 of 4 animals.
Whereas all animals treated with L6-ricin survived, all mice
treated with anti-CD5-ricin died of acute ricin toxicity within
7 days after treatment. These data indicate that lactose plays a
role in preventing ricin toxicity in vivo.
Tumor Morphology. H2981-T3 tumors grew as rigid fibrotic
nodules over the flanks of nude mice. One day after treatment
with IT, the consistency and outline of the tumors softened. By
day 3 after treatment, the tumors appeared to be bruised and
greenish in color. By day 4-5 these areas turned black and
necrotic. The tumors appeared flattened. Most of the observed
relapses were derived from tumor tissue surrounding the ne
crotic areas. Initial complete tumor necrosis resulted in a com
plete tumor remission. Only a white fibrotic scar remained after
40-80 days.
Reexpression of L6 Antigen in a Relapsing Tumor. A relapsing
H2981-T3 tumor which had initially regressed following L6ricin treatment was removed and trypsinized. The tumor cells
obtained from the trypsin suspension were assayed for the
presence of L6 antigen by FACS. A total of 62.4% of the cells
were L6 positive. As a positive control, cells from an untreated
H2981-T3 tumor were 44.0% L6 positive.
DISCUSSION
The antigen defined by L6 MoAb is expressed on the surface
of neoplastic lung, breast, colon, and ovarian tissue. L6-ricin
IT may be used for in vitro purging of residual tumor cells from
the bone marrow of carcinoma patients with marrow involve
ment, prior to autologous bone marrow transplantation. This
technique is used to allow more aggressive radiochemotherapy
for high risk patients, followed by a rescue dose of their own
marrow. L6-ricin may prove useful for this purpose, since our
studies showed that, at a concentration (1 ^g/m\) that was
Table 6 Survival and antitumor response after treatment of established H298I-T3 tumors with ricin IT
H2981-T3 tumors established in the flanks of nude mice were treated with various doses of IT in the presence of lactose. The table shows the number of animals
treated with IT, the survivors after 30 days, and the number of survivors who had complete responses (no tumor detectable at day 30), partial responses (tumor size
< 50% of control on day 30). or tumor progression (tumor size > 50% of control). Data are pooled from three experiments. Control mice treated with anti-CD3-ricin
and anti-CDS-ricin were evaluated in different experiments.
animalsDose
Number of
responsesComplete
response0200310Partial
Immunotoxin0
(/ig)
response0668631Progression13120110
None5
L6-ricinAnti-CDS-ricinAnti-CD3-ricin10
L6-ricinAnti-CD5-ricinAnti-CD3-ricinTreated13998131012Survivors1098810S1Antitumor
3254
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-RICIN ANTITUMOR
sufficient to inhibit about 99.99% of clonogenic growth, there
was still ample growth of normal bone marrow progenitors. It
should be noted that the 37-64% reduction of CFU-GM could
be due to expression of the L6 antigen on accessory cells that
promote CFU-GM growth in vitro or due to Fc-mediated L6
ricin toxicity.
Gel analysis showed that both L6-ricin and anti-CD5-ricin
displayed four major bands in excess of M, 150,000. Scanning
showed that the most abundant band was at M, 220,000, likely
representing one ricin molecule linked to an antibody molecule.
However, a small amount (0.1-6.0%) of M, 260,000 species,
likely representing two ricins linked to one antibody, was also
present. Contaminating free antibody and ricin were minimal.
L6 bound with an affinity of 1.6 x IO8 M~' to the adenocarcinoma cell line H2981-T3. An average of 1.7 x IO5 receptors/
cell were expressed. Similar values have been reported for L6
MoAb on Ch-27 cells (17). Ricin binding studies showed that
ricin bound similarly to H2981-T3 and CEM cells. Thus,
differences obtained with the two cell lines could not be attrib
uted to ricin binding differences.
Binding of L6-ricin to H2981-T3 cells incubated with lactose
was specific, as shown by immunofluorescence. In contrast,
inhibition of protein synthesis by L6-ricin in the presence of
lactose was not entirely selective, and blocking studies with
unconjugated L6 and human IgG showed that both a specific
and a nonspecific component of L6-ricin were active. Approx
imately 5 times greater specific than nonspecific activity was
apparent. An irrelevant anti-CD5 IT showed a 10-fold higher
IC50 value for inhibition of protein synthesis of H2981-T3 than
did L6-ricin, even in the presence of lactose. The inhibition of
the irrelevant IT could be entirely blocked by human IgG. Taken
together, these findings indicate that some part of the antibody
moiety other than the variable region, probably the Fc region,
can also mediate IT activity. Others have reported nonspecific
binding of abrin IT on human melanoma cells that did not
involve the galactose binding site of the B-chain (37). However,
in contrast to our findings, preincubation with human 7-globulin had no effect in their studies, and binding was dependent
on the cell line used.
We observed low modulation of L6 antigen on H2981-T3
cells, as compared to the modulation of CDS on CEM and
PBMC. Since the translocation of ricin A-chain into the cytosol
is a prerequisite for killing, perhaps the antitumor efficacy of
L6-ricin is attributed to the translocation-enhancing capabilities
of ricin B-chain (15). Notably, L6 IT made with A-chain devoid
of B-chain were not cytotoxic (data not shown).
There are several reasons for evaluating IT in a lung adenocarcinoma growing in nude mice, (a) Intratumoral injection
obviates two major problems associated with IT therapy: limited
tumor localization (38) and rapid elimination of IT from the
bloodstream by the liver and the spleen (11, 39, 40). (b) IT
constructed with intact ricin can be used, which has higher
cytotoxicity but lower specificity than A-chain IT (8, 15, 41).
(c) Intratumoral injections may be used as an adjunct therapy
for treating carcinomas localized to particular areas, e.g., ovar
ian carcinomas in the peritoneal cavity.
A tumoricidal effect was observed with intratumoral injec
tions of L6-ricin IT, corresponding to our in vitro findings.
Regressions were noted after a single injection of 5 or 10 ng of
IT. The antitumor effect in this system was more impressive
than in our previously published work using anti-CD5-ricin IT
with CEM lymphomas (8), since there was a higher incidence
of cures with tumors of similar size. However, the L6 model
showed a higher incidence of tumor regressions induced by
treatment with nonspecific anti-CD5-ricin IT. Perhaps Fc bind
ACTIVITY
ing caused indiscriminate tumor kill. Alternatively, H2981-T3
carcinomas are not as well vascularized as the soft CEM lym
phomas, the former being whitish and the latter being reddish
tumors. The carcinomas may not be as easily saturated after
i.v. injection of lactose.
Although selectivity in tumor regression was not absolute,
the systemic administration of lactose is important in our
model. In the absence of lactose prior to IT treatment, all
animals treated with 10 /ug of anti-CD5-ricin died. In contrast,
only 50% of the mice died who had received lactose prior to
anti-CD5-ricin. This indicates that lactose reduces the toxicity
of the irrelevant IT. The toxicity of L6-ricin was not affected
by lactose, suggesting that the L6 determinant kept the IT
localized to the tumor.
We also studied the effects of ricin alone on tumor growth.
Efficacious treatment of neoplasms in humans (42) and in
animal models with unconjugated ricin has been reported (43).
We were able to identify a dose of ricin (0.4 ¿¿g)
which could
induce a significant antitumor effect without significant toxic
ity. However, the effect was partial, since relapse occurred.
Higher concentrations were toxic, while lower concentrations
were not effective at all. In the CEM system, lower concentra
tions (0.3 ¿¿g)
of ricin were lethal without a detectable tumori
cidal effect (8). The different effects in the two systems cannot
be explained by a difference in numbers of galactose receptors,
since Scatchard analysis of ricin binding showed only slight
differences. There may be differences in tumor vascularization
and, therefore, different clearances of ricin in CEM and H2981T3.
A comparison of the toxicity caused by the relevant (L6ricin) or irrelevant (anti-CD5-ricin or anti-CD3-ricin) IT at a
dose of 10 i/g reveals a higher systemic toxicity for the irrelevant
IT (8% survivors for anti-CD3-ricin and 50% survivors for antiCD5-ricin) than for L6-ricin (77% survivors). This may be
explained by the ability of L6-ricin to bind to the tumor, while
the control IT displayed nonspecific binding. The higher tox
icity seen with 10 ¿iganti-CD3-ricin in comparison to antiCD5-ricin might be due to a difference in subclasses of the two
antibodies and differences in Fc binding properties. The treat
ment of tumors with L6-ricin did not lead to a selection of
antigen-negative cells in relapsing tumors, as could be shown
by staining a relapsed tumor with L6 MoAb. Thus, it may be
possible to increase the therapeutic efficacy of L6-ricin by
repeated applications of the IT.
In conclusion, we observed that L6-ricin is a relatively selec
tive IT with a high cytotoxic potency. Thus, L6-ricin may be
useful for the in vitro removal of residual tumor cells from
autologous bone marrow for patients with solid tumors. Com
parisons of the therapeutic and toxic effects of intratumoral
treatment using L6-ricin, the irrelevant IT, and free ricin show
that the toxicity of L6-ricin in this system can be mediated by
at least three different sites on the molecule: the specific anti
body binding site, a nonspecific antibody binding site (probably
Fc), and the ricin B-chain. The specific and nonspecific binding
of ricin IT can contribute to the overall therapeutic effectiveness
of intratumoral treatment, a situation that might be advanta
geous from interstitial tumor treatment of carcinomas but un
desirable for systemic treatment. Clearly, the lack of absolute
specificity in vivo remains a concern. For the systemic use of
L6 IT, it may be necessary to synthesize more selective agents,
such as an IT containing ricin with a manipulated B-chain (4446) or Fab fragments coupled to toxin.
ACKNOWLEDGMENTS
We are grateful to Hybritech, Inc. (La Jolla, CA) for T101 and to
Dr. Peter Beverley for LJCHT-1. We gratefully acknowledge the receipt
3255
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
L6-R1C1N ANTITUMOR
of L6 antibody and tumor cells from Drs. Ingegerd and Karl Erik
Hellström and ONCOGEN (Seattle, WA). We thank M. J. Hildreth
for exceptional editorial assistance.
23.
REFERENCES
24.
1. Vitella, E. S.. Fulton. R. J„May. R. D., Till, M., and Uhr, J. W. Redesigning
nature's poisons to create anti-tumor reagents. Science (Wash. DC), 238:
1098-1104. 1987.
2. Youle. R. J.. and Neville. D. M.. Jr. Anti-Thy-1.2 monoclonal antibody
linked to ricin is a potent cell type specific toxin. Proc. Nati. Acad. Sci. USA,
77:5483-5486, 1980.
3. Gregg, E. O., Bridges, S. H., Youle, R. J., Longo. D. L., Houston, L. L..
Glennie, M. J., Stevenson, F. K., and Green, I. Whole ricin recombinant
ricin A chain idiotype-specific immunotoxins for therapy of the guinea pig
L2C B cell leukemia. J. Immunol., 15:4502-4508, 1987.
4. Leonard, J. E., Wang, Q. C, Kaplan, N. O., and Royston, I. Kinetics of
protein synthesis inactivation in human T-lymphocytes by selective monoclo
nal antibody-ricin conjugates. Cancer Res.. 45: 5263-5269. 1985.
5. Vallera, D. A., Ash, R. C., Zanjani, E. D.. Kersey, J. H.. KeBien, T. W..
Beverley, P. C. L., Neville, D. M., Jr., and Youle, R. J. Anti-T cell reagents
for human bone marrow transplantation: ricin linked to three monoclonal
antibodies. Science (Wash. DC). 222: 8512-8515, 1983.
6. Filipovich, A. H., Vallera. D. A., Youle, R. J., Haake, R.. Blazar. B. R..
Neville, D. M., Jr., Ramsay, N. K. C., McGlave, P.. and Kersey, J. H. Graftvers«.c-host
disease prevention in allogeneic bone marrow transplantation. A
pilot study using immunotoxins for T cell depletion in donor bone marrow.
Transplantation. 44:62-69. 1987.
7. Uckun, F. M., Myers. D. E.. Kersey, J. H.. Filipovich, A. H.. Ramsay, N. K.
C., McGlave. P., Dickson. T.. and Vallera. D. A. Immunotoxins in bone
marrow transplantation: an updated review of the Minnesota experience. In:
B. Bonavida and R. J. Collier (cds.), Membrane-Mediated Cytotoxicity, pp.
231-242. New York: Alan R. Liss, Inc., 1987.
8. Weil-Hillman, G.. Runge, W., Jansen, F. K., and Vallera, D. A. Cytotoxic
effect of unti-M, 67.000 protein immunotoxins on human tumors in a nude
mouse model. Cancer Res.. 45: 1328-1336. 1985.
9. Hara, H., and Seon, B. K. Complete suppression of in vivo growth of human
leukemia cells by specific immunotoxins: nude mouse models. Proc. Nati.
Acad. Sci. USA, 84: 3390-3391, 1987.
10. Weil-Hillman, G., Uckun. F. M.. Manse. J. M.. and Vallera. D. A. Combined
immunochemotherapy of human solid tumors in nude mice. Cancer Res.,
47:579-585, 1987.
11. Manske, J. M., Buchsbaum, D. J., Hanna, D. E., and Vallera, D. A. Cytotoxic
effects of anti-CD5 radioimmunotoxins on human tumors in vitro and in a
nude mouse model. Cancer Res.. 48: 7107-7114. 1988.
12. Buchsbaum. D. J.. Nelson. L. A., Hanna, D. E., and Vallera, D. A. Human
leukemia cell binding and killing by anti-CD5 radioimmunotoxins. Int. J.
Radial. Oncol. Biol. Phys., 13: 1701-1712, 1987.
13. Endo. Y., Mitsui. K., Motizuki. M.. and Tsurugi. K. The mechanisms of
action of ricin and related toxic lectins on eukaryotic ribosomes. J. Biol.
Chem., 262: 5908-5912, 1987.
14. Youle, R. J., and Neville, D. M., Jr. Kinetics of protein synthesis inactivation
by ricin anti-Thy-l.l monoclonal antibody hybrids. J. Biol. Chem., 257:
1598-1601. 1982.
15. Manske. J. M.. Buchsbaum, D. J., and Vallera. D. A. The role of ricin B
chain in the ¡ntracellulartrafficking of anti-CD5 immunotoxins. J. Immunol.,
142: 1755-1766, 1989.
16. Youle. R. J., Murray, G. J., and Neville, D. M., Jr. Studies on the galactosebinding site of ricin and hybrid toxin Man6P-ricin. Cell, 23: 551 -559, 1981.
17. Hellstrom. I.. Horn. D.. Linsley. P.. Brown, J. P., Brankovan, V.. and
Hellstrom. K. E. Monoclonal mouse antibodies raised against human lung
carcinoma. Cancer Res., 46: 3917-3923. 1986.
18. Hellstrom, L, Beaumier, P. L., and Hellstrom, K. E. Antitumor effects of
L6, an IgG2a antibody that reacts with most human carcinomas. Proc. Nati.
Acad. Sci. USA, 83: 7059-7063, 1986.
19. Coombes. R. C., Buckman. R., Forrester, J. A., Shepherd, V.. O'Hare, M.
J., Vincent, M., Powles, T. J., and Neville, A. M. In vitro and in vivo effects
of a monoclonal antibody-toxin conjugate for use in autologous bone marrow
transplantation for patients with breast cancer. Cancer Res., 46:4217-4220,
1986.
20. Mabry, M.. Speak, J. A., Griffin, J. D., Stahel, R. A., and Bernal, S. D. Use
of SM-1 monoclonal antibody and human complement in selective killing of
small cell carcinoma of the lung. J. Clin. Invest., 75: 1690-1695, 1985.
21. Anderson. I., Stipali, E. J., Leslie. D., Daly, L., Nustad, K., Ugelstad, J..
Peters, W.. and Bast, R. Elimination of malignant clonogenic breast cancer
cells with immunomagnetic separation and 4-hydroperoxycyclophosphamide
(4-HC). Proc. Am. Assoc. Cancer Res., 29: 183, 1988.
22. Kemshead, J., Heath, L.. Gibson, F. J., Katz, F., Richmond. F.. Treleaven,
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
ACTIVITY
J.. and Ugelstad, J. Magnetic microspheres and monoclonal antibodies for
the depletion of neuroblastoma cells from bone marrow: experiences, im
provements and observations. Br. J. Cancer. 54: 771-778. 1986.
Zovickian, J.. and Youle, R. J. Efficacy of intrathecal immunotoxin therapy
in an animal model of leptomeningeal neoplasia. J. Neurosurg., 68: 767774, 1988.
Perez. C. A., and Glasgow. G. P. Clinical applications of brachytherapy. In:
C. A. Perez and L. W. Brady (eds.). Principles and Practice of Radiation
Oncology, pp. 252-298. Philadelphia: J. B. Lippincott Company, 1987.
Royston, I., Majda, J. A., Baird, S. M., Meserve. B. L.. and Griffiths, J. C.
Human T cell antigens defined by monoclonal antibodies: the 65.000-dalton
antigen of T cells (T65) is also found on chronic lymphocytic leukemia cells
bearing surface immunoglobulin. J. Immunol.. 125: 725-731. 1980.
Beverly, P. C. L., and Callard, R. C. Distinctive functional characteristics of
human T lymphocytes defined by E rosetting or a monoclonal anti-T cell
antibody. Eur. J. Immunol.. //: 329-333. 1981.
Manske, J. M.. Buchsbaum. D. J.. Azemove, S. M., Hanna. D. E.. and
Vallera, D. A. Antigenic modulation by anti-CD5 immunotoxins. J. Immu
nol., 136: 4721-4728, 1986.
Doran. D. M.. and Spar, I. L. Oxidative iodine monochloride iodination
technique. J. Immunol. Methods. 39: 155-163. 1980.
Titus, J. A., Sharrow, S. O., Connolly, J. M., and Segal, D. M. Fc (IgG)
receptor distributions in homogeneous and heterogeneous cell populations
by How microfluorometry. Proc. Nati. Acad. Sci. USA, 78: 519-523, 1981.
Liberti, P. A., Hackett, C. J., and Askonas. B. A. Influenza virus infection of
mouse lymphoblasts alters the binding affinity of anti-H-2 antibody: require
ment for viral neuraminidase. Eur. J. Immunol.. 9: 751-757. 1979.
Townsend, A. R. M., McMichael, A. J.. Carter, N. P., Huddlcston, J. A.,
and Brownlee. G. G. Cytotoxic T cell recognition of the influenza nucleoprotein and hemagglutinin expressed in transfected mouse L cells. Cell, 39: 1325, 1984.
Leonard, J. E.. Grothaus. C. D.. and Taetle, R. Ricin binding and protein
synthesis inhibition in human hematopoietic cell lines. Blood, 72: 13571363,1988.
Uckun, F. M., Stong, R. C., Youle, R. J., and Vallera, D. A. Combined ex
vivo treatment with immunotoxins and mafosfamid: a novel ¡mmunochemotherapeutic approach for elimination of neoplastic T cells from autologous
marrow grafts. J. Immunol., 134: 3504-3515. 1985.
Ash, R. C., Detrick, R. A., and Zanjani, E. D. Studies of human pluripotential
hemopoietic stem cells (CFU-GEMM) in vitro. Blood, 5«:309-316, 1981.
Schroff. R. W., Farrell. M. M., Klein. R. A.. Oldham. R. K.. and Foon. K.
A. T65 antigen modulation in a phase I monoclonal antibody trial with
chronic lymphocytic leukemia patients. J. Immunol.. 133: 1641-1648. 1984.
Stong, R. C.. Youle. R. J., and Vallera. D. A. Elimination of clonogenic Tleukemic cells from human bone marrow using anti-A/, 65,000 protein
immunotoxins. Cancer Res.. 44: 3000-3006. 1984.
Godal. A.. Fodstad. O.. and Pihl. A. Studies on the mechanism of action of
abrin-9.2.27 immunotoxin in human melanoma cell lines. Cancer Res.. 47:
6243-6247. 1987.
Pimm, M. V. Drug-monoclonal antibody conjugates for cancer therapy:
potentials and limitations. Crit. Rev. Ther. Drug Carrier Syst.. 5: 189-227.
1988.
Bourrie. B. J. P., Casellas. P.. Blythman, H. E.. and Jansen. F. K. Study of
the plasma clearance of antibody—ricin-A-chain immunotoxins: evidence for
specific recognition sites on the A chain that mediate rapid clearance of the
immunotoxin. Eur. J. Biochem., 155: 1-10. 1986.
Blakey. D. C., Watson. G. J.. Knowles. P. P.. and Thorpe. P. E. Effect of
chemical deglycosylation of ricin A chain on the in vivo fate and Cytotoxic
activity of an immunotoxin composed of ricin A chain and anti-Thy 1.1
antibody. Cancer Res.. 47: 947-952, 1987.
Vallera, D. A., Quiñones.R. R.. Azemove, S. M., and Soderling, C. C. B.
Monoclonal antibody toxin conjugates reactive against human T lympho
cytes: a comparison of antibody linked to intact ricin toxin and antibody
linked to ricin A chain. Transplantation. 37: 387-392. 1984.
Fodstad, O., Kvalheim. G., Godal, A.. Lotsverg, J., Aamdal, S., Host. H..
and Pihl. A. Phase I study of the plant protein ricin. Cancer Res., 44: 862865, 1984.
Fodstad, O., Olsnes, S.. and Pihl, A. Inhibitory effect of abrin and ricin on
the growth of transplantable murine tumors and of abrin on human cancers
in nude mice. Cancer Res.. 37: 4559-4567, 1977.
Thorpe, P. E., Ross, W. C. J., Brown. A. N. F., Myers, C. D., Cumber, A.
J.. Foxwell. B. M. J., and Forrester. J. T. Blockade of the galactose-binding
sites of ricin by its linkage to antibody. Eur. J. Biochem., 140: 63-71, 1984.
Wawrzynczak, E. J., Drake, A. F., Watson, G. J., Thrope, P. E.. and Vitetta,
E. S. Ricin B chain-containing immunotoxins prepared with heat-denatured
B chain lacking galactose-binding ability potentiate the cytotoxicity of a cellreactive ricin A chain immunotoxin. Biochim. Biophys. Acta, 971: 55-62,
1988.
Pietersz, G. A., Kanellos, J.. and McKenzie, I. F. C. Novel synthesis and in
vitro characterization of disulfide-linked ricin-monoclonal antibody conju
gates devoid of galactose binding activity. Cancer Res., 48:4469-4476. 1988.
3256
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1990 American Association for Cancer Research.
Antitumor Activity of L6-Ricin Immunotoxin against the
H2981-T3 Lung Adenocarcinoma Cell Line in Vitro and in Vivo
Heinz Schmidberger, Laurel King, Larry C. Lasky, et al.
Cancer Res 1990;50:3249-3256.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/50/11/3249
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. © 1990 American Association for Cancer Research.