| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Lexigen Pharmaceuticals Corp., Lexington, MA 02137; and
Department of Immunology, The Scripps Research Institute, La Jolla, CA 92037
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, but were much less active on mouse spleen cells
than a mouse IL-12 standard. Despite this relatively low activity, the
hybrid fusion protein was as effective in a SCID mouse model as a fully
active Ab-IL-2 fusion protein in eliminating established pulmonary
metastases of CT26 colon carcinoma. Specific targeting of a human IL-12
fusion protein to metastatic prostate carcinoma xenografts was also
shown to be effective in SCID mice transplanted with human
lymphocyte-activated killer cells. These results demonstrate the
importance of directing this potent cytokine to the tumor
microenvironment and suggest an important alternative to systemic IL-12
administration or gene therapy for increasing its therapeutic index. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Another cytokine with potent antitumor activity in several mouse tumor
models is IL-12. This heterodimeric molecule appears to be more potent
in many systems and has both unique and overlapping activities with
IL-2. For example, both cytokines are potent stimulators of NK and
CD8+ CTL, activities for which IL-12 was originally named
(8, 9). IL-12 is a much more potent stimulator of IFN- from both
cell types and of Th1 cell differentiation (10) than IL-2, and together
they show strong synergistic effects (11).
Based on our earlier success with targeting IL-2 to solid tumors of epithelial origin, we attempted to apply this same approach to the delivery of IL-12 to tumors of the prostate and colon. In the first case, we have developed a metastatic model of prostate carcinoma based on the DU-145 carcinoma line growing in SCID mice. Treatment in this case requires the engraftment of human LAK cells as effectors, since the human IL-12 molecule has no activity on the endogenous murine immune cells. Any activity in this model should be due to the transplanted human NK cell population.
In a second model, we have genetically engineered the mouse colon carcinoma, CT26, to express the human EpCAM molecule and have attempted treatment in the syngeneic BALB/c mouse. In this case, we prepared a fusion protein with the mouse p35 subunit for activity in mouse cells; however, the IL-12 activity of the resulting fusion protein was significantly less than that of mouse IL-12. Despite this lower IL-12 activity, the fusion protein, comprised mostly of human Ig sequences, proved to be too immunogenic for studies in immune-competent BALB/c mice. Instead, we tested its activity in SCID mice transplanted with pulmonary metastases, in the absence of exogenous LAK cells. Results show the IL-12 fusion proteins to be effective at treating established tumors of both human prostate carcinoma and mouse colon carcinoma in SCID mice, despite the lack of functional T cells. This is particularly meaningful for the CT26 model, which has proven to be one of the most refractory to systemic IL-12 and gene therapy approaches (12, 13).
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Humanized V regions of the mouse KS-1/4 Ab (14) were modeled using the methods developed by Winter and coworkers (15), which involved the insertion of the complementarity-determining regions of each KS-1/4 V region into the consensus framework sequences of the human V regions with the highest degree of homology. Molecular modeling with a Silicon Graphics Indigo work station and BioSym software confirmed that the shapes of the complementarity-determining regions were maintained between the mouse and humanized forms. The final protein sequences were reverse translated, and the genes were constructed by the ligation of multiple overlapping oligonucleotides.
The V regions were inserted into an expression vector containing the C
regions of the human 
L chain and the human C1
or C3 H chains, essentially as described (16), except
that the metallothionine promoters and IgH chain enhancers were
replaced by the CMV promoter/enhancer for the expression of both
chains. Fusions of the mature sequences of either IL-2 or the p35
subunit of IL-12 to the carboxyl termini of the human H chains were
made as described earlier (16), except that the 3' untranslated regions
of the cytokine genes were derived from the SV40 poly(A) region. Fusion
of p35 to the Fc fragment-encoding region was exactly the same as the
intact Ab fusion, but the normal IgG H chain was replaced by a
truncated C1 Fc-encoding DNA fragment ligated to a
leader sequence from a kappa L chain (17). IL-2 fusion proteins were
expressed by transfection of the plasmids into the NS/0 myeloma cell
line with selection in medium containing methotrexate (0.1 µM). The
IL-12 fusion proteins were expressed by transfection of Ab-p35 subunit
fusion constructs into an NS/0 transfectant already expressing high
levels of the p40 subunit of human IL-12, followed by selection in
methotrexate-containing medium. Expressing clones were identified by Fc
or cytokine ELISA (18), and fusion proteins were purified by binding to
and elution from native (Pharmacia, Piscataway, NJ) or recombinant
(Repligen, Needham, MA) protein A-Sepharose.
A second type of Ab-IL-12 fusion protein was expressed by fusing sequences of a single-chain Fv (scFv) of the humanized KS-1/4 to the p35 subunit, followed by cotransfection with the separate p40 expression vector into human 293 cells. The scFv was made by linking the humanized KS-1/4 L chain V region cDNA to the H chain V region with a flexible (Gly4 Ser)3 peptide, as described by others (19). Plasmid DNAs (5 µg of each) were mixed, coprecipitated with calcium phosphate, and added to monolayers of 293 cells. The medium and remaining DNA precipitate was removed after 16 h and replaced with fresh culture medium. Three to four days later, the culture supernatant was tested for secreted fusion protein by ELISA and analyzed for IL-12 bioactivity. The scFv-IL-12 fusion protein was purified by binding to and elution from an EpCAM-Sepharose column. Recombinant human EpCAM ectodomain, fused to the carboxyl terminus of the mouse Fc (EpCAM-Fc) was coupled to cyanogen bromide-activated Sepharose 4B (Pharmacia) by standard methods supplied by the manufacturer.
Cloning and expression of cell surface EpCAM
The cloning of the Ag recognized by the KS-1/4 Ab has been reported by several groups (20, 21). The known sequence was used to generate PCR primers for cloning the complete protein-encoding sequence. After confirmation of the DNA sequence, the human EpCAM cDNA was inserted in the retroviral expression plasmid, pLNCX (22), as a NotI to XhoI fragment, and the resulting plasmid DNA was transfected into the PA317 amphotropic packaging cell line. Progeny virus was harvested 16 h later and used to infect CT26 mouse colon carcinoma cells (23). Clones were selected in media containing G-418 (1 mg/ml; Life Technologies, Grand Island, NY). Clones expressing EpCAM were identified by immunostaining with humanized KS-1/4 Ab, followed by a goat FITC-conjugated anti-human Ab. Some lines were subcloned further by limiting dilution and retested for cell surface expression of EpCAM and MHC class I molecules. A subclone stably expressing EpCAM, but low levels of MHC class I, CT26-Ep21.6, was used for the studies described in this work.
Cytokine bioassays
The activity of IL-2 was measured in the CTLL-2 assay (24).
Briefly, CTLL-2 cells were starved for 6 h by incubation in
culture medium without IL-2, after which dilutions of the test samples
were added to duplicate wells together with 1 x
105 CTLL-2 cells/well. Twenty-four hours later,
proliferation was measured by pulsing cultures with
[3H]thymidine (1 µCi/ml) for 16 h. IL-12 activity
was tested in two types of assays measuring either cell proliferation
or induction of IFN- (25). In the proliferation assay, resting human
PBMC or mouse spleen cells were cultured with mitogen for 3 days (PHA
or Con A, respectively) and then added to 96-well plates containing
dilutions of the IL-12 samples. Proliferation was measured by pulsing
with [3H]thymidine 48 h later. In the second assay,
human PBMC were activated by growth in medium containing PHA (5
µg/ml) for 3 days. Activated cells were washed by centrifugation to
remove cytokines and plated in new 96-well plates containing dilutions
of IL-12 fusion proteins. Twenty-four hours later, the concentration of
IFN- was measured by ELISA using Ab pairs supplied by Endogen
(Boston, MA).
Ab effector activities
Ab-dependent cellular cytotoxicity (ADCC) was tested as described (26) using 51Cr-labeled human PC-3 or DU-145 prostate carcinoma cells or EpCAM-transfected mouse CT26 cells. Specific lysis was measured after a 4-h incubation with the indicated number of resting human PBMC.
Pharmacokinetic analysis
Abs or fusion proteins were diluted to 250 µg/ml with PBS, and 100 µl (25 µg) was injected into the tail vein of BALB/c mice. Blood samples were taken at the indicated times by retroorbital bleeding, and the amount of circulating Ab or fusion protein was determined by Fc ELISA. Some mouse sera were analyzed further by specific ELISA for the cytokine fusion protein. In this case, the capture Ab was specific for human Fc, while the detection Ab was reactive with the human p40 subunit of IL-12 (PharMingen, San Diego, CA).
CT26-EpCAM experimental pulmonary metastasis model
CB-17 SCID mice were injected in the tail vein with 1 x 105 CT26-Ep21.6 cells in PBS. Once tumors were established at day 3, mice were treated with either PBS, Ab, or fusion protein for 5 consecutive days. On day 28, animals were sacrificed and lungs were removed, weighed, and stained with Bouin fixative for analysis of lung metastases. At this time, it was difficult to count individual lung metastases since they tended to fuse on the surface of the organ. Therefore, the percentage of surface coverage was used as a measure of antitumor effect, together with increases in lung weights.
DU-145 experimental pulmonary metastasis model
SCID mice were injected in the tail vein with 1 x 106 DU-145 cells and examined for lung metastases at different time points. Animals with evidence of tumor growth were used to establish explants from disrupted lung tissue. Cultures were observed for tumor cell growth, and the human origin of established lines was confirmed by immunostaining with the KS-1/4 Ab. Candidate lines were tested for their ability to form pulmonary metastatic foci after a single i.v. injection. One such line, designated DU-145 met, was found to produce from 200 to 500 metastatic foci in 3 to 4 wk after injection of 5 x 105 cells. Treatment of metastatic disease with Abs or fusion proteins was initiated on day 3 after injection of tumor cells and 6 h after the i.p. injection of human LAK cells. These cells were generated from resting human PBMC by incubation for 3 days in media containing human IL-2 (100 U/ml). Animals received Ab or fusion protein diluted in PBS in a volume of 200 µl, or PBS alone, for 6 consecutive days. After a total of 4 wk, mice were sacrificed, the lungs removed, weighed, and the number of metastases counted after first staining with Bouin fixative.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our initial studies utilized a simplified fusion protein system that did not require the coexpression of the Ab L chain. Instead, fusions of either of the two subunits of the heterodimeric IL-12 molecule were fused in-frame to the carboxyl terminus of the Fc fragment of human IgG1. Direct fusions between the terminal lysine and the mature protein sequences (devoid of their leaders) were used to avoid the introduction of potentially immunogenic epitopes. The individual subunits were also cloned into vectors for expression as separate proteins and for mixing in the various possible combinations with the fused forms.
Transient expression analysis in human 293 cells and affinity
purification on protein A-Sepharose are shown in Figure 1
. Gel analysis after reduction of the
purified proteins demonstrates that Fc-p40 dimers are expressed and
secreted, as are Fc-p40/Fc-p35 heterodimers containing approximately
equimolar amounts of each fusion protein. It is unlikely that these
products represent equal amounts of the individual homodimers since
Fc-p35 expression was not detected (lane 1).
Combinations of fusion proteins with individual subunits were also
expressed and secreted as heterodimers (lanes 4 and
5), but unlike the Fc-p35/Fc-p40 heterodimer, they
contain two molar equivalents of IL-12/Fc. Interestingly, the binding
of free p40 subunit to the Fc-p35 fusion protein allowed it to be
secreted from the cell (compare lanes 1 and
4), thus ensuring that only complete heterodimer is
produced. When protein A is used for purification, a homogeneous
product is obtained, devoid of any p40 or p40 dimer. The same strategy
of fusing IL-12 p35 and coexpressing the p40 subunit in the same cell
was applied to whole Ab constructs.
|
The murine Ab, KS-1/4, was humanized and expressed as an intact Ab or fused to either of the potent immune stimulators IL-2 or IL-12. Each of these cytokines can enhance NK and T cell growth as well as their cytotoxic activity. The humanization process did not cause a loss of Ab binding and, in fact, resulted in a twofold higher affinity, as determined by Scatchard analysis (not shown). The abilities of the humanized KS-1/4 (Hu-KS-1/4), as well as the IL-2 and IL-12 fusion proteins, to bind the EpCAM-transfected CT26 mouse carcinoma line were compared by FACS analysis with the original mouse KS-1/4 Ab and found to be equivalent (data not shown). Control CT26 cells were not bound by these molecules.
Expression of the IL-12 fusion protein involved the additional step of
coexpressing the p40 subunit in the same cell as an Ab-p35 subunit
fusion protein. Since the p35 subunit of IL-12 cannot be secreted in
the absence of the p40 subunit (see above), the Ab fusion protein
likewise cannot be secreted without p40 association to the fused p35
subunit. This is much like the inability of the Ab H chain to be
secreted in the absence of the L chain, and ensures that the Ab is
produced with the proper stoichiometry of H and L chains. When all
three chains are coexpressed (L, H-p35, and p40), cells secrete only
fully assembled Ab-IL-12 fusion protein with two complete molecules of
IL-12 per Ab molecule. Analysis on denaturing SDS gels under reducing
and nonreducing conditions shows that an intact molecule of 300 kDa is
purified from transfected cells by protein A binding. Following
reduction, the molecule breaks down into three species of the expected
molecular sizes and relative abundance (Fig. 2A). The minor band
seen above the major p40 band was found to react with anti-p40 Ab
by Western blotting analysis (Fig. 2B) and most
likely represents a form with an altered pattern of glycosylation. The
IL-2 fusion proteins were analyzed accordingly to confirm their
structural integrity.
|
The biologic activities of the cytokines in the Ab-cytokine fusion
proteins were also maintained as reported earlier for IL-2 and other
fusion protein constructs (16, 27). The activity of Hu-KS-IL-2 in the
CTLL-2 assay (data not shown) is typical of several other earlier
reports, and the ability of such molecules to induce antitumor
responses in mouse models is well established (4, 5, 6). We compared the
activity of Ab-fused and control human IL-12 in a PBMC proliferation
assay (Fig. 3A). There
was at most a twofold difference between the IL-12 standard and the
human Ab-IL-12 fusion protein, while the hybrid mouse/human IL-12
fusion protein exhibited more than 10-fold less activity than the human
standard. IL-12 activity of the Hu-KS-IL-12 fusion proteins was also
tested in an IFN- induction assay using mitogen-stimulated human
PBMC or mouse spleen cells. The intact Ab-IL-12 fusion protein
containing human p35 and p40 subunits was approximately 10-fold less
active in the induction of IFN- from human cells than human IL-12
standards (Fig. 3B). The hybrid fusion protein was
active in the same type of IFN- induction assay using mouse spleen
cells (Fig. 3D), albeit much less than the mouse
IL-12 control, while the human fusion protein had little or no
activity. This assay was more sensitive than that based on
proliferation, and revealed greater differences in activity among the
various free and fused molecules.
|
B, the scFV-IL-12
construct with both human subunits had about the same activity as the
whole Ab fusion protein in the human IFN- induction assays. The
scFv-IL-12 made with mouse p35 subunit was as active as the
corresponding whole Ab-IL-12 fusion protein in the mouse splenocyte
assay (Fig. 3D). Thus, there was not a significant
reduction in IL-12 activity that could be attributed to the dimeric
nature of the whole Ab fusion proteins.
The scFv-IL-12 molecules retained their Ab reactivity to EpCAM, as
demonstrated by binding to plate-bound EpCAM, expressed as an Fc fusion
protein (Fig. 3C). The relative binding was
normalized by detecting bound fusion protein with an anti-IL-12 Ab.
The monovalent scFv-IL-12 bound approximately threefold less to EpCAM
than the bivalent whole Ab-IL-12 molecule.
Effector activities
Tests of the effector functions of the Hu-KS-1/4 Ab and whole Ab
fusion proteins indicated that ADCC activity directed against human
PC-3 prostate carcinoma cells was greatly enhanced by the humanization
process (Fig. 4). The mouse Ab had little
activity, while the humanized form had an ED50 of
about 10 ng/ml. Dramatic increases in ADCC activity by Ab chimerization
or humanization have been reported previously (26). Both the IL-2 and
IL-12 fusion proteins retained this potent ADCC activity against
resting human PBMC. In fact, the IL-12 fusion protein showed enhanced
ADCC activity compared with the intact Ab or the IL-2 fusion protein, a
property that is consistent with its original name of NK
cell-stimulating factor. Similar results were obtained using human
DU-145 prostate carcinoma cells as targets (data not shown).
|
We examined the pharmacokinetic properties of the humanized KS-1/4
Ab and its cytokine fusion proteins following i.v. injection. As
reported earlier, Ab-IL-2 fusion proteins have a more rapid
redistribution (
) phase than normal Abs, but remain longer in the
circulation than free IL-2. This is followed by a relatively slow
catabolic (ß) phase (27). We have found that the amount of fusion
protein cleared from the circulation during the initial phase is to
some extent determined by the way it is formulated and may be a
reflection of a protein aggregation effect (data not shown).
Another group of investigators reported superior pharmacokinetic
behavior of an Ab-IL-2 fusion protein made with a human
C3, rather than a C1 H chain C
region (28). However, their study did not directly compare the two
isotypic forms, nor did they measure pharmacokinetics in a single
compartment model as we have. Instead, these investigators injected
radiolabeled fusion protein into the peritoneal cavity and measured
whole body elimination. We have generated both C1 and
C3 forms of Hu-KS-IL-2 and directly compared the
circulating t1/2 in the single compartment
model. As seen in Figure 5A,
both isotypes have similar behavior in the circulation. Since the Ab
isotype made no difference in pharmacokinetic properties, all of the
IL-12 fusion proteins were constructed using the original human
C1 region.
|
distribution phase (Fig. 5A), similar to what we
reported earlier for GM-CSF (27), resulting in significantly higher
sustained circulating concentrations, and an extended ß phase
t1/2. The human/mouse hybrid molecule was used
to measure clearance, since the all-human molecule is not active in
rodents, and thus, most likely does not bind IL-12R on mouse cells. We
confirmed that much of the material in the circulation represents
intact fusion protein by using a specific ELISA that involves capture
with an anti-Ig Ab and detection with an anti-p40 subunit Ab
(Fig. 5B). A portion of the IL-12 does appear to be
proteolytically cleaved from the Ab over time, as confirmed by Western
blotting (not shown), with free Ab returning to the circulation.
Nonetheless, the ß t1/2 of the intact
IL-12 fusion protein was 21 h.
The levels of serum IFN- were also tested in mice injected with the
IL-12 fusion proteins (Table I). Peak
levels of 400 to 500 pg/ml were seen at 48 h after injection of 20
µg of the Hu-KS-IL-12 fusion protein (
10 µg of IL-12) that were
significantly lower than what has been reported by others (29)
following daily doses of submicrogram amounts of IL-12. No signs of
overt toxicity were seen in mice injected with daily doses of the IL-12
fusion proteins in the efficacy studies, described below.
|
Unlike IL-2, IL-12 is highly species specific, and as such, presents problems for the testing of humanized proteins in mouse models. To determine the efficacy of the humanized IL-12 fusion protein, we used a SCID mouse transplantation model in which established human tumors are treated with engrafted human effector cells, administered into the peritoneal cavity, and the humanized fusion proteins. The effector cell population was derived from normal human PBMC after culturing for 3 days in medium containing high levels of IL-2. Both T cell and NK cell populations are expected to be activated by this procedure.
Established DU-145 tumors in the lungs of SCID mice were treated with a
combination of transplanted human effector cells and Ab-IL-12 fusion
proteins. As seen in Figure 6, the
Hu-KS-IL-12 fusion protein was far more effective in killing
established DU-145 lung metastases than a second, irrelevant Ab-IL-12
fusion protein with specificity for the ganglioside GD2 that is not
expressed on these cells. Since this Ab-IL-12 fusion protein has the
same circulating t1/2 and IL-12 sp. act. as the
Hu-KS-IL-12 fusion protein, it was used as a control to demonstrate the
importance of cytokine delivery to tumors in vivo. Although the
nonspecific Ab group had somewhat fewer metastases than the control
group, the results were not statistically significant.
|
are highly species specific, and thus, can only
activate the transplanted human effector cells. It is also likely that
NK cells play the major role in this process since they are more
potently activated by high doses of IL-2 than are T cells cultured in
the absence of their cognate Ag. Nonetheless, a role of human T cells
cannot be ruled out, perhaps based on CD8+ cell
alloreactivity with DU-145 cells. Potent antitumor effects mediated by
both NK and T cells have been reported in similar SCID mouse
transplantation models (30).
Another animal model was developed to test the efficacy of the
Ab-IL-12 fusion proteins, composed of hybrid mouse/human IL-12, that
can activate endogenous mouse effector cells. Mouse colon carcinoma
CT26 cells are highly tumorigenic in BALB/c mice (23), despite the fact
that they express significant levels of MHC class I molecules. After
transfection with the human EpCAM molecule, we isolated CT26 clones
expressing the same level of class I molecules as the parental cells,
as well as a subclone expressing very low levels of both
H2-Dd and H2-Kd. This low expressor subclone,
CT26-Ep21.6, can be induced to express normal levels of both class I
molecules following treatment with IFN- (data not shown), and thus,
closely resembles what is found in human carcinoma (31). Low levels of
MHC class I molecules should not decrease their susceptibility to
cytolysis by activated NK cells (in fact it may increase it), making
the SCID system especially valuable for assessing the role of these
cells in the absence of functional T cells. Both IL-2 and IL-12 can
activate NK cells and lead to the cytolysis of tumor cells in an MHC
class I-independent manner (32).
We tested the possible advantage of selectively targeting these
cytokines to the microenvironment of small, but established, colon
carcinoma metastases in SCID mice. Treatment began 3 days after tumor
injection (1 x 105 cells injected i.v.) and
establishment of lung metastases, and consisted of a single 5-day
course of injections with 10 µg of each fusion protein or control Ab
and cytokine. The benefit of specific tumor targeting was clearly
demonstrated, compared with either no treatment or the injection of Ab
plus nontargeted IL-2 controls (Fig. 7).
Both the human IL-2 and mouse/human hybrid IL-12 fusion proteins
demonstrated potent in vivo activity, although complete eradication of
metastases was not achieved using this dosing regimen. The roughly
equivalent antitumor activity of the two fusion proteins is especially
interesting when one considers that the IL-2 fusion protein has
approximately the same activity as native IL-2, while the hybrid
mouse/human protein used in these experiments was approximately 50-fold
less active than native mouse IL-12. Thus, the Ab-IL-12 fusion protein
appears to be a more potent activator of NK cells in this SCID mouse
model than the Ab-IL-2 fusion protein.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
from human
PBMC. The mouse/human hybrid forms, on the other hand, were far less
potent, especially with mouse effector cells, but the whole Ab fusion
protein was still quite effective in the SCID mouse tumor xenograft
model as a result of tumor targeting. The scFv version has not yet been
tested in the animal model.
The somewhat reduced activity of the Ab IL-12 molecules is
possibly due to the fusion of a relatively large cytokine molecule to
the carboxyl terminus of each Ig H chain. Such a configuration does not
affect the activity of other, smaller cytokines such as IL-2,
lymphotoxin, or TNF- (16, 27). Although we originally reported that
an Ab fusion protein with GM-CSF was somewhat less active compared with
a bacterially expressed standard (27), this reduced activity is well
known to be a result of glycosylation in the mammalian form of human
GM-CSF (33), and neither to the size nor configuration of the fused
molecule. In the case of IL-12, reduced activity is more likely due to
steric hindrance since each IL-12 molecule is approximately threefold
larger than these other cytokines.
A fusion between mouse IL-12 and OVA has already been reported (34), and in this case, biologic activity was reduced 50-fold, as measured in a proliferation assay.
A direct comparison with our human IL-12 fusion protein in the same proliferation assay suggests that it has significantly more activity, as it was only 2-fold rather than 50-fold less active than the corresponding IL-12 standard. However, this comparison is complicated, since there could be significant differences between mouse and human IL-12 regarding receptor/ligand interaction, and the constraints imposed by fusion to another protein. An important difference between the OVA and Ab fusion protein is that OVA was fused to p40 rather than p35, as in our constructs, thus structurally confining the amino terminus of p40, rather than the p35 subunit, and its ability to bind IL-12R. In this regard, it is of interest that a single-chain IL-12 fusion protein (35), consisting of the mouse p35 and p40 subunits linked via a flexible peptide, had normal levels of biologic activity (measured by proliferation) when the p40 subunit was at the amino terminus, but very little activity when the amino terminus was fused to the peptide linker. Thus, it seems that a free p40 amino terminus is important in maintaining biologic activity, and presumably, receptor binding.
A possible explanation that a less active, whole Ab-IL-12 mouse/human
fusion protein was more efficacious than free, systemically
administered IL-12 is that the more active free form is generated by
proteolysis at the tumor site in vivo, resulting in a high local
concentration of active cytokine. Evidence for cleavage in vivo is
provided in our pharmacokinetic analyses, which show a difference in
the amount of intact fusion protein and Ab components. That circulating
fusion protein is less active is borne out by the greatly reduced serum
levels of IFN-, relative to what has been reported previously
(29).
We have also described a useful system for testing the efficacy of therapies targeted against the human EpCAM pan-carcinoma Ag, and have used it to assess the potential of these Ab-IL-12 fusion proteins. This Ag is distributed widely on all human tumors of epithelial origin and particularly well expressed on colon carcinoma. In fact, the 17-1A Ab targeting this Ag has shown efficacy for treatment of colon cancer in an adjuvant setting (36). The mouse CT26 colon carcinoma lines, transfected with human EpCAM, have been selected on the basis of MHC class I expression as well. In this study, we have focused on the low class I-expressing clone, Ep21.6, to test the potential of EpCAM-specific Ab-cytokine fusion proteins in treatment of established pulmonary metastases in SCID mice. This model is designed to reflect an unfavorable clinical setting in which the tumor expresses low levels of class I molecules (as is the case with human colon carcinoma (37)), and there is reduced T cell function as a result of prior therapy, or as a consequence of tumor burden.
The cell lines described in this work are also suitable for testing in
immune-competent, syngeneic BALB/c mice, despite the transfection of a
human Ag. The expression of a foreign protein can itself induce an
immune response. However, we found that high level cell surface
expression of human EpCAM does not induce rejection of the CT26 colon
carcinoma, nor does it slow its ability to kill BALB/c mice. This may
be due, in part, to the fact that the human and mouse forms of this
protein share significant homology (80%). In fact, we have found
that equivalent doses of parental CT26 or EpCAM-transfected cells form
s.c. tumors or metastases with the same kinetics. This is in contrast
to what was reported earlier by others (38), who found that
approximately 200 times more of the EpCAM-transfected cells were
required to form skin tumors, relative to the parental CT26
control.
In summary, we demonstrated that it is possible to genetically engineer and express humanized Ab-IL-12 fusion proteins with relatively high sp. act. Such molecules are effective in SCID mouse xenograft models of human prostate carcinoma, relative to an irrelevant Ab-IL-12 control with the same IL-12 activity and pharmacokinetic properties, but no ability to bind to tumor cells. The hybrid mouse/human IL-12 fusion protein is as effective as an Ab-IL-2 fusion protein, despite its relatively low IL-12 activity, in a very aggressive colon carcinoma model in SCID mice. Taken together, our data suggest that Ab-IL-12 fusion proteins may be applicable for the treatment of epithelial malignancies in an adjuvant setting.
|
Footnotes |
|---|
2 Abbreviations used in this paper: EpCAM, epithelial cell adhesion molecule; ADCC, antibody-dependent cellular cytotoxicity; GM-CSF, granulocyte-macrophage colony-stimulating factor; H chain, heavy chain; L chain, light chain; LAK, lymphocyte-activated killer; sc, single-chain.
Received for publication October 14, 1997. Accepted for publication February 13, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
production by NK cell stimulatory factor (NKSF): characterization of the responder cells and synergy with other inducers. J. Exp. Med. 173:869. in vivo. Int. J. Immunol. 6:157.
This article has been cited by other articles:
|
|
 |
|
  S. D. Gillies, Y. Lan, S. Williams, F. Carr, S. Forman, A. Raubitschek, and K.-M. Lo An anti-CD20-IL-2 immunocytokine is highly efficacious in a SCID mouse model of established human B lymphoma Blood, May 15, 2005; 105(10): 3972 - 3978. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Halin, V. Gafner, M. E. Villani, L. Borsi, A. Berndt, H. Kosmehl, L. Zardi, and D. Neri Synergistic Therapeutic Effects of a Tumor Targeting Antibody Fragment, Fused to Interleukin 12 and to Tumor Necrosis Factor {alpha} Cancer Res., June 15, 2003; 63(12): 3202 - 3210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Gillies, K.-M. Lo, C. Burger, Y. Lan, T. Dahl, and W.-K. Wong Improved Circulating Half-Life and Efficacy of an Antibody-Interleukin 2 Immunocytokine Based on Reduced Intracellular Proteolysis Clin. Cancer Res., January 1, 2002; 8(1): 210 - 216. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Ruehlmann, R. Xiang, A. G. Niethammer, Y. Ba, U. Pertl, C. S. Dolman, S. D. Gillies, and R. A. Reisfeld MIG (CXCL9) Chemokine Gene Therapy Combines with Antibody-Cytokine Fusion Protein to Suppress Growth and Dissemination of Murine Colon Carcinoma Cancer Res., December 1, 2001; 61(23): 8498 - 8503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Xiang, F. J. Primus, J. M. Ruehlmann, A. G. Niethammer, S. Silletti, H. N. Lode, C. S. Dolman, S. D. Gillies, and R. A. Reisfeld A Dual-Function DNA Vaccine Encoding Carcinoembryonic Antigen and CD40 Ligand Trimer Induces T Cell-Mediated Protective Immunity Against Colon Cancer in Carcinoembryonic Antigen-Transgenic Mice J. Immunol., October 15, 2001; 167(8): 4560 - 4565. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Holden, Y. Lan, A. M. Pardo, J. S. Wesolowski, and S. D. Gillies Augmentation of Antitumor Activity of an Antibody-Interleukin 2 Immunocytokine with Chemotherapeutic Agents Clin. Cancer Res., September 1, 2001; 7(9): 2862 - 2869. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Christ, S. Matzku, C. Burger, and M. Zöller Interleukin 2-Antibody and Tumor Necrosis Factor-Antibody Fusion Proteins Induce Different Antitumor Immune Responses in Vivo Clin. Cancer Res., May 1, 2001; 7(5): 1385 - 1397. [Abstract] [Full Text]
|
|
|
|
|
|
O. Christ, S. Seiter, S. Matzku, C. Burger, and M. Zöller Efficacy of Local versus Systemic Application of Antibody-Cytokine Fusion Proteins in Tumor Therapy Clin. Cancer Res., April 1, 2001; 7(4): 985 - 998. [Abstract] [Full Text]
|
|
|
|
|
|
J. G. Turner, A. L. Rakhmilevich, L. Burdelya, Z. Neal, M. Imboden, P. M. Sondel, and H. Yu Anti-CD40 Antibody Induces Antitumor and Antimetastatic Effects: The Role of NK Cells J. Immunol., January 1, 2001; 166(1): 89 - 94. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Xiang, H. N. Lode, S. D. Gillies, and R. A. Reisfeld T Cell Memory Against Colon Carcinoma Is Long-Lived in the Absence of Antigen J. Immunol., October 1, 1999; 163(7): 3676 - 3683. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. S. Schmidt and M. F. Mescher Adjuvant Effect of IL-12: Conversion of Peptide Antigen Administration from Tolerizing to Immunizing for CD8+ T Cells In Vivo J. Immunol., September 1, 1999; 163(5): 2561 - 2567. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. S. Peng, M. L. Penichet, and S. L. Morrison A Single-Chain IL-12 IgG3 Antibody Fusion Protein Retains Antibody Specificity and IL-12 Bioactivity and Demonstrates Antitumor Activity J. Immunol., July 1, 1999; 163(1): 250 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. D. Gillies, Y. Lan, K.-M. Lo, M. Super, and J. Wesolowski Improving the Efficacy of Antibody-Interleukin 2 Fusion Proteins by Reducing Their Interaction with Fc Receptors Cancer Res., May 1, 1999; 59(9): 2159 - 2166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. N. Lode, T. Moehler, R. Xiang, A. Jonczyk, S. D. Gillies, D. A. Cheresh, and R. A. Reisfeld Synergy between an antiangiogenic integrin alpha v antagonist and an antibody-cytokine fusion protein eradicates spontaneous tumor metastases PNAS, February 16, 1999; 96(4): 1591 - 1596. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
