| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Lymphoma Research Unit, Tenovus Laboratory, General Hospital, Southampton, United Kingdom;
Wellcome Trust Immunology Center, University of Cambridge School of Clinical Medicine, Cambridge, United Kingdom;
Institute for Animal Health, Pirbright, United Kingdom; and
§
National Institute for Medical Research, London, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As an alternative explanation for the clinical success of some mAb, many researchers are now considering that mAb might have a direct regulatory effect on tumor growth. A signaling role for therapeutic mAb was shown by experiments in which anti-µ Ab was able to effect cytotoxicity by cross-linking the surface Ig complex (BCR) on certain B cell lines and on some fresh lymphoid tumors (12, 13). Recent clinical results from Vuist and colleagues (14) now indicate that the therapeutic activity of anti-Id mAb in B cell lymphoma correlates with the ability to induce Ig signal transduction. This work revealed a striking correlation between the tendency of anti-Id mAb to produce clinical responses in non-Hodgkin’s lymphoma patients and their ability to induce transmembrane signaling in a patient’s lymphoma cells, as measured by the phosphorylation of intracellular proteins.
While tailor-made anti-Id mAb are not considered practical for general application in lymphoma treatment, they may still have much to teach us about the process of controlling cell growth in vivo. Although Ab-mediated signaling appears critical in anti-Id mAb treatment, it may also play an important role in many other mAb treatments, involving both neoplastic and normal cells. A number of mAb that have been used, or at least considered, for treating lymphoma, such as anti-CD19 (15, 16), anti-CD20 (1, 4), anti-CD22 (17), anti-CD38 (11), and anti-CD40 (18, 19), are capable of generating transmembrane signals in normal and neoplastic B cells. In the current work we have investigated a panel of anti-mouse B cell mAb in two syngeneic B cell lymphoma models, BCL1 (20) and A31 (21), to try to establish whether it is direct growth regulation, ADCC, or complement that is of over-riding importance in therapy. Our results clearly show that results from in vitro assays do not correlate with activity in vivo and that, at least for B cell lymphoma, therapeutically successful mAb appear to be those directed at key receptor molecules (Id, CD19, and CD40) involved in transmembrane signaling during B cell responses.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
BALB/c and CBA/H mice and Louvain (LOU/OlaHsd) rats were all supplied by Harlan UK (Blackthorn, U.K.), and maintained in local animal facilities.
BCL1 (20) and A31 (21) mouse B lymphoma lines were maintained by in vivo, i.p. passage in BALB/c and CBA mice, respectively. Enlarged spleens were taken at the terminal stage of disease, and single cell suspensions were prepared as described previously (22). The BCL1-3B3, cell line (23) is a variant of the BCL1 tumor that is maintained in culture using standard medium (see below).
All cell culture was performed in supplemented DMEM containing glutamine (2 mM), pyruvate (1 mM), penicillin and streptomycin (100 IU/ml), fungizone (2 µg/ml), and 5 or 10% FCS (Myoclone; Life Technologies, Paisley, Scotland) or in supplemented RPMI (Life Technologies) containing the same supplements but with the addition of 50 µM 2-ME (BDH, Poole, U.K.).
Abs and fragments
All mAb used in this study are shown in Table I
together with their sources. Three new
mAb, Mc39-12, Mc39-16, and TI2-3, were generated during the project
using standard somatic fusion technology as discussed previously (28).
Mc39-12 and Mc39-16 mAb reacted with mouse µ-chain and A31 Id,
respectively, and were raised by immunizing LOU rats with mouse IgM
that had been immunoprecipitated from the A31 lymphoma according to the
method described by Hamblin et al. (29). Three days after the final
booster, spleen cells from the rat were fused with NS-1 mouse myeloma
cells, and supernatants from the resulting hybridomas were screened by
ELISA, selecting mAb that bound all mouse IgM (Mc39-12) or only A31
idiotypic IgM (Mc39-16).
|
Hybridoma cells were expanded in stationary culture using 5% supplemented DMEM. To purify the IgG mAb, the culture supernatants were concentrated 20 times by membrane filtration (Amicon, Beverley, MA), precipitated with saturated ammonium sulfate, and then dialyzed and fractionated on protein G (Pharmacia, Piscataway, NJ) according to the manufacturer’s instructions. Four of the rat mAb (Mc39-12, Mc39-16, Mc10-6A5, and M5-114) were prepared by ion-exchange chromatography on DEAE (Whatman, Clifton, NJ) as described by Elliott et al. (28). The purity of all IgG preparations was checked by electrophoresis (EP system, Beckman, Palo Alto, CA) and HPLC using a Zorbax GF250 Bio Series column (Du Pont, Wilmington, DE) (31).
Flow cytometry
Mouse lymphoma cells (splenic) were analyzed by direct immunofluorescence staining using a FACS vantage (Becton Dickinson, Mountain View, CA) as described previously (31).
Binding of [125I]mAb to the surface of lymphoma cells
mAb were trace radiolabeled for binding studies using carrier-free 125I (Amersham International, Aylesbury, U.K.) and Iodo-Beads (Pierce, Rockford, IL) as the oxidizing reagent and were extensively dialyzed to remove unbound 125I.
The binding of radiolabeled mAb to cells was determined as described by Elliott et al. (28). Radiolabeled mAb were serially diluted before incubation with cells (0.5–1.25 x 106/ml; final volume, 1 ml) in RPMI medium containing 10% FCS for 2 h at 37°C. Endocytosis of cell-bound mAb was prevented by inclusion of sodium azide (15 mM) and 2-deoxyglucose (50 mM). The cells were separated from the aqueous phase by rapid centrifugation through a 1.1/1 (v/v) mixture of dibutyl phthalate/dioctyl phthalate according to the method of Dower et al. (32). Radioactivity was measured in a Rackgamma spectrometer (LKB, Gaithersburg, MD).
Complement-dependent cytotoxicity (CDC) assay
The ability of the mAb to kill A31, BCL1 and BCL1-3B3 cells in a CDC assay was determined as previously described (24). Briefly, 51Cr-labeled target cells were exposed to various concentrations of each mAb and control reagents on ice for 30 min. Fresh rat serum (final dilution, 1/5 with supplemented DMEM) was then added as a source of complement, and the samples were warmed to 37°C. After 45 min the samples were centrifuged for 5 min to sediment cells, and specific 51Cr release was assessed by counting the supernatant. All samples were run in duplicate, and the maximum release was determined by measuring 51Cr release obtained after addition of 1% Nonidet P-40 to the cells.
ADCC assay
The ADCC assay was modified from previous work (33). 51Cr-labeled A31, BCL1, and BCL1-3B3 target cells were prepared as described for the CDC assays and resuspended at a final concentration of 105 cells/ml in supplemented DMEM. The effectors were peritoneal exudate cells washed from the peritoneal cavity of CBA mice 5 days after they had received an i.p. injection of thioglycolate broth (Becton Dickinson, Oxford, U.K.) (9). The exudate cells were washed once and resuspended in supplemented DMEM at the appropriate concentration.
For the assay, all samples were made in supplemented DMEM. First, 50-µl aliquots of target cells (5 x 103) and mAb were mixed in individual wells of 96-well U-bottomed culture plates (Life Technologies, Paisley, U.K.) and left on ice for 15 min. Aliquots of 100 µl of effector cells were then added at an E:T cell ratio of 50:1. The plates were centrifuged at 200 x g for 5 min at room temperature, incubated for 6 h at 37°C in a CO2 incubator, and then centrifuged again, this time at 500 x g for 5 min, before finally harvesting 100 µl of the supernatant to estimate the released 51Cr.
All determinations were performed in triplicate, and the maximum release of radioactivity was calculated using target cells to which 150 µl of 1% Nonidet P-40 had been added. The percentage of specific 51Cr release was calculated using the standard formula: % specific release = [(sample release - background release)/(maximum release - background release)] x 100.
Growth and clonogenic assays for Ab-treated cells
In the short term growth assay to assess direct effects of mAb on tumor cells in vitro, BCL1-3B3 cells were exposed to mAb while measuring DNA synthesis (uptake of [3H]thymidine). Culture plates (96-well, flat-bottom; Life Technologies) were first precoated for 2 h at 37°C with each test mAb in sterile PBS (50-µl aliquots/well), 5 x 104 BCL1-3B3 target cells were then added to each well in a volume of 150 µl of supplemented RPMI, and the plates were incubated for 24 h at 37°C before pulsing with 0.5 µCi/well of [3H]thymidine (Amersham) for an additional 16 h. The DNA-incorporated radioactivity was harvested onto glass filters (Whatman) and counted as described previously (33).
In the clonogenic assay, BCL1-3B3 cells were plated in 96-well flat-bottom culture plates (Life Technologies) in supplemented RPMI (100 µl/well) at a frequency of two cells per well. mAb was included in the assay at 10 µg/ml throughout the culture period. For each assay, a complete 96-well plate of cells was set up per mAb. Colonies were grown for approximately 14 days and then counted and expressed as a percentage of those obtained in plates containing isotype-matched control mAb.
Ab immunotherapy
Groups of age-matched CBA and BALB/c mice were injected i.v.
with compatible lymphoma cells. CBA mice received an i.v. inoculum of
105 A31 cells on day 0, followed by mAb treatments on days
3 to 6 (0.5 mg/day, 2 mg total). The first mAb treatment was given
i.v.; the next three were given i.p. BALB/c mice received a similar
protocol with 105 i.v. BCL1 cells on day 0
followed by four doses of mAb on days 9 to 12, again giving the first
mAb treatment i.v. and the next three by the i.p. route. Both tumors
develop primarily in the spleen, with a leukemic overspill toward the
end of the disease (20, 21). Survival was monitored daily, and the
results were analyzed using the 
2 test of Peto (34).
Tracking tumor cells in vivo
To understand more fully the mechanism of action of anti-tumor mAb in vivo we developed two methods of tracking cells after therapy. In the first, fresh A31 cells were tagged in vitro with 5-(and 6-)carboxyfluorescein diacetate, succinimidyl ester (CFSE; Molecular Probes, Leiden, The Netherlands). Stock CFSE (5 mM) was prepared by dissolving CFSE in DMSO and was stored frozen. A31 cells were resuspended in serum-free RPMI 1640 (Life Technologies) at a concentration of 2 x 107/ml and prewarmed to 37°C. Stock CFSE was then added to the cell suspension to give a final concentration of 5 µM and was mixed thoroughly by inversion. The suspension was incubated for 10 min at 37°C, and the cells were washed twice with 4 vol of ice-chilled RPMI 1640 containing 10% FCS. Finally, the cells were suspended in PBS at a concentration of 2.5 x 108/ml for i.v. injection in mice.
In the second method, groups of mice were inoculated i.v. with 5 x 107 A31 or BCL1 cells in 200 µl of PBS on day 0 and then, when the tumor cells were growing in log phase (day 3 for A31 and day 4 for BCL1), were treated with the appropriate mAb (2 mg i.v.). The total yield of splenic tumor cells was estimated each day by staining spleen homogenates with a combination of PE-anti-Id and FITC-anti-CD19 (or FITC-anti-CD22), which allowed tumor cells to be identified by flow cytometry.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Using the phthalate oil method of Dower et al. (31), we have
characterized a range of rat anti-mouse B cell mAb (Table I
) that
bind to the two lymphomas, A31 and BCL1, and might be
considered for treatment of these tumors. Only the anti-Id mAb are
tumor specific, but, as discussed in the introduction, all the other
specificities have at some time been considered for treatment of human
lymphoma. Figure 1 shows typical binding
curves performed at 37°C. Interestingly, the saturation levels of
binding for this panel of mAb are essentially the same on both cell
lines and BCL1-3B3 (data not shown), giving binding levels
in the order anti-CD74 
anti-MHC class II >
anti-Id > anti-CD22 > anti-CD19 =
anti-CD40. The BCL1-3B3 cell line is a variant of
BCL1, which, unlike the parent line, can be maintained in
long term culture (23). The invariant chain of class II, CD74, is very
strongly expressed, particularly on A31 (
2.6 x 105
anti-CD74 molecules bound/cell). CD19 and CD40 are expressed at
almost equal levels and at a relatively low density (<4 x
104 mAb molecules bound/cell) compared with the other
surface molecules. The affinities of the mAb (Table I) are as expected,
with apparent Ka values falling
in the range 108 to 109 M-1.
|
In an attempt to correlate in vitro assays with the performance of
mAb in immunotherapy, we first performed CDC and ADCC assays. Figure 2 shows that the anti-MHC class II
and anti-Id mAb produced marked cell lysis in CDC assays with all
three cell lines. Likewise, anti-CD74 mAb was active on A31 cells.
However, on the BCL1 and its variant, and despite the high
levels of expression of CD74, the anti-CD74 mAb failed to induce
appreciable levels of lysis. This is a somewhat surprising observation
that remains to be explained. The other three mAb, CD19, CD22, and
CD40, which were expressed at lower levels, were relatively inactive on
all cell lines under the conditions investigated.
|
show
that all the mAb mediated ADCC against A31, BCL1, and
BCL1-3B3. In general, the A31 tumor was more sensitive to
lysis than was BCL1 and its variant. Again, Ag density
appeared to be the over-riding feature determining the level of
cytotoxic activity achieved, with anti-MHC class II and
anti-CD74 mAb being consistently the most active. The main message
from both CDC and ADCC assays is that Ag density appears to be the key
factor in determining Ab potency within this group of reagents, there
being no obvious preference in terms of rat isotype or Ab specificity.
|
We next considered direct Ab toxicity on neoplastic B cells in the
absence of any effector systems. This work was performed on the
BCL1 culture subline, BCL1-3B3, since neither
A31 nor the parent BCL1 line can be sustained in culture
for more than a few hours. In the initial assays the uptake of
[3H]thymidine by cultured cells was used as a measure of
proliferation in the presence of the various mAb. Figure 4A shows a striking difference
in the ability of mAb to influence the growth of BCL1-3B3
cells. Two of the mAb, anti-Id and anti-CD74, had a direct
antiproliferative effect. Anti-CD74 was always more active than the
anti-Id mAb. In addition, the anti-CD19 mAb, 1D3, consistently
gave a slight inhibition of proliferation. However, the most effective
inhibitor of proliferation was the anti-µ mAb, Mc39-12, which
appears to deliver a strong growth inhibitory signal, perhaps because
of its ability to cross-link surface IgM (35). Because anti-Id mAb
is less active in this assay, these results are consistent with it
being relatively poor at cross-linking surface IgM. One explanation for
such observations is that anti-Id mAb binds bivalently, but in what
is sometimes called a monogamous union, in which each anti-Id mAb
engages surface Ig symmetrically (IgM/D to IgG mAb), providing intra-
rather than intermolecular cross-linking (28). Two other mAb,
anti-CD22 and anti-MHC class II, had no measurable influence on
BCL1-3B3 growth, while, as noted by others for activated
normal B cells (36), the anti-CD40 mAb tended to promote rather
than inhibit growth. From these in vitro assays we can conclude that
certain mAb specificities have a profound influence on the viability of
mouse lymphoma cells in culture.
|
B). BCL1-3B3 cells were plated at two
cells per well in a 96-well culture plate in the presence of mAb at 10
µg/ml. After approximately 14 days, colonies were counted and
compared with cells grown in isotype-matched controls. The results
again show that anti-µ and anti-Id mAb reduce or stop cell
growth. Interestingly, despite the marked inhibitory effect of
anti-µ mAb in the short term growth assay (Fig. 4A),
here we found that the anti-Id mAb was most potent and completely
prevented any development of BCL1-3B3 clones. The
anti-µ mAb reduced cloning efficiency by about 70%; however,
those clones that did develop were relatively small compared with
controls. The other mAb, including anti-CD40 mAb, did not influence
growth of the BCL1-3B3 cells, except that anti-MHC
class II, anti-CD74, and anti-CD19 mAb tended to cause the
cells to clump into very tight colonies. Finally, we have no
explanation for why anti-CD74 mAb, which causes marked growth
arrest in the short term assay (Fig. 4A), had no effect in
the clonogenic assay. Immunotherapy of A31 and BCL1 cells
Having established the binding and cytotoxic activity of this
panel of rat anti-mouse B cell mAb, we next performed immunotherapy
in tumor-bearing mice. Experiments in the A31 and
BCL1 models were conducted to determine an effective dosing
regimen (data not shown). With the exception of anti-Id mAb, which
was therapeutic at doses of <100 µg/animal, other mAb specificities
required relatively high doses of around 2 mg/mouse to achieve
near-maximum effectiveness. This high level ensured that all tumor
cells were saturated and exposed to mAb for a sustained period,
allowing weakly therapeutic mAb a chance to achieve a therapeutic
effect. Figure 5 shows the results of a
typical therapy in the A31 model in which mice received 105
fresh cells i.v. on day 0 and then 2 mg of mAb between days 3 and 6 at
a dose of 0.5 mg/day. Only two mAb, anti-CD40 and anti-Id,
consistently prolonged the survival of A31-bearing mice and resulted in
increases in survival of about 20 and 45 days, respectively.
Surprisingly, as little as 5 µg/mouse of anti-Id mAb was enough
to obtain 45 days of protection from the A31 lymphoma, far less than
would be needed to saturate the tumor cells in vivo (data not shown).
All the other mAb, despite binding to tumor cells in vivo, failed to
show protection.
|
shows a typical immunotherapy
experiment using 105 BCL1 cells and 2 mg of
each mAb given over 4 days (0.5 mg/day). Treatment of the
BCL1 tumor was delayed until day 9 after tumor inoculation
because this is a less aggressive malignancy than the A31. In the
BCL1 tumor, mAb to CD19, CD40, and Id were clearly
therapeutic, with 80% of those mice receiving anti-CD40 mAb going
on to become long term survivors (>100 days). Although the anti-Id
and anti-CD19 mAb were less effective, there was nevertheless
significant prolongation of survival in all treated animals
(p < 0.01). Interestingly, the therapeutic
effect of anti-CD19 mAb was highly dose dependent, and while
treatment with 2 mg of mAb gave a clear survival benefit, a dose of 100
µg/animal did not (data not shown). The anti-MHC class II mAb had
a small but statistically significant therapeutic effect
(p < 0.05), and the other mAb were without
detectable effect.
|
Tracking A31 and BCL1 cells in vivo following mAb treatment
In our initial attempts to follows tumor cells in vivo we made use
of the stable cell dye, CFSE, which tags cells by irreversibly
fluorescenating the cytoplasm. Once labeled with CFSE in vitro, the
cells and their progeny remain detectable by flow cytometry for 6 to 7
days in vivo. The experimental protocol involved labeling freshly
prepared tumor cells in vitro, injecting them i.v. into groups of mice
(5 x 107 cells/mouse), and then either leaving the
mice untreated or treating them with 2 mg of mAb (i.v.) 1 day later.
Figure 7 shows the results from a typical
experiment with CFSE-labeled A31 cells in which the splenic lymphocytes
have been recovered between days 1 and 8 and then stained with
PE-labeled anti-Id and analyzed by two-color flow cytometry.
Following i.v. injection, CFSE-labeled cells accumulated in the spleen
and began to divide, so that by 2 days after inoculation the tumor
cells (gated) represented about 2 to 3% of the total splenic
lymphocytes (Fig. 7). In untreated mice (Fig. 7, first
column), over the next 4 days these labeled tumor cells divided
rapidly, so that they comprised about half the splenic lymphocytes by 6
days after inoculation. The most interesting aspect of this labeling
technique was that the level of CFSE label dropped progressively as the
tumor population expanded. Thus, the CFSE marks cell division.
|
, second column), the tumor cells in the spleen
continued to proliferate for 2 or 3 days following mAb treatment. This
is shown by both the increase in the percentage of cells labeled with
CFSE and PE-anti-Id, and the progressive reduction in the intensity
of CFSE staining. The failure of the treatment to perturb cell growth
was most clearly seen 2 days after the mAb was given, as at this time
there appeared to be as many if not more tumor cells in the
anti-CD40 mAb-treated animal (Fig. 7, day 3; 9.9% gated) than in
the untreated control (Fig. 7, day 3; 6.2% gated). However, by day 4
differences started to emerge, with a slight reduction in the
percentage of tumor cells after anti-CD40 mAb treatment. By day 6
this reduction was marked (Fig. 7, second column; 3.6%
after anti-CD40 mAb compared with 54% in the untreated control),
and by day 8 tumor cells were no longer detectable following
anti-CD40 mAb treatment.
The profile of tumor cell-growth obtained after anti-Id mAb
treatment was clearly different (Fig. 7, third column).
First, as expected, the PE-anti-Id label was unable to detect the
tumor cells due to blocking and internalization of their surface Ig by
the treatment mAb. This introduced some difficulties in determining a
precise gate for the tumor cells in flow cytometry. Second, anti-Id
mAb treatment appeared to have a quicker effect on tumor growth than
anti-CD40 mAb and resulted in arrested tumor cell division between
days 3 and 4, as shown by the comparative lack of increase in the
percentage of tumor cells. Interestingly, the tumor cells that remained
after treatment showed a relatively small reduction in their levels of
CFSE staining between days 3 and 6, consistent with the belief that
they are growth arrested. Comparison of the intensity of CFSE staining
in tumor cells on day 6 showed a clear distinction between cells taken
from untreated controls and anti-CD40 mAb-treated animals (Fig. 7, first and second columns) and those recovered
from mice given anti-Id mAb (third column). This
difference in mean fluorescence intensity (CFSE) for the two groups was
about 10-fold. Finally, and despite impressive immunotherapy by
anti-Id in A31 lymphoma (Fig. 5), this treatment did not eradicate
the tumor cells, and they remained detectable as a CFSE-labeled
population throughout the study period. In recent experiments we have
been able to extend this period to 10 days, at which point small
numbers of splenic CFSE-labeled tumor cells can still be detected after
anti-Id mAb treatment.
Experiments with CFSE-labeled BCL1 tumor also showed that
anti-CD40 mAb treatment eradicated tumor (data not shown), and that
the kinetics were very similar to those seen in the A31 lymphoma (Fig. 7), i.e., 2 to 3 days of proliferation followed by a rapid decline
(data not shown). In addition, anti-Id and anti-CD19 mAb
produced results similar to those seen with anti-Id in the A31
tumor (Fig. 7, third column). These two mAb prevented the
growth of BCL1, as shown by cell numbers and the rate of
loss of CFSE stain. However, the growth-arresting effects of
anti-Id and anti-CD19 mAb were less evident than those seen
following anti-Id mAb in the A31 tumor, probably because these two
mAb are comparatively poor in immunotherapy of BCL1 (Fig. 6) and consequently less able to contend with the large tumor load that
resulted from injecting 5 x 107 cells. Finally, as
expected, nontherapeutic mAb (Figs. 5 and 6), such as anti-CD22
mAb, failed to have a significant impact on CFSE-labeled tumor growth
in either lymphoma model (see below).
To confirm and extend these observations we next investigated the
development of A31 and BCL1 lymphoma cells without
prelabeling in vitro by calculating the total yield of splenic tumor
cells recovered following mAb treatment. In these studies mice were
given 5 x 107 fresh A31 or BCL1 cells on
day 0 and then treated with 2 mg of mAb on day 3 (A31) or day 4
(BCL1) at a time when the lymphoma cells were growing in
log phase. Tumor cells were detected using flow cytometry after double
staining spleen cells with PE-anti-Id mAb and FITC-anti-CD19
or, in one case, FITC-anti-CD22 mAb (see below). The results in
Figure 8 confirm the rapid growth of both
tumors and show that mAb, such as anti-CD22 and anti-MHC class
II, that were not effective in standard therapy experiments (Figs. 5 and 6) had no impact on the yield of tumor cells recovered from the
spleens of tumor-bearing animals. They also show that anti-CD40 and
anti-Id mAb in A31 and anti-CD40 mAb in BCL1 had a
profound impact on tumor growth, which reflected the effects seen in
CFSE labeling experiments. In neither model did anti-CD40 mAb cause
any inhibition of tumor cell expansion for the 3 days following
treatment, but then tumor cell number dropped markedly as the mAb had
its therapeutic effect. This reduction in tumor cell number was
particularly dramatic in the BCL1 model, where tumor cells
were almost completely eradicated between days 7 and 8. In contrast to
this delayed therapeutic effect, anti-Id mAb treatment of A31
lymphoma caused an immediate halt to tumor cell proliferation, but did
not eliminate tumor over a 5-day period, a result that confirms the
growth arrest seen in the CFSE experiments (Fig. 7). Measuring tumor
load in the A31 tumor following anti-Id mAb treatment was achieved
due to an unusual level of CD22 Ag on the tumor cells compared with
that on the normal B cells in CBA mice. Thus, by staining with
FITC-anti-CD22 mAb we were able to distinguish the tumor cells
despite their surface Ig Id being blocked or internalized by the
treatment mAb. Unfortunately, a similar distinction could not be
achieved with the BCL1 tumor in BALB/c mice, and
consequently, this experiment could not be performed with anti-Id
mAb in the BCL1 model. Interestingly, when anti-CD19
mAb was used for treatment of BCL1 it failed to control
tumor growth and, as shown in the example in Figure 8, sometimes
appeared to cause modest growth promotion of tumor for a short period
after its administration.
|
). In the standard
immunotherapy systems shown in Figure 6, anti-Id and anti-CD19
mAb were therapeutically active, but the BCL1 tumor load at
the time of treatment was less than 1/30th of that present in
these tracking experiments (data not shown). Together, these in vivo
tracking experiments clearly show that two therapeutic mAb, anti-Id
and anti-CD40, appear to be working in very different ways:
anti-Id mAb cause a growth arrest that is almost immediate and does
not eliminate cells over a period of 5 or 6 days, and anti-CD40 mAb
have a delayed effect after 2 to 3 days of uninterrupted growth, but
then eradicate lymphoma cells very effectively. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
; anti-Id, -CD19, -CD22, -CD40, -CD74, and
-MHC class II). In vivo the results were very clear, with only three
mAb, anti-Id and anti-CD40 in both tumor models and
anti-CD19 in BCL1, showing good therapeutic activity.
All the other mAb showed little or no therapeutic activity despite
being given at doses of 2 mg/mouse at a time when tumor burden was low.
Interestingly, these in vivo results could not have been predicted from
the performance of the same mAb in standard in vitro cytotoxicity
assays. In both CDC and ADCC, activity appeared to correlate with the
level of mAb binding. Thus, mAb such as anti-MHC class II (M5-114),
anti-CD74 (TI2-3), and anti-Id, that bound at high levels
tended to give good cytotoxicity, while those that bound at lower
levels (anti-CD19, anti-CD22, and anti-CD40) were usually
less active. One exception to this general observation was
anti-CD74 mAb on BCL1. This bound at very high levels
in both models (>105 molecules/cell) and was fully active
on the A31 tumor, but failed to induce marked complement cytotoxicity
of BCL1 or its in vitro variant BCL1-3B3. We
are currently unable to explain this discrepancy. In growth and clonogenic assays using the in vitro cell line, BCL1-3B3, although showing the profound effect of "naked" mAb on cell growth, we were again unable to see a correlation between mAb behavior in vitro with the treatment results achieved in mice. In both assays the mAb directed at the BCR (anti-Id and anti-µ) caused growth inhibition, particularly the anti-Id mAb in the clonogenic assay, which stopped all cell growth. In addition, the few colonies seen in the clonogenic assay with anti-µ were very small and fragile compared with those in controls. Other mAb, such as anti-CD74 and anti-CD40, did not produce results that were consistent between the two assays. Thus, while anti-CD74 caused a marked growth arrest in the short term assay, it had no effect in the clonogenic assay, and the slight growth promotion seen with anti-CD40 mAb in the short term assay did not result in increased cloning efficiency in the second assay. Therefore, it is our belief that the current work casts serious doubt on the ability of conventional cytotoxicity or growth assays to identify useful mAb for the treatment of many types of cancer (5, 6, 7, 8, 9) and leaves open the question of what are the main mechanisms operating during mAb treatment of neoplastic cells in vivo.
The most important results obtained in the current work to help explain the mechanisms of action of mAb come from in vivo experiments in which lymphoma cells were either tagged with CFSE dye or monitored by flow cytometry following mAb treatment. This work clearly showed that two therapeutically useful mAb, anti-Id and anti-CD40, control tumor growth in completely different ways: anti-Id mAb caused immediate growth arrest without tumor eradication, and anti-CD40 mAb had no effect on lymphoma growth for 2 or 3 days and then caused a dramatic loss of dividing cells. Both these patterns of tumor control are consistent with metabolic changes in the target cells that could have resulted from transmembrane signaling.
A number or recent investigations have suggested that certain therapeutic mAb, particularly anti-Id, might operate at least in part via a direct effect on cell proliferation, and that such activity probably requires cross-linking activity by the mAb to generate intracellular signals (14, 15, 18, 37). Anti-Id mAb in animals and patients appears to deliver signals that inhibit the growth of B cell tumors and, provided the BCR becomes sufficiently cross-linked by Ab, might even drive cells into apoptosis, as has been shown following treatment of lymphoma-bearing SCID mice treated with polyclonal anti-µ Ab (35, 38). Racila et al. (38), working in the BCL1 model, have presented convincing data showing that anti-Id and anti-µ Ab, given passively or produced following active immunization (39), provoke intracellular signals that regulate tumor growth and leave cells in a state of dormancy for extended periods (40). The current work also shows rapid growth arrest of lymphoma cells following anti-Id mAb, especially in the A31 model, and it is quite conceivable that these cells, having not been destroyed by the treatment, might remain dormant for an extended period. In addition, maximum protection in the A31 model requires very little anti-Id mAb (5 µg/mouse), far less than would be required to saturate the target cells for recruitment of effector systems, but perhaps enough to allow signaling for metabolic perturbation.
It has been known for some time that anti-CD19 mAb is therapeutic in xenografts (15) and patients (16). Furthermore, considerable evidence now points to an important signaling component within the therapeutic activity of this mAb specificity (37, 41). This signaling role of anti-CD19 mAb is supported by data showing that when given at very high levels it functions therapeutically as a F(ab')2, a result that clearly rules out FcR-bearing effectors in this system (15). While our data show that anti-CD19 mAb is therapeutic in the BCL1 model, we were unable to obtain a clear indication from the tracking experiments of how this reagent was working. Further investigations will be required to show whether tumor cells treated with anti-CD19 mAb follow a similar response pattern as those receiving anti-Id mAb. Interestingly, recent data from normal B cells suggest that the BCR and surface CD19 may be part of a common signaling unit and cooperate in optimizing responses to Ag (42).
While treatment with anti-CD19 and anti-Id mAb may have
features in common, our data show that anti-CD40 is operating via a
completely different pathway. CD40 is a major costimulatory molecule on
normal B cells. It is normally engaged by CD40 ligand on helper T cells
during Ab responses and appears to be necessary for driving B cell
proliferation and rescuing them from the tolerogenic signals delivered
via the BCR (43, 44, 45). However, work from Funakoshi et al. (18, 19) has
recently shown that the situation is slightly different in many B cell
malignancies in which anti-CD40 mAb, if cross-linked appropriately,
can result in profound growth arrest of cultured cells. Furthermore,
when used in SCID/xenograft models, the same anti-CD40 mAb is
effective in protecting mice from tumor development. At least part of
this in vivo activity appears to depend on signaling activity of the
anti-CD40 mAb because it is partially effective even in the absence
of FcR-expressing effector cells. Our current in vivo work confirms and
extends these observations and shows that even in syngeneic lymphomas
anti-CD40 is therapeutically active. Interestingly, in vitro we
found no evidence of growth arrest, but, rather, the anti-CD40
tended to stimulate growth at least in short term culture. This result
does not agree with those of Funakoshi et al. (18, 19), who reported
only growth arrest in various lymphoma lines. It is quite possible that
the BCL1-3B3 line used here behaves differently from
the human B cell lines used previously, or that the assay conditions,
such as the level of cross-linking, are critically different in the two
studies. Whatever the explanation, it is clear that anti-CD40 mAb
can be therapeutically very active and that part of its mechanism of
action may result from inhibitory signaling. It is important to note
that in normal B cells, signaling via CD40 leads to profound phenotypic
changes, including up-regulation of Fas, and that in certain
circumstances such surface changes can result in B cell depletion (43, 46). It is possible that similar changes occur following anti-CD40
mAb treatment of neoplastic B cells and that these are also important
in the removal of tumor. The current work shows that anti-CD40 mAb
operates with very interesting kinetics, in which tumor cells continue
to proliferate normally for 2 or 3 days after mAb treatment; only then
do their numbers start to decline compared with those in untreated
controls (Figs. 7 and 8). This picture is very different from the
situation following anti-Id mAb treatment, where cells underwent
rapid growth arrest. Investigations are underway to establish the
mechanism of anti-CD40 mAb and whether in vivo it works indirectly,
perhaps killing cells via the Fas system. The relatively slow kinetics
of 2 to 3 days are consistent with up-regulation and activation of
surface Fas; however, it is not clear how this Fas would be engaged,
since Fas ligand is generally on activated T cells and, as
yet, information is not available to show whether such
activation occurs during mAb therapy (43, 47). As an alternative
explanation we will also be considering whether in vivo the
anti-CD40 mAb, rather than delivering a direct or indirect
cytotoxic effect, operates by blocking critical growth signals that are
dependent on the expression of surface CD40.
In conclusion, it is probably more than coincidence that those mAb, such as anti-Id, anti-19, and anti-CD40, that have been successful in the treatment of two mouse B cell lymphomas are all directed at key signaling molecules. Under normal circumstances these same receptors would be required to drive B cell activation and differentiation following interaction with Ag (BCR) and Th cells (CD40). It is now clear that such responses are highly coordinated and only occur when an array of signals is delivered following the appropriate receptor-ligand interactions. It appears that cells that are unable to receive a full complement of these signals are probably rendered anergic or may be deleted and thereby prevented from responding inappropriately to Ag. We speculate that a somewhat similar process may occur when these receptors are heavily cross-linked by individual mAb on lymphoma cells. Such treatment would be expected to deliver an intense, but inappropriate, signal via a single receptor. The target cell, "knowing" that it was not receiving all the signals required for a specific immune response, may be triggered into growth arrest and apoptosis. At present it appears that there are at least two pathways to such inappropriate signaling, one via the BCR and possibly CD19 which is unlikely to involve Fas (48), and a second through CD40, where Fas up-regulation is a distinct possibility (43, 47). If correct, then the implications of such a mechanism on the design of future mAb and vaccine design would be profound. No longer should we think of anti-cancer Ab simply as glycoproteins that recruit cytotoxic effectors; special attention must be given to Ab specificity and their signaling activity on target cells. A major part of this new thinking should include reassessing the in vitro assays normally used for selecting potential therapeutic agents, with perhaps more emphasis being placed on systems that detect Ab-induced transmembrane signaling and triggering of metabolic changes in cell growth and gene expression.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Martin J. Glennie, Lymphoma Research Unit, Tenovus Laboratory, General Hospital, Southampton, United Kingdom SO16 6YD. E-mail address:
3 Abbreviations used in this paper: ADCC, antigen-dependent cellular cytotoxicity; BCR, B cell receptor for antigen; CDC, complement-dependent cytotoxicity; PE, phycoerythrin; CFSE, 5-(and 6-)carboxyfluorescein diacetate succinimidyl ester.
Received for publication December 22, 1997. Accepted for publication May 12, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/ß x anti-CD2) F(ab')2 antibody. J. Immunol. 155:2960.[Abstract]
This article has been cited by other articles:
|
|
 |
|
indicates
the control IgG; 
indicates anti-Id; • indicates
anti-CD19 (1D3); 
indicates anti-CD22 (NIMR6);
• indicates anti-CD40 (3/23); 
indicates anti-CD74
(TI2-3); 
indicates anti-MHC class II (M5-114). One of two
similar experiments is shown. The anti-CD19 and anti-CD40
results overlapped each other on both occasions and are therefore shown
as a single line for clarity of presentation. The control IgG for each
tumor was the inappropriate, nonbinding, anti-Id mAb (i.e.,
anti-BCL1 Id mAb on A31 and anti-A31 Id mAb on
BCL1).
indicates
anti-CD40 (3/23); 
indicates
anti-µ (Mc39-12); and 
