| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics |
Auckland Cancer Society Research Centre, The University of Auckland, Private Bag 92019, Auckland, New Zealand
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The context in which bystander effects in tumors have been most thoroughly studied is in GDEPT3 (1, 2, 3, 4) . This is a form of suicide gene therapy in which the therapeutic gene codes for a PAE; the best known examples are herpes simplex thymidine kinase or bacterial/yeast cytosine deaminase, which activate ganciclovir and 5-fluorocytosine, respectively, to cytotoxic antimetabolites (5) . The latter provide a bystander effect by diffusing from the cell of origin (activator cell) to kill non-PAE-expressing (target) cells in the microenvironment.
One of the more promising enzyme/prodrug systems for GDEPT uses the aerobic NTR (a product of the nfsB/nfnB gene) to activate the dinitrobenzamide aziridine derivative CB 1954 (compound 1, Fig. 4A
) to a bifunctional DNA cross-linking agent (6
, 7) . An advantage of this system over antimetabolite-based GDEPT is its activity against noncycling tumor cells (8)
. NTR expression results in dramatic sensitization of a wide variety of cell lines to CB 1954 in culture (8, 9, 10, 11)
and in human tumor xenografts (11
, 12)
. The NTR/CB 1954 enzyme/prodrug system is currently in clinical trial in conjunction with a replication-defective adenoviral vector (13)
.
|
We have developed a 3D cell culture system in which tumor cells are grown on microporous support membranes to form multicellular layers (16, 17, 18)
, as illustrated in Fig. 1a
. These multilayers become diffusion-limited and, as a result, model key aspects of the extravascular compartment of tumors with the development of noncycling and hypoxic cells and, eventually, central necrosis. Importantly, multilayers can readily be established from mixed cell types to provide steady-state cocultures at tissue-like cell densities.
|
), which was originally designed to be activated in hypoxic tumor cells by endogenous one-electron reductases (19)
but was subsequently shown also to be activated efficiently by NTR (20)
. A potential advantage of SN 23862 in NTR/GDEPT is that it is not activated by DT-diaphorase (19)
, which is the major endogenous CB 1954 reductase in aerobic mammalian cells (21
, 22)
. In the present context, SN 23862 was also of interest because its cytotoxic metabolites are expected to be more lipophilic than those for CB 1954, which we hypothesized would lead to a more efficient bystander effect if the latter requires passive diffusion through plasma membranes. |
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
were synthesized in this laboratory as described previously (19
, 23)
. All prodrugs had >97% purity by high-performance liquid chromatography, monitoring at the longest wavelength absorbance maximum. Octanol:water (pH 7.4; 21°C) partition coefficients (P) were measured as described previously (24)
. Prodrug stock solutions were stored frozen in DMSO and diluted into culture medium immediately before addition to cell cultures. For administration to mice, prodrugs were dissolved in DMSO or DMA and in some cases diluted by sequential addition of PEG (Mr 400) and water. The final composition of the vehicle was 10% DMA, 40% PEG, and 50% water (v/v/v) for compounds 1, 2, and 12, 10% DMSO, 40% PEG, and 50% water for compound 15, saline for compound 3, sodium acetate buffer (0.2 M; pH 4.0) for compound 5, and neat DMSO for compounds 6 and 8.
Cell Lines and Monolayer Cultures.
Cell lines were maintained in monolayer culture in 
-MEM containing 5% fetal bovine serum. V79-NTRpuro (originally named T79-A3), a stable NTR transfectant from the Chinese hamster V79-379A cell line (23)
, was passaged in 15 µM puromycin. V79oua is a spontaneous ouabain-resistant mutant cloned from V79-379A by selection in 1 mM ouabain. WiDr-NTRneo (originally named WC14; Ref. 10
) and Skov3-NTRneo (originally named SC3; (10)
are stable NTR transfectants from WiDr and Skov3 human colon and ovarian carcinoma lines and were passaged in 0.3 and 0.6 mg/ml G418, respectively. Monolayer cultures for determination of bystander killing were prepared by seeding cell mixtures (1.5 x 104 cells/well in 1 ml of medium) in 24-well trays and growing for 3 days to give 0.8–1.0 x 106 cells/well. Prodrugs were added, and the plates were returned to the 37°C incubator for 5 h. The monolayers were then washed with fresh medium, harvested by trypsinization, and plated to determine clonogenic cell survival as described below.
Growth of Multicellular Layers.
Multilayers were grown as described previously (16)
, except that pairs of cell lines were mixed before seeding to provide cocultures. Briefly, the Teflon microporous membranes of Millicell CM cell culture inserts (Millipore, Bedford, MA) were coated with collagen to facilitate cell attachment, and inserts were placed in a ring of polyethylene closed cell foam to allow floatation. Trypsinized single cell suspensions of the appropriate composition (106 cells total in 0.5 ml of -MEM with 5% fetal bovine serum, penicillin, streptomycin, and 1 µg/ml amphotericin B) were seeded onto the collagen-coated inserts. Inserts were incubated in a 5% CO2 incubator overnight to allow cell attachment and then submerged by floating beneath a wide-mesh stainless steel grid in a jar containing a large reservoir of the same medium (80 ml/insert, stirred magnetically in a 37°C incubator) and left to grow for 3 days (Chinese hamster cell lines) or 4–5 days (human cell lines). Trypsinization of V79 or WiDr multilayers typically gave 6–8 x 106 cells, whereas Skov3 multilayers yielded approximately 3 x 106 cells.
Immunohistochemistry.
Monolayers (on glass coverslips), multilayers, and tumors were fixed in neutral buffered formalin, and multilayers and tumors were frozen and sectioned (10-µm thickness). NTR was immunostained using a sheep antibody against NTR (Cobra Therapeutics Ltd., Keele, United Kingdom; 1:2000 dilution; 1 h). We confirmed that this antibody detects a single polypeptide of Mr 24,000, identical to purified NTR (supplied by CAMR, Porton Down, United Kingdom), in V79-NTRpuro, WiDr-NTRneo, and Skov3-NTRneo cells by denaturing 10% PAGE. The primary antibody was visualized using FITC-conjugated donkey antirabbit IgG (Vecta Laboratories, Burlingame, CA; 1:50 dilution; 30 min). Sections were viewed at 40x magnification with a Leica TCS 4D confocal microscope using a krypton/argon laser with a FITC filter, and fluorescence intensity was determined using Scion Image (Scion Corp.). Adjoining sections were stained with H&E.
Drug Treatment of Multicellular Layer Cultures.
Multilayers were submerged in 10 ml of medium containing prodrug in glass bottles with lids containing ports allowing continuous gassing. Each insert was held in place using a stainless steel loop fixed to the lid, and the medium was stirred magnetically in a 37°C water bath. The gas phase was 5% CO2:95% O2 to prevent hypoxia in the center of the multilayers (17)
. After incubation for 5 h, multilayers were trypsinized in 0.07% trypsin (Difco Laboratories, Detroit, MI) in citrate-saline with magnetic stirring for 10 min at 37°C. Cells were centrifuged, resuspended in fresh medium, counted, and plated to determine clonogenic cell survival.
Clonogenic Cell Survival.
To discriminate clonogenic activator (NTR+) from target (NTR-) cells, trypsinized single cell suspensions containing V79-NTRpuro and V79oua cells were plated (up to 105 cells/60-mm dish) in nonselective medium (total cells) or in medium containing 1 mM ouabain (V79oua cells) or 15 µM puromycin (V79-NTRpuro cells). Colonies were grown for 6 days and stained with methylene blue (0.2% in 50% ethanol), and colonies containing >50 cells were counted. The selection conditions were validated by performing reconstruction experiments with known ratios of activator and target cells; both cell types could be quantified accurately in the presence of up to a 104-fold excess of the other line (data not shown). The surviving fraction after drug treatment was determined as the plating efficiency for the treated population divided by that for the untreated population from the same set of monolayers or multilayers. The human lines were assayed similarly, using a growth time of 14 days before staining, plating in nonselective media to determine total clonogens, and plating in medium containing G418 (0.3 and 0.6 mg/ml) to determine WiDr-NTRneo or Skov3-NTRneo activators. Survival of targets was determined from the difference between total clonogens and activators. In some experiments, cells were plated in medium containing the diol derivative of CB 1954 (compound 3; 3 µM) to eliminate activator cells. This compound was chosen because of its weak bystander effect (Fig. 4B)
; absence of bystander killing was confirmed at the cell densities used in the clonogenic assay. The two methods generally provided results in close agreement, but use of compound 3 was preferred when the number of surviving cells was similar for targets and activators.
Antitumor Activity.
Tumors were grown in CD-1 homozygous nude mice of either sex by injection of 107 cells (grown as monolayers in culture) s.c. on the dorsum, using WiDr, WiDr-NTRneo, or 1:1 mixtures of the two cell lines. Mice were individually ear tagged and randomized to treatment when the mean of the two largest orthogonal tumor diameters reached 8 mm. Drugs were administered as single i.p. doses at doses up to the MTD, defined as the highest dose causing no drug-related deaths in a group of six mice with mean body weight loss at day 5 of <10%. The injection volume corresponded to 1 µl of DMSO or DMA per gram of body weight or up to 30 µl/g for aqueous vehicles; controls received equivalent doses of the formulation vehicle. For clonogenic assays, mice were killed 18 h after treatment, and tumors were dissected and dissociated to single cells for determination of clonogenic cell survival (as described above) by incubation in Pronase (2.5 mg/ml), collagenase (1 mg/ml) and DNase (0.2 mg/ml) in culture medium for 45 min with magnetic stirring. For determination of tumor growth inhibition, the two largest tumor diameters were measured twice weekly after treatment. The end point was time to regrowth of the tumors to 15 mm in mean diameter. Animals with no evidence of tumor 100 days after treatment were classed as cures. Significance of treatment effects or differences between tumor types were assessed using Cox’s proportional hazard survival analysis.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The proportion of activator and target cells in the mixed cultures of the V79 and WiDr pairs remained constant during growth and was essentially identical to the input ratio, as determined by plating in selective media [e.g., for multilayers initiated using 10% activators, the multilayers comprised 12 ± 1% V79-NTRpuro (n = 11) after 3 days or 13 ± 2% WiDr-NTRneo (n = 14) after 4–5 days]. When grown as s.c. tumors in nude mice, the take rates of WiDr, WiDr-NTRneo, and 1:1 mixtures were >95%, the median times to reach treatment size (0.3 g) were similar (9, 12, and 12 days, respectively), and the proportion of G418-resistant cells determined by plating from the mixed tumors (43 ± 4%; n = 37) was essentially the same as that at inoculation.
The level and distribution of NTR protein were assessed by immunostaining of WiDr-NTRneo monolayers (Fig. 1b)
, multilayers (Fig. 1e)
, and tumors (Fig. 1h)
. Although some heterogeneity between cells was apparent, multilayers grown from 1:1 mixtures of WiDr-NTRneo and WiDr showed obviously greater heterogeneity; NTR mosaicism was readily identifiable as a "Swiss cheese" staining pattern in the multilayer cocultures, with stained regions having typical dimensions of 2–3 cell diameters (Fig. 1d)
. Absolute immunostaining intensity was similar in multilayers and tumors and was approximately twice as high when multilayers and tumors were grown from WiDr-NTRneo only rather than from 1:1 mixtures of WiDr-NTRneo and WiDr (Fig. 1i)
; these conclusions were supported by quantitative Western immunoblotting (data not shown).
Bystander Effects for NTR/CB 1954 in V79 Monolayer and Multilayer Cocultures.
Use of multilayer cocultures to quantify bystander effects is illustrated for the NTR prodrug CB 1954 (1) in Fig. 2
. The method exploits the puromycin resistance of V79-NTRpuro and the ouabain resistance of V79oua to allow independent quantification of survival for both the activator and target populations. After exposing multilayers to CB 1954 for 5 h, multilayers were disaggregated and plated in selective media (Fig. 2A)
. Cytotoxicity was quantified as the prodrug concentration to reduce survival to 10% of controls (C10). Multilayers grown from activators alone were 140-fold more sensitive to CB 1954 than those grown from targets alone (C10 = 10 versus 1400 µM). In cocultures seeded with 3% activators:97% targets, the targets were 8-fold more sensitive than in multilayers of targets alone, demonstrating a clear bystander effect. When the proportion of activators was increased, killing of target cells increased progressively (Fig. 2B)
. This was also true, to a lesser extent, for the activator population, indicating an autobystander effect at high activator density. The magnitude of the bystander effect against NTR- target cells was quantified as the ratio of C10 values for activators:targets (Fig. 2B
, right axis), which reached approximately 0.8 at high activator:target ratios.
|
. Also shown in this figure, bystander killing in monolayers was decreased further by stirring (using monolayers grown on the same permeable support membranes as multilayers). The effect of stirring is due to dilution of drug metabolite(s) into the bulk medium rather than any contribution of hypoxia to activation of CB 1954 in the absence of stirring because there was no difference in the unstirred C10 using 20% versus 95% O2 in the gas phase (data not shown). Comparison of stirred and unstirred multilayers was not undertaken because of the complicating effects of hypoxia in the absence of stirring, but increasing the volume of medium in which the multilayers were stirred (from 10 to 100 ml) caused no change in target cell killing (Fig. 2C)
, despite a higher medium:cell ratio for multilayers in 100 ml than for the monolayers.
Greater Bystander Effect for SN 23862 than CB 1954 in Multilayer Cocultures.
The bystander effect for activation, by NTR, of CB 1954 (1) and the corresponding nitrogen mustard SN 23862 (2) was compared in multilayer cocultures of the three activator/target cell line pairs (Fig. 3)
. In each case, mixtures were seeded with 10% activator cells, giving cocultures with an average of 9.5%, 13%, or 3.5% activators for V79, WiDr, and Skov3 cells, respectively. In all three genetic backgrounds, CB 1954 provided a bystander effect, but with C10(activator:target) ratios well below unity (values of 0.4, 0.18, and 0.12 in the V79, WiDr, and Skov3 models, respectively). SN 23862 also gave very large differentials between activators and pure cultures of targets. However, in the cocultures, the target cells were as essentially as sensitive to SN 23862 as the activators, with C10(activator:target) ratios of 0.9, 1.0, and 1.0 for the V79, WiDr, and Skov3 models, respectively. Thus, SN 23862 (2) shows a much more efficient bystander effect than CB 1954 (1) in multilayer cocultures in all three genetic backgrounds.
|
except the basic CB 1954 analogue 5) with a wide range of measured octanol:water partition coefficients (P). Because efficient cytotoxic activation by NTR is a precondition for testing bystander efficiency, only compounds showing a large differential in cytotoxicity between NTR+ and NTR- cells were included in the study. (For all 15 compounds in Fig. 4A
, ratios of antiproliferative potency (IC50) were at least 40-fold for V79-NTRpuro and WiDr-NTRneo versus their NTR- counterparts; data not shown). Bystander effects were measured as described above, in V79 multilayers containing 3% activators, and (for a subset of seven compounds) for WiDr multilayers containing 10% activators. In both cases, the bystander effect, quantified as log C10(activators:targets), showed a highly significant linear correlation with logP, with prodrug lipophilicity accounting for 80% and 86% of the variance in the bystander effect in the two cell line pairs, respectively (Fig. 4B)
.
Relationship between Bystander Effects in Multilayers and Tumors.
We tested whether bystander effects in the multilayer model are predictive of those in tumors by measuring killing of activators and targets in mixed WiDr/WiDr-NTRneo tumors, using a subset of prodrugs with widely differing bystander efficiency in multilayers. To facilitate comparison with the in vitro data, tumors were treated using a single dose of prodrug, and response was assessed using identical clonogenic assays to quantify activators and targets [with removal of tumors 18 h after treatment; additional experiments (not shown here) demonstrated that killing by CB 1954 was complete within 12 h]. In the in vivo studies, variability between individual tumors made it difficult to define dose-response relationships as precisely as in multilayers. Therefore, activator and target cell killing was compared by plotting the surviving fraction for targets against that for activators in the same tumor as illustrated for four prodrugs in Fig. 5A
, right panels. To facilitate comparison, the WiDr/WiDr-NTRneo multilayer data are plotted in the same manner (Fig. 5A
, left panels). A similar pattern is seen for both multilayers and tumors, with a relatively low bystander efficiency (inefficient killing of targets relative to activators) for the hydrophilic CB 1954 and mesylate mustard 8 and a higher bystander efficiency for the lipophilic chloromustard 2 (SN 23862) and bromomustard 6 (SN 24927).
|
, with an explained variance of 84%. Thus, the multilayer model has considerable predictive power in the selection of prodrugs with high bystander efficiency in tumors.
Measured in this way, the bystander effect was always greater in tumors than in multilayers. However, in these studies, the tumors were grown using 1:1 mixtures of activators and targets, whereas the multilayers were grown using a 1:9 ratio. In an additional experiment with WiDr multilayers, the activator:target seeding ratio was changed to 1:1, giving 63 ± 2% puromycin-resistant activator cells after 5 days. Treatment with CB 1954 as described above gave C10 = 12 µM for activators and C10 = 22 µM for targets (compare 33 and 250 µM, respectively, in the 1:9 ratio multilayers), with a SF(A/T) of 0.14. The latter value is indistinguishable from that in 1:1 mixed tumors (95% confidence interval, 0.10–0.24; Fig. 5B
).
Identification of an NTR Prodrug with Antitumor Activity Superior to CB 1954.
The above experiments were designed specifically to compare bystander effect efficiency in isolation from other considerations (such as potency or therapeutic ratio), and the analysis excluded tumors for which killing in the activator population was too large to measure by clonogenic assay. To compare bystander effects over a wider dynamic range of cell killing and to evaluate overall antitumor activity in relation to host toxicity, tumor growth inhibition was examined for the four prodrugs giving widely different bystander efficiencies in the clonogenic assay in Fig. 5A
. Each was administered as a single i.p. injection at the MTD. None significantly inhibited the growth of WiDr tumors (data not shown), whereas all showed marked activity against 100% WiDr-NTRneo tumors; at least 80% of animals were free of tumors 100 days after treatment by CB 1954 or its bromo or mesylate mustard analogues 6 or 8 (Fig. 6)
. The chloromustard SN 23862 (2) showed lesser but still highly significant (P < 0.01) activity.
|
8), although the striking curative activity of 6 in mixed as well as pure WiDr-NTR tumors means that its bystander efficiency could not be quantified in the growth inhibition assay. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A difficulty with attempting to model bystander effects in conventional cell cultures is that the extracellular volume fraction is typically orders of magnitude greater than that in tumors. In such a setting, bystander effects are likely to be dominated by accumulation of metabolites in the bulk medium. The multicellular layer model developed by ourselves (16 , 18 , 26) and others (27 , 28) provides an opportunity to investigate bystander effects at tissue-like cell densities. Multilayers faithfully represent key aspects of the extravascular compartment of tumors, including the presence of noncycling cells arising through nutrient and oxygen depletion (27) , and explicitly test the ability of activated metabolites to diffuse through tumor tissue.
Expression of NTR at levels providing substantial sensitization to CB 1954 (see Figs. 5
and 6
) did not significantly modify cell growth in WiDr monolayers, multilayers, or tumors; lack of effect of NTR expression on tumorigenicity of human tumor xenografts has been noted previously (12)
. In addition, NTR expression levels were unchanged by growth as multilayers or tumors (Fig. 1)
. These features facilitate comparison of bystander effects in the three experimental models. Use of selectable markers on both cell types in mixed cultures made it possible to quantify the sensitivity of the activators as well as the target cells, thereby providing a clear picture of bystander efficiency (i.e., the extent to which the survival curve of target cells is shifted toward that of the activators themselves).
Importantly, the bystander efficiency of CB 1954 is much greater in multilayers than in monolayers under equivalent conditions (Fig. 2C)
. The independence of bystander efficiency in multilayers on the cell:medium ratio demonstrates that bystander killing in multilayers is due to local diffusion of cytotoxic metabolite(s) rather than accumulation of metabolites in the bulk medium. This provides a technical advantage with respect to monolayer cultures in which the magnitude of bystander effects will depend on the extent of convective disturbance and mixing, as illustrated by the further lowering of bystander efficiency when the medium was stirred in monolayer experiments (Fig. 2C)
. At a more fundamental level, multilayers provide a physiologically relevant model in which the tissue diffusion ability of the metabolite(s) responsible for the bystander effect is tested explicitly (rather than simply its accumulation in extracellular medium). In this regard, we note that an earlier study of dinitrobenzamide prodrugs using monolayer cocultures of V79 and V79-NTRpuro lines ranked the bystander effect in the order 2 < 1 < 6 (23)
, in contrast to the order 1 < 6 2 based on multilayer cocultures in the present study.
The bystander efficiency for dinitrobenzamide activation by NTR in the multilayer model showed a strong correlation with logP of the prodrugs in both the V79 and WiDr backgrounds (Fig. 4B)
. This supports the hypothesis that the lipophilicity of the diffusing metabolite (for which the logP of the prodrug is assumed to be a surrogate) is a major determinant of the bystander effect in this series, with more lipophilic metabolites better able to diffuse passively through the plasma membrane. This is consistent with earlier suggestions that the bystander effect for CB 1954/NTR is due to passive diffusion of metabolites (14)
. We are currently identifying extracellular metabolites of dinitrobenzamide prodrugs and measuring their logP values, reactivities, and stabilities to extend understanding of the physicochemical determinants of these bystander effects.
A key objective of the present study was to assess whether bystander effects in multicellular layers are predictive of those in tumor xenografts grown from the same cell lines. The bystander effect, quantified by the parameter SF(A/T), which is a measure of the target cell killing at constant (2 log) activator killing, showed highly significant correlation between multilayers and tumors (Fig. 5B)
. When differences in the proportion of activators was taken into account (see "Results"), even the absolute values for bystander efficiency showed close agreement in multilayers and tumors; this is surprising given that multilayers cannot be expected to model tumors quantitatively in every relevant aspect. In particular, washout of activated metabolites (from perivascular cells by blood flow in tumors and from cells at the medium interface for multilayers) could differ. Furthermore, the absolute magnitude of bystander effects must depend in part on the spatial heterogeneity of NTR expression. NTR immunostaining patterns suggest clearer demarcation between NTR+ and NTR- regions in multilayers than in tumors (Fig. 1
, d versus g), possibly because of greater spatial mixing during the longer growth time of tumors. Nonetheless, the good agreement in absolute values and ability to rank prodrugs for bystander effects in tumors gives confidence that there are no major factors being ignored by the multilayer model in the NTR/dinitrobenzamide system. We plan to test the utility of the model further by examining other enzyme/prodrug paradigms with a range of potential bystander mechanisms, including gap-junction-mediated transfer of metabolites (2)
. In situations where immune or vascular mechanisms contribute to bystander killing (29)
, a simple correlation between multilayers and tumors cannot be expected, but the multilayer model would nonetheless be of value in quantifying the component due to cytotoxic metabolite diffusion.
Diffusion of activated metabolites is only one determinant of the therapeutic utility of a GDEPT prodrug; overall antitumor activity will also depend on the efficiency of metabolic activation in transduced cells at well-tolerated prodrug doses (determined by host toxicity, tumor pharmacokinetics, enzyme kinetics of prodrug activation, and potency of activated metabolites). Overall therapeutic activity, assessed as tumor growth inhibition in WiDr xenografts (Fig. 6)
, confirmed the excellent activity of CB 1954 against WiDr-NTRneo tumors, as reported previously for other NTR-transfected human tumor xenografts (11
, 12)
, and confirmed the existence of a substantial bystander effect in tumors (12)
, as demonstrated by large antitumor effects against tumors comprising 50% NTR- cells. The tumor growth assays also demonstrated, for the first time, the strong therapeutic activity of the nitrogen mustard analogues of CB 1954 (especially 6 and 8) against NTR-expressing tumors. For the four prodrugs evaluated in this assay, bystander effects (as judged by comparison of activity against 100% WiDr-NTRneo tumors versus WiDr/WiDr-NTRneo mixtures) were consistent with the clonogenic assay experiments, with 2 superior to 1 and 8. The bromomustard SN 24927 (6) also showed a large bystander effect (efficient killing of target cells in 1:1 mixed tumors) with 7 of 7 animals tumor free at 100 days (versus 15 of 24 animals cured with CB 1954). This improvement presumably reflects the high bystander effect of 6 (Figs. 4
and 5
) as well as the efficiency with which it is activated by NTR as demonstrated previously by its high potency and selectivity for NTR-expressing cells (23)
. SN 24927 (6) is thus a key lead compound for NTR-based GDEPT.
In conclusion, the current study demonstrates the utility of the multicellular layer model for quantifying bystander effects in GDEPT and has assisted in the identification of a dinitrobenzamide prodrug of NTR (SN 24927; 6) with superior therapeutic activity in human tumor xenografts comprising mixtures of NTR+ and NTR- cells. More generally, genetic and environmental heterogeneity in tumors confounds development of selective anticancer agents, making optimization of bystander effects of central importance in experimental oncology. Multicellular layer cultures provide a technically simple and robust method for growing 3D cocultures with a well-defined geometry in which heterogeneity can be controlled, thus providing an ideal system for quantifying bystander effects. In addition, multilayers can be used to test other critical steps in drug action not represented in low-density cultures including the ability of the prodrug to diffuse in the extravascular compartment to reach activator cells (16 , 18 , 30 , 31) . With the increasing use of high-throughput screening systems for front-line drug discovery (32 , 33) , more physiological tissue culture systems, such as the multicellular layer model, are needed to avoid the developing in vivo bottleneck in anticancer drug development as well as to minimize use of animals in research.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by grants from the Health Research Council of New Zealand and Cancer Research Campaign Technology Ltd., United Kingdom. 
2 To whom requests for reprints should be addressed. Phone: 64-9-373-7599, ext. 6883; Fax: 64-9-373-7571; E-mail: wr.wilson{at}auckland.ac.nz.
3 The abbreviations used are: GDEPT, gene-dependent enzyme-prodrug therapy; DMA, dimethylacetamide; MTD, maximum tolerated dose; NTR, Escherichia coli nitroreductase; PAE, prodrug-activating enzyme; PEG, polyethylene glycol; 3D, three-dimensional; SF(A/T), surviving fraction(activators:targets).
Received 10/16/01. Accepted 1/ 4/02.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. A. Helsby, M. A. Goldthorpe, M. H. Y. Tang, G. J. Atwell, E. M. Smith, W. R. Wilson, and M. D. Tingle Influence of Mustard Group Structure on Pathways of in Vitro Metabolism of Anticancer N-(2-Hydroxyethyl)-3,5-dinitrobenzamide 2-Mustard Prodrugs Drug Metab. Dispos., February 1, 2008; 36(2): 353 - 360. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. V. Patterson, D. M. Ferry, S. J. Edmunds, Y. Gu, R. S. Singleton, K. Patel, S. M. Pullen, K. O. Hicks, S. P. Syddall, G. J. Atwell, et al. Mechanism of Action and Preclinical Antitumor Activity of the Novel Hypoxia-Activated DNA Cross-Linking Agent PR-104 Clin. Cancer Res., July 1, 2007; 13(13): 3922 - 3932. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-Q. Ling, B. Li, C. Zhang, D.-Z. Zhu, X.-Q. Huang, W. Gu, and S.-X. Li Inhibitory effect of recombinant adenovirus carrying melittin gene on hepatocellular carcinoma Ann. Onc., January 1, 2005; 16(1): 109 - 115. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Kunz-Schughart, J. P. Freyer, F. Hofstaedter, and R. Ebner The Use of 3-D Cultures for High-Throughput Screening: The Multicellular Spheroid Model J Biomol Screen, June 1, 2004; 9(4): 273 - 285. [Abstract] [PDF]
|
|
|
|
 |
|
 J. I. Grove, A. L. Lovering, C. Guise, P. R. Race, C. J. Wrighton, S. A. White, E. I. Hyde, and P. F. Searle Generation of Escherichia Coli Nitroreductase Mutants Conferring Improved Cell Sensitization to the Prodrug CB1954 Cancer Res., September 1, 2003; 63(17): 5532 - 5537. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
, WiDr; 
, 1:1 mixture of WiDr/WiDr-NTRneo; 
, WiDr-NTRneo.
