Clinical Cancer Research CR Surrogrates AACR Membership
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
Clinical Cancer Research 14, 4259, July 1, 2008. doi: 10.1158/1078-0432.CCR-07-4800
© 2008 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Friedlos, F.
Right arrow Articles by Springer, C. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Friedlos, F.
Right arrow Articles by Springer, C. J.

Cancer Therapy: Preclinical

Attenuated Salmonella Targets Prodrug Activating Enzyme Carboxypeptidase G2 to Mouse Melanoma and Human Breast and Colon Carcinomas for Effective Suicide Gene Therapy

Frank Friedlos1, Panos Lehouritis1, Lesley Ogilvie1, Douglas Hedley1, Lawrence Davies1, David Bermudes3, Ivan King3, Jan Martin1, Richard Marais2 and Caroline J. Springer1

Authors' Affiliations: 1 Cancer Research UK Centre for Cancer Therapeutics at The Institute of Cancer Research, Sutton, Surrey, United Kingdom; 2 Cancer Research UK Centre for Cell and Molecular Biology at The Institute of Cancer Research, London, United Kingdom; and 3 Vion Pharmaceuticals, Inc., New Haven, Connecticut

Requests for reprints: Caroline J. Springer, Cancer Research UK Centre for Cancer Therapeutics at The Institute of Cancer Research, 15 Cotswold Road, Sutton, Surrey SM2 5NG, United Kingdom. Phone: 44-20-8722-4214; Fax: 44-20-8722-4046; E-mail: caroline.springer{at}icr.ac.uk.


   Abstract
Top
 Abstract
Materials and Methods
 Results and Discussion
 Disclosure of Potential...
 References
 
Purpose: We engineered the oncolytic Salmonella typhimurium–derived bacterium VNP20009 as a vector to target delivery to tumors of the prodrug-activating enzyme carboxypeptidase G2 (CPG2) and to show enhanced antitumor efficacy on administration of different prodrugs.

Experimental Design: We characterized CPG2 expression in vectors by immunoblotting, immunofluorescence, and enzyme activity. We assessed prodrug activation by high-performance liquid chromatography. Target human tumor cell and bacterial vector cell cytotoxicity was measured by flow cytometry and colony-forming assays. Therapy was shown in two human tumor xenografts and one mouse allograft with postmortem analysis of bacterial and CPG2 concentration in the tumors.

Results: CPG2 is expressed within the bacterial periplasm. It activates prodrugs and induces cytotoxicity in human tumor cells but not in host bacteria. Following systemic administration, bacteria multiply within xenografts reaching 2 x 107/g to 2 x 108/g at 40 days postinoculation. The concentration of CPG2 in these tumors increases steadily to therapeutic levels of 1 to 6 units/g. The bacteria alone reduce the growth of the tumors. Subsequent administration of prodrugs further reduces significantly the growth of the xenografts.

Conclusions: The bacteria multiply within tumors, resulting in a selective expression of CPG2. The CPG2-expressing bacteria alone reduce the growth of tumors. However, in the presence of prodrugs activated by CPG2, this oncolytic effect is greatly increased. We conclude that bacterial oncolytic therapy, combined with CPG2-mediated prodrug activation, has great potential in the treatment of a range of cancers.

In suicide gene therapy, a variety of vectors has been used to introduce the genes encoding prodrug-activating enzymes selectively to tumor cells, rendering them sensitive to a prodrug, avoiding the toxicity usually associated with cytotoxic chemotherapy. These include viruses, chemical complexes, naked DNA, and bacteria (1, 2). Most studies of bacterial vectors have involved Clostridium sp. spores with the innate property of germinating and proliferating only in the hypoxic milieu of the necrotic center of tumors (3, 4), although the bacterial species Bifidobacteria, Lactobacillus, and Caulobacter have also been employed in this way (5, 6). We report the use in an analogous context of the facultative anaerobic bacterium VNP20009 (7, 8). A similar Salmonella-derived vector has been used to target small interfering RNA (9). Our VNP20009 vector is a derivative of Salmonella typhimurium (ATCC 14028), attenuated by disruption of the genes msbB and purI. The msbB gene regulates the addition of a terminal myristoyl group to lipid A, which otherwise induces tumor necrosis factor-{alpha}, causing septic shock in infected patients (10). The purI gene is involved in purine biosynthesis. Thus, purI deletion mutants are auxotrophic for purines but have been shown to proliferate in tumors (11), although the mechanism is unknown. Animal studies (12) and phase I clinical studies (13) have established the safe doses for systemic injection of VNP20009. The maximum tolerated dose was 3 x 108 colony-forming units (cfu)/m2, with dose-limiting toxicities encountered in patients receiving 1 x 109 cfu/m2, and focal tumor colonization observed in 3 of 24 patients (13). In the preclinical studies, intratumoral bacterial proliferation occurred to levels of 109 cfu/g, 10,000-fold higher than in the liver (the next most colonized tissue; ref. 10), these levels being too low to generate systemic toxicity. As a consequence of this selective colonization, perhaps involving immune responses (14), VNP20009 exerts an oncolytic effect and is effective in dogs with spontaneous cancers (15). We sought to enhance the efficacy further by engineering the bacteria to express the prodrug-activating enzyme carboxypeptidase G2 (CPG2; E.C. 3.4.17.11).

The gene encoding CPG2 was originally cloned from Pseudomonas sp. strain RS-16 (16). The enzyme is a Zn2+-dependent 83-kDa homodimer. CPG2 cleaves folic acid to pteroic acid and also activates a range of bifunctional alkylating agent and antibiotic prodrugs (17–22). CPG2 possesses several properties that make it suitable for suicide gene therapy. There is no human analogue, and no cofactors are required; it catalyzes a one-step activation, produces drugs that are cytotoxic to both cycling and quiescent cells, and can activate a large range of prodrugs. CPG2 catalyzes the hydrolysis of a deactivating glutamate moiety from various nitrogen mustard prodrugs to yield bifunctional alkylating agents that are more reactive than the prodrug and, being less polar, have greater penetration into cells where they produce DNA-DNA interstrand cross-links and cell death. A large number of these aromatic nitrogen mustard prodrugs have been synthesized. They vary in the leaving group of the mustard moiety, the nature of the moiety para to the mustard (phenol, aniline, or benzoate), and substitutions to the aromatic ring. Here, we focus on these prodrugs that are cleaved to benzoic acid mustards that were effective against xenografts engineered to express CPG2 (23). They vary in the mustard-leaving groups and in degrees of fluorination of the aromatic ring.

The potential of CPG2 in combination with prodrug 2 (4-[(2-chloroethyl)(2-methylsulphonyloxyethyl)amino]benzoyl-L-glutamic acid) was first shown in antibody-directed enzyme prodrug therapy (24). CPG2 plus another prodrug, 1 (4-[bis(2-iodoethyl)amino]-phenyloxycarbonyl-L-glutamic acid), is the second antibody-directed enzyme prodrug therapy system to reach clinical trials (25). The safe dose of 4-[bis(2-iodoethyl)amino]-phenyloxycarbonyl-L-glutamic acid depended on dose of CPG2 antibody and time allowed for blood clearance, but typically 537.6 mg/m2 was safe when circulating enzyme was <0.005 units CPG2/mL, giving rise to an AUCinf of 3,408 µg/mL min, dose-limiting toxicities being liver, kidney, and marrow (25). Gene-directed enzyme prodrug therapy with CPG2 has also been developed and we have recently described a conditionally replicating oncolytic adenoviral vector that leads to expression of CPG2 selectively in telomerase-positive tumor cells. In combination with prodrug 1, this adenovirus induced tumor regressions and cures in xenograft models of human hepatocellular carcinoma (26) and colon carcinoma (27).

A crucial beneficial aspect of these therapies is the presence of a "bystander effect" where nontargeted tumor cells are killed by activated drug diffusing away from cells expressing CPG2. This is a powerful feature of the CPG2-activated drugs and was shown both in vitro with prodrugs 2 (28) and 1 (29) and in vivo in both transfected xenografts with prodrug 2 (30) and prodrugs 1, 3, and 4 (23) and virally targeted xenografts with prodrug 1 (27).

We rationalized that VNP20009 could deliver high concentrations of CPG2 selectively to a wide range of tumor types for activation of a range of prodrugs. Here, we describe the generation, properties, and efficacy of the oncolytic bacterium VNP20009, engineered as a vector to express CPG2 selectively in tumors, where activation of prodrugs produces a therapeutic benefit.


   Materials and Methods
Top
 Abstract
 Materials and Methods
Results and Discussion
 Disclosure of Potential...
 References
 
Prodrugs
The synthesis and properties of the prodrugs 4-[bis(2-iodoethyl)amino]-phenyloxycarbonyl-L-glutamic acid (1), 4-[(2-chloroethyl)(2-methylsulphonyloxyethyl)amino]benzoyl-L-glutamic acid (2), 3-fluoro-4-[bis(2-chlorethyl)amino]benzoyl]-L-glutamic acid (3), and 3,5-difluoro-4-[bis(2-iodoethyl)amino]benzoyl-L-glutamic acid (4) and the cognate drugs 4-[(2-chloroethyl)(2-methylsulfonyloxyethyl)amino]benzoic acid (2a) and 3,5-difluoro-4-[bis(2-iodoethyl)amino]benzoic acid (4a) have been described (22, 31–33).

Cell lines
MDA-MB-361 (human breast carcinoma), WiDr (human colon carcinoma), and B16-F10 (mouse melanoma) cell lines were all cultured in DMEM (Life Technologies) plus 10% fetal bovine serum (Life Technologies) at 37°C in a 5% CO2 atmosphere and passaged conventionally.

Bacterial strains and culture conditions
A S. typhimurium derivative named VNP20009 (purI-, msbB-), selected to be comparatively nontoxic and to preferentially colonize tumors, was initially employed to express periplasmic CPG2 (ppCPG2). An asd- derivative of VNP20009, VNP20009asd- (purI-, msbB-, asd-; Vion Pharmaceuticals), provided effectively stable expression of ppCPG2 in the absence of antibiotic. These strains were propagated aerobically in special Luria-Bertani broth (without NaCl and supplemented with 2 mmol/L CaCl2 and 2 mmol/L MgCl2) at 37°C shaken at 300 rpm or on special Luria-Bertani agar plates. If required, the broth was supplemented with the antibiotic carbenicillin at 50 µg/mL. VNP20009asd- lacking pYA3332 was grown by supplementing the culture with diaminopimelic acid 50 µg/mL. A Salmonella strain deficient in restriction endonucleases with intact methylation (YS501) was used to passage plasmids to provide the methylation necessary for transforming VNP20009.

Generation of constructs
The expression vector for ppCPG2 was constructed using the plasmid pTrc99A containing the lac/trp promoter (Pharmacia; ref. 34) and the β-lactamase gene, which mediates ampicillin/carbenicillin resistance (pTrcOMPAppCPG2). CPG2 resides in the periplasm of Pseudomonas sp. and is directed to this location by a 20–amino acid leader sequence (35). We replaced the wild-type CPG2 leader sequence with the leader sequence of the gene ompA that has previously been used very efficiently to direct recombinant proteins to the periplasm of Gram-negative bacteria. This was done to ensure that CPG2 was correctly exported to the periplasm of Salmonella, as the Pseudomonas signal peptide has not been documented to work in these cells (36). The plasmid was used to generate a CPG2-expressing strain (ppCPG2-VNP20009) and a nonexpressing empty vector (EV-VNP20009). An improved version designed for selection-free stable expression was based on a balanced lethal system (37). The host, VNP20009asd-, lacks the vital asd gene and only survives on medium supplemented with diaminopimelic acid or when transformed by the plasmid pYA3332, which supplies the missing asd gene. We engineered pYA3332 to express ppCPG2 and transformed it into VNP20009asd- producing the new strain ppCPG2-VNP20009asd-.

Immunoblotting
SDS-PAGE gels were transferred to an Immobilon-P membrane (Millipore) and the membranes were blocked with 5% nonfat dried milk powder. Immunodetection of CPG2 by a previously described (29) rabbit polyclonal antiserum was followed by a donkey anti-rabbit antibody conjugated to horseradish peroxidase. Detection was by enhanced chemiluminescence reagent (Amersham Pharmacia).

Determination of CPG2 enzymic activity in bacteria
Bacteria were lysed with a nonionic detergent ("Bug Buster" Novagen) and CPG2 activity in the extract was determined by a kinetic method using methotrexate as substrate (29). Briefly, aliquots were added to cuvettes containing assay buffer and methotrexate, and the rate of decrease in A320 was measured spectrophotometrically. CPG2 (1 unit) is defined as the amount that degrades 1 mmol methotrexate/min.

Immunofluorescence
Bacteria were harvested by centrifugation, washed once in PBS, and fixed with 3% paraformaldehyde. For permeabilization, cells were resuspended in 0.1% Triton X-100, washed, and resuspended in PBS/lysozyme (100 µg/mL)/5 mmol/L EDTA. Samples were incubated with anti-CPG2 rabbit antiserum followed by labeled anti-rabbit antibody (Alexa 488 nm; Molecular Probes/Invitrogen). The DNA was counterstained with TO-PRO-3 dye (Molecular Probes/Invitrogen) and the samples were viewed using a confocal microscope.

Cell fractionation
Bacteria were lysed using osmotic shock. Briefly, bacteria were resuspended in a hyperosmotic solution containing 20% buffered sucrose (pH 8) on ice for 10 min, centrifugally pelleted, resuspended in plain buffer, and incubated on ice for another 10 min at which stage the outer cell membrane ruptures. The inner cell membrane and cytoplasm (spheroplasts) were centrifugally harvested at 4,000 rpm. The periplasm and outer membrane in the supernatant were harvested at 50,000 x g for 30 min. The amount of CPG2 in each subcellular fraction was determined by immunoblotting. Gel loading was standardized by adjusting the percentage of the total harvest for each subcellular fraction.

Assessment of cytotoxicity
Cell cycle analysis by flow cytometry. WiDr cells were seeded in six-well plates (1 x 106 per well) 16 h before addition of prodrug 1 (40 µmol/L) and bacterial vectors at concentrations as indicated. After 90 min, the cells were washed three times with fresh medium containing penicillin (12 µg/mL), streptomycin (20 µg/mL), and chloramphenicol (10 µg/mL) and incubated for a further 2 days. This dose-time of 40 µmol/L x 90 min = 3600 µmol/L min is commensurate with the AUC of 5,763 µmol/L min resulting from the safe dose of 537.6 mg/m2 used in phase I clinical trial (25). Adherent and detached cells were harvested, washed in PBS, and fixed in 70% ethanol. The cells were washed in PBS and resuspended in PBS containing RNase (10 µg/mL) and propidium iodide (40 µg/mL). The samples were incubated at 37°C for 30 min and analyzed by flow cytometry (FACSCalibur; Becton Dickinson; software package: CellQuest). Data from 20,000 cells were plotted as counts (abscissa) versus fluorescence intensity (ordinate).

Colony-forming assay. WiDr cells were exposed to prodrug in the presence or absence of vectors as above. After the initial 90-min exposure, cells were replated in antibiotic-containing medium at limiting dilutions, allowed 14 days growth, fixed in 70% cold ethanol, and stained with crystal violet (1%, w/v). Colonies were scored and cell survival was expressed as percentage of control.

Determination of prodrug activation in ppCPG2-VNP20009asd-
VNP20009 or ppCPG2-VNP20009asd- was suspended in PBS/1% glucose/2 mmol/L MgSO4, and prodrug 2 was added to 1 mmol/L. Aliquots were spun down and the supernatants were frozen in liquid nitrogen. The samples were quickly thawed, and aliquots (50 µL) were injected onto a C18 column (Whatman Partisphere 4.6 x 100 mm), eluted with a gradient of 20% to 90% methanol/100 mmol/L ammonium acetate (pH 7) over 15 min, and monitored from 250 to 350 nm. Chromatograms were integrated at 300 nm.

In vivo therapy
All experiments were conducted in accordance with UK Home Office regulations and UK Coordinating Committee on Cancer Research guidelines (38). Xenografts or allografts were established in nude (nu/nu) female CD1 or C57 black female mice (20-22 g) by s.c. inoculation (0.2 mL) in the right flank of MDA-MB-361 human breast carcinoma or WiDr human colon carcinoma cells (107 and 8 x 106, respectively) or B16-F10 mouse melanoma cells (4 x 105) in PBS. After 8 (WiDr), 10 (B16-F10), or 12 (MDA-MB-361) days, mice were allocated by stratified distribution into control and treated groups. ppCPG2-VNP20009asd- (2 x 106 cfu) were diluted in PBS/2 mmol/L MgSO4 and injected into the tail veins. Prodrugs were dissolved in DMSO and diluted 20-fold in 1.26% (w/v) sodium bicarbonate just before injection. Each course of prodrug treatment consisted of three i.p. injections over a 24-h period to a total preestablished maximum tolerated dose of 1,500, 1,200, and 600 mg/kg for prodrugs 2, 3, and 4, respectively. Courses of prodrug were administered on days 7, 14, 28, 35, and 42 (±1 day) after vector. Animals were culled if the tumor exceeded 1.5 cm in any dimension, or if a scab formed, as demanded by the UK Home Office license requirement (38). At cull, tumors were excised and analyzed for CPG2 enzyme activity and for the presence of colony-forming bacterial cells. Similar measurements were performed in separate time-course experiments and the data were pooled.

Determination of CPG2 enzyme activity and bacterial cfu in tumors
Tumors were homogenized in PBS/2 mmol/L MgSO4 at 10% (w/v). CPG2 enzyme activity was determined as described previously (39). Briefly, homogenates were incubated with methotrexate for 30 min, and the reaction was stopped with acidified methanol. The concentration of the reaction product diaminomethylpteroic acid in the centrifuged supernatant was determined by high-performance liquid chromatography. An external calibration curve was constructed by spiking control homogenates with standard enzyme. Bacterial cfu were determined by serially diluting homogenates in growth medium and plating on agar plates. Dilutions that gave 10 to 200 colonies after overnight culture were scored. The results were expressed as enzyme units or cfu per wet weight gram of tumor.

Statistical analysis
All analyses were performed using the functions described in Results using GraphPad Prism version 4.02 for Windows, GraphPad Software. Computed values are quoted with their 95% confidence intervals in brackets.


   Results and Discussion
Top
 Abstract
 Materials and Methods
 Results and Discussion
Disclosure of Potential...
 References
 
Characterization of ppCPG2. We transformed VNP20009 with the plasmid pTrcOMPAppCPG2 encoding ppCPG2 and ampicillin resistance, generating ppCPG2-VNP20009. Expression of CPG2 of expected monomeric apparent molecular weight (~42 kDa) in transformed bacteria was confirmed by immunoblotting (Fig. 1A ) and enzymic activity by methotrexate digestion (Fig. 1B). The subcellular localization of CPG2 in ppCPG2-VNP20009 was verified by immunofluorescence of intact or permeabilized cells (Fig. 1C). Staining was observed only after permeabilization, indicating internal expression of ppCPG2. The majority of the CPG2 was present in the periplasm as confirmed by immunoblotting of the cytoplasmic, periplasmic, and outer membrane subcellular fractions (Fig. 1D).


Figure 1
View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 1. A, immunoblot of CPG2 expression. A band with the expected monomeric molecular weight of 42 kDa is seen in lysates of ppCPG2-VNP20009 but not in EV-VNP20009. B, CPG2 enzyme activity against methotrexate as substrate in extracts of ppCPG2-VNP20009 and EV-VNP20009. C, immunofluorescence micrographs of anti-CPG2 staining in ppCPG2-VNP20009 and EV-VNP20009 with or without cell permeabilization. Green, anti-rabbit Alexa 488; blue, TO-PRO-3 nuclear counterstaining. D, immunoblots of subcellular fractions from a pellet of ppCPG2-VNP20009. Gel loadings were 5% of the total preparation for the cytoplasm and outer membrane fractions and 2.5% for the periplasm fraction.

 
Induction of cytotoxicity by ppCPG2-VNP20009 plus prodrug 1. The ability of ppCPG2-VNP20009 to mediate prodrug-specific cytotoxicity was assessed by flow cytometry. Human colon carcinoma cells (WiDr) were treated with increasing doses of ppCPG2-VNP20009 or the empty vector EV-VNP20009 with or without prodrug 1. The cells were subsequently stained with propidium iodide for flow cytometric cell cycle analysis. The majority of the control cells or cells treated with EV-VNP20009 were in G1 (Fig. 2A, left ). There was a shift in the profile of cells treated with ppCPG2-VNP20009 plus prodrug 1 (Fig. 2A, right). The majority of the cells are in mid-late S-G2-M, with an increased proportion seen in a sub-G1 population arising from dead cells. This pattern is characteristic of the disruption of the cell cycle that precedes cell death in cells exposed to DNA cross-linking agents (40). The perturbation was responsive to bacterial doses from 1 x 106 to 2.5 x 107 cfu. At higher doses of ppCPG2-VNP20009 from 2.5 x 107 to 6.25 x 108 cfu, the majority of cells (>95%) were in non-G1 compartments (Fig. 2B). EV-VNP20009 bacteria produced no such effect. No cell cycle shift was observed with supernatant bacterial medium, showing that CPG2 is not shed into the medium (data not shown).


Figure 2
View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 2. A, flow cytograms of DNA content (abscissa) against frequency (ordinate) in WiDr human colorectal carcinoma cells that have been exposed to ppCPG2-VNP20009 plus prodrug 1. The regions corresponding to the cell cycle compartments of interest are marked. B, proportion of WiDr cells in non-G1 cell cycle compartments as a function of dose of ppCPG2-VNP20009 (shaded columns) or EV-VNP20009 (hatched columns) when treated with ppCPG2-VNP20009 plus prodrug 1 as in A. The additional two grayed out columns at 6.25 x 108 bacteria per well indicate the results of the corresponding bacterium without prodrug. C, percentage of WiDr cells giving rise to colonies 14 d after treatment as in A followed by replating in antibiotic-containing medium. The additional grayed out columns at 6.25 x 108 bacteria per well indicate results of ppCPG2-VNP20009 without prodrug (shaded column) or EV-VNP20009 plus prodrug (hatched column).

 
We then performed a clonogenic survival assay. WiDr cells were treated with ppCPG2-VNP20009 in the presence or absence of prodrug 1, replated at limiting dilutions, and allowed to form colonies. Tumor cells treated with prodrug alone or with EV-VNP20009 in the presence of prodrug 1 showed no effect on cell survival (Fig. 2C). In cells treated with ppCPG2-VNP20009 plus prodrug 1, we detected a bacterial dose-responsive reduction in cell viability, illustrating prodrug activation-specific cytotoxicity. Thus, we show that ppCPG2-VNP20009 can act as a vector for CPG2 and induce prodrug sensitivity in infected mammalian cell cultures.

Production of active drug and effect on bacterial viability. ppCPG2-VNP20009 requires ampicillin selection to maintain CPG2 expression. This would be a drawback in vivo where selection would be difficult. We therefore moved the ompACPG2 gene construct to the plasmid pYA3332 and used the resulting plasmid pYA3332 ompA-CPG2 to transform the bacterium VNP20009asd-, generating the effectively stable vector ppCPG2-VNP20009asd-, which produced similar levels of ppCPG2 activity to ppCPG2-VNP200009. To investigate the in vivo therapeutic efficacy of the suicide gene therapy, we selected the benzoyl glutamate prodrugs 2 to 4 that have previously been proven effective in transfected xenograft models (23) and investigated prodrug activation with ppCPG2-VNP200009asd-. Prodrug 2 has the benefits for chromatographic studies of being more stable than prodrug 1, easier to separate, and detect by UV absorbance and yielding a readily detectable drug (2a). Both prodrug 2 and drug 2a are unstable in aqueous medium, but their respective degradation product peaks could be identified by incubation of authentic compounds in plain buffer. When we incubated prodrug 2 with cultures of VNP20009 or ppCPG2-VNP20009asd-, a time-dependent loss was observed (Fig. 3A ), which is faster for cultures of ppCPG2-VNP20009asd-. However, whereas prodrug 2 and its degradation products are seen in supernatant fractions from both cultures of VNP20009 and ppCPG2-VNP20009asd-, drug 2a and its degradation products are seen only in the supernatant from ppCPG2-VNP20009asd-. All retention times (2, 3.8 min, products 2.5 and 7.0 min; 2a, 5.5 min, products 4.1 and 9.2 min) and spectra matched those from authentic compound. We conclude that prodrug 2 accesses the periplasm and is converted to the cognate drug 2a, which is then released into the surrounding medium. The rates of loss of the prodrug are plotted against time and fitted to an exponential decay (Fig. 3A). The rate constants differ significantly from each other [F(DFn,DFd): 5.9 (1,8); P = 0.041]. From the difference, we then calculated the extent of activation-specific prodrug 2 loss (Fig. 3A).


Figure 3
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 3. A, rates of loss of prodrug 2 in the supernatant medium from cultures of ppCPG2-VNP20009asd- (filled circles) or VNP20009 (filled squares) and the CPG2-specific difference between them (stars), determined by high-performance liquid chromatography. B, recovery of viable bacterial cfu from cultures exposed to prodrug 1 under conditions as for Fig. 2. C, survival of VNP20009 following exposure to a range of doses of drug 4a for 2 h. Broken lines, 95% confidence interval.

 
To establish whether prodrug activation would adversely affect the bacteria, two assays were performed. The CPG2-expressing bacteria ppCPG2-VNP20009 were incubated with the dose of prodrug 1 used to induce cytotoxicity in WiDr cells (40 µmol/L). No significant difference in bacterial viability between treated cells and those untreated or treated with prodrug vehicle alone is found (Fig. 3B). In a separate experiment, VNP20009 bacteria were exposed to various concentrations of the active drug 4a arising from one of the prodrugs (4) used in vivo (Fig. 3C). There is very little reduction in the viability of the bacteria, although fitting a linear regression forced through 0,100 indicates a significant cytotoxic response (t,DF: 2.8,5; P = 0.038). The therapeutic in vivo dose of prodrug 4 (200 mg/kg) even if immediately converted to drug 4a would only correspond to a concentration of 300 µmol/L. By interpolation from Fig. 3C, this would result in negligible toxicity to the bacteria. We conclude therefore that doses of prodrug giving rise to cytotoxicity in tumor cells in vivo would not affect the bacterial vectors.

In vivo tumor colonization by ppCPG2-VNP20009asd- and CPG2 expression. We determined the numbers of live ppCPG2-VNP20009asd- bacteria recovered from the MDA-MB-361, WiDr, and B16-F10 tumors and plotted them against the time from the original inoculation. Bacterial colonization of the tumors proceeds similarly in all three tumor types, initially increasing rapidly then tending toward a maximum (Fig. 4 ). The best fit to the data was a hyperbola, enabling a prediction of the maximum bacterial load (nonoverlapping 95% confidence intervals in brackets), which was approximately 7 times higher in B16-F10 (1.7 x 108 cfu/g; 8.3 x 107-3.5 x 108) than in WiDr (2.5 x 107; 1.0 x 107-6.3 x 107) but with the corresponding value for MDA-MB-361 (2.1 x 108), having a very broad 95% confidence interval. In WiDr, these values for bacterial load per gram are significantly lower than those per milliliter, producing the greatest cell kill in six-well plates (Fig. 2C). However, in vitro, the bacteria are distributed throughout the medium with only a small proportion being in close contact with the cells. The time to half-maximal colonization is longer in MDA-MB-361 [6.2 days (3.3-9.2)] than in WiDr [1.8 days (1.2-2.5)] or B16-F10 [1.2 days (0.8-1.6)]. During the same period, the amount of CPG2 enzyme activity in the tumors was also determined (Fig. 5 ). After an average initial lag (expressed as the X intercept when Y = 0) of 4.2 days that did not differ significantly among the three cell lines, the CPG2 activity increased linearly. The greatest rate of accumulation of CPG2, calculated as the slope, was seen in B16-F10 [Fig. 5C; 0.29 units/g/d (0.22-0.35)] followed by WiDr [Fig. 5B; 0.11 units/g/d (0.069-0.153)] and MDA-MB-361 [Fig. 5A; 0.037 units/g/d (0.015-0.058)]. The lack of a plateau in CPG2 activity to mirror that in bacterial load suggests that either CPG2 expression continues to build in the existing bacteria after they have become stationary or there is death and release of CPG2 from some bacteria but sufficient division to replace their number. The levels of CPG2 in the tumors after 20 days derived from these regressions approximated those shown previously to be sufficient to eradicate tumors in a transfected model [MDA-MB-361 0.59 units/g measured, 0.6 units/g historical (30); 2.0 units/g; WiDr 1.57 units/g measured, 2.0 units/g historical (23)] and are higher in B16-F10 tumors (5.0 units/g).


Figure 4
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 4. Recovery of viable cfu of ppCPG2-VNP20009asd- from human tumor xenografts (A, breast carcinoma MDA-MB-361; B, colon carcinoma WiDr) and a mouse allograft (C, mouse melanoma B16-F10) following a single i.v. dose of 2 x 106 cfu per mouse. Broken lines, 95% confidence interval.

 

Figure 5
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 5. CPG2 enzyme activity in human tumor xenografts (A, breast carcinoma MDA-MB-361; B, colon carcinoma WiDr) and a mouse allograft (C, mouse melanoma B16-F10) following a single i.v. dose of ppCPG2-VNP20009asd-. Broken lines, 95% confidence interval.

 
In vivo therapy of human breast and colon carcinoma and mouse melanoma tumors. Tumor doubling times were calculated from fits to rising exponentials (nonoverlapping 95% confidence intervals in brackets). As expected for an oncolytic vector, ppCPG2-VNP20009asd- alone inhibits the growth of all three tumor types. However, in MDA-MB-361 (Fig. 6A ), it takes until day 14 for there to be any apparent difference between the tumor growth in the controls and any of the treated groups, which probably reflects the slower rate of bacterial colonization and CPG2 accumulation seen in this tumor type (Fig. 4A) compared with WiDr (Fig. 4B) and B16-F10 (Fig. 4C). Subsequent to this initial nonresponsive phase, doubling times from day 14 were calculated by fitting the data to a rising exponential, constraining the start value to the group average for day 0. In this later phase, the control tumors grow with a doubling time of 20.8 days (24.3-18.1) compared with 31.0 days (37.9-26.2) in the vector alone treated and a much slower average of 83 days (64.5-120.0) when tumors additionally received any of the three prodrugs 2, 3, or 4, there being no significant difference between the effects of these prodrugs. Significant difference exists between the curves where indicated [F(DFn,DFd): 51.4 (2,356); P < 0.0001]. In the WiDr (Fig. 6B) and B16-F10 (Fig. 6C) tumors, there was, by comparison, no initial unresponsive phase (Fig. 4B and C); hence, the rising exponentials used to generate doubling times were applied from the first time point, constraining the start value to the group average for day 0. This more immediate response is commensurate with the more rapid onset of bacterial colonization and CPG2 accumulation in WiDr and B16-F10 (Figs. 3B and C and 4B and C). In WiDr xenografts, the control tumor or those treated with ppCPG2-VNP20009asd- alone or with ppCPG2-VNP20009asd- plus prodrug 4 do not give significantly different doubling times, having a joint average of 9.9 days (8.2-12.2) and are shown as a single group. However, those treated with ppCPG2-VNP20009asd- plus prodrug 3 have a longer doubling time of 15.6 days (14.6-16.9). Significant difference exists between the curves where indicated [F(DFn,DFd): 50.1 (3,323); P < 0.0001]. These three prodrugs alone have no effect on the growth of either MDA-MB-361 or WiDr xenografts (23). B16-F10 control tumors grew more rapidly than human tumors, the doubling time being 2.3 days (2.5-2.1), with a more marked reduction in growth rate following treatment with ppCPG2-VNP20009asd- alone of 5.3 days (5.5-5.0). Treatment with ppCPG2-VNP20009asd- plus prodrugs produced further significant extensions of doubling times of 6.2 days (6.6-5.8) for prodrug 4 and to 9.4 days (10.4-8.5) for prodrug 3. Significant difference exists between the curves [F(DFn,DFd): 90.6(3,307); P < 0.0001]. Toxicity defined by >10% body weight reduction was not encountered. Evidence of therapeutic tumor necrosis was observed as frequent scab formation; however, according to UK Home Office license requirements on humane regulations (38), this necessitated premature culling before the scientifically relevant endpoint of tumor size had been reached; hence, meaningful survival data could not be acquired.


Figure 6
View larger version (15K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 6. Growth of human tumor xenografts (A, breast carcinoma MDA-MB-361; B, colon carcinoma WiDr) and a mouse allograft [C, mouse melanoma B16-F10; controls (open circles)] after i.v. injection of ppCPG2-VNP20009asd- (open squares) followed by five weekly i.p. doses of prodrugs 2 (filled inverted triangle), 3 (filled diamond), or 4 (filled circle). Broken lines, 95% confidence interval.

 
In WiDr and B16-F10 but not in MDA-MB-361, in combination with ppCPG2-VNP20009asd-, prodrug 3 is more effective than prodrug 4. In a separate experiment (data not shown), prodrugs 3 and 4 alone were shown to have no effect on B16-F10 tumor growth, with all six of the treated and control tumors reaching the humane limit (1,000 mm3) by 15 to 18 days, having grown slightly slower than in the present experiment. We have shown previously in transfected xenograft models that the choice of prodrug to produce maximum effect varies between tumor types and degree of CPG2 expression within that tumor (23). We observed the same pattern here in a targeted system, B16-F10 especially, which emphasizes the versatility of CPG2 as a prodrug-activating enzyme. In each case, a comparison of the tumor volume to cfu per gram enables an approximation of the total bacterial load, indicative of at least 100-fold more bacteria in the tumor than that administered to the whole animal, showing that the observed increase is a result of bacterial growth in the tumors.

In summary, we have shown that attenuated oncolytic Salmonella can act as a vector for the prodrug-activating enzyme CPG2. We have shown that the enzyme, located in the bacterial periplasm, induces cytotoxicity in mammalian tumor cells by converting prodrugs to their active form, but the bacteria themselves are not killed by the activated drugs. The bacteria colonize tumors and multiply within them, resulting in a selective accumulation of CPG2. As is the case with the nonexpressing bacterium from which they were derived, the CPG2-expressing bacteria alone have a significant effect on the growth of tumors. However, in the presence of prodrugs for CPG2, this oncolytic effect is considerably enhanced. We conclude that bacterial oncolytic therapy, combined with CPG2-mediated prodrug activation, has great potential in the treatment of a range of cancers.


   Disclosure of Potential Conflicts of Interest
Top
 Abstract
 Materials and Methods
 Results and Discussion
 Disclosure of Potential...
References
 
F. Friedlos, P. Lehouritis, L. Ogilvie, D. Hedley, L. Davies, J. Martin, R. Marais, and C. Springer are employed by The Institute of Cancer Research. These authors are party to the rewards to investors scheme that reward contributions to licensed programmers. I. King and D. Bermudes are employed by Vion Pharmaceuticals.


   Footnotes
 
Grant support: Cancer Research UK grants C309/A2187 and C309/A8274 (F. Friedlos, P. Lehouritis, L. Ogilvie, D. Hedley, L. Davies, J. Martin, R. Marais, and C.J. Springer), and Vion Pharmaceuticals, Inc. (D. Bermudes and I. King).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: F. Friedlos and P. Lehouritis contributed equally to this work.

Current address for D. Bermudes: Celator Pharmaceuticals Corp., Vancouver, British Columbia, Canada V6P6P2.

Received 11/ 5/07; revised 1/18/08; accepted 2/12/08.


   References
Top
 Abstract
 Materials and Methods
 Results and Discussion
 Disclosure of Potential...
 References
 

  1. Springer CJ. Suicide gene therapy methods and reviews. Totowa (NJ): Humana Press; 2004.
  2. Hedley D, Ogilvie L, Springer C. Carboxypeptidase G2-based gene-directed enzyme-prodrug therapy: a new weapon in the GDEPT armoury. Nat Rev Cancer 2007;7:870–9.[CrossRef][Medline]
  3. Theys J, Pennington O, Dubois L, et al. Repeated cycles of Clostridium-directed enzyme prodrug therapy result in sustained antitumour effects in vivo. Br J Cancer 2006;95:1212–9.[CrossRef][Medline]
  4. Barbe S, Van Mellaert L, Anne J. The use of clostridial spores for cancer treatment. J Appl Microbiol 2006;101:571–8.[CrossRef][Medline]
  5. Dang LH, Bettegowda C, Huso DL, Kinzler KW, Vogelstein B. Combination bacteriolytic therapy for the treatment of experimental tumors. Proc Natl Acad Sci U S A 2001;98:15155–60.[Abstract/Free Full Text]
  6. Bhatnagar PK, Awasthi A, Nomellini JF, Smit J, Suresh MR. Anti-tumor effects of the bacterium Caulobacter crescentus in murine tumor models. Cancer Biol Ther 2006;5:485–91.[Medline]
  7. Clairmont C, Bermudes D, Low KB, et al. VNP20009, a genetically modified Salmonella typhimurium: anti-tumor efficacy, toxicology, and biodistribution in preclinical models. Clin Cancer Res 1999;5:3830.
  8. Low KB, Ittensohn M, Luo X, et al. Construction of VNP20009, a novel, genetically stable antibiotic-sensitive strain of tumor-targeting Salmonella for parenteral administration in humans. Methods Mol Med 2004;90:47–60.[Medline]
  9. Zhang L, Gao LF, Zhao LJ, et al. Intratumoral delivery and suppression of prostate tumor growth by attenuated Salmonella enterica serovar typhimurium carrying plasmid-based small interfering RNAs. Cancer Res 2007;67:5859–64.[Abstract/Free Full Text]
  10. Low KB, Ittensohn M, Le T, et al. Lipid A mutant Salmonella with suppressed virulence and TNF {alpha} induction retain tumor-targeting in vivo. Nat Biotechnol 1999;17:37–41.[Medline]
  11. Pawelek JM, Low KB, Bermudes D. Tumor-targeted Salmonella as a novel anticancer vector. Cancer Res 1997;57:4537–44.[Abstract/Free Full Text]
  12. Lee KC, Zheng LM, Margitich D, Almassian B, King I. Evaluation of the acute and subchronic toxic effects in mice, rats, and monkeys of the genetically engineered and Escherichia coli cytosine deaminase gene-incorporated Salmonella strain, TAPET-CD, being developed as an antitumor agent. Int J Toxicol 2001;20:207–17.[CrossRef][Medline]
  13. Toso JF, Gill VJ, Hwu P, et al. Phase I study of the intravenous administration of attenuated Salmonella typhimurium to patients with metastatic melanoma. J Clin Oncol 2002;20:142–52.[Abstract/Free Full Text]
  14. Avogadri F, Martinoli C, Petrovska L, et al. Cancer immunotherapy based on killing of Salmonella-infected tumor cells. Cancer Res 2005;65:3920–7.[Abstract/Free Full Text]
  15. Thamm DH, Kurzman ID, King I, et al. Systemic administration of an attenuated, tumor-targeting Salmonella typhimurium to dogs with spontaneous neoplasia: phase I evaluation. Clin Cancer Res 2005;11:4827–34.[Abstract/Free Full Text]
  16. Minton NP, Atkinson T, Sherwood RF. Molecular-cloning of the Pseudomonas carboxypeptidase-G2 gene and its expression in Escherichia coli and Pseudomonas putida. J Bacteriol 1983;156:1222–7.[Abstract/Free Full Text]
  17. Niculescu-Duvaz D, Niculescu-Duvaz I, Friedlos F, et al. Self-immolative nitrogen mustards prodrugs cleavable by carboxypeptidase G2 (CPG2) showing large cytotoxicity differentials in GDEPT. J Med Chem 2003;46:1690–705.[CrossRef][Medline]
  18. Niculescu-Duvaz D, Niculescu-Duvaz I, Friedlos F, et al. Self-immolative nitrogen mustard prodrugs for suicide gene therapy. J Med Chem 1998;41:5297–309.[CrossRef][Medline]
  19. Niculescu-Duvaz I, Friedlos F, Niculescu-Duvaz D, Davies L, Springer CJ. Prodrugs for antibody- and gene-directed enzyme prodrug therapies (ADEPT and GDEPT). Anticancer Drug Des 1999;14:517–38.[Medline]
  20. Niculescu-Duvaz I, Niculescu-Duvaz D, Friedlos F, et al. Self-immolative anthracycline prodrugs for suicide gene therapy. J Med Chem 1999;42:2485–9.[CrossRef][Medline]
  21. Niculescu-Duvaz I, Scanlon I, Niculescu-Duvaz D, et al. Significant differences in biological parameters between prodrugs cleavable by carboxypeptidase G2 that generate 3,5-difluoro-phenol and -aniline nitrogen mustards in gene-directed enzyme prodrug therapy systems. J Med Chem 2004;47:2651–8.[CrossRef][Medline]
  22. Davies LC, Friedlos F, Hedley D, et al. Novel fluorinated prodrugs for activation by carboxypeptidase G2 showing good in vivo anti-tumor activity in gene directed enzyme prodrug therapy (GDEPT). J Med Chem 2005;48:5321–8.[CrossRef][Medline]
  23. Friedlos F, Davies L, Scanlon I, et al. Three new prodrugs for suicide gene therapy using carboxypeptidase G2 elicit bystander efficacy in two xenograft models. Cancer Res 2002;62:1724–9.[Abstract/Free Full Text]
  24. Francis RJ, Sharma SK, Springer CJ, et al. A phase I trial of antibody directed enzyme prodrug therapy (ADEPT) in patients with advanced colorectal cancer or other CEA producing tumours. Br J Cancer 2002;87:600–7.[CrossRef][Medline]
  25. Mayer A, Francis RJ, Sharma SK, et al. A phase I study of single administration of antibody-directed enzyme prodrug therapy with the recombinant anti-carcinoembryonic antigen antibody-enzyme fusion protein MFECP1 and a bis-iodo phenol mustard prodrug. Clin Cancer Res 2006;12:6509–16.[Abstract/Free Full Text]
  26. Schepelmann S, Hallenbeck P, Ogilvie LM, et al. Systemic gene-directed enzyme prodrug therapy of hepatocellular carcinoma using a targeted adenovirus armed with carboxypeptidase G2. Cancer Res 2005;65:5003–8.[Abstract/Free Full Text]
  27. Schepelmann S, Ogilvie LM, Hedley D, et al. Suicide gene therapy of human colon carcinoma xenografts using an armed oncolytic adenovirus expressing carboxypeptidase G2. Cancer Res 2007;67:4949–55.[Abstract/Free Full Text]
  28. Marais R, Spooner R, Light Y, Martin J, Springer CJ. Gene-directed enzyme prodrug therapy with a mustard prodrug/carboxypeptidase G2 combination. Cancer Res 1996;56:4735–42.[Abstract/Free Full Text]
  29. Marais R, Spooner R, Stribbling SM, Light Y, Martin J, Springer CJ. A cell surface tethered enzyme improves efficiency in gene-directed enzyme prodrug therapy. Nat Biotechnol 1997;15:1373–7.[CrossRef][Medline]
  30. Stribbling SM, Friedlos F, Martin J, et al. Regressions of established breast carcinoma xenografts by carboxypeptidase G2 suicide gene therapy and the prodrug CMDA are due to a bystander effect. Hum Gene Ther 2000;11:285–92.[CrossRef][Medline]
  31. Springer CJ. CMDA. Drugs Future 1993;18:212–5.
  32. Springer CJ, Dowell R, Burke PJ, et al. Optimization of alkyating agent prodrugs derived from phenol and aniline mustards: a new clinical candidate prodrug (ZD2767P) for antibody-directed enzyme prodrug therapy (ADEPT). J Med Chem 1995;38:5051–65.[Medline]
  33. Springer CJ, Niculescu-Duvaz I, Pedley RB. Novel prodrugs of alkylating agents derived from 2-fluoro- and 3-fluorobenzoic acids for antibody-directed enzyme prodrug therapy. J Med Chem 1994;37:2361–70.[CrossRef][Medline]
  34. Amann E, Ochs B, Abel KJ. Tightly regulated tac promoter vectors useful for the expression of unfused and fused proteins in Escherichia coli. Gene 1988;69:301–15.[CrossRef][Medline]
  35. Minton NP, Atkinson T, Bruton CJ, Sherwood RF. The complete nucleotide sequence of the Pseudomonas gene coding for carboxypeptidase G2. Gene 1984;31:31–8.[CrossRef][Medline]
  36. Bolla JM, Lazdunski C, Inouye M, Pages JM. Export and secretion of overproduced Ompa-β-lactamase in Escherichia coli. FEBS Lett 1987;224:213–8.[CrossRef][Medline]
  37. Galan JE, Nakayama K, Curtiss R. Cloning and characterization of the Asd gene of Salmonella typhimurium—use in stable maintenance of recombinant plasmids in Salmonella vaccine strains. Gene 1990;94:29–35.[CrossRef][Medline]
  38. Workman P, Twentyman P, Balkwill F, et al. United Kingdom Co-ordinating Committee on Cancer Research (UKCCCR) guidelines for the welfare of animals in experimental neoplasia (second edition). Br J Cancer 1998;77:1–10.[Medline]
  39. Sherwood RF, Melton RG, Alwan SM, Hughes P. Purification and properties of carboxypeptidase G2 from Pseudomonas sp strain Rs-16—use of a novel triazine dye affinity method. Eur J Biochem 1985;148:447–53.[Medline]
  40. Roberts JJ, Friedlos F, Scott D, Ormerod MG, Rawlings CJ. The unique sensitivity of Walker rat tumor cells to difunctional agents is associated with a failure to recover from inhibition of DNA synthesis and increased chromosome damage. Mutat Res 1986;166:169–81.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Friedlos, F.
Right arrow Articles by Springer, C. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Friedlos, F.
Right arrow Articles by Springer, C. J.

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