Antitumour activity of Salmonella typhimurium VNP 20047 in B 16 ( F 10 ) murine melanoma model

A tumour therapy is proposed based on attenuated Salmonella typhimurium VNP20047 expressing the Escherichia coli cytosine deaminase gene. VNP20047 was administered intravenously to B16(F10) melanoma-bearing C57BL/6 mice. VNP20047 proliferated within tumours and livers regardless of the initial inoculum dose. After 10 days the number of bacteria increased in livers up to 4.2 10 cfu/g and decreased in tumours down to 5.9 10 cfu/g. VNP20047 at 1 10 cfu/mouse, when combined with 5-fluorocytosine, inhibited tumour growth by 85% without prolonging animal survival. Histology studies revealed severe lesions in tumours and livers. These data suggest that S. typhimurium VNP20047 induced inflammatory responses, even though the strain was attenuated.

rial vectors including Clostridium, Bifidobacterium and Salmonella have been known to specifically accumulate and replicate within tumours (Yazawa et al., 2000;Theys et al., 2001;Bermudes et al., 2001).They can cause tumour cell death as a result of competing for nutrients with cancer cells and/or secreting therapeutic proteins (King et al., 2002, Theys et al., 2003).The problem with bacteriolytic therapy is its toxicity.Even after deleting the genes that induce immune response, bacteriolytic therapy led to marked mortality in mice (Dang et al., 2001).
Salmonella typhimurium strain VNP20047 carries deletions in the msbB and purI loci, which results in diminished virulence.The purI deletion creates a requirement for an external purine source for survival (Pawelek et al., 1997).The deletion of the msbB gene reduces the toxicity associated with lipopolysaccharide through preventing the addition of the terminal myristyl group to the lipid A domain (Low et al., 1999).Additionally, cytosine deaminase gene from Eschericha coli was incorporated into S. typhimurium VNP20047 to enhance its tumouricidal effect upon subsequent i.p. administration of 5-fluorocytosine (King et al., 2002).The inserted suicide gene converts the nontoxic 5-fluorocytosine (5-FC) to a highly toxic drug, 5-fluorouracil (5-FU).Its metabolites ultimately interfere with DNA and RNA synthesis (Bermudes et al., 2001).
The aim of our study was to investigate if attenuated S. typhimurium VNP20047 strain carrying E. coli cytosine deaminase gene localised specifically in B16(F10) murine melanoma tumours causing inhibition of its growth and improved animals survival.
Animals.Six-to eight-week-old C57BL/6 mice were bred and maintained at the on-site animal facility.Consent for experiments involving animals was obtained from the appropriate local Ethics Committee (Medical Academy, Warszawa, Poland).
Preparation of S. typhimurium VNP-20047.VNP20047 bacteria were obtained from Vion Pharmaceuticals, Inc. Cultures were grown overnight at 37°C in LB broth, then centrifuged for 5 min at 4000 rpm, and the pellet washed twice in 2 ml PBS -.Bacterial suspension was appropriately diluted in PBS -before making the dilution for injection (1 A 600 = 1 ´10 9 cfu/ml) (Zheng et al., 2000).Actual dose was determined by plating appropriate solutions on LB agar plates and counting colonies after overnight incubation.

Quantitation of S. typhimurium VNP-20047 accumulation.
Ten days after melanoma cell injection the mice were injected with three different doses of VNP20047 (1 10 4 , 1 ´10 5 and 1 ´10 6 cfu) in 0.2 ml PBS - were tail-injected into tumour-bearing mice.The mice were sacrificed at different times after the injection.Tumours and livers were aseptically weighed and homogenised in PBS -.Serial dilutions were performed to count the number of bacteria per gram of tissue.
Antitumour activity of S. typhimurium VNP20047.When B16(F10) tumours reached approx.150 mm 3 , seven days after inoculation with the tumour cells, the mice were divided into four groups of five animals each.Group 1 consisted of mice injected with B16(F10) melanoma cells only.Group 2 received PBS -(0.2 ml).Groups 3 and 4 received VNP20047 at 1 ´10 5 cfu per mouse.Groups 2 and 3 received two intraperitoneal injections of 5-FC daily at a dose of 300 mg/kg on days 10 through 19.Tumour dimensions were measured three times a week and their vol-ume calculated from the following formula: width 2 ´length ´0.52.Tumour volume is reported as mean ± standard deviation.The results were analysed with Student's t-test.Histological analysis.To observe potential side-effects of the therapy in the treated mice, 10 days after injection of bacteria, livers and tumours were fixed in 10% neutral buffered formalin solution and embedded in paraffin.Sections, 8 mm-thick, were stained with hematoxylin and eosin (H&E).

Localization of VNP20047 bacteria in tumour and liver
C57Bl/6 mice that had been inoculated with B16(F10) melanoma cells were intravenously administered different doses of VNP20047.Tumour and liver samples, obtained four days after bacteria injection, were evaluated for the presence of S. typhimurium.Figure 1 shows that similar levels of accumulation were achieved regardless of the initial inoculum size.The dose of 10 5 cfu/mice was chosen for subsequent experiments.
The accumulation of VNP20047 was evaluated in tissues at specified times after injec-tion of bacteria.Figure 2 shows that at 2 h after injection VNP20047 accumulated in livers at nearly the same levels as in B16(F10) melanoma tumours.The number of VNP20047 bacteria increased in tumours and decreased in livers within 24 h after administration.By day 4 the number of VNP20047 in tumours increased significantly and the tumour-to-liver ratio amounted to 1000 : 1 on average.However, the levels of VNP20047 did not remain stable.Ten days after injection the number of bacteria increased in livers and decreased in tumours, reaching in both tissues the same level approx.5 ´10 6 cfu/g tissue.

Antitumour efficacy of VNP20047
Subcutaneous B16(F10) tumours were induced in C57Bl/6 mice by injecting 2 ´10 5 viable tumour cells.Seven days later, when the tumours were about 4-5 mm in diameter, 10 5 cfu of S. typhimurium VNP20047, or PBS - (control), was injected intravenously.On the tenth day, three days after bacteria injection, the animals received intraperitoneally 300 mg/kg of the 5-FC prodrug per dose, twice a day for ten consecutive days.Tumour growth and survival were evaluated (Fig. 3 A and B).In the group treated with VNP20047/5-FC, all tumours showed 85% inhibition of growth Vol. 51 Salmonella typhimurium VNP20047 in B16(F10) murine melanoma model 853  during the second week of treatment.The tumours in the VNP20047 group (without 5-FC) were also significantly smaller than those in the control groups (70% inhibition).
Nevertheless, despite the significant tumour growth inhibition, the treatment with bacteria and 5-FC did not prolong mice survival, as compared to control groups (Fig. 3B).Five independent experiments were carried out with similar results.

Histology analysis of tumours and livers in mice receiving VNP20047
Livers and spleens of mice treated with VNP20047 and examined after ten days were grossly enlarged and swollen (hepatosplenomegaly). Numerous grey-white lesions were visible in the liver.Histology sections of tu-mours and livers showed significant differences between control animals and mice treated with the attenuated Salmonella VNP20047.Ten days following the injection of bacteria several necrotic regions with surrounding inflammation were visible in the tumours (Fig. 4A).Tumours from control animals showed less necrotic zones that also had no significant inflammatory infiltrate.Sections of livers from treated animals showed numerous inflammatory foci localised mainly centrolobularly as well as several necrotic regions (Fig. 4B).

DISCUSSION
Salmonella are Gram-negative bacteria from the Enterobacteriacea family.They are faculta-  tive anaerobes and grow readily in simple culture media, under both aerobic and anaerobic conditions.All the bacteria from the Salmonella genus are considered potential pathogens since they cause gastroenteritis in most animal species.S. typhimurium has several advantages that enable it to exert an antitumour effect (King et al., 2000).It is able to grow in hypoxic or necrotic areas that occur in many kinds of tumours.It is capable of invading and killing host cells and is highly sensitive to many antibiotics (Luo et al., 2001).Salmonella can be easily engineered in the laboratory to express foreign genes as well as to attenuate its pathogenicity (Sznol et al., 2000).Pawelek et al. (1997) reported that attenuated Salmonella preferentially accumulated in tumour tissue of B16(F10) melanoma and M019 lung carcinoma-bearing mice with the tumour-to-liver ratio amounting to 1000 : 1 on average.The results obtained in those studies provided the basis for Phase I clinical trial (Toso et al., 2002).The attenuated S. typhimurium VNP20047, carrying deletions of the msbB and purI genes, was developed as a vector suitable for tumour-selective delivery of therapeutic proteins (Pawelek et al., 1997).Available data suggested that tumour-targeting Salmonella could be used selectively to deliver a prodrug-converting enzyme to tumours, producing therapeutic effects (King et al., 2002).Our studies demonstrate that S. typhimurium VNP20047 localises not only to melanoma tu-mours but also to livers.Quantitative analysis demonstrated rapid growth of the bacteria in tumours within a few days after injection of VNP20047, but ultimately their number did not remain stable.After 10 days the number of bacteria decreased in the tumours and increased in livers, reaching approx.5 ´10 6 cfu/g of either tissue.A single dose (1 ´10 5 cfu/ mouse) of VNP20047, together with a systemically administered prodrug (5-FC), inhibited tumour growth but did not prolong mice survival as compared to control groups.In most cases mice treated with VNP20047 died before the onset of progressive tumour growth.
The premature death of treated mice could be caused by an excessive accumulation of the bacteria in normal tissues, resulting in inflammation foci and greatly expanded areas of necrosis.Our data suggest that S. typhimurium VNP20047 induced inflammatory responses, even though the strain was attenuated.These results indicate that cancer therapy using VNP20047 as a tumour selective vector is not very effective.Although the definite causes of for the death of the treated animals are not clear, additional modifications of the bacteria seem necessary to develop a still safer strain.Authors thank Dr. A. Sochanik for helpful review of the manuscript; Dr. S. Krzyzowska-Gruca and Dr. E. Malusecka for their help with microscopic analysis.

Figure
Figure 1.Accumulation of bacteria in tissues of C57BL/6 mice bearing B16(F10) melanoma injected i.v.various dose of VNP20047.On the tenth day mice were administered different doses of VNP20047 (1 ´10 4 , 1 ´105 and 1 ´10 6 cfu).After subsequent four days the tumours and livers were collected, weighed and homogenised in PBS -.Bacteria were counted by plating serial dilutions of the homogenates.

Figure 2 .
Figure 2. Colonisation of B16(F10) tumours and livers by VNP20047 following a single i.v.injection into C57BL/6 tumour-bearing mice.On the tenth day mice were administered a single dose of VNP20047 (1 ´10 5 cfu/mouse).The tumours and livers were collected at various time points, weighed and homogenised in PBS -.Bacteria were counted by plating serial dilutions of the homogenates.
Figure 3. Antitumour effect of VNP20047 injection in B16(F10) melanoma tumour-bearing C57BL/6 mice.(A) Average tumour growth rate and (B) mice survival.On the seventh day mice were administered a single dose of VNP20047 (1 ´10 5 cfu/mouse) expressing cytosine deaminase gene.On days 10 through 19 mice received i.p. twice a day 300 mg/kg of 5-FC.