Tumour-targeted delivery of TRAIL using Salmonella typhimurium enhances breast cancer survival in mice

Abstract

Background: An effective cancer therapeutic must selectively target tumours with minimal systemic toxicity. Expression of a cytotoxic protein using Salmonella typhimurium would enable spatial and temporal control of delivery because these bacteria preferentially target tumours over normal tissue.

Methods: We engineered non-pathogenic S. typhimurium to secrete murine TNF-related apoptosis-inducing ligand (TRAIL) under the control of the prokaryotic radiation-inducible RecA promoter. The response of the RecA promoter to radiation was measured using fluorometry and immunoblotting. TRAIL toxicity was determined using flow cytometry and by measuring caspase-3 activation. A syngeneic murine tumour model was used to determine bacterial accumulation and the response to expressed TRAIL.

Results: After irradiation, engineered S. typhimurium secreted TRAIL, which caused caspase-3-mediated apoptosis and death in 4T1 mammary carcinoma cells in culture. Systemic injection of Salmonella and induction of TRAIL expression using 2 Gy gamma-irradiation caused a significant delay in mammary tumour growth and reduced the risk of death by 76% when compared with irradiated controls. Repeated dosing with TRAIL-bearing Salmonella in conjunction with radiation improved the 30-day survival from 0 to 100%.

Conclusion: These results show the pre-clinical utility of S. typhimurium as a TRAIL expression vector that effectively reduces tumour growth and extends host survival.

Figure 1

Radiation-inducible prokaryotic expression constructs for ZsGreen and TRAIL. (A) The plasmid construct, pRA-ZsG, has the S. typhimurium promoter for RecA upstream of the green fluorescent protein, ZsGreen. The plasmid construct, pRA-TR, substitutes ZsGreen with the apoptosis-inducing peptide, mTRAIL. (B) Fluorescence imaging shows green fluorescence from bacteria electroporated with pZsGreen and pRA-ZsG plasmid constructs. (C) Fluorometry showing relative induction by the RecA promoter. S. typhimurium VNP20009 electroporated with pZsGreen and pRA-ZsG were induced with 2 Gy γ-irradiation or 5 J m⁻² UV irradiation, and then grown at 37 °C, 250 r.p.m. for 4 h. Data were normalised to the bacterial absorbance at 600 nm, and are reported as relative fluorescence of VNP pRA-ZsG compared with VNP pZsGreen. Gene expression is significantly increased with genotoxic damage in comparison with the un-induced control (^*P<0.05). (D) Immunoblot showing TRAIL secretion. 15% SDS–PAGE was performed on 20 μg protein obtained from cytosolic (c) and supernatant (s) fractions of overnight bacterial culture transformed with plasmid constructs. After transfer to PVDF, an immunoblot was performed using 1 : 200 Rabbit anti-TRAIL polyclonal antibody.

Figure 2

In vitro effects of secreted proteins from transformed S. typhimurium on 4T1 mammary carcinoma cells. Filtered supernatants were acquired from VNP pRA-ZsG (ZsGreen sup.) and VNP pRA-TR (TRAIL sup.) after growth for 4 h in minimal medium after induction with 5 J m⁻² UV irradiation. Control supernatants were acquired from sterile medium. (A) Caspase activity assays were conducted on 4T1 cells after application of experimental supernatants or recombinant mouse TNF-α at 50 ng ml⁻¹ for 24 h. Significant increases in caspase-3 and caspase-8 activities were observed after treatments with TNF-α and VNP pRA-TR supernatants when compared with controls (^*P<0.05). Addition of the caspase-3 inhibitor, DEVD-fmk, significantly reduced activity. (B, C) Flow cytometry for annexin-V-FITC and propidium iodide was conducted on 10 000 4T1 cells per treatment in triplicate after application of experimental supernatants or TNF-α at 50 ng ml⁻¹ for 48 h. (B) Flow cytometry dot plots after treatment with control (left) and VNP pRA-TR supernatants (right) indicate cell death (annexin-V and propidium iodide positive) proportions of 5.7 and 12.6%, respectively. (C) Results of flow cytometry indicate cell fractions undergoing cell death (annexin-V positive, propidium iodide positive) and early apoptosis (annexin-V positive, propidium iodide negative). Significant increases in cell death and early apoptosis were observed in 4T1 cells after treatment with the VNP pRA-TR supernatant (^*P<0.05).

Figure 3

Spatial distribution of S. typhimurium after systemic administration. (A) Bacterial concentration (cfu g⁻¹) in 4T1 mammary tumours and livers of BALB/c mice at 48 h after systemic injection of 100 000 cfu g⁻¹ VNP20009. (B) Band of bacteria in a 4T1 tumour (stained red with white arrows) identified by anti-Salmonella immunohistochemistry. Scale bar is 100 μm. (C) Composite image of 4T1 tumour stained using Salmonella immunohistochemistry to visualise bacteria (red with white arrows). Scale bar is 5 mm.

Figure 4

Reduced tumour growth and enhanced survival from VNP pRA-TR with induction by γ-irradiation. Balb/c mice received intravenous injection of 100 000 cfu g⁻¹ VNP pRA-ZsG, VNP pRA-TR, or PBS at 21 days after establishing 4T1/red tumours. (A) Tumour volumes (mm³) after experimental treatments in mice receiving no irradiation, and in mice receiving 2 Gy irradiation on day 2 (arrow). No significant differences were observed in initial tumour volume between treatment groups. Mean tumour volume did not exceed 1000 mm³ in mice treated with VNP pRA-TR and 2 Gy at 1 month. (B) Tumour doubling time and (C) growth delay, as determined from regression analysis of exponential tumour growth curves. Growth delay was determined from the calculated time to tumour volume of 1000 mm³, normalised against the PBS control. ^*P<0.05 compared with PBS. ^†P<0.05 compared with PBS and 2 Gy. (D) Kaplan–Meier survival curves after experimental treatments in mice receiving no irradiation, and in mice receiving 2 Gy irradiation on day 2 (arrow). Survival analysis was based on follow-up until death or killing. Significant differences in 30-day survival were observed between mice receiving VNP pRA-TR and PBS, and between mice receiving VNP pRA-TR and 2 Gy and PBS and 2 Gy (log-rank test, P<0.05).

Figure 5

Suppression of tumour growth and enhanced survival by repeated dosing of VNP pRA-TR and γ-irradiation. (A) Regression curves for tumour growth after treatment with two doses of VNP pRA-TR and radiation (n=4), two doses of PBS and radiation (n=4), and PBS alone (n=11). Experimental groups received intravenous injection with PBS or VNP pRA-TR at days 0 and 6 (small arrows), followed by 2 Gy irradiation 2 days after experimental treatments (days 2 and 8; large arrows). Significant differences in tumour volume after redosing VNP pRA-TR and radiation were observed at all time points except days 0 and 2 in comparison with the PBS control group (P<0.05). (B) Estimates of time to tumour volume of 1000 mm³ show a delay of 30 days after two doses of VNP pRA-TR with irradiation in comparison with the PBS control (^*P<0.05; redosed VNP pRA-TR+2 Gy, in comparison with all other groups). (C) Kaplan–Meier survival curves for the treatment groups in (A). At 30 days of follow-up, 100% of mice survived with two treatments of VNP pRA-TR with 2 Gy irradiation, compared with 25% after repeated dosing with PBS and 2 Gy, and no survival from PBS controls (log-rank test, P<0.05).