Blockage of autophagy pathway enhances Salmonella tumor-targeting

Previous studies have shown that strains of Salmonella typhimurium specifically target tumors in mouse models of cancer. In this study, we report that tumor-targeting Salmonella typhimurium A1-R (A1-R) or VNP20009 induced autophagy in human cancer cells, which serves as a defense response. Functionally, by knockdown of essential autophagy genes Atg5 or Beclin1 in bacteria-infected cancer cells, the autophagy pathway was blocked, which led to a significant increase of intracellular bacteria multiplication in cancer cells. Genetic inactivation of the autophagy pathway enhanced A1-R or VNP20009-mediated cancer cell killing by increasing apoptotic activity. We also demonstrate that the combination of pharmacological autophagy inhibitors chloroquine (CQ) or bafilomycin A1 (Baf A1) with tumor-targeting A1-R or VNP20009 significantly enhanced cancer-cell killing compared with Salmonella infection alone. These findings provide a proof-of-concept of combining autophagy inhibitors and tumor-targeting Salmonella to enhance cancer-cell killing.


INTRODUCTION
The role of bacteria as an anticancer agent has been recognized from very early studies. The German physicians W. Busch and F. Fehleisen separately observed that cancers regressed following accidental erysipelas (Streptococcus pyogenes) infections in patients [1]. Later William B. Coley found that some cancer patients who developed post-operative infection of bacteria were cured from their tumors [2]. Salmonella typhimurium, a facultive anaerobe which can grow in both necrotic and live tumor tissue, is one of the most broadly studied tumor-targeting bacteria. Salmonella can colonize primary tumors as well as accumulate within metastases, rendering S. typhimurium an ideal anticancer agent [3][4][5].
VNP20009, a strain of Salmonella typhimurium, is genetically attenuated by deleting both purI and msbB and has been widely investigated [6,7]. The tumor-targeting and efficacy of VNP20009 has been demonstrated in a variety of animal models [3,8,9] and has shown to be safe in a Phase I clinical trial [10]. VNP20009 has also been used as a vector to deliver potentially therapeutic genes such as pro-drug converting enzymes, cytokines and other genes [11,12].
S. typhimurium A1 obtained after nitrosoguanidine (NTG) mutagenesis to induce auxotrophic mutations for leu and arg, selectively grew in tumor xenografts with rapid clearance in normal tissues [13]. S. typhimurium A1 is able to receive sufficient nutritional support from tumors. S. typhimurium A1-R was further isolated from A1 infection of a human colon tumor in nude mice. Compared to S. typhimurium A1, S. typhimurium A1-R has increased tumor-targeting capacity and antitumor efficacy against major types of cancers [4,[14][15][16]. Mechanistically, it was previously reported that tumor-targeting Salmonella could suppress tumor growth by inducing cell apoptosis, tumor necrosis as well as suppressing tumor angiogenesis [17,18].
Autophagy (macroautophagy) is an important cellular response that targets long-lived proteins, damaged organelles and pathogens for lysosomemediated degradation [19,20]. Previous studies showed that Salmonella typhimurium, an intracellular pathogen, could be targeted by autophagy after infecting mammalian cells such as macrophages and epithelial cells [21,22]. Autophagy targeting of S. typhimurium appears to protect host cells, which has been shown to inhibit the replication of the bacteria in infected cells [21,22]. Moreover, autophagy inhibitors such as wortmannin promote the release of S. typhimurium from vacuoles, and allow them to grow rapidly in the cytosol of cells [23]. More permissive intracellular growth of S. typhimurium in autophagy-deficient MEF-Atg5(−/−) cells than wildtype fibroblasts MEF-Atg5(+/+) also demonstrated that autophagy plays a role in restricting the replication of S. typhimurium in the cytosol [24].
Based on these findings, we hypothesized that tumor-targeting Salmonella such as A1-R or VNP20009 may induce autophagy in human cancer cells, and inactivation of autophagy pathway may significantly enhance Salmonella tumor-targeting. In this study, we verified this hypothesis, highlighting an innovative combination therapy of autophagy blockage with tumortargeting Salmonella.

Tumor-targeting Salmonella A1-R and VNP20009 induce autophagy in human cancer cells
To determine whether tumor-targeting Salmonella A1-R and VNP20009 induce autophagy in cancer cells, we first determined the effect of S. typhimurium A1-R or VNP20009 infection on the formation of autophagosome membranes in cancer cells by detection of the conversion of LC3 I (microtubule-associated protein 1 light chain 3) to lipidated LC3 II, a classical marker of autophagy induced at the early stage and degraded at the late stage of autophagy [24,25]. As shown in Figure 1A, S. typhimurium A1-R or VNP20009 infection induced conversion of LC3 I to LC3 II in HepG2 and Huh7 human liver cancer cells and AGS human gastric carcinoma cells.
Then we determined whether bacterial infection induces formation of puncta, another marker of autophagy [25,26], in EGFP-LC3-expressing HepG2 cells (HepG2-EGFP-LC3). We found that VNP20009 induced a classical puncta formation in infected cells compared with control cells ( Figure 1B). Furthermore, in the acridine-orange staining assay and FACS analysis for autophagy detection, we demonstrated that 17.6% and 13.0% of HepG2 cells infected with A1-R and VNP20009, respectively, had the accumulation of acidic vesicular organelles (AVO, a marker of autophagy), while only 2.5% of uninfected cells had AVOs ( Figure 1C). These results demonstrated that tumor-targeting Salmonella A1-R and VNP20009 induced autophagy in human cancer cells.

Autophagy restricts the growth of tumortargeting Salmonella in cancer cells
To determine the biological role of tumor-targeting Salmonella-induced autophagy, we blocked the autophagy pathway via siRNA, silencing the autophagy-essential genes Atg5 or Beclin1 and determined the effect on the intracellular bacterial multiplication. As shown in Figure 2A and Figure 2B, suppression of autophagy via Atg5 and Beclin1 knockdown significantly enhanced bacterial replication in both HepG2 ( Figure 2A) and Huh 7 ( Figure 2B) cells. To confirm this, we further utilized a MEF cell line with a knock-out of Atg5 locus (Atg5 −/− ) which leads to deficiency in autophagy, since Atg5 is essential for the early steps in autophagosome formation [27]. By determining the conversion of LC3-I to LC3-II with Western blot analysis, we firstly confirmed that S. typhimurium A1-R or VNP20009 induced autophagy in MEF-Atg5(+/+) (MEF cells expressing wild type Atg5) but not MEF-Atg5(−/−) ( Figure 2C). Then, we investigated the role of autophagy in the growth of S. typhimurium A1-R and VNP20009 by infecting the pair of MEF cell lines. As shown in Figure 2D, autophagy-deficient MEF-Atg5(−/−) showed greater intracellular S. typhimurium A1-R and VNP20009 proliferation when compared with wide-type MEF-Atg5(+/+). Collectively, these results demonstrate that the autophagic response induced by tumor-targeting Salmonella plays a protective role for infected cancer cells.

Salmonella-mediated anticancer therapy
Having established that autophagy restricts the growth of tumor-targeting Salmonella in cancer cells, we next investigated a hypothesis that blockage of autophagy pathway enhances the anticancer efficacy of S. typhimurium A1-R or VNP20009 and further retards cancer cell growth. As shown in Figure 3A and 3B, blocking the autophagy pathway via knocking down of Atg5 or Beclin1 significantly enhanced bacterial inhibition of proliferation in both HepG2 and Huh7 cells, while siRNA silencing of Atg5 or Beclin1 alone did not notably affect cell proliferation. Moreover, the enhanced cytotoxic effect of autophagy inhibition combined with bacterial treatment was further confirmed by using infected MEF-Atg5(+/+) (autophagy-competent) versus MEF-Atg5(−/−) (autophagy-deficient) cells. As shown in Figure 3C, morphologic observation and cell counting showed that MEFs-Atg5(−/−) showed greater cell death compared with wide-type MEFs-Atg5(+/+) cells upon Salmonella infection. These findings demonstrate that blockage of the autophagic pathway significantly enhances Salmonella tumor-targeting.

Blockage of autophagy pathway enhances Salmonella-induced apoptosis
Having established that blockage of the autophagy pathway combined with S. typhimurium A1-R or VNP20009 infection leads to higher intracellular bacteria colonization ( Figure 2) and enhanced cell killing (Figure 3), we then investigated how autophagy inhibition sensitizes cells to bacterial infection. PI staining and FACS analysis revealed that the blockage of autophagy via Atg5 or Beclin1 siRNA silencing significantly enhanced S. typhimurium A1-R and VNP20009-induced apoptosis, as demonstrated by the appearance of a sub-G1 peak in FACS analysis ( Figure 4A and 4B). To confirm the effect of the combination therapy on apoptotic induction, S. typhimurium A1-R or VNP20009 infected MEFs-Atg5(+/+) and MEFs-Atg5(−/−) were subjected to apoptosis analysis by PI staining and FACS analysis. Similarly, we found that S. typhimurium A1-R infection of MEFs-Atg5(−/−) induced 22.6% apoptotic cells, but only 3.8% in infected MEFs-Atg5(+/+), while VNP20009 infection led to 29.5% MEFs-Atg5(−/−) cells undergoing apoptosis, but only 2.3% in infected MEFs-Atg5(+/+) ( Figure 4C). These findings demonstrated that blockage of the autophagy pathway significantly enhanced tumor-targeting Salmonella-induced apoptosis.

Pharmaceutical inhibition of autophagy pathway enhances the anticancer efficacy of tumortargeting Salmonella
Since genetic inhibition of the autophagy pathway via siRNA silencing significantly enhanced cytotoxicity of tumor-targeting Salmonella, we further hypothesized that pharmacological inhibition of autophagy would enhance bacteria-induced cell-growth suppression. For this purpose, we used two classical autophagy inhibitors CQ and BafA1 which prevent the late stages of autophagic-flux by inhibiting the fusion of autophagosomes with lysosomes and subsequent lysosomal protein degradation [24,26]. We performed autophagic-flux analysis by treating bacterial infected cells with CQ and BafA1 respectively, then found the accumulation of LC3 II was remarkably enhanced compared with either drug-treated or bacteriallyinfected alone, indicating that autophagic flux induced by Salminella was intact ( Figure 5A). As a result, autophagy inhibition with either CQ or BafA1 significantly enhanced Salmonella-induced cell growth suppression compared with bacterial infection alone ( Figure 5B and 5C). Similarly, both CQ and BafA1 treatment greatly enhanced the cytotoxicity of S. typhimurium A1-R and VNP20009 in Huh7 cells ( Figure 5D-5F). These findings indicate that pharmacological inhibition of autophagy significantly enhances Salminella tumor-targeting.
In summary, we report for the first time that tumortargeting Salmonella induces autophagy in human cancer cells. The autophagy response of cancer cells upon bacterial infection serves as a cellular defense signal. More importantly, genetic or pharmaceutical inhibition of the autophagy pathway significantly enhances intracellular replication of tumor-targeting Salmonella and cytotoxicity. These findings revealed a novel interaction mechanism  Cells infected with VNP for 4 h or uninfected and treated with or without 1 nM Baf A1. Cell counting was conducted at the indicated time points. All results are representative of at least two independent experiments. All cell-counting results are presented as mean value ± SE from three independent experiments, each running in triplicate. ***p < 0.001. between cancer cells and tumor-targeting Salmonella from the perspective of autophagy and provide proofof-concept evidence for the combination therapy of autophagy inhibitors with tumor-targeting Salmonella, further demonstrating its future clinical potential [31].

Establishment of HepG2-EGFP-LC3 stable cell lines
HepG2 cells stably expressing an EGFP-LC3 fusion protein were established as described [28]. Briefly, cells were seeded in 6-well plates and transfected with 3 μg pEGFP-LC3 plasmid using Lipofectamine 2000 (Invitrogen). Cells with enhanced EGFP fluorescence were selected with an MoFlo TM XDP Cell Sorter (Beckman Coulter) and cultured in complete cell culture medium containing G418 at 200 μg/mL. The autophagy induced by A1-R or VNP20009 infection was measured by the appearance of punctate vesicle structure and photographed under fluorescence microscopy (Leica, Wetzlar, Germany).

Salmonella strains and bacterial infection
Tumor-targeting Salmonella typhimurium A1-R (AntiCancer, Inc., San Diego, CA) and Lipid A-modified (msbB -) and auxotrophic ( purI -) Salmonella typhimurium VNP20009 (ATCC, USA) were cultured in modified Luria-Bertani (LB) medium to late-log phase as previously described [4,6]. Cancer cells were cultured in DMEM with 10% FBS and antibiotics (pencillin and streptomycin), for approximately 16-24 hours (h) before bacterial infection. Medium was changed to antibiotic-free medium 2 h before infection. The bacteria were diluted in DMEM and added to the cancer cells in a ratio of 1000 MOI, and placed in an incubator at 37°C. After 2 h, the cells were rinsed three times with DMEM and then cultured in medium containing 10% FBS with gentamicin sulfate (150 μg/mL) to kill external bacteria. After incubating 2 h, the cells were washed three times with DMEM, and then cultured in medium containing 10% FBS. For siRNA experiments, cells were transfected with the indicated siRNA and infection was performed at indicated hours after gene knockdown. (please see below).

Colony-forming unit (cfu) assay
Cells were plated in 24-well plates at 7 × 10 4 cells/ well and cultured to approximately 70% confluence 16-24 h before bacterial infection. In order to determine the number of intracellular bacteria, these cells were washed by PBS twice, digested and lysed by 0.25% trypsin and 0.1% TritonX-100, respectively. Then, the diluted lysates were plated on agar. After incubated overnight at 37°C, CFUs were counted to estimate the intracellular titer of A1-R or VNP20009. This assay was repeated at least three times.

Propidium iodide (PI) staining
Cells were harvested and fixed in 70% ethanol at −20°C overnight, then stained with propidium iodide (36 μg/mL, Sigma) containing RNase ( 10 mg/mL; Sigma) at 37°C for 15 minutes, and analyzed by flow cytometry (CyAn TM ADP, Beckman Coulter) for cell-cycle profiling and apoptosis. Apoptosis was measured by the percentage of cells in the sub-G1 population. Data were analyzed with ModFit LT software.

Cell counting
Cells that were transfected with the indicated siRNAs were split 48 h post transfection and seeded into 12-well plates with 3 × 10 4 cells/well (VNP20009) or 4 × 10 5 cells/60 mm dish S. typimurium (A1-R). For cell counting at indicated time points, cell culture medium was removed and PBS was used to wash cells three times. The cells were then treated with 0.25% trypsin and resuspended with DMEM containing 10% FBS. The cell number was then counted. This assay was repeated at least three times.

Statistical analysis
All data are presented as mean ± SEM. The unpaired 2-tailed t test was used for the comparison of parameters between two groups and the statistical significance of differences between groups was assessed using GraphPad Prism5 software. Three levels of significance (*P < 0.05, **P < 0.01, ***P < 0.001) were used.