The facultative anaerobic, invasive Salmonella enterica serovar typhimurium (S. typhimurium) has been shown to retard the growth of established tumors. We wondered if a more effective antitumor response could be achieved in vivo if these bacteria were used as tools for delivering specific molecular antitumor therapeutics. Constitutively activated transcription factor signal transducer and activator of transcription 3 (STAT3) promotes the survival of a number of human tumors. In this study, we investigated the relative efficacies of attenuated S. typhimurium alone or combined with Stat3-specific small interfering RNA (siRNA) in terms of tumor growth and metastasis. The bacteria preferentially homed into tumors over normal liver and spleen tissues in vivo. S. typhimurium expressing plasmid-based Stat3-specific siRNAs significantly inhibited tumor growth, reduced the number of metastastic organs, and extended the life time for C57BL6 mice bearing an implanted prostate tumor, versus bacterial treatment alone. These results suggest that attenuated S. typhimurium combined with an RNA interference approach might be more effective for the treatment of primary as well as metastatic cancer. [Cancer Res 2007;67(12):5859–64]
- prostate tumor
- tumor regression
RNA interference (RNAi) is an evolutionarily conserved, posttranscriptional gene-silencing mechanism wherein a small interfering double-stranded RNA (siRNA) directs a sequence-specific degradation of its target mRNA ( 1). Because of their unparalleled target specificity, there has been an intensive effort to use siRNAs as therapeutics for various diseases, especially for cancer therapy. Because synthetic siRNAs can only transiently decrease the target gene expression in proliferating cancer cells ( 2), a sustained, localized supply of anticancer siRNAs is critical for imparting a strong therapeutic benefit. Plasmid-based expression of gene-specific small hairpin RNAs (shRNA), under the control of RNA polymerase III–dependent promoters (e.g., U6 and H1), produces a sustainable and economical source of siRNAs for therapeutic purposes. The shRNAs are processed intracellularly by the enzyme Dicer into siRNAs. Some groups have reported successful application in vivo following systemic administration with shRNA-encoding plasmid DNA ( 3– 5). Unfortunately, in most of these approaches, the therapeutics do not reach the tumors in effective doses, or distribution to unwanted sites and degradation by nucleases result in limited antitumor effect. The success of siRNAs as cancer therapeutics relies on the development of safe, economical, and efficacious in vivo delivery systems into tumor cells. Although siRNAs can be used as therapeutics in vivo, their intratumoral delivery, specifically across the plasma membrane of cells, is not achieved easily. Furthermore, they are ineffective at killing quiescent tumor cells that are distantly located from the vasculature and metastatic tumors because of their heterogeneous microenvironments. The ideal delivery system would be (a) nontoxic to normal cells and (b) able to deliver the therapeutic efficiently and specifically to the tumor.
The discovery that genes vectored by bacteria can be functionally transferred to mammalian cells has suggested the possible use of bacterial vectors as vehicles for gene therapy. Genetically modified, nonpathogenic bacteria have been used as potential antitumor agents, either to elicit direct tumoricidal effects or to deliver tumoricidal molecules ( 6– 9). Bioengineered attenuated strains of Salmonella enterica serovar typhimurium (S. typhimurium) have been shown to accumulate preferentially >1,000-fold greater in tumors than in normal tissues and to disperse homogeneously in tumor tissues ( 10, 11). Preferential replication allows the bacteria to produce and deliver a variety of anticancer therapeutic agents at high concentrations directly within the tumor, while minimizing toxicity to normal tissues. These attenuated bacteria have been found to be safe in mice, pigs, and monkeys when administered i.v. ( 8, 9, 12, 13), and certain live attenuated Salmonella strains have been shown to be well tolerated after oral administration in human clinical trials ( 14– 17). The S. typhimurium phoP/phoQ operon is a typical bacterial two-component regulatory system composed of a membrane-associated sensor kinase (PhoQ) and a cytoplasmic transcriptional regulator (PhoP; refs. 18, 19). phoP/phoQ is required for virulence, and its deletion results in poor survival of this bacterium in macrophages and a marked attenuation in mice and humans ( 18– 21). phoP/phoQ deletion strains have been employed as effective vaccine delivery vehicles ( 20– 22). More recently, attenuated salmonellae have been used for targeted delivery of tumoricidal proteins ( 7, 12). We report here the use of an attenuated phoP/phoQ null S. typhimurium as a delivery system for siRNA-based tumor therapy.
Materials and Methods
Construction of siRNA expression vectors. A siRNA target located in the SH2 domain of human signal transducer and activator of transcription 3 (Stat3; nucleotides 2144–2162; Genbank accession no. NM00315) was chosen for use herein based upon our previous study ( 23). The sequence of Stat3-specific hairpin RNA is given as follows: GCAGCAGCTGAACAACATGTTCAAGAGACATGTTGTTCAGCTGCTGCTTTTT. This oligonucleotide contains a sense strand of 20 nucleotides followed by a short spacer (loop sequence: TTCAAGAGA), the antisense strand, and five Ts (terminator). A scrambled siRNA (Ambion) was used as a negative control. Double-stranded DNA oligonucleotides were cloned into pGCsilencerU6/Neo/GFP, which also expresses a green fluorescent protein (GFP) gene (Jikai Chemical, Inc.), to generate plasmids pSi-Stat3 and pSi-Scramble ( Fig. 1A ).
Bacteria, cell culture, and stable cell line establishment. The attenuated S. typhimurium phoP/phoQ null strain LH430 was kindly provided by Dr. E.L. Hohmann ( 16). This strain was created from S. typhimurium strain SL1344 by deletion of the phoP/phoQ locus ( 24). Plasmids were electroporated into Salmonella before use. The mouse prostate cancer cell line RM-1 was obtained from the Shanghai Institute of Cellular Research. The cells were grown in Iscove's modified Dulbecco's medium (Invitrogen) with 10% fetal bovine serum. Cells were cocultured with recombinant bacteria (1 × 108 cfu) at 37°C for 30 min. Cell lines were washed and treated first with 100 μg/mL gentamicin to kill all extracellular bacteria and then with 5 μg/mL of tetracycline to prevent intracellular bacterial multiplication. Stable RM-1 clones, containing integrated plasmids, were selected and maintained by treating the cells with 200 μg/mL G418.
Tumorigenic assays. C57BL6 mice (n = 10 per group) were injected s.c. with the cell lines described above (2 × 106 cfu) into the upper flank. Tumor development was followed for 60 days. All animal studies were conducted in accordance with the principles and procedures outlined in the NIH Guide for the Care and Use of Laboratory Animals under assurance no. A3873-1.
Northern and Western blotting. Cell lysis, protein quantification, and Western blot analyses were carried out as described previously ( 23). Antibodies against Stat3, phosphorylated Tyr705-Stat3 (p-Stat3), cyclin D1, c-Myc, VEGF, and antimouse were obtained from Santa Cruz Biotechnology. Antibody against Bcl-2 was obtained from DAKO Biotech. Antibody against Ki-67 was obtained from Biogenex. Protein bands were detected using enhanced chemiluminescence (Amersham). Total RNA (20 μg) and 32P-labeled cDNAs of Stat3 and actin were used as probes. mRNA level was quantified using a Molecular Dynamics PhosphorImager.
Cell cycle, apoptosis, and proliferation assays. Cell cycle phase distribution was determined by flow cytometry. An Annexin V-CY3 apoptosis detection kit (Sigma) was used for detecting apoptosis. Tumor tissue sections from animals were used for H&E staining and terminal deoxynucleotidyl transferase–mediated nick-end labeling (TUNEL) assays, as described previously ( 23). Cell proliferation was assayed using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) staining kit (Sigma) as per manufacturer's protocol; the cell growth inhibition rate was calculated as follows: A = (1 − absorbance of experimental group / absorbance of control group) × 100%.
Antitumor activity of recombinant S. typhimurium on established prostate tumors. RM-1 cells were transplanted into mice s.c. to generate a primary tumor. After the development of a palpable tumor at the site of inoculation, tumors were excised, and the primary tumor fragments (1.5 mm3) were implanted by surgical orthotopic implantation in between two lobes of the prostatic gland in a recipient group of C57BL6 mice according to methods described previously ( 24, 25). Five days after implantation, mice were divided into three groups (n = 10 per group) and injected i.v. with 1 × 107 cfu of attenuated S. typhimurium carrying different plasmids. One set of mice was sacrificed 18 days after administration of bacteria, and tumors were excised, weighed, and measured diameter. Tumor metastases were counted in the liver, lung, spleen, kidney, and lymph nodes. The remaining mice were followed over 70 days for survival after treatment with different plasmids.
Analysis of bacterial distribution. Tissue samples from the primary tumor, the liver, the spleen, and from other sets of tumor-bearing mice were used for bacterial distribution and clearance studies. Normal and tumor tissues were excised, weighed, minced thoroughly, and homogenized. The diluted tissue homogenates were plated onto Luria-Bertani agar containing ampicillin in triplicate, and the colony count was determined on the next day. The tissues were also observed under a fluorescence microscope to determine the extent of bacterial infection. A portion of the tissues was also prepared for histochemical analyses.
Gelatin zymography assay. The gelatinolytic activities of matrix metalloproteinase-2 (MMP-2) were examined according to the method described previously ( 26).
Data analyses. The significance of the in vitro and in vivo data was determined using the Student's two-tailed t test. The significance of the differences between median data values was determined using the two-tailed Mann Test. P < 0.05 was deemed statistically significant. Data are presented as mean ± SD.
Results and Discussion
Construction of shRNA expression vectors and cell infection. To show the utility of attenuated S. typhimurium–carried, Stat3-specific siRNA for tumor therapy, we first generated plasmid vectors that express a Stat3-specific siRNA (Si-Stat3) and a control scrambled siRNA (Si-Scramble). Because of its potent antitumor effects, the target for Si-Stat3 was chosen from the SH2 domain of human Stat3 based on our earlier study ( 23). Synthetic oligonucleotides (20 bp) capable of coding for the Si-Stat3 and Si-Scramble siRNAs were cloned into pGCsilencerU6/Neo/GFP, a plasmid containing the GFP gene. The resultant plasmids pSi-Stat3 and pSi-Scramble were transformed into S. typhimurium and used for transfection into RM-1 cells ( Fig. 1A). Virtually identical transfection efficiencies were observed for each plasmid as determined by the expression of GFP in RM-1 cells ( Fig. 1B).
In this study, the invasive recombinant S. typhimurium carrying either the pSi-Stat3 or pSi-Scramble plasmids were directly cocultured with a mouse prostate carcinoma cell line (RM-1), and stable cell lines RM-Si-Stat3 and RM-Si-Scramble were established after G418 selection. The continued expression of GFP, in the absence of bacteria in the cell lines, indicates that the siRNA expression vectors were stably integrated into the host cell genome.
Effects of bacterially delivered shRNAs on cell growth and cycling. The ability of these constructs to silence Stat3 was determined next using Western and Northern blot analyses. The Stat3 mRNA level in RM-Si-Stat3 was reduced to ∼13% of that observed in RM-Si-Scramble ( Fig. 2A and B ). Western blot analyses with native Stat3 (Stat3)– and phosphorylated Tyr705 Stat3 (p-Stat3)–specific antibodies also showed a strong inhibition of Stat3 or p-Stat3 proteins to ∼18% or 10%, respectively, in RM-Si-Stat3 ( Fig. 2C and D) compared with RM-Si-Scramble. The percentages of STAT3 knockdown observed in Northern versus Western blot analyses are similar and statistically significant. Thus, the bacterially introduced Si-Stat3 specifically knocks down the expression of Stat3. We also examined the effects of siRNAs on cell growth and cycling. Cells were stained with acridine orange and subjected to flow cytometry. Stat3-siRNA induced significant apoptosis (∼23-fold) compared with the pSi-Scrambled control ( Table 1 ). A further analysis of the flow cytometric data also showed that cells transfected with pSi-Stat3 accumulated significantly in G1 phase compared with the control ( Table 1). These findings indicate that inhibition of Stat3 promotes both cessation of cell growth and enhancement of cell death. Because the Salmonella have been eliminated from the stable cell line by treatment with antibiotics, the effects on cell growth and cycling from the Si-Scramble control was equivalent to the uninfected mock group. Cells transfected with pSi-Stat3 grew slower and showed strong apoptosis ( Fig. 3A ) compared with those transfected with pSi-Scramble. Cells transfected with pSi-Stat3 became confluent 6 days after seeding, in contrast to the control group, which reached confluence by 4 days. In a separate experiment, cellular metabolic activity (as an indicator of cell viability) was measured using MTT assays in RM-1 cells transfected with the various plasmids. MTT data, expressed as tumor cell viability, were significantly decreased in the cells treated with the pSi-Stat3 compared with the control groups at day 6 (P < 0.05, n = 3; Fig. 3B). Stat3 has been shown to play a key role in promoting the cell cycle, proliferation, differentiation, and inhibition of apoptosis ( 27). Persistently active Stat3 and its overexpression have been detected in a wide variety of human tumors ( 28), including prostate cancer ( 29). Constitutively active Stat3 promotes cell growth and survival via an overexpression of downstream targeted genes, such as the antiapoptotic Bcl-2, cell cycle regulators cyclin D1 and c-Myc, and inducers of tumor angiogenesis VEGF and MMP-2 ( 30– 35). We, therefore, examined if the expression of these genes was altered by Si-Stat3. The expression of Bcl-2, cyclin D1, c-Myc, VEGF, and MMP-2 was significantly knocked down in the presence of Si-Stat3 but not Si-Scramble ( Fig. 3C and D). Thus, the Stat3-specific shRNA interferes with the expression of tumor growth–promoting factors and decreases tumor cell survival.
Tumorigenic properties of RM-Si-Stat3 cells in vivo. We next examined the tumorigenic properties of RM-Si-Stat3 cells in vivo. C57BL6 mice (n = 10) were injected with 2 × 106 cells via the s.c. route into the upper flank, and tumor growth was monitored for 60 days. Mice transplanted with RM-Si-Scramble cells developed tumors at the injection sites by 21 ± 3.6 days. In contrast, no tumors formed in the group injected with RM-Si-Stat3. Thus, the blockade of Stat3 reverses tumorigenicity of RM-1 prostate cancer cells.
Inhibition of prostate tumor growth and metastasis in vivo by bacterially delivered shRNAs. Although salmonellae have been effective in retarding the growth of established tumors, complete tumor regression has never been proven. We, therefore, first studied the effects of S. typhimurium alone or combined with Stat3-specific siRNA in terms of prostate tumor growth and metastasis. To this end, we employed a C57BL6 mouse tumor implant model. A primary tumor was first established with RM-1 prostate carcinoma cells. Upon development of palpable s.c. tumors at the sites of inoculation, the tumor was excised and used for initiating primary prostate tumor development via an orthotopic surgical implantation of tumor tissues into recipient naive mouse prostates. Five days after tumor implantation, mice were divided into four groups (n = 10 per group) and then injected with 1 × 107 cfu of attenuated Salmonella carrying different plasmids via the tail vein. Eighteen days after bacterial injection, mice were sacrificed, and the tumors were excised, weighed, and measured. As shown in Table 2 , mice treated with buffer alone (mock control) developed primary tumors with a mean volume of 2,458.51 ± 602.18 mm3. In mice treated with Salmonella-Si-scramble, tumors grew to a volume of 589.22 ± 380.34 mm3. In mice treated with Salmonella without any plasmid, the tumor grew to a comparable volume of 585.44 ± 220.21 mm3. Thus, the bacteria carrying the scrambled-siRNA did not significantly affect tumor growth any differently compared with the Salmonella vector alone. However, mice treated with Salmonella-Si-Stat3 developed tumors with a median reduced volume of 216.42 ± 134.15 mm3. Remarkably, tumors completely disappeared in one third of mice in this group over 18 days. The differences in tumor size between buffer control versus Salmonella-Si-scramble (P < 0.05) and buffer control versus the Salmonella-Si-Stat3 group (P < 0.01) were statistically very significant. The differences between Salmonella-Si-scramble or Salmonella alone versus Salmonella-Si-Stat3 group were also statistically significant (P < 0.05). In summary, ∼3.9-fold higher tumor suppressive effect can be achieved with a single dose of bacteria transformed with a siRNA expression vector than those treated with Salmonella alone or Salmonella carrying Si-Scramble control, and ∼11.4-fold higher than those treated with buffer control ( Fig. 4A, yellow arrowhead ; Table 2). Thus, attenuated Salmonella alone exert an antitumor effect, which can be further enhanced by genetically modifying these organisms in combination with Stat3-specific siRNA expression.
In addition to the primary tumor, metastases into liver, lung, spleen, kidney, and lymph nodes were examined in the recipient mice. A robust 84% reduction (P < 0.01) in the numbers of metastases in the Salmonella-Si-Stat3–treated mice (Supplementary Table S3) was observed. Tumor metastases occur primarily through tumor angiogenesis, aggressive growth of the primary tumor, and an extravasation of the tumor cells ( 36). Secretion of extracellular proteases by the tumor plays an important role in metastasis ( 36, 37). Among these, MMP-2/gelatinase A is believed to be essential for malignant behavior of cancer cells, such as rapid growth, tissue invasion, and metastasis ( 36, 37). Consistent with this observation, we found that the MMP-2 activity in RM-1 cells significantly decreased after treatment with pSi-Stat3 compared with mock or pSi-Scramble (P < 0.05; Supplementary Fig. S1). Furthermore, blockade of Stat3 correlated with a reduction of expression of the Ki-67 protein, a proliferation-associated antigen ( 38). Immunohistochemical analyses for Stat3 and Ki-67 expression in the RM-1 tumor cells after transfection with Si-Scramble and in untreated RM-1 tumor cells were highly positive for Stat3 and Ki-67. In contrast, RM-1 tumor cells treated with pSi-Stat3 stained weakly for Stat3 and Ki-67 ( Fig. 4B).
Tumors from mice treated with pSi-Scramble or pSi-Stat3 were excised for H&E staining and analyzed with TUNEL assays ( Fig. 4C). pSi-Stat3–treated tumors show massive apoptosis with sparsely dispersed chromatin, several TUNEL-positive cells, and some necrotic regions compared with the Si-scramble control, which showed a finely granular cytoplasm with evenly dispersed chromatin and no TUNEL-positive cells. These data show that the Stat3 siRNA carried by Salmonella exerts a strong apoptotic antitumor effect in vivo.
The attenuated S. typhimurium expressing a Stat3-specific siRNA exerts a robust antitumor effect. To further show the therapeutic utility of Salmonella-delivered siRNAs, tumor-bearing mice were injected with Salmonella carrying various plasmids or buffer. Mice were observed for 70 days. As shown in Supplementary Fig. S2, all mice (n = 10) injected with buffer were dead before 30 days. In contrast, the mice injected with Salmonella-Si-Stat3 and Salmonella-Si-Scramble had nine and six surviving mice at 70 days, respectively. These data clearly show that the attenuated Salmonella expressing a Stat3-specific siRNA exerts a robust antitumor effect.
Recombinant bacterial distribution in C57BL6 tumor-bearing mice. To determine if the potent antitumor effects of Salmonella with shRNA vectors was due to a preferential homing of bacteria into tumor tissue, we monitored the kinetics of bacterial distribution in C57BL6 tumor-bearing mice at specified times after injection of bacteria (Supplementary Fig. S3A). Twenty-four hours after injection, similar numbers of bacteria were found in the liver, spleen, and tumors in tumor-bearing mice. The bacterial count (cfu) increased in tumors and decreased in the liver and spleen within 48 h after administration. By day 5, the number of bacteria in tumors increased significantly; the tumor to liver or tumor to spleen cfu ratio was 1,000:1 and 5,000:1, respectively, on average. By day 15, far more bacteria could be seen in the tumor compared with the liver, and no bacteria could be found in spleen tissues. On day 10, by using GFP expression as a marker, the bacterial distribution was also observed as markedly high in tumor tissue sections compared with those in spleen and liver tissue sections (Supplementary Fig. S3B). At present, it is not clear why or how Salmonella specifically home to the tumor. Both characteristics of Salmonella and the heterogeneous microenvironments in solid tumors may combine to allow these bacteria to deliver therapeutic molecules preferentially to tumors. These characteristics may include (a) bacterial motility leading to uniform penetration within tumors; (b) hypoxic regions, an environment to which facultative anaerobic salmonellae are well adapted and can multiply, and in which macrophages, neutrophils, and granulocytes, effectors of bacterial clearance, are reduced in number ( 39); (c) both antibodies and serum complement components, which together can be lytic to salmonellae, are greatly restricted from the tumor environment by the irregular vasculature and positive pressure that exist inside tumors ( 40); (d) nutrients, such as high availability of glucose in aggressively growing tumors, may promote locally increased bacterial growth ( 41); and (e) Salmonella may induce apoptosis in infected macrophages ( 42) at the tumor margins leading to increased antitumor inflammatory responses. An important recent advance in this field is the development of live, attenuated Salmonella vectors for DNA vaccine delivery ( 43). The mechanisms involved in Salmonella delivery of DNA vaccine plasmids to the cytosol of mammalian cells is yet unclear ( 44). However, several lines of evidence suggest that this bacterium can deliver nucleic acid vaccines in vivo, which elicit impressive levels of specific antibody response, T-cell proliferation, and CTL responses ( 45). Lastly, live Salmonella infection, but not Escherichia coli, induces the expression GRIM-19 ( 46), a protein inhibitor of STAT3 ( 47, 48). Thus, the potent antitumor effect of Salmonella can, in part, be due to an inhibition of STAT3 activity by increased GRIM-19 in the tumor. When these bacteria are combined with Stat3-specific siRNAs, a double-edged inhibitory effect may be exerted on STAT3 in vivo.
Our results provide the first convincing evidence that Salmonella can be used for delivering plasmid-based siRNAs into tumors growing in vivo. The Stat3-siRNAs carried by an attenuated S. typhimurium exhibit tumor suppressive effects not only on the growth of the primary tumor but also on the development of metastases, suggesting that an appropriate attenuated S. typhimurium combined with the RNAi approach may offer a clinically feasible approach for cancer therapy. Ultimately, a live, attenuated Salmonella parenteral delivery system would likely be endotoxic in humans unless an msbB mutation was introduced, as reported previously ( 11).
Grant support: Grants-Specialized Programs of Research in Prostate Cancer from the Ministry of Science and Technology of the People's Republic of China (2004DFB02000), International Cooperation Agency (JICA), and National Cancer Institute grants CA78282 and CA105005 (D.V. Kalvakolanu).
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.
We thank Dr. E.L. Hohmann (Massachusetts General Hospital, Harvard Medical School, Boston, MA) for supplying the S. typhimurium strain LH430.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received January 8, 2007.
- Revision received February 28, 2007.
- Accepted March 26, 2007.
- ©2007 American Association for Cancer Research.