Efficacy of tumor-targeting Salmonella typhimurium A1-R in combination with anti-angiogenesis therapy on a pancreatic cancer patient-derived orthotopic xenograft (PDOX) and cell line mouse models

The aim of the present study was to examine the efficacy of tumor-targeting Salmonella typhimurium A1-R treatment following anti-vascular endothelial growth factor (VEGF) therapy on VEGF-positive human pancreatic cancer. A pancreatic cancer patient-derived orthotopic xenograft (PDOX) that was VEGF-positive and an orthotopic VEGF-positive human pancreatic cancer cell line (MiaPaCa-2-GFP) as well as a VEGF-negative cell line (Panc-1) were tested. Nude mice with these tumors were treated with gemcitabine (GEM), bevacizumab (BEV), and S. typhimurium A1-R. BEV/GEM followed by S. typhimurium A1-R significantly reduced tumor weight compared to BEV/GEM treatment alone in the PDOX and MiaPaCa-2 models. Neither treatment was as effective in the VEGF-negative model as in the VEGF-positive models. These results demonstrate that S. typhimurium A1-R following anti-angiogenic therapy is effective on pancreatic cancer including the PDOX model, suggesting its clinical potential.


INTRODUCTION
Pancreatic cancer is one of the most aggressive malignant tumors with a 23 % 1-year survival rate and < 2 % 5-year survival rate. The two most commonly used chemotherapy drugs approved for the treatment of pancreatic cancer are gemcitabine (GEM) and 5-fluorouracil . In recent years, little progress has been made in understanding and treatment of this disease [1].
In pancreatic cancer, overexpression of vascular endothelial growth factor (VEGF) and its receptors is associated with poor prognosis and increased metastatic potential [2,3]. Bevacizumab (BEV) is a humanized monoclonal VEGF-neutralizing antibody that many tumors become resistant to after a short period of response [4].
Our laboratory has previously developed a genetically modified bacterial strain, Salmonella typhimurium A1, selected for anticancer activity in vivo. S. typhimurium A1 is auxotrophic (leu/arg-dependent) [5]. The strain targets and grows in tumors. In contrast, normal tissue is cleared of these bacterial even in immunodeficient athymic mice. In order to increase the tumor-targeting capability of A1, the strain was re-isolated after infection of a human colon tumor growing in nude mice. The tumor-isolated strain, termed S. typhimurium A1-R, had increased targeting for cells in vivo as well as in vitro [6].
In the present study, we demonstrate the efficacy of S. typhimurium A1-R following antiangiogenic therapy with bevacizumab/gemcitabine (BEV/GEM) in patientderived orthotopic xenograft (PDOX) and cell line nudemouse models of pancreatic cancer.

S. typhimurium A1-R killed MiaPaCa-2 and Panc-1 pancreatic cancer cells in vitro
GFP-expressing S. typhimurium A1-R invaded MiaPaCa-2 and Panc-1 pancreatic cancer cells as early as 60 min, and replicated in the cells 120 min after infection. Both cancer cell types appeared to die via apoptosis 24 hr after bacterial infection ( Fig. 2A). In the clonogenic assay, the average colony area of MiaPaCa-2 treated with S. typhimurium A1-R was 2.95 ± 0.84 mm 2 compared to the untreated control, 6.03 ± 0.86 mm 2 . The average colony area of Panc-1 treated with S. typhimurium A1-R was 0.93 ± 0.31 mm 2 , compared to the untreated control, 1.91 ± 0.10 mm 2 . S. typhimurium A1-R significantly reduced colony formation of both pancreatic cancer cell lines compared to the control (MiaPaCa-2: p = 0.001 and Panc-1: p < 0.001) ( Fig. 2B and 2C).

Differential sensitivity to BEV in pancreatic cancer cell lines growing subcutaneously in nude mice
Real-time RT-PCR of VEGF-related gene expression ( Fig. 1) predicted that MiaPaCa-2 was BEV-sensitive and Panc-1 was BEV-resistant. The efficacy of BEV on these cell lines was first determined using a subcutaneous tumor mouse model (Fig. 3). The average tumor volume expression was determined with real-time RT-PCR. MiaPaCa-2 significantly expressed VEGFA more than the other cell lines (p < 0.001) except for BxPC-3 (p = 0.558) (A). MiaPaCa-2 significantly expressed VEGFR2 more than other cell lines (BxPC-3: p = 0.005, Capan-1: p < 0.001; Hs766T: p = 0.005; and Panc-1: p = 0.006) (C). VEGFR1 expression was not detected in MiaPaCa-2 and Capan-1 cell lines (B). Data for each treatment are represented as the mean ± SD. ** p < 0.01. of the MiaPaCa-2 tumors treated with BEV was 1.04 ± 0.24 mm 3 compared to the control which was 4.19 ± 1.21 mm 3 on Day 22. The average tumor volume of the Panc-1 tumors treated with BEV was 5.50 ± 2.62 mm 3 compared to the control which was 5.28 ± 0.99 mm 3 on Day 22. BEV significantly reduced the growth of MiaPaCa-2 compared to the untreated control group on Day 22 (p < 0.001) but did not reduce the growth of Panc-1 (Fig. 3).

Efficacy of BEV on microvessel density in pancreatic cell lines growing subcutaneously in mice
Subcutaneous tumors (MiaPaCa-2 or Panc-1) were treated with BEV (5 mg/kg, twice a week for 2 weeks) and tumor samples were removed 7 days after the last treatment. Frozen sections from each tumor were stained with anti-mouse CD31 antibody, and the MVD was determined by counting three fields at ×100 magnification of the highest vascular density. The average MVD of the MiaPaCa-2 tumors treated with BEV was 27.6 ± 7.45 compared to the control which was 65.1 ± 16.5. The average MVD of the Panc-1 tumors treated with BEV was 52.4 ± 8.43 compared to the control which was 57.4 ± 5.81. BEV significantly reduced the MVD of the MiaPaCa-2 tumor compared to the control (p = 0.002) (Fig. 4A, B, E) but did not significantly reduce MVD of the Panc-1 tumor (Fig. 4C, D, F). These results are consistent with the expression levels of VEGF-related genes ( Fig. 1), indicating that MiaPaCa-2 is BEV-sensitive and Panc-1 is BEV-resistant.
Furthermore, GFP-labeled S. typhimurium A1-R was detected in the MiaPaCa-2 tumor after BEV → S. typhimurium A1-R treatment (Fig. 5). Our data suggest that S. typhimurium A1-R is able to survive and multiply even in the hypo-vascular area of the tumor treated with BEV and cause tumor shrinkage.
A pancreatic cancer PDOX model was used to determine the efficacy of S. typhimurium A1-R treatment following anti-VEGF therapy. The patient tumor was a moderately-differentiated adenocarcinoma which expressed VEGF (Fig. 7A). Twenty mice with PDOXs were established and randomized to 4 groups and treated in the same way as the GFP MiaPaCa-2 orthotopic model. A large primary tumor and some metastases occurred in the control group. A few metastases were found in the The results demonstrated in the PDOX model, BEV/GEM → S. typhimurium A1-R sequential combination therapy was also more effective than the BEV/GEM combination.

Cell culture and establishment of a green fluorescent protein-labeled cancer cell line
Human pancreatic cancer cell lines Panc-1, MiaPaCa-2 and Hs766T were maintained in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA, USA). Human pancreatic cancer cell lines BxPC-3 and Capan-1 were maintained in RPMI-1640 (Invitrogen, Carlsbad, CA, USA). All cell lines were incubated at 37°C in a humidified atmosphere containing 5% CO 2 . Each medium was supplemented with 10% fetal bovine serum, streptomycin, and penicillin (complete medium). The cells were collected after trypsinization and stained with trypan blue (Sigma-Aldrich, St. Louis, MO). Only viable cells which excluded trypan blue were counted For GFP gene transduction of cancer cells, 70% confluent human pancreatic cancer (MiaPaCa-2) cells were used. In brief, cells were incubated with a 1:1 precipitated mixture of retroviral supernatants of PT67-GFP packaging cells which express the GFP gene linked to the G418 resistance gene and RPMI 1640 (Irvine Scientific, Santa Ana, CA) containing 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT) for 72 h. Fresh medium was replenished at this time. Cells were harvested with trypsin/ EDTA 72 h post-transduction and subcultured at a ratio of 1:15 into medium, which contained 200 µg/ml of the selective agent G418. The level of G418 was increased stepwise up to 800 µg/ml [35,36].

Preparation of bacteria
S. typhimurium A1-R was grown overnight on LB medium and then diluted 1:10 in LB medium. Bacteria were harvested at late-log phase, washed with PBS, and then diluted in PBS for use in experiments [6].

Confocal imaging of cancer cells infected with S. typhimurium A1-R in vitro
Both MiaPaCa-2 and Panc-1 pancreatic cancer cell lines were infected with S. typhimurium A1-R GFP in vitro. Pancreatic cancer cells were grown in 24-well tissue culture plates to a density of approximately 10 4 cells/well. S. typhimurium A1-R GFP were grown to late log in LB broth, diluted in cell culture medium and added to the cancer cells (1×10 7 CFU/ml) and incubated at 37°C. After 40 min, the cells were rinsed and cultured in medium containing gentamycin sulfate (20 μg/ml) to kill external, but not internal bacteria. The interaction of S. typhimurium A1-R GFP with cancer cells in vitro was observed with confocal microscopy (Fluoview FV1000, Olympus, Tokyo, Japan). The excitation source was a semiconductor laser at 473 nm for GFP. Fluorescence images were obtained using the 20x/1.0 XLUMPLFLN objective [37].

Clonogenic assay
MiaPaCa-2 and Panc-1 cells (1×10 3 ) were seeded in 35 mm dishes. S. typhimurium A1-R (1×10 7 CFU/ml) was added to the cancer cells and incubated at 37°C. After 40 min, the cells were rinsed and cultured in medium containing gentamycin sulfate (20 μg/ml). After 7-days culture, the cancer-cell colonies were fixed with ethanol and then stained with crystal violet. ImageJ was used to quantify the cell colonies.

Animals
Male athymic (nu/nu) nude mice (AntiCancer, Inc., San Diego) (4-6 weeks) were used in this study. Mice were kept in a barrier facility under HEPA filtration. Mice were fed with autoclaved laboratory rodent diet. All surgical procedures and imaging were performed with the animals anesthetized by intramuscular injection of a solution of 50% ketamine, 38% xylazine, and 12% acepromazine maleate (0.02 ml). All animal studies were conducted in accordance with the principals and procedures outlined in the NIH Guide for the Care and Use of Laboratory Animals under PHS Assurance Number A3873-1.

Subcutaneous pancreatic cancer cell implantation
Panc-1 and MiaPaCa-2 pancreatic cancer cells were harvested by trypsinization and washed twice with serum-free medium. Cells (2×10 6 in 100 μl serum-free medium) were injected subcutaneously within 30 min of harvesting, over the right and left flanks in male nude mice. Subcutaneous tumors were allowed to grow for 2-4 weeks until large enough for subsequent experiments or orthotopic implantation.

Assessment of microvessel density (MVD) in xenograft tumors
Frozen tumor sections (7 μm) were fixed with methanol. The sections were then treated for 30 min with hydrogen peroxide (0.3%) to block endogenous peroxidase activity. After incubation with normal goat serum 15%, the sections were incubated with anti-mouse CD31 (1:100; BD Pharmigen, San Jose, CA, USA) for 1 hour at room temperature. The primary antibodies were detected using anti-rat secondary antibodies and avidin/biotin/horseradish peroxidase complex (Vector Laboratories, Burlingame, CA, USA) for 30 min at room temperature. The labeled antigens were visualized with the DAB kit (DAKO Cytomation, Kyoto, Japan). The sections were counterstained with hematoxylin and examined using a BH-2 microscope (Olympus) equipped with an INFINITY1 2.0 megapixel CMOS digital camera (Lumenera Corporation, Ottawa, Canada). All images were acquired using INFINITY ANALYZE software (Lumenera Corporation) without post-acquisition processing. MVD was determined by counting three fields at ×100 magnification of the highest vascular density.

Specimen collection
Patient pancreatic tumor samples were procured with informed written consent and the study was conducted under the approval of the Institutional Review Board of the UC San Diego Medical Center.

Orthotopic tumor implantation
A small transverse incision (6-to 10-mm) was made on the left flank of the mouse through the skin and peritoneum. The tail of the pancreas was exposed through this incision, and a single tumor fragment (3mm 3. ) harvested from a subcutaneous tumor was sutured to the tail of the pancreas using 8-0 nylon surgical sutures (Ethilon; Ethicon Inc., NJ, USA). On completion, the tail of the pancreas was returned to the abdomen, and the incision was closed in one layer using 6-0 nylon surgical sutures [19,24].

Tissue histology
Tumor samples were removed with surrounding normal tissues at the time of resection. Fresh tissue samples were fixed in 10% formalin and embedded in paraffin before sectioning and staining. Tissue sections (3 μm) were deparaffinized in xylene and rehydrated in an ethanol series. Hematoxylin and eosin (H & E) staining was performed according to standard protocols.

Fluorescence in vivo imaging
The OV100 Small Animal Imaging System, containing an MT-20 light source (Olympus Biosystems, Planegg, Germany) and DP70 CCD camera (Olympus) were used for imaging GFP-labeled S. typhimurium A1-R and orthotopic tumors in live mice [38].

Detection of GFP-labeled S. typhimurium A1-R bacteria in tumors and organs
Tissues from subcutaneous tumors and normal organs (blood, spleen and liver) were removed at termination from the nude mice with subcutaneous tumors. S. typhimurium A1-R was extracted from the tumors and organs and cultured in LB agar for 24 hours, and imaged with the Olympus OV100.

Statistical analysis
PASW Statistics 18.0 (SPSS, Inc) was used for all statistical analyses. The Student's t-test was used to compare continuous variables between two groups. Analysis of variance models were used to compare multiple groups. A p value of ≤ 0.05 was considered statistically significant for all comparisons.