Efficacy of Tumor-Targeting Salmonella typhimurium A1-R against Malignancies in Patient-Derived Orthotopic Xenograft (PDOX) Murine Models

We developed tumor-targeting Salmonella typhimurium (S. typhimurium) A1-R, a facultative anaerobe that is an auxotroph of leucine and arginine. The tumor-targeting efficacy of S. typhimurium A1-R was demonstrated in vivo and vitro using several malignant cell lines including melanoma, sarcoma, glioma, breast, pancreatic, colon, cervical, prostate, and ovarian cancers. Our laboratory also developed a patient-derived orthotopic xenograft (PDOX) model by implanting patient-derived malignant tumor fragments into orthotopic sites in mice. We reviewed studies of S. typhimurium A1-R against recalcitrant cancers. S. typhimurium A1-R was effective against all PDOX tumor models tested and showed stronger efficacies than chemotherapy or molecular-targeting therapy against some tumors. Furthermore, the synergistic efficacy of S. typhimurium A1-R when combined with chemotherapeutic agents, molecular-targeting agents, or recombinant methioninase was also demonstrated. We suggest potential clinical uses of this S. typhimurium A1-R treatment.


Introduction
Dr. William B. Coley began bacterial therapy of cancer using Streptococcus pyogenes (S. pyogenes). He then developed Coley's toxin, a mixture of killed S. pyogenes and Serratia marcescens, achieving clinical responses for many malignant tumors [1].
Live bacteria can actively penetrate tumors to reach lesions distant from blood vessels, where chemotherapeutic drugs cannot be delivered, and damage malignant cells by several cytotoxic mechanisms [2]. Obligate anaerobes and facultative anaerobes can have intrinsic tumor-targeting ability because they can survive intratumor hypoxia. By contrast, traditional treatments such as chemotherapy or radiation therapy have reduced efficacy in hypoxic regions in tumors where malignant cells are generally quiescent. Moreover, the immune-suppressive tumor microenvironment is conducive to bacteria, and host immune responses against administered bacteria may enhance antitumor immunity [3][4][5]. These specific advantages in bacterial therapy can overcome the limit of traditional treatments, even exerting synergistic efficacy in combination with these treatments.
In clinical studies, limited antitumor efficacy has been shown thus far. Clostridium. butyricum (C. butyricum) M-55 resulted in oncolysis and accumulation in treated tumors [6,7]. Intravenous or intratumoral administration of C. novyi-NT induced intratumor infection and necrosis [7]. Moreover, Salmonella typhimurium (S. typhimurium) VNP20009 attenuated by msbB and purI mutations, showed bacterial colonization of treated melanomas [8,9]. Objective tumor response was not observed in any of these studies, however, several studies using S. typhimurium showed bacterial colonization in treated tumors after local or systemic administration [7].
S. typhimurium is a facultative anaerobe. Green fluorescence protein (GFP)-labeled S. typhimurium A1-R developed by our laboratory has high tumor-targeting efficacy, due to the leucine-arginine auxotroph, resulting in broad antitumor efficacy and limited adverse effects [10]. In a CT26 colon cancer-bearing BALB/c mouse experiment, S. typhimurium A1-R was quickly eliminated from normal organs including the liver and spleen seven days after intravenous administration [11]. In contrast, S. typhimurium A1-R remained at high density in CT26 tumors. In addition, tumors treated with S. typhimurium A1-R were significantly smaller than those treated with S. typhimurium VNP20009. The efficacy of S. typhimurium A1-R was demonstrated in orthotopic nude mouse models of prostate [12], breast [13,14], pancreatic [15,16], and ovarian cancer [17], as well as in sarcomas [18] and gliomas [19,20]. S. typhimurium A1-R was also effective in metastatic cancer models [21,22].
We review in the present report the therapeutic efficacy of S. typhimurium A1-R against malignancies in patient-derived orthotopic xenograft (PDOX) nude mouse models, in which human tumors are orthotopically implanted in mice, to examine future clinical applicability.

Overview
We established PDOX tumors as follows: when the original tumors derived from primary sites, the PDOX tumors were implanted into the same primary sites in mice; when the original tumors derived from recurrent or metastatic sites, the PDOX tumors were implanted into original, recurrent, or metastatic sites in mice.

Establishment of PDOX Models
Fresh patient-derived tumor tissue samples were immediately transported to the lab on ice. The samples were cut into 5 mm fragments and implanted subcutaneously into nude mice to establish. They were then passaged when they grew to approximately 10 mm. The grown tumors were cut into small fragments and implanted orthotopically into nude mice.  [27] i.t. 5 × 10 7 CFU Monotherapy Polytherapy A1-R > Ct A1-R → DOX > Ct (Ewing's sarcoma) [28] i.v./i.t. 5 × 10 7 CFU Monotherapy Polytherapy A1-R > Ct A1-R + DOX > Ct (Pleomorphic liposarcoma) [33] i.v. 5 × 10 7 CFU Monotherapy A1-R > DOX or Ct (USTS) [37] i.v. 5 × 10 7 CFU Monotherapy A1-R > DOX > Ct (Myxofibrosarcoma) [39] i.v. 5 × 10 7 CFU Monotherapy A1-R > DOX or Ct For example, procedures for the establishment of a Ewing's sarcoma PDOX model are shown in Figure 1 [28]. The established subcutaneous tumor was cut into small fragments ( Figure 1A,B). A seven mm skin incision was made on the right chest wall; then, a single tumor fragment was implanted orthotopically into the layer between the pectoral and intercostal muscles in the right chest wall of the nude mouse to establish a PDOX model ( Figure 1C-E). The implanted tumor was grown in the right chest wall in the PDOX model ( Figure 1F). Procedures to establish a Ewing's sarcoma patient-derived orthotopic xenograft (PDOX) model. A subcutaneously grown patient-derived Ewing's sarcoma was resected (A) and cut into small fragments on a dish (B). A 7 mm skin incision was made on the right chest wall and then a single tumor fragment was implanted orthotopically into the layer between the pectoral and intercostal muscles in the right chest wall of nude mouse (C). After the pectoral muscle was closed by 6-0 nylon sutures (D), the skin incision was sutured (E). (F) The established PDOX model 4 weeks after the orthotopic implantation [28]. White arrowheads indicate the subcutaneously grown tumor. Black arrowheads indicate the established PDOX tumor.

Cancer of Unknown Primary
For STSs, tumor tissue was implanted into the upper or lower extremities in five experiments ( Figure 2A) while the metastatic Ewing's sarcoma shown above was implanted into the chest wall of mice [25,27,28,33,37]. A metastatic melanoma was also implanted into the chest wall of mice [26,29,30,34]. Pancreatic cancers were implanted into the tail of the pancreas of mice ( Figure 2B) [23,24,36]. Osteosarcoma lung metastases originating from a primary lesion in the femur were implanted both in the femur and the lung [31,35]. A GIST tumor derived from recurrent lymph nodes and originating from a primary gastric tumor was implanted orthotopically into the gastric wall [32]. A PDOX of a cancer of unknown primary origin was established by implanting tumor fragments into the neck lymph node of mice, which was the metastatic site of the patient [38]. Procedures to establish a Ewing's sarcoma patient-derived orthotopic xenograft (PDOX) model. A subcutaneously grown patient-derived Ewing's sarcoma was resected (A) and cut into small fragments on a dish (B). A 7 mm skin incision was made on the right chest wall and then a single tumor fragment was implanted orthotopically into the layer between the pectoral and intercostal muscles in the right chest wall of nude mouse (C). After the pectoral muscle was closed by 6-0 nylon sutures (D), the skin incision was sutured (E). (F) The established PDOX model 4 weeks after the orthotopic implantation [28]. White arrowheads indicate the subcutaneously grown tumor. Black arrowheads indicate the established PDOX tumor.
For STSs, tumor tissue was implanted into the upper or lower extremities in five experiments ( Figure 2A) while the metastatic Ewing's sarcoma shown above was implanted into the chest wall of mice [25,27,28,33,37]. A metastatic melanoma was also implanted into the chest wall of mice [26,29,30,34]. Pancreatic cancers were implanted into the tail of the pancreas of mice ( Figure 2B) [23,24,36]. Osteosarcoma lung metastases originating from a primary lesion in the femur were implanted both in the femur and the lung [31,35]. A GIST tumor derived from recurrent lymph nodes and originating from a primary gastric tumor was implanted orthotopically into the gastric wall [32]. A PDOX of a cancer of unknown primary origin was established by implanting tumor fragments into the neck lymph node of mice, which was the metastatic site of the patient [38].  [27]. The tumor was exposed by cutting the skin. (B) A pancreatic cancer PDOX model was established by implanting a tumor into the tail of the pancreas of a red fluorescent protein (RFP)-expressing nude mouse [24]. Black arrowheads indicate grown tumors. White arrows indicate the RFP-expressing pancreas.

Antitumor Efficacy of S. typhimurium A1-R Compared to Untreated Control
The treatment efficacy of S. typhimurium A1-R was confirmed in all PDOX models ( Table 2). Significant tumor growth inhibition occurred after i.v., i.p., i.t., and i.a. injection of S. typhimurium A1-R. Interestingly, i.v. administration tended to be more effective compared to i.t. administration in the Ewing's sarcoma PDOX models [28]. In addition, S. typhimurium A1-R showed stronger efficacy administered i.a. than i.v. injection in the osteosarcoma lung metastasis PDOX models [31].

Histological Effects
Established tumors in the PDOX model had a similar morphologic appearance to the original patient tumor (Figure 4A,B). S. typhimurium A1-R-treated tumors showed extended necrosis compared to untreated tumors. As an example, S. typhimurium A1-R caused central tumor necrosis to a large extent in the Ewing's sarcoma PDOX model, while the untreated tumors grew without necrosis ( Figure 4C-G). In a study of an osteosarcoma lung metastasis PDOX model, S. typhimurium A1-R treatment resulted in changes in sarcoma cell shape but not necrosis [35]. S. typhimurium A1-R treatment in combination with cisplatinum and rMETase resulted in tumor necrosis. Moreover, S. typhimurium A1-R showed more extensive necrosis when combined with chemotherapy or

Antitumor Efficacy of S. typhimurium A1-R Compared to Untreated Control
The treatment efficacy of S. typhimurium A1-R was confirmed in all PDOX models ( Table 2). Significant tumor growth inhibition occurred after i.v., i.p., i.t., and i.a. injection of S. typhimurium A1-R. Interestingly, i.v. administration tended to be more effective compared to i.t. administration in the Ewing's sarcoma PDOX models [28]. In addition, S. typhimurium A1-R showed stronger efficacy administered i.a. than i.v. injection in the osteosarcoma lung metastasis PDOX models [31].

Histological Effects
Established tumors in the PDOX model had a similar morphologic appearance to the original patient tumor (Figure 4A,B). S. typhimurium A1-R-treated tumors showed extended necrosis compared to untreated tumors. As an example, S. typhimurium A1-R caused central tumor necrosis to a large extent in the Ewing's sarcoma PDOX model, while the untreated tumors grew without necrosis ( Figure 4C-G). In a study of an osteosarcoma lung metastasis PDOX model, S. typhimurium A1-R treatment resulted in changes in sarcoma cell shape but not necrosis [35]. S. typhimurium A1-R treatment in combination with cisplatinum and rMETase resulted in tumor necrosis. Moreover, S. typhimurium A1-R showed more extensive necrosis when combined with chemotherapy or molecular-targeting agents than S. typhimurium A1-R monotherapy on undifferentiated STS and melanoma PDOX models [27,29,30]. When compared to standard treatment, S. typhimurium A1-R induced a higher degree of necrosis in several PDOX models including pancreatic cancer, STS, osteosarcoma, and GIST [24,25,[31][32][33]37]. The i.a. administration of S. typhimurium A1-R led to more extensive necrosis than intravenous administration in the osteosarcoma PDOX model [31]. These results indicate that tumor necrosis was generally associated with tumor growth suppression. molecular-targeting agents than S. typhimurium A1-R monotherapy on undifferentiated STS and melanoma PDOX models [27,29,30]. When compared to standard treatment, S. typhimurium A1-R induced a higher degree of necrosis in several PDOX models including pancreatic cancer, STS, osteosarcoma, and GIST [24,25,[31][32][33]37]. The i.a. administration of S. typhimurium A1-R led to more extensive necrosis than intravenous administration in the osteosarcoma PDOX model [31]. These results indicate that tumor necrosis was generally associated with tumor growth suppression.

Adverse Effects Caused by S. typhimurium A1-R Treatment
None of the PDOX experiments showed adverse effects, in terms of significant weight loss, in mice treated with S. typhimurium A1-R compared to the untreated control.

Adverse Effects Caused by S. typhimurium A1-R Treatment
None of the PDOX experiments showed adverse effects, in terms of significant weight loss, in mice treated with S. typhimurium A1-R compared to the untreated control.

Conclusions and Future Perspectives
PDOX models are theoretically better at mimicking the human disease than heterotopic tumors, increasing the robustness of drug discovery studies. The present review demonstrates the strong antitumor efficacy of S. typhimurium A1-R against recalcitrant-caner PDOX models, indicating advantages that S. typhimurium A1-R may have over chemotherapy. Therefore, for rare malignancies or cancers of unknown primary origin, for which effective treatments have not been established, S. typhimurium A1-R treatment can be a good candidate. Moreover, for highly aggressive malignancies such as pancreatic cancer or melanomas, S. typhimurium A1-R was highly effective when combined with chemotherapy or molecular-targeting therapy. In addition, adverse effects were shown to be limited. Therefore, S. typhimurium A1-R treatment has clinical potential.
S. typhimurium penetrated the cancer cells in vitro by being attracted to small molecules such as ribose and serine [41,42]. The present review demonstrated how S. typhimurium A1-R targeted tumors in several PDOX mouse models. The antitumor efficacy of S. typhimurium A1-R against many kinds of cancer cell lines was demonstrated and suggested that S. typhimurium A1-R kills cancer cells directly [13,43,44]. Unchugonova et al. demonstrated that cancer cells infected by S. typhimurium A1-R expanded and burst, resulting in loss of viability [45]. Importantly, S. typhimurium A1-R, a facultative anaerobe, can grow under anaerobic condition [13,44]. As a result, S. typhimurium A1-R induces central tumor necrosis (Figure 4). In addition, Salmonella plays the role of inducing antitumor immune responses in an immunocompetent model [46]. Salmonella enhances both innate and adaptive immunity. Salmonella induces cytokine production, including interferon-γ, via Toll-like receptor 4 signaling [5]. Upregulated cytokines contribute to the recruitment of peripheral immune cells to the tumor [5,47]. Avogadri et al. demonstrated that intratumoral injection of S. typhimurium resulted in recruitment of CD8 + lymphocytes, CD4 + lymphocytes and B lymphocytes as well as macrophages and granulocytes in the tumor [3]. We also demonstrated that the antitumor efficacy of S. typhimurium A1-R was correlated with CD8 + lymphocyte infiltration into treated tumors in a pancreatic cancer syngeneic immunocompetent mouse model [48]. Moreover, S. typhimurium A1-R acts as a decoy. It induces the cancer cells to leave the chemo-sensitive state of the cell cycle, making the cancer cells highly sensitive to chemotherapy [49]. These facts suggest that various mechanisms are involved in the antitumor efficacy of S. typhimurium A1-R.
The present review discusses the antitumor efficacy of S. typhimurium A1-R against recalcitrant cancer PDOX models. Pre-clinical efficacy studies of S. typhimurium A1-R were completed and only a toxicity test needs to be performed to enable S. typhimurium A1-R to begin phase I clinical studies.