An Engineered Probiotic Platform for Cancer Epitope‐Independent Targeted Radionuclide Therapy of Solid Tumors

Abstract

Targeted radionuclide therapy (TRT) is an emerging therapeutic modality for the treatment of various solid cancers. Current approaches rely on the presence of cancer‐specific epitopes and receptors against which a radiolabeled ligand is systemically administered to specifically deliver cytotoxic doses of α and β particles to tumors. In this proof‐of‐concept study, tumor‐colonizing Escherichia coli Nissle 1917 (EcN) is utilized to deliver a bacteria‐specific radiopharmaceutical to solid tumors in a cancer‐epitope independent manner. In this microbe‐based pretargeted approach, the siderophore‐mediated metal uptake pathway is leveraged to selectively concentrate copper radioisotopes, ⁶⁴Cu and ⁶⁷Cu, complexed to yersiniabactin (YbT) in the genetically modified bacteria. ⁶⁴Cu‐YbT facilitates positron emission tomography (PET) imaging of the intratumoral bacteria, whereas ⁶⁷Cu‐YbT delivers a cytotoxic dose to the surrounding cancer cells. PET imaging with ⁶⁴Cu‐YbT reveals persistence and sustained growth of the bioengineered microbes in the tumor microenvironment. Survival studies with ⁶⁷Cu‐YbT reveals significant attenuation of tumor growth and extends survival of both MC38 and 4T1  tumor‐bearing mice harboring the microbes. Tumor response to this pretargeted approach correlates with promising anti‐tumor immunity, with noticeable CD8⁺ T:T_reg cell ratio. Their strategy offers a pathway to target and ablate multiple solid tumors independent of their epitope and receptor phenotype.

Engineered Escherichia coli Nissle capable of internalizing ⁶⁴Cu and ⁶⁷Cu‐tagged siderophores for imaging and targeted therapy of tumors is developed. This microbial platform attracts systemically administered radiolabeled siderophores to the tumor microenvironment inducing immunomodulatory and cancer cell killing effects. Their approach can potentially be used to image and treat tumors regardless of the presence of targetable cancer‐specific epitopes.

1. Introduction

Significant developments in novel chemotherapeutics, biologics, immune checkpoint inhibitors, and cell‐therapies have been made in the last few decades; however, many patients with cancer have progression refractory to conventional therapies. In addition, off‐target toxicities present a formidable challenge in the curative and non‐curative settings, particularly in patients with solid tumors. While radiotherapy has been used to treat solid cancers for more than a century, it has primarily involved administration of radiation from outside the body to kill malignant cells inside. Targeted radionuclide therapy (TRT) is an emerging therapeutic modality owing to its numerous favorable attributes over existing approaches.^{[ 1} , ² , ^{3 ]} Unlike external beam radiation therapy, which is unable to address systemic metastases, TRT involves administration of specific radiopharmaceuticals that selectively bind to cancer cells.^{[ 4} , ^{5 ]} Once bound, the radioactive nuclides emit cytotoxic doses of α and β particles to the surrounding cancer cells. In addition, TRT facilitates visualization of the radiopharmaceutical probes by imaging techniques such as positron emission tomography (PET) and single‐photon emission computed tomography, for confirming probe localization, and enables an image‐guided theranostic approach to monitoring treatment outcomes.^{[ 5 ]}

Although TRT has proven to be effective when other standard approaches have failed, it should be noted that this radiotherapeutic modality has been employed only for cancers with available target epitopes. For instance, even though ²²⁷Th‐conjugated anti‐HER2/neu demonstrates excellent anti‐tumor efficacy,^{[ 6 ]} it is unlikely to be effective in patients where the HER2 receptor undergoes endocytosis, and hence is not available on the cancer cell membrane for binding.^{[ 7 ]} Even when the target antigens are available, the specificity of the therapeutic probe has to be high to avoid delivery of cytotoxic radiation to non‐cancerous cells that have a basal level of expression of the same epitopes, albeit not as high as the cancer cells. For example, a radiolabeled GD2‐(a major ganglioside present in human neuroblastoma) targeting antibody, that demonstrates favorable tumor: non‐tumor ratio in a pretargeted approach,^{[ 8 ]} might still induce significant central nervous system toxicities to pediatric patients suffering from neuroblastoma due to shared expression of GD2 in vital normal tissues.^{[ 9 ]} Furthermore, in traditional TRT, since the emitted radiation originates from the cancer cell‐bound radionuclide, the probability of killing cells depends on the number of target cells. In solid cancers, such as pancreatic cancer, with high density of extracellular matrix, fewer cancer cells potentially lead to a smaller fraction of the emitted energy deposited onto target cells. Thus, there is a need to investigate alternative targeted and pretargeted strategies to address shortcomings of the current TRT paradigm.

A multitude of studies have revealed that some tumors can host unique microbiomes.^{[ 10 ]} Indeed, the immune‐privileged tumor cores have been known to be hypoxic and hence allow facultative anaerobic bacteria to colonize.^{[ 11 ]} Taking advantage of these inherent properties, we and others recently engineered tumor‐colonizing Escherichia coli (E. coli) Nissle 1917 (EcN) to act as intratumoral bioreactors that continually produce a range of anticancer payloads to promote tumor regression in preclinical models.^{[ 12} , ¹³ , ^{14 ]} Besides its ability to deliver therapeutic molecules, EcN is also an excellent template for molecular imaging in vivo. We recently demonstrated in a preclinical study that EcN can be noninvasively tracked via PET imaging, using a ⁶⁴Cu‐labeled bacterial siderophore, yersiniabactin (YbT).^{[ 15 ]} The ⁶⁴Cu‐YbT probe specifically accumulates in bacteria like EcN that express the ferric YbT uptake receptor (FyuA). We believe that there is a strong rationale to integrate aspects of microbiology to nuclear medicine and explore a pretargeted theranostic strategy using tumor‐homing microbes and their cognate siderophores. We, therefore, postulated that by switching to a therapeutic radionuclide, ⁶⁷Cu, we could deliver ⁶⁷Cu‐YbT to probiotic EcN localized in solid tumors and elicit a “bystander” cytotoxic effect on the cancer cells, irrespective of their biomarker profile. ⁶⁴Cu‐YbT and ⁶⁷Cu‐YbT can thus be used interchangeably to facilitate image‐guided TRT of solid tumors (Figure  1A,B).

Figure 1.

Engineered EcN bind and retain Cu‐YbT specifically via FyuA. A) Reaction mechanism depicting complexation of ^64/67Cu by yersiniabactin (YbT). B) Schematic highlighting FyuA‐specific delivery of Cu‐YbT radiopharmaceuticals by EcN‐based pretargeted cancer theranostic platform. C) Immunoblot confirming FyuA expression by engineered EcN. D) Schematic of EcN constructs used for E) quantitative (regression curve) assessment of FyuA overexpression. F) Dissociation curve of ⁶⁴Cu‐YbT from EcN‐fyuA↑Data in (E) and (F) presented as mean ± s.d. (n = 3); data in (E) (10⁹ cfu) analyzed by one‐way ANOVA and Dunnett's T3 multiple comparison test with Brown–Forsythe and Welch's correction, alpha = 0.05. *p < 0.05, **p < 0.01. Source Data: Unprocessed Western Blot for (C) is shown in Figure S6, Supporting Information.

2. Results

2.1. EcN Internalization of ^64/67Cu Can Be Maximized by Overexpression of FyuA

We had previously observed that E. coli UTI89 (pathogen) exhibited an approximately threefold higher uptake of ⁶⁴Cu‐YbT compared to EcN, a nonpathogen.^{[ 15 ]} An FyuA knockout (KO) strain of EcN, EcNΔfyuA, showed minimal uptake of ⁶⁴Cu‐YbT in vivo, which was consistent with in vitro analyses from independent studies.^{[ 16} , ^{17 ]} Since FyuA facilitates internalization of the metal‐siderophore complex, we postulated that overexpressing FyuA in EcN will maximize uptake of Cu‐YbT. We redesigned the pSF‐OXB20 plasmid by inserting the FyuA gBlock downstream of the constitutive OXB20 promoter and transformed EcN with our new plasmid (pZVS1) to overexpress FyuA (Figure S1, Supporting Information). Since wild‐type (WT) EcN naturally expresses FyuA to some extent, we relied on the FLAG tag to confirm plasmid‐based expression of FyuA by the positive transformants (EcN‐fyuA↑) (Figure 1C). While immunoblotting allowed us to qualitatively confirm plasmid‐based expression of FyuA, we emphasized on radioactive Cu‐YbT uptake studies to prove both expression and functionality of additional FyuA in our engineered probiotic bacterium. We incubated variable colony forming units (cfu) of EcNΔfyuA, EcN WT, or EcN‐fyuA↑ in growth media supplemented with the same radioactive amount of ⁶⁴Cu‐YbT (0.04–0.06 Mbq) for 2 h (Figure 1D). Afterward, we determined cell‐associated ⁶⁴Cu‐YbT for each group of EcN. The underlying assumption was that the total number of FyuA receptors available for ⁶⁴Cu‐YbT binding will positively correlate with the bacterial cfu. We observed that the KO strain accumulated the least amount of ⁶⁴Cu‐YbT at all cfu (Figure 1E). Minimal cell‐associated ⁶⁴Cu‐YbT in EcNΔfyuA conformed to the results of previous studies and indicated that FyuA is required for bacterial uptake of Cu‐YbT.^{[ 15} , ¹⁶ , ^{17 ]} Although EcN WT and EcN‐fyuA↑ demonstrated similar uptake at low cfu, the latter had a significantly higher (p < 0.05) uptake of the radioactive metal complex when the bacterial population exceeded 10⁷ cfu. Since EcN WT genomically expresses FyuA, it accumulated ⁶⁴Cu‐YbT at similar levels to EcN‐fyuA↑ when the bacterial population was low, though the latter appeared to outcompete the wild‐type strain at cfu that are typically observed in the tumor microenvironment following administration of the engineered probiotic for cancer therapy.^{[ 18 ]} Thus, overexpressing FyuA increases cellular uptake of Cu‐YbT in EcN. Finally, we determined the dissociation profile of ⁶⁴Cu‐YbT from EcN‐fyuA↑ over a 24‐h period (Figure 1F). At least 80% of the intact probe was found associated to the cells, which indicates that the genetically encoded E. coli construct bind and retain its cognate Cu‐labeled probe, with high specificity and affinity.

2.2. Engineered EcN Colonize and Persist in the Tumor Microenvironment

We performed imaging studies to recapitulate our in vitro observations in vivo. We administered EcN‐fyuA↑ and EcNΔfyuA intratumorally in a syngeneic MC38/C57BL6 colon cancer model; then we administered ⁶⁴Cu‐YbT retro‐orbitally for PET/CT imaging. Transformation of EcN‐fyuA↑ with pGEN‐luxCDABE (luciferase) allowed us to confirm bacterial localization in the tumor via bioluminescence imaging (BLI) (Figure  2A). Following PET/CT imaging, we noticed significantly higher signals in tumors with EcN‐fyuA↑ compared to those with EcNΔfyuA. Ex vivo biodistribution (BioD) analyses also revealed statistically pronounced (p < 0.01) accumulations of ⁶⁴Cu‐YbT in tumors hosting FyuA‐overexpressing EcN. In accordance with our previous observations,^{[ 15 ]} the probe was cleared primarily by the liver and kidneys with minimal accumulation in rest of the major organs (Figure 2B).

Figure 2.

Engineered EcN localize and persist in tumor microenvironment. A) BLI, PET/CT, and ex vivo BioD analyses depicting EcN‐fyuA↑ localization in MC38 tumors and FyuA‐specific retention of ⁶⁴Cu‐YbT 1‐day post intratumoral administration of bacteria. B) General biodistribution of ⁶⁴Cu‐YbT. C) qRT‐PCR assay confirming presence of EcN‐fyuA↑ in MC38 tumors 1‐ and 7‐days post intratumoral administration of bacteria. D) BLI, ex vivo BioD, and PET/CT demonstrating presence of bacteria in MC38 tumors 18‐days after intratumoral injection. Data of bar‐chart in (A) presented as mean ± s.d. (n = 3–4) analyzed by Welch's t‐test. **p < 0.01.

To determine the extent of colonization, persistence, and growth of intratumorally administered bacteria, we performed longitudinal genomic, cell, and imaging analysis. We administered EcN‐fyuA↑‐lux in MC38 tumors before we euthanized the mice and harvested the major organs to check for bacterial presence 1‐ and 7‐days post‐administration via quantitative real‐time polymerase chain reaction (qRT‐PCR). Using lux‐specific primers, we were able to confirm significant bacterial presence exclusively in the tumors at both time‐points (Figure 2C). All readings for the major organs were below the limit of detection for our analysis, which was 2.5 ng µL⁻¹. Moreover, the qRT‐PCR analyses indicated negligible bacterial extravasation out of tumors, thus preventing localization in off‐target tissues, which is consistent with previous reports of engineered EcN constructs.^{[ 18 ]} We used a separate cohort of mice to determine EcN‐fyuA↑ population in the tumors 2‐days after bacterial administration. After euthanizing the mice, we disaggregated the tumors and plated serial dilutions of the homogenates on antibiotic supplemented LB plates. We observed that there were approximately ten times more bacteria (5.9 ± 2.8 × 10⁷ cfu) in the tumors than the amount injected (5 × 10⁶ cfu) 48 h before. We then performed imaging to check for microbial presence on day 18, when the tumor volume was fairly large. BLI revealed varying levels of EcN‐fyuA↑, which correlated positively with the amount of ⁶⁴Cu‐YbT accumulation in the tumors (Figure 2D). PET/CT imaging corroborated the fact that our engineered bacteria are not only able to maintain their population inside solid tumors, but also achieve a sustained growth pattern, consistent with tumor growth.

2.3. EcN‐fyuA↑ Concentrates ⁶⁷Cu‐YbT in Solid Tumors to Elicit Anti‐Tumor Effects

After validating the specific interactions of EcN‐fyuA↑ with ⁶⁴Cu‐YbT, we switched to the high energy β ⁻‐emitting radiotherapeutic isotope, ⁶⁷Cu, and radiolabeled the ligand using the same technique used for ⁶⁴Cu. The yield and purity of ⁶⁷Cu‐YbT was >95% (Figure S2A, Supporting Information); the stability of the probe was >80% over a 24‐h period in mouse serum (Figure S2B, Supporting Information). For therapy studies, we chose two different murine subcutaneous tumor models—colon cancer MC38 and breast cancer 4T1 (Figure  3A). We injected saline, or EcN‐fyuA↑ intratumorally on day 0, followed by saline or two fractionated doses of ⁶⁷Cu‐YbT on day 1 and 4 retro‐orbitally. For both tumor models, the groups of mice that received a combination of EcN‐fyuA↑ with ⁶⁷Cu‐YbT survived significantly longer than those that received saline, bacteria only, or ⁶⁷Cu‐YbT only (Figure 3 and Figures S3 and S4, Supporting Information). In C57BL6/J model, the combination treatment extended the median survival of mice with highly aggressive MC38 tumors from 8 days in the control groups to 13 days, after initiation of treatment. On the other hand, the median survival in 4T1 tumor‐bearing mice, also a syngeneic highly aggressive model, improved from 11 days in the control groups to 18 days in the combination group (though not statistically significantly when compared to all control groups together). Importantly, EcN‐fyuA↑+⁶⁷Cu‐YbT halted tumor progression significantly more than ⁶⁷Cu‐YbT alone (MC38: p < 0.05). This conforms to our PET/CT imaging data and indicates that the presence of the engineered microbe in the tumor microenvironment allowed higher retention of therapeutic ⁶⁷Cu to attenuate tumor growth, in the colon cancer model.

Figure 3.

EcN traps ⁶⁷Cu‐YbT in solid tumors to elicit anti‐tumor effects. A) Schematic of the in vivo therapeutic plan. Tumor progression during the initial stages of therapy and survival curves of B) MC38 tumor‐bearing C57BL6/J mice (n = 4–8) and C) 4T1 tumor‐bearing BALB/cJ mice (n = 4) in the four treatment regimens. Data in tumor growth curves presented as mean ± s.d.; the final tumor volumes were analyzed by one‐way ANOVA and Dunnett's T3 multiple comparison test with Brown–Forsythe and Welch's correction, alpha = 0.05. Survival curves were analyzed by Kaplan–Meier with log‐rank (Mantel–Cox) test by comparing two groups at a time and presenting the p‐value at which ⁶⁷Cu‐YbT+EcN‐fyuA↑ is significantly different from the other three treatment strategies. ns = not significant, *p < 0.05, **p < 0.01.

2.4. EcN‐fyuA↑ and ⁶⁷Cu‐YbT Remodels Immune Landscape of Tumors without Systemic Toxicity

Next, we evaluated the immune cell profile of the MC38 tumor microenvironment 7 days post‐treatment. Based on preliminary screening there were no significant differences in total immune cell infiltrates (CD45⁺) (Figure  4A), though the global T cell (Figure 4B) tumor infiltration appeared to increase following ⁶⁷Cu‐YbT administration. Further analyses revealed that CD4⁺ T cells (Figure 4C), which typically activate macrophages and B cells to clear extracellular pathogens, decreased in tumors that contained EcN‐fyuA↑. This likely indicates that the immune cells do not recognize EcN‐fyuA↑ as a nonself‐antigen and organism. This might also explain the sustained presence of the bacteria in the tumor microenvironment (Figure 2C,D). On the other hand, the drop in CD4⁺ T cell population might be attributed to the concordant reduction in its subset, the regulatory T cell (T_reg) population (Figure 4D). Cytotoxic CD8⁺ T cell (Figure 4E) infiltration increased following systemic administration of ⁶⁷Cu‐YbT regardless of the presence of bacteria. This indicates that ⁶⁷Cu‐YbT, during its transit in the tumor microenvironment, elicited sufficient mutations in cancer cells for them to present neoantigens to activate CD8⁺ T cells. This resulted in a surge in cytotoxic T cell population to clear the malignant cells. Importantly, the simultaneous egress of immunosuppressive T_reg cells meant that the CD8⁺ T:T_reg cell ratio was significantly higher in tumors of mice that received a combination of EcN‐fyuA↑ and ⁶⁷Cu‐YbT compared to all other three groups combined (Figure 4F). This phenomenon has been observed in a recent TRT investigation as well,^{[ 19 ]} thus affirming EcN‐guided ⁶⁷Cu‐YbT delivery to solid tumors as a potential microbe‐based pretargeted immunomodulatory TRT platform.

Figure 4.

⁶⁷Cu‐YbT+EcN‐fyuA↑ induce changes in tumor immune infiltrate. Flow cytometry analyses of tumor immune cell infiltrates A) total immune cells (CD45⁺), B) total T cells, C) CD4⁺ T cells, D) T_regs (CD4⁺CD25⁺FOXP3⁺) cells, E) CD8⁺ T cells as a percent of total live cells, and F) CD8⁺ T:T_reg ratio 7 days after ⁶⁷Cu‐YbT administration in MC38 tumor‐bearing mice with or without probiotic administration. All data presented as mean ± s.d. (n = 3); data in (F) analyzed by one‐way ANOVA and Dunnett's T3 multiple comparison test with Brown–Forsythe and Welch's correction, alpha = 0.05. *p < 0.05, **p < 0.01, ***p < 0.001.

There are several reports that show that EcN has good preclinical biocompatibility.^{[ 13} , ^{14 ]} Additionally, we have shown (Figure 2D) that our engineered bacteria, which is administered intratumorally, remain primarily confined in the solid tumors with no evidence of localizing in metabolic organs like the liver or kidneys. However, ⁶⁷Cu‐YbT, which is administered systemically is likely to be cleared by liver and kidneys to some extent as revealed by our PET imaging.^{[ 15 ]} Therefore, it was pertinent to evaluate the safety of the radiopharmaceutical. In order to determine whether ⁶⁷Cu‐YbT negatively alters hepatic and renal function in mice, we analyzed serum samples from mice treated with and without ⁶⁴Cu‐YbT, using standard toxicity screening. Hepatic enzyme analyses did not reveal any significant differences in alanine transaminase, alkaline phosphatase, or aspartate transaminase between the two groups of mice (Figure  5A). Bilirubin and total protein concentrations did not vary between the two treatment groups either. Since liver damage is typically associated with elevated levels of metabolic enzymes and total protein concentrations in the serum, we can infer that ⁶⁷Cu‐YbT did not result in off‐target liver toxicity in the mice. Likewise, with renal function we observed no changes in creatinine concentrations and blood urea nitrogen levels from normal values (Figure 5B). Finally, to assess the overall wellbeing of our mice, we monitored their weights every 2 days. While the general trend was a slight increase in body weight, which we attributed to progressive tumor growth, the differences were not statistically significant (Figure 5C).

Figure 5.

⁶⁷Cu‐YbT exhibits good biosafety profile. A) Liver and B) kidney function analyses of serum obtained from ⁶⁷Cu‐YbT treated and untreated C57BL6/J mice. C) Weight of mice monitored every 2 days following administration of ⁶⁷Cu‐YbT. Data presented as mean ± s.d. (n = 3).

3. Discussion

Prior studies using engineered EcN have relied on synthesizing and secreting therapeutic proteins in the tumor microenvironment, in the form of nanobodies, peptides, and enzymes. Our approach allows the use of engineered bacteria essentially as an “adapter” that can be incorporated in tumors, which carry the necessary receptors to attract a therapeutic radioisotope (⁶⁷Cu) with high degree of selectivity and precision. In this proof‐of‐concept study, we demonstrated that instead of secreting proteins, the engineered bacteria could be utilized as a pretargeted platform to attract systemically administered therapeutic radionuclides into the tumor microenvironment that is devoid of targetable biomarkers.

Due to tumor heterogeneity, the biomarker profile of tumors is often indeterminate making it difficult to devise targeted strategies for imaging and therapy. Therefore, there is a need to investigate and design novel targeted systems that can be used to ablate multiple tumor types irrespective of their antigen and receptor phenotype. We engineered E. coli Nissle to overexpress the metal uptake receptor, FyuA, on the bacterial cell‐surface. We demonstrated that the engineered bacterium binds to and internalizes its cognate probes ^64/67Cu‐YbT with high specificity. The probes are highly switchable as the ligand (YbT) can chelate ⁶⁴Cu (imaging) and ⁶⁷Cu (therapy) radioisotopes. We showed that our engineered bacteria are not only able to populate the tumors, but also thrive and maintain sustainable growth kinetics in tumors, which can be tracked longitudinally using the ⁶⁴Cu‐labeled probe. In addition, we revealed that FyuA could serve as a foreign epitope that can then be superimposed on the cancer cells in the tumor matrix, to attract and bind ⁶⁷Cu‐labeled YbT to initiate therapy. Since the surface proteins are genetically encoded by the bacteria, the density and number of these artificial epitopes would be amplified as a function of time and size of tumors. Our results also indicated that EcN‐fyuA↑ sequestered ⁶⁷Cu‐YbT from the systemic circulation and induced anticancer effects by modulating the local immune cell populations, which are consistent with previous studies that illustrated the remodeling of the tumor microenvironment by intratumoral EcN.^{[ 13} , ^{20 ]} We observed that CD8⁺ T:T_reg cell ratio was significantly higher in the tumor microenvironment following ⁶⁷Cu‐YbT+EcN‐fyuA↑ therapy, though further studies are warranted. For instance, generation of memory T cells could be evaluated by analyzing local lymph node population following tumor engraftment (or a failure thereof) in a different location. Whether these memory T cells can elicit an abscopal effect to eradicate distant metastases would also be an intriguing investigation. Moreover, the ability of our radiomicrobial platform to act as an adjuvant to current immunotherapeutic approaches for targeting solid tumors is certainly worth further attention.

A potential limitation of our platform could be the non‐specific tumor accumulation of ⁶⁷Cu‐YbT alone. Since mammalian cells do not express the bacterial FyuA protein, we do not anticipate the radiopharmaceutical to remain in the tumor microenvironment long enough to generate damaging effects to the cancer cells, unless a unique cancer microbiota retains the probe via non‐FyuA mediated pathways. This raises the concern of potential and excess probe accumulation in off‐target organs, which we observed via PET imaging. Nonetheless, most of our serum test results were in the range similar to those obtained from naïve C57BL6 mice,^{[ 21} , ^{22 ]} which should alleviate concerns of potential radiotoxicity of this experimental therapeutic platform.

To our knowledge, this is the first time a bacterium has been used for TRT. In addition, we have also revealed for the first time that such a platform has immunomodulatory (increased anti‐tumor CD8⁺ T cell population, decreased anti‐inflammatory T_reg population) properties. Future studies will focus on comprehensive analysis of cancer cell death mechanisms and immunological pathways triggered by our platform, including establishment of immune memory to be effective in recurrent or relapsed tumors. This microbial receptor‐mediated radiopharmaceutical delivery to tumors could potentially be extended to mammalian receptors for which there are FDA‐approved radiopharmaceuticals. Comparing and contrasting the various ligand–receptor interactions using bacteria as an in situ living template in diverse solid tumors would help us identify the types and subtypes of cancer that would be amenable to this novel form of TRT.

4. Experimental Section

Chemicals

All reagents were purchased from commercial sources as analytical grade and used without further purification. YbT was purchased from EMC Microcollections GmbH (Tuebingen, Germany). ⁶⁴Cu was obtained from Mallinckrodt Institute of Radiology, Washington University School of Medicine, while ⁶⁷Cu from Idaho Accelerator Center, Idaho State University.

Bacterial Strain and Plasmids

E. coli Nissle (Mutaflor) was used in this study. Plasmids were constructed using standard restriction enzyme‐mediated cloning methods. To construct pZVS1, the fyuA gBlock (Figure S1A, Supporting Information) flanked by appropriate restriction sites was purchased from Integrated DNA Technologies. This sequence was cloned into pSF‐OXB20 (Millipore Sigma) using EagI and HindIII sites. Following plasmid construction at each stage, the sequences were verified via Sanger sequencing before EcN was transformed. FyuA KO was performed in EcN using the standard red recombinase method as previously described.^{[ 15} , ^{23 ]} For BLI, EcN‐fyuA↑ was transformed with pGEN‐luxCDABE (Addgene plasmid # 44918, a gift from Harry Mobley) for constitutive expression of luciferase and its substrates. EcN transformants were grown on ampicillin and kanamycin supplemented LB agar or in broth as required.

Immunoblotting

To qualitatively confirm plasmid‐based expression of FyuA, samples of overnight cultures were grown at 1:100 in fresh antibiotic‐supplemented media until the OD₆₀₀ reached 0.8. Subsequently, the bacteria were centrifuged, the spent media was discarded, and the pellets were washed with PBS (×3). The pellets were lysed with B‐PER Bacterial Protein Extraction Reagent (Thermo Fisher Scientific) supplemented with proteases and phosphatases to extract the target proteins for western blot analyses. After quantifying protein concentration via BCA assay, a 1:1 mixture of lysed fraction and Laemmli sample buffer was boiled at 95 °C for 10 min. The proteins were separated by SDS–PAGE and transferred to 0.45‐µm Amersham nitrocellulose membranes (GE Healthcare). Membranes were then blocked in 5% milk/TBST for 1 h at room temperature and incubated with Direct‐Blot HRP anti‐DYKDDDDK (FLAG) tag primary antibody (BioLegend) at 4 °C overnight. Immunoblots were developed using SuperSignal enhanced chemiluminescence (Thermo Fisher Scientific) and imaged via C‐DiGit Blot Scanner (LI‐COR Biosciences).

Radiolabeling

10 µL of ⁶⁴Cu or ⁶⁷Cu was mixed with 10 µg (10 µL) of YbT in 0.1 m ammonium acetate (pH 7) to bring the reaction volume to 100 µL. This mixture was vortexed for 10–15 s before incubating in a thermomixer with 800 rpm agitation at 37 °C for 1 h. Radiochemical purity and stability studies were determined using radio‐HPLC (Agilent 1260 Infinity II Quaternary System with a Flow‐RAM radio‐HPLC detector) on a C18 column (Agilent Poroshell 120 EC‐C18 column, 3 × 50 mm, 4 µm) as described before.^{[ 15 ]}

Functional Characterization of EcN‐fyuA↑

10⁴–10⁹ cfu of EcNΔfyuA, EcN WT, or EcN‐fyuA↑ were incubated with ⁶⁴Cu‐YbT (0.04–0.06 Mbq) supplemented LB for 2 h. Subsequently, the samples were centrifuged, and the pellets washed with PBS (×3). Cell‐associated ⁶⁴Cu levels were measured from pelleted bacteria using a gamma counter. Experiments were repeated three times from independent bacterial cultures.

Mammalian Cell Culture

Murine breast cancer 4T1 cells (ATCC CRL‐2539) and MC38 colon cancer cells (Kerafast#ENH204‐FP) were cultured in RPMI 1640 and DMEM respectively containing 10% fetal bovine serum and 5% penicillin–streptomycin solution. All media were purchased from Gibco, Thermo Fisher Scientific. The cells were maintained at 37 °C with 5% CO₂ in air and sub‐cultured twice weekly.

Animal Models

4–8‐week‐old female BALB/cJ and male C57BL6/J mice (Jackson Laboratory) were used in this study. For tumor induction, 1.5 × 10⁶ cancer cells (100 µL saline) were injected subcutaneously in the shaved right flank of anesthetized (2% isoflurane) mice. Tumor volume was calculated by measuring the length and width of each tumor every 2–3 days using calipers, where volume = 0.5 × length × width².^{[ 2 ]} The tumor volume of each mouse was monitored every 2–3 days until the animals reached one of the defined end‐points: (i) the longest dimension of tumor >1.2 cm, (ii) tumor became necrotic or ulcerated, (iii) tumor started to hamper movement of the mouse, or (iv) mouse lost >10% of its body weight.

Bacterial strains were grown overnight in LB media, which contained the appropriate antibiotics. A 1:100 dilution into growth media with antibiotics was started the day of injection and grown to an OD₆₀₀ of ≈ 0.8. Bacteria were centrifuged and washed with PBS (×3) before 5 × 10⁶ cfu (100 µL saline) were intratumorally injected in mice.

A single dose of 3.7–5.55 Mbq of ⁶⁴Cu‐YbT (for imaging) or two fractionated doses of 9.25 Mbq of ⁶⁷Cu‐YbT (for therapy) was retro‐orbitally administered in each mouse.

For serum analysis, terminal intra‐cardiac blood collection was performed. Serum samples were prepared and shipped to IDEXX BioAnalytics for analyses.

All animal experiments were conducted by following a protocol approved by the University of Cincinnati Biosafety, Radiation Safety, and Animal Care and Use Committees (protocol#: 20‐05‐16‐01).

Imaging and Ex Vivo Biodistribution Studies

Small‐animal PET scan was performed 24 h post injection of ⁶⁴Cu probes on a µPET scanner (Siemens Inveon). Animals were placed in the supine position on the imaging gantry with continued warming for the duration of the scan. CT scan (80 kVp, 500 µA, at 120 projections) was acquired for anatomical reference overlay with PET image for a 15‐min acquisition with real‐time reconstruction. PET images were acquired over an additional 15 min and spatial resolution in the entire field of view was determined by ordered subset expectation maximization in 2D. Histogramming and reconstruction were applied using Siemens Inveon software. Post‐processing was carried out with Inveon Research Workplace and general analysis was used for contouring volume of interest (VOI). These VOI values were considered active infection volumes and used for further analyses. Bioluminescence images were acquired for 5 min using an IVIS Imaging System for quantification of radiance (total flux, photons per second, p s⁻¹) of the bioluminescent signals from the regions of interest. After the imaging studies, the mice were euthanized via carbon dioxide inhalation and cervical dislocation. Organs and tissues of interest were removed and weighed. Residual radioactivity in the samples was measured with a gamma counter and results expressed as percentage of injected dose per gram of organ (% ID/g). Tumors with bacteria were homogenized and serially diluted in fresh LB media, before plating each dilution on antibiotic supplemented LB agar plates to enumerate the bacteria.

Quantitative PCR with Reverse Transcription

Total DNA extraction from organs and tumors was performed with DNeasy Blood & Tissue Kit (QIAGEN). pGEN‐luxCDABE‐specific primers (forward: ATGAAATTTGGAAACTTTTTGCTTACATAC and reverse: GGGGTTTACTTTTACCTTATGGAACT) and Luna Universal qPCR Master Mix (NEB) were used to perform qRT‐PCR on QuantStudio 3 (Thermo Fisher Scientific) in a 96‐well format. Standards were analyzed based on pure bacterial plasmids and used to determine pGEN‐luxCDABE concentration from animal tissues.

Flow Cytometry

Harvested tumors were minced into 1 mm pieces with a razor blade and digested in HBSS containing 2 mg/ml Collagenase IV (Gibco) and 20 µg mL⁻¹ DNase I (Sigma‐Aldrich) for 45 min at 37 °C under agitation. Tumor suspensions were filtered with 70 µm strainer and debris were removed using the Debris removal solution (Miltenyi) according to the manufacturer's protocol. Five million cells per tumor were used for immunophenotyping by flow cytometry. The single cell suspensions were first labeled with the Fixable Viability Dye eFluor506 (eBioscience) in order to separate live from dead cells during analyses. Fc receptors were blocked by incubation with the mouse FC blocker solution (Miltenyi) for 10 min at 4 °C. Samples were incubated 20 min at 4 °C with the following antibodies: anti‐CD45‐AlexaFluor488 (eBioscience, 30‐F11), Anti‐CD3‐PE‐Cy5 (eBioscience, 145‐2C11), anti‐CD4‐APC (eBioscience, RM4‐5), anti‐CD8a‐PerCP‐Cy5.5 (eBiosciences, 53‐6.7), anti‐CD25‐PE (eBioscience, PC61.5). Following two washes with PBS the samples were fixed and permeabilized using the Foxp3 staining kit (eBioscience) according to the manufacturer's protocol. Samples were incubated after permeabilization with an anti‐Foxp3‐PE‐Cy7 (eBioscience, FJK‐16s) for 30 min at 4 °C. Samples were acquired on a BD LSRFortessa 2 in the Cincinnati Children's Hospital Medical Center Flow Cytometry Core. Single stained samples were used to calculate the compensation parameters between the different fluorochrome using Diva software (BD Bioscience). Sample analyses were performed using FlowJo software (BD Bioscience).

Statistical Analyses

Pre‐processing of data (e.g., transformation, normalization, and evaluation of outliers) was not performed for any of the experiments unless otherwise stated in the figure legends. All numerical data are presented as mean ± s.d. The sample size and statistical methods used for each experiment are stated in detail in the figure legends. All data were analyzed using GraphPad Prism 9.0.0 software, unless noted otherwise.

Conflict of Interest

N.A.S. and N.K. have filed a provisional patent application with the US Patent and Trademark Office related to this work.

Author Contributions

N.K. conceived the study. N.A.S. constructed the genetically encoded microbial systems and designed all in vitro and in vivo experiments. N.A.S. and A.J.V. performed the in vitro experiments. N.A.S., A.R.H., T.K., N.S.K., S.C.T., and T.M. carried out the in vivo studies. M.G.S performed flow cytometry analyses. N.A.S. and N.K. wrote and revised the manuscript with input from all authors.

Supporting information

Supporting Information

Siddiqui N. A., Ventrola A. J., Hartman A. R., Konare T., Kamble N. S., Thomas S. C., Madaan T., Kharofa J., Sertorio M. G., Kotagiri N., An Engineered Probiotic Platform for Cancer Epitope‐Independent Targeted Radionuclide Therapy of Solid Tumors. Adv. Healthcare Mater. 2023, 12, 2202870. 10.1002/adhm.202202870

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.