Recent Progress in the Molecular Imaging of Tumor-Treating Bacteria

Abstract

Bacterial cancer therapy (BCT) approaches have been extensively investigated because bacteria can show unique features of strong tropism for cancer, proliferation inside tumors, and antitumor immunity, while bacteria are also possible agents for drug delivery. Despite the rapidly increasing number of preclinical studies using BCT to overcome the limitations of conventional cancer treatments, very few BCT studies have advanced to clinical trials. In patients undergoing BCT, the precise localization and quantification of bacterial density in different body locations is important; however, most clinical trials have used subjective clinical signs and invasive sampling to confirm bacterial colonization. There is therefore a need to improve the visualization of bacterial densities using noninvasive and repetitive in vivo imaging techniques that can facilitate the clinical translation of BCT. In vivo optical imaging techniques using bioluminescence and fluorescence, which are extensively employed to image the therapeutic process of BCT in small animal research, are hard to apply to the human body because of their low penetrative power. Thus, new imaging techniques need to be developed for clinical trials. In this review, we provide an overview of the various in vivo bacteria-specific imaging techniques available for visualizing tumor-treating bacteria in BCT studies.

Introduction

The incidence of cancers and the mortality from them continue to increase worldwide, despite all the efforts made in developing anticancer drugs. According to GLOBOCAN data reporting on the global burden of cancer, there were an estimated 14.1 million new cases and 8.2 million deaths from cancer in 2012 [1], and 18.1 million new cases and 9.6 million deaths in 2018 [2]. Conventional anticancer drugs often encounter challenges such as low tumor specificity and high refractoriness in hypoxic tumor regions. As a novel anticancer strategy to overcome some of the challenges facing conventional therapies, bacterial cancer therapy (BCT) takes advantage of some unique features of facultative or obligate anaerobic bacteria, such as strong tumor tropism, tumor-specific proliferation in both hypoxic and oxic tumor regions, and antitumor immunity, as well as investigating bacteria as drug delivery agents. However, the results of recent human clinical trials of BCT are controversial; although the administration of certain bacterial strains was shown to be safe, their efficacy was not demonstrated [3–5].

In clinical trials, bacterial colonization has been confirmed by clinical signs of inflammation in the target tumor (pain, erythema, swelling), systemic signs of infection (fever, laboratory findings), blood culture, or tumor tissue sampling. However, clinical signs and invasive tissue sampling may not accurately represent bacterial colonization and proliferation at tumor sites, or unintended bacterial colonization of normal tissues, because the bacterial distribution may be heterogeneous both within tumors and in normal tissues. Thus, there is a need to improve the visualization process for BCT to permit noninvasive and repetitive in vivo imaging that can facilitate the clinical translation of BCT.

Initial BCT-specific imaging techniques used in vivo optical imaging with genetically engineered bacteria expressing a fluorescence or bioluminescence reporter gene [6–10]. In recent decades, diverse imaging technologies using more translatable imaging systems, such as positron emission tomography (PET) and magnetic resonance imaging (MRI), have been developed, and the tumor-treating bacterial strains available for BCT include Gram-positive (Listeria monocytogenes, Bifidobacterium spp., Clostridium spp.) and Gram-negative bacteria (Escherichia coli, Salmonella spp.). This overview provides comprehensive information on the in vivo bacteria-specific imaging techniques available for visualizing tumor-treating bacteria in BCT studies, as well as information on the bacterial strains available for BCT studies.

Fluorescence Imaging

Fluorescence imaging (FLI) was the first imaging-based assay used for tumor-treating bacteria. Fluorescent proteins such as green fluorescent protein (GFP) and red fluorescent protein (RFP) are transfected into tumor-treating bacteria [11]. Recently, fluorescently labeled ligands specifically targeting bacteria have been used for the diagnosis of infectious disease (Table 1). Such ligand-based probe strategies have the advantage of not requiring bacterial engineering. Fluorophores have different excitation and emission wavelengths, and recent advances in near-infrared (NIR) fluorophores have allowed the detection of fluorescence signals from deeper tissue depths, which raises the possibility for clinical translation. The various fluorescent probes available include bacterial surface-targeted fluorophores, metabolism-based probes, enzyme-activated probes, fluorescent antibiotics, bacteriophages, and antibodies (Table 1).

Table 1.

Fluorescence imaging probes with the potential to image tumor-treating bacteria

Class of tracers	Probe ligand	Labeling compounds	Bacteria	References
Bacterial surface	Zinc(II) dipicolylamine (Zn-DPA)	NIR fluorophores (IR783, Cy5, streptavidin-coated CdSe/ZnS quantum dots, PSVue®794)	Gram-positive, Gram-negative	[13–16]
Bacterial surface	Concanavalin A	NIR fluorophore (IR750)	Gram-positive, Gram-negative	[17]
Bacterial surface	Mannose	Fluorescein	E. coli	[18]
Antibiotics	Vancomycin	IRDye800CW	Gram-positive	[19]
Antimicrobial peptides	Polymyxin B	7-nitrobenz-2-oxa-1,3-diazole (NBD)	Gram-negative	[20]
Antimicrobial peptides	Ubiquicidin (UBI29-41)	ICG02	Gram-positive, Gram-negative	[21, 22]
Bacteriophages	M13 bacteriophage	Alexa 750	Bacteria expressing F-pili	[24]
Metabolic substrate	Maltodextrin	IR786	Gram-positive, Gram-negative	[25]
Iron transporter	Siderophore	Cy5.5	Gram-negative	[27]
Enzyme-activated probes	Self-quenched NIR dye CytoCy5S	Reduced NIR dye CytoCy5S	Most Gram-positive and Gram-negative	[28]

NIR fluorophores linked to zinc(II) dipicolylamine (Zn-DPA) will bind to the negatively charged anionic surface of bacteria, while showing very little binding to the zwitterionic surface of healthy animal cells [12]. As a negative surface charge is a common feature of nearly all bacterial membranes, these probes should target most types of bacteria [12–16]. Concanavalin A is a plant-derived lectin that can be used as a bacterial-targeting ligand, having a high affinity to the surface mannose residues of both Gram-positive and Gram-negative bacteria [17]. Mannose-functionalized fluorescent polymers can be used to image Escherichia coli, as these bacteria use the surface mannose of host cells as a receptor to facilitate cell–cell adhesion [18].

Fluorescently labeled antibiotics and antimicrobial peptides either bind to or are taken up by specific bacteria [19–22]. Vancomycin binds specifically to the D-Ala-D-Ala that is a terminal moiety of the peptidoglycans of the cell wall of Gram-positive bacteria [19]. The antimicrobial peptide polymyxin binds selectively to the lipid A of the lipopolysaccharide on the outer membrane of Gram-negative bacteria [20]. UBI29-41, a cationic antimicrobial peptide fragment, targets the anionic surfaces of Gram-positive and Gram-negative bacterial cells by nonspecific electrostatic interactions [21, 22].

The M13 bacteriophage has a natural binding affinity to the F-pili of certain strains of E. coli [23]. Furthermore, the M13 bacteriophage can be used as a scaffold for an antibody-based probe against specific bacteria [24]. A Staphylococcus aureus-specific antibody-based probe successfully detected S. aureus infection with high binding affinity [24]. However, it should be considered that the large molecular size of antibody-based probes presents slow tissue diffusion and low blood clearance rates resulting in a low-contrast imaging.

Bacteria have specific metabolic pathways distinct from those of eukaryotic cells, and fluorophore-labeled metabolic substrates may be specifically incorporated by bacteria, thereby allowing their direct imaging. Metabolizable maltodextrin, which has a bacteria-specific transport pathway, can be labeled with fluorophores to image both Gram-positive and Gram-negative bacteria [25]. Siderophores, which are high-affinity iron-chelators able to solubilize and acquire iron, a growth-limiting nutrient, are biosynthesized and secreted by most bacteria and fungi, as well as by some plants [26]. Fluorescently labeled tetrapodal 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic amide (DOTAM), a siderophore mimic, can image Gram-positive and Gram-negative bacteria through the binding of iron and its translocation across the outer and inner cell membranes [27].

The enzyme-activated probe CytoCy5S is reduced to a fluorescent form by bacterial-specific endogenous nitroreductase enzyme when the probe encounters appropriate bacteria [28], with nitroreductase enzymes being present in a wide range of Gram-positive and Gram-negative bacteria. However, as nitroreductases are also expressed in eukaryotic cells, this technique may be subject to high background activity [29].

Bioluminescence Imaging

The bioluminescence imaging (BLI) approach requires genetic engineering of bacteria to incorporate a bioluminescence reporter gene such as luciferase. The luciferase reporter gene from bacterial insect pathogens (lux from Photorhabdus luminescens or Photobacterium leiognathi) is the mainstay of current in vivo bacterial optical imaging, although the luciferase reporter gene from sea pansy Renilla reniformis (rluc) has also been successfully expressed in Salmonella typhimurium [8]. The genes for PpyRE-TS and PpyGR-TS Photinus pyralis firefly luciferase mutants have been expressed in E. coli [30]. Although the luciferase reporter gene from the marine copepod Gaussia princeps (gluc) has been expressed in Mycobacterium smegmatis and used for in vivo imaging, it has not yet been expressed in tumor-treating bacteria. However, because of poor spatial resolution, a limited penetration depth, and low quantification accuracy due to loss and scattering of light within the body, it is difficult to perform such optical imaging in human bodies [31].

Photoacoustic Imaging

Photoacoustic imaging (PAI) uses a pulsed laser to excite a fluorophore, which then induces thermoelastic expansion and the generation of ultrasonic pressure waves that can be detected using ultrasound transducers [32]. PAI offers superior depth penetration (up to 7 cm) to optical imaging (FLI and BLI) and superior spatial resolution to other functional imaging modalities, such as PET and single photon emission computed tomography (SPECT) [33, 34]. Recent technological advances mean that the clinical application of this technique is to be expected. Cy7-1-maltotriose, which is a maltodextrin-based probe for photoacoustic and fluorescence imaging, was shown to accumulate in multiple Gram-positive and Gram-negative bacterial strains but not in mammalian cells [35]. E. coli expressing photoswitchable chromoproteins (e.g., sGPC2, BphP1) were injected into the subcutaneous layer of hairless mice and successfully imaged by PAI [36].

Nuclear Medicine Imaging

PET and SPECT might be strong candidates for overcoming the limitations of other imaging modalities such as optical imaging and PAI, as they show high sensitivity with an unlimited depth of penetration (Table 2).

Table 2.

Nuclear medicine imaging with the potential to image tumor-treating bacteria

Class of tracers	Radiopharmaceuticals	Bacteria	References
Bacterial surface	111In-DOTA-biotin/SA/biotin-Zn-DPA	Gram-positive, Gram-negative	[16]
	89Zr-labeled antibody specific for lipoteichoic acid (89Zr-SAC55)	Gram-positive	[76]
Antibiotics	99mTc-labeled fluoroquinolones	Gram-positive, Gram-negative	[44, 51, 55–57, 59–61]
	99mTc-labeled cephalosporins	Most Gram-positive and Gram-negative	[62–66]
	18F-fluoropropyl-trimethoprim (FPTMP)	Gram-positive, Gram-negative	[67]
Antimicrobial peptides	99mTc-Ubiquicidin (UBI), 68Ga-UBI, 111In-DTPA-Cy5-UBI29-41	Gram-positive, Gram-negative	[73, 75]
Nucleoside analog metabolism	125I-fialuridine (FIAU), 124I-FIAU, 18F-FEAU	E. coli, C. novyi-NT, S. aureus	[38–42]
Sugar metabolism	18F-fluoromaltose, 18F-maltohexaose, 18F-fluoromaltotriose	Gram-positive, Gram-negative	[77–79]
Sugar alcohol metabolism	18F-fluorodeoxysorbitol	Enterobacteriaceae family	[80]
Amino sugar metabolism	18F-fluoroacetamido-D-glucopyranose (FAG)	Gram-positive, Gram-negative	[82]
Folate metabolism (synthesis)	11C-para-aminobenzoic acid (PABA), 18F-PABA	Gram-positive, Gram-negative	[83, 84]
Amino acid metabolism	11C-D-methionine	Gram-positive, Gram-negative	[85]
	11C-D-alanine	Gram-positive, Gram-negative	[86]
Iron transporter	Siderophore	Gram-positive, Gram-negative	[27, 88]
Biotin	111In-biotin, 18F-labeled biotin derivatives	Gram-positive, Gram-negative	[90–93]

Herpes simplex virus 1 thymidine kinase (HSV1-TK) gene [37, 38] and bacterial endogenous TK gene [39] have been shown to be effective reporter genes for nuclear medicine imaging. These reporter genes have been addressed to radiolabeled nucleoside analogs for the noninvasive imaging of tumor-targeting bacteria. Several tumor-targeting bacteria were successfully imaged with these radiolabeled nucleoside analogs. Attenuated S. typhimurium expressing HSV1-TK was successfully imaged with ¹²⁴I-2’-fluoro-1-β-D-arabino-furanosyl-5-iodo-uracil (¹²⁴I-FIAU) PET. E. coli Nissle 1917 with endogenous TK was successfully imaged with ¹⁸F-2′-Fluoro-2′deoxy-1-β-D-arabinofuranosyl-5-ethyl-uracil (¹⁸F-FEAU) PET. ¹²⁴I-FIAU and ¹²⁵I-FIAU were also used to detect bacterial infection [40–42].

Bacteria-specific nuclear imaging for the detection of infection initially focused on radiolabeled antibiotics. Although various radiolabeled antibiotics have been studied for infections, considering the bacterial strains targeted by each of these antibiotics, radiotracers using fluoroquinolones [43–61], cephalosporins [62–66], and trimethoprim [67] could be utilized for BCT. However, such radiolabeled antibiotics typically demonstrate low specificity [49, 50, 53–55, 58, 59], which is not surprising given that the drugs are designed to kill or inhibit bacteria at the lowest possible concentration. Recently, antimicrobial peptides, which are positively charged host defense peptides present in virtually all organisms, have emerged as novel anti-infective therapeutics. Antimicrobial peptides display remarkable structural diversity, functional diversity, and immune-modulating properties, making them interesting novel therapeutic agents against antibiotic-resistant bacteria. Interactions between cationic antimicrobial peptides and mammalian cell membranes are weak because of the zwitterionic phospholipids of mammalian cell membranes [68]. The human antimicrobial peptide ubiquicidin (UBI), labeled with ^99mTc, ¹¹¹In, and ⁶⁸Ga, showed higher specificity than radiolabeled antibiotics in the detection of Gram-positive and Gram-negative bacterial infections [69–75].

The live bacterial cell surface has been successfully targeted by nuclear imaging probes and optical imaging probes. Zn-DPA radiolabeled with ¹¹¹In via streptavidin (¹¹¹In-DOTA-biotin/SA/biotin-Zn-DPA) targeted bacterial infection in a similar manner to fluorescent Zn-DPA [16]. Zn-DPA, an exogenous synthetic ligand, exhibits selective affinity for anionic cell membranes of Gram-positive and Gram-negative bacteria. Lipoteichoic acid (LTA), the surface molecule of the cell wall of Gram-positive bacteria, was successfully targeted by an ⁸⁹Zr-labeled monoclonal antibody specific for LTA (⁸⁹Zr-SAC55), allowing visualization of Gram-positive bacterial infection. [76]

Other approaches have sought to develop novel imaging tracers utilizing metabolic pathways that are selectively expressed by bacteria but not by mammalian cells. Bacteria-specific sugar metabolism-based PET tracers have been developed to identify infections caused by specific classes of bacteria. The maltose- and maltodextrin-based PET tracers ¹⁸F-fluoromaltose [77], ¹⁸F-maltohexaose [78], and ¹⁸F-maltotriose [79] accumulate in multiple strains of Gram-positive and Gram-negative bacteria. These probes target the maltose-maltodextrin transporter that is expressed in Gram-positive and Gram-negative bacteria. ¹⁸F-fluoromaltotriose showed an improved pharmacokinetic profile in comparison with other probes for the maltose-maltodextrin transporter [79]. Sorbitol, a sugar alcohol, is a metabolic substrate for Gram-negative Enterobacteriaceae such as E. coli and Salmonella. ¹⁸F-fluorodeoxysorbitol (FDS) has the advantage that it is an easy and quick synthetic strategy involving the simple reduction of ¹⁸F-fluorodeoxyglucose (FDG); this occurs in 30 min and gives only a low background signal in comparison with sugar metabolism-based PET tracers [80, 81]. ¹⁸F-fluoroacetamido-D-glucopyranose (FAG), a radiolabeled amino sugar, was investigated as a PET tracer for the broad-spectrum imaging of bacteria. N-acetylglucosamine forms the peptidoglycan layer of the cell wall in both Gram-positive and Gram-negative bacteria. ¹⁸F-FAG, a radiolabeled N-acetylglucosamine derivative, was incorporated into the cell surface structures of bacteria and used to distinguish bacterial infection from nonbacterial inflammation [82]. Unlike mammals, bacteria produce their own folate, which is a crucial component of various metabolic pathways. Para-aminobenzoic acid (PABA) is the key precursor to folate, and ¹¹C or ¹⁸F labeled PABA showed a broad sensitivity for Gram-positive and Gram-negative bacteria [83, 84]. Radiolabeled D-amino acid can label the peptidoglycan layer of the bacterial wall; D-amino acids are incorporated into and protect the bacterial wall but are not metabolic substrates in mammals. D-[methyl-¹¹C]methionine [85] and D-[3-¹¹C]alanine [86] were shown to accumulate in living bacteria.

After ⁶⁸Ga-labeled siderophores were recently established for imaging of the fungal pathogen Aspergillus fumigatus [87], SPECT imaging with ⁶⁷Ga-labeled deferoxamine-type siderophores showed in vivo localization of S. aureus infection [88]. Further PET imaging with ⁶⁸Ga-labeled pyoverdine, which is a siderophore produced by Pseudomonas aeruginosa, showed in vivo localization of P. aeruginosa but not the location of E. coli [89].

Streptavidin and avidin are tetrameric biotin-binding proteins that have been purified from Streptomyces avidinii and egg whites, respectively. These proteins accumulate at infected and inflamed sites, probably as a result of increased vascular permeability. Biotin has extraordinarily high affinity for streptavidin and avidin. ¹¹¹In-biotin and ¹⁸F-labeled biotin derivative scans after intravenous administration of streptavidin or avidin showed high sensitivity and specificity for the detection of Gram-positive and Gram-negative infection [90–93].

Magnetic Resonance Imaging

MRI has the advantages of a lack of ionizing radiation burden, a high spatial resolution, and unlimited tissue penetration. It has been shown that MRI can be used for imaging a few species of tumor targeting bacteria. Magnetite-producing bacteria (Magnetospirillum magneticum AMB-1) [94], Clostridium novyi-NT spores labeled with iron-oxide nanoclusters [95], and E. coli Nissle 1917 expressing bacterial ferritin-like (iron storage) proteins [96] were detected using MRI, but MRI-based bacteria-specific imaging is still in its infancy.

Conclusions and Future Perspectives

The successful clinical translation of BCT requires novel therapeutic monitoring methods to replace invasive tissue sampling and subjective physical examinations. Bacteria-specific molecular imaging using various imaging techniques could provide quantitative and repeated information on the anatomical location of bacterial colonization and bacterial densities during BCT, facilitating research into strategies for the clinical translation of BCT to human patients. As a whole-body imaging modality with high sensitivity and unlimited tissue penetration, PET is promising for the visualization of BCT in in-human clinical trials. There have been significant efforts to develop PET radiotracers specifically targeting bacteria that will not accumulate in mammalian cells, particularly for the detection of infection. We may well be able to selectively apply these tracers to the imaging of tumor-treating bacteria.

Although PET is a promising imaging modality for the clinical translation of BCT, it is necessary to develop novel tracers with better specificity for the bacterial species used in BCT. Other than the reporter gene imaging approach, most radiotracers target a wide range of bacteria. As such, a generic imaging approach may mean that it is hard to discriminate between therapeutically administered bacteria, normal microbiota, and pathogenic bacteria causing infection, and further research on imaging of the specific bacterial species used for BCT is needed.

Author Contribution

S.R.K and J.J.M. collected data and wrote the manuscript.

Funding

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2016R1D1A1A02937289), and by National Research Foundation of Korea (NRF) grants funded by the Ministry of Science and ICT (NRF-2017R1A2B3012157, No. 2018R1A5A2024181 and NRF-2018M3A9H3024850).

Data Availability

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Declarations

Ethics Approval and Consent to Participate

For this type of study formal consent is not required

Consent for Publication

Not applicable

Conflict of interest

Sae-Ryung Kang and Jung-Joon Min declare that they have no conflict of interest.