key: cord-102809-06izdji8 authors: Siddiqui, Nabil A.; Houson, Hailey A.; Thomas, Shindu C.; Blanco, Jose R.; O’Donnell, Robert E.; Hassett, Daniel J.; Lapi, Suzanne E.; Kotagiri, Nalinikanth title: Radiolabelled Bacterial Metallophores as Targeted PET Imaging Contrast Agents for Accurate Identification of Bacteria and Outer Membrane Vesicles in vivo date: 2020-08-06 journal: bioRxiv DOI: 10.1101/2020.08.06.240119 sha: doc_id: 102809 cord_uid: 06izdji8 Modern technologies such as 16s DNA sequencing capable of identifying microbes and provide taxonomic resolution at species and strain-specific levels is destined to be transformative1. Likewise, there is an emerging need to accurately identify both infectious and non-infectious microbes non-invasively in the body at the genus and species level to guide diagnosis and treatment strategies. Here, we report development of radiometal-labelled bacterial chelators, knowns as metallophores that allow non-invasive and selective imaging of bacteria and bacterial products in vivo. We show that these novel contrast agents are able to identify E. coli with strain level specificity and other bacteria, such as K. pneumoniae, based on expression of distinct cognate transporters on the bacterial surface. The probe is also capable of tracking probiotic, engineered bacteria and bacterial products, outer membrane vesicles (OMVs), in unique niches such as tumours. Moreover, we report that this novel targeted imaging approach has impactful applicability in monitoring antibiotic treatment outcomes in patients with pulmonary infections, thereby providing the ability to optimize individualized therapeutic approaches. Compared to traditional techniques used to manufacture probes, this strategy simplifies the process considerably by combining the function of metal attachment and cell recognition into a single molecule. Thus, we anticipate that these probes will be widely used in both clinical and investigative settings in living systems for non-invasive imaging of infectious and non-infectious organisms. the body, and distinguish pathogenic from commensal species to minimize indiscriminate use of broad-spectrum antibiotics thereby adopting a more focused treatment strategy targeting specific pathogens. An application of equal importance is the precise mapping of resident microbiota, tracking of engineered bacteria and their products such as virulence factorharbouring outer membrane vesicles (OMVs) for therapeutic applications. The role of microbiota in human disease is increasingly being recognized to play a major role in aetiology and treatment modifications 3 . The emergence of next-generation sequencing techniques has allowed extensive characterization of the microbiome in various pathologic conditions, from neurologic and endocrine, to cancer [4] [5] [6] . However, there is an urgent need for non-invasive tools to detect and identify bacteria using endogenous markers. Bacteria are also commonly used for desirable, biotechnological applications due to the ease of genetic manipulation and heterologous protein expression. Several broad categories of bacteria-based approaches have been explored in preclinical research focusing on a wide range of diseases such as cancer, diabetes, obesity, hypertension, infection and inflammatory dysfunction [7] [8] [9] [10] . However, functional stability is important for biomedical applications of engineered bacteria. Development of convenient and realistic in vivo testing environments will be important for achieving improved accuracy. Such monitoring would provide information regarding presence, location, quantity, proliferation, survival and status of the current microbiome, engineered bacteria and OMVs. Numerous molecular imaging techniques such as ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), single-photon emission CT (SPECT), and positron emission tomography (PET) have been developed for preclinical and clinical research in the last three decades. However, current clinical probes cannot reliably distinguish between bacteria from mammalian cells in vivo 11 . Thus, a comprehensive understanding of bacterial physiology and genetics is required to develop probes for targeted imaging. As bacteria are evolutionarily and phylogenetically distinct from mammalian cells, fundamental differences in metabolism and cellular structures can be leveraged to develop bacteria-specific imaging agents. Some of the recent approaches that have focused on targeting metabolic pathways and proteins that are only present in bacteria include -11 C-para-aminobenzoic acid (PABA) 12 and 18 F-2-PABA 13 , which are involved exclusively in the bacterial folate pathway; 18 Fmaltohexaose 14 and 6-18 F-fluoromaltotriose 15 , which specifically target the maltodextrin transporter in bacteria; and 18 F-fluorodeoxysorbitol (FDS), a synthetic analogue of 18 F-FDG that selectively localizes in Gram-negative Enterobacteriaceae 16 . However, these radiotracers do not have the ability to distinguish one genus/species/strain from another. Aside from requiring high specificity and selectivity to the target bacterial population, labelling techniques for generating the reporter probes should be simple, rapid and inexpensive to qualify as an ideal imaging agent. Current techniques employ a multi-step process to label metabolites and ligands, which can lead to unnecessary delays and avoidable expenses considering the short decay half-lives (t1/2) of some radionuclides. Our study takes advantage of unique natural molecules that serve as both metal chelators as well as ligands, thus minimizing and simplifying the labelling protocol to just one step. Metal transport is a distinct pathway in bacteria that can be exploited to develop specific probes. Bacteria have developed sophisticated mechanisms for metal acquisition and transport to maintain metal homeostasis within a specified microenvironment. Entire pathways for metal acquisition, comprised of -(i) de novo synthesis of metal-chelating "metallophores, Mtps"; and (ii) dedicated membrane transporters that selectively bind the Mtpshave evolved to precisely regulate this process 17 . Mtps are small peptide-like molecules that possess a high affinity for transition metals 18 . The central role that Mtps play as chemical ligands in shuttling metals and the unique biology facilitating this transport into bacteria offers a unique platform that can be harnessed to develop highly versatile and specific contrast agents. Recently, two independent research teams underscored the roles of transport proteins, FyuA and CntA, in copper homeostasis of Escherichia coli and Staphylococcus aureus respectively. In the S. aureus study, a novel metallophore, staphylopine (StP) (Fig. 1a) was shown to chelate metals before the complex is selectively "captured" by the CntA domain and transported into the bacteria by the CntBCDF domains of this ABC transporter protein 19 . On the other hand, pathogenic E. coli UTI89 (urinary tract infection isolate) uses the metallophore yersiniabactin (YbT) (Fig. 1b) to sequester Cu (II) from the extracellular environment inside the bacteria 20 . Metal bound YbT (Fig. 1c ) is first selectively "recognized" by its cognate outer membrane protein receptor, FyuA (ferric yersiniabactin uptake A), before the inner membrane ATP-binding cassette transporters, YbTPQ allow cytoplasmic entry. We surmised that these bacterial metal transporters can be targeted by clinically relevant radioisotope labelled metallophores (Fig. 1d) . Since PET is considered the most sensitive among SPECT, PET, MRI, US and CT imaging technologies 11 , we focused on developing targeted PET imaging contrast agents. To optimize the one-step radiolabelling process, an accurate selection of radiometals, Mtps and buffering conditions is crucial for an efficient metal-metallophore complexation. Consequently, we performed in vitro radiolabelling of StP and YbT in various buffered and temperaturecontrolled conditions with four different transition metals -55 Co, 64 Cu, 68 Ga and 89 Zr (Fig. 1e) . We observed that YbT had the highest complexation with both 64 Cu and 89 Zr. Since StP manifested moderately high affinity for 89 Zr, we proceeded to investigate the stability of 64 Cu and 89 Zr labelled StP and YbT. When we incubated 89 Zr-YbT in saline (0.9% NaCl) and in diethylenetriamine pentaacetate (DTPA, a metal-chelator) supplemented saline, 89 Zr remained bound to YbT in both media for 24 hrs (Fig. 1f) , but the radiometal dissociated significantly from StP within the first hour in the same conditions (Fig. 1g) in vitro. We then assessed the stability of YbT probes in vivo and observed that both 64 Cu (Fig. 1h ) and 89 Zr (Fig. 1i) labelled YbT probes accumulated more significantly in the liver than any other body location, a feature that is likely attributable to the lipophilic nature of YbT. Since 64 Cu and 89 Zr-labeled YbT proved to have the highest stability amongst all the probes investigated thus far, we wanted to examine whether the YbT probes could selectively differentiate its target, E. coli UTI89 (Gram-negative), from representative Gram-negative and Gram-positive bacteria that lack the FyuA receptor in vivo. We executed our one-step radiolabelling protocol to demonstrate the ease, simplicity and rapidity of the radiolabelling process from radionuclide acquisition to animal injection (Extended Data Fig. 1) . We selected Pseudomonas aeruginosa (Gram-negative) and S. aureus (Gram-positive) as control bacteria, since both are widely studied and highly infectious nosocomial bacterial pathogens. Indeed, both P. aeruginosa and S. aureus lack the cognate FyuA transporters. Tail-Extended Data Figure 1 . Schematic representing the experimental approach to establish the suitability of radiolabelled metallophores for bacteria-specific imaging. vein injections of 64 Cu-YbT revealed significantly higher accumulation of the probe in UTI89infected muscles than in S. aureus (Fig. 2a) or P. aeruginosa (Fig. 2b) infected muscles, indicating FyuA-mediated E. coli species-specific imaging. However, it is likely that there could be a limited degree of crosstalk between strains, particularly within E. coli sp. that have the most genetic/pathogenic variants. To determine whether E. coli may possess alternative mechanisms to import 64 Cu-YbT other than the FyuA, we elected to compare the localization of 64 Cu-YbT in another highly pathogenic strain of E. coli, O157:H7 (enterohemorrhagic, EHEC), which does not encode for the FyuA receptor 21 . Intravenous (i.v.) administration of the 64 Cu-YbT revealed PET signals (Fig. 2c) , again exclusively in UTI89, thus confirming strain-specific imaging ability of the probe as well. Furthermore, when we administered 64 Cu-YbT in mice infected with live and heat-killed (90°C for 30 min) UTI89, we observed signals only from muscles harbouring live bacteria (Fig. 2d) . Ex vivo biodistribution revealed concordant results to all the PET/CT images ( Fig. 2e) . Any new diagnostic probe must be evaluated against a reliable "gold standard", so we compared the specificity of 64 Cu-YbT with the clinically available PET probe, Fluorine-18 labelled deoxyglucose ( 18 FDG). We noticed that the latter accumulated in almost equal concentrations in all the pathogens we have investigated to date ( Fig. 2g-i) . It is well-known that 18 F-FDG is taken up by cells that are involved in both septic and aseptic inflammation [22] [23] [24] . Hence, we believe that 18 F-FDG accumulated within immune cells that infiltrated the infected muscles. The lack of a substantial signal from the muscle injected with the non-pathogenic commensal E. coli K12 (Fig. 2h) indicates that the bacteria were not able to cause an inflammatory response. When we attempted to repeat our study with 89 Zr-YbT, our results showed that the probe failed to selectively identify UTI89 and accumulated mostly in the bones of diseased mice ( Fig. 3a and b) , though previous results indicated sufficient in vivo stability of 89 Zr-YbT in naïve mice ( Fig. 1f and i) . This led us to question whether the intact probe might have dissociated within and eventually effluxed from the bacteria. To confirm this postulate, we performed in vitro uptake and retention studies for both the probes. We grew an overnight culture of UTI89 in RPMI supplemented with 64 Cu-YbT or 89 Zr-YbT for 2 hrs to mimic the initial bacterial uptake of probes in a physiologically relevant environment. Several studies have shown that bacteria require 2 hrs to accumulate nearly all of the incubated probes 16, 20, 25 . Subsequently, we pelleted the bacteria and washed with PBS multiple times before we resuspended the pellets in fresh RPMI media for 2 and 24 hrs to investigate how the relative amount of each probe that was retained by the bacteria. Fig. 3c shows that while UTI89 was able to retain most of the 64 Cu-YbT over a period of 24 hrs, the bacteria lost a significant amount of 89 Zr-YbT over the same period. Besides E. coli UTI89, probiotic E. coli Nissle and pathogenic Klebsiella pneumoniae also transport metals using YbT via the FyuA receptor [26] [27] [28] . We next reasoned that if radiometallabelled YbT is truly selective for FyuA only, then our probes should be able to accumulate in Nissle and in K. pneumoniae as well. To test this hypothesis, we injected 64 Cu-YbT in mice that received intranasal administration of PBS (Fig. 4a) , P. aeruginosa (Fig. 4b) , E. coli Nissle (Fig. 4c ) and K. pneumoniae (Fig. 4d) , with the former two serving as negative controls. The goal was to selectively identify FyuA-expressing bacteria in mice with infected respiratory tracts. After we performed PET/CT imaging and ex vivo biodistribution (Fig. 4e) of various harvested tissues, we observed significantly higher signals from lungs and trachea of mice that received K. pneumoniae and Nissle compared to those that received P. aeruginosa and PBS. This illustrates the ability of our probe to trace and identify live bacteria in the pulmonary niche as well. Clinically, healthcare professionals would prefer to have a "real-time" assessment of therapeutic success or failure for the management of seriously ill patients. To demonstrate whether our probe can potentially be used in such circumstances, we injected 64 Cu-YbT in mice infected with E. coli UTI89 and K. pneumoniae to image changes in signal intensity following the administration of the antibiotic, ciprofloxacin. We generated UTI89 clones with a luciferase reporter to track bacterial growth and burden using bioluminescence imaging (BLI). BLI was accurately able to indicate decreases in bacterial burden in response to antibiotic treatment. We used 64 Cu-YbT to co-register PET signals with BLI, and were able to show a decrease in PET signal corresponding to a proportional decrease in bacterial burden in mice that received two doses of ciprofloxacin (Fig. 5a ) compared to untreated mice (Fig. 5b) . When we harvested the thigh muscles from mice post-euthanasia, gamma counter analysis revealed significantly lower radioactive counts from tissues of mice that received ciprofloxacin compared to the control (Fig. 5c ). The specificity of these distinct bacterial metal transporters for their corresponding metal bound Mtps can be leveraged for in vivo visualization of therapeutic bacteria and bacterial virulence factor carrying outer membrane vesicles (OMVs) as well. Since FyuA-expressing E. coli species have been proven to naturally possess the metal transporter in their secreted OMVs 29 , we postulated that both Nissle and its secreted OMVs can be selectively tracked by radiolabelled YbT without the need for further genetic manipulation of the bacteria. We tested our theory by injecting Nissle i.v. in subcutaneously developed 4T1 tumour-bearing mice. Usually, greater than 99% of the administered bacteria are cleared from the animals, leaving only a small percentage to colonize the tumor 30, 31 . Hence, we allowed 3 days for the bacteria to localize and proliferate in the hypoxic core of the tumour before administering 64 Cu-YbT in the mice. The control mice did not receive Nissle (Fig. 6a) . PET/CT and ex vivo biodistribution analysis revealed strong signals in tumours of Nissle-administered mice only (Fig. 6b and c) . To achieve our next goal of imaging OMVs in tumours, which accumulate as a result of enhanced permeability and retention (EPR) effect in leaky solid tumours, we first incubated Nissle OMVs and bovine milk exosomes, as the negative control lacking FyuA transporters, with 64 Cu-YbT to compare the difference in probe uptake between the two types of nanoparticulates. Subsequent purification to eliminate unbound PET probe yielded activities of approximately 200 µCi/ml and 180 µCi/ml for radiolabelled OMVs and exosomes, respectively. After 4 hrs postadministration of the radiolabelled vesicles in our mouse models, we performed PET/CT imaging where we observed radioactive concentration in tumours of the mice injected with OMVs ( Fig. 6d) compared to those in exosome administered subjects (Fig. 6e) . We believe that the presence of FyuA on the lipid bilayer of the OMVs allowed selective entry and statistically significant retention of 64 Cu-YbT within the OMVs (Fig. 6f) . Since the exosomes lack the metal transporter, 64 Cu-YbT is merely bound to the outer regions of the nanostructure without being selectively incorporated inside, and eventually had dissociated in the bloodstream in vivo (Fig. 6g ). In this study, we performed a comprehensive analysis of multiple metal-Mtps and successfully identified YbT complexed to 64 Cu as a highly stable PET probe that can selectively target bacterial metal transport proteins. Though numerous bacteria-specific PET probes have been developed within the last decade, most of these involve complex and time-consuming reaction mechanisms that eventually yield in tracer synthesis 32, 33 . Furthermore, most tracers do not have high complexation (>90%) with the radioisotope, which necessitates an additional purification step before administration. These might prove as potential barriers to clinical translation as hospital/clinical staff members would prefer to be able to prepare the probes in a facile manner. In some of the initial preclinical and clinical investigations, radiolabelled antibiotics such as 99m Tc-ciprofloxacin seemed promising, not only for their ability to specifically kill (or disable) bacteria while being nontoxic to human cells, but also for the ease of probe preparation using manufactured kits 34 . However, in later investigations, these SPECT probes proved to not only accumulate in bacterial lesions, but also in sterile inflammatory sites 35 . Metallophore-based probes eliminate the need for sophisticated synthetic chemical reactions since these molecules are readily synthesized and secreted by bacteria, and can easily be obtained using simple purification strategies from culture supernatants 19, 20 (Extended Data Fig. 2 ). Moreover, we succeeded in optimizing the radiolabelling procedure to a simple and single step of mixing radiometals with these highly selective bioinorganic ligands. thus critical to ensure that 64 Cu-YbT was taken up by live and not non-specifically by dead bacteria. Firstly, we displayed how the heat-killed E. coli UTI89 never imported our probe. Next, we showed that live bacteria that initially took up the probe, were later neutralized by an antibiotic (ciprofloxacin) regimen and eventually cleared from the infected site. This information could be particularly important in a healthcare setting, where a false positive signal by dead bacteria could mislead physicians to over-prescribe antibiotics in patients, which might eventually lead to the development of multidrug-resistant (MDR) bacteria. There is precedent for using metallophores as nuclear contrast agents. 68 Ga coordinated to ferricrocin and desferrioxamine (DFO) has been used in the past to image fungal infections, including Aspergillus fumigatus 36, 37 . 89 Zr coordinated to DFO and coupled to various targeting vectors, including antibodies, peptides and nanoparticles, is a useful strategy for imaging tumor specific receptors. Recently, Pyoverdine-68 Ga probes were used to selectively image P. aeruginosa infections in animal models 38 . However, due to the short half-life of 68 Ga (t1/2 = 68 min) the ability of 68 Ga-pyoverdine to track bacteria longitudinally in vivo can be challenging. The 12.7-hours half-life of 64 Cu provides the flexibility to image at both shorter and longer (upto 48h) time scales that allows for optimal clearance of probe and obtaining images with high contrast. Logistically, the longer half-life also allows 64 Cu-radionuclides to be easily distributed for PET imaging studies at sites remote to the production facility with the loss of approximately one half-life 39 . 89 Zr has a significantly longer t1/2 of 3.3 d which means that some levels of radioactivity can remain in the body for up to a month. This is of particular concern with 89 Zrbased probes as preclinical studies often report dissociation of the metal from the tracer leading to free 89 Zr accumulation in the bones 40 . Although 89 Zr-YbT showed promise in our stability investigations in naïve mice, the bacterial presence in infection models destabilized the probe, which resulted in significant uptake in the bones and joints in subsequent studies. Indeed YbTPQ transporter complex has been shown to be essential for metal-YbT complex dissociation to yield metal-free YbT 20 . Mtps have generally been studied for their affinity for and import of metals such as cobalt, copper, iron, manganese, nickel and zinc, which are deemed "nutritional" by bacteria 19, 20, 41 . However, zirconium has no known nutritional value to bacteria. Hence it is highly likely that only 64 Cu was retained inside bacteria and yielded significant PET signals, whereas 89 Zr was effluxed after 24 hrs from bacteria resulting in loss of PET signal from infection site and subsequent uptake in the bones (Fig. 3d) . This is interesting because 89 Zr is known to be a residualizing radionuclide in mammalian cells, which means that upon internalization in cells it is trapped inside the cell 42 CoV-2, millions of patients globally have required ventilator-assisted breathing in intensive care units (ICUs). Hospitals in developing countries and make-shift ICUs even in developed countries run the risk of exposing patients to common pathogens such as K. pneumoniae, P. aeruginosa and methicillin-resistant S. aureus that are known to cause ventilator-associated pneumonia (VAP) 47 . This pathology can be exacerbated by carbapenemase producing K. pneumoniae (KPC) and often prove fatal to patients. As scientists begin to unravel the exact pathogenesis of SARS-CoV-2 and emerging respiratory illnesses, early diagnosis of secondary infections caused by bacteria in the respiratory tract would be critical in saving significant numbers of lives globally. We have also validated how 64 Cu-YbT can potentially be used to advance bacterial therapies using engineered bacteria. Advances in genetic engineering have enabled us to explore the potential of using bacteria, particularly for cancer therapy in the last two decades 31,48-51 . One of the most notable initial studies tested genetically engineered Salmonella for its anticancer activity in mouse models of subcutaneously implanted B16F10 melanoma tumors 51 . The results were so promising that a phase II clinical trial was conducted in patients with metastatic melanoma. However, the research ended at that stage because the engineered bacteria did not yield sufficient tumour-targeting efficacy in humans. Selective tracking could have allowed scientists to better map its biodistribution, and hence optimize the therapeutic potential of the engineered Salmonella. Whilst numerous optical imaging studies have been conducted, very few nuclear imaging studies have selectively located bacteria in tumors 30 . Optical imaging techniques based on bioluminescence and fluorescence have the inherent drawback of limited light penetration, unlike PET imaging which is depth independent. Recently, tumour localization of OMVs from E. coli was confirmed via photoacoustic imaging 48 . While the study described a novel strategy to specifically image the presence of OMVs in the body, it involved an additional step of vesicle design to make it amenable for optoacoustic imaging. Our study illustrates that 64 Cu-YbT can image bacteria and their products using their endogenous proteins as reporters, without the need for additional genetic modifications. Moreover, the probe can detect bacteria in unique niches, such as tumours, with minimal nonspecific uptake in tumour tissue, thus maximizing the signal to background ratio. Current imaging agents such as 18 FDG, 68 Ga-citrate lack specificity as they are unable to distinguish between infection and inflammation 52 . Importantly, imaging agents such as 18 FDG would not be suitable in this scenario since tumour cells would also uptake the tracer and it would be impossible to delineate signals from bacteria and tumour. In conclusion, we have demonstrated the successful use of bacterial metallophores as "dual-role" compounds -(i) as a chelator as well as a targeting ligand for imaging and (ii) tracking of pathogenic and commensal bacteria as well as OMVs using a single probe. We have shown the versatility of the probe in detecting bacteria in three different modelsmyositis, pulmonary infection and an intratumoural niche. Importantly, this technique facilitates precise labelling of "live" bacteria in vivo. The small probe size (~400-500 Da), when compared to conventional antibody or peptide-based PET probes, will potentially allow access through the lining of blood vessels to analyse extravascular structures with relative ease. We anticipate that this will open opportunities to explore the possibility of using the diverse array of natural metallophores, with over 500 examples known to date, as tailored contrast agents for imaging a wide range of wild-type and engineered bacteria. The same probes can be used across diagnostic platforms, for in vitro assays using highly sensitive liquid scintillation and gamma counters as well as in vivo PET/SPECT/MRI imaging using appropriate radionuclides and metals. For instance, Mtp complexation with radiometals such as 99m Tc and 111 In could facilitate imaging with SPECT scanners that are more ubiquitous and less expensive to operate; and complexation with Mn metal or 52 Mn radionuclide could facilitate bacterial imaging with MRI or PET/MRI scanners, respectively, to improve spatial resolution. In addition, combining metallophores with novel radiometals can provide insights into unique radiochemistry techniques. Such innovations will reveal new and invaluable information based on metallophoremetal, metallophore-bacteria, metallophore-host and bacteria-host interactions in living systems, which in turn will help design advanced tools and future therapeutic strategies. All reagents were purchased from commercial sources as analytical grade and used without A sequential differential centrifugation protocol was developed to isolate exosomes and OMVs. In brief, high molecular weight (HMW) proteins and fats were removed from raw bovine milk and the supernatant was centrifuged at 91000g for 4 hrs to pellet the exosomes. In a similar manner, an overnight culture of E. coli Nissle was centrifuged to remove the bacteria followed by concentrating the supernatant using Pierce™ Concentrators (30 kDa, Thermo Scientific™). The Overnight culture of UTI89 was diluted to an OD600 of 0.8 in RPMI supplemented with 64 Cu-YbT or 89 Zr-YbT (4 µCi each) and grown for 2 hrs while shaking. Bacteria were subsequently pelleted at 8,000 r.p.m for 2 mins and washed with 1× PBS (Sigma) multiple times. Cell-associated 64 Cu or 89 Zr levels were measured from pelleted bacteria using a gamma counter. The pellets were resuspended in fresh RPMI media for 2 and 24 hrs after which the centrifugation and washing steps were repeated as above. Experiments were repeated three times from independent bacterial cultures. Animal models 6 -8 weeks old BALB/cJ and athymic nude (NU/J) mice were used for all experiments. For infection studies, 5 x 10 6 cfu of bacteria were administered either intramuscularly or intranasally 6 hrs or 24 hrs respectively before probe administration. To monitor antibiotic treatment efficacy, ciprofloxacin at a dose of 10mg/kg was given to mice once every 12 hrs via oral gavage. This regimen was thought to provide a plasma peak level in the range that was above those required to achieve efficacy after an oral administration of 500 mg of ciprofloxacin in humans 54 . To image tumour-infiltrating bacteria, 1.5 x 10 6 4T1 cells were subcutaneously injected and allowed to develop into a noticeable tumour (size > 200 mm 3 ) on the right flank of the mice. This was followed by a single i.v. administration of E. coli Nissle (5 x 10 6 cfu) 3 days before probe PET probe administration. To track OMVs and exosomes in 4T1 tumours, radiolabelled nanovesicles were injected immediately after the tumour reached the appropriate size. All animal experiments were performed under anaesthesia (2% isoflurane) by following protocols approved by the University of Cincinnati Biosafety, Radiation Safety, and Animal Care and Use Committees. 100 -150 µCi of PET probes were injected i.v. to selectively image bacteria in both infection and tumour models. 10-20 µCi of radiolabelled vesicles were injected i.v. to study their tumour localization. Small-animal PET scan was performed 4 or 24 hrs post injection 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. A CT scan (80 kVp, 500 μA, at 120 projections) was acquired for anatomical reference overlay with PET images for a 15-min acquisition with real-time reconstruction. The PET images were acquired over 15 mins as well and spatial resolution in the entire field of view was determined by ordered subset expectation maximization in 2 dimensions. Histogramming and reconstruction were applied using Siemens Inveon software. Post-processing was carried out with Inveon Research Workplace and general analysis and 3D visualization 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 mins using an IVIS Imaging System for quantification of radiance (total flux, photons per second, p/s) 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). All statistical analyses were performed using the GraphPad Prism 7 software. Two-tailed unpaired student's t-test was performed to compare the means between two groups, whereas one-way ANOVA was used to compare the means among three or more groups. Values of p < 0.05 was considered statistically significant. The data that support the findings of this study are available from the corresponding author upon request. National Institute Of General Medical Sciences of the National Institutes of Health (R21GM137321) and National Institutes of performed all in vitro and in vivo bacterial experiments. H.A.H. performed radiometal-metallophore complexation and stability studies. S.C.T. prepared exosomes and OMVs analysed data and wrote the manuscript with input from all authors Evaluation of 16S rRNA gene sequencing for species and strainlevel microbiome analysis The antibiotic resistance crisis: part 2: management strategies and new agents. P & T : a peer-reviewed journal for formulary management Cancer and the microbiota How gut microbes talk to organs: The role of endocrine and nervous routes Microglial control of astrocytes in response to microbial metabolites Two-step enhanced cancer immunotherapy with engineered Salmonella typhimurium secreting heterologous flagellin Reversal of Diabetes in NOD Mice by Clinical-Grade Proinsulin and IL-10-Secreting Lactococcus lactis in Combination With Low-Dose Anti-CD3 Depends on the Induction of Foxp3-Positive T Cells A phase I trial with transgenic bacteria expressing interleukin-10 in Crohn's disease. Clinical gastroenterology and hepatology : the official clinical practice journal of the Prevention of vaginal SHIV transmission in macaques by a live recombinant Lactobacillus Molecular imaging of bacterial infections: Overcoming the barriers to clinical translation )C]Para-Aminobenzoic Acid: A Positron Emission Tomography Tracer Targeting Bacteria-Specific Metabolism Positron Emission Tomography Imaging with 2-[(18)F]F-p-Aminobenzoic Acid Detects Staphylococcus aureus Infections and Monitors Drug Response Investigation of 6-[¹⁸F]-fluoromaltose as a novel PET tracer for imaging bacterial infection PET imaging of bacterial infections with fluorine-18-labeled maltohexaose Imaging Enterobacteriaceae infection in vivo with 18F-fluorodeoxysorbitol positron emission tomography Structural biology of bacterial iron uptake Biosynthesis of a broad-spectrum nicotianamine-like metallophore in Staphylococcus aureusKlebsiella pneumoniae Yersiniabactin Promotes Respiratory Tract Infection through Evasion of Lipocalin 2 Probiotic bacteria reduce salmonella typhimurium intestinal colonization by competing for iron Comparative analysis of the uropathogenic Escherichia coli surface proteome by tandem massspectrometry of artificially induced outer membrane vesicles Escherichia coli Nissle 1917 facilitates tumor detection by positron emission tomography and optical imaging Tumor-specific colonization, tissue distribution, and gene induction by probiotic Escherichia coli Nissle 1917 in live mice Bacterial infection imaging with [(18)F]fluoropropyl-trimethoprim In vivo detection of Staphylococcus aureus endocarditis by targeting pathogen-specific prothrombin activation Comparison of 99mTc infecton imaging with radiolabelled white-cell imaging in the evaluation of bacterial infection Inability of 99mTc-ciprofloxacin scintigraphy to discriminate between septic and sterile osteoarticular diseases An iron-mimicking, Trojan horse-entering fungi--has the time come for molecular imaging of fungal infections? PLoS pathogens 11 Siderophore-mediated mechanism of gallium uptake demonstrated in the microorganism Ustilago sphaerogena Imaging of Pseudomonas aeruginosa infection with Ga-68 labelled pyoverdine for positron emission tomography Copper-64 radiopharmaceuticals for PET imaging of cancer: advances in preclinical and clinical research In vivo biodistribution and accumulation of 89Zr in mice Metal selectivity by the virulence-associated yersiniabactin metallophore system Long-lived positron emitters zirconium-89 and iodine-124 for scouting of therapeutic radioimmunoconjugates with PET Identification of an ABC transporter required for iron acquisition and virulence in Mycobacterium tuberculosis The ferrichrome uptake pathway in Pseudomonas aeruginosa involves an iron release mechanism with acylation of the siderophore and recycling of the modified desferrichrome The siderophore yersiniabactin binds copper to protect pathogens during infection Uropathogenic enterobacteria use the yersiniabactin metallophore system to acquire nickel Executive Summary: Management of Adults With Hospital-acquired and Ventilator-associated Pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the Bioengineered bacterial vesicles as biological nano-heaters for optoacoustic imaging Bacterial outer membrane vesicles suppress tumor by interferon-γmediated antitumor response Programmable bacteria induce durable tumor regression and systemic antitumor immunity Lipid A mutant Salmonella with suppressed virulence and TNFalpha induction retain tumor-targeting in vivo FDG PET of Infection and Inflammation Broad host range fluorescence and bioluminescence expression vectors for Gram-negative bacteria Ciprofloxacin Treatment Failure in a Murine Model of Pyelonephritis Due to an AAC(6′)-Ib-cr-Producing Escherichia coli Strain Susceptible to Ciprofloxacin In Vitro