Radiolabeled Cationic Peptides for Targeted Imaging of Infection

Molecular probes targeting bacteria provide opportunities to target bacterial infections in vivo for both imaging and therapy. In the current study, we report the development of positron emission tomography (PET) probes for imaging of live bacterial infection based on the small molecules HLys-DOTA, a polycationic peptide synthesized as the D-isomer (RYWVAWRNRG) conjugated to 1, 4, 7, 10-tetraazacyclododecane-N′,N″,N‴,N-tetraacetic acid (DOTA) and AB1-HLys-DOTA, which includes an unnatural amino acid AB1 that preferentially binds to bacteria membrane lipids with amine groups via formation of iminoboronates. HLys-DOTA and AB1-HLys-DOTA peptides were radiolabeled with 64Cu and investigated as PET imaging agents to track bacterial infection in vitro and in intramuscularly infected (IM) mice models. Cell uptake studies at 37°C in Staphylococcus aureus (SA) show higher uptake of 64Cu-AB1-HLys-DOTA; 98.47 ± 3.54% vs 64Cu-HLys-DOTA; 39.12 ± 3.27% at 24 h. Standard uptake values (SUV) analysis of the PET images resulted in mean SUV of 0.70 ± 0.08, 0.49 ± 0.04, and 0.31 ± 0.01 for 64Cu-AB1-HLys-DOTA and 0.17 ± 0.06, 0.16 ± 0.02, and 0.13 ± 0.01 for 64Cu-HLys-DOTA at 1, 4, and 24 h post injection, respectively, in the infected muscles. Similarly, in the biodistribution studies, dose uptake in the infected muscles was 4 times higher in the targeted 64Cu-AB1-HLys-DOTA group than in the 64Cu-HLys-DOTA group and 2‐3 times higher than in the PBS control group at 1, 4, and 24 h post injection. 64Cu-AB1-HLys-DOTA was able to distinguish between SA-infected muscle and Pseudomonas aeruginosa (PA) infected muscle with lower mean SUV of 0.28 ± 0.10 at 1 h post injection. This illustrates the utility of the AB1 covalently targeting group in synergy with the HLys peptide, which noncovalently binds to bacterial membranes. These results suggest that 64Cu-labeled AB1-HLys-DOTA peptide could be used as an imaging probe for detection of bacterial infection in vivo with specificity for Gram-positive bacteria.


Introduction
Infectious diseases are one of the top ten leading causes of death in the world and the number one cause of death in low-income developing countries. Current diagnostic methods for bacterial infections are time-consuming as they require samples from the patient to be cultured for hours to weeks before enough bacteria are isolated to run destructive diagnostic tests [1,2]. As proper identification of the disease often requires long diagnostic processes and bacterial infections tend to progress rapidly, fast diagnostic techniques would allow physicians to identify what types of antimicrobial therapy may help a patient in a clinically relevant timeframe. e ability to image bacterial disease would permit rapid and specific diagnosis of infection. In addition, the ability to image infections would allow clinicians to determine whether a therapy is effective and to monitor patient response to therapy.
Recently, there has been an increase in research into noninvasive imaging of bacterial infection using PET/CT, SPECT, MRI, and other imaging modalities in preclinical studies of infectious diseases using natural and synthetic molecules [3]. e use of radionuclides for detecting and localizing infection has been employed for decades [4,5]. Ideally, the radiopharmaceutical for detecting infection should be specific, sensitive, and clear from the body rapidly to facilitate early image acquisition and clear delineation of the infected area. ese radionuclides have been utilized as salts or small molecules such as 99m Tc-methylene diphosphonate, 67 Ga-citrate, and 18 F-FDG [6][7][8][9][10]. 111 Inoxine and 99m Tc-exametazime have also been used to label leukocytes for imaging in diagnosis of infection [7,11,12].
is approach, though valuable, has some drawbacks, requiring high levels of leukocytes, and detection of noninfectious inflammations and is thus not specific to the infection or bacteria [13,14].
Another radioisotope of interest is 68 Ga, with a shorter half-life of 68 min compared to its counterpart 67 Ga: half-life of 78.3 h (used in SPECT), has undergone an increased utilization in preclinical and clinical PET imaging particularly in oncology. Although, [ 68 Ga]citrate PET-CT of bone and joint infection offered lower radiation dose, as well as earlier and shorter imaging times, and its sensitivity and specificity was found to be lower than [ 67 Ga ]citrate [15]. In other infection and inflammation imaging studies, 68 Ga has been used to radiolabel ligands, small molecules, and peptides to target a host of receptors, inhibitors, leukocytes, and pathways [16]. 68 Ga-apo-transferrin ([ 68 Ga]TF) has been used to image Staphylococcus aureus (SA) and found to detect the infected lesions better than [ 68 Ga]Cl 3 ; additionally, [ 68 Ga]TF was also found to detect the Gram-negative bacteria, Proteus mirabilis [17]. In other studies, the antibiotic ciprofloxacin was labeled with 68 Ga and used to distinguish inflamed muscle from SA-infected muscle [18]. e use of 64 Cu radioisotope for imaging purposes has increased over the years, and several studies have utilized this isotope to label compounds for detecting regions of infection [19]. In order to label these compounds, several suitable chelators for 64 Cu have been developed with significant progress over the years; these chelates range from acyclic to cage-like bifunctional chelators [20,21]. 64 Cu, with its relatively long half-life (12.7 hours) compared to 18 F or 99m Tc has been used to label peptides, antibody fragments, and whole antibodies for imaging [22]. For instance, necrotic pulmonary tuberculosis lesions in chronically infected mice were detected and demonstrated to be hypoxic using 64 Cu(II)-diacetyl-bis(N 4 -methyl-thiosemicarbazone), [ 64 Cu]ATSM [23]. In mice models of S. aureus endocarditis (heart valve infection), robust localization of [ 64 Cu]DTPAprothrombin was used to noninvasively detect infection lesions as compared to the bacteria-free control mice, which had no accumulation at the site of endothelial trauma [24].
Radiolabeled molecular probes specifically targeting bacterial lipids provide opportunities to target bacterial infections in vivo for both imaging and therapy. In this study, synthetic peptides that preferentially bind to bacterial surfaces with amine-presenting lipids were radiolabeled and studied in vitro and in vivo. Two phospholipids abundant on Gram-positive bacteria, phosphatidylethanolamine (PE) and lysylphosphatidylglycerol (Lys-PG), were targeted for recognition by appropriately constructed short peptides used as low molecular weight probes [25]. For our studies, the D-isomer of a polycationic peptide HLys (RYW-VAWRNRG), which binds noncovalently to bacteria surfaces, was conjugated to 1, 4, 7, 10-tetraazacyclododecane-N′, N″, N‴, N-tetraacetic acid (DOTA) to make HLys-DOTA. To improve the binding of the peptide to the bacterial surface, an unnatural amino acid AB1, which preferentially binds covalently to lipids with amine groups via formation of iminoboronates under physiological conditions, was conjugated to HLys-DOTA, as shown in Figure 1 [26]. By targeting PE and Lys-PG with AB1, the iminoboronate chemistry allows potent labeling of Gram-positive bacteria such as Staphylococcus aureus (SA) even in the presence of serum proteins, while bypassing mammalian cells and Gram-negative bacteria [26]. HLys-DOTA and AB1-HLys-DOTA peptide conjugates were radiolabeled with 64 Cu and studied as PET imaging agents to track bacterial infection in vitro and in intramuscularly infected (IM) mice models.

Materials and Methods
64 Cu was produced in-house on an ACSI TR-24 Cyclotron (Richmond, Canada) located in the UAB Cyclotron facility. Reverse-phase chromatography analysis was performed on a C18 column (PA) with an infinity diode-array UV detector (Agilent, Lake Forest, CA) and a PMT/NaI remote radioactive detector (LabLogic Systems Ltd, Brandon, FL). Laura radiochromatography software (LabLogic Systems Ltd, Brandon, FL) was used to quantify chromatograms by integration. Radioactive samples were counted using a 2480 Wizard II automatic gamma counter (PerkinElmer, Downers Grove, IL). Imaging studies were done by using

Synthesis and Structural
Analysis of HLys-DOTA and AB1-HLys-DOTA Peptides. HLys-DOTA and AB1-HLys-DOTA were synthesized by following a previously described protocol for solid-phase peptide synthesis [26], and the DOTA moiety was conjugated to the N-terminus of peptides on resin using DOTA-NHS-ester as a precursor [27]. To ensure in vivo stability of the peptides, the D-isomer of all amino acids except AB1 was used for peptide synthesis. Literature reports have shown that the D-isomers of antimicrobial peptides behave similarly to their L-counterparts in bacterial binding [28]. e peptides were purified using HPLC, and their purity and integrity was confirmed by LC-MS (Supporting Information Figure 1).
e two peptides, HLys-DOTA and AB1-HLys-DOTA, were radiolabeled with 64 CuCl 2 buffered in 0.1 M NH 4 OAc, pH 6, at 56°C according to modified previously published methods [31,32]. Briefly, radiolabeling of the peptides was achieved by adding 50 μg of HLys-DOTA or AB1-HLys-DOTA into 120-130 MBq (3.2-3.5 mCi) of 64 CuCl 2 in 1450 μL of 0.1 M NH 4 OAc, pH 6. e reactions were incubated on a mixer with 800 rpm agitation at 56°C for 45 min, and radiolabeling was also attempted at lower temperature. Radiolabeling yield and radiochemical purity were assessed using silica gel plates developed in 50 : 50 of methanol: 1 M ammonium acetate and high-performance liquid chromatography, respectively, with gradient: 0-12 min: 95% A to 20% A (A � water, 0.1% TFA, B � acetonitrile, 0.1% TFA), 12-15 min: 20% A to 95% A. 64 Cu-AB1-HLys-DOTA was purified according to the Jacobson method [33]. A SepPak C18-cartridge was activated with 5 mL of ethanol and 10 mL of water. e reaction mixture of 64 Cu-AB1-HLys-DOTA was diluted with 5 mL of water and loaded slowly onto the activated SepPak C18cartridge using a syringe. e cartridge was washed with 10 mL of water followed by elution of the desired labeled peptide with 1 mL of 10 mM HCl in ethanol. e ethanol was evaporated, and the radiolabeled peptide was reformulated with saline.

In Vitro Studies
2.5.1. Bacterial Cell Culture. Bacterial in vitro uptake and imaging experiments were performed using Gram-positive bacteria, Staphylococcus aureus (SA), and Gram-negative bacteria, Pseudomonas aeruginosa (PA), as control. Bacterial cells from a single colony were grown overnight in LB broth at 37°C with agitation until the cells reached the mid logarithmic phase (OD 600 ∼0.6-0.7). For the in vitro experiments, 750-1000 μL of the bacterial cell culture was spun down at 7000 rpm for 7 min in a 1.5 mL centrifuge tube.

Uptake Studies of 64 Cu-Labeled Peptides in Staphylococcus aureus (SA) and Pseudomonas aeruginosa (PA).
Bacteria cell uptake studies of 64 Cu-HLys-DOTA and 64 Cu-AB1-HLys-DOTA in SA and PA bacteria were performed at 37°C. LB broth was inoculated with SA or PA bacteria and after 18-22 hours, cells were harvested by spinning down and washing twice in 750-1000 μL of PBS (50 mM sodium phosphate, pH � 7.4), and the OD 600 , OD 230 , and OD 260 determined. 750 μL of the PBS suspended bacteria (0.9-1.3 × 10 7 CFU/mL) was incubated with 0.925-0.999 MBq (25-27 μCi) of 64 Cu-HLys-DOTA and 64 Cu-AB1-HLys-DOTA at 37°C in triplicates at different time points: 5, 40, 60, 240, and 1440 min. At each time point, the mixture was spun down, washed twice with PBS, and the final bacterial pellets were measured using a gamma counter to determine the percent of associated radiolabeled peptide. Mammalian cell uptake was verified by following a similar procedure in a nonbacterial cell, SKBR3 breast cancer cell line.

Small Animal Imaging and Biodistribution Studies.
All animal experiments were performed according to animal use protocols approved by the University of Alabama at Birmingham Institutional Animal Care and Use Committee (IACUC). Animals were housed under controlled conditions with a natural light-dark cycle. ey were allowed to adapt to the housing environment for at least 48 h prior to study. Biodistribution and PET imaging studies with 64 Cu-HLys-DOTA and 64 Cu-AB1-HLys-DOTA were conducted in mice intramuscularly (IM) infected in the thigh muscle with SA and PBS as a control. irty-six male mice (25-30 g, 4-5 weeks old) infected with SA were used for animal studies. For each peptide, 4 mice were used for small animal PET imaging at 1, 4, and 24 h, and then sacrificed at 24 h for a post-PET biodistribution. Another set of 4 mice each was used for biodistribution studies at 1 and 4 h. A similar set of mice was used for blocking studies with 64 Cu-AB1-HLys-DOTA peptide. In order to check the specificity of 64 Cu-AB1-HLys-DOTA for Gram-positive bacteria, imaging and post-PET biodistribution studies were conducted in another set of mice (four) intramuscularly (IM) infected in the thigh muscle with PA and PBS.

Infection Animal Model.
Mice for infection studies were anesthetized prior to the intramuscular infection of SA or PA. Animals were immobilized with isoflurane. Bacteria cells concentrated in sterile PBS were injected into the thigh muscle and the site of injection pinched to minimize bleeding. SA or PA suspension (50 μL, 2.6-2.8 × 10 7 CFU per mouse) was intramuscularly injected into the right thigh muscle, and sterile PBS (50 μL) was injected into the left thigh muscle as control. Mice were placed back into the cages and allowed to recover with frequent monitoring to identify any sign of extreme distress. e infection was allowed to develop for 3-5 days before intravenous injection of the radiolabeled probe. e muscles on the left and right thighs were harvested, homogenized, and plated on mannitol agar plates or cultured in LB broth to verify the presence of infection. In addition, [ 18 F]FDG imaging was also performed in infected animal models as described in the PET imaging section below to further confirm the existence of infection in the right thigh and the absence thereof in the left thigh. In a similar set of experiment, excess of the unlabeled peptide, 150 nmol, was coinjected with the radiolabeled peptides per mouse. e organs of interest were collected, weighed, and measured for radioactivity content using a gamma counter. e data were background-corrected and decay-corrected. e biodistribution data were expressed as percentage of injected radioactive dose per gram of tissue (%ID/g) for selected organs as the mean value of four mice. Regions of interest (ROI) were manually drawn over organs of interest with CT anatomical guidelines, and the associated radioactivity was measured using Inveon Research Workstation software. Standard uptake values (SUV) were calculated as nCi/cc × animal weight/injected dose, and comparisons in pharmacokinetics of radiolabeled peptides were assessed.

Statistical Analysis.
Statistical calculations were carried out using Prism 7 (GraphPad Software) and expressed as mean ± SD. One-way analysis of standard deviation at 95% confidence level (p < 0.05) were considered statistically significant.

Stability and In Vitro Cell Uptake Studies.
Excellent stability of 64 Cu-AB1-HLys-DOTA was observed in the PBS, human serum, and LB broth with >98% intact peptide observed at 24 h post incubation in all solutions, as shown in Figure 2(a). Bacteria cell uptake of 64 Cu-AB1-HLys-DOTA conjugate at 37°C in Staphylococcus aureus (SA) was significantly higher than that of 64 Cu-HLys-DOTA at all time points: 0.08, 0.67, 1, 4, and 24 h; 98.5 ± 3.5% vs 39.1 ± 3.3% at 24 h, Figure 2(b). Specific uptake in Gram-positive bacteria SA was confirmed by the lower uptake of the peptides in Pseudomonas aeruginosa (PA), a Gram-negative bacterium. At 5 min after incubation, the uptake of 64 Cu-AB1-HLys-DOTA and 64 Cu-HLys-DOTA showed 9.8 ± 4.4% and 1.9 ± 0.4% binding to PA, respectively, which increased slightly to 13.3 ± 1.1% and 5.2 ± 0.3%, respectively, at 24 h after incubation. is indicates a 5-10 times lower uptake of the peptides in PA as compared to SA confirming the specificity of AB1 for Gram-positive bacteria. Specific bacterial cell uptake was confirmed by the low uptake (9.4 ± 2.3%) of 64 Cu-AB1-HLys-DOTA in mammalian SKBR3 breast cancer cells at 24 h post incubation.

Confirmation of Infection in Animal Model.
e development of infection was evident by presence of a palpable mass filled with abscess observed during dissection and harvesting. e infected muscle and control muscle were harvested, homogenized, and the CFU of bacteria counted. In the infected muscle, 4.8 ± 2.5 × 10 6 cfu/g of bacteria was found and only one

Biodistribution Studies.
e biodistribution and pharmacokinetics of the peptides 64 Cu-AB1-HLys-DOTA and 64 Cu-HLys-DOTA were assessed in mice intramuscularly infected with S. aureus on the right thigh muscle and PBS injected in the left thigh muscle as control.
e biodistribution showed that the uptake of both peptides in the infected muscle was significantly higher than the PBS-injected control muscle of the same mouse at all time points: 1, 4, and 24 h post injection, as shown in Figure 3. e uptake of 64 Cu-AB1-HLys-DOTA conjugate in the infected thigh was 2-3 times higher than the control muscle at later time points, although no significant difference was observed at 1 h (4.8 ± 2.5 %ID/g vs. 2.5 ± 0.5 %ID/g). At 4 h, the dose uptake of 64 Cu-AB1-HLys-DOTA in the infected muscle was 3.6 ± 0.5 %ID/g vs. 1.1 ± 0.4 %ID/g in the control muscle, p < 0.0005. Similarly, at 24 h, the uptake of 64 Cu-AB1-HLys-DOTA in the infected muscle was 2.6 ± 0.6 %ID/g vs. 0.8 ± 0.1 %ID/g in the control muscle, p < 0.005. In contrast, the dose accumulation of 64 Cu-HLys-DOTA in the infected muscle was only 1-or 2-fold higher than the control muscle in the same mouse, except at 24 h where no significant difference was observed between them. At 1 and 4 h, the uptake of 64 Cu-HLys-DOTA in the infected muscle vs. control muscle was 1.3 ± 0.5 %ID/g vs. 0.6 ± 0.1 %ID/g, p < 0.05, and 0.8 ± 0.1 %ID/g vs. 0.4 ± 0.1 %ID/g, p < 0.005, respectively. We also observed that uptake in the infected muscle was 4 times higher in the targeted 64 Cu-AB1-HLys-DOTA group compared to that in the 64 Cu-HLys-DOTA group, as shown in Figure 3(c). e uptake of 64 Cu-AB1-HLys-DOTA in the infected muscle was 4.8 ± 2.5, 3.6 ± 0.5, and 2.6 ± 0.6 %ID/g vs. 1.3 ± 0.5, 0.8 ± 0.1, and 0.6 ± 0.1 %ID/ g for 64 Cu-HLys-DOTA, respectively, at 1, 4, and 24 h post injection, p < 0.005 at 4 and 24 h. e highest accumulation of the radiolabeled peptides was observed in the liver and kidney with 64 Cu-AB1-HLys-DOTA also showing some uptake in the lungs. 64 Cu-AB1-HLys-DOTA also showed higher retention in the blood and heart than 64 Cu-HLys-DOTA at 1 and 4 h post injection but at 24 h, both peptides had similar activity in the blood. A similar experiment using an excess of the unlabeled peptides, 150 nmol, did not show a statistical difference (data not shown) in the uptake of the peptides at the infection site.
is may indicate that the binding sites at the bacterial surface cannot be saturated.

PET/CT Imaging.
e uptake of 64 Cu-AB1-HLys-DOTA in the infected muscle was visibly observed via small animal PET/CT imaging as early as 1 and 4 h post injection,  at 2.64 ± 0.57 %ID/g vs. 0.49 ± 0.11 %ID/g in the PA-infected muscle, p � 0.0039.
[ 18 F]FDG uptake in the SA-infected muscle was significantly higher than that in the control muscle at 4 h post injection with mean SUV of 1.96 ± 0.19 vs. 0.41 ± 0.09, p < 0.0001, which confirmed the presence of the SA infection [34][35][36], as shown in Figures 5(a) and 5(b).

Discussion
Antimicrobial peptides produced by phagocytes and other cell types in the body are part of the innate immunity systems against infection resulting from pathogens. e cationic domains of these peptides interact with the negatively charged surface of the microorganisms eliciting an antimicrobial reaction when the peptide is inserted into the microbial membranes [37]. Antimicrobial peptides, due to their diversity, elicit antimicrobial activity through other mechanisms such as the Shai-Matsuzaki-Huang, albeit at micromolar concentration [38]. ese peptides and other synthetic peptide can be radiolabeled with specificity for detecting localized infections using a variety of chemistries [39]. In this study, by targeting the membrane lipids enriched in bacterial cells, namely PE and Lys-PG, the iminoboronate chemistry allows selective labeling of bacterial cells over mammalian cells [26]. Radiolabeling of AB1-HLys-DOTA and HLys-DOTA peptides with 64 Cu allowed monitoring of the infection up to 24 h and observation of their biological clearance. is was evident in the superior uptake of 64 Cu-AB1-HLys-DOTA conjugate in Staphylococcus aureus bacteria in comparison to 64 Cu-HLys-DOTA. Specific bacterial cell uptake of 64 Cu-AB1-HLys-DOTA was also confirmed by the low uptake (less than 10%) in mammalian SKBR3 breast cancer cells, while specificity in Gram-positive bacteria was confirmed with lower uptake in Pseudomonas aeruginosa. Previous studies have shown specific uptake of 68 Ga-radiolabeled probes at sites of infection [17]. Similarly, in this study, the biodistribution numbers show that 64 Cu-AB1-HLys-DOTA has higher affinity to the infection and persists within the infection longer than 64 Cu-HLys-DOTA that lacks the AB1 group.
e difference in accumulation of the peptides shows that the noncovalent attraction of HLys for bacterial cell surface was greatly enhanced by the covalent binding of AB1 to the lipids on the surface. In the biodistribution and PET images, clear uptake of 64 Cu-AB1-HLys-DOTA was observed at 1 h post injection, which may be due in part to its higher bioavailability (higher %ID/g in the blood) and slower clearance. 64 Cu-AB1-HLys-DOTA was able to distinguish between Gram-positive bacterial infection and Gram-negative bacterial infection. is class of peptides offers selective synthetic targets for bacterial lipids, which may give rise to new imaging methods of bacterial infection [26]. Although we are yet to ascertain its utility in distinguishing infection from sterile inflammation, but as reported by Sellmyer et al., it is possible to distinguish infection from background and other noninfection inflammation sites. 18 Flabeled small-molecule antibiotic trimethoprim [ 18 F] FPTMP showed high uptake at the infection and low background signal in normal tissues and other noninfection inflammation sites but did not differentiate between Gram-negative and Gram-positive strains [40]. us, future effort in work would focus on using the peptide to distinguish sterile inflammation from bacterial infection. Additionally, optimizing the peptides with functional groups that can modulate their clearance rates and therefore increase uptake at the infection sites would be implemented in future work.

Conclusion
e purpose of this study was to investigate the potential of AB1-HLys-DOTA peptide to image and distinguish infection due to Gram-positive bacteria from Gram-negative bacteria with higher specificity than HLys-DOTA peptide. 64 Cu-AB1-HLys-DOTA showed higher uptake in S. aureus bacteria cells in vitro and improved accumulation at the infection site of SA-inoculated mice compared with the noncovalently targeting 64 Cu-HLys-DOTA. In the smallanimal PET images, the dose uptake of 64 Cu-AB1-HLys-DOTA at the infected site was distinguishable as early as 1 h after administration, indicating its potential for fast detection of infection. ese results illustrate that the 64 Culabeled AB1-HLys-DOTA peptide could be used as imaging probe for detection of bacterial infection in vivo with specificity for Gram-positive bacterial infection.
Data Availability e cell uptake, biodistribution, and small animal PET/CT data used to support the findings of this study are included within the article.