Info
Background:
Early and accurate identification of bacterial infection remains a major clinical challenge, as current imaging modalities such as CT, MRI, and [18F]FDG PET primarily detect anatomic changes or the host inflammatory response that often occur late in the disease process. Over the past decade, numerous imaging strategies have taken advantage of bacteria-specific metabolic pathways. However there is no stand-alone imaging agent for routine differentiation of infection from sterile inflammation in clinical practice [1].

 

Hypothesis/Objective:
Unlike mammalian cells, many bacteria utilize extracellular gluconate through gluconate permease, followed by phosphorylation via gluconate kinase [2]. By leveraging unique bacterial gluconate metabolism, we hypothesized that a gluconate-derived radiotracer may enable pathogen-specific PET imaging with low off-target signals. Meanwhile, glucose oxidase (GOx) is a widely available enzyme that catalyzes the oxidation of glucose to gluconolactone, which is then spontaneously hydrolyzed to gluconic acid. Interestingly, previous studies indicate that GOx can recognize 2-deoxy-D-glucose as a substrate, even though its specificity is lower than that of D-glucose (25-30%) [3]. We therefore investigated a GOx-mediated transformation of [18F]FDG to 2-deoxy-2-fluoro-D-gluconic acid ([18F]FGA) for pathogen-specific PET imaging.

 

Methods:
Chemoenzymatic radiosynthesis of [18F]FGA was performed by oxidation of clinical [18F]FDG using glucose oxidase (GOx) at 35°C for 30 min. To assess the bacterial specificity, [18F]FGA was incubated with non-radioactive FGA or gluconic acid (1mM) in S. aureus. In vitro assays were conducted in diverse bacterial species to determine the scope of bacterial uptake [4]. Metabolic specificity was validated through comparative in vitro and in vivo studies using S. aureus transposon mutants (ΔgntP and ΔgntK) and their parental JE2 strain [5]. The maltose analog [18F]FDM, which exploits maltose/maltodextrin transport system, was used as a control to support metabolic specificity of [18F]FGA for bacterial gluconate metabolism. List-mode data were acquired to obtain biodistribution and excretion profiles of [18F]FGA in uninfected mice. Bacterial specificity and selectivity in vivo were evaluated using murine models of myositis (S. aureus, E. coli, and LPS) and K. pneumoniae pulmonary infection [6,7].

 

Results:
The chemoenzymatic radiosynthesis of [18F]FGA was simple, only requiring readily available [18F]FDG and glucose oxidase. With a total synthesis time of 1 h, [18F]FGA was synthesized in 42 ± 6% decay corrected radiochemical yield (RCY, n = 16). [18F]FGA showed broad detection sensitivity across multiple bacterial species, including Staphylococcus and Enterobacteriaceae. [18F]FGA uptake in S. aureus in vitro was markedly inhibited by co-incubation with either 1 mM of gluconic acid or a nonradioactive FGA, suggest that [18F]FGA uptake is specific and may be closely related to bacterial gluconate metabolism. We further investigated the specificity of [18F]FGA for bacterial gluconate metabolism. S. aureus transposon mutant studies showed substantial loss of [18F]FGA uptake in ΔgntP and ΔgntK mutants, supporting high specificity for bacterial gluconate metabolism. [18F]FGA PET showed rapid clearance and low off-target background signals in uninfected mice. We further validated the specificity of [18F]FGA PET between infection and inflammation in murine models of myositis. [18F]FGA PET selectively highlighted live bacterial infection, with high target-to-nontarget ratios in S. aureus (36-fold) and E. coli (30-fold) in a murine model of myositis. Unlike [18F]FDG, no appreciable uptake of [18F]FGA was detected at sites of sterile inflammation (LPS-induced). In K. pneumoniae infected mice, significantly higher [18F]FGA accumulation was observed in the infected lung (5.2 ± 0.8 %IA/cc) compared to the uninfected controls (0.1 ± 0.0 %IA/cc).

 

Conclusion:
By leveraging bacterial gluconate metabolism, we demonstrated the translational potential of [18F]FGA as a pathogen-specific PET tracer. The synthetic ease of glucose oxidase-mediated [18F]FDG transformation, together with the rapid cartridge-based formulation, may enable on-demand production of [18F]FGA in acute care settings. Studies in clinically relevant musculoskeletal infection models are currently underway to further validate translational potential of [18F]FGA PET in clinical practice.

 

Image/Figure:
Click to view full size
Click to view full size

Image/Figure Caption:
[18F]FGA is readily available via simple chemoenzymatic radiosynthesis using [18F]FDG and GOx, enabling rapid-clearance and high-contrast imaging of live bacterial infection over sterile inflammation.

 

Author
Sanghee Lee
University of California, San Francisco